From:	Edward Hill <ha AT nochance DOT com>
Newsgroups:	comp.os.msdos.djgpp
Subject:	Re: A Structured Problem ?
Date:	Tue, 20 Jul 1999 11:35:08 +0100
Organization:	GEC-Marconi
Lines:	22
Message-ID:	<379450DC.E0B363B5@nochance.com>
References:	<7mv0te$gq2$1 AT news6 DOT svr DOT pol DOT co DOT uk> <Pine DOT SUN DOT 3 DOT 96 DOT 990719090429 DOT 29235A-100000 AT silver DOT cs DOT umanitoba DOT ca> <7n09p0$9db$1 AT news5 DOT svr DOT pol DOT co DOT uk> <7n1c6l$14s$1 AT taliesin DOT netcom DOT net DOT uk>
NNTP-Posting-Host:	pc02372.gmsws.gecm.com
Mime-Version:	1.0
X-Mailer:	Mozilla 4.51 [en] (Win95; I)
X-Accept-Language:	en
To:	djgpp AT delorie DOT com
DJ-Gateway:	from newsgroup comp.os.msdos.djgpp
Reply-To:	djgpp AT delorie DOT com

Michael Stewart wrote:
: The code I posted used a cast to tell the compiler that the allocated
memory
: will be used as float*. Ed didn't use a cast.

It's good practice not to cast, malloc doesn't require a cast and a cast
can sometimes mask errors.
 
: malloc returns void* and by using a cast it'll prevent the compiler
: complaining about converting between types. It should be quite safe to
use
: the memory without the cast though.

If the comiler is complaining about not casting malloc's return, this
is one of the cases where casting it will mask an error.
If my compiler moans about not casting the return of malloc it's
because I have forgotten to #include <stdlib.h>

You should only ever need to cast malloc when dynamically allocating
a mutidimensional array.

Ed

- Raw text -