From mboxrd@z Thu Jan  1 00:00:00 1970
X-Spam-Checker-Version: SpamAssassin 3.4.4 (2020-01-24) on polar.synack.me
X-Spam-Level: 
X-Spam-Status: No, score=-1.9 required=5.0 tests=BAYES_00 autolearn=ham
	autolearn_force=no version=3.4.4
X-Google-Thread: 103376,fc647120984c9ad2
X-Google-Attributes: gid103376,public
X-Google-Language: ENGLISH,ASCII-7-bit
Path: 
 g2news2.google.com!news3.google.com!border1.nntp.dca.giganews.com!nntp.giganews.com!newscon06.news.prodigy.net!prodigy.net!uns-out.usenetserver.com!news.usenetserver.com!pc02.usenetserver.com!news.flashnewsgroups.com-b7.4zTQh5tI3A!not-for-mail
Newsgroups: comp.lang.ada
Subject: Re: Way to use Ada Mod function on floats?
References: <1164987168.477589.240140@j44g2000cwa.googlegroups.com>
 <uu00e363d.fsf@stephe-leake.org>
  <4571ac78_6@news.bluewin.ch>
From: Stephen Leake <stephen_leake@stephe-leake.org>
Date: Sun, 03 Dec 2006 10:14:59 -0500
Message-ID: <u3b7x2eq4.fsf@stephe-leake.org>
User-Agent: Gnus/5.1006 (Gnus v5.10.6) Emacs/21.3 (windows-nt)
Cancel-Lock: sha1:Ksyq4L6aBy1wUQlG2uxobAciMlM=
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
X-Complaints-To: abuse@flashnewsgroups.com
Organization: FlashNewsgroups.com
X-Trace: 815b84572ea2363d2952e03907
Xref: g2news2.google.com comp.lang.ada:7795
Date: 2006-12-03T10:14:59-05:00
List-Id: <comp.lang.ada>

Gautier <gautier@fakeaddress.nil> writes:

> Stephen Leake:
>
>> SAL (http://stephe-leake.org/ada/sal.html) has:
>>    function Modulo (Dividend, Divisor : in Real_Type) return
>> Real_Type
>>    is
>>       --  match names in Ada LRM 4.5.5
>>       A : Real_Type renames Dividend;
>>       B : Real_Type renames Divisor;
>>       N : constant Integer := Integer (Real_Type'Floor (A / B));
>>    begin
>>       return A - B * Real_Type (N);
>>    end Modulo;
>
> Out of curiosity, what is the need of converting Real_Type'Floor (A /
> B) to an Integer and back to a Real_Type ?

Since so many others claim to be able to read my mind, I almost feel
it is unnecessary to answer!

It is a mistake to convert to Integer; my tests for this pass when I
delete that.

This code was first written for Ada 83. Hmm; the Ada 95 LRM says
'Floor _did_ exist then (it is not listed as an extension to Ada 83),
but apparently I was unaware of it, or it didn't exist in the DEC Ada
compiler, or something. So I used 'round to Integer' instead:

function Modulo (Dividend, Divisor : in FLOAT_TYPE) return FLOAT_TYPE
is
    -- match names in LRM 4.5.5

    A : FLOAT_TYPE renames Dividend;
    B : FLOAT_TYPE renames Divisor;
    Quotient : FLOAT_TYPE := Dividend / Divisor;
    N : INT_32_TYPE := INT_32_TYPE (Quotient - 0.5);
begin
    return A - B * FLOAT_TYPE (N);
end Modulo;

Apparently when I converted to using 'Float, I did not fully think it
thru. Thanks for pointing this out.

It is always useful to expose code to review by experts :).

Using 'Truncation gives different sign results; 'Floor matches the
integer definition of "mod":

Mod (10.0, 5.0) :   0.00000
Mod (11.0, 5.0) :   1.00000
Mod (14.0, 5.0) :   4.00000
Mod (10.0, -5.0) :   0.00000
Mod (11.0, -5.0) :  -4.00000
Mod (14.0, -5.0) :  -1.00000
Mod (-10.0, 5.0) :   0.00000
Mod (-11.0, 5.0) :   4.00000
Mod (-14.0, 5.0) :   1.00000
Mod (-10.0, -5.0) :   0.00000
Mod (-11.0, -5.0) :  -1.00000
Mod (-14.0, -5.0) :  -4.00000
Mod (0.0, 2.0) :   0.00000

Perhaps 'Truncation would match "rem"? I haven't looked at it in
detail.

-- 
-- Stephe