Re: [css3-ui] Moving pseudo-classes and pseudo-elements to the Selectors spec

Le 22/01/2013 16:15, Reece Dunn a écrit :
> On 22 January 2013 14:40, Simon Sapin <simon.sapin@kozea.fr
> <mailto:simon.sapin@kozea.fr>> wrote:
>
>     Hi,
>
>     Section 4 of css3-ui defines a few selector pseudo-classes and
>     pseudo-elements.
>
>     The pseudo-classes’ definitions (but at this point not the
>     pseudo-elements’) are now also in Selectors 4:
>     http://dev.w3.org/csswg/__selectors4/#ui-states-pseudos
>     <http://dev.w3.org/csswg/selectors4/#ui-states-pseudos>
>
>     To avoid divergence in the future, only one definition of everything
>     should be kept. I’d rather have all selectors defined in the
>     Selectors spec, but this is not a strong opinion.
>
>
> Wouldn't that create a dependency of css3-ui functionality on
> css4-selectors? That is, you would not be able to implement all css3-ui
> functionality unless css4-selectors is published and (at least
> partially) implemented as well.

I’m fine with keeping these in css3-ui, only noticing that they were 
more recently added to selectors4.

My point is that right now we have two definitions of the same thing. We 
should have one and only one. Otherwise they might diverge in the future 
if we’re not careful.


> Also, does this then also apply to the css3-values spec that references
> (but does not define) types listed in other specs (section 7 listing the
> color, image and position types). Does this mean that css3-values should
> define those types properly from the other specs as well (to be
> consistent with your proposal)?

I think css3-values is fine. For <color> it says:

> The <color> data type is defined in [CSS21] and extended in
> [CSS3COLOR]. UAs that support CSS Color Level 3 or its successor must
> interpret <color> as defined therein.

The second sentence makes clear to me that an UA can implement 
css3-value and CSS 2.1 but not css3-color, and be conformant. (Similarly 
for <image> and <position>)


> It also makes it harder to introduce new types, pseudo-elements and
> pseudo-classes in other specs. NOTE: I could see a similar issue for
> values for the display property (e.g. ruby or math) in the box
> model.

Flexbox already extends 'display':
http://www.w3.org/TR/css3-flexbox/#flex-containers

-- 
Simon Sapin

Received on Tuesday, 22 January 2013 15:50:11 UTC