From mboxrd@z Thu Jan  1 00:00:00 1970
Return-Path: <sid-return-329-listarch-sid=sources.redhat.com@sources.redhat.com>
Received: (qmail 16099 invoked by alias); 18 Apr 2002 16:51:46 -0000
Mailing-List: contact sid-help@sources.redhat.com; run by ezmlm
Precedence: bulk
List-Subscribe: <mailto:sid-subscribe@sources.redhat.com>
List-Archive: <http://sources.redhat.com/ml/sid/>
List-Post: <mailto:sid@sources.redhat.com>
List-Help: <mailto:sid-help@sources.redhat.com>, <http://sources.redhat.com/lists.html#faqs>
Sender: sid-owner@sources.redhat.com
Received: (qmail 16061 invoked from network); 18 Apr 2002 16:51:44 -0000
Received: from unknown (HELO touchme.toronto.redhat.com) (216.138.202.10)
  by sources.redhat.com with SMTP; 18 Apr 2002 16:51:44 -0000
Received: from toenail.toronto.redhat.com (unknown [172.16.14.211])
	by touchme.toronto.redhat.com (Postfix) with ESMTP
	id 9A272B8013; Thu, 18 Apr 2002 12:51:43 -0400 (EDT)
Received: (from bje@localhost)
	by toenail.toronto.redhat.com (8.11.6/8.11.6) id g3IGphd08510;
	Thu, 18 Apr 2002 12:51:43 -0400
From: Ben Elliston <bje@redhat.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Message-ID: <15550.63903.473019.742005@toenail.toronto.redhat.com>
Date: Thu, 18 Apr 2002 09:51:00 -0000
To: Andrew Cagney <ac131313@cygnus.com>
Cc: sid@sources.redhat.com
Subject: Re: (gdb) target remote |foo-elf-sid --stdio
In-Reply-To: <3CBEBC77.6090205@cygnus.com>
References: <3CBEBC77.6090205@cygnus.com>
X-SW-Source: 2002-q2/txt/msg00007.txt.bz2

>>>>> "Andrew" == Andrew Cagney <ac131313@cygnus.com> writes:

  Andrew> GDB's remote command supports a pipe mode (it's supported it for some 
  Andrew> years now).  It lets GDB communicate with a sub-process using the remote 
  Andrew> protocol via that processes stdin/stdout.  The command syntax is:

  Andrew> 	(gdb) target remote |command to run with arguments

  Andrew> Can SID be modified to support this mode? It is easier to script then 
  Andrew> the current: start sid; start gdb; connect.

Yes.  In fact, I believe it was done once as a proof of concept.
Naturally, the standard input/output from a process emulation
component would need to be diverted to a Tk terminal window or some
other location so that it doesn't get interleaved with GDB protocol
chatter.

  Andrew> (PS: As always, there could be teething problems in GDB)

Agreed.

Ben