Date:	Sun, 29 Jul 2001 10:36:06 +0300 (IDT)
From:	Eli Zaretskii <eliz AT is DOT elta DOT co DOT il>
X-Sender:	eliz AT is
To:	djgpp-workers AT delorie DOT com
Subject:	Re: [kettenis AT wins DOT uva DOT nl: Re: i386 register numbering]
In-Reply-To:	<200107290158.VAA11721@envy.delorie.com>
Message-ID:	<Pine.SUN.3.91.1010729103520.680H-100000@is>
MIME-Version:	1.0
Reply-To:	djgpp-workers AT delorie DOT com
Errors-To:	nobody AT delorie DOT com
X-Mailing-List:	djgpp-workers AT delorie DOT com
X-Unsubscribes-To:	listserv AT delorie DOT com

On Sat, 28 Jul 2001, DJ Delorie wrote:

> Isn't our default still coff debug, perhaps with stabs?

The default is still COFF debugging, without stabs.  You can get stabs
with -gstabs and (with GCC 3.0) DWARF2 with -gdwarf-2.

> I didn't think dwarf2 was even used yet.

It isn't used, since no released port of GDB supports it.  But if you
build a recent snapshot, you will have it supported in GDB.

I expect DWARF2 to become quite popular once GCC 3.0 and GDB 5.1 are
released, since DWARF2 is so much better, especially for debugging
C++.  That's why I think we should try to fix the problem with
register naming.

> I would recommend using the same mapping as linux uses, which should
> be the "popular" one for ix86.

I believe that's what Mark was saying as well.  I agree, which makes
three of us ;-)

- Raw text -