forked from golang/go
-
Notifications
You must be signed in to change notification settings - Fork 0
Expand file tree
/
Copy pathparse.go
More file actions
1977 lines (1820 loc) · 47.7 KB
/
parse.go
File metadata and controls
1977 lines (1820 loc) · 47.7 KB
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
// Copyright 2011 The Go Authors. All rights reserved.
// Use of this source code is governed by a BSD-style
// license that can be found in the LICENSE file.
package syntax
import (
"sort"
"strings"
"unicode"
"unicode/utf8"
)
// An Error describes a failure to parse a regular expression
// and gives the offending expression.
type Error struct {
Code ErrorCode
Expr string
}
func (e *Error) Error() string {
return "error parsing regexp: " + e.Code.String() + ": `" + e.Expr + "`"
}
// An ErrorCode describes a failure to parse a regular expression.
type ErrorCode string
const (
// Unexpected error
ErrInternalError ErrorCode = "regexp/syntax: internal error"
// Parse errors
ErrInvalidCharClass ErrorCode = "invalid character class"
ErrInvalidCharRange ErrorCode = "invalid character class range"
ErrInvalidEscape ErrorCode = "invalid escape sequence"
ErrInvalidNamedCapture ErrorCode = "invalid named capture"
ErrInvalidPerlOp ErrorCode = "invalid or unsupported Perl syntax"
ErrInvalidRepeatOp ErrorCode = "invalid nested repetition operator"
ErrInvalidRepeatSize ErrorCode = "invalid repeat count"
ErrInvalidUTF8 ErrorCode = "invalid UTF-8"
ErrMissingBracket ErrorCode = "missing closing ]"
ErrMissingParen ErrorCode = "missing closing )"
ErrMissingRepeatArgument ErrorCode = "missing argument to repetition operator"
ErrTrailingBackslash ErrorCode = "trailing backslash at end of expression"
ErrUnexpectedParen ErrorCode = "unexpected )"
ErrNestingDepth ErrorCode = "expression nests too deeply"
)
func (e ErrorCode) String() string {
return string(e)
}
// Flags control the behavior of the parser and record information about regexp context.
type Flags uint16
const (
FoldCase Flags = 1 << iota // case-insensitive match
Literal // treat pattern as literal string
ClassNL // allow character classes like [^a-z] and [[:space:]] to match newline
DotNL // allow . to match newline
OneLine // treat ^ and $ as only matching at beginning and end of text
NonGreedy // make repetition operators default to non-greedy
PerlX // allow Perl extensions
UnicodeGroups // allow \p{Han}, \P{Han} for Unicode group and negation
WasDollar // regexp OpEndText was $, not \z
Simple // regexp contains no counted repetition
MatchNL = ClassNL | DotNL
Perl = ClassNL | OneLine | PerlX | UnicodeGroups // as close to Perl as possible
POSIX Flags = 0 // POSIX syntax
)
// Pseudo-ops for parsing stack.
const (
opLeftParen = opPseudo + iota
opVerticalBar
)
// maxHeight is the maximum height of a regexp parse tree.
// It is somewhat arbitrarily chosen, but the idea is to be large enough
// that no one will actually hit in real use but at the same time small enough
// that recursion on the Regexp tree will not hit the 1GB Go stack limit.
// The maximum amount of stack for a single recursive frame is probably
// closer to 1kB, so this could potentially be raised, but it seems unlikely
// that people have regexps nested even this deeply.
// We ran a test on Google's C++ code base and turned up only
// a single use case with depth > 100; it had depth 128.
// Using depth 1000 should be plenty of margin.
// As an optimization, we don't even bother calculating heights
// until we've allocated at least maxHeight Regexp structures.
const maxHeight = 1000
type parser struct {
flags Flags // parse mode flags
stack []*Regexp // stack of parsed expressions
free *Regexp
numCap int // number of capturing groups seen
wholeRegexp string
tmpClass []rune // temporary char class work space
numRegexp int // number of regexps allocated
height map[*Regexp]int // regexp height for height limit check
}
func (p *parser) newRegexp(op Op) *Regexp {
re := p.free
if re != nil {
p.free = re.Sub0[0]
*re = Regexp{}
} else {
re = new(Regexp)
p.numRegexp++
}
re.Op = op
return re
}
func (p *parser) reuse(re *Regexp) {
if p.height != nil {
delete(p.height, re)
}
re.Sub0[0] = p.free
p.free = re
}
func (p *parser) checkHeight(re *Regexp) {
if p.numRegexp < maxHeight {
return
}
if p.height == nil {
p.height = make(map[*Regexp]int)
for _, re := range p.stack {
p.checkHeight(re)
}
}
if p.calcHeight(re, true) > maxHeight {
panic(ErrNestingDepth)
}
}
func (p *parser) calcHeight(re *Regexp, force bool) int {
if !force {
if h, ok := p.height[re]; ok {
return h
}
}
h := 1
for _, sub := range re.Sub {
hsub := p.calcHeight(sub, false)
if h < 1+hsub {
h = 1 + hsub
}
}
p.height[re] = h
return h
}
// Parse stack manipulation.
// push pushes the regexp re onto the parse stack and returns the regexp.
func (p *parser) push(re *Regexp) *Regexp {
if re.Op == OpCharClass && len(re.Rune) == 2 && re.Rune[0] == re.Rune[1] {
// Single rune.
if p.maybeConcat(re.Rune[0], p.flags&^FoldCase) {
return nil
}
re.Op = OpLiteral
re.Rune = re.Rune[:1]
re.Flags = p.flags &^ FoldCase
} else if re.Op == OpCharClass && len(re.Rune) == 4 &&
re.Rune[0] == re.Rune[1] && re.Rune[2] == re.Rune[3] &&
unicode.SimpleFold(re.Rune[0]) == re.Rune[2] &&
unicode.SimpleFold(re.Rune[2]) == re.Rune[0] ||
re.Op == OpCharClass && len(re.Rune) == 2 &&
re.Rune[0]+1 == re.Rune[1] &&
unicode.SimpleFold(re.Rune[0]) == re.Rune[1] &&
unicode.SimpleFold(re.Rune[1]) == re.Rune[0] {
// Case-insensitive rune like [Aa] or [Δδ].
if p.maybeConcat(re.Rune[0], p.flags|FoldCase) {
return nil
}
// Rewrite as (case-insensitive) literal.
re.Op = OpLiteral
re.Rune = re.Rune[:1]
re.Flags = p.flags | FoldCase
} else {
// Incremental concatenation.
p.maybeConcat(-1, 0)
}
p.stack = append(p.stack, re)
p.checkHeight(re)
return re
}
// maybeConcat implements incremental concatenation
// of literal runes into string nodes. The parser calls this
// before each push, so only the top fragment of the stack
// might need processing. Since this is called before a push,
// the topmost literal is no longer subject to operators like *
// (Otherwise ab* would turn into (ab)*.)
// If r >= 0 and there's a node left over, maybeConcat uses it
// to push r with the given flags.
// maybeConcat reports whether r was pushed.
func (p *parser) maybeConcat(r rune, flags Flags) bool {
n := len(p.stack)
if n < 2 {
return false
}
re1 := p.stack[n-1]
re2 := p.stack[n-2]
if re1.Op != OpLiteral || re2.Op != OpLiteral || re1.Flags&FoldCase != re2.Flags&FoldCase {
return false
}
// Push re1 into re2.
re2.Rune = append(re2.Rune, re1.Rune...)
// Reuse re1 if possible.
if r >= 0 {
re1.Rune = re1.Rune0[:1]
re1.Rune[0] = r
re1.Flags = flags
return true
}
p.stack = p.stack[:n-1]
p.reuse(re1)
return false // did not push r
}
// literal pushes a literal regexp for the rune r on the stack.
func (p *parser) literal(r rune) {
re := p.newRegexp(OpLiteral)
re.Flags = p.flags
if p.flags&FoldCase != 0 {
r = minFoldRune(r)
}
re.Rune0[0] = r
re.Rune = re.Rune0[:1]
p.push(re)
}
// minFoldRune returns the minimum rune fold-equivalent to r.
func minFoldRune(r rune) rune {
if r < minFold || r > maxFold {
return r
}
min := r
r0 := r
for r = unicode.SimpleFold(r); r != r0; r = unicode.SimpleFold(r) {
if min > r {
min = r
}
}
return min
}
// op pushes a regexp with the given op onto the stack
// and returns that regexp.
func (p *parser) op(op Op) *Regexp {
re := p.newRegexp(op)
re.Flags = p.flags
return p.push(re)
}
// repeat replaces the top stack element with itself repeated according to op, min, max.
// before is the regexp suffix starting at the repetition operator.
// after is the regexp suffix following after the repetition operator.
// repeat returns an updated 'after' and an error, if any.
func (p *parser) repeat(op Op, min, max int, before, after, lastRepeat string) (string, error) {
flags := p.flags
if p.flags&PerlX != 0 {
if len(after) > 0 && after[0] == '?' {
after = after[1:]
flags ^= NonGreedy
}
if lastRepeat != "" {
// In Perl it is not allowed to stack repetition operators:
// a** is a syntax error, not a doubled star, and a++ means
// something else entirely, which we don't support!
return "", &Error{ErrInvalidRepeatOp, lastRepeat[:len(lastRepeat)-len(after)]}
}
}
n := len(p.stack)
if n == 0 {
return "", &Error{ErrMissingRepeatArgument, before[:len(before)-len(after)]}
}
sub := p.stack[n-1]
if sub.Op >= opPseudo {
return "", &Error{ErrMissingRepeatArgument, before[:len(before)-len(after)]}
}
re := p.newRegexp(op)
re.Min = min
re.Max = max
re.Flags = flags
re.Sub = re.Sub0[:1]
re.Sub[0] = sub
p.stack[n-1] = re
p.checkHeight(re)
if op == OpRepeat && (min >= 2 || max >= 2) && !repeatIsValid(re, 1000) {
return "", &Error{ErrInvalidRepeatSize, before[:len(before)-len(after)]}
}
return after, nil
}
// repeatIsValid reports whether the repetition re is valid.
// Valid means that the combination of the top-level repetition
// and any inner repetitions does not exceed n copies of the
// innermost thing.
// This function rewalks the regexp tree and is called for every repetition,
// so we have to worry about inducing quadratic behavior in the parser.
// We avoid this by only calling repeatIsValid when min or max >= 2.
// In that case the depth of any >= 2 nesting can only get to 9 without
// triggering a parse error, so each subtree can only be rewalked 9 times.
func repeatIsValid(re *Regexp, n int) bool {
if re.Op == OpRepeat {
m := re.Max
if m == 0 {
return true
}
if m < 0 {
m = re.Min
}
if m > n {
return false
}
if m > 0 {
n /= m
}
}
for _, sub := range re.Sub {
if !repeatIsValid(sub, n) {
return false
}
}
return true
}
// concat replaces the top of the stack (above the topmost '|' or '(') with its concatenation.
func (p *parser) concat() *Regexp {
p.maybeConcat(-1, 0)
// Scan down to find pseudo-operator | or (.
i := len(p.stack)
for i > 0 && p.stack[i-1].Op < opPseudo {
i--
}
subs := p.stack[i:]
p.stack = p.stack[:i]
// Empty concatenation is special case.
if len(subs) == 0 {
return p.push(p.newRegexp(OpEmptyMatch))
}
return p.push(p.collapse(subs, OpConcat))
}
// alternate replaces the top of the stack (above the topmost '(') with its alternation.
func (p *parser) alternate() *Regexp {
// Scan down to find pseudo-operator (.
// There are no | above (.
i := len(p.stack)
for i > 0 && p.stack[i-1].Op < opPseudo {
i--
}
subs := p.stack[i:]
p.stack = p.stack[:i]
// Make sure top class is clean.
// All the others already are (see swapVerticalBar).
if len(subs) > 0 {
cleanAlt(subs[len(subs)-1])
}
// Empty alternate is special case
// (shouldn't happen but easy to handle).
if len(subs) == 0 {
return p.push(p.newRegexp(OpNoMatch))
}
return p.push(p.collapse(subs, OpAlternate))
}
// cleanAlt cleans re for eventual inclusion in an alternation.
func cleanAlt(re *Regexp) {
switch re.Op {
case OpCharClass:
re.Rune = cleanClass(&re.Rune)
if len(re.Rune) == 2 && re.Rune[0] == 0 && re.Rune[1] == unicode.MaxRune {
re.Rune = nil
re.Op = OpAnyChar
return
}
if len(re.Rune) == 4 && re.Rune[0] == 0 && re.Rune[1] == '\n'-1 && re.Rune[2] == '\n'+1 && re.Rune[3] == unicode.MaxRune {
re.Rune = nil
re.Op = OpAnyCharNotNL
return
}
if cap(re.Rune)-len(re.Rune) > 100 {
// re.Rune will not grow any more.
// Make a copy or inline to reclaim storage.
re.Rune = append(re.Rune0[:0], re.Rune...)
}
}
}
// collapse returns the result of applying op to sub.
// If sub contains op nodes, they all get hoisted up
// so that there is never a concat of a concat or an
// alternate of an alternate.
func (p *parser) collapse(subs []*Regexp, op Op) *Regexp {
if len(subs) == 1 {
return subs[0]
}
re := p.newRegexp(op)
re.Sub = re.Sub0[:0]
for _, sub := range subs {
if sub.Op == op {
re.Sub = append(re.Sub, sub.Sub...)
p.reuse(sub)
} else {
re.Sub = append(re.Sub, sub)
}
}
if op == OpAlternate {
re.Sub = p.factor(re.Sub)
if len(re.Sub) == 1 {
old := re
re = re.Sub[0]
p.reuse(old)
}
}
return re
}
// factor factors common prefixes from the alternation list sub.
// It returns a replacement list that reuses the same storage and
// frees (passes to p.reuse) any removed *Regexps.
//
// For example,
//
// ABC|ABD|AEF|BCX|BCY
//
// simplifies by literal prefix extraction to
//
// A(B(C|D)|EF)|BC(X|Y)
//
// which simplifies by character class introduction to
//
// A(B[CD]|EF)|BC[XY]
func (p *parser) factor(sub []*Regexp) []*Regexp {
if len(sub) < 2 {
return sub
}
// Round 1: Factor out common literal prefixes.
var str []rune
var strflags Flags
start := 0
out := sub[:0]
for i := 0; i <= len(sub); i++ {
// Invariant: the Regexps that were in sub[0:start] have been
// used or marked for reuse, and the slice space has been reused
// for out (len(out) <= start).
//
// Invariant: sub[start:i] consists of regexps that all begin
// with str as modified by strflags.
var istr []rune
var iflags Flags
if i < len(sub) {
istr, iflags = p.leadingString(sub[i])
if iflags == strflags {
same := 0
for same < len(str) && same < len(istr) && str[same] == istr[same] {
same++
}
if same > 0 {
// Matches at least one rune in current range.
// Keep going around.
str = str[:same]
continue
}
}
}
// Found end of a run with common leading literal string:
// sub[start:i] all begin with str[0:len(str)], but sub[i]
// does not even begin with str[0].
//
// Factor out common string and append factored expression to out.
if i == start {
// Nothing to do - run of length 0.
} else if i == start+1 {
// Just one: don't bother factoring.
out = append(out, sub[start])
} else {
// Construct factored form: prefix(suffix1|suffix2|...)
prefix := p.newRegexp(OpLiteral)
prefix.Flags = strflags
prefix.Rune = append(prefix.Rune[:0], str...)
for j := start; j < i; j++ {
sub[j] = p.removeLeadingString(sub[j], len(str))
}
suffix := p.collapse(sub[start:i], OpAlternate) // recurse
re := p.newRegexp(OpConcat)
re.Sub = append(re.Sub[:0], prefix, suffix)
out = append(out, re)
}
// Prepare for next iteration.
start = i
str = istr
strflags = iflags
}
sub = out
// Round 2: Factor out common simple prefixes,
// just the first piece of each concatenation.
// This will be good enough a lot of the time.
//
// Complex subexpressions (e.g. involving quantifiers)
// are not safe to factor because that collapses their
// distinct paths through the automaton, which affects
// correctness in some cases.
start = 0
out = sub[:0]
var first *Regexp
for i := 0; i <= len(sub); i++ {
// Invariant: the Regexps that were in sub[0:start] have been
// used or marked for reuse, and the slice space has been reused
// for out (len(out) <= start).
//
// Invariant: sub[start:i] consists of regexps that all begin with ifirst.
var ifirst *Regexp
if i < len(sub) {
ifirst = p.leadingRegexp(sub[i])
if first != nil && first.Equal(ifirst) &&
// first must be a character class OR a fixed repeat of a character class.
(isCharClass(first) || (first.Op == OpRepeat && first.Min == first.Max && isCharClass(first.Sub[0]))) {
continue
}
}
// Found end of a run with common leading regexp:
// sub[start:i] all begin with first but sub[i] does not.
//
// Factor out common regexp and append factored expression to out.
if i == start {
// Nothing to do - run of length 0.
} else if i == start+1 {
// Just one: don't bother factoring.
out = append(out, sub[start])
} else {
// Construct factored form: prefix(suffix1|suffix2|...)
prefix := first
for j := start; j < i; j++ {
reuse := j != start // prefix came from sub[start]
sub[j] = p.removeLeadingRegexp(sub[j], reuse)
}
suffix := p.collapse(sub[start:i], OpAlternate) // recurse
re := p.newRegexp(OpConcat)
re.Sub = append(re.Sub[:0], prefix, suffix)
out = append(out, re)
}
// Prepare for next iteration.
start = i
first = ifirst
}
sub = out
// Round 3: Collapse runs of single literals into character classes.
start = 0
out = sub[:0]
for i := 0; i <= len(sub); i++ {
// Invariant: the Regexps that were in sub[0:start] have been
// used or marked for reuse, and the slice space has been reused
// for out (len(out) <= start).
//
// Invariant: sub[start:i] consists of regexps that are either
// literal runes or character classes.
if i < len(sub) && isCharClass(sub[i]) {
continue
}
// sub[i] is not a char or char class;
// emit char class for sub[start:i]...
if i == start {
// Nothing to do - run of length 0.
} else if i == start+1 {
out = append(out, sub[start])
} else {
// Make new char class.
// Start with most complex regexp in sub[start].
max := start
for j := start + 1; j < i; j++ {
if sub[max].Op < sub[j].Op || sub[max].Op == sub[j].Op && len(sub[max].Rune) < len(sub[j].Rune) {
max = j
}
}
sub[start], sub[max] = sub[max], sub[start]
for j := start + 1; j < i; j++ {
mergeCharClass(sub[start], sub[j])
p.reuse(sub[j])
}
cleanAlt(sub[start])
out = append(out, sub[start])
}
// ... and then emit sub[i].
if i < len(sub) {
out = append(out, sub[i])
}
start = i + 1
}
sub = out
// Round 4: Collapse runs of empty matches into a single empty match.
start = 0
out = sub[:0]
for i := range sub {
if i+1 < len(sub) && sub[i].Op == OpEmptyMatch && sub[i+1].Op == OpEmptyMatch {
continue
}
out = append(out, sub[i])
}
sub = out
return sub
}
// leadingString returns the leading literal string that re begins with.
// The string refers to storage in re or its children.
func (p *parser) leadingString(re *Regexp) ([]rune, Flags) {
if re.Op == OpConcat && len(re.Sub) > 0 {
re = re.Sub[0]
}
if re.Op != OpLiteral {
return nil, 0
}
return re.Rune, re.Flags & FoldCase
}
// removeLeadingString removes the first n leading runes
// from the beginning of re. It returns the replacement for re.
func (p *parser) removeLeadingString(re *Regexp, n int) *Regexp {
if re.Op == OpConcat && len(re.Sub) > 0 {
// Removing a leading string in a concatenation
// might simplify the concatenation.
sub := re.Sub[0]
sub = p.removeLeadingString(sub, n)
re.Sub[0] = sub
if sub.Op == OpEmptyMatch {
p.reuse(sub)
switch len(re.Sub) {
case 0, 1:
// Impossible but handle.
re.Op = OpEmptyMatch
re.Sub = nil
case 2:
old := re
re = re.Sub[1]
p.reuse(old)
default:
copy(re.Sub, re.Sub[1:])
re.Sub = re.Sub[:len(re.Sub)-1]
}
}
return re
}
if re.Op == OpLiteral {
re.Rune = re.Rune[:copy(re.Rune, re.Rune[n:])]
if len(re.Rune) == 0 {
re.Op = OpEmptyMatch
}
}
return re
}
// leadingRegexp returns the leading regexp that re begins with.
// The regexp refers to storage in re or its children.
func (p *parser) leadingRegexp(re *Regexp) *Regexp {
if re.Op == OpEmptyMatch {
return nil
}
if re.Op == OpConcat && len(re.Sub) > 0 {
sub := re.Sub[0]
if sub.Op == OpEmptyMatch {
return nil
}
return sub
}
return re
}
// removeLeadingRegexp removes the leading regexp in re.
// It returns the replacement for re.
// If reuse is true, it passes the removed regexp (if no longer needed) to p.reuse.
func (p *parser) removeLeadingRegexp(re *Regexp, reuse bool) *Regexp {
if re.Op == OpConcat && len(re.Sub) > 0 {
if reuse {
p.reuse(re.Sub[0])
}
re.Sub = re.Sub[:copy(re.Sub, re.Sub[1:])]
switch len(re.Sub) {
case 0:
re.Op = OpEmptyMatch
re.Sub = nil
case 1:
old := re
re = re.Sub[0]
p.reuse(old)
}
return re
}
if reuse {
p.reuse(re)
}
return p.newRegexp(OpEmptyMatch)
}
func literalRegexp(s string, flags Flags) *Regexp {
re := &Regexp{Op: OpLiteral}
re.Flags = flags
re.Rune = re.Rune0[:0] // use local storage for small strings
for _, c := range s {
if len(re.Rune) >= cap(re.Rune) {
// string is too long to fit in Rune0. let Go handle it
re.Rune = []rune(s)
break
}
re.Rune = append(re.Rune, c)
}
return re
}
// Parsing.
// Parse parses a regular expression string s, controlled by the specified
// Flags, and returns a regular expression parse tree. The syntax is
// described in the top-level comment.
func Parse(s string, flags Flags) (*Regexp, error) {
return parse(s, flags)
}
func parse(s string, flags Flags) (_ *Regexp, err error) {
defer func() {
switch r := recover(); r {
default:
panic(r)
case nil:
// ok
case ErrNestingDepth:
err = &Error{Code: ErrNestingDepth, Expr: s}
}
}()
if flags&Literal != 0 {
// Trivial parser for literal string.
if err := checkUTF8(s); err != nil {
return nil, err
}
return literalRegexp(s, flags), nil
}
// Otherwise, must do real work.
var (
p parser
c rune
op Op
lastRepeat string
)
p.flags = flags
p.wholeRegexp = s
t := s
for t != "" {
repeat := ""
BigSwitch:
switch t[0] {
default:
if c, t, err = nextRune(t); err != nil {
return nil, err
}
p.literal(c)
case '(':
if p.flags&PerlX != 0 && len(t) >= 2 && t[1] == '?' {
// Flag changes and non-capturing groups.
if t, err = p.parsePerlFlags(t); err != nil {
return nil, err
}
break
}
p.numCap++
p.op(opLeftParen).Cap = p.numCap
t = t[1:]
case '|':
if err = p.parseVerticalBar(); err != nil {
return nil, err
}
t = t[1:]
case ')':
if err = p.parseRightParen(); err != nil {
return nil, err
}
t = t[1:]
case '^':
if p.flags&OneLine != 0 {
p.op(OpBeginText)
} else {
p.op(OpBeginLine)
}
t = t[1:]
case '$':
if p.flags&OneLine != 0 {
p.op(OpEndText).Flags |= WasDollar
} else {
p.op(OpEndLine)
}
t = t[1:]
case '.':
if p.flags&DotNL != 0 {
p.op(OpAnyChar)
} else {
p.op(OpAnyCharNotNL)
}
t = t[1:]
case '[':
if t, err = p.parseClass(t); err != nil {
return nil, err
}
case '*', '+', '?':
before := t
switch t[0] {
case '*':
op = OpStar
case '+':
op = OpPlus
case '?':
op = OpQuest
}
after := t[1:]
if after, err = p.repeat(op, 0, 0, before, after, lastRepeat); err != nil {
return nil, err
}
repeat = before
t = after
case '{':
op = OpRepeat
before := t
min, max, after, ok := p.parseRepeat(t)
if !ok {
// If the repeat cannot be parsed, { is a literal.
p.literal('{')
t = t[1:]
break
}
if min < 0 || min > 1000 || max > 1000 || max >= 0 && min > max {
// Numbers were too big, or max is present and min > max.
return nil, &Error{ErrInvalidRepeatSize, before[:len(before)-len(after)]}
}
if after, err = p.repeat(op, min, max, before, after, lastRepeat); err != nil {
return nil, err
}
repeat = before
t = after
case '\\':
if p.flags&PerlX != 0 && len(t) >= 2 {
switch t[1] {
case 'A':
p.op(OpBeginText)
t = t[2:]
break BigSwitch
case 'b':
p.op(OpWordBoundary)
t = t[2:]
break BigSwitch
case 'B':
p.op(OpNoWordBoundary)
t = t[2:]
break BigSwitch
case 'C':
// any byte; not supported
return nil, &Error{ErrInvalidEscape, t[:2]}
case 'Q':
// \Q ... \E: the ... is always literals
var lit string
lit, t, _ = strings.Cut(t[2:], `\E`)
for lit != "" {
c, rest, err := nextRune(lit)
if err != nil {
return nil, err
}
p.literal(c)
lit = rest
}
break BigSwitch
case 'z':
p.op(OpEndText)
t = t[2:]
break BigSwitch
}
}
re := p.newRegexp(OpCharClass)
re.Flags = p.flags
// Look for Unicode character group like \p{Han}
if len(t) >= 2 && (t[1] == 'p' || t[1] == 'P') {
r, rest, err := p.parseUnicodeClass(t, re.Rune0[:0])
if err != nil {
return nil, err
}
if r != nil {
re.Rune = r
t = rest
p.push(re)
break BigSwitch
}
}
// Perl character class escape.
if r, rest := p.parsePerlClassEscape(t, re.Rune0[:0]); r != nil {
re.Rune = r
t = rest
p.push(re)
break BigSwitch
}
p.reuse(re)
// Ordinary single-character escape.
if c, t, err = p.parseEscape(t); err != nil {
return nil, err
}
p.literal(c)
}
lastRepeat = repeat
}
p.concat()
if p.swapVerticalBar() {
// pop vertical bar
p.stack = p.stack[:len(p.stack)-1]
}
p.alternate()
n := len(p.stack)
if n != 1 {
return nil, &Error{ErrMissingParen, s}
}
return p.stack[0], nil
}
// parseRepeat parses {min} (max=min) or {min,} (max=-1) or {min,max}.
// If s is not of that form, it returns ok == false.
// If s has the right form but the values are too big, it returns min == -1, ok == true.
func (p *parser) parseRepeat(s string) (min, max int, rest string, ok bool) {
if s == "" || s[0] != '{' {
return
}
s = s[1:]
var ok1 bool
if min, s, ok1 = p.parseInt(s); !ok1 {
return
}
if s == "" {
return
}
if s[0] != ',' {
max = min
} else {
s = s[1:]
if s == "" {
return
}
if s[0] == '}' {
max = -1
} else if max, s, ok1 = p.parseInt(s); !ok1 {
return
} else if max < 0 {
// parseInt found too big a number
min = -1
}
}
if s == "" || s[0] != '}' {
return
}
rest = s[1:]