lhc/web/wiklou.git
11 years agoMerge "Fixed use of __METHOD__ in a closure."
jenkins-bot [Thu, 11 Apr 2013 21:31:44 +0000 (21:31 +0000)]
Merge "Fixed use of __METHOD__ in a closure."

11 years agoMerge "phpcs: Fix Squiz.WhiteSpace.ScopeClosingBrace.ContentBefore violations"
jenkins-bot [Thu, 11 Apr 2013 21:20:58 +0000 (21:20 +0000)]
Merge "phpcs: Fix Squiz.WhiteSpace.ScopeClosingBrace.ContentBefore violations"

11 years agoFixed use of __METHOD__ in a closure.
Aaron Schulz [Thu, 11 Apr 2013 21:04:23 +0000 (14:04 -0700)]
Fixed use of __METHOD__ in a closure.

Change-Id: I41dc0ba4aff39d5a65cd1ccb3fba6c102974e6c2

11 years agoMerge "Localisation updates from http://translatewiki.net."
Translation updater bot [Thu, 11 Apr 2013 20:29:10 +0000 (20:29 +0000)]
Merge "Localisation updates from translatewiki.net."

11 years agoMerge "Disable Special:PasswordReset when $wgEnableEmail is false"
jenkins-bot [Thu, 11 Apr 2013 20:02:08 +0000 (20:02 +0000)]
Merge "Disable Special:PasswordReset when $wgEnableEmail is false"

11 years agoLocalisation updates from http://translatewiki.net.
Translation updater bot [Thu, 11 Apr 2013 19:56:07 +0000 (19:56 +0000)]
Localisation updates from translatewiki.net.

Change-Id: I6c19864b175b2cbb84273326a981467af2db8610

11 years agoDisable Special:PasswordReset when $wgEnableEmail is false
JuneHyeon Bae [Tue, 12 Mar 2013 06:10:17 +0000 (15:10 +0900)]
Disable Special:PasswordReset when $wgEnableEmail is false

Previously one could still navigate to the page by entering its URL directly,
and from there be able to request a password reset e-mail for any user who has
an e-mail address configured. This patch adds a check for $wgEnableEmail in the
SpecialPage code, replacing the interface with an error message when e-mail is
not enabled.

Change-Id: I42be68c6a068aafdf45623eb35bb21aa684f54f6

11 years agoMerge "Minor fix to redis queue docs."
jenkins-bot [Thu, 11 Apr 2013 19:50:21 +0000 (19:50 +0000)]
Merge "Minor fix to redis queue docs."

11 years agoMerge "Consistency tweak: Wrap variable names into <code></code>"
Yuvipanda [Thu, 11 Apr 2013 19:27:21 +0000 (19:27 +0000)]
Merge "Consistency tweak: Wrap variable names into <code></code>"

11 years agoMinor fix to redis queue docs.
Aaron Schulz [Thu, 11 Apr 2013 18:50:19 +0000 (11:50 -0700)]
Minor fix to redis queue docs.

Change-Id: I27394ddc8187ab3e9c275e2d72911318cc496604

11 years agoMerge "Added a quick test for getAllQueuedJobs()."
jenkins-bot [Thu, 11 Apr 2013 17:25:43 +0000 (17:25 +0000)]
Merge "Added a quick test for getAllQueuedJobs()."

11 years agoMerge "Remove externaledit/externaldiff preference"
jenkins-bot [Thu, 11 Apr 2013 16:15:27 +0000 (16:15 +0000)]
Merge "Remove externaledit/externaldiff preference"

11 years agoUpdate documentation and styling
Siebrand Mazeland [Mon, 8 Apr 2013 12:09:48 +0000 (14:09 +0200)]
Update documentation and styling

Remove some trailing whitespace, break long lines, make PHPDoc consistent
and update it in places.

Change-Id: I340b43a47e121a44ebd4aaa2a45dff12e945e645

11 years agoRemove externaledit/externaldiff preference
Chad Horohoe [Thu, 11 Apr 2013 13:20:36 +0000 (09:20 -0400)]
Remove externaledit/externaldiff preference

Change-Id: Ic3791d188688f556753ab5556a4d9733bbf2e85e

11 years agogiven name correction in CREDITS
Wikinaut [Thu, 11 Apr 2013 10:39:14 +0000 (12:39 +0200)]
given name correction in CREDITS

Change-Id: I1afdad0129980a988e5626351cff94c38d01688f

11 years agophpcs: Fix Squiz.WhiteSpace.ScopeClosingBrace.ContentBefore violations
Timo Tijhof [Thu, 11 Apr 2013 05:29:05 +0000 (07:29 +0200)]
phpcs: Fix Squiz.WhiteSpace.ScopeClosingBrace.ContentBefore violations

ERROR: Closing brace must be on a line by itself
Squiz.WhiteSpace.ScopeClosingBrace.ContentBefore

- For non-empty scopes it means the closing brace must be on a separate
  line. This is already the case in most classes in some cases the "lazy
  closing" is still used.
  array(
   'x' ); // Moved } to next line
  function () { return 'x'; } // Moved } to next line
  case 1:
    stuff; break; // Moved break to next line
- For empty function it serves as a visual distinction between there not
  being a scope block / function body and there being an empty function
  body.
  function foo(); // No body
  function foo() {} // Empty body - violation
  function bar() { // Empty body corrected
  }

Change-Id: I0310ec379c6d41cc7d483671994d027a49f32164

11 years agoAdd @since line for User::groupHasPermission()
Kevin Israel [Thu, 11 Apr 2013 02:18:43 +0000 (22:18 -0400)]
Add @since line for User::groupHasPermission()

Change-Id: Ibcd20a3921058a010334e0e0983e2d9541aea734
Follows-up: I41edb091fa35c8c68b6f95cc5fd208ea99418cdb

11 years agoMerge "Minor coding style whitespace cleanup"
jenkins-bot [Wed, 10 Apr 2013 22:35:11 +0000 (22:35 +0000)]
Merge "Minor coding style whitespace cleanup"

11 years agoMinor coding style whitespace cleanup
Amir E. Aharoni [Thu, 14 Mar 2013 22:23:30 +0000 (00:23 +0200)]
Minor coding style whitespace cleanup

Change-Id: Ic066eda3a1a2db51b2a953a81829ce7f92568125

11 years agoLocalisation updates from http://translatewiki.net.
Translation updater bot [Wed, 10 Apr 2013 20:02:19 +0000 (20:02 +0000)]
Localisation updates from translatewiki.net.

Change-Id: Id27f47305ed695264d422df9fa51d6492b838b88

11 years agoVector: Move 'editwarning' feature from Vector extension
rahul21 [Sun, 31 Mar 2013 21:55:52 +0000 (03:25 +0530)]
Vector: Move 'editwarning' feature from Vector extension

* Created a new Resource Loader module
* Created an entry for the RL module in the Resources.php
* Added messages to languages/MessagesEn.php and languages/MessagesQqq.php
* Called the new RL module from the EditPage.php
* Added required entries in the messages.inc, DefaultSettings.php

Corresponding change in the Vector extension: I164f17255bc927

Bug: 46514
Change-Id: I7bdbd09f4083ccb316156307400ccfacdfec79e1

11 years agoMerge "Use HTMLForm::setWrapperLegendMsg() in FormSpecialPage::getForm()"
jenkins-bot [Wed, 10 Apr 2013 17:13:09 +0000 (17:13 +0000)]
Merge "Use HTMLForm::setWrapperLegendMsg() in FormSpecialPage::getForm()"

11 years agoAdd release notes for Iea1f7842
Brad Jorsch [Wed, 10 Apr 2013 13:38:56 +0000 (09:38 -0400)]
Add release notes for Iea1f7842

Change Iea1f7842 (Add timezone support to Language::sprintfDate) did not
include an update for RELEASE-NOTES-1.22.

Change-Id: Icad1228b90b1e30130cc820faaba7c0c8f2254a2

11 years agoUse HTMLForm::setWrapperLegendMsg() in FormSpecialPage::getForm()
Alexandre Emsenhuber [Wed, 10 Apr 2013 10:04:48 +0000 (12:04 +0200)]
Use HTMLForm::setWrapperLegendMsg() in FormSpecialPage::getForm()

There is no need to call $this->msg() directly for the fieldset legend
as there is already a method in HTMLForm to do it.

Change-Id: I771520ea230301b51cece685f47e6a606ad860be

11 years agoMerge "Remove dependency on ORMTable from ORMRow"
jenkins-bot [Wed, 10 Apr 2013 09:09:20 +0000 (09:09 +0000)]
Merge "Remove dependency on ORMTable from ORMRow"

11 years agomake error message pretty
JuneHyeon Bae [Wed, 20 Mar 2013 05:14:49 +0000 (14:14 +0900)]
make error message pretty

Change-Id: I45d04691aca8a1e2c7bb94537da8e04d2bbf0481

11 years agoAdd nosummary, minor and watchthis to Special:MyPage and friends
umherirrender [Sat, 6 Apr 2013 17:42:04 +0000 (19:42 +0200)]
Add nosummary, minor and watchthis to Special:MyPage and friends

These are options usable with action=edit and allows better options
for the inital edit view from urls or the InputBox extension.

Change-Id: I696c47ddf9f29764bc8cc6136c618c56dbc2d726

11 years agoRemove dependency on ORMTable from ORMRow
jeroendedauw [Mon, 18 Feb 2013 19:43:03 +0000 (20:43 +0100)]
Remove dependency on ORMTable from ORMRow

IORMRow implementing objects take a IORMTable object in their constructor.
The later is a hard to construct service object while the former ideally
should just be a simple wrapper around a database row. This means IORMRow
objects are tightly coupled with IORMTable objects, which makes them
inflexible and makes various things such as testing logic contained in
them needlessly difficult.

This commit gets rid of this nonsense by allowing for construction of
ORMRow objects without providing an ORMTable. All methods dependent on
the table field have been deprecated. Most of these methods have a new
alternative in ORMTable. For instance, saving an ORMRow can now be done
by passing an instance of ORMRow to the updateRow method of an ORMTable
instance, rather then calling save on the ORMRow instance.

Backwards compatibility has been retained except for the fields passed
in the constructor no longer undergoing magical unserialization if it
looks like this is needed. I do not expect this will affect any existing
code though.

Change-Id: I86368821fc2cd0729df5342b8572eb470c0f77a0

11 years agostartup: Lower blacklist for Firefox from 4 to 3.
Timo Tijhof [Wed, 10 Apr 2013 07:37:19 +0000 (09:37 +0200)]
startup: Lower blacklist for Firefox from 4 to 3.

Though Firefox is 3.6 is both unsupported by Mozilla, by jQuery
and by MediaWiki JS, it is not known to have fatal errors in
javascript run time execution so moving from Grade B (blacklist)
to Grade X (unsupported but given Grade A source code at own risk).

This is done following a thread in English Wikipedia's Village
pump requesting to get the Grade A resources again.

Follows-up 02fe025e6364ebf6f.

Change-Id: Ia54dd738b3ce0995fab6aae7a2729cc7b1a0dcb1

11 years agoFix message cache expiry semantics
Tim Starling [Wed, 3 Apr 2013 10:54:34 +0000 (21:54 +1100)]
Fix message cache expiry semantics

* Use the stale message cache while the new one is being generated
* Revert I811755d4 (make message cache load failure fatal). This
  escalated several very plausible temporary site issues from barely
  noticeable to complete downtime -- for example, memcached being down
  on a site with only one memcached server.
* Remove $wgLocalMessageCacheSerialized, it's always been pointless
* Clarify a couple of comments.
* Increased lock wait timeout to 30s
* Make lock() fail immediately on memcached connection refused

Tests done:
* With local cache enabled: normal cold refill; refill local from
  global cache; use stale local cache during remote refill; use stale
  global cache during remote refill; cold cache wait for remote refill;
  saveToCaches() failure; memcached connection refused.
* With local cache disabled: saveToCaches() failure; cache disabled due
  to "error" status key; memcached connection refused.

Setting a 1-day expiry in memcached, with a ~10s CPU cost to replace, is
not the best idea since it inevitably leads to a cache stampede. Dealing
with the stampede by waiting for a lock is not ideal, even if it were
implemented properly, since it's not necessary to deliver perfectly
fresh message cache data to all clients.

This is especially obvious when you note that barring bugs, expiry and
regeneration always gives you back the exact same data, because we have
incremental updates (MessageCache::replace()). Keeping all clients
waiting for 10s just to give them the data they have already is pretty
pointless.

So, continue to serve the site from the stale message cache while the
new one is being generated.

One caveat: if local caching enabled, when the message cache becomes
stale, a sudden spike in network bandwidth may result due to the full
array (also typically stale) being fetched from the shared cache.

Bug: 43516
Change-Id: Ia145fd90da33956d8aac127634606aaecfaa176b

11 years agoApiFormatXml: remove broken xmldoublequote param
Kevin Israel [Tue, 9 Apr 2013 00:25:47 +0000 (20:25 -0400)]
ApiFormatXml: remove broken xmldoublequote param

r55641 was a refactoring of ApiFormatXml::recXmlPrint() that added
a $doublequote parameter, yet the parameter was not actually passed
along in recursive calls, rendering the formatter's xmldoublequote
parameter useless.

XPath support was the sole reason this parameter was added. It was
argued that because in the XPath 1.0 grammar, a Literal is:

    '"' [^"]* '"' | "'" [^']* "'"

it is not possible to match attribute values that contain both an
apostrophe and a double quote.

However, conforming XPath implementations include a concat() function
that can be used to work around this limitation. Furthermore, XPath 2.0
allows escaping a double-quoted " as "" (and a single-quoted ' as '').

Because there was no Bugzilla report (other than mine) filed in the
last 3-1/2 years, and working around the XPath flaw is possible, I
think it's safe to remove the parameter. The worst that could happen
is that a not-very-robust API client chokes on the "Unrecognized
parameter" warning.

Bug: 46626
Change-Id: I9a4d8c13b78ffd2634d03c8c8b4bbeff69f4bc08

11 years agoMerge "add htmlform-no and htmlform-yes messages for generic yes-no questions"
jenkins-bot [Tue, 9 Apr 2013 23:48:06 +0000 (23:48 +0000)]
Merge "add htmlform-no and htmlform-yes messages for generic yes-no questions"

11 years agoMerge "Added a script to copy jobs from one queue to another."
jenkins-bot [Tue, 9 Apr 2013 23:39:55 +0000 (23:39 +0000)]
Merge "Added a script to copy jobs from one queue to another."

11 years agoMerge "Honor transform limit for images with many old revisions"
jenkins-bot [Tue, 9 Apr 2013 23:30:42 +0000 (23:30 +0000)]
Merge "Honor transform limit for images with many old revisions"

11 years agoMerge "(bug 33454) Add timezone support to Language::sprintfDate"
jenkins-bot [Tue, 9 Apr 2013 23:23:06 +0000 (23:23 +0000)]
Merge "(bug 33454) Add timezone support to Language::sprintfDate"

11 years agoAdded a script to copy jobs from one queue to another.
Aaron Schulz [Tue, 9 Apr 2013 07:01:32 +0000 (00:01 -0700)]
Added a script to copy jobs from one queue to another.

Change-Id: I9173f21563089e094d67deaf203797d90aa84990

11 years agoMerge "ResourceLoader: Don't cache ob_get_level"
jenkins-bot [Tue, 9 Apr 2013 23:08:04 +0000 (23:08 +0000)]
Merge "ResourceLoader: Don't cache ob_get_level"

11 years agoHonor transform limit for images with many old revisions
Brad Jorsch [Wed, 3 Apr 2013 16:03:44 +0000 (12:03 -0400)]
Honor transform limit for images with many old revisions

When preparing Id0ec6a0a, for some reason I thought the thumbnailing
only applied to the current revision. Fix that oversight.

Change-Id: I2e8aecc76a1190bac353a2b9855bc6a19ec06dec

11 years ago(bug 33454) Add timezone support to Language::sprintfDate
Brad Jorsch [Fri, 22 Mar 2013 16:41:03 +0000 (12:41 -0400)]
(bug 33454) Add timezone support to Language::sprintfDate

Add an optional timezone parameter to Language::sprintfDate, add format
characters eIOPTZ, and correct crU.

While we're at it, remove backwards-compatability code for 'N' and then
merge the existing switch cases for cr and wNzWtLoU that are basically
identical, since all those cases need to be changed anyway.

Bug: 33454
Change-Id: Iea1f78428bc0d32d6395818311dbe4b94d776c42

11 years agoMerge "Fixes to Special:PagesWithProp"
jenkins-bot [Tue, 9 Apr 2013 22:26:08 +0000 (22:26 +0000)]
Merge "Fixes to Special:PagesWithProp"

11 years agoMerge "Add SVG version of user icon in Vector personal portlet"
jenkins-bot [Tue, 9 Apr 2013 22:22:09 +0000 (22:22 +0000)]
Merge "Add SVG version of user icon in Vector personal portlet"

11 years agoResourceLoader: Don't cache ob_get_level
Timo Tijhof [Tue, 9 Apr 2013 22:16:22 +0000 (00:16 +0200)]
ResourceLoader: Don't cache ob_get_level

Follows-up I0eba7fcc9ce. This was br

Bug: 46836
Change-Id: Ida0e48931781be77327073dc4ed3966949f4b082

11 years agoMerge "mediawiki.jqueryMsg: Implement HTML support and improve HTML escaping."
jenkins-bot [Tue, 9 Apr 2013 22:14:40 +0000 (22:14 +0000)]
Merge "mediawiki.jqueryMsg: Implement HTML support and improve HTML escaping."

11 years agoAdd SVG version of user icon in Vector personal portlet
Pau Giner [Fri, 22 Feb 2013 09:36:55 +0000 (10:36 +0100)]
Add SVG version of user icon in Vector personal portlet

An SVG version of the user icon is provided only for browsers supporting
SVG. To ensure browser compatibility, the SVG version is provided in a
two-layer background where the first layer is a transparent gradient.

The fact that browsers supporting CSS gradients are a strict subset of
those supporting SVG guarantees an appropriate fallback.
Embedding is used to avoid extra http requests.

Bug: 35341
Change-Id: I914da0649459744ccca9e1a78e9f48fe66e1a77f

11 years agoAdded a quick test for getAllQueuedJobs().
Aaron Schulz [Tue, 9 Apr 2013 21:32:20 +0000 (14:32 -0700)]
Added a quick test for getAllQueuedJobs().

Change-Id: Ia39461bc1e2356da1a694bf1898b8ca8e4200ff8

11 years agoMerge "Localisation updates from http://translatewiki.net."
Translation updater bot [Tue, 9 Apr 2013 20:40:33 +0000 (20:40 +0000)]
Merge "Localisation updates from translatewiki.net."

11 years agoMerge "Cleared ResourceLoader blob store after update.php finishes."
jenkins-bot [Tue, 9 Apr 2013 19:54:47 +0000 (19:54 +0000)]
Merge "Cleared ResourceLoader blob store after update.php finishes."

11 years agoLocalisation updates from http://translatewiki.net.
Translation updater bot [Tue, 9 Apr 2013 19:23:54 +0000 (19:23 +0000)]
Localisation updates from translatewiki.net.

Change-Id: I83961608dc6bc3c0cea008592300972c027b5de1

11 years agoBatch all StatCounter stats into one datagram in flush().
Aaron Schulz [Tue, 2 Apr 2013 17:22:17 +0000 (10:22 -0700)]
Batch all StatCounter stats into one datagram in flush().

Change-Id: I50a2cd24764e0a284c4a1138f8424ebf3990792e

11 years agoMerge "Make sure a job is logged as OK if the status is NULL for sanity."
jenkins-bot [Tue, 9 Apr 2013 18:24:43 +0000 (18:24 +0000)]
Merge "Make sure a job is logged as OK if the status is NULL for sanity."

11 years agoCleared ResourceLoader blob store after update.php finishes.
Tyler Anthony Romeo [Tue, 9 Apr 2013 18:14:57 +0000 (18:14 +0000)]
Cleared ResourceLoader blob store after update.php finishes.

Update.php is supposed to clear all localization caches. This
makes sure to clear the ResourceLoader's own message cache
as well.

Bug: 44524
Change-Id: I15ceb60f5e2e4d3ed9ba818fe6cd619f9adce5b2

11 years agoMerge "jsduck-gen: Add --version parameter"
jenkins-bot [Tue, 9 Apr 2013 09:44:41 +0000 (09:44 +0000)]
Merge "jsduck-gen: Add --version parameter"

11 years agoMerge "startup: Blacklist incompatible mobile browsers"
jenkins-bot [Tue, 9 Apr 2013 09:34:05 +0000 (09:34 +0000)]
Merge "startup: Blacklist incompatible mobile browsers"

11 years agojsduck-gen: Add --version parameter
Timo Tijhof [Sat, 6 Apr 2013 01:21:37 +0000 (03:21 +0200)]
jsduck-gen: Add --version parameter

Follows-up I1d5b626.

Change-Id: I7a48e549e6acd2bdb7be9f6f2c34ff0554bb791c

11 years agoMerge "Revert "Remove is_numeric check from Title::checkUserBlock""
Hashar [Tue, 9 Apr 2013 08:45:11 +0000 (08:45 +0000)]
Merge "Revert "Remove is_numeric check from Title::checkUserBlock""

11 years agoRevert "Remove is_numeric check from Title::checkUserBlock"
Hashar [Tue, 9 Apr 2013 08:20:12 +0000 (08:20 +0000)]
Revert "Remove is_numeric check from Title::checkUserBlock"

Breaks unit testing (see bug 47031). The commit managed to land in
the branch because of a bug in Jenkins (bug 46723).

This reverts commit 8cc0b601aa2db6db09ac0e4d70847293d75875aa

Change-Id: I4b3fadccaae9c35964a0c47d63b22c4f35148a24

11 years agoMerge "Remove is_numeric check from Title::checkUserBlock"
jenkins-bot [Tue, 9 Apr 2013 02:54:44 +0000 (02:54 +0000)]
Merge "Remove is_numeric check from Title::checkUserBlock"

11 years agoRemove is_numeric check from Title::checkUserBlock
Kevin Israel [Wed, 3 Apr 2013 21:44:00 +0000 (17:44 -0400)]
Remove is_numeric check from Title::checkUserBlock

This should allow the usernames of administrators such as "7"
to show correctly on permissions error pages.

I extracted the working code from UserBlockedError::__construct
into a separate method Block::getPermissionsError, called from
both places with context provided as an argument.

Bug: 46768
Change-Id: Ic3fa926a5a4c109faff35fffbccb60fb06ea4a18

11 years agoGet rid of preg_replace( '/.../e', ... )
Kevin Israel [Mon, 18 Mar 2013 01:09:02 +0000 (21:09 -0400)]
Get rid of preg_replace( '/.../e', ... )

This is deprecated as of PHP 5.5, and the remaining uses are quite
silly. Tim said I should remove his easter egg from Special:Version,
as it already was broken, and a new one can be added in a separate
commit.

Change-Id: I0f09f4efc7afe5933c8317462026a475530a5324

11 years agoAllow search backends to return statuses instead of result sets
Chad Horohoe [Wed, 3 Apr 2013 19:16:33 +0000 (15:16 -0400)]
Allow search backends to return statuses instead of result sets

This will allow a backend to propogate more useful errors up
the stack.

Bug: 43544
Change-Id: I54ab520762381e8f46d5ca17c0bad8820eacb964

11 years agostartup: Blacklist incompatible mobile browsers
Max Semenik [Fri, 22 Mar 2013 23:58:12 +0000 (03:58 +0400)]
startup: Blacklist incompatible mobile browsers

These user agents have been blacklisted on the mobile site for some time.

They have been re-reviewed alongside the A grade phones supported by
jQuery mobile [1] and jQuery [2]

[1] http://jquerymobile.com/gbs/
[2] http://jquery.com/browser-support/

Change-Id: If7743c476e95002174763514428c17dfcbe7da4c

11 years agoMerge "bug 42649 - PHP Fatal error: Call to a member function isLocal() on a non...
jenkins-bot [Mon, 8 Apr 2013 23:19:55 +0000 (23:19 +0000)]
Merge "bug 42649 - PHP Fatal error: Call to a member function isLocal() on a non-object in Title.php"

11 years agoRemove first letters that have an overlapping prefix.
Brian Wolff [Sun, 24 Mar 2013 03:09:43 +0000 (00:09 -0300)]
Remove first letters that have an overlapping prefix.

First letters are supposed to be primary collation elements.
However, we do not want expansions to be considered
as firstletters (aka thorn "þ" -> "th" which isn't
the same as any other first letter (since "t" !== "th" )
however if þ was a first letter, the word "the" and
even worse the word "too" would be sorted under it, which
is wrong.

Looking for feedback if this all sounds sane. I have tested
it, it got rid of the contractions while at the same time
not removing any letter it wasn't supposed to.

Once this is merged, we could get rid of all the
-<langcode> entries. The other firstLetter array
entries for tailorings could be merged into
generateCollationData.php too, since incorrect
things would get pruned automatically, which
would probably make the logic in Collation.php
simpler.

Bug: 43740
Change-Id: I4bd3d39ec2938a53e2c6728adc48ee6cf9778d74

11 years agobug 42649 - PHP Fatal error: Call to a member function isLocal() on a non-object...
Reedy [Mon, 8 Apr 2013 22:29:57 +0000 (23:29 +0100)]
bug 42649 - PHP Fatal error: Call to a member function isLocal() on a non-object in Title.php

Change-Id: Ie6027c818785b12f6be9c47caab77ffa4a587a79

11 years agoMerge "Fix doc and deprecated warning of Title::fixUrlQueryArgs"
jenkins-bot [Mon, 8 Apr 2013 20:55:15 +0000 (20:55 +0000)]
Merge "Fix doc and deprecated warning of Title::fixUrlQueryArgs"

11 years agoMerge "mw.loader: Fix regression that caused CSS load after scripts."
jenkins-bot [Mon, 8 Apr 2013 20:52:26 +0000 (20:52 +0000)]
Merge "mw.loader: Fix regression that caused CSS load after scripts."

11 years agoMerge "Localisation updates from http://translatewiki.net."
jenkins-bot [Mon, 8 Apr 2013 20:38:48 +0000 (20:38 +0000)]
Merge "Localisation updates from translatewiki.net."

11 years agomediawiki.jqueryMsg: Implement HTML support and improve HTML escaping.
Matthew Flaschen [Sun, 10 Mar 2013 08:23:43 +0000 (04:23 -0400)]
mediawiki.jqueryMsg: Implement HTML support and improve HTML escaping.

* Whitelisted elements.
* Whitelisted attributes allowed on all elements, and additional ones
  per element.
* Currently, only b (bold) and i (italics) and the common attributes are
  implemented, but more can be added.
* Treat error message as text
* Implement appendWithoutParsing as an extra safeguard to ensure
  jQuery never implicitly parses HTML.  Reuse this where a similar
  case was handled one-off.
* Rename some of the link functions and variables to be clearer about
  relationships.

Bug: 44525
Change-Id: Id8902af9568092d0318d0ac9d1ca6c9c195d36fe

11 years agoFix doc and deprecated warning of Title::fixUrlQueryArgs
umherirrender [Mon, 8 Apr 2013 18:30:06 +0000 (20:30 +0200)]
Fix doc and deprecated warning of Title::fixUrlQueryArgs

Break long line

Change-Id: Iaeef8ce87c3ffd4609651f7cc0e8340a0f4bebfc

11 years agomw.loader: Fix regression that caused CSS load after scripts.
Timo Tijhof [Fri, 29 Mar 2013 20:09:06 +0000 (21:09 +0100)]
mw.loader: Fix regression that caused CSS load after scripts.

Follows-up 705d50c which introduced cssText buffer (yielding
one tick of the event queue to reduce the number of forced
repaints).

However since that made CSS loading asynchronoous (be it only for
a split second) it caused some nasty side-effects.

This was also reflected in the unit tests that had to resort
to doing setTimeout from within the mw.loader implemented script
to assert the styles. This is now sane again: Scripts execute
after styles are inserted.

Bug: 46401
Change-Id: Ib54a3e78710b7f798c6730d3f540d3284001e2de

11 years agoRevert "mw.loader: Fix regression that caused CSS load after scripts."
Hashar [Mon, 8 Apr 2013 19:04:31 +0000 (19:04 +0000)]
Revert "mw.loader: Fix regression that caused CSS load after scripts."

Broke master tests since it causes a JSDuck issue. See bug 47018.

This reverts commit be5f73b595b8fef0697bf83e9ebd13e85756c099

Change-Id: I2d945b7ffb7cfbcc5024a1c3bdcf0c9506b0cf76

11 years agoLocalisation updates from http://translatewiki.net.
Translation updater bot [Mon, 8 Apr 2013 19:03:40 +0000 (19:03 +0000)]
Localisation updates from translatewiki.net.

Change-Id: I3f39cc877ae087d2833fbdce1778ac4487f7090f

11 years agoAdd missing wfProfileOut before throwing an exception
umherirrender [Sun, 7 Apr 2013 16:34:58 +0000 (18:34 +0200)]
Add missing wfProfileOut before throwing an exception

Change-Id: I1d830da0597f19efd0b2ae48642389975e736e23

11 years agoConsistency tweak: Wrap variable names into <code></code>
raymond [Mon, 8 Apr 2013 17:54:36 +0000 (19:54 +0200)]
Consistency tweak: Wrap variable names into <code></code>

Followup to Ia6efdf931b0b6be7556fc16c040aa7493845cdfc

Change-Id: I51eea3de9adb0be6695ef6ff18059d536dce801b

11 years agoAdd missing wfProfileOut to LocalisationCache
umherirrender [Mon, 8 Apr 2013 17:37:17 +0000 (19:37 +0200)]
Add missing wfProfileOut to LocalisationCache

Follow up Ib94a8c18c4e270a7bd46faa17eb27a22ff950f75

Moved some empty lines to match the places of other profile calls in the
file.

Change-Id: I6616b18027cdb9e70b43948aaff999a2e25b07e0

11 years agoMerge "LocalisationCache: add profiling, fix doxygen warnings"
Aaron Schulz [Mon, 8 Apr 2013 17:12:20 +0000 (17:12 +0000)]
Merge "LocalisationCache: add profiling, fix doxygen warnings"

11 years agoMerge "FormatJson: microoptimizations for UTF8_OK mode"
jenkins-bot [Mon, 8 Apr 2013 16:57:03 +0000 (16:57 +0000)]
Merge "FormatJson: microoptimizations for UTF8_OK mode"

11 years agoMerge "mw.loader: Fix regression that caused CSS load after scripts."
Trevor Parscal [Mon, 8 Apr 2013 16:52:49 +0000 (16:52 +0000)]
Merge "mw.loader: Fix regression that caused CSS load after scripts."

11 years agoMerge "Call ApiPageSet::getFinalParamDescription"
jenkins-bot [Mon, 8 Apr 2013 14:54:00 +0000 (14:54 +0000)]
Merge "Call ApiPageSet::getFinalParamDescription"

11 years agoMerge "Normalize as and or language names"
jenkins-bot [Mon, 8 Apr 2013 14:39:58 +0000 (14:39 +0000)]
Merge "Normalize as and or language names"

11 years agoMerge "Avoid fatal errors with uselang=s"
jenkins-bot [Mon, 8 Apr 2013 14:36:17 +0000 (14:36 +0000)]
Merge "Avoid fatal errors with uselang=s"

11 years agoFormatJson: microoptimizations for UTF8_OK mode
Kevin Israel [Fri, 5 Apr 2013 19:15:02 +0000 (15:15 -0400)]
FormatJson: microoptimizations for UTF8_OK mode

* Replace strtr with str_replace where faster.
* Use addcslashes to escape json_encode's output. Because no control
  characters are included, the only characters that have to be
  escaped are \ and ". (irrelevant for PHP 5.4+ installations)

Re-encoding a ~1.5 MB API response from the Chinese Wikipedia:
* PHP 5.3: 32% faster (from 347 ms to 239 ms)
* PHP 5.4: 70% faster (from 51 ms to 15 ms)
* HHVM: 42% faster (from 326 ms to 191 ms)

Change-Id: I7c9342682986d40a2f2436ac978390b6018a3521

11 years agoFix release notes for I5e6f7d48d6904.
Timo Tijhof [Sun, 7 Apr 2013 23:00:09 +0000 (01:00 +0200)]
Fix release notes for I5e6f7d48d6904.

Change-Id: Id442e6fcbdc410798579bc0bf1088ccede08e21b

11 years agoLocalisation updates from http://translatewiki.net.
Translation updater bot [Sun, 7 Apr 2013 19:26:28 +0000 (19:26 +0000)]
Localisation updates from translatewiki.net.

Change-Id: I182906ee394da739d6b8fb64bc7743c05b3b6ac6

11 years agoMerge "Update formatting and docs"
jenkins-bot [Sun, 7 Apr 2013 15:32:52 +0000 (15:32 +0000)]
Merge "Update formatting and docs"

11 years agoUpdate formatting and docs
Siebrand Mazeland [Mon, 1 Apr 2013 21:28:27 +0000 (23:28 +0200)]
Update formatting and docs

Also remove some unneeded else{} statements and reverse if/else if that
reduces indentation.

Change-Id: I1f7bd26119850fc944fbe84ae78bdd95cafc7bbb

11 years agoMerge "startup: Add unit tests and add Firefox < 4 to blacklist."
MaxSem [Sun, 7 Apr 2013 08:18:03 +0000 (08:18 +0000)]
Merge "startup: Add unit tests and add Firefox < 4 to blacklist."

11 years agoRemove the "Quickbar" feature and all things related to it
MatmaRex [Sat, 6 Apr 2013 15:13:45 +0000 (17:13 +0200)]
Remove the "Quickbar" feature and all things related to it

"Quickbar" was a feature of the Standard and CologneBlue skins that
allowed the sidebar to be displayed on left or right side of the page,
floated or fixed, and hidden on diff pages.

Standard was removed (Ia6d73c2d), and CologneBlue doesn't support this
anymore (bug 41246), so all things quickbar can now be safely removed.

* Removed user prefs option + interface
* Removed related messages
* Removed code for this in SkinLegacy
* Removed dead code in DifferenceEngine and Language

Change-Id: I5e6f7d48d6904a052a3a11547d3ebe6161463018

11 years agoLocalisation updates from http://translatewiki.net.
Translation updater bot [Sat, 6 Apr 2013 19:29:48 +0000 (19:29 +0000)]
Localisation updates from translatewiki.net.

Change-Id: Icab7dc96cc3e81918512ea4f4b55fea2bd6cf7d5

11 years agoMerge "remove references to removed skins from mediawiki.util.js"
jenkins-bot [Sat, 6 Apr 2013 15:35:35 +0000 (15:35 +0000)]
Merge "remove references to removed skins from mediawiki.util.js"

11 years agoremove references to removed skins from mediawiki.util.js
MatmaRex [Sat, 6 Apr 2013 15:09:56 +0000 (17:09 +0200)]
remove references to removed skins from mediawiki.util.js

Dead code removal. Followup to Ia6d73c2d.

Change-Id: Ic43462d6775af682626e43a1bdfab3f8077f0101

11 years agoMerge "Remove link to Special:ActiveUsers from Special:Statistics"
Alex Monk [Sat, 6 Apr 2013 13:10:03 +0000 (13:10 +0000)]
Merge "Remove link to Special:ActiveUsers from Special:Statistics"

11 years agoMerge "fixed php doc in ApiBase"
Alex Monk [Sat, 6 Apr 2013 12:26:16 +0000 (12:26 +0000)]
Merge "fixed php doc in ApiBase"

11 years agoRemove link to Special:ActiveUsers from Special:Statistics
Niklas Laxström [Wed, 3 Apr 2013 10:39:54 +0000 (10:39 +0000)]
Remove link to Special:ActiveUsers from Special:Statistics

Besides being broken, causes PHP warnings.

Bug: 46719
Change-Id: I23d9c9bdc9d38eb4256d0f7061e225cf86b576f8

11 years agoLocalisation updates from http://translatewiki.net.
Translation updater bot [Fri, 5 Apr 2013 21:44:25 +0000 (21:44 +0000)]
Localisation updates from translatewiki.net.

Change-Id: I222f66b12c0d6bdb03c7105b3d598304f9ffd61c

11 years agoLocalisationCache: add profiling, fix doxygen warnings
Max Semenik [Fri, 5 Apr 2013 20:07:16 +0000 (00:07 +0400)]
LocalisationCache: add profiling, fix doxygen warnings

Change-Id: Ib94a8c18c4e270a7bd46faa17eb27a22ff950f75

11 years agoMerge "Protected function UploadBase->validateName changed to public"
jenkins-bot [Fri, 5 Apr 2013 21:11:08 +0000 (21:11 +0000)]
Merge "Protected function UploadBase->validateName changed to public"

11 years agoProtected function UploadBase->validateName changed to public
rahul21 [Fri, 29 Mar 2013 10:38:32 +0000 (16:08 +0530)]
Protected function UploadBase->validateName changed to public

*Any class that is not part of the UploadBase class
(so it can't call protected functions or members on an UploadBase object).
If you, for example,build an extension that uses an UploadBase instance,
there is no way to get the title validation error.

Bug: 38221

Change-Id: Ie6c3f95a09ef84777313c3ec2edd31c50c397e28

11 years agoMerge "Fixed fatal when using fake load balancers."
jenkins-bot [Fri, 5 Apr 2013 20:29:29 +0000 (20:29 +0000)]
Merge "Fixed fatal when using fake load balancers."

11 years agostartup: Add unit tests and add Firefox < 4 to blacklist.
Timo Tijhof [Fri, 5 Apr 2013 00:35:37 +0000 (02:35 +0200)]
startup: Add unit tests and add Firefox < 4 to blacklist.

Bug: 35906
Change-Id: Id8214c6a54f1bed2f9e1abb668248c46dcf7317e