| 1 | /*! |
| 2 | Lower level primitive types that are useful in a variety of circumstances. |
| 3 | |
| 4 | # Overview |
| 5 | |
| 6 | This list represents the principle types in this module and briefly describes |
| 7 | when you might want to use them. |
| 8 | |
| 9 | * [`PatternID`] - A type that represents the identifier of a regex pattern. |
| 10 | This is probably the most widely used type in this module (which is why it's |
| 11 | also re-exported in the crate root). |
| 12 | * [`StateID`] - A type the represents the identifier of a finite automaton |
| 13 | state. This is used for both NFAs and DFAs, with the notable exception of |
| 14 | the hybrid NFA/DFA. (The hybrid NFA/DFA uses a special purpose "lazy" state |
| 15 | identifier.) |
| 16 | * [`SmallIndex`] - The internal representation of both a `PatternID` and a |
| 17 | `StateID`. Its purpose is to serve as a type that can index memory without |
| 18 | being as big as a `usize` on 64-bit targets. The main idea behind this type |
| 19 | is that there are many things in regex engines that will, in practice, never |
| 20 | overflow a 32-bit integer. (For example, like the number of patterns in a regex |
| 21 | or the number of states in an NFA.) Thus, a `SmallIndex` can be used to index |
| 22 | memory without peppering `as` casts everywhere. Moreover, it forces callers |
| 23 | to handle errors in the case where, somehow, the value would otherwise overflow |
| 24 | either a 32-bit integer or a `usize` (e.g., on 16-bit targets). |
| 25 | */ |
| 26 | |
| 27 | // The macro we use to define some types below adds methods that we don't |
| 28 | // use on some of the types. There isn't much, so we just squash the warning. |
| 29 | #![allow (dead_code)] |
| 30 | |
| 31 | use alloc::vec::Vec; |
| 32 | |
| 33 | use crate::util::int::{Usize, U16, U32, U64}; |
| 34 | |
| 35 | /// A type that represents a "small" index. |
| 36 | /// |
| 37 | /// The main idea of this type is to provide something that can index memory, |
| 38 | /// but uses less memory than `usize` on 64-bit systems. Specifically, its |
| 39 | /// representation is always a `u32` and has `repr(transparent)` enabled. (So |
| 40 | /// it is safe to transmute between a `u32` and a `SmallIndex`.) |
| 41 | /// |
| 42 | /// A small index is typically useful in cases where there is no practical way |
| 43 | /// that the index will overflow a 32-bit integer. A good example of this is |
| 44 | /// an NFA state. If you could somehow build an NFA with `2^30` states, its |
| 45 | /// memory usage would be exorbitant and its runtime execution would be so |
| 46 | /// slow as to be completely worthless. Therefore, this crate generally deems |
| 47 | /// it acceptable to return an error if it would otherwise build an NFA that |
| 48 | /// requires a slice longer than what a 32-bit integer can index. In exchange, |
| 49 | /// we can use 32-bit indices instead of 64-bit indices in various places. |
| 50 | /// |
| 51 | /// This type ensures this by providing a constructor that will return an error |
| 52 | /// if its argument cannot fit into the type. This makes it much easier to |
| 53 | /// handle these sorts of boundary cases that are otherwise extremely subtle. |
| 54 | /// |
| 55 | /// On all targets, this type guarantees that its value will fit in a `u32`, |
| 56 | /// `i32`, `usize` and an `isize`. This means that on 16-bit targets, for |
| 57 | /// example, this type's maximum value will never overflow an `isize`, |
| 58 | /// which means it will never overflow a `i16` even though its internal |
| 59 | /// representation is still a `u32`. |
| 60 | /// |
| 61 | /// The purpose for making the type fit into even signed integer types like |
| 62 | /// `isize` is to guarantee that the difference between any two small indices |
| 63 | /// is itself also a small index. This is useful in certain contexts, e.g., |
| 64 | /// for delta encoding. |
| 65 | /// |
| 66 | /// # Other types |
| 67 | /// |
| 68 | /// The following types wrap `SmallIndex` to provide a more focused use case: |
| 69 | /// |
| 70 | /// * [`PatternID`] is for representing the identifiers of patterns. |
| 71 | /// * [`StateID`] is for representing the identifiers of states in finite |
| 72 | /// automata. It is used for both NFAs and DFAs. |
| 73 | /// |
| 74 | /// # Representation |
| 75 | /// |
| 76 | /// This type is always represented internally by a `u32` and is marked as |
| 77 | /// `repr(transparent)`. Thus, this type always has the same representation as |
| 78 | /// a `u32`. It is thus safe to transmute between a `u32` and a `SmallIndex`. |
| 79 | /// |
| 80 | /// # Indexing |
| 81 | /// |
| 82 | /// For convenience, callers may use a `SmallIndex` to index slices. |
| 83 | /// |
| 84 | /// # Safety |
| 85 | /// |
| 86 | /// While a `SmallIndex` is meant to guarantee that its value fits into `usize` |
| 87 | /// without using as much space as a `usize` on all targets, callers must |
| 88 | /// not rely on this property for safety. Callers may choose to rely on this |
| 89 | /// property for correctness however. For example, creating a `SmallIndex` with |
| 90 | /// an invalid value can be done in entirely safe code. This may in turn result |
| 91 | /// in panics or silent logical errors. |
| 92 | #[derive ( |
| 93 | Clone, Copy, Debug, Default, Eq, Hash, PartialEq, PartialOrd, Ord, |
| 94 | )] |
| 95 | #[repr (transparent)] |
| 96 | pub(crate) struct SmallIndex(u32); |
| 97 | |
| 98 | impl SmallIndex { |
| 99 | /// The maximum index value. |
| 100 | #[cfg (any(target_pointer_width = "32" , target_pointer_width = "64" ))] |
| 101 | pub const MAX: SmallIndex = |
| 102 | // FIXME: Use as_usize() once const functions in traits are stable. |
| 103 | SmallIndex::new_unchecked(core::i32::MAX as usize - 1); |
| 104 | |
| 105 | /// The maximum index value. |
| 106 | #[cfg (target_pointer_width = "16" )] |
| 107 | pub const MAX: SmallIndex = |
| 108 | SmallIndex::new_unchecked(core::isize::MAX - 1); |
| 109 | |
| 110 | /// The total number of values that can be represented as a small index. |
| 111 | pub const LIMIT: usize = SmallIndex::MAX.as_usize() + 1; |
| 112 | |
| 113 | /// The zero index value. |
| 114 | pub const ZERO: SmallIndex = SmallIndex::new_unchecked(0); |
| 115 | |
| 116 | /// The number of bytes that a single small index uses in memory. |
| 117 | pub const SIZE: usize = core::mem::size_of::<SmallIndex>(); |
| 118 | |
| 119 | /// Create a new small index. |
| 120 | /// |
| 121 | /// If the given index exceeds [`SmallIndex::MAX`], then this returns |
| 122 | /// an error. |
| 123 | #[inline ] |
| 124 | pub fn new(index: usize) -> Result<SmallIndex, SmallIndexError> { |
| 125 | SmallIndex::try_from(index) |
| 126 | } |
| 127 | |
| 128 | /// Create a new small index without checking whether the given value |
| 129 | /// exceeds [`SmallIndex::MAX`]. |
| 130 | /// |
| 131 | /// Using this routine with an invalid index value will result in |
| 132 | /// unspecified behavior, but *not* undefined behavior. In particular, an |
| 133 | /// invalid index value is likely to cause panics or possibly even silent |
| 134 | /// logical errors. |
| 135 | /// |
| 136 | /// Callers must never rely on a `SmallIndex` to be within a certain range |
| 137 | /// for memory safety. |
| 138 | #[inline ] |
| 139 | pub const fn new_unchecked(index: usize) -> SmallIndex { |
| 140 | // FIXME: Use as_u32() once const functions in traits are stable. |
| 141 | SmallIndex::from_u32_unchecked(index as u32) |
| 142 | } |
| 143 | |
| 144 | /// Create a new small index from a `u32` without checking whether the |
| 145 | /// given value exceeds [`SmallIndex::MAX`]. |
| 146 | /// |
| 147 | /// Using this routine with an invalid index value will result in |
| 148 | /// unspecified behavior, but *not* undefined behavior. In particular, an |
| 149 | /// invalid index value is likely to cause panics or possibly even silent |
| 150 | /// logical errors. |
| 151 | /// |
| 152 | /// Callers must never rely on a `SmallIndex` to be within a certain range |
| 153 | /// for memory safety. |
| 154 | #[inline ] |
| 155 | pub const fn from_u32_unchecked(index: u32) -> SmallIndex { |
| 156 | SmallIndex(index) |
| 157 | } |
| 158 | |
| 159 | /// Like [`SmallIndex::new`], but panics if the given index is not valid. |
| 160 | #[inline ] |
| 161 | pub fn must(index: usize) -> SmallIndex { |
| 162 | SmallIndex::new(index).expect("invalid small index" ) |
| 163 | } |
| 164 | |
| 165 | /// Return this small index as a `usize`. This is guaranteed to never |
| 166 | /// overflow `usize`. |
| 167 | #[inline ] |
| 168 | pub const fn as_usize(&self) -> usize { |
| 169 | // FIXME: Use as_usize() once const functions in traits are stable. |
| 170 | self.0 as usize |
| 171 | } |
| 172 | |
| 173 | /// Return this small index as a `u64`. This is guaranteed to never |
| 174 | /// overflow. |
| 175 | #[inline ] |
| 176 | pub const fn as_u64(&self) -> u64 { |
| 177 | // FIXME: Use u64::from() once const functions in traits are stable. |
| 178 | self.0 as u64 |
| 179 | } |
| 180 | |
| 181 | /// Return the internal `u32` of this small index. This is guaranteed to |
| 182 | /// never overflow `u32`. |
| 183 | #[inline ] |
| 184 | pub const fn as_u32(&self) -> u32 { |
| 185 | self.0 |
| 186 | } |
| 187 | |
| 188 | /// Return the internal `u32` of this small index represented as an `i32`. |
| 189 | /// This is guaranteed to never overflow an `i32`. |
| 190 | #[inline ] |
| 191 | pub const fn as_i32(&self) -> i32 { |
| 192 | // This is OK because we guarantee that our max value is <= i32::MAX. |
| 193 | self.0 as i32 |
| 194 | } |
| 195 | |
| 196 | /// Returns one more than this small index as a usize. |
| 197 | /// |
| 198 | /// Since a small index has constraints on its maximum value, adding `1` to |
| 199 | /// it will always fit in a `usize`, `isize`, `u32` and a `i32`. |
| 200 | #[inline ] |
| 201 | pub fn one_more(&self) -> usize { |
| 202 | self.as_usize() + 1 |
| 203 | } |
| 204 | |
| 205 | /// Decode this small index from the bytes given using the native endian |
| 206 | /// byte order for the current target. |
| 207 | /// |
| 208 | /// If the decoded integer is not representable as a small index for the |
| 209 | /// current target, then this returns an error. |
| 210 | #[inline ] |
| 211 | pub fn from_ne_bytes( |
| 212 | bytes: [u8; 4], |
| 213 | ) -> Result<SmallIndex, SmallIndexError> { |
| 214 | let id = u32::from_ne_bytes(bytes); |
| 215 | if id > SmallIndex::MAX.as_u32() { |
| 216 | return Err(SmallIndexError { attempted: u64::from(id) }); |
| 217 | } |
| 218 | Ok(SmallIndex::new_unchecked(id.as_usize())) |
| 219 | } |
| 220 | |
| 221 | /// Decode this small index from the bytes given using the native endian |
| 222 | /// byte order for the current target. |
| 223 | /// |
| 224 | /// This is analogous to [`SmallIndex::new_unchecked`] in that is does not |
| 225 | /// check whether the decoded integer is representable as a small index. |
| 226 | #[inline ] |
| 227 | pub fn from_ne_bytes_unchecked(bytes: [u8; 4]) -> SmallIndex { |
| 228 | SmallIndex::new_unchecked(u32::from_ne_bytes(bytes).as_usize()) |
| 229 | } |
| 230 | |
| 231 | /// Return the underlying small index integer as raw bytes in native endian |
| 232 | /// format. |
| 233 | #[inline ] |
| 234 | pub fn to_ne_bytes(&self) -> [u8; 4] { |
| 235 | self.0.to_ne_bytes() |
| 236 | } |
| 237 | } |
| 238 | |
| 239 | impl<T> core::ops::Index<SmallIndex> for [T] { |
| 240 | type Output = T; |
| 241 | |
| 242 | #[inline ] |
| 243 | fn index(&self, index: SmallIndex) -> &T { |
| 244 | &self[index.as_usize()] |
| 245 | } |
| 246 | } |
| 247 | |
| 248 | impl<T> core::ops::IndexMut<SmallIndex> for [T] { |
| 249 | #[inline ] |
| 250 | fn index_mut(&mut self, index: SmallIndex) -> &mut T { |
| 251 | &mut self[index.as_usize()] |
| 252 | } |
| 253 | } |
| 254 | |
| 255 | impl<T> core::ops::Index<SmallIndex> for Vec<T> { |
| 256 | type Output = T; |
| 257 | |
| 258 | #[inline ] |
| 259 | fn index(&self, index: SmallIndex) -> &T { |
| 260 | &self[index.as_usize()] |
| 261 | } |
| 262 | } |
| 263 | |
| 264 | impl<T> core::ops::IndexMut<SmallIndex> for Vec<T> { |
| 265 | #[inline ] |
| 266 | fn index_mut(&mut self, index: SmallIndex) -> &mut T { |
| 267 | &mut self[index.as_usize()] |
| 268 | } |
| 269 | } |
| 270 | |
| 271 | impl From<StateID> for SmallIndex { |
| 272 | fn from(sid: StateID) -> SmallIndex { |
| 273 | sid.0 |
| 274 | } |
| 275 | } |
| 276 | |
| 277 | impl From<PatternID> for SmallIndex { |
| 278 | fn from(pid: PatternID) -> SmallIndex { |
| 279 | pid.0 |
| 280 | } |
| 281 | } |
| 282 | |
| 283 | impl From<u8> for SmallIndex { |
| 284 | fn from(index: u8) -> SmallIndex { |
| 285 | SmallIndex::new_unchecked(index:usize::from(index)) |
| 286 | } |
| 287 | } |
| 288 | |
| 289 | impl TryFrom<u16> for SmallIndex { |
| 290 | type Error = SmallIndexError; |
| 291 | |
| 292 | fn try_from(index: u16) -> Result<SmallIndex, SmallIndexError> { |
| 293 | if u32::from(index) > SmallIndex::MAX.as_u32() { |
| 294 | return Err(SmallIndexError { attempted: u64::from(index) }); |
| 295 | } |
| 296 | Ok(SmallIndex::new_unchecked(index.as_usize())) |
| 297 | } |
| 298 | } |
| 299 | |
| 300 | impl TryFrom<u32> for SmallIndex { |
| 301 | type Error = SmallIndexError; |
| 302 | |
| 303 | fn try_from(index: u32) -> Result<SmallIndex, SmallIndexError> { |
| 304 | if index > SmallIndex::MAX.as_u32() { |
| 305 | return Err(SmallIndexError { attempted: u64::from(index) }); |
| 306 | } |
| 307 | Ok(SmallIndex::new_unchecked(index.as_usize())) |
| 308 | } |
| 309 | } |
| 310 | |
| 311 | impl TryFrom<u64> for SmallIndex { |
| 312 | type Error = SmallIndexError; |
| 313 | |
| 314 | fn try_from(index: u64) -> Result<SmallIndex, SmallIndexError> { |
| 315 | if index > SmallIndex::MAX.as_u64() { |
| 316 | return Err(SmallIndexError { attempted: index }); |
| 317 | } |
| 318 | Ok(SmallIndex::new_unchecked(index.as_usize())) |
| 319 | } |
| 320 | } |
| 321 | |
| 322 | impl TryFrom<usize> for SmallIndex { |
| 323 | type Error = SmallIndexError; |
| 324 | |
| 325 | fn try_from(index: usize) -> Result<SmallIndex, SmallIndexError> { |
| 326 | if index > SmallIndex::MAX.as_usize() { |
| 327 | return Err(SmallIndexError { attempted: index.as_u64() }); |
| 328 | } |
| 329 | Ok(SmallIndex::new_unchecked(index)) |
| 330 | } |
| 331 | } |
| 332 | |
| 333 | /// This error occurs when a small index could not be constructed. |
| 334 | /// |
| 335 | /// This occurs when given an integer exceeding the maximum small index value. |
| 336 | /// |
| 337 | /// When the `std` feature is enabled, this implements the `Error` trait. |
| 338 | #[derive (Clone, Debug, Eq, PartialEq)] |
| 339 | pub struct SmallIndexError { |
| 340 | attempted: u64, |
| 341 | } |
| 342 | |
| 343 | impl SmallIndexError { |
| 344 | /// Returns the value that could not be converted to a small index. |
| 345 | pub fn attempted(&self) -> u64 { |
| 346 | self.attempted |
| 347 | } |
| 348 | } |
| 349 | |
| 350 | #[cfg (feature = "std" )] |
| 351 | impl std::error::Error for SmallIndexError {} |
| 352 | |
| 353 | impl core::fmt::Display for SmallIndexError { |
| 354 | fn fmt(&self, f: &mut core::fmt::Formatter) -> core::fmt::Result { |
| 355 | write!( |
| 356 | f, |
| 357 | "failed to create small index from {:?}, which exceeds {:?}" , |
| 358 | self.attempted(), |
| 359 | SmallIndex::MAX, |
| 360 | ) |
| 361 | } |
| 362 | } |
| 363 | |
| 364 | #[derive (Clone, Debug)] |
| 365 | pub(crate) struct SmallIndexIter { |
| 366 | rng: core::ops::Range<usize>, |
| 367 | } |
| 368 | |
| 369 | impl Iterator for SmallIndexIter { |
| 370 | type Item = SmallIndex; |
| 371 | |
| 372 | fn next(&mut self) -> Option<SmallIndex> { |
| 373 | if self.rng.start >= self.rng.end { |
| 374 | return None; |
| 375 | } |
| 376 | let next_id: usize = self.rng.start + 1; |
| 377 | let id: usize = core::mem::replace(&mut self.rng.start, src:next_id); |
| 378 | // new_unchecked is OK since we asserted that the number of |
| 379 | // elements in this iterator will fit in an ID at construction. |
| 380 | Some(SmallIndex::new_unchecked(index:id)) |
| 381 | } |
| 382 | } |
| 383 | |
| 384 | macro_rules! index_type_impls { |
| 385 | ($name:ident, $err:ident, $iter:ident, $withiter:ident) => { |
| 386 | impl $name { |
| 387 | /// The maximum value. |
| 388 | pub const MAX: $name = $name(SmallIndex::MAX); |
| 389 | |
| 390 | /// The total number of values that can be represented. |
| 391 | pub const LIMIT: usize = SmallIndex::LIMIT; |
| 392 | |
| 393 | /// The zero value. |
| 394 | pub const ZERO: $name = $name(SmallIndex::ZERO); |
| 395 | |
| 396 | /// The number of bytes that a single value uses in memory. |
| 397 | pub const SIZE: usize = SmallIndex::SIZE; |
| 398 | |
| 399 | /// Create a new value that is represented by a "small index." |
| 400 | /// |
| 401 | /// If the given index exceeds the maximum allowed value, then this |
| 402 | /// returns an error. |
| 403 | #[inline] |
| 404 | pub fn new(value: usize) -> Result<$name, $err> { |
| 405 | SmallIndex::new(value).map($name).map_err($err) |
| 406 | } |
| 407 | |
| 408 | /// Create a new value without checking whether the given argument |
| 409 | /// exceeds the maximum. |
| 410 | /// |
| 411 | /// Using this routine with an invalid value will result in |
| 412 | /// unspecified behavior, but *not* undefined behavior. In |
| 413 | /// particular, an invalid ID value is likely to cause panics or |
| 414 | /// possibly even silent logical errors. |
| 415 | /// |
| 416 | /// Callers must never rely on this type to be within a certain |
| 417 | /// range for memory safety. |
| 418 | #[inline] |
| 419 | pub const fn new_unchecked(value: usize) -> $name { |
| 420 | $name(SmallIndex::new_unchecked(value)) |
| 421 | } |
| 422 | |
| 423 | /// Create a new value from a `u32` without checking whether the |
| 424 | /// given value exceeds the maximum. |
| 425 | /// |
| 426 | /// Using this routine with an invalid value will result in |
| 427 | /// unspecified behavior, but *not* undefined behavior. In |
| 428 | /// particular, an invalid ID value is likely to cause panics or |
| 429 | /// possibly even silent logical errors. |
| 430 | /// |
| 431 | /// Callers must never rely on this type to be within a certain |
| 432 | /// range for memory safety. |
| 433 | #[inline] |
| 434 | pub const fn from_u32_unchecked(index: u32) -> $name { |
| 435 | $name(SmallIndex::from_u32_unchecked(index)) |
| 436 | } |
| 437 | |
| 438 | /// Like `new`, but panics if the given value is not valid. |
| 439 | #[inline] |
| 440 | pub fn must(value: usize) -> $name { |
| 441 | $name::new(value).expect(concat!( |
| 442 | "invalid " , |
| 443 | stringify!($name), |
| 444 | " value" |
| 445 | )) |
| 446 | } |
| 447 | |
| 448 | /// Return the internal value as a `usize`. This is guaranteed to |
| 449 | /// never overflow `usize`. |
| 450 | #[inline] |
| 451 | pub const fn as_usize(&self) -> usize { |
| 452 | self.0.as_usize() |
| 453 | } |
| 454 | |
| 455 | /// Return the internal value as a `u64`. This is guaranteed to |
| 456 | /// never overflow. |
| 457 | #[inline] |
| 458 | pub const fn as_u64(&self) -> u64 { |
| 459 | self.0.as_u64() |
| 460 | } |
| 461 | |
| 462 | /// Return the internal value as a `u32`. This is guaranteed to |
| 463 | /// never overflow `u32`. |
| 464 | #[inline] |
| 465 | pub const fn as_u32(&self) -> u32 { |
| 466 | self.0.as_u32() |
| 467 | } |
| 468 | |
| 469 | /// Return the internal value as a `i32`. This is guaranteed to |
| 470 | /// never overflow an `i32`. |
| 471 | #[inline] |
| 472 | pub const fn as_i32(&self) -> i32 { |
| 473 | self.0.as_i32() |
| 474 | } |
| 475 | |
| 476 | /// Returns one more than this value as a usize. |
| 477 | /// |
| 478 | /// Since values represented by a "small index" have constraints |
| 479 | /// on their maximum value, adding `1` to it will always fit in a |
| 480 | /// `usize`, `u32` and a `i32`. |
| 481 | #[inline] |
| 482 | pub fn one_more(&self) -> usize { |
| 483 | self.0.one_more() |
| 484 | } |
| 485 | |
| 486 | /// Decode this value from the bytes given using the native endian |
| 487 | /// byte order for the current target. |
| 488 | /// |
| 489 | /// If the decoded integer is not representable as a small index |
| 490 | /// for the current target, then this returns an error. |
| 491 | #[inline] |
| 492 | pub fn from_ne_bytes(bytes: [u8; 4]) -> Result<$name, $err> { |
| 493 | SmallIndex::from_ne_bytes(bytes).map($name).map_err($err) |
| 494 | } |
| 495 | |
| 496 | /// Decode this value from the bytes given using the native endian |
| 497 | /// byte order for the current target. |
| 498 | /// |
| 499 | /// This is analogous to `new_unchecked` in that is does not check |
| 500 | /// whether the decoded integer is representable as a small index. |
| 501 | #[inline] |
| 502 | pub fn from_ne_bytes_unchecked(bytes: [u8; 4]) -> $name { |
| 503 | $name(SmallIndex::from_ne_bytes_unchecked(bytes)) |
| 504 | } |
| 505 | |
| 506 | /// Return the underlying integer as raw bytes in native endian |
| 507 | /// format. |
| 508 | #[inline] |
| 509 | pub fn to_ne_bytes(&self) -> [u8; 4] { |
| 510 | self.0.to_ne_bytes() |
| 511 | } |
| 512 | |
| 513 | /// Returns an iterator over all values from 0 up to and not |
| 514 | /// including the given length. |
| 515 | /// |
| 516 | /// If the given length exceeds this type's limit, then this |
| 517 | /// panics. |
| 518 | pub(crate) fn iter(len: usize) -> $iter { |
| 519 | $iter::new(len) |
| 520 | } |
| 521 | } |
| 522 | |
| 523 | // We write our own Debug impl so that we get things like PatternID(5) |
| 524 | // instead of PatternID(SmallIndex(5)). |
| 525 | impl core::fmt::Debug for $name { |
| 526 | fn fmt(&self, f: &mut core::fmt::Formatter) -> core::fmt::Result { |
| 527 | f.debug_tuple(stringify!($name)).field(&self.as_u32()).finish() |
| 528 | } |
| 529 | } |
| 530 | |
| 531 | impl<T> core::ops::Index<$name> for [T] { |
| 532 | type Output = T; |
| 533 | |
| 534 | #[inline] |
| 535 | fn index(&self, index: $name) -> &T { |
| 536 | &self[index.as_usize()] |
| 537 | } |
| 538 | } |
| 539 | |
| 540 | impl<T> core::ops::IndexMut<$name> for [T] { |
| 541 | #[inline] |
| 542 | fn index_mut(&mut self, index: $name) -> &mut T { |
| 543 | &mut self[index.as_usize()] |
| 544 | } |
| 545 | } |
| 546 | |
| 547 | impl<T> core::ops::Index<$name> for Vec<T> { |
| 548 | type Output = T; |
| 549 | |
| 550 | #[inline] |
| 551 | fn index(&self, index: $name) -> &T { |
| 552 | &self[index.as_usize()] |
| 553 | } |
| 554 | } |
| 555 | |
| 556 | impl<T> core::ops::IndexMut<$name> for Vec<T> { |
| 557 | #[inline] |
| 558 | fn index_mut(&mut self, index: $name) -> &mut T { |
| 559 | &mut self[index.as_usize()] |
| 560 | } |
| 561 | } |
| 562 | |
| 563 | impl From<SmallIndex> for $name { |
| 564 | fn from(index: SmallIndex) -> $name { |
| 565 | $name(index) |
| 566 | } |
| 567 | } |
| 568 | |
| 569 | impl From<u8> for $name { |
| 570 | fn from(value: u8) -> $name { |
| 571 | $name(SmallIndex::from(value)) |
| 572 | } |
| 573 | } |
| 574 | |
| 575 | impl TryFrom<u16> for $name { |
| 576 | type Error = $err; |
| 577 | |
| 578 | fn try_from(value: u16) -> Result<$name, $err> { |
| 579 | SmallIndex::try_from(value).map($name).map_err($err) |
| 580 | } |
| 581 | } |
| 582 | |
| 583 | impl TryFrom<u32> for $name { |
| 584 | type Error = $err; |
| 585 | |
| 586 | fn try_from(value: u32) -> Result<$name, $err> { |
| 587 | SmallIndex::try_from(value).map($name).map_err($err) |
| 588 | } |
| 589 | } |
| 590 | |
| 591 | impl TryFrom<u64> for $name { |
| 592 | type Error = $err; |
| 593 | |
| 594 | fn try_from(value: u64) -> Result<$name, $err> { |
| 595 | SmallIndex::try_from(value).map($name).map_err($err) |
| 596 | } |
| 597 | } |
| 598 | |
| 599 | impl TryFrom<usize> for $name { |
| 600 | type Error = $err; |
| 601 | |
| 602 | fn try_from(value: usize) -> Result<$name, $err> { |
| 603 | SmallIndex::try_from(value).map($name).map_err($err) |
| 604 | } |
| 605 | } |
| 606 | |
| 607 | /// This error occurs when an ID could not be constructed. |
| 608 | /// |
| 609 | /// This occurs when given an integer exceeding the maximum allowed |
| 610 | /// value. |
| 611 | /// |
| 612 | /// When the `std` feature is enabled, this implements the `Error` |
| 613 | /// trait. |
| 614 | #[derive(Clone, Debug, Eq, PartialEq)] |
| 615 | pub struct $err(SmallIndexError); |
| 616 | |
| 617 | impl $err { |
| 618 | /// Returns the value that could not be converted to an ID. |
| 619 | pub fn attempted(&self) -> u64 { |
| 620 | self.0.attempted() |
| 621 | } |
| 622 | } |
| 623 | |
| 624 | #[cfg(feature = "std" )] |
| 625 | impl std::error::Error for $err {} |
| 626 | |
| 627 | impl core::fmt::Display for $err { |
| 628 | fn fmt(&self, f: &mut core::fmt::Formatter) -> core::fmt::Result { |
| 629 | write!( |
| 630 | f, |
| 631 | "failed to create {} from {:?}, which exceeds {:?}" , |
| 632 | stringify!($name), |
| 633 | self.attempted(), |
| 634 | $name::MAX, |
| 635 | ) |
| 636 | } |
| 637 | } |
| 638 | |
| 639 | #[derive(Clone, Debug)] |
| 640 | pub(crate) struct $iter(SmallIndexIter); |
| 641 | |
| 642 | impl $iter { |
| 643 | fn new(len: usize) -> $iter { |
| 644 | assert!( |
| 645 | len <= $name::LIMIT, |
| 646 | "cannot create iterator for {} when number of \ |
| 647 | elements exceed {:?}" , |
| 648 | stringify!($name), |
| 649 | $name::LIMIT, |
| 650 | ); |
| 651 | $iter(SmallIndexIter { rng: 0..len }) |
| 652 | } |
| 653 | } |
| 654 | |
| 655 | impl Iterator for $iter { |
| 656 | type Item = $name; |
| 657 | |
| 658 | fn next(&mut self) -> Option<$name> { |
| 659 | self.0.next().map($name) |
| 660 | } |
| 661 | } |
| 662 | |
| 663 | /// An iterator adapter that is like std::iter::Enumerate, but attaches |
| 664 | /// small index values instead. It requires `ExactSizeIterator`. At |
| 665 | /// construction, it ensures that the index of each element in the |
| 666 | /// iterator is representable in the corresponding small index type. |
| 667 | #[derive(Clone, Debug)] |
| 668 | pub(crate) struct $withiter<I> { |
| 669 | it: I, |
| 670 | ids: $iter, |
| 671 | } |
| 672 | |
| 673 | impl<I: Iterator + ExactSizeIterator> $withiter<I> { |
| 674 | fn new(it: I) -> $withiter<I> { |
| 675 | let ids = $name::iter(it.len()); |
| 676 | $withiter { it, ids } |
| 677 | } |
| 678 | } |
| 679 | |
| 680 | impl<I: Iterator + ExactSizeIterator> Iterator for $withiter<I> { |
| 681 | type Item = ($name, I::Item); |
| 682 | |
| 683 | fn next(&mut self) -> Option<($name, I::Item)> { |
| 684 | let item = self.it.next()?; |
| 685 | // Number of elements in this iterator must match, according |
| 686 | // to contract of ExactSizeIterator. |
| 687 | let id = self.ids.next().unwrap(); |
| 688 | Some((id, item)) |
| 689 | } |
| 690 | } |
| 691 | }; |
| 692 | } |
| 693 | |
| 694 | /// The identifier of a pattern in an Aho-Corasick automaton. |
| 695 | /// |
| 696 | /// It is represented by a `u32` even on 64-bit systems in order to conserve |
| 697 | /// space. Namely, on all targets, this type guarantees that its value will |
| 698 | /// fit in a `u32`, `i32`, `usize` and an `isize`. This means that on 16-bit |
| 699 | /// targets, for example, this type's maximum value will never overflow an |
| 700 | /// `isize`, which means it will never overflow a `i16` even though its |
| 701 | /// internal representation is still a `u32`. |
| 702 | /// |
| 703 | /// # Safety |
| 704 | /// |
| 705 | /// While a `PatternID` is meant to guarantee that its value fits into `usize` |
| 706 | /// without using as much space as a `usize` on all targets, callers must |
| 707 | /// not rely on this property for safety. Callers may choose to rely on this |
| 708 | /// property for correctness however. For example, creating a `StateID` with an |
| 709 | /// invalid value can be done in entirely safe code. This may in turn result in |
| 710 | /// panics or silent logical errors. |
| 711 | #[derive (Clone, Copy, Default, Eq, Hash, PartialEq, PartialOrd, Ord)] |
| 712 | #[repr (transparent)] |
| 713 | pub struct PatternID(SmallIndex); |
| 714 | |
| 715 | /// The identifier of a finite automaton state. |
| 716 | /// |
| 717 | /// It is represented by a `u32` even on 64-bit systems in order to conserve |
| 718 | /// space. Namely, on all targets, this type guarantees that its value will |
| 719 | /// fit in a `u32`, `i32`, `usize` and an `isize`. This means that on 16-bit |
| 720 | /// targets, for example, this type's maximum value will never overflow an |
| 721 | /// `isize`, which means it will never overflow a `i16` even though its |
| 722 | /// internal representation is still a `u32`. |
| 723 | /// |
| 724 | /// # Safety |
| 725 | /// |
| 726 | /// While a `StateID` is meant to guarantee that its value fits into `usize` |
| 727 | /// without using as much space as a `usize` on all targets, callers must |
| 728 | /// not rely on this property for safety. Callers may choose to rely on this |
| 729 | /// property for correctness however. For example, creating a `StateID` with an |
| 730 | /// invalid value can be done in entirely safe code. This may in turn result in |
| 731 | /// panics or silent logical errors. |
| 732 | #[derive (Clone, Copy, Default, Eq, Hash, PartialEq, PartialOrd, Ord)] |
| 733 | #[repr (transparent)] |
| 734 | pub struct StateID(SmallIndex); |
| 735 | |
| 736 | index_type_impls!(PatternID, PatternIDError, PatternIDIter, WithPatternIDIter); |
| 737 | index_type_impls!(StateID, StateIDError, StateIDIter, WithStateIDIter); |
| 738 | |
| 739 | /// A utility trait that defines a couple of adapters for making it convenient |
| 740 | /// to access indices as "small index" types. We require ExactSizeIterator so |
| 741 | /// that iterator construction can do a single check to make sure the index of |
| 742 | /// each element is representable by its small index type. |
| 743 | pub(crate) trait IteratorIndexExt: Iterator { |
| 744 | fn with_pattern_ids(self) -> WithPatternIDIter<Self> |
| 745 | where |
| 746 | Self: Sized + ExactSizeIterator, |
| 747 | { |
| 748 | WithPatternIDIter::new(self) |
| 749 | } |
| 750 | |
| 751 | fn with_state_ids(self) -> WithStateIDIter<Self> |
| 752 | where |
| 753 | Self: Sized + ExactSizeIterator, |
| 754 | { |
| 755 | WithStateIDIter::new(self) |
| 756 | } |
| 757 | } |
| 758 | |
| 759 | impl<I: Iterator> IteratorIndexExt for I {} |
| 760 | |