Some questions

[Doug Gwyn]

In article <4919 at muffin.cme.nist.gov> libes at cme.nist.gov (Don Libes) writes:
>Guy Harris has informed me that I was wrong about signal(SIGCHLD,SIG_IGN)
>reaping child processes automatically on SunOS.  He is correct.
>I had the unfortunate luck to be calling a subroutine that someone
>else wrote which did the old wait in a loop trick, making me believe
>that this long-standing behavior had been changed on my system.

I can't tell from the sparse information, but if you were using the
System V environment it is possible you were getting an emulation of
the System V SIGCLD behavior.  I know it was present in my original
System V emulation for 4BSD, even though I personally think it is a
horrible design and that the originator of that kludge in UNIX System
V should be hanged by its thumbs.