Mark A. Hershberger [Fri, 4 Nov 2011 13:08:44 +0000 (13:08 +0000)]
eol w/s
Sam Reed [Fri, 4 Nov 2011 11:53:22 +0000 (11:53 +0000)]
Fix history versions
Sam Reed [Fri, 4 Nov 2011 11:51:06 +0000 (11:51 +0000)]
1.17 -> 1.18
Jure Kajzer [Fri, 4 Nov 2011 11:35:10 +0000 (11:35 +0000)]
* fixed ipblocks.ipb_by_text field, removed default blank not null (fixed install&update)
* fixed Block->insert; ipblocks.ipb_id is autoincrement field (should use sequences for compatibility)
Alexandre Emsenhuber [Fri, 4 Nov 2011 08:58:10 +0000 (08:58 +0000)]
Use WikiPage instead of Article to call the hasViewableContent() method
Jure Kajzer [Fri, 4 Nov 2011 07:17:13 +0000 (07:17 +0000)]
* fixed a typo in oracle/tables.sql
* readded the sha1 fields (installer & upgrader)
Alexandre Emsenhuber [Fri, 4 Nov 2011 06:59:36 +0000 (06:59 +0000)]
Per Platonides, fix for r101630: correct class name
Aaron Schulz [Fri, 4 Nov 2011 01:56:51 +0000 (01:56 +0000)]
Removed debug code from r101944
Brion Vibber [Fri, 4 Nov 2011 01:31:15 +0000 (01:31 +0000)]
Partial revert of r101912 -- restores r99315 'includexmlnamespace' parameter for API, but keeps the logic to keep that from clobbering alternate XML output formats with their own namespaces.
Sam Reed [Fri, 4 Nov 2011 01:28:52 +0000 (01:28 +0000)]
alpha -> beta
Merge RELEASE-NOTES changes
Sam Reed [Fri, 4 Nov 2011 01:00:06 +0000 (01:00 +0000)]
Documentation
Fix mime-type on image
Sam Reed [Fri, 4 Nov 2011 00:59:11 +0000 (00:59 +0000)]
Resync 1.18 RELEASE-NOTES files
Sam Reed [Fri, 4 Nov 2011 00:32:45 +0000 (00:32 +0000)]
Add Central Dusun language code to RELEASE-NOTES-1.18
Brion Vibber [Thu, 3 Nov 2011 23:14:58 +0000 (23:14 +0000)]
* (bug 31878, bug 31542) Fix XML namespace output in API; removed now-unneeded includexmlnamespace parameter.
Any existing uses of includexmlnamespace should be ignored and harmless.
If output data already includes a specific default namespace as an xmlns attribute, it is now retained -- so for example the RSD API discovery module outputs the appropriate namespace instead of the generic API one.
Partial revert of r99135 which added the includexmlnamespace parameter as a temporary requirement to get the xmlns attribute included.
Followup to r88007 which added the xmlns originally, but in the wrong order so it overrode existing output data.
Platonides [Thu, 3 Nov 2011 22:58:16 +0000 (22:58 +0000)]
Follow-up r101808. Use canonical class name.
Aaron Schulz [Thu, 3 Nov 2011 22:21:19 +0000 (22:21 +0000)]
FU r96015: removed & in MakeGlobalVariablesScript hook
Mark A. Hershberger [Thu, 3 Nov 2011 21:22:34 +0000 (21:22 +0000)]
eol w/s
OverlordQ [Thu, 3 Nov 2011 21:00:14 +0000 (21:00 +0000)]
Bug 28172 - wfGetDB called when it shouldn't be
Sam Reed [Thu, 3 Nov 2011 20:55:57 +0000 (20:55 +0000)]
Fix fixme on r101177
Chad Horohoe [Thu, 3 Nov 2011 20:17:49 +0000 (20:17 +0000)]
Whoops, need a !
Raimond Spekking [Thu, 3 Nov 2011 20:16:37 +0000 (20:16 +0000)]
Localisation updates for core and extension messages from translatewiki.net
Chad Horohoe [Thu, 3 Nov 2011 20:15:27 +0000 (20:15 +0000)]
*grumble grumble* know it alls think I'm abusing $wgMiserMode. So fine, have a SHINY NEW GLOBAL INSTEAD. bug 32072, r92109, yada-yada.
This is not a "useless" error message, regardless of what the bug claims.
Alexandre Emsenhuber [Thu, 3 Nov 2011 19:25:30 +0000 (19:25 +0000)]
Fix for r101858: forgot the override those methods
Brion Vibber [Thu, 3 Nov 2011 19:18:49 +0000 (19:18 +0000)]
bug 32183: remove the client-* classes added from user-agent-sniffing onto the <html> element.
Browser sniffing is bad, and having these here encourages unpleasantness.
Keeping the js/nojs ones as those are useful and non-evil (they're an example of feature detection, exposed via CSS here rather than through JS.)
Alexandre Emsenhuber [Thu, 3 Nov 2011 19:16:09 +0000 (19:16 +0000)]
Follow-up r101858: forgot to commit AutoLoader.php
Alexandre Emsenhuber [Thu, 3 Nov 2011 19:14:46 +0000 (19:14 +0000)]
Per Aaron, fix for r89405: introduced DerivativeRequest to allow to override the request parameters, but not headers, cookies and session. Updated ApiEditPage (for the bug) and ApiLogin (for future use)
Brion Vibber [Thu, 3 Nov 2011 18:33:11 +0000 (18:33 +0000)]
Add to jquery.client's qunit tests by actually testing the test() function against the sample user-agents.
Also changed from hardcoded constant of how many user-agents to test to counting them so it won't get out of sync.
Added Firefox 10 and IE 10 sample user-agent strings in to make sure they test correctly.
They *do* test correctly *if and only if* you use raw numbers for a major-version test (eg > 7). If you tried > "7.0" you would find it fails on the version 10 browsers.
Niklas Laxström [Thu, 3 Nov 2011 17:09:29 +0000 (17:09 +0000)]
Followup r101834 - this time tested
Niklas Laxström [Thu, 3 Nov 2011 17:07:29 +0000 (17:07 +0000)]
Quick ugly fix to avoid duplicate log entries in rc feed.
I didn't think that save() would send lines, since there was function to do that (added in r45725 - duplicating existing code).
This part of codebase could use much TLC
Antoine Musso [Thu, 3 Nov 2011 17:06:44 +0000 (17:06 +0000)]
let us force thumb rendering even if previous file exist
The new FORCE_RENDER flag should always be used with FORCE_NOW. It let us
override the file_exist() check and thus trigger a thumb rendering even if
the file exist.
Will be needed by the vipscaler extension.
Mark A. Hershberger [Thu, 3 Nov 2011 16:47:42 +0000 (16:47 +0000)]
w/s fixup
Antoine Musso [Thu, 3 Nov 2011 16:18:48 +0000 (16:18 +0000)]
add \n to a wfDebug() call
follow up r48417
Mark A. Hershberger [Thu, 3 Nov 2011 15:14:02 +0000 (15:14 +0000)]
typo
Antoine Musso [Thu, 3 Nov 2011 15:06:52 +0000 (15:06 +0000)]
comment that isValidURI must reject file:// URI
Mark A. Hershberger [Thu, 3 Nov 2011 14:04:01 +0000 (14:04 +0000)]
use isValidURI for redirect check
Antoine Musso [Thu, 3 Nov 2011 14:01:00 +0000 (14:01 +0000)]
man page help pages generation
Tweaked mwdocgen.php to let us generates man pages. They will be saved to
./docs/man so make sure to update your MANPATH if you use this feature.
Usage:
php mwdocgen.php --all --generate-man
or
make man
Mark A. Hershberger [Thu, 3 Nov 2011 13:58:29 +0000 (13:58 +0000)]
eol w/s
Mark A. Hershberger [Thu, 3 Nov 2011 13:57:25 +0000 (13:57 +0000)]
followup r101490, remove extra tabs after talking to Nikerabbit.
Brian Wolff [Thu, 3 Nov 2011 13:10:14 +0000 (13:10 +0000)]
(bug 31944). As a work around for bug 31944, don't extract tiff:YCbCrSubSampling from XMP.
Several files have wrong type for that property, XMPReader currently marks entire file as invalid instead of just specific property, if there is a type mismatch in XMP data.
Alexandre Emsenhuber [Thu, 3 Nov 2011 09:21:45 +0000 (09:21 +0000)]
* Set constant default values in class definition and not constructor
* Added getTitle()
* Mark constructor and getArticle() as public
* Removed a PHP4-ism in constructor
Alexandre Emsenhuber [Thu, 3 Nov 2011 08:55:04 +0000 (08:55 +0000)]
Cleanup to deletion related code:
* Use the new 'cannotdelete-title' message for the title on error instead of "internal error"
* Don't use OutputPage::showFatalError() but add the content by "normal" methods (related to the item above)
* Don't use the $DeleteReason and $DeleteReasonList member variables in Article and FileDeleteForm but local variables since they are not used anywhere else
* Moved down some variable definitions near where they are used so that they are only set when they will actually be used
* Same for the setPageTitle() call with 'delete-confirm' message in Article::delete() that is now in Article::confirmDelete()
* Added missing "true" as fourth parameter of WikiPage::doDeleteArticle() call that was missing
* Factorised getTitle() and getUser() calls
* Salt token with array( 'delete', 'page title' )
Alexandre Emsenhuber [Thu, 3 Nov 2011 08:16:15 +0000 (08:16 +0000)]
* Use WikiPage instead of Article and call doDeleteArticle() with a custom error message instead of calling doDelete()
* Call Title::getUserPermissionsErrors() instead of User::isAllowed() to know whether the user can delete the page
Chad Horohoe [Thu, 3 Nov 2011 03:30:27 +0000 (03:30 +0000)]
Typofix in comment
Chad Horohoe [Thu, 3 Nov 2011 00:41:12 +0000 (00:41 +0000)]
Address fixme on r101606, should also call from $this
Aaron Schulz [Wed, 2 Nov 2011 20:55:08 +0000 (20:55 +0000)]
Make use of Title::isSpecialPage() in various places
Bryan Tong Minh [Wed, 2 Nov 2011 20:48:50 +0000 (20:48 +0000)]
Per comments on r99911 move $wgMaxImageArea check back to normaliseParams(). Added hook BitmapHandlerCheckImageArea to override the area check. I'm not very happy with this overly specific hook, but I don't see a clear way to obtain the functionallity required otherwise.
Remove the width and height params from BitmapHandler::getImageArea(). There is really no reason for them to be there.
Raimond Spekking [Wed, 2 Nov 2011 20:45:29 +0000 (20:45 +0000)]
Localisation updates for core and extension messages from translatewiki.net
Alexandre Emsenhuber [Wed, 2 Nov 2011 20:18:43 +0000 (20:18 +0000)]
Early return if the page exists and redlink parameter was provided instead of going through all the normal process and then redirect; also moved a part of outside the "else" condition for consistency with the code below
Mark A. Hershberger [Wed, 2 Nov 2011 20:15:28 +0000 (20:15 +0000)]
avoid E_STRICT warnings when no --installdbpass is given
Aaron Schulz [Wed, 2 Nov 2011 20:01:22 +0000 (20:01 +0000)]
* Let HTMLFileCache constructor take in a Title *or* a the prefixed dbkey itself.
* Tweaked filecache fallback in fileCachedPage() to try the raw title param. If the DB is down, we can get most views of titles with colons in them to work this way. Previously, it could fail on an interwiki lookup.
Brion Vibber [Wed, 2 Nov 2011 19:36:49 +0000 (19:36 +0000)]
* (bug 32126) Fix 1.18 regression in watchlist editor when items already removed from watchlist
Overriding validation in HTMLForm bits so the extras can be safely ignored.
Aaron Schulz [Wed, 2 Nov 2011 18:47:04 +0000 (18:47 +0000)]
Alias NS_MEDIA page views to NS_FILE. Fixes bug 32032.
Chad Horohoe [Wed, 2 Nov 2011 18:03:46 +0000 (18:03 +0000)]
(bug 31822) Error during upgrade due to output buffer reset in stdout.
In various places in Maintenance, we're overwriting our output buffer rather than appending to it. This doesn't matter too much in the CLI, but it can corrupt headers when upgrading from the web. Thanks Woozle for pointing this out.
Aaron Schulz [Wed, 2 Nov 2011 17:33:00 +0000 (17:33 +0000)]
FU r101629, made Pager only call doBatchLookups() if there are rows
Alexandre Emsenhuber [Wed, 2 Nov 2011 16:36:20 +0000 (16:36 +0000)]
* Use WikiPage instance of Article
* Changed to call to directly create a Revision instead of going through WikiPage or Article
Alexandre Emsenhuber [Wed, 2 Nov 2011 15:30:55 +0000 (15:30 +0000)]
* Made PermissionsError exception accept an optional second parameter for the description of the errors (as returned by Title::getUserPermissionsErrors())
* PermissionsError now calls OutputPage::showPermissionsErrorPage() to display the error (this is needed to make the item above work correctly)
* Removed the override of the HTML title in OutputPage::showPermissionsErrorPage() so that it shows "Permission errors - Sitename" instead of simply "Permission errors" for consistency with the other things
* Pass the error array returned by Title::getUserPermissionsErrors() to PermissionsError where available
* Converted direct calls to OutputPage::showPermissionsErrorPage() to throw an PermissionsError error instead
* Added 'action-rollback' message that will be displayed when accessing action=rollback without sufficient rights
* Changed getRestriction() in subclasses of Action to return null when they previously returned 'read' so that user rights can be check with Title::getUserPermissionsErrors()
* Reordered checks to do first user rights, then block (if needed) and finally read only (also if needed) so that users don't think the error is temporary when they both don't have right and the database is locked
John Du Hart [Wed, 2 Nov 2011 15:15:10 +0000 (15:15 +0000)]
Followup r100305, don't try to do a batch query for an empty result (which would cause bug 32134)
Roan Kattouw [Wed, 2 Nov 2011 15:04:34 +0000 (15:04 +0000)]
Allow overrides of ResourceLoaderWikiModule::getDB() (introduced in r93760) to return null, needed for ForeignAPIGadgetRepo support (introduced in r101627)
Ariel Glenn [Wed, 2 Nov 2011 11:08:06 +0000 (11:08 +0000)]
clean up duplicated code, for fixme in r97178
Ariel Glenn [Wed, 2 Nov 2011 09:55:43 +0000 (09:55 +0000)]
corrections for fixme in r96486
Ariel Glenn [Wed, 2 Nov 2011 07:58:43 +0000 (07:58 +0000)]
export specified range of revisions (as stubs)
Sam Reed [Tue, 1 Nov 2011 23:48:09 +0000 (23:48 +0000)]
arrray -> array
Whitespace
Other documentation improvements
Comment out some unused code (which has a fixme left with it already)
Siebrand Mazeland [Tue, 1 Nov 2011 22:20:34 +0000 (22:20 +0000)]
Localisation updates from translatewiki.net.
Brion Vibber [Tue, 1 Nov 2011 22:08:33 +0000 (22:08 +0000)]
Bug 31933: fix 1.18 regression in Monobook sidebar: huge spacing between portlets on IE 7 and IE 8/9 in compatibility view.
Followup r74966: drop some lines ported from MonoBook's old rtl.css. The floats don't seem to be needed and were breaking IE 7. Also contained some IE-Mac-related code that seems unnecessary since we no longer support that ancient browser. :)
Tested in:
* Firefox 7 LTR <- no regression
* Firefox 7 RTL <- no regression
* IE 6 LTR <- no regression
* IE 6 RTL <- no regression
* IE 7 LTR <- fixes regression
* IE 7 RTL <- no regression
* IE 8 regular LTR <- no regression
* IE 8 regular LTR <- no regression
* IE 8 regular RTL <- no regression
* IE 8 compatibility LTR <- fixes regression
* IE 8 compatibility RTL <- no regression
* IE 9 regular LTR <- no regression
* IE 9 regular RTL <- no regression
* IE 9 compatibility mode LTR <- fixes regression
* IE 9 compatibility mode RTL <- no regression
Raimond Spekking [Tue, 1 Nov 2011 21:04:38 +0000 (21:04 +0000)]
Localisation updates for core and extension messages from translatewiki.net
Brion Vibber [Tue, 1 Nov 2011 20:17:04 +0000 (20:17 +0000)]
Bug 32125: remove stray $wgLivePasswordStrengthChecks leftover from a reverted feature
Followup to r86157 revert of r70520
Mark A. Hershberger [Tue, 1 Nov 2011 19:02:13 +0000 (19:02 +0000)]
Follow up r101151 — store hex representation of binary data in text
Mark A. Hershberger [Tue, 1 Nov 2011 18:49:38 +0000 (18:49 +0000)]
follow up r101496 — ^demon's comments re style
Mark A. Hershberger [Tue, 1 Nov 2011 18:32:05 +0000 (18:32 +0000)]
Meant to include these: feed the OCD: eol w/s
Mark A. Hershberger [Tue, 1 Nov 2011 18:31:38 +0000 (18:31 +0000)]
feed the OCD: eol w/s
Mark A. Hershberger [Tue, 1 Nov 2011 18:31:04 +0000 (18:31 +0000)]
Re-apply Vitaliy Filippov's patch from Bug #29283 with a modification
to make it work when no ACCEL cache is available.
Roan Kattouw [Tue, 1 Nov 2011 16:51:31 +0000 (16:51 +0000)]
Revert r101492, broken, see CR. Also revert followup r101496.
Alexandre Emsenhuber [Tue, 1 Nov 2011 16:41:36 +0000 (16:41 +0000)]
* Use WikiPage instead of Article since we only need to call doEdit()
* Pass the User object to doEdit() in addition to the override of $wgUser
Mark A. Hershberger [Tue, 1 Nov 2011 16:29:49 +0000 (16:29 +0000)]
followup 101492 — make it actually work.
Roan Kattouw [Tue, 1 Nov 2011 16:25:26 +0000 (16:25 +0000)]
Revert r101488, breaks parser test. Add a comment explaining why there is no caching currently being done, and document the fact that the hook is called a zillion times in hooks.txt
Alexandre Emsenhuber [Tue, 1 Nov 2011 16:19:44 +0000 (16:19 +0000)]
* Use local context to get messages
* Parse 'activeusers-noresult' and 'activeusers-intro' messages
Mark A. Hershberger [Tue, 1 Nov 2011 16:10:44 +0000 (16:10 +0000)]
Apply Vitaliy Filippov's patch from Bug #29283 with some small style mods
Alexandre Emsenhuber [Tue, 1 Nov 2011 16:09:27 +0000 (16:09 +0000)]
* Create all items in $nav_urls at the beginning so that we don't need to bother if they are set or not after that
* Don't show 'whatlinkshere', 'recentchangeslinked' and 'trackbacklink' items if OutputPage::isArticleRelated() is false
* Simplify the code to check if we show show user related items since the actual code is just to check if we have an User object
* Moved the showEmailUser() check to the same location as other related user checks
* Made Skin::showEmailUser() accept an User object instead of doing object -> int -> object conversion
* Check if the items are present in BaseTemplate::getToolbox() instead of 'notspecialpage' and removed some empty() or isset() checks since all of these items are now set
Mark A. Hershberger [Tue, 1 Nov 2011 16:07:09 +0000 (16:07 +0000)]
Style fixups
Antoine Musso [Tue, 1 Nov 2011 16:00:49 +0000 (16:00 +0000)]
space to tab
follow up r101480
Roan Kattouw [Tue, 1 Nov 2011 15:55:57 +0000 (15:55 +0000)]
Cache the result of User::getDefaultOptions(), so the UserGetDefaultOptions hook (introduced in r97365) is run only once rather than on every call. Discovered this issue after seeing an insane number of memcached fetches in my RL2 code, caused by a UserGetDefaultOptions hook accessing the Gadgets storage backend (which has a memc layer) and being called a zillion times due to this bug.
Alexandre Emsenhuber [Tue, 1 Nov 2011 15:45:52 +0000 (15:45 +0000)]
Some cleanup to ProtectionForm:
* When the user doesn't have permission to protect/unprotect the page, display a page with the title set to "View protection level for ...", then the permissions errors and the current protection levels. Except for the title, this was already the case for permissions errors, but when the database was read-only this was throwing an exception due to the change in OutputPage::readOnlyPage()
* Moved check for NS_MEDIAWIKI a level up, so that is it also catched on page save, and make the output in such case much more user friendly
* Changed edit token to be salted with array( 'protect', 'page name' ) (as for action=watch)
* Allow extensions to return false in ProtectionForm::save hook for consistency with other hooks and in that case fill the error message with 'hookaborted' if the extension does not provide an customised error message.
* Changed ProtectionForm::show() to accept an array of message name and its parameters in addition of a string with raw HTML and converted calls to use the new method. This also means that ProtectionForm::save can use this
* Removed the setSubtitle() call for the 'formerror' message in case of error since it's overriden by the "back link"
Antoine Musso [Tue, 1 Nov 2011 15:09:52 +0000 (15:09 +0000)]
(bug 32022) phpunit.php can not execute from another directory
Adds tests/phpunit to the include_path to let PHPUnit find out our files.
Also add to change the suite.xml paths from './path' to 'path' so
stream_resolve_path() correctly find out the suite by using include_path
Antoine Musso [Tue, 1 Nov 2011 14:21:42 +0000 (14:21 +0000)]
bug 32086 Special:Permanentlink require a parameter
When Special:Permanentlink was not given any parameter it assumed oldid=0
and redirected to the Main_Page. This make it throw an error page instead.
Antoine Musso [Tue, 1 Nov 2011 14:05:53 +0000 (14:05 +0000)]
bug 32100 installer complains about suhosin GET limit
Our logic was to warn whenever the suhosin GET limit was set.
This patch skip the warning if the limit is 1024 or more.
Also added 'qqq' message for 'config-suhosin-max-value-length'
Antoine Musso [Tue, 1 Nov 2011 13:14:59 +0000 (13:14 +0000)]
Fix GROUP BY clause in Special:MostLinkedTemplates
This only cause a trouble for Microsoft SQL which require the column to be
referenced by their input name, not the output one (AS alias). Might trigger
on Oracle database.
This is part of bug 31913: 'GROUP BY clauses incorrect in SQL generated
for special pages'
Need backport in REL1_18
Amir E. Aharoni [Tue, 1 Nov 2011 12:04:54 +0000 (12:04 +0000)]
Updated the title of Special:Linksearch to "External links search" and updated qqq documentation accordingly.
Antoine Musso [Tue, 1 Nov 2011 09:28:51 +0000 (09:28 +0000)]
checkbox could not be checked through URL parameter
On [[Special:EmailUser]], someone ought to be able to pass the checkboxes
names to overrides the default. As an example from bug 31770, the email
user page has a checkbox wpCCMe which let the user as for a copy of the
email being send. This is a user preference.
One change that checkbox state by appending ?wpCCMe=<boolean> ie:
Special:EmailUser/Hashar?wpCCMe=0
Special:EmailUser/Hashar?wpCCMe=1
The logic added in r84814 could have allowed checkboxes to be overriden
for GET form. Unfortunately, HTMLForm is mostly with the default POST.
Hence, when appending the query parameter, we would never honor it!
The fix is to still unconditionally look at the value if the form was
correctly submitted (ie has wpEditToken) and additionaly whenever
there is a checkbox name appearing in the query (wherever it is false
or true, hence the use of getVal()).
This is a regression in REL1_18. I dont think it deserves a release notes
since it is not fixing anything compared to 1.17.
Bug fixed:
==========
* (bug 31770) Allow URL parameter wpCCMe on Special:EmailUser
* (bug 30909) URL parameters for checkboxes in Special:Block no longer work
Sam Reed [Tue, 1 Nov 2011 07:34:56 +0000 (07:34 +0000)]
Followup r101449, just changing forceRecache isn't actually enough
Add a method to set force to true on the options array to RebuildLocalisationCache
Sam Reed [Tue, 1 Nov 2011 06:17:40 +0000 (06:17 +0000)]
Bug 30120 - Interwiki links error after upgrade
Per Emufarmers, in wgDBTableOptions replace TYPE with ENGINE
Should at least stop the errors. Do need to do something in the updaters (can we have some "delayed messages" or something?
Needed more for the CLI updater/update.php, as putting it in the middle of the updaters is just pointless, as it won't be seen. Needs to be at the end...
Doing it in the webupdater is saner/easier, as it'll be displayed
Certainly, this does suffice for the moment
Fine on MySQL 4, as it accepts {TYPE|ENGINE}
Sam Reed [Tue, 1 Nov 2011 05:33:57 +0000 (05:33 +0000)]
Followup r101449, bit better message output
Sam Reed [Tue, 1 Nov 2011 05:23:08 +0000 (05:23 +0000)]
Bug 30934 - Warning: Invalid argument supplied for foreach() in /languages/Language.php on line 423
Doing a run of doRebuildLocalisationCache on each run of update.php
Robin Pepermans [Tue, 1 Nov 2011 00:17:37 +0000 (00:17 +0000)]
Follow-up r96261: remove the !important again, and then do the script defaults differently (so that those css rules are only for the default and don't affect when the user selected "never" or "always" underline)
Aaron Schulz [Tue, 1 Nov 2011 00:16:35 +0000 (00:16 +0000)]
FR r101442: forgot to replace another getBlocker()->getName() with getByName()
Aaron Schulz [Tue, 1 Nov 2011 00:11:53 +0000 (00:11 +0000)]
(bug 31739) Made Block code support ipb_by = 0 convention with for foreign users again, as it did pre 1.18. The byText param has been restored in the Block() constructor as CentralAuth still uses it. Some callers have been updated to reflect the fact that getBlocker() may be a local User or a string username.
Ideally, we might have a ForeignUser class and a generic User interface...but this will do for now.
Aaron Schulz [Mon, 31 Oct 2011 23:37:14 +0000 (23:37 +0000)]
Made some callers of getBlocker()->getName() just use getByName()
Platonides [Mon, 31 Oct 2011 22:24:26 +0000 (22:24 +0000)]
Another old, uncommitted, test (passing on NtPP)
Platonides [Mon, 31 Oct 2011 22:21:46 +0000 (22:21 +0000)]
Template as parameter name.
I had this test pending longtime, as it is failing in NativePreprocessor.
Krinkle [Mon, 31 Oct 2011 22:20:25 +0000 (22:20 +0000)]
jquery.tablesorter: More selector fixes
* Limit to rowSpans (td or th) within tableRows that are a direct child of this table's tBody
* Follows-up r99092, r101417
Krinkle [Mon, 31 Oct 2011 22:12:26 +0000 (22:12 +0000)]
jquery.tablesorter: Selector fixes
* Various fixes to limit the selector to the current table and avoid selecting similar elements from other, nested, tables.
* Should fix bug 32049
* Follows-up r97150 and likely some other refactoring commits
dépôts / lhc / web / wiklou.git / log