###!!! ASSERTION: Did not find ourselves on the placeholder's ancestor chain: 'f', file /Users/admin/trunk/mozilla/layout/generic/nsFrame.cpp, line 1302 The only mention of this assertion I can find is bug 317375 comment 51.
In the case, the placeholder for aChild is a child of the next-in-flow of aFrame (because of the linebreak and because we stick the abs pos frames in the first-in-flow of the inline). roc, does this function need changes to handle this case and similar cases? Or does the assert need to change?
Related to bug 288357 comment 19?
That's the block version of this, yes. I was talking about it when I mentioned "similar cases" in comment 2. Is there a bug on us not rendering the text in this case? That's a trunk regression from bug 317375 and should probably be blocking 1.9 alphas... It should probably also block this bug.
This refactors the MarkOutOfFlow stuff, including the caret usage. The main thing is that nsDisplayListBuilder now records all the frames that were marked and unmarks them when the builder is destroyed (i.e. after all painting is done). This is actually a slight simplification and also fixes this bug by allowing the first-in-flow to mark a set of frames to be descended into even though we don't reach those frames until we paint one of its next-in-flows.
Assignee: nobody → roc
Status: NEW → ASSIGNED
Attachment #219824 - Flags: review?(mrbkap)
Comment on attachment 219824 [details] [diff] [review] fix Looks good.
Attachment #219824 - Flags: review?(mrbkap) → review+
Status: ASSIGNED → RESOLVED
Last Resolved: 13 years ago
Resolution: --- → FIXED
Crashtest checked in.
You need to log in before you can comment on or make changes to this bug.