From owner-freebsd-hackers  Sat Feb 15 17:12:37 1997
Return-Path: <owner-hackers>
Received: (from root@localhost)
          by freefall.freebsd.org (8.8.5/8.8.5) id RAA00818
          for hackers-outgoing; Sat, 15 Feb 1997 17:12:37 -0800 (PST)
Received: from phaeton.artisoft.com (phaeton.Artisoft.COM [198.17.250.50])
          by freefall.freebsd.org (8.8.5/8.8.5) with SMTP id RAA00809
          for <hackers@FreeBSD.ORG>; Sat, 15 Feb 1997 17:12:31 -0800 (PST)
Received: (from terry@localhost) by phaeton.artisoft.com (8.6.11/8.6.9) id SAA06537; Sat, 15 Feb 1997 18:09:52 -0700
From: Terry Lambert <terry@lambert.org>
Message-Id: <199702160109.SAA06537@phaeton.artisoft.com>
Subject: Re: Sun Workshop compiler vs. GCC?
To: hamby@aris.jpl.nasa.gov (Jake Hamby)
Date: Sat, 15 Feb 1997 18:09:52 -0700 (MST)
Cc: davem@jenolan.rutgers.edu, asami@vader.cs.berkeley.edu,
        jmb@freefall.freebsd.org, hackers@FreeBSD.ORG
In-Reply-To: <Pine.GSO.3.95.970215155419.18262A-100000@aris> from "Jake Hamby" at Feb 15, 97 04:01:19 pm
X-Mailer: ELM [version 2.4 PL24]
MIME-Version: 1.0
Content-Type: text/plain; charset=US-ASCII
Content-Transfer-Encoding: 7bit
Sender: owner-hackers@FreeBSD.ORG
X-Loop: FreeBSD.org
Precedence: bulk

[ ... SunSoft C optimizer bug ... ]

> In my programming career, I've discovered one other optimizer bug, in
> Metrowerks Codewarrior (BeOS/PowerPC).  It was rather gratifying to see it
> fixed in the very next patch release, with my example code listed in the
> README under bugs fixed.  :)

Go to a platform using the "portable C compiler" (like Ultrix).

Compile a program that does:

char *
func2( p, len)
char	*p;
int	len;
{
	return( p + len);
}


char	*
func( param)
char	*param;
{
	register char *p = param;

	p = func2( p, 3);	/* XXX*/

	return( p);
}

main()
{
	char	*str = "long string";

	printf( "<%s>\n", str);

	str = func( str);
	printf( "<%s>\n", str);
}


What will happen is that the assignment of the return value will occur
before the pop, and both strings will print the same as the pop overwrites
the return value.  This is the "Berkeley pop order bug".

Now try compiling an unmodified Motif 1.x on Ultrix.  8-).


					Regards,
					Terry Lambert
					terry@lambert.org
---
Any opinions in this posting are my own and not those of my present
or previous employers.