On Sun, 10 Feb 2008, Jim Razmus wrote:
> I'm told that math.h should do this for me. Moreover, I think NAN is a
> machine dependent value.
See /usr/include/i386/ieee.h for some hints.
> Adding the line you mention would break on VAX (assuming I understand
> this correctly). Although I don't think anyone would run this program
> on a VAX, if it's in the ports tree there's the possibility.
Probably. Vaxes didn't use IEEE floating point.
> Worst case, I could add those defines though.
I'd do some more research. Examine the source code for "isnan()"
This is the i386 version
#include <sys/types.h>
#include <machine/ieee.h>
#include <math.h>
int
isnan(d)
double d;
{
struct ieee_double *p = (struct ieee_double *)&d;
return (p->dbl_exp == DBL_EXP_INFNAN &&
(p->dbl_frach != 0 || p->dbl_fracl != 0));
}
We notice the || in the comparison. There is more than one NaN,
in other words. DBL_EXP_INFNAN is defined in ieee.h
So there is no unique "NaN".
Dave
--
The president of the United States is the commander-in-chief of
the armed forces. He is not the commander-in-chief of the
government, nor is he the commander-in-chief of the country.
