Date: Wed, 11 Jun 1997 11:53:57 +0200 (METDST)
From: Robert Hoehne <robert DOT hoehne AT mathematik DOT tu-chemnitz DOT de>
Reply-To: Robert Hoehne <robert DOT hoehne AT mathematik DOT tu-chemnitz DOT de>
To: Eli Zaretskii <eliz AT is DOT elta DOT co DOT il>
Cc: DJGPP workers <djgpp-workers AT delorie DOT com>
Subject: Re: Latest stub
In-Reply-To: <Pine.SUN.3.91.970609135020.618J-100000@is>
Message-Id: <Pine.HPP.3.95q.970611114419.19296C-100000@newton.mathematik.tu-chemnitz.de>
Mime-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII
Precedence: bulk

On Mon, 9 Jun 1997, Eli Zaretskii wrote:

> 
> other day.  I think the solution is just to code explicit 0 where a 0 is 
> (incorrectly) assumed.

Since this was the only answer to my originally question I will
apply the the following patch to the stub and then include it 
in the binutils:

--- stub.asm~	Wed Jun 11 11:39:57 1997
+++ stub.asm	Wed Jun 11 11:41:28 1997
@@ -229,7 +229,7 @@
 @f1:
 	movd	[bx+0], 0x4558452e	; append ".EXE"
 	add	bx, 4
-	mov	[bx], al		; al = 0
+	mov	[bx], 0			; store nul
 	mov	[loadname_nul], bx	; remember nul so we can change it to $
 
 no_symlink:


And a note for future changes at the stub. My current patches
to the binutils assume a FIXED size of the stub of 2048 bytes.

That means, any changes on the stub, which will exceed this limit
will be incompatible with binutils 2.8.1!!

The current binutils 2.7 can handle also different sized stubs,
but these patches were not good (the coding style) and the 
assumption to a fixed size is also much better to implement.

If someone cannot agree with this let me know because I will
send the final patches in the next days.

Robert