Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Posting-Version: version B 2.10.2 9/18/84; site tymix.UUCP
Path: utzoo!watmath!clyde!burl!ulysses!gamma!epsilon!zeta!sabre!petrus!bellcore!decvax!decwrl!glacier!oliveb!tymix!kanner
From: kanner@tymix.UUCP (Herb Kanner)
Newsgroups: net.lang.c
Subject: Re: Right shift vs. divide
Message-ID: <605@tymix.UUCP>
Date: Wed, 25-Dec-85 16:09:09 EST
Article-I.D.: tymix.605
Posted: Wed Dec 25 16:09:09 1985
Date-Received: Sat, 28-Dec-85 00:57:34 EST
References: <974@brl-tgr.ARPA>
Reply-To: kanner@tymix.UUCP (Herb Kanner)
Organization: Tymnet Inc., Cupertino CA
Lines: 20

In article <974@brl-tgr.ARPA> Schauble@mit-multics.arpa (Paul Schauble) writes:

>BTW, is >> defined to be a sign-extended shift or a zero extended shift
>and under what circumstances.
>  
I'm doing this at home from memory, without a copy of K & R at hand, so,
please, no flames if I get it slightly wrong.

My recollection of their definition of >> is that if the shifted variable
is of type unsigned, it must be a zero-extended shift.  If the shifted
variable is of type int, it may be sign extended if the underlying
hardware possesses such an instruction--I think this is a relatively
non-commital way of saying that it will probably be implemented as a
sign-extended shift on machines using two-complement notation for negative
numbers.

-- 
Herb Kanner
Tymnet, Inc.
...!hplabs!oliveb!tymix!kanner