From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (qmail 10814 invoked by alias); 15 May 2006 12:42:23 -0000 Received: (qmail 10734 invoked by alias); 15 May 2006 12:42:16 -0000 Date: Mon, 15 May 2006 12:42:00 -0000 Message-ID: <20060515124216.10733.qmail@sourceware.org> X-Bugzilla-Reason: CC References: Subject: [Bug tree-optimization/25737] [4.1/4.2 Regression] ACATS c974001 c974013 hang with struct aliasing In-Reply-To: Reply-To: gcc-bugzilla@gcc.gnu.org To: gcc-bugs@gcc.gnu.org From: "dberlin at dberlin dot org" Mailing-List: contact gcc-bugs-help@gcc.gnu.org; run by ezmlm Precedence: bulk List-Archive: List-Post: List-Help: Sender: gcc-bugs-owner@gcc.gnu.org X-SW-Source: 2006-05/txt/msg01453.txt.bz2 List-Id: ------- Comment #26 from dberlin at gcc dot gnu dot org 2006-05-15 12:42 ------- Subject: Re: [4.1/4.2 Regression] ACATS c974001 c974013 hang with struct aliasing ebotcazou at gcc dot gnu dot org wrote: > ------- Comment #25 from ebotcazou at gcc dot gnu dot org 2006-05-15 05:36 ------- >> (I've not fully understood the discussion here, but if everything in Ada is >> addressable, then it seems to me that everything should have TREE_ADDRESSABLE >> set.) > > Grumpf... you read a little too quickly, that is exactly the opposite. The > struct aliasing machinery essentially assumes that everything is addressable > (modulo bit-fields) whereas *nothing* is addressable by default in Ada. > Read comment 20 again. If nothing was addressable, and we relied on this, your problem would get worse, *not better*. It will cause *strictly more* things to be marked unaliased, and your problem is that this really is aliasable. -- http://gcc.gnu.org/bugzilla/show_bug.cgi?id=25737