[CSSWG] Minutes and Resolutions 2010-10-27

Summary:
   - New snapshot build of CSS 2.1 test suite published; not all bugs fixed yet.
   - Deferring CSS2.1 Issue 101 to see if change is actually needed for web
     compat; IE9 implements per spec.
   - RESOLVED: CSS2.1 Issue 154 (images to clarify inline measurements) closed
               wontfix; prose edits have improved the text already
   - RESOLVED: Proposal accepted for CSS2.1 Issue 159 (margin collapsing rewrite)
               with Bert's "no padding or margin areas" change and "drop 'vertical'"
               changes; fantasai to write example clarifying adjoining behavior
               of collapsed margins
   - RESOLVED: Publish CSS3 Backgrounds and Borders as CR. Disposition of Comments:
                 http://dev.w3.org/csswg/css3-background/issues-lc-2010
   - Reviewed Bert's follow-up comments on edits to CSS2.1 Issue 120 and 142.

====== Full minutes below ======

Present:
   Tab Atkins
   David Baron
   Bert Bos
   Beth Dakin
   Arron Eicholz
   Elika J. Etemad
   Simon Fraser
   Sylvain Galineau
   Koji Ishii
   John Jansen
   Brad Kemper
   Chris Lilley
   Peter Linss
   Alex Mogilevsky

Scribe: Sylvain

<RRSAgent> logging to http://www.w3.org/2010/10/27-CSS-irc

Administrative
--------------

   plinss: any other agenda items ?
   howcome: yes, I'd like to talk about multicol

CSS2.1 test suite
-----------------

   fantasai: I published a build last night
   fantasai: still need to write release notes detailing changes
   arronei: I didn't have time to finish all the testcases. I changed
            700 files from feedback. remaining changes are about 30 files.
   fantasai: I haven't gone through all feedback so I'm not sure how much
             is left. but everything is on the mailing list
   fantasai: so this latest update is just a snapshot
   arronei: also we should start planning on locking/stabilitizing the
            test suite about a month before PR i.e. no more functional
            changes, only small editorial updates
   fantasai: if a test is wrong we should update it though
   arronei: we'll have to agree on which tests get fixed vs. handled in
            the next release of the test suite
   fantasai: I think it'd be harmful to have erroneous testcases in the suite
   arronei: we might be able to remove them; we'll have to evaluate by then.
            I don't expect this to be a frequent problem.
   plinss: I'd actually like to lock down functional changes earlier if possible
   arronei: sooner is better. Middle of November ?
   <dbaron> (I think some of the discussion above also discussed what tests
            should be there for the version of the test suite used to qualify
            the spec, not what tests should/shouldn't be there forever.)
   fantasai: first we need to go through all the feedback on the mailing list
   plinss: what is your ETA to go through the feedback ?
   fantasai/arronei: unsure
   plinss: we need to stop adding test or making major changes so the sooner
           we can lock down on functional changes the better
   plinss: mid-november is aggressive given TPAC
   arronei: let's target that
   ChrisL: the sooner the better. yes we need to process the feedback and fix
           updates. but we could keep doing this for a long time.
   plinss: yes we're getting to the point where we'd remove erroneous testcases
           unless they absolutely have to be fixed

CSS2.1 issues
-------------

   <plinss> http://wiki.csswg.org/spec/css2.1#issue-101
   Bert: my opinion is the same. we shouldn't change the spec. this is
         implemented incorrectly
   TabAtkins: it's implemented incorrectly the same by everyone
   TabAtkins: the change is a matter of making the model described in
              the spec match reality
   dbaron: the change I had made that caused regressions was anologous
           to these rules but not directly related to them
   johnjan: IE9 now passes these testcases so we should investigate the
            actual compat impact first
   johnjan: I'd rather assert a regression tail before changing the spec
   (johnjan will try to get compat data to discuss at TPAC)
   <dbaron> https://bugzilla.mozilla.org/show_bug.cgi?id=472252 was the
            bug I had the next day when I broke one of these cases

   http://wiki.csswg.org/spec/css2.1#issue-154
   arronei: I haven't had time to update these.
   arronei: the prose edits have already improved clarity substantially
   plinss,ChrisL: not worth waiting for then. Resolved.

   http://wiki.csswg.org/spec/css2.1#issue-159
   Bert: I like the text; it's clearer than the current version
   Bert: I have suggested minor improvements; box and elements are often
         mentioned in the same sentence without distinguishing them. The
         things that adjoin are boxes, the things that are siblings are
         the elements
   (Bert discusses the wording described in his review)
   Bert's comments: http://lists.w3.org/Archives/Public/www-style/2010Oct/0819.html
   <Bert> -> http://www.w3.org/mid/201010271802.31842.bert@w3.org my review of 159
   <dbaron> I suggested making "adjoining" not transitive in
            http://lists.w3.org/Archives/Public/www-style/2010Jul/0507.html
   <dbaron> but it's technically ok either way; just a matter of how we
            define the terms
   * dsinger_ seems odd that A can adjoin B while B does not adjoin A
   (discussion of adjoining margin definition that the scribe cannot keep up with)
<Zakim> -bradk
   fantasai: I like Bert's other suggested changes; but I think the way the
             terms adjoining and collapse are currently defined is fine
   <fantasai> >   | A collapsed margin is considered adjoining to another margin if
   <fantasai> >   | any of its component margins is adjoining to that margin.
   <fantasai> ACTION fantasai: write example to clarify this
   <trackbot> Created ACTION-270
   <trackbot> Created ACTION-271
   RESOLVED: Proposal accepted for CSS2.1 Issue 159 with Bert's
             "no padding or margin areas" change and "drop 'vertical'"
             changes; fantasai to write example clarifying adjoining

   http://wiki.csswg.org/spec/css2.1#issue-199
   <fantasai> I would suggest s/auto position/static position/
   <fantasai> And replace "the top of the box in the vertical direction,
              and the same position horizontally that" with "the same as"
   <fantasai> Since it's zero-height, it doesn't matter.
   RESOLVED: Proposal accepted for CSS2.1 Issue 199 with changes above
   (Bert and Tab discuss the change's exact insertion point in the spec)
   <dbaron> Issue 199 is changing text in 9.4.2

agenda items for TPAC
---------------------

   plinss: please email your items, review the wiki and update as necessary
   plinss: let's try to work out Sunday night dinner plans over email

Publishing CSS3 Backgrounds & Borders CR
----------------------------------------

   <fantasai> http://dev.w3.org/csswg/css3-background/issues-lc-2010
   fantasai: I don't see anything blocking publishing
   ChrisL: transition call to exit LC won't happen this week or next but
           we can get the process started
   RESOLVED: publish CSS3 Backgrounds & Borders CR

CSS2.1 Issues Review from Bert
------------------------------

   http://wiki.csswg.org/spec/css2.1#issue-120
   Bert: this is the large rewrite that updates containing block definitions
   Bert: I noted that it unintentionally no longer makes tables containing
         blocks and we need to fix that
   Bert: changes are in the editor's draft, 9.2.1.
   fantasai walks through the spec and points out that there's no actual
     error, due to the distinction between table wrapper box and table box;
     however a clarification might be useful; filed as Issue 212

   http://wiki.csswg.org/spec/css2.1#issue-142
   <Bert> http://lists.w3.org/Archives/Public/www-style/2010Oct/0042.html
   Out of time; actual discussion deferred to next week.

Meeting closed.

Received on Wednesday, 27 October 2010 22:16:02 UTC