-
Notifications
You must be signed in to change notification settings - Fork 3
/
parse.jai
2917 lines (2578 loc) · 85.2 KB
/
parse.jai
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
// This parser is based on RE2’s parser
// https://github.com/google/re2
ParseFlags :: enum_flags u16 {
NoParseFlags :: 0;
FoldCase :: 1<<0; // Fold case during matching (case-insensitive).
// Literal :: 1<<1; // Treat s as literal string instead of a regexp.
ClassNL :: 1<<2; // Allow char classes like [^a-z] and \D and \s
// and [[:space:]] to match newline.
DotNL :: 1<<3; // Allow . to match newline.
MatchNL :: ClassNL | DotNL;
OneLine :: 1<<4; // Treat ^ and $ as only matching at beginning and
// end of text, not around embedded newlines.
// (Perl's default)
// Latin1 :: 1<<5; // Regexp_Node and text are in Latin1, not UTF-8.
NonGreedy :: 1<<6; // Repetition operators are non-greedy by default.
PerlClasses :: 1<<7; // Allow Perl character classes like \d.
PerlB :: 1<<8; // Allow Perl's \b and \B.
PerlX :: 1<<9; // Perl extensions:
// non-capturing parens - (?: )
// non-greedy operators - *? +? ?? {}?
// flag edits - (?i) (?-i) (?i: )
// i - FoldCase
// m - !OneLine
// s - DotNL
// U - NonGreedy
// line ends: \A \z
// \Q and \E to disable/enable metacharacters
// (?P<name>expr) for named captures
// \C to match any single byte
UnicodeGroups :: 1<<10; // Allow \p{Han} for Unicode Han group
// and \P{Han} for its negation.
NeverNL :: 1<<11; // Never match NL, even if the regexp mentions
// it explicitly.
NeverCapture :: 1<<12; // Parse all parens as non-capturing.
// As close to Perl as we can get.
LikePerl :: ClassNL | OneLine | PerlClasses | PerlB | PerlX |
UnicodeGroups;
// Internal use only.
WasDollar :: 1<<13; // on EndText: was $ in regexp text
AllParseFlags :: (1<<14)-1;
};
Regexp_Node :: struct {
// Operator.
op: RegexpOp;
// Is this regexp structure already simple
// (has it been returned by Simplify)?
simple: bool;
// Flags saved from parsing and used during execution.
// (Only FoldCase is used.)
parse_flags: ParseFlags;
// Subexpressions.
// Concat and Alternate handle larger numbers of subexpressions
// by building concatenation or alternation trees.
// Other routines should call Concat or Alternate instead of
// filling in sub() by hand.
MAX_SUBS :: 0xFFFF;
subs: [] *Regexp_Node;
// Extra space for parse and teardown stacks.
down: *Regexp_Node;
// Arguments to operator.
using args: union {
struct { // Repeat
max: s16;
min: s16;
}
struct { // Capture
cap: s32;
name: string;
}
runes: [..] Rune; // LiteralString
struct { // CharClass
char_class: *CharClass;
char_class_finished: bool;
}
rune: Rune; // Literal
// @ToDo Verify size is ok
match_id: int; // HaveMatch
// void *the_union_[2]; // as big as any other element, for memset
};
}
RegexpOp :: enum {
// @ToDo: Rename all options
// Matches no strings.
NoMatch :: 1;
// Matches empty string.
EmptyMatch;
// Matches rune_.
Literal;
// Matches runes_.
LiteralString;
// Matches concatenation of sub_[0..nsub-1].
Concat;
// Matches union of sub_[0..nsub-1].
Alternate;
// Matches sub_[0] zero or more times.
Star;
// Matches sub_[0] one or more times.
Plus;
// Matches sub_[0] zero or one times.
Quest;
// Matches sub_[0] at least min_ times, at most max_ times.
// max_ == -1 means no upper limit.
Repeat;
// Parenthesized (capturing) subexpression. Index is cap_.
// Optionally, capturing name is name_.
Capture;
// Matches any character.
AnyChar;
// Matches any byte [sic].
AnyByte;
// Matches empty string at beginning of line.
BeginLine;
// Matches empty string at end of line.
EndLine;
// Matches word boundary "\b".
WordBoundary;
// Matches not-a-word boundary "\B".
NoWordBoundary;
// Matches empty string at beginning of text.
BeginText;
// Matches empty string at end of text.
EndText;
// Matches character class given by cc_.
CharClass;
// Forces match of entire expression right now,
// with match ID match_id_ (used by RE2::Set).
HaveMatch;
// The ones below are only used internally by the parser and don’t occur in finshed RegExps
kLeftParen;
kVerticalBar;
};
Regexp_Pool :: struct {
pool: *Pool;
bucket: Bucket_Array(Regexp_Node, 64);
}
init :: (pool: *Regexp_Pool) {
pool.pool = New(Pool);
set_allocators(pool.pool);
#if REGEXP_DEBUG pool.pool.overwrite_memory = true;
pool.bucket.allocator.proc = pool_allocator_proc;
pool.bucket.allocator.data = pool.pool;
}
uninit :: (pool: *Regexp_Pool) {
release(pool.pool);
free(pool.pool);
}
#scope_module
new_regexp :: (pool: *Regexp_Pool, op: RegexpOp, flags: ParseFlags) -> *Regexp_Node {
re := find_and_occupy_empty_slot(*pool.bucket);
// Need to clear, in case we re-use a bucket slot
// @ToDo, @Speed: This is unnecessary when its a fresh slot. Find a better way
memset(re, 0, size_of(Regexp_Node));
re.op = op;
re.parse_flags = flags;
return re;
}
remove_regexp :: (pool: *Regexp_Pool, re: *Regexp_Node) {
// @ToDo: Delete subs!
// @ToDO: Delete char_class!
for * pool.bucket {
if it == re {
remove it;
break;
}
}
}
// Tests equality of all top-level structure but not subregexps.
regexp_top_equal :: (a: *Regexp_Node, b: *Regexp_Node) -> bool {
if a.op != b.op return false;
if a.op == {
case .NoMatch; #through;
case .EmptyMatch; #through;
case .AnyChar; #through;
case .AnyByte; #through;
case .BeginLine; #through;
case .EndLine; #through;
case .WordBoundary; #through;
case .NoWordBoundary; #through;
case .BeginText;
return true;
case .EndText;
// The parse flags remember whether it's \z or (?-m:$),
// which matters when testing against PCRE.
return ((a.parse_flags ^ b.parse_flags) & .WasDollar) == 0;
case .Literal;
return a.rune == b.rune && ((a.parse_flags ^ b.parse_flags) & .FoldCase) == 0;
case .LiteralString;
return a.runes.count == b.runes.count && ((a.parse_flags ^ b.parse_flags) & .FoldCase) == 0 &&
memcmp(a.runes.data, b.runes.data, a.runes.count * size_of(Rune)) == 0;
case .Alternate; #through;
case .Concat;
return a.subs.count == b.subs.count;
case .Star; #through;
case .Plus; #through;
case .Quest;
return ((a.parse_flags ^ b.parse_flags) & .NonGreedy) == 0;
case .Repeat;
return ((a.parse_flags ^ b.parse_flags) & .NonGreedy) == 0 && a.min == b.min && a.max == b.max;
case .Capture;
return a.cap == b.cap && a.name == b.name;
case .HaveMatch;
return a.match_id == b.match_id;
case .CharClass;
return a.char_class.nrunes == b.char_class.nrunes &&
a.char_class.ranges.count == b.char_class.ranges.count &&
memcmp(a.char_class.ranges.data, b.char_class.ranges.data, a.char_class.ranges.count * size_of(RuneRange)) == 0;
case;
assert(false, "Unexpected op: %", a.op);
return false;
}
}
regexp_equal :: (a: *Regexp_Node, b: *Regexp_Node) -> bool {
if a == null || b == null return a == b;
if !regexp_top_equal(a, b) return false;
// The stack has pairs of regexps waiting to
// be compared. The regexps are only equal if
// all the pairs end up being equal.
stack: [..] *Regexp_Node;
stack.allocator = __temporary_allocator;
a1 := a;
b1 := b;
while true {
// Invariant: TopEqual(a1, b1) == true.
a2: *Regexp_Node;
b2: *Regexp_Node;
if a1.op == {
case .Alternate; #through;
case .Concat;
for i: 0..a1.subs.count-1 {
a2 = a1.subs[i];
b2 = b1.subs[i];
if !regexp_top_equal(a2, b2) return false;
array_add(*stack, a2);
array_add(*stack, b2);
}
case .Star; #through;
case .Plus; #through;
case .Quest; #through;
case .Repeat; #through;
case .Capture;
a2 = a1.subs[0];
b2 = b1.subs[0];
if !regexp_top_equal(a2, b2) return false;
// Really:
// stack.push_back(a2);
// stack.push_back(b2);
// break;
// but faster to assign directly and loop.
a1 = a2;
b1 = b2;
continue;
case;
return true;
}
if !stack.count return true;
a1 = stack[stack.count-2];
b1 = stack[stack.count-1];
stack.count -= 2;
}
return true;
}
// Determines whether regexp matches must be anchored
// with a fixed string prefix. If so, returns the prefix and
// the regexp that remains after the prefix. The prefix might
// be ASCII case-insensitive.
required_prefix :: (re: *Regexp_Node, pool: *Regexp_Pool) -> prefix: string, foldcase: bool, suffix: *Regexp_Node {
// No need for a walker: the regexp must be of the form
// 1. some number of ^ anchors
// 2. a literal char or string
// 3. the rest
suffix := re;
if re.op != .Concat return "", false, suffix;
i := 0;
while i < re.subs.count && re.subs[i].op == .BeginText {
i += 1;
}
if i == 0 || i >= re.subs.count return "", false, suffix;
to_analyze := re.subs[i];
if to_analyze.op != .Literal && to_analyze.op != .LiteralString return "", false, suffix;
i += 1;
if (i < re.subs.count) {
suffix = new_concat(pool, array_view(re.subs, i, re.subs.count - i), re.parse_flags);
} else {
suffix = new_regexp(pool, .EmptyMatch, re.parse_flags);
}
runes: [] Rune;
if to_analyze.op == .Literal {
runes.data = *to_analyze.rune;
runes.count = 1;
} else {
runes = to_analyze.runes;
}
{
push_allocator(pool_allocator_proc, pool.pool);
prefix := string_from_runes(runes);
return prefix, (to_analyze.parse_flags & .FoldCase) != 0, suffix;
}
}
required_prefix_for_accel :: (re: Regexp_Node) -> prefix: string, foldcase: bool {
// No need for a walker: the regexp must either begin with or be
// a literal char or string.
to_analyze := *re;
if re.op == .Concat && re.subs.count > 0 {
to_analyze = re.subs[0];
}
if to_analyze.op != .Literal && to_analyze.op != .LiteralString return "", false;
runes: [] Rune;
if to_analyze.op == .Literal {
runes.data = *to_analyze.rune;
runes.count = 1;
} else {
runes = to_analyze.runes;
}
prefix := string_from_runes(runes);
return prefix, (to_analyze.parse_flags & .FoldCase) != 0;
}
// @ToDo: rename these
StatusCode :: enum {
// No error
Success :: 0;
// Unexpected error
InternalError;
// Parse errors
BadEscape; // bad escape sequence
BadCharClass; // bad character class
BadCharRange; // bad character class range
MissingBracket; // missing closing ]
MissingParen; // missing closing )
UnexpectedParen; // unexpected closing )
TrailingBackslash; // at end of regexp
RepeatArgument; // repeat argument missing, e.g. "*"
RepeatSize; // bad repetition argument
RepeatOp; // bad repetition operator
BadPerlOp; // bad perl operator
BadUTF8; // invalid UTF-8 in regexp
BadNamedCapture; // bad named capture
};
// Regular expression parse state.
// The list of parsed regexps so far is maintained as a vector of
// Regexp_Node pointers called the stack. Left parenthesis and vertical
// bar markers are also placed on the stack, as Regexps with
// non-standard opcodes.
// Scanning a left parenthesis causes the parser to push a left parenthesis
// marker on the stack.
// Scanning a vertical bar causes the parser to pop the stack until it finds a
// vertical bar or left parenthesis marker (not popping the marker),
// concatenate all the popped results, and push them back on
// the stack (DoConcatenation).
// Scanning a right parenthesis causes the parser to act as though it
// has seen a vertical bar, which then leaves the top of the stack in the
// form LeftParen regexp VerticalBar regexp VerticalBar ... regexp VerticalBar.
// The parser pops all this off the stack and creates an alternation of the
// regexps (DoAlternation).
ParseState :: struct {
flags: ParseFlags;
whole_regexp: string;
status: Status;
stacktop: *Regexp_Node;
ncap: s32; // number of capturing parens seen
rune_max := Runemax; // maximum char value for this encoding
pool: Regexp_Pool;
}
Status :: struct {
code: StatusCode;
error_arg: string;
}
make_status :: (code := StatusCode.Success, error_arg := "") -> Status {
status: Status;
status.code = code;
status.error_arg = error_arg;
return status;
}
rep :: (ps: *ParseState, op: RegexpOp, t: *string, lastunary: string) -> isunary: string, success: bool {
opstr := <<t;
nongreedy := false;
advance(t); // '*' or '+' or '?'
if (ps.flags & .PerlX) {
if (t.count && (<<t)[0] == #char "?") {
nongreedy = true;
advance(t); // '?'
}
if (lastunary) {
// In Perl it is not allowed to stack repetition operators:
// a** is a syntax error, not a double-star.
// (and a++ means something else entirely, which we don't support!)
ps.status.code = .RepeatOp;
ps.status.error_arg.data = lastunary.data;
ps.status.error_arg.count = t.data - lastunary.data;
return "", false;
}
}
opstr.count -= t.count;
if !push_repeat_op(ps, op, opstr, nongreedy)
return "", false;
return opstr, true;
}
// Parses the regular expression given by s,
// returning the corresponding Regexp_Node tree.
// Returns null on error.
parse :: (s: string, global_flags: ParseFlags = .NoParseFlags) -> *Regexp_Node, status: Status, pool: Regexp_Pool {
ps: ParseState;
ps.whole_regexp = s;
ps.flags = global_flags;
init(*ps.pool);
push_allocator(pool_allocator_proc, ps.pool.pool);
lastunary: string;
t := s;
while t.count {
isunary: string;
if t[0] == {
case; {
r, len, status := consume_rune(*t);
if !len return null, status, ps.pool;
push_literal(*ps, r);
}
case #char "(";
// "(?" introduces Perl escape.
if ps.flags & .PerlX && t.count >= 2 && t[1] == #char "?" {
// Flag changes and non-capturing groups.
success: bool;
ps.status, success = parse_perl_flags(*ps, *t);
if !success return null, ps.status, ps.pool;
} else if ps.flags & .NeverCapture {
do_left_paren_no_capture(*ps);
advance(*t); // '('
} else {
do_left_paren(*ps);
advance(*t); // '('
}
case #char "|";
do_vertical_bar(*ps);
advance(*t); // '|'
case #char ")";
if !do_right_paren(*ps) return null, ps.status, ps.pool;
advance(*t); // ')'
case #char "^"; // Beginning of line.
push_caret(*ps);
advance(*t); // '^'
case #char "$"; // End of line.
push_dollar(*ps);
advance(*t); // '$'
case #char "."; // Any character (possibly except newline).
push_dot(*ps);
advance(*t); // '.'
case #char "["; // Character class.
re, status, success := parse_char_class(*ps, *t);
if !success return null, status, ps.pool;
push_regexp(*ps, re);
case #char "*"; // Zero or more.
success: bool;
isunary, success = rep(*ps, .Star, *t, lastunary);
if !success return null, ps.status, ps.pool;
case #char "+"; // One or more.
success: bool;
isunary, success = rep(*ps, .Plus, *t, lastunary);
if !success return null, ps.status, ps.pool;
case #char "?"; // Zero or one.
success: bool;
isunary, success = rep(*ps, .Quest, *t, lastunary);
if !success return null, ps.status, ps.pool;
case #char "{"; // Counted repetition.
opstr := t;
lo, hi, success := maybe_parse_repetition(*t);
if !success {
// Treat like a literal.
push_literal(*ps, #char "{");
advance(*t); // '{'
} else {
nongreedy := false;
if ps.flags & .PerlX {
if t && t[0] == #char "?" {
nongreedy = true;
advance(*t); // '?'
}
if lastunary {
// Not allowed to stack repetition operators.
ps.status.code = .RepeatOp;
ps.status.error_arg = slice(lastunary, 0, t.data - lastunary.data);
return null, ps.status, ps.pool;
}
}
opstr.count -= t.count;
if !push_repetition(*ps, lo, hi, opstr, nongreedy) {
return null, ps.status, ps.pool;
}
isunary = opstr;
}
case #char "\\"; // Escaped character or Perl sequence.
// \b and \B: word boundary or not
parsed := false;
if ps.flags & .PerlB &&
t.count >= 2 && (t[1] == #char "b" || t[1] == #char "B") {
if t[1] == #char "b" {
push_simple_op(*ps, .WordBoundary);
} else {
push_simple_op(*ps, .NoWordBoundary);
}
advance(*t, 2); // '\\', 'b'
parsed = true;
} else if (ps.flags & .PerlX) && t.count >= 2 {
if (t[1] == #char "A") {
push_simple_op(*ps, .BeginText);
advance(*t, 2); // '\\', 'A'
parsed = true;
} else if (t[1] == #char "z") {
push_simple_op(*ps, .EndText);
advance(*t, 2); // '\\', 'z'
parsed = true;
// Do not recognize \Z, because this library can't
// implement the exact Perl/PCRE semantics.
// (This library treats "(?-m)$" as \z, even though
// in Perl and PCRE it is equivalent to \Z.)
} else if (t[1] == #char "C") { // \C: any byte [sic]
push_simple_op(*ps, .AnyByte);
advance(*t, 2); // '\\', 'C'
parsed = true;
} else if (t[1] == #char "Q") { // \Q ... \E: the ... is always literals
advance(*t, 2); // '\\', 'Q'
while t.count {
if (t.count >= 2 && t[0] == #char "\\" && t[1] == #char "E") {
advance(*t, 2); // '\\', 'E'
parsed = true;
}
r, len, status := consume_rune(*t);
if !len return null, status, ps.pool;
push_literal(*ps, r);
}
parsed = true;
}
}
if (!parsed && t.count >= 2 && (t[1] == #char "p" || t[1] == #char "P")) {
re := new_regexp(*ps.pool, .CharClass, ps.flags & ~.FoldCase);
re.char_class = New(CharClass);
re.char_class_finished = false;
parseStatus, parseResult := maybe_parse_unicode_group(*t, ps.flags, re.char_class);
if parseResult == {
case .kParseOk;
push_regexp(*ps, re);
lastunary = isunary;
parsed = true;
case .kParseError;
remove_regexp(*ps.pool, re);
ps.status = parseStatus;
return null, ps.status, ps.pool;
case .kParseNothing;
remove_regexp(*ps.pool, re);
}
}
if !parsed {
g := maybe_parse_perl_cc_escape(*t, ps.flags);
if g != null {
re := new_regexp(*ps.pool, .CharClass, ps.flags & ~.FoldCase);
re.char_class = New(CharClass);
re.char_class_finished = false;
add_unicode_group(re.char_class, g, g.sign, ps.flags);
push_regexp(*ps, re);
parsed = true;
}
}
if !parsed {
r, status, success := parse_escape(*t, ps.rune_max);
if !success return null, status, ps.pool;
push_literal(*ps, r);
}
}
}
return do_finish(*ps), ps.status, ps.pool;
}
new_literal_string :: (pool: *Regexp_Pool, runes: [] Rune, flags: ParseFlags) -> *Regexp_Node {
if runes.count == {
case 0;
return new_regexp(pool, .EmptyMatch, flags);
case 1;
return new_literal(pool, runes[0], flags);
case;
re := new_regexp(pool, .LiteralString, flags);
for runes {
add_rune_to_string(re, it);
}
return re;
}
}
new_literal :: (pool: *Regexp_Pool, rune: Rune, flags: ParseFlags) -> *Regexp_Node {
re := new_regexp(pool, .Literal, flags);
re.rune = rune;
return re;
}
new_star_plus_quest :: (pool: *Regexp_Pool, op: RegexpOp, sub: *Regexp_Node, flags: ParseFlags) -> *Regexp_Node {
// Squash **, ++ and ??.
if (op == sub.op && flags == sub.parse_flags) return sub;
// Squash *+, *?, +*, +?, ?* and ?+. They all squash to *, so because
// op is Star/Plus/Quest, we just have to check that sub.op() is too.
if ((sub.op == .Star ||
sub.op == .Plus ||
sub.op == .Quest) &&
flags == sub.parse_flags) {
// If sub is Star, no need to rewrite it.
if (sub.op == .Star) return sub;
re := new_regexp(pool, .Star, flags);
array_resize(*re.subs, 1, false);
re.subs[0] = sub.subs[0];
return re;
}
re := new_regexp(pool, op, flags);
array_resize(*re.subs, 1, false);
re.subs[0] = sub;
return re;
}
// @ToDo: Why does #bake_arguments not work here? (Compiler bug I can’t find a simple repro for)
// new_plus :: #bake_arguments new_star_plus_quest(op = .Plus);
// new_star :: #bake_arguments new_star_plus_quest(op = RegexpOp.Star);
// new_quest :: #bake_arguments new_star_plus_quest(op = .Quest);
new_plus:: (pool: *Regexp_Pool, sub: *Regexp_Node, flags: ParseFlags) -> *Regexp_Node {
return new_star_plus_quest(pool, .Plus, sub, flags);
}
new_star :: (pool: *Regexp_Pool, sub: *Regexp_Node, flags: ParseFlags) -> *Regexp_Node {
return new_star_plus_quest(pool, .Star, sub, flags);
}
new_quest :: (pool: *Regexp_Pool, sub: *Regexp_Node, flags: ParseFlags) -> *Regexp_Node {
return new_star_plus_quest(pool, .Quest, sub, flags);
}
new_concat_or_alternate :: (pool: *Regexp_Pool, op: RegexpOp, subs: [] *Regexp_Node, flags: ParseFlags, can_factor: bool) -> *Regexp_Node {
if subs.count == 1 {
sub := subs[0];
return sub;
}
if subs.count == 0 {
if op == .Alternate {
return new_regexp(pool, .NoMatch, flags);
} else {
return new_regexp(pool, .EmptyMatch, flags);
}
}
// @ToDo, @Incomplete: We need this factorization for selecting the right
// execution engine later. But does it matter if we implement only NDA for now?
// PODArray<Regexp_Node*> subcopy;
// if op == .Alternate && can_factor {
// // Going to edit sub; make a copy so we don't step on caller.
// subcopy = PODArray<Regexp_Node*>(nsub);
// memmove(subcopy.data(), sub, nsub * sizeof sub[0]);
// sub = subcopy.data();
// nsub = FactorAlternation(sub, nsub, flags);
// if (nsub == 1) {
// Regexp_Node* re = sub[0];
// return re;
// }
// }
// @ToDo, @Incomplete: Is this necessary for us?
// We use a normal array at the moment, so we waste 6 more bytes on the size and won’t overflow
// if (nsub > MAX_SUBS) {
// // Too many subexpressions to fit in a single Regexp_Node.
// // Make a two-level tree. Two levels gets us to 65535^2.
// int nbigsub = (nsub+MAX_SUBS-1)/MAX_SUBS;
// Regexp_Node* re = new Regexp_Node(op, flags);
// re->AllocSub(nbigsub);
// Regexp_Node** subs = re->sub();
// for (int i = 0; i < nbigsub - 1; i++)
// subs[i] = ConcatOrAlternate(op, sub+i*MAX_SUBS, MAX_SUBS, flags, false);
// subs[nbigsub - 1] = ConcatOrAlternate(op, sub+(nbigsub-1)*MAX_SUBS,
// nsub - (nbigsub-1)*MAX_SUBS, flags,
// false);
// return re;
// }
re := new_regexp(pool, op, flags);
re.subs = subs;
return re;
}
// concat :: #bake_arguments concat_or_alternate(op = .Concat, can_factor = false);
// alternate :: #bake_arguments concat_or_alternate(op = .Alternate, can_factor = true);
// alternate_no_factor :: #bake_arguments concat_or_alternate(op = .Alternate, can_factor = false);
new_concat :: (pool: *Regexp_Pool, subs: [] *Regexp_Node, flags: ParseFlags) -> *Regexp_Node {
return new_concat_or_alternate(pool, .Concat, subs, flags, false);
}
new_alternate :: (pool: *Regexp_Pool, subs: [] *Regexp_Node, flags: ParseFlags) -> *Regexp_Node {
return new_concat_or_alternate(pool, .Alternate, subs, flags, true);
}
new_alternate_no_factor :: (pool: *Regexp_Pool, subs: [] *Regexp_Node, flags: ParseFlags) -> *Regexp_Node {
return new_concat_or_alternate(pool, .Alternate, subs, flags, false);
}
new_capture :: (pool: *Regexp_Pool, sub: *Regexp_Node, flags: ParseFlags, cap: s32) -> *Regexp_Node {
re := new_regexp(pool, .Capture, flags);
array_resize(*re.subs, 1, false);
re.subs[0] = sub;
re.cap = cap;
return re;
}
new_repeat :: (pool: *Regexp_Pool, sub: *Regexp_Node, flags: ParseFlags, min: s16, max: s16) -> *Regexp_Node {
re := new_regexp(pool, .Repeat, flags);
array_resize(*re.subs, 1, false);
re.subs[0] = sub;
re.min = min;
re.max = max;
return re;
}
to_string :: (re: *Regexp_Node) -> string {
// Appends a rune for use in a character class to the string t.
append_cc_char :: (builder: *String_Builder, r: Rune) {
if 0x20 <= r && r <= 0x7E {
if find_index_from_left("[]^-\\", cast(u8)r) != -1 {
append(builder, "\\");
}
append(builder, cast(u8)r);
} else {
if r == {
case #char "\r";
append(builder, "\\r");
return;
case #char "\t";
append(builder, "\\t");
return;
case #char "\n";
append(builder, "\\n");
return;
case 0x0c;
append(builder, "\\f");
return;
}
if (r < 0x100) {
print_to_builder(builder, "\\x%", formatInt(r, base = 16, minimum_digits = 2));
} else {
print_to_builder(builder, "\\x{%}", formatInt(r, base = 16));
}
}
}
append_cc_range :: (builder: *String_Builder, lo: Rune, hi: Rune) {
append_cc_char(builder, lo);
// < would be enough here, but we want to print invalid ranges as well, just in case we have a bug somewhere
if lo != hi {
append(builder, #char "-");
append_cc_char(builder, hi);
}
}
append_literal :: (builder: *String_Builder, r: Rune, foldcase: bool) {
if r != 0 && r < 0x80 && find_index_from_left("(){}[]*+?|.^$\\", cast(u8)r) != -1 {
append(builder, #char "\\");
append(builder, cast(u8)r);
} else if (foldcase && #char "a" <= r && r <= #char "z") {
r -= #char "a" - #char "A";
append(builder, #char "[");
append(builder, cast(u8)r);
append(builder, cast(u8)r + #char "a" - #char "A");
append(builder, #char "]");
} else {
append_cc_range(builder, r, r);
}
}
// Appends ( if needed and passes new precedence to children.
to_string_pre :: (builder: *String_Builder, re: *Regexp_Node, prec: Precedence) -> Precedence, bool {
nprec := Precedence.Atom;
if re.op == {
case .NoMatch; #through;
case .EmptyMatch; #through;
case .Literal; #through;
case .AnyChar; #through;
case .AnyByte; #through;
case .BeginLine; #through;
case .EndLine; #through;
case .BeginText; #through;
case .EndText; #through;
case .WordBoundary; #through;
case .NoWordBoundary; #through;
case .CharClass; #through;
case .HaveMatch;
// Atom is fine
case .Concat; #through;
case .LiteralString;
if prec < .Concat {
append(builder, "(?:");
}
nprec = .Concat;
case .Alternate;
if prec < .Alternate {
append(builder, "(?:");
}
nprec = .Alternate;
case .Capture;
append(builder, #char "(");
assert (re.cap != 0, "Capture 0");
if re.name {
append(builder, "?P<");
append(builder, re.name);
append(builder, #char ">");
}
nprec = .Paren;
case .Star; #through;
case .Plus; #through;
case .Quest; #through;
case .Repeat;
if (prec < .Unary) {
append(builder, "(?:");
}
// The subprecedence here is Atom instead of Unary
// because PCRE treats two unary ops in a row as a parse error.
nprec = .Atom;
}
return nprec, false;
}
to_string_post :: (builder: *String_Builder, re: *Regexp_Node, prec: Precedence, pre_arg: Precedence, child_args: [..] Precedence) -> Precedence {
if re.op == {
case .NoMatch;
// There's no simple symbol for "no match", but
// [^0-Runemax] excludes everything.
append(builder, "[^\\x00-\\x{10ffff}]");
case .EmptyMatch;
// Append (?:) to make empty string visible,
// unless this is already being parenthesized.
if prec < .Empty {
append(builder, "(?:)");
}
case .Literal;
append_literal(builder, re.rune, (re.parse_flags & .FoldCase) != 0);
case .LiteralString;
for re.runes {
append_literal(builder, it, (re.parse_flags & .FoldCase) != 0);
}
if prec < .Concat {
append(builder, #char ")");
}
case .Concat;
if prec < .Concat {
append(builder, #char ")");
}
case .Alternate;
// Clumsy but workable: the children all appended |
// at the end of their strings, so just remove the last one.
// @ToDo: we can’t do that with string_builder, at least not without very clumsy or very slow code
// if ((*t_)[t_->size()-1] == '|')
// t_->erase(t_->size()-1);
// else
// LOG(DFATAL) << "Bad final char: " << t_;
if prec < .Alternate {
append(builder, #char ")");
}
case .Star;
append(builder, #char "*");
if re.parse_flags & .NonGreedy {
append(builder, #char "?");
}
if (prec < .Unary) {
append(builder, #char ")");
}
case .Plus;
append(builder, #char "+");
if re.parse_flags & .NonGreedy {
append(builder, #char "?");
}
if prec < .Unary {
append(builder, #char ")");
}
case .Quest;
append(builder, #char "?");
if re.parse_flags & .NonGreedy {
append(builder, #char "?");
}
if (prec < .Unary) {
append(builder, #char ")");
}
case .Repeat;
if (re.max == -1) {
print_to_builder(builder, "{%,}", re.min);
} else if re.min == re.max {
print_to_builder(builder, "{%}", re.min);
} else {
print_to_builder(builder, "{%,%}", re.min, re.max);
}
if (re.parse_flags & .NonGreedy) {
append(builder, #char "?");
}
if (prec < .Unary) {
append(builder, #char ")");
}
case .AnyChar;
append(builder, #char ".");