Implicit conversion from float -79.51219512195124 to int loses precision