| 1 |
/*! |
| 2 |
This module provides forward and reverse substring search routines. |
| 3 |
|
| 4 |
Unlike the standard library's substring search routines, these work on |
| 5 |
arbitrary bytes. For all non-empty needles, these routines will report exactly |
| 6 |
the same values as the corresponding routines in the standard library. For |
| 7 |
the empty needle, the standard library reports matches only at valid UTF-8 |
| 8 |
boundaries, where as these routines will report matches at every position. |
| 9 |
|
| 10 |
Other than being able to work on arbitrary bytes, the primary reason to prefer |
| 11 |
these routines over the standard library routines is that these will generally |
| 12 |
be faster. In some cases, significantly so. |
| 13 |
|
| 14 |
# Example: iterating over substring matches |
| 15 |
|
| 16 |
This example shows how to use [`find_iter`] to find occurrences of a substring |
| 17 |
in a haystack. |
| 18 |
|
| 19 |
``` |
| 20 |
use memchr::memmem; |
| 21 |
|
| 22 |
let haystack = b"foo bar foo baz foo"; |
| 23 |
|
| 24 |
let mut it = memmem::find_iter(haystack, "foo"); |
| 25 |
assert_eq!(Some(0), it.next()); |
| 26 |
assert_eq!(Some(8), it.next()); |
| 27 |
assert_eq!(Some(16), it.next()); |
| 28 |
assert_eq!(None, it.next()); |
| 29 |
``` |
| 30 |
|
| 31 |
# Example: iterating over substring matches in reverse |
| 32 |
|
| 33 |
This example shows how to use [`rfind_iter`] to find occurrences of a substring |
| 34 |
in a haystack starting from the end of the haystack. |
| 35 |
|
| 36 |
**NOTE:** This module does not implement double ended iterators, so reverse |
| 37 |
searches aren't done by calling `rev` on a forward iterator. |
| 38 |
|
| 39 |
``` |
| 40 |
use memchr::memmem; |
| 41 |
|
| 42 |
let haystack = b"foo bar foo baz foo"; |
| 43 |
|
| 44 |
let mut it = memmem::rfind_iter(haystack, "foo"); |
| 45 |
assert_eq!(Some(16), it.next()); |
| 46 |
assert_eq!(Some(8), it.next()); |
| 47 |
assert_eq!(Some(0), it.next()); |
| 48 |
assert_eq!(None, it.next()); |
| 49 |
``` |
| 50 |
|
| 51 |
# Example: repeating a search for the same needle |
| 52 |
|
| 53 |
It may be possible for the overhead of constructing a substring searcher to be |
| 54 |
measurable in some workloads. In cases where the same needle is used to search |
| 55 |
many haystacks, it is possible to do construction once and thus to avoid it for |
| 56 |
subsequent searches. This can be done with a [`Finder`] (or a [`FinderRev`] for |
| 57 |
reverse searches). |
| 58 |
|
| 59 |
``` |
| 60 |
use memchr::memmem; |
| 61 |
|
| 62 |
let finder = memmem::Finder::new("foo"); |
| 63 |
|
| 64 |
assert_eq!(Some(4), finder.find(b"baz foo quux")); |
| 65 |
assert_eq!(None, finder.find(b"quux baz bar")); |
| 66 |
``` |
| 67 |
*/ |
| 68 |
|
| 69 |
pub use crate::memmem::searcher::PrefilterConfig as Prefilter; |
| 70 |
|
| 71 |
// This is exported here for use in the crate::arch::all::twoway |
| 72 |
// implementation. This is essentially an abstraction breaker. Namely, the |
| 73 |
// public API of twoway doesn't support providing a prefilter, but its crate |
| 74 |
// internal API does. The main reason for this is that I didn't want to do the |
| 75 |
// API design required to support it without a concrete use case. |
| 76 |
pub(crate) use crate::memmem::searcher::Pre; |
| 77 |
|
| 78 |
use crate::{ |
| 79 |
arch::all::{ |
| 80 |
packedpair::{DefaultFrequencyRank, HeuristicFrequencyRank}, |
| 81 |
rabinkarp, |
| 82 |
}, |
| 83 |
cow::CowBytes, |
| 84 |
memmem::searcher::{PrefilterState, Searcher, SearcherRev}, |
| 85 |
}; |
| 86 |
|
| 87 |
mod searcher; |
| 88 |
|
| 89 |
/// Returns an iterator over all non-overlapping occurrences of a substring in |
| 90 |
/// a haystack. |
| 91 |
/// |
| 92 |
/// # Complexity |
| 93 |
/// |
| 94 |
/// This routine is guaranteed to have worst case linear time complexity |
| 95 |
/// with respect to both the needle and the haystack. That is, this runs |
| 96 |
/// in `O(needle.len() + haystack.len())` time. |
| 97 |
/// |
| 98 |
/// This routine is also guaranteed to have worst case constant space |
| 99 |
/// complexity. |
| 100 |
/// |
| 101 |
/// # Examples |
| 102 |
/// |
| 103 |
/// Basic usage: |
| 104 |
/// |
| 105 |
/// ``` |
| 106 |
/// use memchr::memmem; |
| 107 |
/// |
| 108 |
/// let haystack = b"foo bar foo baz foo"; |
| 109 |
/// let mut it = memmem::find_iter(haystack, b"foo"); |
| 110 |
/// assert_eq!(Some(0), it.next()); |
| 111 |
/// assert_eq!(Some(8), it.next()); |
| 112 |
/// assert_eq!(Some(16), it.next()); |
| 113 |
/// assert_eq!(None, it.next()); |
| 114 |
/// ``` |
| 115 |
#[inline] |
| 116 |
pub fn find_iter<'h, 'n, N: 'n + ?Sized + AsRef<[u8]>>( |
| 117 |
haystack: &'h [u8], |
| 118 |
needle: &'n N, |
| 119 |
) -> FindIter<'h, 'n> { |
| 120 |
FindIter::new(haystack, Finder::new(needle)) |
| 121 |
} |
| 122 |
|
| 123 |
/// Returns a reverse iterator over all non-overlapping occurrences of a |
| 124 |
/// substring in a haystack. |
| 125 |
/// |
| 126 |
/// # Complexity |
| 127 |
/// |
| 128 |
/// This routine is guaranteed to have worst case linear time complexity |
| 129 |
/// with respect to both the needle and the haystack. That is, this runs |
| 130 |
/// in `O(needle.len() + haystack.len())` time. |
| 131 |
/// |
| 132 |
/// This routine is also guaranteed to have worst case constant space |
| 133 |
/// complexity. |
| 134 |
/// |
| 135 |
/// # Examples |
| 136 |
/// |
| 137 |
/// Basic usage: |
| 138 |
/// |
| 139 |
/// ``` |
| 140 |
/// use memchr::memmem; |
| 141 |
/// |
| 142 |
/// let haystack = b"foo bar foo baz foo"; |
| 143 |
/// let mut it = memmem::rfind_iter(haystack, b"foo"); |
| 144 |
/// assert_eq!(Some(16), it.next()); |
| 145 |
/// assert_eq!(Some(8), it.next()); |
| 146 |
/// assert_eq!(Some(0), it.next()); |
| 147 |
/// assert_eq!(None, it.next()); |
| 148 |
/// ``` |
| 149 |
#[inline] |
| 150 |
pub fn rfind_iter<'h, 'n, N: 'n + ?Sized + AsRef<[u8]>>( |
| 151 |
haystack: &'h [u8], |
| 152 |
needle: &'n N, |
| 153 |
) -> FindRevIter<'h, 'n> { |
| 154 |
FindRevIter::new(haystack, FinderRev::new(needle)) |
| 155 |
} |
| 156 |
|
| 157 |
/// Returns the index of the first occurrence of the given needle. |
| 158 |
/// |
| 159 |
/// Note that if you're are searching for the same needle in many different |
| 160 |
/// small haystacks, it may be faster to initialize a [`Finder`] once, |
| 161 |
/// and reuse it for each search. |
| 162 |
/// |
| 163 |
/// # Complexity |
| 164 |
/// |
| 165 |
/// This routine is guaranteed to have worst case linear time complexity |
| 166 |
/// with respect to both the needle and the haystack. That is, this runs |
| 167 |
/// in `O(needle.len() + haystack.len())` time. |
| 168 |
/// |
| 169 |
/// This routine is also guaranteed to have worst case constant space |
| 170 |
/// complexity. |
| 171 |
/// |
| 172 |
/// # Examples |
| 173 |
/// |
| 174 |
/// Basic usage: |
| 175 |
/// |
| 176 |
/// ``` |
| 177 |
/// use memchr::memmem; |
| 178 |
/// |
| 179 |
/// let haystack = b"foo bar baz"; |
| 180 |
/// assert_eq!(Some(0), memmem::find(haystack, b"foo")); |
| 181 |
/// assert_eq!(Some(4), memmem::find(haystack, b"bar")); |
| 182 |
/// assert_eq!(None, memmem::find(haystack, b"quux")); |
| 183 |
/// ``` |
| 184 |
#[inline] |
| 185 |
pub fn find(haystack: &[u8], needle: &[u8]) -> Option<usize> { |
| 186 |
if haystack.len() < 64 { |
| 187 |
rabinkarp::Finder::new(needle).find(haystack, needle) |
| 188 |
} else { |
| 189 |
Finder::new(needle).find(haystack) |
| 190 |
} |
| 191 |
} |
| 192 |
|
| 193 |
/// Returns the index of the last occurrence of the given needle. |
| 194 |
/// |
| 195 |
/// Note that if you're are searching for the same needle in many different |
| 196 |
/// small haystacks, it may be faster to initialize a [`FinderRev`] once, |
| 197 |
/// and reuse it for each search. |
| 198 |
/// |
| 199 |
/// # Complexity |
| 200 |
/// |
| 201 |
/// This routine is guaranteed to have worst case linear time complexity |
| 202 |
/// with respect to both the needle and the haystack. That is, this runs |
| 203 |
/// in `O(needle.len() + haystack.len())` time. |
| 204 |
/// |
| 205 |
/// This routine is also guaranteed to have worst case constant space |
| 206 |
/// complexity. |
| 207 |
/// |
| 208 |
/// # Examples |
| 209 |
/// |
| 210 |
/// Basic usage: |
| 211 |
/// |
| 212 |
/// ``` |
| 213 |
/// use memchr::memmem; |
| 214 |
/// |
| 215 |
/// let haystack = b"foo bar baz"; |
| 216 |
/// assert_eq!(Some(0), memmem::rfind(haystack, b"foo")); |
| 217 |
/// assert_eq!(Some(4), memmem::rfind(haystack, b"bar")); |
| 218 |
/// assert_eq!(Some(8), memmem::rfind(haystack, b"ba")); |
| 219 |
/// assert_eq!(None, memmem::rfind(haystack, b"quux")); |
| 220 |
/// ``` |
| 221 |
#[inline] |
| 222 |
pub fn rfind(haystack: &[u8], needle: &[u8]) -> Option<usize> { |
| 223 |
if haystack.len() < 64 { |
| 224 |
rabinkarp::FinderRev::new(needle).rfind(haystack, needle) |
| 225 |
} else { |
| 226 |
FinderRev::new(needle).rfind(haystack) |
| 227 |
} |
| 228 |
} |
| 229 |
|
| 230 |
/// An iterator over non-overlapping substring matches. |
| 231 |
/// |
| 232 |
/// Matches are reported by the byte offset at which they begin. |
| 233 |
/// |
| 234 |
/// `'h` is the lifetime of the haystack while `'n` is the lifetime of the |
| 235 |
/// needle. |
| 236 |
#[derive(Debug, Clone)] |
| 237 |
pub struct FindIter<'h, 'n> { |
| 238 |
haystack: &'h [u8], |
| 239 |
prestate: PrefilterState, |
| 240 |
finder: Finder<'n>, |
| 241 |
pos: usize, |
| 242 |
} |
| 243 |
|
| 244 |
impl<'h, 'n> FindIter<'h, 'n> { |
| 245 |
#[inline(always)] |
| 246 |
pub(crate) fn new( |
| 247 |
haystack: &'h [u8], |
| 248 |
finder: Finder<'n>, |
| 249 |
) -> FindIter<'h, 'n> { |
| 250 |
let prestate = PrefilterState::new(); |
| 251 |
FindIter { haystack, prestate, finder, pos: 0 } |
| 252 |
} |
| 253 |
|
| 254 |
/// Convert this iterator into its owned variant, such that it no longer |
| 255 |
/// borrows the finder and needle. |
| 256 |
/// |
| 257 |
/// If this is already an owned iterator, then this is a no-op. Otherwise, |
| 258 |
/// this copies the needle. |
| 259 |
/// |
| 260 |
/// This is only available when the `alloc` feature is enabled. |
| 261 |
#[cfg(feature = "alloc")] |
| 262 |
#[inline] |
| 263 |
pub fn into_owned(self) -> FindIter<'h, 'static> { |
| 264 |
FindIter { |
| 265 |
haystack: self.haystack, |
| 266 |
prestate: self.prestate, |
| 267 |
finder: self.finder.into_owned(), |
| 268 |
pos: self.pos, |
| 269 |
} |
| 270 |
} |
| 271 |
} |
| 272 |
|
| 273 |
impl<'h, 'n> Iterator for FindIter<'h, 'n> { |
| 274 |
type Item = usize; |
| 275 |
|
| 276 |
fn next(&mut self) -> Option<usize> { |
| 277 |
let needle = self.finder.needle(); |
| 278 |
let haystack = self.haystack.get(self.pos..)?; |
| 279 |
let idx = |
| 280 |
self.finder.searcher.find(&mut self.prestate, haystack, needle)?; |
| 281 |
|
| 282 |
let pos = self.pos + idx; |
| 283 |
self.pos = pos + needle.len().max(1); |
| 284 |
|
| 285 |
Some(pos) |
| 286 |
} |
| 287 |
|
| 288 |
fn size_hint(&self) -> (usize, Option<usize>) { |
| 289 |
// The largest possible number of non-overlapping matches is the |
| 290 |
// quotient of the haystack and the needle (or the length of the |
| 291 |
// haystack, if the needle is empty) |
| 292 |
match self.haystack.len().checked_sub(self.pos) { |
| 293 |
None => (0, Some(0)), |
| 294 |
Some(haystack_len) => match self.finder.needle().len() { |
| 295 |
// Empty needles always succeed and match at every point |
| 296 |
// (including the very end) |
| 297 |
0 => ( |
| 298 |
haystack_len.saturating_add(1), |
| 299 |
haystack_len.checked_add(1), |
| 300 |
), |
| 301 |
needle_len => (0, Some(haystack_len / needle_len)), |
| 302 |
}, |
| 303 |
} |
| 304 |
} |
| 305 |
} |
| 306 |
|
| 307 |
/// An iterator over non-overlapping substring matches in reverse. |
| 308 |
/// |
| 309 |
/// Matches are reported by the byte offset at which they begin. |
| 310 |
/// |
| 311 |
/// `'h` is the lifetime of the haystack while `'n` is the lifetime of the |
| 312 |
/// needle. |
| 313 |
#[derive(Clone, Debug)] |
| 314 |
pub struct FindRevIter<'h, 'n> { |
| 315 |
haystack: &'h [u8], |
| 316 |
finder: FinderRev<'n>, |
| 317 |
/// When searching with an empty needle, this gets set to `None` after |
| 318 |
/// we've yielded the last element at `0`. |
| 319 |
pos: Option<usize>, |
| 320 |
} |
| 321 |
|
| 322 |
impl<'h, 'n> FindRevIter<'h, 'n> { |
| 323 |
#[inline(always)] |
| 324 |
pub(crate) fn new( |
| 325 |
haystack: &'h [u8], |
| 326 |
finder: FinderRev<'n>, |
| 327 |
) -> FindRevIter<'h, 'n> { |
| 328 |
let pos = Some(haystack.len()); |
| 329 |
FindRevIter { haystack, finder, pos } |
| 330 |
} |
| 331 |
|
| 332 |
/// Convert this iterator into its owned variant, such that it no longer |
| 333 |
/// borrows the finder and needle. |
| 334 |
/// |
| 335 |
/// If this is already an owned iterator, then this is a no-op. Otherwise, |
| 336 |
/// this copies the needle. |
| 337 |
/// |
| 338 |
/// This is only available when the `std` feature is enabled. |
| 339 |
#[cfg(feature = "alloc")] |
| 340 |
#[inline] |
| 341 |
pub fn into_owned(self) -> FindRevIter<'h, 'static> { |
| 342 |
FindRevIter { |
| 343 |
haystack: self.haystack, |
| 344 |
finder: self.finder.into_owned(), |
| 345 |
pos: self.pos, |
| 346 |
} |
| 347 |
} |
| 348 |
} |
| 349 |
|
| 350 |
impl<'h, 'n> Iterator for FindRevIter<'h, 'n> { |
| 351 |
type Item = usize; |
| 352 |
|
| 353 |
fn next(&mut self) -> Option<usize> { |
| 354 |
let pos = match self.pos { |
| 355 |
None => return None, |
| 356 |
Some(pos) => pos, |
| 357 |
}; |
| 358 |
let result = self.finder.rfind(&self.haystack[..pos]); |
| 359 |
match result { |
| 360 |
None => None, |
| 361 |
Some(i) => { |
| 362 |
if pos == i { |
| 363 |
self.pos = pos.checked_sub(1); |
| 364 |
} else { |
| 365 |
self.pos = Some(i); |
| 366 |
} |
| 367 |
Some(i) |
| 368 |
} |
| 369 |
} |
| 370 |
} |
| 371 |
} |
| 372 |
|
| 373 |
/// A single substring searcher fixed to a particular needle. |
| 374 |
/// |
| 375 |
/// The purpose of this type is to permit callers to construct a substring |
| 376 |
/// searcher that can be used to search haystacks without the overhead of |
| 377 |
/// constructing the searcher in the first place. This is a somewhat niche |
| 378 |
/// concern when it's necessary to re-use the same needle to search multiple |
| 379 |
/// different haystacks with as little overhead as possible. In general, using |
| 380 |
/// [`find`] is good enough, but `Finder` is useful when you can meaningfully |
| 381 |
/// observe searcher construction time in a profile. |
| 382 |
/// |
| 383 |
/// When the `std` feature is enabled, then this type has an `into_owned` |
| 384 |
/// version which permits building a `Finder` that is not connected to |
| 385 |
/// the lifetime of its needle. |
| 386 |
#[derive(Clone, Debug)] |
| 387 |
pub struct Finder<'n> { |
| 388 |
needle: CowBytes<'n>, |
| 389 |
searcher: Searcher, |
| 390 |
} |
| 391 |
|
| 392 |
impl<'n> Finder<'n> { |
| 393 |
/// Create a new finder for the given needle. |
| 394 |
#[inline] |
| 395 |
pub fn new<B: ?Sized + AsRef<[u8]>>(needle: &'n B) -> Finder<'n> { |
| 396 |
FinderBuilder::new().build_forward(needle) |
| 397 |
} |
| 398 |
|
| 399 |
/// Returns the index of the first occurrence of this needle in the given |
| 400 |
/// haystack. |
| 401 |
/// |
| 402 |
/// # Complexity |
| 403 |
/// |
| 404 |
/// This routine is guaranteed to have worst case linear time complexity |
| 405 |
/// with respect to both the needle and the haystack. That is, this runs |
| 406 |
/// in `O(needle.len() + haystack.len())` time. |
| 407 |
/// |
| 408 |
/// This routine is also guaranteed to have worst case constant space |
| 409 |
/// complexity. |
| 410 |
/// |
| 411 |
/// # Examples |
| 412 |
/// |
| 413 |
/// Basic usage: |
| 414 |
/// |
| 415 |
/// ``` |
| 416 |
/// use memchr::memmem::Finder; |
| 417 |
/// |
| 418 |
/// let haystack = b"foo bar baz"; |
| 419 |
/// assert_eq!(Some(0), Finder::new("foo").find(haystack)); |
| 420 |
/// assert_eq!(Some(4), Finder::new("bar").find(haystack)); |
| 421 |
/// assert_eq!(None, Finder::new("quux").find(haystack)); |
| 422 |
/// ``` |
| 423 |
#[inline] |
| 424 |
pub fn find(&self, haystack: &[u8]) -> Option<usize> { |
| 425 |
let mut prestate = PrefilterState::new(); |
| 426 |
let needle = self.needle.as_slice(); |
| 427 |
self.searcher.find(&mut prestate, haystack, needle) |
| 428 |
} |
| 429 |
|
| 430 |
/// Returns an iterator over all occurrences of a substring in a haystack. |
| 431 |
/// |
| 432 |
/// # Complexity |
| 433 |
/// |
| 434 |
/// This routine is guaranteed to have worst case linear time complexity |
| 435 |
/// with respect to both the needle and the haystack. That is, this runs |
| 436 |
/// in `O(needle.len() + haystack.len())` time. |
| 437 |
/// |
| 438 |
/// This routine is also guaranteed to have worst case constant space |
| 439 |
/// complexity. |
| 440 |
/// |
| 441 |
/// # Examples |
| 442 |
/// |
| 443 |
/// Basic usage: |
| 444 |
/// |
| 445 |
/// ``` |
| 446 |
/// use memchr::memmem::Finder; |
| 447 |
/// |
| 448 |
/// let haystack = b"foo bar foo baz foo"; |
| 449 |
/// let finder = Finder::new(b"foo"); |
| 450 |
/// let mut it = finder.find_iter(haystack); |
| 451 |
/// assert_eq!(Some(0), it.next()); |
| 452 |
/// assert_eq!(Some(8), it.next()); |
| 453 |
/// assert_eq!(Some(16), it.next()); |
| 454 |
/// assert_eq!(None, it.next()); |
| 455 |
/// ``` |
| 456 |
#[inline] |
| 457 |
pub fn find_iter<'a, 'h>( |
| 458 |
&'a self, |
| 459 |
haystack: &'h [u8], |
| 460 |
) -> FindIter<'h, 'a> { |
| 461 |
FindIter::new(haystack, self.as_ref()) |
| 462 |
} |
| 463 |
|
| 464 |
/// Convert this finder into its owned variant, such that it no longer |
| 465 |
/// borrows the needle. |
| 466 |
/// |
| 467 |
/// If this is already an owned finder, then this is a no-op. Otherwise, |
| 468 |
/// this copies the needle. |
| 469 |
/// |
| 470 |
/// This is only available when the `alloc` feature is enabled. |
| 471 |
#[cfg(feature = "alloc")] |
| 472 |
#[inline] |
| 473 |
pub fn into_owned(self) -> Finder<'static> { |
| 474 |
Finder { |
| 475 |
needle: self.needle.into_owned(), |
| 476 |
searcher: self.searcher.clone(), |
| 477 |
} |
| 478 |
} |
| 479 |
|
| 480 |
/// Convert this finder into its borrowed variant. |
| 481 |
/// |
| 482 |
/// This is primarily useful if your finder is owned and you'd like to |
| 483 |
/// store its borrowed variant in some intermediate data structure. |
| 484 |
/// |
| 485 |
/// Note that the lifetime parameter of the returned finder is tied to the |
| 486 |
/// lifetime of `self`, and may be shorter than the `'n` lifetime of the |
| 487 |
/// needle itself. Namely, a finder's needle can be either borrowed or |
| 488 |
/// owned, so the lifetime of the needle returned must necessarily be the |
| 489 |
/// shorter of the two. |
| 490 |
#[inline] |
| 491 |
pub fn as_ref(&self) -> Finder<'_> { |
| 492 |
Finder { |
| 493 |
needle: CowBytes::new(self.needle()), |
| 494 |
searcher: self.searcher.clone(), |
| 495 |
} |
| 496 |
} |
| 497 |
|
| 498 |
/// Returns the needle that this finder searches for. |
| 499 |
/// |
| 500 |
/// Note that the lifetime of the needle returned is tied to the lifetime |
| 501 |
/// of the finder, and may be shorter than the `'n` lifetime. Namely, a |
| 502 |
/// finder's needle can be either borrowed or owned, so the lifetime of the |
| 503 |
/// needle returned must necessarily be the shorter of the two. |
| 504 |
#[inline] |
| 505 |
pub fn needle(&self) -> &[u8] { |
| 506 |
self.needle.as_slice() |
| 507 |
} |
| 508 |
} |
| 509 |
|
| 510 |
/// A single substring reverse searcher fixed to a particular needle. |
| 511 |
/// |
| 512 |
/// The purpose of this type is to permit callers to construct a substring |
| 513 |
/// searcher that can be used to search haystacks without the overhead of |
| 514 |
/// constructing the searcher in the first place. This is a somewhat niche |
| 515 |
/// concern when it's necessary to re-use the same needle to search multiple |
| 516 |
/// different haystacks with as little overhead as possible. In general, |
| 517 |
/// using [`rfind`] is good enough, but `FinderRev` is useful when you can |
| 518 |
/// meaningfully observe searcher construction time in a profile. |
| 519 |
/// |
| 520 |
/// When the `std` feature is enabled, then this type has an `into_owned` |
| 521 |
/// version which permits building a `FinderRev` that is not connected to |
| 522 |
/// the lifetime of its needle. |
| 523 |
#[derive(Clone, Debug)] |
| 524 |
pub struct FinderRev<'n> { |
| 525 |
needle: CowBytes<'n>, |
| 526 |
searcher: SearcherRev, |
| 527 |
} |
| 528 |
|
| 529 |
impl<'n> FinderRev<'n> { |
| 530 |
/// Create a new reverse finder for the given needle. |
| 531 |
#[inline] |
| 532 |
pub fn new<B: ?Sized + AsRef<[u8]>>(needle: &'n B) -> FinderRev<'n> { |
| 533 |
FinderBuilder::new().build_reverse(needle) |
| 534 |
} |
| 535 |
|
| 536 |
/// Returns the index of the last occurrence of this needle in the given |
| 537 |
/// haystack. |
| 538 |
/// |
| 539 |
/// The haystack may be any type that can be cheaply converted into a |
| 540 |
/// `&[u8]`. This includes, but is not limited to, `&str` and `&[u8]`. |
| 541 |
/// |
| 542 |
/// # Complexity |
| 543 |
/// |
| 544 |
/// This routine is guaranteed to have worst case linear time complexity |
| 545 |
/// with respect to both the needle and the haystack. That is, this runs |
| 546 |
/// in `O(needle.len() + haystack.len())` time. |
| 547 |
/// |
| 548 |
/// This routine is also guaranteed to have worst case constant space |
| 549 |
/// complexity. |
| 550 |
/// |
| 551 |
/// # Examples |
| 552 |
/// |
| 553 |
/// Basic usage: |
| 554 |
/// |
| 555 |
/// ``` |
| 556 |
/// use memchr::memmem::FinderRev; |
| 557 |
/// |
| 558 |
/// let haystack = b"foo bar baz"; |
| 559 |
/// assert_eq!(Some(0), FinderRev::new("foo").rfind(haystack)); |
| 560 |
/// assert_eq!(Some(4), FinderRev::new("bar").rfind(haystack)); |
| 561 |
/// assert_eq!(None, FinderRev::new("quux").rfind(haystack)); |
| 562 |
/// ``` |
| 563 |
pub fn rfind<B: AsRef<[u8]>>(&self, haystack: B) -> Option<usize> { |
| 564 |
self.searcher.rfind(haystack.as_ref(), self.needle.as_slice()) |
| 565 |
} |
| 566 |
|
| 567 |
/// Returns a reverse iterator over all occurrences of a substring in a |
| 568 |
/// haystack. |
| 569 |
/// |
| 570 |
/// # Complexity |
| 571 |
/// |
| 572 |
/// This routine is guaranteed to have worst case linear time complexity |
| 573 |
/// with respect to both the needle and the haystack. That is, this runs |
| 574 |
/// in `O(needle.len() + haystack.len())` time. |
| 575 |
/// |
| 576 |
/// This routine is also guaranteed to have worst case constant space |
| 577 |
/// complexity. |
| 578 |
/// |
| 579 |
/// # Examples |
| 580 |
/// |
| 581 |
/// Basic usage: |
| 582 |
/// |
| 583 |
/// ``` |
| 584 |
/// use memchr::memmem::FinderRev; |
| 585 |
/// |
| 586 |
/// let haystack = b"foo bar foo baz foo"; |
| 587 |
/// let finder = FinderRev::new(b"foo"); |
| 588 |
/// let mut it = finder.rfind_iter(haystack); |
| 589 |
/// assert_eq!(Some(16), it.next()); |
| 590 |
/// assert_eq!(Some(8), it.next()); |
| 591 |
/// assert_eq!(Some(0), it.next()); |
| 592 |
/// assert_eq!(None, it.next()); |
| 593 |
/// ``` |
| 594 |
#[inline] |
| 595 |
pub fn rfind_iter<'a, 'h>( |
| 596 |
&'a self, |
| 597 |
haystack: &'h [u8], |
| 598 |
) -> FindRevIter<'h, 'a> { |
| 599 |
FindRevIter::new(haystack, self.as_ref()) |
| 600 |
} |
| 601 |
|
| 602 |
/// Convert this finder into its owned variant, such that it no longer |
| 603 |
/// borrows the needle. |
| 604 |
/// |
| 605 |
/// If this is already an owned finder, then this is a no-op. Otherwise, |
| 606 |
/// this copies the needle. |
| 607 |
/// |
| 608 |
/// This is only available when the `std` feature is enabled. |
| 609 |
#[cfg(feature = "alloc")] |
| 610 |
#[inline] |
| 611 |
pub fn into_owned(self) -> FinderRev<'static> { |
| 612 |
FinderRev { |
| 613 |
needle: self.needle.into_owned(), |
| 614 |
searcher: self.searcher.clone(), |
| 615 |
} |
| 616 |
} |
| 617 |
|
| 618 |
/// Convert this finder into its borrowed variant. |
| 619 |
/// |
| 620 |
/// This is primarily useful if your finder is owned and you'd like to |
| 621 |
/// store its borrowed variant in some intermediate data structure. |
| 622 |
/// |
| 623 |
/// Note that the lifetime parameter of the returned finder is tied to the |
| 624 |
/// lifetime of `self`, and may be shorter than the `'n` lifetime of the |
| 625 |
/// needle itself. Namely, a finder's needle can be either borrowed or |
| 626 |
/// owned, so the lifetime of the needle returned must necessarily be the |
| 627 |
/// shorter of the two. |
| 628 |
#[inline] |
| 629 |
pub fn as_ref(&self) -> FinderRev<'_> { |
| 630 |
FinderRev { |
| 631 |
needle: CowBytes::new(self.needle()), |
| 632 |
searcher: self.searcher.clone(), |
| 633 |
} |
| 634 |
} |
| 635 |
|
| 636 |
/// Returns the needle that this finder searches for. |
| 637 |
/// |
| 638 |
/// Note that the lifetime of the needle returned is tied to the lifetime |
| 639 |
/// of the finder, and may be shorter than the `'n` lifetime. Namely, a |
| 640 |
/// finder's needle can be either borrowed or owned, so the lifetime of the |
| 641 |
/// needle returned must necessarily be the shorter of the two. |
| 642 |
#[inline] |
| 643 |
pub fn needle(&self) -> &[u8] { |
| 644 |
self.needle.as_slice() |
| 645 |
} |
| 646 |
} |
| 647 |
|
| 648 |
/// A builder for constructing non-default forward or reverse memmem finders. |
| 649 |
/// |
| 650 |
/// A builder is primarily useful for configuring a substring searcher. |
| 651 |
/// Currently, the only configuration exposed is the ability to disable |
| 652 |
/// heuristic prefilters used to speed up certain searches. |
| 653 |
#[derive(Clone, Debug, Default)] |
| 654 |
pub struct FinderBuilder { |
| 655 |
prefilter: Prefilter, |
| 656 |
} |
| 657 |
|
| 658 |
impl FinderBuilder { |
| 659 |
/// Create a new finder builder with default settings. |
| 660 |
pub fn new() -> FinderBuilder { |
| 661 |
FinderBuilder::default() |
| 662 |
} |
| 663 |
|
| 664 |
/// Build a forward finder using the given needle from the current |
| 665 |
/// settings. |
| 666 |
pub fn build_forward<'n, B: ?Sized + AsRef<[u8]>>( |
| 667 |
&self, |
| 668 |
needle: &'n B, |
| 669 |
) -> Finder<'n> { |
| 670 |
self.build_forward_with_ranker(DefaultFrequencyRank, needle) |
| 671 |
} |
| 672 |
|
| 673 |
/// Build a forward finder using the given needle and a custom heuristic for |
| 674 |
/// determining the frequency of a given byte in the dataset. |
| 675 |
/// See [`HeuristicFrequencyRank`] for more details. |
| 676 |
pub fn build_forward_with_ranker< |
| 677 |
'n, |
| 678 |
R: HeuristicFrequencyRank, |
| 679 |
B: ?Sized + AsRef<[u8]>, |
| 680 |
>( |
| 681 |
&self, |
| 682 |
ranker: R, |
| 683 |
needle: &'n B, |
| 684 |
) -> Finder<'n> { |
| 685 |
let needle = needle.as_ref(); |
| 686 |
Finder { |
| 687 |
needle: CowBytes::new(needle), |
| 688 |
searcher: Searcher::new(self.prefilter, ranker, needle), |
| 689 |
} |
| 690 |
} |
| 691 |
|
| 692 |
/// Build a reverse finder using the given needle from the current |
| 693 |
/// settings. |
| 694 |
pub fn build_reverse<'n, B: ?Sized + AsRef<[u8]>>( |
| 695 |
&self, |
| 696 |
needle: &'n B, |
| 697 |
) -> FinderRev<'n> { |
| 698 |
let needle = needle.as_ref(); |
| 699 |
FinderRev { |
| 700 |
needle: CowBytes::new(needle), |
| 701 |
searcher: SearcherRev::new(needle), |
| 702 |
} |
| 703 |
} |
| 704 |
|
| 705 |
/// Configure the prefilter setting for the finder. |
| 706 |
/// |
| 707 |
/// See the documentation for [`Prefilter`] for more discussion on why |
| 708 |
/// you might want to configure this. |
| 709 |
pub fn prefilter(&mut self, prefilter: Prefilter) -> &mut FinderBuilder { |
| 710 |
self.prefilter = prefilter; |
| 711 |
self |
| 712 |
} |
| 713 |
} |
| 714 |
|
| 715 |
#[cfg(test)] |
| 716 |
mod tests { |
| 717 |
use super::*; |
| 718 |
|
| 719 |
define_substring_forward_quickcheck!(|h, n| Some(Finder::new(n).find(h))); |
| 720 |
define_substring_reverse_quickcheck!(|h, n| Some( |
| 721 |
FinderRev::new(n).rfind(h) |
| 722 |
)); |
| 723 |
|
| 724 |
#[test] |
| 725 |
fn forward() { |
| 726 |
crate::tests::substring::Runner::new() |
| 727 |
.fwd(|h, n| Some(Finder::new(n).find(h))) |
| 728 |
.run(); |
| 729 |
} |
| 730 |
|
| 731 |
#[test] |
| 732 |
fn reverse() { |
| 733 |
crate::tests::substring::Runner::new() |
| 734 |
.rev(|h, n| Some(FinderRev::new(n).rfind(h))) |
| 735 |
.run(); |
| 736 |
} |
| 737 |
} |