56643 - crash [@ morkRowObject::CloseRowObject]

Status: VERIFIED FIXED

(MailNews Core :: Database, defect, P1)

Description

[David :Bienvenu]

talkback is showing crashes in morkRowObject::CloseRowObject
. I suspect this has to do with leaking message headers that live past the
forced closing of their db's (e.g., when trash is emptied).

Comment 1

[David :Bienvenu]

I don't know how to reproduce this, but I have run into it a couple times.

Comment 2

[selmer (gone)]

Working on a topcrash is good :-)

Comment 3

[Jay Patel [:jay]]

adding [@ morkRowObject::CloseRowObject] for tracking and here are some talkback 
entries and a stack trace:

morkRowObject::CloseRowObject 06b0c83b
        
http://bonsai.mozilla.org/cvsblame.cgi?file=mozilla/db/mork/src/morkRowObject.cp
p line 110
        Build: 2000100909 CrashDate: 2000-10-09 UptimeMinutes: 69  Total: 87 
        OS: Windows NT  4.0 build 1381
        URL: 
        Comment: crash when delete webmail message
         Detailed : http://climate/reports/incidenttemplate.cfm?bbid=18800826
         StackTrace: 
http://climate/reports/stackcommentemail.cfm?dynamicBBID=18800826

    morkRowObject::CloseRowObject 010f0a2d
        
http://bonsai.mozilla.org/cvsblame.cgi?file=mozilla/db/mork/src/morkRowObject.cp
p line 110
        Build: 2000101109 CrashDate: 2000-10-11 UptimeMinutes: 97  Total: 106 
        OS: Windows NT  4.0 build 1381
        URL: 
        Comment: crash wehn close seamonkey after "failed to connect to server" 
when sending msgs from aol to webmail
         Detailed : http://climate/reports/incidenttemplate.cfm?bbid=18929003
         StackTrace: 
http://climate/reports/stackcommentemail.cfm?dynamicBBID=18929003

    morkRowObject::CloseRowObject ed6ae9fd
        
http://bonsai.mozilla.org/cvsblame.cgi?file=mozilla/db/mork/src/morkRowObject.cp
p line 110
        Build: 2000101109 CrashDate: 2000-10-11 UptimeMinutes: 27  Total: 134 
        OS: Windows NT  4.0 build 1381
        URL: 
        Comment: crash when close seamonkey after " fail connect to the server" 
displayed on webmail when sending mail from AOL to Webmail
         Detailed : http://climate/reports/incidenttemplate.cfm?bbid=18929016
         StackTrace: 
http://climate/reports/stackcommentemail.cfm?dynamicBBID=18929016

    morkRowObject::CloseRowObject 010f0a2d
        
http://bonsai.mozilla.org/cvsblame.cgi?file=mozilla/db/mork/src/morkRowObject.cp
p line 110
        Build: 2000101209 CrashDate: 2000-10-12 UptimeMinutes: 527  Total: 527 
        OS: Windows 98  4.10 build 67766446
        URL: 
        Comment: 
         Detailed : http://climate/reports/incidenttemplate.cfm?bbid=18991215
         StackTrace: 
http://climate/reports/stackcommentemail.cfm?dynamicBBID=18991215

    morkRowObject::CloseRowObject ed6ae9fd
        
http://bonsai.mozilla.org/cvsblame.cgi?file=mozilla/db/mork/src/morkRowObject.cp
p line 110
        Build: 2000101213 CrashDate: 2000-10-13 UptimeMinutes: 160  Total: 160 
        OS: Windows NT  4.0 build 1381
        URL: 
        Comment: 
         Detailed : http://climate/reports/incidenttemplate.cfm?bbid=19040104
         StackTrace: 
http://climate/reports/stackcommentemail.cfm?dynamicBBID=19040104

Incident ID 19040104 
morkRowObject::CloseRowObject 
[d:\builds\seamonkey\mozilla\db\mork\src\morkRowObject.cpp, line 110] 
morkRowObject::CloseMorkNode 
[d:\builds\seamonkey\mozilla\db\mork\src\morkRowObject.cpp, line 63] 
morkNode::cut_use_count [d:\builds\seamonkey\mozilla\db\mork\src\morkNode.cpp, 
line 535] 
morkNode::CutStrongRef [d:\builds\seamonkey\mozilla\db\mork\src\morkNode.cpp, 
line 552] 
morkNode::SlotStrongNode [d:\builds\seamonkey\mozilla\db\mork\src\morkNode.cpp, 
line 447] 
morkHandle::CloseHandle [d:\builds\seamonkey\mozilla\db\mork\src\morkHandle.cpp, 
line 122] 
morkHandle::CloseMorkNode 
[d:\builds\seamonkey\mozilla\db\mork\src\morkHandle.cpp, line 59] 
morkNode::cut_use_count [d:\builds\seamonkey\mozilla\db\mork\src\morkNode.cpp, 
line 535] 
morkNode::CutStrongRef [d:\builds\seamonkey\mozilla\db\mork\src\morkNode.cpp, 
line 552] 
morkHandle::Handle_CutStrongRef 
[d:\builds\seamonkey\mozilla\db\mork\src\morkHandle.cpp, line 395] 
orkinStore::CutStrongRef 
[d:\builds\seamonkey\mozilla\db\mork\src\orkinStore.cpp, line 216] 
nsMsgHdr::~nsMsgHdr 
[d:\builds\seamonkey\mozilla\mailnews\db\msgdb\src\nsMsgHdr.cpp, line 134] 
nsMsgHdr::Release 
[d:\builds\seamonkey\mozilla\mailnews\db\msgdb\src\nsMsgHdr.cpp, line 33] 
nsCOMPtr_base::~nsCOMPtr_base 
[d:\builds\seamonkey\mozilla\xpcom\base\nsCOMPtr.cpp, line 50] 
nsMessage::`scalar deleting destructor' 
nsMessage::Release 
[d:\builds\seamonkey\mozilla\mailnews\base\util\nsMessage.cpp, line 53] 
Assertion::~Assertion 
[d:\builds\seamonkey\mozilla\rdf\base\src\nsInMemoryDataSource.cpp, line 177] 
InMemoryDataSource::DeleteForwardArcsEntry 
[d:\builds\seamonkey\mozilla\rdf\base\src\nsInMemoryDataSource.cpp, line 785] 
PL_HashTableEnumerateEntries [plhash.c, line 414] 
InMemoryDataSource::~InMemoryDataSource 
[d:\builds\seamonkey\mozilla\rdf\base\src\nsInMemoryDataSource.cpp, line 757] 
InMemoryDataSource::`scalar deleting destructor' 
InMemoryDataSource::Internal::Release 
[d:\builds\seamonkey\mozilla\rdf\base\src\nsInMemoryDataSource.cpp, line 795] 
InMemoryDataSource::Release 
[d:\builds\seamonkey\mozilla\rdf\base\src\nsInMemoryDataSource.cpp, line 791] 
nsCOMPtr_base::~nsCOMPtr_base 
[d:\builds\seamonkey\mozilla\xpcom\base\nsCOMPtr.cpp, line 50] 
nsRDFSortState::~nsRDFSortState 
nsXULTemplateBuilder::~nsXULTemplateBuilder 
[d:\builds\seamonkey\mozilla\rdf\content\src\nsXULTemplateBuilder.cpp, line 
1504] 
XULPopupListenerImpl::Release 
[d:\builds\seamonkey\mozilla\rdf\content\src\nsXULPopupListener.cpp, line 171] 
nsSupportsArray::Clear 
[d:\builds\seamonkey\mozilla\xpcom\ds\nsSupportsArray.cpp, line 320] 
nsSupportsArray::`vector deleting destructor' 
nsCOMPtr_base::~nsCOMPtr_base 
[d:\builds\seamonkey\mozilla\xpcom\base\nsCOMPtr.cpp, line 50] 
nsXULDocument::~nsXULDocument 
[d:\builds\seamonkey\mozilla\rdf\content\src\nsXULDocument.cpp, line 552] 

Comment 4

[selmer (gone)]

Scott, are you or someone else going to work on this while David's away?

Comment 5

[scottputterman]

I can look into this and see if I can find anything before David returns.

Comment 6

[scottputterman]

Adding [e-mail].

I haven't figured this out yet, but when I change folders does it blow away the
rdfsortstate for the previous folder, or are we caching all of the sort
assertions for all of the previously opened folders. If it's the latter, this is
probably making it so we keep a bunch of headers in memory that we don't need to do.

I'm also wondering, if we delete a message if its assertions get deleted from
the sort state's cache.

Anyway, I'm trying to think of a way to make it so that the cache doesn't hold
onto messages that have been deleted.

Comment 7

[David :Bienvenu]

Attachment: proposed fix for the empty trash-related instances of this

Comment 8

[David :Bienvenu]

this fix makes sure that any nsMessage objects hanging out holding onto
nsMsgHdrs with the about to be blown away db will be forece to drop their
reference to the nsMsgHdr. This happens for all deleted folders, including the
delete that happens when we empty the local trash. Similar code is run when we
compact folders (see bug 51306) I think this will fix most of these talkback
crashes.

Comment 9

[scottputterman]

this looks good to me.  I will run with it.

Comment 10

[Scott MacGregor]

sr=mscott

Comment 11

[David :Bienvenu]

can we get a plus or 2? :-) or checkin to the trunk.

Comment 12

[Phil Peterson]

Marking rtm+ to get on PDT's radar on Wed.

Comment 13

[selmer (gone)]

Given that fix, are these crashes typically caused by deleting a folder?

Comment 14

[David :Bienvenu]

Most talkback reports mention emptying the trash, which is done by deleting the
folder and creating a new one, hence the fix in deleteFolder. Also, there were
some talkback reports that mentioned deleting a folder. Putting the fix in
deleteFolder is more general, though it might have a slight performance impact
for that relatively rare operation.

Comment 15

[selmer (gone)]

rtm++

Comment 16

[David :Bienvenu]

fix checked in on trunk and branch

Comment 17

[esther]

Using build 2000-11-03 on win98, mac and linux I delete a WebMail message and
sent a message from WebMail to AOL mail while server was down (so I could get
the failed error) then closed Seamonkey.  No crash  These are the scenarios in
the talkbacks above and the only thing to go by. Verified

Comment 18

[scottputterman]

I don't think we fixed this. I still see talkback reports from the 11/01 build.

Comment 19

[David :Bienvenu]

I believe I fixed the local mail instances of this bug, but not the imap ones
(the fix I made was only for local mail). Since most of the users encountering
this crash encountered it with local mail, and most of our users will be using
local mail, I think we have fixed most instances of this.

Comment 20

[selmer (gone)]

Most existing users are local mail.  With any luck, a huge number of our new
users will be IMAP via webmail.  Is the IMAP part of this still a topcrash?  Has
another bug been opened to cover that aspect?