Implicit conversion from float -43.219178082191775 to int loses precision