When I enter a search in the quicksearch field while on an IMAP folder that is in threaded mode, today's Linux Thunderbird nightly crashes every time. The crash is not folder-specific or related to the search string itself, but it only happens in threaded mode; it doesn't occur when the folder is unthreaded. Here's a sample crash report: http://crash-stats.mozilla.com/report/index/a9903bc5-94fe-499c-98d5-4c3c20081113

need to fix this for beta 1 - it's not imap specific, I don't think.

quick search's override of AddHdr wasn't setting the resultIndex...I'd love to check this in before the builds are made for the test day...

Comment on attachment 348135 [details] proposed fix nsMsgDBView::AddHdr overrides the result index, is that intentional, or does this belong in the threaded case only?

Signature nsMsgThreadedDBView::MoveThreadAt(unsigned int) UUID a9903bc5-94fe-499c-98d5-4c3c20081113 Frame Module Signature [Expand] Source 0 thunderbird-bin nsMsgThreadedDBView::MoveThreadAt nsMsgThreadedDBView.cpp:798 1 thunderbird-bin nsMsgThreadedDBView::OnNewHeader nsMsgThreadedDBView.cpp:675 2 thunderbird-bin nsMsgQuickSearchDBView::AddHdr nsMsgQuickSearchDBView.cpp:136 3 thunderbird-bin nsMsgQuickSearchDBView::OnSearchHit nsMsgQuickSearchDBView.cpp:293 4 thunderbird-bin nsMsgSearchSession::AddSearchHit nsMsgSearchSession.cpp:580 5 thunderbird-bin nsMsgSearchOfflineMail::AddResultElement nsMsgLocalSearch.cpp:780 6 thunderbird-bin nsMsgSearchOfflineMail::Search nsMsgLocalSearch.cpp:732 7 thunderbird-bin nsMsgSearchSession::TimeSliceSerial nsMsgSearchSession.cpp:692 8 thunderbird-bin nsMsgSearchSession::StartTimer nsMsgSearchSession.cpp:656 9 thunderbird-bin nsMsgSearchSession::BeginSearching nsMsgSearchSession.cpp:443 10 thunderbird-bin nsMsgSearchSession::Search nsMsgSearchSession.cpp:271 11 libxpcom_core.so libxpcom_core.so@0x652f6 12 thunderbird-bin XPCWrappedNative::CallMethod js/src/xpconnect/src/xpcwrappednative.cpp:2405

it probably belongs in the threaded/grouped case only, but for some reason this crash doesn't happen on Windows, which is just strange. I'm investigating that now.

previous patch was completely wrong, except that it stopped the crash :-) With this patch, we'll figure out the newIndex if AddHdr doesn't set it for us.

fix checked in.

oh, for fun, the reason windows (at least, debug builds) didn't have an issue is that the uninitialized local var seemed to always be set to 0.