Implicit conversion from float -119.84455958549225 to int loses precision