X-Authentication-Warning:	delorie.com: mailnull set sender to djgpp-bounces using -f
Date:	Mon, 7 Jan 2002 10:25:25 +0200 (IST)
From:	Eli Zaretskii <eliz AT is DOT elta DOT co DOT il>
X-Sender:	eliz AT is
To:	"Guido S." <sassi AT uni DOT de>
cc:	djgpp AT delorie DOT com
Subject:	Re: GCC 3.0.x and GDB 5.00 - symbol demangling
In-Reply-To:	<a1aoqc$drp$1@newsserver.rrzn.uni-hannover.de>
Message-ID:	<Pine.SUN.3.91.1020107101403.9062C-100000@is>
MIME-Version:	1.0
Reply-To:	djgpp AT delorie DOT com
Errors-To:	nobody AT delorie DOT com
X-Mailing-List:	djgpp AT delorie DOT com
X-Unsubscribes-To:	listserv AT delorie DOT com

On Mon, 7 Jan 2002, Guido S. wrote:

> Since the relase of GCC 3.0.x for djgpp I have the problem of using gdb
> 5.0 with the resulting objects/binaries.
> I intend to set breakpoints in a c++ object's member function with
> b 'MyClass:MyFunc'. but it seems, that if I compile a file with -g3
> option, gdb can not handle the stored debug info.

Does it help to use -gstabs+ instead of -g3?  I think -g3 tells GCC 3.0.x 
to produce DWARF2 debug info, and the DJGPP port of GDB 5.0 doesn't yet 
support DWARF2.

> it steps through the sources correctly, but it then prints only the
> signature of the function, not the demangled name of the function.
> 
> So my question is:
> - Is there going to be an updated GDB for djgpp in the near future?

GDB 5.1 was released a little more than a month ago.  I didn't release a 
DJGPP port of that version because the C++ debugging, especially with the 
latest 3.x series of GCC, is seriously hosed in GDB 5.1.  In particular, I 
think the problems you see with demangling are not yet solved there.  
People who use C++ much more than I do and know much more about GDB tell
me unanimously that GDB 5.1 is not very useful for debugging serious C++ 
programs built with GCC 3.x.

I don't want to release a port that is not useful for C++ debugging, 
because that would trigger a flood of questions from users that I will 
be unable to answer.  So I'm waiting for GDB 5.1.1, due out in a few 
weeks (one hopes; the release cycle is about to start), where many of 
those problems are fixed.

If you want, you could download the official GDB 5.1 release from 
ftp.gnu.org and build it yourself; it should build with DJGPP out of the 
box if you follow instructions in the gdb/config/djgpp/README file.

> or
> - May I get my intended functionality by compiling with any other -g
> option? dwarf(-2), stabs etc?

As I said, try -gstabs+.  DWARF2 is not supported by GDB 5.0 that you
are using.

> additionally I'ld like to ask if there is a GDB binary out there, that
> is compiled for i586 (the one currently on simtel is for 686, and I am
> not sure of, if it works ok on my k6-2 in respect to the incompatible
> breakpoint support of those architectures)

There's nothing 686-specific in the GDB binary you find on SimTel (I 
know, since I've built it ;-), so any problems you see are not due to 
this aspect.  FWIW, I'm using the same binary on a P166, which is i586, 
and never had any problems with breakpoints.

What's incompatible with breakpoint support on K6-2, and why do you think 
it should matter for GDB?

- Raw text -