Re: target-attr() and /reference combinator/ proposals

> I think for 'simple' cases :matches() is better than :: (or :select),
> in particular because it copes with many more cases without explicit
> scoping. For more complex cases, scoping is needed:
> 
>    A:matches(/attr/ (C B))
>    C (A /attr/ B)
>    A /attr/ (C B)
>    A /attr/ (C /attr/ B)
> 
> ...or the equivalent of :matches() but starting at the document root,
> as in (equivalents to the previous four examples):
> 
>    A:matches(/attr/ B:matched(C))
>    C B:matched(A /attr/)
>    A /attr/ B:matched(C)
>    A /attr/ B:matched(C /attr/)
> 
> Hmm. I like it. :-)
Me too. Scoping is not very CSS like, and unclear.
But I think :matched is not a good idea.
Instead I'd like to propose a 'this' indicator inside :matches(), f.e. #

The examples would then become:

    A:matches(/attr/ B:matches(C #))
    C B:matches(A /attr/ #)
    A /attr/ B:matches(C #)
    A /attr/ B:matches(C /attr/ #)

And :matches() should also support multiple selectors:

    B:matches(A # C,C # A)

is equivalent to

    A B:selected C,C B:selected A

Does this sound like a good idea?
 
> > In fact, pseudo-elements have nothing to do with selecting, it's
> > about creating. I would like to see something like:
> > 
> > selector {
> >   color: green;
> >   @pseudo-element before {
> >     content: "[";
> >   }
> >   @pseudo-element after {
> >     content: "]";
> >   }
> > }
> > 
> > That's also because different pseudo-elements will be used together
> > with the same selector at the same time. (often both 'before' and
> > 'after' will be used f.e.)
> 
> Interesting, but I think it's probably too late to change the syntax
> that radically. The above cuts deep into how pseudo-elements are
> treated, and there are already working implementations that have had
> bugs wrung out of them, e.g. Mozilla.
> 
I dunno. It's no structural change.

> [2] http://www.bath.ac.uk/%7Epy8ieh/internet/wwwstyle/selectors.txt
about this:

     NOT EXACT MATCH
          e.g. [attr="value"]

I think a ! is missing here?

Should there be negative pseudo-elements? 
Or should pseudo-elements simply not add to the specificity weight?

Received on Tuesday, 25 January 2000 11:21:03 UTC