Converting an integer to float and back to integer can change the value due to the typical accurarcy errors.

The fraction of a float64 has only 56 bits IEEE 754 format but the int64 has 64 bit for the "fraction" because there is no "exponent".

Test program:

func main() {
	i := 922337203685477580
	f := float64(i)
	fmt.Println(i)
	fmt.Println(f)
	fmt.Println(int(f))
}

Output:

922337203685477580
9.223372036854776e+17
922337203685477632

No. All Int64 require the same resources (64 bits).

Hey. I have a question about efficieny of assigning math.MaxInt64. Do you know how does it impact resources consumption?

I did notice this when implementing the encoder, but I didn't use it :)