summaryrefslogtreecommitdiffstats
path: root/vendor/globset/src/glob.rs
blob: 6e35aeec3913a1c19f003860839985d90cb079db (plain)
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
1073
1074
1075
1076
1077
1078
1079
1080
1081
1082
1083
1084
1085
1086
1087
1088
1089
1090
1091
1092
1093
1094
1095
1096
1097
1098
1099
1100
1101
1102
1103
1104
1105
1106
1107
1108
1109
1110
1111
1112
1113
1114
1115
1116
1117
1118
1119
1120
1121
1122
1123
1124
1125
1126
1127
1128
1129
1130
1131
1132
1133
1134
1135
1136
1137
1138
1139
1140
1141
1142
1143
1144
1145
1146
1147
1148
1149
1150
1151
1152
1153
1154
1155
1156
1157
1158
1159
1160
1161
1162
1163
1164
1165
1166
1167
1168
1169
1170
1171
1172
1173
1174
1175
1176
1177
1178
1179
1180
1181
1182
1183
1184
1185
1186
1187
1188
1189
1190
1191
1192
1193
1194
1195
1196
1197
1198
1199
1200
1201
1202
1203
1204
1205
1206
1207
1208
1209
1210
1211
1212
1213
1214
1215
1216
1217
1218
1219
1220
1221
1222
1223
1224
1225
1226
1227
1228
1229
1230
1231
1232
1233
1234
1235
1236
1237
1238
1239
1240
1241
1242
1243
1244
1245
1246
1247
1248
1249
1250
1251
1252
1253
1254
1255
1256
1257
1258
1259
1260
1261
1262
1263
1264
1265
1266
1267
1268
1269
1270
1271
1272
1273
1274
1275
1276
1277
1278
1279
1280
1281
1282
1283
1284
1285
1286
1287
1288
1289
1290
1291
1292
1293
1294
1295
1296
1297
1298
1299
1300
1301
1302
1303
1304
1305
1306
1307
1308
1309
1310
1311
1312
1313
1314
1315
1316
1317
1318
1319
1320
1321
1322
1323
1324
1325
1326
1327
1328
1329
1330
1331
1332
1333
1334
1335
1336
1337
1338
1339
1340
1341
1342
1343
1344
1345
1346
1347
1348
1349
1350
1351
1352
1353
1354
1355
1356
1357
1358
1359
1360
1361
1362
1363
1364
1365
1366
1367
1368
1369
1370
1371
1372
1373
1374
1375
1376
1377
1378
1379
1380
1381
1382
1383
1384
1385
1386
1387
1388
1389
1390
1391
1392
1393
1394
1395
1396
1397
1398
1399
1400
1401
1402
1403
1404
1405
1406
1407
1408
1409
1410
1411
1412
1413
1414
1415
1416
1417
1418
1419
1420
1421
1422
1423
1424
1425
1426
1427
1428
1429
1430
1431
1432
1433
1434
1435
1436
1437
1438
1439
1440
1441
1442
1443
1444
1445
1446
1447
1448
1449
1450
1451
1452
1453
1454
1455
1456
1457
1458
1459
1460
1461
1462
1463
1464
1465
1466
1467
1468
1469
1470
1471
1472
1473
1474
1475
1476
1477
1478
1479
1480
1481
1482
1483
1484
1485
1486
1487
1488
1489
1490
1491
1492
1493
1494
1495
1496
1497
1498
1499
1500
1501
1502
1503
1504
1505
1506
1507
1508
1509
1510
1511
1512
1513
1514
1515
1516
1517
1518
1519
1520
1521
1522
1523
1524
1525
1526
1527
1528
use std::fmt;
use std::hash;
use std::iter;
use std::ops::{Deref, DerefMut};
use std::path::{is_separator, Path};
use std::str;

use regex;
use regex::bytes::Regex;

use crate::{new_regex, Candidate, Error, ErrorKind};

/// Describes a matching strategy for a particular pattern.
///
/// This provides a way to more quickly determine whether a pattern matches
/// a particular file path in a way that scales with a large number of
/// patterns. For example, if many patterns are of the form `*.ext`, then it's
/// possible to test whether any of those patterns matches by looking up a
/// file path's extension in a hash table.
#[derive(Clone, Debug, Eq, PartialEq)]
pub enum MatchStrategy {
    /// A pattern matches if and only if the entire file path matches this
    /// literal string.
    Literal(String),
    /// A pattern matches if and only if the file path's basename matches this
    /// literal string.
    BasenameLiteral(String),
    /// A pattern matches if and only if the file path's extension matches this
    /// literal string.
    Extension(String),
    /// A pattern matches if and only if this prefix literal is a prefix of the
    /// candidate file path.
    Prefix(String),
    /// A pattern matches if and only if this prefix literal is a prefix of the
    /// candidate file path.
    ///
    /// An exception: if `component` is true, then `suffix` must appear at the
    /// beginning of a file path or immediately following a `/`.
    Suffix {
        /// The actual suffix.
        suffix: String,
        /// Whether this must start at the beginning of a path component.
        component: bool,
    },
    /// A pattern matches only if the given extension matches the file path's
    /// extension. Note that this is a necessary but NOT sufficient criterion.
    /// Namely, if the extension matches, then a full regex search is still
    /// required.
    RequiredExtension(String),
    /// A regex needs to be used for matching.
    Regex,
}

impl MatchStrategy {
    /// Returns a matching strategy for the given pattern.
    pub fn new(pat: &Glob) -> MatchStrategy {
        if let Some(lit) = pat.basename_literal() {
            MatchStrategy::BasenameLiteral(lit)
        } else if let Some(lit) = pat.literal() {
            MatchStrategy::Literal(lit)
        } else if let Some(ext) = pat.ext() {
            MatchStrategy::Extension(ext)
        } else if let Some(prefix) = pat.prefix() {
            MatchStrategy::Prefix(prefix)
        } else if let Some((suffix, component)) = pat.suffix() {
            MatchStrategy::Suffix { suffix: suffix, component: component }
        } else if let Some(ext) = pat.required_ext() {
            MatchStrategy::RequiredExtension(ext)
        } else {
            MatchStrategy::Regex
        }
    }
}

/// Glob represents a successfully parsed shell glob pattern.
///
/// It cannot be used directly to match file paths, but it can be converted
/// to a regular expression string or a matcher.
#[derive(Clone, Debug, Eq)]
pub struct Glob {
    glob: String,
    re: String,
    opts: GlobOptions,
    tokens: Tokens,
}

impl PartialEq for Glob {
    fn eq(&self, other: &Glob) -> bool {
        self.glob == other.glob && self.opts == other.opts
    }
}

impl hash::Hash for Glob {
    fn hash<H: hash::Hasher>(&self, state: &mut H) {
        self.glob.hash(state);
        self.opts.hash(state);
    }
}

impl fmt::Display for Glob {
    fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
        self.glob.fmt(f)
    }
}

impl str::FromStr for Glob {
    type Err = Error;

    fn from_str(glob: &str) -> Result<Self, Self::Err> {
        Self::new(glob)
    }
}

/// A matcher for a single pattern.
#[derive(Clone, Debug)]
pub struct GlobMatcher {
    /// The underlying pattern.
    pat: Glob,
    /// The pattern, as a compiled regex.
    re: Regex,
}

impl GlobMatcher {
    /// Tests whether the given path matches this pattern or not.
    pub fn is_match<P: AsRef<Path>>(&self, path: P) -> bool {
        self.is_match_candidate(&Candidate::new(path.as_ref()))
    }

    /// Tests whether the given path matches this pattern or not.
    pub fn is_match_candidate(&self, path: &Candidate<'_>) -> bool {
        self.re.is_match(&path.path)
    }

    /// Returns the `Glob` used to compile this matcher.
    pub fn glob(&self) -> &Glob {
        &self.pat
    }
}

/// A strategic matcher for a single pattern.
#[cfg(test)]
#[derive(Clone, Debug)]
struct GlobStrategic {
    /// The match strategy to use.
    strategy: MatchStrategy,
    /// The pattern, as a compiled regex.
    re: Regex,
}

#[cfg(test)]
impl GlobStrategic {
    /// Tests whether the given path matches this pattern or not.
    fn is_match<P: AsRef<Path>>(&self, path: P) -> bool {
        self.is_match_candidate(&Candidate::new(path.as_ref()))
    }

    /// Tests whether the given path matches this pattern or not.
    fn is_match_candidate(&self, candidate: &Candidate<'_>) -> bool {
        let byte_path = &*candidate.path;

        match self.strategy {
            MatchStrategy::Literal(ref lit) => lit.as_bytes() == byte_path,
            MatchStrategy::BasenameLiteral(ref lit) => {
                lit.as_bytes() == &*candidate.basename
            }
            MatchStrategy::Extension(ref ext) => {
                ext.as_bytes() == &*candidate.ext
            }
            MatchStrategy::Prefix(ref pre) => {
                starts_with(pre.as_bytes(), byte_path)
            }
            MatchStrategy::Suffix { ref suffix, component } => {
                if component && byte_path == &suffix.as_bytes()[1..] {
                    return true;
                }
                ends_with(suffix.as_bytes(), byte_path)
            }
            MatchStrategy::RequiredExtension(ref ext) => {
                let ext = ext.as_bytes();
                &*candidate.ext == ext && self.re.is_match(byte_path)
            }
            MatchStrategy::Regex => self.re.is_match(byte_path),
        }
    }
}

/// A builder for a pattern.
///
/// This builder enables configuring the match semantics of a pattern. For
/// example, one can make matching case insensitive.
///
/// The lifetime `'a` refers to the lifetime of the pattern string.
#[derive(Clone, Debug)]
pub struct GlobBuilder<'a> {
    /// The glob pattern to compile.
    glob: &'a str,
    /// Options for the pattern.
    opts: GlobOptions,
}

#[derive(Clone, Copy, Debug, Eq, Hash, PartialEq)]
struct GlobOptions {
    /// Whether to match case insensitively.
    case_insensitive: bool,
    /// Whether to require a literal separator to match a separator in a file
    /// path. e.g., when enabled, `*` won't match `/`.
    literal_separator: bool,
    /// Whether or not to use `\` to escape special characters.
    /// e.g., when enabled, `\*` will match a literal `*`.
    backslash_escape: bool,
}

impl GlobOptions {
    fn default() -> GlobOptions {
        GlobOptions {
            case_insensitive: false,
            literal_separator: false,
            backslash_escape: !is_separator('\\'),
        }
    }
}

#[derive(Clone, Debug, Default, Eq, PartialEq)]
struct Tokens(Vec<Token>);

impl Deref for Tokens {
    type Target = Vec<Token>;
    fn deref(&self) -> &Vec<Token> {
        &self.0
    }
}

impl DerefMut for Tokens {
    fn deref_mut(&mut self) -> &mut Vec<Token> {
        &mut self.0
    }
}

#[derive(Clone, Debug, Eq, PartialEq)]
enum Token {
    Literal(char),
    Any,
    ZeroOrMore,
    RecursivePrefix,
    RecursiveSuffix,
    RecursiveZeroOrMore,
    Class { negated: bool, ranges: Vec<(char, char)> },
    Alternates(Vec<Tokens>),
}

impl Glob {
    /// Builds a new pattern with default options.
    pub fn new(glob: &str) -> Result<Glob, Error> {
        GlobBuilder::new(glob).build()
    }

    /// Returns a matcher for this pattern.
    pub fn compile_matcher(&self) -> GlobMatcher {
        let re =
            new_regex(&self.re).expect("regex compilation shouldn't fail");
        GlobMatcher { pat: self.clone(), re: re }
    }

    /// Returns a strategic matcher.
    ///
    /// This isn't exposed because it's not clear whether it's actually
    /// faster than just running a regex for a *single* pattern. If it
    /// is faster, then GlobMatcher should do it automatically.
    #[cfg(test)]
    fn compile_strategic_matcher(&self) -> GlobStrategic {
        let strategy = MatchStrategy::new(self);
        let re =
            new_regex(&self.re).expect("regex compilation shouldn't fail");
        GlobStrategic { strategy: strategy, re: re }
    }

    /// Returns the original glob pattern used to build this pattern.
    pub fn glob(&self) -> &str {
        &self.glob
    }

    /// Returns the regular expression string for this glob.
    ///
    /// Note that regular expressions for globs are intended to be matched on
    /// arbitrary bytes (`&[u8]`) instead of Unicode strings (`&str`). In
    /// particular, globs are frequently used on file paths, where there is no
    /// general guarantee that file paths are themselves valid UTF-8. As a
    /// result, callers will need to ensure that they are using a regex API
    /// that can match on arbitrary bytes. For example, the
    /// [`regex`](https://crates.io/regex)
    /// crate's
    /// [`Regex`](https://docs.rs/regex/*/regex/struct.Regex.html)
    /// API is not suitable for this since it matches on `&str`, but its
    /// [`bytes::Regex`](https://docs.rs/regex/*/regex/bytes/struct.Regex.html)
    /// API is suitable for this.
    pub fn regex(&self) -> &str {
        &self.re
    }

    /// Returns the pattern as a literal if and only if the pattern must match
    /// an entire path exactly.
    ///
    /// The basic format of these patterns is `{literal}`.
    fn literal(&self) -> Option<String> {
        if self.opts.case_insensitive {
            return None;
        }
        let mut lit = String::new();
        for t in &*self.tokens {
            match *t {
                Token::Literal(c) => lit.push(c),
                _ => return None,
            }
        }
        if lit.is_empty() {
            None
        } else {
            Some(lit)
        }
    }

    /// Returns an extension if this pattern matches a file path if and only
    /// if the file path has the extension returned.
    ///
    /// Note that this extension returned differs from the extension that
    /// std::path::Path::extension returns. Namely, this extension includes
    /// the '.'. Also, paths like `.rs` are considered to have an extension
    /// of `.rs`.
    fn ext(&self) -> Option<String> {
        if self.opts.case_insensitive {
            return None;
        }
        let start = match self.tokens.get(0) {
            Some(&Token::RecursivePrefix) => 1,
            Some(_) => 0,
            _ => return None,
        };
        match self.tokens.get(start) {
            Some(&Token::ZeroOrMore) => {
                // If there was no recursive prefix, then we only permit
                // `*` if `*` can match a `/`. For example, if `*` can't
                // match `/`, then `*.c` doesn't match `foo/bar.c`.
                if start == 0 && self.opts.literal_separator {
                    return None;
                }
            }
            _ => return None,
        }
        match self.tokens.get(start + 1) {
            Some(&Token::Literal('.')) => {}
            _ => return None,
        }
        let mut lit = ".".to_string();
        for t in self.tokens[start + 2..].iter() {
            match *t {
                Token::Literal('.') | Token::Literal('/') => return None,
                Token::Literal(c) => lit.push(c),
                _ => return None,
            }
        }
        if lit.is_empty() {
            None
        } else {
            Some(lit)
        }
    }

    /// This is like `ext`, but returns an extension even if it isn't sufficient
    /// to imply a match. Namely, if an extension is returned, then it is
    /// necessary but not sufficient for a match.
    fn required_ext(&self) -> Option<String> {
        if self.opts.case_insensitive {
            return None;
        }
        // We don't care at all about the beginning of this pattern. All we
        // need to check for is if it ends with a literal of the form `.ext`.
        let mut ext: Vec<char> = vec![]; // built in reverse
        for t in self.tokens.iter().rev() {
            match *t {
                Token::Literal('/') => return None,
                Token::Literal(c) => {
                    ext.push(c);
                    if c == '.' {
                        break;
                    }
                }
                _ => return None,
            }
        }
        if ext.last() != Some(&'.') {
            None
        } else {
            ext.reverse();
            Some(ext.into_iter().collect())
        }
    }

    /// Returns a literal prefix of this pattern if the entire pattern matches
    /// if the literal prefix matches.
    fn prefix(&self) -> Option<String> {
        if self.opts.case_insensitive {
            return None;
        }
        let (end, need_sep) = match self.tokens.last() {
            Some(&Token::ZeroOrMore) => {
                if self.opts.literal_separator {
                    // If a trailing `*` can't match a `/`, then we can't
                    // assume a match of the prefix corresponds to a match
                    // of the overall pattern. e.g., `foo/*` with
                    // `literal_separator` enabled matches `foo/bar` but not
                    // `foo/bar/baz`, even though `foo/bar/baz` has a `foo/`
                    // literal prefix.
                    return None;
                }
                (self.tokens.len() - 1, false)
            }
            Some(&Token::RecursiveSuffix) => (self.tokens.len() - 1, true),
            _ => (self.tokens.len(), false),
        };
        let mut lit = String::new();
        for t in &self.tokens[0..end] {
            match *t {
                Token::Literal(c) => lit.push(c),
                _ => return None,
            }
        }
        if need_sep {
            lit.push('/');
        }
        if lit.is_empty() {
            None
        } else {
            Some(lit)
        }
    }

    /// Returns a literal suffix of this pattern if the entire pattern matches
    /// if the literal suffix matches.
    ///
    /// If a literal suffix is returned and it must match either the entire
    /// file path or be preceded by a `/`, then also return true. This happens
    /// with a pattern like `**/foo/bar`. Namely, this pattern matches
    /// `foo/bar` and `baz/foo/bar`, but not `foofoo/bar`. In this case, the
    /// suffix returned is `/foo/bar` (but should match the entire path
    /// `foo/bar`).
    ///
    /// When this returns true, the suffix literal is guaranteed to start with
    /// a `/`.
    fn suffix(&self) -> Option<(String, bool)> {
        if self.opts.case_insensitive {
            return None;
        }
        let mut lit = String::new();
        let (start, entire) = match self.tokens.get(0) {
            Some(&Token::RecursivePrefix) => {
                // We only care if this follows a path component if the next
                // token is a literal.
                if let Some(&Token::Literal(_)) = self.tokens.get(1) {
                    lit.push('/');
                    (1, true)
                } else {
                    (1, false)
                }
            }
            _ => (0, false),
        };
        let start = match self.tokens.get(start) {
            Some(&Token::ZeroOrMore) => {
                // If literal_separator is enabled, then a `*` can't
                // necessarily match everything, so reporting a suffix match
                // as a match of the pattern would be a false positive.
                if self.opts.literal_separator {
                    return None;
                }
                start + 1
            }
            _ => start,
        };
        for t in &self.tokens[start..] {
            match *t {
                Token::Literal(c) => lit.push(c),
                _ => return None,
            }
        }
        if lit.is_empty() || lit == "/" {
            None
        } else {
            Some((lit, entire))
        }
    }

    /// If this pattern only needs to inspect the basename of a file path,
    /// then the tokens corresponding to only the basename match are returned.
    ///
    /// For example, given a pattern of `**/*.foo`, only the tokens
    /// corresponding to `*.foo` are returned.
    ///
    /// Note that this will return None if any match of the basename tokens
    /// doesn't correspond to a match of the entire pattern. For example, the
    /// glob `foo` only matches when a file path has a basename of `foo`, but
    /// doesn't *always* match when a file path has a basename of `foo`. e.g.,
    /// `foo` doesn't match `abc/foo`.
    fn basename_tokens(&self) -> Option<&[Token]> {
        if self.opts.case_insensitive {
            return None;
        }
        let start = match self.tokens.get(0) {
            Some(&Token::RecursivePrefix) => 1,
            _ => {
                // With nothing to gobble up the parent portion of a path,
                // we can't assume that matching on only the basename is
                // correct.
                return None;
            }
        };
        if self.tokens[start..].is_empty() {
            return None;
        }
        for t in &self.tokens[start..] {
            match *t {
                Token::Literal('/') => return None,
                Token::Literal(_) => {} // OK
                Token::Any | Token::ZeroOrMore => {
                    if !self.opts.literal_separator {
                        // In this case, `*` and `?` can match a path
                        // separator, which means this could reach outside
                        // the basename.
                        return None;
                    }
                }
                Token::RecursivePrefix
                | Token::RecursiveSuffix
                | Token::RecursiveZeroOrMore => {
                    return None;
                }
                Token::Class { .. } | Token::Alternates(..) => {
                    // We *could* be a little smarter here, but either one
                    // of these is going to prevent our literal optimizations
                    // anyway, so give up.
                    return None;
                }
            }
        }
        Some(&self.tokens[start..])
    }

    /// Returns the pattern as a literal if and only if the pattern exclusively
    /// matches the basename of a file path *and* is a literal.
    ///
    /// The basic format of these patterns is `**/{literal}`, where `{literal}`
    /// does not contain a path separator.
    fn basename_literal(&self) -> Option<String> {
        let tokens = match self.basename_tokens() {
            None => return None,
            Some(tokens) => tokens,
        };
        let mut lit = String::new();
        for t in tokens {
            match *t {
                Token::Literal(c) => lit.push(c),
                _ => return None,
            }
        }
        Some(lit)
    }
}

impl<'a> GlobBuilder<'a> {
    /// Create a new builder for the pattern given.
    ///
    /// The pattern is not compiled until `build` is called.
    pub fn new(glob: &'a str) -> GlobBuilder<'a> {
        GlobBuilder { glob: glob, opts: GlobOptions::default() }
    }

    /// Parses and builds the pattern.
    pub fn build(&self) -> Result<Glob, Error> {
        let mut p = Parser {
            glob: &self.glob,
            stack: vec![Tokens::default()],
            chars: self.glob.chars().peekable(),
            prev: None,
            cur: None,
            opts: &self.opts,
        };
        p.parse()?;
        if p.stack.is_empty() {
            Err(Error {
                glob: Some(self.glob.to_string()),
                kind: ErrorKind::UnopenedAlternates,
            })
        } else if p.stack.len() > 1 {
            Err(Error {
                glob: Some(self.glob.to_string()),
                kind: ErrorKind::UnclosedAlternates,
            })
        } else {
            let tokens = p.stack.pop().unwrap();
            Ok(Glob {
                glob: self.glob.to_string(),
                re: tokens.to_regex_with(&self.opts),
                opts: self.opts,
                tokens: tokens,
            })
        }
    }

    /// Toggle whether the pattern matches case insensitively or not.
    ///
    /// This is disabled by default.
    pub fn case_insensitive(&mut self, yes: bool) -> &mut GlobBuilder<'a> {
        self.opts.case_insensitive = yes;
        self
    }

    /// Toggle whether a literal `/` is required to match a path separator.
    ///
    /// By default this is false: `*` and `?` will match `/`.
    pub fn literal_separator(&mut self, yes: bool) -> &mut GlobBuilder<'a> {
        self.opts.literal_separator = yes;
        self
    }

    /// When enabled, a back slash (`\`) may be used to escape
    /// special characters in a glob pattern. Additionally, this will
    /// prevent `\` from being interpreted as a path separator on all
    /// platforms.
    ///
    /// This is enabled by default on platforms where `\` is not a
    /// path separator and disabled by default on platforms where `\`
    /// is a path separator.
    pub fn backslash_escape(&mut self, yes: bool) -> &mut GlobBuilder<'a> {
        self.opts.backslash_escape = yes;
        self
    }
}

impl Tokens {
    /// Convert this pattern to a string that is guaranteed to be a valid
    /// regular expression and will represent the matching semantics of this
    /// glob pattern and the options given.
    fn to_regex_with(&self, options: &GlobOptions) -> String {
        let mut re = String::new();
        re.push_str("(?-u)");
        if options.case_insensitive {
            re.push_str("(?i)");
        }
        re.push('^');
        // Special case. If the entire glob is just `**`, then it should match
        // everything.
        if self.len() == 1 && self[0] == Token::RecursivePrefix {
            re.push_str(".*");
            re.push('$');
            return re;
        }
        self.tokens_to_regex(options, &self, &mut re);
        re.push('$');
        re
    }

    fn tokens_to_regex(
        &self,
        options: &GlobOptions,
        tokens: &[Token],
        re: &mut String,
    ) {
        for tok in tokens {
            match *tok {
                Token::Literal(c) => {
                    re.push_str(&char_to_escaped_literal(c));
                }
                Token::Any => {
                    if options.literal_separator {
                        re.push_str("[^/]");
                    } else {
                        re.push_str(".");
                    }
                }
                Token::ZeroOrMore => {
                    if options.literal_separator {
                        re.push_str("[^/]*");
                    } else {
                        re.push_str(".*");
                    }
                }
                Token::RecursivePrefix => {
                    re.push_str("(?:/?|.*/)");
                }
                Token::RecursiveSuffix => {
                    re.push_str("/.*");
                }
                Token::RecursiveZeroOrMore => {
                    re.push_str("(?:/|/.*/)");
                }
                Token::Class { negated, ref ranges } => {
                    re.push('[');
                    if negated {
                        re.push('^');
                    }
                    for r in ranges {
                        if r.0 == r.1 {
                            // Not strictly necessary, but nicer to look at.
                            re.push_str(&char_to_escaped_literal(r.0));
                        } else {
                            re.push_str(&char_to_escaped_literal(r.0));
                            re.push('-');
                            re.push_str(&char_to_escaped_literal(r.1));
                        }
                    }
                    re.push(']');
                }
                Token::Alternates(ref patterns) => {
                    let mut parts = vec![];
                    for pat in patterns {
                        let mut altre = String::new();
                        self.tokens_to_regex(options, &pat, &mut altre);
                        if !altre.is_empty() {
                            parts.push(altre);
                        }
                    }

                    // It is possible to have an empty set in which case the
                    // resulting alternation '()' would be an error.
                    if !parts.is_empty() {
                        re.push('(');
                        re.push_str(&parts.join("|"));
                        re.push(')');
                    }
                }
            }
        }
    }
}

/// Convert a Unicode scalar value to an escaped string suitable for use as
/// a literal in a non-Unicode regex.
fn char_to_escaped_literal(c: char) -> String {
    bytes_to_escaped_literal(&c.to_string().into_bytes())
}

/// Converts an arbitrary sequence of bytes to a UTF-8 string. All non-ASCII
/// code units are converted to their escaped form.
fn bytes_to_escaped_literal(bs: &[u8]) -> String {
    let mut s = String::with_capacity(bs.len());
    for &b in bs {
        if b <= 0x7F {
            s.push_str(&regex::escape(&(b as char).to_string()));
        } else {
            s.push_str(&format!("\\x{:02x}", b));
        }
    }
    s
}

struct Parser<'a> {
    glob: &'a str,
    stack: Vec<Tokens>,
    chars: iter::Peekable<str::Chars<'a>>,
    prev: Option<char>,
    cur: Option<char>,
    opts: &'a GlobOptions,
}

impl<'a> Parser<'a> {
    fn error(&self, kind: ErrorKind) -> Error {
        Error { glob: Some(self.glob.to_string()), kind: kind }
    }

    fn parse(&mut self) -> Result<(), Error> {
        while let Some(c) = self.bump() {
            match c {
                '?' => self.push_token(Token::Any)?,
                '*' => self.parse_star()?,
                '[' => self.parse_class()?,
                '{' => self.push_alternate()?,
                '}' => self.pop_alternate()?,
                ',' => self.parse_comma()?,
                '\\' => self.parse_backslash()?,
                c => self.push_token(Token::Literal(c))?,
            }
        }
        Ok(())
    }

    fn push_alternate(&mut self) -> Result<(), Error> {
        if self.stack.len() > 1 {
            return Err(self.error(ErrorKind::NestedAlternates));
        }
        Ok(self.stack.push(Tokens::default()))
    }

    fn pop_alternate(&mut self) -> Result<(), Error> {
        let mut alts = vec![];
        while self.stack.len() >= 2 {
            alts.push(self.stack.pop().unwrap());
        }
        self.push_token(Token::Alternates(alts))
    }

    fn push_token(&mut self, tok: Token) -> Result<(), Error> {
        if let Some(ref mut pat) = self.stack.last_mut() {
            return Ok(pat.push(tok));
        }
        Err(self.error(ErrorKind::UnopenedAlternates))
    }

    fn pop_token(&mut self) -> Result<Token, Error> {
        if let Some(ref mut pat) = self.stack.last_mut() {
            return Ok(pat.pop().unwrap());
        }
        Err(self.error(ErrorKind::UnopenedAlternates))
    }

    fn have_tokens(&self) -> Result<bool, Error> {
        match self.stack.last() {
            None => Err(self.error(ErrorKind::UnopenedAlternates)),
            Some(ref pat) => Ok(!pat.is_empty()),
        }
    }

    fn parse_comma(&mut self) -> Result<(), Error> {
        // If we aren't inside a group alternation, then don't
        // treat commas specially. Otherwise, we need to start
        // a new alternate.
        if self.stack.len() <= 1 {
            self.push_token(Token::Literal(','))
        } else {
            Ok(self.stack.push(Tokens::default()))
        }
    }

    fn parse_backslash(&mut self) -> Result<(), Error> {
        if self.opts.backslash_escape {
            match self.bump() {
                None => Err(self.error(ErrorKind::DanglingEscape)),
                Some(c) => self.push_token(Token::Literal(c)),
            }
        } else if is_separator('\\') {
            // Normalize all patterns to use / as a separator.
            self.push_token(Token::Literal('/'))
        } else {
            self.push_token(Token::Literal('\\'))
        }
    }

    fn parse_star(&mut self) -> Result<(), Error> {
        let prev = self.prev;
        if self.peek() != Some('*') {
            self.push_token(Token::ZeroOrMore)?;
            return Ok(());
        }
        assert!(self.bump() == Some('*'));
        if !self.have_tokens()? {
            if !self.peek().map_or(true, is_separator) {
                self.push_token(Token::ZeroOrMore)?;
                self.push_token(Token::ZeroOrMore)?;
            } else {
                self.push_token(Token::RecursivePrefix)?;
                assert!(self.bump().map_or(true, is_separator));
            }
            return Ok(());
        }

        if !prev.map(is_separator).unwrap_or(false) {
            if self.stack.len() <= 1
                || (prev != Some(',') && prev != Some('{'))
            {
                self.push_token(Token::ZeroOrMore)?;
                self.push_token(Token::ZeroOrMore)?;
                return Ok(());
            }
        }
        let is_suffix = match self.peek() {
            None => {
                assert!(self.bump().is_none());
                true
            }
            Some(',') | Some('}') if self.stack.len() >= 2 => true,
            Some(c) if is_separator(c) => {
                assert!(self.bump().map(is_separator).unwrap_or(false));
                false
            }
            _ => {
                self.push_token(Token::ZeroOrMore)?;
                self.push_token(Token::ZeroOrMore)?;
                return Ok(());
            }
        };
        match self.pop_token()? {
            Token::RecursivePrefix => {
                self.push_token(Token::RecursivePrefix)?;
            }
            Token::RecursiveSuffix => {
                self.push_token(Token::RecursiveSuffix)?;
            }
            _ => {
                if is_suffix {
                    self.push_token(Token::RecursiveSuffix)?;
                } else {
                    self.push_token(Token::RecursiveZeroOrMore)?;
                }
            }
        }
        Ok(())
    }

    fn parse_class(&mut self) -> Result<(), Error> {
        fn add_to_last_range(
            glob: &str,
            r: &mut (char, char),
            add: char,
        ) -> Result<(), Error> {
            r.1 = add;
            if r.1 < r.0 {
                Err(Error {
                    glob: Some(glob.to_string()),
                    kind: ErrorKind::InvalidRange(r.0, r.1),
                })
            } else {
                Ok(())
            }
        }
        let mut ranges = vec![];
        let negated = match self.chars.peek() {
            Some(&'!') | Some(&'^') => {
                let bump = self.bump();
                assert!(bump == Some('!') || bump == Some('^'));
                true
            }
            _ => false,
        };
        let mut first = true;
        let mut in_range = false;
        loop {
            let c = match self.bump() {
                Some(c) => c,
                // The only way to successfully break this loop is to observe
                // a ']'.
                None => return Err(self.error(ErrorKind::UnclosedClass)),
            };
            match c {
                ']' => {
                    if first {
                        ranges.push((']', ']'));
                    } else {
                        break;
                    }
                }
                '-' => {
                    if first {
                        ranges.push(('-', '-'));
                    } else if in_range {
                        // invariant: in_range is only set when there is
                        // already at least one character seen.
                        let r = ranges.last_mut().unwrap();
                        add_to_last_range(&self.glob, r, '-')?;
                        in_range = false;
                    } else {
                        assert!(!ranges.is_empty());
                        in_range = true;
                    }
                }
                c => {
                    if in_range {
                        // invariant: in_range is only set when there is
                        // already at least one character seen.
                        add_to_last_range(
                            &self.glob,
                            ranges.last_mut().unwrap(),
                            c,
                        )?;
                    } else {
                        ranges.push((c, c));
                    }
                    in_range = false;
                }
            }
            first = false;
        }
        if in_range {
            // Means that the last character in the class was a '-', so add
            // it as a literal.
            ranges.push(('-', '-'));
        }
        self.push_token(Token::Class { negated: negated, ranges: ranges })
    }

    fn bump(&mut self) -> Option<char> {
        self.prev = self.cur;
        self.cur = self.chars.next();
        self.cur
    }

    fn peek(&mut self) -> Option<char> {
        self.chars.peek().map(|&ch| ch)
    }
}

#[cfg(test)]
fn starts_with(needle: &[u8], haystack: &[u8]) -> bool {
    needle.len() <= haystack.len() && needle == &haystack[..needle.len()]
}

#[cfg(test)]
fn ends_with(needle: &[u8], haystack: &[u8]) -> bool {
    if needle.len() > haystack.len() {
        return false;
    }
    needle == &haystack[haystack.len() - needle.len()..]
}

#[cfg(test)]
mod tests {
    use super::Token::*;
    use super::{Glob, GlobBuilder, Token};
    use crate::{ErrorKind, GlobSetBuilder};

    #[derive(Clone, Copy, Debug, Default)]
    struct Options {
        casei: Option<bool>,
        litsep: Option<bool>,
        bsesc: Option<bool>,
    }

    macro_rules! syntax {
        ($name:ident, $pat:expr, $tokens:expr) => {
            #[test]
            fn $name() {
                let pat = Glob::new($pat).unwrap();
                assert_eq!($tokens, pat.tokens.0);
            }
        };
    }

    macro_rules! syntaxerr {
        ($name:ident, $pat:expr, $err:expr) => {
            #[test]
            fn $name() {
                let err = Glob::new($pat).unwrap_err();
                assert_eq!(&$err, err.kind());
            }
        };
    }

    macro_rules! toregex {
        ($name:ident, $pat:expr, $re:expr) => {
            toregex!($name, $pat, $re, Options::default());
        };
        ($name:ident, $pat:expr, $re:expr, $options:expr) => {
            #[test]
            fn $name() {
                let mut builder = GlobBuilder::new($pat);
                if let Some(casei) = $options.casei {
                    builder.case_insensitive(casei);
                }
                if let Some(litsep) = $options.litsep {
                    builder.literal_separator(litsep);
                }
                if let Some(bsesc) = $options.bsesc {
                    builder.backslash_escape(bsesc);
                }
                let pat = builder.build().unwrap();
                assert_eq!(format!("(?-u){}", $re), pat.regex());
            }
        };
    }

    macro_rules! matches {
        ($name:ident, $pat:expr, $path:expr) => {
            matches!($name, $pat, $path, Options::default());
        };
        ($name:ident, $pat:expr, $path:expr, $options:expr) => {
            #[test]
            fn $name() {
                let mut builder = GlobBuilder::new($pat);
                if let Some(casei) = $options.casei {
                    builder.case_insensitive(casei);
                }
                if let Some(litsep) = $options.litsep {
                    builder.literal_separator(litsep);
                }
                if let Some(bsesc) = $options.bsesc {
                    builder.backslash_escape(bsesc);
                }
                let pat = builder.build().unwrap();
                let matcher = pat.compile_matcher();
                let strategic = pat.compile_strategic_matcher();
                let set = GlobSetBuilder::new().add(pat).build().unwrap();
                assert!(matcher.is_match($path));
                assert!(strategic.is_match($path));
                assert!(set.is_match($path));
            }
        };
    }

    macro_rules! nmatches {
        ($name:ident, $pat:expr, $path:expr) => {
            nmatches!($name, $pat, $path, Options::default());
        };
        ($name:ident, $pat:expr, $path:expr, $options:expr) => {
            #[test]
            fn $name() {
                let mut builder = GlobBuilder::new($pat);
                if let Some(casei) = $options.casei {
                    builder.case_insensitive(casei);
                }
                if let Some(litsep) = $options.litsep {
                    builder.literal_separator(litsep);
                }
                if let Some(bsesc) = $options.bsesc {
                    builder.backslash_escape(bsesc);
                }
                let pat = builder.build().unwrap();
                let matcher = pat.compile_matcher();
                let strategic = pat.compile_strategic_matcher();
                let set = GlobSetBuilder::new().add(pat).build().unwrap();
                assert!(!matcher.is_match($path));
                assert!(!strategic.is_match($path));
                assert!(!set.is_match($path));
            }
        };
    }

    fn s(string: &str) -> String {
        string.to_string()
    }

    fn class(s: char, e: char) -> Token {
        Class { negated: false, ranges: vec![(s, e)] }
    }

    fn classn(s: char, e: char) -> Token {
        Class { negated: true, ranges: vec![(s, e)] }
    }

    fn rclass(ranges: &[(char, char)]) -> Token {
        Class { negated: false, ranges: ranges.to_vec() }
    }

    fn rclassn(ranges: &[(char, char)]) -> Token {
        Class { negated: true, ranges: ranges.to_vec() }
    }

    syntax!(literal1, "a", vec![Literal('a')]);
    syntax!(literal2, "ab", vec![Literal('a'), Literal('b')]);
    syntax!(any1, "?", vec![Any]);
    syntax!(any2, "a?b", vec![Literal('a'), Any, Literal('b')]);
    syntax!(seq1, "*", vec![ZeroOrMore]);
    syntax!(seq2, "a*b", vec![Literal('a'), ZeroOrMore, Literal('b')]);
    syntax!(
        seq3,
        "*a*b*",
        vec![ZeroOrMore, Literal('a'), ZeroOrMore, Literal('b'), ZeroOrMore,]
    );
    syntax!(rseq1, "**", vec![RecursivePrefix]);
    syntax!(rseq2, "**/", vec![RecursivePrefix]);
    syntax!(rseq3, "/**", vec![RecursiveSuffix]);
    syntax!(rseq4, "/**/", vec![RecursiveZeroOrMore]);
    syntax!(
        rseq5,
        "a/**/b",
        vec![Literal('a'), RecursiveZeroOrMore, Literal('b'),]
    );
    syntax!(cls1, "[a]", vec![class('a', 'a')]);
    syntax!(cls2, "[!a]", vec![classn('a', 'a')]);
    syntax!(cls3, "[a-z]", vec![class('a', 'z')]);
    syntax!(cls4, "[!a-z]", vec![classn('a', 'z')]);
    syntax!(cls5, "[-]", vec![class('-', '-')]);
    syntax!(cls6, "[]]", vec![class(']', ']')]);
    syntax!(cls7, "[*]", vec![class('*', '*')]);
    syntax!(cls8, "[!!]", vec![classn('!', '!')]);
    syntax!(cls9, "[a-]", vec![rclass(&[('a', 'a'), ('-', '-')])]);
    syntax!(cls10, "[-a-z]", vec![rclass(&[('-', '-'), ('a', 'z')])]);
    syntax!(cls11, "[a-z-]", vec![rclass(&[('a', 'z'), ('-', '-')])]);
    syntax!(
        cls12,
        "[-a-z-]",
        vec![rclass(&[('-', '-'), ('a', 'z'), ('-', '-')]),]
    );
    syntax!(cls13, "[]-z]", vec![class(']', 'z')]);
    syntax!(cls14, "[--z]", vec![class('-', 'z')]);
    syntax!(cls15, "[ --]", vec![class(' ', '-')]);
    syntax!(cls16, "[0-9a-z]", vec![rclass(&[('0', '9'), ('a', 'z')])]);
    syntax!(cls17, "[a-z0-9]", vec![rclass(&[('a', 'z'), ('0', '9')])]);
    syntax!(cls18, "[!0-9a-z]", vec![rclassn(&[('0', '9'), ('a', 'z')])]);
    syntax!(cls19, "[!a-z0-9]", vec![rclassn(&[('a', 'z'), ('0', '9')])]);
    syntax!(cls20, "[^a]", vec![classn('a', 'a')]);
    syntax!(cls21, "[^a-z]", vec![classn('a', 'z')]);

    syntaxerr!(err_unclosed1, "[", ErrorKind::UnclosedClass);
    syntaxerr!(err_unclosed2, "[]", ErrorKind::UnclosedClass);
    syntaxerr!(err_unclosed3, "[!", ErrorKind::UnclosedClass);
    syntaxerr!(err_unclosed4, "[!]", ErrorKind::UnclosedClass);
    syntaxerr!(err_range1, "[z-a]", ErrorKind::InvalidRange('z', 'a'));
    syntaxerr!(err_range2, "[z--]", ErrorKind::InvalidRange('z', '-'));

    const CASEI: Options =
        Options { casei: Some(true), litsep: None, bsesc: None };
    const SLASHLIT: Options =
        Options { casei: None, litsep: Some(true), bsesc: None };
    const NOBSESC: Options =
        Options { casei: None, litsep: None, bsesc: Some(false) };
    const BSESC: Options =
        Options { casei: None, litsep: None, bsesc: Some(true) };

    toregex!(re_casei, "a", "(?i)^a$", &CASEI);

    toregex!(re_slash1, "?", r"^[^/]$", SLASHLIT);
    toregex!(re_slash2, "*", r"^[^/]*$", SLASHLIT);

    toregex!(re1, "a", "^a$");
    toregex!(re2, "?", "^.$");
    toregex!(re3, "*", "^.*$");
    toregex!(re4, "a?", "^a.$");
    toregex!(re5, "?a", "^.a$");
    toregex!(re6, "a*", "^a.*$");
    toregex!(re7, "*a", "^.*a$");
    toregex!(re8, "[*]", r"^[\*]$");
    toregex!(re9, "[+]", r"^[\+]$");
    toregex!(re10, "+", r"^\+$");
    toregex!(re11, "☃", r"^\xe2\x98\x83$");
    toregex!(re12, "**", r"^.*$");
    toregex!(re13, "**/", r"^.*$");
    toregex!(re14, "**/*", r"^(?:/?|.*/).*$");
    toregex!(re15, "**/**", r"^.*$");
    toregex!(re16, "**/**/*", r"^(?:/?|.*/).*$");
    toregex!(re17, "**/**/**", r"^.*$");
    toregex!(re18, "**/**/**/*", r"^(?:/?|.*/).*$");
    toregex!(re19, "a/**", r"^a/.*$");
    toregex!(re20, "a/**/**", r"^a/.*$");
    toregex!(re21, "a/**/**/**", r"^a/.*$");
    toregex!(re22, "a/**/b", r"^a(?:/|/.*/)b$");
    toregex!(re23, "a/**/**/b", r"^a(?:/|/.*/)b$");
    toregex!(re24, "a/**/**/**/b", r"^a(?:/|/.*/)b$");
    toregex!(re25, "**/b", r"^(?:/?|.*/)b$");
    toregex!(re26, "**/**/b", r"^(?:/?|.*/)b$");
    toregex!(re27, "**/**/**/b", r"^(?:/?|.*/)b$");
    toregex!(re28, "a**", r"^a.*.*$");
    toregex!(re29, "**a", r"^.*.*a$");
    toregex!(re30, "a**b", r"^a.*.*b$");
    toregex!(re31, "***", r"^.*.*.*$");
    toregex!(re32, "/a**", r"^/a.*.*$");
    toregex!(re33, "/**a", r"^/.*.*a$");
    toregex!(re34, "/a**b", r"^/a.*.*b$");

    matches!(match1, "a", "a");
    matches!(match2, "a*b", "a_b");
    matches!(match3, "a*b*c", "abc");
    matches!(match4, "a*b*c", "a_b_c");
    matches!(match5, "a*b*c", "a___b___c");
    matches!(match6, "abc*abc*abc", "abcabcabcabcabcabcabc");
    matches!(match7, "a*a*a*a*a*a*a*a*a", "aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa");
    matches!(match8, "a*b[xyz]c*d", "abxcdbxcddd");
    matches!(match9, "*.rs", ".rs");
    matches!(match10, "☃", "☃");

    matches!(matchrec1, "some/**/needle.txt", "some/needle.txt");
    matches!(matchrec2, "some/**/needle.txt", "some/one/needle.txt");
    matches!(matchrec3, "some/**/needle.txt", "some/one/two/needle.txt");
    matches!(matchrec4, "some/**/needle.txt", "some/other/needle.txt");
    matches!(matchrec5, "**", "abcde");
    matches!(matchrec6, "**", "");
    matches!(matchrec7, "**", ".asdf");
    matches!(matchrec8, "**", "/x/.asdf");
    matches!(matchrec9, "some/**/**/needle.txt", "some/needle.txt");
    matches!(matchrec10, "some/**/**/needle.txt", "some/one/needle.txt");
    matches!(matchrec11, "some/**/**/needle.txt", "some/one/two/needle.txt");
    matches!(matchrec12, "some/**/**/needle.txt", "some/other/needle.txt");
    matches!(matchrec13, "**/test", "one/two/test");
    matches!(matchrec14, "**/test", "one/test");
    matches!(matchrec15, "**/test", "test");
    matches!(matchrec16, "/**/test", "/one/two/test");
    matches!(matchrec17, "/**/test", "/one/test");
    matches!(matchrec18, "/**/test", "/test");
    matches!(matchrec19, "**/.*", ".abc");
    matches!(matchrec20, "**/.*", "abc/.abc");
    matches!(matchrec21, "**/foo/bar", "foo/bar");
    matches!(matchrec22, ".*/**", ".abc/abc");
    matches!(matchrec23, "test/**", "test/");
    matches!(matchrec24, "test/**", "test/one");
    matches!(matchrec25, "test/**", "test/one/two");
    matches!(matchrec26, "some/*/needle.txt", "some/one/needle.txt");

    matches!(matchrange1, "a[0-9]b", "a0b");
    matches!(matchrange2, "a[0-9]b", "a9b");
    matches!(matchrange3, "a[!0-9]b", "a_b");
    matches!(matchrange4, "[a-z123]", "1");
    matches!(matchrange5, "[1a-z23]", "1");
    matches!(matchrange6, "[123a-z]", "1");
    matches!(matchrange7, "[abc-]", "-");
    matches!(matchrange8, "[-abc]", "-");
    matches!(matchrange9, "[-a-c]", "b");
    matches!(matchrange10, "[a-c-]", "b");
    matches!(matchrange11, "[-]", "-");
    matches!(matchrange12, "a[^0-9]b", "a_b");

    matches!(matchpat1, "*hello.txt", "hello.txt");
    matches!(matchpat2, "*hello.txt", "gareth_says_hello.txt");
    matches!(matchpat3, "*hello.txt", "some/path/to/hello.txt");
    matches!(matchpat4, "*hello.txt", "some\\path\\to\\hello.txt");
    matches!(matchpat5, "*hello.txt", "/an/absolute/path/to/hello.txt");
    matches!(matchpat6, "*some/path/to/hello.txt", "some/path/to/hello.txt");
    matches!(
        matchpat7,
        "*some/path/to/hello.txt",
        "a/bigger/some/path/to/hello.txt"
    );

    matches!(matchescape, "_[[]_[]]_[?]_[*]_!_", "_[_]_?_*_!_");

    matches!(matchcasei1, "aBcDeFg", "aBcDeFg", CASEI);
    matches!(matchcasei2, "aBcDeFg", "abcdefg", CASEI);
    matches!(matchcasei3, "aBcDeFg", "ABCDEFG", CASEI);
    matches!(matchcasei4, "aBcDeFg", "AbCdEfG", CASEI);

    matches!(matchalt1, "a,b", "a,b");
    matches!(matchalt2, ",", ",");
    matches!(matchalt3, "{a,b}", "a");
    matches!(matchalt4, "{a,b}", "b");
    matches!(matchalt5, "{**/src/**,foo}", "abc/src/bar");
    matches!(matchalt6, "{**/src/**,foo}", "foo");
    matches!(matchalt7, "{[}],foo}", "}");
    matches!(matchalt8, "{foo}", "foo");
    matches!(matchalt9, "{}", "");
    matches!(matchalt10, "{,}", "");
    matches!(matchalt11, "{*.foo,*.bar,*.wat}", "test.foo");
    matches!(matchalt12, "{*.foo,*.bar,*.wat}", "test.bar");
    matches!(matchalt13, "{*.foo,*.bar,*.wat}", "test.wat");

    matches!(matchslash1, "abc/def", "abc/def", SLASHLIT);
    #[cfg(unix)]
    nmatches!(matchslash2, "abc?def", "abc/def", SLASHLIT);
    #[cfg(not(unix))]
    nmatches!(matchslash2, "abc?def", "abc\\def", SLASHLIT);
    nmatches!(matchslash3, "abc*def", "abc/def", SLASHLIT);
    matches!(matchslash4, "abc[/]def", "abc/def", SLASHLIT); // differs
    #[cfg(unix)]
    nmatches!(matchslash5, "abc\\def", "abc/def", SLASHLIT);
    #[cfg(not(unix))]
    matches!(matchslash5, "abc\\def", "abc/def", SLASHLIT);

    matches!(matchbackslash1, "\\[", "[", BSESC);
    matches!(matchbackslash2, "\\?", "?", BSESC);
    matches!(matchbackslash3, "\\*", "*", BSESC);
    matches!(matchbackslash4, "\\[a-z]", "\\a", NOBSESC);
    matches!(matchbackslash5, "\\?", "\\a", NOBSESC);
    matches!(matchbackslash6, "\\*", "\\\\", NOBSESC);
    #[cfg(unix)]
    matches!(matchbackslash7, "\\a", "a");
    #[cfg(not(unix))]
    matches!(matchbackslash8, "\\a", "/a");

    nmatches!(matchnot1, "a*b*c", "abcd");
    nmatches!(matchnot2, "abc*abc*abc", "abcabcabcabcabcabcabca");
    nmatches!(matchnot3, "some/**/needle.txt", "some/other/notthis.txt");
    nmatches!(matchnot4, "some/**/**/needle.txt", "some/other/notthis.txt");
    nmatches!(matchnot5, "/**/test", "test");
    nmatches!(matchnot6, "/**/test", "/one/notthis");
    nmatches!(matchnot7, "/**/test", "/notthis");
    nmatches!(matchnot8, "**/.*", "ab.c");
    nmatches!(matchnot9, "**/.*", "abc/ab.c");
    nmatches!(matchnot10, ".*/**", "a.bc");
    nmatches!(matchnot11, ".*/**", "abc/a.bc");
    nmatches!(matchnot12, "a[0-9]b", "a_b");
    nmatches!(matchnot13, "a[!0-9]b", "a0b");
    nmatches!(matchnot14, "a[!0-9]b", "a9b");
    nmatches!(matchnot15, "[!-]", "-");
    nmatches!(matchnot16, "*hello.txt", "hello.txt-and-then-some");
    nmatches!(matchnot17, "*hello.txt", "goodbye.txt");
    nmatches!(
        matchnot18,
        "*some/path/to/hello.txt",
        "some/path/to/hello.txt-and-then-some"
    );
    nmatches!(
        matchnot19,
        "*some/path/to/hello.txt",
        "some/other/path/to/hello.txt"
    );
    nmatches!(matchnot20, "a", "foo/a");
    nmatches!(matchnot21, "./foo", "foo");
    nmatches!(matchnot22, "**/foo", "foofoo");
    nmatches!(matchnot23, "**/foo/bar", "foofoo/bar");
    nmatches!(matchnot24, "/*.c", "mozilla-sha1/sha1.c");
    nmatches!(matchnot25, "*.c", "mozilla-sha1/sha1.c", SLASHLIT);
    nmatches!(
        matchnot26,
        "**/m4/ltoptions.m4",
        "csharp/src/packages/repositories.config",
        SLASHLIT
    );
    nmatches!(matchnot27, "a[^0-9]b", "a0b");
    nmatches!(matchnot28, "a[^0-9]b", "a9b");
    nmatches!(matchnot29, "[^-]", "-");
    nmatches!(matchnot30, "some/*/needle.txt", "some/needle.txt");
    nmatches!(
        matchrec31,
        "some/*/needle.txt",
        "some/one/two/needle.txt",
        SLASHLIT
    );
    nmatches!(
        matchrec32,
        "some/*/needle.txt",
        "some/one/two/three/needle.txt",
        SLASHLIT
    );
    nmatches!(matchrec33, ".*/**", ".abc");
    nmatches!(matchrec34, "foo/**", "foo");

    macro_rules! extract {
        ($which:ident, $name:ident, $pat:expr, $expect:expr) => {
            extract!($which, $name, $pat, $expect, Options::default());
        };
        ($which:ident, $name:ident, $pat:expr, $expect:expr, $options:expr) => {
            #[test]
            fn $name() {
                let mut builder = GlobBuilder::new($pat);
                if let Some(casei) = $options.casei {
                    builder.case_insensitive(casei);
                }
                if let Some(litsep) = $options.litsep {
                    builder.literal_separator(litsep);
                }
                if let Some(bsesc) = $options.bsesc {
                    builder.backslash_escape(bsesc);
                }
                let pat = builder.build().unwrap();
                assert_eq!($expect, pat.$which());
            }
        };
    }

    macro_rules! literal {
        ($($tt:tt)*) => { extract!(literal, $($tt)*); }
    }

    macro_rules! basetokens {
        ($($tt:tt)*) => { extract!(basename_tokens, $($tt)*); }
    }

    macro_rules! ext {
        ($($tt:tt)*) => { extract!(ext, $($tt)*); }
    }

    macro_rules! required_ext {
        ($($tt:tt)*) => { extract!(required_ext, $($tt)*); }
    }

    macro_rules! prefix {
        ($($tt:tt)*) => { extract!(prefix, $($tt)*); }
    }

    macro_rules! suffix {
        ($($tt:tt)*) => { extract!(suffix, $($tt)*); }
    }

    macro_rules! baseliteral {
        ($($tt:tt)*) => { extract!(basename_literal, $($tt)*); }
    }

    literal!(extract_lit1, "foo", Some(s("foo")));
    literal!(extract_lit2, "foo", None, CASEI);
    literal!(extract_lit3, "/foo", Some(s("/foo")));
    literal!(extract_lit4, "/foo/", Some(s("/foo/")));
    literal!(extract_lit5, "/foo/bar", Some(s("/foo/bar")));
    literal!(extract_lit6, "*.foo", None);
    literal!(extract_lit7, "foo/bar", Some(s("foo/bar")));
    literal!(extract_lit8, "**/foo/bar", None);

    basetokens!(
        extract_basetoks1,
        "**/foo",
        Some(&*vec![Literal('f'), Literal('o'), Literal('o'),])
    );
    basetokens!(extract_basetoks2, "**/foo", None, CASEI);
    basetokens!(
        extract_basetoks3,
        "**/foo",
        Some(&*vec![Literal('f'), Literal('o'), Literal('o'),]),
        SLASHLIT
    );
    basetokens!(extract_basetoks4, "*foo", None, SLASHLIT);
    basetokens!(extract_basetoks5, "*foo", None);
    basetokens!(extract_basetoks6, "**/fo*o", None);
    basetokens!(
        extract_basetoks7,
        "**/fo*o",
        Some(&*vec![Literal('f'), Literal('o'), ZeroOrMore, Literal('o'),]),
        SLASHLIT
    );

    ext!(extract_ext1, "**/*.rs", Some(s(".rs")));
    ext!(extract_ext2, "**/*.rs.bak", None);
    ext!(extract_ext3, "*.rs", Some(s(".rs")));
    ext!(extract_ext4, "a*.rs", None);
    ext!(extract_ext5, "/*.c", None);
    ext!(extract_ext6, "*.c", None, SLASHLIT);
    ext!(extract_ext7, "*.c", Some(s(".c")));

    required_ext!(extract_req_ext1, "*.rs", Some(s(".rs")));
    required_ext!(extract_req_ext2, "/foo/bar/*.rs", Some(s(".rs")));
    required_ext!(extract_req_ext3, "/foo/bar/*.rs", Some(s(".rs")));
    required_ext!(extract_req_ext4, "/foo/bar/.rs", Some(s(".rs")));
    required_ext!(extract_req_ext5, ".rs", Some(s(".rs")));
    required_ext!(extract_req_ext6, "./rs", None);
    required_ext!(extract_req_ext7, "foo", None);
    required_ext!(extract_req_ext8, ".foo/", None);
    required_ext!(extract_req_ext9, "foo/", None);

    prefix!(extract_prefix1, "/foo", Some(s("/foo")));
    prefix!(extract_prefix2, "/foo/*", Some(s("/foo/")));
    prefix!(extract_prefix3, "**/foo", None);
    prefix!(extract_prefix4, "foo/**", Some(s("foo/")));

    suffix!(extract_suffix1, "**/foo/bar", Some((s("/foo/bar"), true)));
    suffix!(extract_suffix2, "*/foo/bar", Some((s("/foo/bar"), false)));
    suffix!(extract_suffix3, "*/foo/bar", None, SLASHLIT);
    suffix!(extract_suffix4, "foo/bar", Some((s("foo/bar"), false)));
    suffix!(extract_suffix5, "*.foo", Some((s(".foo"), false)));
    suffix!(extract_suffix6, "*.foo", None, SLASHLIT);
    suffix!(extract_suffix7, "**/*_test", Some((s("_test"), false)));

    baseliteral!(extract_baselit1, "**/foo", Some(s("foo")));
    baseliteral!(extract_baselit2, "foo", None);
    baseliteral!(extract_baselit3, "*foo", None);
    baseliteral!(extract_baselit4, "*/foo", None);
}