W3C home > Mailing lists > Public > www-style@w3.org > February 2013

[CSSWG] Minutes Tucson F2F Monday Afternoon I: Text Decoration, Multi-col

From: fantasai <fantasai.lists@inkedblade.net>
Date: Thu, 14 Feb 2013 14:15:04 -0800
Message-ID: <511D61E8.4070303@inkedblade.net>
To: "www-style@w3.org" <www-style@w3.org>
Text Decoration
---------------

   - RESOLVED: don't drop text-underline-position
   - RESOLVED: not (currently at least) doing the @pattern proposal
   - Other issues awaiting i18n input. Tracked at
       http://dev.w3.org/csswg/css-text-decor-3/issues-lc-2012

Multi-Col
---------

   - Blocked on testcases and a couple spec issues.
   - RESOLVED: column rules are painted in the inline content layer
               (as described in http://lists.w3.org/Archives/Public/www-style/2012Jul/0546.html),
               but below all inline content inside the multicol

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

Text Decoration
---------------
Scribe: dbaron

   fantasai: I wanted to review status of various issues.
   <fantasai> http://dev.w3.org/csswg/css-text-decor-3/issues-lc-2012
   fantasai: we have 5 issues, the first is the trickiest
   fantasai: ... and needs the WG's input

   <fantasai> http://lists.w3.org/Archives/Public/www-style/2013Jan/0282.html
   fantasai: Sebastian Zardner requested we remove text-underline-position
             and add an @-rule to generically create random text decorations
             using a variety of descriptors
   <fantasai> http://lists.w3.org/Archives/Public/www-style/2012Sep/0462.html
   fantasai: this message includes some examples of what he was thinking
   fantasai: Koji and I propose to request this request
   glazou: for the time being or in general?
   fantasai: in general; we think the text decoration features should be
             restricted to lines
   fantasai: and text-underline-position feature is needed for international
             text
   fantasai: if we want to do something like this it should be a different
             feature
   fantasai: text-underline-position is for handling something very specific;
             it should stay that way

   glazou: What about his complaint that it applies only to underline?
   fantasai: It can in some cases affect the overline.
   fantasai: It can distinguish alphabetic vs. below; can also distinguish
             between underline on left vs right in vertical text (which
             swaps the overline to the other side).

   glenn: This proposal essentially asks for a way to group style properties
          and refererentially refer to them by name in other areas
   fantasai: It's just about decorating text; not a generic feature.
   glenn: Does it generalize to a way to group style properties?
   fantasai: no
   [discussion between fantasai and glenn about whether this is macro-ing]

   fantasai: I don't think text-underline-position either prevents us from
             going in this direction or should be dropped.
   glenn: If this is something new, I don't think we should spec it out
          unless somebody's committed to implementing.  Also seems like
          @-rules have a lot more overhead than properties.

   SimonSapin: Can we see this as a way for images above the text and
               generate procedural images images?
   fantasai: There was some discussion on the list... proposal to duplicate
             background properties.
   fantasai: text-decoration is about text and associated with where text
             is drawn
   fantasai: foreground images would be referencing the box
   smfr: what are the use cases for foreground images?
   fantasai: a "new" banner across image, sparkles
   TabAtkins: "ACTUAL SIZE"
   fantasai: It's easy to spec.
   smfr: painting order is tricky
   fantasai: The proposal is to reject this comment and not drop
             text-underline-position.
   dbaron: I have questions about some of the values of text-underline-position,
           but that's a separate thing.
   dbaron: I have comments about some of the values of text-underline-position,
           but that's separate.
   <SimonSapin> maybe that’s just SVG
   smfr: did we resolve we don't want to do the @-rule right now?
   TabAtkins: I think @pattern should be something SVG invents and we
              *possibly* import; we shouldn't try to innovate with that
              in CSS first.
   RESOLVED: don't drop text-underline-position
   RESOLVED: not (currently at least) doing the @pattern proposal

   fantasai: for reference there are 4 other issues on text-decoration
   fantasai: LC period closed last week
   fantasai: several of those are about combination of emphasis marks
             and ruby
   fantasai: 2 are about text-decoration-skip
   fantasai: We'll bring those to i18n for the correct answers, then
             we'll bring back to this WG for comments.
   fantasai: If anybody else has issues, please raise them this week.
   SteveZ: Implementation status?
   fantasai: Mozilla has quite a few; AntennaHouse probably nearly everything
   Koji: WebKit has a good amount in progress
   fantasai: That's it on text-decoration.

Multi-col
---------

   Bert: I'm trying to find the current status. Do we have any ideas when
         it might go to PR?
   Bert: Are there ways to speed it up?
   fantasai: The testing situation  -- not enough tests to cover the spec.
   fantasai: Also a handful of open issues, most notably one SimonSapin
             raised that we didn't resolve at last f2f.
   fantasai: We need to get all the issues handled and publish a new CR,
             and get more tests.
   fantasai: So I think there's a lot of work left to do.

   Bert: And implementations?
   fantasai: Mozilla's working on areas where we're not compliant; haven't
             unprefixed yet.
   fantasai: Not a super-high priority; hopefully able to be unprefixed
             by end of year.

   Bert: Does Mozilla have break-*?
   fantasai: no
   fantasai: If that's what's holding things up, we'll have fragmentation
             in CR, and can drop from multicol.
   [something there about WebKit too]
   Bert: Prince doesn't do column-span: all; I think others do.
   Rossen: We do, I think.
   fantasai: Can IE submit tests?
   Rossen: I can ask Arron to look into it.
   Rossen: Last time I remember talking to Håkon, he said they had tests
           ready to submit.
   fantasai: Håkon submitted tests, but pretty much useless.
   Peter: shepherd reports 129 tests for multicol
   Bert: are you seeing the gaps?
   fantasai: A lot of things
   Peter: have tests across most of chapters, but would need to drill down
   Bert: so I hear not this year... that's a pity
   fantasai: Primarily gated on test, and a couple of spec issues.

   Rossen: One issue about multicol I wanted to discuss.
   <Rossen> http://lists.w3.org/Archives/Public/www-style/2013Jan/0251.html
   [discussion of IRC bouncer]
   Rossen: So there was an issue with column-rule rendering drawing order;
           didn't get much traction on the list (link above).
   Rossen: So the current spec defines column rules to be rendered just
           above the background.
   [reads spec text]
   Rossen: I think it's a fairly reasonable behavior expectation that when
           you have scrollbars, the column rules should scroll along with
           the columns.
   Rossen: The testcase in that link is an example of taking a fairly
           simple case... and I looked through all implementations
           supporting multicol at the moment... and most implementations
           don't actually scroll the column rules because they treat them
           as part of the background.
   Rossen: If you also combine that with a case of z-index elements,
           some of the implementations appear to start working, but
           then don't... it's fairly messy.
   Rossen: I think the current specification is fairly weak in its requirements.
   Rossen: For us, to implement something that would achieve scrolling
           with the content as well as being at the level of background
           (under z-index), that means we need a new layer
   Rossen: It's going to be a hassle to make that work, for I'm not sure
           how much benefit.
   Rossen: column rules are more of a design-time requirement (visual aid)

   fantasai: I agree column rules should scroll with the columns; should
             stay between the columns.
   fantasai: With regards to what level; I think it should be below the
             content (with no z-index); interesting question as to whether
             above or below borders; above borders might make more sense.
   [smfr shows Rossen a JSFiddle]
   fantasai: If that elements forms a stacking context, at which level
             does it paint?
   fantasai: I'd say immediately before the text and text decorations
             (anonymous text), or immediately above the background.
             As long as it's below the text.
   fantasai: I don't care about above or below negative z-index stuff.
   fantasai: So below anything with a z-index of 0 or auto
   smfr: I'd like to avoid a separate layer just for column rules.
   Rossen: So just lift them up to the content layer... first in the
           content layer.
   smfr: Paint all the rules, then the content of the columns

   dbaron: Does multi-col form a pseudo-stacking context?
   fantasai: Don't know that we thought that far..
   dbaron: If multi-col doesn't establish a pseudo-stacking context, then
           floats from outside the multi-col propagate through the multi-col
   <smfr> http://lists.w3.org/Archives/Public/www-style/2012Jul/0546.html
   fantasai: guessing we want before Group A
   dbaron: If a multi-col doesn't even establish an A-Group (pseudo-stacking
           context)
   dbaron: then only way to put it before group A is to give it its own layer
   dbaron: but probably it should establish a pseudo-stacking context
   [investigation into what forms a pseudo-stacking context]
   dbaron: I argued at one point that anything establishing a BFC should
            create a pseudo-stacking context, but I think I lost.
   dbaron: Was that about tables or about flexbox?
   <fantasai> tables don't form stacking contexts, for historical reasons
   <fantasai> flexboxen don't either, see
              http://dev.w3.org/csswg/css3-flexbox/issues-lc-2012#issue-5
   <fantasai> so multi-col shouldn't either
   RESOLVED: column rules are painted in the inline content layer
             (as described in http://lists.w3.org/Archives/Public/www-style/2012Jul/0546.html),
             but below all inline content inside the multicol
   <fantasai> so we're painting right below inline layer, in the middle
              of Group A
   ACTION Rossen to ask Håkon to edit this resolution for column rule
                 painting order
   Created ACTION-530

   Peter: Should we consider adding Rossen as a co-editor of css3-multicol?
   Peter: Anything we can do moving forward testing-wise?
   fantasai: Gerard working on some of mozilla's tests
   fantasai: for multicol specifically, and backgrounds and borders
   Peter: Anything else on multicol?

   fantasai: SimonSapin's issue
   Bert: I want to talk to Simon before tonight.
Received on Thursday, 14 February 2013 22:15:37 GMT

This archive was generated by hypermail 2.3.1 : Tuesday, 26 March 2013 17:21:05 GMT