Path: utzoo!utgpu!watmath!att!dptg!rutgers!usc!elroy.jpl.nasa.gov!ames!think!bloom-beacon!mit-eddie!uw-beaver!blake!oregon!joe
From: joe@oregon.uoregon.edu (Joe St Sauver)
Newsgroups: comp.lang.postscript
Subject: Postscript code needed to recover LaserWriter VM
Message-ID: <6318@oregon.uoregon.edu>
Date: 9 Aug 89 15:34:42 GMT
Organization: University of Oregon
Lines: 25

Can anyone point me to Postscript code that will allow me to do the software 
equivalent of turning an Apple Laserwriter off and back on again, or otherwise
recover needed virtual memory?

Essentially, in case someone can suggest a better alternative, here's the
story: Postscript output is being generated by TeX/dvips (on a NeXT sitting on
an ethernet). The Postscript is then being transferred (using binary FTP) to a
PC with a LocalTalk card. LWPRINT is then used on the PC to send the 
Postscript to a shared Laserwriter (sitting on the LocalTalk network). 

[Before you die laughing at this arrangement, we *will* be striving to
replace this rather kludgy "solution" with CAP before too long, but for now,
here's what happens...]

The Laserwriter performs like a real soldier *IF* the NeXT output is sent to the
LaserWriter just after it has been powered down and back up. However, if the 
normal hodgepodge of jobs from Macintosh and Novell-network have been pumped 
through ahead of it, the NeXT Postscript output croaks with a VMerror when
sent from the PC with LWPRINT.

Incidentally, before anyone asks, yes, I have reduced the amount of memory 
declared to be available to PS to only 35000 bytes, so the NeXT is *not* 
trying to send Postscript suitable for a memory-rich NeXT laser printer. (I've
tried lower values, but 30000 or below hangs dvips). Remember, it works fine
if the NeXT output is the first job onto the newly powered-up LaserWriter.