bob coyne <[EMAIL PROTECTED]> writes:

> If I call (sleep .1) in the main listener, it sleeps for
> the correct amount of time.  If I call it from a separate
> process, it sleeps for a second.   Nothing else is running.
> Any explanations for this?

The whole multiprocessing stuff is a bit strange.  The implementation
assumes that the "main process" does the scheduling and is otherwise
idle.  mp::startup-idle-and-top-level-loops does the proper setup. I
don't know why this is not started by default.  Perhaps for efficiency
reasons (thread switching is rather inefficient in CMUCL) or because
the debugger doesn't know much about threads.  I've heard SBCL has
native threads, maybe this serves your needs better.

--helmut

Reply via email to