Date: Tue, 29 Dec 1998 15:33:50 -0800 (PST)
From: "Steven M. Schultz" <sms@moe.2bsd.com>
Message-Id: <199812292333.PAA12655@moe.2bsd.com>
To: pups@minnie.cs.adfa.oz.au
Subject: Re: Bob Supnik's Emulator.
Sender: owner-pups@minnie.cs.adfa.edu.au
Precedence: bulk

Robin -

> From robin@falstaf.demon.co.uk Tue Dec 29 15:21:08 1998
> >       Umm, you can't use a 'dd'd image - you have to use 'makesimtape'
> >       (or a similar utility) to add the record/file/bytecount markers that

> Now this is what I didn't realise at first.  All I thought makesimtape
> was doing was packaging up the files, not writing some structure around

	It's writing simulated bytecounts and simulated file and tape marks ;)

> >       I assume you compiled and ran 'makesimtape' on the same system
> >
> This is the big one, no.  I had assumed that as the simulator was

	Ah, ok - so you're running the makesimtape program on an 11.  That
	would tend to point the finger at the program not flipping the 
	'structure' bytes into correct big endian order.

> emulating a PDP that it would accept files generated to look like boot
> files etc built on a pdp so I'm running makesimtape in the standalone
> directory of the 11/73.  Nieve maybe but at least it was logical :-).

	The "data" is PDP-11 specific, but the "structure" bytes need to be
	in a canonical (big endian) form.

	I was pretty sure the endianness was ok but I guess not.  Another
	possibility is that there's an alignment disagreement.  The 11 might
	be putting something on a 2 byte bound where the Sun expects a 4 byte
	alignment.  

> >       There are ifdefs around what I thought were the appropriate places
> >       for flipping bytes - what you'll need to do is get Bob's description
> Back in a mo.

	If you find (and fix ;-)) it let me know and I'll integrate the
	changes into makesimtape.c in the 2.11 tree (and eventually in to
	the PUPS archive).

	Steve