Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Posting-Version: version B 2.10.3 4.3bsd-beta 6/6/85; site hoptoad.uucp
Path: utzoo!watmath!clyde!burl!ulysses!allegra!mit-eddie!genrad!decvax!decwrl!sun!hoptoad!laura
From: laura@hoptoad.uucp (Laura Creighton)
Newsgroups: net.lang.c
Subject: Re: Long Longs
Message-ID: <603@hoptoad.uucp>
Date: Sun, 9-Mar-86 23:55:05 EST
Article-I.D.: hoptoad.603
Posted: Sun Mar  9 23:55:05 1986
Date-Received: Wed, 12-Mar-86 21:44:29 EST
References: <491@faron.UUCP> <7000005@convexs>
Reply-To: laura@hoptoad.UUCP (Laura Creighton)
Organization: Nebula Consultants in San Francisco
Lines: 21

In article <7000005@convexs> hosking@convexs.UUCP writes:
>We ran into exactly this sort of problem with the Convex C-1.  It has 64 bit
>registers, but we didn't want to make "long" be 64 bits, for a number of
>reasons, including "portability" with existing code which assumed that
>sizeof(long) == sizeof(int) (Argh!!), and performance impact on typical
>applications.
>/* Danger, Will Robinson!  Flames from hoptoad!laura approaching! */
>We therefore support "short" "int" "long" and "long long" - where
>sizeof(long long) == 64 bits.

Nope -- the flame from hoptoad!laura comes right after yours --
beside the (Arghh!!).  I think that you did the right thing, given
your constraints.  It is the fact that one of your constraints was
``we can hire someone for a year to take the (sizeof int == sizeof long)s
out of 4.2, and delay shipping for a year, or we can work around it.''

 
-- 
Laura Creighton		
ihnp4!hoptoad!laura  utzoo!hoptoad!laura  sun!hoptoad!laura
toad@lll-crg.arpa