-o. > > Yeah, it's gotta be that. Keep staring at code tyring to see how it could be > happening, so far )obviously) without success.
It would be great to know why it fails consistently on some people's machines but works fine on ours. But I don't know how to do that. What about debug calls right at various points in worker thread shutdown (.e.g called to clean up, completed cleanup and exiting) and then in dll shutdown to see if order of calls varies on different machines?
