1 // Copyright 2012-2014 The Rust Project Developers. See the COPYRIGHT
2 // file at the top-level directory of this distribution and at
3 // http://rust-lang.org/COPYRIGHT.
5 // Licensed under the Apache License, Version 2.0 <LICENSE-APACHE or
6 // http://www.apache.org/licenses/LICENSE-2.0> or the MIT license
7 // <LICENSE-MIT or http://opensource.org/licenses/MIT>, at your
8 // option. This file may not be copied, modified, or distributed
9 // except according to those terms.
11 // ignore-lexer-test FIXME #15679
13 //! String manipulation
15 //! For more details, see std::str
17 #![doc(primitive = "str")]
19 use self::Searcher::{Naive, TwoWay, TwoWayLong};
24 use iter::ExactSizeIterator;
25 use iter::{Map, Iterator, IteratorExt, DoubleEndedIterator};
30 use option::Option::{self, None, Some};
32 use raw::{Repr, Slice};
33 use result::Result::{self, Ok, Err};
34 use slice::{self, SliceExt};
37 macro_rules! delegate_iter {
38 (exact $te:ty in $ti:ty) => {
39 delegate_iter!{$te in $ti}
40 impl<'a> ExactSizeIterator for $ti {
42 fn rposition<P>(&mut self, predicate: P) -> Option<uint> where P: FnMut($te) -> bool{
43 self.0.rposition(predicate)
46 fn len(&self) -> uint {
51 ($te:ty in $ti:ty) => {
52 impl<'a> Iterator for $ti {
56 fn next(&mut self) -> Option<$te> {
60 fn size_hint(&self) -> (uint, Option<uint>) {
64 impl<'a> DoubleEndedIterator for $ti {
66 fn next_back(&mut self) -> Option<$te> {
71 (pattern $te:ty in $ti:ty) => {
72 impl<'a, P: CharEq> Iterator for $ti {
76 fn next(&mut self) -> Option<$te> {
80 fn size_hint(&self) -> (uint, Option<uint>) {
84 impl<'a, P: CharEq> DoubleEndedIterator for $ti {
86 fn next_back(&mut self) -> Option<$te> {
91 (pattern forward $te:ty in $ti:ty) => {
92 impl<'a, P: CharEq> Iterator for $ti {
96 fn next(&mut self) -> Option<$te> {
100 fn size_hint(&self) -> (uint, Option<uint>) {
107 /// A trait to abstract the idea of creating a new instance of a type from a
109 // FIXME(#17307): there should be an `E` associated type for a `Result` return
110 #[unstable = "will return a Result once associated types are working"]
112 /// Parses a string `s` to return an optional value of this type. If the
113 /// string is ill-formatted, the None is returned.
114 fn from_str(s: &str) -> Option<Self>;
117 /// A utility function that just calls FromStr::from_str
118 #[deprecated = "call the .parse() method on the string instead"]
119 pub fn from_str<A: FromStr>(s: &str) -> Option<A> {
123 impl FromStr for bool {
124 /// Parse a `bool` from a string.
126 /// Yields an `Option<bool>`, because `s` may or may not actually be parseable.
131 /// assert_eq!("true".parse(), Some(true));
132 /// assert_eq!("false".parse(), Some(false));
133 /// assert_eq!("not even a boolean".parse::<bool>(), None);
136 fn from_str(s: &str) -> Option<bool> {
138 "true" => Some(true),
139 "false" => Some(false),
146 Section: Creating a string
149 /// Errors which can occur when attempting to interpret a byte slice as a `str`.
150 #[deriving(Copy, Eq, PartialEq, Clone)]
152 /// An invalid byte was detected at the byte offset given.
154 /// The offset is guaranteed to be in bounds of the slice in question, and
155 /// the byte at the specified offset was the first invalid byte in the
156 /// sequence detected.
159 /// The byte slice was invalid because more bytes were needed but no more
160 /// bytes were available.
164 /// Converts a slice of bytes to a string slice without performing any
167 /// Once the slice has been validated as utf-8, it is transmuted in-place and
168 /// returned as a '&str' instead of a '&[u8]'
172 /// Returns `Err` if the slice is not utf-8 with a description as to why the
173 /// provided slice is not utf-8.
174 pub fn from_utf8(v: &[u8]) -> Result<&str, Utf8Error> {
175 try!(run_utf8_validation_iterator(&mut v.iter()));
176 Ok(unsafe { from_utf8_unchecked(v) })
179 /// Converts a slice of bytes to a string slice without checking
180 /// that the string contains valid UTF-8.
182 pub unsafe fn from_utf8_unchecked<'a>(v: &'a [u8]) -> &'a str {
186 /// Constructs a static string slice from a given raw pointer.
188 /// This function will read memory starting at `s` until it finds a 0, and then
189 /// transmute the memory up to that point as a string slice, returning the
190 /// corresponding `&'static str` value.
192 /// This function is unsafe because the caller must ensure the C string itself
193 /// has the static lifetime and that the memory `s` is valid up to and including
194 /// the first null byte.
198 /// This function will panic if the string pointed to by `s` is not valid UTF-8.
199 #[unstable = "may change location based on the outcome of the c_str module"]
200 pub unsafe fn from_c_str(s: *const i8) -> &'static str {
201 let s = s as *const u8;
203 while *s.offset(len as int) != 0 {
206 let v: &'static [u8] = ::mem::transmute(Slice { data: s, len: len });
207 from_utf8(v).ok().expect("from_c_str passed invalid utf-8 data")
210 /// Something that can be used to compare against a character
211 #[unstable = "definition may change as pattern-related methods are stabilized"]
213 /// Determine if the splitter should split at the given character
214 fn matches(&mut self, char) -> bool;
215 /// Indicate if this is only concerned about ASCII characters,
216 /// which can allow for a faster implementation.
217 fn only_ascii(&self) -> bool;
220 impl CharEq for char {
222 fn matches(&mut self, c: char) -> bool { *self == c }
225 fn only_ascii(&self) -> bool { (*self as uint) < 128 }
228 impl<F> CharEq for F where F: FnMut(char) -> bool {
230 fn matches(&mut self, c: char) -> bool { (*self)(c) }
233 fn only_ascii(&self) -> bool { false }
236 impl<'a> CharEq for &'a [char] {
238 fn matches(&mut self, c: char) -> bool {
239 self.iter().any(|&mut m| m.matches(c))
243 fn only_ascii(&self) -> bool {
244 self.iter().all(|m| m.only_ascii())
252 /// Iterator for the char (representing *Unicode Scalar Values*) of a string
254 /// Created with the method `.chars()`.
255 #[deriving(Clone, Copy)]
256 pub struct Chars<'a> {
257 iter: slice::Iter<'a, u8>
260 // Return the initial codepoint accumulator for the first byte.
261 // The first byte is special, only want bottom 5 bits for width 2, 4 bits
262 // for width 3, and 3 bits for width 4
263 macro_rules! utf8_first_byte {
264 ($byte:expr, $width:expr) => (($byte & (0x7F >> $width)) as u32)
267 // return the value of $ch updated with continuation byte $byte
268 macro_rules! utf8_acc_cont_byte {
269 ($ch:expr, $byte:expr) => (($ch << 6) | ($byte & CONT_MASK) as u32)
272 macro_rules! utf8_is_cont_byte {
273 ($byte:expr) => (($byte & !CONT_MASK) == TAG_CONT_U8)
277 fn unwrap_or_0(opt: Option<&u8>) -> u8 {
284 impl<'a> Iterator for Chars<'a> {
288 fn next(&mut self) -> Option<char> {
289 // Decode UTF-8, using the valid UTF-8 invariant
290 let x = match self.iter.next() {
292 Some(&next_byte) if next_byte < 128 => return Some(next_byte as char),
293 Some(&next_byte) => next_byte,
296 // Multibyte case follows
297 // Decode from a byte combination out of: [[[x y] z] w]
298 // NOTE: Performance is sensitive to the exact formulation here
299 let init = utf8_first_byte!(x, 2);
300 let y = unwrap_or_0(self.iter.next());
301 let mut ch = utf8_acc_cont_byte!(init, y);
304 // 5th bit in 0xE0 .. 0xEF is always clear, so `init` is still valid
305 let z = unwrap_or_0(self.iter.next());
306 let y_z = utf8_acc_cont_byte!((y & CONT_MASK) as u32, z);
307 ch = init << 12 | y_z;
310 // use only the lower 3 bits of `init`
311 let w = unwrap_or_0(self.iter.next());
312 ch = (init & 7) << 18 | utf8_acc_cont_byte!(y_z, w);
316 // str invariant says `ch` is a valid Unicode Scalar Value
318 Some(mem::transmute(ch))
323 fn size_hint(&self) -> (uint, Option<uint>) {
324 let (len, _) = self.iter.size_hint();
325 (len.saturating_add(3) / 4, Some(len))
329 impl<'a> DoubleEndedIterator for Chars<'a> {
331 fn next_back(&mut self) -> Option<char> {
332 let w = match self.iter.next_back() {
334 Some(&back_byte) if back_byte < 128 => return Some(back_byte as char),
335 Some(&back_byte) => back_byte,
338 // Multibyte case follows
339 // Decode from a byte combination out of: [x [y [z w]]]
341 let z = unwrap_or_0(self.iter.next_back());
342 ch = utf8_first_byte!(z, 2);
343 if utf8_is_cont_byte!(z) {
344 let y = unwrap_or_0(self.iter.next_back());
345 ch = utf8_first_byte!(y, 3);
346 if utf8_is_cont_byte!(y) {
347 let x = unwrap_or_0(self.iter.next_back());
348 ch = utf8_first_byte!(x, 4);
349 ch = utf8_acc_cont_byte!(ch, y);
351 ch = utf8_acc_cont_byte!(ch, z);
353 ch = utf8_acc_cont_byte!(ch, w);
355 // str invariant says `ch` is a valid Unicode Scalar Value
357 Some(mem::transmute(ch))
362 /// External iterator for a string's characters and their byte offsets.
363 /// Use with the `std::iter` module.
365 pub struct CharIndices<'a> {
370 impl<'a> Iterator for CharIndices<'a> {
371 type Item = (uint, char);
374 fn next(&mut self) -> Option<(uint, char)> {
375 let (pre_len, _) = self.iter.iter.size_hint();
376 match self.iter.next() {
379 let index = self.front_offset;
380 let (len, _) = self.iter.iter.size_hint();
381 self.front_offset += pre_len - len;
388 fn size_hint(&self) -> (uint, Option<uint>) {
389 self.iter.size_hint()
393 impl<'a> DoubleEndedIterator for CharIndices<'a> {
395 fn next_back(&mut self) -> Option<(uint, char)> {
396 match self.iter.next_back() {
399 let (len, _) = self.iter.iter.size_hint();
400 let index = self.front_offset + len;
407 /// External iterator for a string's bytes.
408 /// Use with the `std::iter` module.
410 /// Created with `StrExt::bytes`
413 pub struct Bytes<'a>(Map<&'a u8, u8, slice::Iter<'a, u8>, BytesDeref>);
414 delegate_iter!{exact u8 in Bytes<'a>}
416 /// A temporary fn new type that ensures that the `Bytes` iterator
418 #[deriving(Copy, Clone)]
421 impl<'a> Fn(&'a u8) -> u8 for BytesDeref {
423 extern "rust-call" fn call(&self, (ptr,): (&'a u8,)) -> u8 {
428 /// An iterator over the substrings of a string, separated by `sep`.
430 #[deprecated = "Type is now named `Split` or `SplitTerminator`"]
431 pub struct CharSplits<'a, Sep> {
432 /// The slice remaining to be iterated
435 /// Whether an empty string at the end is allowed
436 allow_trailing_empty: bool,
441 /// An iterator over the substrings of a string, separated by `sep`,
442 /// splitting at most `count` times.
444 #[deprecated = "Type is now named `SplitN` or `RSplitN`"]
445 pub struct CharSplitsN<'a, Sep> {
446 iter: CharSplits<'a, Sep>,
447 /// The number of splits remaining
452 /// An iterator over the lines of a string, separated by `\n`.
454 pub struct Lines<'a> {
455 inner: CharSplits<'a, char>,
458 /// An iterator over the lines of a string, separated by either `\n` or (`\r\n`).
460 pub struct LinesAny<'a> {
461 inner: Map<&'a str, &'a str, Lines<'a>, fn(&str) -> &str>,
464 impl<'a, Sep> CharSplits<'a, Sep> {
466 fn get_end(&mut self) -> Option<&'a str> {
467 if !self.finished && (self.allow_trailing_empty || self.string.len() > 0) {
468 self.finished = true;
476 impl<'a, Sep: CharEq> Iterator for CharSplits<'a, Sep> {
480 fn next(&mut self) -> Option<&'a str> {
481 if self.finished { return None }
483 let mut next_split = None;
485 for (idx, byte) in self.string.bytes().enumerate() {
486 if self.sep.matches(byte as char) && byte < 128u8 {
487 next_split = Some((idx, idx + 1));
492 for (idx, ch) in self.string.char_indices() {
493 if self.sep.matches(ch) {
494 next_split = Some((idx, self.string.char_range_at(idx).next));
500 Some((a, b)) => unsafe {
501 let elt = self.string.slice_unchecked(0, a);
502 self.string = self.string.slice_unchecked(b, self.string.len());
505 None => self.get_end(),
510 impl<'a, Sep: CharEq> DoubleEndedIterator for CharSplits<'a, Sep> {
512 fn next_back(&mut self) -> Option<&'a str> {
513 if self.finished { return None }
515 if !self.allow_trailing_empty {
516 self.allow_trailing_empty = true;
517 match self.next_back() {
518 Some(elt) if !elt.is_empty() => return Some(elt),
519 _ => if self.finished { return None }
522 let len = self.string.len();
523 let mut next_split = None;
526 for (idx, byte) in self.string.bytes().enumerate().rev() {
527 if self.sep.matches(byte as char) && byte < 128u8 {
528 next_split = Some((idx, idx + 1));
533 for (idx, ch) in self.string.char_indices().rev() {
534 if self.sep.matches(ch) {
535 next_split = Some((idx, self.string.char_range_at(idx).next));
541 Some((a, b)) => unsafe {
542 let elt = self.string.slice_unchecked(b, len);
543 self.string = self.string.slice_unchecked(0, a);
546 None => { self.finished = true; Some(self.string) }
551 impl<'a, Sep: CharEq> Iterator for CharSplitsN<'a, Sep> {
555 fn next(&mut self) -> Option<&'a str> {
558 if self.invert { self.iter.next_back() } else { self.iter.next() }
565 /// The internal state of an iterator that searches for matches of a substring
566 /// within a larger string using naive search
568 struct NaiveSearcher {
573 fn new() -> NaiveSearcher {
574 NaiveSearcher { position: 0 }
577 fn next(&mut self, haystack: &[u8], needle: &[u8]) -> Option<(uint, uint)> {
578 while self.position + needle.len() <= haystack.len() {
579 if haystack[self.position .. self.position + needle.len()] == needle {
580 let match_pos = self.position;
581 self.position += needle.len(); // add 1 for all matches
582 return Some((match_pos, match_pos + needle.len()));
591 /// The internal state of an iterator that searches for matches of a substring
592 /// within a larger string using two-way search
594 struct TwoWaySearcher {
606 This is the Two-Way search algorithm, which was introduced in the paper:
607 Crochemore, M., Perrin, D., 1991, Two-way string-matching, Journal of the ACM 38(3):651-675.
609 Here's some background information.
611 A *word* is a string of symbols. The *length* of a word should be a familiar
612 notion, and here we denote it for any word x by |x|.
613 (We also allow for the possibility of the *empty word*, a word of length zero).
615 If x is any non-empty word, then an integer p with 0 < p <= |x| is said to be a
616 *period* for x iff for all i with 0 <= i <= |x| - p - 1, we have x[i] == x[i+p].
617 For example, both 1 and 2 are periods for the string "aa". As another example,
618 the only period of the string "abcd" is 4.
620 We denote by period(x) the *smallest* period of x (provided that x is non-empty).
621 This is always well-defined since every non-empty word x has at least one period,
622 |x|. We sometimes call this *the period* of x.
624 If u, v and x are words such that x = uv, where uv is the concatenation of u and
625 v, then we say that (u, v) is a *factorization* of x.
627 Let (u, v) be a factorization for a word x. Then if w is a non-empty word such
628 that both of the following hold
630 - either w is a suffix of u or u is a suffix of w
631 - either w is a prefix of v or v is a prefix of w
633 then w is said to be a *repetition* for the factorization (u, v).
635 Just to unpack this, there are four possibilities here. Let w = "abc". Then we
638 - w is a suffix of u and w is a prefix of v. ex: ("lolabc", "abcde")
639 - w is a suffix of u and v is a prefix of w. ex: ("lolabc", "ab")
640 - u is a suffix of w and w is a prefix of v. ex: ("bc", "abchi")
641 - u is a suffix of w and v is a prefix of w. ex: ("bc", "a")
643 Note that the word vu is a repetition for any factorization (u,v) of x = uv,
644 so every factorization has at least one repetition.
646 If x is a string and (u, v) is a factorization for x, then a *local period* for
647 (u, v) is an integer r such that there is some word w such that |w| = r and w is
648 a repetition for (u, v).
650 We denote by local_period(u, v) the smallest local period of (u, v). We sometimes
651 call this *the local period* of (u, v). Provided that x = uv is non-empty, this
652 is well-defined (because each non-empty word has at least one factorization, as
655 It can be proven that the following is an equivalent definition of a local period
656 for a factorization (u, v): any positive integer r such that x[i] == x[i+r] for
657 all i such that |u| - r <= i <= |u| - 1 and such that both x[i] and x[i+r] are
658 defined. (i.e. i > 0 and i + r < |x|).
660 Using the above reformulation, it is easy to prove that
662 1 <= local_period(u, v) <= period(uv)
664 A factorization (u, v) of x such that local_period(u,v) = period(x) is called a
665 *critical factorization*.
667 The algorithm hinges on the following theorem, which is stated without proof:
669 **Critical Factorization Theorem** Any word x has at least one critical
670 factorization (u, v) such that |u| < period(x).
672 The purpose of maximal_suffix is to find such a critical factorization.
675 impl TwoWaySearcher {
676 fn new(needle: &[u8]) -> TwoWaySearcher {
677 let (crit_pos1, period1) = TwoWaySearcher::maximal_suffix(needle, false);
678 let (crit_pos2, period2) = TwoWaySearcher::maximal_suffix(needle, true);
682 if crit_pos1 > crit_pos2 {
683 crit_pos = crit_pos1;
686 crit_pos = crit_pos2;
690 // This isn't in the original algorithm, as far as I'm aware.
691 let byteset = needle.iter()
692 .fold(0, |a, &b| (1 << ((b & 0x3f) as uint)) | a);
694 // A particularly readable explanation of what's going on here can be found
695 // in Crochemore and Rytter's book "Text Algorithms", ch 13. Specifically
696 // see the code for "Algorithm CP" on p. 323.
698 // What's going on is we have some critical factorization (u, v) of the
699 // needle, and we want to determine whether u is a suffix of
700 // v[..period]. If it is, we use "Algorithm CP1". Otherwise we use
701 // "Algorithm CP2", which is optimized for when the period of the needle
703 if needle[..crit_pos] == needle[period.. period + crit_pos] {
715 period: cmp::max(crit_pos, needle.len() - crit_pos) + 1,
719 memory: uint::MAX // Dummy value to signify that the period is long
724 // One of the main ideas of Two-Way is that we factorize the needle into
725 // two halves, (u, v), and begin trying to find v in the haystack by scanning
726 // left to right. If v matches, we try to match u by scanning right to left.
727 // How far we can jump when we encounter a mismatch is all based on the fact
728 // that (u, v) is a critical factorization for the needle.
730 fn next(&mut self, haystack: &[u8], needle: &[u8], long_period: bool) -> Option<(uint, uint)> {
732 // Check that we have room to search in
733 if self.position + needle.len() > haystack.len() {
737 // Quickly skip by large portions unrelated to our substring
739 ((haystack[self.position + needle.len() - 1] & 0x3f)
741 self.position += needle.len();
748 // See if the right part of the needle matches
749 let start = if long_period { self.crit_pos }
750 else { cmp::max(self.crit_pos, self.memory) };
751 for i in range(start, needle.len()) {
752 if needle[i] != haystack[self.position + i] {
753 self.position += i - self.crit_pos + 1;
761 // See if the left part of the needle matches
762 let start = if long_period { 0 } else { self.memory };
763 for i in range(start, self.crit_pos).rev() {
764 if needle[i] != haystack[self.position + i] {
765 self.position += self.period;
767 self.memory = needle.len() - self.period;
773 // We have found a match!
774 let match_pos = self.position;
775 self.position += needle.len(); // add self.period for all matches
777 self.memory = 0; // set to needle.len() - self.period for all matches
779 return Some((match_pos, match_pos + needle.len()));
783 // Computes a critical factorization (u, v) of `arr`.
784 // Specifically, returns (i, p), where i is the starting index of v in some
785 // critical factorization (u, v) and p = period(v)
787 fn maximal_suffix(arr: &[u8], reversed: bool) -> (uint, uint) {
788 let mut left = -1; // Corresponds to i in the paper
789 let mut right = 0; // Corresponds to j in the paper
790 let mut offset = 1; // Corresponds to k in the paper
791 let mut period = 1; // Corresponds to p in the paper
793 while right + offset < arr.len() {
797 a = arr[left + offset];
798 b = arr[right + offset];
800 a = arr[right + offset];
801 b = arr[left + offset];
804 // Suffix is smaller, period is entire prefix so far.
807 period = right - left;
809 // Advance through repetition of the current period.
810 if offset == period {
817 // Suffix is larger, start over from current location.
828 /// The internal state of an iterator that searches for matches of a substring
829 /// within a larger string using a dynamically chosen search algorithm
832 Naive(NaiveSearcher),
833 TwoWay(TwoWaySearcher),
834 TwoWayLong(TwoWaySearcher)
838 fn new(haystack: &[u8], needle: &[u8]) -> Searcher {
840 // FIXME(#16715): This unsigned integer addition will probably not
841 // overflow because that would mean that the memory almost solely
842 // consists of the needle. Needs #16715 to be formally fixed.
843 if needle.len() + 20 > haystack.len() {
844 Naive(NaiveSearcher::new())
846 let searcher = TwoWaySearcher::new(needle);
847 if searcher.memory == uint::MAX { // If the period is long
856 /// An iterator over the start and end indices of the matches of a
857 /// substring within a larger string
859 pub struct MatchIndices<'a> {
866 /// An iterator over the substrings of a string separated by a given
869 #[unstable = "Type might get removed"]
870 pub struct SplitStr<'a> {
871 it: MatchIndices<'a>,
877 #[deprecated = "Type is now named `SplitStr`"]
878 pub type StrSplits<'a> = SplitStr<'a>;
880 impl<'a> Iterator for MatchIndices<'a> {
881 type Item = (uint, uint);
884 fn next(&mut self) -> Option<(uint, uint)> {
885 match self.searcher {
886 Naive(ref mut searcher)
887 => searcher.next(self.haystack.as_bytes(), self.needle.as_bytes()),
888 TwoWay(ref mut searcher)
889 => searcher.next(self.haystack.as_bytes(), self.needle.as_bytes(), false),
890 TwoWayLong(ref mut searcher)
891 => searcher.next(self.haystack.as_bytes(), self.needle.as_bytes(), true)
896 impl<'a> Iterator for SplitStr<'a> {
900 fn next(&mut self) -> Option<&'a str> {
901 if self.finished { return None; }
903 match self.it.next() {
904 Some((from, to)) => {
905 let ret = Some(self.it.haystack.slice(self.last_end, from));
910 self.finished = true;
911 Some(self.it.haystack.slice(self.last_end, self.it.haystack.len()))
919 Section: Comparing strings
922 // share the implementation of the lang-item vs. non-lang-item
924 /// NOTE: This function is (ab)used in rustc::middle::trans::_match
925 /// to compare &[u8] byte slices that are not necessarily valid UTF-8.
927 fn eq_slice_(a: &str, b: &str) -> bool {
928 #[allow(improper_ctypes)]
929 extern { fn memcmp(s1: *const i8, s2: *const i8, n: uint) -> i32; }
930 a.len() == b.len() && unsafe {
931 memcmp(a.as_ptr() as *const i8,
932 b.as_ptr() as *const i8,
937 /// Bytewise slice equality
938 /// NOTE: This function is (ab)used in rustc::middle::trans::_match
939 /// to compare &[u8] byte slices that are not necessarily valid UTF-8.
942 fn eq_slice(a: &str, b: &str) -> bool {
950 /// Walk through `iter` checking that it's a valid UTF-8 sequence,
951 /// returning `true` in that case, or, if it is invalid, `false` with
952 /// `iter` reset such that it is pointing at the first byte in the
953 /// invalid sequence.
955 fn run_utf8_validation_iterator(iter: &mut slice::Iter<u8>)
956 -> Result<(), Utf8Error> {
957 let whole = iter.as_slice();
959 // save the current thing we're pointing at.
962 // restore the iterator we had at the start of this codepoint.
963 macro_rules! err (() => { {
965 return Err(Utf8Error::InvalidByte(whole.len() - iter.as_slice().len()))
967 macro_rules! next ( () => {
970 // we needed data, but there was none: error!
971 None => return Err(Utf8Error::TooShort),
975 let first = match iter.next() {
977 // we're at the end of the iterator and a codepoint
978 // boundary at the same time, so this string is valid.
979 None => return Ok(())
982 // ASCII characters are always valid, so only large
983 // bytes need more examination.
985 let w = UTF8_CHAR_WIDTH[first as uint] as uint;
986 let second = next!();
987 // 2-byte encoding is for codepoints \u{0080} to \u{07ff}
988 // first C2 80 last DF BF
989 // 3-byte encoding is for codepoints \u{0800} to \u{ffff}
990 // first E0 A0 80 last EF BF BF
991 // excluding surrogates codepoints \u{d800} to \u{dfff}
992 // ED A0 80 to ED BF BF
993 // 4-byte encoding is for codepoints \u{1000}0 to \u{10ff}ff
994 // first F0 90 80 80 last F4 8F BF BF
996 // Use the UTF-8 syntax from the RFC
998 // https://tools.ietf.org/html/rfc3629
1000 // UTF8-2 = %xC2-DF UTF8-tail
1001 // UTF8-3 = %xE0 %xA0-BF UTF8-tail / %xE1-EC 2( UTF8-tail ) /
1002 // %xED %x80-9F UTF8-tail / %xEE-EF 2( UTF8-tail )
1003 // UTF8-4 = %xF0 %x90-BF 2( UTF8-tail ) / %xF1-F3 3( UTF8-tail ) /
1004 // %xF4 %x80-8F 2( UTF8-tail )
1006 2 => if second & !CONT_MASK != TAG_CONT_U8 {err!()},
1008 match (first, second, next!() & !CONT_MASK) {
1009 (0xE0 , 0xA0 ... 0xBF, TAG_CONT_U8) |
1010 (0xE1 ... 0xEC, 0x80 ... 0xBF, TAG_CONT_U8) |
1011 (0xED , 0x80 ... 0x9F, TAG_CONT_U8) |
1012 (0xEE ... 0xEF, 0x80 ... 0xBF, TAG_CONT_U8) => {}
1017 match (first, second, next!() & !CONT_MASK, next!() & !CONT_MASK) {
1018 (0xF0 , 0x90 ... 0xBF, TAG_CONT_U8, TAG_CONT_U8) |
1019 (0xF1 ... 0xF3, 0x80 ... 0xBF, TAG_CONT_U8, TAG_CONT_U8) |
1020 (0xF4 , 0x80 ... 0x8F, TAG_CONT_U8, TAG_CONT_U8) => {}
1030 /// Determines if a vector of bytes contains valid UTF-8.
1031 #[deprecated = "call from_utf8 instead"]
1032 pub fn is_utf8(v: &[u8]) -> bool {
1033 run_utf8_validation_iterator(&mut v.iter()).is_ok()
1036 /// Deprecated function
1037 #[deprecated = "this function will be removed"]
1038 pub fn truncate_utf16_at_nul<'a>(v: &'a [u16]) -> &'a [u16] {
1039 match v.iter().position(|c| *c == 0) {
1040 // don't include the 0
1046 // https://tools.ietf.org/html/rfc3629
1047 static UTF8_CHAR_WIDTH: [u8; 256] = [
1048 1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,
1049 1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1, // 0x1F
1050 1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,
1051 1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1, // 0x3F
1052 1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,
1053 1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1, // 0x5F
1054 1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,
1055 1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1, // 0x7F
1056 0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,
1057 0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0, // 0x9F
1058 0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,
1059 0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0, // 0xBF
1060 0,0,2,2,2,2,2,2,2,2,2,2,2,2,2,2,
1061 2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2, // 0xDF
1062 3,3,3,3,3,3,3,3,3,3,3,3,3,3,3,3, // 0xEF
1063 4,4,4,4,4,0,0,0,0,0,0,0,0,0,0,0, // 0xFF
1066 /// Given a first byte, determine how many bytes are in this UTF-8 character
1068 #[deprecated = "this function has moved to libunicode"]
1069 pub fn utf8_char_width(b: u8) -> uint {
1070 return UTF8_CHAR_WIDTH[b as uint] as uint;
1073 /// Struct that contains a `char` and the index of the first byte of
1074 /// the next `char` in a string. This can be used as a data structure
1075 /// for iterating over the UTF-8 bytes of a string.
1077 #[unstable = "naming is uncertain with container conventions"]
1078 pub struct CharRange {
1081 /// Index of the first byte of the next `char`
1085 /// Mask of the value bits of a continuation byte
1086 const CONT_MASK: u8 = 0b0011_1111u8;
1087 /// Value of the tag bits (tag mask is !CONT_MASK) of a continuation byte
1088 const TAG_CONT_U8: u8 = 0b1000_0000u8;
1090 /// Unsafe operations
1095 use slice::SliceExt;
1098 /// Converts a slice of bytes to a string slice without checking
1099 /// that the string contains valid UTF-8.
1100 #[deprecated = "renamed to str::from_utf8_unchecked"]
1101 pub unsafe fn from_utf8<'a>(v: &'a [u8]) -> &'a str {
1102 super::from_utf8_unchecked(v)
1105 /// Form a slice from a C string. Unsafe because the caller must ensure the
1106 /// C string has the static lifetime, or else the return value may be
1107 /// invalidated later.
1108 #[deprecated = "renamed to str::from_c_str"]
1109 pub unsafe fn c_str_to_static_slice(s: *const i8) -> &'static str {
1110 let s = s as *const u8;
1113 while *curr != 0u8 {
1115 curr = s.offset(len as int);
1117 let v = Slice { data: s, len: len };
1118 super::from_utf8(::mem::transmute(v)).unwrap()
1121 /// Takes a bytewise (not UTF-8) slice from a string.
1123 /// Returns the substring from [`begin`..`end`).
1127 /// If begin is greater than end.
1128 /// If end is greater than the length of the string.
1130 #[deprecated = "call the slice_unchecked method instead"]
1131 pub unsafe fn slice_bytes<'a>(s: &'a str, begin: uint, end: uint) -> &'a str {
1132 assert!(begin <= end);
1133 assert!(end <= s.len());
1134 s.slice_unchecked(begin, end)
1137 /// Takes a bytewise (not UTF-8) slice from a string.
1139 /// Returns the substring from [`begin`..`end`).
1141 /// Caller must check slice boundaries!
1143 #[deprecated = "this has moved to a method on `str` directly"]
1144 pub unsafe fn slice_unchecked<'a>(s: &'a str, begin: uint, end: uint) -> &'a str {
1145 s.slice_unchecked(begin, end)
1150 Section: Trait implementations
1153 #[allow(missing_docs)]
1155 use cmp::{Ordering, Ord, PartialEq, PartialOrd, Equiv, Eq};
1156 use cmp::Ordering::{Less, Equal, Greater};
1157 use iter::IteratorExt;
1159 use option::Option::Some;
1161 use str::{Str, StrExt, eq_slice};
1166 fn cmp(&self, other: &str) -> Ordering {
1167 for (s_b, o_b) in self.bytes().zip(other.bytes()) {
1168 match s_b.cmp(&o_b) {
1169 Greater => return Greater,
1170 Less => return Less,
1175 self.len().cmp(&other.len())
1180 impl PartialEq for str {
1182 fn eq(&self, other: &str) -> bool {
1183 eq_slice(self, other)
1186 fn ne(&self, other: &str) -> bool { !(*self).eq(other) }
1193 impl PartialOrd for str {
1195 fn partial_cmp(&self, other: &str) -> Option<Ordering> {
1196 Some(self.cmp(other))
1200 #[allow(deprecated)]
1201 #[deprecated = "Use overloaded `core::cmp::PartialEq`"]
1202 impl<S: Str> Equiv<S> for str {
1204 fn equiv(&self, other: &S) -> bool { eq_slice(self, other.as_slice()) }
1207 impl ops::Slice<uint, str> for str {
1209 fn as_slice_<'a>(&'a self) -> &'a str {
1214 fn slice_from_or_fail<'a>(&'a self, from: &uint) -> &'a str {
1215 self.slice_from(*from)
1219 fn slice_to_or_fail<'a>(&'a self, to: &uint) -> &'a str {
1224 fn slice_or_fail<'a>(&'a self, from: &uint, to: &uint) -> &'a str {
1225 self.slice(*from, *to)
1230 /// Any string that can be represented as a slice
1231 #[unstable = "Instead of taking this bound generically, this trait will be \
1232 replaced with one of slicing syntax, deref coercions, or \
1233 a more generic conversion trait"]
1234 pub trait Str for Sized? {
1235 /// Work with `self` as a slice.
1236 fn as_slice<'a>(&'a self) -> &'a str;
1239 #[allow(deprecated)]
1242 fn as_slice<'a>(&'a self) -> &'a str { self }
1245 #[allow(deprecated)]
1246 impl<'a, Sized? S> Str for &'a S where S: Str {
1248 fn as_slice(&self) -> &str { Str::as_slice(*self) }
1251 /// Return type of `StrExt::split`
1254 pub struct Split<'a, P>(CharSplits<'a, P>);
1255 delegate_iter!{pattern &'a str in Split<'a, P>}
1257 /// Return type of `StrExt::split_terminator`
1259 #[unstable = "might get removed in favour of a constructor method on Split"]
1260 pub struct SplitTerminator<'a, P>(CharSplits<'a, P>);
1261 delegate_iter!{pattern &'a str in SplitTerminator<'a, P>}
1263 /// Return type of `StrExt::splitn`
1266 pub struct SplitN<'a, P>(CharSplitsN<'a, P>);
1267 delegate_iter!{pattern forward &'a str in SplitN<'a, P>}
1269 /// Return type of `StrExt::rsplitn`
1272 pub struct RSplitN<'a, P>(CharSplitsN<'a, P>);
1273 delegate_iter!{pattern forward &'a str in RSplitN<'a, P>}
1275 /// Methods for string slices
1276 #[allow(missing_docs)]
1277 pub trait StrExt for Sized? {
1278 // NB there are no docs here are they're all located on the StrExt trait in
1279 // libcollections, not here.
1281 fn contains(&self, pat: &str) -> bool;
1282 fn contains_char<P: CharEq>(&self, pat: P) -> bool;
1283 fn chars<'a>(&'a self) -> Chars<'a>;
1284 fn bytes<'a>(&'a self) -> Bytes<'a>;
1285 fn char_indices<'a>(&'a self) -> CharIndices<'a>;
1286 fn split<'a, P: CharEq>(&'a self, pat: P) -> Split<'a, P>;
1287 fn splitn<'a, P: CharEq>(&'a self, count: uint, pat: P) -> SplitN<'a, P>;
1288 fn split_terminator<'a, P: CharEq>(&'a self, pat: P) -> SplitTerminator<'a, P>;
1289 fn rsplitn<'a, P: CharEq>(&'a self, count: uint, pat: P) -> RSplitN<'a, P>;
1290 fn match_indices<'a>(&'a self, sep: &'a str) -> MatchIndices<'a>;
1291 fn split_str<'a>(&'a self, pat: &'a str) -> SplitStr<'a>;
1292 fn lines<'a>(&'a self) -> Lines<'a>;
1293 fn lines_any<'a>(&'a self) -> LinesAny<'a>;
1294 fn char_len(&self) -> uint;
1295 fn slice<'a>(&'a self, begin: uint, end: uint) -> &'a str;
1296 fn slice_from<'a>(&'a self, begin: uint) -> &'a str;
1297 fn slice_to<'a>(&'a self, end: uint) -> &'a str;
1298 fn slice_chars<'a>(&'a self, begin: uint, end: uint) -> &'a str;
1299 unsafe fn slice_unchecked<'a>(&'a self, begin: uint, end: uint) -> &'a str;
1300 fn starts_with(&self, pat: &str) -> bool;
1301 fn ends_with(&self, pat: &str) -> bool;
1302 fn trim_matches<'a, P: CharEq>(&'a self, pat: P) -> &'a str;
1303 fn trim_left_matches<'a, P: CharEq>(&'a self, pat: P) -> &'a str;
1304 fn trim_right_matches<'a, P: CharEq>(&'a self, pat: P) -> &'a str;
1305 fn is_char_boundary(&self, index: uint) -> bool;
1306 fn char_range_at(&self, start: uint) -> CharRange;
1307 fn char_range_at_reverse(&self, start: uint) -> CharRange;
1308 fn char_at(&self, i: uint) -> char;
1309 fn char_at_reverse(&self, i: uint) -> char;
1310 fn as_bytes<'a>(&'a self) -> &'a [u8];
1311 fn find<P: CharEq>(&self, pat: P) -> Option<uint>;
1312 fn rfind<P: CharEq>(&self, pat: P) -> Option<uint>;
1313 fn find_str(&self, pat: &str) -> Option<uint>;
1314 fn slice_shift_char<'a>(&'a self) -> Option<(char, &'a str)>;
1315 fn subslice_offset(&self, inner: &str) -> uint;
1316 fn as_ptr(&self) -> *const u8;
1317 fn len(&self) -> uint;
1318 fn is_empty(&self) -> bool;
1322 fn slice_error_fail(s: &str, begin: uint, end: uint) -> ! {
1323 assert!(begin <= end);
1324 panic!("index {} and/or {} in `{}` do not lie on character boundary",
1328 impl StrExt for str {
1330 fn contains(&self, needle: &str) -> bool {
1331 self.find_str(needle).is_some()
1335 fn contains_char<P: CharEq>(&self, pat: P) -> bool {
1336 self.find(pat).is_some()
1340 fn chars(&self) -> Chars {
1341 Chars{iter: self.as_bytes().iter()}
1345 fn bytes(&self) -> Bytes {
1346 Bytes(self.as_bytes().iter().map(BytesDeref))
1350 fn char_indices(&self) -> CharIndices {
1351 CharIndices { front_offset: 0, iter: self.chars() }
1355 #[allow(deprecated)] // For using CharSplits
1356 fn split<P: CharEq>(&self, pat: P) -> Split<P> {
1359 only_ascii: pat.only_ascii(),
1361 allow_trailing_empty: true,
1367 #[allow(deprecated)] // For using CharSplitsN
1368 fn splitn<P: CharEq>(&self, count: uint, pat: P) -> SplitN<P> {
1369 SplitN(CharSplitsN {
1370 iter: self.split(pat).0,
1377 #[allow(deprecated)] // For using CharSplits
1378 fn split_terminator<P: CharEq>(&self, pat: P) -> SplitTerminator<P> {
1379 SplitTerminator(CharSplits {
1380 allow_trailing_empty: false,
1386 #[allow(deprecated)] // For using CharSplitsN
1387 fn rsplitn<P: CharEq>(&self, count: uint, pat: P) -> RSplitN<P> {
1388 RSplitN(CharSplitsN {
1389 iter: self.split(pat).0,
1396 fn match_indices<'a>(&'a self, sep: &'a str) -> MatchIndices<'a> {
1397 assert!(!sep.is_empty());
1401 searcher: Searcher::new(self.as_bytes(), sep.as_bytes())
1406 fn split_str<'a>(&'a self, sep: &'a str) -> SplitStr<'a> {
1408 it: self.match_indices(sep),
1415 fn lines(&self) -> Lines {
1416 Lines { inner: self.split_terminator('\n').0 }
1419 fn lines_any(&self) -> LinesAny {
1420 fn f(line: &str) -> &str {
1422 if l > 0 && line.as_bytes()[l - 1] == b'\r' { line.slice(0, l - 1) }
1426 let f: fn(&str) -> &str = f; // coerce to fn pointer
1427 LinesAny { inner: self.lines().map(f) }
1431 fn char_len(&self) -> uint { self.chars().count() }
1434 fn slice(&self, begin: uint, end: uint) -> &str {
1435 // is_char_boundary checks that the index is in [0, .len()]
1437 self.is_char_boundary(begin) &&
1438 self.is_char_boundary(end) {
1439 unsafe { self.slice_unchecked(begin, end) }
1441 slice_error_fail(self, begin, end)
1446 fn slice_from(&self, begin: uint) -> &str {
1447 // is_char_boundary checks that the index is in [0, .len()]
1448 if self.is_char_boundary(begin) {
1449 unsafe { self.slice_unchecked(begin, self.len()) }
1451 slice_error_fail(self, begin, self.len())
1456 fn slice_to(&self, end: uint) -> &str {
1457 // is_char_boundary checks that the index is in [0, .len()]
1458 if self.is_char_boundary(end) {
1459 unsafe { self.slice_unchecked(0, end) }
1461 slice_error_fail(self, 0, end)
1465 fn slice_chars(&self, begin: uint, end: uint) -> &str {
1466 assert!(begin <= end);
1468 let mut begin_byte = None;
1469 let mut end_byte = None;
1471 // This could be even more efficient by not decoding,
1472 // only finding the char boundaries
1473 for (idx, _) in self.char_indices() {
1474 if count == begin { begin_byte = Some(idx); }
1475 if count == end { end_byte = Some(idx); break; }
1478 if begin_byte.is_none() && count == begin { begin_byte = Some(self.len()) }
1479 if end_byte.is_none() && count == end { end_byte = Some(self.len()) }
1481 match (begin_byte, end_byte) {
1482 (None, _) => panic!("slice_chars: `begin` is beyond end of string"),
1483 (_, None) => panic!("slice_chars: `end` is beyond end of string"),
1484 (Some(a), Some(b)) => unsafe { self.slice_unchecked(a, b) }
1489 unsafe fn slice_unchecked(&self, begin: uint, end: uint) -> &str {
1490 mem::transmute(Slice {
1491 data: self.as_ptr().offset(begin as int),
1497 fn starts_with(&self, needle: &str) -> bool {
1498 let n = needle.len();
1499 self.len() >= n && needle.as_bytes() == self.as_bytes()[..n]
1503 fn ends_with(&self, needle: &str) -> bool {
1504 let (m, n) = (self.len(), needle.len());
1505 m >= n && needle.as_bytes() == self.as_bytes()[m-n..]
1509 fn trim_matches<P: CharEq>(&self, mut pat: P) -> &str {
1510 let cur = match self.find(|&mut: c: char| !pat.matches(c)) {
1512 Some(i) => unsafe { self.slice_unchecked(i, self.len()) }
1514 match cur.rfind(|&mut: c: char| !pat.matches(c)) {
1517 let right = cur.char_range_at(i).next;
1518 unsafe { cur.slice_unchecked(0, right) }
1524 fn trim_left_matches<P: CharEq>(&self, mut pat: P) -> &str {
1525 match self.find(|&mut: c: char| !pat.matches(c)) {
1527 Some(first) => unsafe { self.slice_unchecked(first, self.len()) }
1532 fn trim_right_matches<P: CharEq>(&self, mut pat: P) -> &str {
1533 match self.rfind(|&mut: c: char| !pat.matches(c)) {
1536 let next = self.char_range_at(last).next;
1537 unsafe { self.slice_unchecked(0u, next) }
1543 fn is_char_boundary(&self, index: uint) -> bool {
1544 if index == self.len() { return true; }
1545 match self.as_bytes().get(index) {
1547 Some(&b) => b < 128u8 || b >= 192u8,
1552 fn char_range_at(&self, i: uint) -> CharRange {
1553 if self.as_bytes()[i] < 128u8 {
1554 return CharRange {ch: self.as_bytes()[i] as char, next: i + 1 };
1557 // Multibyte case is a fn to allow char_range_at to inline cleanly
1558 fn multibyte_char_range_at(s: &str, i: uint) -> CharRange {
1559 let mut val = s.as_bytes()[i] as u32;
1560 let w = UTF8_CHAR_WIDTH[val as uint] as uint;
1563 val = utf8_first_byte!(val, w);
1564 val = utf8_acc_cont_byte!(val, s.as_bytes()[i + 1]);
1565 if w > 2 { val = utf8_acc_cont_byte!(val, s.as_bytes()[i + 2]); }
1566 if w > 3 { val = utf8_acc_cont_byte!(val, s.as_bytes()[i + 3]); }
1568 return CharRange {ch: unsafe { mem::transmute(val) }, next: i + w};
1571 return multibyte_char_range_at(self, i);
1575 fn char_range_at_reverse(&self, start: uint) -> CharRange {
1576 let mut prev = start;
1578 prev = prev.saturating_sub(1);
1579 if self.as_bytes()[prev] < 128 {
1580 return CharRange{ch: self.as_bytes()[prev] as char, next: prev}
1583 // Multibyte case is a fn to allow char_range_at_reverse to inline cleanly
1584 fn multibyte_char_range_at_reverse(s: &str, mut i: uint) -> CharRange {
1585 // while there is a previous byte == 10......
1586 while i > 0 && s.as_bytes()[i] & !CONT_MASK == TAG_CONT_U8 {
1590 let mut val = s.as_bytes()[i] as u32;
1591 let w = UTF8_CHAR_WIDTH[val as uint] as uint;
1594 val = utf8_first_byte!(val, w);
1595 val = utf8_acc_cont_byte!(val, s.as_bytes()[i + 1]);
1596 if w > 2 { val = utf8_acc_cont_byte!(val, s.as_bytes()[i + 2]); }
1597 if w > 3 { val = utf8_acc_cont_byte!(val, s.as_bytes()[i + 3]); }
1599 return CharRange {ch: unsafe { mem::transmute(val) }, next: i};
1602 return multibyte_char_range_at_reverse(self, prev);
1606 fn char_at(&self, i: uint) -> char {
1607 self.char_range_at(i).ch
1611 fn char_at_reverse(&self, i: uint) -> char {
1612 self.char_range_at_reverse(i).ch
1616 fn as_bytes(&self) -> &[u8] {
1617 unsafe { mem::transmute(self) }
1620 fn find<P: CharEq>(&self, mut pat: P) -> Option<uint> {
1621 if pat.only_ascii() {
1622 self.bytes().position(|b| pat.matches(b as char))
1624 for (index, c) in self.char_indices() {
1625 if pat.matches(c) { return Some(index); }
1631 fn rfind<P: CharEq>(&self, mut pat: P) -> Option<uint> {
1632 if pat.only_ascii() {
1633 self.bytes().rposition(|b| pat.matches(b as char))
1635 for (index, c) in self.char_indices().rev() {
1636 if pat.matches(c) { return Some(index); }
1642 fn find_str(&self, needle: &str) -> Option<uint> {
1643 if needle.is_empty() {
1646 self.match_indices(needle)
1648 .map(|(start, _end)| start)
1653 fn slice_shift_char(&self) -> Option<(char, &str)> {
1654 if self.is_empty() {
1657 let CharRange {ch, next} = self.char_range_at(0u);
1658 let next_s = unsafe { self.slice_unchecked(next, self.len()) };
1663 fn subslice_offset(&self, inner: &str) -> uint {
1664 let a_start = self.as_ptr() as uint;
1665 let a_end = a_start + self.len();
1666 let b_start = inner.as_ptr() as uint;
1667 let b_end = b_start + inner.len();
1669 assert!(a_start <= b_start);
1670 assert!(b_end <= a_end);
1675 fn as_ptr(&self) -> *const u8 {
1680 fn len(&self) -> uint { self.repr().len }
1683 fn is_empty(&self) -> bool { self.len() == 0 }
1687 impl<'a> Default for &'a str {
1689 fn default() -> &'a str { "" }
1692 impl<'a> Iterator for Lines<'a> {
1693 type Item = &'a str;
1696 fn next(&mut self) -> Option<&'a str> { self.inner.next() }
1698 fn size_hint(&self) -> (uint, Option<uint>) { self.inner.size_hint() }
1700 impl<'a> DoubleEndedIterator for Lines<'a> {
1702 fn next_back(&mut self) -> Option<&'a str> { self.inner.next_back() }
1704 impl<'a> Iterator for LinesAny<'a> {
1705 type Item = &'a str;
1708 fn next(&mut self) -> Option<&'a str> { self.inner.next() }
1710 fn size_hint(&self) -> (uint, Option<uint>) { self.inner.size_hint() }
1712 impl<'a> DoubleEndedIterator for LinesAny<'a> {
1714 fn next_back(&mut self) -> Option<&'a str> { self.inner.next_back() }