> Correct me if I'm wrong, but wouldn't Spruce fall under the category of
> event-driven applications that *could* benefit from the non-preemptive
> scheduling model used by Pth? (if Jeff decided to design it to that model)

Yes, actually, he could drive pth using the socket io and gui input.  I
didn't mean
to say it couldn't be done, just that using Pth as a pthread compat layer
usually
doesn't work since apps that use pthreads assume that preemptive switching
is going on.  

> Arrrrgh.  Well, so much for that idea.  *sigh*  And here I thought those GNU
> guys generally had thier sh*t together...

Generally, yea, and in this case too.  Sorry, I didn't mean to imply that
Pth was broken, it's actually a (one of the) good solution(s) to the
portability problem.

Ezra