From:	sandmann AT clio DOT rice DOT edu (Charles Sandmann)
Message-Id:	<10010011658.AA13712@clio.rice.edu>
Subject:	Re: Beta Testers Needed
To:	Martin DOT Stromberg AT lu DOT erisoft DOT se (Martin Stromberg)
Date:	Sun, 1 Oct 2000 11:58:09 -0500 (CDT)
Cc:	djgpp-workers AT delorie DOT com
In-Reply-To:	<200010011100.NAA07258@lws256.lu.erisoft.se> from "Martin Stromberg" at Oct 01, 2000 01:00:31 PM
X-Mailer:	ELM [version 2.5 PL2]
Mime-Version:	1.0
Reply-To:	djgpp-workers AT delorie DOT com
Errors-To:	nobody AT delorie DOT com
X-Mailing-List:	djgpp-workers AT delorie DOT com
X-Unsubscribes-To:	listserv AT delorie DOT com

> Then you might be interested in the fact that I've been messing with
> the stub as well. Mainly fixing an invalid header entry, see
> <http://www.delorie.com/djgpp/mail-archives/browse.cgi?p=djgpp-workers/2000/09/09/12:56:33>.

I read the thread on DJ's web page.  PMODE's stub and CWSDPMI's stub are
built using Borland's tools (ie TLINK) - and it leaves the word at 0x3c empty,
but the word at 0x3e is the start of the relocations.  PMODE's exe header
futzer does something similar to DJASM with a copyright - but needs to 
handle relocations and puts it before the relocations.

With CWSDPMI's imbedded stub I put the copyright immediately after the 
relocations, but it starts well after the first 64 bytes of the header.
Since none of the relocations contain ^Z - it still prints if typed.

When I get around to putting a new base stub up, I'll try to get the new 
DJASM to build it.

I did notice the embedded CWSDPMI doesn't work at all if built with my TC 2.0
instead of BCC 5.0 - so I see some additional debugging ahead :-(

(I also discovered an interesting bug-feature of the system() command 
during testing - I had an invalid directory in my path and system quit ...)

- Raw text -