From xemacs-m  Thu Feb  6 21:43:17 1997
Received: from venus.Sun.COM (venus.Sun.COM [192.9.25.5])
	by xemacs.org (8.8.5/8.8.5) with SMTP id VAA27756
	for <xemacs-beta@xemacs.org>; Thu, 6 Feb 1997 21:43:16 -0600 (CST)
Received: from Eng.Sun.COM ([129.146.1.25]) by venus.Sun.COM (SMI-8.6/mail.byaddr) with SMTP id TAA12561 for <xemacs-beta@xemacs.org>; Thu, 6 Feb 1997 19:42:47 -0800
Received: from kindra.eng.sun.com by Eng.Sun.COM (SMI-8.6/SMI-5.3)
	id TAA24353; Thu, 6 Feb 1997 19:42:41 -0800
Received: from xemacs.eng.sun.com by kindra.eng.sun.com (SMI-8.6/SMI-SVR4)
	id TAA18874; Thu, 6 Feb 1997 19:42:41 -0800
Received: by xemacs.eng.sun.com (SMI-8.6/SMI-SVR4)
	id TAA14213; Thu, 6 Feb 1997 19:42:40 -0800
Date: Thu, 6 Feb 1997 19:42:40 -0800
Message-Id: <199702070342.TAA14213@xemacs.eng.sun.com>
From: Martin Buchholz <mrb@Eng.Sun.COM>
To: Hrvoje Niksic <hniksic@srce.hr>
Cc: xemacs-beta@xemacs.org
Subject: Re: 20.0 --with-mule segfaults when viewing Japanese
In-Reply-To: <kig6808gqbw.fsf@jagor.srce.hr>
References: <yviasp3digdf.fsf@atreides.mindspring.com>
	<199702041227.NAA17114@sen2.ida.liu.se>
	<kig6808gqbw.fsf@jagor.srce.hr>
Reply-To: Martin Buchholz <mrb@Eng.Sun.COM>
Mime-Version: 1.0 (generated by tm-edit 7.100)
Content-Type: text/plain; charset=US-ASCII

>>>>> "Hrv" == Hrvoje Niksic <hniksic@srce.hr> writes:

Hrv> David Byers <davby@ida.liu.se> writes:
>> I think the following patch will prevent the segmentation fault by
>> doing bounds checking each time the display table is consulted. This
>> is a typical example of treating the symptoms, not the disease.

Hrv> Maybe you should add an #ifdef DEBUG or something make the patch issue
Hrv> a warning that the maintainers can use as a pointer to where the
Hrv> problem is.

I agree.

In general, it's BAD to add a patch that fixes a crash by fixing the
symptom, especially when it's a crash the maintainers don't get
themselves.  Reproducible crashes are sufficiently rare (!) that they
are a wonderful opportunity to fix real bugs in the code.  As
suggested elsewhere, if you are going to bandaid the crash, perhaps better is

/* #### comment explaining problem here */
#ifdef DEBUG_XEMACS
code known to crash here...
#else
bandaid solution here...
#endif

The standard form for a comment in the code that indicates brokenness
or TODOs is

/* ### comment */

Martin

