Implicit conversion from float -79.66580976863753 to int loses precision