From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (qmail 17276 invoked by alias); 5 Feb 2008 19:16:48 -0000 Received: (qmail 17266 invoked by uid 22791); 5 Feb 2008 19:16:48 -0000 X-Spam-Check-By: sourceware.org Received: from fk-out-0910.google.com (HELO fk-out-0910.google.com) (209.85.128.186) by sourceware.org (qpsmtpd/0.31) with ESMTP; Tue, 05 Feb 2008 19:16:05 +0000 Received: by fk-out-0910.google.com with SMTP id 26so2460558fkx.10 for ; Tue, 05 Feb 2008 11:16:03 -0800 (PST) Received: by 10.82.162.14 with SMTP id k14mr16066574bue.32.1202238962885; Tue, 05 Feb 2008 11:16:02 -0800 (PST) Received: by 10.82.165.12 with HTTP; Tue, 5 Feb 2008 11:16:02 -0800 (PST) Message-ID: <8f2776cb0802051116g1545efe1ld0d34dc602a8dd22@mail.gmail.com> Date: Tue, 05 Feb 2008 19:16:00 -0000 From: "Jim Blandy" To: "andrzej zaborowski" Subject: Re: [PATCH] Disallow pseudo-registers in agent expression. Cc: gdb-patches@sources.redhat.com In-Reply-To: MIME-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 7bit Content-Disposition: inline References: <8f2776cb0801271007r125463e0la1cf8a7c9d1efef@mail.gmail.com> <8f2776cb0801281728v2ff40d50y2be92664b2c28de5@mail.gmail.com> <8f2776cb0802050734x1f79a335kfa3eaa104f25fedc@mail.gmail.com> X-Google-Sender-Auth: af728fdc15537c2c 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: 2008-02/txt/msg00112.txt.bz2 On Feb 5, 2008 10:22 AM, andrzej zaborowski wrote: > As far as I understand it's still a change that requires implementing > new bits in architecture code, because architectures don't know how to > make bytecode for pseudo-registers right now. That's right. The suggested patch I posted doesn't show this, but the way to do it would be to make it a gdbarch method with a predicate indicating whether the architecture provides the method or not (an 'M' entry in gdbarch.sh; search for "M -> " in there for docs), and then have ax-gdb.c check for the presence of the method, use it if present, and throw an error otherwise. This would allow us to implement the method only on the architectures on which some contributor was motivated to implement it. > Apart from little time I have zero knowledge about 95% of the > architectures in gdb :( Well, there are a lot of architectures. I think many of our contributors know only a few.