use std::borrow::Cow;
use std::default::Default;
use std::iter::repeat;
use std::ops::{Add, BitAnd, BitOr, BitXor, Div, Mul, Neg, Rem, Shl, Shr, Sub,
AddAssign, BitAndAssign, BitOrAssign, BitXorAssign, DivAssign,
MulAssign, RemAssign, ShlAssign, ShrAssign, SubAssign};
use std::str::{self, FromStr};
use std::fmt;
use std::cmp;
use std::mem;
use std::cmp::Ordering::{self, Less, Greater, Equal};
use std::{f32, f64};
use std::{u8, u64};
#[allow(unused_imports)]
use std::ascii::AsciiExt;
#[cfg(feature = "serde")]
use serde;
use integer::Integer;
use traits::{ToPrimitive, FromPrimitive, Float, Num, Unsigned, CheckedAdd, CheckedSub, CheckedMul,
CheckedDiv, Zero, One};
#[path = "algorithms.rs"]
mod algorithms;
#[path = "monty.rs"]
mod monty;
pub use self::algorithms::big_digit;
pub use self::big_digit::{BigDigit, DoubleBigDigit, ZERO_BIG_DIGIT};
use self::algorithms::{mac_with_carry, mul3, scalar_mul, div_rem, div_rem_digit};
use self::algorithms::{__add2, add2, sub2, sub2rev};
use self::algorithms::{biguint_shl, biguint_shr};
use self::algorithms::{cmp_slice, fls, ilog2};
use self::monty::monty_modpow;
use UsizePromotion;
use ParseBigIntError;
#[cfg(test)]
#[path = "tests/biguint.rs"]
mod biguint_tests;
#[derive(Clone, Debug, Hash)]
#[cfg_attr(feature = "rustc-serialize", derive(RustcEncodable, RustcDecodable))]
pub struct BigUint {
data: Vec<BigDigit>,
}
impl PartialEq for BigUint {
#[inline]
fn eq(&self, other: &BigUint) -> bool {
match self.cmp(other) {
Equal => true,
_ => false,
}
}
}
impl Eq for BigUint {}
impl PartialOrd for BigUint {
#[inline]
fn partial_cmp(&self, other: &BigUint) -> Option<Ordering> {
Some(self.cmp(other))
}
}
impl Ord for BigUint {
#[inline]
fn cmp(&self, other: &BigUint) -> Ordering {
cmp_slice(&self.data[..], &other.data[..])
}
}
impl Default for BigUint {
#[inline]
fn default() -> BigUint {
Zero::zero()
}
}
impl fmt::Display for BigUint {
fn fmt(&self, f: &mut fmt::Formatter) -> fmt::Result {
f.pad_integral(true, "", &self.to_str_radix(10))
}
}
impl fmt::LowerHex for BigUint {
fn fmt(&self, f: &mut fmt::Formatter) -> fmt::Result {
f.pad_integral(true, "0x", &self.to_str_radix(16))
}
}
impl fmt::UpperHex for BigUint {
fn fmt(&self, f: &mut fmt::Formatter) -> fmt::Result {
f.pad_integral(true, "0x", &self.to_str_radix(16).to_ascii_uppercase())
}
}
impl fmt::Binary for BigUint {
fn fmt(&self, f: &mut fmt::Formatter) -> fmt::Result {
f.pad_integral(true, "0b", &self.to_str_radix(2))
}
}
impl fmt::Octal for BigUint {
fn fmt(&self, f: &mut fmt::Formatter) -> fmt::Result {
f.pad_integral(true, "0o", &self.to_str_radix(8))
}
}
impl FromStr for BigUint {
type Err = ParseBigIntError;
#[inline]
fn from_str(s: &str) -> Result<BigUint, ParseBigIntError> {
BigUint::from_str_radix(s, 10)
}
}
fn from_bitwise_digits_le(v: &[u8], bits: usize) -> BigUint {
debug_assert!(!v.is_empty() && bits <= 8 && big_digit::BITS % bits == 0);
debug_assert!(v.iter().all(|&c| (c as BigDigit) < (1 << bits)));
let digits_per_big_digit = big_digit::BITS / bits;
let data = v.chunks(digits_per_big_digit)
.map(|chunk| {
chunk.iter().rev().fold(0, |acc, &c| (acc << bits) | c as BigDigit)
})
.collect();
BigUint::new(data)
}
fn from_inexact_bitwise_digits_le(v: &[u8], bits: usize) -> BigUint {
debug_assert!(!v.is_empty() && bits <= 8 && big_digit::BITS % bits != 0);
debug_assert!(v.iter().all(|&c| (c as BigDigit) < (1 << bits)));
let big_digits = (v.len() * bits + big_digit::BITS - 1) / big_digit::BITS;
let mut data = Vec::with_capacity(big_digits);
let mut d = 0;
let mut dbits = 0;
for &c in v {
d |= (c as BigDigit) << dbits;
dbits += bits;
if dbits >= big_digit::BITS {
data.push(d);
dbits -= big_digit::BITS;
d = (c as BigDigit) >> (bits - dbits);
}
}
if dbits > 0 {
debug_assert!(dbits < big_digit::BITS);
data.push(d as BigDigit);
}
BigUint::new(data)
}
fn from_radix_digits_be(v: &[u8], radix: u32) -> BigUint {
debug_assert!(!v.is_empty() && !radix.is_power_of_two());
debug_assert!(v.iter().all(|&c| (c as u32) < radix));
let bits = (radix as f64).log2() * v.len() as f64;
let big_digits = (bits / big_digit::BITS as f64).ceil();
let mut data = Vec::with_capacity(big_digits as usize);
let (base, power) = get_radix_base(radix);
let radix = radix as BigDigit;
let r = v.len() % power;
let i = if r == 0 {
power
} else {
r
};
let (head, tail) = v.split_at(i);
let first = head.iter().fold(0, |acc, &d| acc * radix + d as BigDigit);
data.push(first);
debug_assert!(tail.len() % power == 0);
for chunk in tail.chunks(power) {
if data.last() != Some(&0) {
data.push(0);
}
let mut carry = 0;
for d in data.iter_mut() {
*d = mac_with_carry(0, *d, base, &mut carry);
}
debug_assert!(carry == 0);
let n = chunk.iter().fold(0, |acc, &d| acc * radix + d as BigDigit);
add2(&mut data, &[n]);
}
BigUint::new(data)
}
impl Num for BigUint {
type FromStrRadixErr = ParseBigIntError;
fn from_str_radix(s: &str, radix: u32) -> Result<BigUint, ParseBigIntError> {
assert!(2 <= radix && radix <= 36, "The radix must be within 2...36");
let mut s = s;
if s.starts_with('+') {
let tail = &s[1..];
if !tail.starts_with('+') {
s = tail
}
}
if s.is_empty() {
let e = u64::from_str_radix(s, radix).unwrap_err();
return Err(e.into());
}
if s.starts_with('_') {
let e = u64::from_str_radix(s, radix).unwrap_err();
return Err(e.into());
}
let mut v = Vec::with_capacity(s.len());
for b in s.bytes() {
let d = match b {
b'0'...b'9' => b - b'0',
b'a'...b'z' => b - b'a' + 10,
b'A'...b'Z' => b - b'A' + 10,
b'_' => continue,
_ => u8::MAX,
};
if d < radix as u8 {
v.push(d);
} else {
let i = cmp::max(v.len(), 1) - 1;
let e = u64::from_str_radix(&s[i..], radix).unwrap_err();
return Err(e.into());
}
}
let res = if radix.is_power_of_two() {
let bits = ilog2(radix);
v.reverse();
if big_digit::BITS % bits == 0 {
from_bitwise_digits_le(&v, bits)
} else {
from_inexact_bitwise_digits_le(&v, bits)
}
} else {
from_radix_digits_be(&v, radix)
};
Ok(res)
}
}
forward_all_binop_to_val_ref_commutative!(impl BitAnd for BigUint, bitand);
forward_val_assign!(impl BitAndAssign for BigUint, bitand_assign);
impl<'a> BitAnd<&'a BigUint> for BigUint {
type Output = BigUint;
#[inline]
fn bitand(mut self, other: &BigUint) -> BigUint {
self &= other;
self
}
}
impl<'a> BitAndAssign<&'a BigUint> for BigUint {
#[inline]
fn bitand_assign(&mut self, other: &BigUint) {
for (ai, &bi) in self.data.iter_mut().zip(other.data.iter()) {
*ai &= bi;
}
self.data.truncate(other.data.len());
self.normalize();
}
}
forward_all_binop_to_val_ref_commutative!(impl BitOr for BigUint, bitor);
forward_val_assign!(impl BitOrAssign for BigUint, bitor_assign);
impl<'a> BitOr<&'a BigUint> for BigUint {
type Output = BigUint;
fn bitor(mut self, other: &BigUint) -> BigUint {
self |= other;
self
}
}
impl<'a> BitOrAssign<&'a BigUint> for BigUint {
#[inline]
fn bitor_assign(&mut self, other: &BigUint) {
for (ai, &bi) in self.data.iter_mut().zip(other.data.iter()) {
*ai |= bi;
}
if other.data.len() > self.data.len() {
let extra = &other.data[self.data.len()..];
self.data.extend(extra.iter().cloned());
}
}
}
forward_all_binop_to_val_ref_commutative!(impl BitXor for BigUint, bitxor);
forward_val_assign!(impl BitXorAssign for BigUint, bitxor_assign);
impl<'a> BitXor<&'a BigUint> for BigUint {
type Output = BigUint;
fn bitxor(mut self, other: &BigUint) -> BigUint {
self ^= other;
self
}
}
impl<'a> BitXorAssign<&'a BigUint> for BigUint {
#[inline]
fn bitxor_assign(&mut self, other: &BigUint) {
for (ai, &bi) in self.data.iter_mut().zip(other.data.iter()) {
*ai ^= bi;
}
if other.data.len() > self.data.len() {
let extra = &other.data[self.data.len()..];
self.data.extend(extra.iter().cloned());
}
self.normalize();
}
}
impl Shl<usize> for BigUint {
type Output = BigUint;
#[inline]
fn shl(self, rhs: usize) -> BigUint {
biguint_shl(Cow::Owned(self), rhs)
}
}
impl<'a> Shl<usize> for &'a BigUint {
type Output = BigUint;
#[inline]
fn shl(self, rhs: usize) -> BigUint {
biguint_shl(Cow::Borrowed(self), rhs)
}
}
impl ShlAssign<usize> for BigUint {
#[inline]
fn shl_assign(&mut self, rhs: usize) {
*self = biguint_shl(Cow::Borrowed(&*self), rhs);
}
}
impl Shr<usize> for BigUint {
type Output = BigUint;
#[inline]
fn shr(self, rhs: usize) -> BigUint {
biguint_shr(Cow::Owned(self), rhs)
}
}
impl<'a> Shr<usize> for &'a BigUint {
type Output = BigUint;
#[inline]
fn shr(self, rhs: usize) -> BigUint {
biguint_shr(Cow::Borrowed(self), rhs)
}
}
impl ShrAssign<usize> for BigUint {
#[inline]
fn shr_assign(&mut self, rhs: usize) {
let n = mem::replace(self, BigUint::zero());
*self = n >> rhs;
}
}
impl Zero for BigUint {
#[inline]
fn zero() -> BigUint {
BigUint::new(Vec::new())
}
#[inline]
fn is_zero(&self) -> bool {
self.data.is_empty()
}
}
impl One for BigUint {
#[inline]
fn one() -> BigUint {
BigUint::new(vec![1])
}
}
impl Unsigned for BigUint {}
forward_all_binop_to_val_ref_commutative!(impl Add for BigUint, add);
forward_val_assign!(impl AddAssign for BigUint, add_assign);
impl<'a> Add<&'a BigUint> for BigUint {
type Output = BigUint;
fn add(mut self, other: &BigUint) -> BigUint {
self += other;
self
}
}
impl<'a> AddAssign<&'a BigUint> for BigUint {
#[inline]
fn add_assign(&mut self, other: &BigUint) {
if self.data.len() < other.data.len() {
let extra = other.data.len() - self.data.len();
self.data.extend(repeat(0).take(extra));
}
let carry = __add2(&mut self.data[..], &other.data[..]);
if carry != 0 {
self.data.push(carry);
}
}
}
promote_unsigned_scalars!(impl Add for BigUint, add);
promote_unsigned_scalars_assign!(impl AddAssign for BigUint, add_assign);
forward_all_scalar_binop_to_val_val_commutative!(impl Add<BigDigit> for BigUint, add);
forward_all_scalar_binop_to_val_val_commutative!(impl Add<DoubleBigDigit> for BigUint, add);
impl Add<BigDigit> for BigUint {
type Output = BigUint;
#[inline]
fn add(mut self, other: BigDigit) -> BigUint {
self += other;
self
}
}
impl AddAssign<BigDigit> for BigUint {
#[inline]
fn add_assign(&mut self, other: BigDigit) {
if other != 0 {
if self.data.len() == 0 {
self.data.push(0);
}
let carry = __add2(&mut self.data, &[other]);
if carry != 0 {
self.data.push(carry);
}
}
}
}
impl Add<DoubleBigDigit> for BigUint {
type Output = BigUint;
#[inline]
fn add(mut self, other: DoubleBigDigit) -> BigUint {
self += other;
self
}
}
impl AddAssign<DoubleBigDigit> for BigUint {
#[inline]
fn add_assign(&mut self, other: DoubleBigDigit) {
let (hi, lo) = big_digit::from_doublebigdigit(other);
if hi == 0 {
*self += lo;
} else {
while self.data.len() < 2 {
self.data.push(0);
}
let carry = __add2(&mut self.data, &[lo, hi]);
if carry != 0 {
self.data.push(carry);
}
}
}
}
forward_val_val_binop!(impl Sub for BigUint, sub);
forward_ref_ref_binop!(impl Sub for BigUint, sub);
forward_val_assign!(impl SubAssign for BigUint, sub_assign);
impl<'a> Sub<&'a BigUint> for BigUint {
type Output = BigUint;
fn sub(mut self, other: &BigUint) -> BigUint {
self -= other;
self
}
}
impl<'a> SubAssign<&'a BigUint> for BigUint {
fn sub_assign(&mut self, other: &'a BigUint) {
sub2(&mut self.data[..], &other.data[..]);
self.normalize();
}
}
impl<'a> Sub<BigUint> for &'a BigUint {
type Output = BigUint;
fn sub(self, mut other: BigUint) -> BigUint {
if other.data.len() < self.data.len() {
let extra = self.data.len() - other.data.len();
other.data.extend(repeat(0).take(extra));
}
sub2rev(&self.data[..], &mut other.data[..]);
other.normalized()
}
}
promote_unsigned_scalars!(impl Sub for BigUint, sub);
promote_unsigned_scalars_assign!(impl SubAssign for BigUint, sub_assign);
forward_all_scalar_binop_to_val_val!(impl Sub<BigDigit> for BigUint, sub);
forward_all_scalar_binop_to_val_val!(impl Sub<DoubleBigDigit> for BigUint, sub);
impl Sub<BigDigit> for BigUint {
type Output = BigUint;
#[inline]
fn sub(mut self, other: BigDigit) -> BigUint {
self -= other;
self
}
}
impl SubAssign<BigDigit> for BigUint {
fn sub_assign(&mut self, other: BigDigit) {
sub2(&mut self.data[..], &[other]);
self.normalize();
}
}
impl Sub<BigUint> for BigDigit {
type Output = BigUint;
#[inline]
fn sub(self, mut other: BigUint) -> BigUint {
if other.data.len() == 0 {
other.data.push(0);
}
sub2rev(&[self], &mut other.data[..]);
other.normalized()
}
}
impl Sub<DoubleBigDigit> for BigUint {
type Output = BigUint;
#[inline]
fn sub(mut self, other: DoubleBigDigit) -> BigUint {
self -= other;
self
}
}
impl SubAssign<DoubleBigDigit> for BigUint {
fn sub_assign(&mut self, other: DoubleBigDigit) {
let (hi, lo) = big_digit::from_doublebigdigit(other);
sub2(&mut self.data[..], &[lo, hi]);
self.normalize();
}
}
impl Sub<BigUint> for DoubleBigDigit {
type Output = BigUint;
#[inline]
fn sub(self, mut other: BigUint) -> BigUint {
while other.data.len() < 2 {
other.data.push(0);
}
let (hi, lo) = big_digit::from_doublebigdigit(self);
sub2rev(&[lo, hi], &mut other.data[..]);
other.normalized()
}
}
forward_all_binop_to_ref_ref!(impl Mul for BigUint, mul);
forward_val_assign!(impl MulAssign for BigUint, mul_assign);
impl<'a, 'b> Mul<&'b BigUint> for &'a BigUint {
type Output = BigUint;
#[inline]
fn mul(self, other: &BigUint) -> BigUint {
mul3(&self.data[..], &other.data[..])
}
}
impl<'a> MulAssign<&'a BigUint> for BigUint {
#[inline]
fn mul_assign(&mut self, other: &'a BigUint) {
*self = &*self * other
}
}
promote_unsigned_scalars!(impl Mul for BigUint, mul);
promote_unsigned_scalars_assign!(impl MulAssign for BigUint, mul_assign);
forward_all_scalar_binop_to_val_val_commutative!(impl Mul<BigDigit> for BigUint, mul);
forward_all_scalar_binop_to_val_val_commutative!(impl Mul<DoubleBigDigit> for BigUint, mul);
impl Mul<BigDigit> for BigUint {
type Output = BigUint;
#[inline]
fn mul(mut self, other: BigDigit) -> BigUint {
self *= other;
self
}
}
impl MulAssign<BigDigit> for BigUint {
#[inline]
fn mul_assign(&mut self, other: BigDigit) {
if other == 0 {
self.data.clear();
} else {
let carry = scalar_mul(&mut self.data[..], other);
if carry != 0 {
self.data.push(carry);
}
}
}
}
impl Mul<DoubleBigDigit> for BigUint {
type Output = BigUint;
#[inline]
fn mul(mut self, other: DoubleBigDigit) -> BigUint {
self *= other;
self
}
}
impl MulAssign<DoubleBigDigit> for BigUint {
#[inline]
fn mul_assign(&mut self, other: DoubleBigDigit) {
if other == 0 {
self.data.clear();
} else if other <= BigDigit::max_value() as DoubleBigDigit {
*self *= other as BigDigit
} else {
let (hi, lo) = big_digit::from_doublebigdigit(other);
*self = mul3(&self.data[..], &[lo, hi])
}
}
}
forward_all_binop_to_ref_ref!(impl Div for BigUint, div);
forward_val_assign!(impl DivAssign for BigUint, div_assign);
impl<'a, 'b> Div<&'b BigUint> for &'a BigUint {
type Output = BigUint;
#[inline]
fn div(self, other: &BigUint) -> BigUint {
let (q, _) = self.div_rem(other);
q
}
}
impl<'a> DivAssign<&'a BigUint> for BigUint {
#[inline]
fn div_assign(&mut self, other: &'a BigUint) {
*self = &*self / other;
}
}
promote_unsigned_scalars!(impl Div for BigUint, div);
promote_unsigned_scalars_assign!(impl DivAssign for BigUint, div_assign);
forward_all_scalar_binop_to_val_val!(impl Div<BigDigit> for BigUint, div);
forward_all_scalar_binop_to_val_val!(impl Div<DoubleBigDigit> for BigUint, div);
impl Div<BigDigit> for BigUint {
type Output = BigUint;
#[inline]
fn div(self, other: BigDigit) -> BigUint {
let (q, _) = div_rem_digit(self, other);
q
}
}
impl DivAssign<BigDigit> for BigUint {
#[inline]
fn div_assign(&mut self, other: BigDigit) {
*self = &*self / other;
}
}
impl Div<BigUint> for BigDigit {
type Output = BigUint;
#[inline]
fn div(self, other: BigUint) -> BigUint {
match other.data.len() {
0 => panic!(),
1 => From::from(self / other.data[0]),
_ => Zero::zero(),
}
}
}
impl Div<DoubleBigDigit> for BigUint {
type Output = BigUint;
#[inline]
fn div(self, other: DoubleBigDigit) -> BigUint {
let (q, _) = self.div_rem(&From::from(other));
q
}
}
impl DivAssign<DoubleBigDigit> for BigUint {
#[inline]
fn div_assign(&mut self, other: DoubleBigDigit) {
*self = &*self / other;
}
}
impl Div<BigUint> for DoubleBigDigit {
type Output = BigUint;
#[inline]
fn div(self, other: BigUint) -> BigUint {
match other.data.len() {
0 => panic!(),
1 => From::from(self / other.data[0] as u64),
2 => From::from(self / big_digit::to_doublebigdigit(other.data[1], other.data[0])),
_ => Zero::zero(),
}
}
}
forward_all_binop_to_ref_ref!(impl Rem for BigUint, rem);
forward_val_assign!(impl RemAssign for BigUint, rem_assign);
impl<'a, 'b> Rem<&'b BigUint> for &'a BigUint {
type Output = BigUint;
#[inline]
fn rem(self, other: &BigUint) -> BigUint {
let (_, r) = self.div_rem(other);
r
}
}
impl<'a> RemAssign<&'a BigUint> for BigUint {
#[inline]
fn rem_assign(&mut self, other: &BigUint) {
*self = &*self % other;
}
}
promote_unsigned_scalars!(impl Rem for BigUint, rem);
promote_unsigned_scalars_assign!(impl RemAssign for BigUint, rem_assign);
forward_all_scalar_binop_to_val_val!(impl Rem<BigDigit> for BigUint, rem);
forward_all_scalar_binop_to_val_val!(impl Rem<DoubleBigDigit> for BigUint, rem);
impl Rem<BigDigit> for BigUint {
type Output = BigUint;
#[inline]
fn rem(self, other: BigDigit) -> BigUint {
let (_, r) = div_rem_digit(self, other);
From::from(r)
}
}
impl RemAssign<BigDigit> for BigUint {
#[inline]
fn rem_assign(&mut self, other: BigDigit) {
*self = &*self % other;
}
}
impl Rem<BigUint> for BigDigit {
type Output = BigUint;
#[inline]
fn rem(mut self, other: BigUint) -> BigUint {
self %= other;
From::from(self)
}
}
macro_rules! impl_rem_assign_scalar {
($scalar:ty, $to_scalar:ident) => {
forward_val_assign_scalar!(impl RemAssign for BigUint, $scalar, rem_assign);
impl<'a> RemAssign<&'a BigUint> for $scalar {
#[inline]
fn rem_assign(&mut self, other: &BigUint) {
*self = match other.$to_scalar() {
None => *self,
Some(0) => panic!(),
Some(v) => *self % v
};
}
}
}
}
impl_rem_assign_scalar!(usize, to_usize);
impl_rem_assign_scalar!(u64, to_u64);
impl_rem_assign_scalar!(u32, to_u32);
impl_rem_assign_scalar!(u16, to_u16);
impl_rem_assign_scalar!(u8, to_u8);
impl_rem_assign_scalar!(isize, to_isize);
impl_rem_assign_scalar!(i64, to_i64);
impl_rem_assign_scalar!(i32, to_i32);
impl_rem_assign_scalar!(i16, to_i16);
impl_rem_assign_scalar!(i8, to_i8);
impl Rem<DoubleBigDigit> for BigUint {
type Output = BigUint;
#[inline]
fn rem(self, other: DoubleBigDigit) -> BigUint {
let (_, r) = self.div_rem(&From::from(other));
r
}
}
impl RemAssign<DoubleBigDigit> for BigUint {
#[inline]
fn rem_assign(&mut self, other: DoubleBigDigit) {
*self = &*self % other;
}
}
impl Rem<BigUint> for DoubleBigDigit {
type Output = BigUint;
#[inline]
fn rem(mut self, other: BigUint) -> BigUint {
self %= other;
From::from(self)
}
}
impl Neg for BigUint {
type Output = BigUint;
#[inline]
fn neg(self) -> BigUint {
panic!()
}
}
impl<'a> Neg for &'a BigUint {
type Output = BigUint;
#[inline]
fn neg(self) -> BigUint {
panic!()
}
}
impl CheckedAdd for BigUint {
#[inline]
fn checked_add(&self, v: &BigUint) -> Option<BigUint> {
return Some(self.add(v));
}
}
impl CheckedSub for BigUint {
#[inline]
fn checked_sub(&self, v: &BigUint) -> Option<BigUint> {
match self.cmp(v) {
Less => None,
Equal => Some(Zero::zero()),
Greater => Some(self.sub(v)),
}
}
}
impl CheckedMul for BigUint {
#[inline]
fn checked_mul(&self, v: &BigUint) -> Option<BigUint> {
return Some(self.mul(v));
}
}
impl CheckedDiv for BigUint {
#[inline]
fn checked_div(&self, v: &BigUint) -> Option<BigUint> {
if v.is_zero() {
return None;
}
return Some(self.div(v));
}
}
impl Integer for BigUint {
#[inline]
fn div_rem(&self, other: &BigUint) -> (BigUint, BigUint) {
div_rem(self, other)
}
#[inline]
fn div_floor(&self, other: &BigUint) -> BigUint {
let (d, _) = div_rem(self, other);
d
}
#[inline]
fn mod_floor(&self, other: &BigUint) -> BigUint {
let (_, m) = div_rem(self, other);
m
}
#[inline]
fn div_mod_floor(&self, other: &BigUint) -> (BigUint, BigUint) {
div_rem(self, other)
}
#[inline]
fn gcd(&self, other: &Self) -> Self {
if self.is_zero() {
return other.clone();
}
if other.is_zero() {
return self.clone();
}
let mut m = self.clone();
let mut n = other.clone();
let shift = cmp::min(
n.trailing_zeros(),
m.trailing_zeros()
);
n >>= n.trailing_zeros();
while !m.is_zero() {
m >>= m.trailing_zeros();
if n > m { mem::swap(&mut n, &mut m) }
m -= &n;
}
n << shift
}
#[inline]
fn lcm(&self, other: &BigUint) -> BigUint {
self / self.gcd(other) * other
}
#[inline]
fn divides(&self, other: &BigUint) -> bool {
self.is_multiple_of(other)
}
#[inline]
fn is_multiple_of(&self, other: &BigUint) -> bool {
(self % other).is_zero()
}
#[inline]
fn is_even(&self) -> bool {
match self.data.first() {
Some(x) => x.is_even(),
None => true,
}
}
#[inline]
fn is_odd(&self) -> bool {
!self.is_even()
}
}
fn high_bits_to_u64(v: &BigUint) -> u64 {
match v.data.len() {
0 => 0,
1 => v.data[0] as u64,
_ => {
let mut bits = v.bits();
let mut ret = 0u64;
let mut ret_bits = 0;
for d in v.data.iter().rev() {
let digit_bits = (bits - 1) % big_digit::BITS + 1;
let bits_want = cmp::min(64 - ret_bits, digit_bits);
if bits_want != 64 {
ret <<= bits_want;
}
ret |= *d as u64 >> (digit_bits - bits_want);
ret_bits += bits_want;
bits -= bits_want;
if ret_bits == 64 {
break;
}
}
ret
}
}
}
impl ToPrimitive for BigUint {
#[inline]
fn to_i64(&self) -> Option<i64> {
self.to_u64().and_then(|n| {
if n >> 63 == 0 {
Some(n as i64)
} else {
None
}
})
}
#[inline]
fn to_u64(&self) -> Option<u64> {
let mut ret: u64 = 0;
let mut bits = 0;
for i in self.data.iter() {
if bits >= 64 {
return None;
}
ret += (*i as u64) << bits;
bits += big_digit::BITS;
}
Some(ret)
}
#[inline]
fn to_f32(&self) -> Option<f32> {
let mantissa = high_bits_to_u64(self);
let exponent = self.bits() - fls(mantissa);
if exponent > f32::MAX_EXP as usize {
None
} else {
let ret = (mantissa as f32) * 2.0f32.powi(exponent as i32);
if ret.is_infinite() {
None
} else {
Some(ret)
}
}
}
#[inline]
fn to_f64(&self) -> Option<f64> {
let mantissa = high_bits_to_u64(self);
let exponent = self.bits() - fls(mantissa);
if exponent > f64::MAX_EXP as usize {
None
} else {
let ret = (mantissa as f64) * 2.0f64.powi(exponent as i32);
if ret.is_infinite() {
None
} else {
Some(ret)
}
}
}
}
impl FromPrimitive for BigUint {
#[inline]
fn from_i64(n: i64) -> Option<BigUint> {
if n >= 0 {
Some(BigUint::from(n as u64))
} else {
None
}
}
#[inline]
fn from_u64(n: u64) -> Option<BigUint> {
Some(BigUint::from(n))
}
#[inline]
fn from_f64(mut n: f64) -> Option<BigUint> {
if !n.is_finite() {
return None;
}
n = n.trunc();
if n.is_zero() {
return Some(BigUint::zero());
}
let (mantissa, exponent, sign) = Float::integer_decode(n);
if sign == -1 {
return None;
}
let mut ret = BigUint::from(mantissa);
if exponent > 0 {
ret = ret << exponent as usize;
} else if exponent < 0 {
ret = ret >> (-exponent) as usize;
}
Some(ret)
}
}
impl From<u64> for BigUint {
#[inline]
fn from(mut n: u64) -> Self {
let mut ret: BigUint = Zero::zero();
while n != 0 {
ret.data.push(n as BigDigit);
n = (n >> 1) >> (big_digit::BITS - 1);
}
ret
}
}
macro_rules! impl_biguint_from_uint {
($T:ty) => {
impl From<$T> for BigUint {
#[inline]
fn from(n: $T) -> Self {
BigUint::from(n as u64)
}
}
}
}
impl_biguint_from_uint!(u8);
impl_biguint_from_uint!(u16);
impl_biguint_from_uint!(u32);
impl_biguint_from_uint!(usize);
pub trait ToBigUint {
fn to_biguint(&self) -> Option<BigUint>;
}
impl ToBigUint for BigUint {
#[inline]
fn to_biguint(&self) -> Option<BigUint> {
Some(self.clone())
}
}
macro_rules! impl_to_biguint {
($T:ty, $from_ty:path) => {
impl ToBigUint for $T {
#[inline]
fn to_biguint(&self) -> Option<BigUint> {
$from_ty(*self)
}
}
}
}
impl_to_biguint!(isize, FromPrimitive::from_isize);
impl_to_biguint!(i8, FromPrimitive::from_i8);
impl_to_biguint!(i16, FromPrimitive::from_i16);
impl_to_biguint!(i32, FromPrimitive::from_i32);
impl_to_biguint!(i64, FromPrimitive::from_i64);
impl_to_biguint!(usize, FromPrimitive::from_usize);
impl_to_biguint!(u8, FromPrimitive::from_u8);
impl_to_biguint!(u16, FromPrimitive::from_u16);
impl_to_biguint!(u32, FromPrimitive::from_u32);
impl_to_biguint!(u64, FromPrimitive::from_u64);
impl_to_biguint!(f32, FromPrimitive::from_f32);
impl_to_biguint!(f64, FromPrimitive::from_f64);
fn to_bitwise_digits_le(u: &BigUint, bits: usize) -> Vec<u8> {
debug_assert!(!u.is_zero() && bits <= 8 && big_digit::BITS % bits == 0);
let last_i = u.data.len() - 1;
let mask: BigDigit = (1 << bits) - 1;
let digits_per_big_digit = big_digit::BITS / bits;
let digits = (u.bits() + bits - 1) / bits;
let mut res = Vec::with_capacity(digits);
for mut r in u.data[..last_i].iter().cloned() {
for _ in 0..digits_per_big_digit {
res.push((r & mask) as u8);
r >>= bits;
}
}
let mut r = u.data[last_i];
while r != 0 {
res.push((r & mask) as u8);
r >>= bits;
}
res
}
fn to_inexact_bitwise_digits_le(u: &BigUint, bits: usize) -> Vec<u8> {
debug_assert!(!u.is_zero() && bits <= 8 && big_digit::BITS % bits != 0);
let mask: BigDigit = (1 << bits) - 1;
let digits = (u.bits() + bits - 1) / bits;
let mut res = Vec::with_capacity(digits);
let mut r = 0;
let mut rbits = 0;
for c in &u.data {
r |= *c << rbits;
rbits += big_digit::BITS;
while rbits >= bits {
res.push((r & mask) as u8);
r >>= bits;
if rbits > big_digit::BITS {
r = *c >> (big_digit::BITS - (rbits - bits));
}
rbits -= bits;
}
}
if rbits != 0 {
res.push(r as u8);
}
while let Some(&0) = res.last() {
res.pop();
}
res
}
#[inline(always)]
fn to_radix_digits_le(u: &BigUint, radix: u32) -> Vec<u8> {
debug_assert!(!u.is_zero() && !radix.is_power_of_two());
let radix_digits = ((u.bits() as f64) / (radix as f64).log2()).ceil();
let mut res = Vec::with_capacity(radix_digits as usize);
let mut digits = u.clone();
let (base, power) = get_radix_base(radix);
let radix = radix as BigDigit;
while digits.data.len() > 1 {
let (q, mut r) = div_rem_digit(digits, base);
for _ in 0..power {
res.push((r % radix) as u8);
r /= radix;
}
digits = q;
}
let mut r = digits.data[0];
while r != 0 {
res.push((r % radix) as u8);
r /= radix;
}
res
}
pub fn to_radix_le(u: &BigUint, radix: u32) -> Vec<u8> {
if u.is_zero() {
vec![0]
} else if radix.is_power_of_two() {
let bits = ilog2(radix);
if big_digit::BITS % bits == 0 {
to_bitwise_digits_le(u, bits)
} else {
to_inexact_bitwise_digits_le(u, bits)
}
} else if radix == 10 {
to_radix_digits_le(u, 10)
} else {
to_radix_digits_le(u, radix)
}
}
pub fn to_str_radix_reversed(u: &BigUint, radix: u32) -> Vec<u8> {
assert!(2 <= radix && radix <= 36, "The radix must be within 2...36");
if u.is_zero() {
return vec![b'0'];
}
let mut res = to_radix_le(u, radix);
for r in &mut res {
debug_assert!((*r as u32) < radix);
if *r < 10 {
*r += b'0';
} else {
*r += b'a' - 10;
}
}
res
}
impl BigUint {
#[inline]
pub fn new(digits: Vec<BigDigit>) -> BigUint {
BigUint { data: digits }.normalized()
}
#[inline]
pub fn from_slice(slice: &[BigDigit]) -> BigUint {
BigUint::new(slice.to_vec())
}
#[inline]
pub fn assign_from_slice(&mut self, slice: &[BigDigit]) {
self.data.resize(slice.len(), 0);
self.data.clone_from_slice(slice);
self.normalize();
}
#[inline]
pub fn from_bytes_be(bytes: &[u8]) -> BigUint {
if bytes.is_empty() {
Zero::zero()
} else {
let mut v = bytes.to_vec();
v.reverse();
BigUint::from_bytes_le(&*v)
}
}
#[inline]
pub fn from_bytes_le(bytes: &[u8]) -> BigUint {
if bytes.is_empty() {
Zero::zero()
} else {
from_bitwise_digits_le(bytes, 8)
}
}
#[inline]
pub fn parse_bytes(buf: &[u8], radix: u32) -> Option<BigUint> {
str::from_utf8(buf).ok().and_then(|s| BigUint::from_str_radix(s, radix).ok())
}
pub fn from_radix_be(buf: &[u8], radix: u32) -> Option<BigUint> {
assert!(2 <= radix && radix <= 256, "The radix must be within 2...256");
if radix != 256 && buf.iter().any(|&b| b >= radix as u8) {
return None;
}
let res = if radix.is_power_of_two() {
let bits = ilog2(radix);
let mut v = Vec::from(buf);
v.reverse();
if big_digit::BITS % bits == 0 {
from_bitwise_digits_le(&v, bits)
} else {
from_inexact_bitwise_digits_le(&v, bits)
}
} else {
from_radix_digits_be(buf, radix)
};
Some(res)
}
pub fn from_radix_le(buf: &[u8], radix: u32) -> Option<BigUint> {
assert!(2 <= radix && radix <= 256, "The radix must be within 2...256");
if radix != 256 && buf.iter().any(|&b| b >= radix as u8) {
return None;
}
let res = if radix.is_power_of_two() {
let bits = ilog2(radix);
if big_digit::BITS % bits == 0 {
from_bitwise_digits_le(buf, bits)
} else {
from_inexact_bitwise_digits_le(buf, bits)
}
} else {
let mut v = Vec::from(buf);
v.reverse();
from_radix_digits_be(&v, radix)
};
Some(res)
}
#[inline]
pub fn to_bytes_be(&self) -> Vec<u8> {
let mut v = self.to_bytes_le();
v.reverse();
v
}
#[inline]
pub fn to_bytes_le(&self) -> Vec<u8> {
if self.is_zero() {
vec![0]
} else {
to_bitwise_digits_le(self, 8)
}
}
#[inline]
pub fn to_str_radix(&self, radix: u32) -> String {
let mut v = to_str_radix_reversed(self, radix);
v.reverse();
unsafe { String::from_utf8_unchecked(v) }
}
#[inline]
pub fn to_radix_be(&self, radix: u32) -> Vec<u8> {
let mut v = to_radix_le(self, radix);
v.reverse();
v
}
#[inline]
pub fn to_radix_le(&self, radix: u32) -> Vec<u8> {
to_radix_le(self, radix)
}
#[inline]
pub fn bits(&self) -> usize {
if self.is_zero() {
return 0;
}
let zeros = self.data.last().unwrap().leading_zeros();
return self.data.len() * big_digit::BITS - zeros as usize;
}
fn trailing_zeros(&self) -> usize {
self.data
.iter()
.enumerate()
.find(|&(_, &digit)| digit != 0)
.map(|(i, digit)| i * big_digit::BITS + digit.trailing_zeros() as usize)
.unwrap_or(0)
}
#[inline]
fn normalize(&mut self) {
while let Some(&0) = self.data.last() {
self.data.pop();
}
}
#[inline]
fn normalized(mut self) -> BigUint {
self.normalize();
self
}
pub fn modpow(&self, exponent: &Self, modulus: &Self) -> Self {
assert!(!modulus.is_zero(), "divide by zero!");
if modulus.is_odd() {
return monty_modpow(self, exponent, modulus);
}
let one = BigUint::one();
if exponent.is_zero() { return one; }
let mut base = self % modulus;
let mut exp = exponent.clone();
while exp.is_even() {
base = &base * &base % modulus;
exp >>= 1;
}
if exp == one { return base }
let mut acc = base.clone();
while exp > one {
exp >>= 1;
base = &base * &base % modulus;
if exp.is_odd() {
acc = acc * &base % modulus;
}
}
acc
}
}
#[cfg(feature = "serde")]
impl serde::Serialize for BigUint {
fn serialize<S>(&self, serializer: &mut S) -> Result<(), S::Error>
where S: serde::Serializer
{
self.data.serialize(serializer)
}
}
#[cfg(feature = "serde")]
impl serde::Deserialize for BigUint {
fn deserialize<D>(deserializer: &mut D) -> Result<Self, D::Error>
where D: serde::Deserializer
{
let data = try!(Vec::deserialize(deserializer));
Ok(BigUint { data: data })
}
}
#[inline]
fn get_radix_base(radix: u32) -> (BigDigit, usize) {
debug_assert!(2 <= radix && radix <= 256, "The radix must be within 2...256");
debug_assert!(!radix.is_power_of_two());
match big_digit::BITS {
32 => {
const BASES: [(u32, usize); 257] = [
( 0, 0),
( 0, 0),
( 0, 0),
(3486784401, 20),
( 0, 0),
(1220703125, 13),
(2176782336, 12),
(1977326743, 11),
( 0, 0),
(3486784401, 10),
(1000000000, 9),
(2357947691, 9),
( 429981696, 8),
( 815730721, 8),
(1475789056, 8),
(2562890625, 8),
( 0, 0),
( 410338673, 7),
( 612220032, 7),
( 893871739, 7),
(1280000000, 7),
(1801088541, 7),
(2494357888, 7),
(3404825447, 7),
( 191102976, 6),
( 244140625, 6),
( 308915776, 6),
( 387420489, 6),
( 481890304, 6),
( 594823321, 6),
( 729000000, 6),
( 887503681, 6),
( 0, 0),
(1291467969, 6),
(1544804416, 6),
(1838265625, 6),
(2176782336, 6),
(2565726409, 6),
(3010936384, 6),
(3518743761, 6),
(4096000000, 6),
( 115856201, 5),
( 130691232, 5),
( 147008443, 5),
( 164916224, 5),
( 184528125, 5),
( 205962976, 5),
( 229345007, 5),
( 254803968, 5),
( 282475249, 5),
( 312500000, 5),
( 345025251, 5),
( 380204032, 5),
( 418195493, 5),
( 459165024, 5),
( 503284375, 5),
( 550731776, 5),
( 601692057, 5),
( 656356768, 5),
( 714924299, 5),
( 777600000, 5),
( 844596301, 5),
( 916132832, 5),
( 992436543, 5),
( 0, 0),
(1160290625, 5),
(1252332576, 5),
(1350125107, 5),
(1453933568, 5),
(1564031349, 5),
(1680700000, 5),
(1804229351, 5),
(1934917632, 5),
(2073071593, 5),
(2219006624, 5),
(2373046875, 5),
(2535525376, 5),
(2706784157, 5),
(2887174368, 5),
(3077056399, 5),
(3276800000, 5),
(3486784401, 5),
(3707398432, 5),
(3939040643, 5),
(4182119424, 5),
( 52200625, 4),
( 54700816, 4),
( 57289761, 4),
( 59969536, 4),
( 62742241, 4),
( 65610000, 4),
( 68574961, 4),
( 71639296, 4),
( 74805201, 4),
( 78074896, 4),
( 81450625, 4),
( 84934656, 4),
( 88529281, 4),
( 92236816, 4),
( 96059601, 4),
( 100000000, 4),
( 104060401, 4),
( 108243216, 4),
( 112550881, 4),
( 116985856, 4),
( 121550625, 4),
( 126247696, 4),
( 131079601, 4),
( 136048896, 4),
( 141158161, 4),
( 146410000, 4),
( 151807041, 4),
( 157351936, 4),
( 163047361, 4),
( 168896016, 4),
( 174900625, 4),
( 181063936, 4),
( 187388721, 4),
( 193877776, 4),
( 200533921, 4),
( 207360000, 4),
( 214358881, 4),
( 221533456, 4),
( 228886641, 4),
( 236421376, 4),
( 244140625, 4),
( 252047376, 4),
( 260144641, 4),
( 0, 0),
( 276922881, 4),
( 285610000, 4),
( 294499921, 4),
( 303595776, 4),
( 312900721, 4),
( 322417936, 4),
( 332150625, 4),
( 342102016, 4),
( 352275361, 4),
( 362673936, 4),
( 373301041, 4),
( 384160000, 4),
( 395254161, 4),
( 406586896, 4),
( 418161601, 4),
( 429981696, 4),
( 442050625, 4),
( 454371856, 4),
( 466948881, 4),
( 479785216, 4),
( 492884401, 4),
( 506250000, 4),
( 519885601, 4),
( 533794816, 4),
( 547981281, 4),
( 562448656, 4),
( 577200625, 4),
( 592240896, 4),
( 607573201, 4),
( 623201296, 4),
( 639128961, 4),
( 655360000, 4),
( 671898241, 4),
( 688747536, 4),
( 705911761, 4),
( 723394816, 4),
( 741200625, 4),
( 759333136, 4),
( 777796321, 4),
( 796594176, 4),
( 815730721, 4),
( 835210000, 4),
( 855036081, 4),
( 875213056, 4),
( 895745041, 4),
( 916636176, 4),
( 937890625, 4),
( 959512576, 4),
( 981506241, 4),
(1003875856, 4),
(1026625681, 4),
(1049760000, 4),
(1073283121, 4),
(1097199376, 4),
(1121513121, 4),
(1146228736, 4),
(1171350625, 4),
(1196883216, 4),
(1222830961, 4),
(1249198336, 4),
(1275989841, 4),
(1303210000, 4),
(1330863361, 4),
(1358954496, 4),
(1387488001, 4),
(1416468496, 4),
(1445900625, 4),
(1475789056, 4),
(1506138481, 4),
(1536953616, 4),
(1568239201, 4),
(1600000000, 4),
(1632240801, 4),
(1664966416, 4),
(1698181681, 4),
(1731891456, 4),
(1766100625, 4),
(1800814096, 4),
(1836036801, 4),
(1871773696, 4),
(1908029761, 4),
(1944810000, 4),
(1982119441, 4),
(2019963136, 4),
(2058346161, 4),
(2097273616, 4),
(2136750625, 4),
(2176782336, 4),
(2217373921, 4),
(2258530576, 4),
(2300257521, 4),
(2342560000, 4),
(2385443281, 4),
(2428912656, 4),
(2472973441, 4),
(2517630976, 4),
(2562890625, 4),
(2608757776, 4),
(2655237841, 4),
(2702336256, 4),
(2750058481, 4),
(2798410000, 4),
(2847396321, 4),
(2897022976, 4),
(2947295521, 4),
(2998219536, 4),
(3049800625, 4),
(3102044416, 4),
(3154956561, 4),
(3208542736, 4),
(3262808641, 4),
(3317760000, 4),
(3373402561, 4),
(3429742096, 4),
(3486784401, 4),
(3544535296, 4),
(3603000625, 4),
(3662186256, 4),
(3722098081, 4),
(3782742016, 4),
(3844124001, 4),
(3906250000, 4),
(3969126001, 4),
(4032758016, 4),
(4097152081, 4),
(4162314256, 4),
(4228250625, 4),
( 0, 0),
];
let (base, power) = BASES[radix as usize];
(base as BigDigit, power)
}
64 => {
const BASES: [(u64, usize); 257] = [
( 0, 0),
( 0, 0),
( 9223372036854775808, 63),
(12157665459056928801, 40),
( 4611686018427387904, 31),
( 7450580596923828125, 27),
( 4738381338321616896, 24),
( 3909821048582988049, 22),
( 9223372036854775808, 21),
(12157665459056928801, 20),
(10000000000000000000, 19),
( 5559917313492231481, 18),
( 2218611106740436992, 17),
( 8650415919381337933, 17),
( 2177953337809371136, 16),
( 6568408355712890625, 16),
( 1152921504606846976, 15),
( 2862423051509815793, 15),
( 6746640616477458432, 15),
(15181127029874798299, 15),
( 1638400000000000000, 14),
( 3243919932521508681, 14),
( 6221821273427820544, 14),
(11592836324538749809, 14),
( 876488338465357824, 13),
( 1490116119384765625, 13),
( 2481152873203736576, 13),
( 4052555153018976267, 13),
( 6502111422497947648, 13),
(10260628712958602189, 13),
(15943230000000000000, 13),
( 787662783788549761, 12),
( 1152921504606846976, 12),
( 1667889514952984961, 12),
( 2386420683693101056, 12),
( 3379220508056640625, 12),
( 4738381338321616896, 12),
( 6582952005840035281, 12),
( 9065737908494995456, 12),
(12381557655576425121, 12),
(16777216000000000000, 12),
( 550329031716248441, 11),
( 717368321110468608, 11),
( 929293739471222707, 11),
( 1196683881290399744, 11),
( 1532278301220703125, 11),
( 1951354384207722496, 11),
( 2472159215084012303, 11),
( 3116402981210161152, 11),
( 3909821048582988049, 11),
( 4882812500000000000, 11),
( 6071163615208263051, 11),
( 7516865509350965248, 11),
( 9269035929372191597, 11),
(11384956040305711104, 11),
(13931233916552734375, 11),
(16985107389382393856, 11),
( 362033331456891249, 10),
( 430804206899405824, 10),
( 511116753300641401, 10),
( 604661760000000000, 10),
( 713342911662882601, 10),
( 839299365868340224, 10),
( 984930291881790849, 10),
( 1152921504606846976, 10),
( 1346274334462890625, 10),
( 1568336880910795776, 10),
( 1822837804551761449, 10),
( 2113922820157210624, 10),
( 2446194060654759801, 10),
( 2824752490000000000, 10),
( 3255243551009881201, 10),
( 3743906242624487424, 10),
( 4297625829703557649, 10),
( 4923990397355877376, 10),
( 5631351470947265625, 10),
( 6428888932339941376, 10),
( 7326680472586200649, 10),
( 8335775831236199424, 10),
( 9468276082626847201, 10),
(10737418240000000000, 10),
(12157665459056928801, 10),
(13744803133596058624, 10),
(15516041187205853449, 10),
(17490122876598091776, 10),
( 231616946283203125, 9),
( 257327417311663616, 9),
( 285544154243029527, 9),
( 316478381828866048, 9),
( 350356403707485209, 9),
( 387420489000000000, 9),
( 427929800129788411, 9),
( 472161363286556672, 9),
( 520411082988487293, 9),
( 572994802228616704, 9),
( 630249409724609375, 9),
( 692533995824480256, 9),
( 760231058654565217, 9),
( 833747762130149888, 9),
( 913517247483640899, 9),
( 1000000000000000000, 9),
( 1093685272684360901, 9),
( 1195092568622310912, 9),
( 1304773183829244583, 9),
( 1423311812421484544, 9),
( 1551328215978515625, 9),
( 1689478959002692096, 9),
( 1838459212420154507, 9),
( 1999004627104432128, 9),
( 2171893279442309389, 9),
( 2357947691000000000, 9),
( 2558036924386500591, 9),
( 2773078757450186752, 9),
( 3004041937984268273, 9),
( 3251948521156637184, 9),
( 3517876291919921875, 9),
( 3802961274698203136, 9),
( 4108400332687853397, 9),
( 4435453859151328768, 9),
( 4785448563124474679, 9),
( 5159780352000000000, 9),
( 5559917313492231481, 9),
( 5987402799531080192, 9),
( 6443858614676334363, 9),
( 6930988311686938624, 9),
( 7450580596923828125, 9),
( 8004512848309157376, 9),
( 8594754748609397887, 9),
( 9223372036854775808, 9),
( 9892530380752880769, 9),
(10604499373000000000, 9),
(11361656654439817571, 9),
(12166492167065567232, 9),
(13021612539908538853, 9),
(13929745610903012864, 9),
(14893745087865234375, 9),
(15916595351771938816, 9),
(17001416405572203977, 9),
(18151468971815029248, 9),
( 139353667211683681, 8),
( 147578905600000000, 8),
( 156225851787813921, 8),
( 165312903998914816, 8),
( 174859124550883201, 8),
( 184884258895036416, 8),
( 195408755062890625, 8),
( 206453783524884736, 8),
( 218041257467152161, 8),
( 230193853492166656, 8),
( 242935032749128801, 8),
( 256289062500000000, 8),
( 270281038127131201, 8),
( 284936905588473856, 8),
( 300283484326400961, 8),
( 316348490636206336, 8),
( 333160561500390625, 8),
( 350749278894882816, 8),
( 369145194573386401, 8),
( 388379855336079616, 8),
( 408485828788939521, 8),
( 429496729600000000, 8),
( 451447246258894081, 8),
( 474373168346071296, 8),
( 498311414318121121, 8),
( 523300059815673856, 8),
( 549378366500390625, 8),
( 576586811427594496, 8),
( 604967116961135041, 8),
( 634562281237118976, 8),
( 665416609183179841, 8),
( 697575744100000000, 8),
( 731086699811838561, 8),
( 765997893392859136, 8),
( 802359178476091681, 8),
( 840221879151902976, 8),
( 879638824462890625, 8),
( 920664383502155776, 8),
( 963354501121950081, 8),
( 1007766734259732736, 8),
( 1053960288888713761, 8),
( 1101996057600000000, 8),
( 1151936657823500641, 8),
( 1203846470694789376, 8),
( 1257791680575160641, 8),
( 1313840315232157696, 8),
( 1372062286687890625, 8),
( 1432529432742502656, 8),
( 1495315559180183521, 8),
( 1560496482665168896, 8),
( 1628150074335205281, 8),
( 1698356304100000000, 8),
( 1771197285652216321, 8),
( 1846757322198614016, 8),
( 1925122952918976001, 8),
( 2006383000160502016, 8),
( 2090628617375390625, 8),
( 2177953337809371136, 8),
( 2268453123948987361, 8),
( 2362226417735475456, 8),
( 2459374191553118401, 8),
( 2560000000000000000, 8),
( 2664210032449121601, 8),
( 2772113166407885056, 8),
( 2883821021683985761, 8),
( 2999448015365799936, 8),
( 3119111417625390625, 8),
( 3242931408352297216, 8),
( 3371031134626313601, 8),
( 3503536769037500416, 8),
( 3640577568861717121, 8),
( 3782285936100000000, 8),
( 3928797478390152481, 8),
( 4080251070798954496, 8),
( 4236788918503437921, 8),
( 4398556620369715456, 8),
( 4565703233437890625, 8),
( 4738381338321616896, 8),
( 4916747105530914241, 8),
( 5100960362726891776, 8),
( 5291184662917065441, 8),
( 5487587353600000000, 8),
( 5690339646868044961, 8),
( 5899616690476974336, 8),
( 6115597639891380481, 8),
( 6338465731314712576, 8),
( 6568408355712890625, 8),
( 6805617133840466176, 8),
( 7050287992278341281, 8),
( 7302621240492097536, 8),
( 7562821648920027361, 8),
( 7831098528100000000, 8),
( 8107665808844335041, 8),
( 8392742123471896576, 8),
( 8686550888106661441, 8),
( 8989320386052055296, 8),
( 9301283852250390625, 8),
( 9622679558836781056, 8),
( 9953750901796946721, 8),
(10294746488738365696, 8),
(10645920227784266881, 8),
(11007531417600000000, 8),
(11379844838561358721, 8),
(11763130845074473216, 8),
(12157665459056928801, 8),
(12563730464589807616, 8),
(12981613503750390625, 8),
(13411608173635297536, 8),
(13854014124583882561, 8),
(14309137159611744256, 8),
(14777289335064248001, 8),
(15258789062500000000, 8),
(15753961211814252001, 8),
(16263137215612256256, 8),
(16786655174842630561, 8),
(17324859965700833536, 8),
(17878103347812890625, 8),
( 72057594037927936, 7),
];
let (base, power) = BASES[radix as usize];
(base as BigDigit, power)
}
_ => panic!("Invalid bigdigit size")
}
}