From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (qmail 26305 invoked by alias); 27 Sep 2004 07:31:50 -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 26296 invoked from network); 27 Sep 2004 07:31:49 -0000 Received: from unknown (HELO smtp.hispeed.ch) (62.2.95.247) by sourceware.org with SMTP; 27 Sep 2004 07:31:49 -0000 Received: from indel.ch (217-162-27-127.dclient.hispeed.ch [217.162.27.127]) by smtp.hispeed.ch (8.12.6/8.12.6/tornado-1.0) with SMTP id i8R7VlfW004011 for ; Mon, 27 Sep 2004 09:31:48 +0200 Received: from fabi.indel.ch [192.168.1.19] by indel.ch [127.0.0.1] with SMTP (MDaemon.v2.7.SP5.R) for ; Mon, 27 Sep 2004 09:30:50 +0200 Message-Id: <5.2.0.9.1.20040927092234.01d6ebf0@NT_SERVER> X-Sender: cenedese@NT_SERVER (Unverified) Date: Mon, 27 Sep 2004 07:31:00 -0000 To: gdb@sources.redhat.com From: Fabian Cenedese Subject: Re: Unimplemented MI commands In-Reply-To: <20040923114602.GA408@white> References: <5.2.0.9.1.20040923113200.01d1f670@NT_SERVER> <5.2.0.9.1.20040923093618.01d3f068@NT_SERVER> <5.2.0.9.1.20040923113200.01d1f670@NT_SERVER> Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii" X-MDaemon-Deliver-To: gdb@sources.redhat.com X-Return-Path: cenedese@indel.ch X-SW-Source: 2004-09/txt/msg00235.txt.bz2 >> >I've seen in mi/mi-cmds.c that about half of the mi commands are not >> >implemented yet. Are there plans for them? Or do I need to use the >> >normal console interface for these commands? I'm especially >> >interested in the -symbol-* functions where only one is available now. >> >> Replying to myself. >> >> I've found this mail with a patch, why wasn't that included then? >> >> http://sources.redhat.com/ml/gdb-patches/2002-09/msg00302.html > >Yes, only half of the functions are implemented. Not speaking for the >GDB people, but only for myself, I don't think they have any intention >on implementing the rest of the commands. However, if you need a certain >command maybe it would be done for you. However, I think that if you >need certain commands, it would be much better to add an mi command, >rather than use the CLI output. I have started at updating the above patch to my 6.1 version (I guess the actual cvs-gdb won't be much different in this regard). I already mailed the original author about this patch but didn't get a reply yet. If it's accepted I can send a patch with the updated source to be included. But I have no idea about testcases, cleanup handling etc. There sure will be some not-yet-right stuff in it which I would need help with. While thinking about it I wondered why the MI functions need to have an implementation of their own. Couldn't gdb be changed that the CLI and the MI functions just consist of outputting the results in their respecting format and use the same sub function to get the results? Like that the behaviour would be consistent between CLI and MI. Errors would be fixed for both simultaneously. And a new functionality would have to be implemented only once. I think that should be the next task after you have unified the MI. That would also get rid of some code and complete the now quite unimplemented MI at the same time. bye Fabi