From:	sandmann AT clio DOT rice DOT edu (Charles Sandmann)
Message-Id:	<9803301459.AA14053@clio.rice.edu>
Subject:	Re: Auto-symified traceback
To:	eliz AT is DOT elta DOT co DOT il (Eli Zaretskii)
Date:	Mon, 30 Mar 1998 08:59:46 -0600 (CST)
Cc:	djgpp-workers AT delorie DOT com
In-Reply-To:	<Pine.SUN.3.91.980330174620.20293I-100000@is> from "Eli Zaretskii" at Mar 30, 98 05:49:11 pm

> Is it a good idea to make our traceback symify itself (if the debug info 
> is present)?

This was discussed for V1.11, and again for V2.0 in beta.  The issues were
image bloat, complexity, and worse case behavior.

> If bloating the code of each program because of this is too much, maybe
> `abort' should do that?  If that is still too much, how about a special
> debugging version of abort, for those who'd like that? 

The early discussions included a special library to link against, or
potentially exec()'ing symify.

The worse case issues are when the stack is overrun, or a wild pointer
has scrogged code, or the exception table is corrupt.  You may have a very
unstable enviroment to execute symify type code.  The worry was this might
cause some hangs itself (a nesting type situation).

- Raw text -