nsAppCoresManager:::Shutdown static casts each of the appcores in its list of active cores to an nsIDOMBaseAppCore, then calls Release() on the resulting pointer. Of all our appcores, only one of them (nsBaseAppCore) actually inherits from nsIDOMBaseAppCore. Oddly enough, Release() seems to be called correctly on Windows, at least for some appcores. Other platforms are not so ... forgiving. They seem to just hang up after getting a wrong number. How civilized. Still, seems like there's a problem here. I'm making this a bug, rather than just fixing it, because Scott's work to change appcores into components should neatly sidestep this problem. So I'm ignoring it for now, but making a note...
I'll take care of this as part of converting AppCores to components
move to later milestone?
this has moved to m8
Moving all Apprunner bugs past and present to Other component temporarily whilst don and I set correct component. Apprunner component will be deleted/retired shortly.
Moving to M9, when the rest of this change can be checked in.
Moving to M11 per today's bug triage.
following bug #13021, as usual
According to Bill Law in bug #13021, AppCores are no longer used (though they are still part of the build till he gets some makefile stuff checked in). Therefore, this bad code in |nsAppCoresManager::Shutdown| is now dead code and so no longer a problem.
According to developer's comments this bug is marked as verified. If you think the resolution is wrong please reopen.