> To: "Brian C. Grayson" <bgrayson@marvin.ece.utexas.edu>
> Cc: "Chris G. Demetriou" <cgd@netbsd.org>, lukem@cs.rmit.edu.au,
>         tech-userlevel@netbsd.org
> Subject: Re: sanity checking arguments to library functions...
> cc: proff@iq.org
> From: Julian Assange <proff@iq.org>
> Date: 16 Apr 1999 23:49:58 +1000
> User-Agent: Gnus/5.070066 (Pterodactyl Gnus v0.66) XEmacs/21.0(beta67) (20 minutes to Nikko)
>
> "Brian C. Grayson" <bgrayson@marvin.ece.utexas.edu> writes:
>
> >   I've implemented a stack dump in such a way that it works
> > for any gcc/egcs-compiled program, and I'm sure others have done
> > similar stuff.  Interested parties can contact me off-line for
> > more details.  With this functionality, an abort actually
> > provides _some_ useful info.
> > 
> >   Brian
>
> Can we see this in the tree, brian?
>
>

If the thing works by using gcc builtins like:

        __builtin_return_address(frameno)

Well...actually...that doesn't work on "any gcc/egcs-compiled program"
because the builtin isn't implemented on all architectures. (And, for  
those that it isn't implemented for, I don't think it errors out, so
you just get a run-time exception.)

	Ross.Harvey@Computer.Org