Lightweight, high-performance C++ port of https://github.com/Patashu/break_infinity.js

Relies heavily on constexpr optimizations.

BigNums are represented internally as m × 10^e, using double m and uintmax_t e.

Note: e is unsigned. This means that BigNums can only store numbers as small as regular doubles, but the upper limit is higher.

Maximum representable value is std::nextafter(10.0, 0.0) × 10^numeric_limits<uintmax_t>::max(), effectively 1e(2^64).

See the example Makefile for build requirements.