Thijs/wiki.techinc.nl
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions
wiki.techinc.nl/includes/OutputPage.php

4619 lines
140 KiB
PHP
Raw Normal View History

Fix for compatibility with short_open_tag = Off
2004-02-18 02:15:00 +00:00
<?php
Added missing GPLv2 headers in some places. Also made file/class documentation more consistent. Change-Id: Ia45d0756cd84138da267bf1cbd1a58be7f28c6d1
2012-05-11 08:34:29 +00:00
/**
* Preparation for the final page rendering.
*
* This program is free software; you can redistribute it and/or modify
* it under the terms of the GNU General Public License as published by
* the Free Software Foundation; either version 2 of the License, or
* (at your option) any later version.
*
* This program is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
* GNU General Public License for more details.
*
* You should have received a copy of the GNU General Public License along
* with this program; if not, write to the Free Software Foundation, Inc.,
* 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
* http://www.gnu.org/copyleft/gpl.html
*
* @file
*/
Hooks::run() call site migration Migrate all callers of Hooks::run() to use the new HookContainer/HookRunner system. General principles: * Use DI if it is already used. We're not changing the way state is managed in this patch. * HookContainer is always injected, not HookRunner. HookContainer is a service, it's a more generic interface, it is the only thing that provides isRegistered() which is needed in some cases, and a HookRunner can be efficiently constructed from it (confirmed by benchmark). Because HookContainer is needed for object construction, it is also needed by all factories. * "Ask your friendly local base class". Big hierarchies like SpecialPage and ApiBase have getHookContainer() and getHookRunner() methods in the base class, and classes that extend that base class are not expected to know or care where the base class gets its HookContainer from. * ProtectedHookAccessorTrait provides protected getHookContainer() and getHookRunner() methods, getting them from the global service container. The point of this is to ease migration to DI by ensuring that call sites ask their local friendly base class rather than getting a HookRunner from the service container directly. * Private $this->hookRunner. In some smaller classes where accessor methods did not seem warranted, there is a private HookRunner property which is accessed directly. Very rarely (two cases), there is a protected property, for consistency with code that conventionally assumes protected=private, but in cases where the class might actually be overridden, a protected accessor is preferred over a protected property. * The last resort: Hooks::runner(). Mostly for static, file-scope and global code. In a few cases it was used for objects with broken construction schemes, out of horror or laziness. Constructors with new required arguments: * AuthManager * BadFileLookup * BlockManager * ClassicInterwikiLookup * ContentHandlerFactory * ContentSecurityPolicy * DefaultOptionsManager * DerivedPageDataUpdater * FullSearchResultWidget * HtmlCacheUpdater * LanguageFactory * LanguageNameUtils * LinkRenderer * LinkRendererFactory * LocalisationCache * MagicWordFactory * MessageCache * NamespaceInfo * PageEditStash * PageHandlerFactory * PageUpdater * ParserFactory * PermissionManager * RevisionStore * RevisionStoreFactory * SearchEngineConfig * SearchEngineFactory * SearchFormWidget * SearchNearMatcher * SessionBackend * SpecialPageFactory * UserNameUtils * UserOptionsManager * WatchedItemQueryService * WatchedItemStore Constructors with new optional arguments: * DefaultPreferencesFactory * Language * LinkHolderArray * MovePage * Parser * ParserCache * PasswordReset * Router setHookContainer() now required after construction: * AuthenticationProvider * ResourceLoaderModule * SearchEngine Change-Id: Id442b0dbe43aba84bd5cf801d86dedc768b082c7
2020-03-19 02:42:09 +00:00
use MediaWiki\HookContainer\ProtectedHookAccessorTrait;
Reorg: Move HTML-related classes out of includes/ to Html/ Bug: T321882 Change-Id: I5dc1f7e9c303cd3f5b9dd7010d6bb470d8400a18
2023-02-16 19:27:21 +00:00
use MediaWiki\Html\Html;
Generalize ResourceLoader 'excludepage' functionality There has long been a hack for previewing edits to user JS/CSS, where OutputPage would pass an 'excludepage' parameter to ResourceLoaderUserModule to tell it not to load one particular page and would instead embed that page statically. That's nice, but there are other places where we could use the same thing. This patch generalizes it: * DerivativeResourceLoaderContext may now contain a callback for mapping titles to replacement Content objects. * ResourceLoaderWikiModule::getContent() uses the overrides, and requests embedding when they're used. All subclasses in Gerrit should pick it up automatically. * OutputPage gains methods for callers to add to the override mapping, which it passes on to RL. It loses a bunch of the special casing it had for the 'user' and 'user.styles' modules. * EditPage sets the overrides on OutputPage when doing the preview, as does ApiParse for prop=headhtml. TemplateSandbox does too in I83fa0856. * OutputPage::userCanPreview() gets less specific to editing user CSS and JS, since RL now handles the embedding based on the actual modules' dependencies and EditPage only requests it on preview. ApiParse also gets a new hook to support TemplateSandbox's API integration (used in I83fa0856). Bug: T112474 Change-Id: Ib9d2ce42931c1de8372e231314a1f672d7e2ac0e
2017-02-28 20:52:17 +00:00
use MediaWiki\Linker\LinkTarget;
Use name constants to access config settings. Use name constants instead of string literals in calls to Config::get and ServiceOptions::get, when referring to core configuration variables. This protects against typos and makes the decumentation and schema declaration of the config settings discoverable. This is the first batch, only touching files directly under /includes/ Change-Id: I7252e636c7c86d950d9257b33491af492c6dd5eb
2022-04-01 15:58:32 +00:00
use MediaWiki\MainConfigNames;
Replace Linker::link() with LinkRenderer in includes directory * CategoryViewer * OutputPage * Preferences Bug: T149346 Change-Id: I9f9bcd9b461884817e8ceefbc6757c436221e331
2016-11-30 21:28:55 +00:00
use MediaWiki\MediaWikiServices;
Remove Title from public interface of OutputPage Bug: T278459 Change-Id: If455682540faf8d42fd8a14e0ef09bac24f72db4
2021-04-08 19:11:06 +00:00
use MediaWiki\Page\PageRecord;
use MediaWiki\Page\PageReference;
Use ParserOutputFlags::SHOW_TOC in OutputPage Parser output flags are intended to be "set only". That means that extensions can *set* the SHOW_TOC flag, but once set SHOW_TOC can't be reset, and the TOC will be shown in the result if *any* of the merged ParserOutputs on the OutputPage contain the SHOW_TOC flag set. That's consistent with prior merge behavior. There's a possibility extensions are playing other games with the Table Of Contents (TOC) in the OutputPageParserOutput hook; be compatible with prior behavior for now but we may need to provide additional functionality (perhaps a way to merge new section data) if needed. Bug: T310083 Change-Id: I0909ac85c6c785d9089b077a16923c61d6a09996
2022-06-09 12:59:50 +00:00
use MediaWiki\Parser\ParserOutputFlags;
Convert second batch of special pages to authority Depends-On: Ifc82d29a00d3fd136c44e0699e6bbfa11b8cf2a6 Change-Id: Iaf559a72cc47ec5d2481d7f740ad0eb303c94002
2021-03-04 20:16:50 +00:00
use MediaWiki\Permissions\PermissionStatus;
Reorg: Move CSP class to Request/ Bug: T321882 Change-Id: Ic809656a31e08626d0e5a1172da02b2397096e14
2023-02-16 12:36:41 +00:00
use MediaWiki\Request\ContentSecurityPolicy;
Reorg: Move some of request related classes to MediaWiki/Request Redoing I5ea70120d74 but without moving WebRequest that caused issues with phan-taint-plugin. Moving: - DerivativeRequest - FauxRequest - FauxRequestUpload - PathRouter - WebRequestUpload Bug: T321882 Change-Id: I832b133aaf61ee9f6190b0227d2f3de99bd1717b
2022-10-28 10:04:25 +00:00
use MediaWiki\Request\FauxRequest;
ResourceLoader namespace (attempt 2) Move ResourceLoader classes to their own namespace. Strip the "ResourceLoader" prefix from all except ResourceLoader itself. Move the tests by analogy. I used a namespace alias "RL" in some callers since RL\Module is less ambiguous at the call site than just "Module". I did not address DependencyStore which continues to have a non-standard location and namespace. Revert of a241d83e0a6dabedf. Bug: T308718 Change-Id: Id08a220e1d6085e2b33f3f6c9d0e3935a4204659
2022-05-06 09:09:56 +00:00
use MediaWiki\ResourceLoader as RL;
use MediaWiki\ResourceLoader\ResourceLoader;
Revert "Remove SessionManager, temporarily" This reverts commit 823db5d63dd5200d04c63da50ba6bf16f928e70b. Change-Id: Ibb3e023e4eb6715295586dea87d0725c344a8271
2016-02-01 20:44:03 +00:00
use MediaWiki\Session\SessionManager;
Reorg: Namespace the Title class This is moderately messy. Process was principally: * xargs rg --files-with-matches '^use Title;' | grep 'php$' | \ xargs -P 1 -n 1 sed -i -z 's/use Title;/use MediaWiki\\Title\\Title;/1' * rg --files-without-match 'MediaWiki\\Title\\Title;' . | grep 'php$' | \ xargs rg --files-with-matches 'Title\b' | \ xargs -P 1 -n 1 sed -i -z 's/\nuse /\nuse MediaWiki\\Title\\Title;\nuse /1' * composer fix Then manual fix-ups for a few files that don't have any use statements. Bug: T166010 Follows-Up: Ia5d8cb759dc3bc9e9bbe217d0fb109e2f8c4101a Change-Id: If8fc9d0d95fc1a114021e282a706fc3e7da3524b
2023-03-01 20:33:26 +00:00
use MediaWiki\Title\Title;
Replace some more usages of Wikimedia\(suppress|restore)Warnings() Change-Id: I2eb133a9e32116cd155f59086245bc4d15ecbfcc
2022-02-24 20:17:53 +00:00
use Wikimedia\AtEase\AtEase;
Generate/set/get TOCData/SectionMetadata objects instead of arrays * ParserOutput::setSections()/::getSections() are expected to be deprecated. Uses in extensions and skins will need to be migrated in follow up patches once the new interface has stabilized. * In the skins code, the metadata is converted back to an array. Downstream skin TOC consumers will need to be migrated as well before we can remove the toLegacy() conversion. * Fixed SerializationTestTrait's validation method - Not sure if this is overkill but should handle all future complex objects we might stuff into the ParserCache. * This patch emits a backward-compatible Sections property in order to avoid changing the parser cache serialization format. T327439 has been filed to eventually use the JsonCodec support for object serialization, but for this initial patch it makes sense to avoid the need for a concurrent ParserCache format migration by using a backward-compatible serialization. * TOCData is nullable because the intent is that ParserOutput::setTOCData() is MW_MERGE_STRATEGY_WRITE_ONCE; that is, only the top-level fragment composing a page will set the TOCData. This will be enforced in the future via wfDeprecated() (T327429), but again our first patch is as backward-compatible as possible. Bug: T296025 Depends-On: I1b267d23cf49d147c5379b914531303744481b68 Co-Authored-By: C. Scott Ananian <cananian@wikimedia.org> Co-Authored-By: Subramanya Sastry <ssastry@wikimedia.org> Change-Id: I8329864535f0b1dd5f9163868a08d6cb1ffcb78f
2022-09-01 23:07:29 +00:00
use Wikimedia\Parsoid\Core\TOCData;
Add missing use statements Swap some ResultWrapper/Database to interface IResultWrapper/IDatabase Change-Id: I312cb3cca5742f72efdb0965ba35362f2eb1129d
2018-06-27 18:49:23 +00:00
use Wikimedia\Rdbms\IResultWrapper;
Update RelPath calls Bug: T182273 Change-Id: Ifd04e3b3ab72cc56bef48a1fc18fccb34b0af618
2018-01-21 04:33:38 +00:00
use Wikimedia\RelPath;
Update WrappedString use statements Bug: T182273 Change-Id: Ia9515863491fa9c8684698fa33612eec5eb6d937
2018-01-21 04:39:17 +00:00
use Wikimedia\WrappedString;
use Wikimedia\WrappedStringList;
resourceloader: Make ResourceLoader logger aware Change-Id: Ifa197cce1906bc0530dfa873dfde91abbe540637
2015-06-04 03:52:45 +00:00
Changing comments layout preparing for generated documentation with Phpdocumentor
2004-09-02 23:28:24 +00:00
/**
OutputPage: Change outdated comment to point to mediawiki.org page. Comment had outdated TODO for documentation which has since been created. Change-Id: Id0df26c979a6d1bad84e456a923721688d51617f
2019-06-28 14:54:40 +00:00
* This is one of the Core classes and should
* be read at least once by any new developers. Also documented at
* https://www.mediawiki.org/wiki/Manual:Architectural_modules/OutputPage
OutputPage properties documentation This is following a pair session with Zak Greant on november 12th 2010. Please review, ammend and commit your changes if you have anything useful to add. Please note you will find some humoristics quotes.
2011-01-13 21:08:59 +00:00
*
* This class is used to prepare the final rendering. A skin is then
* applied to the output parameters (links, javascript, html, categories ...).
random w/s cleanup
2011-04-23 19:28:35 +00:00
*
Use consistent notation for "@todo FIXME". Should update http://svn.wikimedia.org/doc/todo.html nicely.
2011-05-17 22:03:20 +00:00
* @todo FIXME: Another class handles sending the whole page to the client.
random w/s cleanup
2011-04-23 19:28:35 +00:00
*
using real identity for hashar Converting my pseudonym to use my real identity instead: Ashar Voultoiz -> Antoine Musso
2011-10-24 09:08:13 +00:00
* Some comments comes from a pairing session between Zak Greant and Antoine Musso
Minor: removed some typos in comment lines.
2011-03-26 15:18:16 +00:00
* in November 2010.
OutputPage properties documentation This is following a pair session with Zak Greant on november 12th 2010. Please review, ammend and commit your changes if you have anything useful to add. Please note you will find some humoristics quotes.
2011-01-13 21:08:59 +00:00
*
Changing comments layout preparing for generated documentation with Phpdocumentor
2004-09-02 23:28:24 +00:00
* @todo document
*/
* Made OutputPage extend ContextSource instead of duplicating its code; this also adds getLang() that was missing * Use getLang() instead of $wgLang
2011-07-30 19:37:19 +00:00
class OutputPage extends ContextSource {
Hooks::run() call site migration Migrate all callers of Hooks::run() to use the new HookContainer/HookRunner system. General principles: * Use DI if it is already used. We're not changing the way state is managed in this patch. * HookContainer is always injected, not HookRunner. HookContainer is a service, it's a more generic interface, it is the only thing that provides isRegistered() which is needed in some cases, and a HookRunner can be efficiently constructed from it (confirmed by benchmark). Because HookContainer is needed for object construction, it is also needed by all factories. * "Ask your friendly local base class". Big hierarchies like SpecialPage and ApiBase have getHookContainer() and getHookRunner() methods in the base class, and classes that extend that base class are not expected to know or care where the base class gets its HookContainer from. * ProtectedHookAccessorTrait provides protected getHookContainer() and getHookRunner() methods, getting them from the global service container. The point of this is to ease migration to DI by ensuring that call sites ask their local friendly base class rather than getting a HookRunner from the service container directly. * Private $this->hookRunner. In some smaller classes where accessor methods did not seem warranted, there is a private HookRunner property which is accessed directly. Very rarely (two cases), there is a protected property, for consistency with code that conventionally assumes protected=private, but in cases where the class might actually be overridden, a protected accessor is preferred over a protected property. * The last resort: Hooks::runner(). Mostly for static, file-scope and global code. In a few cases it was used for objects with broken construction schemes, out of horror or laziness. Constructors with new required arguments: * AuthManager * BadFileLookup * BlockManager * ClassicInterwikiLookup * ContentHandlerFactory * ContentSecurityPolicy * DefaultOptionsManager * DerivedPageDataUpdater * FullSearchResultWidget * HtmlCacheUpdater * LanguageFactory * LanguageNameUtils * LinkRenderer * LinkRendererFactory * LocalisationCache * MagicWordFactory * MessageCache * NamespaceInfo * PageEditStash * PageHandlerFactory * PageUpdater * ParserFactory * PermissionManager * RevisionStore * RevisionStoreFactory * SearchEngineConfig * SearchEngineFactory * SearchFormWidget * SearchNearMatcher * SessionBackend * SpecialPageFactory * UserNameUtils * UserOptionsManager * WatchedItemQueryService * WatchedItemStore Constructors with new optional arguments: * DefaultPreferencesFactory * Language * LinkHolderArray * MovePage * Parser * ParserCache * PasswordReset * Router setHookContainer() now required after construction: * AuthenticationProvider * ResourceLoaderModule * SearchEngine Change-Id: Id442b0dbe43aba84bd5cf801d86dedc768b082c7
2020-03-19 02:42:09 +00:00
use ProtectedHookAccessorTrait;
OutputPage: Support loading some mw.config vars late, at end of body JS config vars listed in the 'LateJSConfigVarNames' extension attribute or in the OutputPage::CORE_LATE_JS_CONFIG_VAR_NAMES constant will now be added to the end of the HTML response, instead of early in the HTML head where it blocks rendering/painting. Bug: T41813 Change-Id: I86d461756398780dc24949013f35b7730a481052
2022-10-25 22:36:20 +00:00
// Constants for getJSVars()
private const JS_VAR_EARLY = 1;
private const JS_VAR_LATE = 2;
// Core config vars that opt-in to JS_VAR_LATE.
// Extensions use the 'LateJSConfigVarNames' attribute instead.
private const CORE_LATE_JS_CONFIG_VAR_NAMES = [];
Unsuppress other phan issues (part 4) Bug: T231636 Depends-On: I58e67c2b38389df874438deada4239510d21654f Change-Id: I6e5fba7bd273219b1206559420b5bdb78734aa84
2019-08-30 16:01:28 +00:00
/** @var string[][] Should be private. Used with addMeta() which adds "<meta>" */
Convert all array() syntax to [] Per wikitech-l consensus: https://lists.wikimedia.org/pipermail/wikitech-l/2016-February/084821.html Notes: * Disabled CallTimePassByReference due to false positives (T127163) Change-Id: I2c8ce713ce6600a0bb7bf67537c87044c7a45c4b
2016-02-17 09:09:32 +00:00
protected $mMetatags = [];
OutputPage properties documentation This is following a pair session with Zak Greant on november 12th 2010. Please review, ammend and commit your changes if you have anything useful to add. Please note you will find some humoristics quotes.
2011-01-13 21:08:59 +00:00
Pass php-strict for OutputPage Change-Id: Ia5d03277fff543ca672dc8efc7eaaeca4888ac23
2014-05-11 01:54:15 +00:00
/** @var array */
Convert all array() syntax to [] Per wikitech-l consensus: https://lists.wikimedia.org/pipermail/wikitech-l/2016-February/084821.html Notes: * Disabled CallTimePassByReference due to false positives (T127163) Change-Id: I2c8ce713ce6600a0bb7bf67537c87044c7a45c4b
2016-02-17 09:09:32 +00:00
protected $mLinktags = [];
OutputPage properties documentation This is following a pair session with Zak Greant on november 12th 2010. Please review, ammend and commit your changes if you have anything useful to add. Please note you will find some humoristics quotes.
2011-01-13 21:08:59 +00:00
Various doc fixes about false on method arguments/return types Doc-only changes Change-Id: I5177f582ae7ee70c357e9389fed14819faf79463
2022-11-09 00:11:22 +00:00
/** @var string|false */
Pass php-strict for OutputPage Change-Id: Ia5d03277fff543ca672dc8efc7eaaeca4888ac23
2014-05-11 01:54:15 +00:00
protected $mCanonicalUrl = false;
Fixing up spaces to tabs in r80192 Removing trailing whitespace
2011-01-13 22:08:13 +00:00
Pass php-strict for OutputPage Change-Id: Ia5d03277fff543ca672dc8efc7eaaeca4888ac23
2014-05-11 01:54:15 +00:00
/**
Clean up spacing of doc comments Align the doc stars and normalize start and end tokens Change-Id: Ib0d92e128e7b882bb5b838bd00c74fc16ef14303
2019-08-05 17:00:00 +00:00
* @var string The contents of <h1>
*/
Make OutputPage::mPagetitle private Code search revealed only one outside access, which I've submitted a patch for. Depends-On: I001deecad241b091070483d882088d2d6361ed5a Change-Id: I0f7e565f4eacc1074c619772b2d11a47a05e3584
2018-07-24 13:38:44 +00:00
private $mPageTitle = '';
OutputPage properties documentation This is following a pair session with Zak Greant on november 12th 2010. Please review, ammend and commit your changes if you have anything useful to add. Please note you will find some humoristics quotes.
2011-01-13 21:08:59 +00:00
Improve page display title handling for category pages One use case of display title is to localise page names with Translate extension or without. While the page title changes, the subheadings still say something like "Pages in category Foo/de". Also converted one raw HTML message to be a parsed message. First version of this patch caused an issue when previewing because page title in h1 is not the same as page display title. This issue is fixed by promoting page display title as it's own member in OutputPage. Also added getUnprefixedDisplayTitle that attempts to strip away the namespace prefix to mimic Title::getText() but which works with display title instead. Bug: T43720 Bug: T46197 Change-Id: I6097a873297eb57759252fc56ad6d02c44e4c366
2016-11-09 06:24:57 +00:00
/**
* @var string The displayed title of the page. Different from page title
* if overridden by display title magic word or hooks. Can contain safe
* HTML. Different from page title which may contain messages such as
* "Editing X" which is displayed in h1. This can be used for other places
* where the page name is referred on the page.
*/
private $displayTitle;
block: Allow cookie-block tracking from any uncached web request This was previously hardcoded from three places: 1) Upon viewing EditPage, 2) Upon viewing SpecialCreateAccount, 3) For any url if the user is logged-in (User::loadFromSession/isLoggedIn). == User::loadFromSession Performing cookie blocks from here created a circular dependency because Block may need the user language for localisation, which is determined by asking the User object. This was previously worked around by using a DeferredUpdate (T180050, T226777). Moving this logic explicitly to the end of the pre-send cycle in MediaWiki::preOutputCommit breaks the cycle. This is also where other request-specific handling resides already. == Limited effect on unregistered users When an unregistered user performs an edit, and gets blocked, the cookie block is not applied until they open built-in editor or CreateAccount page. This makes it more likely for a user's IP to change meanwhile. Either intentionally, or simply due to IPs varying naturally (e.g. between mobile locations, or when going on/off WiFi). By applying it throughout sessioned page views for unregistered users, it is more likely to get set. Similar to what was already done for logged-in users. This commit also makes the intent of not caching EditPage and SpecialCreateAccount explicit. This was previously implicit through nothing having called setCdnMaxage() and/or due to Session::persist being checked for by OutputPage::sendCacheControl. Bug: T233594 Change-Id: Icf5a00f9b41d31bb6d4742c049feca0039d0c9d9
2019-09-07 23:44:46 +00:00
/** @var bool See OutputPage::couldBePublicCached. */
private $cacheIsFinal = false;
Pass php-strict for OutputPage Change-Id: Ia5d03277fff543ca672dc8efc7eaaeca4888ac23
2014-05-11 01:54:15 +00:00
/**
* @var string Contains all of the "<body>" content. Should be private we
* got set/get accessors and the append() method.
*/
public $mBodytext = '';
Document mDebugtext
2010-07-08 08:12:19 +00:00
Pass php-strict for OutputPage Change-Id: Ia5d03277fff543ca672dc8efc7eaaeca4888ac23
2014-05-11 01:54:15 +00:00
/** @var string Stores contents of "<title>" tag */
private $mHTMLtitle = '';
* Changed OutputPage::$mIsArticle flag to be false by default. A lot of actions don't change that flag when they should, since about only action=view should have it to true. * Put OutputPage::$mIsArticleRelated declaration just below $mIsArticle's one since they are related * Made DifferenceEngine consistent with action=view, i.e. setting that flag to true when displaying any version of the page, current or not * Made Skin::getCopyright() always use history_copyright when displaying an old version, regardless to the diff parameter * Changed some checks from namespace != NS_SPECIAL && action=view to isArticle() so that they don't get executed when e.g. throwing an Exception (or similar things)
2011-07-27 13:14:28 +00:00
/**
Pass php-strict for OutputPage Change-Id: Ia5d03277fff543ca672dc8efc7eaaeca4888ac23
2014-05-11 01:54:15 +00:00
* @var bool Is the displayed content related to the source of the
* corresponding wiki article.
* Changed OutputPage::$mIsArticle flag to be false by default. A lot of actions don't change that flag when they should, since about only action=view should have it to true. * Put OutputPage::$mIsArticleRelated declaration just below $mIsArticle's one since they are related * Made DifferenceEngine consistent with action=view, i.e. setting that flag to true when displaying any version of the page, current or not * Made Skin::getCopyright() always use history_copyright when displaying an old version, regardless to the diff parameter * Changed some checks from namespace != NS_SPECIAL && action=view to isArticle() so that they don't get executed when e.g. throwing an Exception (or similar things)
2011-07-27 13:14:28 +00:00
*/
Improve OutputPage test coverage more Improved locally from 26.81% to 37.18%. Depends-On: Ib564624c644ee6620ac06872f5684831acaaaadb Change-Id: Icdc0288c04b8c4ba841f9fbb3e05a0cdc8a20fa5
2018-07-25 18:41:42 +00:00
private $mIsArticle = false;
Pass php-strict for OutputPage Change-Id: Ia5d03277fff543ca672dc8efc7eaaeca4888ac23
2014-05-11 01:54:15 +00:00
/** @var bool Stores "article flag" toggle. */
private $mIsArticleRelated = true;
OutputPage properties documentation This is following a pair session with Zak Greant on november 12th 2010. Please review, ammend and commit your changes if you have anything useful to add. Please note you will find some humoristics quotes.
2011-01-13 21:08:59 +00:00
Show copyright based on $output->hasCopyright() Introducing setCopyright/hasCopyright in OutputPage to make the showing of the copyright message controllable by pages other than articles. hasCopyright defaults to isArticle() to preserve the rule that all article should show the copyright. This is used by Flow to show the copyright on various pages that contain user-generated content. Bug: T184960 Change-Id: I3a50dbcedc6b119b9262c50cb3a84b0dd230fb3d
2018-10-03 13:12:37 +00:00
/** @var bool Is the content subject to copyright */
private $mHasCopyright = false;
Follow-up r80192, tweak docs
2011-01-14 07:54:28 +00:00
/**
Pass php-strict for OutputPage Change-Id: Ia5d03277fff543ca672dc8efc7eaaeca4888ac23
2014-05-11 01:54:15 +00:00
* @var bool We have to set isPrintable(). Some pages should
Follow-up r80192, tweak docs
2011-01-14 07:54:28 +00:00
* never be printed (ex: redirections).
OutputPage properties documentation This is following a pair session with Zak Greant on november 12th 2010. Please review, ammend and commit your changes if you have anything useful to add. Please note you will find some humoristics quotes.
2011-01-13 21:08:59 +00:00
*/
Pass php-strict for OutputPage Change-Id: Ia5d03277fff543ca672dc8efc7eaaeca4888ac23
2014-05-11 01:54:15 +00:00
private $mPrintable = false;
OutputPage properties documentation This is following a pair session with Zak Greant on november 12th 2010. Please review, ammend and commit your changes if you have anything useful to add. Please note you will find some humoristics quotes.
2011-01-13 21:08:59 +00:00
Give skins more flexibility over table of contents render * Do not store table of contents in parser output * Instead inject table of contents via strpos where needed inside Article based on Skin "toc" option * Use <mw:tocplace> as a TOC placeholder; for Parsoid compatibility this will be replaced with a <meta> tag in a followup patch. Bug: T287767 Change-Id: I44045b3b9e78e7ab793da3f37e3c0dbc91cd7d39
2021-09-15 01:00:06 +00:00
/**
Generate/set/get TOCData/SectionMetadata objects instead of arrays * ParserOutput::setSections()/::getSections() are expected to be deprecated. Uses in extensions and skins will need to be migrated in follow up patches once the new interface has stabilized. * In the skins code, the metadata is converted back to an array. Downstream skin TOC consumers will need to be migrated as well before we can remove the toLegacy() conversion. * Fixed SerializationTestTrait's validation method - Not sure if this is overkill but should handle all future complex objects we might stuff into the ParserCache. * This patch emits a backward-compatible Sections property in order to avoid changing the parser cache serialization format. T327439 has been filed to eventually use the JsonCodec support for object serialization, but for this initial patch it makes sense to avoid the need for a concurrent ParserCache format migration by using a backward-compatible serialization. * TOCData is nullable because the intent is that ParserOutput::setTOCData() is MW_MERGE_STRATEGY_WRITE_ONCE; that is, only the top-level fragment composing a page will set the TOCData. This will be enforced in the future via wfDeprecated() (T327429), but again our first patch is as backward-compatible as possible. Bug: T296025 Depends-On: I1b267d23cf49d147c5379b914531303744481b68 Co-Authored-By: C. Scott Ananian <cananian@wikimedia.org> Co-Authored-By: Subramanya Sastry <ssastry@wikimedia.org> Change-Id: I8329864535f0b1dd5f9163868a08d6cb1ffcb78f
2022-09-01 23:07:29 +00:00
* @var ?TOCData Table of Contents information from ParserOutput, or
* null if no TOCData was ever set.
Give skins more flexibility over table of contents render * Do not store table of contents in parser output * Instead inject table of contents via strpos where needed inside Article based on Skin "toc" option * Use <mw:tocplace> as a TOC placeholder; for Parsoid compatibility this will be replaced with a <meta> tag in a followup patch. Bug: T287767 Change-Id: I44045b3b9e78e7ab793da3f37e3c0dbc91cd7d39
2021-09-15 01:00:06 +00:00
*/
Generate/set/get TOCData/SectionMetadata objects instead of arrays * ParserOutput::setSections()/::getSections() are expected to be deprecated. Uses in extensions and skins will need to be migrated in follow up patches once the new interface has stabilized. * In the skins code, the metadata is converted back to an array. Downstream skin TOC consumers will need to be migrated as well before we can remove the toLegacy() conversion. * Fixed SerializationTestTrait's validation method - Not sure if this is overkill but should handle all future complex objects we might stuff into the ParserCache. * This patch emits a backward-compatible Sections property in order to avoid changing the parser cache serialization format. T327439 has been filed to eventually use the JsonCodec support for object serialization, but for this initial patch it makes sense to avoid the need for a concurrent ParserCache format migration by using a backward-compatible serialization. * TOCData is nullable because the intent is that ParserOutput::setTOCData() is MW_MERGE_STRATEGY_WRITE_ONCE; that is, only the top-level fragment composing a page will set the TOCData. This will be enforced in the future via wfDeprecated() (T327429), but again our first patch is as backward-compatible as possible. Bug: T296025 Depends-On: I1b267d23cf49d147c5379b914531303744481b68 Co-Authored-By: C. Scott Ananian <cananian@wikimedia.org> Co-Authored-By: Subramanya Sastry <ssastry@wikimedia.org> Change-Id: I8329864535f0b1dd5f9163868a08d6cb1ffcb78f
2022-09-01 23:07:29 +00:00
private $tocData;
Give skins more flexibility over table of contents render * Do not store table of contents in parser output * Instead inject table of contents via strpos where needed inside Article based on Skin "toc" option * Use <mw:tocplace> as a TOC placeholder; for Parsoid compatibility this will be replaced with a <meta> tag in a followup patch. Bug: T287767 Change-Id: I44045b3b9e78e7ab793da3f37e3c0dbc91cd7d39
2021-09-15 01:00:06 +00:00
Follow-up r80192, tweak docs
2011-01-14 07:54:28 +00:00
/**
Pass php-strict for OutputPage Change-Id: Ia5d03277fff543ca672dc8efc7eaaeca4888ac23
2014-05-11 01:54:15 +00:00
* @var array Contains the page subtitle. Special pages usually have some
* links here. Don't confuse with site subtitle added by skins.
Followup r80192 which had an unclosed comment causing some php notices resulting from hidden var statements.
2011-01-13 22:20:04 +00:00
*/
Convert all array() syntax to [] Per wikitech-l consensus: https://lists.wikimedia.org/pipermail/wikitech-l/2016-February/084821.html Notes: * Disabled CallTimePassByReference due to false positives (T127163) Change-Id: I2c8ce713ce6600a0bb7bf67537c87044c7a45c4b
2016-02-17 09:09:32 +00:00
private $mSubtitle = [];
OutputPage properties documentation This is following a pair session with Zak Greant on november 12th 2010. Please review, ammend and commit your changes if you have anything useful to add. Please note you will find some humoristics quotes.
2011-01-13 21:08:59 +00:00
Pass php-strict for OutputPage Change-Id: Ia5d03277fff543ca672dc8efc7eaaeca4888ac23
2014-05-11 01:54:15 +00:00
/** @var string */
public $mRedirect = '';
/** @var int */
protected $mStatusCode;
OutputPage properties documentation This is following a pair session with Zak Greant on november 12th 2010. Please review, ammend and commit your changes if you have anything useful to add. Please note you will find some humoristics quotes.
2011-01-13 21:08:59 +00:00
Follow-up r80192, tweak docs
2011-01-14 07:54:28 +00:00
/**
Remove incomplete and dated 'wgUseETag' experiment Follows-up: * 8ff36ebd (r9793) Set ETag header on article views. * 8167b89e (r16205) Add $wgUseETag, false by default (broke Squid; T9098). ETag is a useful alternative for If-Modified-Since (see T33639, T49529 about use cases and bugs it may fix). However the current incomplete implementation isn't useful. It's a feature flag that has been disabled since 2006. Enabling it stores ParserCache::getETag in OutputPage and sends it out as a response header. It's ignored on incoming requests. We don't check If-None-Match for '304 Not Modified' anywhere. ResourceLoader has a working implementation that could be leveraged. To be re-evaluated in the future. ParserCache::getETag() could use improvement as well. Change-Id: I37fd9d0ce1ce9bffb79243c6ae340403fdfd769f
2016-06-27 11:21:14 +00:00
* @var string Used for sending cache control.
Pass php-strict for OutputPage Change-Id: Ia5d03277fff543ca672dc8efc7eaaeca4888ac23
2014-05-11 01:54:15 +00:00
* The whole caching system should probably be moved into its own class.
Follow-up r80192, tweak docs
2011-01-14 07:54:28 +00:00
*/
Pass php-strict for OutputPage Change-Id: Ia5d03277fff543ca672dc8efc7eaaeca4888ac23
2014-05-11 01:54:15 +00:00
protected $mLastModified = '';
Fixing up spaces to tabs in r80192 Removing trailing whitespace
2011-01-13 22:08:13 +00:00
Deprecate protected and public properties of OutputPage We plan to refactor these properties and move them into a value object instance. Mark them deprecated for 1.38 so we can move/remove them in a future release. Bug: T301020 Change-Id: I8ec2e903bd7cde8d54fe0366707cbcf625f1f1a1
2022-02-08 16:11:32 +00:00
/**
* @var string[][]
* @deprecated since 1.38; will be made private (T301020)
*/
Convert all array() syntax to [] Per wikitech-l consensus: https://lists.wikimedia.org/pipermail/wikitech-l/2016-February/084821.html Notes: * Disabled CallTimePassByReference due to false positives (T127163) Change-Id: I2c8ce713ce6600a0bb7bf67537c87044c7a45c4b
2016-02-17 09:09:32 +00:00
protected $mCategoryLinks = [];
OutputPage properties documentation This is following a pair session with Zak Greant on november 12th 2010. Please review, ammend and commit your changes if you have anything useful to add. Please note you will find some humoristics quotes.
2011-01-13 21:08:59 +00:00
Deprecate protected and public properties of OutputPage We plan to refactor these properties and move them into a value object instance. Mark them deprecated for 1.38 so we can move/remove them in a future release. Bug: T301020 Change-Id: I8ec2e903bd7cde8d54fe0366707cbcf625f1f1a1
2022-02-08 16:11:32 +00:00
/**
* @var string[][]
* @deprecated since 1.38, will be made private (T301020)
*/
OutputPage::getCategories(): Add a possibility to distinguish "normal" and "hidden" categories There's currently no way to get, e.g., all categories except the hidden ones just as text. The OutputPage::getCategories() method always returns all categories as an array of strings (titles) and the getCategoryLinks() method returns the result of Linker::link but with the distinction between "normal" and "hidden" categories. This change adds a new parameter to OutputPage::getCategories(), $type, which can be used to define, what categories should be returned. The default value is "all", which means, that all categories are returned (the current result of the method). With the value "normal" and "hidden", the method will return the respective values. This could be used in I97d7de723fe72da26c7dbde0a559a13704c7099a to remove the stupid Linker::link() and isset workaround. Change-Id: Iadda9ae362a21fbee770240234b8f55326219932
2016-07-08 19:11:53 +00:00
protected $mCategories = [
'hidden' => [],
'normal' => [],
];
OutputPage properties documentation This is following a pair session with Zak Greant on november 12th 2010. Please review, ammend and commit your changes if you have anything useful to add. Please note you will find some humoristics quotes.
2011-01-13 21:08:59 +00:00
Deprecate protected and public properties of OutputPage We plan to refactor these properties and move them into a value object instance. Mark them deprecated for 1.38 so we can move/remove them in a future release. Bug: T301020 Change-Id: I8ec2e903bd7cde8d54fe0366707cbcf625f1f1a1
2022-02-08 16:11:32 +00:00
/**
* @var string[]
* @deprecated since 1.38; will be made private (T301020)
*/
Convert all array() syntax to [] Per wikitech-l consensus: https://lists.wikimedia.org/pipermail/wikitech-l/2016-February/084821.html Notes: * Disabled CallTimePassByReference due to false positives (T127163) Change-Id: I2c8ce713ce6600a0bb7bf67537c87044c7a45c4b
2016-02-17 09:09:32 +00:00
protected $mIndicators = [];
Implement page status indicators Page status indicators are icons (or short text snippets) usually displayed in the top-right corner of the page, outside of the main content. Basically, <indicator name="foo">[[File:Foo.svg|20px]]</indicator> may be used on a page to place the icon in the indicator area. They are also known as top icons, page icons, heading icons or title icons. I found the discussion on bug 23796 highly illuminating. I suggest that everyone read it before suggesting different design choices. I spent some time with a thesaurus pondering the name. "Emblems" and "badges" were also considered, but the former has a much more limited meaning and the latter is already taken by Wikidata, with a similar but subtly different feature set. I am not aware of any naming conflicts ;) besides new talk page message "indicator" (used by core and Echo in some documents) and OOjs UI indicators (tiny icons like the arrow on a dropdown form element), which shouldn't be confusing. Potential use cases include: * "Lock" indicators for page protection levels * Featured/good article indicators * Redirect shortcuts display ("WP:VPT") * Links to help/manual for special pages * Coordinates?… or globe icon for inline pop-up maps Design features: * Skin-customizable. Skins can fully control where and how indicators are shown, or may just do <?php echo $this->getIndicators(); ?> to output the default structure. By default they are not shown at all. * Extension-customizable. Extensions can call ParserOutput::addIndicator() to insert an indicator from one of the numerous parser hooks. * Wiki-customizable. In addition to just using the parser functions, on-wiki styles and scripts can use the provided classes and ids (.mw-indicator, #mw-indicator-<name>) to customize their display. Design limitations: * Every indicator must have a unique identifier (name). It's not possible to create arrays, or to have several indicators with the same name. In case of duplicates, the latest occurrence of the parser function wins. * Indicators are displayed ordered by their names (and not occurrence order). This ensures consistency across pages and provides a simple means of ordering or grouping them. * Indicators are not stored, tracked or accessible outside of ParserOutput (in particular they're not in the page_props table). They are intended to merely reflect the content or metadata that is already present on the page, and not be data themselves. If you ever think you need to list pages with a given status indicator, instead figure out what it means and use the appropriate tracking category, special page report, already existing page_prop, or other means. Corresponding patch in Vector: I90a8ae15ac8275d084ea5f47b6b2684d5e6c7412. I'll implement support in the other three skins included in the tarball and document it on mediawiki.org after this is merged. Bug: 23796 Change-Id: I2389ff9a5332a2b1d033eb75f0946e5241cfaaf4
2014-09-24 10:44:16 +00:00
Deprecate protected and public properties of OutputPage We plan to refactor these properties and move them into a value object instance. Mark them deprecated for 1.38 so we can move/remove them in a future release. Bug: T301020 Change-Id: I8ec2e903bd7cde8d54fe0366707cbcf625f1f1a1
2022-02-08 16:11:32 +00:00
/**
Replace various `array` type hints with more specific `string[]` There are many, many more. I touch only a few where I'm sure it's never anything but an array of strings. Change-Id: I8b798f2e9d48f07a241b95ce0ace8fa9d981695d
2022-09-27 07:22:15 +00:00
* @var string[] Array of Interwiki Prefixed (non DB key) Titles (e.g. 'fr:Test page')
Deprecate protected and public properties of OutputPage We plan to refactor these properties and move them into a value object instance. Mark them deprecated for 1.38 so we can move/remove them in a future release. Bug: T301020 Change-Id: I8ec2e903bd7cde8d54fe0366707cbcf625f1f1a1
2022-02-08 16:11:32 +00:00
*/
Convert all array() syntax to [] Per wikitech-l consensus: https://lists.wikimedia.org/pipermail/wikitech-l/2016-February/084821.html Notes: * Disabled CallTimePassByReference due to false positives (T127163) Change-Id: I2c8ce713ce6600a0bb7bf67537c87044c7a45c4b
2016-02-17 09:09:32 +00:00
private $mLanguageLinks = [];
OutputPage properties documentation This is following a pair session with Zak Greant on november 12th 2010. Please review, ammend and commit your changes if you have anything useful to add. Please note you will find some humoristics quotes.
2011-01-13 21:08:59 +00:00
Follow-up r80192, tweak docs
2011-01-14 07:54:28 +00:00
/**
resourceloader: Consistently refer to the framework as ResourceLoader Change-Id: Ia59e4eac9662723e80d62f7cfcb9e4292e3ee4de
2015-10-28 03:24:40 +00:00
* Used for JavaScript (predates ResourceLoader)
Pass php-strict for OutputPage Change-Id: Ia5d03277fff543ca672dc8efc7eaaeca4888ac23
2014-05-11 01:54:15 +00:00
* @todo We should split JS / CSS.
escape tags and entity in doxygen comments When inserting XML elements inline <such as this one>, doxygen chokes about it not being known. Simply enclosing the tag in double quotes prevents doxygen from emitting a warning. Also enclosed a few invalid functions calls such as \. and double quoted the HTML entities such as &foobar; Change-Id: I4019637145e683c2bec3d17b2fd98b0c50a932f1
2012-07-10 12:48:06 +00:00
* mScripts content is inserted as is in "<head>" by Skin. This might
Pass php-strict for OutputPage Change-Id: Ia5d03277fff543ca672dc8efc7eaaeca4888ac23
2014-05-11 01:54:15 +00:00
* contain either a link to a stylesheet or inline CSS.
OutputPage properties documentation This is following a pair session with Zak Greant on november 12th 2010. Please review, ammend and commit your changes if you have anything useful to add. Please note you will find some humoristics quotes.
2011-01-13 21:08:59 +00:00
*/
Pass php-strict for OutputPage Change-Id: Ia5d03277fff543ca672dc8efc7eaaeca4888ac23
2014-05-11 01:54:15 +00:00
private $mScripts = '';
Follow-up r80192, tweak docs
2011-01-14 07:54:28 +00:00
Pass php-strict for OutputPage Change-Id: Ia5d03277fff543ca672dc8efc7eaaeca4888ac23
2014-05-11 01:54:15 +00:00
/** @var string Inline CSS styles. Use addInlineStyle() sparingly */
protected $mInlineStyles = '';
OutputPage properties documentation This is following a pair session with Zak Greant on november 12th 2010. Please review, ammend and commit your changes if you have anything useful to add. Please note you will find some humoristics quotes.
2011-01-13 21:08:59 +00:00
/**
Pass php-strict for OutputPage Change-Id: Ia5d03277fff543ca672dc8efc7eaaeca4888ac23
2014-05-11 01:54:15 +00:00
* @var string Used by skin template.
OutputPage properties documentation This is following a pair session with Zak Greant on november 12th 2010. Please review, ammend and commit your changes if you have anything useful to add. Please note you will find some humoristics quotes.
2011-01-13 21:08:59 +00:00
* Example: $tpl->set( 'displaytitle', $out->mPageLinkTitle );
*/
Pass php-strict for OutputPage Change-Id: Ia5d03277fff543ca672dc8efc7eaaeca4888ac23
2014-05-11 01:54:15 +00:00
public $mPageLinkTitle = '';
OutputPage: allow additional CSS classes to be added to <html> In some cases CSS classes must be on the <html> element and not the <body>. An example is the DarkMode extension. Due to how the stacking context works, the CSS `invert()` filter must be on <html>. We can't use JavaScript to apply the styles because there would be a brief delay after page load before the JS runs, causing a flashing effect. Bug: T241925 Change-Id: I4b994e081d9582801fe9d198efd68afc46a14703
2019-05-21 20:26:53 +00:00
/**
* Additional <html> classes; This should be rarely modified; prefer mAdditionalBodyClasses.
* @var array
*/
protected $mAdditionalHtmlClasses = [];
Deprecate protected and public properties of OutputPage We plan to refactor these properties and move them into a value object instance. Mark them deprecated for 1.38 so we can move/remove them in a future release. Bug: T301020 Change-Id: I8ec2e903bd7cde8d54fe0366707cbcf625f1f1a1
2022-02-08 16:11:32 +00:00
/**
Replace various `array` type hints with more specific `string[]` There are many, many more. I touch only a few where I'm sure it's never anything but an array of strings. Change-Id: I8b798f2e9d48f07a241b95ce0ace8fa9d981695d
2022-09-27 07:22:15 +00:00
* @var string[] Array of elements in "<head>". Parser might add its own headers!
Deprecate protected and public properties of OutputPage We plan to refactor these properties and move them into a value object instance. Mark them deprecated for 1.38 so we can move/remove them in a future release. Bug: T301020 Change-Id: I8ec2e903bd7cde8d54fe0366707cbcf625f1f1a1
2022-02-08 16:11:32 +00:00
* @deprecated since 1.38; will be made private (T301020)
*/
Convert all array() syntax to [] Per wikitech-l consensus: https://lists.wikimedia.org/pipermail/wikitech-l/2016-February/084821.html Notes: * Disabled CallTimePassByReference due to false positives (T127163) Change-Id: I2c8ce713ce6600a0bb7bf67537c87044c7a45c4b
2016-02-17 09:09:32 +00:00
protected $mHeadItems = [];
Pass php-strict for OutputPage Change-Id: Ia5d03277fff543ca672dc8efc7eaaeca4888ac23
2014-05-11 01:54:15 +00:00
Add <body> class marking whether RCFilters is enabled Also added OutputPage->addBodyClasses to allow this Bug: T175062 Change-Id: I2111d0762328f2e59038c92d63583a9bb267fc9f
2017-09-11 07:12:03 +00:00
/** @var array Additional <body> classes; there are also <body> classes from other sources */
protected $mAdditionalBodyClasses = [];
Deprecate protected and public properties of OutputPage We plan to refactor these properties and move them into a value object instance. Mark them deprecated for 1.38 so we can move/remove them in a future release. Bug: T301020 Change-Id: I8ec2e903bd7cde8d54fe0366707cbcf625f1f1a1
2022-02-08 16:11:32 +00:00
/**
* @var array
* @deprecated since 1.38; will be made private (T301020)
*/
Convert all array() syntax to [] Per wikitech-l consensus: https://lists.wikimedia.org/pipermail/wikitech-l/2016-February/084821.html Notes: * Disabled CallTimePassByReference due to false positives (T127163) Change-Id: I2c8ce713ce6600a0bb7bf67537c87044c7a45c4b
2016-02-17 09:09:32 +00:00
protected $mModules = [];
Pass php-strict for OutputPage Change-Id: Ia5d03277fff543ca672dc8efc7eaaeca4888ac23
2014-05-11 01:54:15 +00:00
Deprecate protected and public properties of OutputPage We plan to refactor these properties and move them into a value object instance. Mark them deprecated for 1.38 so we can move/remove them in a future release. Bug: T301020 Change-Id: I8ec2e903bd7cde8d54fe0366707cbcf625f1f1a1
2022-02-08 16:11:32 +00:00
/**
* @var array
* @deprecated since 1.38; will be made private (T301020)
*/
Convert all array() syntax to [] Per wikitech-l consensus: https://lists.wikimedia.org/pipermail/wikitech-l/2016-February/084821.html Notes: * Disabled CallTimePassByReference due to false positives (T127163) Change-Id: I2c8ce713ce6600a0bb7bf67537c87044c7a45c4b
2016-02-17 09:09:32 +00:00
protected $mModuleStyles = [];
Document mDebugtext
2010-07-08 08:12:19 +00:00
Pass php-strict for OutputPage Change-Id: Ia5d03277fff543ca672dc8efc7eaaeca4888ac23
2014-05-11 01:54:15 +00:00
/** @var ResourceLoader */
protected $mResourceLoader;
Fixing up spaces to tabs in r80192 Removing trailing whitespace
2011-01-13 22:08:13 +00:00
ResourceLoader namespace (attempt 2) Move ResourceLoader classes to their own namespace. Strip the "ResourceLoader" prefix from all except ResourceLoader itself. Move the tests by analogy. I used a namespace alias "RL" in some callers since RL\Module is less ambiguous at the call site than just "Module". I did not address DependencyStore which continues to have a non-standard location and namespace. Revert of a241d83e0a6dabedf. Bug: T308718 Change-Id: Id08a220e1d6085e2b33f3f6c9d0e3935a4204659
2022-05-06 09:09:56 +00:00
/** @var RL\ClientHtml */
resourceloader: Move queue formatting out of OutputPage HTML formatting of the queue was distributed over several OutputPage methods. Each method demanding a snippet of HTML by calling makeResourceLoaderLink() with a limited amount of information. As such, makeResourceLoaderLink() was unable to provide the client with the proper state information. Centralising it also allows it to better reduce duplication in HTML output and maintain a more accurate state. Problems fixed by centralising: 1. The 'user' module is special (due to per-user 'version' and 'user' params). It is manually requested via script-src. To avoid a separate (and wrong) request from something that requires it, we set state=loading directly. However, because the module is in the bottom, the old HTML formatter could only put state=loading in the bottom also. This sometimes caused a wrong request to be fired for modules=user if something in the top queue triggered a requirement for it. 2. Since a464d1d4 (T87871) we track states of page-style modules, with purpose of allowing dependencies on style modules without risking duplicate loading on pages where the styles are loaded already. This didn't work, because the state information about page-style modules is output near the stylesheet, which is after the script tag with mw.loader.load(). That runs first, and mw.loader would still make a duplicate request before it learns the state. Changes: * Document reasons for style/script tag order in getHeadHtml (per 09537e83). * Pass $type from getModuleStyles() to getAllowedModules(). This wasn't needed before since a duplicate check in makeResourceLoaderLink() verified the origin a second time. * Declare explicit position 'top' on 'user.options' and 'user.tokens' module. Previously, OutputPage hardcoded them in the top. The new formatter doesn't. * Remove getHeadScripts(). * Remove getInlineHeadScripts(). * Remove getExternalHeadScripts(). * Remove buildCssLinks(). * Remove getScriptsForBottomQueue(). * Change where Skin::setupSkinUserCss() is called. This methods lets the skin add modules to the queue. Previously it was called from buildCssLinks(), via headElement(), via prepareQuickTemplate(), via OutputPage::output(). It's now in OutputPage::output() directly (slightly earlier). This is needed because prepareQuickTemplate() calls bottomScripts() before headElement(). And bottomScript() would lazy-initialise the queue and lock it before setupSkinUserCss() is called from headElement(). This makes execution order more predictable instead of being dependent on the arbitrary order of data extraction in prepareQuickTemplate (which varies from one skin to another). * Compute isUserModulePreview() and isKnownEmpty() for the 'user' module early on so. This avoids wrongful loading and fixes problem 1. Effective changes in output: * mw.loader.state() is now before mw.loader.load(). This fixes problem 2. * mw.loader.state() now sets 'user.options' and 'user.tokens' to "loading". * mw.loader.state() now sets 'user' (as "loading" or "ready"). Fixes problem 1. * The <script async src> tag for 'startup' changed position (slightly). Previously it was after all inline scripts and stylesheets. It's still after all inline scripts and after most stylesheets, but before any user styles. Since the queue is now formatted outside OutputPage, it can't inject the meta-ResourceLoaderDynamicStyles tag and user-stylesheet hack in the middle of existing output. This shouldn't have any noticable impact. Bug: T87871 Change-Id: I605b8cd1e1fc009b4662a0edbc54d09dd65ee1df
2016-07-15 14:13:09 +00:00
private $rlClient;
ResourceLoader namespace (attempt 2) Move ResourceLoader classes to their own namespace. Strip the "ResourceLoader" prefix from all except ResourceLoader itself. Move the tests by analogy. I used a namespace alias "RL" in some callers since RL\Module is less ambiguous at the call site than just "Module". I did not address DependencyStore which continues to have a non-standard location and namespace. Revert of a241d83e0a6dabedf. Bug: T308718 Change-Id: Id08a220e1d6085e2b33f3f6c9d0e3935a4204659
2022-05-06 09:09:56 +00:00
/** @var RL\Context */
resourceloader: Move queue formatting out of OutputPage HTML formatting of the queue was distributed over several OutputPage methods. Each method demanding a snippet of HTML by calling makeResourceLoaderLink() with a limited amount of information. As such, makeResourceLoaderLink() was unable to provide the client with the proper state information. Centralising it also allows it to better reduce duplication in HTML output and maintain a more accurate state. Problems fixed by centralising: 1. The 'user' module is special (due to per-user 'version' and 'user' params). It is manually requested via script-src. To avoid a separate (and wrong) request from something that requires it, we set state=loading directly. However, because the module is in the bottom, the old HTML formatter could only put state=loading in the bottom also. This sometimes caused a wrong request to be fired for modules=user if something in the top queue triggered a requirement for it. 2. Since a464d1d4 (T87871) we track states of page-style modules, with purpose of allowing dependencies on style modules without risking duplicate loading on pages where the styles are loaded already. This didn't work, because the state information about page-style modules is output near the stylesheet, which is after the script tag with mw.loader.load(). That runs first, and mw.loader would still make a duplicate request before it learns the state. Changes: * Document reasons for style/script tag order in getHeadHtml (per 09537e83). * Pass $type from getModuleStyles() to getAllowedModules(). This wasn't needed before since a duplicate check in makeResourceLoaderLink() verified the origin a second time. * Declare explicit position 'top' on 'user.options' and 'user.tokens' module. Previously, OutputPage hardcoded them in the top. The new formatter doesn't. * Remove getHeadScripts(). * Remove getInlineHeadScripts(). * Remove getExternalHeadScripts(). * Remove buildCssLinks(). * Remove getScriptsForBottomQueue(). * Change where Skin::setupSkinUserCss() is called. This methods lets the skin add modules to the queue. Previously it was called from buildCssLinks(), via headElement(), via prepareQuickTemplate(), via OutputPage::output(). It's now in OutputPage::output() directly (slightly earlier). This is needed because prepareQuickTemplate() calls bottomScripts() before headElement(). And bottomScript() would lazy-initialise the queue and lock it before setupSkinUserCss() is called from headElement(). This makes execution order more predictable instead of being dependent on the arbitrary order of data extraction in prepareQuickTemplate (which varies from one skin to another). * Compute isUserModulePreview() and isKnownEmpty() for the 'user' module early on so. This avoids wrongful loading and fixes problem 1. Effective changes in output: * mw.loader.state() is now before mw.loader.load(). This fixes problem 2. * mw.loader.state() now sets 'user.options' and 'user.tokens' to "loading". * mw.loader.state() now sets 'user' (as "loading" or "ready"). Fixes problem 1. * The <script async src> tag for 'startup' changed position (slightly). Previously it was after all inline scripts and stylesheets. It's still after all inline scripts and after most stylesheets, but before any user styles. Since the queue is now formatted outside OutputPage, it can't inject the meta-ResourceLoaderDynamicStyles tag and user-stylesheet hack in the middle of existing output. This shouldn't have any noticable impact. Bug: T87871 Change-Id: I605b8cd1e1fc009b4662a0edbc54d09dd65ee1df
2016-07-15 14:13:09 +00:00
private $rlClientContext;
OutputPage: Make ResourceLoader position exemption more generic Follows-up 80e5b160e which moved queue formatting out of OutputPage into a a separate ResourceLoaderClientHtml class. The special handling for 'user' and 'user.styles' modules, and the exempt module groups was kept in OutputPage. However the handling for it was hardcoded for the modules in that group by default. It did not account for modules with a group of 'user' loaded by an extension (e.g. GlobalCssJs). GlobalCssJs modules were wrongly loaded in the regular style queue (still in a separate request group, but not in the right cascading order below the DynamicSyles marker). This commit generalises the handling previously put in buildExemptModules and moves it to getRlClient() so that it may apply to all style modules. This commit should be a no-op besides the moving of any <link rel=stylesheet> for non-core modules in group 'site' or 'user' now being one line lower in the <head> HTML (after the DynamicStyles marker). Bug: T143357 Change-Id: I1d6ea10b42293acfc535578172ad7ab2369f6299
2016-08-19 02:04:21 +00:00
/** @var array */
private $rlExemptStyleModules;
Deprecate protected and public properties of OutputPage We plan to refactor these properties and move them into a value object instance. Mark them deprecated for 1.38 so we can move/remove them in a future release. Bug: T301020 Change-Id: I8ec2e903bd7cde8d54fe0366707cbcf625f1f1a1
2022-02-08 16:11:32 +00:00
/**
* @var array
* @deprecated since 1.38; will be made private (T301020)
*/
Convert all array() syntax to [] Per wikitech-l consensus: https://lists.wikimedia.org/pipermail/wikitech-l/2016-February/084821.html Notes: * Disabled CallTimePassByReference due to false positives (T127163) Change-Id: I2c8ce713ce6600a0bb7bf67537c87044c7a45c4b
2016-02-17 09:09:32 +00:00
protected $mJsConfigVars = [];
* Enhanced OutputPage to support js2 style javascript without scriptloader enabled ** loadGM output ** unique request id based on wikiTitle revision / file time * Moved jsAutoLoadClasses to setup.php ( made conditional )
2009-09-17 01:19:02 +00:00
Deprecate protected and public properties of OutputPage We plan to refactor these properties and move them into a value object instance. Mark them deprecated for 1.38 so we can move/remove them in a future release. Bug: T301020 Change-Id: I8ec2e903bd7cde8d54fe0366707cbcf625f1f1a1
2022-02-08 16:11:32 +00:00
/**
* @var array
* @deprecated since 1.38; will be made private (T301020)
*/
Convert all array() syntax to [] Per wikitech-l consensus: https://lists.wikimedia.org/pipermail/wikitech-l/2016-February/084821.html Notes: * Disabled CallTimePassByReference due to false positives (T127163) Change-Id: I2c8ce713ce6600a0bb7bf67537c87044c7a45c4b
2016-02-17 09:09:32 +00:00
protected $mTemplateIds = [];
Revert to r14165 . Did too many changes, didnt even run parserTests (i am bad)
2006-05-11 22:40:38 +00:00
Pass php-strict for OutputPage Change-Id: Ia5d03277fff543ca672dc8efc7eaaeca4888ac23
2014-05-11 01:54:15 +00:00
/** @var array */
Convert all array() syntax to [] Per wikitech-l consensus: https://lists.wikimedia.org/pipermail/wikitech-l/2016-February/084821.html Notes: * Disabled CallTimePassByReference due to false positives (T127163) Change-Id: I2c8ce713ce6600a0bb7bf67537c87044c7a45c4b
2016-02-17 09:09:32 +00:00
protected $mImageTimeKeys = [];
Declared some member fields
2011-07-06 17:55:44 +00:00
Pass php-strict for OutputPage Change-Id: Ia5d03277fff543ca672dc8efc7eaaeca4888ac23
2014-05-11 01:54:15 +00:00
/** @var string */
public $mRedirectCode = '';
Declared some member fields
2011-07-06 17:55:44 +00:00
Pass php-strict for OutputPage Change-Id: Ia5d03277fff543ca672dc8efc7eaaeca4888ac23
2014-05-11 01:54:15 +00:00
protected $mFeedLinksAppendQuery = null;
Make allowing site-wide styles on restricted special pages a config option This mostly reverts commit 614d7e5c274d927f99bfc52ac3a1e6c7e5902408. Many wikis use MediaWiki:Common.css and associated pages to create a custom "theme" for their wiki, which would no longer load on login or preference pages, creating an inconsistent UI. This re-adds the difference in module origin for different types (styles, scripts, etc.), and now OutputPage::disallowUserJs() checks the value of the "AllowSiteCSSOnRestrictedPages" config setting to determine whether to allow site-wide CSS styles or not. By default this feature is disabled to be secure by default. Bug: 71621 Change-Id: I1bf4dd1845b6952c3985e179fbea48181ffb8907
2014-10-10 06:46:12 +00:00
/** @var array
* What level of 'untrustworthiness' is allowed in CSS/JS modules loaded on this page?
ResourceLoader namespace (attempt 2) Move ResourceLoader classes to their own namespace. Strip the "ResourceLoader" prefix from all except ResourceLoader itself. Move the tests by analogy. I used a namespace alias "RL" in some callers since RL\Module is less ambiguous at the call site than just "Module". I did not address DependencyStore which continues to have a non-standard location and namespace. Revert of a241d83e0a6dabedf. Bug: T308718 Change-Id: Id08a220e1d6085e2b33f3f6c9d0e3935a4204659
2022-05-06 09:09:56 +00:00
* @see RL\Module::$origin
* RL\Module::ORIGIN_ALL is assumed unless overridden;
Pass php-strict for OutputPage Change-Id: Ia5d03277fff543ca672dc8efc7eaaeca4888ac23
2014-05-11 01:54:15 +00:00
*/
Convert all array() syntax to [] Per wikitech-l consensus: https://lists.wikimedia.org/pipermail/wikitech-l/2016-February/084821.html Notes: * Disabled CallTimePassByReference due to false positives (T127163) Change-Id: I2c8ce713ce6600a0bb7bf67537c87044c7a45c4b
2016-02-17 09:09:32 +00:00
protected $mAllowedModules = [
ResourceLoader namespace (attempt 2) Move ResourceLoader classes to their own namespace. Strip the "ResourceLoader" prefix from all except ResourceLoader itself. Move the tests by analogy. I used a namespace alias "RL" in some callers since RL\Module is less ambiguous at the call site than just "Module". I did not address DependencyStore which continues to have a non-standard location and namespace. Revert of a241d83e0a6dabedf. Bug: T308718 Change-Id: Id08a220e1d6085e2b33f3f6c9d0e3935a4204659
2022-05-06 09:09:56 +00:00
RL\Module::TYPE_COMBINED => RL\Module::ORIGIN_ALL,
Convert all array() syntax to [] Per wikitech-l consensus: https://lists.wikimedia.org/pipermail/wikitech-l/2016-February/084821.html Notes: * Disabled CallTimePassByReference due to false positives (T127163) Change-Id: I2c8ce713ce6600a0bb7bf67537c87044c7a45c4b
2016-02-17 09:09:32 +00:00
];
OutputPage properties documentation This is following a pair session with Zak Greant on november 12th 2010. Please review, ammend and commit your changes if you have anything useful to add. Please note you will find some humoristics quotes.
2011-01-13 21:08:59 +00:00
Pass php-strict for OutputPage Change-Id: Ia5d03277fff543ca672dc8efc7eaaeca4888ac23
2014-05-11 01:54:15 +00:00
/** @var bool Whether output is disabled. If this is true, the 'output' method will do nothing. */
protected $mDoNothing = false;
// Parser related.
Add, update, tweak documentation Fix document comment blocks Tweak some returns
2011-10-26 03:44:47 +00:00
/**
* lazy initialised, use parserOptions()
* @var ParserOptions
*/
OutputPage properties documentation This is following a pair session with Zak Greant on november 12th 2010. Please review, ammend and commit your changes if you have anything useful to add. Please note you will find some humoristics quotes.
2011-01-13 21:08:59 +00:00
protected $mParserOptions = null;
Follow-up r80192, tweak docs
2011-01-14 07:54:28 +00:00
/**
Pass php-strict for OutputPage Change-Id: Ia5d03277fff543ca672dc8efc7eaaeca4888ac23
2014-05-11 01:54:15 +00:00
* Handles the Atom / RSS links.
* We probably only support Atom in 2011.
Follow-up r80192, tweak docs
2011-01-14 07:54:28 +00:00
* @see $wgAdvertisedFeedTypes
OutputPage properties documentation This is following a pair session with Zak Greant on november 12th 2010. Please review, ammend and commit your changes if you have anything useful to add. Please note you will find some humoristics quotes.
2011-01-13 21:08:59 +00:00
*/
Convert all array() syntax to [] Per wikitech-l consensus: https://lists.wikimedia.org/pipermail/wikitech-l/2016-February/084821.html Notes: * Disabled CallTimePassByReference due to false positives (T127163) Change-Id: I2c8ce713ce6600a0bb7bf67537c87044c7a45c4b
2016-02-17 09:09:32 +00:00
private $mFeedLinks = [];
(missing .= part) r58512#c4384
2009-11-07 09:26:02 +00:00
Add OutputPage::disableClientCache() and deprecate ::enableClientCache(false) As far as I can tell, outside of test cases, every single time we call OutputPage::enableClientCache() we pass it `false` and actually mean to *disable* the client cache. Create a new less-confusingly-named method and deprecate the old one. Change-Id: I7c89e20528a0d91173f0edcb997dcae631935ee5
2022-02-04 19:12:11 +00:00
/**
* @var bool Set to false to send no-cache headers, disabling
* client-side caching. (This variable should really be named
* in the opposite sense; see ::disableClientCache().)
Deprecate protected and public properties of OutputPage We plan to refactor these properties and move them into a value object instance. Mark them deprecated for 1.38 so we can move/remove them in a future release. Bug: T301020 Change-Id: I8ec2e903bd7cde8d54fe0366707cbcf625f1f1a1
2022-02-08 16:11:32 +00:00
* @deprecated since 1.38; will be made private (T301020)
Add OutputPage::disableClientCache() and deprecate ::enableClientCache(false) As far as I can tell, outside of test cases, every single time we call OutputPage::enableClientCache() we pass it `false` and actually mean to *disable* the client cache. Create a new less-confusingly-named method and deprecate the old one. Change-Id: I7c89e20528a0d91173f0edcb997dcae631935ee5
2022-02-04 19:12:11 +00:00
*/
Pass php-strict for OutputPage Change-Id: Ia5d03277fff543ca672dc8efc7eaaeca4888ac23
2014-05-11 01:54:15 +00:00
protected $mEnableClientCache = true;
OutputPage properties documentation This is following a pair session with Zak Greant on november 12th 2010. Please review, ammend and commit your changes if you have anything useful to add. Please note you will find some humoristics quotes.
2011-01-13 21:08:59 +00:00
Pass php-strict for OutputPage Change-Id: Ia5d03277fff543ca672dc8efc7eaaeca4888ac23
2014-05-11 01:54:15 +00:00
/** @var bool Flag if output should only contain the body of the article. */
private $mArticleBodyOnly = false;
Deprecate protected and public properties of OutputPage We plan to refactor these properties and move them into a value object instance. Mark them deprecated for 1.38 so we can move/remove them in a future release. Bug: T301020 Change-Id: I8ec2e903bd7cde8d54fe0366707cbcf625f1f1a1
2022-02-08 16:11:32 +00:00
/**
* @var bool
* @deprecated since 1.38; will be made private (T301020)
*/
Pass php-strict for OutputPage Change-Id: Ia5d03277fff543ca672dc8efc7eaaeca4888ac23
2014-05-11 01:54:15 +00:00
protected $mNewSectionLink = false;
* remove end of line whitespace * remove empty lines at end of file * remove "?>" where still present
2008-04-14 07:45:50 +00:00
Deprecate protected and public properties of OutputPage We plan to refactor these properties and move them into a value object instance. Mark them deprecated for 1.38 so we can move/remove them in a future release. Bug: T301020 Change-Id: I8ec2e903bd7cde8d54fe0366707cbcf625f1f1a1
2022-02-08 16:11:32 +00:00
/**
* @var bool
* @deprecated since 1.38; will be made private (T301020)
*/
Pass php-strict for OutputPage Change-Id: Ia5d03277fff543ca672dc8efc7eaaeca4888ac23
2014-05-11 01:54:15 +00:00
protected $mHideNewSectionLink = false;
OutputPage properties documentation This is following a pair session with Zak Greant on november 12th 2010. Please review, ammend and commit your changes if you have anything useful to add. Please note you will find some humoristics quotes.
2011-01-13 21:08:59 +00:00
Follow-up r80192, tweak docs
2011-01-14 07:54:28 +00:00
/**
Pass php-strict for OutputPage Change-Id: Ia5d03277fff543ca672dc8efc7eaaeca4888ac23
2014-05-11 01:54:15 +00:00
* @var bool Comes from the parser. This was probably made to load CSS/JS
* only if we had "<gallery>". Used directly in CategoryPage.php.
resourceloader: Consistently refer to the framework as ResourceLoader Change-Id: Ia59e4eac9662723e80d62f7cfcb9e4292e3ee4de
2015-10-28 03:24:40 +00:00
* Looks like ResourceLoader can replace this.
Deprecate protected and public properties of OutputPage We plan to refactor these properties and move them into a value object instance. Mark them deprecated for 1.38 so we can move/remove them in a future release. Bug: T301020 Change-Id: I8ec2e903bd7cde8d54fe0366707cbcf625f1f1a1
2022-02-08 16:11:32 +00:00
* @deprecated since 1.38; will be made private (T301020)
OutputPage properties documentation This is following a pair session with Zak Greant on november 12th 2010. Please review, ammend and commit your changes if you have anything useful to add. Please note you will find some humoristics quotes.
2011-01-13 21:08:59 +00:00
*/
Pass php-strict for OutputPage Change-Id: Ia5d03277fff543ca672dc8efc7eaaeca4888ac23
2014-05-11 01:54:15 +00:00
public $mNoGallery = false;
/** @var int Cache stuff. Looks like mEnableClientCache */
Rename setSquidMaxage() => setCdnMaxage() Also updated the OutputPage member variable. Change-Id: I0b41c66da58c6126c1bd96170c02d084a067383c
2015-12-10 01:07:05 +00:00
protected $mCdnMaxage = 0;
/** @var int Upper limit on mCdnMaxage */
Lower CDN cache TTL when slave lag is high * $wgCdnMaxageLagged controls exactly what that TTL is and the usual "max lag" settings determine what "high" is for lag (which already makes the site read-only). * This helps avoids stale content getting stuck in CDN for a month just because a slave was lagged for a minute. Of course race conditions with normal slave lag and WAN cache relay purges can still lead to this problem, though the scope of it is reduced. Bug: T113204 Change-Id: I7ff0a8d88665f4e557566e7b412e75edee2627fe
2015-10-01 07:24:18 +00:00
protected $mCdnMaxageLimit = INF;
OutputPage properties documentation This is following a pair session with Zak Greant on november 12th 2010. Please review, ammend and commit your changes if you have anything useful to add. Please note you will find some humoristics quotes.
2011-01-13 21:08:59 +00:00
Pass php-strict for OutputPage Change-Id: Ia5d03277fff543ca672dc8efc7eaaeca4888ac23
2014-05-11 01:54:15 +00:00
/**
Document OutputPage::$mPreventClickjacking Change-Id: I3b4f967743c810c5cbef094b3e30807cf0a668ca
2015-01-01 11:19:06 +00:00
* @var bool Controls if anti-clickjacking / frame-breaking headers will
* be sent. This should be done for pages where edit actions are possible.
Rename ParserOutput::{allow,prevent}Clickjacking() -> ::{get,set}PreventClickjacking() This name is consist with the rest of the setter and getter methods in ParserOutput. Renamed the methods in OutputPage, ImageHistoryList, ImageHistoryPseudoPager, and ContribsPager as well for consistency; it also makes chasing down lingering references in codesearch easier. Soft-deprecated the old name for 1.38. Hard-deprecation will follow, but there are a number of users in production that should be chased down first. Code search: https://codesearch.https://codesearch.wmcloud.org/deployed/?q=(allow%7Cprevent)Clickjacking&i=nope&files=&excludeFiles=&repos= Bug: T287216 Change-Id: I9822c60c180d204bd30cb4447a1120155d456da4
2021-09-29 20:58:59 +00:00
* Setter: $this->setPreventClickjacking()
Pass php-strict for OutputPage Change-Id: Ia5d03277fff543ca672dc8efc7eaaeca4888ac23
2014-05-11 01:54:15 +00:00
*/
protected $mPreventClickjacking = true;
Avoid null query in Skin::lastModified For action=history there is a query: SELECT rev_timestamp FROM `revision` WHERE rev_id IS NULL LIMIT 1 to generate the footer links Change-Id: I8b0f30f874eb73544693dff9d12d45585bd9a110
2020-08-18 11:41:53 +00:00
/** @var int|null To include the variable {{REVISIONID}} */
Pass php-strict for OutputPage Change-Id: Ia5d03277fff543ca672dc8efc7eaaeca4888ac23
2014-05-11 01:54:15 +00:00
private $mRevisionId = null;
OutputPage properties documentation This is following a pair session with Zak Greant on november 12th 2010. Please review, ammend and commit your changes if you have anything useful to add. Please note you will find some humoristics quotes.
2011-01-13 21:08:59 +00:00
OutputPage: Provide consistent info about whether the revision shown is current Previously, OutputPage::isRevisionCurrent() would check a replica database, potentially different from the database used to fetch the revision being displayed, according to OutputPage::getRevisionId(). Now the data for both functions is provided in the same way. Bug: T314684 Change-Id: I266d643d1eed931df346889f43d72d42e5f4a3ba
2022-08-06 14:40:47 +00:00
/** @var bool|null */
private $mRevisionIsCurrent = null;
Pass php-strict for OutputPage Change-Id: Ia5d03277fff543ca672dc8efc7eaaeca4888ac23
2014-05-11 01:54:15 +00:00
/** @var string */
* Added revision's timestamp to OutputPage along with revision ID; avoid a DB hit in Skin::lastModified() when showing parser cache's content. This changed with the removal of $wgArticle in Skin since now it's a different WikiPage object and thus WikiPage::setTimetstamp() call is useless (but still kept). * Added ParserOutput::(get|set)Timestamp() and the $mTimestamp member; avoid messing with isset()
2011-12-10 16:30:40 +00:00
private $mRevisionTimestamp = null;
OutputPage properties documentation This is following a pair session with Zak Greant on november 12th 2010. Please review, ammend and commit your changes if you have anything useful to add. Please note you will find some humoristics quotes.
2011-01-13 21:08:59 +00:00
Pass php-strict for OutputPage Change-Id: Ia5d03277fff543ca672dc8efc7eaaeca4888ac23
2014-05-11 01:54:15 +00:00
/** @var array */
protected $mFileVersion = null;
* Added getFileVersion()/setFileVersion() functions to OutputPage * Removed getDisplayedFile() from FlaggedPage and simplified getFile() * Cleaned up getIncludeParams() to just do formatting * Made template/file IDs mandatory for RevisionReviewFormUI
2011-07-05 05:23:26 +00:00
Tweaks for skins: * Moved SkinTemplate::addStyle() and related stuff to OutputPage so that it can be used non-SkinTemplate skins and avoid duplication with the actual OutputPage::addStyle() (the two functions have the same format). * Non-SkinTemplate skins now also load their CSS with <link> tags instead of @import. * Moved SkinTemplate::setupUserCss() to Skin. * Merged action=raw&gen=(js|css) for SkinTemplate and non-SkinTemplate skins, renamed functions to Skin::generateUserJs() and Skin::generateUserStyleSheet() and dropped a lot of cascading call which is a bit incomprehensible.
2008-08-21 14:09:57 +00:00
/**
Pass php-strict for OutputPage Change-Id: Ia5d03277fff543ca672dc8efc7eaaeca4888ac23
2014-05-11 01:54:15 +00:00
* @var array An array of stylesheet filenames (relative from skins path),
* with options for CSS media, IE conditions, and RTL/LTR direction.
Tweaks for skins: * Moved SkinTemplate::addStyle() and related stuff to OutputPage so that it can be used non-SkinTemplate skins and avoid duplication with the actual OutputPage::addStyle() (the two functions have the same format). * Non-SkinTemplate skins now also load their CSS with <link> tags instead of @import. * Moved SkinTemplate::setupUserCss() to Skin. * Merged action=raw&gen=(js|css) for SkinTemplate and non-SkinTemplate skins, renamed functions to Skin::generateUserJs() and Skin::generateUserStyleSheet() and dropped a lot of cascading call which is a bit incomprehensible.
2008-08-21 14:09:57 +00:00
* For internal use; add settings in the skin via $this->addStyle()
OutputPage properties documentation This is following a pair session with Zak Greant on november 12th 2010. Please review, ammend and commit your changes if you have anything useful to add. Please note you will find some humoristics quotes.
2011-01-13 21:08:59 +00:00
*
* Style again! This seems like a code duplication since we already have
Pass php-strict for OutputPage Change-Id: Ia5d03277fff543ca672dc8efc7eaaeca4888ac23
2014-05-11 01:54:15 +00:00
* mStyles. This is what makes Open Source amazing.
Tweaks for skins: * Moved SkinTemplate::addStyle() and related stuff to OutputPage so that it can be used non-SkinTemplate skins and avoid duplication with the actual OutputPage::addStyle() (the two functions have the same format). * Non-SkinTemplate skins now also load their CSS with <link> tags instead of @import. * Moved SkinTemplate::setupUserCss() to Skin. * Merged action=raw&gen=(js|css) for SkinTemplate and non-SkinTemplate skins, renamed functions to Skin::generateUserJs() and Skin::generateUserStyleSheet() and dropped a lot of cascading call which is a bit incomprehensible.
2008-08-21 14:09:57 +00:00
*/
Convert all array() syntax to [] Per wikitech-l consensus: https://lists.wikimedia.org/pipermail/wikitech-l/2016-February/084821.html Notes: * Disabled CallTimePassByReference due to false positives (T127163) Change-Id: I2c8ce713ce6600a0bb7bf67537c87044c7a45c4b
2016-02-17 09:09:32 +00:00
protected $styles = [];
Tweaks for skins: * Moved SkinTemplate::addStyle() and related stuff to OutputPage so that it can be used non-SkinTemplate skins and avoid duplication with the actual OutputPage::addStyle() (the two functions have the same format). * Non-SkinTemplate skins now also load their CSS with <link> tags instead of @import. * Moved SkinTemplate::setupUserCss() to Skin. * Merged action=raw&gen=(js|css) for SkinTemplate and non-SkinTemplate skins, renamed functions to Skin::generateUserJs() and Skin::generateUserStyleSheet() and dropped a lot of cascading call which is a bit incomprehensible.
2008-08-21 14:09:57 +00:00
Refactor a bit preparatory to fixing bug 8068: rewrite the robot policy stuff in OutputPage to allow index and follow policy to be set separately. Also now validates input to setRobotPolicy(). And renamed setRobotpolicy to setRobotPolicy, too. If anyone was accessing $mRobotpolicy directly they're out of luck, though.
2008-07-23 19:05:43 +00:00
private $mIndexPolicy = 'index';
private $mFollowPolicy = 'follow';
Comments and code cleanup in OutputPage Change-Id: I8c44f30801792e8b84ab3d556cb5b1b5feb9baf3
2018-08-02 18:26:55 +00:00
Add option to amend robots meta tag - Add private var, helper methods to format and set robots options. - Add reference to new method in OutputPage::getHeadLinksArray(). Bug: T301584 Change-Id: I2652f20c9fe5f2d3e28e329d4eddc2a2ff88cf76
2022-03-10 21:27:21 +00:00
/** @var array */
Allow large image previews by adding max image preview tag - This is across all skins - Adjusted unit tests to check for value Bug: T310267 Change-Id: If61b8bb3bbbe990c8babb0742f747372004dc10c
2022-10-05 13:22:15 +00:00
private $mRobotsOptions = [ 'max-image-preview' => 'standard' ];
Add option to amend robots meta tag - Add private var, helper methods to format and set robots options. - Add reference to new method in OutputPage::getHeadLinksArray(). Bug: T301584 Change-Id: I2652f20c9fe5f2d3e28e329d4eddc2a2ff88cf76
2022-03-10 21:27:21 +00:00
Comments and code cleanup in OutputPage Change-Id: I8c44f30801792e8b84ab3d556cb5b1b5feb9baf3
2018-08-02 18:26:55 +00:00
/**
Remove $wgUseKeyHeader and OutputPage::getKeyHeader(), deprecated in 1.32 These implemented a since-abandoned draft IETF spec, and the code was broken due to (1) case-(in)sensitivity issues with the Accept-Language header and (2) the BCP47 language code compatibility workaround we use. Change-Id: Ia53d07cd8ce8ab1497294ea244c13c7499f632c7
2019-06-19 18:22:42 +00:00
* @var array Headers that cause the cache to vary. Key is header name,
* value should always be null. (Value was an array of options for
* the `Key` header, which was deprecated in 1.32 and removed in 1.34.)
Comments and code cleanup in OutputPage Change-Id: I8c44f30801792e8b84ab3d556cb5b1b5feb9baf3
2018-08-02 18:26:55 +00:00
*/
Convert all array() syntax to [] Per wikitech-l consensus: https://lists.wikimedia.org/pipermail/wikitech-l/2016-February/084821.html Notes: * Disabled CallTimePassByReference due to false positives (T127163) Change-Id: I2c8ce713ce6600a0bb7bf67537c87044c7a45c4b
2016-02-17 09:09:32 +00:00
private $mVaryHeader = [
Remove $wgUseKeyHeader and OutputPage::getKeyHeader(), deprecated in 1.32 These implemented a since-abandoned draft IETF spec, and the code was broken due to (1) case-(in)sensitivity issues with the Accept-Language header and (2) the BCP47 language code compatibility workaround we use. Change-Id: Ia53d07cd8ce8ab1497294ea244c13c7499f632c7
2019-06-19 18:22:42 +00:00
'Accept-Encoding' => null,
Convert all array() syntax to [] Per wikitech-l consensus: https://lists.wikimedia.org/pipermail/wikitech-l/2016-February/084821.html Notes: * Disabled CallTimePassByReference due to false positives (T127163) Change-Id: I2c8ce713ce6600a0bb7bf67537c87044c7a45c4b
2016-02-17 09:09:32 +00:00
];
Step one in fixing OutputPage's documentation: * Groupped methods by function * Added documentation * Fixed some doxygen warnings * Some little style fixes (more to come)
2010-01-28 21:58:50 +00:00
(bug 32230) Expose wgRedirectedFrom (title of the page we were redirected from, if any) in JavaScript. Patch by MrBlueSky
2011-11-30 12:43:10 +00:00
/**
Remove Title from public interface of OutputPage Bug: T278459 Change-Id: If455682540faf8d42fd8a14e0ef09bac24f72db4
2021-04-08 19:11:06 +00:00
* If the current page was reached through a redirect, $mRedirectedFrom contains the title
(bug 32230) Expose wgRedirectedFrom (title of the page we were redirected from, if any) in JavaScript. Patch by MrBlueSky
2011-11-30 12:43:10 +00:00
* of the redirect.
*
Remove Title from public interface of OutputPage Bug: T278459 Change-Id: If455682540faf8d42fd8a14e0ef09bac24f72db4
2021-04-08 19:11:06 +00:00
* @var PageReference
(bug 32230) Expose wgRedirectedFrom (title of the page we were redirected from, if any) in JavaScript. Patch by MrBlueSky
2011-11-30 12:43:10 +00:00
*/
private $mRedirectedFrom = null;
Allow adding arbitrary properties to OutputPage Helps to pass extension-specific data Change-Id: Iea98510abc100128733361f2d2287802f35bb359
2013-02-08 22:58:38 +00:00
/**
* Additional key => value data
*/
Convert all array() syntax to [] Per wikitech-l consensus: https://lists.wikimedia.org/pipermail/wikitech-l/2016-February/084821.html Notes: * Disabled CallTimePassByReference due to false positives (T127163) Change-Id: I2c8ce713ce6600a0bb7bf67537c87044c7a45c4b
2016-02-17 09:09:32 +00:00
private $mProperties = [];
Allow adding arbitrary properties to OutputPage Helps to pass extension-specific data Change-Id: Iea98510abc100128733361f2d2287802f35bb359
2013-02-08 22:58:38 +00:00
Add RL targets support to OutputPage The target URI parameter is already used by ResourceLoaderStartUpModule to filter modules, this change introduces the ability to produce mobile-aware load.php links in the core and filter out modules not suitable for current target. Change-Id: Ie03990e11d9a0ac8cd81dd0aae5c5a95f1d23e81
2013-02-14 17:05:22 +00:00
/**
Removed unneeded spaces and colons in @param and friends Also swapped some "$var type" to "type $var" or added missing types before the $var. Changed some other types to match the more common spelling. Makes beginning of some text in captial. Change-Id: Ic36c8c7820a6c2d603f1138130670c6bf6a1ca59
2014-04-08 15:29:17 +00:00
* @var string|null ResourceLoader target for load.php links. If null, will be omitted
Add RL targets support to OutputPage The target URI parameter is already used by ResourceLoaderStartUpModule to filter modules, this change introduces the ability to produce mobile-aware load.php links in the core and filter out modules not suitable for current target. Change-Id: Ie03990e11d9a0ac8cd81dd0aae5c5a95f1d23e81
2013-02-14 17:05:22 +00:00
*/
private $mTarget = null;
Make TOC hideable Currently, if an extension doesn't want a TOC, it has to remove it manually. This change wraps the TOC in markers that make it easy to remove it in ParserOutput on demand without fragmenting the parser cache with stuff like "use/not use TOC". Change-Id: I2889bcb9eb999c9049601e92440132118e1a8a41
2013-08-22 22:22:03 +00:00
/**
Move loading of mediawiki.toc from Parser to Skin This was the only addModules() call ever to be inside Parser. Introduced in a54ef1a203d. Prior to that, mediawiki.toc had always been loaded by OutputPage (via mediawiki.util; and before that, via wikibits). This patch restores that, and also fixes T130632 by making OutputPage get it from the Skin, instead of hardcoding this somewhere in addParserOutput(). * Remove deprecated method OutputPage::enableTOC(). * Move mEnableTOC to addParserOutputText(). Bug: T130632 Change-Id: Iaad84d241a4c4348c712ac1087a664b8c9c46da4
2017-05-20 13:39:40 +00:00
* @var bool Whether parser output contains a table of contents
Make TOC hideable Currently, if an extension doesn't want a TOC, it has to remove it manually. This change wraps the TOC in markers that make it easy to remove it in ParserOutput on demand without fragmenting the parser cache with stuff like "use/not use TOC". Change-Id: I2889bcb9eb999c9049601e92440132118e1a8a41
2013-08-22 22:22:03 +00:00
*/
Move loading of mediawiki.toc from Parser to Skin This was the only addModules() call ever to be inside Parser. Introduced in a54ef1a203d. Prior to that, mediawiki.toc had always been loaded by OutputPage (via mediawiki.util; and before that, via wikibits). This patch restores that, and also fixes T130632 by making OutputPage get it from the Skin, instead of hardcoding this somewhere in addParserOutput(). * Remove deprecated method OutputPage::enableTOC(). * Move mEnableTOC to addParserOutputText(). Bug: T130632 Change-Id: Iaad84d241a4c4348c712ac1087a664b8c9c46da4
2017-05-20 13:39:40 +00:00
private $mEnableTOC = false;
Make TOC hideable Currently, if an extension doesn't want a TOC, it has to remove it manually. This change wraps the TOC in markers that make it easy to remove it in ParserOutput on demand without fragmenting the parser cache with stuff like "use/not use TOC". Change-Id: I2889bcb9eb999c9049601e92440132118e1a8a41
2013-08-22 22:22:03 +00:00
Fix rel="copyright" for ApiHelp Bug: T94038 Bug: T93995 Change-Id: I06e9bb6797c4e4ee00e5d72eda1e273800b2c7f4
2015-06-08 13:39:06 +00:00
/**
Use <link rel="license"> instead of <link rel="copyright"> "copyright" is not in the list of the allowed values for the attribut rel: https://w3c.github.io/html/single-page.html#allowed-keywords-and-their-meanings Bug: T167284 Change-Id: I09205bb146001d1d1d90e6a0f25d21851f3f697c
2017-06-07 12:11:08 +00:00
* @var string|null The URL to send in a <link> element with rel=license
Fix rel="copyright" for ApiHelp Bug: T94038 Bug: T93995 Change-Id: I06e9bb6797c4e4ee00e5d72eda1e273800b2c7f4
2015-06-08 13:39:06 +00:00
*/
private $copyrightUrl;
Include JS variable for NewPP report Adapted from reverted commit b7c4c8717f9. Bug: T110763 Change-Id: If249b679c534879bfac622592a1d2fa913a0cf9d
2016-11-10 20:29:27 +00:00
/** @var array Profiling data */
private $limitReportJSData = [];
Generalize ResourceLoader 'excludepage' functionality There has long been a hack for previewing edits to user JS/CSS, where OutputPage would pass an 'excludepage' parameter to ResourceLoaderUserModule to tell it not to load one particular page and would instead embed that page statically. That's nice, but there are other places where we could use the same thing. This patch generalizes it: * DerivativeResourceLoaderContext may now contain a callback for mapping titles to replacement Content objects. * ResourceLoaderWikiModule::getContent() uses the overrides, and requests embedding when they're used. All subclasses in Gerrit should pick it up automatically. * OutputPage gains methods for callers to add to the override mapping, which it passes on to RL. It loses a bunch of the special casing it had for the 'user' and 'user.styles' modules. * EditPage sets the overrides on OutputPage when doing the preview, as does ApiParse for prop=headhtml. TemplateSandbox does too in I83fa0856. * OutputPage::userCanPreview() gets less specific to editing user CSS and JS, since RL now handles the embedding based on the actual modules' dependencies and EditPage only requests it on preview. ApiParse also gets a new hook to support TemplateSandbox's API integration (used in I83fa0856). Bug: T112474 Change-Id: Ib9d2ce42931c1de8372e231314a1f672d7e2ac0e
2017-02-28 20:52:17 +00:00
/** @var array Map Title to Content */
private $contentOverrides = [];
/** @var callable[] */
private $contentOverrideCallbacks = [];
Preload the logo using link rel="preload" http header This greatly increases the priority of loading the logo on browsers that support rel="preload". Bug: T100999 Change-Id: I0738fcc0a575153dab65016fa87faaa9b8b97a9d
2015-06-01 16:58:42 +00:00
/**
* Link: header contents
*/
private $mLinkHeader = [];
Initial support for Content Security Policy, disabled by default The primary goal here is a defense in depth measure to stop an attacker who found a bug in the parser allowing them to insert malicious attributes. This wouldn't stop someone who could insert a full script tag (since at current it can't distinguish between malicious and legit user js). It also would not prevent DOM-based or reflected XSS for anons, as the nonce value is guessable for anons when receiving a response cached by varnish. However, the limited protection of just stopping stored XSS where the attacker only has control of attributes, is still a big win in my opinion. (But it wouldn't prevent someone who has that type of xss from abusing things like data-ooui attribute). This will likely break many gadgets. Its expected that any sort of rollout on Wikimedia will be done very slowly, with lots of testing and the report-only option to begin with. This is behind feature flags that are off by default, so merging this patch should not cause any change in default behaviour. This may break some extensions (The most obvious one is charinsert (See fe648d41005), but will probably need some testing in report-only mode to see if anything else breaks) This uses the unsafe-eval option of CSP, in order to support RL's local storage thingy. For better security, we may want to remove some of the sillier uses of eval (e.g. jquery.ui.datepicker.js). For more info, see spec: https://www.w3.org/TR/CSP2/ Additionally see: https://www.mediawiki.org/wiki/Requests_for_comment/Content-Security-Policy Bug: T135963 Change-Id: I80f6f469ba4c0b608385483457df96ccb7429ae5
2016-02-29 04:13:10 +00:00
/**
Mild refactoring of ContentSecurityPolicy This is to make it behave in a more object orientied way. The goal is to make it be easier to allow extensions to mark certain pages as requiring a different policy (For example, CodeEditor extension uses a blob: url with a WebWorker. We don't want to include that on the policy of every page, but allow the extension to mark it as required whenever needed). This commit does not change code behaviour in any way. Change-Id: I4bf53dabb6e6c5446cea99a64db68b300cef2fd4
2019-10-28 05:01:17 +00:00
* @var ContentSecurityPolicy
Initial support for Content Security Policy, disabled by default The primary goal here is a defense in depth measure to stop an attacker who found a bug in the parser allowing them to insert malicious attributes. This wouldn't stop someone who could insert a full script tag (since at current it can't distinguish between malicious and legit user js). It also would not prevent DOM-based or reflected XSS for anons, as the nonce value is guessable for anons when receiving a response cached by varnish. However, the limited protection of just stopping stored XSS where the attacker only has control of attributes, is still a big win in my opinion. (But it wouldn't prevent someone who has that type of xss from abusing things like data-ooui attribute). This will likely break many gadgets. Its expected that any sort of rollout on Wikimedia will be done very slowly, with lots of testing and the report-only option to begin with. This is behind feature flags that are off by default, so merging this patch should not cause any change in default behaviour. This may break some extensions (The most obvious one is charinsert (See fe648d41005), but will probably need some testing in report-only mode to see if anything else breaks) This uses the unsafe-eval option of CSP, in order to support RL's local storage thingy. For better security, we may want to remove some of the sillier uses of eval (e.g. jquery.ui.datepicker.js). For more info, see spec: https://www.w3.org/TR/CSP2/ Additionally see: https://www.mediawiki.org/wiki/Requests_for_comment/Content-Security-Policy Bug: T135963 Change-Id: I80f6f469ba4c0b608385483457df96ccb7429ae5
2016-02-29 04:13:10 +00:00
*/
Mild refactoring of ContentSecurityPolicy This is to make it behave in a more object orientied way. The goal is to make it be easier to allow extensions to mark certain pages as requiring a different policy (For example, CodeEditor extension uses a blob: url with a WebWorker. We don't want to include that on the policy of every page, but allow the extension to mark it as required whenever needed). This commit does not change code behaviour in any way. Change-Id: I4bf53dabb6e6c5446cea99a64db68b300cef2fd4
2019-10-28 05:01:17 +00:00
private $CSP;
Initial support for Content Security Policy, disabled by default The primary goal here is a defense in depth measure to stop an attacker who found a bug in the parser allowing them to insert malicious attributes. This wouldn't stop someone who could insert a full script tag (since at current it can't distinguish between malicious and legit user js). It also would not prevent DOM-based or reflected XSS for anons, as the nonce value is guessable for anons when receiving a response cached by varnish. However, the limited protection of just stopping stored XSS where the attacker only has control of attributes, is still a big win in my opinion. (But it wouldn't prevent someone who has that type of xss from abusing things like data-ooui attribute). This will likely break many gadgets. Its expected that any sort of rollout on Wikimedia will be done very slowly, with lots of testing and the report-only option to begin with. This is behind feature flags that are off by default, so merging this patch should not cause any change in default behaviour. This may break some extensions (The most obvious one is charinsert (See fe648d41005), but will probably need some testing in report-only mode to see if anything else breaks) This uses the unsafe-eval option of CSP, in order to support RL's local storage thingy. For better security, we may want to remove some of the sillier uses of eval (e.g. jquery.ui.datepicker.js). For more info, see spec: https://www.w3.org/TR/CSP2/ Additionally see: https://www.mediawiki.org/wiki/Requests_for_comment/Content-Security-Policy Bug: T135963 Change-Id: I80f6f469ba4c0b608385483457df96ccb7429ae5
2016-02-29 04:13:10 +00:00
Increase OutputPage test coverage to >45% Also stop returning a value from OutputPage::adaptCdnTTL(). There are no users and the value doesn't seem very helpful: one would probably expect it to return the new value of mCdnMaxage, but instead it returns the new value of mCdnMaxageLimit. Better to have no return value than one that's easily misunderstood (*and* which nobody uses). Change-Id: Ia9dab86923b839334eab9f6fde17c4aed52130ec
2018-08-02 18:42:17 +00:00
/**
* @var array A cache of the names of the cookies that will influence the cache
*/
private static $cacheVaryCookies = null;
Instead of creating an OutputPage and then setting a context start initializing OutputPages 'with' a context and send depreciated calls to extensions directly creating OutputPage instances. Now that we've taken care of the mess of mTitle inside OutputPage and Skin and made RequestContext track Skin instead of $wgUser we don't need the hack in SpecialPage anymore.
2011-04-04 00:37:42 +00:00
/**
* Constructor for OutputPage. This should not be called directly.
* Instead a new RequestContext should be created and it will implicitly create
* a OutputPage tied to that context.
Simplify OutputPage constructor Make it impossible to instantiate this class without passing it context. This appears to be the oldest wfDeprecated() in MediaWiki. All callers have been updated. Change-Id: I00b9e283ad22c6274b358063b9f9e60fe93d5ee7
2017-09-09 02:12:37 +00:00
* @param IContextSource $context
Instead of creating an OutputPage and then setting a context start initializing OutputPages 'with' a context and send depreciated calls to extensions directly creating OutputPage instances. Now that we've taken care of the mess of mTitle inside OutputPage and Skin and made RequestContext track Skin instead of $wgUser we don't need the hack in SpecialPage anymore.
2011-04-04 00:37:42 +00:00
*/
OutputPage: Add visibility to various methods & do some micro-optimizations Methods that visibility was added to are; `addMeta()`, `addLink()`, `setCanonicalUrl()`, `addScript()`, `getHeadItemsArray()`, `addParserOutput()`, `getCacheVaryCookies()` and `haveCacheVaryCookies()`. Last but not lease, did a few micro-optimizations to `addMeta()` and `addLink()`. Change-Id: I94d037a5edc7131627724fd1d864000128077b0c
2019-09-14 10:26:31 +00:00
public function __construct( IContextSource $context ) {
Simplify OutputPage constructor Make it impossible to instantiate this class without passing it context. This appears to be the oldest wfDeprecated() in MediaWiki. All callers have been updated. Change-Id: I00b9e283ad22c6274b358063b9f9e60fe93d5ee7
2017-09-09 02:12:37 +00:00
$this->setContext( $context );
Mild refactoring of ContentSecurityPolicy This is to make it behave in a more object orientied way. The goal is to make it be easier to allow extensions to mark certain pages as requiring a different policy (For example, CodeEditor extension uses a blob: url with a WebWorker. We don't want to include that on the policy of every page, but allow the extension to mark it as required whenever needed). This commit does not change code behaviour in any way. Change-Id: I4bf53dabb6e6c5446cea99a64db68b300cef2fd4
2019-10-28 05:01:17 +00:00
$this->CSP = new ContentSecurityPolicy(
$context->getRequest()->response(),
Hooks::run() call site migration Migrate all callers of Hooks::run() to use the new HookContainer/HookRunner system. General principles: * Use DI if it is already used. We're not changing the way state is managed in this patch. * HookContainer is always injected, not HookRunner. HookContainer is a service, it's a more generic interface, it is the only thing that provides isRegistered() which is needed in some cases, and a HookRunner can be efficiently constructed from it (confirmed by benchmark). Because HookContainer is needed for object construction, it is also needed by all factories. * "Ask your friendly local base class". Big hierarchies like SpecialPage and ApiBase have getHookContainer() and getHookRunner() methods in the base class, and classes that extend that base class are not expected to know or care where the base class gets its HookContainer from. * ProtectedHookAccessorTrait provides protected getHookContainer() and getHookRunner() methods, getting them from the global service container. The point of this is to ease migration to DI by ensuring that call sites ask their local friendly base class rather than getting a HookRunner from the service container directly. * Private $this->hookRunner. In some smaller classes where accessor methods did not seem warranted, there is a private HookRunner property which is accessed directly. Very rarely (two cases), there is a protected property, for consistency with code that conventionally assumes protected=private, but in cases where the class might actually be overridden, a protected accessor is preferred over a protected property. * The last resort: Hooks::runner(). Mostly for static, file-scope and global code. In a few cases it was used for objects with broken construction schemes, out of horror or laziness. Constructors with new required arguments: * AuthManager * BadFileLookup * BlockManager * ClassicInterwikiLookup * ContentHandlerFactory * ContentSecurityPolicy * DefaultOptionsManager * DerivedPageDataUpdater * FullSearchResultWidget * HtmlCacheUpdater * LanguageFactory * LanguageNameUtils * LinkRenderer * LinkRendererFactory * LocalisationCache * MagicWordFactory * MessageCache * NamespaceInfo * PageEditStash * PageHandlerFactory * PageUpdater * ParserFactory * PermissionManager * RevisionStore * RevisionStoreFactory * SearchEngineConfig * SearchEngineFactory * SearchFormWidget * SearchNearMatcher * SessionBackend * SpecialPageFactory * UserNameUtils * UserOptionsManager * WatchedItemQueryService * WatchedItemStore Constructors with new optional arguments: * DefaultPreferencesFactory * Language * LinkHolderArray * MovePage * Parser * ParserCache * PasswordReset * Router setHookContainer() now required after construction: * AuthenticationProvider * ResourceLoaderModule * SearchEngine Change-Id: Id442b0dbe43aba84bd5cf801d86dedc768b082c7
2020-03-19 02:42:09 +00:00
$context->getConfig(),
$this->getHookContainer()
Mild refactoring of ContentSecurityPolicy This is to make it behave in a more object orientied way. The goal is to make it be easier to allow extensions to mark certain pages as requiring a different policy (For example, CodeEditor extension uses a blob: url with a WebWorker. We don't want to include that on the policy of every page, but allow the extension to mark it as required whenever needed). This commit does not change code behaviour in any way. Change-Id: I4bf53dabb6e6c5446cea99a64db68b300cef2fd4
2019-10-28 05:01:17 +00:00
);
Instead of creating an OutputPage and then setting a context start initializing OutputPages 'with' a context and send depreciated calls to extensions directly creating OutputPage instances. Now that we've taken care of the mess of mTitle inside OutputPage and Skin and made RequestContext track Skin instead of $wgUser we don't need the hack in SpecialPage anymore.
2011-04-04 00:37:42 +00:00
}
Step one in fixing OutputPage's documentation: * Groupped methods by function * Added documentation * Fixed some doxygen warnings * Some little style fixes (more to come)
2010-01-28 21:58:50 +00:00
/**
* Redirect to $url rather than displaying the normal page
*
Remove @param comments that literally repeat what the code says These comments do not add anything. I argue they are worse than having no comments, because I have to read them first to understand they actually don't explain anything. Removing them makes room for actual improvements in the future (if needed). Change-Id: Iee70aad681b3385e9af282d5581c10addbb91ac4
2017-12-28 15:06:10 +00:00
* @param string $url
Fix more scalar types in docblocks Change-Id: I574d4e261ab986e028c3ce26c4f0ec648b88a2ac
2019-12-08 12:37:24 +00:00
* @param string|int $responsecode HTTP status code
Step one in fixing OutputPage's documentation: * Groupped methods by function * Added documentation * Fixed some doxygen warnings * Some little style fixes (more to come)
2010-01-28 21:58:50 +00:00
*/
Add some documentation to OutputPage.php and specify public/private for most of its member functions (practically all public anyway). And replace one gratuitous use of strcmp with good old ===.
2006-11-07 05:37:31 +00:00
public function redirect( $url, $responsecode = '302' ) {
Added check for newlines in redirects as a paranoia guard against header injection in PHP<5.1.2. Removed OutputPage::addHeader(), has been non-functioning for as long as I can remember.
2006-06-23 03:56:03 +00:00
# Strip newlines as a paranoia check for header injection in PHP<5.1.2
$this->mRedirect = str_replace( "\n", '', $url );
Fix more scalar types in docblocks Change-Id: I574d4e261ab986e028c3ce26c4f0ec648b88a2ac
2019-12-08 12:37:24 +00:00
$this->mRedirectCode = (string)$responsecode;
Added check for newlines in redirects as a paranoia guard against header injection in PHP<5.1.2. Removed OutputPage::addHeader(), has been non-functioning for as long as I can remember.
2006-06-23 03:56:03 +00:00
}
* remove end of line whitespace * remove empty lines at end of file * remove "?>" where still present
2008-04-14 07:45:50 +00:00
Step one in fixing OutputPage's documentation: * Groupped methods by function * Added documentation * Fixed some doxygen warnings * Some little style fixes (more to come)
2010-01-28 21:58:50 +00:00
/**
* Get the URL to redirect to, or an empty string if not redirect URL set
*
Fixed some @params documentation Swapped some "$var type" to "type $var" or added missing types before the $var. Changed some other types to match the more common spelling. Makes beginning of some text in captial. Change-Id: I64e8cfe478cb0ba438f40b0631d6e9049cdab567
2014-04-10 18:50:10 +00:00
* @return string
Step one in fixing OutputPage's documentation: * Groupped methods by function * Added documentation * Fixed some doxygen warnings * Some little style fixes (more to come)
2010-01-28 21:58:50 +00:00
*/
The uncontroversial Liquid Threads hooks. The hook SkinTemplateTabAction got left in from a previous attempted commit, but the documentation is added in this commit. The other new hooks here are ChangesListInsertArticleLink, MediaWikiPerformAction, and BeforeWatchlist.
2007-11-06 01:16:25 +00:00
public function getRedirect() {
return $this->mRedirect;
}
Initial revision
2003-04-14 23:10:40 +00:00
Fix rel="copyright" for ApiHelp Bug: T94038 Bug: T93995 Change-Id: I06e9bb6797c4e4ee00e5d72eda1e273800b2c7f4
2015-06-08 13:39:06 +00:00
/**
* Set the copyright URL to send with the output.
* Empty string to omit, null to reset.
*
* @since 1.26
*
* @param string|null $url
*/
public function setCopyrightUrl( $url ) {
$this->copyrightUrl = $url;
}
Add some documentation to OutputPage.php and specify public/private for most of its member functions (practically all public anyway). And replace one gratuitous use of strcmp with good old ===.
2006-11-07 05:37:31 +00:00
/**
* Set the HTTP status code to send with the output.
*
Fixed some @params documentation Swapped some "$var type" to "type $var" or added missing types before the $var. Changed some other types to match the more common spelling. Makes beginning of some text in captial. Change-Id: I64e8cfe478cb0ba438f40b0631d6e9049cdab567
2014-04-10 18:50:10 +00:00
* @param int $statusCode
Add some documentation to OutputPage.php and specify public/private for most of its member functions (practically all public anyway). And replace one gratuitous use of strcmp with good old ===.
2006-11-07 05:37:31 +00:00
*/
Step one in fixing OutputPage's documentation: * Groupped methods by function * Added documentation * Fixed some doxygen warnings * Some little style fixes (more to come)
2010-01-28 21:58:50 +00:00
public function setStatusCode( $statusCode ) {
$this->mStatusCode = $statusCode;
}
Tweaks for skins: * Moved SkinTemplate::addStyle() and related stuff to OutputPage so that it can be used non-SkinTemplate skins and avoid duplication with the actual OutputPage::addStyle() (the two functions have the same format). * Non-SkinTemplate skins now also load their CSS with <link> tags instead of @import. * Moved SkinTemplate::setupUserCss() to Skin. * Merged action=raw&gen=(js|css) for SkinTemplate and non-SkinTemplate skins, renamed functions to Skin::generateUserJs() and Skin::generateUserStyleSheet() and dropped a lot of cascading call which is a bit incomprehensible.
2008-08-21 14:09:57 +00:00
/**
escape tags and entity in doxygen comments When inserting XML elements inline <such as this one>, doxygen chokes about it not being known. Simply enclosing the tag in double quotes prevents doxygen from emitting a warning. Also enclosed a few invalid functions calls such as \. and double quoted the HTML entities such as &foobar; Change-Id: I4019637145e683c2bec3d17b2fd98b0c50a932f1
2012-07-10 12:48:06 +00:00
* Add a new "<meta>" tag
Tweaks for skins: * Moved SkinTemplate::addStyle() and related stuff to OutputPage so that it can be used non-SkinTemplate skins and avoid duplication with the actual OutputPage::addStyle() (the two functions have the same format). * Non-SkinTemplate skins now also load their CSS with <link> tags instead of @import. * Moved SkinTemplate::setupUserCss() to Skin. * Merged action=raw&gen=(js|css) for SkinTemplate and non-SkinTemplate skins, renamed functions to Skin::generateUserJs() and Skin::generateUserStyleSheet() and dropped a lot of cascading call which is a bit incomprehensible.
2008-08-21 14:09:57 +00:00
* To add an http-equiv meta tag, precede the name with "http:"
*
Improve (weak and strict) type hints invarious places Change-Id: I3a42ec1547fae971d7b495c99dd144739d8521ab
2017-12-28 15:31:56 +00:00
* @param string $name Name of the meta tag
* @param string $val Value of the meta tag
Tweaks for skins: * Moved SkinTemplate::addStyle() and related stuff to OutputPage so that it can be used non-SkinTemplate skins and avoid duplication with the actual OutputPage::addStyle() (the two functions have the same format). * Non-SkinTemplate skins now also load their CSS with <link> tags instead of @import. * Moved SkinTemplate::setupUserCss() to Skin. * Merged action=raw&gen=(js|css) for SkinTemplate and non-SkinTemplate skins, renamed functions to Skin::generateUserJs() and Skin::generateUserStyleSheet() and dropped a lot of cascading call which is a bit incomprehensible.
2008-08-21 14:09:57 +00:00
*/
OutputPage: Add visibility to various methods & do some micro-optimizations Methods that visibility was added to are; `addMeta()`, `addLink()`, `setCanonicalUrl()`, `addScript()`, `getHeadItemsArray()`, `addParserOutput()`, `getCacheVaryCookies()` and `haveCacheVaryCookies()`. Last but not lease, did a few micro-optimizations to `addMeta()` and `addLink()`. Change-Id: I94d037a5edc7131627724fd1d864000128077b0c
2019-09-14 10:26:31 +00:00
public function addMeta( $name, $val ) {
$this->mMetatags[] = [ $name, $val ];
Tweaks for skins: * Moved SkinTemplate::addStyle() and related stuff to OutputPage so that it can be used non-SkinTemplate skins and avoid duplication with the actual OutputPage::addStyle() (the two functions have the same format). * Non-SkinTemplate skins now also load their CSS with <link> tags instead of @import. * Moved SkinTemplate::setupUserCss() to Skin. * Merged action=raw&gen=(js|css) for SkinTemplate and non-SkinTemplate skins, renamed functions to Skin::generateUserJs() and Skin::generateUserStyleSheet() and dropped a lot of cascading call which is a bit incomprehensible.
2008-08-21 14:09:57 +00:00
}
OutputPage: Add accessors for some protected properties Bug: T76168 Change-Id: Iee11c217b6b0d35acfdfac916814521acda2b1a4
2014-12-03 17:40:15 +00:00
/**
* Returns the current <meta> tags
*
* @since 1.25
* @return array
*/
public function getMetaTags() {
return $this->mMetatags;
}
Step one in fixing OutputPage's documentation: * Groupped methods by function * Added documentation * Fixed some doxygen warnings * Some little style fixes (more to come)
2010-01-28 21:58:50 +00:00
/**
Introduce $wgEnableCanonicalServerLink (bug 43466) If enabled, send a rel=canonical link on every page indicating the correct server to use. I tested all three callers: Article, ImagePage and the variant feature in getHeadLinksArray(). Change-Id: I59b7c1d5589825ea390967036190218d5ce2db88
2012-12-31 01:11:43 +00:00
* Add a new \<link\> tag to the page header.
*
* Note: use setCanonicalUrl() for rel=canonical.
Step one in fixing OutputPage's documentation: * Groupped methods by function * Added documentation * Fixed some doxygen warnings * Some little style fixes (more to come)
2010-01-28 21:58:50 +00:00
*
Cleanup some docs (includes/*.php) - Swap "$variable type" to "type $variable" - Added missing types - Fixed spacing inside docs - Makes beginning of @param/@return/@var/@throws in capital - Changed some types to match the more common spelling Change-Id: I783e4dbfe5f6f98b32b9a03ccf6439e13e132bcc
2014-07-24 17:42:24 +00:00
* @param array $linkarr Associative array of attributes.
Step one in fixing OutputPage's documentation: * Groupped methods by function * Added documentation * Fixed some doxygen warnings * Some little style fixes (more to come)
2010-01-28 21:58:50 +00:00
*/
OutputPage: Add visibility to various methods & do some micro-optimizations Methods that visibility was added to are; `addMeta()`, `addLink()`, `setCanonicalUrl()`, `addScript()`, `getHeadItemsArray()`, `addParserOutput()`, `getCacheVaryCookies()` and `haveCacheVaryCookies()`. Last but not lease, did a few micro-optimizations to `addMeta()` and `addLink()`. Change-Id: I94d037a5edc7131627724fd1d864000128077b0c
2019-09-14 10:26:31 +00:00
public function addLink( array $linkarr ) {
$this->mLinktags[] = $linkarr;
Step one in fixing OutputPage's documentation: * Groupped methods by function * Added documentation * Fixed some doxygen warnings * Some little style fixes (more to come)
2010-01-28 21:58:50 +00:00
}
OutputPage: Add accessors for some protected properties Bug: T76168 Change-Id: Iee11c217b6b0d35acfdfac916814521acda2b1a4
2014-12-03 17:40:15 +00:00
/**
* Returns the current <link> tags
*
* @since 1.25
* @return array
*/
public function getLinkTags() {
return $this->mLinktags;
}
Introduce $wgEnableCanonicalServerLink (bug 43466) If enabled, send a rel=canonical link on every page indicating the correct server to use. I tested all three callers: Article, ImagePage and the variant feature in getHeadLinksArray(). Change-Id: I59b7c1d5589825ea390967036190218d5ce2db88
2012-12-31 01:11:43 +00:00
/**
* Set the URL to be used for the <link rel=canonical>. This should be used
* in preference to addLink(), to avoid duplicate link tags.
Add missing @param to function docs Change-Id: I47fa96a976f55a1a93cb75397285edb8c7f4cd8a
2014-08-14 18:22:52 +00:00
* @param string $url
Introduce $wgEnableCanonicalServerLink (bug 43466) If enabled, send a rel=canonical link on every page indicating the correct server to use. I tested all three callers: Article, ImagePage and the variant feature in getHeadLinksArray(). Change-Id: I59b7c1d5589825ea390967036190218d5ce2db88
2012-12-31 01:11:43 +00:00
*/
OutputPage: Add visibility to various methods & do some micro-optimizations Methods that visibility was added to are; `addMeta()`, `addLink()`, `setCanonicalUrl()`, `addScript()`, `getHeadItemsArray()`, `addParserOutput()`, `getCacheVaryCookies()` and `haveCacheVaryCookies()`. Last but not lease, did a few micro-optimizations to `addMeta()` and `addLink()`. Change-Id: I94d037a5edc7131627724fd1d864000128077b0c
2019-09-14 10:26:31 +00:00
public function setCanonicalUrl( $url ) {
Introduce $wgEnableCanonicalServerLink (bug 43466) If enabled, send a rel=canonical link on every page indicating the correct server to use. I tested all three callers: Article, ImagePage and the variant feature in getHeadLinksArray(). Change-Id: I59b7c1d5589825ea390967036190218d5ce2db88
2012-12-31 01:11:43 +00:00
$this->mCanonicalUrl = $url;
}
OutputPage: Add accessors for some protected properties Bug: T76168 Change-Id: Iee11c217b6b0d35acfdfac916814521acda2b1a4
2014-12-03 17:40:15 +00:00
/**
* Returns the URL to be used for the <link rel=canonical> if
* one is set.
*
* @since 1.25
* @return bool|string
*/
public function getCanonicalUrl() {
return $this->mCanonicalUrl;
}
Step one in fixing OutputPage's documentation: * Groupped methods by function * Added documentation * Fixed some doxygen warnings * Some little style fixes (more to come)
2010-01-28 21:58:50 +00:00
/**
* Add raw HTML to the list of scripts (including \<script\> tag, etc.)
OutputPage: documentation for (semi-)deprecated methods Deprecated: addExtensionStyle, getExtStyle These had been deprecated unofficially for a while now. ResourceLoader is preferred. Internal/Unpreferred: addHeadItem, addScript, addScriptFile, addInlineScript, addStyle, addInlineStyle. All should be used as littel as possible, in favor of ResourceLoader and a few other higher level accessors. They are not truly deprecated, but should not really be used unless you have a really good reason. Also added a few missing wfDeprecated calls to methods marked as deprecated in the phpdoc. Bug: T89955 Change-Id: I472d906b6de16257fa34a2f9bfe9e4eb284d02b4
2015-10-31 15:16:58 +00:00
* Internal use only. Use OutputPage::addModules() or OutputPage::addJsConfigVars()
* if possible.
Step one in fixing OutputPage's documentation: * Groupped methods by function * Added documentation * Fixed some doxygen warnings * Some little style fixes (more to come)
2010-01-28 21:58:50 +00:00
*
Cleanup some docs (includes/*.php) - Swap "$variable type" to "type $variable" - Added missing types - Fixed spacing inside docs - Makes beginning of @param/@return/@var/@throws in capital - Changed some types to match the more common spelling Change-Id: I783e4dbfe5f6f98b32b9a03ccf6439e13e132bcc
2014-07-24 17:42:24 +00:00
* @param string $script Raw HTML
Step one in fixing OutputPage's documentation: * Groupped methods by function * Added documentation * Fixed some doxygen warnings * Some little style fixes (more to come)
2010-01-28 21:58:50 +00:00
*/
OutputPage: Add visibility to various methods & do some micro-optimizations Methods that visibility was added to are; `addMeta()`, `addLink()`, `setCanonicalUrl()`, `addScript()`, `getHeadItemsArray()`, `addParserOutput()`, `getCacheVaryCookies()` and `haveCacheVaryCookies()`. Last but not lease, did a few micro-optimizations to `addMeta()` and `addLink()`. Change-Id: I94d037a5edc7131627724fd1d864000128077b0c
2019-09-14 10:26:31 +00:00
public function addScript( $script ) {
OutputPage: Minor clean up of <head> and HTML * Remove trailing space in self-closing tag. Brings parserTest output of Parser and Parsoid closer together. * Remove various line breaks at begin and end of script contents. * Remove FILTER_NOMIN from makeConfigSetScript() output. This isn't part of any user- or page-dependent module and not minified. And Xml::encodeJsCall already ensures compact output for prod mode. Bug: T127328 Change-Id: I85a5a59fd0955c1a112e8b24b933f0d9e983a156
2016-02-18 16:33:15 +00:00
$this->mScripts .= $script;
fixed bug in addClass append
2009-08-07 00:16:54 +00:00
}
here it is ... the upload-api, script-server, js2 (javascript phase2) branch merge 1st attempt. Here is a short overview of changes and associated default configuration variables (most everything is off by default) also see ~soon to be updated~: http://www.mediawiki.org/wiki/Media_Projects_Overview = Upload Improvements = ==Upload API == * Based on the early work of Bryan Tong and others it adds the upload option to the api. * We rewrite Special:Upload page to include use the new refactoring * Added in token checks in both the SpecialUpload.php page so avoids DOS / xss copy-by-url JavaScript based cross site POST file submissions == Copy by URL== $wgAllowCopyUploads = false; * http class rewrite includes a new http background download see: includes/HttpFunctions.php * spins off a php process that calls: maintenance/http_session_download.php * pushes updates to the session and gives the user a progress bar on http copy uploads from other server progress (using js2 upload interface) (if not using the js2 upload interface it does the request in-place but the download is limited to the php ini timeout time) == Firefogg == * Firefogg enables resumable upload by chunks * progress indicators and conditional invokation (js2 system) * and of-course client side transcoding. = Script Server = $wgEnableScriptLoader = false; * off by default if $wgEnableScriptLoader is turned on script files are grouped, gziped, cached etc. for more info see: http://www.mediawiki.org/wiki/Extension:ScriptLoader * Includes some early skin js include fixes (skin/script system still lots of love) * Includes a "javascript class autoloader" this is packaged into mwEmbed so that the mwEmbed library can work in stand alone mode (while retaining localization and script serving) (one such application is the make page for firefogg.org : http://www.firefogg.org/make/index.html ) * The file that contains the autojavascript loading classes is: js2/php/jsAutoloadLocalClasses.php * One can use this auto class loading dependency system with extensions and add-ons but I need to better document that. = js2 system / mwEmbed= $wgEnableJS2system = false * includes initial rewrite towards more jquery based javascript code * especially for the Special:Upload page. * Also the edit page include support for the "add-media-wizard" * includes dependency loader for javascript that optionally takes advantage of the script-loader * remote embedding of javascript interfaces (like embedding video, or commons media searching) * $wgDebugJavaScript = false; .. .this variable lets you always get "always fresh javascript". When used with the script-loader it does not minify the script-loader output. = mwEmbed = * Will commit a separate patch to oggHandler that conditionally outputs <video tag> to use the new javascript video player. ** mv_embed player includes: play-head, volume control, remote embedding, oggz-chop support across plugins. * add-media-wizard adds easy inserts of media to pages (with import) == jQuery== * we include a base install of jQuery, jQuery ui and some plugins. * all the javascript classes are in the scriptloader so its easy to load any set of jquery ui components that you may need using the script-server. You get a callback so you can then execute js with dependencies loaded. == other stuff == there is a bit more code in js2 that pertains to sequence editing, timed text display and basic image editing. We include a base import of pixastic-lib & pixastic-editor... will work with the pixastic developer to try and ensure upstream compatibility on our usage of the library for in-browser photo and sequence manipulation.
2009-07-14 23:52:14 +00:00
* add OutputPage::addScriptFile() for easier adding of JS files, similar to OutputPage::addStyle() * broke Special:Allmessages-specific JS out of wikibits.js to allmessages.js
2008-05-09 20:18:35 +00:00
/**
OutputPage: Remove support for non-existent /w/skins/common directory Test Plan: * Add calls `$this->addScriptFile( 'example.js' )` and `$this->addScriptFile( '/example.js' )` to top of OutputPage::output(). * Without this change, two `<script>` are added. * With this change, only for the latter a script is added, and the former triggers a deprecation warning. Bug: T71277 Bug: T181318 Change-Id: I0576ef09fafa4ba34d52d75f4211fcfa28f4f3b0
2018-05-21 21:08:02 +00:00
* Add a JavaScript file to be loaded as `<script>` on this page.
*
OutputPage: documentation for (semi-)deprecated methods Deprecated: addExtensionStyle, getExtStyle These had been deprecated unofficially for a while now. ResourceLoader is preferred. Internal/Unpreferred: addHeadItem, addScript, addScriptFile, addInlineScript, addStyle, addInlineStyle. All should be used as littel as possible, in favor of ResourceLoader and a few other higher level accessors. They are not truly deprecated, but should not really be used unless you have a really good reason. Also added a few missing wfDeprecated calls to methods marked as deprecated in the phpdoc. Bug: T89955 Change-Id: I472d906b6de16257fa34a2f9bfe9e4eb284d02b4
2015-10-31 15:16:58 +00:00
* Internal use only. Use OutputPage::addModules() if possible.
Step one in fixing OutputPage's documentation: * Groupped methods by function * Added documentation * Fixed some doxygen warnings * Some little style fixes (more to come)
2010-01-28 21:58:50 +00:00
*
OutputPage: Remove support for non-existent /w/skins/common directory Test Plan: * Add calls `$this->addScriptFile( 'example.js' )` and `$this->addScriptFile( '/example.js' )` to top of OutputPage::output(). * Without this change, two `<script>` are added. * With this change, only for the latter a script is added, and the former triggers a deprecation warning. Bug: T71277 Bug: T181318 Change-Id: I0576ef09fafa4ba34d52d75f4211fcfa28f4f3b0
2018-05-21 21:08:02 +00:00
* @param string $file URL to file (absolute path, protocol-relative, or full url)
Fix PhanTypeMismatchDeclaredParam Auto fix MediaWiki.Commenting.FunctionComment.DefaultNullTypeParam sniff Change-Id: I865323fd0295aabd06f3e3c75e0e5043fb31069e
2018-06-26 21:14:43 +00:00
* @param string|null $unused Previously used to change the cache-busting query parameter
* add OutputPage::addScriptFile() for easier adding of JS files, similar to OutputPage::addStyle() * broke Special:Allmessages-specific JS out of wikibits.js to allmessages.js
2008-05-09 20:18:35 +00:00
*/
OutputPage: Remove appending of wgStyleVersion to legacy resources For addScriptFile(), just remove the appending of wgStyleVersion. Going forward, anyone still using this, should simply append a query parameter themselves in a way that is specific to that one url (instead of relying on a generic global variable). Alternatively, one could use OutputPage::transformResourcePath if the file is in /w/. For addStyles(), also remove the appending of wgStyleVersion. Since this method takes paths relative to /w/skins, we can easily update this to automatically use transformResourcePath(), so that file-hash based query parameters are automatically added. Test Plan: * Add calls to top of OutputPage::output(): `$this->addStyle( 'Vector/README.md' );` `$this->addScriptFile( "{$GLOBALS['wgScriptPath']}/composer.json" );` * Before, they are both inserted as `<link>` (head) and `<script>` (body) with a query parameter based on wgStyleVersion. * After, the `<script>` (end of body) has no query. After, the stylesheet (head) has a SHA1 content hash as query. Bug: T181318 Change-Id: Ie5ab5066ef7d07279086bde838d7305e9e4eabaf
2018-05-21 21:21:28 +00:00
public function addScriptFile( $file, $unused = null ) {
Mild refactoring of ContentSecurityPolicy This is to make it behave in a more object orientied way. The goal is to make it be easier to allow extensions to mark certain pages as requiring a different policy (For example, CodeEditor extension uses a blob: url with a WebWorker. We don't want to include that on the policy of every page, but allow the extension to mark it as required whenever needed). This commit does not change code behaviour in any way. Change-Id: I4bf53dabb6e6c5446cea99a64db68b300cef2fd4
2019-10-28 05:01:17 +00:00
$this->addScript( Html::linkedScript( $file, $this->CSP->getNonce() ) );
* add OutputPage::addScriptFile() for easier adding of JS files, similar to OutputPage::addStyle() * broke Special:Allmessages-specific JS out of wikibits.js to allmessages.js
2008-05-09 20:18:35 +00:00
}
follow-up to r53282: *coding style tweaks *added __METHOD__ to some wfDebug calls *removed php ending tag from mwScriptLoader.php *some doxygen docs added
2009-07-15 00:55:58 +00:00
* Added redirect to section feature. Use it wisely. * Added a configuration variable allowing the "break out of framesets" feature to be switched on and off. Off by default -- there's all sorts of flashy frameset gadgets around, outright abuse is more rare, especially for small default installs. * Refactored URL fragment handling. * Made Title::secureAndSplit() slightly more legible. * Refactored makeGlobalVariablesScript() -- it's not particularly convenient or maintainable to pass all global variables through the template array. The array is there for BC, not flexibility.
2006-12-08 06:09:15 +00:00
/**
* Add a self-contained script tag with the given contents
OutputPage: documentation for (semi-)deprecated methods Deprecated: addExtensionStyle, getExtStyle These had been deprecated unofficially for a while now. ResourceLoader is preferred. Internal/Unpreferred: addHeadItem, addScript, addScriptFile, addInlineScript, addStyle, addInlineStyle. All should be used as littel as possible, in favor of ResourceLoader and a few other higher level accessors. They are not truly deprecated, but should not really be used unless you have a really good reason. Also added a few missing wfDeprecated calls to methods marked as deprecated in the phpdoc. Bug: T89955 Change-Id: I472d906b6de16257fa34a2f9bfe9e4eb284d02b4
2015-10-31 15:16:58 +00:00
* Internal use only. Use OutputPage::addModules() if possible.
Step one in fixing OutputPage's documentation: * Groupped methods by function * Added documentation * Fixed some doxygen warnings * Some little style fixes (more to come)
2010-01-28 21:58:50 +00:00
*
resourceloader: Move queue formatting out of OutputPage HTML formatting of the queue was distributed over several OutputPage methods. Each method demanding a snippet of HTML by calling makeResourceLoaderLink() with a limited amount of information. As such, makeResourceLoaderLink() was unable to provide the client with the proper state information. Centralising it also allows it to better reduce duplication in HTML output and maintain a more accurate state. Problems fixed by centralising: 1. The 'user' module is special (due to per-user 'version' and 'user' params). It is manually requested via script-src. To avoid a separate (and wrong) request from something that requires it, we set state=loading directly. However, because the module is in the bottom, the old HTML formatter could only put state=loading in the bottom also. This sometimes caused a wrong request to be fired for modules=user if something in the top queue triggered a requirement for it. 2. Since a464d1d4 (T87871) we track states of page-style modules, with purpose of allowing dependencies on style modules without risking duplicate loading on pages where the styles are loaded already. This didn't work, because the state information about page-style modules is output near the stylesheet, which is after the script tag with mw.loader.load(). That runs first, and mw.loader would still make a duplicate request before it learns the state. Changes: * Document reasons for style/script tag order in getHeadHtml (per 09537e83). * Pass $type from getModuleStyles() to getAllowedModules(). This wasn't needed before since a duplicate check in makeResourceLoaderLink() verified the origin a second time. * Declare explicit position 'top' on 'user.options' and 'user.tokens' module. Previously, OutputPage hardcoded them in the top. The new formatter doesn't. * Remove getHeadScripts(). * Remove getInlineHeadScripts(). * Remove getExternalHeadScripts(). * Remove buildCssLinks(). * Remove getScriptsForBottomQueue(). * Change where Skin::setupSkinUserCss() is called. This methods lets the skin add modules to the queue. Previously it was called from buildCssLinks(), via headElement(), via prepareQuickTemplate(), via OutputPage::output(). It's now in OutputPage::output() directly (slightly earlier). This is needed because prepareQuickTemplate() calls bottomScripts() before headElement(). And bottomScript() would lazy-initialise the queue and lock it before setupSkinUserCss() is called from headElement(). This makes execution order more predictable instead of being dependent on the arbitrary order of data extraction in prepareQuickTemplate (which varies from one skin to another). * Compute isUserModulePreview() and isKnownEmpty() for the 'user' module early on so. This avoids wrongful loading and fixes problem 1. Effective changes in output: * mw.loader.state() is now before mw.loader.load(). This fixes problem 2. * mw.loader.state() now sets 'user.options' and 'user.tokens' to "loading". * mw.loader.state() now sets 'user' (as "loading" or "ready"). Fixes problem 1. * The <script async src> tag for 'startup' changed position (slightly). Previously it was after all inline scripts and stylesheets. It's still after all inline scripts and after most stylesheets, but before any user styles. Since the queue is now formatted outside OutputPage, it can't inject the meta-ResourceLoaderDynamicStyles tag and user-stylesheet hack in the middle of existing output. This shouldn't have any noticable impact. Bug: T87871 Change-Id: I605b8cd1e1fc009b4662a0edbc54d09dd65ee1df
2016-07-15 14:13:09 +00:00
* @param string $script JavaScript text, no script tags
* Added redirect to section feature. Use it wisely. * Added a configuration variable allowing the "break out of framesets" feature to be switched on and off. Off by default -- there's all sorts of flashy frameset gadgets around, outright abuse is more rare, especially for small default installs. * Refactored URL fragment handling. * Made Title::secureAndSplit() slightly more legible. * Refactored makeGlobalVariablesScript() -- it's not particularly convenient or maintainable to pass all global variables through the template array. The array is there for BC, not flexibility.
2006-12-08 06:09:15 +00:00
*/
Step one in fixing OutputPage's documentation: * Groupped methods by function * Added documentation * Fixed some doxygen warnings * Some little style fixes (more to come)
2010-01-28 21:58:50 +00:00
public function addInlineScript( $script ) {
Mild refactoring of ContentSecurityPolicy This is to make it behave in a more object orientied way. The goal is to make it be easier to allow extensions to mark certain pages as requiring a different policy (For example, CodeEditor extension uses a blob: url with a WebWorker. We don't want to include that on the policy of every page, but allow the extension to mark it as required whenever needed). This commit does not change code behaviour in any way. Change-Id: I4bf53dabb6e6c5446cea99a64db68b300cef2fd4
2019-10-28 05:01:17 +00:00
$this->mScripts .= Html::inlineScript( "\n$script\n", $this->CSP->getNonce() ) . "\n";
* Added redirect to section feature. Use it wisely. * Added a configuration variable allowing the "break out of framesets" feature to be switched on and off. Off by default -- there's all sorts of flashy frameset gadgets around, outright abuse is more rare, especially for small default installs. * Refactored URL fragment handling. * Made Title::secureAndSplit() slightly more legible. * Refactored makeGlobalVariablesScript() -- it's not particularly convenient or maintainable to pass all global variables through the template array. The array is there for BC, not flexibility.
2006-12-08 06:09:15 +00:00
}
Follow-up r64670 (bug22929): cleaner implementation of security for script (and potentially CSS) files. ResourceLoader *already* knows where each module has come from, so all we need to do is filter them in OutputPage according to the desired level of 'trustworthiness'. TODO: * Are there instances where we might want to restrict CSS as well as JS? * Would a $wg config option and/or user preference and/or index.php GET parameter to limit inclusion be useful? * Can we deprecate any of the existing $wg config options? * What's going on with the duplicated code between OutputPage and SkinTemplate?
2011-02-04 16:39:17 +00:00
/**
Follow-up r81524: fix fatal when modules are deregistered
2011-03-19 21:19:50 +00:00
* Filter an array of modules to remove insufficiently trustworthy members, and modules
* which are no longer registered (eg a page is cached before an extension is disabled)
Update a lot of unspecific "array" types in PHPDocs This includes fixing some mistakes, as well as removing redundant text that doesn't add new information, either because it literally repeats what the code already says, or is actually duplicated. Change-Id: I3a8dd8ce57192deda8916cc444c87d7ab1a36515
2020-10-28 10:01:33 +00:00
* @param string[] $modules
resourceloader: Remove ResourceLoaderModule::getPosition @deprecated since 1.29 Bug: T184257 Change-Id: I1a59a1f0a5fbedfea9f0b28bee8c298710557a8d
2018-02-12 21:06:05 +00:00
* @param string|null $position Unused
Fixed some @params documentation Swapped some "$var type" to "type $var" or added missing types before the $var. Changed some other types to match the more common spelling. Makes beginning of some text in captial. Change-Id: I64e8cfe478cb0ba438f40b0631d6e9049cdab567
2014-04-10 18:50:10 +00:00
* @param string $type
Update a lot of unspecific "array" types in PHPDocs This includes fixing some mistakes, as well as removing redundant text that doesn't add new information, either because it literally repeats what the code already says, or is actually duplicated. Change-Id: I3a8dd8ce57192deda8916cc444c87d7ab1a36515
2020-10-28 10:01:33 +00:00
* @return string[]
Follow-up r64670 (bug22929): cleaner implementation of security for script (and potentially CSS) files. ResourceLoader *already* knows where each module has come from, so all we need to do is filter them in OutputPage according to the desired level of 'trustworthiness'. TODO: * Are there instances where we might want to restrict CSS as well as JS? * Would a $wg config option and/or user preference and/or index.php GET parameter to limit inclusion be useful? * Can we deprecate any of the existing $wg config options? * What's going on with the duplicated code between OutputPage and SkinTemplate?
2011-02-04 16:39:17 +00:00
*/
Enforce array type hinting in OutputPage.php Change-Id: I2b37a95e517e120471c709823ea319f60165be0e
2014-07-23 02:49:30 +00:00
protected function filterModules( array $modules, $position = null,
ResourceLoader namespace (attempt 2) Move ResourceLoader classes to their own namespace. Strip the "ResourceLoader" prefix from all except ResourceLoader itself. Move the tests by analogy. I used a namespace alias "RL" in some callers since RL\Module is less ambiguous at the call site than just "Module". I did not address DependencyStore which continues to have a non-standard location and namespace. Revert of a241d83e0a6dabedf. Bug: T308718 Change-Id: Id08a220e1d6085e2b33f3f6c9d0e3935a4204659
2022-05-06 09:09:56 +00:00
$type = RL\Module::TYPE_COMBINED
Pass php-strict for OutputPage Change-Id: Ia5d03277fff543ca672dc8efc7eaaeca4888ac23
2014-05-11 01:54:15 +00:00
) {
Follow-up r64670 (bug22929): cleaner implementation of security for script (and potentially CSS) files. ResourceLoader *already* knows where each module has come from, so all we need to do is filter them in OutputPage according to the desired level of 'trustworthiness'. TODO: * Are there instances where we might want to restrict CSS as well as JS? * Would a $wg config option and/or user preference and/or index.php GET parameter to limit inclusion be useful? * Can we deprecate any of the existing $wg config options? * What's going on with the duplicated code between OutputPage and SkinTemplate?
2011-02-04 16:39:17 +00:00
$resourceLoader = $this->getResourceLoader();
Convert all array() syntax to [] Per wikitech-l consensus: https://lists.wikimedia.org/pipermail/wikitech-l/2016-February/084821.html Notes: * Disabled CallTimePassByReference due to false positives (T127163) Change-Id: I2c8ce713ce6600a0bb7bf67537c87044c7a45c4b
2016-02-17 09:09:32 +00:00
$filteredModules = [];
else if -> elseif Also fixed some spacing while at it Added some braces for one line statements Change-Id: Iebfa2b17091509daabc76248121c7763e2295d20
2013-04-17 14:52:47 +00:00
foreach ( $modules as $val ) {
Follow-up r64670 (bug22929): cleaner implementation of security for script (and potentially CSS) files. ResourceLoader *already* knows where each module has come from, so all we need to do is filter them in OutputPage according to the desired level of 'trustworthiness'. TODO: * Are there instances where we might want to restrict CSS as well as JS? * Would a $wg config option and/or user preference and/or index.php GET parameter to limit inclusion be useful? * Can we deprecate any of the existing $wg config options? * What's going on with the duplicated code between OutputPage and SkinTemplate?
2011-02-04 16:39:17 +00:00
$module = $resourceLoader->getModule( $val );
ResourceLoader namespace (attempt 2) Move ResourceLoader classes to their own namespace. Strip the "ResourceLoader" prefix from all except ResourceLoader itself. Move the tests by analogy. I used a namespace alias "RL" in some callers since RL\Module is less ambiguous at the call site than just "Module". I did not address DependencyStore which continues to have a non-standard location and namespace. Revert of a241d83e0a6dabedf. Bug: T308718 Change-Id: Id08a220e1d6085e2b33f3f6c9d0e3935a4204659
2022-05-06 09:09:56 +00:00
if ( $module instanceof RL\Module
For bug 27488: move the startup script, jquery+mediawiki and the mw.config.set() call for configuration variables back to the <head> . Let modules control whether they're loaded in the <head> ('top') or at the bottom of the <body> ('bottom') through the position parameter/property Also rearranges the loading order a little bit such that only=messages comes before only=scripts, and config comes before everything except startup and jquery+mediawiki
2011-04-07 12:07:25 +00:00
&& $module->getOrigin() <= $this->getAllowedModules( $type )
Move closing parenthesis from multi line if and function to own line The Line continuation Coding conventions prefers the closing parenthesis on the same line than the beginning curly braces. This is done for ifs and functions. Also move some boolean operator from the end of a line to the beginning and changed some indentation to make the condition hopefully better readable. Change-Id: Id0437b06bde86eb5a75bc59eefa19e7edb624426
2013-12-01 20:39:00 +00:00
) {
OutputPage: Log violations of module 'target' filter Bug: T140675 Change-Id: Icf9a2e72b663e363e8c87d5d464892c2bbbe78b2
2017-02-15 00:25:39 +00:00
if ( $this->mTarget && !in_array( $this->mTarget, $module->getTargets() ) ) {
$this->warnModuleTargetFilter( $module->getName() );
continue;
}
Follow-up r64670 (bug22929): cleaner implementation of security for script (and potentially CSS) files. ResourceLoader *already* knows where each module has come from, so all we need to do is filter them in OutputPage according to the desired level of 'trustworthiness'. TODO: * Are there instances where we might want to restrict CSS as well as JS? * Would a $wg config option and/or user preference and/or index.php GET parameter to limit inclusion be useful? * Can we deprecate any of the existing $wg config options? * What's going on with the duplicated code between OutputPage and SkinTemplate?
2011-02-04 16:39:17 +00:00
$filteredModules[] = $val;
}
}
return $filteredModules;
}
OutputPage: Log violations of module 'target' filter Bug: T140675 Change-Id: Icf9a2e72b663e363e8c87d5d464892c2bbbe78b2
2017-02-15 00:25:39 +00:00
private function warnModuleTargetFilter( $moduleName ) {
static $warnings = [];
if ( isset( $warnings[$this->mTarget][$moduleName] ) ) {
return;
}
$warnings[$this->mTarget][$moduleName] = true;
Revert "resourceloader: Raise severity of targets violation to WARNING" This reverts commit aa2a3574076d00b45b1174ec01488cb84e56abda. Reason for revert: This was a temporary change to collect data around target violations. One week should be enough data so we can revert this in the following train. Bug: T235712 Change-Id: I3a3db293299771ab9465258d1a2a73c62c652a9b
2022-12-05 17:56:03 +00:00
$this->getResourceLoader()->getLogger()->debug(
OutputPage: Log violations of module 'target' filter Bug: T140675 Change-Id: Icf9a2e72b663e363e8c87d5d464892c2bbbe78b2
2017-02-15 00:25:39 +00:00
'Module "{module}" not loadable on target "{target}".',
[
'module' => $moduleName,
'target' => $this->mTarget,
]
);
}
Merging resourceloader branch into trunk. Full documentation is at http://www.mediawiki.org/wiki/ResourceLoader and a general overview has been posted on wikitech-li <http://lists.wikimedia.org/pipermail/wikitech-l/2010-September/049253.html>. One important change is that all JS is now loaded at the bottom, so any scripts assuming things from wikibits or whatever are present will fail.
2010-09-04 04:00:09 +00:00
/**
* Get the list of modules to include on this page
Fixed some doxygen warnings
2010-09-07 21:05:09 +00:00
*
Fixed some @params documentation Swapped some "$var type" to "type $var" or added missing types before the $var. Changed some other types to match the more common spelling. Makes beginning of some text in captial. Change-Id: I64e8cfe478cb0ba438f40b0631d6e9049cdab567
2014-04-10 18:50:10 +00:00
* @param bool $filter Whether to filter out insufficiently trustworthy modules
resourceloader: Remove ResourceLoaderModule::getPosition @deprecated since 1.29 Bug: T184257 Change-Id: I1a59a1f0a5fbedfea9f0b28bee8c298710557a8d
2018-02-12 21:06:05 +00:00
* @param string|null $position Unused
Fixed some @params documentation Swapped some "$var type" to "type $var" or added missing types before the $var. Changed some other types to match the more common spelling. Makes beginning of some text in captial. Change-Id: I64e8cfe478cb0ba438f40b0631d6e9049cdab567
2014-04-10 18:50:10 +00:00
* @param string $param
Add missing @param and @return documentation Change-Id: I1d1098eec3933df6561cceef646576013ddc08c8
2017-08-11 18:04:11 +00:00
* @param string $type
Update a lot of unspecific "array" types in PHPDocs This includes fixing some mistakes, as well as removing redundant text that doesn't add new information, either because it literally repeats what the code already says, or is actually duplicated. Change-Id: I3a8dd8ce57192deda8916cc444c87d7ab1a36515
2020-10-28 10:01:33 +00:00
* @return string[] Array of module names
Merging resourceloader branch into trunk. Full documentation is at http://www.mediawiki.org/wiki/ResourceLoader and a general overview has been posted on wikitech-li <http://lists.wikimedia.org/pipermail/wikitech-l/2010-September/049253.html>. One important change is that all JS is now loaded at the bottom, so any scripts assuming things from wikibits or whatever are present will fail.
2010-09-04 04:00:09 +00:00
*/
resourceloader: Move queue formatting out of OutputPage HTML formatting of the queue was distributed over several OutputPage methods. Each method demanding a snippet of HTML by calling makeResourceLoaderLink() with a limited amount of information. As such, makeResourceLoaderLink() was unable to provide the client with the proper state information. Centralising it also allows it to better reduce duplication in HTML output and maintain a more accurate state. Problems fixed by centralising: 1. The 'user' module is special (due to per-user 'version' and 'user' params). It is manually requested via script-src. To avoid a separate (and wrong) request from something that requires it, we set state=loading directly. However, because the module is in the bottom, the old HTML formatter could only put state=loading in the bottom also. This sometimes caused a wrong request to be fired for modules=user if something in the top queue triggered a requirement for it. 2. Since a464d1d4 (T87871) we track states of page-style modules, with purpose of allowing dependencies on style modules without risking duplicate loading on pages where the styles are loaded already. This didn't work, because the state information about page-style modules is output near the stylesheet, which is after the script tag with mw.loader.load(). That runs first, and mw.loader would still make a duplicate request before it learns the state. Changes: * Document reasons for style/script tag order in getHeadHtml (per 09537e83). * Pass $type from getModuleStyles() to getAllowedModules(). This wasn't needed before since a duplicate check in makeResourceLoaderLink() verified the origin a second time. * Declare explicit position 'top' on 'user.options' and 'user.tokens' module. Previously, OutputPage hardcoded them in the top. The new formatter doesn't. * Remove getHeadScripts(). * Remove getInlineHeadScripts(). * Remove getExternalHeadScripts(). * Remove buildCssLinks(). * Remove getScriptsForBottomQueue(). * Change where Skin::setupSkinUserCss() is called. This methods lets the skin add modules to the queue. Previously it was called from buildCssLinks(), via headElement(), via prepareQuickTemplate(), via OutputPage::output(). It's now in OutputPage::output() directly (slightly earlier). This is needed because prepareQuickTemplate() calls bottomScripts() before headElement(). And bottomScript() would lazy-initialise the queue and lock it before setupSkinUserCss() is called from headElement(). This makes execution order more predictable instead of being dependent on the arbitrary order of data extraction in prepareQuickTemplate (which varies from one skin to another). * Compute isUserModulePreview() and isKnownEmpty() for the 'user' module early on so. This avoids wrongful loading and fixes problem 1. Effective changes in output: * mw.loader.state() is now before mw.loader.load(). This fixes problem 2. * mw.loader.state() now sets 'user.options' and 'user.tokens' to "loading". * mw.loader.state() now sets 'user' (as "loading" or "ready"). Fixes problem 1. * The <script async src> tag for 'startup' changed position (slightly). Previously it was after all inline scripts and stylesheets. It's still after all inline scripts and after most stylesheets, but before any user styles. Since the queue is now formatted outside OutputPage, it can't inject the meta-ResourceLoaderDynamicStyles tag and user-stylesheet hack in the middle of existing output. This shouldn't have any noticable impact. Bug: T87871 Change-Id: I605b8cd1e1fc009b4662a0edbc54d09dd65ee1df
2016-07-15 14:13:09 +00:00
public function getModules( $filter = false, $position = null, $param = 'mModules',
ResourceLoader namespace (attempt 2) Move ResourceLoader classes to their own namespace. Strip the "ResourceLoader" prefix from all except ResourceLoader itself. Move the tests by analogy. I used a namespace alias "RL" in some callers since RL\Module is less ambiguous at the call site than just "Module". I did not address DependencyStore which continues to have a non-standard location and namespace. Revert of a241d83e0a6dabedf. Bug: T308718 Change-Id: Id08a220e1d6085e2b33f3f6c9d0e3935a4204659
2022-05-06 09:09:56 +00:00
$type = RL\Module::TYPE_COMBINED
resourceloader: Move queue formatting out of OutputPage HTML formatting of the queue was distributed over several OutputPage methods. Each method demanding a snippet of HTML by calling makeResourceLoaderLink() with a limited amount of information. As such, makeResourceLoaderLink() was unable to provide the client with the proper state information. Centralising it also allows it to better reduce duplication in HTML output and maintain a more accurate state. Problems fixed by centralising: 1. The 'user' module is special (due to per-user 'version' and 'user' params). It is manually requested via script-src. To avoid a separate (and wrong) request from something that requires it, we set state=loading directly. However, because the module is in the bottom, the old HTML formatter could only put state=loading in the bottom also. This sometimes caused a wrong request to be fired for modules=user if something in the top queue triggered a requirement for it. 2. Since a464d1d4 (T87871) we track states of page-style modules, with purpose of allowing dependencies on style modules without risking duplicate loading on pages where the styles are loaded already. This didn't work, because the state information about page-style modules is output near the stylesheet, which is after the script tag with mw.loader.load(). That runs first, and mw.loader would still make a duplicate request before it learns the state. Changes: * Document reasons for style/script tag order in getHeadHtml (per 09537e83). * Pass $type from getModuleStyles() to getAllowedModules(). This wasn't needed before since a duplicate check in makeResourceLoaderLink() verified the origin a second time. * Declare explicit position 'top' on 'user.options' and 'user.tokens' module. Previously, OutputPage hardcoded them in the top. The new formatter doesn't. * Remove getHeadScripts(). * Remove getInlineHeadScripts(). * Remove getExternalHeadScripts(). * Remove buildCssLinks(). * Remove getScriptsForBottomQueue(). * Change where Skin::setupSkinUserCss() is called. This methods lets the skin add modules to the queue. Previously it was called from buildCssLinks(), via headElement(), via prepareQuickTemplate(), via OutputPage::output(). It's now in OutputPage::output() directly (slightly earlier). This is needed because prepareQuickTemplate() calls bottomScripts() before headElement(). And bottomScript() would lazy-initialise the queue and lock it before setupSkinUserCss() is called from headElement(). This makes execution order more predictable instead of being dependent on the arbitrary order of data extraction in prepareQuickTemplate (which varies from one skin to another). * Compute isUserModulePreview() and isKnownEmpty() for the 'user' module early on so. This avoids wrongful loading and fixes problem 1. Effective changes in output: * mw.loader.state() is now before mw.loader.load(). This fixes problem 2. * mw.loader.state() now sets 'user.options' and 'user.tokens' to "loading". * mw.loader.state() now sets 'user' (as "loading" or "ready"). Fixes problem 1. * The <script async src> tag for 'startup' changed position (slightly). Previously it was after all inline scripts and stylesheets. It's still after all inline scripts and after most stylesheets, but before any user styles. Since the queue is now formatted outside OutputPage, it can't inject the meta-ResourceLoaderDynamicStyles tag and user-stylesheet hack in the middle of existing output. This shouldn't have any noticable impact. Bug: T87871 Change-Id: I605b8cd1e1fc009b4662a0edbc54d09dd65ee1df
2016-07-15 14:13:09 +00:00
) {
Follow-up r64670 (bug22929): cleaner implementation of security for script (and potentially CSS) files. ResourceLoader *already* knows where each module has come from, so all we need to do is filter them in OutputPage according to the desired level of 'trustworthiness'. TODO: * Are there instances where we might want to restrict CSS as well as JS? * Would a $wg config option and/or user preference and/or index.php GET parameter to limit inclusion be useful? * Can we deprecate any of the existing $wg config options? * What's going on with the duplicated code between OutputPage and SkinTemplate?
2011-02-04 16:39:17 +00:00
$modules = array_values( array_unique( $this->$param ) );
return $filter
resourceloader: Remove ResourceLoaderModule::getPosition @deprecated since 1.29 Bug: T184257 Change-Id: I1a59a1f0a5fbedfea9f0b28bee8c298710557a8d
2018-02-12 21:06:05 +00:00
? $this->filterModules( $modules, null, $type )
Follow-up r64670 (bug22929): cleaner implementation of security for script (and potentially CSS) files. ResourceLoader *already* knows where each module has come from, so all we need to do is filter them in OutputPage according to the desired level of 'trustworthiness'. TODO: * Are there instances where we might want to restrict CSS as well as JS? * Would a $wg config option and/or user preference and/or index.php GET parameter to limit inclusion be useful? * Can we deprecate any of the existing $wg config options? * What's going on with the duplicated code between OutputPage and SkinTemplate?
2011-02-04 16:39:17 +00:00
: $modules;
Merging resourceloader branch into trunk. Full documentation is at http://www.mediawiki.org/wiki/ResourceLoader and a general overview has been posted on wikitech-li <http://lists.wikimedia.org/pipermail/wikitech-l/2010-September/049253.html>. One important change is that all JS is now loaded at the bottom, so any scripts assuming things from wikibits or whatever are present will fail.
2010-09-04 04:00:09 +00:00
}
/**
resourceloader: Deprecate OutputPage::addModuleScripts() And the counter-part in ParserOutput as well. This follows-up on 80e5b160e0985, which removed the last use of this method. It also deprecated ResourceLoaderClientHtml::setModuleScripts() but that isn't usually used directly. The public interfaces are mainly in OutputPage/ParserOutput, which are now deprecated as well. Bug: T188689 Change-Id: I90baafdfc089ac03c4c3abc12065c0b60d43994e
2018-03-02 01:21:32 +00:00
* Load one or more ResourceLoader modules on this page.
Fixed some doxygen warnings
2010-09-07 21:05:09 +00:00
*
Removed unneeded spaces and colons in @param and friends Also swapped some "$var type" to "type $var" or added missing types before the $var. Changed some other types to match the more common spelling. Makes beginning of some text in captial. Change-Id: Ic36c8c7820a6c2d603f1138130670c6bf6a1ca59
2014-04-08 15:29:17 +00:00
* @param string|array $modules Module name (string) or array of module names
Merging resourceloader branch into trunk. Full documentation is at http://www.mediawiki.org/wiki/ResourceLoader and a general overview has been posted on wikitech-li <http://lists.wikimedia.org/pipermail/wikitech-l/2010-September/049253.html>. One important change is that all JS is now loaded at the bottom, so any scripts assuming things from wikibits or whatever are present will fail.
2010-09-04 04:00:09 +00:00
*/
public function addModules( $modules ) {
$this->mModules = array_merge( $this->mModules, (array)$modules );
}
/**
resourceloader: Deprecate OutputPage::addModuleScripts() And the counter-part in ParserOutput as well. This follows-up on 80e5b160e0985, which removed the last use of this method. It also deprecated ResourceLoaderClientHtml::setModuleScripts() but that isn't usually used directly. The public interfaces are mainly in OutputPage/ParserOutput, which are now deprecated as well. Bug: T188689 Change-Id: I90baafdfc089ac03c4c3abc12065c0b60d43994e
2018-03-02 01:21:32 +00:00
* Get the list of style-only modules to load on this page.
Fixed some doxygen warnings
2010-09-07 21:05:09 +00:00
*
Fixed some @params documentation Swapped some "$var type" to "type $var" or added missing types before the $var. Changed some other types to match the more common spelling. Makes beginning of some text in captial. Change-Id: I64e8cfe478cb0ba438f40b0631d6e9049cdab567
2014-04-10 18:50:10 +00:00
* @param bool $filter
resourceloader: Remove ResourceLoaderModule::getPosition @deprecated since 1.29 Bug: T184257 Change-Id: I1a59a1f0a5fbedfea9f0b28bee8c298710557a8d
2018-02-12 21:06:05 +00:00
* @param string|null $position Unused
Update a lot of unspecific "array" types in PHPDocs This includes fixing some mistakes, as well as removing redundant text that doesn't add new information, either because it literally repeats what the code already says, or is actually duplicated. Change-Id: I3a8dd8ce57192deda8916cc444c87d7ab1a36515
2020-10-28 10:01:33 +00:00
* @return string[] Array of module names
Merging resourceloader branch into trunk. Full documentation is at http://www.mediawiki.org/wiki/ResourceLoader and a general overview has been posted on wikitech-li <http://lists.wikimedia.org/pipermail/wikitech-l/2010-September/049253.html>. One important change is that all JS is now loaded at the bottom, so any scripts assuming things from wikibits or whatever are present will fail.
2010-09-04 04:00:09 +00:00
*/
For bug 27488: move the startup script, jquery+mediawiki and the mw.config.set() call for configuration variables back to the <head> . Let modules control whether they're loaded in the <head> ('top') or at the bottom of the <body> ('bottom') through the position parameter/property Also rearranges the loading order a little bit such that only=messages comes before only=scripts, and config comes before everything except startup and jquery+mediawiki
2011-04-07 12:07:25 +00:00
public function getModuleStyles( $filter = false, $position = null ) {
resourceloader: Remove ResourceLoaderModule::getPosition @deprecated since 1.29 Bug: T184257 Change-Id: I1a59a1f0a5fbedfea9f0b28bee8c298710557a8d
2018-02-12 21:06:05 +00:00
return $this->getModules( $filter, null, 'mModuleStyles',
ResourceLoader namespace (attempt 2) Move ResourceLoader classes to their own namespace. Strip the "ResourceLoader" prefix from all except ResourceLoader itself. Move the tests by analogy. I used a namespace alias "RL" in some callers since RL\Module is less ambiguous at the call site than just "Module". I did not address DependencyStore which continues to have a non-standard location and namespace. Revert of a241d83e0a6dabedf. Bug: T308718 Change-Id: Id08a220e1d6085e2b33f3f6c9d0e3935a4204659
2022-05-06 09:09:56 +00:00
RL\Module::TYPE_STYLES
resourceloader: Move queue formatting out of OutputPage HTML formatting of the queue was distributed over several OutputPage methods. Each method demanding a snippet of HTML by calling makeResourceLoaderLink() with a limited amount of information. As such, makeResourceLoaderLink() was unable to provide the client with the proper state information. Centralising it also allows it to better reduce duplication in HTML output and maintain a more accurate state. Problems fixed by centralising: 1. The 'user' module is special (due to per-user 'version' and 'user' params). It is manually requested via script-src. To avoid a separate (and wrong) request from something that requires it, we set state=loading directly. However, because the module is in the bottom, the old HTML formatter could only put state=loading in the bottom also. This sometimes caused a wrong request to be fired for modules=user if something in the top queue triggered a requirement for it. 2. Since a464d1d4 (T87871) we track states of page-style modules, with purpose of allowing dependencies on style modules without risking duplicate loading on pages where the styles are loaded already. This didn't work, because the state information about page-style modules is output near the stylesheet, which is after the script tag with mw.loader.load(). That runs first, and mw.loader would still make a duplicate request before it learns the state. Changes: * Document reasons for style/script tag order in getHeadHtml (per 09537e83). * Pass $type from getModuleStyles() to getAllowedModules(). This wasn't needed before since a duplicate check in makeResourceLoaderLink() verified the origin a second time. * Declare explicit position 'top' on 'user.options' and 'user.tokens' module. Previously, OutputPage hardcoded them in the top. The new formatter doesn't. * Remove getHeadScripts(). * Remove getInlineHeadScripts(). * Remove getExternalHeadScripts(). * Remove buildCssLinks(). * Remove getScriptsForBottomQueue(). * Change where Skin::setupSkinUserCss() is called. This methods lets the skin add modules to the queue. Previously it was called from buildCssLinks(), via headElement(), via prepareQuickTemplate(), via OutputPage::output(). It's now in OutputPage::output() directly (slightly earlier). This is needed because prepareQuickTemplate() calls bottomScripts() before headElement(). And bottomScript() would lazy-initialise the queue and lock it before setupSkinUserCss() is called from headElement(). This makes execution order more predictable instead of being dependent on the arbitrary order of data extraction in prepareQuickTemplate (which varies from one skin to another). * Compute isUserModulePreview() and isKnownEmpty() for the 'user' module early on so. This avoids wrongful loading and fixes problem 1. Effective changes in output: * mw.loader.state() is now before mw.loader.load(). This fixes problem 2. * mw.loader.state() now sets 'user.options' and 'user.tokens' to "loading". * mw.loader.state() now sets 'user' (as "loading" or "ready"). Fixes problem 1. * The <script async src> tag for 'startup' changed position (slightly). Previously it was after all inline scripts and stylesheets. It's still after all inline scripts and after most stylesheets, but before any user styles. Since the queue is now formatted outside OutputPage, it can't inject the meta-ResourceLoaderDynamicStyles tag and user-stylesheet hack in the middle of existing output. This shouldn't have any noticable impact. Bug: T87871 Change-Id: I605b8cd1e1fc009b4662a0edbc54d09dd65ee1df
2016-07-15 14:13:09 +00:00
);
Merging resourceloader branch into trunk. Full documentation is at http://www.mediawiki.org/wiki/ResourceLoader and a general overview has been posted on wikitech-li <http://lists.wikimedia.org/pipermail/wikitech-l/2010-September/049253.html>. One important change is that all JS is now loaded at the bottom, so any scripts assuming things from wikibits or whatever are present will fail.
2010-09-04 04:00:09 +00:00
}
/**
Clarify that addModuleStyles is for style-only modules “the styles of … ResourceLoader modules” suggests to me that this method can be used with any ResourceLoader module(s), and their style parts will be loaded as <link rel=stylesheet> while JavaScript parts will load as usual, but this is not in fact the case: if ResourceLoaderClientHtml encounters a module type other than LOAD_STYLES in the moduleStyles modules, it logs an error and skips the module. Change-Id: If0e709311c1b777f6d99cb714f19307c1015ec01
2020-06-19 15:27:18 +00:00
* Load the styles of one or more style-only ResourceLoader modules on this page.
Improve documentation for OutputPage->addModuleStyles Change-Id: I56392cdb4e0d9fd5cac15ac947962364ae96381f
2013-02-27 06:09:26 +00:00
*
resourceloader: Deprecate OutputPage::addModuleScripts() And the counter-part in ParserOutput as well. This follows-up on 80e5b160e0985, which removed the last use of this method. It also deprecated ResourceLoaderClientHtml::setModuleScripts() but that isn't usually used directly. The public interfaces are mainly in OutputPage/ParserOutput, which are now deprecated as well. Bug: T188689 Change-Id: I90baafdfc089ac03c4c3abc12065c0b60d43994e
2018-03-02 01:21:32 +00:00
* Module styles added through this function will be loaded as a stylesheet,
* using a standard `<link rel=stylesheet>` HTML tag, rather than as a combined
* Javascript and CSS package. Thus, they will even load when JavaScript is disabled.
Fixed some doxygen warnings
2010-09-07 21:05:09 +00:00
*
Removed unneeded spaces and colons in @param and friends Also swapped some "$var type" to "type $var" or added missing types before the $var. Changed some other types to match the more common spelling. Makes beginning of some text in captial. Change-Id: Ic36c8c7820a6c2d603f1138130670c6bf6a1ca59
2014-04-08 15:29:17 +00:00
* @param string|array $modules Module name (string) or array of module names
Merging resourceloader branch into trunk. Full documentation is at http://www.mediawiki.org/wiki/ResourceLoader and a general overview has been posted on wikitech-li <http://lists.wikimedia.org/pipermail/wikitech-l/2010-September/049253.html>. One important change is that all JS is now loaded at the bottom, so any scripts assuming things from wikibits or whatever are present will fail.
2010-09-04 04:00:09 +00:00
*/
public function addModuleStyles( $modules ) {
$this->mModuleStyles = array_merge( $this->mModuleStyles, (array)$modules );
}
Add RL targets support to OutputPage The target URI parameter is already used by ResourceLoaderStartUpModule to filter modules, this change introduces the ability to produce mobile-aware load.php links in the core and filter out modules not suitable for current target. Change-Id: Ie03990e11d9a0ac8cd81dd0aae5c5a95f1d23e81
2013-02-14 17:05:22 +00:00
/**
Removed unneeded spaces and colons in @param and friends Also swapped some "$var type" to "type $var" or added missing types before the $var. Changed some other types to match the more common spelling. Makes beginning of some text in captial. Change-Id: Ic36c8c7820a6c2d603f1138130670c6bf6a1ca59
2014-04-08 15:29:17 +00:00
* @return null|string ResourceLoader target
Add RL targets support to OutputPage The target URI parameter is already used by ResourceLoaderStartUpModule to filter modules, this change introduces the ability to produce mobile-aware load.php links in the core and filter out modules not suitable for current target. Change-Id: Ie03990e11d9a0ac8cd81dd0aae5c5a95f1d23e81
2013-02-14 17:05:22 +00:00
*/
public function getTarget() {
return $this->mTarget;
}
/**
* Sets ResourceLoader target for load.php links. If null, will be omitted
*
Fixed some @params documentation Swapped some "$var type" to "type $var" or added missing types before the $var. Changed some other types to match the more common spelling. Makes beginning of some text in captial. Change-Id: I64e8cfe478cb0ba438f40b0631d6e9049cdab567
2014-04-10 18:50:10 +00:00
* @param string|null $target
Add RL targets support to OutputPage The target URI parameter is already used by ResourceLoaderStartUpModule to filter modules, this change introduces the ability to produce mobile-aware load.php links in the core and filter out modules not suitable for current target. Change-Id: Ie03990e11d9a0ac8cd81dd0aae5c5a95f1d23e81
2013-02-14 17:05:22 +00:00
*/
public function setTarget( $target ) {
$this->mTarget = $target;
}
Generalize ResourceLoader 'excludepage' functionality There has long been a hack for previewing edits to user JS/CSS, where OutputPage would pass an 'excludepage' parameter to ResourceLoaderUserModule to tell it not to load one particular page and would instead embed that page statically. That's nice, but there are other places where we could use the same thing. This patch generalizes it: * DerivativeResourceLoaderContext may now contain a callback for mapping titles to replacement Content objects. * ResourceLoaderWikiModule::getContent() uses the overrides, and requests embedding when they're used. All subclasses in Gerrit should pick it up automatically. * OutputPage gains methods for callers to add to the override mapping, which it passes on to RL. It loses a bunch of the special casing it had for the 'user' and 'user.styles' modules. * EditPage sets the overrides on OutputPage when doing the preview, as does ApiParse for prop=headhtml. TemplateSandbox does too in I83fa0856. * OutputPage::userCanPreview() gets less specific to editing user CSS and JS, since RL now handles the embedding based on the actual modules' dependencies and EditPage only requests it on preview. ApiParse also gets a new hook to support TemplateSandbox's API integration (used in I83fa0856). Bug: T112474 Change-Id: Ib9d2ce42931c1de8372e231314a1f672d7e2ac0e
2017-02-28 20:52:17 +00:00
/**
resourceloader: Remove Title from public interfaces Bug: T278459 Depends-On: I9bc371fd299e7a5401ebc1b8d7456d9a4ecde90d Change-Id: Ib3a0591766f873492ec5a24721e7840f4e1a6407
2021-05-05 20:58:49 +00:00
* Force the given Content object for the given page, for things like page preview.
Generalize ResourceLoader 'excludepage' functionality There has long been a hack for previewing edits to user JS/CSS, where OutputPage would pass an 'excludepage' parameter to ResourceLoaderUserModule to tell it not to load one particular page and would instead embed that page statically. That's nice, but there are other places where we could use the same thing. This patch generalizes it: * DerivativeResourceLoaderContext may now contain a callback for mapping titles to replacement Content objects. * ResourceLoaderWikiModule::getContent() uses the overrides, and requests embedding when they're used. All subclasses in Gerrit should pick it up automatically. * OutputPage gains methods for callers to add to the override mapping, which it passes on to RL. It loses a bunch of the special casing it had for the 'user' and 'user.styles' modules. * EditPage sets the overrides on OutputPage when doing the preview, as does ApiParse for prop=headhtml. TemplateSandbox does too in I83fa0856. * OutputPage::userCanPreview() gets less specific to editing user CSS and JS, since RL now handles the embedding based on the actual modules' dependencies and EditPage only requests it on preview. ApiParse also gets a new hook to support TemplateSandbox's API integration (used in I83fa0856). Bug: T112474 Change-Id: Ib9d2ce42931c1de8372e231314a1f672d7e2ac0e
2017-02-28 20:52:17 +00:00
* @see self::addContentOverrideCallback()
* @since 1.32
resourceloader: Remove Title from public interfaces Bug: T278459 Depends-On: I9bc371fd299e7a5401ebc1b8d7456d9a4ecde90d Change-Id: Ib3a0591766f873492ec5a24721e7840f4e1a6407
2021-05-05 20:58:49 +00:00
* @param LinkTarget|PageReference $target
Generalize ResourceLoader 'excludepage' functionality There has long been a hack for previewing edits to user JS/CSS, where OutputPage would pass an 'excludepage' parameter to ResourceLoaderUserModule to tell it not to load one particular page and would instead embed that page statically. That's nice, but there are other places where we could use the same thing. This patch generalizes it: * DerivativeResourceLoaderContext may now contain a callback for mapping titles to replacement Content objects. * ResourceLoaderWikiModule::getContent() uses the overrides, and requests embedding when they're used. All subclasses in Gerrit should pick it up automatically. * OutputPage gains methods for callers to add to the override mapping, which it passes on to RL. It loses a bunch of the special casing it had for the 'user' and 'user.styles' modules. * EditPage sets the overrides on OutputPage when doing the preview, as does ApiParse for prop=headhtml. TemplateSandbox does too in I83fa0856. * OutputPage::userCanPreview() gets less specific to editing user CSS and JS, since RL now handles the embedding based on the actual modules' dependencies and EditPage only requests it on preview. ApiParse also gets a new hook to support TemplateSandbox's API integration (used in I83fa0856). Bug: T112474 Change-Id: Ib9d2ce42931c1de8372e231314a1f672d7e2ac0e
2017-02-28 20:52:17 +00:00
* @param Content $content
*/
resourceloader: Remove Title from public interfaces Bug: T278459 Depends-On: I9bc371fd299e7a5401ebc1b8d7456d9a4ecde90d Change-Id: Ib3a0591766f873492ec5a24721e7840f4e1a6407
2021-05-05 20:58:49 +00:00
public function addContentOverride( $target, Content $content ) {
Generalize ResourceLoader 'excludepage' functionality There has long been a hack for previewing edits to user JS/CSS, where OutputPage would pass an 'excludepage' parameter to ResourceLoaderUserModule to tell it not to load one particular page and would instead embed that page statically. That's nice, but there are other places where we could use the same thing. This patch generalizes it: * DerivativeResourceLoaderContext may now contain a callback for mapping titles to replacement Content objects. * ResourceLoaderWikiModule::getContent() uses the overrides, and requests embedding when they're used. All subclasses in Gerrit should pick it up automatically. * OutputPage gains methods for callers to add to the override mapping, which it passes on to RL. It loses a bunch of the special casing it had for the 'user' and 'user.styles' modules. * EditPage sets the overrides on OutputPage when doing the preview, as does ApiParse for prop=headhtml. TemplateSandbox does too in I83fa0856. * OutputPage::userCanPreview() gets less specific to editing user CSS and JS, since RL now handles the embedding based on the actual modules' dependencies and EditPage only requests it on preview. ApiParse also gets a new hook to support TemplateSandbox's API integration (used in I83fa0856). Bug: T112474 Change-Id: Ib9d2ce42931c1de8372e231314a1f672d7e2ac0e
2017-02-28 20:52:17 +00:00
if ( !$this->contentOverrides ) {
// Register a callback for $this->contentOverrides on the first call
resourceloader: Remove Title from public interfaces Bug: T278459 Depends-On: I9bc371fd299e7a5401ebc1b8d7456d9a4ecde90d Change-Id: Ib3a0591766f873492ec5a24721e7840f4e1a6407
2021-05-05 20:58:49 +00:00
$this->addContentOverrideCallback( function ( $target ) {
Generalize ResourceLoader 'excludepage' functionality There has long been a hack for previewing edits to user JS/CSS, where OutputPage would pass an 'excludepage' parameter to ResourceLoaderUserModule to tell it not to load one particular page and would instead embed that page statically. That's nice, but there are other places where we could use the same thing. This patch generalizes it: * DerivativeResourceLoaderContext may now contain a callback for mapping titles to replacement Content objects. * ResourceLoaderWikiModule::getContent() uses the overrides, and requests embedding when they're used. All subclasses in Gerrit should pick it up automatically. * OutputPage gains methods for callers to add to the override mapping, which it passes on to RL. It loses a bunch of the special casing it had for the 'user' and 'user.styles' modules. * EditPage sets the overrides on OutputPage when doing the preview, as does ApiParse for prop=headhtml. TemplateSandbox does too in I83fa0856. * OutputPage::userCanPreview() gets less specific to editing user CSS and JS, since RL now handles the embedding based on the actual modules' dependencies and EditPage only requests it on preview. ApiParse also gets a new hook to support TemplateSandbox's API integration (used in I83fa0856). Bug: T112474 Change-Id: Ib9d2ce42931c1de8372e231314a1f672d7e2ac0e
2017-02-28 20:52:17 +00:00
$key = $target->getNamespace() . ':' . $target->getDBkey();
Use PHP 7 '??' operator instead of '?:' with 'isset()' where convenient Find: /isset\(\s*([^()]+?)\s*\)\s*\?\s*\1\s*:\s*/ Replace with: '\1 ?? ' (Everywhere except includes/PHPVersionCheck.php) (Then, manually fix some line length and indentation issues) Then manually reviewed the replacements for cases where confusing operator precedence would result in incorrect results (fixing those in I478db046a1cc162c6767003ce45c9b56270f3372). Change-Id: I33b421c8cb11cdd4ce896488c9ff5313f03a38cf
2017-10-06 22:17:58 +00:00
return $this->contentOverrides[$key] ?? null;
Generalize ResourceLoader 'excludepage' functionality There has long been a hack for previewing edits to user JS/CSS, where OutputPage would pass an 'excludepage' parameter to ResourceLoaderUserModule to tell it not to load one particular page and would instead embed that page statically. That's nice, but there are other places where we could use the same thing. This patch generalizes it: * DerivativeResourceLoaderContext may now contain a callback for mapping titles to replacement Content objects. * ResourceLoaderWikiModule::getContent() uses the overrides, and requests embedding when they're used. All subclasses in Gerrit should pick it up automatically. * OutputPage gains methods for callers to add to the override mapping, which it passes on to RL. It loses a bunch of the special casing it had for the 'user' and 'user.styles' modules. * EditPage sets the overrides on OutputPage when doing the preview, as does ApiParse for prop=headhtml. TemplateSandbox does too in I83fa0856. * OutputPage::userCanPreview() gets less specific to editing user CSS and JS, since RL now handles the embedding based on the actual modules' dependencies and EditPage only requests it on preview. ApiParse also gets a new hook to support TemplateSandbox's API integration (used in I83fa0856). Bug: T112474 Change-Id: Ib9d2ce42931c1de8372e231314a1f672d7e2ac0e
2017-02-28 20:52:17 +00:00
} );
}
$key = $target->getNamespace() . ':' . $target->getDBkey();
$this->contentOverrides[$key] = $content;
}
/**
* Add a callback for mapping from a Title to a Content object, for things
* like page preview.
ResourceLoader namespace (attempt 2) Move ResourceLoader classes to their own namespace. Strip the "ResourceLoader" prefix from all except ResourceLoader itself. Move the tests by analogy. I used a namespace alias "RL" in some callers since RL\Module is less ambiguous at the call site than just "Module". I did not address DependencyStore which continues to have a non-standard location and namespace. Revert of a241d83e0a6dabedf. Bug: T308718 Change-Id: Id08a220e1d6085e2b33f3f6c9d0e3935a4204659
2022-05-06 09:09:56 +00:00
* @see RL\Context::getContentOverrideCallback()
Generalize ResourceLoader 'excludepage' functionality There has long been a hack for previewing edits to user JS/CSS, where OutputPage would pass an 'excludepage' parameter to ResourceLoaderUserModule to tell it not to load one particular page and would instead embed that page statically. That's nice, but there are other places where we could use the same thing. This patch generalizes it: * DerivativeResourceLoaderContext may now contain a callback for mapping titles to replacement Content objects. * ResourceLoaderWikiModule::getContent() uses the overrides, and requests embedding when they're used. All subclasses in Gerrit should pick it up automatically. * OutputPage gains methods for callers to add to the override mapping, which it passes on to RL. It loses a bunch of the special casing it had for the 'user' and 'user.styles' modules. * EditPage sets the overrides on OutputPage when doing the preview, as does ApiParse for prop=headhtml. TemplateSandbox does too in I83fa0856. * OutputPage::userCanPreview() gets less specific to editing user CSS and JS, since RL now handles the embedding based on the actual modules' dependencies and EditPage only requests it on preview. ApiParse also gets a new hook to support TemplateSandbox's API integration (used in I83fa0856). Bug: T112474 Change-Id: Ib9d2ce42931c1de8372e231314a1f672d7e2ac0e
2017-02-28 20:52:17 +00:00
* @since 1.32
* @param callable $callback
*/
public function addContentOverrideCallback( callable $callback ) {
$this->contentOverrideCallbacks[] = $callback;
}
OutputPage: allow additional CSS classes to be added to <html> In some cases CSS classes must be on the <html> element and not the <body>. An example is the DarkMode extension. Due to how the stacking context works, the CSS `invert()` filter must be on <html>. We can't use JavaScript to apply the styles because there would be a brief delay after page load before the JS runs, causing a flashing effect. Bug: T241925 Change-Id: I4b994e081d9582801fe9d198efd68afc46a14703
2019-05-21 20:26:53 +00:00
/**
* Add a class to the <html> element. This should rarely be used.
* Instead use OutputPage::addBodyClasses() if possible.
*
OutputPage: Use @unstable instead of @internal for addHtmlClasses() Follows-up 9d3254d18e7e. Methods using @internal are designed for use internal to the component that defines the method, and are generally considered safe to remove anytime if they are unused. This method however is not for internal use, it is designed for public use (nothing in core needs it), but we're not yet ready to commit to it being a stable API that would receive no breaking changes or be removed without deprecation. The appropiate tag for such experimental feature that allows early adoption (with caution, knowing it might get renamed, moved or dropped in the next major release) – is @unstable. <https://www.mediawiki.org/wiki/Stable_interface_policy> Bug: T247862 Change-Id: Iceaaf520979afa1c3ee806262083f94d67660174
2020-07-02 01:01:39 +00:00
* @unstable Experimental since 1.35. Prefer OutputPage::addBodyClasses()
OutputPage: allow additional CSS classes to be added to <html> In some cases CSS classes must be on the <html> element and not the <body>. An example is the DarkMode extension. Due to how the stacking context works, the CSS `invert()` filter must be on <html>. We can't use JavaScript to apply the styles because there would be a brief delay after page load before the JS runs, causing a flashing effect. Bug: T241925 Change-Id: I4b994e081d9582801fe9d198efd68afc46a14703
2019-05-21 20:26:53 +00:00
* @param string|string[] $classes One or more classes to add
*/
public function addHtmlClasses( $classes ) {
$this->mAdditionalHtmlClasses = array_merge( $this->mAdditionalHtmlClasses, (array)$classes );
}
* (bug 15558) Allow includable special pages to be parameterized using wiki syntax * Changed SpecialPageFactory::capturePath() to take the same parameters as SpecialPageFactory::executePath() (except the last one) and made it always return a bool instead of bool-or-string. The result HTML can be fetched from the OutputPage object of the context. * Added module styles, scritps and messages members to ParserOutput in addition to modules. The first one is needed to display Special:RecentChanges correctly when transcluded since EnhancedChangesList::beginRecentChangesList() calls addModuleStyles( 'mediawiki.special.changeslist' ) Yes, this means that you can use {{Special:Recentchanges|enhanced=0}} to use the old changes list. For the ones that wonder, {{Special:Recentchanges|uselang=something}} will not work since the Language object is forced to the one used by the parser.
2011-08-02 15:40:03 +00:00
/**
* Get an array of head items
*
Replace various `array` type hints with more specific `string[]` There are many, many more. I touch only a few where I'm sure it's never anything but an array of strings. Change-Id: I8b798f2e9d48f07a241b95ce0ace8fa9d981695d
2022-09-27 07:22:15 +00:00
* @return string[]
* (bug 15558) Allow includable special pages to be parameterized using wiki syntax * Changed SpecialPageFactory::capturePath() to take the same parameters as SpecialPageFactory::executePath() (except the last one) and made it always return a bool instead of bool-or-string. The result HTML can be fetched from the OutputPage object of the context. * Added module styles, scritps and messages members to ParserOutput in addition to modules. The first one is needed to display Special:RecentChanges correctly when transcluded since EnhancedChangesList::beginRecentChangesList() calls addModuleStyles( 'mediawiki.special.changeslist' ) Yes, this means that you can use {{Special:Recentchanges|enhanced=0}} to use the old changes list. For the ones that wonder, {{Special:Recentchanges|uselang=something}} will not work since the Language object is forced to the one used by the parser.
2011-08-02 15:40:03 +00:00
*/
OutputPage: Add visibility to various methods & do some micro-optimizations Methods that visibility was added to are; `addMeta()`, `addLink()`, `setCanonicalUrl()`, `addScript()`, `getHeadItemsArray()`, `addParserOutput()`, `getCacheVaryCookies()` and `haveCacheVaryCookies()`. Last but not lease, did a few micro-optimizations to `addMeta()` and `addLink()`. Change-Id: I94d037a5edc7131627724fd1d864000128077b0c
2019-09-14 10:26:31 +00:00
public function getHeadItemsArray() {
* (bug 15558) Allow includable special pages to be parameterized using wiki syntax * Changed SpecialPageFactory::capturePath() to take the same parameters as SpecialPageFactory::executePath() (except the last one) and made it always return a bool instead of bool-or-string. The result HTML can be fetched from the OutputPage object of the context. * Added module styles, scritps and messages members to ParserOutput in addition to modules. The first one is needed to display Special:RecentChanges correctly when transcluded since EnhancedChangesList::beginRecentChangesList() calls addModuleStyles( 'mediawiki.special.changeslist' ) Yes, this means that you can use {{Special:Recentchanges|enhanced=0}} to use the old changes list. For the ones that wonder, {{Special:Recentchanges|uselang=something}} will not work since the Language object is forced to the one used by the parser.
2011-08-02 15:40:03 +00:00
return $this->mHeadItems;
}
Step one in fixing OutputPage's documentation: * Groupped methods by function * Added documentation * Fixed some doxygen warnings * Some little style fixes (more to come)
2010-01-28 21:58:50 +00:00
/**
OutputPage: Add addHeadItems() method Right now OutputPage requires head items to have a name when using the addHeadItem() method. This in contrary to most of the helper methods, which don't require a name. ParserOutput::addHeadItem() also doesn't require a name. Since the existing signature isn't easily changed without adding confusion, add a separate method that allow numerical indexes to be pushed. This ability was already exposed via addParserOutputMetadata() and its array_merge. This method exposes it directly. Change-Id: I433afd124b57526caa13a540cda48ba2b99a9bde
2016-07-25 23:54:13 +00:00
* Add or replace a head item to the output
Step one in fixing OutputPage's documentation: * Groupped methods by function * Added documentation * Fixed some doxygen warnings * Some little style fixes (more to come)
2010-01-28 21:58:50 +00:00
*
OutputPage: documentation for (semi-)deprecated methods Deprecated: addExtensionStyle, getExtStyle These had been deprecated unofficially for a while now. ResourceLoader is preferred. Internal/Unpreferred: addHeadItem, addScript, addScriptFile, addInlineScript, addStyle, addInlineStyle. All should be used as littel as possible, in favor of ResourceLoader and a few other higher level accessors. They are not truly deprecated, but should not really be used unless you have a really good reason. Also added a few missing wfDeprecated calls to methods marked as deprecated in the phpdoc. Bug: T89955 Change-Id: I472d906b6de16257fa34a2f9bfe9e4eb284d02b4
2015-10-31 15:16:58 +00:00
* Whenever possible, use more specific options like ResourceLoader modules,
Fix method name in comment There is no addMetaLink() method; I assume it’s meant to be addMeta(). Change-Id: I4f59d5dd96d3a59522fbc7cf38badc93698e29fa
2020-06-08 10:30:33 +00:00
* OutputPage::addLink(), OutputPage::addMeta() and OutputPage::addFeedLink()
OutputPage: documentation for (semi-)deprecated methods Deprecated: addExtensionStyle, getExtStyle These had been deprecated unofficially for a while now. ResourceLoader is preferred. Internal/Unpreferred: addHeadItem, addScript, addScriptFile, addInlineScript, addStyle, addInlineStyle. All should be used as littel as possible, in favor of ResourceLoader and a few other higher level accessors. They are not truly deprecated, but should not really be used unless you have a really good reason. Also added a few missing wfDeprecated calls to methods marked as deprecated in the phpdoc. Bug: T89955 Change-Id: I472d906b6de16257fa34a2f9bfe9e4eb284d02b4
2015-10-31 15:16:58 +00:00
* Fallback options for those are: OutputPage::addStyle, OutputPage::addScript(),
* OutputPage::addInlineScript() and OutputPage::addInlineStyle()
* This would be your very LAST fallback.
*
Cleanup some docs (includes/*.php) - Swap "$variable type" to "type $variable" - Added missing types - Fixed spacing inside docs - Makes beginning of @param/@return/@var/@throws in capital - Changed some types to match the more common spelling Change-Id: I783e4dbfe5f6f98b32b9a03ccf6439e13e132bcc
2014-07-24 17:42:24 +00:00
* @param string $name Item name
* @param string $value Raw HTML
Step one in fixing OutputPage's documentation: * Groupped methods by function * Added documentation * Fixed some doxygen warnings * Some little style fixes (more to come)
2010-01-28 21:58:50 +00:00
*/
public function addHeadItem( $name, $value ) {
addHeadItem() method
2007-04-29 10:15:05 +00:00
$this->mHeadItems[$name] = $value;
}
This commit implements Josh Cogliati's "info" feature (e.g. http://en.wikipedia.org/w/wiki.phtml?title=Wikipedia&action=info), fixes the page headings for protect/unprotect, and does a lot of whitespace changes (sorry about that). I've also fixed confusing indentation somewhere.
2004-07-08 14:53:54 +00:00
OutputPage: Add addHeadItems() method Right now OutputPage requires head items to have a name when using the addHeadItem() method. This in contrary to most of the helper methods, which don't require a name. ParserOutput::addHeadItem() also doesn't require a name. Since the existing signature isn't easily changed without adding confusion, add a separate method that allow numerical indexes to be pushed. This ability was already exposed via addParserOutputMetadata() and its array_merge. This method exposes it directly. Change-Id: I433afd124b57526caa13a540cda48ba2b99a9bde
2016-07-25 23:54:13 +00:00
/**
* Add one or more head items to the output
*
* @since 1.28
Use correct variable name in @param documentation For some varargs a variable name is added with suffix ,... as seen for many other varargs Some @param are swapped, because there are in the wrong order Enable Sniff MediaWiki.Commenting.FunctionComment.ParamNameNoMatch Change-Id: I60fec6025bce824d5c67563ab7b65ad6cd628ad8
2017-08-11 16:09:41 +00:00
* @param string|string[] $values Raw HTML
OutputPage: Add addHeadItems() method Right now OutputPage requires head items to have a name when using the addHeadItem() method. This in contrary to most of the helper methods, which don't require a name. ParserOutput::addHeadItem() also doesn't require a name. Since the existing signature isn't easily changed without adding confusion, add a separate method that allow numerical indexes to be pushed. This ability was already exposed via addParserOutputMetadata() and its array_merge. This method exposes it directly. Change-Id: I433afd124b57526caa13a540cda48ba2b99a9bde
2016-07-25 23:54:13 +00:00
*/
public function addHeadItems( $values ) {
$this->mHeadItems = array_merge( $this->mHeadItems, (array)$values );
}
Step one in fixing OutputPage's documentation: * Groupped methods by function * Added documentation * Fixed some doxygen warnings * Some little style fixes (more to come)
2010-01-28 21:58:50 +00:00
/**
* Check if the header item $name is already set
*
Fixed some @params documentation Swapped some "$var type" to "type $var" or added missing types before the $var. Changed some other types to match the more common spelling. Makes beginning of some text in captial. Change-Id: I64e8cfe478cb0ba438f40b0631d6e9049cdab567
2014-04-10 18:50:10 +00:00
* @param string $name Item name
* @return bool
Step one in fixing OutputPage's documentation: * Groupped methods by function * Added documentation * Fixed some doxygen warnings * Some little style fixes (more to come)
2010-01-28 21:58:50 +00:00
*/
public function hasHeadItem( $name ) {
Basic integrated audio/video support, with Ogg implementation. * JavaScript video player based loosely on Greg Maxwell's player * Image page text snippet customisation * Abstraction of transform parameters in the parser. Introduced Linker::makeImageLink2(). * Made canRender(), mustRender() depend on file, not just on handler. Moved width=0, height=0 checking to ImageHandler::canRender(), since audio streams have width=height=0 but should be rendered. Also: * Automatic upgrade for oldimage rows on image page view, allows media handler selection based on oi_*_mime * oi_*_mime unconditionally referenced, REQUIRES SCHEMA UPGRADE * Don't destroy file info for missing files on upgrade * Simple, centralised extension message file handling * Made MessageCache::loadAllMessages non-static, optimised for repeated-call case due to abuse in User.php * Support for lightweight parser output hooks, with callback whitelist for security * Moved Linker::formatSize() to Language, to join the new formatTimePeriod() and formatBitrate() * Introduced MagicWordArray, regex capture trick requires that magic word IDs DO NOT CONTAIN HYPHENS.
2007-08-15 10:50:09 +00:00
return isset( $this->mHeadItems[$name] );
}
Add <body> class marking whether RCFilters is enabled Also added OutputPage->addBodyClasses to allow this Bug: T175062 Change-Id: I2111d0762328f2e59038c92d63583a9bb267fc9f
2017-09-11 07:12:03 +00:00
/**
* Add a class to the <body> element
*
* @since 1.30
* @param string|string[] $classes One or more classes to add
*/
public function addBodyClasses( $classes ) {
$this->mAdditionalBodyClasses = array_merge( $this->mAdditionalBodyClasses, (array)$classes );
}
Step one in fixing OutputPage's documentation: * Groupped methods by function * Added documentation * Fixed some doxygen warnings * Some little style fixes (more to come)
2010-01-28 21:58:50 +00:00
/**
* Set whether the output should only contain the body of the article,
* without any skin, sidebar, etc.
Fixes to OutputPage's documentation * action=raw -> action=render * escape tags in setPageTitle()'s documentation so that it's displayed correctly (not treated as real tags)
2010-02-01 18:04:17 +00:00
* Used e.g. when calling with "action=render".
Step one in fixing OutputPage's documentation: * Groupped methods by function * Added documentation * Fixed some doxygen warnings * Some little style fixes (more to come)
2010-01-28 21:58:50 +00:00
*
Removed unneeded spaces and colons in @param and friends Also swapped some "$var type" to "type $var" or added missing types before the $var. Changed some other types to match the more common spelling. Makes beginning of some text in captial. Change-Id: Ic36c8c7820a6c2d603f1138130670c6bf6a1ca59
2014-04-08 15:29:17 +00:00
* @param bool $only Whether to output only the body of the article
Step one in fixing OutputPage's documentation: * Groupped methods by function * Added documentation * Fixed some doxygen warnings * Some little style fixes (more to come)
2010-01-28 21:58:50 +00:00
*/
public function setArticleBodyOnly( $only ) {
$this->mArticleBodyOnly = $only;
Another go at extension CSS order
2008-08-10 07:14:08 +00:00
}
Initial revision
2003-04-14 23:10:40 +00:00
Step one in fixing OutputPage's documentation: * Groupped methods by function * Added documentation * Fixed some doxygen warnings * Some little style fixes (more to come)
2010-01-28 21:58:50 +00:00
/**
* Return whether the output will contain only the body of the article
*
Fixed some @params documentation Swapped some "$var type" to "type $var" or added missing types before the $var. Changed some other types to match the more common spelling. Makes beginning of some text in captial. Change-Id: I64e8cfe478cb0ba438f40b0631d6e9049cdab567
2014-04-10 18:50:10 +00:00
* @return bool
Step one in fixing OutputPage's documentation: * Groupped methods by function * Added documentation * Fixed some doxygen warnings * Some little style fixes (more to come)
2010-01-28 21:58:50 +00:00
*/
public function getArticleBodyOnly() {
return $this->mArticleBodyOnly;
Whitespace & coding standards adjustments to Evan's last commits. Metadata.php seems to be missing, so the new code doesn't work yet.
2004-04-04 22:33:11 +00:00
}
Added code to implement Dublin Core and Creative Commons RDF metadata. If these features are turned on (with global flags), a metadata link tag is provided to the appropriate action ('dublincore' or 'creativecommons'). When the wiki script receives these actions, it creates RDF files on the fly. Also includes two valuable general functions: one to show an HTTP error (used when the flags are turned off), and one to do content negotiation (used to determine the MIME type for RDF files). A helper function is also added to page output to display metadata links.
2004-04-04 21:58:05 +00:00
Allow adding arbitrary properties to OutputPage Helps to pass extension-specific data Change-Id: Iea98510abc100128733361f2d2287802f35bb359
2013-02-08 22:58:38 +00:00
/**
* Set an additional output property
* @since 1.21
*
* @param string $name
* @param mixed $value
*/
public function setProperty( $name, $value ) {
$this->mProperties[$name] = $value;
}
/**
* Get an additional output property
* @since 1.21
*
Removed unneeded spaces and colons in @param and friends Also swapped some "$var type" to "type $var" or added missing types before the $var. Changed some other types to match the more common spelling. Makes beginning of some text in captial. Change-Id: Ic36c8c7820a6c2d603f1138130670c6bf6a1ca59
2014-04-08 15:29:17 +00:00
* @param string $name
* @return mixed Property value or null if not found
Allow adding arbitrary properties to OutputPage Helps to pass extension-specific data Change-Id: Iea98510abc100128733361f2d2287802f35bb359
2013-02-08 22:58:38 +00:00
*/
public function getProperty( $name ) {
Use PHP 7 '??' operator instead of if-then-else Change-Id: I790b86e2e9e3e41386144637659516a4bfca1cfe
2018-06-12 20:44:33 +00:00
return $this->mProperties[$name] ?? null;
Allow adding arbitrary properties to OutputPage Helps to pass extension-specific data Change-Id: Iea98510abc100128733361f2d2287802f35bb359
2013-02-08 22:58:38 +00:00
}
Changing comments layout preparing for generated documentation with Phpdocumentor
2004-09-02 23:28:24 +00:00
/**
* checkLastModified tells the client to use the client-cached page if
Fixing typo Change-Id: I930c37ad7550c7dd0c4856c22921155dec6ccb3b
2012-12-01 16:02:15 +00:00
* possible. If successful, the OutputPage is disabled so that
Add some documentation to OutputPage.php and specify public/private for most of its member functions (practically all public anyway). And replace one gratuitous use of strcmp with good old ===.
2006-11-07 05:37:31 +00:00
* any future call to OutputPage->output() have no effect.
*
* Generalised OutputPage::checkLastModified() to allow it to easily handle more than three types of modified timestamp. Added OutputPageCheckLastModified hook. * Flattened out the structure of that function with early returns, to avoid ugly deep nesting.
2008-08-29 08:40:13 +00:00
* Side effect: sets mLastModified for Last-Modified header
*
Removed unneeded spaces and colons in @param and friends Also swapped some "$var type" to "type $var" or added missing types before the $var. Changed some other types to match the more common spelling. Makes beginning of some text in captial. Change-Id: Ic36c8c7820a6c2d603f1138130670c6bf6a1ca59
2014-04-08 15:29:17 +00:00
* @param string $timestamp
More adding, updating and tweaking of documentation
2011-05-21 19:07:24 +00:00
*
Removed unneeded spaces and colons in @param and friends Also swapped some "$var type" to "type $var" or added missing types before the $var. Changed some other types to match the more common spelling. Makes beginning of some text in captial. Change-Id: Ic36c8c7820a6c2d603f1138130670c6bf6a1ca59
2014-04-08 15:29:17 +00:00
* @return bool True if cache-ok headers was sent.
Changing comments layout preparing for generated documentation with Phpdocumentor
2004-09-02 23:28:24 +00:00
*/
Step one in fixing OutputPage's documentation: * Groupped methods by function * Added documentation * Fixed some doxygen warnings * Some little style fixes (more to come)
2010-01-28 21:58:50 +00:00
public function checkLastModified( $timestamp ) {
Fixed warning due to default timestamp from Article being passed to wfTimestamp Warning was: gmmktime(): Windows does not support negative values for this function in <b>c:\\htdocs\\w\\includes\\GlobalFunctions.php</b> on line <b>1017</b>
2005-03-20 03:59:06 +00:00
if ( !$timestamp || $timestamp == '19700101000000' ) {
Remove terminating line breaks from debug messages A terminating line break has not been required in wfDebug() since 2014, however no migration was done. Some of these line breaks found their way into LoggerInterface::debug() calls, where they mess up the formatting of the debug log. So, remove terminating line breaks from wfDebug() and LoggerInterface::debug() calls. Also: * Fix the stripping of leading line breaks from the log header emitted by Setup.php. This feature, accidentally broken in 2014, allows requests to be distinguished in the log file. * Avoid using the global variable $self. * Move the logging of the client IP back to Setup.php. It was moved to WebRequest in the hopes that it would not always be needed, however $wgRequest->getIP() is now called unconditionally a few lines up in Setup.php. This means that it is put in its proper place after the "start request" message. * Wrap the log header code in a closure so that variables like $name do not leak into global scope. * In Linker.php, remove a few instances of an unnecessary second parameter to wfDebug(). Change-Id: I96651d3044a95b9d210b51cb8368edc76bebbb9e
2020-06-01 05:00:39 +00:00
wfDebug( __METHOD__ . ": CACHE DISABLED, NO TIMESTAMP" );
* Generalised OutputPage::checkLastModified() to allow it to easily handle more than three types of modified timestamp. Added OutputPageCheckLastModified hook. * Flattened out the structure of that function with early returns, to avoid ugly deep nesting.
2008-08-29 08:40:13 +00:00
return false;
Fixed warning due to default timestamp from Article being passed to wfTimestamp Warning was: gmmktime(): Windows does not support negative values for this function in <b>c:\\htdocs\\w\\includes\\GlobalFunctions.php</b> on line <b>1017</b>
2005-03-20 03:59:06 +00:00
}
Use config instead of globals for OutputPage Change-Id: I5e0ebc173631d1d1052de7ccee4ef839c7c1767f
2014-08-23 08:52:41 +00:00
$config = $this->getConfig();
Use name constants to access config settings. Use name constants instead of string literals in calls to Config::get and ServiceOptions::get, when referring to core configuration variables. This protects against typos and makes the decumentation and schema declaration of the config settings discoverable. This is the first batch, only touching files directly under /includes/ Change-Id: I7252e636c7c86d950d9257b33491af492c6dd5eb
2022-04-01 15:58:32 +00:00
if ( !$config->get( MainConfigNames::CachePages ) ) {
Remove terminating line breaks from debug messages A terminating line break has not been required in wfDebug() since 2014, however no migration was done. Some of these line breaks found their way into LoggerInterface::debug() calls, where they mess up the formatting of the debug log. So, remove terminating line breaks from wfDebug() and LoggerInterface::debug() calls. Also: * Fix the stripping of leading line breaks from the log header emitted by Setup.php. This feature, accidentally broken in 2014, allows requests to be distinguished in the log file. * Avoid using the global variable $self. * Move the logging of the client IP back to Setup.php. It was moved to WebRequest in the hopes that it would not always be needed, however $wgRequest->getIP() is now called unconditionally a few lines up in Setup.php. This means that it is put in its proper place after the "start request" message. * Wrap the log header code in a closure so that variables like $name do not leak into global scope. * In Linker.php, remove a few instances of an unnecessary second parameter to wfDebug(). Change-Id: I96651d3044a95b9d210b51cb8368edc76bebbb9e
2020-06-01 05:00:39 +00:00
wfDebug( __METHOD__ . ": CACHE DISABLED" );
* Generalised OutputPage::checkLastModified() to allow it to easily handle more than three types of modified timestamp. Added OutputPageCheckLastModified hook. * Flattened out the structure of that function with early returns, to avoid ugly deep nesting.
2008-08-29 08:40:13 +00:00
return false;
cache tweaks
2003-07-10 04:55:41 +00:00
}
* Changed inclusion syntax to allow e.g. {{stub}} * Split MediaWiki namespace into MediaWiki and Template (requires changes to all language files) * Purge links to on edit of Template namespace * General refactoring of purging and cache invalidation code
2004-03-20 15:03:26 +00:00
* Generalised OutputPage::checkLastModified() to allow it to easily handle more than three types of modified timestamp. Added OutputPageCheckLastModified hook. * Flattened out the structure of that function with early returns, to avoid ugly deep nesting.
2008-08-29 08:40:13 +00:00
$timestamp = wfTimestamp( TS_MW, $timestamp );
Convert all array() syntax to [] Per wikitech-l consensus: https://lists.wikimedia.org/pipermail/wikitech-l/2016-February/084821.html Notes: * Disabled CallTimePassByReference due to false positives (T127163) Change-Id: I2c8ce713ce6600a0bb7bf67537c87044c7a45c4b
2016-02-17 09:09:32 +00:00
$modifiedTimes = [
else if -> elseif Also fixed some spacing while at it Added some braces for one line statements Change-Id: Iebfa2b17091509daabc76248121c7763e2295d20
2013-04-17 14:52:47 +00:00
'page' => $timestamp,
'user' => $this->getUser()->getTouched(),
Use name constants to access config settings. Use name constants instead of string literals in calls to Config::get and ServiceOptions::get, when referring to core configuration variables. This protects against typos and makes the decumentation and schema declaration of the config settings discoverable. This is the first batch, only touching files directly under /includes/ Change-Id: I7252e636c7c86d950d9257b33491af492c6dd5eb
2022-04-01 15:58:32 +00:00
'epoch' => $config->get( MainConfigNames::CacheEpoch )
Convert all array() syntax to [] Per wikitech-l consensus: https://lists.wikimedia.org/pipermail/wikitech-l/2016-February/084821.html Notes: * Disabled CallTimePassByReference due to false positives (T127163) Change-Id: I2c8ce713ce6600a0bb7bf67537c87044c7a45c4b
2016-02-17 09:09:32 +00:00
];
Use name constants to access config settings. Use name constants instead of string literals in calls to Config::get and ServiceOptions::get, when referring to core configuration variables. This protects against typos and makes the decumentation and schema declaration of the config settings discoverable. This is the first batch, only touching files directly under /includes/ Change-Id: I7252e636c7c86d950d9257b33491af492c6dd5eb
2022-04-01 15:58:32 +00:00
if ( $config->get( MainConfigNames::UseCdn ) ) {
OutputPage: Factor out CdnCacheEpoch logic and cover with tests Bug: T178629 Change-Id: Ife7dd79677c2b5353317e06ac7ed521edd6193cc
2018-04-20 15:00:32 +00:00
$modifiedTimes['sepoch'] = wfTimestamp( TS_MW, $this->getCdnCacheEpoch(
time(),
Use name constants to access config settings. Use name constants instead of string literals in calls to Config::get and ServiceOptions::get, when referring to core configuration variables. This protects against typos and makes the decumentation and schema declaration of the config settings discoverable. This is the first batch, only touching files directly under /includes/ Change-Id: I7252e636c7c86d950d9257b33491af492c6dd5eb
2022-04-01 15:58:32 +00:00
$config->get( MainConfigNames::CdnMaxAge )
OutputPage: Factor out CdnCacheEpoch logic and cover with tests Bug: T178629 Change-Id: Ife7dd79677c2b5353317e06ac7ed521edd6193cc
2018-04-20 15:00:32 +00:00
) );
Account for $wgSquidMaxage in OutputPage::checkLastModified(). * This deals with the fact that seldom edited pages can end up cached with very stale resource (JS/CSS) references since the response to IMS GET requests will be 304 Not Modified if page_touched is ancient. When squid revalidates its stale cache it will keep getting 304s and renewing the TTL on the stale cache. Bug: 44570 Change-Id: I3889f300012aeabd37e228653279ad19b296e4ae
2013-04-09 20:55:36 +00:00
}
Hooks::run() call site migration Migrate all callers of Hooks::run() to use the new HookContainer/HookRunner system. General principles: * Use DI if it is already used. We're not changing the way state is managed in this patch. * HookContainer is always injected, not HookRunner. HookContainer is a service, it's a more generic interface, it is the only thing that provides isRegistered() which is needed in some cases, and a HookRunner can be efficiently constructed from it (confirmed by benchmark). Because HookContainer is needed for object construction, it is also needed by all factories. * "Ask your friendly local base class". Big hierarchies like SpecialPage and ApiBase have getHookContainer() and getHookRunner() methods in the base class, and classes that extend that base class are not expected to know or care where the base class gets its HookContainer from. * ProtectedHookAccessorTrait provides protected getHookContainer() and getHookRunner() methods, getting them from the global service container. The point of this is to ease migration to DI by ensuring that call sites ask their local friendly base class rather than getting a HookRunner from the service container directly. * Private $this->hookRunner. In some smaller classes where accessor methods did not seem warranted, there is a private HookRunner property which is accessed directly. Very rarely (two cases), there is a protected property, for consistency with code that conventionally assumes protected=private, but in cases where the class might actually be overridden, a protected accessor is preferred over a protected property. * The last resort: Hooks::runner(). Mostly for static, file-scope and global code. In a few cases it was used for objects with broken construction schemes, out of horror or laziness. Constructors with new required arguments: * AuthManager * BadFileLookup * BlockManager * ClassicInterwikiLookup * ContentHandlerFactory * ContentSecurityPolicy * DefaultOptionsManager * DerivedPageDataUpdater * FullSearchResultWidget * HtmlCacheUpdater * LanguageFactory * LanguageNameUtils * LinkRenderer * LinkRendererFactory * LocalisationCache * MagicWordFactory * MessageCache * NamespaceInfo * PageEditStash * PageHandlerFactory * PageUpdater * ParserFactory * PermissionManager * RevisionStore * RevisionStoreFactory * SearchEngineConfig * SearchEngineFactory * SearchFormWidget * SearchNearMatcher * SessionBackend * SpecialPageFactory * UserNameUtils * UserOptionsManager * WatchedItemQueryService * WatchedItemStore Constructors with new optional arguments: * DefaultPreferencesFactory * Language * LinkHolderArray * MovePage * Parser * ParserCache * PasswordReset * Router setHookContainer() now required after construction: * AuthenticationProvider * ResourceLoaderModule * SearchEngine Change-Id: Id442b0dbe43aba84bd5cf801d86dedc768b082c7
2020-03-19 02:42:09 +00:00
$this->getHookRunner()->onOutputPageCheckLastModified( $modifiedTimes, $this );
* remove end of line whitespace * remove empty lines at end of file * remove "?>" where still present
2008-04-14 07:45:50 +00:00
* Generalised OutputPage::checkLastModified() to allow it to easily handle more than three types of modified timestamp. Added OutputPageCheckLastModified hook. * Flattened out the structure of that function with early returns, to avoid ugly deep nesting.
2008-08-29 08:40:13 +00:00
$maxModified = max( $modifiedTimes );
$this->mLastModified = wfTimestamp( TS_RFC2822, $maxModified );
Use WebRequest instead of $_SERVER in OutputPage. Also moved debug string generation a bit lower so that it's executed only when it will actually be used. Change-Id: Ia01e7fb1dba1f2f55c156d3ac097acf2a68dca70
2012-07-19 16:35:25 +00:00
$clientHeader = $this->getRequest()->getHeader( 'If-Modified-Since' );
if ( $clientHeader === false ) {
Remove use of $dest = 'log' for wfDebugLog The functionality for marking a debug log message for delivery to a log file but not to MWDebug was removed in 1.25 with the introduction of the PSR-3 logging system. Convert the messages that were marked for this special handling in a7a0883 to use 'private' instead. Bug: T122644 Change-Id: Iefaac07a5922c16c2899904c7b678199c5b1efe9
2015-12-30 19:48:01 +00:00
wfDebug( __METHOD__ . ": client did not send If-Modified-Since header", 'private' );
* Generalised OutputPage::checkLastModified() to allow it to easily handle more than three types of modified timestamp. Added OutputPageCheckLastModified hook. * Flattened out the structure of that function with early returns, to avoid ugly deep nesting.
2008-08-29 08:40:13 +00:00
return false;
Initial revision
2003-04-14 23:10:40 +00:00
}
* Generalised OutputPage::checkLastModified() to allow it to easily handle more than three types of modified timestamp. Added OutputPageCheckLastModified hook. * Flattened out the structure of that function with early returns, to avoid ugly deep nesting.
2008-08-29 08:40:13 +00:00
# IE sends sizes after the date like this:
# Wed, 20 Aug 2003 06:51:19 GMT; length=5202
# this breaks strtotime().
Use WebRequest instead of $_SERVER in OutputPage. Also moved debug string generation a bit lower so that it's executed only when it will actually be used. Change-Id: Ia01e7fb1dba1f2f55c156d3ac097acf2a68dca70
2012-07-19 16:35:25 +00:00
$clientHeader = preg_replace( '/;.*$/', '', $clientHeader );
* Generalised OutputPage::checkLastModified() to allow it to easily handle more than three types of modified timestamp. Added OutputPageCheckLastModified hook. * Flattened out the structure of that function with early returns, to avoid ugly deep nesting.
2008-08-29 08:40:13 +00:00
Replace some more usages of Wikimedia\(suppress|restore)Warnings() Change-Id: I2eb133a9e32116cd155f59086245bc4d15ecbfcc
2022-02-24 20:17:53 +00:00
// E_STRICT system time warnings
AtEase::suppressWarnings();
Revert r44801 "Tweaks from profiling" Mostly seems to be formatting tweaks, loss of useful debug log output, and mysterious unexplained changes. If some of these tweaks actually are based on profiling data, please provide details, such as "rearranging this call reduces service time from 80ms to 50ms for client cache hits on my machine" or whatever.
2008-12-23 19:39:00 +00:00
$clientHeaderTime = strtotime( $clientHeader );
Replace some more usages of Wikimedia\(suppress|restore)Warnings() Change-Id: I2eb133a9e32116cd155f59086245bc4d15ecbfcc
2022-02-24 20:17:53 +00:00
AtEase::restoreWarnings();
Revert r44801 "Tweaks from profiling" Mostly seems to be formatting tweaks, loss of useful debug log output, and mysterious unexplained changes. If some of these tweaks actually are based on profiling data, please provide details, such as "rearranging this call reduces service time from 80ms to 50ms for client cache hits on my machine" or whatever.
2008-12-23 19:39:00 +00:00
if ( !$clientHeaderTime ) {
Pass php-strict for OutputPage Change-Id: Ia5d03277fff543ca672dc8efc7eaaeca4888ac23
2014-05-11 01:54:15 +00:00
wfDebug( __METHOD__
Remove terminating line breaks from debug messages A terminating line break has not been required in wfDebug() since 2014, however no migration was done. Some of these line breaks found their way into LoggerInterface::debug() calls, where they mess up the formatting of the debug log. So, remove terminating line breaks from wfDebug() and LoggerInterface::debug() calls. Also: * Fix the stripping of leading line breaks from the log header emitted by Setup.php. This feature, accidentally broken in 2014, allows requests to be distinguished in the log file. * Avoid using the global variable $self. * Move the logging of the client IP back to Setup.php. It was moved to WebRequest in the hopes that it would not always be needed, however $wgRequest->getIP() is now called unconditionally a few lines up in Setup.php. This means that it is put in its proper place after the "start request" message. * Wrap the log header code in a closure so that variables like $name do not leak into global scope. * In Linker.php, remove a few instances of an unnecessary second parameter to wfDebug(). Change-Id: I96651d3044a95b9d210b51cb8368edc76bebbb9e
2020-06-01 05:00:39 +00:00
. ": unable to parse the client's If-Modified-Since header: $clientHeader" );
* Generalised OutputPage::checkLastModified() to allow it to easily handle more than three types of modified timestamp. Added OutputPageCheckLastModified hook. * Flattened out the structure of that function with early returns, to avoid ugly deep nesting.
2008-08-29 08:40:13 +00:00
return false;
}
$clientHeaderTime = wfTimestamp( TS_MW, $clientHeaderTime );
Use WebRequest instead of $_SERVER in OutputPage. Also moved debug string generation a bit lower so that it's executed only when it will actually be used. Change-Id: Ia01e7fb1dba1f2f55c156d3ac097acf2a68dca70
2012-07-19 16:35:25 +00:00
# Make debug info
$info = '';
foreach ( $modifiedTimes as $name => $value ) {
if ( $info !== '' ) {
$info .= ', ';
}
$info .= "$name=" . wfTimestamp( TS_ISO_8601, $value );
}
here it is ... the upload-api, script-server, js2 (javascript phase2) branch merge 1st attempt. Here is a short overview of changes and associated default configuration variables (most everything is off by default) also see ~soon to be updated~: http://www.mediawiki.org/wiki/Media_Projects_Overview = Upload Improvements = ==Upload API == * Based on the early work of Bryan Tong and others it adds the upload option to the api. * We rewrite Special:Upload page to include use the new refactoring * Added in token checks in both the SpecialUpload.php page so avoids DOS / xss copy-by-url JavaScript based cross site POST file submissions == Copy by URL== $wgAllowCopyUploads = false; * http class rewrite includes a new http background download see: includes/HttpFunctions.php * spins off a php process that calls: maintenance/http_session_download.php * pushes updates to the session and gives the user a progress bar on http copy uploads from other server progress (using js2 upload interface) (if not using the js2 upload interface it does the request in-place but the download is limited to the php ini timeout time) == Firefogg == * Firefogg enables resumable upload by chunks * progress indicators and conditional invokation (js2 system) * and of-course client side transcoding. = Script Server = $wgEnableScriptLoader = false; * off by default if $wgEnableScriptLoader is turned on script files are grouped, gziped, cached etc. for more info see: http://www.mediawiki.org/wiki/Extension:ScriptLoader * Includes some early skin js include fixes (skin/script system still lots of love) * Includes a "javascript class autoloader" this is packaged into mwEmbed so that the mwEmbed library can work in stand alone mode (while retaining localization and script serving) (one such application is the make page for firefogg.org : http://www.firefogg.org/make/index.html ) * The file that contains the autojavascript loading classes is: js2/php/jsAutoloadLocalClasses.php * One can use this auto class loading dependency system with extensions and add-ons but I need to better document that. = js2 system / mwEmbed= $wgEnableJS2system = false * includes initial rewrite towards more jquery based javascript code * especially for the Special:Upload page. * Also the edit page include support for the "add-media-wizard" * includes dependency loader for javascript that optionally takes advantage of the script-loader * remote embedding of javascript interfaces (like embedding video, or commons media searching) * $wgDebugJavaScript = false; .. .this variable lets you always get "always fresh javascript". When used with the script-loader it does not minify the script-loader output. = mwEmbed = * Will commit a separate patch to oggHandler that conditionally outputs <video tag> to use the new javascript video player. ** mv_embed player includes: play-head, volume control, remote embedding, oggz-chop support across plugins. * add-media-wizard adds easy inserts of media to pages (with import) == jQuery== * we include a base install of jQuery, jQuery ui and some plugins. * all the javascript classes are in the scriptloader so its easy to load any set of jquery ui components that you may need using the script-server. You get a callback so you can then execute js with dependencies loaded. == other stuff == there is a bit more code in js2 that pertains to sequence editing, timed text display and basic image editing. We include a base import of pixastic-lib & pixastic-editor... will work with the pixastic developer to try and ensure upstream compatibility on our usage of the library for in-browser photo and sequence manipulation.
2009-07-14 23:52:14 +00:00
wfDebug( __METHOD__ . ": client sent If-Modified-Since: " .
Remove use of $dest = 'log' for wfDebugLog The functionality for marking a debug log message for delivery to a log file but not to MWDebug was removed in 1.25 with the introduction of the PSR-3 logging system. Convert the messages that were marked for this special handling in a7a0883 to use 'private' instead. Bug: T122644 Change-Id: Iefaac07a5922c16c2899904c7b678199c5b1efe9
2015-12-30 19:48:01 +00:00
wfTimestamp( TS_ISO_8601, $clientHeaderTime ), 'private' );
here it is ... the upload-api, script-server, js2 (javascript phase2) branch merge 1st attempt. Here is a short overview of changes and associated default configuration variables (most everything is off by default) also see ~soon to be updated~: http://www.mediawiki.org/wiki/Media_Projects_Overview = Upload Improvements = ==Upload API == * Based on the early work of Bryan Tong and others it adds the upload option to the api. * We rewrite Special:Upload page to include use the new refactoring * Added in token checks in both the SpecialUpload.php page so avoids DOS / xss copy-by-url JavaScript based cross site POST file submissions == Copy by URL== $wgAllowCopyUploads = false; * http class rewrite includes a new http background download see: includes/HttpFunctions.php * spins off a php process that calls: maintenance/http_session_download.php * pushes updates to the session and gives the user a progress bar on http copy uploads from other server progress (using js2 upload interface) (if not using the js2 upload interface it does the request in-place but the download is limited to the php ini timeout time) == Firefogg == * Firefogg enables resumable upload by chunks * progress indicators and conditional invokation (js2 system) * and of-course client side transcoding. = Script Server = $wgEnableScriptLoader = false; * off by default if $wgEnableScriptLoader is turned on script files are grouped, gziped, cached etc. for more info see: http://www.mediawiki.org/wiki/Extension:ScriptLoader * Includes some early skin js include fixes (skin/script system still lots of love) * Includes a "javascript class autoloader" this is packaged into mwEmbed so that the mwEmbed library can work in stand alone mode (while retaining localization and script serving) (one such application is the make page for firefogg.org : http://www.firefogg.org/make/index.html ) * The file that contains the autojavascript loading classes is: js2/php/jsAutoloadLocalClasses.php * One can use this auto class loading dependency system with extensions and add-ons but I need to better document that. = js2 system / mwEmbed= $wgEnableJS2system = false * includes initial rewrite towards more jquery based javascript code * especially for the Special:Upload page. * Also the edit page include support for the "add-media-wizard" * includes dependency loader for javascript that optionally takes advantage of the script-loader * remote embedding of javascript interfaces (like embedding video, or commons media searching) * $wgDebugJavaScript = false; .. .this variable lets you always get "always fresh javascript". When used with the script-loader it does not minify the script-loader output. = mwEmbed = * Will commit a separate patch to oggHandler that conditionally outputs <video tag> to use the new javascript video player. ** mv_embed player includes: play-head, volume control, remote embedding, oggz-chop support across plugins. * add-media-wizard adds easy inserts of media to pages (with import) == jQuery== * we include a base install of jQuery, jQuery ui and some plugins. * all the javascript classes are in the scriptloader so its easy to load any set of jquery ui components that you may need using the script-server. You get a callback so you can then execute js with dependencies loaded. == other stuff == there is a bit more code in js2 that pertains to sequence editing, timed text display and basic image editing. We include a base import of pixastic-lib & pixastic-editor... will work with the pixastic developer to try and ensure upstream compatibility on our usage of the library for in-browser photo and sequence manipulation.
2009-07-14 23:52:14 +00:00
wfDebug( __METHOD__ . ": effective Last-Modified: " .
Remove use of $dest = 'log' for wfDebugLog The functionality for marking a debug log message for delivery to a log file but not to MWDebug was removed in 1.25 with the introduction of the PSR-3 logging system. Convert the messages that were marked for this special handling in a7a0883 to use 'private' instead. Bug: T122644 Change-Id: Iefaac07a5922c16c2899904c7b678199c5b1efe9
2015-12-30 19:48:01 +00:00
wfTimestamp( TS_ISO_8601, $maxModified ), 'private' );
else if -> elseif Also fixed some spacing while at it Added some braces for one line statements Change-Id: Iebfa2b17091509daabc76248121c7763e2295d20
2013-04-17 14:52:47 +00:00
if ( $clientHeaderTime < $maxModified ) {
Remove use of $dest = 'log' for wfDebugLog The functionality for marking a debug log message for delivery to a log file but not to MWDebug was removed in 1.25 with the introduction of the PSR-3 logging system. Convert the messages that were marked for this special handling in a7a0883 to use 'private' instead. Bug: T122644 Change-Id: Iefaac07a5922c16c2899904c7b678199c5b1efe9
2015-12-30 19:48:01 +00:00
wfDebug( __METHOD__ . ": STALE, $info", 'private' );
* Generalised OutputPage::checkLastModified() to allow it to easily handle more than three types of modified timestamp. Added OutputPageCheckLastModified hook. * Flattened out the structure of that function with early returns, to avoid ugly deep nesting.
2008-08-29 08:40:13 +00:00
return false;
}
Revert r44801 "Tweaks from profiling" Mostly seems to be formatting tweaks, loss of useful debug log output, and mysterious unexplained changes. If some of these tweaks actually are based on profiling data, please provide details, such as "rearranging this call reduces service time from 80ms to 50ms for client cache hits on my machine" or whatever.
2008-12-23 19:39:00 +00:00
# Not modified
WebResponse: Implement statusHeader() using the new HttpStatus::header() * Convert existing use of WebResponse::header() for HTTP status headers to use this new statusHeader() method. * Extend unit test forFauxResponse. I'm not calling HttpStatus::header directly in code. We keep the abstraction layer of WebResponse so that responses can continue to be mocked/fauxed without affecting the outer HTTP response. Change-Id: I8a536e16659fa88b54cffa1457efb889efa5fcd6
2015-05-24 12:31:11 +00:00
# Give a 304 Not Modified response code and disable body output
Remove use of $dest = 'log' for wfDebugLog The functionality for marking a debug log message for delivery to a log file but not to MWDebug was removed in 1.25 with the introduction of the PSR-3 logging system. Convert the messages that were marked for this special handling in a7a0883 to use 'private' instead. Bug: T122644 Change-Id: Iefaac07a5922c16c2899904c7b678199c5b1efe9
2015-12-30 19:48:01 +00:00
wfDebug( __METHOD__ . ": NOT MODIFIED, $info", 'private' );
phan: Disable scalar_implicit_cast setting Make phan stricter about scalar types by setting scalar_implicit_cast to false (the default in mediawiki-phan-config) Bug: T242536 Bug: T301991 Change-Id: Ia2fe30b17804186571722e728578121c8b75d455
2022-03-16 23:34:23 +00:00
// @phan-suppress-next-line PhanTypeMismatchArgumentInternal Scalar okay with php8.1
*coding style cleanup (added braces, spacing tweaks etc.) *removed some unused globals
2010-05-22 12:18:22 +00:00
ini_set( 'zlib.output_compression', 0 );
WebResponse: Implement statusHeader() using the new HttpStatus::header() * Convert existing use of WebResponse::header() for HTTP status headers to use this new statusHeader() method. * Extend unit test forFauxResponse. I'm not calling HttpStatus::header directly in code. We keep the abstraction layer of WebResponse so that responses can continue to be mocked/fauxed without affecting the outer HTTP response. Change-Id: I8a536e16659fa88b54cffa1457efb889efa5fcd6
2015-05-24 12:31:11 +00:00
$this->getRequest()->response()->statusHeader( 304 );
* Generalised OutputPage::checkLastModified() to allow it to easily handle more than three types of modified timestamp. Added OutputPageCheckLastModified hook. * Flattened out the structure of that function with early returns, to avoid ugly deep nesting.
2008-08-29 08:40:13 +00:00
$this->sendCacheControl();
$this->disable();
// Don't output a compressed blob when using ob_gzhandler;
// it's technically against HTTP spec and seems to confuse
// Firefox when the response gets split over two packets.
Deprecate a bunch of global functions * wfAcceptToPrefs * wfClearOutputBuffers * wfConfiguredReadOnlyReason * wfDebugMem * wfGetPrecompiledData * wfNegotiateType Bug: T264976 Bug: T264979 Bug: T264981 Bug: T264983 Bug: T264984 Bug: T264985 Change-Id: Ia05bc84e4d1be7c8a02472f32e2c009e4bb32032
2020-12-18 19:57:09 +00:00
wfResetOutputBuffers( false );
* Generalised OutputPage::checkLastModified() to allow it to easily handle more than three types of modified timestamp. Added OutputPageCheckLastModified hook. * Flattened out the structure of that function with early returns, to avoid ugly deep nesting.
2008-08-29 08:40:13 +00:00
return true;
Initial revision
2003-04-14 23:10:40 +00:00
}
OutputPage: Factor out CdnCacheEpoch logic and cover with tests Bug: T178629 Change-Id: Ife7dd79677c2b5353317e06ac7ed521edd6193cc
2018-04-20 15:00:32 +00:00
/**
* @param int $reqTime Time of request (eg. now)
* @param int $maxAge Cache TTL in seconds
* @return int Timestamp
*/
private function getCdnCacheEpoch( $reqTime, $maxAge ) {
Rename CDN config variables to be generic, deprecating the old names Hook, methods, classes still to rename where appropriate. Bug: T104148 Depends-On: Id34339dff88bc6d1863378ac94b96b2d590b891d Depends-On: I4e2938395bcbf7956b83fff00978f09c61dcfa36 Change-Id: I7a725dae551c867a4fa7c213838d52c7fb862756
2017-11-01 20:55:24 +00:00
// Ensure Last-Modified is never more than $wgCdnMaxAge in the past,
OutputPage: Factor out CdnCacheEpoch logic and cover with tests Bug: T178629 Change-Id: Ife7dd79677c2b5353317e06ac7ed521edd6193cc
2018-04-20 15:00:32 +00:00
// because even if the wiki page content hasn't changed since, static
// resources may have changed (skin HTML, interface messages, urls, etc.)
// and must roll-over in a timely manner (T46570)
return $reqTime - $maxAge;
}
* (bug 21916) Last-Modified header is now correct when outputting cached feed
2010-04-05 16:14:58 +00:00
/**
* Override the last modified timestamp
*
Cleanup some docs (includes/*.php) - Swap "$variable type" to "type $variable" - Added missing types - Fixed spacing inside docs - Makes beginning of @param/@return/@var/@throws in capital - Changed some types to match the more common spelling Change-Id: I783e4dbfe5f6f98b32b9a03ccf6439e13e132bcc
2014-07-24 17:42:24 +00:00
* @param string $timestamp New timestamp, in a format readable by
* (bug 21916) Last-Modified header is now correct when outputting cached feed
2010-04-05 16:14:58 +00:00
* wfTimestamp()
*/
public function setLastModified( $timestamp ) {
$this->mLastModified = wfTimestamp( TS_RFC2822, $timestamp );
}
Refactor a bit preparatory to fixing bug 8068: rewrite the robot policy stuff in OutputPage to allow index and follow policy to be set separately. Also now validates input to setRobotPolicy(). And renamed setRobotpolicy to setRobotPolicy, too. If anyone was accessing $mRobotpolicy directly they're out of luck, though.
2008-07-23 19:05:43 +00:00
/**
* Set the robot policy for the page: <http://www.robotstxt.org/meta.html>
*
Cleanup some docs (includes/*.php) - Swap "$variable type" to "type $variable" - Added missing types - Fixed spacing inside docs - Makes beginning of @param/@return/@var/@throws in capital - Changed some types to match the more common spelling Change-Id: I783e4dbfe5f6f98b32b9a03ccf6439e13e132bcc
2014-07-24 17:42:24 +00:00
* @param string $policy The literal string to output as the contents of
Revert "Add $robotPolicy parameter to SpecialPage::setHeaders()" This reverts commit 8c7541da3b12d4727743ed1d3447f6ac6c3b6846. Bug: 57883 Change-Id: I582088528286ff34da164fa4457dd203f11df7ef
2013-12-02 21:20:11 +00:00
* the meta tag. Will be parsed according to the spec and output in
* standardized form.
Refactor a bit preparatory to fixing bug 8068: rewrite the robot policy stuff in OutputPage to allow index and follow policy to be set separately. Also now validates input to setRobotPolicy(). And renamed setRobotpolicy to setRobotPolicy, too. If anyone was accessing $mRobotpolicy directly they're out of luck, though.
2008-07-23 19:05:43 +00:00
*/
public function setRobotPolicy( $policy ) {
(bug14900) Make __INDEX__ and __NOINDEX__ not override $wgArticleRobotPolicies where the two conflict.
2009-08-31 19:19:12 +00:00
$policy = Article::formatRobotPolicy( $policy );
Refactor a bit preparatory to fixing bug 8068: rewrite the robot policy stuff in OutputPage to allow index and follow policy to be set separately. Also now validates input to setRobotPolicy(). And renamed setRobotpolicy to setRobotPolicy, too. If anyone was accessing $mRobotpolicy directly they're out of luck, though.
2008-07-23 19:05:43 +00:00
else if -> elseif Also fixed some spacing while at it Added some braces for one line statements Change-Id: Iebfa2b17091509daabc76248121c7763e2295d20
2013-04-17 14:52:47 +00:00
if ( isset( $policy['index'] ) ) {
(bug14900) Make __INDEX__ and __NOINDEX__ not override $wgArticleRobotPolicies where the two conflict.
2009-08-31 19:19:12 +00:00
$this->setIndexPolicy( $policy['index'] );
Step one in fixing OutputPage's documentation: * Groupped methods by function * Added documentation * Fixed some doxygen warnings * Some little style fixes (more to come)
2010-01-28 21:58:50 +00:00
}
else if -> elseif Also fixed some spacing while at it Added some braces for one line statements Change-Id: Iebfa2b17091509daabc76248121c7763e2295d20
2013-04-17 14:52:47 +00:00
if ( isset( $policy['follow'] ) ) {
(bug14900) Make __INDEX__ and __NOINDEX__ not override $wgArticleRobotPolicies where the two conflict.
2009-08-31 19:19:12 +00:00
$this->setFollowPolicy( $policy['follow'] );
Step one in fixing OutputPage's documentation: * Groupped methods by function * Added documentation * Fixed some doxygen warnings * Some little style fixes (more to come)
2010-01-28 21:58:50 +00:00
}
Refactor a bit preparatory to fixing bug 8068: rewrite the robot policy stuff in OutputPage to allow index and follow policy to be set separately. Also now validates input to setRobotPolicy(). And renamed setRobotpolicy to setRobotPolicy, too. If anyone was accessing $mRobotpolicy directly they're out of luck, though.
2008-07-23 19:05:43 +00:00
}
Add get{Robot,Index,Follow}Policy methods to OutputPage Inspired by wikiHow's RobotPolicy extension, which depends on the existence of OutputPage#getRobotPolicy, as the extension needs to know the current robot policy state. Change-Id: I764f5584a9f04a6b66c53193e9118dd1a5248284
2019-12-11 12:23:51 +00:00
/**
* Get the current robot policy for the page as a string in the form
* <index policy>,<follow policy>.
*
* @return string
*/
public function getRobotPolicy() {
return "{$this->mIndexPolicy},{$this->mFollowPolicy}";
}
Add option to amend robots meta tag - Add private var, helper methods to format and set robots options. - Add reference to new method in OutputPage::getHeadLinksArray(). Bug: T301584 Change-Id: I2652f20c9fe5f2d3e28e329d4eddc2a2ff88cf76
2022-03-10 21:27:21 +00:00
/**
* Format an array of robots options as a string of directives.
*
* @return string The robots policy options.
*/
private function formatRobotsOptions(): string {
$options = $this->mRobotsOptions;
// Check if options array has any non-integer keys.
if ( count( array_filter( array_keys( $options ), 'is_string' ) ) > 0 ) {
// Robots meta tags can have directives that are single strings or
// have parameters that should be formatted like <directive>:<setting>.
// If the options keys are strings, format them accordingly.
// https://developers.google.com/search/docs/advanced/robots/robots_meta_tag
array_walk( $options, static function ( &$value, $key ) {
$value = is_string( $key ) ? "{$key}:{$value}" : "{$value}";
} );
}
return implode( ',', $options );
}
/**
* Set the robots policy with options for the page.
*
* @since 1.38
* @param array $options An array of key-value pairs or a string
* to populate the robots meta tag content attribute as a string.
*/
public function setRobotsOptions( array $options = [] ): void {
$this->mRobotsOptions = array_merge( $this->mRobotsOptions, $options );
}
/**
* Get the robots policy content attribute for the page
* as a string in the form <index policy>,<follow policy>,<options>.
*
* @return string
*/
private function getRobotsContent(): string {
$robotOptionString = $this->formatRobotsOptions();
$robotArgs = ( $this->mIndexPolicy === 'index' &&
$this->mFollowPolicy === 'follow' ) ?
[] :
[
$this->mIndexPolicy,
$this->mFollowPolicy,
];
if ( $robotOptionString ) {
$robotArgs[] = $robotOptionString;
}
return implode( ',', $robotArgs );
}
Refactor a bit preparatory to fixing bug 8068: rewrite the robot policy stuff in OutputPage to allow index and follow policy to be set separately. Also now validates input to setRobotPolicy(). And renamed setRobotpolicy to setRobotPolicy, too. If anyone was accessing $mRobotpolicy directly they're out of luck, though.
2008-07-23 19:05:43 +00:00
/**
* Set the index policy for the page, but leave the follow policy un-
* touched.
*
Fixed @param tags to conform with Doxygen format. Doxygen expects parameter types to come before the parameter name in @param tags. Used a quick regex to switch everything around where possible. This only fixes cases where a primitve variable (or a primitive followed by other types) is the variable type. Other cases will need to be fixed manually. Change-Id: Ic59fd20856eb0489d70f3469a56ebce0efb3db13
2013-03-11 17:15:01 +00:00
* @param string $policy Either 'index' or 'noindex'.
Refactor a bit preparatory to fixing bug 8068: rewrite the robot policy stuff in OutputPage to allow index and follow policy to be set separately. Also now validates input to setRobotPolicy(). And renamed setRobotpolicy to setRobotPolicy, too. If anyone was accessing $mRobotpolicy directly they're out of luck, though.
2008-07-23 19:05:43 +00:00
*/
public function setIndexPolicy( $policy ) {
$policy = trim( $policy );
Convert all array() syntax to [] Per wikitech-l consensus: https://lists.wikimedia.org/pipermail/wikitech-l/2016-February/084821.html Notes: * Disabled CallTimePassByReference due to false positives (T127163) Change-Id: I2c8ce713ce6600a0bb7bf67537c87044c7a45c4b
2016-02-17 09:09:32 +00:00
if ( in_array( $policy, [ 'index', 'noindex' ] ) ) {
Refactor a bit preparatory to fixing bug 8068: rewrite the robot policy stuff in OutputPage to allow index and follow policy to be set separately. Also now validates input to setRobotPolicy(). And renamed setRobotpolicy to setRobotPolicy, too. If anyone was accessing $mRobotpolicy directly they're out of luck, though.
2008-07-23 19:05:43 +00:00
$this->mIndexPolicy = $policy;
}
}
Add get{Robot,Index,Follow}Policy methods to OutputPage Inspired by wikiHow's RobotPolicy extension, which depends on the existence of OutputPage#getRobotPolicy, as the extension needs to know the current robot policy state. Change-Id: I764f5584a9f04a6b66c53193e9118dd1a5248284
2019-12-11 12:23:51 +00:00
/**
* Get the current index policy for the page as a string.
*
* @return string
*/
public function getIndexPolicy() {
return $this->mIndexPolicy;
}
Refactor a bit preparatory to fixing bug 8068: rewrite the robot policy stuff in OutputPage to allow index and follow policy to be set separately. Also now validates input to setRobotPolicy(). And renamed setRobotpolicy to setRobotPolicy, too. If anyone was accessing $mRobotpolicy directly they're out of luck, though.
2008-07-23 19:05:43 +00:00
/**
* Set the follow policy for the page, but leave the index policy un-
* touched.
*
Cleanup some docs (includes/*.php) - Swap "$variable type" to "type $variable" - Added missing types - Fixed spacing inside docs - Makes beginning of @param/@return/@var/@throws in capital - Changed some types to match the more common spelling Change-Id: I783e4dbfe5f6f98b32b9a03ccf6439e13e132bcc
2014-07-24 17:42:24 +00:00
* @param string $policy Either 'follow' or 'nofollow'.
Refactor a bit preparatory to fixing bug 8068: rewrite the robot policy stuff in OutputPage to allow index and follow policy to be set separately. Also now validates input to setRobotPolicy(). And renamed setRobotpolicy to setRobotPolicy, too. If anyone was accessing $mRobotpolicy directly they're out of luck, though.
2008-07-23 19:05:43 +00:00
*/
public function setFollowPolicy( $policy ) {
$policy = trim( $policy );
Convert all array() syntax to [] Per wikitech-l consensus: https://lists.wikimedia.org/pipermail/wikitech-l/2016-February/084821.html Notes: * Disabled CallTimePassByReference due to false positives (T127163) Change-Id: I2c8ce713ce6600a0bb7bf67537c87044c7a45c4b
2016-02-17 09:09:32 +00:00
if ( in_array( $policy, [ 'follow', 'nofollow' ] ) ) {
Refactor a bit preparatory to fixing bug 8068: rewrite the robot policy stuff in OutputPage to allow index and follow policy to be set separately. Also now validates input to setRobotPolicy(). And renamed setRobotpolicy to setRobotPolicy, too. If anyone was accessing $mRobotpolicy directly they're out of luck, though.
2008-07-23 19:05:43 +00:00
$this->mFollowPolicy = $policy;
}
}
Add get{Robot,Index,Follow}Policy methods to OutputPage Inspired by wikiHow's RobotPolicy extension, which depends on the existence of OutputPage#getRobotPolicy, as the extension needs to know the current robot policy state. Change-Id: I764f5584a9f04a6b66c53193e9118dd1a5248284
2019-12-11 12:23:51 +00:00
/**
* Get the current follow policy for the page as a string.
*
* @return string
*/
public function getFollowPolicy() {
return $this->mFollowPolicy;
}
(bug 12998) Allow <sup>, <sub>, etc. in DISPLAYTITLE
2009-04-09 05:15:43 +00:00
/**
escape tags and entity in doxygen comments When inserting XML elements inline <such as this one>, doxygen chokes about it not being known. Simply enclosing the tag in double quotes prevents doxygen from emitting a warning. Also enclosed a few invalid functions calls such as \. and double quoted the HTML entities such as &foobar; Change-Id: I4019637145e683c2bec3d17b2fd98b0c50a932f1
2012-07-10 12:48:06 +00:00
* "HTML title" means the contents of "<title>".
Fix bug 23115 again. Follow up r64821, r64823 and r64827. Rewrite the converted title to HTML title if the HTML title haven't set by other messages.
2010-04-10 05:46:01 +00:00
* It is stored as plain, unescaped text and will be run through htmlspecialchars in the skin file.
More adding, updating and tweaking of documentation
2011-05-21 19:07:24 +00:00
*
Fix param doc of OutputPage::setHTMLTitle Method also accept Message object Change-Id: I3a5254a3776f0215667e64ff9d01d60b80375aca
2014-07-27 15:21:01 +00:00
* @param string|Message $name
(bug 12998) Allow <sup>, <sub>, etc. in DISPLAYTITLE
2009-04-09 05:15:43 +00:00
*/
Proposed cleanup of recent LanguageConverter-related commits: * Moved the responsibility for calling $wgOut->setPageTitle() from OutputPage::addParserOutputNoText() to the OutputPage/Parser caller. Previously, every call to $wgOut->addWikiText() (or any other message parsing function) was resulting in the title being reset to a converted version of $wgTitle, producing bug 23124. Moving responsibility to the caller seems to work fairly well, since there are apparently only two callers that really want {{DISPLAYTITLE}} etc. to work (EditPage and Article::view). * Reverted data flow obfuscation in OutputPage::setHTMLTitle() from r64851, replaced by the above. The caller decides what overrides what. * Reverted inappropriate, cache-polluting references to $wgUser and $wgRequest in Parser::parse(), introduced in r64819. * Reverted incomprehensible boolean parameter to Language::convert() from r64851, reintroduced Language::convertTitle() instead. Gave it a simpler implementation than before. * Fixed broken {{DISPLAYTITLE}} feature, was being unconditionally overwritten by the ParserOutput::setTitleText() call in Parser::parse(). Give {{DISPLAYTITLE}} precedence over autoconverted title, like we do for -{T|...}-. * Tested extensively (perhaps not exhaustively)
2010-04-10 13:38:50 +00:00
public function setHTMLTitle( $name ) {
Per Nikerabbit's comment on r100621: * Removed OutputPage::setPageTitleMsg() and OutputPage::setHTMLTitleMsg() and make OutputPage::setPageTitle() and OutputPage::setHTMLTitle() accept a Message object * Updated core calls (including some that I missed last time because of non-matching case) * Added Message::setContext() and use it in RequestContext so that I don't need to duplicate the call in OutputPage * Yes, I'm calling $this->msg() on places and then setting the context one more time in OutputPage::setPageTitle() or OutputPage::setHTMLTitle(), but at least I won't be confused about which objects $
2011-10-27 20:23:16 +00:00
if ( $name instanceof Message ) {
$this->mHTMLtitle = $name->setContext( $this->getContext() )->text();
} else {
$this->mHTMLtitle = $name;
}
Added OutputPage::setPageTitleMsg() and OutputPage::setHTMLTitleMsg() as modified versions of OutputPage::setPageTitle() and OutputPage::setHTMLTitle() that take an message name and its parameters; updated all calls in core that can use the two new functions
2011-10-24 17:24:38 +00:00
}
Step one in fixing OutputPage's documentation: * Groupped methods by function * Added documentation * Fixed some doxygen warnings * Some little style fixes (more to come)
2010-01-28 21:58:50 +00:00
/**
escape tags and entity in doxygen comments When inserting XML elements inline <such as this one>, doxygen chokes about it not being known. Simply enclosing the tag in double quotes prevents doxygen from emitting a warning. Also enclosed a few invalid functions calls such as \. and double quoted the HTML entities such as &foobar; Change-Id: I4019637145e683c2bec3d17b2fd98b0c50a932f1
2012-07-10 12:48:06 +00:00
* Return the "HTML title", i.e. the content of the "<title>" tag.
Step one in fixing OutputPage's documentation: * Groupped methods by function * Added documentation * Fixed some doxygen warnings * Some little style fixes (more to come)
2010-01-28 21:58:50 +00:00
*
Fixed some @params documentation Swapped some "$var type" to "type $var" or added missing types before the $var. Changed some other types to match the more common spelling. Makes beginning of some text in captial. Change-Id: I64e8cfe478cb0ba438f40b0631d6e9049cdab567
2014-04-10 18:50:10 +00:00
* @return string
Step one in fixing OutputPage's documentation: * Groupped methods by function * Added documentation * Fixed some doxygen warnings * Some little style fixes (more to come)
2010-01-28 21:58:50 +00:00
*/
public function getHTMLTitle() {
return $this->mHTMLtitle;
}
(bug 32230) Expose wgRedirectedFrom (title of the page we were redirected from, if any) in JavaScript. Patch by MrBlueSky
2011-11-30 12:43:10 +00:00
/**
Remove Title from public interface of OutputPage Bug: T278459 Change-Id: If455682540faf8d42fd8a14e0ef09bac24f72db4
2021-04-08 19:11:06 +00:00
* Set $mRedirectedFrom, the page which redirected us to the current page.
(bug 32230) Expose wgRedirectedFrom (title of the page we were redirected from, if any) in JavaScript. Patch by MrBlueSky
2011-11-30 12:43:10 +00:00
*
Remove Title from public interface of OutputPage Bug: T278459 Change-Id: If455682540faf8d42fd8a14e0ef09bac24f72db4
2021-04-08 19:11:06 +00:00
* @param PageReference $t
(bug 32230) Expose wgRedirectedFrom (title of the page we were redirected from, if any) in JavaScript. Patch by MrBlueSky
2011-11-30 12:43:10 +00:00
*/
Remove Title from public interface of OutputPage Bug: T278459 Change-Id: If455682540faf8d42fd8a14e0ef09bac24f72db4
2021-04-08 19:11:06 +00:00
public function setRedirectedFrom( PageReference $t ) {
(bug 32230) Expose wgRedirectedFrom (title of the page we were redirected from, if any) in JavaScript. Patch by MrBlueSky
2011-11-30 12:43:10 +00:00
$this->mRedirectedFrom = $t;
}
(bug 12998) Allow <sup>, <sub>, etc. in DISPLAYTITLE
2009-04-09 05:15:43 +00:00
/**
Pass php-strict for OutputPage Change-Id: Ia5d03277fff543ca672dc8efc7eaaeca4888ac23
2014-05-11 01:54:15 +00:00
* "Page title" means the contents of \<h1\>. It is stored as a valid HTML
* fragment. This function allows good tags like \<sup\> in the \<h1\> tag,
* but not bad tags like \<script\>. This function automatically sets
* \<title\> to the same content as \<h1\> but with all tags removed. Bad
* tags that were escaped in \<h1\> will still be escaped in \<title\>, and
* good tags like \<i\> will be dropped entirely.
More adding, updating and tweaking of documentation
2011-05-21 19:07:24 +00:00
*
Fixed some @params documentation Swapped some "$var type" to "type $var" or added missing types before the $var. Changed some other types to match the more common spelling. Makes beginning of some text in captial. Change-Id: I64e8cfe478cb0ba438f40b0631d6e9049cdab567
2014-04-10 18:50:10 +00:00
* @param string|Message $name
Various fixes to make phan-taint-check happier Bug: T216348 Change-Id: If4071e689f476d2138d8964598c5a02b09448677
2019-02-17 11:32:50 +00:00
* @param-taint $name tainted
* Phan-taint-check gets very confused by $name being either a string or a Message
Code cleanup, covert leading spaces into tabs per coding style
2009-11-14 21:27:13 +00:00
*/
Add some documentation to OutputPage.php and specify public/private for most of its member functions (practically all public anyway). And replace one gratuitous use of strcmp with good old ===.
2006-11-07 05:37:31 +00:00
public function setPageTitle( $name ) {
Per Nikerabbit's comment on r100621: * Removed OutputPage::setPageTitleMsg() and OutputPage::setHTMLTitleMsg() and make OutputPage::setPageTitle() and OutputPage::setHTMLTitle() accept a Message object * Updated core calls (including some that I missed last time because of non-matching case) * Added Message::setContext() and use it in RequestContext so that I don't need to duplicate the call in OutputPage * Yes, I'm calling $this->msg() on places and then setting the context one more time in OutputPage::setPageTitle() or OutputPage::setHTMLTitle(), but at least I won't be confused about which objects $
2011-10-27 20:23:16 +00:00
if ( $name instanceof Message ) {
$name = $name->setContext( $this->getContext() )->text();
}
(bug 12998) Allow <sup>, <sub>, etc. in DISPLAYTITLE
2009-04-09 05:15:43 +00:00
# change "<script>foo&bar</script>" to "&lt;script&gt;foo&amp;bar&lt;/script&gt;"
# but leave "<i>foobar</i>" alone
Add Sanitizer::removeSomeTags() which uses Remex to tokenize The existing Sanitizer::removeHTMLtags() method, in addition to having dodgy capitalization, uses regular expressions to parse the HTML. That produces corner cases like T298401 and T67747 and is not guaranteed to yield balanced or well-formed HTML. Instead, introduce and use a new Sanitizer::removeSomeTags() method which is guaranteed to always return balanced and well-formed HTML. Note that Sanitizer::removeHTMLtags()/::removeSomeTags() take a callback argument which (as far as I can tell) is never used outside core. Mark that argument as @internal, and clean up the version used by ::removeSomeTags(). Use the new ::removeSomeTags() method in the two places where DISPLAYTITLE is handled (following up on T67747). The use by the legacy parser is more difficult to replace (and would have a performace cost), so leave the old ::removeHTMLtags() method in place for that call site for now: when the legacy parser is replaced by Parsoid the need for the old ::removeHTMLtags() will go away. In a follow-up patch we'll rename ::removeHTMLtags() and mark it @internal so that we can deprecate ::removeHTMLtags() for external use. Some benchmarking code added. On my machine, with PHP 7.4, the new method tidies short 30-character title strings at a rate of about 6764/s while the tidy-based method being replaced here managed 6384/s. Sanitizer::removeHTMLtags blazes through short strings 20x faster (120,915/s); some of this difference is due to the set up cost of creating the tag whitelist and the Remex pipeline, so further optimizations could doubtless be done if Sanitizer::removeSomeTags() is more widely used. Bug: T299722 Bug: T67747 Change-Id: Ic864c01471c292f11799c4fbdac4d7d30b8bc50f
2022-01-21 22:03:26 +00:00
$nameWithTags = Sanitizer::removeSomeTags( $name );
Make OutputPage::mPagetitle private Code search revealed only one outside access, which I've submitted a patch for. Depends-On: I001deecad241b091070483d882088d2d6361ed5a Change-Id: I0f7e565f4eacc1074c619772b2d11a47a05e3584
2018-07-24 13:38:44 +00:00
$this->mPageTitle = $nameWithTags;
Removed check for empty( $action ), seems to be old register_globals code, but since $action is set in index.php with $action = $wgRequest->getVal( 'action', 'view' ); it can't match empty() in "normal" conditions
2009-02-04 19:43:30 +00:00
(bug 12998) Allow <sup>, <sub>, etc. in DISPLAYTITLE
2009-04-09 05:15:43 +00:00
# change "<i>foo&amp;bar</i>" to "foo&bar"
Use 'pagetitle' in content language It often happens that wiki sysops want to change it site wide, especially when they want to change the format (eg, have {{SITENAME}} removed, or replace the hyphen with a middot). Bug: 48701 Change-Id: Iaf00fca1e89ae022c348c4fa0de32b998d7921a1
2013-11-12 08:36:00 +00:00
$this->setHTMLTitle(
Various fixes to make phan-taint-check happier Bug: T216348 Change-Id: If4071e689f476d2138d8964598c5a02b09448677
2019-02-17 11:32:50 +00:00
$this->msg( 'pagetitle' )->plaintextParams( Sanitizer::stripAllTags( $nameWithTags ) )
Use 'pagetitle' in content language It often happens that wiki sysops want to change it site wide, especially when they want to change the format (eg, have {{SITENAME}} removed, or replace the hyphen with a middot). Bug: 48701 Change-Id: Iaf00fca1e89ae022c348c4fa0de32b998d7921a1
2013-11-12 08:36:00 +00:00
->inContentLanguage()
);
Use the set page title and HTML title in MonoBook skin, which was ignoring them. Moved the action name appending into Output to work with all skins.
2004-04-25 00:32:24 +00:00
}
here it is ... the upload-api, script-server, js2 (javascript phase2) branch merge 1st attempt. Here is a short overview of changes and associated default configuration variables (most everything is off by default) also see ~soon to be updated~: http://www.mediawiki.org/wiki/Media_Projects_Overview = Upload Improvements = ==Upload API == * Based on the early work of Bryan Tong and others it adds the upload option to the api. * We rewrite Special:Upload page to include use the new refactoring * Added in token checks in both the SpecialUpload.php page so avoids DOS / xss copy-by-url JavaScript based cross site POST file submissions == Copy by URL== $wgAllowCopyUploads = false; * http class rewrite includes a new http background download see: includes/HttpFunctions.php * spins off a php process that calls: maintenance/http_session_download.php * pushes updates to the session and gives the user a progress bar on http copy uploads from other server progress (using js2 upload interface) (if not using the js2 upload interface it does the request in-place but the download is limited to the php ini timeout time) == Firefogg == * Firefogg enables resumable upload by chunks * progress indicators and conditional invokation (js2 system) * and of-course client side transcoding. = Script Server = $wgEnableScriptLoader = false; * off by default if $wgEnableScriptLoader is turned on script files are grouped, gziped, cached etc. for more info see: http://www.mediawiki.org/wiki/Extension:ScriptLoader * Includes some early skin js include fixes (skin/script system still lots of love) * Includes a "javascript class autoloader" this is packaged into mwEmbed so that the mwEmbed library can work in stand alone mode (while retaining localization and script serving) (one such application is the make page for firefogg.org : http://www.firefogg.org/make/index.html ) * The file that contains the autojavascript loading classes is: js2/php/jsAutoloadLocalClasses.php * One can use this auto class loading dependency system with extensions and add-ons but I need to better document that. = js2 system / mwEmbed= $wgEnableJS2system = false * includes initial rewrite towards more jquery based javascript code * especially for the Special:Upload page. * Also the edit page include support for the "add-media-wizard" * includes dependency loader for javascript that optionally takes advantage of the script-loader * remote embedding of javascript interfaces (like embedding video, or commons media searching) * $wgDebugJavaScript = false; .. .this variable lets you always get "always fresh javascript". When used with the script-loader it does not minify the script-loader output. = mwEmbed = * Will commit a separate patch to oggHandler that conditionally outputs <video tag> to use the new javascript video player. ** mv_embed player includes: play-head, volume control, remote embedding, oggz-chop support across plugins. * add-media-wizard adds easy inserts of media to pages (with import) == jQuery== * we include a base install of jQuery, jQuery ui and some plugins. * all the javascript classes are in the scriptloader so its easy to load any set of jquery ui components that you may need using the script-server. You get a callback so you can then execute js with dependencies loaded. == other stuff == there is a bit more code in js2 that pertains to sequence editing, timed text display and basic image editing. We include a base import of pixastic-lib & pixastic-editor... will work with the pixastic developer to try and ensure upstream compatibility on our usage of the library for in-browser photo and sequence manipulation.
2009-07-14 23:52:14 +00:00
Step one in fixing OutputPage's documentation: * Groupped methods by function * Added documentation * Fixed some doxygen warnings * Some little style fixes (more to come)
2010-01-28 21:58:50 +00:00
/**
Follow-up r61806: also escape tag in getPageTitle()
2010-02-02 16:26:52 +00:00
* Return the "page title", i.e. the content of the \<h1\> tag.
Step one in fixing OutputPage's documentation: * Groupped methods by function * Added documentation * Fixed some doxygen warnings * Some little style fixes (more to come)
2010-01-28 21:58:50 +00:00
*
Fixed some @params documentation Swapped some "$var type" to "type $var" or added missing types before the $var. Changed some other types to match the more common spelling. Makes beginning of some text in captial. Change-Id: I64e8cfe478cb0ba438f40b0631d6e9049cdab567
2014-04-10 18:50:10 +00:00
* @return string
Step one in fixing OutputPage's documentation: * Groupped methods by function * Added documentation * Fixed some doxygen warnings * Some little style fixes (more to come)
2010-01-28 21:58:50 +00:00
*/
public function getPageTitle() {
Make OutputPage::mPagetitle private Code search revealed only one outside access, which I've submitted a patch for. Depends-On: I001deecad241b091070483d882088d2d6361ed5a Change-Id: I0f7e565f4eacc1074c619772b2d11a47a05e3584
2018-07-24 13:38:44 +00:00
return $this->mPageTitle;
Step one in fixing OutputPage's documentation: * Groupped methods by function * Added documentation * Fixed some doxygen warnings * Some little style fixes (more to come)
2010-01-28 21:58:50 +00:00
}
Improve page display title handling for category pages One use case of display title is to localise page names with Translate extension or without. While the page title changes, the subheadings still say something like "Pages in category Foo/de". Also converted one raw HTML message to be a parsed message. First version of this patch caused an issue when previewing because page title in h1 is not the same as page display title. This issue is fixed by promoting page display title as it's own member in OutputPage. Also added getUnprefixedDisplayTitle that attempts to strip away the namespace prefix to mimic Title::getText() but which works with display title instead. Bug: T43720 Bug: T46197 Change-Id: I6097a873297eb57759252fc56ad6d02c44e4c366
2016-11-09 06:24:57 +00:00
/**
* Same as page title but only contains name of the page, not any other text.
*
* @since 1.32
* @param string $html Page title text.
* @see OutputPage::setPageTitle
*/
public function setDisplayTitle( $html ) {
$this->displayTitle = $html;
}
/**
* Returns page display title.
*
* Performs some normalization, but this not as strict the magic word.
*
* @since 1.32
* @return string HTML
*/
public function getDisplayTitle() {
$html = $this->displayTitle;
if ( $html === null ) {
Encode & to &amp; in displaytitle fallback The value in the attribute displaytitle must contain valid HTML. The sanitizer of the {{DISPLAYTITLE}} parser ensures that only valid HTML is accepted. If there is no {{DISPLAYTITLE}} in the wikitext then displaytitle falls back to $title->getPrefixedText(). Here an HTML encoding of special characters is necessary. This affects only the replacement of & by &amp; because other special characters like < and > are not allowed in the title. This change affects the displaytitle fallback on the following places: * ApiParse * ApiQueryInfo * InfoAction * Parser The displaytitle fallback in OutputPage is also updated to this behavior although Sanitizer::normalizeCharReferences( Sanitizer::removeHTMLtags( $html ) also replaces & by &amp;. Also add test cases with & in the displaytitle to: * ApiParseTest * ApiQueryInfoTest * parserTests Bug: T291985 Change-Id: I8ee1e2731d9bfa49725d663b34986e7e3073e4ca
2021-09-28 18:47:20 +00:00
return htmlspecialchars( $this->getTitle()->getPrefixedText(), ENT_NOQUOTES );
Improve page display title handling for category pages One use case of display title is to localise page names with Translate extension or without. While the page title changes, the subheadings still say something like "Pages in category Foo/de". Also converted one raw HTML message to be a parsed message. First version of this patch caused an issue when previewing because page title in h1 is not the same as page display title. This issue is fixed by promoting page display title as it's own member in OutputPage. Also added getUnprefixedDisplayTitle that attempts to strip away the namespace prefix to mimic Title::getText() but which works with display title instead. Bug: T43720 Bug: T46197 Change-Id: I6097a873297eb57759252fc56ad6d02c44e4c366
2016-11-09 06:24:57 +00:00
}
Add Sanitizer::removeSomeTags() which uses Remex to tokenize The existing Sanitizer::removeHTMLtags() method, in addition to having dodgy capitalization, uses regular expressions to parse the HTML. That produces corner cases like T298401 and T67747 and is not guaranteed to yield balanced or well-formed HTML. Instead, introduce and use a new Sanitizer::removeSomeTags() method which is guaranteed to always return balanced and well-formed HTML. Note that Sanitizer::removeHTMLtags()/::removeSomeTags() take a callback argument which (as far as I can tell) is never used outside core. Mark that argument as @internal, and clean up the version used by ::removeSomeTags(). Use the new ::removeSomeTags() method in the two places where DISPLAYTITLE is handled (following up on T67747). The use by the legacy parser is more difficult to replace (and would have a performace cost), so leave the old ::removeHTMLtags() method in place for that call site for now: when the legacy parser is replaced by Parsoid the need for the old ::removeHTMLtags() will go away. In a follow-up patch we'll rename ::removeHTMLtags() and mark it @internal so that we can deprecate ::removeHTMLtags() for external use. Some benchmarking code added. On my machine, with PHP 7.4, the new method tidies short 30-character title strings at a rate of about 6764/s while the tidy-based method being replaced here managed 6384/s. Sanitizer::removeHTMLtags blazes through short strings 20x faster (120,915/s); some of this difference is due to the set up cost of creating the tag whitelist and the Remex pipeline, so further optimizations could doubtless be done if Sanitizer::removeSomeTags() is more widely used. Bug: T299722 Bug: T67747 Change-Id: Ic864c01471c292f11799c4fbdac4d7d30b8bc50f
2022-01-21 22:03:26 +00:00
return Sanitizer::removeSomeTags( $html );
Improve page display title handling for category pages One use case of display title is to localise page names with Translate extension or without. While the page title changes, the subheadings still say something like "Pages in category Foo/de". Also converted one raw HTML message to be a parsed message. First version of this patch caused an issue when previewing because page title in h1 is not the same as page display title. This issue is fixed by promoting page display title as it's own member in OutputPage. Also added getUnprefixedDisplayTitle that attempts to strip away the namespace prefix to mimic Title::getText() but which works with display title instead. Bug: T43720 Bug: T46197 Change-Id: I6097a873297eb57759252fc56ad6d02c44e4c366
2016-11-09 06:24:57 +00:00
}
/**
* Returns page display title without namespace prefix if possible.
*
Add markup to page titles to distinguish the namespace and the main text Pages outside of the main namespace now have the following markup in their <h1> page titles, using 'Talk:Hello' as an example: <h1> <span class="mw-page-title-namespace">Talk</span> <span class="mw-page-title-separator">:</span> <span class="mw-page-title-main">Hello</span> </h1> (line breaks and spaces added for readability) Pages in the main namespace only have the last part, e.g. for 'Hello': <h1> <span class="mw-page-title-main">Hello</span> </h1> The change is motivated by a desire to style the titles differently on talk pages in the DiscussionTools extension (T313636), but it could also be used for other things: * Language-specific tweaks (e.g. adding typographically-correct spaces around the colon separator: T249149, or replacing it with a different character: T36295) * Site-specific tweaks (e.g. de-emphasize or emphasize specific namespaces like 'Draft': T62973 / T236215) The markup is also added to automatically language-converted titles. It is not added when the title is overridden using the wikitext `{{DISPLAYTITLE:…}}` or `-{T|…}-` forms. I think this is a small limitation, as those forms mostly used in the main namespace, where the extra markup isn't very helpful anyway. This may be improved in the future. As a workaround, users could also just add the same HTML markup to their wikitext (as those forms accept it). It is not also added when the title is overridden by an extension like Translate. Maybe we'll have a better API before anyone wants to do that. If not, one could un-mark Parser::formatPageTitle() as @internal, and use that method to add the markup themselves. Bug: T306440 Change-Id: I62b17ef22de3606d736e6c261e542a34b58b5a05
2022-08-09 02:52:53 +00:00
* This method is unreliable and best avoided. (T314399)
*
Improve page display title handling for category pages One use case of display title is to localise page names with Translate extension or without. While the page title changes, the subheadings still say something like "Pages in category Foo/de". Also converted one raw HTML message to be a parsed message. First version of this patch caused an issue when previewing because page title in h1 is not the same as page display title. This issue is fixed by promoting page display title as it's own member in OutputPage. Also added getUnprefixedDisplayTitle that attempts to strip away the namespace prefix to mimic Title::getText() but which works with display title instead. Bug: T43720 Bug: T46197 Change-Id: I6097a873297eb57759252fc56ad6d02c44e4c366
2016-11-09 06:24:57 +00:00
* @since 1.32
* @return string HTML
*/
public function getUnprefixedDisplayTitle() {
OutputPage: Handle language variants in getUnprefixedDisplayTitle When the namespace prefix is converted to a specific language variant, the function getUnprefixedDisplayTitle() takes no effect. This issue is visible on category pages, where the unprefixed display title is used for subheadings. Bug: T302161 Change-Id: I94684ca0760aafb7c41470210652abfffb7e0d51
2022-02-20 23:24:49 +00:00
$service = MediaWikiServices::getInstance();
$languageConverter = $service->getLanguageConverterFactory()
->getLanguageConverter( $service->getContentLanguage() );
Improve page display title handling for category pages One use case of display title is to localise page names with Translate extension or without. While the page title changes, the subheadings still say something like "Pages in category Foo/de". Also converted one raw HTML message to be a parsed message. First version of this patch caused an issue when previewing because page title in h1 is not the same as page display title. This issue is fixed by promoting page display title as it's own member in OutputPage. Also added getUnprefixedDisplayTitle that attempts to strip away the namespace prefix to mimic Title::getText() but which works with display title instead. Bug: T43720 Bug: T46197 Change-Id: I6097a873297eb57759252fc56ad6d02c44e4c366
2016-11-09 06:24:57 +00:00
$text = $this->getDisplayTitle();
Add markup to page titles to distinguish the namespace and the main text Pages outside of the main namespace now have the following markup in their <h1> page titles, using 'Talk:Hello' as an example: <h1> <span class="mw-page-title-namespace">Talk</span> <span class="mw-page-title-separator">:</span> <span class="mw-page-title-main">Hello</span> </h1> (line breaks and spaces added for readability) Pages in the main namespace only have the last part, e.g. for 'Hello': <h1> <span class="mw-page-title-main">Hello</span> </h1> The change is motivated by a desire to style the titles differently on talk pages in the DiscussionTools extension (T313636), but it could also be used for other things: * Language-specific tweaks (e.g. adding typographically-correct spaces around the colon separator: T249149, or replacing it with a different character: T36295) * Site-specific tweaks (e.g. de-emphasize or emphasize specific namespaces like 'Draft': T62973 / T236215) The markup is also added to automatically language-converted titles. It is not added when the title is overridden using the wikitext `{{DISPLAYTITLE:…}}` or `-{T|…}-` forms. I think this is a small limitation, as those forms mostly used in the main namespace, where the extra markup isn't very helpful anyway. This may be improved in the future. As a workaround, users could also just add the same HTML markup to their wikitext (as those forms accept it). It is not also added when the title is overridden by an extension like Translate. Maybe we'll have a better API before anyone wants to do that. If not, one could un-mark Parser::formatPageTitle() as @internal, and use that method to add the markup themselves. Bug: T306440 Change-Id: I62b17ef22de3606d736e6c261e542a34b58b5a05
2022-08-09 02:52:53 +00:00
// Create a regexp with matching groups as placeholders for the namespace, separator and main text
$pageTitleRegexp = "/^" . str_replace(
preg_quote( '(.+?)', '/' ),
'(.+?)',
preg_quote( Parser::formatPageTitle( '(.+?)', '(.+?)', '(.+?)' ), '/' )
) . "$/";
$matches = [];
if ( preg_match( $pageTitleRegexp, $text, $matches ) ) {
// The regexp above could be manipulated by malicious user input,
// sanitize the result just in case
return Sanitizer::removeSomeTags( $matches[3] );
}
OutputPage: Handle language variants in getUnprefixedDisplayTitle When the namespace prefix is converted to a specific language variant, the function getUnprefixedDisplayTitle() takes no effect. This issue is visible on category pages, where the unprefixed display title is used for subheadings. Bug: T302161 Change-Id: I94684ca0760aafb7c41470210652abfffb7e0d51
2022-02-20 23:24:49 +00:00
$nsPrefix = $languageConverter->convertNamespace(
$this->getTitle()->getNamespace()
) . ':';
Improve page display title handling for category pages One use case of display title is to localise page names with Translate extension or without. While the page title changes, the subheadings still say something like "Pages in category Foo/de". Also converted one raw HTML message to be a parsed message. First version of this patch caused an issue when previewing because page title in h1 is not the same as page display title. This issue is fixed by promoting page display title as it's own member in OutputPage. Also added getUnprefixedDisplayTitle that attempts to strip away the namespace prefix to mimic Title::getText() but which works with display title instead. Bug: T43720 Bug: T46197 Change-Id: I6097a873297eb57759252fc56ad6d02c44e4c366
2016-11-09 06:24:57 +00:00
$prefix = preg_quote( $nsPrefix, '/' );
return preg_replace( "/^$prefix/i", '', $text );
}
Start better utalizing OutputPage as the focal point for things related to the output of the page. Like getTitle use $out->get{User,Skin} instead of $wgUser and $wgUser->getSkin() for things where the user or skin is being used for output generation back to $out instead of using globals with bad structure.
2011-04-03 03:59:47 +00:00
/**
Implement the RequestContext class. Some credit to IAlex, ;) other credit for me and that plethora of bugs and hicoughs I had to deal with in impelenting it. http://www.mediawiki.org/wiki/Requests_for_comment/Context_object (it's little different though)
2011-04-03 10:41:14 +00:00
* Set the Title object to use
Start better utalizing OutputPage as the focal point for things related to the output of the page. Like getTitle use $out->get{User,Skin} instead of $wgUser and $wgUser->getSkin() for things where the user or skin is being used for output generation back to $out instead of using globals with bad structure.
2011-04-03 03:59:47 +00:00
*
Remove Title from public interface of OutputPage Bug: T278459 Change-Id: If455682540faf8d42fd8a14e0ef09bac24f72db4
2021-04-08 19:11:06 +00:00
* @param PageReference $t
Start better utalizing OutputPage as the focal point for things related to the output of the page. Like getTitle use $out->get{User,Skin} instead of $wgUser and $wgUser->getSkin() for things where the user or skin is being used for output generation back to $out instead of using globals with bad structure.
2011-04-03 03:59:47 +00:00
*/
Remove Title from public interface of OutputPage Bug: T278459 Change-Id: If455682540faf8d42fd8a14e0ef09bac24f72db4
2021-04-08 19:11:06 +00:00
public function setTitle( PageReference $t ) {
$t = Title::castFromPageReference( $t );
Unsuppress another phan issue (part 7) Bug: T231636 Depends-On: I2cd24e73726394e3200a570c45d5e86b6849bfa9 Depends-On: I4fa3e6aad872434ca397325ed7a83f94973661d0 Change-Id: Ie6233561de78457cae5e4e44e220feec2d1272d8
2019-08-31 16:14:38 +00:00
// @phan-suppress-next-next-line PhanUndeclaredMethod
// @fixme Not all implementations of IContextSource have this method!
* (bug 28532) wfMsgExt() and wfMsgWikiHtml() use $wgOut->parse() * (bug 16129) Transcluded special pages expose strip markers when they output parsed messages Also adding some related documentation during my travels around the code
2011-04-18 12:43:53 +00:00
$this->getContext()->setTitle( $t );
Implement the RequestContext class. Some credit to IAlex, ;) other credit for me and that plethora of bugs and hicoughs I had to deal with in impelenting it. http://www.mediawiki.org/wiki/Requests_for_comment/Context_object (it's little different though)
2011-04-03 10:41:14 +00:00
}
Step one in fixing OutputPage's documentation: * Groupped methods by function * Added documentation * Fixed some doxygen warnings * Some little style fixes (more to come)
2010-01-28 21:58:50 +00:00
/**
Spellchecked /includes directory * Ran spell-checker over code comments in /includes/ * A few spellchecking fixes for wfDebug() calls Found one very strange (NOOP?) line in Linker.php - see "TODO: BUG?" Change-Id: Ibb86b51073b980eda9ecce2cf0b8dd33f058adbf
2013-03-13 07:42:41 +00:00
* Replace the subtitle with $str
Step one in fixing OutputPage's documentation: * Groupped methods by function * Added documentation * Fixed some doxygen warnings * Some little style fixes (more to come)
2010-01-28 21:58:50 +00:00
*
Cleanup some docs (includes/*.php) - Swap "$variable type" to "type $variable" - Added missing types - Fixed spacing inside docs - Makes beginning of @param/@return/@var/@throws in capital - Changed some types to match the more common spelling Change-Id: I783e4dbfe5f6f98b32b9a03ccf6439e13e132bcc
2014-07-24 17:42:24 +00:00
* @param string|Message $str New value of the subtitle. String should be safe HTML.
Step one in fixing OutputPage's documentation: * Groupped methods by function * Added documentation * Fixed some doxygen warnings * Some little style fixes (more to come)
2010-01-28 21:58:50 +00:00
*/
public function setSubtitle( $str ) {
Fix for r102417: commiting OutputPage.php is actually a good idea
2011-11-08 18:02:26 +00:00
$this->clearSubtitle();
$this->addSubtitle( $str );
Step one in fixing OutputPage's documentation: * Groupped methods by function * Added documentation * Fixed some doxygen warnings * Some little style fixes (more to come)
2010-01-28 21:58:50 +00:00
}
Fix for r102417: commiting OutputPage.php is actually a good idea
2011-11-08 18:02:26 +00:00
/**
* Add $str to the subtitle
*
Cleanup some docs (includes/*.php) - Swap "$variable type" to "type $variable" - Added missing types - Fixed spacing inside docs - Makes beginning of @param/@return/@var/@throws in capital - Changed some types to match the more common spelling Change-Id: I783e4dbfe5f6f98b32b9a03ccf6439e13e132bcc
2014-07-24 17:42:24 +00:00
* @param string|Message $str String or Message to add to the subtitle. String should be safe HTML.
Fix for r102417: commiting OutputPage.php is actually a good idea
2011-11-08 18:02:26 +00:00
*/
public function addSubtitle( $str ) {
if ( $str instanceof Message ) {
$this->mSubtitle[] = $str->setContext( $this->getContext() )->parse();
} else {
$this->mSubtitle[] = $str;
}
}
/**
Fix backlink for RevertAction RevertAction::getDescription cannot set subtitle on OutputPage, because the subtitle on OutputPage gets cleared before the result of getDescription is added and than the subtitle is gone. Refactored the code for building the backlink into a static function and use it. Change-Id: Iedad0b8e040035a9a10a0b140d2322357e6b539a
2014-08-03 20:04:45 +00:00
* Build message object for a subtitle containing a backlink to a page
Fix for r102417: commiting OutputPage.php is actually a good idea
2011-11-08 18:02:26 +00:00
*
actions: Add missing ->text() in RevertAction::getPageTitle() It currently becomes a string implicitly much later on based on the default format. All other implementations return a string and for messages that typically is formatted as text(). This should be done explicitly as it is the responsibility of RevertAction to know this message and how it is meant to be formatted to get the string that the "getPageTitle" contract calls for. This bug has been here since 2012 with change I2c6fad30e2c4eaebf122, where the code changed from wfMsg (string, unlike wfMessage) to `$this->msg` without calling a format. Bug: T252946 Change-Id: Ibfa335952eb752515a31be6feb6749534bb2a7c4
2022-04-12 21:20:38 +00:00
* @since 1.25
Remove Title from public interface of OutputPage Bug: T278459 Change-Id: If455682540faf8d42fd8a14e0ef09bac24f72db4
2021-04-08 19:11:06 +00:00
* @param PageReference $page Title to link to
Add $query parameter to OutputPage::addBacklinkSubtitle Change-Id: If6145b5512657bbd9e165b0c614e5061615c294d
2014-08-07 08:22:57 +00:00
* @param array $query Array of additional parameters to include in the link
Fix backlink for RevertAction RevertAction::getDescription cannot set subtitle on OutputPage, because the subtitle on OutputPage gets cleared before the result of getDescription is added and than the subtitle is gone. Refactored the code for building the backlink into a static function and use it. Change-Id: Iedad0b8e040035a9a10a0b140d2322357e6b539a
2014-08-03 20:04:45 +00:00
* @return Message
Fix for r102417: commiting OutputPage.php is actually a good idea
2011-11-08 18:02:26 +00:00
*/
Remove Title from public interface of OutputPage Bug: T278459 Change-Id: If455682540faf8d42fd8a14e0ef09bac24f72db4
2021-04-08 19:11:06 +00:00
public static function buildBacklinkSubtitle( PageReference $page, $query = [] ) {
if ( $page instanceof PageRecord || $page instanceof Title ) {
// Callers will typically have a PageRecord
if ( $page->isRedirect() ) {
$query['redirect'] = 'no';
}
} elseif ( $page->getNamespace() !== NS_SPECIAL ) {
// We don't know whether it's a redirect, so add the parameter, just to be sure.
Fix for r102417: commiting OutputPage.php is actually a good idea
2011-11-08 18:02:26 +00:00
$query['redirect'] = 'no';
}
Remove Title from public interface of OutputPage Bug: T278459 Change-Id: If455682540faf8d42fd8a14e0ef09bac24f72db4
2021-04-08 19:11:06 +00:00
Replace Linker::link() with LinkRenderer in includes directory * CategoryViewer * OutputPage * Preferences Bug: T149346 Change-Id: I9f9bcd9b461884817e8ceefbc6757c436221e331
2016-11-30 21:28:55 +00:00
$linkRenderer = MediaWikiServices::getInstance()->getLinkRenderer();
Fix backlink for RevertAction RevertAction::getDescription cannot set subtitle on OutputPage, because the subtitle on OutputPage gets cleared before the result of getDescription is added and than the subtitle is gone. Refactored the code for building the backlink into a static function and use it. Change-Id: Iedad0b8e040035a9a10a0b140d2322357e6b539a
2014-08-03 20:04:45 +00:00
return wfMessage( 'backlinksubtitle' )
Remove unneeded calls of TitleValue::castPageToLinkTarget - ManualLogEntry::setTarget accepts PageReference - LinkRenderer::makeLink accepts PageReference - PermissionManager::isBlockedFrom accepts PageReference Change-Id: I41901e10c3df22145e92d12de74f0382976bb598
2022-03-10 19:32:03 +00:00
->rawParams( $linkRenderer->makeLink( $page, null, [], $query ) );
Fix backlink for RevertAction RevertAction::getDescription cannot set subtitle on OutputPage, because the subtitle on OutputPage gets cleared before the result of getDescription is added and than the subtitle is gone. Refactored the code for building the backlink into a static function and use it. Change-Id: Iedad0b8e040035a9a10a0b140d2322357e6b539a
2014-08-03 20:04:45 +00:00
}
/**
* Add a subtitle containing a backlink to a page
*
Remove Title from public interface of OutputPage Bug: T278459 Change-Id: If455682540faf8d42fd8a14e0ef09bac24f72db4
2021-04-08 19:11:06 +00:00
* @param PageReference $title Title to link to
Fix backlink for RevertAction RevertAction::getDescription cannot set subtitle on OutputPage, because the subtitle on OutputPage gets cleared before the result of getDescription is added and than the subtitle is gone. Refactored the code for building the backlink into a static function and use it. Change-Id: Iedad0b8e040035a9a10a0b140d2322357e6b539a
2014-08-03 20:04:45 +00:00
* @param array $query Array of additional parameters to include in the link
*/
Remove Title from public interface of OutputPage Bug: T278459 Change-Id: If455682540faf8d42fd8a14e0ef09bac24f72db4
2021-04-08 19:11:06 +00:00
public function addBacklinkSubtitle( PageReference $title, $query = [] ) {
Fix backlink for RevertAction RevertAction::getDescription cannot set subtitle on OutputPage, because the subtitle on OutputPage gets cleared before the result of getDescription is added and than the subtitle is gone. Refactored the code for building the backlink into a static function and use it. Change-Id: Iedad0b8e040035a9a10a0b140d2322357e6b539a
2014-08-03 20:04:45 +00:00
$this->addSubtitle( self::buildBacklinkSubtitle( $title, $query ) );
Fix for r102417: commiting OutputPage.php is actually a good idea
2011-11-08 18:02:26 +00:00
}
/**
* Clear the subtitles
*/
public function clearSubtitle() {
Convert all array() syntax to [] Per wikitech-l consensus: https://lists.wikimedia.org/pipermail/wikitech-l/2016-February/084821.html Notes: * Disabled CallTimePassByReference due to false positives (T127163) Change-Id: I2c8ce713ce6600a0bb7bf67537c87044c7a45c4b
2016-02-17 09:09:32 +00:00
$this->mSubtitle = [];
Step one in fixing OutputPage's documentation: * Groupped methods by function * Added documentation * Fixed some doxygen warnings * Some little style fixes (more to come)
2010-01-28 21:58:50 +00:00
}
/**
Fixed some @params documentation Swapped some "$var type" to "type $var" or added missing types before the $var. Changed some other types to match the more common spelling. Makes beginning of some text in captial. Change-Id: I64e8cfe478cb0ba438f40b0631d6e9049cdab567
2014-04-10 18:50:10 +00:00
* @return string
Step one in fixing OutputPage's documentation: * Groupped methods by function * Added documentation * Fixed some doxygen warnings * Some little style fixes (more to come)
2010-01-28 21:58:50 +00:00
*/
public function getSubtitle() {
Fix for r102417: commiting OutputPage.php is actually a good idea
2011-11-08 18:02:26 +00:00
return implode( "<br />\n\t\t\t\t", $this->mSubtitle );
Step one in fixing OutputPage's documentation: * Groupped methods by function * Added documentation * Fixed some doxygen warnings * Some little style fixes (more to come)
2010-01-28 21:58:50 +00:00
}
/**
Fix some stuttering in comments and documentation Change-Id: I9c0088b9aab37335203cad45a1d6fa8ac3f43321
2014-12-16 00:41:45 +00:00
* Set the page as printable, i.e. it'll be displayed with all
Step one in fixing OutputPage's documentation: * Groupped methods by function * Added documentation * Fixed some doxygen warnings * Some little style fixes (more to come)
2010-01-28 21:58:50 +00:00
* print styles included
*/
public function setPrintable() {
$this->mPrintable = true;
}
/**
* Return whether the page is "printable"
*
Fixed some @params documentation Swapped some "$var type" to "type $var" or added missing types before the $var. Changed some other types to match the more common spelling. Makes beginning of some text in captial. Change-Id: I64e8cfe478cb0ba438f40b0631d6e9049cdab567
2014-04-10 18:50:10 +00:00
* @return bool
Step one in fixing OutputPage's documentation: * Groupped methods by function * Added documentation * Fixed some doxygen warnings * Some little style fixes (more to come)
2010-01-28 21:58:50 +00:00
*/
public function isPrintable() {
return $this->mPrintable;
}
/**
* Disable output completely, i.e. calling output() will have no effect
*/
public function disable() {
$this->mDoNothing = true;
}
/**
* Return whether the output will be completely disabled
*
Fixed some @params documentation Swapped some "$var type" to "type $var" or added missing types before the $var. Changed some other types to match the more common spelling. Makes beginning of some text in captial. Change-Id: I64e8cfe478cb0ba438f40b0631d6e9049cdab567
2014-04-10 18:50:10 +00:00
* @return bool
Step one in fixing OutputPage's documentation: * Groupped methods by function * Added documentation * Fixed some doxygen warnings * Some little style fixes (more to come)
2010-01-28 21:58:50 +00:00
*/
public function isDisabled() {
return $this->mDoNothing;
}
Step two in fixing OutputPage's documentation, more or less the same as r61647
2010-01-29 21:32:45 +00:00
/**
* Show an "add new section" link?
*
Fixed some @params documentation Swapped some "$var type" to "type $var" or added missing types before the $var. Changed some other types to match the more common spelling. Makes beginning of some text in captial. Change-Id: I64e8cfe478cb0ba438f40b0631d6e9049cdab567
2014-04-10 18:50:10 +00:00
* @return bool
Step two in fixing OutputPage's documentation, more or less the same as r61647
2010-01-29 21:32:45 +00:00
*/
public function showNewSectionLink() {
return $this->mNewSectionLink;
}
/**
* Forcibly hide the new section link?
*
Fixed some @params documentation Swapped some "$var type" to "type $var" or added missing types before the $var. Changed some other types to match the more common spelling. Makes beginning of some text in captial. Change-Id: I64e8cfe478cb0ba438f40b0631d6e9049cdab567
2014-04-10 18:50:10 +00:00
* @return bool
Step two in fixing OutputPage's documentation, more or less the same as r61647
2010-01-29 21:32:45 +00:00
*/
public function forceHideNewSectionLink() {
return $this->mHideNewSectionLink;
}
Follow-up r61021: added some documentation per Tim's request
2010-01-26 20:47:34 +00:00
/**
* Add or remove feed links in the page header
* This is mainly kept for backward compatibility, see OutputPage::addFeedLink()
* for the new version
* @see addFeedLink()
*
Cleanup some docs (includes/*.php) - Swap "$variable type" to "type $variable" - Added missing types - Fixed spacing inside docs - Makes beginning of @param/@return/@var/@throws in capital - Changed some types to match the more common spelling Change-Id: I783e4dbfe5f6f98b32b9a03ccf6439e13e132bcc
2014-07-24 17:42:24 +00:00
* @param bool $show True: add default feeds, false: remove all feeds
Follow-up r61021: added some documentation per Tim's request
2010-01-26 20:47:34 +00:00
*/
* (bug 21738) Fix for r57119: OutputPage::setSyndicated() now works again (was breaking Special:RecentChanges' feeds, ...)
2010-01-13 20:23:14 +00:00
public function setSyndicated( $show = true ) {
if ( $show ) {
$this->setFeedAppendQuery( false );
} else {
Convert all array() syntax to [] Per wikitech-l consensus: https://lists.wikimedia.org/pipermail/wikitech-l/2016-February/084821.html Notes: * Disabled CallTimePassByReference due to false positives (T127163) Change-Id: I2c8ce713ce6600a0bb7bf67537c87044c7a45c4b
2016-02-17 09:09:32 +00:00
$this->mFeedLinks = [];
* (bug 21738) Fix for r57119: OutputPage::setSyndicated() now works again (was breaking Special:RecentChanges' feeds, ...)
2010-01-13 20:23:14 +00:00
}
}
* improved script loader language support ** supports url language code parameter * improved performance on cached hit ( apache bench mark goes from old: Rps: 23.40 [#/sec] new: Rps: 378.61 [#/sec] ** we don't check existence of wiki title until after cache file check. ** we don't load all includes/WebStart.php before checking cache * some remote search bug ** improved speed and accuracy of language swapping ( token check no breakage with }) in msg text limitation * improved core js2 performance in parsing class defines from mv_embed.js ** only reads 8k of head at a time. * reduced expire time to one day (instead of one year on script requests) * removed $wgEnableScriptMinify (just use debug or ?debug instead)
2009-10-24 21:32:53 +00:00
Advertise feeds only if $wgFeed is enabled Wikis with $wgFeed=false should not present an Atom icon on the RecentChanges/Watchlist. Bug: T116145 Change-Id: Ic64f7f5115a77c42ed2b336f6996fe711b3822ba
2018-01-28 02:21:51 +00:00
/**
* Return effective list of advertised feed types
* @see addFeedLink()
*
Update a lot of unspecific "array" types in PHPDocs This includes fixing some mistakes, as well as removing redundant text that doesn't add new information, either because it literally repeats what the code already says, or is actually duplicated. Change-Id: I3a8dd8ce57192deda8916cc444c87d7ab1a36515
2020-10-28 10:01:33 +00:00
* @return string[] Array of feed type names ( 'rss', 'atom' )
Advertise feeds only if $wgFeed is enabled Wikis with $wgFeed=false should not present an Atom icon on the RecentChanges/Watchlist. Bug: T116145 Change-Id: Ic64f7f5115a77c42ed2b336f6996fe711b3822ba
2018-01-28 02:21:51 +00:00
*/
protected function getAdvertisedFeedTypes() {
Use name constants to access config settings. Use name constants instead of string literals in calls to Config::get and ServiceOptions::get, when referring to core configuration variables. This protects against typos and makes the decumentation and schema declaration of the config settings discoverable. This is the first batch, only touching files directly under /includes/ Change-Id: I7252e636c7c86d950d9257b33491af492c6dd5eb
2022-04-01 15:58:32 +00:00
if ( $this->getConfig()->get( MainConfigNames::Feed ) ) {
return $this->getConfig()->get( MainConfigNames::AdvertisedFeedTypes );
Advertise feeds only if $wgFeed is enabled Wikis with $wgFeed=false should not present an Atom icon on the RecentChanges/Watchlist. Bug: T116145 Change-Id: Ic64f7f5115a77c42ed2b336f6996fe711b3822ba
2018-01-28 02:21:51 +00:00
} else {
return [];
}
}
Follow-up r61021: added some documentation per Tim's request
2010-01-26 20:47:34 +00:00
/**
* Add default feeds to the page header
* This is mainly kept for backward compatibility, see OutputPage::addFeedLink()
* for the new version
* @see addFeedLink()
*
Fix various documentation related to false Some function already document "False on failure", but does not document it on the used type Found by phan strict checks Change-Id: I12eb8bbc99179833ee3e42c1a7d1dc1443682ca6
2022-03-03 20:33:11 +00:00
* @param string|false $val Query to append to feed links or false to output
Follow-up r61021: added some documentation per Tim's request
2010-01-26 20:47:34 +00:00
* default links
*/
Re-work feed exposure system. * Includes a new, more flexible system which allows you to provide an array of format => URL, but also supports the older system of providing a URL suffix. * Implemented it for the watchlist if a user has set their own token.
2009-09-30 17:35:41 +00:00
public function setFeedAppendQuery( $val ) {
Convert all array() syntax to [] Per wikitech-l consensus: https://lists.wikimedia.org/pipermail/wikitech-l/2016-February/084821.html Notes: * Disabled CallTimePassByReference due to false positives (T127163) Change-Id: I2c8ce713ce6600a0bb7bf67537c87044c7a45c4b
2016-02-17 09:09:32 +00:00
$this->mFeedLinks = [];
* improved script loader language support ** supports url language code parameter * improved performance on cached hit ( apache bench mark goes from old: Rps: 23.40 [#/sec] new: Rps: 378.61 [#/sec] ** we don't check existence of wiki title until after cache file check. ** we don't load all includes/WebStart.php before checking cache * some remote search bug ** improved speed and accuracy of language swapping ( token check no breakage with }) in msg text limitation * improved core js2 performance in parsing class defines from mv_embed.js ** only reads 8k of head at a time. * reduced expire time to one day (instead of one year on script requests) * removed $wgEnableScriptMinify (just use debug or ?debug instead)
2009-10-24 21:32:53 +00:00
Advertise feeds only if $wgFeed is enabled Wikis with $wgFeed=false should not present an Atom icon on the RecentChanges/Watchlist. Bug: T116145 Change-Id: Ic64f7f5115a77c42ed2b336f6996fe711b3822ba
2018-01-28 02:21:51 +00:00
foreach ( $this->getAdvertisedFeedTypes() as $type ) {
* (bug 21738) Fix for r57119: OutputPage::setSyndicated() now works again (was breaking Special:RecentChanges' feeds, ...)
2010-01-13 20:23:14 +00:00
$query = "feed=$type";
Only advertise RSS feeds by default, not Atom Advertising both is pointless, since every feed reader in the universe supports both. Their functionality is essentially identical. And users might be given a choice of multiple feeds to read when there's really only one distinct feed. Plus, it cuts a line from the <head>. The Atom feed still exists, and you can get the <head> link back if you really want it with $wgAdvertisedFeedTypes.
2010-01-13 21:29:47 +00:00
if ( is_string( $val ) ) {
* (bug 21738) Fix for r57119: OutputPage::setSyndicated() now works again (was breaking Special:RecentChanges' feeds, ...)
2010-01-13 20:23:14 +00:00
$query .= '&' . $val;
Only advertise RSS feeds by default, not Atom Advertising both is pointless, since every feed reader in the universe supports both. Their functionality is essentially identical. And users might be given a choice of multiple feeds to read when there's really only one distinct feed. Plus, it cuts a line from the <head>. The Atom feed still exists, and you can get the <head> link back if you really want it with $wgAdvertisedFeedTypes.
2010-01-13 21:29:47 +00:00
}
Re-work feed exposure system. * Includes a new, more flexible system which allows you to provide an array of format => URL, but also supports the older system of providing a URL suffix. * Implemented it for the watchlist if a user has set their own token.
2009-09-30 17:35:41 +00:00
$this->mFeedLinks[$type] = $this->getTitle()->getLocalURL( $query );
}
}
* improved script loader language support ** supports url language code parameter * improved performance on cached hit ( apache bench mark goes from old: Rps: 23.40 [#/sec] new: Rps: 378.61 [#/sec] ** we don't check existence of wiki title until after cache file check. ** we don't load all includes/WebStart.php before checking cache * some remote search bug ** improved speed and accuracy of language swapping ( token check no breakage with }) in msg text limitation * improved core js2 performance in parsing class defines from mv_embed.js ** only reads 8k of head at a time. * reduced expire time to one day (instead of one year on script requests) * removed $wgEnableScriptMinify (just use debug or ?debug instead)
2009-10-24 21:32:53 +00:00
Follow-up r61021: added some documentation per Tim's request
2010-01-26 20:47:34 +00:00
/**
* Add a feed link to the page header
*
Cleanup some docs (includes/*.php) - Swap "$variable type" to "type $variable" - Added missing types - Fixed spacing inside docs - Makes beginning of @param/@return/@var/@throws in capital - Changed some types to match the more common spelling Change-Id: I783e4dbfe5f6f98b32b9a03ccf6439e13e132bcc
2014-07-24 17:42:24 +00:00
* @param string $format Feed type, should be a key of $wgFeedClasses
Fixed @param tags to conform with Doxygen format. Doxygen expects parameter types to come before the parameter name in @param tags. Used a quick regex to switch everything around where possible. This only fixes cases where a primitve variable (or a primitive followed by other types) is the variable type. Other cases will need to be fixed manually. Change-Id: Ic59fd20856eb0489d70f3469a56ebce0efb3db13
2013-03-11 17:15:01 +00:00
* @param string $href URL
Follow-up r61021: added some documentation per Tim's request
2010-01-26 20:47:34 +00:00
*/
Re-work feed exposure system. * Includes a new, more flexible system which allows you to provide an array of format => URL, but also supports the older system of providing a URL suffix. * Implemented it for the watchlist if a user has set their own token.
2009-09-30 17:35:41 +00:00
public function addFeedLink( $format, $href ) {
Advertise feeds only if $wgFeed is enabled Wikis with $wgFeed=false should not present an Atom icon on the RecentChanges/Watchlist. Bug: T116145 Change-Id: Ic64f7f5115a77c42ed2b336f6996fe711b3822ba
2018-01-28 02:21:51 +00:00
if ( in_array( $format, $this->getAdvertisedFeedTypes() ) ) {
Keep formats that are not in $wgAdvertisedFeedTypes from being advertised. Useful for things like LiquidThreads that use this function to provide links without modifying third party modules to support wgAdvertisedFeedTypes.
2010-05-11 19:36:47 +00:00
$this->mFeedLinks[$format] = $href;
}
Re-work feed exposure system. * Includes a new, more flexible system which allows you to provide an array of format => URL, but also supports the older system of providing a URL suffix. * Implemented it for the watchlist if a user has set their own token.
2009-09-30 17:35:41 +00:00
}
* improved script loader language support ** supports url language code parameter * improved performance on cached hit ( apache bench mark goes from old: Rps: 23.40 [#/sec] new: Rps: 378.61 [#/sec] ** we don't check existence of wiki title until after cache file check. ** we don't load all includes/WebStart.php before checking cache * some remote search bug ** improved speed and accuracy of language swapping ( token check no breakage with }) in msg text limitation * improved core js2 performance in parsing class defines from mv_embed.js ** only reads 8k of head at a time. * reduced expire time to one day (instead of one year on script requests) * removed $wgEnableScriptMinify (just use debug or ?debug instead)
2009-10-24 21:32:53 +00:00
Follow-up r61021: added some documentation per Tim's request
2010-01-26 20:47:34 +00:00
/**
Fix isSyndicated() description too
2010-01-27 00:11:13 +00:00
* Should we output feed links for this page?
Fixed some @params documentation Swapped some "$var type" to "type $var" or added missing types before the $var. Changed some other types to match the more common spelling. Makes beginning of some text in captial. Change-Id: I64e8cfe478cb0ba438f40b0631d6e9049cdab567
2014-04-10 18:50:10 +00:00
* @return bool
Follow-up r61021: added some documentation per Tim's request
2010-01-26 20:47:34 +00:00
*/
Step one in fixing OutputPage's documentation: * Groupped methods by function * Added documentation * Fixed some doxygen warnings * Some little style fixes (more to come)
2010-01-28 21:58:50 +00:00
public function isSyndicated() {
return count( $this->mFeedLinks ) > 0;
}
Add some documentation to OutputPage.php and specify public/private for most of its member functions (practically all public anyway). And replace one gratuitous use of strcmp with good old ===.
2006-11-07 05:37:31 +00:00
Step two in fixing OutputPage's documentation, more or less the same as r61647
2010-01-29 21:32:45 +00:00
/**
* Return URLs for each supported syndication format for this page.
Cleanup some docs (includes/*.php) - Swap "$variable type" to "type $variable" - Added missing types - Fixed spacing inside docs - Makes beginning of @param/@return/@var/@throws in capital - Changed some types to match the more common spelling Change-Id: I783e4dbfe5f6f98b32b9a03ccf6439e13e132bcc
2014-07-24 17:42:24 +00:00
* @return array Associating format keys with URLs
Step two in fixing OutputPage's documentation, more or less the same as r61647
2010-01-29 21:32:45 +00:00
*/
public function getSyndicationLinks() {
return $this->mFeedLinks;
}
Step one in fixing OutputPage's documentation: * Groupped methods by function * Added documentation * Fixed some doxygen warnings * Some little style fixes (more to come)
2010-01-28 21:58:50 +00:00
/**
* Will currently always return null
*
* @return null
*/
public function getFeedAppendQuery() {
return $this->mFeedLinksAppendQuery;
}
Follow-up r61021: added some documentation per Tim's request
2010-01-26 20:47:34 +00:00
Step one in fixing OutputPage's documentation: * Groupped methods by function * Added documentation * Fixed some doxygen warnings * Some little style fixes (more to come)
2010-01-28 21:58:50 +00:00
/**
* Set whether the displayed content is related to the source of the
* corresponding article on the wiki
* Setting true will cause the change "article related" toggle to true
*
Improve OutputPage test coverage more Improved locally from 26.81% to 37.18%. Depends-On: Ib564624c644ee6620ac06872f5684831acaaaadb Change-Id: Icdc0288c04b8c4ba841f9fbb3e05a0cdc8a20fa5
2018-07-25 18:41:42 +00:00
* @param bool $newVal
Step one in fixing OutputPage's documentation: * Groupped methods by function * Added documentation * Fixed some doxygen warnings * Some little style fixes (more to come)
2010-01-28 21:58:50 +00:00
*/
Improve OutputPage test coverage more Improved locally from 26.81% to 37.18%. Depends-On: Ib564624c644ee6620ac06872f5684831acaaaadb Change-Id: Icdc0288c04b8c4ba841f9fbb3e05a0cdc8a20fa5
2018-07-25 18:41:42 +00:00
public function setArticleFlag( $newVal ) {
$this->mIsArticle = $newVal;
if ( $newVal ) {
$this->mIsArticleRelated = $newVal;
Step one in fixing OutputPage's documentation: * Groupped methods by function * Added documentation * Fixed some doxygen warnings * Some little style fixes (more to come)
2010-01-28 21:58:50 +00:00
}
}
/**
* Return whether the content displayed page is related to the source of
* the corresponding article on the wiki
*
Fixed some @params documentation Swapped some "$var type" to "type $var" or added missing types before the $var. Changed some other types to match the more common spelling. Makes beginning of some text in captial. Change-Id: I64e8cfe478cb0ba438f40b0631d6e9049cdab567
2014-04-10 18:50:10 +00:00
* @return bool
Step one in fixing OutputPage's documentation: * Groupped methods by function * Added documentation * Fixed some doxygen warnings * Some little style fixes (more to come)
2010-01-28 21:58:50 +00:00
*/
public function isArticle() {
Improve OutputPage test coverage more Improved locally from 26.81% to 37.18%. Depends-On: Ib564624c644ee6620ac06872f5684831acaaaadb Change-Id: Icdc0288c04b8c4ba841f9fbb3e05a0cdc8a20fa5
2018-07-25 18:41:42 +00:00
return $this->mIsArticle;
Step one in fixing OutputPage's documentation: * Groupped methods by function * Added documentation * Fixed some doxygen warnings * Some little style fixes (more to come)
2010-01-28 21:58:50 +00:00
}
/**
* Set whether this page is related an article on the wiki
* Setting false will cause the change of "article flag" toggle to false
*
Improve OutputPage test coverage more Improved locally from 26.81% to 37.18%. Depends-On: Ib564624c644ee6620ac06872f5684831acaaaadb Change-Id: Icdc0288c04b8c4ba841f9fbb3e05a0cdc8a20fa5
2018-07-25 18:41:42 +00:00
* @param bool $newVal
Step one in fixing OutputPage's documentation: * Groupped methods by function * Added documentation * Fixed some doxygen warnings * Some little style fixes (more to come)
2010-01-28 21:58:50 +00:00
*/
Improve OutputPage test coverage more Improved locally from 26.81% to 37.18%. Depends-On: Ib564624c644ee6620ac06872f5684831acaaaadb Change-Id: Icdc0288c04b8c4ba841f9fbb3e05a0cdc8a20fa5
2018-07-25 18:41:42 +00:00
public function setArticleRelated( $newVal ) {
$this->mIsArticleRelated = $newVal;
if ( !$newVal ) {
$this->mIsArticle = false;
Bug fixes: watchlist and recent changes linked were broken in the last commit
2004-01-17 09:49:43 +00:00
}
}
Step one in fixing OutputPage's documentation: * Groupped methods by function * Added documentation * Fixed some doxygen warnings * Some little style fixes (more to come)
2010-01-28 21:58:50 +00:00
/**
* Return whether this page is related an article on the wiki
*
Fixed some @params documentation Swapped some "$var type" to "type $var" or added missing types before the $var. Changed some other types to match the more common spelling. Makes beginning of some text in captial. Change-Id: I64e8cfe478cb0ba438f40b0631d6e9049cdab567
2014-04-10 18:50:10 +00:00
* @return bool
Step one in fixing OutputPage's documentation: * Groupped methods by function * Added documentation * Fixed some doxygen warnings * Some little style fixes (more to come)
2010-01-28 21:58:50 +00:00
*/
public function isArticleRelated() {
return $this->mIsArticleRelated;
Bug fixes: watchlist and recent changes linked were broken in the last commit
2004-01-17 09:49:43 +00:00
}
Show copyright based on $output->hasCopyright() Introducing setCopyright/hasCopyright in OutputPage to make the showing of the copyright message controllable by pages other than articles. hasCopyright defaults to isArticle() to preserve the rule that all article should show the copyright. This is used by Flow to show the copyright on various pages that contain user-generated content. Bug: T184960 Change-Id: I3a50dbcedc6b119b9262c50cb3a84b0dd230fb3d
2018-10-03 13:12:37 +00:00
/**
* Set whether the standard copyright should be shown for the current page.
*
* @param bool $hasCopyright
*/
public function setCopyright( $hasCopyright ) {
$this->mHasCopyright = $hasCopyright;
}
/**
* Return whether the standard copyright should be shown for the current page.
* By default, it is true for all articles but other pages
* can signal it by using setCopyright( true ).
*
* Used by SkinTemplate to decided whether to show the copyright.
*
* @return bool
*/
public function showsCopyright() {
return $this->isArticle() || $this->mHasCopyright;
}
Step one in fixing OutputPage's documentation: * Groupped methods by function * Added documentation * Fixed some doxygen warnings * Some little style fixes (more to come)
2010-01-28 21:58:50 +00:00
/**
* Add new language links
*
Fix docs for OutputPage::addLanguageLinks and OutputPage::setLanguageLinks Per what has been described and fixed in Ie9c42ac2b4ff143e36d07642f57cca769e8c00e7. Change-Id: I2af28ae97805f3259ca038942a84b43f89b55150
2016-11-07 13:13:51 +00:00
* @param string[] $newLinkArray Array of interwiki-prefixed (non DB key) titles
* (e.g. 'fr:Test page')
Step one in fixing OutputPage's documentation: * Groupped methods by function * Added documentation * Fixed some doxygen warnings * Some little style fixes (more to come)
2010-01-28 21:58:50 +00:00
*/
Enforce array type hinting in OutputPage.php Change-Id: I2b37a95e517e120471c709823ea319f60165be0e
2014-07-23 02:49:30 +00:00
public function addLanguageLinks( array $newLinkArray ) {
Use array_merge() for OutputPage::$mLanguageLinks, not + Based on documentation together with inspection of some callers, the intent seems to be that this is an indexed array, not associative. + will therefore do totally the wrong thing, ignoring any new values that have the same key as an existing item (e.g., '0' or '1'). Even if it was an associative array, + keeps the values on the left-hand side, which means you normally want to do $foo = $bar + $foo instead of $foo += $bar if you want to overwrite old values with the new ones. Before this change, calling addLanguageLinks() or addParserOutputMetadata() would generally not add all of the links it was supposed to if there were already links defined. (It could still work if the arrays' keys didn't conflict for some reason, e.g., something passed an associative array or an indexed array with a hole.) I don't know if anything actually hits this bug, because it's likely that callers usually add all their links at once. I find no uses of addLanguageLinks() at all. I found this bug while working on adding more tests for OutputPage, and the tests for this change will be submitted later in Icdc0288c04b8c4ba841f9fbb3e05a0cdc8a20fa5. Change-Id: I53f6e7ea94417b0034371e56e733e8c86af21658
2018-07-25 18:36:56 +00:00
$this->mLanguageLinks = array_merge( $this->mLanguageLinks, $newLinkArray );
addLanguageLinks and setLanguageLinks
2004-06-01 18:29:31 +00:00
}
Step one in fixing OutputPage's documentation: * Groupped methods by function * Added documentation * Fixed some doxygen warnings * Some little style fixes (more to come)
2010-01-28 21:58:50 +00:00
/**
* Reset the language links and add new language links
*
Fix docs for OutputPage::addLanguageLinks and OutputPage::setLanguageLinks Per what has been described and fixed in Ie9c42ac2b4ff143e36d07642f57cca769e8c00e7. Change-Id: I2af28ae97805f3259ca038942a84b43f89b55150
2016-11-07 13:13:51 +00:00
* @param string[] $newLinkArray Array of interwiki-prefixed (non DB key) titles
* (e.g. 'fr:Test page')
Step one in fixing OutputPage's documentation: * Groupped methods by function * Added documentation * Fixed some doxygen warnings * Some little style fixes (more to come)
2010-01-28 21:58:50 +00:00
*/
Enforce array type hinting in OutputPage.php Change-Id: I2b37a95e517e120471c709823ea319f60165be0e
2014-07-23 02:49:30 +00:00
public function setLanguageLinks( array $newLinkArray ) {
addLanguageLinks and setLanguageLinks
2004-06-01 18:29:31 +00:00
$this->mLanguageLinks = $newLinkArray;
}
some comments / code foobar () { schemas
2004-09-17 15:24:43 +00:00
Step one in fixing OutputPage's documentation: * Groupped methods by function * Added documentation * Fixed some doxygen warnings * Some little style fixes (more to come)
2010-01-28 21:58:50 +00:00
/**
* Get the list of language links
*
Fix docs for OutputPage::addLanguageLinks and OutputPage::setLanguageLinks Per what has been described and fixed in Ie9c42ac2b4ff143e36d07642f57cca769e8c00e7. Change-Id: I2af28ae97805f3259ca038942a84b43f89b55150
2016-11-07 13:13:51 +00:00
* @return string[] Array of interwiki-prefixed (non DB key) titles (e.g. 'fr:Test page')
Step one in fixing OutputPage's documentation: * Groupped methods by function * Added documentation * Fixed some doxygen warnings * Some little style fixes (more to come)
2010-01-28 21:58:50 +00:00
*/
public function getLanguageLinks() {
return $this->mLanguageLinks;
This commit implements Josh Cogliati's "info" feature (e.g. http://en.wikipedia.org/w/wiki.phtml?title=Wikipedia&action=info), fixes the page headings for protect/unprotect, and does a lot of whitespace changes (sorry about that). I've also fixed confusing indentation somewhere.
2004-07-08 14:53:54 +00:00
}
* Added templatelinks table. The table currently represents a literal list of templates included from each article. That is, the table contains pages which were actually loaded during parsing, not the markup which went into resolving their names. * Ended the role of $wgLinkCache in link updates. Instead, links (and related entities) are registered in the ParserOutput object during a parse. The LinksUpdate constructor now takes a ParserOutput object as a parameter. $wgLinkCache is still used, but only as a cache of article IDs. * Because the link list is now saved and restored in the parser cache, meta tag keywords now work on parser cache hits. Some refactoring took place in this area. * Rendering of the HTML for category links has moved from Parser to OutputPage. * Did some general pottering around in Article.php, such as allowing an Article object to be created with a specified revision ID, thereby optionally removing the dependence on $wgRequest. Not used at the current time. * A few documentation tweaks.
2005-12-30 09:33:11 +00:00
/**
* Add an array of categories, with names in the keys
Step one in fixing OutputPage's documentation: * Groupped methods by function * Added documentation * Fixed some doxygen warnings * Some little style fixes (more to come)
2010-01-28 21:58:50 +00:00
*
Cleanup some docs (includes/*.php) - Swap "$variable type" to "type $variable" - Added missing types - Fixed spacing inside docs - Makes beginning of @param/@return/@var/@throws in capital - Changed some types to match the more common spelling Change-Id: I783e4dbfe5f6f98b32b9a03ccf6439e13e132bcc
2014-07-24 17:42:24 +00:00
* @param array $categories Mapping category name => sort key
* Added templatelinks table. The table currently represents a literal list of templates included from each article. That is, the table contains pages which were actually loaded during parsing, not the markup which went into resolving their names. * Ended the role of $wgLinkCache in link updates. Instead, links (and related entities) are registered in the ParserOutput object during a parse. The LinksUpdate constructor now takes a ParserOutput object as a parameter. $wgLinkCache is still used, but only as a cache of article IDs. * Because the link list is now saved and restored in the parser cache, meta tag keywords now work on parser cache hits. Some refactoring took place in this area. * Rendering of the HTML for category links has moved from Parser to OutputPage. * Did some general pottering around in Article.php, such as allowing an Article object to be created with a specified revision ID, thereby optionally removing the dependence on $wgRequest. Not used at the current time. * A few documentation tweaks.
2005-12-30 09:33:11 +00:00
*/
Enforce array type hinting in OutputPage.php Change-Id: I2b37a95e517e120471c709823ea319f60165be0e
2014-07-23 02:49:30 +00:00
public function addCategoryLinks( array $categories ) {
Improve OutputPage test coverage more Improved locally from 26.81% to 37.18%. Depends-On: Ib564624c644ee6620ac06872f5684831acaaaadb Change-Id: Icdc0288c04b8c4ba841f9fbb3e05a0cdc8a20fa5
2018-07-25 18:41:42 +00:00
if ( !$categories ) {
* Added __HIDDENCAT__ feature, to hide categories from the box at the bottom of the member pages depending on special text on the category page. * Added page_props backend, generic parser-driven page properties which, when changed, invalidate the cache of backlinked pages on any links table. Could be used to add overlays to images depending on their deletion status, or to add icons to articles based on category membership. * Refactored double-underscore handling in the parser. * Moved CoreParserFunctions registration to CoreParserFunctions.
2008-02-20 08:53:12 +00:00
return;
}
* Put both hidden categories and normal categories into the view page HTML, but with hidden categories being unconditionally hidden with CSS. A JS show/hide toggle can be added in user/site JS. * Add user preference to always show hidden categories * Add all hidden categories to [[Category:Hidden categories]], localised by hidden-category-category * Add wgVariantArticlePath and wgActionPaths to the JS variables script, needed to determine title from link href.
2008-02-25 16:38:25 +00:00
OutputPage::getCategories(): Add a possibility to distinguish "normal" and "hidden" categories There's currently no way to get, e.g., all categories except the hidden ones just as text. The OutputPage::getCategories() method always returns all categories as an array of strings (titles) and the getCategoryLinks() method returns the result of Linker::link but with the distinction between "normal" and "hidden" categories. This change adds a new parameter to OutputPage::getCategories(), $type, which can be used to define, what categories should be returned. The default value is "all", which means, that all categories are returned (the current result of the method). With the value "normal" and "hidden", the method will return the respective values. This could be used in I97d7de723fe72da26c7dbde0a559a13704c7099a to remove the stupid Linker::link() and isset workaround. Change-Id: Iadda9ae362a21fbee770240234b8f55326219932
2016-07-08 19:11:53 +00:00
$res = $this->addCategoryLinksToLBAndGetResult( $categories );
* Added __HIDDENCAT__ feature, to hide categories from the box at the bottom of the member pages depending on special text on the category page. * Added page_props backend, generic parser-driven page properties which, when changed, invalidate the cache of backlinked pages on any links table. Could be used to add overlays to images depending on their deletion status, or to add icons to articles based on category membership. * Refactored double-underscore handling in the parser. * Moved CoreParserFunctions registration to CoreParserFunctions.
2008-02-20 08:53:12 +00:00
Change addCategoryLinks() to use array_fill_keys Previously it used an ugly array_combine hack. Change-Id: I6bc4fb5cfffda468c61037588ebb4399b5121db5
2014-07-23 02:41:50 +00:00
# Set all the values to 'normal'.
$categories = array_fill_keys( array_keys( $categories ), 'normal' );
Reuse the query result in addCategoryLinks instead of relying on cache This is an implicit logic that it first loads everything as a batch and then adds them to the LinkCache and then just hopes it works. During passing arround of Title object, it reaches Title::getArticleID() (in Title::exists()) and that now reads it from PageStore (by using ::getFieldFromPageStore) and since the Page object doesn't exist yet (we only populated the LinkCache), it loads the page again. Bug: T297669 Change-Id: Icca81c97a15d90f593090711cdccdee3ea404ae2
2021-12-14 07:41:29 +00:00
$pageData = [];
* Put both hidden categories and normal categories into the view page HTML, but with hidden categories being unconditionally hidden with CSS. A JS show/hide toggle can be added in user/site JS. * Add user preference to always show hidden categories * Add all hidden categories to [[Category:Hidden categories]], localised by hidden-category-category * Add wgVariantArticlePath and wgActionPaths to the JS variables script, needed to determine title from link href.
2008-02-25 16:38:25 +00:00
# Mark hidden categories
* Added __HIDDENCAT__ feature, to hide categories from the box at the bottom of the member pages depending on special text on the category page. * Added page_props backend, generic parser-driven page properties which, when changed, invalidate the cache of backlinked pages on any links table. Could be used to add overlays to images depending on their deletion status, or to add icons to articles based on category membership. * Refactored double-underscore handling in the parser. * Moved CoreParserFunctions registration to CoreParserFunctions.
2008-02-20 08:53:12 +00:00
foreach ( $res as $row ) {
* Put both hidden categories and normal categories into the view page HTML, but with hidden categories being unconditionally hidden with CSS. A JS show/hide toggle can be added in user/site JS. * Add user preference to always show hidden categories * Add all hidden categories to [[Category:Hidden categories]], localised by hidden-category-category * Add wgVariantArticlePath and wgActionPaths to the JS variables script, needed to determine title from link href.
2008-02-25 16:38:25 +00:00
if ( isset( $row->pp_value ) ) {
$categories[$row->page_title] = 'hidden';
* Added __HIDDENCAT__ feature, to hide categories from the box at the bottom of the member pages depending on special text on the category page. * Added page_props backend, generic parser-driven page properties which, when changed, invalidate the cache of backlinked pages on any links table. Could be used to add overlays to images depending on their deletion status, or to add icons to articles based on category membership. * Refactored double-underscore handling in the parser. * Moved CoreParserFunctions registration to CoreParserFunctions.
2008-02-20 08:53:12 +00:00
}
Reuse the query result in addCategoryLinks instead of relying on cache This is an implicit logic that it first loads everything as a batch and then adds them to the LinkCache and then just hopes it works. During passing arround of Title object, it reaches Title::getArticleID() (in Title::exists()) and that now reads it from PageStore (by using ::getFieldFromPageStore) and since the Page object doesn't exist yet (we only populated the LinkCache), it loads the page again. Bug: T297669 Change-Id: Icca81c97a15d90f593090711cdccdee3ea404ae2
2021-12-14 07:41:29 +00:00
// Page exists, cache results
if ( isset( $row->page_id ) ) {
$pageData[$row->page_title] = $row;
}
* Added __HIDDENCAT__ feature, to hide categories from the box at the bottom of the member pages depending on special text on the category page. * Added page_props backend, generic parser-driven page properties which, when changed, invalidate the cache of backlinked pages on any links table. Could be used to add overlays to images depending on their deletion status, or to add icons to articles based on category membership. * Refactored double-underscore handling in the parser. * Moved CoreParserFunctions registration to CoreParserFunctions.
2008-02-20 08:53:12 +00:00
}
* Put both hidden categories and normal categories into the view page HTML, but with hidden categories being unconditionally hidden with CSS. A JS show/hide toggle can be added in user/site JS. * Add user preference to always show hidden categories * Add all hidden categories to [[Category:Hidden categories]], localised by hidden-category-category * Add wgVariantArticlePath and wgActionPaths to the JS variables script, needed to determine title from link href.
2008-02-25 16:38:25 +00:00
* Added __HIDDENCAT__ feature, to hide categories from the box at the bottom of the member pages depending on special text on the category page. * Added page_props backend, generic parser-driven page properties which, when changed, invalidate the cache of backlinked pages on any links table. Could be used to add overlays to images depending on their deletion status, or to add icons to articles based on category membership. * Refactored double-underscore handling in the parser. * Moved CoreParserFunctions registration to CoreParserFunctions.
2008-02-20 08:53:12 +00:00
# Add the remaining categories to the skin
Hooks::run() call site migration Migrate all callers of Hooks::run() to use the new HookContainer/HookRunner system. General principles: * Use DI if it is already used. We're not changing the way state is managed in this patch. * HookContainer is always injected, not HookRunner. HookContainer is a service, it's a more generic interface, it is the only thing that provides isRegistered() which is needed in some cases, and a HookRunner can be efficiently constructed from it (confirmed by benchmark). Because HookContainer is needed for object construction, it is also needed by all factories. * "Ask your friendly local base class". Big hierarchies like SpecialPage and ApiBase have getHookContainer() and getHookRunner() methods in the base class, and classes that extend that base class are not expected to know or care where the base class gets its HookContainer from. * ProtectedHookAccessorTrait provides protected getHookContainer() and getHookRunner() methods, getting them from the global service container. The point of this is to ease migration to DI by ensuring that call sites ask their local friendly base class rather than getting a HookRunner from the service container directly. * Private $this->hookRunner. In some smaller classes where accessor methods did not seem warranted, there is a private HookRunner property which is accessed directly. Very rarely (two cases), there is a protected property, for consistency with code that conventionally assumes protected=private, but in cases where the class might actually be overridden, a protected accessor is preferred over a protected property. * The last resort: Hooks::runner(). Mostly for static, file-scope and global code. In a few cases it was used for objects with broken construction schemes, out of horror or laziness. Constructors with new required arguments: * AuthManager * BadFileLookup * BlockManager * ClassicInterwikiLookup * ContentHandlerFactory * ContentSecurityPolicy * DefaultOptionsManager * DerivedPageDataUpdater * FullSearchResultWidget * HtmlCacheUpdater * LanguageFactory * LanguageNameUtils * LinkRenderer * LinkRendererFactory * LocalisationCache * MagicWordFactory * MessageCache * NamespaceInfo * PageEditStash * PageHandlerFactory * PageUpdater * ParserFactory * PermissionManager * RevisionStore * RevisionStoreFactory * SearchEngineConfig * SearchEngineFactory * SearchFormWidget * SearchNearMatcher * SessionBackend * SpecialPageFactory * UserNameUtils * UserOptionsManager * WatchedItemQueryService * WatchedItemStore Constructors with new optional arguments: * DefaultPreferencesFactory * Language * LinkHolderArray * MovePage * Parser * ParserCache * PasswordReset * Router setHookContainer() now required after construction: * AuthenticationProvider * ResourceLoaderModule * SearchEngine Change-Id: Id442b0dbe43aba84bd5cf801d86dedc768b082c7
2020-03-19 02:42:09 +00:00
if ( $this->getHookRunner()->onOutputPageMakeCategoryLinks(
$this, $categories, $this->mCategoryLinks )
Pass php-strict for OutputPage Change-Id: Ia5d03277fff543ca672dc8efc7eaaeca4888ac23
2014-05-11 01:54:15 +00:00
) {
Mass conversion of $wgContLang to service Brought to you by vim macros. Bug: T200246 Change-Id: I79e919f4553e3bd3eb714073fed7a43051b4fb2a
2018-07-29 12:24:54 +00:00
$services = MediaWikiServices::getInstance();
$linkRenderer = $services->getLinkRenderer();
Hard deprecation of Language::convertTitle(), ::findVariantLink(), ::updateConversionTable() Co-authored-by: C. Scott Ananian <cananian@wikimedia.org> Bug: T226832 Change-Id: I41a3b67490fc6b9d4c484f566d346a0d10c670e9
2020-05-06 00:26:09 +00:00
$languageConverter = $services->getLanguageConverterFactory()
->getLanguageConverter( $services->getContentLanguage() );
Add hook to replace logic for generating category links. May be used by CategoryTree one day.
2008-07-02 20:02:51 +00:00
foreach ( $categories as $category => $type ) {
Cast numeric cat names back to string in api output This could be considered a breaking change even though it unifies behaviour. Change-Id: Ib6bc7386f689cc6c2812e9e9406f18b20caa8fee
2015-12-03 14:15:59 +00:00
// array keys will cast numeric category names to ints, so cast back to string
$category = (string)$category;
1. Use param $ignoreOtherCond instead of $forTemplate and $ignoreSubmitCond in findVariantLink(). 2. Drop some duplicate categories produced by findVariantLink().
2009-02-02 07:54:43 +00:00
$origcategory = $category;
Reuse the query result in addCategoryLinks instead of relying on cache This is an implicit logic that it first loads everything as a batch and then adds them to the LinkCache and then just hopes it works. During passing arround of Title object, it reaches Title::getArticleID() (in Title::exists()) and that now reads it from PageStore (by using ::getFieldFromPageStore) and since the Page object doesn't exist yet (we only populated the LinkCache), it loads the page again. Bug: T297669 Change-Id: Icca81c97a15d90f593090711cdccdee3ea404ae2
2021-12-14 07:41:29 +00:00
if ( array_key_exists( $category, $pageData ) ) {
$title = Title::newFromRow( $pageData[$category] );
} else {
$title = Title::makeTitleSafe( NS_CATEGORY, $category );
}
Call to a member function getText() on a non-object Bug: 68394 Change-Id: I78a9cd940fed9134263f086af89c15b5777b57fc
2014-07-22 19:18:07 +00:00
if ( !$title ) {
continue;
}
Hard deprecation of Language::convertTitle(), ::findVariantLink(), ::updateConversionTable() Co-authored-by: C. Scott Ananian <cananian@wikimedia.org> Bug: T226832 Change-Id: I41a3b67490fc6b9d4c484f566d346a0d10c670e9
2020-05-06 00:26:09 +00:00
$languageConverter->findVariantLink( $category, $title, true );
Collapse nested if statements Change-Id: I1cecfe5884edb98c8ad0a441f4d82288d597f631
2014-07-22 19:19:41 +00:00
if ( $category != $origcategory && array_key_exists( $category, $categories ) ) {
continue;
*coding style cleanup (added braces, spacing tweaks etc.) *removed some unused globals
2010-05-22 12:18:22 +00:00
}
Hard deprecation of Language::convertTitle(), ::findVariantLink(), ::updateConversionTable() Co-authored-by: C. Scott Ananian <cananian@wikimedia.org> Bug: T226832 Change-Id: I41a3b67490fc6b9d4c484f566d346a0d10c670e9
2020-05-06 00:26:09 +00:00
$text = $languageConverter->convertHtml( $title->getText() );
OutputPage::getCategories(): Add a possibility to distinguish "normal" and "hidden" categories There's currently no way to get, e.g., all categories except the hidden ones just as text. The OutputPage::getCategories() method always returns all categories as an array of strings (titles) and the getCategoryLinks() method returns the result of Linker::link but with the distinction between "normal" and "hidden" categories. This change adds a new parameter to OutputPage::getCategories(), $type, which can be used to define, what categories should be returned. The default value is "all", which means, that all categories are returned (the current result of the method). With the value "normal" and "hidden", the method will return the respective values. This could be used in I97d7de723fe72da26c7dbde0a559a13704c7099a to remove the stupid Linker::link() and isset workaround. Change-Id: Iadda9ae362a21fbee770240234b8f55326219932
2016-07-08 19:11:53 +00:00
$this->mCategories[$type][] = $title->getText();
Replace Linker::link() with LinkRenderer in includes directory * CategoryViewer * OutputPage * Preferences Bug: T149346 Change-Id: I9f9bcd9b461884817e8ceefbc6757c436221e331
2016-11-30 21:28:55 +00:00
$this->mCategoryLinks[$type][] = $linkRenderer->makeLink( $title, new HtmlArmor( $text ) );
Add hook to replace logic for generating category links. May be used by CategoryTree one day.
2008-07-02 20:02:51 +00:00
}
* Added templatelinks table. The table currently represents a literal list of templates included from each article. That is, the table contains pages which were actually loaded during parsing, not the markup which went into resolving their names. * Ended the role of $wgLinkCache in link updates. Instead, links (and related entities) are registered in the ParserOutput object during a parse. The LinksUpdate constructor now takes a ParserOutput object as a parameter. $wgLinkCache is still used, but only as a cache of article IDs. * Because the link list is now saved and restored in the parser cache, meta tag keywords now work on parser cache hits. Some refactoring took place in this area. * Rendering of the HTML for category links has moved from Parser to OutputPage. * Did some general pottering around in Article.php, such as allowing an Article object to be created with a specified revision ID, thereby optionally removing the dependence on $wgRequest. Not used at the current time. * A few documentation tweaks.
2005-12-30 09:33:11 +00:00
}
Show category links in preview
2004-06-19 06:46:54 +00:00
}
* Added templatelinks table. The table currently represents a literal list of templates included from each article. That is, the table contains pages which were actually loaded during parsing, not the markup which went into resolving their names. * Ended the role of $wgLinkCache in link updates. Instead, links (and related entities) are registered in the ParserOutput object during a parse. The LinksUpdate constructor now takes a ParserOutput object as a parameter. $wgLinkCache is still used, but only as a cache of article IDs. * Because the link list is now saved and restored in the parser cache, meta tag keywords now work on parser cache hits. Some refactoring took place in this area. * Rendering of the HTML for category links has moved from Parser to OutputPage. * Did some general pottering around in Article.php, such as allowing an Article object to be created with a specified revision ID, thereby optionally removing the dependence on $wgRequest. Not used at the current time. * A few documentation tweaks.
2005-12-30 09:33:11 +00:00
OutputPage::getCategories(): Add a possibility to distinguish "normal" and "hidden" categories There's currently no way to get, e.g., all categories except the hidden ones just as text. The OutputPage::getCategories() method always returns all categories as an array of strings (titles) and the getCategoryLinks() method returns the result of Linker::link but with the distinction between "normal" and "hidden" categories. This change adds a new parameter to OutputPage::getCategories(), $type, which can be used to define, what categories should be returned. The default value is "all", which means, that all categories are returned (the current result of the method). With the value "normal" and "hidden", the method will return the respective values. This could be used in I97d7de723fe72da26c7dbde0a559a13704c7099a to remove the stupid Linker::link() and isset workaround. Change-Id: Iadda9ae362a21fbee770240234b8f55326219932
2016-07-08 19:11:53 +00:00
/**
* @param array $categories
Add missing use statements Swap some ResultWrapper/Database to interface IResultWrapper/IDatabase Change-Id: I312cb3cca5742f72efdb0965ba35362f2eb1129d
2018-06-27 18:49:23 +00:00
* @return bool|IResultWrapper
OutputPage::getCategories(): Add a possibility to distinguish "normal" and "hidden" categories There's currently no way to get, e.g., all categories except the hidden ones just as text. The OutputPage::getCategories() method always returns all categories as an array of strings (titles) and the getCategoryLinks() method returns the result of Linker::link but with the distinction between "normal" and "hidden" categories. This change adds a new parameter to OutputPage::getCategories(), $type, which can be used to define, what categories should be returned. The default value is "all", which means, that all categories are returned (the current result of the method). With the value "normal" and "hidden", the method will return the respective values. This could be used in I97d7de723fe72da26c7dbde0a559a13704c7099a to remove the stupid Linker::link() and isset workaround. Change-Id: Iadda9ae362a21fbee770240234b8f55326219932
2016-07-08 19:11:53 +00:00
*/
protected function addCategoryLinksToLBAndGetResult( array $categories ) {
# Add the links to a LinkBatch
$arr = [ NS_CATEGORY => $categories ];
Use LinkBatchFactory Change-Id: Ib82f91048faad8cf8be961d038f30c9261072af5
2020-08-07 22:22:37 +00:00
$linkBatchFactory = MediaWikiServices::getInstance()->getLinkBatchFactory();
$lb = $linkBatchFactory->newLinkBatch();
OutputPage::getCategories(): Add a possibility to distinguish "normal" and "hidden" categories There's currently no way to get, e.g., all categories except the hidden ones just as text. The OutputPage::getCategories() method always returns all categories as an array of strings (titles) and the getCategoryLinks() method returns the result of Linker::link but with the distinction between "normal" and "hidden" categories. This change adds a new parameter to OutputPage::getCategories(), $type, which can be used to define, what categories should be returned. The default value is "all", which means, that all categories are returned (the current result of the method). With the value "normal" and "hidden", the method will return the respective values. This could be used in I97d7de723fe72da26c7dbde0a559a13704c7099a to remove the stupid Linker::link() and isset workaround. Change-Id: Iadda9ae362a21fbee770240234b8f55326219932
2016-07-08 19:11:53 +00:00
$lb->setArray( $arr );
# Fetch existence plus the hiddencat property
$dbr = wfGetDB( DB_REPLICA );
$fields = array_merge(
LinkCache::getSelectFields(),
Stop merging LinkCache select fields with page_title and page_namespace LinkCache::getSelectFields() includes page_title and page_namespace since If77c2f9879d7bae71eb59944efd8b3798d16aa46, so we don't need to add those two specifically. Change-Id: I8fa9d563af3f9da90fb96369dab0ee5bde860081
2022-07-05 22:24:26 +00:00
[ 'pp_value' ]
OutputPage::getCategories(): Add a possibility to distinguish "normal" and "hidden" categories There's currently no way to get, e.g., all categories except the hidden ones just as text. The OutputPage::getCategories() method always returns all categories as an array of strings (titles) and the getCategoryLinks() method returns the result of Linker::link but with the distinction between "normal" and "hidden" categories. This change adds a new parameter to OutputPage::getCategories(), $type, which can be used to define, what categories should be returned. The default value is "all", which means, that all categories are returned (the current result of the method). With the value "normal" and "hidden", the method will return the respective values. This could be used in I97d7de723fe72da26c7dbde0a559a13704c7099a to remove the stupid Linker::link() and isset workaround. Change-Id: Iadda9ae362a21fbee770240234b8f55326219932
2016-07-08 19:11:53 +00:00
);
Use SelectQueryBuilder in OutputPage This makes the LEFT JOIN and the WHERE conditions somewhat more readable, IMHO. Bug: T311866 Change-Id: I76ef805530187ae9958b5a0fd8e5cda20eb969c6
2022-07-05 11:50:49 +00:00
$res = $dbr->newSelectQueryBuilder()
->select( $fields )
->from( 'page' )
->leftJoin( 'page_props', null, [
OutputPage::getCategories(): Add a possibility to distinguish "normal" and "hidden" categories There's currently no way to get, e.g., all categories except the hidden ones just as text. The OutputPage::getCategories() method always returns all categories as an array of strings (titles) and the getCategoryLinks() method returns the result of Linker::link but with the distinction between "normal" and "hidden" categories. This change adds a new parameter to OutputPage::getCategories(), $type, which can be used to define, what categories should be returned. The default value is "all", which means, that all categories are returned (the current result of the method). With the value "normal" and "hidden", the method will return the respective values. This could be used in I97d7de723fe72da26c7dbde0a559a13704c7099a to remove the stupid Linker::link() and isset workaround. Change-Id: Iadda9ae362a21fbee770240234b8f55326219932
2016-07-08 19:11:53 +00:00
'pp_propname' => 'hiddencat',
Use SelectQueryBuilder in OutputPage This makes the LEFT JOIN and the WHERE conditions somewhat more readable, IMHO. Bug: T311866 Change-Id: I76ef805530187ae9958b5a0fd8e5cda20eb969c6
2022-07-05 11:50:49 +00:00
'pp_page = page_id',
] )
->where( $lb->constructSet( 'page', $dbr ) )
->caller( __METHOD__ )
->fetchResultSet();
OutputPage::getCategories(): Add a possibility to distinguish "normal" and "hidden" categories There's currently no way to get, e.g., all categories except the hidden ones just as text. The OutputPage::getCategories() method always returns all categories as an array of strings (titles) and the getCategoryLinks() method returns the result of Linker::link but with the distinction between "normal" and "hidden" categories. This change adds a new parameter to OutputPage::getCategories(), $type, which can be used to define, what categories should be returned. The default value is "all", which means, that all categories are returned (the current result of the method). With the value "normal" and "hidden", the method will return the respective values. This could be used in I97d7de723fe72da26c7dbde0a559a13704c7099a to remove the stupid Linker::link() and isset workaround. Change-Id: Iadda9ae362a21fbee770240234b8f55326219932
2016-07-08 19:11:53 +00:00
# Add the results to the link cache
Avoid deprecated LinkCache::singleton() Change-Id: Ie0e5c4ef0fe6ec896378bb2433af0898655dd907
2018-06-11 06:55:11 +00:00
$linkCache = MediaWikiServices::getInstance()->getLinkCache();
$lb->addResultToCache( $linkCache, $res );
OutputPage::getCategories(): Add a possibility to distinguish "normal" and "hidden" categories There's currently no way to get, e.g., all categories except the hidden ones just as text. The OutputPage::getCategories() method always returns all categories as an array of strings (titles) and the getCategoryLinks() method returns the result of Linker::link but with the distinction between "normal" and "hidden" categories. This change adds a new parameter to OutputPage::getCategories(), $type, which can be used to define, what categories should be returned. The default value is "all", which means, that all categories are returned (the current result of the method). With the value "normal" and "hidden", the method will return the respective values. This could be used in I97d7de723fe72da26c7dbde0a559a13704c7099a to remove the stupid Linker::link() and isset workaround. Change-Id: Iadda9ae362a21fbee770240234b8f55326219932
2016-07-08 19:11:53 +00:00
return $res;
}
Step one in fixing OutputPage's documentation: * Groupped methods by function * Added documentation * Fixed some doxygen warnings * Some little style fixes (more to come)
2010-01-28 21:58:50 +00:00
/**
* Reset the category links (but not the category list) and add $categories
*
Cleanup some docs (includes/*.php) - Swap "$variable type" to "type $variable" - Added missing types - Fixed spacing inside docs - Makes beginning of @param/@return/@var/@throws in capital - Changed some types to match the more common spelling Change-Id: I783e4dbfe5f6f98b32b9a03ccf6439e13e132bcc
2014-07-24 17:42:24 +00:00
* @param array $categories Mapping category name => sort key
Step one in fixing OutputPage's documentation: * Groupped methods by function * Added documentation * Fixed some doxygen warnings * Some little style fixes (more to come)
2010-01-28 21:58:50 +00:00
*/
Enforce array type hinting in OutputPage.php Change-Id: I2b37a95e517e120471c709823ea319f60165be0e
2014-07-23 02:49:30 +00:00
public function setCategoryLinks( array $categories ) {
Convert all array() syntax to [] Per wikitech-l consensus: https://lists.wikimedia.org/pipermail/wikitech-l/2016-February/084821.html Notes: * Disabled CallTimePassByReference due to false positives (T127163) Change-Id: I2c8ce713ce6600a0bb7bf67537c87044c7a45c4b
2016-02-17 09:09:32 +00:00
$this->mCategoryLinks = [];
Step one in fixing OutputPage's documentation: * Groupped methods by function * Added documentation * Fixed some doxygen warnings * Some little style fixes (more to come)
2010-01-28 21:58:50 +00:00
$this->addCategoryLinks( $categories );
}
/**
* Get the list of category links, in a 2-D array with the following format:
* $arr[$type][] = $link, where $type is either "normal" or "hidden" (for
* hidden categories) and $link a HTML fragment with a link to the category
* page
*
Update a lot of unspecific "array" types in PHPDocs This includes fixing some mistakes, as well as removing redundant text that doesn't add new information, either because it literally repeats what the code already says, or is actually duplicated. Change-Id: I3a8dd8ce57192deda8916cc444c87d7ab1a36515
2020-10-28 10:01:33 +00:00
* @return string[][]
Revert "Revert "Categories are modelled as a portlet"" This reverts commit c360439219dfa1385572f86932513c1888e10d75. Reason for revert: The issue was fixed elsewhere, see T299352. Bug: T299352 Bug: T298302 Bug: T298801 Change-Id: I67f699e57e47b36d183e5d90ba407b48dbe5e425
2022-01-18 02:00:16 +00:00
* @return-taint none
Step one in fixing OutputPage's documentation: * Groupped methods by function * Added documentation * Fixed some doxygen warnings * Some little style fixes (more to come)
2010-01-28 21:58:50 +00:00
*/
public function getCategoryLinks() {
return $this->mCategoryLinks;
}
/**
OutputPage::getCategories(): Add a possibility to distinguish "normal" and "hidden" categories There's currently no way to get, e.g., all categories except the hidden ones just as text. The OutputPage::getCategories() method always returns all categories as an array of strings (titles) and the getCategoryLinks() method returns the result of Linker::link but with the distinction between "normal" and "hidden" categories. This change adds a new parameter to OutputPage::getCategories(), $type, which can be used to define, what categories should be returned. The default value is "all", which means, that all categories are returned (the current result of the method). With the value "normal" and "hidden", the method will return the respective values. This could be used in I97d7de723fe72da26c7dbde0a559a13704c7099a to remove the stupid Linker::link() and isset workaround. Change-Id: Iadda9ae362a21fbee770240234b8f55326219932
2016-07-08 19:11:53 +00:00
* Get the list of category names this page belongs to.
Step one in fixing OutputPage's documentation: * Groupped methods by function * Added documentation * Fixed some doxygen warnings * Some little style fixes (more to come)
2010-01-28 21:58:50 +00:00
*
OutputPage::getCategories(): Add a possibility to distinguish "normal" and "hidden" categories There's currently no way to get, e.g., all categories except the hidden ones just as text. The OutputPage::getCategories() method always returns all categories as an array of strings (titles) and the getCategoryLinks() method returns the result of Linker::link but with the distinction between "normal" and "hidden" categories. This change adds a new parameter to OutputPage::getCategories(), $type, which can be used to define, what categories should be returned. The default value is "all", which means, that all categories are returned (the current result of the method). With the value "normal" and "hidden", the method will return the respective values. This could be used in I97d7de723fe72da26c7dbde0a559a13704c7099a to remove the stupid Linker::link() and isset workaround. Change-Id: Iadda9ae362a21fbee770240234b8f55326219932
2016-07-08 19:11:53 +00:00
* @param string $type The type of categories which should be returned. Possible values:
* * all: all categories of all types
* * hidden: only the hidden categories
* * normal: all categories, except hidden categories
Update a lot of unspecific "array" types in PHPDocs This includes fixing some mistakes, as well as removing redundant text that doesn't add new information, either because it literally repeats what the code already says, or is actually duplicated. Change-Id: I3a8dd8ce57192deda8916cc444c87d7ab1a36515
2020-10-28 10:01:33 +00:00
* @return string[]
Step one in fixing OutputPage's documentation: * Groupped methods by function * Added documentation * Fixed some doxygen warnings * Some little style fixes (more to come)
2010-01-28 21:58:50 +00:00
*/
OutputPage::getCategories(): Add a possibility to distinguish "normal" and "hidden" categories There's currently no way to get, e.g., all categories except the hidden ones just as text. The OutputPage::getCategories() method always returns all categories as an array of strings (titles) and the getCategoryLinks() method returns the result of Linker::link but with the distinction between "normal" and "hidden" categories. This change adds a new parameter to OutputPage::getCategories(), $type, which can be used to define, what categories should be returned. The default value is "all", which means, that all categories are returned (the current result of the method). With the value "normal" and "hidden", the method will return the respective values. This could be used in I97d7de723fe72da26c7dbde0a559a13704c7099a to remove the stupid Linker::link() and isset workaround. Change-Id: Iadda9ae362a21fbee770240234b8f55326219932
2016-07-08 19:11:53 +00:00
public function getCategories( $type = 'all' ) {
if ( $type === 'all' ) {
$allCategories = [];
foreach ( $this->mCategories as $categories ) {
$allCategories = array_merge( $allCategories, $categories );
}
return $allCategories;
}
if ( !isset( $this->mCategories[$type] ) ) {
throw new InvalidArgumentException( 'Invalid category type given: ' . $type );
}
return $this->mCategories[$type];
Show category links in preview
2004-06-19 06:46:54 +00:00
}
Implement page status indicators Page status indicators are icons (or short text snippets) usually displayed in the top-right corner of the page, outside of the main content. Basically, <indicator name="foo">[[File:Foo.svg|20px]]</indicator> may be used on a page to place the icon in the indicator area. They are also known as top icons, page icons, heading icons or title icons. I found the discussion on bug 23796 highly illuminating. I suggest that everyone read it before suggesting different design choices. I spent some time with a thesaurus pondering the name. "Emblems" and "badges" were also considered, but the former has a much more limited meaning and the latter is already taken by Wikidata, with a similar but subtly different feature set. I am not aware of any naming conflicts ;) besides new talk page message "indicator" (used by core and Echo in some documents) and OOjs UI indicators (tiny icons like the arrow on a dropdown form element), which shouldn't be confusing. Potential use cases include: * "Lock" indicators for page protection levels * Featured/good article indicators * Redirect shortcuts display ("WP:VPT") * Links to help/manual for special pages * Coordinates?… or globe icon for inline pop-up maps Design features: * Skin-customizable. Skins can fully control where and how indicators are shown, or may just do <?php echo $this->getIndicators(); ?> to output the default structure. By default they are not shown at all. * Extension-customizable. Extensions can call ParserOutput::addIndicator() to insert an indicator from one of the numerous parser hooks. * Wiki-customizable. In addition to just using the parser functions, on-wiki styles and scripts can use the provided classes and ids (.mw-indicator, #mw-indicator-<name>) to customize their display. Design limitations: * Every indicator must have a unique identifier (name). It's not possible to create arrays, or to have several indicators with the same name. In case of duplicates, the latest occurrence of the parser function wins. * Indicators are displayed ordered by their names (and not occurrence order). This ensures consistency across pages and provides a simple means of ordering or grouping them. * Indicators are not stored, tracked or accessible outside of ParserOutput (in particular they're not in the page_props table). They are intended to merely reflect the content or metadata that is already present on the page, and not be data themselves. If you ever think you need to list pages with a given status indicator, instead figure out what it means and use the appropriate tracking category, special page report, already existing page_prop, or other means. Corresponding patch in Vector: I90a8ae15ac8275d084ea5f47b6b2684d5e6c7412. I'll implement support in the other three skins included in the tarball and document it on mediawiki.org after this is merged. Bug: 23796 Change-Id: I2389ff9a5332a2b1d033eb75f0946e5241cfaaf4
2014-09-24 10:44:16 +00:00
/**
Add top help link to MediaWiki.org in several pages via indicator All the chosen targets are translatable public domain help pages on MediaWiki.org. Mostly special pages and actions for privileged users for now. Adapted from the Translate extension, credit to Niklas Laxström (TranslateUtils::addSpecialHelpLink). Depends on 6f5b29ff4e6fdf21b5a8cacaf10d6aceaee26a7d, whose commit message has a typo addIndicator() instead of setIndicator(). Bug: T45591 Change-Id: I2934b1708a0d207dcf3d940264f140613646f203
2014-08-31 10:25:18 +00:00
* Add an array of indicators, with their identifiers as array
* keys and HTML contents as values.
Implement page status indicators Page status indicators are icons (or short text snippets) usually displayed in the top-right corner of the page, outside of the main content. Basically, <indicator name="foo">[[File:Foo.svg|20px]]</indicator> may be used on a page to place the icon in the indicator area. They are also known as top icons, page icons, heading icons or title icons. I found the discussion on bug 23796 highly illuminating. I suggest that everyone read it before suggesting different design choices. I spent some time with a thesaurus pondering the name. "Emblems" and "badges" were also considered, but the former has a much more limited meaning and the latter is already taken by Wikidata, with a similar but subtly different feature set. I am not aware of any naming conflicts ;) besides new talk page message "indicator" (used by core and Echo in some documents) and OOjs UI indicators (tiny icons like the arrow on a dropdown form element), which shouldn't be confusing. Potential use cases include: * "Lock" indicators for page protection levels * Featured/good article indicators * Redirect shortcuts display ("WP:VPT") * Links to help/manual for special pages * Coordinates?… or globe icon for inline pop-up maps Design features: * Skin-customizable. Skins can fully control where and how indicators are shown, or may just do <?php echo $this->getIndicators(); ?> to output the default structure. By default they are not shown at all. * Extension-customizable. Extensions can call ParserOutput::addIndicator() to insert an indicator from one of the numerous parser hooks. * Wiki-customizable. In addition to just using the parser functions, on-wiki styles and scripts can use the provided classes and ids (.mw-indicator, #mw-indicator-<name>) to customize their display. Design limitations: * Every indicator must have a unique identifier (name). It's not possible to create arrays, or to have several indicators with the same name. In case of duplicates, the latest occurrence of the parser function wins. * Indicators are displayed ordered by their names (and not occurrence order). This ensures consistency across pages and provides a simple means of ordering or grouping them. * Indicators are not stored, tracked or accessible outside of ParserOutput (in particular they're not in the page_props table). They are intended to merely reflect the content or metadata that is already present on the page, and not be data themselves. If you ever think you need to list pages with a given status indicator, instead figure out what it means and use the appropriate tracking category, special page report, already existing page_prop, or other means. Corresponding patch in Vector: I90a8ae15ac8275d084ea5f47b6b2684d5e6c7412. I'll implement support in the other three skins included in the tarball and document it on mediawiki.org after this is merged. Bug: 23796 Change-Id: I2389ff9a5332a2b1d033eb75f0946e5241cfaaf4
2014-09-24 10:44:16 +00:00
*
* In case of duplicate keys, existing values are overwritten.
*
parse: Wrap indicators with a div.mw-parser-output element Alternative to having this in the skin, is to do it for each separate indicator. A counter argument against this was inline vs block mode elements, specifically the reason why we stripped the p element. But the p element has a margin in each and every skin. It's rather unlikely that div's have such a standard layout. The wrapping div around each indicator at the skin level is already inline-block, so interactions between the indicators are already decoupled via that skin element. In this patch the wrapping is done in OutputPage when a ParserOutput is added. OutputPage::addHelpLink() also adds indicators, which would not be wrapped by this patch; this is reasonable since the content of ::addHelpLink() is not parsed wikitext. It's possible external users may also call OutputPage::setIndicators() directly, which would also add unwrapped indicators; the caller of OutputPage::setIndicators() is responsible for wrapping any indicators which might come from parsed wikitext. Bug: T188443 Co-Authored-By: C. Scott Ananian <cananian@wikimedia.org> Change-Id: I42be893101f7d7887480fe635143bea535ee5634
2021-12-22 00:19:02 +00:00
* @note External code which calls this method should ensure that
* any indicators sourced from parsed wikitext are wrapped with
* the appropriate class; see note in ::getIndicators().
*
Update a lot of unspecific "array" types in PHPDocs This includes fixing some mistakes, as well as removing redundant text that doesn't add new information, either because it literally repeats what the code already says, or is actually duplicated. Change-Id: I3a8dd8ce57192deda8916cc444c87d7ab1a36515
2020-10-28 10:01:33 +00:00
* @param string[] $indicators
Implement page status indicators Page status indicators are icons (or short text snippets) usually displayed in the top-right corner of the page, outside of the main content. Basically, <indicator name="foo">[[File:Foo.svg|20px]]</indicator> may be used on a page to place the icon in the indicator area. They are also known as top icons, page icons, heading icons or title icons. I found the discussion on bug 23796 highly illuminating. I suggest that everyone read it before suggesting different design choices. I spent some time with a thesaurus pondering the name. "Emblems" and "badges" were also considered, but the former has a much more limited meaning and the latter is already taken by Wikidata, with a similar but subtly different feature set. I am not aware of any naming conflicts ;) besides new talk page message "indicator" (used by core and Echo in some documents) and OOjs UI indicators (tiny icons like the arrow on a dropdown form element), which shouldn't be confusing. Potential use cases include: * "Lock" indicators for page protection levels * Featured/good article indicators * Redirect shortcuts display ("WP:VPT") * Links to help/manual for special pages * Coordinates?… or globe icon for inline pop-up maps Design features: * Skin-customizable. Skins can fully control where and how indicators are shown, or may just do <?php echo $this->getIndicators(); ?> to output the default structure. By default they are not shown at all. * Extension-customizable. Extensions can call ParserOutput::addIndicator() to insert an indicator from one of the numerous parser hooks. * Wiki-customizable. In addition to just using the parser functions, on-wiki styles and scripts can use the provided classes and ids (.mw-indicator, #mw-indicator-<name>) to customize their display. Design limitations: * Every indicator must have a unique identifier (name). It's not possible to create arrays, or to have several indicators with the same name. In case of duplicates, the latest occurrence of the parser function wins. * Indicators are displayed ordered by their names (and not occurrence order). This ensures consistency across pages and provides a simple means of ordering or grouping them. * Indicators are not stored, tracked or accessible outside of ParserOutput (in particular they're not in the page_props table). They are intended to merely reflect the content or metadata that is already present on the page, and not be data themselves. If you ever think you need to list pages with a given status indicator, instead figure out what it means and use the appropriate tracking category, special page report, already existing page_prop, or other means. Corresponding patch in Vector: I90a8ae15ac8275d084ea5f47b6b2684d5e6c7412. I'll implement support in the other three skins included in the tarball and document it on mediawiki.org after this is merged. Bug: 23796 Change-Id: I2389ff9a5332a2b1d033eb75f0946e5241cfaaf4
2014-09-24 10:44:16 +00:00
* @since 1.25
*/
public function setIndicators( array $indicators ) {
$this->mIndicators = $indicators + $this->mIndicators;
// Keep ordered by key
ksort( $this->mIndicators );
}
/**
* Get the indicators associated with this page.
*
* The array will be internally ordered by item keys.
*
Update a lot of unspecific "array" types in PHPDocs This includes fixing some mistakes, as well as removing redundant text that doesn't add new information, either because it literally repeats what the code already says, or is actually duplicated. Change-Id: I3a8dd8ce57192deda8916cc444c87d7ab1a36515
2020-10-28 10:01:33 +00:00
* @return string[] Keys: identifiers, values: HTML contents
Implement page status indicators Page status indicators are icons (or short text snippets) usually displayed in the top-right corner of the page, outside of the main content. Basically, <indicator name="foo">[[File:Foo.svg|20px]]</indicator> may be used on a page to place the icon in the indicator area. They are also known as top icons, page icons, heading icons or title icons. I found the discussion on bug 23796 highly illuminating. I suggest that everyone read it before suggesting different design choices. I spent some time with a thesaurus pondering the name. "Emblems" and "badges" were also considered, but the former has a much more limited meaning and the latter is already taken by Wikidata, with a similar but subtly different feature set. I am not aware of any naming conflicts ;) besides new talk page message "indicator" (used by core and Echo in some documents) and OOjs UI indicators (tiny icons like the arrow on a dropdown form element), which shouldn't be confusing. Potential use cases include: * "Lock" indicators for page protection levels * Featured/good article indicators * Redirect shortcuts display ("WP:VPT") * Links to help/manual for special pages * Coordinates?… or globe icon for inline pop-up maps Design features: * Skin-customizable. Skins can fully control where and how indicators are shown, or may just do <?php echo $this->getIndicators(); ?> to output the default structure. By default they are not shown at all. * Extension-customizable. Extensions can call ParserOutput::addIndicator() to insert an indicator from one of the numerous parser hooks. * Wiki-customizable. In addition to just using the parser functions, on-wiki styles and scripts can use the provided classes and ids (.mw-indicator, #mw-indicator-<name>) to customize their display. Design limitations: * Every indicator must have a unique identifier (name). It's not possible to create arrays, or to have several indicators with the same name. In case of duplicates, the latest occurrence of the parser function wins. * Indicators are displayed ordered by their names (and not occurrence order). This ensures consistency across pages and provides a simple means of ordering or grouping them. * Indicators are not stored, tracked or accessible outside of ParserOutput (in particular they're not in the page_props table). They are intended to merely reflect the content or metadata that is already present on the page, and not be data themselves. If you ever think you need to list pages with a given status indicator, instead figure out what it means and use the appropriate tracking category, special page report, already existing page_prop, or other means. Corresponding patch in Vector: I90a8ae15ac8275d084ea5f47b6b2684d5e6c7412. I'll implement support in the other three skins included in the tarball and document it on mediawiki.org after this is merged. Bug: 23796 Change-Id: I2389ff9a5332a2b1d033eb75f0946e5241cfaaf4
2014-09-24 10:44:16 +00:00
* @since 1.25
*/
public function getIndicators() {
parse: Wrap indicators with a div.mw-parser-output element Alternative to having this in the skin, is to do it for each separate indicator. A counter argument against this was inline vs block mode elements, specifically the reason why we stripped the p element. But the p element has a margin in each and every skin. It's rather unlikely that div's have such a standard layout. The wrapping div around each indicator at the skin level is already inline-block, so interactions between the indicators are already decoupled via that skin element. In this patch the wrapping is done in OutputPage when a ParserOutput is added. OutputPage::addHelpLink() also adds indicators, which would not be wrapped by this patch; this is reasonable since the content of ::addHelpLink() is not parsed wikitext. It's possible external users may also call OutputPage::setIndicators() directly, which would also add unwrapped indicators; the caller of OutputPage::setIndicators() is responsible for wrapping any indicators which might come from parsed wikitext. Bug: T188443 Co-Authored-By: C. Scott Ananian <cananian@wikimedia.org> Change-Id: I42be893101f7d7887480fe635143bea535ee5634
2021-12-22 00:19:02 +00:00
// Note that some -- but not all -- indicators will be wrapped
// with a class appropriate for user-generated wikitext content
// (usually .mw-parser-output). The exceptions would be an
// indicator added via ::addHelpLink() below, which adds content
// which don't come from the parser and is not user-generated;
// and any indicators added by extensions which may call
// OutputPage::setIndicators() directly. In the latter case the
// caller is responsible for wrapping any parser-generated
// indicators.
Implement page status indicators Page status indicators are icons (or short text snippets) usually displayed in the top-right corner of the page, outside of the main content. Basically, <indicator name="foo">[[File:Foo.svg|20px]]</indicator> may be used on a page to place the icon in the indicator area. They are also known as top icons, page icons, heading icons or title icons. I found the discussion on bug 23796 highly illuminating. I suggest that everyone read it before suggesting different design choices. I spent some time with a thesaurus pondering the name. "Emblems" and "badges" were also considered, but the former has a much more limited meaning and the latter is already taken by Wikidata, with a similar but subtly different feature set. I am not aware of any naming conflicts ;) besides new talk page message "indicator" (used by core and Echo in some documents) and OOjs UI indicators (tiny icons like the arrow on a dropdown form element), which shouldn't be confusing. Potential use cases include: * "Lock" indicators for page protection levels * Featured/good article indicators * Redirect shortcuts display ("WP:VPT") * Links to help/manual for special pages * Coordinates?… or globe icon for inline pop-up maps Design features: * Skin-customizable. Skins can fully control where and how indicators are shown, or may just do <?php echo $this->getIndicators(); ?> to output the default structure. By default they are not shown at all. * Extension-customizable. Extensions can call ParserOutput::addIndicator() to insert an indicator from one of the numerous parser hooks. * Wiki-customizable. In addition to just using the parser functions, on-wiki styles and scripts can use the provided classes and ids (.mw-indicator, #mw-indicator-<name>) to customize their display. Design limitations: * Every indicator must have a unique identifier (name). It's not possible to create arrays, or to have several indicators with the same name. In case of duplicates, the latest occurrence of the parser function wins. * Indicators are displayed ordered by their names (and not occurrence order). This ensures consistency across pages and provides a simple means of ordering or grouping them. * Indicators are not stored, tracked or accessible outside of ParserOutput (in particular they're not in the page_props table). They are intended to merely reflect the content or metadata that is already present on the page, and not be data themselves. If you ever think you need to list pages with a given status indicator, instead figure out what it means and use the appropriate tracking category, special page report, already existing page_prop, or other means. Corresponding patch in Vector: I90a8ae15ac8275d084ea5f47b6b2684d5e6c7412. I'll implement support in the other three skins included in the tarball and document it on mediawiki.org after this is merged. Bug: 23796 Change-Id: I2389ff9a5332a2b1d033eb75f0946e5241cfaaf4
2014-09-24 10:44:16 +00:00
return $this->mIndicators;
}
Add top help link to MediaWiki.org in several pages via indicator All the chosen targets are translatable public domain help pages on MediaWiki.org. Mostly special pages and actions for privileged users for now. Adapted from the Translate extension, credit to Niklas Laxström (TranslateUtils::addSpecialHelpLink). Depends on 6f5b29ff4e6fdf21b5a8cacaf10d6aceaee26a7d, whose commit message has a typo addIndicator() instead of setIndicator(). Bug: T45591 Change-Id: I2934b1708a0d207dcf3d940264f140613646f203
2014-08-31 10:25:18 +00:00
/**
* Adds help link with an icon via page indicators.
Allow to customise addHelpLink() target via system message Method similar to SpecialPage::outputHeader() to avoid registering tons of system messages and to have -summary and -helppage tidily listed together in Special:AllMessages by default. Bug: T45591 Change-Id: Ic849dde00be7379c1909a8486cf20f48c5aea5cf
2015-05-02 11:37:19 +00:00
* Link target can be overridden by a local message containing a wikilink:
* the message key is: lowercase action or special page name + '-helppage'.
* @param string $to Target MediaWiki.org page title or encoded URL.
* @param bool $overrideBaseUrl Whether $url is a full URL, to avoid MW.o.
Add top help link to MediaWiki.org in several pages via indicator All the chosen targets are translatable public domain help pages on MediaWiki.org. Mostly special pages and actions for privileged users for now. Adapted from the Translate extension, credit to Niklas Laxström (TranslateUtils::addSpecialHelpLink). Depends on 6f5b29ff4e6fdf21b5a8cacaf10d6aceaee26a7d, whose commit message has a typo addIndicator() instead of setIndicator(). Bug: T45591 Change-Id: I2934b1708a0d207dcf3d940264f140613646f203
2014-08-31 10:25:18 +00:00
* @since 1.25
*/
public function addHelpLink( $to, $overrideBaseUrl = false ) {
Don't require JavaScript for addHelpLink styles And avoid wfMessage. Followup to 376c0a13df05f30367b2a89c2cac7bdab22600ed Bug: T45591 Change-Id: I2f02c2a4db609356ff8be30f260490d19f3226d1
2015-03-07 17:36:41 +00:00
$this->addModuleStyles( 'mediawiki.helplink' );
$text = $this->msg( 'helppage-top-gethelp' )->escaped();
Add top help link to MediaWiki.org in several pages via indicator All the chosen targets are translatable public domain help pages on MediaWiki.org. Mostly special pages and actions for privileged users for now. Adapted from the Translate extension, credit to Niklas Laxström (TranslateUtils::addSpecialHelpLink). Depends on 6f5b29ff4e6fdf21b5a8cacaf10d6aceaee26a7d, whose commit message has a typo addIndicator() instead of setIndicator(). Bug: T45591 Change-Id: I2934b1708a0d207dcf3d940264f140613646f203
2014-08-31 10:25:18 +00:00
if ( $overrideBaseUrl ) {
$helpUrl = $to;
} else {
Use underscores in help links Replace spaces by underscore to build correct links to wiki pages. IE11 will show %20 for spaces. Also use urlencode to make the url safe. Follow-Up: I2934b1708a0d207dcf3d940264f140613646f203 Change-Id: I5ef08441406e96aa9749476af0a81fc11fa4e4d6
2015-03-10 18:26:39 +00:00
$toUrlencoded = wfUrlencode( str_replace( ' ', '_', $to ) );
Use HTTPS instead of protocol relative for WMF links The WMF websites redirect from HTTP to HTTPS. Change-Id: I916c97697b4ffa9774eec9c8e6ab650a08bb6edd
2019-05-17 05:16:16 +00:00
$helpUrl = "https://www.mediawiki.org/wiki/Special:MyLanguage/$toUrlencoded";
Add top help link to MediaWiki.org in several pages via indicator All the chosen targets are translatable public domain help pages on MediaWiki.org. Mostly special pages and actions for privileged users for now. Adapted from the Translate extension, credit to Niklas Laxström (TranslateUtils::addSpecialHelpLink). Depends on 6f5b29ff4e6fdf21b5a8cacaf10d6aceaee26a7d, whose commit message has a typo addIndicator() instead of setIndicator(). Bug: T45591 Change-Id: I2934b1708a0d207dcf3d940264f140613646f203
2014-08-31 10:25:18 +00:00
}
Allow to customise addHelpLink() target via system message Method similar to SpecialPage::outputHeader() to avoid registering tons of system messages and to have -summary and -helppage tidily listed together in Special:AllMessages by default. Bug: T45591 Change-Id: Ic849dde00be7379c1909a8486cf20f48c5aea5cf
2015-05-02 11:37:19 +00:00
Add top help link to MediaWiki.org in several pages via indicator All the chosen targets are translatable public domain help pages on MediaWiki.org. Mostly special pages and actions for privileged users for now. Adapted from the Translate extension, credit to Niklas Laxström (TranslateUtils::addSpecialHelpLink). Depends on 6f5b29ff4e6fdf21b5a8cacaf10d6aceaee26a7d, whose commit message has a typo addIndicator() instead of setIndicator(). Bug: T45591 Change-Id: I2934b1708a0d207dcf3d940264f140613646f203
2014-08-31 10:25:18 +00:00
$link = Html::rawElement(
'a',
Convert all array() syntax to [] Per wikitech-l consensus: https://lists.wikimedia.org/pipermail/wikitech-l/2016-February/084821.html Notes: * Disabled CallTimePassByReference due to false positives (T127163) Change-Id: I2c8ce713ce6600a0bb7bf67537c87044c7a45c4b
2016-02-17 09:09:32 +00:00
[
Add top help link to MediaWiki.org in several pages via indicator All the chosen targets are translatable public domain help pages on MediaWiki.org. Mostly special pages and actions for privileged users for now. Adapted from the Translate extension, credit to Niklas Laxström (TranslateUtils::addSpecialHelpLink). Depends on 6f5b29ff4e6fdf21b5a8cacaf10d6aceaee26a7d, whose commit message has a typo addIndicator() instead of setIndicator(). Bug: T45591 Change-Id: I2934b1708a0d207dcf3d940264f140613646f203
2014-08-31 10:25:18 +00:00
'href' => $helpUrl,
'target' => '_blank',
'class' => 'mw-helplink',
Convert all array() syntax to [] Per wikitech-l consensus: https://lists.wikimedia.org/pipermail/wikitech-l/2016-February/084821.html Notes: * Disabled CallTimePassByReference due to false positives (T127163) Change-Id: I2c8ce713ce6600a0bb7bf67537c87044c7a45c4b
2016-02-17 09:09:32 +00:00
],
Add top help link to MediaWiki.org in several pages via indicator All the chosen targets are translatable public domain help pages on MediaWiki.org. Mostly special pages and actions for privileged users for now. Adapted from the Translate extension, credit to Niklas Laxström (TranslateUtils::addSpecialHelpLink). Depends on 6f5b29ff4e6fdf21b5a8cacaf10d6aceaee26a7d, whose commit message has a typo addIndicator() instead of setIndicator(). Bug: T45591 Change-Id: I2934b1708a0d207dcf3d940264f140613646f203
2014-08-31 10:25:18 +00:00
$text
);
parse: Wrap indicators with a div.mw-parser-output element Alternative to having this in the skin, is to do it for each separate indicator. A counter argument against this was inline vs block mode elements, specifically the reason why we stripped the p element. But the p element has a margin in each and every skin. It's rather unlikely that div's have such a standard layout. The wrapping div around each indicator at the skin level is already inline-block, so interactions between the indicators are already decoupled via that skin element. In this patch the wrapping is done in OutputPage when a ParserOutput is added. OutputPage::addHelpLink() also adds indicators, which would not be wrapped by this patch; this is reasonable since the content of ::addHelpLink() is not parsed wikitext. It's possible external users may also call OutputPage::setIndicators() directly, which would also add unwrapped indicators; the caller of OutputPage::setIndicators() is responsible for wrapping any indicators which might come from parsed wikitext. Bug: T188443 Co-Authored-By: C. Scott Ananian <cananian@wikimedia.org> Change-Id: I42be893101f7d7887480fe635143bea535ee5634
2021-12-22 00:19:02 +00:00
// See note in ::getIndicators() above -- unlike wikitext-generated
// indicators which come from ParserOutput, this indicator will not
// be wrapped.
Convert all array() syntax to [] Per wikitech-l consensus: https://lists.wikimedia.org/pipermail/wikitech-l/2016-February/084821.html Notes: * Disabled CallTimePassByReference due to false positives (T127163) Change-Id: I2c8ce713ce6600a0bb7bf67537c87044c7a45c4b
2016-02-17 09:09:32 +00:00
$this->setIndicators( [ 'mw-helplink' => $link ] );
Add top help link to MediaWiki.org in several pages via indicator All the chosen targets are translatable public domain help pages on MediaWiki.org. Mostly special pages and actions for privileged users for now. Adapted from the Translate extension, credit to Niklas Laxström (TranslateUtils::addSpecialHelpLink). Depends on 6f5b29ff4e6fdf21b5a8cacaf10d6aceaee26a7d, whose commit message has a typo addIndicator() instead of setIndicator(). Bug: T45591 Change-Id: I2934b1708a0d207dcf3d940264f140613646f203
2014-08-31 10:25:18 +00:00
}
Step two in fixing OutputPage's documentation, more or less the same as r61647
2010-01-29 21:32:45 +00:00
/**
Make allowing site-wide styles on restricted special pages a config option This mostly reverts commit 614d7e5c274d927f99bfc52ac3a1e6c7e5902408. Many wikis use MediaWiki:Common.css and associated pages to create a custom "theme" for their wiki, which would no longer load on login or preference pages, creating an inconsistent UI. This re-adds the difference in module origin for different types (styles, scripts, etc.), and now OutputPage::disallowUserJs() checks the value of the "AllowSiteCSSOnRestrictedPages" config setting to determine whether to allow site-wide CSS styles or not. By default this feature is disabled to be secure by default. Bug: 71621 Change-Id: I1bf4dd1845b6952c3985e179fbea48181ffb8907
2014-10-10 06:46:12 +00:00
* Do not allow scripts which can be modified by wiki users to load on this page;
* only allow scripts bundled with, or generated by, the software.
* Site-wide styles are controlled by a config setting, since they can be
* used to create a custom skin/theme, but not user-specific ones.
SECURITY: OutputPage: Remove separation of css and js module allowance * No longer segment module origin allowance by an "only=" content type. Both can be sensitive security-wise and there's no valid use case for allowing CSS anywhere you want to disallow JS. Both can significantly impact the user interface and cause unintended actions to be taken on the user's behalf, or desired actions to be made practically impossible. * While at it, also remove the ability to set the module allowance directly. The reduceAllowedModuleOrigin method is all we need. I couldn't find usage or mention of setAllowedModules() in mediawiki-core nor in any other Wikimedia-hosted repository. Bug: 70672 Change-Id: I308e794daca27a9380c67be350f8ab51f9c2de34
2014-09-11 23:14:16 +00:00
*
Make allowing site-wide styles on restricted special pages a config option This mostly reverts commit 614d7e5c274d927f99bfc52ac3a1e6c7e5902408. Many wikis use MediaWiki:Common.css and associated pages to create a custom "theme" for their wiki, which would no longer load on login or preference pages, creating an inconsistent UI. This re-adds the difference in module origin for different types (styles, scripts, etc.), and now OutputPage::disallowUserJs() checks the value of the "AllowSiteCSSOnRestrictedPages" config setting to determine whether to allow site-wide CSS styles or not. By default this feature is disabled to be secure by default. Bug: 71621 Change-Id: I1bf4dd1845b6952c3985e179fbea48181ffb8907
2014-10-10 06:46:12 +00:00
* @todo this should be given a more accurate name
Step two in fixing OutputPage's documentation, more or less the same as r61647
2010-01-29 21:32:45 +00:00
*/
public function disallowUserJs() {
Make allowing site-wide styles on restricted special pages a config option This mostly reverts commit 614d7e5c274d927f99bfc52ac3a1e6c7e5902408. Many wikis use MediaWiki:Common.css and associated pages to create a custom "theme" for their wiki, which would no longer load on login or preference pages, creating an inconsistent UI. This re-adds the difference in module origin for different types (styles, scripts, etc.), and now OutputPage::disallowUserJs() checks the value of the "AllowSiteCSSOnRestrictedPages" config setting to determine whether to allow site-wide CSS styles or not. By default this feature is disabled to be secure by default. Bug: 71621 Change-Id: I1bf4dd1845b6952c3985e179fbea48181ffb8907
2014-10-10 06:46:12 +00:00
$this->reduceAllowedModules(
ResourceLoader namespace (attempt 2) Move ResourceLoader classes to their own namespace. Strip the "ResourceLoader" prefix from all except ResourceLoader itself. Move the tests by analogy. I used a namespace alias "RL" in some callers since RL\Module is less ambiguous at the call site than just "Module". I did not address DependencyStore which continues to have a non-standard location and namespace. Revert of a241d83e0a6dabedf. Bug: T308718 Change-Id: Id08a220e1d6085e2b33f3f6c9d0e3935a4204659
2022-05-06 09:09:56 +00:00
RL\Module::TYPE_SCRIPTS,
RL\Module::ORIGIN_CORE_INDIVIDUAL
Make allowing site-wide styles on restricted special pages a config option This mostly reverts commit 614d7e5c274d927f99bfc52ac3a1e6c7e5902408. Many wikis use MediaWiki:Common.css and associated pages to create a custom "theme" for their wiki, which would no longer load on login or preference pages, creating an inconsistent UI. This re-adds the difference in module origin for different types (styles, scripts, etc.), and now OutputPage::disallowUserJs() checks the value of the "AllowSiteCSSOnRestrictedPages" config setting to determine whether to allow site-wide CSS styles or not. By default this feature is disabled to be secure by default. Bug: 71621 Change-Id: I1bf4dd1845b6952c3985e179fbea48181ffb8907
2014-10-10 06:46:12 +00:00
);
includes: Replace implicit Bugzilla bug numbers with Phab ones It's unreasonable to expect newbies to know that "bug 12345" means "Task T14345" except where it doesn't, so let's just standardise on the real numbers. Change-Id: I6f59febaf8fc96e80f8cfc11f4356283f461142a
2017-02-20 22:44:19 +00:00
// Site-wide styles are controlled by a config setting, see T73621
Make allowing site-wide styles on restricted special pages a config option This mostly reverts commit 614d7e5c274d927f99bfc52ac3a1e6c7e5902408. Many wikis use MediaWiki:Common.css and associated pages to create a custom "theme" for their wiki, which would no longer load on login or preference pages, creating an inconsistent UI. This re-adds the difference in module origin for different types (styles, scripts, etc.), and now OutputPage::disallowUserJs() checks the value of the "AllowSiteCSSOnRestrictedPages" config setting to determine whether to allow site-wide CSS styles or not. By default this feature is disabled to be secure by default. Bug: 71621 Change-Id: I1bf4dd1845b6952c3985e179fbea48181ffb8907
2014-10-10 06:46:12 +00:00
// for background on why. User styles are never allowed.
Use name constants to access config settings. Use name constants instead of string literals in calls to Config::get and ServiceOptions::get, when referring to core configuration variables. This protects against typos and makes the decumentation and schema declaration of the config settings discoverable. This is the first batch, only touching files directly under /includes/ Change-Id: I7252e636c7c86d950d9257b33491af492c6dd5eb
2022-04-01 15:58:32 +00:00
if ( $this->getConfig()->get( MainConfigNames::AllowSiteCSSOnRestrictedPages ) ) {
ResourceLoader namespace (attempt 2) Move ResourceLoader classes to their own namespace. Strip the "ResourceLoader" prefix from all except ResourceLoader itself. Move the tests by analogy. I used a namespace alias "RL" in some callers since RL\Module is less ambiguous at the call site than just "Module". I did not address DependencyStore which continues to have a non-standard location and namespace. Revert of a241d83e0a6dabedf. Bug: T308718 Change-Id: Id08a220e1d6085e2b33f3f6c9d0e3935a4204659
2022-05-06 09:09:56 +00:00
$styleOrigin = RL\Module::ORIGIN_USER_SITEWIDE;
Make allowing site-wide styles on restricted special pages a config option This mostly reverts commit 614d7e5c274d927f99bfc52ac3a1e6c7e5902408. Many wikis use MediaWiki:Common.css and associated pages to create a custom "theme" for their wiki, which would no longer load on login or preference pages, creating an inconsistent UI. This re-adds the difference in module origin for different types (styles, scripts, etc.), and now OutputPage::disallowUserJs() checks the value of the "AllowSiteCSSOnRestrictedPages" config setting to determine whether to allow site-wide CSS styles or not. By default this feature is disabled to be secure by default. Bug: 71621 Change-Id: I1bf4dd1845b6952c3985e179fbea48181ffb8907
2014-10-10 06:46:12 +00:00
} else {
ResourceLoader namespace (attempt 2) Move ResourceLoader classes to their own namespace. Strip the "ResourceLoader" prefix from all except ResourceLoader itself. Move the tests by analogy. I used a namespace alias "RL" in some callers since RL\Module is less ambiguous at the call site than just "Module". I did not address DependencyStore which continues to have a non-standard location and namespace. Revert of a241d83e0a6dabedf. Bug: T308718 Change-Id: Id08a220e1d6085e2b33f3f6c9d0e3935a4204659
2022-05-06 09:09:56 +00:00
$styleOrigin = RL\Module::ORIGIN_CORE_INDIVIDUAL;
Make allowing site-wide styles on restricted special pages a config option This mostly reverts commit 614d7e5c274d927f99bfc52ac3a1e6c7e5902408. Many wikis use MediaWiki:Common.css and associated pages to create a custom "theme" for their wiki, which would no longer load on login or preference pages, creating an inconsistent UI. This re-adds the difference in module origin for different types (styles, scripts, etc.), and now OutputPage::disallowUserJs() checks the value of the "AllowSiteCSSOnRestrictedPages" config setting to determine whether to allow site-wide CSS styles or not. By default this feature is disabled to be secure by default. Bug: 71621 Change-Id: I1bf4dd1845b6952c3985e179fbea48181ffb8907
2014-10-10 06:46:12 +00:00
}
$this->reduceAllowedModules(
ResourceLoader namespace (attempt 2) Move ResourceLoader classes to their own namespace. Strip the "ResourceLoader" prefix from all except ResourceLoader itself. Move the tests by analogy. I used a namespace alias "RL" in some callers since RL\Module is less ambiguous at the call site than just "Module". I did not address DependencyStore which continues to have a non-standard location and namespace. Revert of a241d83e0a6dabedf. Bug: T308718 Change-Id: Id08a220e1d6085e2b33f3f6c9d0e3935a4204659
2022-05-06 09:09:56 +00:00
RL\Module::TYPE_STYLES,
Make allowing site-wide styles on restricted special pages a config option This mostly reverts commit 614d7e5c274d927f99bfc52ac3a1e6c7e5902408. Many wikis use MediaWiki:Common.css and associated pages to create a custom "theme" for their wiki, which would no longer load on login or preference pages, creating an inconsistent UI. This re-adds the difference in module origin for different types (styles, scripts, etc.), and now OutputPage::disallowUserJs() checks the value of the "AllowSiteCSSOnRestrictedPages" config setting to determine whether to allow site-wide CSS styles or not. By default this feature is disabled to be secure by default. Bug: 71621 Change-Id: I1bf4dd1845b6952c3985e179fbea48181ffb8907
2014-10-10 06:46:12 +00:00
$styleOrigin
);
Step two in fixing OutputPage's documentation, more or less the same as r61647
2010-01-29 21:32:45 +00:00
}
Follow-up r64670 (bug22929): cleaner implementation of security for script (and potentially CSS) files. ResourceLoader *already* knows where each module has come from, so all we need to do is filter them in OutputPage according to the desired level of 'trustworthiness'. TODO: * Are there instances where we might want to restrict CSS as well as JS? * Would a $wg config option and/or user preference and/or index.php GET parameter to limit inclusion be useful? * Can we deprecate any of the existing $wg config options? * What's going on with the duplicated code between OutputPage and SkinTemplate?
2011-02-04 16:39:17 +00:00
/**
Make allowing site-wide styles on restricted special pages a config option This mostly reverts commit 614d7e5c274d927f99bfc52ac3a1e6c7e5902408. Many wikis use MediaWiki:Common.css and associated pages to create a custom "theme" for their wiki, which would no longer load on login or preference pages, creating an inconsistent UI. This re-adds the difference in module origin for different types (styles, scripts, etc.), and now OutputPage::disallowUserJs() checks the value of the "AllowSiteCSSOnRestrictedPages" config setting to determine whether to allow site-wide CSS styles or not. By default this feature is disabled to be secure by default. Bug: 71621 Change-Id: I1bf4dd1845b6952c3985e179fbea48181ffb8907
2014-10-10 06:46:12 +00:00
* Show what level of JavaScript / CSS untrustworthiness is allowed on this page
ResourceLoader namespace (attempt 2) Move ResourceLoader classes to their own namespace. Strip the "ResourceLoader" prefix from all except ResourceLoader itself. Move the tests by analogy. I used a namespace alias "RL" in some callers since RL\Module is less ambiguous at the call site than just "Module". I did not address DependencyStore which continues to have a non-standard location and namespace. Revert of a241d83e0a6dabedf. Bug: T308718 Change-Id: Id08a220e1d6085e2b33f3f6c9d0e3935a4204659
2022-05-06 09:09:56 +00:00
* @see RL\Module::$origin
* @param string $type RL\Module TYPE_ constant
* @return int Module ORIGIN_ class constant
Follow-up r64670 (bug22929): cleaner implementation of security for script (and potentially CSS) files. ResourceLoader *already* knows where each module has come from, so all we need to do is filter them in OutputPage according to the desired level of 'trustworthiness'. TODO: * Are there instances where we might want to restrict CSS as well as JS? * Would a $wg config option and/or user preference and/or index.php GET parameter to limit inclusion be useful? * Can we deprecate any of the existing $wg config options? * What's going on with the duplicated code between OutputPage and SkinTemplate?
2011-02-04 16:39:17 +00:00
*/
Make allowing site-wide styles on restricted special pages a config option This mostly reverts commit 614d7e5c274d927f99bfc52ac3a1e6c7e5902408. Many wikis use MediaWiki:Common.css and associated pages to create a custom "theme" for their wiki, which would no longer load on login or preference pages, creating an inconsistent UI. This re-adds the difference in module origin for different types (styles, scripts, etc.), and now OutputPage::disallowUserJs() checks the value of the "AllowSiteCSSOnRestrictedPages" config setting to determine whether to allow site-wide CSS styles or not. By default this feature is disabled to be secure by default. Bug: 71621 Change-Id: I1bf4dd1845b6952c3985e179fbea48181ffb8907
2014-10-10 06:46:12 +00:00
public function getAllowedModules( $type ) {
ResourceLoader namespace (attempt 2) Move ResourceLoader classes to their own namespace. Strip the "ResourceLoader" prefix from all except ResourceLoader itself. Move the tests by analogy. I used a namespace alias "RL" in some callers since RL\Module is less ambiguous at the call site than just "Module". I did not address DependencyStore which continues to have a non-standard location and namespace. Revert of a241d83e0a6dabedf. Bug: T308718 Change-Id: Id08a220e1d6085e2b33f3f6c9d0e3935a4204659
2022-05-06 09:09:56 +00:00
if ( $type == RL\Module::TYPE_COMBINED ) {
Make allowing site-wide styles on restricted special pages a config option This mostly reverts commit 614d7e5c274d927f99bfc52ac3a1e6c7e5902408. Many wikis use MediaWiki:Common.css and associated pages to create a custom "theme" for their wiki, which would no longer load on login or preference pages, creating an inconsistent UI. This re-adds the difference in module origin for different types (styles, scripts, etc.), and now OutputPage::disallowUserJs() checks the value of the "AllowSiteCSSOnRestrictedPages" config setting to determine whether to allow site-wide CSS styles or not. By default this feature is disabled to be secure by default. Bug: 71621 Change-Id: I1bf4dd1845b6952c3985e179fbea48181ffb8907
2014-10-10 06:46:12 +00:00
return min( array_values( $this->mAllowedModules ) );
} else {
ResourceLoader namespace (attempt 2) Move ResourceLoader classes to their own namespace. Strip the "ResourceLoader" prefix from all except ResourceLoader itself. Move the tests by analogy. I used a namespace alias "RL" in some callers since RL\Module is less ambiguous at the call site than just "Module". I did not address DependencyStore which continues to have a non-standard location and namespace. Revert of a241d83e0a6dabedf. Bug: T308718 Change-Id: Id08a220e1d6085e2b33f3f6c9d0e3935a4204659
2022-05-06 09:09:56 +00:00
return $this->mAllowedModules[$type] ?? RL\Module::ORIGIN_ALL;
Make allowing site-wide styles on restricted special pages a config option This mostly reverts commit 614d7e5c274d927f99bfc52ac3a1e6c7e5902408. Many wikis use MediaWiki:Common.css and associated pages to create a custom "theme" for their wiki, which would no longer load on login or preference pages, creating an inconsistent UI. This re-adds the difference in module origin for different types (styles, scripts, etc.), and now OutputPage::disallowUserJs() checks the value of the "AllowSiteCSSOnRestrictedPages" config setting to determine whether to allow site-wide CSS styles or not. By default this feature is disabled to be secure by default. Bug: 71621 Change-Id: I1bf4dd1845b6952c3985e179fbea48181ffb8907
2014-10-10 06:46:12 +00:00
}
Follow-up r64670 (bug22929): cleaner implementation of security for script (and potentially CSS) files. ResourceLoader *already* knows where each module has come from, so all we need to do is filter them in OutputPage according to the desired level of 'trustworthiness'. TODO: * Are there instances where we might want to restrict CSS as well as JS? * Would a $wg config option and/or user preference and/or index.php GET parameter to limit inclusion be useful? * Can we deprecate any of the existing $wg config options? * What's going on with the duplicated code between OutputPage and SkinTemplate?
2011-02-04 16:39:17 +00:00
}
SECURITY: OutputPage: Remove separation of css and js module allowance * No longer segment module origin allowance by an "only=" content type. Both can be sensitive security-wise and there's no valid use case for allowing CSS anywhere you want to disallow JS. Both can significantly impact the user interface and cause unintended actions to be taken on the user's behalf, or desired actions to be made practically impossible. * While at it, also remove the ability to set the module allowance directly. The reduceAllowedModuleOrigin method is all we need. I couldn't find usage or mention of setAllowedModules() in mediawiki-core nor in any other Wikimedia-hosted repository. Bug: 70672 Change-Id: I308e794daca27a9380c67be350f8ab51f9c2de34
2014-09-11 23:14:16 +00:00
/**
* Limit the highest level of CSS/JS untrustworthiness allowed.
*
* If passed the same or a higher level than the current level of untrustworthiness set, the
* level will remain unchanged.
*
Make allowing site-wide styles on restricted special pages a config option This mostly reverts commit 614d7e5c274d927f99bfc52ac3a1e6c7e5902408. Many wikis use MediaWiki:Common.css and associated pages to create a custom "theme" for their wiki, which would no longer load on login or preference pages, creating an inconsistent UI. This re-adds the difference in module origin for different types (styles, scripts, etc.), and now OutputPage::disallowUserJs() checks the value of the "AllowSiteCSSOnRestrictedPages" config setting to determine whether to allow site-wide CSS styles or not. By default this feature is disabled to be secure by default. Bug: 71621 Change-Id: I1bf4dd1845b6952c3985e179fbea48181ffb8907
2014-10-10 06:46:12 +00:00
* @param string $type
ResourceLoader namespace (attempt 2) Move ResourceLoader classes to their own namespace. Strip the "ResourceLoader" prefix from all except ResourceLoader itself. Move the tests by analogy. I used a namespace alias "RL" in some callers since RL\Module is less ambiguous at the call site than just "Module". I did not address DependencyStore which continues to have a non-standard location and namespace. Revert of a241d83e0a6dabedf. Bug: T308718 Change-Id: Id08a220e1d6085e2b33f3f6c9d0e3935a4204659
2022-05-06 09:09:56 +00:00
* @param int $level RL\Module class constant
SECURITY: OutputPage: Remove separation of css and js module allowance * No longer segment module origin allowance by an "only=" content type. Both can be sensitive security-wise and there's no valid use case for allowing CSS anywhere you want to disallow JS. Both can significantly impact the user interface and cause unintended actions to be taken on the user's behalf, or desired actions to be made practically impossible. * While at it, also remove the ability to set the module allowance directly. The reduceAllowedModuleOrigin method is all we need. I couldn't find usage or mention of setAllowedModules() in mediawiki-core nor in any other Wikimedia-hosted repository. Bug: 70672 Change-Id: I308e794daca27a9380c67be350f8ab51f9c2de34
2014-09-11 23:14:16 +00:00
*/
Make allowing site-wide styles on restricted special pages a config option This mostly reverts commit 614d7e5c274d927f99bfc52ac3a1e6c7e5902408. Many wikis use MediaWiki:Common.css and associated pages to create a custom "theme" for their wiki, which would no longer load on login or preference pages, creating an inconsistent UI. This re-adds the difference in module origin for different types (styles, scripts, etc.), and now OutputPage::disallowUserJs() checks the value of the "AllowSiteCSSOnRestrictedPages" config setting to determine whether to allow site-wide CSS styles or not. By default this feature is disabled to be secure by default. Bug: 71621 Change-Id: I1bf4dd1845b6952c3985e179fbea48181ffb8907
2014-10-10 06:46:12 +00:00
public function reduceAllowedModules( $type, $level ) {
$this->mAllowedModules[$type] = min( $this->getAllowedModules( $type ), $level );
Step two in fixing OutputPage's documentation, more or less the same as r61647
2010-01-29 21:32:45 +00:00
}
/**
* Prepend $text to the body HTML
*
Fixed @param tags to conform with Doxygen format. Doxygen expects parameter types to come before the parameter name in @param tags. Used a quick regex to switch everything around where possible. This only fixes cases where a primitve variable (or a primitive followed by other types) is the variable type. Other cases will need to be fixed manually. Change-Id: Ic59fd20856eb0489d70f3469a56ebce0efb3db13
2013-03-11 17:15:01 +00:00
* @param string $text HTML
Step two in fixing OutputPage's documentation, more or less the same as r61647
2010-01-29 21:32:45 +00:00
*/
public function prependHTML( $text ) {
$this->mBodytext = $text . $this->mBodytext;
}
/**
* Append $text to the body HTML
*
Fixed @param tags to conform with Doxygen format. Doxygen expects parameter types to come before the parameter name in @param tags. Used a quick regex to switch everything around where possible. This only fixes cases where a primitve variable (or a primitive followed by other types) is the variable type. Other cases will need to be fixed manually. Change-Id: Ic59fd20856eb0489d70f3469a56ebce0efb3db13
2013-03-11 17:15:01 +00:00
* @param string $text HTML
Step two in fixing OutputPage's documentation, more or less the same as r61647
2010-01-29 21:32:45 +00:00
*/
public function addHTML( $text ) {
$this->mBodytext .= $text;
}
Fixup some assignments in conditionals Add/normalise some more return statements
2011-10-16 03:27:12 +00:00
added shortcut method for commonly used method combination
2011-10-10 21:05:02 +00:00
/**
* Shortcut for adding an Html::element via addHTML.
Fixup some assignments in conditionals Add/normalise some more return statements
2011-10-16 03:27:12 +00:00
*
added shortcut method for commonly used method combination
2011-10-10 21:05:02 +00:00
* @since 1.19
Fixup some assignments in conditionals Add/normalise some more return statements
2011-10-16 03:27:12 +00:00
*
Fixed some @params documentation Swapped some "$var type" to "type $var" or added missing types before the $var. Changed some other types to match the more common spelling. Makes beginning of some text in captial. Change-Id: I64e8cfe478cb0ba438f40b0631d6e9049cdab567
2014-04-10 18:50:10 +00:00
* @param string $element
* @param array $attribs
* @param string $contents
added shortcut method for commonly used method combination
2011-10-10 21:05:02 +00:00
*/
Convert all array() syntax to [] Per wikitech-l consensus: https://lists.wikimedia.org/pipermail/wikitech-l/2016-February/084821.html Notes: * Disabled CallTimePassByReference due to false positives (T127163) Change-Id: I2c8ce713ce6600a0bb7bf67537c87044c7a45c4b
2016-02-17 09:09:32 +00:00
public function addElement( $element, array $attribs = [], $contents = '' ) {
added shortcut method for commonly used method combination
2011-10-10 21:05:02 +00:00
$this->addHTML( Html::element( $element, $attribs, $contents ) );
}
Step two in fixing OutputPage's documentation, more or less the same as r61647
2010-01-29 21:32:45 +00:00
/**
* Clear the body HTML
*/
public function clearHTML() {
$this->mBodytext = '';
}
/**
* Get the body HTML
*
Removed unneeded spaces and colons in @param and friends Also swapped some "$var type" to "type $var" or added missing types before the $var. Changed some other types to match the more common spelling. Makes beginning of some text in captial. Change-Id: Ic36c8c7820a6c2d603f1138130670c6bf6a1ca59
2014-04-08 15:29:17 +00:00
* @return string HTML
Step two in fixing OutputPage's documentation, more or less the same as r61647
2010-01-29 21:32:45 +00:00
*/
public function getHTML() {
return $this->mBodytext;
}
/**
* Get/set the ParserOptions object to use for wikitext parsing
*
Fixed some @params documentation Swapped some "$var type" to "type $var" or added missing types before the $var. Changed some other types to match the more common spelling. Makes beginning of some text in captial. Change-Id: I64e8cfe478cb0ba438f40b0631d6e9049cdab567
2014-04-10 18:50:10 +00:00
* @return ParserOptions
Step two in fixing OutputPage's documentation, more or less the same as r61647
2010-01-29 21:32:45 +00:00
*/
Remove $options parameter from OutputPage::parserOptions No callers passing a parameter https://codesearch.wmflabs.org/search/?q=-%3EparserOptions%5C(%5B%5E%5C)%5D&i=nope&files=&repos= and no dynamic calls AFAICS. Change-Id: I526b1e8e310c1ad3d14402b67e6c967923f3609f
2020-03-08 13:22:13 +00:00
public function parserOptions() {
Merged localisation-work branch: * Made lines from initialiseMessages() appear as list items during installation * Moved the bulk of the localisation data from the Language*.php files to the Messages*.php files. Deleted most of the Languages*.php files. * Introduced "stub global" framework to provide deferred initialisation of core modules. * Removed placeholder values for $wgTitle and $wgArticle, these variables will now be null during the initialisation process, until they are set by index.php or another entry point. * Added DBA cache type, for BDB-style caches. * Removed custom date format functions, replacing them with a format string in the style of PHP's date(). Used string identifiers instead of integer identifiers, in both the language files and user preferences. Migration should be transparent in most cases. * Simplified the initialisation API for LoadBalancer objects. * Removed the broken altencoding feature. * Moved default user options and toggles from Language to User. Language objects are still able to define default preference overrides and extra user toggles, via a slightly different interface. * Don't include the date option in the parser cache rendering hash unless $wgUseDynamicDates is enabled. * Merged LanguageUtf8 with Language. Removed LanguageUtf8.php. * Removed inclusion of language files from the bottom of Language.php. This is now consistently done from Language::factory(). * Add the name of the executing maintenance script to the debug log. Start the profiler during maintenance scripts. * Added "serialized" directory, for storing precompiled data in serialized form.
2006-07-26 07:15:39 +00:00
if ( !$this->mParserOptions ) {
OutputPage: Use $this->getUser() Use $this->getUser() instead of $this->getContext()->getUser(). ContextSource::getUser is a shortcut for $this->getContext()->getUser(). Change-Id: I7d19116b634500cd1a435bebe235c7562889f7bb
2019-03-11 16:50:46 +00:00
if ( !$this->getUser()->isSafeToLoad() ) {
Reduce some references to $wgUser Should be a no-op Bug: T243708 Change-Id: I7b204cdc629c6c33e6a84b5c7a4e0e7ac321f19d
2020-10-09 19:43:28 +00:00
// Context user isn't unstubbable yet, so don't try to get a
OutputPage: Don't get a ParserOptions for $wgUser before the end of Setup.php Until Setup.php completes, we can't access the "current" user and trying to do so logs a warning. So don't try. Bug: T124367 Change-Id: I1acc82760c368a24448f3c90e268b24414e43a30
2016-02-03 20:18:52 +00:00
// ParserOptions for it. And don't cache this ParserOptions
// either.
$po = ParserOptions::newFromAnon();
SECURITY: Disable <html> tag on system messages despite $wgRawHtml = true; System messages may take parameters from untrusted sources. This may include taking parameters from urls given by unauthenticated users even if the wiki is a read-only wiki. Allowing <html> tags in such a context seems like an accident waiting to happen. Bug: T156184 Change-Id: I661f482986d319cf41da1d3e7b20a0f028a42e90
2017-02-06 03:00:39 +00:00
$po->setAllowUnsafeRawHtml( false );
OutputPage: Don't get a ParserOptions for $wgUser before the end of Setup.php Until Setup.php completes, we can't access the "current" user and trying to do so logs a warning. So don't try. Bug: T124367 Change-Id: I1acc82760c368a24448f3c90e268b24414e43a30
2016-02-03 20:18:52 +00:00
return $po;
}
* Changed ParserOptions to store a Language object instead of only a string, avoids object -> string -> object conversion * ParserOptions::getUserLang() will still return a string for compatibility, added ParserOptions::getUserLangObj() to get the object * Added ParserOptions::newFromUserAndLang() and ParserOptions::newFromContext() to easily get a ParserOptions object when a context is available or when someone wants to force the language * Updated OutputPage and Preferences to use newFromContext() and WikiPage to use newFromUserAndLang() * ParserOptions::setUserLang() still accepts either a string or a Language object, but changed the calls to pass an object instead of a string * Changed Parser::getFunctionLang() to return the Language object from ParserOptions when parsing interface messages rather than $wgLang directly and updated the documentation to say that $wgLang should not be used directly (as $wgUser, $wgTitle and $wgRequest)
2011-10-19 14:16:01 +00:00
$this->mParserOptions = ParserOptions::newFromContext( $this->getContext() );
SECURITY: Disable <html> tag on system messages despite $wgRawHtml = true; System messages may take parameters from untrusted sources. This may include taking parameters from urls given by unauthenticated users even if the wiki is a read-only wiki. Allowing <html> tags in such a context seems like an accident waiting to happen. Bug: T156184 Change-Id: I661f482986d319cf41da1d3e7b20a0f028a42e90
2017-02-06 03:00:39 +00:00
$this->mParserOptions->setAllowUnsafeRawHtml( false );
Merged localisation-work branch: * Made lines from initialiseMessages() appear as list items during installation * Moved the bulk of the localisation data from the Language*.php files to the Messages*.php files. Deleted most of the Languages*.php files. * Introduced "stub global" framework to provide deferred initialisation of core modules. * Removed placeholder values for $wgTitle and $wgArticle, these variables will now be null during the initialisation process, until they are set by index.php or another entry point. * Added DBA cache type, for BDB-style caches. * Removed custom date format functions, replacing them with a format string in the style of PHP's date(). Used string identifiers instead of integer identifiers, in both the language files and user preferences. Migration should be transparent in most cases. * Simplified the initialisation API for LoadBalancer objects. * Removed the broken altencoding feature. * Moved default user options and toggles from Language to User. Language objects are still able to define default preference overrides and extra user toggles, via a slightly different interface. * Don't include the date option in the parser cache rendering hash unless $wgUseDynamicDates is enabled. * Merged LanguageUtf8 with Language. Removed LanguageUtf8.php. * Removed inclusion of language files from the bottom of Language.php. This is now consistently done from Language::factory(). * Add the name of the executing maintenance script to the debug log. Start the profiler during maintenance scripts. * Added "serialized" directory, for storing precompiled data in serialized form.
2006-07-26 07:15:39 +00:00
}
OutputPage: Don't get a ParserOptions for $wgUser before the end of Setup.php Until Setup.php completes, we can't access the "current" user and trying to do so logs a warning. So don't try. Bug: T124367 Change-Id: I1acc82760c368a24448f3c90e268b24414e43a30
2016-02-03 20:18:52 +00:00
Remove $options parameter from OutputPage::parserOptions No callers passing a parameter https://codesearch.wmflabs.org/search/?q=-%3EparserOptions%5C(%5B%5E%5C)%5D&i=nope&files=&repos= and no dynamic calls AFAICS. Change-Id: I526b1e8e310c1ad3d14402b67e6c967923f3609f
2020-03-08 13:22:13 +00:00
return $this->mParserOptions;
Parser improvements: global variable destruction
2004-02-29 08:43:29 +00:00
}
* s~\t+$~~
2006-01-07 13:31:29 +00:00
* (bug 3979) Save correct {{REVISIONID}} into parser cache on edit * Substitute {{REVISIONID}} correctly in diff display Removed hook for revision-id override, avar will change his extension to use the new system
2005-11-27 06:04:41 +00:00
/**
* Set the revision ID which will be seen by the wiki text parser
* for things such as embedded {{REVISIONID}} variable use.
Step two in fixing OutputPage's documentation, more or less the same as r61647
2010-01-29 21:32:45 +00:00
*
Improve OutputPage test coverage more Improved locally from 26.81% to 37.18%. Depends-On: Ib564624c644ee6620ac06872f5684831acaaaadb Change-Id: Icdc0288c04b8c4ba841f9fbb3e05a0cdc8a20fa5
2018-07-25 18:41:42 +00:00
* @param int|null $revid A positive integer, or null
Removed unneeded spaces and colons in @param and friends Also swapped some "$var type" to "type $var" or added missing types before the $var. Changed some other types to match the more common spelling. Makes beginning of some text in captial. Change-Id: Ic36c8c7820a6c2d603f1138130670c6bf6a1ca59
2014-04-08 15:29:17 +00:00
* @return mixed Previous value
* (bug 3979) Save correct {{REVISIONID}} into parser cache on edit * Substitute {{REVISIONID}} correctly in diff display Removed hook for revision-id override, avar will change his extension to use the new system
2005-11-27 06:04:41 +00:00
*/
Add some documentation to OutputPage.php and specify public/private for most of its member functions (practically all public anyway). And replace one gratuitous use of strcmp with good old ===.
2006-11-07 05:37:31 +00:00
public function setRevisionId( $revid ) {
Coding style: Auto-fix MediaWiki.Usage.IsNull.IsNull Change-Id: I90cfe8366c0245c9c67e598d17800684897a4e27
2020-01-09 23:48:34 +00:00
$val = $revid === null ? null : intval( $revid );
Improve OutputPage test coverage more Improved locally from 26.81% to 37.18%. Depends-On: Ib564624c644ee6620ac06872f5684831acaaaadb Change-Id: Icdc0288c04b8c4ba841f9fbb3e05a0cdc8a20fa5
2018-07-25 18:41:42 +00:00
return wfSetVar( $this->mRevisionId, $val, true );
* (bug 3979) Save correct {{REVISIONID}} into parser cache on edit * Substitute {{REVISIONID}} correctly in diff display Removed hook for revision-id override, avar will change his extension to use the new system
2005-11-27 06:04:41 +00:00
}
here it is ... the upload-api, script-server, js2 (javascript phase2) branch merge 1st attempt. Here is a short overview of changes and associated default configuration variables (most everything is off by default) also see ~soon to be updated~: http://www.mediawiki.org/wiki/Media_Projects_Overview = Upload Improvements = ==Upload API == * Based on the early work of Bryan Tong and others it adds the upload option to the api. * We rewrite Special:Upload page to include use the new refactoring * Added in token checks in both the SpecialUpload.php page so avoids DOS / xss copy-by-url JavaScript based cross site POST file submissions == Copy by URL== $wgAllowCopyUploads = false; * http class rewrite includes a new http background download see: includes/HttpFunctions.php * spins off a php process that calls: maintenance/http_session_download.php * pushes updates to the session and gives the user a progress bar on http copy uploads from other server progress (using js2 upload interface) (if not using the js2 upload interface it does the request in-place but the download is limited to the php ini timeout time) == Firefogg == * Firefogg enables resumable upload by chunks * progress indicators and conditional invokation (js2 system) * and of-course client side transcoding. = Script Server = $wgEnableScriptLoader = false; * off by default if $wgEnableScriptLoader is turned on script files are grouped, gziped, cached etc. for more info see: http://www.mediawiki.org/wiki/Extension:ScriptLoader * Includes some early skin js include fixes (skin/script system still lots of love) * Includes a "javascript class autoloader" this is packaged into mwEmbed so that the mwEmbed library can work in stand alone mode (while retaining localization and script serving) (one such application is the make page for firefogg.org : http://www.firefogg.org/make/index.html ) * The file that contains the autojavascript loading classes is: js2/php/jsAutoloadLocalClasses.php * One can use this auto class loading dependency system with extensions and add-ons but I need to better document that. = js2 system / mwEmbed= $wgEnableJS2system = false * includes initial rewrite towards more jquery based javascript code * especially for the Special:Upload page. * Also the edit page include support for the "add-media-wizard" * includes dependency loader for javascript that optionally takes advantage of the script-loader * remote embedding of javascript interfaces (like embedding video, or commons media searching) * $wgDebugJavaScript = false; .. .this variable lets you always get "always fresh javascript". When used with the script-loader it does not minify the script-loader output. = mwEmbed = * Will commit a separate patch to oggHandler that conditionally outputs <video tag> to use the new javascript video player. ** mv_embed player includes: play-head, volume control, remote embedding, oggz-chop support across plugins. * add-media-wizard adds easy inserts of media to pages (with import) == jQuery== * we include a base install of jQuery, jQuery ui and some plugins. * all the javascript classes are in the scriptloader so its easy to load any set of jquery ui components that you may need using the script-server. You get a callback so you can then execute js with dependencies loaded. == other stuff == there is a bit more code in js2 that pertains to sequence editing, timed text display and basic image editing. We include a base import of pixastic-lib & pixastic-editor... will work with the pixastic developer to try and ensure upstream compatibility on our usage of the library for in-browser photo and sequence manipulation.
2009-07-14 23:52:14 +00:00
Step two in fixing OutputPage's documentation, more or less the same as r61647
2010-01-29 21:32:45 +00:00
/**
* Added getFileVersion()/setFileVersion() functions to OutputPage * Removed getDisplayedFile() from FlaggedPage and simplified getFile() * Cleaned up getIncludeParams() to just do formatting * Made template/file IDs mandatory for RevisionReviewFormUI
2011-07-05 05:23:26 +00:00
* Get the displayed revision ID
Step two in fixing OutputPage's documentation, more or less the same as r61647
2010-01-29 21:32:45 +00:00
*
Avoid null query in Skin::lastModified For action=history there is a query: SELECT rev_timestamp FROM `revision` WHERE rev_id IS NULL LIMIT 1 to generate the footer links Change-Id: I8b0f30f874eb73544693dff9d12d45585bd9a110
2020-08-18 11:41:53 +00:00
* @return int|null
Step two in fixing OutputPage's documentation, more or less the same as r61647
2010-01-29 21:32:45 +00:00
*/
Add getRevisionId()
2008-08-11 13:23:45 +00:00
public function getRevisionId() {
return $this->mRevisionId;
}
Parser improvements: global variable destruction
2004-02-29 08:43:29 +00:00
OutputPage: Provide consistent info about whether the revision shown is current Previously, OutputPage::isRevisionCurrent() would check a replica database, potentially different from the database used to fetch the revision being displayed, according to OutputPage::getRevisionId(). Now the data for both functions is provided in the same way. Bug: T314684 Change-Id: I266d643d1eed931df346889f43d72d42e5f4a3ba
2022-08-06 14:40:47 +00:00
/**
* Set whether the revision displayed (as set in ::setRevisionId())
* is the latest revision of the page.
*
* @param bool $isCurrent
*/
public function setRevisionIsCurrent( bool $isCurrent ): void {
$this->mRevisionIsCurrent = $isCurrent;
}
Deprecate skin functions that are not skin responsibilities Skin shouldn't be responsible for providing requested revisionId nor if that revision is the current revision. The OutputPage object has all required information (both the currentRevisionID and the current Title object). Change-Id: I2dbae4c6968a2b3b3cea3e09977e9579609b4cc5
2019-08-21 12:10:09 +00:00
/**
* Whether the revision displayed is the latest revision of the page
*
* @since 1.34
* @return bool
*/
OutputPage: Provide consistent info about whether the revision shown is current Previously, OutputPage::isRevisionCurrent() would check a replica database, potentially different from the database used to fetch the revision being displayed, according to OutputPage::getRevisionId(). Now the data for both functions is provided in the same way. Bug: T314684 Change-Id: I266d643d1eed931df346889f43d72d42e5f4a3ba
2022-08-06 14:40:47 +00:00
public function isRevisionCurrent(): bool {
return $this->mRevisionId == 0 || (
$this->mRevisionIsCurrent ?? (
$this->mRevisionId == $this->getTitle()->getLatestRevID()
)
);
Deprecate skin functions that are not skin responsibilities Skin shouldn't be responsible for providing requested revisionId nor if that revision is the current revision. The OutputPage object has all required information (both the currentRevisionID and the current Title object). Change-Id: I2dbae4c6968a2b3b3cea3e09977e9579609b4cc5
2019-08-21 12:10:09 +00:00
}
* Added revision's timestamp to OutputPage along with revision ID; avoid a DB hit in Skin::lastModified() when showing parser cache's content. This changed with the removal of $wgArticle in Skin since now it's a different WikiPage object and thus WikiPage::setTimetstamp() call is useless (but still kept). * Added ParserOutput::(get|set)Timestamp() and the $mTimestamp member; avoid messing with isset()
2011-12-10 16:30:40 +00:00
/**
* Set the timestamp of the revision which will be displayed. This is used
* to avoid a extra DB call in Skin::lastModified().
*
Removed unneeded spaces and colons in @param and friends Also swapped some "$var type" to "type $var" or added missing types before the $var. Changed some other types to match the more common spelling. Makes beginning of some text in captial. Change-Id: Ic36c8c7820a6c2d603f1138130670c6bf6a1ca59
2014-04-08 15:29:17 +00:00
* @param string|null $timestamp
* @return mixed Previous value
* Added revision's timestamp to OutputPage along with revision ID; avoid a DB hit in Skin::lastModified() when showing parser cache's content. This changed with the removal of $wgArticle in Skin since now it's a different WikiPage object and thus WikiPage::setTimetstamp() call is useless (but still kept). * Added ParserOutput::(get|set)Timestamp() and the $mTimestamp member; avoid messing with isset()
2011-12-10 16:30:40 +00:00
*/
fix some spacing * Removed spaces around array index * Removed double spaces or added spaces to begin or end of function calls, method signature, conditions or foreachs * Added braces to one-line ifs * Changed multi line conditions to one line conditions * Realigned some arrays Change-Id: Ia04d2a99d663b07101013c2d53b3b2e872fd9cc3
2013-03-24 10:01:51 +00:00
public function setRevisionTimestamp( $timestamp ) {
Improve OutputPage test coverage more Improved locally from 26.81% to 37.18%. Depends-On: Ib564624c644ee6620ac06872f5684831acaaaadb Change-Id: Icdc0288c04b8c4ba841f9fbb3e05a0cdc8a20fa5
2018-07-25 18:41:42 +00:00
return wfSetVar( $this->mRevisionTimestamp, $timestamp, true );
* Added revision's timestamp to OutputPage along with revision ID; avoid a DB hit in Skin::lastModified() when showing parser cache's content. This changed with the removal of $wgArticle in Skin since now it's a different WikiPage object and thus WikiPage::setTimetstamp() call is useless (but still kept). * Added ParserOutput::(get|set)Timestamp() and the $mTimestamp member; avoid messing with isset()
2011-12-10 16:30:40 +00:00
}
/**
* Get the timestamp of displayed revision.
* This will be null if not filled by setRevisionTimestamp().
*
Fixed some @params documentation Swapped some "$var type" to "type $var" or added missing types before the $var. Changed some other types to match the more common spelling. Makes beginning of some text in captial. Change-Id: I64e8cfe478cb0ba438f40b0631d6e9049cdab567
2014-04-10 18:50:10 +00:00
* @return string|null
* Added revision's timestamp to OutputPage along with revision ID; avoid a DB hit in Skin::lastModified() when showing parser cache's content. This changed with the removal of $wgArticle in Skin since now it's a different WikiPage object and thus WikiPage::setTimetstamp() call is useless (but still kept). * Added ParserOutput::(get|set)Timestamp() and the $mTimestamp member; avoid messing with isset()
2011-12-10 16:30:40 +00:00
*/
public function getRevisionTimestamp() {
return $this->mRevisionTimestamp;
}
* Added getFileVersion()/setFileVersion() functions to OutputPage * Removed getDisplayedFile() from FlaggedPage and simplified getFile() * Cleaned up getIncludeParams() to just do formatting * Made template/file IDs mandatory for RevisionReviewFormUI
2011-07-05 05:23:26 +00:00
/**
* Set the displayed file version
*
Improve OutputPage test coverage more Improved locally from 26.81% to 37.18%. Depends-On: Ib564624c644ee6620ac06872f5684831acaaaadb Change-Id: Icdc0288c04b8c4ba841f9fbb3e05a0cdc8a20fa5
2018-07-25 18:41:42 +00:00
* @param File|null $file
Removed unneeded spaces and colons in @param and friends Also swapped some "$var type" to "type $var" or added missing types before the $var. Changed some other types to match the more common spelling. Makes beginning of some text in captial. Change-Id: Ic36c8c7820a6c2d603f1138130670c6bf6a1ca59
2014-04-08 15:29:17 +00:00
* @return mixed Previous value
* Added getFileVersion()/setFileVersion() functions to OutputPage * Removed getDisplayedFile() from FlaggedPage and simplified getFile() * Cleaned up getIncludeParams() to just do formatting * Made template/file IDs mandatory for RevisionReviewFormUI
2011-07-05 05:23:26 +00:00
*/
public function setFileVersion( $file ) {
setFileVersion (r91435) was missing default $val
2011-07-06 17:27:33 +00:00
$val = null;
* Added getFileVersion()/setFileVersion() functions to OutputPage * Removed getDisplayedFile() from FlaggedPage and simplified getFile() * Cleaned up getIncludeParams() to just do formatting * Made template/file IDs mandatory for RevisionReviewFormUI
2011-07-05 05:23:26 +00:00
if ( $file instanceof File && $file->exists() ) {
Convert all array() syntax to [] Per wikitech-l consensus: https://lists.wikimedia.org/pipermail/wikitech-l/2016-February/084821.html Notes: * Disabled CallTimePassByReference due to false positives (T127163) Change-Id: I2c8ce713ce6600a0bb7bf67537c87044c7a45c4b
2016-02-17 09:09:32 +00:00
$val = [ 'time' => $file->getTimestamp(), 'sha1' => $file->getSha1() ];
* Added getFileVersion()/setFileVersion() functions to OutputPage * Removed getDisplayedFile() from FlaggedPage and simplified getFile() * Cleaned up getIncludeParams() to just do formatting * Made template/file IDs mandatory for RevisionReviewFormUI
2011-07-05 05:23:26 +00:00
}
setFileVersion (r91435) was missing default $val
2011-07-06 17:27:33 +00:00
return wfSetVar( $this->mFileVersion, $val, true );
* Added getFileVersion()/setFileVersion() functions to OutputPage * Removed getDisplayedFile() from FlaggedPage and simplified getFile() * Cleaned up getIncludeParams() to just do formatting * Made template/file IDs mandatory for RevisionReviewFormUI
2011-07-05 05:23:26 +00:00
}
/**
* Get the displayed file version
*
Fixed some @params documentation Swapped some "$var type" to "type $var" or added missing types before the $var. Changed some other types to match the more common spelling. Makes beginning of some text in captial. Change-Id: I64e8cfe478cb0ba438f40b0631d6e9049cdab567
2014-04-10 18:50:10 +00:00
* @return array|null ('time' => MW timestamp, 'sha1' => sha1)
* Added getFileVersion()/setFileVersion() functions to OutputPage * Removed getDisplayedFile() from FlaggedPage and simplified getFile() * Cleaned up getIncludeParams() to just do formatting * Made template/file IDs mandatory for RevisionReviewFormUI
2011-07-05 05:23:26 +00:00
*/
public function getFileVersion() {
return $this->mFileVersion;
}
* Removed isset() checks - redundant given Parser::version * For sanity in insertOn(), if the template/file version arrays are null, don't hit the DB - just assume they are empty * Refactored mFlags handling in FlaggedRevision * Use accessors for ParserOutput template/file versions * Added OutputPage accessors for template/file versions * Changed instance of NS_IMAGE -> NS_FILE
2011-04-03 05:53:07 +00:00
/**
* Get the templates used on this page
*
Fixed some @params documentation Swapped some "$var type" to "type $var" or added missing types before the $var. Changed some other types to match the more common spelling. Makes beginning of some text in captial. Change-Id: I64e8cfe478cb0ba438f40b0631d6e9049cdab567
2014-04-10 18:50:10 +00:00
* @return array (namespace => dbKey => revId)
Follow-up changes to r84610: * Cleaned up mImageTimeKeys format * ImageMap: Removed redundant addImage call (makeImage handes this) * ParserFunctions: added time/sha1 to addImage() call * Removed excess ampersands in hooks * Added some function doc comments
2011-04-04 01:22:08 +00:00
* @since 1.18
* Removed isset() checks - redundant given Parser::version * For sanity in insertOn(), if the template/file version arrays are null, don't hit the DB - just assume they are empty * Refactored mFlags handling in FlaggedRevision * Use accessors for ParserOutput template/file versions * Added OutputPage accessors for template/file versions * Changed instance of NS_IMAGE -> NS_FILE
2011-04-03 05:53:07 +00:00
*/
public function getTemplateIds() {
return $this->mTemplateIds;
}
/**
* Get the files used on this page
*
Use [...] instead of array(...) in PHP comments and documentation Change-Id: I0c83783051bf35fe785bc01644eeb2946902b6b2
2017-01-16 15:42:53 +00:00
* @return array [ dbKey => [ 'time' => MW timestamp or null, 'sha1' => sha1 or '' ] ]
Follow-up changes to r84610: * Cleaned up mImageTimeKeys format * ImageMap: Removed redundant addImage call (makeImage handes this) * ParserFunctions: added time/sha1 to addImage() call * Removed excess ampersands in hooks * Added some function doc comments
2011-04-04 01:22:08 +00:00
* @since 1.18
* Removed isset() checks - redundant given Parser::version * For sanity in insertOn(), if the template/file version arrays are null, don't hit the DB - just assume they are empty * Refactored mFlags handling in FlaggedRevision * Use accessors for ParserOutput template/file versions * Added OutputPage accessors for template/file versions * Changed instance of NS_IMAGE -> NS_FILE
2011-04-03 05:53:07 +00:00
*/
FU r84610: Renamed getImageTimeKeys() to getFileSearchOptions()
2011-09-14 19:05:43 +00:00
public function getFileSearchOptions() {
* Removed isset() checks - redundant given Parser::version * For sanity in insertOn(), if the template/file version arrays are null, don't hit the DB - just assume they are empty * Refactored mFlags handling in FlaggedRevision * Use accessors for ParserOutput template/file versions * Added OutputPage accessors for template/file versions * Changed instance of NS_IMAGE -> NS_FILE
2011-04-03 05:53:07 +00:00
return $this->mImageTimeKeys;
}
Deprecate and rename OutputPage::addWikiText* methods Tidy will always be enabled with our future parsers, and it is fast and pure PHP now with the Remex implementation, so deprecate all the untidy variants of 'OutputPage::addWikiText*()' and add new methods which tidy by default. Clarify the content language/interface language distinction while we're at it by adding 'AsInterface' to the name of methods which use the "interface language" by default, and renaming the 'addWikiText*Tidy' methods to 'addWikiTextAsContent'. The 'OutputPage::addWikiTextTitle' method has been deprecated, but it is still used internally as the implementation for the newly-added methods. It is expected that the shared implementation will move in the future to a new private method. Setting the `$tidy` parameter of `OutputPage::addWikiTextTitle` to false is independently deprecated; for backwards-compatibility with old MW releases you may wish to continue to invoke OutputPage::addWikiTextTitle() but set $tidy=true; this will result in the same tidied output that the newly added methods would produce. Bug: T198214 Change-Id: Ia58910164baaca608cea3b24333b7d13ed773339
2018-09-21 16:24:57 +00:00
/**
* Convert wikitext *in the user interface language* to HTML and
* add it to the buffer. The result will not be
* language-converted, as user interface messages are already
* localized into a specific variant. Assumes that the current
* page title will be used if optional $title is not
* provided. Output will be tidy.
*
* @param string $text Wikitext in the user interface language
* @param bool $linestart Is this the start of a line? (Defaults to true)
Remove Title from public interface of OutputPage Bug: T278459 Change-Id: If455682540faf8d42fd8a14e0ef09bac24f72db4
2021-04-08 19:11:06 +00:00
* @param PageReference|null $title Optional title to use; default of `null`
Deprecate and rename OutputPage::addWikiText* methods Tidy will always be enabled with our future parsers, and it is fast and pure PHP now with the Remex implementation, so deprecate all the untidy variants of 'OutputPage::addWikiText*()' and add new methods which tidy by default. Clarify the content language/interface language distinction while we're at it by adding 'AsInterface' to the name of methods which use the "interface language" by default, and renaming the 'addWikiText*Tidy' methods to 'addWikiTextAsContent'. The 'OutputPage::addWikiTextTitle' method has been deprecated, but it is still used internally as the implementation for the newly-added methods. It is expected that the shared implementation will move in the future to a new private method. Setting the `$tidy` parameter of `OutputPage::addWikiTextTitle` to false is independently deprecated; for backwards-compatibility with old MW releases you may wish to continue to invoke OutputPage::addWikiTextTitle() but set $tidy=true; this will result in the same tidied output that the newly added methods would produce. Bug: T198214 Change-Id: Ia58910164baaca608cea3b24333b7d13ed773339
2018-09-21 16:24:57 +00:00
* means use current page title.
* @throws MWException if $title is not provided and OutputPage::getTitle()
* is null
* @since 1.32
*/
public function addWikiTextAsInterface(
Remove Title from public interface of OutputPage Bug: T278459 Change-Id: If455682540faf8d42fd8a14e0ef09bac24f72db4
2021-04-08 19:11:06 +00:00
$text, $linestart = true, PageReference $title = null
Deprecate and rename OutputPage::addWikiText* methods Tidy will always be enabled with our future parsers, and it is fast and pure PHP now with the Remex implementation, so deprecate all the untidy variants of 'OutputPage::addWikiText*()' and add new methods which tidy by default. Clarify the content language/interface language distinction while we're at it by adding 'AsInterface' to the name of methods which use the "interface language" by default, and renaming the 'addWikiText*Tidy' methods to 'addWikiTextAsContent'. The 'OutputPage::addWikiTextTitle' method has been deprecated, but it is still used internally as the implementation for the newly-added methods. It is expected that the shared implementation will move in the future to a new private method. Setting the `$tidy` parameter of `OutputPage::addWikiTextTitle` to false is independently deprecated; for backwards-compatibility with old MW releases you may wish to continue to invoke OutputPage::addWikiTextTitle() but set $tidy=true; this will result in the same tidied output that the newly added methods would produce. Bug: T198214 Change-Id: Ia58910164baaca608cea3b24333b7d13ed773339
2018-09-21 16:24:57 +00:00
) {
Make use of ??= in more places New feature from PHP 7.4 Change-Id: Ifa7a9bc7b2ec415ad7ecb23f4c1776f51f58fd6b
2022-12-16 23:48:27 +00:00
$title ??= $this->getTitle();
Deprecate and rename OutputPage::addWikiText* methods Tidy will always be enabled with our future parsers, and it is fast and pure PHP now with the Remex implementation, so deprecate all the untidy variants of 'OutputPage::addWikiText*()' and add new methods which tidy by default. Clarify the content language/interface language distinction while we're at it by adding 'AsInterface' to the name of methods which use the "interface language" by default, and renaming the 'addWikiText*Tidy' methods to 'addWikiTextAsContent'. The 'OutputPage::addWikiTextTitle' method has been deprecated, but it is still used internally as the implementation for the newly-added methods. It is expected that the shared implementation will move in the future to a new private method. Setting the `$tidy` parameter of `OutputPage::addWikiTextTitle` to false is independently deprecated; for backwards-compatibility with old MW releases you may wish to continue to invoke OutputPage::addWikiTextTitle() but set $tidy=true; this will result in the same tidied output that the newly added methods would produce. Bug: T198214 Change-Id: Ia58910164baaca608cea3b24333b7d13ed773339
2018-09-21 16:24:57 +00:00
if ( !$title ) {
throw new MWException( 'Title is null' );
}
Remove OutputPage::addWikitext()* functions, hard-deprecated in 1.32 This starts cleaning up the programmer-visible API for OutputPage and removed some deprecated untidy parser modes. Change-Id: Ib464b57248f114b68424ec1175d36ad86d1319ad
2019-06-19 15:57:36 +00:00
$this->addWikiTextTitleInternal( $text, $title, $linestart, /*interface*/true );
Deprecate and rename OutputPage::addWikiText* methods Tidy will always be enabled with our future parsers, and it is fast and pure PHP now with the Remex implementation, so deprecate all the untidy variants of 'OutputPage::addWikiText*()' and add new methods which tidy by default. Clarify the content language/interface language distinction while we're at it by adding 'AsInterface' to the name of methods which use the "interface language" by default, and renaming the 'addWikiText*Tidy' methods to 'addWikiTextAsContent'. The 'OutputPage::addWikiTextTitle' method has been deprecated, but it is still used internally as the implementation for the newly-added methods. It is expected that the shared implementation will move in the future to a new private method. Setting the `$tidy` parameter of `OutputPage::addWikiTextTitle` to false is independently deprecated; for backwards-compatibility with old MW releases you may wish to continue to invoke OutputPage::addWikiTextTitle() but set $tidy=true; this will result in the same tidied output that the newly added methods would produce. Bug: T198214 Change-Id: Ia58910164baaca608cea3b24333b7d13ed773339
2018-09-21 16:24:57 +00:00
}
Add OutputPage::wrapWikiTextAsInterface() to safely wrap wikitext This patch introduces a new method, OutputPage::wrapWikiTextAsInterface(), which wraps the result of OutputPage::addWikiTextAsInterface() in a tidy and robust way that won't break if the wrapped message contains double newlines, extra </div>, or other nasties. This replaces a common unsafe pattern: $output->addWikiText( '<div class="..."' .... '</div>' ); Bug: T205624 Change-Id: I1040c7cf0ec1f5c4bef7c06d4486f50d85f2dc0f
2018-09-27 15:04:45 +00:00
/**
* Convert wikitext *in the user interface language* to HTML and
* add it to the buffer with a `<div class="$wrapperClass">`
* wrapper. The result will not be language-converted, as user
* interface messages as already localized into a specific
* variant. The $text will be parsed in start-of-line context.
* Output will be tidy.
*
* @param string $wrapperClass The class attribute value for the <div>
* wrapper in the output HTML
* @param string $text Wikitext in the user interface language
* @since 1.32
*/
public function wrapWikiTextAsInterface(
$wrapperClass, $text
) {
$this->addWikiTextTitleInternal(
$text, $this->getTitle(),
Remove OutputPage::addWikitext()* functions, hard-deprecated in 1.32 This starts cleaning up the programmer-visible API for OutputPage and removed some deprecated untidy parser modes. Change-Id: Ib464b57248f114b68424ec1175d36ad86d1319ad
2019-06-19 15:57:36 +00:00
/*linestart*/true, /*interface*/true,
Add OutputPage::wrapWikiTextAsInterface() to safely wrap wikitext This patch introduces a new method, OutputPage::wrapWikiTextAsInterface(), which wraps the result of OutputPage::addWikiTextAsInterface() in a tidy and robust way that won't break if the wrapped message contains double newlines, extra </div>, or other nasties. This replaces a common unsafe pattern: $output->addWikiText( '<div class="..."' .... '</div>' ); Bug: T205624 Change-Id: I1040c7cf0ec1f5c4bef7c06d4486f50d85f2dc0f
2018-09-27 15:04:45 +00:00
$wrapperClass
);
}
Deprecate and rename OutputPage::addWikiText* methods Tidy will always be enabled with our future parsers, and it is fast and pure PHP now with the Remex implementation, so deprecate all the untidy variants of 'OutputPage::addWikiText*()' and add new methods which tidy by default. Clarify the content language/interface language distinction while we're at it by adding 'AsInterface' to the name of methods which use the "interface language" by default, and renaming the 'addWikiText*Tidy' methods to 'addWikiTextAsContent'. The 'OutputPage::addWikiTextTitle' method has been deprecated, but it is still used internally as the implementation for the newly-added methods. It is expected that the shared implementation will move in the future to a new private method. Setting the `$tidy` parameter of `OutputPage::addWikiTextTitle` to false is independently deprecated; for backwards-compatibility with old MW releases you may wish to continue to invoke OutputPage::addWikiTextTitle() but set $tidy=true; this will result in the same tidied output that the newly added methods would produce. Bug: T198214 Change-Id: Ia58910164baaca608cea3b24333b7d13ed773339
2018-09-21 16:24:57 +00:00
/**
* Convert wikitext *in the page content language* to HTML and add
* it to the buffer. The result with be language-converted to the
* user's preferred variant. Assumes that the current page title
* will be used if optional $title is not provided. Output will be
* tidy.
*
* @param string $text Wikitext in the page content language
* @param bool $linestart Is this the start of a line? (Defaults to true)
Remove Title from public interface of OutputPage Bug: T278459 Change-Id: If455682540faf8d42fd8a14e0ef09bac24f72db4
2021-04-08 19:11:06 +00:00
* @param PageReference|null $title Optional title to use; default of `null`
Deprecate and rename OutputPage::addWikiText* methods Tidy will always be enabled with our future parsers, and it is fast and pure PHP now with the Remex implementation, so deprecate all the untidy variants of 'OutputPage::addWikiText*()' and add new methods which tidy by default. Clarify the content language/interface language distinction while we're at it by adding 'AsInterface' to the name of methods which use the "interface language" by default, and renaming the 'addWikiText*Tidy' methods to 'addWikiTextAsContent'. The 'OutputPage::addWikiTextTitle' method has been deprecated, but it is still used internally as the implementation for the newly-added methods. It is expected that the shared implementation will move in the future to a new private method. Setting the `$tidy` parameter of `OutputPage::addWikiTextTitle` to false is independently deprecated; for backwards-compatibility with old MW releases you may wish to continue to invoke OutputPage::addWikiTextTitle() but set $tidy=true; this will result in the same tidied output that the newly added methods would produce. Bug: T198214 Change-Id: Ia58910164baaca608cea3b24333b7d13ed773339
2018-09-21 16:24:57 +00:00
* means use current page title.
* @throws MWException if $title is not provided and OutputPage::getTitle()
* is null
* @since 1.32
*/
public function addWikiTextAsContent(
Remove Title from public interface of OutputPage Bug: T278459 Change-Id: If455682540faf8d42fd8a14e0ef09bac24f72db4
2021-04-08 19:11:06 +00:00
$text, $linestart = true, PageReference $title = null
Deprecate and rename OutputPage::addWikiText* methods Tidy will always be enabled with our future parsers, and it is fast and pure PHP now with the Remex implementation, so deprecate all the untidy variants of 'OutputPage::addWikiText*()' and add new methods which tidy by default. Clarify the content language/interface language distinction while we're at it by adding 'AsInterface' to the name of methods which use the "interface language" by default, and renaming the 'addWikiText*Tidy' methods to 'addWikiTextAsContent'. The 'OutputPage::addWikiTextTitle' method has been deprecated, but it is still used internally as the implementation for the newly-added methods. It is expected that the shared implementation will move in the future to a new private method. Setting the `$tidy` parameter of `OutputPage::addWikiTextTitle` to false is independently deprecated; for backwards-compatibility with old MW releases you may wish to continue to invoke OutputPage::addWikiTextTitle() but set $tidy=true; this will result in the same tidied output that the newly added methods would produce. Bug: T198214 Change-Id: Ia58910164baaca608cea3b24333b7d13ed773339
2018-09-21 16:24:57 +00:00
) {
Make use of ??= in more places New feature from PHP 7.4 Change-Id: Ifa7a9bc7b2ec415ad7ecb23f4c1776f51f58fd6b
2022-12-16 23:48:27 +00:00
$title ??= $this->getTitle();
Deprecate and rename OutputPage::addWikiText* methods Tidy will always be enabled with our future parsers, and it is fast and pure PHP now with the Remex implementation, so deprecate all the untidy variants of 'OutputPage::addWikiText*()' and add new methods which tidy by default. Clarify the content language/interface language distinction while we're at it by adding 'AsInterface' to the name of methods which use the "interface language" by default, and renaming the 'addWikiText*Tidy' methods to 'addWikiTextAsContent'. The 'OutputPage::addWikiTextTitle' method has been deprecated, but it is still used internally as the implementation for the newly-added methods. It is expected that the shared implementation will move in the future to a new private method. Setting the `$tidy` parameter of `OutputPage::addWikiTextTitle` to false is independently deprecated; for backwards-compatibility with old MW releases you may wish to continue to invoke OutputPage::addWikiTextTitle() but set $tidy=true; this will result in the same tidied output that the newly added methods would produce. Bug: T198214 Change-Id: Ia58910164baaca608cea3b24333b7d13ed773339
2018-09-21 16:24:57 +00:00
if ( !$title ) {
throw new MWException( 'Title is null' );
}
Remove OutputPage::addWikitext()* functions, hard-deprecated in 1.32 This starts cleaning up the programmer-visible API for OutputPage and removed some deprecated untidy parser modes. Change-Id: Ib464b57248f114b68424ec1175d36ad86d1319ad
2019-06-19 15:57:36 +00:00
$this->addWikiTextTitleInternal( $text, $title, $linestart, /*interface*/false );
Hard deprecate unused OutputPage::addWikiText* methods Codesearch shows no users of `OutputPage::addWikiTextWithTitle()`, and no users of `OutputPage::addWikiTextTitle()` (other than the implicit uses by the other `OutputPage::addWikiText*()` methods). These methods produce untidy output, which future parsers won't support. Bug: T198214 Change-Id: Id5ee3bdfa6c464e3a92af82af7bc7317ca9d07a9
2018-09-21 17:16:42 +00:00
}
/**
* Add wikitext with a custom Title object.
* Output is unwrapped.
*
* @param string $text Wikitext
Remove Title from public interface of OutputPage Bug: T278459 Change-Id: If455682540faf8d42fd8a14e0ef09bac24f72db4
2021-04-08 19:11:06 +00:00
* @param PageReference $title
* @param bool $linestart Is this the start of a line?@param
Hard deprecate unused OutputPage::addWikiText* methods Codesearch shows no users of `OutputPage::addWikiTextWithTitle()`, and no users of `OutputPage::addWikiTextTitle()` (other than the implicit uses by the other `OutputPage::addWikiText*()` methods). These methods produce untidy output, which future parsers won't support. Bug: T198214 Change-Id: Id5ee3bdfa6c464e3a92af82af7bc7317ca9d07a9
2018-09-21 17:16:42 +00:00
* @param bool $interface Whether it is an interface message
* (for example disables conversion)
Improve param docs Change-Id: I746a69f6ed01c3ff000da125457df62b02d13b34
2019-11-23 22:28:57 +00:00
* @param string|null $wrapperClass if not empty, wraps the output in
Add OutputPage::wrapWikiTextAsInterface() to safely wrap wikitext This patch introduces a new method, OutputPage::wrapWikiTextAsInterface(), which wraps the result of OutputPage::addWikiTextAsInterface() in a tidy and robust way that won't break if the wrapped message contains double newlines, extra </div>, or other nasties. This replaces a common unsafe pattern: $output->addWikiText( '<div class="..."' .... '</div>' ); Bug: T205624 Change-Id: I1040c7cf0ec1f5c4bef7c06d4486f50d85f2dc0f
2018-09-27 15:04:45 +00:00
* a `<div class="$wrapperClass">`
Hard deprecate unused OutputPage::addWikiText* methods Codesearch shows no users of `OutputPage::addWikiTextWithTitle()`, and no users of `OutputPage::addWikiTextTitle()` (other than the implicit uses by the other `OutputPage::addWikiText*()` methods). These methods produce untidy output, which future parsers won't support. Bug: T198214 Change-Id: Id5ee3bdfa6c464e3a92af82af7bc7317ca9d07a9
2018-09-21 17:16:42 +00:00
*/
private function addWikiTextTitleInternal(
Remove Title from public interface of OutputPage Bug: T278459 Change-Id: If455682540faf8d42fd8a14e0ef09bac24f72db4
2021-04-08 19:11:06 +00:00
$text, PageReference $title, $linestart, $interface, $wrapperClass = null
Pass php-strict for OutputPage Change-Id: Ia5d03277fff543ca672dc8efc7eaaeca4888ac23
2014-05-11 01:54:15 +00:00
) {
Deprecate OutputPage::parse() and OutputPage::parseInline() The OutputPage::parse() and OutputPage::parseInline() methods behave misleadingly different from the OutputPage::addWikitext*() methods: they don't tidy their output, they have different defaults for interface/content language selection, and they (sometimes) add wrapper divs. Deprecate these and add new methods with tidy output, clear language selection, and consistent defaults: OutputPage::parseAsContent(), OutputPage::parseAsInterface(), and OutputPage::parseInlineAsInterface(). Unify the implementation of the parse* methods with the addWikiText* methods, to reduce the likelihood that the behavior will diverge again in the future. Bug: T198214 Change-Id: Ica79c2acbc542ef37f971c0be2582ae771a23bd0
2018-10-26 15:14:01 +00:00
$parserOutput = $this->parseInternal(
Remove deprecated non-tidy OutputPage::parse/parseInline methods The non-tidy parsing modes were deprecated in 1.32. Remove the deprecated parse/parseInline methods which use them, as part of a general clean up of nontidy parsing modes. Bug: T198214 Change-Id: I04ea82dd20eac8b0ce07a8d0e6b0bf2c4b03dbe8
2020-01-15 19:03:33 +00:00
$text, $title, $linestart, $interface
*coding style cleanup (added braces, spacing tweaks etc.) *removed some unused globals
2010-05-22 12:18:22 +00:00
);
* remove end of line whitespace * remove empty lines at end of file * remove "?>" where still present
2008-04-14 07:45:50 +00:00
Use ParserOutput stateless transforms We still set the state in many cases for benefit of extensions, but all calls within core should no longer be using non-default state. Change-Id: I78b62ec33fcb8273acb9b3b4e9012215442be94c Depends-On: I140ff32373430b61b92226689ef9b58cca317450
2017-11-22 20:07:51 +00:00
$this->addParserOutput( $parserOutput, [
'enableSectionEditLinks' => false,
Add OutputPage::wrapWikiTextAsInterface() to safely wrap wikitext This patch introduces a new method, OutputPage::wrapWikiTextAsInterface(), which wraps the result of OutputPage::addWikiTextAsInterface() in a tidy and robust way that won't break if the wrapped message contains double newlines, extra </div>, or other nasties. This replaces a common unsafe pattern: $output->addWikiText( '<div class="..."' .... '</div>' ); Bug: T205624 Change-Id: I1040c7cf0ec1f5c4bef7c06d4486f50d85f2dc0f
2018-09-27 15:04:45 +00:00
'wrapperDivClass' => $wrapperClass ?? '',
Use ParserOutput stateless transforms We still set the state in many cases for benefit of extensions, but all calls within core should no longer be using non-default state. Change-Id: I78b62ec33fcb8273acb9b3b4e9012215442be94c Depends-On: I140ff32373430b61b92226689ef9b58cca317450
2017-11-22 20:07:51 +00:00
] );
fix for broken categories on preview
2006-01-01 20:08:08 +00:00
}
Give skins more flexibility over table of contents render * Do not store table of contents in parser output * Instead inject table of contents via strpos where needed inside Article based on Skin "toc" option * Use <mw:tocplace> as a TOC placeholder; for Parsoid compatibility this will be replaced with a <meta> tag in a followup patch. Bug: T287767 Change-Id: I44045b3b9e78e7ab793da3f37e3c0dbc91cd7d39
2021-09-15 01:00:06 +00:00
/**
Generate/set/get TOCData/SectionMetadata objects instead of arrays * ParserOutput::setSections()/::getSections() are expected to be deprecated. Uses in extensions and skins will need to be migrated in follow up patches once the new interface has stabilized. * In the skins code, the metadata is converted back to an array. Downstream skin TOC consumers will need to be migrated as well before we can remove the toLegacy() conversion. * Fixed SerializationTestTrait's validation method - Not sure if this is overkill but should handle all future complex objects we might stuff into the ParserCache. * This patch emits a backward-compatible Sections property in order to avoid changing the parser cache serialization format. T327439 has been filed to eventually use the JsonCodec support for object serialization, but for this initial patch it makes sense to avoid the need for a concurrent ParserCache format migration by using a backward-compatible serialization. * TOCData is nullable because the intent is that ParserOutput::setTOCData() is MW_MERGE_STRATEGY_WRITE_ONCE; that is, only the top-level fragment composing a page will set the TOCData. This will be enforced in the future via wfDeprecated() (T327429), but again our first patch is as backward-compatible as possible. Bug: T296025 Depends-On: I1b267d23cf49d147c5379b914531303744481b68 Co-Authored-By: C. Scott Ananian <cananian@wikimedia.org> Co-Authored-By: Subramanya Sastry <ssastry@wikimedia.org> Change-Id: I8329864535f0b1dd5f9163868a08d6cb1ffcb78f
2022-09-01 23:07:29 +00:00
* Adds Table of Contents data to OutputPage from ParserOutput
* @param TOCData $tocData
Give skins more flexibility over table of contents render * Do not store table of contents in parser output * Instead inject table of contents via strpos where needed inside Article based on Skin "toc" option * Use <mw:tocplace> as a TOC placeholder; for Parsoid compatibility this will be replaced with a <meta> tag in a followup patch. Bug: T287767 Change-Id: I44045b3b9e78e7ab793da3f37e3c0dbc91cd7d39
2021-09-15 01:00:06 +00:00
* @internal For use by Article.php
*/
Generate/set/get TOCData/SectionMetadata objects instead of arrays * ParserOutput::setSections()/::getSections() are expected to be deprecated. Uses in extensions and skins will need to be migrated in follow up patches once the new interface has stabilized. * In the skins code, the metadata is converted back to an array. Downstream skin TOC consumers will need to be migrated as well before we can remove the toLegacy() conversion. * Fixed SerializationTestTrait's validation method - Not sure if this is overkill but should handle all future complex objects we might stuff into the ParserCache. * This patch emits a backward-compatible Sections property in order to avoid changing the parser cache serialization format. T327439 has been filed to eventually use the JsonCodec support for object serialization, but for this initial patch it makes sense to avoid the need for a concurrent ParserCache format migration by using a backward-compatible serialization. * TOCData is nullable because the intent is that ParserOutput::setTOCData() is MW_MERGE_STRATEGY_WRITE_ONCE; that is, only the top-level fragment composing a page will set the TOCData. This will be enforced in the future via wfDeprecated() (T327429), but again our first patch is as backward-compatible as possible. Bug: T296025 Depends-On: I1b267d23cf49d147c5379b914531303744481b68 Co-Authored-By: C. Scott Ananian <cananian@wikimedia.org> Co-Authored-By: Subramanya Sastry <ssastry@wikimedia.org> Change-Id: I8329864535f0b1dd5f9163868a08d6cb1ffcb78f
2022-09-01 23:07:29 +00:00
public function setTOCData( TOCData $tocData ) {
$this->tocData = $tocData;
Give skins more flexibility over table of contents render * Do not store table of contents in parser output * Instead inject table of contents via strpos where needed inside Article based on Skin "toc" option * Use <mw:tocplace> as a TOC placeholder; for Parsoid compatibility this will be replaced with a <meta> tag in a followup patch. Bug: T287767 Change-Id: I44045b3b9e78e7ab793da3f37e3c0dbc91cd7d39
2021-09-15 01:00:06 +00:00
}
/**
Generate/set/get TOCData/SectionMetadata objects instead of arrays * ParserOutput::setSections()/::getSections() are expected to be deprecated. Uses in extensions and skins will need to be migrated in follow up patches once the new interface has stabilized. * In the skins code, the metadata is converted back to an array. Downstream skin TOC consumers will need to be migrated as well before we can remove the toLegacy() conversion. * Fixed SerializationTestTrait's validation method - Not sure if this is overkill but should handle all future complex objects we might stuff into the ParserCache. * This patch emits a backward-compatible Sections property in order to avoid changing the parser cache serialization format. T327439 has been filed to eventually use the JsonCodec support for object serialization, but for this initial patch it makes sense to avoid the need for a concurrent ParserCache format migration by using a backward-compatible serialization. * TOCData is nullable because the intent is that ParserOutput::setTOCData() is MW_MERGE_STRATEGY_WRITE_ONCE; that is, only the top-level fragment composing a page will set the TOCData. This will be enforced in the future via wfDeprecated() (T327429), but again our first patch is as backward-compatible as possible. Bug: T296025 Depends-On: I1b267d23cf49d147c5379b914531303744481b68 Co-Authored-By: C. Scott Ananian <cananian@wikimedia.org> Co-Authored-By: Subramanya Sastry <ssastry@wikimedia.org> Change-Id: I8329864535f0b1dd5f9163868a08d6cb1ffcb78f
2022-09-01 23:07:29 +00:00
* @internal For usage in Skin::getTOCData() only.
* @return ?TOCData Table of Contents data, or
* null if OutputPage::setTOCData() has not been called.
Give skins more flexibility over table of contents render * Do not store table of contents in parser output * Instead inject table of contents via strpos where needed inside Article based on Skin "toc" option * Use <mw:tocplace> as a TOC placeholder; for Parsoid compatibility this will be replaced with a <meta> tag in a followup patch. Bug: T287767 Change-Id: I44045b3b9e78e7ab793da3f37e3c0dbc91cd7d39
2021-09-15 01:00:06 +00:00
*/
Generate/set/get TOCData/SectionMetadata objects instead of arrays * ParserOutput::setSections()/::getSections() are expected to be deprecated. Uses in extensions and skins will need to be migrated in follow up patches once the new interface has stabilized. * In the skins code, the metadata is converted back to an array. Downstream skin TOC consumers will need to be migrated as well before we can remove the toLegacy() conversion. * Fixed SerializationTestTrait's validation method - Not sure if this is overkill but should handle all future complex objects we might stuff into the ParserCache. * This patch emits a backward-compatible Sections property in order to avoid changing the parser cache serialization format. T327439 has been filed to eventually use the JsonCodec support for object serialization, but for this initial patch it makes sense to avoid the need for a concurrent ParserCache format migration by using a backward-compatible serialization. * TOCData is nullable because the intent is that ParserOutput::setTOCData() is MW_MERGE_STRATEGY_WRITE_ONCE; that is, only the top-level fragment composing a page will set the TOCData. This will be enforced in the future via wfDeprecated() (T327429), but again our first patch is as backward-compatible as possible. Bug: T296025 Depends-On: I1b267d23cf49d147c5379b914531303744481b68 Co-Authored-By: C. Scott Ananian <cananian@wikimedia.org> Co-Authored-By: Subramanya Sastry <ssastry@wikimedia.org> Change-Id: I8329864535f0b1dd5f9163868a08d6cb1ffcb78f
2022-09-01 23:07:29 +00:00
public function getTOCData(): ?TOCData {
return $this->tocData;
Give skins more flexibility over table of contents render * Do not store table of contents in parser output * Instead inject table of contents via strpos where needed inside Article based on Skin "toc" option * Use <mw:tocplace> as a TOC placeholder; for Parsoid compatibility this will be replaced with a <meta> tag in a followup patch. Bug: T287767 Change-Id: I44045b3b9e78e7ab793da3f37e3c0dbc91cd7d39
2021-09-15 01:00:06 +00:00
}
OutputPage: Add addParserOutputContent() for more fine-grained control We previously had addParserOutput(), which added everything and did some other magic, and addParserOutputNoText() which, as the same says, added everything but the text. I renamed addParserOutputNoText() to addParserOutputMetadata() and created two more functions: * addParserOutputText(): This is almost identical to adding the raw HTML, but calls the OutputPageBeforeHTML hook like other addParserOutput*() methods. * addParserOutputContent(): Like addParserOutputText(), but also adds the ResourceLoader modules and variables associated with the parser output. This is important especially for some extensions like TemplateData or SyntaxHighlight which add styles to the page to enhance the display. Change-Id: Iead541886fd1ccdbdf1cb06af71b34cd04644985
2014-05-08 20:10:13 +00:00
/**
* Add all metadata associated with a ParserOutput object, but without the actual HTML. This
* includes categories, language links, ResourceLoader modules, effects of certain magic words,
Give skins more flexibility over table of contents render * Do not store table of contents in parser output * Instead inject table of contents via strpos where needed inside Article based on Skin "toc" option * Use <mw:tocplace> as a TOC placeholder; for Parsoid compatibility this will be replaced with a <meta> tag in a followup patch. Bug: T287767 Change-Id: I44045b3b9e78e7ab793da3f37e3c0dbc91cd7d39
2021-09-15 01:00:06 +00:00
* and so on. It does *not* include section information.
OutputPage: Add addParserOutputContent() for more fine-grained control We previously had addParserOutput(), which added everything and did some other magic, and addParserOutputNoText() which, as the same says, added everything but the text. I renamed addParserOutputNoText() to addParserOutputMetadata() and created two more functions: * addParserOutputText(): This is almost identical to adding the raw HTML, but calls the OutputPageBeforeHTML hook like other addParserOutput*() methods. * addParserOutputContent(): Like addParserOutputText(), but also adds the ResourceLoader modules and variables associated with the parser output. This is important especially for some extensions like TemplateData or SyntaxHighlight which add styles to the page to enhance the display. Change-Id: Iead541886fd1ccdbdf1cb06af71b34cd04644985
2014-05-08 20:10:13 +00:00
*
* @since 1.24
* @param ParserOutput $parserOutput
*/
Add type hint for ParserOutput EditPage::getPreviewLimitReport is called by EditPage::showEditForm with $output = null. Specify this in the @param tag and allow this by a default value. Change-Id: Iec8905aab736a1f254a57853c7cab935d008653e
2018-07-27 20:20:39 +00:00
public function addParserOutputMetadata( ParserOutput $parserOutput ) {
Deprecate protected and public properties of OutputPage We plan to refactor these properties and move them into a value object instance. Mark them deprecated for 1.38 so we can move/remove them in a future release. Bug: T301020 Change-Id: I8ec2e903bd7cde8d54fe0366707cbcf625f1f1a1
2022-02-08 16:11:32 +00:00
// T301020 This should eventually use the standard "merge ParserOutput"
// function between $parserOutput and $this->metadata.
Use array_merge() for OutputPage::$mLanguageLinks, not + Based on documentation together with inspection of some callers, the intent seems to be that this is an indexed array, not associative. + will therefore do totally the wrong thing, ignoring any new values that have the same key as an existing item (e.g., '0' or '1'). Even if it was an associative array, + keeps the values on the left-hand side, which means you normally want to do $foo = $bar + $foo instead of $foo += $bar if you want to overwrite old values with the new ones. Before this change, calling addLanguageLinks() or addParserOutputMetadata() would generally not add all of the links it was supposed to if there were already links defined. (It could still work if the arrays' keys didn't conflict for some reason, e.g., something passed an associative array or an indexed array with a hole.) I don't know if anything actually hits this bug, because it's likely that callers usually add all their links at once. I find no uses of addLanguageLinks() at all. I found this bug while working on adding more tests for OutputPage, and the tests for this change will be submitted later in Icdc0288c04b8c4ba841f9fbb3e05a0cdc8a20fa5. Change-Id: I53f6e7ea94417b0034371e56e733e8c86af21658
2018-07-25 18:36:56 +00:00
$this->mLanguageLinks =
array_merge( $this->mLanguageLinks, $parserOutput->getLanguageLinks() );
* Added templatelinks table. The table currently represents a literal list of templates included from each article. That is, the table contains pages which were actually loaded during parsing, not the markup which went into resolving their names. * Ended the role of $wgLinkCache in link updates. Instead, links (and related entities) are registered in the ParserOutput object during a parse. The LinksUpdate constructor now takes a ParserOutput object as a parameter. $wgLinkCache is still used, but only as a cache of article IDs. * Because the link list is now saved and restored in the parser cache, meta tag keywords now work on parser cache hits. Some refactoring took place in this area. * Rendering of the HTML for category links has moved from Parser to OutputPage. * Did some general pottering around in Article.php, such as allowing an Article object to be created with a specified revision ID, thereby optionally removing the dependence on $wgRequest. Not used at the current time. * A few documentation tweaks.
2005-12-30 09:33:11 +00:00
$this->addCategoryLinks( $parserOutput->getCategories() );
parse: Wrap indicators with a div.mw-parser-output element Alternative to having this in the skin, is to do it for each separate indicator. A counter argument against this was inline vs block mode elements, specifically the reason why we stripped the p element. But the p element has a margin in each and every skin. It's rather unlikely that div's have such a standard layout. The wrapping div around each indicator at the skin level is already inline-block, so interactions between the indicators are already decoupled via that skin element. In this patch the wrapping is done in OutputPage when a ParserOutput is added. OutputPage::addHelpLink() also adds indicators, which would not be wrapped by this patch; this is reasonable since the content of ::addHelpLink() is not parsed wikitext. It's possible external users may also call OutputPage::setIndicators() directly, which would also add unwrapped indicators; the caller of OutputPage::setIndicators() is responsible for wrapping any indicators which might come from parsed wikitext. Bug: T188443 Co-Authored-By: C. Scott Ananian <cananian@wikimedia.org> Change-Id: I42be893101f7d7887480fe635143bea535ee5634
2021-12-22 00:19:02 +00:00
// Parser-generated indicators get wrapped like other parser output.
$wrapClass = $parserOutput->getWrapperDivClass();
$result = [];
foreach ( $parserOutput->getIndicators() as $name => $html ) {
if ( $html !== '' && $wrapClass !== '' ) {
$html = Html::rawElement( 'div', [ 'class' => $wrapClass ], $html );
}
$result[$name] = $html;
}
$this->setIndicators( $result );
Give skins more flexibility over table of contents render * Do not store table of contents in parser output * Instead inject table of contents via strpos where needed inside Article based on Skin "toc" option * Use <mw:tocplace> as a TOC placeholder; for Parsoid compatibility this will be replaced with a <meta> tag in a followup patch. Bug: T287767 Change-Id: I44045b3b9e78e7ab793da3f37e3c0dbc91cd7d39
2021-09-15 01:00:06 +00:00
Generate/set/get TOCData/SectionMetadata objects instead of arrays * ParserOutput::setSections()/::getSections() are expected to be deprecated. Uses in extensions and skins will need to be migrated in follow up patches once the new interface has stabilized. * In the skins code, the metadata is converted back to an array. Downstream skin TOC consumers will need to be migrated as well before we can remove the toLegacy() conversion. * Fixed SerializationTestTrait's validation method - Not sure if this is overkill but should handle all future complex objects we might stuff into the ParserCache. * This patch emits a backward-compatible Sections property in order to avoid changing the parser cache serialization format. T327439 has been filed to eventually use the JsonCodec support for object serialization, but for this initial patch it makes sense to avoid the need for a concurrent ParserCache format migration by using a backward-compatible serialization. * TOCData is nullable because the intent is that ParserOutput::setTOCData() is MW_MERGE_STRATEGY_WRITE_ONCE; that is, only the top-level fragment composing a page will set the TOCData. This will be enforced in the future via wfDeprecated() (T327429), but again our first patch is as backward-compatible as possible. Bug: T296025 Depends-On: I1b267d23cf49d147c5379b914531303744481b68 Co-Authored-By: C. Scott Ananian <cananian@wikimedia.org> Co-Authored-By: Subramanya Sastry <ssastry@wikimedia.org> Change-Id: I8329864535f0b1dd5f9163868a08d6cb1ffcb78f
2022-09-01 23:07:29 +00:00
$tocData = $parserOutput->getTOCData();
if ( $tocData !== null ) {
$this->setTOCData( $tocData );
}
OutputPage: Call setSections() from addParserOutputMetadata() Remove calls elsewhere which now became redundant. Bug: T325562 Bug: T325600 Change-Id: I1611f3bdea46a729ecf9daf699d761b018d49b55
2022-12-19 22:43:43 +00:00
Give skins more flexibility over table of contents render * Do not store table of contents in parser output * Instead inject table of contents via strpos where needed inside Article based on Skin "toc" option * Use <mw:tocplace> as a TOC placeholder; for Parsoid compatibility this will be replaced with a <meta> tag in a followup patch. Bug: T287767 Change-Id: I44045b3b9e78e7ab793da3f37e3c0dbc91cd7d39
2021-09-15 01:00:06 +00:00
// FIXME: Best practice is for OutputPage to be an accumulator, as
// addParserOutputMetadata() may be called multiple times, but the
// following lines overwrite any previous data. These should
Deprecate protected and public properties of OutputPage We plan to refactor these properties and move them into a value object instance. Mark them deprecated for 1.38 so we can move/remove them in a future release. Bug: T301020 Change-Id: I8ec2e903bd7cde8d54fe0366707cbcf625f1f1a1
2022-02-08 16:11:32 +00:00
// be migrated to an injection pattern. (T301020, T300979)
(bug 4876) Add __NEWSECTIONLINK__ magic word to force the "new section" link/tab to show up on specific pages on demand
2006-05-01 20:35:08 +00:00
$this->mNewSectionLink = $parserOutput->getNewSection();
* (bug 16335) __NONEWSECTIONLINK__ magic word to suppress new section link. Patch by Carlin: https://bugzilla.wikimedia.org/attachment.cgi?id=5680 With slight whitespace tweaks.
2009-02-19 22:14:59 +00:00
$this->mHideNewSectionLink = $parserOutput->getHideNewSection();
Give skins more flexibility over table of contents render * Do not store table of contents in parser output * Instead inject table of contents via strpos where needed inside Article based on Skin "toc" option * Use <mw:tocplace> as a TOC placeholder; for Parsoid compatibility this will be replaced with a <meta> tag in a followup patch. Bug: T287767 Change-Id: I44045b3b9e78e7ab793da3f37e3c0dbc91cd7d39
2021-09-15 01:00:06 +00:00
$this->mNoGallery = $parserOutput->getNoGallery();
Instead of last commit, use $wgExemptFromUserRobotsControl to control __INDEX__/__NOINDEX__. This defaults to null, which means $wgContentNamespaces. Pages whose namespaces are in the array will be unaffected by __INDEX__/__NOINDEX__. This will prevent serious vandalism, so the feature is again enabled by default (with content namespaces exempted). As was pointed out, it's probably better anyway that the vandal noindex a page and have nothing turn up in searches if the search engine arrives at that exact time, than to have the vandal replace the page with "ARTICLE SUBJECT IS A POOPY-HEAD" and have that turn up in searches if the search engine arrives at that exact time. :) At any rate, this should solve the issue.
2008-07-24 18:02:20 +00:00
allow parser/extensions to control for how long a ParserOutput gets cached. Introduced updateCacheExpiry(), Cleaned up use of getCacheTime() for marking uncacheable objects.
2010-06-01 14:28:51 +00:00
if ( !$parserOutput->isCacheable() ) {
Add OutputPage::disableClientCache() and deprecate ::enableClientCache(false) As far as I can tell, outside of test cases, every single time we call OutputPage::enableClientCache() we pass it `false` and actually mean to *disable* the client cache. Create a new less-confusingly-named method and deprecate the old one. Change-Id: I7c89e20528a0d91173f0edcb997dcae631935ee5
2022-02-04 19:12:11 +00:00
$this->disableClientCache();
basic support for special page inclusions
2005-05-28 11:09:22 +00:00
}
* (bug 22061) API: add prop=headitems to action=parse Patch contributed by Paul Copperman
2010-01-16 13:07:58 +00:00
$this->mHeadItems = array_merge( $this->mHeadItems, $parserOutput->getHeadItems() );
Merging resourceloader branch into trunk. Full documentation is at http://www.mediawiki.org/wiki/ResourceLoader and a general overview has been posted on wikitech-li <http://lists.wikimedia.org/pipermail/wikitech-l/2010-September/049253.html>. One important change is that all JS is now loaded at the bottom, so any scripts assuming things from wikibits or whatever are present will fail.
2010-09-04 04:00:09 +00:00
$this->addModules( $parserOutput->getModules() );
* (bug 15558) Allow includable special pages to be parameterized using wiki syntax * Changed SpecialPageFactory::capturePath() to take the same parameters as SpecialPageFactory::executePath() (except the last one) and made it always return a bool instead of bool-or-string. The result HTML can be fetched from the OutputPage object of the context. * Added module styles, scritps and messages members to ParserOutput in addition to modules. The first one is needed to display Special:RecentChanges correctly when transcluded since EnhancedChangesList::beginRecentChangesList() calls addModuleStyles( 'mediawiki.special.changeslist' ) Yes, this means that you can use {{Special:Recentchanges|enhanced=0}} to use the old changes list. For the ones that wonder, {{Special:Recentchanges|uselang=something}} will not work since the Language object is forced to the one used by the parser.
2011-08-02 15:40:03 +00:00
$this->addModuleStyles( $parserOutput->getModuleStyles() );
Add ParserOutput::addJsConfigVars The OutputPage has variables for modules, moduleScripts, moduleStyles, moduleMessages and the config vars, but the ParserOutput is missing the last one. With ParserOutput::addJsConfigVars it is possible to add scripts and it config vars at one place and have not use the MakeGlobalVariablesScript hook or other ways to get the needed javascript variable in the output. Change-Id: I6ad61cca76805f6b9d76e053c98c7509c323d5da
2014-01-19 15:39:46 +00:00
$this->addJsConfigVars( $parserOutput->getJsConfigVars() );
SECURITY: Copy prevent-clickjacking between OutputPage and ParserOutput Special page transclusion returns an OutputPage, whose metadata is copied into the ParserOutput, and then later back into an OutputPage. The "preventClickjacking" flag should be part of that metadata. Bug: 65778 Change-Id: I17d2720fb94bb383a92059e5adbf6c16ee3e9ef4
2014-07-10 19:16:29 +00:00
$this->mPreventClickjacking = $this->mPreventClickjacking
Rename ParserOutput::{allow,prevent}Clickjacking() -> ::{get,set}PreventClickjacking() This name is consist with the rest of the setter and getter methods in ParserOutput. Renamed the methods in OutputPage, ImageHistoryList, ImageHistoryPseudoPager, and ContribsPager as well for consistency; it also makes chasing down lingering references in codesearch easier. Soft-deprecated the old name for 1.38. Hard-deprecation will follow, but there are a number of users in production that should be chased down first. Code search: https://codesearch.https://codesearch.wmcloud.org/deployed/?q=(allow%7Cprevent)Clickjacking&i=nope&files=&excludeFiles=&repos= Bug: T287216 Change-Id: I9822c60c180d204bd30cb4447a1120155d456da4
2021-09-29 20:58:59 +00:00
|| $parserOutput->getPreventClickjacking();
Allow storing additional CSP sources in ParserOutput This adds methods to ParserOutput ::addExtraCSPStyleSrc, ::addExtraCSPDefaultSrc, and ::addExtraCSPScriptSrc, to easily allow parser tags/functions to add additional CSP sources if their tag needs it. Previously such an extension would need to use and OutputPage hook. This is modeled on how addModules() works. The immediate use case is for Kartographer (T240960), although its expected that lots of extensions might do something like this, especially extensions used outside of Wikimedia. Change-Id: I24e5f0b4edff58025a0c2a3e1a9aa3f62eb7db7b
2020-02-03 09:50:14 +00:00
$scriptSrcs = $parserOutput->getExtraCSPScriptSrcs();
foreach ( $scriptSrcs as $src ) {
$this->getCSP()->addScriptSrc( $src );
}
$defaultSrcs = $parserOutput->getExtraCSPDefaultSrcs();
foreach ( $defaultSrcs as $src ) {
$this->getCSP()->addDefaultSrc( $src );
}
$styleSrcs = $parserOutput->getExtraCSPStyleSrcs();
foreach ( $styleSrcs as $src ) {
$this->getCSP()->addStyleSrc( $src );
}
* Put parser output file version tracking to core * Added some ParserOutput accessors * A few cleanups to fetchFile()
2011-03-23 17:35:40 +00:00
OutputPage: Fix undefined `['host']` in ImagePreconnect code Instead of checking the result of uncondiional `['host']` access in the foreach loop afterward, check the result of url parsing directly. Also: * Switch to plain parse_url(), we don't support exotic input here. * Move parsing to loop, minor optimisation, and will often thus mean we don't even need to parse or check the LocalRepo case, since the feature is generally only enabled when a cross-domain ForeignRepo exists. * Improve docs. * Remove experimental warning. Bug: T317329 Change-Id: I6492f5142861513e4a7eb65fe4c14fa12c46977a
2022-09-08 17:17:00 +00:00
// If $wgImagePreconnect is true, and if the output contains images, give the user-agent
// a hint about a remote hosts from which images may be served. Launched in T123582.
Use name constants to access config settings. Use name constants instead of string literals in calls to Config::get and ServiceOptions::get, when referring to core configuration variables. This protects against typos and makes the decumentation and schema declaration of the config settings discoverable. This is the first batch, only touching files directly under /includes/ Change-Id: I7252e636c7c86d950d9257b33491af492c6dd5eb
2022-04-01 15:58:32 +00:00
if ( $this->getConfig()->get( MainConfigNames::ImagePreconnect ) && count( $parserOutput->getImages() ) ) {
OutputPage: Add experimental preconnect resource hint for thumbnails Adds <link rel="preconnect"> for the first valid foreign or local file repo, on pages containing images. This is a hint to the browser that it should open a connection to the other host (e.g. upload.wikimedia.org), if it doesn't have one already. Doing so allows the connection setup to happen before page parsing is complete, and should result in faster image loading, especially on slow connections. Bug: T123582 Change-Id: I2dcc14f05012570a3e41ed8c7064969a4cbfb6db
2020-02-07 05:31:00 +00:00
$preconnect = [];
OutputPage: Fix undefined `['host']` in ImagePreconnect code Instead of checking the result of uncondiional `['host']` access in the foreach loop afterward, check the result of url parsing directly. Also: * Switch to plain parse_url(), we don't support exotic input here. * Move parsing to loop, minor optimisation, and will often thus mean we don't even need to parse or check the LocalRepo case, since the feature is generally only enabled when a cross-domain ForeignRepo exists. * Improve docs. * Remove experimental warning. Bug: T317329 Change-Id: I6492f5142861513e4a7eb65fe4c14fa12c46977a
2022-09-08 17:17:00 +00:00
// Optimization: Instead of processing each image, assume that wikis either serve both
// foreign and local from the same remote hostname (e.g. public wikis at WMF), or that
// foreign images are common enough to be worth the preconnect (e.g. private wikis).
OutputPage: Add experimental preconnect resource hint for thumbnails Adds <link rel="preconnect"> for the first valid foreign or local file repo, on pages containing images. This is a hint to the browser that it should open a connection to the other host (e.g. upload.wikimedia.org), if it doesn't have one already. Doing so allows the connection setup to happen before page parsing is complete, and should result in faster image loading, especially on slow connections. Bug: T123582 Change-Id: I2dcc14f05012570a3e41ed8c7064969a4cbfb6db
2020-02-07 05:31:00 +00:00
$repoGroup = MediaWikiServices::getInstance()->getRepoGroup();
Use static closures where safe to use This is micro-optimization of closure code to avoid binding the closure to $this where it is not needed. Created by I25a17fb22b6b669e817317a0f45051ae9c608208 Change-Id: I0ffc6200f6c6693d78a3151cb8cea7dce7c21653
2021-02-10 22:31:02 +00:00
$repoGroup->forEachForeignRepo( static function ( $repo ) use ( &$preconnect ) {
OutputPage: Fix undefined `['host']` in ImagePreconnect code Instead of checking the result of uncondiional `['host']` access in the foreach loop afterward, check the result of url parsing directly. Also: * Switch to plain parse_url(), we don't support exotic input here. * Move parsing to loop, minor optimisation, and will often thus mean we don't even need to parse or check the LocalRepo case, since the feature is generally only enabled when a cross-domain ForeignRepo exists. * Improve docs. * Remove experimental warning. Bug: T317329 Change-Id: I6492f5142861513e4a7eb65fe4c14fa12c46977a
2022-09-08 17:17:00 +00:00
$preconnect[] = $repo->getZoneUrl( 'thumb' );
OutputPage: Add experimental preconnect resource hint for thumbnails Adds <link rel="preconnect"> for the first valid foreign or local file repo, on pages containing images. This is a hint to the browser that it should open a connection to the other host (e.g. upload.wikimedia.org), if it doesn't have one already. Doing so allows the connection setup to happen before page parsing is complete, and should result in faster image loading, especially on slow connections. Bug: T123582 Change-Id: I2dcc14f05012570a3e41ed8c7064969a4cbfb6db
2020-02-07 05:31:00 +00:00
} );
OutputPage: Fix undefined `['host']` in ImagePreconnect code Instead of checking the result of uncondiional `['host']` access in the foreach loop afterward, check the result of url parsing directly. Also: * Switch to plain parse_url(), we don't support exotic input here. * Move parsing to loop, minor optimisation, and will often thus mean we don't even need to parse or check the LocalRepo case, since the feature is generally only enabled when a cross-domain ForeignRepo exists. * Improve docs. * Remove experimental warning. Bug: T317329 Change-Id: I6492f5142861513e4a7eb65fe4c14fa12c46977a
2022-09-08 17:17:00 +00:00
// Consider both foreign and local repos. While LocalRepo by default uses a relative
// path on the same domain, wiki farms may configure it to use a dedicated hostname.
$preconnect[] = $repoGroup->getLocalRepo()->getZoneUrl( 'thumb' );
foreach ( $preconnect as $url ) {
$host = parse_url( $url, PHP_URL_HOST );
// It is expected that file URLs are often path-only, without hostname (T317329).
OutputPage: Add experimental preconnect resource hint for thumbnails Adds <link rel="preconnect"> for the first valid foreign or local file repo, on pages containing images. This is a hint to the browser that it should open a connection to the other host (e.g. upload.wikimedia.org), if it doesn't have one already. Doing so allows the connection setup to happen before page parsing is complete, and should result in faster image loading, especially on slow connections. Bug: T123582 Change-Id: I2dcc14f05012570a3e41ed8c7064969a4cbfb6db
2020-02-07 05:31:00 +00:00
if ( $host ) {
$this->addLink( [ 'rel' => 'preconnect', 'href' => '//' . $host ] );
break;
}
}
}
* Put parser output file version tracking to core * Added some ParserOutput accessors * A few cleanups to fetchFile()
2011-03-23 17:35:40 +00:00
// Template versioning...
foreach ( (array)$parserOutput->getTemplateIds() as $ns => $dbks ) {
* Rewrote wfArrayMerge() in terms of array plus * Two callers of wfArrayMerge() were bugs, both assuming strange and complex behaviour in wfArrayMerge() which has never been present or documented. * Introduced wfMergeErrorArrays() to remove duplicates from merged error arrays, e.g. from getUserPermissionsErrors(). * Rewrote the remaining callers of wfArrayMerge() to use array plus. It makes the code clearer, assuming the reader knows more about basic PHP operators than GlobalFunctions.php. Considering the two bugs discussed above, this seems like a fair assumption. If you don't know PHP, you shouldn't be writing MediaWiki code.
2008-11-01 23:20:25 +00:00
if ( isset( $this->mTemplateIds[$ns] ) ) {
$this->mTemplateIds[$ns] = $dbks + $this->mTemplateIds[$ns];
} else {
$this->mTemplateIds[$ns] = $dbks;
}
}
* Put parser output file version tracking to core * Added some ParserOutput accessors * A few cleanups to fetchFile()
2011-03-23 17:35:40 +00:00
// File versioning...
FU r84610: Renamed getImageTimeKeys() to getFileSearchOptions()
2011-09-14 19:05:43 +00:00
foreach ( (array)$parserOutput->getFileSearchOptions() as $dbk => $data ) {
* Put parser output file version tracking to core * Added some ParserOutput accessors * A few cleanups to fetchFile()
2011-03-23 17:35:40 +00:00
$this->mImageTimeKeys[$dbk] = $data;
}
Basic integrated audio/video support, with Ogg implementation. * JavaScript video player based loosely on Greg Maxwell's player * Image page text snippet customisation * Abstraction of transform parameters in the parser. Introduced Linker::makeImageLink2(). * Made canRender(), mustRender() depend on file, not just on handler. Moved width=0, height=0 checking to ImageHandler::canRender(), since audio streams have width=height=0 but should be rendered. Also: * Automatic upgrade for oldimage rows on image page view, allows media handler selection based on oi_*_mime * oi_*_mime unconditionally referenced, REQUIRES SCHEMA UPGRADE * Don't destroy file info for missing files on upgrade * Simple, centralised extension message file handling * Made MessageCache::loadAllMessages non-static, optimised for repeated-call case due to abuse in User.php * Support for lightweight parser output hooks, with callback whitelist for security * Moved Linker::formatSize() to Language, to join the new formatTimePeriod() and formatBitrate() * Introduced MagicWordArray, regex capture trick requires that magic word IDs DO NOT CONTAIN HYPHENS.
2007-08-15 10:50:09 +00:00
* Added __HIDDENCAT__ feature, to hide categories from the box at the bottom of the member pages depending on special text on the category page. * Added page_props backend, generic parser-driven page properties which, when changed, invalidate the cache of backlinked pages on any links table. Could be used to add overlays to images depending on their deletion status, or to add icons to articles based on category membership. * Refactored double-underscore handling in the parser. * Moved CoreParserFunctions registration to CoreParserFunctions.
2008-02-20 08:53:12 +00:00
// Hooks registered in the object
Deprecate the ParserOutputHook functionality These hooks should be implemented in the OutputPageParserOutput hook instead. Bug: T292321 Change-Id: Ib6f457596ea9d193bc03e15a48f135db4f4a6b27
2022-02-08 16:12:38 +00:00
// Deprecated! See T292321; should be done in the OutputPageParserOutput
// hook instead.
Use name constants to access config settings. Use name constants instead of string literals in calls to Config::get and ServiceOptions::get, when referring to core configuration variables. This protects against typos and makes the decumentation and schema declaration of the config settings discoverable. This is the first batch, only touching files directly under /includes/ Change-Id: I7252e636c7c86d950d9257b33491af492c6dd5eb
2022-04-01 15:58:32 +00:00
$parserOutputHooks = $this->getConfig()->get( MainConfigNames::ParserOutputHooks );
Basic integrated audio/video support, with Ogg implementation. * JavaScript video player based loosely on Greg Maxwell's player * Image page text snippet customisation * Abstraction of transform parameters in the parser. Introduced Linker::makeImageLink2(). * Made canRender(), mustRender() depend on file, not just on handler. Moved width=0, height=0 checking to ImageHandler::canRender(), since audio streams have width=height=0 but should be rendered. Also: * Automatic upgrade for oldimage rows on image page view, allows media handler selection based on oi_*_mime * oi_*_mime unconditionally referenced, REQUIRES SCHEMA UPGRADE * Don't destroy file info for missing files on upgrade * Simple, centralised extension message file handling * Made MessageCache::loadAllMessages non-static, optimised for repeated-call case due to abuse in User.php * Support for lightweight parser output hooks, with callback whitelist for security * Moved Linker::formatSize() to Language, to join the new formatTimePeriod() and formatBitrate() * Introduced MagicWordArray, regex capture trick requires that magic word IDs DO NOT CONTAIN HYPHENS.
2007-08-15 10:50:09 +00:00
foreach ( $parserOutput->getOutputHooks() as $hookInfo ) {
Use name constants to access config settings. Use name constants instead of string literals in calls to Config::get and ServiceOptions::get, when referring to core configuration variables. This protects against typos and makes the decumentation and schema declaration of the config settings discoverable. This is the first batch, only touching files directly under /includes/ Change-Id: I7252e636c7c86d950d9257b33491af492c6dd5eb
2022-04-01 15:58:32 +00:00
[ $hookName, $data ] = $hookInfo;
Use config instead of globals for OutputPage Change-Id: I5e0ebc173631d1d1052de7ccee4ef839c7c1767f
2014-08-23 08:52:41 +00:00
if ( isset( $parserOutputHooks[$hookName] ) ) {
Comments and code cleanup in OutputPage Change-Id: I8c44f30801792e8b84ab3d556cb5b1b5feb9baf3
2018-08-02 18:26:55 +00:00
$parserOutputHooks[$hookName]( $this, $parserOutput, $data );
Basic integrated audio/video support, with Ogg implementation. * JavaScript video player based loosely on Greg Maxwell's player * Image page text snippet customisation * Abstraction of transform parameters in the parser. Introduced Linker::makeImageLink2(). * Made canRender(), mustRender() depend on file, not just on handler. Moved width=0, height=0 checking to ImageHandler::canRender(), since audio streams have width=height=0 but should be rendered. Also: * Automatic upgrade for oldimage rows on image page view, allows media handler selection based on oi_*_mime * oi_*_mime unconditionally referenced, REQUIRES SCHEMA UPGRADE * Don't destroy file info for missing files on upgrade * Simple, centralised extension message file handling * Made MessageCache::loadAllMessages non-static, optimised for repeated-call case due to abuse in User.php * Support for lightweight parser output hooks, with callback whitelist for security * Moved Linker::formatSize() to Language, to join the new formatTimePeriod() and formatBitrate() * Introduced MagicWordArray, regex capture trick requires that magic word IDs DO NOT CONTAIN HYPHENS.
2007-08-15 10:50:09 +00:00
}
}
Include JS variable for NewPP report Adapted from reverted commit b7c4c8717f9. Bug: T110763 Change-Id: If249b679c534879bfac622592a1d2fa913a0cf9d
2016-11-10 20:29:27 +00:00
// Enable OOUI if requested via ParserOutput
Allow to enable OOUI via a parser tag extension This change adds the possibility to enable OOUI out of the parser, which enabled parser tag functions to easily enable OOUI, if they need it, for every page view out of the function that handles the parser tag. Bug: T106949 Change-Id: If1e139d4f07be98e418e11470794ea42e8a9b2eb
2015-07-25 15:32:08 +00:00
if ( $parserOutput->getEnableOOUI() ) {
$this->enableOOUI();
}
Include JS variable for NewPP report Adapted from reverted commit b7c4c8717f9. Bug: T110763 Change-Id: If249b679c534879bfac622592a1d2fa913a0cf9d
2016-11-10 20:29:27 +00:00
// Include parser limit report
Give skins more flexibility over table of contents render * Do not store table of contents in parser output * Instead inject table of contents via strpos where needed inside Article based on Skin "toc" option * Use <mw:tocplace> as a TOC placeholder; for Parsoid compatibility this will be replaced with a <meta> tag in a followup patch. Bug: T287767 Change-Id: I44045b3b9e78e7ab793da3f37e3c0dbc91cd7d39
2021-09-15 01:00:06 +00:00
// FIXME: This should append, rather than overwrite, or else this
// data should be injected into the OutputPage like is done for the
Generate/set/get TOCData/SectionMetadata objects instead of arrays * ParserOutput::setSections()/::getSections() are expected to be deprecated. Uses in extensions and skins will need to be migrated in follow up patches once the new interface has stabilized. * In the skins code, the metadata is converted back to an array. Downstream skin TOC consumers will need to be migrated as well before we can remove the toLegacy() conversion. * Fixed SerializationTestTrait's validation method - Not sure if this is overkill but should handle all future complex objects we might stuff into the ParserCache. * This patch emits a backward-compatible Sections property in order to avoid changing the parser cache serialization format. T327439 has been filed to eventually use the JsonCodec support for object serialization, but for this initial patch it makes sense to avoid the need for a concurrent ParserCache format migration by using a backward-compatible serialization. * TOCData is nullable because the intent is that ParserOutput::setTOCData() is MW_MERGE_STRATEGY_WRITE_ONCE; that is, only the top-level fragment composing a page will set the TOCData. This will be enforced in the future via wfDeprecated() (T327429), but again our first patch is as backward-compatible as possible. Bug: T296025 Depends-On: I1b267d23cf49d147c5379b914531303744481b68 Co-Authored-By: C. Scott Ananian <cananian@wikimedia.org> Co-Authored-By: Subramanya Sastry <ssastry@wikimedia.org> Change-Id: I8329864535f0b1dd5f9163868a08d6cb1ffcb78f
2022-09-01 23:07:29 +00:00
// other page-level things (like OutputPage::setTOCData()).
Include JS variable for NewPP report Adapted from reverted commit b7c4c8717f9. Bug: T110763 Change-Id: If249b679c534879bfac622592a1d2fa913a0cf9d
2016-11-10 20:29:27 +00:00
if ( !$this->limitReportJSData ) {
$this->limitReportJSData = $parserOutput->getLimitReportJSData();
}
(bug 45535) Hook for changing language links. This adds a new hook called LanguageLinks which is called whenever a list of language links is returned to the user. This gives extensions the option to manipulate the links on the fly. Note that this change only covers the language links used in OutputPage and by ApiParse. Adapting ApiQueryLangLinks is left as a follow-up task. As explained on bugzilla, this is a precondition to allowing Wikibase/Wikidata to update languagelinks without forcing a (redundant) re-parse of the page content. This change also introduces the notion of link flags that can be used to associate flags with language links. This will be integrated with ParserOutput and OutputPage in a follow-up. Change-Id: Iaec0faa131413a291fc8f77496e4f371addb3b99
2013-04-19 11:19:06 +00:00
// Link flags are ignored for now, but may in the future be
// used to mark individual language links.
Convert all array() syntax to [] Per wikitech-l consensus: https://lists.wikimedia.org/pipermail/wikitech-l/2016-February/084821.html Notes: * Disabled CallTimePassByReference due to false positives (T127163) Change-Id: I2c8ce713ce6600a0bb7bf67537c87044c7a45c4b
2016-02-17 09:09:32 +00:00
$linkFlags = [];
Hooks::run() call site migration Migrate all callers of Hooks::run() to use the new HookContainer/HookRunner system. General principles: * Use DI if it is already used. We're not changing the way state is managed in this patch. * HookContainer is always injected, not HookRunner. HookContainer is a service, it's a more generic interface, it is the only thing that provides isRegistered() which is needed in some cases, and a HookRunner can be efficiently constructed from it (confirmed by benchmark). Because HookContainer is needed for object construction, it is also needed by all factories. * "Ask your friendly local base class". Big hierarchies like SpecialPage and ApiBase have getHookContainer() and getHookRunner() methods in the base class, and classes that extend that base class are not expected to know or care where the base class gets its HookContainer from. * ProtectedHookAccessorTrait provides protected getHookContainer() and getHookRunner() methods, getting them from the global service container. The point of this is to ease migration to DI by ensuring that call sites ask their local friendly base class rather than getting a HookRunner from the service container directly. * Private $this->hookRunner. In some smaller classes where accessor methods did not seem warranted, there is a private HookRunner property which is accessed directly. Very rarely (two cases), there is a protected property, for consistency with code that conventionally assumes protected=private, but in cases where the class might actually be overridden, a protected accessor is preferred over a protected property. * The last resort: Hooks::runner(). Mostly for static, file-scope and global code. In a few cases it was used for objects with broken construction schemes, out of horror or laziness. Constructors with new required arguments: * AuthManager * BadFileLookup * BlockManager * ClassicInterwikiLookup * ContentHandlerFactory * ContentSecurityPolicy * DefaultOptionsManager * DerivedPageDataUpdater * FullSearchResultWidget * HtmlCacheUpdater * LanguageFactory * LanguageNameUtils * LinkRenderer * LinkRendererFactory * LocalisationCache * MagicWordFactory * MessageCache * NamespaceInfo * PageEditStash * PageHandlerFactory * PageUpdater * ParserFactory * PermissionManager * RevisionStore * RevisionStoreFactory * SearchEngineConfig * SearchEngineFactory * SearchFormWidget * SearchNearMatcher * SessionBackend * SpecialPageFactory * UserNameUtils * UserOptionsManager * WatchedItemQueryService * WatchedItemStore Constructors with new optional arguments: * DefaultPreferencesFactory * Language * LinkHolderArray * MovePage * Parser * ParserCache * PasswordReset * Router setHookContainer() now required after construction: * AuthenticationProvider * ResourceLoaderModule * SearchEngine Change-Id: Id442b0dbe43aba84bd5cf801d86dedc768b082c7
2020-03-19 02:42:09 +00:00
$this->getHookRunner()->onLanguageLinks( $this->getTitle(), $this->mLanguageLinks, $linkFlags );
Use ParserOutputFlags::SHOW_TOC in OutputPage Parser output flags are intended to be "set only". That means that extensions can *set* the SHOW_TOC flag, but once set SHOW_TOC can't be reset, and the TOC will be shown in the result if *any* of the merged ParserOutputs on the OutputPage contain the SHOW_TOC flag set. That's consistent with prior merge behavior. There's a possibility extensions are playing other games with the Table Of Contents (TOC) in the OutputPageParserOutput hook; be compatible with prior behavior for now but we may need to provide additional functionality (perhaps a way to merge new section data) if needed. Bug: T310083 Change-Id: I0909ac85c6c785d9089b077a16923c61d6a09996
2022-06-09 12:59:50 +00:00
Hooks::run() call site migration Migrate all callers of Hooks::run() to use the new HookContainer/HookRunner system. General principles: * Use DI if it is already used. We're not changing the way state is managed in this patch. * HookContainer is always injected, not HookRunner. HookContainer is a service, it's a more generic interface, it is the only thing that provides isRegistered() which is needed in some cases, and a HookRunner can be efficiently constructed from it (confirmed by benchmark). Because HookContainer is needed for object construction, it is also needed by all factories. * "Ask your friendly local base class". Big hierarchies like SpecialPage and ApiBase have getHookContainer() and getHookRunner() methods in the base class, and classes that extend that base class are not expected to know or care where the base class gets its HookContainer from. * ProtectedHookAccessorTrait provides protected getHookContainer() and getHookRunner() methods, getting them from the global service container. The point of this is to ease migration to DI by ensuring that call sites ask their local friendly base class rather than getting a HookRunner from the service container directly. * Private $this->hookRunner. In some smaller classes where accessor methods did not seem warranted, there is a private HookRunner property which is accessed directly. Very rarely (two cases), there is a protected property, for consistency with code that conventionally assumes protected=private, but in cases where the class might actually be overridden, a protected accessor is preferred over a protected property. * The last resort: Hooks::runner(). Mostly for static, file-scope and global code. In a few cases it was used for objects with broken construction schemes, out of horror or laziness. Constructors with new required arguments: * AuthManager * BadFileLookup * BlockManager * ClassicInterwikiLookup * ContentHandlerFactory * ContentSecurityPolicy * DefaultOptionsManager * DerivedPageDataUpdater * FullSearchResultWidget * HtmlCacheUpdater * LanguageFactory * LanguageNameUtils * LinkRenderer * LinkRendererFactory * LocalisationCache * MagicWordFactory * MessageCache * NamespaceInfo * PageEditStash * PageHandlerFactory * PageUpdater * ParserFactory * PermissionManager * RevisionStore * RevisionStoreFactory * SearchEngineConfig * SearchEngineFactory * SearchFormWidget * SearchNearMatcher * SessionBackend * SpecialPageFactory * UserNameUtils * UserOptionsManager * WatchedItemQueryService * WatchedItemStore Constructors with new optional arguments: * DefaultPreferencesFactory * Language * LinkHolderArray * MovePage * Parser * ParserCache * PasswordReset * Router setHookContainer() now required after construction: * AuthenticationProvider * ResourceLoaderModule * SearchEngine Change-Id: Id442b0dbe43aba84bd5cf801d86dedc768b082c7
2020-03-19 02:42:09 +00:00
$this->getHookRunner()->onOutputPageParserOutput( $this, $parserOutput );
Move loading of mediawiki.toc from Parser to Skin This was the only addModules() call ever to be inside Parser. Introduced in a54ef1a203d. Prior to that, mediawiki.toc had always been loaded by OutputPage (via mediawiki.util; and before that, via wikibits). This patch restores that, and also fixes T130632 by making OutputPage get it from the Skin, instead of hardcoding this somewhere in addParserOutput(). * Remove deprecated method OutputPage::enableTOC(). * Move mEnableTOC to addParserOutputText(). Bug: T130632 Change-Id: Iaad84d241a4c4348c712ac1087a664b8c9c46da4
2017-05-20 13:39:40 +00:00
// This check must be after 'OutputPageParserOutput' runs in addParserOutputMetadata
// so that extensions may modify ParserOutput to toggle TOC.
// This cannot be moved to addParserOutputText because that is not
// called by EditPage for Preview.
Use ParserOutputFlags::SHOW_TOC in OutputPage Parser output flags are intended to be "set only". That means that extensions can *set* the SHOW_TOC flag, but once set SHOW_TOC can't be reset, and the TOC will be shown in the result if *any* of the merged ParserOutputs on the OutputPage contain the SHOW_TOC flag set. That's consistent with prior merge behavior. There's a possibility extensions are playing other games with the Table Of Contents (TOC) in the OutputPageParserOutput hook; be compatible with prior behavior for now but we may need to provide additional functionality (perhaps a way to merge new section data) if needed. Bug: T310083 Change-Id: I0909ac85c6c785d9089b077a16923c61d6a09996
2022-06-09 12:59:50 +00:00
// T294950/T293513: ParserOutput::getTOCHTML() will be
Generate/set/get TOCData/SectionMetadata objects instead of arrays * ParserOutput::setSections()/::getSections() are expected to be deprecated. Uses in extensions and skins will need to be migrated in follow up patches once the new interface has stabilized. * In the skins code, the metadata is converted back to an array. Downstream skin TOC consumers will need to be migrated as well before we can remove the toLegacy() conversion. * Fixed SerializationTestTrait's validation method - Not sure if this is overkill but should handle all future complex objects we might stuff into the ParserCache. * This patch emits a backward-compatible Sections property in order to avoid changing the parser cache serialization format. T327439 has been filed to eventually use the JsonCodec support for object serialization, but for this initial patch it makes sense to avoid the need for a concurrent ParserCache format migration by using a backward-compatible serialization. * TOCData is nullable because the intent is that ParserOutput::setTOCData() is MW_MERGE_STRATEGY_WRITE_ONCE; that is, only the top-level fragment composing a page will set the TOCData. This will be enforced in the future via wfDeprecated() (T327429), but again our first patch is as backward-compatible as possible. Bug: T296025 Depends-On: I1b267d23cf49d147c5379b914531303744481b68 Co-Authored-By: C. Scott Ananian <cananian@wikimedia.org> Co-Authored-By: Subramanya Sastry <ssastry@wikimedia.org> Change-Id: I8329864535f0b1dd5f9163868a08d6cb1ffcb78f
2022-09-01 23:07:29 +00:00
// replaced by ParserOutput::getTOCData(), and
Use ParserOutputFlags::SHOW_TOC in OutputPage Parser output flags are intended to be "set only". That means that extensions can *set* the SHOW_TOC flag, but once set SHOW_TOC can't be reset, and the TOC will be shown in the result if *any* of the merged ParserOutputs on the OutputPage contain the SHOW_TOC flag set. That's consistent with prior merge behavior. There's a possibility extensions are playing other games with the Table Of Contents (TOC) in the OutputPageParserOutput hook; be compatible with prior behavior for now but we may need to provide additional functionality (perhaps a way to merge new section data) if needed. Bug: T310083 Change-Id: I0909ac85c6c785d9089b077a16923c61d6a09996
2022-06-09 12:59:50 +00:00
// ParserOutputFlags::SHOW_TOC is used to indicate whether the TOC
// should be shown (or hidden) in the output.
$this->mEnableTOC = $this->mEnableTOC ||
$parserOutput->getOutputFlag( ParserOutputFlags::SHOW_TOC );
// But extensions used to be able to modify ParserOutput::setTOCHTML()
// to toggle TOC in the OutputPageParserOutput hook; so for backward
// compatibility check to see if that happened.
$isTocPresent = (bool)$parserOutput->getTOCHTML();
if ( $isTocPresent && !$this->mEnableTOC ) {
// Eventually we'll emit a deprecation message here (T293513)
Move loading of mediawiki.toc from Parser to Skin This was the only addModules() call ever to be inside Parser. Introduced in a54ef1a203d. Prior to that, mediawiki.toc had always been loaded by OutputPage (via mediawiki.util; and before that, via wikibits). This patch restores that, and also fixes T130632 by making OutputPage get it from the Skin, instead of hardcoding this somewhere in addParserOutput(). * Remove deprecated method OutputPage::enableTOC(). * Move mEnableTOC to addParserOutputText(). Bug: T130632 Change-Id: Iaad84d241a4c4348c712ac1087a664b8c9c46da4
2017-05-20 13:39:40 +00:00
$this->mEnableTOC = true;
}
fix for broken categories on preview
2006-01-01 20:08:08 +00:00
}
* s~\t+$~~
2006-01-07 13:31:29 +00:00
Add some documentation to OutputPage.php and specify public/private for most of its member functions (practically all public anyway). And replace one gratuitous use of strcmp with good old ===.
2006-11-07 05:37:31 +00:00
/**
OutputPage: Add addParserOutputContent() for more fine-grained control We previously had addParserOutput(), which added everything and did some other magic, and addParserOutputNoText() which, as the same says, added everything but the text. I renamed addParserOutputNoText() to addParserOutputMetadata() and created two more functions: * addParserOutputText(): This is almost identical to adding the raw HTML, but calls the OutputPageBeforeHTML hook like other addParserOutput*() methods. * addParserOutputContent(): Like addParserOutputText(), but also adds the ResourceLoader modules and variables associated with the parser output. This is important especially for some extensions like TemplateData or SyntaxHighlight which add styles to the page to enhance the display. Change-Id: Iead541886fd1ccdbdf1cb06af71b34cd04644985
2014-05-08 20:10:13 +00:00
* Add the HTML and enhancements for it (like ResourceLoader modules) associated with a
* ParserOutput object, without any other metadata.
*
* @since 1.24
* @param ParserOutput $parserOutput
Use ParserOutput stateless transforms We still set the state in many cases for benefit of extensions, but all calls within core should no longer be using non-default state. Change-Id: I78b62ec33fcb8273acb9b3b4e9012215442be94c Depends-On: I140ff32373430b61b92226689ef9b58cca317450
2017-11-22 20:07:51 +00:00
* @param array $poOptions Options to ParserOutput::getText()
OutputPage: Add addParserOutputContent() for more fine-grained control We previously had addParserOutput(), which added everything and did some other magic, and addParserOutputNoText() which, as the same says, added everything but the text. I renamed addParserOutputNoText() to addParserOutputMetadata() and created two more functions: * addParserOutputText(): This is almost identical to adding the raw HTML, but calls the OutputPageBeforeHTML hook like other addParserOutput*() methods. * addParserOutputContent(): Like addParserOutputText(), but also adds the ResourceLoader modules and variables associated with the parser output. This is important especially for some extensions like TemplateData or SyntaxHighlight which add styles to the page to enhance the display. Change-Id: Iead541886fd1ccdbdf1cb06af71b34cd04644985
2014-05-08 20:10:13 +00:00
*/
Add type hint for ParserOutput EditPage::getPreviewLimitReport is called by EditPage::showEditForm with $output = null. Specify this in the @param tag and allow this by a default value. Change-Id: Iec8905aab736a1f254a57853c7cab935d008653e
2018-07-27 20:20:39 +00:00
public function addParserOutputContent( ParserOutput $parserOutput, $poOptions = [] ) {
Use ParserOutput stateless transforms We still set the state in many cases for benefit of extensions, but all calls within core should no longer be using non-default state. Change-Id: I78b62ec33fcb8273acb9b3b4e9012215442be94c Depends-On: I140ff32373430b61b92226689ef9b58cca317450
2017-11-22 20:07:51 +00:00
$this->addParserOutputText( $parserOutput, $poOptions );
OutputPage: Add addParserOutputContent() for more fine-grained control We previously had addParserOutput(), which added everything and did some other magic, and addParserOutputNoText() which, as the same says, added everything but the text. I renamed addParserOutputNoText() to addParserOutputMetadata() and created two more functions: * addParserOutputText(): This is almost identical to adding the raw HTML, but calls the OutputPageBeforeHTML hook like other addParserOutput*() methods. * addParserOutputContent(): Like addParserOutputText(), but also adds the ResourceLoader modules and variables associated with the parser output. This is important especially for some extensions like TemplateData or SyntaxHighlight which add styles to the page to enhance the display. Change-Id: Iead541886fd1ccdbdf1cb06af71b34cd04644985
2014-05-08 20:10:13 +00:00
$this->addModules( $parserOutput->getModules() );
$this->addModuleStyles( $parserOutput->getModuleStyles() );
$this->addJsConfigVars( $parserOutput->getJsConfigVars() );
}
/**
* Add the HTML associated with a ParserOutput object, without any metadata.
*
* @since 1.24
* @param ParserOutput $parserOutput
Use ParserOutput stateless transforms We still set the state in many cases for benefit of extensions, but all calls within core should no longer be using non-default state. Change-Id: I78b62ec33fcb8273acb9b3b4e9012215442be94c Depends-On: I140ff32373430b61b92226689ef9b58cca317450
2017-11-22 20:07:51 +00:00
* @param array $poOptions Options to ParserOutput::getText()
OutputPage: Add addParserOutputContent() for more fine-grained control We previously had addParserOutput(), which added everything and did some other magic, and addParserOutputNoText() which, as the same says, added everything but the text. I renamed addParserOutputNoText() to addParserOutputMetadata() and created two more functions: * addParserOutputText(): This is almost identical to adding the raw HTML, but calls the OutputPageBeforeHTML hook like other addParserOutput*() methods. * addParserOutputContent(): Like addParserOutputText(), but also adds the ResourceLoader modules and variables associated with the parser output. This is important especially for some extensions like TemplateData or SyntaxHighlight which add styles to the page to enhance the display. Change-Id: Iead541886fd1ccdbdf1cb06af71b34cd04644985
2014-05-08 20:10:13 +00:00
*/
Add type hint for ParserOutput EditPage::getPreviewLimitReport is called by EditPage::showEditForm with $output = null. Specify this in the @param tag and allow this by a default value. Change-Id: Iec8905aab736a1f254a57853c7cab935d008653e
2018-07-27 20:20:39 +00:00
public function addParserOutputText( ParserOutput $parserOutput, $poOptions = [] ) {
OutputPage::addParserOutputText(): use default ParserOutput options from skin This addresses the common case patched by I530d71d0f9279b40a263cd62467d3ef8c76975c3, If6267f3389b166043fc94d7f952bc54122b1a378 and probably the code in Article.php from I44045b3b9e78e7ab793da3f37e3c0dbc91cd7d39 by ensuring that "injectTOC" in the options passed to ParserOutput::getText() defaults to the correct value based on the skin being used by OutputPage. Bug: T317333 Change-Id: Ica30569efbb5730eff5b807e8fc34beb2e13e74f
2022-09-08 19:29:52 +00:00
// Add default options from the skin
$skin = $this->getSkin();
$skinOptions = $skin->getOptions();
$poOptions += [
'skin' => $skin,
'injectTOC' => $skinOptions['toc'],
];
Use ParserOutput stateless transforms We still set the state in many cases for benefit of extensions, but all calls within core should no longer be using non-default state. Change-Id: I78b62ec33fcb8273acb9b3b4e9012215442be94c Depends-On: I140ff32373430b61b92226689ef9b58cca317450
2017-11-22 20:07:51 +00:00
$text = $parserOutput->getText( $poOptions );
Hooks::run() call site migration Migrate all callers of Hooks::run() to use the new HookContainer/HookRunner system. General principles: * Use DI if it is already used. We're not changing the way state is managed in this patch. * HookContainer is always injected, not HookRunner. HookContainer is a service, it's a more generic interface, it is the only thing that provides isRegistered() which is needed in some cases, and a HookRunner can be efficiently constructed from it (confirmed by benchmark). Because HookContainer is needed for object construction, it is also needed by all factories. * "Ask your friendly local base class". Big hierarchies like SpecialPage and ApiBase have getHookContainer() and getHookRunner() methods in the base class, and classes that extend that base class are not expected to know or care where the base class gets its HookContainer from. * ProtectedHookAccessorTrait provides protected getHookContainer() and getHookRunner() methods, getting them from the global service container. The point of this is to ease migration to DI by ensuring that call sites ask their local friendly base class rather than getting a HookRunner from the service container directly. * Private $this->hookRunner. In some smaller classes where accessor methods did not seem warranted, there is a private HookRunner property which is accessed directly. Very rarely (two cases), there is a protected property, for consistency with code that conventionally assumes protected=private, but in cases where the class might actually be overridden, a protected accessor is preferred over a protected property. * The last resort: Hooks::runner(). Mostly for static, file-scope and global code. In a few cases it was used for objects with broken construction schemes, out of horror or laziness. Constructors with new required arguments: * AuthManager * BadFileLookup * BlockManager * ClassicInterwikiLookup * ContentHandlerFactory * ContentSecurityPolicy * DefaultOptionsManager * DerivedPageDataUpdater * FullSearchResultWidget * HtmlCacheUpdater * LanguageFactory * LanguageNameUtils * LinkRenderer * LinkRendererFactory * LocalisationCache * MagicWordFactory * MessageCache * NamespaceInfo * PageEditStash * PageHandlerFactory * PageUpdater * ParserFactory * PermissionManager * RevisionStore * RevisionStoreFactory * SearchEngineConfig * SearchEngineFactory * SearchFormWidget * SearchNearMatcher * SessionBackend * SpecialPageFactory * UserNameUtils * UserOptionsManager * WatchedItemQueryService * WatchedItemStore Constructors with new optional arguments: * DefaultPreferencesFactory * Language * LinkHolderArray * MovePage * Parser * ParserCache * PasswordReset * Router setHookContainer() now required after construction: * AuthenticationProvider * ResourceLoaderModule * SearchEngine Change-Id: Id442b0dbe43aba84bd5cf801d86dedc768b082c7
2020-03-19 02:42:09 +00:00
$this->getHookRunner()->onOutputPageBeforeHTML( $this, $text );
OutputPage: Add addParserOutputContent() for more fine-grained control We previously had addParserOutput(), which added everything and did some other magic, and addParserOutputNoText() which, as the same says, added everything but the text. I renamed addParserOutputNoText() to addParserOutputMetadata() and created two more functions: * addParserOutputText(): This is almost identical to adding the raw HTML, but calls the OutputPageBeforeHTML hook like other addParserOutput*() methods. * addParserOutputContent(): Like addParserOutputText(), but also adds the ResourceLoader modules and variables associated with the parser output. This is important especially for some extensions like TemplateData or SyntaxHighlight which add styles to the page to enhance the display. Change-Id: Iead541886fd1ccdbdf1cb06af71b34cd04644985
2014-05-08 20:10:13 +00:00
$this->addHTML( $text );
}
/**
* Add everything from a ParserOutput object.
Step two in fixing OutputPage's documentation, more or less the same as r61647
2010-01-29 21:32:45 +00:00
*
Fixed some @params documentation Swapped some "$var type" to "type $var" or added missing types before the $var. Changed some other types to match the more common spelling. Makes beginning of some text in captial. Change-Id: I64e8cfe478cb0ba438f40b0631d6e9049cdab567
2014-04-10 18:50:10 +00:00
* @param ParserOutput $parserOutput
Use ParserOutput stateless transforms We still set the state in many cases for benefit of extensions, but all calls within core should no longer be using non-default state. Change-Id: I78b62ec33fcb8273acb9b3b4e9012215442be94c Depends-On: I140ff32373430b61b92226689ef9b58cca317450
2017-11-22 20:07:51 +00:00
* @param array $poOptions Options to ParserOutput::getText()
Add some documentation to OutputPage.php and specify public/private for most of its member functions (practically all public anyway). And replace one gratuitous use of strcmp with good old ===.
2006-11-07 05:37:31 +00:00
*/
OutputPage: Add visibility to various methods & do some micro-optimizations Methods that visibility was added to are; `addMeta()`, `addLink()`, `setCanonicalUrl()`, `addScript()`, `getHeadItemsArray()`, `addParserOutput()`, `getCacheVaryCookies()` and `haveCacheVaryCookies()`. Last but not lease, did a few micro-optimizations to `addMeta()` and `addLink()`. Change-Id: I94d037a5edc7131627724fd1d864000128077b0c
2019-09-14 10:26:31 +00:00
public function addParserOutput( ParserOutput $parserOutput, $poOptions = [] ) {
OutputPage: Add addParserOutputContent() for more fine-grained control We previously had addParserOutput(), which added everything and did some other magic, and addParserOutputNoText() which, as the same says, added everything but the text. I renamed addParserOutputNoText() to addParserOutputMetadata() and created two more functions: * addParserOutputText(): This is almost identical to adding the raw HTML, but calls the OutputPageBeforeHTML hook like other addParserOutput*() methods. * addParserOutputContent(): Like addParserOutputText(), but also adds the ResourceLoader modules and variables associated with the parser output. This is important especially for some extensions like TemplateData or SyntaxHighlight which add styles to the page to enhance the display. Change-Id: Iead541886fd1ccdbdf1cb06af71b34cd04644985
2014-05-08 20:10:13 +00:00
$this->addParserOutputMetadata( $parserOutput );
Use ParserOutput stateless transforms We still set the state in many cases for benefit of extensions, but all calls within core should no longer be using non-default state. Change-Id: I78b62ec33fcb8273acb9b3b4e9012215442be94c Depends-On: I140ff32373430b61b92226689ef9b58cca317450
2017-11-22 20:07:51 +00:00
$this->addParserOutputText( $parserOutput, $poOptions );
remove wfdebug in pagehistory set ETag header on article views
2005-07-01 00:03:31 +00:00
}
Separate the HTML guts of the user login form into a template to keep the code more or less legibile.
2004-11-25 06:04:16 +00:00
/**
* Add the output of a QuickTemplate to the output buffer
Add some documentation to OutputPage.php and specify public/private for most of its member functions (practically all public anyway). And replace one gratuitous use of strcmp with good old ===.
2006-11-07 05:37:31 +00:00
*
Add missing & to @param documentation to match functon call Change-Id: I81e68310abcbc59964b22e0e74842d509f6b1fb9
2017-08-11 00:23:16 +00:00
* @param QuickTemplate &$template
Separate the HTML guts of the user login form into a template to keep the code more or less legibile.
2004-11-25 06:04:16 +00:00
*/
Add some documentation to OutputPage.php and specify public/private for most of its member functions (practically all public anyway). And replace one gratuitous use of strcmp with good old ===.
2006-11-07 05:37:31 +00:00
public function addTemplate( &$template ) {
New getHTML() method for QuickTemplate to get the HTML of a template. OutputPage's addTemplate() is now a wrapper around QuickTemplate. This allows more flexible usage of templated HTML, as required by some third-party extensions. Change-Id: I943e8e50438c716b7b56d2f908da38a4bf73d9e2
2013-11-21 07:47:39 +00:00
$this->addHTML( $template->getHTML() );
Separate the HTML guts of the user login form into a template to keep the code more or less legibile.
2004-11-25 06:04:16 +00:00
}
parse() function, simplifying calling of the parser by supplying options from OutputPage
2004-11-29 11:26:24 +00:00
Deprecate OutputPage::parse() and OutputPage::parseInline() The OutputPage::parse() and OutputPage::parseInline() methods behave misleadingly different from the OutputPage::addWikitext*() methods: they don't tidy their output, they have different defaults for interface/content language selection, and they (sometimes) add wrapper divs. Deprecate these and add new methods with tidy output, clear language selection, and consistent defaults: OutputPage::parseAsContent(), OutputPage::parseAsInterface(), and OutputPage::parseInlineAsInterface(). Unify the implementation of the parse* methods with the addWikiText* methods, to reduce the likelihood that the behavior will diverge again in the future. Bug: T198214 Change-Id: Ica79c2acbc542ef37f971c0be2582ae771a23bd0
2018-10-26 15:14:01 +00:00
/**
* Parse wikitext *in the page content language* and return the HTML.
* The result will be language-converted to the user's preferred variant.
* Output will be tidy.
*
* @param string $text Wikitext in the page content language
* @param bool $linestart Is this the start of a line? (Defaults to true)
* @throws MWException
* @return string HTML
Update deprecations after OutputPage::parse*As*() backport to 1.32 Originally landed in 1.33 in a3f63785eebfa2f729d42da880b6722736ef1e7d. Backported to 1.32 in cff9ab4b5c26dc87137aa2c04efd5d1f31226fe8. Change-Id: I402c31b57904f6a6fd0b3a2c32a1c0a73c553254
2018-11-01 20:30:57 +00:00
* @since 1.32
Deprecate OutputPage::parse() and OutputPage::parseInline() The OutputPage::parse() and OutputPage::parseInline() methods behave misleadingly different from the OutputPage::addWikitext*() methods: they don't tidy their output, they have different defaults for interface/content language selection, and they (sometimes) add wrapper divs. Deprecate these and add new methods with tidy output, clear language selection, and consistent defaults: OutputPage::parseAsContent(), OutputPage::parseAsInterface(), and OutputPage::parseInlineAsInterface(). Unify the implementation of the parse* methods with the addWikiText* methods, to reduce the likelihood that the behavior will diverge again in the future. Bug: T198214 Change-Id: Ica79c2acbc542ef37f971c0be2582ae771a23bd0
2018-10-26 15:14:01 +00:00
*/
public function parseAsContent( $text, $linestart = true ) {
return $this->parseInternal(
Remove deprecated non-tidy OutputPage::parse/parseInline methods The non-tidy parsing modes were deprecated in 1.32. Remove the deprecated parse/parseInline methods which use them, as part of a general clean up of nontidy parsing modes. Bug: T198214 Change-Id: I04ea82dd20eac8b0ce07a8d0e6b0bf2c4b03dbe8
2020-01-15 19:03:33 +00:00
$text, $this->getTitle(), $linestart, /*interface*/false
Fix OutputPage::parseInternal() by stripping <div> wrapper We should probably strip the <div> wrapper in OutputPage::parse() as well: this behavior was added in 1.30, but it only applies when $interface is false. However, that requires a more careful audit (a lot more places call parse() than parseInline()) and so I'll defer that for now. Change-Id: Iad5412f03af29c04deb653969dd71f6c86f0ae50
2018-10-26 14:05:34 +00:00
)->getText( [
Pass appropriate userLang to ParserOutput::getText() The rules for how ParserOptions::getUserLang() are set are a little baroque. When the 'uselang' option is set on the request, for example, RequestContext::getLanguage() can differ from the user language; we also deliberately force the "user language" to the content language during the post-edit path to avoid having the user language leak into the links table. The changes in this patch make the ParserOutput::getText() calls "more correct" but really they are just removing some uses of the global RequestContext::getMain() from the code path, since in all of these cases the main request context (which is what we use by default) does have the correct language set. For consistency, this patch also adds `allowTOC=false` in many of the places where we already had `enableSectionEditLinks=false`; for example, user interface messages, wikitext content injected on special pages, etc. Change-Id: Ib0d8c5e538e472ad672c45e5b68907242afb1e14
2023-01-25 19:50:36 +00:00
'allowTOC' => false,
Fix OutputPage::parseInternal() by stripping <div> wrapper We should probably strip the <div> wrapper in OutputPage::parse() as well: this behavior was added in 1.30, but it only applies when $interface is false. However, that requires a more careful audit (a lot more places call parse() than parseInline()) and so I'll defer that for now. Change-Id: Iad5412f03af29c04deb653969dd71f6c86f0ae50
2018-10-26 14:05:34 +00:00
'enableSectionEditLinks' => false,
Pass appropriate userLang to ParserOutput::getText() The rules for how ParserOptions::getUserLang() are set are a little baroque. When the 'uselang' option is set on the request, for example, RequestContext::getLanguage() can differ from the user language; we also deliberately force the "user language" to the content language during the post-edit path to avoid having the user language leak into the links table. The changes in this patch make the ParserOutput::getText() calls "more correct" but really they are just removing some uses of the global RequestContext::getMain() from the code path, since in all of these cases the main request context (which is what we use by default) does have the correct language set. For consistency, this patch also adds `allowTOC=false` in many of the places where we already had `enableSectionEditLinks=false`; for example, user interface messages, wikitext content injected on special pages, etc. Change-Id: Ib0d8c5e538e472ad672c45e5b68907242afb1e14
2023-01-25 19:50:36 +00:00
'wrapperDivClass' => '',
'userLang' => $this->getContext()->getLanguage(),
Fix OutputPage::parseInternal() by stripping <div> wrapper We should probably strip the <div> wrapper in OutputPage::parse() as well: this behavior was added in 1.30, but it only applies when $interface is false. However, that requires a more careful audit (a lot more places call parse() than parseInline()) and so I'll defer that for now. Change-Id: Iad5412f03af29c04deb653969dd71f6c86f0ae50
2018-10-26 14:05:34 +00:00
] );
}
Deprecate OutputPage::parse() and OutputPage::parseInline() The OutputPage::parse() and OutputPage::parseInline() methods behave misleadingly different from the OutputPage::addWikitext*() methods: they don't tidy their output, they have different defaults for interface/content language selection, and they (sometimes) add wrapper divs. Deprecate these and add new methods with tidy output, clear language selection, and consistent defaults: OutputPage::parseAsContent(), OutputPage::parseAsInterface(), and OutputPage::parseInlineAsInterface(). Unify the implementation of the parse* methods with the addWikiText* methods, to reduce the likelihood that the behavior will diverge again in the future. Bug: T198214 Change-Id: Ica79c2acbc542ef37f971c0be2582ae771a23bd0
2018-10-26 15:14:01 +00:00
/**
* Parse wikitext *in the user interface language* and return the HTML.
* The result will not be language-converted, as user interface messages
* are already localized into a specific variant.
* Output will be tidy.
*
* @param string $text Wikitext in the user interface language
* @param bool $linestart Is this the start of a line? (Defaults to true)
* @throws MWException
* @return string HTML
Update deprecations after OutputPage::parse*As*() backport to 1.32 Originally landed in 1.33 in a3f63785eebfa2f729d42da880b6722736ef1e7d. Backported to 1.32 in cff9ab4b5c26dc87137aa2c04efd5d1f31226fe8. Change-Id: I402c31b57904f6a6fd0b3a2c32a1c0a73c553254
2018-11-01 20:30:57 +00:00
* @since 1.32
Deprecate OutputPage::parse() and OutputPage::parseInline() The OutputPage::parse() and OutputPage::parseInline() methods behave misleadingly different from the OutputPage::addWikitext*() methods: they don't tidy their output, they have different defaults for interface/content language selection, and they (sometimes) add wrapper divs. Deprecate these and add new methods with tidy output, clear language selection, and consistent defaults: OutputPage::parseAsContent(), OutputPage::parseAsInterface(), and OutputPage::parseInlineAsInterface(). Unify the implementation of the parse* methods with the addWikiText* methods, to reduce the likelihood that the behavior will diverge again in the future. Bug: T198214 Change-Id: Ica79c2acbc542ef37f971c0be2582ae771a23bd0
2018-10-26 15:14:01 +00:00
*/
public function parseAsInterface( $text, $linestart = true ) {
return $this->parseInternal(
Remove deprecated non-tidy OutputPage::parse/parseInline methods The non-tidy parsing modes were deprecated in 1.32. Remove the deprecated parse/parseInline methods which use them, as part of a general clean up of nontidy parsing modes. Bug: T198214 Change-Id: I04ea82dd20eac8b0ce07a8d0e6b0bf2c4b03dbe8
2020-01-15 19:03:33 +00:00
$text, $this->getTitle(), $linestart, /*interface*/true
Deprecate OutputPage::parse() and OutputPage::parseInline() The OutputPage::parse() and OutputPage::parseInline() methods behave misleadingly different from the OutputPage::addWikitext*() methods: they don't tidy their output, they have different defaults for interface/content language selection, and they (sometimes) add wrapper divs. Deprecate these and add new methods with tidy output, clear language selection, and consistent defaults: OutputPage::parseAsContent(), OutputPage::parseAsInterface(), and OutputPage::parseInlineAsInterface(). Unify the implementation of the parse* methods with the addWikiText* methods, to reduce the likelihood that the behavior will diverge again in the future. Bug: T198214 Change-Id: Ica79c2acbc542ef37f971c0be2582ae771a23bd0
2018-10-26 15:14:01 +00:00
)->getText( [
Pass appropriate userLang to ParserOutput::getText() The rules for how ParserOptions::getUserLang() are set are a little baroque. When the 'uselang' option is set on the request, for example, RequestContext::getLanguage() can differ from the user language; we also deliberately force the "user language" to the content language during the post-edit path to avoid having the user language leak into the links table. The changes in this patch make the ParserOutput::getText() calls "more correct" but really they are just removing some uses of the global RequestContext::getMain() from the code path, since in all of these cases the main request context (which is what we use by default) does have the correct language set. For consistency, this patch also adds `allowTOC=false` in many of the places where we already had `enableSectionEditLinks=false`; for example, user interface messages, wikitext content injected on special pages, etc. Change-Id: Ib0d8c5e538e472ad672c45e5b68907242afb1e14
2023-01-25 19:50:36 +00:00
'allowTOC' => false,
Deprecate OutputPage::parse() and OutputPage::parseInline() The OutputPage::parse() and OutputPage::parseInline() methods behave misleadingly different from the OutputPage::addWikitext*() methods: they don't tidy their output, they have different defaults for interface/content language selection, and they (sometimes) add wrapper divs. Deprecate these and add new methods with tidy output, clear language selection, and consistent defaults: OutputPage::parseAsContent(), OutputPage::parseAsInterface(), and OutputPage::parseInlineAsInterface(). Unify the implementation of the parse* methods with the addWikiText* methods, to reduce the likelihood that the behavior will diverge again in the future. Bug: T198214 Change-Id: Ica79c2acbc542ef37f971c0be2582ae771a23bd0
2018-10-26 15:14:01 +00:00
'enableSectionEditLinks' => false,
Pass appropriate userLang to ParserOutput::getText() The rules for how ParserOptions::getUserLang() are set are a little baroque. When the 'uselang' option is set on the request, for example, RequestContext::getLanguage() can differ from the user language; we also deliberately force the "user language" to the content language during the post-edit path to avoid having the user language leak into the links table. The changes in this patch make the ParserOutput::getText() calls "more correct" but really they are just removing some uses of the global RequestContext::getMain() from the code path, since in all of these cases the main request context (which is what we use by default) does have the correct language set. For consistency, this patch also adds `allowTOC=false` in many of the places where we already had `enableSectionEditLinks=false`; for example, user interface messages, wikitext content injected on special pages, etc. Change-Id: Ib0d8c5e538e472ad672c45e5b68907242afb1e14
2023-01-25 19:50:36 +00:00
'wrapperDivClass' => '',
'userLang' => $this->getContext()->getLanguage(),
Deprecate OutputPage::parse() and OutputPage::parseInline() The OutputPage::parse() and OutputPage::parseInline() methods behave misleadingly different from the OutputPage::addWikitext*() methods: they don't tidy their output, they have different defaults for interface/content language selection, and they (sometimes) add wrapper divs. Deprecate these and add new methods with tidy output, clear language selection, and consistent defaults: OutputPage::parseAsContent(), OutputPage::parseAsInterface(), and OutputPage::parseInlineAsInterface(). Unify the implementation of the parse* methods with the addWikiText* methods, to reduce the likelihood that the behavior will diverge again in the future. Bug: T198214 Change-Id: Ica79c2acbc542ef37f971c0be2582ae771a23bd0
2018-10-26 15:14:01 +00:00
] );
}
/**
* Parse wikitext *in the user interface language*, strip
* paragraph wrapper, and return the HTML.
* The result will not be language-converted, as user interface messages
* are already localized into a specific variant.
* Output will be tidy. Outer paragraph wrapper will only be stripped
* if the result is a single paragraph.
*
* @param string $text Wikitext in the user interface language
* @param bool $linestart Is this the start of a line? (Defaults to true)
* @throws MWException
* @return string HTML
Update deprecations after OutputPage::parse*As*() backport to 1.32 Originally landed in 1.33 in a3f63785eebfa2f729d42da880b6722736ef1e7d. Backported to 1.32 in cff9ab4b5c26dc87137aa2c04efd5d1f31226fe8. Change-Id: I402c31b57904f6a6fd0b3a2c32a1c0a73c553254
2018-11-01 20:30:57 +00:00
* @since 1.32
Deprecate OutputPage::parse() and OutputPage::parseInline() The OutputPage::parse() and OutputPage::parseInline() methods behave misleadingly different from the OutputPage::addWikitext*() methods: they don't tidy their output, they have different defaults for interface/content language selection, and they (sometimes) add wrapper divs. Deprecate these and add new methods with tidy output, clear language selection, and consistent defaults: OutputPage::parseAsContent(), OutputPage::parseAsInterface(), and OutputPage::parseInlineAsInterface(). Unify the implementation of the parse* methods with the addWikiText* methods, to reduce the likelihood that the behavior will diverge again in the future. Bug: T198214 Change-Id: Ica79c2acbc542ef37f971c0be2582ae771a23bd0
2018-10-26 15:14:01 +00:00
*/
public function parseInlineAsInterface( $text, $linestart = true ) {
return Parser::stripOuterParagraph(
$this->parseAsInterface( $text, $linestart )
);
}
Fix OutputPage::parseInternal() by stripping <div> wrapper We should probably strip the <div> wrapper in OutputPage::parse() as well: this behavior was added in 1.30, but it only applies when $interface is false. However, that requires a more careful audit (a lot more places call parse() than parseInline()) and so I'll defer that for now. Change-Id: Iad5412f03af29c04deb653969dd71f6c86f0ae50
2018-10-26 14:05:34 +00:00
/**
* Parse wikitext and return the HTML (internal implementation helper)
*
* @param string $text
Remove Title from public interface of OutputPage Bug: T278459 Change-Id: If455682540faf8d42fd8a14e0ef09bac24f72db4
2021-04-08 19:11:06 +00:00
* @param PageReference $title The title to use
Fix OutputPage::parseInternal() by stripping <div> wrapper We should probably strip the <div> wrapper in OutputPage::parse() as well: this behavior was added in 1.30, but it only applies when $interface is false. However, that requires a more careful audit (a lot more places call parse() than parseInline()) and so I'll defer that for now. Change-Id: Iad5412f03af29c04deb653969dd71f6c86f0ae50
2018-10-26 14:05:34 +00:00
* @param bool $linestart Is this the start of a line?
* @param bool $interface Use interface language (instead of content language) while parsing
* language sensitive magic words like GRAMMAR and PLURAL. This also disables
* LanguageConverter.
* @throws MWException
* @return ParserOutput
*/
Remove deprecated non-tidy OutputPage::parse/parseInline methods The non-tidy parsing modes were deprecated in 1.32. Remove the deprecated parse/parseInline methods which use them, as part of a general clean up of nontidy parsing modes. Bug: T198214 Change-Id: I04ea82dd20eac8b0ce07a8d0e6b0bf2c4b03dbe8
2020-01-15 19:03:33 +00:00
private function parseInternal( $text, $title, $linestart, $interface ) {
Coding style: Auto-fix MediaWiki.Usage.IsNull.IsNull Change-Id: I90cfe8366c0245c9c67e598d17800684897a4e27
2020-01-09 23:48:34 +00:00
if ( $title === null ) {
Stage 2 of war on $wgTitle!! Make OutputPage, Skin and children rely on mTitle rather than $wgTitle. In theory, you could have an OutputPage/Skin that was referring to some title other than $wgTitle, unlikely though. In any case, make getTitle() return $wgTitle for now, just in case.
2009-04-09 02:22:36 +00:00
throw new MWException( 'Empty $mTitle in ' . __METHOD__ );
* Add guard exception in OutputPage::parse() for failure case where $wgTitle is null; that causes a fatal error which doesn't show a backtrace if we don't catch it here. * Temporarily set $wgTitle while doing article creation/edit updates during XML import. A null $wgTitle as in importDump.php caused a fatal error when special pages were transcluded if they did any parsing via OutputPage::parse()
2008-11-24 01:58:15 +00:00
}
Allow OutputPage::parse() to parse in any langauge, modified wfMsgExt() and wfMessage() accordingly
2011-01-05 12:24:39 +00:00
Merged localisation-work branch: * Made lines from initialiseMessages() appear as list items during installation * Moved the bulk of the localisation data from the Language*.php files to the Messages*.php files. Deleted most of the Languages*.php files. * Introduced "stub global" framework to provide deferred initialisation of core modules. * Removed placeholder values for $wgTitle and $wgArticle, these variables will now be null during the initialisation process, until they are set by index.php or another entry point. * Added DBA cache type, for BDB-style caches. * Removed custom date format functions, replacing them with a format string in the style of PHP's date(). Used string identifiers instead of integer identifiers, in both the language files and user preferences. Migration should be transparent in most cases. * Simplified the initialisation API for LoadBalancer objects. * Removed the broken altencoding feature. * Moved default user options and toggles from Language to User. Language objects are still able to define default preference overrides and extra user toggles, via a slightly different interface. * Don't include the date option in the parser cache rendering hash unless $wgUseDynamicDates is enabled. * Merged LanguageUtf8 with Language. Removed LanguageUtf8.php. * Removed inclusion of language files from the bottom of Language.php. This is now consistently done from Language::factory(). * Add the name of the executing maintenance script to the debug log. Start the profiler during maintenance scripts. * Added "serialized" directory, for storing precompiled data in serialized form.
2006-07-26 07:15:39 +00:00
$popts = $this->parserOptions();
Deprecate ParserOptions::getTidy() and ParserOptions::setTidy() These options no longer have any effect. Bug: T198214 Change-Id: Icc3eaed7ab8a3070c4339b272d580328ba40912d
2020-04-01 21:36:23 +00:00
Deprecate OutputPage::parse() and OutputPage::parseInline() The OutputPage::parse() and OutputPage::parseInline() methods behave misleadingly different from the OutputPage::addWikitext*() methods: they don't tidy their output, they have different defaults for interface/content language selection, and they (sometimes) add wrapper divs. Deprecate these and add new methods with tidy output, clear language selection, and consistent defaults: OutputPage::parseAsContent(), OutputPage::parseAsInterface(), and OutputPage::parseInlineAsInterface(). Unify the implementation of the parse* methods with the addWikiText* methods, to reduce the likelihood that the behavior will diverge again in the future. Bug: T198214 Change-Id: Ica79c2acbc542ef37f971c0be2582ae771a23bd0
2018-10-26 15:14:01 +00:00
$oldInterface = $popts->setInterfaceMessage( (bool)$interface );
Deprecate Parser::getFreshParser() Following up on the comment I made at Ibbc1423166f4804a5122, make Parser instance management a ParserFactory responsibility. It is weird for Parser to have a ParserFactory proxy aspect. * Add ParserFactory::getMainInstance(), which is equivalent to the old MediaWikiServices::getParser() and $wgParser. * Add ParserFactory::getInstance(), which is equivalent to $wgParser->getFreshInstance(), returning the main instance if it is free, or a new instance otherwise. The naming is supposed to encourage it as the default way to get a parser, which will help with the linked bug. * Deprecate Parser::getFreshParser() and migrate all core callers. I left the entry in ServiceWiring.php so that it's not immediately necessary to migrate ObjectFactory specs that ask for Parser. Bug: T310948 Change-Id: I762b191e978c2d1bbc9f332c9cfa047888ce2e67
2022-06-20 03:48:44 +00:00
$parserOutput = MediaWikiServices::getInstance()->getParserFactory()->getInstance()
->parse(
$text, $title, $popts,
$linestart, true, $this->mRevisionId
);
Allow OutputPage::parse() to parse in any langauge, modified wfMsgExt() and wfMessage() accordingly
2011-01-05 12:24:39 +00:00
Deprecate OutputPage::parse() and OutputPage::parseInline() The OutputPage::parse() and OutputPage::parseInline() methods behave misleadingly different from the OutputPage::addWikitext*() methods: they don't tidy their output, they have different defaults for interface/content language selection, and they (sometimes) add wrapper divs. Deprecate these and add new methods with tidy output, clear language selection, and consistent defaults: OutputPage::parseAsContent(), OutputPage::parseAsInterface(), and OutputPage::parseInlineAsInterface(). Unify the implementation of the parse* methods with the addWikiText* methods, to reduce the likelihood that the behavior will diverge again in the future. Bug: T198214 Change-Id: Ica79c2acbc542ef37f971c0be2582ae771a23bd0
2018-10-26 15:14:01 +00:00
$popts->setInterfaceMessage( $oldInterface );
Fix OutputPage::parseInternal() by stripping <div> wrapper We should probably strip the <div> wrapper in OutputPage::parse() as well: this behavior was added in 1.30, but it only applies when $interface is false. However, that requires a more careful audit (a lot more places call parse() than parseInline()) and so I'll defer that for now. Change-Id: Iad5412f03af29c04deb653969dd71f6c86f0ae50
2018-10-26 14:05:34 +00:00
return $parserOutput;
Allow af_public_comments to be wikitext. Relatedly, add parseInline function to OutputPage, to avoid duplicating the same awful regex in wfMsgExt.
2009-01-31 01:59:13 +00:00
}
Changing comments layout preparing for generated documentation with Phpdocumentor
2004-09-02 23:28:24 +00:00
/**
Step two in fixing OutputPage's documentation, more or less the same as r61647
2010-01-29 21:32:45 +00:00
* Set the value of the "s-maxage" part of the "Cache-control" HTTP header
*
Rename setSquidMaxage() => setCdnMaxage() Also updated the OutputPage member variable. Change-Id: I0b41c66da58c6126c1bd96170c02d084a067383c
2015-12-10 01:07:05 +00:00
* @param int $maxage Maximum cache time on the CDN, in seconds.
Changing comments layout preparing for generated documentation with Phpdocumentor
2004-09-02 23:28:24 +00:00
*/
Rename setSquidMaxage() => setCdnMaxage() Also updated the OutputPage member variable. Change-Id: I0b41c66da58c6126c1bd96170c02d084a067383c
2015-12-10 01:07:05 +00:00
public function setCdnMaxage( $maxage ) {
$this->mCdnMaxage = min( $maxage, $this->mCdnMaxageLimit );
Lower CDN cache TTL when slave lag is high * $wgCdnMaxageLagged controls exactly what that TTL is and the usual "max lag" settings determine what "high" is for lag (which already makes the site read-only). * This helps avoids stale content getting stuck in CDN for a month just because a slave was lagged for a minute. Of course race conditions with normal slave lag and WAN cache relay purges can still lead to this problem, though the scope of it is reduced. Bug: T113204 Change-Id: I7ff0a8d88665f4e557566e7b412e75edee2627fe
2015-10-01 07:24:18 +00:00
}
/**
Comments and code cleanup in OutputPage Change-Id: I8c44f30801792e8b84ab3d556cb5b1b5feb9baf3
2018-08-02 18:26:55 +00:00
* Set the value of the "s-maxage" part of the "Cache-control" HTTP header to $maxage if that is
* lower than the current s-maxage. Either way, $maxage is now an upper limit on s-maxage, so
* that future calls to setCdnMaxage() will no longer be able to raise the s-maxage above
* $maxage.
Lower CDN cache TTL when slave lag is high * $wgCdnMaxageLagged controls exactly what that TTL is and the usual "max lag" settings determine what "high" is for lag (which already makes the site read-only). * This helps avoids stale content getting stuck in CDN for a month just because a slave was lagged for a minute. Of course race conditions with normal slave lag and WAN cache relay purges can still lead to this problem, though the scope of it is reduced. Bug: T113204 Change-Id: I7ff0a8d88665f4e557566e7b412e75edee2627fe
2015-10-01 07:24:18 +00:00
*
Rename setSquidMaxage() => setCdnMaxage() Also updated the OutputPage member variable. Change-Id: I0b41c66da58c6126c1bd96170c02d084a067383c
2015-12-10 01:07:05 +00:00
* @param int $maxage Maximum cache time on the CDN, in seconds
Lower CDN cache TTL when slave lag is high * $wgCdnMaxageLagged controls exactly what that TTL is and the usual "max lag" settings determine what "high" is for lag (which already makes the site read-only). * This helps avoids stale content getting stuck in CDN for a month just because a slave was lagged for a minute. Of course race conditions with normal slave lag and WAN cache relay purges can still lead to this problem, though the scope of it is reduced. Bug: T113204 Change-Id: I7ff0a8d88665f4e557566e7b412e75edee2627fe
2015-10-01 07:24:18 +00:00
* @since 1.27
*/
public function lowerCdnMaxage( $maxage ) {
Fix broken min() call in OutputPage Change-Id: Ia2d895371f6f114c1cb821dd1d2aaaef7935033f
2015-10-10 03:55:47 +00:00
$this->mCdnMaxageLimit = min( $maxage, $this->mCdnMaxageLimit );
Rename setSquidMaxage() => setCdnMaxage() Also updated the OutputPage member variable. Change-Id: I0b41c66da58c6126c1bd96170c02d084a067383c
2015-12-10 01:07:05 +00:00
$this->setCdnMaxage( $this->mCdnMaxage );
added function setSquidMaxage to adjust the Cache timeout for special pages
2004-02-08 21:12:07 +00:00
}
This commit implements Josh Cogliati's "info" feature (e.g. http://en.wikipedia.org/w/wiki.phtml?title=Wikipedia&action=info), fixes the page headings for protect/unprotect, and does a lot of whitespace changes (sorry about that). I've also fixed confusing indentation somewhere.
2004-07-08 14:53:54 +00:00
Use adaptive CDN TTLs for page views This makes frequently changed pages be less likely to be seen in stale forms if purges are delayed or lost. Pages that have not been edited for somewhat longer than the nominal CDN cache TTL will retain the nominal TTL. Category pages will adapt based on page_touched rather than the last revision, given the possibility of constaintly changing membership. With their lesser overall usage, this is less risky than for Article, and also more useful. Change-Id: If621aca2fb68e9f87a50c891dac8dc6ec7641f5c
2016-08-23 04:10:13 +00:00
/**
docs: Avoid some scalar juggling Phan can treat scalar types as non-interchangeable with `scalar_implicit_cast` set to false. This patch fixes some of those issues (which are in total >1000), namely the ones with alphabetic order < includes/actions. Change-Id: Ib1c6573ab899088bc319b9da9ceaffc850da3dbe
2019-09-15 14:40:42 +00:00
* Get TTL in [$minTTL,$maxTTL] and pass it to lowerCdnMaxage()
Use adaptive CDN TTLs for page views This makes frequently changed pages be less likely to be seen in stale forms if purges are delayed or lost. Pages that have not been edited for somewhat longer than the nominal CDN cache TTL will retain the nominal TTL. Category pages will adapt based on page_touched rather than the last revision, given the possibility of constaintly changing membership. With their lesser overall usage, this is less risky than for Article, and also more useful. Change-Id: If621aca2fb68e9f87a50c891dac8dc6ec7641f5c
2016-08-23 04:10:13 +00:00
*
* This sets and returns $minTTL if $mtime is false or null. Otherwise,
* the TTL is higher the older the $mtime timestamp is. Essentially, the
* TTL is 90% of the age of the object, subject to the min and max.
*
Use short type bool/int in param documentation Enable the phpcs sniffs for this and used phpcbf Change-Id: Iaa36687154ddd2bf663b9dd519f5c99409d37925
2017-08-20 11:20:59 +00:00
* @param string|int|float|bool|null $mtime Last-Modified timestamp
Comments and code cleanup in OutputPage Change-Id: I8c44f30801792e8b84ab3d556cb5b1b5feb9baf3
2018-08-02 18:26:55 +00:00
* @param int $minTTL Minimum TTL in seconds [default: 1 minute]
Rename CDN config variables to be generic, deprecating the old names Hook, methods, classes still to rename where appropriate. Bug: T104148 Depends-On: Id34339dff88bc6d1863378ac94b96b2d590b891d Depends-On: I4e2938395bcbf7956b83fff00978f09c61dcfa36 Change-Id: I7a725dae551c867a4fa7c213838d52c7fb862756
2017-11-01 20:55:24 +00:00
* @param int $maxTTL Maximum TTL in seconds [default: $wgCdnMaxAge]
Use adaptive CDN TTLs for page views This makes frequently changed pages be less likely to be seen in stale forms if purges are delayed or lost. Pages that have not been edited for somewhat longer than the nominal CDN cache TTL will retain the nominal TTL. Category pages will adapt based on page_touched rather than the last revision, given the possibility of constaintly changing membership. With their lesser overall usage, this is less risky than for Article, and also more useful. Change-Id: If621aca2fb68e9f87a50c891dac8dc6ec7641f5c
2016-08-23 04:10:13 +00:00
* @since 1.28
*/
public function adaptCdnTTL( $mtime, $minTTL = 0, $maxTTL = 0 ) {
$minTTL = $minTTL ?: IExpiringStore::TTL_MINUTE;
Use name constants to access config settings. Use name constants instead of string literals in calls to Config::get and ServiceOptions::get, when referring to core configuration variables. This protects against typos and makes the decumentation and schema declaration of the config settings discoverable. This is the first batch, only touching files directly under /includes/ Change-Id: I7252e636c7c86d950d9257b33491af492c6dd5eb
2022-04-01 15:58:32 +00:00
$maxTTL = $maxTTL ?: $this->getConfig()->get( MainConfigNames::CdnMaxAge );
Use adaptive CDN TTLs for page views This makes frequently changed pages be less likely to be seen in stale forms if purges are delayed or lost. Pages that have not been edited for somewhat longer than the nominal CDN cache TTL will retain the nominal TTL. Category pages will adapt based on page_touched rather than the last revision, given the possibility of constaintly changing membership. With their lesser overall usage, this is less risky than for Article, and also more useful. Change-Id: If621aca2fb68e9f87a50c891dac8dc6ec7641f5c
2016-08-23 04:10:13 +00:00
if ( $mtime === null || $mtime === false ) {
docs: Avoid some scalar juggling Phan can treat scalar types as non-interchangeable with `scalar_implicit_cast` set to false. This patch fixes some of those issues (which are in total >1000), namely the ones with alphabetic order < includes/actions. Change-Id: Ib1c6573ab899088bc319b9da9ceaffc850da3dbe
2019-09-15 14:40:42 +00:00
return; // entity does not exist
Use adaptive CDN TTLs for page views This makes frequently changed pages be less likely to be seen in stale forms if purges are delayed or lost. Pages that have not been edited for somewhat longer than the nominal CDN cache TTL will retain the nominal TTL. Category pages will adapt based on page_touched rather than the last revision, given the possibility of constaintly changing membership. With their lesser overall usage, this is less risky than for Article, and also more useful. Change-Id: If621aca2fb68e9f87a50c891dac8dc6ec7641f5c
2016-08-23 04:10:13 +00:00
}
docs: Avoid some scalar juggling Phan can treat scalar types as non-interchangeable with `scalar_implicit_cast` set to false. This patch fixes some of those issues (which are in total >1000), namely the ones with alphabetic order < includes/actions. Change-Id: Ib1c6573ab899088bc319b9da9ceaffc850da3dbe
2019-09-15 14:40:42 +00:00
$age = MWTimestamp::time() - (int)wfTimestamp( TS_UNIX, $mtime );
build: Updating mediawiki/mediawiki-codesniffer to 0.12.0 Change-Id: Iefaae5043fa77d5d556c31079549dab8f61bd3ef
2017-09-10 19:11:37 +00:00
$adaptiveTTL = max( 0.9 * $age, $minTTL );
Use adaptive CDN TTLs for page views This makes frequently changed pages be less likely to be seen in stale forms if purges are delayed or lost. Pages that have not been edited for somewhat longer than the nominal CDN cache TTL will retain the nominal TTL. Category pages will adapt based on page_touched rather than the last revision, given the possibility of constaintly changing membership. With their lesser overall usage, this is less risky than for Article, and also more useful. Change-Id: If621aca2fb68e9f87a50c891dac8dc6ec7641f5c
2016-08-23 04:10:13 +00:00
$adaptiveTTL = min( $adaptiveTTL, $maxTTL );
$this->lowerCdnMaxage( (int)$adaptiveTTL );
}
Changing comments layout preparing for generated documentation with Phpdocumentor
2004-09-02 23:28:24 +00:00
/**
BadTitleError: Enable 1 hour CDN cache for this error response Varnish is unable to cache these currently. However since pages with an invalid title will never have any variable/editable content, we can cache them unconditionally. Also, undeprecate OutputPage::enableClientCache for this since it is clearly useful here, but the parameter is being removed since OutputPage::disableClientCache should be used to disable client side caching instead. Bug: T316932 Change-Id: Ib98fd165bffd9ca3dcbbcb8c6f06964dba7c90b7
2022-09-02 16:13:30 +00:00
* Do not send nocache headers
Changing comments layout preparing for generated documentation with Phpdocumentor
2004-09-02 23:28:24 +00:00
*/
BadTitleError: Enable 1 hour CDN cache for this error response Varnish is unable to cache these currently. However since pages with an invalid title will never have any variable/editable content, we can cache them unconditionally. Also, undeprecate OutputPage::enableClientCache for this since it is clearly useful here, but the parameter is being removed since OutputPage::disableClientCache should be used to disable client side caching instead. Bug: T316932 Change-Id: Ib98fd165bffd9ca3dcbbcb8c6f06964dba7c90b7
2022-09-02 16:13:30 +00:00
public function enableClientCache(): void {
$this->mEnableClientCache = true;
don't cache error pages
2004-03-23 10:19:31 +00:00
}
This commit implements Josh Cogliati's "info" feature (e.g. http://en.wikipedia.org/w/wiki.phtml?title=Wikipedia&action=info), fixes the page headings for protect/unprotect, and does a lot of whitespace changes (sorry about that). I've also fixed confusing indentation somewhere.
2004-07-08 14:53:54 +00:00
Add OutputPage::disableClientCache() and deprecate ::enableClientCache(false) As far as I can tell, outside of test cases, every single time we call OutputPage::enableClientCache() we pass it `false` and actually mean to *disable* the client cache. Create a new less-confusingly-named method and deprecate the old one. Change-Id: I7c89e20528a0d91173f0edcb997dcae631935ee5
2022-02-04 19:12:11 +00:00
/**
* Force the page to send nocache headers
* @since 1.38
*/
public function disableClientCache(): void {
$this->mEnableClientCache = false;
}
block: Allow cookie-block tracking from any uncached web request This was previously hardcoded from three places: 1) Upon viewing EditPage, 2) Upon viewing SpecialCreateAccount, 3) For any url if the user is logged-in (User::loadFromSession/isLoggedIn). == User::loadFromSession Performing cookie blocks from here created a circular dependency because Block may need the user language for localisation, which is determined by asking the User object. This was previously worked around by using a DeferredUpdate (T180050, T226777). Moving this logic explicitly to the end of the pre-send cycle in MediaWiki::preOutputCommit breaks the cycle. This is also where other request-specific handling resides already. == Limited effect on unregistered users When an unregistered user performs an edit, and gets blocked, the cookie block is not applied until they open built-in editor or CreateAccount page. This makes it more likely for a user's IP to change meanwhile. Either intentionally, or simply due to IPs varying naturally (e.g. between mobile locations, or when going on/off WiFi). By applying it throughout sessioned page views for unregistered users, it is more likely to get set. Similar to what was already done for logged-in users. This commit also makes the intent of not caching EditPage and SpecialCreateAccount explicit. This was previously implicit through nothing having called setCdnMaxage() and/or due to Session::persist being checked for by OutputPage::sendCacheControl. Bug: T233594 Change-Id: Icf5a00f9b41d31bb6d4742c049feca0039d0c9d9
2019-09-07 23:44:46 +00:00
/**
* Whether the output might become publicly cached.
*
* @since 1.34
* @return bool
*/
public function couldBePublicCached() {
if ( !$this->cacheIsFinal ) {
// - The entry point handles its own caching and/or doesn't use OutputPage.
Remove ActionAjax Bug: T42786 Change-Id: I8bda0c281e1f4abbffbddb80ac74a6d61a034d28
2021-11-25 16:37:36 +00:00
// (such as load.php, or MediaWiki\Rest\EntryPoint).
block: Allow cookie-block tracking from any uncached web request This was previously hardcoded from three places: 1) Upon viewing EditPage, 2) Upon viewing SpecialCreateAccount, 3) For any url if the user is logged-in (User::loadFromSession/isLoggedIn). == User::loadFromSession Performing cookie blocks from here created a circular dependency because Block may need the user language for localisation, which is determined by asking the User object. This was previously worked around by using a DeferredUpdate (T180050, T226777). Moving this logic explicitly to the end of the pre-send cycle in MediaWiki::preOutputCommit breaks the cycle. This is also where other request-specific handling resides already. == Limited effect on unregistered users When an unregistered user performs an edit, and gets blocked, the cookie block is not applied until they open built-in editor or CreateAccount page. This makes it more likely for a user's IP to change meanwhile. Either intentionally, or simply due to IPs varying naturally (e.g. between mobile locations, or when going on/off WiFi). By applying it throughout sessioned page views for unregistered users, it is more likely to get set. Similar to what was already done for logged-in users. This commit also makes the intent of not caching EditPage and SpecialCreateAccount explicit. This was previously implicit through nothing having called setCdnMaxage() and/or due to Session::persist being checked for by OutputPage::sendCacheControl. Bug: T233594 Change-Id: Icf5a00f9b41d31bb6d4742c049feca0039d0c9d9
2019-09-07 23:44:46 +00:00
//
// - Or, we haven't finished processing the main part of the request yet
// (e.g. Action::show, SpecialPage::execute), and the state may still
// change via enableClientCache().
return true;
}
// e.g. various error-type pages disable all client caching
return $this->mEnableClientCache;
}
/**
* Set the expectation that cache control will not change after this point.
*
* This should be called after the main processing logic has completed
* (e.g. Action::show or SpecialPage::execute), but may be called
* before Skin output has started (OutputPage::output).
*
* @since 1.34
*/
public function considerCacheSettingsFinal() {
$this->cacheIsFinal = true;
}
Step two in fixing OutputPage's documentation, more or less the same as r61647
2010-01-29 21:32:45 +00:00
/**
Comments and code cleanup in OutputPage Change-Id: I8c44f30801792e8b84ab3d556cb5b1b5feb9baf3
2018-08-02 18:26:55 +00:00
* Get the list of cookie names that will influence the cache
Step two in fixing OutputPage's documentation, more or less the same as r61647
2010-01-29 21:32:45 +00:00
*
Fixed some @params documentation Swapped some "$var type" to "type $var" or added missing types before the $var. Changed some other types to match the more common spelling. Makes beginning of some text in captial. Change-Id: I64e8cfe478cb0ba438f40b0631d6e9049cdab567
2014-04-10 18:50:10 +00:00
* @return array
Step two in fixing OutputPage's documentation, more or less the same as r61647
2010-01-29 21:32:45 +00:00
*/
OutputPage: Add visibility to various methods & do some micro-optimizations Methods that visibility was added to are; `addMeta()`, `addLink()`, `setCanonicalUrl()`, `addScript()`, `getHeadItemsArray()`, `addParserOutput()`, `getCacheVaryCookies()` and `haveCacheVaryCookies()`. Last but not lease, did a few micro-optimizations to `addMeta()` and `addLink()`. Change-Id: I94d037a5edc7131627724fd1d864000128077b0c
2019-09-14 10:26:31 +00:00
public function getCacheVaryCookies() {
Increase OutputPage test coverage to >45% Also stop returning a value from OutputPage::adaptCdnTTL(). There are no users and the value doesn't seem very helpful: one would probably expect it to return the new value of mCdnMaxage, but instead it returns the new value of mCdnMaxageLimit. Better to have no return value than one that's easily misunderstood (*and* which nobody uses). Change-Id: Ia9dab86923b839334eab9f6fde17c4aed52130ec
2018-08-02 18:42:17 +00:00
if ( self::$cacheVaryCookies === null ) {
Use config instead of globals for OutputPage Change-Id: I5e0ebc173631d1d1052de7ccee4ef839c7c1767f
2014-08-23 08:52:41 +00:00
$config = $this->getConfig();
Increase OutputPage test coverage to >45% Also stop returning a value from OutputPage::adaptCdnTTL(). There are no users and the value doesn't seem very helpful: one would probably expect it to return the new value of mCdnMaxage, but instead it returns the new value of mCdnMaxageLimit. Better to have no return value than one that's easily misunderstood (*and* which nobody uses). Change-Id: Ia9dab86923b839334eab9f6fde17c4aed52130ec
2018-08-02 18:42:17 +00:00
self::$cacheVaryCookies = array_values( array_unique( array_merge(
Revert "Remove SessionManager, temporarily" This reverts commit 823db5d63dd5200d04c63da50ba6bf16f928e70b. Change-Id: Ibb3e023e4eb6715295586dea87d0725c344a8271
2016-02-01 20:44:03 +00:00
SessionManager::singleton()->getVaryCookies(),
Convert all array() syntax to [] Per wikitech-l consensus: https://lists.wikimedia.org/pipermail/wikitech-l/2016-February/084821.html Notes: * Disabled CallTimePassByReference due to false positives (T127163) Change-Id: I2c8ce713ce6600a0bb7bf67537c87044c7a45c4b
2016-02-17 09:09:32 +00:00
[
Revert "Remove SessionManager, temporarily" This reverts commit 823db5d63dd5200d04c63da50ba6bf16f928e70b. Change-Id: Ibb3e023e4eb6715295586dea87d0725c344a8271
2016-02-01 20:44:03 +00:00
'forceHTTPS',
Convert all array() syntax to [] Per wikitech-l consensus: https://lists.wikimedia.org/pipermail/wikitech-l/2016-February/084821.html Notes: * Disabled CallTimePassByReference due to false positives (T127163) Change-Id: I2c8ce713ce6600a0bb7bf67537c87044c7a45c4b
2016-02-17 09:09:32 +00:00
],
Use name constants to access config settings. Use name constants instead of string literals in calls to Config::get and ServiceOptions::get, when referring to core configuration variables. This protects against typos and makes the decumentation and schema declaration of the config settings discoverable. This is the first batch, only touching files directly under /includes/ Change-Id: I7252e636c7c86d950d9257b33491af492c6dd5eb
2022-04-01 15:58:32 +00:00
$config->get( MainConfigNames::CacheVaryCookies )
Increase OutputPage test coverage to >45% Also stop returning a value from OutputPage::adaptCdnTTL(). There are no users and the value doesn't seem very helpful: one would probably expect it to return the new value of mCdnMaxage, but instead it returns the new value of mCdnMaxageLimit. Better to have no return value than one that's easily misunderstood (*and* which nobody uses). Change-Id: Ia9dab86923b839334eab9f6fde17c4aed52130ec
2018-08-02 18:42:17 +00:00
) ) );
Revert "OutputPage: Temporary hack to avoid taint-check crash" This reverts commit dad17c82da0dfba8348fb03ec8b58202a4860d5d. Reason for revert: now working with newer taint-check. Change-Id: I394c04b5583f718d5d809fdd15ae723b1564d551
2020-06-26 09:33:33 +00:00
$this->getHookRunner()->onGetCacheVaryCookies( $this, self::$cacheVaryCookies );
Add CentralAuth cookies to the XVO header
2008-04-10 08:42:36 +00:00
}
Increase OutputPage test coverage to >45% Also stop returning a value from OutputPage::adaptCdnTTL(). There are no users and the value doesn't seem very helpful: one would probably expect it to return the new value of mCdnMaxage, but instead it returns the new value of mCdnMaxageLimit. Better to have no return value than one that's easily misunderstood (*and* which nobody uses). Change-Id: Ia9dab86923b839334eab9f6fde17c4aed52130ec
2018-08-02 18:42:17 +00:00
return self::$cacheVaryCookies;
Vary the cache on session, Token and LoggedOut cookies, but not UserID cookie. The session and Token cookies are necessary because they can represent credentials. The UserID cookie is sometimes present after the credential cookies are expired and thus the request is anonymous. Need to vary on the LoggedOut cookie to prevent Squid from returning a 304 response for a logged-in page view cached on the client. Presumably this could be better dealt with with ETags. Always send Cache-Control: private when there is a cache-varying cookie present, to prevent pollution of the logged-in user cache with logged-out page views. This missing logic is why people have been reporting getting logged out.
2008-02-20 04:13:24 +00:00
}
/**
* Check if the request has a cache-varying cookie header
* If it does, it's very important that we don't allow public caching
Step two in fixing OutputPage's documentation, more or less the same as r61647
2010-01-29 21:32:45 +00:00
*
Fixed some @params documentation Swapped some "$var type" to "type $var" or added missing types before the $var. Changed some other types to match the more common spelling. Makes beginning of some text in captial. Change-Id: I64e8cfe478cb0ba438f40b0631d6e9049cdab567
2014-04-10 18:50:10 +00:00
* @return bool
Vary the cache on session, Token and LoggedOut cookies, but not UserID cookie. The session and Token cookies are necessary because they can represent credentials. The UserID cookie is sometimes present after the credential cookies are expired and thus the request is anonymous. Need to vary on the LoggedOut cookie to prevent Squid from returning a 304 response for a logged-in page view cached on the client. Presumably this could be better dealt with with ETags. Always send Cache-Control: private when there is a cache-varying cookie present, to prevent pollution of the logged-in user cache with logged-out page views. This missing logic is why people have been reporting getting logged out.
2008-02-20 04:13:24 +00:00
*/
OutputPage: Add visibility to various methods & do some micro-optimizations Methods that visibility was added to are; `addMeta()`, `addLink()`, `setCanonicalUrl()`, `addScript()`, `getHeadItemsArray()`, `addParserOutput()`, `getCacheVaryCookies()` and `haveCacheVaryCookies()`. Last but not lease, did a few micro-optimizations to `addMeta()` and `addLink()`. Change-Id: I94d037a5edc7131627724fd1d864000128077b0c
2019-09-14 10:26:31 +00:00
public function haveCacheVaryCookies() {
Replace XVO with support for the Key HTTP header MediaWiki currently has support for a header called X-Vary-Options (XVO), used to communicate to upstream caches more granular cache variance options than the Vary header can. The header was envisioned by Tim Starling back in 2008 and implemented into MediaWiki and Squid 2.0, with those patches submitted to the squid-dev mailing list at the time: http://www.squid-cache.org/mail-archive/squid-dev/200802/0085.html The patches never actually made it into an upstream Squid release, however, and Squid has since evolved in potentially significant ways. Wikimedia has since switched to Varnish but XVO was not ported over as it was deemed too complex at the time; custom VCL was used instead. To our knowledge, noone else is using XVO in production and certainly not with recent, up-to-date MediaWiki releases. There is currently work at IETF's httpbis working group for a "Key" header that is in concept and implementation very similar to Tim's XVO header: https://datatracker.ietf.org/doc/draft-fielding-http-key/ Rather than rip XVO out of MediaWiki, replace it with support for the Key header, as preliminary defined by the draft spec. This is an almost straight search-and-replace. No other software (caching proxy or user-agent) currently implements Key to my knowledge, so this is essentially untested. Change-Id: I949fc289dd5d48bd34f3b37f7739e2b9cd5db277
2015-10-06 07:01:10 +00:00
$request = $this->getRequest();
foreach ( $this->getCacheVaryCookies() as $cookieName ) {
Fix-up for I949fc289d: handle falsy cookie values correctly Ensure that OutputPage::haveVaryCacheCookies() handles falsey values correctly, and that it requires a cookie to have a nonempty value before it declares a match. Change-Id: I2afe54b62c940187a427498cb4037e1dd0e78dd9
2015-10-08 04:45:26 +00:00
if ( $request->getCookie( $cookieName, '', '' ) !== '' ) {
Remove terminating line breaks from debug messages A terminating line break has not been required in wfDebug() since 2014, however no migration was done. Some of these line breaks found their way into LoggerInterface::debug() calls, where they mess up the formatting of the debug log. So, remove terminating line breaks from wfDebug() and LoggerInterface::debug() calls. Also: * Fix the stripping of leading line breaks from the log header emitted by Setup.php. This feature, accidentally broken in 2014, allows requests to be distinguished in the log file. * Avoid using the global variable $self. * Move the logging of the client IP back to Setup.php. It was moved to WebRequest in the hopes that it would not always be needed, however $wgRequest->getIP() is now called unconditionally a few lines up in Setup.php. This means that it is put in its proper place after the "start request" message. * Wrap the log header code in a closure so that variables like $name do not leak into global scope. * In Linker.php, remove a few instances of an unnecessary second parameter to wfDebug(). Change-Id: I96651d3044a95b9d210b51cb8368edc76bebbb9e
2020-06-01 05:00:39 +00:00
wfDebug( __METHOD__ . ": found $cookieName" );
Vary the cache on session, Token and LoggedOut cookies, but not UserID cookie. The session and Token cookies are necessary because they can represent credentials. The UserID cookie is sometimes present after the credential cookies are expired and thus the request is anonymous. Need to vary on the LoggedOut cookie to prevent Squid from returning a 304 response for a logged-in page view cached on the client. Presumably this could be better dealt with with ETags. Always send Cache-Control: private when there is a cache-varying cookie present, to prevent pollution of the logged-in user cache with logged-out page views. This missing logic is why people have been reporting getting logged out.
2008-02-20 04:13:24 +00:00
return true;
}
}
Remove terminating line breaks from debug messages A terminating line break has not been required in wfDebug() since 2014, however no migration was done. Some of these line breaks found their way into LoggerInterface::debug() calls, where they mess up the formatting of the debug log. So, remove terminating line breaks from wfDebug() and LoggerInterface::debug() calls. Also: * Fix the stripping of leading line breaks from the log header emitted by Setup.php. This feature, accidentally broken in 2014, allows requests to be distinguished in the log file. * Avoid using the global variable $self. * Move the logging of the client IP back to Setup.php. It was moved to WebRequest in the hopes that it would not always be needed, however $wgRequest->getIP() is now called unconditionally a few lines up in Setup.php. This means that it is put in its proper place after the "start request" message. * Wrap the log header code in a closure so that variables like $name do not leak into global scope. * In Linker.php, remove a few instances of an unnecessary second parameter to wfDebug(). Change-Id: I96651d3044a95b9d210b51cb8368edc76bebbb9e
2020-06-01 05:00:39 +00:00
wfDebug( __METHOD__ . ": no cache-varying cookies found" );
Vary the cache on session, Token and LoggedOut cookies, but not UserID cookie. The session and Token cookies are necessary because they can represent credentials. The UserID cookie is sometimes present after the credential cookies are expired and thus the request is anonymous. Need to vary on the LoggedOut cookie to prevent Squid from returning a 304 response for a logged-in page view cached on the client. Presumably this could be better dealt with with ETags. Always send Cache-Control: private when there is a cache-varying cookie present, to prevent pollution of the logged-in user cache with logged-out page views. This missing logic is why people have been reporting getting logged out.
2008-02-20 04:13:24 +00:00
return false;
}
Step two in fixing OutputPage's documentation, more or less the same as r61647
2010-01-29 21:32:45 +00:00
/**
* Add an HTTP header that will influence on the cache
*
Cleanup some docs (includes/*.php) - Swap "$variable type" to "type $variable" - Added missing types - Fixed spacing inside docs - Makes beginning of @param/@return/@var/@throws in capital - Changed some types to match the more common spelling Change-Id: I783e4dbfe5f6f98b32b9a03ccf6439e13e132bcc
2014-07-24 17:42:24 +00:00
* @param string $header Header name
Remove $wgUseKeyHeader and OutputPage::getKeyHeader(), deprecated in 1.32 These implemented a since-abandoned draft IETF spec, and the code was broken due to (1) case-(in)sensitivity issues with the Accept-Language header and (2) the BCP47 language code compatibility workaround we use. Change-Id: Ia53d07cd8ce8ab1497294ea244c13c7499f632c7
2019-06-19 18:22:42 +00:00
* @param string[]|null $option Deprecated; formerly options for the
* Key header, deprecated in 1.32 and removed in 1.34. See
* https://datatracker.ietf.org/doc/draft-fielding-http-key/
* for the list of formerly-valid options.
Step two in fixing OutputPage's documentation, more or less the same as r61647
2010-01-29 21:32:45 +00:00
*/
Rewrite OutputPage::addVaryHeader Rewrite OutputPage::addVaryHeader which had a very confusing structure. There is one breaking change: the $option argument was declared as array|null, but the function accepted everything and showed inconsistent behavior; e.g. $op->addVaryHeader( 'Foo', 'bar' ) resulted in 'X-Vary-Options: Foo;bar' but $op->addVaryHeader( 'Foo' ) $op->addVaryHeader( 'Foo', 'bar' ) resulted in 'X-Vary-Options: Foo'. With the patch, non-array arguments (other than null) result in an error. Change-Id: Id31d95fe27b01b00ec8a1d7a3996275fc0aacf3c
2015-09-08 21:59:45 +00:00
public function addVaryHeader( $header, array $option = null ) {
Remove $wgUseKeyHeader and OutputPage::getKeyHeader(), deprecated in 1.32 These implemented a since-abandoned draft IETF spec, and the code was broken due to (1) case-(in)sensitivity issues with the Accept-Language header and (2) the BCP47 language code compatibility workaround we use. Change-Id: Ia53d07cd8ce8ab1497294ea244c13c7499f632c7
2019-06-19 18:22:42 +00:00
if ( $option !== null && count( $option ) > 0 ) {
Introduce wfDeprecatedMsg() Deprecating something means to say something nasty about it, or to draw its character into question. For example, "this function is lazy and good for nothing". Deprecatory remarks by a developer are generally taken as a warning that violence will soon be done against the function in question. Other developers are thus warned to avoid associating with the deprecated function. However, since wfDeprecated() was introduced, it has become obvious that the targets of deprecation are not limited to functions. Developers can deprecate literally anything: a parameter, a return value, a file format, Mondays, the concept of being, etc. wfDeprecated() requires every deprecatory statement to begin with "use of", leading to some awkward sentences. For example, one might say: "Use of your mouth to cough without it being covered by your arm is deprecated since 2020." So, introduce wfDeprecatedMsg(), which allows deprecation messages to be specified in plain text, with the caller description being optionally appended. Migrate incorrect or gramatically awkward uses of wfDeprecated() to wfDeprecatedMsg(). Change-Id: Ib3dd2fe37677d98425d0f3692db5c9e988943ae8
2020-06-12 04:18:35 +00:00
wfDeprecatedMsg(
'The $option parameter to addVaryHeader is ignored since MediaWiki 1.34',
'1.34' );
Rewrite OutputPage::addVaryHeader Rewrite OutputPage::addVaryHeader which had a very confusing structure. There is one breaking change: the $option argument was declared as array|null, but the function accepted everything and showed inconsistent behavior; e.g. $op->addVaryHeader( 'Foo', 'bar' ) resulted in 'X-Vary-Options: Foo;bar' but $op->addVaryHeader( 'Foo' ) $op->addVaryHeader( 'Foo', 'bar' ) resulted in 'X-Vary-Options: Foo'. With the patch, non-array arguments (other than null) result in an error. Change-Id: Id31d95fe27b01b00ec8a1d7a3996275fc0aacf3c
2015-09-08 21:59:45 +00:00
}
Remove $wgUseKeyHeader and OutputPage::getKeyHeader(), deprecated in 1.32 These implemented a since-abandoned draft IETF spec, and the code was broken due to (1) case-(in)sensitivity issues with the Accept-Language header and (2) the BCP47 language code compatibility workaround we use. Change-Id: Ia53d07cd8ce8ab1497294ea244c13c7499f632c7
2019-06-19 18:22:42 +00:00
if ( !array_key_exists( $header, $this->mVaryHeader ) ) {
$this->mVaryHeader[$header] = null;
bug 21672: Add Accept-Language to Vary and XVO headers to help Squid to determine user's perferred local language which has been eagerly wanted in Chinese Wikipedia.
2009-11-28 19:13:23 +00:00
}
}
Make OutputPage's mechanism for gathering Vary headers reusable Maybe WebResponse would be a better place for this, but this'll work for now. * Add getVaryHeader() to get the serialized version of $mVaryHeader. * Remove 'Cookie' from the initializer, and add it explicitly later ** This allows reuse by code that doesn't want 'Cookie' in there Change-Id: Ifa8f74a21f832154b05e54ced95b9f9eb6052fcd
2012-06-01 10:23:53 +00:00
/**
* Return a Vary: header on which to vary caches. Based on the keys of $mVaryHeader,
* such as Accept-Encoding or Cookie
Follow-up I47412b67: pass an array of query parameters to Linker::link() Spotted by Daniel Werner in that change. Change-Id: I8ded52b8285b2af0f69cc586abe532ea56cc18bb
2012-08-08 08:42:17 +00:00
*
Fixed some @params documentation Swapped some "$var type" to "type $var" or added missing types before the $var. Changed some other types to match the more common spelling. Makes beginning of some text in captial. Change-Id: I64e8cfe478cb0ba438f40b0631d6e9049cdab567
2014-04-10 18:50:10 +00:00
* @return string
Make OutputPage's mechanism for gathering Vary headers reusable Maybe WebResponse would be a better place for this, but this'll work for now. * Add getVaryHeader() to get the serialized version of $mVaryHeader. * Remove 'Cookie' from the initializer, and add it explicitly later ** This allows reuse by code that doesn't want 'Cookie' in there Change-Id: Ifa8f74a21f832154b05e54ced95b9f9eb6052fcd
2012-06-01 10:23:53 +00:00
*/
public function getVaryHeader() {
SECURITY: RawAction: Vary on the usual headers This avoids edge cases where the user isn't logged in but we still need varying for proper cache behavior. Bug: T125283 Change-Id: I43cde3a48371e62a16bda1291b1b51986e60fe4c Signed-off-by: Chad Horohoe <chadh@wikimedia.org>
2016-01-31 20:43:00 +00:00
// If we vary on cookies, let's make sure it's always included here too.
if ( $this->getCacheVaryCookies() ) {
$this->addVaryHeader( 'Cookie' );
}
Revert "Remove SessionManager, temporarily" This reverts commit 823db5d63dd5200d04c63da50ba6bf16f928e70b. Change-Id: Ibb3e023e4eb6715295586dea87d0725c344a8271
2016-02-01 20:44:03 +00:00
foreach ( SessionManager::singleton()->getVaryHeaders() as $header => $options ) {
$this->addVaryHeader( $header, $options );
}
Replace uses of join() by implode() All of core uses implode() consistently now. Change-Id: Iba50898c64c43f356d1caf8869f484e90d9ff651
2016-03-08 08:13:12 +00:00
return 'Vary: ' . implode( ', ', array_keys( $this->mVaryHeader ) );
Make OutputPage's mechanism for gathering Vary headers reusable Maybe WebResponse would be a better place for this, but this'll work for now. * Add getVaryHeader() to get the serialized version of $mVaryHeader. * Remove 'Cookie' from the initializer, and add it explicitly later ** This allows reuse by code that doesn't want 'Cookie' in there Change-Id: Ifa8f74a21f832154b05e54ced95b9f9eb6052fcd
2012-06-01 10:23:53 +00:00
}
Preload the logo using link rel="preload" http header This greatly increases the priority of loading the logo on browsers that support rel="preload". Bug: T100999 Change-Id: I0738fcc0a575153dab65016fa87faaa9b8b97a9d
2015-06-01 16:58:42 +00:00
/**
* Add an HTTP Link: header
*
* @param string $header Header value
*/
public function addLinkHeader( $header ) {
$this->mLinkHeader[] = $header;
}
/**
* Return a Link: header. Based on the values of $mLinkHeader.
*
docs: Avoid some scalar juggling Phan can treat scalar types as non-interchangeable with `scalar_implicit_cast` set to false. This patch fixes some of those issues (which are in total >1000), namely the ones with alphabetic order < includes/actions. Change-Id: Ib1c6573ab899088bc319b9da9ceaffc850da3dbe
2019-09-15 14:40:42 +00:00
* @return string|false
Preload the logo using link rel="preload" http header This greatly increases the priority of loading the logo on browsers that support rel="preload". Bug: T100999 Change-Id: I0738fcc0a575153dab65016fa87faaa9b8b97a9d
2015-06-01 16:58:42 +00:00
*/
public function getLinkHeader() {
if ( !$this->mLinkHeader ) {
return false;
}
return 'Link: ' . implode( ',', $this->mLinkHeader );
}
Step two in fixing OutputPage's documentation, more or less the same as r61647
2010-01-29 21:32:45 +00:00
/**
Follow up to 'Remove support for the deprecated Key header' Clean up a few more code paths and documentation bits left behind by Ia53d07cd8ce8ab1497294ea244c13c7499f632c7. Change-Id: I2bb1749c45bb79b27c5a3b2e1b8ed3395e8c11e0
2018-10-18 16:01:23 +00:00
* T23672: Add Accept-Language to Vary header if there's no 'variant' parameter in GET.
Step two in fixing OutputPage's documentation, more or less the same as r61647
2010-01-29 21:32:45 +00:00
*
* For example:
Comments and code cleanup in OutputPage Change-Id: I8c44f30801792e8b84ab3d556cb5b1b5feb9baf3
2018-08-02 18:26:55 +00:00
* /w/index.php?title=Main_page will vary based on Accept-Language; but
* /w/index.php?title=Main_page&variant=zh-cn will not.
Step two in fixing OutputPage's documentation, more or less the same as r61647
2010-01-29 21:32:45 +00:00
*/
Comments and code cleanup in OutputPage Change-Id: I8c44f30801792e8b84ab3d556cb5b1b5feb9baf3
2018-08-02 18:26:55 +00:00
private function addAcceptLanguage() {
Apparently for certain (API) requests $this->getTitle() doesn't return a valid Title. Usually these are API siteinfo requests (api.php?action=query&meta=siteinfo). In that case, this code used to produce a fatal ("Call to a member function getPageLanguage() on a non-object"). Change-Id: Ifb6f99fe554890ff2719220f58d1b6c1a7a95ddc
2014-06-11 20:00:24 +00:00
$title = $this->getTitle();
if ( !$title instanceof Title ) {
return;
}
Replace deprecated function in OutputPage Now mocking LanguageConverter Change-Id: I25f2f90d8dcf2cb09429e01873f1982dca199976
2020-11-03 17:18:15 +00:00
$languageConverter = MediaWikiServices::getInstance()->getLanguageConverterFactory()
->getLanguageConverter( $title->getPageLanguage() );
if ( !$this->getRequest()->getCheck( 'variant' ) && $languageConverter->hasVariants() ) {
Remove $wgUseKeyHeader and OutputPage::getKeyHeader(), deprecated in 1.32 These implemented a since-abandoned draft IETF spec, and the code was broken due to (1) case-(in)sensitivity issues with the Accept-Language header and (2) the BCP47 language code compatibility workaround we use. Change-Id: Ia53d07cd8ce8ab1497294ea244c13c7499f632c7
2019-06-19 18:22:42 +00:00
$this->addVaryHeader( 'Accept-Language' );
follow-up r59522. follow Tim's suggestion to serve same vary and XVO headers for the same URL. Now the vary headers should be consistent. Feel free to test it.
2009-12-21 18:55:42 +00:00
}
}
Fix for bug 26561: clickjacking attacks. See the bug report for full documentation.
2011-01-04 06:12:33 +00:00
/**
random w/s cleanup
2011-04-23 19:28:35 +00:00
* Set a flag which will cause an X-Frame-Options header appropriate for
* edit pages to be sent. The header value is controlled by
Fix for bug 26561: clickjacking attacks. See the bug report for full documentation.
2011-01-04 06:12:33 +00:00
* $wgEditPageFrameOptions.
*
random w/s cleanup
2011-04-23 19:28:35 +00:00
* This is the default for special pages. If you display a CSRF-protected
Fix for bug 26561: clickjacking attacks. See the bug report for full documentation.
2011-01-04 06:12:33 +00:00
* form on an ordinary view page, then you need to call this function.
More adding, updating and tweaking of documentation
2011-05-21 19:07:24 +00:00
*
Fixed some @params documentation Swapped some "$var type" to "type $var" or added missing types before the $var. Changed some other types to match the more common spelling. Makes beginning of some text in captial. Change-Id: I64e8cfe478cb0ba438f40b0631d6e9049cdab567
2014-04-10 18:50:10 +00:00
* @param bool $enable
Rename ParserOutput::{allow,prevent}Clickjacking() -> ::{get,set}PreventClickjacking() This name is consist with the rest of the setter and getter methods in ParserOutput. Renamed the methods in OutputPage, ImageHistoryList, ImageHistoryPseudoPager, and ContribsPager as well for consistency; it also makes chasing down lingering references in codesearch easier. Soft-deprecated the old name for 1.38. Hard-deprecation will follow, but there are a number of users in production that should be chased down first. Code search: https://codesearch.https://codesearch.wmcloud.org/deployed/?q=(allow%7Cprevent)Clickjacking&i=nope&files=&excludeFiles=&repos= Bug: T287216 Change-Id: I9822c60c180d204bd30cb4447a1120155d456da4
2021-09-29 20:58:59 +00:00
* @deprecated since 1.38, use ::setPreventClickjacking( true )
Fix for bug 26561: clickjacking attacks. See the bug report for full documentation.
2011-01-04 06:12:33 +00:00
*/
public function preventClickjacking( $enable = true ) {
$this->mPreventClickjacking = $enable;
}
/**
* Turn off frame-breaking. Alias for $this->preventClickjacking(false).
* This can be called from pages which do not contain any CSRF-protected
* HTML form.
Rename ParserOutput::{allow,prevent}Clickjacking() -> ::{get,set}PreventClickjacking() This name is consist with the rest of the setter and getter methods in ParserOutput. Renamed the methods in OutputPage, ImageHistoryList, ImageHistoryPseudoPager, and ContribsPager as well for consistency; it also makes chasing down lingering references in codesearch easier. Soft-deprecated the old name for 1.38. Hard-deprecation will follow, but there are a number of users in production that should be chased down first. Code search: https://codesearch.https://codesearch.wmcloud.org/deployed/?q=(allow%7Cprevent)Clickjacking&i=nope&files=&excludeFiles=&repos= Bug: T287216 Change-Id: I9822c60c180d204bd30cb4447a1120155d456da4
2021-09-29 20:58:59 +00:00
*
* @deprecated since 1.38, use ::setPreventClickjacking( false )
Fix for bug 26561: clickjacking attacks. See the bug report for full documentation.
2011-01-04 06:12:33 +00:00
*/
public function allowClickjacking() {
$this->mPreventClickjacking = false;
}
Rename ParserOutput::{allow,prevent}Clickjacking() -> ::{get,set}PreventClickjacking() This name is consist with the rest of the setter and getter methods in ParserOutput. Renamed the methods in OutputPage, ImageHistoryList, ImageHistoryPseudoPager, and ContribsPager as well for consistency; it also makes chasing down lingering references in codesearch easier. Soft-deprecated the old name for 1.38. Hard-deprecation will follow, but there are a number of users in production that should be chased down first. Code search: https://codesearch.https://codesearch.wmcloud.org/deployed/?q=(allow%7Cprevent)Clickjacking&i=nope&files=&excludeFiles=&repos= Bug: T287216 Change-Id: I9822c60c180d204bd30cb4447a1120155d456da4
2021-09-29 20:58:59 +00:00
/**
* Set the prevent-clickjacking flag.
*
* If true, will cause an X-Frame-Options header appropriate for
* edit pages to be sent. The header value is controlled by
* $wgEditPageFrameOptions. This is the default for special
* pages. If you display a CSRF-protected form on an ordinary view
* page, then you need to call this function.
*
* Setting this flag to false will turn off frame-breaking. This
* can be called from pages which do not contain any
* CSRF-protected HTML form.
*
* @param bool $enable If true, will cause an X-Frame-Options header
* appropriate for edit pages to be sent.
*
* @since 1.38
*/
public function setPreventClickjacking( bool $enable ) {
$this->mPreventClickjacking = $enable;
}
SECURITY: Copy prevent-clickjacking between OutputPage and ParserOutput Special page transclusion returns an OutputPage, whose metadata is copied into the ParserOutput, and then later back into an OutputPage. The "preventClickjacking" flag should be part of that metadata. Bug: 65778 Change-Id: I17d2720fb94bb383a92059e5adbf6c16ee3e9ef4
2014-07-10 19:16:29 +00:00
/**
* Get the prevent-clickjacking flag
*
* @since 1.24
Fixed docs - Use short form of boolean - Use capital at begin of doc text Change-Id: Ic5afacfa7298b1938d3b45ffd0cac5ce01f2f9db
2014-08-04 10:00:15 +00:00
* @return bool
SECURITY: Copy prevent-clickjacking between OutputPage and ParserOutput Special page transclusion returns an OutputPage, whose metadata is copied into the ParserOutput, and then later back into an OutputPage. The "preventClickjacking" flag should be part of that metadata. Bug: 65778 Change-Id: I17d2720fb94bb383a92059e5adbf6c16ee3e9ef4
2014-07-10 19:16:29 +00:00
*/
public function getPreventClickjacking() {
return $this->mPreventClickjacking;
}
Fix for bug 26561: clickjacking attacks. See the bug report for full documentation.
2011-01-04 06:12:33 +00:00
/**
random w/s cleanup
2011-04-23 19:28:35 +00:00
* Get the X-Frame-Options header value (without the name part), or false
* if there isn't one. This is used by Skin to determine whether to enable
Fix for bug 26561: clickjacking attacks. See the bug report for full documentation.
2011-01-04 06:12:33 +00:00
* JavaScript frame-breaking, for clients that don't support X-Frame-Options.
More adding, updating and tweaking of documentation
2011-05-21 19:07:24 +00:00
*
Cleanup some incorrect return annotations Most of these are simply changing annotations to reflect reality. If a function can return false to indicate failure the @return should indicate it. Some are fixing preg_match calls, preg match returns 1, 0 or false, but the functions all claim to return booleans. This is far from all the incorrect return types in mediawiki, there are around 250 detected by phan, but have to start somewhere. Change-Id: I1bbdfee6190747bde460f8a7084212ccafe169ef
2016-12-08 05:04:53 +00:00
* @return string|false
Fix for bug 26561: clickjacking attacks. See the bug report for full documentation.
2011-01-04 06:12:33 +00:00
*/
public function getFrameOptions() {
Use config instead of globals for OutputPage Change-Id: I5e0ebc173631d1d1052de7ccee4ef839c7c1767f
2014-08-23 08:52:41 +00:00
$config = $this->getConfig();
Use name constants to access config settings. Use name constants instead of string literals in calls to Config::get and ServiceOptions::get, when referring to core configuration variables. This protects against typos and makes the decumentation and schema declaration of the config settings discoverable. This is the first batch, only touching files directly under /includes/ Change-Id: I7252e636c7c86d950d9257b33491af492c6dd5eb
2022-04-01 15:58:32 +00:00
if ( $config->get( MainConfigNames::BreakFrames ) ) {
Fix for bug 26561: clickjacking attacks. See the bug report for full documentation.
2011-01-04 06:12:33 +00:00
return 'DENY';
Use name constants to access config settings. Use name constants instead of string literals in calls to Config::get and ServiceOptions::get, when referring to core configuration variables. This protects against typos and makes the decumentation and schema declaration of the config settings discoverable. This is the first batch, only touching files directly under /includes/ Change-Id: I7252e636c7c86d950d9257b33491af492c6dd5eb
2022-04-01 15:58:32 +00:00
} elseif ( $this->mPreventClickjacking && $config->get( MainConfigNames::EditPageFrameOptions ) ) {
return $config->get( MainConfigNames::EditPageFrameOptions );
Fix for bug 26561: clickjacking attacks. See the bug report for full documentation.
2011-01-04 06:12:33 +00:00
}
Fixup some assignments in conditionals Add/normalise some more return statements
2011-10-16 03:27:12 +00:00
return false;
Fix for bug 26561: clickjacking attacks. See the bug report for full documentation.
2011-01-04 06:12:33 +00:00
}
Reporting API and Feature Policy reporting support These are experimental APIs that lets one report policy violations: https://www.w3.org/TR/reporting/ https://github.com/w3c/webappsec-feature-policy/blob/master/reporting.md Bug: T209572 Change-Id: I002e7802000ec37b3320d8ae761cc1888d4f6edb
2019-03-22 05:43:30 +00:00
private function getReportTo() {
$config = $this->getConfig();
Use name constants to access config settings. Use name constants instead of string literals in calls to Config::get and ServiceOptions::get, when referring to core configuration variables. This protects against typos and makes the decumentation and schema declaration of the config settings discoverable. This is the first batch, only touching files directly under /includes/ Change-Id: I7252e636c7c86d950d9257b33491af492c6dd5eb
2022-04-01 15:58:32 +00:00
$expiry = $config->get( MainConfigNames::ReportToExpiry );
Reporting API and Feature Policy reporting support These are experimental APIs that lets one report policy violations: https://www.w3.org/TR/reporting/ https://github.com/w3c/webappsec-feature-policy/blob/master/reporting.md Bug: T209572 Change-Id: I002e7802000ec37b3320d8ae761cc1888d4f6edb
2019-03-22 05:43:30 +00:00
if ( !$expiry ) {
return false;
}
Use name constants to access config settings. Use name constants instead of string literals in calls to Config::get and ServiceOptions::get, when referring to core configuration variables. This protects against typos and makes the decumentation and schema declaration of the config settings discoverable. This is the first batch, only touching files directly under /includes/ Change-Id: I7252e636c7c86d950d9257b33491af492c6dd5eb
2022-04-01 15:58:32 +00:00
$endpoints = $config->get( MainConfigNames::ReportToEndpoints );
Reporting API and Feature Policy reporting support These are experimental APIs that lets one report policy violations: https://www.w3.org/TR/reporting/ https://github.com/w3c/webappsec-feature-policy/blob/master/reporting.md Bug: T209572 Change-Id: I002e7802000ec37b3320d8ae761cc1888d4f6edb
2019-03-22 05:43:30 +00:00
if ( !$endpoints ) {
return false;
}
$output = [ 'max_age' => $expiry, 'endpoints' => [] ];
foreach ( $endpoints as $endpoint ) {
$output['endpoints'][] = [ 'url' => $endpoint ];
}
return json_encode( $output, JSON_UNESCAPED_SLASHES );
}
private function getFeaturePolicyReportOnly() {
$config = $this->getConfig();
Use name constants to access config settings. Use name constants instead of string literals in calls to Config::get and ServiceOptions::get, when referring to core configuration variables. This protects against typos and makes the decumentation and schema declaration of the config settings discoverable. This is the first batch, only touching files directly under /includes/ Change-Id: I7252e636c7c86d950d9257b33491af492c6dd5eb
2022-04-01 15:58:32 +00:00
$features = $config->get( MainConfigNames::FeaturePolicyReportOnly );
Reporting API and Feature Policy reporting support These are experimental APIs that lets one report policy violations: https://www.w3.org/TR/reporting/ https://github.com/w3c/webappsec-feature-policy/blob/master/reporting.md Bug: T209572 Change-Id: I002e7802000ec37b3320d8ae761cc1888d4f6edb
2019-03-22 05:43:30 +00:00
return implode( ';', $features );
}
Step two in fixing OutputPage's documentation, more or less the same as r61647
2010-01-29 21:32:45 +00:00
/**
* Send cache control HTTP headers
*/
Add some documentation to OutputPage.php and specify public/private for most of its member functions (practically all public anyway). And replace one gratuitous use of strcmp with good old ===.
2006-11-07 05:37:31 +00:00
public function sendCacheControl() {
Continue with r85240; Move getSkin from User to RequestContext, do it without globals, strip out the non-functional $title related stuff, and update Skin to use a RequestContext.
2011-04-03 12:46:36 +00:00
$response = $this->getRequest()->response();
Use config instead of globals for OutputPage Change-Id: I5e0ebc173631d1d1052de7ccee4ef839c7c1767f
2014-08-23 08:52:41 +00:00
$config = $this->getConfig();
Revert r44801 "Tweaks from profiling" Mostly seems to be formatting tweaks, loss of useful debug log output, and mysterious unexplained changes. If some of these tweaks actually are based on profiling data, please provide details, such as "rearranging this call reduces service time from 80ms to 50ms for client cache hits on my machine" or whatever.
2008-12-23 19:39:00 +00:00
Make OutputPage's mechanism for gathering Vary headers reusable Maybe WebResponse would be a better place for this, but this'll work for now. * Add getVaryHeader() to get the serialized version of $mVaryHeader. * Remove 'Cookie' from the initializer, and add it explicitly later ** This allows reuse by code that doesn't want 'Cookie' in there Change-Id: Ifa8f74a21f832154b05e54ced95b9f9eb6052fcd
2012-06-01 10:23:53 +00:00
$this->addVaryHeader( 'Cookie' );
follow-up r59522. follow Tim's suggestion to serve same vary and XVO headers for the same URL. Now the vary headers should be consistent. Feel free to test it.
2009-12-21 18:55:42 +00:00
$this->addAcceptLanguage();
Revert r44801 "Tweaks from profiling" Mostly seems to be formatting tweaks, loss of useful debug log output, and mysterious unexplained changes. If some of these tweaks actually are based on profiling data, please provide details, such as "rearranging this call reduces service time from 80ms to 50ms for client cache hits on my machine" or whatever.
2008-12-23 19:39:00 +00:00
# don't serve compressed data to clients who can't handle it
update comment from live copy
2005-03-08 02:58:43 +00:00
# maintain different caches for logged-in users and non-logged in ones
Make OutputPage's mechanism for gathering Vary headers reusable Maybe WebResponse would be a better place for this, but this'll work for now. * Add getVaryHeader() to get the serialized version of $mVaryHeader. * Remove 'Cookie' from the initializer, and add it explicitly later ** This allows reuse by code that doesn't want 'Cookie' in there Change-Id: Ifa8f74a21f832154b05e54ced95b9f9eb6052fcd
2012-06-01 10:23:53 +00:00
$response->header( $this->getVaryHeader() );
Add an X-Vary-Options header, as per my proposed squid patch
2008-02-08 07:12:38 +00:00
else if -> elseif Also fixed some spacing while at it Added some braces for one line statements Change-Id: Iebfa2b17091509daabc76248121c7763e2295d20
2013-04-17 14:52:47 +00:00
if ( $this->mEnableClientCache ) {
Use name constants to access config settings. Use name constants instead of string literals in calls to Config::get and ServiceOptions::get, when referring to core configuration variables. This protects against typos and makes the decumentation and schema declaration of the config settings discoverable. This is the first batch, only touching files directly under /includes/ Change-Id: I7252e636c7c86d950d9257b33491af492c6dd5eb
2022-04-01 15:58:32 +00:00
if ( !$config->get( MainConfigNames::UseCdn ) ) {
Log the reason for private caching In OutputPage::sendCacheControl(), break up the multi-line conditional and include in the debug log message the reason for private cache headers being sent. I think this improves the readability of the code. Change-Id: I0229e04c44771245b7ac8cd45d2c83062a4d78f2
2020-11-05 02:12:42 +00:00
$privateReason = 'config';
} elseif ( $response->hasCookies() ) {
$privateReason = 'set-cookies';
// The client might use methods other than cookies to appear logged-in.
// E.g. HTTP headers, or query parameter tokens, OAuth, etc.
} elseif ( SessionManager::getGlobalSession()->isPersistent() ) {
$privateReason = 'session';
} elseif ( $this->isPrintable() ) {
$privateReason = 'printable';
} elseif ( $this->mCdnMaxage == 0 ) {
$privateReason = 'no-maxage';
} elseif ( $this->haveCacheVaryCookies() ) {
$privateReason = 'cache-vary-cookies';
} else {
$privateReason = false;
}
if ( $privateReason === false ) {
Remove support for wgUseESI Experimental feature deprecated since 1.33 (d63684a064b09). Change-Id: I9bab4ad1cd1f2c80a3dfcccdf99ed011c9c8a440
2019-09-08 04:11:22 +00:00
# We'll purge the proxy cache for anons explicitly, but require end user agents
# to revalidate against the proxy on each visit.
# IMPORTANT! The CDN needs to replace the Cache-Control header with
# Cache-Control: s-maxage=0, must-revalidate, max-age=0
wfDebug( __METHOD__ .
": local proxy caching; {$this->mLastModified} **", 'private' );
# start with a shorter timeout for initial testing
# header( "Cache-Control: s-maxage=2678400, must-revalidate, max-age=0" );
$response->header( "Cache-Control: " .
"s-maxage={$this->mCdnMaxage}, must-revalidate, max-age=0" );
Some cleanup to header output; use a forced expires header again to force some older browsers (?); use private caching; separate out the local squid case from the general case
2004-01-31 10:29:31 +00:00
} else {
# We do want clients to cache if they can, but they *must* check for updates
OutputPage: Remove $wgLoggedOutMaxAge experiment and related tests Bug: T293848 Change-Id: I5e5e52fac7fe389b36715dae11cc30cdb2b9136a
2021-10-20 16:16:03 +00:00
# on revisiting the page.
Log the reason for private caching In OutputPage::sendCacheControl(), break up the multi-line conditional and include in the debug log message the reason for private cache headers being sent. I think this improves the readability of the code. Change-Id: I0229e04c44771245b7ac8cd45d2c83062a4d78f2
2020-11-05 02:12:42 +00:00
wfDebug( __METHOD__ . ": private caching ($privateReason); {$this->mLastModified} **", 'private' );
Make max-age configurable for logged-out users For small wikis. When a CDN is present, the value is overwritten by the CDN. Bug: T231184 Change-Id: I45a53571d0f5f6316eb97b396c18c0b224586526
2019-08-26 13:16:43 +00:00
OutputPage: Remove $wgLoggedOutMaxAge experiment and related tests Bug: T293848 Change-Id: I5e5e52fac7fe389b36715dae11cc30cdb2b9136a
2021-10-20 16:16:03 +00:00
$response->header( 'Expires: ' . gmdate( 'D, d M Y H:i:s', 0 ) . ' GMT' );
$response->header( "Cache-Control: private, must-revalidate, max-age=0" );
Some cleanup to header output; use a forced expires header again to force some older browsers (?); use private caching; separate out the local squid case from the general case
2004-01-31 10:29:31 +00:00
}
else if -> elseif Also fixed some spacing while at it Added some braces for one line statements Change-Id: Iebfa2b17091509daabc76248121c7763e2295d20
2013-04-17 14:52:47 +00:00
if ( $this->mLastModified ) {
* Generalised OutputPage::checkLastModified() to allow it to easily handle more than three types of modified timestamp. Added OutputPageCheckLastModified hook. * Flattened out the structure of that function with early returns, to avoid ugly deep nesting.
2008-08-29 08:40:13 +00:00
$response->header( "Last-Modified: {$this->mLastModified}" );
}
Fixes missing cache-control and last-modified headers with file cached pages. HOWEVER, the session cookie seems to override our cache control. May need to mess with this to get things working smoothly. Also, caching + gzip may be a difficult combination.
2003-07-03 10:18:07 +00:00
} else {
Remove use of $dest = 'log' for wfDebugLog The functionality for marking a debug log message for delivery to a log file but not to MWDebug was removed in 1.25 with the introduction of the PSR-3 logging system. Convert the messages that were marked for this special handling in a7a0883 to use 'private' instead. Bug: T122644 Change-Id: Iefaac07a5922c16c2899904c7b678199c5b1efe9
2015-12-30 19:48:01 +00:00
wfDebug( __METHOD__ . ": no caching **", 'private' );
don't cache error pages
2004-03-23 10:19:31 +00:00
# In general, the absence of a last modified header should be enough to prevent
# the client from using its cache. We send a few other things just to make sure.
Add an X-Vary-Options header, as per my proposed squid patch
2008-02-08 07:12:38 +00:00
$response->header( 'Expires: ' . gmdate( 'D, d M Y H:i:s', 0 ) . ' GMT' );
$response->header( 'Cache-Control: no-cache, no-store, max-age=0, must-revalidate' );
$response->header( 'Pragma: no-cache' );
Fixes missing cache-control and last-modified headers with file cached pages. HOWEVER, the session cookie seems to override our cache control. May need to mess with this to get things working smoothly. Also, caching + gzip may be a difficult combination.
2003-07-03 10:18:07 +00:00
}
}
This commit implements Josh Cogliati's "info" feature (e.g. http://en.wikipedia.org/w/wiki.phtml?title=Wikipedia&action=info), fixes the page headings for protect/unprotect, and does a lot of whitespace changes (sorry about that). I've also fixed confusing indentation somewhere.
2004-07-08 14:53:54 +00:00
Skins: getDefaultStyles can now define render blocking CSS This optimisation attempts to minimise loading the styles in places they are not needed. The logic is kept inside Skin::getDefaultModules to avoid fragmentation of where modules get defined. Update ApiParse to avoid repetition of code. Bug: T42792 Bug: T42812 Change-Id: I59f02a7bab3baa9d43f6bc2ef1f549d9d31d8456
2018-04-10 01:22:13 +00:00
/**
* Transfer styles and JavaScript modules from skin.
*
* @param Skin $sk to load modules for
*/
public function loadSkinModules( $sk ) {
foreach ( $sk->getDefaultModules() as $group => $modules ) {
if ( $group === 'styles' ) {
Remove unused key variable from foreach loops Change-Id: Id2d91e30a6f7cc4eb93427b50efc1c5c77f14b75
2022-09-21 19:05:03 +00:00
foreach ( $modules as $moduleMembers ) {
Skins: getDefaultStyles can now define render blocking CSS This optimisation attempts to minimise loading the styles in places they are not needed. The logic is kept inside Skin::getDefaultModules to avoid fragmentation of where modules get defined. Update ApiParse to avoid repetition of code. Bug: T42792 Bug: T42812 Change-Id: I59f02a7bab3baa9d43f6bc2ef1f549d9d31d8456
2018-04-10 01:22:13 +00:00
$this->addModuleStyles( $moduleMembers );
}
} else {
$this->addModules( $modules );
}
}
}
Changing comments layout preparing for generated documentation with Phpdocumentor
2004-09-02 23:28:24 +00:00
/**
* Finally, all the text has been munged and accumulated into
* the object, let's actually output it:
Support masking the WRITE_SYNC latency from ChronologyProtector * Use OutputPage::output() as the method to mask latency, since it takes a good while to run. By the time it runs, cache replication should have caught up, so the reap call will likely not block. * For redirects emitted after changes in POST, instead of masking with OutputPage, add a parameter to the redirect and block on the positions appearing. This uses the redirection RTT to mask the replication latency. Change-Id: Ib23690c302e8033610fef9a0ef451dafe8a5803e
2016-08-24 21:22:11 +00:00
*
* @param bool $return Set to true to get the result as a string rather than sending it
* @return string|null
* @throws Exception
* @throws FatalError
* @throws MWException
Changing comments layout preparing for generated documentation with Phpdocumentor
2004-09-02 23:28:24 +00:00
*/
Support masking the WRITE_SYNC latency from ChronologyProtector * Use OutputPage::output() as the method to mask latency, since it takes a good while to run. By the time it runs, cache replication should have caught up, so the reap call will likely not block. * For redirects emitted after changes in POST, instead of masking with OutputPage, add a parameter to the redirect and block on the positions appearing. This uses the redirection RTT to mask the replication latency. Change-Id: Ib23690c302e8033610fef9a0ef451dafe8a5803e
2016-08-24 21:22:11 +00:00
public function output( $return = false ) {
else if -> elseif Also fixed some spacing while at it Added some braces for one line statements Change-Id: Iebfa2b17091509daabc76248121c7763e2295d20
2013-04-17 14:52:47 +00:00
if ( $this->mDoNothing ) {
Support masking the WRITE_SYNC latency from ChronologyProtector * Use OutputPage::output() as the method to mask latency, since it takes a good while to run. By the time it runs, cache replication should have caught up, so the reap call will likely not block. * For redirects emitted after changes in POST, instead of masking with OutputPage, add a parameter to the redirect and block on the positions appearing. This uses the redirection RTT to mask the replication latency. Change-Id: Ib23690c302e8033610fef9a0ef451dafe8a5803e
2016-08-24 21:22:11 +00:00
return $return ? '' : null;
Revert r45063 "Move doNothing return check above global declarations" Change makes the code harder to read by hiding a control statement in the middle of a global declaraction chunk.
2008-12-31 17:56:04 +00:00
}
* Moved all <link> definitions in OutputPage::getHeadLinks() instead of having them in a *lot* of different functions * Also moved there generic <meta> and removed OutputPage::addDefaultMeta() with its $called static local variable which was breaking the output when generating multiple pages on the same request (rebuildFileCache.php, dumpHTML.php) since that function could only be executed completely once for all instances, and not once per instance * Moved default module from OutputPage::output() to its own function and don't call it when executing a body only request, since it's useless in that case * Call Skin::setMembers() from Skin::initPage() instead of Skin::outputPage()
2011-04-02 18:38:42 +00:00
Continue with r85240; Move getSkin from User to RequestContext, do it without globals, strip out the non-functional $title related stuff, and update Skin to use a RequestContext.
2011-04-03 12:46:36 +00:00
$response = $this->getRequest()->response();
Use config instead of globals for OutputPage Change-Id: I5e0ebc173631d1d1052de7ccee4ef839c7c1767f
2014-08-23 08:52:41 +00:00
$config = $this->getConfig();
* Moved all <link> definitions in OutputPage::getHeadLinks() instead of having them in a *lot* of different functions * Also moved there generic <meta> and removed OutputPage::addDefaultMeta() with its $called static local variable which was breaking the output when generating multiple pages on the same request (rebuildFileCache.php, dumpHTML.php) since that function could only be executed completely once for all instances, and not once per instance * Moved default module from OutputPage::output() to its own function and don't call it when executing a body only request, since it's useless in that case * Call Skin::setMembers() from Skin::initPage() instead of Skin::outputPage()
2011-04-02 18:38:42 +00:00
Attempt at normalistion of comparison styles - empty string on left and right hand side normalised to it being on the rhs Before this change, there were (? being regex 0 or 1) "" ===? 1 '' ===? 24 "" !==? 8 '' !==? 32 == "" 14 == '' 344 != "" 9 !== "" 4 != '' 151 !== '' 85 Rhs was the much more common, and the preferred style by many developers.. (Was a similar discussion in #mediawiki recently.. After that lolbugreport i think) Where there is a string (non empty) on the lhs, and variable/method call on the rhs still need normalising
2010-01-06 19:59:42 +00:00
if ( $this->mRedirect != '' ) {
* (bug 12935, 12981) Fully-qualify archive URLs in delete, revert messages Replaced a bunch of instances of 'if-this-is-not-fully-qualified-prepend-$wgServer' with a function, wfExpandUrl()
2008-02-13 05:59:14 +00:00
# Standards require redirect URLs to be absolute
Make the PROTO_ parameter to wfExpandUrl() explicit in all calls. There are now no calls to wfExpandUrl() in core that rely on PROTO_CURRENT being the default. Also switch to using getCanonicalUrl() for.... rel="canonical" :D (I picked the right name, I KNEW IT)
2011-08-19 15:46:08 +00:00
$this->mRedirect = wfExpandUrl( $this->mRedirect, PROTO_CURRENT );
Add BeforePageRedirect hook to OutputPage, allowing extensions to override redirect output. This is needed by MobileFrontend to normalize some redirects to the mobile site, such as on login.
2011-12-14 00:38:21 +00:00
$redirect = $this->mRedirect;
$code = $this->mRedirectCode;
OutputPage: Make $wgDebugRedirects work again This was broken by I2dd2b71. In that patch MediaWiki::outputResponsePayload() was introduced and it added another output buffer layer on top of Outputpage::output( true ) which directly outputs and return empty string if $wgDebugRedirects is true (and the location is redirect). Because Outputpage::output( true ) returns empty string in that case, the output buffer started by MediaWiki::outputResponsePayload() will always end up with empty string, and the print() in Outputpage cannot send the output because it's now nested inside another buffer layer with call to ob_end_flush(), which discards contents of inner buffers. Previously it worked because the return value of Outputpage::output() was not wrapped in another output buffer then, so the direct print of the html was enough to make it work. This patch modified Outputpage::output() to return the html string if $wgDebugRedirects is on, and return value is requested --instead of outputting it-- that way the debug html will now be captured in the topmost buffer layer inside MediaWiki::outputResponsePayload() Bug: T236639 Change-Id: Ibe4e771b0530b7c6bb7132c88c8455b0ef4fde0e
2020-12-04 06:02:19 +00:00
$content = '';
Add BeforePageRedirect hook to OutputPage, allowing extensions to override redirect output. This is needed by MobileFrontend to normalize some redirects to the mobile site, such as on login.
2011-12-14 00:38:21 +00:00
Hooks::run() call site migration Migrate all callers of Hooks::run() to use the new HookContainer/HookRunner system. General principles: * Use DI if it is already used. We're not changing the way state is managed in this patch. * HookContainer is always injected, not HookRunner. HookContainer is a service, it's a more generic interface, it is the only thing that provides isRegistered() which is needed in some cases, and a HookRunner can be efficiently constructed from it (confirmed by benchmark). Because HookContainer is needed for object construction, it is also needed by all factories. * "Ask your friendly local base class". Big hierarchies like SpecialPage and ApiBase have getHookContainer() and getHookRunner() methods in the base class, and classes that extend that base class are not expected to know or care where the base class gets its HookContainer from. * ProtectedHookAccessorTrait provides protected getHookContainer() and getHookRunner() methods, getting them from the global service container. The point of this is to ease migration to DI by ensuring that call sites ask their local friendly base class rather than getting a HookRunner from the service container directly. * Private $this->hookRunner. In some smaller classes where accessor methods did not seem warranted, there is a private HookRunner property which is accessed directly. Very rarely (two cases), there is a protected property, for consistency with code that conventionally assumes protected=private, but in cases where the class might actually be overridden, a protected accessor is preferred over a protected property. * The last resort: Hooks::runner(). Mostly for static, file-scope and global code. In a few cases it was used for objects with broken construction schemes, out of horror or laziness. Constructors with new required arguments: * AuthManager * BadFileLookup * BlockManager * ClassicInterwikiLookup * ContentHandlerFactory * ContentSecurityPolicy * DefaultOptionsManager * DerivedPageDataUpdater * FullSearchResultWidget * HtmlCacheUpdater * LanguageFactory * LanguageNameUtils * LinkRenderer * LinkRendererFactory * LocalisationCache * MagicWordFactory * MessageCache * NamespaceInfo * PageEditStash * PageHandlerFactory * PageUpdater * ParserFactory * PermissionManager * RevisionStore * RevisionStoreFactory * SearchEngineConfig * SearchEngineFactory * SearchFormWidget * SearchNearMatcher * SessionBackend * SpecialPageFactory * UserNameUtils * UserOptionsManager * WatchedItemQueryService * WatchedItemStore Constructors with new optional arguments: * DefaultPreferencesFactory * Language * LinkHolderArray * MovePage * Parser * ParserCache * PasswordReset * Router setHookContainer() now required after construction: * AuthenticationProvider * ResourceLoaderModule * SearchEngine Change-Id: Id442b0dbe43aba84bd5cf801d86dedc768b082c7
2020-03-19 02:42:09 +00:00
if ( $this->getHookRunner()->onBeforePageRedirect( $this, $redirect, $code ) ) {
else if -> elseif Also fixed some spacing while at it Added some braces for one line statements Change-Id: Iebfa2b17091509daabc76248121c7763e2295d20
2013-04-17 14:52:47 +00:00
if ( $code == '301' || $code == '303' ) {
Use name constants to access config settings. Use name constants instead of string literals in calls to Config::get and ServiceOptions::get, when referring to core configuration variables. This protects against typos and makes the decumentation and schema declaration of the config settings discoverable. This is the first batch, only touching files directly under /includes/ Change-Id: I7252e636c7c86d950d9257b33491af492c6dd5eb
2022-04-01 15:58:32 +00:00
if ( !$config->get( MainConfigNames::DebugRedirects ) ) {
Add explicit casts between scalar types * Some functions accept only string, cast ints and floats to string * After preg_matches or explode() casts numbers to int to do maths * Cast unix timestamps to int to do maths * Cast return values from timestamp format function to int * Cast bitwise operator to bool when needed as bool * php internal functions like floor/round/ceil documented to return float, most cases the result is used as int, added casts Found by phan strict checks Change-Id: Icb2de32107f43817acc45fe296fb77acf65c1786
2022-02-26 16:28:48 +00:00
$response->statusHeader( (int)$code );
Add BeforePageRedirect hook to OutputPage, allowing extensions to override redirect output. This is needed by MobileFrontend to normalize some redirects to the mobile site, such as on login.
2011-12-14 00:38:21 +00:00
}
$this->mLastModified = wfTimestamp( TS_RFC2822 );
}
Use name constants to access config settings. Use name constants instead of string literals in calls to Config::get and ServiceOptions::get, when referring to core configuration variables. This protects against typos and makes the decumentation and schema declaration of the config settings discoverable. This is the first batch, only touching files directly under /includes/ Change-Id: I7252e636c7c86d950d9257b33491af492c6dd5eb
2022-04-01 15:58:32 +00:00
if ( $config->get( MainConfigNames::VaryOnXFP ) ) {
Add BeforePageRedirect hook to OutputPage, allowing extensions to override redirect output. This is needed by MobileFrontend to normalize some redirects to the mobile site, such as on login.
2011-12-14 00:38:21 +00:00
$this->addVaryHeader( 'X-Forwarded-Proto' );
}
$this->sendCacheControl();
exception: Simplify MWExceptionRenderer to reduce influence of config Remove use of two configuration globals. This follows-up to commit 47adb6d65a (I1a691f01cd82e60) which aimed to access all config via MediaWikiServices, but that isn't safe during error handling. These two were only used in the low-level "plain text" and "basic HTML" error pages, which already can't have any branding, skinning or localisation; and are not how most exceptions are rendered. in those edge cases, we can use the name of the software instead of (trying) to use the name of the site. == Remove use of $wgSitename == In msg() and reportHTML(), remove use of $wgSitename in favour of a generic fallback matching DefaulSettings.php. This does not affect the common case of runtime fatals and timeouts, as we're only changing the fallback after wfMessage() fails in msg(), or when OutputPage is unavailable, which is typically only if services, localisation or DB are also down (or not yet loaded). Most exceptions happen when and after those have initialised fine. Test case 1: (Clean state) Edit ViewAction::show() to add `throw new RuntimeException();` as its first statement, then try to view the main page. This error page is unchanged. It is skinned, localised, and still uses the configured sitename in the doc title. Test case 2: Edit index.php to call foo() instead of wfIndexMain(), then try to view the main page. Before, this "minimal HTML" error page would have a doc title of "Internal error - MyWiki", and now "Internal error - MediaWiki". Test case 3: (Clean state) Edit Message::text() to add `throw new RuntimeException();`, then try to view the main page. This results in a "plain text" error page that doesn't even have an HTML doc, and is also unchanged. == Remove use of $wgMimeType == In output(), remove use of $wgMimeType and remove the Content-Type header that it was used for. This was redundant because the next statements (reportHTML) already outputs Content-type. In reportHTML, we sometimes delegate to OutputPage (if safe) and that honours $wgMimeType already. In other cases, it is handled inline in with a basic HTML error page, and that branch also sets the Content-Type header already. In one case (reportOutageHTML) a header was not yet set. For that one, I've added the missing header call and made it explicitly text/html. This is technically a bugfix, because our basic HTML error page is HTML5, whereas $wgMimeType (which exists to allow enabling XHTML) can be XHTML which we weren't following. In OutputPage and Html::htmlHeader that would normally result in outputting `<?xml`. Test case: Edit ViewAction::show(), and add `throw new RuntimeException();` as its first statement, then try to view the main page. In devtools>network, there is still a proper Content-Type and charset on the error page document. Change-Id: I03cfa2b6155fb711582164852e7cab4c325a1b92
2022-01-12 23:05:34 +00:00
$response->header( 'Content-Type: text/html; charset=UTF-8' );
Use name constants to access config settings. Use name constants instead of string literals in calls to Config::get and ServiceOptions::get, when referring to core configuration variables. This protects against typos and makes the decumentation and schema declaration of the config settings discoverable. This is the first batch, only touching files directly under /includes/ Change-Id: I7252e636c7c86d950d9257b33491af492c6dd5eb
2022-04-01 15:58:32 +00:00
if ( $config->get( MainConfigNames::DebugRedirects ) ) {
Add BeforePageRedirect hook to OutputPage, allowing extensions to override redirect output. This is needed by MobileFrontend to normalize some redirects to the mobile site, such as on login.
2011-12-14 00:38:21 +00:00
$url = htmlspecialchars( $redirect );
OutputPage: Make $wgDebugRedirects work again This was broken by I2dd2b71. In that patch MediaWiki::outputResponsePayload() was introduced and it added another output buffer layer on top of Outputpage::output( true ) which directly outputs and return empty string if $wgDebugRedirects is true (and the location is redirect). Because Outputpage::output( true ) returns empty string in that case, the output buffer started by MediaWiki::outputResponsePayload() will always end up with empty string, and the print() in Outputpage cannot send the output because it's now nested inside another buffer layer with call to ob_end_flush(), which discards contents of inner buffers. Previously it worked because the return value of Outputpage::output() was not wrapped in another output buffer then, so the direct print of the html was enough to make it work. This patch modified Outputpage::output() to return the html string if $wgDebugRedirects is on, and return value is requested --instead of outputting it-- that way the debug html will now be captured in the topmost buffer layer inside MediaWiki::outputResponsePayload() Bug: T236639 Change-Id: Ibe4e771b0530b7c6bb7132c88c8455b0ef4fde0e
2020-12-04 06:02:19 +00:00
$content = "<!DOCTYPE html>\n<html>\n<head>\n"
. "<title>Redirect</title>\n</head>\n<body>\n"
. "<p>Location: <a href=\"$url\">$url</a></p>\n"
. "</body>\n</html>\n";
if ( !$return ) {
print $content;
}
Add BeforePageRedirect hook to OutputPage, allowing extensions to override redirect output. This is needed by MobileFrontend to normalize some redirects to the mobile site, such as on login.
2011-12-14 00:38:21 +00:00
} else {
$response->header( 'Location: ' . $redirect );
Add $wgDebugRedirects to assist in debugging some problems hidden by automated http redirects.
2004-03-05 03:18:31 +00:00
}
}
Add BeforePageRedirect hook to OutputPage, allowing extensions to override redirect output. This is needed by MobileFrontend to normalize some redirects to the mobile site, such as on login.
2011-12-14 00:38:21 +00:00
OutputPage: Make $wgDebugRedirects work again This was broken by I2dd2b71. In that patch MediaWiki::outputResponsePayload() was introduced and it added another output buffer layer on top of Outputpage::output( true ) which directly outputs and return empty string if $wgDebugRedirects is true (and the location is redirect). Because Outputpage::output( true ) returns empty string in that case, the output buffer started by MediaWiki::outputResponsePayload() will always end up with empty string, and the print() in Outputpage cannot send the output because it's now nested inside another buffer layer with call to ob_end_flush(), which discards contents of inner buffers. Previously it worked because the return value of Outputpage::output() was not wrapped in another output buffer then, so the direct print of the html was enough to make it work. This patch modified Outputpage::output() to return the html string if $wgDebugRedirects is on, and return value is requested --instead of outputting it-- that way the debug html will now be captured in the topmost buffer layer inside MediaWiki::outputResponsePayload() Bug: T236639 Change-Id: Ibe4e771b0530b7c6bb7132c88c8455b0ef4fde0e
2020-12-04 06:02:19 +00:00
return $return ? $content : null;
* (bug 21574) Redirects can now have "303 See Other" HTTP status Based on a patch by denny vrandecic - http://bug-attachment.wikimedia.org/attachment.cgi?id=6805
2009-12-01 20:56:43 +00:00
} elseif ( $this->mStatusCode ) {
WebResponse: Implement statusHeader() using the new HttpStatus::header() * Convert existing use of WebResponse::header() for HTTP status headers to use this new statusHeader() method. * Extend unit test forFauxResponse. I'm not calling HttpStatus::header directly in code. We keep the abstraction layer of WebResponse so that responses can continue to be mocked/fauxed without affecting the outer HTTP response. Change-Id: I8a536e16659fa88b54cffa1457efb889efa5fcd6
2015-05-24 12:31:11 +00:00
$response->statusHeader( $this->mStatusCode );
* integrating UniversalEditButton extension into core (enabled by default, set $wgUniversalEditButton = false; in LocalSettings to disable)
2008-10-10 01:15:11 +00:00
}
trim trailing spaces
2010-05-22 11:47:56 +00:00
(bug 26) Defer sending cache headers until the skin runs, so that the skin can declare the page as uncacheable
2004-11-12 09:34:11 +00:00
# Buffer output; final headers may depend on later processing
Addendum to r40837: only validate/clean the body text, as we can assume the rest of the skin is valid.
2008-09-15 01:27:22 +00:00
ob_start();
(bug 26) Defer sending cache headers until the skin runs, so that the skin can declare the page as uncacheable
2004-11-12 09:34:11 +00:00
Use name constants to access config settings. Use name constants instead of string literals in calls to Config::get and ServiceOptions::get, when referring to core configuration variables. This protects against typos and makes the decumentation and schema declaration of the config settings discoverable. This is the first batch, only touching files directly under /includes/ Change-Id: I7252e636c7c86d950d9257b33491af492c6dd5eb
2022-04-01 15:58:32 +00:00
$response->header( 'Content-type: ' . $config->get( MainConfigNames::MimeType ) . '; charset=UTF-8' );
Mass conversion of $wgContLang to service Brought to you by vim macros. Bug: T200246 Change-Id: I79e919f4553e3bd3eb714073fed7a43051b4fb2a
2018-07-29 12:24:54 +00:00
$response->header( 'Content-language: ' .
MediaWikiServices::getInstance()->getContentLanguage()->getHtmlCode() );
Initial revision
2003-04-14 23:10:40 +00:00
Preload the logo using link rel="preload" http header This greatly increases the priority of loading the logo on browsers that support rel="preload". Bug: T100999 Change-Id: I0738fcc0a575153dab65016fa87faaa9b8b97a9d
2015-06-01 16:58:42 +00:00
$linkHeader = $this->getLinkHeader();
if ( $linkHeader ) {
$response->header( $linkHeader );
}
Fix for bug 26561: clickjacking attacks. See the bug report for full documentation.
2011-01-04 06:12:33 +00:00
// Prevent framing, if requested
$frameOptions = $this->getFrameOptions();
if ( $frameOptions ) {
* Moved all <link> definitions in OutputPage::getHeadLinks() instead of having them in a *lot* of different functions * Also moved there generic <meta> and removed OutputPage::addDefaultMeta() with its $called static local variable which was breaking the output when generating multiple pages on the same request (rebuildFileCache.php, dumpHTML.php) since that function could only be executed completely once for all instances, and not once per instance * Moved default module from OutputPage::output() to its own function and don't call it when executing a body only request, since it's useless in that case * Call Skin::setMembers() from Skin::initPage() instead of Skin::outputPage()
2011-04-02 18:38:42 +00:00
$response->header( "X-Frame-Options: $frameOptions" );
Fix for bug 26561: clickjacking attacks. See the bug report for full documentation.
2011-01-04 06:12:33 +00:00
}
Cleanup: Remove some tiny private methods in favor of inlining the code HistoryPager: remove ::preventClickjacking OutputPage: remove ::getOriginTrials RevisionStore: remove ::getDBLoadBalancer Change-Id: Id60d9861ffd86d9a82ee7f56da939274b1be82e0
2020-05-18 03:12:44 +00:00
// Get the Origin-Trial header values. This is used to enable Chrome Origin
// Trials: https://github.com/GoogleChrome/OriginTrials
$originTrials = $config->get( MainConfigNames::OriginTrials );
Chrome Origin Trial support Bug: T216595 Change-Id: I74b8824a87e6f52a84d2f8a9708501fc05f73266
2019-02-20 08:49:28 +00:00
foreach ( $originTrials as $originTrial ) {
$response->header( "Origin-Trial: $originTrial", false );
}
Reporting API and Feature Policy reporting support These are experimental APIs that lets one report policy violations: https://www.w3.org/TR/reporting/ https://github.com/w3c/webappsec-feature-policy/blob/master/reporting.md Bug: T209572 Change-Id: I002e7802000ec37b3320d8ae761cc1888d4f6edb
2019-03-22 05:43:30 +00:00
$reportTo = $this->getReportTo();
if ( $reportTo ) {
$response->header( "Report-To: $reportTo" );
}
$featurePolicyReportOnly = $this->getFeaturePolicyReportOnly();
if ( $featurePolicyReportOnly ) {
$response->header( "Feature-Policy-Report-Only: $featurePolicyReportOnly" );
}
*coding style cleanup (added braces, spacing tweaks etc.) *removed some unused globals
2010-05-22 12:18:22 +00:00
if ( $this->mArticleBodyOnly ) {
Allow adjusting CSP headers from BeforePageDisplay hook Previously this hook ran after the header was sent, so it was very confusing why CSP adjustments were being ignored by it. This was especially the case, as it is a very natural point to adjust the CSP policy, as commonly it is used to adjust what modules are loaded. Bug: T246614 Change-Id: I38f8f43fdc1d9bbd2f9af04305aad7a370b107b6
2020-03-02 09:11:14 +00:00
$this->CSP->sendHeaders();
Deprecate OutputPage::out. OutputPage is supposed to be a container for output. It should NOT be used as a replacement for echo. Only one seemingly unmaintained extension uses this method. This method is deprecated now and should be removed in the next release. Change-Id: I82711cee7204604a47cfbb5e4496b4cc737a837c
2013-05-12 09:00:16 +00:00
echo $this->mBodytext;
implement "action=render"
2005-07-03 04:00:33 +00:00
} else {
resourceloader: Apply safemode to startup module registry This effectively applies safemode to the mw.loader client, without the client itself needing specific knowledge of safemode. Test Plan: * Unchanged: When viewing a page in safemode, the 'user' and 'site' modules are still not queued by OutputPage. * New: mw.loader.getState('site'), previously would yield 'registered', but will now yield null. * New: mw.loader.load('site'), previously loaded the module, it now logs a dependency warning for unknown module, like for any other unknown module. * New: mw.loader.using('site'), previously resolved, it is now rejected. Bug: T185303 Change-Id: I672e3891c8e1b3c2d13655fa134d0f1d031b8247
2018-03-28 00:57:06 +00:00
// Enable safe mode if requested (T152169)
Add ?safemode=1 to disable user JS/CSS Bug: T152169 Change-Id: I112b7abf1cfa1330b2d4e6150018aad14c00b654
2017-03-29 21:01:11 +00:00
if ( $this->getRequest()->getBool( 'safemode' ) ) {
$this->disallowUserJs();
}
Updated calls to Linker to call them statically and removed useless parameter to OutputPage::getSkin()
2011-06-22 18:24:58 +00:00
$sk = $this->getSkin();
Skins: getDefaultStyles can now define render blocking CSS This optimisation attempts to minimise loading the styles in places they are not needed. The logic is kept inside Skin::getDefaultModules to avoid fragmentation of where modules get defined. Update ApiParse to avoid repetition of code. Bug: T42792 Bug: T42812 Change-Id: I59f02a7bab3baa9d43f6bc2ef1f549d9d31d8456
2018-04-10 01:22:13 +00:00
$this->loadSkinModules( $sk );
OutputPage: Move hardcoded default modules to Skin::getDefaultModules These modules should not be hardcoded in OutputPage::output() which makes them impossible to override and also very hard to retrieve through the API for action=parse. Move these instead to Skin which is where all other default module loading happens already. Moving these modules is in preparation for customising ApiParse to support "really" returning all would-be loaded modules on a page when setting 'useskin', which is also needed for Live Preview and in theory for ajax navigation and other scenarios where there is a delay between the "initial" page rendering, and a later re-render which may not have all all the necessary modules. Bug: T130632 Change-Id: Ic4afccf0cd0d428d7fbc36d4e747415af3ab49f5
2017-05-20 13:07:37 +00:00
Give skin more control of OutputPage modules Remove addDefaultModules from OutputPage Instead only enforce mediawiki.page.startup Add a method getDefaultModules which groups modules by type allowing a skin to tweak Change-Id: I89d529f0378d90af0fe0a5adea5d5dbdca83a86e
2013-05-13 22:09:33 +00:00
MWDebug::addModules( $this );
Only call User::getSkin() when necessary. (With this, a parser cache hit on action=render may not load the skin at all)
2011-03-11 14:31:12 +00:00
(bug 12064) BeforePageDisplay hook is now called for all skins
2008-04-06 20:25:45 +00:00
// Hook that allows last minute changes to the output page, e.g.
Allow adjusting CSP headers from BeforePageDisplay hook Previously this hook ran after the header was sent, so it was very confusing why CSP adjustments were being ignored by it. This was especially the case, as it is a very natural point to adjust the CSP policy, as commonly it is used to adjust what modules are loaded. Bug: T246614 Change-Id: I38f8f43fdc1d9bbd2f9af04305aad7a370b107b6
2020-03-02 09:11:14 +00:00
// adding of CSS or Javascript by extensions, adding CSP sources.
Hooks::run() call site migration Migrate all callers of Hooks::run() to use the new HookContainer/HookRunner system. General principles: * Use DI if it is already used. We're not changing the way state is managed in this patch. * HookContainer is always injected, not HookRunner. HookContainer is a service, it's a more generic interface, it is the only thing that provides isRegistered() which is needed in some cases, and a HookRunner can be efficiently constructed from it (confirmed by benchmark). Because HookContainer is needed for object construction, it is also needed by all factories. * "Ask your friendly local base class". Big hierarchies like SpecialPage and ApiBase have getHookContainer() and getHookRunner() methods in the base class, and classes that extend that base class are not expected to know or care where the base class gets its HookContainer from. * ProtectedHookAccessorTrait provides protected getHookContainer() and getHookRunner() methods, getting them from the global service container. The point of this is to ease migration to DI by ensuring that call sites ask their local friendly base class rather than getting a HookRunner from the service container directly. * Private $this->hookRunner. In some smaller classes where accessor methods did not seem warranted, there is a private HookRunner property which is accessed directly. Very rarely (two cases), there is a protected property, for consistency with code that conventionally assumes protected=private, but in cases where the class might actually be overridden, a protected accessor is preferred over a protected property. * The last resort: Hooks::runner(). Mostly for static, file-scope and global code. In a few cases it was used for objects with broken construction schemes, out of horror or laziness. Constructors with new required arguments: * AuthManager * BadFileLookup * BlockManager * ClassicInterwikiLookup * ContentHandlerFactory * ContentSecurityPolicy * DefaultOptionsManager * DerivedPageDataUpdater * FullSearchResultWidget * HtmlCacheUpdater * LanguageFactory * LanguageNameUtils * LinkRenderer * LinkRendererFactory * LocalisationCache * MagicWordFactory * MessageCache * NamespaceInfo * PageEditStash * PageHandlerFactory * PageUpdater * ParserFactory * PermissionManager * RevisionStore * RevisionStoreFactory * SearchEngineConfig * SearchEngineFactory * SearchFormWidget * SearchNearMatcher * SessionBackend * SpecialPageFactory * UserNameUtils * UserOptionsManager * WatchedItemQueryService * WatchedItemStore Constructors with new optional arguments: * DefaultPreferencesFactory * Language * LinkHolderArray * MovePage * Parser * ParserCache * PasswordReset * Router setHookContainer() now required after construction: * AuthenticationProvider * ResourceLoaderModule * SearchEngine Change-Id: Id442b0dbe43aba84bd5cf801d86dedc768b082c7
2020-03-19 02:42:09 +00:00
$this->getHookRunner()->onBeforePageDisplay( $this, $sk );
(bug 12064) BeforePageDisplay hook is now called for all skins
2008-04-06 20:25:45 +00:00
Allow adjusting CSP headers from BeforePageDisplay hook Previously this hook ran after the header was sent, so it was very confusing why CSP adjustments were being ignored by it. This was especially the case, as it is a very natural point to adjust the CSP policy, as commonly it is used to adjust what modules are loaded. Bug: T246614 Change-Id: I38f8f43fdc1d9bbd2f9af04305aad7a370b107b6
2020-03-02 09:11:14 +00:00
$this->CSP->sendHeaders();
Avoid double-rendering on late exceptions Bug: T129657 Change-Id: I86af834b842bc056b57b2a55a9e1385481d5781d
2016-03-18 19:11:45 +00:00
try {
$sk->outputPage();
} catch ( Exception $e ) {
ob_end_clean(); // bug T129657
throw $e;
}
implement "action=render"
2005-07-03 04:00:33 +00:00
}
remove wfdebug in pagehistory set ETag header on article views
2005-07-01 00:03:31 +00:00
Avoid double-rendering on late exceptions Bug: T129657 Change-Id: I86af834b842bc056b57b2a55a9e1385481d5781d
2016-03-18 19:11:45 +00:00
try {
// This hook allows last minute changes to final overall output by modifying output buffer
Hooks::run() call site migration Migrate all callers of Hooks::run() to use the new HookContainer/HookRunner system. General principles: * Use DI if it is already used. We're not changing the way state is managed in this patch. * HookContainer is always injected, not HookRunner. HookContainer is a service, it's a more generic interface, it is the only thing that provides isRegistered() which is needed in some cases, and a HookRunner can be efficiently constructed from it (confirmed by benchmark). Because HookContainer is needed for object construction, it is also needed by all factories. * "Ask your friendly local base class". Big hierarchies like SpecialPage and ApiBase have getHookContainer() and getHookRunner() methods in the base class, and classes that extend that base class are not expected to know or care where the base class gets its HookContainer from. * ProtectedHookAccessorTrait provides protected getHookContainer() and getHookRunner() methods, getting them from the global service container. The point of this is to ease migration to DI by ensuring that call sites ask their local friendly base class rather than getting a HookRunner from the service container directly. * Private $this->hookRunner. In some smaller classes where accessor methods did not seem warranted, there is a private HookRunner property which is accessed directly. Very rarely (two cases), there is a protected property, for consistency with code that conventionally assumes protected=private, but in cases where the class might actually be overridden, a protected accessor is preferred over a protected property. * The last resort: Hooks::runner(). Mostly for static, file-scope and global code. In a few cases it was used for objects with broken construction schemes, out of horror or laziness. Constructors with new required arguments: * AuthManager * BadFileLookup * BlockManager * ClassicInterwikiLookup * ContentHandlerFactory * ContentSecurityPolicy * DefaultOptionsManager * DerivedPageDataUpdater * FullSearchResultWidget * HtmlCacheUpdater * LanguageFactory * LanguageNameUtils * LinkRenderer * LinkRendererFactory * LocalisationCache * MagicWordFactory * MessageCache * NamespaceInfo * PageEditStash * PageHandlerFactory * PageUpdater * ParserFactory * PermissionManager * RevisionStore * RevisionStoreFactory * SearchEngineConfig * SearchEngineFactory * SearchFormWidget * SearchNearMatcher * SessionBackend * SpecialPageFactory * UserNameUtils * UserOptionsManager * WatchedItemQueryService * WatchedItemStore Constructors with new optional arguments: * DefaultPreferencesFactory * Language * LinkHolderArray * MovePage * Parser * ParserCache * PasswordReset * Router setHookContainer() now required after construction: * AuthenticationProvider * ResourceLoaderModule * SearchEngine Change-Id: Id442b0dbe43aba84bd5cf801d86dedc768b082c7
2020-03-19 02:42:09 +00:00
$this->getHookRunner()->onAfterFinalPageOutput( $this );
Avoid double-rendering on late exceptions Bug: T129657 Change-Id: I86af834b842bc056b57b2a55a9e1385481d5781d
2016-03-18 19:11:45 +00:00
} catch ( Exception $e ) {
ob_end_clean(); // bug T129657
throw $e;
}
New hook 'AfterFinalPageOutput' which is called at the end of OutputPage::output() Called before final ob_end_flush() which will send the buffered output to the client. This allows for last-minute modification of the output within the buffer by using ob_get_clean(). - patch-set2: moved the hook before OutputPage::sendCacheControl() is called, as suggested by Reedy. Change-Id: I264a37ac74e0c18ff141f82156e669ece3212c2d
2012-07-05 14:21:21 +00:00
(bug 26) Defer sending cache headers until the skin runs, so that the skin can declare the page as uncacheable
2004-11-12 09:34:11 +00:00
$this->sendCacheControl();
We need this 'AfterFinalPageOutput' hook for the STTLanguage extension. Change-Id: I1f131e87324cb7a62355dc296f0b4d411a9a0800
2012-07-04 21:23:56 +00:00
Support masking the WRITE_SYNC latency from ChronologyProtector * Use OutputPage::output() as the method to mask latency, since it takes a good while to run. By the time it runs, cache replication should have caught up, so the reap call will likely not block. * For redirects emitted after changes in POST, instead of masking with OutputPage, add a parameter to the redirect and block on the positions appearing. This uses the redirection RTT to mask the replication latency. Change-Id: Ib23690c302e8033610fef9a0ef451dafe8a5803e
2016-08-24 21:22:11 +00:00
if ( $return ) {
return ob_get_clean();
} else {
ob_end_flush();
return null;
}
Initial revision
2003-04-14 23:10:40 +00:00
}
* Factorise the code used to prepare the OutputPage object to display an error page in OutputPage::prepareErrorPage() * Changed default value of OutputPage::getPageTitle() to match the value of the 'internalerror' message
2011-10-28 15:45:54 +00:00
/**
* Prepare this object to display an error page; disable caching and
* indexing, clear the current text and redirect, set the page's title
escape tags and entity in doxygen comments When inserting XML elements inline <such as this one>, doxygen chokes about it not being known. Simply enclosing the tag in double quotes prevents doxygen from emitting a warning. Also enclosed a few invalid functions calls such as \. and double quoted the HTML entities such as &foobar; Change-Id: I4019637145e683c2bec3d17b2fd98b0c50a932f1
2012-07-10 12:48:06 +00:00
* and optionally an custom HTML title (content of the "<title>" tag).
* Factorise the code used to prepare the OutputPage object to display an error page in OutputPage::prepareErrorPage() * Changed default value of OutputPage::getPageTitle() to match the value of the 'internalerror' message
2011-10-28 15:45:54 +00:00
*
Cleanup some docs (includes/*.php) - Swap "$variable type" to "type $variable" - Added missing types - Fixed spacing inside docs - Makes beginning of @param/@return/@var/@throws in capital - Changed some types to match the more common spelling Change-Id: I783e4dbfe5f6f98b32b9a03ccf6439e13e132bcc
2014-07-24 17:42:24 +00:00
* @param string|Message $pageTitle Will be passed directly to setPageTitle()
docs: Avoid some scalar juggling Phan can treat scalar types as non-interchangeable with `scalar_implicit_cast` set to false. This patch fixes some of those issues (which are in total >1000), namely the ones with alphabetic order < includes/actions. Change-Id: Ib1c6573ab899088bc319b9da9ceaffc850da3dbe
2019-09-15 14:40:42 +00:00
* @param string|Message|false $htmlTitle Will be passed directly to setHTMLTitle();
escape tags and entity in doxygen comments When inserting XML elements inline <such as this one>, doxygen chokes about it not being known. Simply enclosing the tag in double quotes prevents doxygen from emitting a warning. Also enclosed a few invalid functions calls such as \. and double quoted the HTML entities such as &foobar; Change-Id: I4019637145e683c2bec3d17b2fd98b0c50a932f1
2012-07-10 12:48:06 +00:00
* optional, if not passed the "<title>" attribute will be
* Factorise the code used to prepare the OutputPage object to display an error page in OutputPage::prepareErrorPage() * Changed default value of OutputPage::getPageTitle() to match the value of the 'internalerror' message
2011-10-28 15:45:54 +00:00
* based on $pageTitle
*/
public function prepareErrorPage( $pageTitle, $htmlTitle = false ) {
$this->setPageTitle( $pageTitle );
if ( $htmlTitle !== false ) {
$this->setHTMLTitle( $htmlTitle );
}
$this->setRobotPolicy( 'noindex,nofollow' );
$this->setArticleRelated( false );
Add OutputPage::disableClientCache() and deprecate ::enableClientCache(false) As far as I can tell, outside of test cases, every single time we call OutputPage::enableClientCache() we pass it `false` and actually mean to *disable* the client cache. Create a new less-confusingly-named method and deprecate the old one. Change-Id: I7c89e20528a0d91173f0edcb997dcae631935ee5
2022-02-04 19:12:11 +00:00
$this->disableClientCache();
* Factorise the code used to prepare the OutputPage object to display an error page in OutputPage::prepareErrorPage() * Changed default value of OutputPage::getPageTitle() to match the value of the 'internalerror' message
2011-10-28 15:45:54 +00:00
$this->mRedirect = '';
Fix for r102417: commiting OutputPage.php is actually a good idea
2011-11-08 18:02:26 +00:00
$this->clearSubtitle();
* Factorise the code used to prepare the OutputPage object to display an error page in OutputPage::prepareErrorPage() * Changed default value of OutputPage::getPageTitle() to match the value of the 'internalerror' message
2011-10-28 15:45:54 +00:00
$this->clearHTML();
}
Changing comments layout preparing for generated documentation with Phpdocumentor
2004-09-02 23:28:24 +00:00
/**
* Support passing message arguments to OutputPage::showErrorPage() * (bug 8818) Expose "wpDestFile" as parameter $1 to "uploaddisabledtext"
2007-05-18 20:46:42 +00:00
* Output a standard error page
Add some documentation to OutputPage.php and specify public/private for most of its member functions (practically all public anyway). And replace one gratuitous use of strcmp with good old ===.
2006-11-07 05:37:31 +00:00
*
Improve OutputPage::showErrorPage method documentation - Added the most usual case of call - Fixed the incoherence between -Object and -Obj Change-Id: I81207f4ef46a43529974e3ada194f2948a1e9017
2014-01-29 15:30:40 +00:00
* showErrorPage( 'titlemsg', 'pagetextmsg' );
Clean up array() syntax in docs, part VII Last part Change-Id: I38f015e2122ef4fd2d2141718bd889794c29f06c
2016-09-27 03:15:08 +00:00
* showErrorPage( 'titlemsg', 'pagetextmsg', [ 'param1', 'param2' ] );
Followup r85994: eliminate code duplication introduced by this revision, instead make OutputPage::showErrorPage() accept Message objects
2011-08-25 10:13:30 +00:00
* showErrorPage( 'titlemsg', $messageObject );
Improve OutputPage::showErrorPage method documentation - Added the most usual case of call - Fixed the incoherence between -Object and -Obj Change-Id: I81207f4ef46a43529974e3ada194f2948a1e9017
2014-01-29 15:30:40 +00:00
* showErrorPage( $titleMessageObject, $messageObject );
Tidy up w/c Remove trailing whitespace, add some documentation
2011-08-31 14:47:08 +00:00
*
Removed unneeded spaces and colons in @param and friends Also swapped some "$var type" to "type $var" or added missing types before the $var. Changed some other types to match the more common spelling. Makes beginning of some text in captial. Change-Id: Ic36c8c7820a6c2d603f1138130670c6bf6a1ca59
2014-04-08 15:29:17 +00:00
* @param string|Message $title Message key (string) for page title, or a Message object
* @param string|Message $msg Message key (string) for page text, or a Message object
* @param array $params Message parameters; ignored if $msg is a Message object
Allow customizing OutputPage::showErrorPage() return URL Change-Id: Id35574600e0118f0b6b823bdc84c8c4783d72d12
2022-09-01 01:30:10 +00:00
* @param PageReference|LinkTarget|string|null $returnto Page to show a return link to;
* defaults to the 'returnto' URL parameter
* @param string|null $returntoquery Query string for the return to link;
* defaults to the 'returntoquery' URL parameter
Changing comments layout preparing for generated documentation with Phpdocumentor
2004-09-02 23:28:24 +00:00
*/
Allow customizing OutputPage::showErrorPage() return URL Change-Id: Id35574600e0118f0b6b823bdc84c8c4783d72d12
2022-09-01 01:30:10 +00:00
public function showErrorPage(
$title, $msg, $params = [], $returnto = null, $returntoquery = null
) {
else if -> elseif Also fixed some spacing while at it Added some braces for one line statements Change-Id: Iebfa2b17091509daabc76248121c7763e2295d20
2013-04-17 14:52:47 +00:00
if ( !$title instanceof Message ) {
Allowing Message objects to be passed as error page title. Changed OutputPage::showErrorPage() so that the title parameter can be either a string or a Message object, just like the message parameter. This allows titles that take parameters. Change-Id: Icb7f7b0db8599984774f438333b60d66956b6075 Signed-off-by: Tyler Anthony Romeo <tylerromeo@gmail.com>
2012-08-06 21:07:24 +00:00
$title = $this->msg( $title );
}
Bug 26069 <title> element is "Error" for all error pages added more meaningful heading Change-Id: If59e729df9aa9d559302d6ebe5d4ee11999808ea
2012-08-18 23:17:57 +00:00
$this->prepareErrorPage( $title );
* remove end of line whitespace * remove empty lines at end of file * remove "?>" where still present
2008-04-14 07:45:50 +00:00
fix some spacing Change-Id: Ie7bb35871cc99237f3a655f7db22ca1f0646df5e
2013-01-26 18:32:03 +00:00
if ( $msg instanceof Message ) {
Convert all array() syntax to [] Per wikitech-l consensus: https://lists.wikimedia.org/pipermail/wikitech-l/2016-February/084821.html Notes: * Disabled CallTimePassByReference due to false positives (T127163) Change-Id: I2c8ce713ce6600a0bb7bf67537c87044c7a45c4b
2016-02-17 09:09:32 +00:00
if ( $params !== [] ) {
Pass php-strict for OutputPage Change-Id: Ia5d03277fff543ca672dc8efc7eaaeca4888ac23
2014-05-11 01:54:15 +00:00
trigger_error( 'Argument ignored: $params. The message parameters argument '
. 'is discarded when the $msg argument is a Message object instead of '
. 'a string.', E_USER_NOTICE );
OutputPage::showErrorPage raises a notice if arguments are incoherent The method has the following signature: OutputPage::showErrorPage( $title, $msg, $params = array() ) $msg can be a string or a Message object. If it's a string, a Message object is built, with $params as parameters. If it's a Message object, $params is ignored. The core now triggers a notice in the case a call is made with $msg an instance of Message object, and a (non-empty array) $params argument is given. Change-Id: I227a416f088fc1acd6a04345ed0e24d06f967ecc
2014-01-29 15:31:03 +00:00
}
Use parseAsBlock() instead of parse() to show the error message in OutputPage::showErrorPage() When a message name is passed, addWikiMsgArray() is used, so it adds a <p> arround the result. Do the same when a Message object is passed to that method. Change-Id: Ie97ad9a921ae3fc6d327531b0ace64b6a08e1fa2
2012-10-06 15:28:09 +00:00
$this->addHTML( $msg->parseAsBlock() );
Followup r85994: eliminate code duplication introduced by this revision, instead make OutputPage::showErrorPage() accept Message objects
2011-08-25 10:13:30 +00:00
} else {
Fix stupid copypaste mistake in r95470
2011-08-25 10:56:46 +00:00
$this->addWikiMsgArray( $msg, $params );
Followup r85994: eliminate code duplication introduced by this revision, instead make OutputPage::showErrorPage() accept Message objects
2011-08-25 10:13:30 +00:00
}
* remove end of line whitespace * remove empty lines at end of file * remove "?>" where still present
2008-04-14 07:45:50 +00:00
Allow customizing OutputPage::showErrorPage() return URL Change-Id: Id35574600e0118f0b6b823bdc84c8c4783d72d12
2022-09-01 01:30:10 +00:00
$this->returnToMain( null, $returnto, $returntoquery );
Initial revision
2003-04-14 23:10:40 +00:00
}
* Tweaking Werdnas new functions 'showPermissionsErrorPage' and 'formatPermissionsErrorMessage' a bit: ** Removing parameter $title from 'showPermissionsErrorPage', seems unused ** The message 'permissionserrorstext' uses {{PLURAL:}}, so pass the number of errors to the message. ** Show message 'permissionserrorstext' only once. ** Documentation * First use of 'showPermissionsErrorPage' in SpecialMovepage.php * Formatting MessagesEn.php a bit
2007-08-06 11:30:40 +00:00
/**
* Output a standard permission error page
*
Relax phpdoc of PermissionError to match actual usage Also fix OutputPage::showPermissionsErrorPage to handle everything it claims to handle. Change-Id: I8ed4a6f2f038fe70084ab673300e24af97e935db
2016-10-14 22:59:11 +00:00
* @param array $errors Error message keys or [key, param...] arrays
Fix PhanTypeMismatchDeclaredParam Auto fix MediaWiki.Commenting.FunctionComment.DefaultNullTypeParam sniff Change-Id: I865323fd0295aabd06f3e3c75e0e5043fb31069e
2018-06-26 21:14:43 +00:00
* @param string|null $action Action that was denied or null if unknown
* Tweaking Werdnas new functions 'showPermissionsErrorPage' and 'formatPermissionsErrorMessage' a bit: ** Removing parameter $title from 'showPermissionsErrorPage', seems unused ** The message 'permissionserrorstext' uses {{PLURAL:}}, so pass the number of errors to the message. ** Show message 'permissionserrorstext' only once. ** Documentation * First use of 'showPermissionsErrorPage' in SpecialMovepage.php * Formatting MessagesEn.php a bit
2007-08-06 11:30:40 +00:00
*/
Enforce array type hinting in OutputPage.php Change-Id: I2b37a95e517e120471c709823ea319f60165be0e
2014-07-23 02:49:30 +00:00
public function showPermissionsErrorPage( array $errors, $action = null ) {
Remove deprecated User groups/permission static calls. Bug: T220191 Change-Id: Ifa8afa90c432723b0bba0033a46b6a499c77e6fc
2019-08-22 02:22:26 +00:00
$services = MediaWikiServices::getInstance();
Replace deprecated PermissionManager::groupHasPermission Change-Id: I1e89c5361255c55f71ac2d4008a88475b4b04447
2022-11-13 20:20:37 +00:00
$groupPermissionsLookup = $services->getGroupPermissionsLookup();
Relax phpdoc of PermissionError to match actual usage Also fix OutputPage::showPermissionsErrorPage to handle everything it claims to handle. Change-Id: I8ed4a6f2f038fe70084ab673300e24af97e935db
2016-10-14 22:59:11 +00:00
foreach ( $errors as $key => $error ) {
$errors[$key] = (array)$error;
}
* Made OuputPage::showPermissionsErrorPage() show a different messages for 'read', 'edit', 'create' and 'upload' actions to saying "You need to log in to do this action" when 1) The user is not logged in 2) The only error is a permissions error (no block or something else) and 3) The error can simply be avoided by logging in * This replaces OuputPage::loginToUse() functionnality, made it simply throw a PermissionsEror exception and updated all calls in core * Same for the check in SpecialUpload::execute(), EditPage::userNotLoggedInPage() and EditPage::noCreatePermission() * Throw the same exception in EditPage::attemptSave() whether the user is logged in or not and let OuputPage::showPermissionsErrorPage() decide which message to display * Replaced call to deprecated OutputPage::blockedPage() in SpecialUpload * Displayed messages are the same as now, except the title is always "loginreqtitle" * 'nocreatetitle' and 'uploadnologin' messages are still used by extensions, so I kept them, but the message 'whitelistedittitle' is not used anymore and has been removed
2011-11-05 19:51:05 +00:00
// For some action (read, edit, create and upload), display a "login to do this action"
// error if all of the following conditions are met:
// 1. the user is not logged in
// 2. the only error is insufficient permissions (i.e. no block or something else)
// 3. the error can be avoided simply by logging in
Remove deprecated User groups/permission static calls. Bug: T220191 Change-Id: Ifa8afa90c432723b0bba0033a46b6a499c77e6fc
2019-08-22 02:22:26 +00:00
Convert all array() syntax to [] Per wikitech-l consensus: https://lists.wikimedia.org/pipermail/wikitech-l/2016-February/084821.html Notes: * Disabled CallTimePassByReference due to false positives (T127163) Change-Id: I2c8ce713ce6600a0bb7bf67537c87044c7a45c4b
2016-02-17 09:09:32 +00:00
if ( in_array( $action, [ 'read', 'edit', 'createpage', 'createtalk', 'upload' ] )
* Made OuputPage::showPermissionsErrorPage() show a different messages for 'read', 'edit', 'create' and 'upload' actions to saying "You need to log in to do this action" when 1) The user is not logged in 2) The only error is a permissions error (no block or something else) and 3) The error can simply be avoided by logging in * This replaces OuputPage::loginToUse() functionnality, made it simply throw a PermissionsEror exception and updated all calls in core * Same for the check in SpecialUpload::execute(), EditPage::userNotLoggedInPage() and EditPage::noCreatePermission() * Throw the same exception in EditPage::attemptSave() whether the user is logged in or not and let OuputPage::showPermissionsErrorPage() decide which message to display * Replaced call to deprecated OutputPage::blockedPage() in SpecialUpload * Displayed messages are the same as now, except the title is always "loginreqtitle" * 'nocreatetitle' and 'uploadnologin' messages are still used by extensions, so I kept them, but the message 'whitelistedittitle' is not used anymore and has been removed
2011-11-05 19:51:05 +00:00
&& $this->getUser()->isAnon() && count( $errors ) == 1 && isset( $errors[0][0] )
&& ( $errors[0][0] == 'badaccess-groups' || $errors[0][0] == 'badaccess-group0' )
Replace deprecated PermissionManager::groupHasPermission Change-Id: I1e89c5361255c55f71ac2d4008a88475b4b04447
2022-11-13 20:20:37 +00:00
&& ( $groupPermissionsLookup->groupHasPermission( 'user', $action )
|| $groupPermissionsLookup->groupHasPermission( 'autoconfirmed', $action ) )
* Made OuputPage::showPermissionsErrorPage() show a different messages for 'read', 'edit', 'create' and 'upload' actions to saying "You need to log in to do this action" when 1) The user is not logged in 2) The only error is a permissions error (no block or something else) and 3) The error can simply be avoided by logging in * This replaces OuputPage::loginToUse() functionnality, made it simply throw a PermissionsEror exception and updated all calls in core * Same for the check in SpecialUpload::execute(), EditPage::userNotLoggedInPage() and EditPage::noCreatePermission() * Throw the same exception in EditPage::attemptSave() whether the user is logged in or not and let OuputPage::showPermissionsErrorPage() decide which message to display * Replaced call to deprecated OutputPage::blockedPage() in SpecialUpload * Displayed messages are the same as now, except the title is always "loginreqtitle" * 'nocreatetitle' and 'uploadnologin' messages are still used by extensions, so I kept them, but the message 'whitelistedittitle' is not used anymore and has been removed
2011-11-05 19:51:05 +00:00
) {
$displayReturnto = null;
* (bug 32276) Skins were generating output using the internal page title which would allow anonymous users to determine wheter a page exists, potentially leaking private data. In fact, the curid and oldid request parameters would allow page titles to be enumerated even when they are not guessable. * (bug 32616) action=ajax requests were dispatched to the relevant internal functions without any read permission checks being done. This could lead to data leakage on private wikis.
2011-11-28 23:18:55 +00:00
includes: Replace implicit Bugzilla bug numbers with Phab ones It's unreasonable to expect newbies to know that "bug 12345" means "Task T14345" except where it doesn't, so let's just standardise on the real numbers. Change-Id: I6f59febaf8fc96e80f8cfc11f4356283f461142a
2017-02-20 22:44:19 +00:00
# Due to T34276, if a user does not have read permissions,
Moar documentations
2011-11-29 21:04:20 +00:00
# $this->getTitle() will just give Special:Badtitle, which is
# not especially useful as a returnto parameter. Use the title
* (bug 32276) Skins were generating output using the internal page title which would allow anonymous users to determine wheter a page exists, potentially leaking private data. In fact, the curid and oldid request parameters would allow page titles to be enumerated even when they are not guessable. * (bug 32616) action=ajax requests were dispatched to the relevant internal functions without any read permission checks being done. This could lead to data leakage on private wikis.
2011-11-28 23:18:55 +00:00
# from the request instead, if there was one.
$request = $this->getRequest();
Change trivial use of getVal('action') to getRawVal Per docs added in I18767cd809f67b, these don't need normalization as they are only compared against predefined strings, and besides are generally entered manually in a form, and even then would not require the kinds of Unicode chars that have multiple/non-normalized forms. In nearby areas to also fix some trivial cases: * getVal('title') obviously needs normalization. Use getText() to make this more obvious. * getVal() compared against simple string literals within the code obviously don't need normalization (e.g. printable === 'no'). * Change hot code in MediaWiki checking for whether 'diff' or 'oldid' are set to getCheck (which uses getRawVal) instead of getVal. As a bonus this means it now handles values like "0" correctly, which could theoretically have caused bad behaviour before. Change-Id: Ied721cfdf59c7ba11d1afa6f4cc59ede1381238e
2021-08-10 21:59:39 +00:00
$returnto = Title::newFromText( $request->getText( 'title' ) );
* Made OuputPage::showPermissionsErrorPage() show a different messages for 'read', 'edit', 'create' and 'upload' actions to saying "You need to log in to do this action" when 1) The user is not logged in 2) The only error is a permissions error (no block or something else) and 3) The error can simply be avoided by logging in * This replaces OuputPage::loginToUse() functionnality, made it simply throw a PermissionsEror exception and updated all calls in core * Same for the check in SpecialUpload::execute(), EditPage::userNotLoggedInPage() and EditPage::noCreatePermission() * Throw the same exception in EditPage::attemptSave() whether the user is logged in or not and let OuputPage::showPermissionsErrorPage() decide which message to display * Replaced call to deprecated OutputPage::blockedPage() in SpecialUpload * Displayed messages are the same as now, except the title is always "loginreqtitle" * 'nocreatetitle' and 'uploadnologin' messages are still used by extensions, so I kept them, but the message 'whitelistedittitle' is not used anymore and has been removed
2011-11-05 19:51:05 +00:00
if ( $action == 'edit' ) {
$msg = 'whitelistedittext';
$displayReturnto = $returnto;
} elseif ( $action == 'createpage' || $action == 'createtalk' ) {
$msg = 'nocreatetext';
} elseif ( $action == 'upload' ) {
$msg = 'uploadnologintext';
} else { # Read
$msg = 'loginreqpagetext';
$displayReturnto = Title::newMainPage();
}
Convert all array() syntax to [] Per wikitech-l consensus: https://lists.wikimedia.org/pipermail/wikitech-l/2016-February/084821.html Notes: * Disabled CallTimePassByReference due to false positives (T127163) Change-Id: I2c8ce713ce6600a0bb7bf67537c87044c7a45c4b
2016-02-17 09:09:32 +00:00
$query = [];
* (bug 32276) Skins were generating output using the internal page title which would allow anonymous users to determine wheter a page exists, potentially leaking private data. In fact, the curid and oldid request parameters would allow page titles to be enumerated even when they are not guessable. * (bug 32616) action=ajax requests were dispatched to the relevant internal functions without any read permission checks being done. This could lead to data leakage on private wikis.
2011-11-28 23:18:55 +00:00
* Made OuputPage::showPermissionsErrorPage() show a different messages for 'read', 'edit', 'create' and 'upload' actions to saying "You need to log in to do this action" when 1) The user is not logged in 2) The only error is a permissions error (no block or something else) and 3) The error can simply be avoided by logging in * This replaces OuputPage::loginToUse() functionnality, made it simply throw a PermissionsEror exception and updated all calls in core * Same for the check in SpecialUpload::execute(), EditPage::userNotLoggedInPage() and EditPage::noCreatePermission() * Throw the same exception in EditPage::attemptSave() whether the user is logged in or not and let OuputPage::showPermissionsErrorPage() decide which message to display * Replaced call to deprecated OutputPage::blockedPage() in SpecialUpload * Displayed messages are the same as now, except the title is always "loginreqtitle" * 'nocreatetitle' and 'uploadnologin' messages are still used by extensions, so I kept them, but the message 'whitelistedittitle' is not used anymore and has been removed
2011-11-05 19:51:05 +00:00
if ( $returnto ) {
$query['returnto'] = $returnto->getPrefixedText();
* (bug 32276) Skins were generating output using the internal page title which would allow anonymous users to determine wheter a page exists, potentially leaking private data. In fact, the curid and oldid request parameters would allow page titles to be enumerated even when they are not guessable. * (bug 32616) action=ajax requests were dispatched to the relevant internal functions without any read permission checks being done. This could lead to data leakage on private wikis.
2011-11-28 23:18:55 +00:00
* Made OuputPage::showPermissionsErrorPage() show a different messages for 'read', 'edit', 'create' and 'upload' actions to saying "You need to log in to do this action" when 1) The user is not logged in 2) The only error is a permissions error (no block or something else) and 3) The error can simply be avoided by logging in * This replaces OuputPage::loginToUse() functionnality, made it simply throw a PermissionsEror exception and updated all calls in core * Same for the check in SpecialUpload::execute(), EditPage::userNotLoggedInPage() and EditPage::noCreatePermission() * Throw the same exception in EditPage::attemptSave() whether the user is logged in or not and let OuputPage::showPermissionsErrorPage() decide which message to display * Replaced call to deprecated OutputPage::blockedPage() in SpecialUpload * Displayed messages are the same as now, except the title is always "loginreqtitle" * 'nocreatetitle' and 'uploadnologin' messages are still used by extensions, so I kept them, but the message 'whitelistedittitle' is not used anymore and has been removed
2011-11-05 19:51:05 +00:00
if ( !$request->wasPosted() ) {
$returntoquery = $request->getValues();
unset( $returntoquery['title'] );
unset( $returntoquery['returnto'] );
unset( $returntoquery['returntoquery'] );
(bug 36537) Rename calls to wfArrayToCGI to wfArrayToCgi Done with this command: grep wfArrayToCGI * -R -l -Z | xargs -0 -l sed -i -e 's/wfArrayToCGI/wfArrayToCgi/g' && git checkout -- HISTORY Change-Id: If7f4d4bfc199289c11c43cf519c2415a1aad1c51
2013-01-28 18:04:20 +00:00
$query['returntoquery'] = wfArrayToCgi( $returntoquery );
* Made OuputPage::showPermissionsErrorPage() show a different messages for 'read', 'edit', 'create' and 'upload' actions to saying "You need to log in to do this action" when 1) The user is not logged in 2) The only error is a permissions error (no block or something else) and 3) The error can simply be avoided by logging in * This replaces OuputPage::loginToUse() functionnality, made it simply throw a PermissionsEror exception and updated all calls in core * Same for the check in SpecialUpload::execute(), EditPage::userNotLoggedInPage() and EditPage::noCreatePermission() * Throw the same exception in EditPage::attemptSave() whether the user is logged in or not and let OuputPage::showPermissionsErrorPage() decide which message to display * Replaced call to deprecated OutputPage::blockedPage() in SpecialUpload * Displayed messages are the same as now, except the title is always "loginreqtitle" * 'nocreatetitle' and 'uploadnologin' messages are still used by extensions, so I kept them, but the message 'whitelistedittitle' is not used anymore and has been removed
2011-11-05 19:51:05 +00:00
}
}
Migrate remaining usages of Title::userCan() to PermissionManager T208768 introduced the PermissionManager service that can now be used for page specific permission checks. This change replaces remaining calls to Title::userCan() with the new service in MediaWiki core. Bug: T220191 Change-Id: Ie45e0cb6aa49a8c66147b470946161fc18160fc1
2019-06-03 10:48:02 +00:00
Login-related messages should get the URL as a parameter System messages along the lines of "you can't view this page, please log in" get the login link (with returnto parameters) as a HTML string, but sometimes one might want to customize the link on one page without affecting all the other pages so it's nice to have the raw link available as well. Change-Id: Ib71791ea28718455d2925e4424147e87ef901f33
2018-12-07 00:18:47 +00:00
$title = SpecialPage::getTitleFor( 'Userlogin' );
Migrate remaining usages of Title::userCan() to PermissionManager T208768 introduced the PermissionManager service that can now be used for page specific permission checks. This change replaces remaining calls to Title::userCan() with the new service in MediaWiki core. Bug: T220191 Change-Id: Ie45e0cb6aa49a8c66147b470946161fc18160fc1
2019-06-03 10:48:02 +00:00
$linkRenderer = $services->getLinkRenderer();
Login-related messages should get the URL as a parameter System messages along the lines of "you can't view this page, please log in" get the login link (with returnto parameters) as a HTML string, but sometimes one might want to customize the link on one page without affecting all the other pages so it's nice to have the raw link available as well. Change-Id: Ib71791ea28718455d2925e4424147e87ef901f33
2018-12-07 00:18:47 +00:00
$loginUrl = $title->getLinkURL( $query, false, PROTO_RELATIVE );
Replaced all deprecated Linker methods with proper ones in core(1) Change-Id: Ie3a718dc1eae1507f8829fcf419c64c6846d2cb6
2017-01-18 20:55:20 +00:00
$loginLink = $linkRenderer->makeKnownLink(
Login-related messages should get the URL as a parameter System messages along the lines of "you can't view this page, please log in" get the login link (with returnto parameters) as a HTML string, but sometimes one might want to customize the link on one page without affecting all the other pages so it's nice to have the raw link available as well. Change-Id: Ib71791ea28718455d2925e4424147e87ef901f33
2018-12-07 00:18:47 +00:00
$title,
Replaced all deprecated Linker methods with proper ones in core(1) Change-Id: Ie3a718dc1eae1507f8829fcf419c64c6846d2cb6
2017-01-18 20:55:20 +00:00
$this->msg( 'loginreqlink' )->text(),
Convert all array() syntax to [] Per wikitech-l consensus: https://lists.wikimedia.org/pipermail/wikitech-l/2016-February/084821.html Notes: * Disabled CallTimePassByReference due to false positives (T127163) Change-Id: I2c8ce713ce6600a0bb7bf67537c87044c7a45c4b
2016-02-17 09:09:32 +00:00
[],
* Made OuputPage::showPermissionsErrorPage() show a different messages for 'read', 'edit', 'create' and 'upload' actions to saying "You need to log in to do this action" when 1) The user is not logged in 2) The only error is a permissions error (no block or something else) and 3) The error can simply be avoided by logging in * This replaces OuputPage::loginToUse() functionnality, made it simply throw a PermissionsEror exception and updated all calls in core * Same for the check in SpecialUpload::execute(), EditPage::userNotLoggedInPage() and EditPage::noCreatePermission() * Throw the same exception in EditPage::attemptSave() whether the user is logged in or not and let OuputPage::showPermissionsErrorPage() decide which message to display * Replaced call to deprecated OutputPage::blockedPage() in SpecialUpload * Displayed messages are the same as now, except the title is always "loginreqtitle" * 'nocreatetitle' and 'uploadnologin' messages are still used by extensions, so I kept them, but the message 'whitelistedittitle' is not used anymore and has been removed
2011-11-05 19:51:05 +00:00
$query
);
$this->prepareErrorPage( $this->msg( 'loginreqtitle' ) );
Login-related messages should get the URL as a parameter System messages along the lines of "you can't view this page, please log in" get the login link (with returnto parameters) as a HTML string, but sometimes one might want to customize the link on one page without affecting all the other pages so it's nice to have the raw link available as well. Change-Id: Ib71791ea28718455d2925e4424147e87ef901f33
2018-12-07 00:18:47 +00:00
$this->addHTML( $this->msg( $msg )->rawParams( $loginLink )->params( $loginUrl )->parse() );
* Factorise the code used to prepare the OutputPage object to display an error page in OutputPage::prepareErrorPage() * Changed default value of OutputPage::getPageTitle() to match the value of the 'internalerror' message
2011-10-28 15:45:54 +00:00
* Made OuputPage::showPermissionsErrorPage() show a different messages for 'read', 'edit', 'create' and 'upload' actions to saying "You need to log in to do this action" when 1) The user is not logged in 2) The only error is a permissions error (no block or something else) and 3) The error can simply be avoided by logging in * This replaces OuputPage::loginToUse() functionnality, made it simply throw a PermissionsEror exception and updated all calls in core * Same for the check in SpecialUpload::execute(), EditPage::userNotLoggedInPage() and EditPage::noCreatePermission() * Throw the same exception in EditPage::attemptSave() whether the user is logged in or not and let OuputPage::showPermissionsErrorPage() decide which message to display * Replaced call to deprecated OutputPage::blockedPage() in SpecialUpload * Displayed messages are the same as now, except the title is always "loginreqtitle" * 'nocreatetitle' and 'uploadnologin' messages are still used by extensions, so I kept them, but the message 'whitelistedittitle' is not used anymore and has been removed
2011-11-05 19:51:05 +00:00
# Don't return to a page the user can't read otherwise
# we'll end up in a pointless loop
OutputPage: move to authority where relevant Change-Id: Iae42cc747f80cb0215f862d633e9af32fc7c8571
2021-03-04 15:55:20 +00:00
if ( $displayReturnto && $this->getAuthority()->probablyCan( 'read', $displayReturnto ) ) {
* Made OuputPage::showPermissionsErrorPage() show a different messages for 'read', 'edit', 'create' and 'upload' actions to saying "You need to log in to do this action" when 1) The user is not logged in 2) The only error is a permissions error (no block or something else) and 3) The error can simply be avoided by logging in * This replaces OuputPage::loginToUse() functionnality, made it simply throw a PermissionsEror exception and updated all calls in core * Same for the check in SpecialUpload::execute(), EditPage::userNotLoggedInPage() and EditPage::noCreatePermission() * Throw the same exception in EditPage::attemptSave() whether the user is logged in or not and let OuputPage::showPermissionsErrorPage() decide which message to display * Replaced call to deprecated OutputPage::blockedPage() in SpecialUpload * Displayed messages are the same as now, except the title is always "loginreqtitle" * 'nocreatetitle' and 'uploadnologin' messages are still used by extensions, so I kept them, but the message 'whitelistedittitle' is not used anymore and has been removed
2011-11-05 19:51:05 +00:00
$this->returnToMain( null, $displayReturnto );
}
} else {
$this->prepareErrorPage( $this->msg( 'permissionserrors' ) );
Use OutputPage::addWikiTextAsInterface() instead of untidy addWikiText() This change ensures that the output is tidy, and is necessary to support future parsers which will not be able to produce untidy output. Bug: T198214 Change-Id: I743f4185a03403f8d9b9db010ff1ee4e9342e062
2018-09-25 15:02:07 +00:00
$this->addWikiTextAsInterface( $this->formatPermissionsErrorMessage( $errors, $action ) );
* Made OuputPage::showPermissionsErrorPage() show a different messages for 'read', 'edit', 'create' and 'upload' actions to saying "You need to log in to do this action" when 1) The user is not logged in 2) The only error is a permissions error (no block or something else) and 3) The error can simply be avoided by logging in * This replaces OuputPage::loginToUse() functionnality, made it simply throw a PermissionsEror exception and updated all calls in core * Same for the check in SpecialUpload::execute(), EditPage::userNotLoggedInPage() and EditPage::noCreatePermission() * Throw the same exception in EditPage::attemptSave() whether the user is logged in or not and let OuputPage::showPermissionsErrorPage() decide which message to display * Replaced call to deprecated OutputPage::blockedPage() in SpecialUpload * Displayed messages are the same as now, except the title is always "loginreqtitle" * 'nocreatetitle' and 'uploadnologin' messages are still used by extensions, so I kept them, but the message 'whitelistedittitle' is not used anymore and has been removed
2011-11-05 19:51:05 +00:00
}
userCan changes, which have been done for about 3 weeks. Still TODO includes making a version of OutputPage::showPermissionsErrorPage that returns, rather than prints the output (for editing pages and such), and integrating the new functionality into the software. However, I haven't had the time to do this.
2007-08-01 10:19:26 +00:00
}
* Added a versionRequired() function to OutputPage, useful for extension writers that want to control what version of MediaWiki their extension can be used with
2005-06-26 06:49:56 +00:00
/**
* Display an error page indicating that a given version of MediaWiki is
* required to use it
*
Removed unneeded spaces and colons in @param and friends Also swapped some "$var type" to "type $var" or added missing types before the $var. Changed some other types to match the more common spelling. Makes beginning of some text in captial. Change-Id: Ic36c8c7820a6c2d603f1138130670c6bf6a1ca59
2014-04-08 15:29:17 +00:00
* @param mixed $version The version of MediaWiki needed to use the page
* Added a versionRequired() function to OutputPage, useful for extension writers that want to control what version of MediaWiki their extension can be used with
2005-06-26 06:49:56 +00:00
*/
Add some documentation to OutputPage.php and specify public/private for most of its member functions (practically all public anyway). And replace one gratuitous use of strcmp with good old ===.
2006-11-07 05:37:31 +00:00
public function versionRequired( $version ) {
* Factorise the code used to prepare the OutputPage object to display an error page in OutputPage::prepareErrorPage() * Changed default value of OutputPage::getPageTitle() to match the value of the 'internalerror' message
2011-10-28 15:45:54 +00:00
$this->prepareErrorPage( $this->msg( 'versionrequired', $version ) );
* Added a versionRequired() function to OutputPage, useful for extension writers that want to control what version of MediaWiki their extension can be used with
2005-06-26 06:49:56 +00:00
* Introduced OutputPage::addWikiMsg() and OutputPage::wrapWikiMsg(), to make it easier to avoid double-parsing. Double-parsing introduces differences between the old and new parsers. * Used the new functions in the relevant places throughout the core. Where this was not possible, used wfMsgNoTrans() to avoid double-parsing. * Calling wfEscapeWikiText() on a title before passing it to a message is unnecessary and causes bugs. The best idea is to let the message do its own escaping. Removed wfEscapeWikiText() from the caller of addedwatchtext, removedwatchtext, deletedtext, protect-text, nosuchusershort and movepage-moved. Updated all message files to use <nowiki>$1</nowiki> where appropriate in these messages. * Removed unnecessary empty message handling from semiprotectedpagewarning * Removed nonsense warning about the output of wfMsg() not being safe for inclusion in HTML. * Changed wfShellExec() to use passthru/ob_get_contents instead of exec(). This avoids corruption of whitespace in the response. * Added $wgValidateAllHtml feature, to optionally validate HTML on output. * Fixed missing newline in cascadeprotected in MessagesEn.php * Belatedly updated addedwatchtext to take account of the new "monobook" skin. No unwatch link in the sidebar anymore.
2008-02-18 07:25:35 +00:00
$this->addWikiMsg( 'versionrequiredtext', $version );
* Added a versionRequired() function to OutputPage, useful for extension writers that want to control what version of MediaWiki their extension can be used with
2005-06-26 06:49:56 +00:00
$this->returnToMain();
}
remove wfdebug in pagehistory set ETag header on article views
2005-07-01 00:03:31 +00:00
Convert second batch of special pages to authority Depends-On: Ifc82d29a00d3fd136c44e0699e6bbfa11b8cf2a6 Change-Id: Iaf559a72cc47ec5d2481d7f740ad0eb303c94002
2021-03-04 20:16:50 +00:00
/**
* Format permission $status obtained from Authority for display.
*
* @param PermissionStatus $status
* @param string|null $action that was denied or null if unknown
* @return string
*/
public function formatPermissionStatus( PermissionStatus $status, string $action = null ): string {
if ( $status->isGood() ) {
return '';
}
return $this->formatPermissionsErrorMessage( $status->toLegacyErrorArray(), $action );
}
Use the new userCan changes to display better, clearer error messages when a permissions error is encountered.
2007-08-03 09:27:28 +00:00
/**
Step two in fixing OutputPage's documentation, more or less the same as r61647
2010-01-29 21:32:45 +00:00
* Format a list of error messages
*
Convert second batch of special pages to authority Depends-On: Ifc82d29a00d3fd136c44e0699e6bbfa11b8cf2a6 Change-Id: Iaf559a72cc47ec5d2481d7f740ad0eb303c94002
2021-03-04 20:16:50 +00:00
* @deprecated since 1.36. Use ::formatPermissionStatus instead
Hard deprecate Title::getUserPermissionsErrors Remove final uses as well Bug: T244929 Change-Id: I65e937c7b9904b1e93f649508b14148849589f82
2020-02-26 04:26:00 +00:00
* @param array $errors Array of arrays returned by PermissionManager::getPermissionErrors
Fixes for the phan upgrade, part 1 Mainly, document some parameters as non-empty-array so that phan knows the list of arguments won't be empty when unpacking. In EditPage, account for hooks potentially unsetting the copyright notice. Also rewrite some code in LogPager, so it's hopefully easier for phan to understand what's going on. Change-Id: Ic0638571554424098d0743db32dd46723a08e103
2022-10-08 12:15:06 +00:00
* @phan-param non-empty-array[] $errors
Fix PhanTypeMismatchDeclaredParam Auto fix MediaWiki.Commenting.FunctionComment.DefaultNullTypeParam sniff Change-Id: I865323fd0295aabd06f3e3c75e0e5043fb31069e
2018-06-26 21:14:43 +00:00
* @param string|null $action Action that was denied or null if unknown
Removed unneeded spaces and colons in @param and friends Also swapped some "$var type" to "type $var" or added missing types before the $var. Changed some other types to match the more common spelling. Makes beginning of some text in captial. Change-Id: Ic36c8c7820a6c2d603f1138130670c6bf6a1ca59
2014-04-08 15:29:17 +00:00
* @return string The wikitext error-messages, formatted into a list.
Use the new userCan changes to display better, clearer error messages when a permissions error is encountered.
2007-08-03 09:27:28 +00:00
*/
Enforce array type hinting in OutputPage.php Change-Id: I2b37a95e517e120471c709823ea319f60165be0e
2014-07-23 02:49:30 +00:00
public function formatPermissionsErrorMessage( array $errors, $action = null ) {
*coding style cleanup (added braces, spacing tweaks etc.) *removed some unused globals
2010-05-22 12:18:22 +00:00
if ( $action == null ) {
* Use local context to get messages * Simply use getLang() instead of getContext()->getLang() now that this method exists * Also changed addWikiMsgArray() to use a Message object with addWikiText() for the parse operation; removed its $options parameter, no caller was using it
2011-10-23 08:13:52 +00:00
$text = $this->msg( 'permissionserrorstext', count( $errors ) )->plain() . "\n\n";
Core changes for GlobalBlocking and TorBlock extensions, plus some core refactoring work: * Instead of saying 'do that' in a permissions error, actually list what the action is (drawn from the right-$1 messages). This isn't perfect - it says you don't have permission to edit pages when you can't edit a single page, but it's better than 'do that'. * Refactor out some code from various block files into Block::formatExpiry and Block::parseExpiryInput. * Don't display 'you cannot edit special pages' when you're trying to execute, or create an account, or something like that. * New AbortAutoblock hook (for use in TorBlock extension), which allows extensions to cancel autoblocks.
2008-05-23 10:34:11 +00:00
} else {
* Use local context to get messages * Simply use getLang() instead of getContext()->getLang() now that this method exists * Also changed addWikiMsgArray() to use a Message object with addWikiText() for the parse operation; removed its $options parameter, no caller was using it
2011-10-23 08:13:52 +00:00
$action_desc = $this->msg( "action-$action" )->plain();
$text = $this->msg(
*coding style cleanup (added braces, spacing tweaks etc.) *removed some unused globals
2010-05-22 12:18:22 +00:00
'permissionserrorstext-withaction',
count( $errors ),
$action_desc
* Use local context to get messages * Simply use getLang() instead of getContext()->getLang() now that this method exists * Also changed addWikiMsgArray() to use a Message object with addWikiText() for the parse operation; removed its $options parameter, no caller was using it
2011-10-23 08:13:52 +00:00
)->plain() . "\n\n";
Core changes for GlobalBlocking and TorBlock extensions, plus some core refactoring work: * Instead of saying 'do that' in a permissions error, actually list what the action is (drawn from the right-$1 messages). This isn't perfect - it says you don't have permission to edit pages when you can't edit a single page, but it's better than 'do that'. * Refactor out some code from various block files into Block::formatExpiry and Block::parseExpiryInput. * Don't display 'you cannot edit special pages' when you're trying to execute, or create an account, or something like that. * New AbortAutoblock hook (for use in TorBlock extension), which allows extensions to cancel autoblocks.
2008-05-23 10:34:11 +00:00
}
Revert incorrectly committed file, again.
2008-01-02 01:16:44 +00:00
*coding style cleanup (added braces, spacing tweaks etc.) *removed some unused globals
2010-05-22 12:18:22 +00:00
if ( count( $errors ) > 1 ) {
* Fix bug with Protected namespace message when protected namespace is NS_MAIN. * Clean up formatPermissionsErrorMessage .. no point building a list if there's only one item, and it casues lots of grossness on complicated custom messages (removing linebreaks, screwing up formatting, etc.).
2007-08-16 07:05:38 +00:00
$text .= '<ul class="permissions-errors">' . "\n";
else if -> elseif Also fixed some spacing while at it Added some braces for one line statements Change-Id: Iebfa2b17091509daabc76248121c7763e2295d20
2013-04-17 14:52:47 +00:00
foreach ( $errors as $error ) {
* Fix bug with Protected namespace message when protected namespace is NS_MAIN. * Clean up formatPermissionsErrorMessage .. no point building a list if there's only one item, and it casues lots of grossness on complicated custom messages (removing linebreaks, screwing up formatting, etc.).
2007-08-16 07:05:38 +00:00
$text .= '<li>';
Replace call_user_func_array(), part 2 Uses new PHP 5.6 syntax like ...parameter unpacking and calling anything looking like a callback to make the code more readable. There are much more occurrences but this commit is intentionally limited to an easily reviewable size. In one occurrence, a simple conditional instead of trickery was much more readable. This patch finishes all the easy stuf in the core, the remainder is either unobvious or would result in smaller readability gains. It will be carefully dealt with in further commits. Change-Id: I79a16c48bfb98b75e5b99f2f6f4fa07b3ae02c5b
2018-06-08 02:58:35 +00:00
$text .= $this->msg( ...$error )->plain();
* Fix bug with Protected namespace message when protected namespace is NS_MAIN. * Clean up formatPermissionsErrorMessage .. no point building a list if there's only one item, and it casues lots of grossness on complicated custom messages (removing linebreaks, screwing up formatting, etc.).
2007-08-16 07:05:38 +00:00
$text .= "</li>\n";
}
$text .= '</ul>';
} else {
*coding style cleanup (added braces, spacing tweaks etc.) *removed some unused globals
2010-05-22 12:18:22 +00:00
$text .= "<div class=\"permissions-errors\">\n" .
build: Update mediawiki/mediawiki-phan-config to 0.12.0 This patch only adds and removes suppressions, which must be done in the same patch as the version bump. Bug: T298571 Change-Id: I4044d4d9ce82b3dae7ba0af85bf04f22cb1dd347
2022-10-08 12:12:34 +00:00
// @phan-suppress-next-line PhanParamTooFewUnpack Elements of $errors already annotated as non-empty
Replace call_user_func_array(), part 2 Uses new PHP 5.6 syntax like ...parameter unpacking and calling anything looking like a callback to make the code more readable. There are much more occurrences but this commit is intentionally limited to an easily reviewable size. In one occurrence, a simple conditional instead of trickery was much more readable. This patch finishes all the easy stuf in the core, the remainder is either unobvious or would result in smaller readability gains. It will be carefully dealt with in further commits. Change-Id: I79a16c48bfb98b75e5b99f2f6f4fa07b3ae02c5b
2018-06-08 02:58:35 +00:00
$this->msg( ...reset( $errors ) )->plain() .
*coding style cleanup (added braces, spacing tweaks etc.) *removed some unused globals
2010-05-22 12:18:22 +00:00
"\n</div>";
Use the new userCan changes to display better, clearer error messages when a permissions error is encountered.
2007-08-03 09:27:28 +00:00
}
return $text;
}
Moved rateLimited() and showLagWarning() near other similar functions
2011-04-07 16:04:08 +00:00
/**
Change "slave" => "replica DB" in /includes Change-Id: Icb716219c9335ff8fa447b1733d04b71d9712bf9
2016-09-05 20:21:26 +00:00
* Show a warning about replica DB lag
Moved rateLimited() and showLagWarning() near other similar functions
2011-04-07 16:04:08 +00:00
*
Rename $wgSlaveLagWarning and $wgSlaveLagCritical to match DB_REPLICA These are now named $wgDatabaseReplicaLagWarning and $wgDatabaseReplicaLagCritical, respectively. Bug: T277987 Change-Id: I62d8c2028fdfee00904f1058f8387499f2b2e977
2021-03-20 10:48:02 +00:00
* If the lag is higher than $wgDatabaseReplicaLagCritical seconds,
Moved rateLimited() and showLagWarning() near other similar functions
2011-04-07 16:04:08 +00:00
* then the warning is a bit more obvious. If the lag is
Rename $wgSlaveLagWarning and $wgSlaveLagCritical to match DB_REPLICA These are now named $wgDatabaseReplicaLagWarning and $wgDatabaseReplicaLagCritical, respectively. Bug: T277987 Change-Id: I62d8c2028fdfee00904f1058f8387499f2b2e977
2021-03-20 10:48:02 +00:00
* lower than $wgDatabaseReplicaLagWarning, then no warning is shown.
Moved rateLimited() and showLagWarning() near other similar functions
2011-04-07 16:04:08 +00:00
*
doc: Modernise parameter names and documentation for 'replica' DBs Non-breaking change. Remaining uses are public interfaces (a constant, two globals, a config sub-parameter, SQL queries, storage function names), one i18n message key, and a whole lot of maintenance scripts with calls to the deprecated function wfWaitForSlaves(). Change-Id: I6ee5ca92ccf6a80c08f53d9efe38ebb4b05064d7
2018-10-31 17:36:48 +00:00
* @param int $lag Replica lag
Moved rateLimited() and showLagWarning() near other similar functions
2011-04-07 16:04:08 +00:00
*/
public function showLagWarning( $lag ) {
Use config instead of globals for OutputPage Change-Id: I5e0ebc173631d1d1052de7ccee4ef839c7c1767f
2014-08-23 08:52:41 +00:00
$config = $this->getConfig();
Use name constants to access config settings. Use name constants instead of string literals in calls to Config::get and ServiceOptions::get, when referring to core configuration variables. This protects against typos and makes the decumentation and schema declaration of the config settings discoverable. This is the first batch, only touching files directly under /includes/ Change-Id: I7252e636c7c86d950d9257b33491af492c6dd5eb
2022-04-01 15:58:32 +00:00
if ( $lag >= $config->get( MainConfigNames::DatabaseReplicaLagWarning ) ) {
Floor slave lag before display Seeing slave lags like 15.309758901596 looks scary Change-Id: Ie564670448311364b8492af6b82943312118dc29
2016-08-21 15:01:04 +00:00
$lag = floor( $lag ); // floor to avoid nano seconds to display
Use name constants to access config settings. Use name constants instead of string literals in calls to Config::get and ServiceOptions::get, when referring to core configuration variables. This protects against typos and makes the decumentation and schema declaration of the config settings discoverable. This is the first batch, only touching files directly under /includes/ Change-Id: I7252e636c7c86d950d9257b33491af492c6dd5eb
2022-04-01 15:58:32 +00:00
$message = $lag < $config->get( MainConfigNames::DatabaseReplicaLagCritical )
Moved rateLimited() and showLagWarning() near other similar functions
2011-04-07 16:04:08 +00:00
? 'lag-warn-normal'
: 'lag-warn-high';
OutputPage: Make the DB replica lag warning CSS classes greppable mw-lag-warn-normal and mw-lag-warn-high. Change-Id: I92c5cba64059a42d1b325858debf10e0fc33dc85
2019-12-14 20:07:58 +00:00
// For grep: mw-lag-warn-normal, mw-lag-warn-high
Convert all array() syntax to [] Per wikitech-l consensus: https://lists.wikimedia.org/pipermail/wikitech-l/2016-February/084821.html Notes: * Disabled CallTimePassByReference due to false positives (T127163) Change-Id: I2c8ce713ce6600a0bb7bf67537c87044c7a45c4b
2016-02-17 09:09:32 +00:00
$wrap = Html::rawElement( 'div', [ 'class' => "mw-{$message}" ], "\n$1\n" );
$this->wrapWikiMsg( "$wrap\n", [ $message, $this->getLanguage()->formatNum( $lag ) ] );
Moved rateLimited() and showLagWarning() near other similar functions
2011-04-07 16:04:08 +00:00
}
}
Deprecate specialized file errors in OutputPage and fix escaping OutputPage has a number of specialized error reporting methods related to file handling. With exception of showFileDeleteError, they are all unused. In my opinion such specialized error handling does not belong in OutputPage, but in whatever class is generating the error. Futhermore, these functions do not appropriately escape their arguments or their i18n messages. I replaced the one usage in SpecialUpload with an equivalent that does escape properly. This is not exploitable as the attacker is not in control of the temporary file name, but it is very bad practice. This deprecates the following methods: * OutputPage::showFileDeleteError() * OutputPage::showFileNotFoundError() * OutputPage::showFileRenameError() * OutputPage::showFileCopyError() * OutputPage::showUnexpectedValueError() [Discovered with the help of an experimental phan plugin] Change-Id: I9e7aaa59ded66f32c78cfdfed1e59e073ffd5051
2017-10-16 05:53:30 +00:00
/**
* Output an error page
*
* @note FatalError exception class provides an alternative.
* @param string $message Error to output. Must be escaped for HTML.
*/
Add some documentation to OutputPage.php and specify public/private for most of its member functions (practically all public anyway). And replace one gratuitous use of strcmp with good old ===.
2006-11-07 05:37:31 +00:00
public function showFatalError( $message ) {
* Factorise the code used to prepare the OutputPage object to display an error page in OutputPage::prepareErrorPage() * Changed default value of OutputPage::getPageTitle() to match the value of the 'internalerror' message
2011-10-28 15:45:54 +00:00
$this->prepareErrorPage( $this->msg( 'internalerror' ) );
$this->addHTML( $message );
Initial revision
2003-04-14 23:10:40 +00:00
}
Changing comments layout preparing for generated documentation with Phpdocumentor
2004-09-02 23:28:24 +00:00
/**
Introduce OutputPage::addReturnTo(), with a view to replacing OutputPage::returnToMain() [cleaner calling convention], and document the latter
2007-07-29 23:53:45 +00:00
* Add a "return to" link pointing to a specified title
*
output: Narrow Title type hint to LinkTarget The code never needs a full Title object, but only what the (more narrow) interface provides. Change-Id: I3a7aa35c7e90fffaf37f4cb444549237c20e0bd9
2019-09-13 09:34:33 +00:00
* @param LinkTarget $title Title to link
Fixed some @params documentation Swapped some "$var type" to "type $var" or added missing types before the $var. Changed some other types to match the more common spelling. Makes beginning of some text in captial. Change-Id: I64e8cfe478cb0ba438f40b0631d6e9049cdab567
2014-04-10 18:50:10 +00:00
* @param array $query Query string parameters
Fix PhanTypeMismatchDeclaredParam Auto fix MediaWiki.Commenting.FunctionComment.DefaultNullTypeParam sniff Change-Id: I865323fd0295aabd06f3e3c75e0e5043fb31069e
2018-06-26 21:14:43 +00:00
* @param string|null $text Text of the link (input is not escaped)
Fixed some @params documentation Swapped some "$var type" to "type $var" or added missing types before the $var. Changed some other types to match the more common spelling. Makes beginning of some text in captial. Change-Id: I64e8cfe478cb0ba438f40b0631d6e9049cdab567
2014-04-10 18:50:10 +00:00
* @param array $options Options array to pass to Linker
Introduce OutputPage::addReturnTo(), with a view to replacing OutputPage::returnToMain() [cleaner calling convention], and document the latter
2007-07-29 23:53:45 +00:00
*/
Convert all array() syntax to [] Per wikitech-l consensus: https://lists.wikimedia.org/pipermail/wikitech-l/2016-February/084821.html Notes: * Disabled CallTimePassByReference due to false positives (T127163) Change-Id: I2c8ce713ce6600a0bb7bf67537c87044c7a45c4b
2016-02-17 09:09:32 +00:00
public function addReturnTo( $title, array $query = [], $text = null, $options = [] ) {
Replace Linker::link() with LinkRenderer in includes directory * CategoryViewer * OutputPage * Preferences Bug: T149346 Change-Id: I9f9bcd9b461884817e8ceefbc6757c436221e331
2016-11-30 21:28:55 +00:00
$linkRenderer = MediaWikiServices::getInstance()
->getLinkRendererFactory()->createFromLegacyOptions( $options );
* Use local context to get messages * Simply use getLang() instead of getContext()->getLang() now that this method exists * Also changed addWikiMsgArray() to use a Message object with addWikiText() for the parse operation; removed its $options parameter, no caller was using it
2011-10-23 08:13:52 +00:00
$link = $this->msg( 'returnto' )->rawParams(
Replace Linker::link() with LinkRenderer in includes directory * CategoryViewer * OutputPage * Preferences Bug: T149346 Change-Id: I9f9bcd9b461884817e8ceefbc6757c436221e331
2016-11-30 21:28:55 +00:00
$linkRenderer->makeLink( $title, $text, [], $query ) )->escaped();
Follow-up to r50822. Add id="mw-returnto" to allow styling.
2009-06-15 11:11:17 +00:00
$this->addHTML( "<p id=\"mw-returnto\">{$link}</p>\n" );
Introduce OutputPage::addReturnTo(), with a view to replacing OutputPage::returnToMain() [cleaner calling convention], and document the latter
2007-07-29 23:53:45 +00:00
}
/**
* Add a "return to" link pointing to a specified title,
* or the title indicated in the request, or else the main page
*
Fix PhanTypeMismatchDeclaredParam Auto fix MediaWiki.Commenting.FunctionComment.DefaultNullTypeParam sniff Change-Id: I865323fd0295aabd06f3e3c75e0e5043fb31069e
2018-06-26 21:14:43 +00:00
* @param mixed|null $unused
Remove Title from public interface of OutputPage Bug: T278459 Change-Id: If455682540faf8d42fd8a14e0ef09bac24f72db4
2021-04-08 19:11:06 +00:00
* @param PageReference|LinkTarget|string|null $returnto Page to return to
Fix PhanTypeMismatchDeclaredParam Auto fix MediaWiki.Commenting.FunctionComment.DefaultNullTypeParam sniff Change-Id: I865323fd0295aabd06f3e3c75e0e5043fb31069e
2018-06-26 21:14:43 +00:00
* @param string|null $returntoquery Query string for the return to link
Changing comments layout preparing for generated documentation with Phpdocumentor
2004-09-02 23:28:24 +00:00
*/
Add returntoquery= parameter to Special:Userlogin which adds a query string to the "Return to" link. Also using this parameter in the Log in and Log out links on top of the screen, so that people who click Log in from e.g. an edit page will be taken back to the edit page rather than to the page view.
2009-07-14 21:25:33 +00:00
public function returnToMain( $unused = null, $returnto = null, $returntoquery = null ) {
Make use of ??= in more places New feature from PHP 7.4 Change-Id: Ifa7a9bc7b2ec415ad7ecb23f4c1776f51f58fd6b
2022-12-16 23:48:27 +00:00
$returnto ??= $this->getRequest()->getText( 'returnto' );
fixed bug in addClass append
2009-08-07 00:16:54 +00:00
Make use of ??= in more places New feature from PHP 7.4 Change-Id: Ifa7a9bc7b2ec415ad7ecb23f4c1776f51f58fd6b
2022-12-16 23:48:27 +00:00
$returntoquery ??= $this->getRequest()->getText( 'returntoquery' );
* remove end of line whitespace * remove empty lines at end of file * remove "?>" where still present
2008-04-14 07:45:50 +00:00
Attempt at normalistion of comparison styles - empty string on left and right hand side normalised to it being on the rhs Before this change, there were (? being regex 0 or 1) "" ===? 1 '' ===? 24 "" !==? 8 '' !==? 32 == "" 14 == '' 344 != "" 9 !== "" 4 != '' 151 !== '' 85 Rhs was the much more common, and the preferred style by many developers.. (Was a similar discussion in #mediawiki recently.. After that lolbugreport i think) Where there is a string (non empty) on the lhs, and variable/method call on the rhs still need normalising
2010-01-06 19:59:42 +00:00
if ( $returnto === '' ) {
Use Title::newMainPage() in various places instead of calling wfMsgForContent( 'mainpage' ) directly. Also add release notes for previous commit.
2006-12-03 00:22:14 +00:00
$returnto = Title::newMainPage();
Initial revision
2003-04-14 23:10:40 +00:00
}
Saner handling of title objects in OutputPage::returnToMain()
2006-06-21 16:28:04 +00:00
if ( is_object( $returnto ) ) {
Remove Title from public interface of OutputPage Bug: T278459 Change-Id: If455682540faf8d42fd8a14e0ef09bac24f72db4
2021-04-08 19:11:06 +00:00
$linkTarget = TitleValue::castPageToLinkTarget( $returnto );
Saner handling of title objects in OutputPage::returnToMain()
2006-06-21 16:28:04 +00:00
} else {
Remove Title from public interface of OutputPage Bug: T278459 Change-Id: If455682540faf8d42fd8a14e0ef09bac24f72db4
2021-04-08 19:11:06 +00:00
$linkTarget = Title::newFromText( $returnto );
Saner handling of title objects in OutputPage::returnToMain()
2006-06-21 16:28:04 +00:00
}
Remove Title from public interface of OutputPage Bug: T278459 Change-Id: If455682540faf8d42fd8a14e0ef09bac24f72db4
2021-04-08 19:11:06 +00:00
SECURITY: Do not directly redirect to interwikis, but use splash page Directly redirecting based on a url paramter might potentially be used in a phishing attack to confuse users. Bug: T109140 Bug: T122209 Change-Id: I6c604439320fa876719933cc7f3a3ff04fb1a6ad
2016-02-07 13:07:20 +00:00
// We don't want people to return to external interwiki. That
// might potentially be used as part of a phishing scheme
Remove Title from public interface of OutputPage Bug: T278459 Change-Id: If455682540faf8d42fd8a14e0ef09bac24f72db4
2021-04-08 19:11:06 +00:00
if ( !is_object( $linkTarget ) || $linkTarget->isExternal() ) {
$linkTarget = Title::newMainPage();
Check for bad title in returnToMain()
2006-06-23 06:53:05 +00:00
}
Saner handling of title objects in OutputPage::returnToMain()
2006-06-21 16:28:04 +00:00
Remove Title from public interface of OutputPage Bug: T278459 Change-Id: If455682540faf8d42fd8a14e0ef09bac24f72db4
2021-04-08 19:11:06 +00:00
$this->addReturnTo( $linkTarget, wfCgiToArray( $returntoquery ) );
Initial revision
2003-04-14 23:10:40 +00:00
}
resourceloader: Move queue formatting out of OutputPage HTML formatting of the queue was distributed over several OutputPage methods. Each method demanding a snippet of HTML by calling makeResourceLoaderLink() with a limited amount of information. As such, makeResourceLoaderLink() was unable to provide the client with the proper state information. Centralising it also allows it to better reduce duplication in HTML output and maintain a more accurate state. Problems fixed by centralising: 1. The 'user' module is special (due to per-user 'version' and 'user' params). It is manually requested via script-src. To avoid a separate (and wrong) request from something that requires it, we set state=loading directly. However, because the module is in the bottom, the old HTML formatter could only put state=loading in the bottom also. This sometimes caused a wrong request to be fired for modules=user if something in the top queue triggered a requirement for it. 2. Since a464d1d4 (T87871) we track states of page-style modules, with purpose of allowing dependencies on style modules without risking duplicate loading on pages where the styles are loaded already. This didn't work, because the state information about page-style modules is output near the stylesheet, which is after the script tag with mw.loader.load(). That runs first, and mw.loader would still make a duplicate request before it learns the state. Changes: * Document reasons for style/script tag order in getHeadHtml (per 09537e83). * Pass $type from getModuleStyles() to getAllowedModules(). This wasn't needed before since a duplicate check in makeResourceLoaderLink() verified the origin a second time. * Declare explicit position 'top' on 'user.options' and 'user.tokens' module. Previously, OutputPage hardcoded them in the top. The new formatter doesn't. * Remove getHeadScripts(). * Remove getInlineHeadScripts(). * Remove getExternalHeadScripts(). * Remove buildCssLinks(). * Remove getScriptsForBottomQueue(). * Change where Skin::setupSkinUserCss() is called. This methods lets the skin add modules to the queue. Previously it was called from buildCssLinks(), via headElement(), via prepareQuickTemplate(), via OutputPage::output(). It's now in OutputPage::output() directly (slightly earlier). This is needed because prepareQuickTemplate() calls bottomScripts() before headElement(). And bottomScript() would lazy-initialise the queue and lock it before setupSkinUserCss() is called from headElement(). This makes execution order more predictable instead of being dependent on the arbitrary order of data extraction in prepareQuickTemplate (which varies from one skin to another). * Compute isUserModulePreview() and isKnownEmpty() for the 'user' module early on so. This avoids wrongful loading and fixes problem 1. Effective changes in output: * mw.loader.state() is now before mw.loader.load(). This fixes problem 2. * mw.loader.state() now sets 'user.options' and 'user.tokens' to "loading". * mw.loader.state() now sets 'user' (as "loading" or "ready"). Fixes problem 1. * The <script async src> tag for 'startup' changed position (slightly). Previously it was after all inline scripts and stylesheets. It's still after all inline scripts and after most stylesheets, but before any user styles. Since the queue is now formatted outside OutputPage, it can't inject the meta-ResourceLoaderDynamicStyles tag and user-stylesheet hack in the middle of existing output. This shouldn't have any noticable impact. Bug: T87871 Change-Id: I605b8cd1e1fc009b4662a0edbc54d09dd65ee1df
2016-07-15 14:13:09 +00:00
private function getRlClientContext() {
if ( !$this->rlClientContext ) {
$query = ResourceLoader::makeLoaderQuery(
[], // modules; not relevant
$this->getLanguage()->getCode(),
$this->getSkin()->getSkinName(),
Swap out uses of User->isLoggedIn() with ->isRegistered() Bug: T270450 Change-Id: I90ff94a553ebae9a4ef97c25c333f067f19c327d
2020-12-17 23:10:11 +00:00
$this->getUser()->isRegistered() ? $this->getUser()->getName() : null,
resourceloader: Move queue formatting out of OutputPage HTML formatting of the queue was distributed over several OutputPage methods. Each method demanding a snippet of HTML by calling makeResourceLoaderLink() with a limited amount of information. As such, makeResourceLoaderLink() was unable to provide the client with the proper state information. Centralising it also allows it to better reduce duplication in HTML output and maintain a more accurate state. Problems fixed by centralising: 1. The 'user' module is special (due to per-user 'version' and 'user' params). It is manually requested via script-src. To avoid a separate (and wrong) request from something that requires it, we set state=loading directly. However, because the module is in the bottom, the old HTML formatter could only put state=loading in the bottom also. This sometimes caused a wrong request to be fired for modules=user if something in the top queue triggered a requirement for it. 2. Since a464d1d4 (T87871) we track states of page-style modules, with purpose of allowing dependencies on style modules without risking duplicate loading on pages where the styles are loaded already. This didn't work, because the state information about page-style modules is output near the stylesheet, which is after the script tag with mw.loader.load(). That runs first, and mw.loader would still make a duplicate request before it learns the state. Changes: * Document reasons for style/script tag order in getHeadHtml (per 09537e83). * Pass $type from getModuleStyles() to getAllowedModules(). This wasn't needed before since a duplicate check in makeResourceLoaderLink() verified the origin a second time. * Declare explicit position 'top' on 'user.options' and 'user.tokens' module. Previously, OutputPage hardcoded them in the top. The new formatter doesn't. * Remove getHeadScripts(). * Remove getInlineHeadScripts(). * Remove getExternalHeadScripts(). * Remove buildCssLinks(). * Remove getScriptsForBottomQueue(). * Change where Skin::setupSkinUserCss() is called. This methods lets the skin add modules to the queue. Previously it was called from buildCssLinks(), via headElement(), via prepareQuickTemplate(), via OutputPage::output(). It's now in OutputPage::output() directly (slightly earlier). This is needed because prepareQuickTemplate() calls bottomScripts() before headElement(). And bottomScript() would lazy-initialise the queue and lock it before setupSkinUserCss() is called from headElement(). This makes execution order more predictable instead of being dependent on the arbitrary order of data extraction in prepareQuickTemplate (which varies from one skin to another). * Compute isUserModulePreview() and isKnownEmpty() for the 'user' module early on so. This avoids wrongful loading and fixes problem 1. Effective changes in output: * mw.loader.state() is now before mw.loader.load(). This fixes problem 2. * mw.loader.state() now sets 'user.options' and 'user.tokens' to "loading". * mw.loader.state() now sets 'user' (as "loading" or "ready"). Fixes problem 1. * The <script async src> tag for 'startup' changed position (slightly). Previously it was after all inline scripts and stylesheets. It's still after all inline scripts and after most stylesheets, but before any user styles. Since the queue is now formatted outside OutputPage, it can't inject the meta-ResourceLoaderDynamicStyles tag and user-stylesheet hack in the middle of existing output. This shouldn't have any noticable impact. Bug: T87871 Change-Id: I605b8cd1e1fc009b4662a0edbc54d09dd65ee1df
2016-07-15 14:13:09 +00:00
null, // version; not relevant
ResourceLoader::inDebugMode(),
null, // only; not relevant
resourceloader: Remove obsolete logic for handheld=yes pageviews This feature was removed in 2013 (Ia8d79b4a, T49504), with further remnants removed in (1965df8df3, I7c24128f7b1). * Simplify the code of transformCssMedia() as it now only needs to support printable=yes. * Remove query parameter abstraction from test cases, use query array directly. * Remove internal use of the $handheld parameter. Function signature of makeLoaderQuery() left in-tact for compatibility with $extraData. This will be deprecated and removed in a follow-up commit. Bug: T49504 Bug: T32956 Change-Id: Id16ef3f4630538de1887ac2327eb317c6c6b1f9d
2021-10-09 01:41:15 +00:00
$this->isPrintable()
resourceloader: Move queue formatting out of OutputPage HTML formatting of the queue was distributed over several OutputPage methods. Each method demanding a snippet of HTML by calling makeResourceLoaderLink() with a limited amount of information. As such, makeResourceLoaderLink() was unable to provide the client with the proper state information. Centralising it also allows it to better reduce duplication in HTML output and maintain a more accurate state. Problems fixed by centralising: 1. The 'user' module is special (due to per-user 'version' and 'user' params). It is manually requested via script-src. To avoid a separate (and wrong) request from something that requires it, we set state=loading directly. However, because the module is in the bottom, the old HTML formatter could only put state=loading in the bottom also. This sometimes caused a wrong request to be fired for modules=user if something in the top queue triggered a requirement for it. 2. Since a464d1d4 (T87871) we track states of page-style modules, with purpose of allowing dependencies on style modules without risking duplicate loading on pages where the styles are loaded already. This didn't work, because the state information about page-style modules is output near the stylesheet, which is after the script tag with mw.loader.load(). That runs first, and mw.loader would still make a duplicate request before it learns the state. Changes: * Document reasons for style/script tag order in getHeadHtml (per 09537e83). * Pass $type from getModuleStyles() to getAllowedModules(). This wasn't needed before since a duplicate check in makeResourceLoaderLink() verified the origin a second time. * Declare explicit position 'top' on 'user.options' and 'user.tokens' module. Previously, OutputPage hardcoded them in the top. The new formatter doesn't. * Remove getHeadScripts(). * Remove getInlineHeadScripts(). * Remove getExternalHeadScripts(). * Remove buildCssLinks(). * Remove getScriptsForBottomQueue(). * Change where Skin::setupSkinUserCss() is called. This methods lets the skin add modules to the queue. Previously it was called from buildCssLinks(), via headElement(), via prepareQuickTemplate(), via OutputPage::output(). It's now in OutputPage::output() directly (slightly earlier). This is needed because prepareQuickTemplate() calls bottomScripts() before headElement(). And bottomScript() would lazy-initialise the queue and lock it before setupSkinUserCss() is called from headElement(). This makes execution order more predictable instead of being dependent on the arbitrary order of data extraction in prepareQuickTemplate (which varies from one skin to another). * Compute isUserModulePreview() and isKnownEmpty() for the 'user' module early on so. This avoids wrongful loading and fixes problem 1. Effective changes in output: * mw.loader.state() is now before mw.loader.load(). This fixes problem 2. * mw.loader.state() now sets 'user.options' and 'user.tokens' to "loading". * mw.loader.state() now sets 'user' (as "loading" or "ready"). Fixes problem 1. * The <script async src> tag for 'startup' changed position (slightly). Previously it was after all inline scripts and stylesheets. It's still after all inline scripts and after most stylesheets, but before any user styles. Since the queue is now formatted outside OutputPage, it can't inject the meta-ResourceLoaderDynamicStyles tag and user-stylesheet hack in the middle of existing output. This shouldn't have any noticable impact. Bug: T87871 Change-Id: I605b8cd1e1fc009b4662a0edbc54d09dd65ee1df
2016-07-15 14:13:09 +00:00
);
ResourceLoader namespace (attempt 2) Move ResourceLoader classes to their own namespace. Strip the "ResourceLoader" prefix from all except ResourceLoader itself. Move the tests by analogy. I used a namespace alias "RL" in some callers since RL\Module is less ambiguous at the call site than just "Module". I did not address DependencyStore which continues to have a non-standard location and namespace. Revert of a241d83e0a6dabedf. Bug: T308718 Change-Id: Id08a220e1d6085e2b33f3f6c9d0e3935a4204659
2022-05-06 09:09:56 +00:00
$this->rlClientContext = new RL\Context(
resourceloader: Move queue formatting out of OutputPage HTML formatting of the queue was distributed over several OutputPage methods. Each method demanding a snippet of HTML by calling makeResourceLoaderLink() with a limited amount of information. As such, makeResourceLoaderLink() was unable to provide the client with the proper state information. Centralising it also allows it to better reduce duplication in HTML output and maintain a more accurate state. Problems fixed by centralising: 1. The 'user' module is special (due to per-user 'version' and 'user' params). It is manually requested via script-src. To avoid a separate (and wrong) request from something that requires it, we set state=loading directly. However, because the module is in the bottom, the old HTML formatter could only put state=loading in the bottom also. This sometimes caused a wrong request to be fired for modules=user if something in the top queue triggered a requirement for it. 2. Since a464d1d4 (T87871) we track states of page-style modules, with purpose of allowing dependencies on style modules without risking duplicate loading on pages where the styles are loaded already. This didn't work, because the state information about page-style modules is output near the stylesheet, which is after the script tag with mw.loader.load(). That runs first, and mw.loader would still make a duplicate request before it learns the state. Changes: * Document reasons for style/script tag order in getHeadHtml (per 09537e83). * Pass $type from getModuleStyles() to getAllowedModules(). This wasn't needed before since a duplicate check in makeResourceLoaderLink() verified the origin a second time. * Declare explicit position 'top' on 'user.options' and 'user.tokens' module. Previously, OutputPage hardcoded them in the top. The new formatter doesn't. * Remove getHeadScripts(). * Remove getInlineHeadScripts(). * Remove getExternalHeadScripts(). * Remove buildCssLinks(). * Remove getScriptsForBottomQueue(). * Change where Skin::setupSkinUserCss() is called. This methods lets the skin add modules to the queue. Previously it was called from buildCssLinks(), via headElement(), via prepareQuickTemplate(), via OutputPage::output(). It's now in OutputPage::output() directly (slightly earlier). This is needed because prepareQuickTemplate() calls bottomScripts() before headElement(). And bottomScript() would lazy-initialise the queue and lock it before setupSkinUserCss() is called from headElement(). This makes execution order more predictable instead of being dependent on the arbitrary order of data extraction in prepareQuickTemplate (which varies from one skin to another). * Compute isUserModulePreview() and isKnownEmpty() for the 'user' module early on so. This avoids wrongful loading and fixes problem 1. Effective changes in output: * mw.loader.state() is now before mw.loader.load(). This fixes problem 2. * mw.loader.state() now sets 'user.options' and 'user.tokens' to "loading". * mw.loader.state() now sets 'user' (as "loading" or "ready"). Fixes problem 1. * The <script async src> tag for 'startup' changed position (slightly). Previously it was after all inline scripts and stylesheets. It's still after all inline scripts and after most stylesheets, but before any user styles. Since the queue is now formatted outside OutputPage, it can't inject the meta-ResourceLoaderDynamicStyles tag and user-stylesheet hack in the middle of existing output. This shouldn't have any noticable impact. Bug: T87871 Change-Id: I605b8cd1e1fc009b4662a0edbc54d09dd65ee1df
2016-07-15 14:13:09 +00:00
$this->getResourceLoader(),
new FauxRequest( $query )
);
Generalize ResourceLoader 'excludepage' functionality There has long been a hack for previewing edits to user JS/CSS, where OutputPage would pass an 'excludepage' parameter to ResourceLoaderUserModule to tell it not to load one particular page and would instead embed that page statically. That's nice, but there are other places where we could use the same thing. This patch generalizes it: * DerivativeResourceLoaderContext may now contain a callback for mapping titles to replacement Content objects. * ResourceLoaderWikiModule::getContent() uses the overrides, and requests embedding when they're used. All subclasses in Gerrit should pick it up automatically. * OutputPage gains methods for callers to add to the override mapping, which it passes on to RL. It loses a bunch of the special casing it had for the 'user' and 'user.styles' modules. * EditPage sets the overrides on OutputPage when doing the preview, as does ApiParse for prop=headhtml. TemplateSandbox does too in I83fa0856. * OutputPage::userCanPreview() gets less specific to editing user CSS and JS, since RL now handles the embedding based on the actual modules' dependencies and EditPage only requests it on preview. ApiParse also gets a new hook to support TemplateSandbox's API integration (used in I83fa0856). Bug: T112474 Change-Id: Ib9d2ce42931c1de8372e231314a1f672d7e2ac0e
2017-02-28 20:52:17 +00:00
if ( $this->contentOverrideCallbacks ) {
ResourceLoader namespace (attempt 2) Move ResourceLoader classes to their own namespace. Strip the "ResourceLoader" prefix from all except ResourceLoader itself. Move the tests by analogy. I used a namespace alias "RL" in some callers since RL\Module is less ambiguous at the call site than just "Module". I did not address DependencyStore which continues to have a non-standard location and namespace. Revert of a241d83e0a6dabedf. Bug: T308718 Change-Id: Id08a220e1d6085e2b33f3f6c9d0e3935a4204659
2022-05-06 09:09:56 +00:00
$this->rlClientContext = new RL\DerivativeContext( $this->rlClientContext );
resourceloader: Remove Title from public interfaces Bug: T278459 Depends-On: I9bc371fd299e7a5401ebc1b8d7456d9a4ecde90d Change-Id: Ib3a0591766f873492ec5a24721e7840f4e1a6407
2021-05-05 20:58:49 +00:00
$this->rlClientContext->setContentOverrideCallback( function ( $title ) {
Generalize ResourceLoader 'excludepage' functionality There has long been a hack for previewing edits to user JS/CSS, where OutputPage would pass an 'excludepage' parameter to ResourceLoaderUserModule to tell it not to load one particular page and would instead embed that page statically. That's nice, but there are other places where we could use the same thing. This patch generalizes it: * DerivativeResourceLoaderContext may now contain a callback for mapping titles to replacement Content objects. * ResourceLoaderWikiModule::getContent() uses the overrides, and requests embedding when they're used. All subclasses in Gerrit should pick it up automatically. * OutputPage gains methods for callers to add to the override mapping, which it passes on to RL. It loses a bunch of the special casing it had for the 'user' and 'user.styles' modules. * EditPage sets the overrides on OutputPage when doing the preview, as does ApiParse for prop=headhtml. TemplateSandbox does too in I83fa0856. * OutputPage::userCanPreview() gets less specific to editing user CSS and JS, since RL now handles the embedding based on the actual modules' dependencies and EditPage only requests it on preview. ApiParse also gets a new hook to support TemplateSandbox's API integration (used in I83fa0856). Bug: T112474 Change-Id: Ib9d2ce42931c1de8372e231314a1f672d7e2ac0e
2017-02-28 20:52:17 +00:00
foreach ( $this->contentOverrideCallbacks as $callback ) {
Comments and code cleanup in OutputPage Change-Id: I8c44f30801792e8b84ab3d556cb5b1b5feb9baf3
2018-08-02 18:26:55 +00:00
$content = $callback( $title );
Generalize ResourceLoader 'excludepage' functionality There has long been a hack for previewing edits to user JS/CSS, where OutputPage would pass an 'excludepage' parameter to ResourceLoaderUserModule to tell it not to load one particular page and would instead embed that page statically. That's nice, but there are other places where we could use the same thing. This patch generalizes it: * DerivativeResourceLoaderContext may now contain a callback for mapping titles to replacement Content objects. * ResourceLoaderWikiModule::getContent() uses the overrides, and requests embedding when they're used. All subclasses in Gerrit should pick it up automatically. * OutputPage gains methods for callers to add to the override mapping, which it passes on to RL. It loses a bunch of the special casing it had for the 'user' and 'user.styles' modules. * EditPage sets the overrides on OutputPage when doing the preview, as does ApiParse for prop=headhtml. TemplateSandbox does too in I83fa0856. * OutputPage::userCanPreview() gets less specific to editing user CSS and JS, since RL now handles the embedding based on the actual modules' dependencies and EditPage only requests it on preview. ApiParse also gets a new hook to support TemplateSandbox's API integration (used in I83fa0856). Bug: T112474 Change-Id: Ib9d2ce42931c1de8372e231314a1f672d7e2ac0e
2017-02-28 20:52:17 +00:00
if ( $content !== null ) {
Avoid using ContentHandler::getContentText() ContentHandler::getContentText() is deprecated and should be replaced with Content::getText() for TextContent instances. Change-Id: I556d3d3f64fafd1d54c4a0c5021efaff2d9c3ce8
2021-06-02 07:19:41 +00:00
$text = ( $content instanceof TextContent ) ? $content->getText() : '';
resourceloader: Refuse to preview content with </script> Bug: T200506 Change-Id: I4ab5fbb0f5413aad24360169ba635672ce8d9c8e
2018-08-20 00:14:46 +00:00
if ( strpos( $text, '</script>' ) !== false ) {
// Proactively replace this so that we can display a message
// to the user, instead of letting it go to Html::inlineScript(),
// where it would be considered a server-side issue.
$content = new JavaScriptContent(
Xml::encodeJsCall( 'mw.log.error', [
resourceloader: Remove Title from public interfaces Bug: T278459 Depends-On: I9bc371fd299e7a5401ebc1b8d7456d9a4ecde90d Change-Id: Ib3a0591766f873492ec5a24721e7840f4e1a6407
2021-05-05 20:58:49 +00:00
"Cannot preview $title due to script-closing tag."
resourceloader: Refuse to preview content with </script> Bug: T200506 Change-Id: I4ab5fbb0f5413aad24360169ba635672ce8d9c8e
2018-08-20 00:14:46 +00:00
] )
);
}
Generalize ResourceLoader 'excludepage' functionality There has long been a hack for previewing edits to user JS/CSS, where OutputPage would pass an 'excludepage' parameter to ResourceLoaderUserModule to tell it not to load one particular page and would instead embed that page statically. That's nice, but there are other places where we could use the same thing. This patch generalizes it: * DerivativeResourceLoaderContext may now contain a callback for mapping titles to replacement Content objects. * ResourceLoaderWikiModule::getContent() uses the overrides, and requests embedding when they're used. All subclasses in Gerrit should pick it up automatically. * OutputPage gains methods for callers to add to the override mapping, which it passes on to RL. It loses a bunch of the special casing it had for the 'user' and 'user.styles' modules. * EditPage sets the overrides on OutputPage when doing the preview, as does ApiParse for prop=headhtml. TemplateSandbox does too in I83fa0856. * OutputPage::userCanPreview() gets less specific to editing user CSS and JS, since RL now handles the embedding based on the actual modules' dependencies and EditPage only requests it on preview. ApiParse also gets a new hook to support TemplateSandbox's API integration (used in I83fa0856). Bug: T112474 Change-Id: Ib9d2ce42931c1de8372e231314a1f672d7e2ac0e
2017-02-28 20:52:17 +00:00
return $content;
}
}
return null;
} );
}
resourceloader: Move queue formatting out of OutputPage HTML formatting of the queue was distributed over several OutputPage methods. Each method demanding a snippet of HTML by calling makeResourceLoaderLink() with a limited amount of information. As such, makeResourceLoaderLink() was unable to provide the client with the proper state information. Centralising it also allows it to better reduce duplication in HTML output and maintain a more accurate state. Problems fixed by centralising: 1. The 'user' module is special (due to per-user 'version' and 'user' params). It is manually requested via script-src. To avoid a separate (and wrong) request from something that requires it, we set state=loading directly. However, because the module is in the bottom, the old HTML formatter could only put state=loading in the bottom also. This sometimes caused a wrong request to be fired for modules=user if something in the top queue triggered a requirement for it. 2. Since a464d1d4 (T87871) we track states of page-style modules, with purpose of allowing dependencies on style modules without risking duplicate loading on pages where the styles are loaded already. This didn't work, because the state information about page-style modules is output near the stylesheet, which is after the script tag with mw.loader.load(). That runs first, and mw.loader would still make a duplicate request before it learns the state. Changes: * Document reasons for style/script tag order in getHeadHtml (per 09537e83). * Pass $type from getModuleStyles() to getAllowedModules(). This wasn't needed before since a duplicate check in makeResourceLoaderLink() verified the origin a second time. * Declare explicit position 'top' on 'user.options' and 'user.tokens' module. Previously, OutputPage hardcoded them in the top. The new formatter doesn't. * Remove getHeadScripts(). * Remove getInlineHeadScripts(). * Remove getExternalHeadScripts(). * Remove buildCssLinks(). * Remove getScriptsForBottomQueue(). * Change where Skin::setupSkinUserCss() is called. This methods lets the skin add modules to the queue. Previously it was called from buildCssLinks(), via headElement(), via prepareQuickTemplate(), via OutputPage::output(). It's now in OutputPage::output() directly (slightly earlier). This is needed because prepareQuickTemplate() calls bottomScripts() before headElement(). And bottomScript() would lazy-initialise the queue and lock it before setupSkinUserCss() is called from headElement(). This makes execution order more predictable instead of being dependent on the arbitrary order of data extraction in prepareQuickTemplate (which varies from one skin to another). * Compute isUserModulePreview() and isKnownEmpty() for the 'user' module early on so. This avoids wrongful loading and fixes problem 1. Effective changes in output: * mw.loader.state() is now before mw.loader.load(). This fixes problem 2. * mw.loader.state() now sets 'user.options' and 'user.tokens' to "loading". * mw.loader.state() now sets 'user' (as "loading" or "ready"). Fixes problem 1. * The <script async src> tag for 'startup' changed position (slightly). Previously it was after all inline scripts and stylesheets. It's still after all inline scripts and after most stylesheets, but before any user styles. Since the queue is now formatted outside OutputPage, it can't inject the meta-ResourceLoaderDynamicStyles tag and user-stylesheet hack in the middle of existing output. This shouldn't have any noticable impact. Bug: T87871 Change-Id: I605b8cd1e1fc009b4662a0edbc54d09dd65ee1df
2016-07-15 14:13:09 +00:00
}
return $this->rlClientContext;
}
/**
* Call this to freeze the module queue and JS config and create a formatter.
*
* Depending on the Skin, this may get lazy-initialised in either headElement() or
* getBottomScripts(). See SkinTemplate::prepareQuickTemplate(). Calling this too early may
* cause unexpected side-effects since disallowUserJs() may be called at any time to change
* the module filters retroactively. Skins and extension hooks may also add modules until very
* late in the request lifecycle.
*
ResourceLoader namespace (attempt 2) Move ResourceLoader classes to their own namespace. Strip the "ResourceLoader" prefix from all except ResourceLoader itself. Move the tests by analogy. I used a namespace alias "RL" in some callers since RL\Module is less ambiguous at the call site than just "Module". I did not address DependencyStore which continues to have a non-standard location and namespace. Revert of a241d83e0a6dabedf. Bug: T308718 Change-Id: Id08a220e1d6085e2b33f3f6c9d0e3935a4204659
2022-05-06 09:09:56 +00:00
* @return RL\ClientHtml
resourceloader: Move queue formatting out of OutputPage HTML formatting of the queue was distributed over several OutputPage methods. Each method demanding a snippet of HTML by calling makeResourceLoaderLink() with a limited amount of information. As such, makeResourceLoaderLink() was unable to provide the client with the proper state information. Centralising it also allows it to better reduce duplication in HTML output and maintain a more accurate state. Problems fixed by centralising: 1. The 'user' module is special (due to per-user 'version' and 'user' params). It is manually requested via script-src. To avoid a separate (and wrong) request from something that requires it, we set state=loading directly. However, because the module is in the bottom, the old HTML formatter could only put state=loading in the bottom also. This sometimes caused a wrong request to be fired for modules=user if something in the top queue triggered a requirement for it. 2. Since a464d1d4 (T87871) we track states of page-style modules, with purpose of allowing dependencies on style modules without risking duplicate loading on pages where the styles are loaded already. This didn't work, because the state information about page-style modules is output near the stylesheet, which is after the script tag with mw.loader.load(). That runs first, and mw.loader would still make a duplicate request before it learns the state. Changes: * Document reasons for style/script tag order in getHeadHtml (per 09537e83). * Pass $type from getModuleStyles() to getAllowedModules(). This wasn't needed before since a duplicate check in makeResourceLoaderLink() verified the origin a second time. * Declare explicit position 'top' on 'user.options' and 'user.tokens' module. Previously, OutputPage hardcoded them in the top. The new formatter doesn't. * Remove getHeadScripts(). * Remove getInlineHeadScripts(). * Remove getExternalHeadScripts(). * Remove buildCssLinks(). * Remove getScriptsForBottomQueue(). * Change where Skin::setupSkinUserCss() is called. This methods lets the skin add modules to the queue. Previously it was called from buildCssLinks(), via headElement(), via prepareQuickTemplate(), via OutputPage::output(). It's now in OutputPage::output() directly (slightly earlier). This is needed because prepareQuickTemplate() calls bottomScripts() before headElement(). And bottomScript() would lazy-initialise the queue and lock it before setupSkinUserCss() is called from headElement(). This makes execution order more predictable instead of being dependent on the arbitrary order of data extraction in prepareQuickTemplate (which varies from one skin to another). * Compute isUserModulePreview() and isKnownEmpty() for the 'user' module early on so. This avoids wrongful loading and fixes problem 1. Effective changes in output: * mw.loader.state() is now before mw.loader.load(). This fixes problem 2. * mw.loader.state() now sets 'user.options' and 'user.tokens' to "loading". * mw.loader.state() now sets 'user' (as "loading" or "ready"). Fixes problem 1. * The <script async src> tag for 'startup' changed position (slightly). Previously it was after all inline scripts and stylesheets. It's still after all inline scripts and after most stylesheets, but before any user styles. Since the queue is now formatted outside OutputPage, it can't inject the meta-ResourceLoaderDynamicStyles tag and user-stylesheet hack in the middle of existing output. This shouldn't have any noticable impact. Bug: T87871 Change-Id: I605b8cd1e1fc009b4662a0edbc54d09dd65ee1df
2016-07-15 14:13:09 +00:00
*/
public function getRlClient() {
if ( !$this->rlClient ) {
$context = $this->getRlClientContext();
OutputPage: Make ResourceLoader position exemption more generic Follows-up 80e5b160e which moved queue formatting out of OutputPage into a a separate ResourceLoaderClientHtml class. The special handling for 'user' and 'user.styles' modules, and the exempt module groups was kept in OutputPage. However the handling for it was hardcoded for the modules in that group by default. It did not account for modules with a group of 'user' loaded by an extension (e.g. GlobalCssJs). GlobalCssJs modules were wrongly loaded in the regular style queue (still in a separate request group, but not in the right cascading order below the DynamicSyles marker). This commit generalises the handling previously put in buildExemptModules and moves it to getRlClient() so that it may apply to all style modules. This commit should be a no-op besides the moving of any <link rel=stylesheet> for non-core modules in group 'site' or 'user' now being one line lower in the <head> HTML (after the DynamicStyles marker). Bug: T143357 Change-Id: I1d6ea10b42293acfc535578172ad7ab2369f6299
2016-08-19 02:04:21 +00:00
$rl = $this->getResourceLoader();
resourceloader: Move queue formatting out of OutputPage HTML formatting of the queue was distributed over several OutputPage methods. Each method demanding a snippet of HTML by calling makeResourceLoaderLink() with a limited amount of information. As such, makeResourceLoaderLink() was unable to provide the client with the proper state information. Centralising it also allows it to better reduce duplication in HTML output and maintain a more accurate state. Problems fixed by centralising: 1. The 'user' module is special (due to per-user 'version' and 'user' params). It is manually requested via script-src. To avoid a separate (and wrong) request from something that requires it, we set state=loading directly. However, because the module is in the bottom, the old HTML formatter could only put state=loading in the bottom also. This sometimes caused a wrong request to be fired for modules=user if something in the top queue triggered a requirement for it. 2. Since a464d1d4 (T87871) we track states of page-style modules, with purpose of allowing dependencies on style modules without risking duplicate loading on pages where the styles are loaded already. This didn't work, because the state information about page-style modules is output near the stylesheet, which is after the script tag with mw.loader.load(). That runs first, and mw.loader would still make a duplicate request before it learns the state. Changes: * Document reasons for style/script tag order in getHeadHtml (per 09537e83). * Pass $type from getModuleStyles() to getAllowedModules(). This wasn't needed before since a duplicate check in makeResourceLoaderLink() verified the origin a second time. * Declare explicit position 'top' on 'user.options' and 'user.tokens' module. Previously, OutputPage hardcoded them in the top. The new formatter doesn't. * Remove getHeadScripts(). * Remove getInlineHeadScripts(). * Remove getExternalHeadScripts(). * Remove buildCssLinks(). * Remove getScriptsForBottomQueue(). * Change where Skin::setupSkinUserCss() is called. This methods lets the skin add modules to the queue. Previously it was called from buildCssLinks(), via headElement(), via prepareQuickTemplate(), via OutputPage::output(). It's now in OutputPage::output() directly (slightly earlier). This is needed because prepareQuickTemplate() calls bottomScripts() before headElement(). And bottomScript() would lazy-initialise the queue and lock it before setupSkinUserCss() is called from headElement(). This makes execution order more predictable instead of being dependent on the arbitrary order of data extraction in prepareQuickTemplate (which varies from one skin to another). * Compute isUserModulePreview() and isKnownEmpty() for the 'user' module early on so. This avoids wrongful loading and fixes problem 1. Effective changes in output: * mw.loader.state() is now before mw.loader.load(). This fixes problem 2. * mw.loader.state() now sets 'user.options' and 'user.tokens' to "loading". * mw.loader.state() now sets 'user' (as "loading" or "ready"). Fixes problem 1. * The <script async src> tag for 'startup' changed position (slightly). Previously it was after all inline scripts and stylesheets. It's still after all inline scripts and after most stylesheets, but before any user styles. Since the queue is now formatted outside OutputPage, it can't inject the meta-ResourceLoaderDynamicStyles tag and user-stylesheet hack in the middle of existing output. This shouldn't have any noticable impact. Bug: T87871 Change-Id: I605b8cd1e1fc009b4662a0edbc54d09dd65ee1df
2016-07-15 14:13:09 +00:00
$this->addModules( [
Generalize ResourceLoader 'excludepage' functionality There has long been a hack for previewing edits to user JS/CSS, where OutputPage would pass an 'excludepage' parameter to ResourceLoaderUserModule to tell it not to load one particular page and would instead embed that page statically. That's nice, but there are other places where we could use the same thing. This patch generalizes it: * DerivativeResourceLoaderContext may now contain a callback for mapping titles to replacement Content objects. * ResourceLoaderWikiModule::getContent() uses the overrides, and requests embedding when they're used. All subclasses in Gerrit should pick it up automatically. * OutputPage gains methods for callers to add to the override mapping, which it passes on to RL. It loses a bunch of the special casing it had for the 'user' and 'user.styles' modules. * EditPage sets the overrides on OutputPage when doing the preview, as does ApiParse for prop=headhtml. TemplateSandbox does too in I83fa0856. * OutputPage::userCanPreview() gets less specific to editing user CSS and JS, since RL now handles the embedding based on the actual modules' dependencies and EditPage only requests it on preview. ApiParse also gets a new hook to support TemplateSandbox's API integration (used in I83fa0856). Bug: T112474 Change-Id: Ib9d2ce42931c1de8372e231314a1f672d7e2ac0e
2017-02-28 20:52:17 +00:00
'user',
resourceloader: Move queue formatting out of OutputPage HTML formatting of the queue was distributed over several OutputPage methods. Each method demanding a snippet of HTML by calling makeResourceLoaderLink() with a limited amount of information. As such, makeResourceLoaderLink() was unable to provide the client with the proper state information. Centralising it also allows it to better reduce duplication in HTML output and maintain a more accurate state. Problems fixed by centralising: 1. The 'user' module is special (due to per-user 'version' and 'user' params). It is manually requested via script-src. To avoid a separate (and wrong) request from something that requires it, we set state=loading directly. However, because the module is in the bottom, the old HTML formatter could only put state=loading in the bottom also. This sometimes caused a wrong request to be fired for modules=user if something in the top queue triggered a requirement for it. 2. Since a464d1d4 (T87871) we track states of page-style modules, with purpose of allowing dependencies on style modules without risking duplicate loading on pages where the styles are loaded already. This didn't work, because the state information about page-style modules is output near the stylesheet, which is after the script tag with mw.loader.load(). That runs first, and mw.loader would still make a duplicate request before it learns the state. Changes: * Document reasons for style/script tag order in getHeadHtml (per 09537e83). * Pass $type from getModuleStyles() to getAllowedModules(). This wasn't needed before since a duplicate check in makeResourceLoaderLink() verified the origin a second time. * Declare explicit position 'top' on 'user.options' and 'user.tokens' module. Previously, OutputPage hardcoded them in the top. The new formatter doesn't. * Remove getHeadScripts(). * Remove getInlineHeadScripts(). * Remove getExternalHeadScripts(). * Remove buildCssLinks(). * Remove getScriptsForBottomQueue(). * Change where Skin::setupSkinUserCss() is called. This methods lets the skin add modules to the queue. Previously it was called from buildCssLinks(), via headElement(), via prepareQuickTemplate(), via OutputPage::output(). It's now in OutputPage::output() directly (slightly earlier). This is needed because prepareQuickTemplate() calls bottomScripts() before headElement(). And bottomScript() would lazy-initialise the queue and lock it before setupSkinUserCss() is called from headElement(). This makes execution order more predictable instead of being dependent on the arbitrary order of data extraction in prepareQuickTemplate (which varies from one skin to another). * Compute isUserModulePreview() and isKnownEmpty() for the 'user' module early on so. This avoids wrongful loading and fixes problem 1. Effective changes in output: * mw.loader.state() is now before mw.loader.load(). This fixes problem 2. * mw.loader.state() now sets 'user.options' and 'user.tokens' to "loading". * mw.loader.state() now sets 'user' (as "loading" or "ready"). Fixes problem 1. * The <script async src> tag for 'startup' changed position (slightly). Previously it was after all inline scripts and stylesheets. It's still after all inline scripts and after most stylesheets, but before any user styles. Since the queue is now formatted outside OutputPage, it can't inject the meta-ResourceLoaderDynamicStyles tag and user-stylesheet hack in the middle of existing output. This shouldn't have any noticable impact. Bug: T87871 Change-Id: I605b8cd1e1fc009b4662a0edbc54d09dd65ee1df
2016-07-15 14:13:09 +00:00
'user.options',
] );
OutputPage: Make ResourceLoader position exemption more generic Follows-up 80e5b160e which moved queue formatting out of OutputPage into a a separate ResourceLoaderClientHtml class. The special handling for 'user' and 'user.styles' modules, and the exempt module groups was kept in OutputPage. However the handling for it was hardcoded for the modules in that group by default. It did not account for modules with a group of 'user' loaded by an extension (e.g. GlobalCssJs). GlobalCssJs modules were wrongly loaded in the regular style queue (still in a separate request group, but not in the right cascading order below the DynamicSyles marker). This commit generalises the handling previously put in buildExemptModules and moves it to getRlClient() so that it may apply to all style modules. This commit should be a no-op besides the moving of any <link rel=stylesheet> for non-core modules in group 'site' or 'user' now being one line lower in the <head> HTML (after the DynamicStyles marker). Bug: T143357 Change-Id: I1d6ea10b42293acfc535578172ad7ab2369f6299
2016-08-19 02:04:21 +00:00
$this->addModuleStyles( [
'site.styles',
'noscript',
'user.styles',
] );
// Prepare exempt modules for buildExemptModules()
resourceloader: Use named constants for groups Makes the code easier to follow, in particular to identify functionality of the various groups. Change-Id: Ib2dae5757e9325f9706ff2a9f760f311475fe615
2022-01-12 14:34:02 +00:00
$exemptGroups = [
ResourceLoader namespace (attempt 2) Move ResourceLoader classes to their own namespace. Strip the "ResourceLoader" prefix from all except ResourceLoader itself. Move the tests by analogy. I used a namespace alias "RL" in some callers since RL\Module is less ambiguous at the call site than just "Module". I did not address DependencyStore which continues to have a non-standard location and namespace. Revert of a241d83e0a6dabedf. Bug: T308718 Change-Id: Id08a220e1d6085e2b33f3f6c9d0e3935a4204659
2022-05-06 09:09:56 +00:00
RL\Module::GROUP_SITE => [],
RL\Module::GROUP_NOSCRIPT => [],
RL\Module::GROUP_PRIVATE => [],
RL\Module::GROUP_USER => []
resourceloader: Use named constants for groups Makes the code easier to follow, in particular to identify functionality of the various groups. Change-Id: Ib2dae5757e9325f9706ff2a9f760f311475fe615
2022-01-12 14:34:02 +00:00
];
OutputPage: Make ResourceLoader position exemption more generic Follows-up 80e5b160e which moved queue formatting out of OutputPage into a a separate ResourceLoaderClientHtml class. The special handling for 'user' and 'user.styles' modules, and the exempt module groups was kept in OutputPage. However the handling for it was hardcoded for the modules in that group by default. It did not account for modules with a group of 'user' loaded by an extension (e.g. GlobalCssJs). GlobalCssJs modules were wrongly loaded in the regular style queue (still in a separate request group, but not in the right cascading order below the DynamicSyles marker). This commit generalises the handling previously put in buildExemptModules and moves it to getRlClient() so that it may apply to all style modules. This commit should be a no-op besides the moving of any <link rel=stylesheet> for non-core modules in group 'site' or 'user' now being one line lower in the <head> HTML (after the DynamicStyles marker). Bug: T143357 Change-Id: I1d6ea10b42293acfc535578172ad7ab2369f6299
2016-08-19 02:04:21 +00:00
$exemptStates = [];
Preload ResourceLoaderWikiModule::getTitleInfo in OutputPage This avoids a separate query for each module. Bug: T46362 Change-Id: Ie109a8776cbdcd5928cbb59351f2cf94088c0c95
2016-09-08 08:38:26 +00:00
$moduleStyles = $this->getModuleStyles( /*filter*/ true );
ResourceLoader namespace (attempt 2) Move ResourceLoader classes to their own namespace. Strip the "ResourceLoader" prefix from all except ResourceLoader itself. Move the tests by analogy. I used a namespace alias "RL" in some callers since RL\Module is less ambiguous at the call site than just "Module". I did not address DependencyStore which continues to have a non-standard location and namespace. Revert of a241d83e0a6dabedf. Bug: T308718 Change-Id: Id08a220e1d6085e2b33f3f6c9d0e3935a4204659
2022-05-06 09:09:56 +00:00
// Preload getTitleInfo for isKnownEmpty calls below and in RL\ClientHtml
OutputPage: Reduce getTitleInfo() calls (improve preloading) * Remove filter between $moduleStyles and preloadTitleInfo(). Previously, this was filtering out wiki modules created by gadgets and extensions. Still causing a couple of direct queries from getTitleInfo(). * Store __METHOD__ in $fname outside getWithSetCallback. This way queries are logged as either getTitleInfo or preloadTitleInfo. This regressed in 0852a000a5, after which the latter was logged as "::{closure}". Change-Id: I454e43f43f8ad3270b86f4cfdbd68192c305cc07
2016-10-20 21:43:57 +00:00
// Separate user-specific batch for improved cache-hit ratio.
$userBatch = [ 'user.styles', 'user' ];
$siteBatch = array_diff( $moduleStyles, $userBatch );
$dbr = wfGetDB( DB_REPLICA );
ResourceLoader namespace (attempt 2) Move ResourceLoader classes to their own namespace. Strip the "ResourceLoader" prefix from all except ResourceLoader itself. Move the tests by analogy. I used a namespace alias "RL" in some callers since RL\Module is less ambiguous at the call site than just "Module". I did not address DependencyStore which continues to have a non-standard location and namespace. Revert of a241d83e0a6dabedf. Bug: T308718 Change-Id: Id08a220e1d6085e2b33f3f6c9d0e3935a4204659
2022-05-06 09:09:56 +00:00
RL\WikiModule::preloadTitleInfo( $context, $dbr, $siteBatch );
RL\WikiModule::preloadTitleInfo( $context, $dbr, $userBatch );
Preload ResourceLoaderWikiModule::getTitleInfo in OutputPage This avoids a separate query for each module. Bug: T46362 Change-Id: Ie109a8776cbdcd5928cbb59351f2cf94088c0c95
2016-09-08 08:38:26 +00:00
// Filter out modules handled by buildExemptModules()
$moduleStyles = array_filter( $moduleStyles,
Use static closures where safe to use This is micro-optimization of closure code to avoid binding the closure to $this where it is not needed. Created by I25a17fb22b6b669e817317a0f45051ae9c608208 Change-Id: I0ffc6200f6c6693d78a3151cb8cea7dce7c21653
2021-02-10 22:31:02 +00:00
static function ( $name ) use ( $rl, $context, &$exemptGroups, &$exemptStates ) {
OutputPage: Make ResourceLoader position exemption more generic Follows-up 80e5b160e which moved queue formatting out of OutputPage into a a separate ResourceLoaderClientHtml class. The special handling for 'user' and 'user.styles' modules, and the exempt module groups was kept in OutputPage. However the handling for it was hardcoded for the modules in that group by default. It did not account for modules with a group of 'user' loaded by an extension (e.g. GlobalCssJs). GlobalCssJs modules were wrongly loaded in the regular style queue (still in a separate request group, but not in the right cascading order below the DynamicSyles marker). This commit generalises the handling previously put in buildExemptModules and moves it to getRlClient() so that it may apply to all style modules. This commit should be a no-op besides the moving of any <link rel=stylesheet> for non-core modules in group 'site' or 'user' now being one line lower in the <head> HTML (after the DynamicStyles marker). Bug: T143357 Change-Id: I1d6ea10b42293acfc535578172ad7ab2369f6299
2016-08-19 02:04:21 +00:00
$module = $rl->getModule( $name );
if ( $module ) {
Preload ResourceLoaderWikiModule::getTitleInfo in OutputPage This avoids a separate query for each module. Bug: T46362 Change-Id: Ie109a8776cbdcd5928cbb59351f2cf94088c0c95
2016-09-08 08:38:26 +00:00
$group = $module->getGroup();
Add various null checks when null is not possible to use as argument Also check for false if needed Found by phan strict checks Change-Id: I298204653dfb788515a87978dd8705b6e4f9c775
2022-03-09 19:14:13 +00:00
if ( $group !== null && isset( $exemptGroups[$group] ) ) {
resourceloader: omit `noscript` module from client registry In all cases where the client registry is used, the `noscript` module is inapplicable. Additionally, make it illegal to have a module include the 'noscript' module as a dependency. Bug: T291735 Change-Id: I5846f2c19b003989b56b12628b385fd37049483b
2021-09-28 01:37:54 +00:00
// The `noscript` module is excluded from the client
// side registry, no need to set its state either.
// But we still output it. See T291735
ResourceLoader namespace (attempt 2) Move ResourceLoader classes to their own namespace. Strip the "ResourceLoader" prefix from all except ResourceLoader itself. Move the tests by analogy. I used a namespace alias "RL" in some callers since RL\Module is less ambiguous at the call site than just "Module". I did not address DependencyStore which continues to have a non-standard location and namespace. Revert of a241d83e0a6dabedf. Bug: T308718 Change-Id: Id08a220e1d6085e2b33f3f6c9d0e3935a4204659
2022-05-06 09:09:56 +00:00
if ( $group !== RL\Module::GROUP_NOSCRIPT ) {
resourceloader: omit `noscript` module from client registry In all cases where the client registry is used, the `noscript` module is inapplicable. Additionally, make it illegal to have a module include the 'noscript' module as a dependency. Bug: T291735 Change-Id: I5846f2c19b003989b56b12628b385fd37049483b
2021-09-28 01:37:54 +00:00
$exemptStates[$name] = 'ready';
}
OutputPage: Make ResourceLoader position exemption more generic Follows-up 80e5b160e which moved queue formatting out of OutputPage into a a separate ResourceLoaderClientHtml class. The special handling for 'user' and 'user.styles' modules, and the exempt module groups was kept in OutputPage. However the handling for it was hardcoded for the modules in that group by default. It did not account for modules with a group of 'user' loaded by an extension (e.g. GlobalCssJs). GlobalCssJs modules were wrongly loaded in the regular style queue (still in a separate request group, but not in the right cascading order below the DynamicSyles marker). This commit generalises the handling previously put in buildExemptModules and moves it to getRlClient() so that it may apply to all style modules. This commit should be a no-op besides the moving of any <link rel=stylesheet> for non-core modules in group 'site' or 'user' now being one line lower in the <head> HTML (after the DynamicStyles marker). Bug: T143357 Change-Id: I1d6ea10b42293acfc535578172ad7ab2369f6299
2016-08-19 02:04:21 +00:00
if ( !$module->isKnownEmpty( $context ) ) {
// E.g. Don't output empty <styles>
$exemptGroups[$group][] = $name;
}
return false;
}
}
return true;
}
);
$this->rlExemptStyleModules = $exemptGroups;
ResourceLoader: Follow-up to adding ResourceLoaderClientPreferences Minor clean up: * Fix broken `@see` in MainConfigSchema. * Add missing `@since`. * Doc experimental nature, as per the Ic3b6eec1995393 msg. * Doc anonymous scope, and mention general strategy for elsewhere. * Dependency inject, to separate concerns and keep ClientHtml decoupled from MediaWiki settings like wgCookiePrefix, which otherwise break testing this class with only RL-specific config. * Apply JS conventions. Made easier by using "JS" as the heredoc identifier, which IDEs recognise as for highlighting. * Move code together with the other documentElement.className statement. This helps both with understanding the PHP side in terms of related logic and how it interacts, as well as the frontend as it literally brings <html script>, client-js, and clientpref all next to each other. HTML weight of default Main_Page on localhost, logged-out: * vector : 23.21 kB / 140.48 kB * vector-2022 : 24.17 kB / 146.43 kB +6.0 kB * vector-2022 ClientPref=true before : 24.28 kB / 146.70 kB +6.3 kB * vector-2022 ClientPref=true after : 24.27 kB / 146.68 kB +6.2 kB Given: * $wgResourceLoaderClientPreferences = true; Test plan: 1. View /wiki/Main_Page?useskin=vector-2022 while logged-out, confirm the inline script is there, and "vector-feature-limited-width-content" is set on <html class>, layout width appears fixed. 2. Run mw.cookie.set('mwclientprefs', 'vector-feature-limited-width-content'); 3. Refresh and confirm the class is changed at runtime and layout width is fluid. Use mw.cookie.set('mwclientprefs', null) to undo. Bug: T321498 Change-Id: I07f471b815ffadfca9eb4f7bd228cb72dfd1ec9b
2023-01-25 01:21:41 +00:00
$config = $this->getConfig();
$clientPrefEnabled = (
$config->get( MainConfigNames::ResourceLoaderClientPreferences ) &&
!$this->getUser()->isRegistered()
);
$clientPrefCookiePrefix = $config->get( MainConfigNames::CookiePrefix );
ResourceLoader namespace (attempt 2) Move ResourceLoader classes to their own namespace. Strip the "ResourceLoader" prefix from all except ResourceLoader itself. Move the tests by analogy. I used a namespace alias "RL" in some callers since RL\Module is less ambiguous at the call site than just "Module". I did not address DependencyStore which continues to have a non-standard location and namespace. Revert of a241d83e0a6dabedf. Bug: T308718 Change-Id: Id08a220e1d6085e2b33f3f6c9d0e3935a4204659
2022-05-06 09:09:56 +00:00
$rlClient = new RL\ClientHtml( $context, [
resourceloader: Replace ClientHtml 'target' param with 'options' array In preparation for passing down 'safemode' from OutputPage. Only used in one place in Wikimedia Git: OutputPage::getRlClient(). Bug: T185303 Change-Id: If01eca96986ff8d7dcdaab6910bf183ba7c7311f
2018-03-27 21:28:03 +00:00
'target' => $this->getTarget(),
Mild refactoring of ContentSecurityPolicy This is to make it behave in a more object orientied way. The goal is to make it be easier to allow extensions to mark certain pages as requiring a different policy (For example, CodeEditor extension uses a blob: url with a WebWorker. We don't want to include that on the policy of every page, but allow the extension to mark it as required whenever needed). This commit does not change code behaviour in any way. Change-Id: I4bf53dabb6e6c5446cea99a64db68b300cef2fd4
2019-10-28 05:01:17 +00:00
'nonce' => $this->CSP->getNonce(),
resourceloader: Apply safemode to startup module registry This effectively applies safemode to the mw.loader client, without the client itself needing specific knowledge of safemode. Test Plan: * Unchanged: When viewing a page in safemode, the 'user' and 'site' modules are still not queued by OutputPage. * New: mw.loader.getState('site'), previously would yield 'registered', but will now yield null. * New: mw.loader.load('site'), previously loaded the module, it now logs a dependency warning for unknown module, like for any other unknown module. * New: mw.loader.using('site'), previously resolved, it is now rejected. Bug: T185303 Change-Id: I672e3891c8e1b3c2d13655fa134d0f1d031b8247
2018-03-28 00:57:06 +00:00
// When 'safemode', disallowUserJs(), or reduceAllowedModules() is used
// to only restrict modules to ORIGIN_CORE (ie. disallow ORIGIN_USER), the list of
Fix typos in comments (N-R) Change-Id: I2d1bdb7531ff5126114a391550c2615ea6e244b3
2022-01-09 17:44:44 +00:00
// modules enqueued for loading on this page is filtered to just those.
resourceloader: Apply safemode to startup module registry This effectively applies safemode to the mw.loader client, without the client itself needing specific knowledge of safemode. Test Plan: * Unchanged: When viewing a page in safemode, the 'user' and 'site' modules are still not queued by OutputPage. * New: mw.loader.getState('site'), previously would yield 'registered', but will now yield null. * New: mw.loader.load('site'), previously loaded the module, it now logs a dependency warning for unknown module, like for any other unknown module. * New: mw.loader.using('site'), previously resolved, it is now rejected. Bug: T185303 Change-Id: I672e3891c8e1b3c2d13655fa134d0f1d031b8247
2018-03-28 00:57:06 +00:00
// However, to make sure we also apply the restriction to dynamic dependencies and
// lazy-loaded modules at run-time on the client-side, pass 'safemode' down to the
// StartupModule so that the client-side registry will not contain any restricted
// modules either. (T152169, T185303)
ResourceLoader namespace (attempt 2) Move ResourceLoader classes to their own namespace. Strip the "ResourceLoader" prefix from all except ResourceLoader itself. Move the tests by analogy. I used a namespace alias "RL" in some callers since RL\Module is less ambiguous at the call site than just "Module". I did not address DependencyStore which continues to have a non-standard location and namespace. Revert of a241d83e0a6dabedf. Bug: T308718 Change-Id: Id08a220e1d6085e2b33f3f6c9d0e3935a4204659
2022-05-06 09:09:56 +00:00
'safemode' => ( $this->getAllowedModules( RL\Module::TYPE_COMBINED )
<= RL\Module::ORIGIN_CORE_INDIVIDUAL
resourceloader: Apply safemode to startup module registry This effectively applies safemode to the mw.loader client, without the client itself needing specific knowledge of safemode. Test Plan: * Unchanged: When viewing a page in safemode, the 'user' and 'site' modules are still not queued by OutputPage. * New: mw.loader.getState('site'), previously would yield 'registered', but will now yield null. * New: mw.loader.load('site'), previously loaded the module, it now logs a dependency warning for unknown module, like for any other unknown module. * New: mw.loader.using('site'), previously resolved, it is now rejected. Bug: T185303 Change-Id: I672e3891c8e1b3c2d13655fa134d0f1d031b8247
2018-03-28 00:57:06 +00:00
) ? '1' : null,
ResourceLoader: Follow-up to adding ResourceLoaderClientPreferences Minor clean up: * Fix broken `@see` in MainConfigSchema. * Add missing `@since`. * Doc experimental nature, as per the Ic3b6eec1995393 msg. * Doc anonymous scope, and mention general strategy for elsewhere. * Dependency inject, to separate concerns and keep ClientHtml decoupled from MediaWiki settings like wgCookiePrefix, which otherwise break testing this class with only RL-specific config. * Apply JS conventions. Made easier by using "JS" as the heredoc identifier, which IDEs recognise as for highlighting. * Move code together with the other documentElement.className statement. This helps both with understanding the PHP side in terms of related logic and how it interacts, as well as the frontend as it literally brings <html script>, client-js, and clientpref all next to each other. HTML weight of default Main_Page on localhost, logged-out: * vector : 23.21 kB / 140.48 kB * vector-2022 : 24.17 kB / 146.43 kB +6.0 kB * vector-2022 ClientPref=true before : 24.28 kB / 146.70 kB +6.3 kB * vector-2022 ClientPref=true after : 24.27 kB / 146.68 kB +6.2 kB Given: * $wgResourceLoaderClientPreferences = true; Test plan: 1. View /wiki/Main_Page?useskin=vector-2022 while logged-out, confirm the inline script is there, and "vector-feature-limited-width-content" is set on <html class>, layout width appears fixed. 2. Run mw.cookie.set('mwclientprefs', 'vector-feature-limited-width-content'); 3. Refresh and confirm the class is changed at runtime and layout width is fluid. Use mw.cookie.set('mwclientprefs', null) to undo. Bug: T321498 Change-Id: I07f471b815ffadfca9eb4f7bd228cb72dfd1ec9b
2023-01-25 01:21:41 +00:00
'clientPrefEnabled' => $clientPrefEnabled,
'clientPrefCookiePrefix' => $clientPrefCookiePrefix,
resourceloader: Replace ClientHtml 'target' param with 'options' array In preparation for passing down 'safemode' from OutputPage. Only used in one place in Wikimedia Git: OutputPage::getRlClient(). Bug: T185303 Change-Id: If01eca96986ff8d7dcdaab6910bf183ba7c7311f
2018-03-27 21:28:03 +00:00
] );
OutputPage: Support loading some mw.config vars late, at end of body JS config vars listed in the 'LateJSConfigVarNames' extension attribute or in the OutputPage::CORE_LATE_JS_CONFIG_VAR_NAMES constant will now be added to the end of the HTML response, instead of early in the HTML head where it blocks rendering/painting. Bug: T41813 Change-Id: I86d461756398780dc24949013f35b7730a481052
2022-10-25 22:36:20 +00:00
$rlClient->setConfig( $this->getJSVars( self::JS_VAR_EARLY ) );
resourceloader: Move queue formatting out of OutputPage HTML formatting of the queue was distributed over several OutputPage methods. Each method demanding a snippet of HTML by calling makeResourceLoaderLink() with a limited amount of information. As such, makeResourceLoaderLink() was unable to provide the client with the proper state information. Centralising it also allows it to better reduce duplication in HTML output and maintain a more accurate state. Problems fixed by centralising: 1. The 'user' module is special (due to per-user 'version' and 'user' params). It is manually requested via script-src. To avoid a separate (and wrong) request from something that requires it, we set state=loading directly. However, because the module is in the bottom, the old HTML formatter could only put state=loading in the bottom also. This sometimes caused a wrong request to be fired for modules=user if something in the top queue triggered a requirement for it. 2. Since a464d1d4 (T87871) we track states of page-style modules, with purpose of allowing dependencies on style modules without risking duplicate loading on pages where the styles are loaded already. This didn't work, because the state information about page-style modules is output near the stylesheet, which is after the script tag with mw.loader.load(). That runs first, and mw.loader would still make a duplicate request before it learns the state. Changes: * Document reasons for style/script tag order in getHeadHtml (per 09537e83). * Pass $type from getModuleStyles() to getAllowedModules(). This wasn't needed before since a duplicate check in makeResourceLoaderLink() verified the origin a second time. * Declare explicit position 'top' on 'user.options' and 'user.tokens' module. Previously, OutputPage hardcoded them in the top. The new formatter doesn't. * Remove getHeadScripts(). * Remove getInlineHeadScripts(). * Remove getExternalHeadScripts(). * Remove buildCssLinks(). * Remove getScriptsForBottomQueue(). * Change where Skin::setupSkinUserCss() is called. This methods lets the skin add modules to the queue. Previously it was called from buildCssLinks(), via headElement(), via prepareQuickTemplate(), via OutputPage::output(). It's now in OutputPage::output() directly (slightly earlier). This is needed because prepareQuickTemplate() calls bottomScripts() before headElement(). And bottomScript() would lazy-initialise the queue and lock it before setupSkinUserCss() is called from headElement(). This makes execution order more predictable instead of being dependent on the arbitrary order of data extraction in prepareQuickTemplate (which varies from one skin to another). * Compute isUserModulePreview() and isKnownEmpty() for the 'user' module early on so. This avoids wrongful loading and fixes problem 1. Effective changes in output: * mw.loader.state() is now before mw.loader.load(). This fixes problem 2. * mw.loader.state() now sets 'user.options' and 'user.tokens' to "loading". * mw.loader.state() now sets 'user' (as "loading" or "ready"). Fixes problem 1. * The <script async src> tag for 'startup' changed position (slightly). Previously it was after all inline scripts and stylesheets. It's still after all inline scripts and after most stylesheets, but before any user styles. Since the queue is now formatted outside OutputPage, it can't inject the meta-ResourceLoaderDynamicStyles tag and user-stylesheet hack in the middle of existing output. This shouldn't have any noticable impact. Bug: T87871 Change-Id: I605b8cd1e1fc009b4662a0edbc54d09dd65ee1df
2016-07-15 14:13:09 +00:00
$rlClient->setModules( $this->getModules( /*filter*/ true ) );
OutputPage: Make ResourceLoader position exemption more generic Follows-up 80e5b160e which moved queue formatting out of OutputPage into a a separate ResourceLoaderClientHtml class. The special handling for 'user' and 'user.styles' modules, and the exempt module groups was kept in OutputPage. However the handling for it was hardcoded for the modules in that group by default. It did not account for modules with a group of 'user' loaded by an extension (e.g. GlobalCssJs). GlobalCssJs modules were wrongly loaded in the regular style queue (still in a separate request group, but not in the right cascading order below the DynamicSyles marker). This commit generalises the handling previously put in buildExemptModules and moves it to getRlClient() so that it may apply to all style modules. This commit should be a no-op besides the moving of any <link rel=stylesheet> for non-core modules in group 'site' or 'user' now being one line lower in the <head> HTML (after the DynamicStyles marker). Bug: T143357 Change-Id: I1d6ea10b42293acfc535578172ad7ab2369f6299
2016-08-19 02:04:21 +00:00
$rlClient->setModuleStyles( $moduleStyles );
$rlClient->setExemptStates( $exemptStates );
resourceloader: Move queue formatting out of OutputPage HTML formatting of the queue was distributed over several OutputPage methods. Each method demanding a snippet of HTML by calling makeResourceLoaderLink() with a limited amount of information. As such, makeResourceLoaderLink() was unable to provide the client with the proper state information. Centralising it also allows it to better reduce duplication in HTML output and maintain a more accurate state. Problems fixed by centralising: 1. The 'user' module is special (due to per-user 'version' and 'user' params). It is manually requested via script-src. To avoid a separate (and wrong) request from something that requires it, we set state=loading directly. However, because the module is in the bottom, the old HTML formatter could only put state=loading in the bottom also. This sometimes caused a wrong request to be fired for modules=user if something in the top queue triggered a requirement for it. 2. Since a464d1d4 (T87871) we track states of page-style modules, with purpose of allowing dependencies on style modules without risking duplicate loading on pages where the styles are loaded already. This didn't work, because the state information about page-style modules is output near the stylesheet, which is after the script tag with mw.loader.load(). That runs first, and mw.loader would still make a duplicate request before it learns the state. Changes: * Document reasons for style/script tag order in getHeadHtml (per 09537e83). * Pass $type from getModuleStyles() to getAllowedModules(). This wasn't needed before since a duplicate check in makeResourceLoaderLink() verified the origin a second time. * Declare explicit position 'top' on 'user.options' and 'user.tokens' module. Previously, OutputPage hardcoded them in the top. The new formatter doesn't. * Remove getHeadScripts(). * Remove getInlineHeadScripts(). * Remove getExternalHeadScripts(). * Remove buildCssLinks(). * Remove getScriptsForBottomQueue(). * Change where Skin::setupSkinUserCss() is called. This methods lets the skin add modules to the queue. Previously it was called from buildCssLinks(), via headElement(), via prepareQuickTemplate(), via OutputPage::output(). It's now in OutputPage::output() directly (slightly earlier). This is needed because prepareQuickTemplate() calls bottomScripts() before headElement(). And bottomScript() would lazy-initialise the queue and lock it before setupSkinUserCss() is called from headElement(). This makes execution order more predictable instead of being dependent on the arbitrary order of data extraction in prepareQuickTemplate (which varies from one skin to another). * Compute isUserModulePreview() and isKnownEmpty() for the 'user' module early on so. This avoids wrongful loading and fixes problem 1. Effective changes in output: * mw.loader.state() is now before mw.loader.load(). This fixes problem 2. * mw.loader.state() now sets 'user.options' and 'user.tokens' to "loading". * mw.loader.state() now sets 'user' (as "loading" or "ready"). Fixes problem 1. * The <script async src> tag for 'startup' changed position (slightly). Previously it was after all inline scripts and stylesheets. It's still after all inline scripts and after most stylesheets, but before any user styles. Since the queue is now formatted outside OutputPage, it can't inject the meta-ResourceLoaderDynamicStyles tag and user-stylesheet hack in the middle of existing output. This shouldn't have any noticable impact. Bug: T87871 Change-Id: I605b8cd1e1fc009b4662a0edbc54d09dd65ee1df
2016-07-15 14:13:09 +00:00
$this->rlClient = $rlClient;
}
return $this->rlClient;
}
Changing comments layout preparing for generated documentation with Phpdocumentor
2004-09-02 23:28:24 +00:00
/**
Removed unneeded spaces and colons in @param and friends Also swapped some "$var type" to "type $var" or added missing types before the $var. Changed some other types to match the more common spelling. Makes beginning of some text in captial. Change-Id: Ic36c8c7820a6c2d603f1138130670c6bf6a1ca59
2014-04-08 15:29:17 +00:00
* @param Skin $sk The given Skin
* @param bool $includeStyle Unused
* @return string The doctype, opening "<html>", and head element.
Changing comments layout preparing for generated documentation with Phpdocumentor
2004-09-02 23:28:24 +00:00
*/
* (bug 20631) Preview of personal JavaScript and CSS pages now works again * introduced SkinTemplate::$useHeadElement as switch for backward compatibility for extension skins using the old way of generating the <head> element: ** false (default): no change from previous version ** true: <head> specific items set in SkinTemplate::outputPage() are no longer generated (avoid double execution of some functions) and the result of OutputPage::headElement() is stored in the 'headelement' item * updated all core skin to use this new method, some extensions using MonoBookTemplate but not extending SkinMonoBook (or for other core skins) will need to set $useHeadElement to true to work properly though * Made Skin::userCanPreview() public since it's needed in OutputPage::getHeadScripts() * Pass the Skin object from OutputPage::headElement() to OutputPage::getHeadScripts() rather than getting it from $wgUser
2009-09-22 16:52:40 +00:00
public function headElement( Skin $sk, $includeStyle = true ) {
OutputPage: Load html5shiv without indirection of load.php This library was introduced in 3a30e03645f, to make sure Grade C stylesheets that apply to HTML5 elements that older browsers might not know yet, work as expected in old IE. It originally committed a minified version and loaded it directly as an HTML script tag in a conditional comment. This had minimal impact on anything else, and was easy to maintain. In 68237fb1a74, this was changed to commit the unminified version instead because Debian maintainers don't like packaging software that contain minified files, despite being a simple file, unmodified from the original (upstream publishes it in minified form, with license header), published under a compatible free license, and embedded in a license-compliant manner. We then registered it as an unused ResourceLoader module, to be minified on-the-fly at run-time. Support for "server-registered client-unregistered" modules was removed last week in c554ee8e64e because nothing needed it anymore (except html5shiv apparently), which resulted in this module being registered client-side on all page views for all users (in latest master). This doesn't break anything, but it is a regression performance-wise. Restore this by (mostly) going to how it was before: A simple static file, committed to Git, served as-is. Not related to, served by, pulled through, nor registered with, ResourceLoader in any way. Only difference with the original approach is that it is no longer minified, which means a few more bytes transferred on old IE page views, which is considered an acceptable compromise. Bug: T201483 Change-Id: Ib0020b6bd015679b61f63eaa8109ed9b83d8ad15
2019-07-04 15:54:24 +00:00
$config = $this->getConfig();
Bug 29524 - Rename RequestContext::getLang to getLanguage I'll be amazed if this doens't break any tests.
2011-11-21 16:13:21 +00:00
$userdir = $this->getLanguage()->getDir();
Replace deprecated function in OutputPage Now mocking LanguageConverter Change-Id: I25f2f90d8dcf2cb09429e01873f1982dca199976
2020-11-03 17:18:15 +00:00
$services = MediaWikiServices::getInstance();
$sitedir = $services->getContentLanguage()->getDir();
Initial revision
2003-04-14 23:10:40 +00:00
OutputPage: Make use of WrappedStringList in headElement() Right now, getInlineHeadScripts(), buildCssLinks() and getExternalHeadScripts() all return WrappedString::join(). But because they don't know about each other and because they need to return strings (not arrays), headElement() has no way of merging them. WrappedStringList allows this array to be kept, whilst still being backward-compatible with code that calls these methods and assumig a string (since it will lazy-join the array if the object is treated like a string). To be used by I8b6c6a10d965e7396. Output is not changed in this commit. Merely refactoring. Change-Id: Iae08345473bd93cc0948d51b62c48aeb1ea460a3
2016-07-07 19:26:36 +00:00
$pieces = [];
OutputPage: allow additional CSS classes to be added to <html> In some cases CSS classes must be on the <html> element and not the <body>. An example is the DarkMode extension. Due to how the stacking context works, the CSS `invert()` filter must be on <html>. We can't use JavaScript to apply the styles because there would be a brief delay after page load before the JS runs, causing a flashing effect. Bug: T241925 Change-Id: I4b994e081d9582801fe9d198efd68afc46a14703
2019-05-21 20:26:53 +00:00
$htmlAttribs = Sanitizer::mergeAttributes( Sanitizer::mergeAttributes(
resourceloader: Move queue formatting out of OutputPage HTML formatting of the queue was distributed over several OutputPage methods. Each method demanding a snippet of HTML by calling makeResourceLoaderLink() with a limited amount of information. As such, makeResourceLoaderLink() was unable to provide the client with the proper state information. Centralising it also allows it to better reduce duplication in HTML output and maintain a more accurate state. Problems fixed by centralising: 1. The 'user' module is special (due to per-user 'version' and 'user' params). It is manually requested via script-src. To avoid a separate (and wrong) request from something that requires it, we set state=loading directly. However, because the module is in the bottom, the old HTML formatter could only put state=loading in the bottom also. This sometimes caused a wrong request to be fired for modules=user if something in the top queue triggered a requirement for it. 2. Since a464d1d4 (T87871) we track states of page-style modules, with purpose of allowing dependencies on style modules without risking duplicate loading on pages where the styles are loaded already. This didn't work, because the state information about page-style modules is output near the stylesheet, which is after the script tag with mw.loader.load(). That runs first, and mw.loader would still make a duplicate request before it learns the state. Changes: * Document reasons for style/script tag order in getHeadHtml (per 09537e83). * Pass $type from getModuleStyles() to getAllowedModules(). This wasn't needed before since a duplicate check in makeResourceLoaderLink() verified the origin a second time. * Declare explicit position 'top' on 'user.options' and 'user.tokens' module. Previously, OutputPage hardcoded them in the top. The new formatter doesn't. * Remove getHeadScripts(). * Remove getInlineHeadScripts(). * Remove getExternalHeadScripts(). * Remove buildCssLinks(). * Remove getScriptsForBottomQueue(). * Change where Skin::setupSkinUserCss() is called. This methods lets the skin add modules to the queue. Previously it was called from buildCssLinks(), via headElement(), via prepareQuickTemplate(), via OutputPage::output(). It's now in OutputPage::output() directly (slightly earlier). This is needed because prepareQuickTemplate() calls bottomScripts() before headElement(). And bottomScript() would lazy-initialise the queue and lock it before setupSkinUserCss() is called from headElement(). This makes execution order more predictable instead of being dependent on the arbitrary order of data extraction in prepareQuickTemplate (which varies from one skin to another). * Compute isUserModulePreview() and isKnownEmpty() for the 'user' module early on so. This avoids wrongful loading and fixes problem 1. Effective changes in output: * mw.loader.state() is now before mw.loader.load(). This fixes problem 2. * mw.loader.state() now sets 'user.options' and 'user.tokens' to "loading". * mw.loader.state() now sets 'user' (as "loading" or "ready"). Fixes problem 1. * The <script async src> tag for 'startup' changed position (slightly). Previously it was after all inline scripts and stylesheets. It's still after all inline scripts and after most stylesheets, but before any user styles. Since the queue is now formatted outside OutputPage, it can't inject the meta-ResourceLoaderDynamicStyles tag and user-stylesheet hack in the middle of existing output. This shouldn't have any noticable impact. Bug: T87871 Change-Id: I605b8cd1e1fc009b4662a0edbc54d09dd65ee1df
2016-07-15 14:13:09 +00:00
$this->getRlClient()->getDocumentAttributes(),
$sk->getHtmlElementAttributes()
OutputPage: allow additional CSS classes to be added to <html> In some cases CSS classes must be on the <html> element and not the <body>. An example is the DarkMode extension. Due to how the stacking context works, the CSS `invert()` filter must be on <html>. We can't use JavaScript to apply the styles because there would be a brief delay after page load before the JS runs, causing a flashing effect. Bug: T241925 Change-Id: I4b994e081d9582801fe9d198efd68afc46a14703
2019-05-21 20:26:53 +00:00
), [ 'class' => implode( ' ', $this->mAdditionalHtmlClasses ) ] );
resourceloader: Compile documentElement.className server-side Reduces output by not needlessly performing a change client-side for which we already know the result server-side. Bug: T231168 Change-Id: I4b8749f976d04d24f85236ddd641c7a4c7f6c23a
2019-08-25 16:24:55 +00:00
$pieces[] = Html::htmlHeader( $htmlAttribs );
resourceloader: Move queue formatting out of OutputPage HTML formatting of the queue was distributed over several OutputPage methods. Each method demanding a snippet of HTML by calling makeResourceLoaderLink() with a limited amount of information. As such, makeResourceLoaderLink() was unable to provide the client with the proper state information. Centralising it also allows it to better reduce duplication in HTML output and maintain a more accurate state. Problems fixed by centralising: 1. The 'user' module is special (due to per-user 'version' and 'user' params). It is manually requested via script-src. To avoid a separate (and wrong) request from something that requires it, we set state=loading directly. However, because the module is in the bottom, the old HTML formatter could only put state=loading in the bottom also. This sometimes caused a wrong request to be fired for modules=user if something in the top queue triggered a requirement for it. 2. Since a464d1d4 (T87871) we track states of page-style modules, with purpose of allowing dependencies on style modules without risking duplicate loading on pages where the styles are loaded already. This didn't work, because the state information about page-style modules is output near the stylesheet, which is after the script tag with mw.loader.load(). That runs first, and mw.loader would still make a duplicate request before it learns the state. Changes: * Document reasons for style/script tag order in getHeadHtml (per 09537e83). * Pass $type from getModuleStyles() to getAllowedModules(). This wasn't needed before since a duplicate check in makeResourceLoaderLink() verified the origin a second time. * Declare explicit position 'top' on 'user.options' and 'user.tokens' module. Previously, OutputPage hardcoded them in the top. The new formatter doesn't. * Remove getHeadScripts(). * Remove getInlineHeadScripts(). * Remove getExternalHeadScripts(). * Remove buildCssLinks(). * Remove getScriptsForBottomQueue(). * Change where Skin::setupSkinUserCss() is called. This methods lets the skin add modules to the queue. Previously it was called from buildCssLinks(), via headElement(), via prepareQuickTemplate(), via OutputPage::output(). It's now in OutputPage::output() directly (slightly earlier). This is needed because prepareQuickTemplate() calls bottomScripts() before headElement(). And bottomScript() would lazy-initialise the queue and lock it before setupSkinUserCss() is called from headElement(). This makes execution order more predictable instead of being dependent on the arbitrary order of data extraction in prepareQuickTemplate (which varies from one skin to another). * Compute isUserModulePreview() and isKnownEmpty() for the 'user' module early on so. This avoids wrongful loading and fixes problem 1. Effective changes in output: * mw.loader.state() is now before mw.loader.load(). This fixes problem 2. * mw.loader.state() now sets 'user.options' and 'user.tokens' to "loading". * mw.loader.state() now sets 'user' (as "loading" or "ready"). Fixes problem 1. * The <script async src> tag for 'startup' changed position (slightly). Previously it was after all inline scripts and stylesheets. It's still after all inline scripts and after most stylesheets, but before any user styles. Since the queue is now formatted outside OutputPage, it can't inject the meta-ResourceLoaderDynamicStyles tag and user-stylesheet hack in the middle of existing output. This shouldn't have any noticable impact. Bug: T87871 Change-Id: I605b8cd1e1fc009b4662a0edbc54d09dd65ee1df
2016-07-15 14:13:09 +00:00
$pieces[] = Html::openElement( 'head' );
This commit implements Josh Cogliati's "info" feature (e.g. http://en.wikipedia.org/w/wiki.phtml?title=Wikipedia&action=info), fixes the page headings for protect/unprotect, and does a lot of whitespace changes (sorry about that). I've also fixed confusing indentation somewhere.
2004-07-08 14:53:54 +00:00
Attempt at normalistion of comparison styles - empty string on left and right hand side normalised to it being on the rhs Before this change, there were (? being regex 0 or 1) "" ===? 1 '' ===? 24 "" !==? 8 '' !==? 32 == "" 14 == '' 344 != "" 9 !== "" 4 != '' 151 !== '' 85 Rhs was the much more common, and the preferred style by many developers.. (Was a similar discussion in #mediawiki recently.. After that lolbugreport i think) Where there is a string (non empty) on the lhs, and variable/method call on the rhs still need normalising
2010-01-06 19:59:42 +00:00
if ( $this->getHTMLTitle() == '' ) {
Use 'pagetitle' in content language It often happens that wiki sysops want to change it site wide, especially when they want to change the format (eg, have {{SITENAME}} removed, or replace the hyphen with a middot). Bug: 48701 Change-Id: Iaf00fca1e89ae022c348c4fa0de32b998d7921a1
2013-11-12 08:36:00 +00:00
$this->setHTMLTitle( $this->msg( 'pagetitle', $this->getPageTitle() )->inContentLanguage() );
Initial revision
2003-04-14 23:10:40 +00:00
}
Always include xmlns and xml:lang even when serving non-XHTML MIME type
2005-02-07 14:11:59 +00:00
Use name constants to access config settings. Use name constants instead of string literals in calls to Config::get and ServiceOptions::get, when referring to core configuration variables. This protects against typos and makes the decumentation and schema declaration of the config settings discoverable. This is the first batch, only touching files directly under /includes/ Change-Id: I7252e636c7c86d950d9257b33491af492c6dd5eb
2022-04-01 15:58:32 +00:00
if ( !Html::isXmlMimeType( $config->get( MainConfigNames::MimeType ) ) ) {
OutputPage: Move our <meta charset="...> so it's placed above the title It's recommended for the meta charset to be placed before the <title> since <title> contains text which is inside the character set defined by the meta charset. Use of meta charset inside XHTML also seems to be redundant, not recommended, and is very likely completely ignored. Change-Id: I335b0598a9615540dc5e917682508b4a8d32d96e
2013-05-11 09:31:30 +00:00
// Add <meta charset="UTF-8">
// This should be before <title> since it defines the charset used by
// text including the text inside <title>.
// The spec recommends defining XHTML5's charset using the XML declaration
// instead of meta.
// Our XML declaration is output by Html::htmlHeader.
Update weblinks in comments from HTTP to HTTPS Use HTTPS instead of HTTP where the HTTP link is a redirect to the HTTPS link. Also update some defect links. Change-Id: Ic3a5eac910d098ed5c2a21e9f47c9b6ee06b2643
2016-10-13 05:34:26 +00:00
// https://html.spec.whatwg.org/multipage/semantics.html#attr-meta-http-equiv-content-type
// https://html.spec.whatwg.org/multipage/semantics.html#charset
OutputPage: Make use of WrappedStringList in headElement() Right now, getInlineHeadScripts(), buildCssLinks() and getExternalHeadScripts() all return WrappedString::join(). But because they don't know about each other and because they need to return strings (not arrays), headElement() has no way of merging them. WrappedStringList allows this array to be kept, whilst still being backward-compatible with code that calls these methods and assumig a string (since it will lazy-join the array if the object is treated like a string). To be used by I8b6c6a10d965e7396. Output is not changed in this commit. Merely refactoring. Change-Id: Iae08345473bd93cc0948d51b62c48aeb1ea460a3
2016-07-07 19:26:36 +00:00
$pieces[] = Html::element( 'meta', [ 'charset' => 'UTF-8' ] );
OutputPage: Move our <meta charset="...> so it's placed above the title It's recommended for the meta charset to be placed before the <title> since <title> contains text which is inside the character set defined by the meta charset. Use of meta charset inside XHTML also seems to be redundant, not recommended, and is very likely completely ignored. Change-Id: I335b0598a9615540dc5e917682508b4a8d32d96e
2013-05-11 09:31:30 +00:00
}
Change null to [] for html attributes on call of Html functions Change-Id: Ia84f5b1f210396cba5daa7b9e8bb8c4069e0171d
2021-09-22 20:47:28 +00:00
$pieces[] = Html::element( 'title', [], $this->getHTMLTitle() );
resourceloader: Compile documentElement.className server-side Reduces output by not needlessly performing a change client-side for which we already know the result server-side. Bug: T231168 Change-Id: I4b8749f976d04d24f85236ddd641c7a4c7f6c23a
2019-08-25 16:24:55 +00:00
$pieces[] = $this->getRlClient()->getHeadHtml( $htmlAttribs['class'] ?? null );
resourceloader: Move queue formatting out of OutputPage HTML formatting of the queue was distributed over several OutputPage methods. Each method demanding a snippet of HTML by calling makeResourceLoaderLink() with a limited amount of information. As such, makeResourceLoaderLink() was unable to provide the client with the proper state information. Centralising it also allows it to better reduce duplication in HTML output and maintain a more accurate state. Problems fixed by centralising: 1. The 'user' module is special (due to per-user 'version' and 'user' params). It is manually requested via script-src. To avoid a separate (and wrong) request from something that requires it, we set state=loading directly. However, because the module is in the bottom, the old HTML formatter could only put state=loading in the bottom also. This sometimes caused a wrong request to be fired for modules=user if something in the top queue triggered a requirement for it. 2. Since a464d1d4 (T87871) we track states of page-style modules, with purpose of allowing dependencies on style modules without risking duplicate loading on pages where the styles are loaded already. This didn't work, because the state information about page-style modules is output near the stylesheet, which is after the script tag with mw.loader.load(). That runs first, and mw.loader would still make a duplicate request before it learns the state. Changes: * Document reasons for style/script tag order in getHeadHtml (per 09537e83). * Pass $type from getModuleStyles() to getAllowedModules(). This wasn't needed before since a duplicate check in makeResourceLoaderLink() verified the origin a second time. * Declare explicit position 'top' on 'user.options' and 'user.tokens' module. Previously, OutputPage hardcoded them in the top. The new formatter doesn't. * Remove getHeadScripts(). * Remove getInlineHeadScripts(). * Remove getExternalHeadScripts(). * Remove buildCssLinks(). * Remove getScriptsForBottomQueue(). * Change where Skin::setupSkinUserCss() is called. This methods lets the skin add modules to the queue. Previously it was called from buildCssLinks(), via headElement(), via prepareQuickTemplate(), via OutputPage::output(). It's now in OutputPage::output() directly (slightly earlier). This is needed because prepareQuickTemplate() calls bottomScripts() before headElement(). And bottomScript() would lazy-initialise the queue and lock it before setupSkinUserCss() is called from headElement(). This makes execution order more predictable instead of being dependent on the arbitrary order of data extraction in prepareQuickTemplate (which varies from one skin to another). * Compute isUserModulePreview() and isKnownEmpty() for the 'user' module early on so. This avoids wrongful loading and fixes problem 1. Effective changes in output: * mw.loader.state() is now before mw.loader.load(). This fixes problem 2. * mw.loader.state() now sets 'user.options' and 'user.tokens' to "loading". * mw.loader.state() now sets 'user' (as "loading" or "ready"). Fixes problem 1. * The <script async src> tag for 'startup' changed position (slightly). Previously it was after all inline scripts and stylesheets. It's still after all inline scripts and after most stylesheets, but before any user styles. Since the queue is now formatted outside OutputPage, it can't inject the meta-ResourceLoaderDynamicStyles tag and user-stylesheet hack in the middle of existing output. This shouldn't have any noticable impact. Bug: T87871 Change-Id: I605b8cd1e1fc009b4662a0edbc54d09dd65ee1df
2016-07-15 14:13:09 +00:00
$pieces[] = $this->buildExemptModules();
$pieces = array_merge( $pieces, array_values( $this->getHeadLinksArray() ) );
$pieces = array_merge( $pieces, array_values( $this->mHeadItems ) );
OutputPage: Support HTML5 elements in older browsers using html5shiv Bug: T122965 Change-Id: Ic11ca687a5459b2dc5d942cf1dc784daa0ddc2f3
2017-08-03 15:35:10 +00:00
resourceloader: Move queue formatting out of OutputPage HTML formatting of the queue was distributed over several OutputPage methods. Each method demanding a snippet of HTML by calling makeResourceLoaderLink() with a limited amount of information. As such, makeResourceLoaderLink() was unable to provide the client with the proper state information. Centralising it also allows it to better reduce duplication in HTML output and maintain a more accurate state. Problems fixed by centralising: 1. The 'user' module is special (due to per-user 'version' and 'user' params). It is manually requested via script-src. To avoid a separate (and wrong) request from something that requires it, we set state=loading directly. However, because the module is in the bottom, the old HTML formatter could only put state=loading in the bottom also. This sometimes caused a wrong request to be fired for modules=user if something in the top queue triggered a requirement for it. 2. Since a464d1d4 (T87871) we track states of page-style modules, with purpose of allowing dependencies on style modules without risking duplicate loading on pages where the styles are loaded already. This didn't work, because the state information about page-style modules is output near the stylesheet, which is after the script tag with mw.loader.load(). That runs first, and mw.loader would still make a duplicate request before it learns the state. Changes: * Document reasons for style/script tag order in getHeadHtml (per 09537e83). * Pass $type from getModuleStyles() to getAllowedModules(). This wasn't needed before since a duplicate check in makeResourceLoaderLink() verified the origin a second time. * Declare explicit position 'top' on 'user.options' and 'user.tokens' module. Previously, OutputPage hardcoded them in the top. The new formatter doesn't. * Remove getHeadScripts(). * Remove getInlineHeadScripts(). * Remove getExternalHeadScripts(). * Remove buildCssLinks(). * Remove getScriptsForBottomQueue(). * Change where Skin::setupSkinUserCss() is called. This methods lets the skin add modules to the queue. Previously it was called from buildCssLinks(), via headElement(), via prepareQuickTemplate(), via OutputPage::output(). It's now in OutputPage::output() directly (slightly earlier). This is needed because prepareQuickTemplate() calls bottomScripts() before headElement(). And bottomScript() would lazy-initialise the queue and lock it before setupSkinUserCss() is called from headElement(). This makes execution order more predictable instead of being dependent on the arbitrary order of data extraction in prepareQuickTemplate (which varies from one skin to another). * Compute isUserModulePreview() and isKnownEmpty() for the 'user' module early on so. This avoids wrongful loading and fixes problem 1. Effective changes in output: * mw.loader.state() is now before mw.loader.load(). This fixes problem 2. * mw.loader.state() now sets 'user.options' and 'user.tokens' to "loading". * mw.loader.state() now sets 'user' (as "loading" or "ready"). Fixes problem 1. * The <script async src> tag for 'startup' changed position (slightly). Previously it was after all inline scripts and stylesheets. It's still after all inline scripts and after most stylesheets, but before any user styles. Since the queue is now formatted outside OutputPage, it can't inject the meta-ResourceLoaderDynamicStyles tag and user-stylesheet hack in the middle of existing output. This shouldn't have any noticable impact. Bug: T87871 Change-Id: I605b8cd1e1fc009b4662a0edbc54d09dd65ee1df
2016-07-15 14:13:09 +00:00
$pieces[] = Html::closeElement( 'head' );
Merge all skins' output of opening <body> tag This fixes a few minor discrepancies, like Vector outputting dir="" (redundant to the one on <html>), and non-Monobook-based skins omitting the capitalize-all-nouns class (!). This adds Html::openElement() and refactors Html::rawElement() accordingly, so I checked that all parser tests still pass. I wasn't able to figure out if I broke some feature of right-floating quickbars in the Standard skin, because I wasn't able to figure out what the feature was in the first place. Hopefully either it works, or nobody cares, or someone else will figure out what it was supposed to do. (This is the stuff in getBodyOptions() in Standard.php I deleted; I'm not sure the addition to sticky.js does what I want.)
2010-01-15 01:16:52 +00:00
Skins can define bodyClasses at skin registration Bug: T298734 Change-Id: I55fbe6c8d90fb2c13a5661c4f328f26fa48f1245
2022-01-12 21:21:31 +00:00
$skinOptions = $sk->getOptions();
$bodyClasses = array_merge( $this->mAdditionalBodyClasses, $skinOptions['bodyClasses'] );
Clean up the class building logic in OutputPage Build the classes using an array that is finally imploded, instead of concatenating strings repeatedly. Change-Id: I2a09282d5ba33f131a4311c58e49dab5eefce418
2013-04-22 13:57:44 +00:00
$bodyClasses[] = 'mediawiki';
Merge all skins' output of opening <body> tag This fixes a few minor discrepancies, like Vector outputting dir="" (redundant to the one on <html>), and non-Monobook-based skins omitting the capitalize-all-nouns class (!). This adds Html::openElement() and refactors Html::rawElement() accordingly, so I checked that all parser tests still pass. I wasn't able to figure out if I broke some feature of right-floating quickbars in the Standard skin, because I wasn't able to figure out what the feature was in the first place. Hopefully either it works, or nobody cares, or someone else will figure out what it was supposed to do. (This is the stuff in getBodyOptions() in Standard.php I deleted; I'm not sure the addition to sticky.js does what I want.)
2010-01-15 01:16:52 +00:00
(bug 6100; follow-up to r91315) Being bold and removing $wgBetterDirectionality (and dependent wfUILang) in core, as most or all work is finished. Also: * Introduce classes mw-float-end, mw-float-start so we don't have to use inline css depending on wfUILang()/$wgLang (see HistoryPage and SpecialFileDuplicateSearch) * Add direction mark to protection log * Remove specialpageattributes as it is obsoleted by this commit (also fixes bug 28572) * Add two direction marks in wfSpecialList, which makes ltr links on rtl wiki (and vice versa) display nicely as well (only on those special pages however) * Revert r91340 partially: use mw-content-ltr/rtl class anyway in shared.css. Both ways have their [dis]advantages... * Set the direction of input fields by default to the content language direction (except buttons etc.) in shared.css
2011-07-06 02:26:06 +00:00
# Classes for LTR/RTL directionality support
Clean up the class building logic in OutputPage Build the classes using an array that is finally imploded, instead of concatenating strings repeatedly. Change-Id: I2a09282d5ba33f131a4311c58e49dab5eefce418
2013-04-22 13:57:44 +00:00
$bodyClasses[] = $userdir;
$bodyClasses[] = "sitedir-$sitedir";
Merge all skins' output of opening <body> tag This fixes a few minor discrepancies, like Vector outputting dir="" (redundant to the one on <html>), and non-Monobook-based skins omitting the capitalize-all-nouns class (!). This adds Html::openElement() and refactors Html::rawElement() accordingly, so I checked that all parser tests still pass. I wasn't able to figure out if I broke some feature of right-floating quickbars in the Standard skin, because I wasn't able to figure out what the feature was in the first place. Hopefully either it works, or nobody cares, or someone else will figure out what it was supposed to do. (This is the stuff in getBodyOptions() in Standard.php I deleted; I'm not sure the addition to sticky.js does what I want.)
2010-01-15 01:16:52 +00:00
Replace deprecated function in OutputPage Now mocking LanguageConverter Change-Id: I25f2f90d8dcf2cb09429e01873f1982dca199976
2020-11-03 17:18:15 +00:00
$underline = $services->getUserOptionsLookup()->getOption( $this->getUser(), 'underline' );
Move link-underline user preference rule to body class ResourceLoaderUserCSSPrefsModule is now empty and can be removed. Change-Id: I806e117131765b859e990c14735bb9494db9a37e
2016-09-07 22:51:48 +00:00
if ( $underline < 2 ) {
// The following classes can be used here:
// * mw-underline-always
// * mw-underline-never
$bodyClasses[] = 'mw-underline-' . ( $underline ? 'always' : 'never' );
}
Hide marked empty elements by default (stage 1) We originally imagined rolling out the display of empty elements simultaneously with the Html5Depurate, but now we have added support for marking empty elements to Html5Depurate and plan on having some sort of longer migration period. So, move the relevant CSS to content.css, and remove the concept of CSS dependant on tidy driver. Add a body class which will allow the effect to be toggled in a gadget or extension. Actual toggling in the CSS will be in the stage 2 patch, to be deployed after the varnish cache and parser cache have expired. I originally imagined that there would be a gadget that overrides the rule with an !important selector, but that method does not allow you to recover the original display property, which is often overridden by the style attribute or site CSS to be "inline". Also, in RaggettWrapper, switch to the new class mw-empty-elt, following Html5Depurate, instead of mw-empty-li. The old class will be removed in the stage 2 patch. Change-Id: Ic0f432c43a006629ca5a1a7c2dda3552ceb4dc4f
2016-07-13 02:01:59 +00:00
// Parser feature migration class
// The idea is that this will eventually be removed, after the wikitext
// which requires it is cleaned up.
$bodyClasses[] = 'mw-hide-empty-elt';
Clean up the class building logic in OutputPage Build the classes using an array that is finally imploded, instead of concatenating strings repeatedly. Change-Id: I2a09282d5ba33f131a4311c58e49dab5eefce418
2013-04-22 13:57:44 +00:00
$bodyClasses[] = $sk->getPageClasses( $this->getTitle() );
$bodyClasses[] = 'skin-' . Sanitizer::escapeClass( $sk->getSkinName() );
Pass php-strict for OutputPage Change-Id: Ia5d03277fff543ca672dc8efc7eaaeca4888ac23
2014-05-11 01:54:15 +00:00
$bodyClasses[] =
context: Add a cached RequestContext::getActionName method This method is dependent on, and inherently must depend on, all of Title, WikiPage, and WebRequest. And, like Title and WikiPage, which also have getters in RequestContext, Action is also derived from a query parameter that is widely recognised in almost all web requests to index.php. The status quo in core and extensions, is to obtain this value via Action::getActionName(), which as a static method that bypasses dependency injection and also has the problem of not being cached. Caching it within ActionFactory seems hard and awkward, due to varying by context. In change I61d66211bd (22f9a32853f) a cached wrapper method was added internally to the Skin class. In change I8cbc4bba4d248d9 (235820d631) another cached wrapper was added in the Gadgets extension. This change takes this approach further by making it a stable public method on RequestContext. To facilitate testing and to offer basic confidence in this working correctly, this commit also adopts the new method in two place that are considered "safe" (Skin, and OutputPage). Both of these are called relatively late in the PHP proccess and well after any Setup code and overrides (such as in MediaWiki.php), during which it is more complex to call this. I'll audit and update those in a subsequent change. Change-Id: I1e259b54dca48a32be5a8c6cbb8eb69aec2da115
2021-11-06 00:28:23 +00:00
'action-' . Sanitizer::escapeClass( $this->getContext()->getActionName() );
Clean up the class building logic in OutputPage Build the classes using an array that is finally imploded, instead of concatenating strings repeatedly. Change-Id: I2a09282d5ba33f131a4311c58e49dab5eefce418
2013-04-22 13:57:44 +00:00
Skin: Add a skin option for responsiveness Skins can now specify responsive in the skin option, which will add a meta tag with the value width=device-width, initial-scale=1.0, user-scalable=yes, minimum-scale=0.25, maximum-scale=5.0. If responsiveness is not constant, but based on, for example, a user preference, then skins can override isResponsive to indicate when the skin is responsive. When enabled, it will also add the skin--responsive class to <body>. Bug: T258290 Change-Id: I6e69b482cdbc49c09adc6f7cee54f5e17b7c92c4
2020-08-03 18:54:44 +00:00
if ( $sk->isResponsive() ) {
$bodyClasses[] = 'skin--responsive';
}
Convert all array() syntax to [] Per wikitech-l consensus: https://lists.wikimedia.org/pipermail/wikitech-l/2016-February/084821.html Notes: * Disabled CallTimePassByReference due to false positives (T127163) Change-Id: I2c8ce713ce6600a0bb7bf67537c87044c7a45c4b
2016-02-17 09:09:32 +00:00
$bodyAttrs = [];
Clean up the class building logic in OutputPage Build the classes using an array that is finally imploded, instead of concatenating strings repeatedly. Change-Id: I2a09282d5ba33f131a4311c58e49dab5eefce418
2013-04-22 13:57:44 +00:00
// While the implode() is not strictly needed, it's used for backwards compatibility
// (this used to be built as a string and hooks likely still expect that).
$bodyAttrs['class'] = implode( ' ', $bodyClasses );
// Allow skins and extensions to add body attributes they need
Deprecate using Skin method to add body attributes. - Remove the parent method. There's no skin that's relying on this method directly and since it's no-op declaration, removing it is also no-op. - Raise deprecation warning if the body attributes are modified with the method (as implemented by the relevant skin that made call to the OutputPage) - The recommended way to add body attributes now is to use the OutputPageBodyAttributes hook Bug: T255698 Depends-On: Idab13c37002ac5663f292e4f4b4ff51eb7573fda Change-Id: Ibc56bf40c1ea70b2d7f7860ed85fbeb7b5d2bbf5
2020-06-19 08:09:21 +00:00
// Get ones from deprecated method
if ( method_exists( $sk, 'addToBodyAttributes' ) ) {
/** @phan-suppress-next-line PhanUndeclaredMethod */
$sk->addToBodyAttributes( $this, $bodyAttrs );
wfDeprecated( 'Skin::addToBodyAttributes method to add body attributes', '1.35' );
}
// Then run the hook, the recommended way of adding body attributes now
Hooks::run() call site migration Migrate all callers of Hooks::run() to use the new HookContainer/HookRunner system. General principles: * Use DI if it is already used. We're not changing the way state is managed in this patch. * HookContainer is always injected, not HookRunner. HookContainer is a service, it's a more generic interface, it is the only thing that provides isRegistered() which is needed in some cases, and a HookRunner can be efficiently constructed from it (confirmed by benchmark). Because HookContainer is needed for object construction, it is also needed by all factories. * "Ask your friendly local base class". Big hierarchies like SpecialPage and ApiBase have getHookContainer() and getHookRunner() methods in the base class, and classes that extend that base class are not expected to know or care where the base class gets its HookContainer from. * ProtectedHookAccessorTrait provides protected getHookContainer() and getHookRunner() methods, getting them from the global service container. The point of this is to ease migration to DI by ensuring that call sites ask their local friendly base class rather than getting a HookRunner from the service container directly. * Private $this->hookRunner. In some smaller classes where accessor methods did not seem warranted, there is a private HookRunner property which is accessed directly. Very rarely (two cases), there is a protected property, for consistency with code that conventionally assumes protected=private, but in cases where the class might actually be overridden, a protected accessor is preferred over a protected property. * The last resort: Hooks::runner(). Mostly for static, file-scope and global code. In a few cases it was used for objects with broken construction schemes, out of horror or laziness. Constructors with new required arguments: * AuthManager * BadFileLookup * BlockManager * ClassicInterwikiLookup * ContentHandlerFactory * ContentSecurityPolicy * DefaultOptionsManager * DerivedPageDataUpdater * FullSearchResultWidget * HtmlCacheUpdater * LanguageFactory * LanguageNameUtils * LinkRenderer * LinkRendererFactory * LocalisationCache * MagicWordFactory * MessageCache * NamespaceInfo * PageEditStash * PageHandlerFactory * PageUpdater * ParserFactory * PermissionManager * RevisionStore * RevisionStoreFactory * SearchEngineConfig * SearchEngineFactory * SearchFormWidget * SearchNearMatcher * SessionBackend * SpecialPageFactory * UserNameUtils * UserOptionsManager * WatchedItemQueryService * WatchedItemStore Constructors with new optional arguments: * DefaultPreferencesFactory * Language * LinkHolderArray * MovePage * Parser * ParserCache * PasswordReset * Router setHookContainer() now required after construction: * AuthenticationProvider * ResourceLoaderModule * SearchEngine Change-Id: Id442b0dbe43aba84bd5cf801d86dedc768b082c7
2020-03-19 02:42:09 +00:00
$this->getHookRunner()->onOutputPageBodyAttributes( $this, $sk, $bodyAttrs );
Add to OutputPage::headElement a OutputPageBodyAttrs hook for extensions and a bodyAttrs callback to Skin so that Extensions and Skins are able to add attributes and classes to the <body> tag without neeing to do insane things.
2010-10-03 14:12:41 +00:00
OutputPage: Make use of WrappedStringList in headElement() Right now, getInlineHeadScripts(), buildCssLinks() and getExternalHeadScripts() all return WrappedString::join(). But because they don't know about each other and because they need to return strings (not arrays), headElement() has no way of merging them. WrappedStringList allows this array to be kept, whilst still being backward-compatible with code that calls these methods and assumig a string (since it will lazy-join the array if the object is treated like a string). To be used by I8b6c6a10d965e7396. Output is not changed in this commit. Merely refactoring. Change-Id: Iae08345473bd93cc0948d51b62c48aeb1ea460a3
2016-07-07 19:26:36 +00:00
$pieces[] = Html::openElement( 'body', $bodyAttrs );
Merge all skins' output of opening <body> tag This fixes a few minor discrepancies, like Vector outputting dir="" (redundant to the one on <html>), and non-Monobook-based skins omitting the capitalize-all-nouns class (!). This adds Html::openElement() and refactors Html::rawElement() accordingly, so I checked that all parser tests still pass. I wasn't able to figure out if I broke some feature of right-floating quickbars in the Standard skin, because I wasn't able to figure out what the feature was in the first place. Hopefully either it works, or nobody cares, or someone else will figure out what it was supposed to do. (This is the stuff in getBodyOptions() in Standard.php I deleted; I'm not sure the addition to sticky.js does what I want.)
2010-01-15 01:16:52 +00:00
resourceloader: Move queue formatting out of OutputPage HTML formatting of the queue was distributed over several OutputPage methods. Each method demanding a snippet of HTML by calling makeResourceLoaderLink() with a limited amount of information. As such, makeResourceLoaderLink() was unable to provide the client with the proper state information. Centralising it also allows it to better reduce duplication in HTML output and maintain a more accurate state. Problems fixed by centralising: 1. The 'user' module is special (due to per-user 'version' and 'user' params). It is manually requested via script-src. To avoid a separate (and wrong) request from something that requires it, we set state=loading directly. However, because the module is in the bottom, the old HTML formatter could only put state=loading in the bottom also. This sometimes caused a wrong request to be fired for modules=user if something in the top queue triggered a requirement for it. 2. Since a464d1d4 (T87871) we track states of page-style modules, with purpose of allowing dependencies on style modules without risking duplicate loading on pages where the styles are loaded already. This didn't work, because the state information about page-style modules is output near the stylesheet, which is after the script tag with mw.loader.load(). That runs first, and mw.loader would still make a duplicate request before it learns the state. Changes: * Document reasons for style/script tag order in getHeadHtml (per 09537e83). * Pass $type from getModuleStyles() to getAllowedModules(). This wasn't needed before since a duplicate check in makeResourceLoaderLink() verified the origin a second time. * Declare explicit position 'top' on 'user.options' and 'user.tokens' module. Previously, OutputPage hardcoded them in the top. The new formatter doesn't. * Remove getHeadScripts(). * Remove getInlineHeadScripts(). * Remove getExternalHeadScripts(). * Remove buildCssLinks(). * Remove getScriptsForBottomQueue(). * Change where Skin::setupSkinUserCss() is called. This methods lets the skin add modules to the queue. Previously it was called from buildCssLinks(), via headElement(), via prepareQuickTemplate(), via OutputPage::output(). It's now in OutputPage::output() directly (slightly earlier). This is needed because prepareQuickTemplate() calls bottomScripts() before headElement(). And bottomScript() would lazy-initialise the queue and lock it before setupSkinUserCss() is called from headElement(). This makes execution order more predictable instead of being dependent on the arbitrary order of data extraction in prepareQuickTemplate (which varies from one skin to another). * Compute isUserModulePreview() and isKnownEmpty() for the 'user' module early on so. This avoids wrongful loading and fixes problem 1. Effective changes in output: * mw.loader.state() is now before mw.loader.load(). This fixes problem 2. * mw.loader.state() now sets 'user.options' and 'user.tokens' to "loading". * mw.loader.state() now sets 'user' (as "loading" or "ready"). Fixes problem 1. * The <script async src> tag for 'startup' changed position (slightly). Previously it was after all inline scripts and stylesheets. It's still after all inline scripts and after most stylesheets, but before any user styles. Since the queue is now formatted outside OutputPage, it can't inject the meta-ResourceLoaderDynamicStyles tag and user-stylesheet hack in the middle of existing output. This shouldn't have any noticable impact. Bug: T87871 Change-Id: I605b8cd1e1fc009b4662a0edbc54d09dd65ee1df
2016-07-15 14:13:09 +00:00
return self::combineWrappedStrings( $pieces );
Include the meta-robots, meta-keywords, and link-RSS tags, and the set MIME type, in MonoBook output.
2004-04-09 04:53:52 +00:00
}
Removed trailing whitespaces
2010-10-02 13:45:35 +00:00
* Made Resources.php return a pure-data array instead of an ugly mix of data and code. This allows the class code to be lazy-loaded with the autoloader, for a performance advantage especially on non-APC installs. And using the convention where if the class is omitted, ResourceLoaderFileModule is assumed, the registration code becomes shorter and simpler. * Modified ResourceLoader to lazy-initialise module objects, for a further performance advantage. * Deleted ResourceLoader::getModules(), provided getModuleNames() instead. Although the startup module needs this functionality, it's slow to generate, so to avoid misuse, it's better to provide a foolproof fast interface and let the startup module do the slow thing itself. * Modified ResourceLoader::register() to optionally accept an info array instead of an object. * Added $wgResourceModules, allowing extensions to efficiently define their own resource loader modules. The trouble with hooks is that they contain code, and code is slow. We've been through all this before with i18n. Hooks are useful as a performance tool only if you call them very rarely. * Moved ResourceLoader settings to their own section in DefaultSettings.php * Added options to ResourceLoaderFileModule equivalent to the $localBasePath and $remoteBasePath parameters, to allow it to be instantiated via the new array style. Also added remoteExtPath, which allows modules to be registered before $wgExtensionAssetsPath is known. * Added OutputPage::getResourceLoader(), mostly for debugging. * The time saving at the moment is about 5ms per request with no extensions, which is significant already with 6 load.php requests for a cold cache page view. This is a much more scalable interface; the relative saving will grow as more extensions are added which use this interface, especially for non-APC installs. Although the interface is backwards compatible, extension updates will follow in a subsequent commit.
2010-11-19 10:41:06 +00:00
/**
* Get a ResourceLoader object associated with this OutputPage
Fixup and add another return type
2011-02-21 17:12:33 +00:00
*
* @return ResourceLoader
* Made Resources.php return a pure-data array instead of an ugly mix of data and code. This allows the class code to be lazy-loaded with the autoloader, for a performance advantage especially on non-APC installs. And using the convention where if the class is omitted, ResourceLoaderFileModule is assumed, the registration code becomes shorter and simpler. * Modified ResourceLoader to lazy-initialise module objects, for a further performance advantage. * Deleted ResourceLoader::getModules(), provided getModuleNames() instead. Although the startup module needs this functionality, it's slow to generate, so to avoid misuse, it's better to provide a foolproof fast interface and let the startup module do the slow thing itself. * Modified ResourceLoader::register() to optionally accept an info array instead of an object. * Added $wgResourceModules, allowing extensions to efficiently define their own resource loader modules. The trouble with hooks is that they contain code, and code is slow. We've been through all this before with i18n. Hooks are useful as a performance tool only if you call them very rarely. * Moved ResourceLoader settings to their own section in DefaultSettings.php * Added options to ResourceLoaderFileModule equivalent to the $localBasePath and $remoteBasePath parameters, to allow it to be instantiated via the new array style. Also added remoteExtPath, which allows modules to be registered before $wgExtensionAssetsPath is known. * Added OutputPage::getResourceLoader(), mostly for debugging. * The time saving at the moment is about 5ms per request with no extensions, which is significant already with 6 load.php requests for a cold cache page view. This is a much more scalable interface; the relative saving will grow as more extensions are added which use this interface, especially for non-APC installs. Although the interface is backwards compatible, extension updates will follow in a subsequent commit.
2010-11-19 10:41:06 +00:00
*/
public function getResourceLoader() {
Coding style: Auto-fix MediaWiki.Usage.IsNull.IsNull Change-Id: I90cfe8366c0245c9c67e598d17800684897a4e27
2020-01-09 23:48:34 +00:00
if ( $this->mResourceLoader === null ) {
resourceloader: Instantiate main class via ServiceWiring It also removes some code duplication which is nice. This unlocks various future changes, including: * Making the `$config` parameter mandatory for the ResourceLoader class constructor, which currently falls back to global state. This should be deprecated and removed. * Making it possible to instantiate the ResourceLoader class without all the default MW modules being registered from global state. E.g. move MW module registration from main class constructor to ServiceWiring, and remove the 'EmptyResourceLoader' class hack from unit tests, and use regular 'new ResourceLoader' instead. * Making ResourceLoader a standalone library (some day), e.g. allowing it to be instantiated from a basic PHP script, in a way that is still useful and perhaps able to serve (most) RL modules without MW itself. Bug: T32956 Change-Id: I4939f296c705b268e9cf8de635e923a739410470
2019-02-16 23:16:09 +00:00
// Lazy-initialise as needed
$this->mResourceLoader = MediaWikiServices::getInstance()->getResourceLoader();
* Made Resources.php return a pure-data array instead of an ugly mix of data and code. This allows the class code to be lazy-loaded with the autoloader, for a performance advantage especially on non-APC installs. And using the convention where if the class is omitted, ResourceLoaderFileModule is assumed, the registration code becomes shorter and simpler. * Modified ResourceLoader to lazy-initialise module objects, for a further performance advantage. * Deleted ResourceLoader::getModules(), provided getModuleNames() instead. Although the startup module needs this functionality, it's slow to generate, so to avoid misuse, it's better to provide a foolproof fast interface and let the startup module do the slow thing itself. * Modified ResourceLoader::register() to optionally accept an info array instead of an object. * Added $wgResourceModules, allowing extensions to efficiently define their own resource loader modules. The trouble with hooks is that they contain code, and code is slow. We've been through all this before with i18n. Hooks are useful as a performance tool only if you call them very rarely. * Moved ResourceLoader settings to their own section in DefaultSettings.php * Added options to ResourceLoaderFileModule equivalent to the $localBasePath and $remoteBasePath parameters, to allow it to be instantiated via the new array style. Also added remoteExtPath, which allows modules to be registered before $wgExtensionAssetsPath is known. * Added OutputPage::getResourceLoader(), mostly for debugging. * The time saving at the moment is about 5ms per request with no extensions, which is significant already with 6 load.php requests for a cold cache page view. This is a much more scalable interface; the relative saving will grow as more extensions are added which use this interface, especially for non-APC installs. Although the interface is backwards compatible, extension updates will follow in a subsequent commit.
2010-11-19 10:41:06 +00:00
}
return $this->mResourceLoader;
random w/s cleanup
2011-04-23 19:28:35 +00:00
}
* Made Resources.php return a pure-data array instead of an ugly mix of data and code. This allows the class code to be lazy-loaded with the autoloader, for a performance advantage especially on non-APC installs. And using the convention where if the class is omitted, ResourceLoaderFileModule is assumed, the registration code becomes shorter and simpler. * Modified ResourceLoader to lazy-initialise module objects, for a further performance advantage. * Deleted ResourceLoader::getModules(), provided getModuleNames() instead. Although the startup module needs this functionality, it's slow to generate, so to avoid misuse, it's better to provide a foolproof fast interface and let the startup module do the slow thing itself. * Modified ResourceLoader::register() to optionally accept an info array instead of an object. * Added $wgResourceModules, allowing extensions to efficiently define their own resource loader modules. The trouble with hooks is that they contain code, and code is slow. We've been through all this before with i18n. Hooks are useful as a performance tool only if you call them very rarely. * Moved ResourceLoader settings to their own section in DefaultSettings.php * Added options to ResourceLoaderFileModule equivalent to the $localBasePath and $remoteBasePath parameters, to allow it to be instantiated via the new array style. Also added remoteExtPath, which allows modules to be registered before $wgExtensionAssetsPath is known. * Added OutputPage::getResourceLoader(), mostly for debugging. * The time saving at the moment is about 5ms per request with no extensions, which is significant already with 6 load.php requests for a cold cache page view. This is a much more scalable interface; the relative saving will grow as more extensions are added which use this interface, especially for non-APC installs. Although the interface is backwards compatible, extension updates will follow in a subsequent commit.
2010-11-19 10:41:06 +00:00
Remove checks for debug mode and module count on every makeResourceLoaderLink() invocation. Moved inside makeResourceLoaderLink() The only difference would be having now 'user' and 'user.options' modules split on debug mode.
2010-11-06 18:52:32 +00:00
/**
Fix typos in comments (N-R) Change-Id: I2d1bdb7531ff5126114a391550c2615ea6e244b3
2022-01-09 17:44:44 +00:00
* Explicitly load or embed modules on a page.
resourceloader: Async all the way Page startup: * Due to the startup module and top queue being asynchronous now, move client-nojs/client-js class handling to OutputPage to ensure there is no flashes of wrongly styled or unstyled content. To preserve compatibility for unsupported browsers, undo the class swap at runtime after the isCompatible() check. ResourceLoader startup module: * Load the startup module with <script async>. * Use DOM methods instead of 'document.write' to create base module request (jquery|mediawiki). mw.loader: * Drop 'async' parameter from mw.loader.load(). * Remove the now-unused code paths for synchronous requests. OutputPage: * Drop '$loadCall' parameter from makeResourceLoaderLink(). Asynchronous is now the default and only way to load JavaScript. This means the 'user' module "conditional document-write scripts" are now a simple "mw.loader.load( url )" call. * Fix incorrect @return of makeResourceLoaderLink(). This returns an array not a string. * Improve documentation of makeResourceLoaderLink(). * Drop '$inHead' parameter from getScriptsForBottomQueue(). No longer used. Compatibility with the $wgResourceLoaderExperimentalAsyncLoading feature is maintained. It just no longer needs to change the way the queue works since it's always asynchronous. The feature flag now only controls whether the bottom queue starts at the bottom or starts at the top. * Remove jQuery.ready() optimisation. This was mostly there to avoid the setTimeout() loop jQuery does to detect dom-ready in IE6/IE7 (which we no longer serve JavaScript at all). And for a bug in Firefox with document.write (which is no longer used as of this commit). Bug: T107399 Change-Id: Icba6d7a87b239bf127a221bc6bc432cfa71a4a72
2015-07-28 02:46:00 +00:00
*
resourceloader: Refactor module links output Changes: * Removed hardcoded logic in OutputPage regarding modules being "enabled". Previously we would always output state=loading and use $wgAllowUserJs (and others) to decide whether to output state=ready or makeResourceLoaderLink. Now, we no longer unconditionally output state=loading and simply always call makeResourceLoaderLink. That method takes care of checking whether modules are enabled and non-empty and returns state=ready when that is the case. This cleans up cases where the duplicated and incomplete logic in OutputPage thought the module was non-empty but turned out to be empty and thus would output both state=loading and later state=ready for the same module. * Clean up documentation for makeResourceLoaderLink (inconsistent ordering of type hint and $var, and @return was missing the fact that the returned html can also contain <link>). * makeResourceLoaderLink now returns an array of html and module states. This allows the consumer of this method to combine the states in 1 larger script tag on top of multiple makeResourceLoaderLink calls (e.g. one state script followed by multiple <script src=load.php>). This isn't to reduce html output, but to make sure we inform mw.loader about modules before the <script src=load.php>. If we were to mix/alternate the state script and load.php requests (which are blocking in html), it is possible for those scripts to request other modules. We need to prevent duplicate loading of modules we already know are going to be requested by the HTML output futher down. * Removed spurious new line. Example of change in HTML output: * The output has been reduced from: - loader.state( site: loading, user: loading, user.groups: loading ) - loader.load( .. ) - <script src="load.php?modules=site .."> - loader.state( user: ready, user.groups: ready ) to: - loader.state( site: loading, user: ready, user.groups: ready ) - loader.load( .. ) - <script src="load.php?modules=site .."> Change-Id: I91754ce5fae3d05b4bfa7372372eba81ee2fc579
2013-11-14 16:54:19 +00:00
* @param array|string $modules One or more module names
ResourceLoader namespace (attempt 2) Move ResourceLoader classes to their own namespace. Strip the "ResourceLoader" prefix from all except ResourceLoader itself. Move the tests by analogy. I used a namespace alias "RL" in some callers since RL\Module is less ambiguous at the call site than just "Module". I did not address DependencyStore which continues to have a non-standard location and namespace. Revert of a241d83e0a6dabedf. Bug: T308718 Change-Id: Id08a220e1d6085e2b33f3f6c9d0e3935a4204659
2022-05-06 09:09:56 +00:00
* @param string $only RL\Module TYPE_ class constant
resourceloader: Async all the way Page startup: * Due to the startup module and top queue being asynchronous now, move client-nojs/client-js class handling to OutputPage to ensure there is no flashes of wrongly styled or unstyled content. To preserve compatibility for unsupported browsers, undo the class swap at runtime after the isCompatible() check. ResourceLoader startup module: * Load the startup module with <script async>. * Use DOM methods instead of 'document.write' to create base module request (jquery|mediawiki). mw.loader: * Drop 'async' parameter from mw.loader.load(). * Remove the now-unused code paths for synchronous requests. OutputPage: * Drop '$loadCall' parameter from makeResourceLoaderLink(). Asynchronous is now the default and only way to load JavaScript. This means the 'user' module "conditional document-write scripts" are now a simple "mw.loader.load( url )" call. * Fix incorrect @return of makeResourceLoaderLink(). This returns an array not a string. * Improve documentation of makeResourceLoaderLink(). * Drop '$inHead' parameter from getScriptsForBottomQueue(). No longer used. Compatibility with the $wgResourceLoaderExperimentalAsyncLoading feature is maintained. It just no longer needs to change the way the queue works since it's always asynchronous. The feature flag now only controls whether the bottom queue starts at the bottom or starts at the top. * Remove jQuery.ready() optimisation. This was mostly there to avoid the setTimeout() loop jQuery does to detect dom-ready in IE6/IE7 (which we no longer serve JavaScript at all). And for a bug in Firefox with document.write (which is no longer used as of this commit). Bug: T107399 Change-Id: Icba6d7a87b239bf127a221bc6bc432cfa71a4a72
2015-07-28 02:46:00 +00:00
* @param array $extraQuery [optional] Array with extra query parameters for the request
resourceloader: Move queue formatting out of OutputPage HTML formatting of the queue was distributed over several OutputPage methods. Each method demanding a snippet of HTML by calling makeResourceLoaderLink() with a limited amount of information. As such, makeResourceLoaderLink() was unable to provide the client with the proper state information. Centralising it also allows it to better reduce duplication in HTML output and maintain a more accurate state. Problems fixed by centralising: 1. The 'user' module is special (due to per-user 'version' and 'user' params). It is manually requested via script-src. To avoid a separate (and wrong) request from something that requires it, we set state=loading directly. However, because the module is in the bottom, the old HTML formatter could only put state=loading in the bottom also. This sometimes caused a wrong request to be fired for modules=user if something in the top queue triggered a requirement for it. 2. Since a464d1d4 (T87871) we track states of page-style modules, with purpose of allowing dependencies on style modules without risking duplicate loading on pages where the styles are loaded already. This didn't work, because the state information about page-style modules is output near the stylesheet, which is after the script tag with mw.loader.load(). That runs first, and mw.loader would still make a duplicate request before it learns the state. Changes: * Document reasons for style/script tag order in getHeadHtml (per 09537e83). * Pass $type from getModuleStyles() to getAllowedModules(). This wasn't needed before since a duplicate check in makeResourceLoaderLink() verified the origin a second time. * Declare explicit position 'top' on 'user.options' and 'user.tokens' module. Previously, OutputPage hardcoded them in the top. The new formatter doesn't. * Remove getHeadScripts(). * Remove getInlineHeadScripts(). * Remove getExternalHeadScripts(). * Remove buildCssLinks(). * Remove getScriptsForBottomQueue(). * Change where Skin::setupSkinUserCss() is called. This methods lets the skin add modules to the queue. Previously it was called from buildCssLinks(), via headElement(), via prepareQuickTemplate(), via OutputPage::output(). It's now in OutputPage::output() directly (slightly earlier). This is needed because prepareQuickTemplate() calls bottomScripts() before headElement(). And bottomScript() would lazy-initialise the queue and lock it before setupSkinUserCss() is called from headElement(). This makes execution order more predictable instead of being dependent on the arbitrary order of data extraction in prepareQuickTemplate (which varies from one skin to another). * Compute isUserModulePreview() and isKnownEmpty() for the 'user' module early on so. This avoids wrongful loading and fixes problem 1. Effective changes in output: * mw.loader.state() is now before mw.loader.load(). This fixes problem 2. * mw.loader.state() now sets 'user.options' and 'user.tokens' to "loading". * mw.loader.state() now sets 'user' (as "loading" or "ready"). Fixes problem 1. * The <script async src> tag for 'startup' changed position (slightly). Previously it was after all inline scripts and stylesheets. It's still after all inline scripts and after most stylesheets, but before any user styles. Since the queue is now formatted outside OutputPage, it can't inject the meta-ResourceLoaderDynamicStyles tag and user-stylesheet hack in the middle of existing output. This shouldn't have any noticable impact. Bug: T87871 Change-Id: I605b8cd1e1fc009b4662a0edbc54d09dd65ee1df
2016-07-15 14:13:09 +00:00
* @return string|WrappedStringList HTML
resourceloader: Async all the way Page startup: * Due to the startup module and top queue being asynchronous now, move client-nojs/client-js class handling to OutputPage to ensure there is no flashes of wrongly styled or unstyled content. To preserve compatibility for unsupported browsers, undo the class swap at runtime after the isCompatible() check. ResourceLoader startup module: * Load the startup module with <script async>. * Use DOM methods instead of 'document.write' to create base module request (jquery|mediawiki). mw.loader: * Drop 'async' parameter from mw.loader.load(). * Remove the now-unused code paths for synchronous requests. OutputPage: * Drop '$loadCall' parameter from makeResourceLoaderLink(). Asynchronous is now the default and only way to load JavaScript. This means the 'user' module "conditional document-write scripts" are now a simple "mw.loader.load( url )" call. * Fix incorrect @return of makeResourceLoaderLink(). This returns an array not a string. * Improve documentation of makeResourceLoaderLink(). * Drop '$inHead' parameter from getScriptsForBottomQueue(). No longer used. Compatibility with the $wgResourceLoaderExperimentalAsyncLoading feature is maintained. It just no longer needs to change the way the queue works since it's always asynchronous. The feature flag now only controls whether the bottom queue starts at the bottom or starts at the top. * Remove jQuery.ready() optimisation. This was mostly there to avoid the setTimeout() loop jQuery does to detect dom-ready in IE6/IE7 (which we no longer serve JavaScript at all). And for a bug in Firefox with document.write (which is no longer used as of this commit). Bug: T107399 Change-Id: Icba6d7a87b239bf127a221bc6bc432cfa71a4a72
2015-07-28 02:46:00 +00:00
*/
Convert all array() syntax to [] Per wikitech-l consensus: https://lists.wikimedia.org/pipermail/wikitech-l/2016-February/084821.html Notes: * Disabled CallTimePassByReference due to false positives (T127163) Change-Id: I2c8ce713ce6600a0bb7bf67537c87044c7a45c4b
2016-02-17 09:09:32 +00:00
public function makeResourceLoaderLink( $modules, $only, array $extraQuery = [] ) {
OutputPage: Apply target and origin filter to exempt modules Follows-up 80e5b160e, which moved filter logic to getRlClient() so that ResourceLoaderClientHtml doesn't have to deal with this. While that worked fine for most modules (including 'site'), it no longer applied the filter for exempt/hardcoded modules (such as 'site.styles' and 'startup'). Bug: T143066 Change-Id: Iec924003873bc47484a0dc2f1a215f87aa4afdfb
2016-08-16 22:25:08 +00:00
// Apply 'target' and 'origin' filters
$modules = $this->filterModules( (array)$modules, null, $only );
ResourceLoader namespace (attempt 2) Move ResourceLoader classes to their own namespace. Strip the "ResourceLoader" prefix from all except ResourceLoader itself. Move the tests by analogy. I used a namespace alias "RL" in some callers since RL\Module is less ambiguous at the call site than just "Module". I did not address DependencyStore which continues to have a non-standard location and namespace. Revert of a241d83e0a6dabedf. Bug: T308718 Change-Id: Id08a220e1d6085e2b33f3f6c9d0e3935a4204659
2022-05-06 09:09:56 +00:00
return RL\ClientHtml::makeLoad(
resourceloader: Move queue formatting out of OutputPage HTML formatting of the queue was distributed over several OutputPage methods. Each method demanding a snippet of HTML by calling makeResourceLoaderLink() with a limited amount of information. As such, makeResourceLoaderLink() was unable to provide the client with the proper state information. Centralising it also allows it to better reduce duplication in HTML output and maintain a more accurate state. Problems fixed by centralising: 1. The 'user' module is special (due to per-user 'version' and 'user' params). It is manually requested via script-src. To avoid a separate (and wrong) request from something that requires it, we set state=loading directly. However, because the module is in the bottom, the old HTML formatter could only put state=loading in the bottom also. This sometimes caused a wrong request to be fired for modules=user if something in the top queue triggered a requirement for it. 2. Since a464d1d4 (T87871) we track states of page-style modules, with purpose of allowing dependencies on style modules without risking duplicate loading on pages where the styles are loaded already. This didn't work, because the state information about page-style modules is output near the stylesheet, which is after the script tag with mw.loader.load(). That runs first, and mw.loader would still make a duplicate request before it learns the state. Changes: * Document reasons for style/script tag order in getHeadHtml (per 09537e83). * Pass $type from getModuleStyles() to getAllowedModules(). This wasn't needed before since a duplicate check in makeResourceLoaderLink() verified the origin a second time. * Declare explicit position 'top' on 'user.options' and 'user.tokens' module. Previously, OutputPage hardcoded them in the top. The new formatter doesn't. * Remove getHeadScripts(). * Remove getInlineHeadScripts(). * Remove getExternalHeadScripts(). * Remove buildCssLinks(). * Remove getScriptsForBottomQueue(). * Change where Skin::setupSkinUserCss() is called. This methods lets the skin add modules to the queue. Previously it was called from buildCssLinks(), via headElement(), via prepareQuickTemplate(), via OutputPage::output(). It's now in OutputPage::output() directly (slightly earlier). This is needed because prepareQuickTemplate() calls bottomScripts() before headElement(). And bottomScript() would lazy-initialise the queue and lock it before setupSkinUserCss() is called from headElement(). This makes execution order more predictable instead of being dependent on the arbitrary order of data extraction in prepareQuickTemplate (which varies from one skin to another). * Compute isUserModulePreview() and isKnownEmpty() for the 'user' module early on so. This avoids wrongful loading and fixes problem 1. Effective changes in output: * mw.loader.state() is now before mw.loader.load(). This fixes problem 2. * mw.loader.state() now sets 'user.options' and 'user.tokens' to "loading". * mw.loader.state() now sets 'user' (as "loading" or "ready"). Fixes problem 1. * The <script async src> tag for 'startup' changed position (slightly). Previously it was after all inline scripts and stylesheets. It's still after all inline scripts and after most stylesheets, but before any user styles. Since the queue is now formatted outside OutputPage, it can't inject the meta-ResourceLoaderDynamicStyles tag and user-stylesheet hack in the middle of existing output. This shouldn't have any noticable impact. Bug: T87871 Change-Id: I605b8cd1e1fc009b4662a0edbc54d09dd65ee1df
2016-07-15 14:13:09 +00:00
$this->getRlClientContext(),
OutputPage: Apply target and origin filter to exempt modules Follows-up 80e5b160e, which moved filter logic to getRlClient() so that ResourceLoaderClientHtml doesn't have to deal with this. While that worked fine for most modules (including 'site'), it no longer applied the filter for exempt/hardcoded modules (such as 'site.styles' and 'startup'). Bug: T143066 Change-Id: Iec924003873bc47484a0dc2f1a215f87aa4afdfb
2016-08-16 22:25:08 +00:00
$modules,
resourceloader: Move queue formatting out of OutputPage HTML formatting of the queue was distributed over several OutputPage methods. Each method demanding a snippet of HTML by calling makeResourceLoaderLink() with a limited amount of information. As such, makeResourceLoaderLink() was unable to provide the client with the proper state information. Centralising it also allows it to better reduce duplication in HTML output and maintain a more accurate state. Problems fixed by centralising: 1. The 'user' module is special (due to per-user 'version' and 'user' params). It is manually requested via script-src. To avoid a separate (and wrong) request from something that requires it, we set state=loading directly. However, because the module is in the bottom, the old HTML formatter could only put state=loading in the bottom also. This sometimes caused a wrong request to be fired for modules=user if something in the top queue triggered a requirement for it. 2. Since a464d1d4 (T87871) we track states of page-style modules, with purpose of allowing dependencies on style modules without risking duplicate loading on pages where the styles are loaded already. This didn't work, because the state information about page-style modules is output near the stylesheet, which is after the script tag with mw.loader.load(). That runs first, and mw.loader would still make a duplicate request before it learns the state. Changes: * Document reasons for style/script tag order in getHeadHtml (per 09537e83). * Pass $type from getModuleStyles() to getAllowedModules(). This wasn't needed before since a duplicate check in makeResourceLoaderLink() verified the origin a second time. * Declare explicit position 'top' on 'user.options' and 'user.tokens' module. Previously, OutputPage hardcoded them in the top. The new formatter doesn't. * Remove getHeadScripts(). * Remove getInlineHeadScripts(). * Remove getExternalHeadScripts(). * Remove buildCssLinks(). * Remove getScriptsForBottomQueue(). * Change where Skin::setupSkinUserCss() is called. This methods lets the skin add modules to the queue. Previously it was called from buildCssLinks(), via headElement(), via prepareQuickTemplate(), via OutputPage::output(). It's now in OutputPage::output() directly (slightly earlier). This is needed because prepareQuickTemplate() calls bottomScripts() before headElement(). And bottomScript() would lazy-initialise the queue and lock it before setupSkinUserCss() is called from headElement(). This makes execution order more predictable instead of being dependent on the arbitrary order of data extraction in prepareQuickTemplate (which varies from one skin to another). * Compute isUserModulePreview() and isKnownEmpty() for the 'user' module early on so. This avoids wrongful loading and fixes problem 1. Effective changes in output: * mw.loader.state() is now before mw.loader.load(). This fixes problem 2. * mw.loader.state() now sets 'user.options' and 'user.tokens' to "loading". * mw.loader.state() now sets 'user' (as "loading" or "ready"). Fixes problem 1. * The <script async src> tag for 'startup' changed position (slightly). Previously it was after all inline scripts and stylesheets. It's still after all inline scripts and after most stylesheets, but before any user styles. Since the queue is now formatted outside OutputPage, it can't inject the meta-ResourceLoaderDynamicStyles tag and user-stylesheet hack in the middle of existing output. This shouldn't have any noticable impact. Bug: T87871 Change-Id: I605b8cd1e1fc009b4662a0edbc54d09dd65ee1df
2016-07-15 14:13:09 +00:00
$only,
Initial support for Content Security Policy, disabled by default The primary goal here is a defense in depth measure to stop an attacker who found a bug in the parser allowing them to insert malicious attributes. This wouldn't stop someone who could insert a full script tag (since at current it can't distinguish between malicious and legit user js). It also would not prevent DOM-based or reflected XSS for anons, as the nonce value is guessable for anons when receiving a response cached by varnish. However, the limited protection of just stopping stored XSS where the attacker only has control of attributes, is still a big win in my opinion. (But it wouldn't prevent someone who has that type of xss from abusing things like data-ooui attribute). This will likely break many gadgets. Its expected that any sort of rollout on Wikimedia will be done very slowly, with lots of testing and the report-only option to begin with. This is behind feature flags that are off by default, so merging this patch should not cause any change in default behaviour. This may break some extensions (The most obvious one is charinsert (See fe648d41005), but will probably need some testing in report-only mode to see if anything else breaks) This uses the unsafe-eval option of CSP, in order to support RL's local storage thingy. For better security, we may want to remove some of the sillier uses of eval (e.g. jquery.ui.datepicker.js). For more info, see spec: https://www.w3.org/TR/CSP2/ Additionally see: https://www.mediawiki.org/wiki/Requests_for_comment/Content-Security-Policy Bug: T135963 Change-Id: I80f6f469ba4c0b608385483457df96ccb7429ae5
2016-02-29 04:13:10 +00:00
$extraQuery,
Mild refactoring of ContentSecurityPolicy This is to make it behave in a more object orientied way. The goal is to make it be easier to allow extensions to mark certain pages as requiring a different policy (For example, CodeEditor extension uses a blob: url with a WebWorker. We don't want to include that on the policy of every page, but allow the extension to mark it as required whenever needed). This commit does not change code behaviour in any way. Change-Id: I4bf53dabb6e6c5446cea99a64db68b300cef2fd4
2019-10-28 05:01:17 +00:00
$this->CSP->getNonce()
resourceloader: Move queue formatting out of OutputPage HTML formatting of the queue was distributed over several OutputPage methods. Each method demanding a snippet of HTML by calling makeResourceLoaderLink() with a limited amount of information. As such, makeResourceLoaderLink() was unable to provide the client with the proper state information. Centralising it also allows it to better reduce duplication in HTML output and maintain a more accurate state. Problems fixed by centralising: 1. The 'user' module is special (due to per-user 'version' and 'user' params). It is manually requested via script-src. To avoid a separate (and wrong) request from something that requires it, we set state=loading directly. However, because the module is in the bottom, the old HTML formatter could only put state=loading in the bottom also. This sometimes caused a wrong request to be fired for modules=user if something in the top queue triggered a requirement for it. 2. Since a464d1d4 (T87871) we track states of page-style modules, with purpose of allowing dependencies on style modules without risking duplicate loading on pages where the styles are loaded already. This didn't work, because the state information about page-style modules is output near the stylesheet, which is after the script tag with mw.loader.load(). That runs first, and mw.loader would still make a duplicate request before it learns the state. Changes: * Document reasons for style/script tag order in getHeadHtml (per 09537e83). * Pass $type from getModuleStyles() to getAllowedModules(). This wasn't needed before since a duplicate check in makeResourceLoaderLink() verified the origin a second time. * Declare explicit position 'top' on 'user.options' and 'user.tokens' module. Previously, OutputPage hardcoded them in the top. The new formatter doesn't. * Remove getHeadScripts(). * Remove getInlineHeadScripts(). * Remove getExternalHeadScripts(). * Remove buildCssLinks(). * Remove getScriptsForBottomQueue(). * Change where Skin::setupSkinUserCss() is called. This methods lets the skin add modules to the queue. Previously it was called from buildCssLinks(), via headElement(), via prepareQuickTemplate(), via OutputPage::output(). It's now in OutputPage::output() directly (slightly earlier). This is needed because prepareQuickTemplate() calls bottomScripts() before headElement(). And bottomScript() would lazy-initialise the queue and lock it before setupSkinUserCss() is called from headElement(). This makes execution order more predictable instead of being dependent on the arbitrary order of data extraction in prepareQuickTemplate (which varies from one skin to another). * Compute isUserModulePreview() and isKnownEmpty() for the 'user' module early on so. This avoids wrongful loading and fixes problem 1. Effective changes in output: * mw.loader.state() is now before mw.loader.load(). This fixes problem 2. * mw.loader.state() now sets 'user.options' and 'user.tokens' to "loading". * mw.loader.state() now sets 'user' (as "loading" or "ready"). Fixes problem 1. * The <script async src> tag for 'startup' changed position (slightly). Previously it was after all inline scripts and stylesheets. It's still after all inline scripts and after most stylesheets, but before any user styles. Since the queue is now formatted outside OutputPage, it can't inject the meta-ResourceLoaderDynamicStyles tag and user-stylesheet hack in the middle of existing output. This shouldn't have any noticable impact. Bug: T87871 Change-Id: I605b8cd1e1fc009b4662a0edbc54d09dd65ee1df
2016-07-15 14:13:09 +00:00
);
Merging resourceloader branch into trunk. Full documentation is at http://www.mediawiki.org/wiki/ResourceLoader and a general overview has been posted on wikitech-li <http://lists.wikimedia.org/pipermail/wikitech-l/2010-September/049253.html>. One important change is that all JS is now loaded at the bottom, so any scripts assuming things from wikibits or whatever are present will fail.
2010-09-04 04:00:09 +00:00
}
Removed trailing whitespaces
2010-10-02 13:45:35 +00:00
resourceloader: Refactor module links output Changes: * Removed hardcoded logic in OutputPage regarding modules being "enabled". Previously we would always output state=loading and use $wgAllowUserJs (and others) to decide whether to output state=ready or makeResourceLoaderLink. Now, we no longer unconditionally output state=loading and simply always call makeResourceLoaderLink. That method takes care of checking whether modules are enabled and non-empty and returns state=ready when that is the case. This cleans up cases where the duplicated and incomplete logic in OutputPage thought the module was non-empty but turned out to be empty and thus would output both state=loading and later state=ready for the same module. * Clean up documentation for makeResourceLoaderLink (inconsistent ordering of type hint and $var, and @return was missing the fact that the returned html can also contain <link>). * makeResourceLoaderLink now returns an array of html and module states. This allows the consumer of this method to combine the states in 1 larger script tag on top of multiple makeResourceLoaderLink calls (e.g. one state script followed by multiple <script src=load.php>). This isn't to reduce html output, but to make sure we inform mw.loader about modules before the <script src=load.php>. If we were to mix/alternate the state script and load.php requests (which are blocking in html), it is possible for those scripts to request other modules. We need to prevent duplicate loading of modules we already know are going to be requested by the HTML output futher down. * Removed spurious new line. Example of change in HTML output: * The output has been reduced from: - loader.state( site: loading, user: loading, user.groups: loading ) - loader.load( .. ) - <script src="load.php?modules=site .."> - loader.state( user: ready, user.groups: ready ) to: - loader.state( site: loading, user: ready, user.groups: ready ) - loader.load( .. ) - <script src="load.php?modules=site .."> Change-Id: I91754ce5fae3d05b4bfa7372372eba81ee2fc579
2013-11-14 16:54:19 +00:00
/**
resourceloader: Move queue formatting out of OutputPage HTML formatting of the queue was distributed over several OutputPage methods. Each method demanding a snippet of HTML by calling makeResourceLoaderLink() with a limited amount of information. As such, makeResourceLoaderLink() was unable to provide the client with the proper state information. Centralising it also allows it to better reduce duplication in HTML output and maintain a more accurate state. Problems fixed by centralising: 1. The 'user' module is special (due to per-user 'version' and 'user' params). It is manually requested via script-src. To avoid a separate (and wrong) request from something that requires it, we set state=loading directly. However, because the module is in the bottom, the old HTML formatter could only put state=loading in the bottom also. This sometimes caused a wrong request to be fired for modules=user if something in the top queue triggered a requirement for it. 2. Since a464d1d4 (T87871) we track states of page-style modules, with purpose of allowing dependencies on style modules without risking duplicate loading on pages where the styles are loaded already. This didn't work, because the state information about page-style modules is output near the stylesheet, which is after the script tag with mw.loader.load(). That runs first, and mw.loader would still make a duplicate request before it learns the state. Changes: * Document reasons for style/script tag order in getHeadHtml (per 09537e83). * Pass $type from getModuleStyles() to getAllowedModules(). This wasn't needed before since a duplicate check in makeResourceLoaderLink() verified the origin a second time. * Declare explicit position 'top' on 'user.options' and 'user.tokens' module. Previously, OutputPage hardcoded them in the top. The new formatter doesn't. * Remove getHeadScripts(). * Remove getInlineHeadScripts(). * Remove getExternalHeadScripts(). * Remove buildCssLinks(). * Remove getScriptsForBottomQueue(). * Change where Skin::setupSkinUserCss() is called. This methods lets the skin add modules to the queue. Previously it was called from buildCssLinks(), via headElement(), via prepareQuickTemplate(), via OutputPage::output(). It's now in OutputPage::output() directly (slightly earlier). This is needed because prepareQuickTemplate() calls bottomScripts() before headElement(). And bottomScript() would lazy-initialise the queue and lock it before setupSkinUserCss() is called from headElement(). This makes execution order more predictable instead of being dependent on the arbitrary order of data extraction in prepareQuickTemplate (which varies from one skin to another). * Compute isUserModulePreview() and isKnownEmpty() for the 'user' module early on so. This avoids wrongful loading and fixes problem 1. Effective changes in output: * mw.loader.state() is now before mw.loader.load(). This fixes problem 2. * mw.loader.state() now sets 'user.options' and 'user.tokens' to "loading". * mw.loader.state() now sets 'user' (as "loading" or "ready"). Fixes problem 1. * The <script async src> tag for 'startup' changed position (slightly). Previously it was after all inline scripts and stylesheets. It's still after all inline scripts and after most stylesheets, but before any user styles. Since the queue is now formatted outside OutputPage, it can't inject the meta-ResourceLoaderDynamicStyles tag and user-stylesheet hack in the middle of existing output. This shouldn't have any noticable impact. Bug: T87871 Change-Id: I605b8cd1e1fc009b4662a0edbc54d09dd65ee1df
2016-07-15 14:13:09 +00:00
* Combine WrappedString chunks and filter out empty ones
*
* @param array $chunks
OutputPage: Make use of WrappedStringList in headElement() Right now, getInlineHeadScripts(), buildCssLinks() and getExternalHeadScripts() all return WrappedString::join(). But because they don't know about each other and because they need to return strings (not arrays), headElement() has no way of merging them. WrappedStringList allows this array to be kept, whilst still being backward-compatible with code that calls these methods and assumig a string (since it will lazy-join the array if the object is treated like a string). To be used by I8b6c6a10d965e7396. Output is not changed in this commit. Merely refactoring. Change-Id: Iae08345473bd93cc0948d51b62c48aeb1ea460a3
2016-07-07 19:26:36 +00:00
* @return string|WrappedStringList HTML
resourceloader: Refactor module links output Changes: * Removed hardcoded logic in OutputPage regarding modules being "enabled". Previously we would always output state=loading and use $wgAllowUserJs (and others) to decide whether to output state=ready or makeResourceLoaderLink. Now, we no longer unconditionally output state=loading and simply always call makeResourceLoaderLink. That method takes care of checking whether modules are enabled and non-empty and returns state=ready when that is the case. This cleans up cases where the duplicated and incomplete logic in OutputPage thought the module was non-empty but turned out to be empty and thus would output both state=loading and later state=ready for the same module. * Clean up documentation for makeResourceLoaderLink (inconsistent ordering of type hint and $var, and @return was missing the fact that the returned html can also contain <link>). * makeResourceLoaderLink now returns an array of html and module states. This allows the consumer of this method to combine the states in 1 larger script tag on top of multiple makeResourceLoaderLink calls (e.g. one state script followed by multiple <script src=load.php>). This isn't to reduce html output, but to make sure we inform mw.loader about modules before the <script src=load.php>. If we were to mix/alternate the state script and load.php requests (which are blocking in html), it is possible for those scripts to request other modules. We need to prevent duplicate loading of modules we already know are going to be requested by the HTML output futher down. * Removed spurious new line. Example of change in HTML output: * The output has been reduced from: - loader.state( site: loading, user: loading, user.groups: loading ) - loader.load( .. ) - <script src="load.php?modules=site .."> - loader.state( user: ready, user.groups: ready ) to: - loader.state( site: loading, user: ready, user.groups: ready ) - loader.load( .. ) - <script src="load.php?modules=site .."> Change-Id: I91754ce5fae3d05b4bfa7372372eba81ee2fc579
2013-11-14 16:54:19 +00:00
*/
resourceloader: Move queue formatting out of OutputPage HTML formatting of the queue was distributed over several OutputPage methods. Each method demanding a snippet of HTML by calling makeResourceLoaderLink() with a limited amount of information. As such, makeResourceLoaderLink() was unable to provide the client with the proper state information. Centralising it also allows it to better reduce duplication in HTML output and maintain a more accurate state. Problems fixed by centralising: 1. The 'user' module is special (due to per-user 'version' and 'user' params). It is manually requested via script-src. To avoid a separate (and wrong) request from something that requires it, we set state=loading directly. However, because the module is in the bottom, the old HTML formatter could only put state=loading in the bottom also. This sometimes caused a wrong request to be fired for modules=user if something in the top queue triggered a requirement for it. 2. Since a464d1d4 (T87871) we track states of page-style modules, with purpose of allowing dependencies on style modules without risking duplicate loading on pages where the styles are loaded already. This didn't work, because the state information about page-style modules is output near the stylesheet, which is after the script tag with mw.loader.load(). That runs first, and mw.loader would still make a duplicate request before it learns the state. Changes: * Document reasons for style/script tag order in getHeadHtml (per 09537e83). * Pass $type from getModuleStyles() to getAllowedModules(). This wasn't needed before since a duplicate check in makeResourceLoaderLink() verified the origin a second time. * Declare explicit position 'top' on 'user.options' and 'user.tokens' module. Previously, OutputPage hardcoded them in the top. The new formatter doesn't. * Remove getHeadScripts(). * Remove getInlineHeadScripts(). * Remove getExternalHeadScripts(). * Remove buildCssLinks(). * Remove getScriptsForBottomQueue(). * Change where Skin::setupSkinUserCss() is called. This methods lets the skin add modules to the queue. Previously it was called from buildCssLinks(), via headElement(), via prepareQuickTemplate(), via OutputPage::output(). It's now in OutputPage::output() directly (slightly earlier). This is needed because prepareQuickTemplate() calls bottomScripts() before headElement(). And bottomScript() would lazy-initialise the queue and lock it before setupSkinUserCss() is called from headElement(). This makes execution order more predictable instead of being dependent on the arbitrary order of data extraction in prepareQuickTemplate (which varies from one skin to another). * Compute isUserModulePreview() and isKnownEmpty() for the 'user' module early on so. This avoids wrongful loading and fixes problem 1. Effective changes in output: * mw.loader.state() is now before mw.loader.load(). This fixes problem 2. * mw.loader.state() now sets 'user.options' and 'user.tokens' to "loading". * mw.loader.state() now sets 'user' (as "loading" or "ready"). Fixes problem 1. * The <script async src> tag for 'startup' changed position (slightly). Previously it was after all inline scripts and stylesheets. It's still after all inline scripts and after most stylesheets, but before any user styles. Since the queue is now formatted outside OutputPage, it can't inject the meta-ResourceLoaderDynamicStyles tag and user-stylesheet hack in the middle of existing output. This shouldn't have any noticable impact. Bug: T87871 Change-Id: I605b8cd1e1fc009b4662a0edbc54d09dd65ee1df
2016-07-15 14:13:09 +00:00
protected static function combineWrappedStrings( array $chunks ) {
resourceloader: Use WrappedString library to merge RLQ inline scripts Update unit test to account for the internal 'html' prop now being an array instead of string. And update expected values to no longer have a trailing line break. Bug: T27202 Change-Id: I105b6ef2e64ab8b891562e16940edb88592bd415
2015-07-31 00:13:04 +00:00
// Filter out empty values
resourceloader: Move queue formatting out of OutputPage HTML formatting of the queue was distributed over several OutputPage methods. Each method demanding a snippet of HTML by calling makeResourceLoaderLink() with a limited amount of information. As such, makeResourceLoaderLink() was unable to provide the client with the proper state information. Centralising it also allows it to better reduce duplication in HTML output and maintain a more accurate state. Problems fixed by centralising: 1. The 'user' module is special (due to per-user 'version' and 'user' params). It is manually requested via script-src. To avoid a separate (and wrong) request from something that requires it, we set state=loading directly. However, because the module is in the bottom, the old HTML formatter could only put state=loading in the bottom also. This sometimes caused a wrong request to be fired for modules=user if something in the top queue triggered a requirement for it. 2. Since a464d1d4 (T87871) we track states of page-style modules, with purpose of allowing dependencies on style modules without risking duplicate loading on pages where the styles are loaded already. This didn't work, because the state information about page-style modules is output near the stylesheet, which is after the script tag with mw.loader.load(). That runs first, and mw.loader would still make a duplicate request before it learns the state. Changes: * Document reasons for style/script tag order in getHeadHtml (per 09537e83). * Pass $type from getModuleStyles() to getAllowedModules(). This wasn't needed before since a duplicate check in makeResourceLoaderLink() verified the origin a second time. * Declare explicit position 'top' on 'user.options' and 'user.tokens' module. Previously, OutputPage hardcoded them in the top. The new formatter doesn't. * Remove getHeadScripts(). * Remove getInlineHeadScripts(). * Remove getExternalHeadScripts(). * Remove buildCssLinks(). * Remove getScriptsForBottomQueue(). * Change where Skin::setupSkinUserCss() is called. This methods lets the skin add modules to the queue. Previously it was called from buildCssLinks(), via headElement(), via prepareQuickTemplate(), via OutputPage::output(). It's now in OutputPage::output() directly (slightly earlier). This is needed because prepareQuickTemplate() calls bottomScripts() before headElement(). And bottomScript() would lazy-initialise the queue and lock it before setupSkinUserCss() is called from headElement(). This makes execution order more predictable instead of being dependent on the arbitrary order of data extraction in prepareQuickTemplate (which varies from one skin to another). * Compute isUserModulePreview() and isKnownEmpty() for the 'user' module early on so. This avoids wrongful loading and fixes problem 1. Effective changes in output: * mw.loader.state() is now before mw.loader.load(). This fixes problem 2. * mw.loader.state() now sets 'user.options' and 'user.tokens' to "loading". * mw.loader.state() now sets 'user' (as "loading" or "ready"). Fixes problem 1. * The <script async src> tag for 'startup' changed position (slightly). Previously it was after all inline scripts and stylesheets. It's still after all inline scripts and after most stylesheets, but before any user styles. Since the queue is now formatted outside OutputPage, it can't inject the meta-ResourceLoaderDynamicStyles tag and user-stylesheet hack in the middle of existing output. This shouldn't have any noticable impact. Bug: T87871 Change-Id: I605b8cd1e1fc009b4662a0edbc54d09dd65ee1df
2016-07-15 14:13:09 +00:00
$chunks = array_filter( $chunks, 'strlen' );
return WrappedString::join( "\n", $chunks );
resourceloader: Refactor module links output Changes: * Removed hardcoded logic in OutputPage regarding modules being "enabled". Previously we would always output state=loading and use $wgAllowUserJs (and others) to decide whether to output state=ready or makeResourceLoaderLink. Now, we no longer unconditionally output state=loading and simply always call makeResourceLoaderLink. That method takes care of checking whether modules are enabled and non-empty and returns state=ready when that is the case. This cleans up cases where the duplicated and incomplete logic in OutputPage thought the module was non-empty but turned out to be empty and thus would output both state=loading and later state=ready for the same module. * Clean up documentation for makeResourceLoaderLink (inconsistent ordering of type hint and $var, and @return was missing the fact that the returned html can also contain <link>). * makeResourceLoaderLink now returns an array of html and module states. This allows the consumer of this method to combine the states in 1 larger script tag on top of multiple makeResourceLoaderLink calls (e.g. one state script followed by multiple <script src=load.php>). This isn't to reduce html output, but to make sure we inform mw.loader about modules before the <script src=load.php>. If we were to mix/alternate the state script and load.php requests (which are blocking in html), it is possible for those scripts to request other modules. We need to prevent duplicate loading of modules we already know are going to be requested by the HTML output futher down. * Removed spurious new line. Example of change in HTML output: * The output has been reduced from: - loader.state( site: loading, user: loading, user.groups: loading ) - loader.load( .. ) - <script src="load.php?modules=site .."> - loader.state( user: ready, user.groups: ready ) to: - loader.state( site: loading, user: ready, user.groups: ready ) - loader.load( .. ) - <script src="load.php?modules=site .."> Change-Id: I91754ce5fae3d05b4bfa7372372eba81ee2fc579
2013-11-14 16:54:19 +00:00
}
Optimize the order of styles and scripts in <head> Emit styles and scripts in the following order: * Inline <script>...</script> tags. * External stylesheets. * External script tags. I have attempted this before in I98d383a6, and subsequently reverted (in I151f74a41) when it failed to show a performance improvement. Our ability to measure performance were substantially poorer then, so I would like to give this another try and see if it makes a dent in first render times. Change-Id: I5e6c79c7041aa77bcdc6130e91e77aa538334c42
2015-08-13 21:30:31 +00:00
/**
resourceloader: Remove ResourceLoaderModule::getPosition @deprecated since 1.29 Bug: T184257 Change-Id: I1a59a1f0a5fbedfea9f0b28bee8c298710557a8d
2018-02-12 21:06:05 +00:00
* JS stuff to put at the bottom of the `<body>`.
* These are legacy scripts ($this->mScripts), and user JS.
Optimize the order of styles and scripts in <head> Emit styles and scripts in the following order: * Inline <script>...</script> tags. * External stylesheets. * External script tags. I have attempted this before in I98d383a6, and subsequently reverted (in I151f74a41) when it failed to show a performance improvement. Our ability to measure performance were substantially poorer then, so I would like to give this another try and see if it makes a dent in first render times. Change-Id: I5e6c79c7041aa77bcdc6130e91e77aa538334c42
2015-08-13 21:30:31 +00:00
*
Soft deprecate Skin::bottomScripts Also prepare OutputPage::getBottomScripts to handle the work of SkinAfterBottomScripts hook in the future. Bug: T257704 Change-Id: Ib0a8b72faa88259c24b5fb7e9d5a908423494247
2021-08-30 10:32:02 +00:00
* @param string $extraHtml (only for use by this->tailElement(); will be removed in future)
OutputPage: Make use of WrappedStringList in headElement() Right now, getInlineHeadScripts(), buildCssLinks() and getExternalHeadScripts() all return WrappedString::join(). But because they don't know about each other and because they need to return strings (not arrays), headElement() has no way of merging them. WrappedStringList allows this array to be kept, whilst still being backward-compatible with code that calls these methods and assumig a string (since it will lazy-join the array if the object is treated like a string). To be used by I8b6c6a10d965e7396. Output is not changed in this commit. Merely refactoring. Change-Id: Iae08345473bd93cc0948d51b62c48aeb1ea460a3
2016-07-07 19:26:36 +00:00
* @return string|WrappedStringList HTML
Optimize the order of styles and scripts in <head> Emit styles and scripts in the following order: * Inline <script>...</script> tags. * External stylesheets. * External script tags. I have attempted this before in I98d383a6, and subsequently reverted (in I151f74a41) when it failed to show a performance improvement. Our ability to measure performance were substantially poorer then, so I would like to give this another try and see if it makes a dent in first render times. Change-Id: I5e6c79c7041aa77bcdc6130e91e77aa538334c42
2015-08-13 21:30:31 +00:00
*/
Soft deprecate Skin::bottomScripts Also prepare OutputPage::getBottomScripts to handle the work of SkinAfterBottomScripts hook in the future. Bug: T257704 Change-Id: Ib0a8b72faa88259c24b5fb7e9d5a908423494247
2021-08-30 10:32:02 +00:00
public function getBottomScripts( $extraHtml = '' ) {
resourceloader: Move queue formatting out of OutputPage HTML formatting of the queue was distributed over several OutputPage methods. Each method demanding a snippet of HTML by calling makeResourceLoaderLink() with a limited amount of information. As such, makeResourceLoaderLink() was unable to provide the client with the proper state information. Centralising it also allows it to better reduce duplication in HTML output and maintain a more accurate state. Problems fixed by centralising: 1. The 'user' module is special (due to per-user 'version' and 'user' params). It is manually requested via script-src. To avoid a separate (and wrong) request from something that requires it, we set state=loading directly. However, because the module is in the bottom, the old HTML formatter could only put state=loading in the bottom also. This sometimes caused a wrong request to be fired for modules=user if something in the top queue triggered a requirement for it. 2. Since a464d1d4 (T87871) we track states of page-style modules, with purpose of allowing dependencies on style modules without risking duplicate loading on pages where the styles are loaded already. This didn't work, because the state information about page-style modules is output near the stylesheet, which is after the script tag with mw.loader.load(). That runs first, and mw.loader would still make a duplicate request before it learns the state. Changes: * Document reasons for style/script tag order in getHeadHtml (per 09537e83). * Pass $type from getModuleStyles() to getAllowedModules(). This wasn't needed before since a duplicate check in makeResourceLoaderLink() verified the origin a second time. * Declare explicit position 'top' on 'user.options' and 'user.tokens' module. Previously, OutputPage hardcoded them in the top. The new formatter doesn't. * Remove getHeadScripts(). * Remove getInlineHeadScripts(). * Remove getExternalHeadScripts(). * Remove buildCssLinks(). * Remove getScriptsForBottomQueue(). * Change where Skin::setupSkinUserCss() is called. This methods lets the skin add modules to the queue. Previously it was called from buildCssLinks(), via headElement(), via prepareQuickTemplate(), via OutputPage::output(). It's now in OutputPage::output() directly (slightly earlier). This is needed because prepareQuickTemplate() calls bottomScripts() before headElement(). And bottomScript() would lazy-initialise the queue and lock it before setupSkinUserCss() is called from headElement(). This makes execution order more predictable instead of being dependent on the arbitrary order of data extraction in prepareQuickTemplate (which varies from one skin to another). * Compute isUserModulePreview() and isKnownEmpty() for the 'user' module early on so. This avoids wrongful loading and fixes problem 1. Effective changes in output: * mw.loader.state() is now before mw.loader.load(). This fixes problem 2. * mw.loader.state() now sets 'user.options' and 'user.tokens' to "loading". * mw.loader.state() now sets 'user' (as "loading" or "ready"). Fixes problem 1. * The <script async src> tag for 'startup' changed position (slightly). Previously it was after all inline scripts and stylesheets. It's still after all inline scripts and after most stylesheets, but before any user styles. Since the queue is now formatted outside OutputPage, it can't inject the meta-ResourceLoaderDynamicStyles tag and user-stylesheet hack in the middle of existing output. This shouldn't have any noticable impact. Bug: T87871 Change-Id: I605b8cd1e1fc009b4662a0edbc54d09dd65ee1df
2016-07-15 14:13:09 +00:00
$chunks = [];
$chunks[] = $this->getRlClient()->getBodyHtml();
// Legacy non-ResourceLoader scripts
$chunks[] = $this->mScripts;
OutputPage: Support loading some mw.config vars late, at end of body JS config vars listed in the 'LateJSConfigVarNames' extension attribute or in the OutputPage::CORE_LATE_JS_CONFIG_VAR_NAMES constant will now be added to the end of the HTML response, instead of early in the HTML head where it blocks rendering/painting. Bug: T41813 Change-Id: I86d461756398780dc24949013f35b7730a481052
2022-10-25 22:36:20 +00:00
$vars = $this->getJSVars( self::JS_VAR_LATE );
Include JS variable for NewPP report Adapted from reverted commit b7c4c8717f9. Bug: T110763 Change-Id: If249b679c534879bfac622592a1d2fa913a0cf9d
2016-11-10 20:29:27 +00:00
if ( $this->limitReportJSData ) {
OutputPage: Support loading some mw.config vars late, at end of body JS config vars listed in the 'LateJSConfigVarNames' extension attribute or in the OutputPage::CORE_LATE_JS_CONFIG_VAR_NAMES constant will now be added to the end of the HTML response, instead of early in the HTML head where it blocks rendering/painting. Bug: T41813 Change-Id: I86d461756398780dc24949013f35b7730a481052
2022-10-25 22:36:20 +00:00
$vars['wgPageParseReport'] = $this->limitReportJSData;
}
if ( $vars ) {
ResourceLoader: Permit partial failure in OutputPage::getBottomScripts Up until recently, `wgPageParseReport` was the only property in the bottom payload for mw.config. However, with the introduction of "late variables" in I86d46175639 (e3784468fc, T41813) this can now contains variables from extension as well. When a variable moves from one to the other, it should have no difference other than its position. Currently, this is not the case because the encoding in `<head>` silently tolerates invalid UTF-8 and has for several years. The bottom one, however, produces a fatal error. Change this to make the two more similar. As of I89832142b55e, encodeJson() improves debugging by logging the specific error that json_encode encountered as well. Bug: T329330 Bug: T41813 Change-Id: I1b78e81d08e11dce3426dfc195db5499efdc2dcf
2023-02-14 23:44:37 +00:00
$rlContext = $this->getRlClientContext();
Include JS variable for NewPP report Adapted from reverted commit b7c4c8717f9. Bug: T110763 Change-Id: If249b679c534879bfac622592a1d2fa913a0cf9d
2016-11-10 20:29:27 +00:00
$chunks[] = ResourceLoader::makeInlineScript(
ResourceLoader: Permit partial failure in OutputPage::getBottomScripts Up until recently, `wgPageParseReport` was the only property in the bottom payload for mw.config. However, with the introduction of "late variables" in I86d46175639 (e3784468fc, T41813) this can now contains variables from extension as well. When a variable moves from one to the other, it should have no difference other than its position. Currently, this is not the case because the encoding in `<head>` silently tolerates invalid UTF-8 and has for several years. The bottom one, however, produces a fatal error. Change this to make the two more similar. As of I89832142b55e, encodeJson() improves debugging by logging the specific error that json_encode encountered as well. Bug: T329330 Bug: T41813 Change-Id: I1b78e81d08e11dce3426dfc195db5499efdc2dcf
2023-02-14 23:44:37 +00:00
'mw.config.set(' . $rlContext->encodeJson( $vars ) . ');',
Mild refactoring of ContentSecurityPolicy This is to make it behave in a more object orientied way. The goal is to make it be easier to allow extensions to mark certain pages as requiring a different policy (For example, CodeEditor extension uses a blob: url with a WebWorker. We don't want to include that on the policy of every page, but allow the extension to mark it as required whenever needed). This commit does not change code behaviour in any way. Change-Id: I4bf53dabb6e6c5446cea99a64db68b300cef2fd4
2019-10-28 05:01:17 +00:00
$this->CSP->getNonce()
Include JS variable for NewPP report Adapted from reverted commit b7c4c8717f9. Bug: T110763 Change-Id: If249b679c534879bfac622592a1d2fa913a0cf9d
2016-11-10 20:29:27 +00:00
);
}
OutputPage: Support loading some mw.config vars late, at end of body JS config vars listed in the 'LateJSConfigVarNames' extension attribute or in the OutputPage::CORE_LATE_JS_CONFIG_VAR_NAMES constant will now be added to the end of the HTML response, instead of early in the HTML head where it blocks rendering/painting. Bug: T41813 Change-Id: I86d461756398780dc24949013f35b7730a481052
2022-10-25 22:36:20 +00:00
Prepare for deprecation of Skin::bottomScripts Bug: T62846 Change-Id: I203881d7ab035bc953d113106f21348187b51374
2022-04-06 15:32:49 +00:00
// Keep the hook appendage separate to preserve WrappedString objects.
// This enables BaseTemplate::getTrail() to merge them where possible.
$this->getHookRunner()->onSkinAfterBottomScripts( $this->getSkin(), $extraHtml );
$chunks = [ self::combineWrappedStrings( $chunks ) ];
if ( $extraHtml !== '' ) {
$chunks[] = $extraHtml;
}
Include JS variable for NewPP report Adapted from reverted commit b7c4c8717f9. Bug: T110763 Change-Id: If249b679c534879bfac622592a1d2fa913a0cf9d
2016-11-10 20:29:27 +00:00
Prepare for deprecation of Skin::bottomScripts Bug: T62846 Change-Id: I203881d7ab035bc953d113106f21348187b51374
2022-04-06 15:32:49 +00:00
return WrappedString::join( "\n", $chunks );
ResourceLoader: Add an experimental option to move the main module loading queue (the bottom queue) from the bottom of the <body> up into the <head> , while still being loaded asynchronously. This makes them load earlier, which should make the page load faster. This is the product of a long discussion on bug 27488 * Added a "blocking" state to mw.loader . When loading scripts while the document is not ready, the loader will use document.write() if blocking is true, and append to the <body> or the <head> if blocking is false. If the document is ready, the loader will always append to the <body> * Enable blocking mode while loading the top queue, and disable it after. This ensures that modules in the top queue are still loaded in a blocking way as they were before * If $wgResourceLoaderExperimentalAsyncLoading is true, the bottom queue is also loaded in the head, but with blocking mode disabled. Otherwise, it's loaded at the bottom of the <body> as before * scripts-only and messages-only requests need special treatment: ** in the top queue, they can continue to use <script src="..."> tags because they are blocking ** if the bottom queue is at the bottom of the <body> (experimental async loading disabled), they can continue to use <script src="..."> tags as before ** if the bottom queue is in the <head> (experimental async loading enabled), they cannot use <script src="..."> tags, because those would block. Instead, call mw.loader.load() on the load.php URL
2012-01-05 23:32:41 +00:00
}
Add ParserOutput::addJsConfigVars The OutputPage has variables for modules, moduleScripts, moduleStyles, moduleMessages and the config vars, but the ParserOutput is missing the last one. With ParserOutput::addJsConfigVars it is possible to add scripts and it config vars at one place and have not use the MakeGlobalVariablesScript hook or other ways to get the needed javascript variable in the output. Change-Id: I6ad61cca76805f6b9d76e053c98c7509c323d5da
2014-01-19 15:39:46 +00:00
/**
* Get the javascript config vars to include on this page
*
Fixed some @params documentation Swapped some "$var type" to "type $var" or added missing types before the $var. Changed some other types to match the more common spelling. Makes beginning of some text in captial. Change-Id: I64e8cfe478cb0ba438f40b0631d6e9049cdab567
2014-04-10 18:50:10 +00:00
* @return array Array of javascript config vars
Add ParserOutput::addJsConfigVars The OutputPage has variables for modules, moduleScripts, moduleStyles, moduleMessages and the config vars, but the ParserOutput is missing the last one. With ParserOutput::addJsConfigVars it is possible to add scripts and it config vars at one place and have not use the MakeGlobalVariablesScript hook or other ways to get the needed javascript variable in the output. Change-Id: I6ad61cca76805f6b9d76e053c98c7509c323d5da
2014-01-19 15:39:46 +00:00
* @since 1.23
*/
public function getJsConfigVars() {
return $this->mJsConfigVars;
}
Moved Skin::makeGlobalVariablesScript() to OutputPage::getJSVars() * merged <script> and if ( window.mediaWiki ) block with the one of mediaWiki.loader as stated in the doc * removed dependency of $wgTitle * the two only calls to this functions are in SemanticForms, but will not affect current version of MediaWiki: ** specials/SF_UploadWindow.php: this file is only used before 1.16 ** specials/SF_UploadWindow2.php: the call is part of the else branch of a "method_exists( $wgOut, 'addModules' )" check, which means it's not called since 1.17
2011-02-16 18:25:44 +00:00
/**
Fixed some @params documentation Swapped some "$var type" to "type $var" or added missing types before the $var. Changed some other types to match the more common spelling. Makes beginning of some text in captial. Change-Id: I64e8cfe478cb0ba438f40b0631d6e9049cdab567
2014-04-10 18:50:10 +00:00
* Add one or more variables to be set in mw.config in JavaScript
random w/s cleanup
2011-04-23 19:28:35 +00:00
*
Fixed some @params documentation Swapped some "$var type" to "type $var" or added missing types before the $var. Changed some other types to match the more common spelling. Makes beginning of some text in captial. Change-Id: I64e8cfe478cb0ba438f40b0631d6e9049cdab567
2014-04-10 18:50:10 +00:00
* @param string|array $keys Key or array of key/value pairs
Fix PhanTypeMismatchDeclaredParam Auto fix MediaWiki.Commenting.FunctionComment.DefaultNullTypeParam sniff Change-Id: I865323fd0295aabd06f3e3c75e0e5043fb31069e
2018-06-26 21:14:43 +00:00
* @param mixed|null $value [optional] Value of the configuration variable
New OutputPage::addJsConfigVars() method (bug 31233) * to make the output page specific mw.config map extendable. * fixes (bug 31233) Make OutputPage's mw.config array extendable internally instead of hard coded * fixes weird unfinished comment from r61690
2011-09-28 22:08:08 +00:00
*/
[Core JS] More fixing of global config variable usage * mw.config is the new way, and global config variable lookups are deprecated * Based on two phase3-wide quick searches: -- of " wg": http://toolserver.org/~krinkle/wikimedia-svn-search/view.php?id=321&hash=81700bf7486e4fee3b7bc1f83eb9eba6 -- of "!wg": http://toolserver.org/~krinkle/wikimedia-svn-search/view.php?id=327&hash=47c9d54a7a1d5d58a724dd834585f40d Related changes: * Changed some php comments mentioning "wg" variables to include the dollar sign, and a typo when the wf function prefix was meant. * Removed TODO comment in wikibits.js and made it use the JS equivalent of wfUrlencode, which we have now, mw.util.wikiUrlencode * SpecialUpload.php: use OutputPage::addJsConfigVars instead of creating a new script tag through OutputPage::addScript(Skin::makeVariablesScript(..)) * Renamed wgUploadSetup in upload.js and made it local. Not used anywhere in ./trunk/phase3 and ./trunk/extensions * Fix OutputPage::addJsConfigVars so that it can actually be called with an array instead of two arguments for key/value * Some minor whitespace/convention stuff around the same line
2011-12-31 21:25:00 +00:00
public function addJsConfigVars( $keys, $value = null ) {
New OutputPage::addJsConfigVars() method (bug 31233) * to make the output page specific mw.config map extendable. * fixes (bug 31233) Make OutputPage's mw.config array extendable internally instead of hard coded * fixes weird unfinished comment from r61690
2011-09-28 22:08:08 +00:00
if ( is_array( $keys ) ) {
foreach ( $keys as $key => $value ) {
$this->mJsConfigVars[$key] = $value;
}
return;
}
$this->mJsConfigVars[$keys] = $value;
}
/**
* Get an array containing the variables to be set in mw.config in JavaScript.
*
OutputPage: Support loading some mw.config vars late, at end of body JS config vars listed in the 'LateJSConfigVarNames' extension attribute or in the OutputPage::CORE_LATE_JS_CONFIG_VAR_NAMES constant will now be added to the end of the HTML response, instead of early in the HTML head where it blocks rendering/painting. Bug: T41813 Change-Id: I86d461756398780dc24949013f35b7730a481052
2022-10-25 22:36:20 +00:00
* Do not add things here which can be evaluated in RL\StartUpModule,
* in other words, page-independent/site-wide variables (without state).
* These would add a blocking HTML cost to page rendering time, and require waiting for
* HTTP caches to expire before configuration changes take effect everywhere.
*
* By default, these are loaded in the HTML head and block page rendering.
* Config variable names can be set in CORE_LATE_JS_CONFIG_VAR_NAMES, or
* for extensions via the 'LateJSConfigVarNames' attribute, to opt-in to
* being sent from the end of the HTML body instead, to improve page load time.
* In JavaScript, late variables should be accessed via mw.hook('wikipage.content').
*
* @param int|null $flag Return only the specified kind of variables: self::JS_VAR_EARLY or self::JS_VAR_LATE.
* For internal use only.
Moar documentations
2011-11-29 21:04:20 +00:00
* @return array
Moved Skin::makeGlobalVariablesScript() to OutputPage::getJSVars() * merged <script> and if ( window.mediaWiki ) block with the one of mediaWiki.loader as stated in the doc * removed dependency of $wgTitle * the two only calls to this functions are in SemanticForms, but will not affect current version of MediaWiki: ** specials/SF_UploadWindow.php: this file is only used before 1.16 ** specials/SF_UploadWindow2.php: the call is part of the else branch of a "method_exists( $wgOut, 'addModules' )" check, which means it's not called since 1.17
2011-02-16 18:25:44 +00:00
*/
OutputPage: Support loading some mw.config vars late, at end of body JS config vars listed in the 'LateJSConfigVarNames' extension attribute or in the OutputPage::CORE_LATE_JS_CONFIG_VAR_NAMES constant will now be added to the end of the HTML response, instead of early in the HTML head where it blocks rendering/painting. Bug: T41813 Change-Id: I86d461756398780dc24949013f35b7730a481052
2022-10-25 22:36:20 +00:00
public function getJSVars( ?int $flag = null ) {
OutputPage->getJSVars: Maintain symmetry between JS & PHP vars The function is more difficult to read than it need be by dint of some things two names. I can't be the only one who finds 'wgCurRevisionId' => $latestRevID ugly, right? Anyhow, lots of JS code depends on the JS variable names, whereas the PHP variables exist only in the scope of that method, so it's clear that it's the PHP names that should be brought in line. Change-Id: Ibddd5d2fc8d75e0ade18ff3433714d52605f2911
2013-09-26 05:03:16 +00:00
$curRevisionId = 0;
$articleId = 0;
includes: Replace implicit Bugzilla bug numbers with Phab ones It's unreasonable to expect newbies to know that "bug 12345" means "Task T14345" except where it doesn't, so let's just standardise on the real numbers. Change-Id: I6f59febaf8fc96e80f8cfc11f4356283f461142a
2017-02-20 22:44:19 +00:00
$canonicalSpecialPageName = false; # T23115
Reduce calls to MediaWikiServices::getInstance() In some functions MediaWikiServices::getInstance() was called twices or in loops. Extract the variable to reduce calls. Change-Id: I2705db11d7a9ea73efb9b5a5c40747ab0b3ea36f
2018-08-18 04:02:39 +00:00
$services = MediaWikiServices::getInstance();
Use the WikiPage object object instead of Title to get the page id and latest revision
2012-01-29 20:35:32 +00:00
Moved Skin::makeGlobalVariablesScript() to OutputPage::getJSVars() * merged <script> and if ( window.mediaWiki ) block with the one of mediaWiki.loader as stated in the doc * removed dependency of $wgTitle * the two only calls to this functions are in SemanticForms, but will not affect current version of MediaWiki: ** specials/SF_UploadWindow.php: this file is only used before 1.16 ** specials/SF_UploadWindow2.php: the call is part of the else branch of a "method_exists( $wgOut, 'addModules' )" check, which means it's not called since 1.17
2011-02-16 18:25:44 +00:00
$title = $this->getTitle();
$ns = $title->getNamespace();
Use local variabe $services instead of MediaWikiServices::getInstance() Change-Id: Ieb6230231d9b916fd1366e7b473db00c87ac4428
2019-08-26 14:08:10 +00:00
$nsInfo = $services->getNamespaceInfo();
Mass conversion to NamespaceInfo Change-Id: I2fef157ceec772f304c0923a1cd8c0eef2e82a0f
2018-08-05 17:58:51 +00:00
$canonicalNamespace = $nsInfo->exists( $ns )
? $nsInfo->getCanonicalName( $ns )
Pass php-strict for OutputPage Change-Id: Ia5d03277fff543ca672dc8efc7eaaeca4888ac23
2014-05-11 01:54:15 +00:00
: $title->getNsText();
Use the WikiPage object object instead of Title to get the page id and latest revision
2012-01-29 20:35:32 +00:00
Add "wgRelevantUserName" from skin to mw.config A skin can have a relevant user, then some help links in the sidebar are shown. When a user want extend this informationen with userjs, he has to parse the existing items or the title param of the url to get the name of the user for which this help links are shown. Having the name as javascript variable makes it easier to add more links in the sidebar. Change-Id: I17a75902b6e739d5149d332b6a94a6568b79501f
2013-10-29 20:49:17 +00:00
$sk = $this->getSkin();
User:PleaseStand patch fixes bug 34972
2012-03-13 17:57:54 +00:00
// Get the relevant title so that AJAX features can use the correct page name
includes: Replace implicit Bugzilla bug numbers with Phab ones It's unreasonable to expect newbies to know that "bug 12345" means "Task T14345" except where it doesn't, so let's just standardise on the real numbers. Change-Id: I6f59febaf8fc96e80f8cfc11f4356283f461142a
2017-02-20 22:44:19 +00:00
// when making API requests from certain special pages (T36972).
Add "wgRelevantUserName" from skin to mw.config A skin can have a relevant user, then some help links in the sidebar are shown. When a user want extend this informationen with userjs, he has to parse the existing items or the title param of the url to get the name of the user for which this help links are shown. Having the name as javascript variable makes it easier to add more links in the sidebar. Change-Id: I17a75902b6e739d5149d332b6a94a6568b79501f
2013-10-29 20:49:17 +00:00
$relevantTitle = $sk->getRelevantTitle();
User:PleaseStand patch fixes bug 34972
2012-03-13 17:57:54 +00:00
Use strict (in)equality with namespaces constants when LHS is definitely an integer Change-Id: I8fede00dfe1270d93c5d78d3c36e788cddfc8a99
2020-07-22 17:29:48 +00:00
if ( $ns === NS_SPECIAL ) {
Use name constants to access config settings. Use name constants instead of string literals in calls to Config::get and ServiceOptions::get, when referring to core configuration variables. This protects against typos and makes the decumentation and schema declaration of the config settings discoverable. This is the first batch, only touching files directly under /includes/ Change-Id: I7252e636c7c86d950d9257b33491af492c6dd5eb
2022-04-01 15:58:32 +00:00
[ $canonicalSpecialPageName, /*...*/ ] =
Reduce calls to MediaWikiServices::getInstance() In some functions MediaWikiServices::getInstance() was called twices or in loops. Extract the variable to reduce calls. Change-Id: I2705db11d7a9ea73efb9b5a5c40747ab0b3ea36f
2018-08-18 04:02:39 +00:00
$services->getSpecialPageFactory()->
Mass conversion to SpecialPageFactory service Change-Id: Ia6e1e819ec6cbe8bf75b820109f51d47863e31fc
2018-08-12 09:08:58 +00:00
resolveAlias( $title->getDBkey() );
Revert "Replace deprecated methods IContextSource::getWikiPage && IContextSource::canUseWikiPage" This reverts commit ef458e894888d5d984f6b8dcf75cfae008d544d0. Reason for revert: Causes page tabs to disappear on Special:WhatLinksHere. Bug: T297744 Change-Id: I0ee282a9f7a5a9b2cfdc3261d800d9e27eaf977e
2021-12-14 20:34:46 +00:00
} elseif ( $this->canUseWikiPage() ) {
$wikiPage = $this->getWikiPage();
$curRevisionId = $wikiPage->getLatest();
$articleId = $wikiPage->getId();
* (bug 27680) Fix for r82273: wgCanonicalSpecialPageName no longer false when requesting a special page with subpage parameter The problem is that in this case $wgOut's Title object is not exactely the same $wgTitle because of line 600 of SpecialPage: $wgTitle = $page->getTitle(); which strips the subpage parameter. Now using SpecialPage::resolveAliasWithSubpage() so that it works in all cases. Also PHP strikes again for not allowing SpecialPage::resolveAliasWithSubpage( $title->getDBkey() )[0] syntax. YAY :)
2011-02-25 11:48:14 +00:00
}
Moved Skin::makeGlobalVariablesScript() to OutputPage::getJSVars() * merged <script> and if ( window.mediaWiki ) block with the one of mediaWiki.loader as stated in the doc * removed dependency of $wgTitle * the two only calls to this functions are in SemanticForms, but will not affect current version of MediaWiki: ** specials/SF_UploadWindow.php: this file is only used before 1.16 ** specials/SF_UploadWindow2.php: the call is part of the else branch of a "method_exists( $wgOut, 'addModules' )" check, which means it's not called since 1.17
2011-02-16 18:25:44 +00:00
OutputPage: Set wgPageContentLanguage to pageViewLang instead of pageLang Follows-up 80a090ea7 (r104483). I'm assuming the difference is that pageLanguage will be the canonical variant for the page (as authored in wikitext, presumably the wiki's default), whereas pageViewLanguage will be the variant of the pageLanguage based on the current user-interface language (if and only if the user interface is related). Viewing zh by default has: * userLanguage: zh * pageLanguage: zh * pageViewLanguage: zh Viewing zh-cn by default has: * userLanguage: zh * pageLanguage: zh * pageViewLanguage: zh-cn Viewing zh-cn with uselang=en has: * userLanguage: en * pageLanguage: zh * pageViewLanguage: zh-cn Viewing zh-cn with uselang=zh-tw has: * userLanguage: zh-tw * pageLanguage: zh * pageViewLanguage: zh-cn The main use case behind wgPageContentLanguage in 80a090ea7 was tablesorter which should match the language and variant of the rendered wikipage content. The parser output and div#mw-content-text element use pageViewLanguage, so let's set wgPageContentLanguage to that as well. Change-Id: Ibf399dfe5ca60fb55e7acd75a5c0238d8b903807
2016-03-08 23:44:18 +00:00
$lang = $title->getPageViewLanguage();
* (bug 27721) Make JavaScript variables wgSeparatorTransformTable and wgDigitTransformTable depend on page content language so the sort script sorts correctly more often. It will e.g. sort according to the user language on special pages, where localised numbers are usually used. * New JavaScript variable wgPageContentLanguage (independent from the bug, but I assume this var can be useful)
2011-11-28 20:53:26 +00:00
// Pre-process information
$separatorTransTable = $lang->separatorTransformTable();
PHP: Use short ternary operator (?:) where possible Change-Id: Idcc7e4fcdd4d8302ceda44bf6d294fa8c2219381
2018-06-11 09:16:48 +00:00
$separatorTransTable = $separatorTransTable ?: [];
Convert all array() syntax to [] Per wikitech-l consensus: https://lists.wikimedia.org/pipermail/wikitech-l/2016-February/084821.html Notes: * Disabled CallTimePassByReference due to false positives (T127163) Change-Id: I2c8ce713ce6600a0bb7bf67537c87044c7a45c4b
2016-02-17 09:09:32 +00:00
$compactSeparatorTransTable = [
* (bug 27721) Make JavaScript variables wgSeparatorTransformTable and wgDigitTransformTable depend on page content language so the sort script sorts correctly more often. It will e.g. sort according to the user language on special pages, where localised numbers are usually used. * New JavaScript variable wgPageContentLanguage (independent from the bug, but I assume this var can be useful)
2011-11-28 20:53:26 +00:00
implode( "\t", array_keys( $separatorTransTable ) ),
implode( "\t", $separatorTransTable ),
Convert all array() syntax to [] Per wikitech-l consensus: https://lists.wikimedia.org/pipermail/wikitech-l/2016-February/084821.html Notes: * Disabled CallTimePassByReference due to false positives (T127163) Change-Id: I2c8ce713ce6600a0bb7bf67537c87044c7a45c4b
2016-02-17 09:09:32 +00:00
];
* (bug 27721) Make JavaScript variables wgSeparatorTransformTable and wgDigitTransformTable depend on page content language so the sort script sorts correctly more often. It will e.g. sort according to the user language on special pages, where localised numbers are usually used. * New JavaScript variable wgPageContentLanguage (independent from the bug, but I assume this var can be useful)
2011-11-28 20:53:26 +00:00
$digitTransTable = $lang->digitTransformTable();
PHP: Use short ternary operator (?:) where possible Change-Id: Idcc7e4fcdd4d8302ceda44bf6d294fa8c2219381
2018-06-11 09:16:48 +00:00
$digitTransTable = $digitTransTable ?: [];
Convert all array() syntax to [] Per wikitech-l consensus: https://lists.wikimedia.org/pipermail/wikitech-l/2016-February/084821.html Notes: * Disabled CallTimePassByReference due to false positives (T127163) Change-Id: I2c8ce713ce6600a0bb7bf67537c87044c7a45c4b
2016-02-17 09:09:32 +00:00
$compactDigitTransTable = [
* (bug 27721) Make JavaScript variables wgSeparatorTransformTable and wgDigitTransformTable depend on page content language so the sort script sorts correctly more often. It will e.g. sort according to the user language on special pages, where localised numbers are usually used. * New JavaScript variable wgPageContentLanguage (independent from the bug, but I assume this var can be useful)
2011-11-28 20:53:26 +00:00
implode( "\t", array_keys( $digitTransTable ) ),
implode( "\t", $digitTransTable ),
Convert all array() syntax to [] Per wikitech-l consensus: https://lists.wikimedia.org/pipermail/wikitech-l/2016-February/084821.html Notes: * Disabled CallTimePassByReference due to false positives (T127163) Change-Id: I2c8ce713ce6600a0bb7bf67537c87044c7a45c4b
2016-02-17 09:09:32 +00:00
];
* (bug 27721) Make JavaScript variables wgSeparatorTransformTable and wgDigitTransformTable depend on page content language so the sort script sorts correctly more often. It will e.g. sort according to the user language on special pages, where localised numbers are usually used. * New JavaScript variable wgPageContentLanguage (independent from the bug, but I assume this var can be useful)
2011-11-28 20:53:26 +00:00
Expose additional user attributes as JS vars This change adds user ID, registration time, and edit count to the default set of JavaScript configuration variables. These values are all included in $mCacheVars and so require no additional database lookups. Change-Id: Id0adb795567d4ba932104938f2793cdb4084883c
2013-01-22 22:44:22 +00:00
$user = $this->getUser();
OutputPage,ResourceLoaderStartupModule: Separate internal from public config vars Change-Id: Ic54d3b36fb379d77139b21b46db7da9f78869e37
2019-10-12 00:18:43 +00:00
// Internal variables for MediaWiki core
Convert all array() syntax to [] Per wikitech-l consensus: https://lists.wikimedia.org/pipermail/wikitech-l/2016-February/084821.html Notes: * Disabled CallTimePassByReference due to false positives (T127163) Change-Id: I2c8ce713ce6600a0bb7bf67537c87044c7a45c4b
2016-02-17 09:09:32 +00:00
$vars = [
mediawiki.page.startup: Merge with mediawiki.page.ready module The distinction between the two was lost several years years ago when the "position top" queue ceased to be a thing. Since then, the two have been loaded and executed together in the same batch, and are doing similar things. mediawiki.page.ready is publicly used in several places as dependendency, but mediawiki.page.startup is entirely internal to core, which makes it the easier one of the two to dissolve. Keep an alias for two weeks for cache compat to avoid console warnings about unknown modules. Although even some cache still refers to it, this is harmless since the errors are recoverable and the correct module was also loaded by the cached pages already. Bug: T260210 Change-Id: Ic418c23a7400abba22fd07b17f173d3c5f1d1d10
2020-02-22 01:01:24 +00:00
// @internal For mediawiki.page.ready
OutputPage,ResourceLoaderStartupModule: Separate internal from public config vars Change-Id: Ic54d3b36fb379d77139b21b46db7da9f78869e37
2019-10-12 00:18:43 +00:00
'wgBreakFrames' => $this->getFrameOptions() == 'DENY',
// @internal For jquery.tablesorter
'wgSeparatorTransformTable' => $compactSeparatorTransTable,
'wgDigitTransformTable' => $compactDigitTransTable,
'wgDefaultDateFormat' => $lang->getDefaultDateFormat(),
'wgMonthNames' => $lang->getMonthNamesArray(),
// @internal For debugging purposes
'wgRequestId' => WebRequest::getRequestId(),
// @internal For mw.loader
'wgCSPNonce' => $this->CSP->getNonce(),
];
// Start of supported and stable config vars (for use by extensions/gadgets).
$vars += [
OutputPage->getJSVars: Maintain symmetry between JS & PHP vars The function is more difficult to read than it need be by dint of some things two names. I can't be the only one who finds 'wgCurRevisionId' => $latestRevID ugly, right? Anyhow, lots of JS code depends on the JS variable names, whereas the PHP variables exist only in the scope of that method, so it's clear that it's the PHP names that should be brought in line. Change-Id: Ibddd5d2fc8d75e0ade18ff3433714d52605f2911
2013-09-26 05:03:16 +00:00
'wgCanonicalNamespace' => $canonicalNamespace,
'wgCanonicalSpecialPageName' => $canonicalSpecialPageName,
Moved Skin::makeGlobalVariablesScript() to OutputPage::getJSVars() * merged <script> and if ( window.mediaWiki ) block with the one of mediaWiki.loader as stated in the doc * removed dependency of $wgTitle * the two only calls to this functions are in SemanticForms, but will not affect current version of MediaWiki: ** specials/SF_UploadWindow.php: this file is only used before 1.16 ** specials/SF_UploadWindow2.php: the call is part of the else branch of a "method_exists( $wgOut, 'addModules' )" check, which means it's not called since 1.17
2011-02-16 18:25:44 +00:00
'wgNamespaceNumber' => $title->getNamespace(),
Correct case of Title::getPrefixedDBkey() calls Change-Id: Ic9ef0cc6dbb0a8eb0ee93432299ef59759223e2c
2013-03-09 20:14:22 +00:00
'wgPageName' => $title->getPrefixedDBkey(),
Moved Skin::makeGlobalVariablesScript() to OutputPage::getJSVars() * merged <script> and if ( window.mediaWiki ) block with the one of mediaWiki.loader as stated in the doc * removed dependency of $wgTitle * the two only calls to this functions are in SemanticForms, but will not affect current version of MediaWiki: ** specials/SF_UploadWindow.php: this file is only used before 1.16 ** specials/SF_UploadWindow2.php: the call is part of the else branch of a "method_exists( $wgOut, 'addModules' )" check, which means it's not called since 1.17
2011-02-16 18:25:44 +00:00
'wgTitle' => $title->getText(),
OutputPage->getJSVars: Maintain symmetry between JS & PHP vars The function is more difficult to read than it need be by dint of some things two names. I can't be the only one who finds 'wgCurRevisionId' => $latestRevID ugly, right? Anyhow, lots of JS code depends on the JS variable names, whereas the PHP variables exist only in the scope of that method, so it's clear that it's the PHP names that should be brought in line. Change-Id: Ibddd5d2fc8d75e0ade18ff3433714d52605f2911
2013-09-26 05:03:16 +00:00
'wgCurRevisionId' => $curRevisionId,
Add wgRevisionId variable to ResourceLoader mw.config provides wgCurRevisionId, which has the latest revision number. This adds wgRevisionId, which indicates the revision number currently being viewed. This is useful when looking at old revisions. Bug: 51594 Change-Id: Ie0d00f3a9a8af8533ab28204b5bb483a0092c710
2013-08-26 18:52:23 +00:00
'wgRevisionId' => (int)$this->getRevisionId(),
OutputPage->getJSVars: Maintain symmetry between JS & PHP vars The function is more difficult to read than it need be by dint of some things two names. I can't be the only one who finds 'wgCurRevisionId' => $latestRevID ugly, right? Anyhow, lots of JS code depends on the JS variable names, whereas the PHP variables exist only in the scope of that method, so it's clear that it's the PHP names that should be brought in line. Change-Id: Ibddd5d2fc8d75e0ade18ff3433714d52605f2911
2013-09-26 05:03:16 +00:00
'wgArticleId' => $articleId,
Moved Skin::makeGlobalVariablesScript() to OutputPage::getJSVars() * merged <script> and if ( window.mediaWiki ) block with the one of mediaWiki.loader as stated in the doc * removed dependency of $wgTitle * the two only calls to this functions are in SemanticForms, but will not affect current version of MediaWiki: ** specials/SF_UploadWindow.php: this file is only used before 1.16 ** specials/SF_UploadWindow2.php: the call is part of the else branch of a "method_exists( $wgOut, 'addModules' )" check, which means it's not called since 1.17
2011-02-16 18:25:44 +00:00
'wgIsArticle' => $this->isArticle(),
mw.config: Add wgIsRedirect Change-Id: I6e365f20b513ded10ede51d4b34de0736653a3dc
2013-07-23 19:11:17 +00:00
'wgIsRedirect' => $title->isRedirect(),
context: Add a cached RequestContext::getActionName method This method is dependent on, and inherently must depend on, all of Title, WikiPage, and WebRequest. And, like Title and WikiPage, which also have getters in RequestContext, Action is also derived from a query parameter that is widely recognised in almost all web requests to index.php. The status quo in core and extensions, is to obtain this value via Action::getActionName(), which as a static method that bypasses dependency injection and also has the problem of not being cached. Caching it within ActionFactory seems hard and awkward, due to varying by context. In change I61d66211bd (22f9a32853f) a cached wrapper method was added internally to the Skin class. In change I8cbc4bba4d248d9 (235820d631) another cached wrapper was added in the Gadgets extension. This change takes this approach further by making it a stable public method on RequestContext. To facilitate testing and to offer basic confidence in this working correctly, this commit also adopts the new method in two place that are considered "safe" (Skin, and OutputPage). Both of these are called relatively late in the PHP proccess and well after any Setup code and overrides (such as in MediaWiki.php), during which it is more complex to call this. I'll audit and update those in a subsequent change. Change-Id: I1e259b54dca48a32be5a8c6cbb8eb69aec2da115
2021-11-06 00:28:23 +00:00
'wgAction' => $this->getContext()->getActionName(),
Expose additional user attributes as JS vars This change adds user ID, registration time, and edit count to the default set of JavaScript configuration variables. These values are all included in $mCacheVars and so require no additional database lookups. Change-Id: Id0adb795567d4ba932104938f2793cdb4084883c
2013-01-22 22:44:22 +00:00
'wgUserName' => $user->isAnon() ? null : $user->getName(),
Replace deprecated function in OutputPage Now mocking LanguageConverter Change-Id: I25f2f90d8dcf2cb09429e01873f1982dca199976
2020-11-03 17:18:15 +00:00
'wgUserGroups' => $services->getUserGroupManager()->getUserEffectiveGroups( $user ),
Moved Skin::makeGlobalVariablesScript() to OutputPage::getJSVars() * merged <script> and if ( window.mediaWiki ) block with the one of mediaWiki.loader as stated in the doc * removed dependency of $wgTitle * the two only calls to this functions are in SemanticForms, but will not affect current version of MediaWiki: ** specials/SF_UploadWindow.php: this file is only used before 1.16 ** specials/SF_UploadWindow2.php: the call is part of the else branch of a "method_exists( $wgOut, 'addModules' )" check, which means it's not called since 1.17
2011-02-16 18:25:44 +00:00
'wgCategories' => $this->getCategories(),
* (bug 27721) Make JavaScript variables wgSeparatorTransformTable and wgDigitTransformTable depend on page content language so the sort script sorts correctly more often. It will e.g. sort according to the user language on special pages, where localised numbers are usually used. * New JavaScript variable wgPageContentLanguage (independent from the bug, but I assume this var can be useful)
2011-11-28 20:53:26 +00:00
'wgPageContentLanguage' => $lang->getCode(),
mw.config: Add wgPageContentModel for content model of current page This field of the page table can be useful to determine the corrosponding model for a page. Change-Id: I0525ad764540b213b2697e1e9e64f7da73f6cf83
2013-05-07 19:59:40 +00:00
'wgPageContentModel' => $title->getContentModel(),
Correct case of Title::getPrefixedDBkey() calls Change-Id: Ic9ef0cc6dbb0a8eb0ee93432299ef59759223e2c
2013-03-09 20:14:22 +00:00
'wgRelevantPageName' => $relevantTitle->getPrefixedDBkey(),
Many more function case mismatches Change-Id: I5d3a5eb8adea1ecbf136415bb9fd7a162633ccca
2016-03-19 00:08:06 +00:00
'wgRelevantArticleId' => $relevantTitle->getArticleID(),
Convert all array() syntax to [] Per wikitech-l consensus: https://lists.wikimedia.org/pipermail/wikitech-l/2016-February/084821.html Notes: * Disabled CallTimePassByReference due to false positives (T127163) Change-Id: I2c8ce713ce6600a0bb7bf67537c87044c7a45c4b
2016-02-17 09:09:32 +00:00
];
Swap out uses of User->isLoggedIn() with ->isRegistered() Bug: T270450 Change-Id: I90ff94a553ebae9a4ef97c25c333f067f19c327d
2020-12-17 23:10:11 +00:00
if ( $user->isRegistered() ) {
Expose additional user attributes as JS vars This change adds user ID, registration time, and edit count to the default set of JavaScript configuration variables. These values are all included in $mCacheVars and so require no additional database lookups. Change-Id: Id0adb795567d4ba932104938f2793cdb4084883c
2013-01-22 22:44:22 +00:00
$vars['wgUserId'] = $user->getId();
mediawiki.user: Add methods for isNamed and isTemp Bug: T327317 Change-Id: Ie3e817a6ebb4c468772c600a02257650750c3387
2023-01-23 10:00:23 +00:00
$vars['wgUserIsTemp'] = $user->isTemp();
Expose additional user attributes as JS vars This change adds user ID, registration time, and edit count to the default set of JavaScript configuration variables. These values are all included in $mCacheVars and so require no additional database lookups. Change-Id: Id0adb795567d4ba932104938f2793cdb4084883c
2013-01-22 22:44:22 +00:00
$vars['wgUserEditCount'] = $user->getEditCount();
Clean up misuse of wfTimestampOrNull() It's primarily meant for nullable timestamps in the database. And for the mere purpose of an is_null() check, it's more confusion than useful to cause this function since in these two cases its return value is not used as-is but further conditionalized. * Also fix fragile link between getRegistration()'s false return value never being passed to wfTimestamp (to mean "now") because of the isLoggedIn() guard. The ternary would now make those follow the 'else' branch toward null. Change-Id: I9a1ee2a56e1767bfb750e27b1f37bdaeb5e6378d
2016-09-23 23:59:56 +00:00
$userReg = $user->getRegistration();
docs: Avoid some scalar juggling Phan can treat scalar types as non-interchangeable with `scalar_implicit_cast` set to false. This patch fixes some of those issues (which are in total >1000), namely the ones with alphabetic order < includes/actions. Change-Id: Ib1c6573ab899088bc319b9da9ceaffc850da3dbe
2019-09-15 14:40:42 +00:00
$vars['wgUserRegistration'] = $userReg ? (int)wfTimestamp( TS_UNIX, $userReg ) * 1000 : null;
Adding GetNewMessagesAlert hook and wgUserNewMsgRevisionId JS global This hook allows extensions to disable or modify the new messages alert ('orange bar of doom') while still allowing the user_newtalk table to be updated. The wgUserNewMsgRevisionId JS global allows gadgets and extensions to create their own new message alerts on the client side. I also threw in a few comment updates for good measure! See also Echo change I3f35a56b which utilizes this. Bug: 47962 Change-Id: I2105bdd2bcd5b27f6f36ec8d8fa7fa99d60a2d82
2013-05-05 01:08:58 +00:00
// Get the revision ID of the oldest new message on the user's talk
// page. This can be used for constructing new message alerts on
// the client side.
Deprecate User::getNewMessageRevisionId Uses User::getNewMessageLinks, which uses Revision objects Only caller updated to use a new private method Bug: T253949 Change-Id: I04c0558916e7216540dd7bd12a2a391a1ed7660a
2020-05-29 05:37:46 +00:00
$userNewMsgRevId = $this->getLastSeenUserTalkRevId();
OutputPage: Only export wgUserNewMsgRevisionId if non-null Bug: T53640 Bug: T233676 Bug: T231168 Change-Id: Ia4f07bb90e6ed029d87947e7cf62c5a5d69726e9
2020-03-13 22:26:16 +00:00
// Only occupy precious space in the <head> when it is non-null (T53640)
// mw.config.get returns null by default.
if ( $userNewMsgRevId ) {
OutputPage: Fix warning when setting wgUserNewMsgRevisionId Follow-up to e08e9609ffa0cc8b52a5752acf665861e3857603, which contained a typo ($$ instead of $) Bug: T248049 Change-Id: I7206cb61a61cad528ece880cffcbdd7b4e04e935
2020-03-19 03:12:48 +00:00
$vars['wgUserNewMsgRevisionId'] = $userNewMsgRevId;
OutputPage: Only export wgUserNewMsgRevisionId if non-null Bug: T53640 Bug: T233676 Bug: T231168 Change-Id: Ia4f07bb90e6ed029d87947e7cf62c5a5d69726e9
2020-03-13 22:26:16 +00:00
}
Expose additional user attributes as JS vars This change adds user ID, registration time, and edit count to the default set of JavaScript configuration variables. These values are all included in $mCacheVars and so require no additional database lookups. Change-Id: Id0adb795567d4ba932104938f2793cdb4084883c
2013-01-22 22:44:22 +00:00
}
Hard deprecation of Language::convertTitle(), ::findVariantLink(), ::updateConversionTable() Co-authored-by: C. Scott Ananian <cananian@wikimedia.org> Bug: T226832 Change-Id: I41a3b67490fc6b9d4c484f566d346a0d10c670e9
2020-05-06 00:26:09 +00:00
$languageConverter = $services->getLanguageConverterFactory()
Support variants of page content language in live preview Before this change variants in live preview are only supported if the content language of the Wiki project is a language with variants. This change allows to preview variants in live preview also if the page content language for a specific page in the database is a language with variants. Bug: T294052 Change-Id: I204171837072a347a75837ea63991125f2891a7b
2021-10-21 18:53:51 +00:00
->getLanguageConverter( $title->getPageLanguage() );
Hard deprecation of Language::convertTitle(), ::findVariantLink(), ::updateConversionTable() Co-authored-by: C. Scott Ananian <cananian@wikimedia.org> Bug: T226832 Change-Id: I41a3b67490fc6b9d4c484f566d346a0d10c670e9
2020-05-06 00:26:09 +00:00
if ( $languageConverter->hasVariants() ) {
$vars['wgUserVariant'] = $languageConverter->getPreferredVariant();
Fix indentation whitespace errors Change-Id: Ie268bee2098c589c050e1b5b0e93fe1b3feca86f
2012-10-26 15:42:13 +00:00
}
Add wgIsProbablyEditable for editability based on quickUserCan. It uses quickUserCan like SkinTemplate, for performance reasons. Bug: 43790 Change-Id: I2f77b69624cc1a4cd74c8178eea53681f3f2b107
2013-05-23 04:31:11 +00:00
// Same test as SkinTemplate
Remove the fallback for create rights (covered in PM edit rights check). Remove key to i18n message, no longer being used. Bug: T272079 Change-Id: Id6566e5241b06ba377c2eca083221ee02556365f
2021-06-24 14:10:43 +00:00
$vars['wgIsProbablyEditable'] = $this->getAuthority()->probablyCan( 'edit', $title );
Remove usages of Title::quickUserCan Change-Id: Ifa53e0ec800e23dc4184d133a100fb9378dfee9e
2019-08-23 23:53:15 +00:00
$vars['wgRelevantPageIsProbablyEditable'] = $relevantTitle &&
Remove the fallback for create rights (covered in PM edit rights check). Remove key to i18n message, no longer being used. Bug: T272079 Change-Id: Id6566e5241b06ba377c2eca083221ee02556365f
2021-06-24 14:10:43 +00:00
$this->getAuthority()->probablyCan( 'edit', $relevantTitle );
Remove usage of protection related deprecated Title function Bug: T306131 Change-Id: I487a12a88ae82c367d1cbb2e52083fe20b27d4ce
2022-04-14 17:12:27 +00:00
$restrictionStore = $services->getRestrictionStore();
foreach ( $restrictionStore->listApplicableRestrictionTypes( $title ) as $type ) {
Add grep comment for js vars wgRestriction* Change-Id: I974841acf97e9426ed2338f210821e91d7f94286
2017-12-27 02:04:45 +00:00
// Following keys are set in $vars:
// wgRestrictionCreate, wgRestrictionEdit, wgRestrictionMove, wgRestrictionUpload
Remove usage of protection related deprecated Title function Bug: T306131 Change-Id: I487a12a88ae82c367d1cbb2e52083fe20b27d4ce
2022-04-14 17:12:27 +00:00
$vars['wgRestriction' . ucfirst( $type )] = $restrictionStore->getRestrictions( $title, $type );
Moved Skin::makeGlobalVariablesScript() to OutputPage::getJSVars() * merged <script> and if ( window.mediaWiki ) block with the one of mediaWiki.loader as stated in the doc * removed dependency of $wgTitle * the two only calls to this functions are in SemanticForms, but will not affect current version of MediaWiki: ** specials/SF_UploadWindow.php: this file is only used before 1.16 ** specials/SF_UploadWindow2.php: the call is part of the else branch of a "method_exists( $wgOut, 'addModules' )" check, which means it's not called since 1.17
2011-02-16 18:25:44 +00:00
}
Only define 'isMainPage' if we are on the main page, the new mw.config library is specifically built for this efficient way of exporting variables as it will prevent any TypeError exceptions and simply return null (it's not a real global anymore). This is a follow-up on Brion's CR at r87212.
2011-06-14 23:29:45 +00:00
if ( $title->isMainPage() ) {
Using true instead of calling isMainPage() again. (Follow-up r90099 CR, Poke r87212)
2011-06-15 01:11:16 +00:00
$vars['wgIsMainPage'] = true;
(bug 32230) Expose wgRedirectedFrom (title of the page we were redirected from, if any) in JavaScript. Patch by MrBlueSky
2011-11-30 12:43:10 +00:00
}
Enforce concealment of hidden RelevantUser at Skin level Skin is what takes a user or extract it from a title to set relevant user and so this is the right place to handle permission check too because if the user is hidden, it should be made known for all callers, instead of hoping each caller will do similar check which is duplication and not guranteed. Skin:setRelevantUser always takes user and this is now typehinted to avoid errors. Callers need to give what's required. Bug: T120883 Change-Id: I177cd83a9c4f87c27a63d8f530c40895d0f3f5dd
2020-12-19 02:57:26 +00:00
$relevantUser = $sk->getRelevantUser();
if ( $relevantUser ) {
OutputPage,ResourceLoaderStartupModule: Separate internal from public config vars Change-Id: Ic54d3b36fb379d77139b21b46db7da9f78869e37
2019-10-12 00:18:43 +00:00
$vars['wgRelevantUserName'] = $relevantUser->getName();
}
// End of stable config vars
Pass php-strict for OutputPage Change-Id: Ia5d03277fff543ca672dc8efc7eaaeca4888ac23
2014-05-11 01:54:15 +00:00
Remove Title from public interface of OutputPage Bug: T278459 Change-Id: If455682540faf8d42fd8a14e0ef09bac24f72db4
2021-04-08 19:11:06 +00:00
$titleFormatter = $services->getTitleFormatter();
(bug 32230) Expose wgRedirectedFrom (title of the page we were redirected from, if any) in JavaScript. Patch by MrBlueSky
2011-11-30 12:43:10 +00:00
if ( $this->mRedirectedFrom ) {
OutputPage,ResourceLoaderStartupModule: Separate internal from public config vars Change-Id: Ic54d3b36fb379d77139b21b46db7da9f78869e37
2019-10-12 00:18:43 +00:00
// @internal For skin JS
Remove Title from public interface of OutputPage Bug: T278459 Change-Id: If455682540faf8d42fd8a14e0ef09bac24f72db4
2021-04-08 19:11:06 +00:00
$vars['wgRedirectedFrom'] = $titleFormatter->getPrefixedDBkey( $this->mRedirectedFrom );
Only define 'isMainPage' if we are on the main page, the new mw.config library is specifically built for this efficient way of exporting variables as it will prevent any TypeError exceptions and simply return null (it's not a real global anymore). This is a follow-up on Brion's CR at r87212.
2011-06-14 23:29:45 +00:00
}
Pass php-strict for OutputPage Change-Id: Ia5d03277fff543ca672dc8efc7eaaeca4888ac23
2014-05-11 01:54:15 +00:00
New OutputPage::addJsConfigVars() method (bug 31233) * to make the output page specific mw.config map extendable. * fixes (bug 31233) Make OutputPage's mw.config array extendable internally instead of hard coded * fixes weird unfinished comment from r61690
2011-09-28 22:08:08 +00:00
// Allow extensions to add their custom variables to the mw.config map.
// Use the 'ResourceLoaderGetConfigVars' hook if the variable is not
docs: Fix 'dependant' typos The intended word in all these cases was the adjective "dependent". Whilst the "dependant" does exist, it is a noun and generally refers to a person. The word is rare used in general, but especially so in a technology context. Change-Id: Ic7e2d2ea6a566f4139ff1fdb77f38b0e962ccd9c
2021-02-18 06:38:52 +00:00
// page-dependent but site-wide (without state).
New OutputPage::addJsConfigVars() method (bug 31233) * to make the output page specific mw.config map extendable. * fixes (bug 31233) Make OutputPage's mw.config array extendable internally instead of hard coded * fixes weird unfinished comment from r61690
2011-09-28 22:08:08 +00:00
// Alternatively, you may want to use OutputPage->addJsConfigVars() instead.
Hooks::run() call site migration Migrate all callers of Hooks::run() to use the new HookContainer/HookRunner system. General principles: * Use DI if it is already used. We're not changing the way state is managed in this patch. * HookContainer is always injected, not HookRunner. HookContainer is a service, it's a more generic interface, it is the only thing that provides isRegistered() which is needed in some cases, and a HookRunner can be efficiently constructed from it (confirmed by benchmark). Because HookContainer is needed for object construction, it is also needed by all factories. * "Ask your friendly local base class". Big hierarchies like SpecialPage and ApiBase have getHookContainer() and getHookRunner() methods in the base class, and classes that extend that base class are not expected to know or care where the base class gets its HookContainer from. * ProtectedHookAccessorTrait provides protected getHookContainer() and getHookRunner() methods, getting them from the global service container. The point of this is to ease migration to DI by ensuring that call sites ask their local friendly base class rather than getting a HookRunner from the service container directly. * Private $this->hookRunner. In some smaller classes where accessor methods did not seem warranted, there is a private HookRunner property which is accessed directly. Very rarely (two cases), there is a protected property, for consistency with code that conventionally assumes protected=private, but in cases where the class might actually be overridden, a protected accessor is preferred over a protected property. * The last resort: Hooks::runner(). Mostly for static, file-scope and global code. In a few cases it was used for objects with broken construction schemes, out of horror or laziness. Constructors with new required arguments: * AuthManager * BadFileLookup * BlockManager * ClassicInterwikiLookup * ContentHandlerFactory * ContentSecurityPolicy * DefaultOptionsManager * DerivedPageDataUpdater * FullSearchResultWidget * HtmlCacheUpdater * LanguageFactory * LanguageNameUtils * LinkRenderer * LinkRendererFactory * LocalisationCache * MagicWordFactory * MessageCache * NamespaceInfo * PageEditStash * PageHandlerFactory * PageUpdater * ParserFactory * PermissionManager * RevisionStore * RevisionStoreFactory * SearchEngineConfig * SearchEngineFactory * SearchFormWidget * SearchNearMatcher * SessionBackend * SpecialPageFactory * UserNameUtils * UserOptionsManager * WatchedItemQueryService * WatchedItemStore Constructors with new optional arguments: * DefaultPreferencesFactory * Language * LinkHolderArray * MovePage * Parser * ParserCache * PasswordReset * Router setHookContainer() now required after construction: * AuthenticationProvider * ResourceLoaderModule * SearchEngine Change-Id: Id442b0dbe43aba84bd5cf801d86dedc768b082c7
2020-03-19 02:42:09 +00:00
$this->getHookRunner()->onMakeGlobalVariablesScript( $vars, $this );
random w/s cleanup
2011-04-23 19:28:35 +00:00
New OutputPage::addJsConfigVars() method (bug 31233) * to make the output page specific mw.config map extendable. * fixes (bug 31233) Make OutputPage's mw.config array extendable internally instead of hard coded * fixes weird unfinished comment from r61690
2011-09-28 22:08:08 +00:00
// Merge in variables from addJsConfigVars last
OutputPage: Support loading some mw.config vars late, at end of body JS config vars listed in the 'LateJSConfigVarNames' extension attribute or in the OutputPage::CORE_LATE_JS_CONFIG_VAR_NAMES constant will now be added to the end of the HTML response, instead of early in the HTML head where it blocks rendering/painting. Bug: T41813 Change-Id: I86d461756398780dc24949013f35b7730a481052
2022-10-25 22:36:20 +00:00
$vars = array_merge( $vars, $this->getJsConfigVars() );
// Return only early or late vars if requested
if ( $flag !== null ) {
$lateVarNames =
array_fill_keys( self::CORE_LATE_JS_CONFIG_VAR_NAMES, true ) +
array_fill_keys( ExtensionRegistry::getInstance()->getAttribute( 'LateJSConfigVarNames' ), true );
foreach ( array_keys( $vars ) as $name ) {
// If the variable's late flag doesn't match the requested late flag, unset it
if ( isset( $lateVarNames[ $name ] ) !== ( $flag === self::JS_VAR_LATE ) ) {
unset( $vars[ $name ] );
}
}
}
return $vars;
Moved Skin::makeGlobalVariablesScript() to OutputPage::getJSVars() * merged <script> and if ( window.mediaWiki ) block with the one of mediaWiki.loader as stated in the doc * removed dependency of $wgTitle * the two only calls to this functions are in SemanticForms, but will not affect current version of MediaWiki: ** specials/SF_UploadWindow.php: this file is only used before 1.16 ** specials/SF_UploadWindow2.php: the call is part of the else branch of a "method_exists( $wgOut, 'addModules' )" check, which means it's not called since 1.17
2011-02-16 18:25:44 +00:00
}
Step two in fixing OutputPage's documentation, more or less the same as r61647
2010-01-29 21:32:45 +00:00
/**
Deprecate User::getNewMessageRevisionId Uses User::getNewMessageLinks, which uses Revision objects Only caller updated to use a new private method Bug: T253949 Change-Id: I04c0558916e7216540dd7bd12a2a391a1ed7660a
2020-05-29 05:37:46 +00:00
* Get the revision ID for the last user talk page revision viewed by the talk page owner.
*
* @return int|null
*/
private function getLastSeenUserTalkRevId() {
$services = MediaWikiServices::getInstance();
$user = $this->getUser();
$userHasNewMessages = $services
->getTalkPageNotificationManager()
->userHasNewMessages( $user );
if ( !$userHasNewMessages ) {
return null;
}
$timestamp = $services
->getTalkPageNotificationManager()
->getLatestSeenMessageTimestamp( $user );
if ( !$timestamp ) {
return null;
}
$revRecord = $services->getRevisionLookup()->getRevisionByTimestamp(
$user->getTalkPage(),
$timestamp
);
Remove/simplify not trivial and unnecessary null checks Change-Id: I101dec0be2bd10e280acf6f5a0c8d7a3a81c6b1f
2023-02-14 08:01:09 +00:00
return $revRecord ? $revRecord->getId() : null;
Deprecate User::getNewMessageRevisionId Uses User::getNewMessageLinks, which uses Revision objects Only caller updated to use a new private method Bug: T253949 Change-Id: I04c0558916e7216540dd7bd12a2a391a1ed7660a
2020-05-29 05:37:46 +00:00
}
/**
Reduce a bit the coupling between Skin and OutputPage: * Removed Skin::setupUserCss() and merged it in OutputPage::buildCssLinks() and OutputPage::buildCssLinksArray(), also removed its call from ApiParse.php * Moved Skin::userCanPreview() to OutputPage since two of the three calls to that function are made from there and the last one in SkinTemplate.php is in a somewhat deprecated function * Removed the Skin parameter from OutputPage::buildCssLinks(), OutputPage::getHeadScripts(), OutputPage::getBottomScripts() and OutputPage::makeResourceLoaderLink() since we now have a context Also made ApiParse.php call createContext() instead of create a new RequestContext manually and set its Title so so that it does not rely on $wgRequest and $wgTitle to get that member objects
2011-07-22 10:45:07 +00:00
* To make it harder for someone to slip a user a fake
Generalize ResourceLoader 'excludepage' functionality There has long been a hack for previewing edits to user JS/CSS, where OutputPage would pass an 'excludepage' parameter to ResourceLoaderUserModule to tell it not to load one particular page and would instead embed that page statically. That's nice, but there are other places where we could use the same thing. This patch generalizes it: * DerivativeResourceLoaderContext may now contain a callback for mapping titles to replacement Content objects. * ResourceLoaderWikiModule::getContent() uses the overrides, and requests embedding when they're used. All subclasses in Gerrit should pick it up automatically. * OutputPage gains methods for callers to add to the override mapping, which it passes on to RL. It loses a bunch of the special casing it had for the 'user' and 'user.styles' modules. * EditPage sets the overrides on OutputPage when doing the preview, as does ApiParse for prop=headhtml. TemplateSandbox does too in I83fa0856. * OutputPage::userCanPreview() gets less specific to editing user CSS and JS, since RL now handles the embedding based on the actual modules' dependencies and EditPage only requests it on preview. ApiParse also gets a new hook to support TemplateSandbox's API integration (used in I83fa0856). Bug: T112474 Change-Id: Ib9d2ce42931c1de8372e231314a1f672d7e2ac0e
2017-02-28 20:52:17 +00:00
* JavaScript or CSS preview, a random token
Reduce a bit the coupling between Skin and OutputPage: * Removed Skin::setupUserCss() and merged it in OutputPage::buildCssLinks() and OutputPage::buildCssLinksArray(), also removed its call from ApiParse.php * Moved Skin::userCanPreview() to OutputPage since two of the three calls to that function are made from there and the last one in SkinTemplate.php is in a somewhat deprecated function * Removed the Skin parameter from OutputPage::buildCssLinks(), OutputPage::getHeadScripts(), OutputPage::getBottomScripts() and OutputPage::makeResourceLoaderLink() since we now have a context Also made ApiParse.php call createContext() instead of create a new RequestContext manually and set its Title so so that it does not rely on $wgRequest and $wgTitle to get that member objects
2011-07-22 10:45:07 +00:00
* is associated with the login session. If it's not
* passed back with the preview request, we won't render
* the code.
*
* @return bool
*/
public function userCanPreview() {
Small clean-ups for OutputPage::userCanPreview Change-Id: Ibd973750b60cbcc8d1289686de6cabcfdca5c5d9
2015-10-08 06:04:59 +00:00
$request = $this->getRequest();
Don't "preview" (i.e. execute) user JS on non-preview requests If you edit your user JS page, change the JS, and click Show preview, the new JS will be loaded ("previewed"). However, this "preview" is also engaged when you click Show changes, or click "Save page" and fail to save. Don't do that: only "preview" JS when the user has actually chosen to preview it. Change-Id: Id9bcd235d3414b68de6e5d491b7b0c4f9b16e05f
2015-11-15 01:28:25 +00:00
if (
Change trivial use of getVal('action') to getRawVal Per docs added in I18767cd809f67b, these don't need normalization as they are only compared against predefined strings, and besides are generally entered manually in a form, and even then would not require the kinds of Unicode chars that have multiple/non-normalized forms. In nearby areas to also fix some trivial cases: * getVal('title') obviously needs normalization. Use getText() to make this more obvious. * getVal() compared against simple string literals within the code obviously don't need normalization (e.g. printable === 'no'). * Change hot code in MediaWiki checking for whether 'diff' or 'oldid' are set to getCheck (which uses getRawVal) instead of getVal. As a bonus this means it now handles values like "0" correctly, which could theoretically have caused bad behaviour before. Change-Id: Ied721cfdf59c7ba11d1afa6f4cc59ede1381238e
2021-08-10 21:59:39 +00:00
$request->getRawVal( 'action' ) !== 'submit' ||
Don't "preview" (i.e. execute) user JS on non-preview requests If you edit your user JS page, change the JS, and click Show preview, the new JS will be loaded ("previewed"). However, this "preview" is also engaged when you click Show changes, or click "Save page" and fail to save. Don't do that: only "preview" JS when the user has actually chosen to preview it. Change-Id: Id9bcd235d3414b68de6e5d491b7b0c4f9b16e05f
2015-11-15 01:28:25 +00:00
!$request->wasPosted()
) {
Small clean-ups for OutputPage::userCanPreview Change-Id: Ibd973750b60cbcc8d1289686de6cabcfdca5c5d9
2015-10-08 06:04:59 +00:00
return false;
}
$user = $this->getUser();
SECURITY: Require login to preview user CSS pages Anon users have predictable edit tokens, hence someone could force an anon to execute arbitrary CSS by means of a CSRF. Bug: T133147 Change-Id: I442b2b46cadb967aaa1f35648eff183fc7eaa475
2016-08-19 20:53:52 +00:00
Swap out uses of User->isLoggedIn() with ->isRegistered() Bug: T270450 Change-Id: I90ff94a553ebae9a4ef97c25c333f067f19c327d
2020-12-17 23:10:11 +00:00
if ( !$user->isRegistered() ) {
SECURITY: Require login to preview user CSS pages Anon users have predictable edit tokens, hence someone could force an anon to execute arbitrary CSS by means of a CSRF. Bug: T133147 Change-Id: I442b2b46cadb967aaa1f35648eff183fc7eaa475
2016-08-19 20:53:52 +00:00
// Anons have predictable edit tokens
return false;
}
Revert "Use CsrfTokenSet as CSRF token source" This reverts commit 0d75fdb4f73ddbf4ad5f29e14258d58e507d55b4. Bug: T287542 Change-Id: Iedd3461869f973f8d621a39e6ad4674cbb577551
2021-08-05 06:54:11 +00:00
if ( !$user->matchEditToken( $request->getVal( 'wpEditToken' ) ) ) {
Reduce a bit the coupling between Skin and OutputPage: * Removed Skin::setupUserCss() and merged it in OutputPage::buildCssLinks() and OutputPage::buildCssLinksArray(), also removed its call from ApiParse.php * Moved Skin::userCanPreview() to OutputPage since two of the three calls to that function are made from there and the last one in SkinTemplate.php is in a somewhat deprecated function * Removed the Skin parameter from OutputPage::buildCssLinks(), OutputPage::getHeadScripts(), OutputPage::getBottomScripts() and OutputPage::makeResourceLoaderLink() since we now have a context Also made ApiParse.php call createContext() instead of create a new RequestContext manually and set its Title so so that it does not rely on $wgRequest and $wgTitle to get that member objects
2011-07-22 10:45:07 +00:00
return false;
}
Small clean-ups for OutputPage::userCanPreview Change-Id: Ibd973750b60cbcc8d1289686de6cabcfdca5c5d9
2015-10-08 06:04:59 +00:00
$title = $this->getTitle();
OutputPage: move to authority where relevant Change-Id: Iae42cc747f80cb0215f862d633e9af32fc7c8571
2021-03-04 15:55:20 +00:00
if ( !$this->getAuthority()->probablyCan( 'edit', $title ) ) {
Small clean-ups for OutputPage::userCanPreview Change-Id: Ibd973750b60cbcc8d1289686de6cabcfdca5c5d9
2015-10-08 06:04:59 +00:00
return false;
}
return true;
Reduce a bit the coupling between Skin and OutputPage: * Removed Skin::setupUserCss() and merged it in OutputPage::buildCssLinks() and OutputPage::buildCssLinksArray(), also removed its call from ApiParse.php * Moved Skin::userCanPreview() to OutputPage since two of the three calls to that function are made from there and the last one in SkinTemplate.php is in a somewhat deprecated function * Removed the Skin parameter from OutputPage::buildCssLinks(), OutputPage::getHeadScripts(), OutputPage::getBottomScripts() and OutputPage::makeResourceLoaderLink() since we now have a context Also made ApiParse.php call createContext() instead of create a new RequestContext manually and set its Title so so that it does not rely on $wgRequest and $wgTitle to get that member objects
2011-07-22 10:45:07 +00:00
}
/**
Cleanup some docs (includes/*.php) - Swap "$variable type" to "type $variable" - Added missing types - Fixed spacing inside docs - Makes beginning of @param/@return/@var/@throws in capital - Changed some types to match the more common spelling Change-Id: I783e4dbfe5f6f98b32b9a03ccf6439e13e132bcc
2014-07-24 17:42:24 +00:00
* @return array Array in format "link name or number => 'link html'".
Step two in fixing OutputPage's documentation, more or less the same as r61647
2010-01-29 21:32:45 +00:00
*/
OutputPage: Move our <meta charset="...> so it's placed above the title It's recommended for the meta charset to be placed before the <title> since <title> contains text which is inside the character set defined by the meta charset. Use of meta charset inside XHTML also seems to be redundant, not recommended, and is very likely completely ignored. Change-Id: I335b0598a9615540dc5e917682508b4a8d32d96e
2013-05-11 09:31:30 +00:00
public function getHeadLinksArray() {
Convert all array() syntax to [] Per wikitech-l consensus: https://lists.wikimedia.org/pipermail/wikitech-l/2016-February/084821.html Notes: * Disabled CallTimePassByReference due to false positives (T127163) Change-Id: I2c8ce713ce6600a0bb7bf67537c87044c7a45c4b
2016-02-17 09:09:32 +00:00
$tags = [];
Use config instead of globals for OutputPage Change-Id: I5e0ebc173631d1d1052de7ccee4ef839c7c1767f
2014-08-23 08:52:41 +00:00
$config = $this->getConfig();
Unify MonoBook <head> generation with legacy skins This was less work than I thought it would be. The only possible practical difference in MonoBook should be that some IE fixes are moved down; at a glance, I don't see how this would hurt anything, but if it does then some more hacks can be added to move them back up. A bunch of whitespace was changed in both MonoBook and legacy skins. Legacy skins had some stuff moved around in the <head>, but mainly to better match MonoBook, so they should work better if anything, not worse. dir= was added to the <body> in legacy skins, matching MonoBook. There should be no other practical differences. I would strongly encourage people to port the Modern and Vector skins to at least share the <head>-generation code like this. I'd even more strongly encourage people to totally rewrite the skin system, but that's a bigger job. :)
2009-07-07 21:49:45 +00:00
Convert all array() syntax to [] Per wikitech-l consensus: https://lists.wikimedia.org/pipermail/wikitech-l/2016-February/084821.html Notes: * Disabled CallTimePassByReference due to false positives (T127163) Change-Id: I2c8ce713ce6600a0bb7bf67537c87044c7a45c4b
2016-02-17 09:09:32 +00:00
$tags['meta-generator'] = Html::element( 'meta', [
* Moved all <link> definitions in OutputPage::getHeadLinks() instead of having them in a *lot* of different functions * Also moved there generic <meta> and removed OutputPage::addDefaultMeta() with its $called static local variable which was breaking the output when generating multiple pages on the same request (rebuildFileCache.php, dumpHTML.php) since that function could only be executed completely once for all instances, and not once per instance * Moved default module from OutputPage::output() to its own function and don't call it when executing a body only request, since it's useless in that case * Call Skin::setMembers() from Skin::initPage() instead of Skin::outputPage()
2011-04-02 18:38:42 +00:00
'name' => 'generator',
Update all use of $wgVersion to MW_VERSION Follows-up I04628de4152dd5. Bug: T212738 Change-Id: I718474ec0d9fd29ac2c05477f0f2493615d8aff5
2020-02-25 01:33:18 +00:00
'content' => 'MediaWiki ' . MW_VERSION,
Convert all array() syntax to [] Per wikitech-l consensus: https://lists.wikimedia.org/pipermail/wikitech-l/2016-February/084821.html Notes: * Disabled CallTimePassByReference due to false positives (T127163) Change-Id: I2c8ce713ce6600a0bb7bf67537c87044c7a45c4b
2016-02-17 09:09:32 +00:00
] );
here it is ... the upload-api, script-server, js2 (javascript phase2) branch merge 1st attempt. Here is a short overview of changes and associated default configuration variables (most everything is off by default) also see ~soon to be updated~: http://www.mediawiki.org/wiki/Media_Projects_Overview = Upload Improvements = ==Upload API == * Based on the early work of Bryan Tong and others it adds the upload option to the api. * We rewrite Special:Upload page to include use the new refactoring * Added in token checks in both the SpecialUpload.php page so avoids DOS / xss copy-by-url JavaScript based cross site POST file submissions == Copy by URL== $wgAllowCopyUploads = false; * http class rewrite includes a new http background download see: includes/HttpFunctions.php * spins off a php process that calls: maintenance/http_session_download.php * pushes updates to the session and gives the user a progress bar on http copy uploads from other server progress (using js2 upload interface) (if not using the js2 upload interface it does the request in-place but the download is limited to the php ini timeout time) == Firefogg == * Firefogg enables resumable upload by chunks * progress indicators and conditional invokation (js2 system) * and of-course client side transcoding. = Script Server = $wgEnableScriptLoader = false; * off by default if $wgEnableScriptLoader is turned on script files are grouped, gziped, cached etc. for more info see: http://www.mediawiki.org/wiki/Extension:ScriptLoader * Includes some early skin js include fixes (skin/script system still lots of love) * Includes a "javascript class autoloader" this is packaged into mwEmbed so that the mwEmbed library can work in stand alone mode (while retaining localization and script serving) (one such application is the make page for firefogg.org : http://www.firefogg.org/make/index.html ) * The file that contains the autojavascript loading classes is: js2/php/jsAutoloadLocalClasses.php * One can use this auto class loading dependency system with extensions and add-ons but I need to better document that. = js2 system / mwEmbed= $wgEnableJS2system = false * includes initial rewrite towards more jquery based javascript code * especially for the Special:Upload page. * Also the edit page include support for the "add-media-wizard" * includes dependency loader for javascript that optionally takes advantage of the script-loader * remote embedding of javascript interfaces (like embedding video, or commons media searching) * $wgDebugJavaScript = false; .. .this variable lets you always get "always fresh javascript". When used with the script-loader it does not minify the script-loader output. = mwEmbed = * Will commit a separate patch to oggHandler that conditionally outputs <video tag> to use the new javascript video player. ** mv_embed player includes: play-head, volume control, remote embedding, oggz-chop support across plugins. * add-media-wizard adds easy inserts of media to pages (with import) == jQuery== * we include a base install of jQuery, jQuery ui and some plugins. * all the javascript classes are in the scriptloader so its easy to load any set of jquery ui components that you may need using the script-server. You get a callback so you can then execute js with dependencies loaded. == other stuff == there is a bit more code in js2 that pertains to sequence editing, timed text display and basic image editing. We include a base import of pixastic-lib & pixastic-editor... will work with the pixastic developer to try and ensure upstream compatibility on our usage of the library for in-browser photo and sequence manipulation.
2009-07-14 23:52:14 +00:00
Use name constants to access config settings. Use name constants instead of string literals in calls to Config::get and ServiceOptions::get, when referring to core configuration variables. This protects against typos and makes the decumentation and schema declaration of the config settings discoverable. This is the first batch, only touching files directly under /includes/ Change-Id: I7252e636c7c86d950d9257b33491af492c6dd5eb
2022-04-01 15:58:32 +00:00
if ( $config->get( MainConfigNames::ReferrerPolicy ) !== false ) {
Allow fallbacks for $wgReferrerPolicy For browsers that support older versions of the Referrer Policy specification (Edge and Safari), using a value from a newer version of the specification will result in those browsers falling back to a default of "default". So allow $wgReferrerPolicy to have fallbacks if the browser does not recognize the currently set value. It will emit <meta> tags for each value in the array, but in reverse order, as browsers will use the last one that they recognize. Bug: T180921 Change-Id: Ie0f523fc6937c9ecffc8a6fc791c6b54d5a1cb06
2017-12-23 17:35:59 +00:00
// Per https://w3c.github.io/webappsec-referrer-policy/#unknown-policy-values
// fallbacks should come before the primary value so we need to reverse the array.
Use name constants to access config settings. Use name constants instead of string literals in calls to Config::get and ServiceOptions::get, when referring to core configuration variables. This protects against typos and makes the decumentation and schema declaration of the config settings discoverable. This is the first batch, only touching files directly under /includes/ Change-Id: I7252e636c7c86d950d9257b33491af492c6dd5eb
2022-04-01 15:58:32 +00:00
foreach ( array_reverse( (array)$config->get( MainConfigNames::ReferrerPolicy ) ) as $i => $policy ) {
Allow fallbacks for $wgReferrerPolicy For browsers that support older versions of the Referrer Policy specification (Edge and Safari), using a value from a newer version of the specification will result in those browsers falling back to a default of "default". So allow $wgReferrerPolicy to have fallbacks if the browser does not recognize the currently set value. It will emit <meta> tags for each value in the array, but in reverse order, as browsers will use the last one that they recognize. Bug: T180921 Change-Id: Ie0f523fc6937c9ecffc8a6fc791c6b54d5a1cb06
2017-12-23 17:35:59 +00:00
$tags["meta-referrer-$i"] = Html::element( 'meta', [
'name' => 'referrer',
'content' => $policy,
] );
}
Add option to output referrer policy meta tag Bug: T87276 Change-Id: I9ea4797c4d292cd4565f47b88e605b4e7afdb1d1
2015-01-20 23:41:41 +00:00
}
Add option to amend robots meta tag - Add private var, helper methods to format and set robots options. - Add reference to new method in OutputPage::getHeadLinksArray(). Bug: T301584 Change-Id: I2652f20c9fe5f2d3e28e329d4eddc2a2ff88cf76
2022-03-10 21:27:21 +00:00
$p = $this->getRobotsContent();
if ( $p ) {
Attempt to clean up some of the insanity in creating meta and link tags in the headers. Values are now escaped consistently, which should be a good thing. :)
2008-07-02 22:52:22 +00:00
// http://www.robotstxt.org/wc/meta-user.html
// Only show if it's different from the default robots policy
Convert all array() syntax to [] Per wikitech-l consensus: https://lists.wikimedia.org/pipermail/wikitech-l/2016-February/084821.html Notes: * Disabled CallTimePassByReference due to false positives (T127163) Change-Id: I2c8ce713ce6600a0bb7bf67537c87044c7a45c4b
2016-02-17 09:09:32 +00:00
$tags['meta-robots'] = Html::element( 'meta', [
* Moved all <link> definitions in OutputPage::getHeadLinks() instead of having them in a *lot* of different functions * Also moved there generic <meta> and removed OutputPage::addDefaultMeta() with its $called static local variable which was breaking the output when generating multiple pages on the same request (rebuildFileCache.php, dumpHTML.php) since that function could only be executed completely once for all instances, and not once per instance * Moved default module from OutputPage::output() to its own function and don't call it when executing a body only request, since it's useless in that case * Call Skin::setMembers() from Skin::initPage() instead of Skin::outputPage()
2011-04-02 18:38:42 +00:00
'name' => 'robots',
'content' => $p,
Convert all array() syntax to [] Per wikitech-l consensus: https://lists.wikimedia.org/pipermail/wikitech-l/2016-February/084821.html Notes: * Disabled CallTimePassByReference due to false positives (T127163) Change-Id: I2c8ce713ce6600a0bb7bf67537c87044c7a45c4b
2016-02-17 09:09:32 +00:00
] );
Attempt to clean up some of the insanity in creating meta and link tags in the headers. Values are now escaped consistently, which should be a good thing. :)
2008-07-02 22:52:22 +00:00
}
Disable browser data detection of telephone numbers Number ranges are common and easily mistaken as telephone numbers by the data detectors of popular browsers. To link a telephone number, use the explicit [tel:phonenumber display value] syntax. If you want to re-enable, set the BrowserFormatDetection configuration variable to false. Bug: T256758 Change-Id: I49dff3c084d8c9a66defcdae5a1ffb8f35ba8a93
2020-07-21 20:00:23 +00:00
# Browser based phonenumber detection
Use name constants to access config settings. Use name constants instead of string literals in calls to Config::get and ServiceOptions::get, when referring to core configuration variables. This protects against typos and makes the decumentation and schema declaration of the config settings discoverable. This is the first batch, only touching files directly under /includes/ Change-Id: I7252e636c7c86d950d9257b33491af492c6dd5eb
2022-04-01 15:58:32 +00:00
if ( $config->get( MainConfigNames::BrowserFormatDetection ) !== false ) {
Disable browser data detection of telephone numbers Number ranges are common and easily mistaken as telephone numbers by the data detectors of popular browsers. To link a telephone number, use the explicit [tel:phonenumber display value] syntax. If you want to re-enable, set the BrowserFormatDetection configuration variable to false. Bug: T256758 Change-Id: I49dff3c084d8c9a66defcdae5a1ffb8f35ba8a93
2020-07-21 20:00:23 +00:00
$tags['meta-format-detection'] = Html::element( 'meta', [
'name' => 'format-detection',
Use name constants to access config settings. Use name constants instead of string literals in calls to Config::get and ServiceOptions::get, when referring to core configuration variables. This protects against typos and makes the decumentation and schema declaration of the config settings discoverable. This is the first batch, only touching files directly under /includes/ Change-Id: I7252e636c7c86d950d9257b33491af492c6dd5eb
2022-04-01 15:58:32 +00:00
'content' => $config->get( MainConfigNames::BrowserFormatDetection ),
Disable browser data detection of telephone numbers Number ranges are common and easily mistaken as telephone numbers by the data detectors of popular browsers. To link a telephone number, use the explicit [tel:phonenumber display value] syntax. If you want to re-enable, set the BrowserFormatDetection configuration variable to false. Bug: T256758 Change-Id: I49dff3c084d8c9a66defcdae5a1ffb8f35ba8a93
2020-07-21 20:00:23 +00:00
] );
}
Initial revision
2003-04-14 23:10:40 +00:00
foreach ( $this->mMetatags as $tag ) {
Allow OpenGraph RFDa meta tags See: http://stackoverflow.com/questions/22350105/whats-the-difference-between-meta-name-and-meta-property And https://developers.facebook.com/docs/sharing/webmasters#markup Bug: T51859 Change-Id: Ie8fd697e8588435a69f02e779038ae5a11d67be6
2017-01-24 12:01:47 +00:00
if ( strncasecmp( $tag[0], 'http:', 5 ) === 0 ) {
Revert RDFa and Microdata improvements to addMeta in r111439 till after the git migration.
2012-02-20 22:48:43 +00:00
$a = 'http-equiv';
$tag[0] = substr( $tag[0], 5 );
Allow OpenGraph RFDa meta tags See: http://stackoverflow.com/questions/22350105/whats-the-difference-between-meta-name-and-meta-property And https://developers.facebook.com/docs/sharing/webmasters#markup Bug: T51859 Change-Id: Ie8fd697e8588435a69f02e779038ae5a11d67be6
2017-01-24 12:01:47 +00:00
} elseif ( strncasecmp( $tag[0], 'og:', 3 ) === 0 ) {
$a = 'property';
Revert RDFa and Microdata improvements to addMeta in r111439 till after the git migration.
2012-02-20 22:48:43 +00:00
} else {
$a = 'name';
Initial revision
2003-04-14 23:10:40 +00:00
}
Add function OutputPage::getHeadLinksArray() ...that returns head tags as associative array for the benefit of skin makers who don't want to output every <meta> tag in the universe. Patchset 2: make sure that array keys are unique Change-Id: I6fa3c954d603e0e401cbdb68975d536cf07e11ab
2012-04-24 17:25:06 +00:00
$tagName = "meta-{$tag[0]}";
if ( isset( $tags[$tagName] ) ) {
$tagName .= $tag[1];
}
$tags[$tagName] = Html::element( 'meta',
Convert all array() syntax to [] Per wikitech-l consensus: https://lists.wikimedia.org/pipermail/wikitech-l/2016-February/084821.html Notes: * Disabled CallTimePassByReference due to false positives (T127163) Change-Id: I2c8ce713ce6600a0bb7bf67537c87044c7a45c4b
2016-02-17 09:09:32 +00:00
[
Attempt to clean up some of the insanity in creating meta and link tags in the headers. Values are now escaped consistently, which should be a good thing. :)
2008-07-02 22:52:22 +00:00
$a => $tag[0],
*coding style cleanup (added braces, spacing tweaks etc.) *removed some unused globals
2010-05-22 12:18:22 +00:00
'content' => $tag[1]
Convert all array() syntax to [] Per wikitech-l consensus: https://lists.wikimedia.org/pipermail/wikitech-l/2016-February/084821.html Notes: * Disabled CallTimePassByReference due to false positives (T127163) Change-Id: I2c8ce713ce6600a0bb7bf67537c87044c7a45c4b
2016-02-17 09:09:32 +00:00
]
*coding style cleanup (added braces, spacing tweaks etc.) *removed some unused globals
2010-05-22 12:18:22 +00:00
);
Initial revision
2003-04-14 23:10:40 +00:00
}
* Moved all <link> definitions in OutputPage::getHeadLinks() instead of having them in a *lot* of different functions * Also moved there generic <meta> and removed OutputPage::addDefaultMeta() with its $called static local variable which was breaking the output when generating multiple pages on the same request (rebuildFileCache.php, dumpHTML.php) since that function could only be executed completely once for all instances, and not once per instance * Moved default module from OutputPage::output() to its own function and don't call it when executing a body only request, since it's useless in that case * Call Skin::setMembers() from Skin::initPage() instead of Skin::outputPage()
2011-04-02 18:38:42 +00:00
Initial revision
2003-04-14 23:10:40 +00:00
foreach ( $this->mLinktags as $tag ) {
Use Html:: functions
2009-09-18 20:10:25 +00:00
$tags[] = Html::element( 'link', $tag );
Initial revision
2003-04-14 23:10:40 +00:00
}
* remove end of line whitespace * remove empty lines at end of file * remove "?>" where still present
2008-04-14 07:45:50 +00:00
Use name constants to access config settings. Use name constants instead of string literals in calls to Config::get and ServiceOptions::get, when referring to core configuration variables. This protects against typos and makes the decumentation and schema declaration of the config settings discoverable. This is the first batch, only touching files directly under /includes/ Change-Id: I7252e636c7c86d950d9257b33491af492c6dd5eb
2022-04-01 15:58:32 +00:00
if ( $config->get( MainConfigNames::UniversalEditButton ) && $this->isArticleRelated() ) {
Remove the fallback for create rights (covered in PM edit rights check). Remove key to i18n message, no longer being used. Bug: T272079 Change-Id: Id6566e5241b06ba377c2eca083221ee02556365f
2021-06-24 14:10:43 +00:00
if ( $this->getAuthority()->probablyCan( 'edit', $this->getTitle() ) ) {
* Use local context to get messages * Simply use getLang() instead of getContext()->getLang() now that this method exists * Also changed addWikiMsgArray() to use a Message object with addWikiText() for the parse operation; removed its $options parameter, no caller was using it
2011-10-23 08:13:52 +00:00
$msg = $this->msg( 'edit' )->text();
OutputPage: Remove redundant "rel=edit" head link after "x-wiki" link Bug: T21165 Change-Id: I53f4b8cfc117012abecdfb5e751886be8ef6cda6
2021-09-21 03:15:22 +00:00
// Use mime type per https://phabricator.wikimedia.org/T21165#6946526
Convert all array() syntax to [] Per wikitech-l consensus: https://lists.wikimedia.org/pipermail/wikitech-l/2016-February/084821.html Notes: * Disabled CallTimePassByReference due to false positives (T127163) Change-Id: I2c8ce713ce6600a0bb7bf67537c87044c7a45c4b
2016-02-17 09:09:32 +00:00
$tags['universal-edit-button'] = Html::element( 'link', [
* Moved all <link> definitions in OutputPage::getHeadLinks() instead of having them in a *lot* of different functions * Also moved there generic <meta> and removed OutputPage::addDefaultMeta() with its $called static local variable which was breaking the output when generating multiple pages on the same request (rebuildFileCache.php, dumpHTML.php) since that function could only be executed completely once for all instances, and not once per instance * Moved default module from OutputPage::output() to its own function and don't call it when executing a body only request, since it's useless in that case * Call Skin::setMembers() from Skin::initPage() instead of Skin::outputPage()
2011-04-02 18:38:42 +00:00
'rel' => 'alternate',
'type' => 'application/x-wiki',
'title' => $msg,
opportunity to reuse Title::getEditURL() This will be helpful if we want to make changes to the edit URL in the future. Change-Id: I6e877d23b041a2bd1672ae804e06bf6a52e16c9f
2014-05-18 20:54:46 +00:00
'href' => $this->getTitle()->getEditURL(),
Convert all array() syntax to [] Per wikitech-l consensus: https://lists.wikimedia.org/pipermail/wikitech-l/2016-February/084821.html Notes: * Disabled CallTimePassByReference due to false positives (T127163) Change-Id: I2c8ce713ce6600a0bb7bf67537c87044c7a45c4b
2016-02-17 09:09:32 +00:00
] );
* Moved all <link> definitions in OutputPage::getHeadLinks() instead of having them in a *lot* of different functions * Also moved there generic <meta> and removed OutputPage::addDefaultMeta() with its $called static local variable which was breaking the output when generating multiple pages on the same request (rebuildFileCache.php, dumpHTML.php) since that function could only be executed completely once for all instances, and not once per instance * Moved default module from OutputPage::output() to its own function and don't call it when executing a body only request, since it's useless in that case * Call Skin::setMembers() from Skin::initPage() instead of Skin::outputPage()
2011-04-02 18:38:42 +00:00
}
}
# Generally the order of the favicon and apple-touch-icon links
# should not matter, but Konqueror (3.5.9 at least) incorrectly
# uses whichever one appears later in the HTML source. Make sure
# apple-touch-icon is specified first to avoid this.
Use name constants to access config settings. Use name constants instead of string literals in calls to Config::get and ServiceOptions::get, when referring to core configuration variables. This protects against typos and makes the decumentation and schema declaration of the config settings discoverable. This is the first batch, only touching files directly under /includes/ Change-Id: I7252e636c7c86d950d9257b33491af492c6dd5eb
2022-04-01 15:58:32 +00:00
if ( $config->get( MainConfigNames::AppleTouchIcon ) !== false ) {
Convert all array() syntax to [] Per wikitech-l consensus: https://lists.wikimedia.org/pipermail/wikitech-l/2016-February/084821.html Notes: * Disabled CallTimePassByReference due to false positives (T127163) Change-Id: I2c8ce713ce6600a0bb7bf67537c87044c7a45c4b
2016-02-17 09:09:32 +00:00
$tags['apple-touch-icon'] = Html::element( 'link', [
Pass php-strict for OutputPage Change-Id: Ia5d03277fff543ca672dc8efc7eaaeca4888ac23
2014-05-11 01:54:15 +00:00
'rel' => 'apple-touch-icon',
Use name constants to access config settings. Use name constants instead of string literals in calls to Config::get and ServiceOptions::get, when referring to core configuration variables. This protects against typos and makes the decumentation and schema declaration of the config settings discoverable. This is the first batch, only touching files directly under /includes/ Change-Id: I7252e636c7c86d950d9257b33491af492c6dd5eb
2022-04-01 15:58:32 +00:00
'href' => $config->get( MainConfigNames::AppleTouchIcon )
Convert all array() syntax to [] Per wikitech-l consensus: https://lists.wikimedia.org/pipermail/wikitech-l/2016-February/084821.html Notes: * Disabled CallTimePassByReference due to false positives (T127163) Change-Id: I2c8ce713ce6600a0bb7bf67537c87044c7a45c4b
2016-02-17 09:09:32 +00:00
] );
* Moved all <link> definitions in OutputPage::getHeadLinks() instead of having them in a *lot* of different functions * Also moved there generic <meta> and removed OutputPage::addDefaultMeta() with its $called static local variable which was breaking the output when generating multiple pages on the same request (rebuildFileCache.php, dumpHTML.php) since that function could only be executed completely once for all instances, and not once per instance * Moved default module from OutputPage::output() to its own function and don't call it when executing a body only request, since it's useless in that case * Call Skin::setMembers() from Skin::initPage() instead of Skin::outputPage()
2011-04-02 18:38:42 +00:00
}
Use name constants to access config settings. Use name constants instead of string literals in calls to Config::get and ServiceOptions::get, when referring to core configuration variables. This protects against typos and makes the decumentation and schema declaration of the config settings discoverable. This is the first batch, only touching files directly under /includes/ Change-Id: I7252e636c7c86d950d9257b33491af492c6dd5eb
2022-04-01 15:58:32 +00:00
if ( $config->get( MainConfigNames::Favicon ) !== false ) {
Convert all array() syntax to [] Per wikitech-l consensus: https://lists.wikimedia.org/pipermail/wikitech-l/2016-February/084821.html Notes: * Disabled CallTimePassByReference due to false positives (T127163) Change-Id: I2c8ce713ce6600a0bb7bf67537c87044c7a45c4b
2016-02-17 09:09:32 +00:00
$tags['favicon'] = Html::element( 'link', [
OutputPage: replace legacy link type "shortcut icon" with standard Bug: T19980 Change-Id: I6ed3417fc9efa9bab9533f155dcfe23a78d19e3b
2022-08-19 12:02:22 +00:00
'rel' => 'icon',
Use name constants to access config settings. Use name constants instead of string literals in calls to Config::get and ServiceOptions::get, when referring to core configuration variables. This protects against typos and makes the decumentation and schema declaration of the config settings discoverable. This is the first batch, only touching files directly under /includes/ Change-Id: I7252e636c7c86d950d9257b33491af492c6dd5eb
2022-04-01 15:58:32 +00:00
'href' => $config->get( MainConfigNames::Favicon )
Convert all array() syntax to [] Per wikitech-l consensus: https://lists.wikimedia.org/pipermail/wikitech-l/2016-February/084821.html Notes: * Disabled CallTimePassByReference due to false positives (T127163) Change-Id: I2c8ce713ce6600a0bb7bf67537c87044c7a45c4b
2016-02-17 09:09:32 +00:00
] );
* Moved all <link> definitions in OutputPage::getHeadLinks() instead of having them in a *lot* of different functions * Also moved there generic <meta> and removed OutputPage::addDefaultMeta() with its $called static local variable which was breaking the output when generating multiple pages on the same request (rebuildFileCache.php, dumpHTML.php) since that function could only be executed completely once for all instances, and not once per instance * Moved default module from OutputPage::output() to its own function and don't call it when executing a body only request, since it's useless in that case * Call Skin::setMembers() from Skin::initPage() instead of Skin::outputPage()
2011-04-02 18:38:42 +00:00
}
# OpenSearch description link
Convert all array() syntax to [] Per wikitech-l consensus: https://lists.wikimedia.org/pipermail/wikitech-l/2016-February/084821.html Notes: * Disabled CallTimePassByReference due to false positives (T127163) Change-Id: I2c8ce713ce6600a0bb7bf67537c87044c7a45c4b
2016-02-17 09:09:32 +00:00
$tags['opensearch'] = Html::element( 'link', [
* Moved all <link> definitions in OutputPage::getHeadLinks() instead of having them in a *lot* of different functions * Also moved there generic <meta> and removed OutputPage::addDefaultMeta() with its $called static local variable which was breaking the output when generating multiple pages on the same request (rebuildFileCache.php, dumpHTML.php) since that function could only be executed completely once for all instances, and not once per instance * Moved default module from OutputPage::output() to its own function and don't call it when executing a body only request, since it's useless in that case * Call Skin::setMembers() from Skin::initPage() instead of Skin::outputPage()
2011-04-02 18:38:42 +00:00
'rel' => 'search',
'type' => 'application/opensearchdescription+xml',
'href' => wfScript( 'opensearch_desc' ),
* Use local context to get messages * Simply use getLang() instead of getContext()->getLang() now that this method exists * Also changed addWikiMsgArray() to use a Message object with addWikiText() for the parse operation; removed its $options parameter, no caller was using it
2011-10-23 08:13:52 +00:00
'title' => $this->msg( 'opensearch-desc' )->inContentLanguage()->text(),
Convert all array() syntax to [] Per wikitech-l consensus: https://lists.wikimedia.org/pipermail/wikitech-l/2016-February/084821.html Notes: * Disabled CallTimePassByReference due to false positives (T127163) Change-Id: I2c8ce713ce6600a0bb7bf67537c87044c7a45c4b
2016-02-17 09:09:32 +00:00
] );
* Moved all <link> definitions in OutputPage::getHeadLinks() instead of having them in a *lot* of different functions * Also moved there generic <meta> and removed OutputPage::addDefaultMeta() with its $called static local variable which was breaking the output when generating multiple pages on the same request (rebuildFileCache.php, dumpHTML.php) since that function could only be executed completely once for all instances, and not once per instance * Moved default module from OutputPage::output() to its own function and don't call it when executing a body only request, since it's useless in that case * Call Skin::setMembers() from Skin::initPage() instead of Skin::outputPage()
2011-04-02 18:38:42 +00:00
Drop deprecated EnableAPI and EnableWriteAPI settings The siteinfo API response's 'writeapi' value is now hard-set to true, as are the ResourceLoader variables wgEnableAPI and wgEnableWriteAPI, to be deprecated later. Bug: T115414 Change-Id: I54ff9428b247ba203d67aba079149393f323d5a9
2017-11-20 23:50:22 +00:00
# Real Simple Discovery link, provides auto-discovery information
# for the MediaWiki API (and potentially additional custom API
# support such as WordPress or Twitter-compatible APIs for a
# blogging extension, etc)
$tags['rsd'] = Html::element( 'link', [
'rel' => 'EditURI',
'type' => 'application/rsd+xml',
// Output a protocol-relative URL here if $wgServer is protocol-relative.
// Whether RSD accepts relative or protocol-relative URLs is completely
// undocumented, though.
'href' => wfExpandUrl( wfAppendQuery(
wfScript( 'api' ),
[ 'action' => 'rsd' ] ),
PROTO_RELATIVE
),
] );
* Moved all <link> definitions in OutputPage::getHeadLinks() instead of having them in a *lot* of different functions * Also moved there generic <meta> and removed OutputPage::addDefaultMeta() with its $called static local variable which was breaking the output when generating multiple pages on the same request (rebuildFileCache.php, dumpHTML.php) since that function could only be executed completely once for all instances, and not once per instance * Moved default module from OutputPage::output() to its own function and don't call it when executing a body only request, since it's useless in that case * Call Skin::setMembers() from Skin::initPage() instead of Skin::outputPage()
2011-04-02 18:38:42 +00:00
OutputPage: Reduce line complexity of getHeadLinksArray Split this into multiple functions with multiple concerns for easier maintenance and code review. Bug: T330777 Change-Id: I80d93ec09be2df5c8b94a16140ba11da4827a42d
2023-02-08 16:52:57 +00:00
$tags = array_merge(
$tags,
$this->getHeadLinksCanonicalURLArray( $config ),
$this->getHeadLinksAlternateURLsArray(),
$this->getHeadLinksCopyrightArray( $config ),
$this->getHeadLinksSyndicationArray( $config ),
);
// Allow extensions to add, remove and/or otherwise manipulate these links
// If you want only to *add* <head> links, please use the addHeadItem()
// (or addHeadItems() for multiple items) method instead.
// This hook is provided as a last resort for extensions to modify these
// links before the output is sent to client.
$this->getHookRunner()->onOutputPageAfterGetHeadLinksArray( $tags, $this );
return $tags;
}
/**
OutputPage: Fix the behavior for canonical URL and alternate URLs This fix is for the behavior of language variants, including: * Fix the language variant behavior in meta alternate hreflang URLs. meta alternate URLs should point to respective canonical URLs. meta alternate URLs should be kept even in current page view language * Make meta canonical URL language-variant-aware. This only take effect when there's variant or uselang specified in URL, so it doesn't affect variant-non-specified (variant-auto- detected) URLs. See `LanguageConverter::getURLVariant()` . Bug: T54429 Bug: T108443 Bug: T305540 Change-Id: I5e6c8c3f7671d27aecdb52be271b34ba524fadf7 Co-Authored-by: Func <Funcer@outlook.com> Co-Authored-by: Jdlrobson <jrobson@wikimedia.org>
2023-02-09 10:43:55 +00:00
* Canonical URL and alternate URLs
OutputPage: Reduce line complexity of getHeadLinksArray Split this into multiple functions with multiple concerns for easier maintenance and code review. Bug: T330777 Change-Id: I80d93ec09be2df5c8b94a16140ba11da4827a42d
2023-02-08 16:52:57 +00:00
*
* isCanonicalUrlAction affects all requests where "setArticleRelated" is true.
* This is typically all requests that show content (query title, curid, oldid, diff),
* and all wikipage actions (edit, delete, purge, info, history etc.).
* It does not apply to file pages and special pages.
* 'history' and 'info' actions address page metadata rather than the page
* content itself, so they may not be canonicalized to the view page url.
* TODO: this logic should be owned by Action subclasses.
* See T67402
OutputPage: Fix the behavior for canonical URL and alternate URLs This fix is for the behavior of language variants, including: * Fix the language variant behavior in meta alternate hreflang URLs. meta alternate URLs should point to respective canonical URLs. meta alternate URLs should be kept even in current page view language * Make meta canonical URL language-variant-aware. This only take effect when there's variant or uselang specified in URL, so it doesn't affect variant-non-specified (variant-auto- detected) URLs. See `LanguageConverter::getURLVariant()` . Bug: T54429 Bug: T108443 Bug: T305540 Change-Id: I5e6c8c3f7671d27aecdb52be271b34ba524fadf7 Co-Authored-by: Func <Funcer@outlook.com> Co-Authored-by: Jdlrobson <jrobson@wikimedia.org>
2023-02-09 10:43:55 +00:00
*/
/**
* Get head links relating to the canonical URL
* Note: There should only be one canonical URL.
OutputPage: Reduce line complexity of getHeadLinksArray Split this into multiple functions with multiple concerns for easier maintenance and code review. Bug: T330777 Change-Id: I80d93ec09be2df5c8b94a16140ba11da4827a42d
2023-02-08 16:52:57 +00:00
* @param Config $config
* @return array
*/
private function getHeadLinksCanonicalURLArray( Config $config ) {
$tags = [];
$canonicalUrl = $this->mCanonicalUrl;
if ( $config->get( MainConfigNames::EnableCanonicalServerLink ) ) {
$query = [];
$action = $this->getContext()->getActionName();
$isCanonicalUrlAction = in_array( $action, [ 'history', 'info' ] );
OutputPage: Fix the behavior for canonical URL and alternate URLs This fix is for the behavior of language variants, including: * Fix the language variant behavior in meta alternate hreflang URLs. meta alternate URLs should point to respective canonical URLs. meta alternate URLs should be kept even in current page view language * Make meta canonical URL language-variant-aware. This only take effect when there's variant or uselang specified in URL, so it doesn't affect variant-non-specified (variant-auto- detected) URLs. See `LanguageConverter::getURLVariant()` . Bug: T54429 Bug: T108443 Bug: T305540 Change-Id: I5e6c8c3f7671d27aecdb52be271b34ba524fadf7 Co-Authored-by: Func <Funcer@outlook.com> Co-Authored-by: Jdlrobson <jrobson@wikimedia.org>
2023-02-09 10:43:55 +00:00
$services = MediaWikiServices::getInstance();
$languageConverterFactory = $services->getLanguageConverterFactory();
$isLangConversionDisabled = $languageConverterFactory->isConversionDisabled();
$pageLang = $this->getTitle()->getPageLanguage();
$pageLanguageConverter = $languageConverterFactory->getLanguageConverter( $pageLang );
$urlVariant = $pageLanguageConverter->getURLVariant();
OutputPage: Reduce line complexity of getHeadLinksArray Split this into multiple functions with multiple concerns for easier maintenance and code review. Bug: T330777 Change-Id: I80d93ec09be2df5c8b94a16140ba11da4827a42d
2023-02-08 16:52:57 +00:00
if ( $canonicalUrl !== false ) {
$canonicalUrl = wfExpandUrl( $canonicalUrl, PROTO_CANONICAL );
} elseif ( $this->isArticleRelated() ) {
if ( $isCanonicalUrlAction ) {
$query['action'] = $action;
OutputPage: Fix the behavior for canonical URL and alternate URLs This fix is for the behavior of language variants, including: * Fix the language variant behavior in meta alternate hreflang URLs. meta alternate URLs should point to respective canonical URLs. meta alternate URLs should be kept even in current page view language * Make meta canonical URL language-variant-aware. This only take effect when there's variant or uselang specified in URL, so it doesn't affect variant-non-specified (variant-auto- detected) URLs. See `LanguageConverter::getURLVariant()` . Bug: T54429 Bug: T108443 Bug: T305540 Change-Id: I5e6c8c3f7671d27aecdb52be271b34ba524fadf7 Co-Authored-by: Func <Funcer@outlook.com> Co-Authored-by: Jdlrobson <jrobson@wikimedia.org>
2023-02-09 10:43:55 +00:00
} elseif ( !$isLangConversionDisabled && $urlVariant ) {
# T54429, T108443: Making canonical URL language-variant-aware.
$query['variant'] = $urlVariant;
OutputPage: Reduce line complexity of getHeadLinksArray Split this into multiple functions with multiple concerns for easier maintenance and code review. Bug: T330777 Change-Id: I80d93ec09be2df5c8b94a16140ba11da4827a42d
2023-02-08 16:52:57 +00:00
}
$canonicalUrl = $this->getTitle()->getCanonicalURL( $query );
} else {
$reqUrl = $this->getRequest()->getRequestURL();
$canonicalUrl = wfExpandUrl( $reqUrl, PROTO_CANONICAL );
}
}
if ( $canonicalUrl !== false ) {
$tags['link-canonical'] = Html::element( 'link', [
'rel' => 'canonical',
'href' => $canonicalUrl
] );
}
return $tags;
}
/**
* Get head links relating to alternate URL(s) in languages including language variants
* Output fully-qualified URL since meta alternate URLs must be fully-qualified
* Per https://developers.google.com/search/docs/advanced/crawling/localized-versions
* See T294716
*
* @return array
*/
private function getHeadLinksAlternateURLsArray() {
$tags = [];
$languageUrls = [];
OutputPage: Fix the behavior for canonical URL and alternate URLs This fix is for the behavior of language variants, including: * Fix the language variant behavior in meta alternate hreflang URLs. meta alternate URLs should point to respective canonical URLs. meta alternate URLs should be kept even in current page view language * Make meta canonical URL language-variant-aware. This only take effect when there's variant or uselang specified in URL, so it doesn't affect variant-non-specified (variant-auto- detected) URLs. See `LanguageConverter::getURLVariant()` . Bug: T54429 Bug: T108443 Bug: T305540 Change-Id: I5e6c8c3f7671d27aecdb52be271b34ba524fadf7 Co-Authored-by: Func <Funcer@outlook.com> Co-Authored-by: Jdlrobson <jrobson@wikimedia.org>
2023-02-09 10:43:55 +00:00
$action = $this->getContext()->getActionName();
$isCanonicalUrlAction = in_array( $action, [ 'history', 'info' ] );
Replace $wgDisable{Lang,Title}Conversion with LanguageConverterFactory methods Replace direct access to $wgDisableLangConversion with LanguageConverterFactory::isConversionDisabled(), and replace direct access to $wgDisableTitleConversion with LanguageConverterFactory::isTitleConversionDisabled(). However, most places that check ::isTitleConversionDisabled() actually want ::isLinkConversionDisabled(), so add that too (and deprecate isTitleConversionDisabled()). Code search: https://codesearch.wmcloud.org/search/?q=Disable%28Lang|Title%29Conversion&i=nope&files=&repos= This change removes a number of spurious dependencies on the global configuration and reduces code duplication (for example, if the logic for disabling language conversion were ever to change). Depends-On: I6fa8230ae97b0e34c381003548e61f9b7387d363 Change-Id: Icc4687638ff1815003dd903854efdbd904854f1e
2020-11-25 04:35:49 +00:00
$services = MediaWikiServices::getInstance();
Replace deprecated function in OutputPage Now mocking LanguageConverter Change-Id: I25f2f90d8dcf2cb09429e01873f1982dca199976
2020-11-03 17:18:15 +00:00
$languageConverterFactory = $services->getLanguageConverterFactory();
OutputPage: Reduce line complexity of getHeadLinksArray Split this into multiple functions with multiple concerns for easier maintenance and code review. Bug: T330777 Change-Id: I80d93ec09be2df5c8b94a16140ba11da4827a42d
2023-02-08 16:52:57 +00:00
$isLangConversionDisabled = $languageConverterFactory->isConversionDisabled();
$pageLang = $this->getTitle()->getPageLanguage();
$pageLanguageConverter = $languageConverterFactory->getLanguageConverter( $pageLang );
# Language variants
OutputPage: Fix the behavior for canonical URL and alternate URLs This fix is for the behavior of language variants, including: * Fix the language variant behavior in meta alternate hreflang URLs. meta alternate URLs should point to respective canonical URLs. meta alternate URLs should be kept even in current page view language * Make meta canonical URL language-variant-aware. This only take effect when there's variant or uselang specified in URL, so it doesn't affect variant-non-specified (variant-auto- detected) URLs. See `LanguageConverter::getURLVariant()` . Bug: T54429 Bug: T108443 Bug: T305540 Change-Id: I5e6c8c3f7671d27aecdb52be271b34ba524fadf7 Co-Authored-by: Func <Funcer@outlook.com> Co-Authored-by: Jdlrobson <jrobson@wikimedia.org>
2023-02-09 10:43:55 +00:00
if (
$this->isArticleRelated() &&
!$isCanonicalUrlAction &&
$pageLanguageConverter->hasVariants() &&
!$isLangConversionDisabled
) {
OutputPage: Reduce line complexity of getHeadLinksArray Split this into multiple functions with multiple concerns for easier maintenance and code review. Bug: T330777 Change-Id: I80d93ec09be2df5c8b94a16140ba11da4827a42d
2023-02-08 16:52:57 +00:00
$variants = $pageLanguageConverter->getVariants();
foreach ( $variants as $variant ) {
$bcp47 = LanguageCode::bcp47( $variant );
$languageUrls[$bcp47] = $this->getTitle()
->getFullURL( [ 'variant' => $variant ], false, PROTO_CURRENT );
}
}
# TODO: Interlanguage links
if ( $languageUrls ) {
# Force the alternate URL of page language code to be self.
# T123901, T305540, T108443: Override mixed-variant variant link in language variant links.
$currentUrl = $this->getTitle()->getFullURL( [], false, PROTO_CURRENT );
$pageLangCodeBcp47 = LanguageCode::bcp47( $pageLang->getCode() );
$languageUrls[$pageLangCodeBcp47] = $currentUrl;
ksort( $languageUrls );
# Also add x-default link per https://support.google.com/webmasters/answer/189077?hl=en
$languageUrls['x-default'] = $currentUrl;
# Process all of language variants and interlanguage links
foreach ( $languageUrls as $bcp47 => $languageUrl ) {
$bcp47lowercase = strtolower( $bcp47 );
$tags['link-alternate-language-' . $bcp47lowercase] = Html::element( 'link', [
OutputPage: Only set <link hreflang=x-default> on wiki/pages with variants Follows-up 91d8a51825. Having the default variant advertised seems useful, however for wikis and/or pages that have no variants or translated versions, outputting this header doesn't seem useful. https://support.google.com/webmasters/answer/189077 doesn't justify sending it unconditionally. To try out, setting $wgLanguageCode = 'zh'; will result in all pages having hreflang=x-default and all variants. And setting to 'en' or 'de' will result in none of those link tags. Change-Id: I21cd072534ae1df960209e657b19c96889ece27c
2015-07-07 10:32:03 +00:00
'rel' => 'alternate',
OutputPage: Reduce line complexity of getHeadLinksArray Split this into multiple functions with multiple concerns for easier maintenance and code review. Bug: T330777 Change-Id: I80d93ec09be2df5c8b94a16140ba11da4827a42d
2023-02-08 16:52:57 +00:00
'hreflang' => $bcp47,
'href' => $languageUrl,
] );
* Moved all <link> definitions in OutputPage::getHeadLinks() instead of having them in a *lot* of different functions * Also moved there generic <meta> and removed OutputPage::addDefaultMeta() with its $called static local variable which was breaking the output when generating multiple pages on the same request (rebuildFileCache.php, dumpHTML.php) since that function could only be executed completely once for all instances, and not once per instance * Moved default module from OutputPage::output() to its own function and don't call it when executing a body only request, since it's useless in that case * Call Skin::setMembers() from Skin::initPage() instead of Skin::outputPage()
2011-04-02 18:38:42 +00:00
}
}
OutputPage: Reduce line complexity of getHeadLinksArray Split this into multiple functions with multiple concerns for easier maintenance and code review. Bug: T330777 Change-Id: I80d93ec09be2df5c8b94a16140ba11da4827a42d
2023-02-08 16:52:57 +00:00
return $tags;
}
/**
* Get head links relating to copyright
*
* @param Config $config
* @return array
*/
private function getHeadLinksCopyrightArray( Config $config ) {
$tags = [];
Fix rel="copyright" for ApiHelp Bug: T94038 Bug: T93995 Change-Id: I06e9bb6797c4e4ee00e5d72eda1e273800b2c7f4
2015-06-08 13:39:06 +00:00
if ( $this->copyrightUrl !== null ) {
$copyright = $this->copyrightUrl;
} else {
$copyright = '';
Use name constants to access config settings. Use name constants instead of string literals in calls to Config::get and ServiceOptions::get, when referring to core configuration variables. This protects against typos and makes the decumentation and schema declaration of the config settings discoverable. This is the first batch, only touching files directly under /includes/ Change-Id: I7252e636c7c86d950d9257b33491af492c6dd5eb
2022-04-01 15:58:32 +00:00
if ( $config->get( MainConfigNames::RightsPage ) ) {
$copy = Title::newFromText( $config->get( MainConfigNames::RightsPage ) );
* Moved all <link> definitions in OutputPage::getHeadLinks() instead of having them in a *lot* of different functions * Also moved there generic <meta> and removed OutputPage::addDefaultMeta() with its $called static local variable which was breaking the output when generating multiple pages on the same request (rebuildFileCache.php, dumpHTML.php) since that function could only be executed completely once for all instances, and not once per instance * Moved default module from OutputPage::output() to its own function and don't call it when executing a body only request, since it's useless in that case * Call Skin::setMembers() from Skin::initPage() instead of Skin::outputPage()
2011-04-02 18:38:42 +00:00
Fix rel="copyright" for ApiHelp Bug: T94038 Bug: T93995 Change-Id: I06e9bb6797c4e4ee00e5d72eda1e273800b2c7f4
2015-06-08 13:39:06 +00:00
if ( $copy ) {
$copyright = $copy->getLocalURL();
}
* Moved all <link> definitions in OutputPage::getHeadLinks() instead of having them in a *lot* of different functions * Also moved there generic <meta> and removed OutputPage::addDefaultMeta() with its $called static local variable which was breaking the output when generating multiple pages on the same request (rebuildFileCache.php, dumpHTML.php) since that function could only be executed completely once for all instances, and not once per instance * Moved default module from OutputPage::output() to its own function and don't call it when executing a body only request, since it's useless in that case * Call Skin::setMembers() from Skin::initPage() instead of Skin::outputPage()
2011-04-02 18:38:42 +00:00
}
Use name constants to access config settings. Use name constants instead of string literals in calls to Config::get and ServiceOptions::get, when referring to core configuration variables. This protects against typos and makes the decumentation and schema declaration of the config settings discoverable. This is the first batch, only touching files directly under /includes/ Change-Id: I7252e636c7c86d950d9257b33491af492c6dd5eb
2022-04-01 15:58:32 +00:00
if ( !$copyright && $config->get( MainConfigNames::RightsUrl ) ) {
$copyright = $config->get( MainConfigNames::RightsUrl );
Fix rel="copyright" for ApiHelp Bug: T94038 Bug: T93995 Change-Id: I06e9bb6797c4e4ee00e5d72eda1e273800b2c7f4
2015-06-08 13:39:06 +00:00
}
* Moved all <link> definitions in OutputPage::getHeadLinks() instead of having them in a *lot* of different functions * Also moved there generic <meta> and removed OutputPage::addDefaultMeta() with its $called static local variable which was breaking the output when generating multiple pages on the same request (rebuildFileCache.php, dumpHTML.php) since that function could only be executed completely once for all instances, and not once per instance * Moved default module from OutputPage::output() to its own function and don't call it when executing a body only request, since it's useless in that case * Call Skin::setMembers() from Skin::initPage() instead of Skin::outputPage()
2011-04-02 18:38:42 +00:00
}
if ( $copyright ) {
Convert all array() syntax to [] Per wikitech-l consensus: https://lists.wikimedia.org/pipermail/wikitech-l/2016-February/084821.html Notes: * Disabled CallTimePassByReference due to false positives (T127163) Change-Id: I2c8ce713ce6600a0bb7bf67537c87044c7a45c4b
2016-02-17 09:09:32 +00:00
$tags['copyright'] = Html::element( 'link', [
Use <link rel="license"> instead of <link rel="copyright"> "copyright" is not in the list of the allowed values for the attribut rel: https://w3c.github.io/html/single-page.html#allowed-keywords-and-their-meanings Bug: T167284 Change-Id: I09205bb146001d1d1d90e6a0f25d21851f3f697c
2017-06-07 12:11:08 +00:00
'rel' => 'license',
OutputPage: Reduce line complexity of getHeadLinksArray Split this into multiple functions with multiple concerns for easier maintenance and code review. Bug: T330777 Change-Id: I80d93ec09be2df5c8b94a16140ba11da4827a42d
2023-02-08 16:52:57 +00:00
'href' => $copyright
] );
* Moved all <link> definitions in OutputPage::getHeadLinks() instead of having them in a *lot* of different functions * Also moved there generic <meta> and removed OutputPage::addDefaultMeta() with its $called static local variable which was breaking the output when generating multiple pages on the same request (rebuildFileCache.php, dumpHTML.php) since that function could only be executed completely once for all instances, and not once per instance * Moved default module from OutputPage::output() to its own function and don't call it when executing a body only request, since it's useless in that case * Call Skin::setMembers() from Skin::initPage() instead of Skin::outputPage()
2011-04-02 18:38:42 +00:00
}
OutputPage: Reduce line complexity of getHeadLinksArray Split this into multiple functions with multiple concerns for easier maintenance and code review. Bug: T330777 Change-Id: I80d93ec09be2df5c8b94a16140ba11da4827a42d
2023-02-08 16:52:57 +00:00
return $tags;
}
Add a new hook AfterBuildFeedLinks to manipulate feed links Some extensions may want to replace, remove or manipulate a specific feed link after it was build. For this purpose: Add a new hook to enable this. The hook isn't meant for adding new feed links, this is already easily possible using OutputPage::addFeedLink(). Bug: T121774 Change-Id: If48bd07a0e7e67fd1c1a94ebaa28ca9dae146c55
2015-12-18 19:09:34 +00:00
OutputPage: Reduce line complexity of getHeadLinksArray Split this into multiple functions with multiple concerns for easier maintenance and code review. Bug: T330777 Change-Id: I80d93ec09be2df5c8b94a16140ba11da4827a42d
2023-02-08 16:52:57 +00:00
/**
* Get head links relating to syndication feeds.
*
* @param Config $config
* @return array
*/
private function getHeadLinksSyndicationArray( Config $config ) {
if ( !$config->get( MainConfigNames::Feed ) ) {
return [];
}
* remove end of line whitespace * remove empty lines at end of file * remove "?>" where still present
2008-04-14 07:45:50 +00:00
OutputPage: Reduce line complexity of getHeadLinksArray Split this into multiple functions with multiple concerns for easier maintenance and code review. Bug: T330777 Change-Id: I80d93ec09be2df5c8b94a16140ba11da4827a42d
2023-02-08 16:52:57 +00:00
$tags = [];
$feedLinks = [];
foreach ( $this->getSyndicationLinks() as $format => $link ) {
# Use the page name for the title. In principle, this could
# lead to issues with having the same name for different feeds
# corresponding to the same page, but we can't avoid that at
# this low a level.
$feedLinks[] = $this->feedLink(
$format,
$link,
# Used messages: 'page-rss-feed' and 'page-atom-feed' (for an easier grep)
$this->msg(
"page-{$format}-feed", $this->getTitle()->getPrefixedText()
)->text()
);
}
# Recent changes feed should appear on every page (except recentchanges,
# that would be redundant). Put it after the per-page feed to avoid
# changing existing behavior. It's still available, probably via a
# menu in your browser. Some sites might have a different feed they'd
# like to promote instead of the RC feed (maybe like a "Recent New Articles"
# or "Breaking news" one). For this, we see if $wgOverrideSiteFeed is defined.
# If so, use it instead.
$sitename = $config->get( MainConfigNames::Sitename );
$overrideSiteFeed = $config->get( MainConfigNames::OverrideSiteFeed );
if ( $overrideSiteFeed ) {
foreach ( $overrideSiteFeed as $type => $feedUrl ) {
// Note, this->feedLink escapes the url.
Add a new hook AfterBuildFeedLinks to manipulate feed links Some extensions may want to replace, remove or manipulate a specific feed link after it was build. For this purpose: Add a new hook to enable this. The hook isn't meant for adding new feed links, this is already easily possible using OutputPage::addFeedLink(). Bug: T121774 Change-Id: If48bd07a0e7e67fd1c1a94ebaa28ca9dae146c55
2015-12-18 19:09:34 +00:00
$feedLinks[] = $this->feedLink(
OutputPage: Reduce line complexity of getHeadLinksArray Split this into multiple functions with multiple concerns for easier maintenance and code review. Bug: T330777 Change-Id: I80d93ec09be2df5c8b94a16140ba11da4827a42d
2023-02-08 16:52:57 +00:00
$type,
$feedUrl,
$this->msg( "site-{$type}-feed", $sitename )->text()
*coding style cleanup (added braces, spacing tweaks etc.) *removed some unused globals
2010-05-22 12:18:22 +00:00
);
Introducing $wgFeed variable. Allows tuning sydication feeds off, when desired.
2008-02-22 12:33:51 +00:00
}
OutputPage: Reduce line complexity of getHeadLinksArray Split this into multiple functions with multiple concerns for easier maintenance and code review. Bug: T330777 Change-Id: I80d93ec09be2df5c8b94a16140ba11da4827a42d
2023-02-08 16:52:57 +00:00
} elseif ( !$this->getTitle()->isSpecial( 'Recentchanges' ) ) {
$rctitle = SpecialPage::getTitleFor( 'Recentchanges' );
foreach ( $this->getAdvertisedFeedTypes() as $format ) {
$feedLinks[] = $this->feedLink(
$format,
$rctitle->getLocalURL( [ 'feed' => $format ] ),
# For grep: 'site-rss-feed', 'site-atom-feed'
$this->msg( "site-{$format}-feed", $sitename )->text()
);
(bug 14500) - The sitefeed (RC) shouldn't show up on Recentchanges itself.
2008-06-12 00:51:04 +00:00
}
Autodiscovery <link> for RSS feed. Added helper functions for query stuff: appendQuery() and escapeAppendQuery() to WebRequest.
2004-03-19 08:05:36 +00:00
}
Introduce $wgEnableCanonicalServerLink (bug 43466) If enabled, send a rel=canonical link on every page indicating the correct server to use. I tested all three callers: Article, ImagePage and the variant feature in getHeadLinksArray(). Change-Id: I59b7c1d5589825ea390967036190218d5ce2db88
2012-12-31 01:11:43 +00:00
OutputPage: Reduce line complexity of getHeadLinksArray Split this into multiple functions with multiple concerns for easier maintenance and code review. Bug: T330777 Change-Id: I80d93ec09be2df5c8b94a16140ba11da4827a42d
2023-02-08 16:52:57 +00:00
# Allow extensions to change the list pf feeds. This hook is primarily for changing,
# manipulating or removing existing feed tags. If you want to add new feeds, you should
# use OutputPage::addFeedLink() instead.
$this->getHookRunner()->onAfterBuildFeedLinks( $feedLinks );
Introduce $wgEnableCanonicalServerLink (bug 43466) If enabled, send a rel=canonical link on every page indicating the correct server to use. I tested all three callers: Article, ImagePage and the variant feature in getHeadLinksArray(). Change-Id: I59b7c1d5589825ea390967036190218d5ce2db88
2012-12-31 01:11:43 +00:00
OutputPage: Reduce line complexity of getHeadLinksArray Split this into multiple functions with multiple concerns for easier maintenance and code review. Bug: T330777 Change-Id: I80d93ec09be2df5c8b94a16140ba11da4827a42d
2023-02-08 16:52:57 +00:00
$tags += $feedLinks;
New 'OutputPageAfterGetHeadLinksArray' hook, allowing extensions to modify the return value of OutputPage#getHeadLinksArray Code from wikiHow codebase, where this hook is used by the following extensions: * AlternateDomain -- used to remove certain links altogether and change the contents of other elements (e.g. <meta description="..." />) * hooks (PageHooks) -- used to hide certain links for anons on noindexed pages to avoid leaking article info to Googlebot * QADomain -- used to remove certain elements and correct <meta keywords="..." /> tags not to mention "wikiHow" if that string is present * search (LSearch) -- used to remove canonical URL on Special:LSearch for SEO Change-Id: I4a9ceb343bb5c0b4eb79e4589d36c3790938f8a9
2018-07-14 22:54:40 +00:00
Add function OutputPage::getHeadLinksArray() ...that returns head tags as associative array for the benefit of skin makers who don't want to output every <meta> tag in the universe. Patchset 2: make sure that array keys are unique Change-Id: I6fa3c954d603e0e401cbdb68975d536cf07e11ab
2012-04-24 17:25:06 +00:00
return $tags;
}
* Use only default options when generating RSS and Atom syndication links. This should help prevent infinite link loops that some software may follow, and will generally keep feed behavior cleaner. Combined duplicate link expansion from OutputPage and SkinTemplate into OutputPage::getSyndicationLinks()
2008-01-12 00:13:53 +00:00
/**
escape tags and entity in doxygen comments When inserting XML elements inline <such as this one>, doxygen chokes about it not being known. Simply enclosing the tag in double quotes prevents doxygen from emitting a warning. Also enclosed a few invalid functions calls such as \. and double quoted the HTML entities such as &foobar; Change-Id: I4019637145e683c2bec3d17b2fd98b0c50a932f1
2012-07-10 12:48:06 +00:00
* Generate a "<link rel/>" for a feed.
Step two in fixing OutputPage's documentation, more or less the same as r61647
2010-01-29 21:32:45 +00:00
*
Removed unneeded spaces and colons in @param and friends Also swapped some "$var type" to "type $var" or added missing types before the $var. Changed some other types to match the more common spelling. Makes beginning of some text in captial. Change-Id: Ic36c8c7820a6c2d603f1138130670c6bf6a1ca59
2014-04-08 15:29:17 +00:00
* @param string $type Feed type
Fixed @param tags to conform with Doxygen format. Doxygen expects parameter types to come before the parameter name in @param tags. Used a quick regex to switch everything around where possible. This only fixes cases where a primitve variable (or a primitive followed by other types) is the variable type. Other cases will need to be fixed manually. Change-Id: Ic59fd20856eb0489d70f3469a56ebce0efb3db13
2013-03-11 17:15:01 +00:00
* @param string $url URL to the feed
Removed unneeded spaces and colons in @param and friends Also swapped some "$var type" to "type $var" or added missing types before the $var. Changed some other types to match the more common spelling. Makes beginning of some text in captial. Change-Id: Ic36c8c7820a6c2d603f1138130670c6bf6a1ca59
2014-04-08 15:29:17 +00:00
* @param string $text Value of the "title" attribute
* @return string HTML fragment
Clean up r26058, r26059 a bit... * We've got these nice functions to do escaping of XML for us, use them. :) * Swapped order so per-page feeds (history, newpages, etc) show up first I'm still not too thrilled with the message texts on the titles.
2007-09-24 18:25:56 +00:00
*/
private function feedLink( $type, $url, $text ) {
Convert all array() syntax to [] Per wikitech-l consensus: https://lists.wikimedia.org/pipermail/wikitech-l/2016-February/084821.html Notes: * Disabled CallTimePassByReference due to false positives (T127163) Change-Id: I2c8ce713ce6600a0bb7bf67537c87044c7a45c4b
2016-02-17 09:09:32 +00:00
return Html::element( 'link', [
Clean up r26058, r26059 a bit... * We've got these nice functions to do escaping of XML for us, use them. :) * Swapped order so per-page feeds (history, newpages, etc) show up first I'm still not too thrilled with the message texts on the titles.
2007-09-24 18:25:56 +00:00
'rel' => 'alternate',
'type' => "application/$type+xml",
'title' => $text,
Convert all array() syntax to [] Per wikitech-l consensus: https://lists.wikimedia.org/pipermail/wikitech-l/2016-February/084821.html Notes: * Disabled CallTimePassByReference due to false positives (T127163) Change-Id: I2c8ce713ce6600a0bb7bf67537c87044c7a45c4b
2016-02-17 09:09:32 +00:00
'href' => $url ]
*coding style cleanup (added braces, spacing tweaks etc.) *removed some unused globals
2010-05-22 12:18:22 +00:00
);
Clean up r26058, r26059 a bit... * We've got these nice functions to do escaping of XML for us, use them. :) * Swapped order so per-page feeds (history, newpages, etc) show up first I'm still not too thrilled with the message texts on the titles.
2007-09-24 18:25:56 +00:00
}
remove wfdebug in pagehistory set ETag header on article views
2005-07-01 00:03:31 +00:00
Tweaks for skins: * Moved SkinTemplate::addStyle() and related stuff to OutputPage so that it can be used non-SkinTemplate skins and avoid duplication with the actual OutputPage::addStyle() (the two functions have the same format). * Non-SkinTemplate skins now also load their CSS with <link> tags instead of @import. * Moved SkinTemplate::setupUserCss() to Skin. * Merged action=raw&gen=(js|css) for SkinTemplate and non-SkinTemplate skins, renamed functions to Skin::generateUserJs() and Skin::generateUserStyleSheet() and dropped a lot of cascading call which is a bit incomprehensible.
2008-08-21 14:09:57 +00:00
/**
* Add a local or specified stylesheet, with the given media options.
OutputPage: documentation for (semi-)deprecated methods Deprecated: addExtensionStyle, getExtStyle These had been deprecated unofficially for a while now. ResourceLoader is preferred. Internal/Unpreferred: addHeadItem, addScript, addScriptFile, addInlineScript, addStyle, addInlineStyle. All should be used as littel as possible, in favor of ResourceLoader and a few other higher level accessors. They are not truly deprecated, but should not really be used unless you have a really good reason. Also added a few missing wfDeprecated calls to methods marked as deprecated in the phpdoc. Bug: T89955 Change-Id: I472d906b6de16257fa34a2f9bfe9e4eb284d02b4
2015-10-31 15:16:58 +00:00
* Internal use only. Use OutputPage::addModuleStyles() if possible.
Tweaks for skins: * Moved SkinTemplate::addStyle() and related stuff to OutputPage so that it can be used non-SkinTemplate skins and avoid duplication with the actual OutputPage::addStyle() (the two functions have the same format). * Non-SkinTemplate skins now also load their CSS with <link> tags instead of @import. * Moved SkinTemplate::setupUserCss() to Skin. * Merged action=raw&gen=(js|css) for SkinTemplate and non-SkinTemplate skins, renamed functions to Skin::generateUserJs() and Skin::generateUserStyleSheet() and dropped a lot of cascading call which is a bit incomprehensible.
2008-08-21 14:09:57 +00:00
*
Fixed @param tags to conform with Doxygen format. Doxygen expects parameter types to come before the parameter name in @param tags. Used a quick regex to switch everything around where possible. This only fixes cases where a primitve variable (or a primitive followed by other types) is the variable type. Other cases will need to be fixed manually. Change-Id: Ic59fd20856eb0489d70f3469a56ebce0efb3db13
2013-03-11 17:15:01 +00:00
* @param string $style URL to the file
Cleanup some docs (includes/*.php) - Swap "$variable type" to "type $variable" - Added missing types - Fixed spacing inside docs - Makes beginning of @param/@return/@var/@throws in capital - Changed some types to match the more common spelling Change-Id: I783e4dbfe5f6f98b32b9a03ccf6439e13e132bcc
2014-07-24 17:42:24 +00:00
* @param string $media To specify a media type, 'screen', 'printable', 'handheld' or any.
* @param string $condition For IE conditional comments, specifying an IE version
* @param string $dir Set to 'rtl' or 'ltr' for direction-specific sheets
Tweaks for skins: * Moved SkinTemplate::addStyle() and related stuff to OutputPage so that it can be used non-SkinTemplate skins and avoid duplication with the actual OutputPage::addStyle() (the two functions have the same format). * Non-SkinTemplate skins now also load their CSS with <link> tags instead of @import. * Moved SkinTemplate::setupUserCss() to Skin. * Merged action=raw&gen=(js|css) for SkinTemplate and non-SkinTemplate skins, renamed functions to Skin::generateUserJs() and Skin::generateUserStyleSheet() and dropped a lot of cascading call which is a bit incomprehensible.
2008-08-21 14:09:57 +00:00
*/
*coding style cleanup (added braces, spacing tweaks etc.) *removed some unused globals
2010-05-22 12:18:22 +00:00
public function addStyle( $style, $media = '', $condition = '', $dir = '' ) {
Convert all array() syntax to [] Per wikitech-l consensus: https://lists.wikimedia.org/pipermail/wikitech-l/2016-February/084821.html Notes: * Disabled CallTimePassByReference due to false positives (T127163) Change-Id: I2c8ce713ce6600a0bb7bf67537c87044c7a45c4b
2016-02-17 09:09:32 +00:00
$options = [];
else if -> elseif Also fixed some spacing while at it Added some braces for one line statements Change-Id: Iebfa2b17091509daabc76248121c7763e2295d20
2013-04-17 14:52:47 +00:00
if ( $media ) {
Tweaks for skins: * Moved SkinTemplate::addStyle() and related stuff to OutputPage so that it can be used non-SkinTemplate skins and avoid duplication with the actual OutputPage::addStyle() (the two functions have the same format). * Non-SkinTemplate skins now also load their CSS with <link> tags instead of @import. * Moved SkinTemplate::setupUserCss() to Skin. * Merged action=raw&gen=(js|css) for SkinTemplate and non-SkinTemplate skins, renamed functions to Skin::generateUserJs() and Skin::generateUserStyleSheet() and dropped a lot of cascading call which is a bit incomprehensible.
2008-08-21 14:09:57 +00:00
$options['media'] = $media;
*coding style cleanup (added braces, spacing tweaks etc.) *removed some unused globals
2010-05-22 12:18:22 +00:00
}
else if -> elseif Also fixed some spacing while at it Added some braces for one line statements Change-Id: Iebfa2b17091509daabc76248121c7763e2295d20
2013-04-17 14:52:47 +00:00
if ( $condition ) {
Tweaks for skins: * Moved SkinTemplate::addStyle() and related stuff to OutputPage so that it can be used non-SkinTemplate skins and avoid duplication with the actual OutputPage::addStyle() (the two functions have the same format). * Non-SkinTemplate skins now also load their CSS with <link> tags instead of @import. * Moved SkinTemplate::setupUserCss() to Skin. * Merged action=raw&gen=(js|css) for SkinTemplate and non-SkinTemplate skins, renamed functions to Skin::generateUserJs() and Skin::generateUserStyleSheet() and dropped a lot of cascading call which is a bit incomprehensible.
2008-08-21 14:09:57 +00:00
$options['condition'] = $condition;
*coding style cleanup (added braces, spacing tweaks etc.) *removed some unused globals
2010-05-22 12:18:22 +00:00
}
else if -> elseif Also fixed some spacing while at it Added some braces for one line statements Change-Id: Iebfa2b17091509daabc76248121c7763e2295d20
2013-04-17 14:52:47 +00:00
if ( $dir ) {
Tweaks for skins: * Moved SkinTemplate::addStyle() and related stuff to OutputPage so that it can be used non-SkinTemplate skins and avoid duplication with the actual OutputPage::addStyle() (the two functions have the same format). * Non-SkinTemplate skins now also load their CSS with <link> tags instead of @import. * Moved SkinTemplate::setupUserCss() to Skin. * Merged action=raw&gen=(js|css) for SkinTemplate and non-SkinTemplate skins, renamed functions to Skin::generateUserJs() and Skin::generateUserStyleSheet() and dropped a lot of cascading call which is a bit incomprehensible.
2008-08-21 14:09:57 +00:00
$options['dir'] = $dir;
*coding style cleanup (added braces, spacing tweaks etc.) *removed some unused globals
2010-05-22 12:18:22 +00:00
}
Tweaks for skins: * Moved SkinTemplate::addStyle() and related stuff to OutputPage so that it can be used non-SkinTemplate skins and avoid duplication with the actual OutputPage::addStyle() (the two functions have the same format). * Non-SkinTemplate skins now also load their CSS with <link> tags instead of @import. * Moved SkinTemplate::setupUserCss() to Skin. * Merged action=raw&gen=(js|css) for SkinTemplate and non-SkinTemplate skins, renamed functions to Skin::generateUserJs() and Skin::generateUserStyleSheet() and dropped a lot of cascading call which is a bit incomprehensible.
2008-08-21 14:09:57 +00:00
$this->styles[$style] = $options;
}
follow-up to r53282: *coding style tweaks *added __METHOD__ to some wfDebug calls *removed php ending tag from mwScriptLoader.php *some doxygen docs added
2009-07-15 00:55:58 +00:00
/**
* Adds inline CSS styles
OutputPage: documentation for (semi-)deprecated methods Deprecated: addExtensionStyle, getExtStyle These had been deprecated unofficially for a while now. ResourceLoader is preferred. Internal/Unpreferred: addHeadItem, addScript, addScriptFile, addInlineScript, addStyle, addInlineStyle. All should be used as littel as possible, in favor of ResourceLoader and a few other higher level accessors. They are not truly deprecated, but should not really be used unless you have a really good reason. Also added a few missing wfDeprecated calls to methods marked as deprecated in the phpdoc. Bug: T89955 Change-Id: I472d906b6de16257fa34a2f9bfe9e4eb284d02b4
2015-10-31 15:16:58 +00:00
* Internal use only. Use OutputPage::addModuleStyles() if possible.
*
Removed unneeded spaces and colons in @param and friends Also swapped some "$var type" to "type $var" or added missing types before the $var. Changed some other types to match the more common spelling. Makes beginning of some text in captial. Change-Id: Ic36c8c7820a6c2d603f1138130670c6bf6a1ca59
2014-04-08 15:29:17 +00:00
* @param mixed $style_css Inline CSS
Fixed @param tags to conform with Doxygen format. Doxygen expects parameter types to come before the parameter name in @param tags. Used a quick regex to switch everything around where possible. This only fixes cases where a primitve variable (or a primitive followed by other types) is the variable type. Other cases will need to be fixed manually. Change-Id: Ic59fd20856eb0489d70f3469a56ebce0efb3db13
2013-03-11 17:15:01 +00:00
* @param string $flip Set to 'flip' to flip the CSS if needed
follow-up to r53282: *coding style tweaks *added __METHOD__ to some wfDebug calls *removed php ending tag from mwScriptLoader.php *some doxygen docs added
2009-07-15 00:55:58 +00:00
*/
* r96737: set OutputPage->addWikiText() as an interface message by default * r96760: use 'noflip' for $flip in OutputPage->addInlineStyle() so it's always a string * Add some parameter documentation
2011-09-11 21:07:17 +00:00
public function addInlineStyle( $style_css, $flip = 'noflip' ) {
else if -> elseif Also fixed some spacing while at it Added some braces for one line statements Change-Id: Iebfa2b17091509daabc76248121c7763e2295d20
2013-04-17 14:52:47 +00:00
if ( $flip === 'flip' && $this->getLanguage()->isRTL() ) {
* Mark underline css as !important. It should override other css (such as the css added in r91432) if the user specifically selected that option. * Fix css flipping: for Simple, it still depended on the content language (however, no css actually needs flipping there). For Standard/CologneBlue, I re-added flipping even though it's not needed there either. Just in case css is added that does need flipping, so it is future-proof :). Consequently, I added noflip to the existing css. * For this flipping I added a parameter to addInlineStyle(), which I assume to be useful in other cases as well
2011-09-04 21:50:02 +00:00
# If wanted, and the interface is right-to-left, flip the CSS
$style_css = CSSJanus::transform( $style_css, true, false );
}
OutputPage: Minor clean up of <head> and HTML * Remove trailing space in self-closing tag. Brings parserTest output of Parser and Parsoid closer together. * Remove various line breaks at begin and end of script contents. * Remove FILTER_NOMIN from makeConfigSetScript() output. This isn't part of any user- or page-dependent module and not minified. And Xml::encodeJsCall already ensures compact output for prod mode. Bug: T127328 Change-Id: I85a5a59fd0955c1a112e8b24b933f0d9e983a156
2016-02-18 16:33:15 +00:00
$this->mInlineStyles .= Html::inlineStyle( $style_css );
here it is ... the upload-api, script-server, js2 (javascript phase2) branch merge 1st attempt. Here is a short overview of changes and associated default configuration variables (most everything is off by default) also see ~soon to be updated~: http://www.mediawiki.org/wiki/Media_Projects_Overview = Upload Improvements = ==Upload API == * Based on the early work of Bryan Tong and others it adds the upload option to the api. * We rewrite Special:Upload page to include use the new refactoring * Added in token checks in both the SpecialUpload.php page so avoids DOS / xss copy-by-url JavaScript based cross site POST file submissions == Copy by URL== $wgAllowCopyUploads = false; * http class rewrite includes a new http background download see: includes/HttpFunctions.php * spins off a php process that calls: maintenance/http_session_download.php * pushes updates to the session and gives the user a progress bar on http copy uploads from other server progress (using js2 upload interface) (if not using the js2 upload interface it does the request in-place but the download is limited to the php ini timeout time) == Firefogg == * Firefogg enables resumable upload by chunks * progress indicators and conditional invokation (js2 system) * and of-course client side transcoding. = Script Server = $wgEnableScriptLoader = false; * off by default if $wgEnableScriptLoader is turned on script files are grouped, gziped, cached etc. for more info see: http://www.mediawiki.org/wiki/Extension:ScriptLoader * Includes some early skin js include fixes (skin/script system still lots of love) * Includes a "javascript class autoloader" this is packaged into mwEmbed so that the mwEmbed library can work in stand alone mode (while retaining localization and script serving) (one such application is the make page for firefogg.org : http://www.firefogg.org/make/index.html ) * The file that contains the autojavascript loading classes is: js2/php/jsAutoloadLocalClasses.php * One can use this auto class loading dependency system with extensions and add-ons but I need to better document that. = js2 system / mwEmbed= $wgEnableJS2system = false * includes initial rewrite towards more jquery based javascript code * especially for the Special:Upload page. * Also the edit page include support for the "add-media-wizard" * includes dependency loader for javascript that optionally takes advantage of the script-loader * remote embedding of javascript interfaces (like embedding video, or commons media searching) * $wgDebugJavaScript = false; .. .this variable lets you always get "always fresh javascript". When used with the script-loader it does not minify the script-loader output. = mwEmbed = * Will commit a separate patch to oggHandler that conditionally outputs <video tag> to use the new javascript video player. ** mv_embed player includes: play-head, volume control, remote embedding, oggz-chop support across plugins. * add-media-wizard adds easy inserts of media to pages (with import) == jQuery== * we include a base install of jQuery, jQuery ui and some plugins. * all the javascript classes are in the scriptloader so its easy to load any set of jquery ui components that you may need using the script-server. You get a callback so you can then execute js with dependencies loaded. == other stuff == there is a bit more code in js2 that pertains to sequence editing, timed text display and basic image editing. We include a base import of pixastic-lib & pixastic-editor... will work with the pixastic developer to try and ensure upstream compatibility on our usage of the library for in-browser photo and sequence manipulation.
2009-07-14 23:52:14 +00:00
}
follow-up to r53282: *coding style tweaks *added __METHOD__ to some wfDebug calls *removed php ending tag from mwScriptLoader.php *some doxygen docs added
2009-07-15 00:55:58 +00:00
Tweaks for skins: * Moved SkinTemplate::addStyle() and related stuff to OutputPage so that it can be used non-SkinTemplate skins and avoid duplication with the actual OutputPage::addStyle() (the two functions have the same format). * Non-SkinTemplate skins now also load their CSS with <link> tags instead of @import. * Moved SkinTemplate::setupUserCss() to Skin. * Merged action=raw&gen=(js|css) for SkinTemplate and non-SkinTemplate skins, renamed functions to Skin::generateUserJs() and Skin::generateUserStyleSheet() and dropped a lot of cascading call which is a bit incomprehensible.
2008-08-21 14:09:57 +00:00
/**
resourceloader: Move queue formatting out of OutputPage HTML formatting of the queue was distributed over several OutputPage methods. Each method demanding a snippet of HTML by calling makeResourceLoaderLink() with a limited amount of information. As such, makeResourceLoaderLink() was unable to provide the client with the proper state information. Centralising it also allows it to better reduce duplication in HTML output and maintain a more accurate state. Problems fixed by centralising: 1. The 'user' module is special (due to per-user 'version' and 'user' params). It is manually requested via script-src. To avoid a separate (and wrong) request from something that requires it, we set state=loading directly. However, because the module is in the bottom, the old HTML formatter could only put state=loading in the bottom also. This sometimes caused a wrong request to be fired for modules=user if something in the top queue triggered a requirement for it. 2. Since a464d1d4 (T87871) we track states of page-style modules, with purpose of allowing dependencies on style modules without risking duplicate loading on pages where the styles are loaded already. This didn't work, because the state information about page-style modules is output near the stylesheet, which is after the script tag with mw.loader.load(). That runs first, and mw.loader would still make a duplicate request before it learns the state. Changes: * Document reasons for style/script tag order in getHeadHtml (per 09537e83). * Pass $type from getModuleStyles() to getAllowedModules(). This wasn't needed before since a duplicate check in makeResourceLoaderLink() verified the origin a second time. * Declare explicit position 'top' on 'user.options' and 'user.tokens' module. Previously, OutputPage hardcoded them in the top. The new formatter doesn't. * Remove getHeadScripts(). * Remove getInlineHeadScripts(). * Remove getExternalHeadScripts(). * Remove buildCssLinks(). * Remove getScriptsForBottomQueue(). * Change where Skin::setupSkinUserCss() is called. This methods lets the skin add modules to the queue. Previously it was called from buildCssLinks(), via headElement(), via prepareQuickTemplate(), via OutputPage::output(). It's now in OutputPage::output() directly (slightly earlier). This is needed because prepareQuickTemplate() calls bottomScripts() before headElement(). And bottomScript() would lazy-initialise the queue and lock it before setupSkinUserCss() is called from headElement(). This makes execution order more predictable instead of being dependent on the arbitrary order of data extraction in prepareQuickTemplate (which varies from one skin to another). * Compute isUserModulePreview() and isKnownEmpty() for the 'user' module early on so. This avoids wrongful loading and fixes problem 1. Effective changes in output: * mw.loader.state() is now before mw.loader.load(). This fixes problem 2. * mw.loader.state() now sets 'user.options' and 'user.tokens' to "loading". * mw.loader.state() now sets 'user' (as "loading" or "ready"). Fixes problem 1. * The <script async src> tag for 'startup' changed position (slightly). Previously it was after all inline scripts and stylesheets. It's still after all inline scripts and after most stylesheets, but before any user styles. Since the queue is now formatted outside OutputPage, it can't inject the meta-ResourceLoaderDynamicStyles tag and user-stylesheet hack in the middle of existing output. This shouldn't have any noticable impact. Bug: T87871 Change-Id: I605b8cd1e1fc009b4662a0edbc54d09dd65ee1df
2016-07-15 14:13:09 +00:00
* Build exempt modules and legacy non-ResourceLoader styles.
More adding, updating and tweaking of documentation
2011-05-21 19:07:24 +00:00
*
OutputPage: Make use of WrappedStringList in headElement() Right now, getInlineHeadScripts(), buildCssLinks() and getExternalHeadScripts() all return WrappedString::join(). But because they don't know about each other and because they need to return strings (not arrays), headElement() has no way of merging them. WrappedStringList allows this array to be kept, whilst still being backward-compatible with code that calls these methods and assumig a string (since it will lazy-join the array if the object is treated like a string). To be used by I8b6c6a10d965e7396. Output is not changed in this commit. Merely refactoring. Change-Id: Iae08345473bd93cc0948d51b62c48aeb1ea460a3
2016-07-07 19:26:36 +00:00
* @return string|WrappedStringList HTML
Tweaks for skins: * Moved SkinTemplate::addStyle() and related stuff to OutputPage so that it can be used non-SkinTemplate skins and avoid duplication with the actual OutputPage::addStyle() (the two functions have the same format). * Non-SkinTemplate skins now also load their CSS with <link> tags instead of @import. * Moved SkinTemplate::setupUserCss() to Skin. * Merged action=raw&gen=(js|css) for SkinTemplate and non-SkinTemplate skins, renamed functions to Skin::generateUserJs() and Skin::generateUserStyleSheet() and dropped a lot of cascading call which is a bit incomprehensible.
2008-08-21 14:09:57 +00:00
*/
resourceloader: Move queue formatting out of OutputPage HTML formatting of the queue was distributed over several OutputPage methods. Each method demanding a snippet of HTML by calling makeResourceLoaderLink() with a limited amount of information. As such, makeResourceLoaderLink() was unable to provide the client with the proper state information. Centralising it also allows it to better reduce duplication in HTML output and maintain a more accurate state. Problems fixed by centralising: 1. The 'user' module is special (due to per-user 'version' and 'user' params). It is manually requested via script-src. To avoid a separate (and wrong) request from something that requires it, we set state=loading directly. However, because the module is in the bottom, the old HTML formatter could only put state=loading in the bottom also. This sometimes caused a wrong request to be fired for modules=user if something in the top queue triggered a requirement for it. 2. Since a464d1d4 (T87871) we track states of page-style modules, with purpose of allowing dependencies on style modules without risking duplicate loading on pages where the styles are loaded already. This didn't work, because the state information about page-style modules is output near the stylesheet, which is after the script tag with mw.loader.load(). That runs first, and mw.loader would still make a duplicate request before it learns the state. Changes: * Document reasons for style/script tag order in getHeadHtml (per 09537e83). * Pass $type from getModuleStyles() to getAllowedModules(). This wasn't needed before since a duplicate check in makeResourceLoaderLink() verified the origin a second time. * Declare explicit position 'top' on 'user.options' and 'user.tokens' module. Previously, OutputPage hardcoded them in the top. The new formatter doesn't. * Remove getHeadScripts(). * Remove getInlineHeadScripts(). * Remove getExternalHeadScripts(). * Remove buildCssLinks(). * Remove getScriptsForBottomQueue(). * Change where Skin::setupSkinUserCss() is called. This methods lets the skin add modules to the queue. Previously it was called from buildCssLinks(), via headElement(), via prepareQuickTemplate(), via OutputPage::output(). It's now in OutputPage::output() directly (slightly earlier). This is needed because prepareQuickTemplate() calls bottomScripts() before headElement(). And bottomScript() would lazy-initialise the queue and lock it before setupSkinUserCss() is called from headElement(). This makes execution order more predictable instead of being dependent on the arbitrary order of data extraction in prepareQuickTemplate (which varies from one skin to another). * Compute isUserModulePreview() and isKnownEmpty() for the 'user' module early on so. This avoids wrongful loading and fixes problem 1. Effective changes in output: * mw.loader.state() is now before mw.loader.load(). This fixes problem 2. * mw.loader.state() now sets 'user.options' and 'user.tokens' to "loading". * mw.loader.state() now sets 'user' (as "loading" or "ready"). Fixes problem 1. * The <script async src> tag for 'startup' changed position (slightly). Previously it was after all inline scripts and stylesheets. It's still after all inline scripts and after most stylesheets, but before any user styles. Since the queue is now formatted outside OutputPage, it can't inject the meta-ResourceLoaderDynamicStyles tag and user-stylesheet hack in the middle of existing output. This shouldn't have any noticable impact. Bug: T87871 Change-Id: I605b8cd1e1fc009b4662a0edbc54d09dd65ee1df
2016-07-15 14:13:09 +00:00
protected function buildExemptModules() {
$chunks = [];
Reduce a bit the coupling between Skin and OutputPage: * Removed Skin::setupUserCss() and merged it in OutputPage::buildCssLinks() and OutputPage::buildCssLinksArray(), also removed its call from ApiParse.php * Moved Skin::userCanPreview() to OutputPage since two of the three calls to that function are made from there and the last one in SkinTemplate.php is in a somewhat deprecated function * Removed the Skin parameter from OutputPage::buildCssLinks(), OutputPage::getHeadScripts(), OutputPage::getBottomScripts() and OutputPage::makeResourceLoaderLink() since we now have a context Also made ApiParse.php call createContext() instead of create a new RequestContext manually and set its Title so so that it does not rely on $wgRequest and $wgTitle to get that member objects
2011-07-22 10:45:07 +00:00
resourceloader: Only output ResourceLoaderDynamicStyles when needed In mediawiki.js, this marker has always been optional, falling back to appending to <head>. When no stylesheets need to be after the marker (e.g. no site styles on the wiki, and user is not logged-in), then there is no need for the marker to exist. In a previous refactor, I was going to do this and created an "$append" variable in the function to do what this commit does, but I forgot to actually use it for anything. Test Plan: * Local wiki, with no MediaWiki:Common/{Skinname}.css pages existing. * When logged-out, before this change, there is a marker, now there is not. * When creating "MediaWiki:Group-user.css" and logging in, there is still a marker, and it is still above the <link> for that user styles request in the <head>. Bug: T219342 Change-Id: I2e9657f318088860916823efeb96ae4f1532974c
2019-06-26 23:20:33 +00:00
// Requirements:
// - Within modules provided by the software (core, skin, extensions),
// styles from skin stylesheets should be overridden by styles
// from modules dynamically loaded with JavaScript.
// - Styles from site-specific, private, and user modules should override
// both of the above.
//
// The effective order for stylesheets must thus be:
ResourceLoader namespace (attempt 2) Move ResourceLoader classes to their own namespace. Strip the "ResourceLoader" prefix from all except ResourceLoader itself. Move the tests by analogy. I used a namespace alias "RL" in some callers since RL\Module is less ambiguous at the call site than just "Module". I did not address DependencyStore which continues to have a non-standard location and namespace. Revert of a241d83e0a6dabedf. Bug: T308718 Change-Id: Id08a220e1d6085e2b33f3f6c9d0e3935a4204659
2022-05-06 09:09:56 +00:00
// 1. Page style modules, formatted server-side by RL\ClientHtml.
resourceloader: Only output ResourceLoaderDynamicStyles when needed In mediawiki.js, this marker has always been optional, falling back to appending to <head>. When no stylesheets need to be after the marker (e.g. no site styles on the wiki, and user is not logged-in), then there is no need for the marker to exist. In a previous refactor, I was going to do this and created an "$append" variable in the function to do what this commit does, but I forgot to actually use it for anything. Test Plan: * Local wiki, with no MediaWiki:Common/{Skinname}.css pages existing. * When logged-out, before this change, there is a marker, now there is not. * When creating "MediaWiki:Group-user.css" and logging in, there is still a marker, and it is still above the <link> for that user styles request in the <head>. Bug: T219342 Change-Id: I2e9657f318088860916823efeb96ae4f1532974c
2019-06-26 23:20:33 +00:00
// 2. Dynamically-loaded styles, inserted client-side by mw.loader.
// 3. Styles that are site-specific, private or from the user, formatted
// server-side by this function.
//
// The 'ResourceLoaderDynamicStyles' marker helps JavaScript know where
// point #2 is.
resourceloader: Move queue formatting out of OutputPage HTML formatting of the queue was distributed over several OutputPage methods. Each method demanding a snippet of HTML by calling makeResourceLoaderLink() with a limited amount of information. As such, makeResourceLoaderLink() was unable to provide the client with the proper state information. Centralising it also allows it to better reduce duplication in HTML output and maintain a more accurate state. Problems fixed by centralising: 1. The 'user' module is special (due to per-user 'version' and 'user' params). It is manually requested via script-src. To avoid a separate (and wrong) request from something that requires it, we set state=loading directly. However, because the module is in the bottom, the old HTML formatter could only put state=loading in the bottom also. This sometimes caused a wrong request to be fired for modules=user if something in the top queue triggered a requirement for it. 2. Since a464d1d4 (T87871) we track states of page-style modules, with purpose of allowing dependencies on style modules without risking duplicate loading on pages where the styles are loaded already. This didn't work, because the state information about page-style modules is output near the stylesheet, which is after the script tag with mw.loader.load(). That runs first, and mw.loader would still make a duplicate request before it learns the state. Changes: * Document reasons for style/script tag order in getHeadHtml (per 09537e83). * Pass $type from getModuleStyles() to getAllowedModules(). This wasn't needed before since a duplicate check in makeResourceLoaderLink() verified the origin a second time. * Declare explicit position 'top' on 'user.options' and 'user.tokens' module. Previously, OutputPage hardcoded them in the top. The new formatter doesn't. * Remove getHeadScripts(). * Remove getInlineHeadScripts(). * Remove getExternalHeadScripts(). * Remove buildCssLinks(). * Remove getScriptsForBottomQueue(). * Change where Skin::setupSkinUserCss() is called. This methods lets the skin add modules to the queue. Previously it was called from buildCssLinks(), via headElement(), via prepareQuickTemplate(), via OutputPage::output(). It's now in OutputPage::output() directly (slightly earlier). This is needed because prepareQuickTemplate() calls bottomScripts() before headElement(). And bottomScript() would lazy-initialise the queue and lock it before setupSkinUserCss() is called from headElement(). This makes execution order more predictable instead of being dependent on the arbitrary order of data extraction in prepareQuickTemplate (which varies from one skin to another). * Compute isUserModulePreview() and isKnownEmpty() for the 'user' module early on so. This avoids wrongful loading and fixes problem 1. Effective changes in output: * mw.loader.state() is now before mw.loader.load(). This fixes problem 2. * mw.loader.state() now sets 'user.options' and 'user.tokens' to "loading". * mw.loader.state() now sets 'user' (as "loading" or "ready"). Fixes problem 1. * The <script async src> tag for 'startup' changed position (slightly). Previously it was after all inline scripts and stylesheets. It's still after all inline scripts and after most stylesheets, but before any user styles. Since the queue is now formatted outside OutputPage, it can't inject the meta-ResourceLoaderDynamicStyles tag and user-stylesheet hack in the middle of existing output. This shouldn't have any noticable impact. Bug: T87871 Change-Id: I605b8cd1e1fc009b4662a0edbc54d09dd65ee1df
2016-07-15 14:13:09 +00:00
// Add legacy styles added through addStyle()/addInlineStyle() here
$chunks[] = implode( '', $this->buildCssLinksArray() ) . $this->mInlineStyles;
resourceloader: Only output ResourceLoaderDynamicStyles when needed In mediawiki.js, this marker has always been optional, falling back to appending to <head>. When no stylesheets need to be after the marker (e.g. no site styles on the wiki, and user is not logged-in), then there is no need for the marker to exist. In a previous refactor, I was going to do this and created an "$append" variable in the function to do what this commit does, but I forgot to actually use it for anything. Test Plan: * Local wiki, with no MediaWiki:Common/{Skinname}.css pages existing. * When logged-out, before this change, there is a marker, now there is not. * When creating "MediaWiki:Group-user.css" and logging in, there is still a marker, and it is still above the <link> for that user styles request in the <head>. Bug: T219342 Change-Id: I2e9657f318088860916823efeb96ae4f1532974c
2019-06-26 23:20:33 +00:00
// Things that go after the ResourceLoaderDynamicStyles marker
$append = [];
resourceloader: Ensure user.styles and site.styles having their own request Regardless of whether other modules exist with group=user or group=site, these two modules in particular must always be in their own request for legacy reasons. This has already always been the case because even in the few cases where an extension uses this group (eg. MobileFrontend's custom site module) it would load it instead of another module in that group, never at the same time. There is one notable exception, which is GlobalCssJs. However the ext.globalCssJs.user.styles module is usually served from another wiki which is why that went unnoticed as well. This commit fixes that so that even if you're viewing a page on the central wiki, the modules are still in separate requests. Aside from this one existing edge case, there is also need to add group=site to gadgets by default so that they load after the DynamicStyles marker instead of before, which is currently causing problems with the cascading order (gadget apply before core and skin styles due to being in the same request group and alphabetically sorting before them). Semantically, the appropiate solution is group=site, but this wasn't possible due to core putting "all" group=site modules in the same request (under the assumption there is only one such module). This commit removes that fragile assumption. Bug: T147667 Change-Id: I9eb725c083124d22a9af3bf3d075ade6f3b970a3
2017-05-12 22:26:36 +00:00
$separateReq = [ 'site.styles', 'user.styles' ];
Remove unused key variable from foreach loops Change-Id: Id2d91e30a6f7cc4eb93427b50efc1c5c77f14b75
2022-09-21 19:05:03 +00:00
foreach ( $this->rlExemptStyleModules as $moduleNames ) {
resourceloader: Only output ResourceLoaderDynamicStyles when needed In mediawiki.js, this marker has always been optional, falling back to appending to <head>. When no stylesheets need to be after the marker (e.g. no site styles on the wiki, and user is not logged-in), then there is no need for the marker to exist. In a previous refactor, I was going to do this and created an "$append" variable in the function to do what this commit does, but I forgot to actually use it for anything. Test Plan: * Local wiki, with no MediaWiki:Common/{Skinname}.css pages existing. * When logged-out, before this change, there is a marker, now there is not. * When creating "MediaWiki:Group-user.css" and logging in, there is still a marker, and it is still above the <link> for that user styles request in the <head>. Bug: T219342 Change-Id: I2e9657f318088860916823efeb96ae4f1532974c
2019-06-26 23:20:33 +00:00
if ( $moduleNames ) {
$append[] = $this->makeResourceLoaderLink(
array_diff( $moduleNames, $separateReq ),
ResourceLoader namespace (attempt 2) Move ResourceLoader classes to their own namespace. Strip the "ResourceLoader" prefix from all except ResourceLoader itself. Move the tests by analogy. I used a namespace alias "RL" in some callers since RL\Module is less ambiguous at the call site than just "Module". I did not address DependencyStore which continues to have a non-standard location and namespace. Revert of a241d83e0a6dabedf. Bug: T308718 Change-Id: Id08a220e1d6085e2b33f3f6c9d0e3935a4204659
2022-05-06 09:09:56 +00:00
RL\Module::TYPE_STYLES
resourceloader: Ensure user.styles and site.styles having their own request Regardless of whether other modules exist with group=user or group=site, these two modules in particular must always be in their own request for legacy reasons. This has already always been the case because even in the few cases where an extension uses this group (eg. MobileFrontend's custom site module) it would load it instead of another module in that group, never at the same time. There is one notable exception, which is GlobalCssJs. However the ext.globalCssJs.user.styles module is usually served from another wiki which is why that went unnoticed as well. This commit fixes that so that even if you're viewing a page on the central wiki, the modules are still in separate requests. Aside from this one existing edge case, there is also need to add group=site to gadgets by default so that they load after the DynamicStyles marker instead of before, which is currently causing problems with the cascading order (gadget apply before core and skin styles due to being in the same request group and alphabetically sorting before them). Semantically, the appropiate solution is group=site, but this wasn't possible due to core putting "all" group=site modules in the same request (under the assumption there is only one such module). This commit removes that fragile assumption. Bug: T147667 Change-Id: I9eb725c083124d22a9af3bf3d075ade6f3b970a3
2017-05-12 22:26:36 +00:00
);
resourceloader: Only output ResourceLoaderDynamicStyles when needed In mediawiki.js, this marker has always been optional, falling back to appending to <head>. When no stylesheets need to be after the marker (e.g. no site styles on the wiki, and user is not logged-in), then there is no need for the marker to exist. In a previous refactor, I was going to do this and created an "$append" variable in the function to do what this commit does, but I forgot to actually use it for anything. Test Plan: * Local wiki, with no MediaWiki:Common/{Skinname}.css pages existing. * When logged-out, before this change, there is a marker, now there is not. * When creating "MediaWiki:Group-user.css" and logging in, there is still a marker, and it is still above the <link> for that user styles request in the <head>. Bug: T219342 Change-Id: I2e9657f318088860916823efeb96ae4f1532974c
2019-06-26 23:20:33 +00:00
foreach ( array_intersect( $moduleNames, $separateReq ) as $name ) {
// These require their own dedicated request in order to support "@import"
// syntax, which is incompatible with concatenation. (T147667, T37562)
$append[] = $this->makeResourceLoaderLink( $name,
ResourceLoader namespace (attempt 2) Move ResourceLoader classes to their own namespace. Strip the "ResourceLoader" prefix from all except ResourceLoader itself. Move the tests by analogy. I used a namespace alias "RL" in some callers since RL\Module is less ambiguous at the call site than just "Module". I did not address DependencyStore which continues to have a non-standard location and namespace. Revert of a241d83e0a6dabedf. Bug: T308718 Change-Id: Id08a220e1d6085e2b33f3f6c9d0e3935a4204659
2022-05-06 09:09:56 +00:00
RL\Module::TYPE_STYLES
resourceloader: Only output ResourceLoaderDynamicStyles when needed In mediawiki.js, this marker has always been optional, falling back to appending to <head>. When no stylesheets need to be after the marker (e.g. no site styles on the wiki, and user is not logged-in), then there is no need for the marker to exist. In a previous refactor, I was going to do this and created an "$append" variable in the function to do what this commit does, but I forgot to actually use it for anything. Test Plan: * Local wiki, with no MediaWiki:Common/{Skinname}.css pages existing. * When logged-out, before this change, there is a marker, now there is not. * When creating "MediaWiki:Group-user.css" and logging in, there is still a marker, and it is still above the <link> for that user styles request in the <head>. Bug: T219342 Change-Id: I2e9657f318088860916823efeb96ae4f1532974c
2019-06-26 23:20:33 +00:00
);
}
resourceloader: Ensure user.styles and site.styles having their own request Regardless of whether other modules exist with group=user or group=site, these two modules in particular must always be in their own request for legacy reasons. This has already always been the case because even in the few cases where an extension uses this group (eg. MobileFrontend's custom site module) it would load it instead of another module in that group, never at the same time. There is one notable exception, which is GlobalCssJs. However the ext.globalCssJs.user.styles module is usually served from another wiki which is why that went unnoticed as well. This commit fixes that so that even if you're viewing a page on the central wiki, the modules are still in separate requests. Aside from this one existing edge case, there is also need to add group=site to gadgets by default so that they load after the DynamicStyles marker instead of before, which is currently causing problems with the cascading order (gadget apply before core and skin styles due to being in the same request group and alphabetically sorting before them). Semantically, the appropiate solution is group=site, but this wasn't possible due to core putting "all" group=site modules in the same request (under the assumption there is only one such module). This commit removes that fragile assumption. Bug: T147667 Change-Id: I9eb725c083124d22a9af3bf3d075ade6f3b970a3
2017-05-12 22:26:36 +00:00
}
(bug 27468) Move private styles out to their own section, so they override everything except user styles. Moved from one makeResourceLoaderLink() call with array_merge() to multiple calls, because makeResourceLoaderLink() orders the modules, then groups them, so the input order is not preserved. Also fixed an issue in makeResourceLoaderLink() where the $uery variable was reused in each iteration but not cleared, allowing pollution to occur in theory (doesn't seem to have happened in practice, though).
2011-02-20 17:24:42 +00:00
}
resourceloader: Only output ResourceLoaderDynamicStyles when needed In mediawiki.js, this marker has always been optional, falling back to appending to <head>. When no stylesheets need to be after the marker (e.g. no site styles on the wiki, and user is not logged-in), then there is no need for the marker to exist. In a previous refactor, I was going to do this and created an "$append" variable in the function to do what this commit does, but I forgot to actually use it for anything. Test Plan: * Local wiki, with no MediaWiki:Common/{Skinname}.css pages existing. * When logged-out, before this change, there is a marker, now there is not. * When creating "MediaWiki:Group-user.css" and logging in, there is still a marker, and it is still above the <link> for that user styles request in the <head>. Bug: T219342 Change-Id: I2e9657f318088860916823efeb96ae4f1532974c
2019-06-26 23:20:33 +00:00
if ( $append ) {
$chunks[] = Html::element(
'meta',
[ 'name' => 'ResourceLoaderDynamicStyles', 'content' => '' ]
);
$chunks = array_merge( $chunks, $append );
}
Tidy up w/c Remove trailing whitespace, add some documentation
2011-08-31 14:47:08 +00:00
resourceloader: Only output ResourceLoaderDynamicStyles when needed In mediawiki.js, this marker has always been optional, falling back to appending to <head>. When no stylesheets need to be after the marker (e.g. no site styles on the wiki, and user is not logged-in), then there is no need for the marker to exist. In a previous refactor, I was going to do this and created an "$append" variable in the function to do what this commit does, but I forgot to actually use it for anything. Test Plan: * Local wiki, with no MediaWiki:Common/{Skinname}.css pages existing. * When logged-out, before this change, there is a marker, now there is not. * When creating "MediaWiki:Group-user.css" and logging in, there is still a marker, and it is still above the <link> for that user styles request in the <head>. Bug: T219342 Change-Id: I2e9657f318088860916823efeb96ae4f1532974c
2019-06-26 23:20:33 +00:00
return self::combineWrappedStrings( $chunks );
bug 18608 done in a fashion much closer to how I should've done it originally. Latching onto bug 22061 output Following up r62195 and r61125
2010-05-24 16:35:23 +00:00
}
* r96737: set OutputPage->addWikiText() as an interface message by default * r96760: use 'noflip' for $flip in OutputPage->addInlineStyle() so it's always a string * Add some parameter documentation
2011-09-11 21:07:17 +00:00
/**
Fixed some @params documentation Swapped some "$var type" to "type $var" or added missing types before the $var. Changed some other types to match the more common spelling. Makes beginning of some text in captial. Change-Id: I64e8cfe478cb0ba438f40b0631d6e9049cdab567
2014-04-10 18:50:10 +00:00
* @return array
* r96737: set OutputPage->addWikiText() as an interface message by default * r96760: use 'noflip' for $flip in OutputPage->addInlineStyle() so it's always a string * Add some parameter documentation
2011-09-11 21:07:17 +00:00
*/
bug 18608 done in a fashion much closer to how I should've done it originally. Latching onto bug 22061 output Following up r62195 and r61125
2010-05-24 16:35:23 +00:00
public function buildCssLinksArray() {
Convert all array() syntax to [] Per wikitech-l consensus: https://lists.wikimedia.org/pipermail/wikitech-l/2016-February/084821.html Notes: * Disabled CallTimePassByReference due to false positives (T127163) Change-Id: I2c8ce713ce6600a0bb7bf67537c87044c7a45c4b
2016-02-17 09:09:32 +00:00
$links = [];
Reduce a bit the coupling between Skin and OutputPage: * Removed Skin::setupUserCss() and merged it in OutputPage::buildCssLinks() and OutputPage::buildCssLinksArray(), also removed its call from ApiParse.php * Moved Skin::userCanPreview() to OutputPage since two of the three calls to that function are made from there and the last one in SkinTemplate.php is in a somewhat deprecated function * Removed the Skin parameter from OutputPage::buildCssLinks(), OutputPage::getHeadScripts(), OutputPage::getBottomScripts() and OutputPage::makeResourceLoaderLink() since we now have a context Also made ApiParse.php call createContext() instead of create a new RequestContext manually and set its Title so so that it does not rely on $wgRequest and $wgTitle to get that member objects
2011-07-22 10:45:07 +00:00
else if -> elseif Also fixed some spacing while at it Added some braces for one line statements Change-Id: Iebfa2b17091509daabc76248121c7763e2295d20
2013-04-17 14:52:47 +00:00
foreach ( $this->styles as $file => $options ) {
Tweaks for skins: * Moved SkinTemplate::addStyle() and related stuff to OutputPage so that it can be used non-SkinTemplate skins and avoid duplication with the actual OutputPage::addStyle() (the two functions have the same format). * Non-SkinTemplate skins now also load their CSS with <link> tags instead of @import. * Moved SkinTemplate::setupUserCss() to Skin. * Merged action=raw&gen=(js|css) for SkinTemplate and non-SkinTemplate skins, renamed functions to Skin::generateUserJs() and Skin::generateUserStyleSheet() and dropped a lot of cascading call which is a bit incomprehensible.
2008-08-21 14:09:57 +00:00
$link = $this->styleLink( $file, $options );
else if -> elseif Also fixed some spacing while at it Added some braces for one line statements Change-Id: Iebfa2b17091509daabc76248121c7763e2295d20
2013-04-17 14:52:47 +00:00
if ( $link ) {
bug 18608 done in a fashion much closer to how I should've done it originally. Latching onto bug 22061 output Following up r62195 and r61125
2010-05-24 16:35:23 +00:00
$links[$file] = $link;
*coding style cleanup (added braces, spacing tweaks etc.) *removed some unused globals
2010-05-22 12:18:22 +00:00
}
Tweaks for skins: * Moved SkinTemplate::addStyle() and related stuff to OutputPage so that it can be used non-SkinTemplate skins and avoid duplication with the actual OutputPage::addStyle() (the two functions have the same format). * Non-SkinTemplate skins now also load their CSS with <link> tags instead of @import. * Moved SkinTemplate::setupUserCss() to Skin. * Merged action=raw&gen=(js|css) for SkinTemplate and non-SkinTemplate skins, renamed functions to Skin::generateUserJs() and Skin::generateUserStyleSheet() and dropped a lot of cascading call which is a bit incomprehensible.
2008-08-21 14:09:57 +00:00
}
bug 18608 done in a fashion much closer to how I should've done it originally. Latching onto bug 22061 output Following up r62195 and r61125
2010-05-24 16:35:23 +00:00
return $links;
Tweaks for skins: * Moved SkinTemplate::addStyle() and related stuff to OutputPage so that it can be used non-SkinTemplate skins and avoid duplication with the actual OutputPage::addStyle() (the two functions have the same format). * Non-SkinTemplate skins now also load their CSS with <link> tags instead of @import. * Moved SkinTemplate::setupUserCss() to Skin. * Merged action=raw&gen=(js|css) for SkinTemplate and non-SkinTemplate skins, renamed functions to Skin::generateUserJs() and Skin::generateUserStyleSheet() and dropped a lot of cascading call which is a bit incomprehensible.
2008-08-21 14:09:57 +00:00
}
Step two in fixing OutputPage's documentation, more or less the same as r61647
2010-01-29 21:32:45 +00:00
/**
* Generate \<link\> tags for stylesheets
*
Fixed @param tags to conform with Doxygen format. Doxygen expects parameter types to come before the parameter name in @param tags. Used a quick regex to switch everything around where possible. This only fixes cases where a primitve variable (or a primitive followed by other types) is the variable type. Other cases will need to be fixed manually. Change-Id: Ic59fd20856eb0489d70f3469a56ebce0efb3db13
2013-03-11 17:15:01 +00:00
* @param string $style URL to the file
Removed unneeded spaces and colons in @param and friends Also swapped some "$var type" to "type $var" or added missing types before the $var. Changed some other types to match the more common spelling. Makes beginning of some text in captial. Change-Id: Ic36c8c7820a6c2d603f1138130670c6bf6a1ca59
2014-04-08 15:29:17 +00:00
* @param array $options Option, can contain 'condition', 'dir', 'media' keys
* @return string HTML fragment
Step two in fixing OutputPage's documentation, more or less the same as r61647
2010-01-29 21:32:45 +00:00
*/
Enforce array type hinting in OutputPage.php Change-Id: I2b37a95e517e120471c709823ea319f60165be0e
2014-07-23 02:49:30 +00:00
protected function styleLink( $style, array $options ) {
Collapse some nested if statements Change-Id: I9a97325d738d09370d29d35d5254bc0dadc57ff4
2019-03-29 20:12:24 +00:00
if ( isset( $options['dir'] ) && $this->getLanguage()->getDir() != $options['dir'] ) {
return '';
Tweaks for skins: * Moved SkinTemplate::addStyle() and related stuff to OutputPage so that it can be used non-SkinTemplate skins and avoid duplication with the actual OutputPage::addStyle() (the two functions have the same format). * Non-SkinTemplate skins now also load their CSS with <link> tags instead of @import. * Moved SkinTemplate::setupUserCss() to Skin. * Merged action=raw&gen=(js|css) for SkinTemplate and non-SkinTemplate skins, renamed functions to Skin::generateUserJs() and Skin::generateUserStyleSheet() and dropped a lot of cascading call which is a bit incomprehensible.
2008-08-21 14:09:57 +00:00
}
else if -> elseif Also fixed some spacing while at it Added some braces for one line statements Change-Id: Iebfa2b17091509daabc76248121c7763e2295d20
2013-04-17 14:52:47 +00:00
if ( isset( $options['media'] ) ) {
(bug 26497) printable=yes and handheld=yes request parameters were broken. Fixed this in a bit of an ugly way, by propagating these parameters to load.php and running OutputPage::transformCssMedia() on it there. This revision also fixes the fact that ResourceLoader ignored $wgHandheldForIPhone, and no longer wraps styles in @media all { } because it's useless and breaks @import (which can't be used inside @media ; this is bug 26478)
2011-01-07 20:22:50 +00:00
$media = self::transformCssMedia( $options['media'] );
Coding style: Auto-fix MediaWiki.Usage.IsNull.IsNull Change-Id: I90cfe8366c0245c9c67e598d17800684897a4e27
2020-01-09 23:48:34 +00:00
if ( $media === null ) {
Tweaks for skins: * Moved SkinTemplate::addStyle() and related stuff to OutputPage so that it can be used non-SkinTemplate skins and avoid duplication with the actual OutputPage::addStyle() (the two functions have the same format). * Non-SkinTemplate skins now also load their CSS with <link> tags instead of @import. * Moved SkinTemplate::setupUserCss() to Skin. * Merged action=raw&gen=(js|css) for SkinTemplate and non-SkinTemplate skins, renamed functions to Skin::generateUserJs() and Skin::generateUserStyleSheet() and dropped a lot of cascading call which is a bit incomprehensible.
2008-08-21 14:09:57 +00:00
return '';
}
} else {
Refactor redundant attrib dropping into new method This saves code in a few places on the caller's side, and will reduce the size of output HTML more consistently.
2009-09-07 00:21:55 +00:00
$media = 'all';
Tweaks for skins: * Moved SkinTemplate::addStyle() and related stuff to OutputPage so that it can be used non-SkinTemplate skins and avoid duplication with the actual OutputPage::addStyle() (the two functions have the same format). * Non-SkinTemplate skins now also load their CSS with <link> tags instead of @import. * Moved SkinTemplate::setupUserCss() to Skin. * Merged action=raw&gen=(js|css) for SkinTemplate and non-SkinTemplate skins, renamed functions to Skin::generateUserJs() and Skin::generateUserStyleSheet() and dropped a lot of cascading call which is a bit incomprehensible.
2008-08-21 14:09:57 +00:00
}
else if -> elseif Also fixed some spacing while at it Added some braces for one line statements Change-Id: Iebfa2b17091509daabc76248121c7763e2295d20
2013-04-17 14:52:47 +00:00
if ( substr( $style, 0, 1 ) == '/' ||
Tweaks for skins: * Moved SkinTemplate::addStyle() and related stuff to OutputPage so that it can be used non-SkinTemplate skins and avoid duplication with the actual OutputPage::addStyle() (the two functions have the same format). * Non-SkinTemplate skins now also load their CSS with <link> tags instead of @import. * Moved SkinTemplate::setupUserCss() to Skin. * Merged action=raw&gen=(js|css) for SkinTemplate and non-SkinTemplate skins, renamed functions to Skin::generateUserJs() and Skin::generateUserStyleSheet() and dropped a lot of cascading call which is a bit incomprehensible.
2008-08-21 14:09:57 +00:00
substr( $style, 0, 5 ) == 'http:' ||
substr( $style, 0, 6 ) == 'https:' ) {
$url = $style;
} else {
Use config instead of globals for OutputPage Change-Id: I5e0ebc173631d1d1052de7ccee4ef839c7c1767f
2014-08-23 08:52:41 +00:00
$config = $this->getConfig();
OutputPage: Remove appending of wgStyleVersion to legacy resources For addScriptFile(), just remove the appending of wgStyleVersion. Going forward, anyone still using this, should simply append a query parameter themselves in a way that is specific to that one url (instead of relying on a generic global variable). Alternatively, one could use OutputPage::transformResourcePath if the file is in /w/. For addStyles(), also remove the appending of wgStyleVersion. Since this method takes paths relative to /w/skins, we can easily update this to automatically use transformResourcePath(), so that file-hash based query parameters are automatically added. Test Plan: * Add calls to top of OutputPage::output(): `$this->addStyle( 'Vector/README.md' );` `$this->addScriptFile( "{$GLOBALS['wgScriptPath']}/composer.json" );` * Before, they are both inserted as `<link>` (head) and `<script>` (body) with a query parameter based on wgStyleVersion. * After, the `<script>` (end of body) has no query. After, the stylesheet (head) has a SHA1 content hash as query. Bug: T181318 Change-Id: Ie5ab5066ef7d07279086bde838d7305e9e4eabaf
2018-05-21 21:21:28 +00:00
// Append file hash as query parameter
$url = self::transformResourcePath(
$config,
Use name constants to access config settings. Use name constants instead of string literals in calls to Config::get and ServiceOptions::get, when referring to core configuration variables. This protects against typos and makes the decumentation and schema declaration of the config settings discoverable. This is the first batch, only touching files directly under /includes/ Change-Id: I7252e636c7c86d950d9257b33491af492c6dd5eb
2022-04-01 15:58:32 +00:00
$config->get( MainConfigNames::StylePath ) . '/' . $style
OutputPage: Remove appending of wgStyleVersion to legacy resources For addScriptFile(), just remove the appending of wgStyleVersion. Going forward, anyone still using this, should simply append a query parameter themselves in a way that is specific to that one url (instead of relying on a generic global variable). Alternatively, one could use OutputPage::transformResourcePath if the file is in /w/. For addStyles(), also remove the appending of wgStyleVersion. Since this method takes paths relative to /w/skins, we can easily update this to automatically use transformResourcePath(), so that file-hash based query parameters are automatically added. Test Plan: * Add calls to top of OutputPage::output(): `$this->addStyle( 'Vector/README.md' );` `$this->addScriptFile( "{$GLOBALS['wgScriptPath']}/composer.json" );` * Before, they are both inserted as `<link>` (head) and `<script>` (body) with a query parameter based on wgStyleVersion. * After, the `<script>` (end of body) has no query. After, the stylesheet (head) has a SHA1 content hash as query. Bug: T181318 Change-Id: Ie5ab5066ef7d07279086bde838d7305e9e4eabaf
2018-05-21 21:21:28 +00:00
);
Tweaks for skins: * Moved SkinTemplate::addStyle() and related stuff to OutputPage so that it can be used non-SkinTemplate skins and avoid duplication with the actual OutputPage::addStyle() (the two functions have the same format). * Non-SkinTemplate skins now also load their CSS with <link> tags instead of @import. * Moved SkinTemplate::setupUserCss() to Skin. * Merged action=raw&gen=(js|css) for SkinTemplate and non-SkinTemplate skins, renamed functions to Skin::generateUserJs() and Skin::generateUserStyleSheet() and dropped a lot of cascading call which is a bit incomprehensible.
2008-08-21 14:09:57 +00:00
}
For HTML 5, drop type="" attributes for CSS/JS This time done in a nice, centralized fashion, reducing LOC for callers even if HTML 5 is disabled. The implementation is a new Html class, similar to Xml but intended to be HTML-specific from the beginning instead of half-heartedly attempting to provide generic XML services but actually with lots of HTML-specific stuff tacked on. As part of the new Html class, a global config option $wgWellFormedXml is added. It's set to true by default, but if set to false, the Html class will drop some things that HTML 5 doesn't require, like self-closing " />" syntax and attribute quotation marks (sometimes).
2009-08-11 00:09:24 +00:00
$link = Html::linkedStyle( $url, $media );
Tweaks for skins: * Moved SkinTemplate::addStyle() and related stuff to OutputPage so that it can be used non-SkinTemplate skins and avoid duplication with the actual OutputPage::addStyle() (the two functions have the same format). * Non-SkinTemplate skins now also load their CSS with <link> tags instead of @import. * Moved SkinTemplate::setupUserCss() to Skin. * Merged action=raw&gen=(js|css) for SkinTemplate and non-SkinTemplate skins, renamed functions to Skin::generateUserJs() and Skin::generateUserStyleSheet() and dropped a lot of cascading call which is a bit incomprehensible.
2008-08-21 14:09:57 +00:00
else if -> elseif Also fixed some spacing while at it Added some braces for one line statements Change-Id: Iebfa2b17091509daabc76248121c7763e2295d20
2013-04-17 14:52:47 +00:00
if ( isset( $options['condition'] ) ) {
Tweaks for skins: * Moved SkinTemplate::addStyle() and related stuff to OutputPage so that it can be used non-SkinTemplate skins and avoid duplication with the actual OutputPage::addStyle() (the two functions have the same format). * Non-SkinTemplate skins now also load their CSS with <link> tags instead of @import. * Moved SkinTemplate::setupUserCss() to Skin. * Merged action=raw&gen=(js|css) for SkinTemplate and non-SkinTemplate skins, renamed functions to Skin::generateUserJs() and Skin::generateUserStyleSheet() and dropped a lot of cascading call which is a bit incomprehensible.
2008-08-21 14:09:57 +00:00
$condition = htmlspecialchars( $options['condition'] );
$link = "<!--[if $condition]>$link<![endif]-->";
}
return $link;
}
Centralise url handling for urls to static resources Keep in CSSMin as-is for back-compat and to ensure library remains independent of MediaWiki. Moved down a few lines as there is no need to compute the md5 hash when we're returning a data URI. Previously md5_file was called twice during module builds (once for the fallback url, and another time when producing the embedded data uri). Applied to logo in SkinModule as example. To be applied elsewhere as needed. Without it, fallback is current behaviour (no cache invalidation). Bug: T99096 Change-Id: I7f38bfc1bea5c241bc4f8ec4f4b640fd65f2c04f
2016-01-22 22:58:03 +00:00
/**
* Transform path to web-accessible static resource.
*
* This is used to add a validation hash as query string.
* This aids various behaviors:
*
* - Put long Cache-Control max-age headers on responses for improved
* cache performance.
* - Get the correct version of a file as expected by the current page.
* - Instantly get the updated version of a file after deployment.
*
* Avoid using this for urls included in HTML as otherwise clients may get different
* versions of a resource when navigating the site depending on when the page was cached.
* If changes to the url propagate, this is not a problem (e.g. if the url is in
* an external stylesheet).
*
* @since 1.27
* @param Config $config
* @param string $path Path-absolute URL to file (from document root, must start with "/")
* @return string URL
*/
public static function transformResourcePath( Config $config, $path ) {
Use name constants to access config settings. Use name constants instead of string literals in calls to Config::get and ServiceOptions::get, when referring to core configuration variables. This protects against typos and makes the decumentation and schema declaration of the config settings discoverable. This is the first batch, only touching files directly under /includes/ Change-Id: I7252e636c7c86d950d9257b33491af492c6dd5eb
2022-04-01 15:58:32 +00:00
$localDir = $config->get( MainConfigNames::BaseDirectory );
$remotePathPrefix = $config->get( MainConfigNames::ResourceBasePath );
Fix resource path check when ResourceBasePath is an empty string If you have MediaWiki installed in the root of the domain, then $wgScriptPath and $wgResourceBasePath is an empty string. In HHVM and PHP, passing an empty string as the second parameter of strpos() causes a warning and returns false, which will cause the condition to fail, as if the path were not within the base path. So, normalize such paths. Using substr() instead of strpos() for a "starts with" check would have worked except that RelPath also fails when given an empty string. Bug: T127652 Change-Id: If7e94ae638d6834f7cc0f31f67a5fe6a2f74771c
2016-02-22 03:51:19 +00:00
if ( $remotePathPrefix === '' ) {
// The configured base path is required to be empty string for
// wikis in the domain root
$remotePath = '/';
} else {
$remotePath = $remotePathPrefix;
}
OutputPage: Ignore protocol-relative urls in transformResourcePath() transformResourcePath is only supposed to be called with something that starts with a '/' (absolute path from document root). While the primary caller (CSSMin) only calls this method if the path makes sense as a local file path, ResourceLoaderSkinModule did not, and wgLogo may be set to a full url that includes a domain. While chance made it so that protocol-including urls were already discarded, protocol-relative urls were mistaken for being a path inside the file system root with a duplicate slash (e.g. '//tmp/foo.txt', as 'foo.txt' in /fmp, instead of '/foo.txt' at http://tmp). This should be fixed upstream in the wikimedia/relpath library, but workaround it for now since it really shouldn't be called with urls in the first place. Bug: T155310 Change-Id: I9b063f1219ddeca5cc2c8a48832cdf8c9eaffe58
2017-01-18 06:07:05 +00:00
if ( strpos( $path, $remotePath ) !== 0 || substr( $path, 0, 2 ) === '//' ) {
// - Path is outside wgResourceBasePath, ignore.
// - Path is protocol-relative. Fixes T155310. Not supported by RelPath lib.
Centralise url handling for urls to static resources Keep in CSSMin as-is for back-compat and to ensure library remains independent of MediaWiki. Moved down a few lines as there is no need to compute the md5 hash when we're returning a data URI. Previously md5_file was called twice during module builds (once for the fallback url, and another time when producing the embedded data uri). Applied to logo in SkinModule as example. To be applied elsewhere as needed. Without it, fallback is current behaviour (no cache invalidation). Bug: T99096 Change-Id: I7f38bfc1bea5c241bc4f8ec4f4b640fd65f2c04f
2016-01-22 22:58:03 +00:00
return $path;
}
OutputPage: Support UploadPath in testTransformResourcePath() Updated tests to reflect this use case. Currently we assume all web-accessible paths within ResourceBasePath (e.g "/w") to exist on disk at the same path in $IP (e.g. "/var/www/mw"). While in theory any number of web server rewrites or aliases could exist, there is one case in particular that we should support since the information is available in the configuration: UploadDir and UploadPath. This path may be rewritten in a way that varies by wiki in multi-wiki installs that share the same source code. E.g. a server may rewrite "/w/images" to somewhere else, which means it will not match the directory on disk that is shared between wikis. Bug: T155146 Change-Id: I320478c9c262cc012f08b585b48d290594ec2420
2017-02-10 00:03:06 +00:00
// For files in resources, extensions/ or skins/, ResourceBasePath is preferred here.
// For other misc files in $IP, we'll fallback to that as well. There is, however, a fourth
// supported dir/path pair in the configuration (wgUploadDirectory, wgUploadPath)
// which is not expected to be in wgResourceBasePath on CDNs. (T155146)
Use name constants to access config settings. Use name constants instead of string literals in calls to Config::get and ServiceOptions::get, when referring to core configuration variables. This protects against typos and makes the decumentation and schema declaration of the config settings discoverable. This is the first batch, only touching files directly under /includes/ Change-Id: I7252e636c7c86d950d9257b33491af492c6dd5eb
2022-04-01 15:58:32 +00:00
$uploadPath = $config->get( MainConfigNames::UploadPath );
OutputPage: Support UploadPath in testTransformResourcePath() Updated tests to reflect this use case. Currently we assume all web-accessible paths within ResourceBasePath (e.g "/w") to exist on disk at the same path in $IP (e.g. "/var/www/mw"). While in theory any number of web server rewrites or aliases could exist, there is one case in particular that we should support since the information is available in the configuration: UploadDir and UploadPath. This path may be rewritten in a way that varies by wiki in multi-wiki installs that share the same source code. E.g. a server may rewrite "/w/images" to somewhere else, which means it will not match the directory on disk that is shared between wikis. Bug: T155146 Change-Id: I320478c9c262cc012f08b585b48d290594ec2420
2017-02-10 00:03:06 +00:00
if ( strpos( $path, $uploadPath ) === 0 ) {
Use name constants to access config settings. Use name constants instead of string literals in calls to Config::get and ServiceOptions::get, when referring to core configuration variables. This protects against typos and makes the decumentation and schema declaration of the config settings discoverable. This is the first batch, only touching files directly under /includes/ Change-Id: I7252e636c7c86d950d9257b33491af492c6dd5eb
2022-04-01 15:58:32 +00:00
$localDir = $config->get( MainConfigNames::UploadDirectory );
OutputPage: Support UploadPath in testTransformResourcePath() Updated tests to reflect this use case. Currently we assume all web-accessible paths within ResourceBasePath (e.g "/w") to exist on disk at the same path in $IP (e.g. "/var/www/mw"). While in theory any number of web server rewrites or aliases could exist, there is one case in particular that we should support since the information is available in the configuration: UploadDir and UploadPath. This path may be rewritten in a way that varies by wiki in multi-wiki installs that share the same source code. E.g. a server may rewrite "/w/images" to somewhere else, which means it will not match the directory on disk that is shared between wikis. Bug: T155146 Change-Id: I320478c9c262cc012f08b585b48d290594ec2420
2017-02-10 00:03:06 +00:00
$remotePathPrefix = $remotePath = $uploadPath;
}
Update RelPath calls Bug: T182273 Change-Id: Ifd04e3b3ab72cc56bef48a1fc18fccb34b0af618
2018-01-21 04:33:38 +00:00
$path = RelPath::getRelativePath( $path, $remotePath );
OutputPage: Support UploadPath in testTransformResourcePath() Updated tests to reflect this use case. Currently we assume all web-accessible paths within ResourceBasePath (e.g "/w") to exist on disk at the same path in $IP (e.g. "/var/www/mw"). While in theory any number of web server rewrites or aliases could exist, there is one case in particular that we should support since the information is available in the configuration: UploadDir and UploadPath. This path may be rewritten in a way that varies by wiki in multi-wiki installs that share the same source code. E.g. a server may rewrite "/w/images" to somewhere else, which means it will not match the directory on disk that is shared between wikis. Bug: T155146 Change-Id: I320478c9c262cc012f08b585b48d290594ec2420
2017-02-10 00:03:06 +00:00
return self::transformFilePath( $remotePathPrefix, $localDir, $path );
Centralise url handling for urls to static resources Keep in CSSMin as-is for back-compat and to ensure library remains independent of MediaWiki. Moved down a few lines as there is no need to compute the md5 hash when we're returning a data URI. Previously md5_file was called twice during module builds (once for the fallback url, and another time when producing the embedded data uri). Applied to logo in SkinModule as example. To be applied elsewhere as needed. Without it, fallback is current behaviour (no cache invalidation). Bug: T99096 Change-Id: I7f38bfc1bea5c241bc4f8ec4f4b640fd65f2c04f
2016-01-22 22:58:03 +00:00
}
/**
* Utility method for transformResourceFilePath().
*
* Caller is responsible for ensuring the file exists. Emits a PHP warning otherwise.
*
* @since 1.27
Use correct variable name in @param documentation For some varargs a variable name is added with suffix ,... as seen for many other varargs Some @param are swapped, because there are in the wrong order Enable Sniff MediaWiki.Commenting.FunctionComment.ParamNameNoMatch Change-Id: I60fec6025bce824d5c67563ab7b65ad6cd628ad8
2017-08-11 16:09:41 +00:00
* @param string $remotePathPrefix URL path prefix that points to $localPath
Centralise url handling for urls to static resources Keep in CSSMin as-is for back-compat and to ensure library remains independent of MediaWiki. Moved down a few lines as there is no need to compute the md5 hash when we're returning a data URI. Previously md5_file was called twice during module builds (once for the fallback url, and another time when producing the embedded data uri). Applied to logo in SkinModule as example. To be applied elsewhere as needed. Without it, fallback is current behaviour (no cache invalidation). Bug: T99096 Change-Id: I7f38bfc1bea5c241bc4f8ec4f4b640fd65f2c04f
2016-01-22 22:58:03 +00:00
* @param string $localPath File directory exposed at $remotePath
* @param string $file Path to target file relative to $localPath
* @return string URL
*/
Fix resource path check when ResourceBasePath is an empty string If you have MediaWiki installed in the root of the domain, then $wgScriptPath and $wgResourceBasePath is an empty string. In HHVM and PHP, passing an empty string as the second parameter of strpos() causes a warning and returns false, which will cause the condition to fail, as if the path were not within the base path. So, normalize such paths. Using substr() instead of strpos() for a "starts with" check would have worked except that RelPath also fails when given an empty string. Bug: T127652 Change-Id: If7e94ae638d6834f7cc0f31f67a5fe6a2f74771c
2016-02-22 03:51:19 +00:00
public static function transformFilePath( $remotePathPrefix, $localPath, $file ) {
resourceloader: Remove wfExpandUrl() coupling from CSSMin There are three cases in CSSMin::remap where performs path resolution. 1. Absolute local path URLs to full URL. Example: url(/static/foo.png), url(/w/index.php?…), etc. These previously used wfExpandUrl(), which got the server name and protocol directly from $wgServer. We will now use the $remote parameter to get this information instead, which is generally set to something like https://wiki/w/resources/foo, and thus naturally contains the server name and protocol. The resolution is powered by the Net_URL2 library, allowing this to work outside Mediawiki as well. Some tests needed to change because they were calling CSSMin::remap with an incomplete $remote dummy values like "/" or "/w, because the test author (past me) was trying to be clever by not supplying it, knowing MW would ignore it. Now that it is consistently used, like normal calls from ResourceLoader would, the expected values will always be based on http://localhost/w, instead of sometimes the competing $wgServer value of `https://expand.example`. 2. Relative local path to full URL Example: url(foo.png), url(../foo.png), url(bar/foo.png) These were already using $remote. The only change is that they now use Net_URL2 library instead of blind string concatenation. One of the benefits of this is that we will no longer need to call wfRemoveDotSegments() to get rid of things like double slashes or redundant "../" sequences. Previously, thing like "foo//bar" or "foo/../bar" were cleaned up only due to wfRemoveDotSegments(). This is now naturally handled by Net_URL2. 3. Remote URLs Example: url(http://example.org/bar.png), url(//example.org/bar.png) This is generally not used in source code, but gadgets may use this, e.g. for upload.wikimedia.org or cross-wiki imports. Other changes: * One test case used spaces within the URL string in CSS, which the net_url2 library represents with percent-encoding instead. Same thing either way. Bug: T88914 Change-Id: Ibef70cc934c0ee8260a244c51bca9fb88c1c0d88
2020-09-11 21:35:37 +00:00
// This MUST match the equivalent logic in CSSMin::remapOne()
Stop logging and md5'ing if the file does not exist Also bring this more in line with CSSMin::remapOne() Bug: T283387 Change-Id: Iea5e68f48be96fae5b5640fe26c2095ca5196e88
2021-05-21 18:15:51 +00:00
$localFile = "$localPath/$file";
$url = "$remotePathPrefix/$file";
Prefer is_dir/is_file to file_exists Mirco-optimisation for performance, part II. Change-Id: Ie8d5332a510bd18456ebc4a851f4d47f6a010dcb
2022-01-04 10:03:09 +00:00
if ( is_file( $localFile ) ) {
Stop logging and md5'ing if the file does not exist Also bring this more in line with CSSMin::remapOne() Bug: T283387 Change-Id: Iea5e68f48be96fae5b5640fe26c2095ca5196e88
2021-05-21 18:15:51 +00:00
$hash = md5_file( $localFile );
if ( $hash === false ) {
wfLogWarning( __METHOD__ . ": Failed to hash $localFile" );
$hash = '';
}
$url .= '?' . substr( $hash, 0, 5 );
Centralise url handling for urls to static resources Keep in CSSMin as-is for back-compat and to ensure library remains independent of MediaWiki. Moved down a few lines as there is no need to compute the md5 hash when we're returning a data URI. Previously md5_file was called twice during module builds (once for the fallback url, and another time when producing the embedded data uri). Applied to logo in SkinModule as example. To be applied elsewhere as needed. Without it, fallback is current behaviour (no cache invalidation). Bug: T99096 Change-Id: I7f38bfc1bea5c241bc4f8ec4f4b640fd65f2c04f
2016-01-22 22:58:03 +00:00
}
Stop logging and md5'ing if the file does not exist Also bring this more in line with CSSMin::remapOne() Bug: T283387 Change-Id: Iea5e68f48be96fae5b5640fe26c2095ca5196e88
2021-05-21 18:15:51 +00:00
return $url;
Centralise url handling for urls to static resources Keep in CSSMin as-is for back-compat and to ensure library remains independent of MediaWiki. Moved down a few lines as there is no need to compute the md5 hash when we're returning a data URI. Previously md5_file was called twice during module builds (once for the fallback url, and another time when producing the embedded data uri). Applied to logo in SkinModule as example. To be applied elsewhere as needed. Without it, fallback is current behaviour (no cache invalidation). Bug: T99096 Change-Id: I7f38bfc1bea5c241bc4f8ec4f4b640fd65f2c04f
2016-01-22 22:58:03 +00:00
}
Step two in fixing OutputPage's documentation, more or less the same as r61647
2010-01-29 21:32:45 +00:00
/**
* Transform "media" attribute based on request parameters
*
Removed unneeded spaces and colons in @param and friends Also swapped some "$var type" to "type $var" or added missing types before the $var. Changed some other types to match the more common spelling. Makes beginning of some text in captial. Change-Id: Ic36c8c7820a6c2d603f1138130670c6bf6a1ca59
2014-04-08 15:29:17 +00:00
* @param string $media Current value of the "media" attribute
resourceloader: Remove obsolete logic for handheld=yes pageviews This feature was removed in 2013 (Ia8d79b4a, T49504), with further remnants removed in (1965df8df3, I7c24128f7b1). * Simplify the code of transformCssMedia() as it now only needs to support printable=yes. * Remove query parameter abstraction from test cases, use query array directly. * Remove internal use of the $handheld parameter. Function signature of makeLoaderQuery() left in-tact for compatibility with $extraData. This will be deprecated and removed in a follow-up commit. Bug: T49504 Bug: T32956 Change-Id: Id16ef3f4630538de1887ac2327eb317c6c6b1f9d
2021-10-09 01:41:15 +00:00
* @return string|null Modified value of the "media" attribute, or null to disable
(bug 43942) Skip screen sheets with media queries when printing * Add test suite, with regression tests as well as tests for the fix. Change-Id: I66d40673cc610370b66af0129412bc6495673f8d
2013-01-14 03:26:15 +00:00
* this stylesheet
Step two in fixing OutputPage's documentation, more or less the same as r61647
2010-01-29 21:32:45 +00:00
*/
(bug 26497) printable=yes and handheld=yes request parameters were broken. Fixed this in a bit of an ugly way, by propagating these parameters to load.php and running OutputPage::transformCssMedia() on it there. This revision also fixes the fact that ResourceLoader ignored $wgHandheldForIPhone, and no longer wraps styles in @media all { } because it's useless and breaks @import (which can't be used inside @media ; this is bug 26478)
2011-01-07 20:22:50 +00:00
public static function transformCssMedia( $media ) {
Remove the $wgHandheldForIPhone config variable entirely It only really made sense in pair with $wgHandheldStyle, and that has been removed in Ia8d79b4a. Remove irrelevant tests, adjust still relevant ones. Change-Id: I7c24128f7b148d0244538ad95bb60bf09ec4b5cb
2013-05-11 19:05:43 +00:00
global $wgRequest;
Tweaks for skins: * Moved SkinTemplate::addStyle() and related stuff to OutputPage so that it can be used non-SkinTemplate skins and avoid duplication with the actual OutputPage::addStyle() (the two functions have the same format). * Non-SkinTemplate skins now also load their CSS with <link> tags instead of @import. * Moved SkinTemplate::setupUserCss() to Skin. * Merged action=raw&gen=(js|css) for SkinTemplate and non-SkinTemplate skins, renamed functions to Skin::generateUserJs() and Skin::generateUserStyleSheet() and dropped a lot of cascading call which is a bit incomprehensible.
2008-08-21 14:09:57 +00:00
resourceloader: Remove obsolete logic for handheld=yes pageviews This feature was removed in 2013 (Ia8d79b4a, T49504), with further remnants removed in (1965df8df3, I7c24128f7b1). * Simplify the code of transformCssMedia() as it now only needs to support printable=yes. * Remove query parameter abstraction from test cases, use query array directly. * Remove internal use of the $handheld parameter. Function signature of makeLoaderQuery() left in-tact for compatibility with $extraData. This will be deprecated and removed in a follow-up commit. Bug: T49504 Bug: T32956 Change-Id: Id16ef3f4630538de1887ac2327eb317c6c6b1f9d
2021-10-09 01:41:15 +00:00
if ( $wgRequest->getBool( 'printable' ) ) {
// When browsing with printable=yes, apply "print" media styles
// as if they are screen styles (no media, media="").
if ( $media === 'print' ) {
return '';
}
(bug 43942) Skip screen sheets with media queries when printing * Add test suite, with regression tests as well as tests for the fix. Change-Id: I66d40673cc610370b66af0129412bc6495673f8d
2013-01-14 03:26:15 +00:00
resourceloader: Remove obsolete logic for handheld=yes pageviews This feature was removed in 2013 (Ia8d79b4a, T49504), with further remnants removed in (1965df8df3, I7c24128f7b1). * Simplify the code of transformCssMedia() as it now only needs to support printable=yes. * Remove query parameter abstraction from test cases, use query array directly. * Remove internal use of the $handheld parameter. Function signature of makeLoaderQuery() left in-tact for compatibility with $extraData. This will be deprecated and removed in a follow-up commit. Bug: T49504 Bug: T32956 Change-Id: Id16ef3f4630538de1887ac2327eb317c6c6b1f9d
2021-10-09 01:41:15 +00:00
// https://www.w3.org/TR/css3-mediaqueries/#syntax
//
// This regex will not attempt to understand a comma-separated media_query_list
// Example supported values for $media:
//
// 'screen', 'only screen', 'screen and (min-width: 982px)' ),
//
// Example NOT supported value for $media:
//
// '3d-glasses, screen, print and resolution > 90dpi'
//
// If it's a "printable" request, we disable all screen stylesheets.
$screenMediaQueryRegex = '/^(?:only\s+)?screen\b/i';
if ( preg_match( $screenMediaQueryRegex, $media ) === 1 ) {
return null;
Tweaks for skins: * Moved SkinTemplate::addStyle() and related stuff to OutputPage so that it can be used non-SkinTemplate skins and avoid duplication with the actual OutputPage::addStyle() (the two functions have the same format). * Non-SkinTemplate skins now also load their CSS with <link> tags instead of @import. * Moved SkinTemplate::setupUserCss() to Skin. * Merged action=raw&gen=(js|css) for SkinTemplate and non-SkinTemplate skins, renamed functions to Skin::generateUserJs() and Skin::generateUserStyleSheet() and dropped a lot of cascading call which is a bit incomprehensible.
2008-08-21 14:09:57 +00:00
}
}
return $media;
}
* Introduced OutputPage::addWikiMsg() and OutputPage::wrapWikiMsg(), to make it easier to avoid double-parsing. Double-parsing introduces differences between the old and new parsers. * Used the new functions in the relevant places throughout the core. Where this was not possible, used wfMsgNoTrans() to avoid double-parsing. * Calling wfEscapeWikiText() on a title before passing it to a message is unnecessary and causes bugs. The best idea is to let the message do its own escaping. Removed wfEscapeWikiText() from the caller of addedwatchtext, removedwatchtext, deletedtext, protect-text, nosuchusershort and movepage-moved. Updated all message files to use <nowiki>$1</nowiki> where appropriate in these messages. * Removed unnecessary empty message handling from semiprotectedpagewarning * Removed nonsense warning about the output of wfMsg() not being safe for inclusion in HTML. * Changed wfShellExec() to use passthru/ob_get_contents instead of exec(). This avoids corruption of whitespace in the response. * Added $wgValidateAllHtml feature, to optionally validate HTML on output. * Fixed missing newline in cascadeprotected in MessagesEn.php * Belatedly updated addedwatchtext to take account of the new "monobook" skin. No unwatch link in the sidebar anymore.
2008-02-18 07:25:35 +00:00
/**
* Add a wikitext-formatted message to the output.
* This is equivalent to:
*
Use wfMessage instead of deprecated wfMsg* Or $this->msg in special pages. Change-Id: I774a89d646615053c8424050e42ad95601f92543
2012-07-24 01:04:15 +00:00
* $wgOut->addWikiText( wfMessage( ... )->plain() )
Use real varargs in OutputPage Change-Id: I24821df0299d18a81e89d2e10671937f65c88bd8
2019-10-11 18:17:48 +00:00
*
* @param mixed ...$args
* Introduced OutputPage::addWikiMsg() and OutputPage::wrapWikiMsg(), to make it easier to avoid double-parsing. Double-parsing introduces differences between the old and new parsers. * Used the new functions in the relevant places throughout the core. Where this was not possible, used wfMsgNoTrans() to avoid double-parsing. * Calling wfEscapeWikiText() on a title before passing it to a message is unnecessary and causes bugs. The best idea is to let the message do its own escaping. Removed wfEscapeWikiText() from the caller of addedwatchtext, removedwatchtext, deletedtext, protect-text, nosuchusershort and movepage-moved. Updated all message files to use <nowiki>$1</nowiki> where appropriate in these messages. * Removed unnecessary empty message handling from semiprotectedpagewarning * Removed nonsense warning about the output of wfMsg() not being safe for inclusion in HTML. * Changed wfShellExec() to use passthru/ob_get_contents instead of exec(). This avoids corruption of whitespace in the response. * Added $wgValidateAllHtml feature, to optionally validate HTML on output. * Fixed missing newline in cascadeprotected in MessagesEn.php * Belatedly updated addedwatchtext to take account of the new "monobook" skin. No unwatch link in the sidebar anymore.
2008-02-18 07:25:35 +00:00
*/
Use real varargs in OutputPage Change-Id: I24821df0299d18a81e89d2e10671937f65c88bd8
2019-10-11 18:17:48 +00:00
public function addWikiMsg( ...$args ) {
* Introduced OutputPage::addWikiMsg() and OutputPage::wrapWikiMsg(), to make it easier to avoid double-parsing. Double-parsing introduces differences between the old and new parsers. * Used the new functions in the relevant places throughout the core. Where this was not possible, used wfMsgNoTrans() to avoid double-parsing. * Calling wfEscapeWikiText() on a title before passing it to a message is unnecessary and causes bugs. The best idea is to let the message do its own escaping. Removed wfEscapeWikiText() from the caller of addedwatchtext, removedwatchtext, deletedtext, protect-text, nosuchusershort and movepage-moved. Updated all message files to use <nowiki>$1</nowiki> where appropriate in these messages. * Removed unnecessary empty message handling from semiprotectedpagewarning * Removed nonsense warning about the output of wfMsg() not being safe for inclusion in HTML. * Changed wfShellExec() to use passthru/ob_get_contents instead of exec(). This avoids corruption of whitespace in the response. * Added $wgValidateAllHtml feature, to optionally validate HTML on output. * Fixed missing newline in cascadeprotected in MessagesEn.php * Belatedly updated addedwatchtext to take account of the new "monobook" skin. No unwatch link in the sidebar anymore.
2008-02-18 07:25:35 +00:00
$name = array_shift( $args );
$this->addWikiMsgArray( $name, $args );
}
/**
* Add a wikitext-formatted message to the output.
* Like addWikiMsg() except the parameters are taken as an array
* instead of a variable argument list.
*
Fixed some @params documentation Swapped some "$var type" to "type $var" or added missing types before the $var. Changed some other types to match the more common spelling. Makes beginning of some text in captial. Change-Id: I64e8cfe478cb0ba438f40b0631d6e9049cdab567
2014-04-10 18:50:10 +00:00
* @param string $name
* @param array $args
* Introduced OutputPage::addWikiMsg() and OutputPage::wrapWikiMsg(), to make it easier to avoid double-parsing. Double-parsing introduces differences between the old and new parsers. * Used the new functions in the relevant places throughout the core. Where this was not possible, used wfMsgNoTrans() to avoid double-parsing. * Calling wfEscapeWikiText() on a title before passing it to a message is unnecessary and causes bugs. The best idea is to let the message do its own escaping. Removed wfEscapeWikiText() from the caller of addedwatchtext, removedwatchtext, deletedtext, protect-text, nosuchusershort and movepage-moved. Updated all message files to use <nowiki>$1</nowiki> where appropriate in these messages. * Removed unnecessary empty message handling from semiprotectedpagewarning * Removed nonsense warning about the output of wfMsg() not being safe for inclusion in HTML. * Changed wfShellExec() to use passthru/ob_get_contents instead of exec(). This avoids corruption of whitespace in the response. * Added $wgValidateAllHtml feature, to optionally validate HTML on output. * Fixed missing newline in cascadeprotected in MessagesEn.php * Belatedly updated addedwatchtext to take account of the new "monobook" skin. No unwatch link in the sidebar anymore.
2008-02-18 07:25:35 +00:00
*/
* Use local context to get messages * Simply use getLang() instead of getContext()->getLang() now that this method exists * Also changed addWikiMsgArray() to use a Message object with addWikiText() for the parse operation; removed its $options parameter, no caller was using it
2011-10-23 08:13:52 +00:00
public function addWikiMsgArray( $name, $args ) {
* (bug 35014) Fix for r100529: including a special page no longer sets the page's title to the included page
2012-03-08 20:06:49 +00:00
$this->addHTML( $this->msg( $name, $args )->parseAsBlock() );
* Introduced OutputPage::addWikiMsg() and OutputPage::wrapWikiMsg(), to make it easier to avoid double-parsing. Double-parsing introduces differences between the old and new parsers. * Used the new functions in the relevant places throughout the core. Where this was not possible, used wfMsgNoTrans() to avoid double-parsing. * Calling wfEscapeWikiText() on a title before passing it to a message is unnecessary and causes bugs. The best idea is to let the message do its own escaping. Removed wfEscapeWikiText() from the caller of addedwatchtext, removedwatchtext, deletedtext, protect-text, nosuchusershort and movepage-moved. Updated all message files to use <nowiki>$1</nowiki> where appropriate in these messages. * Removed unnecessary empty message handling from semiprotectedpagewarning * Removed nonsense warning about the output of wfMsg() not being safe for inclusion in HTML. * Changed wfShellExec() to use passthru/ob_get_contents instead of exec(). This avoids corruption of whitespace in the response. * Added $wgValidateAllHtml feature, to optionally validate HTML on output. * Fixed missing newline in cascadeprotected in MessagesEn.php * Belatedly updated addedwatchtext to take account of the new "monobook" skin. No unwatch link in the sidebar anymore.
2008-02-18 07:25:35 +00:00
}
/**
* remove end of line whitespace * remove empty lines at end of file * remove "?>" where still present
2008-04-14 07:45:50 +00:00
* This function takes a number of message/argument specifications, wraps them in
* Introduced OutputPage::addWikiMsg() and OutputPage::wrapWikiMsg(), to make it easier to avoid double-parsing. Double-parsing introduces differences between the old and new parsers. * Used the new functions in the relevant places throughout the core. Where this was not possible, used wfMsgNoTrans() to avoid double-parsing. * Calling wfEscapeWikiText() on a title before passing it to a message is unnecessary and causes bugs. The best idea is to let the message do its own escaping. Removed wfEscapeWikiText() from the caller of addedwatchtext, removedwatchtext, deletedtext, protect-text, nosuchusershort and movepage-moved. Updated all message files to use <nowiki>$1</nowiki> where appropriate in these messages. * Removed unnecessary empty message handling from semiprotectedpagewarning * Removed nonsense warning about the output of wfMsg() not being safe for inclusion in HTML. * Changed wfShellExec() to use passthru/ob_get_contents instead of exec(). This avoids corruption of whitespace in the response. * Added $wgValidateAllHtml feature, to optionally validate HTML on output. * Fixed missing newline in cascadeprotected in MessagesEn.php * Belatedly updated addedwatchtext to take account of the new "monobook" skin. No unwatch link in the sidebar anymore.
2008-02-18 07:25:35 +00:00
* some overall structure, and then parses the result and adds it to the output.
*
Typofix in wrapWikiMsg() comment, smoother sentence Change-Id: Ic0ebc94cec68ee1515d4c98d49b26a6f9148e95e
2015-01-08 17:09:09 +00:00
* In the $wrap, $1 is replaced with the first message, $2 with the second,
* and so on. The subsequent arguments may be either
* 1) strings, in which case they are message names, or
* 2) arrays, in which case, within each array, the first element is the message
* name, and subsequent elements are the parameters to that message.
* Introduced OutputPage::addWikiMsg() and OutputPage::wrapWikiMsg(), to make it easier to avoid double-parsing. Double-parsing introduces differences between the old and new parsers. * Used the new functions in the relevant places throughout the core. Where this was not possible, used wfMsgNoTrans() to avoid double-parsing. * Calling wfEscapeWikiText() on a title before passing it to a message is unnecessary and causes bugs. The best idea is to let the message do its own escaping. Removed wfEscapeWikiText() from the caller of addedwatchtext, removedwatchtext, deletedtext, protect-text, nosuchusershort and movepage-moved. Updated all message files to use <nowiki>$1</nowiki> where appropriate in these messages. * Removed unnecessary empty message handling from semiprotectedpagewarning * Removed nonsense warning about the output of wfMsg() not being safe for inclusion in HTML. * Changed wfShellExec() to use passthru/ob_get_contents instead of exec(). This avoids corruption of whitespace in the response. * Added $wgValidateAllHtml feature, to optionally validate HTML on output. * Fixed missing newline in cascadeprotected in MessagesEn.php * Belatedly updated addedwatchtext to take account of the new "monobook" skin. No unwatch link in the sidebar anymore.
2008-02-18 07:25:35 +00:00
*
Typofix in wrapWikiMsg() comment, smoother sentence Change-Id: Ic0ebc94cec68ee1515d4c98d49b26a6f9148e95e
2015-01-08 17:09:09 +00:00
* Don't use this for messages that are not in the user's interface language.
* Parse messages in correct language, so plurals and grammar might even work correctly * Deprecated one unneeded parameter in one message
2008-04-18 06:31:37 +00:00
*
* Introduced OutputPage::addWikiMsg() and OutputPage::wrapWikiMsg(), to make it easier to avoid double-parsing. Double-parsing introduces differences between the old and new parsers. * Used the new functions in the relevant places throughout the core. Where this was not possible, used wfMsgNoTrans() to avoid double-parsing. * Calling wfEscapeWikiText() on a title before passing it to a message is unnecessary and causes bugs. The best idea is to let the message do its own escaping. Removed wfEscapeWikiText() from the caller of addedwatchtext, removedwatchtext, deletedtext, protect-text, nosuchusershort and movepage-moved. Updated all message files to use <nowiki>$1</nowiki> where appropriate in these messages. * Removed unnecessary empty message handling from semiprotectedpagewarning * Removed nonsense warning about the output of wfMsg() not being safe for inclusion in HTML. * Changed wfShellExec() to use passthru/ob_get_contents instead of exec(). This avoids corruption of whitespace in the response. * Added $wgValidateAllHtml feature, to optionally validate HTML on output. * Fixed missing newline in cascadeprotected in MessagesEn.php * Belatedly updated addedwatchtext to take account of the new "monobook" skin. No unwatch link in the sidebar anymore.
2008-02-18 07:25:35 +00:00
* For example:
*
Drop messagebox, warningbox, errorbox classes This can be done when relevant communities have been notified and the classes have rolled out of cached HTML. Bug: T270796 Change-Id: If35ebe88702bf97f2306c77b26d8b3cab4c0420d
2022-01-21 17:32:49 +00:00
* $wgOut->wrapWikiMsg( "<div class='customclass'>\n$1\n</div>", 'some-msg-key' );
* remove end of line whitespace * remove empty lines at end of file * remove "?>" where still present
2008-04-14 07:45:50 +00:00
*
* Introduced OutputPage::addWikiMsg() and OutputPage::wrapWikiMsg(), to make it easier to avoid double-parsing. Double-parsing introduces differences between the old and new parsers. * Used the new functions in the relevant places throughout the core. Where this was not possible, used wfMsgNoTrans() to avoid double-parsing. * Calling wfEscapeWikiText() on a title before passing it to a message is unnecessary and causes bugs. The best idea is to let the message do its own escaping. Removed wfEscapeWikiText() from the caller of addedwatchtext, removedwatchtext, deletedtext, protect-text, nosuchusershort and movepage-moved. Updated all message files to use <nowiki>$1</nowiki> where appropriate in these messages. * Removed unnecessary empty message handling from semiprotectedpagewarning * Removed nonsense warning about the output of wfMsg() not being safe for inclusion in HTML. * Changed wfShellExec() to use passthru/ob_get_contents instead of exec(). This avoids corruption of whitespace in the response. * Added $wgValidateAllHtml feature, to optionally validate HTML on output. * Fixed missing newline in cascadeprotected in MessagesEn.php * Belatedly updated addedwatchtext to take account of the new "monobook" skin. No unwatch link in the sidebar anymore.
2008-02-18 07:25:35 +00:00
* Is equivalent to:
*
Drop messagebox, warningbox, errorbox classes This can be done when relevant communities have been notified and the classes have rolled out of cached HTML. Bug: T270796 Change-Id: If35ebe88702bf97f2306c77b26d8b3cab4c0420d
2022-01-21 17:32:49 +00:00
* $wgOut->addWikiTextAsInterface( "<div class='customclass'>\n"
* . wfMessage( 'some-msg-key' )->plain() . "\n</div>" );
(bug 19226) First line renders differently on many UI messages Added newlines before $1 to wrapWikiMsgs of <div>s Also fixes r61052 utf-8 screw-up on RELEASE-NOTES.
2010-01-14 17:14:49 +00:00
*
includes: Replace implicit Bugzilla bug numbers with Phab ones It's unreasonable to expect newbies to know that "bug 12345" means "Task T14345" except where it doesn't, so let's just standardise on the real numbers. Change-Id: I6f59febaf8fc96e80f8cfc11f4356283f461142a
2017-02-20 22:44:19 +00:00
* The newline after the opening div is needed in some wikitext. See T21226.
More adding, updating and tweaking of documentation
2011-05-21 19:07:24 +00:00
*
Fixed some @params documentation Swapped some "$var type" to "type $var" or added missing types before the $var. Changed some other types to match the more common spelling. Makes beginning of some text in captial. Change-Id: I64e8cfe478cb0ba438f40b0631d6e9049cdab567
2014-04-10 18:50:10 +00:00
* @param string $wrap
Use real varargs in OutputPage Change-Id: I24821df0299d18a81e89d2e10671937f65c88bd8
2019-10-11 18:17:48 +00:00
* @param mixed ...$msgSpecs
* Introduced OutputPage::addWikiMsg() and OutputPage::wrapWikiMsg(), to make it easier to avoid double-parsing. Double-parsing introduces differences between the old and new parsers. * Used the new functions in the relevant places throughout the core. Where this was not possible, used wfMsgNoTrans() to avoid double-parsing. * Calling wfEscapeWikiText() on a title before passing it to a message is unnecessary and causes bugs. The best idea is to let the message do its own escaping. Removed wfEscapeWikiText() from the caller of addedwatchtext, removedwatchtext, deletedtext, protect-text, nosuchusershort and movepage-moved. Updated all message files to use <nowiki>$1</nowiki> where appropriate in these messages. * Removed unnecessary empty message handling from semiprotectedpagewarning * Removed nonsense warning about the output of wfMsg() not being safe for inclusion in HTML. * Changed wfShellExec() to use passthru/ob_get_contents instead of exec(). This avoids corruption of whitespace in the response. * Added $wgValidateAllHtml feature, to optionally validate HTML on output. * Fixed missing newline in cascadeprotected in MessagesEn.php * Belatedly updated addedwatchtext to take account of the new "monobook" skin. No unwatch link in the sidebar anymore.
2008-02-18 07:25:35 +00:00
*/
Use real varargs in OutputPage Change-Id: I24821df0299d18a81e89d2e10671937f65c88bd8
2019-10-11 18:17:48 +00:00
public function wrapWikiMsg( $wrap, ...$msgSpecs ) {
* Introduced OutputPage::addWikiMsg() and OutputPage::wrapWikiMsg(), to make it easier to avoid double-parsing. Double-parsing introduces differences between the old and new parsers. * Used the new functions in the relevant places throughout the core. Where this was not possible, used wfMsgNoTrans() to avoid double-parsing. * Calling wfEscapeWikiText() on a title before passing it to a message is unnecessary and causes bugs. The best idea is to let the message do its own escaping. Removed wfEscapeWikiText() from the caller of addedwatchtext, removedwatchtext, deletedtext, protect-text, nosuchusershort and movepage-moved. Updated all message files to use <nowiki>$1</nowiki> where appropriate in these messages. * Removed unnecessary empty message handling from semiprotectedpagewarning * Removed nonsense warning about the output of wfMsg() not being safe for inclusion in HTML. * Changed wfShellExec() to use passthru/ob_get_contents instead of exec(). This avoids corruption of whitespace in the response. * Added $wgValidateAllHtml feature, to optionally validate HTML on output. * Fixed missing newline in cascadeprotected in MessagesEn.php * Belatedly updated addedwatchtext to take account of the new "monobook" skin. No unwatch link in the sidebar anymore.
2008-02-18 07:25:35 +00:00
$s = $wrap;
foreach ( $msgSpecs as $n => $spec ) {
if ( is_array( $spec ) ) {
$args = $spec;
$name = array_shift( $args );
fix some spacing Change-Id: I8f976013f33c5818e4402604fe8610aa3f43b0c6
2013-02-03 20:05:24 +00:00
} else {
Convert all array() syntax to [] Per wikitech-l consensus: https://lists.wikimedia.org/pipermail/wikitech-l/2016-February/084821.html Notes: * Disabled CallTimePassByReference due to false positives (T127163) Change-Id: I2c8ce713ce6600a0bb7bf67537c87044c7a45c4b
2016-02-17 09:09:32 +00:00
$args = [];
* Introduced OutputPage::addWikiMsg() and OutputPage::wrapWikiMsg(), to make it easier to avoid double-parsing. Double-parsing introduces differences between the old and new parsers. * Used the new functions in the relevant places throughout the core. Where this was not possible, used wfMsgNoTrans() to avoid double-parsing. * Calling wfEscapeWikiText() on a title before passing it to a message is unnecessary and causes bugs. The best idea is to let the message do its own escaping. Removed wfEscapeWikiText() from the caller of addedwatchtext, removedwatchtext, deletedtext, protect-text, nosuchusershort and movepage-moved. Updated all message files to use <nowiki>$1</nowiki> where appropriate in these messages. * Removed unnecessary empty message handling from semiprotectedpagewarning * Removed nonsense warning about the output of wfMsg() not being safe for inclusion in HTML. * Changed wfShellExec() to use passthru/ob_get_contents instead of exec(). This avoids corruption of whitespace in the response. * Added $wgValidateAllHtml feature, to optionally validate HTML on output. * Fixed missing newline in cascadeprotected in MessagesEn.php * Belatedly updated addedwatchtext to take account of the new "monobook" skin. No unwatch link in the sidebar anymore.
2008-02-18 07:25:35 +00:00
$name = $spec;
}
Fixes for Id2ac68ef (a6e8b13) - drop no longer used $options variable - use $this->msg() instead of wfMessage() Change-Id: Ib690f3a8e63f3ede993e9cd34e2831ef18098ed9
2012-09-03 19:38:59 +00:00
$s = str_replace( '$' . ( $n + 1 ), $this->msg( $name, $args )->plain(), $s );
* Introduced OutputPage::addWikiMsg() and OutputPage::wrapWikiMsg(), to make it easier to avoid double-parsing. Double-parsing introduces differences between the old and new parsers. * Used the new functions in the relevant places throughout the core. Where this was not possible, used wfMsgNoTrans() to avoid double-parsing. * Calling wfEscapeWikiText() on a title before passing it to a message is unnecessary and causes bugs. The best idea is to let the message do its own escaping. Removed wfEscapeWikiText() from the caller of addedwatchtext, removedwatchtext, deletedtext, protect-text, nosuchusershort and movepage-moved. Updated all message files to use <nowiki>$1</nowiki> where appropriate in these messages. * Removed unnecessary empty message handling from semiprotectedpagewarning * Removed nonsense warning about the output of wfMsg() not being safe for inclusion in HTML. * Changed wfShellExec() to use passthru/ob_get_contents instead of exec(). This avoids corruption of whitespace in the response. * Added $wgValidateAllHtml feature, to optionally validate HTML on output. * Fixed missing newline in cascadeprotected in MessagesEn.php * Belatedly updated addedwatchtext to take account of the new "monobook" skin. No unwatch link in the sidebar anymore.
2008-02-18 07:25:35 +00:00
}
Use OutputPage::addWikiTextAsInterface() instead of untidy addWikiText() This change ensures that the output is tidy, and is necessary to support future parsers which will not be able to produce untidy output. Bug: T198214 Change-Id: I743f4185a03403f8d9b9db010ff1ee4e9342e062
2018-09-25 15:02:07 +00:00
$this->addWikiTextAsInterface( $s );
* Introduced OutputPage::addWikiMsg() and OutputPage::wrapWikiMsg(), to make it easier to avoid double-parsing. Double-parsing introduces differences between the old and new parsers. * Used the new functions in the relevant places throughout the core. Where this was not possible, used wfMsgNoTrans() to avoid double-parsing. * Calling wfEscapeWikiText() on a title before passing it to a message is unnecessary and causes bugs. The best idea is to let the message do its own escaping. Removed wfEscapeWikiText() from the caller of addedwatchtext, removedwatchtext, deletedtext, protect-text, nosuchusershort and movepage-moved. Updated all message files to use <nowiki>$1</nowiki> where appropriate in these messages. * Removed unnecessary empty message handling from semiprotectedpagewarning * Removed nonsense warning about the output of wfMsg() not being safe for inclusion in HTML. * Changed wfShellExec() to use passthru/ob_get_contents instead of exec(). This avoids corruption of whitespace in the response. * Added $wgValidateAllHtml feature, to optionally validate HTML on output. * Fixed missing newline in cascadeprotected in MessagesEn.php * Belatedly updated addedwatchtext to take account of the new "monobook" skin. No unwatch link in the sidebar anymore.
2008-02-18 07:25:35 +00:00
}
Part 2 at solving conflict with vector and js edit Add OutputPage::includeJQuery() for loading the jQuery currently shipped with core. Extension can use this function to avoid loading it multiple times. No support for plugins etc. Changes Translate extension to use that function, restored compatability with 1.15 for at least this part of code.
2010-01-23 20:11:42 +00:00
Make TOC hideable Currently, if an extension doesn't want a TOC, it has to remove it manually. This change wraps the TOC in markers that make it easy to remove it in ParserOutput on demand without fragmenting the parser cache with stuff like "use/not use TOC". Change-Id: I2889bcb9eb999c9049601e92440132118e1a8a41
2013-08-22 22:22:03 +00:00
/**
Move loading of mediawiki.toc from Parser to Skin This was the only addModules() call ever to be inside Parser. Introduced in a54ef1a203d. Prior to that, mediawiki.toc had always been loaded by OutputPage (via mediawiki.util; and before that, via wikibits). This patch restores that, and also fixes T130632 by making OutputPage get it from the Skin, instead of hardcoding this somewhere in addParserOutput(). * Remove deprecated method OutputPage::enableTOC(). * Move mEnableTOC to addParserOutputText(). Bug: T130632 Change-Id: Iaad84d241a4c4348c712ac1087a664b8c9c46da4
2017-05-20 13:39:40 +00:00
* Whether the output has a table of contents
Make TOC hideable Currently, if an extension doesn't want a TOC, it has to remove it manually. This change wraps the TOC in markers that make it easy to remove it in ParserOutput on demand without fragmenting the parser cache with stuff like "use/not use TOC". Change-Id: I2889bcb9eb999c9049601e92440132118e1a8a41
2013-08-22 22:22:03 +00:00
* @return bool
* @since 1.22
*/
public function isTOCEnabled() {
return $this->mEnableTOC;
}
Suppress section edit links with action=render action=render is often used to pull the rendered HTML of an article for inclusion in a different context. More often than not, the edit links are not desired in that context. This reverts commit a04b5d5dcb1cef9cff45170b3eee6651337678be and redoes Iab02cbd42f2f93f0f5264a74691ae1b84f296f12. Bug: 19415 Change-Id: I26213653c017c2e19a6f6799f3ea0676ff8524d1
2014-03-11 20:00:02 +00:00
Add OutputPage::enableOOUI() to set up OOUI stuff when needed Initialize OOUI theme and set OOUI element directionality, load necessary RL modules. Bug: T85767 Change-Id: I9aa430969e4250bd2bd6c396b6133a35ca5bf46b
2015-01-05 20:18:57 +00:00
/**
Allow to enable OOUI via a parser tag extension This change adds the possibility to enable OOUI out of the parser, which enabled parser tag functions to easily enable OOUI, if they need it, for every page view out of the function that handles the parser tag. Bug: T106949 Change-Id: If1e139d4f07be98e418e11470794ea42e8a9b2eb
2015-07-25 15:32:08 +00:00
* Helper function to setup the PHP implementation of OOUI to use in this request.
Add OutputPage::enableOOUI() to set up OOUI stuff when needed Initialize OOUI theme and set OOUI element directionality, load necessary RL modules. Bug: T85767 Change-Id: I9aa430969e4250bd2bd6c396b6133a35ca5bf46b
2015-01-05 20:18:57 +00:00
*
Allow to enable OOUI via a parser tag extension This change adds the possibility to enable OOUI out of the parser, which enabled parser tag functions to easily enable OOUI, if they need it, for every page view out of the function that handles the parser tag. Bug: T106949 Change-Id: If1e139d4f07be98e418e11470794ea42e8a9b2eb
2015-07-25 15:32:08 +00:00
* @since 1.26
Fix PHPDoc type for instance variables and methods Should be "string" not "String" and "array" not "Array" in @param, @return and @var use cases. Also, minor typo fixes. Change-Id: I9d5ebc5b741c6560907b95f7c0c4039da2861f4a
2018-10-21 11:38:39 +00:00
* @param string $skinName The Skin name to determine the correct OOUI theme
* @param string $dir Language direction
Add OutputPage::enableOOUI() to set up OOUI stuff when needed Initialize OOUI theme and set OOUI element directionality, load necessary RL modules. Bug: T85767 Change-Id: I9aa430969e4250bd2bd6c396b6133a35ca5bf46b
2015-01-05 20:18:57 +00:00
*/
Refactor registration of OOjs UI resource modules includes/resourceloader/ResourceLoaderOOUIModule.php * New trait centralizing some logic for dealing with OOjs UI themes, previously duplicated in OutputPage, ResourcesOOUI.php and ResourceLoaderOOUIImageModule. * Follow-up change I74362f0fc215b26f1f104ce7bdbbac1e106736ad uses this as a base to allow skins/extensions to define new OOjs UI themes. resources/Resources.php resources/ResourcesOOUI.php includes/resourceloader/ResourceLoader.php * OOjs UI resource module definitions are moved back to their rightly place in Resources.php. They are again (almost) normal and static. * Theme-specific logic is now handled by the module code, definitions only specify 'themeScripts'/'themeStyles'/'themeImages'. * ResourcesOOUI.php is deleted and no longer loaded by ResourceLoader. includes/resourceloader/ResourceLoaderOOUIFileModule.php includes/resourceloader/ResourceLoaderOOUIImageModule.php * Glue code previously existing in ResourcesOOUI.php now lives here. * Use the ResourceLoaderOOUIModule trait to avoid code duplication. Change-Id: I39cc2a735d9625c87bf4ede6f5fb0ec441d47dcc
2017-03-17 01:32:59 +00:00
public static function setupOOUI( $skinName = 'default', $dir = 'ltr' ) {
ResourceLoader namespace (attempt 2) Move ResourceLoader classes to their own namespace. Strip the "ResourceLoader" prefix from all except ResourceLoader itself. Move the tests by analogy. I used a namespace alias "RL" in some callers since RL\Module is less ambiguous at the call site than just "Module". I did not address DependencyStore which continues to have a non-standard location and namespace. Revert of a241d83e0a6dabedf. Bug: T308718 Change-Id: Id08a220e1d6085e2b33f3f6c9d0e3935a4204659
2022-05-06 09:09:56 +00:00
$themes = RL\OOUIFileModule::getSkinThemeMap();
Use PHP 7 '??' operator instead of '?:' with 'isset()' where convenient Find: /isset\(\s*([^()]+?)\s*\)\s*\?\s*\1\s*:\s*/ Replace with: '\1 ?? ' (Everywhere except includes/PHPVersionCheck.php) (Then, manually fix some line length and indentation issues) Then manually reviewed the replacements for cases where confusing operator precedence would result in incorrect results (fixing those in I478db046a1cc162c6767003ce45c9b56270f3372). Change-Id: I33b421c8cb11cdd4ce896488c9ff5313f03a38cf
2017-10-06 22:17:58 +00:00
$theme = $themes[$skinName] ?? $themes['default'];
Refactor registration of OOjs UI resource modules includes/resourceloader/ResourceLoaderOOUIModule.php * New trait centralizing some logic for dealing with OOjs UI themes, previously duplicated in OutputPage, ResourcesOOUI.php and ResourceLoaderOOUIImageModule. * Follow-up change I74362f0fc215b26f1f104ce7bdbbac1e106736ad uses this as a base to allow skins/extensions to define new OOjs UI themes. resources/Resources.php resources/ResourcesOOUI.php includes/resourceloader/ResourceLoader.php * OOjs UI resource module definitions are moved back to their rightly place in Resources.php. They are again (almost) normal and static. * Theme-specific logic is now handled by the module code, definitions only specify 'themeScripts'/'themeStyles'/'themeImages'. * ResourcesOOUI.php is deleted and no longer loaded by ResourceLoader. includes/resourceloader/ResourceLoaderOOUIFileModule.php includes/resourceloader/ResourceLoaderOOUIImageModule.php * Glue code previously existing in ResourcesOOUI.php now lives here. * Use the ResourceLoaderOOUIModule trait to avoid code duplication. Change-Id: I39cc2a735d9625c87bf4ede6f5fb0ec441d47dcc
2017-03-17 01:32:59 +00:00
// For example, 'OOUI\WikimediaUITheme'.
OutputPage: Load skin-appropriate OOUI theme We allow skins to choose the appropriate OOUI theme for it using 'SkinOOUIThemes', but we ignored that here. Luckily this was not a problem for the Apex theme (the only one other than the default, MediaWiki theme), the additional CSS classes this generated were just ignored. Depends on Apex and MonoBook skins being updated to use canonical case for theme names, see I1a88092f0ceb210c642bbee875e368ded3c186f8. Change-Id: I368804cca141acf9cc15a348d9b7244046395a70
2015-07-23 16:37:53 +00:00
$themeClass = "OOUI\\{$theme}Theme";
OOUI\Theme::setSingleton( new $themeClass() );
Allow to enable OOUI via a parser tag extension This change adds the possibility to enable OOUI out of the parser, which enabled parser tag functions to easily enable OOUI, if they need it, for every page view out of the function that handles the parser tag. Bug: T106949 Change-Id: If1e139d4f07be98e418e11470794ea42e8a9b2eb
2015-07-25 15:32:08 +00:00
OOUI\Element::setDefaultDir( $dir );
}
/**
* Add ResourceLoader module styles for OOUI and set up the PHP implementation of it for use with
* MediaWiki and this OutputPage instance.
*
* @since 1.25
*/
public function enableOOUI() {
self::setupOOUI(
strtolower( $this->getSkin()->getSkinName() ),
$this->getLanguage()->getDir()
);
Convert all array() syntax to [] Per wikitech-l consensus: https://lists.wikimedia.org/pipermail/wikitech-l/2016-February/084821.html Notes: * Disabled CallTimePassByReference due to false positives (T127163) Change-Id: I2c8ce713ce6600a0bb7bf67537c87044c7a45c4b
2016-02-17 09:09:32 +00:00
$this->addModuleStyles( [
resources: Load OOjs UI from its four parts See the task for more details. This is a backwards-compatible change. If your script only needs a subset of OOjs UI functionality, you can use one of the new smaller modules instead of the old big one. New modules: oojs-ui-core The core JavaScript library. oojs-ui-widgets Additional widgets and layouts module. oojs-ui-toolbars Toolbar and tools module. oojs-ui-windows Windows and dialogs module. Changed modules: oojs-ui.styles Now correctly only loads the styles needed by OOjs UI PHP. oojs-ui Now just loads core+widgets+toolbars+windows as dependencies. Using the new modules in I58799e22f9c0a2f78c1b4a02c4b7af576157883a. Bug: T113677 Change-Id: I0a3bf8fb25fb82325705a473cebd883e20b3ab8d
2016-02-01 22:28:13 +00:00
'oojs-ui-core.styles',
Use ResourceLoaderImageModule to serve icons for OOjs UI Benefits: * Full per-language icons support. Icons that differ for each language (such as the 'Bold' icon) will now always display correctly according to user interface language, even on old browsers. * MediaWiki UI icons support. When the 'mediawiki.ui.icon' module is loaded, you can use syntaxes such as below to display any OOUI icons (from the packs that were loaded) without involving OOUI itself. <div class="mw-ui-icon mw-ui-icon-before mw-ui-icon-check">OK</div> <div class="mw-ui-icon mw-ui-icon-after mw-ui-icon-check">OK</div> <div class="mw-ui-icon mw-ui-icon-element mw-ui-icon-check">OK</div> Summary of changes: * Resources.php: * Remove icons CSS files. Include image data JSON files instead. * Split the images from 'oojs-ui.styles' module to separate ones. * OutputPage: Update enableOOUI() method for newly split modules. * ResourceLoaderImageModule: Make it possible to load image data from a JSON file. * update-oojs-ui.sh: Copy source files rather than distribution for icon packs. This is not an improvement when it comes to code quality, though :( Issues include some nasty code duplication, using "source code" (image definitions) from OOUI rather than just distribution files, and hacky methods to load image data from JSON files live. Bug: T92551 Change-Id: Id369ecaec7048dcf68ba1e4df748362760533782
2015-04-07 16:02:46 +00:00
'oojs-ui.styles.indicators',
Implement NamespaceInputWidget * Add PHP version of NamespaceInputWidget, co-authored by Florian, which consists of a DropdownInputWidget offering a choice of namespaces and two CheckboxInputWidgets allowing to also match associated namespace (talk/content) or to invert the choice. * Add an incomplete JS version of NamespaceInputWidget, which is only really functional when infused from the PHP version (it can't generate the dropdown by itself, for example). Implement some JS to improve the experience of selecting the "all namespaces" option in the dropdown (by disabling the checkboxes when this happens). * Split off a 'mediawiki.widgets.styles' module, which has the basic styles for PHP widgets which are to be loaded in the head. Make OutputPage::enableOOUI() also add this module (which should stay reasonably small). * Use the new widget in HTMLForm's HTMLSelectNamespace field. It can be seen in action on Special:LinkSearch, for example. Co-Authored-By: Florian <florian.schmidt.welzow@t-online.de> Co-Authored-By: Bartosz Dziewoński <matma.rex@gmail.com> Change-Id: I5cbfa9d0f6a8641148ce476b7dbe65e9096b4485
2015-07-14 20:30:06 +00:00
'mediawiki.widgets.styles',
Allow loading styles for arbitrary OOUI icon packs You can now create ResourceLoader modules for arbitrary sets of OOUI icons. This is an alternative to I8af783666a2b23a938af93c1b56fee619219eaf5. Update dependencies of OOUI's modules to use custom icon packs instead of default icon packs. Bug: T160690 Change-Id: Icf9560da79c91e56c7a3f4c0de01dd057f5aa00d
2018-02-12 19:41:44 +00:00
'oojs-ui-core.icons',
Convert all array() syntax to [] Per wikitech-l consensus: https://lists.wikimedia.org/pipermail/wikitech-l/2016-February/084821.html Notes: * Disabled CallTimePassByReference due to false positives (T127163) Change-Id: I2c8ce713ce6600a0bb7bf67537c87044c7a45c4b
2016-02-17 09:09:32 +00:00
] );
Add OutputPage::enableOOUI() to set up OOUI stuff when needed Initialize OOUI theme and set OOUI element directionality, load necessary RL modules. Bug: T85767 Change-Id: I9aa430969e4250bd2bd6c396b6133a35ca5bf46b
2015-01-05 20:18:57 +00:00
}
Preload the logo using link rel="preload" http header This greatly increases the priority of loading the logo on browsers that support rel="preload". Bug: T100999 Change-Id: I0738fcc0a575153dab65016fa87faaa9b8b97a9d
2015-06-01 16:58:42 +00:00
Initial support for Content Security Policy, disabled by default The primary goal here is a defense in depth measure to stop an attacker who found a bug in the parser allowing them to insert malicious attributes. This wouldn't stop someone who could insert a full script tag (since at current it can't distinguish between malicious and legit user js). It also would not prevent DOM-based or reflected XSS for anons, as the nonce value is guessable for anons when receiving a response cached by varnish. However, the limited protection of just stopping stored XSS where the attacker only has control of attributes, is still a big win in my opinion. (But it wouldn't prevent someone who has that type of xss from abusing things like data-ooui attribute). This will likely break many gadgets. Its expected that any sort of rollout on Wikimedia will be done very slowly, with lots of testing and the report-only option to begin with. This is behind feature flags that are off by default, so merging this patch should not cause any change in default behaviour. This may break some extensions (The most obvious one is charinsert (See fe648d41005), but will probably need some testing in report-only mode to see if anything else breaks) This uses the unsafe-eval option of CSP, in order to support RL's local storage thingy. For better security, we may want to remove some of the sillier uses of eval (e.g. jquery.ui.datepicker.js). For more info, see spec: https://www.w3.org/TR/CSP2/ Additionally see: https://www.mediawiki.org/wiki/Requests_for_comment/Content-Security-Policy Bug: T135963 Change-Id: I80f6f469ba4c0b608385483457df96ccb7429ae5
2016-02-29 04:13:10 +00:00
/**
* Get (and set if not yet set) the CSP nonce.
*
* This value needs to be included in any <script> tags on the
* page.
*
Fix various documentation related to false Some function already document "False on failure", but does not document it on the used type Found by phan strict checks Change-Id: I12eb8bbc99179833ee3e42c1a7d1dc1443682ca6
2022-03-03 20:33:11 +00:00
* @return string|false Nonce or false to mean don't output nonce
Initial support for Content Security Policy, disabled by default The primary goal here is a defense in depth measure to stop an attacker who found a bug in the parser allowing them to insert malicious attributes. This wouldn't stop someone who could insert a full script tag (since at current it can't distinguish between malicious and legit user js). It also would not prevent DOM-based or reflected XSS for anons, as the nonce value is guessable for anons when receiving a response cached by varnish. However, the limited protection of just stopping stored XSS where the attacker only has control of attributes, is still a big win in my opinion. (But it wouldn't prevent someone who has that type of xss from abusing things like data-ooui attribute). This will likely break many gadgets. Its expected that any sort of rollout on Wikimedia will be done very slowly, with lots of testing and the report-only option to begin with. This is behind feature flags that are off by default, so merging this patch should not cause any change in default behaviour. This may break some extensions (The most obvious one is charinsert (See fe648d41005), but will probably need some testing in report-only mode to see if anything else breaks) This uses the unsafe-eval option of CSP, in order to support RL's local storage thingy. For better security, we may want to remove some of the sillier uses of eval (e.g. jquery.ui.datepicker.js). For more info, see spec: https://www.w3.org/TR/CSP2/ Additionally see: https://www.mediawiki.org/wiki/Requests_for_comment/Content-Security-Policy Bug: T135963 Change-Id: I80f6f469ba4c0b608385483457df96ccb7429ae5
2016-02-29 04:13:10 +00:00
* @since 1.32
Mild refactoring of ContentSecurityPolicy This is to make it behave in a more object orientied way. The goal is to make it be easier to allow extensions to mark certain pages as requiring a different policy (For example, CodeEditor extension uses a blob: url with a WebWorker. We don't want to include that on the policy of every page, but allow the extension to mark it as required whenever needed). This commit does not change code behaviour in any way. Change-Id: I4bf53dabb6e6c5446cea99a64db68b300cef2fd4
2019-10-28 05:01:17 +00:00
* @deprecated Since 1.35 use getCSP()->getNonce() instead
Initial support for Content Security Policy, disabled by default The primary goal here is a defense in depth measure to stop an attacker who found a bug in the parser allowing them to insert malicious attributes. This wouldn't stop someone who could insert a full script tag (since at current it can't distinguish between malicious and legit user js). It also would not prevent DOM-based or reflected XSS for anons, as the nonce value is guessable for anons when receiving a response cached by varnish. However, the limited protection of just stopping stored XSS where the attacker only has control of attributes, is still a big win in my opinion. (But it wouldn't prevent someone who has that type of xss from abusing things like data-ooui attribute). This will likely break many gadgets. Its expected that any sort of rollout on Wikimedia will be done very slowly, with lots of testing and the report-only option to begin with. This is behind feature flags that are off by default, so merging this patch should not cause any change in default behaviour. This may break some extensions (The most obvious one is charinsert (See fe648d41005), but will probably need some testing in report-only mode to see if anything else breaks) This uses the unsafe-eval option of CSP, in order to support RL's local storage thingy. For better security, we may want to remove some of the sillier uses of eval (e.g. jquery.ui.datepicker.js). For more info, see spec: https://www.w3.org/TR/CSP2/ Additionally see: https://www.mediawiki.org/wiki/Requests_for_comment/Content-Security-Policy Bug: T135963 Change-Id: I80f6f469ba4c0b608385483457df96ccb7429ae5
2016-02-29 04:13:10 +00:00
*/
public function getCSPNonce() {
Mild refactoring of ContentSecurityPolicy This is to make it behave in a more object orientied way. The goal is to make it be easier to allow extensions to mark certain pages as requiring a different policy (For example, CodeEditor extension uses a blob: url with a WebWorker. We don't want to include that on the policy of every page, but allow the extension to mark it as required whenever needed). This commit does not change code behaviour in any way. Change-Id: I4bf53dabb6e6c5446cea99a64db68b300cef2fd4
2019-10-28 05:01:17 +00:00
return $this->CSP->getNonce();
Initial support for Content Security Policy, disabled by default The primary goal here is a defense in depth measure to stop an attacker who found a bug in the parser allowing them to insert malicious attributes. This wouldn't stop someone who could insert a full script tag (since at current it can't distinguish between malicious and legit user js). It also would not prevent DOM-based or reflected XSS for anons, as the nonce value is guessable for anons when receiving a response cached by varnish. However, the limited protection of just stopping stored XSS where the attacker only has control of attributes, is still a big win in my opinion. (But it wouldn't prevent someone who has that type of xss from abusing things like data-ooui attribute). This will likely break many gadgets. Its expected that any sort of rollout on Wikimedia will be done very slowly, with lots of testing and the report-only option to begin with. This is behind feature flags that are off by default, so merging this patch should not cause any change in default behaviour. This may break some extensions (The most obvious one is charinsert (See fe648d41005), but will probably need some testing in report-only mode to see if anything else breaks) This uses the unsafe-eval option of CSP, in order to support RL's local storage thingy. For better security, we may want to remove some of the sillier uses of eval (e.g. jquery.ui.datepicker.js). For more info, see spec: https://www.w3.org/TR/CSP2/ Additionally see: https://www.mediawiki.org/wiki/Requests_for_comment/Content-Security-Policy Bug: T135963 Change-Id: I80f6f469ba4c0b608385483457df96ccb7429ae5
2016-02-29 04:13:10 +00:00
}
Show copyright based on $output->hasCopyright() Introducing setCopyright/hasCopyright in OutputPage to make the showing of the copyright message controllable by pages other than articles. hasCopyright defaults to isArticle() to preserve the rule that all article should show the copyright. This is used by Flow to show the copyright on various pages that contain user-generated content. Bug: T184960 Change-Id: I3a50dbcedc6b119b9262c50cb3a84b0dd230fb3d
2018-10-03 13:12:37 +00:00
Mild refactoring of ContentSecurityPolicy This is to make it behave in a more object orientied way. The goal is to make it be easier to allow extensions to mark certain pages as requiring a different policy (For example, CodeEditor extension uses a blob: url with a WebWorker. We don't want to include that on the policy of every page, but allow the extension to mark it as required whenever needed). This commit does not change code behaviour in any way. Change-Id: I4bf53dabb6e6c5446cea99a64db68b300cef2fd4
2019-10-28 05:01:17 +00:00
/**
* Get the ContentSecurityPolicy object
*
* @since 1.35
* @return ContentSecurityPolicy
*/
public function getCSP() {
return $this->CSP;
}
Introduce OutputPage::tailElement * SkinMustache::tailElement is removed, SkinMustache now calls the OutputPage method Bug: T257704 Change-Id: Ief537d2c317595d192a37e8493369e08e1fb2f45
2021-08-26 17:55:20 +00:00
/**
* The final bits that go to the bottom of a page
* HTML document including the closing tags
*
* @internal
* @since 1.37
* @param Skin $skin
* @return string
*/
public function tailElement( $skin ) {
Soft deprecate Skin::bottomScripts Also prepare OutputPage::getBottomScripts to handle the work of SkinAfterBottomScripts hook in the future. Bug: T257704 Change-Id: Ib0a8b72faa88259c24b5fb7e9d5a908423494247
2021-08-30 10:32:02 +00:00
// T257704: Temporarily run skin hook here pending
// creation dedicated outputpage hook for this
$extraHtml = '';
$this->getHookRunner()->onSkinAfterBottomScripts( $skin, $extraHtml );
Introduce OutputPage::tailElement * SkinMustache::tailElement is removed, SkinMustache now calls the OutputPage method Bug: T257704 Change-Id: Ief537d2c317595d192a37e8493369e08e1fb2f45
2021-08-26 17:55:20 +00:00
$tail = [
MWDebug::getDebugHTML( $skin ),
Soft deprecate Skin::bottomScripts Also prepare OutputPage::getBottomScripts to handle the work of SkinAfterBottomScripts hook in the future. Bug: T257704 Change-Id: Ib0a8b72faa88259c24b5fb7e9d5a908423494247
2021-08-30 10:32:02 +00:00
$this->getBottomScripts( $extraHtml ),
Introduce OutputPage::tailElement * SkinMustache::tailElement is removed, SkinMustache now calls the OutputPage method Bug: T257704 Change-Id: Ief537d2c317595d192a37e8493369e08e1fb2f45
2021-08-26 17:55:20 +00:00
wfReportTime( $this->getCSP()->getNonce() ),
Use WrappedString to build strings in MWDebug This gives a small performance benefit on developer wikis when the html page is build. Change-Id: Ic90bde09500776ae24952d627e582fbce688e36c
2021-09-29 12:43:22 +00:00
MWDebug::getHTMLDebugLog(),
Html::closeElement( 'body' ),
Html::closeElement( 'html' ),
Introduce OutputPage::tailElement * SkinMustache::tailElement is removed, SkinMustache now calls the OutputPage method Bug: T257704 Change-Id: Ief537d2c317595d192a37e8493369e08e1fb2f45
2021-08-26 17:55:20 +00:00
];
return WrappedStringList::join( "\n", $tail );
}
* Fix bug with Protected namespace message when protected namespace is NS_MAIN. * Clean up formatPermissionsErrorMessage .. no point building a list if there's only one item, and it casues lots of grossness on complicated custom messages (removing linebreaks, screwing up formatting, etc.).
2007-08-16 07:05:38 +00:00
}
Reference in a new issue Copy permalink