src/libcollections/string.rs

   1 // Copyright 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.
   4 //
   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.
  10
  11 //! An owned, growable string that enforces that its contents are valid UTF-8.
  12
  13 #![stable(feature = "rust1", since = "1.0.0")]
  14
  15 use core::fmt;
  16 use core::hash;
  17 use core::iter::FromIterator;
  18 use core::mem;
  19 use core::ops::{self, Deref, Add, Index};
  20 use core::ptr;
  21 use core::slice;
  22 use core::str::pattern::Pattern;
  23 use rustc_unicode::char::{decode_utf16, REPLACEMENT_CHARACTER};
  24 use rustc_unicode::str as unicode_str;
  25
  26 use borrow::{Cow, IntoCow};
  27 use range::RangeArgument;
  28 use str::{self, FromStr, Utf8Error, Chars};
  29 use vec::Vec;
  30 use boxed::Box;
  31
  32 /// A growable string stored as a UTF-8 encoded buffer.
  33 #[derive(Clone, PartialOrd, Eq, Ord)]
  34 #[stable(feature = "rust1", since = "1.0.0")]
  35 pub struct String {
  36     vec: Vec<u8>,
  37 }
  38
  39 /// A possible error value from the `String::from_utf8` function.
  40 #[stable(feature = "rust1", since = "1.0.0")]
  41 #[derive(Debug)]
  42 pub struct FromUtf8Error {
  43     bytes: Vec<u8>,
  44     error: Utf8Error,
  45 }
  46
  47 /// A possible error value from the `String::from_utf16` function.
  48 #[stable(feature = "rust1", since = "1.0.0")]
  49 #[derive(Debug)]
  50 pub struct FromUtf16Error(());
  51
  52 impl String {
  53     /// Creates a new string buffer initialized with the empty string.
  54     ///
  55     /// # Examples
  56     ///
  57     /// ```
  58     /// let mut s = String::new();
  59     /// ```
  60     #[inline]
  61     #[stable(feature = "rust1", since = "1.0.0")]
  62     pub fn new() -> String {
  63         String {
  64             vec: Vec::new(),
  65         }
  66     }
  67
  68     /// Creates a new string buffer with the given capacity.
  69     /// The string will be able to hold exactly `capacity` bytes without
  70     /// reallocating. If `capacity` is 0, the string will not allocate.
  71     ///
  72     /// # Examples
  73     ///
  74     /// ```
  75     /// let mut s = String::with_capacity(10);
  76     /// ```
  77     #[inline]
  78     #[stable(feature = "rust1", since = "1.0.0")]
  79     pub fn with_capacity(capacity: usize) -> String {
  80         String {
  81             vec: Vec::with_capacity(capacity),
  82         }
  83     }
  84
  85     // HACK(japaric): with cfg(test) the inherent `[T]::to_vec` method, which is
  86     // required for this method definition, is not available. Since we don't
  87     // require this method for testing purposes, I'll just stub it
  88     // NB see the slice::hack module in slice.rs for more information
  89     #[inline]
  90     #[cfg(test)]
  91     pub fn from_str(_: &str) -> String {
  92         panic!("not available with cfg(test)");
  93     }
  94
  95     /// Returns the vector as a string buffer, if possible, taking care not to
  96     /// copy it.
  97     ///
  98     /// # Failure
  99     ///
 100     /// If the given vector is not valid UTF-8, then the original vector and the
 101     /// corresponding error is returned.
 102     ///
 103     /// # Examples
 104     ///
 105     /// ```
 106     /// let hello_vec = vec![104, 101, 108, 108, 111];
 107     /// let s = String::from_utf8(hello_vec).unwrap();
 108     /// assert_eq!(s, "hello");
 109     ///
 110     /// let invalid_vec = vec![240, 144, 128];
 111     /// let s = String::from_utf8(invalid_vec).err().unwrap();
 112     /// let err = s.utf8_error();
 113     /// assert_eq!(s.into_bytes(), [240, 144, 128]);
 114     /// ```
 115     #[inline]
 116     #[stable(feature = "rust1", since = "1.0.0")]
 117     pub fn from_utf8(vec: Vec<u8>) -> Result<String, FromUtf8Error> {
 118         match str::from_utf8(&vec) {
 119             Ok(..) => Ok(String { vec: vec }),
 120             Err(e) => Err(FromUtf8Error { bytes: vec, error: e })
 121         }
 122     }
 123
 124     /// Converts a vector of bytes to a new UTF-8 string.
 125     /// Any invalid UTF-8 sequences are replaced with U+FFFD REPLACEMENT CHARACTER.
 126     ///
 127     /// # Examples
 128     ///
 129     /// ```
 130     /// let input = b"Hello \xF0\x90\x80World";
 131     /// let output = String::from_utf8_lossy(input);
 132     /// assert_eq!(output, "Hello \u{FFFD}World");
 133     /// ```
 134     #[stable(feature = "rust1", since = "1.0.0")]
 135     pub fn from_utf8_lossy<'a>(v: &'a [u8]) -> Cow<'a, str> {
 136         let mut i;
 137         match str::from_utf8(v) {
 138             Ok(s) => return Cow::Borrowed(s),
 139             Err(e) => i = e.valid_up_to(),
 140         }
 141
 142         const TAG_CONT_U8: u8 = 128;
 143         const REPLACEMENT: &'static [u8] = b"\xEF\xBF\xBD"; // U+FFFD in UTF-8
 144         let total = v.len();
 145         fn unsafe_get(xs: &[u8], i: usize) -> u8 {
 146             unsafe { *xs.get_unchecked(i) }
 147         }
 148         fn safe_get(xs: &[u8], i: usize, total: usize) -> u8 {
 149             if i >= total {
 150                 0
 151             } else {
 152                 unsafe_get(xs, i)
 153             }
 154         }
 155
 156         let mut res = String::with_capacity(total);
 157
 158         if i > 0 {
 159             unsafe {
 160                 res.as_mut_vec().push_all(&v[..i])
 161             };
 162         }
 163
 164         // subseqidx is the index of the first byte of the subsequence we're
 165         // looking at.  It's used to copy a bunch of contiguous good codepoints
 166         // at once instead of copying them one by one.
 167         let mut subseqidx = i;
 168
 169         while i < total {
 170             let i_ = i;
 171             let byte = unsafe_get(v, i);
 172             i += 1;
 173
 174             macro_rules! error { () => ({
 175                 unsafe {
 176                     if subseqidx != i_ {
 177                         res.as_mut_vec().push_all(&v[subseqidx..i_]);
 178                     }
 179                     subseqidx = i;
 180                     res.as_mut_vec().push_all(REPLACEMENT);
 181                 }
 182             })}
 183
 184             if byte < 128 {
 185                 // subseqidx handles this
 186             } else {
 187                 let w = unicode_str::utf8_char_width(byte);
 188
 189                 match w {
 190                     2 => {
 191                         if safe_get(v, i, total) & 192 != TAG_CONT_U8 {
 192                             error!();
 193                             continue;
 194                         }
 195                         i += 1;
 196                     }
 197                     3 => {
 198                         match (byte, safe_get(v, i, total)) {
 199                             (0xE0         , 0xA0 ... 0xBF) => (),
 200                             (0xE1 ... 0xEC, 0x80 ... 0xBF) => (),
 201                             (0xED         , 0x80 ... 0x9F) => (),
 202                             (0xEE ... 0xEF, 0x80 ... 0xBF) => (),
 203                             _ => {
 204                                 error!();
 205                                 continue;
 206                             }
 207                         }
 208                         i += 1;
 209                         if safe_get(v, i, total) & 192 != TAG_CONT_U8 {
 210                             error!();
 211                             continue;
 212                         }
 213                         i += 1;
 214                     }
 215                     4 => {
 216                         match (byte, safe_get(v, i, total)) {
 217                             (0xF0         , 0x90 ... 0xBF) => (),
 218                             (0xF1 ... 0xF3, 0x80 ... 0xBF) => (),
 219                             (0xF4         , 0x80 ... 0x8F) => (),
 220                             _ => {
 221                                 error!();
 222                                 continue;
 223                             }
 224                         }
 225                         i += 1;
 226                         if safe_get(v, i, total) & 192 != TAG_CONT_U8 {
 227                             error!();
 228                             continue;
 229                         }
 230                         i += 1;
 231                         if safe_get(v, i, total) & 192 != TAG_CONT_U8 {
 232                             error!();
 233                             continue;
 234                         }
 235                         i += 1;
 236                     }
 237                     _ => {
 238                         error!();
 239                         continue;
 240                     }
 241                 }
 242             }
 243         }
 244         if subseqidx < total {
 245             unsafe {
 246                 res.as_mut_vec().push_all(&v[subseqidx..total])
 247             };
 248         }
 249         Cow::Owned(res)
 250     }
 251
 252     /// Decode a UTF-16 encoded vector `v` into a `String`, returning `None`
 253     /// if `v` contains any invalid data.
 254     ///
 255     /// # Examples
 256     ///
 257     /// ```
 258     /// // 𝄞music
 259     /// let mut v = &mut [0xD834, 0xDD1E, 0x006d, 0x0075,
 260     ///                   0x0073, 0x0069, 0x0063];
 261     /// assert_eq!(String::from_utf16(v).unwrap(),
 262     ///            "𝄞music".to_string());
 263     ///
 264     /// // 𝄞mu<invalid>ic
 265     /// v[4] = 0xD800;
 266     /// assert!(String::from_utf16(v).is_err());
 267     /// ```
 268     #[stable(feature = "rust1", since = "1.0.0")]
 269     pub fn from_utf16(v: &[u16]) -> Result<String, FromUtf16Error> {
 270         decode_utf16(v.iter().cloned()).collect::<Result<_, _>>().map_err(|_| FromUtf16Error(()))
 271     }
 272
 273     /// Decode a UTF-16 encoded vector `v` into a string, replacing
 274     /// invalid data with the replacement character (U+FFFD).
 275     ///
 276     /// # Examples
 277     ///
 278     /// ```
 279     /// // 𝄞mus<invalid>ic<invalid>
 280     /// let v = &[0xD834, 0xDD1E, 0x006d, 0x0075,
 281     ///           0x0073, 0xDD1E, 0x0069, 0x0063,
 282     ///           0xD834];
 283     ///
 284     /// assert_eq!(String::from_utf16_lossy(v),
 285     ///            "𝄞mus\u{FFFD}ic\u{FFFD}".to_string());
 286     /// ```
 287     #[inline]
 288     #[stable(feature = "rust1", since = "1.0.0")]
 289     pub fn from_utf16_lossy(v: &[u16]) -> String {
 290         decode_utf16(v.iter().cloned()).map(|r| r.unwrap_or(REPLACEMENT_CHARACTER)).collect()
 291     }
 292
 293     /// Creates a new `String` from a length, capacity, and pointer.
 294     ///
 295     /// # Unsafety
 296     ///
 297     /// This is _very_ unsafe because:
 298     ///
 299     /// * We call `Vec::from_raw_parts` to get a `Vec<u8>`. Therefore, this
 300     ///   function inherits all of its unsafety, see [its
 301     ///   documentation](../vec/struct.Vec.html#method.from_raw_parts)
 302     ///   for the invariants it expects, they also apply to this function.
 303     /// * We assume that the `Vec` contains valid UTF-8.
 304     #[inline]
 305     #[stable(feature = "rust1", since = "1.0.0")]
 306     pub unsafe fn from_raw_parts(buf: *mut u8, length: usize, capacity: usize) -> String {
 307         String {
 308             vec: Vec::from_raw_parts(buf, length, capacity),
 309         }
 310     }
 311
 312     /// Converts a vector of bytes to a new `String` without checking if
 313     /// it contains valid UTF-8. This is unsafe because it assumes that
 314     /// the UTF-8-ness of the vector has already been validated.
 315     #[inline]
 316     #[stable(feature = "rust1", since = "1.0.0")]
 317     pub unsafe fn from_utf8_unchecked(bytes: Vec<u8>) -> String {
 318         String { vec: bytes }
 319     }
 320
 321     /// Returns the underlying byte buffer, encoded as UTF-8.
 322     ///
 323     /// # Examples
 324     ///
 325     /// ```
 326     /// let s = String::from("hello");
 327     /// let bytes = s.into_bytes();
 328     /// assert_eq!(bytes, [104, 101, 108, 108, 111]);
 329     /// ```
 330     #[inline]
 331     #[stable(feature = "rust1", since = "1.0.0")]
 332     pub fn into_bytes(self) -> Vec<u8> {
 333         self.vec
 334     }
 335
 336     /// Extracts a string slice containing the entire string.
 337     #[inline]
 338     #[unstable(feature = "convert",
 339                reason = "waiting on RFC revision",
 340                issue = "27729")]
 341     pub fn as_str(&self) -> &str {
 342         self
 343     }
 344
 345     /// Pushes the given string onto this string buffer.
 346     ///
 347     /// # Examples
 348     ///
 349     /// ```
 350     /// let mut s = String::from("foo");
 351     /// s.push_str("bar");
 352     /// assert_eq!(s, "foobar");
 353     /// ```
 354     #[inline]
 355     #[stable(feature = "rust1", since = "1.0.0")]
 356     pub fn push_str(&mut self, string: &str) {
 357         self.vec.push_all(string.as_bytes())
 358     }
 359
 360     /// Returns the number of bytes that this string buffer can hold without
 361     /// reallocating.
 362     ///
 363     /// # Examples
 364     ///
 365     /// ```
 366     /// let s = String::with_capacity(10);
 367     /// assert!(s.capacity() >= 10);
 368     /// ```
 369     #[inline]
 370     #[stable(feature = "rust1", since = "1.0.0")]
 371     pub fn capacity(&self) -> usize {
 372         self.vec.capacity()
 373     }
 374
 375     /// Reserves capacity for at least `additional` more bytes to be inserted
 376     /// in the given `String`. The collection may reserve more space to avoid
 377     /// frequent reallocations.
 378     ///
 379     /// # Panics
 380     ///
 381     /// Panics if the new capacity overflows `usize`.
 382     ///
 383     /// # Examples
 384     ///
 385     /// ```
 386     /// let mut s = String::new();
 387     /// s.reserve(10);
 388     /// assert!(s.capacity() >= 10);
 389     /// ```
 390     #[inline]
 391     #[stable(feature = "rust1", since = "1.0.0")]
 392     pub fn reserve(&mut self, additional: usize) {
 393         self.vec.reserve(additional)
 394     }
 395
 396     /// Reserves the minimum capacity for exactly `additional` more bytes to be
 397     /// inserted in the given `String`. Does nothing if the capacity is already
 398     /// sufficient.
 399     ///
 400     /// Note that the allocator may give the collection more space than it
 401     /// requests. Therefore capacity can not be relied upon to be precisely
 402     /// minimal. Prefer `reserve` if future insertions are expected.
 403     ///
 404     /// # Panics
 405     ///
 406     /// Panics if the new capacity overflows `usize`.
 407     ///
 408     /// # Examples
 409     ///
 410     /// ```
 411     /// let mut s = String::new();
 412     /// s.reserve_exact(10);
 413     /// assert!(s.capacity() >= 10);
 414     /// ```
 415     #[inline]
 416     #[stable(feature = "rust1", since = "1.0.0")]
 417     pub fn reserve_exact(&mut self, additional: usize) {
 418         self.vec.reserve_exact(additional)
 419     }
 420
 421     /// Shrinks the capacity of this string buffer to match its length.
 422     ///
 423     /// # Examples
 424     ///
 425     /// ```
 426     /// let mut s = String::from("foo");
 427     /// s.reserve(100);
 428     /// assert!(s.capacity() >= 100);
 429     /// s.shrink_to_fit();
 430     /// assert_eq!(s.capacity(), 3);
 431     /// ```
 432     #[inline]
 433     #[stable(feature = "rust1", since = "1.0.0")]
 434     pub fn shrink_to_fit(&mut self) {
 435         self.vec.shrink_to_fit()
 436     }
 437
 438     /// Adds the given character to the end of the string.
 439     ///
 440     /// # Examples
 441     ///
 442     /// ```
 443     /// let mut s = String::from("abc");
 444     /// s.push('1');
 445     /// s.push('2');
 446     /// s.push('3');
 447     /// assert_eq!(s, "abc123");
 448     /// ```
 449     #[inline]
 450     #[stable(feature = "rust1", since = "1.0.0")]
 451     pub fn push(&mut self, ch: char) {
 452         match ch.len_utf8() {
 453             1 => self.vec.push(ch as u8),
 454             ch_len => {
 455                 let cur_len = self.len();
 456                 // This may use up to 4 bytes.
 457                 self.vec.reserve(ch_len);
 458
 459                 unsafe {
 460                     // Attempt to not use an intermediate buffer by just pushing bytes
 461                     // directly onto this string.
 462                     let slice = slice::from_raw_parts_mut (
 463                         self.vec.as_mut_ptr().offset(cur_len as isize),
 464                         ch_len
 465                     );
 466                     let used = ch.encode_utf8(slice).unwrap_or(0);
 467                     self.vec.set_len(cur_len + used);
 468                 }
 469             }
 470         }
 471     }
 472
 473     /// Works with the underlying buffer as a byte slice.
 474     ///
 475     /// # Examples
 476     ///
 477     /// ```
 478     /// let s = String::from("hello");
 479     /// assert_eq!(s.as_bytes(), [104, 101, 108, 108, 111]);
 480     /// ```
 481     #[inline]
 482     #[stable(feature = "rust1", since = "1.0.0")]
 483     pub fn as_bytes(&self) -> &[u8] {
 484         &self.vec
 485     }
 486
 487     /// Shortens a string to the specified length.
 488     ///
 489     /// # Panics
 490     ///
 491     /// Panics if `new_len` > current length,
 492     /// or if `new_len` is not a character boundary.
 493     ///
 494     /// # Examples
 495     ///
 496     /// ```
 497     /// let mut s = String::from("hello");
 498     /// s.truncate(2);
 499     /// assert_eq!(s, "he");
 500     /// ```
 501     #[inline]
 502     #[stable(feature = "rust1", since = "1.0.0")]
 503     pub fn truncate(&mut self, new_len: usize) {
 504         assert!(self.is_char_boundary(new_len));
 505         self.vec.truncate(new_len)
 506     }
 507
 508     /// Removes the last character from the string buffer and returns it.
 509     /// Returns `None` if this string buffer is empty.
 510     ///
 511     /// # Examples
 512     ///
 513     /// ```
 514     /// let mut s = String::from("foo");
 515     /// assert_eq!(s.pop(), Some('o'));
 516     /// assert_eq!(s.pop(), Some('o'));
 517     /// assert_eq!(s.pop(), Some('f'));
 518     /// assert_eq!(s.pop(), None);
 519     /// ```
 520     #[inline]
 521     #[stable(feature = "rust1", since = "1.0.0")]
 522     pub fn pop(&mut self) -> Option<char> {
 523         let len = self.len();
 524         if len == 0 {
 525             return None
 526         }
 527
 528         let ch = self.char_at_reverse(len);
 529         unsafe {
 530             self.vec.set_len(len - ch.len_utf8());
 531         }
 532         Some(ch)
 533     }
 534
 535     /// Removes the character from the string buffer at byte position `idx` and
 536     /// returns it.
 537     ///
 538     /// # Warning
 539     ///
 540     /// This is an O(n) operation as it requires copying every element in the
 541     /// buffer.
 542     ///
 543     /// # Panics
 544     ///
 545     /// If `idx` does not lie on a character boundary, or if it is out of
 546     /// bounds, then this function will panic.
 547     ///
 548     /// # Examples
 549     ///
 550     /// ```
 551     /// let mut s = String::from("foo");
 552     /// assert_eq!(s.remove(0), 'f');
 553     /// assert_eq!(s.remove(1), 'o');
 554     /// assert_eq!(s.remove(0), 'o');
 555     /// ```
 556     #[inline]
 557     #[stable(feature = "rust1", since = "1.0.0")]
 558     pub fn remove(&mut self, idx: usize) -> char {
 559         let len = self.len();
 560         assert!(idx <= len);
 561
 562         let ch = self.char_at(idx);
 563         let next = idx + ch.len_utf8();
 564         unsafe {
 565             ptr::copy(self.vec.as_ptr().offset(next as isize),
 566                       self.vec.as_mut_ptr().offset(idx as isize),
 567                       len - next);
 568             self.vec.set_len(len - (next - idx));
 569         }
 570         ch
 571     }
 572
 573     /// Inserts a character into the string buffer at byte position `idx`.
 574     ///
 575     /// # Warning
 576     ///
 577     /// This is an O(n) operation as it requires copying every element in the
 578     /// buffer.
 579     ///
 580     /// # Panics
 581     ///
 582     /// If `idx` does not lie on a character boundary or is out of bounds, then
 583     /// this function will panic.
 584     #[inline]
 585     #[stable(feature = "rust1", since = "1.0.0")]
 586     pub fn insert(&mut self, idx: usize, ch: char) {
 587         let len = self.len();
 588         assert!(idx <= len);
 589         assert!(self.is_char_boundary(idx));
 590         self.vec.reserve(4);
 591         let mut bits = [0; 4];
 592         let amt = ch.encode_utf8(&mut bits).unwrap();
 593
 594         unsafe {
 595             ptr::copy(self.vec.as_ptr().offset(idx as isize),
 596                       self.vec.as_mut_ptr().offset((idx + amt) as isize),
 597                       len - idx);
 598             ptr::copy(bits.as_ptr(),
 599                       self.vec.as_mut_ptr().offset(idx as isize),
 600                       amt);
 601             self.vec.set_len(len + amt);
 602         }
 603     }
 604
 605     /// Views the string buffer as a mutable sequence of bytes.
 606     ///
 607     /// This is unsafe because it does not check
 608     /// to ensure that the resulting string will be valid UTF-8.
 609     ///
 610     /// # Examples
 611     ///
 612     /// ```
 613     /// let mut s = String::from("hello");
 614     /// unsafe {
 615     ///     let vec = s.as_mut_vec();
 616     ///     assert!(vec == &[104, 101, 108, 108, 111]);
 617     ///     vec.reverse();
 618     /// }
 619     /// assert_eq!(s, "olleh");
 620     /// ```
 621     #[inline]
 622     #[stable(feature = "rust1", since = "1.0.0")]
 623     pub unsafe fn as_mut_vec(&mut self) -> &mut Vec<u8> {
 624         &mut self.vec
 625     }
 626
 627     /// Returns the number of bytes in this string.
 628     ///
 629     /// # Examples
 630     ///
 631     /// ```
 632     /// let a = "foo".to_string();
 633     /// assert_eq!(a.len(), 3);
 634     /// ```
 635     #[inline]
 636     #[stable(feature = "rust1", since = "1.0.0")]
 637     pub fn len(&self) -> usize { self.vec.len() }
 638
 639     /// Returns true if the string contains no bytes
 640     ///
 641     /// # Examples
 642     ///
 643     /// ```
 644     /// let mut v = String::new();
 645     /// assert!(v.is_empty());
 646     /// v.push('a');
 647     /// assert!(!v.is_empty());
 648     /// ```
 649     #[inline]
 650     #[stable(feature = "rust1", since = "1.0.0")]
 651     pub fn is_empty(&self) -> bool { self.len() == 0 }
 652
 653     /// Truncates the string, returning it to 0 length.
 654     ///
 655     /// # Examples
 656     ///
 657     /// ```
 658     /// let mut s = "foo".to_string();
 659     /// s.clear();
 660     /// assert!(s.is_empty());
 661     /// ```
 662     #[inline]
 663     #[stable(feature = "rust1", since = "1.0.0")]
 664     pub fn clear(&mut self) {
 665         self.vec.clear()
 666     }
 667
 668     /// Create a draining iterator that removes the specified range in the string
 669     /// and yields the removed chars from start to end. The element range is
 670     /// removed even if the iterator is not consumed until the end.
 671     ///
 672     /// # Panics
 673     ///
 674     /// Panics if the starting point or end point are not on character boundaries,
 675     /// or if they are out of bounds.
 676     ///
 677     /// # Examples
 678     ///
 679     /// ```
 680     /// #![feature(drain)]
 681     ///
 682     /// let mut s = String::from("α is alpha, β is beta");
 683     /// let beta_offset = s.find('β').unwrap_or(s.len());
 684     ///
 685     /// // Remove the range up until the β from the string
 686     /// let t: String = s.drain(..beta_offset).collect();
 687     /// assert_eq!(t, "α is alpha, ");
 688     /// assert_eq!(s, "β is beta");
 689     ///
 690     /// // A full range clears the string
 691     /// s.drain(..);
 692     /// assert_eq!(s, "");
 693     /// ```
 694     #[unstable(feature = "drain",
 695                reason = "recently added, matches RFC",
 696                issue = "27711")]
 697     pub fn drain<R>(&mut self, range: R) -> Drain where R: RangeArgument<usize> {
 698         // Memory safety
 699         //
 700         // The String version of Drain does not have the memory safety issues
 701         // of the vector version. The data is just plain bytes.
 702         // Because the range removal happens in Drop, if the Drain iterator is leaked,
 703         // the removal will not happen.
 704         let len = self.len();
 705         let start = *range.start().unwrap_or(&0);
 706         let end = *range.end().unwrap_or(&len);
 707
 708         // Take out two simultaneous borrows. The &mut String won't be accessed
 709         // until iteration is over, in Drop.
 710         let self_ptr = self as *mut _;
 711         // slicing does the appropriate bounds checks
 712         let chars_iter = self[start..end].chars();
 713
 714         Drain {
 715             start: start,
 716             end: end,
 717             iter: chars_iter,
 718             string: self_ptr,
 719         }
 720     }
 721
 722     /// Converts the string into `Box<str>`.
 723     ///
 724     /// Note that this will drop any excess capacity.
 725     #[stable(feature = "box_str", since = "1.4.0")]
 726     pub fn into_boxed_str(self) -> Box<str> {
 727         let slice = self.vec.into_boxed_slice();
 728         unsafe { mem::transmute::<Box<[u8]>, Box<str>>(slice) }
 729     }
 730
 731     /// Converts the string into `Box<str>`.
 732     ///
 733     /// Note that this will drop any excess capacity.
 734     #[unstable(feature = "box_str2",
 735                reason = "recently added, matches RFC",
 736                issue = "27785")]
 737     #[deprecated(since = "1.4.0", reason = "renamed to `into_boxed_str`")]
 738     pub fn into_boxed_slice(self) -> Box<str> {
 739         self.into_boxed_str()
 740     }
 741 }
 742
 743 impl FromUtf8Error {
 744     /// Consumes this error, returning the bytes that were attempted to make a
 745     /// `String` with.
 746     #[stable(feature = "rust1", since = "1.0.0")]
 747     pub fn into_bytes(self) -> Vec<u8> { self.bytes }
 748
 749     /// Access the underlying UTF8-error that was the cause of this error.
 750     #[stable(feature = "rust1", since = "1.0.0")]
 751     pub fn utf8_error(&self) -> Utf8Error { self.error }
 752 }
 753
 754 #[stable(feature = "rust1", since = "1.0.0")]
 755 impl fmt::Display for FromUtf8Error {
 756     fn fmt(&self, f: &mut fmt::Formatter) -> fmt::Result {
 757         fmt::Display::fmt(&self.error, f)
 758     }
 759 }
 760
 761 #[stable(feature = "rust1", since = "1.0.0")]
 762 impl fmt::Display for FromUtf16Error {
 763     fn fmt(&self, f: &mut fmt::Formatter) -> fmt::Result {
 764         fmt::Display::fmt("invalid utf-16: lone surrogate found", f)
 765     }
 766 }
 767
 768 #[stable(feature = "rust1", since = "1.0.0")]
 769 impl FromIterator<char> for String {
 770     fn from_iter<I: IntoIterator<Item=char>>(iterable: I) -> String {
 771         let mut buf = String::new();
 772         buf.extend(iterable);
 773         buf
 774     }
 775 }
 776
 777 #[stable(feature = "rust1", since = "1.0.0")]
 778 impl<'a> FromIterator<&'a str> for String {
 779     fn from_iter<I: IntoIterator<Item=&'a str>>(iterable: I) -> String {
 780         let mut buf = String::new();
 781         buf.extend(iterable);
 782         buf
 783     }
 784 }
 785
 786 #[stable(feature = "extend_string", since = "1.4.0")]
 787 impl FromIterator<String> for String {
 788     fn from_iter<I: IntoIterator<Item=String>>(iterable: I) -> String {
 789         let mut buf = String::new();
 790         buf.extend(iterable);
 791         buf
 792     }
 793 }
 794
 795 #[stable(feature = "rust1", since = "1.0.0")]
 796 impl Extend<char> for String {
 797     fn extend<I: IntoIterator<Item=char>>(&mut self, iterable: I) {
 798         let iterator = iterable.into_iter();
 799         let (lower_bound, _) = iterator.size_hint();
 800         self.reserve(lower_bound);
 801         for ch in iterator {
 802             self.push(ch)
 803         }
 804     }
 805 }
 806
 807 #[stable(feature = "extend_ref", since = "1.2.0")]
 808 impl<'a> Extend<&'a char> for String {
 809     fn extend<I: IntoIterator<Item=&'a char>>(&mut self, iterable: I) {
 810         self.extend(iterable.into_iter().cloned());
 811     }
 812 }
 813
 814 #[stable(feature = "rust1", since = "1.0.0")]
 815 impl<'a> Extend<&'a str> for String {
 816     fn extend<I: IntoIterator<Item=&'a str>>(&mut self, iterable: I) {
 817         for s in iterable {
 818             self.push_str(s)
 819         }
 820     }
 821 }
 822
 823 #[stable(feature = "extend_string", since = "1.4.0")]
 824 impl Extend<String> for String {
 825     fn extend<I: IntoIterator<Item=String>>(&mut self, iterable: I) {
 826         for s in iterable {
 827             self.push_str(&s)
 828         }
 829     }
 830 }
 831
 832 /// A convenience impl that delegates to the impl for `&str`
 833 impl<'a, 'b> Pattern<'a> for &'b String {
 834     type Searcher = <&'b str as Pattern<'a>>::Searcher;
 835
 836     fn into_searcher(self, haystack: &'a str) -> <&'b str as Pattern<'a>>::Searcher {
 837         self[..].into_searcher(haystack)
 838     }
 839
 840     #[inline]
 841     fn is_contained_in(self, haystack: &'a str) -> bool {
 842         self[..].is_contained_in(haystack)
 843     }
 844
 845     #[inline]
 846     fn is_prefix_of(self, haystack: &'a str) -> bool {
 847         self[..].is_prefix_of(haystack)
 848     }
 849 }
 850
 851 #[stable(feature = "rust1", since = "1.0.0")]
 852 impl PartialEq for String {
 853     #[inline]
 854     fn eq(&self, other: &String) -> bool { PartialEq::eq(&self[..], &other[..]) }
 855     #[inline]
 856     fn ne(&self, other: &String) -> bool { PartialEq::ne(&self[..], &other[..]) }
 857 }
 858
 859 macro_rules! impl_eq {
 860     ($lhs:ty, $rhs: ty) => {
 861         #[stable(feature = "rust1", since = "1.0.0")]
 862         impl<'a> PartialEq<$rhs> for $lhs {
 863             #[inline]
 864             fn eq(&self, other: &$rhs) -> bool { PartialEq::eq(&self[..], &other[..]) }
 865             #[inline]
 866             fn ne(&self, other: &$rhs) -> bool { PartialEq::ne(&self[..], &other[..]) }
 867         }
 868
 869         #[stable(feature = "rust1", since = "1.0.0")]
 870         impl<'a> PartialEq<$lhs> for $rhs {
 871             #[inline]
 872             fn eq(&self, other: &$lhs) -> bool { PartialEq::eq(&self[..], &other[..]) }
 873             #[inline]
 874             fn ne(&self, other: &$lhs) -> bool { PartialEq::ne(&self[..], &other[..]) }
 875         }
 876
 877     }
 878 }
 879
 880 impl_eq! { String, str }
 881 impl_eq! { String, &'a str }
 882 impl_eq! { Cow<'a, str>, str }
 883 impl_eq! { Cow<'a, str>, String }
 884
 885 #[stable(feature = "rust1", since = "1.0.0")]
 886 impl<'a, 'b> PartialEq<&'b str> for Cow<'a, str> {
 887     #[inline]
 888     fn eq(&self, other: &&'b str) -> bool { PartialEq::eq(&self[..], &other[..]) }
 889     #[inline]
 890     fn ne(&self, other: &&'b str) -> bool { PartialEq::ne(&self[..], &other[..]) }
 891 }
 892
 893 #[stable(feature = "rust1", since = "1.0.0")]
 894 impl<'a, 'b> PartialEq<Cow<'a, str>> for &'b str {
 895     #[inline]
 896     fn eq(&self, other: &Cow<'a, str>) -> bool { PartialEq::eq(&self[..], &other[..]) }
 897     #[inline]
 898     fn ne(&self, other: &Cow<'a, str>) -> bool { PartialEq::ne(&self[..], &other[..]) }
 899 }
 900
 901 #[stable(feature = "rust1", since = "1.0.0")]
 902 impl Default for String {
 903     #[inline]
 904     #[stable(feature = "rust1", since = "1.0.0")]
 905     fn default() -> String {
 906         String::new()
 907     }
 908 }
 909
 910 #[stable(feature = "rust1", since = "1.0.0")]
 911 impl fmt::Display for String {
 912     #[inline]
 913     fn fmt(&self, f: &mut fmt::Formatter) -> fmt::Result {
 914         fmt::Display::fmt(&**self, f)
 915     }
 916 }
 917
 918 #[stable(feature = "rust1", since = "1.0.0")]
 919 impl fmt::Debug for String {
 920     #[inline]
 921     fn fmt(&self, f: &mut fmt::Formatter) -> fmt::Result {
 922         fmt::Debug::fmt(&**self, f)
 923     }
 924 }
 925
 926 #[stable(feature = "rust1", since = "1.0.0")]
 927 impl hash::Hash for String {
 928     #[inline]
 929     fn hash<H: hash::Hasher>(&self, hasher: &mut H) {
 930         (**self).hash(hasher)
 931     }
 932 }
 933
 934 #[stable(feature = "rust1", since = "1.0.0")]
 935 impl<'a> Add<&'a str> for String {
 936     type Output = String;
 937
 938     #[inline]
 939     fn add(mut self, other: &str) -> String {
 940         self.push_str(other);
 941         self
 942     }
 943 }
 944
 945 #[stable(feature = "rust1", since = "1.0.0")]
 946 impl ops::Index<ops::Range<usize>> for String {
 947     type Output = str;
 948
 949     #[inline]
 950     fn index(&self, index: ops::Range<usize>) -> &str {
 951         &self[..][index]
 952     }
 953 }
 954 #[stable(feature = "rust1", since = "1.0.0")]
 955 impl ops::Index<ops::RangeTo<usize>> for String {
 956     type Output = str;
 957
 958     #[inline]
 959     fn index(&self, index: ops::RangeTo<usize>) -> &str {
 960         &self[..][index]
 961     }
 962 }
 963 #[stable(feature = "rust1", since = "1.0.0")]
 964 impl ops::Index<ops::RangeFrom<usize>> for String {
 965     type Output = str;
 966
 967     #[inline]
 968     fn index(&self, index: ops::RangeFrom<usize>) -> &str {
 969         &self[..][index]
 970     }
 971 }
 972 #[stable(feature = "rust1", since = "1.0.0")]
 973 impl ops::Index<ops::RangeFull> for String {
 974     type Output = str;
 975
 976     #[inline]
 977     fn index(&self, _index: ops::RangeFull) -> &str {
 978         unsafe { str::from_utf8_unchecked(&self.vec) }
 979     }
 980 }
 981
 982 #[stable(feature = "derefmut_for_string", since = "1.2.0")]
 983 impl ops::IndexMut<ops::Range<usize>> for String {
 984     #[inline]
 985     fn index_mut(&mut self, index: ops::Range<usize>) -> &mut str {
 986         &mut self[..][index]
 987     }
 988 }
 989 #[stable(feature = "derefmut_for_string", since = "1.2.0")]
 990 impl ops::IndexMut<ops::RangeTo<usize>> for String {
 991     #[inline]
 992     fn index_mut(&mut self, index: ops::RangeTo<usize>) -> &mut str {
 993         &mut self[..][index]
 994     }
 995 }
 996 #[stable(feature = "derefmut_for_string", since = "1.2.0")]
 997 impl ops::IndexMut<ops::RangeFrom<usize>> for String {
 998     #[inline]
 999     fn index_mut(&mut self, index: ops::RangeFrom<usize>) -> &mut str {
1000         &mut self[..][index]
1001     }
1002 }
1003 #[stable(feature = "derefmut_for_string", since = "1.2.0")]
1004 impl ops::IndexMut<ops::RangeFull> for String {
1005     #[inline]
1006     fn index_mut(&mut self, _index: ops::RangeFull) -> &mut str {
1007         unsafe { mem::transmute(&mut *self.vec) }
1008     }
1009 }
1010
1011 #[stable(feature = "rust1", since = "1.0.0")]
1012 impl ops::Deref for String {
1013     type Target = str;
1014
1015     #[inline]
1016     fn deref(&self) -> &str {
1017         unsafe { str::from_utf8_unchecked(&self.vec) }
1018     }
1019 }
1020
1021 #[stable(feature = "derefmut_for_string", since = "1.2.0")]
1022 impl ops::DerefMut for String {
1023     #[inline]
1024     fn deref_mut(&mut self) -> &mut str {
1025         unsafe { mem::transmute(&mut *self.vec) }
1026     }
1027 }
1028
1029 /// Error returned from `String::from`
1030 #[unstable(feature = "str_parse_error", reason = "may want to be replaced with \
1031                                                   Void if it ever exists",
1032            issue = "27734")]
1033 #[derive(Copy, Clone, Debug, Eq, PartialEq)]
1034 pub struct ParseError(());
1035
1036 #[stable(feature = "rust1", since = "1.0.0")]
1037 impl FromStr for String {
1038     type Err = ParseError;
1039     #[inline]
1040     fn from_str(s: &str) -> Result<String, ParseError> {
1041         Ok(String::from(s))
1042     }
1043 }
1044
1045 /// A generic trait for converting a value to a string
1046 #[stable(feature = "rust1", since = "1.0.0")]
1047 pub trait ToString {
1048     /// Converts the value of `self` to an owned string
1049     #[stable(feature = "rust1", since = "1.0.0")]
1050     fn to_string(&self) -> String;
1051 }
1052
1053 #[stable(feature = "rust1", since = "1.0.0")]
1054 impl<T: fmt::Display + ?Sized> ToString for T {
1055     #[inline]
1056     fn to_string(&self) -> String {
1057         use core::fmt::Write;
1058         let mut buf = String::new();
1059         let _ = buf.write_fmt(format_args!("{}", self));
1060         buf.shrink_to_fit();
1061         buf
1062     }
1063 }
1064
1065 #[stable(feature = "rust1", since = "1.0.0")]
1066 impl AsRef<str> for String {
1067     #[inline]
1068     fn as_ref(&self) -> &str {
1069         self
1070     }
1071 }
1072
1073 #[stable(feature = "rust1", since = "1.0.0")]
1074 impl AsRef<[u8]> for String {
1075     #[inline]
1076     fn as_ref(&self) -> &[u8] {
1077         self.as_bytes()
1078     }
1079 }
1080
1081 #[stable(feature = "rust1", since = "1.0.0")]
1082 impl<'a> From<&'a str> for String {
1083     #[cfg(not(test))]
1084     #[inline]
1085     fn from(s: &'a str) -> String {
1086         String { vec: <[_]>::to_vec(s.as_bytes()) }
1087     }
1088
1089     // HACK(japaric): with cfg(test) the inherent `[T]::to_vec` method, which is
1090     // required for this method definition, is not available. Since we don't
1091     // require this method for testing purposes, I'll just stub it
1092     // NB see the slice::hack module in slice.rs for more information
1093     #[inline]
1094     #[cfg(test)]
1095     fn from(_: &str) -> String {
1096         panic!("not available with cfg(test)");
1097     }
1098 }
1099
1100 #[stable(feature = "rust1", since = "1.0.0")]
1101 impl<'a> From<&'a str> for Cow<'a, str> {
1102     #[inline]
1103     fn from(s: &'a str) -> Cow<'a, str> {
1104         Cow::Borrowed(s)
1105     }
1106 }
1107
1108 #[stable(feature = "rust1", since = "1.0.0")]
1109 impl<'a> From<String> for Cow<'a, str> {
1110     #[inline]
1111     fn from(s: String) -> Cow<'a, str> {
1112         Cow::Owned(s)
1113     }
1114 }
1115
1116 #[stable(feature = "rust1", since = "1.0.0")]
1117 impl Into<Vec<u8>> for String {
1118     fn into(self) -> Vec<u8> {
1119         self.into_bytes()
1120     }
1121 }
1122
1123 #[unstable(feature = "into_cow", reason = "may be replaced by `convert::Into`",
1124            issue= "27735")]
1125 impl IntoCow<'static, str> for String {
1126     #[inline]
1127     fn into_cow(self) -> Cow<'static, str> {
1128         Cow::Owned(self)
1129     }
1130 }
1131
1132 #[unstable(feature = "into_cow", reason = "may be replaced by `convert::Into`",
1133            issue = "27735")]
1134 impl<'a> IntoCow<'a, str> for &'a str {
1135     #[inline]
1136     fn into_cow(self) -> Cow<'a, str> {
1137         Cow::Borrowed(self)
1138     }
1139 }
1140
1141 #[stable(feature = "rust1", since = "1.0.0")]
1142 impl fmt::Write for String {
1143     #[inline]
1144     fn write_str(&mut self, s: &str) -> fmt::Result {
1145         self.push_str(s);
1146         Ok(())
1147     }
1148
1149     #[inline]
1150     fn write_char(&mut self, c: char) -> fmt::Result {
1151         self.push(c);
1152         Ok(())
1153     }
1154 }
1155
1156 /// A draining iterator for `String`.
1157 #[unstable(feature = "drain", reason = "recently added", issue = "27711")]
1158 pub struct Drain<'a> {
1159     /// Will be used as &'a mut String in the destructor
1160     string: *mut String,
1161     /// Start of part to remove
1162     start: usize,
1163     /// End of part to remove
1164     end: usize,
1165     /// Current remaining range to remove
1166     iter: Chars<'a>,
1167 }
1168
1169 unsafe impl<'a> Sync for Drain<'a> {}
1170 unsafe impl<'a> Send for Drain<'a> {}
1171
1172 #[unstable(feature = "drain", reason = "recently added", issue = "27711")]
1173 impl<'a> Drop for Drain<'a> {
1174     fn drop(&mut self) {
1175         unsafe {
1176             // Use Vec::drain. "Reaffirm" the bounds checks to avoid
1177             // panic code being inserted again.
1178             let self_vec = (*self.string).as_mut_vec();
1179             if self.start <= self.end && self.end <= self_vec.len() {
1180                 self_vec.drain(self.start..self.end);
1181             }
1182         }
1183     }
1184 }
1185
1186 #[unstable(feature = "drain", reason = "recently added", issue = "27711")]
1187 impl<'a> Iterator for Drain<'a> {
1188     type Item = char;
1189
1190     #[inline]
1191     fn next(&mut self) -> Option<char> {
1192         self.iter.next()
1193     }
1194
1195     fn size_hint(&self) -> (usize, Option<usize>) {
1196         self.iter.size_hint()
1197     }
1198 }
1199
1200 #[unstable(feature = "drain", reason = "recently added", issue = "27711")]
1201 impl<'a> DoubleEndedIterator for Drain<'a> {
1202     #[inline]
1203     fn next_back(&mut self) -> Option<char> {
1204         self.iter.next_back()
1205     }
1206 }