* Don't create a WebRequest instance to put in $wgRequest in test since it is meant for HTTP requests, not command line ones
* Force $wgAlwaysUseTidy to false in ExtraParserTest, having it to true breaks testBug8689() and testParse()
* wfGetIP() now support resetting its internal static variable. Thanks to
Platonides which introduced this trick with r92960.
* Various tests for $_SERVER['REMOTE_ADDR'] and $wgCommandLineMode.
revert r94575:
- reenable testGetFromServerRemoteAddr() which was not an issue
reintroduce r94558:
- per CR on r94558 by Aaron use meaningful parameter to wfGetIP() when
resetting the static variable ( 'reset' instead of true).
- keep testLackOfRemoteAddrThrowAnException() test in the broken group
with a comment for later fixing.
TODO:
- implements tests for XFF headers.
TEST PLAN:
$ ./phpunit.php --filter wfGetIP --testdox
PHPUnit 3.5.14 by Sebastian Bergmann.
wfGetIP
[x] Get loopback address when in command line
[x] Get from server remote addr
[x] Lack of remote addr throw an exception
$
* Don't assume $ as a global, but alias it to jQuery instead. The code is already wrapped in a closure so this is easy
* Remove unused constants TYPE_SIMPLEFUNC and TYPE_OBJCONSTRFUNC
* Fix spelling in comments, mostly brake->break
* Fix whitespace
* Kill useless return at the end of a void function
* Correct comment in hasTests() which claims the function does the opposite of what it really does
Check to make sure the title object actual exists, before checking its namespace,
since alll the does this have illegal chars check is done after the does it
start with file: check.
For future note though, when using upload by url feature, the suggested
destination name when you put http://upload.wikimedia.org/wikipedia/commons/e/e0/Petrorhagia_prolifera_%281%29.JPG is illegal, which is probably not very user-friendly (filed as bug 30390)
- reverts ./includes/ProxyTools.php
- marks tests broken
In our test suite, the first call to wfGetIP() set the static variable.
Hence the remaining of the code is only tested on the first call to it.
Resetting the static variable enlight a bug somewhere in our code where
we are calling wfGetIP() but can not reliably get an IP, somehow REMOTE_ADDR
does not exist and $wgCommandLineMode is disabled.
Will have to track this bug further when I got time.
This code appears to be trying to output something like 'HTTP/1.1 500 MediaWiki Error', but using $_SERVER['SERVER_PROTOCOL']. And it seems to output it on things that run in phpunit tests.
Seems pretty broken?
* wfGetIP() now support resetting its internal static variable. Thanks to
Platonides which introduced this trick with r92960.
* Various tests for $_SERVER['REMOTE_ADDR'] and $wgCommandLineMode.
TODO:
- implements tests for XFF headers.
TEST PLAN:
$ ./phpunit.php --filter wfGetIP --testdox
PHPUnit 3.5.14 by Sebastian Bergmann.
wfGetIP
[x] Get loopback address when in command line
[x] Get from server remote addr
[x] Lack of remote addr throw an exception
$
Per CR on r92234, this correctly test hasSubpages independently from your
local configuration. Also test altering the global and having static
methods reacting accordingly.
* Introduce a boolean parameter to wfUrlProtocols() which, if set to false, will cause '//' to be dropped from the returned regex so it doesn't match protocol-relative URLs
* Introduce wfUrlProtocolsWithoutProtRel() as a wrapper for wfUrlProtocols( false ). The latter should not be used directly because the former is much clearer
* Use this new function in Parser::doMagicLinks() to fix the original bug. Also use it in ApiFormatBase::formatHTML() and CodeCommentLinker::link(), which probably had similar bugs
Simple use case (PHP 5.3+) that will work show the expand text passed to a <preprocess> tag:
$wgHooks['ParserFirstCallInit'][] = function( $parser ) {
$parser->setHook( 'preprocess', function( $text, $attr, $parser, $frame ) {
return $parser->recursivePreprocess( $text, $frame );
} );
return true;
};
Note that there were some attributes excluded. Namely stuff on img and object, and the table cellspacing and cellpading which aren't easily converted into inline css.
Since it's only used right now for paranoia in cases you currently can't actually exploit a browser we let it slide.
However this thing needs a big fat warning message next to it to avoid someone thinking this is actually a good idea for security and ending up later on using it and opening up an XSS hole in core.
