From: horst DOT kraemer AT t-online DOT de (Horst Kraemer)
Newsgroups: comp.os.msdos.djgpp
Subject: Re: new vs malloc
Date: Sun, 20 Feb 2000 00:43:08 GMT
Organization: T-Online
Lines: 57
Message-ID: <38aebc31.19427507@news.btx.dtag.de>
References: <38AD8622 DOT AE067F97 AT tiscalinet DOT it> <004101bf7a97$06ff2ca0$cff0fea9 AT stevenhe> <88lt6e$9pj$1 AT nets3 DOT rz DOT RWTH-Aachen DOT DE> <01bf7ba9$8fb5a980$c3247d81 AT default>
Mime-Version: 1.0
Content-Type: text/plain; charset=ISO-8859-1
Content-Transfer-Encoding: 8bit
X-Trace: news07.btx.dtag.de 951007354 19017 0306239354-0001 000220 00:42:34
X-Complaints-To: abuse AT t-online DOT de
X-Sender: 0306239354-0001 AT t-dialin DOT net
X-Newsreader: Forte Free Agent 1.11/32.235
To: djgpp AT delorie DOT com
DJ-Gateway: from newsgroup comp.os.msdos.djgpp
Reply-To: djgpp AT delorie DOT com

On 19 Feb 2000 13:44:37 GMT, "doug" <deleveld AT dds DOT nl> wrote:

> > In a nutshell: if you write C++, use new. If you write C, use
> > malloc(), and *don't* cast the return value, as it will actually hide
> > potential problems, instead of fixing any.
> 
> Can you give me an example.   I like my C programs to compile
> with as few warnings as possible, even if they are compiled as C++

Sorry. If you compile a program "as C++" it is a C++ program and not a
C program.

And there is no point compiling a valid C program "as C++" although
since Charles Petzold (in his excellent book "Programming windows")
recommended compiling C programs "as C++" people ruminate that this is
GOOD...

Imho the only profit you gain is that you have to do the malloc-cast
which is only a syntactical "ceremony" without any practical
usefulness.

But if you prefer C++ then stick to C++ by all means - but don't call
your programs C programs ;-) C and C++ are distinct languages with a
lot of similarities. And this "lot" is the dangerous part because it
makes you forget that there _are_ differences.

> and I cast the returned value from malloc to achieve this.  Can you
> give me an example of where this casting may cause problems in
> a C program?

int main()
{
  char* p = malloc(50);
}

Now you will get a diagnostic from a C compiler. As there is no
declaration for malloc in scope (because you forgot to #include
stdlib.h) the compiler will assume that malloc returns an int and
there is no implicit conversion from int to char*.

If you cast

    char* p = (char*)malloc(50);

there is no diagnostic. The compiler still assumes that malloc returns
an int and assumes that the value returned by malloc is an int to be
casted to a char*. This is desastrous if int and char* have different
sizes and/or representations.

In C++ you have no option because you always have to cast explicitly.
The stdlib problem doesn't exist in C++ because in C++ it is an error
to call malloc without #including stdlib, while it is permitted in C.


Regards
Horst