I have experienced the same thing.

Because Windows tends to be an event-driven environment, and we are used to jumping from task to task randomly, that can cause problems if a procedure is programmed in a way that requires a specific exit in order to completely turn off the procedure and reinitialize the normal state of things.

When it happens to me, I always figured that I inadvertently jumped to something else without completing all the requirements for finishing what I had started.

The fact that some of us experience certain problems repeatedly while others never experience them at all leads me to believe that things like this come from problematic navigation habits.

This may not be the case in RB, but I know its a possibility, because more than one of the systems where I work require specific conclusions to procedures, or problems start happening mysteriously.

Exiting and coming back into the program is the easy and guaranteed way to reinitialize everything. Identifying and correcting the exact navigation habit that causes the problem is probably going to be a bit more time consuming