From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (qmail 19971 invoked by alias); 9 Jul 2002 03:53:10 -0000 Mailing-List: contact gdb-help@sources.redhat.com; run by ezmlm Precedence: bulk List-Subscribe: List-Archive: List-Post: List-Help: , Sender: gdb-owner@sources.redhat.com Received: (qmail 19901 invoked from network); 9 Jul 2002 03:53:06 -0000 Received: from unknown (HELO www.dberlin.org) (138.88.46.115) by sources.redhat.com with SMTP; 9 Jul 2002 03:53:06 -0000 Received: by www.dberlin.org (Postfix, from userid 503) id E7C41181BE81; Mon, 8 Jul 2002 23:53:05 -0400 (EDT) Received: from Daniel-Berlins-Computer (unknown [10.0.0.254]) by www.dberlin.org (Postfix) with ESMTP id 95A6B181BE80; Mon, 8 Jul 2002 23:52:59 -0400 (EDT) Date: Mon, 08 Jul 2002 20:53:00 -0000 Subject: Re: ``struct location'' Content-Type: text/plain; charset=US-ASCII; format=flowed Mime-Version: 1.0 (Apple Message framework v530) Cc: gdb@sources.redhat.com To: Andrew Cagney From: Daniel Berlin In-Reply-To: <3D2A51AE.3080102@ges.redhat.com> Message-Id: <5C273806-92EF-11D6-867A-000393575BCC@dberlin.org> Content-Transfer-Encoding: 7bit X-Spam-Status: No, hits=-1.6 required=5.0 tests=IN_REP_TO,NO_MX_FOR_FROM,AWL version=2.31 X-Spam-Level: X-SW-Source: 2002-07/txt/msg00089.txt.bz2 On Monday, July 8, 2002, at 10:59 PM, Andrew Cagney wrote: > Hello, > > This is a hunch. > > I think there should be a ``struct location'' object to go with the > ``struct type'' object. The ``struct value'' object would then be > made of, the TYPE, LOCATION and, once the value has been read from > memory, its VALUE. > > This would make it possible for location expression evaluators and > similar (e.g., unwind functions describing register locations) to > return a ``location'' without any need to refer to types. > > Having the location as an abstraction would also make it possible for > GDB's ``location'' code to evolve independantly of the ``struct > value'' - this is something we know we need since at present GDB can't > describe things like values with a location that spreads across > multiple registers and memory. > > thoughts? > Were you gone during April (seriously)? Check the mailing list archives. The whole loc_computed patch I submitted was the result of a discussion on how to do this. that's what the struct location_funcs are meant to do. " /* A structure of function pointers describing the location of a variable, structure member, or structure base class. " > Andrew > >