Implicit conversion from float -110.2659574468085 to int loses precision