Initial project import

tools/v/vlib/rand/sys/system_rng.c.v

@@ -0,0 +1,124 @@
+// Copyright (c) 2019-2024 Alexander Medvednikov. All rights reserved.
+// Use of this source code is governed by an MIT license
+// that can be found in the LICENSE file.
+module sys
+
+import math.bits
+import rand.buffer
+import rand.seed
+
+// Implementation note:
+// ====================
+// C.rand returns a pseudorandom integer from 0 (inclusive) to C.RAND_MAX (exclusive)
+// C.rand() is okay to use within its defined range.
+// (See: https://web.archive.org/web/20180801210127/http://eternallyconfuzzled.com/arts/jsw_art_rand.aspx)
+// The problem is, this value varies with the libc implementation. On windows,
+// for example, RAND_MAX is usually a measly 32767, whereas on (newer) linux it's generally
+// 2147483647. The repetition period also varies wildly. In order to provide more entropy
+// without altering the underlying algorithm too much, this implementation simply
+// requests for more random bits until the necessary width for the integers is achieved.
+
+pub const seed_len = 1
+
+const rand_limit = u64(C.RAND_MAX)
+const rand_bitsize = bits.len_64(rand_limit)
+const rand_bytesize = rand_bitsize / 8
+const u16_iter_count = calculate_iterations_for(16)
+const u32_iter_count = calculate_iterations_for(32)
+const u64_iter_count = calculate_iterations_for(64)
+
+fn calculate_iterations_for(bits_ int) int {
+	base := bits_ / rand_bitsize
+	extra := if bits_ % rand_bitsize == 0 { 0 } else { 1 }
+	return base + extra
+}
+
+// SysRNG is the PRNG provided by default in the libc implementiation that V uses.
+pub struct SysRNG {
+	buffer.PRNGBuffer
+mut:
+	seed u32 = seed.time_seed_32()
+}
+
+// r.seed() sets the seed of the accepting SysRNG to the given data.
+pub fn (mut r SysRNG) seed(seed_data []u32) {
+	if seed_data.len != 1 {
+		eprintln('SysRNG needs one 32-bit unsigned integer as the seed.')
+		exit(1)
+	}
+	r.seed = seed_data[0]
+	C.srand(r.seed)
+}
+
+// r.default_rand() exposes the default behavior of the system's RNG
+// (equivalent to calling C.rand()). Recommended for testing/comparison
+// b/w V and other languages using libc and not for regular use.
+// This is also a one-off feature of SysRNG, similar to the global seed
+// situation. Other generators will not have this.
+@[inline]
+pub fn (r SysRNG) default_rand() int {
+	return C.rand()
+}
+
+// byte returns a uniformly distributed pseudorandom 8-bit unsigned positive `byte`.
+@[inline]
+pub fn (mut r SysRNG) u8() u8 {
+	if r.bytes_left >= 1 {
+		r.bytes_left -= 1
+		value := u8(r.buffer)
+		r.buffer >>= 8
+		return value
+	}
+	r.buffer = u64(r.default_rand())
+	r.bytes_left = rand_bytesize - 1
+	value := u8(r.buffer)
+	r.buffer >>= 8
+	return value
+}
+
+// u16 returns a uniformly distributed pseudorandom 16-bit unsigned positive `u16`.
+@[inline]
+pub fn (mut r SysRNG) u16() u16 {
+	if r.bytes_left >= 2 {
+		r.bytes_left -= 2
+		value := u16(r.buffer)
+		r.buffer >>= 16
+		return value
+	}
+	mut result := u16(C.rand())
+	for i in 1 .. u16_iter_count {
+		result = result ^ (u16(C.rand()) << (rand_bitsize * i))
+	}
+	return result
+}
+
+// u32 returns a uniformly distributed pseudorandom 32-bit unsigned positive `u32`.
+@[inline]
+pub fn (r SysRNG) u32() u32 {
+	mut result := u32(C.rand())
+	for i in 1 .. u32_iter_count {
+		result = result ^ (u32(C.rand()) << (rand_bitsize * i))
+	}
+	return result
+}
+
+// u64 returns a uniformly distributed pseudorandom 64-bit unsigned positive `u64`.
+@[inline]
+pub fn (r SysRNG) u64() u64 {
+	mut result := u64(C.rand())
+	for i in 1 .. u64_iter_count {
+		result = result ^ (u64(C.rand()) << (rand_bitsize * i))
+	}
+	return result
+}
+
+// block_size returns the number of bits that the RNG can produce in a single iteration.
+@[inline]
+pub fn (r SysRNG) block_size() int {
+	return rand_bitsize
+}
+
+// free should be called when the generator is no longer needed.
+@[unsafe]
+pub fn (mut rng SysRNG) free() {
+}

tools/v/vlib/rand/sys/system_rng.js.v

@@ -0,0 +1,15 @@
+// Copyright (c) 2019-2024 Alexander Medvednikov. All rights reserved.
+// Use of this source code is governed by an MIT license
+// that can be found in the LICENSE file.
+module sys
+
+// Until there's a portable, JS has a seeded way to produce random numbers
+// and not just Math.random(), use any of the existing implementations
+// as the System's RNG
+type SysRNG = WyRandRNG
+
+// In the JS version, we simply return the same int as is normally generated.
+@[inline]
+pub fn (r SysRNG) default_rand() int {
+	return r.int()
+}

tools/v/vlib/rand/sys/system_rng_test.v

@@ -0,0 +1,381 @@
+import math
+import rand
+import rand.sys
+
+const range_limit = 40
+const value_count = 1000
+const seeds = [u32(42), 256]
+
+const sample_size = 1000
+const stats_epsilon = 0.05
+const inv_sqrt_12 = 1.0 / math.sqrt(12)
+
+fn get_n_randoms(n int, mut r rand.PRNG) []int {
+	mut ints := []int{cap: n}
+	for _ in 0 .. n {
+		ints << r.int()
+	}
+	return ints
+}
+
+fn test_sys_rng_reproducibility() {
+	// Note that C.srand() sets the seed globally.
+	// So the order of seeding matters. It is recommended
+	// to obtain all necessary data first, then set the
+	// seed for another batch of data.
+	for seed in seeds {
+		seed_data := [seed]
+		mut r1 := &rand.PRNG(&sys.SysRNG{})
+		mut r2 := &rand.PRNG(&sys.SysRNG{})
+		r1.seed(seed_data)
+		ints1 := get_n_randoms(value_count, mut r1)
+		r2.seed(seed_data)
+		ints2 := get_n_randoms(value_count, mut r2)
+		assert ints1 == ints2
+	}
+}
+
+fn test_sys_rng_variability() {
+	// If this test fails and if it is certainly not the implementation
+	// at fault, try changing the seed values. Repeated values are
+	// improbable but not impossible.
+	for seed in seeds {
+		seed_data := [seed]
+		mut rng := &rand.PRNG(&sys.SysRNG{})
+		rng.seed(seed_data)
+		mut values := []u64{cap: value_count}
+		for i in 0 .. value_count {
+			value := rng.u64()
+			assert value !in values
+			assert values.len == i
+			values << value
+		}
+	}
+}
+
+fn check_uniformity_u64(mut rng rand.PRNG, range u64) {
+	range_f64 := f64(range)
+	expected_mean := range_f64 / 2.0
+	mut variance := 0.0
+	for _ in 0 .. sample_size {
+		diff := f64(rng.u64n(range) or { panic("Couldn't obtain u64") }) - expected_mean
+		variance += diff * diff
+	}
+	variance /= sample_size - 1
+	sigma := math.sqrt(variance)
+	expected_sigma := range_f64 * inv_sqrt_12
+	error := (sigma - expected_sigma) / expected_sigma
+	assert math.abs(error) < stats_epsilon
+}
+
+fn test_sys_rng_uniformity_u64() {
+	$if windows {
+		// On Windows, C.rand() has RAND_MAX = 32767 (15 bits). The XOR-based
+		// extension to u64 is inherently biased, so this uniformity check
+		// cannot pass with the system RNG.
+		return
+	}
+	// This assumes that C.rand() produces uniform results to begin with.
+	// If the failure persists, report an issue on GitHub
+	ranges := [14019545, 80240, 130]
+	for seed in seeds {
+		seed_data := [seed]
+		mut rng := &rand.PRNG(&sys.SysRNG{})
+		rng.seed(seed_data)
+		for range in ranges {
+			check_uniformity_u64(mut rng, u64(range))
+		}
+	}
+}
+
+fn check_uniformity_f64(mut rng rand.PRNG) {
+	expected_mean := 0.5
+	mut variance := 0.0
+	for _ in 0 .. sample_size {
+		diff := rng.f64() - expected_mean
+		variance += diff * diff
+	}
+	variance /= sample_size - 1
+	sigma := math.sqrt(variance)
+	expected_sigma := inv_sqrt_12
+	error := (sigma - expected_sigma) / expected_sigma
+	assert math.abs(error) < stats_epsilon
+}
+
+fn test_sys_rng_uniformity_f64() {
+	$if windows {
+		// See test_sys_rng_uniformity_u64 above: Windows C.rand() has a
+		// 15-bit range, so the derived f64 distribution is not uniform enough
+		// to pass the statistical check.
+		return
+	}
+	// The f64 version
+	for seed in seeds {
+		seed_data := [seed]
+		mut rng := &rand.PRNG(&sys.SysRNG{})
+		rng.seed(seed_data)
+		check_uniformity_f64(mut rng)
+	}
+}
+
+fn test_sys_rng_u32n() {
+	max := u32(16384)
+	for seed in seeds {
+		seed_data := [seed]
+		mut rng := &rand.PRNG(&sys.SysRNG{})
+		rng.seed(seed_data)
+		for _ in 0 .. range_limit {
+			value := rng.u32n(max) or { panic("Couldn't obtain u32") }
+			assert value >= 0
+			assert value < max
+		}
+	}
+}
+
+fn test_sys_rng_u64n() {
+	max := u64(379091181005)
+	for seed in seeds {
+		seed_data := [seed]
+		mut rng := &rand.PRNG(&sys.SysRNG{})
+		rng.seed(seed_data)
+		for _ in 0 .. range_limit {
+			value := rng.u64n(max) or { panic("Couldn't obtain u64") }
+			assert value >= 0
+			assert value < max
+		}
+	}
+}
+
+fn test_sys_rng_u32_in_range() {
+	max := u32(484468466)
+	min := u32(316846)
+	for seed in seeds {
+		seed_data := [seed]
+		mut rng := &rand.PRNG(&sys.SysRNG{})
+		rng.seed(seed_data)
+		for _ in 0 .. range_limit {
+			value := rng.u32_in_range(min, max) or { panic("Couldn't obtain u32 in range") }
+			assert value >= min
+			assert value < max
+		}
+	}
+}
+
+fn test_sys_rng_u64_in_range() {
+	max := u64(216468454685163)
+	min := u64(6848646868)
+	for seed in seeds {
+		seed_data := [seed]
+		mut rng := &rand.PRNG(&sys.SysRNG{})
+		rng.seed(seed_data)
+		for _ in 0 .. range_limit {
+			value := rng.u64_in_range(min, max) or { panic("Couldn't obtain u64 in range") }
+			assert value >= min
+			assert value < max
+		}
+	}
+}
+
+fn test_sys_rng_intn() {
+	max := 2525642
+	for seed in seeds {
+		seed_data := [seed]
+		mut rng := &rand.PRNG(&sys.SysRNG{})
+		rng.seed(seed_data)
+		for _ in 0 .. range_limit {
+			value := rng.intn(max) or { panic("Couldn't obtain int") }
+			assert value >= 0
+			assert value < max
+		}
+	}
+}
+
+fn test_sys_rng_i32n() {
+	max := i32(2525642)
+	for seed in seeds {
+		seed_data := [seed]
+		mut rng := &rand.PRNG(&sys.SysRNG{})
+		rng.seed(seed_data)
+		for _ in 0 .. range_limit {
+			value := rng.i32n(max) or { panic("Couldn't obtain i32") }
+			assert value >= 0
+			assert value < max
+		}
+	}
+}
+
+fn test_sys_rng_i64n() {
+	max := i64(3246727724653636)
+	for seed in seeds {
+		seed_data := [seed]
+		mut rng := &rand.PRNG(&sys.SysRNG{})
+		rng.seed(seed_data)
+		for _ in 0 .. range_limit {
+			value := rng.i64n(max) or { panic("Couldn't obtain i64") }
+			assert value >= 0
+			assert value < max
+		}
+	}
+}
+
+fn test_sys_rng_int_in_range() {
+	min := -4252
+	max := 23054962
+	for seed in seeds {
+		seed_data := [seed]
+		mut rng := &rand.PRNG(&sys.SysRNG{})
+		rng.seed(seed_data)
+		for _ in 0 .. range_limit {
+			value := rng.int_in_range(min, max) or { panic("Couldn't obtain int in range") }
+			assert value >= min
+			assert value < max
+		}
+	}
+}
+
+fn test_sys_rng_i32_in_range() {
+	min := i32(-4252)
+	max := i32(23054962)
+	for seed in seeds {
+		seed_data := [seed]
+		mut rng := &rand.PRNG(&sys.SysRNG{})
+		rng.seed(seed_data)
+		for _ in 0 .. range_limit {
+			value := rng.i32_in_range(min, max) or { panic("Couldn't obtain i32 in range") }
+			assert value >= min
+			assert value < max
+		}
+	}
+}
+
+fn test_sys_rng_i64_in_range() {
+	min := i64(-24095)
+	max := i64(324058)
+	for seed in seeds {
+		seed_data := [seed]
+		mut rng := &rand.PRNG(&sys.SysRNG{})
+		rng.seed(seed_data)
+		for _ in 0 .. range_limit {
+			value := rng.i64_in_range(min, max) or { panic("Couldn't obtain i64 in range") }
+			assert value >= min
+			assert value < max
+		}
+	}
+}
+
+fn test_sys_rng_int31() {
+	max_u31 := int(0x7FFFFFFF)
+	sign_mask := int(u32(0x80000000))
+	for seed in seeds {
+		seed_data := [seed]
+		mut rng := &rand.PRNG(&sys.SysRNG{})
+		rng.seed(seed_data)
+		for _ in 0 .. range_limit {
+			value := rng.int31()
+			assert value >= 0
+			assert value <= max_u31
+			// This statement ensures that the sign bit is zero
+			assert (value & sign_mask) == 0
+		}
+	}
+}
+
+fn test_sys_rng_int63() {
+	max_u63 := i64(0x7FFFFFFFFFFFFFFF)
+	sign_mask := i64(u64(0x8000000000000000))
+	for seed in seeds {
+		seed_data := [seed]
+		mut rng := &rand.PRNG(&sys.SysRNG{})
+		rng.seed(seed_data)
+		for _ in 0 .. range_limit {
+			value := rng.int63()
+			assert value >= 0
+			assert value <= max_u63
+			assert (value & sign_mask) == 0
+		}
+	}
+}
+
+fn test_sys_rng_f32() {
+	for seed in seeds {
+		seed_data := [seed]
+		mut rng := &rand.PRNG(&sys.SysRNG{})
+		rng.seed(seed_data)
+		for _ in 0 .. range_limit {
+			value := rng.f32()
+			assert value >= 0.0
+			assert value < 1.0
+		}
+	}
+}
+
+fn test_sys_rng_f64() {
+	for seed in seeds {
+		seed_data := [seed]
+		mut rng := &rand.PRNG(&sys.SysRNG{})
+		rng.seed(seed_data)
+		for _ in 0 .. range_limit {
+			value := rng.f64()
+			assert value >= 0.0
+			assert value < 1.0
+		}
+	}
+}
+
+fn test_sys_rng_f32n() {
+	max := f32(357.0)
+	for seed in seeds {
+		seed_data := [seed]
+		mut rng := &rand.PRNG(&sys.SysRNG{})
+		rng.seed(seed_data)
+		for _ in 0 .. range_limit {
+			value := rng.f32n(max) or { panic("Couldn't obtain f32") }
+			assert value >= 0.0
+			assert value < max
+		}
+	}
+}
+
+fn test_sys_rng_f64n() {
+	max := 1.52e6
+	for seed in seeds {
+		seed_data := [seed]
+		mut rng := &rand.PRNG(&sys.SysRNG{})
+		rng.seed(seed_data)
+		for _ in 0 .. range_limit {
+			value := rng.f64n(max) or { panic("Couldn't obtain f64") }
+			assert value >= 0.0
+			assert value < max
+		}
+	}
+}
+
+fn test_sys_rng_f32_in_range() {
+	min := f32(-24.0)
+	max := f32(125.0)
+	for seed in seeds {
+		seed_data := [seed]
+		mut rng := &rand.PRNG(&sys.SysRNG{})
+		rng.seed(seed_data)
+		for _ in 0 .. range_limit {
+			value := rng.f32_in_range(min, max) or { panic("Couldn't obtain f32 in range") }
+			assert value >= min
+			assert value < max
+		}
+	}
+}
+
+fn test_sys_rng_f64_in_range() {
+	min := -548.7
+	max := 5015.2
+	for seed in seeds {
+		seed_data := [seed]
+		mut rng := &rand.PRNG(&sys.SysRNG{})
+		rng.seed(seed_data)
+		for _ in 0 .. range_limit {
+			value := rng.f64_in_range(min, max) or { panic("Couldn't obtain f64 in range") }
+			assert value >= min
+			assert value < max
+		}
+	}
+}