jenkins-bot [Sun, 6 Apr 2014 16:19:47 +0000 (16:19 +0000)]
Merge "Remove unused 'prefs-beta' message"
jenkins-bot [Sun, 6 Apr 2014 13:01:36 +0000 (13:01 +0000)]
Merge "mediawiki.action.edit.js: Remove WikiEditor iframe hack from 2010"
jenkins-bot [Sun, 6 Apr 2014 11:43:18 +0000 (11:43 +0000)]
Merge "mediawiki.util: Init tooltipAccessKeyPrefix immediately instead of on domready"
Fomafix [Sat, 5 Apr 2014 18:46:53 +0000 (18:46 +0000)]
mediawiki.util: Init tooltipAccessKeyPrefix immediately instead of on domready
Change-Id: I60e30e1d029d29b713b229bf2b0313e92a59f120
jenkins-bot [Sun, 6 Apr 2014 09:26:10 +0000 (09:26 +0000)]
Merge "Make chunked upload jobs robust in face of exceptions."
jenkins-bot [Sun, 6 Apr 2014 09:26:03 +0000 (09:26 +0000)]
Merge "When uploading a new file, save to memcached directly after commit"
jenkins-bot [Sun, 6 Apr 2014 09:12:02 +0000 (09:12 +0000)]
Merge "Make doEditContent call $dbw->rollback() if exception happens"
TheDJ [Sun, 6 Apr 2014 08:53:03 +0000 (08:53 +0000)]
Merge "shared.css: Move sup, sub styling from commonElements to shared styles"
Brian Wolff [Sun, 6 Apr 2014 03:16:37 +0000 (00:16 -0300)]
Make doEditContent call $dbw->rollback() if exception happens
doEditContent calls a variety of methods in the middle of its
database transaction that can throw exceptions. For example if
$revision->insertOn() throws an exception, it can cause
referential integrity issues unless the transaction is rolled
back by something further up the chain (Thus see I5807e, I8f1da5).
It seems like it would be more reliable if doEditContent rolled
back its own transaction in exceptional circumstances. So this
patch puts everything in a try block, and does rollback if
an exception happens.
Bug: 63145
Bug: 32551
Change-Id: Icc44687da4edb1a72f13d2b95bfee2eea3ad6808
Bartosz Dziewoński [Sat, 8 Feb 2014 13:48:37 +0000 (14:48 +0100)]
Optimize all SVGs
Once more, with feeling. First reverted attempt:
a9b855ee.
Passed with svgo 0.4.2 <https://github.com/svg/svgo>, then prepended
'<?xml version="1.0" encoding="UTF-8"?>' prolog to each file since
svgo sees it fit to remove it and that breaks just about everything.
Previous attempt used a '<?xml version="1.0" ?>' prolog, which
apparently is too short to ensure the file is picked up as an SVG in
some versions of the 'file' utility; see PleaseStand's comments on
bug 59234. We could just add some spaces instead of 'encoding="UTF-8"',
but it seems like a reasonable thing to add.
Skipped the ones in tests/, there's some weird stuff there and they
didn't shrink much when I tried anyway.
Total file size reduced from ~30 KB to ~10 KB.
---
For future reference, the prolog appending was done with the following
script:
for i in `find . -name '*.svg'`
do
mv $i $i-tmp
(echo -n '<?xml version="1.0" encoding="UTF-8"?>'; cat $i-tmp) > $i
rm $i-tmp
done
---
Bug: 59744
Change-Id: Ic9fdd081003b694dbd7e413ca25fffe782b912b5
jenkins-bot [Sun, 6 Apr 2014 02:31:43 +0000 (02:31 +0000)]
Merge "Rollback transaction in case of uncaught exceptions during job"
Brian Wolff [Sun, 6 Apr 2014 02:50:10 +0000 (23:50 -0300)]
When uploading a new file, save to memcached directly after commit
Previously if an exception was thrown when creating the description
page, the old (negative) memcached entry would not be purged, and
the file would appear not to exist until 24 hours later.
This way, as soon as we have the image row committed to db, we
change the memcache entry, so memcache of image row always looks
like actual image row.
Bug: 32551
Change-Id: I9c260e9a4a1e0d6ade1771542dd834f969904cef
Brian Wolff [Sun, 6 Apr 2014 02:15:26 +0000 (23:15 -0300)]
Make chunked upload jobs robust in face of exceptions.
If an exception occurs when creating the new image page during
PublishStashedFile (for example, if ExternalStorage servers
become overloaded), the job is aborted, but the current
transaction is not aborted. Then the next job comes along, does
$dbw->begin(), which implicitly commits the current open
transactions, causing database corruption (entries in page table
with page_latest = 0).
This patch ensures that if some unknown exception occurs during
the job, that the database gets rolled back to a known good state.
I also plan to make a separate patch to make the recordUpload
code handle exceptions properly. Nonetheless, I think its
important that the general job code be as robust as possible.
Bug: 32551
Change-Id: Icb5e06e81700e5dcf0d4e739c7b2bb3221e718b8
Brian Wolff [Sun, 6 Apr 2014 02:02:02 +0000 (23:02 -0300)]
Rollback transaction in case of uncaught exceptions during job
If there is an unhandled exception during a job, rollback the
transaction to be safe. Otherwise its likely that the
transaction will be commited when next job is run, as $dbw->begin()
implicitly commits any open pending transactions. This can result
in database referential integrity issues.
Change-Id: I5807e64440ff6c6651fbbb4924645d05d843b98e
jenkins-bot [Sun, 6 Apr 2014 01:12:18 +0000 (01:12 +0000)]
Merge "WatchAction: Add 'returnto' and 'returntoquery' to the login link"
konarak [Fri, 4 Apr 2014 08:35:28 +0000 (14:05 +0530)]
WatchAction: Add 'returnto' and 'returntoquery' to the login link
Bug: 60594
Change-Id: Ic8c0ea5b203c63f9273992cb70bfb06c0d461311
Bartosz Dziewoński [Sat, 5 Apr 2014 23:49:03 +0000 (01:49 +0200)]
Remove unused 'prefs-beta' message
Unused since change Id8b9a47a in WikiEditor (9 months ago).
Bug: 28555
Change-Id: I76667a15ccbb1f6bdc5e5a67ffcb0a65ce55d124
Bartosz Dziewoński [Tue, 4 Mar 2014 20:31:47 +0000 (21:31 +0100)]
mediawiki.special: Kill width: 98% on borderless tables
For Special:Prefixindex this dates back to the stone age, I didn't
bother finding out what the reasoning was. width: 100% was clearly
intended.
For Special:EmailUser there isn't even such an element anywhere.
Change-Id: Ie3a7e7fa00961b62e46e3f8cd05fd2e7d3c25f06
jrobson [Fri, 4 Apr 2014 00:58:56 +0000 (17:58 -0700)]
Vector: Simply style external links
These external link icons are extremely costly with little gain.
During the typography experiment, these were removed, no one
raised this as an issue during the 5 months it was live, suggesting
that explicit pdf icons etc are not as useful as we like to believe.
As a result of these findings posting this patch upstream.
Bug: 54604
Change-Id: If985b16c4f682d737683597ed80951c6d6644c8f
Max Semenik [Sat, 5 Apr 2014 21:51:50 +0000 (01:51 +0400)]
Per wikitech-l, remove premature wfDeprecated()
This way of deprecation is way too rushed, we generally don't mark as
@deprecated and put wfDeprecated() in the same release.
Change-Id: I5dc07d9819b6fb77fcf3194ed16676eaacb834a8
umherirrender [Sat, 5 Apr 2014 19:54:24 +0000 (21:54 +0200)]
Fix spacing after @param and friends in comments
Searched for:
\@(param|return|throws|since|deprecated|access|todo|var)[ \t]{2,}
Change-Id: Icce22ba9fe0635455691ca58d9872d618151f346
Translation updater bot [Sat, 5 Apr 2014 19:57:38 +0000 (19:57 +0000)]
Merge "Localisation updates from https://translatewiki.net."
Translation updater bot [Sat, 5 Apr 2014 19:35:17 +0000 (19:35 +0000)]
Localisation updates from https://translatewiki.net.
Change-Id: I44b9547b386f7c169a7add27096e2708d5a3cce9
umherirrender [Sat, 5 Apr 2014 19:13:38 +0000 (21:13 +0200)]
fix typo in comment in Preferences.php
The name of the preference was spelled wrong.
Change-Id: I443d0a8fc67c89da36b5c725dd3f3513d4170357
jenkins-bot [Sat, 5 Apr 2014 19:11:24 +0000 (19:11 +0000)]
Merge "Remove trailing line break check from MemcachedBagOStuff::debugLog()"
Alexandre Emsenhuber [Sat, 5 Apr 2014 09:06:37 +0000 (11:06 +0200)]
Correctly order AutoLoader class definitions
- Group per directory
- Sort by class name for each directory
Change-Id: If4771f2868c2183f60a6321533142a1ad5cdc7cd
Alexandre Emsenhuber [Sat, 5 Apr 2014 18:28:06 +0000 (20:28 +0200)]
Remove trailing line break check from MemcachedBagOStuff::debugLog()
wfDebugLog() already does the same, so it's unnecessary here.
Change-Id: Ia7bbbd2806fcd638861d010835c33969322d262a
jeroendedauw [Sat, 5 Apr 2014 17:49:21 +0000 (19:49 +0200)]
Fix path after file got moved
Change-Id: Ic67f590fb53f2b96cfd8a0192968cb4953bb0d2b
Rob Moen [Sat, 5 Apr 2014 03:01:42 +0000 (20:01 -0700)]
Fix documentation typo
Change-Id: I535dd23eeccd6686c64a885f06e030b756e67fec
Raimond Spekking [Fri, 4 Apr 2014 20:34:27 +0000 (20:34 +0000)]
Revert "Localisation updates from https://translatewiki.net."
wtf?
This reverts commit
ff744475d7d37d4887ebaba3d69884317a736c34.
Change-Id: Ie7ea50de88f87aa527205a816837dbfa5950bb0b
Translation updater bot [Fri, 4 Apr 2014 20:32:02 +0000 (20:32 +0000)]
Localisation updates from https://translatewiki.net.
Change-Id: Idfe1499c79cd459377b31ca022fc45a9e0b3e833
jenkins-bot [Fri, 4 Apr 2014 19:43:07 +0000 (19:43 +0000)]
Merge "mw.hook: Make hook.fire actually chainable"
jenkins-bot [Fri, 4 Apr 2014 19:43:05 +0000 (19:43 +0000)]
Merge "Update jquery.fullscreen to newer version (fix IE11 support)"
Timo Tijhof [Fri, 4 Apr 2014 01:35:33 +0000 (18:35 -0700)]
mw.hook: Make hook.fire actually chainable
It was being called without being bound to the mw.hook instance,
as such the (detachable) callbackList.fireWith method (which
returns its given 'this') returned the callbackList instance
instead of the mw.hook object literal.
The tests weren't catching it because it was just checking one
can call .fire() without it crashing, but it was in fact calling
list.fire instead of hook.fire (which unlike #add and #remove are
not compatible, and we might add new methods later on).
Change-Id: If3d4dfbed494e7ff9f32514e539cfa089aea18ac
jenkins-bot [Fri, 4 Apr 2014 08:11:44 +0000 (08:11 +0000)]
Merge "Add a dot at the end of the deprecate messages"
Timo Tijhof [Thu, 14 Nov 2013 16:54:19 +0000 (17:54 +0100)]
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
jenkins-bot [Fri, 4 Apr 2014 00:54:54 +0000 (00:54 +0000)]
Merge "Callers of ResourceLoader::getModule should check for null return"
Ori Livneh [Thu, 3 Apr 2014 23:09:09 +0000 (16:09 -0700)]
Callers of ResourceLoader::getModule should check for null return
ResourceLoader::getModule returns null if cannot find a ResourceLoaderModule
instance or a module configuration array matching the requested name. Existing
callers of that method chain ResourceLoaderModule method calls to the return
value of ResourceLoader::getModule without bothering to check for a null
result. This patch makes the documentation for the method more explicit, and it
adds checks for the return value to existing calls of the method. In some
cases, it is possible to reason to that the call will succeed, on the
assumption that missing modules would have been filtered out by earlier
iterations through the modules array. But in the interest of robustness, it is
always good to check.
Bug: 63310
Change-Id: Ic32f1e61ffee0383f7a8761423099041e3b6b8cc
James D. Forrester [Thu, 3 Apr 2014 23:56:26 +0000 (16:56 -0700)]
Update OOjs UI to v0.1.0-pre (
eaa1b7f06d)
New changes:
468b1c8 Localisation updates from https://translatewiki.net.
3e2a5e1 Update OOjs to v1.0.9
eaa1b7f Add pressed state to option widgets and use instead of initializing items
Change-Id: I6e93d9ab0e4a926c09c5e08e3db045cff99d2098
jenkins-bot [Thu, 3 Apr 2014 23:07:58 +0000 (23:07 +0000)]
Merge "New API module, list=prefixsearch"
Aaron Schulz [Thu, 3 Apr 2014 00:25:14 +0000 (17:25 -0700)]
Added per-wiki queue stats information
Change-Id: I25f42408c9a7a368aa99f4ceb04214f188a74be6
Bartosz Dziewoński [Thu, 3 Apr 2014 20:34:08 +0000 (22:34 +0200)]
mediawiki.action.edit.js: Remove WikiEditor iframe hack from 2010
Introduced in
aa69029e2 (r62242) by Roan.
The iframe this refers to is going away for good in I2d736911.
Change-Id: If2f225c51791b46e25468deb4081ecac67db409a
Brian Wolff [Sat, 1 Mar 2014 07:18:45 +0000 (03:18 -0400)]
Use rgba() for gallery caption background instead of opacity
Opacity applies to child elements, and can't be unset.
Change-Id: Id4b58ccf82cf3b6d23f67408c5311da0eb92fbce
Bartosz Dziewoński [Thu, 3 Apr 2014 17:43:03 +0000 (19:43 +0200)]
shared.css: Move sup, sub styling from commonElements to shared styles
This really should be an even more basic style than I thought when
moving this CSS here from Vector-specific stylesheet.
Follow-up to Ie98f8b57 and Ic5ba8363.
Bug: 49965
Change-Id: I80f8279e10277d1a7389c821037d733ccfc4e166
jenkins-bot [Thu, 3 Apr 2014 17:41:19 +0000 (17:41 +0000)]
Merge "mediawiki.api: Refactor getToken and postWithToken methods"
jenkins-bot [Thu, 3 Apr 2014 14:59:24 +0000 (14:59 +0000)]
Merge "ApiFeedRecentChanges: Validate param target"
Fomafix [Thu, 3 Apr 2014 14:52:35 +0000 (14:52 +0000)]
Add a dot at the end of the deprecate messages
Change-Id: I73b4bc296051bf506683434de9bc7b56f3ff825f
Niklas Laxström [Sat, 29 Mar 2014 16:00:18 +0000 (16:00 +0000)]
ApiFeedRecentChanges: Validate param target
Change-Id: I6933177d47c25d48b9b55d479afa4dedcc64299d
jenkins-bot [Thu, 3 Apr 2014 07:42:44 +0000 (07:42 +0000)]
Merge "mediawiki.html: Add missing @return documentation"
Timo Tijhof [Thu, 3 Apr 2014 01:05:13 +0000 (18:05 -0700)]
mediawiki.html: Add missing @return documentation
Change-Id: I63f12ee44aace6cf4f7762dd67244648b95d7301
jenkins-bot [Thu, 3 Apr 2014 00:52:23 +0000 (00:52 +0000)]
Merge "Update OOjs UI to v0.1.0-pre (
8986c46d35)"
S Page [Wed, 2 Apr 2014 23:25:22 +0000 (16:25 -0700)]
Adapt signup form numbers to new Vector typography
The .mw-number-text CSS was less specific than Vector's typography
styling of 'div#content h3'. Match Vector's serif font stack.
Bug: 63450
Change-Id: I23041c356ad4356dcf085479721cb950a89b33d0
James D. Forrester [Wed, 2 Apr 2014 21:40:24 +0000 (14:40 -0700)]
Update OOjs to v1.0.9
Code:
https://oojs.github.io/releases/oojs-1.0.9.js
Release notes:
https://git.wikimedia.org/blob/oojs%2Fcore.git/v1.0.9/History.md
Change-Id: Ic1fea7dd3a07af32a29511191c45da90bc306b17
Timo Tijhof [Wed, 5 Feb 2014 20:54:20 +0000 (12:54 -0800)]
mediawiki.api: Refactor getToken and postWithToken methods
* getToken now keeps promise objects around so that it can:
1) Re-use already retrieved tokens instead of always making
a new request (this was previously done locally inside
one of the callers, namely #postWithToken).
2) By storing promises instead of the tokens, we can re-use
it asynchronously. While it is unlikely that one module
calls getToken twice in a short period amount of time,
it is quite likely that two separate modules (both using
mediawiki.api) request similar tokens from the API at the
same time (e.g. at page load).
This cache has to be additionally keyed by API endpoint since
mw.Api is a generic class that can be (and in practice, is)
instantied for foreign wiki as well.
* Initialise the cache with pre-resolved promises for values
from user.tokens from the local wiki (embedded in the page
output, no point in requesting these again from the API).
* postWithToken no longer has its own token cache, it now
unconditionally uses getToken.
* postWithToken already took care of clearing the token cache when
the server responds with a 'badtoken' error. However, it then
changed the request to delete the 'token' query and try again
without any token. That seems rather silly. Change this to
instead have getToken fetch a fresh one and re-submit with that.
* Add unit tests.
Bug: 34733
Change-Id: I5c25ae5ea4bf3336899bc60fd94ec3b0948050e1
Erik Bernhardson [Wed, 2 Apr 2014 20:22:33 +0000 (13:22 -0700)]
Better memory reporting when running in HHVM
Change-Id: If06b7b3ff4dc7ba8b873361f7878f8abb718f571
Erik Bernhardson [Sat, 28 Dec 2013 22:08:32 +0000 (14:08 -0800)]
Visualize profile output in debug toolbar
This adds a new pane for the debug toolbar labeled Profiler. It groups
together events sent to wfProfileIn/Out and draws a timeline giving a
visual representation of what the page was doing at a particular point
in time. Hovering over a timeline brings up a flyout about the contained
events.
Depends on I08854d80f57286129f19cb14566f84ed3c7fa21c for proper positioning
of the flyout.
Change-Id: I23a0ce4137abcd74ad87967fb263214c08cb9685
jenkins-bot [Wed, 2 Apr 2014 21:50:27 +0000 (21:50 +0000)]
Merge "Made BagOStuff fail fast in cas/lock on certain errors"
jenkins-bot [Wed, 2 Apr 2014 21:23:29 +0000 (21:23 +0000)]
Merge "mediawiki.ui: Do not depend on divs being present inside .mw-ui-vform"
jenkins-bot [Wed, 2 Apr 2014 19:20:54 +0000 (19:20 +0000)]
Merge "Fix language statistics after messages have been moved to JSON"
Erik Bernhardson [Mon, 30 Dec 2013 20:03:50 +0000 (12:03 -0800)]
Respect sizing from user-supplied class
A user-supplied class provided to tipsy that affects the
height or width of the flyout breaks tipsys positioning.
Adjust tipsy to apply the user-supplied class earlier so
its css rules are taken into account during positioning.
Also clean up trailing whitespace.
Change-Id: I08854d80f57286129f19cb14566f84ed3c7fa21c
jenkins-bot [Wed, 2 Apr 2014 18:29:40 +0000 (18:29 +0000)]
Merge "Cleaning up function definitions in Language.php"
kaldari [Tue, 1 Apr 2014 00:31:06 +0000 (17:31 -0700)]
Cleaning up function definitions in Language.php
Some of these were confusing or incorrect.
Change-Id: I2dd02669a7cf9d8b92e2d5c1b86ab8c5e0cbf6bd
Gilles Dubuc [Tue, 1 Apr 2014 06:49:37 +0000 (08:49 +0200)]
Update jquery.fullscreen to newer version (fix IE11 support)
jquery.fullscreen wouldn't work properly in IE11. This commit fixes the
issues that were preventing it from working in that browser.
Source
* https://github.com/theopolisme/jquery-fullscreen/blob/
9f8f97d127/jquery.fullscreen.js
Change log
* https://github.com/theopolisme/jquery-fullscreen/compare/v2.0.0...
9f8f97d127
Change-Id: I6cb853f05ef9a464990a785827074c78ac61ac7f
Aaron Schulz [Wed, 2 Apr 2014 18:08:22 +0000 (11:08 -0700)]
Re-organized PoolCounter files
* No actual changes to the classes
Change-Id: I4e35a2b097e4bc72315688d0390d3f5c4c1ca4ed
Alexandre Emsenhuber [Wed, 2 Apr 2014 18:05:51 +0000 (20:05 +0200)]
Fix language statistics after messages have been moved to JSON
Statistics are now showing always no messages since languages.inc
does not manage to load the JSON messages.
Had to make LocalisationCache::readJSONFile() public so that it can
be used in languages.inc; since all methods from LocalisationCache
return localisations merged with fallbacks, which is not what we
want here.
Fix for I918cfdc46c (
0dd91d5).
Change-Id: Ib52287db618b9d072e847130070d165a3e7ae44b
jenkins-bot [Wed, 2 Apr 2014 17:26:33 +0000 (17:26 +0000)]
Merge "Adding debugging for bug 62241"
aude [Wed, 2 Apr 2014 15:37:25 +0000 (17:37 +0200)]
Rename $p to $parts in Special:Search
Change-Id: I794fca2d24fcebb59067d5e619b68d86cbaa2a8f
aude [Wed, 2 Apr 2014 15:31:06 +0000 (17:31 +0200)]
Fix variable naming in SpecialSearch, $t => $title
Change-Id: I70ec1e147f99a73bc0b12511669ab5a799018695
Max Semenik [Tue, 1 Apr 2014 22:35:20 +0000 (02:35 +0400)]
New API module, list=prefixsearch
It's pretty much like action=opensearch but can be used as a generator
which can be handy when you need to retrieve both the list of pages and
some information about them.
Change-Id: Iaffe30a0f7402e1316c4885a805692a34bbe1a6a
aude [Wed, 2 Apr 2014 14:54:52 +0000 (16:54 +0200)]
Rename $t to $title in SpecialSearch
$title is a bit more informative than $t
Change-Id: I66deeb6672dc45161e0fcc30988bcc1efcb93f56
shirayuki [Wed, 2 Apr 2014 11:51:37 +0000 (20:51 +0900)]
i18n: Fix the parser function name
- Fix the parser function name "#ifexist"
- Put the parser function in code tags
Spotted by Siebrand
https://translatewiki.net/wiki/Thread:Support/About_MediaWiki:Expensive-parserfunction-category-desc/en
Change-Id: If29585ceda7f409c668804ff13c821b4eece95fd
Siebrand Mazeland [Wed, 2 Apr 2014 09:24:05 +0000 (11:24 +0200)]
Add punctuation to "pool-servererror"
Follow-up on
eb03cf3f7dda4.
Change-Id: Iee1fd4ca9f4baad0d03998dbc1ea1e24f2210ba7
jenkins-bot [Wed, 2 Apr 2014 04:37:46 +0000 (04:37 +0000)]
Merge "Put LESS test back into extension suite in XML"
James D. Forrester [Wed, 2 Apr 2014 02:33:44 +0000 (19:33 -0700)]
Update OOjs UI to v0.1.0-pre (
8986c46d35)
New changes:
735cebc Filter calls to selectItem and highlightItem on SelectWidget
897b30f Return null in Element#getDocument
b5ede88 Add Element#isElementAttached
68df19f Disable tool groups when all its tools are disabled
1aaa507 Grunt-watch less files
e4dd7b2 build: Update devDependencies
64d6c32 Localisation updates from https://translatewiki.net.
b5c6db5 Localisation updates from https://translatewiki.net.
b9460cb apex/button: Reset border-color of disabled button when not hovered
32c4799 Localisation updates from https://translatewiki.net.
674165e Enable jscs for OOjs UI repo
d98b338 grunt: Add messageDocumentation task to ensure each 'en' message has a 'qqq'
d421dd6 demos/tools: Indentation fix
Change-Id: Ic9b7ff01fe066eb5cba6a0fd0fc515ff3c62b66a
Aaron Schulz [Sat, 15 Mar 2014 21:17:16 +0000 (14:17 -0700)]
Added a Redis pool counter class
* This should be easier to set up for typical installs
Change-Id: Icb4a7481b944fa0818c4635e3edbe12d08af9924
Timo Tijhof [Tue, 1 Apr 2014 19:47:32 +0000 (12:47 -0700)]
vector: Remove redundant margin-top reset
There is nothing setting this to anything else. The only other
preceeding relevant rule inherited before this is "h1, h2, h3 .."
which sets "margin: 0;" already.
Change-Id: If6531221891ae3c7eb792c5bdc930ddaaf613f52
Matthew Flaschen [Tue, 1 Apr 2014 21:27:35 +0000 (17:27 -0400)]
Put LESS test back into extension suite in XML
Followup to
13f3e21e9c7ddcef0d59982aeb3bb4b9e240e5f4
Change-Id: I1a15a0d8aba05b774f072673787b712f6ba97fcd
Siebrand Mazeland [Tue, 1 Apr 2014 13:36:07 +0000 (15:36 +0200)]
Migrate translations for core to JSON format
Largest patch set in the history of MediaWiki!
Change-Id: I918cfdc46c099017ca3622d1bcbbdd30c4fd00f6
Siebrand Mazeland [Tue, 1 Apr 2014 10:32:00 +0000 (12:32 +0200)]
Preparations for migrating core to use JSON based i18n
LocalisationCache and Language have to take the JSON files into account
in deciding if a language is present or not.
Standardizing language validity checking with isSupportedLanguage
and isValidBuiltInCode.
Co-Authored-By: Niklas Laxström <niklas.laxstrom@gmail.com>
Co-Authored-By: Siebrand Mazeland <siebrand@kitano.nl>
Change-Id: I35bbb3a7a145fc48d14fff620407dff5ecfdd4fc
Aaron Schulz [Tue, 1 Apr 2014 20:25:42 +0000 (13:25 -0700)]
Made BagOStuff fail fast in cas/lock on certain errors
* This avoids wasting time waiting on down servers
bug: 63058
Change-Id: Ia78a2036ffa7a20102ec9e8bf941cacb7fa63435
Alexandre Emsenhuber [Tue, 1 Apr 2014 12:52:33 +0000 (14:52 +0200)]
Send profiled items under the correct name
Fix for Ie8481a2e13 (
811a084). The "-total" item
was send under the name "close" since it's ended
by close().
Bug: 63347
Change-Id: Iad91f1d12303e5cfb92c111f79c0ce5a8187f53b
jenkins-bot [Tue, 1 Apr 2014 16:49:56 +0000 (16:49 +0000)]
Merge "Include parsed revision ID in parser cache"
Erik Bernhardson [Mon, 3 Mar 2014 21:26:54 +0000 (13:26 -0800)]
New Hook rc/watchlist hook ChangesListInitRows
An extension hooking into recent changes may need to load additional
data that did not fit into the recentchanges table. This hook gives
extensions an opportunity to see the full result prior to rendering
and batch load where approprite rather than loading piecemeal during
the render process.
This is implemented as an optional method called by the ChangesList
consumer, since the ChangesList never sees the full result set. Hook
implementers must be able to work regardless of the hook being called,
they just have the oportunity to be more efficient when it is called.
Change-Id: If74ae600ffba949364dd381dd3d466cbbaa27286
Brad Jorsch [Tue, 1 Apr 2014 14:52:24 +0000 (10:52 -0400)]
Include parsed revision ID in parser cache
One theory for what's behind bug 46014 is that the vandal submits the
edit, then someone (maybe the vandal) gets into the branch of
Article::view that uses PoolWorkArticleView, then ClueBot comes along
and reverts before the PoolWorkArticleView actually executes. Once that
PoolWorkArticleView actually does execute, it overwrites the parser
cache entry from ClueBot's revert with the one from the old edit.
To detect this sort of thing, let's include the revision id in the
parser cache entry and consider it expired if that doesn't match. Which
makes sense to do anyway.
And for good measure, let's have PoolWorkArticleView not save to the
parser cache if !$isCurrent.
Bug: 46014
Change-Id: Ifcc4d2f67f3b77f990eb2fa45417a25bd6c7b790
jenkins-bot [Tue, 1 Apr 2014 12:09:53 +0000 (12:09 +0000)]
Merge "i18n: add "Special:MyLanguage/""
jenkins-bot [Tue, 1 Apr 2014 12:04:45 +0000 (12:04 +0000)]
Merge "Fix order of expected, actual in some tests"
Niklas Laxström [Tue, 1 Apr 2014 11:59:46 +0000 (11:59 +0000)]
Fix order of expected, actual in some tests
Change-Id: Iabd431517586c06d411a0894e5d3621a4c4ddf05
shirayuki [Mon, 31 Mar 2014 10:42:39 +0000 (19:42 +0900)]
i18n: add "Special:MyLanguage/"
Change-Id: I791399e1a9bdff0bc672fc6ae69f4c2374690636
Niklas Laxström [Tue, 1 Apr 2014 08:03:05 +0000 (08:03 +0000)]
Add newline at the end of converted json files
1) Some text editors add missing newline at the end of the file
automatically.
2) It's generally a nice thing to do to avoid messing up display
for example when cat'ing a file
Change-Id: Icf636721c4edcf4ae706c9a346fe36233928cae8
Aaron Schulz [Mon, 31 Mar 2014 20:36:14 +0000 (13:36 -0700)]
Speed up LocalFile locking behavior
* Lowered the lock() timeout
* Wrap the destination DB write with memcached locks for move
ops since it is easier to control that lock timeout than the
DB lock timeouts.
* Cleanup locks in the destructor (e.g. on exceptions). Other
process can pile up more if locks have to expire.
bug: 63058
Change-Id: I38b28d81ec96daa80ece2354db284a614289ba0b
Antoine Musso [Tue, 1 Apr 2014 09:02:03 +0000 (11:02 +0200)]
Adjust debug log for caches being used
I always found out the debug message listing the cache being in use to
be slightly confusing:
MemcachedPeclBagOStuff[main] MemcachedPeclBagOStuff[message]
MemcachedPhpBagOStuff[parser]
Swap the class name and identifier to output:
main: MemcachedPeclBagOStuff, message: MemcachedPeclBagOStuff,
parser: MemcachedPhpBagOStuff
Change-Id: If069dec70eb31ed2dafe0e88ec4d4653005ff7e3
jenkins-bot [Tue, 1 Apr 2014 03:14:43 +0000 (03:14 +0000)]
Merge "Add a LESS test suite"
Sam Smith [Mon, 17 Mar 2014 13:24:52 +0000 (13:24 +0000)]
Add a LESS test suite
Add the LessFileCompilationTest test case class, which represents the
validation of a LESS file by compilation.
Add the LessTestSuite test suite, which tests all LESS files registered
with the ResourceLoader, and use this to rewrite the checkLess.php
maintenance script.
Bug: 54665
Change-Id: Iedb8dc31e4817d8b4e40b655cf9b8fb092979e90
jenkins-bot [Mon, 31 Mar 2014 22:24:59 +0000 (22:24 +0000)]
Merge "mediawiki.log: Fix unknown method "apply" error in IE9 and below"
Gilles Dubuc [Mon, 31 Mar 2014 13:57:05 +0000 (15:57 +0200)]
mediawiki.log: Fix unknown method "apply" error in IE9 and below
In IE < 10, the console.log function is implemented as a non-standard
host function (not inheriting the regular Function prototype) which
doesn't provide an "apply" method.
Change-Id: I6ecf933589ec430ff4d1748a5f3c511b692e3b9a
kaldari [Mon, 31 Mar 2014 20:50:03 +0000 (13:50 -0700)]
Adding debugging for bug 62241
In the event of a bad filename warning record the relavent variables
so that we can figure out what is causing the failure.
See also change I532f8ee7c7455e43862b0204a7fff23d6792412e
Bug: 62241
Change-Id: If5d8b9e7f34c0f9bd09684d7dada8fafd7f5f7c2
jenkins-bot [Mon, 31 Mar 2014 20:56:39 +0000 (20:56 +0000)]
Merge "MWException: Fix sitename appearing twice in <title>"
Translation updater bot [Mon, 31 Mar 2014 20:11:50 +0000 (20:11 +0000)]
Localisation updates from https://translatewiki.net.
Change-Id: I54a8e52634c1e36fadd7a84f236747765fc98684
Timo Tijhof [Fri, 14 Mar 2014 05:02:45 +0000 (06:02 +0100)]
MWException: Fix sitename appearing twice in <title>
The MWException::getPageTitle method is wrapping the page title
in the 'pagetitle' ($1 - sitename) message, but OutputPage does
this as well when outputing <title>.
OutputPage::prepareErrorPage takes the first parameter as page
title (unwrapped, appears in <h1>,) and the second optional one
is for <title>, which (if omitted, like before) results in
pagetitle being applied by OutputPage.
Now that we specify them separately, this also removes the
sitename from the <h1> on the page, which always looked weird.
It also fixes DBQueryError (extends MWException) where previously
no sitename was added, it just returns 'databaseerror' from
its DBQueryError::getPageTitle() method.
Change-Id: I85d65d89e72b4497a050f440524ff6f959518b88
jenkins-bot [Mon, 31 Mar 2014 18:23:06 +0000 (18:23 +0000)]
Merge "Adjust margin between h2 and special page table element"