[plt-scheme] Problems With Large Files in .plt Files

From: Williams, M. Douglas (M.DOUGLAS.WILLIAMS at saic.com)
Date: Sat Jan 28 12:35:41 EST 2006

I rebuilt the .plt files without the pdfs and it loads just fine on a 256MB
machine.  I will submit the new ones to PLaneT.  Unfortunately, I really
liked distributing the reference manual (as a pdf) with the collection.


If someone wants the old one with the pdfs in order to replicate the
problem, I can send it to them or they can download it from PLaneT before
the new one is made available.





From: plt-scheme-bounces at list.cs.brown.edu
[mailto:plt-scheme-bounces at list.cs.brown.edu] On Behalf Of Williams, M.
Sent: Saturday, January 28, 2006 9:46 AM
To: plt-scheme at list.cs.brown.edu
Subject: [plt-scheme] Problems With Large Files in .plt Files


When I built the science and simulation collections for distribution with
PLaneT, I included pdfs of the reference manuals - for example, the
science-ref.pdf file size is 677KB.  I built and tested it on my development
machines (both Windows and Linux), which have at least 1GB each, and they
work great.


As a test, try '(require (planet "science.ss" ("williams" "science.plt" 2
0)))' or 'planet -i williams science.plt 2 0' from the command line.  On a
1GB machine. a minute or so later you have the science collection installed
and compiled on your machine.


I went to demonstrate it on another machine, which had only 512MB, and five
minutes later we had a very non-responsive machine and Drscheme eventually
died - with a corrupt installation of the collection.  I tried it on various
machines with the same result - any machine with 1GB or greater it works and
anything 512MB or less it fails.  I tried various combinations of virtual
memory settings and had the same result.


Watching the memory usage on task manager - yes, a very scientific way to do
it :-) - shows that the process grows slowly until it fills memory; then it
shrinks back to a couple tens of megs before finally dying several minutes
later.  No errors are reported in the PLaneT install-log.  The
science-ref.pdf file is corrupt (I assume truncated, but I didn't check) and
nothing after it is installed.


I will rebuild the .plts without the pdfs and that should fix the problem.
But, in looking over the unpack module in setup, it seems to go out of its
way to avoid having the entire file in memory during the deflating process.
So I'm curious as to what is going on.



-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.racket-lang.org/users/archive/attachments/20060128/b5a38f4a/attachment.html>

Posted on the users mailing list.