From: Masahiro Tanaka <masa@...>
Date: 2000-09-29T02:05:16+09:00
Subject: [ruby-talk:5177] Re: Proposal: Add constants to Math

Hi,

>From: Masahiro Tanaka <masa@stars.gsfc.nasa.gov>
>Subject: [ruby-talk:5138] Re: Proposal: Add constants to Math

> IMHO, A practical way is, to check IEEE 754 support at the
> configure time, and then define Float::NAN with 0.0/0.0 or
> the bit pattern of NaN.

My idea was to insert such a script like below in configure.in, 
and surround the definition of Float::NaN/Inf with

#ifdef HAVE_IEEE_NAN_INF
... 
#endif

Because some machine raises exception when calculating 0.0/0.0,
and then the program terminates.

Regards,
Masa Tanaka

--- configure.in ---

AC_CACHE_CHECK(whether IEEE754 NaN/Inf support, rb_cv_ieee_nan_inf,
    [AC_TRY_RUN([
int
main()
{
  double zero=0.0, one=1.0, nan, inf;

  nan = zero/zero;
  inf = one/zero;

  if (nan!=nan && inf==inf+1)
    return 0;
  return 1;
}
],
        rb_cv_ieee_nan_inf=yes,
        rb_cv_ieee_nan_inf=no,
        rb_cv_ieee_nan_inf=no)])
if test "$rb_cv_ieee_nan_inf" = yes; then
  AC_DEFINE(HAVE_IEEE_NAN_INF)
fi

---