From mboxrd@z Thu Jan  1 00:00:00 1970
Return-Path: <gcc-return-74287-listarch-gcc=gcc.gnu.org@gcc.gnu.org>
Received: (qmail 28297 invoked by alias); 14 May 2003 21:47:35 -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 28235 invoked from network); 14 May 2003 21:47:33 -0000
Received: from unknown (HELO crack.them.org) (146.82.138.56)
  by sources.redhat.com with SMTP; 14 May 2003 21:47:33 -0000
Received: from nevyn.them.org ([66.93.61.169] ident=mail)
	by crack.them.org with asmtp (Exim 3.12 #1 (Debian))
	id 19G45y-0001CQ-00; Wed, 14 May 2003 16:47:30 -0500
Received: from drow by nevyn.them.org with local (Exim 3.36 #1 (Debian))
	id 19G45U-00059f-00; Wed, 14 May 2003 17:47:00 -0400
Date: Wed, 14 May 2003 21:47:00 -0000
From: Daniel Jacobowitz <drow@mvista.com>
To: Gabriel Dos Reis <gdr@integrable-solutions.net>
Cc: Andreas Schwab <schwab@suse.de>, Brad Lucier <lucier@math.purdue.edu>,
	gcc@gcc.gnu.org
Subject: Re: strict-aliasing and typedefs
Message-ID: <20030514214700.GA19731@nevyn.them.org>
Mail-Followup-To: Gabriel Dos Reis <gdr@integrable-solutions.net>,
	Andreas Schwab <schwab@suse.de>,
	Brad Lucier <lucier@math.purdue.edu>, gcc@gcc.gnu.org
References: <200305142056.h4EKujuX025820@banach.math.purdue.edu> <jed6ilb4xo.fsf@sykes.suse.de> <m3ptmli5a2.fsf@uniton.integrable-solutions.net>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <m3ptmli5a2.fsf@uniton.integrable-solutions.net>
User-Agent: Mutt/1.5.1i
X-SW-Source: 2003-05/txt/msg01459.txt.bz2

On Wed, May 14, 2003 at 11:31:49PM +0200, Gabriel Dos Reis wrote:
> Andreas Schwab <schwab@suse.de> writes:
> 
> [...]
> 
> | int *sp;
> | int *hp;
> | 
> | |> I'm wondering if a Scheme->C compiler
> | |> can use typedefs and ISO C's aliasing rules to tell gcc that certain
> | |> memory locations cannot alias each other.
> | 
> | One way to do this is to define differently named structures with a single
> | member, but they may pessimize the generated code due to ABI
> | peculiarities.
> 
> As I understand it, he would like an object-based non-alias
> optimization, I believe restrict is a good candidate.

That depends.  Brad's initial example could also be used for type-based
non-alias optimization, for different sets of int objects - which is
somewhat looser than restrict allows.

-- 
Daniel Jacobowitz
MontaVista Software                         Debian GNU/Linux Developer