Message-Id:	<199907290027.AAA82496@out5.ibm.net>
From:	"Mark E." <snowball3 AT bigfoot DOT com>
To:	djgpp-workers AT delorie DOT com, binutils AT sourceware DOT cygnus DOT com
Date:	Wed, 28 Jul 1999 20:27:45 -0400
MIME-Version:	1.0
Subject:	Re: Another script for building binutils
In-reply-to:	<B0000096305@stargate.astr.lu.lv>
X-mailer:	Pegasus Mail for Win32 (v3.11)
Reply-To:	djgpp-workers AT delorie DOT com
X-Mailing-List:	djgpp-workers AT delorie DOT com
X-Unsubscribes-To:	listserv AT delorie DOT com

> Last test were done with binutils snapshots (990727 appears to be 
> bad: broken debug info, so I'll perhaps return to 990710 which I 
> used for some time).

Hi Andris,
I'm also sending this to binutils so perhaps we can get some help. I 
did some testing and found that compiling with -g (for COFF 
debugging) doesn't work properly while using -gstabs+ will work with 
my sample C program.

How I tested:
I compiled a small C program to an .s file using 'gcc -gstabs+ -S 
main.c'. I then used the as.exe from Dec '98 to generate a .o file and an 
as.exe from July 26th to do the same but with a different filename. I 
was able to step through the generated executables derived from both 
.o files with rhgdb.exe.

I then generated another .s file using 'gcc -g -S main.c' and repeated 
the generation of the two .o files. I could debug the main.exe derived 
from the Dec. '98 as.exe generated .o file. However, I could only step 
one line into the main.exe derived from the Jul. 26th generated .o file.

Andris, is this something like what you found?

If this is something that only happened very recently, the cause should 
be easily traceable.

BTW, I made sure to disable long section support and my alignment 
patch #2 before testing.

Mark
--- 
Mark Elbrecht, snowball3 AT bigfoot DOT com
http://snowball.frogspace.net/

- Raw text -