Implicit conversion from float -110.46082949308754 to int loses precision