Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Path: utzoo!linus!decvax!genrad!panda!talcott!harvard!seismo!brl-adm!brl-smoke!smoke!jon@cit-vax.ARPA
From: jon@cit-vax.ARPA (Jonathan P. Leech)
Newsgroups: net.lang.c
Subject: Re: Long Longs
Message-ID: <1434@brl-smoke.ARPA>
Date: Sun, 2-Mar-86 21:24:47 EST
Article-I.D.: brl-smok.1434
Posted: Sun Mar  2 21:24:47 1986
Date-Received: Tue, 4-Mar-86 03:55:23 EST
Sender: news@brl-smoke.ARPA
Lines: 47

>   From:    "Robert D. Silverman" <bs%faron.uucp@BRL.ARPA>
>   Subject: Long Longs
>
>   ... Various reasons for wanting 'long longs' deleted
>
>   Does C++ have long longs? Does anyone know of ANY language that does?
>   (Bignums in Lisp don't qualify)
>
>   Bob Silverman

    Some C compilers have them, especially if  the  hardware  supports
it. I know that C on ELXSI Unix and UTS (Amdahl) Unix has them.   Note
that (even if they were added to the ANSI standard) there  is  nothing
prohibiting the COMPILER from generating function calls  just  as  you
would - many compilers already do this for floating  point  operations
if there is no hardware f.p.  - so there's no guaranteed speedup.

    You could  define  a  class  in  C++  containg  64-bit  longs  and
overload the normal arithmetic operations to work on them; getting  it
to generate the 'proper' code for

	int b, c, d;
	longlong result;
	result = (b * c) / d;

would be more difficult. My limited experience with C++ makes me think
that you would have to define a new class  equivalent  to  int	except
that result of '*' is a longlong (or whatever you want	to  call  it).
C++ could generate such code inline.

    Final cautionary note: [ from K&R Appendix	A  Sec.   7  (pg  185)
Expressions ]:

   "... Expressions involving a commutative and  associative  operator
    (*, +, &, |,  ^)  may  be  rearranged  arbitrarily,  even  in  the
    presence of parentheses; to force a particular order of evaluation
    an explicit temporary must be used."

    I believe the ANSI draft lets  you	specify  order	of  evaluation
with the unary '+' operator, so you would want

    a = +(b * c) / d;

whenever ANSI compilers finally arrive.

    -- Jon Leech (jon@csvax.caltech.edu)
    __@/