This reverts most of commit 2d842f1425,
leaving only the test added in it, and reimplements the same
functionality better.
Instead of stripping /*@noflip*/ annotations in CSSJanus, which is
incompatible with other implementations that preserve it, extend
CSSMin to allow other CSS comments to be present before the
rule-global @embed annotation. (This required making the regex logic
in it even worse than it was, but it's actually slightly less terrible
than I expected it would be. Good thing we have tests!)
Bug: 69698
Change-Id: I58603ef64f7d7cdc6461b34721a4d6b15f15ad79
The output of Message::text() should always be acceptable to pass into
external html escaping, such as when the response is returned over an
API request and escaped by the client side code. Calling ->text() on a
non-existent key was returning the entity encoded value which leads to
double encoding down the line, this patch fixes that oversight.
Bug: 66199
Change-Id: Ieec94d4e4c7e5c36e5e68bbf01792e96368e54e0
Although 'added' to the database, the user name was cleared
in User::loadDefaults() and not added correctly to the database.
Then if one has BetaFeatures extension, then the test fails.
Via a hook, BetaFeatures does "User::newFromName( $user->getName() );",
that produces a null object as name is '127.0.0.1' and then
$user->getOption() fails.
loadDefaults() is automatically called by addToDatabase() with correct
parameters, so we can just remove the call.
Bug: 68626
Change-Id: Ibb5c60192eb83b2608f72b59d7705854549a5dac
Custom LESS functions are problematic for us for a number of reasons,
as outlined by Timo on bug 67368. We should get rid of them.
The only use case was implementing CSSMin data: URI embedding in LESS,
which used to be impossible due to lessc not preserving comments (bug
54673). However, thanks to new syntax added in f3779e06 we can insert
the annotations in such a way that the compiler won't mess with them.
The same technique is used in OOjs UI since 584ed144.
The LESS-function-based embedding implementation also meant that we
were unable to flip images for RTL (bug 66091 and friends: bug 66773,
bug 68326). The annotation one doesn't have this limitation.
Bug: 67368
Bug: 66091
Bug: 66773
Bug: 68326
Change-Id: I3062346ed63272a1c22b5df27b4cc1de2a699d9a
Follows-up b36d883.
By far most data providers are static (and PHPUnit expects them
to be static and calls them that way).
Most of these classes already had their data providers static
but additional commits sloppily introduced non-static ones.
* ResourceLoaderWikiModuleTest, 8968d8787f.
* TitleTest, 545f1d3a73.
Odd unused method 'dataTestIsValidMoveOperation' was introduced
in 550b878e63.
* GlobalVarConfigTest, a3e18c3670.
Change-Id: I5da99f7cd3da68c550ae507ffe1f725d31e7666f
This enables factory functions to be registered for special
pages, as an alterative to giving a class name. This follows the
same rationale as Ieb85493a7765, which introduced factory functions
for API modules.
Change-Id: Ia2107dc5af7869187ba5dc02a1bef46d6801e138
Since Id44d566a, the text passed to the parser when parsing a
&redirect=no page no longer contains the #REDIRECT directive. For the
benefit of extensions that want to know the redirect target from various
parser hooks, record the target on the ParserOptions object associated
with the parse.
Bug: 62856
Change-Id: Icd1da9911a43eabacbd9e9a369a8326f67f270ff
This change adds MultiConfig and HashConfig classes,
but does not actually use them anywhere. In a future
change, we can convert DefaultSettings.php into
a HashConfig instance and use MultiConfig as the
'main' config instance.
Bug: 69418
Change-Id: I0ef2fbb86d5c27602d70240219ee08be31e2d09b
Several unit tests access the database, but did not declare
themselves in the Database @group. This causes
"make databaseless" to fail needlessly. Add the
missing tags.
tests/phpunit/includes/LinkerTest.php
tests/phpunit/includes/actions/ActionTest.php
tests/phpunit/includes/specials/ImageListPagerTest.php
tests/phpunit/includes/specials/SpecialMIMESearchTest.php
tests/phpunit/includes/title/MediaWikiPageLinkRendererTest.php
Other than ActionTest.php, these also are problematic in 1.23
Change-Id: I7c1c957e2194c13e48b7ba68d7529e5d89901875
The previous implementation would unescape '&', '=', '+', and '%'. The
first three will break the URL when unescaped in the query string, and
the last will break when unescaped anywhere.
The code is now changed to treat the path, query, and fragment parts of
the URL separately when unescaping. We also escape any unsafe characters
and ensure all percent-encodings use uppercase hexits.
And since the old name is no longer accurate,
Parser::replaceUnusualEscapes is deprecated in favor of
Parser::normalizeLinkUrl.
Bug: 57909
Change-Id: I77dc308d0d016c395ad737c08cf10a7711e25bbd
This includes the extension name, and it also does much
more stringent validation. In the (now rather unlikely)
event of a duplicate name, it will append a number.
This is important, as it is very confusing when this bug strikes.
There exists extensions like CharRangeSpan which will trigger this bug.
Bug: 42174
Change-Id: Idf14b4cbdb8ec103340d48855e0361acf707b101
* All content handlers that deal with code/data tend to have
English as their page language & pageview language, so moved common
code to the abstract CodeContentHandler class.
* Renamed JSONContent & JSONContentHandler into JsonContent*
Change-Id: I46819a0572ef5becc211d0d82471ff7102edaa3c
BitmapHandler has a lot of generic-ish functionality that could
be re-usable by extension classes (Such as how it organizes
$scalerParams array, or various image magick escaping methods).
However it's combined with a lot of very format specific things,
such as the shell-out call to image magick.
Try to separate out the more generic stuff into
TransformationalImageHandler. In order to do this, I also made
canRotate, autoRotateEnabled, and getScalerType non-static. No
extensions in our repo appeared to be using these methods, and they
don't really make sense to be static (imo).
In particular, I think code duplication can be reduced in
PagedTiffHandler by extending this new class. See comments
on I1b9a77a4a56eeb65.
Change-Id: Id3a8b25a598942572cb5791a95e86054d7784961
Can't be implemented as a test case because that's not being
run before MediaWikiTestCase::tearDown.
Also there's no parent destructor, so no need to call one.
Also fixed BagOStuffTest.
Change-Id: Ifd8659b6c8a748c6716099f8822e2c50ab51bda7
Introduces a new interface for Config types
that are mutable. Not all Config instances
should be mutable, like MultiConfig.
Change-Id: I56e193cbbf72b7afdf551f60ea635fc347e14b3e
Follows-up 9272bc6c47, 03c503da22, 1e063f6078.
One can't wrap arbitrary JavaScript in an if-statement and have
its inner-body mean exactly the same.
Certain statements are only allowed in the top of a scope (such
as hoisted function declarations). These are not allowed inside
a block. They're fine in both global scope and local function
scope, but not inside an if-block of any scope.
The ECMAScript spec only describes what is an allowed token.
Any unexpected token should result in a SyntaxError.
Chrome's implementation (V8) allows function declarations in
blocks and hoists them to *outside* the condition. Firefox's
SpiderMonkey silently ignores the statement. Neither throw a
SyntaxError.
Rgular ResourceLoader responses only contain mw.loader.implement()
and mw.loader.state() call which could be wrapped without issues.
However such responses don't need wrapping as they're only made
by mediawiki.js (in which case mw is obviously loaded). The
wrapping is for legacy scripts that execute in the global scope.
For those, let's wrap the script tag itself (instead of the
response). That seems like the most water-tight and semantically
correct solution.
Had to bring in $isRaw from ResourceLoader.php, else the startup
module would have been wrapped as well (added regression test).
Bug: 69924
Change-Id: Iedda0464f734ba5f7a884726487f6c7e07d444f1
In cases where an extension has multiple API modules that
have the same exact factory requirements, it would let them
use the same factory function and check with the module manager
what the class name to construct should be. This is already possible
using getNamesWithClasses, but this makes it much more
straightforward.
Change-Id: I889e3e6f15907896a8df376913125d903debe984
