Re: [css3-regions] regions forming stacking contexts

On Thu, Feb 2, 2012 at 9:02 PM, Alex Mogilevsky <alexmog@microsoft.com>wrote:

> It seems there are some optimizations that become a problem with
> fragmentation, but somehow are not a problem in print preview. I don't know
> how Mozilla rendering works (and I don't need to know...)
>

It's not about optimizations, but about making sure that all layout and
rendering is well-defined.

Printing and print preview behave mostly like multicol (except that pages
can have different widths and columns can't). Regions add region styling,
and the ability to break an element into parts that have different ancestor
chains in layout and the DOM. These break existing implicit invariants such
as "all boxes for an element have the same computed 'background' style" and
"the containing block boxes for each box for an element all belong to the
same element (and thus have the same computed style for most properties)".

Here's another example of questions arising from the latter. Consider two
regions R1 and R2, where R1 is position:relative and R2 is position:static,
and we flow content into the regions containing an element that's
position:absolute. How do we determine the abs-pos containing block of the
element? Is it possible for the element to split into multiple boxes with
different containing blocks? Could a simple layout change, such adding some
text to the DOM before the abs-pos element, push the abs-pos element from
R1 to R2 and thus change its containing block?

Really, the people implementing regions, not me, are in a better position
to make sure all these issues are raised and addressed, because they'll run
into them creating and testing the implementation.

Rob
-- 
"If we claim to be without sin, we deceive ourselves and the truth is not
in us. If we confess our sins, he is faithful and just and will forgive us
our sins and purify us from all unrighteousness. If we claim we have not
sinned, we make him out to be a liar and his word is not in us." [1 John
1:8-10]