Created attachment 250736 [details] testcase Clicking the button in the testcase triggers an assertion: ###!!! ASSERTION: Shouldn't hit this case -- we should a be a combobox if we have no size set and no multiple set!: 'Not Reached', file /Users/admin/trunk/mozilla/layout/forms/nsListControlFrame.cpp, line 1988 This assertion appears to have been added by the landing of the reflow branch.
I can't seem to reproduce this. Can you post the stack to the assertion?
Created attachment 251033 [details] stack trace (mac debug / gdb) Hmm, I had to load it from file:/// to reproduce the assertion.
Created attachment 251149 [details] [diff] [review] Fix, sort of This seems like the right thing to do, and fixes this bug for me (I could also reproduce when loading locally only). However, this makes the "wasDirty lied" assertion to appear a _lot_ more often (3-4 times during startup for me)... I blame scrollbars, but maybe we want to hold off on taking this patch until the scrollbar thing is resolved?
Comment on attachment 251149 [details] [diff] [review] Fix, sort of r+sr=dbaron. I think the scrollbar thing is resolved now, no?
Comment on attachment 251149 [details] [diff] [review] Fix, sort of I actually thought about this some more, and I'm not happy with this patch... this makes a reflow-only flush also flush out style reresolves, which means it becomes unsafe to do those flushes from frame code. Sadly, we have frames which do that. :(
WFM, Mac trunk debug. bz, you had a patch here, but I'm guessing based on your comments that it's no longer useful.
Status: NEW → RESOLVED
Last Resolved: 11 years ago
Resolution: --- → WORKSFORME
Crashtest checked in.
Flags: in-testsuite? → in-testsuite+
You need to log in before you can comment on or make changes to this bug.