From: pavenis AT lanet DOT lv
To: Eli Zaretskii <eliz AT is DOT elta DOT co DOT il>
Date: Tue, 27 Mar 2001 16:30:58 +0300
MIME-Version: 1.0
Content-type: text/plain; charset=US-ASCII
Content-transfer-encoding: 7BIT
Subject: Re: gcc2953 and RHIDE debugging
Cc: djgpp-workers AT delorie DOT com
Message-ID: <3AC0C042.22771.17023A0@localhost>
References: <3ABF29BF DOT 8067 DOT F04DC AT localhost>
In-reply-to: <Pine.SUN.3.91.1010326130917.8356B-100000@is>
X-mailer: Pegasus Mail for Win32 (v3.12c)
Reply-To: djgpp-workers AT delorie DOT com

On 26 Mar 2001, at 13:13, Eli Zaretskii wrote:

> 
> On Mon, 26 Mar 2001 pavenis AT lanet DOT lv wrote:
> 
> > For comparisson: my build of gcc-3.0 20010314 development version 
> > (from CVS) for DJGPP seems to have correct line numbers in debug 
> > info with -gcoff.
> 
> Perhaps it would be possible to identify the changes in the
> development sources that bring -gcoff support back, and retrofit them 
> into GCC 2.95.3.  Then we could make an updated release.
> 
> It is IMHO unfortunate that the stable release doesn't support COFF 
> debug info.  I think this makes its value for DJGPP users much less that 
> it could have been.  I also anticipate a flood of FAQs as more people 
> discover this problem.  I think it's worth some effort to try to correct 
> this as quickly as possible.

I have put binaries of updated cc1.exe and cc1plus.exe at
	http://www.ltn.lv/~pavenis/cc1.exe
	http://www.ltn.lv/~pavenis/cc1plus.exe

First impression is that -gcoff problem is fixed, but perhaps more
testing is needed. 

There were 2 possibilities:

undo one patch in gcc/function.c (easy, get diffs from cvs, ...)

don't undo it, but try to update it from 3.0 CVS sources (all related 
changes are located in a part of a single procedure). I tried 2nd way.
Seems that works for me. So please test it.

Andris