From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (qmail 16219 invoked by alias); 12 Mar 2003 15:27:09 -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 16204 invoked from network); 12 Mar 2003 15:27:04 -0000 Received: from unknown (HELO aldan.algebra.com) (216.254.65.224) by 172.16.49.205 with SMTP; 12 Mar 2003 15:27:04 -0000 Received: from aldan.algebra.com (mi@localhost [127.0.0.1]) by aldan.algebra.com (8.12.7/8.12.7) with ESMTP id h2CFR24b086393 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=NO); Wed, 12 Mar 2003 10:27:02 -0500 (EST) (envelope-from mi@aldan.algebra.com) Received: from localhost (localhost [[UNIX: localhost]]) by aldan.algebra.com (8.12.7/8.12.7/Submit) id h2CFR16X086392; Wed, 12 Mar 2003 10:27:01 -0500 (EST) From: Mikhail Teterin To: Andrew Cagney Subject: Re: More on the Solaris 64 bit troubles Date: Wed, 12 Mar 2003 15:27:00 -0000 User-Agent: KMail/1.5 Cc: gdb@sources.redhat.com References: <200303112155.48503.mi+mx@aldan.algebra.com> <3E6F5066.6030405@redhat.com> In-Reply-To: <3E6F5066.6030405@redhat.com> MIME-Version: 1.0 Content-Type: text/plain; charset="koi8-u" Content-Transfer-Encoding: 7bit Content-Disposition: inline Message-Id: <200303121027.01420@aldan> X-SW-Source: 2003-03/txt/msg00178.txt.bz2 On Wednesday 12 March 2003 10:21 am, Andrew Cagney wrote: = > It happens when I try to use the 64-bit compiled gdb executable to debug = > a 32-bit executable. I don't see why this should not work... = > = > It dies, because when the second gdbarch structure is initialized = > (don't why it does it twice) the set_gdbarch_num_regs is not called and = > num_regs fields stays at the unitialized -1 :-( = = At present, on Solaris, the only way to debug a 32 bit executable is to = build a 32 bit GDB. GDB needs something (thread library?) and that only = works when built/linked 32 bit on a 32 bit executable. Thanks! I'll try that. Too bad, because all my Tcl (itcl, tix, etc.) are built for 64-bit, so I'll have to build a standalone insight-only distributions instead of sharing them. However, it seems to be dying long before getting to the thread library. While trying to use the second `struct gdbarch' -- why does it need more than one anyway? Also I have a bunch of patches for 5.3, which remove warnings raised by the Solaris' compiler (and a couple of bugs uncovered through them). Should I post them here, or send to someone directly? -mi