to <div class="center">. <center> is perfectly valid in the target doctype
(XHTML 1.0 Transitional), while the replacement didn't catch all cases and
could even result in invalid output from valid input.
* (bug 2607) Fix for section editing with mix of wiki and HTML headings
* (bug 3342) Fix for section editing with headings wrapped in <noinclude>
* (bug 3476) Fix for section editing with faux headings in extensions
* (bug 5272) Fix for section editing with HTML-heading subsections
* Fix for bogus wiki headings improperly detected with following text
* Fix for HTML headings improperly not detected with preceding/following text
* Section extraction and replacement functions merged into one implementation
on the Parser object, so they can't get out of sync with each other.
* remove require_once() throughout whole code, yet left in few places
* move global functions in HttpUtils, ProxyTools, Credits to class methods
* php5 only: __autoload() now used, combined with class->file map and require()
* move initialization of $wgValidSkinNames to Skin::getSkinNames()
* few more changes that will surely break stuff.
* Nesting of different tag extensions and comments should now work more
consistently and more safely. A cleaner, one-pass tag strip lets the
'outer' tag either take source (<nowiki>-style) or pass it down to
further parsing (<ref>-style). There should no longer be surprise
expansion of foreign extensions inside HTML output, or differences
in behavior based on the order tags are loaded.
* Move parameter replacement before brace transformations in most of the
wfMsg() family (except for wfMsgHtml). This allows things like {{plural:}},
{{urlencode:}}, and {{fullurl:}} to be used in most cases correctly.
The content or UI language will be used accordingly for (forContent)?.
* Work around glitch with the above in {{int:}}; skip redundant transformations,
allow the including parser to do it.
* Fix {{int:}} to use content language, so it won't break caches and links
tables and randomly include data from the wrong language.
Uncertain what the cause is of the breakage, I have a hard time duplicating it
and it looks massively wrong (should not be possible). Failure on Parser.php
line 590, claims var not array, but the check two lines above ensures that
it is first...
Did similar for other unstrip a while ago because it started mysteriously
blowing up; now getting mysteriuos blowups on these too, not sure why.
Only seems ot be affecting zhwiki for some reason
allow the including parser to do it.
* Fix {{int:}} to use content language, so it won't break caches and links
tables and randomly include data from the wrong language.
* Only the first __TOC__ is now turned into a TOC.
The table change doesn't disrupt either the parser tests or the en.wikipedia main page. Hopefully it won't break other real content...
