From xemacs-m  Sun Nov 24 22:47:31 1996
Received: from deliverator.sgi.com (deliverator.sgi.com [204.94.214.10]) by xemacs.cs.uiuc.edu (8.7.6/8.7.3) with SMTP id WAA02118 for <xemacs-beta@xemacs.org>; Sun, 24 Nov 1996 22:47:30 -0600 (CST)
Received: from odin.corp.sgi.com by deliverator.sgi.com via ESMTP (950413.SGI.8.6.12/951211.SGI.AUTO)
	 id UAA22640; Sun, 24 Nov 1996 20:47:30 -0800
Received: from cardboard.mti.sgi.com by odin.corp.sgi.com via ESMTP (951211.SGI.8.6.12.PATCH1502/951211.SGI)
	 id UAA05806; Sun, 24 Nov 1996 20:47:30 -0800
Received: from litotes.mti.sgi.com by cardboard.mti.sgi.com via ESMTP (950413.SGI.8.6.12/930416.SGI.AUTO-MTI)
	 id UAA25600; Sun, 24 Nov 1996 20:47:57 -0800
Received: by litotes.mti.sgi.com (950413.SGI.8.6.12/930416.SGI.AUTO-MTI)
	 id UAA28459; Sun, 24 Nov 1996 20:47:32 -0800
To: turner@lanl.gov
Cc: xemacs-beta@xemacs.org
Subject: Re: 19.14 on Irix 6.2
References: <199611220616.XAA19851@branagh.lanl.gov>
From: Rick Braumoeller <rickb@litotes.mti.sgi.com>
Date: 24 Nov 1996 20:47:32 -0800
In-Reply-To: John Turner's message of Thu, 21 Nov 1996 23:16:27 -0700
Message-ID: <wl7g21yu6fv.fsf@litotes.mti.sgi.com>
Lines: 40
X-Mailer: Gnus v5.2.25/XEmacs 19.14

John Turner <turner@xdiv.lanl.gov> writes:

   A colleague has been trying to build 19.14 on Irix 6.2 for both
   an R4400 machine and an R10k box.

   In the following, note that:

   terre is a 4x200MHz R10k box
   ciel is a 150MHz R4400 box

   He reports:

   ------- begin -------

   I built xemacs on terre.  No, I can not get a binary that works on
   both terre and and the r4400 irix 6.2 machines.  Most programs are
   no problem, but xemacs just will not work on both.

   The binary I build on ciel runs on the r4400 machines, but drops
   core on terre.  The binary I build on terre drops core on an r4400
   machine.  I know that's not much to go on.  Ever seen similar
   behavior?  -32 -mips2 was used, although I've tried other
   combinations as well.

   ------- end -------

   I know there are some SGI folks on xemacs-beta, so I waas wondering
   if any of you know anything about this.

This has been on my rainy-day (or couple of rainy-days) list for a
while. I suspect that some DSO-specific offset information is leaking
into the undumped xemacs, and that some (but not all) changes to
libc.so are provoking the core dumps. I spent a few minutes talking
with someone in our compiler group about it, but I need to investigate
more to understand the problem.

If somebody knows which bogons are slipping into the undumped binary,
that might help. Or, next time our file server crashes, I'll dig in.

  -- Rick

