Re: [css-selectors] Proposal: Logical Combinators / Sets

> My point is that Brian original request can be answered by allowing 
> complex selectors in :not() and :matches(), without adding redundant 
> pseudo-classes.

Yes, this is true and was already stated earlier in the thread, I think ;-)
http://lists.w3.org/Archives/Public/www-style/2013Mar/0259.html

I think the thread's first point is more about relaxing the complexity 
constraints on the existing pseudos than creating new ones.



> That’s why a "one argument" :matches() is useful (assuming the "complete" 
> profile), it provides grouping.

Oh, yes, I don't care much about the naming, ":matches" seems pretty good to 
me.

Indeed, if you use ":matches" as a name, it feels more familiar to use 
juxtaposition to have the 'and' operator than if you use ':any-of'.



> > .list .item:any-of(  .start,  :all-of(.start ~ *, :not(.end ~ *))   )
> .list .item:matches(.start, :matches(.start ~ *):not(.end ~ *))

Indeed, you can just juxtaposate the :matches to create an AND, that's 
probably good enough for most purposes. 

Received on Saturday, 6 April 2013 10:27:10 UTC