From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (qmail 21156 invoked by alias); 29 May 2008 15:41:21 -0000 Received: (qmail 21091 invoked by uid 22791); 29 May 2008 15:41:20 -0000 X-Spam-Check-By: sourceware.org Received: from NaN.false.org (HELO nan.false.org) (208.75.86.248) by sourceware.org (qpsmtpd/0.31) with ESMTP; Thu, 29 May 2008 15:41:00 +0000 Received: from nan.false.org (localhost [127.0.0.1]) by nan.false.org (Postfix) with ESMTP id C60D2983D6; Thu, 29 May 2008 15:40:58 +0000 (GMT) Received: from caradoc.them.org (22.svnf5.xdsl.nauticom.net [209.195.183.55]) by nan.false.org (Postfix) with ESMTP id 8AAEB9810F; Thu, 29 May 2008 15:40:58 +0000 (GMT) Received: from drow by caradoc.them.org with local (Exim 4.69) (envelope-from ) id 1K1kFN-00028g-MH; Thu, 29 May 2008 11:40:57 -0400 Date: Thu, 29 May 2008 21:33:00 -0000 From: Daniel Jacobowitz To: Ulrich Weigand Cc: Andreas Schwab , gdb-patches@sourceware.org Subject: Re: Fix frame-args.exp failure Message-ID: <20080529154057.GA7904@caradoc.them.org> Mail-Followup-To: Ulrich Weigand , Andreas Schwab , gdb-patches@sourceware.org References: <20080527183701.GA21155@caradoc.them.org> <200805291200.m4TC0VYx011880@d12av02.megacenter.de.ibm.com> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <200805291200.m4TC0VYx011880@d12av02.megacenter.de.ibm.com> User-Agent: Mutt/1.5.17 (2008-05-11) X-IsSubscribed: yes Mailing-List: contact gdb-patches-help@sourceware.org; run by ezmlm Precedence: bulk List-Id: List-Subscribe: List-Archive: List-Post: List-Help: , Sender: gdb-patches-owner@sourceware.org X-SW-Source: 2008-05/txt/msg00757.txt.bz2 On Thu, May 29, 2008 at 02:00:31PM +0200, Ulrich Weigand wrote: > Now that this is in, I'm wondering whether we should also add a > SYMBOL_IS_REFERENCE flag (or maybe make SYMBOL_IS_ARGUMENT triple- > valued) to indicate that the location described by address_class > holds a pointer to value instead of the value itself. > > Then we could get rid of LOC_REF_ARG and LOC_REGPARM_ADDR as well ... Making it triple-valued would not be nice, since at least in theory the two concepts are orthogonal. I'd rather see every special case move to LOC_COMPUTED. For space savings, Andrew suggested using the address class as an index into an array of symbol operations. -- Daniel Jacobowitz CodeSourcery