Crate big_int

source ·
Expand description

big_int - Arbitrary precision, arbitrary base integer arithmetic library.

use big_int::prelude::*;

let mut a: Loose<10> = "9000000000000000000000000000000000000000".parse().unwrap();
a /= 13.into();
assert_eq!(a, "692307692307692307692307692307692307692".parse().unwrap());

let mut b: Loose<16> = a.convert();
assert_eq!(b, "208D59C8D8669EDC306F76344EC4EC4EC".parse().unwrap());
b >>= 16.into();

let c: Loose<2> = b.convert();
assert_eq!(c, "100000100011010101100111001000110110000110011010011110110111000011".parse().unwrap());

let mut d: Tight<256> = c.convert();
d += vec![15, 90, 0].into();
assert_eq!(d, vec![2, 8, 213, 156, 141, 134, 121, 71, 195].into());

let e: Tight<10> = d.convert();
assert_eq!(format!("{e}"), "37530075201422313411".to_string());

This crate contains five primary big int implementations:

  • LooseBytes<BASE> - A collection of loosely packed 8-bit byte values representing each digit. Slightly memory inefficient, but with minimal performance overhead. Capable of representing any base from 2-256.
  • LooseShorts<BASE> - A collection of loosely packed 16-bit short values representing each digit. Somewhat memory inefficient, but with minimal performance overhead. Capable of representing any base from 2-65536.
  • LooseWords<BASE> - A collection of loosely packed 32-bit word values representing each digit. Fairly memory inefficient, but with minimal performance overhead. Capable of representing any base from 2-2^32.
  • Loose<BASE> - A collection of loosely packed 64-bit ints representing each digit. Very memory inefficient, but with minimal performance overhead. Capable of representing any base from 2-2^64.
  • Tight<BASE> - A collection of tightly packed bits representing each digit. Maximally memory efficient, and capable of representing any base from 2-2^64. However, the additional indirection adds some performance overhead.

Ints support most basic arithmetic operations, including addition, subtraction, multiplication, division, exponentiation, logarithm, nth root, and left/right shifting. Notably, shifting acts on the BASE of the associated number, increasing or decreasing the magnitude by powers of BASE as opposed to powers of 2.

Modules

  • base64 encoding & decoding, baked into the library :)
  • Denormalized numbers.
  • big int errors.
  • Safely retrieve items from a collection with negative indexing.
  • Loosely packed big int implementations.
  • Default exports: includes Loose, Tight, Sign, & Denormal
  • tightly packed big int implementation, for better memory efficiency.

Macros

  • Format out a vec of bytes as a list of binary numbers.
  • Safely create a bitmask of n bits in size shifted to the right side of the number without overflowing.
  • dbg! but don’t multiline-print
  • Create a list of pairs of randomly generated ints, constrained by the sizes of the associated int types passed.
  • Create a list of randomly generated ints, constrained by the sizes of the associated int types passed.

Structs

Enums

  • Represents the sign of a big int; either Positive or Negative.

Constants

Traits

  • A big int.
  • A builder for a big int. Use this to construct a big int one digit at a time, then call .into() to construct the final int.
  • A conversion that may only be performed unsafely.
  • A value that may be unwrapped.

Type Aliases

  • Size of an individual big int digit.

Derive Macros