Implicit conversion from float -58.82260596546308 to int loses precision