Path: utzoo!utgpu!news-server.csri.toronto.edu!rpi!zaphod.mps.ohio-state.edu!uakari.primate.wisc.edu!caen!uwm.edu!ogicse!ucsd!dog.ee.lbl.gov!elf.ee.lbl.gov!torek
From: torek@elf.ee.lbl.gov (Chris Torek)
Newsgroups: comp.lang.c
Subject: Re: Memory allocation / data access
Message-ID: <11121@dog.ee.lbl.gov>
Date: 19 Mar 91 19:53:23 GMT
References: <29491@cs.yale.edu> <5654@male.EBay.Sun.COM>
Reply-To: torek@elf.ee.lbl.gov (Chris Torek)
Distribution: na
Organization: Lawrence Berkeley Laboratory, Berkeley
Lines: 19
X-Local-Date: Tue, 19 Mar 91 11:53:23 PST

In article <5654@male.EBay.Sun.COM> harry.protoolis@uk.1.com writes:
>... why do you use malloc the first time and calloc the
>next ?, as they are both array allocations you should, strictly speaking, be
>using calloc for both.

Well, no: calloc(n,s) is *exactly* the same as

	malloc(n * s)

plus (if the malloc succeeds) a call to memset to set all n*s `char's
to '\0'.  If the objects you are allocating hold something other than
integral types, this `zeroing' is not useful since other types
(i.e., pointers and floating types) may not have all-bits-zero `0'
values.  For instance, some machines use `ring 7' for all nil pointers,
or a special bit for all floating point values meaning `do not trap
when this floating point value is examined'.
-- 
In-Real-Life: Chris Torek, Lawrence Berkeley Lab CSE/EE (+1 415 486 5427)
Berkeley, CA		Domain:	torek@ee.lbl.gov