From:	pavenis AT lanet DOT lv
To:	Eli Zaretskii <eliz AT is DOT elta DOT co DOT il>, djgpp-workers AT delorie DOT com
Date:	Wed, 26 Jun 2002 15:13:06 +0300
MIME-Version:	1.0
Subject:	Re: gcc-3.1.X and DWARF2 debugging info for DJGPP
Message-ID:	<3D19DA02.11875.20CAF8@localhost>
References:	<3D19C2DD DOT 1004 DOT 8F32B0 AT localhost>
In-reply-to:	<Pine.SUN.3.91.1020626143939.13169A-100000@is>
X-mailer:	Pegasus Mail for Windows (v4.01)
Reply-To:	djgpp-workers AT delorie DOT com

On 26 Jun 2002 at 14:41, Eli Zaretskii wrote:

> 
> On Wed, 26 Jun 2002 pavenis AT lanet DOT lv wrote:
> 
> > Also one should rebuild also bfdstubify with newer libbfd to be useable with
> > DWARF2 debug info.
> 
> Newer than the one included with GDB 5.2?

One from GDB-5.2 is OK.

> > Of course it cannot be done with -gcoff.  I thought the default -g. One can 
> > still use -gcoff of course.
> 
> Personally, I'm not sure we are ready to switch to DWARF2 as the default 
> format.  But if others think otherwise, I won't object.

From first impression it seems to be not so bad. But perhaps more testing is 
needed. I'm using practically only DWARF2 debugging info for my own
programs  (mostly C++ with often rather heavy use of STL etc.) already for some 
time.

Well, I put archives of my build of gcc-3.1.1 20020625 (prerelease) at
	http://www.ltn.lv/~pavenis/gcctest.html

It would be nice if others would test my binaries and report about problems 
here. 

My package of gdb-5.2 is also there. My latest builds of RHIDE also uses
internally GDB-5.2, so they should be OK.

> > Also change to DWARF2 debugging info as the default would force to 
> > put bfdsymify instead of symify. Maybe it's worth to release it in a separate
> > package (not as a part of gdbXXXb.zip)
> 
> Unfortunately, I don't have time to do that.

Andris

- Raw text -