Path: utzoo!utgpu!utstat!jarvis.csri.toronto.edu!mailrus!tut.cis.ohio-state.edu!ucbvax!agate!eos!ptolemy!raymond
From: raymond@ptolemy.arc.nasa.gov (Eric A. Raymond)
Newsgroups: comp.sys.ibm.pc
Subject: Heap Expander breaks 640K? !!!
Message-ID: <1086@ptolemy.arc.nasa.gov>
Date: 14 Apr 89 18:50:05 GMT
Reply-To: raymond@ptolemy.arc.nasa.gov (Eric A. Raymond)
Organization: NASA Ames Research Center, Moffett Field, CA
Lines: 27

Has anyone had any experience with using the Heap Expander from the
Tool Makers (408-458-0690 for info, 800-248-1045 x100 to order)?

Apparently, it includes library and source (for Turbo C & Pascal, MSC)
to extend your heap using extended, expanded or disk memory.  The ad
claims to provide a virtually unlimited heap (limited by disk space?).

Does it work in the Huge memory model?
Does it work with my interrupt routines?
Does it work with my 386 protected mode device drivers?
Will it destroy the debbuger?

This really sounds to good to be true.  I'm just about hitting my head
on the 640K (actually 736k using '386 backfilling) memory ceiling.
The other alternative would be to go to Unix (which may not be
possible as this is a space bound application and hard disk
availability is uncertain) or a protected mode system (i.e. 16M, Phar
Lap, OS/386 (from AI Architects, not MS)).  All of these are
expensive ($500 - $5000) and make development cumbersome if I want to
retain my exisisting graphics libararies and development environment.
Certainly '386 protected mode would be faster (32 bit pointers in
small memory model), but seems awfully painful right now.

Any chance of a protected mode Turbo C being on the horizon?
-- 
Eric A. Raymond  (raymond@ptolemy.arc.nasa.gov)
"A hungry mob is an angry mob"