- From: David Hyatt <hyatt@apple.com>
- Date: Sun, 23 May 2010 16:05:24 -0500
- To: MURAKAMI Shinyu <murakami@antenna.co.jp>
- Cc: Hakon Wium Lie <howcome@opera.com>, Bert Bos <bert@w3.org>, "www-style@w3.org list" <www-style@w3.org>
On May 23, 2010, at 9:14 AM, MURAKAMI Shinyu wrote: > I prefer the left image, but I don't think the margins should collapse > between spanning elements and their non-spanning siblings. > I think the margins of spanning elements should be kept and > the margins of non-spanning siblings should be discarded. > I'm fine with this I guess, but note that spanning elements still need to be able to collapse their margins with the multi-column block itself. In this example: <p style="column-count:2; "> <span style="column-span: all">HHHHHHH</span> </p> You would not expect to see the top margin of the <span> inside the columns block. It needs to collapse with the <p>. You also don't want to just discard it, since imagine putting two of those together: <p style="column-count:2; "> <span style="column-span: all">HHHHHHH</span> </p> <p style="column-count:2; "> <span style="column-span: all">HHHHHHH</span> </p> You'd expect for the spans to collapse with their enclosing paragraphs and ultimately with one another as well. I really think the simplest rule is just to allow margin collapsing with both spanning and non-spanning siblings, but I'd also be fine with allowing margin collapsing with spanning children only and discarding the margins of non-spanning elements at the top of the first column (and after both forced and unforced breaks). If my proposal for how to implement column-span is used, then note that the margin collapsing of spanning elements happens naturally as a consequence of the block splitting and alterations that occur, since the enclosing block of the column-span is no longer considered multi-column (and thus no longer establishes a BFC). dave (hyatt@apple.com)
Received on Sunday, 23 May 2010 21:05:59 UTC