Implicit conversion from float -119.68992248062014 to int loses precision