| 1 | // Copyright 2018 Developers of the Rand project. |
| 2 | // |
| 3 | // Licensed under the Apache License, Version 2.0 <LICENSE-APACHE or |
| 4 | // https://www.apache.org/licenses/LICENSE-2.0> or the MIT license |
| 5 | // <LICENSE-MIT or https://opensource.org/licenses/MIT>, at your |
| 6 | // option. This file may not be copied, modified, or distributed |
| 7 | // except according to those terms. |
| 8 | |
| 9 | //! The implementations of the `Standard` distribution for other built-in types. |
| 10 | |
| 11 | use core::char; |
| 12 | use core::num::Wrapping; |
| 13 | #[cfg (feature = "alloc" )] |
| 14 | use alloc::string::String; |
| 15 | |
| 16 | use crate::distributions::{Distribution, Standard, Uniform}; |
| 17 | #[cfg (feature = "alloc" )] |
| 18 | use crate::distributions::DistString; |
| 19 | use crate::Rng; |
| 20 | |
| 21 | #[cfg (feature = "serde1" )] |
| 22 | use serde::{Serialize, Deserialize}; |
| 23 | #[cfg (feature = "min_const_gen" )] |
| 24 | use core::mem::{self, MaybeUninit}; |
| 25 | |
| 26 | |
| 27 | // ----- Sampling distributions ----- |
| 28 | |
| 29 | /// Sample a `u8`, uniformly distributed over ASCII letters and numbers: |
| 30 | /// a-z, A-Z and 0-9. |
| 31 | /// |
| 32 | /// # Example |
| 33 | /// |
| 34 | /// ``` |
| 35 | /// use rand::{Rng, thread_rng}; |
| 36 | /// use rand::distributions::Alphanumeric; |
| 37 | /// |
| 38 | /// let mut rng = thread_rng(); |
| 39 | /// let chars: String = (0..7).map(|_| rng.sample(Alphanumeric) as char).collect(); |
| 40 | /// println!("Random chars: {}" , chars); |
| 41 | /// ``` |
| 42 | /// |
| 43 | /// The [`DistString`] trait provides an easier method of generating |
| 44 | /// a random `String`, and offers more efficient allocation: |
| 45 | /// ``` |
| 46 | /// use rand::distributions::{Alphanumeric, DistString}; |
| 47 | /// let string = Alphanumeric.sample_string(&mut rand::thread_rng(), 16); |
| 48 | /// println!("Random string: {}" , string); |
| 49 | /// ``` |
| 50 | /// |
| 51 | /// # Passwords |
| 52 | /// |
| 53 | /// Users sometimes ask whether it is safe to use a string of random characters |
| 54 | /// as a password. In principle, all RNGs in Rand implementing `CryptoRng` are |
| 55 | /// suitable as a source of randomness for generating passwords (if they are |
| 56 | /// properly seeded), but it is more conservative to only use randomness |
| 57 | /// directly from the operating system via the `getrandom` crate, or the |
| 58 | /// corresponding bindings of a crypto library. |
| 59 | /// |
| 60 | /// When generating passwords or keys, it is important to consider the threat |
| 61 | /// model and in some cases the memorability of the password. This is out of |
| 62 | /// scope of the Rand project, and therefore we defer to the following |
| 63 | /// references: |
| 64 | /// |
| 65 | /// - [Wikipedia article on Password Strength](https://en.wikipedia.org/wiki/Password_strength) |
| 66 | /// - [Diceware for generating memorable passwords](https://en.wikipedia.org/wiki/Diceware) |
| 67 | #[derive (Debug, Clone, Copy)] |
| 68 | #[cfg_attr (feature = "serde1" , derive(Serialize, Deserialize))] |
| 69 | pub struct Alphanumeric; |
| 70 | |
| 71 | |
| 72 | // ----- Implementations of distributions ----- |
| 73 | |
| 74 | impl Distribution<char> for Standard { |
| 75 | #[inline ] |
| 76 | fn sample<R: Rng + ?Sized>(&self, rng: &mut R) -> char { |
| 77 | // A valid `char` is either in the interval `[0, 0xD800)` or |
| 78 | // `(0xDFFF, 0x11_0000)`. All `char`s must therefore be in |
| 79 | // `[0, 0x11_0000)` but not in the "gap" `[0xD800, 0xDFFF]` which is |
| 80 | // reserved for surrogates. This is the size of that gap. |
| 81 | const GAP_SIZE: u32 = 0xDFFF - 0xD800 + 1; |
| 82 | |
| 83 | // Uniform::new(0, 0x11_0000 - GAP_SIZE) can also be used but it |
| 84 | // seemed slower. |
| 85 | let range: Uniform = Uniform::new(GAP_SIZE, high:0x11_0000); |
| 86 | |
| 87 | let mut n: u32 = range.sample(rng); |
| 88 | if n <= 0xDFFF { |
| 89 | n -= GAP_SIZE; |
| 90 | } |
| 91 | unsafe { char::from_u32_unchecked(n) } |
| 92 | } |
| 93 | } |
| 94 | |
| 95 | /// Note: the `String` is potentially left with excess capacity; optionally the |
| 96 | /// user may call `string.shrink_to_fit()` afterwards. |
| 97 | #[cfg (feature = "alloc" )] |
| 98 | impl DistString for Standard { |
| 99 | fn append_string<R: Rng + ?Sized>(&self, rng: &mut R, s: &mut String, len: usize) { |
| 100 | // A char is encoded with at most four bytes, thus this reservation is |
| 101 | // guaranteed to be sufficient. We do not shrink_to_fit afterwards so |
| 102 | // that repeated usage on the same `String` buffer does not reallocate. |
| 103 | s.reserve(additional:4 * len); |
| 104 | s.extend(iter:Distribution::<char>::sample_iter(self, rng).take(len)); |
| 105 | } |
| 106 | } |
| 107 | |
| 108 | impl Distribution<u8> for Alphanumeric { |
| 109 | fn sample<R: Rng + ?Sized>(&self, rng: &mut R) -> u8 { |
| 110 | const RANGE: u32 = 26 + 26 + 10; |
| 111 | const GEN_ASCII_STR_CHARSET: &[u8] = b"ABCDEFGHIJKLMNOPQRSTUVWXYZ\ |
| 112 | abcdefghijklmnopqrstuvwxyz\ |
| 113 | 0123456789" ; |
| 114 | // We can pick from 62 characters. This is so close to a power of 2, 64, |
| 115 | // that we can do better than `Uniform`. Use a simple bitshift and |
| 116 | // rejection sampling. We do not use a bitmask, because for small RNGs |
| 117 | // the most significant bits are usually of higher quality. |
| 118 | loop { |
| 119 | let var: u32 = rng.next_u32() >> (32 - 6); |
| 120 | if var < RANGE { |
| 121 | return GEN_ASCII_STR_CHARSET[var as usize]; |
| 122 | } |
| 123 | } |
| 124 | } |
| 125 | } |
| 126 | |
| 127 | #[cfg (feature = "alloc" )] |
| 128 | impl DistString for Alphanumeric { |
| 129 | fn append_string<R: Rng + ?Sized>(&self, rng: &mut R, string: &mut String, len: usize) { |
| 130 | unsafe { |
| 131 | let v: &mut Vec = string.as_mut_vec(); |
| 132 | v.extend(self.sample_iter(rng).take(len)); |
| 133 | } |
| 134 | } |
| 135 | } |
| 136 | |
| 137 | impl Distribution<bool> for Standard { |
| 138 | #[inline ] |
| 139 | fn sample<R: Rng + ?Sized>(&self, rng: &mut R) -> bool { |
| 140 | // We can compare against an arbitrary bit of an u32 to get a bool. |
| 141 | // Because the least significant bits of a lower quality RNG can have |
| 142 | // simple patterns, we compare against the most significant bit. This is |
| 143 | // easiest done using a sign test. |
| 144 | (rng.next_u32() as i32) < 0 |
| 145 | } |
| 146 | } |
| 147 | |
| 148 | macro_rules! tuple_impl { |
| 149 | // use variables to indicate the arity of the tuple |
| 150 | ($($tyvar:ident),* ) => { |
| 151 | // the trailing commas are for the 1 tuple |
| 152 | impl< $( $tyvar ),* > |
| 153 | Distribution<( $( $tyvar ),* , )> |
| 154 | for Standard |
| 155 | where $( Standard: Distribution<$tyvar> ),* |
| 156 | { |
| 157 | #[inline] |
| 158 | fn sample<R: Rng + ?Sized>(&self, _rng: &mut R) -> ( $( $tyvar ),* , ) { |
| 159 | ( |
| 160 | // use the $tyvar's to get the appropriate number of |
| 161 | // repeats (they're not actually needed) |
| 162 | $( |
| 163 | _rng.gen::<$tyvar>() |
| 164 | ),* |
| 165 | , |
| 166 | ) |
| 167 | } |
| 168 | } |
| 169 | } |
| 170 | } |
| 171 | |
| 172 | impl Distribution<()> for Standard { |
| 173 | #[allow (clippy::unused_unit)] |
| 174 | #[inline ] |
| 175 | fn sample<R: Rng + ?Sized>(&self, _: &mut R) -> () { |
| 176 | () |
| 177 | } |
| 178 | } |
| 179 | tuple_impl! {A} |
| 180 | tuple_impl! {A, B} |
| 181 | tuple_impl! {A, B, C} |
| 182 | tuple_impl! {A, B, C, D} |
| 183 | tuple_impl! {A, B, C, D, E} |
| 184 | tuple_impl! {A, B, C, D, E, F} |
| 185 | tuple_impl! {A, B, C, D, E, F, G} |
| 186 | tuple_impl! {A, B, C, D, E, F, G, H} |
| 187 | tuple_impl! {A, B, C, D, E, F, G, H, I} |
| 188 | tuple_impl! {A, B, C, D, E, F, G, H, I, J} |
| 189 | tuple_impl! {A, B, C, D, E, F, G, H, I, J, K} |
| 190 | tuple_impl! {A, B, C, D, E, F, G, H, I, J, K, L} |
| 191 | |
| 192 | #[cfg (feature = "min_const_gen" )] |
| 193 | #[cfg_attr (doc_cfg, doc(cfg(feature = "min_const_gen" )))] |
| 194 | impl<T, const N: usize> Distribution<[T; N]> for Standard |
| 195 | where Standard: Distribution<T> |
| 196 | { |
| 197 | #[inline ] |
| 198 | fn sample<R: Rng + ?Sized>(&self, _rng: &mut R) -> [T; N] { |
| 199 | let mut buff: [MaybeUninit<T>; N] = unsafe { MaybeUninit::uninit().assume_init() }; |
| 200 | |
| 201 | for elem in &mut buff { |
| 202 | *elem = MaybeUninit::new(_rng.gen()); |
| 203 | } |
| 204 | |
| 205 | unsafe { mem::transmute_copy::<_, _>(&buff) } |
| 206 | } |
| 207 | } |
| 208 | |
| 209 | #[cfg (not(feature = "min_const_gen" ))] |
| 210 | macro_rules! array_impl { |
| 211 | // recursive, given at least one type parameter: |
| 212 | {$n:expr, $t:ident, $($ts:ident,)*} => { |
| 213 | array_impl!{($n - 1), $($ts,)*} |
| 214 | |
| 215 | impl<T> Distribution<[T; $n]> for Standard where Standard: Distribution<T> { |
| 216 | #[inline] |
| 217 | fn sample<R: Rng + ?Sized>(&self, _rng: &mut R) -> [T; $n] { |
| 218 | [_rng.gen::<$t>(), $(_rng.gen::<$ts>()),*] |
| 219 | } |
| 220 | } |
| 221 | }; |
| 222 | // empty case: |
| 223 | {$n:expr,} => { |
| 224 | impl<T> Distribution<[T; $n]> for Standard { |
| 225 | fn sample<R: Rng + ?Sized>(&self, _rng: &mut R) -> [T; $n] { [] } |
| 226 | } |
| 227 | }; |
| 228 | } |
| 229 | |
| 230 | #[cfg (not(feature = "min_const_gen" ))] |
| 231 | array_impl! {32, T, T, T, T, T, T, T, T, T, T, T, T, T, T, T, T, T, T, T, T, T, T, T, T, T, T, T, T, T, T, T, T,} |
| 232 | |
| 233 | impl<T> Distribution<Option<T>> for Standard |
| 234 | where Standard: Distribution<T> |
| 235 | { |
| 236 | #[inline ] |
| 237 | fn sample<R: Rng + ?Sized>(&self, rng: &mut R) -> Option<T> { |
| 238 | // UFCS is needed here: https://github.com/rust-lang/rust/issues/24066 |
| 239 | if rng.gen::<bool>() { |
| 240 | Some(rng.gen()) |
| 241 | } else { |
| 242 | None |
| 243 | } |
| 244 | } |
| 245 | } |
| 246 | |
| 247 | impl<T> Distribution<Wrapping<T>> for Standard |
| 248 | where Standard: Distribution<T> |
| 249 | { |
| 250 | #[inline ] |
| 251 | fn sample<R: Rng + ?Sized>(&self, rng: &mut R) -> Wrapping<T> { |
| 252 | Wrapping(rng.gen()) |
| 253 | } |
| 254 | } |
| 255 | |
| 256 | |
| 257 | #[cfg (test)] |
| 258 | mod tests { |
| 259 | use super::*; |
| 260 | use crate::RngCore; |
| 261 | #[cfg (feature = "alloc" )] use alloc::string::String; |
| 262 | |
| 263 | #[test ] |
| 264 | fn test_misc() { |
| 265 | let rng: &mut dyn RngCore = &mut crate::test::rng(820); |
| 266 | |
| 267 | rng.sample::<char, _>(Standard); |
| 268 | rng.sample::<bool, _>(Standard); |
| 269 | } |
| 270 | |
| 271 | #[cfg (feature = "alloc" )] |
| 272 | #[test ] |
| 273 | fn test_chars() { |
| 274 | use core::iter; |
| 275 | let mut rng = crate::test::rng(805); |
| 276 | |
| 277 | // Test by generating a relatively large number of chars, so we also |
| 278 | // take the rejection sampling path. |
| 279 | let word: String = iter::repeat(()) |
| 280 | .map(|()| rng.gen::<char>()) |
| 281 | .take(1000) |
| 282 | .collect(); |
| 283 | assert!(!word.is_empty()); |
| 284 | } |
| 285 | |
| 286 | #[test ] |
| 287 | fn test_alphanumeric() { |
| 288 | let mut rng = crate::test::rng(806); |
| 289 | |
| 290 | // Test by generating a relatively large number of chars, so we also |
| 291 | // take the rejection sampling path. |
| 292 | let mut incorrect = false; |
| 293 | for _ in 0..100 { |
| 294 | let c: char = rng.sample(Alphanumeric).into(); |
| 295 | incorrect |= !(('0' ..='9' ).contains(&c) || |
| 296 | ('A' ..='Z' ).contains(&c) || |
| 297 | ('a' ..='z' ).contains(&c) ); |
| 298 | } |
| 299 | assert!(!incorrect); |
| 300 | } |
| 301 | |
| 302 | #[test ] |
| 303 | fn value_stability() { |
| 304 | fn test_samples<T: Copy + core::fmt::Debug + PartialEq, D: Distribution<T>>( |
| 305 | distr: &D, zero: T, expected: &[T], |
| 306 | ) { |
| 307 | let mut rng = crate::test::rng(807); |
| 308 | let mut buf = [zero; 5]; |
| 309 | for x in &mut buf { |
| 310 | *x = rng.sample(&distr); |
| 311 | } |
| 312 | assert_eq!(&buf, expected); |
| 313 | } |
| 314 | |
| 315 | test_samples(&Standard, 'a' , &[ |
| 316 | ' \u{8cdac}' , |
| 317 | ' \u{a346a}' , |
| 318 | ' \u{80120}' , |
| 319 | ' \u{ed692}' , |
| 320 | ' \u{35888}' , |
| 321 | ]); |
| 322 | test_samples(&Alphanumeric, 0, &[104, 109, 101, 51, 77]); |
| 323 | test_samples(&Standard, false, &[true, true, false, true, false]); |
| 324 | test_samples(&Standard, None as Option<bool>, &[ |
| 325 | Some(true), |
| 326 | None, |
| 327 | Some(false), |
| 328 | None, |
| 329 | Some(false), |
| 330 | ]); |
| 331 | test_samples(&Standard, Wrapping(0i32), &[ |
| 332 | Wrapping(-2074640887), |
| 333 | Wrapping(-1719949321), |
| 334 | Wrapping(2018088303), |
| 335 | Wrapping(-547181756), |
| 336 | Wrapping(838957336), |
| 337 | ]); |
| 338 | |
| 339 | // We test only sub-sets of tuple and array impls |
| 340 | test_samples(&Standard, (), &[(), (), (), (), ()]); |
| 341 | test_samples(&Standard, (false,), &[ |
| 342 | (true,), |
| 343 | (true,), |
| 344 | (false,), |
| 345 | (true,), |
| 346 | (false,), |
| 347 | ]); |
| 348 | test_samples(&Standard, (false, false), &[ |
| 349 | (true, true), |
| 350 | (false, true), |
| 351 | (false, false), |
| 352 | (true, false), |
| 353 | (false, false), |
| 354 | ]); |
| 355 | |
| 356 | test_samples(&Standard, [0u8; 0], &[[], [], [], [], []]); |
| 357 | test_samples(&Standard, [0u8; 3], &[ |
| 358 | [9, 247, 111], |
| 359 | [68, 24, 13], |
| 360 | [174, 19, 194], |
| 361 | [172, 69, 213], |
| 362 | [149, 207, 29], |
| 363 | ]); |
| 364 | } |
| 365 | } |
| 366 | |