From mboxrd@z Thu Jan  1 00:00:00 1970
Return-Path: <gcc-help-return-16324-listarch-gcc-help=gcc.gnu.org@gcc.gnu.org>
Received: (qmail 26054 invoked by alias); 15 Apr 2004 13:27:46 -0000
Mailing-List: contact gcc-help-help@gcc.gnu.org; run by ezmlm
Precedence: bulk
List-Archive: <http://gcc.gnu.org/ml/gcc-help/>
List-Post: <mailto:gcc-help@gcc.gnu.org>
List-Help: <mailto:gcc-help-help@gcc.gnu.org>
Sender: gcc-help-owner@gcc.gnu.org
Received: (qmail 26032 invoked from network); 15 Apr 2004 13:27:45 -0000
Received: from unknown (HELO mgr2.xmission.com) (198.60.22.202)
  by sources.redhat.com with SMTP; 15 Apr 2004 13:27:45 -0000
Received: from [198.60.22.201] (helo=mgr1.xmission.com)
	by mgr2.xmission.com with esmtp (Exim 3.35 #1)
	id 1BE6u8-0002jw-02; Thu, 15 Apr 2004 07:27:44 -0600
Received: from [198.60.22.20] (helo=xmission.xmission.com)
	by mgr1.xmission.com with esmtp (Exim 4.30)
	id 1BE6u8-0002wG-R4; Thu, 15 Apr 2004 07:27:44 -0600
Received: from llewelly by xmission.xmission.com with local (Exim 3.35 #1 (Debian))
	id 1BE6u8-00037d-00; Thu, 15 Apr 2004 07:27:44 -0600
To: "Hamilton, Ian" <Ian.Hamilton@gbr.xerox.com>
Cc: "'gcc-help@gcc.gnu.org'" <gcc-help@gcc.gnu.org>
References: <8EAC52A94CD8D411A01000805FBB37760615AFE3@gbrwgcms02.wgc.gbr.xerox.com>
From: llewelly@xmission.com
Date: Thu, 15 Apr 2004 13:27:00 -0000
In-Reply-To: <8EAC52A94CD8D411A01000805FBB37760615AFE3@gbrwgcms02.wgc.gbr.xerox.com>
Message-ID: <s3rk70h74mn.fsf@xmission.xmission.com>
User-Agent: Gnus/5.0808 (Gnus v5.8.8) Emacs/21.2
MIME-Version: 1.0
Subject: Re: Self-referring const array initialisation
Content-Type: text/plain; charset=us-ascii
X-Spam-Checker-Version: SpamAssassin 2.63 (2004-01-11) on mgr1.xmission.com
X-Spam-Level: 
X-Spam-Status: No, hits=0.3 required=8.0 tests=NO_REAL_NAME autolearn=no 
	version=2.63
X-SA-Exim-Mail-From: llewelly@xmission.com
X-SA-Exim-Version: 3.1 (built Wed Aug 20 09:38:54 PDT 2003)
X-SA-Exim-Scanned: Yes
X-SW-Source: 2004-04/txt/msg00183.txt.bz2

"Hamilton, Ian" <Ian.Hamilton@gbr.xerox.com> writes:

> Hi there
> 
> Here is a test program to illustrate a problem in some code
> I have inherited:
> 
> #include <stdio.h>
> 
> const int array1[5]=
> {
>   1,
>   2,
>   3,
>   4,
>   5
> };
> 
> const int array2[5]=
> {
>   1,
>   array2[0] + 1,
>   array2[1] + 1,
>   array2[2] + 1,
>   array2[3] + 1,
> };
> 
> int main()
> {
>   printf ("Array values:\n\n");
>   printf("array1  array2\n");
> 
>   printf("  %i       %i\n", array1[0], array2[0]);
>   printf("  %i       %i\n", array1[1], array2[1]);
>   printf("  %i       %i\n", array1[2], array2[2]);
>   printf("  %i       %i\n", array1[3], array2[3]);
>   printf("  %i       %i\n", array1[4], array2[4]);
> }
> 
> 
> If I build and run this on a Sun/Sparc/Solaris system using
> gcc version 2.95.2, I get the following output:
> 
> Array values:
> 
> array1  array2
>   1       1
>   2       2
>   3       3
>   4       4
>   5       5
> 
> 
> However, if I use gcc version 3.3.2, I get this output:
> 
> Array values:
> 
> array1  array2
>   1       1
>   2       1
>   3       1
>   4       1
>   5       1

FWIW, on i386-freebsd5.2, gcc 3.4 release cannidate gives the same
    output as gcc 2.95.2 . So I think this bug is fixed in 3.4 (but it is
    still present in 3.3.3). I can't find the bug report, however.

[snip]
> My question is whether this sort of self-reference is allowed by
> the C and/or C++ standards. In other words, is this a compiler bug,
> or a bug in the above code?

I am nearly certain C++ allows this. I can't find anything in 8.5.1
    that disallows it.