[CSSWG] Minutes Seattle F2F Mon 2014-01-27 PM II: CSS3 Backgrounds, CSS2.2, Counter Styles, -webkit-print-color-adjust

CSS3 Backgrounds and Borders
----------------------------

   - RESOLVED: Publish css3-background as LC with missing edit, 3-week LC period

CSS2.2
------

   - Waiting on either tests or Process change before publication.

Counter Styles
--------------

   - RESOLVED: Drop definition of extended CJK counter styles
               (i.e. numbering beyond 10K) because ppl can't agree
               on when to drop ones and zeroes.
   - RESOLVED: Counter Styles to CR, once Tab finish off the DoC

-webkit-color-adjust
--------------------

   - RESOLVED: Add this property/functionality to Color 4 with names TBD

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


Backgrounds and Borders
-----------------------

   fantasai: Anyone have issues on css3-background?
   krit: putting values between <box> values in background shorthand?
   fantasai: should be fixed
   fantasai: nope, not fixed, I'll fix it...
   RESOLVED: Publish css3-background as LC with above edit, 3-week LC period

CSS2.2
------

   Bert: Discussing updating CSS2.1
   Bert: The main thing keeping us from that was some tests
   Bert: I wrote some tests
   Bert: Found that things were not implemented
   Bert: specifically, the scinot parsing is not implemented
   Bert: I haven't checked other errata
   Bert: Wondering if they also have no implementations
   SimonSapin: I have 2 implementations, but they are not independent
   SimonSapin: WeasyPrint and Servo
   dbaron: Should take about 5m to implement in Gecko
   dbaron: need to remove an SVG check
   ChrisL: SciNot was added to SVG everywhere, asked CSS to add it, and
           they said no. And then was recently added
   <ChrisL> so the implementations need to remove their special checks
            that disallow scinot in properties

   discussion of going to LC
   ChrisL: There's a bunch of tests someone posted as well
   [discussion of where to find tests, etc.]
   <Bert> -> http://wiki.csswg.org/spec/css2.2 volunteers for CSS2 errata
          tests

   plinss: Is the document ready to publish as LC?
   Bert: yes, just have to change the status
   fantasai: Did you kill the entire changes section, other than the most
             recent set?
   ACTION Bert Cut down changes section to just the changes from CSS2.1REC2011
   <trackbot> Created ACTION-606

   ChrisL: What about testing undetectable changes?
   fantasai: Just say they're undetectable in the impl report?
   ChrisL: Also, how do we find existing tests that might need changes?
   ChrisL: Will need a new edition of the test suite
   plinss: I can build one

   plinss: So, publish LC?
   fantasai: Concerned some people will freak out, especially if we stay
             in LC for 6months+.
   fantasai: Maybe put everything together and then go to LC very shortly
             once it's all set to move forward
   dbaron: What if we put it at shortname CSS22, then nobody will
           look at it
   Tantek proposes waiting for the Process change
   florian: Once we have the tests, then we can rush.
   ChrisL: Once we have the tests we know where we are
   tantek: excellent, we've rationalized procrastination

   <dbaron> filed https://bugzilla.mozilla.org/show_bug.cgi?id=964529 (with patch)
            for supporting scientific notation in Gecko

Counter Styles
--------------

   TabAtkins: Only one issue open
   TabAtkins: Nobody can give me consistent answers on the extended cjk
              numbering styles
   TabAtkins: Already have one that goes to 10^5. One that goes to 10^16,
              optional
   TabAtkins: I think it's sufficient, what we have already
   TabAtkins: Can address in a future level if someone can give me a real
              answer
   TabAtkins: Are people ok with that?
   Kawabata-san nods
   fantasai: I'm ok, if it's undefined above 10k. Mozilla implements
             beyond that, don't want to make them incompliant
   dbaron: Were there disagreements on informal or formal?
   plh: Maybe ask i18n?
   TabAtkins: Haven't, no. Asked some native speakers
   fantasai: Say beyond 10K could either switch to cjk-decimal, or extend
             beyond (you figure out how)
   TabAtkins: The main issues are around when you drop ones and zeroes
   TabAtkins: Everyone agrees on the first 4 digits
   TabAtkins: beyond 10K becomes a problem
   plinss proposes scientific notation
   fantasai proposes engineering notation
   RESOLVED: Drop extended algorithm for CJK (10K+)

   plh: should i18n review this?
   fantasai: it's optional anyway
   RESOLVED: Counter Styles to CR, once Tab finish off the DoC

-webkit-print-color-adjust
--------------------------

   florian: We discussed this awhile ago
   <florian> http://wiki.csswg.org/ideas/print-backgrounds
   florian: we concluded 2-6 doesn't work
   florian: 7 is what webkit does
   florian: 8 I don't remember
   florian: Suggest we go with what webkit does, with a non-silly name

   florian: By default browsers will adjust colors to have white background,
            to save ink
   florian: This property can switch that behavior off, print what's specified
   florian: gives author ability to say when things are important
   TabAtkins, florian: Doesn't affect user's ability to control things;
                       handle that via user stylesheet mechanism

   florian: We need a name, a spec, and a description
   fantasai: Colors 4
   florian: So, inherits, applies anywhere, and mode where browser does
            whatever it wants
   ChrisL: i.e. initial value is auto
   florian: suggest 'economy', to be a little more explicit than auto...
   ChrisL: Need on, off, and auto
   TabAtkins: Default behavior is whatever user wants, which happens to
              be save-ink by default
   ChrisL: Right so, that's auto
   fantasai: No, just two. 'economy' is initial value, but user can set
             to 'true-colors' in user style sheet
   discussion of how the user-stylesheet cascade works

   Bert: Aren't you now giving the user 3 choices rather than 2?
   TabAtkins: ...
   Bert: User has options Save, Don't Save, Do what author said
   Bert: Property only has 2 values, but user has 3 options
   TabAtkins: Blink doesn't give user any options. Depends on what the
              browser wants to expose
   florian: We don't mandate what the UA puts in its prefs.
   florian: we just give them more info to work with, if they want to
   Bert: Wonder what the user thinks, seeing these options
   TabAtkins: That's not our job, that's the UX designers' job.
   [discussion of what customization options are appropriate for a browser to expose]
   * fantasai is not minuting this
   Bert: How do I say, I really want to save ink?
   ...
   dbaron: I don't think we'd expose a 3-way toggle
   dbaron: The reason we have this magic initial behavior is because
           authors generally don't think about printing
   dbaron: This property is only used if an author *is* thinking about printing.
   dbaron: If an author uses it, we trust them to have thought about printing
   plinss: We thought of other heuristics to see whether author thought
           about printing. They weren't good enough heuristics. But this
           could be.

   plinss: Does anyone object to adding this property?
   Bert: I disagree with this. Why aren't print style sheet enough?
   Tab rants about "future legacy"
   tantek, TabAtkins: Authors right now use borders as backgrounds,
                      to force the printing of "backgrounds"
   Basically lots of hacks. This would allow a clean way to solve the problem
   RESOLVED: Add this property with names TBD

   <dbaron> print-economy: auto | as-specified ?
   <ChrisL> economy | premium
   <dbaron> economy | premium | super94 ??

Received on Thursday, 30 January 2014 13:42:07 UTC