Implicit conversion from float -111.49253731343282 to int loses precision