RE: [CSS OM] list of issues

[Anne van Kesteren:] 
> On Wed, 01 Jun 2011 08:37:57 +0200, Sylvain Galineau
> <sylvaing <at> microsoft.com> wrote:
> > [Daniel Glazman:]
> >> That's undefined at this time  But if we read the CSS OM spec with
> >> a strict eye, it says the second parameter of getComputedStyle() is a
> >> pseudo-element. "first-line" is not a pseudo-element; ":first-line"
> >> and "::first-line" are...
> >
> > If we do read it with a strict eye then only ::before and ::after are
> > supported [1]...so I'd very much argue against reading this
> > normatively at this point :)
> 
> Maybe :first-line and :first-letter ought to be added. But :selection for
> instance does not make much I think. To what element does it apply? If you
> really want answers to all of that you need to expose the render tree to
> script, but I am not sure we want to go there.
> 
> 
> > [1] http://dev.w3.org/csswg/cssom/#extensions-to-the-window-interface


The issue I raised was actually far more pedestrian than which pseudo-elements
are exposed through the OM or whether the render tree should be available. The
question was: given a pseudo-element ::foo exposed through the second optional
argument to getComputedStyle(), what are the string values that map to ::foo ?

Is it: "::foo" only ? "::foo" or ":foo" ? "::foo" or ":foo" or "foo" ? In other
words, what constitutes the name of ::foo for this argument ? 
not clear that the