1//! String manipulation.
2//!
3//! For more details, see the [`std::str`] module.
4//!
5//! [`std::str`]: ../../std/str/index.html
6
7#![stable(feature = "rust1", since = "1.0.0")]
8
9mod converts;
10mod count;
11mod error;
12mod iter;
13mod traits;
14mod validations;
15
16use self::pattern::Pattern;
17use self::pattern::{DoubleEndedSearcher, ReverseSearcher, Searcher};
18
19use crate::ascii;
20use crate::char::{self, EscapeDebugExtArgs};
21use crate::mem;
22use crate::slice::{self, SliceIndex};
23
24pub mod pattern;
25
26mod lossy;
27#[unstable(feature = "utf8_chunks", issue = "99543")]
28pub use lossy::{Utf8Chunk, Utf8Chunks};
29
30#[stable(feature = "rust1", since = "1.0.0")]
31pub use converts::{from_utf8, from_utf8_unchecked};
32
33#[stable(feature = "str_mut_extras", since = "1.20.0")]
34pub use converts::{from_utf8_mut, from_utf8_unchecked_mut};
35
36#[unstable(feature = "str_from_raw_parts", issue = "119206")]
37pub use converts::{from_raw_parts, from_raw_parts_mut};
38
39#[stable(feature = "rust1", since = "1.0.0")]
40pub use error::{ParseBoolError, Utf8Error};
41
42#[stable(feature = "rust1", since = "1.0.0")]
43pub use traits::FromStr;
44
45#[stable(feature = "rust1", since = "1.0.0")]
46pub use iter::{Bytes, CharIndices, Chars, Lines, SplitWhitespace};
47
48#[stable(feature = "rust1", since = "1.0.0")]
49#[allow(deprecated)]
50pub use iter::LinesAny;
51
52#[stable(feature = "rust1", since = "1.0.0")]
53pub use iter::{RSplit, RSplitTerminator, Split, SplitTerminator};
54
55#[stable(feature = "rust1", since = "1.0.0")]
56pub use iter::{RSplitN, SplitN};
57
58#[stable(feature = "str_matches", since = "1.2.0")]
59pub use iter::{Matches, RMatches};
60
61#[stable(feature = "str_match_indices", since = "1.5.0")]
62pub use iter::{MatchIndices, RMatchIndices};
63
64#[stable(feature = "encode_utf16", since = "1.8.0")]
65pub use iter::EncodeUtf16;
66
67#[stable(feature = "str_escape", since = "1.34.0")]
68pub use iter::{EscapeDebug, EscapeDefault, EscapeUnicode};
69
70#[stable(feature = "split_ascii_whitespace", since = "1.34.0")]
71pub use iter::SplitAsciiWhitespace;
72
73#[stable(feature = "split_inclusive", since = "1.51.0")]
74pub use iter::SplitInclusive;
75
76#[unstable(feature = "str_internals", issue = "none")]
77pub use validations::{next_code_point, utf8_char_width};
78
79use iter::MatchIndicesInternal;
80use iter::SplitInternal;
81use iter::{MatchesInternal, SplitNInternal};
82
83#[inline(never)]
84#[cold]
85#[track_caller]
86#[rustc_allow_const_fn_unstable(const_eval_select)]
87#[cfg(not(feature = "panic_immediate_abort"))]
88const fn slice_error_fail(s: &str, begin: usize, end: usize) -> ! {
89 // SAFETY: panics for both branches
90 unsafe {
91 crate::intrinsics::const_eval_select(
92 (s, begin, end),
93 called_in_const:slice_error_fail_ct,
94 called_at_rt:slice_error_fail_rt,
95 )
96 }
97}
98
99#[cfg(feature = "panic_immediate_abort")]
100const fn slice_error_fail(s: &str, begin: usize, end: usize) -> ! {
101 slice_error_fail_ct(s, begin, end)
102}
103
104#[track_caller]
105const fn slice_error_fail_ct(_: &str, _: usize, _: usize) -> ! {
106 panic!("failed to slice string");
107}
108
109#[track_caller]
110fn slice_error_fail_rt(s: &str, begin: usize, end: usize) -> ! {
111 const MAX_DISPLAY_LENGTH: usize = 256;
112 let trunc_len = s.floor_char_boundary(MAX_DISPLAY_LENGTH);
113 let s_trunc = &s[..trunc_len];
114 let ellipsis = if trunc_len < s.len() { "[...]" } else { "" };
115
116 // 1. out of bounds
117 if begin > s.len() || end > s.len() {
118 let oob_index = if begin > s.len() { begin } else { end };
119 panic!("byte index {oob_index} is out of bounds of `{s_trunc}`{ellipsis}");
120 }
121
122 // 2. begin <= end
123 assert!(
124 begin <= end,
125 "begin <= end ({} <= {}) when slicing `{}`{}",
126 begin,
127 end,
128 s_trunc,
129 ellipsis
130 );
131
132 // 3. character boundary
133 let index = if !s.is_char_boundary(begin) { begin } else { end };
134 // find the character
135 let char_start = s.floor_char_boundary(index);
136 // `char_start` must be less than len and a char boundary
137 let ch = s[char_start..].chars().next().unwrap();
138 let char_range = char_start..char_start + ch.len_utf8();
139 panic!(
140 "byte index {} is not a char boundary; it is inside {:?} (bytes {:?}) of `{}`{}",
141 index, ch, char_range, s_trunc, ellipsis
142 );
143}
144
145#[cfg(not(test))]
146impl str {
147 /// Returns the length of `self`.
148 ///
149 /// This length is in bytes, not [`char`]s or graphemes. In other words,
150 /// it might not be what a human considers the length of the string.
151 ///
152 /// [`char`]: prim@char
153 ///
154 /// # Examples
155 ///
156 /// ```
157 /// let len = "foo".len();
158 /// assert_eq!(3, len);
159 ///
160 /// assert_eq!("ƒoo".len(), 4); // fancy f!
161 /// assert_eq!("ƒoo".chars().count(), 3);
162 /// ```
163 #[stable(feature = "rust1", since = "1.0.0")]
164 #[rustc_const_stable(feature = "const_str_len", since = "1.39.0")]
165 #[must_use]
166 #[inline]
167 pub const fn len(&self) -> usize {
168 self.as_bytes().len()
169 }
170
171 /// Returns `true` if `self` has a length of zero bytes.
172 ///
173 /// # Examples
174 ///
175 /// ```
176 /// let s = "";
177 /// assert!(s.is_empty());
178 ///
179 /// let s = "not empty";
180 /// assert!(!s.is_empty());
181 /// ```
182 #[stable(feature = "rust1", since = "1.0.0")]
183 #[rustc_const_stable(feature = "const_str_is_empty", since = "1.39.0")]
184 #[must_use]
185 #[inline]
186 pub const fn is_empty(&self) -> bool {
187 self.len() == 0
188 }
189
190 /// Checks that `index`-th byte is the first byte in a UTF-8 code point
191 /// sequence or the end of the string.
192 ///
193 /// The start and end of the string (when `index == self.len()`) are
194 /// considered to be boundaries.
195 ///
196 /// Returns `false` if `index` is greater than `self.len()`.
197 ///
198 /// # Examples
199 ///
200 /// ```
201 /// let s = "Löwe 老虎 Léopard";
202 /// assert!(s.is_char_boundary(0));
203 /// // start of `老`
204 /// assert!(s.is_char_boundary(6));
205 /// assert!(s.is_char_boundary(s.len()));
206 ///
207 /// // second byte of `ö`
208 /// assert!(!s.is_char_boundary(2));
209 ///
210 /// // third byte of `老`
211 /// assert!(!s.is_char_boundary(8));
212 /// ```
213 #[must_use]
214 #[stable(feature = "is_char_boundary", since = "1.9.0")]
215 #[inline]
216 pub fn is_char_boundary(&self, index: usize) -> bool {
217 // 0 is always ok.
218 // Test for 0 explicitly so that it can optimize out the check
219 // easily and skip reading string data for that case.
220 // Note that optimizing `self.get(..index)` relies on this.
221 if index == 0 {
222 return true;
223 }
224
225 match self.as_bytes().get(index) {
226 // For `None` we have two options:
227 //
228 // - index == self.len()
229 // Empty strings are valid, so return true
230 // - index > self.len()
231 // In this case return false
232 //
233 // The check is placed exactly here, because it improves generated
234 // code on higher opt-levels. See PR #84751 for more details.
235 None => index == self.len(),
236
237 Some(&b) => b.is_utf8_char_boundary(),
238 }
239 }
240
241 /// Finds the closest `x` not exceeding `index` where `is_char_boundary(x)` is `true`.
242 ///
243 /// This method can help you truncate a string so that it's still valid UTF-8, but doesn't
244 /// exceed a given number of bytes. Note that this is done purely at the character level
245 /// and can still visually split graphemes, even though the underlying characters aren't
246 /// split. For example, the emoji 🧑‍🔬 (scientist) could be split so that the string only
247 /// includes 🧑 (person) instead.
248 ///
249 /// # Examples
250 ///
251 /// ```
252 /// #![feature(round_char_boundary)]
253 /// let s = "❤️🧡💛💚💙💜";
254 /// assert_eq!(s.len(), 26);
255 /// assert!(!s.is_char_boundary(13));
256 ///
257 /// let closest = s.floor_char_boundary(13);
258 /// assert_eq!(closest, 10);
259 /// assert_eq!(&s[..closest], "❤️🧡");
260 /// ```
261 #[unstable(feature = "round_char_boundary", issue = "93743")]
262 #[inline]
263 pub fn floor_char_boundary(&self, index: usize) -> usize {
264 if index >= self.len() {
265 self.len()
266 } else {
267 let lower_bound = index.saturating_sub(3);
268 let new_index = self.as_bytes()[lower_bound..=index]
269 .iter()
270 .rposition(|b| b.is_utf8_char_boundary());
271
272 // SAFETY: we know that the character boundary will be within four bytes
273 unsafe { lower_bound + new_index.unwrap_unchecked() }
274 }
275 }
276
277 /// Finds the closest `x` not below `index` where `is_char_boundary(x)` is `true`.
278 ///
279 /// If `index` is greater than the length of the string, this returns the length of the string.
280 ///
281 /// This method is the natural complement to [`floor_char_boundary`]. See that method
282 /// for more details.
283 ///
284 /// [`floor_char_boundary`]: str::floor_char_boundary
285 ///
286 ///
287 /// # Examples
288 ///
289 /// ```
290 /// #![feature(round_char_boundary)]
291 /// let s = "❤️🧡💛💚💙💜";
292 /// assert_eq!(s.len(), 26);
293 /// assert!(!s.is_char_boundary(13));
294 ///
295 /// let closest = s.ceil_char_boundary(13);
296 /// assert_eq!(closest, 14);
297 /// assert_eq!(&s[..closest], "❤️🧡💛");
298 /// ```
299 #[unstable(feature = "round_char_boundary", issue = "93743")]
300 #[inline]
301 pub fn ceil_char_boundary(&self, index: usize) -> usize {
302 if index > self.len() {
303 self.len()
304 } else {
305 let upper_bound = Ord::min(index + 4, self.len());
306 self.as_bytes()[index..upper_bound]
307 .iter()
308 .position(|b| b.is_utf8_char_boundary())
309 .map_or(upper_bound, |pos| pos + index)
310 }
311 }
312
313 /// Converts a string slice to a byte slice. To convert the byte slice back
314 /// into a string slice, use the [`from_utf8`] function.
315 ///
316 /// # Examples
317 ///
318 /// ```
319 /// let bytes = "bors".as_bytes();
320 /// assert_eq!(b"bors", bytes);
321 /// ```
322 #[stable(feature = "rust1", since = "1.0.0")]
323 #[rustc_const_stable(feature = "str_as_bytes", since = "1.39.0")]
324 #[must_use]
325 #[inline(always)]
326 #[allow(unused_attributes)]
327 pub const fn as_bytes(&self) -> &[u8] {
328 // SAFETY: const sound because we transmute two types with the same layout
329 unsafe { mem::transmute(self) }
330 }
331
332 /// Converts a mutable string slice to a mutable byte slice.
333 ///
334 /// # Safety
335 ///
336 /// The caller must ensure that the content of the slice is valid UTF-8
337 /// before the borrow ends and the underlying `str` is used.
338 ///
339 /// Use of a `str` whose contents are not valid UTF-8 is undefined behavior.
340 ///
341 /// # Examples
342 ///
343 /// Basic usage:
344 ///
345 /// ```
346 /// let mut s = String::from("Hello");
347 /// let bytes = unsafe { s.as_bytes_mut() };
348 ///
349 /// assert_eq!(b"Hello", bytes);
350 /// ```
351 ///
352 /// Mutability:
353 ///
354 /// ```
355 /// let mut s = String::from("🗻∈🌏");
356 ///
357 /// unsafe {
358 /// let bytes = s.as_bytes_mut();
359 ///
360 /// bytes[0] = 0xF0;
361 /// bytes[1] = 0x9F;
362 /// bytes[2] = 0x8D;
363 /// bytes[3] = 0x94;
364 /// }
365 ///
366 /// assert_eq!("🍔∈🌏", s);
367 /// ```
368 #[stable(feature = "str_mut_extras", since = "1.20.0")]
369 #[must_use]
370 #[inline(always)]
371 pub unsafe fn as_bytes_mut(&mut self) -> &mut [u8] {
372 // SAFETY: the cast from `&str` to `&[u8]` is safe since `str`
373 // has the same layout as `&[u8]` (only std can make this guarantee).
374 // The pointer dereference is safe since it comes from a mutable reference which
375 // is guaranteed to be valid for writes.
376 unsafe { &mut *(self as *mut str as *mut [u8]) }
377 }
378
379 /// Converts a string slice to a raw pointer.
380 ///
381 /// As string slices are a slice of bytes, the raw pointer points to a
382 /// [`u8`]. This pointer will be pointing to the first byte of the string
383 /// slice.
384 ///
385 /// The caller must ensure that the returned pointer is never written to.
386 /// If you need to mutate the contents of the string slice, use [`as_mut_ptr`].
387 ///
388 /// [`as_mut_ptr`]: str::as_mut_ptr
389 ///
390 /// # Examples
391 ///
392 /// ```
393 /// let s = "Hello";
394 /// let ptr = s.as_ptr();
395 /// ```
396 #[stable(feature = "rust1", since = "1.0.0")]
397 #[rustc_const_stable(feature = "rustc_str_as_ptr", since = "1.32.0")]
398 #[rustc_never_returns_null_ptr]
399 #[must_use]
400 #[inline(always)]
401 pub const fn as_ptr(&self) -> *const u8 {
402 self as *const str as *const u8
403 }
404
405 /// Converts a mutable string slice to a raw pointer.
406 ///
407 /// As string slices are a slice of bytes, the raw pointer points to a
408 /// [`u8`]. This pointer will be pointing to the first byte of the string
409 /// slice.
410 ///
411 /// It is your responsibility to make sure that the string slice only gets
412 /// modified in a way that it remains valid UTF-8.
413 #[stable(feature = "str_as_mut_ptr", since = "1.36.0")]
414 #[rustc_never_returns_null_ptr]
415 #[must_use]
416 #[inline(always)]
417 pub fn as_mut_ptr(&mut self) -> *mut u8 {
418 self as *mut str as *mut u8
419 }
420
421 /// Returns a subslice of `str`.
422 ///
423 /// This is the non-panicking alternative to indexing the `str`. Returns
424 /// [`None`] whenever equivalent indexing operation would panic.
425 ///
426 /// # Examples
427 ///
428 /// ```
429 /// let v = String::from("🗻∈🌏");
430 ///
431 /// assert_eq!(Some("🗻"), v.get(0..4));
432 ///
433 /// // indices not on UTF-8 sequence boundaries
434 /// assert!(v.get(1..).is_none());
435 /// assert!(v.get(..8).is_none());
436 ///
437 /// // out of bounds
438 /// assert!(v.get(..42).is_none());
439 /// ```
440 #[stable(feature = "str_checked_slicing", since = "1.20.0")]
441 #[inline]
442 pub fn get<I: SliceIndex<str>>(&self, i: I) -> Option<&I::Output> {
443 i.get(self)
444 }
445
446 /// Returns a mutable subslice of `str`.
447 ///
448 /// This is the non-panicking alternative to indexing the `str`. Returns
449 /// [`None`] whenever equivalent indexing operation would panic.
450 ///
451 /// # Examples
452 ///
453 /// ```
454 /// let mut v = String::from("hello");
455 /// // correct length
456 /// assert!(v.get_mut(0..5).is_some());
457 /// // out of bounds
458 /// assert!(v.get_mut(..42).is_none());
459 /// assert_eq!(Some("he"), v.get_mut(0..2).map(|v| &*v));
460 ///
461 /// assert_eq!("hello", v);
462 /// {
463 /// let s = v.get_mut(0..2);
464 /// let s = s.map(|s| {
465 /// s.make_ascii_uppercase();
466 /// &*s
467 /// });
468 /// assert_eq!(Some("HE"), s);
469 /// }
470 /// assert_eq!("HEllo", v);
471 /// ```
472 #[stable(feature = "str_checked_slicing", since = "1.20.0")]
473 #[inline]
474 pub fn get_mut<I: SliceIndex<str>>(&mut self, i: I) -> Option<&mut I::Output> {
475 i.get_mut(self)
476 }
477
478 /// Returns an unchecked subslice of `str`.
479 ///
480 /// This is the unchecked alternative to indexing the `str`.
481 ///
482 /// # Safety
483 ///
484 /// Callers of this function are responsible that these preconditions are
485 /// satisfied:
486 ///
487 /// * The starting index must not exceed the ending index;
488 /// * Indexes must be within bounds of the original slice;
489 /// * Indexes must lie on UTF-8 sequence boundaries.
490 ///
491 /// Failing that, the returned string slice may reference invalid memory or
492 /// violate the invariants communicated by the `str` type.
493 ///
494 /// # Examples
495 ///
496 /// ```
497 /// let v = "🗻∈🌏";
498 /// unsafe {
499 /// assert_eq!("🗻", v.get_unchecked(0..4));
500 /// assert_eq!("∈", v.get_unchecked(4..7));
501 /// assert_eq!("🌏", v.get_unchecked(7..11));
502 /// }
503 /// ```
504 #[stable(feature = "str_checked_slicing", since = "1.20.0")]
505 #[inline]
506 pub unsafe fn get_unchecked<I: SliceIndex<str>>(&self, i: I) -> &I::Output {
507 // SAFETY: the caller must uphold the safety contract for `get_unchecked`;
508 // the slice is dereferenceable because `self` is a safe reference.
509 // The returned pointer is safe because impls of `SliceIndex` have to guarantee that it is.
510 unsafe { &*i.get_unchecked(self) }
511 }
512
513 /// Returns a mutable, unchecked subslice of `str`.
514 ///
515 /// This is the unchecked alternative to indexing the `str`.
516 ///
517 /// # Safety
518 ///
519 /// Callers of this function are responsible that these preconditions are
520 /// satisfied:
521 ///
522 /// * The starting index must not exceed the ending index;
523 /// * Indexes must be within bounds of the original slice;
524 /// * Indexes must lie on UTF-8 sequence boundaries.
525 ///
526 /// Failing that, the returned string slice may reference invalid memory or
527 /// violate the invariants communicated by the `str` type.
528 ///
529 /// # Examples
530 ///
531 /// ```
532 /// let mut v = String::from("🗻∈🌏");
533 /// unsafe {
534 /// assert_eq!("🗻", v.get_unchecked_mut(0..4));
535 /// assert_eq!("∈", v.get_unchecked_mut(4..7));
536 /// assert_eq!("🌏", v.get_unchecked_mut(7..11));
537 /// }
538 /// ```
539 #[stable(feature = "str_checked_slicing", since = "1.20.0")]
540 #[inline]
541 pub unsafe fn get_unchecked_mut<I: SliceIndex<str>>(&mut self, i: I) -> &mut I::Output {
542 // SAFETY: the caller must uphold the safety contract for `get_unchecked_mut`;
543 // the slice is dereferenceable because `self` is a safe reference.
544 // The returned pointer is safe because impls of `SliceIndex` have to guarantee that it is.
545 unsafe { &mut *i.get_unchecked_mut(self) }
546 }
547
548 /// Creates a string slice from another string slice, bypassing safety
549 /// checks.
550 ///
551 /// This is generally not recommended, use with caution! For a safe
552 /// alternative see [`str`] and [`Index`].
553 ///
554 /// [`Index`]: crate::ops::Index
555 ///
556 /// This new slice goes from `begin` to `end`, including `begin` but
557 /// excluding `end`.
558 ///
559 /// To get a mutable string slice instead, see the
560 /// [`slice_mut_unchecked`] method.
561 ///
562 /// [`slice_mut_unchecked`]: str::slice_mut_unchecked
563 ///
564 /// # Safety
565 ///
566 /// Callers of this function are responsible that three preconditions are
567 /// satisfied:
568 ///
569 /// * `begin` must not exceed `end`.
570 /// * `begin` and `end` must be byte positions within the string slice.
571 /// * `begin` and `end` must lie on UTF-8 sequence boundaries.
572 ///
573 /// # Examples
574 ///
575 /// ```
576 /// let s = "Löwe 老虎 Léopard";
577 ///
578 /// unsafe {
579 /// assert_eq!("Löwe 老虎 Léopard", s.slice_unchecked(0, 21));
580 /// }
581 ///
582 /// let s = "Hello, world!";
583 ///
584 /// unsafe {
585 /// assert_eq!("world", s.slice_unchecked(7, 12));
586 /// }
587 /// ```
588 #[stable(feature = "rust1", since = "1.0.0")]
589 #[deprecated(since = "1.29.0", note = "use `get_unchecked(begin..end)` instead")]
590 #[must_use]
591 #[inline]
592 pub unsafe fn slice_unchecked(&self, begin: usize, end: usize) -> &str {
593 // SAFETY: the caller must uphold the safety contract for `get_unchecked`;
594 // the slice is dereferenceable because `self` is a safe reference.
595 // The returned pointer is safe because impls of `SliceIndex` have to guarantee that it is.
596 unsafe { &*(begin..end).get_unchecked(self) }
597 }
598
599 /// Creates a string slice from another string slice, bypassing safety
600 /// checks.
601 /// This is generally not recommended, use with caution! For a safe
602 /// alternative see [`str`] and [`IndexMut`].
603 ///
604 /// [`IndexMut`]: crate::ops::IndexMut
605 ///
606 /// This new slice goes from `begin` to `end`, including `begin` but
607 /// excluding `end`.
608 ///
609 /// To get an immutable string slice instead, see the
610 /// [`slice_unchecked`] method.
611 ///
612 /// [`slice_unchecked`]: str::slice_unchecked
613 ///
614 /// # Safety
615 ///
616 /// Callers of this function are responsible that three preconditions are
617 /// satisfied:
618 ///
619 /// * `begin` must not exceed `end`.
620 /// * `begin` and `end` must be byte positions within the string slice.
621 /// * `begin` and `end` must lie on UTF-8 sequence boundaries.
622 #[stable(feature = "str_slice_mut", since = "1.5.0")]
623 #[deprecated(since = "1.29.0", note = "use `get_unchecked_mut(begin..end)` instead")]
624 #[inline]
625 pub unsafe fn slice_mut_unchecked(&mut self, begin: usize, end: usize) -> &mut str {
626 // SAFETY: the caller must uphold the safety contract for `get_unchecked_mut`;
627 // the slice is dereferenceable because `self` is a safe reference.
628 // The returned pointer is safe because impls of `SliceIndex` have to guarantee that it is.
629 unsafe { &mut *(begin..end).get_unchecked_mut(self) }
630 }
631
632 /// Divide one string slice into two at an index.
633 ///
634 /// The argument, `mid`, should be a byte offset from the start of the
635 /// string. It must also be on the boundary of a UTF-8 code point.
636 ///
637 /// The two slices returned go from the start of the string slice to `mid`,
638 /// and from `mid` to the end of the string slice.
639 ///
640 /// To get mutable string slices instead, see the [`split_at_mut`]
641 /// method.
642 ///
643 /// [`split_at_mut`]: str::split_at_mut
644 ///
645 /// # Panics
646 ///
647 /// Panics if `mid` is not on a UTF-8 code point boundary, or if it is past
648 /// the end of the last code point of the string slice. For a non-panicking
649 /// alternative see [`split_at_checked`](str::split_at_checked).
650 ///
651 /// # Examples
652 ///
653 /// ```
654 /// let s = "Per Martin-Löf";
655 ///
656 /// let (first, last) = s.split_at(3);
657 ///
658 /// assert_eq!("Per", first);
659 /// assert_eq!(" Martin-Löf", last);
660 /// ```
661 #[inline]
662 #[must_use]
663 #[stable(feature = "str_split_at", since = "1.4.0")]
664 pub fn split_at(&self, mid: usize) -> (&str, &str) {
665 match self.split_at_checked(mid) {
666 None => slice_error_fail(self, 0, mid),
667 Some(pair) => pair,
668 }
669 }
670
671 /// Divide one mutable string slice into two at an index.
672 ///
673 /// The argument, `mid`, should be a byte offset from the start of the
674 /// string. It must also be on the boundary of a UTF-8 code point.
675 ///
676 /// The two slices returned go from the start of the string slice to `mid`,
677 /// and from `mid` to the end of the string slice.
678 ///
679 /// To get immutable string slices instead, see the [`split_at`] method.
680 ///
681 /// [`split_at`]: str::split_at
682 ///
683 /// # Panics
684 ///
685 /// Panics if `mid` is not on a UTF-8 code point boundary, or if it is past
686 /// the end of the last code point of the string slice. For a non-panicking
687 /// alternative see [`split_at_mut_checked`](str::split_at_mut_checked).
688 ///
689 /// # Examples
690 ///
691 /// ```
692 /// let mut s = "Per Martin-Löf".to_string();
693 /// {
694 /// let (first, last) = s.split_at_mut(3);
695 /// first.make_ascii_uppercase();
696 /// assert_eq!("PER", first);
697 /// assert_eq!(" Martin-Löf", last);
698 /// }
699 /// assert_eq!("PER Martin-Löf", s);
700 /// ```
701 #[inline]
702 #[must_use]
703 #[stable(feature = "str_split_at", since = "1.4.0")]
704 pub fn split_at_mut(&mut self, mid: usize) -> (&mut str, &mut str) {
705 // is_char_boundary checks that the index is in [0, .len()]
706 if self.is_char_boundary(mid) {
707 // SAFETY: just checked that `mid` is on a char boundary.
708 unsafe { self.split_at_mut_unchecked(mid) }
709 } else {
710 slice_error_fail(self, 0, mid)
711 }
712 }
713
714 /// Divide one string slice into two at an index.
715 ///
716 /// The argument, `mid`, should be a valid byte offset from the start of the
717 /// string. It must also be on the boundary of a UTF-8 code point. The
718 /// method returns `None` if that’s not the case.
719 ///
720 /// The two slices returned go from the start of the string slice to `mid`,
721 /// and from `mid` to the end of the string slice.
722 ///
723 /// To get mutable string slices instead, see the [`split_at_mut_checked`]
724 /// method.
725 ///
726 /// [`split_at_mut_checked`]: str::split_at_mut_checked
727 ///
728 /// # Examples
729 ///
730 /// ```
731 /// #![feature(split_at_checked)]
732 ///
733 /// let s = "Per Martin-Löf";
734 ///
735 /// let (first, last) = s.split_at_checked(3).unwrap();
736 /// assert_eq!("Per", first);
737 /// assert_eq!(" Martin-Löf", last);
738 ///
739 /// assert_eq!(None, s.split_at_checked(13)); // Inside “ö”
740 /// assert_eq!(None, s.split_at_checked(16)); // Beyond the string length
741 /// ```
742 #[inline]
743 #[must_use]
744 #[unstable(feature = "split_at_checked", reason = "new API", issue = "119128")]
745 pub fn split_at_checked(&self, mid: usize) -> Option<(&str, &str)> {
746 // is_char_boundary checks that the index is in [0, .len()]
747 if self.is_char_boundary(mid) {
748 // SAFETY: just checked that `mid` is on a char boundary.
749 Some(unsafe { (self.get_unchecked(0..mid), self.get_unchecked(mid..self.len())) })
750 } else {
751 None
752 }
753 }
754
755 /// Divide one mutable string slice into two at an index.
756 ///
757 /// The argument, `mid`, should be a valid byte offset from the start of the
758 /// string. It must also be on the boundary of a UTF-8 code point. The
759 /// method returns `None` if that’s not the case.
760 ///
761 /// The two slices returned go from the start of the string slice to `mid`,
762 /// and from `mid` to the end of the string slice.
763 ///
764 /// To get immutable string slices instead, see the [`split_at_checked`] method.
765 ///
766 /// [`split_at_checked`]: str::split_at_checked
767 ///
768 /// # Examples
769 ///
770 /// ```
771 /// #![feature(split_at_checked)]
772 ///
773 /// let mut s = "Per Martin-Löf".to_string();
774 /// if let Some((first, last)) = s.split_at_mut_checked(3) {
775 /// first.make_ascii_uppercase();
776 /// assert_eq!("PER", first);
777 /// assert_eq!(" Martin-Löf", last);
778 /// }
779 /// assert_eq!("PER Martin-Löf", s);
780 ///
781 /// assert_eq!(None, s.split_at_mut_checked(13)); // Inside “ö”
782 /// assert_eq!(None, s.split_at_mut_checked(16)); // Beyond the string length
783 /// ```
784 #[inline]
785 #[must_use]
786 #[unstable(feature = "split_at_checked", reason = "new API", issue = "119128")]
787 pub fn split_at_mut_checked(&mut self, mid: usize) -> Option<(&mut str, &mut str)> {
788 // is_char_boundary checks that the index is in [0, .len()]
789 if self.is_char_boundary(mid) {
790 // SAFETY: just checked that `mid` is on a char boundary.
791 Some(unsafe { self.split_at_mut_unchecked(mid) })
792 } else {
793 None
794 }
795 }
796
797 /// Divide one string slice into two at an index.
798 ///
799 /// # Safety
800 ///
801 /// The caller must ensure that `mid` is a valid byte offset from the start
802 /// of the string and falls on the boundary of a UTF-8 code point.
803 unsafe fn split_at_mut_unchecked(&mut self, mid: usize) -> (&mut str, &mut str) {
804 let len = self.len();
805 let ptr = self.as_mut_ptr();
806 // SAFETY: caller guarantees `mid` is on a char boundary.
807 unsafe {
808 (
809 from_utf8_unchecked_mut(slice::from_raw_parts_mut(ptr, mid)),
810 from_utf8_unchecked_mut(slice::from_raw_parts_mut(ptr.add(mid), len - mid)),
811 )
812 }
813 }
814
815 /// Returns an iterator over the [`char`]s of a string slice.
816 ///
817 /// As a string slice consists of valid UTF-8, we can iterate through a
818 /// string slice by [`char`]. This method returns such an iterator.
819 ///
820 /// It's important to remember that [`char`] represents a Unicode Scalar
821 /// Value, and might not match your idea of what a 'character' is. Iteration
822 /// over grapheme clusters may be what you actually want. This functionality
823 /// is not provided by Rust's standard library, check crates.io instead.
824 ///
825 /// # Examples
826 ///
827 /// Basic usage:
828 ///
829 /// ```
830 /// let word = "goodbye";
831 ///
832 /// let count = word.chars().count();
833 /// assert_eq!(7, count);
834 ///
835 /// let mut chars = word.chars();
836 ///
837 /// assert_eq!(Some('g'), chars.next());
838 /// assert_eq!(Some('o'), chars.next());
839 /// assert_eq!(Some('o'), chars.next());
840 /// assert_eq!(Some('d'), chars.next());
841 /// assert_eq!(Some('b'), chars.next());
842 /// assert_eq!(Some('y'), chars.next());
843 /// assert_eq!(Some('e'), chars.next());
844 ///
845 /// assert_eq!(None, chars.next());
846 /// ```
847 ///
848 /// Remember, [`char`]s might not match your intuition about characters:
849 ///
850 /// [`char`]: prim@char
851 ///
852 /// ```
853 /// let y = "y̆";
854 ///
855 /// let mut chars = y.chars();
856 ///
857 /// assert_eq!(Some('y'), chars.next()); // not 'y̆'
858 /// assert_eq!(Some('\u{0306}'), chars.next());
859 ///
860 /// assert_eq!(None, chars.next());
861 /// ```
862 #[stable(feature = "rust1", since = "1.0.0")]
863 #[inline]
864 pub fn chars(&self) -> Chars<'_> {
865 Chars { iter: self.as_bytes().iter() }
866 }
867
868 /// Returns an iterator over the [`char`]s of a string slice, and their
869 /// positions.
870 ///
871 /// As a string slice consists of valid UTF-8, we can iterate through a
872 /// string slice by [`char`]. This method returns an iterator of both
873 /// these [`char`]s, as well as their byte positions.
874 ///
875 /// The iterator yields tuples. The position is first, the [`char`] is
876 /// second.
877 ///
878 /// # Examples
879 ///
880 /// Basic usage:
881 ///
882 /// ```
883 /// let word = "goodbye";
884 ///
885 /// let count = word.char_indices().count();
886 /// assert_eq!(7, count);
887 ///
888 /// let mut char_indices = word.char_indices();
889 ///
890 /// assert_eq!(Some((0, 'g')), char_indices.next());
891 /// assert_eq!(Some((1, 'o')), char_indices.next());
892 /// assert_eq!(Some((2, 'o')), char_indices.next());
893 /// assert_eq!(Some((3, 'd')), char_indices.next());
894 /// assert_eq!(Some((4, 'b')), char_indices.next());
895 /// assert_eq!(Some((5, 'y')), char_indices.next());
896 /// assert_eq!(Some((6, 'e')), char_indices.next());
897 ///
898 /// assert_eq!(None, char_indices.next());
899 /// ```
900 ///
901 /// Remember, [`char`]s might not match your intuition about characters:
902 ///
903 /// [`char`]: prim@char
904 ///
905 /// ```
906 /// let yes = "y̆es";
907 ///
908 /// let mut char_indices = yes.char_indices();
909 ///
910 /// assert_eq!(Some((0, 'y')), char_indices.next()); // not (0, 'y̆')
911 /// assert_eq!(Some((1, '\u{0306}')), char_indices.next());
912 ///
913 /// // note the 3 here - the previous character took up two bytes
914 /// assert_eq!(Some((3, 'e')), char_indices.next());
915 /// assert_eq!(Some((4, 's')), char_indices.next());
916 ///
917 /// assert_eq!(None, char_indices.next());
918 /// ```
919 #[stable(feature = "rust1", since = "1.0.0")]
920 #[inline]
921 pub fn char_indices(&self) -> CharIndices<'_> {
922 CharIndices { front_offset: 0, iter: self.chars() }
923 }
924
925 /// An iterator over the bytes of a string slice.
926 ///
927 /// As a string slice consists of a sequence of bytes, we can iterate
928 /// through a string slice by byte. This method returns such an iterator.
929 ///
930 /// # Examples
931 ///
932 /// ```
933 /// let mut bytes = "bors".bytes();
934 ///
935 /// assert_eq!(Some(b'b'), bytes.next());
936 /// assert_eq!(Some(b'o'), bytes.next());
937 /// assert_eq!(Some(b'r'), bytes.next());
938 /// assert_eq!(Some(b's'), bytes.next());
939 ///
940 /// assert_eq!(None, bytes.next());
941 /// ```
942 #[stable(feature = "rust1", since = "1.0.0")]
943 #[inline]
944 pub fn bytes(&self) -> Bytes<'_> {
945 Bytes(self.as_bytes().iter().copied())
946 }
947
948 /// Splits a string slice by whitespace.
949 ///
950 /// The iterator returned will return string slices that are sub-slices of
951 /// the original string slice, separated by any amount of whitespace.
952 ///
953 /// 'Whitespace' is defined according to the terms of the Unicode Derived
954 /// Core Property `White_Space`. If you only want to split on ASCII whitespace
955 /// instead, use [`split_ascii_whitespace`].
956 ///
957 /// [`split_ascii_whitespace`]: str::split_ascii_whitespace
958 ///
959 /// # Examples
960 ///
961 /// Basic usage:
962 ///
963 /// ```
964 /// let mut iter = "A few words".split_whitespace();
965 ///
966 /// assert_eq!(Some("A"), iter.next());
967 /// assert_eq!(Some("few"), iter.next());
968 /// assert_eq!(Some("words"), iter.next());
969 ///
970 /// assert_eq!(None, iter.next());
971 /// ```
972 ///
973 /// All kinds of whitespace are considered:
974 ///
975 /// ```
976 /// let mut iter = " Mary had\ta\u{2009}little \n\t lamb".split_whitespace();
977 /// assert_eq!(Some("Mary"), iter.next());
978 /// assert_eq!(Some("had"), iter.next());
979 /// assert_eq!(Some("a"), iter.next());
980 /// assert_eq!(Some("little"), iter.next());
981 /// assert_eq!(Some("lamb"), iter.next());
982 ///
983 /// assert_eq!(None, iter.next());
984 /// ```
985 ///
986 /// If the string is empty or all whitespace, the iterator yields no string slices:
987 /// ```
988 /// assert_eq!("".split_whitespace().next(), None);
989 /// assert_eq!(" ".split_whitespace().next(), None);
990 /// ```
991 #[must_use = "this returns the split string as an iterator, \
992 without modifying the original"]
993 #[stable(feature = "split_whitespace", since = "1.1.0")]
994 #[cfg_attr(not(test), rustc_diagnostic_item = "str_split_whitespace")]
995 #[inline]
996 pub fn split_whitespace(&self) -> SplitWhitespace<'_> {
997 SplitWhitespace { inner: self.split(IsWhitespace).filter(IsNotEmpty) }
998 }
999
1000 /// Splits a string slice by ASCII whitespace.
1001 ///
1002 /// The iterator returned will return string slices that are sub-slices of
1003 /// the original string slice, separated by any amount of ASCII whitespace.
1004 ///
1005 /// To split by Unicode `Whitespace` instead, use [`split_whitespace`].
1006 ///
1007 /// [`split_whitespace`]: str::split_whitespace
1008 ///
1009 /// # Examples
1010 ///
1011 /// Basic usage:
1012 ///
1013 /// ```
1014 /// let mut iter = "A few words".split_ascii_whitespace();
1015 ///
1016 /// assert_eq!(Some("A"), iter.next());
1017 /// assert_eq!(Some("few"), iter.next());
1018 /// assert_eq!(Some("words"), iter.next());
1019 ///
1020 /// assert_eq!(None, iter.next());
1021 /// ```
1022 ///
1023 /// All kinds of ASCII whitespace are considered:
1024 ///
1025 /// ```
1026 /// let mut iter = " Mary had\ta little \n\t lamb".split_ascii_whitespace();
1027 /// assert_eq!(Some("Mary"), iter.next());
1028 /// assert_eq!(Some("had"), iter.next());
1029 /// assert_eq!(Some("a"), iter.next());
1030 /// assert_eq!(Some("little"), iter.next());
1031 /// assert_eq!(Some("lamb"), iter.next());
1032 ///
1033 /// assert_eq!(None, iter.next());
1034 /// ```
1035 ///
1036 /// If the string is empty or all ASCII whitespace, the iterator yields no string slices:
1037 /// ```
1038 /// assert_eq!("".split_ascii_whitespace().next(), None);
1039 /// assert_eq!(" ".split_ascii_whitespace().next(), None);
1040 /// ```
1041 #[must_use = "this returns the split string as an iterator, \
1042 without modifying the original"]
1043 #[stable(feature = "split_ascii_whitespace", since = "1.34.0")]
1044 #[inline]
1045 pub fn split_ascii_whitespace(&self) -> SplitAsciiWhitespace<'_> {
1046 let inner =
1047 self.as_bytes().split(IsAsciiWhitespace).filter(BytesIsNotEmpty).map(UnsafeBytesToStr);
1048 SplitAsciiWhitespace { inner }
1049 }
1050
1051 /// An iterator over the lines of a string, as string slices.
1052 ///
1053 /// Lines are split at line endings that are either newlines (`\n`) or
1054 /// sequences of a carriage return followed by a line feed (`\r\n`).
1055 ///
1056 /// Line terminators are not included in the lines returned by the iterator.
1057 ///
1058 /// Note that any carriage return (`\r`) not immediately followed by a
1059 /// line feed (`\n`) does not split a line. These carriage returns are
1060 /// thereby included in the produced lines.
1061 ///
1062 /// The final line ending is optional. A string that ends with a final line
1063 /// ending will return the same lines as an otherwise identical string
1064 /// without a final line ending.
1065 ///
1066 /// # Examples
1067 ///
1068 /// Basic usage:
1069 ///
1070 /// ```
1071 /// let text = "foo\r\nbar\n\nbaz\r";
1072 /// let mut lines = text.lines();
1073 ///
1074 /// assert_eq!(Some("foo"), lines.next());
1075 /// assert_eq!(Some("bar"), lines.next());
1076 /// assert_eq!(Some(""), lines.next());
1077 /// // Trailing carriage return is included in the last line
1078 /// assert_eq!(Some("baz\r"), lines.next());
1079 ///
1080 /// assert_eq!(None, lines.next());
1081 /// ```
1082 ///
1083 /// The final line does not require any ending:
1084 ///
1085 /// ```
1086 /// let text = "foo\nbar\n\r\nbaz";
1087 /// let mut lines = text.lines();
1088 ///
1089 /// assert_eq!(Some("foo"), lines.next());
1090 /// assert_eq!(Some("bar"), lines.next());
1091 /// assert_eq!(Some(""), lines.next());
1092 /// assert_eq!(Some("baz"), lines.next());
1093 ///
1094 /// assert_eq!(None, lines.next());
1095 /// ```
1096 #[stable(feature = "rust1", since = "1.0.0")]
1097 #[inline]
1098 pub fn lines(&self) -> Lines<'_> {
1099 Lines(self.split_inclusive('\n').map(LinesMap))
1100 }
1101
1102 /// An iterator over the lines of a string.
1103 #[stable(feature = "rust1", since = "1.0.0")]
1104 #[deprecated(since = "1.4.0", note = "use lines() instead now", suggestion = "lines")]
1105 #[inline]
1106 #[allow(deprecated)]
1107 pub fn lines_any(&self) -> LinesAny<'_> {
1108 LinesAny(self.lines())
1109 }
1110
1111 /// Returns an iterator of `u16` over the string encoded as UTF-16.
1112 ///
1113 /// # Examples
1114 ///
1115 /// ```
1116 /// let text = "Zażółć gęślą jaźń";
1117 ///
1118 /// let utf8_len = text.len();
1119 /// let utf16_len = text.encode_utf16().count();
1120 ///
1121 /// assert!(utf16_len <= utf8_len);
1122 /// ```
1123 #[must_use = "this returns the encoded string as an iterator, \
1124 without modifying the original"]
1125 #[stable(feature = "encode_utf16", since = "1.8.0")]
1126 pub fn encode_utf16(&self) -> EncodeUtf16<'_> {
1127 EncodeUtf16 { chars: self.chars(), extra: 0 }
1128 }
1129
1130 /// Returns `true` if the given pattern matches a sub-slice of
1131 /// this string slice.
1132 ///
1133 /// Returns `false` if it does not.
1134 ///
1135 /// The [pattern] can be a `&str`, [`char`], a slice of [`char`]s, or a
1136 /// function or closure that determines if a character matches.
1137 ///
1138 /// [`char`]: prim@char
1139 /// [pattern]: self::pattern
1140 ///
1141 /// # Examples
1142 ///
1143 /// ```
1144 /// let bananas = "bananas";
1145 ///
1146 /// assert!(bananas.contains("nana"));
1147 /// assert!(!bananas.contains("apples"));
1148 /// ```
1149 #[stable(feature = "rust1", since = "1.0.0")]
1150 #[inline]
1151 pub fn contains<'a, P: Pattern<'a>>(&'a self, pat: P) -> bool {
1152 pat.is_contained_in(self)
1153 }
1154
1155 /// Returns `true` if the given pattern matches a prefix of this
1156 /// string slice.
1157 ///
1158 /// Returns `false` if it does not.
1159 ///
1160 /// The [pattern] can be a `&str`, [`char`], a slice of [`char`]s, or a
1161 /// function or closure that determines if a character matches.
1162 ///
1163 /// [`char`]: prim@char
1164 /// [pattern]: self::pattern
1165 ///
1166 /// # Examples
1167 ///
1168 /// ```
1169 /// let bananas = "bananas";
1170 ///
1171 /// assert!(bananas.starts_with("bana"));
1172 /// assert!(!bananas.starts_with("nana"));
1173 /// ```
1174 #[stable(feature = "rust1", since = "1.0.0")]
1175 pub fn starts_with<'a, P: Pattern<'a>>(&'a self, pat: P) -> bool {
1176 pat.is_prefix_of(self)
1177 }
1178
1179 /// Returns `true` if the given pattern matches a suffix of this
1180 /// string slice.
1181 ///
1182 /// Returns `false` if it does not.
1183 ///
1184 /// The [pattern] can be a `&str`, [`char`], a slice of [`char`]s, or a
1185 /// function or closure that determines if a character matches.
1186 ///
1187 /// [`char`]: prim@char
1188 /// [pattern]: self::pattern
1189 ///
1190 /// # Examples
1191 ///
1192 /// ```
1193 /// let bananas = "bananas";
1194 ///
1195 /// assert!(bananas.ends_with("anas"));
1196 /// assert!(!bananas.ends_with("nana"));
1197 /// ```
1198 #[stable(feature = "rust1", since = "1.0.0")]
1199 pub fn ends_with<'a, P>(&'a self, pat: P) -> bool
1200 where
1201 P: Pattern<'a, Searcher: ReverseSearcher<'a>>,
1202 {
1203 pat.is_suffix_of(self)
1204 }
1205
1206 /// Returns the byte index of the first character of this string slice that
1207 /// matches the pattern.
1208 ///
1209 /// Returns [`None`] if the pattern doesn't match.
1210 ///
1211 /// The [pattern] can be a `&str`, [`char`], a slice of [`char`]s, or a
1212 /// function or closure that determines if a character matches.
1213 ///
1214 /// [`char`]: prim@char
1215 /// [pattern]: self::pattern
1216 ///
1217 /// # Examples
1218 ///
1219 /// Simple patterns:
1220 ///
1221 /// ```
1222 /// let s = "Löwe 老虎 Léopard Gepardi";
1223 ///
1224 /// assert_eq!(s.find('L'), Some(0));
1225 /// assert_eq!(s.find('é'), Some(14));
1226 /// assert_eq!(s.find("pard"), Some(17));
1227 /// ```
1228 ///
1229 /// More complex patterns using point-free style and closures:
1230 ///
1231 /// ```
1232 /// let s = "Löwe 老虎 Léopard";
1233 ///
1234 /// assert_eq!(s.find(char::is_whitespace), Some(5));
1235 /// assert_eq!(s.find(char::is_lowercase), Some(1));
1236 /// assert_eq!(s.find(|c: char| c.is_whitespace() || c.is_lowercase()), Some(1));
1237 /// assert_eq!(s.find(|c: char| (c < 'o') && (c > 'a')), Some(4));
1238 /// ```
1239 ///
1240 /// Not finding the pattern:
1241 ///
1242 /// ```
1243 /// let s = "Löwe 老虎 Léopard";
1244 /// let x: &[_] = &['1', '2'];
1245 ///
1246 /// assert_eq!(s.find(x), None);
1247 /// ```
1248 #[stable(feature = "rust1", since = "1.0.0")]
1249 #[inline]
1250 pub fn find<'a, P: Pattern<'a>>(&'a self, pat: P) -> Option<usize> {
1251 pat.into_searcher(self).next_match().map(|(i, _)| i)
1252 }
1253
1254 /// Returns the byte index for the first character of the last match of the pattern in
1255 /// this string slice.
1256 ///
1257 /// Returns [`None`] if the pattern doesn't match.
1258 ///
1259 /// The [pattern] can be a `&str`, [`char`], a slice of [`char`]s, or a
1260 /// function or closure that determines if a character matches.
1261 ///
1262 /// [`char`]: prim@char
1263 /// [pattern]: self::pattern
1264 ///
1265 /// # Examples
1266 ///
1267 /// Simple patterns:
1268 ///
1269 /// ```
1270 /// let s = "Löwe 老虎 Léopard Gepardi";
1271 ///
1272 /// assert_eq!(s.rfind('L'), Some(13));
1273 /// assert_eq!(s.rfind('é'), Some(14));
1274 /// assert_eq!(s.rfind("pard"), Some(24));
1275 /// ```
1276 ///
1277 /// More complex patterns with closures:
1278 ///
1279 /// ```
1280 /// let s = "Löwe 老虎 Léopard";
1281 ///
1282 /// assert_eq!(s.rfind(char::is_whitespace), Some(12));
1283 /// assert_eq!(s.rfind(char::is_lowercase), Some(20));
1284 /// ```
1285 ///
1286 /// Not finding the pattern:
1287 ///
1288 /// ```
1289 /// let s = "Löwe 老虎 Léopard";
1290 /// let x: &[_] = &['1', '2'];
1291 ///
1292 /// assert_eq!(s.rfind(x), None);
1293 /// ```
1294 #[stable(feature = "rust1", since = "1.0.0")]
1295 #[inline]
1296 pub fn rfind<'a, P>(&'a self, pat: P) -> Option<usize>
1297 where
1298 P: Pattern<'a, Searcher: ReverseSearcher<'a>>,
1299 {
1300 pat.into_searcher(self).next_match_back().map(|(i, _)| i)
1301 }
1302
1303 /// An iterator over substrings of this string slice, separated by
1304 /// characters matched by a pattern.
1305 ///
1306 /// The [pattern] can be a `&str`, [`char`], a slice of [`char`]s, or a
1307 /// function or closure that determines if a character matches.
1308 ///
1309 /// [`char`]: prim@char
1310 /// [pattern]: self::pattern
1311 ///
1312 /// # Iterator behavior
1313 ///
1314 /// The returned iterator will be a [`DoubleEndedIterator`] if the pattern
1315 /// allows a reverse search and forward/reverse search yields the same
1316 /// elements. This is true for, e.g., [`char`], but not for `&str`.
1317 ///
1318 /// If the pattern allows a reverse search but its results might differ
1319 /// from a forward search, the [`rsplit`] method can be used.
1320 ///
1321 /// [`rsplit`]: str::rsplit
1322 ///
1323 /// # Examples
1324 ///
1325 /// Simple patterns:
1326 ///
1327 /// ```
1328 /// let v: Vec<&str> = "Mary had a little lamb".split(' ').collect();
1329 /// assert_eq!(v, ["Mary", "had", "a", "little", "lamb"]);
1330 ///
1331 /// let v: Vec<&str> = "".split('X').collect();
1332 /// assert_eq!(v, [""]);
1333 ///
1334 /// let v: Vec<&str> = "lionXXtigerXleopard".split('X').collect();
1335 /// assert_eq!(v, ["lion", "", "tiger", "leopard"]);
1336 ///
1337 /// let v: Vec<&str> = "lion::tiger::leopard".split("::").collect();
1338 /// assert_eq!(v, ["lion", "tiger", "leopard"]);
1339 ///
1340 /// let v: Vec<&str> = "abc1def2ghi".split(char::is_numeric).collect();
1341 /// assert_eq!(v, ["abc", "def", "ghi"]);
1342 ///
1343 /// let v: Vec<&str> = "lionXtigerXleopard".split(char::is_uppercase).collect();
1344 /// assert_eq!(v, ["lion", "tiger", "leopard"]);
1345 /// ```
1346 ///
1347 /// If the pattern is a slice of chars, split on each occurrence of any of the characters:
1348 ///
1349 /// ```
1350 /// let v: Vec<&str> = "2020-11-03 23:59".split(&['-', ' ', ':', '@'][..]).collect();
1351 /// assert_eq!(v, ["2020", "11", "03", "23", "59"]);
1352 /// ```
1353 ///
1354 /// A more complex pattern, using a closure:
1355 ///
1356 /// ```
1357 /// let v: Vec<&str> = "abc1defXghi".split(|c| c == '1' || c == 'X').collect();
1358 /// assert_eq!(v, ["abc", "def", "ghi"]);
1359 /// ```
1360 ///
1361 /// If a string contains multiple contiguous separators, you will end up
1362 /// with empty strings in the output:
1363 ///
1364 /// ```
1365 /// let x = "||||a||b|c".to_string();
1366 /// let d: Vec<_> = x.split('|').collect();
1367 ///
1368 /// assert_eq!(d, &["", "", "", "", "a", "", "b", "c"]);
1369 /// ```
1370 ///
1371 /// Contiguous separators are separated by the empty string.
1372 ///
1373 /// ```
1374 /// let x = "(///)".to_string();
1375 /// let d: Vec<_> = x.split('/').collect();
1376 ///
1377 /// assert_eq!(d, &["(", "", "", ")"]);
1378 /// ```
1379 ///
1380 /// Separators at the start or end of a string are neighbored
1381 /// by empty strings.
1382 ///
1383 /// ```
1384 /// let d: Vec<_> = "010".split("0").collect();
1385 /// assert_eq!(d, &["", "1", ""]);
1386 /// ```
1387 ///
1388 /// When the empty string is used as a separator, it separates
1389 /// every character in the string, along with the beginning
1390 /// and end of the string.
1391 ///
1392 /// ```
1393 /// let f: Vec<_> = "rust".split("").collect();
1394 /// assert_eq!(f, &["", "r", "u", "s", "t", ""]);
1395 /// ```
1396 ///
1397 /// Contiguous separators can lead to possibly surprising behavior
1398 /// when whitespace is used as the separator. This code is correct:
1399 ///
1400 /// ```
1401 /// let x = " a b c".to_string();
1402 /// let d: Vec<_> = x.split(' ').collect();
1403 ///
1404 /// assert_eq!(d, &["", "", "", "", "a", "", "b", "c"]);
1405 /// ```
1406 ///
1407 /// It does _not_ give you:
1408 ///
1409 /// ```,ignore
1410 /// assert_eq!(d, &["a", "b", "c"]);
1411 /// ```
1412 ///
1413 /// Use [`split_whitespace`] for this behavior.
1414 ///
1415 /// [`split_whitespace`]: str::split_whitespace
1416 #[stable(feature = "rust1", since = "1.0.0")]
1417 #[inline]
1418 pub fn split<'a, P: Pattern<'a>>(&'a self, pat: P) -> Split<'a, P> {
1419 Split(SplitInternal {
1420 start: 0,
1421 end: self.len(),
1422 matcher: pat.into_searcher(self),
1423 allow_trailing_empty: true,
1424 finished: false,
1425 })
1426 }
1427
1428 /// An iterator over substrings of this string slice, separated by
1429 /// characters matched by a pattern. Differs from the iterator produced by
1430 /// `split` in that `split_inclusive` leaves the matched part as the
1431 /// terminator of the substring.
1432 ///
1433 /// The [pattern] can be a `&str`, [`char`], a slice of [`char`]s, or a
1434 /// function or closure that determines if a character matches.
1435 ///
1436 /// [`char`]: prim@char
1437 /// [pattern]: self::pattern
1438 ///
1439 /// # Examples
1440 ///
1441 /// ```
1442 /// let v: Vec<&str> = "Mary had a little lamb\nlittle lamb\nlittle lamb."
1443 /// .split_inclusive('\n').collect();
1444 /// assert_eq!(v, ["Mary had a little lamb\n", "little lamb\n", "little lamb."]);
1445 /// ```
1446 ///
1447 /// If the last element of the string is matched,
1448 /// that element will be considered the terminator of the preceding substring.
1449 /// That substring will be the last item returned by the iterator.
1450 ///
1451 /// ```
1452 /// let v: Vec<&str> = "Mary had a little lamb\nlittle lamb\nlittle lamb.\n"
1453 /// .split_inclusive('\n').collect();
1454 /// assert_eq!(v, ["Mary had a little lamb\n", "little lamb\n", "little lamb.\n"]);
1455 /// ```
1456 #[stable(feature = "split_inclusive", since = "1.51.0")]
1457 #[inline]
1458 pub fn split_inclusive<'a, P: Pattern<'a>>(&'a self, pat: P) -> SplitInclusive<'a, P> {
1459 SplitInclusive(SplitInternal {
1460 start: 0,
1461 end: self.len(),
1462 matcher: pat.into_searcher(self),
1463 allow_trailing_empty: false,
1464 finished: false,
1465 })
1466 }
1467
1468 /// An iterator over substrings of the given string slice, separated by
1469 /// characters matched by a pattern and yielded in reverse order.
1470 ///
1471 /// The [pattern] can be a `&str`, [`char`], a slice of [`char`]s, or a
1472 /// function or closure that determines if a character matches.
1473 ///
1474 /// [`char`]: prim@char
1475 /// [pattern]: self::pattern
1476 ///
1477 /// # Iterator behavior
1478 ///
1479 /// The returned iterator requires that the pattern supports a reverse
1480 /// search, and it will be a [`DoubleEndedIterator`] if a forward/reverse
1481 /// search yields the same elements.
1482 ///
1483 /// For iterating from the front, the [`split`] method can be used.
1484 ///
1485 /// [`split`]: str::split
1486 ///
1487 /// # Examples
1488 ///
1489 /// Simple patterns:
1490 ///
1491 /// ```
1492 /// let v: Vec<&str> = "Mary had a little lamb".rsplit(' ').collect();
1493 /// assert_eq!(v, ["lamb", "little", "a", "had", "Mary"]);
1494 ///
1495 /// let v: Vec<&str> = "".rsplit('X').collect();
1496 /// assert_eq!(v, [""]);
1497 ///
1498 /// let v: Vec<&str> = "lionXXtigerXleopard".rsplit('X').collect();
1499 /// assert_eq!(v, ["leopard", "tiger", "", "lion"]);
1500 ///
1501 /// let v: Vec<&str> = "lion::tiger::leopard".rsplit("::").collect();
1502 /// assert_eq!(v, ["leopard", "tiger", "lion"]);
1503 /// ```
1504 ///
1505 /// A more complex pattern, using a closure:
1506 ///
1507 /// ```
1508 /// let v: Vec<&str> = "abc1defXghi".rsplit(|c| c == '1' || c == 'X').collect();
1509 /// assert_eq!(v, ["ghi", "def", "abc"]);
1510 /// ```
1511 #[stable(feature = "rust1", since = "1.0.0")]
1512 #[inline]
1513 pub fn rsplit<'a, P>(&'a self, pat: P) -> RSplit<'a, P>
1514 where
1515 P: Pattern<'a, Searcher: ReverseSearcher<'a>>,
1516 {
1517 RSplit(self.split(pat).0)
1518 }
1519
1520 /// An iterator over substrings of the given string slice, separated by
1521 /// characters matched by a pattern.
1522 ///
1523 /// The [pattern] can be a `&str`, [`char`], a slice of [`char`]s, or a
1524 /// function or closure that determines if a character matches.
1525 ///
1526 /// [`char`]: prim@char
1527 /// [pattern]: self::pattern
1528 ///
1529 /// Equivalent to [`split`], except that the trailing substring
1530 /// is skipped if empty.
1531 ///
1532 /// [`split`]: str::split
1533 ///
1534 /// This method can be used for string data that is _terminated_,
1535 /// rather than _separated_ by a pattern.
1536 ///
1537 /// # Iterator behavior
1538 ///
1539 /// The returned iterator will be a [`DoubleEndedIterator`] if the pattern
1540 /// allows a reverse search and forward/reverse search yields the same
1541 /// elements. This is true for, e.g., [`char`], but not for `&str`.
1542 ///
1543 /// If the pattern allows a reverse search but its results might differ
1544 /// from a forward search, the [`rsplit_terminator`] method can be used.
1545 ///
1546 /// [`rsplit_terminator`]: str::rsplit_terminator
1547 ///
1548 /// # Examples
1549 ///
1550 /// ```
1551 /// let v: Vec<&str> = "A.B.".split_terminator('.').collect();
1552 /// assert_eq!(v, ["A", "B"]);
1553 ///
1554 /// let v: Vec<&str> = "A..B..".split_terminator(".").collect();
1555 /// assert_eq!(v, ["A", "", "B", ""]);
1556 ///
1557 /// let v: Vec<&str> = "A.B:C.D".split_terminator(&['.', ':'][..]).collect();
1558 /// assert_eq!(v, ["A", "B", "C", "D"]);
1559 /// ```
1560 #[stable(feature = "rust1", since = "1.0.0")]
1561 #[inline]
1562 pub fn split_terminator<'a, P: Pattern<'a>>(&'a self, pat: P) -> SplitTerminator<'a, P> {
1563 SplitTerminator(SplitInternal { allow_trailing_empty: false, ..self.split(pat).0 })
1564 }
1565
1566 /// An iterator over substrings of `self`, separated by characters
1567 /// matched by a pattern and yielded in reverse order.
1568 ///
1569 /// The [pattern] can be a `&str`, [`char`], a slice of [`char`]s, or a
1570 /// function or closure that determines if a character matches.
1571 ///
1572 /// [`char`]: prim@char
1573 /// [pattern]: self::pattern
1574 ///
1575 /// Equivalent to [`split`], except that the trailing substring is
1576 /// skipped if empty.
1577 ///
1578 /// [`split`]: str::split
1579 ///
1580 /// This method can be used for string data that is _terminated_,
1581 /// rather than _separated_ by a pattern.
1582 ///
1583 /// # Iterator behavior
1584 ///
1585 /// The returned iterator requires that the pattern supports a
1586 /// reverse search, and it will be double ended if a forward/reverse
1587 /// search yields the same elements.
1588 ///
1589 /// For iterating from the front, the [`split_terminator`] method can be
1590 /// used.
1591 ///
1592 /// [`split_terminator`]: str::split_terminator
1593 ///
1594 /// # Examples
1595 ///
1596 /// ```
1597 /// let v: Vec<&str> = "A.B.".rsplit_terminator('.').collect();
1598 /// assert_eq!(v, ["B", "A"]);
1599 ///
1600 /// let v: Vec<&str> = "A..B..".rsplit_terminator(".").collect();
1601 /// assert_eq!(v, ["", "B", "", "A"]);
1602 ///
1603 /// let v: Vec<&str> = "A.B:C.D".rsplit_terminator(&['.', ':'][..]).collect();
1604 /// assert_eq!(v, ["D", "C", "B", "A"]);
1605 /// ```
1606 #[stable(feature = "rust1", since = "1.0.0")]
1607 #[inline]
1608 pub fn rsplit_terminator<'a, P>(&'a self, pat: P) -> RSplitTerminator<'a, P>
1609 where
1610 P: Pattern<'a, Searcher: ReverseSearcher<'a>>,
1611 {
1612 RSplitTerminator(self.split_terminator(pat).0)
1613 }
1614
1615 /// An iterator over substrings of the given string slice, separated by a
1616 /// pattern, restricted to returning at most `n` items.
1617 ///
1618 /// If `n` substrings are returned, the last substring (the `n`th substring)
1619 /// will contain the remainder of the string.
1620 ///
1621 /// The [pattern] can be a `&str`, [`char`], a slice of [`char`]s, or a
1622 /// function or closure that determines if a character matches.
1623 ///
1624 /// [`char`]: prim@char
1625 /// [pattern]: self::pattern
1626 ///
1627 /// # Iterator behavior
1628 ///
1629 /// The returned iterator will not be double ended, because it is
1630 /// not efficient to support.
1631 ///
1632 /// If the pattern allows a reverse search, the [`rsplitn`] method can be
1633 /// used.
1634 ///
1635 /// [`rsplitn`]: str::rsplitn
1636 ///
1637 /// # Examples
1638 ///
1639 /// Simple patterns:
1640 ///
1641 /// ```
1642 /// let v: Vec<&str> = "Mary had a little lambda".splitn(3, ' ').collect();
1643 /// assert_eq!(v, ["Mary", "had", "a little lambda"]);
1644 ///
1645 /// let v: Vec<&str> = "lionXXtigerXleopard".splitn(3, "X").collect();
1646 /// assert_eq!(v, ["lion", "", "tigerXleopard"]);
1647 ///
1648 /// let v: Vec<&str> = "abcXdef".splitn(1, 'X').collect();
1649 /// assert_eq!(v, ["abcXdef"]);
1650 ///
1651 /// let v: Vec<&str> = "".splitn(1, 'X').collect();
1652 /// assert_eq!(v, [""]);
1653 /// ```
1654 ///
1655 /// A more complex pattern, using a closure:
1656 ///
1657 /// ```
1658 /// let v: Vec<&str> = "abc1defXghi".splitn(2, |c| c == '1' || c == 'X').collect();
1659 /// assert_eq!(v, ["abc", "defXghi"]);
1660 /// ```
1661 #[stable(feature = "rust1", since = "1.0.0")]
1662 #[inline]
1663 pub fn splitn<'a, P: Pattern<'a>>(&'a self, n: usize, pat: P) -> SplitN<'a, P> {
1664 SplitN(SplitNInternal { iter: self.split(pat).0, count: n })
1665 }
1666
1667 /// An iterator over substrings of this string slice, separated by a
1668 /// pattern, starting from the end of the string, restricted to returning
1669 /// at most `n` items.
1670 ///
1671 /// If `n` substrings are returned, the last substring (the `n`th substring)
1672 /// will contain the remainder of the string.
1673 ///
1674 /// The [pattern] can be a `&str`, [`char`], a slice of [`char`]s, or a
1675 /// function or closure that determines if a character matches.
1676 ///
1677 /// [`char`]: prim@char
1678 /// [pattern]: self::pattern
1679 ///
1680 /// # Iterator behavior
1681 ///
1682 /// The returned iterator will not be double ended, because it is not
1683 /// efficient to support.
1684 ///
1685 /// For splitting from the front, the [`splitn`] method can be used.
1686 ///
1687 /// [`splitn`]: str::splitn
1688 ///
1689 /// # Examples
1690 ///
1691 /// Simple patterns:
1692 ///
1693 /// ```
1694 /// let v: Vec<&str> = "Mary had a little lamb".rsplitn(3, ' ').collect();
1695 /// assert_eq!(v, ["lamb", "little", "Mary had a"]);
1696 ///
1697 /// let v: Vec<&str> = "lionXXtigerXleopard".rsplitn(3, 'X').collect();
1698 /// assert_eq!(v, ["leopard", "tiger", "lionX"]);
1699 ///
1700 /// let v: Vec<&str> = "lion::tiger::leopard".rsplitn(2, "::").collect();
1701 /// assert_eq!(v, ["leopard", "lion::tiger"]);
1702 /// ```
1703 ///
1704 /// A more complex pattern, using a closure:
1705 ///
1706 /// ```
1707 /// let v: Vec<&str> = "abc1defXghi".rsplitn(2, |c| c == '1' || c == 'X').collect();
1708 /// assert_eq!(v, ["ghi", "abc1def"]);
1709 /// ```
1710 #[stable(feature = "rust1", since = "1.0.0")]
1711 #[inline]
1712 pub fn rsplitn<'a, P>(&'a self, n: usize, pat: P) -> RSplitN<'a, P>
1713 where
1714 P: Pattern<'a, Searcher: ReverseSearcher<'a>>,
1715 {
1716 RSplitN(self.splitn(n, pat).0)
1717 }
1718
1719 /// Splits the string on the first occurrence of the specified delimiter and
1720 /// returns prefix before delimiter and suffix after delimiter.
1721 ///
1722 /// # Examples
1723 ///
1724 /// ```
1725 /// assert_eq!("cfg".split_once('='), None);
1726 /// assert_eq!("cfg=".split_once('='), Some(("cfg", "")));
1727 /// assert_eq!("cfg=foo".split_once('='), Some(("cfg", "foo")));
1728 /// assert_eq!("cfg=foo=bar".split_once('='), Some(("cfg", "foo=bar")));
1729 /// ```
1730 #[stable(feature = "str_split_once", since = "1.52.0")]
1731 #[inline]
1732 pub fn split_once<'a, P: Pattern<'a>>(&'a self, delimiter: P) -> Option<(&'a str, &'a str)> {
1733 let (start, end) = delimiter.into_searcher(self).next_match()?;
1734 // SAFETY: `Searcher` is known to return valid indices.
1735 unsafe { Some((self.get_unchecked(..start), self.get_unchecked(end..))) }
1736 }
1737
1738 /// Splits the string on the last occurrence of the specified delimiter and
1739 /// returns prefix before delimiter and suffix after delimiter.
1740 ///
1741 /// # Examples
1742 ///
1743 /// ```
1744 /// assert_eq!("cfg".rsplit_once('='), None);
1745 /// assert_eq!("cfg=foo".rsplit_once('='), Some(("cfg", "foo")));
1746 /// assert_eq!("cfg=foo=bar".rsplit_once('='), Some(("cfg=foo", "bar")));
1747 /// ```
1748 #[stable(feature = "str_split_once", since = "1.52.0")]
1749 #[inline]
1750 pub fn rsplit_once<'a, P>(&'a self, delimiter: P) -> Option<(&'a str, &'a str)>
1751 where
1752 P: Pattern<'a, Searcher: ReverseSearcher<'a>>,
1753 {
1754 let (start, end) = delimiter.into_searcher(self).next_match_back()?;
1755 // SAFETY: `Searcher` is known to return valid indices.
1756 unsafe { Some((self.get_unchecked(..start), self.get_unchecked(end..))) }
1757 }
1758
1759 /// An iterator over the disjoint matches of a pattern within the given string
1760 /// slice.
1761 ///
1762 /// The [pattern] can be a `&str`, [`char`], a slice of [`char`]s, or a
1763 /// function or closure that determines if a character matches.
1764 ///
1765 /// [`char`]: prim@char
1766 /// [pattern]: self::pattern
1767 ///
1768 /// # Iterator behavior
1769 ///
1770 /// The returned iterator will be a [`DoubleEndedIterator`] if the pattern
1771 /// allows a reverse search and forward/reverse search yields the same
1772 /// elements. This is true for, e.g., [`char`], but not for `&str`.
1773 ///
1774 /// If the pattern allows a reverse search but its results might differ
1775 /// from a forward search, the [`rmatches`] method can be used.
1776 ///
1777 /// [`rmatches`]: str::rmatches
1778 ///
1779 /// # Examples
1780 ///
1781 /// ```
1782 /// let v: Vec<&str> = "abcXXXabcYYYabc".matches("abc").collect();
1783 /// assert_eq!(v, ["abc", "abc", "abc"]);
1784 ///
1785 /// let v: Vec<&str> = "1abc2abc3".matches(char::is_numeric).collect();
1786 /// assert_eq!(v, ["1", "2", "3"]);
1787 /// ```
1788 #[stable(feature = "str_matches", since = "1.2.0")]
1789 #[inline]
1790 pub fn matches<'a, P: Pattern<'a>>(&'a self, pat: P) -> Matches<'a, P> {
1791 Matches(MatchesInternal(pat.into_searcher(self)))
1792 }
1793
1794 /// An iterator over the disjoint matches of a pattern within this string slice,
1795 /// yielded in reverse order.
1796 ///
1797 /// The [pattern] can be a `&str`, [`char`], a slice of [`char`]s, or a
1798 /// function or closure that determines if a character matches.
1799 ///
1800 /// [`char`]: prim@char
1801 /// [pattern]: self::pattern
1802 ///
1803 /// # Iterator behavior
1804 ///
1805 /// The returned iterator requires that the pattern supports a reverse
1806 /// search, and it will be a [`DoubleEndedIterator`] if a forward/reverse
1807 /// search yields the same elements.
1808 ///
1809 /// For iterating from the front, the [`matches`] method can be used.
1810 ///
1811 /// [`matches`]: str::matches
1812 ///
1813 /// # Examples
1814 ///
1815 /// ```
1816 /// let v: Vec<&str> = "abcXXXabcYYYabc".rmatches("abc").collect();
1817 /// assert_eq!(v, ["abc", "abc", "abc"]);
1818 ///
1819 /// let v: Vec<&str> = "1abc2abc3".rmatches(char::is_numeric).collect();
1820 /// assert_eq!(v, ["3", "2", "1"]);
1821 /// ```
1822 #[stable(feature = "str_matches", since = "1.2.0")]
1823 #[inline]
1824 pub fn rmatches<'a, P>(&'a self, pat: P) -> RMatches<'a, P>
1825 where
1826 P: Pattern<'a, Searcher: ReverseSearcher<'a>>,
1827 {
1828 RMatches(self.matches(pat).0)
1829 }
1830
1831 /// An iterator over the disjoint matches of a pattern within this string
1832 /// slice as well as the index that the match starts at.
1833 ///
1834 /// For matches of `pat` within `self` that overlap, only the indices
1835 /// corresponding to the first match are returned.
1836 ///
1837 /// The [pattern] can be a `&str`, [`char`], a slice of [`char`]s, or a
1838 /// function or closure that determines if a character matches.
1839 ///
1840 /// [`char`]: prim@char
1841 /// [pattern]: self::pattern
1842 ///
1843 /// # Iterator behavior
1844 ///
1845 /// The returned iterator will be a [`DoubleEndedIterator`] if the pattern
1846 /// allows a reverse search and forward/reverse search yields the same
1847 /// elements. This is true for, e.g., [`char`], but not for `&str`.
1848 ///
1849 /// If the pattern allows a reverse search but its results might differ
1850 /// from a forward search, the [`rmatch_indices`] method can be used.
1851 ///
1852 /// [`rmatch_indices`]: str::rmatch_indices
1853 ///
1854 /// # Examples
1855 ///
1856 /// ```
1857 /// let v: Vec<_> = "abcXXXabcYYYabc".match_indices("abc").collect();
1858 /// assert_eq!(v, [(0, "abc"), (6, "abc"), (12, "abc")]);
1859 ///
1860 /// let v: Vec<_> = "1abcabc2".match_indices("abc").collect();
1861 /// assert_eq!(v, [(1, "abc"), (4, "abc")]);
1862 ///
1863 /// let v: Vec<_> = "ababa".match_indices("aba").collect();
1864 /// assert_eq!(v, [(0, "aba")]); // only the first `aba`
1865 /// ```
1866 #[stable(feature = "str_match_indices", since = "1.5.0")]
1867 #[inline]
1868 pub fn match_indices<'a, P: Pattern<'a>>(&'a self, pat: P) -> MatchIndices<'a, P> {
1869 MatchIndices(MatchIndicesInternal(pat.into_searcher(self)))
1870 }
1871
1872 /// An iterator over the disjoint matches of a pattern within `self`,
1873 /// yielded in reverse order along with the index of the match.
1874 ///
1875 /// For matches of `pat` within `self` that overlap, only the indices
1876 /// corresponding to the last match are returned.
1877 ///
1878 /// The [pattern] can be a `&str`, [`char`], a slice of [`char`]s, or a
1879 /// function or closure that determines if a character matches.
1880 ///
1881 /// [`char`]: prim@char
1882 /// [pattern]: self::pattern
1883 ///
1884 /// # Iterator behavior
1885 ///
1886 /// The returned iterator requires that the pattern supports a reverse
1887 /// search, and it will be a [`DoubleEndedIterator`] if a forward/reverse
1888 /// search yields the same elements.
1889 ///
1890 /// For iterating from the front, the [`match_indices`] method can be used.
1891 ///
1892 /// [`match_indices`]: str::match_indices
1893 ///
1894 /// # Examples
1895 ///
1896 /// ```
1897 /// let v: Vec<_> = "abcXXXabcYYYabc".rmatch_indices("abc").collect();
1898 /// assert_eq!(v, [(12, "abc"), (6, "abc"), (0, "abc")]);
1899 ///
1900 /// let v: Vec<_> = "1abcabc2".rmatch_indices("abc").collect();
1901 /// assert_eq!(v, [(4, "abc"), (1, "abc")]);
1902 ///
1903 /// let v: Vec<_> = "ababa".rmatch_indices("aba").collect();
1904 /// assert_eq!(v, [(2, "aba")]); // only the last `aba`
1905 /// ```
1906 #[stable(feature = "str_match_indices", since = "1.5.0")]
1907 #[inline]
1908 pub fn rmatch_indices<'a, P>(&'a self, pat: P) -> RMatchIndices<'a, P>
1909 where
1910 P: Pattern<'a, Searcher: ReverseSearcher<'a>>,
1911 {
1912 RMatchIndices(self.match_indices(pat).0)
1913 }
1914
1915 /// Returns a string slice with leading and trailing whitespace removed.
1916 ///
1917 /// 'Whitespace' is defined according to the terms of the Unicode Derived
1918 /// Core Property `White_Space`, which includes newlines.
1919 ///
1920 /// # Examples
1921 ///
1922 /// ```
1923 /// let s = "\n Hello\tworld\t\n";
1924 ///
1925 /// assert_eq!("Hello\tworld", s.trim());
1926 /// ```
1927 #[inline]
1928 #[must_use = "this returns the trimmed string as a slice, \
1929 without modifying the original"]
1930 #[stable(feature = "rust1", since = "1.0.0")]
1931 #[cfg_attr(not(test), rustc_diagnostic_item = "str_trim")]
1932 pub fn trim(&self) -> &str {
1933 self.trim_matches(|c: char| c.is_whitespace())
1934 }
1935
1936 /// Returns a string slice with leading whitespace removed.
1937 ///
1938 /// 'Whitespace' is defined according to the terms of the Unicode Derived
1939 /// Core Property `White_Space`, which includes newlines.
1940 ///
1941 /// # Text directionality
1942 ///
1943 /// A string is a sequence of bytes. `start` in this context means the first
1944 /// position of that byte string; for a left-to-right language like English or
1945 /// Russian, this will be left side, and for right-to-left languages like
1946 /// Arabic or Hebrew, this will be the right side.
1947 ///
1948 /// # Examples
1949 ///
1950 /// Basic usage:
1951 ///
1952 /// ```
1953 /// let s = "\n Hello\tworld\t\n";
1954 /// assert_eq!("Hello\tworld\t\n", s.trim_start());
1955 /// ```
1956 ///
1957 /// Directionality:
1958 ///
1959 /// ```
1960 /// let s = " English ";
1961 /// assert!(Some('E') == s.trim_start().chars().next());
1962 ///
1963 /// let s = " עברית ";
1964 /// assert!(Some('ע') == s.trim_start().chars().next());
1965 /// ```
1966 #[inline]
1967 #[must_use = "this returns the trimmed string as a new slice, \
1968 without modifying the original"]
1969 #[stable(feature = "trim_direction", since = "1.30.0")]
1970 #[cfg_attr(not(test), rustc_diagnostic_item = "str_trim_start")]
1971 pub fn trim_start(&self) -> &str {
1972 self.trim_start_matches(|c: char| c.is_whitespace())
1973 }
1974
1975 /// Returns a string slice with trailing whitespace removed.
1976 ///
1977 /// 'Whitespace' is defined according to the terms of the Unicode Derived
1978 /// Core Property `White_Space`, which includes newlines.
1979 ///
1980 /// # Text directionality
1981 ///
1982 /// A string is a sequence of bytes. `end` in this context means the last
1983 /// position of that byte string; for a left-to-right language like English or
1984 /// Russian, this will be right side, and for right-to-left languages like
1985 /// Arabic or Hebrew, this will be the left side.
1986 ///
1987 /// # Examples
1988 ///
1989 /// Basic usage:
1990 ///
1991 /// ```
1992 /// let s = "\n Hello\tworld\t\n";
1993 /// assert_eq!("\n Hello\tworld", s.trim_end());
1994 /// ```
1995 ///
1996 /// Directionality:
1997 ///
1998 /// ```
1999 /// let s = " English ";
2000 /// assert!(Some('h') == s.trim_end().chars().rev().next());
2001 ///
2002 /// let s = " עברית ";
2003 /// assert!(Some('ת') == s.trim_end().chars().rev().next());
2004 /// ```
2005 #[inline]
2006 #[must_use = "this returns the trimmed string as a new slice, \
2007 without modifying the original"]
2008 #[stable(feature = "trim_direction", since = "1.30.0")]
2009 #[cfg_attr(not(test), rustc_diagnostic_item = "str_trim_end")]
2010 pub fn trim_end(&self) -> &str {
2011 self.trim_end_matches(|c: char| c.is_whitespace())
2012 }
2013
2014 /// Returns a string slice with leading whitespace removed.
2015 ///
2016 /// 'Whitespace' is defined according to the terms of the Unicode Derived
2017 /// Core Property `White_Space`.
2018 ///
2019 /// # Text directionality
2020 ///
2021 /// A string is a sequence of bytes. 'Left' in this context means the first
2022 /// position of that byte string; for a language like Arabic or Hebrew
2023 /// which are 'right to left' rather than 'left to right', this will be
2024 /// the _right_ side, not the left.
2025 ///
2026 /// # Examples
2027 ///
2028 /// Basic usage:
2029 ///
2030 /// ```
2031 /// let s = " Hello\tworld\t";
2032 ///
2033 /// assert_eq!("Hello\tworld\t", s.trim_left());
2034 /// ```
2035 ///
2036 /// Directionality:
2037 ///
2038 /// ```
2039 /// let s = " English";
2040 /// assert!(Some('E') == s.trim_left().chars().next());
2041 ///
2042 /// let s = " עברית";
2043 /// assert!(Some('ע') == s.trim_left().chars().next());
2044 /// ```
2045 #[must_use = "this returns the trimmed string as a new slice, \
2046 without modifying the original"]
2047 #[inline]
2048 #[stable(feature = "rust1", since = "1.0.0")]
2049 #[deprecated(since = "1.33.0", note = "superseded by `trim_start`", suggestion = "trim_start")]
2050 pub fn trim_left(&self) -> &str {
2051 self.trim_start()
2052 }
2053
2054 /// Returns a string slice with trailing whitespace removed.
2055 ///
2056 /// 'Whitespace' is defined according to the terms of the Unicode Derived
2057 /// Core Property `White_Space`.
2058 ///
2059 /// # Text directionality
2060 ///
2061 /// A string is a sequence of bytes. 'Right' in this context means the last
2062 /// position of that byte string; for a language like Arabic or Hebrew
2063 /// which are 'right to left' rather than 'left to right', this will be
2064 /// the _left_ side, not the right.
2065 ///
2066 /// # Examples
2067 ///
2068 /// Basic usage:
2069 ///
2070 /// ```
2071 /// let s = " Hello\tworld\t";
2072 ///
2073 /// assert_eq!(" Hello\tworld", s.trim_right());
2074 /// ```
2075 ///
2076 /// Directionality:
2077 ///
2078 /// ```
2079 /// let s = "English ";
2080 /// assert!(Some('h') == s.trim_right().chars().rev().next());
2081 ///
2082 /// let s = "עברית ";
2083 /// assert!(Some('ת') == s.trim_right().chars().rev().next());
2084 /// ```
2085 #[must_use = "this returns the trimmed string as a new slice, \
2086 without modifying the original"]
2087 #[inline]
2088 #[stable(feature = "rust1", since = "1.0.0")]
2089 #[deprecated(since = "1.33.0", note = "superseded by `trim_end`", suggestion = "trim_end")]
2090 pub fn trim_right(&self) -> &str {
2091 self.trim_end()
2092 }
2093
2094 /// Returns a string slice with all prefixes and suffixes that match a
2095 /// pattern repeatedly removed.
2096 ///
2097 /// The [pattern] can be a [`char`], a slice of [`char`]s, or a function
2098 /// or closure that determines if a character matches.
2099 ///
2100 /// [`char`]: prim@char
2101 /// [pattern]: self::pattern
2102 ///
2103 /// # Examples
2104 ///
2105 /// Simple patterns:
2106 ///
2107 /// ```
2108 /// assert_eq!("11foo1bar11".trim_matches('1'), "foo1bar");
2109 /// assert_eq!("123foo1bar123".trim_matches(char::is_numeric), "foo1bar");
2110 ///
2111 /// let x: &[_] = &['1', '2'];
2112 /// assert_eq!("12foo1bar12".trim_matches(x), "foo1bar");
2113 /// ```
2114 ///
2115 /// A more complex pattern, using a closure:
2116 ///
2117 /// ```
2118 /// assert_eq!("1foo1barXX".trim_matches(|c| c == '1' || c == 'X'), "foo1bar");
2119 /// ```
2120 #[must_use = "this returns the trimmed string as a new slice, \
2121 without modifying the original"]
2122 #[stable(feature = "rust1", since = "1.0.0")]
2123 pub fn trim_matches<'a, P>(&'a self, pat: P) -> &'a str
2124 where
2125 P: Pattern<'a, Searcher: DoubleEndedSearcher<'a>>,
2126 {
2127 let mut i = 0;
2128 let mut j = 0;
2129 let mut matcher = pat.into_searcher(self);
2130 if let Some((a, b)) = matcher.next_reject() {
2131 i = a;
2132 j = b; // Remember earliest known match, correct it below if
2133 // last match is different
2134 }
2135 if let Some((_, b)) = matcher.next_reject_back() {
2136 j = b;
2137 }
2138 // SAFETY: `Searcher` is known to return valid indices.
2139 unsafe { self.get_unchecked(i..j) }
2140 }
2141
2142 /// Returns a string slice with all prefixes that match a pattern
2143 /// repeatedly removed.
2144 ///
2145 /// The [pattern] can be a `&str`, [`char`], a slice of [`char`]s, or a
2146 /// function or closure that determines if a character matches.
2147 ///
2148 /// [`char`]: prim@char
2149 /// [pattern]: self::pattern
2150 ///
2151 /// # Text directionality
2152 ///
2153 /// A string is a sequence of bytes. `start` in this context means the first
2154 /// position of that byte string; for a left-to-right language like English or
2155 /// Russian, this will be left side, and for right-to-left languages like
2156 /// Arabic or Hebrew, this will be the right side.
2157 ///
2158 /// # Examples
2159 ///
2160 /// ```
2161 /// assert_eq!("11foo1bar11".trim_start_matches('1'), "foo1bar11");
2162 /// assert_eq!("123foo1bar123".trim_start_matches(char::is_numeric), "foo1bar123");
2163 ///
2164 /// let x: &[_] = &['1', '2'];
2165 /// assert_eq!("12foo1bar12".trim_start_matches(x), "foo1bar12");
2166 /// ```
2167 #[must_use = "this returns the trimmed string as a new slice, \
2168 without modifying the original"]
2169 #[stable(feature = "trim_direction", since = "1.30.0")]
2170 pub fn trim_start_matches<'a, P: Pattern<'a>>(&'a self, pat: P) -> &'a str {
2171 let mut i = self.len();
2172 let mut matcher = pat.into_searcher(self);
2173 if let Some((a, _)) = matcher.next_reject() {
2174 i = a;
2175 }
2176 // SAFETY: `Searcher` is known to return valid indices.
2177 unsafe { self.get_unchecked(i..self.len()) }
2178 }
2179
2180 /// Returns a string slice with the prefix removed.
2181 ///
2182 /// If the string starts with the pattern `prefix`, returns substring after the prefix, wrapped
2183 /// in `Some`. Unlike `trim_start_matches`, this method removes the prefix exactly once.
2184 ///
2185 /// If the string does not start with `prefix`, returns `None`.
2186 ///
2187 /// The [pattern] can be a `&str`, [`char`], a slice of [`char`]s, or a
2188 /// function or closure that determines if a character matches.
2189 ///
2190 /// [`char`]: prim@char
2191 /// [pattern]: self::pattern
2192 ///
2193 /// # Examples
2194 ///
2195 /// ```
2196 /// assert_eq!("foo:bar".strip_prefix("foo:"), Some("bar"));
2197 /// assert_eq!("foo:bar".strip_prefix("bar"), None);
2198 /// assert_eq!("foofoo".strip_prefix("foo"), Some("foo"));
2199 /// ```
2200 #[must_use = "this returns the remaining substring as a new slice, \
2201 without modifying the original"]
2202 #[stable(feature = "str_strip", since = "1.45.0")]
2203 pub fn strip_prefix<'a, P: Pattern<'a>>(&'a self, prefix: P) -> Option<&'a str> {
2204 prefix.strip_prefix_of(self)
2205 }
2206
2207 /// Returns a string slice with the suffix removed.
2208 ///
2209 /// If the string ends with the pattern `suffix`, returns the substring before the suffix,
2210 /// wrapped in `Some`. Unlike `trim_end_matches`, this method removes the suffix exactly once.
2211 ///
2212 /// If the string does not end with `suffix`, returns `None`.
2213 ///
2214 /// The [pattern] can be a `&str`, [`char`], a slice of [`char`]s, or a
2215 /// function or closure that determines if a character matches.
2216 ///
2217 /// [`char`]: prim@char
2218 /// [pattern]: self::pattern
2219 ///
2220 /// # Examples
2221 ///
2222 /// ```
2223 /// assert_eq!("bar:foo".strip_suffix(":foo"), Some("bar"));
2224 /// assert_eq!("bar:foo".strip_suffix("bar"), None);
2225 /// assert_eq!("foofoo".strip_suffix("foo"), Some("foo"));
2226 /// ```
2227 #[must_use = "this returns the remaining substring as a new slice, \
2228 without modifying the original"]
2229 #[stable(feature = "str_strip", since = "1.45.0")]
2230 pub fn strip_suffix<'a, P>(&'a self, suffix: P) -> Option<&'a str>
2231 where
2232 P: Pattern<'a>,
2233 <P as Pattern<'a>>::Searcher: ReverseSearcher<'a>,
2234 {
2235 suffix.strip_suffix_of(self)
2236 }
2237
2238 /// Returns a string slice with all suffixes that match a pattern
2239 /// repeatedly removed.
2240 ///
2241 /// The [pattern] can be a `&str`, [`char`], a slice of [`char`]s, or a
2242 /// function or closure that determines if a character matches.
2243 ///
2244 /// [`char`]: prim@char
2245 /// [pattern]: self::pattern
2246 ///
2247 /// # Text directionality
2248 ///
2249 /// A string is a sequence of bytes. `end` in this context means the last
2250 /// position of that byte string; for a left-to-right language like English or
2251 /// Russian, this will be right side, and for right-to-left languages like
2252 /// Arabic or Hebrew, this will be the left side.
2253 ///
2254 /// # Examples
2255 ///
2256 /// Simple patterns:
2257 ///
2258 /// ```
2259 /// assert_eq!("11foo1bar11".trim_end_matches('1'), "11foo1bar");
2260 /// assert_eq!("123foo1bar123".trim_end_matches(char::is_numeric), "123foo1bar");
2261 ///
2262 /// let x: &[_] = &['1', '2'];
2263 /// assert_eq!("12foo1bar12".trim_end_matches(x), "12foo1bar");
2264 /// ```
2265 ///
2266 /// A more complex pattern, using a closure:
2267 ///
2268 /// ```
2269 /// assert_eq!("1fooX".trim_end_matches(|c| c == '1' || c == 'X'), "1foo");
2270 /// ```
2271 #[must_use = "this returns the trimmed string as a new slice, \
2272 without modifying the original"]
2273 #[stable(feature = "trim_direction", since = "1.30.0")]
2274 pub fn trim_end_matches<'a, P>(&'a self, pat: P) -> &'a str
2275 where
2276 P: Pattern<'a, Searcher: ReverseSearcher<'a>>,
2277 {
2278 let mut j = 0;
2279 let mut matcher = pat.into_searcher(self);
2280 if let Some((_, b)) = matcher.next_reject_back() {
2281 j = b;
2282 }
2283 // SAFETY: `Searcher` is known to return valid indices.
2284 unsafe { self.get_unchecked(0..j) }
2285 }
2286
2287 /// Returns a string slice with all prefixes that match a pattern
2288 /// repeatedly removed.
2289 ///
2290 /// The [pattern] can be a `&str`, [`char`], a slice of [`char`]s, or a
2291 /// function or closure that determines if a character matches.
2292 ///
2293 /// [`char`]: prim@char
2294 /// [pattern]: self::pattern
2295 ///
2296 /// # Text directionality
2297 ///
2298 /// A string is a sequence of bytes. 'Left' in this context means the first
2299 /// position of that byte string; for a language like Arabic or Hebrew
2300 /// which are 'right to left' rather than 'left to right', this will be
2301 /// the _right_ side, not the left.
2302 ///
2303 /// # Examples
2304 ///
2305 /// ```
2306 /// assert_eq!("11foo1bar11".trim_left_matches('1'), "foo1bar11");
2307 /// assert_eq!("123foo1bar123".trim_left_matches(char::is_numeric), "foo1bar123");
2308 ///
2309 /// let x: &[_] = &['1', '2'];
2310 /// assert_eq!("12foo1bar12".trim_left_matches(x), "foo1bar12");
2311 /// ```
2312 #[stable(feature = "rust1", since = "1.0.0")]
2313 #[deprecated(
2314 since = "1.33.0",
2315 note = "superseded by `trim_start_matches`",
2316 suggestion = "trim_start_matches"
2317 )]
2318 pub fn trim_left_matches<'a, P: Pattern<'a>>(&'a self, pat: P) -> &'a str {
2319 self.trim_start_matches(pat)
2320 }
2321
2322 /// Returns a string slice with all suffixes that match a pattern
2323 /// repeatedly removed.
2324 ///
2325 /// The [pattern] can be a `&str`, [`char`], a slice of [`char`]s, or a
2326 /// function or closure that determines if a character matches.
2327 ///
2328 /// [`char`]: prim@char
2329 /// [pattern]: self::pattern
2330 ///
2331 /// # Text directionality
2332 ///
2333 /// A string is a sequence of bytes. 'Right' in this context means the last
2334 /// position of that byte string; for a language like Arabic or Hebrew
2335 /// which are 'right to left' rather than 'left to right', this will be
2336 /// the _left_ side, not the right.
2337 ///
2338 /// # Examples
2339 ///
2340 /// Simple patterns:
2341 ///
2342 /// ```
2343 /// assert_eq!("11foo1bar11".trim_right_matches('1'), "11foo1bar");
2344 /// assert_eq!("123foo1bar123".trim_right_matches(char::is_numeric), "123foo1bar");
2345 ///
2346 /// let x: &[_] = &['1', '2'];
2347 /// assert_eq!("12foo1bar12".trim_right_matches(x), "12foo1bar");
2348 /// ```
2349 ///
2350 /// A more complex pattern, using a closure:
2351 ///
2352 /// ```
2353 /// assert_eq!("1fooX".trim_right_matches(|c| c == '1' || c == 'X'), "1foo");
2354 /// ```
2355 #[stable(feature = "rust1", since = "1.0.0")]
2356 #[deprecated(
2357 since = "1.33.0",
2358 note = "superseded by `trim_end_matches`",
2359 suggestion = "trim_end_matches"
2360 )]
2361 pub fn trim_right_matches<'a, P>(&'a self, pat: P) -> &'a str
2362 where
2363 P: Pattern<'a, Searcher: ReverseSearcher<'a>>,
2364 {
2365 self.trim_end_matches(pat)
2366 }
2367
2368 /// Parses this string slice into another type.
2369 ///
2370 /// Because `parse` is so general, it can cause problems with type
2371 /// inference. As such, `parse` is one of the few times you'll see
2372 /// the syntax affectionately known as the 'turbofish': `::<>`. This
2373 /// helps the inference algorithm understand specifically which type
2374 /// you're trying to parse into.
2375 ///
2376 /// `parse` can parse into any type that implements the [`FromStr`] trait.
2377
2378 ///
2379 /// # Errors
2380 ///
2381 /// Will return [`Err`] if it's not possible to parse this string slice into
2382 /// the desired type.
2383 ///
2384 /// [`Err`]: FromStr::Err
2385 ///
2386 /// # Examples
2387 ///
2388 /// Basic usage
2389 ///
2390 /// ```
2391 /// let four: u32 = "4".parse().unwrap();
2392 ///
2393 /// assert_eq!(4, four);
2394 /// ```
2395 ///
2396 /// Using the 'turbofish' instead of annotating `four`:
2397 ///
2398 /// ```
2399 /// let four = "4".parse::<u32>();
2400 ///
2401 /// assert_eq!(Ok(4), four);
2402 /// ```
2403 ///
2404 /// Failing to parse:
2405 ///
2406 /// ```
2407 /// let nope = "j".parse::<u32>();
2408 ///
2409 /// assert!(nope.is_err());
2410 /// ```
2411 #[inline]
2412 #[stable(feature = "rust1", since = "1.0.0")]
2413 pub fn parse<F: FromStr>(&self) -> Result<F, F::Err> {
2414 FromStr::from_str(self)
2415 }
2416
2417 /// Checks if all characters in this string are within the ASCII range.
2418 ///
2419 /// # Examples
2420 ///
2421 /// ```
2422 /// let ascii = "hello!\n";
2423 /// let non_ascii = "Grüße, Jürgen ❤";
2424 ///
2425 /// assert!(ascii.is_ascii());
2426 /// assert!(!non_ascii.is_ascii());
2427 /// ```
2428 #[stable(feature = "ascii_methods_on_intrinsics", since = "1.23.0")]
2429 #[rustc_const_stable(feature = "const_slice_is_ascii", since = "1.74.0")]
2430 #[must_use]
2431 #[inline]
2432 pub const fn is_ascii(&self) -> bool {
2433 // We can treat each byte as character here: all multibyte characters
2434 // start with a byte that is not in the ASCII range, so we will stop
2435 // there already.
2436 self.as_bytes().is_ascii()
2437 }
2438
2439 /// If this string slice [`is_ascii`](Self::is_ascii), returns it as a slice
2440 /// of [ASCII characters](`ascii::Char`), otherwise returns `None`.
2441 #[unstable(feature = "ascii_char", issue = "110998")]
2442 #[must_use]
2443 #[inline]
2444 pub const fn as_ascii(&self) -> Option<&[ascii::Char]> {
2445 // Like in `is_ascii`, we can work on the bytes directly.
2446 self.as_bytes().as_ascii()
2447 }
2448
2449 /// Checks that two strings are an ASCII case-insensitive match.
2450 ///
2451 /// Same as `to_ascii_lowercase(a) == to_ascii_lowercase(b)`,
2452 /// but without allocating and copying temporaries.
2453 ///
2454 /// # Examples
2455 ///
2456 /// ```
2457 /// assert!("Ferris".eq_ignore_ascii_case("FERRIS"));
2458 /// assert!("Ferrös".eq_ignore_ascii_case("FERRöS"));
2459 /// assert!(!"Ferrös".eq_ignore_ascii_case("FERRÖS"));
2460 /// ```
2461 #[stable(feature = "ascii_methods_on_intrinsics", since = "1.23.0")]
2462 #[must_use]
2463 #[inline]
2464 pub fn eq_ignore_ascii_case(&self, other: &str) -> bool {
2465 self.as_bytes().eq_ignore_ascii_case(other.as_bytes())
2466 }
2467
2468 /// Converts this string to its ASCII upper case equivalent in-place.
2469 ///
2470 /// ASCII letters 'a' to 'z' are mapped to 'A' to 'Z',
2471 /// but non-ASCII letters are unchanged.
2472 ///
2473 /// To return a new uppercased value without modifying the existing one, use
2474 /// [`to_ascii_uppercase()`].
2475 ///
2476 /// [`to_ascii_uppercase()`]: #method.to_ascii_uppercase
2477 ///
2478 /// # Examples
2479 ///
2480 /// ```
2481 /// let mut s = String::from("Grüße, Jürgen ❤");
2482 ///
2483 /// s.make_ascii_uppercase();
2484 ///
2485 /// assert_eq!("GRüßE, JüRGEN ❤", s);
2486 /// ```
2487 #[stable(feature = "ascii_methods_on_intrinsics", since = "1.23.0")]
2488 #[inline]
2489 pub fn make_ascii_uppercase(&mut self) {
2490 // SAFETY: changing ASCII letters only does not invalidate UTF-8.
2491 let me = unsafe { self.as_bytes_mut() };
2492 me.make_ascii_uppercase()
2493 }
2494
2495 /// Converts this string to its ASCII lower case equivalent in-place.
2496 ///
2497 /// ASCII letters 'A' to 'Z' are mapped to 'a' to 'z',
2498 /// but non-ASCII letters are unchanged.
2499 ///
2500 /// To return a new lowercased value without modifying the existing one, use
2501 /// [`to_ascii_lowercase()`].
2502 ///
2503 /// [`to_ascii_lowercase()`]: #method.to_ascii_lowercase
2504 ///
2505 /// # Examples
2506 ///
2507 /// ```
2508 /// let mut s = String::from("GRÜßE, JÜRGEN ❤");
2509 ///
2510 /// s.make_ascii_lowercase();
2511 ///
2512 /// assert_eq!("grÜße, jÜrgen ❤", s);
2513 /// ```
2514 #[stable(feature = "ascii_methods_on_intrinsics", since = "1.23.0")]
2515 #[inline]
2516 pub fn make_ascii_lowercase(&mut self) {
2517 // SAFETY: changing ASCII letters only does not invalidate UTF-8.
2518 let me = unsafe { self.as_bytes_mut() };
2519 me.make_ascii_lowercase()
2520 }
2521
2522 /// Returns a string slice with leading ASCII whitespace removed.
2523 ///
2524 /// 'Whitespace' refers to the definition used by
2525 /// [`u8::is_ascii_whitespace`].
2526 ///
2527 /// [`u8::is_ascii_whitespace`]: u8::is_ascii_whitespace
2528 ///
2529 /// # Examples
2530 ///
2531 /// ```
2532 /// #![feature(byte_slice_trim_ascii)]
2533 ///
2534 /// assert_eq!(" \t \u{3000}hello world\n".trim_ascii_start(), "\u{3000}hello world\n");
2535 /// assert_eq!(" ".trim_ascii_start(), "");
2536 /// assert_eq!("".trim_ascii_start(), "");
2537 /// ```
2538 #[unstable(feature = "byte_slice_trim_ascii", issue = "94035")]
2539 #[must_use = "this returns the trimmed string as a new slice, \
2540 without modifying the original"]
2541 #[inline]
2542 pub const fn trim_ascii_start(&self) -> &str {
2543 // SAFETY: Removing ASCII characters from a `&str` does not invalidate
2544 // UTF-8.
2545 unsafe { core::str::from_utf8_unchecked(self.as_bytes().trim_ascii_start()) }
2546 }
2547
2548 /// Returns a string slice with trailing ASCII whitespace removed.
2549 ///
2550 /// 'Whitespace' refers to the definition used by
2551 /// [`u8::is_ascii_whitespace`].
2552 ///
2553 /// [`u8::is_ascii_whitespace`]: u8::is_ascii_whitespace
2554 ///
2555 /// # Examples
2556 ///
2557 /// ```
2558 /// #![feature(byte_slice_trim_ascii)]
2559 ///
2560 /// assert_eq!("\r hello world\u{3000}\n ".trim_ascii_end(), "\r hello world\u{3000}");
2561 /// assert_eq!(" ".trim_ascii_end(), "");
2562 /// assert_eq!("".trim_ascii_end(), "");
2563 /// ```
2564 #[unstable(feature = "byte_slice_trim_ascii", issue = "94035")]
2565 #[must_use = "this returns the trimmed string as a new slice, \
2566 without modifying the original"]
2567 #[inline]
2568 pub const fn trim_ascii_end(&self) -> &str {
2569 // SAFETY: Removing ASCII characters from a `&str` does not invalidate
2570 // UTF-8.
2571 unsafe { core::str::from_utf8_unchecked(self.as_bytes().trim_ascii_end()) }
2572 }
2573
2574 /// Returns a string slice with leading and trailing ASCII whitespace
2575 /// removed.
2576 ///
2577 /// 'Whitespace' refers to the definition used by
2578 /// [`u8::is_ascii_whitespace`].
2579 ///
2580 /// [`u8::is_ascii_whitespace`]: u8::is_ascii_whitespace
2581 ///
2582 /// # Examples
2583 ///
2584 /// ```
2585 /// #![feature(byte_slice_trim_ascii)]
2586 ///
2587 /// assert_eq!("\r hello world\n ".trim_ascii(), "hello world");
2588 /// assert_eq!(" ".trim_ascii(), "");
2589 /// assert_eq!("".trim_ascii(), "");
2590 /// ```
2591 #[unstable(feature = "byte_slice_trim_ascii", issue = "94035")]
2592 #[must_use = "this returns the trimmed string as a new slice, \
2593 without modifying the original"]
2594 #[inline]
2595 pub const fn trim_ascii(&self) -> &str {
2596 // SAFETY: Removing ASCII characters from a `&str` does not invalidate
2597 // UTF-8.
2598 unsafe { core::str::from_utf8_unchecked(self.as_bytes().trim_ascii()) }
2599 }
2600
2601 /// Return an iterator that escapes each char in `self` with [`char::escape_debug`].
2602 ///
2603 /// Note: only extended grapheme codepoints that begin the string will be
2604 /// escaped.
2605 ///
2606 /// # Examples
2607 ///
2608 /// As an iterator:
2609 ///
2610 /// ```
2611 /// for c in "❤\n!".escape_debug() {
2612 /// print!("{c}");
2613 /// }
2614 /// println!();
2615 /// ```
2616 ///
2617 /// Using `println!` directly:
2618 ///
2619 /// ```
2620 /// println!("{}", "❤\n!".escape_debug());
2621 /// ```
2622 ///
2623 ///
2624 /// Both are equivalent to:
2625 ///
2626 /// ```
2627 /// println!("❤\\n!");
2628 /// ```
2629 ///
2630 /// Using `to_string`:
2631 ///
2632 /// ```
2633 /// assert_eq!("❤\n!".escape_debug().to_string(), "❤\\n!");
2634 /// ```
2635 #[must_use = "this returns the escaped string as an iterator, \
2636 without modifying the original"]
2637 #[stable(feature = "str_escape", since = "1.34.0")]
2638 pub fn escape_debug(&self) -> EscapeDebug<'_> {
2639 let mut chars = self.chars();
2640 EscapeDebug {
2641 inner: chars
2642 .next()
2643 .map(|first| first.escape_debug_ext(EscapeDebugExtArgs::ESCAPE_ALL))
2644 .into_iter()
2645 .flatten()
2646 .chain(chars.flat_map(CharEscapeDebugContinue)),
2647 }
2648 }
2649
2650 /// Return an iterator that escapes each char in `self` with [`char::escape_default`].
2651 ///
2652 /// # Examples
2653 ///
2654 /// As an iterator:
2655 ///
2656 /// ```
2657 /// for c in "❤\n!".escape_default() {
2658 /// print!("{c}");
2659 /// }
2660 /// println!();
2661 /// ```
2662 ///
2663 /// Using `println!` directly:
2664 ///
2665 /// ```
2666 /// println!("{}", "❤\n!".escape_default());
2667 /// ```
2668 ///
2669 ///
2670 /// Both are equivalent to:
2671 ///
2672 /// ```
2673 /// println!("\\u{{2764}}\\n!");
2674 /// ```
2675 ///
2676 /// Using `to_string`:
2677 ///
2678 /// ```
2679 /// assert_eq!("❤\n!".escape_default().to_string(), "\\u{2764}\\n!");
2680 /// ```
2681 #[must_use = "this returns the escaped string as an iterator, \
2682 without modifying the original"]
2683 #[stable(feature = "str_escape", since = "1.34.0")]
2684 pub fn escape_default(&self) -> EscapeDefault<'_> {
2685 EscapeDefault { inner: self.chars().flat_map(CharEscapeDefault) }
2686 }
2687
2688 /// Return an iterator that escapes each char in `self` with [`char::escape_unicode`].
2689 ///
2690 /// # Examples
2691 ///
2692 /// As an iterator:
2693 ///
2694 /// ```
2695 /// for c in "❤\n!".escape_unicode() {
2696 /// print!("{c}");
2697 /// }
2698 /// println!();
2699 /// ```
2700 ///
2701 /// Using `println!` directly:
2702 ///
2703 /// ```
2704 /// println!("{}", "❤\n!".escape_unicode());
2705 /// ```
2706 ///
2707 ///
2708 /// Both are equivalent to:
2709 ///
2710 /// ```
2711 /// println!("\\u{{2764}}\\u{{a}}\\u{{21}}");
2712 /// ```
2713 ///
2714 /// Using `to_string`:
2715 ///
2716 /// ```
2717 /// assert_eq!("❤\n!".escape_unicode().to_string(), "\\u{2764}\\u{a}\\u{21}");
2718 /// ```
2719 #[must_use = "this returns the escaped string as an iterator, \
2720 without modifying the original"]
2721 #[stable(feature = "str_escape", since = "1.34.0")]
2722 pub fn escape_unicode(&self) -> EscapeUnicode<'_> {
2723 EscapeUnicode { inner: self.chars().flat_map(CharEscapeUnicode) }
2724 }
2725}
2726
2727#[stable(feature = "rust1", since = "1.0.0")]
2728impl AsRef<[u8]> for str {
2729 #[inline]
2730 fn as_ref(&self) -> &[u8] {
2731 self.as_bytes()
2732 }
2733}
2734
2735#[stable(feature = "rust1", since = "1.0.0")]
2736impl Default for &str {
2737 /// Creates an empty str
2738 #[inline]
2739 fn default() -> Self {
2740 ""
2741 }
2742}
2743
2744#[stable(feature = "default_mut_str", since = "1.28.0")]
2745impl Default for &mut str {
2746 /// Creates an empty mutable str
2747 #[inline]
2748 fn default() -> Self {
2749 // SAFETY: The empty string is valid UTF-8.
2750 unsafe { from_utf8_unchecked_mut(&mut []) }
2751 }
2752}
2753
2754impl_fn_for_zst! {
2755 /// A nameable, cloneable fn type
2756 #[derive(Clone)]
2757 struct LinesMap impl<'a> Fn = |line: &'a str| -> &'a str {
2758 let Some(line) = line.strip_suffix('\n') else { return line };
2759 let Some(line) = line.strip_suffix('\r') else { return line };
2760 line
2761 };
2762
2763 #[derive(Clone)]
2764 struct CharEscapeDebugContinue impl Fn = |c: char| -> char::EscapeDebug {
2765 c.escape_debug_ext(EscapeDebugExtArgs {
2766 escape_grapheme_extended: false,
2767 escape_single_quote: true,
2768 escape_double_quote: true
2769 })
2770 };
2771
2772 #[derive(Clone)]
2773 struct CharEscapeUnicode impl Fn = |c: char| -> char::EscapeUnicode {
2774 c.escape_unicode()
2775 };
2776 #[derive(Clone)]
2777 struct CharEscapeDefault impl Fn = |c: char| -> char::EscapeDefault {
2778 c.escape_default()
2779 };
2780
2781 #[derive(Clone)]
2782 struct IsWhitespace impl Fn = |c: char| -> bool {
2783 c.is_whitespace()
2784 };
2785
2786 #[derive(Clone)]
2787 struct IsAsciiWhitespace impl Fn = |byte: &u8| -> bool {
2788 byte.is_ascii_whitespace()
2789 };
2790
2791 #[derive(Clone)]
2792 struct IsNotEmpty impl<'a, 'b> Fn = |s: &'a &'b str| -> bool {
2793 !s.is_empty()
2794 };
2795
2796 #[derive(Clone)]
2797 struct BytesIsNotEmpty impl<'a, 'b> Fn = |s: &'a &'b [u8]| -> bool {
2798 !s.is_empty()
2799 };
2800
2801 #[derive(Clone)]
2802 struct UnsafeBytesToStr impl<'a> Fn = |bytes: &'a [u8]| -> &'a str {
2803 // SAFETY: not safe
2804 unsafe { from_utf8_unchecked(bytes) }
2805 };
2806}
2807
2808// This is required to make `impl From<&str> for Box<dyn Error>` and `impl<E> From<E> for Box<dyn Error>` not overlap.
2809#[stable(feature = "rust1", since = "1.0.0")]
2810impl !crate::error::Error for &str {}
2811