I've got a change coming in the next build which should reduce the chance of what caused it without it breaking anything new (am trying to get the current output time once & cache it instead of querying it every time it's needed which has been coming up in some other crash reports so a change was needed).
The driver aspect seems to have been due to the OS trying to close the hung process & how it was reacting so that's looking much less likely a problem to follow-up on.
-dro
