This will better preserve a default behavior of the more commonly used int being encoded as an int32 instead of being encoded as an int64 right now. This will make it easier to not pass around an EncodeContext with MinSize set. This should cover negative and positive values.
It might be worth decoding ints in the same vein as well.