> I wonder if utrace could be more helpful here. If an engine just > wants to single-step from some other callback, forcing it to have a > quiesce callback just to re-re-re-repeat what it already asked for > seems like make-work. Could utrace instead remember the last > instruction received either via utrace_control() or from a real > callback, and have utrace supply a dummy echo-bot quiesce?
That is essentially like the old API where these were sticky flag bits. The point is that if other engines are doing other things, you really need to see what they are doing and react.
