[webkit-reviews] review requested: [Bug 12123] REGRESSION: Incomplete repaint of floats' overflows : [Attachment 13090] Unify floats with overflow

[bugzilla-request-daemon at macosforge.org]

mitz at webkit.org has asked  for review:
Bug 12123: REGRESSION: Incomplete repaint of floats' overflows
http://bugs.webkit.org/show_bug.cgi?id=12123

Attachment 13090: Unify floats with overflow
http://bugs.webkit.org/attachment.cgi?id=13090&action=edit

------- Additional Comments from mitz at webkit.org
This patch takes a different approach, as suggested by Hyatt. It includes a
change log, but the gist of it is to treat floats as any other overflow, while
establishing that overflow is strictly visual (by removing the only case where
overflow affects layout in WebKit, which is table cell's expand to encompass
overflow behavior). Overhanging floats whose painting is propagated to an
ancestor only factor into the overflow of the ancestor painting them. This is
done by deferring the addition of floats to the overflow until a parent gets a
chance to adopt overhanging floats. This can be changed, in the spirit of the
previous patch, to have parents adopt also overflowing floats, but for now I've
left things the way they are. Blocks that are not in a block formatting context
(including table cells) always paint their own floats and therefore their
floats are always added to their overflow.