109815 - IMAP: threadsafety assertions on shutdown

Status: VERIFIED DUPLICATE of bug 123457

(MailNews Core :: Networking: IMAP, defect)

Description

[Navin Gupta]

Comment 1

[Navin Gupta]

###!!! ASSERTION: nsDeviceContextGTK not thread-safe: 'owningThread ==
NS_CurrentThread()', file /builds/cssorg/mozilla/xpcom/base/nsDebug.cpp,
line 528
###!!! Break: at file /builds/cssorg/mozilla/xpcom/base/nsDebug.cpp, line
528
nsDebug::Assertion(char const *, char const *, char const *,
int)+0x000000A9 [/builds/cssorg/obj/debug/dist/bin/libxpcom.so
+0x000FE689]
NS_CheckThreadSafe+0x000000A4
[/builds/cssorg/obj/debug/dist/bin/libxpcom.so +0x000FED10]
nsDeviceContextGTK::Release(void)+0x00000050
[/builds/cssorg/obj/debug/dist/bin/components/libgfx_gtk.so +0x0001FFA4]
nsCOMPtr<nsIDeviceContext>::assign_assuming_AddRef(nsIDeviceContext
*)+0x00000067 [/builds/cssorg/obj/debug/dist/bin/libgkgfx.so +0x000277C3]
nsCOMPtr<nsIDe
vic
eContext>::assign_with_AddRef(nsISupports *)+0x0000003C
[/builds/cssorg/obj/debug/dist/bin/libgkgfx.so +0x00027560]
nsCOMPtr<nsIDeviceContext>::operator=(nsIDeviceContext *)+0x00000021
[/builds/cssorg/obj/debug/dist/bin/libgkgfx.so +0x00027201]
nsWebShell::~nsWebShell(void)+0x000001FF
[/builds/cssorg/obj/debug/dist/bin/components/libdocshell.so +0x000554C3]
nsDocShell::Release(void)+0x0000009A
[/builds/cssorg/obj/debug/dist/bin/components/libdocshell.so +0x0003A87E]
nsWebShell::Release(void)+0x0000001F
[/builds/cssorg/obj/debug/dist/bin/components/libdocshell.so +0x0005569B]
nsCOMPtr<nsIInterfaceRequestor>::~nsCOMPtr(void)+0x0000004A [./mozilla-bin
+0x00019ACA]
nsSocketTransport::~nsSocketTransport(void)+0x0000026F
[/builds/cssorg/obj/debug/dist/bin/components/libnecko.so +0x000A2797]
nsSocketTransport::Release(void)+0x00000095
[/builds/cssorg/obj/debug/dist/bin/components/libnecko.so +0x000A4EED]
nsSocketR
equest:
:~nsSocketRequest(void)+0x0000003F
[/builds/cssorg/obj/debug/dist/bin/components/libnecko.so +0x000A8EFB]
nsSocketReadRequest::~nsSocketReadRequest(void)+0x0000006D
[/builds/cssorg/obj/debug/dist/bin/components/libnecko.so +0x000A954D]
nsSocketRequest::Release(void)+0x00000095
[/builds/cssorg/obj/debug/dist/bin/components/libnecko.so +0x000A8D09]
nsCOMPtr<nsIRequest>::~nsCOMPtr(void)+0x0000004A
[/builds/cssorg/obj/debug/dist/bin/components/libnecko.so +0x001136DE]
nsARequestObserverEvent::~nsARequestObserverEvent(void)+0x00000045
[/builds/cssorg/obj/debug/dist/bin/components/libnecko.so +0x0011C4AD]
nsOnStopRequestEvent::~nsOnStopRequestEvent(void)+0x0000006A
[/builds/cssorg/obj/debug/dist/bin/components/libnecko.so +0x0011C8E6]
nsARequestObserverEvent::DestroyPLEvent(PLEvent *)+0x00000066
[/builds/cssorg/obj/debug/dist/bin/components/libnecko.so +0x000A0422]
PL_DestroyEvent+0x00000076 [/builds/cssorg/obj/debug/dist/b
in/libxpcom
.so
+0x000EF37E]
PL_HandleEvent+0x000000BF [/builds/cssorg/obj/debug/dist/bin/libxpcom.so
+0x000EF2FF]
PL_ProcessPendingEvents+0x000000E9
[/builds/cssorg/obj/debug/dist/bin/libxpcom.so +0x000EF0AD]
nsEventQueueImpl::ProcessPendingEvents(void)+0x00000072
[/builds/cssorg/obj/debug/dist/bin/libxpcom.so +0x000F12DE]
nsImapProtocol::Run(void)+0x00000294
[/builds/cssorg/obj/debug/dist/bin/components/libmsgimap.so +0x000A755C]
nsThread::Main(void *)+0x000000C1
[/builds/cssorg/obj/debug/dist/bin/libxpcom.so +0x000F319D]
PR_Select+0x000004FE [/builds/cssorg/obj/debug/dist/bin/libnspr4.so
+0x0003070E]pthread_detach+0x000005CD [/lib/i686/libpthread.so.0
+0x00006BFD]

Comment 2

[Navin Gupta]

ok, it is not due to my checkin from last week. I think these are not new.

Comment 3

[David Baron :dbaron: (⌚️UTC-5, no longer working on Mozilla)]

Actually, now that I look more closely, the nsDocShell implements threadsafe
refcounting, which implies (or at least strongly suggests) it is usable from
multiple threads.  That would suggest that this is then a docshell bug, since
the nsWebShell shouldn't release the device context, etc., on a thread other
than the main thread.

Comment 4

[scottputterman]

does this need to be reassigned?

Comment 5

[David :Bienvenu]

this is a dup of a bug scott fixed

*** This bug has been marked as a duplicate of 123457 ***

Comment 6

[Karen Huang]

Verified as dup.