Path: utzoo!utgpu!jarvis.csri.toronto.edu!mailrus!tut.cis.ohio-state.edu!purdue!ames!ig!arizona!cjeffery
From: cjeffery@arizona.edu (Clinton Jeffery)
Newsgroups: comp.sys.ibm.pc
Subject: Re: Rational Sys. DOS Extender?
Message-ID: <10607@megaron.arizona.edu>
Date: 4 May 89 07:20:38 GMT
References: <1135@ptolemy.arc.nasa.gov>
Organization: U of Arizona CS Dept, Tucson
Lines: 22

From an article, by raymond@ptolemy.arc.nasa.gov (Eric A. Raymond):
> Of course if your code is less than 640K (or whatever size your LOW
> memory is), you can use a malloc which handles extended/expanded
> memory for a LOT LESS HASSLE and $'s.

This statement is a bit naive.  There is no such thing as
a "malloc that handles extended/expanded memory" under MSDOS proper.
If you are willing to turn every memory ACCESS (read/write) into a
procedure call you can implement swapping using extended/expanded
memory (or disk) easily.  If you only need another 64K and have no other
application using extended memory you can directly address 64K of
the extended memory space from DOS (and simply never do any swapping).
If you have an AT you can get an (almost) real operating system which
breaks the 640K barrier.  If you write your application assuming file i/o
instead of memory then you can run as big a cache as you can afford and
most of the reads will be to memory.  But no malloc can "handle"
extended/expanded memory under normal MSDOS in the usual C fashion.
Correct me if I am wrong.
-- 
| Clint Jeffery, University of Arizona Department of Computer Science
| cjeffery@arizona.edu -or- {noao allegra}!arizona!cjeffery
--