Path: utzoo!utgpu!jarvis.csri.toronto.edu!cs.utexas.edu!tut.cis.ohio-state.edu!snorkelwacker!mintaka!ogicse!ucsd!ucbvax!CORONA.UOREGON.EDU!parker
From: parker@CORONA.UOREGON.EDU
Newsgroups: comp.sys.apple
Subject: IIGS Memory Manager:  Help!
Message-ID: <9002230935.AA06607@corona.uoregon.edu.noname>
Date: 23 Feb 90 09:35:46 GMT
Sender: daemon@ucbvax.BERKELEY.EDU
Lines: 37

A question for the GS Programming Gurus out there--but first, the Good
News and the Bad News:

The good news:  I have written a CDA which cleans up the GS's memory by
purging all purgeable memory blocks and compacting whatever's left over.
This works wonders with Appleworks GS--I no longer have to worry about
it locking up on my 1-Meg machine due to Creeping Memory Loss Disease.

The bad news:  In order to find the handles of all purgeable blocks, I
had to write a routine to scan the Memory Manager's handle list.  This
requires the use of undocumented information about the Memory Manager's
data structures (I learned about it by disassembling the Memory Peeker
CDA), so my program will probably die a horrible death if Apple ever
rewrites the Memory Manager.

And now, the question:  I would like very much to accomplish this task by
"legal" means, but a thorough search of the available documentation
reveals no such "legal" means.  Is there any Apple-approved way to purge
ALL purgeable blocks regardless of memory address or owner ID?  For that
matter, is there any Apple-approved way to do ANYTHING to a set of
handles for which the owner IDs and addresses are not known?

If this information is documented in the new Toolbox Ref. Vol. 3, I
apologize--I have no access to this information, since I'm just a poor
grad student, and I can't currently afford APDA's prices.  ($35 each for
the beta drafts, when the final editions of Vols. 1 and 2 only cost
$26.95?!?  You've got to be kidding!)

	  - Neil Parker

P.S.  If anybody out there is interested in a DA to free up all freeable
memory in the GS, let me know.  If there's enough interest, I'll post my
CDA to the net.
--
Neil Parker   parker@corona.uoregon.edu (also nparker@cie.uoregon.edu or
parker@astro.uoregon.edu)   DISCLAIMER:  Nobody anywhere believes in
these opinions.