Path: utzoo!censor!geac!sq!msb
From: msb@sq.sq.com (Mark Brader)
Newsgroups: comp.lang.c
Subject: Re: Casting malloc() in ANSI C?
Message-ID: <1989Dec19.081011.5594@sq.sq.com>
Date: 19 Dec 89 08:10:11 GMT
References: <24969@cup.portal.com>
Reply-To: msb@sq.com (Mark Brader)
Organization: SoftQuad Inc., Toronto
Lines: 32

An article which, curiously, was posted to the USA only, asks:

> In ANSI C, assigning a void pointer to any pointer variable is
> legal, so casting the value returned by malloc() to the
> appropriate pointer type is no longer necessary.  As a matter of
> recommended style, should the redundant cast be retained?

I would advise retaining the cast for a few years until (a) there *is*
such a thing as ANSI C, and (b) its use is widespread.  Before that time
there is too much probability of wanting to port back to older compilers.

Once such porting really is not a problem, then I would strongly recommend
removal of the cast.  The statements

	foop = malloc (sizeof *foop);
	foolistp = malloc (n * sizeof *foolistp);

contain NO MENTION of the type of *foop or *foolistp, which makes it
impossible to get it wrong.

Those who say "implicit type conversion is a bug, not a feature" will
disagree.  My opinion is that it is a feature.

A followup suggested that the cast is required in C++ and should there-
fore be used in C.  My opinion is that C++ should "therefore" be fixed.

-- 
Mark Brader		    "'A matter of opinion'[?]  I have to say you are
SoftQuad Inc., Toronto	      right.  There['s] your opinion, which is wrong,
utzoo!sq!msb, msb@sq.com      and mine, which is right."  -- Gene Ward Smith

This article is in the public domain.