Implicit conversion from float -118.45921450151059 to int loses precision