Set the render ID for each parse stored into cache so that we are able
to identify a specific parse when there are dependencies (for example
in an edit based on that parse). This is recorded as a property added
to the ParserOutput, not the parent CacheTime interface. Even though
the render ID is /related/ to the CacheTime interface, CacheTime is
also used directly as a parser cache key, and the UUID should not be
part of the lookup key.
In general we are trying to move the location where these cache
properties are set as early as possible, so we check at each location
to ensure we don't overwrite a previously-set value. Eventually we
can convert most of these checks into assertions that the cache
properties have already been set (T350538). The primary location for
setting cache properties is the ContentRenderer.
Moved setting the revision timestamp into ContentRenderer as well, as
it was set along the same code paths. An extra parameter was added to
ContentRenderer::getParserOutput() to support this.
Added merge code to ParserOutput::mergeInternalMetaDataFrom() which
should ensure that cache time, revision, timestamp, and render id are
all set properly when multiple slots are combined together in MCR.
In order to ensure the render ID is set on all codepaths we needed to
plumb the GlobalIdGenerator service into ContentRenderer, ParserCache,
ParserCacheFactory, and RevisionOutputCache. Eventually (T350538) it
should only be necessary in the ContentRenderer.
Bug: T350538
Bug: T349868
Followup-To: Ic9b7cc0fcf365e772b7d080d76a065e3fd585f80
Change-Id: I72c5e6f86b7f081ab5ce7a56f5365d2f75067a78
This reverts commit 82da9cf14b.
Passing through Remex seems to have unexpected consequences to be
investigated but, for the sake of unbreaking the UBN, let's revert this
first.
Bug: T353920
Change-Id: Iaac7942aa77aee5ab525852ac5b41dd516ff13c9
The previous implementation was using an ad-hoc regular expression which
was matching inside the data-mw attribute of Parsoid output, eg:
<sup about="#mwt42" [...] typeof="mw:Extension/ref mw:Error" data-mw="{"name":"ref","attrs":{"name":"infobox_stats_ref_rail"},"body":{"html":"<style data-mw-deduplicate=\"TemplateStyles:r1133582631\" typeof=\"...">
After substitution, the <link> element inserted contained " instead of
" and so broke out of the attribute.
Instead use a proper HTML tokenizer (via wikimedia/remex-html) so that
we don't allow bogus matches inside attribute values.
To fix up tests:
* Don't deduplicate styles when parsing UX messages (also helps performance)
* Don't deduplicate styles in ContentHandler integration tests
* Don't deduplicate styles by default in parser tests
(unless explicit option is set)
Depends-On: Id9801a9ff540bd818a32bc6fa35c48a9cff12d3a
Depends-On: I5111f1fdb7140948b82113adbc774af286174ab3
Followup-To: Ic0b17e361bf6eb0e71c498abc17f5f67f82318f8
Change-Id: I32d3d1772243c3819e1e1486351d16871b6e21c4
Broadened the argument type to allow passing LinkTarget to:
* ParserOutput::addCategory()
* ParserOutput::addLanguageLink()
* ParserOutput::addLink()
* ParserOutput::addImage()
* ParserOutput::addTemplate()
This allows for a tighter interface with Parsoid's
ContentMetadataCollector class and avoids errors caused by passing the
wrong form of string title ("text" with spaces versus "dbkey" with
underscores).
There are a few performance problems remaining after this patch, which
only apply to use by Parsoid (not the legacy parser):
1. ::addLink() does inefficient db requests to fetch the page id for
each link if the optional $id parameter is not passed. These lookups
should be deferred and a LinkBatch used. (The legacy parser always
passes $id.)
2. ::addTemplate() similarly requires $page_id (and $rev_id) to be
passed, so is not currently usable by Parsoid.
3. ::addLanguageLink() uses Title::getFullText() which is not present
in LinkTarget and is currently implemented as a full Title lookup.
This is not an issue for the legacy parser, because it already has a
Title object so the lookup is a no-op, but could be improved for
Parsoid's use.
Bug: T296023
Change-Id: If21ec8563c8a619bdde7c0cb6534bb9009480a21
Pages that are fast to render can be omitted from the parser cache
to preserve disk space and cache write operations.
The threshold is configurable per namespace, so the tradeoff can
be evaluated based on different access patterns. For example, pages
that are accessed rarely, like file description pages on commons,
may have a high threshold configured, while pages that are read
frequently, like wikipedia articles, may be configured to be always
cached, using a 0 threshold.
Filtering is based on a time profile recorded in the ParserOutput.
A generic mechanism for capturing the timing profile is implemented
in the ContentHandler base class. Subclasses may implement a more
rigorous capture mechanism.
Bug: T346765
Change-Id: I38a6f3ef064f98f3ad6a7c60856b0248a94fe9ac
== Skin::wrapHTML ==
Skin::wrapHTML no longer has to perform any guessing of the
ParserOutput language. Nor does it have to special wiki pages vs
special pages in this regard. Yay, code removal.
== ImagePage ==
On URLs like /wiki/File:Example.jpg, the main output handler is
ImagePage::view. This calls the parent Article::view to handle most of
its output. Article::view obtains the ParserOptions, and then fetches
ParserOutput, and then adds `<div class=mw-parser-output>` and its
metadata to OutputPage.
Before this change, ImagePage::view was creating a wrapper based
on "predicting" what language the ParserOutput will contain. It
couldn't call the new OutputPage::getContentLanguage or some
equivalent as Article::view wouldn't have populated that yet.
This leaky abstraction is fixed by this change as now the `<div>`
from ParserOutput no longer comes with a "please wrap it properly"
contract that Article subclasses couldn't possibly implement correctly
(it coudln't wrap it after the fact because Article::view writes to
OutputPage directly).
RECENT (T310445):
A special case was recently added for file pages about translated SVGs.
For those, we decide which language to use for the "fullMedia" thumb
atop the page. This was recently changed as part of T310445 from a
hardcoded $wgLanguageCode (site content lang) to new problematic
Title::getPageViewLanguage, which tries to guestimate the page
language of the rendered ParserOutput and then gets the preferred
variant for the current user. The motivation for this was to support
language variants but used Title::getPageViewLanguage as a kitchen
sink to achieve that minor side-effect. The only part of this
now-deprecated method that we actually need is
LanguageConverter::getPreferredVariant().
Test plan: Covered by ImagePageTest.
== Skin mainpage-title ==
RECENT (T331095, T298715):
A special case was added to Skin::getTemplateData that powers the
mainpage-title interface message feature. This is empty by default,
but when created via MediaWiki:mainpage-title allows interface admins
to replace the H1 with a custom and localised page heading.
A few months ago, in Ifc9f0a7174, Title::getPageViewLanguage was
applied here to support language variants. Replace with the same
fix as for ImagePage. Revert back to Message::inContentLanguage()
but refactor to inLanguage() via MediaWikiServices::getContentLanguage
so that LanguageConverter::getPreferredVariant can be applied.
== EditPage ==
This was doing similar "predicting" of the ParserOutput language to
create an empty preview placeholder for use by preview.js. Now that
ApiParse (via ParserOutput::getText) returns a usable element without
any secret "you magically know the right class, lang, and dir" contract,
this placeholder is no longer needed.
Test Plan:
* EditPage: Default preview
1. index.php?title=Main_Page&action=edit
2. Show preview
3. Assert <div class="mw-content-ltr mw-parser-output" lang=en dir=ltr>
* EditPage: JS preview
1. Preferences > Editing > Show preview without reload
2. index.php?title=Main_Page&action=edit
3. Show preview
4. Assert <div class="mw-content-ltr mw-parser-output" lang=en dir=ltr>
5. Type something and 'Show preview' again
6. Assert old element gone, new text is shown, and new element
attributes are the same as the above.
== McrUndoAction ==
Same as EditPage basically, but without the JS preview use case.
== DifferenceEngine ==
Test:
1. Open /w/index.php?title=Main_Page&diff=0
(this shows the latest diff, can do manually by viewing
/wiki/Main_Page, click "View history", click "Compare selected revisions")
2. Assert <div class="mw-content-ltr mw-parser-output" lang=en dir=ltr>
3. Open /w/index.php?title=Main_Page&diff=0&action=render
4. Assert <div class="mw-content-ltr mw-parser-output" lang=en dir=ltr>
== Special:ExpandTemplates ==
Test:
1. /wiki/Special:ExpandTemplates
2. Write "Hello".
3. "OK"
4. Assert <div class="mw-content-ltr mw-parser-output" lang=en dir=ltr>
Bug: T341244
Depends-On: Icd9c079f5896ee83d86b9c2699636dc81d25a14c
Depends-On: I4e7484b3b94f1cb6062e7cef9f20626b650bb4b1
Depends-On: I90b88f3b3a3bbeba4f48d118f92f54864997e105
Change-Id: Ib130a055e46764544af0f1a46d2bc2b3a7ee85b7
This matches the behavior of parserTests.txt again (in which
the fallback skin is used by ParserTestRunner::runLegacyTest).
The extra <span> wrappers were added by the Vector skin
(and could be affected by future changes to the Vector skin).
Follow-up to Ief6a6ee03ada8207fc5c60ea438412fa2d529022.
Change-Id: I33729b5026fcfbdbacc0e3fdfef91c9e6b461e6c
The SkinMustache class now accepts a skin option that allows
callers to specify a template that can be used to render
the edit section link.
Additional change:
* Parser tests updated as now edit link label is wrapped
as a span when rendered in Vector 2022 consistent with other
links.
Bug: T346944
Change-Id: Ief6a6ee03ada8207fc5c60ea438412fa2d529022
This also introduces the ephemeral field "$mTransformedText" to store
the result of transformation in ParserOutput.
This is a first step before the transformation uses HtmlHolder as input
and output.
Bug: T348253
Change-Id: I312f3748ebfb0373ee3542ba0abdeefe7db1d488
The ::setTOCHTML() and ::getTOCHTML() method have been deprecated
since 1.40; there's no reason we should be updating ::$mTOCHTML
behind their backs.
Bug: T348134
Change-Id: I9396bc0a2caeb974a06c5b47075b3e2bb9f4278a
It is difficult to distinguish this method from OutputPage::addJsConfigVars()
in code search:
https://codesearch.wmcloud.org/deployed/?q=%5BOo%5Dut%28put%29%3F%28%5C%28%5C%29%29%3F-%3EgetCategories%5C%28&files=&excludeFiles=&repos=
We generally try to replace $output with $parserOutput or $pOutput
as we touch code to improve the ability of codesearch to dig up
deprecated ParserOutput methods.
Bug: T305161
Depends-On: I02dd4f61c43c225b0ef6dc51c3e4f9d967a0a272
Depends-On: I61d2d77591579d825ad9d37f902e40366be55dd6
Depends-On: I91155106b7a9e10d3334f95ba4936d02851bfb11
Depends-On: Iaca745c79d9587571af03b23b21d76a6cba0ebf1
Depends-On: Id10a171c44411b1233ee4d6cf8fbd3dc57744eef
Depends-On: I47a25c011d9bd4b1a15dda4e673e32c25eb64f2b
Depends-On: I683fc768aba50b801f46467fcfa1668fa8731ea6
Change-Id: I5a2ac1c99b8b199102e12f0d32dd6ec5cdc24054
ParserOutput::addOutputHook() has been deprecated since 1.38, and without
any calls to ::addOutputHook() the associated ::getOutputHooks() and
$wgParserOutputHooks configuration do nothing.
Bug: T292321
Bug: T305161
Change-Id: Ib770c680d5e0697980e7e36a323ec56ba1d806b8
Tweaked the pluralization of the newly-added
ParserOutput::appendOutputString() method (now ::appendOutputStrings()
and ::getOutputStrings()), and name of the ParserOutputStrings class
(now ParserOutputStringSets), in an effort to continue repainting
bikesheds until the color is juuuust right.
Also extended the new method to cover ::addModules() and ::addModuleStyles()
and added support for these string sets in ::collectMetadata().
(These methods and the enumeration class were originally added in
b2cfa31eb6173e9f5e8607eadd126c33f8ce440b.)
Depends-On: I8bdffa55498d90e990af5bfc3332e3028b0a3539
Change-Id: Ibd41485d5db7779f01642e2144c50ed49d409812
This aims at providing an interface similar to setOutputFlag for string
sets, such as the ones used in CSP properties.
Change-Id: I6f103bd88802e66611e483403a2f8a540d54aae9
Syntactical leftover with no significance in modern web.
Bug: T309150
Depends-On: I3a029ca950db42b938962b2452ad136ae8ddea6f
Depends-On: Id0557ac19583de36d7226b14a4c06933da47fe97
Depends-On: I17580a72e4a9384d7d774866e610197e950900cb
Change-Id: I4bbfa47fbf6e30fb90d920d6d02cdf6e0b1cdb46
Just methods where adding "static" to the declaration was enough, I
didn't do anything with providers that used $this.
Initially by search and replace. There were many mistakes which I
found mostly by running the PHPStorm inspection which searches for
$this usage in a static method. Later I used the PHPStorm "make static"
action which avoids the more obvious mistakes.
Bug: T332865
Change-Id: I47ed6692945607dfa5c139d42edbd934fa4f3a36
Provide a way for backend code to determine the primary language of a
ParserOutput, eg for setting the Content-Language header of an API
response.
This is read-only and backed by extension data at the moment for
transition purposes; if this API sticks we'll graduate it to a
"real" property in the future, with appropriate serialization
to/from JSON (T303329).
Similarly, this patch only includes the most basic code to handle
the various ParserOutput merge cases in
ParserOutput::merge{Internal,Html,Tracking}MetaDataFrom(),
ParserOutput::collectMetadata(), and
OutputPage::addParserOutput{Content,Metadata,Text,}(); mostly
inherited from the fact that the storage is backed by extension
data at the moment.
Generally only the "top-level" parser output gets to set the
primary language; we'll presumably need to ensure that the
language is consistent during merge.
Change-Id: I767daba22805a877d9b806fd77334e508902844b
Before 1.39 we used <mw:toc> and in 1.39 we switched to <mw:tocplace/>
(commit 24949480eb). This was changed
to a <meta> tag in 1.40 (commit
0b10563895 and
fa8646ca7b) and the old content has long
since expired from the ParserCache. Clean up the old ParserCache
transition code.
Change-Id: I3254d0acba31e107b50767797a2b0ad28aba59ee
* Rather than computing TOC HTML in Parser and setting it in
ParserOutput, compute it on demand based on section metadata.
This will let Parsoid set section metadata in ParserOutput
and have the TOC generated automatically.
* This required fixing some "bugs" in Linker's generateTOC
which didn't properly close tags and relied on Tidy to fix
up unclosed li and ul tags.
* This patch relies on converting section metadata objects to
array objects, but Linker::generateTOC could be converted to
use TOC data instead.
* Since TOC generation is now moved to getText(), this is done
post-PC load and this eliminates the parser cache split on
user language for TOC heading localization.
Bug: T293513
Change-Id: Ief1bba326d3612b40930440c872a61abadffab10
* 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
Split out from the I44045b3b9e78e change.
This is consistent with what Parsoid will use for the TOC marker.
Bug: T287767
Bug: T270199
Bug: T311502
Depends-On: I1f607cf1ef1b61fb4d2e1880de756fb94d5a6b22
Change-Id: Ie63eed07b9bca1bfa07d4c256aba3728cedd8f93
Split out from the I44045b3b9e78e and Ie63eed07b9bca changes. We
first add code to handle the new tag as well as the old tag in
ParserCache contents. This will allow us to safely rollback if needed
when deploying the follow-on patch which actually changes the tag
used.
Bug: T287767
Bug: T270199
Bug: T311502
Change-Id: Ib3e5e010b9f5ca2c4ea7c4fe28080170b6a88812
Previously:
* It was unclear that generate-html is an optional optimization
* Most of MediaWiki core was doing $parserOutput->setText('') if
html wasn't generated. However this is wrong and will cause
$parserOutput->hasText() to return true and also potentially cause
cache pollution if a content handler both does that and supports
parser cache (Like MassMessage; see T299896)
* The default value of mText in the constructor was '', and most
of the time MW used that default. This doesn't seem right. If
setText() is never called, the ParserOutput should not be considered
to have text
* It was impossible to set mText to null, as $parserOutput->setText(null)
was a no-op. Docs implied you were supposed to do this, so it was very
confusing.
This patch clarifies docs, changes the default value for ParserOutput::$mText
from '' to null, and makes $parserOutput->setText(null) do what you
expect it to. The last two are arguably breaking changes, although
the previous behaviours were unexpected, mostly undocumented and
based on a code search do not appear to be relied on.
It seems like the main reason this only broke MassMessage is most
content handlers either don't support generateHtml, or they don't
support parser cache.
Bug: T306591
Change-Id: I49cdf21411c6b02ac9a221a13393bebe17c7871e
Depends-On: I68ad491735b2df13951399312a4f9c37b63a08fa
The old ParserOutput::getProperty() method returned `false` when a property
was missing. This requires callers to use the `?:` syntax to supply default
values, which then causes any falsey value to be treated as missing.
So, for example, setting the defaultsort to '0' will cause the default
sort to be ignored.
Modern php convention is to use `null` for missing values, and the `??`
syntax is a better/more restrictive alternative to `?:`.
We renamed `ParserOutput::getProperty()` to `::getPageProperty()` in
1.38 (Ie963eea5aa0f0e984ced7c4dfa0fd65d57313cfa/T287216) but kept the
return value convention. Before this actually makes it into a 1.38
release, take the opportunity to fix the return value for the new
`ParserOutput::getPageProperty()` method to return `null` when the
property is missing.
We need to do some temporary workarounds to the places we'd
already swapped over to use the new `::getPageProperty()` method
to allow them to handle either `false` or `null` as a return value;
we'll clean that up once this is merged.
Code search:
https://codesearch.wmcloud.org/deployed/?q=-%3EgetPageProperty%5C%28|T301915&i=nope&files=&excludeFiles=&repos=
Bug: T301915
Depends-On: I3f11ce604970e47b41fc1c123792df8c3045626f
Depends-On: Ie7533f49fe4cad01ebfda29760d23c61e9867b10
Depends-On: Ic5c09f5caa4c897bc553c614fbae9cee159566a2
Depends-On: I0278b2eafd90e77e4fee41c45a1165fb79ddf47e
Depends-On: I383abb6b7dc5e96c0061af13957609f6e31a1065
Depends-On: I79f9f4078e415284af29b15047bafd1c823d7f5b
Depends-On: I02276c48c49f5d2d241a69eb0a6cdf439b572d8b
Depends-On: I71628661b4539a4e35ae32846e719f92bcf782e0
Depends-On: I7e215cb43de0ce150a6bcc00f92481dcdcfed383
Change-Id: Iaa25c390118d2db2b6578cdd558f2defd5351d15
There is a common and reasonable need for longer lines in tests.
The nudge for shorter lines doesn't seem valuable here. The natural
breaks will likely still fall in 80-100 given the enforced practice
for non-test code, e.g. whether through habit, or 80-100 column markers
in text editors, or the finite width of diff and code review
interfaces.
Change-Id: I879479e13551789a67624ce66f0946d2f185e6ee
Soft-deprecate the use of ::setExtensionData() to destructively update
the value stored under a single key. Add the new
::appendExtensionData() method to use where multiple values are
desired. This accomodates the asynchronous and incremental parsing
goals on the Parsoid roadmap.
Bug: T300981
Change-Id: I2dea4ba71ea506428854a9983c1abd906b2efd5f
Deprecate ParserOutput::addJsConfigVars() and add setter methods which
better ensure that the ParserOutput contents are independent of parse
order. This accomodates the asynchronous and incremental parsing goals
on the Parsoid roadmap.
Bug: T300307
Change-Id: I4f08d1098da211f7bf5c43c08c620de224cbf37f
* 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
Encourage localization and factor out common code by taking a message
key as the first argument to ::addWarningMsg() instead of a wikitext
string. This also plays nicer with Parsoid by separating out the
localization code from the parse.
Bug: T293515
Change-Id: I6a7c04c67ac586ab00d4edcbb3d09485a7794e23
