Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Posting-Version: version B 2.10.1 6/24/83; site rabbit.UUCP
Path: utzoo!watmath!clyde!burl!ulysses!allegra!alice!rabbit!ark
From: ark@rabbit.UUCP (Andrew Koenig)
Newsgroups: net.lang.c
Subject: neophyte's pointer question
Message-ID: <2692@rabbit.UUCP>
Date: Tue, 10-Apr-84 13:54:45 EST
Article-I.D.: rabbit.2692
Posted: Tue Apr 10 13:54:45 1984
Date-Received: Wed, 11-Apr-84 07:02:36 EST
Organization: AT&T Bell Laboratories, Murray Hill
Lines: 24

If s is a character pointer then

	*s++ = (s & 0x7f);

is definitely undefined, as the meaning of the "and" operator applied
to a pointer is undefined (i. e. implementation dependent).  Perhaps the
question was intended this way:  Is it OK to write:

	*s++ = (*s & 0x7f);

The answer here is still no, because there is no guarantee as to
whether s will be incremented before or after the right-hand-side
is evaluated.  If the author assumed that the right-hand-side would
be evaluated first, the result desired would be to turn off all but
the low-order seven bits of the character addressed by s and then
increment s.  This can be done as follows:

	*s++ &= 0x7f;

Alternatively, one can write:

	*s = (*s & 0x7f);		/* parens not really needed */
	s++;