> Please note this came from Jim Seagrave who has been developing the
> RADIUS code at Demon Internet Ltd NOT myself. I am just posting this
> FYI so hopefully others wont be caught by this.
Thanks for passing it on - I'm playing with radius at present...
> > The fix I made with the Ascend code was to move the entire body of the
> > SIGCHLD handler out of the signal handler and into the main event
> > loop. SIGCHLD now is a simple minded handler:
The pd-ksh had the exact same problem a few years ago when I was
porting it from Sun3 to a Sparc2, the faster machine was unable to run
quite simple pipe lines due to too much being done in the SIGCHLD handler
and being interrupted part way though. But the shell then goes on to
wait for ever... my solution was much the same as Jim's.
--sjg [still months behind...]