If int64 is the default type for int, then any int will be able to contain values not representable as a float64 (you'd need a float80 or float128 to make it work). To safely convert, a run-time check will be needed.