Implicit conversion from float -119.2250922509225 to int loses precision