Re: [css3-multicol] should spanner margins collapse?

On 23/11/2011 16:44, Florian Rivoal wrote:
> On Mon, 21 Nov 2011 21:11:36 +0100, Rossen Atanassov
> <Rossen.Atanassov@microsoft.com> wrote:
>
>> Why not option C.
>> Margin collapsing only makes sense on a boundary between two blocks
>> elements belonging to the same normal flow.
>
> Two consecutive spanners are in the same normal flow with respect to
> each other. The combined group of consecutive spanners isn't in normal
> flow with regards to the rest of the page, so not having margin
> collapsing between spanners and non spanning content is reasonable, but
> that's not what we are discussing.

Hang on, that's only the case in Option A where there is an anonymous 
BFC wrapper placed around the spanners.  In other cases the spanner is 
in the same flow as non-spanner content and can theoretically collapse 
with it.

However, I think I'm seeing one of Rossen's concerns more clearly now: 
it seems to be undefined how spanners interact with the columns whose 
breakage they cause.  Margin collapsing merely exacerbates the problem, 
so I can see why we might want to run away from Option C.  At least the 
other two options sidestep the problem by providing a fixed barrier 
between the spanners and their surrounding content.  I guess this puts 
me in favour of Option A again.

I think that the difficulty we're having is just a symptom of a problem 
at a deeper layer; the flow of content into columns and the behaviour of 
content at either side of column boundaries is not sufficiently defined. 
  I don't think we can hope to finalize spanners until that's cleared up.

Cheers,
Anton Prowse
http://dev.moonhenge.net

Received on Monday, 28 November 2011 16:00:37 UTC