From owner-cvs-gnu  Sat Sep 20 23:00:15 1997
Return-Path: <owner-cvs-gnu>
Received: (from root@localhost)
          by hub.freebsd.org (8.8.7/8.8.7) id XAA15597
          for cvs-gnu-outgoing; Sat, 20 Sep 1997 23:00:15 -0700 (PDT)
Received: from godzilla.zeta.org.au (godzilla.zeta.org.au [203.2.228.19])
          by hub.freebsd.org (8.8.7/8.8.7) with ESMTP id XAA15579;
          Sat, 20 Sep 1997 23:00:04 -0700 (PDT)
Received: (from bde@localhost) by godzilla.zeta.org.au (8.8.5/8.6.9) id PAA20728; Sun, 21 Sep 1997 15:53:28 +1000
Date: Sun, 21 Sep 1997 15:53:28 +1000
From: Bruce Evans <bde@zeta.org.au>
Message-Id: <199709210553.PAA20728@godzilla.zeta.org.au>
To: mike@smith.net.au, perhaps@yes.no
Subject: Re: cvs commit: src/gnu/usr.bin/as/config atof-ieee.c src/gnu/usr.bin/ld/rtld rtld.c src/gnu/usr.bin/man/man man.c
Cc: cvs-all@FreeBSD.ORG, cvs-committers@FreeBSD.ORG, cvs-gnu@FreeBSD.ORG,
        jonny@coppe.ufrj.br, phk@FreeBSD.ORG
Sender: owner-cvs-gnu@FreeBSD.ORG
X-Loop: FreeBSD.org
Precedence: bulk

><C LAWYER>((void*)0) is a correct definition of NULL.  The other
                                                            ^^^^^another
>correct definition is 0.</C LAWYER>

Other correct definitions:

	0L		/* This has a different type than plain 0.
			 * It may be useful for hiding bugs on machines
			 * where sizeof(long) == sizeof(void *) &&
			 * sizeof(int) < sizeof(void *), and for finding
			 * bugs on machines where sizeof(long) >
			 * sizeof(void *).
			 */
	0UL		/* A weirder type.  This may be useful for
			 * finding misuses of NULL in integer context
			 */
	(2 + 2 != 4)	/* A gratuitously complicated constant integral
			 * expression with value 0.  This definition is
			 * not useful.
			 */
	(sizeof(char) - (int)1.1)
			/* A gratuitously complicated not-obviously
			 * constant integral expression with value 0.
			 * This definition is not useful.
			 */

	Any of the above cast to `void *'.
			/* Such definitions are just silly ways of writing
			 * ((void *)0), since they all have the same type
			 * and value.
			 */

Other not-quite correct definitions:

	0LL		/* Will probably be correct in C9X, since C9X has
			 * long long (blech), and may already be correct
			 * in Gnu C.  This may be useful for hiding bugs
			 * on machines where sizeof(long long) ==
			 * sizeof(void *) && sizeof(long) < sizeof(void *),
			 * and for finding bugs on machines where
			 * sizeof(long long) > * sizeof(void *).
			 */

Bruce