From: Phil Galbiati <Philip DOT S DOT Galbiati AT Tek DOT com>
Newsgroups: comp.os.msdos.djgpp
Subject: Re: sqrt() problem
Date: Mon, 09 Dec 1996 19:37:08 -0800
Organization: Tektronix
Lines: 84
Message-ID: <32ACDAE4.33FB@Tek.com>
References: <32aa47fd DOT 8683086 AT nntp DOT southeast DOT net> <32AABA23 DOT 15FC AT cs DOT com> <32ab4462 DOT 73306026 AT nntp DOT southeast DOT net> <32AB49F3 DOT 65AF AT cornell DOT edu>
NNTP-Posting-Host: philipga.cse.tek.com
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
To: murray AT southeast DOT net
CC: "A. Sinan Unur" <asu1 AT cornell DOT edu>
DJ-Gateway: from newsgroup comp.os.msdos.djgpp

A. Sinan Unur wrote:
> 
> Murray Stokely wrote:
> > it seems sqrt is returning a realy
> > weird number
>   ^^^^^
>   no it is not.
> 
> >
> >         s=sqrt((double)(r*r) - (double)(m*m));
> >         // also used pow(x,2) - same result
> >         printf("%d\n",s);                    ^^^
You are attempting to print a double (8-byte floating point data)
as a 4-byte decimal integer.

> >         // also used %i - same result
> >
> > r and m are integers (10 and 50), yet s keeps returning a huge
> > negative number (should be like 51 or something)
> >
> > Murray Stokely ( murray AT southeast DOT net )
> 
> do you realize that 10*10 - 50*50 = -2400 and there is no real
> number whose square will give you -2400? you are feeding sqrt
> a totally invalid argument. GIGO (garbage-in-garbage-out).

<SNIP>

> the ansi standard states that "...a domain error occurs if the
> argument is negative." elsewhere, it states that "On a domain
> error, the function returns an implementation-defined value;
> the value of the macro EDOM is stored in errno.


The fact that he (Murray) got anything to print suggests to
me that what he posted was *not* what he compiled, since
sqrt()ing a negative number should cause a floating point
exception (at least it does on my machine), and unless he has
a handler for SIGFPE, his program would have crashed before
printing the "huge negative number" he described. I *suspect*
that he compiled:

     s = sqrt((double)(r*r) + (double)(m*m));
                           ^^^

Change the '-' to a '+', and fix the printf() statement and it
should work better. If you make both changes it should print:

50.990195

Here's the whole program with changes hi-lited

/********************* CUT HERE *******************/
#include <stdio.h>
#include <math.h>

int
main ()
{
    double s;
    int r;
    int m;

    r = 10;
    m = 50;

    s = sqrt((double)(r*r) + (double)(m*m));
/*                        ^^^                     */
    printf("%lf\n",s);
/*           ^^^                                  */
    return (0);
}
/********************* CUT HERE *******************/

Compile like this:

   gxx -ofoo foo.cpp

Hope this helps.
--Phil Galbiati
=================================================
     Any opinions expressed here reflect the
     ignorance of the author, NOT Tektronix.
=================================================