From fadushin@blackcat.cat.syr.edu  Tue May 25 23:25:30 1999
Received: from blackcat.cat.syr.edu (IDENT:root@blackcat.cat.syr.edu [128.230.59.12])
	by swi.swi.psy.uva.nl (8.9.3/8.9.3) with ESMTP id XAA02947
	for <prolog@swi.psy.uva.nl>; Tue, 25 May 1999 23:25:29 +0200 (MET DST)
Received: from blackcat.cat.syr.edu (blackcat.cat.syr.edu [128.230.59.12]) by blackcat.cat.syr.edu (8.7.4/8.7.3) with SMTP id QAA29199 for <prolog@swi.psy.uva.nl>; Tue, 25 May 1999 16:28:17 -0500
Date: Tue, 25 May 1999 17:28:16 -0400 (EDT)
From: Fred Dushin <fadushin@blackcat.cat.syr.edu>
Reply-To: fadushin@blackcat.cat.syr.edu
To: Recipients of SWI-Prolog Mailing List <prolog@swi.psy.uva.nl>
Subject: FATAL ERRORs
Message-ID: <Pine.LNX.4.00.9905251723090.28843-100000@blackcat.cat.syr.edu>
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII

Anyone have any ideas about how to diagnose FATAL ERRORs
(in this case, a segmentation violation) in SWI?

This is a non-interactive program (embedded), and
it's fairly small, though I admit it is being pressed
fairly hard into service (lots of queries to facts
set via assert in a threaded environment, though 
the queries are made sequentially).

The error is:

[FATAL ERROR:
        Received signal 11 (segv) while in 0-th garbage collection]

I have the -L flag set to 0 (-L0), though fiddling with -L and -G
don't seem to improve things.

Is SWI known to be robust under heavy loads?

--
Fred Dushin                      mailto:fadushin@syr.edu
PGP public key available at      http://blackcat.cat.syr.edu/~fadushin/pgp

