1
   2
   3
   4
   5
   6
   7
   8
   9
  10
  11
  12
  13
  14
  15
  16
  17
  18
  19
  20
  21
  22
  23
  24
  25
  26
  27
  28
  29
  30
  31
  32
  33
  34
  35
  36
  37
  38
  39
  40
  41
  42
  43
  44
  45
  46
  47
  48
  49
  50
  51
  52
  53
  54
  55
  56
  57
  58
  59
  60
  61
  62
  63
  64
  65
  66
  67
  68
  69
  70
  71
  72
  73
  74
  75
  76
  77
  78
  79
  80
  81
  82
  83
  84
  85
  86
  87
  88
  89
  90
  91
  92
  93
  94
  95
  96
  97
  98
  99
 100
 101
 102
 103
 104
 105
 106
 107
 108
 109
 110
 111
 112
 113
 114
 115
 116
 117
 118
 119
 120
 121
 122
 123
 124
 125
 126
 127
 128
 129
 130
 131
 132
 133
 134
 135
 136
 137
 138
 139
 140
 141
 142
 143
 144
 145
 146
 147
 148
 149
 150
 151
 152
 153
 154
 155
 156
 157
 158
 159
 160
 161
 162
 163
 164
 165
 166
 167
 168
 169
 170
 171
 172
 173
 174
 175
 176
 177
 178
 179
 180
 181
 182
 183
 184
 185
 186
 187
 188
 189
 190
 191
 192
 193
 194
 195
 196
 197
 198
 199
 200
 201
 202
 203
 204
 205
 206
 207
 208
 209
 210
 211
 212
 213
 214
 215
 216
 217
 218
 219
 220
 221
 222
 223
 224
 225
 226
 227
 228
 229
 230
 231
 232
 233
 234
 235
 236
 237
 238
 239
 240
 241
 242
 243
 244
 245
 246
 247
 248
 249
 250
 251
 252
 253
 254
 255
 256
 257
 258
 259
 260
 261
 262
 263
 264
 265
 266
 267
 268
 269
 270
 271
 272
 273
 274
 275
 276
 277
 278
 279
 280
 281
 282
 283
 284
 285
 286
 287
 288
 289
 290
 291
 292
 293
 294
 295
 296
 297
 298
 299
 300
 301
 302
 303
 304
 305
 306
 307
 308
 309
 310
 311
 312
 313
 314
 315
 316
 317
 318
 319
 320
 321
 322
 323
 324
 325
 326
 327
 328
 329
 330
 331
 332
 333
 334
 335
 336
 337
 338
 339
 340
 341
 342
 343
 344
 345
 346
 347
 348
 349
 350
 351
 352
 353
 354
 355
 356
 357
 358
 359
 360
 361
 362
 363
 364
 365
 366
 367
 368
 369
 370
 371
 372
 373
 374
 375
 376
 377
 378
 379
 380
 381
 382
 383
 384
 385
 386
 387
 388
 389
 390
 391
 392
 393
 394
 395
 396
 397
 398
 399
 400
 401
 402
 403
 404
 405
 406
 407
 408
 409
 410
 411
 412
 413
 414
 415
 416
 417
 418
 419
 420
 421
 422
 423
 424
 425
 426
 427
 428
 429
 430
 431
 432
 433
 434
 435
 436
 437
 438
 439
 440
 441
 442
 443
 444
 445
 446
 447
 448
 449
 450
 451
 452
 453
 454
 455
 456
 457
 458
 459
 460
 461
 462
 463
 464
 465
 466
 467
 468
 469
 470
 471
 472
 473
 474
 475
 476
 477
 478
 479
 480
 481
 482
 483
 484
 485
 486
 487
 488
 489
 490
 491
 492
 493
 494
 495
 496
 497
 498
 499
 500
 501
 502
 503
 504
 505
 506
 507
 508
 509
 510
 511
 512
 513
 514
 515
 516
 517
 518
 519
 520
 521
 522
 523
 524
 525
 526
 527
 528
 529
 530
 531
 532
 533
 534
 535
 536
 537
 538
 539
 540
 541
 542
 543
 544
 545
 546
 547
 548
 549
 550
 551
 552
 553
 554
 555
 556
 557
 558
 559
 560
 561
 562
 563
 564
 565
 566
 567
 568
 569
 570
 571
 572
 573
 574
 575
 576
 577
 578
 579
 580
 581
 582
 583
 584
 585
 586
 587
 588
 589
 590
 591
 592
 593
 594
 595
 596
 597
 598
 599
 600
 601
 602
 603
 604
 605
 606
 607
 608
 609
 610
 611
 612
 613
 614
 615
 616
 617
 618
 619
 620
 621
 622
 623
 624
 625
 626
 627
 628
 629
 630
 631
 632
 633
 634
 635
 636
 637
 638
 639
 640
 641
 642
 643
 644
 645
 646
 647
 648
 649
 650
 651
 652
 653
 654
 655
 656
 657
 658
 659
 660
 661
 662
 663
 664
 665
 666
 667
 668
 669
 670
 671
 672
 673
 674
 675
 676
 677
 678
 679
 680
 681
 682
 683
 684
 685
 686
 687
 688
 689
 690
 691
 692
 693
 694
 695
 696
 697
 698
 699
 700
 701
 702
 703
 704
 705
 706
 707
 708
 709
 710
 711
 712
 713
 714
 715
 716
 717
 718
 719
 720
 721
 722
 723
 724
 725
 726
 727
 728
 729
 730
 731
 732
 733
 734
 735
 736
 737
 738
 739
 740
 741
 742
 743
 744
 745
 746
 747
 748
 749
 750
 751
 752
 753
 754
 755
 756
 757
 758
 759
 760
 761
 762
 763
 764
 765
 766
 767
 768
 769
 770
 771
 772
 773
 774
 775
 776
 777
 778
 779
 780
 781
 782
 783
 784
 785
 786
 787
 788
 789
 790
 791
 792
 793
 794
 795
 796
 797
 798
 799
 800
 801
 802
 803
 804
 805
 806
 807
 808
 809
 810
 811
 812
 813
 814
 815
 816
 817
 818
 819
 820
 821
 822
 823
 824
 825
 826
 827
 828
 829
 830
 831
 832
 833
 834
 835
 836
 837
 838
 839
 840
 841
 842
 843
 844
 845
 846
 847
 848
 849
 850
 851
 852
 853
 854
 855
 856
 857
 858
 859
 860
 861
 862
 863
 864
 865
 866
 867
 868
 869
 870
 871
 872
 873
 874
 875
 876
 877
 878
 879
 880
 881
 882
 883
 884
 885
 886
 887
 888
 889
 890
 891
 892
 893
 894
 895
 896
 897
 898
 899
 900
 901
 902
 903
 904
 905
 906
 907
 908
 909
 910
 911
 912
 913
 914
 915
 916
 917
 918
 919
 920
 921
 922
 923
 924
 925
 926
 927
 928
 929
 930
 931
 932
 933
 934
 935
 936
 937
 938
 939
 940
 941
 942
 943
 944
 945
 946
 947
 948
 949
 950
 951
 952
 953
 954
 955
 956
 957
 958
 959
 960
 961
 962
 963
 964
 965
 966
 967
 968
 969
 970
 971
 972
 973
 974
 975
 976
 977
 978
 979
 980
 981
 982
 983
 984
 985
 986
 987
 988
 989
 990
 991
 992
 993
 994
 995
 996
 997
 998
 999
1000
1001
1002
1003
1004
1005
1006
1007
1008
1009
1010
1011
1012
1013
1014
1015
1016
1017
1018
1019
1020
1021
1022
1023
1024
1025
1026
1027
1028
1029
1030
1031
1032
1033
1034
1035
1036
1037
1038
1039
1040
1041
1042
1043
1044
1045
1046
1047
1048
1049
1050
1051
1052
1053
1054
1055
1056
1057
1058
1059
1060
1061
1062
1063
1064
1065
1066
1067
1068
1069
1070
1071
1072
// Copyright 2014-2015 The Rust Project Developers. See the COPYRIGHT
// file at the top-level directory of this distribution and at
// http://rust-lang.org/COPYRIGHT.
//
// Licensed under the Apache License, Version 2.0 <LICENSE-APACHE or
// http://www.apache.org/licenses/LICENSE-2.0> or the MIT license
// <LICENSE-MIT or http://opensource.org/licenses/MIT>, at your
// option. This file may not be copied, modified, or distributed
// except according to those terms.

use std::borrow::Cow;
use std::collections::HashMap;
use std::fmt;
use std::ops::Index;
use std::str::FromStr;
use std::sync::Arc;

use memchr::memchr;

use exec::{Exec, ExecNoSync};
use expand::expand_bytes;
use error::Error;
use re_builder::bytes::RegexBuilder;
use re_trait::{self, RegularExpression, Locations, SubCapturesPosIter};

/// Match represents a single match of a regex in a haystack.
///
/// The lifetime parameter `'t` refers to the lifetime of the matched text.
#[derive(Copy, Clone, Debug, Eq, PartialEq)]
pub struct Match<'t> {
    text: &'t [u8],
    start: usize,
    end: usize,
}

impl<'t> Match<'t> {
    /// Returns the starting byte offset of the match in the haystack.
    #[inline]
    pub fn start(&self) -> usize {
        self.start
    }

    /// Returns the ending byte offset of the match in the haystack.
    #[inline]
    pub fn end(&self) -> usize {
        self.end
    }

    /// Returns the matched text.
    #[inline]
    pub fn as_bytes(&self) -> &'t [u8] {
        &self.text[self.start..self.end]
    }

    /// Creates a new match from the given haystack and byte offsets.
    #[inline]
    fn new(haystack: &'t [u8], start: usize, end: usize) -> Match<'t> {
        Match {
            text: haystack,
            start: start,
            end: end,
        }
    }
}

/// A compiled regular expression for matching arbitrary bytes.
///
/// It can be used to search, split or replace text. All searching is done with
/// an implicit `.*?` at the beginning and end of an expression. To force an
/// expression to match the whole string (or a prefix or a suffix), you must
/// use an anchor like `^` or `$` (or `\A` and `\z`).
///
/// Like the `Regex` type in the parent module, matches with this regex return
/// byte offsets into the search text. **Unlike** the parent `Regex` type,
/// these byte offsets may not correspond to UTF-8 sequence boundaries since
/// the regexes in this module can match arbitrary bytes.
#[derive(Clone)]
pub struct Regex(Exec);

impl fmt::Display for Regex {
    /// Shows the original regular expression.
    fn fmt(&self, f: &mut fmt::Formatter) -> fmt::Result {
        write!(f, "{}", self.as_str())
    }
}

impl fmt::Debug for Regex {
    /// Shows the original regular expression.
    fn fmt(&self, f: &mut fmt::Formatter) -> fmt::Result {
        fmt::Display::fmt(self, f)
    }
}

/// A constructor for Regex from an Exec.
///
/// This is hidden because Exec isn't actually part of the public API.
#[doc(hidden)]
impl From<Exec> for Regex {
    fn from(exec: Exec) -> Regex {
        Regex(exec)
    }
}

impl FromStr for Regex {
    type Err = Error;

    /// Attempts to parse a string into a regular expression
    fn from_str(s: &str) -> Result<Regex, Error> {
        Regex::new(s)
    }
}

/// Core regular expression methods.
impl Regex {
    /// Compiles a regular expression. Once compiled, it can be used repeatedly
    /// to search, split or replace text in a string.
    ///
    /// If an invalid expression is given, then an error is returned.
    pub fn new(re: &str) -> Result<Regex, Error> {
        RegexBuilder::new(re).build()
    }

    /// Returns true if and only if the regex matches the string given.
    ///
    /// It is recommended to use this method if all you need to do is test
    /// a match, since the underlying matching engine may be able to do less
    /// work.
    ///
    /// # Example
    ///
    /// Test if some text contains at least one word with exactly 13 ASCII word
    /// bytes:
    ///
    /// ```rust
    /// # extern crate regex; use regex::bytes::Regex;
    /// # fn main() {
    /// let text = b"I categorically deny having triskaidekaphobia.";
    /// assert!(Regex::new(r"\b\w{13}\b").unwrap().is_match(text));
    /// # }
    /// ```
    pub fn is_match(&self, text: &[u8]) -> bool {
        self.is_match_at(text, 0)
    }

    /// Returns the start and end byte range of the leftmost-first match in
    /// `text`. If no match exists, then `None` is returned.
    ///
    /// Note that this should only be used if you want to discover the position
    /// of the match. Testing the existence of a match is faster if you use
    /// `is_match`.
    ///
    /// # Example
    ///
    /// Find the start and end location of the first word with exactly 13
    /// ASCII word bytes:
    ///
    /// ```rust
    /// # extern crate regex; use regex::bytes::Regex;
    /// # fn main() {
    /// let text = b"I categorically deny having triskaidekaphobia.";
    /// let mat = Regex::new(r"\b\w{13}\b").unwrap().find(text).unwrap();
    /// assert_eq!((mat.start(), mat.end()), (2, 15));
    /// # }
    /// ```
    pub fn find<'t>(&self, text: &'t [u8]) -> Option<Match<'t>> {
        self.find_at(text, 0)
    }

    /// Returns an iterator for each successive non-overlapping match in
    /// `text`, returning the start and end byte indices with respect to
    /// `text`.
    ///
    /// # Example
    ///
    /// Find the start and end location of every word with exactly 13 ASCII
    /// word bytes:
    ///
    /// ```rust
    /// # extern crate regex; use regex::bytes::Regex;
    /// # fn main() {
    /// let text = b"Retroactively relinquishing remunerations is reprehensible.";
    /// for mat in Regex::new(r"\b\w{13}\b").unwrap().find_iter(text) {
    ///     println!("{:?}", mat);
    /// }
    /// # }
    /// ```
    pub fn find_iter<'r, 't>(&'r self, text: &'t [u8]) -> Matches<'r, 't> {
        Matches(self.0.searcher().find_iter(text))
    }

    /// Returns the capture groups corresponding to the leftmost-first
    /// match in `text`. Capture group `0` always corresponds to the entire
    /// match. If no match is found, then `None` is returned.
    ///
    /// You should only use `captures` if you need access to the location of
    /// capturing group matches. Otherwise, `find` is faster for discovering
    /// the location of the overall match.
    ///
    /// # Examples
    ///
    /// Say you have some text with movie names and their release years,
    /// like "'Citizen Kane' (1941)". It'd be nice if we could search for text
    /// looking like that, while also extracting the movie name and its release
    /// year separately.
    ///
    /// ```rust
    /// # extern crate regex; use regex::bytes::Regex;
    /// # fn main() {
    /// let re = Regex::new(r"'([^']+)'\s+\((\d{4})\)").unwrap();
    /// let text = b"Not my favorite movie: 'Citizen Kane' (1941).";
    /// let caps = re.captures(text).unwrap();
    /// assert_eq!(&caps[1], &b"Citizen Kane"[..]);
    /// assert_eq!(&caps[2], &b"1941"[..]);
    /// assert_eq!(&caps[0], &b"'Citizen Kane' (1941)"[..]);
    /// // You can also access the groups by index using the Index notation.
    /// // Note that this will panic on an invalid index.
    /// assert_eq!(&caps[1], b"Citizen Kane");
    /// assert_eq!(&caps[2], b"1941");
    /// assert_eq!(&caps[0], b"'Citizen Kane' (1941)");
    /// # }
    /// ```
    ///
    /// Note that the full match is at capture group `0`. Each subsequent
    /// capture group is indexed by the order of its opening `(`.
    ///
    /// We can make this example a bit clearer by using *named* capture groups:
    ///
    /// ```rust
    /// # extern crate regex; use regex::bytes::Regex;
    /// # fn main() {
    /// let re = Regex::new(r"'(?P<title>[^']+)'\s+\((?P<year>\d{4})\)")
    ///                .unwrap();
    /// let text = b"Not my favorite movie: 'Citizen Kane' (1941).";
    /// let caps = re.captures(text).unwrap();
    /// assert_eq!(&caps["title"], &b"Citizen Kane"[..]);
    /// assert_eq!(&caps["year"], &b"1941"[..]);
    /// assert_eq!(&caps[0], &b"'Citizen Kane' (1941)"[..]);
    /// // You can also access the groups by name using the Index notation.
    /// // Note that this will panic on an invalid group name.
    /// assert_eq!(&caps["title"], b"Citizen Kane");
    /// assert_eq!(&caps["year"], b"1941");
    /// assert_eq!(&caps[0], b"'Citizen Kane' (1941)");
    ///
    /// # }
    /// ```
    ///
    /// Here we name the capture groups, which we can access with the `name`
    /// method or the `Index` notation with a `&str`. Note that the named
    /// capture groups are still accessible with `get` or the `Index` notation
    /// with a `usize`.
    ///
    /// The `0`th capture group is always unnamed, so it must always be
    /// accessed with `get(0)` or `[0]`.
    pub fn captures<'t>(&self, text: &'t [u8]) -> Option<Captures<'t>> {
        let mut locs = self.locations();
        self.read_captures_at(&mut locs, text, 0).map(|_| Captures {
            text: text,
            locs: locs,
            named_groups: self.0.capture_name_idx().clone(),
        })
    }

    /// Returns an iterator over all the non-overlapping capture groups matched
    /// in `text`. This is operationally the same as `find_iter`, except it
    /// yields information about capturing group matches.
    ///
    /// # Example
    ///
    /// We can use this to find all movie titles and their release years in
    /// some text, where the movie is formatted like "'Title' (xxxx)":
    ///
    /// ```rust
    /// # extern crate regex; use std::str; use regex::bytes::Regex;
    /// # fn main() {
    /// let re = Regex::new(r"'(?P<title>[^']+)'\s+\((?P<year>\d{4})\)")
    ///                .unwrap();
    /// let text = b"'Citizen Kane' (1941), 'The Wizard of Oz' (1939), 'M' (1931).";
    /// for caps in re.captures_iter(text) {
    ///     let title = str::from_utf8(&caps["title"]).unwrap();
    ///     let year = str::from_utf8(&caps["year"]).unwrap();
    ///     println!("Movie: {:?}, Released: {:?}", title, year);
    /// }
    /// // Output:
    /// // Movie: Citizen Kane, Released: 1941
    /// // Movie: The Wizard of Oz, Released: 1939
    /// // Movie: M, Released: 1931
    /// # }
    /// ```
    pub fn captures_iter<'r, 't>(
        &'r self,
        text: &'t [u8],
    ) -> CaptureMatches<'r, 't> {
        CaptureMatches(self.0.searcher().captures_iter(text))
    }

    /// Returns an iterator of substrings of `text` delimited by a match of the
    /// regular expression. Namely, each element of the iterator corresponds to
    /// text that *isn't* matched by the regular expression.
    ///
    /// This method will *not* copy the text given.
    ///
    /// # Example
    ///
    /// To split a string delimited by arbitrary amounts of spaces or tabs:
    ///
    /// ```rust
    /// # extern crate regex; use regex::bytes::Regex;
    /// # fn main() {
    /// let re = Regex::new(r"[ \t]+").unwrap();
    /// let fields: Vec<&[u8]> = re.split(b"a b \t  c\td    e").collect();
    /// assert_eq!(fields, vec![
    ///     &b"a"[..], &b"b"[..], &b"c"[..], &b"d"[..], &b"e"[..],
    /// ]);
    /// # }
    /// ```
    pub fn split<'r, 't>(&'r self, text: &'t [u8]) -> Split<'r, 't> {
        Split {
            finder: self.find_iter(text),
            last: 0,
        }
    }

    /// Returns an iterator of at most `limit` substrings of `text` delimited
    /// by a match of the regular expression. (A `limit` of `0` will return no
    /// substrings.) Namely, each element of the iterator corresponds to text
    /// that *isn't* matched by the regular expression. The remainder of the
    /// string that is not split will be the last element in the iterator.
    ///
    /// This method will *not* copy the text given.
    ///
    /// # Example
    ///
    /// Get the first two words in some text:
    ///
    /// ```rust
    /// # extern crate regex; use regex::bytes::Regex;
    /// # fn main() {
    /// let re = Regex::new(r"\W+").unwrap();
    /// let fields: Vec<&[u8]> = re.splitn(b"Hey! How are you?", 3).collect();
    /// assert_eq!(fields, vec![&b"Hey"[..], &b"How"[..], &b"are you?"[..]]);
    /// # }
    /// ```
    pub fn splitn<'r, 't>(
        &'r self,
        text: &'t [u8],
        limit: usize,
    ) -> SplitN<'r, 't> {
        SplitN {
            splits: self.split(text),
            n: limit,
        }
    }

    /// Replaces the leftmost-first match with the replacement provided. The
    /// replacement can be a regular byte string (where `$N` and `$name` are
    /// expanded to match capture groups) or a function that takes the matches'
    /// `Captures` and returns the replaced byte string.
    ///
    /// If no match is found, then a copy of the byte string is returned
    /// unchanged.
    ///
    /// # Replacement string syntax
    ///
    /// All instances of `$name` in the replacement text is replaced with the
    /// corresponding capture group `name`.
    ///
    /// `name` may be an integer corresponding to the index of the
    /// capture group (counted by order of opening parenthesis where `0` is the
    /// entire match) or it can be a name (consisting of letters, digits or
    /// underscores) corresponding to a named capture group.
    ///
    /// If `name` isn't a valid capture group (whether the name doesn't exist
    /// or isn't a valid index), then it is replaced with the empty string.
    ///
    /// The longest possible name is used. e.g., `$1a` looks up the capture
    /// group named `1a` and not the capture group at index `1`. To exert more
    /// precise control over the name, use braces, e.g., `${1}a`.
    ///
    /// To write a literal `$` use `$$`.
    ///
    /// # Examples
    ///
    /// Note that this function is polymorphic with respect to the replacement.
    /// In typical usage, this can just be a normal byte string:
    ///
    /// ```rust
    /// # extern crate regex; use regex::bytes::Regex;
    /// # fn main() {
    /// let re = Regex::new("[^01]+").unwrap();
    /// assert_eq!(re.replace(b"1078910", &b""[..]), &b"1010"[..]);
    /// # }
    /// ```
    ///
    /// But anything satisfying the `Replacer` trait will work. For example, a
    /// closure of type `|&Captures| -> Vec<u8>` provides direct access to the
    /// captures corresponding to a match. This allows one to access capturing
    /// group matches easily:
    ///
    /// ```rust
    /// # extern crate regex; use regex::bytes::Regex;
    /// # use regex::bytes::Captures; fn main() {
    /// let re = Regex::new(r"([^,\s]+),\s+(\S+)").unwrap();
    /// let result = re.replace(b"Springsteen, Bruce", |caps: &Captures| {
    ///     let mut replacement = caps[2].to_owned();
    ///     replacement.push(b' ');
    ///     replacement.extend(&caps[1]);
    ///     replacement
    /// });
    /// assert_eq!(result, &b"Bruce Springsteen"[..]);
    /// # }
    /// ```
    ///
    /// But this is a bit cumbersome to use all the time. Instead, a simple
    /// syntax is supported that expands `$name` into the corresponding capture
    /// group. Here's the last example, but using this expansion technique
    /// with named capture groups:
    ///
    /// ```rust
    /// # extern crate regex; use regex::bytes::Regex;
    /// # fn main() {
    /// let re = Regex::new(r"(?P<last>[^,\s]+),\s+(?P<first>\S+)").unwrap();
    /// let result = re.replace(b"Springsteen, Bruce", &b"$first $last"[..]);
    /// assert_eq!(result, &b"Bruce Springsteen"[..]);
    /// # }
    /// ```
    ///
    /// Note that using `$2` instead of `$first` or `$1` instead of `$last`
    /// would produce the same result. To write a literal `$` use `$$`.
    ///
    /// Sometimes the replacement string requires use of curly braces to
    /// delineate a capture group replacement and surrounding literal text.
    /// For example, if we wanted to join two words together with an
    /// underscore:
    ///
    /// ```rust
    /// # extern crate regex; use regex::bytes::Regex;
    /// # fn main() {
    /// let re = Regex::new(r"(?P<first>\w+)\s+(?P<second>\w+)").unwrap();
    /// let result = re.replace(b"deep fried", &b"${first}_$second"[..]);
    /// assert_eq!(result, &b"deep_fried"[..]);
    /// # }
    /// ```
    ///
    /// Without the curly braces, the capture group name `first_` would be
    /// used, and since it doesn't exist, it would be replaced with the empty
    /// string.
    ///
    /// Finally, sometimes you just want to replace a literal string with no
    /// regard for capturing group expansion. This can be done by wrapping a
    /// byte string with `NoExpand`:
    ///
    /// ```rust
    /// # extern crate regex; use regex::bytes::Regex;
    /// # fn main() {
    /// use regex::bytes::NoExpand;
    ///
    /// let re = Regex::new(r"(?P<last>[^,\s]+),\s+(\S+)").unwrap();
    /// let result = re.replace(b"Springsteen, Bruce", NoExpand(b"$2 $last"));
    /// assert_eq!(result, &b"$2 $last"[..]);
    /// # }
    /// ```
    pub fn replace<'t, R: Replacer>(
        &self,
        text: &'t [u8],
        rep: R,
    ) -> Cow<'t, [u8]> {
        self.replacen(text, 1, rep)
    }

    /// Replaces all non-overlapping matches in `text` with the replacement
    /// provided. This is the same as calling `replacen` with `limit` set to
    /// `0`.
    ///
    /// See the documentation for `replace` for details on how to access
    /// capturing group matches in the replacement text.
    pub fn replace_all<'t, R: Replacer>(
        &self,
        text: &'t [u8],
        rep: R,
    ) -> Cow<'t, [u8]> {
        self.replacen(text, 0, rep)
    }

    /// Replaces at most `limit` non-overlapping matches in `text` with the
    /// replacement provided. If `limit` is 0, then all non-overlapping matches
    /// are replaced.
    ///
    /// See the documentation for `replace` for details on how to access
    /// capturing group matches in the replacement text.
    pub fn replacen<'t, R: Replacer>(
        &self,
        text: &'t [u8],
        limit: usize,
        mut rep: R,
    ) -> Cow<'t, [u8]> {
        if let Some(rep) = rep.no_expansion() {
            let mut it = self.find_iter(text).enumerate().peekable();
            if it.peek().is_none() {
                return Cow::Borrowed(text);
            }
            let mut new = Vec::with_capacity(text.len());
            let mut last_match = 0;
            for (i, m) in it {
                if limit > 0 && i >= limit {
                    break
                }
                new.extend_from_slice(&text[last_match..m.start()]);
                new.extend_from_slice(&rep);
                last_match = m.end();
            }
            new.extend_from_slice(&text[last_match..]);
            return Cow::Owned(new);
        }

        // The slower path, which we use if the replacement needs access to
        // capture groups.
        let mut it = self.captures_iter(text).enumerate().peekable();
        if it.peek().is_none() {
            return Cow::Borrowed(text);
        }
        let mut new = Vec::with_capacity(text.len());
        let mut last_match = 0;
        for (i, cap) in it {
            if limit > 0 && i >= limit {
                break
            }
            // unwrap on 0 is OK because captures only reports matches
            let m = cap.get(0).unwrap();
            new.extend_from_slice(&text[last_match..m.start()]);
            rep.replace_append(&cap, &mut new);
            last_match = m.end();
        }
        new.extend_from_slice(&text[last_match..]);
        Cow::Owned(new)
    }
}

/// Advanced or "lower level" search methods.
impl Regex {
    /// Returns the end location of a match in the text given.
    ///
    /// This method may have the same performance characteristics as
    /// `is_match`, except it provides an end location for a match. In
    /// particular, the location returned *may be shorter* than the proper end
    /// of the leftmost-first match.
    ///
    /// # Example
    ///
    /// Typically, `a+` would match the entire first sequence of `a` in some
    /// text, but `shortest_match` can give up as soon as it sees the first
    /// `a`.
    ///
    /// ```rust
    /// # extern crate regex; use regex::bytes::Regex;
    /// # fn main() {
    /// let text = b"aaaaa";
    /// let pos = Regex::new(r"a+").unwrap().shortest_match(text);
    /// assert_eq!(pos, Some(1));
    /// # }
    /// ```
    pub fn shortest_match(&self, text: &[u8]) -> Option<usize> {
        self.shortest_match_at(text, 0)
    }

    /// Returns the same as shortest_match, but starts the search at the given
    /// offset.
    ///
    /// The significance of the starting point is that it takes the surrounding
    /// context into consideration. For example, the `\A` anchor can only
    /// match when `start == 0`.
    #[doc(hidden)]
    pub fn shortest_match_at(
        &self,
        text: &[u8],
        start: usize,
    ) -> Option<usize> {
        self.0.searcher().shortest_match_at(text, start)
    }

    /// Returns the same as is_match, but starts the search at the given
    /// offset.
    ///
    /// The significance of the starting point is that it takes the surrounding
    /// context into consideration. For example, the `\A` anchor can only
    /// match when `start == 0`.
    #[doc(hidden)]
    pub fn is_match_at(&self, text: &[u8], start: usize) -> bool {
        self.shortest_match_at(text, start).is_some()
    }

    /// Returns the same as find, but starts the search at the given
    /// offset.
    ///
    /// The significance of the starting point is that it takes the surrounding
    /// context into consideration. For example, the `\A` anchor can only
    /// match when `start == 0`.
    #[doc(hidden)]
    pub fn find_at<'t>(
        &self,
        text: &'t [u8],
        start: usize,
    ) -> Option<Match<'t>> {
        self.0.searcher().find_at(text, start)
            .map(|(s, e)| Match::new(text, s, e))
    }

    /// Returns the same as captures, but starts the search at the given
    /// offset and populates the capture locations given.
    ///
    /// The significance of the starting point is that it takes the surrounding
    /// context into consideration. For example, the `\A` anchor can only
    /// match when `start == 0`.
    #[doc(hidden)]
    pub fn read_captures_at<'t>(
        &self,
        locs: &mut Locations,
        text: &'t [u8],
        start: usize,
    ) -> Option<Match<'t>> {
        self.0.searcher().read_captures_at(locs, text, start)
            .map(|(s, e)| Match::new(text, s, e))
    }
}

/// Auxiliary methods.
impl Regex {
    /// Returns the original string of this regex.
    pub fn as_str(&self) -> &str {
        &self.0.regex_strings()[0]
    }

    /// Returns an iterator over the capture names.
    pub fn capture_names(&self) -> CaptureNames {
        CaptureNames(self.0.capture_names().iter())
    }

    /// Returns the number of captures.
    pub fn captures_len(&self) -> usize {
        self.0.capture_names().len()
    }

    /// Returns an empty set of locations that can be reused in multiple calls
    /// to `read_captures`.
    #[doc(hidden)]
    pub fn locations(&self) -> Locations {
        self.0.searcher().locations()
    }
}

/// An iterator over all non-overlapping matches for a particular string.
///
/// The iterator yields a tuple of integers corresponding to the start and end
/// of the match. The indices are byte offsets. The iterator stops when no more
/// matches can be found.
///
/// `'r` is the lifetime of the compiled regular expression and `'t` is the
/// lifetime of the matched byte string.
pub struct Matches<'r, 't>(re_trait::Matches<'t, ExecNoSync<'r>>);

impl<'r, 't> Iterator for Matches<'r, 't> {
    type Item = Match<'t>;

    fn next(&mut self) -> Option<Match<'t>> {
        let text = self.0.text();
        self.0.next().map(|(s, e)| Match::new(text, s, e))
    }
}

/// An iterator that yields all non-overlapping capture groups matching a
/// particular regular expression.
///
/// The iterator stops when no more matches can be found.
///
/// `'r` is the lifetime of the compiled regular expression and `'t` is the
/// lifetime of the matched byte string.
pub struct CaptureMatches<'r, 't>(re_trait::CaptureMatches<'t, ExecNoSync<'r>>);

impl<'r, 't> Iterator for CaptureMatches<'r, 't> {
    type Item = Captures<'t>;

    fn next(&mut self) -> Option<Captures<'t>> {
        self.0.next().map(|locs| Captures {
            text: self.0.text(),
            locs: locs,
            named_groups: self.0.regex().capture_name_idx().clone(),
        })
    }
}

/// Yields all substrings delimited by a regular expression match.
///
/// `'r` is the lifetime of the compiled regular expression and `'t` is the
/// lifetime of the byte string being split.
pub struct Split<'r, 't> {
    finder: Matches<'r, 't>,
    last: usize,
}

impl<'r, 't> Iterator for Split<'r, 't> {
    type Item = &'t [u8];

    fn next(&mut self) -> Option<&'t [u8]> {
        let text = self.finder.0.text();
        match self.finder.next() {
            None => {
                if self.last >= text.len() {
                    None
                } else {
                    let s = &text[self.last..];
                    self.last = text.len();
                    Some(s)
                }
            }
            Some(m) => {
                let matched = &text[self.last..m.start()];
                self.last = m.end();
                Some(matched)
            }
        }
    }
}

/// Yields at most `N` substrings delimited by a regular expression match.
///
/// The last substring will be whatever remains after splitting.
///
/// `'r` is the lifetime of the compiled regular expression and `'t` is the
/// lifetime of the byte string being split.
pub struct SplitN<'r, 't> {
    splits: Split<'r, 't>,
    n: usize,
}

impl<'r, 't> Iterator for SplitN<'r, 't> {
    type Item = &'t [u8];

    fn next(&mut self) -> Option<&'t [u8]> {
        if self.n == 0 {
            return None
        }
        self.n -= 1;
        if self.n == 0 {
            let text = self.splits.finder.0.text();
            Some(&text[self.splits.last..])
        } else {
            self.splits.next()
        }
    }
}

/// An iterator over the names of all possible captures.
///
/// `None` indicates an unnamed capture; the first element (capture 0, the
/// whole matched region) is always unnamed.
///
/// `'r` is the lifetime of the compiled regular expression.
pub struct CaptureNames<'r>(::std::slice::Iter<'r, Option<String>>);

impl<'r> Iterator for CaptureNames<'r> {
    type Item = Option<&'r str>;

    fn next(&mut self) -> Option<Option<&'r str>> {
        self.0.next().as_ref()
            .map(|slot| slot.as_ref().map(|name| name.as_ref()))
    }

    fn size_hint(&self) -> (usize, Option<usize>) {
        self.0.size_hint()
    }
}

/// Captures represents a group of captured byte strings for a single match.
///
/// The 0th capture always corresponds to the entire match. Each subsequent
/// index corresponds to the next capture group in the regex. If a capture
/// group is named, then the matched byte string is *also* available via the
/// `name` method. (Note that the 0th capture is always unnamed and so must be
/// accessed with the `get` method.)
///
/// Positions returned from a capture group are always byte indices.
///
/// `'t` is the lifetime of the matched text.
pub struct Captures<'t> {
    text: &'t [u8],
    locs: Locations,
    named_groups: Arc<HashMap<String, usize>>,
}

impl<'t> Captures<'t> {
    /// Returns the match associated with the capture group at index `i`. If
    /// `i` does not correspond to a capture group, or if the capture group
    /// did not participate in the match, then `None` is returned.
    ///
    /// # Examples
    ///
    /// Get the text of the match with a default of an empty string if this
    /// group didn't participate in the match:
    ///
    /// ```rust
    /// # use regex::bytes::Regex;
    /// let re = Regex::new(r"[a-z]+(?:([0-9]+)|([A-Z]+))").unwrap();
    /// let caps = re.captures(b"abc123").unwrap();
    ///
    /// let text1 = caps.get(1).map_or(&b""[..], |m| m.as_bytes());
    /// let text2 = caps.get(2).map_or(&b""[..], |m| m.as_bytes());
    /// assert_eq!(text1, &b"123"[..]);
    /// assert_eq!(text2, &b""[..]);
    /// ```
    pub fn get(&self, i: usize) -> Option<Match<'t>> {
        self.locs.pos(i).map(|(s, e)| Match::new(self.text, s, e))
    }

    /// Returns the match for the capture group named `name`. If `name` isn't a
    /// valid capture group or didn't match anything, then `None` is returned.
    pub fn name(&self, name: &str) -> Option<Match<'t>> {
        self.named_groups.get(name).and_then(|&i| self.get(i))
    }

    /// An iterator that yields all capturing matches in the order in which
    /// they appear in the regex. If a particular capture group didn't
    /// participate in the match, then `None` is yielded for that capture.
    ///
    /// The first match always corresponds to the overall match of the regex.
    pub fn iter<'c>(&'c self) -> SubCaptureMatches<'c, 't> {
        SubCaptureMatches {
            caps: self,
            it: self.locs.iter(),
        }
    }

    /// Expands all instances of `$name` in `replacement` to the corresponding
    /// capture group `name`, and writes them to the `dst` buffer given.
    ///
    /// `name` may be an integer corresponding to the index of the
    /// capture group (counted by order of opening parenthesis where `0` is the
    /// entire match) or it can be a name (consisting of letters, digits or
    /// underscores) corresponding to a named capture group.
    ///
    /// If `name` isn't a valid capture group (whether the name doesn't exist
    /// or isn't a valid index), then it is replaced with the empty string.
    ///
    /// The longest possible name is used. e.g., `$1a` looks up the capture
    /// group named `1a` and not the capture group at index `1`. To exert more
    /// precise control over the name, use braces, e.g., `${1}a`.
    ///
    /// To write a literal `$` use `$$`.
    pub fn expand(&self, replacement: &[u8], dst: &mut Vec<u8>) {
        expand_bytes(self, replacement, dst)
    }

    /// Returns the number of captured groups.
    ///
    /// This is always at least `1`, since every regex has at least one capture
    /// group that corresponds to the full match.
    #[inline]
    pub fn len(&self) -> usize {
        self.locs.len()
    }
}

impl<'t> fmt::Debug for Captures<'t> {
    fn fmt(&self, f: &mut fmt::Formatter) -> fmt::Result {
        f.debug_tuple("Captures").field(&CapturesDebug(self)).finish()
    }
}

struct CapturesDebug<'c, 't: 'c>(&'c Captures<'t>);

impl<'c, 't> fmt::Debug for CapturesDebug<'c, 't> {
    fn fmt(&self, f: &mut fmt::Formatter) -> fmt::Result {
        fn escape_bytes(bytes: &[u8]) -> String {
            let mut s = String::new();
            for &b in bytes {
                s.push_str(&escape_byte(b));
            }
            s
        }

        fn escape_byte(byte: u8) -> String {
            use std::ascii::escape_default;

            let escaped: Vec<u8> = escape_default(byte).collect();
            String::from_utf8_lossy(&escaped).into_owned()
        }

        // We'd like to show something nice here, even if it means an
        // allocation to build a reverse index.
        let slot_to_name: HashMap<&usize, &String> =
            self.0.named_groups.iter().map(|(a, b)| (b, a)).collect();
        let mut map = f.debug_map();
        for (slot, m) in self.0.locs.iter().enumerate() {
            let m = m.map(|(s, e)| escape_bytes(&self.0.text[s..e]));
            if let Some(name) = slot_to_name.get(&slot) {
                map.entry(&name, &m);
            } else {
                map.entry(&slot, &m);
            }
        }
        map.finish()
    }
}

/// Get a group by index.
///
/// `'t` is the lifetime of the matched text.
///
/// The text can't outlive the `Captures` object if this method is
/// used, because of how `Index` is defined (normally `a[i]` is part
/// of `a` and can't outlive it); to do that, use `get()` instead.
///
/// # Panics
///
/// If there is no group at the given index.
impl<'t> Index<usize> for Captures<'t> {
    type Output = [u8];

    fn index(&self, i: usize) -> &[u8] {
        self.get(i).map(|m| m.as_bytes())
            .unwrap_or_else(|| panic!("no group at index '{}'", i))
    }
}

/// Get a group by name.
///
/// `'t` is the lifetime of the matched text and `'i` is the lifetime
/// of the group name (the index).
///
/// The text can't outlive the `Captures` object if this method is
/// used, because of how `Index` is defined (normally `a[i]` is part
/// of `a` and can't outlive it); to do that, use `name` instead.
///
/// # Panics
///
/// If there is no group named by the given value.
impl<'t, 'i> Index<&'i str> for Captures<'t> {
    type Output = [u8];

    fn index<'a>(&'a self, name: &'i str) -> &'a [u8] {
        self.name(name).map(|m| m.as_bytes())
            .unwrap_or_else(|| panic!("no group named '{}'", name))
    }
}

/// An iterator that yields all capturing matches in the order in which they
/// appear in the regex.
///
/// If a particular capture group didn't participate in the match, then `None`
/// is yielded for that capture. The first match always corresponds to the
/// overall match of the regex.
///
/// The lifetime `'c` corresponds to the lifetime of the `Captures` value, and
/// the lifetime `'t` corresponds to the originally matched text.
pub struct SubCaptureMatches<'c, 't: 'c> {
    caps: &'c Captures<'t>,
    it: SubCapturesPosIter<'c>,
}

impl<'c, 't> Iterator for SubCaptureMatches<'c, 't> {
    type Item = Option<Match<'t>>;

    fn next(&mut self) -> Option<Option<Match<'t>>> {
        self.it.next()
            .map(|cap| cap.map(|(s, e)| Match::new(self.caps.text, s, e)))
    }
}

/// Replacer describes types that can be used to replace matches in a byte
/// string.
///
/// In general, users of this crate shouldn't need to implement this trait,
/// since implementations are already provided for `&[u8]` and
/// `FnMut(&Captures) -> Vec<u8>`, which covers most use cases.
pub trait Replacer {
    /// Appends text to `dst` to replace the current match.
    ///
    /// The current match is represented by `caps`, which is guaranteed to
    /// have a match at capture group `0`.
    ///
    /// For example, a no-op replacement would be
    /// `dst.extend(&caps[0])`.
    fn replace_append(&mut self, caps: &Captures, dst: &mut Vec<u8>);

    /// Return a fixed unchanging replacement byte string.
    ///
    /// When doing replacements, if access to `Captures` is not needed (e.g.,
    /// the replacement byte string does not need `$` expansion), then it can
    /// be beneficial to avoid finding sub-captures.
    ///
    /// In general, this is called once for every call to `replacen`.
    fn no_expansion<'r>(&'r mut self) -> Option<Cow<'r, [u8]>> {
        None
    }

    /// Return a `Replacer` that borrows and wraps this `Replacer`.
    ///
    /// This is useful when you want to take a generic `Replacer` (which might
    /// not be cloneable) and use it without consuming it, so it can be used
    /// more than once.
    ///
    /// # Example
    ///
    /// ```
    /// use regex::bytes::{Regex, Replacer};
    ///
    /// fn replace_all_twice<R: Replacer>(
    ///     re: Regex,
    ///     src: &[u8],
    ///     mut rep: R,
    /// ) -> Vec<u8> {
    ///     let dst = re.replace_all(src, rep.by_ref());
    ///     let dst = re.replace_all(&dst, rep.by_ref());
    ///     dst.into_owned()
    /// }
    /// ```
    fn by_ref<'r>(&'r mut self) -> ReplacerRef<'r, Self> {
        ReplacerRef(self)
    }
}

/// By-reference adaptor for a `Replacer`
///
/// Returned by [`Replacer::by_ref`](trait.Replacer.html#method.by_ref).
#[derive(Debug)]
pub struct ReplacerRef<'a, R: ?Sized + 'a>(&'a mut R);

impl<'a, R: Replacer + ?Sized + 'a> Replacer for ReplacerRef<'a, R> {
    fn replace_append(&mut self, caps: &Captures, dst: &mut Vec<u8>) {
        self.0.replace_append(caps, dst)
    }
    fn no_expansion<'r>(&'r mut self) -> Option<Cow<'r, [u8]>> {
        self.0.no_expansion()
    }
}

impl<'a> Replacer for &'a [u8] {
    fn replace_append(&mut self, caps: &Captures, dst: &mut Vec<u8>) {
        caps.expand(*self, dst);
    }

    fn no_expansion(&mut self) -> Option<Cow<[u8]>> {
        match memchr(b'$', *self) {
            Some(_) => None,
            None => Some(Cow::Borrowed(*self)),
        }
    }
}

impl<F> Replacer for F where F: FnMut(&Captures) -> Vec<u8> {
    fn replace_append(&mut self, caps: &Captures, dst: &mut Vec<u8>) {
        dst.extend_from_slice(&(*self)(caps));
    }
}

/// `NoExpand` indicates literal byte string replacement.
///
/// It can be used with `replace` and `replace_all` to do a literal byte string
/// replacement without expanding `$name` to their corresponding capture
/// groups. This can be both convenient (to avoid escaping `$`, for example)
/// and performant (since capture groups don't need to be found).
///
/// `'t` is the lifetime of the literal text.
pub struct NoExpand<'t>(pub &'t [u8]);

impl<'t> Replacer for NoExpand<'t> {
    fn replace_append(&mut self, _: &Captures, dst: &mut Vec<u8>) {
        dst.extend_from_slice(self.0);
    }

    fn no_expansion(&mut self) -> Option<Cow<[u8]>> {
        Some(Cow::Borrowed(self.0))
    }
}