This is the mail archive of the
libstdc++@gcc.gnu.org
mailing list for the libstdc++ project.
RE: debugging programs that use stl containers
- To: "'libstdc++ at gcc dot gnu dot org'" <libstdc++ at gcc dot gnu dot org>
- Subject: RE: debugging programs that use stl containers
- From: "Craveiro, Marco" <Marco dot Craveiro at solvay dot com>
- Date: Tue, 3 Apr 2001 16:38:52 +0200
> Thanks for your replies!
>
> I just have two more questions: in terms of the debbuging info, what
> should I use instead of STABS
> and how can I get gcc to do it?
>
> Also, I got a "straight from the shelves" red hat 7.0.98. I do a normal
> #include <map>. Does this mean
> I'm still using g++ and not libstdc++? in other words, is the support that
> you are talking about
> already there or do I need to download and compile gcc w/ libstdc++?
>
> Thanks again,
>
> Marco
> -----Original Message-----
> From: Daniel Berlin [SMTP:dan@www.cgsoftware.com]
> Sent: Tuesday, April 03, 2001 12:20 AM
> To: Benjamin Kosnik
> Cc: Craveiro, Marco; 'libstdc++@gcc.gnu.org'
> Subject: Re: debugging programs that use stl containers
>
> OYeah, it's slightly flaky. I've got it goin fine on the new abi, except
> on objects that have multiple shared virtual baseclasses.
>
> YOu also *should* be able to do p my_vector[0], p my_map["fred"], etc,
> just not with STABS. The typing on stabs is too wrong because we have to
> parse them.
>
>
> On Mon, 2 Apr 2001, Benjamin Kosnik wrote:
>
> >
> > not really. You can do
> >
> > (gdb) ptype my_vector
> >
> > to see what gdb thinks of the type, and
> >
> > (gdb) p my_vector
> >
> > to see the internals of the object. This is supposed to work, but in my
> > experience is a bit flaky.
> >
> > -benjamin
> >
>