From bug 314199: Steps to reproduce: (1) Start SeaMonkey trunk to Navigator (2) From Navigator, launch Composer. (3) Focus on Navigator. (Do not right-click to go to Close directly for the next step.) (4) Close Navigator. (5) Without letting focus get back to the Composer window, close Composer. (To do this in Windows, right-click on the taskbar item for Composer and click Close.) This testcase is 100% reproducible! 00 ntdll!DbgBreakPoint 01 xpcom_core!nsDebugImpl::Break(char * aFile = 0x01a8b4b4 "m:/mozilla/content/base/src/nsGenericElement.cpp", int aLine = 838)+0x7e [m:\mozilla\xpcom\base\nsdebugimpl.cpp @ 350] 02 xpcom_core!nsDebugImpl::Assertion(char * aStr = 0x01a8b518 "nsGenericElement's event listener manager hash not empty at shutdown!", char * aExpr = 0x01a8b4e8 "sEventListenerManagersHash.entryCount == 0", char * aFile = 0x01a8b4b4 "m:/mozilla/content/base/src/nsGenericElement.cpp", int aLine = 838)+0x29e [m:\mozilla\xpcom\base\nsdebugimpl.cpp @ 266] 03 xpcom_core!NSGlue_Assertion(char * aStr = 0x01a8b518 "nsGenericElement's event listener manager hash not empty at shutdown!", char * aExpr = 0x01a8b4e8 "sEventListenerManagersHash.entryCount == 0", char * aFile = 0x01a8b4b4 "m:/mozilla/content/base/src/nsGenericElement.cpp", int aLine = 838)+0x55 [m:\builds\mozdebug\xpcom\build\nsdebug.cpp @ 111] 04 gklayout!nsGenericElement::Shutdown(void)+0xa1 [m:\mozilla\content\base\src\nsgenericelement.cpp @ 838] 05 gklayout!Shutdown(void)+0x56 [m:\mozilla\layout\build\nslayoutmodule.cpp @ 399] 06 gklayout!LayoutShutdownObserver::Observe(class nsISupports * aSubject = 0x002eb3ac, char * aTopic = 0x100d3858 "xpcom-shutdown", unsigned short * someData = 0x00000000)+0x1d [m:\mozilla\layout\build\nslayoutmodule.cpp @ 266] 07 xpcom_core!nsObserverService::NotifyObservers(class nsISupports * aSubject = 0x002eb3ac, char * aTopic = 0x100d3858 "xpcom-shutdown", unsigned short * someData = 0x00000000)+0x1a5 [m:\mozilla\xpcom\ds\nsobserverservice.cpp @ 235] 08 xpcom_core!NS_ShutdownXPCOM_P(class nsIServiceManager * servMgr = 0x00000000)+0x9b [m:\mozilla\xpcom\build\nsxpcominit.cpp @ 808] 09 seamonkey!NS_ShutdownXPCOM(class nsIServiceManager * servMgr = 0x00000000)+0x1d [m:\mozilla\xpcom\glue\standalone\nsxpcomglue.cpp @ 171] 0a seamonkey!GRE_Shutdown(void)+0xa [m:\mozilla\xpcom\glue\standalone\nsxpcomglue.cpp @ 515] 0b seamonkey!main(int argc = 1, char ** argv = 0x002e40d8)+0x1b3 [m:\mozilla\xpfe\bootstrap\nsapprunner.cpp @ 1745] 0c seamonkey!mainCRTStartup(void)+0x12c [f:\vs70builds\3077\vc\crtbld\crt\src\crtexe.c @ 398] 0d kernel32!BaseProcessStart+0x23 cc'ing bzbarsky at his request.

I can't reproduce this reliably... if you can, would you mind doing some leak debugging per the mozilla.org docs? I can help as needed.

Immediately following dbaron's fix for bug 241518 (which was for the same assertion), I was unable to reproduce this bug at all. Marking as a DUPE. *** This bug has been marked as a duplicate of 241518 ***