[CSSWG] Minutes and Resolutions 2008-09-03

Summary:

   Reviewed status of LC specs:
     - CSS3 Color has some editorial comments. Needs implementation reports
       to go to PR.
     - Selectors needs some changes in response to substantive issues
     - 2007 Snapshot is waiting for Selectors
     - CSS3 Paged Media will be republished as LC soonish
     - Print Profile is waiting to be updated to match Paged Media
     - CSS3 Fonts is in LC, technically, but is going back to WD
     - Marquee also in LC

   Discussed grammar of @page wrt @margin-box rules. Created CSS 2.1 Issue 71
     in response.
   RESOLVED: @rules inside declaration blocks are ignored if between declarations
             (i.e. at the beginning of a declaration before the property),
             invalid if within declarations. (CSS2.1 Issue 71)

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

Attendees:
   David Baron
   Arron Eicholz
   Elika Etemad (scribe)
   Sylvain Galineau
   Daniel Glazman (chair)
   Melinda Grant
   Håkon Wium Lie
   Peter Linss
   Alex Mogilevsky
   Saloni Mira Rai
   David Singer
   Jeffrey Wilson
   Mohamed Zergaoui

Status of LC Specs
------------------

   Peter: I wanted to track where we are on things that are in LC
   Peter: Any comments, any progress on test suites
   Peter: Status check to see what we need to do to move forward
   Daniel: XSLFO has editorial comments on CSS3 Color
   fantasai: We should let dbaron handle those, and only bring up comments
             on the telecon if they require discussion. editorial comments
             usually don't
   Daniel: What else is in LC?
   fantasai: Selectors
   Daniel: Still a few changes to the document to make
   Daniel: I still need to make those changes
   ACTION Daniel make edits to Selectors
   http://www.w3.org/Style/CSS/current-work
   fantasai: Snapshot is in LC, waiting for Selectors
   fantasai: Paged Media is at LC, need to publish another LC
   Daniel: Marquee
   fantasai: CSS3 Fonts, but should go back to WD next
   Daniel: Print Profile
   Melinda: Needs to go back out, but waiting for Paged Media
   Daniel: David, did you see editorial comments on css3-color?
   David: LC period ended 2 days ago, so I'll try and gather them all up
   David: Not sure I can do that for next week
   ACTION dbaron Evaluate LC comments for css3-color due two weeks
   Peter: Status of test suite?
   David: Has test suite. Wrote impl report for FF on Linux. Need more
          implementation reports

Other Topics
------------

   Daniel: Next two items on agenda, we are missing the right people
   * anne glazou, see http://lists.w3.org/Archives/Member/w3c-css-wg/2008JulSep/0179.html
          I can't make :/ I have guests coming over

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

   fantasai: I haven't looked into them since the F2F, so I don't think
             we have any open
   Daniel: bolder/lighter issue?
   fantasai: We closed it as undefined for CSS2.1 and deferred to CSS3
   Daniel: Should we review mediaqueries issues without Anne?
   <glazou> anne: postponed, we need you here
   howcome: We should wait until he is here
   Daniel: That doesn't leave much on the agenda for today.. we are done.
   Daniel: Anything else?
   dsinger: We'll have a new version of Animations and Transforms soon
   dsinger: There's a discussion in HTML5 about accessibility of timed media
            (audio/video), might affect this group.
   dsigner: because Media Queries might be a good way to handle that
   <glazou> sorry, me is muted, sneezing and sneezing
   Daniel: Anything else?

Grammar for @margin-box rules inside @page
------------------------------------------

   fantasai: Issue for Paged Media we could discuss
   fantasai: It's a grammar issue
   fantasai: The @page rule in CSS2.1 takes a declaration block
   fantasai: In CSS3 it can also contain @rules
   fantasai: In the grammar in css3-page it says you have to have a semicolon
             at the end of the @rule
   fantasai: that allows declarations after the @rule to be parsed and
             recognized after the semicolon
   Melinda: But several examples don't include that semicolon, and several
            implementations don't require
   Melinda: There are two proposals for fixing this
   Melinda: First is to require semicolons after @blocks
   Melinda: Second is to require all margin box statements to be after all
            other declarations.
   <dbaron> oops
   <dbaron> The appendix G grammar or the ch 4 grammar?
   fantasai: Either way we will need to allow an exception so that UAs that
             parse and recognize @rules without semicolons are conforming
             if their market requires this behavior, but such style sheets
             are not valid CSS
   fantasai: This market is mostly for printers, so browsers should be able
             to parse strictly from the beginning
   <fantasai> http://www.w3.org/TR/CSS21/syndata.html#parsing-errors
   <sylvaing> are they allowed in CSS2.1 ?
   fantasai: Melinda says the DLNA test suite doesn't have semicolons, but
             does place the @rules at the end of the @page rule.
   fantasai: So I would recommend speccing that.
   sylvaing, no they are not
   <sylvaing> so a CSS3 stylesheet may not parse properly with a CSS2.1 UA
   howcome, peter don't like requiring semicolons
   <glazou> howcome does not like pushing nested @rules at the end either
   Peter: do we have a problem with that currently in browsers? Do they drop
          the declaration after the @page rule?
   howcome: It's true that Opera swallows the declaration if it comes after.
   Melinda: who raised this issue?
   sylvaing(?): Microsoft, we noticed it and posted about it.
   howcome: If we say the at-rules have to come after all declarations, then
            we'd have to drop all the declarations after the at-page rule.
   howcome: I think it makes sense to parse at-rules in declaration lists.
            It's a key for extending CSS in the future.
   peter: I think forward-compatible rules should always win, and anytime
          you see an @keyword then you parse it as an @rule.
   <dbaron> I agree with peter (line above).
   Melinda: So we need to make the semicolon optional in CSS3 Page
   dbaron: I don't think this is incompatible with the forward-compatible
           grammar
   Peter: I don't think we need to change the grammar, we need to change
          the parsing rules for errors.
   <fantasai> http://csswg.inkedblade.net/spec/css2.1#issue-71
   howcome: http://www.w3.org/TR/CSS21/syndata.html#at-rules
   fantasai: http://www.w3.org/TR/CSS21/syndata.html#declaration
   fantasai: The problem isn't where does the @rule end, it's does the @rule
             end the declaration.
   property: value; @key { ... } property2: value;
   fantasai: The question isn't whether @key { ... } stays together, it's
             whether it's considered part of the property2 declaration or not
   peter: "User agents must ignore an invalid at-keyword together with
          everything following it, up to and including the next semicolon
          (;) or block ({...}), whichever comes first."
   fantasai: If you take that, then
   <fantasai> p { color: @key { ... } green; }
   <dsinger> ow
   <fantasai> If you take that rule like that, then in this rule my paragraph
              is going to be green.
   <fantasai> So.. I think we've just proven that there's a problem with this
              spec :)
   <glazou> really really ugly
   fantasai: So we don't want to just parse and "ignore" the @rule.
   peter: If it's in between declarations then we want to ignore it, if it's
          within a declaration then we want to throw it out
   fantasai: So someone needs to draft text for this.
   fantasai: it's CSS2.1 Issue 71, which needs a better description
   <sylvaing> invalid if value, ignored if declaration ?
   <glazou> yes
   Daniel: So who is going to draft that?
   <glazou> Bert...
   howcome: What's the resolution?
   fantasai: So my summary would be, @rules inside declaration blocks are
             ignored if between declarations (i.e. at the beginning of a
             declaration before the property), invalid if within declarations.
   fantasai: { property: value @rule; } /* invalid */
   fantasai: { @rule {} property: value } /* ignored */
   Melinda: that should apply only to @rules that are invalid inside declaration
            blocks, not to valid ones
   (in CSS2.1 all @rules inside declaration blocks are invalid, in CSS3 Page
    @marginbox rules are valid inside @page's declaration block)
   RESOLVED: @rules inside declaration blocks are ignored if between declarations
             (i.e. at the beginning of a declaration before the property),
             invalid if within declarations.
   Daniel: Anything else?

TPAC F2F
--------

   Daniel: One more administrative issue
   Daniel: We need to start collecting items for agenda for TP
   * MoZ announce that he will be there on sunday at least
   <dsinger> I deeply regret that I'll miss the TP, but Dean Jackson (ex-w3c)
             will be there
   Daniel notes that there are other hotels within easy walking distance of
          the sofitel
   Daniel: Arcadia or something
   <dsinger> I have stayed at a small hotel in walking, just on the canal,
             can't remember its name
   dbaron, fantasai usually stay at Hotel la Calanque
   <fantasai> http://csswg.inkedblade.net/planning/mandelieu-2008
   <fantasai> Add topics
   <Bert> Sorry, the other telcon was rather long.
   <fantasai> hey Bert
   <Bert> What did you assign to me in my absence?
   <fantasai> we wanted to ask you to write a propsal for CSS2.1 issue 71
   <fantasai> or proposed text rather
   <dbaron> See http://www.mandelieu.com/accommodation/french-riviera-hotel/hotel-accommodation.html
            for a list of hotels in the Mandelieu area

Received on Tuesday, 9 September 2008 01:58:46 UTC