D  A float of long-double precision in native format.
   (Long doubles are available only if your system supports
   long double values. Raises an exception otherwise.
   Note that there are different long double formats.)