beta 20.1 and 19 and 18 'rint' function with gcc is broken

[Jeff Deifik <jdeifik at weasel dot com>]

I have reported a bug with the math function rint.
It is still not fixed with the latest beta-20.1

I am running windows nt 4.0 server, sp3.
I have a library regression test that I run on the compilers I use.
It showed a bug with the 'rint' function.
This function is supposed to convert a double to an integral double
value. It isn't working, as the enclosed test program shows.

weasel_in_a_box:/e/j/util/test:505: uname -a
CYGWIN_NT-4.0 WEASEL_IN_A_BOX 20.1 (0.3/1/1) 1998-12-3 20:39:18 i686 unknown

Compile with no options, i.e.
gcc yy.c
yy

The output should be nothing, but it is:
47.4 47.125000
47.6 47.625000
-47.4 -47.125000
-47.6 -47.625000

	Jeff Deifik


/* y.c, demonstrates a bug with gnu-w32 beta 18 and 19 and 20.1 */
/* Written 11/06/96 by Jeff Deifik	*/
/* Modified 05/11/97 by Jeff Deifik	*/

#include	<stdio.h>
#include	<math.h>

typedef enum { False = 0, True = 1 } Boolean;

float FLOAT_TO_ROUNDED_FLOAT(float);
int flo_close(double value,double f,double clo);

int main(int argc, char *argv[])
{
Boolean	fail = False;
double	d;
float	f;

    d = rint(47);
    if (!flo_close(47.0f,d,0.0001f))
      { (void)printf("47 %f\n",d); fail = True; }

    f = FLOAT_TO_ROUNDED_FLOAT(47.0);
    if (!flo_close(47.0f,f,0.0001f))
      { (void)printf("47 %f\n",f); fail = True; }

    d = rint(47.4);
    if (!flo_close(47.0f,d,0.0001f)) { (void)printf("47.4 %f\n",d); fail =
True; }
    d = rint(47.6);
    if (!flo_close(48.0f,d,0.0001f)) { (void)printf("47.6 %f\n",d); fail =
True; }
    d = rint(-47);
    if (!flo_close(-47.0f,d,0.0001f)) { (void)printf("-47 %f\n",d); fail =
True; }
    d = rint(-47.4);
    if (!flo_close(-47.0f,d,0.0001f)) { (void)printf("-47.4 %f\n",d); fail
= True; }
    d = rint(-47.6);
    if (!flo_close(-48.0f,d,0.0001f)) { (void)printf("-47.6 %f\n",d); fail
= True; }

    return 0;
}


/* Value is the ideal number, f is what it really is.*/
int	flo_close(double value,double f,double clo)
{
    return(((f+clo) >= value) && ((f-clo) <= value));
}

float	FLOAT_TO_ROUNDED_FLOAT(float x)		{ return(floor(x+0.5)); }
-
For help on using this list (especially unsubscribing), send a message to
"gnu-win32-request@cygnus.com" with one line of text: "help".