Date: Wed, 1 Apr 1998 11:17:37 +0300 (IDT)
From: Eli Zaretskii <eliz AT is DOT elta DOT co DOT il>
To: George Foot <george DOT foot AT merton DOT oxford DOT ac DOT uk>
cc: Martin Stromberg <Martin DOT Stromberg AT lu DOT erisoft DOT se>,
        djgpp-workers AT delorie DOT com, Charles Sandmann <sandmann AT clio DOT rice DOT edu>
Subject: Re: Auto-symified traceback
In-Reply-To: <199804010030.BAA24868@sable.ox.ac.uk>
Message-ID: <Pine.SUN.3.91.980401111252.2919E-100000@is>
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII
Precedence: bulk


On Wed, 1 Apr 1998, George Foot wrote:

> I think the average DOS user would be very annoyed at having large
> files dumped to disk.

``Large''?  Come on, an average program seldom takes more than 20 stack 
frames, which generates traceback that is a few KB long.  That's a single 
cluster on most modern disks.

> IMHO proper core file support (if any) should
> be an optional extra, possibly enabled by the environment or the
> programmer.  Unless GDB can read them I don't expect many people
> would be interested in it though.

Reading them is not the problem.  The real problem is how to recreate the 
memory layout which was in effect when the program crashed.  Charles once 
explained that this might be very hard in some cases.  Charles, could you 
please elaborate?

> If GDB support isn't feasible, people can simply use their own signal 
> handlers to generate their core files and use a custom debugger to 
> browse them.

Generating the core file is quite simple, I think.  Emacs does something 
very similar when it dumps its image as part of the build process.