1 // ignore-tidy-filelength
2 // ignore-tidy-undocumented-unsafe
4 //! String manipulation.
6 //! For more details, see the `std::str` module.
8 #![stable(feature = "rust1", since = "1.0.0")]
10 use self::pattern::Pattern;
11 use self::pattern::{Searcher, SearchStep, ReverseSearcher, DoubleEndedSearcher};
14 use crate::fmt::{self, Write};
15 use crate::iter::{Map, Cloned, FusedIterator, TrustedLen, TrustedRandomAccess, Filter};
16 use crate::iter::{Flatten, FlatMap, Chain};
17 use crate::slice::{self, SliceIndex, Split as SliceSplit};
24 #[unstable(feature = "str_internals", issue = "0")]
25 #[allow(missing_docs)]
28 /// Parse a value from a string
30 /// `FromStr`'s [`from_str`] method is often used implicitly, through
31 /// [`str`]'s [`parse`] method. See [`parse`]'s documentation for examples.
33 /// [`from_str`]: #tymethod.from_str
34 /// [`str`]: ../../std/primitive.str.html
35 /// [`parse`]: ../../std/primitive.str.html#method.parse
37 /// `FromStr` does not have a lifetime parameter, and so you can only parse types
38 /// that do not contain a lifetime parameter themselves. In other words, you can
39 /// parse an `i32` with `FromStr`, but not a `&i32`. You can parse a struct that
40 /// contains an `i32`, but not one that contains an `&i32`.
44 /// Basic implementation of `FromStr` on an example `Point` type:
47 /// use std::str::FromStr;
48 /// use std::num::ParseIntError;
50 /// #[derive(Debug, PartialEq)]
56 /// impl FromStr for Point {
57 /// type Err = ParseIntError;
59 /// fn from_str(s: &str) -> Result<Self, Self::Err> {
60 /// let coords: Vec<&str> = s.trim_matches(|p| p == '(' || p == ')' )
64 /// let x_fromstr = coords[0].parse::<i32>()?;
65 /// let y_fromstr = coords[1].parse::<i32>()?;
67 /// Ok(Point { x: x_fromstr, y: y_fromstr })
71 /// let p = Point::from_str("(1,2)");
72 /// assert_eq!(p.unwrap(), Point{ x: 1, y: 2} )
74 #[stable(feature = "rust1", since = "1.0.0")]
75 pub trait FromStr: Sized {
76 /// The associated error which can be returned from parsing.
77 #[stable(feature = "rust1", since = "1.0.0")]
80 /// Parses a string `s` to return a value of this type.
82 /// If parsing succeeds, return the value inside [`Ok`], otherwise
83 /// when the string is ill-formatted return an error specific to the
84 /// inside [`Err`]. The error type is specific to implementation of the trait.
86 /// [`Ok`]: ../../std/result/enum.Result.html#variant.Ok
87 /// [`Err`]: ../../std/result/enum.Result.html#variant.Err
91 /// Basic usage with [`i32`][ithirtytwo], a type that implements `FromStr`:
93 /// [ithirtytwo]: ../../std/primitive.i32.html
96 /// use std::str::FromStr;
99 /// let x = i32::from_str(s).unwrap();
101 /// assert_eq!(5, x);
103 #[stable(feature = "rust1", since = "1.0.0")]
104 fn from_str(s: &str) -> Result<Self, Self::Err>;
107 #[stable(feature = "rust1", since = "1.0.0")]
108 impl FromStr for bool {
109 type Err = ParseBoolError;
111 /// Parse a `bool` from a string.
113 /// Yields a `Result<bool, ParseBoolError>`, because `s` may or may not
114 /// actually be parseable.
119 /// use std::str::FromStr;
121 /// assert_eq!(FromStr::from_str("true"), Ok(true));
122 /// assert_eq!(FromStr::from_str("false"), Ok(false));
123 /// assert!(<bool as FromStr>::from_str("not even a boolean").is_err());
126 /// Note, in many cases, the `.parse()` method on `str` is more proper.
129 /// assert_eq!("true".parse(), Ok(true));
130 /// assert_eq!("false".parse(), Ok(false));
131 /// assert!("not even a boolean".parse::<bool>().is_err());
134 fn from_str(s: &str) -> Result<bool, ParseBoolError> {
137 "false" => Ok(false),
138 _ => Err(ParseBoolError { _priv: () }),
143 /// An error returned when parsing a `bool` using [`from_str`] fails
145 /// [`from_str`]: ../../std/primitive.bool.html#method.from_str
146 #[derive(Debug, Clone, PartialEq, Eq)]
147 #[stable(feature = "rust1", since = "1.0.0")]
148 pub struct ParseBoolError { _priv: () }
150 #[stable(feature = "rust1", since = "1.0.0")]
151 impl fmt::Display for ParseBoolError {
152 fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
153 "provided string was not `true` or `false`".fmt(f)
158 Section: Creating a string
161 /// Errors which can occur when attempting to interpret a sequence of [`u8`]
164 /// [`u8`]: ../../std/primitive.u8.html
166 /// As such, the `from_utf8` family of functions and methods for both [`String`]s
167 /// and [`&str`]s make use of this error, for example.
169 /// [`String`]: ../../std/string/struct.String.html#method.from_utf8
170 /// [`&str`]: ../../std/str/fn.from_utf8.html
174 /// This error type’s methods can be used to create functionality
175 /// similar to `String::from_utf8_lossy` without allocating heap memory:
178 /// fn from_utf8_lossy<F>(mut input: &[u8], mut push: F) where F: FnMut(&str) {
180 /// match std::str::from_utf8(input) {
186 /// let (valid, after_valid) = input.split_at(error.valid_up_to());
188 /// push(std::str::from_utf8_unchecked(valid))
190 /// push("\u{FFFD}");
192 /// if let Some(invalid_sequence_length) = error.error_len() {
193 /// input = &after_valid[invalid_sequence_length..]
202 #[derive(Copy, Eq, PartialEq, Clone, Debug)]
203 #[stable(feature = "rust1", since = "1.0.0")]
204 pub struct Utf8Error {
206 error_len: Option<u8>,
210 /// Returns the index in the given string up to which valid UTF-8 was
213 /// It is the maximum index such that `from_utf8(&input[..index])`
214 /// would return `Ok(_)`.
223 /// // some invalid bytes, in a vector
224 /// let sparkle_heart = vec![0, 159, 146, 150];
226 /// // std::str::from_utf8 returns a Utf8Error
227 /// let error = str::from_utf8(&sparkle_heart).unwrap_err();
229 /// // the second byte is invalid here
230 /// assert_eq!(1, error.valid_up_to());
232 #[stable(feature = "utf8_error", since = "1.5.0")]
233 pub fn valid_up_to(&self) -> usize { self.valid_up_to }
235 /// Provides more information about the failure:
237 /// * `None`: the end of the input was reached unexpectedly.
238 /// `self.valid_up_to()` is 1 to 3 bytes from the end of the input.
239 /// If a byte stream (such as a file or a network socket) is being decoded incrementally,
240 /// this could be a valid `char` whose UTF-8 byte sequence is spanning multiple chunks.
242 /// * `Some(len)`: an unexpected byte was encountered.
243 /// The length provided is that of the invalid byte sequence
244 /// that starts at the index given by `valid_up_to()`.
245 /// Decoding should resume after that sequence
246 /// (after inserting a [`U+FFFD REPLACEMENT CHARACTER`][U+FFFD]) in case of
249 /// [U+FFFD]: ../../std/char/constant.REPLACEMENT_CHARACTER.html
250 #[stable(feature = "utf8_error_error_len", since = "1.20.0")]
251 pub fn error_len(&self) -> Option<usize> {
252 self.error_len.map(|len| len as usize)
256 /// Converts a slice of bytes to a string slice.
258 /// A string slice ([`&str`]) is made of bytes ([`u8`]), and a byte slice
259 /// ([`&[u8]`][byteslice]) is made of bytes, so this function converts between
260 /// the two. Not all byte slices are valid string slices, however: [`&str`] requires
261 /// that it is valid UTF-8. `from_utf8()` checks to ensure that the bytes are valid
262 /// UTF-8, and then does the conversion.
264 /// [`&str`]: ../../std/primitive.str.html
265 /// [`u8`]: ../../std/primitive.u8.html
266 /// [byteslice]: ../../std/primitive.slice.html
268 /// If you are sure that the byte slice is valid UTF-8, and you don't want to
269 /// incur the overhead of the validity check, there is an unsafe version of
270 /// this function, [`from_utf8_unchecked`][fromutf8u], which has the same
271 /// behavior but skips the check.
273 /// [fromutf8u]: fn.from_utf8_unchecked.html
275 /// If you need a `String` instead of a `&str`, consider
276 /// [`String::from_utf8`][string].
278 /// [string]: ../../std/string/struct.String.html#method.from_utf8
280 /// Because you can stack-allocate a `[u8; N]`, and you can take a
281 /// [`&[u8]`][byteslice] of it, this function is one way to have a
282 /// stack-allocated string. There is an example of this in the
283 /// examples section below.
285 /// [byteslice]: ../../std/primitive.slice.html
289 /// Returns `Err` if the slice is not UTF-8 with a description as to why the
290 /// provided slice is not UTF-8.
299 /// // some bytes, in a vector
300 /// let sparkle_heart = vec![240, 159, 146, 150];
302 /// // We know these bytes are valid, so just use `unwrap()`.
303 /// let sparkle_heart = str::from_utf8(&sparkle_heart).unwrap();
305 /// assert_eq!("💖", sparkle_heart);
313 /// // some invalid bytes, in a vector
314 /// let sparkle_heart = vec![0, 159, 146, 150];
316 /// assert!(str::from_utf8(&sparkle_heart).is_err());
319 /// See the docs for [`Utf8Error`][error] for more details on the kinds of
320 /// errors that can be returned.
322 /// [error]: struct.Utf8Error.html
324 /// A "stack allocated string":
329 /// // some bytes, in a stack-allocated array
330 /// let sparkle_heart = [240, 159, 146, 150];
332 /// // We know these bytes are valid, so just use `unwrap()`.
333 /// let sparkle_heart = str::from_utf8(&sparkle_heart).unwrap();
335 /// assert_eq!("💖", sparkle_heart);
337 #[stable(feature = "rust1", since = "1.0.0")]
338 pub fn from_utf8(v: &[u8]) -> Result<&str, Utf8Error> {
339 run_utf8_validation(v)?;
340 Ok(unsafe { from_utf8_unchecked(v) })
343 /// Converts a mutable slice of bytes to a mutable string slice.
352 /// // "Hello, Rust!" as a mutable vector
353 /// let mut hellorust = vec![72, 101, 108, 108, 111, 44, 32, 82, 117, 115, 116, 33];
355 /// // As we know these bytes are valid, we can use `unwrap()`
356 /// let outstr = str::from_utf8_mut(&mut hellorust).unwrap();
358 /// assert_eq!("Hello, Rust!", outstr);
366 /// // Some invalid bytes in a mutable vector
367 /// let mut invalid = vec![128, 223];
369 /// assert!(str::from_utf8_mut(&mut invalid).is_err());
371 /// See the docs for [`Utf8Error`][error] for more details on the kinds of
372 /// errors that can be returned.
374 /// [error]: struct.Utf8Error.html
375 #[stable(feature = "str_mut_extras", since = "1.20.0")]
376 pub fn from_utf8_mut(v: &mut [u8]) -> Result<&mut str, Utf8Error> {
377 run_utf8_validation(v)?;
378 Ok(unsafe { from_utf8_unchecked_mut(v) })
381 /// Converts a slice of bytes to a string slice without checking
382 /// that the string contains valid UTF-8.
384 /// See the safe version, [`from_utf8`][fromutf8], for more information.
386 /// [fromutf8]: fn.from_utf8.html
390 /// This function is unsafe because it does not check that the bytes passed to
391 /// it are valid UTF-8. If this constraint is violated, undefined behavior
392 /// results, as the rest of Rust assumes that [`&str`]s are valid UTF-8.
394 /// [`&str`]: ../../std/primitive.str.html
403 /// // some bytes, in a vector
404 /// let sparkle_heart = vec![240, 159, 146, 150];
406 /// let sparkle_heart = unsafe {
407 /// str::from_utf8_unchecked(&sparkle_heart)
410 /// assert_eq!("💖", sparkle_heart);
413 #[stable(feature = "rust1", since = "1.0.0")]
414 pub unsafe fn from_utf8_unchecked(v: &[u8]) -> &str {
415 &*(v as *const [u8] as *const str)
418 /// Converts a slice of bytes to a string slice without checking
419 /// that the string contains valid UTF-8; mutable version.
421 /// See the immutable version, [`from_utf8_unchecked()`][fromutf8], for more information.
423 /// [fromutf8]: fn.from_utf8_unchecked.html
432 /// let mut heart = vec![240, 159, 146, 150];
433 /// let heart = unsafe { str::from_utf8_unchecked_mut(&mut heart) };
435 /// assert_eq!("💖", heart);
438 #[stable(feature = "str_mut_extras", since = "1.20.0")]
439 pub unsafe fn from_utf8_unchecked_mut(v: &mut [u8]) -> &mut str {
440 &mut *(v as *mut [u8] as *mut str)
443 #[stable(feature = "rust1", since = "1.0.0")]
444 impl fmt::Display for Utf8Error {
445 fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
446 if let Some(error_len) = self.error_len {
447 write!(f, "invalid utf-8 sequence of {} bytes from index {}",
448 error_len, self.valid_up_to)
450 write!(f, "incomplete utf-8 byte sequence from index {}", self.valid_up_to)
459 /// An iterator over the [`char`]s of a string slice.
461 /// [`char`]: ../../std/primitive.char.html
463 /// This struct is created by the [`chars`] method on [`str`].
464 /// See its documentation for more.
466 /// [`chars`]: ../../std/primitive.str.html#method.chars
467 /// [`str`]: ../../std/primitive.str.html
469 #[stable(feature = "rust1", since = "1.0.0")]
470 pub struct Chars<'a> {
471 iter: slice::Iter<'a, u8>
474 /// Returns the initial codepoint accumulator for the first byte.
475 /// The first byte is special, only want bottom 5 bits for width 2, 4 bits
476 /// for width 3, and 3 bits for width 4.
478 fn utf8_first_byte(byte: u8, width: u32) -> u32 { (byte & (0x7F >> width)) as u32 }
480 /// Returns the value of `ch` updated with continuation byte `byte`.
482 fn utf8_acc_cont_byte(ch: u32, byte: u8) -> u32 { (ch << 6) | (byte & CONT_MASK) as u32 }
484 /// Checks whether the byte is a UTF-8 continuation byte (i.e., starts with the
487 fn utf8_is_cont_byte(byte: u8) -> bool { (byte & !CONT_MASK) == TAG_CONT_U8 }
490 fn unwrap_or_0(opt: Option<&u8>) -> u8 {
497 /// Reads the next code point out of a byte iterator (assuming a
498 /// UTF-8-like encoding).
499 #[unstable(feature = "str_internals", issue = "0")]
501 pub fn next_code_point<'a, I: Iterator<Item = &'a u8>>(bytes: &mut I) -> Option<u32> {
503 let x = *bytes.next()?;
505 return Some(x as u32)
508 // Multibyte case follows
509 // Decode from a byte combination out of: [[[x y] z] w]
510 // NOTE: Performance is sensitive to the exact formulation here
511 let init = utf8_first_byte(x, 2);
512 let y = unwrap_or_0(bytes.next());
513 let mut ch = utf8_acc_cont_byte(init, y);
516 // 5th bit in 0xE0 .. 0xEF is always clear, so `init` is still valid
517 let z = unwrap_or_0(bytes.next());
518 let y_z = utf8_acc_cont_byte((y & CONT_MASK) as u32, z);
519 ch = init << 12 | y_z;
522 // use only the lower 3 bits of `init`
523 let w = unwrap_or_0(bytes.next());
524 ch = (init & 7) << 18 | utf8_acc_cont_byte(y_z, w);
531 /// Reads the last code point out of a byte iterator (assuming a
532 /// UTF-8-like encoding).
534 fn next_code_point_reverse<'a, I>(bytes: &mut I) -> Option<u32>
535 where I: DoubleEndedIterator<Item = &'a u8>,
538 let w = match *bytes.next_back()? {
539 next_byte if next_byte < 128 => return Some(next_byte as u32),
540 back_byte => back_byte,
543 // Multibyte case follows
544 // Decode from a byte combination out of: [x [y [z w]]]
546 let z = unwrap_or_0(bytes.next_back());
547 ch = utf8_first_byte(z, 2);
548 if utf8_is_cont_byte(z) {
549 let y = unwrap_or_0(bytes.next_back());
550 ch = utf8_first_byte(y, 3);
551 if utf8_is_cont_byte(y) {
552 let x = unwrap_or_0(bytes.next_back());
553 ch = utf8_first_byte(x, 4);
554 ch = utf8_acc_cont_byte(ch, y);
556 ch = utf8_acc_cont_byte(ch, z);
558 ch = utf8_acc_cont_byte(ch, w);
563 #[stable(feature = "rust1", since = "1.0.0")]
564 impl<'a> Iterator for Chars<'a> {
568 fn next(&mut self) -> Option<char> {
569 next_code_point(&mut self.iter).map(|ch| {
570 // str invariant says `ch` is a valid Unicode Scalar Value
572 char::from_u32_unchecked(ch)
578 fn count(self) -> usize {
579 // length in `char` is equal to the number of non-continuation bytes
580 let bytes_len = self.iter.len();
581 let mut cont_bytes = 0;
582 for &byte in self.iter {
583 cont_bytes += utf8_is_cont_byte(byte) as usize;
585 bytes_len - cont_bytes
589 fn size_hint(&self) -> (usize, Option<usize>) {
590 let len = self.iter.len();
591 // `(len + 3)` can't overflow, because we know that the `slice::Iter`
592 // belongs to a slice in memory which has a maximum length of
593 // `isize::MAX` (that's well below `usize::MAX`).
594 ((len + 3) / 4, Some(len))
598 fn last(mut self) -> Option<char> {
599 // No need to go through the entire string.
604 #[stable(feature = "chars_debug_impl", since = "1.38.0")]
605 impl fmt::Debug for Chars<'_> {
606 fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
607 write!(f, "Chars(")?;
608 f.debug_list().entries(self.clone()).finish()?;
614 #[stable(feature = "rust1", since = "1.0.0")]
615 impl<'a> DoubleEndedIterator for Chars<'a> {
617 fn next_back(&mut self) -> Option<char> {
618 next_code_point_reverse(&mut self.iter).map(|ch| {
619 // str invariant says `ch` is a valid Unicode Scalar Value
621 char::from_u32_unchecked(ch)
627 #[stable(feature = "fused", since = "1.26.0")]
628 impl FusedIterator for Chars<'_> {}
631 /// Views the underlying data as a subslice of the original data.
633 /// This has the same lifetime as the original slice, and so the
634 /// iterator can continue to be used while this exists.
639 /// let mut chars = "abc".chars();
641 /// assert_eq!(chars.as_str(), "abc");
643 /// assert_eq!(chars.as_str(), "bc");
646 /// assert_eq!(chars.as_str(), "");
648 #[stable(feature = "iter_to_slice", since = "1.4.0")]
650 pub fn as_str(&self) -> &'a str {
651 unsafe { from_utf8_unchecked(self.iter.as_slice()) }
655 /// An iterator over the [`char`]s of a string slice, and their positions.
657 /// [`char`]: ../../std/primitive.char.html
659 /// This struct is created by the [`char_indices`] method on [`str`].
660 /// See its documentation for more.
662 /// [`char_indices`]: ../../std/primitive.str.html#method.char_indices
663 /// [`str`]: ../../std/primitive.str.html
664 #[derive(Clone, Debug)]
665 #[stable(feature = "rust1", since = "1.0.0")]
666 pub struct CharIndices<'a> {
671 #[stable(feature = "rust1", since = "1.0.0")]
672 impl<'a> Iterator for CharIndices<'a> {
673 type Item = (usize, char);
676 fn next(&mut self) -> Option<(usize, char)> {
677 let pre_len = self.iter.iter.len();
678 match self.iter.next() {
681 let index = self.front_offset;
682 let len = self.iter.iter.len();
683 self.front_offset += pre_len - len;
690 fn count(self) -> usize {
695 fn size_hint(&self) -> (usize, Option<usize>) {
696 self.iter.size_hint()
700 fn last(mut self) -> Option<(usize, char)> {
701 // No need to go through the entire string.
706 #[stable(feature = "rust1", since = "1.0.0")]
707 impl<'a> DoubleEndedIterator for CharIndices<'a> {
709 fn next_back(&mut self) -> Option<(usize, char)> {
710 self.iter.next_back().map(|ch| {
711 let index = self.front_offset + self.iter.iter.len();
717 #[stable(feature = "fused", since = "1.26.0")]
718 impl FusedIterator for CharIndices<'_> {}
720 impl<'a> CharIndices<'a> {
721 /// Views the underlying data as a subslice of the original data.
723 /// This has the same lifetime as the original slice, and so the
724 /// iterator can continue to be used while this exists.
725 #[stable(feature = "iter_to_slice", since = "1.4.0")]
727 pub fn as_str(&self) -> &'a str {
732 /// An iterator over the bytes of a string slice.
734 /// This struct is created by the [`bytes`] method on [`str`].
735 /// See its documentation for more.
737 /// [`bytes`]: ../../std/primitive.str.html#method.bytes
738 /// [`str`]: ../../std/primitive.str.html
739 #[stable(feature = "rust1", since = "1.0.0")]
740 #[derive(Clone, Debug)]
741 pub struct Bytes<'a>(Cloned<slice::Iter<'a, u8>>);
743 #[stable(feature = "rust1", since = "1.0.0")]
744 impl Iterator for Bytes<'_> {
748 fn next(&mut self) -> Option<u8> {
753 fn size_hint(&self) -> (usize, Option<usize>) {
758 fn count(self) -> usize {
763 fn last(self) -> Option<Self::Item> {
768 fn nth(&mut self, n: usize) -> Option<Self::Item> {
773 fn all<F>(&mut self, f: F) -> bool where F: FnMut(Self::Item) -> bool {
778 fn any<F>(&mut self, f: F) -> bool where F: FnMut(Self::Item) -> bool {
783 fn find<P>(&mut self, predicate: P) -> Option<Self::Item> where
784 P: FnMut(&Self::Item) -> bool
786 self.0.find(predicate)
790 fn position<P>(&mut self, predicate: P) -> Option<usize> where
791 P: FnMut(Self::Item) -> bool
793 self.0.position(predicate)
797 fn rposition<P>(&mut self, predicate: P) -> Option<usize> where
798 P: FnMut(Self::Item) -> bool
800 self.0.rposition(predicate)
804 #[stable(feature = "rust1", since = "1.0.0")]
805 impl DoubleEndedIterator for Bytes<'_> {
807 fn next_back(&mut self) -> Option<u8> {
812 fn nth_back(&mut self, n: usize) -> Option<Self::Item> {
817 fn rfind<P>(&mut self, predicate: P) -> Option<Self::Item> where
818 P: FnMut(&Self::Item) -> bool
820 self.0.rfind(predicate)
824 #[stable(feature = "rust1", since = "1.0.0")]
825 impl ExactSizeIterator for Bytes<'_> {
827 fn len(&self) -> usize {
832 fn is_empty(&self) -> bool {
837 #[stable(feature = "fused", since = "1.26.0")]
838 impl FusedIterator for Bytes<'_> {}
840 #[unstable(feature = "trusted_len", issue = "37572")]
841 unsafe impl TrustedLen for Bytes<'_> {}
844 unsafe impl TrustedRandomAccess for Bytes<'_> {
845 unsafe fn get_unchecked(&mut self, i: usize) -> u8 {
846 self.0.get_unchecked(i)
848 fn may_have_side_effect() -> bool { false }
851 /// This macro generates a Clone impl for string pattern API
852 /// wrapper types of the form X<'a, P>
853 macro_rules! derive_pattern_clone {
854 (clone $t:ident with |$s:ident| $e:expr) => {
855 impl<'a, P> Clone for $t<'a, P>
857 P: Pattern<'a, Searcher: Clone>,
859 fn clone(&self) -> Self {
867 /// This macro generates two public iterator structs
868 /// wrapping a private internal one that makes use of the `Pattern` API.
870 /// For all patterns `P: Pattern<'a>` the following items will be
871 /// generated (generics omitted):
873 /// struct $forward_iterator($internal_iterator);
874 /// struct $reverse_iterator($internal_iterator);
876 /// impl Iterator for $forward_iterator
877 /// { /* internal ends up calling Searcher::next_match() */ }
879 /// impl DoubleEndedIterator for $forward_iterator
880 /// where P::Searcher: DoubleEndedSearcher
881 /// { /* internal ends up calling Searcher::next_match_back() */ }
883 /// impl Iterator for $reverse_iterator
884 /// where P::Searcher: ReverseSearcher
885 /// { /* internal ends up calling Searcher::next_match_back() */ }
887 /// impl DoubleEndedIterator for $reverse_iterator
888 /// where P::Searcher: DoubleEndedSearcher
889 /// { /* internal ends up calling Searcher::next_match() */ }
891 /// The internal one is defined outside the macro, and has almost the same
892 /// semantic as a DoubleEndedIterator by delegating to `pattern::Searcher` and
893 /// `pattern::ReverseSearcher` for both forward and reverse iteration.
895 /// "Almost", because a `Searcher` and a `ReverseSearcher` for a given
896 /// `Pattern` might not return the same elements, so actually implementing
897 /// `DoubleEndedIterator` for it would be incorrect.
898 /// (See the docs in `str::pattern` for more details)
900 /// However, the internal struct still represents a single ended iterator from
901 /// either end, and depending on pattern is also a valid double ended iterator,
902 /// so the two wrapper structs implement `Iterator`
903 /// and `DoubleEndedIterator` depending on the concrete pattern type, leading
904 /// to the complex impls seen above.
905 macro_rules! generate_pattern_iterators {
909 $(#[$forward_iterator_attribute:meta])*
910 struct $forward_iterator:ident;
914 $(#[$reverse_iterator_attribute:meta])*
915 struct $reverse_iterator:ident;
917 // Stability of all generated items
919 $(#[$common_stability_attribute:meta])*
921 // Internal almost-iterator that is being delegated to
923 $internal_iterator:ident yielding ($iterty:ty);
925 // Kind of delegation - either single ended or double ended
928 $(#[$forward_iterator_attribute])*
929 $(#[$common_stability_attribute])*
930 pub struct $forward_iterator<'a, P: Pattern<'a>>($internal_iterator<'a, P>);
932 $(#[$common_stability_attribute])*
933 impl<'a, P> fmt::Debug for $forward_iterator<'a, P>
935 P: Pattern<'a, Searcher: fmt::Debug>,
937 fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
938 f.debug_tuple(stringify!($forward_iterator))
944 $(#[$common_stability_attribute])*
945 impl<'a, P: Pattern<'a>> Iterator for $forward_iterator<'a, P> {
949 fn next(&mut self) -> Option<$iterty> {
954 $(#[$common_stability_attribute])*
955 impl<'a, P> Clone for $forward_iterator<'a, P>
957 P: Pattern<'a, Searcher: Clone>,
959 fn clone(&self) -> Self {
960 $forward_iterator(self.0.clone())
964 $(#[$reverse_iterator_attribute])*
965 $(#[$common_stability_attribute])*
966 pub struct $reverse_iterator<'a, P: Pattern<'a>>($internal_iterator<'a, P>);
968 $(#[$common_stability_attribute])*
969 impl<'a, P> fmt::Debug for $reverse_iterator<'a, P>
971 P: Pattern<'a, Searcher: fmt::Debug>,
973 fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
974 f.debug_tuple(stringify!($reverse_iterator))
980 $(#[$common_stability_attribute])*
981 impl<'a, P> Iterator for $reverse_iterator<'a, P>
983 P: Pattern<'a, Searcher: ReverseSearcher<'a>>,
988 fn next(&mut self) -> Option<$iterty> {
993 $(#[$common_stability_attribute])*
994 impl<'a, P> Clone for $reverse_iterator<'a, P>
996 P: Pattern<'a, Searcher: Clone>,
998 fn clone(&self) -> Self {
999 $reverse_iterator(self.0.clone())
1003 #[stable(feature = "fused", since = "1.26.0")]
1004 impl<'a, P: Pattern<'a>> FusedIterator for $forward_iterator<'a, P> {}
1006 #[stable(feature = "fused", since = "1.26.0")]
1007 impl<'a, P> FusedIterator for $reverse_iterator<'a, P>
1009 P: Pattern<'a, Searcher: ReverseSearcher<'a>>,
1012 generate_pattern_iterators!($($t)* with $(#[$common_stability_attribute])*,
1014 $reverse_iterator, $iterty);
1017 double ended; with $(#[$common_stability_attribute:meta])*,
1018 $forward_iterator:ident,
1019 $reverse_iterator:ident, $iterty:ty
1021 $(#[$common_stability_attribute])*
1022 impl<'a, P> DoubleEndedIterator for $forward_iterator<'a, P>
1024 P: Pattern<'a, Searcher: DoubleEndedSearcher<'a>>,
1027 fn next_back(&mut self) -> Option<$iterty> {
1032 $(#[$common_stability_attribute])*
1033 impl<'a, P> DoubleEndedIterator for $reverse_iterator<'a, P>
1035 P: Pattern<'a, Searcher: DoubleEndedSearcher<'a>>,
1038 fn next_back(&mut self) -> Option<$iterty> {
1044 single ended; with $(#[$common_stability_attribute:meta])*,
1045 $forward_iterator:ident,
1046 $reverse_iterator:ident, $iterty:ty
1050 derive_pattern_clone!{
1052 with |s| SplitInternal { matcher: s.matcher.clone(), ..*s }
1055 struct SplitInternal<'a, P: Pattern<'a>> {
1058 matcher: P::Searcher,
1059 allow_trailing_empty: bool,
1063 impl<'a, P> fmt::Debug for SplitInternal<'a, P>
1065 P: Pattern<'a, Searcher: fmt::Debug>,
1067 fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
1068 f.debug_struct("SplitInternal")
1069 .field("start", &self.start)
1070 .field("end", &self.end)
1071 .field("matcher", &self.matcher)
1072 .field("allow_trailing_empty", &self.allow_trailing_empty)
1073 .field("finished", &self.finished)
1078 impl<'a, P: Pattern<'a>> SplitInternal<'a, P> {
1080 fn get_end(&mut self) -> Option<&'a str> {
1081 if !self.finished && (self.allow_trailing_empty || self.end - self.start > 0) {
1082 self.finished = true;
1084 let string = self.matcher.haystack().get_unchecked(self.start..self.end);
1093 fn next(&mut self) -> Option<&'a str> {
1094 if self.finished { return None }
1096 let haystack = self.matcher.haystack();
1097 match self.matcher.next_match() {
1098 Some((a, b)) => unsafe {
1099 let elt = haystack.get_unchecked(self.start..a);
1103 None => self.get_end(),
1108 fn next_back(&mut self) -> Option<&'a str>
1109 where P::Searcher: ReverseSearcher<'a>
1111 if self.finished { return None }
1113 if !self.allow_trailing_empty {
1114 self.allow_trailing_empty = true;
1115 match self.next_back() {
1116 Some(elt) if !elt.is_empty() => return Some(elt),
1117 _ => if self.finished { return None }
1121 let haystack = self.matcher.haystack();
1122 match self.matcher.next_match_back() {
1123 Some((a, b)) => unsafe {
1124 let elt = haystack.get_unchecked(b..self.end);
1129 self.finished = true;
1130 Some(haystack.get_unchecked(self.start..self.end))
1136 generate_pattern_iterators! {
1138 /// Created with the method [`split`].
1140 /// [`split`]: ../../std/primitive.str.html#method.split
1143 /// Created with the method [`rsplit`].
1145 /// [`rsplit`]: ../../std/primitive.str.html#method.rsplit
1148 #[stable(feature = "rust1", since = "1.0.0")]
1150 SplitInternal yielding (&'a str);
1151 delegate double ended;
1154 generate_pattern_iterators! {
1156 /// Created with the method [`split_terminator`].
1158 /// [`split_terminator`]: ../../std/primitive.str.html#method.split_terminator
1159 struct SplitTerminator;
1161 /// Created with the method [`rsplit_terminator`].
1163 /// [`rsplit_terminator`]: ../../std/primitive.str.html#method.rsplit_terminator
1164 struct RSplitTerminator;
1166 #[stable(feature = "rust1", since = "1.0.0")]
1168 SplitInternal yielding (&'a str);
1169 delegate double ended;
1172 derive_pattern_clone!{
1173 clone SplitNInternal
1174 with |s| SplitNInternal { iter: s.iter.clone(), ..*s }
1177 struct SplitNInternal<'a, P: Pattern<'a>> {
1178 iter: SplitInternal<'a, P>,
1179 /// The number of splits remaining
1183 impl<'a, P> fmt::Debug for SplitNInternal<'a, P>
1185 P: Pattern<'a, Searcher: fmt::Debug>,
1187 fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
1188 f.debug_struct("SplitNInternal")
1189 .field("iter", &self.iter)
1190 .field("count", &self.count)
1195 impl<'a, P: Pattern<'a>> SplitNInternal<'a, P> {
1197 fn next(&mut self) -> Option<&'a str> {
1200 1 => { self.count = 0; self.iter.get_end() }
1201 _ => { self.count -= 1; self.iter.next() }
1206 fn next_back(&mut self) -> Option<&'a str>
1207 where P::Searcher: ReverseSearcher<'a>
1211 1 => { self.count = 0; self.iter.get_end() }
1212 _ => { self.count -= 1; self.iter.next_back() }
1217 generate_pattern_iterators! {
1219 /// Created with the method [`splitn`].
1221 /// [`splitn`]: ../../std/primitive.str.html#method.splitn
1224 /// Created with the method [`rsplitn`].
1226 /// [`rsplitn`]: ../../std/primitive.str.html#method.rsplitn
1229 #[stable(feature = "rust1", since = "1.0.0")]
1231 SplitNInternal yielding (&'a str);
1232 delegate single ended;
1235 derive_pattern_clone!{
1236 clone MatchIndicesInternal
1237 with |s| MatchIndicesInternal(s.0.clone())
1240 struct MatchIndicesInternal<'a, P: Pattern<'a>>(P::Searcher);
1242 impl<'a, P> fmt::Debug for MatchIndicesInternal<'a, P>
1244 P: Pattern<'a, Searcher: fmt::Debug>,
1246 fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
1247 f.debug_tuple("MatchIndicesInternal")
1253 impl<'a, P: Pattern<'a>> MatchIndicesInternal<'a, P> {
1255 fn next(&mut self) -> Option<(usize, &'a str)> {
1256 self.0.next_match().map(|(start, end)| unsafe {
1257 (start, self.0.haystack().get_unchecked(start..end))
1262 fn next_back(&mut self) -> Option<(usize, &'a str)>
1263 where P::Searcher: ReverseSearcher<'a>
1265 self.0.next_match_back().map(|(start, end)| unsafe {
1266 (start, self.0.haystack().get_unchecked(start..end))
1271 generate_pattern_iterators! {
1273 /// Created with the method [`match_indices`].
1275 /// [`match_indices`]: ../../std/primitive.str.html#method.match_indices
1276 struct MatchIndices;
1278 /// Created with the method [`rmatch_indices`].
1280 /// [`rmatch_indices`]: ../../std/primitive.str.html#method.rmatch_indices
1281 struct RMatchIndices;
1283 #[stable(feature = "str_match_indices", since = "1.5.0")]
1285 MatchIndicesInternal yielding ((usize, &'a str));
1286 delegate double ended;
1289 derive_pattern_clone!{
1290 clone MatchesInternal
1291 with |s| MatchesInternal(s.0.clone())
1294 struct MatchesInternal<'a, P: Pattern<'a>>(P::Searcher);
1296 impl<'a, P> fmt::Debug for MatchesInternal<'a, P>
1298 P: Pattern<'a, Searcher: fmt::Debug>,
1300 fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
1301 f.debug_tuple("MatchesInternal")
1307 impl<'a, P: Pattern<'a>> MatchesInternal<'a, P> {
1309 fn next(&mut self) -> Option<&'a str> {
1310 self.0.next_match().map(|(a, b)| unsafe {
1311 // Indices are known to be on utf8 boundaries
1312 self.0.haystack().get_unchecked(a..b)
1317 fn next_back(&mut self) -> Option<&'a str>
1318 where P::Searcher: ReverseSearcher<'a>
1320 self.0.next_match_back().map(|(a, b)| unsafe {
1321 // Indices are known to be on utf8 boundaries
1322 self.0.haystack().get_unchecked(a..b)
1327 generate_pattern_iterators! {
1329 /// Created with the method [`matches`].
1331 /// [`matches`]: ../../std/primitive.str.html#method.matches
1334 /// Created with the method [`rmatches`].
1336 /// [`rmatches`]: ../../std/primitive.str.html#method.rmatches
1339 #[stable(feature = "str_matches", since = "1.2.0")]
1341 MatchesInternal yielding (&'a str);
1342 delegate double ended;
1345 /// An iterator over the lines of a string, as string slices.
1347 /// This struct is created with the [`lines`] method on [`str`].
1348 /// See its documentation for more.
1350 /// [`lines`]: ../../std/primitive.str.html#method.lines
1351 /// [`str`]: ../../std/primitive.str.html
1352 #[stable(feature = "rust1", since = "1.0.0")]
1353 #[derive(Clone, Debug)]
1354 pub struct Lines<'a>(Map<SplitTerminator<'a, char>, LinesAnyMap>);
1356 #[stable(feature = "rust1", since = "1.0.0")]
1357 impl<'a> Iterator for Lines<'a> {
1358 type Item = &'a str;
1361 fn next(&mut self) -> Option<&'a str> {
1366 fn size_hint(&self) -> (usize, Option<usize>) {
1371 fn last(mut self) -> Option<&'a str> {
1376 #[stable(feature = "rust1", since = "1.0.0")]
1377 impl<'a> DoubleEndedIterator for Lines<'a> {
1379 fn next_back(&mut self) -> Option<&'a str> {
1384 #[stable(feature = "fused", since = "1.26.0")]
1385 impl FusedIterator for Lines<'_> {}
1387 /// Created with the method [`lines_any`].
1389 /// [`lines_any`]: ../../std/primitive.str.html#method.lines_any
1390 #[stable(feature = "rust1", since = "1.0.0")]
1391 #[rustc_deprecated(since = "1.4.0", reason = "use lines()/Lines instead now")]
1392 #[derive(Clone, Debug)]
1393 #[allow(deprecated)]
1394 pub struct LinesAny<'a>(Lines<'a>);
1397 /// A nameable, cloneable fn type
1399 struct LinesAnyMap impl<'a> Fn = |line: &'a str| -> &'a str {
1401 if l > 0 && line.as_bytes()[l - 1] == b'\r' { &line[0 .. l - 1] }
1406 #[stable(feature = "rust1", since = "1.0.0")]
1407 #[allow(deprecated)]
1408 impl<'a> Iterator for LinesAny<'a> {
1409 type Item = &'a str;
1412 fn next(&mut self) -> Option<&'a str> {
1417 fn size_hint(&self) -> (usize, Option<usize>) {
1422 #[stable(feature = "rust1", since = "1.0.0")]
1423 #[allow(deprecated)]
1424 impl<'a> DoubleEndedIterator for LinesAny<'a> {
1426 fn next_back(&mut self) -> Option<&'a str> {
1431 #[stable(feature = "fused", since = "1.26.0")]
1432 #[allow(deprecated)]
1433 impl FusedIterator for LinesAny<'_> {}
1436 Section: UTF-8 validation
1439 // use truncation to fit u64 into usize
1440 const NONASCII_MASK: usize = 0x80808080_80808080u64 as usize;
1442 /// Returns `true` if any byte in the word `x` is nonascii (>= 128).
1444 fn contains_nonascii(x: usize) -> bool {
1445 (x & NONASCII_MASK) != 0
1448 /// Walks through `v` checking that it's a valid UTF-8 sequence,
1449 /// returning `Ok(())` in that case, or, if it is invalid, `Err(err)`.
1451 fn run_utf8_validation(v: &[u8]) -> Result<(), Utf8Error> {
1455 let usize_bytes = mem::size_of::<usize>();
1456 let ascii_block_size = 2 * usize_bytes;
1457 let blocks_end = if len >= ascii_block_size { len - ascii_block_size + 1 } else { 0 };
1458 let align = v.as_ptr().align_offset(usize_bytes);
1461 let old_offset = index;
1463 ($error_len: expr) => {
1464 return Err(Utf8Error {
1465 valid_up_to: old_offset,
1466 error_len: $error_len,
1471 macro_rules! next { () => {{
1473 // we needed data, but there was none: error!
1480 let first = v[index];
1482 let w = UTF8_CHAR_WIDTH[first as usize];
1483 // 2-byte encoding is for codepoints \u{0080} to \u{07ff}
1484 // first C2 80 last DF BF
1485 // 3-byte encoding is for codepoints \u{0800} to \u{ffff}
1486 // first E0 A0 80 last EF BF BF
1487 // excluding surrogates codepoints \u{d800} to \u{dfff}
1488 // ED A0 80 to ED BF BF
1489 // 4-byte encoding is for codepoints \u{1000}0 to \u{10ff}ff
1490 // first F0 90 80 80 last F4 8F BF BF
1492 // Use the UTF-8 syntax from the RFC
1494 // https://tools.ietf.org/html/rfc3629
1496 // UTF8-2 = %xC2-DF UTF8-tail
1497 // UTF8-3 = %xE0 %xA0-BF UTF8-tail / %xE1-EC 2( UTF8-tail ) /
1498 // %xED %x80-9F UTF8-tail / %xEE-EF 2( UTF8-tail )
1499 // UTF8-4 = %xF0 %x90-BF 2( UTF8-tail ) / %xF1-F3 3( UTF8-tail ) /
1500 // %xF4 %x80-8F 2( UTF8-tail )
1502 2 => if next!() & !CONT_MASK != TAG_CONT_U8 {
1506 match (first, next!()) {
1507 (0xE0 , 0xA0 ..= 0xBF) |
1508 (0xE1 ..= 0xEC, 0x80 ..= 0xBF) |
1509 (0xED , 0x80 ..= 0x9F) |
1510 (0xEE ..= 0xEF, 0x80 ..= 0xBF) => {}
1513 if next!() & !CONT_MASK != TAG_CONT_U8 {
1518 match (first, next!()) {
1519 (0xF0 , 0x90 ..= 0xBF) |
1520 (0xF1 ..= 0xF3, 0x80 ..= 0xBF) |
1521 (0xF4 , 0x80 ..= 0x8F) => {}
1524 if next!() & !CONT_MASK != TAG_CONT_U8 {
1527 if next!() & !CONT_MASK != TAG_CONT_U8 {
1535 // Ascii case, try to skip forward quickly.
1536 // When the pointer is aligned, read 2 words of data per iteration
1537 // until we find a word containing a non-ascii byte.
1538 if align != usize::max_value() && align.wrapping_sub(index) % usize_bytes == 0 {
1539 let ptr = v.as_ptr();
1540 while index < blocks_end {
1542 let block = ptr.add(index) as *const usize;
1543 // break if there is a nonascii byte
1544 let zu = contains_nonascii(*block);
1545 let zv = contains_nonascii(*block.offset(1));
1550 index += ascii_block_size;
1552 // step from the point where the wordwise loop stopped
1553 while index < len && v[index] < 128 {
1565 // https://tools.ietf.org/html/rfc3629
1566 static UTF8_CHAR_WIDTH: [u8; 256] = [
1567 1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,
1568 1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1, // 0x1F
1569 1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,
1570 1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1, // 0x3F
1571 1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,
1572 1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1, // 0x5F
1573 1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,
1574 1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1, // 0x7F
1575 0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,
1576 0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0, // 0x9F
1577 0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,
1578 0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0, // 0xBF
1579 0,0,2,2,2,2,2,2,2,2,2,2,2,2,2,2,
1580 2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2, // 0xDF
1581 3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3, // 0xEF
1582 4,4,4,4,4,0,0,0,0,0,0,0,0,0,0,0, // 0xFF
1585 /// Given a first byte, determines how many bytes are in this UTF-8 character.
1586 #[unstable(feature = "str_internals", issue = "0")]
1588 pub fn utf8_char_width(b: u8) -> usize {
1589 UTF8_CHAR_WIDTH[b as usize] as usize
1592 /// Mask of the value bits of a continuation byte.
1593 const CONT_MASK: u8 = 0b0011_1111;
1594 /// Value of the tag bits (tag mask is !CONT_MASK) of a continuation byte.
1595 const TAG_CONT_U8: u8 = 0b1000_0000;
1598 Section: Trait implementations
1602 use crate::cmp::Ordering;
1604 use crate::slice::{self, SliceIndex};
1606 /// Implements ordering of strings.
1608 /// Strings are ordered lexicographically by their byte values. This orders Unicode code
1609 /// points based on their positions in the code charts. This is not necessarily the same as
1610 /// "alphabetical" order, which varies by language and locale. Sorting strings according to
1611 /// culturally-accepted standards requires locale-specific data that is outside the scope of
1613 #[stable(feature = "rust1", since = "1.0.0")]
1616 fn cmp(&self, other: &str) -> Ordering {
1617 self.as_bytes().cmp(other.as_bytes())
1621 #[stable(feature = "rust1", since = "1.0.0")]
1622 impl PartialEq for str {
1624 fn eq(&self, other: &str) -> bool {
1625 self.as_bytes() == other.as_bytes()
1628 fn ne(&self, other: &str) -> bool { !(*self).eq(other) }
1631 #[stable(feature = "rust1", since = "1.0.0")]
1634 /// Implements comparison operations on strings.
1636 /// Strings are compared lexicographically by their byte values. This compares Unicode code
1637 /// points based on their positions in the code charts. This is not necessarily the same as
1638 /// "alphabetical" order, which varies by language and locale. Comparing strings according to
1639 /// culturally-accepted standards requires locale-specific data that is outside the scope of
1641 #[stable(feature = "rust1", since = "1.0.0")]
1642 impl PartialOrd for str {
1644 fn partial_cmp(&self, other: &str) -> Option<Ordering> {
1645 Some(self.cmp(other))
1649 #[stable(feature = "rust1", since = "1.0.0")]
1650 impl<I> ops::Index<I> for str
1654 type Output = I::Output;
1657 fn index(&self, index: I) -> &I::Output {
1662 #[stable(feature = "rust1", since = "1.0.0")]
1663 impl<I> ops::IndexMut<I> for str
1668 fn index_mut(&mut self, index: I) -> &mut I::Output {
1669 index.index_mut(self)
1675 fn str_index_overflow_fail() -> ! {
1676 panic!("attempted to index str up to maximum usize");
1679 /// Implements substring slicing with syntax `&self[..]` or `&mut self[..]`.
1681 /// Returns a slice of the whole string, i.e., returns `&self` or `&mut
1682 /// self`. Equivalent to `&self[0 .. len]` or `&mut self[0 .. len]`. Unlike
1683 /// other indexing operations, this can never panic.
1685 /// This operation is `O(1)`.
1687 /// Prior to 1.20.0, these indexing operations were still supported by
1688 /// direct implementation of `Index` and `IndexMut`.
1690 /// Equivalent to `&self[0 .. len]` or `&mut self[0 .. len]`.
1691 #[stable(feature = "str_checked_slicing", since = "1.20.0")]
1692 impl SliceIndex<str> for ops::RangeFull {
1695 fn get(self, slice: &str) -> Option<&Self::Output> {
1699 fn get_mut(self, slice: &mut str) -> Option<&mut Self::Output> {
1703 unsafe fn get_unchecked(self, slice: &str) -> &Self::Output {
1707 unsafe fn get_unchecked_mut(self, slice: &mut str) -> &mut Self::Output {
1711 fn index(self, slice: &str) -> &Self::Output {
1715 fn index_mut(self, slice: &mut str) -> &mut Self::Output {
1720 /// Implements substring slicing with syntax `&self[begin .. end]` or `&mut
1721 /// self[begin .. end]`.
1723 /// Returns a slice of the given string from the byte range
1724 /// [`begin`, `end`).
1726 /// This operation is `O(1)`.
1728 /// Prior to 1.20.0, these indexing operations were still supported by
1729 /// direct implementation of `Index` and `IndexMut`.
1733 /// Panics if `begin` or `end` does not point to the starting byte offset of
1734 /// a character (as defined by `is_char_boundary`), if `begin > end`, or if
1740 /// let s = "Löwe 老虎 Léopard";
1741 /// assert_eq!(&s[0 .. 1], "L");
1743 /// assert_eq!(&s[1 .. 9], "öwe 老");
1745 /// // these will panic:
1746 /// // byte 2 lies within `ö`:
1749 /// // byte 8 lies within `老`
1752 /// // byte 100 is outside the string
1753 /// // &s[3 .. 100];
1755 #[stable(feature = "str_checked_slicing", since = "1.20.0")]
1756 impl SliceIndex<str> for ops::Range<usize> {
1759 fn get(self, slice: &str) -> Option<&Self::Output> {
1760 if self.start <= self.end &&
1761 slice.is_char_boundary(self.start) &&
1762 slice.is_char_boundary(self.end) {
1763 Some(unsafe { self.get_unchecked(slice) })
1769 fn get_mut(self, slice: &mut str) -> Option<&mut Self::Output> {
1770 if self.start <= self.end &&
1771 slice.is_char_boundary(self.start) &&
1772 slice.is_char_boundary(self.end) {
1773 Some(unsafe { self.get_unchecked_mut(slice) })
1779 unsafe fn get_unchecked(self, slice: &str) -> &Self::Output {
1780 let ptr = slice.as_ptr().add(self.start);
1781 let len = self.end - self.start;
1782 super::from_utf8_unchecked(slice::from_raw_parts(ptr, len))
1785 unsafe fn get_unchecked_mut(self, slice: &mut str) -> &mut Self::Output {
1786 let ptr = slice.as_mut_ptr().add(self.start);
1787 let len = self.end - self.start;
1788 super::from_utf8_unchecked_mut(slice::from_raw_parts_mut(ptr, len))
1791 fn index(self, slice: &str) -> &Self::Output {
1792 let (start, end) = (self.start, self.end);
1793 self.get(slice).unwrap_or_else(|| super::slice_error_fail(slice, start, end))
1796 fn index_mut(self, slice: &mut str) -> &mut Self::Output {
1797 // is_char_boundary checks that the index is in [0, .len()]
1798 // cannot reuse `get` as above, because of NLL trouble
1799 if self.start <= self.end &&
1800 slice.is_char_boundary(self.start) &&
1801 slice.is_char_boundary(self.end) {
1802 unsafe { self.get_unchecked_mut(slice) }
1804 super::slice_error_fail(slice, self.start, self.end)
1809 /// Implements substring slicing with syntax `&self[.. end]` or `&mut
1812 /// Returns a slice of the given string from the byte range [`0`, `end`).
1813 /// Equivalent to `&self[0 .. end]` or `&mut self[0 .. end]`.
1815 /// This operation is `O(1)`.
1817 /// Prior to 1.20.0, these indexing operations were still supported by
1818 /// direct implementation of `Index` and `IndexMut`.
1822 /// Panics if `end` does not point to the starting byte offset of a
1823 /// character (as defined by `is_char_boundary`), or if `end > len`.
1824 #[stable(feature = "str_checked_slicing", since = "1.20.0")]
1825 impl SliceIndex<str> for ops::RangeTo<usize> {
1828 fn get(self, slice: &str) -> Option<&Self::Output> {
1829 if slice.is_char_boundary(self.end) {
1830 Some(unsafe { self.get_unchecked(slice) })
1836 fn get_mut(self, slice: &mut str) -> Option<&mut Self::Output> {
1837 if slice.is_char_boundary(self.end) {
1838 Some(unsafe { self.get_unchecked_mut(slice) })
1844 unsafe fn get_unchecked(self, slice: &str) -> &Self::Output {
1845 let ptr = slice.as_ptr();
1846 super::from_utf8_unchecked(slice::from_raw_parts(ptr, self.end))
1849 unsafe fn get_unchecked_mut(self, slice: &mut str) -> &mut Self::Output {
1850 let ptr = slice.as_mut_ptr();
1851 super::from_utf8_unchecked_mut(slice::from_raw_parts_mut(ptr, self.end))
1854 fn index(self, slice: &str) -> &Self::Output {
1856 self.get(slice).unwrap_or_else(|| super::slice_error_fail(slice, 0, end))
1859 fn index_mut(self, slice: &mut str) -> &mut Self::Output {
1860 // is_char_boundary checks that the index is in [0, .len()]
1861 if slice.is_char_boundary(self.end) {
1862 unsafe { self.get_unchecked_mut(slice) }
1864 super::slice_error_fail(slice, 0, self.end)
1869 /// Implements substring slicing with syntax `&self[begin ..]` or `&mut
1870 /// self[begin ..]`.
1872 /// Returns a slice of the given string from the byte range [`begin`,
1873 /// `len`). Equivalent to `&self[begin .. len]` or `&mut self[begin ..
1876 /// This operation is `O(1)`.
1878 /// Prior to 1.20.0, these indexing operations were still supported by
1879 /// direct implementation of `Index` and `IndexMut`.
1883 /// Panics if `begin` does not point to the starting byte offset of
1884 /// a character (as defined by `is_char_boundary`), or if `begin >= len`.
1885 #[stable(feature = "str_checked_slicing", since = "1.20.0")]
1886 impl SliceIndex<str> for ops::RangeFrom<usize> {
1889 fn get(self, slice: &str) -> Option<&Self::Output> {
1890 if slice.is_char_boundary(self.start) {
1891 Some(unsafe { self.get_unchecked(slice) })
1897 fn get_mut(self, slice: &mut str) -> Option<&mut Self::Output> {
1898 if slice.is_char_boundary(self.start) {
1899 Some(unsafe { self.get_unchecked_mut(slice) })
1905 unsafe fn get_unchecked(self, slice: &str) -> &Self::Output {
1906 let ptr = slice.as_ptr().add(self.start);
1907 let len = slice.len() - self.start;
1908 super::from_utf8_unchecked(slice::from_raw_parts(ptr, len))
1911 unsafe fn get_unchecked_mut(self, slice: &mut str) -> &mut Self::Output {
1912 let ptr = slice.as_mut_ptr().add(self.start);
1913 let len = slice.len() - self.start;
1914 super::from_utf8_unchecked_mut(slice::from_raw_parts_mut(ptr, len))
1917 fn index(self, slice: &str) -> &Self::Output {
1918 let (start, end) = (self.start, slice.len());
1919 self.get(slice).unwrap_or_else(|| super::slice_error_fail(slice, start, end))
1922 fn index_mut(self, slice: &mut str) -> &mut Self::Output {
1923 // is_char_boundary checks that the index is in [0, .len()]
1924 if slice.is_char_boundary(self.start) {
1925 unsafe { self.get_unchecked_mut(slice) }
1927 super::slice_error_fail(slice, self.start, slice.len())
1932 /// Implements substring slicing with syntax `&self[begin ..= end]` or `&mut
1933 /// self[begin ..= end]`.
1935 /// Returns a slice of the given string from the byte range
1936 /// [`begin`, `end`]. Equivalent to `&self [begin .. end + 1]` or `&mut
1937 /// self[begin .. end + 1]`, except if `end` has the maximum value for
1940 /// This operation is `O(1)`.
1944 /// Panics if `begin` does not point to the starting byte offset of
1945 /// a character (as defined by `is_char_boundary`), if `end` does not point
1946 /// to the ending byte offset of a character (`end + 1` is either a starting
1947 /// byte offset or equal to `len`), if `begin > end`, or if `end >= len`.
1948 #[stable(feature = "inclusive_range", since = "1.26.0")]
1949 impl SliceIndex<str> for ops::RangeInclusive<usize> {
1952 fn get(self, slice: &str) -> Option<&Self::Output> {
1953 if *self.end() == usize::max_value() { None }
1954 else { (*self.start()..self.end()+1).get(slice) }
1957 fn get_mut(self, slice: &mut str) -> Option<&mut Self::Output> {
1958 if *self.end() == usize::max_value() { None }
1959 else { (*self.start()..self.end()+1).get_mut(slice) }
1962 unsafe fn get_unchecked(self, slice: &str) -> &Self::Output {
1963 (*self.start()..self.end()+1).get_unchecked(slice)
1966 unsafe fn get_unchecked_mut(self, slice: &mut str) -> &mut Self::Output {
1967 (*self.start()..self.end()+1).get_unchecked_mut(slice)
1970 fn index(self, slice: &str) -> &Self::Output {
1971 if *self.end() == usize::max_value() { str_index_overflow_fail(); }
1972 (*self.start()..self.end()+1).index(slice)
1975 fn index_mut(self, slice: &mut str) -> &mut Self::Output {
1976 if *self.end() == usize::max_value() { str_index_overflow_fail(); }
1977 (*self.start()..self.end()+1).index_mut(slice)
1981 /// Implements substring slicing with syntax `&self[..= end]` or `&mut
1984 /// Returns a slice of the given string from the byte range [0, `end`].
1985 /// Equivalent to `&self [0 .. end + 1]`, except if `end` has the maximum
1986 /// value for `usize`.
1988 /// This operation is `O(1)`.
1992 /// Panics if `end` does not point to the ending byte offset of a character
1993 /// (`end + 1` is either a starting byte offset as defined by
1994 /// `is_char_boundary`, or equal to `len`), or if `end >= len`.
1995 #[stable(feature = "inclusive_range", since = "1.26.0")]
1996 impl SliceIndex<str> for ops::RangeToInclusive<usize> {
1999 fn get(self, slice: &str) -> Option<&Self::Output> {
2000 if self.end == usize::max_value() { None }
2001 else { (..self.end+1).get(slice) }
2004 fn get_mut(self, slice: &mut str) -> Option<&mut Self::Output> {
2005 if self.end == usize::max_value() { None }
2006 else { (..self.end+1).get_mut(slice) }
2009 unsafe fn get_unchecked(self, slice: &str) -> &Self::Output {
2010 (..self.end+1).get_unchecked(slice)
2013 unsafe fn get_unchecked_mut(self, slice: &mut str) -> &mut Self::Output {
2014 (..self.end+1).get_unchecked_mut(slice)
2017 fn index(self, slice: &str) -> &Self::Output {
2018 if self.end == usize::max_value() { str_index_overflow_fail(); }
2019 (..self.end+1).index(slice)
2022 fn index_mut(self, slice: &mut str) -> &mut Self::Output {
2023 if self.end == usize::max_value() { str_index_overflow_fail(); }
2024 (..self.end+1).index_mut(slice)
2029 // truncate `&str` to length at most equal to `max`
2030 // return `true` if it were truncated, and the new str.
2031 fn truncate_to_char_boundary(s: &str, mut max: usize) -> (bool, &str) {
2035 while !s.is_char_boundary(max) {
2044 fn slice_error_fail(s: &str, begin: usize, end: usize) -> ! {
2045 const MAX_DISPLAY_LENGTH: usize = 256;
2046 let (truncated, s_trunc) = truncate_to_char_boundary(s, MAX_DISPLAY_LENGTH);
2047 let ellipsis = if truncated { "[...]" } else { "" };
2050 if begin > s.len() || end > s.len() {
2051 let oob_index = if begin > s.len() { begin } else { end };
2052 panic!("byte index {} is out of bounds of `{}`{}", oob_index, s_trunc, ellipsis);
2056 assert!(begin <= end, "begin <= end ({} <= {}) when slicing `{}`{}",
2057 begin, end, s_trunc, ellipsis);
2059 // 3. character boundary
2060 let index = if !s.is_char_boundary(begin) { begin } else { end };
2061 // find the character
2062 let mut char_start = index;
2063 while !s.is_char_boundary(char_start) {
2066 // `char_start` must be less than len and a char boundary
2067 let ch = s[char_start..].chars().next().unwrap();
2068 let char_range = char_start .. char_start + ch.len_utf8();
2069 panic!("byte index {} is not a char boundary; it is inside {:?} (bytes {:?}) of `{}`{}",
2070 index, ch, char_range, s_trunc, ellipsis);
2076 /// Returns the length of `self`.
2078 /// This length is in bytes, not [`char`]s or graphemes. In other words,
2079 /// it may not be what a human considers the length of the string.
2086 /// let len = "foo".len();
2087 /// assert_eq!(3, len);
2089 /// assert_eq!("ƒoo".len(), 4); // fancy f!
2090 /// assert_eq!("ƒoo".chars().count(), 3);
2092 #[stable(feature = "rust1", since = "1.0.0")]
2094 pub const fn len(&self) -> usize {
2095 self.as_bytes().len()
2098 /// Returns `true` if `self` has a length of zero bytes.
2106 /// assert!(s.is_empty());
2108 /// let s = "not empty";
2109 /// assert!(!s.is_empty());
2112 #[stable(feature = "rust1", since = "1.0.0")]
2113 pub const fn is_empty(&self) -> bool {
2117 /// Checks that `index`-th byte lies at the start and/or end of a
2118 /// UTF-8 code point sequence.
2120 /// The start and end of the string (when `index == self.len()`) are
2121 /// considered to be
2124 /// Returns `false` if `index` is greater than `self.len()`.
2129 /// let s = "Löwe 老虎 Léopard";
2130 /// assert!(s.is_char_boundary(0));
2132 /// assert!(s.is_char_boundary(6));
2133 /// assert!(s.is_char_boundary(s.len()));
2135 /// // second byte of `ö`
2136 /// assert!(!s.is_char_boundary(2));
2138 /// // third byte of `老`
2139 /// assert!(!s.is_char_boundary(8));
2141 #[stable(feature = "is_char_boundary", since = "1.9.0")]
2143 pub fn is_char_boundary(&self, index: usize) -> bool {
2144 // 0 and len are always ok.
2145 // Test for 0 explicitly so that it can optimize out the check
2146 // easily and skip reading string data for that case.
2147 if index == 0 || index == self.len() { return true; }
2148 match self.as_bytes().get(index) {
2150 // This is bit magic equivalent to: b < 128 || b >= 192
2151 Some(&b) => (b as i8) >= -0x40,
2155 /// Converts a string slice to a byte slice. To convert the byte slice back
2156 /// into a string slice, use the [`str::from_utf8`] function.
2158 /// [`str::from_utf8`]: ./str/fn.from_utf8.html
2165 /// let bytes = "bors".as_bytes();
2166 /// assert_eq!(b"bors", bytes);
2168 #[stable(feature = "rust1", since = "1.0.0")]
2170 // SAFETY: const sound because we transmute two types with the same layout
2171 #[allow(unused_attributes)]
2172 #[allow_internal_unstable(const_fn_union)]
2173 pub const fn as_bytes(&self) -> &[u8] {
2179 unsafe { Slices { str: self }.slice }
2182 /// Converts a mutable string slice to a mutable byte slice. To convert the
2183 /// mutable byte slice back into a mutable string slice, use the
2184 /// [`str::from_utf8_mut`] function.
2186 /// [`str::from_utf8_mut`]: ./str/fn.from_utf8_mut.html
2193 /// let mut s = String::from("Hello");
2194 /// let bytes = unsafe { s.as_bytes_mut() };
2196 /// assert_eq!(b"Hello", bytes);
2202 /// let mut s = String::from("🗻∈🌏");
2205 /// let bytes = s.as_bytes_mut();
2207 /// bytes[0] = 0xF0;
2208 /// bytes[1] = 0x9F;
2209 /// bytes[2] = 0x8D;
2210 /// bytes[3] = 0x94;
2213 /// assert_eq!("🍔∈🌏", s);
2215 #[stable(feature = "str_mut_extras", since = "1.20.0")]
2217 pub unsafe fn as_bytes_mut(&mut self) -> &mut [u8] {
2218 &mut *(self as *mut str as *mut [u8])
2221 /// Converts a string slice to a raw pointer.
2223 /// As string slices are a slice of bytes, the raw pointer points to a
2224 /// [`u8`]. This pointer will be pointing to the first byte of the string
2227 /// The caller must ensure that the returned pointer is never written to.
2228 /// If you need to mutate the contents of the string slice, use [`as_mut_ptr`].
2230 /// [`u8`]: primitive.u8.html
2231 /// [`as_mut_ptr`]: #method.as_mut_ptr
2238 /// let s = "Hello";
2239 /// let ptr = s.as_ptr();
2241 #[stable(feature = "rust1", since = "1.0.0")]
2243 pub const fn as_ptr(&self) -> *const u8 {
2244 self as *const str as *const u8
2247 /// Converts a mutable string slice to a raw pointer.
2249 /// As string slices are a slice of bytes, the raw pointer points to a
2250 /// [`u8`]. This pointer will be pointing to the first byte of the string
2253 /// It is your responsibility to make sure that the string slice only gets
2254 /// modified in a way that it remains valid UTF-8.
2256 /// [`u8`]: primitive.u8.html
2257 #[stable(feature = "str_as_mut_ptr", since = "1.36.0")]
2259 pub fn as_mut_ptr(&mut self) -> *mut u8 {
2260 self as *mut str as *mut u8
2263 /// Returns a subslice of `str`.
2265 /// This is the non-panicking alternative to indexing the `str`. Returns
2266 /// [`None`] whenever equivalent indexing operation would panic.
2268 /// [`None`]: option/enum.Option.html#variant.None
2273 /// let v = String::from("🗻∈🌏");
2275 /// assert_eq!(Some("🗻"), v.get(0..4));
2277 /// // indices not on UTF-8 sequence boundaries
2278 /// assert!(v.get(1..).is_none());
2279 /// assert!(v.get(..8).is_none());
2281 /// // out of bounds
2282 /// assert!(v.get(..42).is_none());
2284 #[stable(feature = "str_checked_slicing", since = "1.20.0")]
2286 pub fn get<I: SliceIndex<str>>(&self, i: I) -> Option<&I::Output> {
2290 /// Returns a mutable subslice of `str`.
2292 /// This is the non-panicking alternative to indexing the `str`. Returns
2293 /// [`None`] whenever equivalent indexing operation would panic.
2295 /// [`None`]: option/enum.Option.html#variant.None
2300 /// let mut v = String::from("hello");
2301 /// // correct length
2302 /// assert!(v.get_mut(0..5).is_some());
2303 /// // out of bounds
2304 /// assert!(v.get_mut(..42).is_none());
2305 /// assert_eq!(Some("he"), v.get_mut(0..2).map(|v| &*v));
2307 /// assert_eq!("hello", v);
2309 /// let s = v.get_mut(0..2);
2310 /// let s = s.map(|s| {
2311 /// s.make_ascii_uppercase();
2314 /// assert_eq!(Some("HE"), s);
2316 /// assert_eq!("HEllo", v);
2318 #[stable(feature = "str_checked_slicing", since = "1.20.0")]
2320 pub fn get_mut<I: SliceIndex<str>>(&mut self, i: I) -> Option<&mut I::Output> {
2324 /// Returns a unchecked subslice of `str`.
2326 /// This is the unchecked alternative to indexing the `str`.
2330 /// Callers of this function are responsible that these preconditions are
2333 /// * The starting index must come before the ending index;
2334 /// * Indexes must be within bounds of the original slice;
2335 /// * Indexes must lie on UTF-8 sequence boundaries.
2337 /// Failing that, the returned string slice may reference invalid memory or
2338 /// violate the invariants communicated by the `str` type.
2345 /// assert_eq!("🗻", v.get_unchecked(0..4));
2346 /// assert_eq!("∈", v.get_unchecked(4..7));
2347 /// assert_eq!("🌏", v.get_unchecked(7..11));
2350 #[stable(feature = "str_checked_slicing", since = "1.20.0")]
2352 pub unsafe fn get_unchecked<I: SliceIndex<str>>(&self, i: I) -> &I::Output {
2353 i.get_unchecked(self)
2356 /// Returns a mutable, unchecked subslice of `str`.
2358 /// This is the unchecked alternative to indexing the `str`.
2362 /// Callers of this function are responsible that these preconditions are
2365 /// * The starting index must come before the ending index;
2366 /// * Indexes must be within bounds of the original slice;
2367 /// * Indexes must lie on UTF-8 sequence boundaries.
2369 /// Failing that, the returned string slice may reference invalid memory or
2370 /// violate the invariants communicated by the `str` type.
2375 /// let mut v = String::from("🗻∈🌏");
2377 /// assert_eq!("🗻", v.get_unchecked_mut(0..4));
2378 /// assert_eq!("∈", v.get_unchecked_mut(4..7));
2379 /// assert_eq!("🌏", v.get_unchecked_mut(7..11));
2382 #[stable(feature = "str_checked_slicing", since = "1.20.0")]
2384 pub unsafe fn get_unchecked_mut<I: SliceIndex<str>>(&mut self, i: I) -> &mut I::Output {
2385 i.get_unchecked_mut(self)
2388 /// Creates a string slice from another string slice, bypassing safety
2391 /// This is generally not recommended, use with caution! For a safe
2392 /// alternative see [`str`] and [`Index`].
2394 /// [`str`]: primitive.str.html
2395 /// [`Index`]: ops/trait.Index.html
2397 /// This new slice goes from `begin` to `end`, including `begin` but
2398 /// excluding `end`.
2400 /// To get a mutable string slice instead, see the
2401 /// [`slice_mut_unchecked`] method.
2403 /// [`slice_mut_unchecked`]: #method.slice_mut_unchecked
2407 /// Callers of this function are responsible that three preconditions are
2410 /// * `begin` must come before `end`.
2411 /// * `begin` and `end` must be byte positions within the string slice.
2412 /// * `begin` and `end` must lie on UTF-8 sequence boundaries.
2419 /// let s = "Löwe 老虎 Léopard";
2422 /// assert_eq!("Löwe 老虎 Léopard", s.slice_unchecked(0, 21));
2425 /// let s = "Hello, world!";
2428 /// assert_eq!("world", s.slice_unchecked(7, 12));
2431 #[stable(feature = "rust1", since = "1.0.0")]
2432 #[rustc_deprecated(since = "1.29.0", reason = "use `get_unchecked(begin..end)` instead")]
2434 pub unsafe fn slice_unchecked(&self, begin: usize, end: usize) -> &str {
2435 (begin..end).get_unchecked(self)
2438 /// Creates a string slice from another string slice, bypassing safety
2440 /// This is generally not recommended, use with caution! For a safe
2441 /// alternative see [`str`] and [`IndexMut`].
2443 /// [`str`]: primitive.str.html
2444 /// [`IndexMut`]: ops/trait.IndexMut.html
2446 /// This new slice goes from `begin` to `end`, including `begin` but
2447 /// excluding `end`.
2449 /// To get an immutable string slice instead, see the
2450 /// [`slice_unchecked`] method.
2452 /// [`slice_unchecked`]: #method.slice_unchecked
2456 /// Callers of this function are responsible that three preconditions are
2459 /// * `begin` must come before `end`.
2460 /// * `begin` and `end` must be byte positions within the string slice.
2461 /// * `begin` and `end` must lie on UTF-8 sequence boundaries.
2462 #[stable(feature = "str_slice_mut", since = "1.5.0")]
2463 #[rustc_deprecated(since = "1.29.0", reason = "use `get_unchecked_mut(begin..end)` instead")]
2465 pub unsafe fn slice_mut_unchecked(&mut self, begin: usize, end: usize) -> &mut str {
2466 (begin..end).get_unchecked_mut(self)
2469 /// Divide one string slice into two at an index.
2471 /// The argument, `mid`, should be a byte offset from the start of the
2472 /// string. It must also be on the boundary of a UTF-8 code point.
2474 /// The two slices returned go from the start of the string slice to `mid`,
2475 /// and from `mid` to the end of the string slice.
2477 /// To get mutable string slices instead, see the [`split_at_mut`]
2480 /// [`split_at_mut`]: #method.split_at_mut
2484 /// Panics if `mid` is not on a UTF-8 code point boundary, or if it is
2485 /// beyond the last code point of the string slice.
2492 /// let s = "Per Martin-Löf";
2494 /// let (first, last) = s.split_at(3);
2496 /// assert_eq!("Per", first);
2497 /// assert_eq!(" Martin-Löf", last);
2500 #[stable(feature = "str_split_at", since = "1.4.0")]
2501 pub fn split_at(&self, mid: usize) -> (&str, &str) {
2502 // is_char_boundary checks that the index is in [0, .len()]
2503 if self.is_char_boundary(mid) {
2505 (self.get_unchecked(0..mid),
2506 self.get_unchecked(mid..self.len()))
2509 slice_error_fail(self, 0, mid)
2513 /// Divide one mutable string slice into two at an index.
2515 /// The argument, `mid`, should be a byte offset from the start of the
2516 /// string. It must also be on the boundary of a UTF-8 code point.
2518 /// The two slices returned go from the start of the string slice to `mid`,
2519 /// and from `mid` to the end of the string slice.
2521 /// To get immutable string slices instead, see the [`split_at`] method.
2523 /// [`split_at`]: #method.split_at
2527 /// Panics if `mid` is not on a UTF-8 code point boundary, or if it is
2528 /// beyond the last code point of the string slice.
2535 /// let mut s = "Per Martin-Löf".to_string();
2537 /// let (first, last) = s.split_at_mut(3);
2538 /// first.make_ascii_uppercase();
2539 /// assert_eq!("PER", first);
2540 /// assert_eq!(" Martin-Löf", last);
2542 /// assert_eq!("PER Martin-Löf", s);
2545 #[stable(feature = "str_split_at", since = "1.4.0")]
2546 pub fn split_at_mut(&mut self, mid: usize) -> (&mut str, &mut str) {
2547 // is_char_boundary checks that the index is in [0, .len()]
2548 if self.is_char_boundary(mid) {
2549 let len = self.len();
2550 let ptr = self.as_mut_ptr();
2552 (from_utf8_unchecked_mut(slice::from_raw_parts_mut(ptr, mid)),
2553 from_utf8_unchecked_mut(slice::from_raw_parts_mut(
2559 slice_error_fail(self, 0, mid)
2563 /// Returns an iterator over the [`char`]s of a string slice.
2565 /// As a string slice consists of valid UTF-8, we can iterate through a
2566 /// string slice by [`char`]. This method returns such an iterator.
2568 /// It's important to remember that [`char`] represents a Unicode Scalar
2569 /// Value, and may not match your idea of what a 'character' is. Iteration
2570 /// over grapheme clusters may be what you actually want.
2577 /// let word = "goodbye";
2579 /// let count = word.chars().count();
2580 /// assert_eq!(7, count);
2582 /// let mut chars = word.chars();
2584 /// assert_eq!(Some('g'), chars.next());
2585 /// assert_eq!(Some('o'), chars.next());
2586 /// assert_eq!(Some('o'), chars.next());
2587 /// assert_eq!(Some('d'), chars.next());
2588 /// assert_eq!(Some('b'), chars.next());
2589 /// assert_eq!(Some('y'), chars.next());
2590 /// assert_eq!(Some('e'), chars.next());
2592 /// assert_eq!(None, chars.next());
2595 /// Remember, [`char`]s may not match your human intuition about characters:
2600 /// let mut chars = y.chars();
2602 /// assert_eq!(Some('y'), chars.next()); // not 'y̆'
2603 /// assert_eq!(Some('\u{0306}'), chars.next());
2605 /// assert_eq!(None, chars.next());
2607 #[stable(feature = "rust1", since = "1.0.0")]
2609 pub fn chars(&self) -> Chars<'_> {
2610 Chars{iter: self.as_bytes().iter()}
2613 /// Returns an iterator over the [`char`]s of a string slice, and their
2616 /// As a string slice consists of valid UTF-8, we can iterate through a
2617 /// string slice by [`char`]. This method returns an iterator of both
2618 /// these [`char`]s, as well as their byte positions.
2620 /// The iterator yields tuples. The position is first, the [`char`] is
2628 /// let word = "goodbye";
2630 /// let count = word.char_indices().count();
2631 /// assert_eq!(7, count);
2633 /// let mut char_indices = word.char_indices();
2635 /// assert_eq!(Some((0, 'g')), char_indices.next());
2636 /// assert_eq!(Some((1, 'o')), char_indices.next());
2637 /// assert_eq!(Some((2, 'o')), char_indices.next());
2638 /// assert_eq!(Some((3, 'd')), char_indices.next());
2639 /// assert_eq!(Some((4, 'b')), char_indices.next());
2640 /// assert_eq!(Some((5, 'y')), char_indices.next());
2641 /// assert_eq!(Some((6, 'e')), char_indices.next());
2643 /// assert_eq!(None, char_indices.next());
2646 /// Remember, [`char`]s may not match your human intuition about characters:
2649 /// let yes = "y̆es";
2651 /// let mut char_indices = yes.char_indices();
2653 /// assert_eq!(Some((0, 'y')), char_indices.next()); // not (0, 'y̆')
2654 /// assert_eq!(Some((1, '\u{0306}')), char_indices.next());
2656 /// // note the 3 here - the last character took up two bytes
2657 /// assert_eq!(Some((3, 'e')), char_indices.next());
2658 /// assert_eq!(Some((4, 's')), char_indices.next());
2660 /// assert_eq!(None, char_indices.next());
2662 #[stable(feature = "rust1", since = "1.0.0")]
2664 pub fn char_indices(&self) -> CharIndices<'_> {
2665 CharIndices { front_offset: 0, iter: self.chars() }
2668 /// An iterator over the bytes of a string slice.
2670 /// As a string slice consists of a sequence of bytes, we can iterate
2671 /// through a string slice by byte. This method returns such an iterator.
2678 /// let mut bytes = "bors".bytes();
2680 /// assert_eq!(Some(b'b'), bytes.next());
2681 /// assert_eq!(Some(b'o'), bytes.next());
2682 /// assert_eq!(Some(b'r'), bytes.next());
2683 /// assert_eq!(Some(b's'), bytes.next());
2685 /// assert_eq!(None, bytes.next());
2687 #[stable(feature = "rust1", since = "1.0.0")]
2689 pub fn bytes(&self) -> Bytes<'_> {
2690 Bytes(self.as_bytes().iter().cloned())
2693 /// Splits a string slice by whitespace.
2695 /// The iterator returned will return string slices that are sub-slices of
2696 /// the original string slice, separated by any amount of whitespace.
2698 /// 'Whitespace' is defined according to the terms of the Unicode Derived
2699 /// Core Property `White_Space`. If you only want to split on ASCII whitespace
2700 /// instead, use [`split_ascii_whitespace`].
2702 /// [`split_ascii_whitespace`]: #method.split_ascii_whitespace
2709 /// let mut iter = "A few words".split_whitespace();
2711 /// assert_eq!(Some("A"), iter.next());
2712 /// assert_eq!(Some("few"), iter.next());
2713 /// assert_eq!(Some("words"), iter.next());
2715 /// assert_eq!(None, iter.next());
2718 /// All kinds of whitespace are considered:
2721 /// let mut iter = " Mary had\ta\u{2009}little \n\t lamb".split_whitespace();
2722 /// assert_eq!(Some("Mary"), iter.next());
2723 /// assert_eq!(Some("had"), iter.next());
2724 /// assert_eq!(Some("a"), iter.next());
2725 /// assert_eq!(Some("little"), iter.next());
2726 /// assert_eq!(Some("lamb"), iter.next());
2728 /// assert_eq!(None, iter.next());
2730 #[stable(feature = "split_whitespace", since = "1.1.0")]
2732 pub fn split_whitespace(&self) -> SplitWhitespace<'_> {
2733 SplitWhitespace { inner: self.split(IsWhitespace).filter(IsNotEmpty) }
2736 /// Splits a string slice by ASCII whitespace.
2738 /// The iterator returned will return string slices that are sub-slices of
2739 /// the original string slice, separated by any amount of ASCII whitespace.
2741 /// To split by Unicode `Whitespace` instead, use [`split_whitespace`].
2743 /// [`split_whitespace`]: #method.split_whitespace
2750 /// let mut iter = "A few words".split_ascii_whitespace();
2752 /// assert_eq!(Some("A"), iter.next());
2753 /// assert_eq!(Some("few"), iter.next());
2754 /// assert_eq!(Some("words"), iter.next());
2756 /// assert_eq!(None, iter.next());
2759 /// All kinds of ASCII whitespace are considered:
2762 /// let mut iter = " Mary had\ta little \n\t lamb".split_ascii_whitespace();
2763 /// assert_eq!(Some("Mary"), iter.next());
2764 /// assert_eq!(Some("had"), iter.next());
2765 /// assert_eq!(Some("a"), iter.next());
2766 /// assert_eq!(Some("little"), iter.next());
2767 /// assert_eq!(Some("lamb"), iter.next());
2769 /// assert_eq!(None, iter.next());
2771 #[stable(feature = "split_ascii_whitespace", since = "1.34.0")]
2773 pub fn split_ascii_whitespace(&self) -> SplitAsciiWhitespace<'_> {
2776 .split(IsAsciiWhitespace)
2777 .filter(BytesIsNotEmpty)
2778 .map(UnsafeBytesToStr);
2779 SplitAsciiWhitespace { inner }
2782 /// An iterator over the lines of a string, as string slices.
2784 /// Lines are ended with either a newline (`\n`) or a carriage return with
2785 /// a line feed (`\r\n`).
2787 /// The final line ending is optional.
2794 /// let text = "foo\r\nbar\n\nbaz\n";
2795 /// let mut lines = text.lines();
2797 /// assert_eq!(Some("foo"), lines.next());
2798 /// assert_eq!(Some("bar"), lines.next());
2799 /// assert_eq!(Some(""), lines.next());
2800 /// assert_eq!(Some("baz"), lines.next());
2802 /// assert_eq!(None, lines.next());
2805 /// The final line ending isn't required:
2808 /// let text = "foo\nbar\n\r\nbaz";
2809 /// let mut lines = text.lines();
2811 /// assert_eq!(Some("foo"), lines.next());
2812 /// assert_eq!(Some("bar"), lines.next());
2813 /// assert_eq!(Some(""), lines.next());
2814 /// assert_eq!(Some("baz"), lines.next());
2816 /// assert_eq!(None, lines.next());
2818 #[stable(feature = "rust1", since = "1.0.0")]
2820 pub fn lines(&self) -> Lines<'_> {
2821 Lines(self.split_terminator('\n').map(LinesAnyMap))
2824 /// An iterator over the lines of a string.
2825 #[stable(feature = "rust1", since = "1.0.0")]
2826 #[rustc_deprecated(since = "1.4.0", reason = "use lines() instead now")]
2828 #[allow(deprecated)]
2829 pub fn lines_any(&self) -> LinesAny<'_> {
2830 LinesAny(self.lines())
2833 /// Returns an iterator of `u16` over the string encoded as UTF-16.
2840 /// let text = "Zażółć gęślą jaźń";
2842 /// let utf8_len = text.len();
2843 /// let utf16_len = text.encode_utf16().count();
2845 /// assert!(utf16_len <= utf8_len);
2847 #[stable(feature = "encode_utf16", since = "1.8.0")]
2848 pub fn encode_utf16(&self) -> EncodeUtf16<'_> {
2849 EncodeUtf16 { chars: self.chars(), extra: 0 }
2852 /// Returns `true` if the given pattern matches a sub-slice of
2853 /// this string slice.
2855 /// Returns `false` if it does not.
2862 /// let bananas = "bananas";
2864 /// assert!(bananas.contains("nana"));
2865 /// assert!(!bananas.contains("apples"));
2867 #[stable(feature = "rust1", since = "1.0.0")]
2869 pub fn contains<'a, P: Pattern<'a>>(&'a self, pat: P) -> bool {
2870 pat.is_contained_in(self)
2873 /// Returns `true` if the given pattern matches a prefix of this
2876 /// Returns `false` if it does not.
2883 /// let bananas = "bananas";
2885 /// assert!(bananas.starts_with("bana"));
2886 /// assert!(!bananas.starts_with("nana"));
2888 #[stable(feature = "rust1", since = "1.0.0")]
2889 pub fn starts_with<'a, P: Pattern<'a>>(&'a self, pat: P) -> bool {
2890 pat.is_prefix_of(self)
2893 /// Returns `true` if the given pattern matches a suffix of this
2896 /// Returns `false` if it does not.
2903 /// let bananas = "bananas";
2905 /// assert!(bananas.ends_with("anas"));
2906 /// assert!(!bananas.ends_with("nana"));
2908 #[stable(feature = "rust1", since = "1.0.0")]
2909 pub fn ends_with<'a, P>(&'a self, pat: P) -> bool
2911 P: Pattern<'a, Searcher: ReverseSearcher<'a>>,
2913 pat.is_suffix_of(self)
2916 /// Returns the byte index of the first character of this string slice that
2917 /// matches the pattern.
2919 /// Returns [`None`] if the pattern doesn't match.
2921 /// The pattern can be a `&str`, [`char`], or a closure that determines if
2922 /// a character matches.
2924 /// [`None`]: option/enum.Option.html#variant.None
2928 /// Simple patterns:
2931 /// let s = "Löwe 老虎 Léopard";
2933 /// assert_eq!(s.find('L'), Some(0));
2934 /// assert_eq!(s.find('é'), Some(14));
2935 /// assert_eq!(s.find("Léopard"), Some(13));
2938 /// More complex patterns using point-free style and closures:
2941 /// let s = "Löwe 老虎 Léopard";
2943 /// assert_eq!(s.find(char::is_whitespace), Some(5));
2944 /// assert_eq!(s.find(char::is_lowercase), Some(1));
2945 /// assert_eq!(s.find(|c: char| c.is_whitespace() || c.is_lowercase()), Some(1));
2946 /// assert_eq!(s.find(|c: char| (c < 'o') && (c > 'a')), Some(4));
2949 /// Not finding the pattern:
2952 /// let s = "Löwe 老虎 Léopard";
2953 /// let x: &[_] = &['1', '2'];
2955 /// assert_eq!(s.find(x), None);
2957 #[stable(feature = "rust1", since = "1.0.0")]
2959 pub fn find<'a, P: Pattern<'a>>(&'a self, pat: P) -> Option<usize> {
2960 pat.into_searcher(self).next_match().map(|(i, _)| i)
2963 /// Returns the byte index of the last character of this string slice that
2964 /// matches the pattern.
2966 /// Returns [`None`] if the pattern doesn't match.
2968 /// The pattern can be a `&str`, [`char`], or a closure that determines if
2969 /// a character matches.
2971 /// [`None`]: option/enum.Option.html#variant.None
2975 /// Simple patterns:
2978 /// let s = "Löwe 老虎 Léopard";
2980 /// assert_eq!(s.rfind('L'), Some(13));
2981 /// assert_eq!(s.rfind('é'), Some(14));
2984 /// More complex patterns with closures:
2987 /// let s = "Löwe 老虎 Léopard";
2989 /// assert_eq!(s.rfind(char::is_whitespace), Some(12));
2990 /// assert_eq!(s.rfind(char::is_lowercase), Some(20));
2993 /// Not finding the pattern:
2996 /// let s = "Löwe 老虎 Léopard";
2997 /// let x: &[_] = &['1', '2'];
2999 /// assert_eq!(s.rfind(x), None);
3001 #[stable(feature = "rust1", since = "1.0.0")]
3003 pub fn rfind<'a, P>(&'a self, pat: P) -> Option<usize>
3005 P: Pattern<'a, Searcher: ReverseSearcher<'a>>,
3007 pat.into_searcher(self).next_match_back().map(|(i, _)| i)
3010 /// An iterator over substrings of this string slice, separated by
3011 /// characters matched by a pattern.
3013 /// The pattern can be any type that implements the Pattern trait. Notable
3014 /// examples are `&str`, [`char`], and closures that determines the split.
3016 /// # Iterator behavior
3018 /// The returned iterator will be a [`DoubleEndedIterator`] if the pattern
3019 /// allows a reverse search and forward/reverse search yields the same
3020 /// elements. This is true for, e.g., [`char`], but not for `&str`.
3022 /// [`DoubleEndedIterator`]: iter/trait.DoubleEndedIterator.html
3024 /// If the pattern allows a reverse search but its results might differ
3025 /// from a forward search, the [`rsplit`] method can be used.
3027 /// [`rsplit`]: #method.rsplit
3031 /// Simple patterns:
3034 /// let v: Vec<&str> = "Mary had a little lamb".split(' ').collect();
3035 /// assert_eq!(v, ["Mary", "had", "a", "little", "lamb"]);
3037 /// let v: Vec<&str> = "".split('X').collect();
3038 /// assert_eq!(v, [""]);
3040 /// let v: Vec<&str> = "lionXXtigerXleopard".split('X').collect();
3041 /// assert_eq!(v, ["lion", "", "tiger", "leopard"]);
3043 /// let v: Vec<&str> = "lion::tiger::leopard".split("::").collect();
3044 /// assert_eq!(v, ["lion", "tiger", "leopard"]);
3046 /// let v: Vec<&str> = "abc1def2ghi".split(char::is_numeric).collect();
3047 /// assert_eq!(v, ["abc", "def", "ghi"]);
3049 /// let v: Vec<&str> = "lionXtigerXleopard".split(char::is_uppercase).collect();
3050 /// assert_eq!(v, ["lion", "tiger", "leopard"]);
3053 /// A more complex pattern, using a closure:
3056 /// let v: Vec<&str> = "abc1defXghi".split(|c| c == '1' || c == 'X').collect();
3057 /// assert_eq!(v, ["abc", "def", "ghi"]);
3060 /// If a string contains multiple contiguous separators, you will end up
3061 /// with empty strings in the output:
3064 /// let x = "||||a||b|c".to_string();
3065 /// let d: Vec<_> = x.split('|').collect();
3067 /// assert_eq!(d, &["", "", "", "", "a", "", "b", "c"]);
3070 /// Contiguous separators are separated by the empty string.
3073 /// let x = "(///)".to_string();
3074 /// let d: Vec<_> = x.split('/').collect();
3076 /// assert_eq!(d, &["(", "", "", ")"]);
3079 /// Separators at the start or end of a string are neighbored
3080 /// by empty strings.
3083 /// let d: Vec<_> = "010".split("0").collect();
3084 /// assert_eq!(d, &["", "1", ""]);
3087 /// When the empty string is used as a separator, it separates
3088 /// every character in the string, along with the beginning
3089 /// and end of the string.
3092 /// let f: Vec<_> = "rust".split("").collect();
3093 /// assert_eq!(f, &["", "r", "u", "s", "t", ""]);
3096 /// Contiguous separators can lead to possibly surprising behavior
3097 /// when whitespace is used as the separator. This code is correct:
3100 /// let x = " a b c".to_string();
3101 /// let d: Vec<_> = x.split(' ').collect();
3103 /// assert_eq!(d, &["", "", "", "", "a", "", "b", "c"]);
3106 /// It does _not_ give you:
3109 /// assert_eq!(d, &["a", "b", "c"]);
3112 /// Use [`split_whitespace`] for this behavior.
3114 /// [`split_whitespace`]: #method.split_whitespace
3115 #[stable(feature = "rust1", since = "1.0.0")]
3117 pub fn split<'a, P: Pattern<'a>>(&'a self, pat: P) -> Split<'a, P> {
3118 Split(SplitInternal {
3121 matcher: pat.into_searcher(self),
3122 allow_trailing_empty: true,
3127 /// An iterator over substrings of the given string slice, separated by
3128 /// characters matched by a pattern and yielded in reverse order.
3130 /// The pattern can be any type that implements the Pattern trait. Notable
3131 /// examples are `&str`, [`char`], and closures that determines the split.
3133 /// # Iterator behavior
3135 /// The returned iterator requires that the pattern supports a reverse
3136 /// search, and it will be a [`DoubleEndedIterator`] if a forward/reverse
3137 /// search yields the same elements.
3139 /// [`DoubleEndedIterator`]: iter/trait.DoubleEndedIterator.html
3141 /// For iterating from the front, the [`split`] method can be used.
3143 /// [`split`]: #method.split
3147 /// Simple patterns:
3150 /// let v: Vec<&str> = "Mary had a little lamb".rsplit(' ').collect();
3151 /// assert_eq!(v, ["lamb", "little", "a", "had", "Mary"]);
3153 /// let v: Vec<&str> = "".rsplit('X').collect();
3154 /// assert_eq!(v, [""]);
3156 /// let v: Vec<&str> = "lionXXtigerXleopard".rsplit('X').collect();
3157 /// assert_eq!(v, ["leopard", "tiger", "", "lion"]);
3159 /// let v: Vec<&str> = "lion::tiger::leopard".rsplit("::").collect();
3160 /// assert_eq!(v, ["leopard", "tiger", "lion"]);
3163 /// A more complex pattern, using a closure:
3166 /// let v: Vec<&str> = "abc1defXghi".rsplit(|c| c == '1' || c == 'X').collect();
3167 /// assert_eq!(v, ["ghi", "def", "abc"]);
3169 #[stable(feature = "rust1", since = "1.0.0")]
3171 pub fn rsplit<'a, P>(&'a self, pat: P) -> RSplit<'a, P>
3173 P: Pattern<'a, Searcher: ReverseSearcher<'a>>,
3175 RSplit(self.split(pat).0)
3178 /// An iterator over substrings of the given string slice, separated by
3179 /// characters matched by a pattern.
3181 /// The pattern can be any type that implements the Pattern trait. Notable
3182 /// examples are `&str`, [`char`], and closures that determines the split.
3184 /// Equivalent to [`split`], except that the trailing substring
3185 /// is skipped if empty.
3187 /// [`split`]: #method.split
3189 /// This method can be used for string data that is _terminated_,
3190 /// rather than _separated_ by a pattern.
3192 /// # Iterator behavior
3194 /// The returned iterator will be a [`DoubleEndedIterator`] if the pattern
3195 /// allows a reverse search and forward/reverse search yields the same
3196 /// elements. This is true for, e.g., [`char`], but not for `&str`.
3198 /// [`DoubleEndedIterator`]: iter/trait.DoubleEndedIterator.html
3200 /// If the pattern allows a reverse search but its results might differ
3201 /// from a forward search, the [`rsplit_terminator`] method can be used.
3203 /// [`rsplit_terminator`]: #method.rsplit_terminator
3210 /// let v: Vec<&str> = "A.B.".split_terminator('.').collect();
3211 /// assert_eq!(v, ["A", "B"]);
3213 /// let v: Vec<&str> = "A..B..".split_terminator(".").collect();
3214 /// assert_eq!(v, ["A", "", "B", ""]);
3216 #[stable(feature = "rust1", since = "1.0.0")]
3218 pub fn split_terminator<'a, P: Pattern<'a>>(&'a self, pat: P) -> SplitTerminator<'a, P> {
3219 SplitTerminator(SplitInternal {
3220 allow_trailing_empty: false,
3225 /// An iterator over substrings of `self`, separated by characters
3226 /// matched by a pattern and yielded in reverse order.
3228 /// The pattern can be any type that implements the Pattern trait. Notable
3229 /// examples are `&str`, [`char`], and closures that determines the split.
3230 /// Additional libraries might provide more complex patterns like
3231 /// regular expressions.
3233 /// Equivalent to [`split`], except that the trailing substring is
3234 /// skipped if empty.
3236 /// [`split`]: #method.split
3238 /// This method can be used for string data that is _terminated_,
3239 /// rather than _separated_ by a pattern.
3241 /// # Iterator behavior
3243 /// The returned iterator requires that the pattern supports a
3244 /// reverse search, and it will be double ended if a forward/reverse
3245 /// search yields the same elements.
3247 /// For iterating from the front, the [`split_terminator`] method can be
3250 /// [`split_terminator`]: #method.split_terminator
3255 /// let v: Vec<&str> = "A.B.".rsplit_terminator('.').collect();
3256 /// assert_eq!(v, ["B", "A"]);
3258 /// let v: Vec<&str> = "A..B..".rsplit_terminator(".").collect();
3259 /// assert_eq!(v, ["", "B", "", "A"]);
3261 #[stable(feature = "rust1", since = "1.0.0")]
3263 pub fn rsplit_terminator<'a, P>(&'a self, pat: P) -> RSplitTerminator<'a, P>
3265 P: Pattern<'a, Searcher: ReverseSearcher<'a>>,
3267 RSplitTerminator(self.split_terminator(pat).0)
3270 /// An iterator over substrings of the given string slice, separated by a
3271 /// pattern, restricted to returning at most `n` items.
3273 /// If `n` substrings are returned, the last substring (the `n`th substring)
3274 /// will contain the remainder of the string.
3276 /// The pattern can be any type that implements the Pattern trait. Notable
3277 /// examples are `&str`, [`char`], and closures that determines the split.
3279 /// # Iterator behavior
3281 /// The returned iterator will not be double ended, because it is
3282 /// not efficient to support.
3284 /// If the pattern allows a reverse search, the [`rsplitn`] method can be
3287 /// [`rsplitn`]: #method.rsplitn
3291 /// Simple patterns:
3294 /// let v: Vec<&str> = "Mary had a little lambda".splitn(3, ' ').collect();
3295 /// assert_eq!(v, ["Mary", "had", "a little lambda"]);
3297 /// let v: Vec<&str> = "lionXXtigerXleopard".splitn(3, "X").collect();
3298 /// assert_eq!(v, ["lion", "", "tigerXleopard"]);
3300 /// let v: Vec<&str> = "abcXdef".splitn(1, 'X').collect();
3301 /// assert_eq!(v, ["abcXdef"]);
3303 /// let v: Vec<&str> = "".splitn(1, 'X').collect();
3304 /// assert_eq!(v, [""]);
3307 /// A more complex pattern, using a closure:
3310 /// let v: Vec<&str> = "abc1defXghi".splitn(2, |c| c == '1' || c == 'X').collect();
3311 /// assert_eq!(v, ["abc", "defXghi"]);
3313 #[stable(feature = "rust1", since = "1.0.0")]
3315 pub fn splitn<'a, P: Pattern<'a>>(&'a self, n: usize, pat: P) -> SplitN<'a, P> {
3316 SplitN(SplitNInternal {
3317 iter: self.split(pat).0,
3322 /// An iterator over substrings of this string slice, separated by a
3323 /// pattern, starting from the end of the string, restricted to returning
3324 /// at most `n` items.
3326 /// If `n` substrings are returned, the last substring (the `n`th substring)
3327 /// will contain the remainder of the string.
3329 /// The pattern can be any type that implements the Pattern trait. Notable
3330 /// examples are `&str`, [`char`], and closures that determines the split.
3332 /// # Iterator behavior
3334 /// The returned iterator will not be double ended, because it is not
3335 /// efficient to support.
3337 /// For splitting from the front, the [`splitn`] method can be used.
3339 /// [`splitn`]: #method.splitn
3343 /// Simple patterns:
3346 /// let v: Vec<&str> = "Mary had a little lamb".rsplitn(3, ' ').collect();
3347 /// assert_eq!(v, ["lamb", "little", "Mary had a"]);
3349 /// let v: Vec<&str> = "lionXXtigerXleopard".rsplitn(3, 'X').collect();
3350 /// assert_eq!(v, ["leopard", "tiger", "lionX"]);
3352 /// let v: Vec<&str> = "lion::tiger::leopard".rsplitn(2, "::").collect();
3353 /// assert_eq!(v, ["leopard", "lion::tiger"]);
3356 /// A more complex pattern, using a closure:
3359 /// let v: Vec<&str> = "abc1defXghi".rsplitn(2, |c| c == '1' || c == 'X').collect();
3360 /// assert_eq!(v, ["ghi", "abc1def"]);
3362 #[stable(feature = "rust1", since = "1.0.0")]
3364 pub fn rsplitn<'a, P>(&'a self, n: usize, pat: P) -> RSplitN<'a, P>
3366 P: Pattern<'a, Searcher: ReverseSearcher<'a>>,
3368 RSplitN(self.splitn(n, pat).0)
3371 /// An iterator over the disjoint matches of a pattern within the given string
3374 /// The pattern can be any type that implements the Pattern trait. Notable
3375 /// examples are `&str`, [`char`], and closures that determines the split.
3377 /// # Iterator behavior
3379 /// The returned iterator will be a [`DoubleEndedIterator`] if the pattern
3380 /// allows a reverse search and forward/reverse search yields the same
3381 /// elements. This is true for, e.g., [`char`], but not for `&str`.
3383 /// [`DoubleEndedIterator`]: iter/trait.DoubleEndedIterator.html
3385 /// If the pattern allows a reverse search but its results might differ
3386 /// from a forward search, the [`rmatches`] method can be used.
3388 /// [`rmatches`]: #method.rmatches
3395 /// let v: Vec<&str> = "abcXXXabcYYYabc".matches("abc").collect();
3396 /// assert_eq!(v, ["abc", "abc", "abc"]);
3398 /// let v: Vec<&str> = "1abc2abc3".matches(char::is_numeric).collect();
3399 /// assert_eq!(v, ["1", "2", "3"]);
3401 #[stable(feature = "str_matches", since = "1.2.0")]
3403 pub fn matches<'a, P: Pattern<'a>>(&'a self, pat: P) -> Matches<'a, P> {
3404 Matches(MatchesInternal(pat.into_searcher(self)))
3407 /// An iterator over the disjoint matches of a pattern within this string slice,
3408 /// yielded in reverse order.
3410 /// The pattern can be a `&str`, [`char`], or a closure that determines if
3411 /// a character matches.
3413 /// # Iterator behavior
3415 /// The returned iterator requires that the pattern supports a reverse
3416 /// search, and it will be a [`DoubleEndedIterator`] if a forward/reverse
3417 /// search yields the same elements.
3419 /// [`DoubleEndedIterator`]: iter/trait.DoubleEndedIterator.html
3421 /// For iterating from the front, the [`matches`] method can be used.
3423 /// [`matches`]: #method.matches
3430 /// let v: Vec<&str> = "abcXXXabcYYYabc".rmatches("abc").collect();
3431 /// assert_eq!(v, ["abc", "abc", "abc"]);
3433 /// let v: Vec<&str> = "1abc2abc3".rmatches(char::is_numeric).collect();
3434 /// assert_eq!(v, ["3", "2", "1"]);
3436 #[stable(feature = "str_matches", since = "1.2.0")]
3438 pub fn rmatches<'a, P>(&'a self, pat: P) -> RMatches<'a, P>
3440 P: Pattern<'a, Searcher: ReverseSearcher<'a>>,
3442 RMatches(self.matches(pat).0)
3445 /// An iterator over the disjoint matches of a pattern within this string
3446 /// slice as well as the index that the match starts at.
3448 /// For matches of `pat` within `self` that overlap, only the indices
3449 /// corresponding to the first match are returned.
3451 /// The pattern can be a `&str`, [`char`], or a closure that determines
3452 /// if a character matches.
3454 /// # Iterator behavior
3456 /// The returned iterator will be a [`DoubleEndedIterator`] if the pattern
3457 /// allows a reverse search and forward/reverse search yields the same
3458 /// elements. This is true for, e.g., [`char`], but not for `&str`.
3460 /// [`DoubleEndedIterator`]: iter/trait.DoubleEndedIterator.html
3462 /// If the pattern allows a reverse search but its results might differ
3463 /// from a forward search, the [`rmatch_indices`] method can be used.
3465 /// [`rmatch_indices`]: #method.rmatch_indices
3472 /// let v: Vec<_> = "abcXXXabcYYYabc".match_indices("abc").collect();
3473 /// assert_eq!(v, [(0, "abc"), (6, "abc"), (12, "abc")]);
3475 /// let v: Vec<_> = "1abcabc2".match_indices("abc").collect();
3476 /// assert_eq!(v, [(1, "abc"), (4, "abc")]);
3478 /// let v: Vec<_> = "ababa".match_indices("aba").collect();
3479 /// assert_eq!(v, [(0, "aba")]); // only the first `aba`
3481 #[stable(feature = "str_match_indices", since = "1.5.0")]
3483 pub fn match_indices<'a, P: Pattern<'a>>(&'a self, pat: P) -> MatchIndices<'a, P> {
3484 MatchIndices(MatchIndicesInternal(pat.into_searcher(self)))
3487 /// An iterator over the disjoint matches of a pattern within `self`,
3488 /// yielded in reverse order along with the index of the match.
3490 /// For matches of `pat` within `self` that overlap, only the indices
3491 /// corresponding to the last match are returned.
3493 /// The pattern can be a `&str`, [`char`], or a closure that determines if a
3494 /// character matches.
3496 /// # Iterator behavior
3498 /// The returned iterator requires that the pattern supports a reverse
3499 /// search, and it will be a [`DoubleEndedIterator`] if a forward/reverse
3500 /// search yields the same elements.
3502 /// [`DoubleEndedIterator`]: iter/trait.DoubleEndedIterator.html
3504 /// For iterating from the front, the [`match_indices`] method can be used.
3506 /// [`match_indices`]: #method.match_indices
3513 /// let v: Vec<_> = "abcXXXabcYYYabc".rmatch_indices("abc").collect();
3514 /// assert_eq!(v, [(12, "abc"), (6, "abc"), (0, "abc")]);
3516 /// let v: Vec<_> = "1abcabc2".rmatch_indices("abc").collect();
3517 /// assert_eq!(v, [(4, "abc"), (1, "abc")]);
3519 /// let v: Vec<_> = "ababa".rmatch_indices("aba").collect();
3520 /// assert_eq!(v, [(2, "aba")]); // only the last `aba`
3522 #[stable(feature = "str_match_indices", since = "1.5.0")]
3524 pub fn rmatch_indices<'a, P>(&'a self, pat: P) -> RMatchIndices<'a, P>
3526 P: Pattern<'a, Searcher: ReverseSearcher<'a>>,
3528 RMatchIndices(self.match_indices(pat).0)
3531 /// Returns a string slice with leading and trailing whitespace removed.
3533 /// 'Whitespace' is defined according to the terms of the Unicode Derived
3534 /// Core Property `White_Space`.
3541 /// let s = " Hello\tworld\t";
3543 /// assert_eq!("Hello\tworld", s.trim());
3545 #[must_use = "this returns the trimmed string as a slice, \
3546 without modifying the original"]
3547 #[stable(feature = "rust1", since = "1.0.0")]
3548 pub fn trim(&self) -> &str {
3549 self.trim_matches(|c: char| c.is_whitespace())
3552 /// Returns a string slice with leading whitespace removed.
3554 /// 'Whitespace' is defined according to the terms of the Unicode Derived
3555 /// Core Property `White_Space`.
3557 /// # Text directionality
3559 /// A string is a sequence of bytes. `start` in this context means the first
3560 /// position of that byte string; for a left-to-right language like English or
3561 /// Russian, this will be left side, and for right-to-left languages like
3562 /// Arabic or Hebrew, this will be the right side.
3569 /// let s = " Hello\tworld\t";
3570 /// assert_eq!("Hello\tworld\t", s.trim_start());
3576 /// let s = " English ";
3577 /// assert!(Some('E') == s.trim_start().chars().next());
3579 /// let s = " עברית ";
3580 /// assert!(Some('ע') == s.trim_start().chars().next());
3582 #[must_use = "this returns the trimmed string as a new slice, \
3583 without modifying the original"]
3584 #[stable(feature = "trim_direction", since = "1.30.0")]
3585 pub fn trim_start(&self) -> &str {
3586 self.trim_start_matches(|c: char| c.is_whitespace())
3589 /// Returns a string slice with trailing whitespace removed.
3591 /// 'Whitespace' is defined according to the terms of the Unicode Derived
3592 /// Core Property `White_Space`.
3594 /// # Text directionality
3596 /// A string is a sequence of bytes. `end` in this context means the last
3597 /// position of that byte string; for a left-to-right language like English or
3598 /// Russian, this will be right side, and for right-to-left languages like
3599 /// Arabic or Hebrew, this will be the left side.
3606 /// let s = " Hello\tworld\t";
3607 /// assert_eq!(" Hello\tworld", s.trim_end());
3613 /// let s = " English ";
3614 /// assert!(Some('h') == s.trim_end().chars().rev().next());
3616 /// let s = " עברית ";
3617 /// assert!(Some('ת') == s.trim_end().chars().rev().next());
3619 #[must_use = "this returns the trimmed string as a new slice, \
3620 without modifying the original"]
3621 #[stable(feature = "trim_direction", since = "1.30.0")]
3622 pub fn trim_end(&self) -> &str {
3623 self.trim_end_matches(|c: char| c.is_whitespace())
3626 /// Returns a string slice with leading whitespace removed.
3628 /// 'Whitespace' is defined according to the terms of the Unicode Derived
3629 /// Core Property `White_Space`.
3631 /// # Text directionality
3633 /// A string is a sequence of bytes. 'Left' in this context means the first
3634 /// position of that byte string; for a language like Arabic or Hebrew
3635 /// which are 'right to left' rather than 'left to right', this will be
3636 /// the _right_ side, not the left.
3643 /// let s = " Hello\tworld\t";
3645 /// assert_eq!("Hello\tworld\t", s.trim_left());
3651 /// let s = " English";
3652 /// assert!(Some('E') == s.trim_left().chars().next());
3654 /// let s = " עברית";
3655 /// assert!(Some('ע') == s.trim_left().chars().next());
3657 #[stable(feature = "rust1", since = "1.0.0")]
3660 reason = "superseded by `trim_start`",
3661 suggestion = "trim_start",
3663 pub fn trim_left(&self) -> &str {
3667 /// Returns a string slice with trailing whitespace removed.
3669 /// 'Whitespace' is defined according to the terms of the Unicode Derived
3670 /// Core Property `White_Space`.
3672 /// # Text directionality
3674 /// A string is a sequence of bytes. 'Right' in this context means the last
3675 /// position of that byte string; for a language like Arabic or Hebrew
3676 /// which are 'right to left' rather than 'left to right', this will be
3677 /// the _left_ side, not the right.
3684 /// let s = " Hello\tworld\t";
3686 /// assert_eq!(" Hello\tworld", s.trim_right());
3692 /// let s = "English ";
3693 /// assert!(Some('h') == s.trim_right().chars().rev().next());
3695 /// let s = "עברית ";
3696 /// assert!(Some('ת') == s.trim_right().chars().rev().next());
3698 #[stable(feature = "rust1", since = "1.0.0")]
3701 reason = "superseded by `trim_end`",
3702 suggestion = "trim_end",
3704 pub fn trim_right(&self) -> &str {
3708 /// Returns a string slice with all prefixes and suffixes that match a
3709 /// pattern repeatedly removed.
3711 /// The pattern can be a [`char`] or a closure that determines if a
3712 /// character matches.
3716 /// Simple patterns:
3719 /// assert_eq!("11foo1bar11".trim_matches('1'), "foo1bar");
3720 /// assert_eq!("123foo1bar123".trim_matches(char::is_numeric), "foo1bar");
3722 /// let x: &[_] = &['1', '2'];
3723 /// assert_eq!("12foo1bar12".trim_matches(x), "foo1bar");
3726 /// A more complex pattern, using a closure:
3729 /// assert_eq!("1foo1barXX".trim_matches(|c| c == '1' || c == 'X'), "foo1bar");
3731 #[must_use = "this returns the trimmed string as a new slice, \
3732 without modifying the original"]
3733 #[stable(feature = "rust1", since = "1.0.0")]
3734 pub fn trim_matches<'a, P>(&'a self, pat: P) -> &'a str
3736 P: Pattern<'a, Searcher: DoubleEndedSearcher<'a>>,
3740 let mut matcher = pat.into_searcher(self);
3741 if let Some((a, b)) = matcher.next_reject() {
3743 j = b; // Remember earliest known match, correct it below if
3744 // last match is different
3746 if let Some((_, b)) = matcher.next_reject_back() {
3750 // Searcher is known to return valid indices
3751 self.get_unchecked(i..j)
3755 /// Returns a string slice with all prefixes that match a pattern
3756 /// repeatedly removed.
3758 /// The pattern can be a `&str`, [`char`], or a closure that determines if
3759 /// a character matches.
3761 /// # Text directionality
3763 /// A string is a sequence of bytes. `start` in this context means the first
3764 /// position of that byte string; for a left-to-right language like English or
3765 /// Russian, this will be left side, and for right-to-left languages like
3766 /// Arabic or Hebrew, this will be the right side.
3773 /// assert_eq!("11foo1bar11".trim_start_matches('1'), "foo1bar11");
3774 /// assert_eq!("123foo1bar123".trim_start_matches(char::is_numeric), "foo1bar123");
3776 /// let x: &[_] = &['1', '2'];
3777 /// assert_eq!("12foo1bar12".trim_start_matches(x), "foo1bar12");
3779 #[must_use = "this returns the trimmed string as a new slice, \
3780 without modifying the original"]
3781 #[stable(feature = "trim_direction", since = "1.30.0")]
3782 pub fn trim_start_matches<'a, P: Pattern<'a>>(&'a self, pat: P) -> &'a str {
3783 let mut i = self.len();
3784 let mut matcher = pat.into_searcher(self);
3785 if let Some((a, _)) = matcher.next_reject() {
3789 // Searcher is known to return valid indices
3790 self.get_unchecked(i..self.len())
3794 /// Returns a string slice with the prefix removed.
3796 /// If the string starts with the pattern `prefix`, `Some` is returned with the substring where
3797 /// the prefix is removed. Unlike `trim_start_matches`, this method removes the prefix exactly
3800 /// If the string does not start with `prefix`, it is removed.
3805 /// #![feature(str_strip)]
3807 /// assert_eq!("foobar".strip_prefix("foo"), Some("bar"));
3808 /// assert_eq!("foobar".strip_prefix("bar"), None);
3809 /// assert_eq!("foofoo".strip_prefix("foo"), Some("foo"));
3811 #[must_use = "this returns the remaining substring as a new slice, \
3812 without modifying the original"]
3813 #[unstable(feature = "str_strip", reason = "newly added", issue = "0")]
3814 pub fn strip_prefix<'a, P: Pattern<'a>>(&'a self, prefix: P) -> Option<&'a str> {
3815 let mut matcher = prefix.into_searcher(self);
3816 if let SearchStep::Match(start, len) = matcher.next() {
3817 debug_assert_eq!(start, 0);
3819 Some(self.get_unchecked(len..))
3826 /// Returns a string slice with the suffix removed.
3828 /// If the string starts with the pattern `suffix`, `Some` is returned with the substring where
3829 /// the suffix is removed. Unlike `trim_end_matches`, this method removes the suffix exactly
3832 /// If the string does not start with `suffix`, it is removed.
3837 /// #![feature(str_strip)]
3838 /// assert_eq!("barfoo".strip_suffix("foo"), Some("bar"));
3839 /// assert_eq!("barfoo".strip_suffix("bar"), None);
3840 /// assert_eq!("foofoo".strip_suffix("foo"), Some("foo"));
3842 #[must_use = "this returns the remaining substring as a new slice, \
3843 without modifying the original"]
3844 #[unstable(feature = "str_strip", reason = "newly added", issue = "0")]
3845 pub fn strip_suffix<'a, P>(&'a self, suffix: P) -> Option<&'a str>
3848 <P as Pattern<'a>>::Searcher: ReverseSearcher<'a>,
3850 let mut matcher = suffix.into_searcher(self);
3851 if let SearchStep::Match(start, end) = matcher.next_back() {
3852 debug_assert_eq!(end, self.len());
3854 Some(self.get_unchecked(..start))
3861 /// Returns a string slice with all suffixes that match a pattern
3862 /// repeatedly removed.
3864 /// The pattern can be a `&str`, [`char`], or a closure that
3865 /// determines if a character matches.
3867 /// # Text directionality
3869 /// A string is a sequence of bytes. `end` in this context means the last
3870 /// position of that byte string; for a left-to-right language like English or
3871 /// Russian, this will be right side, and for right-to-left languages like
3872 /// Arabic or Hebrew, this will be the left side.
3876 /// Simple patterns:
3879 /// assert_eq!("11foo1bar11".trim_end_matches('1'), "11foo1bar");
3880 /// assert_eq!("123foo1bar123".trim_end_matches(char::is_numeric), "123foo1bar");
3882 /// let x: &[_] = &['1', '2'];
3883 /// assert_eq!("12foo1bar12".trim_end_matches(x), "12foo1bar");
3886 /// A more complex pattern, using a closure:
3889 /// assert_eq!("1fooX".trim_end_matches(|c| c == '1' || c == 'X'), "1foo");
3891 #[must_use = "this returns the trimmed string as a new slice, \
3892 without modifying the original"]
3893 #[stable(feature = "trim_direction", since = "1.30.0")]
3894 pub fn trim_end_matches<'a, P>(&'a self, pat: P) -> &'a str
3896 P: Pattern<'a, Searcher: ReverseSearcher<'a>>,
3899 let mut matcher = pat.into_searcher(self);
3900 if let Some((_, b)) = matcher.next_reject_back() {
3904 // Searcher is known to return valid indices
3905 self.get_unchecked(0..j)
3909 /// Returns a string slice with all prefixes that match a pattern
3910 /// repeatedly removed.
3912 /// The pattern can be a `&str`, [`char`], or a closure that determines if
3913 /// a character matches.
3915 /// [`char`]: primitive.char.html
3917 /// # Text directionality
3919 /// A string is a sequence of bytes. 'Left' in this context means the first
3920 /// position of that byte string; for a language like Arabic or Hebrew
3921 /// which are 'right to left' rather than 'left to right', this will be
3922 /// the _right_ side, not the left.
3929 /// assert_eq!("11foo1bar11".trim_left_matches('1'), "foo1bar11");
3930 /// assert_eq!("123foo1bar123".trim_left_matches(char::is_numeric), "foo1bar123");
3932 /// let x: &[_] = &['1', '2'];
3933 /// assert_eq!("12foo1bar12".trim_left_matches(x), "foo1bar12");
3935 #[stable(feature = "rust1", since = "1.0.0")]
3938 reason = "superseded by `trim_start_matches`",
3939 suggestion = "trim_start_matches",
3941 pub fn trim_left_matches<'a, P: Pattern<'a>>(&'a self, pat: P) -> &'a str {
3942 self.trim_start_matches(pat)
3945 /// Returns a string slice with all suffixes that match a pattern
3946 /// repeatedly removed.
3948 /// The pattern can be a `&str`, [`char`], or a closure that
3949 /// determines if a character matches.
3951 /// [`char`]: primitive.char.html
3953 /// # Text directionality
3955 /// A string is a sequence of bytes. 'Right' in this context means the last
3956 /// position of that byte string; for a language like Arabic or Hebrew
3957 /// which are 'right to left' rather than 'left to right', this will be
3958 /// the _left_ side, not the right.
3962 /// Simple patterns:
3965 /// assert_eq!("11foo1bar11".trim_right_matches('1'), "11foo1bar");
3966 /// assert_eq!("123foo1bar123".trim_right_matches(char::is_numeric), "123foo1bar");
3968 /// let x: &[_] = &['1', '2'];
3969 /// assert_eq!("12foo1bar12".trim_right_matches(x), "12foo1bar");
3972 /// A more complex pattern, using a closure:
3975 /// assert_eq!("1fooX".trim_right_matches(|c| c == '1' || c == 'X'), "1foo");
3977 #[stable(feature = "rust1", since = "1.0.0")]
3980 reason = "superseded by `trim_end_matches`",
3981 suggestion = "trim_end_matches",
3983 pub fn trim_right_matches<'a, P>(&'a self, pat: P) -> &'a str
3985 P: Pattern<'a, Searcher: ReverseSearcher<'a>>,
3987 self.trim_end_matches(pat)
3990 /// Parses this string slice into another type.
3992 /// Because `parse` is so general, it can cause problems with type
3993 /// inference. As such, `parse` is one of the few times you'll see
3994 /// the syntax affectionately known as the 'turbofish': `::<>`. This
3995 /// helps the inference algorithm understand specifically which type
3996 /// you're trying to parse into.
3998 /// `parse` can parse any type that implements the [`FromStr`] trait.
4000 /// [`FromStr`]: str/trait.FromStr.html
4004 /// Will return [`Err`] if it's not possible to parse this string slice into
4005 /// the desired type.
4007 /// [`Err`]: str/trait.FromStr.html#associatedtype.Err
4014 /// let four: u32 = "4".parse().unwrap();
4016 /// assert_eq!(4, four);
4019 /// Using the 'turbofish' instead of annotating `four`:
4022 /// let four = "4".parse::<u32>();
4024 /// assert_eq!(Ok(4), four);
4027 /// Failing to parse:
4030 /// let nope = "j".parse::<u32>();
4032 /// assert!(nope.is_err());
4035 #[stable(feature = "rust1", since = "1.0.0")]
4036 pub fn parse<F: FromStr>(&self) -> Result<F, F::Err> {
4037 FromStr::from_str(self)
4040 /// Checks if all characters in this string are within the ASCII range.
4045 /// let ascii = "hello!\n";
4046 /// let non_ascii = "Grüße, Jürgen ❤";
4048 /// assert!(ascii.is_ascii());
4049 /// assert!(!non_ascii.is_ascii());
4051 #[stable(feature = "ascii_methods_on_intrinsics", since = "1.23.0")]
4053 pub fn is_ascii(&self) -> bool {
4054 // We can treat each byte as character here: all multibyte characters
4055 // start with a byte that is not in the ascii range, so we will stop
4057 self.bytes().all(|b| b.is_ascii())
4060 /// Checks that two strings are an ASCII case-insensitive match.
4062 /// Same as `to_ascii_lowercase(a) == to_ascii_lowercase(b)`,
4063 /// but without allocating and copying temporaries.
4068 /// assert!("Ferris".eq_ignore_ascii_case("FERRIS"));
4069 /// assert!("Ferrös".eq_ignore_ascii_case("FERRöS"));
4070 /// assert!(!"Ferrös".eq_ignore_ascii_case("FERRÖS"));
4072 #[stable(feature = "ascii_methods_on_intrinsics", since = "1.23.0")]
4074 pub fn eq_ignore_ascii_case(&self, other: &str) -> bool {
4075 self.as_bytes().eq_ignore_ascii_case(other.as_bytes())
4078 /// Converts this string to its ASCII upper case equivalent in-place.
4080 /// ASCII letters 'a' to 'z' are mapped to 'A' to 'Z',
4081 /// but non-ASCII letters are unchanged.
4083 /// To return a new uppercased value without modifying the existing one, use
4084 /// [`to_ascii_uppercase`].
4086 /// [`to_ascii_uppercase`]: #method.to_ascii_uppercase
4091 /// let mut s = String::from("Grüße, Jürgen ❤");
4093 /// s.make_ascii_uppercase();
4095 /// assert_eq!("GRüßE, JüRGEN ❤", s);
4097 #[stable(feature = "ascii_methods_on_intrinsics", since = "1.23.0")]
4098 pub fn make_ascii_uppercase(&mut self) {
4099 let me = unsafe { self.as_bytes_mut() };
4100 me.make_ascii_uppercase()
4103 /// Converts this string to its ASCII lower case equivalent in-place.
4105 /// ASCII letters 'A' to 'Z' are mapped to 'a' to 'z',
4106 /// but non-ASCII letters are unchanged.
4108 /// To return a new lowercased value without modifying the existing one, use
4109 /// [`to_ascii_lowercase`].
4111 /// [`to_ascii_lowercase`]: #method.to_ascii_lowercase
4116 /// let mut s = String::from("GRÜßE, JÜRGEN ❤");
4118 /// s.make_ascii_lowercase();
4120 /// assert_eq!("grÜße, jÜrgen ❤", s);
4122 #[stable(feature = "ascii_methods_on_intrinsics", since = "1.23.0")]
4123 pub fn make_ascii_lowercase(&mut self) {
4124 let me = unsafe { self.as_bytes_mut() };
4125 me.make_ascii_lowercase()
4128 /// Return an iterator that escapes each char in `self` with [`char::escape_debug`].
4130 /// Note: only extended grapheme codepoints that begin the string will be
4133 /// [`char::escape_debug`]: ../std/primitive.char.html#method.escape_debug
4140 /// for c in "❤\n!".escape_debug() {
4141 /// print!("{}", c);
4146 /// Using `println!` directly:
4149 /// println!("{}", "❤\n!".escape_debug());
4153 /// Both are equivalent to:
4156 /// println!("❤\\n!");
4159 /// Using `to_string`:
4162 /// assert_eq!("❤\n!".escape_debug().to_string(), "❤\\n!");
4164 #[stable(feature = "str_escape", since = "1.34.0")]
4165 pub fn escape_debug(&self) -> EscapeDebug<'_> {
4166 let mut chars = self.chars();
4169 .map(|first| first.escape_debug_ext(true))
4172 .chain(chars.flat_map(CharEscapeDebugContinue))
4176 /// Return an iterator that escapes each char in `self` with [`char::escape_default`].
4178 /// [`char::escape_default`]: ../std/primitive.char.html#method.escape_default
4185 /// for c in "❤\n!".escape_default() {
4186 /// print!("{}", c);
4191 /// Using `println!` directly:
4194 /// println!("{}", "❤\n!".escape_default());
4198 /// Both are equivalent to:
4201 /// println!("\\u{{2764}}\\n!");
4204 /// Using `to_string`:
4207 /// assert_eq!("❤\n!".escape_default().to_string(), "\\u{2764}\\n!");
4209 #[stable(feature = "str_escape", since = "1.34.0")]
4210 pub fn escape_default(&self) -> EscapeDefault<'_> {
4211 EscapeDefault { inner: self.chars().flat_map(CharEscapeDefault) }
4214 /// Return an iterator that escapes each char in `self` with [`char::escape_unicode`].
4216 /// [`char::escape_unicode`]: ../std/primitive.char.html#method.escape_unicode
4223 /// for c in "❤\n!".escape_unicode() {
4224 /// print!("{}", c);
4229 /// Using `println!` directly:
4232 /// println!("{}", "❤\n!".escape_unicode());
4236 /// Both are equivalent to:
4239 /// println!("\\u{{2764}}\\u{{a}}\\u{{21}}");
4242 /// Using `to_string`:
4245 /// assert_eq!("❤\n!".escape_unicode().to_string(), "\\u{2764}\\u{a}\\u{21}");
4247 #[stable(feature = "str_escape", since = "1.34.0")]
4248 pub fn escape_unicode(&self) -> EscapeUnicode<'_> {
4249 EscapeUnicode { inner: self.chars().flat_map(CharEscapeUnicode) }
4255 struct CharEscapeDebugContinue impl Fn = |c: char| -> char::EscapeDebug {
4256 c.escape_debug_ext(false)
4260 struct CharEscapeUnicode impl Fn = |c: char| -> char::EscapeUnicode {
4264 struct CharEscapeDefault impl Fn = |c: char| -> char::EscapeDefault {
4269 #[stable(feature = "rust1", since = "1.0.0")]
4270 impl AsRef<[u8]> for str {
4272 fn as_ref(&self) -> &[u8] {
4277 #[stable(feature = "rust1", since = "1.0.0")]
4278 impl Default for &str {
4279 /// Creates an empty str
4280 fn default() -> Self { "" }
4283 #[stable(feature = "default_mut_str", since = "1.28.0")]
4284 impl Default for &mut str {
4285 /// Creates an empty mutable str
4286 fn default() -> Self { unsafe { from_utf8_unchecked_mut(&mut []) } }
4289 /// An iterator over the non-whitespace substrings of a string,
4290 /// separated by any amount of whitespace.
4292 /// This struct is created by the [`split_whitespace`] method on [`str`].
4293 /// See its documentation for more.
4295 /// [`split_whitespace`]: ../../std/primitive.str.html#method.split_whitespace
4296 /// [`str`]: ../../std/primitive.str.html
4297 #[stable(feature = "split_whitespace", since = "1.1.0")]
4298 #[derive(Clone, Debug)]
4299 pub struct SplitWhitespace<'a> {
4300 inner: Filter<Split<'a, IsWhitespace>, IsNotEmpty>,
4303 /// An iterator over the non-ASCII-whitespace substrings of a string,
4304 /// separated by any amount of ASCII whitespace.
4306 /// This struct is created by the [`split_ascii_whitespace`] method on [`str`].
4307 /// See its documentation for more.
4309 /// [`split_ascii_whitespace`]: ../../std/primitive.str.html#method.split_ascii_whitespace
4310 /// [`str`]: ../../std/primitive.str.html
4311 #[stable(feature = "split_ascii_whitespace", since = "1.34.0")]
4312 #[derive(Clone, Debug)]
4313 pub struct SplitAsciiWhitespace<'a> {
4314 inner: Map<Filter<SliceSplit<'a, u8, IsAsciiWhitespace>, BytesIsNotEmpty>, UnsafeBytesToStr>,
4319 struct IsWhitespace impl Fn = |c: char| -> bool {
4324 struct IsAsciiWhitespace impl Fn = |byte: &u8| -> bool {
4325 byte.is_ascii_whitespace()
4329 struct IsNotEmpty impl<'a, 'b> Fn = |s: &'a &'b str| -> bool {
4334 struct BytesIsNotEmpty impl<'a, 'b> Fn = |s: &'a &'b [u8]| -> bool {
4339 struct UnsafeBytesToStr impl<'a> Fn = |bytes: &'a [u8]| -> &'a str {
4340 unsafe { from_utf8_unchecked(bytes) }
4344 #[stable(feature = "split_whitespace", since = "1.1.0")]
4345 impl<'a> Iterator for SplitWhitespace<'a> {
4346 type Item = &'a str;
4349 fn next(&mut self) -> Option<&'a str> {
4354 fn size_hint(&self) -> (usize, Option<usize>) {
4355 self.inner.size_hint()
4359 fn last(mut self) -> Option<&'a str> {
4364 #[stable(feature = "split_whitespace", since = "1.1.0")]
4365 impl<'a> DoubleEndedIterator for SplitWhitespace<'a> {
4367 fn next_back(&mut self) -> Option<&'a str> {
4368 self.inner.next_back()
4372 #[stable(feature = "fused", since = "1.26.0")]
4373 impl FusedIterator for SplitWhitespace<'_> {}
4375 #[stable(feature = "split_ascii_whitespace", since = "1.34.0")]
4376 impl<'a> Iterator for SplitAsciiWhitespace<'a> {
4377 type Item = &'a str;
4380 fn next(&mut self) -> Option<&'a str> {
4385 fn size_hint(&self) -> (usize, Option<usize>) {
4386 self.inner.size_hint()
4390 fn last(mut self) -> Option<&'a str> {
4395 #[stable(feature = "split_ascii_whitespace", since = "1.34.0")]
4396 impl<'a> DoubleEndedIterator for SplitAsciiWhitespace<'a> {
4398 fn next_back(&mut self) -> Option<&'a str> {
4399 self.inner.next_back()
4403 #[stable(feature = "split_ascii_whitespace", since = "1.34.0")]
4404 impl FusedIterator for SplitAsciiWhitespace<'_> {}
4406 /// An iterator of [`u16`] over the string encoded as UTF-16.
4408 /// [`u16`]: ../../std/primitive.u16.html
4410 /// This struct is created by the [`encode_utf16`] method on [`str`].
4411 /// See its documentation for more.
4413 /// [`encode_utf16`]: ../../std/primitive.str.html#method.encode_utf16
4414 /// [`str`]: ../../std/primitive.str.html
4416 #[stable(feature = "encode_utf16", since = "1.8.0")]
4417 pub struct EncodeUtf16<'a> {
4422 #[stable(feature = "collection_debug", since = "1.17.0")]
4423 impl fmt::Debug for EncodeUtf16<'_> {
4424 fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
4425 f.pad("EncodeUtf16 { .. }")
4429 #[stable(feature = "encode_utf16", since = "1.8.0")]
4430 impl<'a> Iterator for EncodeUtf16<'a> {
4434 fn next(&mut self) -> Option<u16> {
4435 if self.extra != 0 {
4436 let tmp = self.extra;
4441 let mut buf = [0; 2];
4442 self.chars.next().map(|ch| {
4443 let n = ch.encode_utf16(&mut buf).len();
4445 self.extra = buf[1];
4452 fn size_hint(&self) -> (usize, Option<usize>) {
4453 let (low, high) = self.chars.size_hint();
4454 // every char gets either one u16 or two u16,
4455 // so this iterator is between 1 or 2 times as
4456 // long as the underlying iterator.
4457 (low, high.and_then(|n| n.checked_mul(2)))
4461 #[stable(feature = "fused", since = "1.26.0")]
4462 impl FusedIterator for EncodeUtf16<'_> {}
4464 /// The return type of [`str::escape_debug`].
4466 /// [`str::escape_debug`]: ../../std/primitive.str.html#method.escape_debug
4467 #[stable(feature = "str_escape", since = "1.34.0")]
4468 #[derive(Clone, Debug)]
4469 pub struct EscapeDebug<'a> {
4471 Flatten<option::IntoIter<char::EscapeDebug>>,
4472 FlatMap<Chars<'a>, char::EscapeDebug, CharEscapeDebugContinue>
4476 /// The return type of [`str::escape_default`].
4478 /// [`str::escape_default`]: ../../std/primitive.str.html#method.escape_default
4479 #[stable(feature = "str_escape", since = "1.34.0")]
4480 #[derive(Clone, Debug)]
4481 pub struct EscapeDefault<'a> {
4482 inner: FlatMap<Chars<'a>, char::EscapeDefault, CharEscapeDefault>,
4485 /// The return type of [`str::escape_unicode`].
4487 /// [`str::escape_unicode`]: ../../std/primitive.str.html#method.escape_unicode
4488 #[stable(feature = "str_escape", since = "1.34.0")]
4489 #[derive(Clone, Debug)]
4490 pub struct EscapeUnicode<'a> {
4491 inner: FlatMap<Chars<'a>, char::EscapeUnicode, CharEscapeUnicode>,
4494 macro_rules! escape_types_impls {
4495 ($( $Name: ident ),+) => {$(
4496 #[stable(feature = "str_escape", since = "1.34.0")]
4497 impl<'a> fmt::Display for $Name<'a> {
4498 fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
4499 self.clone().try_for_each(|c| f.write_char(c))
4503 #[stable(feature = "str_escape", since = "1.34.0")]
4504 impl<'a> Iterator for $Name<'a> {
4508 fn next(&mut self) -> Option<char> { self.inner.next() }
4511 fn size_hint(&self) -> (usize, Option<usize>) { self.inner.size_hint() }
4514 fn try_fold<Acc, Fold, R>(&mut self, init: Acc, fold: Fold) -> R where
4515 Self: Sized, Fold: FnMut(Acc, Self::Item) -> R, R: Try<Ok=Acc>
4517 self.inner.try_fold(init, fold)
4521 fn fold<Acc, Fold>(self, init: Acc, fold: Fold) -> Acc
4522 where Fold: FnMut(Acc, Self::Item) -> Acc,
4524 self.inner.fold(init, fold)
4528 #[stable(feature = "str_escape", since = "1.34.0")]
4529 impl<'a> FusedIterator for $Name<'a> {}
4533 escape_types_impls!(EscapeDebug, EscapeDefault, EscapeUnicode);