Bartosz Dziewoński [Sat, 11 Jul 2015 15:10:12 +0000 (17:10 +0200)]
SpecialLinkSearch: Use 'type' rather than 'class' for HTMLForm's HTMLSelectNamespace
That seems to be the usual convention.
Change-Id: I3afe7510be73178f82c9507ec68964f6e81567d4
umherirrender [Fri, 10 Jul 2015 21:03:52 +0000 (23:03 +0200)]
Pass function name in DatabaseMysqlBase::setBigSelects
This allow to see the correct function name in the debug logs
Change-Id: I913acbc39e8e7b21f647b544ef161574c5f720f8
jenkins-bot [Fri, 10 Jul 2015 20:55:14 +0000 (20:55 +0000)]
Merge "Pass function name in Database::selectRowCount"
umherirrender [Fri, 10 Jul 2015 20:44:03 +0000 (22:44 +0200)]
Pass function name in Database::selectRowCount
This allow to see the correct function name in the debug logs
Change-Id: I84c94edce60efb8f15a7eb178065ee03e85fc282
Kunal Mehta [Thu, 9 Jul 2015 22:56:17 +0000 (15:56 -0700)]
SECURITY: Do not allow data-ooui attributes in wikitext
We now automatically infuse any element with a data-ooui attribute, so
allowing them in wikitext allows rendering any arbitrary OOUI widget,
some of which (ButtonWidget) are unsafe and can lead to XSS.
By blacklisting data-ooui, widgets cannot be created in wikitext.
T101666 will enable a safe-subset of them.
Bug: T105413
Change-Id: I3f63594a41e9cac3219791e181a2f93818178263
Chad Horohoe [Fri, 10 Jul 2015 18:12:15 +0000 (11:12 -0700)]
Protect against non-text output from StripState going into Title::newFromText()
Non-string input shouldn't be fed into newFromText(). We currently handle this
indirectly with relying on Title to do it. Instead just return earlier and not
try to construct a title from bad input.
Bug: T102321
Change-Id: I9bc96111378d9d4ed5981bffc6f150cbd0c1e331
Translation updater bot [Fri, 10 Jul 2015 18:59:32 +0000 (20:59 +0200)]
Localisation updates from https://translatewiki.net.
Change-Id: I2c337f29ca8c60e6b1c912d6ee828b1632ba89b8
Florian [Fri, 10 Jul 2015 14:31:33 +0000 (16:31 +0200)]
Don't try to call load() on non-object
FileRepo::newFile() can return null or false, in addition to a File
object. Calling load() on the returned value in FileRepo::findFile()
can result in an exception. Prevent this with an additional check after
FileRepo::newFile() was called.
Follow up: I1aa4b096c0cad5f5ca34321cc897019005c53a76
Follow up: I60c106b5b27db067b1884af95f5fb74a0a682a9a
Bug: T105497
Change-Id: Iabdae8fd960d19a467dc08a193228978575017ce
Brad Jorsch [Fri, 10 Jul 2015 17:38:19 +0000 (13:38 -0400)]
Fix sql.php behavior on error
When an erroneous query is entered into sql.php, it doesn't clear its
line buffer, so the intended next query (often fixing a typo in the
original) winds up concatenated onto the erroneous query. Usually
leading to another error, which repeats the process.
The solution is simple enough: clear the line buffer unconditionally,
not only when no exception is caught.
Change-Id: Ia78e2df8b9c6698c0a225bfb3135274ed8da5303
jenkins-bot [Fri, 10 Jul 2015 17:50:30 +0000 (17:50 +0000)]
Merge "ApiFeedWatchlist: Handle revdel"
jenkins-bot [Fri, 10 Jul 2015 16:54:02 +0000 (16:54 +0000)]
Merge "Fixed BadMethodCallException in FileRepo for bad titles"
Brad Jorsch [Fri, 10 Jul 2015 14:04:10 +0000 (10:04 -0400)]
ApiFeedWatchlist: Handle revdel
We might get back rows with no title (log entries with deleted action)
or no user (anything with deleted user). Don't raise warnings for those.
Bug: T105367
Change-Id: I64b4fba478b3e1e86854843965b2ea2b5c1d7933
Kunal Mehta [Fri, 10 Jul 2015 03:33:53 +0000 (20:33 -0700)]
Gruntfile.js: Remove unused 'pkg'
It's unused and ends up being copied into a bunch of repositories that
also don't need it.
Change-Id: Icf81b3ae13e510670bde38c1c8d7ef42917b7c2a
Antoine Musso [Fri, 10 Jul 2015 10:41:47 +0000 (12:41 +0200)]
rubocop: ignore node_modules and fix a single offense
Make rubocop pass:
- fix a trivial double quotes error
- ignore 'node_modules' useful when one ran 'npm install'
Change-Id: Id1ab83df0bec4fc7b223f33016c8f18316509b7a
jenkins-bot [Fri, 10 Jul 2015 07:11:00 +0000 (07:11 +0000)]
Merge "Ignore tagfilter url parameter on Special:Undelete"
jenkins-bot [Fri, 10 Jul 2015 07:10:56 +0000 (07:10 +0000)]
Merge "Sort $wgLogActionsHandlers/$wgLogActions/$wgLogHeaders"
jenkins-bot [Thu, 9 Jul 2015 23:41:35 +0000 (23:41 +0000)]
Merge "Check install user's password as sysop/bureaucrat"
Aaron Schulz [Thu, 9 Jul 2015 22:59:48 +0000 (15:59 -0700)]
Fixed BadMethodCallException in FileRepo for bad titles
* Fixes regression from
7508b86ad3
Change-Id: I60c106b5b27db067b1884af95f5fb74a0a682a9a
jenkins-bot [Thu, 9 Jul 2015 21:52:32 +0000 (21:52 +0000)]
Merge "Made LoadMonitor use wfGlobalCacheKey() and tweaked the key name"
jenkins-bot [Thu, 9 Jul 2015 21:49:26 +0000 (21:49 +0000)]
Merge "Formally defined IDatabase and made a few Database methods protected"
jenkins-bot [Thu, 9 Jul 2015 21:47:47 +0000 (21:47 +0000)]
Merge "importScopedSession() comment updates"
Aaron Schulz [Thu, 9 Jul 2015 21:32:04 +0000 (14:32 -0700)]
importScopedSession() comment updates
Change-Id: I9c545a2c255d2af0826499754555b2e0d95159fb
Aaron Schulz [Thu, 9 Jul 2015 16:07:52 +0000 (09:07 -0700)]
jobqueue: Enable job retries by default in $wgJobTypeConf
* This setting works fine for WMF, including various extensions.
It seems a better default than just pruning out failed jobs.
Change-Id: I635880a49f50544433559ef2cbfb218783b32534
jenkins-bot [Thu, 9 Jul 2015 21:08:51 +0000 (21:08 +0000)]
Merge "Convert Special:Search to OOUI"
Aaron Schulz [Thu, 9 Jul 2015 20:27:44 +0000 (13:27 -0700)]
Made LoadMonitor use wfGlobalCacheKey() and tweaked the key name
* Changing the key name makes sense after
9cf6637 to avoid
problems with het deploy and competing key formats.
Bug: T105373
Change-Id: I2e6888198b1ee4fb4f758070b03d7fe3eac547fc
jenkins-bot [Thu, 9 Jul 2015 19:38:22 +0000 (19:38 +0000)]
Merge "jobqueue: Fix typo in claimTTL documentation"
Translation updater bot [Thu, 9 Jul 2015 19:28:41 +0000 (21:28 +0200)]
Localisation updates from https://translatewiki.net.
Change-Id: I91cbe91552b47cb189f6d86574cbb7ee4ddff320
C. Scott Ananian [Thu, 9 Jul 2015 17:47:52 +0000 (13:47 -0400)]
Sync up with Parsoid parserTests.
This now aligns with Parsoid commit
ece4412e84bdfc7960b69c38207f6981b9f6b596
Change-Id: I5beed6dce381cfbd6ead8b280697c2e044cae889
Bartosz Dziewoński [Sat, 30 May 2015 16:13:00 +0000 (18:13 +0200)]
Convert Special:Search to OOUI
Limiting this to the primary form for now, which was using old
MediaWiki UI styles. There are some native elements tucked in the
advanced form which I am not changing here yet.
Also, use new search suggestions provided by TitleInputWidget.
Bug: T100898
Change-Id: I699cf27a686f2615a22fce70f43b9f89092485c9
jenkins-bot [Thu, 9 Jul 2015 17:06:26 +0000 (17:06 +0000)]
Merge "build: Move PHPCS configuration to phpcs.xml"
Timo Tijhof [Thu, 9 Jul 2015 16:52:07 +0000 (17:52 +0100)]
jobqueue: Fix typo in claimTTL documentation
Change-Id: I0232492db424a0dd6bea23d04772f14980a73fd4
jenkins-bot [Thu, 9 Jul 2015 16:42:17 +0000 (16:42 +0000)]
Merge "MovePage: Don't try to pass non-strings into Title::newFromText"
jenkins-bot [Thu, 9 Jul 2015 14:54:47 +0000 (14:54 +0000)]
Merge "T105242: U is for urlencode"
jenkins-bot [Thu, 9 Jul 2015 14:20:24 +0000 (14:20 +0000)]
Merge "objectcache: Ensure an integer put into memcached comes out an integer"
jenkins-bot [Thu, 9 Jul 2015 13:42:50 +0000 (13:42 +0000)]
Merge "Handle no pageids response field in MessagePoster factory"
jenkins-bot [Thu, 9 Jul 2015 13:32:36 +0000 (13:32 +0000)]
Merge "APIQueryImageInfo: Set META_BC_BOOLS for 'value' in 'metadata' block"
Kunal Mehta [Mon, 29 Jun 2015 21:35:38 +0000 (14:35 -0700)]
build: Move PHPCS configuration to phpcs.xml
"composer phpcs" now runs PHPCS against the relevant directories.
Change-Id: Ic28eaff2318fa57977689957781538acb2b1893c
Alex Monk [Mon, 6 Jul 2015 20:30:32 +0000 (21:30 +0100)]
MovePage: Don't try to pass non-strings into Title::newFromText
Bug: T104852
Change-Id: I240c0993977a8e7e584b93e95ace0907f7b7825c
jenkins-bot [Thu, 9 Jul 2015 11:06:22 +0000 (11:06 +0000)]
Merge "Do not encode '~' as %7E. Fixes redirect loop in chrome."
Brian Wolff [Thu, 9 Jul 2015 05:19:35 +0000 (23:19 -0600)]
Do not encode '~' as %7E. Fixes redirect loop in chrome.
As of
155d555b83eca640, we now redirect variations on hex escapes
into their canonical form. This was causing '~' to be redirected
to %7E. However google chrome seems to canonicalize %7E back
into ~, causing a redirect loop.
RFC 3986 says ~ is unreserved, so not hex encoding it should be
fine. To quote: "For consistency, percent-encoded octets in the
ranges of...tilde (%7E) should not be created by URI producers"
Bug: T105265
Change-Id: I01556eee496e2fb540de8ff09c082c1fedddb5f7
jenkins-bot [Thu, 9 Jul 2015 09:40:59 +0000 (09:40 +0000)]
Merge "Allow detected external links to linebreak at any place"
Derk-Jan Hartman [Tue, 7 Jul 2015 20:46:23 +0000 (22:46 +0200)]
Allow detected external links to linebreak at any place
These so called 'bare urls', which are not wrapped in [] can be very
long at times and because they are not normal text usually don't wrap
to a new line when needed, causing annoying overflows,
especially on smaller screens and often on talk pages where they are a
bit more common. This is CSS3.
Change-Id: Ied3f5fb52d79318c40d541f2c47ae36ea2dfe504
jenkins-bot [Thu, 9 Jul 2015 06:50:57 +0000 (06:50 +0000)]
Merge "Fix inaccurate documentation re $wgEnotifMinorEdits"
Arlo Breault [Fri, 3 Jul 2015 18:53:06 +0000 (11:53 -0700)]
Cleanup in doTableStuff
Change-Id: I75c0a943b24f96a30c6ee1efc3f0b11388f892b7
jenkins-bot [Thu, 9 Jul 2015 01:45:42 +0000 (01:45 +0000)]
Merge "Sync up with Parsoid parserTests."
Matthew Flaschen [Thu, 9 Jul 2015 00:29:39 +0000 (20:29 -0400)]
Fix inaccurate documentation re $wgEnotifMinorEdits
In actuallyNotifyOnPageChange, the $wgEnotifUserTalk is inside
a check involving $wgEnotifMinorEdits, so it won't fire if
it's a minor edit and $wgEnotifMinorEdits is false.
Change-Id: I6576cb1735db5d9288257e6877b8755450d0dacd
Kunal Mehta [Wed, 8 Jul 2015 22:52:52 +0000 (15:52 -0700)]
APIQueryImageInfo: Set META_BC_BOOLS for 'value' in 'metadata' block
The 'value' property can be a boolean, which was previously output as a
raw boolean.
Bug: T103781
Change-Id: I788f13f33885b7c8f694470e44478907f28cfbfd
Arlo Breault [Wed, 7 Jan 2015 20:03:02 +0000 (12:03 -0800)]
T105242: U is for urlencode
* Urlencode and extension tags don't play nice. The current behaviour
is pretty broken. For example,
{{urlencode:<pre id="one">two</pre>}}
outputs
<pre id="one">two</pre>
which isn't exactly urlencoded.
* Just drop the unexpanded content; anything more seems unsafe. There's
precedent for this already in anchorencode.
* Noticed in the infobox caption for enwiki/%22F%22_Is_for_Fugitive
Change-Id: I0ec5d2fdeb71f762036464f65029afd029e8b4a1
C. Scott Ananian [Wed, 8 Jul 2015 21:56:54 +0000 (17:56 -0400)]
Sync up with Parsoid parserTests.
This now aligns with Parsoid commit
34bd76a1747816ed96952ce81ce2fafa751dbb21
Change-Id: Iabf427db488251e26d55a0eb68b282eccd9eb8df
Kunal Mehta [Wed, 8 Jul 2015 21:57:20 +0000 (14:57 -0700)]
Unbreak missing flags in enhanced RC
$this->recentChangesFlags() was being called on the data twice, so the
second time it was processing the string, causing the flags go missing.
Follows-up
94f153db6a0.
Bug: T105237
Change-Id: I04465d0317eecc1c12e47bc9a74f11986eba99a4
Translation updater bot [Wed, 8 Jul 2015 19:28:18 +0000 (21:28 +0200)]
Localisation updates from https://translatewiki.net.
Change-Id: I5277704a2e9af694cf5b5cd70e3fb291238deade
jenkins-bot [Wed, 8 Jul 2015 18:23:27 +0000 (18:23 +0000)]
Merge "Allow using createAndPromote.php with custom groups"
Brad Jorsch [Wed, 15 Apr 2015 17:44:28 +0000 (13:44 -0400)]
Parser: Avoid producing <span></span> in the TOC
If someone renames a section but wants old targeted links to still work,
<span id="old-anchor"></span> is the usual solution. And sometimes
people put it inside the section header markup, like
== <span id="old-anchor"></span>New name ==
since putting it before makes it be considered part of the previous section
while putting it after causes the browser to scroll the section header
off the screen.
But this has the unfortunate side effect that the TOC text for that
section will be "<span></span>New name". We should strip that useless
empty span.
Bug: T96153
Change-Id: I47a33ceb79d48f6d0c38fa3b3814a378feb5e31e
umherirrender [Wed, 8 Jul 2015 16:56:56 +0000 (18:56 +0200)]
Sort $wgLogActionsHandlers/$wgLogActions/$wgLogHeaders
For better readability sort the keys of the log config vars in
DefaultSettings.php
Change-Id: I2db37f4987fc791402368813c39361e04a83e7e8
Željko Filipin [Wed, 8 Jul 2015 15:56:59 +0000 (17:56 +0200)]
Update environments.yml file according to the documentation
Bug: T105174
Change-Id: Ib049d12f24444029f5aeb8ef3149c731c2b9c810
Timo Tijhof [Wed, 8 Jul 2015 13:45:42 +0000 (14:45 +0100)]
mediawiki.api: Include 'mobile' target in mediawiki.api.parse module
Bug: T104940
Change-Id: Ic654f0e38b27b41df13c0423460c1c12c9aec9c3
Timo Tijhof [Mon, 6 Jul 2015 14:12:30 +0000 (15:12 +0100)]
mediawiki.api: Refactor to use server.respondImmediately
* Simplifies code a lot and removes the need to explicitly flush
and handle the request queue with respond().
This was especially annoying when a request spawned others as
they wouldn't be in the queue yet.
* Make tests more explicit and resilient by specifying what they
respond to instead of assigning to requests[i] directly.
This also makes the failure better if one request isn't made,
instead of throwing for accessing properties on undefined objects.
* Avoid relying on test order for badToken().
Follows-up
7b05096bcae0. Tests should be atomic and not rely on
order. This is already important as QUnit re-orders failed tests.
And in the future they may run concurrently.
Resolve using a unique name (like the other tests).
Also, the previous test wasn't asserting that badToken() works,
it was merely asserting that getToken() works and that it wasn't
yet cached. The new test will fetch and purge its own token.
Change-Id: I26d22ace6c5df19d7144779be1a625aede79749f
jenkins-bot [Wed, 8 Jul 2015 09:22:31 +0000 (09:22 +0000)]
Merge "HTMLTextAreaField: Honor 'rows' setting in OOUI format"
jenkins-bot [Wed, 8 Jul 2015 06:58:06 +0000 (06:58 +0000)]
Merge "OutputPage: Only set <link hreflang=x-default> on wiki/pages with variants"
jenkins-bot [Wed, 8 Jul 2015 06:44:42 +0000 (06:44 +0000)]
Merge "Improved fail-over in ReplicatedBagOStuff for redis"
jenkins-bot [Wed, 8 Jul 2015 06:37:28 +0000 (06:37 +0000)]
Merge "Move `$.extend( OO.ui, mw.widgets )` hack to 'mediawiki.widgets' module"
jenkins-bot [Wed, 8 Jul 2015 06:29:51 +0000 (06:29 +0000)]
Merge "Add Content-Length header to thumb.php error responses"
jenkins-bot [Wed, 8 Jul 2015 06:29:47 +0000 (06:29 +0000)]
Merge "qunit: Restore sinon.sandbox after teardown instead of before"
jenkins-bot [Wed, 8 Jul 2015 06:29:25 +0000 (06:29 +0000)]
Merge "Update Sinon.JS to 1.15.4"
jenkins-bot [Wed, 8 Jul 2015 06:26:43 +0000 (06:26 +0000)]
Merge "Update QUnit to v1.18.0"
James D. Forrester [Wed, 8 Jul 2015 01:40:19 +0000 (18:40 -0700)]
Update OOjs UI to v0.11.8
Release notes:
https://git.wikimedia.org/blob/oojs%2Fui.git/v0.11.8/History.md
Change-Id: I701c504acf0d527916f1925bbf49177cc9c0d883
Kunal Mehta [Wed, 8 Jul 2015 01:27:03 +0000 (18:27 -0700)]
HTMLTextAreaField: Honor 'rows' setting in OOUI format
And throw an exception if 'cols' is set, since it won't work.
Requires
eed0f5294b0080 in OOUI.
Bug: T104682
Change-Id: I8e09402a01cecac8a90497d31b3b1ca15ff2d949
jenkins-bot [Wed, 8 Jul 2015 00:42:00 +0000 (00:42 +0000)]
Merge "Explain API action=parse&disablepp option more clearly"
This, that and the other [Wed, 8 Jul 2015 00:30:52 +0000 (10:30 +1000)]
Explain API action=parse&disablepp option more clearly
Brought up by Purodha on translatewiki.net Support page.
Also change a few ungrammatical "Disable"s to "Omit"s.
Change-Id: I9a14fa16cc166471d66c53489e8186ec09297399
Matthew Flaschen [Wed, 1 Jul 2015 07:21:53 +0000 (03:21 -0400)]
Allow using createAndPromote.php with custom groups
This lets any group(s) be added using a comma-separated argument,
in addition to the --sysop, --bureaucrat, and --bot that already
existed.
Bug: T105079
Change-Id: I1274b065e3dad917e545f9278b996da014d87ae9
Translation updater bot [Tue, 7 Jul 2015 19:42:13 +0000 (21:42 +0200)]
Localisation updates from https://translatewiki.net.
Change-Id: I3d2c6254a862d635419018c2c4fe3f824ae43a93
jenkins-bot [Tue, 7 Jul 2015 18:53:35 +0000 (18:53 +0000)]
Merge "resourceloader: Set a TTL for minification cache entries"
Timo Tijhof [Tue, 7 Jul 2015 18:42:43 +0000 (19:42 +0100)]
backupTextPassTest: Disable flaky test 'testCheckpointPlain'
Follows-up
9107fb539111de.
Ref T70653
Change-Id: I493c0a97bd9b5213a3f1f6ed0b2a6c591fd67d3b
Brad Jorsch [Tue, 7 Jul 2015 17:46:17 +0000 (13:46 -0400)]
API: Move stats update for API module execution to when the module is actually executed
The place it was before could easily be reached without actually
executing (or even constructing) the module.
Change-Id: I6f807adc9cbf71c5d7b83c7eec43965dce1d2a16
raymond [Tue, 7 Jul 2015 14:16:23 +0000 (16:16 +0200)]
Add alias for Special:ChangeContentModel
Follow up to
fa31bf1675af725.
Change-Id: I2589b8a48c82484f8ef561d9562b28f613fa15b1
Timo Tijhof [Tue, 7 Jul 2015 16:40:00 +0000 (17:40 +0100)]
resourceloader: Set a TTL for minification cache entries
Follows-up
458e7cabbbafd. HHVM's APC cache doesn't have any
limitation or eviction logic. Thus entries without a TTL will
linger indefinitely.
At time of writing over 99% of APC entries were from "resourceloader:filter"
because there are far more unpopular resources than popular ones.
A fixed TTL is suboptimal as it will also cause popular resources
(e.g. startup module, or jquery) to be re-minified every day.
Beware this isn't one single minification. It's one minification,
for every JS/CSS resource, on every app server, for many different
lang/skin/config variations.
Bug: T104769
Change-Id: Ib62c320da0eeede68aad3f2294006ec7dcf1df55
Timo Tijhof [Tue, 7 Jul 2015 14:46:32 +0000 (15:46 +0100)]
qunit: Restore sinon.sandbox after teardown instead of before
This was causing pending (mocked) XHR requests to sometimes be
cleaned up prematurely.
As a general rule, setup always calls its parent first. And teardown
always calls its parent last. The inner callback runs within the
scope of the outer one and should still have acces to the sinon mocks.
Change-Id: Ic4d9243c04af73e529b689e650d35ddabaa7f4b9
Timo Tijhof [Tue, 7 Jul 2015 09:28:16 +0000 (10:28 +0100)]
Update Sinon.JS to 1.15.4
* Code
http://sinonjs.org/releases/sinon-1.15.4.js
http://sinonjs.org/releases/sinon-ie-1.15.4.js
* Change log
http://sinonjs.org/Changelog.txt
https://github.com/cjohansen/Sinon.JS/commits/v1.15.4
Mainly:
* Support XHR.responseType 'json'.
* Remove deprecation warnings from sinon.mock().
Change-Id: I4e0d8f044eb2841ab361b1efc76fb36797f15e42
Timo Tijhof [Tue, 7 Jul 2015 09:22:35 +0000 (10:22 +0100)]
Update QUnit to v1.18.0
* Code
https://code.jquery.com/qunit/qunit-1.18.0.css
https://code.jquery.com/qunit/qunit-1.18.0.js
* Change log
https://github.com/jquery/qunit/blob/1.18.0/History.md
Change-Id: I9da13eedb2f9b5ee936aaa658c3da6ce222cfa0a
jenkins-bot [Tue, 7 Jul 2015 14:05:59 +0000 (14:05 +0000)]
Merge "API: Use ApiBase::PARAM_HELP_MSG_PER_VALUE for 'prop' in query+watchlist"
jenkins-bot [Tue, 7 Jul 2015 13:47:50 +0000 (13:47 +0000)]
Merge "mediawiki.searchSuggest: Allow multiple suggestion fields"
Bartosz Dziewoński [Tue, 7 Jul 2015 13:23:25 +0000 (15:23 +0200)]
Move `$.extend( OO.ui, mw.widgets )` hack to 'mediawiki.widgets' module
OO.ui.infuse assumes all widgets are in the OO.ui. namespace.
We previously made that happen on document-ready in
'mediawiki.page.ready' module, but that's too late: document-ready
handlers defined by other modules' can run earlier.
Let's just do it in 'mediawiki.widgets' after all widgets are defined.
Change-Id: Ib2184bf87d72db64d1565d742ef9f1be37f24b6f
Timo Tijhof [Tue, 7 Jul 2015 10:32:03 +0000 (11:32 +0100)]
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
jenkins-bot [Tue, 7 Jul 2015 04:45:24 +0000 (04:45 +0000)]
Merge "mediawiki.ui: Add @activeColor argument to the .button-colors() mixin"
Prateek Saxena [Fri, 27 Mar 2015 06:26:07 +0000 (11:56 +0530)]
mediawiki.ui: Add @activeColor argument to the .button-colors() mixin
Have kept the colors same for now except for .mw-ui-button:active and
.mw-ui-button.mw-ui-checked . Can be changed later once the design
team has decided.
Change-Id: I7c418b970c5e5fa95f740ecf3d90622bf7f02364
Bryan Davis [Mon, 6 Jul 2015 22:46:45 +0000 (16:46 -0600)]
Update composer-merge-plugin to v1.2.1
Fixes bug in optimized autoloader generation on initial plugin install.
Change-Id: If637d2b5583b32895fe8839478e7e553927f1d68
Requires: I22d1fd4044320613c6ba61ac5e84263e5af4ff27
Bug: T104037
jenkins-bot [Mon, 6 Jul 2015 22:41:31 +0000 (22:41 +0000)]
Merge "TitleInputWidget: Add 'relative' option"
Ori Livneh [Mon, 6 Jul 2015 20:49:45 +0000 (13:49 -0700)]
Add Content-Length header to thumb.php error responses
Without the Content-Length header, the response is sent with Transfer-Encoding:
Chunked, which is somehow mangled by mod_fastcgi. Varnish then claims the
response is malformed and declines to process it, sending the client a 503
instead.
Change-Id: I977387bf3e039926df58e3e5a14d0d0bd55201c4
jenkins-bot [Mon, 6 Jul 2015 20:29:23 +0000 (20:29 +0000)]
Merge "Allow hooks to abort lines in EnhancedRC"
jenkins-bot [Mon, 6 Jul 2015 20:17:39 +0000 (20:17 +0000)]
Merge "API: Avoid caching when prop=info&intestactions=... is used"
Kunal Mehta [Tue, 23 Jun 2015 21:23:07 +0000 (14:23 -0700)]
API: Use ApiBase::PARAM_HELP_MSG_PER_VALUE for 'prop' in query+watchlist
Change-Id: I5e956f4b4dccb2b109d514332e4347252ce846c4
jenkins-bot [Mon, 6 Jul 2015 20:04:58 +0000 (20:04 +0000)]
Merge "Add hook EnhancedChangesListModifyBlockLineData"
jenkins-bot [Mon, 6 Jul 2015 19:30:12 +0000 (19:30 +0000)]
Merge "Add tests for Special:BlankPage"
Translation updater bot [Mon, 6 Jul 2015 19:29:52 +0000 (19:29 +0000)]
Merge "Localisation updates from https://translatewiki.net."
jenkins-bot [Mon, 6 Jul 2015 19:29:01 +0000 (19:29 +0000)]
Merge "Add SpecialPageTestBase to simplify testing of special pages"
addshore [Mon, 6 Jul 2015 10:49:58 +0000 (12:49 +0200)]
Add tests for Special:BlankPage
This uses SpecialPageTestBase and demonstrates how that class
should be used.
Change-Id: Ibbf3fdcb172d7c1c71491d27b534faa65fd9be63
addshore [Sat, 9 Aug 2014 00:37:44 +0000 (01:37 +0100)]
Add SpecialPageTestBase to simplify testing of special pages
This is copied from Wikibase. A follow up change adds a test
for Special:BlankPage to demonstrate how it works.
Change-Id: I3c34fd17ceb0049b160ec4f821474de457533983
Translation updater bot [Mon, 6 Jul 2015 19:05:31 +0000 (21:05 +0200)]
Localisation updates from https://translatewiki.net.
Change-Id: Ifcf5396c7f9a8a6abb43dd9233dc6c55b6ca00c1
jenkins-bot [Mon, 6 Jul 2015 16:57:53 +0000 (16:57 +0000)]
Merge "Add {{bidi:}} syntax for directionality-safe arguments"
Brad Jorsch [Mon, 6 Jul 2015 16:07:45 +0000 (12:07 -0400)]
API: Avoid caching when prop=info&intestactions=... is used
Whether the actions pass or not depends on the current user, including
the results of several hooks and whether the current user is blocked.
Also, we can clean up the test for public inprops slightly.
Change-Id: I7e2610dd609c52711355b640778f795520a4ff8c