From mboxrd@z Thu Jan  1 00:00:00 1970
Return-Path: <gcc-return-83597-listarch-gcc=gcc.gnu.org@gcc.gnu.org>
Received: (qmail 28638 invoked by alias); 6 Oct 2003 12:46:48 -0000
Mailing-List: contact gcc-help@gcc.gnu.org; run by ezmlm
Precedence: bulk
List-Archive: <http://gcc.gnu.org/ml/gcc/>
List-Post: <mailto:gcc@gcc.gnu.org>
List-Help: <http://gcc.gnu.org/ml/>
Sender: gcc-owner@gcc.gnu.org
Received: (qmail 28629 invoked from network); 6 Oct 2003 12:46:46 -0000
Received: from unknown (HELO mx3.informatik.uni-tuebingen.de) (134.2.12.26)
  by sources.redhat.com with SMTP; 6 Oct 2003 12:46:46 -0000
Received: from juist (semeai [134.2.15.66])
	by mx3.informatik.uni-tuebingen.de (Postfix) with ESMTP id 7CF10141
	for <gcc@gcc.gnu.org>; Mon,  6 Oct 2003 14:46:45 +0200 (DFT)
Received: from falk by juist with local (Exim 3.36 #1 (Debian))
	id 1A6UlA-0006It-00
	for <gcc@gcc.gnu.org>; Mon, 06 Oct 2003 14:46:44 +0200
To: gcc@gcc.gnu.org
Subject: Suggested warning: "negating an expression of unsigned type does not yield a negative value"
From: Falk Hueffner <falk.hueffner@student.uni-tuebingen.de>
Date: Mon, 06 Oct 2003 12:46:00 -0000
Message-ID: <87zngev8i4.fsf@student.uni-tuebingen.de>
User-Agent: Gnus/5.0808 (Gnus v5.8.8) XEmacs/21.5 (cabbage)
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
X-SW-Source: 2003-10/txt/msg00152.txt.bz2

Hi,

I just found yet another bug of the kind:

int f (int *p, unsigned x) { return p[-x]; }

which only manifests on 64 bit platforms, because most (all?)
platforms have wrapping address arithmetic.

So I was wondering about a general warning about negating unsigned
values, since I couldn't really think of a legitimate application. A
quick check with the gcc source turned up:

gengtype-lex.l:
  char *namestart;
  size_t namelen;
  [...]
  for (namelen = 1; !ISSPACE (namestart[-namelen]); namelen++)

This looks actually invalid to me, although it will probably work
everywhere.

In fold_const.c, there's

case RSHIFT_EXPR:
      int2l = -int2l;

also "invalid but works" since it's later passed to a function taking
int.

Then there's everybody's favourite idiom "x &= -x", but it can be
expressed clearer as "x &= ~x + 1".

Then there's constant folding in neg_double. Hm. Damn. I can't think
of any reformulation which does not obscure the code. So this warning
should probably not be turned on by -W. But it seems generally useful.

Any opinions?

-- 
	Falk