Implicit conversion from float -115.57377049180326 to int loses precision