>> I think you mean "Elements [that generate a local stacking context] have
>> 2 stack *levels*".
> Again, I'm trying to describe the issue in plain language. I could just
> recite the spec, but the problems is that not everybody understands what
> it sais however many times they seem to read/quote it.
Woops sorry. Then ignore the stuff inside the square brackets and read
that as "Elements have 2 stack levels" Stack levels and stacking contexts
are very different terms, and you shouldn't confuse between them.
>> No, the application user interface should not be involved with this at
>> all. It is a big mistake for Mozilla to be putting the application into
>> the picture when rendering a document.
> No, the problem is that you don't realize that Mozilla has the only
> consitant in real life working model that follow the spec.
Can you please explain? CSS is not a language for styling a user
interface. *Nothing* should be rendered beneath the application's canvas.
Yes, it's probably OK to let boxes go under the root element, but the
outside of the UI is out of bounds.
Looking back at your first post, you say that the root's parent is the
canvas of the app itself. Firstly, the root has no parent, at least in
the CSS sense. If you mean parent in a broader sense, then you are talking
about something outside the scope of the rendering engine. If it's outside
the scope of the renderer, then CSS boxes cannot interact with it. Period.
Anyway, this discussion is completely off-topic for this list and should
never have started. I'll stop now.
Kevin W :-)