>>What isn't in the call stack is interesting, there's not a reference to ExitBtn1.Click nor another to QueryUnload(). It's as if there's no user intervention at all. That's another thing that makes me wonder if the Windows shutdown sequence wasn't at work here.
>
>Thats a normal thing. Which is what I was explaining earlier (or maybe I didn't post that), wether you called Release() or the user clicked the "X" button won't show up, because after you call release(), it will also do any code after Release() in the whole call stack up until where the form was called before the form.
>
>Then it will internally do Destroy() so the call stack in Destroy() would be blank. But using ReleaseType, you should be able to tell how it was killed.

Mike,

I'm not so sure that this would apply in an error handling sequence. If for example, you did the trace in the form's Unload event, You'd only see the main program and unload without anything else. What I was getting at was that there's no indication of what called Destroy, which is present in the call stack and normally wouldn't be. It doesn't come from the main that's for sure. I tried yesterday to do a shutdown with the app running only to have the machine lock down. I could logoff without a problem. I was doing a lot of stuff yesterday, with different servers and other stuff. Something may have been futzed there. I'm going to try again in a little while.