1 /**********************************************************************
2
3 re.c -
4
5 $Author: nagachika $
6 created at: Mon Aug 9 18:24:49 JST 1993
7
8 Copyright (C) 1993-2007 Yukihiro Matsumoto
9
10 **********************************************************************/
11
12 #include "ruby/encoding.h"
13 #include "ruby/re.h"
14 #include "ruby/util.h"
15 #include "internal.h"
16 #include "regint.h"
17 #include "encindex.h"
18 #include <ctype.h>
19
20 VALUE rb_eRegexpError;
21
22 typedef char onig_errmsg_buffer[ONIG_MAX_ERROR_MESSAGE_LEN];
23 #define errcpy(err, msg) strlcpy((err), (msg), ONIG_MAX_ERROR_MESSAGE_LEN)
24
25 #define BEG(no) (regs->beg[(no)])
26 #define END(no) (regs->end[(no)])
27
28 #if 'a' == 97 /* it's ascii */
29 static const char casetable[] = {
30 '\000', '\001', '\002', '\003', '\004', '\005', '\006', '\007',
31 '\010', '\011', '\012', '\013', '\014', '\015', '\016', '\017',
32 '\020', '\021', '\022', '\023', '\024', '\025', '\026', '\027',
33 '\030', '\031', '\032', '\033', '\034', '\035', '\036', '\037',
34 /* ' ' '!' '"' '#' '$' '%' '&' ''' */
35 '\040', '\041', '\042', '\043', '\044', '\045', '\046', '\047',
36 /* '(' ')' '*' '+' ',' '-' '.' '/' */
37 '\050', '\051', '\052', '\053', '\054', '\055', '\056', '\057',
38 /* '0' '1' '2' '3' '4' '5' '6' '7' */
39 '\060', '\061', '\062', '\063', '\064', '\065', '\066', '\067',
40 /* '8' '9' ':' ';' '<' '=' '>' '?' */
41 '\070', '\071', '\072', '\073', '\074', '\075', '\076', '\077',
42 /* '@' 'A' 'B' 'C' 'D' 'E' 'F' 'G' */
43 '\100', '\141', '\142', '\143', '\144', '\145', '\146', '\147',
44 /* 'H' 'I' 'J' 'K' 'L' 'M' 'N' 'O' */
45 '\150', '\151', '\152', '\153', '\154', '\155', '\156', '\157',
46 /* 'P' 'Q' 'R' 'S' 'T' 'U' 'V' 'W' */
47 '\160', '\161', '\162', '\163', '\164', '\165', '\166', '\167',
48 /* 'X' 'Y' 'Z' '[' '\' ']' '^' '_' */
49 '\170', '\171', '\172', '\133', '\134', '\135', '\136', '\137',
50 /* '`' 'a' 'b' 'c' 'd' 'e' 'f' 'g' */
51 '\140', '\141', '\142', '\143', '\144', '\145', '\146', '\147',
52 /* 'h' 'i' 'j' 'k' 'l' 'm' 'n' 'o' */
53 '\150', '\151', '\152', '\153', '\154', '\155', '\156', '\157',
54 /* 'p' 'q' 'r' 's' 't' 'u' 'v' 'w' */
55 '\160', '\161', '\162', '\163', '\164', '\165', '\166', '\167',
56 /* 'x' 'y' 'z' '{' '|' '}' '~' */
57 '\170', '\171', '\172', '\173', '\174', '\175', '\176', '\177',
58 '\200', '\201', '\202', '\203', '\204', '\205', '\206', '\207',
59 '\210', '\211', '\212', '\213', '\214', '\215', '\216', '\217',
60 '\220', '\221', '\222', '\223', '\224', '\225', '\226', '\227',
61 '\230', '\231', '\232', '\233', '\234', '\235', '\236', '\237',
62 '\240', '\241', '\242', '\243', '\244', '\245', '\246', '\247',
63 '\250', '\251', '\252', '\253', '\254', '\255', '\256', '\257',
64 '\260', '\261', '\262', '\263', '\264', '\265', '\266', '\267',
65 '\270', '\271', '\272', '\273', '\274', '\275', '\276', '\277',
66 '\300', '\301', '\302', '\303', '\304', '\305', '\306', '\307',
67 '\310', '\311', '\312', '\313', '\314', '\315', '\316', '\317',
68 '\320', '\321', '\322', '\323', '\324', '\325', '\326', '\327',
69 '\330', '\331', '\332', '\333', '\334', '\335', '\336', '\337',
70 '\340', '\341', '\342', '\343', '\344', '\345', '\346', '\347',
71 '\350', '\351', '\352', '\353', '\354', '\355', '\356', '\357',
72 '\360', '\361', '\362', '\363', '\364', '\365', '\366', '\367',
73 '\370', '\371', '\372', '\373', '\374', '\375', '\376', '\377',
74 };
75 #else
76 # error >>> "You lose. You will need a translation table for your character set." <<<
77 #endif
78
79 int
rb_memcicmp(const void * x,const void * y,long len)80 rb_memcicmp(const void *x, const void *y, long len)
81 {
82 const unsigned char *p1 = x, *p2 = y;
83 int tmp;
84
85 while (len--) {
86 if ((tmp = casetable[(unsigned)*p1++] - casetable[(unsigned)*p2++]))
87 return tmp;
88 }
89 return 0;
90 }
91
92 #ifdef HAVE_MEMMEM
93 static inline long
rb_memsearch_ss(const unsigned char * xs,long m,const unsigned char * ys,long n)94 rb_memsearch_ss(const unsigned char *xs, long m, const unsigned char *ys, long n)
95 {
96 const unsigned char *y;
97
98 if ((y = memmem(ys, n, xs, m)) != NULL)
99 return y - ys;
100 else
101 return -1;
102 }
103 #else
104 static inline long
rb_memsearch_ss(const unsigned char * xs,long m,const unsigned char * ys,long n)105 rb_memsearch_ss(const unsigned char *xs, long m, const unsigned char *ys, long n)
106 {
107 const unsigned char *x = xs, *xe = xs + m;
108 const unsigned char *y = ys, *ye = ys + n;
109 #define VALUE_MAX ((VALUE)~(VALUE)0)
110 VALUE hx, hy, mask = VALUE_MAX >> ((SIZEOF_VALUE - m) * CHAR_BIT);
111
112 if (m > SIZEOF_VALUE)
113 rb_bug("!!too long pattern string!!");
114
115 if (!(y = memchr(y, *x, n - m + 1)))
116 return -1;
117
118 /* Prepare hash value */
119 for (hx = *x++, hy = *y++; x < xe; ++x, ++y) {
120 hx <<= CHAR_BIT;
121 hy <<= CHAR_BIT;
122 hx |= *x;
123 hy |= *y;
124 }
125 /* Searching */
126 while (hx != hy) {
127 if (y == ye)
128 return -1;
129 hy <<= CHAR_BIT;
130 hy |= *y;
131 hy &= mask;
132 y++;
133 }
134 return y - ys - m;
135 }
136 #endif
137
138 static inline long
rb_memsearch_qs(const unsigned char * xs,long m,const unsigned char * ys,long n)139 rb_memsearch_qs(const unsigned char *xs, long m, const unsigned char *ys, long n)
140 {
141 const unsigned char *x = xs, *xe = xs + m;
142 const unsigned char *y = ys;
143 VALUE i, qstable[256];
144
145 /* Preprocessing */
146 for (i = 0; i < 256; ++i)
147 qstable[i] = m + 1;
148 for (; x < xe; ++x)
149 qstable[*x] = xe - x;
150 /* Searching */
151 for (; y + m <= ys + n; y += *(qstable + y[m])) {
152 if (*xs == *y && memcmp(xs, y, m) == 0)
153 return y - ys;
154 }
155 return -1;
156 }
157
158 static inline unsigned int
rb_memsearch_qs_utf8_hash(const unsigned char * x)159 rb_memsearch_qs_utf8_hash(const unsigned char *x)
160 {
161 register const unsigned int mix = 8353;
162 register unsigned int h = *x;
163 if (h < 0xC0) {
164 return h + 256;
165 }
166 else if (h < 0xE0) {
167 h *= mix;
168 h += x[1];
169 }
170 else if (h < 0xF0) {
171 h *= mix;
172 h += x[1];
173 h *= mix;
174 h += x[2];
175 }
176 else if (h < 0xF5) {
177 h *= mix;
178 h += x[1];
179 h *= mix;
180 h += x[2];
181 h *= mix;
182 h += x[3];
183 }
184 else {
185 return h + 256;
186 }
187 return (unsigned char)h;
188 }
189
190 static inline long
rb_memsearch_qs_utf8(const unsigned char * xs,long m,const unsigned char * ys,long n)191 rb_memsearch_qs_utf8(const unsigned char *xs, long m, const unsigned char *ys, long n)
192 {
193 const unsigned char *x = xs, *xe = xs + m;
194 const unsigned char *y = ys;
195 VALUE i, qstable[512];
196
197 /* Preprocessing */
198 for (i = 0; i < 512; ++i) {
199 qstable[i] = m + 1;
200 }
201 for (; x < xe; ++x) {
202 qstable[rb_memsearch_qs_utf8_hash(x)] = xe - x;
203 }
204 /* Searching */
205 for (; y + m <= ys + n; y += qstable[rb_memsearch_qs_utf8_hash(y+m)]) {
206 if (*xs == *y && memcmp(xs, y, m) == 0)
207 return y - ys;
208 }
209 return -1;
210 }
211
212 static inline long
rb_memsearch_wchar(const unsigned char * xs,long m,const unsigned char * ys,long n)213 rb_memsearch_wchar(const unsigned char *xs, long m, const unsigned char *ys, long n)
214 {
215 const unsigned char *x = xs, x0 = *xs, *y = ys;
216 enum {char_size = 2};
217
218 for (n -= m; n >= 0; n -= char_size, y += char_size) {
219 if (x0 == *y && memcmp(x+1, y+1, m-1) == 0)
220 return y - ys;
221 }
222 return -1;
223 }
224
225 static inline long
rb_memsearch_qchar(const unsigned char * xs,long m,const unsigned char * ys,long n)226 rb_memsearch_qchar(const unsigned char *xs, long m, const unsigned char *ys, long n)
227 {
228 const unsigned char *x = xs, x0 = *xs, *y = ys;
229 enum {char_size = 4};
230
231 for (n -= m; n >= 0; n -= char_size, y += char_size) {
232 if (x0 == *y && memcmp(x+1, y+1, m-1) == 0)
233 return y - ys;
234 }
235 return -1;
236 }
237
238 long
rb_memsearch(const void * x0,long m,const void * y0,long n,rb_encoding * enc)239 rb_memsearch(const void *x0, long m, const void *y0, long n, rb_encoding *enc)
240 {
241 const unsigned char *x = x0, *y = y0;
242
243 if (m > n) return -1;
244 else if (m == n) {
245 return memcmp(x0, y0, m) == 0 ? 0 : -1;
246 }
247 else if (m < 1) {
248 return 0;
249 }
250 else if (m == 1) {
251 const unsigned char *ys = memchr(y, *x, n);
252
253 if (ys)
254 return ys - y;
255 else
256 return -1;
257 }
258 else if (LIKELY(rb_enc_mbminlen(enc) == 1)) {
259 if (m <= SIZEOF_VALUE) {
260 return rb_memsearch_ss(x0, m, y0, n);
261 }
262 else if (enc == rb_utf8_encoding()){
263 return rb_memsearch_qs_utf8(x0, m, y0, n);
264 }
265 }
266 else if (LIKELY(rb_enc_mbminlen(enc) == 2)) {
267 return rb_memsearch_wchar(x0, m, y0, n);
268 }
269 else if (LIKELY(rb_enc_mbminlen(enc) == 4)) {
270 return rb_memsearch_qchar(x0, m, y0, n);
271 }
272 return rb_memsearch_qs(x0, m, y0, n);
273 }
274
275 #define REG_LITERAL FL_USER5
276 #define REG_ENCODING_NONE FL_USER6
277
278 #define KCODE_FIXED FL_USER4
279
280 #define ARG_REG_OPTION_MASK \
281 (ONIG_OPTION_IGNORECASE|ONIG_OPTION_MULTILINE|ONIG_OPTION_EXTEND)
282 #define ARG_ENCODING_FIXED 16
283 #define ARG_ENCODING_NONE 32
284
285 static int
char_to_option(int c)286 char_to_option(int c)
287 {
288 int val;
289
290 switch (c) {
291 case 'i':
292 val = ONIG_OPTION_IGNORECASE;
293 break;
294 case 'x':
295 val = ONIG_OPTION_EXTEND;
296 break;
297 case 'm':
298 val = ONIG_OPTION_MULTILINE;
299 break;
300 default:
301 val = 0;
302 break;
303 }
304 return val;
305 }
306
307 static char *
option_to_str(char str[4],int options)308 option_to_str(char str[4], int options)
309 {
310 char *p = str;
311 if (options & ONIG_OPTION_MULTILINE) *p++ = 'm';
312 if (options & ONIG_OPTION_IGNORECASE) *p++ = 'i';
313 if (options & ONIG_OPTION_EXTEND) *p++ = 'x';
314 *p = 0;
315 return str;
316 }
317
318 extern int
rb_char_to_option_kcode(int c,int * option,int * kcode)319 rb_char_to_option_kcode(int c, int *option, int *kcode)
320 {
321 *option = 0;
322
323 switch (c) {
324 case 'n':
325 *kcode = rb_ascii8bit_encindex();
326 return (*option = ARG_ENCODING_NONE);
327 case 'e':
328 *kcode = ENCINDEX_EUC_JP;
329 break;
330 case 's':
331 *kcode = ENCINDEX_Windows_31J;
332 break;
333 case 'u':
334 *kcode = rb_utf8_encindex();
335 break;
336 default:
337 *kcode = -1;
338 return (*option = char_to_option(c));
339 }
340 *option = ARG_ENCODING_FIXED;
341 return 1;
342 }
343
344 static void
rb_reg_check(VALUE re)345 rb_reg_check(VALUE re)
346 {
347 if (!RREGEXP_PTR(re) || !RREGEXP_SRC(re) || !RREGEXP_SRC_PTR(re)) {
348 rb_raise(rb_eTypeError, "uninitialized Regexp");
349 }
350 }
351
352 static void
rb_reg_expr_str(VALUE str,const char * s,long len,rb_encoding * enc,rb_encoding * resenc,int term)353 rb_reg_expr_str(VALUE str, const char *s, long len,
354 rb_encoding *enc, rb_encoding *resenc, int term)
355 {
356 const char *p, *pend;
357 int cr = ENC_CODERANGE_UNKNOWN;
358 int need_escape = 0;
359 int c, clen;
360
361 p = s; pend = p + len;
362 rb_str_coderange_scan_restartable(p, pend, enc, &cr);
363 if (rb_enc_asciicompat(enc) && ENC_CODERANGE_CLEAN_P(cr)) {
364 while (p < pend) {
365 c = rb_enc_ascget(p, pend, &clen, enc);
366 if (c == -1) {
367 if (enc == resenc) {
368 p += mbclen(p, pend, enc);
369 }
370 else {
371 need_escape = 1;
372 break;
373 }
374 }
375 else if (c != term && rb_enc_isprint(c, enc)) {
376 p += clen;
377 }
378 else {
379 need_escape = 1;
380 break;
381 }
382 }
383 }
384 else {
385 need_escape = 1;
386 }
387
388 if (!need_escape) {
389 rb_str_buf_cat(str, s, len);
390 }
391 else {
392 int unicode_p = rb_enc_unicode_p(enc);
393 p = s;
394 while (p<pend) {
395 c = rb_enc_ascget(p, pend, &clen, enc);
396 if (c == '\\' && p+clen < pend) {
397 int n = clen + mbclen(p+clen, pend, enc);
398 rb_str_buf_cat(str, p, n);
399 p += n;
400 continue;
401 }
402 else if (c == -1) {
403 clen = rb_enc_precise_mbclen(p, pend, enc);
404 if (!MBCLEN_CHARFOUND_P(clen)) {
405 c = (unsigned char)*p;
406 clen = 1;
407 goto hex;
408 }
409 if (resenc) {
410 unsigned int c = rb_enc_mbc_to_codepoint(p, pend, enc);
411 rb_str_buf_cat_escaped_char(str, c, unicode_p);
412 }
413 else {
414 clen = MBCLEN_CHARFOUND_LEN(clen);
415 rb_str_buf_cat(str, p, clen);
416 }
417 }
418 else if (c == term) {
419 char c = '\\';
420 rb_str_buf_cat(str, &c, 1);
421 rb_str_buf_cat(str, p, clen);
422 }
423 else if (rb_enc_isprint(c, enc)) {
424 rb_str_buf_cat(str, p, clen);
425 }
426 else if (!rb_enc_isspace(c, enc)) {
427 char b[8];
428
429 hex:
430 snprintf(b, sizeof(b), "\\x%02X", c);
431 rb_str_buf_cat(str, b, 4);
432 }
433 else {
434 rb_str_buf_cat(str, p, clen);
435 }
436 p += clen;
437 }
438 }
439 }
440
441 static VALUE
rb_reg_desc(const char * s,long len,VALUE re)442 rb_reg_desc(const char *s, long len, VALUE re)
443 {
444 rb_encoding *enc = rb_enc_get(re);
445 VALUE str = rb_str_buf_new2("/");
446 rb_encoding *resenc = rb_default_internal_encoding();
447 if (resenc == NULL) resenc = rb_default_external_encoding();
448
449 if (re && rb_enc_asciicompat(enc)) {
450 rb_enc_copy(str, re);
451 }
452 else {
453 rb_enc_associate(str, rb_usascii_encoding());
454 }
455 rb_reg_expr_str(str, s, len, enc, resenc, '/');
456 rb_str_buf_cat2(str, "/");
457 if (re) {
458 char opts[4];
459 rb_reg_check(re);
460 if (*option_to_str(opts, RREGEXP_PTR(re)->options))
461 rb_str_buf_cat2(str, opts);
462 if (RBASIC(re)->flags & REG_ENCODING_NONE)
463 rb_str_buf_cat2(str, "n");
464 }
465 OBJ_INFECT(str, re);
466 return str;
467 }
468
469
470 /*
471 * call-seq:
472 * rxp.source -> str
473 *
474 * Returns the original string of the pattern.
475 *
476 * /ab+c/ix.source #=> "ab+c"
477 *
478 * Note that escape sequences are retained as is.
479 *
480 * /\x20\+/.source #=> "\\x20\\+"
481 *
482 */
483
484 static VALUE
rb_reg_source(VALUE re)485 rb_reg_source(VALUE re)
486 {
487 VALUE str;
488
489 rb_reg_check(re);
490 str = rb_str_dup(RREGEXP_SRC(re));
491 if (OBJ_TAINTED(re)) OBJ_TAINT(str);
492 return str;
493 }
494
495 /*
496 * call-seq:
497 * rxp.inspect -> string
498 *
499 * Produce a nicely formatted string-version of _rxp_. Perhaps surprisingly,
500 * <code>#inspect</code> actually produces the more natural version of
501 * the string than <code>#to_s</code>.
502 *
503 * /ab+c/ix.inspect #=> "/ab+c/ix"
504 *
505 */
506
507 static VALUE
rb_reg_inspect(VALUE re)508 rb_reg_inspect(VALUE re)
509 {
510 if (!RREGEXP_PTR(re) || !RREGEXP_SRC(re) || !RREGEXP_SRC_PTR(re)) {
511 return rb_any_to_s(re);
512 }
513 return rb_reg_desc(RREGEXP_SRC_PTR(re), RREGEXP_SRC_LEN(re), re);
514 }
515
516 static VALUE rb_reg_str_with_term(VALUE re, int term);
517
518 /*
519 * call-seq:
520 * rxp.to_s -> str
521 *
522 * Returns a string containing the regular expression and its options (using the
523 * <code>(?opts:source)</code> notation. This string can be fed back in to
524 * <code>Regexp::new</code> to a regular expression with the same semantics as
525 * the original. (However, <code>Regexp#==</code> may not return true when
526 * comparing the two, as the source of the regular expression itself may
527 * differ, as the example shows). <code>Regexp#inspect</code> produces a
528 * generally more readable version of <i>rxp</i>.
529 *
530 * r1 = /ab+c/ix #=> /ab+c/ix
531 * s1 = r1.to_s #=> "(?ix-m:ab+c)"
532 * r2 = Regexp.new(s1) #=> /(?ix-m:ab+c)/
533 * r1 == r2 #=> false
534 * r1.source #=> "ab+c"
535 * r2.source #=> "(?ix-m:ab+c)"
536 */
537
538 static VALUE
rb_reg_to_s(VALUE re)539 rb_reg_to_s(VALUE re)
540 {
541 return rb_reg_str_with_term(re, '/');
542 }
543
544 static VALUE
rb_reg_str_with_term(VALUE re,int term)545 rb_reg_str_with_term(VALUE re, int term)
546 {
547 int options, opt;
548 const int embeddable = ONIG_OPTION_MULTILINE|ONIG_OPTION_IGNORECASE|ONIG_OPTION_EXTEND;
549 long len;
550 const UChar* ptr;
551 VALUE str = rb_str_buf_new2("(?");
552 char optbuf[5];
553 rb_encoding *enc = rb_enc_get(re);
554
555 rb_reg_check(re);
556
557 rb_enc_copy(str, re);
558 options = RREGEXP_PTR(re)->options;
559 ptr = (UChar*)RREGEXP_SRC_PTR(re);
560 len = RREGEXP_SRC_LEN(re);
561 again:
562 if (len >= 4 && ptr[0] == '(' && ptr[1] == '?') {
563 int err = 1;
564 ptr += 2;
565 if ((len -= 2) > 0) {
566 do {
567 opt = char_to_option((int )*ptr);
568 if (opt != 0) {
569 options |= opt;
570 }
571 else {
572 break;
573 }
574 ++ptr;
575 } while (--len > 0);
576 }
577 if (len > 1 && *ptr == '-') {
578 ++ptr;
579 --len;
580 do {
581 opt = char_to_option((int )*ptr);
582 if (opt != 0) {
583 options &= ~opt;
584 }
585 else {
586 break;
587 }
588 ++ptr;
589 } while (--len > 0);
590 }
591 if (*ptr == ')') {
592 --len;
593 ++ptr;
594 goto again;
595 }
596 if (*ptr == ':' && ptr[len-1] == ')') {
597 Regexp *rp;
598 VALUE verbose = ruby_verbose;
599 ruby_verbose = Qfalse;
600
601 ++ptr;
602 len -= 2;
603 err = onig_new(&rp, ptr, ptr + len, options,
604 enc, OnigDefaultSyntax, NULL);
605 onig_free(rp);
606 ruby_verbose = verbose;
607 }
608 if (err) {
609 options = RREGEXP_PTR(re)->options;
610 ptr = (UChar*)RREGEXP_SRC_PTR(re);
611 len = RREGEXP_SRC_LEN(re);
612 }
613 }
614
615 if (*option_to_str(optbuf, options)) rb_str_buf_cat2(str, optbuf);
616
617 if ((options & embeddable) != embeddable) {
618 optbuf[0] = '-';
619 option_to_str(optbuf + 1, ~options);
620 rb_str_buf_cat2(str, optbuf);
621 }
622
623 rb_str_buf_cat2(str, ":");
624 if (rb_enc_asciicompat(enc)) {
625 rb_reg_expr_str(str, (char*)ptr, len, enc, NULL, term);
626 rb_str_buf_cat2(str, ")");
627 }
628 else {
629 const char *s, *e;
630 char *paren;
631 ptrdiff_t n;
632 rb_str_buf_cat2(str, ")");
633 rb_enc_associate(str, rb_usascii_encoding());
634 str = rb_str_encode(str, rb_enc_from_encoding(enc), 0, Qnil);
635
636 /* backup encoded ")" to paren */
637 s = RSTRING_PTR(str);
638 e = RSTRING_END(str);
639 s = rb_enc_left_char_head(s, e-1, e, enc);
640 n = e - s;
641 paren = ALLOCA_N(char, n);
642 memcpy(paren, s, n);
643 rb_str_resize(str, RSTRING_LEN(str) - n);
644
645 rb_reg_expr_str(str, (char*)ptr, len, enc, NULL, term);
646 rb_str_buf_cat(str, paren, n);
647 }
648 rb_enc_copy(str, re);
649
650 OBJ_INFECT(str, re);
651 return str;
652 }
653
654 NORETURN(static void rb_reg_raise(const char *s, long len, const char *err, VALUE re));
655
656 static void
rb_reg_raise(const char * s,long len,const char * err,VALUE re)657 rb_reg_raise(const char *s, long len, const char *err, VALUE re)
658 {
659 VALUE desc = rb_reg_desc(s, len, re);
660
661 rb_raise(rb_eRegexpError, "%s: %"PRIsVALUE, err, desc);
662 }
663
664 static VALUE
rb_enc_reg_error_desc(const char * s,long len,rb_encoding * enc,int options,const char * err)665 rb_enc_reg_error_desc(const char *s, long len, rb_encoding *enc, int options, const char *err)
666 {
667 char opts[6];
668 VALUE desc = rb_str_buf_new2(err);
669 rb_encoding *resenc = rb_default_internal_encoding();
670 if (resenc == NULL) resenc = rb_default_external_encoding();
671
672 rb_enc_associate(desc, enc);
673 rb_str_buf_cat2(desc, ": /");
674 rb_reg_expr_str(desc, s, len, enc, resenc, '/');
675 opts[0] = '/';
676 option_to_str(opts + 1, options);
677 rb_str_buf_cat2(desc, opts);
678 return rb_exc_new3(rb_eRegexpError, desc);
679 }
680
681 NORETURN(static void rb_enc_reg_raise(const char *s, long len, rb_encoding *enc, int options, const char *err));
682
683 static void
rb_enc_reg_raise(const char * s,long len,rb_encoding * enc,int options,const char * err)684 rb_enc_reg_raise(const char *s, long len, rb_encoding *enc, int options, const char *err)
685 {
686 rb_exc_raise(rb_enc_reg_error_desc(s, len, enc, options, err));
687 }
688
689 static VALUE
rb_reg_error_desc(VALUE str,int options,const char * err)690 rb_reg_error_desc(VALUE str, int options, const char *err)
691 {
692 return rb_enc_reg_error_desc(RSTRING_PTR(str), RSTRING_LEN(str),
693 rb_enc_get(str), options, err);
694 }
695
696 NORETURN(static void rb_reg_raise_str(VALUE str, int options, const char *err));
697
698 static void
rb_reg_raise_str(VALUE str,int options,const char * err)699 rb_reg_raise_str(VALUE str, int options, const char *err)
700 {
701 rb_exc_raise(rb_reg_error_desc(str, options, err));
702 }
703
704
705 /*
706 * call-seq:
707 * rxp.casefold? -> true or false
708 *
709 * Returns the value of the case-insensitive flag.
710 *
711 * /a/.casefold? #=> false
712 * /a/i.casefold? #=> true
713 * /(?i:a)/.casefold? #=> false
714 */
715
716 static VALUE
rb_reg_casefold_p(VALUE re)717 rb_reg_casefold_p(VALUE re)
718 {
719 rb_reg_check(re);
720 if (RREGEXP_PTR(re)->options & ONIG_OPTION_IGNORECASE) return Qtrue;
721 return Qfalse;
722 }
723
724
725 /*
726 * call-seq:
727 * rxp.options -> integer
728 *
729 * Returns the set of bits corresponding to the options used when creating this
730 * Regexp (see <code>Regexp::new</code> for details. Note that additional bits
731 * may be set in the returned options: these are used internally by the regular
732 * expression code. These extra bits are ignored if the options are passed to
733 * <code>Regexp::new</code>.
734 *
735 * Regexp::IGNORECASE #=> 1
736 * Regexp::EXTENDED #=> 2
737 * Regexp::MULTILINE #=> 4
738 *
739 * /cat/.options #=> 0
740 * /cat/ix.options #=> 3
741 * Regexp.new('cat', true).options #=> 1
742 * /\xa1\xa2/e.options #=> 16
743 *
744 * r = /cat/ix
745 * Regexp.new(r.source, r.options) #=> /cat/ix
746 */
747
748 static VALUE
rb_reg_options_m(VALUE re)749 rb_reg_options_m(VALUE re)
750 {
751 int options = rb_reg_options(re);
752 return INT2NUM(options);
753 }
754
755 static int
reg_names_iter(const OnigUChar * name,const OnigUChar * name_end,int back_num,int * back_refs,OnigRegex regex,void * arg)756 reg_names_iter(const OnigUChar *name, const OnigUChar *name_end,
757 int back_num, int *back_refs, OnigRegex regex, void *arg)
758 {
759 VALUE ary = (VALUE)arg;
760 rb_ary_push(ary, rb_enc_str_new((const char *)name, name_end-name, regex->enc));
761 return 0;
762 }
763
764 /*
765 * call-seq:
766 * rxp.names -> [name1, name2, ...]
767 *
768 * Returns a list of names of captures as an array of strings.
769 *
770 * /(?<foo>.)(?<bar>.)(?<baz>.)/.names
771 * #=> ["foo", "bar", "baz"]
772 *
773 * /(?<foo>.)(?<foo>.)/.names
774 * #=> ["foo"]
775 *
776 * /(.)(.)/.names
777 * #=> []
778 */
779
780 static VALUE
rb_reg_names(VALUE re)781 rb_reg_names(VALUE re)
782 {
783 VALUE ary;
784 rb_reg_check(re);
785 ary = rb_ary_new_capa(onig_number_of_names(RREGEXP_PTR(re)));
786 onig_foreach_name(RREGEXP_PTR(re), reg_names_iter, (void*)ary);
787 return ary;
788 }
789
790 static int
reg_named_captures_iter(const OnigUChar * name,const OnigUChar * name_end,int back_num,int * back_refs,OnigRegex regex,void * arg)791 reg_named_captures_iter(const OnigUChar *name, const OnigUChar *name_end,
792 int back_num, int *back_refs, OnigRegex regex, void *arg)
793 {
794 VALUE hash = (VALUE)arg;
795 VALUE ary = rb_ary_new2(back_num);
796 int i;
797
798 for (i = 0; i < back_num; i++)
799 rb_ary_store(ary, i, INT2NUM(back_refs[i]));
800
801 rb_hash_aset(hash, rb_str_new((const char*)name, name_end-name),ary);
802
803 return 0;
804 }
805
806 /*
807 * call-seq:
808 * rxp.named_captures -> hash
809 *
810 * Returns a hash representing information about named captures of <i>rxp</i>.
811 *
812 * A key of the hash is a name of the named captures.
813 * A value of the hash is an array which is list of indexes of corresponding
814 * named captures.
815 *
816 * /(?<foo>.)(?<bar>.)/.named_captures
817 * #=> {"foo"=>[1], "bar"=>[2]}
818 *
819 * /(?<foo>.)(?<foo>.)/.named_captures
820 * #=> {"foo"=>[1, 2]}
821 *
822 * If there are no named captures, an empty hash is returned.
823 *
824 * /(.)(.)/.named_captures
825 * #=> {}
826 */
827
828 static VALUE
rb_reg_named_captures(VALUE re)829 rb_reg_named_captures(VALUE re)
830 {
831 regex_t *reg = (rb_reg_check(re), RREGEXP_PTR(re));
832 VALUE hash = rb_hash_new_with_size(onig_number_of_names(reg));
833 onig_foreach_name(reg, reg_named_captures_iter, (void*)hash);
834 return hash;
835 }
836
837 static int
onig_new_with_source(regex_t ** reg,const UChar * pattern,const UChar * pattern_end,OnigOptionType option,OnigEncoding enc,const OnigSyntaxType * syntax,OnigErrorInfo * einfo,const char * sourcefile,int sourceline)838 onig_new_with_source(regex_t** reg, const UChar* pattern, const UChar* pattern_end,
839 OnigOptionType option, OnigEncoding enc, const OnigSyntaxType* syntax,
840 OnigErrorInfo* einfo, const char *sourcefile, int sourceline)
841 {
842 int r;
843
844 *reg = (regex_t* )malloc(sizeof(regex_t));
845 if (IS_NULL(*reg)) return ONIGERR_MEMORY;
846
847 r = onig_reg_init(*reg, option, ONIGENC_CASE_FOLD_DEFAULT, enc, syntax);
848 if (r) goto err;
849
850 r = onig_compile_ruby(*reg, pattern, pattern_end, einfo, sourcefile, sourceline);
851 if (r) {
852 err:
853 onig_free(*reg);
854 *reg = NULL;
855 }
856 return r;
857 }
858
859 static Regexp*
make_regexp(const char * s,long len,rb_encoding * enc,int flags,onig_errmsg_buffer err,const char * sourcefile,int sourceline)860 make_regexp(const char *s, long len, rb_encoding *enc, int flags, onig_errmsg_buffer err,
861 const char *sourcefile, int sourceline)
862 {
863 Regexp *rp;
864 int r;
865 OnigErrorInfo einfo;
866
867 /* Handle escaped characters first. */
868
869 /* Build a copy of the string (in dest) with the
870 escaped characters translated, and generate the regex
871 from that.
872 */
873
874 r = onig_new_with_source(&rp, (UChar*)s, (UChar*)(s + len), flags,
875 enc, OnigDefaultSyntax, &einfo, sourcefile, sourceline);
876 if (r) {
877 onig_error_code_to_str((UChar*)err, r, &einfo);
878 return 0;
879 }
880 return rp;
881 }
882
883
884 /*
885 * Document-class: MatchData
886 *
887 * <code>MatchData</code> encapsulates the result of matching a Regexp against
888 * string. It is returned by Regexp#match and
889 * String#match, and also stored in a global variable returned by
890 * Regexp.last_match.
891 *
892 * Usage:
893 *
894 * url = 'https://docs.ruby-lang.org/en/2.5.0/MatchData.html'
895 * m = url.match(/(\d\.?)+/) # => #<MatchData "2.5.0" 1:"0">
896 * m.string # => "https://docs.ruby-lang.org/en/2.5.0/MatchData.html"
897 * m.regexp # => /(\d\.?)+/
898 * # entire matched substring:
899 * m[0] # => "2.5.0"
900 *
901 * # Working with unnamed captures
902 * m = url.match(%r{([^/]+)/([^/]+)\.html$})
903 * m.captures # => ["2.5.0", "MatchData"]
904 * m[1] # => "2.5.0"
905 * m.values_at(1, 2) # => ["2.5.0", "MatchData"]
906 *
907 * # Working with named captures
908 * m = url.match(%r{(?<version>[^/]+)/(?<module>[^/]+)\.html$})
909 * m.captures # => ["2.5.0", "MatchData"]
910 * m.named_captures # => {"version"=>"2.5.0", "module"=>"MatchData"}
911 * m[:version] # => "2.5.0"
912 * m.values_at(:version, :module)
913 * # => ["2.5.0", "MatchData"]
914 * # Numerical indexes are working, too
915 * m[1] # => "2.5.0"
916 * m.values_at(1, 2) # => ["2.5.0", "MatchData"]
917 *
918 * == Global variables equivalence
919 *
920 * Parts of last <code>MatchData</code> (returned by Regexp.last_match) are also
921 * aliased as global variables:
922 *
923 * * <code>$~</code> is <code>Regexp.last_match</code>;
924 * * <code>$&</code> is <code>Regexp.last_match[0]</code>;
925 * * <code>$1</code>, <code>$2</code>, and so on are
926 * <code>Regexp.last_match[i]</code> (captures by number);
927 * * <code>$`</code> is <code>Regexp.last_match.pre_match</code>;
928 * * <code>$'</code> is <code>Regexp.last_match.post_match</code>;
929 * * <code>$+</code> is <code>Regexp.last_match[-1]</code> (the last capture).
930 *
931 * See also "Special global variables" section in Regexp documentation.
932 */
933
934 VALUE rb_cMatch;
935
936 static VALUE
match_alloc(VALUE klass)937 match_alloc(VALUE klass)
938 {
939 NEWOBJ_OF(match, struct RMatch, klass, T_MATCH);
940
941 match->str = 0;
942 match->rmatch = 0;
943 match->regexp = 0;
944 match->rmatch = ZALLOC(struct rmatch);
945
946 return (VALUE)match;
947 }
948
949 int
rb_reg_region_copy(struct re_registers * to,const struct re_registers * from)950 rb_reg_region_copy(struct re_registers *to, const struct re_registers *from)
951 {
952 onig_region_copy(to, (OnigRegion *)from);
953 if (to->allocated) return 0;
954 rb_gc();
955 onig_region_copy(to, (OnigRegion *)from);
956 if (to->allocated) return 0;
957 return ONIGERR_MEMORY;
958 }
959
960 typedef struct {
961 long byte_pos;
962 long char_pos;
963 } pair_t;
964
965 static int
pair_byte_cmp(const void * pair1,const void * pair2)966 pair_byte_cmp(const void *pair1, const void *pair2)
967 {
968 long diff = ((pair_t*)pair1)->byte_pos - ((pair_t*)pair2)->byte_pos;
969 #if SIZEOF_LONG > SIZEOF_INT
970 return diff ? diff > 0 ? 1 : -1 : 0;
971 #else
972 return (int)diff;
973 #endif
974 }
975
976 static void
update_char_offset(VALUE match)977 update_char_offset(VALUE match)
978 {
979 struct rmatch *rm = RMATCH(match)->rmatch;
980 struct re_registers *regs;
981 int i, num_regs, num_pos;
982 long c;
983 char *s, *p, *q;
984 rb_encoding *enc;
985 pair_t *pairs;
986
987 if (rm->char_offset_updated)
988 return;
989
990 regs = &rm->regs;
991 num_regs = rm->regs.num_regs;
992
993 if (rm->char_offset_num_allocated < num_regs) {
994 REALLOC_N(rm->char_offset, struct rmatch_offset, num_regs);
995 rm->char_offset_num_allocated = num_regs;
996 }
997
998 enc = rb_enc_get(RMATCH(match)->str);
999 if (rb_enc_mbmaxlen(enc) == 1) {
1000 for (i = 0; i < num_regs; i++) {
1001 rm->char_offset[i].beg = BEG(i);
1002 rm->char_offset[i].end = END(i);
1003 }
1004 rm->char_offset_updated = 1;
1005 return;
1006 }
1007
1008 pairs = ALLOCA_N(pair_t, num_regs*2);
1009 num_pos = 0;
1010 for (i = 0; i < num_regs; i++) {
1011 if (BEG(i) < 0)
1012 continue;
1013 pairs[num_pos++].byte_pos = BEG(i);
1014 pairs[num_pos++].byte_pos = END(i);
1015 }
1016 qsort(pairs, num_pos, sizeof(pair_t), pair_byte_cmp);
1017
1018 s = p = RSTRING_PTR(RMATCH(match)->str);
1019 c = 0;
1020 for (i = 0; i < num_pos; i++) {
1021 q = s + pairs[i].byte_pos;
1022 c += rb_enc_strlen(p, q, enc);
1023 pairs[i].char_pos = c;
1024 p = q;
1025 }
1026
1027 for (i = 0; i < num_regs; i++) {
1028 pair_t key, *found;
1029 if (BEG(i) < 0) {
1030 rm->char_offset[i].beg = -1;
1031 rm->char_offset[i].end = -1;
1032 continue;
1033 }
1034
1035 key.byte_pos = BEG(i);
1036 found = bsearch(&key, pairs, num_pos, sizeof(pair_t), pair_byte_cmp);
1037 rm->char_offset[i].beg = found->char_pos;
1038
1039 key.byte_pos = END(i);
1040 found = bsearch(&key, pairs, num_pos, sizeof(pair_t), pair_byte_cmp);
1041 rm->char_offset[i].end = found->char_pos;
1042 }
1043
1044 rm->char_offset_updated = 1;
1045 }
1046
1047 static void
match_check(VALUE match)1048 match_check(VALUE match)
1049 {
1050 if (!RMATCH(match)->regexp) {
1051 rb_raise(rb_eTypeError, "uninitialized Match");
1052 }
1053 }
1054
1055 /* :nodoc: */
1056 static VALUE
match_init_copy(VALUE obj,VALUE orig)1057 match_init_copy(VALUE obj, VALUE orig)
1058 {
1059 struct rmatch *rm;
1060
1061 if (!OBJ_INIT_COPY(obj, orig)) return obj;
1062
1063 RMATCH(obj)->str = RMATCH(orig)->str;
1064 RMATCH(obj)->regexp = RMATCH(orig)->regexp;
1065
1066 rm = RMATCH(obj)->rmatch;
1067 if (rb_reg_region_copy(&rm->regs, RMATCH_REGS(orig)))
1068 rb_memerror();
1069
1070 if (!RMATCH(orig)->rmatch->char_offset_updated) {
1071 rm->char_offset_updated = 0;
1072 }
1073 else {
1074 if (rm->char_offset_num_allocated < rm->regs.num_regs) {
1075 REALLOC_N(rm->char_offset, struct rmatch_offset, rm->regs.num_regs);
1076 rm->char_offset_num_allocated = rm->regs.num_regs;
1077 }
1078 MEMCPY(rm->char_offset, RMATCH(orig)->rmatch->char_offset,
1079 struct rmatch_offset, rm->regs.num_regs);
1080 rm->char_offset_updated = 1;
1081 RB_GC_GUARD(orig);
1082 }
1083
1084 return obj;
1085 }
1086
1087
1088 /*
1089 * call-seq:
1090 * mtch.regexp -> regexp
1091 *
1092 * Returns the regexp.
1093 *
1094 * m = /a.*b/.match("abc")
1095 * m.regexp #=> /a.*b/
1096 */
1097
1098 static VALUE
match_regexp(VALUE match)1099 match_regexp(VALUE match)
1100 {
1101 VALUE regexp;
1102 match_check(match);
1103 regexp = RMATCH(match)->regexp;
1104 if (NIL_P(regexp)) {
1105 VALUE str = rb_reg_nth_match(0, match);
1106 regexp = rb_reg_regcomp(rb_reg_quote(str));
1107 RMATCH(match)->regexp = regexp;
1108 }
1109 return regexp;
1110 }
1111
1112 /*
1113 * call-seq:
1114 * mtch.names -> [name1, name2, ...]
1115 *
1116 * Returns a list of names of captures as an array of strings.
1117 * It is same as mtch.regexp.names.
1118 *
1119 * /(?<foo>.)(?<bar>.)(?<baz>.)/.match("hoge").names
1120 * #=> ["foo", "bar", "baz"]
1121 *
1122 * m = /(?<x>.)(?<y>.)?/.match("a") #=> #<MatchData "a" x:"a" y:nil>
1123 * m.names #=> ["x", "y"]
1124 */
1125
1126 static VALUE
match_names(VALUE match)1127 match_names(VALUE match)
1128 {
1129 match_check(match);
1130 if (NIL_P(RMATCH(match)->regexp))
1131 return rb_ary_new_capa(0);
1132 return rb_reg_names(RMATCH(match)->regexp);
1133 }
1134
1135 /*
1136 * call-seq:
1137 * mtch.length -> integer
1138 * mtch.size -> integer
1139 *
1140 * Returns the number of elements in the match array.
1141 *
1142 * m = /(.)(.)(\d+)(\d)/.match("THX1138.")
1143 * m.length #=> 5
1144 * m.size #=> 5
1145 */
1146
1147 static VALUE
match_size(VALUE match)1148 match_size(VALUE match)
1149 {
1150 match_check(match);
1151 return INT2FIX(RMATCH_REGS(match)->num_regs);
1152 }
1153
1154 static int name_to_backref_number(struct re_registers *, VALUE, const char*, const char*);
1155
1156 static int
match_backref_number(VALUE match,VALUE backref)1157 match_backref_number(VALUE match, VALUE backref)
1158 {
1159 const char *name;
1160 int num;
1161
1162 struct re_registers *regs = RMATCH_REGS(match);
1163 VALUE regexp = RMATCH(match)->regexp;
1164
1165 match_check(match);
1166 if (SYMBOL_P(backref)) {
1167 backref = rb_sym2str(backref);
1168 }
1169 else if (!RB_TYPE_P(backref, T_STRING)) {
1170 return NUM2INT(backref);
1171 }
1172 name = StringValueCStr(backref);
1173
1174 num = name_to_backref_number(regs, regexp, name, name + strlen(name));
1175
1176 if (num < 1) {
1177 rb_raise(rb_eIndexError, "undefined group name reference: %s", name);
1178 }
1179
1180 return num;
1181 }
1182
1183 int
rb_reg_backref_number(VALUE match,VALUE backref)1184 rb_reg_backref_number(VALUE match, VALUE backref)
1185 {
1186 return match_backref_number(match, backref);
1187 }
1188
1189 /*
1190 * call-seq:
1191 * mtch.offset(n) -> array
1192 *
1193 * Returns a two-element array containing the beginning and ending offsets of
1194 * the <em>n</em>th match.
1195 * <em>n</em> can be a string or symbol to reference a named capture.
1196 *
1197 * m = /(.)(.)(\d+)(\d)/.match("THX1138.")
1198 * m.offset(0) #=> [1, 7]
1199 * m.offset(4) #=> [6, 7]
1200 *
1201 * m = /(?<foo>.)(.)(?<bar>.)/.match("hoge")
1202 * p m.offset(:foo) #=> [0, 1]
1203 * p m.offset(:bar) #=> [2, 3]
1204 *
1205 */
1206
1207 static VALUE
match_offset(VALUE match,VALUE n)1208 match_offset(VALUE match, VALUE n)
1209 {
1210 int i = match_backref_number(match, n);
1211 struct re_registers *regs = RMATCH_REGS(match);
1212
1213 match_check(match);
1214 if (i < 0 || regs->num_regs <= i)
1215 rb_raise(rb_eIndexError, "index %d out of matches", i);
1216
1217 if (BEG(i) < 0)
1218 return rb_assoc_new(Qnil, Qnil);
1219
1220 update_char_offset(match);
1221 return rb_assoc_new(INT2FIX(RMATCH(match)->rmatch->char_offset[i].beg),
1222 INT2FIX(RMATCH(match)->rmatch->char_offset[i].end));
1223 }
1224
1225
1226 /*
1227 * call-seq:
1228 * mtch.begin(n) -> integer
1229 *
1230 * Returns the offset of the start of the <em>n</em>th element of the match
1231 * array in the string.
1232 * <em>n</em> can be a string or symbol to reference a named capture.
1233 *
1234 * m = /(.)(.)(\d+)(\d)/.match("THX1138.")
1235 * m.begin(0) #=> 1
1236 * m.begin(2) #=> 2
1237 *
1238 * m = /(?<foo>.)(.)(?<bar>.)/.match("hoge")
1239 * p m.begin(:foo) #=> 0
1240 * p m.begin(:bar) #=> 2
1241 */
1242
1243 static VALUE
match_begin(VALUE match,VALUE n)1244 match_begin(VALUE match, VALUE n)
1245 {
1246 int i = match_backref_number(match, n);
1247 struct re_registers *regs = RMATCH_REGS(match);
1248
1249 match_check(match);
1250 if (i < 0 || regs->num_regs <= i)
1251 rb_raise(rb_eIndexError, "index %d out of matches", i);
1252
1253 if (BEG(i) < 0)
1254 return Qnil;
1255
1256 update_char_offset(match);
1257 return INT2FIX(RMATCH(match)->rmatch->char_offset[i].beg);
1258 }
1259
1260
1261 /*
1262 * call-seq:
1263 * mtch.end(n) -> integer
1264 *
1265 * Returns the offset of the character immediately following the end of the
1266 * <em>n</em>th element of the match array in the string.
1267 * <em>n</em> can be a string or symbol to reference a named capture.
1268 *
1269 * m = /(.)(.)(\d+)(\d)/.match("THX1138.")
1270 * m.end(0) #=> 7
1271 * m.end(2) #=> 3
1272 *
1273 * m = /(?<foo>.)(.)(?<bar>.)/.match("hoge")
1274 * p m.end(:foo) #=> 1
1275 * p m.end(:bar) #=> 3
1276 */
1277
1278 static VALUE
match_end(VALUE match,VALUE n)1279 match_end(VALUE match, VALUE n)
1280 {
1281 int i = match_backref_number(match, n);
1282 struct re_registers *regs = RMATCH_REGS(match);
1283
1284 match_check(match);
1285 if (i < 0 || regs->num_regs <= i)
1286 rb_raise(rb_eIndexError, "index %d out of matches", i);
1287
1288 if (BEG(i) < 0)
1289 return Qnil;
1290
1291 update_char_offset(match);
1292 return INT2FIX(RMATCH(match)->rmatch->char_offset[i].end);
1293 }
1294
1295 #define MATCH_BUSY FL_USER2
1296
1297 void
rb_match_busy(VALUE match)1298 rb_match_busy(VALUE match)
1299 {
1300 FL_SET(match, MATCH_BUSY);
1301 }
1302
1303 void
rb_match_unbusy(VALUE match)1304 rb_match_unbusy(VALUE match)
1305 {
1306 FL_UNSET(match, MATCH_BUSY);
1307 }
1308
1309 int
rb_match_count(VALUE match)1310 rb_match_count(VALUE match)
1311 {
1312 struct re_registers *regs;
1313 if (NIL_P(match)) return -1;
1314 regs = RMATCH_REGS(match);
1315 if (!regs) return -1;
1316 return regs->num_regs;
1317 }
1318
1319 int
rb_match_nth_defined(int nth,VALUE match)1320 rb_match_nth_defined(int nth, VALUE match)
1321 {
1322 struct re_registers *regs;
1323 if (NIL_P(match)) return FALSE;
1324 regs = RMATCH_REGS(match);
1325 if (!regs) return FALSE;
1326 if (nth >= regs->num_regs) {
1327 return FALSE;
1328 }
1329 if (nth < 0) {
1330 nth += regs->num_regs;
1331 if (nth <= 0) return FALSE;
1332 }
1333 return (BEG(nth) != -1);
1334 }
1335
1336 static void
match_set_string(VALUE m,VALUE string,long pos,long len)1337 match_set_string(VALUE m, VALUE string, long pos, long len)
1338 {
1339 struct RMatch *match = (struct RMatch *)m;
1340 struct rmatch *rmatch = match->rmatch;
1341
1342 match->str = string;
1343 match->regexp = Qnil;
1344 onig_region_resize(&rmatch->regs, 1);
1345 rmatch->regs.beg[0] = pos;
1346 rmatch->regs.end[0] = pos + len;
1347 rmatch->char_offset_updated = 0;
1348 OBJ_INFECT(match, string);
1349 }
1350
1351 void
rb_backref_set_string(VALUE string,long pos,long len)1352 rb_backref_set_string(VALUE string, long pos, long len)
1353 {
1354 VALUE match = rb_backref_get();
1355 if (NIL_P(match) || FL_TEST(match, MATCH_BUSY)) {
1356 match = match_alloc(rb_cMatch);
1357 }
1358 match_set_string(match, string, pos, len);
1359 rb_backref_set(match);
1360 }
1361
1362 /*
1363 * call-seq:
1364 * rxp.fixed_encoding? -> true or false
1365 *
1366 * Returns false if rxp is applicable to
1367 * a string with any ASCII compatible encoding.
1368 * Returns true otherwise.
1369 *
1370 * r = /a/
1371 * r.fixed_encoding? #=> false
1372 * r =~ "\u{6666} a" #=> 2
1373 * r =~ "\xa1\xa2 a".force_encoding("euc-jp") #=> 2
1374 * r =~ "abc".force_encoding("euc-jp") #=> 0
1375 *
1376 * r = /a/u
1377 * r.fixed_encoding? #=> true
1378 * r.encoding #=> #<Encoding:UTF-8>
1379 * r =~ "\u{6666} a" #=> 2
1380 * r =~ "\xa1\xa2".force_encoding("euc-jp") #=> Encoding::CompatibilityError
1381 * r =~ "abc".force_encoding("euc-jp") #=> 0
1382 *
1383 * r = /\u{6666}/
1384 * r.fixed_encoding? #=> true
1385 * r.encoding #=> #<Encoding:UTF-8>
1386 * r =~ "\u{6666} a" #=> 0
1387 * r =~ "\xa1\xa2".force_encoding("euc-jp") #=> Encoding::CompatibilityError
1388 * r =~ "abc".force_encoding("euc-jp") #=> nil
1389 */
1390
1391 static VALUE
rb_reg_fixed_encoding_p(VALUE re)1392 rb_reg_fixed_encoding_p(VALUE re)
1393 {
1394 if (FL_TEST(re, KCODE_FIXED))
1395 return Qtrue;
1396 else
1397 return Qfalse;
1398 }
1399
1400 static VALUE
1401 rb_reg_preprocess(const char *p, const char *end, rb_encoding *enc,
1402 rb_encoding **fixed_enc, onig_errmsg_buffer err);
1403
1404 NORETURN(static void reg_enc_error(VALUE re, VALUE str));
1405
1406 static void
reg_enc_error(VALUE re,VALUE str)1407 reg_enc_error(VALUE re, VALUE str)
1408 {
1409 rb_raise(rb_eEncCompatError,
1410 "incompatible encoding regexp match (%s regexp with %s string)",
1411 rb_enc_name(rb_enc_get(re)),
1412 rb_enc_name(rb_enc_get(str)));
1413 }
1414
1415 static inline int
str_coderange(VALUE str)1416 str_coderange(VALUE str)
1417 {
1418 int cr = ENC_CODERANGE(str);
1419 if (cr == ENC_CODERANGE_UNKNOWN) {
1420 cr = rb_enc_str_coderange(str);
1421 }
1422 return cr;
1423 }
1424
1425 static rb_encoding*
rb_reg_prepare_enc(VALUE re,VALUE str,int warn)1426 rb_reg_prepare_enc(VALUE re, VALUE str, int warn)
1427 {
1428 rb_encoding *enc = 0;
1429 int cr = str_coderange(str);
1430
1431 if (cr == ENC_CODERANGE_BROKEN) {
1432 rb_raise(rb_eArgError,
1433 "invalid byte sequence in %s",
1434 rb_enc_name(rb_enc_get(str)));
1435 }
1436
1437 rb_reg_check(re);
1438 enc = rb_enc_get(str);
1439 if (RREGEXP_PTR(re)->enc == enc) {
1440 }
1441 else if (cr == ENC_CODERANGE_7BIT &&
1442 RREGEXP_PTR(re)->enc == rb_usascii_encoding()) {
1443 enc = RREGEXP_PTR(re)->enc;
1444 }
1445 else if (!rb_enc_asciicompat(enc)) {
1446 reg_enc_error(re, str);
1447 }
1448 else if (rb_reg_fixed_encoding_p(re)) {
1449 if ((!rb_enc_asciicompat(RREGEXP_PTR(re)->enc) ||
1450 cr != ENC_CODERANGE_7BIT)) {
1451 reg_enc_error(re, str);
1452 }
1453 enc = RREGEXP_PTR(re)->enc;
1454 }
1455 else if (warn && (RBASIC(re)->flags & REG_ENCODING_NONE) &&
1456 enc != rb_ascii8bit_encoding() &&
1457 cr != ENC_CODERANGE_7BIT) {
1458 rb_warn("historical binary regexp match /.../n against %s string",
1459 rb_enc_name(enc));
1460 }
1461 return enc;
1462 }
1463
1464 regex_t *
rb_reg_prepare_re0(VALUE re,VALUE str,onig_errmsg_buffer err)1465 rb_reg_prepare_re0(VALUE re, VALUE str, onig_errmsg_buffer err)
1466 {
1467 regex_t *reg = RREGEXP_PTR(re);
1468 int r;
1469 OnigErrorInfo einfo;
1470 const char *pattern;
1471 VALUE unescaped;
1472 rb_encoding *fixed_enc = 0;
1473 rb_encoding *enc = rb_reg_prepare_enc(re, str, 1);
1474
1475 if (reg->enc == enc) return reg;
1476
1477 rb_reg_check(re);
1478 reg = RREGEXP_PTR(re);
1479 pattern = RREGEXP_SRC_PTR(re);
1480
1481 unescaped = rb_reg_preprocess(
1482 pattern, pattern + RREGEXP_SRC_LEN(re), enc,
1483 &fixed_enc, err);
1484
1485 if (unescaped == Qnil) {
1486 rb_raise(rb_eArgError, "regexp preprocess failed: %s", err);
1487 }
1488
1489 r = onig_new(®, (UChar* )RSTRING_PTR(unescaped),
1490 (UChar* )(RSTRING_PTR(unescaped) + RSTRING_LEN(unescaped)),
1491 reg->options, enc,
1492 OnigDefaultSyntax, &einfo);
1493 if (r) {
1494 onig_error_code_to_str((UChar*)err, r, &einfo);
1495 rb_reg_raise(pattern, RREGEXP_SRC_LEN(re), err, re);
1496 }
1497
1498 RB_GC_GUARD(unescaped);
1499 return reg;
1500 }
1501
1502 regex_t *
rb_reg_prepare_re(VALUE re,VALUE str)1503 rb_reg_prepare_re(VALUE re, VALUE str)
1504 {
1505 onig_errmsg_buffer err = "";
1506 return rb_reg_prepare_re0(re, str, err);
1507 }
1508
1509 long
rb_reg_adjust_startpos(VALUE re,VALUE str,long pos,int reverse)1510 rb_reg_adjust_startpos(VALUE re, VALUE str, long pos, int reverse)
1511 {
1512 long range;
1513 rb_encoding *enc;
1514 UChar *p, *string;
1515
1516 enc = rb_reg_prepare_enc(re, str, 0);
1517
1518 if (reverse) {
1519 range = -pos;
1520 }
1521 else {
1522 range = RSTRING_LEN(str) - pos;
1523 }
1524
1525 if (pos > 0 && ONIGENC_MBC_MAXLEN(enc) != 1 && pos < RSTRING_LEN(str)) {
1526 string = (UChar*)RSTRING_PTR(str);
1527
1528 if (range > 0) {
1529 p = onigenc_get_right_adjust_char_head(enc, string, string + pos, string + RSTRING_LEN(str));
1530 }
1531 else {
1532 p = ONIGENC_LEFT_ADJUST_CHAR_HEAD(enc, string, string + pos, string + RSTRING_LEN(str));
1533 }
1534 return p - string;
1535 }
1536
1537 return pos;
1538 }
1539
1540 /* returns byte offset */
1541 long
rb_reg_search0(VALUE re,VALUE str,long pos,int reverse,int set_backref_str)1542 rb_reg_search0(VALUE re, VALUE str, long pos, int reverse, int set_backref_str)
1543 {
1544 long result;
1545 VALUE match;
1546 struct re_registers regi, *regs = ®i;
1547 char *range = RSTRING_PTR(str);
1548 regex_t *reg;
1549 int tmpreg;
1550 onig_errmsg_buffer err = "";
1551
1552 if (pos > RSTRING_LEN(str) || pos < 0) {
1553 rb_backref_set(Qnil);
1554 return -1;
1555 }
1556
1557 reg = rb_reg_prepare_re0(re, str, err);
1558 tmpreg = reg != RREGEXP_PTR(re);
1559 if (!tmpreg) RREGEXP(re)->usecnt++;
1560
1561 match = rb_backref_get();
1562 if (!NIL_P(match)) {
1563 if (FL_TEST(match, MATCH_BUSY)) {
1564 match = Qnil;
1565 }
1566 else {
1567 regs = RMATCH_REGS(match);
1568 }
1569 }
1570 if (NIL_P(match)) {
1571 MEMZERO(regs, struct re_registers, 1);
1572 }
1573 if (!reverse) {
1574 range += RSTRING_LEN(str);
1575 }
1576 result = onig_search(reg,
1577 (UChar*)(RSTRING_PTR(str)),
1578 ((UChar*)(RSTRING_PTR(str)) + RSTRING_LEN(str)),
1579 ((UChar*)(RSTRING_PTR(str)) + pos),
1580 ((UChar*)range),
1581 regs, ONIG_OPTION_NONE);
1582 if (!tmpreg) RREGEXP(re)->usecnt--;
1583 if (tmpreg) {
1584 if (RREGEXP(re)->usecnt) {
1585 onig_free(reg);
1586 }
1587 else {
1588 onig_free(RREGEXP_PTR(re));
1589 RREGEXP_PTR(re) = reg;
1590 }
1591 }
1592 if (result < 0) {
1593 if (regs == ®i)
1594 onig_region_free(regs, 0);
1595 if (result == ONIG_MISMATCH) {
1596 rb_backref_set(Qnil);
1597 return result;
1598 }
1599 else {
1600 onig_error_code_to_str((UChar*)err, (int)result);
1601 rb_reg_raise(RREGEXP_SRC_PTR(re), RREGEXP_SRC_LEN(re), err, re);
1602 }
1603 }
1604
1605 if (NIL_P(match)) {
1606 int err;
1607 match = match_alloc(rb_cMatch);
1608 err = rb_reg_region_copy(RMATCH_REGS(match), regs);
1609 onig_region_free(regs, 0);
1610 if (err) rb_memerror();
1611 }
1612 else {
1613 FL_UNSET(match, FL_TAINT);
1614 }
1615
1616 if (set_backref_str) {
1617 RMATCH(match)->str = rb_str_new4(str);
1618 OBJ_INFECT(match, str);
1619 }
1620
1621 RMATCH(match)->regexp = re;
1622 RMATCH(match)->rmatch->char_offset_updated = 0;
1623 rb_backref_set(match);
1624
1625 OBJ_INFECT(match, re);
1626
1627 return result;
1628 }
1629
1630 long
rb_reg_search(VALUE re,VALUE str,long pos,int reverse)1631 rb_reg_search(VALUE re, VALUE str, long pos, int reverse)
1632 {
1633 return rb_reg_search0(re, str, pos, reverse, 1);
1634 }
1635
1636 bool
rb_reg_start_with_p(VALUE re,VALUE str)1637 rb_reg_start_with_p(VALUE re, VALUE str)
1638 {
1639 long result;
1640 VALUE match;
1641 struct re_registers regi, *regs = ®i;
1642 regex_t *reg;
1643 int tmpreg;
1644 onig_errmsg_buffer err = "";
1645
1646 reg = rb_reg_prepare_re0(re, str, err);
1647 tmpreg = reg != RREGEXP_PTR(re);
1648 if (!tmpreg) RREGEXP(re)->usecnt++;
1649
1650 match = rb_backref_get();
1651 if (!NIL_P(match)) {
1652 if (FL_TEST(match, MATCH_BUSY)) {
1653 match = Qnil;
1654 }
1655 else {
1656 regs = RMATCH_REGS(match);
1657 }
1658 }
1659 if (NIL_P(match)) {
1660 MEMZERO(regs, struct re_registers, 1);
1661 }
1662 result = onig_match(reg,
1663 (UChar*)(RSTRING_PTR(str)),
1664 ((UChar*)(RSTRING_PTR(str)) + RSTRING_LEN(str)),
1665 (UChar*)(RSTRING_PTR(str)),
1666 regs, ONIG_OPTION_NONE);
1667 if (!tmpreg) RREGEXP(re)->usecnt--;
1668 if (tmpreg) {
1669 if (RREGEXP(re)->usecnt) {
1670 onig_free(reg);
1671 }
1672 else {
1673 onig_free(RREGEXP_PTR(re));
1674 RREGEXP_PTR(re) = reg;
1675 }
1676 }
1677 if (result < 0) {
1678 if (regs == ®i)
1679 onig_region_free(regs, 0);
1680 if (result == ONIG_MISMATCH) {
1681 rb_backref_set(Qnil);
1682 return false;
1683 }
1684 else {
1685 onig_error_code_to_str((UChar*)err, (int)result);
1686 rb_reg_raise(RREGEXP_SRC_PTR(re), RREGEXP_SRC_LEN(re), err, re);
1687 }
1688 }
1689
1690 if (NIL_P(match)) {
1691 int err;
1692 match = match_alloc(rb_cMatch);
1693 err = rb_reg_region_copy(RMATCH_REGS(match), regs);
1694 onig_region_free(regs, 0);
1695 if (err) rb_memerror();
1696 }
1697 else {
1698 FL_UNSET(match, FL_TAINT);
1699 }
1700
1701 RMATCH(match)->str = rb_str_new4(str);
1702 OBJ_INFECT(match, str);
1703
1704 RMATCH(match)->regexp = re;
1705 RMATCH(match)->rmatch->char_offset_updated = 0;
1706 rb_backref_set(match);
1707
1708 OBJ_INFECT(match, re);
1709
1710 return true;
1711 }
1712
1713 VALUE
rb_reg_nth_defined(int nth,VALUE match)1714 rb_reg_nth_defined(int nth, VALUE match)
1715 {
1716 struct re_registers *regs;
1717 if (NIL_P(match)) return Qnil;
1718 match_check(match);
1719 regs = RMATCH_REGS(match);
1720 if (nth >= regs->num_regs) {
1721 return Qnil;
1722 }
1723 if (nth < 0) {
1724 nth += regs->num_regs;
1725 if (nth <= 0) return Qnil;
1726 }
1727 if (BEG(nth) == -1) return Qfalse;
1728 return Qtrue;
1729 }
1730
1731 VALUE
rb_reg_nth_match(int nth,VALUE match)1732 rb_reg_nth_match(int nth, VALUE match)
1733 {
1734 VALUE str;
1735 long start, end, len;
1736 struct re_registers *regs;
1737
1738 if (NIL_P(match)) return Qnil;
1739 match_check(match);
1740 regs = RMATCH_REGS(match);
1741 if (nth >= regs->num_regs) {
1742 return Qnil;
1743 }
1744 if (nth < 0) {
1745 nth += regs->num_regs;
1746 if (nth <= 0) return Qnil;
1747 }
1748 start = BEG(nth);
1749 if (start == -1) return Qnil;
1750 end = END(nth);
1751 len = end - start;
1752 str = rb_str_subseq(RMATCH(match)->str, start, len);
1753 OBJ_INFECT(str, match);
1754 return str;
1755 }
1756
1757 VALUE
rb_reg_last_match(VALUE match)1758 rb_reg_last_match(VALUE match)
1759 {
1760 return rb_reg_nth_match(0, match);
1761 }
1762
1763
1764 /*
1765 * call-seq:
1766 * mtch.pre_match -> str
1767 *
1768 * Returns the portion of the original string before the current match.
1769 * Equivalent to the special variable <code>$`</code>.
1770 *
1771 * m = /(.)(.)(\d+)(\d)/.match("THX1138.")
1772 * m.pre_match #=> "T"
1773 */
1774
1775 VALUE
rb_reg_match_pre(VALUE match)1776 rb_reg_match_pre(VALUE match)
1777 {
1778 VALUE str;
1779 struct re_registers *regs;
1780
1781 if (NIL_P(match)) return Qnil;
1782 match_check(match);
1783 regs = RMATCH_REGS(match);
1784 if (BEG(0) == -1) return Qnil;
1785 str = rb_str_subseq(RMATCH(match)->str, 0, BEG(0));
1786 if (OBJ_TAINTED(match)) OBJ_TAINT(str);
1787 return str;
1788 }
1789
1790
1791 /*
1792 * call-seq:
1793 * mtch.post_match -> str
1794 *
1795 * Returns the portion of the original string after the current match.
1796 * Equivalent to the special variable <code>$'</code>.
1797 *
1798 * m = /(.)(.)(\d+)(\d)/.match("THX1138: The Movie")
1799 * m.post_match #=> ": The Movie"
1800 */
1801
1802 VALUE
rb_reg_match_post(VALUE match)1803 rb_reg_match_post(VALUE match)
1804 {
1805 VALUE str;
1806 long pos;
1807 struct re_registers *regs;
1808
1809 if (NIL_P(match)) return Qnil;
1810 match_check(match);
1811 regs = RMATCH_REGS(match);
1812 if (BEG(0) == -1) return Qnil;
1813 str = RMATCH(match)->str;
1814 pos = END(0);
1815 str = rb_str_subseq(str, pos, RSTRING_LEN(str) - pos);
1816 if (OBJ_TAINTED(match)) OBJ_TAINT(str);
1817 return str;
1818 }
1819
1820 VALUE
rb_reg_match_last(VALUE match)1821 rb_reg_match_last(VALUE match)
1822 {
1823 int i;
1824 struct re_registers *regs;
1825
1826 if (NIL_P(match)) return Qnil;
1827 match_check(match);
1828 regs = RMATCH_REGS(match);
1829 if (BEG(0) == -1) return Qnil;
1830
1831 for (i=regs->num_regs-1; BEG(i) == -1 && i > 0; i--)
1832 ;
1833 if (i == 0) return Qnil;
1834 return rb_reg_nth_match(i, match);
1835 }
1836
1837 static VALUE
last_match_getter(void)1838 last_match_getter(void)
1839 {
1840 return rb_reg_last_match(rb_backref_get());
1841 }
1842
1843 static VALUE
prematch_getter(void)1844 prematch_getter(void)
1845 {
1846 return rb_reg_match_pre(rb_backref_get());
1847 }
1848
1849 static VALUE
postmatch_getter(void)1850 postmatch_getter(void)
1851 {
1852 return rb_reg_match_post(rb_backref_get());
1853 }
1854
1855 static VALUE
last_paren_match_getter(void)1856 last_paren_match_getter(void)
1857 {
1858 return rb_reg_match_last(rb_backref_get());
1859 }
1860
1861 static VALUE
match_array(VALUE match,int start)1862 match_array(VALUE match, int start)
1863 {
1864 struct re_registers *regs;
1865 VALUE ary;
1866 VALUE target;
1867 int i;
1868 int taint = OBJ_TAINTED(match);
1869
1870 match_check(match);
1871 regs = RMATCH_REGS(match);
1872 ary = rb_ary_new2(regs->num_regs);
1873 target = RMATCH(match)->str;
1874
1875 for (i=start; i<regs->num_regs; i++) {
1876 if (regs->beg[i] == -1) {
1877 rb_ary_push(ary, Qnil);
1878 }
1879 else {
1880 VALUE str = rb_str_subseq(target, regs->beg[i], regs->end[i]-regs->beg[i]);
1881 if (taint) OBJ_TAINT(str);
1882 rb_ary_push(ary, str);
1883 }
1884 }
1885 return ary;
1886 }
1887
1888
1889 /*
1890 * call-seq:
1891 * mtch.to_a -> anArray
1892 *
1893 * Returns the array of matches.
1894 *
1895 * m = /(.)(.)(\d+)(\d)/.match("THX1138.")
1896 * m.to_a #=> ["HX1138", "H", "X", "113", "8"]
1897 *
1898 * Because <code>to_a</code> is called when expanding
1899 * <code>*</code><em>variable</em>, there's a useful assignment
1900 * shortcut for extracting matched fields. This is slightly slower than
1901 * accessing the fields directly (as an intermediate array is
1902 * generated).
1903 *
1904 * all,f1,f2,f3 = * /(.)(.)(\d+)(\d)/.match("THX1138.")
1905 * all #=> "HX1138"
1906 * f1 #=> "H"
1907 * f2 #=> "X"
1908 * f3 #=> "113"
1909 */
1910
1911 static VALUE
match_to_a(VALUE match)1912 match_to_a(VALUE match)
1913 {
1914 return match_array(match, 0);
1915 }
1916
1917
1918 /*
1919 * call-seq:
1920 * mtch.captures -> array
1921 *
1922 * Returns the array of captures; equivalent to <code>mtch.to_a[1..-1]</code>.
1923 *
1924 * f1,f2,f3,f4 = /(.)(.)(\d+)(\d)/.match("THX1138.").captures
1925 * f1 #=> "H"
1926 * f2 #=> "X"
1927 * f3 #=> "113"
1928 * f4 #=> "8"
1929 */
1930 static VALUE
match_captures(VALUE match)1931 match_captures(VALUE match)
1932 {
1933 return match_array(match, 1);
1934 }
1935
1936 static int
name_to_backref_number(struct re_registers * regs,VALUE regexp,const char * name,const char * name_end)1937 name_to_backref_number(struct re_registers *regs, VALUE regexp, const char* name, const char* name_end)
1938 {
1939 if (NIL_P(regexp)) return -1;
1940 return onig_name_to_backref_number(RREGEXP_PTR(regexp),
1941 (const unsigned char *)name, (const unsigned char *)name_end, regs);
1942 }
1943
1944 NORETURN(static void name_to_backref_error(VALUE name));
1945 static void
name_to_backref_error(VALUE name)1946 name_to_backref_error(VALUE name)
1947 {
1948 rb_raise(rb_eIndexError, "undefined group name reference: % "PRIsVALUE,
1949 name);
1950 }
1951
1952 #define NAME_TO_NUMBER(regs, re, name, name_ptr, name_end) \
1953 (NIL_P(re) ? 0 : \
1954 !rb_enc_compatible(RREGEXP_SRC(re), (name)) ? 0 : \
1955 name_to_backref_number((regs), (re), (name_ptr), (name_end)))
1956
1957 static int
namev_to_backref_number(struct re_registers * regs,VALUE re,VALUE name)1958 namev_to_backref_number(struct re_registers *regs, VALUE re, VALUE name)
1959 {
1960 int num;
1961
1962 if (SYMBOL_P(name)) {
1963 name = rb_sym2str(name);
1964 }
1965 else if (!RB_TYPE_P(name, T_STRING)) {
1966 return -1;
1967 }
1968 num = NAME_TO_NUMBER(regs, re, name,
1969 RSTRING_PTR(name), RSTRING_END(name));
1970 if (num < 1) {
1971 name_to_backref_error(name);
1972 }
1973 return num;
1974 }
1975
1976 static VALUE
match_ary_subseq(VALUE match,long beg,long len,VALUE result)1977 match_ary_subseq(VALUE match, long beg, long len, VALUE result)
1978 {
1979 long olen = RMATCH_REGS(match)->num_regs;
1980 long j, end = olen < beg+len ? olen : beg+len;
1981 if (NIL_P(result)) result = rb_ary_new_capa(len);
1982 if (len == 0) return result;
1983
1984 for (j = beg; j < end; j++) {
1985 rb_ary_push(result, rb_reg_nth_match((int)j, match));
1986 }
1987 if (beg + len > j) {
1988 rb_ary_resize(result, RARRAY_LEN(result) + (beg + len) - j);
1989 }
1990 return result;
1991 }
1992
1993 static VALUE
match_ary_aref(VALUE match,VALUE idx,VALUE result)1994 match_ary_aref(VALUE match, VALUE idx, VALUE result)
1995 {
1996 long beg, len;
1997 int num_regs = RMATCH_REGS(match)->num_regs;
1998
1999 /* check if idx is Range */
2000 switch (rb_range_beg_len(idx, &beg, &len, (long)num_regs, !NIL_P(result))) {
2001 case Qfalse:
2002 if (NIL_P(result)) return rb_reg_nth_match(NUM2INT(idx), match);
2003 rb_ary_push(result, rb_reg_nth_match(NUM2INT(idx), match));
2004 return result;
2005 case Qnil:
2006 return Qnil;
2007 default:
2008 return match_ary_subseq(match, beg, len, result);
2009 }
2010 }
2011
2012 /*
2013 * call-seq:
2014 * mtch[i] -> str or nil
2015 * mtch[start, length] -> array
2016 * mtch[range] -> array
2017 * mtch[name] -> str or nil
2018 *
2019 * Match Reference -- <code>MatchData</code> acts as an array, and may be
2020 * accessed using the normal array indexing techniques. <code>mtch[0]</code>
2021 * is equivalent to the special variable <code>$&</code>, and returns the
2022 * entire matched string. <code>mtch[1]</code>, <code>mtch[2]</code>, and so
2023 * on return the values of the matched backreferences (portions of the
2024 * pattern between parentheses).
2025 *
2026 * m = /(.)(.)(\d+)(\d)/.match("THX1138.")
2027 * m #=> #<MatchData "HX1138" 1:"H" 2:"X" 3:"113" 4:"8">
2028 * m[0] #=> "HX1138"
2029 * m[1, 2] #=> ["H", "X"]
2030 * m[1..3] #=> ["H", "X", "113"]
2031 * m[-3, 2] #=> ["X", "113"]
2032 *
2033 * m = /(?<foo>a+)b/.match("ccaaab")
2034 * m #=> #<MatchData "aaab" foo:"aaa">
2035 * m["foo"] #=> "aaa"
2036 * m[:foo] #=> "aaa"
2037 */
2038
2039 static VALUE
match_aref(int argc,VALUE * argv,VALUE match)2040 match_aref(int argc, VALUE *argv, VALUE match)
2041 {
2042 VALUE idx, length;
2043
2044 match_check(match);
2045 rb_scan_args(argc, argv, "11", &idx, &length);
2046
2047 if (NIL_P(length)) {
2048 if (FIXNUM_P(idx)) {
2049 return rb_reg_nth_match(FIX2INT(idx), match);
2050 }
2051 else {
2052 int num = namev_to_backref_number(RMATCH_REGS(match), RMATCH(match)->regexp, idx);
2053 if (num >= 0) {
2054 return rb_reg_nth_match(num, match);
2055 }
2056 else {
2057 return match_ary_aref(match, idx, Qnil);
2058 }
2059 }
2060 }
2061 else {
2062 long beg = NUM2LONG(idx);
2063 long len = NUM2LONG(length);
2064 long num_regs = RMATCH_REGS(match)->num_regs;
2065 if (len < 0) {
2066 return Qnil;
2067 }
2068 if (beg < 0) {
2069 beg += num_regs;
2070 if (beg < 0) return Qnil;
2071 }
2072 else if (beg > num_regs) {
2073 return Qnil;
2074 }
2075 else if (beg+len > num_regs) {
2076 len = num_regs - beg;
2077 }
2078 return match_ary_subseq(match, beg, len, Qnil);
2079 }
2080 }
2081
2082 /*
2083 * call-seq:
2084 *
2085 * mtch.values_at(index, ...) -> array
2086 *
2087 * Uses each <i>index</i> to access the matching values, returning an array of
2088 * the corresponding matches.
2089 *
2090 * m = /(.)(.)(\d+)(\d)/.match("THX1138: The Movie")
2091 * m.to_a #=> ["HX1138", "H", "X", "113", "8"]
2092 * m.values_at(0, 2, -2) #=> ["HX1138", "X", "113"]
2093 *
2094 * m = /(?<a>\d+) *(?<op>[+\-*\/]) *(?<b>\d+)/.match("1 + 2")
2095 * m.to_a #=> ["1 + 2", "1", "+", "2"]
2096 * m.values_at(:a, :b, :op) #=> ["1", "2", "+"]
2097 */
2098
2099 static VALUE
match_values_at(int argc,VALUE * argv,VALUE match)2100 match_values_at(int argc, VALUE *argv, VALUE match)
2101 {
2102 VALUE result;
2103 int i;
2104
2105 match_check(match);
2106 result = rb_ary_new2(argc);
2107
2108 for (i=0; i<argc; i++) {
2109 if (FIXNUM_P(argv[i])) {
2110 rb_ary_push(result, rb_reg_nth_match(FIX2INT(argv[i]), match));
2111 }
2112 else {
2113 int num = namev_to_backref_number(RMATCH_REGS(match), RMATCH(match)->regexp, argv[i]);
2114 if (num >= 0) {
2115 rb_ary_push(result, rb_reg_nth_match(num, match));
2116 }
2117 else {
2118 match_ary_aref(match, argv[i], result);
2119 }
2120 }
2121 }
2122 return result;
2123 }
2124
2125
2126 /*
2127 * call-seq:
2128 * mtch.to_s -> str
2129 *
2130 * Returns the entire matched string.
2131 *
2132 * m = /(.)(.)(\d+)(\d)/.match("THX1138.")
2133 * m.to_s #=> "HX1138"
2134 */
2135
2136 static VALUE
match_to_s(VALUE match)2137 match_to_s(VALUE match)
2138 {
2139 VALUE str = rb_reg_last_match(match);
2140
2141 match_check(match);
2142 if (NIL_P(str)) str = rb_str_new(0,0);
2143 if (OBJ_TAINTED(match)) OBJ_TAINT(str);
2144 if (OBJ_TAINTED(RMATCH(match)->str)) OBJ_TAINT(str);
2145 return str;
2146 }
2147
2148 static int
match_named_captures_iter(const OnigUChar * name,const OnigUChar * name_end,int back_num,int * back_refs,OnigRegex regex,void * arg)2149 match_named_captures_iter(const OnigUChar *name, const OnigUChar *name_end,
2150 int back_num, int *back_refs, OnigRegex regex, void *arg) {
2151 struct MEMO *memo = MEMO_CAST(arg);
2152 VALUE hash = memo->v1;
2153 VALUE match = memo->v2;
2154
2155 VALUE key = rb_enc_str_new((const char *)name, name_end-name, regex->enc);
2156 VALUE value;
2157
2158 int i;
2159 int found = 0;
2160
2161 for (i = 0; i < back_num; i++) {
2162 value = rb_reg_nth_match(back_refs[i], match);
2163 if (RTEST(value)) {
2164 rb_hash_aset(hash, key, value);
2165 found = 1;
2166 }
2167 }
2168
2169 if (found == 0) {
2170 rb_hash_aset(hash, key, Qnil);
2171 }
2172
2173 return 0;
2174 }
2175
2176 /*
2177 * call-seq:
2178 * mtch.named_captures -> hash
2179 *
2180 * Returns a Hash using named capture.
2181 *
2182 * A key of the hash is a name of the named captures.
2183 * A value of the hash is a string of last successful capture of corresponding
2184 * group.
2185 *
2186 * m = /(?<a>.)(?<b>.)/.match("01")
2187 * m.named_captures #=> {"a" => "0", "b" => "1"}
2188 *
2189 * m = /(?<a>.)(?<b>.)?/.match("0")
2190 * m.named_captures #=> {"a" => "0", "b" => nil}
2191 *
2192 * m = /(?<a>.)(?<a>.)/.match("01")
2193 * m.named_captures #=> {"a" => "1"}
2194 *
2195 * m = /(?<a>x)|(?<a>y)/.match("x")
2196 * m.named_captures #=> {"a" => "x"}
2197 */
2198
2199 static VALUE
match_named_captures(VALUE match)2200 match_named_captures(VALUE match)
2201 {
2202 VALUE hash;
2203 struct MEMO *memo;
2204
2205 match_check(match);
2206 if (NIL_P(RMATCH(match)->regexp))
2207 return rb_hash_new();
2208
2209 hash = rb_hash_new();
2210 memo = MEMO_NEW(hash, match, 0);
2211
2212 onig_foreach_name(RREGEXP(RMATCH(match)->regexp)->ptr, match_named_captures_iter, (void*)memo);
2213
2214 return hash;
2215 }
2216
2217 /*
2218 * call-seq:
2219 * mtch.string -> str
2220 *
2221 * Returns a frozen copy of the string passed in to <code>match</code>.
2222 *
2223 * m = /(.)(.)(\d+)(\d)/.match("THX1138.")
2224 * m.string #=> "THX1138."
2225 */
2226
2227 static VALUE
match_string(VALUE match)2228 match_string(VALUE match)
2229 {
2230 match_check(match);
2231 return RMATCH(match)->str; /* str is frozen */
2232 }
2233
2234 struct backref_name_tag {
2235 const UChar *name;
2236 long len;
2237 };
2238
2239 static int
match_inspect_name_iter(const OnigUChar * name,const OnigUChar * name_end,int back_num,int * back_refs,OnigRegex regex,void * arg0)2240 match_inspect_name_iter(const OnigUChar *name, const OnigUChar *name_end,
2241 int back_num, int *back_refs, OnigRegex regex, void *arg0)
2242 {
2243 struct backref_name_tag *arg = (struct backref_name_tag *)arg0;
2244 int i;
2245
2246 for (i = 0; i < back_num; i++) {
2247 arg[back_refs[i]].name = name;
2248 arg[back_refs[i]].len = name_end - name;
2249 }
2250 return 0;
2251 }
2252
2253 /*
2254 * call-seq:
2255 * mtch.inspect -> str
2256 *
2257 * Returns a printable version of <i>mtch</i>.
2258 *
2259 * puts /.$/.match("foo").inspect
2260 * #=> #<MatchData "o">
2261 *
2262 * puts /(.)(.)(.)/.match("foo").inspect
2263 * #=> #<MatchData "foo" 1:"f" 2:"o" 3:"o">
2264 *
2265 * puts /(.)(.)?(.)/.match("fo").inspect
2266 * #=> #<MatchData "fo" 1:"f" 2:nil 3:"o">
2267 *
2268 * puts /(?<foo>.)(?<bar>.)(?<baz>.)/.match("hoge").inspect
2269 * #=> #<MatchData "hog" foo:"h" bar:"o" baz:"g">
2270 *
2271 */
2272
2273 static VALUE
match_inspect(VALUE match)2274 match_inspect(VALUE match)
2275 {
2276 VALUE cname = rb_class_path(rb_obj_class(match));
2277 VALUE str;
2278 int i;
2279 struct re_registers *regs = RMATCH_REGS(match);
2280 int num_regs = regs->num_regs;
2281 struct backref_name_tag *names;
2282 VALUE regexp = RMATCH(match)->regexp;
2283
2284 if (regexp == 0) {
2285 return rb_sprintf("#<%"PRIsVALUE":%p>", cname, (void*)match);
2286 }
2287 else if (NIL_P(regexp)) {
2288 return rb_sprintf("#<%"PRIsVALUE": %"PRIsVALUE">",
2289 cname, rb_reg_nth_match(0, match));
2290 }
2291
2292 names = ALLOCA_N(struct backref_name_tag, num_regs);
2293 MEMZERO(names, struct backref_name_tag, num_regs);
2294
2295 onig_foreach_name(RREGEXP_PTR(regexp),
2296 match_inspect_name_iter, names);
2297
2298 str = rb_str_buf_new2("#<");
2299 rb_str_append(str, cname);
2300
2301 for (i = 0; i < num_regs; i++) {
2302 VALUE v;
2303 rb_str_buf_cat2(str, " ");
2304 if (0 < i) {
2305 if (names[i].name)
2306 rb_str_buf_cat(str, (const char *)names[i].name, names[i].len);
2307 else {
2308 rb_str_catf(str, "%d", i);
2309 }
2310 rb_str_buf_cat2(str, ":");
2311 }
2312 v = rb_reg_nth_match(i, match);
2313 if (v == Qnil)
2314 rb_str_buf_cat2(str, "nil");
2315 else
2316 rb_str_buf_append(str, rb_str_inspect(v));
2317 }
2318 rb_str_buf_cat2(str, ">");
2319
2320 return str;
2321 }
2322
2323 VALUE rb_cRegexp;
2324
2325 static int
read_escaped_byte(const char ** pp,const char * end,onig_errmsg_buffer err)2326 read_escaped_byte(const char **pp, const char *end, onig_errmsg_buffer err)
2327 {
2328 const char *p = *pp;
2329 int code;
2330 int meta_prefix = 0, ctrl_prefix = 0;
2331 size_t len;
2332
2333 if (p == end || *p++ != '\\') {
2334 errcpy(err, "too short escaped multibyte character");
2335 return -1;
2336 }
2337
2338 again:
2339 if (p == end) {
2340 errcpy(err, "too short escape sequence");
2341 return -1;
2342 }
2343 switch (*p++) {
2344 case '\\': code = '\\'; break;
2345 case 'n': code = '\n'; break;
2346 case 't': code = '\t'; break;
2347 case 'r': code = '\r'; break;
2348 case 'f': code = '\f'; break;
2349 case 'v': code = '\013'; break;
2350 case 'a': code = '\007'; break;
2351 case 'e': code = '\033'; break;
2352
2353 /* \OOO */
2354 case '0': case '1': case '2': case '3':
2355 case '4': case '5': case '6': case '7':
2356 p--;
2357 code = scan_oct(p, end < p+3 ? end-p : 3, &len);
2358 p += len;
2359 break;
2360
2361 case 'x': /* \xHH */
2362 code = scan_hex(p, end < p+2 ? end-p : 2, &len);
2363 if (len < 1) {
2364 errcpy(err, "invalid hex escape");
2365 return -1;
2366 }
2367 p += len;
2368 break;
2369
2370 case 'M': /* \M-X, \M-\C-X, \M-\cX */
2371 if (meta_prefix) {
2372 errcpy(err, "duplicate meta escape");
2373 return -1;
2374 }
2375 meta_prefix = 1;
2376 if (p+1 < end && *p++ == '-' && (*p & 0x80) == 0) {
2377 if (*p == '\\') {
2378 p++;
2379 goto again;
2380 }
2381 else {
2382 code = *p++;
2383 break;
2384 }
2385 }
2386 errcpy(err, "too short meta escape");
2387 return -1;
2388
2389 case 'C': /* \C-X, \C-\M-X */
2390 if (p == end || *p++ != '-') {
2391 errcpy(err, "too short control escape");
2392 return -1;
2393 }
2394 case 'c': /* \cX, \c\M-X */
2395 if (ctrl_prefix) {
2396 errcpy(err, "duplicate control escape");
2397 return -1;
2398 }
2399 ctrl_prefix = 1;
2400 if (p < end && (*p & 0x80) == 0) {
2401 if (*p == '\\') {
2402 p++;
2403 goto again;
2404 }
2405 else {
2406 code = *p++;
2407 break;
2408 }
2409 }
2410 errcpy(err, "too short control escape");
2411 return -1;
2412
2413 default:
2414 errcpy(err, "unexpected escape sequence");
2415 return -1;
2416 }
2417 if (code < 0 || 0xff < code) {
2418 errcpy(err, "invalid escape code");
2419 return -1;
2420 }
2421
2422 if (ctrl_prefix)
2423 code &= 0x1f;
2424 if (meta_prefix)
2425 code |= 0x80;
2426
2427 *pp = p;
2428 return code;
2429 }
2430
2431 static int
unescape_escaped_nonascii(const char ** pp,const char * end,rb_encoding * enc,VALUE buf,rb_encoding ** encp,onig_errmsg_buffer err)2432 unescape_escaped_nonascii(const char **pp, const char *end, rb_encoding *enc,
2433 VALUE buf, rb_encoding **encp, onig_errmsg_buffer err)
2434 {
2435 const char *p = *pp;
2436 int chmaxlen = rb_enc_mbmaxlen(enc);
2437 unsigned char *area = ALLOCA_N(unsigned char, chmaxlen);
2438 char *chbuf = (char *)area;
2439 int chlen = 0;
2440 int byte;
2441 int l;
2442
2443 memset(chbuf, 0, chmaxlen);
2444
2445 byte = read_escaped_byte(&p, end, err);
2446 if (byte == -1) {
2447 return -1;
2448 }
2449
2450 area[chlen++] = byte;
2451 while (chlen < chmaxlen &&
2452 MBCLEN_NEEDMORE_P(rb_enc_precise_mbclen(chbuf, chbuf+chlen, enc))) {
2453 byte = read_escaped_byte(&p, end, err);
2454 if (byte == -1) {
2455 return -1;
2456 }
2457 area[chlen++] = byte;
2458 }
2459
2460 l = rb_enc_precise_mbclen(chbuf, chbuf+chlen, enc);
2461 if (MBCLEN_INVALID_P(l)) {
2462 errcpy(err, "invalid multibyte escape");
2463 return -1;
2464 }
2465 if (1 < chlen || (area[0] & 0x80)) {
2466 rb_str_buf_cat(buf, chbuf, chlen);
2467
2468 if (*encp == 0)
2469 *encp = enc;
2470 else if (*encp != enc) {
2471 errcpy(err, "escaped non ASCII character in UTF-8 regexp");
2472 return -1;
2473 }
2474 }
2475 else {
2476 char escbuf[5];
2477 snprintf(escbuf, sizeof(escbuf), "\\x%02X", area[0]&0xff);
2478 rb_str_buf_cat(buf, escbuf, 4);
2479 }
2480 *pp = p;
2481 return 0;
2482 }
2483
2484 static int
check_unicode_range(unsigned long code,onig_errmsg_buffer err)2485 check_unicode_range(unsigned long code, onig_errmsg_buffer err)
2486 {
2487 if ((0xd800 <= code && code <= 0xdfff) || /* Surrogates */
2488 0x10ffff < code) {
2489 errcpy(err, "invalid Unicode range");
2490 return -1;
2491 }
2492 return 0;
2493 }
2494
2495 static int
append_utf8(unsigned long uv,VALUE buf,rb_encoding ** encp,onig_errmsg_buffer err)2496 append_utf8(unsigned long uv,
2497 VALUE buf, rb_encoding **encp, onig_errmsg_buffer err)
2498 {
2499 if (check_unicode_range(uv, err) != 0)
2500 return -1;
2501 if (uv < 0x80) {
2502 char escbuf[5];
2503 snprintf(escbuf, sizeof(escbuf), "\\x%02X", (int)uv);
2504 rb_str_buf_cat(buf, escbuf, 4);
2505 }
2506 else {
2507 int len;
2508 char utf8buf[6];
2509 len = rb_uv_to_utf8(utf8buf, uv);
2510 rb_str_buf_cat(buf, utf8buf, len);
2511
2512 if (*encp == 0)
2513 *encp = rb_utf8_encoding();
2514 else if (*encp != rb_utf8_encoding()) {
2515 errcpy(err, "UTF-8 character in non UTF-8 regexp");
2516 return -1;
2517 }
2518 }
2519 return 0;
2520 }
2521
2522 static int
unescape_unicode_list(const char ** pp,const char * end,VALUE buf,rb_encoding ** encp,onig_errmsg_buffer err)2523 unescape_unicode_list(const char **pp, const char *end,
2524 VALUE buf, rb_encoding **encp, onig_errmsg_buffer err)
2525 {
2526 const char *p = *pp;
2527 int has_unicode = 0;
2528 unsigned long code;
2529 size_t len;
2530
2531 while (p < end && ISSPACE(*p)) p++;
2532
2533 while (1) {
2534 code = ruby_scan_hex(p, end-p, &len);
2535 if (len == 0)
2536 break;
2537 if (6 < len) { /* max 10FFFF */
2538 errcpy(err, "invalid Unicode range");
2539 return -1;
2540 }
2541 p += len;
2542 if (append_utf8(code, buf, encp, err) != 0)
2543 return -1;
2544 has_unicode = 1;
2545
2546 while (p < end && ISSPACE(*p)) p++;
2547 }
2548
2549 if (has_unicode == 0) {
2550 errcpy(err, "invalid Unicode list");
2551 return -1;
2552 }
2553
2554 *pp = p;
2555
2556 return 0;
2557 }
2558
2559 static int
unescape_unicode_bmp(const char ** pp,const char * end,VALUE buf,rb_encoding ** encp,onig_errmsg_buffer err)2560 unescape_unicode_bmp(const char **pp, const char *end,
2561 VALUE buf, rb_encoding **encp, onig_errmsg_buffer err)
2562 {
2563 const char *p = *pp;
2564 size_t len;
2565 unsigned long code;
2566
2567 if (end < p+4) {
2568 errcpy(err, "invalid Unicode escape");
2569 return -1;
2570 }
2571 code = ruby_scan_hex(p, 4, &len);
2572 if (len != 4) {
2573 errcpy(err, "invalid Unicode escape");
2574 return -1;
2575 }
2576 if (append_utf8(code, buf, encp, err) != 0)
2577 return -1;
2578 *pp = p + 4;
2579 return 0;
2580 }
2581
2582 static int
unescape_nonascii(const char * p,const char * end,rb_encoding * enc,VALUE buf,rb_encoding ** encp,int * has_property,onig_errmsg_buffer err)2583 unescape_nonascii(const char *p, const char *end, rb_encoding *enc,
2584 VALUE buf, rb_encoding **encp, int *has_property,
2585 onig_errmsg_buffer err)
2586 {
2587 unsigned char c;
2588 char smallbuf[2];
2589
2590 while (p < end) {
2591 int chlen = rb_enc_precise_mbclen(p, end, enc);
2592 if (!MBCLEN_CHARFOUND_P(chlen)) {
2593 invalid_multibyte:
2594 errcpy(err, "invalid multibyte character");
2595 return -1;
2596 }
2597 chlen = MBCLEN_CHARFOUND_LEN(chlen);
2598 if (1 < chlen || (*p & 0x80)) {
2599 multibyte:
2600 rb_str_buf_cat(buf, p, chlen);
2601 p += chlen;
2602 if (*encp == 0)
2603 *encp = enc;
2604 else if (*encp != enc) {
2605 errcpy(err, "non ASCII character in UTF-8 regexp");
2606 return -1;
2607 }
2608 continue;
2609 }
2610
2611 switch (c = *p++) {
2612 case '\\':
2613 if (p == end) {
2614 errcpy(err, "too short escape sequence");
2615 return -1;
2616 }
2617 chlen = rb_enc_precise_mbclen(p, end, enc);
2618 if (!MBCLEN_CHARFOUND_P(chlen)) {
2619 goto invalid_multibyte;
2620 }
2621 if ((chlen = MBCLEN_CHARFOUND_LEN(chlen)) > 1) {
2622 /* include the previous backslash */
2623 --p;
2624 ++chlen;
2625 goto multibyte;
2626 }
2627 switch (c = *p++) {
2628 case '1': case '2': case '3':
2629 case '4': case '5': case '6': case '7': /* \O, \OO, \OOO or backref */
2630 {
2631 size_t len = end-(p-1), octlen;
2632 if (ruby_scan_oct(p-1, len < 3 ? len : 3, &octlen) <= 0177) {
2633 /* backref or 7bit octal.
2634 no need to unescape anyway.
2635 re-escaping may break backref */
2636 goto escape_asis;
2637 }
2638 }
2639 /* xxx: How about more than 199 subexpressions? */
2640
2641 case '0': /* \0, \0O, \0OO */
2642
2643 case 'x': /* \xHH */
2644 case 'c': /* \cX, \c\M-X */
2645 case 'C': /* \C-X, \C-\M-X */
2646 case 'M': /* \M-X, \M-\C-X, \M-\cX */
2647 p = p-2;
2648 if (enc == rb_usascii_encoding()) {
2649 const char *pbeg = p;
2650 int byte = read_escaped_byte(&p, end, err);
2651 if (byte == -1) return -1;
2652 c = byte;
2653 rb_str_buf_cat(buf, pbeg, p-pbeg);
2654 }
2655 else {
2656 if (unescape_escaped_nonascii(&p, end, enc, buf, encp, err) != 0)
2657 return -1;
2658 }
2659 break;
2660
2661 case 'u':
2662 if (p == end) {
2663 errcpy(err, "too short escape sequence");
2664 return -1;
2665 }
2666 if (*p == '{') {
2667 /* \u{H HH HHH HHHH HHHHH HHHHHH ...} */
2668 p++;
2669 if (unescape_unicode_list(&p, end, buf, encp, err) != 0)
2670 return -1;
2671 if (p == end || *p++ != '}') {
2672 errcpy(err, "invalid Unicode list");
2673 return -1;
2674 }
2675 break;
2676 }
2677 else {
2678 /* \uHHHH */
2679 if (unescape_unicode_bmp(&p, end, buf, encp, err) != 0)
2680 return -1;
2681 break;
2682 }
2683
2684 case 'p': /* \p{Hiragana} */
2685 case 'P':
2686 if (!*encp) {
2687 *has_property = 1;
2688 }
2689 goto escape_asis;
2690
2691 default: /* \n, \\, \d, \9, etc. */
2692 escape_asis:
2693 smallbuf[0] = '\\';
2694 smallbuf[1] = c;
2695 rb_str_buf_cat(buf, smallbuf, 2);
2696 break;
2697 }
2698 break;
2699
2700 default:
2701 rb_str_buf_cat(buf, (char *)&c, 1);
2702 break;
2703 }
2704 }
2705
2706 return 0;
2707 }
2708
2709 static VALUE
rb_reg_preprocess(const char * p,const char * end,rb_encoding * enc,rb_encoding ** fixed_enc,onig_errmsg_buffer err)2710 rb_reg_preprocess(const char *p, const char *end, rb_encoding *enc,
2711 rb_encoding **fixed_enc, onig_errmsg_buffer err)
2712 {
2713 VALUE buf;
2714 int has_property = 0;
2715
2716 buf = rb_str_buf_new(0);
2717
2718 if (rb_enc_asciicompat(enc))
2719 *fixed_enc = 0;
2720 else {
2721 *fixed_enc = enc;
2722 rb_enc_associate(buf, enc);
2723 }
2724
2725 if (unescape_nonascii(p, end, enc, buf, fixed_enc, &has_property, err) != 0)
2726 return Qnil;
2727
2728 if (has_property && !*fixed_enc) {
2729 *fixed_enc = enc;
2730 }
2731
2732 if (*fixed_enc) {
2733 rb_enc_associate(buf, *fixed_enc);
2734 }
2735
2736 return buf;
2737 }
2738
2739 VALUE
rb_reg_check_preprocess(VALUE str)2740 rb_reg_check_preprocess(VALUE str)
2741 {
2742 rb_encoding *fixed_enc = 0;
2743 onig_errmsg_buffer err = "";
2744 VALUE buf;
2745 char *p, *end;
2746 rb_encoding *enc;
2747
2748 StringValue(str);
2749 p = RSTRING_PTR(str);
2750 end = p + RSTRING_LEN(str);
2751 enc = rb_enc_get(str);
2752
2753 buf = rb_reg_preprocess(p, end, enc, &fixed_enc, err);
2754 RB_GC_GUARD(str);
2755
2756 if (buf == Qnil) {
2757 return rb_reg_error_desc(str, 0, err);
2758 }
2759 return Qnil;
2760 }
2761
2762 static VALUE
rb_reg_preprocess_dregexp(VALUE ary,int options)2763 rb_reg_preprocess_dregexp(VALUE ary, int options)
2764 {
2765 rb_encoding *fixed_enc = 0;
2766 rb_encoding *regexp_enc = 0;
2767 onig_errmsg_buffer err = "";
2768 int i;
2769 VALUE result = 0;
2770 rb_encoding *ascii8bit = rb_ascii8bit_encoding();
2771
2772 if (RARRAY_LEN(ary) == 0) {
2773 rb_raise(rb_eArgError, "no arguments given");
2774 }
2775
2776 for (i = 0; i < RARRAY_LEN(ary); i++) {
2777 VALUE str = RARRAY_AREF(ary, i);
2778 VALUE buf;
2779 char *p, *end;
2780 rb_encoding *src_enc;
2781
2782 src_enc = rb_enc_get(str);
2783 if (options & ARG_ENCODING_NONE &&
2784 src_enc != ascii8bit) {
2785 if (str_coderange(str) != ENC_CODERANGE_7BIT)
2786 rb_raise(rb_eRegexpError, "/.../n has a non escaped non ASCII character in non ASCII-8BIT script");
2787 else
2788 src_enc = ascii8bit;
2789 }
2790
2791 StringValue(str);
2792 p = RSTRING_PTR(str);
2793 end = p + RSTRING_LEN(str);
2794
2795 buf = rb_reg_preprocess(p, end, src_enc, &fixed_enc, err);
2796
2797 if (buf == Qnil)
2798 rb_raise(rb_eArgError, "%s", err);
2799
2800 if (fixed_enc != 0) {
2801 if (regexp_enc != 0 && regexp_enc != fixed_enc) {
2802 rb_raise(rb_eRegexpError, "encoding mismatch in dynamic regexp : %s and %s",
2803 rb_enc_name(regexp_enc), rb_enc_name(fixed_enc));
2804 }
2805 regexp_enc = fixed_enc;
2806 }
2807
2808 if (!result)
2809 result = rb_str_new3(str);
2810 else
2811 rb_str_buf_append(result, str);
2812 }
2813 if (regexp_enc) {
2814 rb_enc_associate(result, regexp_enc);
2815 }
2816
2817 return result;
2818 }
2819
2820 static int
rb_reg_initialize(VALUE obj,const char * s,long len,rb_encoding * enc,int options,onig_errmsg_buffer err,const char * sourcefile,int sourceline)2821 rb_reg_initialize(VALUE obj, const char *s, long len, rb_encoding *enc,
2822 int options, onig_errmsg_buffer err,
2823 const char *sourcefile, int sourceline)
2824 {
2825 struct RRegexp *re = RREGEXP(obj);
2826 VALUE unescaped;
2827 rb_encoding *fixed_enc = 0;
2828 rb_encoding *a_enc = rb_ascii8bit_encoding();
2829
2830 rb_check_frozen(obj);
2831 if (FL_TEST(obj, REG_LITERAL))
2832 rb_raise(rb_eSecurityError, "can't modify literal regexp");
2833 if (re->ptr)
2834 rb_raise(rb_eTypeError, "already initialized regexp");
2835 re->ptr = 0;
2836
2837 if (rb_enc_dummy_p(enc)) {
2838 errcpy(err, "can't make regexp with dummy encoding");
2839 return -1;
2840 }
2841
2842 unescaped = rb_reg_preprocess(s, s+len, enc, &fixed_enc, err);
2843 if (unescaped == Qnil)
2844 return -1;
2845
2846 if (fixed_enc) {
2847 if ((fixed_enc != enc && (options & ARG_ENCODING_FIXED)) ||
2848 (fixed_enc != a_enc && (options & ARG_ENCODING_NONE))) {
2849 errcpy(err, "incompatible character encoding");
2850 return -1;
2851 }
2852 if (fixed_enc != a_enc) {
2853 options |= ARG_ENCODING_FIXED;
2854 enc = fixed_enc;
2855 }
2856 }
2857 else if (!(options & ARG_ENCODING_FIXED)) {
2858 enc = rb_usascii_encoding();
2859 }
2860
2861 rb_enc_associate((VALUE)re, enc);
2862 if ((options & ARG_ENCODING_FIXED) || fixed_enc) {
2863 re->basic.flags |= KCODE_FIXED;
2864 }
2865 if (options & ARG_ENCODING_NONE) {
2866 re->basic.flags |= REG_ENCODING_NONE;
2867 }
2868
2869 re->ptr = make_regexp(RSTRING_PTR(unescaped), RSTRING_LEN(unescaped), enc,
2870 options & ARG_REG_OPTION_MASK, err,
2871 sourcefile, sourceline);
2872 if (!re->ptr) return -1;
2873 RB_GC_GUARD(unescaped);
2874 return 0;
2875 }
2876
2877 static void
reg_set_source(VALUE reg,VALUE str,rb_encoding * enc)2878 reg_set_source(VALUE reg, VALUE str, rb_encoding *enc)
2879 {
2880 rb_encoding *regenc = rb_enc_get(reg);
2881 if (regenc != enc) {
2882 str = rb_enc_associate(rb_str_dup(str), enc = regenc);
2883 }
2884 RB_OBJ_WRITE(reg, &RREGEXP(reg)->src, rb_fstring(str));
2885 }
2886
2887 static int
rb_reg_initialize_str(VALUE obj,VALUE str,int options,onig_errmsg_buffer err,const char * sourcefile,int sourceline)2888 rb_reg_initialize_str(VALUE obj, VALUE str, int options, onig_errmsg_buffer err,
2889 const char *sourcefile, int sourceline)
2890 {
2891 int ret;
2892 rb_encoding *str_enc = rb_enc_get(str), *enc = str_enc;
2893 if (options & ARG_ENCODING_NONE) {
2894 rb_encoding *ascii8bit = rb_ascii8bit_encoding();
2895 if (enc != ascii8bit) {
2896 if (str_coderange(str) != ENC_CODERANGE_7BIT) {
2897 errcpy(err, "/.../n has a non escaped non ASCII character in non ASCII-8BIT script");
2898 return -1;
2899 }
2900 enc = ascii8bit;
2901 }
2902 }
2903 ret = rb_reg_initialize(obj, RSTRING_PTR(str), RSTRING_LEN(str), enc,
2904 options, err, sourcefile, sourceline);
2905 OBJ_INFECT(obj, str);
2906 if (ret == 0) reg_set_source(obj, str, str_enc);
2907 return ret;
2908 }
2909
2910 static VALUE
rb_reg_s_alloc(VALUE klass)2911 rb_reg_s_alloc(VALUE klass)
2912 {
2913 NEWOBJ_OF(re, struct RRegexp, klass, T_REGEXP | (RGENGC_WB_PROTECTED_REGEXP ? FL_WB_PROTECTED : 0));
2914
2915 re->ptr = 0;
2916 RB_OBJ_WRITE(re, &re->src, 0);
2917 re->usecnt = 0;
2918
2919 return (VALUE)re;
2920 }
2921
2922 VALUE
rb_reg_alloc(void)2923 rb_reg_alloc(void)
2924 {
2925 return rb_reg_s_alloc(rb_cRegexp);
2926 }
2927
2928 VALUE
rb_reg_new_str(VALUE s,int options)2929 rb_reg_new_str(VALUE s, int options)
2930 {
2931 return rb_reg_init_str(rb_reg_alloc(), s, options);
2932 }
2933
2934 VALUE
rb_reg_init_str(VALUE re,VALUE s,int options)2935 rb_reg_init_str(VALUE re, VALUE s, int options)
2936 {
2937 onig_errmsg_buffer err = "";
2938
2939 if (rb_reg_initialize_str(re, s, options, err, NULL, 0) != 0) {
2940 rb_reg_raise_str(s, options, err);
2941 }
2942
2943 return re;
2944 }
2945
2946 static VALUE
rb_reg_init_str_enc(VALUE re,VALUE s,rb_encoding * enc,int options)2947 rb_reg_init_str_enc(VALUE re, VALUE s, rb_encoding *enc, int options)
2948 {
2949 onig_errmsg_buffer err = "";
2950
2951 if (rb_reg_initialize(re, RSTRING_PTR(s), RSTRING_LEN(s),
2952 enc, options, err, NULL, 0) != 0) {
2953 rb_reg_raise_str(s, options, err);
2954 }
2955 reg_set_source(re, s, enc);
2956
2957 return re;
2958 }
2959
2960 MJIT_FUNC_EXPORTED VALUE
rb_reg_new_ary(VALUE ary,int opt)2961 rb_reg_new_ary(VALUE ary, int opt)
2962 {
2963 return rb_reg_new_str(rb_reg_preprocess_dregexp(ary, opt), opt);
2964 }
2965
2966 VALUE
rb_enc_reg_new(const char * s,long len,rb_encoding * enc,int options)2967 rb_enc_reg_new(const char *s, long len, rb_encoding *enc, int options)
2968 {
2969 VALUE re = rb_reg_alloc();
2970 onig_errmsg_buffer err = "";
2971
2972 if (rb_reg_initialize(re, s, len, enc, options, err, NULL, 0) != 0) {
2973 rb_enc_reg_raise(s, len, enc, options, err);
2974 }
2975 RB_OBJ_WRITE(re, &RREGEXP(re)->src, rb_fstring(rb_enc_str_new(s, len, enc)));
2976
2977 return re;
2978 }
2979
2980 VALUE
rb_reg_new(const char * s,long len,int options)2981 rb_reg_new(const char *s, long len, int options)
2982 {
2983 return rb_enc_reg_new(s, len, rb_ascii8bit_encoding(), options);
2984 }
2985
2986 VALUE
rb_reg_compile(VALUE str,int options,const char * sourcefile,int sourceline)2987 rb_reg_compile(VALUE str, int options, const char *sourcefile, int sourceline)
2988 {
2989 VALUE re = rb_reg_alloc();
2990 onig_errmsg_buffer err = "";
2991
2992 if (!str) str = rb_str_new(0,0);
2993 if (rb_reg_initialize_str(re, str, options, err, sourcefile, sourceline) != 0) {
2994 rb_set_errinfo(rb_reg_error_desc(str, options, err));
2995 return Qnil;
2996 }
2997 FL_SET(re, REG_LITERAL);
2998 return re;
2999 }
3000
3001 static VALUE reg_cache;
3002
3003 VALUE
rb_reg_regcomp(VALUE str)3004 rb_reg_regcomp(VALUE str)
3005 {
3006 if (reg_cache && RREGEXP_SRC_LEN(reg_cache) == RSTRING_LEN(str)
3007 && ENCODING_GET(reg_cache) == ENCODING_GET(str)
3008 && memcmp(RREGEXP_SRC_PTR(reg_cache), RSTRING_PTR(str), RSTRING_LEN(str)) == 0)
3009 return reg_cache;
3010
3011 return reg_cache = rb_reg_new_str(str, 0);
3012 }
3013
3014 static st_index_t reg_hash(VALUE re);
3015 /*
3016 * call-seq:
3017 * rxp.hash -> integer
3018 *
3019 * Produce a hash based on the text and options of this regular expression.
3020 *
3021 * See also Object#hash.
3022 */
3023
3024 static VALUE
rb_reg_hash(VALUE re)3025 rb_reg_hash(VALUE re)
3026 {
3027 st_index_t hashval = reg_hash(re);
3028 return ST2FIX(hashval);
3029 }
3030
3031 static st_index_t
reg_hash(VALUE re)3032 reg_hash(VALUE re)
3033 {
3034 st_index_t hashval;
3035
3036 rb_reg_check(re);
3037 hashval = RREGEXP_PTR(re)->options;
3038 hashval = rb_hash_uint(hashval, rb_memhash(RREGEXP_SRC_PTR(re), RREGEXP_SRC_LEN(re)));
3039 return rb_hash_end(hashval);
3040 }
3041
3042
3043 /*
3044 * call-seq:
3045 * rxp == other_rxp -> true or false
3046 * rxp.eql?(other_rxp) -> true or false
3047 *
3048 * Equality---Two regexps are equal if their patterns are identical, they have
3049 * the same character set code, and their <code>casefold?</code> values are the
3050 * same.
3051 *
3052 * /abc/ == /abc/x #=> false
3053 * /abc/ == /abc/i #=> false
3054 * /abc/ == /abc/u #=> false
3055 * /abc/u == /abc/n #=> false
3056 */
3057
3058 static VALUE
rb_reg_equal(VALUE re1,VALUE re2)3059 rb_reg_equal(VALUE re1, VALUE re2)
3060 {
3061 if (re1 == re2) return Qtrue;
3062 if (!RB_TYPE_P(re2, T_REGEXP)) return Qfalse;
3063 rb_reg_check(re1); rb_reg_check(re2);
3064 if (FL_TEST(re1, KCODE_FIXED) != FL_TEST(re2, KCODE_FIXED)) return Qfalse;
3065 if (RREGEXP_PTR(re1)->options != RREGEXP_PTR(re2)->options) return Qfalse;
3066 if (RREGEXP_SRC_LEN(re1) != RREGEXP_SRC_LEN(re2)) return Qfalse;
3067 if (ENCODING_GET(re1) != ENCODING_GET(re2)) return Qfalse;
3068 if (memcmp(RREGEXP_SRC_PTR(re1), RREGEXP_SRC_PTR(re2), RREGEXP_SRC_LEN(re1)) == 0) {
3069 return Qtrue;
3070 }
3071 return Qfalse;
3072 }
3073
3074 /*
3075 * call-seq:
3076 * mtch.hash -> integer
3077 *
3078 * Produce a hash based on the target string, regexp and matched
3079 * positions of this matchdata.
3080 *
3081 * See also Object#hash.
3082 */
3083
3084 static VALUE
match_hash(VALUE match)3085 match_hash(VALUE match)
3086 {
3087 const struct re_registers *regs;
3088 st_index_t hashval;
3089
3090 match_check(match);
3091 hashval = rb_hash_start(rb_str_hash(RMATCH(match)->str));
3092 hashval = rb_hash_uint(hashval, reg_hash(match_regexp(match)));
3093 regs = RMATCH_REGS(match);
3094 hashval = rb_hash_uint(hashval, regs->num_regs);
3095 hashval = rb_hash_uint(hashval, rb_memhash(regs->beg, regs->num_regs * sizeof(*regs->beg)));
3096 hashval = rb_hash_uint(hashval, rb_memhash(regs->end, regs->num_regs * sizeof(*regs->end)));
3097 hashval = rb_hash_end(hashval);
3098 return ST2FIX(hashval);
3099 }
3100
3101 /*
3102 * call-seq:
3103 * mtch == mtch2 -> true or false
3104 * mtch.eql?(mtch2) -> true or false
3105 *
3106 * Equality---Two matchdata are equal if their target strings,
3107 * patterns, and matched positions are identical.
3108 */
3109
3110 static VALUE
match_equal(VALUE match1,VALUE match2)3111 match_equal(VALUE match1, VALUE match2)
3112 {
3113 const struct re_registers *regs1, *regs2;
3114
3115 if (match1 == match2) return Qtrue;
3116 if (!RB_TYPE_P(match2, T_MATCH)) return Qfalse;
3117 if (!RMATCH(match1)->regexp || !RMATCH(match2)->regexp) return Qfalse;
3118 if (!rb_str_equal(RMATCH(match1)->str, RMATCH(match2)->str)) return Qfalse;
3119 if (!rb_reg_equal(match_regexp(match1), match_regexp(match2))) return Qfalse;
3120 regs1 = RMATCH_REGS(match1);
3121 regs2 = RMATCH_REGS(match2);
3122 if (regs1->num_regs != regs2->num_regs) return Qfalse;
3123 if (memcmp(regs1->beg, regs2->beg, regs1->num_regs * sizeof(*regs1->beg))) return Qfalse;
3124 if (memcmp(regs1->end, regs2->end, regs1->num_regs * sizeof(*regs1->end))) return Qfalse;
3125 return Qtrue;
3126 }
3127
3128 static VALUE
reg_operand(VALUE s,int check)3129 reg_operand(VALUE s, int check)
3130 {
3131 if (SYMBOL_P(s)) {
3132 return rb_sym2str(s);
3133 }
3134 else if (RB_TYPE_P(s, T_STRING)) {
3135 return s;
3136 }
3137 else {
3138 return check ? rb_str_to_str(s) : rb_check_string_type(s);
3139 }
3140 }
3141
3142 static long
reg_match_pos(VALUE re,VALUE * strp,long pos)3143 reg_match_pos(VALUE re, VALUE *strp, long pos)
3144 {
3145 VALUE str = *strp;
3146
3147 if (NIL_P(str)) {
3148 rb_backref_set(Qnil);
3149 return -1;
3150 }
3151 *strp = str = reg_operand(str, TRUE);
3152 if (pos != 0) {
3153 if (pos < 0) {
3154 VALUE l = rb_str_length(str);
3155 pos += NUM2INT(l);
3156 if (pos < 0) {
3157 return pos;
3158 }
3159 }
3160 pos = rb_str_offset(str, pos);
3161 }
3162 return rb_reg_search(re, str, pos, 0);
3163 }
3164
3165 /*
3166 * call-seq:
3167 * rxp =~ str -> integer or nil
3168 *
3169 * Match---Matches <i>rxp</i> against <i>str</i>.
3170 *
3171 * /at/ =~ "input data" #=> 7
3172 * /ax/ =~ "input data" #=> nil
3173 *
3174 * If <code>=~</code> is used with a regexp literal with named captures,
3175 * captured strings (or nil) is assigned to local variables named by
3176 * the capture names.
3177 *
3178 * /(?<lhs>\w+)\s*=\s*(?<rhs>\w+)/ =~ " x = y "
3179 * p lhs #=> "x"
3180 * p rhs #=> "y"
3181 *
3182 * If it is not matched, nil is assigned for the variables.
3183 *
3184 * /(?<lhs>\w+)\s*=\s*(?<rhs>\w+)/ =~ " x = "
3185 * p lhs #=> nil
3186 * p rhs #=> nil
3187 *
3188 * This assignment is implemented in the Ruby parser.
3189 * The parser detects 'regexp-literal =~ expression' for the assignment.
3190 * The regexp must be a literal without interpolation and placed at left hand side.
3191 *
3192 * The assignment does not occur if the regexp is not a literal.
3193 *
3194 * re = /(?<lhs>\w+)\s*=\s*(?<rhs>\w+)/
3195 * re =~ " x = y "
3196 * p lhs # undefined local variable
3197 * p rhs # undefined local variable
3198 *
3199 * A regexp interpolation, <code>#{}</code>, also disables
3200 * the assignment.
3201 *
3202 * rhs_pat = /(?<rhs>\w+)/
3203 * /(?<lhs>\w+)\s*=\s*#{rhs_pat}/ =~ "x = y"
3204 * p lhs # undefined local variable
3205 *
3206 * The assignment does not occur if the regexp is placed at the right hand side.
3207 *
3208 * " x = y " =~ /(?<lhs>\w+)\s*=\s*(?<rhs>\w+)/
3209 * p lhs, rhs # undefined local variable
3210 *
3211 */
3212
3213 VALUE
rb_reg_match(VALUE re,VALUE str)3214 rb_reg_match(VALUE re, VALUE str)
3215 {
3216 long pos = reg_match_pos(re, &str, 0);
3217 if (pos < 0) return Qnil;
3218 pos = rb_str_sublen(str, pos);
3219 return LONG2FIX(pos);
3220 }
3221
3222 /*
3223 * call-seq:
3224 * rxp === str -> true or false
3225 *
3226 * Case Equality---Used in case statements.
3227 *
3228 * a = "HELLO"
3229 * case a
3230 * when /\A[a-z]*\z/; print "Lower case\n"
3231 * when /\A[A-Z]*\z/; print "Upper case\n"
3232 * else; print "Mixed case\n"
3233 * end
3234 * #=> "Upper case"
3235 *
3236 * Following a regular expression literal with the #=== operator allows you to
3237 * compare against a String.
3238 *
3239 * /^[a-z]*$/ === "HELLO" #=> false
3240 * /^[A-Z]*$/ === "HELLO" #=> true
3241 */
3242
3243 VALUE
rb_reg_eqq(VALUE re,VALUE str)3244 rb_reg_eqq(VALUE re, VALUE str)
3245 {
3246 long start;
3247
3248 str = reg_operand(str, FALSE);
3249 if (NIL_P(str)) {
3250 rb_backref_set(Qnil);
3251 return Qfalse;
3252 }
3253 start = rb_reg_search(re, str, 0, 0);
3254 if (start < 0) {
3255 return Qfalse;
3256 }
3257 return Qtrue;
3258 }
3259
3260
3261 /*
3262 * call-seq:
3263 * ~ rxp -> integer or nil
3264 *
3265 * Match---Matches <i>rxp</i> against the contents of <code>$_</code>.
3266 * Equivalent to <code><i>rxp</i> =~ $_</code>.
3267 *
3268 * $_ = "input data"
3269 * ~ /at/ #=> 7
3270 */
3271
3272 VALUE
rb_reg_match2(VALUE re)3273 rb_reg_match2(VALUE re)
3274 {
3275 long start;
3276 VALUE line = rb_lastline_get();
3277
3278 if (!RB_TYPE_P(line, T_STRING)) {
3279 rb_backref_set(Qnil);
3280 return Qnil;
3281 }
3282
3283 start = rb_reg_search(re, line, 0, 0);
3284 if (start < 0) {
3285 return Qnil;
3286 }
3287 start = rb_str_sublen(line, start);
3288 return LONG2FIX(start);
3289 }
3290
3291
3292 /*
3293 * call-seq:
3294 * rxp.match(str) -> matchdata or nil
3295 * rxp.match(str,pos) -> matchdata or nil
3296 *
3297 * Returns a <code>MatchData</code> object describing the match, or
3298 * <code>nil</code> if there was no match. This is equivalent to retrieving the
3299 * value of the special variable <code>$~</code> following a normal match.
3300 * If the second parameter is present, it specifies the position in the string
3301 * to begin the search.
3302 *
3303 * /(.)(.)(.)/.match("abc")[2] #=> "b"
3304 * /(.)(.)/.match("abc", 1)[2] #=> "c"
3305 *
3306 * If a block is given, invoke the block with MatchData if match succeed, so
3307 * that you can write
3308 *
3309 * /M(.*)/.match("Matz") do |m|
3310 * puts m[0]
3311 * puts m[1]
3312 * end
3313 *
3314 * instead of
3315 *
3316 * if m = /M(.*)/.match("Matz")
3317 * puts m[0]
3318 * puts m[1]
3319 * end
3320 *
3321 * The return value is a value from block execution in this case.
3322 */
3323
3324 static VALUE
rb_reg_match_m(int argc,VALUE * argv,VALUE re)3325 rb_reg_match_m(int argc, VALUE *argv, VALUE re)
3326 {
3327 VALUE result, str, initpos;
3328 long pos;
3329
3330 if (rb_scan_args(argc, argv, "11", &str, &initpos) == 2) {
3331 pos = NUM2LONG(initpos);
3332 }
3333 else {
3334 pos = 0;
3335 }
3336
3337 pos = reg_match_pos(re, &str, pos);
3338 if (pos < 0) {
3339 rb_backref_set(Qnil);
3340 return Qnil;
3341 }
3342 result = rb_backref_get();
3343 rb_match_busy(result);
3344 if (!NIL_P(result) && rb_block_given_p()) {
3345 return rb_yield(result);
3346 }
3347 return result;
3348 }
3349
3350 /*
3351 * call-seq:
3352 * rxp.match?(str) -> true or false
3353 * rxp.match?(str,pos) -> true or false
3354 *
3355 * Returns a <code>true</code> or <code>false</code> indicates whether the
3356 * regexp is matched or not without updating $~ and other related variables.
3357 * If the second parameter is present, it specifies the position in the string
3358 * to begin the search.
3359 *
3360 * /R.../.match?("Ruby") #=> true
3361 * /R.../.match?("Ruby", 1) #=> false
3362 * /P.../.match?("Ruby") #=> false
3363 * $& #=> nil
3364 */
3365
3366 static VALUE
rb_reg_match_m_p(int argc,VALUE * argv,VALUE re)3367 rb_reg_match_m_p(int argc, VALUE *argv, VALUE re)
3368 {
3369 long pos = rb_check_arity(argc, 1, 2) > 1 ? NUM2LONG(argv[1]) : 0;
3370 return rb_reg_match_p(re, argv[0], pos);
3371 }
3372
3373 VALUE
rb_reg_match_p(VALUE re,VALUE str,long pos)3374 rb_reg_match_p(VALUE re, VALUE str, long pos)
3375 {
3376 regex_t *reg;
3377 onig_errmsg_buffer err = "";
3378 OnigPosition result;
3379 const UChar *start, *end;
3380 int tmpreg;
3381
3382 if (NIL_P(str)) return Qfalse;
3383 str = SYMBOL_P(str) ? rb_sym2str(str) : StringValue(str);
3384 if (pos) {
3385 if (pos < 0) {
3386 pos += NUM2LONG(rb_str_length(str));
3387 if (pos < 0) return Qfalse;
3388 }
3389 if (pos > 0) {
3390 long len = 1;
3391 const char *beg = rb_str_subpos(str, pos, &len);
3392 if (!beg) return Qfalse;
3393 pos = beg - RSTRING_PTR(str);
3394 }
3395 }
3396 reg = rb_reg_prepare_re0(re, str, err);
3397 tmpreg = reg != RREGEXP_PTR(re);
3398 if (!tmpreg) RREGEXP(re)->usecnt++;
3399 start = ((UChar*)RSTRING_PTR(str));
3400 end = start + RSTRING_LEN(str);
3401 result = onig_search(reg, start, end, start + pos, end,
3402 NULL, ONIG_OPTION_NONE);
3403 if (!tmpreg) RREGEXP(re)->usecnt--;
3404 if (tmpreg) {
3405 if (RREGEXP(re)->usecnt) {
3406 onig_free(reg);
3407 }
3408 else {
3409 onig_free(RREGEXP_PTR(re));
3410 RREGEXP_PTR(re) = reg;
3411 }
3412 }
3413 if (result < 0) {
3414 if (result == ONIG_MISMATCH) {
3415 return Qfalse;
3416 }
3417 else {
3418 onig_error_code_to_str((UChar*)err, (int)result);
3419 rb_reg_raise(RREGEXP_SRC_PTR(re), RREGEXP_SRC_LEN(re), err, re);
3420 }
3421 }
3422 return Qtrue;
3423 }
3424
3425 /*
3426 * Document-method: compile
3427 *
3428 * Alias for <code>Regexp.new</code>
3429 */
3430
3431 /*
3432 * call-seq:
3433 * Regexp.new(string, [options]) -> regexp
3434 * Regexp.new(regexp) -> regexp
3435 * Regexp.compile(string, [options]) -> regexp
3436 * Regexp.compile(regexp) -> regexp
3437 *
3438 * Constructs a new regular expression from +pattern+, which can be either a
3439 * String or a Regexp (in which case that regexp's options are propagated),
3440 * and new options may not be specified (a change as of Ruby 1.8).
3441 *
3442 * If +options+ is an Integer, it should be one or more of the constants
3443 * Regexp::EXTENDED, Regexp::IGNORECASE, and Regexp::MULTILINE,
3444 * <em>or</em>-ed together. Otherwise, if +options+ is not
3445 * +nil+ or +false+, the regexp will be case insensitive.
3446 *
3447 * r1 = Regexp.new('^a-z+:\\s+\w+') #=> /^a-z+:\s+\w+/
3448 * r2 = Regexp.new('cat', true) #=> /cat/i
3449 * r3 = Regexp.new(r2) #=> /cat/i
3450 * r4 = Regexp.new('dog', Regexp::EXTENDED | Regexp::IGNORECASE) #=> /dog/ix
3451 */
3452
3453 static VALUE
rb_reg_initialize_m(int argc,VALUE * argv,VALUE self)3454 rb_reg_initialize_m(int argc, VALUE *argv, VALUE self)
3455 {
3456 int flags = 0;
3457 VALUE str;
3458 rb_encoding *enc = 0;
3459
3460 rb_check_arity(argc, 1, 3);
3461 if (RB_TYPE_P(argv[0], T_REGEXP)) {
3462 VALUE re = argv[0];
3463
3464 if (argc > 1) {
3465 rb_warn("flags ignored");
3466 }
3467 rb_reg_check(re);
3468 flags = rb_reg_options(re);
3469 str = RREGEXP_SRC(re);
3470 }
3471 else {
3472 if (argc >= 2) {
3473 if (FIXNUM_P(argv[1])) flags = FIX2INT(argv[1]);
3474 else if (RTEST(argv[1])) flags = ONIG_OPTION_IGNORECASE;
3475 }
3476 if (argc == 3 && !NIL_P(argv[2])) {
3477 char *kcode = StringValuePtr(argv[2]);
3478 if (kcode[0] == 'n' || kcode[0] == 'N') {
3479 enc = rb_ascii8bit_encoding();
3480 flags |= ARG_ENCODING_NONE;
3481 }
3482 else {
3483 rb_warn("encoding option is ignored - %s", kcode);
3484 }
3485 }
3486 str = StringValue(argv[0]);
3487 }
3488 if (enc && rb_enc_get(str) != enc)
3489 rb_reg_init_str_enc(self, str, enc, flags);
3490 else
3491 rb_reg_init_str(self, str, flags);
3492 return self;
3493 }
3494
3495 VALUE
rb_reg_quote(VALUE str)3496 rb_reg_quote(VALUE str)
3497 {
3498 rb_encoding *enc = rb_enc_get(str);
3499 char *s, *send, *t;
3500 VALUE tmp;
3501 int c, clen;
3502 int ascii_only = rb_enc_str_asciionly_p(str);
3503
3504 s = RSTRING_PTR(str);
3505 send = s + RSTRING_LEN(str);
3506 while (s < send) {
3507 c = rb_enc_ascget(s, send, &clen, enc);
3508 if (c == -1) {
3509 s += mbclen(s, send, enc);
3510 continue;
3511 }
3512 switch (c) {
3513 case '[': case ']': case '{': case '}':
3514 case '(': case ')': case '|': case '-':
3515 case '*': case '.': case '\\':
3516 case '?': case '+': case '^': case '$':
3517 case ' ': case '#':
3518 case '\t': case '\f': case '\v': case '\n': case '\r':
3519 goto meta_found;
3520 }
3521 s += clen;
3522 }
3523 tmp = rb_str_new3(str);
3524 if (ascii_only) {
3525 rb_enc_associate(tmp, rb_usascii_encoding());
3526 }
3527 return tmp;
3528
3529 meta_found:
3530 tmp = rb_str_new(0, RSTRING_LEN(str)*2);
3531 if (ascii_only) {
3532 rb_enc_associate(tmp, rb_usascii_encoding());
3533 }
3534 else {
3535 rb_enc_copy(tmp, str);
3536 }
3537 t = RSTRING_PTR(tmp);
3538 /* copy upto metacharacter */
3539 memcpy(t, RSTRING_PTR(str), s - RSTRING_PTR(str));
3540 t += s - RSTRING_PTR(str);
3541
3542 while (s < send) {
3543 c = rb_enc_ascget(s, send, &clen, enc);
3544 if (c == -1) {
3545 int n = mbclen(s, send, enc);
3546
3547 while (n--)
3548 *t++ = *s++;
3549 continue;
3550 }
3551 s += clen;
3552 switch (c) {
3553 case '[': case ']': case '{': case '}':
3554 case '(': case ')': case '|': case '-':
3555 case '*': case '.': case '\\':
3556 case '?': case '+': case '^': case '$':
3557 case '#':
3558 t += rb_enc_mbcput('\\', t, enc);
3559 break;
3560 case ' ':
3561 t += rb_enc_mbcput('\\', t, enc);
3562 t += rb_enc_mbcput(' ', t, enc);
3563 continue;
3564 case '\t':
3565 t += rb_enc_mbcput('\\', t, enc);
3566 t += rb_enc_mbcput('t', t, enc);
3567 continue;
3568 case '\n':
3569 t += rb_enc_mbcput('\\', t, enc);
3570 t += rb_enc_mbcput('n', t, enc);
3571 continue;
3572 case '\r':
3573 t += rb_enc_mbcput('\\', t, enc);
3574 t += rb_enc_mbcput('r', t, enc);
3575 continue;
3576 case '\f':
3577 t += rb_enc_mbcput('\\', t, enc);
3578 t += rb_enc_mbcput('f', t, enc);
3579 continue;
3580 case '\v':
3581 t += rb_enc_mbcput('\\', t, enc);
3582 t += rb_enc_mbcput('v', t, enc);
3583 continue;
3584 }
3585 t += rb_enc_mbcput(c, t, enc);
3586 }
3587 rb_str_resize(tmp, t - RSTRING_PTR(tmp));
3588 OBJ_INFECT(tmp, str);
3589 return tmp;
3590 }
3591
3592
3593 /*
3594 * call-seq:
3595 * Regexp.escape(str) -> string
3596 * Regexp.quote(str) -> string
3597 *
3598 * Escapes any characters that would have special meaning in a regular
3599 * expression. Returns a new escaped string, or self if no characters are
3600 * escaped. For any string,
3601 * <code>Regexp.new(Regexp.escape(<i>str</i>))=~<i>str</i></code> will be true.
3602 *
3603 * Regexp.escape('\*?{}.') #=> \\\*\?\{\}\.
3604 *
3605 */
3606
3607 static VALUE
rb_reg_s_quote(VALUE c,VALUE str)3608 rb_reg_s_quote(VALUE c, VALUE str)
3609 {
3610 return rb_reg_quote(reg_operand(str, TRUE));
3611 }
3612
3613 int
rb_reg_options(VALUE re)3614 rb_reg_options(VALUE re)
3615 {
3616 int options;
3617
3618 rb_reg_check(re);
3619 options = RREGEXP_PTR(re)->options & ARG_REG_OPTION_MASK;
3620 if (RBASIC(re)->flags & KCODE_FIXED) options |= ARG_ENCODING_FIXED;
3621 if (RBASIC(re)->flags & REG_ENCODING_NONE) options |= ARG_ENCODING_NONE;
3622 return options;
3623 }
3624
3625 VALUE
rb_check_regexp_type(VALUE re)3626 rb_check_regexp_type(VALUE re)
3627 {
3628 return rb_check_convert_type(re, T_REGEXP, "Regexp", "to_regexp");
3629 }
3630
3631 /*
3632 * call-seq:
3633 * Regexp.try_convert(obj) -> re or nil
3634 *
3635 * Try to convert <i>obj</i> into a Regexp, using to_regexp method.
3636 * Returns converted regexp or nil if <i>obj</i> cannot be converted
3637 * for any reason.
3638 *
3639 * Regexp.try_convert(/re/) #=> /re/
3640 * Regexp.try_convert("re") #=> nil
3641 *
3642 * o = Object.new
3643 * Regexp.try_convert(o) #=> nil
3644 * def o.to_regexp() /foo/ end
3645 * Regexp.try_convert(o) #=> /foo/
3646 *
3647 */
3648 static VALUE
rb_reg_s_try_convert(VALUE dummy,VALUE re)3649 rb_reg_s_try_convert(VALUE dummy, VALUE re)
3650 {
3651 return rb_check_regexp_type(re);
3652 }
3653
3654 static VALUE
rb_reg_s_union(VALUE self,VALUE args0)3655 rb_reg_s_union(VALUE self, VALUE args0)
3656 {
3657 long argc = RARRAY_LEN(args0);
3658
3659 if (argc == 0) {
3660 VALUE args[1];
3661 args[0] = rb_str_new2("(?!)");
3662 return rb_class_new_instance(1, args, rb_cRegexp);
3663 }
3664 else if (argc == 1) {
3665 VALUE arg = rb_ary_entry(args0, 0);
3666 VALUE re = rb_check_regexp_type(arg);
3667 if (!NIL_P(re))
3668 return re;
3669 else {
3670 VALUE quoted;
3671 quoted = rb_reg_s_quote(Qnil, arg);
3672 return rb_reg_new_str(quoted, 0);
3673 }
3674 }
3675 else {
3676 int i;
3677 VALUE source = rb_str_buf_new(0);
3678 rb_encoding *result_enc;
3679
3680 int has_asciionly = 0;
3681 rb_encoding *has_ascii_compat_fixed = 0;
3682 rb_encoding *has_ascii_incompat = 0;
3683
3684 for (i = 0; i < argc; i++) {
3685 volatile VALUE v;
3686 VALUE e = rb_ary_entry(args0, i);
3687
3688 if (0 < i)
3689 rb_str_buf_cat_ascii(source, "|");
3690
3691 v = rb_check_regexp_type(e);
3692 if (!NIL_P(v)) {
3693 rb_encoding *enc = rb_enc_get(v);
3694 if (!rb_enc_asciicompat(enc)) {
3695 if (!has_ascii_incompat)
3696 has_ascii_incompat = enc;
3697 else if (has_ascii_incompat != enc)
3698 rb_raise(rb_eArgError, "incompatible encodings: %s and %s",
3699 rb_enc_name(has_ascii_incompat), rb_enc_name(enc));
3700 }
3701 else if (rb_reg_fixed_encoding_p(v)) {
3702 if (!has_ascii_compat_fixed)
3703 has_ascii_compat_fixed = enc;
3704 else if (has_ascii_compat_fixed != enc)
3705 rb_raise(rb_eArgError, "incompatible encodings: %s and %s",
3706 rb_enc_name(has_ascii_compat_fixed), rb_enc_name(enc));
3707 }
3708 else {
3709 has_asciionly = 1;
3710 }
3711 v = rb_reg_str_with_term(v, -1);
3712 }
3713 else {
3714 rb_encoding *enc;
3715 StringValue(e);
3716 enc = rb_enc_get(e);
3717 if (!rb_enc_asciicompat(enc)) {
3718 if (!has_ascii_incompat)
3719 has_ascii_incompat = enc;
3720 else if (has_ascii_incompat != enc)
3721 rb_raise(rb_eArgError, "incompatible encodings: %s and %s",
3722 rb_enc_name(has_ascii_incompat), rb_enc_name(enc));
3723 }
3724 else if (rb_enc_str_asciionly_p(e)) {
3725 has_asciionly = 1;
3726 }
3727 else {
3728 if (!has_ascii_compat_fixed)
3729 has_ascii_compat_fixed = enc;
3730 else if (has_ascii_compat_fixed != enc)
3731 rb_raise(rb_eArgError, "incompatible encodings: %s and %s",
3732 rb_enc_name(has_ascii_compat_fixed), rb_enc_name(enc));
3733 }
3734 v = rb_reg_s_quote(Qnil, e);
3735 }
3736 if (has_ascii_incompat) {
3737 if (has_asciionly) {
3738 rb_raise(rb_eArgError, "ASCII incompatible encoding: %s",
3739 rb_enc_name(has_ascii_incompat));
3740 }
3741 if (has_ascii_compat_fixed) {
3742 rb_raise(rb_eArgError, "incompatible encodings: %s and %s",
3743 rb_enc_name(has_ascii_incompat), rb_enc_name(has_ascii_compat_fixed));
3744 }
3745 }
3746
3747 if (i == 0) {
3748 rb_enc_copy(source, v);
3749 }
3750 rb_str_append(source, v);
3751 }
3752
3753 if (has_ascii_incompat) {
3754 result_enc = has_ascii_incompat;
3755 }
3756 else if (has_ascii_compat_fixed) {
3757 result_enc = has_ascii_compat_fixed;
3758 }
3759 else {
3760 result_enc = rb_ascii8bit_encoding();
3761 }
3762
3763 rb_enc_associate(source, result_enc);
3764 return rb_class_new_instance(1, &source, rb_cRegexp);
3765 }
3766 }
3767
3768 /*
3769 * call-seq:
3770 * Regexp.union(pat1, pat2, ...) -> new_regexp
3771 * Regexp.union(pats_ary) -> new_regexp
3772 *
3773 * Return a <code>Regexp</code> object that is the union of the given
3774 * <em>pattern</em>s, i.e., will match any of its parts. The <em>pattern</em>s
3775 * can be Regexp objects, in which case their options will be preserved, or
3776 * Strings. If no patterns are given, returns <code>/(?!)/</code>.
3777 * The behavior is unspecified if any given <em>pattern</em> contains capture.
3778 *
3779 * Regexp.union #=> /(?!)/
3780 * Regexp.union("penzance") #=> /penzance/
3781 * Regexp.union("a+b*c") #=> /a\+b\*c/
3782 * Regexp.union("skiing", "sledding") #=> /skiing|sledding/
3783 * Regexp.union(["skiing", "sledding"]) #=> /skiing|sledding/
3784 * Regexp.union(/dogs/, /cats/i) #=> /(?-mix:dogs)|(?i-mx:cats)/
3785 *
3786 * Note: the arguments for ::union will try to be converted into a regular
3787 * expression literal via #to_regexp.
3788 */
3789 static VALUE
rb_reg_s_union_m(VALUE self,VALUE args)3790 rb_reg_s_union_m(VALUE self, VALUE args)
3791 {
3792 VALUE v;
3793 if (RARRAY_LEN(args) == 1 &&
3794 !NIL_P(v = rb_check_array_type(rb_ary_entry(args, 0)))) {
3795 return rb_reg_s_union(self, v);
3796 }
3797 return rb_reg_s_union(self, args);
3798 }
3799
3800 /* :nodoc: */
3801 static VALUE
rb_reg_init_copy(VALUE copy,VALUE re)3802 rb_reg_init_copy(VALUE copy, VALUE re)
3803 {
3804 if (!OBJ_INIT_COPY(copy, re)) return copy;
3805 rb_reg_check(re);
3806 return rb_reg_init_str(copy, RREGEXP_SRC(re), rb_reg_options(re));
3807 }
3808
3809 VALUE
rb_reg_regsub(VALUE str,VALUE src,struct re_registers * regs,VALUE regexp)3810 rb_reg_regsub(VALUE str, VALUE src, struct re_registers *regs, VALUE regexp)
3811 {
3812 VALUE val = 0;
3813 char *p, *s, *e;
3814 int no, clen;
3815 rb_encoding *str_enc = rb_enc_get(str);
3816 rb_encoding *src_enc = rb_enc_get(src);
3817 int acompat = rb_enc_asciicompat(str_enc);
3818 #define ASCGET(s,e,cl) (acompat ? (*(cl)=1,ISASCII((s)[0])?(s)[0]:-1) : rb_enc_ascget((s), (e), (cl), str_enc))
3819
3820 p = s = RSTRING_PTR(str);
3821 e = s + RSTRING_LEN(str);
3822
3823 while (s < e) {
3824 int c = ASCGET(s, e, &clen);
3825 char *ss;
3826
3827 if (c == -1) {
3828 s += mbclen(s, e, str_enc);
3829 continue;
3830 }
3831 ss = s;
3832 s += clen;
3833
3834 if (c != '\\' || s == e) continue;
3835
3836 if (!val) {
3837 val = rb_str_buf_new(ss-p);
3838 }
3839 rb_enc_str_buf_cat(val, p, ss-p, str_enc);
3840
3841 c = ASCGET(s, e, &clen);
3842 if (c == -1) {
3843 s += mbclen(s, e, str_enc);
3844 rb_enc_str_buf_cat(val, ss, s-ss, str_enc);
3845 p = s;
3846 continue;
3847 }
3848 s += clen;
3849
3850 p = s;
3851 switch (c) {
3852 case '1': case '2': case '3': case '4':
3853 case '5': case '6': case '7': case '8': case '9':
3854 if (!NIL_P(regexp) && onig_noname_group_capture_is_active(RREGEXP_PTR(regexp))) {
3855 no = c - '0';
3856 }
3857 else {
3858 continue;
3859 }
3860 break;
3861
3862 case 'k':
3863 if (s < e && ASCGET(s, e, &clen) == '<') {
3864 char *name, *name_end;
3865
3866 name_end = name = s + clen;
3867 while (name_end < e) {
3868 c = ASCGET(name_end, e, &clen);
3869 if (c == '>') break;
3870 name_end += c == -1 ? mbclen(name_end, e, str_enc) : clen;
3871 }
3872 if (name_end < e) {
3873 VALUE n = rb_str_subseq(str, (long)(name - RSTRING_PTR(str)),
3874 (long)(name_end - name));
3875 if ((no = NAME_TO_NUMBER(regs, regexp, n, name, name_end)) < 1) {
3876 name_to_backref_error(n);
3877 }
3878 p = s = name_end + clen;
3879 break;
3880 }
3881 else {
3882 rb_raise(rb_eRuntimeError, "invalid group name reference format");
3883 }
3884 }
3885
3886 rb_enc_str_buf_cat(val, ss, s-ss, str_enc);
3887 continue;
3888
3889 case '0':
3890 case '&':
3891 no = 0;
3892 break;
3893
3894 case '`':
3895 rb_enc_str_buf_cat(val, RSTRING_PTR(src), BEG(0), src_enc);
3896 continue;
3897
3898 case '\'':
3899 rb_enc_str_buf_cat(val, RSTRING_PTR(src)+END(0), RSTRING_LEN(src)-END(0), src_enc);
3900 continue;
3901
3902 case '+':
3903 no = regs->num_regs-1;
3904 while (BEG(no) == -1 && no > 0) no--;
3905 if (no == 0) continue;
3906 break;
3907
3908 case '\\':
3909 rb_enc_str_buf_cat(val, s-clen, clen, str_enc);
3910 continue;
3911
3912 default:
3913 rb_enc_str_buf_cat(val, ss, s-ss, str_enc);
3914 continue;
3915 }
3916
3917 if (no >= 0) {
3918 if (no >= regs->num_regs) continue;
3919 if (BEG(no) == -1) continue;
3920 rb_enc_str_buf_cat(val, RSTRING_PTR(src)+BEG(no), END(no)-BEG(no), src_enc);
3921 }
3922 }
3923
3924 if (!val) return str;
3925 if (p < e) {
3926 rb_enc_str_buf_cat(val, p, e-p, str_enc);
3927 }
3928
3929 return val;
3930 }
3931
3932 static VALUE
kcode_getter(void)3933 kcode_getter(void)
3934 {
3935 rb_warn("variable $KCODE is no longer effective");
3936 return Qnil;
3937 }
3938
3939 static void
kcode_setter(VALUE val,ID id)3940 kcode_setter(VALUE val, ID id)
3941 {
3942 rb_warn("variable $KCODE is no longer effective; ignored");
3943 }
3944
3945 static VALUE
ignorecase_getter(void)3946 ignorecase_getter(void)
3947 {
3948 rb_warn("variable $= is no longer effective");
3949 return Qfalse;
3950 }
3951
3952 static void
ignorecase_setter(VALUE val,ID id)3953 ignorecase_setter(VALUE val, ID id)
3954 {
3955 rb_warn("variable $= is no longer effective; ignored");
3956 }
3957
3958 static VALUE
match_getter(void)3959 match_getter(void)
3960 {
3961 VALUE match = rb_backref_get();
3962
3963 if (NIL_P(match)) return Qnil;
3964 rb_match_busy(match);
3965 return match;
3966 }
3967
3968 static void
match_setter(VALUE val)3969 match_setter(VALUE val)
3970 {
3971 if (!NIL_P(val)) {
3972 Check_Type(val, T_MATCH);
3973 }
3974 rb_backref_set(val);
3975 }
3976
3977 /*
3978 * call-seq:
3979 * Regexp.last_match -> matchdata
3980 * Regexp.last_match(n) -> str
3981 *
3982 * The first form returns the MatchData object generated by the
3983 * last successful pattern match. Equivalent to reading the special global
3984 * variable <code>$~</code> (see Special global variables in Regexp for
3985 * details).
3986 *
3987 * The second form returns the <i>n</i>th field in this MatchData object.
3988 * _n_ can be a string or symbol to reference a named capture.
3989 *
3990 * Note that the last_match is local to the thread and method scope of the
3991 * method that did the pattern match.
3992 *
3993 * /c(.)t/ =~ 'cat' #=> 0
3994 * Regexp.last_match #=> #<MatchData "cat" 1:"a">
3995 * Regexp.last_match(0) #=> "cat"
3996 * Regexp.last_match(1) #=> "a"
3997 * Regexp.last_match(2) #=> nil
3998 *
3999 * /(?<lhs>\w+)\s*=\s*(?<rhs>\w+)/ =~ "var = val"
4000 * Regexp.last_match #=> #<MatchData "var = val" lhs:"var" rhs:"val">
4001 * Regexp.last_match(:lhs) #=> "var"
4002 * Regexp.last_match(:rhs) #=> "val"
4003 */
4004
4005 static VALUE
rb_reg_s_last_match(int argc,VALUE * argv)4006 rb_reg_s_last_match(int argc, VALUE *argv)
4007 {
4008 if (rb_check_arity(argc, 0, 1) == 1) {
4009 VALUE match = rb_backref_get();
4010 int n;
4011 if (NIL_P(match)) return Qnil;
4012 n = match_backref_number(match, argv[0]);
4013 return rb_reg_nth_match(n, match);
4014 }
4015 return match_getter();
4016 }
4017
4018 static void
re_warn(const char * s)4019 re_warn(const char *s)
4020 {
4021 rb_warn("%s", s);
4022 }
4023
4024 /*
4025 * Document-class: RegexpError
4026 *
4027 * Raised when given an invalid regexp expression.
4028 *
4029 * Regexp.new("?")
4030 *
4031 * <em>raises the exception:</em>
4032 *
4033 * RegexpError: target of repeat operator is not specified: /?/
4034 */
4035
4036 /*
4037 * Document-class: Regexp
4038 *
4039 * A <code>Regexp</code> holds a regular expression, used to match a pattern
4040 * against strings. Regexps are created using the <code>/.../</code> and
4041 * <code>%r{...}</code> literals, and by the <code>Regexp::new</code>
4042 * constructor.
4043 *
4044 * :include: doc/regexp.rdoc
4045 */
4046
4047 void
Init_Regexp(void)4048 Init_Regexp(void)
4049 {
4050 rb_eRegexpError = rb_define_class("RegexpError", rb_eStandardError);
4051
4052 onigenc_set_default_encoding(ONIG_ENCODING_ASCII);
4053 onig_set_warn_func(re_warn);
4054 onig_set_verb_warn_func(re_warn);
4055
4056 rb_define_virtual_variable("$~", match_getter, match_setter);
4057 rb_define_virtual_variable("$&", last_match_getter, 0);
4058 rb_define_virtual_variable("$`", prematch_getter, 0);
4059 rb_define_virtual_variable("$'", postmatch_getter, 0);
4060 rb_define_virtual_variable("$+", last_paren_match_getter, 0);
4061
4062 rb_define_virtual_variable("$=", ignorecase_getter, ignorecase_setter);
4063 rb_define_virtual_variable("$KCODE", kcode_getter, kcode_setter);
4064 rb_define_virtual_variable("$-K", kcode_getter, kcode_setter);
4065
4066 rb_cRegexp = rb_define_class("Regexp", rb_cObject);
4067 rb_define_alloc_func(rb_cRegexp, rb_reg_s_alloc);
4068 rb_define_singleton_method(rb_cRegexp, "compile", rb_class_new_instance, -1);
4069 rb_define_singleton_method(rb_cRegexp, "quote", rb_reg_s_quote, 1);
4070 rb_define_singleton_method(rb_cRegexp, "escape", rb_reg_s_quote, 1);
4071 rb_define_singleton_method(rb_cRegexp, "union", rb_reg_s_union_m, -2);
4072 rb_define_singleton_method(rb_cRegexp, "last_match", rb_reg_s_last_match, -1);
4073 rb_define_singleton_method(rb_cRegexp, "try_convert", rb_reg_s_try_convert, 1);
4074
4075 rb_define_method(rb_cRegexp, "initialize", rb_reg_initialize_m, -1);
4076 rb_define_method(rb_cRegexp, "initialize_copy", rb_reg_init_copy, 1);
4077 rb_define_method(rb_cRegexp, "hash", rb_reg_hash, 0);
4078 rb_define_method(rb_cRegexp, "eql?", rb_reg_equal, 1);
4079 rb_define_method(rb_cRegexp, "==", rb_reg_equal, 1);
4080 rb_define_method(rb_cRegexp, "=~", rb_reg_match, 1);
4081 rb_define_method(rb_cRegexp, "===", rb_reg_eqq, 1);
4082 rb_define_method(rb_cRegexp, "~", rb_reg_match2, 0);
4083 rb_define_method(rb_cRegexp, "match", rb_reg_match_m, -1);
4084 rb_define_method(rb_cRegexp, "match?", rb_reg_match_m_p, -1);
4085 rb_define_method(rb_cRegexp, "to_s", rb_reg_to_s, 0);
4086 rb_define_method(rb_cRegexp, "inspect", rb_reg_inspect, 0);
4087 rb_define_method(rb_cRegexp, "source", rb_reg_source, 0);
4088 rb_define_method(rb_cRegexp, "casefold?", rb_reg_casefold_p, 0);
4089 rb_define_method(rb_cRegexp, "options", rb_reg_options_m, 0);
4090 rb_define_method(rb_cRegexp, "encoding", rb_obj_encoding, 0); /* in encoding.c */
4091 rb_define_method(rb_cRegexp, "fixed_encoding?", rb_reg_fixed_encoding_p, 0);
4092 rb_define_method(rb_cRegexp, "names", rb_reg_names, 0);
4093 rb_define_method(rb_cRegexp, "named_captures", rb_reg_named_captures, 0);
4094
4095 /* see Regexp.options and Regexp.new */
4096 rb_define_const(rb_cRegexp, "IGNORECASE", INT2FIX(ONIG_OPTION_IGNORECASE));
4097 /* see Regexp.options and Regexp.new */
4098 rb_define_const(rb_cRegexp, "EXTENDED", INT2FIX(ONIG_OPTION_EXTEND));
4099 /* see Regexp.options and Regexp.new */
4100 rb_define_const(rb_cRegexp, "MULTILINE", INT2FIX(ONIG_OPTION_MULTILINE));
4101 /* see Regexp.options and Regexp.new */
4102 rb_define_const(rb_cRegexp, "FIXEDENCODING", INT2FIX(ARG_ENCODING_FIXED));
4103 /* see Regexp.options and Regexp.new */
4104 rb_define_const(rb_cRegexp, "NOENCODING", INT2FIX(ARG_ENCODING_NONE));
4105
4106 rb_global_variable(®_cache);
4107
4108 rb_cMatch = rb_define_class("MatchData", rb_cObject);
4109 rb_define_alloc_func(rb_cMatch, match_alloc);
4110 rb_undef_method(CLASS_OF(rb_cMatch), "new");
4111
4112 rb_define_method(rb_cMatch, "initialize_copy", match_init_copy, 1);
4113 rb_define_method(rb_cMatch, "regexp", match_regexp, 0);
4114 rb_define_method(rb_cMatch, "names", match_names, 0);
4115 rb_define_method(rb_cMatch, "size", match_size, 0);
4116 rb_define_method(rb_cMatch, "length", match_size, 0);
4117 rb_define_method(rb_cMatch, "offset", match_offset, 1);
4118 rb_define_method(rb_cMatch, "begin", match_begin, 1);
4119 rb_define_method(rb_cMatch, "end", match_end, 1);
4120 rb_define_method(rb_cMatch, "to_a", match_to_a, 0);
4121 rb_define_method(rb_cMatch, "[]", match_aref, -1);
4122 rb_define_method(rb_cMatch, "captures", match_captures, 0);
4123 rb_define_method(rb_cMatch, "named_captures", match_named_captures, 0);
4124 rb_define_method(rb_cMatch, "values_at", match_values_at, -1);
4125 rb_define_method(rb_cMatch, "pre_match", rb_reg_match_pre, 0);
4126 rb_define_method(rb_cMatch, "post_match", rb_reg_match_post, 0);
4127 rb_define_method(rb_cMatch, "to_s", match_to_s, 0);
4128 rb_define_method(rb_cMatch, "inspect", match_inspect, 0);
4129 rb_define_method(rb_cMatch, "string", match_string, 0);
4130 rb_define_method(rb_cMatch, "hash", match_hash, 0);
4131 rb_define_method(rb_cMatch, "eql?", match_equal, 1);
4132 rb_define_method(rb_cMatch, "==", match_equal, 1);
4133 }
4134