* This can be useful for carrying over metadata when copying files around
* Also fixed a bug in sanitizeHdrs() for Swift (broken content-disposition)
Change-Id: I4534e9acac2b306086797b3677f85c05b98e39fc
* Added a new curl multi wrapper under libs/.
* Used the new MultiHttpClient class instead of CloudFiles
* Removed CDN related config and code as it does not work.
well due to mutating files and tiny purge rate limits by CDNs.
* Also removed the unused and esoteric 'swiftAnonUser' parameter.
* Support concurrent 'describe' operations.
* Set ACLs on containers on intial PUT instead of in two steps.
Previously if the second step failed, prepare() would never
set the ACLs correctly since it aborts if the container exists.
* Improve token deletion to only trigger on 401s rather than other
random errors like 50X timeouts.
* Removed duplicative getResponse* code by just using one closure
called for both the concurrent and serial cases. Both cases are
now funneled through doExecuteOpHandlesInternal().
* Add the MD5 to stat info, as it might be useful for speeding up
the copy scripts in some cases (when both backends have md5 set).
* Avoid use of trigger_error() since there is a swift log group.
* A few other small code cleanups.
Change-Id: I1adb3e2df6df8cf01d3ad74158de96ea9a79da2e
assertContainsOnlyInstancesOf has been added in PHPUnit 3.7.0, so we need to
have that version for running the tests.
Bug: 59759
Change-Id: I25dbd377b97ec0b9062d3c8cf495ba3f06956ac4
Sometimes it is desirable to change a class name that is still referenced
in extensions or config files (e.g. for consistency, as in I507ba00a).
PHP's class_alias() function can help in preserving backward compatibility;
however, creating an alias first requires that the class be loaded.
Hence this is implemented in AutoLoader. Lazy loading continues to work,
the list of class names is still maintained in a central location, and
optionally, deprecation warnings can be generated.
Change-Id: I1d3fb04a448647b5be10bed7fec05238b9fc6fc7
There's no need for TestUser to remove all groups only to add some back
on. We should be able to speed things up (see bug 43762) by only
removing the groups that aren't wanted and only adding the ones that
aren't already possessed.
Change-Id: I50dd9b117a8a21c8a6170520b3bbf98e3030ae47
$wgSearchType needs to be controlled for at time the pages get created.
This patch sets the $wgSearchType global.
Also, the 'singleton' is removed from SearchEngine::getSearchTypes, which
appears to provide little or no benefit while binding all tests to the
global variable.
SearchEngine::getSearchTypes() gets called in SearchUpdate which is
run on page edits. This means the first page edit in the entire
test suite causes the global variable to be set and subsequent tests
cannot override it.
These changes allow SearchEngineTest to pass, even if one has CirrusSearch
or other such extensions enabled.
Change-Id: I39050da8659dc69db31125f469f494a5fb4b8fca
* Throw InvalidArgumentException
* Use data provider in unit tests
* Detect associative arrays without copying
Per Tyler's post-merge review of Id2833c5a9.
Change-Id: Iec6b135238ca5da3002944066843102f0ae8d23d
Russian (ru) plural rules have a major change. The 'few' form is
merged with the 'other' form. The current forms are 'one', 'many', 'other'.
In MW ru plural rules were overridden using convertPlural methdod
in LanguagesRu.php with 3 forms.
Effectively forms[1] and forms[2] are swapped.
Followup: I9930b290d004667a3bb09e5c1663ec2c9c27d8a6
Bug: 56931
Change-Id: Ia5779e42315d3f41f52dce2bfffaee0a4297d23b
There is nothing about the class that mandates it be used exclusively for
profiling, or exclusively within the context of MediaWiki. The functionality it
provides is generic. So move it to libs/ to make it easier to discover & reuse.
Change-Id: Iac8f2f8e1c3231ebe76318d72cf6225e25a7b891
Updated plurals.xml with new data from CLDR 24.
This data is according to UTS #35 Rev 33.
Update the CLDRPluralRuleParser.js to version 1.1 from upstream
https://github.com/santhoshtr/CLDRPluralRuleParser
Changes to the plural rules:
* Hebrew override removed since CLDR 24 matches with MW plural rules.
* Updated the syntax of overridden rules to TR35 Rev 33 for
Lower Sorbian (dsb), Upper Sorbian (hsb), Belarusian in Taraskievica
orthography (be_tarask), Old Church Slavonic (cu), Bhojpuri (bho),
Samogitian (sgs).
* Removed Manx (gv) override. See I46ab3dadc7fe08c1e60bbd81a1ee841e166e9608.
* Removed the overriden convertPlural method for Serbian from LanguageSr.php,
since CLDR 24 matches with MW rules. Updated and added more tests.
Tests updated for Serbocroatian (sh), too. Old CLDR versions had 4 plural
rules and MW had only 3. In CLDR 24, the form 'many' was removed and it
became identical to the MW. Same for Bosnian (bs) and Croatian (hr).
Also for variants sr-ec and sr-el
* Macedonian (mk) used to count 11 as 'other' form.
CLDR 24 counts it as 'one'.
Not overriding, using CLDR 24 here.
Updated the tests. MW will not override this.
* Armenian (hy) used to count 0 as 'other'.
Now it is 'one' form.
Updated the tests. MW will not override this.
* Latvian (lv) used to count only 0 as 'zero' form, but CLDR 24, any number
satisifying the following formula is counted as zero:
n % 10 = 0 or n % 100 = 11..19 or v = 2 and f % 100 = 11..19
Examples: 0, 10~20, 30, 40, 50, 60, 100.
Updated the tests accordingly. Not overriding it in MW.
Users will see different plural form for the above numbers.
* Removed Ukranian custom plural rule since it match with MW
* Russian (ru) plural rules have a major change.
The 'few' form is merged with the 'other' form.
The current forms are 'one', 'many', 'other'.
In MW ru plural rules were overridden using convertPlural methdod
in LanguagesRu.php with 3 forms.
Effectively forms[1] and forms[2] are swapped.
This will affect the messages, and such messages
must be reviewed and updated. This change is not included in this patch and
wil be done separately.
Russian is the only remaining language class with convertPlural method overridden.
Notable impact on the exising messages:
* For languages ru, uk, be_tarask, sr, For the special case
of two plural forms and first mapped to 1 and rest to the other form, syntax like
{{plural:$1|1=one|other}} should be used.
For further information regarding each of the above language changes, see
1. http://unicode.org/cldr/trac/ticket/3727
2. http://goo.gl/H2HEz
CLDR 24 can handle fractions. Ideally it should start working
in MW without any code changes, but MW language test suite
does not have enough tests to confirm.
Followup: e571717e06
Bug: 56931
Change-Id: I9930b290d004667a3bb09e5c1663ec2c9c27d8a6
MessagePack is a space-efficient binary data interchange format. I am going to
use it to encode profiling data in ProfilerSimpleUDP. The official PHP
implementation is provided as a C extension, so using it would further
encumber migration to HHVM. This patch adds MWMessagePack, a class implementing
a pack() method for encoding native PHP values as MessagePack byte strings. The
implementation is based on <https://github.com/onlinecity/msgpack-php>, but
revised for clarity and conformity with MediaWiki coding conventions.
Change-Id: Id2833c5a9da659cb13ec1330de9dd57138ada9c8
Instead test that LBFactory::getLBFactoryClass() does the right thing.
In LBFactory::getLBFactoryClass(), the method now requires a configuration
array as a parameter, and returns the class name as a string.
LBFactory::singleton() now passes in $wgLBFactoryConf into getLBFactoryClass(),
while tests can use test configuration.
Only use of LBFactory::getLBFactoryClass() in core is in LBFactory::singleton().
No extension in gerrit uses getLBFactoryClass() and I can find no use of it
when searching places like Github and Google, so it should be safe to make
such change in LBFactory.
Bug: 59105
Change-Id: I71ae7df16bc5c214b9389125140bca5ce68d274c
RunningStat computes the central tendency, shape, and extrema of a set of
points online, in constant space. It uses a neat one-pass algorithm for
calculating variance, described here:
<en.wikipedia.org/wiki/Algorithms_for_calculating_variance#On-line_algorithm>
This particular implementation adapts a sample C++ implementation by John D.
Cook to PHP. See <http://www.johndcook.com/standard_deviation.html> and
<http://www.johndcook.com/skewness_kurtosis.html>.
RunningStat instances can be combined. The resultant RunningStat has the same
state it would have had if it had been used to accumulate each point. This
property is attractive because it allows separate threads of execution to
process a stream in parallel. More importantly, individual points can be
accumulated in stages, without loss of fidelity, at intermediate points in the
aggregation process. JavaScript profiling samples can be accumulated in the
user's browser and be combined with measurements from other browsers on the
profiling data aggregator. Functions that are called multiple times in the
course of a profiled web request can be accumulated in MediaWiki prior to being
transmitted to the profiling data aggregator.
Usage will be introduced in a dependent commit.
Change-Id: Ifedda276dfe8e0783cb8c4a95626e2aedd4ad368
No uses found in core of Gerrit extensions.
- Revision::revText() (deprecated in 1.17)
- SkinTemplate::jstext() (deprecated in 1.21)
- WikiPage::quickEdit() (deprecated in 1.18)
- WikiPage::useParserCache() (deprecated in 1.18)
- WikiPage::viewUpdates() (deprecated in 1.18)
Change-Id: I9eee23ea3ce381b6b8373da899287bcaea4cead4
The Interwiki class caches lookups, so ApiQueryTest.php's use of "en:"
was failing for mediawiki-core-regression-master because something else
had already made a call for that prefix. The easy fix is to use a
different prefix that is very unlikely to ever be used in a different
test.
Change-Id: I43ca554d5b5465d760eae03d1f76b8c3bf589210
ApiQueryBase::titlePartToKey now allows an extra parameter that
indicates the namespace in order to properly capitalize the title part.
This allows list=allcategories, list=allimages, list=alllinks,
list=allpages, list=deletedrevs and list=filearchive to
handle case-sensitivity properly for all parameters.
Bug: 25702
Change-Id: Iaa5a71ec536f3716f54bc84b39f645545dfd8660
* This backend passes all filebackend and parser tests
* Fixed setupUploads() in parser tests to just use create() instead
of using store() and having a race condition in the process
* Fixed 'use-filebackend=' for Parser tests
bug: 58094
Change-Id: Ib0c38183cb7f9f2325da98c8a8a1eb2b8e39a7aa
Page with id 0 does not exist at the time transaction
is committed to the database.
Unit tests with table constraints enabled
fail with:
1) RevisionStorageTest::testUserWasLastToEdit with data set #0 (3, true)
DBQueryError: A database error has occurred. Did you forget to run
maintenance/update.php after upgrading? See:
https://www.mediawiki.org/wiki/Manual:Upgrading#Run_the_update_script
Query: INSERT INTO "unittest_revision"
(rev_id,rev_page,rev_text_id /* more fields */ )
VALUES ('461','0','460' /* more fields */
Function: Revision::insertOn
Error: 23503 ERROR: insert or update on table "unittest_revision"
violates foreign key constraint "ut_revision_rev_page_fkey"
DETAIL: Key (rev_page)=(0) is not present in table "unittest_page".
/usr/home/saper/test/mytest/includes/db/Database.php:1111
/usr/home/saper/test/mytest/includes/db/DatabasePostgres.php:511
/usr/home/saper/test/mytest/includes/db/Database.php:1077
/usr/home/saper/test/mytest/includes/db/DatabasePostgres.php:871
/usr/home/saper/test/mytest/includes/Revision.php:1422
/usr/home/saper/test/mytest/tests/phpunit/includes/RevisionStorageTest.php:488
/usr/home/saper/test/mytest/tests/phpunit/MediaWikiTestCase.php:123
/usr/home/saper/test/mytest/tests/phpunit/MediaWikiPHPUnitCommand.php:80
/usr/home/saper/test/mytest/tests/phpunit/MediaWikiPHPUnitCommand.php:64
/usr/home/saper/test/mytest/tests/phpunit/phpunit.php:115
Change-Id: I653a8bccdaa748a9bea453cd1dbf609a30e1ff6f
DatabaseMysqlBase::fieldType() is used in the maintenance script
convertLinks.php for upgrading from MediaWiki 1.2 or earlier. This
script was failing when the mysqli extension was enabled.
Follows-up b1cc3511fd, b2d64432ee, and 00299ecb95.
Change-Id: I5d888f4dd344ecba15506023d35f6d2d6e24cfd8
* New operands i, v, w, f, t
* New operators =, !=, %
* Ignore "samples", which are basically unit tests embedded in rule
specifications
* Ignore the new "other" rules, which have an empty condition. It
doesn't really makes sense to parse them, since the empty condition
means special handling should be done in the caller, it is not
equivalent to an unconditional true or false.
* Trailing zero support requires that the input number be a string.
Documented this.
* Fixed some comments
* Added test cases for new features
Bug: 56931
Change-Id: I96986c0c664f785e75b0a4ced2ec9e37b72681c1
Introduce new static function, CSSMin::buildUrlValue.
Actually using such values in CSS does not work well because the
URL_REGEX is nowhere near good enough. :(
Change-Id: I04a7078dd0087bcb461fa5e5168c870d37c255f4
We would previously return the path to the local file on the
filesystem, which is useless in all cases and possibly a security
issue in some. Now we return the URL at which the file would be
accessible had it existed.
Also reordered the code around that part to make the control flow
clearer and added a test.
Change-Id: I1d5befb2ea385ae4d316c5d8c5d1fc092b64c4ff
Backported the plural rules from CLDR 24 as an override to CLDR 23 rules
exising in MediaWiki. The syntax for plural rules changed in CLDR 24, so
modified the syntax to fit the CLDR 23 syntax
Once we are ready with the updated parsers for CLDR 24(See bug 56931),
we should remove the override.
Since we remove the custom plural forms in MW in favor of CLDR,
the following changes comes into effect:
1. 'few' form used as 'zero' form in MW. Practially that make 'one' form used
as 'two' form and 'two' used as 'few' form.
This breaks existing gv {{PURAL}} usage as dicussed in the bug report
2. CLDR defines 'few' form as n % 100 = 0,20,40,60 but MW adds 80 also to that
list, ie n % 100 = 0,20,40,60, 80. So with this patch, 80 is no longer considered
as 'few' plural form.
Bug: 47099
Change-Id: I46ab3dadc7fe08c1e60bbd81a1ee841e166e9608
