Found by running tests under a version of PHP patched to report
case mismatches as E_STRICT errors.
User classes:
* MIMEsearchPage
* MostlinkedTemplatesPage
* SpecialBookSources
* UnwatchedpagesPage
Internal classes:
* DOMXPath
* stdClass
* XMLReader
Did not change:
* testautoLoadedcamlCLASS
* testautoloadedserializedclass
Change-Id: Idc8caa82cd6adb7bab44b142af2b02e15f0a89ee
The Tidy extension for HHVM is nominally usable, but does not provide error
text returns. Tests should use TidySupport to determine whether they should use
the Tidy extension or not.
Change-Id: I0f842ddf8484206fc2a8b01289bc0a6750f34fed
Autolinking free external links is clever about making sure that trailing
punctuation isn't included in the link. But if an HTML entity happens to
terminate the URL, the semicolon from the entity is stripped from the url,
breaking it.
Fix this corner case. This also unifies autolink parsing with Parsoid.
See: I5ae8435322c78dd1df170d7a3543fff3642759b1
Change-Id: I5482782c25e12283030b0fd2150ac55092f7979b
wddx_serialize_value() fails to escape the ampersand under HHVM. It has
been fixed upstream https://github.com/facebook/hhvm/issues/4283 but
has not been released yet.
When running under HHVM and detecting the ampersand is not escaped, skip
the test..
Bug: T75531
Change-Id: Ia58ec20b4daf78cd90da1bdf8af6cac86015c5d7
Follows-up d2a82fcb60. These issues weren't previously exposed
as nothing uses JsonContent by default in core, and the extensions
using it (e.g. EventLogging) lock it down very early. As are
most of these methods were never really put to use (they were
called after the extension does its superset of checking, or
too early and WikiPage ignores it).
Bug fixes
* Empty JSON object was converted to an array by PST conversion.
The beautifyJSON method is intended for prettify purposes but
actually modified the content stored in the database and made
it no longer roundtrip ({} != []).
We can't change getJsonData to return an object since it's
a public method and people use it as an array. So we can't cast
it to a PHP object as that would break back-compat.
Turns out the class doesn't even support non-objects anyway (a
primitive in JSON can trivially cause a fatal as it wasn't
consistently considered invalid, though it didn't actually fatal
due to some lucky spaghetti code in WikiPage).
* Fix beautifyJSON by checking for empty objects to prevent
implicit {} to [] conversion.
* Add isValid() check to fillParserOutput() as it's called early
on. Otherwise it throws a warning that 'foreach' (in objectTable)
iterates over null. In practice it doesn't matter since the
entire parser output is rejected when WikiPage eventually
checks isValid (through Content::prepareSave).
* Consider all non- (PHP) array values invalid instead of just
non-null values.
Enhancements
* Display message "Empty object" instead of a completely blank page
for an empty object.
* Display message "Empty object" or "Empty array" instead of an
empty table cell.
* Render arrays as a list of values (without indices).
* Remove italics from table cells for values. The monospace font
should be enough. It also offsets it from the "Empty"
placeholders (which are italicised).
Refactoring and clean up
* Use FormatJson::parse so that we can use Status to distinguish
between null parse result and thus reliably cache it.
Ideally we wouldn't need to cache it, but right now this code
is pulled apart and called in so many strange ways that we end
up calling this several times.
* Improve fairly meaningless test (testBeautifyJson) that was
calling FormatJson in its data provider, exactly what the method
being tested did. It also provided the test with data that could
never end up in normal usage (a PHP-style associated array with
implied numerical indices).
* Document that this class rejects non-array values.
* Document the problem with WikiPage assumming PST can run on any
content. WikiPage fundamentally still assumes wikitext, in that
there's no concept of invalid content.
* Fix incorrect documentation for getJsonData's return value
(It may return null.)
* Fix incorrect documentation for beautifyJSON's return value.
(It never returned boolean.)
Bug: T76553
Change-Id: Ifed379ba4674a8289b554a95953951886bf2cbfd
The tests in the structure PHPUnit testsuite are used to verify the
Autoloader, ResourceLoader definitions and some basic other tests.
They are not run by Jenkins for MediaWiki extensions since we invoke:
phpunit.php --testsuite extensions
Add the structure directory to the 'extensions' testsuite. That will
start enforcing them on all extensions, so need to be merged with care.
We will want to backport this patch on all release branches we support
and make sure they pass for all extensions/branches :-/
Bug: T78798
Change-Id: Icb88fb3f10d203cfcb61c36c9c92864b92d6739c
The config variable itself and the documentation property were added
in MediaWiki 1.19 (r107919 / c447423593).
The testswarm-injectjs propert was added in MediaWiki 1.20 (5e590be3d6).
We never actually ended up using TestSwarm, and this variable
is not used anywhere I can see.
Having a configuration variable for a documentation page seems
odd. I can't find another instance of this. As it's tied to development
(not for users of the wiki), link to mediawiki.org direcly.
Change-Id: Ib16607683a293b6d6661ed0411dad9a3ff551a08
ApiLogin and ApiCreateAccount calling wfSetupSession, which leaks a
session over the test.
The test RequestContextText needs a clear session to work, so the api
tests should avoid leaking the session.
Doing this in the ApiTestCase because some tests calling ApiLogin over
FauxRequest and that also starts a session.
Change-Id: Icf5cb4d4a2c24c96698cac5bf32147c0c9149ef3
Added a TransformTooBigImageAreaError to allow setting an extra message.
Added also size-*pixel messages to show the value of $wgMaxImageArea
with some formatting.
This error is still throwing for all files, to fix bug T34387 this needs
a follow up with a proper check. I am not sure, if a File::isLocal() is
okay, because files from a DBForeignRepo maybe transformed on the same
server, so the check needs to be done also for this. For APIForeignRepo
the check is done on the foreign server.
Change-Id: Ieba12e424c8bddb1961a30d3f9ea5c8ff241abb5
If an image has a title but isn't linked, set the title on the image
itself rather than not setting it at all.
Bug: T23454
Change-Id: I6302246cf28c84ade805a6a7d4a5e131b6f42e86
Ensure that there is a \b boundary before and after RFC, PMID, and ISBN
links. (Previously we enforced \b boundaries only before free external
links and after ISBN links.) Consistency is a good thing!
In addition:
* \b is not a PHP escape sequence, so you don't need to write \\b inside
a string.
* \b before the numeric part of an ISBN is pointless: by the structure
of the regexp there will always be a space on the left and a word
character (a digit) on the right.
Bug: 65278
Change-Id: Ic315b988091a5c7530a8285b9249804db72e55db
The @requires just works for other tests, not for php functions.
This fixed a failure for the hhvm test server:
1) ApiFormatWddxTest::testValidSyntax
Failed asserting that null is of type "array".
Change-Id: Iadd8238ce62d4803cdf56508886ad016f8ac8f1d
* getElementById, getElementsByTagName()[0] and jQuery#get(0) all
return an HTMLElement. Test that only once.
Avoid using random internal QUnit elements like #qunit-header.
These don't always exist (e.g. when running tests via Karma).
If anything, it should use #qunit-fixture. But might as well
create a simple DIV and test that.
* Use 'null' instead of getElementById('not-found'). We're
testing isDomElement, not getElementById.
* Add test for TextNode.
* Separate tests for isDomElement and isEmpty.
Change-Id: If9284fbbd066897561ae8e679f5f7ceab9255e18
We already did this, but it was rather convoluted with lots of
if/elseif sequences checking all the possible values.
Remove this logic from ResourceLoaderStartUpModule. Simplying it
simply create the array and pass it to ResourceLoader::makeLoaderRegisterScript.
In makeLoaderRegisterScript, we apply a filter to the array(s) that
trim empty values.
While at it:
* As with other registration properties' default values (like for dependencies,
group, and skip) also use 'null' for the default value of 'source'.
The mediawiki.js client was already compatible with this, and the server omitted
it if it was the last value in the list. But in all other cases it explicitly
outputs "local". Use null instead of simplicity sake. This also gains us a
few characters in the output, and a relatively larger win after gzip since
there's lots more re-using of "null".
* Remove stray casting of $version to int. This only happened in case of
registering a single module (which don't do anywhere), and is redundant.
Change-Id: I1f321e7b8bd3b5cffc550b51169957a3da9b971d
By using the existing indexes of modules in the array being passed to
mw.loader.register we can reduce the size of the startup module by about
6% after gzip (nearly 20% before) on a wiki with very few modules (such
as my localhost). Comparing data from en.wikipedia.org shows about 9%
after gzip (nearly 30% before).
The technique adds a function to mediawiki.js which resolves the indexes
before registering the modules, which costs a little bit of data in that
payload, but it's negligible (118 bytes after gzip) in comparison to the
overall reduction.
Also, cleaned up lies in documentation and strange use of "m" as an
iterator variable.
Bonus: fix ISO8601 timestamp instead of UNIX timestamp being passed
to custom loader scripts.
Change-Id: If12991413fa6129cd20ceab0e59a3a30a4fdf5ce
* Use time() instead of:
- wfTimestamp()
- wfTimestamp( TS_UNIX )
- wfTimestamp( TS_UNIX, 0 )
- wfTimestamp( TS_UNIX, time() )
- intval( wfTimestamp( TS_UNIX, time() ) )
* Consistently use 1 as default instead of 0. Previously the
unwritten convention was that anything "final" max()'ed with 1,
and any internal method would use 0, but this wasn't applied
consistently made it fragile. There doesn't seem to be any
value in returning 0 only to have it maxed up to 1 (because if
the 0 would ever make it out alive, we'd be in trouble).
* wfTimestamp returns a string for TS_UNIX. In PHP this doesn't
matter much. In fact, max() takes number-like integers so
transparently, it even preserves it:
> max( 1, 3, '2' );
< 3
> max( 1, '3', 2 );
< "3"
Just cast it in one place at the very end (StartupModule)
instead of doing intval( wfTimestamp() ).
* Fix weird documentation claiming getModifiedTime can return
an array, or mixed.
* Remove 'version > 1 ? version : 1' logic in
ResourceLoader::makeLoaderRegisterScript. The client doesn't
have "0 means now" behaviour so this isn't needed. And the method
was only doing it for variadic argument calls.
Removal of quotes around timestamps reduced the size of the startup
module from 26.8KB to 25.9KB before gzip. After gzip the size was
and still is 5.7KB, though. (From 5456 bytes to 5415 bytes.)
Change-Id: If92ca3e7511e78fa779f2f2701e2ab24db78c8a8
- Added/removed spaces around parenthesis
- Added newline in empty blocks
- Added space after switch/foreach/function
- Use tabs at begin of line
- Add newline at end of file
Change-Id: I244cdb2c333489e1020931bf4ac5266a87439f0d
IE12's User-Agent is, unhelpfully:
Mozilla/5.0 (Windows NT 6.4; WOW64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/36.0.1985.143 Safari/537.36 Edge/12.0
Unsurprisingly, this is currently recognized as Chrome 36.
The word "Trident" also doesn't appear any more, because IE12's
new layout engine is called "Edge". So look for that string instead.
Change-Id: I94a4ff2cc37ae6ce0c677ef92eeb6a2dff30c1f4
If $wgMessagesDirs is initially empty, we can optimize when batch loading
extensions:
if ( !$wgMessagesDirs ) {
$wgMessagesDirs = $cache['MessagesDirs'];
}
With APC, this should be O(1) CPU time.
This was suggested by Tim in the code review of I7074b65d07c5.
Change-Id: I66fa907cdaafe18b74b5b9afaa8b6b1db069bea3
Was not completely removed and the remaining bits were mostly
broken. $wgProfiler['output'] = 'db' is the proper way.
Fixes T75917
Change-Id: I36565e2372db2ed49b219cf533ec433e8111c52f
