forwarded message from Niels Hallenberg

[Stephen Weeks]

Return-Path: <nh@nh.itu.dk>
Received: from localhost (localhost [127.0.0.1])
	by eponym.epr.com (8.11.2/8.11.2) with ESMTP id f69LrbJ04496
	for <sweeks@localhost>; Mon, 9 Jul 2001 14:53:37 -0700
Received: from exchange.epr.com [198.3.162.249]
	by localhost with IMAP (fetchmail-5.7.4)
	for sweeks@localhost (single-drop); Mon, 09 Jul 2001 14:53:37 -0700 (PDT)
Received: from maguro.epr.com ([198.3.162.27]) by exchange.epr.com with SMTP (Microsoft Exchange Internet Mail Service Version 5.5.2653.13)
	id NYYCXSSS; Mon, 9 Jul 2001 14:51:00 -0700
Received: from magrathea.epr.com (firewall-user@magrathea.epr.com [198.3.160.1])
	by maguro.epr.com (8.9.3/8.9.3) with ESMTP id OAA13687
	for <sweeks@intertrust.com>; Mon, 9 Jul 2001 14:52:57 -0700 (PDT)
Received: (from uucp@localhost) by magrathea.epr.com (8.9.3/8.7.3) id OAA18953 for <sweeks@intertrust.com>; Mon, 9 Jul 2001 14:52:56 -0700 (PDT)
Received: from nodnsquery(213.237.71.166) by magrathea.epr.com via smap (V5.5)
	id xma018761; Mon, 9 Jul 01 14:52:17 -0700
Received: (from nh@localhost)
	by nh.itu.dk (8.11.2/8.11.2) id f69Ln6F01527;
	Mon, 9 Jul 2001 23:49:06 +0200
X-Authentication-Warning: nh.itu.dk: nh set sender to nh@it-c.dk using -f
References: <15173.65053.727272.519257@eponym.epr.com>
	<lx1ynqv8wi.fsf@nh.itu.dk> <15177.60537.358762.508107@eponym.epr.com>
In-Reply-To: Stephen Weeks's message of "Mon, 9 Jul 2001 10:40:09 -0700"
Message-ID: <lxzoadojsd.fsf@nh.itu.dk>
Lines: 25
User-Agent: Gnus/5.0807 (Gnus v5.8.7) Emacs/20.7
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
From: Niels Hallenberg <nh@it-c.dk>
Sender: nh@nh.itu.dk
To: Stephen Weeks <sweeks@intertrust.com>
CC: nh@it-c.dk, mael@it-c.dk
Subject: Re: seg fault with List.tabulate in 3.9.1
Date: 09 Jul 2001 23:49:06 +0200

Hi Stephen,

> > The cure is to increase the stacksize with the limit and unlimit
> > command:
> 
> Thanks Niels.  That fixes the problem for me.  It would be nice, though, if the
> Kit could report something other than a segfault.  SML programmers aren't
> supposed to see such things :-).

sure. We actually discussed it again today. The problem, that we
haven't been able to solve, is that we can't distinguish a stack-size
segmentation fault from a real segmentation fault (i.e., a
segmentation fault due to bug in the ML Kit). A nice solution would be
to raise an STACK_SIZE_TOO_SMALL exception, but this is only possible
if Linux can tell us how the segmentation fault happened.

We are open to any suggestions about this problem.

-- Niels
-------------------------------------------------------
Niels Hallenberg                 Email: nh@it.edu
The IT University of Copenhagen  Tel.:  +45-38 16 88 24
Glentevej 67                     Fax:   +45-38 16 88 99
DK-2400 Copenhagen NV            WWW:   www.it.edu
-------------------------------------------------------