1841154 - Startup Crash in [@ mozilla::detail::InvalidArrayIndex_CRASH | nsMsgDBView::Sort]

Status: RESOLVED FIXED

(Thunderbird :: General, defect, P1)

Description

[Wayne Mery (:wsmwk)]

Crashes start with 115.0b5. Mac only afaict.

Regression from Bug 1498313 ?

Crash report: https://crash-stats.mozilla.org/report/index/974cbf76-c3c5-4b45-ad2b-6123e0230629

Reason: EXC_BAD_ACCESS / KERN_INVALID_ADDRESS

Top 10 frames of crashing thread:

0  libmozglue.dylib  MOZ_Crash  mfbt/Assertions.h:264
0  libmozglue.dylib  mozilla::detail::InvalidArrayIndex_CRASH  mfbt/Assertions.cpp:50
1  XUL  std::__1::__introsort<std::__1::_ClassicAlgPolicy, nsMsgDBView::Sort  /builds/worker/fetches/MacOSX13.3.sdk/usr/include/c++/v1/__algorithm/sort.h
2  XUL  std::__1::__introsort<std::__1::_ClassicAlgPolicy, nsMsgDBView::Sort  /builds/worker/fetches/MacOSX13.3.sdk/usr/include/c++/v1/__algorithm/sort.h:593
3  XUL  std::__1::__introsort<std::__1::_ClassicAlgPolicy, nsMsgDBView::Sort  /builds/worker/fetches/MacOSX13.3.sdk/usr/include/c++/v1/__algorithm/sort.h:593
4  XUL  std::__1::__introsort<std::__1::_ClassicAlgPolicy, nsMsgDBView::Sort  /builds/worker/fetches/MacOSX13.3.sdk/usr/include/c++/v1/__algorithm/sort.h:590
5  XUL  std::__1::__introsort<std::__1::_ClassicAlgPolicy, nsMsgDBView::Sort  /builds/worker/fetches/MacOSX13.3.sdk/usr/include/c++/v1/__algorithm/sort.h:593
6  XUL  std::__1::__introsort<std::__1::_ClassicAlgPolicy, nsMsgDBView::Sort  /builds/worker/fetches/MacOSX13.3.sdk/usr/include/c++/v1/__algorithm/sort.h:593
7  XUL  std::__1::__sort<nsMsgDBView::Sort  /builds/worker/fetches/MacOSX13.3.sdk/usr/include/c++/v1/__algorithm/sort.h:627
7  XUL  std::__1::__sort_impl[abi:v15006]<std::__1::_ClassicAlgPolicy, mozilla::ArrayIterator<IdUint32*&, nsTArray_Impl<IdUint32, nsTArrayInfallibleAllocator> >, nsMsgDBView::Sort  /builds/worker/fetches/MacOSX13.3.sdk/usr/include/c++/v1/__algorithm/sort.h:687

Comment 1

[Sean Burke [:leftmostcat]]

This seems to be a result of the sorting changes from bug 1498313, yes. We seem to be experiencing fewer crashes overall than when using NS_QuickSort(), at least. I may have an idea of what's going wrong here, but it's very hard to tell as I'm unable to reproduce and we don't have any reporters reproducing. I will put up a patch and we can keep an eye on crash stats in daily.

Comment 2

[Sean Burke [:leftmostcat]]

Attachment: Bug 1841154 - Don't return zero for unequal values in message sort. r=babolivier

Comment 3

[Pulsebot]

Pushed by [e-mail]:
https://hg.mozilla.org/comm-central/rev/5ce9735466f3
Don't return zero for unequal values in message sort. r=babolivier

Comment 4

[Wayne Mery (:wsmwk)]

Approx 25% of crashes are 15 seconds or less uptime, 60% less than 60 seconds.

Comment 5

[Wayne Mery (:wsmwk)]

Comment on attachment 9343342 [details]
Bug 1841154 - Don't return zero for unequal values in message sort. r=babolivier

[Triage Comment]
Approved for esr115

This isn't happening to a ton of users but, for those who encounter it, a significant percentage are stuck crashing at startup crash. If you don't mind respining 115.1.1 this would be a good patch to take.

Comment 6

[Rob Lemley [:rjl]]

Thunderbird 115.1.1:
https://hg.mozilla.org/releases/comm-esr115/rev/3302e2044b9c

Comment 7

[Wayne Mery (:wsmwk)]

Verified crash is gone in 115.1.1

Thanks for the patch and review!