TL;DR Many older parser bugs were pushed to the new parser.  However,
several, mostly newer, bugs were assigned to be worked on now.  To
contribute to the new call for bugs, read <http://hexm.de/20>.

This past Monday, I held a bug triage meeting with WMF developers to
cover Parser bugs that I and other participants on this list had marked
for triage.  There were 27 bugs in total.  Because of the open-ended
nature of the call for bugs, we ended up with a fair number of
“oldie-but-goodie” bugs.  With Brion's parser rewrite currently being
planned, this meant that a fair number of bugs were easy to shrug off.
If we've lived with the pain so far, the rationale is, why go through
they horrible pain of fixing it in the current parser when a more
maintainable parser is just around the corner.

Still, it was helpful to look at these older bugs to see the sort of
problems that need to be addressed in the parser rewrite.  The bugs that
we decided it would be better to address in the new parser were the
following:

    Preceding text and single apostrophes are not included in links
        http://bugzilla.wikimedia.org/468

    Incorrect parsing of table headings and cells on the same line
        http://bugzilla.wikimedia.org/549

    [[#foo|]], [[/bar|]] should be equivalent to [[#foo|foo]], [[/bar|bar]]
    (new use of "pipe trick")
        http://bugzilla.wikimedia.org/845

    Newline as list item terminator is troublesome
        http://bugzilla.wikimedia.org/1115

    pre over multiple lines in lists
        http://bugzilla.wikimedia.org/1581

    Need method for multiparagraph list items, continuing numbered lists,
    and assigning specific numbers to list items
        http://bugzilla.wikimedia.org/1584

    Allow one blank line in list environments
        http://bugzilla.wikimedia.org/9342

    Automatic nbsp is inserted even into XHTML attributes, including style
        http://bugzilla.wikimedia.org/3158

    The newline added to a template, magic word,  variable,  or parser
    function that returns line-start wikicode  formatting (*#:;)  causes
    unexpected parsing
        http://bugzilla.wikimedia.org/12974

    Leading spaces in <pre> block render incorrectly when block preceded by
    another <pre>
        http://bugzilla.wikimedia.org/3230

    Blank lines at the top of an article should be ignored
        http://bugzilla.wikimedia.org/4161

    Single newlines sometimes create paragraphs
        http://bugzilla.wikimedia.org/9207

    Block element written inline splits multiline paragraphs
        http://bugzilla.wikimedia.org/5718

    Linebreaks are mishandled in <blockquote> and <li>
        http://bugzilla.wikimedia.org/6200

    Multiline tags in lists should be output more intelligently
        http://bugzilla.wikimedia.org/9996

    Bold/italic markup handled differently depending on leading whitespace
        http://bugzilla.wikimedia.org/18765

    post expand size counted multiple times for nested transclusions
        http://bugzilla.wikimedia.org/13260

Additionally, Brion punted this bug to the new rich text editor he has
planned since the problem is seen mostly in copy-and-pasted URLs:

    External URL syntax cannot handle square brackets
        http://bugzilla.wikimedia.org/3695

One bug was dismissed as WONTFIX with the justification that the
reporter had a certain behavior in mind for the behavior of the parser
when he wrote ''''lots of quotes'''' but that while the parser acted
consistently, it didn't act in his preferred manner

    Single quote inside triple quote bold (''') parsing error
        http://bugzilla.wikimedia.org/13227

Still, all was not lost.  For example, Neil saw this ancient bug as an
opportunity to get closer to the gnarly internals of MediaWiki.

    tilde signatures inside nowiki tags sometimes get expanded
    (<includeonly><nowiki>~~~~</nowiki></includeonly>)
        http://bugzilla.wikimedia.org/93

Sam saw this bug and decided it looked like it would be easy to test and
apply the included patch:

    Transcluded special pages expose strip markers when they output
    parsed messages
        http://bugzilla.wikimedia.org/16129

Finally, Tim saw these two relatively recent bugs and decided he would
investigate them further and hopefully fix them:

    DOM preprocessor barfs on headings inserted by parser functions
        http://bugzilla.wikimedia.org/21844

    {{fullurl:}} does not urlencode passed querystring
        http://bugzilla.wikimedia.org/27972

To see the notes from the Bug Triage (thanks, Sumana!) visit
http://etherpad.wikimedia.org/BugTriage.

Please see my earlier email to the list (http://hexm.de/20) if you'd
like to contribute to this coming week's triage.

Mark.

_______________________________________________
Wikitech-l mailing list
Wikitech-l@lists.wikimedia.org
https://lists.wikimedia.org/mailman/listinfo/wikitech-l

Reply via email to