Path: utzoo!utgpu!news-server.csri.toronto.edu!cs.utexas.edu!wuarchive!zaphod.mps.ohio-state.edu!swrinde!elroy.jpl.nasa.gov!decwrl!ogicse!zephyr.ens.tek.com!gvgpsa!gold.gvg.tek.com
From: shaunc@gold.gvg.tek.com (Shaun Case)
Newsgroups: comp.os.msdos.programmer
Subject: dynamic mem allocation question
Message-ID: <2106@gold.gvg.tek.com>
Date: 29 Mar 91 23:35:18 GMT
Sender: shaunc@gold.gvg.tek.com
Organization: Grass Valley Group, Grass Valley, CA
Lines: 21


I'm using Borland C++ 2.0 to develop an application that will handle large 
amounts of transient information.  I am using the compact memory model.
I am having some problems, which are almost certainly due to wild pointers;
however, while staring at my code, I suddenly noticed that I am using both
malloc() and farmalloc() in the same program.

The only thing I use malloc() for is to allocate some string space.  That is,
I have an array of char pointers, and for each one, I malloc some space, then
free() it when I'm done.  Everything else (all linked list stuff) uses 
farmalloc() and farfree().

My question is, is this safe?  As I understand it, there is no near heap in the
compact memory model, so what happens when I use malloc()?  Is Borland C++
clever enough to malloc space from the far heap, and return it properly, or
is stuff getting clobbered?

Incidentally, I am not using any of the OOP extensions, nor any of the 
windows stuff, just straight ANSI C.

Shaun.