From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (qmail 9677 invoked by alias); 15 Sep 2004 19:57:07 -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 9657 invoked from network); 15 Sep 2004 19:57:06 -0000 Received: from unknown (HELO mx1.redhat.com) (66.187.233.31) by sourceware.org with SMTP; 15 Sep 2004 19:57:06 -0000 Received: from int-mx1.corp.redhat.com (int-mx1.corp.redhat.com [172.16.52.254]) by mx1.redhat.com (8.12.11/8.12.10) with ESMTP id i8FJv1NH006017 for ; Wed, 15 Sep 2004 15:57:01 -0400 Received: from localhost.redhat.com (porkchop.devel.redhat.com [172.16.58.2]) by int-mx1.corp.redhat.com (8.11.6/8.11.6) with ESMTP id i8FJunr14608; Wed, 15 Sep 2004 15:56:49 -0400 Received: from gnu.org (localhost [127.0.0.1]) by localhost.redhat.com (Postfix) with ESMTP id F149928D2; Wed, 15 Sep 2004 15:54:29 -0400 (EDT) Message-ID: <41489DF5.3020104@gnu.org> Date: Wed, 15 Sep 2004 19:57:00 -0000 From: Andrew Cagney User-Agent: Mozilla/5.0 (X11; U; NetBSD macppc; en-GB; rv:1.4.1) Gecko/20040831 MIME-Version: 1.0 To: Michael Chastain Cc: brobecker@gnat.com, gdb-patches@sources.redhat.com Subject: Re: [RFA/mips] mips32 frame code simplification... References: <20040915062611.GD1622@gnat.com> <414844E4.nail2UE22SZGZ@mindspring.com> <20040915162937.GU5843@gnat.com> <414879E3.nailB7Z1MJFFM@mindspring.com> In-Reply-To: <414879E3.nailB7Z1MJFFM@mindspring.com> Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit X-SW-Source: 2004-09/txt/msg00261.txt.bz2 > Joel Brobecker writes: > >>> I don't think that's the problem. I get an infinite loop inside GDB >>> itself (backtracing from a signal handler I think), and that causes >>> the expect/dejagnu buffer to get full. > > > How about: > > set backtrace limit 1000 > > That would look good in default_gdb_start and mi_gdb_start. Someone somewhere pointed out (with a patch) that set backtrace limit 1000 didn't actually work! I've lost the patch but am working on a tested fix. Now if only I could remember who posted the original patch. Andrew > However a backtrace limit of 1000 would still overflow the > 20000-character buffer. Also I've heard that "set backtrace limit" > has a bug these days. > > We could also change the individual "backtrace" and "bt" calls to > "backtrace 100". That's simpler and more effective, but it fixes > just one nit at a time. > > >>> And apparently, this just disables dejagnu for the rest of the run >>> (can't spawn the compiler nor the debugger anymore for instance). >>> Maybe there is a way of flushing that buffer when we do a gdb_exit? > > > I don't know enough about expect to get into that area. :( >