Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Posting-Version: version B 2.10.2 9/18/84; site harvard.ARPA
Path: utzoo!watmath!clyde!burl!ulysses!allegra!mit-eddie!godot!harvard!breuel
From: breuel@harvard.ARPA (Thomas M. Breuel)
Newsgroups: net.lang.c
Subject: Re: limitations of cast, pointers, &c...
Message-ID: <124@harvard.ARPA>
Date: Thu, 1-Nov-84 08:52:47 EST
Article-I.D.: harvard.124
Posted: Thu Nov  1 08:52:47 1984
Date-Received: Sat, 3-Nov-84 01:50:54 EST
Distribution: net
Organization: Aiken Computation Laboratory, Harvard
Lines: 23

Thanks for your replies. Casting the lhs of an assignment might be
defined in an unambiguous way as, for example:

(type)x = y; <=> *(type *)&x = y;

which would be useful and unambiguous, but different in meaning from
a cast as it is usually used. It might also be defined as:

(ytype)x = y; <=> x = (xtype)y;

where xtype and ytype refer to the types of x and y respectively, and
where the construct is undefined if the cast on the lhs is not to the type
of the expression on the rhs.

In a similar manner it would also be possible to fill 'typedef ref *ref;'
with meaning. At any rate, replacing the dereferencing operator
'*' by '#define deref(x) *(ref *)x' gives the type 'ref' the desired
properties and thereby solves my problem.

				Yours sincerely,
				Thomas.

				(breuel@harvard.arpa)