Path: utzoo!utgpu!attcan!uunet!lll-winken!lll-tis!ames!oliveb!pyramid!voder!kontron!optilink!cramer
From: cramer@optilink.UUCP (Clayton Cramer)
Newsgroups: comp.lang.postscript
Subject: Re: micro-soft word postscript header file.
Message-ID: <322@optilink.UUCP>
Date: 28 Jul 88 22:00:46 GMT
References: <808@wacsvax.OZ>
Organization: Optilink Corporation, Petaluma, CA
Lines: 44

In article <808@wacsvax.OZ>, doug@wacsvax.OZ (Doug Robb) writes:
> 
> I have an ast turbo laser hooked onto (via serial line) a vax750 running
> 4.3BSD. I can spool off files using lwf (posted on net a while
> back) and so flow control etc seems to be working (in fact I know
> its working as I've hung a terminal off the line instead of the printer
> and checked it). 
> 
> The problem:
> 
> To print files from a msword document (nfs'ed up from pc) requires
> the msword postscript prolog file to be sent to printer once.
> After I send this file (the prolog) it refuses to print more than 3 pages
> of any msword document. Documents shorter than this are printed
> ok most of the time but not always. 
> It usually gives me "error: undefined: offending command : flushing
> rest of job " etc at random spots in the file. 
> Now the SAME prolog with SAME printer and SAME test file
> works fine if the printer is hooked directly to a pc and driven through
> ms-word. The prolog writes to non volotile memory so I could understand
> how it could screw up ditrof, lwf etc files but to screw up its own
> files? By the way the header loads up without er> out flow control as I have said so what else could be happening?
> 
> Has anyone encountered problems with loading/running ms-word documents
> from a unix host? Any thoughts about this strange behaviour would be 
> appreciated. Thanks doug@wapsyvax.oz

No.  But thea couple of weeks while debugging my Epson MX-80 to PostScript trans-
lator, and I found that it would generate the "ioerror" messagehow far in it would fail, but how far still varied in some non-obvious
way.

I spoke to other users of the AST TurboLaser/PS, as well as an engineer
at AST (who was having trouble persuading his colleagues), and all were
aware that very long PostScript files would occasionally fail.  It looks
to me like the problem is a dropped or garbled character, and was
readily reproducible with either parallel or serial interface.  I didn't
use the Appletalk interface, but some of my customers say they have
had no problem with Appletalk -- only parallel and serial.

Note that this may not be a problem for all AST TurboLaser/PS printers.

The difference between UNIX and PC may simply be a characteristic of
how fast the data is transmitted to the printer.

Clayton E. Cramer