Path: utzoo!attcan!uunet!lll-winken!lll-tis!helios.ee.lbl.gov!pasteur!ames!haven!adm!smoke!gwyn
From: gwyn@smoke.BRL.MIL (Doug Gwyn )
Newsgroups: comp.lang.c
Subject: Re: question about shift operator
Message-ID: <8820@smoke.BRL.MIL>
Date: 2 Nov 88 17:43:23 GMT
References: <786@gtx.com>
Reply-To: gwyn@brl.arpa (Doug Gwyn (VLD/VMB) <gwyn>)
Distribution: na
Organization: Ballistic Research Lab (BRL), APG, MD.
Lines: 23

In article <786@gtx.com> randy@gtx.UUCP (Randy D. Miller) writes:
>     n <<= 32;
>K&R 1st ed. seems to indicate the operation is undefined, but does
>the proposed standard change this?  

It's undefined because the intention is that a raw machine shift
instruction be usable to implement this, and usually such a shift
count cannot be represented in the instruction.  I did see an
architecture once that was unable to represent a 0 shift count,
or rather you could build such an instruction but it didn't work
right!  However, the proposed C standard allows <<0 and >>0.

(By the way, if the count is written as a constant, the compiler
can sometimes use alternate code sequences, e.g.
	n <<= 16
becomes
	SWAP_HALVES
	CLEAR_LOWER
and
	n <<= 100
could become
	CLEAR
It is when a general expression is used for the count that things
get interesting.)