Implicit conversion from float -30.80665813060179 to int loses precision