Path: utzoo!utgpu!news-server.csri.toronto.edu!cs.utexas.edu!swrinde!zaphod.mps.ohio-state.edu!wuarchive!udel!haven!mimsy!chris
From: chris@mimsy.umd.edu (Chris Torek)
Newsgroups: comp.unix.questions
Subject: Re: How to get Status of processes in C environment
Message-ID: <26503@mimsy.umd.edu>
Date: 12 Sep 90 07:30:38 GMT
References: <25467.26e7ac6a@kuhub.cc.ukans.edu> <4049@auspex.auspex.com>
Organization: U of Maryland, Dept. of Computer Science, Coll. Pk., MD 20742
Lines: 23

In article <4049@auspex.auspex.com> guy@auspex.auspex.com (Guy Harris) writes:
>If you just ignore "union wait" (and the "lint" complaints that doing so
>will give you), and treat the "status" argument to "wait()" as an "int
>*", even on BSD systems, your code will still work (unless the
>implementor of your UNIX system screwed up badly), but will be more
>portable and won't be polluted with #ifdefs. 

Right.

Note, in 4.3BSD-reno the argument to wait (and wait3) has been changed
back to `int *'.  The W* macros in <sys/wait.h> operate on `int' lvalues
(and continue to work on `union wait's by the bletcherous hack of
defining the W* macros as, e.g.,

	#define _W_INT(w)	(*(int *)&(w))
	#define	WEXITSTATUS(x)	(_W_INT(x) >> 8)

The funny _W_INT definition WILL GO AWAY someday, along with `union wait'.)
If compiling with the (semi-nonexistent) `POSIX compiler', the W* macros
operate on int expressions (because _W_INT(w) is defined is (w)).
-- 
In-Real-Life: Chris Torek, Univ of MD Comp Sci Dept (+1 301 405 2750)
Domain:	chris@cs.umd.edu	Path:	uunet!mimsy!chris