Module exact 

Exact / arbitrary-precision arithmetic — the §3.1 “exact computation” foundation.

• bigint::BigInt — arbitrary-precision signed integer (sign + little- endian u32 magnitude, schoolbook multiply, Knuth long division).
• rational::BigRational — exact rational over BigInt, always reduced and sign-normalised.

These are heap-backed and live deliberately off the zero-heap hot path. Fallible operations (zero divisor / zero denominator) fail closed with Option/Result and never fabricate a value.

Re-exports

pub use bigint::BigInt;

pub use rational::BigRational;

Modules

bigint

Arbitrary-precision signed integer (BigInt).

rational

Exact rational numbers (BigRational) over BigInt.