jenkins-bot [Tue, 21 Nov 2017 15:22:23 +0000 (15:22 +0000)]
Merge "Display MAX_CHARS / MAX_BYTES in ApiSandbox"
jenkins-bot [Tue, 21 Nov 2017 15:21:26 +0000 (15:21 +0000)]
Merge "Add string length limits"
WMDE-Fisch [Tue, 7 Nov 2017 11:44:53 +0000 (12:44 +0100)]
Add localised tooltips for moved paragraph indicators
Bug: T166882
Change-Id: I71600acc19319f7b5781b42dc88cd5e8c65c29a0
Kunal Mehta [Sat, 4 Nov 2017 20:15:26 +0000 (13:15 -0700)]
DifferenceEngine: Improve cache invalidation
Invalidate the diff cache if the engine producing the diff changes, or
if a configuration setting that controls the diff output changes. This
is probably what most users expect, that changing the configuration will
result in a change for diffs that may have already been viewed.
For wikidiff2 specifically, a change in version or
$wgWikiDiff2MovedParagraphDetectionCutoff will invalidate the cache.
Refactor engine detection and sanity-checking into a private getEngine()
function.
As part of this getDiffBodyCacheKey() was deprecated, and subclasses
should implement getDiffBodyCacheKeyParams() instead. Drop the
deprecated and unused MW_DIFF_VERSION constant while we're at it, and
bump DIFF_VERSION since we're already changing the cache key format.
Bug: T180043
Change-Id: I4e386ca05bd2a2fb54208d760c131eb42e3a72ab
Gergő Tisza [Mon, 13 Nov 2017 00:37:21 +0000 (00:37 +0000)]
Display MAX_CHARS / MAX_BYTES in ApiSandbox
Change-Id: I0f15afe6656765e92582e4e4cead714b31987b8a
Gergő Tisza [Sun, 12 Nov 2017 09:51:34 +0000 (09:51 +0000)]
Add string length limits
Adds two new ApiBase::getAllowedParams() keys:
PARAM_MAX_BYTES and PARAM_MAX_CHARS, to set a length
limit for a (string-like) parameter.
This makes it easy to document and enforce database
field length limits (where relying on the database
would either result in unfriendly error messages or
silent truncation, depending on DB settings) and
also exposes them in structured form so API clients
can verify the length without doing roundtrips.
Change-Id: I2e784972d7e11cad79fdef887bbcde297dbd9ce0
Aaron Schulz [Tue, 21 Nov 2017 00:04:38 +0000 (16:04 -0800)]
objectcache: Make MemcachedBagOStuff::makeKeyInternal always have a key class
Even if a key is too long and shortened, it should still have some key class.
Change-Id: I006b6b03ad1302e9e49362bbd051332bc6105837
jenkins-bot [Mon, 20 Nov 2017 22:47:54 +0000 (22:47 +0000)]
Merge "Special:UserRights: Prevent FOUC on loading"
jenkins-bot [Mon, 20 Nov 2017 22:27:57 +0000 (22:27 +0000)]
Merge "Crimean Tatar Transliteration"
tjones [Mon, 31 Jul 2017 22:35:39 +0000 (18:35 -0400)]
Crimean Tatar Transliteration
This is a first pass at Latin/Cyrillic translitertion for Crimean
Tatar (crh).
Includes transliteration tables, prefix/suffix mappings, regex
mappings, and exceptions lists for words and abbreviations.
Regularize CRH language name in messages/* files.
Fix "varient" typos in qqq.json.
Add unit tests for CRH transliteration.
Bug: T23582
Change-Id: I424703f99adf837f6217872b882d1ea26bfdd068
Translation updater bot [Mon, 20 Nov 2017 20:56:22 +0000 (21:56 +0100)]
Localisation updates from https://translatewiki.net.
Change-Id: I2541976a8c5c079f0e11d84f843aff0ffd540dab
jenkins-bot [Mon, 20 Nov 2017 20:51:26 +0000 (20:51 +0000)]
Merge "objectcache: fix some makeKey/makeGlobalKey IDEA warnings"
jenkins-bot [Mon, 20 Nov 2017 20:48:47 +0000 (20:48 +0000)]
Merge "Fix some broke cache key generations"
jenkins-bot [Mon, 20 Nov 2017 20:41:16 +0000 (20:41 +0000)]
Merge "objectcache: make sure variant keys are namespaced as keys should be"
Aaron Schulz [Mon, 20 Nov 2017 07:27:34 +0000 (23:27 -0800)]
objectcache: make sure variant keys are namespaced as keys should be
Change-Id: I02597d6dbc2febace25ef6d1981f69dc704efe8d
Moriel Schottlender [Mon, 20 Nov 2017 20:07:43 +0000 (12:07 -0800)]
RCFilters: Followup I08db859c571: Fix typo in "namespaces"
Bug: T180863
Change-Id: I1b65ed5dfceb99bea98cbcadea11ddab367c6ca7
Fomafix [Mon, 20 Nov 2017 07:32:41 +0000 (08:32 +0100)]
Special:UserRights: Prevent FOUC on loading
Implement the show/hide based on the group checkbox with CSS instead of
JavaScript.
Hide the expiry input field initially with CSS instead of JavaScript.
Change-Id: I33a4b0a13dab92156aeea96529573ad49fa2d616
addshore [Mon, 20 Nov 2017 18:52:46 +0000 (18:52 +0000)]
Deprecate access to TitleValue properties
This switches from private to protected access of TitleValue
properties, and marks them as deprecated instead as per the
deprecation policy.
Partial revert of I3d8315ade6aa70bda43d90b0b32b730d8c9cbd2e
Change-Id: Ie290b6a84635ebb5865e82ae24fe4b27224e601a
jenkins-bot [Mon, 20 Nov 2017 18:20:45 +0000 (18:20 +0000)]
Merge "RCFilters: Only apply excluded label to namespace items"
jenkins-bot [Mon, 20 Nov 2017 16:26:48 +0000 (16:26 +0000)]
Merge "Tests for TitleValue::__toString"
jenkins-bot [Mon, 20 Nov 2017 16:15:38 +0000 (16:15 +0000)]
Merge "Set wgCommentTableSchemaMigrationStage for some RevisionTests"
jenkins-bot [Mon, 20 Nov 2017 11:50:13 +0000 (11:50 +0000)]
Merge "Move styles for Special:UserRights to separate style module"
addshore [Mon, 20 Nov 2017 09:39:55 +0000 (09:39 +0000)]
Tests for TitleValue::__toString
Change-Id: Ibb157b8e6e4fdc75bdb27ecaa4a552304465c4b2
addshore [Mon, 20 Nov 2017 09:33:28 +0000 (09:33 +0000)]
Set wgCommentTableSchemaMigrationStage for some RevisionTests
Bug: T180922
Change-Id: Ib88b53e36a66042b3c177cae0663040b4fb14a4c
Reedy [Sun, 19 Nov 2017 18:36:34 +0000 (18:36 +0000)]
Move styles for Special:UserRights to separate style module
Bug: T180914
Change-Id: I0e3cf5e3d69f32a30e5fe6dc0d2fd964c4f5720c
Aaron Schulz [Mon, 20 Nov 2017 05:30:45 +0000 (21:30 -0800)]
Fix some broke cache key generations
Change-Id: Ib724fe0cfa866351caee89d3451b6c4b7b6a4578
Translation updater bot [Sun, 19 Nov 2017 20:55:12 +0000 (21:55 +0100)]
Localisation updates from https://translatewiki.net.
Change-Id: I571e77344da60ffec62d52ea79231ff94562c424
Aaron Schulz [Sat, 18 Nov 2017 20:39:21 +0000 (12:39 -0800)]
objectcache: fix some makeKey/makeGlobalKey IDEA warnings
Change-Id: I5fefda9676a685167326c63c34b2b7df8be2e7cd
jenkins-bot [Sat, 18 Nov 2017 21:44:15 +0000 (21:44 +0000)]
Merge "@since tags & private class properties for TitleValue"
jenkins-bot [Sat, 18 Nov 2017 21:42:34 +0000 (21:42 +0000)]
Merge "Add __toString method to LinkTarget interface"
Translation updater bot [Sat, 18 Nov 2017 20:53:19 +0000 (21:53 +0100)]
Localisation updates from https://translatewiki.net.
Change-Id: If8cc0f3abd7cde509909de175a32c9492f9388a6
addshore [Sat, 18 Nov 2017 17:34:41 +0000 (17:34 +0000)]
@since tags & private class properties for TitleValue
Change-Id: I3d8315ade6aa70bda43d90b0b32b730d8c9cbd2e
addshore [Sat, 18 Nov 2017 17:34:10 +0000 (17:34 +0000)]
Add __toString method to LinkTarget interface
Change-Id: I4f12e3175a308f88e240db01cd4e91e78b283057
daniel [Thu, 16 Nov 2017 19:44:44 +0000 (20:44 +0100)]
Introduce the UserIdentity interface.
This provides a narrow view on a user identity, providing access to
ID and name.
This has been extracted from I140f43a6fb443b for re-use with Actors,
on Anomie's request.
Change-Id: Ief00db5ce382537c5bf992159eae6baf096ae4be
addshore [Sat, 18 Nov 2017 18:22:24 +0000 (18:22 +0000)]
[MCR] tests for Revision::getQueryInfo
Bug: T180210
Change-Id: I5bdb5eed853e22bacd6b4c2546343e9d0f2d8c89
addshore [Sat, 18 Nov 2017 18:08:45 +0000 (18:08 +0000)]
[MCR] tests for Revision::getArchiveQueryInfo
Bug: T180210
Change-Id: Icb016be8d69dfdfa83f44bbc4fb259b5beb30678
addshore [Sat, 18 Nov 2017 17:49:22 +0000 (17:49 +0000)]
[MCR] Readd various field & cond method tests for Revision
This is a partial revert of:
Idcfd15568489d9f03a7ba4460e96610d33bc4089
which removed these tests.
Bug: T180210
Change-Id: Ib0617ee0a7bd4391ed25415b44a8ed077a985eaa
jenkins-bot [Sat, 18 Nov 2017 13:00:39 +0000 (13:00 +0000)]
Merge "Remove box-shadow from preference panels for ooui-apex"
jenkins-bot [Sat, 18 Nov 2017 03:54:22 +0000 (03:54 +0000)]
Merge "Fix RemexCompatMunger infinite recursion"
jenkins-bot [Sat, 18 Nov 2017 02:08:27 +0000 (02:08 +0000)]
Merge "SwiftFileBackend::resolveContainerPath() check the proper length"
Moriel Schottlender [Sat, 18 Nov 2017 00:57:40 +0000 (16:57 -0800)]
RCFilters: Only apply excluded label to namespace items
Bug: T180863
Change-Id: I08db859c571b82ed4e4793d97b99100875ff1e23
Aaron Schulz [Fri, 17 Nov 2017 23:27:49 +0000 (15:27 -0800)]
SwiftFileBackend::resolveContainerPath() check the proper length
The length sanity check should use this instead of urlencode()
see it is rawurlencode() that is actually used.
Change-Id: I5632e30c14c8ab27c8324c3e31311ca8bff7c162
Ed Sanders [Fri, 17 Nov 2017 22:02:58 +0000 (22:02 +0000)]
Follow-Up Iae63b6994: Add missing editfont dependency
Change-Id: I606a81576baf312891f3bb12f575892b00bca823
Ed Sanders [Fri, 17 Nov 2017 21:23:16 +0000 (21:23 +0000)]
Remove box-shadow from preference panels for ooui-apex
Change-Id: I7d42c75053b29cb634b18bae9e06e6b28ae1e967
Translation updater bot [Fri, 17 Nov 2017 21:03:49 +0000 (22:03 +0100)]
Localisation updates from https://translatewiki.net.
Change-Id: Iabf53d3c29d23ea48d35c2a45d88b2b3d937d5c6
addshore [Wed, 15 Nov 2017 10:37:29 +0000 (10:37 +0000)]
Revision::userCanBitfield test fallback to $wgUser
Bug: T180210
Change-Id: Ibda256d6fdd8762e0e90748d71034979e2de106f
Ed Sanders [Fri, 17 Nov 2017 17:15:44 +0000 (17:15 +0000)]
Work around Firefox 57 attribute parsing regression
Bug: T180138
Change-Id: I554d5adf88c448db64e6f135e19ee76a4ec28493
jenkins-bot [Fri, 17 Nov 2017 15:53:52 +0000 (15:53 +0000)]
Merge "Expose string->bool conversion as function"
Tim Starling [Fri, 17 Nov 2017 11:15:59 +0000 (22:15 +1100)]
Fix RemexCompatMunger infinite recursion
When TreeBuilder requests reparenting of all child nodes of a given
element, we do this by removing the existing child nodes, and then
inserting the proposed new parent under the old parent. However, when a
p-wrap diversion is in place, the insertion of the new parent is
diverted into the p-wrap, and the p-wrap then becomes a child of the new
parent, causing a reference loop, and ultimately infinite recursion in
Serializer.
Instead, divert the entire reparent request to the p-wrap, so that the
new parent is a child of the p-wrap. This makes sense since the new
parent is always a formatting element. The only caller of
reparentChildren(), apart from proxies, is AAA step 17, which reparents
children under the formatting element cloned from the AFE list.
Left in some debug code for next time.
Bug: T178632
Change-Id: Id77d21d99748e94c064ef24c43ee0033de627b8e
jenkins-bot [Thu, 16 Nov 2017 23:05:43 +0000 (23:05 +0000)]
Merge "Preferences: Improve visual appearance by “unboxing” sections"
jenkins-bot [Thu, 16 Nov 2017 23:03:20 +0000 (23:03 +0000)]
Merge "Cleanup, removed space"
Volker E [Wed, 15 Nov 2017 00:47:52 +0000 (16:47 -0800)]
Preferences: Improve visual appearance by “unboxing” sections
Instead let's work with whitespace, which also saves virtual space and
makes the appearance not as jarring.
Bug: T180538
Depends-on: I39088107e6ab07399f9826dd925df9e1b8dda006
Change-Id: I24d21eb3c0d188004dacbce8a9bc1ac3ad7e2a8f
Translation updater bot [Thu, 16 Nov 2017 21:34:02 +0000 (22:34 +0100)]
Localisation updates from https://translatewiki.net.
Change-Id: I20f8422f709fe37f212b65f2e995db30776009eb
jenkins-bot [Thu, 16 Nov 2017 20:34:31 +0000 (20:34 +0000)]
Merge "Use Remex in Sanitizer::stripAllTags()"
jenkins-bot [Thu, 16 Nov 2017 17:31:44 +0000 (17:31 +0000)]
Merge "OOUIHTMLForm: Prevent duplicate FieldsetLayout wrapping"
jenkins-bot [Thu, 16 Nov 2017 17:22:39 +0000 (17:22 +0000)]
Merge "HTMLMultiSelectField: Fix OOUI\CheckboxMultiselectInputWidget to be infusable again"
jenkins-bot [Thu, 16 Nov 2017 17:05:14 +0000 (17:05 +0000)]
Merge "HTMLRadioField: Do not automatically infuse our RadioSelectInputWidgets"
Bartosz Dziewoński [Thu, 16 Nov 2017 16:14:29 +0000 (17:14 +0100)]
HTMLRadioField: Do not automatically infuse our RadioSelectInputWidgets
This is really a workaround for an issue in a completely different place:
JS RadioSelectInputWidget internally uses `<input type="hidden">`
rather than real radio buttons, which does not work correctly with the
code in mediawiki.special.preferences.confirmClose.js. Ideally we would
change RadioSelectInputWidget to not do such weird things.
However, I think this is actually a good thing to do in general.
From the user's perspective, PHP RadioSelectInputWidget and JS
RadioSelectInputWidget look and behave the same, so there's no reason
to infuse and rebuild them.
This behavior was implemented in
f50cee1375201a5d3fd76c0c262cfc7e66bd5d42
in which unfortunately I did not document the reason for it. For other
fields it makes obvious sense (the JS widgets have improvements like
autocompletion, or at least look "pretty"), but I have no idea why
I did it for this one.
Bug: T180643
Change-Id: I53e50f8cda39466b2396b374e642c154487888bb
Bartosz Dziewoński [Thu, 16 Nov 2017 10:24:47 +0000 (11:24 +0100)]
OOUIHTMLForm: Prevent duplicate FieldsetLayout wrapping
The code in formatSection() assumed it was only called for the
toplevel section (the whole form), while it's actually called
for every subsection too. I think it was written before we added
support for subsections in OOUIHTMLForm.
Move code for toplevel section wrapping to wrapForm().
As a bonus, this also fixes display of custom headers and error
or warning messages for forms with subsections.
Bug: T180535
Change-Id: I6a88184d302a951be78387490404137acde3fa1a
WMDE-Fisch [Thu, 16 Nov 2017 12:27:11 +0000 (13:27 +0100)]
Fixed hover circle for timeless skin
Bug: T180663
Change-Id: I5112636bcfae6f41e86ccc29524ccf5c9e2a6004
Bartosz Dziewoński [Thu, 16 Nov 2017 10:51:19 +0000 (11:51 +0100)]
HTMLMultiSelectField: Fix OOUI\CheckboxMultiselectInputWidget to be infusable again
Regression from
5a113417e5af9d0d0dbed63429649a9780784d45.
Bug: T180677
Change-Id: Id1b0ebe9d9a56a76d73deb2b4d17213ae5e45a04
jenkins-bot [Thu, 16 Nov 2017 10:04:35 +0000 (10:04 +0000)]
Merge "Hide empty OOUI FieldsetLayout headers"
Huji Lee [Sun, 12 Nov 2017 01:44:00 +0000 (20:44 -0500)]
UserGroupsChanged hook should specify the performer of the change
Otherwise, there will be a unit-testing error when Echo is enabled
Bug: T180292
Change-Id: Ibc185c82ad2a03e06e5727a633e6ab6bccce3345
jenkins-bot [Thu, 16 Nov 2017 01:33:21 +0000 (01:33 +0000)]
Merge "Move Sanitizer.php to includes/parser/"
jenkins-bot [Thu, 16 Nov 2017 01:32:46 +0000 (01:32 +0000)]
Merge "SanitizerTest: Add tests for stripAllTags"
Roan Kattouw [Tue, 14 Nov 2017 22:22:31 +0000 (14:22 -0800)]
Use Remex in Sanitizer::stripAllTags()
Using a real HTML tokenizer fixes bugs when < or > appear in attribute
values. The old implementation used delimiterReplace(), which didn't
handle this case:
> print Sanitizer::stripAllTags( '<p data-foo="a<b>c">Hello</p>' );
c">Hello
We also can't use PHP's built-in strip_tags() because it doesn't handle
<?php and <? correctly:
> print strip_tags('1<span class="<?php">2</span>3');
1
> print strip_tags('1<span class="<?">2</span>3');
1
Bug: T179978
Change-Id: I53b98e6c877c00c03ff110914168b398559c9c3e
Roan Kattouw [Wed, 15 Nov 2017 20:44:48 +0000 (12:44 -0800)]
Move Sanitizer.php to includes/parser/
Change-Id: Id08d91c747ec77d715459b89b03eee247ccd4e1b
Roan Kattouw [Tue, 14 Nov 2017 22:16:14 +0000 (14:16 -0800)]
SanitizerTest: Add tests for stripAllTags
Bug: T179978
Change-Id: I9776cfd51b1b3ec772d4216168fbe466f48f5892
Translation updater bot [Wed, 15 Nov 2017 20:54:46 +0000 (21:54 +0100)]
Localisation updates from https://translatewiki.net.
Change-Id: I746f9a0b5a9ffcfbe198b4d222e476b169bad2dc
jenkins-bot [Wed, 15 Nov 2017 19:27:05 +0000 (19:27 +0000)]
Merge "Preferences: Remove unwise caching of Preferences::getPreferences()"
jenkins-bot [Wed, 15 Nov 2017 19:15:49 +0000 (19:15 +0000)]
Merge "Revert "RCFilters: Remove excluded params from URL""
Catrope [Wed, 15 Nov 2017 18:34:55 +0000 (18:34 +0000)]
Revert "RCFilters: Remove excluded params from URL"
Breaks limit and days selection by also removing these
params from the AJAX request URL
This reverts commit
b8a10e6dcf00da3519ccb9e43d1c2ce0db422557.
Bug: T180577
Change-Id: Ifc6fa8cde8ffce0ac79fc3a2db55291bc2a84e20
Brad Jorsch [Wed, 15 Nov 2017 16:44:53 +0000 (11:44 -0500)]
Clean up RELEASE-NOTES-1.31
A few recent changes introduced entries with bad spacing, and there's no
need for the ellipsis placeholder entries in a section once real entries
have been added.
Change-Id: Ia1f4aaa63c1fc859a5f6b0ec7726d98f81df2c05
Arlo Breault [Wed, 15 Nov 2017 14:42:11 +0000 (09:42 -0500)]
Sync up with Parsoid parserTests.txt
This now aligns with Parsoid commit
3048db625dca69d8a89cde4cbabee0105f2975f5
Change-Id: I2bf07d6582367f25d9f69712dc4350982b627851
jenkins-bot [Wed, 15 Nov 2017 14:21:33 +0000 (14:21 +0000)]
Merge "Remove $wgAuth usage from wrapOldPasswords.php"
jenkins-bot [Wed, 15 Nov 2017 11:13:22 +0000 (11:13 +0000)]
Merge "Preferences: Show preview of edit fonts in edit font selector"
Ed Sanders [Tue, 7 Nov 2017 15:33:03 +0000 (15:33 +0000)]
Preferences: Show preview of edit fonts in edit font selector
Change-Id: Iae63b69940485165b660f51deb864979aefb8cd0
Bartosz Dziewoński [Mon, 13 Nov 2017 16:42:04 +0000 (17:42 +0100)]
Preferences: Remove unwise caching of Preferences::getPreferences()
The result of this function depends on the $user and $context
parameters (e.g. it includes the username from the user, and
localisation messages the language from the context). However,
both of them would be ignored if the result was cached, even
if calling with a different $user or $context.
Rather than make this more complicated just remove the caching.
This is not a hot code path: this function is not called at all
on normal page views, it's called just once when viewing
preferences, and at most twice when saving them.
Change-Id: I92390120a16448383a25e9ba2dd35a434a2f21bf
Bartosz Dziewoński [Mon, 13 Nov 2017 18:42:33 +0000 (19:42 +0100)]
ApiOptionsTest: Do not use ->at()
Quoting PHPUnit docs:
The $index parameter for the at() matcher refers to the index,
starting at zero, in all method invocations for a given mock object.
Exercise caution when using this matcher as it can lead to brittle
tests which are too closely tied to specific implementation details.
Indeed these test cases would break horribly with unintuitive error
messages ("Mocked method does not exist") if anything in preferences
or API code called any additional methods on the mocked user. For
example, it relied on the caching in Preferences::getPreferences(),
which is being removed in I92390120a16448383a25e9ba2dd35a434a2f21bf.
I'm pretty sure all that matters here is that all the setOption()
calls with different arguments happen, so let's test just that.
Change-Id: I30a814151a006e5f147eebb918344049807b2b97
jenkins-bot [Wed, 15 Nov 2017 10:38:42 +0000 (10:38 +0000)]
Merge "MWExceptionRenderer: Wrap error message in a paragraph"
jenkins-bot [Wed, 15 Nov 2017 10:17:32 +0000 (10:17 +0000)]
Merge "Treat langtags in SVG switch case-insensitively"
jenkins-bot [Wed, 15 Nov 2017 07:12:06 +0000 (07:12 +0000)]
Merge "Updated phpunit/phpunit from v4.8.35 to v4.8.36 in correct RELEASE-NOTES"
Stanislav Malyshev [Tue, 31 Oct 2017 21:01:02 +0000 (14:01 -0700)]
Expose string->bool conversion as function
There is code in several places in extensions which converts
setting or parameter string (such as "true", "yes", "false", "no")
to boolean. Since we already have the code that does in global
functions in wfStringToBool(), it makes sense to expose this code
and reuse it.
Change-Id: I88d98b012ff4bf14fd64a05a9135a6e75cf2d4e7
Brian Wolff [Wed, 15 Nov 2017 06:56:38 +0000 (06:56 +0000)]
Follow-up I077d30c50 fix phpcs error
Change-Id: I28cb7060d6149d96ceb0dcad7e2bff2ed3434411
Brian Wolff [Wed, 15 Nov 2017 05:34:10 +0000 (05:34 +0000)]
Fix langauge converter parser test with self-close tags
This fixes an issue in
f21f3942 where if there was an html
element with an alt or title attribute containing an <
entity, an ascii EOT control character (0x04) may become
inserted into the text if language converter was enabled.
Due to a really old bug in language converter, self-closed tags
got turned into non-self closed tags. However due a different
bug which was fixed in
f21f3942 this code path was rarely taken
so nobody noticed until now.
Follow-up Idbc45cac12
Bug: T180552
Change-Id: I077d30c50fcb419837fef937d27caca307153d2d
Brian Wolff [Wed, 15 Nov 2017 04:37:15 +0000 (04:37 +0000)]
Follow-up
5e56f01f1. Rebase failure.
Bug: T180551
Change-Id: I07a8c2555f08c6c731cd7e1917be451ec40d4d0c
Brian Wolff [Wed, 15 Nov 2017 03:42:38 +0000 (03:42 +0000)]
Fix LanguageSrTest for language converter
Language converter seems to maintain state between parser
invocations. Use a more unique string for the test.
This is a follow-up to
98b6635895.
Bug: T180551
Change-Id: I0259b378549f7d9778c27c4bc3efd8d277893f8c
Reedy [Wed, 15 Nov 2017 03:45:30 +0000 (03:45 +0000)]
Ignore long git hashes for eval-stdin.php
This reverts commit
5b205725842edbae6ccde5a13baf20c2eb3ec4ae.
Change-Id: Ib47ffed1174628959b9ef537370ff88f8bcb22f1
Reedy [Wed, 15 Nov 2017 03:37:27 +0000 (03:37 +0000)]
Fix phpcs issues from LanguageConverter patches
Change-Id: I34e57c90ffd40fbd9f8afe3c57dd73fa7f655841
Brian Wolff [Thu, 11 Feb 2016 22:08:03 +0000 (17:08 -0500)]
SECURITY: Handle -{}- syntax in attributes safely
Previously, if one had an attribute with the contents
"-{}-foo-{}-", foo would get replaced by language converter as if
it wasn't in an attribute. This lead to an XSS attack.
This breaks doing manual conversions in url href's (or any
other attribute that goes through an escaping method
other than Sanitizer's). e.g. http://{sr-el:foo';sr-ec:bar}.com
won't work anymore. See also T87332
Bug: T119158
Change-Id: Idbc45cac12c309b0ccb4adeff6474fa527b48edb
Brian Wolff [Sun, 24 Jan 2016 10:29:10 +0000 (05:29 -0500)]
SECURITY: XSS in langconverter when regex hits pcre.backtrack_limit
Adjust regexes for what not to convert to avoid backtracking by
preferring possesive quantifiers
Add check that we really have matched to the end of the string, and
log error if the regex hits some sort of error preventing the
entire string from being matched. Should the regex not match to the
end, then language conversion is disabled for the string.
Bug: T124404
Change-Id: I4f0c171c7da804e9c1508ef1f59556665a318f6a
Brad Jorsch [Tue, 14 Nov 2017 16:17:02 +0000 (11:17 -0500)]
SECURITY: Fix rebase error in
4d38a489
The fix for T125177 from
F4932228 was incorrectly rebased when it was
applied to master as
4d38a489, causing the bug to not actually be fixed.
Bug: T180488
Change-Id: Ie6b87ef2373369987c112c19903c99afb789c1ff
Reedy [Wed, 15 Nov 2017 03:19:00 +0000 (03:19 +0000)]
Shorten git hashes for eval-stdin.php
Ping I5b838686ede9764083c52853cc05c52ea72739df
Change-Id: Ie80a068507444721f8ffdbdc1867555338489283
Kunal Mehta [Sat, 11 Nov 2017 00:53:24 +0000 (16:53 -0800)]
SECURITY: update.php: Remove eval-stdin.php if necessary
If phpunit's eval-stdin.php file exists and is one of the vulnerable
versions, delete it when running update.php as most people should run
that when updating to a new release. If the unlink() call fails, we'll
warn the user but continue with update.php processing and hope they've
mitigated it in some other way.
Bug: T180231
Change-Id: I5b838686ede9764083c52853cc05c52ea72739df
Reedy [Wed, 15 Nov 2017 01:55:22 +0000 (01:55 +0000)]
Add missing ComposerVendorHtaccessCreator class to autoload.php
Change-Id: Ia70324acf3db2df50c6629d705c2c4728c38aaaa
Follow-up: I2cf6541750c90b5708d7cf5f81b914ae2d9d46d1
Kunal Mehta [Sat, 11 Nov 2017 00:05:13 +0000 (16:05 -0800)]
SECURITY: Create a .htaccess in /vendor after composer runs
The /vendor directory does not need to be web accessible, and to reduce
attack surface, it should not be web accessible. We can use the
post-install-cmd and post-update-cmd hooks to create a .htaccess after
the user has run "composer install" or "composer update". On the first
run of composer, this hook will be invoked twice due to the composer
merge plugin.
If the htaccess file already exists, this hook won't do anything.
Bug: T180237
Change-Id: I2cf6541750c90b5708d7cf5f81b914ae2d9d46d1
Brian Wolff [Sun, 24 Sep 2017 00:57:05 +0000 (00:57 +0000)]
SECURITY: Ensure Message::rawParams can't lead to XSS
If you used wfMessage( 'foo' )->rawParams( 'bar"baz' )
there's a possibility of leading to xss, if the foo
message has a $1 in an attribute, as the quote characters
may end the attribute.
To prevent that, we convert $1 to $'"1 for after parameters,
so if any of them end up in attributes, the attribute escaping
will break the parameter name, preventing substitution.
This would of course break if someone intentionally inserted
a raw parameter into an attribute, but that's silly and I
don't think we should allow that.
This is similar to the parser strip marker issue.
Bug: T176247
Change-Id: If83aec01b20e414f9c92be894f145d7df2974866
Gergő Tisza [Wed, 15 Nov 2017 01:58:49 +0000 (17:58 -0800)]
Fix tests for I7afaa955a4b393ef00b11e420709bd62b84fbc71
Change-Id: Id5a2a9768dc64db5259a5988d645e664163ba234
Brian Wolff [Mon, 13 Nov 2017 16:02:50 +0000 (16:02 +0000)]
SECURITY: Do not reveal if user exists during login failure
This is meant for private wikis where the list of users may
be secret. It is only meant to prevent trivial enumeration
of usernames. It is not designed to prevent enumeration
via timing attacks.
Bug: T134100
Change-Id: I7afaa955a4b393ef00b11e420709bd62b84fbc71
Brad Jorsch [Fri, 26 Feb 2016 22:46:07 +0000 (17:46 -0500)]
SECURITY: API: Avoid some silliness with browser-guessed filenames
If someone is both dumb enough to blindly save an API response and to
then execute the resulting file, this can be used to attack their
computer.
We can mitigate this by disallowing PATH_INFO in api.php URLs (because
we don't make any use of them anyway) and by setting a sensible filename
using a Content-Disposition header so the browser won't go guessing at
the filename based on what is in the URL.
Issue reported by: Abdullah Hussam
Bug: T128209
Change-Id: I8526f5cc506c551edb6138d68450b6acea065e93