This is the mail archive of the
gdb-patches@sourceware.org
mailing list for the GDB project.
Re: [RFA/commit] error/internal-error printing local variable during "bt full".
- From: Joel Brobecker <brobecker at adacore dot com>
- To: Pedro Alves <palves at redhat dot com>
- Cc: gdb-patches at sourceware dot org
- Date: Mon, 23 Nov 2015 10:06:46 -0800
- Subject: Re: [RFA/commit] error/internal-error printing local variable during "bt full".
- Authentication-results: sourceware.org; auth=none
- References: <1447093747-21709-1-git-send-email-brobecker at adacore dot com> <564F4851 dot 4070102 at redhat dot com>
> > gdb/ChangeLog:
> >
> > * stack.c (print_frame_local_vars): Temporarily set the selected
> > frame to FRAME while printing the frame's local variables.
[...]
> LGTM.
Thanks, Pedro. The patch has now been pushed.
> > Thinking longer term, I think the best approach is to add a "frame"
> > parameter to value_at_lazy, and make sure it gets passed down all
> > the way to the type resolver. But the number of locations that this
> > change impacts is quite large, so I just wanted to receive the OK from
> > the rest of the group before embarking on such a large change, knowing
> > that once done, the patch would ideally be pushed quickly to avoid
> > having to handle rebasing conflicts.
>
> In general (haven't looked at the locations this particular case involves),
> I've wondered before about adding some "struct execution_context" or some
> such, which would encapsulate frame, thread, target, language, etc.,
> and pass around a pointer to such a thing.
I think it might make sense. There are a number of places where it would
be beneficial to access that info through such a struct, rather than
through the global context... I'll stew on it for a while...
--
Joel