Path: utzoo!utgpu!news-server.csri.toronto.edu!cs.utexas.edu!uwm.edu!ogicse!milton!sumax!amc-gw!mikem
From: mikem@amc.com (Mike McGinnis)
Newsgroups: comp.os.minix
Subject: definition of NULL
Message-ID: <1991Jun25.172459.1142@amc.com>
Date: 25 Jun 91 17:24:59 GMT
Reply-To: mikem@anabol.amc.com ()
Organization: Applied Microsystems, Redmond, WA
Lines: 55


>>*** BUG FIX ***
>>
>>In /usr/include/stdio.h, #define NULL 0
>>In /usr/include/*.h (all others), #define NULL ((void *)0)
>>
>>stdio.h should be updated asap.
>>
>>If I'm wrong about this, or if this will break anything, please let me
>>know.
>
>You are absolutely right. When I made this change, I was able to
>compile cdungeon (ie. Zork) in 16-bit mode with c68 on the Atari ST.
>Before, I had to use Gnu C in 32-bit mode. Let's make this change
>official...
>
>
>David
>

You are absolutely wrong ;^).

OK, One More Time...
This is your brain on (int)0, this is your brain on (pointer)0...

NULL should always be defined as the integer 0. It is the compilers job to
infer the context and treat it accordingly in every instance with the exception
of function parameters; it should then be cast explicitly to the type defined
in the called function i.e. "(char *)NULL".
C compiler writers go to great pains to make this so with pointers and zero.
Please utilize it accordingly.

In fun, to see this context sensitivity, try this:
{
    char *c;

    if (c == 0)	/* integer 0 will draw no warning */
	;
    if (c == 1)	/* integer 1 will draw warning */
	;
}

The VAX compiler produces the following warning for the comparison with 1, but
no warning for the comparison of zero.

foo.c, line 7: warning: illegal combination of pointer and integer, op ==

Most compilers will probably draw a warning on the comparison with 1,
especially if it sizeof(int) != sizeof(pointer).
If it draws a warning on the comparison with 0, it is a weak compiler.

	Michael E. McGinnis
	Applied Microsystems Corporation, Redmond Washington.

"Help Mr. Wizard... I don't want to be an engineer anymore!"