- From: Tab Atkins Jr. <jackalmage@gmail.com>
- Date: Tue, 11 May 2010 16:45:01 -0700
- To: "L. David Baron" <dbaron@dbaron.org>
- Cc: Alex Mogilevsky <alexmog@microsoft.com>, "www-style@w3.org list" <www-style@w3.org>
On Tue, May 11, 2010 at 1:10 PM, L. David Baron <dbaron@dbaron.org> wrote: > On Tuesday 2010-05-11 19:45 +0000, Alex Mogilevsky wrote: >> This should be editorial. I see that some properties (e.g. >> 'box-flex') apply to "in-flow children of box elements". Others >> (e.g. 'box-ordinal-group') apply to "children of box elements". >> >> They should probably all apply to simply "children" as there isn't >> any flow is a box element. > > Well, there's the question of what happens if a child of a flexbox > is floated or absolutely positioned. > > My inclination would probably be that floating simply doesn't do > anything, but that we might want absolute positioning to work... in > which case the absolutely positioned children would be out-of-flow. I am absolutely of the opinion that float only has meaning within normal flow, and is ignored in all other flows. Table layout does something different there, but that's a mistake imo that we might be stuck with. Absposing an element should take it out of the flow entirely, and not interact with its siblings or ancestors in any way. So a floated flexbox child should act exactly like it wasn't floated, while an abspos flexbox child should act like an abspos and not participate in the flexbox layout at all. ~TJ
Received on Tuesday, 11 May 2010 23:45:54 UTC