Re: [css3-multicol] column-span in overflow

On Jun 6, 2010, at 1:58 PM, Alex Mogilevsky wrote:

> What is supposed to happen when “column-span:all” element is encountered in overflow (specifically in the case where additional columns are added outside of the overflowing box)? What should it span?
>  
> We can say that if a “column-span:all” element does not fit in the original box it should be treated as “column-span:1”. Reasonable?
>  
> There will be a boundary case of course, where “column-span:all” creates overflow, and “column-span:1” fits without overflow. I am not sure what to do in that case. I am thinking it should force balancing preceding content anyway, and then get pushed into overflow (and only span one column there).

I don't think this is the best behavior.

I believe the way to answer questions about column-span is to go back to the headers and content example that Bert posted.  I think you should get more or less the same layout if you apply the column styles to the content as you would if you apply the column styles to the container of the headers/content and put column spans on the headers.

If my suggestion to just split blocks is followed, then any height constraint that would have led to horizontal overflow would no longer be relevant to the columns, since the anonymous blocks with the column styles would not have specified heights.  You'd end up just balancing the column content and everything would just spill out of the block vertically.  I think this is more sensible behavior than deliberately ignoring the span, since the span is supposed to constitute a break in the column flow.

WebKit now implements -webkit-column-span by the way, so people should feel free to download a nightly, write test cases for it, try it out, give feedback, etc.  It uses the block splitting strategy that I outlined in the previous column-span thread.

dave
(hyatt@apple.com)

Received on Monday, 7 June 2010 18:28:32 UTC