W3C home > Mailing lists > Public > www-style@w3.org > January 2011

[CSSWG] Minutes and Resolutions 2011-01-26

From: fantasai <fantasai.lists@inkedblade.net>
Date: Thu, 27 Jan 2011 13:50:06 -0800
Message-ID: <4D41E88E.3040808@inkedblade.net>
To: "www-style@w3.org" <www-style@w3.org>

   - RESOLVED: Accept proposal to defer 'display: run-in' from CSS2.1 to CSS3.
   - RESOLVED: Drop 'vertical-align' from the required list of properties that
               apply to ::first-line.
   - RESOLVED: Make interaction of bidi and white-space trimming undefined for
               CSS2.1 due to lack of implementations. (Leave current behavior
               as a suggestion.)
   - RESOLVED: Make containing block formed by positioned inline split across
               multiple lines undefined.
   - RESOLVED: Have spec allow two possible margin collapsing behaviors in the
               margin-collapse-clear case: the hypothetical position before
               clearance is applied may be calculated either with respect to
               the parent block, or with respect to the containing block
               formatting context.
   - RESOLVED: publish css3-images WD after Brad approves issue edits
   - RESOLVED: Publish css3-writing-modes as WD

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

   César Acebal
   Tab Atkins
   David Baron
   Bert Bos
   Tantek Çelik
   Arron Eicholz
   Elika Etemad
   Daniel Glazman
   Koji Ishii
   John Jansen
   Brad Kemper
   Hĺkon Wium Lie
   Peter Linss
   Steve Zilles

Many hours earlier in a timezone far away ...

<karl> http://www.w3.org/TR/html5/rendering.html#css-extensions
<karl> it doesn't seem to have been proposed yet to the CSS WG.
<karl> Maybe in the plans?
<karl> >A CSS user agent that implements the text tracks model must implement
        the '::cue' and '::cue(selector)' pseudo-elements, and the ':past' and
        ':future' pseudo-classes.

Later that morning/afternoon/evening/night ...

ScribeNick: fantasai

CSS2.1 Test Suite

   plinss: Yesterday a bunch of us went through tests. We have passes
           or plans for each one
   plinss: some issues resulting
   plinss: first item, we have a number of run-in tests that fail, no
           plans for implementations for those tests
   plinss: plan is to move run-in and its tests to a css3 module
   plinss: any objections from the phone? we have consensus here
   <Bert> Disappointing that such a basic tupography still doesn't work. :-(
          But in the interest of finally getting rid of CSS 2.1...
   plinss: do not plan to remove it, do not plan to prefix it, just move
           definition and tests to css3
   howcome: I wouldn't take it out.
   plinss: Give us a proposal that gets us out of CR within a month
   RESOLVED: Accept http://lists.w3.org/Archives/Public/www-style/2011Jan/0542.html

   fantasai: Drop vertical-align from required properties for ::first-line
   dbaron: Gecko is the only browser that implements it. Spec doesn't
           particularly describe how it's supposed to work other than
           "it's mostly like an inline element"
   dbaron: Spec lists MUST properties and has MAY for all other properties
   dbaron: so our impl would end up in the MAY category
   RESOLVED: Drop 'vertical-align' from list of ::first-line required properties.

   plinss: Bidi, were going to leave the interaction of spacing undefined
   fantasai: What's failing is that impls trim leading and trailing white
             space on a line before bidi reordering instead of after, so
             we have test bidi-004 failing in everything
   fantasai: But that's clearly the right behavior, so we shouldn't change
             the spec, just leave the definition for CSS3
   fantasai: And make undefined for CSS2.1
   <fantasai> "Then, the block container's inlines are laid out. Inlines are
              laid out, taking bidi reordering into account, and wrapping as
              specified by the 'white-space' property. When wrapping, line
              breaking opportunities are determined based on the text prior
              to the white space collapsing steps above. "
   <fantasai> That's followed by the trimming steps in question.
   <fantasai> So we would remove the mention of bidi reordering and add that
              a sentence saying that the interaction with bidi is undefined
              with the steps below.
   <fantasai> (but not above, 'cuz we have interop on that)
   <dbaron> could also say ... but implementations may follow them, or
            something like that
   RESOLVED: do that

   fantasai: other issue was the containing block formed by an relpos'ed
             inline element
   fantasai: that breaks across multiple lines
   fantasai: proposed to make resulting containing block undefined
   fantasai: so would remain defined if inline is all on one line
   fantasai: just not if it's split across multiple lines
   RESOLVED: make containing block formed by inline split across multiple
             lines undefined

   plinss: We had a test that passes in one impl, but if you split the test
           then we have two impls per piece
   dbaron: Splitting the test doesn't help with our exit criteria because
           it's per-feature, not per test
   dbaron: But we could make a convincing argument about implementability
   dbaron: because there are two impls of each part of the feature
   dbaron: and one impl that is capable of implementing both
   dbaron: Suggestion is to keep the test, but also make split versions
   plinss: Question for Daniel, Bert - do you think we can defend that?
   Daniel: Seems bad to have a red line in the impl report
   dbaron: Our exit criteria are per feature anyway, not per test
   Arron: Doesn't seem like we would get a fixed impl soon enough
   Bert: I think you can argue about test suites. MathWG has done the same,
         for example.
   Bert: You can do that. But it is work: you have to prepare those explanations.
   Bert: So I think it depends how much you want the original tests to stay
         in the test suite.
   plinss: I could go either way.
   plinss: I think we're more honest to not remove the original.
   SteveZ: can you flag the test yellow?
   plinss: Yes, we're adding some concepts to the test suite about
           parent-child tests
   plinss: Daniel, you on board?
   Daniel: Yes, that's fine.
   dbaron: I don't want to remove the parent because I can imagine a case
           where impl passes both children, but fails the parent
   plinss: OK. We have a plan.

   plinss: Other issue is margin-collapse, if we want to discuss that now.
   dbaron: I wrote an email to hixie, which I should probably repurpose to
           an email to www-style
   dbaron: Briefly, hyatt proposed changing the margin collapsing rules in 2007
   dbaron: To prevent a counter-intuitive behavior, of elements moving up
           and margins disappearing.
   dbaron: We resolved to fix that.
   dbaron: However, if you read the spec very carefully, the fix didn't
           actually work
   dbaron: The tests are for what the spec was intended to say
   dbaron: But implementations have not changed
   dbaron: And IE claims there are web-compat issues
   Arron: Also Acid2 is affected.
   dbaron: Although I think if we tell hixie to fix Acid2, he'll fix Acid2
   Arron: We have two implementations of the intended behavior, from Antenna
          House and WebToPDF.
   Arron: So we have the passes we need
   Arron: If we don't change the spec, but change the tests, then we have
          passes from all the browsers.
   fantasai: We also noted that there are existing patches for Gecko and
             Trident that would implement the intended behavior, if that
             was chosen.
   plinss: So we're not blocked on this issue
   dbaron: It's easier to leave as-is, but I'm not sure that's good for authors.
   fantasai: Could also leave it undefined for now -- allow both behaviors --
             and lock it down after testing in a beta cycle
   Arron and John Jansen are in favor of leaving the behavior.
   fantasai is not happy with margins, however large, mysteriously disappearing.
   fantasai: I think we should just list the two behaviors and allow one or
             the other. Decide later once we have better chance to check web
   <dbaron> I think I prefer fantasai's suggestion.
   fantasai: Can have two controls within the same test to test for that.
   RESOLVED: Have spec allow either margin collapsing behavior in this
             margin-collapse-clear test

   plinss: going to split some bidi tests up, and fix a few erros, but other
           than that we're mostly done here

CSS3 Image Values

   Tab: I have some pending notes to put in about gradient issues, I will
        put in as issues in the working draft.
   Tab: Leif raised a few small issues
   Tab: Leif wants object-fit: none;
   dbaron: I really don't want that
   Tab: I would like to go to WD
   Brad: I have some notes I want captured.
   RESOLVED: publish css3-images WD after Brad approves issue edits

CSS3 Writing Modes

   fantasai: No major substantive changes. Mainly lots of editorial edits,
             and lots of tweaking of details e.g. in box model section.
   fantasai: Just want to publish an updated WD for comments
   RESOLVED: Publish css3-writing-modes as WD

CSS3 Text

   fantasai: A lot of changes to this draft, some major, others minor
   fantasai: I updated the changes section last night
   fantasai: CVS logs available (with comments) for anyone who wants more
             detail http://dev.w3.org/cvsweb/csswg/css3-text/Overview.src.html

   howcome: I sent in comments on this one, I don't think we should change
            the names of the hyphenation properties, particularly the main
   howcome: We reached consensus on the name in 2007
   ... ok with changing lesser-used properties, mainly concerned about hyphens ...
   <Bert> (I have the feeling there are too many line/word breaking properties,
          with too many interdependencies, but I haven't had time yet to work
          through all combinations...)
   <fantasai> (It's a mess, and I agree, but there isn't much I can do about it.)
   howcome: I don't think the new names are better or worse, think we should
            hear back from implementers
   fantasai: Murakami-san liked the new hyphenation-limit names. Haven't
             heard back from Michael Day.

   plinss: Where are we?
   fantasai: I think I need to add hyphens property back in and nothing else
   howcome: And clarify word-break behavior for non-CJK scripts
   smfr: we implement some of the hyphenation properties
   * scribe didn't catch everything smfr said
   smfr: hyphenate-character we support auto or a string
   <smfr> webkit supports hyphens, hyphenate-character, hyphenate-locale
          (with prefixes)
   howcome: I don't see any use cases for the second character
   howcome: I didn't find anything
   fantasai: I think there are two advantages to publishing with two strings
             right now and removing it in the next draft. One it's marked as
             an issue, we might get feedback. Other is if an implementor wants
             this functionality, they have some idea of how to do it.

   SteveZ: I have a colleague who believes that text-trim and/or hanging-punct
           doesn't match the Japanese requirements
   Koji: I talked with Nat about this, and I think I explained that the way
         we do it is different from what InDesign does, but we have some
         current level of consensus.
   Koji: CSS3 Text is only subset of what InDesign does
   SteveZ: My concern is not the subsetting, but that the way it was done
           does not match Japanese requirements.
   SteveZ: Feedback was from someone other than Nat, btw.
   SteveZ: Don't object to publishing WD, just wanted to note this issue.
   Koji: I met with Yamamoto-san and Nat on Monday
   SteveZ: My understanding was that values would pick things out of a table.
   Koji: My understanding was the request was a little too much for this level.
   Koji: I will talk about that with Adobe at our next meeting
   plinss: Do we have consensus to publish with suggested changes?
   plinss: Or do we need to come back and review?
   howcome: I would like to have consensus from implementors first.

CSS3 Speech

   plinss: Daniel Weck is preparing to request LCWD for CSS3 Speech, so
           please review

Meeting closed.
Received on Thursday, 27 January 2011 21:50:42 UTC

This archive was generated by hypermail 2.4.0 : Friday, 25 March 2022 10:07:55 UTC