Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Posting-Version: version B 2.10.2 9/18/84 SMI; site sun.uucp
Path: utzoo!linus!decvax!decwrl!sun!guy
From: guy@sun.uucp (Guy Harris)
Newsgroups: net.lang.c
Subject: Re: (void *)
Message-ID: <2500@sun.uucp>
Date: Sat, 27-Jul-85 06:02:53 EDT
Article-I.D.: sun.2500
Posted: Sat Jul 27 06:02:53 1985
Date-Received: Sun, 28-Jul-85 06:21:11 EDT
References: <1035@homxa.UUCP> <921@umcp-cs.UUCP> <2462@sun.uucp> <318@baylor.UUCP>
Organization: Sun Microsystems, Inc.
Lines: 16

> I just realised a major problem with (void *). On machines like the HP-1000
> a pointer to a byte and a pointer to a word don't have the same bit-pattern.
> How are you supposed to deal with this without having a terribly
> in-efficient implementation of 'C'.

This problem already exists in C.  "malloc" returns a "char *"; it requires
work to turn it into an "int *".  After "void *" becomes available, "malloc"
will return a "void *" (which would either be a byte pointer - in which case
the generated code would be exactly the same - or a word pointer, in which
case the code assigning the result to a "char *" would have to convert and
the code assigning it to another kind of pointer probably wouldn't.)

"void *" introduces no problems of this sort that aren't already in C due to
the existence of "char *".

	Guy Harris