- From: Alan Stearns <stearns@adobe.com>
- Date: Thu, 5 Sep 2013 16:42:19 -0700
- To: Alan Stearns <stearns@adobe.com>, W3C WWW Style <www-style@w3.org>
A while back, I made a change to how shapes interact with floats, which included this simplification: On 6/19/13 3:08 PM, "Alan Stearns" <stearns@adobe.com> wrote: > >B: Where there are voids (either a large content box with a small shape, >or a large margin in a block direction) inline content will wrap to the >furthest margin edge of the float. So if you have a left float with a >shape-outside that conforms to its content box and a large margin-top, any >inline content flowing through the top margin area will be constrained by >the float's left margin edge. This avoids some stacking issues where it >could be possible for more than one left float to have a relevant edge to >consider. With this simplification, only the rightmost left float would >need to be considered for a given line box. The implications of this can be seen in the second image in example 4 in the spec. The result isn't ideal, but stacking up floats with shapes is a bit weird to begin with, so I was OK trying this simplification out. Rossen was never that happy with it, though. As it turns out, float positioning has enough complications to make this simplification moot. Add in some float clearing and you can get back into a situation where more than one float with a shape needs to be considered when laying out a given line box. So I'd like to remove the last two sentences of section 3.1 [1] and change example 4 to show the first three lines of text interacting with the shape from the first float. Thanks, Alan
Received on Thursday, 5 September 2013 23:42:45 UTC