jenkins-bot [Fri, 20 Jul 2018 15:13:48 +0000 (15:13 +0000)]
Merge "Send API timing to statsd for all requests, not just writes"
Gergő Tisza [Mon, 28 May 2018 16:25:36 +0000 (16:25 +0000)]
Add maintenance script for resetting login/signup throttle
Bug: T194506
Change-Id: I6dda5f516ba47a452e438946379fcb696c8c87f9
jenkins-bot [Fri, 20 Jul 2018 10:51:35 +0000 (10:51 +0000)]
Merge "filebackend: let any specific backend value of "readOnly" override the default"
jenkins-bot [Fri, 20 Jul 2018 10:30:37 +0000 (10:30 +0000)]
Merge "JobQueueGroup: Allow readOnlyReason to be specified per JQ type"
jenkins-bot [Fri, 20 Jul 2018 10:21:48 +0000 (10:21 +0000)]
Merge "Install cache/integration-tests as require-dev"
addshore [Wed, 18 Jul 2018 10:47:27 +0000 (12:47 +0200)]
Install cache/integration-tests as require-dev
Bug: T199440
Change-Id: I1e50d7202a5d02088319ec9e90e0136519cba1ce
Max Semenik [Fri, 20 Jul 2018 06:54:49 +0000 (23:54 -0700)]
GlobalFunctions: use a closure instead of polluting global namespace
Change-Id: Ia587008de75a8bde90ad45d4932284f0b6401845
Tim Starling [Fri, 20 Jul 2018 06:49:49 +0000 (16:49 +1000)]
Send API timing to statsd for all requests, not just writes
The rate will be about 8k req/s, whereas currently graphite1001
receives about 350k metric updates per second, so we would expect
this to increase statsd load by about 2%.
This would have been very helpful in diagnosing API cluster overload
incidents such as the one on July 11.
Change-Id: I0a248e1bcec05915d98ec735f4cfb9800e64412d
jenkins-bot [Fri, 20 Jul 2018 06:12:44 +0000 (06:12 +0000)]
Merge "language: Document sources of data and example usage"
jenkins-bot [Fri, 20 Jul 2018 04:59:31 +0000 (04:59 +0000)]
Merge "resourceloader: CSSMin::getLocalFileReferences now strips anchors"
jenkins-bot [Fri, 20 Jul 2018 04:56:09 +0000 (04:56 +0000)]
Merge "Language::listToText: Avoid PHP warnings from E_STRICT"
Antoine Musso [Tue, 17 Jul 2018 15:59:17 +0000 (17:59 +0200)]
resourceloader: CSSMin::getLocalFileReferences now strips anchors
When processing:
url( 'foo.svg#anchors' )
getLocalFileReferences() would yield a file 'foo.svg#anchors'. It might
occur with upstream libraries, such as WikiFonts used by the Refreshed
skin.
When a path is found, strip the anchor entirely. Skip the case if that
results in an empty file.
remap() would properly skip the behavior "#default#', also skip url that
are solely an anchor such as '#other'.
Add a few test cases to cover CSSMin::getLocalFileReferences and cover
anchors usage in CSSMin::remap.
Bug: T115436
Change-Id: I1749ddc2b89021807f42d64131931ad7a99a7b43
Timo Tijhof [Fri, 20 Jul 2018 03:06:28 +0000 (20:06 -0700)]
language: Document sources of data and example usage
Change-Id: I0c72069d010285d9b88ff83bc1478410961065d3
Tim Starling [Fri, 20 Jul 2018 02:03:48 +0000 (12:03 +1000)]
parserTests: re-enable save parse and fix MCR errors
* Re-enable parse on save as requested in T159014, instead hide
the feature behind an option.
* Add the MCR tables to the list of tables that need to be cloned. Fixes
unconditional exception.
* Fix some undefined member variables, remove assignment to unused
variable $this->databaseSetupDone.
Bug: T159014
Change-Id: I6be577ebcbadeb4e08079a82543d521fc2bb5948
jenkins-bot [Fri, 20 Jul 2018 00:06:16 +0000 (00:06 +0000)]
Merge "Convert Title::getTitleCache() to using MapCacheLRU (again)"
jenkins-bot [Fri, 20 Jul 2018 00:06:10 +0000 (00:06 +0000)]
Merge "ApiMain: Always create a new printer in getPrinterByName()"
jenkins-bot [Fri, 20 Jul 2018 00:06:05 +0000 (00:06 +0000)]
Merge "Avoid passing "false" as keys to MapCacheLRU in RepoGroup"
Aaron Schulz [Thu, 19 Jul 2018 23:40:02 +0000 (00:40 +0100)]
Convert Title::getTitleCache() to using MapCacheLRU (again)
Re-applies commit
7a25cd388c8adcce6, reverted by
3648e5683c0ff11c0.
Change-Id: I087edb15fbe11ccac7be28540e20d990f5615c4e
Aaron Schulz [Thu, 19 Jul 2018 10:02:57 +0000 (11:02 +0100)]
Avoid passing "false" as keys to MapCacheLRU in RepoGroup
Bug: T200026
Change-Id: I40f6ad2a3d281d06c9b6eaf4f31d9796ea5e9e9e
Translation updater bot [Thu, 19 Jul 2018 20:13:30 +0000 (22:13 +0200)]
Localisation updates from https://translatewiki.net.
Change-Id: I78f8728c250651e132b1104d2a26047ee938ac54
jenkins-bot [Thu, 19 Jul 2018 18:19:56 +0000 (18:19 +0000)]
Merge "Convert Title::getTitleCache() to using MapCacheLRU (again)"
jenkins-bot [Thu, 19 Jul 2018 17:35:01 +0000 (17:35 +0000)]
Merge "In mysql.php ignore SIGINT"
Mark A. Hershberger [Fri, 13 Jul 2018 18:19:51 +0000 (14:19 -0400)]
Language::listToText: Avoid PHP warnings from E_STRICT
Language::listToText() uses for(;;) which assumes the index in a list
starts at 0 and following keys are consecutive integers.
This assumption fails (and results in warnings with E_STRICT) when
SpecialListGroupRights::formatPermissions() uses the key-preserving
array_intersect() to create a subset of $allGroups.
Bug: T199559
Change-Id: I4eb03a95509d69653156a2764d58c0c5d0d1dfbc
jenkins-bot [Thu, 19 Jul 2018 17:02:53 +0000 (17:02 +0000)]
Merge "RCFilters: Work around Chrome browser bug in filter menu"
jenkins-bot [Thu, 19 Jul 2018 16:59:29 +0000 (16:59 +0000)]
Merge "Set default wgMultiContentRevisionSchemaMigrationStage to write-both"
Marko Obrovac [Thu, 19 Jul 2018 16:13:38 +0000 (18:13 +0200)]
JobQueueGroup: Allow readOnlyReason to be specified per JQ type
We use $wgReadOnly for various reasons, one of which is to disallow
writes in the currently-non-active DC. However, we should allow the
readOnlyReason configuration variable to be available per JobQueue type
and have it the code respect that.
In our current set-up, we use JobQueueEventBus which ever only uses the
enqueue execution path and does not depend upon which DC it is executed
in, so this will allow us to enqueue jobs in both DCs.
Note that this is an alternative approach to the one outlined in
Ibbad6063b6b154d7f7d172c79f7be324bf80eb7e
Bug: T199594
Change-Id: I8f1a57a81ea11c1c587c0057fa8bb3454b0e0b56
Aaron Schulz [Thu, 19 Jul 2018 16:11:39 +0000 (17:11 +0100)]
filebackend: let any specific backend value of "readOnly" override the default
Previously, only different reason strings could override, but there was
no way to override the read-only mode to being off.
Change-Id: I4266694034bfbdada90d4be96bcbc1a37f5b5409
jenkins-bot [Thu, 19 Jul 2018 16:11:24 +0000 (16:11 +0000)]
Merge "docs: Remove table-layout documentation from core"
Brad Jorsch [Thu, 19 Jul 2018 13:24:48 +0000 (09:24 -0400)]
ApiMain: Always create a new printer in getPrinterByName()
ApiMain already caches the printer in ->mPrinter, so if
getPrinterByName() is being called more than once that's because we
really want a new printer instance, without any cached errors or other
behavior that results from reusing the same instance.
Bug: T199949
Change-Id: I779cbbaa8aab9b049a8eed732416edd828121ec4
jenkins-bot [Thu, 19 Jul 2018 14:40:26 +0000 (14:40 +0000)]
Merge "config: Add new ConfigRepository"
jenkins-bot [Thu, 19 Jul 2018 14:26:51 +0000 (14:26 +0000)]
Merge "registration: Expose config array to ExtensionRegistry"
Florian Schmidt [Wed, 14 Dec 2016 20:55:56 +0000 (21:55 +0100)]
config: Add new ConfigRepository
This, for now, has the focus for holding metadata of each
configuration option inside of MediaWiki (both extensions and core).
It's very close to the Config interface and friends, and uses these
implementations to retrieve the actual value of the
configuration option.
The goal with this change is to implement a basic architecture to
allow to display the values of the curent configuration of the MediaWiki
installation on-wiki (e.g. on a special page). It also provides a
central point, where the currently known options can be fetched
from.
The long-term goal, of course, would be to get a web interface to really
configure the MediaWiki installation. But, this is more like a dream,
then a plan (from the point of view of this commit).
Next steps would be:
* Enable ExtensionRegistry to "register" the configuration options of
extensions, so they're available in the repo (done).
* Find a good way to get mediawiki/core configurations into this repo
* Work out an overall architecture to display the different possible
values. (I think about something like different formatters for types).
Change-Id: I9419508eaa85ffc55520db7f33b3e9530fc99f00
Florian Schmidt [Fri, 13 Jan 2017 05:21:00 +0000 (06:21 +0100)]
registration: Expose config array to ExtensionRegistry
Currently, the ExtensionProcessor processes the config array of the
extension and just sets the value as a global. In this way, however,
the metadata of the configs (description, public, ...) aren't available.
This change exposes the configurations in the extracted data as the
config key.
Bug: T155908
Change-Id: Ie41e6ecb8d08a5b20b56eb2281af1dfdd808b2c4
Volker E [Thu, 19 Jul 2018 12:39:34 +0000 (14:39 +0200)]
docs: Remove table-layout documentation from core
This doesn't belong here and also doesn't make a lot of sense
with current browser support any more.
We shouldn't provide basic HTML documentation in 2018 as part of
core.
Change-Id: I1a3516e45902d16f0d9efbe97cb2dbd6efd566be
Aaron Schulz [Tue, 17 Jul 2018 07:47:06 +0000 (08:47 +0100)]
Convert Title::getTitleCache() to using MapCacheLRU (again)
Re-appplies commit
7a25cd388c8a, which was reverted because
HTMLTitleTextField was passing `null` as a title key, this was
fixed in I29ecd94cdf9e30.
Bug: T199763
Change-Id: I7fe8bbed3fb24240e47ba22de57da9dc8a9bea22
Petr Pchelko [Thu, 19 Jul 2018 09:42:47 +0000 (11:42 +0200)]
Remove JobQueueSecondTestQueue class.
After the switchover is completed we do not need this any more.
This class was marked deprecated as soon as it was created, so
removing it should be safe.
Change-Id: I3451a7a9d1e78bb2c7c988aeb1e3aac93bfc1b6f
jenkins-bot [Thu, 19 Jul 2018 09:32:52 +0000 (09:32 +0000)]
Merge "registration: Add class docs for ExtensionJsonValidator"
jenkins-bot [Thu, 19 Jul 2018 08:32:11 +0000 (08:32 +0000)]
Merge "Fix handling of already logged-in user in Special:Userlogin"
Max Semenik [Wed, 18 Jul 2018 08:42:52 +0000 (01:42 -0700)]
Deprecate wfArrayFilter() and wfArrayFilterByKey()
Now that all our supported PHP versions have array_filter()
with a third parameter, these functions aren't needed anymore.
Depends-On: I3b097a1a048baabcaca15dc214a3a1bb06e746cc
Depends-On: I0187e27ac47cbab099249572201d1a649226a734
Change-Id: I7cabd0252691a083cb749cf9d3a7a23f1d076c39
jenkins-bot [Thu, 19 Jul 2018 02:25:04 +0000 (02:25 +0000)]
Merge "LogEventsList: Remove options not used by HTMLForm"
jenkins-bot [Thu, 19 Jul 2018 01:53:32 +0000 (01:53 +0000)]
Merge "Make MultiWriteBagOStuff use the native merge() of each backend"
Krinkle [Thu, 19 Jul 2018 01:48:56 +0000 (01:48 +0000)]
Merge "Convert MemcLockManager to using MapCacheLRU"
Timo Tijhof [Thu, 19 Jul 2018 01:20:29 +0000 (18:20 -0700)]
registration: Add class docs for ExtensionJsonValidator
Change-Id: Ic37544de1400600719764057a63cc546c26218d8
Gergő Tisza [Wed, 18 Jul 2018 22:49:50 +0000 (00:49 +0200)]
Fix handling of already logged-in user in Special:Userlogin
Due to a forgotten return clause, the full request handling
is executed after calling successfulAction() when the user
reloads the login page while logged in. With a typical setup
case this is masked by the fact that successfulAction() does
a redirect (although it probably still impacts performance),
for less common configurations this might result in the login
form rendered below a success message, or in calling extension
hooks which should not be called.
Bug: T185702
Change-Id: If19220c4cf17b08f7e03ca0d06381067f26dda1a
Aaron Schulz [Tue, 10 Jul 2018 23:26:43 +0000 (00:26 +0100)]
Make MultiWriteBagOStuff use the native merge() of each backend
This lets backends that support CAS take advantage of it.
Also fix race condition with merge() in asyncWrites mode where two
threads might add to a list/set but the first request finishes after
the second, which would erase an item during the set().
Bug: T198239
Change-Id: Ibd2539429746f12b4de74d0f031089cb80aa8b4b
jenkins-bot [Wed, 18 Jul 2018 21:03:59 +0000 (21:03 +0000)]
Merge "LogEventsList: Add backwards-compatibility for log-show-hide messages"
Translation updater bot [Wed, 18 Jul 2018 19:54:28 +0000 (21:54 +0200)]
Localisation updates from https://translatewiki.net.
Change-Id: Ie2ceb98d5d3a492d4bd82a59e957ac36cbd01f05
addshore [Wed, 18 Jul 2018 14:00:36 +0000 (16:00 +0200)]
Remove unused $links in TablePager::getNavigationBar
Change-Id: I43ba3390c47e4955704cfd82b16627008f9e6bc8
jenkins-bot [Wed, 18 Jul 2018 19:20:08 +0000 (19:20 +0000)]
Merge "WikiPage: Return false from hasDifferencesOutsideMainSlot if no differences at all"
jenkins-bot [Wed, 18 Jul 2018 18:25:56 +0000 (18:25 +0000)]
Merge "Add $maxAge parameter to MapCacheLRU get() and getField() methods"
jenkins-bot [Wed, 18 Jul 2018 18:21:24 +0000 (18:21 +0000)]
Merge "rdbms: give TransactionProfiler write time and row count for ROLLBACK"
jenkins-bot [Wed, 18 Jul 2018 18:18:28 +0000 (18:18 +0000)]
Merge "objectcache: switch WANObjectCache process cache to MapCacheLRU"
jenkins-bot [Wed, 18 Jul 2018 18:10:14 +0000 (18:10 +0000)]
Merge "HTMLTitleTextField, HTMLUserTextField: Check for `null` before using the value"
Aaron Schulz [Wed, 18 Jul 2018 00:01:56 +0000 (01:01 +0100)]
Convert MemcLockManager to using MapCacheLRU
Change-Id: Ia376641f29c1d9acc51268afb9319ab0beab5528
Aaron Schulz [Wed, 11 Jul 2018 13:29:27 +0000 (14:29 +0100)]
Add $maxAge parameter to MapCacheLRU get() and getField() methods
Change-Id: I2dee1ad3a5b4f2ef8c182e3466812777a9a9cf7f
jenkins-bot [Wed, 18 Jul 2018 18:07:02 +0000 (18:07 +0000)]
Merge "LogEventsList: Use GET in HTMLForm"
Aaron Schulz [Mon, 2 Jul 2018 15:49:28 +0000 (16:49 +0100)]
rdbms: give TransactionProfiler write time and row count for ROLLBACK
This will hopefully provide more diagnostic info when DBTransactionSizeError
rollbacks are triggered.
Bug: T190260
Change-Id: Ib8bea5a9ec7d3ffeaf423adb930dc6fb14314449
Aaron Schulz [Wed, 11 Jul 2018 13:38:22 +0000 (14:38 +0100)]
objectcache: switch WANObjectCache process cache to MapCacheLRU
Change-Id: I60d8de99e9590bfef6009f133723bb10d5837c52
Bartosz Dziewoński [Wed, 18 Jul 2018 17:45:40 +0000 (19:45 +0200)]
LogEventsList: Remove options not used by HTMLForm
HTMLForm gets the input value from the request. The values
specified in the form descriptor were never used. (You can specify a
default value, but that uses the 'default' key anyway.)
Change-Id: I27934e1118492371e547dd5e94feb172a0510cc9
Roan Kattouw [Wed, 18 Jul 2018 17:41:18 +0000 (19:41 +0200)]
RCFilters: Work around Chrome browser bug in filter menu
For some reason, the presence of any padding whatsoever makes the menu
items a lot taller, and the amount of padding doesn't actually cause any
visual change at all (it looks the same with 1px and 6px of padding).
Bug: T199932
Change-Id: If294de0f9dfa48583243ba53d74b0ea66d021839
jenkins-bot [Wed, 18 Jul 2018 17:38:04 +0000 (17:38 +0000)]
Merge "Reduce the rate of calls to Category::refreshCounts"
Brad Jorsch [Wed, 18 Jul 2018 15:14:56 +0000 (11:14 -0400)]
WikiPage: Return false from hasDifferencesOutsideMainSlot if no differences at all
If the page has no differences in any slot, the method is incorrectly
returning "true" to claim there are differences in some non-main slot.
Follows up Ifdf9bc9d884844f9ffeb8019d9b13d5737862063.
Bug: T194412
Change-Id: I013bcd9c39d05d9af47c1a29026b14d27af2db15
Matěj Suchánek [Wed, 18 Jul 2018 13:59:04 +0000 (15:59 +0200)]
Provide correct param to Special:Log
Follows up Iba3c6aa5ac40dcdee0792c2d045b238b02d76521.
Bug: T117737
Change-Id: I90bb55eba4c3b4846da2394798ad6a2ac6fc9eca
jenkins-bot [Wed, 18 Jul 2018 10:33:14 +0000 (10:33 +0000)]
Merge "AuthManager: Don't invalidate BotPasswords if a password reset email is sent"
jenkins-bot [Wed, 18 Jul 2018 09:47:07 +0000 (09:47 +0000)]
Merge "registration: Validate no duplicate keys exist in extension.json"
Florian Schmidt [Sun, 22 Jan 2017 15:36:22 +0000 (16:36 +0100)]
registration: Validate no duplicate keys exist in extension.json
The JSON specification permits duplicate keys, but doing so is almost
always nearly a mistake.
We can use the JSON lint library (added as
dev-requirement to composer), which is now also used to decode the
JSON (as it's the same step as validating the JSON).
Bug: T153507
Change-Id: Ia713a1906169333c1aa2aebdc0ed060d26428d72
jenkins-bot [Wed, 18 Jul 2018 09:26:50 +0000 (09:26 +0000)]
Merge "registration: Use the correct key for skin dependencies"
jenkins-bot [Wed, 18 Jul 2018 09:02:28 +0000 (09:02 +0000)]
Merge "RCFilters: Standardize loading indicator appearance"
jenkins-bot [Wed, 18 Jul 2018 07:26:41 +0000 (07:26 +0000)]
Merge "maintenance/interwiki.*: Replace Wikimedia Foundation wiki URL"
Reedy [Tue, 17 Jul 2018 15:31:11 +0000 (16:31 +0100)]
maintenance/interwiki.*: Replace Wikimedia Foundation wiki URL
The wiki is now being moved to foundation.wikimedia.org.
Bug: T199812
Change-Id: I7ff4e74f075b90e81bc61ec3b03fe0884b6043ef
Tim Starling [Wed, 18 Jul 2018 05:47:23 +0000 (15:47 +1000)]
In mysql.php ignore SIGINT
Every foreground process with the terminal open will receive SIGINT when
the interrupt key is pressed. But mysql handles SIGINT and kills the
running query, we don't want the wrapper to exit in that case. So,
ignore SIGINT while mysql is running. This is similar to how a shell
deals with this problem.
The other options would have been:
* Close the filehandles in PHP after forking (probably impossible)
* Use pcntl_exec() so that PHP isn't running (more complicated and
requires the pcntl extension anyway)
* Clear termcap ISIG flag (probably would have broken mysql)
Bug: T199152
Change-Id: I7660cd24e036fbe3fe36d12563ffe1c1fa2360d0
Prateek Saxena [Wed, 18 Jul 2018 05:08:03 +0000 (10:38 +0530)]
LogEventsList: Use GET in HTMLForm
I missed this during conversion.
Follows-up on Iba3c6aa5ac40dcdee0792c2d045b238b02d76521.
Bug: T199856
Change-Id: If3f228425b217c9ccdb7d32e0bc899647797ed70
Prateek Saxena [Wed, 18 Jul 2018 04:19:36 +0000 (09:49 +0530)]
LogEventsList: Add backwards-compatibility for log-show-hide messages
Also updated the release notes.
Bug: T199657
Change-Id: I027effbc7bd797ee241281ddd57334a5dc5c8a57
Aaron Schulz [Tue, 17 Jul 2018 20:22:59 +0000 (21:22 +0100)]
Reduce the rate of calls to Category::refreshCounts
Bug: T199762
Change-Id: I23e2e1ebf187d21ea4bd22304aa622199a8b9c5b
Aaron Schulz [Wed, 11 Jul 2018 12:54:51 +0000 (13:54 +0100)]
Switch callers from ProcessCacheLRU to MapCacheLRU
Change-Id: Iefe5c25e952079420b3241bc26efe1d2d086e6fb
Aaron Schulz [Wed, 11 Jul 2018 12:53:23 +0000 (13:53 +0100)]
filerepo: switch callers from ProcessCacheLRU to MapCacheLRU
Change-Id: I58706b5610e0113a0c8d86eed2662870198f9b46
Prateek Saxena [Tue, 17 Jul 2018 04:54:03 +0000 (10:24 +0530)]
HTMLTitleTextField, HTMLUserTextField: Check for `null` before using the value
Otherwise HTMLTitleTextField sends null to Title::newFromTextThrow(),
which causes an exception when trying to look it up in a cache.
Similar issue is present in HTMLUserTextField, although that one
hasn't caused problems in practice yet.
Follows-up on I93ad51ffe7bee597d2d127f4c5d6b2929ffc8f7e and
I0de4194a37b6ef260d35feb1e6730985775d5351.
Bug: T199763
Change-Id: I29ecd94cdf9e3064e6e9e7f4e65a50f267b5282d
David Causse [Tue, 10 Jul 2018 12:10:09 +0000 (14:10 +0200)]
Deprecate SearchEngine::replacePrefixes
This should be handled internally by SearchEngine implementations.
Bug: T198860
Change-Id: Ifbfd0fcb81fcacf5228bd2ffcac7b80fca872b2a
Depends-On: I7d4ff9498fa1f4ea66835c634b8931f4c29993fb
David Causse [Mon, 9 Jul 2018 16:55:54 +0000 (18:55 +0200)]
Unify SearchEngine normalizeNamespace and parseNamespacePrefixes
These methods are very similar there should be no need to have
two differents way to extract the namespace prefix.
Bug: T198860
Change-Id: I22802278452559d35a3d8f6068549c1fef1a5e86
David Causse [Mon, 25 Jun 2018 17:19:32 +0000 (19:19 +0200)]
Deprecate usage of SearchEngine:transformSearchTerm
This method was introduced in
4115586000a575eb814a80868c817d18ee169d3b
to support the prefix URI param introduced by the InputBox extension.
There are no reasons that this logic is exposed to SearchEngine users
and should be handled internally by SearchEngine implementations
that supports it.
Previously the search query was updated, now the prefix param will passed
along using SpecialSearch::$extraParams.
Bug: T198318
Change-Id: I33518d3f3ddee741ff4f3b47eb4928009bea66d1
Depends-On: I67c7f1886dd6a2d07c12015e2711c138e9f140e9
jenkins-bot [Tue, 17 Jul 2018 21:38:53 +0000 (21:38 +0000)]
Merge "Rephrase 'logeventslist-more-filters'"
jenkins-bot [Tue, 17 Jul 2018 20:40:03 +0000 (20:40 +0000)]
Merge "Introduce preference filters"
Bartosz Dziewoński [Tue, 17 Jul 2018 20:28:20 +0000 (22:28 +0200)]
Rephrase 'logeventslist-more-filters'
Bug: T117737
Change-Id: I6315ac76f8ddfeb024cb16331fc643ba726dfc34
Brad Jorsch [Tue, 17 Jul 2018 20:18:59 +0000 (16:18 -0400)]
AuthManager: Don't invalidate BotPasswords if a password reset email is sent
There's a difference between addition of credentials, which doesn't
need to invaliate BotPasswords, and changing or removal of credentials,
which does.
It seems most straightforward for the caller of
AuthManager::changeAuthenticationData() to know which is intended, so
let's add a flag there.
Bug: T199809
Change-Id: Ib8405734e605b94f3f0b66596ad95784cb365e4f
Max Semenik [Fri, 6 Jul 2018 19:21:22 +0000 (12:21 -0700)]
Introduce preference filters
This normalizes handling of transformations on the boundaries between
preferences and generic form controls and removes the special case
where email-blacklist is passed around as an array internally, leaking
into the API.
As a result of this normalization, meta=userinfo no longer returns an
array of users, using the internal representation like action=options.
Bug: T198935
Change-Id: Iff63da0d215585cfcf083e7f7ec8ed45d5b77301
jenkins-bot [Tue, 17 Jul 2018 20:05:18 +0000 (20:05 +0000)]
Merge "LogEventsList: Remove unused messages"
Translation updater bot [Tue, 17 Jul 2018 19:54:08 +0000 (21:54 +0200)]
Localisation updates from https://translatewiki.net.
Change-Id: Ia5e603f787ffeb5450d8695acee5f8dbde829ca8
jenkins-bot [Tue, 17 Jul 2018 18:52:18 +0000 (18:52 +0000)]
Merge "LogEventsListGetExtraInputs: Keep $input and add $formDescriptor"
Prateek Saxena [Fri, 13 Jul 2018 02:02:13 +0000 (07:32 +0530)]
LogEventsListGetExtraInputs: Keep $input and add $formDescriptor
Follows-up on Iba3c6aa5ac40dcdee0792c2d045b238b02d76521.
Bug: T117737
Bug: T199495
Change-Id: I697e158887fcca1da88763a4c929a981d9211490
jenkins-bot [Tue, 17 Jul 2018 07:15:45 +0000 (07:15 +0000)]
Merge "Revert "Convert Title::getTitleCache() to using MapCacheLRU""
Krinkle [Tue, 17 Jul 2018 07:01:17 +0000 (07:01 +0000)]
Revert "Convert Title::getTitleCache() to using MapCacheLRU"
This reverts commit
7a25cd388c8adcce60f4c055295baed71e730616.
Bug: T199763
Change-Id: I276b833d75e93328a188056bf1eb2491461d6d53
jenkins-bot [Tue, 17 Jul 2018 05:11:31 +0000 (05:11 +0000)]
Merge "Cached HTML/JS shouldn't show two toggle links"
jenkins-bot [Tue, 17 Jul 2018 05:05:44 +0000 (05:05 +0000)]
Merge "Clean up docs around $wgUseSharedUploads"
jenkins-bot [Tue, 17 Jul 2018 05:03:49 +0000 (05:03 +0000)]
Merge "WebRequest: Fix flaky testGetElapsedTime test case"
Fomafix [Wed, 11 Jul 2018 21:02:48 +0000 (14:02 -0700)]
Cached HTML/JS shouldn't show two toggle links
Per https://phabricator.wikimedia.org/T195053#
4417392
Bug: T195053
Change-Id: I43ff902055b338b434658e9bb1d6eacdc5684ba2
jenkins-bot [Tue, 17 Jul 2018 03:44:26 +0000 (03:44 +0000)]
Merge "filebackend: switch callers from ProcessCacheLRU to MapCacheLRU"
Timo Tijhof [Tue, 17 Jul 2018 03:23:30 +0000 (20:23 -0700)]
WebRequest: Fix flaky testGetElapsedTime test case
* Increase tolerance from 0.2s to 60s.
* Update mock to only set 'requestTime' when needed.
* Inject the mock timestamp.
* Increase test coverage by not just ensuring 0.0 is returned
when invoked right after (which would be satisfied if it
always returned 0.0), but use a timestamp that started
slightly in the past to confirm it does (likely) do
computation.
Bug: T199764
Change-Id: Iad9499391eecb4a9d1923d231a1a5f1afe173ecc
Brad Jorsch [Mon, 2 Jul 2018 15:17:06 +0000 (11:17 -0400)]
Parser: Remove style and script tags' content from TOC
We don't want to display the stylesheet as part of the TOC entry if
someone uses TemplateStyles in a heading.
Bug: T198618
Change-Id: I2f7316daaba0cce662b6a4702ab87322e6783655
jenkins-bot [Tue, 17 Jul 2018 02:49:58 +0000 (02:49 +0000)]
Merge "Sync up with Parsoid parserTests.txt"
jenkins-bot [Tue, 17 Jul 2018 01:09:25 +0000 (01:09 +0000)]
Merge "objectcache: make BagOStuff::mergeViaLock() timeout more sensible"