Per feeedback on T152540. Now there's an uniform way to get to
section's <h*> from element with the given ID.
Change-Id: I94b46707dadb1988d963344b02060dcaffdab7b4
It adds the ability to replace the current section ID escaping
schema (.C0.DE) with a HTML5-compliant escaping schema that is
displayed as Unicode in many modern browsers.
See the linked bug for discussion of various options that were
considered before the implementation. A few remarks:
* Because Sanitizer::escapeId() is used in a bunch of places without
escaping, I'm deprecating it without altering its behavior.
* The bug described in comments for Parser::guessLegacySectionNameFromWikiText()
is still there in some Edge versions that display mojibake.
Bug: T152540
Change-Id: Id304010a0342efbb7ef2d56c5b8b244f2e4fb2c5
And auto-fix all errors.
The `<exclude-pattern>` stanzas are now included in the default ruleset
and don't need to be repeated.
Change-Id: I928af549dc88ac2c6cb82058f64c7c7f3111598a
The title, although it refers to a foreign page, is considered local
due to Title::isExternal() being false, hence it's namespace prefix
is localized.
We need to use the canonical namespace name instead.
Bug: T169221
Change-Id: I5e5f7d873b9497ca6c8853e7d8170a4f4f07c051
In 1bf5a652 the id selector was changed to a class selector for toctitle.
The cached HTML has been expired now and the id selector is not necessary
anymore.
Also remove the id selector #toc.tochidden for print style. This is not
necessary because the tochidden gets only added to .toc and not to #toc.
Change-Id: I43cfffdb0807e8ed8f6b7b8732ba857b709bee80
Self-links are still semantically links, and representing them as <strong>s
is inelegant and more important a real pain to work with, especially in
contexts where they may change state (like inside an editor).
Instead, render them as <a>, with no href to avoid user agent style over-
rides and with a class to style them as before, named 'mw-selflink' to go
with 'mw-redirect'. This allows much easier adjustment later. The old CSS
class 'selflink' is retained for backwards compatibility, but deprecated.
Bug: T160480
Change-Id: If058843924c3b30c116df2520aef93a004d98a5d
On Special:Watchlist, Special:Contributions, Special:Recentchanges etc.
there are links to (talk | contribs | block) for the user who did the
contribution. Add CSS class for them. Introduce the following css
classes:
- mw-usertoollinks-contribs
- mw-usertoollinks-talk
- mw-usertoollinks-block
- mw-usertoollinks-mail
Bug: T156879
Change-Id: I85a3b0987a016ff25026f1c047214a31170b0452
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
This is a static function that uses global state through $wgLang, and
provides no benefit than calling Language::formatSize() directly aside
from escaping. But that should generally be handled closer to output
generation.
Change-Id: I35fdbaf2fcaa89f0d4442e1f63ec8ed29e0c339f
Linker::formatTemplates() was a static function that depended upon
global state like $wgLang (explicitly), $wgUser & $wgTitle (implicitly).
Moving it to a separate class allows us to clean it up a little bit and
use modern things like RequestContext and LinkRenderer.
Bug: T145177
Change-Id: Icdea8a2b299b4876feb3df3d66df3e4c104dd928
This is more consistent with LoadBalancer, modern, and inclusive
of master/master mysql, NDB cluster, and MariaDB galera cluster.
The old constant is an alias now.
Change-Id: I0b37299ecb439cc446ffbe8c341365d1eef45849
Usernames can be in any script/language and include weak and neutral
characters like parentheses and dashes - these flip and make things
look very weird when displayed in the opposite directionality.
This fix adds <bdi> tags to isolate the username when we output user
links. This will affect displays like user personal tools, history
and diff view.
Also, fixing LogFormatterTestCase::removeSomeHtml() that expected
a very rigid html input (and thus failed with the new <bdi> wrapping)
to use strip_tags() instead.
Change-Id: I2db5f4b7d3a00726461eb6b699fbdf0ecd47a1cb
Now that SpecialPage::getTitleValueFor() exists, use it so that we can
avoid returning a Title object uncessarily.
Change-Id: I5bf605baf2ecad62e189421a2e5c556cc0b6f6ac
...and add a link to the on-wiki LinkRenderer documentation to the
class. Actual wfDeprecated and replacements in all the uses to come in
follow-up patches.
Change-Id: I2fefb6c38153028d2cc68c13bc62434d3df69cd2
* Rename to getLinkClasses() since it's not really returning colours,
but CSS classes.
* Dependency inject LinkCache into LinkRenderer
* Update all callers of Linker::getLinkColour(), and mark it as
deprecated (no other uses in Gerrit)
* Update a bunch of tests for new dependency
Change-Id: Id178e2dcc60b833ce2dbad4920896b93cabba1bf
This partially reverts commit 9af38c046c.
* The new JS modules continue to exist. While not used by default,
the mediawiki.api.rollback may be used elsewhere by now.
The mediawiki.page.rollback module may still be used on cached
pages and should continue to work.
* The new API behaviour remains.
* To avoid breaking mid-air rollbacks again, the token remains
the same (plain "rollback" salt). We're not re-introducing the
web-specific salt again.
Bug: T136375
Change-Id: Iba17ce55ff9506e838bfc6e70ca280e5a20b77b6
'noclasses' makes more sense as a per-link option rather than an
instance member of the LinkRenderer instance, since it depends entirely
on whether the calling code has preloaded the link classes.
Introduce LinkRenderer::makePreloadedLink() which makes this clear and
requires passing in the classes as a separate parameter. As a side-
effect, due to the way LinkRenderer::mergeAttribs() is implemented, the
'class' attribute will always appear before the 'title' attribute in the
final output.
Change-Id: I0545aa9d7139794bc22f9d3d6d6eccde003b2982
When MW_NO_SESSION is set, getting the user options for $wgUser will
result in an error. Check for that in the service definition, and
reuse that check in Linker.
Also removed an unused use clause.
Bug: T136124
Change-Id: Id6448d12e8f4a8a94852980e5eb6ccb202066822
Similar to WatchAction (converted in commit 77cdf1919).
* Make FormAction::getFormFields not abstract.
In most cases this will just be an empty array.
* Convert RollbackAction from FormlessAction to FormAction and implement the
required error handling scenarios (mostly moved out of from the old method, or
duplicated from the WikiPage method where necessary).
* In most cases the in-between form is never used since a JavaScript handler
takes over the link and uses the API over AJAX instead. In the no-js fallback
(as well as for any existing tokenless rollback links) copy the GET parameters
into the form for re-submission as POST (plus token, added by HTMLForm).
* Remove the distinction between WebUI and API tokens. This stronger token salt made it
unnecessarily complex and was only there because it used GET until now. This streamlining of
tokens matches what we already do for 'watch', 'edit', 'patrol' and other actions.
* Fix form submission bugs when 'from' query parameter is missing.
- Ensure the required 'from' query parameter is present before showing a form.
No need for the user to submit a form we know will fail.
- Plain GET request to action=rollback (with no parameters) is now a 400 Bad Request
instead of a form that would fail when submitted.
- Submitting the form without 'form' field now correctly says why it failed.
Previously it emitted a session error, which was a lie.
Bug: T88044
Change-Id: Ia457802fec2e90573c8e7d552bc1f3cee258f10b
This is a rewrite of Linker::link() to a non-static, LinkTarget-based
interface. Users of plain Linker::link() with no options can use the
LinkRenderer instance provided by MediaWikiServices. Others that
have specific options should create and configure their own instance,
which can be used to create as many links as necessary.
The main entrypoints for making links are:
* ->makeLink( $target, $text, $attribs, $query );
* ->makeKnownLink( $target, $text, $attribs, $query );
* ->makeBrokenLink( $target, $text, $attribs, $query );
The order of the parameters are the same as Linker::link(), except
$options are now part of the LinkRenderer instance, and
known/broken status requires calling the function explicitly.
Additionally, instead of passing in raw $html for the link text, the
$text parameter will automatically be escaped unless it is specially
marked as safe HTML using the MediaWiki\Linker\HtmlArmor class.
The LinkBegin and LinkEnd hooks are now deprecated, but still function
for backwards-compatability. Clients should migrate to the nearly-
equivalent LinkRendererBegin and LinkRendererEnd hooks.
The main differences between the hooks are:
* Passing HtmlPageLinkRenderer object instead of deprecated DummyLinker
* Using LinkTarget instead of Title
* Begin hook can no longer change known/broken status of link. Use the
TitleIsAlwaysKnown hook for that.
* $options are no longer passed, they can be read (but shouldn't be
modified!) from the LinkRenderer object.
Bug: T469
Change-Id: I057cc86ae6404a080aa3c8e0e956ecbb10a897d5
noreferrer is used as support for noopener is very limited.
This is to prevent the attack detailed at
https://mathiasbynens.github.io/rel-noopener/ where you can
navigate the parent window, even if the new window is a cross-origin.
Bug: T133507
Change-Id: I6e4ab938861e246ff44048077b94847e303f1859
Signed-off-by: Chad Horohoe <chadh@wikimedia.org>
The intention for Linker::linkKnown() was to be used when the caller had
already preloaded the target's existence ('known') and called
Linker::getLinkColour() directly ('noclasses'). However, nearly all
usage of linkKnown() only did the first part, and not the latter.
So do what people actually ended up using the function for, and remove
'noclasses' from the default parameters. As long as the target the link
is being created for is already in LinkCache, this shouldn't cause any
extra database queries.
Change-Id: Ia5a4c2f18ec780627146617a1498bd04fcfbb3ee
This is a private function so it doesn't really make much of a
difference, but will allow further refactoring.
Change-Id: Idacad7b1bd4776c04659ed9af64540d6ce4b2f28
This is needed because the call to substr() returns false when the string
is only 1 character long, which occurs when parsing degenerate links like
[[:]] and [[::|foo]].
The seemingly unnecessary test for $match[1] !== '' is for forwards
compatibility with PHP 7.
Bug: T121706
Change-Id: Icc19ee990d01958d64b938d298e9a7e1df7181b5
Linker::linkAttribs() needs the stub threshold to determine the classes
in some cases. For the code path from Parser → LinkHolderArray → Linker,
we can pass down the correct threshold from ParserOptions instead of
pulling a potentially-different threshold from $wgUser.
Bug: T124367
Change-Id: I16b9d6c3044ae60d5a7fd340569c019ffc4b2a55
With I779aa1c017abc9a17047fd5335f2d082148daa6f all public functions from
WikiPage were copied as delegator into Article.
Now remove the unused functions and add a release notes
Change-Id: I9cc79acedecb003048e1add568e8bba96ddbb41a
This method was removed in:
cfaf26e501
so lets remove the reference to it so people don't
go looking for what is not there.
Change-Id: I229f033a3b4553e6eacc460faaf156ba2539d6d5
Redlink tooltips end up in parser cache!
This is not ideal, as ideally it would be in the page's language,
not the content language and it would be user language for things
outside of pages, but that's not easy to do the way the code is
structured, and this is much better than the alternative.
Other possibilities include:
* Splitting parser cache (Previously been reverted, and doesn't
seem worth it given the feature in question)
* Post-processing step after parsing (Really complicated. Doesn't
seem worth it)
Bug: T34686
Change-Id: Id632f8ef59d1c762aed1867a708c569cbff5f0dd
You can now pass parameters to toolbox links' tooltips,
by setting 'tooltip-params' key in a $nav_urls item of
SkinTemplate::buildNavUrls.
This functionality is used to add GENDER support to two
tooltip messages.
Bug: T123365
Change-Id: I45577d54bd78ba1e2ba10fb8f04367244495bfea
* Linker: Follows-up 9ce5ca9886. null is already handled.
* PrefixSearch: Follows-up 337b1e2e. null is already handled.
If other types are bassed, we want to find these.
* UploadBase: Follows-up 11f5ebec28. null is already handled.
WebRequest::getVal($name) can only return string or null.
Change-Id: I3dc3b700cde735f5cd9d497f0867a2b4e3b61f46
Remove empty line comments as found by the
MediaWiki.WhiteSpace.SpaceBeforeSingleLineComment.EmptyComment sniff
Change-Id: I5d694f7a7d3bc97e16300ba03c60ad17f3c912a5
strtr() is marginally faster as it runs through the string only
once. A better fit for one-for-one character translation.
The strtr() function also supports an associative array as second
parameter for entire string replacements. This, too, has the same
performance and predictable behaviour (starts with the longest key).
Whereas str_replace is for more aggressive needs where you want
multiple passes until there are no further matches.
The associative array form is arguably also easier to understand
and harder to mess up since the needle/replacement pairs are
explicitly connected instead of two separate arrays.
Also:
* Use getFormattedNsText instead of strtr( getNsText, .. ) which
reduces duplication of this fact through a more semantic intent.
Change-Id: Ie23e4210a5b6908dd79eebc8a2b931d12fe31af6
Follows-up 6c7480e5f0, d2a6a73d2.
The should stop the § character from being indexed by Google et al.
Bug: T18691
Bug: T93000
Change-Id: Ie9e334e973e3ded270f1897a2c3816d9df739fc0
Changed the Linker::makeHeadline function to additionally generate an
anchor that uses a section symbol. Created mediawiki.sectionAnchor
module to style the anchors and included it in SkinTemplate so that
other skins automatically get those styles.
Bug: T18691
Change-Id: I562e437ec0bd337c9db0406f5dcab504dfa37034
The word-separator is already part of the return value of
Linker::userToolLinks which results in a double space when add an own
word-separator
Change-Id: Id9d3125b9cae4a92f489215dda0b32c487f3fb11
Revision->getRawUser()
=> Revision->getUser( Revision::RAW )
Revision->getRawUserText()
=> Revision->getUserText( Revision::RAW )
Revision->getRawComment()
=> Revision->getComment( Revision::RAW )
The body of Revision->getRawUserText() has been moved
into Revision->getUserText().
Every usage has been replaced.
Change-Id: Ic6fbfbc0507dcf88072fcb2a2e2364ae1436dce7
Affected both the linked text and the actual URL.
For an example please see the ukwiki link on:
https://www.mediawiki.org/wiki/Special:CentralAuth/Admin
Follow-Up: Ibdedf087f85046646450367cbf1811db578d8f4b
Change-Id: Ie73796dcde4e0a096a6baa86a53245e6fe2b48b5
Xhprof generates this data now. Custom profiling of various
sub-function units are kept.
Calls to profiler represented about 3% of page execution
time on Special:BlankPage (1.5% in/out); after this change
it's down to about 0.98% of page execution time.
Change-Id: Id9a1dc9d8f80bbd52e42226b724a1e1213d07af7
Before r39373, all autocomments in an edit summary were formatted. In
fixing a bug with page titles containing "/*" this was accidentally
broken.
To use a single preg_replace_callback call to replace multiple
autocomments, we need to make sure that the match of one autocomment
doesn't overlap the match of another, which means we can't have "(.*)"
before and after. But we do still need to detect whether there is
anything before or after. "(?=(.?))" and "(?<=(.?))" would do nicely,
except the latter isn't actually supported. "(?=(.))?" and "(?<=(.))?"
work too, but older versions of PCRE don't support that. They do,
however, support "(?:(?=(.)))?" and "(?:(?<=(.)))?", so that's what
we'll go with.
This change does change the values for $pre and $post passed to the
FormatAutocomments hook; extensions need to be updated to accept (and
not prepend/append) booleans for these parameters.
Bug: T18530
Bug: T70361
Change-Id: I36c3a9e548a4ef72f93974bb35f9add8c29e9287
The link text for [[/Foo/]] is `Foo` and the link text for
[[../Foo/]] is `Foo`. So far so good.
But the link text for [[/Foo//]] is `Foo` while the link text
for [[../Foo//]] is `Foo/`. We are stripping all trailing slashes
in the first case, but not the second.
Fix the code so that we strip all trailing slashes in both cases.
Update some of the comments in the code while we are at it.
Change-Id: Id61eacafea9820c404699a7902c8eb8102779516
