java.util.Random — C++ Port

Repository: https://github.com/walshyaw/cpp-random

A faithful C++ implementation of Java’s java.util.Random class, using the same Linear Congruential Generator (LCG) algorithm, constants, and seed transformation as the original Java standard library. Given the same seed, this implementation produces identical output to java.util.Random.

Algorithm

This implementation uses a 48-bit Linear Congruential Generator with the following parameters:

Parameter	Value
Multiplier	`0x5DEECE66DL`
Addend	`0xBL`
Modulus	`2^48` (bitmask `(1L << 48) - 1`)

The seed transformation on initialization:

seed = (seed ^ 0x5DEECE66DL) & ((1L << 48) - 1)

Each call to next(bits) advances the state and returns the top bits bits of the new seed:

seed = (seed * 0x5DEECE66DL + 0xBL) & ((1L << 48) - 1)
return (int)(seed >> (48 - bits))

API Reference

Method	Return Type	Description
`Random(i64 seed)`	—	Constructs and seeds the generator
`setSeed(i64 seed)`	`void`	Resets the generator to a new seed
`getSeed()`	`i64`	Returns the current internal seed
`next(int bits)`	`int`	Core LCG step; returns top `bits` bits
`nextInt()`	`int`	Random 32-bit signed integer
`nextBytes(vector<ui8>&)`	`void`	Fills a byte vector with random values
`nextLong()`	`i64`	Random 64-bit signed integer
`nextBoolean()`	`bool`	Random boolean (~50% true)
`nextFloat()`	`float`	Random float in `[0.0, 1.0)`
`nextDouble()`	`double`	Random double in `[0.0, 1.0)`
`nextGaussian()`	`double`	Gaussian-distributed double (mean=0, stddev=1)

Limitations

This port does not implement the Java 8+ stream methods (ints(), longs(), doubles()).

Dependencies

C++17 or later
<cmath> — used by nextGaussian() for std::sqrt and std::log
<algorithm> — used by nextBytes() for std::min
<vector> — used by nextBytes()