Sam Reed [Fri, 6 Jan 2012 15:36:17 +0000 (15:36 +0000)]
Add a few @since entries
Roan Kattouw [Fri, 6 Jan 2012 15:14:04 +0000 (15:14 +0000)]
Followup r108231, remove useless _this
Roan Kattouw [Fri, 6 Jan 2012 15:06:39 +0000 (15:06 +0000)]
Followup r108184: fix loading in Opera. Before, Opera would only begin to render the document after all async scripts were loaded, now it's the other way around (scripts load after the document is ready). This is not an improvement compared to the current situation, but it's not a regression either.
Alexandre Emsenhuber [Fri, 6 Jan 2012 14:21:16 +0000 (14:21 +0000)]
* Added callback to send notices from WikiImporter and use it in ImportReporter so that it can use the context to get messages and send them to OutputPage (also removes on usage of $wgCommandLineMode)
* Early abort on invalid title in ImportReporter::reportPage() since a notice has already been sent
* Localised message saying the title is invalid
Roan Kattouw [Fri, 6 Jan 2012 14:17:03 +0000 (14:17 +0000)]
Fix broken oldParser call in r108230
Roan Kattouw [Fri, 6 Jan 2012 14:11:34 +0000 (14:11 +0000)]
Fix up r108203: just loading mw.jqueryMsg in the bottom queue, then assuming its presence in mw.Message doesn't work, see CR comments.
* Moved message parsing (including $1 replacement) to Message.prototype.parser(), and let jqueryMsg override that when loaded
** Make the Message constructor public to make this possible
** Moved logic for skipping jqueryMsg when the message is simple from mw.Message to mw.jqueryMsg, where it belongs
* Remove mw.jqueryMsg from the default modules list in OutputPage. Modules that require PLURAL/GENDER should depend on mw.jqueryMsg
* TODOs
** The jqueryMsg parser is recreated for every mw.msg() call. It should probably be cached, but the only way I can think of is to add it as a member of the Map object, which is kind of weird
** Because jqueryMsg doesn't support a 'text' mode that expands PLURAL/GENDER but doesn't output HTML (leaves e.g. links alone), mw.Message.plain() and mw.Message.parse() currently behave identically. This is wrong and should be fixed, but that needs support in jqueryMsg too
Alexandre Emsenhuber [Fri, 6 Jan 2012 12:06:13 +0000 (12:06 +0000)]
* Don't allow to import a page in the Media: namespace, this didn't work just because WikiPage::factory() throws an exception when a Title in NS_MEDIA is passed
* Added localised message when trying to import a page with an interwiki prefix
Roan Kattouw [Fri, 6 Jan 2012 12:03:06 +0000 (12:03 +0000)]
Set position => 'top' for wikibits, required to support legacy gadgets. Spotted by Niklas on TranslateWiki. Ping r107327.
Roan Kattouw [Fri, 6 Jan 2012 11:44:41 +0000 (11:44 +0000)]
Remove unneeded dependency from wikibits, and move the showtoc and hidetoc messages to mw.util where they're used
Santhosh Thottingal [Fri, 6 Jan 2012 11:24:57 +0000 (11:24 +0000)]
Prevent #firstHeading overriding the language specific h1 height.
Ref Bug 30809
Santhosh Thottingal [Fri, 6 Jan 2012 09:14:45 +0000 (09:14 +0000)]
Use mw.jqueryMsg parser for message parsing to support PLURAL and GENDER
Follow up r107556 and based on the discussions on wikitech-l about this.
mediawiki.jqueryMsg is now loaded always. mw.msg uses the parser if required.
Add qunit test cases.
Raimond Spekking [Fri, 6 Jan 2012 08:06:05 +0000 (08:06 +0000)]
Fix grammar in r108154.
Spotted by kgh https://translatewiki.net/wiki/Thread:Support/About_MediaWiki:Perfcached/en
Santhosh Thottingal [Fri, 6 Jan 2012 06:56:01 +0000 (06:56 +0000)]
Correction in the test comment, followup r107824
Aaron Schulz [Fri, 6 Jan 2012 06:35:54 +0000 (06:35 +0000)]
r108192: Explicitly set flags for RecursiveDirectoryIterator constructor
Aaron Schulz [Fri, 6 Jan 2012 06:16:05 +0000 (06:16 +0000)]
* Cleanup for r108175: made if easier for extensions to interact with each other in TitleReadWhitelist hook. A handler can set $whitelisted to true to whitelist the page or it can set it to false and return false (most restrictive wins).
* Added some hook doc comments.
Aaron Schulz [Fri, 6 Jan 2012 05:15:51 +0000 (05:15 +0000)]
Fixes for r106752:
* Make sure FSFileIterator::current() directly returns the string path and that it is relative to the directory being searched.
* Fixed silly bug in testGetFileList that masked any bugs.
* Added a few code comments.
Brion Vibber [Fri, 6 Jan 2012 01:38:26 +0000 (01:38 +0000)]
Fix a race condition in classic edit toolbar setup
mediawiki.action.edit used to call mw.toolbar.init on its own during DOMReady -- this required that the module already be loaded *before* DOMReady, *and* that it be fully loaded before the immedate-execution code that sets up the default toolbar items (output from EditPage.php).
This failed since we've gone fully async in the bottom-loading items, and could result in failed calls to mw.toolbar.addButton.
Wrapped those immediate calls in a mediawiki.loader.using call to ensure they're run after the module is loaded, and moved the mw.toolbar.init call into that section so we guarantee it gets run first.
Aaron Schulz [Thu, 5 Jan 2012 23:35:38 +0000 (23:35 +0000)]
Partially reverted r108111: we can't assume subclasses put thumbnails in the...thumbnails zone. They might override the thumbnail path functions.
Roan Kattouw [Thu, 5 Jan 2012 23:32:41 +0000 (23:32 +0000)]
ResourceLoader: Add an experimental option to move the main module loading queue (the bottom queue) from the bottom of the <body> up into the <head> , while still being loaded asynchronously. This makes them load earlier, which should make the page load faster. This is the product of a long discussion on bug 27488
* Added a "blocking" state to mw.loader . When loading scripts while the document is not ready, the loader will use document.write() if blocking is true, and append to the <body> or the <head> if blocking is false. If the document is ready, the loader will always append to the <body>
* Enable blocking mode while loading the top queue, and disable it after. This ensures that modules in the top queue are still loaded in a blocking way as they were before
* If $wgResourceLoaderExperimentalAsyncLoading is true, the bottom queue is also loaded in the head, but with blocking mode disabled. Otherwise, it's loaded at the bottom of the <body> as before
* scripts-only and messages-only requests need special treatment:
** in the top queue, they can continue to use <script src="..."> tags because they are blocking
** if the bottom queue is at the bottom of the <body> (experimental async loading disabled), they can continue to use <script src="..."> tags as before
** if the bottom queue is in the <head> (experimental async loading enabled), they cannot use <script src="..."> tags, because those would block. Instead, call mw.loader.load() on the load.php URL
Brian Wolff [Thu, 5 Jan 2012 23:25:39 +0000 (23:25 +0000)]
Make sure that if we fail to read the App13 (iptc) block of a JPG file, that that doesn't block other metadata from being read. Also makes sure if more then one app13 block is in the file, they are all read, not just the last one that appears in the file (This required some changes to tests since before the intermediate value was just one value, now its an array of all such blocks)
Antoine Musso [Thu, 5 Jan 2012 23:01:24 +0000 (23:01 +0000)]
XCF: remove var_dump() statement
Raimond Spekking [Thu, 5 Jan 2012 21:31:37 +0000 (21:31 +0000)]
Localisation updates for core and extension messages from translatewiki.net
Alexandre Emsenhuber [Thu, 5 Jan 2012 20:29:53 +0000 (20:29 +0000)]
Per request of Aaron Schulz, follow-up r102187: added new 'TitleReadWhitelist' hook to allow extensions to bypass core (and other extensions) checks to allow an user to read the page
Alexandre Emsenhuber [Thu, 5 Jan 2012 19:05:02 +0000 (19:05 +0000)]
Per Nikerabbit, fix for r107637: added @since
Sam Reed [Thu, 5 Jan 2012 18:49:19 +0000 (18:49 +0000)]
Followup r107932
Check random class to be isntantiated is a subclass of DatabaseBase
Max Semenik [Thu, 5 Jan 2012 18:05:54 +0000 (18:05 +0000)]
Follow-up r108137: @since
Brian Wolff [Thu, 5 Jan 2012 17:56:51 +0000 (17:56 +0000)]
Follow-up r108147. PLURAL in message.
Greg Sabino Mullane [Thu, 5 Jan 2012 17:10:07 +0000 (17:10 +0000)]
For new 'export all' feature, cleanup whitepspace, remove redundant check, add documentation. Bug 10574y
Antoine Musso [Thu, 5 Jan 2012 16:33:52 +0000 (16:33 +0000)]
add basic XCF pictures for testing purpose
Brian Wolff [Thu, 5 Jan 2012 16:14:52 +0000 (16:14 +0000)]
(bug 33456) show $wgQueryCacheLimit on cached query pages, so users know that the results are artificially cut off, and its not just that there is only 1000 wanted files (or whatever else) on the wiki.
Brian Wolff [Thu, 5 Jan 2012 15:34:26 +0000 (15:34 +0000)]
(bug 33321. Sort of) Adding a line to MediaWiki:Sidebar that contains a pipe, but doesn't
have any pipes after being transformed by MessageCache, causes exception on
all pages.
This can happen with lines like:
**{{#if:yes|Something}}
Thank you to liangent for figuring out how to escape a | without {{!}} existing and | not working.
Antoine Musso [Thu, 5 Jan 2012 15:06:49 +0000 (15:06 +0000)]
explanations by Erwin Dokter on bug 33496
Antoine Musso [Thu, 5 Jan 2012 14:48:55 +0000 (14:48 +0000)]
rewrite getXCFMetaData() to get ride of im identify
By reading the file header and unpacking the data, we can avoid
shelling out to imagemagick 'identify'. Save up some CPU cycles :D
An XCF is made of a canvas of a given width / height, the various
layers are applied to it which must fit in the canvas. So we just
use the canvas size :-)
I do not think we have any usage for channels count, so I have just
skip that part. I am not sure it makes any sense when the picture
can be made of several layers each using different channels count.
Bits per color is always 8 per definition. Grayscale is 0 - 255
and indexed palette is 256 colors at most.
XCF spec:
http://svn.gnome.org/viewvc/gimp/trunk/devel-docs/xcf.txt?view=markup
pack() / unpack() is familiar to perl monkeys
Fully reimplements r107351
Max Semenik [Thu, 5 Jan 2012 13:00:37 +0000 (13:00 +0000)]
Accessor for LinksUpdate's parser output
Antoine Musso [Thu, 5 Jan 2012 11:03:53 +0000 (11:03 +0000)]
parserTest need to clone user_former_groups table
Not sure why it is suddenly needed.
Siebrand Mazeland [Thu, 5 Jan 2012 09:33:46 +0000 (09:33 +0000)]
Add additional parameter $altUserName to Linker::userLink to make it possible to override the displayed user name.
Renamed $userText to $userName following Nikerabbit's advice.
Antoine Musso [Thu, 5 Jan 2012 09:24:16 +0000 (09:24 +0000)]
additional monospaced font-family hacks
font-family: monospace, monospace;
Is elegant and has my preference :-)
font-family: monospace, DOESNOTEXISTREALLY;
Really make it obvious
follow r108123
Antoine Musso [Thu, 5 Jan 2012 09:16:13 +0000 (09:16 +0000)]
test rendering font-family for monospaced fonts
follow r108112
Aaron Schulz [Thu, 5 Jan 2012 06:18:36 +0000 (06:18 +0000)]
* Made use of FileBackend function 'latest' param in FileOp.
* Added FileBackend process cache for fileExists(), getFileTimestamp(), and getLocalReference().
* Refactored getFileSha1Base36() into parent class and subclass functions.
* Removed some FileBackendMultiWrite comment duplication.
Mark A. Hershberger [Thu, 5 Jan 2012 04:53:05 +0000 (04:53 +0000)]
Bug 33514 - Make category "columns" (table cells) equal width
Author: Erwin Dokter
Mark A. Hershberger [Thu, 5 Jan 2012 02:24:29 +0000 (02:24 +0000)]
Fix Bug 33496 - Misleading comment regarding preformatted font in commonElements.css
Author: Erwin Dokter
Aaron Schulz [Thu, 5 Jan 2012 01:58:05 +0000 (01:58 +0000)]
In SpecialUploadStash:
* Updated outputLocallyScaledThumb() and outputLocalFile() to handle changes in r106752.
In MediaTransformOutput:
* Added a storage path field to the transformation output object and set it in File::maybeDoTransform().
In File:
* Fixed maybeDoTransform() handling if repo member is not set and made it fully respect $wgIgnoreImageErrors.
In BitmapHandler:
* Don't set bogus path if TRANSFORM_LATER in doTransform() for deffered renderings.
Aaron Schulz [Thu, 5 Jan 2012 01:47:00 +0000 (01:47 +0000)]
In FSFile:
* Added getMimeType() and extensionFromPath() functions.
* A few other minor code and comment cleanups.
Aaron Schulz [Thu, 5 Jan 2012 01:33:58 +0000 (01:33 +0000)]
Added a bit to StreamFile docs
Krinkle [Thu, 5 Jan 2012 01:30:07 +0000 (01:30 +0000)]
[mediawiki.loader] fix numerous bugs and edge cases discovered with Roan on Etherpad/Skype
* 'undefined' state removed:
State 'undefined' has been removed. This was half-supported, probably with the intention to support loading of modules that are registered server side after the startup module was loaded (either because the server had a deployment while the user was browsing the page, or because loader.load calls in HTML document or loader.register where cached and out of sync. That's not unlikely to happen after deployment, after which there is a 5 minute window of cached startup modules in peoples browser cache but fresh HTML documents.
Instead of filling in the missing registry with no information, simply don't support this. Shouldn't break anything as it wasn't supported all the way.
* Document state 'missing'. Used by the load.php when a requested module doesn't exist according to the server.
* Fix recurse(). Previously it failed to recognize circular dependencies, because "unresolved" started as an empty array (given by resolve(). See also test cases at bottom of commitmsg.
* Fix recurse() even more. It was using inArray as a parameter to [].splice(), which is bad as inArray can be -1, in which case splice(-1,1) will remove the last item in the array, instead of the one at the index. Fortunately this was never exploited as 'unresolved=[]' is starting point, so nothing to delete.
* Fix resolve(). Don't return an empty array, which is contrary to it's (now documented) behavior of including the requested module in the return array. Instead letting recurse() handle it.
* Add magic filter to filter() named "unregistered" which will reduce the set to only module names that are not in the registry.
* Removing call to request() from the bottom bottom of mw.loader.implement. This can't be needed as before load.php is called the toes implement, all dependencies are put in queue and set to "loading" state, this queue is build first and then the load requests happen. At point implement is called, it's dependencies must already be in the queue. If any unloaded dependencies at moment of implementation (which is very common), the next implement will call execute which does handlePending, which will execute previously implemented modules that were waiting for dependencies to arrive.
* Make load() and using() no longer return meaningless booleans, they were undocumented and meaningless as they are either used as voids or with callbacks.
* Bring consistency in handling input. If direct input to a loader function is invalid or contains inexistent modules, throw an exception. Otherwise use error callback.
-- with one exception, that is the raw module list passed to mw.loader.load, those are not related to each other and should continue even if there is an inexistent module in there.
-- For example if a deployment occurs, adding an extension with modules loaded on a particular page. The user will have a startup module in cache up to 5 minutes that doesn't have this module's registry yet. In that case the new modules should simply be ignored. Before this commit they were also silently ignored, but not by skipping them. previously inexistent/undefined modules would give ok-callback of using() instantly, because filter(['ready'], modules) is the same as modules when it contains nothing.
Daniel Friesen [Thu, 5 Jan 2012 00:26:13 +0000 (00:26 +0000)]
Followup r105809; Split up the tests with a dataProvider and add a round trip test.
Sam Reed [Wed, 4 Jan 2012 21:34:56 +0000 (21:34 +0000)]
Tidy up nested assignments
Sam Reed [Wed, 4 Jan 2012 21:34:02 +0000 (21:34 +0000)]
Tidy up/rearrange RELEASE-NOTES-1.19 slightly
Sam Reed [Wed, 4 Jan 2012 21:30:06 +0000 (21:30 +0000)]
* (bug 33525) clearTagHooks doesn't clear function hooks.
* (bug 33523) Function tag hooks don't appear on Special:Version.
Patches by Nikola Kovacs
Raimond Spekking [Wed, 4 Jan 2012 21:00:47 +0000 (21:00 +0000)]
Localisation updates for core and extension messages from translatewiki.net
Roan Kattouw [Wed, 4 Jan 2012 20:01:29 +0000 (20:01 +0000)]
Redo r107066 properly: apply CSSJanus to user CSS previews when needed. Ping r94421
Krinkle [Wed, 4 Jan 2012 19:14:43 +0000 (19:14 +0000)]
[mediawiki.debug] add implied semi-colon
* poke r107610
Krinkle [Wed, 4 Jan 2012 19:13:45 +0000 (19:13 +0000)]
[JSTesting] rm redundant comment, there are no more "*" behind any entries
Krinkle [Wed, 4 Jan 2012 19:11:45 +0000 (19:11 +0000)]
[JSTesting] fix case typo in file path
* twn:
exception 'MWException' with message 'ResourceLoaderFileModule::readScriptFiles: script file not found: "/www/sandwiki/tests/qunit/suites/resources/mediawiki/mediawiki.title.test.js"' in /www/sandwiki/includes/resourceloader/ResourceLoaderFileModule.php:518
Mark A. Hershberger [Wed, 4 Jan 2012 19:09:03 +0000 (19:09 +0000)]
re r102301 — apply Krinkle's suggested fix.
Ryan Kaldari [Wed, 4 Jan 2012 19:01:12 +0000 (19:01 +0000)]
follow-up to r106912 - NULL to null
Mark A. Hershberger [Wed, 4 Jan 2012 18:01:10 +0000 (18:01 +0000)]
re r106025 — apply follow up from Luca Fulchir
Sam Reed [Wed, 4 Jan 2012 17:30:09 +0000 (17:30 +0000)]
Bug 33494 - Remove redundant CSS from chick/main.css
Patch by Erwin Dokter
Roan Kattouw [Wed, 4 Jan 2012 17:15:25 +0000 (17:15 +0000)]
Revert r107066, see CR comments for full rationale. Basically this moves the previewed CSS to the wrong place in the <head> while not actually applying Janus
Mark A. Hershberger [Wed, 4 Jan 2012 16:43:46 +0000 (16:43 +0000)]
follow up r107932 with w/s fixes
Mark A. Hershberger [Wed, 4 Jan 2012 16:29:01 +0000 (16:29 +0000)]
Attempt to avoid a warning I got during input:
$ 7z x -so ../dewikisource-
20111012-pages-meta-history.xml.7z | php maintenance/importDump.php
7-Zip [64] 9.20 Copyright (c) 1999-2010 Igor Pavlov 2010-11-18
p7zip Version 9.20 (locale=en_US.UTF-8,Utf16=on,HugeFiles=on,4 CPUs)
Processing archive: ../dewikisource-
20111012-pages-meta-history.xml.7z
Extracting dewikisource-
20111012-pages-meta-history.xml
PHP Notice: Undefined index: wgExtensionAliasesFiles in /home/mah/work/code/mediawiki/mw-
svn/includes/cache/CacheDependency.php on line 386
PHP Stack trace:
PHP 1. {main}() /home/mah/work/code/mediawiki/mw-svn/maintenance/importDump.php:0
PHP 2. require_once() /home/mah/work/code/mediawiki/mw-svn/maintenance/importDump.php:287
PHP 3. BackupReader->execute() /home/mah/work/code/mediawiki/mw-svn/maintenance/doMaintenance.php:105
PHP 4. BackupReader->importFromStdin() /home/mah/work/code/mediawiki/mw-svn/maintenance/importDump.php:93
PHP 5. BackupReader->importFromHandle() /home/mah/work/code/mediawiki/mw-svn/maintenance/importDump.php:249
PHP 6. WikiImporter->doImport() /home/mah/work/code/mediawiki/mw-svn/maintenance/importDump.php:282
PHP 7. WikiImporter->handlePage() /home/mah/work/code/mediawiki/mw-svn/includes/Import.php:411
PHP 8. WikiImporter->processTitle() /home/mah/work/code/mediawiki/mw-svn/includes/Import.php:534
PHP 9. Title->userCan() /home/mah/work/code/mediawiki/mw-svn/includes/Import.php:788
PHP 10. Title->getUserPermissionsErrorsInternal() /home/mah/work/code/mediawiki/mw-svn/includes/Title.php:1557
PHP 11. Title->checkSpecialsAndNSPermissions() /home/mah/work/code/mediawiki/mw-svn/includes/Title.php:2118
PHP 12. Title->getNsText() /home/mah/work/code/mediawiki/mw-svn/includes/Title.php:1728
PHP 13. Language->needsGenderDistinction() /home/mah/work/code/mediawiki/mw-svn/includes/Title.php:723
PHP 14. LocalisationCache->getItem() /home/mah/work/code/mediawiki/mw-svn/languages/Language.php:442
PHP 15. LocalisationCache->loadItem() /home/mah/work/code/mediawiki/mw-svn/includes/LocalisationCache.php:218
PHP 16. LocalisationCache->initLanguage() /home/mah/work/code/mediawiki/mw-svn/includes/LocalisationCache.php:283
PHP 17. LocalisationCache->isExpired() /home/mah/work/code/mediawiki/mw-svn/includes/LocalisationCache.php:395
PHP 18. GlobalDependency->isExpired() /home/mah/work/code/mediawiki/mw-svn/includes/LocalisationCache.php:367
Antoine Musso [Wed, 4 Jan 2012 16:24:41 +0000 (16:24 +0000)]
application/vnd.oasis.opendocument.database [OFFICE]
Mark A. Hershberger [Wed, 4 Jan 2012 15:52:48 +0000 (15:52 +0000)]
Fix Bug 29102 - Upgrade fails "Unknown character set: 'mysql4'
Fix suggested by Tremault and tested by maanto. Confirmed by me.
Antoine Musso [Wed, 4 Jan 2012 15:14:08 +0000 (15:14 +0000)]
mime.types: add opendoc database
application/vnd.oasis.opendocument.database odb
Antoine Musso [Wed, 4 Jan 2012 15:12:08 +0000 (15:12 +0000)]
mime.types: sort opendocument entries
Sorted all application/vnd.oasis.opendocument entries, no other change.
Siebrand Mazeland [Wed, 4 Jan 2012 14:52:14 +0000 (14:52 +0000)]
Consistency tweak.
Antoine Musso [Wed, 4 Jan 2012 13:37:28 +0000 (13:37 +0000)]
XCF format: code style/comment
* Saved a level of indentation by returning early on command failure
* $md -> $metadata
Follow r107351
Antoine Musso [Wed, 4 Jan 2012 13:22:01 +0000 (13:22 +0000)]
document getImageSize() return format
Return array should return the same kind of array that PHP
internal method getimagesize() returns. See upstream doc:
http://www.php.net/getimagesize
Antoine Musso [Wed, 4 Jan 2012 11:33:35 +0000 (11:33 +0000)]
enhance size tooltip on changelist
follow r104085
Niklas Laxström [Wed, 4 Jan 2012 11:03:33 +0000 (11:03 +0000)]
Guessing a typo fix, could also be 'is a feed' but neither expression really opens to me
Antoine Musso [Wed, 4 Jan 2012 10:51:57 +0000 (10:51 +0000)]
JSTesting: javascripttest-pagetext-skin msg
* added colon in the message
* reworded message
Follow r107919
Max Semenik [Wed, 4 Jan 2012 10:41:39 +0000 (10:41 +0000)]
Made installer not install on SQLite less than 3.3.7, would prevent stuff like bug 25746. Also, threw MSSQL out of release notes - it can't be installed through the usual means.
Antoine Musso [Wed, 4 Jan 2012 10:35:11 +0000 (10:35 +0000)]
JSTesting: escape message
Replaced call to wfMsg() by wfMsgHtml() just to be safe.
Per cr on r107919.
Antoine Musso [Wed, 4 Jan 2012 10:31:02 +0000 (10:31 +0000)]
JSTesting: make sure wrapSummaryHtml is given a valid state
* wrapSummaryHTML() now really need one of three states or an exception
is thrown.
* Moved a parameter incorrectly passed to addHtml() up to the previous
wrapSummaryHtml() code. Per CR on r107919
Antoine Musso [Wed, 4 Jan 2012 10:18:49 +0000 (10:18 +0000)]
ResourceLoaderTestModules doc: better wording
follow r107993
Max Semenik [Wed, 4 Jan 2012 10:08:59 +0000 (10:08 +0000)]
Made WebInstaller::showStatusMessage() look uniformly for one and multiple errors/warnings per status
Antoine Musso [Wed, 4 Jan 2012 09:08:16 +0000 (09:08 +0000)]
revert r107006
(bug 23427) Introduced {{PAGEID}} variable to expose page.page_id
[patch by Waldir]
Santhosh Thottingal [Wed, 4 Jan 2012 09:06:05 +0000 (09:06 +0000)]
Stylize.
Antoine Musso [Wed, 4 Jan 2012 08:59:57 +0000 (08:59 +0000)]
svn property on .svg file (follow r107940)
Santhosh Thottingal [Wed, 4 Jan 2012 08:56:36 +0000 (08:56 +0000)]
Address an edge case - input being null. Add test cases.
Max Semenik [Wed, 4 Jan 2012 08:38:13 +0000 (08:38 +0000)]
Follow-up r107376: disable test by default, causes failures in some configurations
Niklas Laxström [Wed, 4 Jan 2012 08:29:58 +0000 (08:29 +0000)]
Typos, casing. Still don't understand "'qunit' is feed using tests/qunit/QUnitTestResources.php"
Ping r107919
Alexandre Emsenhuber [Wed, 4 Jan 2012 08:24:11 +0000 (08:24 +0000)]
Fix for r107945: also partially revert r107842
Aaron Schulz [Wed, 4 Jan 2012 08:07:52 +0000 (08:07 +0000)]
* Updated rebuildImages.php per r106752.
* Also added FIXME about old call to bogus renameFile() function in rebuildImages.php.
* Added ContainerShardIterator class from r107980 to autoloader.
* Added a few FileBackend comments and improved FileRepo::enumFiles() comments.
Niklas Laxström [Wed, 4 Jan 2012 07:46:16 +0000 (07:46 +0000)]
consistency and clarification fixes for messages merged in r107919
Niklas Laxström [Wed, 4 Jan 2012 06:52:59 +0000 (06:52 +0000)]
* Add typehint to debug Call to a member function getTitle() on a non-object in /www/w/includes/EditPage.php on line 121
* Remove @todo from constructor, because it's a constructor
* Some typos and code style
Aaron Schulz [Wed, 4 Jan 2012 02:15:07 +0000 (02:15 +0000)]
* Added FileBackendBase::getFileContents() function with a default FileBackend version.
* Added read-only mode to FileBackendBase config.
* Moved FileBackendBase::getFileTimestamp() up slightly.
Krinkle [Wed, 4 Jan 2012 01:29:39 +0000 (01:29 +0000)]
[mediawiki.debug] display: inline-block; work-around for in IE7
* display: inline-block; is not supported by IE7
* Using standard work-around with triggering hasLayout (via zoom: 1;) on an inline element (only for IE7 through *hack)
Krinkle [Wed, 4 Jan 2012 01:28:01 +0000 (01:28 +0000)]
[jquery.footHovzer] new plugin for mw-log-console and mw-debug-toolbar
* Previously mw.log and mw.Debug both were in a fixed container on the bottom, overlapping each other. mw.log did increase the body's padding-bottom to account for the height so that all content is still visible, but it was still a problem when mw.Debug came along.
* This plugin adds a single fixed position element to bottom of the body, to which other stuff like mw.log and mw.Debug can append a non-fixed position container. That will take care of it.
* Method update() will re-check the padding and scroll position and fix where needed
* Reduces code a little bit
Aaron Schulz [Wed, 4 Jan 2012 01:08:33 +0000 (01:08 +0000)]
* Added supported for container sharding via existing FileRepo-style hash paths. For sharded containers, all paths used to store files must be hashed appropriately (or the path will be rejected as invalid).
* Tweaked FileBackend::isValidContainerName() length check to give some head room for shard and segment container names.
* Updated FSFileBackend to use new resolveStoragePathReal() function for file paths.
* Removed useless cache invalidation in FileBackend::concatenateInternal().
Krinkle [Wed, 4 Jan 2012 00:05:08 +0000 (00:05 +0000)]
Fix unused parameter from r107350
* /me is going to run JSLint again instead of JSHint, that would've caught this.
John Du Hart [Tue, 3 Jan 2012 23:52:15 +0000 (23:52 +0000)]
Followup r107954, rm debugging code
Mark A. Hershberger [Tue, 3 Jan 2012 23:42:15 +0000 (23:42 +0000)]
re Bug 31007 - rebuildrecentchanges.php fails for Postgresql with strings for integers
Authors: "Max" and Tim Landscheidt
Unquote 'null'
Brion Vibber [Tue, 3 Jan 2012 23:41:08 +0000 (23:41 +0000)]
Fix regression in r107354: typo caused 'watch' link in Vector to never be hooked up for ajax watch. :)
John Du Hart [Tue, 3 Jan 2012 22:36:35 +0000 (22:36 +0000)]
Followup r107862
Fix bug that caused a PHP notice
Vertical-align:top on callers in the console
Brion Vibber [Tue, 3 Jan 2012 22:29:32 +0000 (22:29 +0000)]
Revert r107054: breaks execution of lines that might test or output '<?php' say when generating a PHP script
Is apparently meant to strip an initial line from PHP script files being piped in, but is insufficient to actually catch all cases.
Rotem Liss [Tue, 3 Jan 2012 22:18:03 +0000 (22:18 +0000)]
* Setting "javascripttest-backlink" message as ignored.
* Adding special page alias to he.
Krinkle [Tue, 3 Jan 2012 22:16:24 +0000 (22:16 +0000)]
[JSTesting] Fix static html tests
* static tests need wgServer as well
* Follows-up r107946, r107919
Krinkle [Tue, 3 Jan 2012 21:53:23 +0000 (21:53 +0000)]
[JSTesting] Fix static html tests
* r107919 merge broke it as it removed all the inline mw.config calls, and the tearDown is resetting them to the initial values. That works fine on a MediaWiki page but on the static page are now no variables at all. Copying a basic sample to it. This should've been in the static test suite all along actually.
* Calculating wgScriptPath based on location.pathname instead of location.href, exclude wgServer
* Follows-up r107919
Brion Vibber [Tue, 3 Jan 2012 21:44:41 +0000 (21:44 +0000)]
Revert r107769, r107771, r107825, r107840, r107927, r107934
Title objects are meant to be dumb value objects; we shouldn't add to their internal state like this, but should be working to remove the bits already in there like the article ID.
Preloading information like this can make sense, but probably belongs in WikiPage, not Title.