[Webkit-unassigned] [Bug 9121] REGRESSION: [Incremental Repaint] DHTML movement test failures

[bugzilla-daemon at opendarwin.org]

http://bugzilla.opendarwin.org/show_bug.cgi?id=9121





------- Comment #3 from opendarwin.org at mitzpettel.com  2006-05-31 15:08 PDT -------
Created an attachment (id=8629)
 --> (http://bugzilla.opendarwin.org/attachment.cgi?id=8629&action=view)
Sketch: remove positioned objects from their old containing block

The layout problem stems from the following situation: if object C is
absolutely positioned, its parent B is static, and B's parent A is relative,
and then you change B to be relative, C isn't removed from A's list of
positioned objects. The patch gives a very rough sketch of a possible fix. It
doesn't address the other direction (B going from relative to static), which
should be easier, it has obvious deficiencies (indicated by the FIXMEs) and it
doesn't mark the positioned objects for layout, which it should.

However, I am not 100% convinced that this approach will also fix the repaint
issue. It might be easier, and acceptable, to simply change Node::diff to
return Detach upon changes in position to/from static.


-- 
Configure bugmail: http://bugzilla.opendarwin.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug, or are watching the assignee.