Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Path: utzoo!mnetor!uunet!husc6!cmcl2!rutgers!princeton!udel!rochester!PT.CS.CMU.EDU!H.GP.CS.CMU.EDU!cef
From: cef@H.GP.CS.CMU.EDU (Charles Fineman)
Newsgroups: comp.lang.c
Subject: Re: void declarations and casts
Message-ID: <220@PT.CS.CMU.EDU>
Date: Fri, 23-Oct-87 03:43:51 EST
Article-I.D.: PT.220
Posted: Fri Oct 23 03:43:51 1987
Date-Received: Sun, 25-Oct-87 15:11:34 EST
References: <359@targon.UUCP>
Sender: netnews@PT.CS.CMU.EDU
Organization: Carnegie-Mellon University, CS/RI
Lines: 36
Keywords: void, casts, declarations, assignments
Summary: Of *course* it wont work!!!

First to the declaration:
  *Whenever* you try to declare something that the C compiler does not
 know the size of, it complains. For example, you can say:

	struct foo;

  as a forward declaration and then say:

	struct foo *ptr;

  later in your code to get a pointer to a structure of type foo
  but you can't say:

	struct foo var;

  because it dosn't know the size of the thing. Its the same exact
  thing for the void type. Since C dosn't know what size a thing
  of type void should be, you can't declare a variable of that
  type. 

    **  One iteresting side note: there was a discusion a while 
    ** ago on some newsgroup about some folks declaring:
    **     typedef void *generic_pointer;

   As for casting to type void, there is no problem here as long 
 as you realize that the *specified* semantics of this is to
 throw away the value. Hence, casting anything to void makes it
 *not* an rvalue (i.e. something we get a value of).

   Sure, you program is sytactically correct, but then so was 
 that bogus fortran program that sent that poor satalite
 careening off to where no man has gone before ;-)

	Charles Fineman
	Carnegie-Mellon University
	cef@h.cs.cmu.edu (via seismo)