gcc supports an extended precison "long double" type typically 10 or 12 bytes
but Windows XP always maps the "long double" to the standard double (typically 8 bytes)
MinGW on windows XP thinks it has a 12-byte "long double" but in fact it only has the 8 byte support on Windows XP. any long double math only has 8 byte precision and must be recast and printed with the standard double floating point format specifier (%f)
does anyone know if Windows Vista or Win 7 has extended precision for long doubles, or does it merely map them to the standard double?