We have several types of change lists (old RC/watchlist/related
changes, enhanced RC/watchlist, history) with slightly different
HTML, each with their own idiosyncracies. JavaScript code trying
to identify lines by log ID / revision ID has to jump through all
kinds of hoops to work with that.
To simplify the lives of frontend / gadget maintainers and provide
something approaching an API for these pages, we now expose the basic
attributes of each change line (revision ID for edits, log type/action
and ID for log events) as data attributes.
The OldChangesListRecentChangesLine, EnhancedChangesListModifyLineData,
EnhancedChangesListModifyBlockLineData, PageHistoryLine,
ContributionsLineEnding and DeletedContributionsLineEnding hooks
are updated accordingly. New hooks (LogEventsListLineEnding and
NewPagesLineEnding) are added for the change list pages which did
not yet have them.
Change-Id: I6dd006d0b1b0fd35c0020f0f9eea9113eca30b35
While most of MediaWiki ignores the shell's/C library's locale setting,
there are some things other than shell execs that use it (e.g. the
luasandbox PHP extension).
To provide a consistent environment, set the locale in Setup.php instead
of letting it be changed mid-request depending on whether something else
happened to have called certain functions.
Bug: T107128
Change-Id: I02943803d26d5b1b3ac00ef9216f69cdfa149585
The following changes are added:
- Created MediawikiStatsdDataFactory interface
- Added hasData() method to see if there are any data to send
- Added getData() method to fetch data
- Made service infrastructure use MediawikiStatsdDataFactory interface
- Made wfLogProfilingData() use MediawikiStatsdDataFactory interface
- Added capability to enable/disable buffering collector
Bug: T166354
Change-Id: I2874175647e987996a9a399829b3319674471aaa
* Move @deprecated tags out of the header so the documentation comment
still has a summary
* makeKey() and co. are part of BagOStuff, not ObjectCache
Follows up 1e5961dbf6.
Change-Id: I168ce97597f288f6345a404354d68371e57522d3
For a while now these functions have called makeKey/makeGlobalKey() on
the default ObjectCache instances, and callers should now use those
functions directly instead.
There are other advantages to doing so, like in cases where the
ObjectCache instance being used has a different makeKey implementation
(e.g. memcache) than the default configured cache.
Change-Id: I9abf6badbf3ba800e591d2503b714433d5faf1e3
Introduce a service to represent wfReadOnly() and friends.
It's necessary to have two service instances, one for wfReadOnly() and
one for wfConfiguredReadOnlyReason(), to avoid a circular dependency,
since LoadBalancer needs the configured reason during construction, but
wfReadOnly() needs to query the currently active load balancer.
Not having a cache of the configuration makes it possible to dynamically
change the configuration. Ideally things would not change the
configuration, and I removed such instances in core, but to support
extensions, I added a test ensuring that the configuration can be changed.
Change-Id: I9bbee946c10742526d3423208efd68cb3cc5a7ee
I was bored. What? Don't look at me that way.
I mostly targetted mixed tabs and spaces, but others were not spared.
Note that some of the whitespace changes are inside HTML output,
extended regexps or SQL snippets.
Change-Id: Ie206cc946459f6befcfc2d520e35ad3ea3c0f1e0
Install the backtrace collector very early, so that we can get the
backtrace even if headers were sent from LocalSettings.php.
Bug: T157392
Change-Id: I9bc732b34481c95afb5362e135a87bd4302498e2
* Verified none of these constants depend on each other.
* Move them to the top of Defines.php for consistency.
Change-Id: I19c77e706d69bff513e4abafd0300207de0dbb1d
The old code was similar to Message::params(), but Message::params()
was unable to handle "special" parameters and received an overhaul
in 7f2663f, yet wfMessage remained broken. To avoid duplication,
wfMessage shall call Message::params() to correctly handle these
parameters.
CategoryMembershipChange::getChangeMessageText and its caller has
been updated so as not to take advantage of this bug.
Bug: T153747
Change-Id: I6667acf7e71c9db07fefc9fbb741c160e15823ff
If iiprop=badfile is set, expose the wfIsBadImage() status as a boolean.
This is mainly intended for use by Parsoid, but could be useful for
other API users too. The context title can be set by using the
badfilecontextitle parameter.
Also be a little more paranoid and make sure wfIsBadImage() always
returns a boolean.
Bug: T148343
Change-Id: I3432ead92b8d2a17de5243285f2760fc39b501ae
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
Use HTTPS instead of HTTP where the HTTP link is a redirect to the HTTPS link.
Also update some defect links.
Change-Id: Ic3a5eac910d098ed5c2a21e9f47c9b6ee06b2643
Parser tests depend upon being able to change the value of
$wgEnableMagicLinks at runtime, so skip caching the replacement patterns
while parser tests or phpunit tests are being run.
Change-Id: Ic9fe4fe898aa353bdefa2bf8207b77d2ac14887f
The ScopedCallback class was moved into a separate library. This updates
all callers to use the namespaced version, and provides a
backwards-compatibility class wrapper under the old name.
Bug: T146258
Change-Id: I2dd0a66fe2f510f26bdfef6b0a975c1beb3fd93c
Depends-On: Iea0c40bdd7776372ccf72db8f088a2abaa4d3721
Apparently it is possible for Parser::mParserOptions
to not be set in some cases. I'll try again later.
This reverts commit bda74bff6e.
Bug: T146433
Change-Id: Idb6d1b20995d5f86b712abb386ab987356c4f560
* This method is analogous to wfTimestamp(). Optimise for the common
idiom of just converting a timestamp without having the caller hold
on to any object.
* Make wfTimestamp() use this (it could already since it didn't
use any MWTimestamp methods). Use via MWTimestamp. While this is
the same as direct access, it allows future changes.
* Add tests covering this new method.
Change-Id: I7f9104f1701d92fe25d72c7943581c64e1d093fa
* Remove wf* function dependencies. This includes wfTempDir().
Callers now should specify the directory, though it will try to do
most of the wfTempDir() logic anyway if they do not.
* Update callers to inject wfTempDir() so $wgTmpDirectory is used by
TempFSFile instead of it probing to find a valid directory itself.
* Move most of the wfTempDir() logic to TempFSFile::getUsableTempDirectory().
* Remove unused getMimeType() method.
Change-Id: Idd55936b07f9448a6c90577708722b7b52b8fe66
wfEscapeWikiText() used $wgEnableMagicLinks, but that could result in an
inconsistency when something modifies the magic link related
ParserOptions.
In general, most uses of wfEscapeWikiText() are in parser functions or
when message parsing, so the Parser is a logical place for it.
A future patch will make it easy to use Parser::escapeWikitext() in
message parameters.
Change-Id: I0fd4d5c135541971b1384a20328f1302b03d715f
The magic link functionality is "old backwards-compatibility baggage"
that we probably want to get rid of eventually. The first step to doing
so would be making it configurable and allowing it to be turned off on
wikis that don't use it.
This adds each of the 3 magic link types as individual parser options,
which can be controlled by the $wgEnableMagicLinks setting.
Additionally, wfEscapeWikiText() was updated to only escape enabled
magic link types.
Bug: T47942
Change-Id: If63965f31d17da4b864510146e0018da1cae188c
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
