> Does this sound as if it could work? Can anyone see any
> obvious gotchas?

Ok, I'm running an iBCS2 shell, I type 'ps' and run the native
ps, then type kill <pid> and the shell will kill the wrong process.

You would certainly need to ensure the mapped pids are otherwise
unallocated.

OTOH how important is iBCS2?
Would restricting kernels that support it to 32k processes
actually be a problem?

(I started this thread by suggesting that NO_PID be changed
to allow pids greater than the compile-time MAX_PID to be
safely allocated.)

	David

-- 
David Laight: david@l8s.co.uk