Implicit conversion from float -112.09580838323353 to int loses precision