the includes/parser/MediaWikiParserTest.php test case
* Drop now unused various test recorder options -- phpunit outputs its data in lots of machine-readable formats, use those if you need to work with
test data
* There's still a lot of duplication between NewParserTest::setUp() and MediaWikiTestCase, but hey one step at a time ;-)
* All tests pass for me (make phpunit && make parser)
* Made 'success', 'successCount', and 'failCount' fields reflect the overall operation in FileBackendMultiWrite::doOperationsInternal(). This makes it match up with single-write backends.
* Made FileBackend::clearCache() part of the public API.
the includes/parser/MediaWikiParserTest.php test case
* Drop now unused various test recorder options -- phpunit outputs its data in lots of machine-readable formats, use those if you need to work with
test data
* There's still a lot of duplication between NewParserTest::setUp() and MediaWikiTestCase, but hey one step at a time ;-)
* All tests pass for me (make phpunit && make parser)
* Added getFileSize()/getFileStat() functions. Refactored some functions to use the stat function for better reuse and caching/consistency.
* Refactored streamFile() to allow for subclasses to avoid local file copying with less duplication. Also make last-modified check actually work since we always get the timestamp of the original file.
* Renamed 'ignoreErrors' parameter to 'force'.
In FileBackendMultiWrite:
* Simplified how read ops are done (use 'master' backend for consistency).
* Added consistency check to doOperationsInternal() to check if the files are synced.
* Various fixes after testing.
In StreamFile:
* Split out prepareForStream() function from stream() in StreamFile for code reuse.
In FileBackendTest:
* Properly cover FileBackendMultiWrite in tests.
* Various test improvements.
* Also pass the line number
* Removed useless usage of $title when throwing the exception about invalid since that variable is always null
* Added $ignoreDuplicate parameter to ParserTest::addArticle()
* Replaced 'media' portion of container names with the repo name. This makes it easy for multiple repos to use the same backend without 'wikiId' hacks. Full container names are now like <wiki>-<repo>-<zone> (or <repo>-<zone> if 'wikiId' is set to an empty string).
* Restricted isValidContainerName() more in light of Azure portability and shorted shard suffix.
* Bumped $maxCacheSize to 75 storage paths.
* Code comment cleanups and additions.
Unit tests:
* Updated related tests and marked testBug29408() as broken (I can't find the problem).
* Reduced leakage in UploadFromUrlTestSuite a bit.
* Make sure FSFileIterator::current() directly returns the string path and that it is relative to the directory being searched.
* Fixed silly bug in testGetFileList that masked any bugs.
* Added a few code comments.
* mw.config is the new way, and global config variable lookups are deprecated
* Based on two phase3-wide quick searches:
-- of " wg": http://toolserver.org/~krinkle/wikimedia-svn-search/view.php?id=321&hash=81700bf7486e4fee3b7bc1f83eb9eba6
-- of "!wg": http://toolserver.org/~krinkle/wikimedia-svn-search/view.php?id=327&hash=47c9d54a7a1d5d58a724dd834585f40d
Related changes:
* Changed some php comments mentioning "wg" variables to include the dollar sign, and a typo when the wf function prefix was meant.
* Removed TODO comment in wikibits.js and made it use the JS equivalent of wfUrlencode, which we have now, mw.util.wikiUrlencode
* SpecialUpload.php: use OutputPage::addJsConfigVars instead of creating a new script tag through OutputPage::addScript(Skin::makeVariablesScript(..))
* Renamed wgUploadSetup in upload.js and made it local. Not used anywhere in ./trunk/phase3 and ./trunk/extensions
* Fix OutputPage::addJsConfigVars so that it can actually be called with an array instead of two arguments for key/value
* Some minor whitespace/convention stuff around the same line
* Refactored DatabaseBase::sourceStream(), made it possible for descendant classes to alter its behaviour w/o having to redo it completely like Oracle does.
* MySQL class now supports specifying DELIMITER.
* Thrown away the mess of catering for double semicolon. If it's a problem, fix your .sql files!
* Haven't actually touched Oracle.
* Tests!
* Changed concatenate to store to a specified temp FS file rather than a final storage destination. This makes it better fit the use case (chunked upload), so we can avoid extra copying around. Subclasses no longer have to implement this function now as well.
* Added extensionFromPath() helper function.
* Moved createInternal() up a bit and fixed @see comments pointing to the wrong functions.
In FSFileBackend:
* Use parent implementation of doConcatenateInternal().
In FileRepo/File:
* Added FileRepo::ALLOW_STALE and made thumbnail transforms use it.
* Refactored FileRepo::initZones() to require the specific zones; nothing was using it differently.
* Removed deleted zone check in deleteBatch(), a similar error will instead trigger with the initZones() call as needed.
* Added $wgLocalFileRepo comment.
* Updated tests.
* Added some new methos such as Watchable/Subpages/Capitalized/Content
* Fixed namespace 0 being converted to NS_ , now show NS_MAIN
* Converted all assertion calls to use the __call() hack
Follow up r105883
This is the wrong configuration format for such a feature, and the wrong interface. We already have certain per-namespace permissions in the Title class, and we didn't need to add extra formal parameters to a whole lot of User methods in order to get them. The feature should be implemented wholly in Title, and the concept of user rights should remain relatively simple and easy to understand, and independent of its many applications, i.e. a user either has a right or doesn't. Rights are just a tool for developing access policies; the complexity should be in the caller.
The revert was mostly done by hand, since there were a lot of conflicts. I tried to preserve the gist of conflicting changes in r102187 and r102873. The test changes are not simple reverts, rather I just edited out the per-namespace tests. I reverted the followups r92589 and r104310.
- 'foo' => '' now outputs "&foo=" instead of the key being omitted
- 'foo' => null and 'foo' => false now omit the key instead of outputting "&foo="
- Added a test to make sure that 'foo' => true outputs "&foo=1"
- Fixed a php notice caused when passing a =value-less bit like "&qwerty" to wfCgiToArray by treating it like php and extracting it as 'qwerty' => ''
- Updated tests
- Update our woefully out of date doc comment for WebRequest::getPathInfo (we haven't simply been extracting a PATH_INFO for ages)
- Make PathRouter::makeWeight protected
- Add more comments to the PathRouter code
- Add two more edge case tests to the PathRouter tests.
A regex in SearchUpdate was built for ancient pure ISO 8859-1 and looked for \xa0-\xff bytes -- this caused the regex to cut off partway through if there was a char containing a byte in the \x80-\x9f range.
Fixed regex to pass \x80-\xff instead.
Added a test case to SearchUpdateTest which checks for this case (example text run through the update squash algo, then run through preg_replace with a /u param to make sure it gets treated as UTF-8 and checking whether it breaks.)
Articles containing [[es:Plantilla:Ficha_de_película]] showed the categories at the bottom, but were
not in categorylinks (solved by a doing a null edit to the template and waiting for the job queue, but
how did that situation happen to begin with?).
r103915 added a parse error for 'more than 2 decimal points' in a number; this is the wrong place to check for that. Should only check whether there's more digits or identifier chars -- identifier chars would be illegal.
Added test cases for the exponent missing fails, tweaked it to be more consistent (only need to check for one e; if we have more we can lump them in with 'not digits' :)
Also cleaned up no-longer-needed suppress/restore warnings around JS parser invocation
* MWNamespace::equals to test equivalence of two namespaces (forward compatible with any changes we may make like introducing namespace keys like 'USER')
* MWNamespace::subjectEquals to test equivalence of the subject of two namespaces e.g.: MWNamespace::subjectEquals( NS_USER, $ns ); instead of testing for equivalence to both NS_USER and NS_USER_TALK
* Title::inNamespace to use instead of $title->getNamespace() == NS_???
* Title::inNamespaces for use like $title->inNamespaces( NS_USER, NS_PROJECT ) when you only care if it's in one of a number of namespaces (also accepts an array)
* Title::hasSubjectNamespace for use instead of testing for equivalence to both the subject and talk such as NS_USER and NS_USER_TALK.
Include phpunit tests for all this new code, and also add some tests for some existing code.
This will trigger 2 test failures, where an exponent in a JS numeric literal gets split over line breaks at the '-' or '+', causing a parse error in the resulting output.
A number with the same string length but without using + or - in the exponent passes through fine, indicating that it's the -/+ that's getting misinterpreted.
bug 32168. This function is the inverse of wfParseUrl and is useful when you
need to modify part of a URL and have to put it back together.
Further, with the addition of this function, there is sufficient code in core
to create a proper URI class.
It might be a good idea to have a commit hook or similar to catch these.
You can find candidates for renaming with this command:
$ find tests/phpunit/includes/ -name '*.php' | grep -Ev 'Test.php$' | xargs grep -l MediaWikiTestCase
bug 32168. This implements RFC3986 Section 5.2.4.
http://tools.ietf.org/html/rfc3986#section-5.2.4
This is important because you need to remove dot segments in order to safely
compare URLs when limiting URLs to a particular path.
tempnam() like, not the two-words filenames of RandomImageGenerator::getRandomFilenames()
They aren't produced by ApiTestCaseUpload::fakeUploadFile() either.
Part of bug 32118: test special pages SQL queries against all supported DB
Still need to add all the other non QueryPage special pages and then setup
jenkins to support other databases.
Some preprocessor could have preprocessToObj as a native type.
PHP Fatal error: Call to a member function __toString() on a non-object in phase3/tests/phpunit/includes/parser/PreprocessorTest.php on line 119
* Make tests work again
* Added ApiTestCase::doApiRequest() and related to pass it to the context and removed override of $wgUser where possible
* Fix ApiLogin to get the correct User object (i.e. the logged in one)
* Fix ApiBase to feed a RequestContext to setContext() so that ApiLogin can call setUser()
For some reason, the ApiUploadTest::testLogin method no more returns a valid
session. Since most, if not all, of the ApiUploadTest tests depends on it
we want that testLogin method to fail so we can skip the rest of the tests.
Test command:
./phpunit.php --filter ApiUpload --tap
BEFORE:
=======
TAP version 13
array(0) {
}
ok 1 - ApiUploadTest::testLogin
ok 2 - ApiUploadTest::testUploadRequiresToken
not ok 3 - Error: ApiUploadTest::testUploadMissingParams
not ok 4 - Error: ApiUploadTest::testUpload
^C
AFTER:
======
TAP version 13
not ok 1 - Failure: ApiUploadTest::testLogin
---
message: 'API Login must return a session'
severity: fail
...
ok 1 - # SKIP This test depends on "ApiUploadTest::testLogin" to pass.
ok 1 - # SKIP This test depends on "ApiUploadTest::testLogin" to pass.
^C ... more skips
bug 31176 mentionned an issue with {{SERVERNAME}} returning the full
URL instead of servername when using relative URLs (//localhost).
The fix was made by r98193.
This patch test three different URL protocols (http, https & relative)
and check we return the expected 'localhost'.
r85856 fixed a CSS injection issue but lacked testing. This
test verify we properly strip out CSS comments even when the
token delimiter '/*' is backslash-escaped : \2f\2a
Using same technique as ApiExpandTemplates to serialize the object tree back to XML, rather than asking for the DOM implementation's internal XML return function.
Have to also perform normalization on the test cases, as they aren't normalized to what libxml2 serializes. :P
Note that there are 4 test failures currently with Preprocessor_Hash, as it makes a separate <equals> element around = which doesn't appear to be in Preprocessor_Dom's output.
There are many cases this method will fail for non-latin languages,
but not added now since there are bugs reported on that already and results
a rewrite of the method as per UAX 31 standard.
Basically, in XMP you can specify partial date (for example 2011-04 for april 2011 with no day).
We were extracting that, and fulling out the day to 01 if it wasn't present (My original resoning when writing
the code was that the nice localized date functions need a full date, but that's a pretty poor rationale for displaying
wrong date). This changes it to just display such year-month things as YYYY:MM.
Followup r91591, r93020: patch to jsminplus to support Unicode chars and char escapes in identifiers
Fast-path check keeps runtime about the same on most scripts (eg jquery.js parsing was abround 4100ms both before and after on my test machine)
Slow-path code kicks in if plain ASCII word chars don't extend all the way to the next whitespace or punctuation char.
Using PCRE's Unicode properties magic to ensure that we're catching everything, following ECMA-262 edition 5.1 spec.
Note that identifiers using escapes don't get normalized to their UTF-8 form; this might be a nice thing to do as it saves a couple bytes, but currently there's no change made to output.
Added QUnit tests to verify that unicode letter & escapes work in identifiers in all supported browsers (ok back to IE 6, yay)
- sets an image's reported width/height to the logical form (portait image reports itself as portait)
- everything works in logical coordinates when sizing -- we don't touch the physical pre-rotation dimensions again until it's actual low-level resize time. This fixes several problems with incorrect thumb sizing (eg getting a 600x800 image when we asked for something that fits in 800x600 box)
- fixes unit test cases in ExifRotationTest that were reporting that the width/height were coming back with the physical form which we don't want
- removes some test cases on ExifRotationTest that tested dimension swapping in a place where we don't want it
- ensures that only logical width/height need be exposed to API etc, making exif-rotated images work via ForeignAPIRepo
Note that this may actually cause file metadata to get loaded twice during File::getPropsFromPath, as the $image parameter it passes in to the handler's getImageSize function is bogus and can't be used to fetch an already-loaded metadata blob. This should not generally be too expensive though; it's not a fast path.
Rotated files that were uploaded under previous versions may still have their width/height reversed; an action=purge on the file page will refresh it and cause thumbs to be regenerated.
Follows up on r79845, r90016, r92246, r92279, r96687, r97651, r97656, r97659.
Needs merge to 1.18.
This should fix issues reported in http://lists.wikimedia.org/pipermail/wikitech-l/2011-September/055212.html Lots of these tests loaded the exif extension dynamically, but then didn't reset $wgShowEXIF var properly.
Also change some of the markTestIncomplete to markTestSkipped as that seemed more correct.
* There should be no two files with the same name
* Test suite files should be named after the main file containing the code being tested followed by "Test"
This is Bryan's patch from bug 30640 with a couple minor related changes, plus some unit tests.
This also addresses an issue with preventing too-big images from being scaled from r92279.
I also noticed that image magick's rotation support is broken in 6.3.7 (the version I had installed locally. I've since upgraded) I'm not sure if we should be doing something about that...
I did test this without both image magick, and gd (although only very briefly with gd) both seemed to work well. I didn't test any other image scalars.
* Moved generateReason() to WikiPage. This fixes the PureWikiDeletion ext too.
* Make Checkpoint ext call getRawText() as doEdit() does. Aalso, WikiPage doesn't have fetchContent().
* Call getAutosummary() statically in doEdit().
* Fixed outdated code comment.
* Fixed title used in unit test.
* Has been suggested since August 2009 in r54767 (doc-comment from rawElement/element function)
* Implements normalization for these attributes (removal of duplicates and redundant space)
* Adds support for arrays (instead of just string) for these attributes.
* String are still supported, and are converted to arrays to get the same normalization.
* Wrote unit tests (which pass locally: $ php phpunit.php includes/HtmlTest.php)
* Not trigger for the media-attribute. Reason: Although some people think it's space-separated, it's actually comma-separated. Treating them as space separated might even destroy the value. [2] [3]. Neither the html4 or html5 spec documents media-attribute as space-separated, and as of HTML5/CSS3 the media attribute may contain "media queries".
[1] "In the future, other HTML-specific features might be added, like allowing arrays for the values of attributes like class= and media=" in r54767 by Simetrical.
[2] http://www.w3.org/TR/1999/REC-html401-19991224/types.html#h-6.13
[3] http://dev.w3.org/csswg/css3-mediaqueries/#background
Implementation note: I choose to have a single list of attributes that trigger this feature. Some of these attributes only support multiple values and/or are documented as space-separated as of html5 (such as accesskey), but since those attributes in general have existed in html4 as well (just different w3c spec), they are not stripped if wgHtml5 is not true. So if this feature would (eg. for accesskey) would only be done if wgHtml5=true, then people could get output like <a accesskey=Array /> depending on a configuration variable, which will get messy and make developers' life hard.
difference.
Otherwise if the test happens to be run on a second barrier, it could
falsely fail (like it did just now when codur said some hurtful things
about my code failing unit-tests where it really didn't ;)
Some of the other block tests seem to itermitantly fail for no appearent
reason (null comes back instead of the block)... I don't know why.
