Re: [css3] Suggestion: Selector variables or “synonyms”

> >   .old = .new;
> >   .new {}
> >   .new, .somethingelse {}
> >   .new, .different {}
> >   .new, .whatever {}
> Saying that it behaves a certain way in a certain case isn't enough
> to define how it behaves in all cases.  You need to do that in order
> to make it clear what it is that you're proposing, since in this
> case there are a *lot* of possibilities.

Right, the proposal's certain no role model for “best practice spec writing”. Yet, hopefully.

> For example, given ".old = .new", does a selector "" match an p
> element with class="old"?

Yes, it should.

> Does it matter if the selector is in the same style sheet as the
> assignment expression?

Yes, the “expression” should be applied.

> Does it matter what order they're in?

That probably depends on the processing, but no.

> What happens when more complex things are on one or both sides of the
> assigment?  Are multiple assignments transitive?

I cannot perfectly state that, but any combination of selectors and “placeholder” should be basically fine:

  E = F; /* Valid */
  E = F, G; /* Valid */
  E, F = G; /* Valid, same as above */
  @E, F = G, H; /* Valid */
  @E = F; /* Valid */

However, the following statement should be either invalid or only valid in conjunction with a statement resolving either “@E” or “@F”:

  @E = @F;

> Note that you don't need to answer all these questions; it would be
> better to have a simple statement of what you propose from which all
> these answers (and others) could be derived. If you're not sure
> what the answers to some of the questions should be, you're probably
> better off proposing a list of requirements for the feature rather
> than the actual details / syntax of the feature.

(Thanks for understanding that indeed, I couldn't write such a spec part all myself yet.)

Well, cases like

  @foo, E = .bar, #baz
  E@foo {}

will certainly be interesting, but I'd love to find answers and help in building requirements. What will be most helpful next?

> (And I'd also remind you that whatever it is has to fit with the
> model of selectors, where the question selectors are used to answer
> is "does selector S match element E?".)

Of course.

Jens Meiert

Received on Wednesday, 6 February 2008 19:58:23 UTC