Dan thanks for the info.  I did manage to get a build of gdb (I can live
without floating point).  Setting breakpoints does work, however
when I execute the code, it terminates saying that
the application received a SIGILL (Illegal instruction). If I remember
correctly, gdb and other debuggers insert an illegal instruction at
the breakpoint address to cause a trap, or am I totally off?
Does your gdb work with breakpoints correctly?

The sample code was simple:

int main ()
{
    x = 0;
    x++;
}

Thanks
Sanjay



Dan Winship wrote:

> > Last I heard, it was possible to get a working gdb running with
> > -current, but it doesn't cooperate well with FPU stuff on the ppc.
>
> That's pretty much correct. "cd /usr/src/gnu/usr.bin/gdb; make && make
> install" will get you gdb, but it doesn't deal with floating point,
> and I'm not sure can read core files...
>
> > Until there's some commercial-scale (which includes open source stuff
> > like the Gimp) production being done on machines running NetBSD/macppc,
> > there won't be a real call for a debugger like gdb, and that won't
> > happen until the systems actually stable and complete, so...
>
> Funny you should mention that. I discovered a few days ago that due to
> a compiler bug, Gtk+ apps will crash almost immediately on the
> powerpc. I'm working on it.
>
> More generally, the problem is that there just aren't that many people
> doing serious development on port-macppc, so it's still rough around a
> lot of edges.
>
> -- Dan