Re: [css-display] FYI: Gecko Intent to ship CSS display:flow-root

On 12/23/2016 11:43 PM, Mats Palmgren wrote:
> On 12/23/2016 07:00 PM, fantasai wrote:
>> That said, I'd love to see you all at least *ask* us to officially
>> stabilize specs you're planning to ship before announcing that
>> you're going to do it anyway.
>
> Well, Tab said it was "stable design-wise" and then implied that
> he wanted to see someone shipping it:
> https://discourse.wicg.io/t/containing-floated-children-with-clear-after-modern-clearfix/1835/6
> I took that as an OK to ship it; did I misunderstand?

His statement there isn't wrong, but it's also not clearance to ship.
Not entirely your fault for taking it the wrong way; I think Tab has
a tendency to forget that we have a process that requires more than
his opinion...

>> In this case, I think we can easily transition a trimmed down version
>> of Display in January. It would help to know which of its features
>> Mozilla has or is planning to implement, and if there are any known
>> spec issues with those features.
>
> Only display:flow-root this time.
> We have previously shipped display:contents and the ruby values
> (and the obvious CSS2 values).  We have *not* implemented 'run-in'
> yet though.  We only support the single-keyword form of 'display'.
>
> AFAICT, the 'display' keywords we haven't implemented yet are:
>   run-in, flow, inline-list-item
> Is the spec for these stable enough to implement and/or ship?

Most of Display has been basically blocked on someone other than me
and Tab to review it for sanity. :) See the email I sent you on 3 Aug
2016. Bzbarsky sent in some comments on the shakier features, which I
still need to fully process, but I'm also waiting on your analysis of
the spec for 'display: contents'!

> I think we (Gecko) plan to implement 'inline-list-item'[1] at some point.
>
> I'm not aware of any plans to implement 'display:flow', but it seems
> kind of trivial (in its single-keyword form) so I don't see any reason
> not to.

Tab and I think it's stable, the WG got a chance to review and approve it,
and nobody's sent any other comments, so I'd agree they're pretty stable.
That said, we haven't issued a transition request yet, which sometimes
surfaces issues... but in this case I'm really not expecting anything.

My current thoughts are that we should cut L3 of Display with just the
new keywords, and push the rest of it to L4, since we can do that pretty
quickly.

It's also possible we can sort out the remaining issues on 'run-in'
and/or the display-or-not property fast enough to get them to CR by
the end of January or so. But I'd be really depending on you/dbaron/
bzbarsky to help make sure the spec is ready, because nobody else I
know is really able to give these the requisite level of review.

> There's a (very old) bug filed on 'run-in' where someone said it was
> *removed* from Chrome/Safari[2] - was that because authors weren't
> very interested in this feature, or because of some other reason?
> Are they interested in adding it again?

It was because the CSS2 box model for it was insane, and therefore the
implementation was very buggy.

~fantasai

Received on Sunday, 25 December 2016 03:26:19 UTC