From:	j DOT aldrich6 AT genie DOT com
Message-Id:	<199607220024.AA286965065@relay1.geis.com>
Date:	Mon, 22 Jul 96 00:03:00 UTC 0000
To:	dclayton AT dircon DOT co DOT uk
Cc:	djgpp AT delorie DOT com
Mime-Version:	1.0
Subject:	Re:Float to integer conversion

Reply to message 0521777    from DCLAYTON AT DIRC on 07/21/96  5:39PM


>Could somebody please tell me the best way to change an integer number to a
>float, and vice versa?

Depends on what you need it for.  Here's a sample program which
demonstrates some basic (and not so basic) concepts of typecasting:

#include <stdio.h>
#include <math.h>    /* defines double sqrt( double n ); */

struct foo
{
    char name[20];
    long socsec;
    char dummy;
};

int main( void )
{
    int x = 10;
    float y = 5.55;
    struct foo bar;
    char *bytes;

    x = (int) y;
    y = (float) x;
    y = (float) sqrt( (double) x );

    printf( " x = %d\n (float) x = %f\n (char) x = %c\n (long double *) x =
%p\n",
        x, (float) x, (char) x, (long double *) x );

    strcpy( bar.name, "John Adams" );
    bar.socsec = 123456789;
    bar.dummy = '\0';

    bytes = (char *) &bar;
    printf( "%s\n", bytes );

    printf( "The social security number is %ld\n", ((struct foo *)
bytes)->socsec );

    bytes = (char *) malloc( 16 * 1024 * 1024 * sizeof *bytes );
    printf( "Address of allocated memory = %p\n", bytes );

    return 0;
}

WARNING:  Anything more complex than char-int-float typecasting
should be examined very carefully to make sure it won't damage
anything!  For example, if I had not created a 'dummy' field in the
above example and set it to zero, the "printf( bytes );" line would
have done some very strange things.  See if you can figure out why.  :)

John

- Raw text -