Path: utzoo!utgpu!news-server.csri.toronto.edu!cs.utexas.edu!usc!jarthur!petunia!news
From: rcfische@polyslo.CalPoly.EDU (Ray Fischer)
Newsgroups: comp.sys.mac.programmer
Subject: Re: Help with bad calloc() call in THINK C 4.02
Message-ID: <26f7f672.65b2@petunia.CalPoly.EDU>
Date: 19 Sep 90 22:51:30 GMT
References: <26250@cs.yale.edu>
Organization: Cal Poly State Univ,CSC Dept,San Luis Obispo,CA 93407
Lines: 24



tarr-michael@CS.YALE.EDU (michael tarr) writes ...
>For some reason using THINK C 4.02 the following calloc (at the end of the
>code segment) breaks. It returns a pointer that is not really valid (but not
>NULL). Here are the data structures used:

I've not copied your code in this reply, but nowhere do I see either a
	#include <stdlib.h>
or a 
	void *calloc();

With Think C (and possibly other C compilers) a definition for the calloc (and
alloc, et al) function is required or you will get bad results.

The reason is that the default return value for a function is an int.  In
Think C an int is just 16 bits.  Thus, without any declaration to the contrary,
Think C will take the 32-bit address returned, expand the low-order 16-bits
to 32 bits, then assign that value to your pointer.  

The solution is simply to tell the compiler that calloc returns a 32-bit result.

Ray Fischer
rcfische@polyslo.calpoly.edu