4396272: Parsing doubles fails to follow IEEE for largest decimal that should yield 0
7039391: Use Math.ulp in FloatingDecimal
Summary: Correct rounding and monotonicity problems in floats and doubles
Reviewed-by: bpb, martin
Contributed-by: NDmitry Nadezhin &lt;dmitry.nadezhin@oracle.com&gt;, Louis Wasserman <lowasser@google.com>