70875 - Hit cancel to Windows Integration dialog crashes application.

Status: VERIFIED WORKSFORME

(SeaMonkey :: UI Design, defect, P1)

Description

[Joseph Elwell]

I was using a commercial build from friday March 2nd 2001. I launched netscp6 -
aim then clicked on the navigator item in the lowest toolbar. hit cancel on the 
windows integration question. *crash*.

Here's the stack trace:
JS_GetFunctionObject(JSFunction * 0x01010100) line 2677 + 3 bytes
nsScriptSecurityManager::GetFunctionObjectPrincipal(nsScriptSecurityManager * 
const 0x0286b650, JSContext * 0x06ff4bb0, JSObject * 0x04bf11d0, nsIPrincipal * 
* 0x0012ebd4) line 941 + 10 bytes
nsScriptSecurityManager::CheckFunctionAccess(nsScriptSecurityManager * const 
0x0286b650, JSContext * 0x06ff4bb0, void * 0x04bf11d0, void * 0x00cf5b10) line 
650 + 44 bytes
nsJSContext::CallEventHandler(nsJSContext * const 0x06ff4d60, void * 
0x00cf5b10, void * 0x04bf11d0, unsigned int 1, void * 0x0012ec7c, int * 
0x0012ec78, int 0) line 934 + 38 bytes
nsJSDOMEventListener::HandleEvent(nsIDOMEvent * 0x077684f4) line 91 + 52 bytes
nsEventListenerManager::HandleEventSubType(nsListenerStruct * 0x075754c0, 
nsIDOMEvent * 0x077684f4, nsIDOMEventTarget * 0x06ff5090, unsigned int 1, 
unsigned int 7) line 838 + 19 bytes
nsEventListenerManager::HandleEvent(nsIPresContext * 0x047cd280, nsEvent * 
0x0012f178, nsIDOMEvent * * 0x0012f13c, nsIDOMEventTarget * 0x06ff5090, 
unsigned int 7, nsEventStatus * 0x0012f19c) line 1415 + 39 bytes
GlobalWindowImpl::HandleDOMEvent(GlobalWindowImpl * const 0x06ff5080, 
nsIPresContext * 0x047cd280, nsEvent * 0x0012f178, nsIDOMEvent * * 0x0012f13c, 
unsigned int 1, nsEventStatus * 0x0012f19c) line 575
DocumentViewerImpl::LoadComplete(DocumentViewerImpl * const 0x047cef20, 
unsigned int 0) line 712 + 47 bytes
nsDocShell::EndPageLoad(nsIWebProgress * 0x06ff7da4, nsIChannel * 0x04786260, 
unsigned int 0) line 2655
nsWebShell::EndPageLoad(nsIWebProgress * 0x06ff7da4, nsIChannel * 0x04786260, 
unsigned int 0) line 992
nsDocShell::OnStateChange(nsDocShell * const 0x06ff7ec0, nsIWebProgress * 
0x06ff7da4, nsIRequest * 0x04786260, int 131088, unsigned int 0) line 2573
nsWebShell::OnStateChange(nsWebShell * const 0x06ff7ec0, nsIWebProgress * 
0x06ff7da4, nsIRequest * 0x04786260, int 131088, unsigned int 0) line 954
nsDocLoaderImpl::FireOnStateChange(nsIWebProgress * 0x06ff7da4, nsIRequest * 
0x04786260, int 131088, unsigned int 0) line 1309
nsDocLoaderImpl::doStopDocumentLoad(nsIRequest * 0x04786260, unsigned int 0) 
line 736
nsDocLoaderImpl::DocLoaderIsEmpty(unsigned int 0) line 632
nsDocLoaderImpl::OnStopRequest(nsDocLoaderImpl * const 0x06ff7d94, nsIRequest * 
0x07708180, nsISupports * 0x00000000, unsigned int 0, const unsigned short * 
0x100b0c00 gCommonEmptyBuffer) line 564
nsLoadGroup::RemoveRequest(nsLoadGroup * const 0x06ff7d20, nsIRequest * 
0x07708180, nsISupports * 0x00000000, unsigned int 0, const unsigned short * 
0x100b0c00 gCommonEmptyBuffer) line 518 + 48 bytes
nsJARChannel::OnStopRequest(nsJARChannel * const 0x07708184, nsIRequest * 
0x0772d4a4, nsISupports * 0x00000000, unsigned int 0, const unsigned short * 
0x100b0c00 gCommonEmptyBuffer) line 612
nsOnStopRequestEvent::HandleEvent(nsOnStopRequestEvent * const 0x076f7870) line 
178 + 67 bytes
nsStreamObserverEvent::HandlePLEvent(PLEvent * 0x076f7874) line 79
PL_HandleEvent(PLEvent * 0x076f7874) line 576 + 10 bytes
PL_ProcessPendingEvents(PLEventQueue * 0x0055f4d0) line 509 + 9 bytes
_md_EventReceiverProc(HWND__ * 0x0003014e, unsigned int 49326, unsigned int 0, 
long 5633232) line 1054 + 9 bytes
USER32! 77e13eb0()
USER32! 77e1401a()
USER32! 77e192da()
nsAppShellService::Run(nsAppShellService * const 0x00585380) line 408
main1(int 2, char * * 0x004b41a0, nsISupports * 0x00000000) line 1004 + 32 bytes
main(int 2, char * * 0x004b41a0) line 1295 + 37 bytes
mainCRTStartup() line 338 + 17 bytes
KERNEL32! 77e87903()

Comment 1

[R.K.Aa.]

the backtrace is the same as some described in topcrash bug 70361

Comment 2

[sairuh (rarely reading bugmail)]

jrgm, d'you see this?

Comment 3

[John Morrison]

It is the same stack as that other top crasher bug 70361, as R.K. notes. 
I don't know if there is anything specific that needs to be done for this
case, or if this should just be closed as a dup.

Comment 4

[Oliver Klee]

I see this if I launch Mozilla Mail only (e.g. if I select "If Mozilla starts
up, open ... Mozilla Mail" in Preferences and open Navigator (from MailNews).

I guess this has nothing to do with the Windows Integration dialog.

Comment 5

[Oliver Klee]

Attachment: Dr. Watson log for my case

Comment 6

[sairuh (rarely reading bugmail)]

oliver, do you still see this? i wasn't able to reproduce this by having the app
launch mail only [2001.03.08.13 on linux, mac and winNT]. should i mark this
wfm? or should it be dup'd?

Comment 7

[Oliver Klee]

I'm still seeing this with 2001031008 Win2k. The crash occurs if I open the
Navigator window. So this should probably be dupe (to bug 31837 or which bug?).

Comment 8

[Paul Chen]

Marking nsbeta1+, p1, mozilla0.9.1

Comment 9

[Samir Gehani]

Taking from Paul.

Comment 10

[Samir Gehani]

Can't reproduce this on a WinNT or on a WinNT system (debug or opt).  There are
2 sets of steps listed here.  I couldn't reproduce the crash with either.  If
you can still reproduce this please list your OS and the exact steps used to
reproduce this crash.  Thanks.

The 2 sets of steps I tried were:

Set A
-----
1> Ensure windows integration dialog is going to pop up by editing the windows
registry key: HKEY_LOCAL_MACHINE\SOFTWARE\Mozilla\Desktop and adding the name
data pair: name set to "showDialog" and data set to "1"
2> Launch with N6 with the -aim flag.
3> Click on the navigator icon in the bottom left toolbar.
4> Click "Cancel" on the windows integration dialog.
5> Navigator doesn't crash.

Set B
-----
1> Launch mozilla with the -mail flag.
2> Click on the navigator icon in the bottom left toolbar.
3> Navigator doesn't crash.

Joseph, Oliver,
Are either of you still seeing this crash?  Thanks.

Comment 11

[Samir Gehani]

Oops, I meant "Can't reproduce this on a Win2K or on a WinNT system"
                                            ^^

Comment 12

[Oliver Klee]

This no WFM, too.

Comment 13

[Oliver Klee]

Sorry for the type, they stole the "W" from my keyboard. =.->

I meant: this _now_ WFM, too.

Comment 14

[sairuh (rarely reading bugmail)]

moving Windows Integration [formerly Desktop Integration] bugs to PaulW as qa
contact.

Comment 15

[Paul Wyskoczka]

Marking verified