Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Path: utzoo!utgpu!water!watmath!clyde!ima!haddock!karl
From: karl@haddock.UUCP
Newsgroups: comp.lang.c
Subject: Re: volatile question
Message-ID: <464@haddock.UUCP>
Date: Tue, 28-Apr-87 13:06:25 EDT
Article-I.D.: haddock.464
Posted: Tue Apr 28 13:06:25 1987
Date-Received: Wed, 29-Apr-87 05:49:54 EDT
References: <7124@brl-adm.ARPA>
Reply-To: karl@haddock.ISC.COM.UUCP (Karl Heuer)
Organization: Interactive Systems, Boston
Lines: 19

In article <7124@brl-adm.ARPA> rbbb@rice.EDU (David Chase) writes:
>How does one make the distinction (in C, as proposed by ANSI) between "a
>volatile pointer to something" and "a pointer to volatile something"?  An
>example of a pointer to a volatile is the address of a memory-mapped
>device register; an example of a volatile pointer is an extern referenced
>in a signal handler.

I believe the former is "something * volatile foo" whereas the latter is
"volatile something * foo".  The same syntactic distinction is made for the
"const" attribute.  I think it looks ugly and confusing.  (But I used to have
trouble with prefix/postfix ++ when I was learning C, and I got that straight;
maybe this is no worse.)

In a program I wrote yesterday, I needed a pointer to a volatile pointer to
char.  (The pointer resided in shared memory, its referent did not.)  This
would seem to require "char * volatile * foo".  I think "typedef char *str;
volatile str * foo" should have worked too, but the compiler didn't like it.

Karl W. Z. Heuer (ima!haddock!karl or karl@haddock.isc.com), The Walking Lint