Crash in [@ mozilla::widget::StatusBarEntry::OnMessage] (diagnostic assert)
Categories
(Core :: Widget: Win32, defect)
Tracking
()
| Tracking | Status | |
|---|---|---|
| firefox-esr115 | --- | unaffected |
| firefox-esr140 | --- | unaffected |
| firefox144 | --- | unaffected |
| firefox145 | --- | unaffected |
| firefox146 | --- | fixed |
People
(Reporter: mccr8, Assigned: emilio)
References
(Regression)
Details
(Keywords: crash, regression)
Crash Data
Attachments
(1 file)
Crash report: https://crash-stats.mozilla.org/report/index/03a04efd-945e-479a-addc-263fd0251015
MOZ_CRASH Reason:
MOZ_DIAGNOSTIC_ASSERT(widget)
Top 10 frames:
0 xul.dll mozilla::widget::StatusBarEntry::OnMessage(HWND__*, unsigned int, unsigned lo... widget/windows/SystemStatusBar.cpp:189
1 xul.dll mozilla::widget::WindowProc(HWND__*, unsigned int, unsigned long long, long l... widget/windows/SystemStatusBar.cpp:246
2 user32.dll UserCallWinProcCheckWow(_ACTIVATION_CONTEXT*, int64_t (*)(tagWND*, unsigned i...
3 user32.dll DispatchClientMessage
4 user32.dll _fnDWORD
5 ntdll.dll KiUserCallbackDispatch
6 win32u.dll ZwUserPeekMessage
7 user32.dll _PeekMessage(tagMSG*, HWND__*, unsigned int, unsigned int, unsigned int, unsi...
8 user32.dll PeekMessageW
9 msctf.dll CThreadInputMgr::PeekMessageW(tagMSG*, HWND__*, unsigned int, unsigned int, u...
This assertion is old, but it seems to have started showing up on Nightly 146, build id 20251014213342. Here is the set of changes in that build.
Emilio, could this be related to bug 1994157? That's the bug I found when I searched for "widget" in that change log. Thanks.
| Assignee | ||
Comment 1•1 month ago
|
||
This probably comes from bug 1933181 instead, but that assert is wrong afaict, it's not flushing and we're now creating the widget a bit later, so there might be a popupframe without a widget more often. Looking at https://hg-edge.mozilla.org/mozilla-central/rev/5823cb0f699866c8672cce50cf8e1cbceadc43aa it wasn't clear these asserts were correct to begin with.
| Assignee | ||
Comment 2•1 month ago
|
||
This spike probably comes from bug 1933181, but that assert is wrong
afaict. It's not flushing and we're now creating the widget a bit later,
so there might be a popupframe without a widget more often.
Looking at the commit that introduced these, it wasn't clear these
asserts were correct to begin with.
https://hg-edge.mozilla.org/mozilla-central/rev/5823cb0f699866c8672cce50cf8e1cbceadc43aa
I think the assert below (every widget hanging from a popup having an
hwnd) should hold.
Updated•1 month ago
|
| Reporter | ||
Updated•1 month ago
|
Comment 4•1 month ago
|
||
| bugherder | ||
Updated•1 month ago
|
Updated•14 days ago
|
Updated•12 days ago
|
Description
•