lhc/web/wiklou.git
13 years agoDocument from r87129 summary
Platonides [Wed, 15 Jun 2011 21:23:47 +0000 (21:23 +0000)]
Document from r87129 summary

13 years agor87164 removed mDataLoaded and broke this test databaseless nature.
Platonides [Wed, 15 Jun 2011 21:22:08 +0000 (21:22 +0000)]
r87164 removed mDataLoaded and broke this test databaseless nature.

13 years agoMake sure to die if there are more MediaWikiLangTestCase::tearDowns than setUp()s...
Platonides [Wed, 15 Jun 2011 21:07:12 +0000 (21:07 +0000)]
Make sure to die if there are more MediaWikiLangTestCase::tearDowns than setUp()s (such as r90150 & r90155)

13 years agoFollow-up r90150, another parent::setUp missing.
Platonides [Wed, 15 Jun 2011 21:03:08 +0000 (21:03 +0000)]
Follow-up r90150, another parent::setUp missing.

13 years agoFollowup r90105, ReflectionMethod::setAccessible() requires PHP 5.3.2. Mark incomplet...
Chad Horohoe [Wed, 15 Jun 2011 20:57:13 +0000 (20:57 +0000)]
Followup r90105, ReflectionMethod::setAccessible() requires PHP 5.3.2. Mark incomplete otherwise

13 years agoFollowup to r90090 -- thou shalt add parent:: calls to both overridden tearDown ...
Brion Vibber [Wed, 15 Jun 2011 20:51:17 +0000 (20:51 +0000)]
Followup to r90090 -- thou shalt add parent:: calls to both overridden tearDown *and* overridden setUp ()

Had been resulting in fatal errors when running TimeAdjustTest, due to trying to use an old $wgContLang value that had never gotten saved properly because the parent class setUp() was not called.

13 years agoFix for BlockTest -- when setup ran multiple times, the block entry wouldn't get...
Brion Vibber [Wed, 15 Jun 2011 20:43:24 +0000 (20:43 +0000)]
Fix for BlockTest -- when setup ran multiple times, the block entry wouldn't get re-saved and we'd end up with null or 0 as our block id.
By first removing any matching block, we can re-insert it at will and have a fresh working block to test.

13 years agoAnother partial fix for Block test: update $this->mId on successful Block::insert()
Brion Vibber [Wed, 15 Jun 2011 20:42:16 +0000 (20:42 +0000)]
Another partial fix for Block test: update $this->mId on successful Block::insert()

13 years agoLocalisation updates for core and extension messages from translatewiki.net (2011...
Raimond Spekking [Wed, 15 Jun 2011 20:04:37 +0000 (20:04 +0000)]
Localisation updates for core and extension messages from translatewiki.net (2011-06-15 19:51:00 UTC)

13 years agoFix for partial regression in r84534: Block::newFromId was no longer handling the...
Brion Vibber [Wed, 15 Jun 2011 19:05:25 +0000 (19:05 +0000)]
Fix for partial regression in r84534: Block::newFromId was no longer handling the case where there was no matching row correctly. Now returns null.

This bug was showing up in test results like this:

  BlockTest::testInitializerFunctionsReturnCorrectBlock
  Trying to get property of non-object

  /home/ci/cruisecontrol-bin-2.8.3/projects/mw/source/includes/Block.php:340
  /home/ci/cruisecontrol-bin-2.8.3/projects/mw/source/includes/Block.php:365
  /home/ci/cruisecontrol-bin-2.8.3/projects/mw/source/includes/Block.php:118
  /home/ci/cruisecontrol-bin-2.8.3/projects/mw/source/tests/phpunit/includes/BlockTest.php:60
  /home/ci/cruisecontrol-bin-2.8.3/projects/mw/source/tests/phpunit/MediaWikiTestCase.php:60
  /home/ci/cruisecontrol-bin-2.8.3/projects/mw/source/tests/phpunit/MediaWikiPHPUnitCommand.php:20
  /home/ci/cruisecontrol-bin-2.8.3/projects/mw/source/tests/phpunit/phpunit.php:60

However, that only triggers because the actual test is failing -- it's expecting to get a return back. This can only be reproduce when using the suite.xml configuration file, and not when running the BlockTest standalone.

13 years agoAlso check indexes in SQLite upgrade tests
Max Semenik [Wed, 15 Jun 2011 17:59:02 +0000 (17:59 +0000)]
Also check indexes in SQLite upgrade tests

13 years agoSQLite support for r86485
Max Semenik [Wed, 15 Jun 2011 17:57:00 +0000 (17:57 +0000)]
SQLite support for r86485

13 years agoBroke some long lines
Aaron Schulz [Wed, 15 Jun 2011 17:03:18 +0000 (17:03 +0000)]
Broke some long lines

13 years agoPer Brion, follow-up r86784: doucment that wfMsg() and related also accept message...
Alexandre Emsenhuber [Wed, 15 Jun 2011 16:33:13 +0000 (16:33 +0000)]
Per Brion, follow-up r86784: doucment that wfMsg() and related also accept message parameters as an array in the second parameter

13 years ago* (bug 28798) Set $wgServer in the default LocalSettings.php
Tim Starling [Wed, 15 Jun 2011 07:35:47 +0000 (07:35 +0000)]
* (bug 28798) Set $wgServer in the default LocalSettings.php
* (bug 14977) When detecting $wgServer, treat IPv6 addresses in $_SERVER['SERVER_NAME'] etc. in a sensible way.
* Tests for the new functions in IP.php and Installer.php

13 years agoUse camelCase, as pointed out by siebrand on r89731
Chad Horohoe [Wed, 15 Jun 2011 01:34:23 +0000 (01:34 +0000)]
Use camelCase, as pointed out by siebrand on r89731

13 years agoUsing true instead of calling isMainPage() again.
Krinkle [Wed, 15 Jun 2011 01:11:16 +0000 (01:11 +0000)]
Using true instead of calling isMainPage() again.

(Follow-up r90099 CR, Poke r87212)

13 years agoRevert r86764, r89134, r86827 -- added a second opaque boolean parameter to Linker...
Brion Vibber [Tue, 14 Jun 2011 23:50:40 +0000 (23:50 +0000)]
Revert r86764, r89134, r86827 -- added a second opaque boolean parameter to Linker::commentBlock() which appeared to mostly just turn it into Linker::formatComment().

commentBlock() exists for the sole purpose of embedding a comment into parentheses if it exists so you can append it to a line of text -- if you're not putting stuff in parentheses, don't use commentBlock() because you're not generating a parenthesized comment block.
Opaque boolean parameters are also very poor form, especially when tacking on multiple ones. There was already a nasty optional '$local' boolean param, forcing all uses of this other parameter to add *two* parameters, making illegible stuff like 'false, false'.

13 years agoOnly define 'isMainPage' if we are on the main page, the new mw.config library is...
Krinkle [Tue, 14 Jun 2011 23:29:45 +0000 (23:29 +0000)]
Only define 'isMainPage' if we are on the main page, the new mw.config library is specifically built for this efficient way of exporting variables as it will prevent any TypeError exceptions and simply return null (it's not a real global anymore).

This is a follow-up on Brion's CR at r87212.

13 years agoMerge conflict fix from r90096.
Krinkle [Tue, 14 Jun 2011 22:51:47 +0000 (22:51 +0000)]
Merge conflict fix from r90096.

13 years agowgNamespaceIds in JavaScript didn't include canonical namespaces. Adding them to...
Krinkle [Tue, 14 Jun 2011 22:49:38 +0000 (22:49 +0000)]
wgNamespaceIds in JavaScript didn't include canonical namespaces. Adding them to it, in a similar way that Language->getNamespaceIds does it for the localized namespaces and the namespace aliases.

Fixes bug in mw.Title constructor when .setNamespace() is used with a canonical namespace on a non-English content-language wiki.
Example: On a German wiki "var foo = new mw.Title('bar').setNamespace('file')" will throw an Error, as wgNamespaceIds only contains localized namespaces + namespace aliases, not canonical ones (in contrary to the assumption that has been made in various places).

(bug 25375) Add canonical namespaces to JavaScript "wgNamespaceIds"

13 years agoFollowup to r86347, r86797: change SpecialUndelete->mFile to mFilename to remove...
Brion Vibber [Tue, 14 Jun 2011 21:47:23 +0000 (21:47 +0000)]
Followup to r86347, r86797: change SpecialUndelete->mFile to mFilename to remove conflict with SpecialPage->mFile which carries different data and is now private.

13 years agofix for bug29371 . regex wordwrap with UTF8: do not use \b metacharacter. The problem...
Thomas Gries [Tue, 14 Jun 2011 21:42:25 +0000 (21:42 +0000)]
fix for bug29371 . regex wordwrap with UTF8: do not use \b metacharacter. The problem is that JavaScript recognizes word boundaries only before/after ASCII letters (and numbers/underscore)

13 years agoBlockTest changes $wgContLang, too.
Platonides [Tue, 14 Jun 2011 21:27:07 +0000 (21:27 +0000)]
BlockTest changes $wgContLang, too.

13 years agoThou shall call parent::tearDown(), MediaWikiLangTestCase child.
Platonides [Tue, 14 Jun 2011 21:21:26 +0000 (21:21 +0000)]
Thou shall call parent::tearDown(), MediaWikiLangTestCase child.
Follow up r87210. This has been hurting my tests for weeks.

Made TimeAdjustTest extend MediaWikiLangTestCase.

13 years agomakeCollapsible fix for jQuery 1.6.1
Krinkle [Tue, 14 Jun 2011 21:20:30 +0000 (21:20 +0000)]
makeCollapsible fix for jQuery 1.6.1
* Pre 1.6, jQuery returned an empty string for attributes that were not set. Although in a way that was wrong, it was the way it was. From jQuery documentation: "As of jQuery 1.6, the .attr() method returns <code>undefined</code> for attributes that have not been set." [1]
** Fixing makeCollapsible by removing empty-string checks and casting to boolean instead
* Wrapped a long line

[1] http://api.jquery.com/attr/

(jQuery 1.6.1 upgrade was in r89866)

13 years agoFix one bad doc comment entry from r86848
Brion Vibber [Tue, 14 Jun 2011 21:10:33 +0000 (21:10 +0000)]
Fix one bad doc comment entry from r86848

13 years agoRevert r86897 in wfMakeUrlIndex() and solve the issue in wfParseUrl()
Platonides [Tue, 14 Jun 2011 20:57:42 +0000 (20:57 +0000)]
Revert r86897 in wfMakeUrlIndex() and solve the issue in wfParseUrl()

13 years agoFollowup to r86942 per CR comments: instanceof Array -> $.isArray (safer, more cross...
Brion Vibber [Tue, 14 Jun 2011 20:55:39 +0000 (20:55 +0000)]
Followup to r86942 per CR comments: instanceof Array -> $.isArray (safer, more cross-browser)

13 years agoLocalisation updates for core and extension messages from translatewiki.net (2011...
Raimond Spekking [Tue, 14 Jun 2011 20:52:22 +0000 (20:52 +0000)]
Localisation updates for core and extension messages from translatewiki.net (2011-06-14 20:30:00 UTC)

13 years agoIncreasing time-out (testswarm reports false positives from swarmiers with a slower...
Krinkle [Tue, 14 Jun 2011 20:02:51 +0000 (20:02 +0000)]
Increasing time-out (testswarm reports false positives from swarmiers with a slower connection). Follow-up r89247

13 years agoRemove redundant mediawiki.util directory
Krinkle [Tue, 14 Jun 2011 17:44:02 +0000 (17:44 +0000)]
Remove redundant mediawiki.util directory

13 years agoPer Nikerabbit, fixes for r89176:
Alexandre Emsenhuber [Tue, 14 Jun 2011 15:08:52 +0000 (15:08 +0000)]
Per Nikerabbit, fixes for r89176:
* correct variable name in preprocessorFuzzTest.php
* Added type hints in Parser's test methods

13 years agoFU r89545: change broke api watching by switching watch and unwatch around.
Niklas Laxström [Tue, 14 Jun 2011 13:10:26 +0000 (13:10 +0000)]
FU r89545: change broke api watching by switching watch and unwatch around.
Also added zero to the article constructor..how many subtle bugs we need to make default article constructor sane?

13 years agoRevert r89925, setting disable_functions in php.ini is definitely the wrong way to...
Tim Starling [Tue, 14 Jun 2011 06:07:20 +0000 (06:07 +0000)]
Revert r89925, setting disable_functions in php.ini is definitely the wrong way to fix hash mismatches.

13 years agoMaintenance script for exporting the preprocessed wikitext from installer document...
Tim Starling [Tue, 14 Jun 2011 03:09:49 +0000 (03:09 +0000)]
Maintenance script for exporting the preprocessed wikitext from installer document pages, plus relevant refactoring. For use in updating mediawiki.org/wiki/Release_notes/1.17 etc.

13 years ago* Remove @static from the last few places it's left in core. Please don't use this...
Chad Horohoe [Tue, 14 Jun 2011 01:56:55 +0000 (01:56 +0000)]
* Remove @static from the last few places it's left in core. Please don't use this silly annotation anymore, that's what the static keyword is for.
* Enforce protected on SpecialAllPages::getNamespaceKeyAndText(), it's not @private like the comment implies
* Drop User::getMaxID(), nothing uses it

13 years agoFixing two broken tests in jquery.autoEllipsis.js for IE6 and Opera:
Krinkle [Tue, 14 Jun 2011 01:38:26 +0000 (01:38 +0000)]
Fixing two broken tests in jquery.autoEllipsis.js for IE6 and Opera:
* Mysterious test failure in IE6 in jquery.autoEllipsis.js finally figured. Way simpler than we thought. In IE6 width behaves like min-width, so adding one or two characters to a <span> with white-space:nowrap inside a <div> with a  fixed width will NOT (as other browsers do) make the span-width wider than the div-width, instead the div-width will happily increase as well.
* 2nd breakage fixed: In some cases adding 1 character did not widen the span, probably because of some edge case with narrow fonts and characters like "i" which are so small that they're just on the edge. Changed the test to add 2 characters instead of 1 to overcome this problem.

This new version of the test suite was already pushed to TestSwarm from my own account (instead of the MediaWiki-SVN account) and confirmed to work in IE6. Yay!

Not-pretty: Browser-sniffing for IE6 in a unit test..

13 years agoConvert a couple of wfDie()s to exceptions
Chad Horohoe [Mon, 13 Jun 2011 23:43:01 +0000 (23:43 +0000)]
Convert a couple of wfDie()s to exceptions

13 years agoRewrite compressOld to subclass maintenanance
Chad Horohoe [Mon, 13 Jun 2011 23:28:59 +0000 (23:28 +0000)]
Rewrite compressOld to subclass maintenanance
* Dropped compressOld.inc
* Moved LS_INDIVIDUAL, LS_CHUNKED to class constants
* Drop unused and undocumented 'h' parameter
* Added long forms for all the short parameters
* All of this so I could get rid of the wfDie()

13 years agoStart on test cases for bug 6672 (Exif orientation support), follow up to r79845.
Brion Vibber [Mon, 13 Jun 2011 22:13:38 +0000 (22:13 +0000)]
Start on test cases for bug 6672 (Exif orientation support), follow up to r79845.

Unfortunately seem to be hitting a wall right off at checking width & height; BitmapHandler reports back the pre-rotation width & height, leading to the oddity that a file that's clearly 1024px tall and 768px wide reports itself as 1024px wide and 768px tall.
Haven't gotten to generating rotated output to confirm file size. :)

Marked test case as broken pending fix so it doesn't add to surprise regressions in test reports.

13 years agoRm some more completely incomplete tests. If you're not going to add the test, don...
Chad Horohoe [Mon, 13 Jun 2011 21:55:22 +0000 (21:55 +0000)]
Rm some more completely incomplete tests. If you're not going to add the test, don't add to suite execution time by adding empty tests like this

13 years agoDrop suite full of incomplete tests. Tired of it cluttering my runs
Chad Horohoe [Mon, 13 Jun 2011 21:49:20 +0000 (21:49 +0000)]
Drop suite full of incomplete tests. Tired of it cluttering my runs

13 years agoFollowup r89562, use 'feedformat' not 'feed'
Sam Reed [Mon, 13 Jun 2011 21:35:23 +0000 (21:35 +0000)]
Followup r89562, use 'feedformat' not 'feed'

13 years ago(no commit message)
Krinkle [Mon, 13 Jun 2011 20:17:39 +0000 (20:17 +0000)]

13 years agoavoid test being marked as incomplete due to lack of assertions
Max Semenik [Mon, 13 Jun 2011 19:22:52 +0000 (19:22 +0000)]
avoid test being marked as incomplete due to lack of assertions

13 years ago(bug 28556) Remove MacBinary
Krinkle [Mon, 13 Jun 2011 19:00:49 +0000 (19:00 +0000)]
(bug 28556) Remove MacBinary
* IE5 for Mac is history.

13 years agoAdded: ActionBeforeFormDisplay, ActionModifyFormFields, ExtensionTypes, MaintenanceRe...
Alexandre Emsenhuber [Mon, 13 Jun 2011 14:55:03 +0000 (14:55 +0000)]
Added: ActionBeforeFormDisplay, ActionModifyFormFields, ExtensionTypes, MaintenanceRefreshLinksInit and SpecialPasswordResetOnSubmit
Removed: AfterUserMessage, FormatUserMessage, MathAfterTexvc, SetupUserMessageArticle and UserLoginMailPassword

13 years agoresinserted context.data.prevText = context.data..val(); which was accidentially...
Thomas Gries [Mon, 13 Jun 2011 11:32:04 +0000 (11:32 +0000)]
resinserted context.data.prevText = context.data..val(); which was accidentially dropped

13 years agofollow up patch to r89950 r89952 r89953 r89955 : moved div clearing inside call back...
Thomas Gries [Mon, 13 Jun 2011 11:16:31 +0000 (11:16 +0000)]
follow up patch to r89950 r89952 r89953 r89955 : moved div clearing inside call back function maybeFetch so that this is also done when coming from timeouts

13 years agoessage=corrected my mistake when comparing if textbox length is empty, or not.
Thomas Gries [Mon, 13 Jun 2011 09:56:54 +0000 (09:56 +0000)]
essage=corrected my mistake when comparing if textbox length is empty, or not.

13 years agocorrected comment
Thomas Gries [Mon, 13 Jun 2011 09:43:19 +0000 (09:43 +0000)]
corrected comment

13 years agocorrected comment
Thomas Gries [Mon, 13 Jun 2011 09:41:02 +0000 (09:41 +0000)]
corrected comment

13 years agofix for bug29368 which may have consequences. Please code reviewers, pls. check for...
Thomas Gries [Mon, 13 Jun 2011 09:38:06 +0000 (09:38 +0000)]
fix for bug29368 which may have consequences. Please code reviewers, pls. check for potential side effects - I cannot detect any, currently.

13 years ago* Fix I18nTags #languagename to use ParserOptions::getUserLang()
Niklas Laxström [Mon, 13 Jun 2011 07:47:39 +0000 (07:47 +0000)]
* Fix I18nTags #languagename to use ParserOptions::getUserLang()
* Fix ParserOptions::getUserLang() to return code as documented

13 years agoReplacing strictEqual-assertion with a more useful QUnit.push that includes the expec...
Krinkle [Mon, 13 Jun 2011 00:47:15 +0000 (00:47 +0000)]
Replacing strictEqual-assertion with a more useful QUnit.push that includes the expected and actual values in the html output. This way debugging problems from the TestSwarm will be easier.

This way we end up with:
"Expected: > 100, Result: 99"
instead of:
"Expected: true, Result: false"
which could virtually mean anything (even null, undefined, string, array, you name it)

* Also fixed issue in jquery.colorUtil.js

13 years agoLocalisation updates for core and extension messages from translatewiki.net (2011...
Raimond Spekking [Sun, 12 Jun 2011 21:05:54 +0000 (21:05 +0000)]
Localisation updates for core and extension messages from translatewiki.net (2011-06-12 22:40:00 UTC)

13 years agoChanged some magic __get method calls to accessors, were throwing notices
Alexandre Emsenhuber [Sun, 12 Jun 2011 19:09:14 +0000 (19:09 +0000)]
Changed some magic __get method calls to accessors, were throwing notices

13 years agoMoved action=deletetrackback and action=markpatrolled to Action class.
Alexandre Emsenhuber [Sun, 12 Jun 2011 18:46:07 +0000 (18:46 +0000)]
Moved action=deletetrackback and action=markpatrolled to Action class.
Also fixed order in DefaultSettings.php

13 years agosafer check if mhash is available (and I'm about to make it unavailable by disabling...
Ariel Glenn [Sun, 12 Jun 2011 12:02:53 +0000 (12:02 +0000)]
safer check if mhash is available (and I'm about to make it unavailable by disabling the fn in the php.ini for lucid builds)

13 years agoUpdate Chinese conversion tables.
Philip Tzou [Sun, 12 Jun 2011 08:02:04 +0000 (08:02 +0000)]
Update Chinese conversion tables.

13 years agoRemove resize/arrow cursors from jquery.makeCollapsible
Krinkle [Sun, 12 Jun 2011 00:17:02 +0000 (00:17 +0000)]
Remove resize/arrow cursors from jquery.makeCollapsible
* In Safari they look nice (a visual clue which direction we're heading on-click) but on other browsers there is only a general arrow-icon that points both up and down, and other browsers have an even weirder cursor for this. Summarized: It's inconsistent and redundant with the text-label we already have.

(Poke r79083 CR)

13 years agoRemove $wgCheckCopyrightUpload from DefaultSettings.php since the feature was removed...
Brian Wolff [Sat, 11 Jun 2011 21:10:26 +0000 (21:10 +0000)]
Remove $wgCheckCopyrightUpload from DefaultSettings.php since the feature was removed about 5 years ago

Looks like the feature was partially removed back in r10168. In any case, its long gone out of trunk now.

13 years agosqlite.php: don't fail after --check-syntax on a non-SQLite DB
Max Semenik [Sat, 11 Jun 2011 18:45:56 +0000 (18:45 +0000)]
sqlite.php: don't fail after --check-syntax on a non-SQLite DB

13 years agoFix jquery.tabIndex even more.
Krinkle [Sat, 11 Jun 2011 10:56:54 +0000 (10:56 +0000)]
Fix jquery.tabIndex even more.
* Discovered through TestSwarm. Aside from IE6/IE7, which already needed extra checking, when IE6/IE7 is on Windows NT 5.2, it can also return NaN.

13 years agomediawiki.util.test.js: IE7 doesn't expose the CSSStyleSheet publicly (like it doesn...
Krinkle [Sat, 11 Jun 2011 10:37:39 +0000 (10:37 +0000)]
mediawiki.util.test.js: IE7 doesn't expose the CSSStyleSheet publicly (like it doesn't for many). Checking object type instead.

13 years agoAnd I forgot to commit this file in r89874
Alexandre Emsenhuber [Sat, 11 Jun 2011 08:40:35 +0000 (08:40 +0000)]
And I forgot to commit this file in r89874

13 years ago* Made Special:RevisionDelete use the local variables stored in the RequestContext...
Alexandre Emsenhuber [Sat, 11 Jun 2011 08:30:19 +0000 (08:30 +0000)]
* Made Special:RevisionDelete use the local variables stored in the RequestContext object instead of global ones and removed the $skin member
* Also updated classes in includes/revisiondelete that were using the $skin member of Special:RevisionDelete
* Handle action=revisiondelete (or rather action=historysubmit&revisiondelete=1) with an Action object that sets the RequestContext object of the SpecialPage object, and in this case get the target Title object through the context instead of $wgTitle

13 years agoRemoved usage of error suppression operator, also removed function_exists() check...
Alexandre Emsenhuber [Sat, 11 Jun 2011 08:03:59 +0000 (08:03 +0000)]
Removed usage of error suppression operator, also removed function_exists() check for timezone_open() since we now require PHP 5.2.3

13 years agoMoved part of action handling functions near other ones
Alexandre Emsenhuber [Sat, 11 Jun 2011 07:43:40 +0000 (07:43 +0000)]
Moved part of action handling functions near other ones

13 years agoFollowup r88873, remove setting of $wgEnableCreativeCommonsRdf in LocalSettingsGenerator
Sam Reed [Sat, 11 Jun 2011 07:33:35 +0000 (07:33 +0000)]
Followup r88873, remove setting of $wgEnableCreativeCommonsRdf in LocalSettingsGenerator

13 years agoChanged more magic __get method calls to accessors, were throwing notices
Alexandre Emsenhuber [Sat, 11 Jun 2011 07:02:34 +0000 (07:02 +0000)]
Changed more magic __get method calls to accessors, were throwing notices

13 years agoRELEASE-NOTES for r89866
Krinkle [Sat, 11 Jun 2011 00:28:20 +0000 (00:28 +0000)]
RELEASE-NOTES for r89866

13 years agoUpgrade jQuery from 1.4.4 to 1.6.1
Krinkle [Sat, 11 Jun 2011 00:25:55 +0000 (00:25 +0000)]
Upgrade jQuery from 1.4.4 to 1.6.1

Poke r88725, r88607, bug 28904;

13 years agoFollow-up r89853: Documentation for completenesstest + minor fixes/whitespace
Krinkle [Sat, 11 Jun 2011 00:24:36 +0000 (00:24 +0000)]
Follow-up r89853: Documentation for completenesstest + minor fixes/whitespace

13 years agoAdd missing language name
Robin Pepermans [Sat, 11 Jun 2011 00:02:22 +0000 (00:02 +0000)]
Add missing language name

13 years agoFollow-up to r75241 / 75241: rename LanguageBat_smg to LanguageSgs
Robin Pepermans [Fri, 10 Jun 2011 22:33:49 +0000 (22:33 +0000)]
Follow-up to r75241 / 75241: rename LanguageBat_smg to LanguageSgs

13 years agoFix for r89821, r89839: we can skip certain tests if the web user is the same as...
Tim Starling [Fri, 10 Jun 2011 22:33:16 +0000 (22:33 +0000)]
Fix for r89821, r89839: we can skip certain tests if the web user is the same as the install user.

13 years agoAdded jquery.qunit.completenessTest.js (A jQuery/QUnit test coverage utility)
Krinkle [Fri, 10 Jun 2011 22:15:14 +0000 (22:15 +0000)]
Added jquery.qunit.completenessTest.js (A jQuery/QUnit test coverage utility)
* Added to /resources
* Conditionally loaded (condition being that the url parameter "completenesstest" has a truthy value)
* Fixed a test that was using === and true
* Setting an added method somewhere back to undefined so it won't be listed as a potential "missing test".

13 years agoFollow-up r89835: Accidently comitted from a deeper dir than the diff. Comitting...
Krinkle [Fri, 10 Jun 2011 20:16:42 +0000 (20:16 +0000)]
Follow-up r89835: Accidently comitted from a deeper dir than the diff. Comitting the rest now

13 years agoFollow-up: r89845: Forgot to commit changes in the other directory
Krinkle [Fri, 10 Jun 2011 20:08:19 +0000 (20:08 +0000)]
Follow-up: r89845: Forgot to commit changes in the other directory

13 years agoReview of unit test suites
Krinkle [Fri, 10 Jun 2011 19:55:40 +0000 (19:55 +0000)]
Review of unit test suites
* There's more than ok(), equal() and deepEqual(). Using the others as appropriate.
** Considered: strictEqual(a,b) tests the same as ok(a===b). But the latter doesn't include the values in the report when it fails. So strictEqual saves a lot of time in debugging (especially on TestSwarm where the report is all you have, there "not okay" or "Expected { foo: 500 }, Given: ['bar', '250']" is a big saver.
* Adding an "expect" to every test.
* Applying whitespace conventions and bringing consistency in the use of mw, mediaWiki, $, $j, jQuery. (except in the initial test where the aliases are being checked)
* Added cleanup for added CSSStyleSheet
* Making IPtest more complete (Based on IPTest.php)

13 years agoMake Pg installer work in the “common” case: follow up r89821 so that $exists can...
Mark A. Hershberger [Fri, 10 Jun 2011 18:48:16 +0000 (18:48 +0000)]
Make Pg installer work in the “common” case: follow up r89821 so that $exists can be set true if same user is the same for install and web.

13 years ago[Installer] .mw-help-hint and #config-live-log
Krinkle [Fri, 10 Jun 2011 17:09:15 +0000 (17:09 +0000)]
[Installer] .mw-help-hint and #config-live-log
* Merged padding rules into one
* Fixed the width issues of the textarea (see also r87975)

13 years agoLocalisation updates for core and extension messages from translatewiki.net (2011...
Raimond Spekking [Fri, 10 Jun 2011 16:48:37 +0000 (16:48 +0000)]
Localisation updates for core and extension messages from translatewiki.net (2011-06-10 16:40:00 UTC)

13 years agoSimplify link construction a bit.
Raimond Spekking [Fri, 10 Jun 2011 16:48:15 +0000 (16:48 +0000)]
Simplify link construction a bit.

13 years agoPostgreSQL install fixes:
Tim Starling [Fri, 10 Jun 2011 11:32:57 +0000 (11:32 +0000)]
PostgreSQL install fixes:
* Made PG throw a DBQueryError when it gets a query error, instead of DBUnexpectedError. Apparently this mistake goes back to r14625, when exceptions were first introduced. Did it by removing reportQueryError(), the DatabaseBase version works fine.
* Fixed several places where there was an attempt to check for a query error by checking if the result of query() was false. This never worked. Used try/catch instead.
* Made the DBConnectionError messages go on one line so that they don't mess up the formatting in the installer.
* In DatabasePostgres::selectDB(), only disconnect and reconnect if the DB name is actually changing.
* Made DatabasePostgres::schemaExists() less weird and scary.
* Added DatabasePostgres::roleExists() for use by the installer.
* Removed the PostgreSQL-specific hack to make _InstallUser have a default other than "root". Made _InstallUser into a proper DBMS-specific internal variable instead, since every DBMS we support so far needs a different default.
* Removed the $dbName parameters from openConnection/getConnection, and got rid of $this->useAdmin. Implemented a more sophisticated caching scheme instead. Partial revert of r89389 and r81440.
* When connecting as the install user before DB creation, and when testing the web user's credentials, try a few different database names and use whichever one works.
* Instead of connecting as the web user to create tables, I used SET ROLE. It seems cleaner and more like what the other DBMSes do during installation. "SET ROLE wikiuser" requires the same privileges as "CREATE SCHEMA ... AUTHORIZATION wikiuser", so it's unlikely to break anything.
* In the area of web account creation, fixed various minor logic errors and introduced more informative error messages at the submit stage, pre-install. Show a helpful error message if the web user exists already and the install user can't do the relevant SET ROLE.
* Split schema creation out to a separate install step.
* When creating an account as a non-superuser, add the administrative account to the new account's group. This is necessary to avoid a fatal error during installation (bug 28845).
* Removed code which alters an existing web user to have appropriate search paths and permissions. This may break other apps and is not necessary. As in other DBMSes, If the web user exists, it is the responsibility of the sysadmin to ensure that it has appropriate permissions.
* Rewrote setupPLpgSQL() to use the query builder functions.

13 years agoFixed typo in comment.
Tim Starling [Fri, 10 Jun 2011 11:32:28 +0000 (11:32 +0000)]
Fixed typo in comment.

13 years agoFix for r86741 which broke PG update due to the math table disappearing
Tim Starling [Fri, 10 Jun 2011 08:03:35 +0000 (08:03 +0000)]
Fix for r86741 which broke PG update due to the math table disappearing

13 years agoFix for logic error in r89529 causing a notice.
Tim Starling [Fri, 10 Jun 2011 07:48:21 +0000 (07:48 +0000)]
Fix for logic error in r89529 causing a notice.

13 years agoReverted r89393. A single Database object certainly should not be attempting to manag...
Tim Starling [Fri, 10 Jun 2011 07:36:22 +0000 (07:36 +0000)]
Reverted r89393. A single Database object certainly should not be attempting to manage multiple database connections, that is the job of LBFactory/LoadBalancer. I would like to see $wgSharedDB managed by LBFactory instead, for all DBMSes. Then the cruft in Database::tableName() can be removed.

r89393 caused tableExists() etc. to be completely broken, so PostgreSQL upgrade was broken too, see CR.

13 years ago(bug 29332) don't let install progress if user requests mediawiki-announce subscripti...
Chad Horohoe [Fri, 10 Jun 2011 03:27:37 +0000 (03:27 +0000)]
(bug 29332) don't let install progress if user requests mediawiki-announce subscription but does not enter email. Thanks Sumana :)

13 years agoSelf revert r89797: At least the block log link is wrong. I should not commit when...
Raimond Spekking [Thu, 9 Jun 2011 21:31:51 +0000 (21:31 +0000)]
Self revert r89797: At least the block log link is wrong. I should not commit when bed calls me :/

13 years agor89767 didn't please overlords
Niklas Laxström [Thu, 9 Jun 2011 21:25:41 +0000 (21:25 +0000)]
r89767 didn't please overlords

13 years agoSimplify link construction a bit.
Raimond Spekking [Thu, 9 Jun 2011 21:22:00 +0000 (21:22 +0000)]
Simplify link construction a bit.

13 years agoMarked Article::doRedirect() as deprecated, was used by Article::insertNewArticle...
Alexandre Emsenhuber [Thu, 9 Jun 2011 21:17:41 +0000 (21:17 +0000)]
Marked Article::doRedirect() as deprecated, was used by Article::insertNewArticle() and Article::updateArticle() that were removed. No usage left in core or extensions.

13 years agoLocalisation updates for core and extension messages from translatewiki.net (2011...
Raimond Spekking [Thu, 9 Jun 2011 20:58:25 +0000 (20:58 +0000)]
Localisation updates for core and extension messages from translatewiki.net (2011-06-09 20:44:00 UTC)

13 years agoRecommit SpecialPage.php and SpecialPageFactory.php parts of r89406 (removal of magic...
Alexandre Emsenhuber [Thu, 9 Jun 2011 20:00:40 +0000 (20:00 +0000)]
Recommit SpecialPage.php and SpecialPageFactory.php parts of r89406 (removal of magic __get() accessor for RequestContext)

13 years agoChange whitespace so it is easier to read.
Platonides [Thu, 9 Jun 2011 19:58:30 +0000 (19:58 +0000)]
Change whitespace so it is easier to read.

13 years agoIf this was ever passed an empty string it would spit "PHP Notice: Uninitialized...
Platonides [Thu, 9 Jun 2011 19:57:10 +0000 (19:57 +0000)]
If this was ever passed an empty string it would spit "PHP Notice:  Uninitialized string offset: 0..."