Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Path: utzoo!watmath!clyde!burl!ulysses!allegra!mit-eddie!genrad!panda!talcott!harvard!seismo!brl-adm!brl-smoke!smoke!cottrell@NBS-VMS.ARPA
From: cottrell@NBS-VMS.ARPA (COTTRELL, JAMES)
Newsgroups: net.lang.c
Subject: Fast Morals & No Exit
Message-ID: <1407@brl-smoke.ARPA>
Date: Fri, 28-Feb-86 14:12:45 EST
Article-I.D.: brl-smok.1407
Posted: Fri Feb 28 14:12:45 1986
Date-Received: Mon, 3-Mar-86 00:33:53 EST
Sender: news@brl-smoke.ARPA
Lines: 28

/* >> is Me, > is Andrew Koenig <ark@alice.uucp>
> > The point is that the VMS C implementation must map its exit codes into
> > what the operating system expects. When the programmer says `exit 0', he
> > means `I succeeded'. The exit() function must support this.
> 
> No way!  exit(n) means "terminate and send the number n back to the system."

Referring to the infamous `April 30 1985' ANSI C Draft (I geuss the NIC
doesn't have their copy online yet :-), Section D.10.4.2, last three 
lines, it says, "Finally, control is returned to the environment. If the
value of status is zero, the status returned is successful termination;
otherwise an implementation-defined form of the status unsuccessful
termination is returned". Whew! Quoting that was worse than reading it.

I would be surprised if the standard has changed on this point.

The Point Is, if `successful termination' means rewinding all the
tapedrives, printing the American flag on the laser printer, or aborting
the space shuttle, when I say `exit(0)', THAT IS WHAT MUST HAPPEN!!!
As you well know, a function _exit(int status) exists to do just what 
you mention. On VMS, after cleanup, exit(int status) should call
_exit(!status) in order to conform. I want my scripts to work right!

You see, Doug, not everyone who disagrees with Andy is a turkey.

	jim		cottrell@nbs
*/
------