same string of issues as before, there is no profile dir, the initializations all failed ealier, NSS doesn't need to shoot a user at the end of a session, at worst it should fail silently, at best it shouldn't be called at all. NTDLL! 77f9f9df() ShutdownCRLCache() line 1038 + 32 bytes NSS_Shutdown() line 546 nsNSSComponent::ShutdownNSS() line 1091 nsNSSComponent::~nsNSSComponent() line 255 nsNSSComponent::`scalar deleting destructor'() + 15 bytes nsNSSComponent::Release(nsNSSComponent * const 0x04e802c8) line 1150 + 151 bytes nsSupportsArray::Clear(nsSupportsArray * const 0x01073f08) line 560 + 54 bytes nsSupportsArray::DeleteArray() line 305 nsSupportsArray::~nsSupportsArray() line 148 nsSupportsArray::`vector deleting destructor'(unsigned int 1) + 81 bytes nsSupportsArray::Release(nsSupportsArray * const 0x01073f08) line 239 + 148 bytes nsCOMPtr<nsISupportsArray>::~nsCOMPtr<nsISupportsArray>() line 491 nsObserverList::~nsObserverList() line 58 + 11 bytes nsObserverList::`scalar deleting destructor'(unsigned int 1) + 15 bytes ReleaseObserverList(nsHashKey * 0x01073ec0, void * 0x01039468, void * 0x00000000) line 109 + 28 bytes hashEnumerateRemove(PLDHashTable * 0x0027ca30, PLDHashEntryHdr * 0x010359ac, unsigned int 23, void * 0x0012fdd8) line 315 + 26 bytes PL_DHashTableEnumerate(PLDHashTable * 0x0027ca30, int (PLDHashTable *, PLDHashEntryHdr *, unsigned int, void *)* 0x100204d0 hashEnumerateRemove(PLDHashTable *, PLDHashEntryHdr *, unsigned int, void *), void * 0x0012fdd8) line 603 + 34 bytes nsHashtable::Reset(int (nsHashKey *, void *, void *)* 0x1001b160 ReleaseObserverList(nsHashKey *, void *, void *), void * 0x00000000) line 336 + 21 bytes nsObjectHashtable::Reset() line 875 nsObjectHashtable::~nsObjectHashtable() line 834 nsObjectHashtable::`vector deleting destructor'(unsigned int 1) + 81 bytes nsObserverService::~nsObserverService() line 84 + 33 bytes nsObserverService::`scalar deleting destructor'(unsigned int 1) + 15 bytes nsObserverService::Release(nsObserverService * const 0x00fc6e48) line 72 + 142 bytes nsCOMPtr_base::assign_assuming_AddRef(nsISupports * 0x00000000) line 436 nsCOMPtr_base::assign_with_AddRef(nsISupports * 0x00000000) line 74 nsCOMPtr<nsISupports>::operator=(nsISupports * 0x00000000) line 796 FreeServiceContractIDEntryEnumerate(PLDHashTable * 0x0027d7ac, PLDHashEntryHdr * 0x00febb64, unsigned int 847, void * 0x00000000) line 1926 PL_DHashTableEnumerate(PLDHashTable * 0x0027d7ac, int (PLDHashTable *, PLDHashEntryHdr *, unsigned int, void *)* 0x1006b1f0 FreeServiceContractIDEntryEnumerate(PLDHashTable *, PLDHashEntryHdr *, unsigned int, void *), void * 0x00000000) line 603 + 34 bytes nsComponentManagerImpl::FreeServices() line 1938 + 19 bytes NS_ShutdownXPCOM(nsIServiceManager * 0x00000000) line 723 main(int 1, char * * 0x002847a0) line 1911 + 8 bytes mainCRTStartup() line 338 + 17 bytes KERNEL32! 77e87903()

Assigned the bug to Julien.

I just checked this into the tip. Do we also want this fix on NSS_3_6_BRANCH ?

this is blocking automated (xpcshell) testing by multiple mozilla developers. please get it onto the mozilla branch whatever that may be.

I have run into such an assertion with an earlier snapshot of NSS (the NSS_CLIENT_TAG before it was switched to NSS 3.6.1 beta). However, in my most recent tests with profile switching and NSS 3.6.1 beta I did no longer see a crash. Anyway, if you think we are at risk this could be triggered in some circumstances, and if the change is safe, I'd lobby for landing in on the branch.

It appears that our release schedule has changed. Mozilla 1.3 will pick up 3.7, not 3.6.x . So this fix doesn't need to go in to the NSS_3_6_BRANCH .