From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (qmail 18637 invoked by alias); 5 Mar 2003 06:15:24 -0000 Mailing-List: contact gdb-patches-help@sources.redhat.com; run by ezmlm Precedence: bulk List-Subscribe: List-Archive: List-Post: List-Help: , Sender: gdb-patches-owner@sources.redhat.com Received: (qmail 18629 invoked from network); 5 Mar 2003 06:15:23 -0000 Received: from unknown (HELO anchor-post-35.mail.demon.net) (194.217.242.85) by 172.16.49.205 with SMTP; 5 Mar 2003 06:15:23 -0000 Received: from tiptree.demon.co.uk ([158.152.73.173]) by anchor-post-35.mail.demon.net with esmtp (Exim 3.36 #2) id 18qSBW-0001Wi-0Z; Wed, 05 Mar 2003 06:15:22 +0000 Date: Wed, 05 Mar 2003 06:15:00 -0000 Subject: Re: [RFA] ObjC Testsuite Content-Type: text/plain; charset=US-ASCII; format=flowed Mime-Version: 1.0 (Apple Message framework v551) Cc: fedor@gnu.org, fnasser@redhat.com, msalter@redhat.com, ezannoni@redhat.com, gdb-patches@sources.redhat.com, Michael Elizabeth Chastain To: Andrew Cagney From: richard@tiptree.demon.co.uk In-Reply-To: <3E64BE1E.90407@redhat.com> Message-Id: Content-Transfer-Encoding: 7bit X-SW-Source: 2003-03/txt/msg00121.txt.bz2 On Tuesday, March 4, 2003, at 02:54 pm, Andrew Cagney wrote: > >> I was getting increasingly worried that, while stuff is being >> reviewed, none >> of it actually seems to be getting into CVS ... even one of the >> earliest patches >> to simply add the existing objc code into the configure/make process >> and >> activate it (which I think one reviewer said was a no-brainer or >> something >> similar) is still outstanding. > > Just FYI, enabling the code for the default build is one of the last > things to happen, not one of the first. Thanks for the explanation, but I don't find it wholly compelling ... > - gdb's language code isn't modula > There isn't a clean way of dropping in new languages (like many > things, a developer got half way through the process, and then > stopped. It has then remained in that state for ~10 years. A > --with-languages=... isn't there. I understand that, yet it is the case that most of the objective-c code has no effect except when debugging an objective-c program, so errors which were not caught by the review process would only effect objective-c developers, and frankly, for them it's probably better in the CVS version of gdb to have partial/unreliable objective-c support than no support at all. The patch I was referring to (Jan 3rd, 'Compile objc-lang.c objc-exp-tab.c') would seem to be safe. I would have thought that the *main* point of the review process was to ensure that any added code did not break existing code, so I assume that the objective-c support files in CVS at present are 'safe' (when enabled they certainly don't seem to break things and I don't see why they should), and having patches in place in CVS as rapidly as possible would seem to make it easier to review later patches ... but you may have found different in practice. > - this specific code came from [very old] a fork. > As such it needs sigificant work before it gets integrated into GDB. > Adam has been doing a briliant job. What's in GDB is noticably > different to what was in Apple's fork. Yes ... I did not intend to criticise/complain about Adams work ... I was just offering to help any way I can, and trying to encourage an efficient way of getting the work he's done into CVS quickly so he doesn't have to do yet more work modifying his patches to cope with changes in gdb. Anyway, please don't feel compelled to waste time responding to this ... unless you have a suggestion for something I can do to help of course.