Re: Procedural (non-technical) point about freezing the cat and hat combinators before they've even been defined (was Re: Shadow DOM: Hat and Cat -- if that's your real name.)

On Tue, Feb 4, 2014 at 6:22 PM, Edward O'Connor <eoconnor@apple.com> wrote:
> +www-archive, replies just to www-archive (since this non-technical)
>
> Hi,
>
> Tab wrote:
>
>> Chrome *will* be shipping Shadow DOM publicly (in conjunction with
>> Moz) in the *very near* future. Whatever API gets shipped will be
>> frozen almost immediately. If you want to suggest name changes, as we
>> brainstormed a bit at the f2f, do so RIGHT NOW or forever hold your
>> peace.
>
> In the CSS WG we've historically allowed implementations to ship
> unprefixed properties when the spec containing those properties hits CR.
> Selector combinators are a funny case—they can't be prefixed—so we
> should be extra careful about shipping them prematurely.
>
> But as far as I can tell, these combinators *aren't even specced*, much
> less in a spec that's hit (or will soon hit) CR. This seems highly
> irregular.
>
> I assumed ^ and ^^ would be defined in Selectors 4. But they're not in
> its latest WD:
>
>                     http://www.w3.org/TR/selectors4/
>
> Nor in its latest ED:
>
>                   http://dev.w3.org/csswg/selectors4/
>
> Nor are they in the latest Shadow DOM WD:
>
>                     http://www.w3.org/TR/shadow-dom/
>
> Nor in its ED:
>
>          http://w3c.github.io/webcomponents/spec/shadow/#styles
>
> No, wait, they're in there. In Chapter 6 "Styles" we find this:
>
>> ISSUE 6
>> Hats, ^, and Cats, ^^, selector combinators should be defined in this
>> section.
>
> I'm left with the conclusion that these combinators are entirely
> undefined. I'm really surprised the Chrome team intends to ship these
> enabled by default in production.

Sorry, this is my fault.  These things *were* defined in the spec
before, but we sliced them out for a separate spec, which I was
supposed to write and haven't gotten finished yet.

That said, my last (voluminous) update from November
<http://lists.w3.org/Archives/Public/www-style/2013Nov/0313.html> is
still completely correct, except for the naming change/functionality
split of :host() into :host() and :ancestor(), which I talked about in
the f2f and which are pretty clear in the minutes.

While this update is not quite spec-worthy, it's fairly close, and
I'll be closing the gap as soon as I can.

~TJ

Received on Wednesday, 5 February 2014 03:03:30 UTC