[css-display] Use "computes to" instead of "behaves as" for display: contents on unusual elements. #2755
Labels
Comments
|
cc @lilles since he implemented the Blink style fixup, but I think he wouldn't have any objection about changing this, given it would adapt the spec to how Blink, Firefox and WebKit implement. |
|
Huh. I thought we did store the original display for computed style before adjusting. I think we're fine with either resolution. |
|
I misunderstood our OriginalDisplay() thing. Not used for getComputedStyle(). I support resolving this as proposed. |
|
Sounds fine to me. @fantasai, any objection? |
|
WFM. |
|
The Working Group just discussed
The full IRC log of that discussion<dael> Topic: Use "computes to" instead of "behaves as" for display: contents on unusual elements<dael> github: https://github.com//issues/2755 <dael> fantasai: We had said that display contents behaves as display:none on certain elements like some SVG. There's an appendix of exactly which ones. emilio filed an issue to jsut say computes to display:none since that's easier from impl then handling that at used style time <dael> fantasai: TabAtkins and I have not much of an opinion. Up to impl. <dael> emilio: Do we clear everyone that impl computes to none? <fantasai> s/Do we/To be/ <astearns> s/Do we/To be/ <dael> Rossen_: Any reason to do anything other than that? Sounds like there's interop. <dael> Rossen_: Objections to serialize it out as none? <dael> RESOLVED: serialize it out as none <dael> RESOLVED: compute to none |
fantasai
added a commit
that referenced
this issue
Jul 25, 2018
…ed elements. #2755 per WG resolution
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
https://drafts.csswg.org/css-display/#unbox says:
Which means that the computed value of
displayshould still becontents. This adds unnecessary complexity, and this is not how all engines implementing this section right work.In particular, it requires a mapping from element to "is really display contents", or something of the sort. Firefox used to have that, but it caused a lot of performance issues for no gain, so I removed it in https://bugzilla.mozilla.org/show_bug.cgi?id=1303605.
(Note that I removed that mapping earlier without realizing the correctness issue it introduced, https://bugzilla.mozilla.org/show_bug.cgi?id=1453702, which I fixed computing to none instead of contents, which also aligns with Blink and WebKit).
The text was updated successfully, but these errors were encountered: