Implicit conversion from float -12.719999999999999 to int loses precision