323497 - ASSERTION: SetParent failed! (nsViewManager::ReparentChildWidgets)

Status: RESOLVED WORKSFORME

Description

[Jesse Ruderman]

###!!! ASSERTION: SetParent failed!: 'NS_SUCCEEDED(rv)', file /Users/admin/trunk/mozilla/view/src/nsViewManager.cpp, line 2630
Break: at file /Users/admin/trunk/mozilla/view/src/nsViewManager.cpp, line 2630

Comment 1

[Jesse Ruderman]

Created attachment 208538 [details]
testcase

Comment 2

[Martijn Wargers (zombie)]

This is wfm in my debug build, which contains the fix for bug 317375, so I guess this depends on bug 317375.

Comment 3

[Jesse Ruderman]

WFM Mozilla/5.0 (Macintosh; U; PPC Mac OS X Mach-O; en-US; rv:1.9a1) Gecko/20060310 Firefox/1.6a1 late

WFM, probably fixed by the switch to frame display lists.

Comment 4

[Jesse Ruderman]

I'm seeing this again now, with the same testcasce.

Comment 5

[Boris Zbarsky [:bzbarsky, bz on IRC]]

So the assert I get (on Linux) is:

###!!! ASSERTION: nsWindow::SetParent - reparenting a non-child window: 'Not Reached', file ../../../../mozilla/widget/src/gtk2/nsWindow.cpp, line 493

It looks like we're calling

#0  nsWindow::SetParent (this=0x8643520, aNewParent=0x85fea48)

where |this| is the widget for the view of the nsListControlFrame.  In particular, it's a floating view, so the widget parent chain is weird, no?

On Mac, I bet we're calling SetParent on an nsWindow, which throws NS_ERROR_NOT_IMPLEMENTED and triggers the assert Jesse sees.

Comment 6

[Jesse Ruderman]

WFM on Mac trunk.  The text is ugly (grayscale antialiasing rather than subpixel antialiasing), but that's not this bug.

bz, are you still seeing the assertion you mentioned in comment 5 (on Linux)?

Comment 7

[Boris Zbarsky [:bzbarsky, bz on IRC]]

No, seems to be worksforme too.

Comment 8

[Jesse Ruderman]

The ugly text is bug 379786.

Comment 9

[Jesse Ruderman]

Crashtest checked in.