If the stack is random memory addresses that don't map to any module in memory, you probably have some code that is overwriting the stack. Look for potential buffer overruns.
-----Original Message----- From: [email protected] [mailto:[email protected]] On Behalf Of [email protected] Sent: Friday, August 07, 2009 2:59 AM To: [email protected] Subject: [hlcoders] Awkward crash dumps I've been working on fixing the crashes introduced in the latest version of our mod. From mdmps, I've got all the obvious ones, but there's a few that I just don't know how to deal with. In fact, it seems to me that there's likely nothing that I CAN do, short of stripping all our features out & letting a small army of testers play with it for a week, then adding in features one at a time and repeating, until these occasional crashes are reintroduced. That sounds like a knightmare, and would likely take several months. Does anyone have any better suggestions on how to deal with: * A crash where the call stack ONLY shows a memory address? Not even any mention of a dll. This is our most common crash, and happens 1-2 times a day on each server. * One pointing to __RTDynamicCast in server.dll, with nothing to say WHICH cast caused issues? And if its a dynamic cast, then ... it shouldn't crash if it fails? This has only come up once, admittedly. * Another crash pointing at this line in Physics_RunThinkFunctions: CBaseEntity **list = (CBaseEntity **)stackalloc( sizeof(CBaseEntity *) * listMax ); I'd guess that could potentially be a stack overflow? If it is, any suggestions on tracking down what from? Its come up a couple of times now, on different servers. Note that we don't currently have a linux client, these are all windows crashes, in case that is relevant. Any suggestions or tips would be greatly appreciated - we're all ready to release an update, apart from these! -- This message was sent on behalf of [email protected] at openSubscriber.com http://www.opensubscriber.com/messages/[email protected]/topic .html _______________________________________________ To unsubscribe, edit your list preferences, or view the list archives, please visit: http://list.valvesoftware.com/mailman/listinfo/hlcoders _______________________________________________ To unsubscribe, edit your list preferences, or view the list archives, please visit: http://list.valvesoftware.com/mailman/listinfo/hlcoders
