Join GitHub today
GitHub is home to over 40 million developers working together to host and review code, manage projects, and build software together.Sign up
[css-writing-modes-3] Dropped definition of automatic block sizing #4220
The CSS Working Group just discussed
The full IRC log of that discussion<dael> Topic: Writing Modes update
<dael> Rossen_: fantasai I think you raised an issue about part of a chapter being dropped and it warrants a republication?
<dael> Rossen_: Is fantasai on?
<florian> github: https://github.com//issues/4220
<dael> fantasai: Section about how to size orthogonal flow was dropped. There is a section on available space, but not sizing. Need to put that section back.
<dael> fantasai: It was dropped with text about multi-col.
<dael> Rossen_: Editorial mistake during the update? Or did we resolve?
<dael> fantasai: Resolution was to drop everything at risk which does not include sizing block elements in orthogonal flow. Without that you interpret rules quite differently
<dael> Rossen_: Don't disagree this is core functionality on how we size. I'm surprised we had it out of the spec so far and just discovering it's out.
<dael> fantasai: Also surprised, but I don't know.
<dael> dbaron: Also feels like we should have a chance to review it again since there has been lots of spec review with it out.
<dael> fantasai: Sure. But what is dropped was related to what's implemented. Without that text we do not match impl
<dael> florian: It's in L4 right?
<dael> fantasai: Text is in L4 but it's...its embedded in multicol text. It says this is what you do if you support multicol and this is what you do if you don't. Then we made multicol a may and the alternative behavior ended up dropped as well
<dael> fantasai: If you want I can do that
<dael> AmeliaBR: I think question was is this missing from both L3 and L4?
<dael> fantasai: L4 includes all the text
<dael> florian: dbaron said we need to review again and we have as part of L4? Given it has been part of a CR spec we don't need to heavily review
<dael> dbaron: L3 and L4 have had different review levels
<dbaron> s/review levels/amounts of review/
<dael> fantasai: If anyone reviewed the spec and paid attention to this part they would have noticed it does not match impl
<tantek> that sounds bad (that nobody noticed)
<dael> dbaron: I think add it back, but I'd like to look at it.
<dael> fantasai: Of course
<dael> Rossen_: Review before merge into L3?
<dael> dbaron: No.
<dael> dbaron: I'd like to look before it's a new CR
<dael> Rossen_: For sure
<dael> Rossen_: Let's add the missing section back.
<dael> Rossen_: We can come back next week. Are we close to republish CR?
<dael> florian: We just did. This is only issue since. Should be close to republish
<dael> Rossen_: What would this section mean from testing PoV?
<dael> florian: THat's the next topic; I have just compled impl report
<dael> florian: I have not checked if the tests aligned with presence or absense of that section. I only reviewed failing tests for correctness.
<dael> florian: We have ~1200 tests and ~25 are mandatory and lack 2 impl
<dael> florian: Some is related to recent fix about prop from body to root at used time. A few are sizing in orthogonal flow and I think there's work planned to fix.
<dael> florian: One large-ish problem is orthogonal table cells have problems. Edge reasonably well, FF with some problems, WK and Blink don't do it.
<dael> florian: There are small corner cases we might decide to oerlook. Not quite PR but may be close
<dael> Rossen_: Thank you for putting this together
<dael> Rossen_: I'm expecting new section might require additional tests
<dael> florian: Have tests for that area, if need verified need to decide
<dael> Rossen_: With that we can see if we need to address the 18 with mandatory pass of 2+
<dael> florian: I encourage impl to look at the report. 1st section with detailed results is the list of thigns that need to fix before move forward. If you can look and see if there's anything easy to fix it would be appriciated
<dael> florian: It's not that many
<dael> Rossen_: That's a reasonable ask
<dael> Rossen_: The things that would have put a pause for me are table related. From impl PoV tables were particularly rusty when we had to impl writing modes. Should be something we can hope to impl in Blink once layout has enough wind behind it for purposes of common-based browsers. Hopefully FF and WK can follow
<dbaron> Firefox table-cell issues are related to https://bugzilla.mozilla.org/show_bug.cgi?id=1231656 and https://bugzilla.mozilla.org/show_bug.cgi?id=1244601
<dael> florian: FF is closer. WK does not support orthogonal table flows. FF only has problems with sizing, rest works.
<dael> Rossen_: Case where orthogonal cell sizing is titchy. Missing section will help
<dael> Rossen_: In summary; we have a number of cases impl should go back and look to see what can be done to address in short term. If they're bug level I hope people can address soon. If major refactor we need another convo about if we insist on those cases to pass or if we push the spec with the fails
<dael> Rossen_: We'll give it a week for dbaron and anyone else to look at the dropped section we're going to add and then we'll have a resolution to add it next call.
<dael> florian: Didn't wesay add now and discuss republish next week?
<dael> Rossen_: We don't gain anything by doing that. I'd prefer anyone has opportunity for feedback before we do editorial
<dael> florian: Easier to read if it's edited in
<chris> yes, easier to read in-place
<dael> Rossen_: Missing section is clear in the PR fantasai pasted. It's self-contained
<dael> florian: What fantasai pasted is a deletion that deletes too much. Only parts need to be added back. I'd rather an editor decide what parts.
<dbaron> I agree with Florian
<dael> Rossen_: Fair. If we can prepare PR for added section that would be great.
<dael> Rossen_: fantasai will you handle this?
<dael> Rossen_: I'm hoping you're saying yes to a muted mic
<dael> fantasai: Yes I'll do it today
<dael> Rossen_: Thanks florian for taking time to piece together impl report. It's super helpful. Hopefully make progress before TPAC.
<dael> Rossen_: One more nudge to impl to look at failing test cases