From: Dan P. <ba...@al...> - 2006-06-12 05:23:05
|
On Jun 11, 2006, at 4:58 PM, Sam Steele wrote: > Aha! Having the input thread fire off while drawing / updating has > been driving me crazy! Thanks for putting this in (and pointing > out the existence of EventCollector). Yeah, I was scratching my head on this one for a bit. :) I was getting an assert for a usage of a null RefPtr, but it wasn't happening at a predictable point in the code when tracing through with the debugger. "It's like it's in a different thread or something..." ..... light bulb illuminates :) The amusing thing is that I had the exact same crash at basically the exact same point while running on the DC. How's that for consistency? 'Course that also led me to realize there's a race condition in the EventCollector, even with the mutex objects. That one's a tough nut to crack. We had a similar problem at work with DCOM callbacks happening after someone unregisters for them. I'll definitely take suggestions on that one if someone has an idea! |