Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Path: utzoo!mnetor!seismo!columbia!rutgers!clyde!cuae2!ihnp4!aicchi!mdb
From: mdb@aicchi.UUCP (Blackwell)
Newsgroups: comp.unix.wizards
Subject: Re: process term bug, sysIII (was text: table is full)
Message-ID: <868@aicchi.UUCP>
Date: Mon, 8-Dec-86 18:29:59 EST
Article-I.D.: aicchi.868
Posted: Mon Dec  8 18:29:59 1986
Date-Received: Tue, 9-Dec-86 06:16:21 EST
References: <867@aicchi.UUCP>
Reply-To: mdb@aicchi.UUCP (Blackwell)
Distribution: net
Organization: Analysts International Corp; Chicago Branch
Lines: 23
Keywords: oops, and more info

In article <867@aicchi.UUCP> mdb@aicchi.UUCP (*ME*) writes:
>
>The bug re zombies not going away exists also in our version of
     ^^^---oops

I shouldn't really have said BUG.  This is documented in wait(2) and
exit(2) and signal(2).  I can also see *why* it happens this way.  
However, it *is* an annoyance, and it *should* be stated in fork(2).
(I know, read the SEE ALSO's; but that would include the whole manual 
  (1/2 :-).  The correct way to fork(2) a process that you never want to
here from again is to first signal(SIGCHD,SIG_IGN) in the parent process.
This is in both the sysIII and SYSVr2 manuals.  The child-death signal
is handled differently in BSD, so I can't speak for that.....
			Mike Blackwell
			mdb%aicchi.UUCP@seismo.css.gov
			...ihnp4!aicchi!mdb

---
HIjol - yInqoq'e' val retlhbw'chu'taH
-- 
--
			Mike Blackwell
			..ihnp4!aicchi!mdb