- From: Tab Atkins Jr. <jackalmage@gmail.com>
- Date: Thu, 29 Mar 2012 13:55:02 -0700
- To: "Kang-Hao (Kenny) Lu" <kennyluck@csail.mit.edu>
- Cc: François REMY <fremycompany_pub@yahoo.fr>, Christoph Päper <christoph.paeper@crissov.de>, www-style CSS <www-style@w3.org>
On Thu, Mar 29, 2012 at 1:36 PM, Kang-Hao (Kenny) Lu
<kennyluck@csail.mit.edu> wrote:
> (12/03/29 17:56), François REMY wrote:
>> Most of your proposal just doesn't make sense.
>>
>> Each pseudoclass has a different meaning, how can you say "I just want
>> one of them"? Seriously, why would you apply the same styling for a
>> :hover element, a :invalid and a :valid element ? Matching any
>> parametric pseudoclass is also pure non-sense. :*(xxx) would match
>> :any(:matches(xxx),:not(xxx)) which mean all elements.
>
> Agreed. But perhaps there might be use cases of *::*, which should
> include all elements and pseudo-elements all together, esp. for
> non-inheritable properties, although I haven't thought of anything that
> I would truly consider useful.
>
> (If 'image-orientation' were not changed to be inheritable.
>
> *::* { image-orientation: from-exif; }
>
> might be useful maybe? Would *::* { clear: both; } be any of use? What
> about *::* { box-decoration-break: clone; } ? )
Yeah, I can see use-cases for this.
This just shows, yet again, that pseudo-elements should be treated
like real elements, not as aspects of another element. If they were
real elements and :: was a combinator or something, you'd just do: "*
:: *" and be done with it.
~TJ
Received on Thursday, 29 March 2012 20:55:58 UTC