Date: Thu, 14 May 1998 15:48:56 +0200 (WET)
From: Andris Pavenis <pavenis AT lanet DOT lv>
To: Michel JULIER <julier AT GES DOT univ-montp2 DOT fr>
cc: djgpp AT delorie DOT com
Subject: Re: BUG: debugging with gcc 2.8.1
In-Reply-To: <1.5.4.16.19980514110938.237f77da@um2gess0>
Message-ID: <Pine.A32.3.91.980514154026.65286A-100000@ieva01.lanet.lv>
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII
Precedence: bulk



On Thu, 14 May 1998, Michel JULIER wrote:

> This is a bug report made with RHIDE Version 1.4 (Sep 30 1997 23:06:59)
> I am running MS-DOS 6.20 (Windows 3.11).
> At the end of the message are some information collected by rhide - bug_report.
> 
> My problem is the following:
> ============================
> 
> Using gcc 2.8.1, I noticed some problems to debug some programs.
> More precisely:
>   -In a program that contains #include "sub_prog.c", it is impossible
> to follow the pointer while stepping with gdb (actually, I use rhide).
> However, I have no problem if I include the sub-program directly into
> the main file.

Perhaps options -gstabs or -gstabs+ instead of -g can help here. (See 
Options-->Compilers-->Debugging)

> 
>   -With another subprogram, I get an error compliling the program with
> "gcc -g -Wall -save-temps -c fit_gene.c"
> >Compiling: fit_gene.c
> >Assembler messages:
> >fit_gene.s(4208) Error: Fatal error: C_EFCN symbol out of scope
> >There were some errors
> 
> This occurs only with the "-g" option, and does not depend on the
> other options. It does the same if I include the subprogram
> directly into the main file.
> 
> Looking at fit_gene.s: "C_EFCN" does not show in this file, and
> the line 4208 is the last line of the file.
> 
> If I compile without the "-g" option, everything is OK (except for
> debugging)
> 

Strange. Haven't met something similar before (I'm working with DJGPP 
port of gcc-2.8.1 already for almost 2 months)

Andris