503811 - crash [@ nsMsgDBView::MarkThreadOfMsgRead(unsigned int, unsigned int, nsTArray<unsigned int>&, int)] marking thread read

Status: VERIFIED FIXED

(MailNews Core :: Backend, defect)

Description

[Wayne Mery (:wsmwk)]

crash [@ nsMsgDBView::MarkThreadOfMsgRead(unsigned int, unsigned int, nsTArray<unsigned int>&, int)] 

I hit this marking thread read in a virtual folder.
based on crash-stats is quite rare - a few per month.
goes back at least as far as 3.0b1

bp-32466d25-3d71-4ec9-902f-ff67b2090712
0	thunderbird.exe	nsMsgDBView::MarkThreadOfMsgRead	 mailnews/base/src/nsMsgDBView.cpp:5815
1	thunderbird.exe	nsMsgDBView::SetThreadOfMsgReadByIndex	mailnews/base/src/nsMsgDBView.cpp:3019
2	thunderbird.exe	nsMsgDBView::ApplyCommandToIndices	mailnews/base/src/nsMsgDBView.cpp:2697
3	thunderbird.exe	nsMsgSearchDBView::DoCommand	mailnews/base/src/nsMsgSearchDBView.cpp:811
4	xpcom_core.dll	NS_InvokeByIndex_P	xpcom/reflect/xptcall/src/md/win32/xptcinvoke.cpp:101
5	thunderbird.exe	XPCWrappedNative::CallMethod	js/src/xpconnect/src/xpcwrappednative.cpp:2454
6	nspr4.dll	PR_Lock	nsprpub/pr/src/threads/combined/prulock.c:233
7	js3250.dll	js3250.dll@0xbff6f	


There is perhaps another issue ... crash-stats for 3.1a1pre crashes does not show the source lines eg. bp-c764df28-ec69-46de-9d93-34e0a2090710
I see also lack of line # in bp-1c57ec5e-4374-4446-a43b-6dfb92090710 for a different crash.  But I don't see this in all 3.1a1pre crashes

Comment 1

[Kent James (:rkent)]

I'm guessing this code:

  5812     nsCOMPtr <nsIMsgDBHdr> firstHdr;

  5813     threadHdr->GetChildAt(0, getter_AddRefs(firstHdr));

  5814     nsMsgKey firstHdrId;

  5815     firstHdr->GetMessageKey(&firstHdrId);

is returning null firstHdr, so checking that should prevent this. The call to GetChildAt returns errors if message is not found, but we are ignoring that here.

I'll do this simple patch if there is no controversy about that.

Comment 2

[Kent James (:rkent)]

Attachment: Add check for null header

Comment 3

[Mark Banner (:standard8)]

Comment on attachment 392433 [details] [diff] [review]
Add check for null header

I guess this comes under the general heading of the fact that we occasionally get issues in threads. In any case, checking the result is definitely a good idea here.

r/sr=Standard8.

Comment 4

[Mark Banner (:standard8)]

Checked in: http://hg.mozilla.org/comm-central/rev/eaf4d5eb5fa1

Comment 5

[Wayne Mery (:wsmwk)]

I think we can safely say this is gone.
not since 2009072300 build, i.e. almost 3 months, despite an average of 2 crashes/month in prior nightly builds.