- From: fantasai <fantasai.lists@inkedblade.net>
- Date: Wed, 09 Feb 2011 11:48:28 -0800
- To: "www-style@w3.org" <www-style@w3.org>
Summary:
- Reviewed blocking tests in CSS2.1 test suite. Needed fixes listed on
blocking page:
http://wiki.csswg.org/test/css2.1/blocking
Opera to investigate fixes for quotes-035, replaced-intrinsic-001, orphans-004a
Still need implementations for allowed-page-breaks-001a, forced-page-breaks-000, uri-016
- Reviewed state of CSS2.1 issues list.
- CSS3 Backgrounds and Borders CR still blocked by Director's approval step due to
lack of Director's representative's attention to scheduling the required telecon.
- Brief discussion about CSS Variables
- Brief discussion about multicol algorithm to determine column count, to
be continued next week.
- <strong>Resolved:</strong> Publish CSS3 Text as WD.
====== Full minutes below ======
Present:
Acebal
Tab Atkins
David Baron
Bert Bos
Arron Eicholz
Elika Etemad
Simon Fraser
Sylvain Galineau
Koji Ishii
John Jansen
Brad Kemper
Håkon Wium Lie
Chris Lilley
David Singer
Geoffrey Sneddon
Steve Zilles
<RRSAgent> logging to http://www.w3.org/2011/02/09-css-irc
Scribe: Tab Atkins
Agenda
------
glazou: Three item are Variables, multicol algo, and CSS3 text.
fantasai: Can we do a quick check on B&B?
glazou: Yes, after 2.1 issues.
CSS2.1
------
glazou: There are a few urls in the agenda with css 2.1 blockers
glazou: Let's review the firs tone, with all the redness inside.
<fantasai> http://test.csswg.org/harness/results?s=CSS21_HTML&t=0&f[]=1&f[]=1
<fantasai> http://wiki.csswg.org/test/css2.1/blocking
<fantasai> http://wiki.csswg.org/test/css2.1/need-data
<fantasai> http://wiki.csswg.org/test/css2.1/invalid
glazou: I think these issues are mostly just awaiting fixes from implementors.
TabAtkins: These bugs are in Webkit's bug report, but I dunno what's been
worked on.
smfr: I haven't been able to work on several. I know that [some involving
intrinsic] are outstanding, but I'm not aware of issues for the others.
glazou: A lot of these are related to print, and we have only one impl
passing for a lot of them.
howcome: If print is blocking here, did ChrisL run any of them through Prince?
ChrisL: I've done a few of them, yes.
fantasai: One problem is that you have to tweak some of the test for Prince,
since it has problems with height:100% on <html>.
<ChrisL> what is the tweak?
<fantasai> For some tests that test pagination
<fantasai> You need to remove height: 100% from the root element
<fantasai> and change the divs' 50% to 11.25cm
fantasai: The tests use 100%/50% so they would work for all page sizes.
fantasai: But Prince has trouble with height:100%
fantasai: The 11.25cm does the same thing, assuming Prince's default page
size.
<ChrisL> ok so those tests are not explicitly testing height 100%
fantasai: Like allowed-page-break-001a.
johnjan: I don't think we need to go through all of these, right? We
already went through these at the mini-f2f.
johnjan: We've already covered all of these, so they're no longer
significantly blocking, right?
glazou: My question as a chair is *when* these will be addressed?
fantasai: If you look at the blocking link, you see that most of the
tests were fixed, and for the rest we're expecting impls.
fantasai: [talk about specific tests and their status in impls]
fantasai: We do need a couple fixes from Opera, though
howcome: Can you list the tests that Opera needs to pass?
<fantasai> quotes-035
<fantasai> replaced-intrinsic-ratio-001
<fantasai> http://wiki.csswg.org/test/css2.1/blocking
<ChrisL> http://test.csswg.org/harness/testcase?s=CSS21_HTML&c=quotes-035
<dbaron> for details on quotes-035 and Opera:
http://lists.w3.org/Archives/Public/public-css-testsuite/2011Jan/0095.html
johnjan: Look at the blocking list - it shows which ones Opera can fix.
fantasai: The open questions are here:
<fantasai> http://wiki.csswg.org/test/css2.1/blocking#tests-needing-breakup-or-implementation
fantasai: The three at the bottom are ones we don't have a plan for.
fantasai: They're passed by Prince, and seem to be correct, but don't
have expected implementations anywhere else.
howcome: What about Antenna House?
johnjan: No, they didn't pass.
glazou: Can we expect implementations, or are they at-risk?
fantasai: In orphans-004a, it's a parsing test. Should be straightforward
to fix.
fantasai: If Opera or IE9 fixes the parsing bug it would pass.
fantasai: I think Opera accepts numbers rather than integers, and IE9
accepts 0.
glazou: So, Moz/Opera/IE, is there something you can do here?
howcome: I'm taking notes here. It's hard to promise, but I'll do my best.
johnjan: I'm taking a look at it right now.
dbaron: We'd have to implement orphans in the first place, so it's not
just a parsing bug for us.
fantasai: When I get a break from writing specs, I'll implement it. :)
glazou: That's the end of the list. Is there anything else about CSS 2.1?
johnjan: There's the open issues list.
<johnjan> http://wiki.csswg.org/spec/css2.1
johnjan: There's like 4 that bert needs to edit, 4 that need proposals,
and 4 that need testcases. 40 or so that have been edited and
updated, but just need to be reviewed.
johnjan: i sent an email hoping that they could get done before the telcon,
but stuff was too busy.
johnjan: So can we get promises to get these done this week?
glazou: Bert, can you do your edits this week?
Bert: Haven't looked at them recently, but I assume so, yes.
glazou: Reviews should be done offline.
fantasai: I can do those this week.
<ChrisL> what is the list of tests needing reviews?
<fantasai> ACTION: Bert to finish outstanding 2.1 edits
<trackbot> Created ACTION-293
johnjan: First that needs proposal is 181.
<fantasai> ACTION: fantasai to review edits
<trackbot> Created ACTION-294
http://wiki.csswg.org/spec/css2.1#issue-181
glazou: Elika, can you summarize?
* fantasai looks
<fantasai> http://lists.w3.org/Archives/Public/www-style/2010Aug/0010.html
<fantasai> Seems there's a bunch of editorial suggestions
<fantasai> the significant ones start at "Alternative construction for line
box height calculation"
fantasai: I think the prose that adjusts the guide boxes (?) makes sense,
and is probably less confusing to our current text.
fantasai: My suggestion is that Bert looks this section over - it's in his
focus area - and go through that email and make whatever edits
seem good, then post a diff for review.
Bert: This won't happen this week, but maybe in time for next telcon.
glazou: Anything about the testcase at the end of the email?
fantasai: That might be an interesting example.
<johnjan> http://wiki.csswg.org/spec/css2.1#issue-192
TabAtkins: 192 is mine. Sorry, I'll get it done this week and post to
the list.
<johnjan> http://wiki.csswg.org/spec/css2.1#issue-199
TabAtkins: I haven't gotten feedback on my 199 proposal yet. I'm not
happy with it conceptually (positioned elements should generate
an inline placeholder, which *should* make a linebox), but it
matches implementations.
TabAtkins asks about bzbarsky reviewing.
dbaron: Email me or him to get us to look at the proposal again.
<johnjan> http://wiki.csswg.org/spec/css2.1#issue-c
fantasai: I guess I can write that text.
ACTION fantasai to write text for issue C.
<trackbot> Created ACTION-295
fantasai: I need to update the issues list, so we may get a few more.
johnjan: Will you have time to go through the mailing list and find all
the open ones?
fantasai: Dunno. The date of my last review is at the top of that page.
If anyone else is doing work, please update that date.
dbaron: It looks like there's a continuation on a separate issues list.
fantasai: Yeah, I'm thinking I should merge the two.
CSS3 Backgrounds and Borders
----------------------------
glazou: Next item. B&B.
fantasai: So, what's the status?
Bert: I failed to get Ralph's attention, so we didn't get a telcon.
Bert: I'm hoping to get another one this Friday so we can try again.
glazou: Peter won't be available this friday. I *hope* to be so, but
won't know for sure until noon thursday.
CSS Variables
-------------
glazou: Next item on the agenda, CSS Variables.
fantasai: I'd like to request CSS3 Text because I'd like to get a draft
out there, and I'd like to get a chance to discuss that.
glazou: I think there is at least one urgent point about variables, first.
glazou: Variables are extremely visible to the public. Webdevs want it,
they request it loudly.
glazou: We have to be extra careful with our communication here.
glazou: A lot of people understood from your email, Tab, that "it's going
to be implemented that way"
glazou: And that's not good from WG perspective
glazou: You need to be clearer in your communication
Tab: That wasn't what I meant. It's just an experimental implementation.
glazou: I suggest you post a draft for discussion.
glazou: Second point, Tab's proposal introduces $ for delimiting variables
<dbaron> (Is $ easily available on keyboards around the world?)
glazou: It changes a lot the way we introduce new idents inside CSS. So
please review and comment.
glazou: Taking comments on variables for next 5 minutes.
Stevez: What I found confusing was, what is the token stream?
Tab: The exact answer for what's a valid value of a variable is is a little
in the air. Need a lot of work to make sure that's very clear.
Tab: High-level answer is, something that's a value in CSS.
fantasai: I think you want some number of component values up to a full
property value
CSS3 Text
---------
glazou: Next topic. CSS3 text.
fantasai: What's blocking an updated WD?
howcome notes that his phone line is staticky.
<fantasai> http://dev.w3.org/csswg/css3-text/
howcome: I expressed in an email that I'd like the GCPM functionality to
move over without changes, and then have changes down afterwards.
<fantasai> http://dev.w3.org/csswg/css3-text/#hyphenation-resource
howcome: In particular, the @hyphenate-resource was added, when it was a
'hyphenate-resource' property in GCPM.
<howcome: both are in the draft right now. I would like one or the other.
howcome: We spent a lot of time on the property as a WG, and we had consensus.
howcome: In the moving process, I don't think we should say one or the other.
We should say "this is what we have", and then add issues.
howcome: But I think putting them on equal footing deteriorates the quality
of the draft.
howcome: I would like to try and preserve what has been done, and
hyphenate-resource has been discussed at length in the WG, been
implemented, and is in use.
howcome: I think the changes should come as a result of discussions, not
as a result of moving.
fantasai: The @-rule came as a result of discussion at the f2f, so I consider
it my duty as an editor to include the proposal.
howcome: I think it can be in the draft but not on equal footing, as I don't
think we made that decision.
glazou: I don't understand you, howcome. It's a WD, not a CR. Everything
can change.
howcome: In the ED that's fine, but in the WD we should try and preserve
consensus.
howcome: The consensus may change in the end, that's fine. But by giving
two proposals equal footing, it makes it less obvious what's going
on.
szilles: Can you suggest wording that makes it not equal footing that fantasai
can use?
howcome: I can - I think we can call it an issue. That's what we've done
before.
glazou: So you're suggetsing a note in the doc saying "Another proposal for
this is an @-rule, etc."?
fantaai: An issue.
howcome: Yes.
glazou: That's fine by me.
howcome: You may want to add another issue, regarding the format of the
hyphenate resource files.
smfr: Already an issue there.
* smfr mumbles something about video codecs
[chatter about the format of hyphenation resources]
* Bert wonders if there are as many patents on hyphenation dicts as on
video codecs :-)
<glazou> or fonts
howcome: I think there's consensus to add an issue about the hyphenation
resource format.
<dsinger> having lots of formats, and making sites somehow supply the
right format for the browser it has in hand, would be ... ugly...
* dbaron has to leave a few minutes early today
Multi-Col
---------
glazou: I don't know if we have time for multicol algos this week. I'll
put it in second position, after css2.1 issues.
howcome: Can we get Alex to call in?
johnjan: I can get him to call in.
* Scribe missed what Sylvain's comment was.
sylvaing: In that branch, the more columns you squeeze in, the less content
you get.
sylvaing: I see the goal is to make more columns and thus more content,
but in practice it reduces the amount of content, due to
pushing in a new column gap.
sylvaing: So I suggest looking at the ratio of whitespace to content
after each change, and see what the right answer is.
howcome: I know there's some weird stuff, but the aim is to make things
more stable. Previous we went from 3->2->1 columns over the
space of two pixels.
* fantasai checked in and would like to know if howcome is satisfied with the
current editors draft and if so if we can publish css3-text
sylvaing: For me the real scenario is people snapping their window to half
of the screen in W7 and suddenly the content area is shrunk in half.
sylvaing: In those scenarios it's better for the user if we can make less
columns and thus less column gap.
sylvaing: My feel is that you were trying to fit in more columns in the
assumption that you get more content, which isn't necessarily true.
howcome: Alex had a suggestion that we go straight to 1 col when we hit that.
Another is that columns get stuck there, and designers learn to not
make wide columns.
howcome: I don't think it's necessarily a huge issue.
glazou: Table the rest of the discussion for next week.
szilles: +1 for what sylvain is saying.
howcome: I'd like to hear more proposals on the multicol thread.
<Bert> (Isn't the issue just that there is no minimum column width in the
algo?)
CSS3 Text Revisited
-------------------
fantasai: I made the changes to the draft. Does howcome find them satisfactory?
<fantasai> http://dev.w3.org/csswg/css3-text/#hyphenation-resource
howcome: That's acceptable to me.
glazou: Is the edit only the redness on the paragraph?
fantasai: That, and I put the explanation of the issue to the top of the
definition.
howcome: I'd like to see "Issue" in there.
fantasai: I can put "Issue: " there.
RESOLVED: Publish CSS3 Text as WD.
<fantasai> whew
* fantasai is glad to be able to publish WD!
Meeting closed.
glazou: Btw, got our rooms in Tokyo all reserved and ready.
Some extra IRC discussion on multicol:
<Bert> But the actual amount is hard to define, maybe just some rule of thumb.
<Bert> Maybe the minimum is 10% of the gap width. I have no typographic
reason for that, but it makes the formula easy. :-)
<TabAtkins> Bert: If the goal is to preserve content readability in degenerate
situations, I don't think a % of the gap width is useful.
Received on Wednesday, 9 February 2011 19:48:58 UTC