Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Path: utzoo!watmath!clyde!burl!ulysses!mhuxl!ihnp4!zehntel!hplabs!sri-unix!lwa@mit-mrclean.ARPA
From: lwa@mit-mrclean.ARPA
Newsgroups: net.unix-wizards
Subject: more on BCPL and C
Message-ID: <1347@sri-arpa.UUCP>
Date: Wed, 27-Jun-84 21:51:00 EDT
Article-I.D.: sri-arpa.1347
Posted: Wed Jun 27 21:51:00 1984
Date-Received: Sun, 1-Jul-84 03:45:12 EDT
Lines: 27

From:  Larry Allen <lwa@mit-mrclean.ARPA>

While we're on the subject of BCPL, I've always missed one of the BCPL
constructs which never found its way into C: the "valof-resultis"
construction.  This allowed any compound statement to return a value
and hence be used in an expression.  As I recall, the basic syntax
was something like:

	variable = valof {
		<statement list>
		resultis (result)
	}

and the statement list could contain declarations, etc.

This construction would be especially useful in conjunction with the
C preprocessor's macro capability, as it would allow one to build
macros with their own local variables and with moderately sophisticated
control structure.  One difficulty with imlementing it in C is that
the type of the result needs to be declared somewhere (not a problem
in BCPL!)

Any comments?
					-Larry Allen

PS. This is probably not the appropriate newsgroup for this, but from the
Internet it's the only one I get, so...