From: Hans-Bernhard Broeker <broeker AT physik DOT rwth-aachen DOT de>
Newsgroups: comp.os.msdos.djgpp
Subject: Re: malloc question - About output casting
Date: 8 Jun 2000 10:24:22 GMT
Organization: Aachen University of Technology (RWTH)
Lines: 29
Message-ID: <8hns8m$bvh$1@nets3.rz.RWTH-Aachen.DE>
References: <393EFFB8 DOT 2D5AA7BB AT fujixerox DOT co DOT jp>
NNTP-Posting-Host: acp3bf.physik.rwth-aachen.de
X-Trace: nets3.rz.RWTH-Aachen.DE 960459862 12273 137.226.32.75 (8 Jun 2000 10:24:22 GMT)
X-Complaints-To: abuse AT rwth-aachen DOT de
NNTP-Posting-Date: 8 Jun 2000 10:24:22 GMT
Originator: broeker@
To: djgpp AT delorie DOT com
DJ-Gateway: from newsgroup comp.os.msdos.djgpp
Reply-To: djgpp AT delorie DOT com

Burgupalli Chaitanya <Chaitanya DOT Burgupalli AT fujixerox DOT co DOT jp> wrote:

> But there was a mailing list thread where in it was recomended that 
> it is better not do that. 

... and by coincidence, the problem you describe is _exactly_ the one
that is the key argument against malloc() casting. You've been
brushing a potential problem under the carpet by casting, which now
surfaces as you removed the casts:

> 	main.c:34: warning: assignment makes pointer from integer without a
> cast

This is the telltale sign of a forgotten #include <stdlib.h>. This
warning is somewhat serious, but with the cast in place, you wouldn't
have gotten it.

Casts between pointer types should be needed only *very* rarely, in a
properly written C program. Whenever you catch yourself writing one,
it's best to sit back and re-think your current strategy: chances are
you're just doing something silly or evil. About the only exceptions
from this rule are arguments to variadic functions (like printf())
that have to be casted to (void *), and casts back from void* to some
other type needed with calls through generalized function pointers
like the qsort() comparison function. 

-- 
Hans-Bernhard Broeker (broeker AT physik DOT rwth-aachen DOT de)
Even if all the snow were burnt, ashes would remain.