Ruby  3.4.0dev (2024-11-22 revision 37a72b0150ec36b4ea27175039afc28c62207b0c)
re.c (37a72b0150ec36b4ea27175039afc28c62207b0c)
1 /**********************************************************************
2 
3  re.c -
4 
5  $Author$
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/internal/config.h"
13 
14 #include <ctype.h>
15 
16 #include "encindex.h"
17 #include "hrtime.h"
18 #include "internal.h"
19 #include "internal/encoding.h"
20 #include "internal/hash.h"
21 #include "internal/imemo.h"
22 #include "internal/re.h"
23 #include "internal/string.h"
24 #include "internal/object.h"
25 #include "internal/ractor.h"
26 #include "internal/variable.h"
27 #include "regint.h"
28 #include "ruby/encoding.h"
29 #include "ruby/re.h"
30 #include "ruby/util.h"
31 
32 VALUE rb_eRegexpError, rb_eRegexpTimeoutError;
33 
34 typedef char onig_errmsg_buffer[ONIG_MAX_ERROR_MESSAGE_LEN];
35 #define errcpy(err, msg) strlcpy((err), (msg), ONIG_MAX_ERROR_MESSAGE_LEN)
36 
37 #define BEG(no) (regs->beg[(no)])
38 #define END(no) (regs->end[(no)])
39 
40 #if 'a' == 97 /* it's ascii */
41 static const char casetable[] = {
42  '\000', '\001', '\002', '\003', '\004', '\005', '\006', '\007',
43  '\010', '\011', '\012', '\013', '\014', '\015', '\016', '\017',
44  '\020', '\021', '\022', '\023', '\024', '\025', '\026', '\027',
45  '\030', '\031', '\032', '\033', '\034', '\035', '\036', '\037',
46  /* ' ' '!' '"' '#' '$' '%' '&' ''' */
47  '\040', '\041', '\042', '\043', '\044', '\045', '\046', '\047',
48  /* '(' ')' '*' '+' ',' '-' '.' '/' */
49  '\050', '\051', '\052', '\053', '\054', '\055', '\056', '\057',
50  /* '0' '1' '2' '3' '4' '5' '6' '7' */
51  '\060', '\061', '\062', '\063', '\064', '\065', '\066', '\067',
52  /* '8' '9' ':' ';' '<' '=' '>' '?' */
53  '\070', '\071', '\072', '\073', '\074', '\075', '\076', '\077',
54  /* '@' 'A' 'B' 'C' 'D' 'E' 'F' 'G' */
55  '\100', '\141', '\142', '\143', '\144', '\145', '\146', '\147',
56  /* 'H' 'I' 'J' 'K' 'L' 'M' 'N' 'O' */
57  '\150', '\151', '\152', '\153', '\154', '\155', '\156', '\157',
58  /* 'P' 'Q' 'R' 'S' 'T' 'U' 'V' 'W' */
59  '\160', '\161', '\162', '\163', '\164', '\165', '\166', '\167',
60  /* 'X' 'Y' 'Z' '[' '\' ']' '^' '_' */
61  '\170', '\171', '\172', '\133', '\134', '\135', '\136', '\137',
62  /* '`' 'a' 'b' 'c' 'd' 'e' 'f' 'g' */
63  '\140', '\141', '\142', '\143', '\144', '\145', '\146', '\147',
64  /* 'h' 'i' 'j' 'k' 'l' 'm' 'n' 'o' */
65  '\150', '\151', '\152', '\153', '\154', '\155', '\156', '\157',
66  /* 'p' 'q' 'r' 's' 't' 'u' 'v' 'w' */
67  '\160', '\161', '\162', '\163', '\164', '\165', '\166', '\167',
68  /* 'x' 'y' 'z' '{' '|' '}' '~' */
69  '\170', '\171', '\172', '\173', '\174', '\175', '\176', '\177',
70  '\200', '\201', '\202', '\203', '\204', '\205', '\206', '\207',
71  '\210', '\211', '\212', '\213', '\214', '\215', '\216', '\217',
72  '\220', '\221', '\222', '\223', '\224', '\225', '\226', '\227',
73  '\230', '\231', '\232', '\233', '\234', '\235', '\236', '\237',
74  '\240', '\241', '\242', '\243', '\244', '\245', '\246', '\247',
75  '\250', '\251', '\252', '\253', '\254', '\255', '\256', '\257',
76  '\260', '\261', '\262', '\263', '\264', '\265', '\266', '\267',
77  '\270', '\271', '\272', '\273', '\274', '\275', '\276', '\277',
78  '\300', '\301', '\302', '\303', '\304', '\305', '\306', '\307',
79  '\310', '\311', '\312', '\313', '\314', '\315', '\316', '\317',
80  '\320', '\321', '\322', '\323', '\324', '\325', '\326', '\327',
81  '\330', '\331', '\332', '\333', '\334', '\335', '\336', '\337',
82  '\340', '\341', '\342', '\343', '\344', '\345', '\346', '\347',
83  '\350', '\351', '\352', '\353', '\354', '\355', '\356', '\357',
84  '\360', '\361', '\362', '\363', '\364', '\365', '\366', '\367',
85  '\370', '\371', '\372', '\373', '\374', '\375', '\376', '\377',
86 };
87 #else
88 # error >>> "You lose. You will need a translation table for your character set." <<<
89 #endif
90 
91 // The process-global timeout for regexp matching
92 rb_hrtime_t rb_reg_match_time_limit = 0;
93 
94 int
95 rb_memcicmp(const void *x, const void *y, long len)
96 {
97  const unsigned char *p1 = x, *p2 = y;
98  int tmp;
99 
100  while (len--) {
101  if ((tmp = casetable[(unsigned)*p1++] - casetable[(unsigned)*p2++]))
102  return tmp;
103  }
104  return 0;
105 }
106 
107 #ifdef HAVE_MEMMEM
108 static inline long
109 rb_memsearch_ss(const unsigned char *xs, long m, const unsigned char *ys, long n)
110 {
111  const unsigned char *y;
112 
113  if ((y = memmem(ys, n, xs, m)) != NULL)
114  return y - ys;
115  else
116  return -1;
117 }
118 #else
119 static inline long
120 rb_memsearch_ss(const unsigned char *xs, long m, const unsigned char *ys, long n)
121 {
122  const unsigned char *x = xs, *xe = xs + m;
123  const unsigned char *y = ys, *ye = ys + n;
124 #define VALUE_MAX ((VALUE)~(VALUE)0)
125  VALUE hx, hy, mask = VALUE_MAX >> ((SIZEOF_VALUE - m) * CHAR_BIT);
126 
127  if (m > SIZEOF_VALUE)
128  rb_bug("!!too long pattern string!!");
129 
130  if (!(y = memchr(y, *x, n - m + 1)))
131  return -1;
132 
133  /* Prepare hash value */
134  for (hx = *x++, hy = *y++; x < xe; ++x, ++y) {
135  hx <<= CHAR_BIT;
136  hy <<= CHAR_BIT;
137  hx |= *x;
138  hy |= *y;
139  }
140  /* Searching */
141  while (hx != hy) {
142  if (y == ye)
143  return -1;
144  hy <<= CHAR_BIT;
145  hy |= *y;
146  hy &= mask;
147  y++;
148  }
149  return y - ys - m;
150 }
151 #endif
152 
153 static inline long
154 rb_memsearch_qs(const unsigned char *xs, long m, const unsigned char *ys, long n)
155 {
156  const unsigned char *x = xs, *xe = xs + m;
157  const unsigned char *y = ys;
158  VALUE i, qstable[256];
159 
160  /* Preprocessing */
161  for (i = 0; i < 256; ++i)
162  qstable[i] = m + 1;
163  for (; x < xe; ++x)
164  qstable[*x] = xe - x;
165  /* Searching */
166  for (; y + m <= ys + n; y += *(qstable + y[m])) {
167  if (*xs == *y && memcmp(xs, y, m) == 0)
168  return y - ys;
169  }
170  return -1;
171 }
172 
173 static inline unsigned int
174 rb_memsearch_qs_utf8_hash(const unsigned char *x)
175 {
176  register const unsigned int mix = 8353;
177  register unsigned int h = *x;
178  if (h < 0xC0) {
179  return h + 256;
180  }
181  else if (h < 0xE0) {
182  h *= mix;
183  h += x[1];
184  }
185  else if (h < 0xF0) {
186  h *= mix;
187  h += x[1];
188  h *= mix;
189  h += x[2];
190  }
191  else if (h < 0xF5) {
192  h *= mix;
193  h += x[1];
194  h *= mix;
195  h += x[2];
196  h *= mix;
197  h += x[3];
198  }
199  else {
200  return h + 256;
201  }
202  return (unsigned char)h;
203 }
204 
205 static inline long
206 rb_memsearch_qs_utf8(const unsigned char *xs, long m, const unsigned char *ys, long n)
207 {
208  const unsigned char *x = xs, *xe = xs + m;
209  const unsigned char *y = ys;
210  VALUE i, qstable[512];
211 
212  /* Preprocessing */
213  for (i = 0; i < 512; ++i) {
214  qstable[i] = m + 1;
215  }
216  for (; x < xe; ++x) {
217  qstable[rb_memsearch_qs_utf8_hash(x)] = xe - x;
218  }
219  /* Searching */
220  for (; y + m <= ys + n; y += qstable[rb_memsearch_qs_utf8_hash(y+m)]) {
221  if (*xs == *y && memcmp(xs, y, m) == 0)
222  return y - ys;
223  }
224  return -1;
225 }
226 
227 static inline long
228 rb_memsearch_with_char_size(const unsigned char *xs, long m, const unsigned char *ys, long n, int char_size)
229 {
230  const unsigned char *x = xs, x0 = *xs, *y = ys;
231 
232  for (n -= m; n >= 0; n -= char_size, y += char_size) {
233  if (x0 == *y && memcmp(x+1, y+1, m-1) == 0)
234  return y - ys;
235  }
236  return -1;
237 }
238 
239 static inline long
240 rb_memsearch_wchar(const unsigned char *xs, long m, const unsigned char *ys, long n)
241 {
242  return rb_memsearch_with_char_size(xs, m, ys, n, 2);
243 }
244 
245 static inline long
246 rb_memsearch_qchar(const unsigned char *xs, long m, const unsigned char *ys, long n)
247 {
248  return rb_memsearch_with_char_size(xs, m, ys, n, 4);
249 }
250 
251 long
252 rb_memsearch(const void *x0, long m, const void *y0, long n, rb_encoding *enc)
253 {
254  const unsigned char *x = x0, *y = y0;
255 
256  if (m > n) return -1;
257  else if (m == n) {
258  return memcmp(x0, y0, m) == 0 ? 0 : -1;
259  }
260  else if (m < 1) {
261  return 0;
262  }
263  else if (m == 1) {
264  const unsigned char *ys = memchr(y, *x, n);
265 
266  if (ys)
267  return ys - y;
268  else
269  return -1;
270  }
271  else if (LIKELY(rb_enc_mbminlen(enc) == 1)) {
272  if (m <= SIZEOF_VALUE) {
273  return rb_memsearch_ss(x0, m, y0, n);
274  }
275  else if (enc == rb_utf8_encoding()){
276  return rb_memsearch_qs_utf8(x0, m, y0, n);
277  }
278  }
279  else if (LIKELY(rb_enc_mbminlen(enc) == 2)) {
280  return rb_memsearch_wchar(x0, m, y0, n);
281  }
282  else if (LIKELY(rb_enc_mbminlen(enc) == 4)) {
283  return rb_memsearch_qchar(x0, m, y0, n);
284  }
285  return rb_memsearch_qs(x0, m, y0, n);
286 }
287 
288 #define REG_ENCODING_NONE FL_USER6
289 
290 #define KCODE_FIXED FL_USER4
291 
292 #define ARG_REG_OPTION_MASK \
293  (ONIG_OPTION_IGNORECASE|ONIG_OPTION_MULTILINE|ONIG_OPTION_EXTEND)
294 #define ARG_ENCODING_FIXED 16
295 #define ARG_ENCODING_NONE 32
296 
297 static int
298 char_to_option(int c)
299 {
300  int val;
301 
302  switch (c) {
303  case 'i':
304  val = ONIG_OPTION_IGNORECASE;
305  break;
306  case 'x':
307  val = ONIG_OPTION_EXTEND;
308  break;
309  case 'm':
310  val = ONIG_OPTION_MULTILINE;
311  break;
312  default:
313  val = 0;
314  break;
315  }
316  return val;
317 }
318 
319 enum { OPTBUF_SIZE = 4 };
320 
321 static char *
322 option_to_str(char str[OPTBUF_SIZE], int options)
323 {
324  char *p = str;
325  if (options & ONIG_OPTION_MULTILINE) *p++ = 'm';
326  if (options & ONIG_OPTION_IGNORECASE) *p++ = 'i';
327  if (options & ONIG_OPTION_EXTEND) *p++ = 'x';
328  *p = 0;
329  return str;
330 }
331 
332 extern int
333 rb_char_to_option_kcode(int c, int *option, int *kcode)
334 {
335  *option = 0;
336 
337  switch (c) {
338  case 'n':
339  *kcode = rb_ascii8bit_encindex();
340  return (*option = ARG_ENCODING_NONE);
341  case 'e':
342  *kcode = ENCINDEX_EUC_JP;
343  break;
344  case 's':
345  *kcode = ENCINDEX_Windows_31J;
346  break;
347  case 'u':
348  *kcode = rb_utf8_encindex();
349  break;
350  default:
351  *kcode = -1;
352  return (*option = char_to_option(c));
353  }
354  *option = ARG_ENCODING_FIXED;
355  return 1;
356 }
357 
358 static void
359 rb_reg_check(VALUE re)
360 {
361  if (!RREGEXP_PTR(re) || !RREGEXP_SRC(re) || !RREGEXP_SRC_PTR(re)) {
362  rb_raise(rb_eTypeError, "uninitialized Regexp");
363  }
364 }
365 
366 static void
367 rb_reg_expr_str(VALUE str, const char *s, long len,
368  rb_encoding *enc, rb_encoding *resenc, int term)
369 {
370  const char *p, *pend;
371  int cr = ENC_CODERANGE_UNKNOWN;
372  int need_escape = 0;
373  int c, clen;
374 
375  p = s; pend = p + len;
376  rb_str_coderange_scan_restartable(p, pend, enc, &cr);
377  if (rb_enc_asciicompat(enc) && ENC_CODERANGE_CLEAN_P(cr)) {
378  while (p < pend) {
379  c = rb_enc_ascget(p, pend, &clen, enc);
380  if (c == -1) {
381  if (enc == resenc) {
382  p += mbclen(p, pend, enc);
383  }
384  else {
385  need_escape = 1;
386  break;
387  }
388  }
389  else if (c != term && rb_enc_isprint(c, enc)) {
390  p += clen;
391  }
392  else {
393  need_escape = 1;
394  break;
395  }
396  }
397  }
398  else {
399  need_escape = 1;
400  }
401 
402  if (!need_escape) {
403  rb_str_buf_cat(str, s, len);
404  }
405  else {
406  int unicode_p = rb_enc_unicode_p(enc);
407  p = s;
408  while (p<pend) {
409  c = rb_enc_ascget(p, pend, &clen, enc);
410  if (c == '\\' && p+clen < pend) {
411  int n = clen + mbclen(p+clen, pend, enc);
412  rb_str_buf_cat(str, p, n);
413  p += n;
414  continue;
415  }
416  else if (c == -1) {
417  clen = rb_enc_precise_mbclen(p, pend, enc);
418  if (!MBCLEN_CHARFOUND_P(clen)) {
419  c = (unsigned char)*p;
420  clen = 1;
421  goto hex;
422  }
423  if (resenc) {
424  unsigned int c = rb_enc_mbc_to_codepoint(p, pend, enc);
425  rb_str_buf_cat_escaped_char(str, c, unicode_p);
426  }
427  else {
428  clen = MBCLEN_CHARFOUND_LEN(clen);
429  rb_str_buf_cat(str, p, clen);
430  }
431  }
432  else if (c == term) {
433  char c = '\\';
434  rb_str_buf_cat(str, &c, 1);
435  rb_str_buf_cat(str, p, clen);
436  }
437  else if (rb_enc_isprint(c, enc)) {
438  rb_str_buf_cat(str, p, clen);
439  }
440  else if (!rb_enc_isspace(c, enc)) {
441  char b[8];
442 
443  hex:
444  snprintf(b, sizeof(b), "\\x%02X", c);
445  rb_str_buf_cat(str, b, 4);
446  }
447  else {
448  rb_str_buf_cat(str, p, clen);
449  }
450  p += clen;
451  }
452  }
453 }
454 
455 static VALUE
456 rb_reg_desc(VALUE re)
457 {
458  rb_encoding *enc = rb_enc_get(re);
459  VALUE str = rb_str_buf_new2("/");
461  if (resenc == NULL) resenc = rb_default_external_encoding();
462 
463  if (re && rb_enc_asciicompat(enc)) {
464  rb_enc_copy(str, re);
465  }
466  else {
468  }
469 
470  VALUE src_str = RREGEXP_SRC(re);
471  rb_reg_expr_str(str, RSTRING_PTR(src_str), RSTRING_LEN(src_str), enc, resenc, '/');
472  RB_GC_GUARD(src_str);
473 
474  rb_str_buf_cat2(str, "/");
475  if (re) {
476  char opts[OPTBUF_SIZE];
477  rb_reg_check(re);
478  if (*option_to_str(opts, RREGEXP_PTR(re)->options))
479  rb_str_buf_cat2(str, opts);
480  if (RBASIC(re)->flags & REG_ENCODING_NONE)
481  rb_str_buf_cat2(str, "n");
482  }
483  return str;
484 }
485 
486 
487 /*
488  * call-seq:
489  * source -> string
490  *
491  * Returns the original string of +self+:
492  *
493  * /ab+c/ix.source # => "ab+c"
494  *
495  * Regexp escape sequences are retained:
496  *
497  * /\x20\+/.source # => "\\x20\\+"
498  *
499  * Lexer escape characters are not retained:
500  *
501  * /\//.source # => "/"
502  *
503  */
504 
505 static VALUE
506 rb_reg_source(VALUE re)
507 {
508  VALUE str;
509 
510  rb_reg_check(re);
511  str = rb_str_dup(RREGEXP_SRC(re));
512  return str;
513 }
514 
515 /*
516  * call-seq:
517  * inspect -> string
518  *
519  * Returns a nicely-formatted string representation of +self+:
520  *
521  * /ab+c/ix.inspect # => "/ab+c/ix"
522  *
523  * Related: Regexp#to_s.
524  */
525 
526 static VALUE
527 rb_reg_inspect(VALUE re)
528 {
529  if (!RREGEXP_PTR(re) || !RREGEXP_SRC(re) || !RREGEXP_SRC_PTR(re)) {
530  return rb_any_to_s(re);
531  }
532  return rb_reg_desc(re);
533 }
534 
535 static VALUE rb_reg_str_with_term(VALUE re, int term);
536 
537 /*
538  * call-seq:
539  * to_s -> string
540  *
541  * Returns a string showing the options and string of +self+:
542  *
543  * r0 = /ab+c/ix
544  * s0 = r0.to_s # => "(?ix-m:ab+c)"
545  *
546  * The returned string may be used as an argument to Regexp.new,
547  * or as interpolated text for a
548  * {Regexp interpolation}[rdoc-ref:Regexp@Interpolation+Mode]:
549  *
550  * r1 = Regexp.new(s0) # => /(?ix-m:ab+c)/
551  * r2 = /#{s0}/ # => /(?ix-m:ab+c)/
552  *
553  * Note that +r1+ and +r2+ are not equal to +r0+
554  * because their original strings are different:
555  *
556  * r0 == r1 # => false
557  * r0.source # => "ab+c"
558  * r1.source # => "(?ix-m:ab+c)"
559  *
560  * Related: Regexp#inspect.
561  *
562  */
563 
564 static VALUE
565 rb_reg_to_s(VALUE re)
566 {
567  return rb_reg_str_with_term(re, '/');
568 }
569 
570 static VALUE
571 rb_reg_str_with_term(VALUE re, int term)
572 {
573  int options, opt;
574  const int embeddable = ONIG_OPTION_MULTILINE|ONIG_OPTION_IGNORECASE|ONIG_OPTION_EXTEND;
575  VALUE str = rb_str_buf_new2("(?");
576  char optbuf[OPTBUF_SIZE + 1]; /* for '-' */
577  rb_encoding *enc = rb_enc_get(re);
578 
579  rb_reg_check(re);
580 
581  rb_enc_copy(str, re);
582  options = RREGEXP_PTR(re)->options;
583  VALUE src_str = RREGEXP_SRC(re);
584  const UChar *ptr = (UChar *)RSTRING_PTR(src_str);
585  long len = RSTRING_LEN(src_str);
586  again:
587  if (len >= 4 && ptr[0] == '(' && ptr[1] == '?') {
588  int err = 1;
589  ptr += 2;
590  if ((len -= 2) > 0) {
591  do {
592  opt = char_to_option((int )*ptr);
593  if (opt != 0) {
594  options |= opt;
595  }
596  else {
597  break;
598  }
599  ++ptr;
600  } while (--len > 0);
601  }
602  if (len > 1 && *ptr == '-') {
603  ++ptr;
604  --len;
605  do {
606  opt = char_to_option((int )*ptr);
607  if (opt != 0) {
608  options &= ~opt;
609  }
610  else {
611  break;
612  }
613  ++ptr;
614  } while (--len > 0);
615  }
616  if (*ptr == ')') {
617  --len;
618  ++ptr;
619  goto again;
620  }
621  if (*ptr == ':' && ptr[len-1] == ')') {
622  Regexp *rp;
623  VALUE verbose = ruby_verbose;
625 
626  ++ptr;
627  len -= 2;
628  err = onig_new(&rp, ptr, ptr + len, options,
629  enc, OnigDefaultSyntax, NULL);
630  onig_free(rp);
631  ruby_verbose = verbose;
632  }
633  if (err) {
634  options = RREGEXP_PTR(re)->options;
635  ptr = (UChar*)RREGEXP_SRC_PTR(re);
636  len = RREGEXP_SRC_LEN(re);
637  }
638  }
639 
640  if (*option_to_str(optbuf, options)) rb_str_buf_cat2(str, optbuf);
641 
642  if ((options & embeddable) != embeddable) {
643  optbuf[0] = '-';
644  option_to_str(optbuf + 1, ~options);
645  rb_str_buf_cat2(str, optbuf);
646  }
647 
648  rb_str_buf_cat2(str, ":");
649  if (rb_enc_asciicompat(enc)) {
650  rb_reg_expr_str(str, (char*)ptr, len, enc, NULL, term);
651  rb_str_buf_cat2(str, ")");
652  }
653  else {
654  const char *s, *e;
655  char *paren;
656  ptrdiff_t n;
657  rb_str_buf_cat2(str, ")");
659  str = rb_str_encode(str, rb_enc_from_encoding(enc), 0, Qnil);
660 
661  /* backup encoded ")" to paren */
662  s = RSTRING_PTR(str);
663  e = RSTRING_END(str);
664  s = rb_enc_left_char_head(s, e-1, e, enc);
665  n = e - s;
666  paren = ALLOCA_N(char, n);
667  memcpy(paren, s, n);
668  rb_str_resize(str, RSTRING_LEN(str) - n);
669 
670  rb_reg_expr_str(str, (char*)ptr, len, enc, NULL, term);
671  rb_str_buf_cat(str, paren, n);
672  }
673  rb_enc_copy(str, re);
674 
675  RB_GC_GUARD(src_str);
676 
677  return str;
678 }
679 
680 NORETURN(static void rb_reg_raise(const char *err, VALUE re));
681 
682 static void
683 rb_reg_raise(const char *err, VALUE re)
684 {
685  VALUE desc = rb_reg_desc(re);
686 
687  rb_raise(rb_eRegexpError, "%s: %"PRIsVALUE, err, desc);
688 }
689 
690 static VALUE
691 rb_enc_reg_error_desc(const char *s, long len, rb_encoding *enc, int options, const char *err)
692 {
693  char opts[OPTBUF_SIZE + 1]; /* for '/' */
694  VALUE desc = rb_str_buf_new2(err);
696  if (resenc == NULL) resenc = rb_default_external_encoding();
697 
698  rb_enc_associate(desc, enc);
699  rb_str_buf_cat2(desc, ": /");
700  rb_reg_expr_str(desc, s, len, enc, resenc, '/');
701  opts[0] = '/';
702  option_to_str(opts + 1, options);
703  rb_str_buf_cat2(desc, opts);
704  return rb_exc_new3(rb_eRegexpError, desc);
705 }
706 
707 NORETURN(static void rb_enc_reg_raise(const char *s, long len, rb_encoding *enc, int options, const char *err));
708 
709 static void
710 rb_enc_reg_raise(const char *s, long len, rb_encoding *enc, int options, const char *err)
711 {
712  rb_exc_raise(rb_enc_reg_error_desc(s, len, enc, options, err));
713 }
714 
715 static VALUE
716 rb_reg_error_desc(VALUE str, int options, const char *err)
717 {
718  return rb_enc_reg_error_desc(RSTRING_PTR(str), RSTRING_LEN(str),
719  rb_enc_get(str), options, err);
720 }
721 
722 NORETURN(static void rb_reg_raise_str(VALUE str, int options, const char *err));
723 
724 static void
725 rb_reg_raise_str(VALUE str, int options, const char *err)
726 {
727  rb_exc_raise(rb_reg_error_desc(str, options, err));
728 }
729 
730 
731 /*
732  * call-seq:
733  * casefold?-> true or false
734  *
735  * Returns +true+ if the case-insensitivity flag in +self+ is set,
736  * +false+ otherwise:
737  *
738  * /a/.casefold? # => false
739  * /a/i.casefold? # => true
740  * /(?i:a)/.casefold? # => false
741  *
742  */
743 
744 static VALUE
745 rb_reg_casefold_p(VALUE re)
746 {
747  rb_reg_check(re);
748  return RBOOL(RREGEXP_PTR(re)->options & ONIG_OPTION_IGNORECASE);
749 }
750 
751 
752 /*
753  * call-seq:
754  * options -> integer
755  *
756  * Returns an integer whose bits show the options set in +self+.
757  *
758  * The option bits are:
759  *
760  * Regexp::IGNORECASE # => 1
761  * Regexp::EXTENDED # => 2
762  * Regexp::MULTILINE # => 4
763  *
764  * Examples:
765  *
766  * /foo/.options # => 0
767  * /foo/i.options # => 1
768  * /foo/x.options # => 2
769  * /foo/m.options # => 4
770  * /foo/mix.options # => 7
771  *
772  * Note that additional bits may be set in the returned integer;
773  * these are maintained internally in +self+, are ignored if passed
774  * to Regexp.new, and may be ignored by the caller:
775  *
776  * Returns the set of bits corresponding to the options used when
777  * creating this regexp (see Regexp::new for details). Note that
778  * additional bits may be set in the returned options: these are used
779  * internally by the regular expression code. These extra bits are
780  * ignored if the options are passed to Regexp::new:
781  *
782  * r = /\xa1\xa2/e # => /\xa1\xa2/
783  * r.source # => "\\xa1\\xa2"
784  * r.options # => 16
785  * Regexp.new(r.source, r.options) # => /\xa1\xa2/
786  *
787  */
788 
789 static VALUE
790 rb_reg_options_m(VALUE re)
791 {
792  int options = rb_reg_options(re);
793  return INT2NUM(options);
794 }
795 
796 static int
797 reg_names_iter(const OnigUChar *name, const OnigUChar *name_end,
798  int back_num, int *back_refs, OnigRegex regex, void *arg)
799 {
800  VALUE ary = (VALUE)arg;
801  rb_ary_push(ary, rb_enc_str_new((const char *)name, name_end-name, regex->enc));
802  return 0;
803 }
804 
805 /*
806  * call-seq:
807  * names -> array_of_names
808  *
809  * Returns an array of names of captures
810  * (see {Named Captures}[rdoc-ref:Regexp@Named+Captures]):
811  *
812  * /(?<foo>.)(?<bar>.)(?<baz>.)/.names # => ["foo", "bar", "baz"]
813  * /(?<foo>.)(?<foo>.)/.names # => ["foo"]
814  * /(.)(.)/.names # => []
815  *
816  */
817 
818 static VALUE
819 rb_reg_names(VALUE re)
820 {
821  VALUE ary;
822  rb_reg_check(re);
823  ary = rb_ary_new_capa(onig_number_of_names(RREGEXP_PTR(re)));
824  onig_foreach_name(RREGEXP_PTR(re), reg_names_iter, (void*)ary);
825  return ary;
826 }
827 
828 static int
829 reg_named_captures_iter(const OnigUChar *name, const OnigUChar *name_end,
830  int back_num, int *back_refs, OnigRegex regex, void *arg)
831 {
832  VALUE hash = (VALUE)arg;
833  VALUE ary = rb_ary_new2(back_num);
834  int i;
835 
836  for (i = 0; i < back_num; i++)
837  rb_ary_store(ary, i, INT2NUM(back_refs[i]));
838 
839  rb_hash_aset(hash, rb_str_new((const char*)name, name_end-name),ary);
840 
841  return 0;
842 }
843 
844 /*
845  * call-seq:
846  * named_captures -> hash
847  *
848  * Returns a hash representing named captures of +self+
849  * (see {Named Captures}[rdoc-ref:Regexp@Named+Captures]):
850  *
851  * - Each key is the name of a named capture.
852  * - Each value is an array of integer indexes for that named capture.
853  *
854  * Examples:
855  *
856  * /(?<foo>.)(?<bar>.)/.named_captures # => {"foo"=>[1], "bar"=>[2]}
857  * /(?<foo>.)(?<foo>.)/.named_captures # => {"foo"=>[1, 2]}
858  * /(.)(.)/.named_captures # => {}
859  *
860  */
861 
862 static VALUE
863 rb_reg_named_captures(VALUE re)
864 {
865  regex_t *reg = (rb_reg_check(re), RREGEXP_PTR(re));
866  VALUE hash = rb_hash_new_with_size(onig_number_of_names(reg));
867  onig_foreach_name(reg, reg_named_captures_iter, (void*)hash);
868  return hash;
869 }
870 
871 static int
872 onig_new_with_source(regex_t** reg, const UChar* pattern, const UChar* pattern_end,
873  OnigOptionType option, OnigEncoding enc, const OnigSyntaxType* syntax,
874  OnigErrorInfo* einfo, const char *sourcefile, int sourceline)
875 {
876  int r;
877 
878  *reg = (regex_t* )malloc(sizeof(regex_t));
879  if (IS_NULL(*reg)) return ONIGERR_MEMORY;
880 
881  r = onig_reg_init(*reg, option, ONIGENC_CASE_FOLD_DEFAULT, enc, syntax);
882  if (r) goto err;
883 
884  r = onig_compile_ruby(*reg, pattern, pattern_end, einfo, sourcefile, sourceline);
885  if (r) {
886  err:
887  onig_free(*reg);
888  *reg = NULL;
889  }
890  return r;
891 }
892 
893 static Regexp*
894 make_regexp(const char *s, long len, rb_encoding *enc, int flags, onig_errmsg_buffer err,
895  const char *sourcefile, int sourceline)
896 {
897  Regexp *rp;
898  int r;
899  OnigErrorInfo einfo;
900 
901  /* Handle escaped characters first. */
902 
903  /* Build a copy of the string (in dest) with the
904  escaped characters translated, and generate the regex
905  from that.
906  */
907 
908  r = onig_new_with_source(&rp, (UChar*)s, (UChar*)(s + len), flags,
909  enc, OnigDefaultSyntax, &einfo, sourcefile, sourceline);
910  if (r) {
911  onig_error_code_to_str((UChar*)err, r, &einfo);
912  return 0;
913  }
914  return rp;
915 }
916 
917 
918 /*
919  * Document-class: MatchData
920  *
921  * MatchData encapsulates the result of matching a Regexp against
922  * string. It is returned by Regexp#match and String#match, and also
923  * stored in a global variable returned by Regexp.last_match.
924  *
925  * Usage:
926  *
927  * url = 'https://docs.ruby-lang.org/en/2.5.0/MatchData.html'
928  * m = url.match(/(\d\.?)+/) # => #<MatchData "2.5.0" 1:"0">
929  * m.string # => "https://docs.ruby-lang.org/en/2.5.0/MatchData.html"
930  * m.regexp # => /(\d\.?)+/
931  * # entire matched substring:
932  * m[0] # => "2.5.0"
933  *
934  * # Working with unnamed captures
935  * m = url.match(%r{([^/]+)/([^/]+)\.html$})
936  * m.captures # => ["2.5.0", "MatchData"]
937  * m[1] # => "2.5.0"
938  * m.values_at(1, 2) # => ["2.5.0", "MatchData"]
939  *
940  * # Working with named captures
941  * m = url.match(%r{(?<version>[^/]+)/(?<module>[^/]+)\.html$})
942  * m.captures # => ["2.5.0", "MatchData"]
943  * m.named_captures # => {"version"=>"2.5.0", "module"=>"MatchData"}
944  * m[:version] # => "2.5.0"
945  * m.values_at(:version, :module)
946  * # => ["2.5.0", "MatchData"]
947  * # Numerical indexes are working, too
948  * m[1] # => "2.5.0"
949  * m.values_at(1, 2) # => ["2.5.0", "MatchData"]
950  *
951  * == Global variables equivalence
952  *
953  * Parts of last MatchData (returned by Regexp.last_match) are also
954  * aliased as global variables:
955  *
956  * * <code>$~</code> is Regexp.last_match;
957  * * <code>$&</code> is Regexp.last_match<code>[ 0 ]</code>;
958  * * <code>$1</code>, <code>$2</code>, and so on are
959  * Regexp.last_match<code>[ i ]</code> (captures by number);
960  * * <code>$`</code> is Regexp.last_match<code>.pre_match</code>;
961  * * <code>$'</code> is Regexp.last_match<code>.post_match</code>;
962  * * <code>$+</code> is Regexp.last_match<code>[ -1 ]</code> (the last capture).
963  *
964  * See also "Special global variables" section in Regexp documentation.
965  */
966 
968 
969 static VALUE
970 match_alloc(VALUE klass)
971 {
972  size_t alloc_size = sizeof(struct RMatch) + sizeof(rb_matchext_t);
974  NEWOBJ_OF(match, struct RMatch, klass, flags, alloc_size, 0);
975 
976  match->str = Qfalse;
977  match->regexp = Qfalse;
978  memset(RMATCH_EXT(match), 0, sizeof(rb_matchext_t));
979 
980  return (VALUE)match;
981 }
982 
983 int
984 rb_reg_region_copy(struct re_registers *to, const struct re_registers *from)
985 {
986  onig_region_copy(to, (OnigRegion *)from);
987  if (to->allocated) return 0;
988  rb_gc();
989  onig_region_copy(to, (OnigRegion *)from);
990  if (to->allocated) return 0;
991  return ONIGERR_MEMORY;
992 }
993 
994 typedef struct {
995  long byte_pos;
996  long char_pos;
997 } pair_t;
998 
999 static int
1000 pair_byte_cmp(const void *pair1, const void *pair2)
1001 {
1002  long diff = ((pair_t*)pair1)->byte_pos - ((pair_t*)pair2)->byte_pos;
1003 #if SIZEOF_LONG > SIZEOF_INT
1004  return diff ? diff > 0 ? 1 : -1 : 0;
1005 #else
1006  return (int)diff;
1007 #endif
1008 }
1009 
1010 static void
1011 update_char_offset(VALUE match)
1012 {
1013  rb_matchext_t *rm = RMATCH_EXT(match);
1014  struct re_registers *regs;
1015  int i, num_regs, num_pos;
1016  long c;
1017  char *s, *p, *q;
1018  rb_encoding *enc;
1019  pair_t *pairs;
1020 
1021  if (rm->char_offset_num_allocated)
1022  return;
1023 
1024  regs = &rm->regs;
1025  num_regs = rm->regs.num_regs;
1026 
1027  if (rm->char_offset_num_allocated < num_regs) {
1028  REALLOC_N(rm->char_offset, struct rmatch_offset, num_regs);
1029  rm->char_offset_num_allocated = num_regs;
1030  }
1031 
1032  enc = rb_enc_get(RMATCH(match)->str);
1033  if (rb_enc_mbmaxlen(enc) == 1) {
1034  for (i = 0; i < num_regs; i++) {
1035  rm->char_offset[i].beg = BEG(i);
1036  rm->char_offset[i].end = END(i);
1037  }
1038  return;
1039  }
1040 
1041  pairs = ALLOCA_N(pair_t, num_regs*2);
1042  num_pos = 0;
1043  for (i = 0; i < num_regs; i++) {
1044  if (BEG(i) < 0)
1045  continue;
1046  pairs[num_pos++].byte_pos = BEG(i);
1047  pairs[num_pos++].byte_pos = END(i);
1048  }
1049  qsort(pairs, num_pos, sizeof(pair_t), pair_byte_cmp);
1050 
1051  s = p = RSTRING_PTR(RMATCH(match)->str);
1052  c = 0;
1053  for (i = 0; i < num_pos; i++) {
1054  q = s + pairs[i].byte_pos;
1055  c += rb_enc_strlen(p, q, enc);
1056  pairs[i].char_pos = c;
1057  p = q;
1058  }
1059 
1060  for (i = 0; i < num_regs; i++) {
1061  pair_t key, *found;
1062  if (BEG(i) < 0) {
1063  rm->char_offset[i].beg = -1;
1064  rm->char_offset[i].end = -1;
1065  continue;
1066  }
1067 
1068  key.byte_pos = BEG(i);
1069  found = bsearch(&key, pairs, num_pos, sizeof(pair_t), pair_byte_cmp);
1070  rm->char_offset[i].beg = found->char_pos;
1071 
1072  key.byte_pos = END(i);
1073  found = bsearch(&key, pairs, num_pos, sizeof(pair_t), pair_byte_cmp);
1074  rm->char_offset[i].end = found->char_pos;
1075  }
1076 }
1077 
1078 static VALUE
1079 match_check(VALUE match)
1080 {
1081  if (!RMATCH(match)->regexp) {
1082  rb_raise(rb_eTypeError, "uninitialized MatchData");
1083  }
1084  return match;
1085 }
1086 
1087 /* :nodoc: */
1088 static VALUE
1089 match_init_copy(VALUE obj, VALUE orig)
1090 {
1091  rb_matchext_t *rm;
1092 
1093  if (!OBJ_INIT_COPY(obj, orig)) return obj;
1094 
1095  RB_OBJ_WRITE(obj, &RMATCH(obj)->str, RMATCH(orig)->str);
1096  RB_OBJ_WRITE(obj, &RMATCH(obj)->regexp, RMATCH(orig)->regexp);
1097 
1098  rm = RMATCH_EXT(obj);
1099  if (rb_reg_region_copy(&rm->regs, RMATCH_REGS(orig)))
1100  rb_memerror();
1101 
1102  if (RMATCH_EXT(orig)->char_offset_num_allocated) {
1103  if (rm->char_offset_num_allocated < rm->regs.num_regs) {
1104  REALLOC_N(rm->char_offset, struct rmatch_offset, rm->regs.num_regs);
1105  rm->char_offset_num_allocated = rm->regs.num_regs;
1106  }
1107  MEMCPY(rm->char_offset, RMATCH_EXT(orig)->char_offset,
1108  struct rmatch_offset, rm->regs.num_regs);
1109  RB_GC_GUARD(orig);
1110  }
1111 
1112  return obj;
1113 }
1114 
1115 
1116 /*
1117  * call-seq:
1118  * regexp -> regexp
1119  *
1120  * Returns the regexp that produced the match:
1121  *
1122  * m = /a.*b/.match("abc") # => #<MatchData "ab">
1123  * m.regexp # => /a.*b/
1124  *
1125  */
1126 
1127 static VALUE
1128 match_regexp(VALUE match)
1129 {
1130  VALUE regexp;
1131  match_check(match);
1132  regexp = RMATCH(match)->regexp;
1133  if (NIL_P(regexp)) {
1134  VALUE str = rb_reg_nth_match(0, match);
1135  regexp = rb_reg_regcomp(rb_reg_quote(str));
1136  RB_OBJ_WRITE(match, &RMATCH(match)->regexp, regexp);
1137  }
1138  return regexp;
1139 }
1140 
1141 /*
1142  * call-seq:
1143  * names -> array_of_names
1144  *
1145  * Returns an array of the capture names
1146  * (see {Named Captures}[rdoc-ref:Regexp@Named+Captures]):
1147  *
1148  * m = /(?<foo>.)(?<bar>.)(?<baz>.)/.match("hoge")
1149  * # => #<MatchData "hog" foo:"h" bar:"o" baz:"g">
1150  * m.names # => ["foo", "bar", "baz"]
1151  *
1152  * m = /foo/.match('foo') # => #<MatchData "foo">
1153  * m.names # => [] # No named captures.
1154  *
1155  * Equivalent to:
1156  *
1157  * m = /(?<foo>.)(?<bar>.)(?<baz>.)/.match("hoge")
1158  * m.regexp.names # => ["foo", "bar", "baz"]
1159  *
1160  */
1161 
1162 static VALUE
1163 match_names(VALUE match)
1164 {
1165  match_check(match);
1166  if (NIL_P(RMATCH(match)->regexp))
1167  return rb_ary_new_capa(0);
1168  return rb_reg_names(RMATCH(match)->regexp);
1169 }
1170 
1171 /*
1172  * call-seq:
1173  * size -> integer
1174  *
1175  * Returns size of the match array:
1176  *
1177  * m = /(.)(.)(\d+)(\d)/.match("THX1138.")
1178  * # => #<MatchData "HX1138" 1:"H" 2:"X" 3:"113" 4:"8">
1179  * m.size # => 5
1180  *
1181  */
1182 
1183 static VALUE
1184 match_size(VALUE match)
1185 {
1186  match_check(match);
1187  return INT2FIX(RMATCH_REGS(match)->num_regs);
1188 }
1189 
1190 static int name_to_backref_number(struct re_registers *, VALUE, const char*, const char*);
1191 NORETURN(static void name_to_backref_error(VALUE name));
1192 
1193 static void
1194 name_to_backref_error(VALUE name)
1195 {
1196  rb_raise(rb_eIndexError, "undefined group name reference: % "PRIsVALUE,
1197  name);
1198 }
1199 
1200 static void
1201 backref_number_check(struct re_registers *regs, int i)
1202 {
1203  if (i < 0 || regs->num_regs <= i)
1204  rb_raise(rb_eIndexError, "index %d out of matches", i);
1205 }
1206 
1207 static int
1208 match_backref_number(VALUE match, VALUE backref)
1209 {
1210  const char *name;
1211  int num;
1212 
1213  struct re_registers *regs = RMATCH_REGS(match);
1214  VALUE regexp = RMATCH(match)->regexp;
1215 
1216  match_check(match);
1217  if (SYMBOL_P(backref)) {
1218  backref = rb_sym2str(backref);
1219  }
1220  else if (!RB_TYPE_P(backref, T_STRING)) {
1221  return NUM2INT(backref);
1222  }
1223  name = StringValueCStr(backref);
1224 
1225  num = name_to_backref_number(regs, regexp, name, name + RSTRING_LEN(backref));
1226 
1227  if (num < 1) {
1228  name_to_backref_error(backref);
1229  }
1230 
1231  return num;
1232 }
1233 
1234 int
1236 {
1237  return match_backref_number(match, backref);
1238 }
1239 
1240 /*
1241  * call-seq:
1242  * offset(n) -> [start_offset, end_offset]
1243  * offset(name) -> [start_offset, end_offset]
1244  *
1245  * :include: doc/matchdata/offset.rdoc
1246  *
1247  */
1248 
1249 static VALUE
1250 match_offset(VALUE match, VALUE n)
1251 {
1252  int i = match_backref_number(match, n);
1253  struct re_registers *regs = RMATCH_REGS(match);
1254 
1255  match_check(match);
1256  backref_number_check(regs, i);
1257 
1258  if (BEG(i) < 0)
1259  return rb_assoc_new(Qnil, Qnil);
1260 
1261  update_char_offset(match);
1262  return rb_assoc_new(LONG2NUM(RMATCH_EXT(match)->char_offset[i].beg),
1263  LONG2NUM(RMATCH_EXT(match)->char_offset[i].end));
1264 }
1265 
1266 /*
1267  * call-seq:
1268  * mtch.byteoffset(n) -> array
1269  *
1270  * Returns a two-element array containing the beginning and ending byte-based offsets of
1271  * the <em>n</em>th match.
1272  * <em>n</em> can be a string or symbol to reference a named capture.
1273  *
1274  * m = /(.)(.)(\d+)(\d)/.match("THX1138.")
1275  * m.byteoffset(0) #=> [1, 7]
1276  * m.byteoffset(4) #=> [6, 7]
1277  *
1278  * m = /(?<foo>.)(.)(?<bar>.)/.match("hoge")
1279  * p m.byteoffset(:foo) #=> [0, 1]
1280  * p m.byteoffset(:bar) #=> [2, 3]
1281  *
1282  */
1283 
1284 static VALUE
1285 match_byteoffset(VALUE match, VALUE n)
1286 {
1287  int i = match_backref_number(match, n);
1288  struct re_registers *regs = RMATCH_REGS(match);
1289 
1290  match_check(match);
1291  backref_number_check(regs, i);
1292 
1293  if (BEG(i) < 0)
1294  return rb_assoc_new(Qnil, Qnil);
1295  return rb_assoc_new(LONG2NUM(BEG(i)), LONG2NUM(END(i)));
1296 }
1297 
1298 
1299 /*
1300  * call-seq:
1301  * bytebegin(n) -> integer
1302  * bytebegin(name) -> integer
1303  *
1304  * :include: doc/matchdata/bytebegin.rdoc
1305  *
1306  */
1307 
1308 static VALUE
1309 match_bytebegin(VALUE match, VALUE n)
1310 {
1311  int i = match_backref_number(match, n);
1312  struct re_registers *regs = RMATCH_REGS(match);
1313 
1314  match_check(match);
1315  backref_number_check(regs, i);
1316 
1317  if (BEG(i) < 0)
1318  return Qnil;
1319  return LONG2NUM(BEG(i));
1320 }
1321 
1322 
1323 /*
1324  * call-seq:
1325  * byteend(n) -> integer
1326  * byteend(name) -> integer
1327  *
1328  * :include: doc/matchdata/byteend.rdoc
1329  *
1330  */
1331 
1332 static VALUE
1333 match_byteend(VALUE match, VALUE n)
1334 {
1335  int i = match_backref_number(match, n);
1336  struct re_registers *regs = RMATCH_REGS(match);
1337 
1338  match_check(match);
1339  backref_number_check(regs, i);
1340 
1341  if (BEG(i) < 0)
1342  return Qnil;
1343  return LONG2NUM(END(i));
1344 }
1345 
1346 
1347 /*
1348  * call-seq:
1349  * begin(n) -> integer
1350  * begin(name) -> integer
1351  *
1352  * :include: doc/matchdata/begin.rdoc
1353  *
1354  */
1355 
1356 static VALUE
1357 match_begin(VALUE match, VALUE n)
1358 {
1359  int i = match_backref_number(match, n);
1360  struct re_registers *regs = RMATCH_REGS(match);
1361 
1362  match_check(match);
1363  backref_number_check(regs, i);
1364 
1365  if (BEG(i) < 0)
1366  return Qnil;
1367 
1368  update_char_offset(match);
1369  return LONG2NUM(RMATCH_EXT(match)->char_offset[i].beg);
1370 }
1371 
1372 
1373 /*
1374  * call-seq:
1375  * end(n) -> integer
1376  * end(name) -> integer
1377  *
1378  * :include: doc/matchdata/end.rdoc
1379  *
1380  */
1381 
1382 static VALUE
1383 match_end(VALUE match, VALUE n)
1384 {
1385  int i = match_backref_number(match, n);
1386  struct re_registers *regs = RMATCH_REGS(match);
1387 
1388  match_check(match);
1389  backref_number_check(regs, i);
1390 
1391  if (BEG(i) < 0)
1392  return Qnil;
1393 
1394  update_char_offset(match);
1395  return LONG2NUM(RMATCH_EXT(match)->char_offset[i].end);
1396 }
1397 
1398 /*
1399  * call-seq:
1400  * match(n) -> string or nil
1401  * match(name) -> string or nil
1402  *
1403  * Returns the matched substring corresponding to the given argument.
1404  *
1405  * When non-negative argument +n+ is given,
1406  * returns the matched substring for the <tt>n</tt>th match:
1407  *
1408  * m = /(.)(.)(\d+)(\d)(\w)?/.match("THX1138.")
1409  * # => #<MatchData "HX1138" 1:"H" 2:"X" 3:"113" 4:"8" 5:nil>
1410  * m.match(0) # => "HX1138"
1411  * m.match(4) # => "8"
1412  * m.match(5) # => nil
1413  *
1414  * When string or symbol argument +name+ is given,
1415  * returns the matched substring for the given name:
1416  *
1417  * m = /(?<foo>.)(.)(?<bar>.+)/.match("hoge")
1418  * # => #<MatchData "hoge" foo:"h" bar:"ge">
1419  * m.match('foo') # => "h"
1420  * m.match(:bar) # => "ge"
1421  *
1422  */
1423 
1424 static VALUE
1425 match_nth(VALUE match, VALUE n)
1426 {
1427  int i = match_backref_number(match, n);
1428  struct re_registers *regs = RMATCH_REGS(match);
1429 
1430  backref_number_check(regs, i);
1431 
1432  long start = BEG(i), end = END(i);
1433  if (start < 0)
1434  return Qnil;
1435 
1436  return rb_str_subseq(RMATCH(match)->str, start, end - start);
1437 }
1438 
1439 /*
1440  * call-seq:
1441  * match_length(n) -> integer or nil
1442  * match_length(name) -> integer or nil
1443  *
1444  * Returns the length (in characters) of the matched substring
1445  * corresponding to the given argument.
1446  *
1447  * When non-negative argument +n+ is given,
1448  * returns the length of the matched substring
1449  * for the <tt>n</tt>th match:
1450  *
1451  * m = /(.)(.)(\d+)(\d)(\w)?/.match("THX1138.")
1452  * # => #<MatchData "HX1138" 1:"H" 2:"X" 3:"113" 4:"8" 5:nil>
1453  * m.match_length(0) # => 6
1454  * m.match_length(4) # => 1
1455  * m.match_length(5) # => nil
1456  *
1457  * When string or symbol argument +name+ is given,
1458  * returns the length of the matched substring
1459  * for the named match:
1460  *
1461  * m = /(?<foo>.)(.)(?<bar>.+)/.match("hoge")
1462  * # => #<MatchData "hoge" foo:"h" bar:"ge">
1463  * m.match_length('foo') # => 1
1464  * m.match_length(:bar) # => 2
1465  *
1466  */
1467 
1468 static VALUE
1469 match_nth_length(VALUE match, VALUE n)
1470 {
1471  int i = match_backref_number(match, n);
1472  struct re_registers *regs = RMATCH_REGS(match);
1473 
1474  match_check(match);
1475  backref_number_check(regs, i);
1476 
1477  if (BEG(i) < 0)
1478  return Qnil;
1479 
1480  update_char_offset(match);
1481  const struct rmatch_offset *const ofs =
1482  &RMATCH_EXT(match)->char_offset[i];
1483  return LONG2NUM(ofs->end - ofs->beg);
1484 }
1485 
1486 #define MATCH_BUSY FL_USER2
1487 
1488 void
1490 {
1491  FL_SET(match, MATCH_BUSY);
1492 }
1493 
1494 void
1495 rb_match_unbusy(VALUE match)
1496 {
1497  FL_UNSET(match, MATCH_BUSY);
1498 }
1499 
1500 int
1501 rb_match_count(VALUE match)
1502 {
1503  struct re_registers *regs;
1504  if (NIL_P(match)) return -1;
1505  regs = RMATCH_REGS(match);
1506  if (!regs) return -1;
1507  return regs->num_regs;
1508 }
1509 
1510 static void
1511 match_set_string(VALUE m, VALUE string, long pos, long len)
1512 {
1513  struct RMatch *match = (struct RMatch *)m;
1514  rb_matchext_t *rmatch = RMATCH_EXT(match);
1515 
1516  RB_OBJ_WRITE(match, &RMATCH(match)->str, string);
1517  RB_OBJ_WRITE(match, &RMATCH(match)->regexp, Qnil);
1518  int err = onig_region_resize(&rmatch->regs, 1);
1519  if (err) rb_memerror();
1520  rmatch->regs.beg[0] = pos;
1521  rmatch->regs.end[0] = pos + len;
1522 }
1523 
1524 void
1525 rb_backref_set_string(VALUE string, long pos, long len)
1526 {
1527  VALUE match = rb_backref_get();
1528  if (NIL_P(match) || FL_TEST(match, MATCH_BUSY)) {
1529  match = match_alloc(rb_cMatch);
1530  }
1531  match_set_string(match, string, pos, len);
1532  rb_backref_set(match);
1533 }
1534 
1535 /*
1536  * call-seq:
1537  * fixed_encoding? -> true or false
1538  *
1539  * Returns +false+ if +self+ is applicable to
1540  * a string with any ASCII-compatible encoding;
1541  * otherwise returns +true+:
1542  *
1543  * r = /a/ # => /a/
1544  * r.fixed_encoding? # => false
1545  * r.match?("\u{6666} a") # => true
1546  * r.match?("\xa1\xa2 a".force_encoding("euc-jp")) # => true
1547  * r.match?("abc".force_encoding("euc-jp")) # => true
1548  *
1549  * r = /a/u # => /a/
1550  * r.fixed_encoding? # => true
1551  * r.match?("\u{6666} a") # => true
1552  * r.match?("\xa1\xa2".force_encoding("euc-jp")) # Raises exception.
1553  * r.match?("abc".force_encoding("euc-jp")) # => true
1554  *
1555  * r = /\u{6666}/ # => /\u{6666}/
1556  * r.fixed_encoding? # => true
1557  * r.encoding # => #<Encoding:UTF-8>
1558  * r.match?("\u{6666} a") # => true
1559  * r.match?("\xa1\xa2".force_encoding("euc-jp")) # Raises exception.
1560  * r.match?("abc".force_encoding("euc-jp")) # => false
1561  *
1562  */
1563 
1564 static VALUE
1565 rb_reg_fixed_encoding_p(VALUE re)
1566 {
1567  return RBOOL(FL_TEST(re, KCODE_FIXED));
1568 }
1569 
1570 static VALUE
1571 rb_reg_preprocess(const char *p, const char *end, rb_encoding *enc,
1572  rb_encoding **fixed_enc, onig_errmsg_buffer err, int options);
1573 
1574 NORETURN(static void reg_enc_error(VALUE re, VALUE str));
1575 
1576 static void
1577 reg_enc_error(VALUE re, VALUE str)
1578 {
1580  "incompatible encoding regexp match (%s regexp with %s string)",
1581  rb_enc_inspect_name(rb_enc_get(re)),
1582  rb_enc_inspect_name(rb_enc_get(str)));
1583 }
1584 
1585 static inline int
1586 str_coderange(VALUE str)
1587 {
1588  int cr = ENC_CODERANGE(str);
1589  if (cr == ENC_CODERANGE_UNKNOWN) {
1590  cr = rb_enc_str_coderange(str);
1591  }
1592  return cr;
1593 }
1594 
1595 static rb_encoding*
1596 rb_reg_prepare_enc(VALUE re, VALUE str, int warn)
1597 {
1598  rb_encoding *enc = 0;
1599  int cr = str_coderange(str);
1600 
1601  if (cr == ENC_CODERANGE_BROKEN) {
1603  "invalid byte sequence in %s",
1605  }
1606 
1607  rb_reg_check(re);
1608  enc = rb_enc_get(str);
1609  if (RREGEXP_PTR(re)->enc == enc) {
1610  }
1611  else if (cr == ENC_CODERANGE_7BIT &&
1612  RREGEXP_PTR(re)->enc == rb_usascii_encoding()) {
1613  enc = RREGEXP_PTR(re)->enc;
1614  }
1615  else if (!rb_enc_asciicompat(enc)) {
1616  reg_enc_error(re, str);
1617  }
1618  else if (rb_reg_fixed_encoding_p(re)) {
1619  if ((!rb_enc_asciicompat(RREGEXP_PTR(re)->enc) ||
1620  cr != ENC_CODERANGE_7BIT)) {
1621  reg_enc_error(re, str);
1622  }
1623  enc = RREGEXP_PTR(re)->enc;
1624  }
1625  else if (warn && (RBASIC(re)->flags & REG_ENCODING_NONE) &&
1626  enc != rb_ascii8bit_encoding() &&
1627  cr != ENC_CODERANGE_7BIT) {
1628  rb_warn("historical binary regexp match /.../n against %s string",
1629  rb_enc_name(enc));
1630  }
1631  return enc;
1632 }
1633 
1634 regex_t *
1636 {
1637  int r;
1638  OnigErrorInfo einfo;
1639  VALUE unescaped;
1640  rb_encoding *fixed_enc = 0;
1641  rb_encoding *enc = rb_reg_prepare_enc(re, str, 1);
1642 
1643  regex_t *reg = RREGEXP_PTR(re);
1644  if (reg->enc == enc) return reg;
1645 
1646  rb_reg_check(re);
1647 
1648  VALUE src_str = RREGEXP_SRC(re);
1649  const char *pattern = RSTRING_PTR(src_str);
1650 
1651  onig_errmsg_buffer err = "";
1652  unescaped = rb_reg_preprocess(
1653  pattern, pattern + RSTRING_LEN(src_str), enc,
1654  &fixed_enc, err, 0);
1655 
1656  if (NIL_P(unescaped)) {
1657  rb_raise(rb_eArgError, "regexp preprocess failed: %s", err);
1658  }
1659 
1660  // inherit the timeout settings
1661  rb_hrtime_t timelimit = reg->timelimit;
1662 
1663  const char *ptr;
1664  long len;
1665  RSTRING_GETMEM(unescaped, ptr, len);
1666 
1667  /* If there are no other users of this regex, then we can directly overwrite it. */
1668  if (RREGEXP(re)->usecnt == 0) {
1669  regex_t tmp_reg;
1670  r = onig_new_without_alloc(&tmp_reg, (UChar *)ptr, (UChar *)(ptr + len),
1671  reg->options, enc,
1672  OnigDefaultSyntax, &einfo);
1673 
1674  if (r) {
1675  /* There was an error so perform cleanups. */
1676  onig_free_body(&tmp_reg);
1677  }
1678  else {
1679  onig_free_body(reg);
1680  /* There are no errors so set reg to tmp_reg. */
1681  *reg = tmp_reg;
1682  }
1683  }
1684  else {
1685  r = onig_new(&reg, (UChar *)ptr, (UChar *)(ptr + len),
1686  reg->options, enc,
1687  OnigDefaultSyntax, &einfo);
1688  }
1689 
1690  if (r) {
1691  onig_error_code_to_str((UChar*)err, r, &einfo);
1692  rb_reg_raise(err, re);
1693  }
1694 
1695  reg->timelimit = timelimit;
1696 
1697  RB_GC_GUARD(unescaped);
1698  RB_GC_GUARD(src_str);
1699  return reg;
1700 }
1701 
1702 OnigPosition
1704  OnigPosition (*match)(regex_t *reg, VALUE str, struct re_registers *regs, void *args),
1705  void *args, struct re_registers *regs)
1706 {
1707  regex_t *reg = rb_reg_prepare_re(re, str);
1708 
1709  bool tmpreg = reg != RREGEXP_PTR(re);
1710  if (!tmpreg) RREGEXP(re)->usecnt++;
1711 
1712  OnigPosition result = match(reg, str, regs, args);
1713 
1714  if (!tmpreg) RREGEXP(re)->usecnt--;
1715  if (tmpreg) {
1716  onig_free(reg);
1717  }
1718 
1719  if (result < 0) {
1720  onig_region_free(regs, 0);
1721 
1722  switch (result) {
1723  case ONIG_MISMATCH:
1724  break;
1725  case ONIGERR_TIMEOUT:
1726  rb_raise(rb_eRegexpTimeoutError, "regexp match timeout");
1727  default: {
1728  onig_errmsg_buffer err = "";
1729  onig_error_code_to_str((UChar*)err, (int)result);
1730  rb_reg_raise(err, re);
1731  }
1732  }
1733  }
1734 
1735  return result;
1736 }
1737 
1738 long
1739 rb_reg_adjust_startpos(VALUE re, VALUE str, long pos, int reverse)
1740 {
1741  long range;
1742  rb_encoding *enc;
1743  UChar *p, *string;
1744 
1745  enc = rb_reg_prepare_enc(re, str, 0);
1746 
1747  if (reverse) {
1748  range = -pos;
1749  }
1750  else {
1751  range = RSTRING_LEN(str) - pos;
1752  }
1753 
1754  if (pos > 0 && ONIGENC_MBC_MAXLEN(enc) != 1 && pos < RSTRING_LEN(str)) {
1755  string = (UChar*)RSTRING_PTR(str);
1756 
1757  if (range > 0) {
1758  p = onigenc_get_right_adjust_char_head(enc, string, string + pos, string + RSTRING_LEN(str));
1759  }
1760  else {
1761  p = ONIGENC_LEFT_ADJUST_CHAR_HEAD(enc, string, string + pos, string + RSTRING_LEN(str));
1762  }
1763  return p - string;
1764  }
1765 
1766  return pos;
1767 }
1768 
1770  long pos;
1771  long range;
1772 };
1773 
1774 static OnigPosition
1775 reg_onig_search(regex_t *reg, VALUE str, struct re_registers *regs, void *args_ptr)
1776 {
1777  struct reg_onig_search_args *args = (struct reg_onig_search_args *)args_ptr;
1778  const char *ptr;
1779  long len;
1780  RSTRING_GETMEM(str, ptr, len);
1781 
1782  return onig_search(
1783  reg,
1784  (UChar *)ptr,
1785  (UChar *)(ptr + len),
1786  (UChar *)(ptr + args->pos),
1787  (UChar *)(ptr + args->range),
1788  regs,
1789  ONIG_OPTION_NONE);
1790 }
1791 
1792 /* returns byte offset */
1793 static long
1794 rb_reg_search_set_match(VALUE re, VALUE str, long pos, int reverse, int set_backref_str, VALUE *set_match)
1795 {
1796  long len = RSTRING_LEN(str);
1797  if (pos > len || pos < 0) {
1799  return -1;
1800  }
1801 
1802  struct reg_onig_search_args args = {
1803  .pos = pos,
1804  .range = reverse ? 0 : len,
1805  };
1806  struct re_registers regs = {0};
1807 
1808  OnigPosition result = rb_reg_onig_match(re, str, reg_onig_search, &args, &regs);
1809 
1810  if (result == ONIG_MISMATCH) {
1812  return ONIG_MISMATCH;
1813  }
1814 
1815  VALUE match = match_alloc(rb_cMatch);
1816  rb_matchext_t *rm = RMATCH_EXT(match);
1817  rm->regs = regs;
1818 
1819  if (set_backref_str) {
1820  RB_OBJ_WRITE(match, &RMATCH(match)->str, rb_str_new4(str));
1821  }
1822  else {
1823  /* Note that a MatchData object with RMATCH(match)->str == 0 is incomplete!
1824  * We need to hide the object from ObjectSpace.each_object.
1825  * https://bugs.ruby-lang.org/issues/19159
1826  */
1827  rb_obj_hide(match);
1828  }
1829 
1830  RB_OBJ_WRITE(match, &RMATCH(match)->regexp, re);
1831  rb_backref_set(match);
1832  if (set_match) *set_match = match;
1833 
1834  return result;
1835 }
1836 
1837 long
1838 rb_reg_search0(VALUE re, VALUE str, long pos, int reverse, int set_backref_str)
1839 {
1840  return rb_reg_search_set_match(re, str, pos, reverse, set_backref_str, NULL);
1841 }
1842 
1843 long
1844 rb_reg_search(VALUE re, VALUE str, long pos, int reverse)
1845 {
1846  return rb_reg_search0(re, str, pos, reverse, 1);
1847 }
1848 
1849 static OnigPosition
1850 reg_onig_match(regex_t *reg, VALUE str, struct re_registers *regs, void *_)
1851 {
1852  const char *ptr;
1853  long len;
1854  RSTRING_GETMEM(str, ptr, len);
1855 
1856  return onig_match(
1857  reg,
1858  (UChar *)ptr,
1859  (UChar *)(ptr + len),
1860  (UChar *)ptr,
1861  regs,
1862  ONIG_OPTION_NONE);
1863 }
1864 
1865 bool
1866 rb_reg_start_with_p(VALUE re, VALUE str)
1867 {
1868  VALUE match = rb_backref_get();
1869  if (NIL_P(match) || FL_TEST(match, MATCH_BUSY)) {
1870  match = match_alloc(rb_cMatch);
1871  }
1872 
1873  struct re_registers *regs = RMATCH_REGS(match);
1874 
1875  if (rb_reg_onig_match(re, str, reg_onig_match, NULL, regs) == ONIG_MISMATCH) {
1877  return false;
1878  }
1879 
1880  RB_OBJ_WRITE(match, &RMATCH(match)->str, rb_str_new4(str));
1881  RB_OBJ_WRITE(match, &RMATCH(match)->regexp, re);
1882  rb_backref_set(match);
1883 
1884  return true;
1885 }
1886 
1887 VALUE
1888 rb_reg_nth_defined(int nth, VALUE match)
1889 {
1890  struct re_registers *regs;
1891  if (NIL_P(match)) return Qnil;
1892  match_check(match);
1893  regs = RMATCH_REGS(match);
1894  if (nth >= regs->num_regs) {
1895  return Qnil;
1896  }
1897  if (nth < 0) {
1898  nth += regs->num_regs;
1899  if (nth <= 0) return Qnil;
1900  }
1901  return RBOOL(BEG(nth) != -1);
1902 }
1903 
1904 VALUE
1905 rb_reg_nth_match(int nth, VALUE match)
1906 {
1907  VALUE str;
1908  long start, end, len;
1909  struct re_registers *regs;
1910 
1911  if (NIL_P(match)) return Qnil;
1912  match_check(match);
1913  regs = RMATCH_REGS(match);
1914  if (nth >= regs->num_regs) {
1915  return Qnil;
1916  }
1917  if (nth < 0) {
1918  nth += regs->num_regs;
1919  if (nth <= 0) return Qnil;
1920  }
1921  start = BEG(nth);
1922  if (start == -1) return Qnil;
1923  end = END(nth);
1924  len = end - start;
1925  str = rb_str_subseq(RMATCH(match)->str, start, len);
1926  return str;
1927 }
1928 
1929 VALUE
1931 {
1932  return rb_reg_nth_match(0, match);
1933 }
1934 
1935 
1936 /*
1937  * call-seq:
1938  * pre_match -> string
1939  *
1940  * Returns the substring of the target string from its beginning
1941  * up to the first match in +self+ (that is, <tt>self[0]</tt>);
1942  * equivalent to regexp global variable <tt>$`</tt>:
1943  *
1944  * m = /(.)(.)(\d+)(\d)/.match("THX1138.")
1945  * # => #<MatchData "HX1138" 1:"H" 2:"X" 3:"113" 4:"8">
1946  * m[0] # => "HX1138"
1947  * m.pre_match # => "T"
1948  *
1949  * Related: MatchData#post_match.
1950  *
1951  */
1952 
1953 VALUE
1955 {
1956  VALUE str;
1957  struct re_registers *regs;
1958 
1959  if (NIL_P(match)) return Qnil;
1960  match_check(match);
1961  regs = RMATCH_REGS(match);
1962  if (BEG(0) == -1) return Qnil;
1963  str = rb_str_subseq(RMATCH(match)->str, 0, BEG(0));
1964  return str;
1965 }
1966 
1967 
1968 /*
1969  * call-seq:
1970  * post_match -> str
1971  *
1972  * Returns the substring of the target string from
1973  * the end of the first match in +self+ (that is, <tt>self[0]</tt>)
1974  * to the end of the string;
1975  * equivalent to regexp global variable <tt>$'</tt>:
1976  *
1977  * m = /(.)(.)(\d+)(\d)/.match("THX1138: The Movie")
1978  * # => #<MatchData "HX1138" 1:"H" 2:"X" 3:"113" 4:"8">
1979  * m[0] # => "HX1138"
1980  * m.post_match # => ": The Movie"\
1981  *
1982  * Related: MatchData.pre_match.
1983  *
1984  */
1985 
1986 VALUE
1988 {
1989  VALUE str;
1990  long pos;
1991  struct re_registers *regs;
1992 
1993  if (NIL_P(match)) return Qnil;
1994  match_check(match);
1995  regs = RMATCH_REGS(match);
1996  if (BEG(0) == -1) return Qnil;
1997  str = RMATCH(match)->str;
1998  pos = END(0);
1999  str = rb_str_subseq(str, pos, RSTRING_LEN(str) - pos);
2000  return str;
2001 }
2002 
2003 static int
2004 match_last_index(VALUE match)
2005 {
2006  int i;
2007  struct re_registers *regs;
2008 
2009  if (NIL_P(match)) return -1;
2010  match_check(match);
2011  regs = RMATCH_REGS(match);
2012  if (BEG(0) == -1) return -1;
2013 
2014  for (i=regs->num_regs-1; BEG(i) == -1 && i > 0; i--)
2015  ;
2016  return i;
2017 }
2018 
2019 VALUE
2021 {
2022  int i = match_last_index(match);
2023  if (i <= 0) return Qnil;
2024  struct re_registers *regs = RMATCH_REGS(match);
2025  return rb_str_subseq(RMATCH(match)->str, BEG(i), END(i) - BEG(i));
2026 }
2027 
2028 VALUE
2029 rb_reg_last_defined(VALUE match)
2030 {
2031  int i = match_last_index(match);
2032  if (i < 0) return Qnil;
2033  return RBOOL(i);
2034 }
2035 
2036 static VALUE
2037 last_match_getter(ID _x, VALUE *_y)
2038 {
2040 }
2041 
2042 static VALUE
2043 prematch_getter(ID _x, VALUE *_y)
2044 {
2045  return rb_reg_match_pre(rb_backref_get());
2046 }
2047 
2048 static VALUE
2049 postmatch_getter(ID _x, VALUE *_y)
2050 {
2052 }
2053 
2054 static VALUE
2055 last_paren_match_getter(ID _x, VALUE *_y)
2056 {
2058 }
2059 
2060 static VALUE
2061 match_array(VALUE match, int start)
2062 {
2063  struct re_registers *regs;
2064  VALUE ary;
2065  VALUE target;
2066  int i;
2067 
2068  match_check(match);
2069  regs = RMATCH_REGS(match);
2070  ary = rb_ary_new2(regs->num_regs);
2071  target = RMATCH(match)->str;
2072 
2073  for (i=start; i<regs->num_regs; i++) {
2074  if (regs->beg[i] == -1) {
2075  rb_ary_push(ary, Qnil);
2076  }
2077  else {
2078  VALUE str = rb_str_subseq(target, regs->beg[i], regs->end[i]-regs->beg[i]);
2079  rb_ary_push(ary, str);
2080  }
2081  }
2082  return ary;
2083 }
2084 
2085 
2086 /*
2087  * call-seq:
2088  * to_a -> array
2089  *
2090  * Returns the array of matches:
2091  *
2092  * m = /(.)(.)(\d+)(\d)/.match("THX1138.")
2093  * # => #<MatchData "HX1138" 1:"H" 2:"X" 3:"113" 4:"8">
2094  * m.to_a # => ["HX1138", "H", "X", "113", "8"]
2095  *
2096  * Related: MatchData#captures.
2097  *
2098  */
2099 
2100 static VALUE
2101 match_to_a(VALUE match)
2102 {
2103  return match_array(match, 0);
2104 }
2105 
2106 
2107 /*
2108  * call-seq:
2109  * captures -> array
2110  *
2111  * Returns the array of captures,
2112  * which are all matches except <tt>m[0]</tt>:
2113  *
2114  * m = /(.)(.)(\d+)(\d)/.match("THX1138.")
2115  * # => #<MatchData "HX1138" 1:"H" 2:"X" 3:"113" 4:"8">
2116  * m[0] # => "HX1138"
2117  * m.captures # => ["H", "X", "113", "8"]
2118  *
2119  * Related: MatchData.to_a.
2120  *
2121  */
2122 static VALUE
2123 match_captures(VALUE match)
2124 {
2125  return match_array(match, 1);
2126 }
2127 
2128 static int
2129 name_to_backref_number(struct re_registers *regs, VALUE regexp, const char* name, const char* name_end)
2130 {
2131  if (NIL_P(regexp)) return -1;
2132  return onig_name_to_backref_number(RREGEXP_PTR(regexp),
2133  (const unsigned char *)name, (const unsigned char *)name_end, regs);
2134 }
2135 
2136 #define NAME_TO_NUMBER(regs, re, name, name_ptr, name_end) \
2137  (NIL_P(re) ? 0 : \
2138  !rb_enc_compatible(RREGEXP_SRC(re), (name)) ? 0 : \
2139  name_to_backref_number((regs), (re), (name_ptr), (name_end)))
2140 
2141 static int
2142 namev_to_backref_number(struct re_registers *regs, VALUE re, VALUE name)
2143 {
2144  int num;
2145 
2146  if (SYMBOL_P(name)) {
2147  name = rb_sym2str(name);
2148  }
2149  else if (!RB_TYPE_P(name, T_STRING)) {
2150  return -1;
2151  }
2152  num = NAME_TO_NUMBER(regs, re, name,
2153  RSTRING_PTR(name), RSTRING_END(name));
2154  if (num < 1) {
2155  name_to_backref_error(name);
2156  }
2157  return num;
2158 }
2159 
2160 static VALUE
2161 match_ary_subseq(VALUE match, long beg, long len, VALUE result)
2162 {
2163  long olen = RMATCH_REGS(match)->num_regs;
2164  long j, end = olen < beg+len ? olen : beg+len;
2165  if (NIL_P(result)) result = rb_ary_new_capa(len);
2166  if (len == 0) return result;
2167 
2168  for (j = beg; j < end; j++) {
2169  rb_ary_push(result, rb_reg_nth_match((int)j, match));
2170  }
2171  if (beg + len > j) {
2172  rb_ary_resize(result, RARRAY_LEN(result) + (beg + len) - j);
2173  }
2174  return result;
2175 }
2176 
2177 static VALUE
2178 match_ary_aref(VALUE match, VALUE idx, VALUE result)
2179 {
2180  long beg, len;
2181  int num_regs = RMATCH_REGS(match)->num_regs;
2182 
2183  /* check if idx is Range */
2184  switch (rb_range_beg_len(idx, &beg, &len, (long)num_regs, !NIL_P(result))) {
2185  case Qfalse:
2186  if (NIL_P(result)) return rb_reg_nth_match(NUM2INT(idx), match);
2187  rb_ary_push(result, rb_reg_nth_match(NUM2INT(idx), match));
2188  return result;
2189  case Qnil:
2190  return Qnil;
2191  default:
2192  return match_ary_subseq(match, beg, len, result);
2193  }
2194 }
2195 
2196 /*
2197  * call-seq:
2198  * matchdata[index] -> string or nil
2199  * matchdata[start, length] -> array
2200  * matchdata[range] -> array
2201  * matchdata[name] -> string or nil
2202  *
2203  * When arguments +index+, +start and +length+, or +range+ are given,
2204  * returns match and captures in the style of Array#[]:
2205  *
2206  * m = /(.)(.)(\d+)(\d)/.match("THX1138.")
2207  * # => #<MatchData "HX1138" 1:"H" 2:"X" 3:"113" 4:"8">
2208  * m[0] # => "HX1138"
2209  * m[1, 2] # => ["H", "X"]
2210  * m[1..3] # => ["H", "X", "113"]
2211  * m[-3, 2] # => ["X", "113"]
2212  *
2213  * When string or symbol argument +name+ is given,
2214  * returns the matched substring for the given name:
2215  *
2216  * m = /(?<foo>.)(.)(?<bar>.+)/.match("hoge")
2217  * # => #<MatchData "hoge" foo:"h" bar:"ge">
2218  * m['foo'] # => "h"
2219  * m[:bar] # => "ge"
2220  *
2221  * If multiple captures have the same name, returns the last matched
2222  * substring.
2223  *
2224  * m = /(?<foo>.)(?<foo>.+)/.match("hoge")
2225  * # => #<MatchData "hoge" foo:"h" foo:"oge">
2226  * m[:foo] #=> "oge"
2227  *
2228  * m = /\W(?<foo>.+)|\w(?<foo>.+)|(?<foo>.+)/.match("hoge")
2229  * #<MatchData "hoge" foo:nil foo:"oge" foo:nil>
2230  * m[:foo] #=> "oge"
2231  *
2232  */
2233 
2234 static VALUE
2235 match_aref(int argc, VALUE *argv, VALUE match)
2236 {
2237  VALUE idx, length;
2238 
2239  match_check(match);
2240  rb_scan_args(argc, argv, "11", &idx, &length);
2241 
2242  if (NIL_P(length)) {
2243  if (FIXNUM_P(idx)) {
2244  return rb_reg_nth_match(FIX2INT(idx), match);
2245  }
2246  else {
2247  int num = namev_to_backref_number(RMATCH_REGS(match), RMATCH(match)->regexp, idx);
2248  if (num >= 0) {
2249  return rb_reg_nth_match(num, match);
2250  }
2251  else {
2252  return match_ary_aref(match, idx, Qnil);
2253  }
2254  }
2255  }
2256  else {
2257  long beg = NUM2LONG(idx);
2258  long len = NUM2LONG(length);
2259  long num_regs = RMATCH_REGS(match)->num_regs;
2260  if (len < 0) {
2261  return Qnil;
2262  }
2263  if (beg < 0) {
2264  beg += num_regs;
2265  if (beg < 0) return Qnil;
2266  }
2267  else if (beg > num_regs) {
2268  return Qnil;
2269  }
2270  if (beg+len > num_regs) {
2271  len = num_regs - beg;
2272  }
2273  return match_ary_subseq(match, beg, len, Qnil);
2274  }
2275 }
2276 
2277 /*
2278  * call-seq:
2279  * values_at(*indexes) -> array
2280  *
2281  * Returns match and captures at the given +indexes+,
2282  * which may include any mixture of:
2283  *
2284  * - Integers.
2285  * - Ranges.
2286  * - Names (strings and symbols).
2287  *
2288  *
2289  * Examples:
2290  *
2291  * m = /(.)(.)(\d+)(\d)/.match("THX1138: The Movie")
2292  * # => #<MatchData "HX1138" 1:"H" 2:"X" 3:"113" 4:"8">
2293  * m.values_at(0, 2, -2) # => ["HX1138", "X", "113"]
2294  * m.values_at(1..2, -1) # => ["H", "X", "8"]
2295  *
2296  * m = /(?<a>\d+) *(?<op>[+\-*\/]) *(?<b>\d+)/.match("1 + 2")
2297  * # => #<MatchData "1 + 2" a:"1" op:"+" b:"2">
2298  * m.values_at(0, 1..2, :a, :b, :op)
2299  * # => ["1 + 2", "1", "+", "1", "2", "+"]
2300  *
2301  */
2302 
2303 static VALUE
2304 match_values_at(int argc, VALUE *argv, VALUE match)
2305 {
2306  VALUE result;
2307  int i;
2308 
2309  match_check(match);
2310  result = rb_ary_new2(argc);
2311 
2312  for (i=0; i<argc; i++) {
2313  if (FIXNUM_P(argv[i])) {
2314  rb_ary_push(result, rb_reg_nth_match(FIX2INT(argv[i]), match));
2315  }
2316  else {
2317  int num = namev_to_backref_number(RMATCH_REGS(match), RMATCH(match)->regexp, argv[i]);
2318  if (num >= 0) {
2319  rb_ary_push(result, rb_reg_nth_match(num, match));
2320  }
2321  else {
2322  match_ary_aref(match, argv[i], result);
2323  }
2324  }
2325  }
2326  return result;
2327 }
2328 
2329 
2330 /*
2331  * call-seq:
2332  * to_s -> string
2333  *
2334  * Returns the matched string:
2335  *
2336  * m = /(.)(.)(\d+)(\d)/.match("THX1138.")
2337  * # => #<MatchData "HX1138" 1:"H" 2:"X" 3:"113" 4:"8">
2338  * m.to_s # => "HX1138"
2339  *
2340  * m = /(?<foo>.)(.)(?<bar>.+)/.match("hoge")
2341  * # => #<MatchData "hoge" foo:"h" bar:"ge">
2342  * m.to_s # => "hoge"
2343  *
2344  * Related: MatchData.inspect.
2345  *
2346  */
2347 
2348 static VALUE
2349 match_to_s(VALUE match)
2350 {
2351  VALUE str = rb_reg_last_match(match_check(match));
2352 
2353  if (NIL_P(str)) str = rb_str_new(0,0);
2354  return str;
2355 }
2356 
2357 static int
2358 match_named_captures_iter(const OnigUChar *name, const OnigUChar *name_end,
2359  int back_num, int *back_refs, OnigRegex regex, void *arg)
2360 {
2361  struct MEMO *memo = MEMO_CAST(arg);
2362  VALUE hash = memo->v1;
2363  VALUE match = memo->v2;
2364  long symbolize = memo->u3.state;
2365 
2366  VALUE key = rb_enc_str_new((const char *)name, name_end-name, regex->enc);
2367 
2368  if (symbolize > 0) {
2369  key = rb_str_intern(key);
2370  }
2371 
2372  VALUE value;
2373 
2374  int i;
2375  int found = 0;
2376 
2377  for (i = 0; i < back_num; i++) {
2378  value = rb_reg_nth_match(back_refs[i], match);
2379  if (RTEST(value)) {
2380  rb_hash_aset(hash, key, value);
2381  found = 1;
2382  }
2383  }
2384 
2385  if (found == 0) {
2386  rb_hash_aset(hash, key, Qnil);
2387  }
2388 
2389  return 0;
2390 }
2391 
2392 /*
2393  * call-seq:
2394  * named_captures(symbolize_names: false) -> hash
2395  *
2396  * Returns a hash of the named captures;
2397  * each key is a capture name; each value is its captured string or +nil+:
2398  *
2399  * m = /(?<foo>.)(.)(?<bar>.+)/.match("hoge")
2400  * # => #<MatchData "hoge" foo:"h" bar:"ge">
2401  * m.named_captures # => {"foo"=>"h", "bar"=>"ge"}
2402  *
2403  * m = /(?<a>.)(?<b>.)/.match("01")
2404  * # => #<MatchData "01" a:"0" b:"1">
2405  * m.named_captures #=> {"a" => "0", "b" => "1"}
2406  *
2407  * m = /(?<a>.)(?<b>.)?/.match("0")
2408  * # => #<MatchData "0" a:"0" b:nil>
2409  * m.named_captures #=> {"a" => "0", "b" => nil}
2410  *
2411  * m = /(?<a>.)(?<a>.)/.match("01")
2412  * # => #<MatchData "01" a:"0" a:"1">
2413  * m.named_captures #=> {"a" => "1"}
2414  *
2415  * If keyword argument +symbolize_names+ is given
2416  * a true value, the keys in the resulting hash are Symbols:
2417  *
2418  * m = /(?<a>.)(?<a>.)/.match("01")
2419  * # => #<MatchData "01" a:"0" a:"1">
2420  * m.named_captures(symbolize_names: true) #=> {:a => "1"}
2421  *
2422  */
2423 
2424 static VALUE
2425 match_named_captures(int argc, VALUE *argv, VALUE match)
2426 {
2427  VALUE hash;
2428  struct MEMO *memo;
2429 
2430  match_check(match);
2431  if (NIL_P(RMATCH(match)->regexp))
2432  return rb_hash_new();
2433 
2434  VALUE opt;
2435  VALUE symbolize_names = 0;
2436 
2437  rb_scan_args(argc, argv, "0:", &opt);
2438 
2439  if (!NIL_P(opt)) {
2440  static ID keyword_ids[1];
2441 
2442  VALUE symbolize_names_val;
2443 
2444  if (!keyword_ids[0]) {
2445  keyword_ids[0] = rb_intern_const("symbolize_names");
2446  }
2447  rb_get_kwargs(opt, keyword_ids, 0, 1, &symbolize_names_val);
2448  if (!UNDEF_P(symbolize_names_val) && RTEST(symbolize_names_val)) {
2449  symbolize_names = 1;
2450  }
2451  }
2452 
2453  hash = rb_hash_new();
2454  memo = MEMO_NEW(hash, match, symbolize_names);
2455 
2456  onig_foreach_name(RREGEXP(RMATCH(match)->regexp)->ptr, match_named_captures_iter, (void*)memo);
2457 
2458  return hash;
2459 }
2460 
2461 /*
2462  * call-seq:
2463  * deconstruct_keys(array_of_names) -> hash
2464  *
2465  * Returns a hash of the named captures for the given names.
2466  *
2467  * m = /(?<hours>\d{2}):(?<minutes>\d{2}):(?<seconds>\d{2})/.match("18:37:22")
2468  * m.deconstruct_keys([:hours, :minutes]) # => {:hours => "18", :minutes => "37"}
2469  * m.deconstruct_keys(nil) # => {:hours => "18", :minutes => "37", :seconds => "22"}
2470  *
2471  * Returns an empty hash if no named captures were defined:
2472  *
2473  * m = /(\d{2}):(\d{2}):(\d{2})/.match("18:37:22")
2474  * m.deconstruct_keys(nil) # => {}
2475  *
2476  */
2477 static VALUE
2478 match_deconstruct_keys(VALUE match, VALUE keys)
2479 {
2480  VALUE h;
2481  long i;
2482 
2483  match_check(match);
2484 
2485  if (NIL_P(RMATCH(match)->regexp)) {
2486  return rb_hash_new_with_size(0);
2487  }
2488 
2489  if (NIL_P(keys)) {
2490  h = rb_hash_new_with_size(onig_number_of_names(RREGEXP_PTR(RMATCH(match)->regexp)));
2491 
2492  struct MEMO *memo;
2493  memo = MEMO_NEW(h, match, 1);
2494 
2495  onig_foreach_name(RREGEXP_PTR(RMATCH(match)->regexp), match_named_captures_iter, (void*)memo);
2496 
2497  return h;
2498  }
2499 
2500  Check_Type(keys, T_ARRAY);
2501 
2502  if (onig_number_of_names(RREGEXP_PTR(RMATCH(match)->regexp)) < RARRAY_LEN(keys)) {
2503  return rb_hash_new_with_size(0);
2504  }
2505 
2506  h = rb_hash_new_with_size(RARRAY_LEN(keys));
2507 
2508  for (i=0; i<RARRAY_LEN(keys); i++) {
2509  VALUE key = RARRAY_AREF(keys, i);
2510  VALUE name;
2511 
2512  Check_Type(key, T_SYMBOL);
2513 
2514  name = rb_sym2str(key);
2515 
2516  int num = NAME_TO_NUMBER(RMATCH_REGS(match), RMATCH(match)->regexp, RMATCH(match)->regexp,
2517  RSTRING_PTR(name), RSTRING_END(name));
2518 
2519  if (num >= 0) {
2520  rb_hash_aset(h, key, rb_reg_nth_match(num, match));
2521  }
2522  else {
2523  return h;
2524  }
2525  }
2526 
2527  return h;
2528 }
2529 
2530 /*
2531  * call-seq:
2532  * string -> string
2533  *
2534  * Returns the target string if it was frozen;
2535  * otherwise, returns a frozen copy of the target string:
2536  *
2537  * m = /(.)(.)(\d+)(\d)/.match("THX1138.")
2538  * # => #<MatchData "HX1138" 1:"H" 2:"X" 3:"113" 4:"8">
2539  * m.string # => "THX1138."
2540  *
2541  */
2542 
2543 static VALUE
2544 match_string(VALUE match)
2545 {
2546  match_check(match);
2547  return RMATCH(match)->str; /* str is frozen */
2548 }
2549 
2551  const UChar *name;
2552  long len;
2553 };
2554 
2555 static int
2556 match_inspect_name_iter(const OnigUChar *name, const OnigUChar *name_end,
2557  int back_num, int *back_refs, OnigRegex regex, void *arg0)
2558 {
2559  struct backref_name_tag *arg = (struct backref_name_tag *)arg0;
2560  int i;
2561 
2562  for (i = 0; i < back_num; i++) {
2563  arg[back_refs[i]].name = name;
2564  arg[back_refs[i]].len = name_end - name;
2565  }
2566  return 0;
2567 }
2568 
2569 /*
2570  * call-seq:
2571  * inspect -> string
2572  *
2573  * Returns a string representation of +self+:
2574  *
2575  * m = /.$/.match("foo")
2576  * # => #<MatchData "o">
2577  * m.inspect # => "#<MatchData \"o\">"
2578  *
2579  * m = /(.)(.)(.)/.match("foo")
2580  * # => #<MatchData "foo" 1:"f" 2:"o" 3:"o">
2581  * m.inspect # => "#<MatchData \"foo\" 1:\"f\" 2:\"o\
2582  *
2583  * m = /(.)(.)?(.)/.match("fo")
2584  * # => #<MatchData "fo" 1:"f" 2:nil 3:"o">
2585  * m.inspect # => "#<MatchData \"fo\" 1:\"f\" 2:nil 3:\"o\">"
2586  *
2587  * Related: MatchData#to_s.
2588  */
2589 
2590 static VALUE
2591 match_inspect(VALUE match)
2592 {
2593  VALUE cname = rb_class_path(rb_obj_class(match));
2594  VALUE str;
2595  int i;
2596  struct re_registers *regs = RMATCH_REGS(match);
2597  int num_regs = regs->num_regs;
2598  struct backref_name_tag *names;
2599  VALUE regexp = RMATCH(match)->regexp;
2600 
2601  if (regexp == 0) {
2602  return rb_sprintf("#<%"PRIsVALUE":%p>", cname, (void*)match);
2603  }
2604  else if (NIL_P(regexp)) {
2605  return rb_sprintf("#<%"PRIsVALUE": %"PRIsVALUE">",
2606  cname, rb_reg_nth_match(0, match));
2607  }
2608 
2609  names = ALLOCA_N(struct backref_name_tag, num_regs);
2610  MEMZERO(names, struct backref_name_tag, num_regs);
2611 
2612  onig_foreach_name(RREGEXP_PTR(regexp),
2613  match_inspect_name_iter, names);
2614 
2615  str = rb_str_buf_new2("#<");
2616  rb_str_append(str, cname);
2617 
2618  for (i = 0; i < num_regs; i++) {
2619  VALUE v;
2620  rb_str_buf_cat2(str, " ");
2621  if (0 < i) {
2622  if (names[i].name)
2623  rb_str_buf_cat(str, (const char *)names[i].name, names[i].len);
2624  else {
2625  rb_str_catf(str, "%d", i);
2626  }
2627  rb_str_buf_cat2(str, ":");
2628  }
2629  v = rb_reg_nth_match(i, match);
2630  if (NIL_P(v))
2631  rb_str_buf_cat2(str, "nil");
2632  else
2634  }
2635  rb_str_buf_cat2(str, ">");
2636 
2637  return str;
2638 }
2639 
2641 
2642 static int
2643 read_escaped_byte(const char **pp, const char *end, onig_errmsg_buffer err)
2644 {
2645  const char *p = *pp;
2646  int code;
2647  int meta_prefix = 0, ctrl_prefix = 0;
2648  size_t len;
2649 
2650  if (p == end || *p++ != '\\') {
2651  errcpy(err, "too short escaped multibyte character");
2652  return -1;
2653  }
2654 
2655 again:
2656  if (p == end) {
2657  errcpy(err, "too short escape sequence");
2658  return -1;
2659  }
2660  switch (*p++) {
2661  case '\\': code = '\\'; break;
2662  case 'n': code = '\n'; break;
2663  case 't': code = '\t'; break;
2664  case 'r': code = '\r'; break;
2665  case 'f': code = '\f'; break;
2666  case 'v': code = '\013'; break;
2667  case 'a': code = '\007'; break;
2668  case 'e': code = '\033'; break;
2669 
2670  /* \OOO */
2671  case '0': case '1': case '2': case '3':
2672  case '4': case '5': case '6': case '7':
2673  p--;
2674  code = scan_oct(p, end < p+3 ? end-p : 3, &len);
2675  p += len;
2676  break;
2677 
2678  case 'x': /* \xHH */
2679  code = scan_hex(p, end < p+2 ? end-p : 2, &len);
2680  if (len < 1) {
2681  errcpy(err, "invalid hex escape");
2682  return -1;
2683  }
2684  p += len;
2685  break;
2686 
2687  case 'M': /* \M-X, \M-\C-X, \M-\cX */
2688  if (meta_prefix) {
2689  errcpy(err, "duplicate meta escape");
2690  return -1;
2691  }
2692  meta_prefix = 1;
2693  if (p+1 < end && *p++ == '-' && (*p & 0x80) == 0) {
2694  if (*p == '\\') {
2695  p++;
2696  goto again;
2697  }
2698  else {
2699  code = *p++;
2700  break;
2701  }
2702  }
2703  errcpy(err, "too short meta escape");
2704  return -1;
2705 
2706  case 'C': /* \C-X, \C-\M-X */
2707  if (p == end || *p++ != '-') {
2708  errcpy(err, "too short control escape");
2709  return -1;
2710  }
2711  case 'c': /* \cX, \c\M-X */
2712  if (ctrl_prefix) {
2713  errcpy(err, "duplicate control escape");
2714  return -1;
2715  }
2716  ctrl_prefix = 1;
2717  if (p < end && (*p & 0x80) == 0) {
2718  if (*p == '\\') {
2719  p++;
2720  goto again;
2721  }
2722  else {
2723  code = *p++;
2724  break;
2725  }
2726  }
2727  errcpy(err, "too short control escape");
2728  return -1;
2729 
2730  default:
2731  errcpy(err, "unexpected escape sequence");
2732  return -1;
2733  }
2734  if (code < 0 || 0xff < code) {
2735  errcpy(err, "invalid escape code");
2736  return -1;
2737  }
2738 
2739  if (ctrl_prefix)
2740  code &= 0x1f;
2741  if (meta_prefix)
2742  code |= 0x80;
2743 
2744  *pp = p;
2745  return code;
2746 }
2747 
2748 static int
2749 unescape_escaped_nonascii(const char **pp, const char *end, rb_encoding *enc,
2750  VALUE buf, rb_encoding **encp, onig_errmsg_buffer err)
2751 {
2752  const char *p = *pp;
2753  int chmaxlen = rb_enc_mbmaxlen(enc);
2754  unsigned char *area = ALLOCA_N(unsigned char, chmaxlen);
2755  char *chbuf = (char *)area;
2756  int chlen = 0;
2757  int byte;
2758  int l;
2759 
2760  memset(chbuf, 0, chmaxlen);
2761 
2762  byte = read_escaped_byte(&p, end, err);
2763  if (byte == -1) {
2764  return -1;
2765  }
2766 
2767  area[chlen++] = byte;
2768  while (chlen < chmaxlen &&
2769  MBCLEN_NEEDMORE_P(rb_enc_precise_mbclen(chbuf, chbuf+chlen, enc))) {
2770  byte = read_escaped_byte(&p, end, err);
2771  if (byte == -1) {
2772  return -1;
2773  }
2774  area[chlen++] = byte;
2775  }
2776 
2777  l = rb_enc_precise_mbclen(chbuf, chbuf+chlen, enc);
2778  if (MBCLEN_INVALID_P(l)) {
2779  errcpy(err, "invalid multibyte escape");
2780  return -1;
2781  }
2782  if (1 < chlen || (area[0] & 0x80)) {
2783  rb_str_buf_cat(buf, chbuf, chlen);
2784 
2785  if (*encp == 0)
2786  *encp = enc;
2787  else if (*encp != enc) {
2788  errcpy(err, "escaped non ASCII character in UTF-8 regexp");
2789  return -1;
2790  }
2791  }
2792  else {
2793  char escbuf[5];
2794  snprintf(escbuf, sizeof(escbuf), "\\x%02X", area[0]&0xff);
2795  rb_str_buf_cat(buf, escbuf, 4);
2796  }
2797  *pp = p;
2798  return 0;
2799 }
2800 
2801 static int
2802 check_unicode_range(unsigned long code, onig_errmsg_buffer err)
2803 {
2804  if ((0xd800 <= code && code <= 0xdfff) || /* Surrogates */
2805  0x10ffff < code) {
2806  errcpy(err, "invalid Unicode range");
2807  return -1;
2808  }
2809  return 0;
2810 }
2811 
2812 static int
2813 append_utf8(unsigned long uv,
2814  VALUE buf, rb_encoding **encp, onig_errmsg_buffer err)
2815 {
2816  if (check_unicode_range(uv, err) != 0)
2817  return -1;
2818  if (uv < 0x80) {
2819  char escbuf[5];
2820  snprintf(escbuf, sizeof(escbuf), "\\x%02X", (int)uv);
2821  rb_str_buf_cat(buf, escbuf, 4);
2822  }
2823  else {
2824  int len;
2825  char utf8buf[6];
2826  len = rb_uv_to_utf8(utf8buf, uv);
2827  rb_str_buf_cat(buf, utf8buf, len);
2828 
2829  if (*encp == 0)
2830  *encp = rb_utf8_encoding();
2831  else if (*encp != rb_utf8_encoding()) {
2832  errcpy(err, "UTF-8 character in non UTF-8 regexp");
2833  return -1;
2834  }
2835  }
2836  return 0;
2837 }
2838 
2839 static int
2840 unescape_unicode_list(const char **pp, const char *end,
2841  VALUE buf, rb_encoding **encp, onig_errmsg_buffer err)
2842 {
2843  const char *p = *pp;
2844  int has_unicode = 0;
2845  unsigned long code;
2846  size_t len;
2847 
2848  while (p < end && ISSPACE(*p)) p++;
2849 
2850  while (1) {
2851  code = ruby_scan_hex(p, end-p, &len);
2852  if (len == 0)
2853  break;
2854  if (6 < len) { /* max 10FFFF */
2855  errcpy(err, "invalid Unicode range");
2856  return -1;
2857  }
2858  p += len;
2859  if (append_utf8(code, buf, encp, err) != 0)
2860  return -1;
2861  has_unicode = 1;
2862 
2863  while (p < end && ISSPACE(*p)) p++;
2864  }
2865 
2866  if (has_unicode == 0) {
2867  errcpy(err, "invalid Unicode list");
2868  return -1;
2869  }
2870 
2871  *pp = p;
2872 
2873  return 0;
2874 }
2875 
2876 static int
2877 unescape_unicode_bmp(const char **pp, const char *end,
2878  VALUE buf, rb_encoding **encp, onig_errmsg_buffer err)
2879 {
2880  const char *p = *pp;
2881  size_t len;
2882  unsigned long code;
2883 
2884  if (end < p+4) {
2885  errcpy(err, "invalid Unicode escape");
2886  return -1;
2887  }
2888  code = ruby_scan_hex(p, 4, &len);
2889  if (len != 4) {
2890  errcpy(err, "invalid Unicode escape");
2891  return -1;
2892  }
2893  if (append_utf8(code, buf, encp, err) != 0)
2894  return -1;
2895  *pp = p + 4;
2896  return 0;
2897 }
2898 
2899 static int
2900 unescape_nonascii0(const char **pp, const char *end, rb_encoding *enc,
2901  VALUE buf, rb_encoding **encp, int *has_property,
2902  onig_errmsg_buffer err, int options, int recurse)
2903 {
2904  const char *p = *pp;
2905  unsigned char c;
2906  char smallbuf[2];
2907  int in_char_class = 0;
2908  int parens = 1; /* ignored unless recurse is true */
2909  int extended_mode = options & ONIG_OPTION_EXTEND;
2910 
2911 begin_scan:
2912  while (p < end) {
2913  int chlen = rb_enc_precise_mbclen(p, end, enc);
2914  if (!MBCLEN_CHARFOUND_P(chlen)) {
2915  invalid_multibyte:
2916  errcpy(err, "invalid multibyte character");
2917  return -1;
2918  }
2919  chlen = MBCLEN_CHARFOUND_LEN(chlen);
2920  if (1 < chlen || (*p & 0x80)) {
2921  multibyte:
2922  rb_str_buf_cat(buf, p, chlen);
2923  p += chlen;
2924  if (*encp == 0)
2925  *encp = enc;
2926  else if (*encp != enc) {
2927  errcpy(err, "non ASCII character in UTF-8 regexp");
2928  return -1;
2929  }
2930  continue;
2931  }
2932 
2933  switch (c = *p++) {
2934  case '\\':
2935  if (p == end) {
2936  errcpy(err, "too short escape sequence");
2937  return -1;
2938  }
2939  chlen = rb_enc_precise_mbclen(p, end, enc);
2940  if (!MBCLEN_CHARFOUND_P(chlen)) {
2941  goto invalid_multibyte;
2942  }
2943  if ((chlen = MBCLEN_CHARFOUND_LEN(chlen)) > 1) {
2944  /* include the previous backslash */
2945  --p;
2946  ++chlen;
2947  goto multibyte;
2948  }
2949  switch (c = *p++) {
2950  case '1': case '2': case '3':
2951  case '4': case '5': case '6': case '7': /* \O, \OO, \OOO or backref */
2952  {
2953  size_t len = end-(p-1), octlen;
2954  if (ruby_scan_oct(p-1, len < 3 ? len : 3, &octlen) <= 0177) {
2955  /* backref or 7bit octal.
2956  no need to unescape anyway.
2957  re-escaping may break backref */
2958  goto escape_asis;
2959  }
2960  }
2961  /* xxx: How about more than 199 subexpressions? */
2962 
2963  case '0': /* \0, \0O, \0OO */
2964 
2965  case 'x': /* \xHH */
2966  case 'c': /* \cX, \c\M-X */
2967  case 'C': /* \C-X, \C-\M-X */
2968  case 'M': /* \M-X, \M-\C-X, \M-\cX */
2969  p = p-2;
2970  if (rb_is_usascii_enc(enc)) {
2971  const char *pbeg = p;
2972  int byte = read_escaped_byte(&p, end, err);
2973  if (byte == -1) return -1;
2974  c = byte;
2975  rb_str_buf_cat(buf, pbeg, p-pbeg);
2976  }
2977  else {
2978  if (unescape_escaped_nonascii(&p, end, enc, buf, encp, err) != 0)
2979  return -1;
2980  }
2981  break;
2982 
2983  case 'u':
2984  if (p == end) {
2985  errcpy(err, "too short escape sequence");
2986  return -1;
2987  }
2988  if (*p == '{') {
2989  /* \u{H HH HHH HHHH HHHHH HHHHHH ...} */
2990  p++;
2991  if (unescape_unicode_list(&p, end, buf, encp, err) != 0)
2992  return -1;
2993  if (p == end || *p++ != '}') {
2994  errcpy(err, "invalid Unicode list");
2995  return -1;
2996  }
2997  break;
2998  }
2999  else {
3000  /* \uHHHH */
3001  if (unescape_unicode_bmp(&p, end, buf, encp, err) != 0)
3002  return -1;
3003  break;
3004  }
3005 
3006  case 'p': /* \p{Hiragana} */
3007  case 'P':
3008  if (!*encp) {
3009  *has_property = 1;
3010  }
3011  goto escape_asis;
3012 
3013  default: /* \n, \\, \d, \9, etc. */
3014 escape_asis:
3015  smallbuf[0] = '\\';
3016  smallbuf[1] = c;
3017  rb_str_buf_cat(buf, smallbuf, 2);
3018  break;
3019  }
3020  break;
3021 
3022  case '#':
3023  if (extended_mode && !in_char_class) {
3024  /* consume and ignore comment in extended regexp */
3025  while ((p < end) && ((c = *p++) != '\n')) {
3026  if ((c & 0x80) && !*encp && enc == rb_utf8_encoding()) {
3027  *encp = enc;
3028  }
3029  }
3030  break;
3031  }
3032  rb_str_buf_cat(buf, (char *)&c, 1);
3033  break;
3034  case '[':
3035  in_char_class++;
3036  rb_str_buf_cat(buf, (char *)&c, 1);
3037  break;
3038  case ']':
3039  if (in_char_class) {
3040  in_char_class--;
3041  }
3042  rb_str_buf_cat(buf, (char *)&c, 1);
3043  break;
3044  case ')':
3045  rb_str_buf_cat(buf, (char *)&c, 1);
3046  if (!in_char_class && recurse) {
3047  if (--parens == 0) {
3048  *pp = p;
3049  return 0;
3050  }
3051  }
3052  break;
3053  case '(':
3054  if (!in_char_class && p + 1 < end && *p == '?') {
3055  if (*(p+1) == '#') {
3056  /* (?# is comment inside any regexp, and content inside should be ignored */
3057  const char *orig_p = p;
3058  int cont = 1;
3059 
3060  while (cont && (p < end)) {
3061  switch (c = *p++) {
3062  default:
3063  if (!(c & 0x80)) break;
3064  if (!*encp && enc == rb_utf8_encoding()) {
3065  *encp = enc;
3066  }
3067  --p;
3068  /* fallthrough */
3069  case '\\':
3070  chlen = rb_enc_precise_mbclen(p, end, enc);
3071  if (!MBCLEN_CHARFOUND_P(chlen)) {
3072  goto invalid_multibyte;
3073  }
3074  p += MBCLEN_CHARFOUND_LEN(chlen);
3075  break;
3076  case ')':
3077  cont = 0;
3078  break;
3079  }
3080  }
3081 
3082  if (cont) {
3083  /* unterminated (?#, rewind so it is syntax error */
3084  p = orig_p;
3085  c = '(';
3086  rb_str_buf_cat(buf, (char *)&c, 1);
3087  }
3088  break;
3089  }
3090  else {
3091  /* potential change of extended option */
3092  int invert = 0;
3093  int local_extend = 0;
3094  const char *s;
3095 
3096  if (recurse) {
3097  parens++;
3098  }
3099 
3100  for (s = p+1; s < end; s++) {
3101  switch(*s) {
3102  case 'x':
3103  local_extend = invert ? -1 : 1;
3104  break;
3105  case '-':
3106  invert = 1;
3107  break;
3108  case ':':
3109  case ')':
3110  if (local_extend == 0 ||
3111  (local_extend == -1 && !extended_mode) ||
3112  (local_extend == 1 && extended_mode)) {
3113  /* no changes to extended flag */
3114  goto fallthrough;
3115  }
3116 
3117  if (*s == ':') {
3118  /* change extended flag until ')' */
3119  int local_options = options;
3120  if (local_extend == 1) {
3121  local_options |= ONIG_OPTION_EXTEND;
3122  }
3123  else {
3124  local_options &= ~ONIG_OPTION_EXTEND;
3125  }
3126 
3127  rb_str_buf_cat(buf, (char *)&c, 1);
3128  int ret = unescape_nonascii0(&p, end, enc, buf, encp,
3129  has_property, err,
3130  local_options, 1);
3131  if (ret < 0) return ret;
3132  goto begin_scan;
3133  }
3134  else {
3135  /* change extended flag for rest of expression */
3136  extended_mode = local_extend == 1;
3137  goto fallthrough;
3138  }
3139  case 'i':
3140  case 'm':
3141  case 'a':
3142  case 'd':
3143  case 'u':
3144  /* other option flags, ignored during scanning */
3145  break;
3146  default:
3147  /* other character, no extended flag change*/
3148  goto fallthrough;
3149  }
3150  }
3151  }
3152  }
3153  else if (!in_char_class && recurse) {
3154  parens++;
3155  }
3156  /* FALLTHROUGH */
3157  default:
3158 fallthrough:
3159  rb_str_buf_cat(buf, (char *)&c, 1);
3160  break;
3161  }
3162  }
3163 
3164  if (recurse) {
3165  *pp = p;
3166  }
3167  return 0;
3168 }
3169 
3170 static int
3171 unescape_nonascii(const char *p, const char *end, rb_encoding *enc,
3172  VALUE buf, rb_encoding **encp, int *has_property,
3173  onig_errmsg_buffer err, int options)
3174 {
3175  return unescape_nonascii0(&p, end, enc, buf, encp, has_property,
3176  err, options, 0);
3177 }
3178 
3179 static VALUE
3180 rb_reg_preprocess(const char *p, const char *end, rb_encoding *enc,
3181  rb_encoding **fixed_enc, onig_errmsg_buffer err, int options)
3182 {
3183  VALUE buf;
3184  int has_property = 0;
3185 
3186  buf = rb_str_buf_new(0);
3187 
3188  if (rb_enc_asciicompat(enc))
3189  *fixed_enc = 0;
3190  else {
3191  *fixed_enc = enc;
3192  rb_enc_associate(buf, enc);
3193  }
3194 
3195  if (unescape_nonascii(p, end, enc, buf, fixed_enc, &has_property, err, options) != 0)
3196  return Qnil;
3197 
3198  if (has_property && !*fixed_enc) {
3199  *fixed_enc = enc;
3200  }
3201 
3202  if (*fixed_enc) {
3203  rb_enc_associate(buf, *fixed_enc);
3204  }
3205 
3206  return buf;
3207 }
3208 
3209 VALUE
3210 rb_reg_check_preprocess(VALUE str)
3211 {
3212  rb_encoding *fixed_enc = 0;
3213  onig_errmsg_buffer err = "";
3214  VALUE buf;
3215  char *p, *end;
3216  rb_encoding *enc;
3217 
3218  StringValue(str);
3219  p = RSTRING_PTR(str);
3220  end = p + RSTRING_LEN(str);
3221  enc = rb_enc_get(str);
3222 
3223  buf = rb_reg_preprocess(p, end, enc, &fixed_enc, err, 0);
3224  RB_GC_GUARD(str);
3225 
3226  if (NIL_P(buf)) {
3227  return rb_reg_error_desc(str, 0, err);
3228  }
3229  return Qnil;
3230 }
3231 
3232 static VALUE
3233 rb_reg_preprocess_dregexp(VALUE ary, int options)
3234 {
3235  rb_encoding *fixed_enc = 0;
3236  rb_encoding *regexp_enc = 0;
3237  onig_errmsg_buffer err = "";
3238  int i;
3239  VALUE result = 0;
3240  rb_encoding *ascii8bit = rb_ascii8bit_encoding();
3241 
3242  if (RARRAY_LEN(ary) == 0) {
3243  rb_raise(rb_eArgError, "no arguments given");
3244  }
3245 
3246  for (i = 0; i < RARRAY_LEN(ary); i++) {
3247  VALUE str = RARRAY_AREF(ary, i);
3248  VALUE buf;
3249  char *p, *end;
3250  rb_encoding *src_enc;
3251 
3252  src_enc = rb_enc_get(str);
3253  if (options & ARG_ENCODING_NONE &&
3254  src_enc != ascii8bit) {
3255  if (str_coderange(str) != ENC_CODERANGE_7BIT)
3256  rb_raise(rb_eRegexpError, "/.../n has a non escaped non ASCII character in non ASCII-8BIT script");
3257  else
3258  src_enc = ascii8bit;
3259  }
3260 
3261  StringValue(str);
3262  p = RSTRING_PTR(str);
3263  end = p + RSTRING_LEN(str);
3264 
3265  buf = rb_reg_preprocess(p, end, src_enc, &fixed_enc, err, options);
3266 
3267  if (NIL_P(buf))
3268  rb_raise(rb_eArgError, "%s", err);
3269 
3270  if (fixed_enc != 0) {
3271  if (regexp_enc != 0 && regexp_enc != fixed_enc) {
3272  rb_raise(rb_eRegexpError, "encoding mismatch in dynamic regexp : %s and %s",
3273  rb_enc_name(regexp_enc), rb_enc_name(fixed_enc));
3274  }
3275  regexp_enc = fixed_enc;
3276  }
3277 
3278  if (!result)
3279  result = rb_str_new3(str);
3280  else
3281  rb_str_buf_append(result, str);
3282  }
3283  if (regexp_enc) {
3284  rb_enc_associate(result, regexp_enc);
3285  }
3286 
3287  return result;
3288 }
3289 
3290 static void
3291 rb_reg_initialize_check(VALUE obj)
3292 {
3293  rb_check_frozen(obj);
3294  if (RREGEXP_PTR(obj)) {
3295  rb_raise(rb_eTypeError, "already initialized regexp");
3296  }
3297 }
3298 
3299 static int
3300 rb_reg_initialize(VALUE obj, const char *s, long len, rb_encoding *enc,
3301  int options, onig_errmsg_buffer err,
3302  const char *sourcefile, int sourceline)
3303 {
3304  struct RRegexp *re = RREGEXP(obj);
3305  VALUE unescaped;
3306  rb_encoding *fixed_enc = 0;
3308 
3309  rb_reg_initialize_check(obj);
3310 
3311  if (rb_enc_dummy_p(enc)) {
3312  errcpy(err, "can't make regexp with dummy encoding");
3313  return -1;
3314  }
3315 
3316  unescaped = rb_reg_preprocess(s, s+len, enc, &fixed_enc, err, options);
3317  if (NIL_P(unescaped))
3318  return -1;
3319 
3320  if (fixed_enc) {
3321  if ((fixed_enc != enc && (options & ARG_ENCODING_FIXED)) ||
3322  (fixed_enc != a_enc && (options & ARG_ENCODING_NONE))) {
3323  errcpy(err, "incompatible character encoding");
3324  return -1;
3325  }
3326  if (fixed_enc != a_enc) {
3327  options |= ARG_ENCODING_FIXED;
3328  enc = fixed_enc;
3329  }
3330  }
3331  else if (!(options & ARG_ENCODING_FIXED)) {
3332  enc = rb_usascii_encoding();
3333  }
3334 
3335  rb_enc_associate((VALUE)re, enc);
3336  if ((options & ARG_ENCODING_FIXED) || fixed_enc) {
3337  re->basic.flags |= KCODE_FIXED;
3338  }
3339  if (options & ARG_ENCODING_NONE) {
3340  re->basic.flags |= REG_ENCODING_NONE;
3341  }
3342 
3343  re->ptr = make_regexp(RSTRING_PTR(unescaped), RSTRING_LEN(unescaped), enc,
3344  options & ARG_REG_OPTION_MASK, err,
3345  sourcefile, sourceline);
3346  if (!re->ptr) return -1;
3347  RB_GC_GUARD(unescaped);
3348  return 0;
3349 }
3350 
3351 static void
3352 reg_set_source(VALUE reg, VALUE str, rb_encoding *enc)
3353 {
3354  rb_encoding *regenc = rb_enc_get(reg);
3355  if (regenc != enc) {
3356  str = rb_enc_associate(rb_str_dup(str), enc = regenc);
3357  }
3358  RB_OBJ_WRITE(reg, &RREGEXP(reg)->src, rb_fstring(str));
3359 }
3360 
3361 static int
3362 rb_reg_initialize_str(VALUE obj, VALUE str, int options, onig_errmsg_buffer err,
3363  const char *sourcefile, int sourceline)
3364 {
3365  int ret;
3366  rb_encoding *str_enc = rb_enc_get(str), *enc = str_enc;
3367  if (options & ARG_ENCODING_NONE) {
3368  rb_encoding *ascii8bit = rb_ascii8bit_encoding();
3369  if (enc != ascii8bit) {
3370  if (str_coderange(str) != ENC_CODERANGE_7BIT) {
3371  errcpy(err, "/.../n has a non escaped non ASCII character in non ASCII-8BIT script");
3372  return -1;
3373  }
3374  enc = ascii8bit;
3375  }
3376  }
3377  ret = rb_reg_initialize(obj, RSTRING_PTR(str), RSTRING_LEN(str), enc,
3378  options, err, sourcefile, sourceline);
3379  if (ret == 0) reg_set_source(obj, str, str_enc);
3380  return ret;
3381 }
3382 
3383 static VALUE
3384 rb_reg_s_alloc(VALUE klass)
3385 {
3386  NEWOBJ_OF(re, struct RRegexp, klass, T_REGEXP | (RGENGC_WB_PROTECTED_REGEXP ? FL_WB_PROTECTED : 0), sizeof(struct RRegexp), 0);
3387 
3388  re->ptr = 0;
3389  RB_OBJ_WRITE(re, &re->src, 0);
3390  re->usecnt = 0;
3391 
3392  return (VALUE)re;
3393 }
3394 
3395 VALUE
3396 rb_reg_alloc(void)
3397 {
3398  return rb_reg_s_alloc(rb_cRegexp);
3399 }
3400 
3401 VALUE
3402 rb_reg_new_str(VALUE s, int options)
3403 {
3404  return rb_reg_init_str(rb_reg_alloc(), s, options);
3405 }
3406 
3407 VALUE
3408 rb_reg_init_str(VALUE re, VALUE s, int options)
3409 {
3410  onig_errmsg_buffer err = "";
3411 
3412  if (rb_reg_initialize_str(re, s, options, err, NULL, 0) != 0) {
3413  rb_reg_raise_str(s, options, err);
3414  }
3415 
3416  return re;
3417 }
3418 
3419 static VALUE
3420 rb_reg_init_str_enc(VALUE re, VALUE s, rb_encoding *enc, int options)
3421 {
3422  onig_errmsg_buffer err = "";
3423 
3424  if (rb_reg_initialize(re, RSTRING_PTR(s), RSTRING_LEN(s),
3425  enc, options, err, NULL, 0) != 0) {
3426  rb_reg_raise_str(s, options, err);
3427  }
3428  reg_set_source(re, s, enc);
3429 
3430  return re;
3431 }
3432 
3433 VALUE
3434 rb_reg_new_ary(VALUE ary, int opt)
3435 {
3436  VALUE re = rb_reg_new_str(rb_reg_preprocess_dregexp(ary, opt), opt);
3437  rb_obj_freeze(re);
3438  return re;
3439 }
3440 
3441 VALUE
3442 rb_enc_reg_new(const char *s, long len, rb_encoding *enc, int options)
3443 {
3444  VALUE re = rb_reg_alloc();
3445  onig_errmsg_buffer err = "";
3446 
3447  if (rb_reg_initialize(re, s, len, enc, options, err, NULL, 0) != 0) {
3448  rb_enc_reg_raise(s, len, enc, options, err);
3449  }
3450  RB_OBJ_WRITE(re, &RREGEXP(re)->src, rb_fstring(rb_enc_str_new(s, len, enc)));
3451 
3452  return re;
3453 }
3454 
3455 VALUE
3456 rb_reg_new(const char *s, long len, int options)
3457 {
3458  return rb_enc_reg_new(s, len, rb_ascii8bit_encoding(), options);
3459 }
3460 
3461 VALUE
3462 rb_reg_compile(VALUE str, int options, const char *sourcefile, int sourceline)
3463 {
3464  VALUE re = rb_reg_alloc();
3465  onig_errmsg_buffer err = "";
3466 
3467  if (!str) str = rb_str_new(0,0);
3468  if (rb_reg_initialize_str(re, str, options, err, sourcefile, sourceline) != 0) {
3469  rb_set_errinfo(rb_reg_error_desc(str, options, err));
3470  return Qnil;
3471  }
3472  rb_obj_freeze(re);
3473  return re;
3474 }
3475 
3476 static VALUE reg_cache;
3477 
3478 VALUE
3480 {
3481  if (reg_cache && RREGEXP_SRC_LEN(reg_cache) == RSTRING_LEN(str)
3482  && ENCODING_GET(reg_cache) == ENCODING_GET(str)
3483  && memcmp(RREGEXP_SRC_PTR(reg_cache), RSTRING_PTR(str), RSTRING_LEN(str)) == 0)
3484  return reg_cache;
3485 
3486  return reg_cache = rb_reg_new_str(str, 0);
3487 }
3488 
3489 static st_index_t reg_hash(VALUE re);
3490 /*
3491  * call-seq:
3492  * hash -> integer
3493  *
3494  * Returns the integer hash value for +self+.
3495  *
3496  * Related: Object#hash.
3497  *
3498  */
3499 
3500 VALUE
3501 rb_reg_hash(VALUE re)
3502 {
3503  st_index_t hashval = reg_hash(re);
3504  return ST2FIX(hashval);
3505 }
3506 
3507 static st_index_t
3508 reg_hash(VALUE re)
3509 {
3510  st_index_t hashval;
3511 
3512  rb_reg_check(re);
3513  hashval = RREGEXP_PTR(re)->options;
3514  hashval = rb_hash_uint(hashval, rb_memhash(RREGEXP_SRC_PTR(re), RREGEXP_SRC_LEN(re)));
3515  return rb_hash_end(hashval);
3516 }
3517 
3518 
3519 /*
3520  * call-seq:
3521  * regexp == object -> true or false
3522  *
3523  * Returns +true+ if +object+ is another \Regexp whose pattern,
3524  * flags, and encoding are the same as +self+, +false+ otherwise:
3525  *
3526  * /foo/ == Regexp.new('foo') # => true
3527  * /foo/ == /foo/i # => false
3528  * /foo/ == Regexp.new('food') # => false
3529  * /foo/ == Regexp.new("abc".force_encoding("euc-jp")) # => false
3530  *
3531  */
3532 
3533 VALUE
3534 rb_reg_equal(VALUE re1, VALUE re2)
3535 {
3536  if (re1 == re2) return Qtrue;
3537  if (!RB_TYPE_P(re2, T_REGEXP)) return Qfalse;
3538  rb_reg_check(re1); rb_reg_check(re2);
3539  if (FL_TEST(re1, KCODE_FIXED) != FL_TEST(re2, KCODE_FIXED)) return Qfalse;
3540  if (RREGEXP_PTR(re1)->options != RREGEXP_PTR(re2)->options) return Qfalse;
3541  if (RREGEXP_SRC_LEN(re1) != RREGEXP_SRC_LEN(re2)) return Qfalse;
3542  if (ENCODING_GET(re1) != ENCODING_GET(re2)) return Qfalse;
3543  return RBOOL(memcmp(RREGEXP_SRC_PTR(re1), RREGEXP_SRC_PTR(re2), RREGEXP_SRC_LEN(re1)) == 0);
3544 }
3545 
3546 /*
3547  * call-seq:
3548  * hash -> integer
3549  *
3550  * Returns the integer hash value for +self+,
3551  * based on the target string, regexp, match, and captures.
3552  *
3553  * See also Object#hash.
3554  *
3555  */
3556 
3557 static VALUE
3558 match_hash(VALUE match)
3559 {
3560  const struct re_registers *regs;
3561  st_index_t hashval;
3562 
3563  match_check(match);
3564  hashval = rb_hash_start(rb_str_hash(RMATCH(match)->str));
3565  hashval = rb_hash_uint(hashval, reg_hash(match_regexp(match)));
3566  regs = RMATCH_REGS(match);
3567  hashval = rb_hash_uint(hashval, regs->num_regs);
3568  hashval = rb_hash_uint(hashval, rb_memhash(regs->beg, regs->num_regs * sizeof(*regs->beg)));
3569  hashval = rb_hash_uint(hashval, rb_memhash(regs->end, regs->num_regs * sizeof(*regs->end)));
3570  hashval = rb_hash_end(hashval);
3571  return ST2FIX(hashval);
3572 }
3573 
3574 /*
3575  * call-seq:
3576  * matchdata == object -> true or false
3577  *
3578  * Returns +true+ if +object+ is another \MatchData object
3579  * whose target string, regexp, match, and captures
3580  * are the same as +self+, +false+ otherwise.
3581  */
3582 
3583 static VALUE
3584 match_equal(VALUE match1, VALUE match2)
3585 {
3586  const struct re_registers *regs1, *regs2;
3587 
3588  if (match1 == match2) return Qtrue;
3589  if (!RB_TYPE_P(match2, T_MATCH)) return Qfalse;
3590  if (!RMATCH(match1)->regexp || !RMATCH(match2)->regexp) return Qfalse;
3591  if (!rb_str_equal(RMATCH(match1)->str, RMATCH(match2)->str)) return Qfalse;
3592  if (!rb_reg_equal(match_regexp(match1), match_regexp(match2))) return Qfalse;
3593  regs1 = RMATCH_REGS(match1);
3594  regs2 = RMATCH_REGS(match2);
3595  if (regs1->num_regs != regs2->num_regs) return Qfalse;
3596  if (memcmp(regs1->beg, regs2->beg, regs1->num_regs * sizeof(*regs1->beg))) return Qfalse;
3597  if (memcmp(regs1->end, regs2->end, regs1->num_regs * sizeof(*regs1->end))) return Qfalse;
3598  return Qtrue;
3599 }
3600 
3601 static VALUE
3602 reg_operand(VALUE s, int check)
3603 {
3604  if (SYMBOL_P(s)) {
3605  return rb_sym2str(s);
3606  }
3607  else if (RB_TYPE_P(s, T_STRING)) {
3608  return s;
3609  }
3610  else {
3611  return check ? rb_str_to_str(s) : rb_check_string_type(s);
3612  }
3613 }
3614 
3615 static long
3616 reg_match_pos(VALUE re, VALUE *strp, long pos, VALUE* set_match)
3617 {
3618  VALUE str = *strp;
3619 
3620  if (NIL_P(str)) {
3622  return -1;
3623  }
3624  *strp = str = reg_operand(str, TRUE);
3625  if (pos != 0) {
3626  if (pos < 0) {
3627  VALUE l = rb_str_length(str);
3628  pos += NUM2INT(l);
3629  if (pos < 0) {
3630  return pos;
3631  }
3632  }
3633  pos = rb_str_offset(str, pos);
3634  }
3635  return rb_reg_search_set_match(re, str, pos, 0, 1, set_match);
3636 }
3637 
3638 /*
3639  * call-seq:
3640  * regexp =~ string -> integer or nil
3641  *
3642  * Returns the integer index (in characters) of the first match
3643  * for +self+ and +string+, or +nil+ if none;
3644  * also sets the
3645  * {rdoc-ref:Regexp global variables}[rdoc-ref:Regexp@Global+Variables]:
3646  *
3647  * /at/ =~ 'input data' # => 7
3648  * $~ # => #<MatchData "at">
3649  * /ax/ =~ 'input data' # => nil
3650  * $~ # => nil
3651  *
3652  * Assigns named captures to local variables of the same names
3653  * if and only if +self+:
3654  *
3655  * - Is a regexp literal;
3656  * see {Regexp Literals}[rdoc-ref:literals.rdoc@Regexp+Literals].
3657  * - Does not contain interpolations;
3658  * see {Regexp interpolation}[rdoc-ref:Regexp@Interpolation+Mode].
3659  * - Is at the left of the expression.
3660  *
3661  * Example:
3662  *
3663  * /(?<lhs>\w+)\s*=\s*(?<rhs>\w+)/ =~ ' x = y '
3664  * p lhs # => "x"
3665  * p rhs # => "y"
3666  *
3667  * Assigns +nil+ if not matched:
3668  *
3669  * /(?<lhs>\w+)\s*=\s*(?<rhs>\w+)/ =~ ' x = '
3670  * p lhs # => nil
3671  * p rhs # => nil
3672  *
3673  * Does not make local variable assignments if +self+ is not a regexp literal:
3674  *
3675  * r = /(?<foo>\w+)\s*=\s*(?<foo>\w+)/
3676  * r =~ ' x = y '
3677  * p foo # Undefined local variable
3678  * p bar # Undefined local variable
3679  *
3680  * The assignment does not occur if the regexp is not at the left:
3681  *
3682  * ' x = y ' =~ /(?<foo>\w+)\s*=\s*(?<foo>\w+)/
3683  * p foo, foo # Undefined local variables
3684  *
3685  * A regexp interpolation, <tt>#{}</tt>, also disables
3686  * the assignment:
3687  *
3688  * r = /(?<foo>\w+)/
3689  * /(?<foo>\w+)\s*=\s*#{r}/ =~ 'x = y'
3690  * p foo # Undefined local variable
3691  *
3692  */
3693 
3694 VALUE
3696 {
3697  long pos = reg_match_pos(re, &str, 0, NULL);
3698  if (pos < 0) return Qnil;
3699  pos = rb_str_sublen(str, pos);
3700  return LONG2FIX(pos);
3701 }
3702 
3703 /*
3704  * call-seq:
3705  * regexp === string -> true or false
3706  *
3707  * Returns +true+ if +self+ finds a match in +string+:
3708  *
3709  * /^[a-z]*$/ === 'HELLO' # => false
3710  * /^[A-Z]*$/ === 'HELLO' # => true
3711  *
3712  * This method is called in case statements:
3713  *
3714  * s = 'HELLO'
3715  * case s
3716  * when /\A[a-z]*\z/; print "Lower case\n"
3717  * when /\A[A-Z]*\z/; print "Upper case\n"
3718  * else print "Mixed case\n"
3719  * end # => "Upper case"
3720  *
3721  */
3722 
3723 static VALUE
3724 rb_reg_eqq(VALUE re, VALUE str)
3725 {
3726  long start;
3727 
3728  str = reg_operand(str, FALSE);
3729  if (NIL_P(str)) {
3731  return Qfalse;
3732  }
3733  start = rb_reg_search(re, str, 0, 0);
3734  return RBOOL(start >= 0);
3735 }
3736 
3737 
3738 /*
3739  * call-seq:
3740  * ~ rxp -> integer or nil
3741  *
3742  * Equivalent to <tt><i>rxp</i> =~ $_</tt>:
3743  *
3744  * $_ = "input data"
3745  * ~ /at/ # => 7
3746  *
3747  */
3748 
3749 VALUE
3751 {
3752  long start;
3753  VALUE line = rb_lastline_get();
3754 
3755  if (!RB_TYPE_P(line, T_STRING)) {
3757  return Qnil;
3758  }
3759 
3760  start = rb_reg_search(re, line, 0, 0);
3761  if (start < 0) {
3762  return Qnil;
3763  }
3764  start = rb_str_sublen(line, start);
3765  return LONG2FIX(start);
3766 }
3767 
3768 
3769 /*
3770  * call-seq:
3771  * match(string, offset = 0) -> matchdata or nil
3772  * match(string, offset = 0) {|matchdata| ... } -> object
3773  *
3774  * With no block given, returns the MatchData object
3775  * that describes the match, if any, or +nil+ if none;
3776  * the search begins at the given character +offset+ in +string+:
3777  *
3778  * /abra/.match('abracadabra') # => #<MatchData "abra">
3779  * /abra/.match('abracadabra', 4) # => #<MatchData "abra">
3780  * /abra/.match('abracadabra', 8) # => nil
3781  * /abra/.match('abracadabra', 800) # => nil
3782  *
3783  * string = "\u{5d0 5d1 5e8 5d0}cadabra"
3784  * /abra/.match(string, 7) #=> #<MatchData "abra">
3785  * /abra/.match(string, 8) #=> nil
3786  * /abra/.match(string.b, 8) #=> #<MatchData "abra">
3787  *
3788  * With a block given, calls the block if and only if a match is found;
3789  * returns the block's value:
3790  *
3791  * /abra/.match('abracadabra') {|matchdata| p matchdata }
3792  * # => #<MatchData "abra">
3793  * /abra/.match('abracadabra', 4) {|matchdata| p matchdata }
3794  * # => #<MatchData "abra">
3795  * /abra/.match('abracadabra', 8) {|matchdata| p matchdata }
3796  * # => nil
3797  * /abra/.match('abracadabra', 8) {|marchdata| fail 'Cannot happen' }
3798  * # => nil
3799  *
3800  * Output (from the first two blocks above):
3801  *
3802  * #<MatchData "abra">
3803  * #<MatchData "abra">
3804  *
3805  * /(.)(.)(.)/.match("abc")[2] # => "b"
3806  * /(.)(.)/.match("abc", 1)[2] # => "c"
3807  *
3808  */
3809 
3810 static VALUE
3811 rb_reg_match_m(int argc, VALUE *argv, VALUE re)
3812 {
3813  VALUE result = Qnil, str, initpos;
3814  long pos;
3815 
3816  if (rb_scan_args(argc, argv, "11", &str, &initpos) == 2) {
3817  pos = NUM2LONG(initpos);
3818  }
3819  else {
3820  pos = 0;
3821  }
3822 
3823  pos = reg_match_pos(re, &str, pos, &result);
3824  if (pos < 0) {
3826  return Qnil;
3827  }
3828  rb_match_busy(result);
3829  if (!NIL_P(result) && rb_block_given_p()) {
3830  return rb_yield(result);
3831  }
3832  return result;
3833 }
3834 
3835 /*
3836  * call-seq:
3837  * match?(string) -> true or false
3838  * match?(string, offset = 0) -> true or false
3839  *
3840  * Returns <code>true</code> or <code>false</code> to indicate whether the
3841  * regexp is matched or not without updating $~ and other related variables.
3842  * If the second parameter is present, it specifies the position in the string
3843  * to begin the search.
3844  *
3845  * /R.../.match?("Ruby") # => true
3846  * /R.../.match?("Ruby", 1) # => false
3847  * /P.../.match?("Ruby") # => false
3848  * $& # => nil
3849  */
3850 
3851 static VALUE
3852 rb_reg_match_m_p(int argc, VALUE *argv, VALUE re)
3853 {
3854  long pos = rb_check_arity(argc, 1, 2) > 1 ? NUM2LONG(argv[1]) : 0;
3855  return rb_reg_match_p(re, argv[0], pos);
3856 }
3857 
3858 VALUE
3859 rb_reg_match_p(VALUE re, VALUE str, long pos)
3860 {
3861  if (NIL_P(str)) return Qfalse;
3862  str = SYMBOL_P(str) ? rb_sym2str(str) : StringValue(str);
3863  if (pos) {
3864  if (pos < 0) {
3865  pos += NUM2LONG(rb_str_length(str));
3866  if (pos < 0) return Qfalse;
3867  }
3868  if (pos > 0) {
3869  long len = 1;
3870  const char *beg = rb_str_subpos(str, pos, &len);
3871  if (!beg) return Qfalse;
3872  pos = beg - RSTRING_PTR(str);
3873  }
3874  }
3875 
3876  struct reg_onig_search_args args = {
3877  .pos = pos,
3878  .range = RSTRING_LEN(str),
3879  };
3880 
3881  return rb_reg_onig_match(re, str, reg_onig_search, &args, NULL) == ONIG_MISMATCH ? Qfalse : Qtrue;
3882 }
3883 
3884 /*
3885  * Document-method: compile
3886  *
3887  * Alias for Regexp.new
3888  */
3889 
3890 static int
3891 str_to_option(VALUE str)
3892 {
3893  int flag = 0;
3894  const char *ptr;
3895  long len;
3896  str = rb_check_string_type(str);
3897  if (NIL_P(str)) return -1;
3898  RSTRING_GETMEM(str, ptr, len);
3899  for (long i = 0; i < len; ++i) {
3900  int f = char_to_option(ptr[i]);
3901  if (!f) {
3902  rb_raise(rb_eArgError, "unknown regexp option: %"PRIsVALUE, str);
3903  }
3904  flag |= f;
3905  }
3906  return flag;
3907 }
3908 
3909 static void
3910 set_timeout(rb_hrtime_t *hrt, VALUE timeout)
3911 {
3912  double timeout_d = NIL_P(timeout) ? 0.0 : NUM2DBL(timeout);
3913  if (!NIL_P(timeout) && timeout_d <= 0) {
3914  rb_raise(rb_eArgError, "invalid timeout: %"PRIsVALUE, timeout);
3915  }
3916  double2hrtime(hrt, timeout_d);
3917 }
3918 
3919 static VALUE
3920 reg_copy(VALUE copy, VALUE orig)
3921 {
3922  int r;
3923  regex_t *re;
3924 
3925  rb_reg_initialize_check(copy);
3926  if ((r = onig_reg_copy(&re, RREGEXP_PTR(orig))) != 0) {
3927  /* ONIGERR_MEMORY only */
3928  rb_raise(rb_eRegexpError, "%s", onig_error_code_to_format(r));
3929  }
3930  RREGEXP_PTR(copy) = re;
3931  RB_OBJ_WRITE(copy, &RREGEXP(copy)->src, RREGEXP(orig)->src);
3932  RREGEXP_PTR(copy)->timelimit = RREGEXP_PTR(orig)->timelimit;
3933  rb_enc_copy(copy, orig);
3934  FL_SET_RAW(copy, FL_TEST_RAW(orig, KCODE_FIXED|REG_ENCODING_NONE));
3935 
3936  return copy;
3937 }
3938 
3940  VALUE str;
3941  VALUE timeout;
3942  rb_encoding *enc;
3943  int flags;
3944 };
3945 
3946 static VALUE reg_extract_args(int argc, VALUE *argv, struct reg_init_args *args);
3947 static VALUE reg_init_args(VALUE self, VALUE str, rb_encoding *enc, int flags);
3948 void rb_warn_deprecated_to_remove(const char *removal, const char *fmt, const char *suggest, ...);
3949 
3950 /*
3951  * call-seq:
3952  * Regexp.new(string, options = 0, timeout: nil) -> regexp
3953  * Regexp.new(regexp, timeout: nil) -> regexp
3954  *
3955  * With argument +string+ given, returns a new regexp with the given string
3956  * and options:
3957  *
3958  * r = Regexp.new('foo') # => /foo/
3959  * r.source # => "foo"
3960  * r.options # => 0
3961  *
3962  * Optional argument +options+ is one of the following:
3963  *
3964  * - A String of options:
3965  *
3966  * Regexp.new('foo', 'i') # => /foo/i
3967  * Regexp.new('foo', 'im') # => /foo/im
3968  *
3969  * - The bit-wise OR of one or more of the constants
3970  * Regexp::EXTENDED, Regexp::IGNORECASE, Regexp::MULTILINE, and
3971  * Regexp::NOENCODING:
3972  *
3973  * Regexp.new('foo', Regexp::IGNORECASE) # => /foo/i
3974  * Regexp.new('foo', Regexp::EXTENDED) # => /foo/x
3975  * Regexp.new('foo', Regexp::MULTILINE) # => /foo/m
3976  * Regexp.new('foo', Regexp::NOENCODING) # => /foo/n
3977  * flags = Regexp::IGNORECASE | Regexp::EXTENDED | Regexp::MULTILINE
3978  * Regexp.new('foo', flags) # => /foo/mix
3979  *
3980  * - +nil+ or +false+, which is ignored.
3981  * - Any other truthy value, in which case the regexp will be
3982  * case-insensitive.
3983  *
3984  * If optional keyword argument +timeout+ is given,
3985  * its float value overrides the timeout interval for the class,
3986  * Regexp.timeout.
3987  * If +nil+ is passed as +timeout, it uses the timeout interval
3988  * for the class, Regexp.timeout.
3989  *
3990  * With argument +regexp+ given, returns a new regexp. The source,
3991  * options, timeout are the same as +regexp+. +options+ and +n_flag+
3992  * arguments are ineffective. The timeout can be overridden by
3993  * +timeout+ keyword.
3994  *
3995  * options = Regexp::MULTILINE
3996  * r = Regexp.new('foo', options, timeout: 1.1) # => /foo/m
3997  * r2 = Regexp.new(r) # => /foo/m
3998  * r2.timeout # => 1.1
3999  * r3 = Regexp.new(r, timeout: 3.14) # => /foo/m
4000  * r3.timeout # => 3.14
4001  *
4002  */
4003 
4004 static VALUE
4005 rb_reg_initialize_m(int argc, VALUE *argv, VALUE self)
4006 {
4007  struct reg_init_args args;
4008  VALUE re = reg_extract_args(argc, argv, &args);
4009 
4010  if (NIL_P(re)) {
4011  reg_init_args(self, args.str, args.enc, args.flags);
4012  }
4013  else {
4014  reg_copy(self, re);
4015  }
4016 
4017  set_timeout(&RREGEXP_PTR(self)->timelimit, args.timeout);
4018 
4019  return self;
4020 }
4021 
4022 static VALUE
4023 reg_extract_args(int argc, VALUE *argv, struct reg_init_args *args)
4024 {
4025  int flags = 0;
4026  rb_encoding *enc = 0;
4027  VALUE str, src, opts = Qundef, kwargs;
4028  VALUE re = Qnil;
4029 
4030  rb_scan_args(argc, argv, "11:", &src, &opts, &kwargs);
4031 
4032  args->timeout = Qnil;
4033  if (!NIL_P(kwargs)) {
4034  static ID keywords[1];
4035  if (!keywords[0]) {
4036  keywords[0] = rb_intern_const("timeout");
4037  }
4038  rb_get_kwargs(kwargs, keywords, 0, 1, &args->timeout);
4039  }
4040 
4041  if (RB_TYPE_P(src, T_REGEXP)) {
4042  re = src;
4043 
4044  if (!NIL_P(opts)) {
4045  rb_warn("flags ignored");
4046  }
4047  rb_reg_check(re);
4048  flags = rb_reg_options(re);
4049  str = RREGEXP_SRC(re);
4050  }
4051  else {
4052  if (!NIL_P(opts)) {
4053  int f;
4054  if (FIXNUM_P(opts)) flags = FIX2INT(opts);
4055  else if ((f = str_to_option(opts)) >= 0) flags = f;
4056  else if (rb_bool_expected(opts, "ignorecase", FALSE))
4057  flags = ONIG_OPTION_IGNORECASE;
4058  }
4059  str = StringValue(src);
4060  }
4061  args->str = str;
4062  args->enc = enc;
4063  args->flags = flags;
4064  return re;
4065 }
4066 
4067 static VALUE
4068 reg_init_args(VALUE self, VALUE str, rb_encoding *enc, int flags)
4069 {
4070  if (enc && rb_enc_get(str) != enc)
4071  rb_reg_init_str_enc(self, str, enc, flags);
4072  else
4073  rb_reg_init_str(self, str, flags);
4074  return self;
4075 }
4076 
4077 VALUE
4079 {
4080  rb_encoding *enc = rb_enc_get(str);
4081  char *s, *send, *t;
4082  VALUE tmp;
4083  int c, clen;
4084  int ascii_only = rb_enc_str_asciionly_p(str);
4085 
4086  s = RSTRING_PTR(str);
4087  send = s + RSTRING_LEN(str);
4088  while (s < send) {
4089  c = rb_enc_ascget(s, send, &clen, enc);
4090  if (c == -1) {
4091  s += mbclen(s, send, enc);
4092  continue;
4093  }
4094  switch (c) {
4095  case '[': case ']': case '{': case '}':
4096  case '(': case ')': case '|': case '-':
4097  case '*': case '.': case '\\':
4098  case '?': case '+': case '^': case '$':
4099  case ' ': case '#':
4100  case '\t': case '\f': case '\v': case '\n': case '\r':
4101  goto meta_found;
4102  }
4103  s += clen;
4104  }
4105  tmp = rb_str_new3(str);
4106  if (ascii_only) {
4108  }
4109  return tmp;
4110 
4111  meta_found:
4112  tmp = rb_str_new(0, RSTRING_LEN(str)*2);
4113  if (ascii_only) {
4115  }
4116  else {
4117  rb_enc_copy(tmp, str);
4118  }
4119  t = RSTRING_PTR(tmp);
4120  /* copy upto metacharacter */
4121  const char *p = RSTRING_PTR(str);
4122  memcpy(t, p, s - p);
4123  t += s - p;
4124 
4125  while (s < send) {
4126  c = rb_enc_ascget(s, send, &clen, enc);
4127  if (c == -1) {
4128  int n = mbclen(s, send, enc);
4129 
4130  while (n--)
4131  *t++ = *s++;
4132  continue;
4133  }
4134  s += clen;
4135  switch (c) {
4136  case '[': case ']': case '{': case '}':
4137  case '(': case ')': case '|': case '-':
4138  case '*': case '.': case '\\':
4139  case '?': case '+': case '^': case '$':
4140  case '#':
4141  t += rb_enc_mbcput('\\', t, enc);
4142  break;
4143  case ' ':
4144  t += rb_enc_mbcput('\\', t, enc);
4145  t += rb_enc_mbcput(' ', t, enc);
4146  continue;
4147  case '\t':
4148  t += rb_enc_mbcput('\\', t, enc);
4149  t += rb_enc_mbcput('t', t, enc);
4150  continue;
4151  case '\n':
4152  t += rb_enc_mbcput('\\', t, enc);
4153  t += rb_enc_mbcput('n', t, enc);
4154  continue;
4155  case '\r':
4156  t += rb_enc_mbcput('\\', t, enc);
4157  t += rb_enc_mbcput('r', t, enc);
4158  continue;
4159  case '\f':
4160  t += rb_enc_mbcput('\\', t, enc);
4161  t += rb_enc_mbcput('f', t, enc);
4162  continue;
4163  case '\v':
4164  t += rb_enc_mbcput('\\', t, enc);
4165  t += rb_enc_mbcput('v', t, enc);
4166  continue;
4167  }
4168  t += rb_enc_mbcput(c, t, enc);
4169  }
4170  rb_str_resize(tmp, t - RSTRING_PTR(tmp));
4171  return tmp;
4172 }
4173 
4174 
4175 /*
4176  * call-seq:
4177  * Regexp.escape(string) -> new_string
4178  *
4179  * Returns a new string that escapes any characters
4180  * that have special meaning in a regular expression:
4181  *
4182  * s = Regexp.escape('\*?{}.') # => "\\\\\\*\\?\\{\\}\\."
4183  *
4184  * For any string +s+, this call returns a MatchData object:
4185  *
4186  * r = Regexp.new(Regexp.escape(s)) # => /\\\\\\\*\\\?\\\{\\\}\\\./
4187  * r.match(s) # => #<MatchData "\\\\\\*\\?\\{\\}\\.">
4188  *
4189  */
4190 
4191 static VALUE
4192 rb_reg_s_quote(VALUE c, VALUE str)
4193 {
4194  return rb_reg_quote(reg_operand(str, TRUE));
4195 }
4196 
4197 int
4199 {
4200  int options;
4201 
4202  rb_reg_check(re);
4203  options = RREGEXP_PTR(re)->options & ARG_REG_OPTION_MASK;
4204  if (RBASIC(re)->flags & KCODE_FIXED) options |= ARG_ENCODING_FIXED;
4205  if (RBASIC(re)->flags & REG_ENCODING_NONE) options |= ARG_ENCODING_NONE;
4206  return options;
4207 }
4208 
4209 static VALUE
4210 rb_check_regexp_type(VALUE re)
4211 {
4212  return rb_check_convert_type(re, T_REGEXP, "Regexp", "to_regexp");
4213 }
4214 
4215 /*
4216  * call-seq:
4217  * Regexp.try_convert(object) -> regexp or nil
4218  *
4219  * Returns +object+ if it is a regexp:
4220  *
4221  * Regexp.try_convert(/re/) # => /re/
4222  *
4223  * Otherwise if +object+ responds to <tt>:to_regexp</tt>,
4224  * calls <tt>object.to_regexp</tt> and returns the result.
4225  *
4226  * Returns +nil+ if +object+ does not respond to <tt>:to_regexp</tt>.
4227  *
4228  * Regexp.try_convert('re') # => nil
4229  *
4230  * Raises an exception unless <tt>object.to_regexp</tt> returns a regexp.
4231  *
4232  */
4233 static VALUE
4234 rb_reg_s_try_convert(VALUE dummy, VALUE re)
4235 {
4236  return rb_check_regexp_type(re);
4237 }
4238 
4239 static VALUE
4240 rb_reg_s_union(VALUE self, VALUE args0)
4241 {
4242  long argc = RARRAY_LEN(args0);
4243 
4244  if (argc == 0) {
4245  VALUE args[1];
4246  args[0] = rb_str_new2("(?!)");
4247  return rb_class_new_instance(1, args, rb_cRegexp);
4248  }
4249  else if (argc == 1) {
4250  VALUE arg = rb_ary_entry(args0, 0);
4251  VALUE re = rb_check_regexp_type(arg);
4252  if (!NIL_P(re))
4253  return re;
4254  else {
4255  VALUE quoted;
4256  quoted = rb_reg_s_quote(Qnil, arg);
4257  return rb_reg_new_str(quoted, 0);
4258  }
4259  }
4260  else {
4261  int i;
4262  VALUE source = rb_str_buf_new(0);
4263  rb_encoding *result_enc;
4264 
4265  int has_asciionly = 0;
4266  rb_encoding *has_ascii_compat_fixed = 0;
4267  rb_encoding *has_ascii_incompat = 0;
4268 
4269  for (i = 0; i < argc; i++) {
4270  volatile VALUE v;
4271  VALUE e = rb_ary_entry(args0, i);
4272 
4273  if (0 < i)
4274  rb_str_buf_cat_ascii(source, "|");
4275 
4276  v = rb_check_regexp_type(e);
4277  if (!NIL_P(v)) {
4278  rb_encoding *enc = rb_enc_get(v);
4279  if (!rb_enc_asciicompat(enc)) {
4280  if (!has_ascii_incompat)
4281  has_ascii_incompat = enc;
4282  else if (has_ascii_incompat != enc)
4283  rb_raise(rb_eArgError, "incompatible encodings: %s and %s",
4284  rb_enc_name(has_ascii_incompat), rb_enc_name(enc));
4285  }
4286  else if (rb_reg_fixed_encoding_p(v)) {
4287  if (!has_ascii_compat_fixed)
4288  has_ascii_compat_fixed = enc;
4289  else if (has_ascii_compat_fixed != enc)
4290  rb_raise(rb_eArgError, "incompatible encodings: %s and %s",
4291  rb_enc_name(has_ascii_compat_fixed), rb_enc_name(enc));
4292  }
4293  else {
4294  has_asciionly = 1;
4295  }
4296  v = rb_reg_str_with_term(v, -1);
4297  }
4298  else {
4299  rb_encoding *enc;
4300  StringValue(e);
4301  enc = rb_enc_get(e);
4302  if (!rb_enc_asciicompat(enc)) {
4303  if (!has_ascii_incompat)
4304  has_ascii_incompat = enc;
4305  else if (has_ascii_incompat != enc)
4306  rb_raise(rb_eArgError, "incompatible encodings: %s and %s",
4307  rb_enc_name(has_ascii_incompat), rb_enc_name(enc));
4308  }
4309  else if (rb_enc_str_asciionly_p(e)) {
4310  has_asciionly = 1;
4311  }
4312  else {
4313  if (!has_ascii_compat_fixed)
4314  has_ascii_compat_fixed = enc;
4315  else if (has_ascii_compat_fixed != enc)
4316  rb_raise(rb_eArgError, "incompatible encodings: %s and %s",
4317  rb_enc_name(has_ascii_compat_fixed), rb_enc_name(enc));
4318  }
4319  v = rb_reg_s_quote(Qnil, e);
4320  }
4321  if (has_ascii_incompat) {
4322  if (has_asciionly) {
4323  rb_raise(rb_eArgError, "ASCII incompatible encoding: %s",
4324  rb_enc_name(has_ascii_incompat));
4325  }
4326  if (has_ascii_compat_fixed) {
4327  rb_raise(rb_eArgError, "incompatible encodings: %s and %s",
4328  rb_enc_name(has_ascii_incompat), rb_enc_name(has_ascii_compat_fixed));
4329  }
4330  }
4331 
4332  if (i == 0) {
4333  rb_enc_copy(source, v);
4334  }
4335  rb_str_append(source, v);
4336  }
4337 
4338  if (has_ascii_incompat) {
4339  result_enc = has_ascii_incompat;
4340  }
4341  else if (has_ascii_compat_fixed) {
4342  result_enc = has_ascii_compat_fixed;
4343  }
4344  else {
4345  result_enc = rb_ascii8bit_encoding();
4346  }
4347 
4348  rb_enc_associate(source, result_enc);
4349  return rb_class_new_instance(1, &source, rb_cRegexp);
4350  }
4351 }
4352 
4353 /*
4354  * call-seq:
4355  * Regexp.union(*patterns) -> regexp
4356  * Regexp.union(array_of_patterns) -> regexp
4357  *
4358  * Returns a new regexp that is the union of the given patterns:
4359  *
4360  * r = Regexp.union(%w[cat dog]) # => /cat|dog/
4361  * r.match('cat') # => #<MatchData "cat">
4362  * r.match('dog') # => #<MatchData "dog">
4363  * r.match('cog') # => nil
4364  *
4365  * For each pattern that is a string, <tt>Regexp.new(pattern)</tt> is used:
4366  *
4367  * Regexp.union('penzance') # => /penzance/
4368  * Regexp.union('a+b*c') # => /a\+b\*c/
4369  * Regexp.union('skiing', 'sledding') # => /skiing|sledding/
4370  * Regexp.union(['skiing', 'sledding']) # => /skiing|sledding/
4371  *
4372  * For each pattern that is a regexp, it is used as is,
4373  * including its flags:
4374  *
4375  * Regexp.union(/foo/i, /bar/m, /baz/x)
4376  * # => /(?i-mx:foo)|(?m-ix:bar)|(?x-mi:baz)/
4377  * Regexp.union([/foo/i, /bar/m, /baz/x])
4378  * # => /(?i-mx:foo)|(?m-ix:bar)|(?x-mi:baz)/
4379  *
4380  * With no arguments, returns <tt>/(?!)/</tt>:
4381  *
4382  * Regexp.union # => /(?!)/
4383  *
4384  * If any regexp pattern contains captures, the behavior is unspecified.
4385  *
4386  */
4387 static VALUE
4388 rb_reg_s_union_m(VALUE self, VALUE args)
4389 {
4390  VALUE v;
4391  if (RARRAY_LEN(args) == 1 &&
4392  !NIL_P(v = rb_check_array_type(rb_ary_entry(args, 0)))) {
4393  return rb_reg_s_union(self, v);
4394  }
4395  return rb_reg_s_union(self, args);
4396 }
4397 
4398 /*
4399  * call-seq:
4400  * Regexp.linear_time?(re)
4401  * Regexp.linear_time?(string, options = 0)
4402  *
4403  * Returns +true+ if matching against <tt>re</tt> can be
4404  * done in linear time to the input string.
4405  *
4406  * Regexp.linear_time?(/re/) # => true
4407  *
4408  * Note that this is a property of the ruby interpreter, not of the argument
4409  * regular expression. Identical regexp can or cannot run in linear time
4410  * depending on your ruby binary. Neither forward nor backward compatibility
4411  * is guaranteed about the return value of this method. Our current algorithm
4412  * is (*1) but this is subject to change in the future. Alternative
4413  * implementations can also behave differently. They might always return
4414  * false for everything.
4415  *
4416  * (*1): https://doi.org/10.1109/SP40001.2021.00032
4417  *
4418  */
4419 static VALUE
4420 rb_reg_s_linear_time_p(int argc, VALUE *argv, VALUE self)
4421 {
4422  struct reg_init_args args;
4423  VALUE re = reg_extract_args(argc, argv, &args);
4424 
4425  if (NIL_P(re)) {
4426  re = reg_init_args(rb_reg_alloc(), args.str, args.enc, args.flags);
4427  }
4428 
4429  return RBOOL(onig_check_linear_time(RREGEXP_PTR(re)));
4430 }
4431 
4432 /* :nodoc: */
4433 static VALUE
4434 rb_reg_init_copy(VALUE copy, VALUE re)
4435 {
4436  if (!OBJ_INIT_COPY(copy, re)) return copy;
4437  rb_reg_check(re);
4438  return reg_copy(copy, re);
4439 }
4440 
4441 VALUE
4442 rb_reg_regsub(VALUE str, VALUE src, struct re_registers *regs, VALUE regexp)
4443 {
4444  VALUE val = 0;
4445  char *p, *s, *e;
4446  int no, clen;
4447  rb_encoding *str_enc = rb_enc_get(str);
4448  rb_encoding *src_enc = rb_enc_get(src);
4449  int acompat = rb_enc_asciicompat(str_enc);
4450  long n;
4451 #define ASCGET(s,e,cl) (acompat ? (*(cl)=1,ISASCII((s)[0])?(s)[0]:-1) : rb_enc_ascget((s), (e), (cl), str_enc))
4452 
4453  RSTRING_GETMEM(str, s, n);
4454  p = s;
4455  e = s + n;
4456 
4457  while (s < e) {
4458  int c = ASCGET(s, e, &clen);
4459  char *ss;
4460 
4461  if (c == -1) {
4462  s += mbclen(s, e, str_enc);
4463  continue;
4464  }
4465  ss = s;
4466  s += clen;
4467 
4468  if (c != '\\' || s == e) continue;
4469 
4470  if (!val) {
4471  val = rb_str_buf_new(ss-p);
4472  }
4473  rb_enc_str_buf_cat(val, p, ss-p, str_enc);
4474 
4475  c = ASCGET(s, e, &clen);
4476  if (c == -1) {
4477  s += mbclen(s, e, str_enc);
4478  rb_enc_str_buf_cat(val, ss, s-ss, str_enc);
4479  p = s;
4480  continue;
4481  }
4482  s += clen;
4483 
4484  p = s;
4485  switch (c) {
4486  case '1': case '2': case '3': case '4':
4487  case '5': case '6': case '7': case '8': case '9':
4488  if (!NIL_P(regexp) && onig_noname_group_capture_is_active(RREGEXP_PTR(regexp))) {
4489  no = c - '0';
4490  }
4491  else {
4492  continue;
4493  }
4494  break;
4495 
4496  case 'k':
4497  if (s < e && ASCGET(s, e, &clen) == '<') {
4498  char *name, *name_end;
4499 
4500  name_end = name = s + clen;
4501  while (name_end < e) {
4502  c = ASCGET(name_end, e, &clen);
4503  if (c == '>') break;
4504  name_end += c == -1 ? mbclen(name_end, e, str_enc) : clen;
4505  }
4506  if (name_end < e) {
4507  VALUE n = rb_str_subseq(str, (long)(name - RSTRING_PTR(str)),
4508  (long)(name_end - name));
4509  if ((no = NAME_TO_NUMBER(regs, regexp, n, name, name_end)) < 1) {
4510  name_to_backref_error(n);
4511  }
4512  p = s = name_end + clen;
4513  break;
4514  }
4515  else {
4516  rb_raise(rb_eRuntimeError, "invalid group name reference format");
4517  }
4518  }
4519 
4520  rb_enc_str_buf_cat(val, ss, s-ss, str_enc);
4521  continue;
4522 
4523  case '0':
4524  case '&':
4525  no = 0;
4526  break;
4527 
4528  case '`':
4529  rb_enc_str_buf_cat(val, RSTRING_PTR(src), BEG(0), src_enc);
4530  continue;
4531 
4532  case '\'':
4533  rb_enc_str_buf_cat(val, RSTRING_PTR(src)+END(0), RSTRING_LEN(src)-END(0), src_enc);
4534  continue;
4535 
4536  case '+':
4537  no = regs->num_regs-1;
4538  while (BEG(no) == -1 && no > 0) no--;
4539  if (no == 0) continue;
4540  break;
4541 
4542  case '\\':
4543  rb_enc_str_buf_cat(val, s-clen, clen, str_enc);
4544  continue;
4545 
4546  default:
4547  rb_enc_str_buf_cat(val, ss, s-ss, str_enc);
4548  continue;
4549  }
4550 
4551  if (no >= 0) {
4552  if (no >= regs->num_regs) continue;
4553  if (BEG(no) == -1) continue;
4554  rb_enc_str_buf_cat(val, RSTRING_PTR(src)+BEG(no), END(no)-BEG(no), src_enc);
4555  }
4556  }
4557 
4558  if (!val) return str;
4559  if (p < e) {
4560  rb_enc_str_buf_cat(val, p, e-p, str_enc);
4561  }
4562 
4563  return val;
4564 }
4565 
4566 static VALUE
4567 ignorecase_getter(ID _x, VALUE *_y)
4568 {
4569  rb_category_warn(RB_WARN_CATEGORY_DEPRECATED, "variable $= is no longer effective");
4570  return Qfalse;
4571 }
4572 
4573 static void
4574 ignorecase_setter(VALUE val, ID id, VALUE *_)
4575 {
4576  rb_category_warn(RB_WARN_CATEGORY_DEPRECATED, "variable $= is no longer effective; ignored");
4577 }
4578 
4579 static VALUE
4580 match_getter(void)
4581 {
4582  VALUE match = rb_backref_get();
4583 
4584  if (NIL_P(match)) return Qnil;
4585  rb_match_busy(match);
4586  return match;
4587 }
4588 
4589 static VALUE
4590 get_LAST_MATCH_INFO(ID _x, VALUE *_y)
4591 {
4592  return match_getter();
4593 }
4594 
4595 static void
4596 match_setter(VALUE val, ID _x, VALUE *_y)
4597 {
4598  if (!NIL_P(val)) {
4599  Check_Type(val, T_MATCH);
4600  }
4601  rb_backref_set(val);
4602 }
4603 
4604 /*
4605  * call-seq:
4606  * Regexp.last_match -> matchdata or nil
4607  * Regexp.last_match(n) -> string or nil
4608  * Regexp.last_match(name) -> string or nil
4609  *
4610  * With no argument, returns the value of <tt>$~</tt>,
4611  * which is the result of the most recent pattern match
4612  * (see {Regexp global variables}[rdoc-ref:Regexp@Global+Variables]):
4613  *
4614  * /c(.)t/ =~ 'cat' # => 0
4615  * Regexp.last_match # => #<MatchData "cat" 1:"a">
4616  * /a/ =~ 'foo' # => nil
4617  * Regexp.last_match # => nil
4618  *
4619  * With non-negative integer argument +n+, returns the _n_th field in the
4620  * matchdata, if any, or nil if none:
4621  *
4622  * /c(.)t/ =~ 'cat' # => 0
4623  * Regexp.last_match(0) # => "cat"
4624  * Regexp.last_match(1) # => "a"
4625  * Regexp.last_match(2) # => nil
4626  *
4627  * With negative integer argument +n+, counts backwards from the last field:
4628  *
4629  * Regexp.last_match(-1) # => "a"
4630  *
4631  * With string or symbol argument +name+,
4632  * returns the string value for the named capture, if any:
4633  *
4634  * /(?<lhs>\w+)\s*=\s*(?<rhs>\w+)/ =~ 'var = val'
4635  * Regexp.last_match # => #<MatchData "var = val" lhs:"var"rhs:"val">
4636  * Regexp.last_match(:lhs) # => "var"
4637  * Regexp.last_match('rhs') # => "val"
4638  * Regexp.last_match('foo') # Raises IndexError.
4639  *
4640  */
4641 
4642 static VALUE
4643 rb_reg_s_last_match(int argc, VALUE *argv, VALUE _)
4644 {
4645  if (rb_check_arity(argc, 0, 1) == 1) {
4646  VALUE match = rb_backref_get();
4647  int n;
4648  if (NIL_P(match)) return Qnil;
4649  n = match_backref_number(match, argv[0]);
4650  return rb_reg_nth_match(n, match);
4651  }
4652  return match_getter();
4653 }
4654 
4655 static void
4656 re_warn(const char *s)
4657 {
4658  rb_warn("%s", s);
4659 }
4660 
4661 // This function is periodically called during regexp matching
4662 bool
4663 rb_reg_timeout_p(regex_t *reg, void *end_time_)
4664 {
4665  rb_hrtime_t *end_time = (rb_hrtime_t *)end_time_;
4666 
4667  if (*end_time == 0) {
4668  // This is the first time to check interrupts;
4669  // just measure the current time and determine the end time
4670  // if timeout is set.
4671  rb_hrtime_t timelimit = reg->timelimit;
4672 
4673  if (!timelimit) {
4674  // no per-object timeout.
4675  timelimit = rb_reg_match_time_limit;
4676  }
4677 
4678  if (timelimit) {
4679  *end_time = rb_hrtime_add(timelimit, rb_hrtime_now());
4680  }
4681  else {
4682  // no timeout is set
4683  *end_time = RB_HRTIME_MAX;
4684  }
4685  }
4686  else {
4687  if (*end_time < rb_hrtime_now()) {
4688  // Timeout has exceeded
4689  return true;
4690  }
4691  }
4692 
4693  return false;
4694 }
4695 
4696 /*
4697  * call-seq:
4698  * Regexp.timeout -> float or nil
4699  *
4700  * It returns the current default timeout interval for Regexp matching in second.
4701  * +nil+ means no default timeout configuration.
4702  */
4703 
4704 static VALUE
4705 rb_reg_s_timeout_get(VALUE dummy)
4706 {
4707  double d = hrtime2double(rb_reg_match_time_limit);
4708  if (d == 0.0) return Qnil;
4709  return DBL2NUM(d);
4710 }
4711 
4712 /*
4713  * call-seq:
4714  * Regexp.timeout = float or nil
4715  *
4716  * It sets the default timeout interval for Regexp matching in second.
4717  * +nil+ means no default timeout configuration.
4718  * This configuration is process-global. If you want to set timeout for
4719  * each Regexp, use +timeout+ keyword for <code>Regexp.new</code>.
4720  *
4721  * Regexp.timeout = 1
4722  * /^a*b?a*$/ =~ "a" * 100000 + "x" #=> regexp match timeout (RuntimeError)
4723  */
4724 
4725 static VALUE
4726 rb_reg_s_timeout_set(VALUE dummy, VALUE timeout)
4727 {
4728  rb_ractor_ensure_main_ractor("can not access Regexp.timeout from non-main Ractors");
4729 
4730  set_timeout(&rb_reg_match_time_limit, timeout);
4731 
4732  return timeout;
4733 }
4734 
4735 /*
4736  * call-seq:
4737  * rxp.timeout -> float or nil
4738  *
4739  * It returns the timeout interval for Regexp matching in second.
4740  * +nil+ means no default timeout configuration.
4741  *
4742  * This configuration is per-object. The global configuration set by
4743  * Regexp.timeout= is ignored if per-object configuration is set.
4744  *
4745  * re = Regexp.new("^a*b?a*$", timeout: 1)
4746  * re.timeout #=> 1.0
4747  * re =~ "a" * 100000 + "x" #=> regexp match timeout (RuntimeError)
4748  */
4749 
4750 static VALUE
4751 rb_reg_timeout_get(VALUE re)
4752 {
4753  rb_reg_check(re);
4754  double d = hrtime2double(RREGEXP_PTR(re)->timelimit);
4755  if (d == 0.0) return Qnil;
4756  return DBL2NUM(d);
4757 }
4758 
4759 /*
4760  * Document-class: RegexpError
4761  *
4762  * Raised when given an invalid regexp expression.
4763  *
4764  * Regexp.new("?")
4765  *
4766  * <em>raises the exception:</em>
4767  *
4768  * RegexpError: target of repeat operator is not specified: /?/
4769  */
4770 
4771 /*
4772  * Document-class: Regexp
4773  *
4774  * :include: doc/_regexp.rdoc
4775  */
4776 
4777 void
4778 Init_Regexp(void)
4779 {
4781 
4782  onigenc_set_default_encoding(ONIG_ENCODING_ASCII);
4783  onig_set_warn_func(re_warn);
4784  onig_set_verb_warn_func(re_warn);
4785 
4786  rb_define_virtual_variable("$~", get_LAST_MATCH_INFO, match_setter);
4787  rb_define_virtual_variable("$&", last_match_getter, 0);
4788  rb_define_virtual_variable("$`", prematch_getter, 0);
4789  rb_define_virtual_variable("$'", postmatch_getter, 0);
4790  rb_define_virtual_variable("$+", last_paren_match_getter, 0);
4791 
4792  rb_gvar_ractor_local("$~");
4793  rb_gvar_ractor_local("$&");
4794  rb_gvar_ractor_local("$`");
4795  rb_gvar_ractor_local("$'");
4796  rb_gvar_ractor_local("$+");
4797 
4798  rb_define_virtual_variable("$=", ignorecase_getter, ignorecase_setter);
4799 
4800  rb_cRegexp = rb_define_class("Regexp", rb_cObject);
4801  rb_define_alloc_func(rb_cRegexp, rb_reg_s_alloc);
4803  rb_define_singleton_method(rb_cRegexp, "quote", rb_reg_s_quote, 1);
4804  rb_define_singleton_method(rb_cRegexp, "escape", rb_reg_s_quote, 1);
4805  rb_define_singleton_method(rb_cRegexp, "union", rb_reg_s_union_m, -2);
4806  rb_define_singleton_method(rb_cRegexp, "last_match", rb_reg_s_last_match, -1);
4807  rb_define_singleton_method(rb_cRegexp, "try_convert", rb_reg_s_try_convert, 1);
4808  rb_define_singleton_method(rb_cRegexp, "linear_time?", rb_reg_s_linear_time_p, -1);
4809 
4810  rb_define_method(rb_cRegexp, "initialize", rb_reg_initialize_m, -1);
4811  rb_define_method(rb_cRegexp, "initialize_copy", rb_reg_init_copy, 1);
4812  rb_define_method(rb_cRegexp, "hash", rb_reg_hash, 0);
4813  rb_define_method(rb_cRegexp, "eql?", rb_reg_equal, 1);
4814  rb_define_method(rb_cRegexp, "==", rb_reg_equal, 1);
4816  rb_define_method(rb_cRegexp, "===", rb_reg_eqq, 1);
4818  rb_define_method(rb_cRegexp, "match", rb_reg_match_m, -1);
4819  rb_define_method(rb_cRegexp, "match?", rb_reg_match_m_p, -1);
4820  rb_define_method(rb_cRegexp, "to_s", rb_reg_to_s, 0);
4821  rb_define_method(rb_cRegexp, "inspect", rb_reg_inspect, 0);
4822  rb_define_method(rb_cRegexp, "source", rb_reg_source, 0);
4823  rb_define_method(rb_cRegexp, "casefold?", rb_reg_casefold_p, 0);
4824  rb_define_method(rb_cRegexp, "options", rb_reg_options_m, 0);
4825  rb_define_method(rb_cRegexp, "encoding", rb_obj_encoding, 0); /* in encoding.c */
4826  rb_define_method(rb_cRegexp, "fixed_encoding?", rb_reg_fixed_encoding_p, 0);
4827  rb_define_method(rb_cRegexp, "names", rb_reg_names, 0);
4828  rb_define_method(rb_cRegexp, "named_captures", rb_reg_named_captures, 0);
4829  rb_define_method(rb_cRegexp, "timeout", rb_reg_timeout_get, 0);
4830 
4831  rb_eRegexpTimeoutError = rb_define_class_under(rb_cRegexp, "TimeoutError", rb_eRegexpError);
4832  rb_define_singleton_method(rb_cRegexp, "timeout", rb_reg_s_timeout_get, 0);
4833  rb_define_singleton_method(rb_cRegexp, "timeout=", rb_reg_s_timeout_set, 1);
4834 
4835  /* see Regexp.options and Regexp.new */
4836  rb_define_const(rb_cRegexp, "IGNORECASE", INT2FIX(ONIG_OPTION_IGNORECASE));
4837  /* see Regexp.options and Regexp.new */
4838  rb_define_const(rb_cRegexp, "EXTENDED", INT2FIX(ONIG_OPTION_EXTEND));
4839  /* see Regexp.options and Regexp.new */
4840  rb_define_const(rb_cRegexp, "MULTILINE", INT2FIX(ONIG_OPTION_MULTILINE));
4841  /* see Regexp.options and Regexp.new */
4842  rb_define_const(rb_cRegexp, "FIXEDENCODING", INT2FIX(ARG_ENCODING_FIXED));
4843  /* see Regexp.options and Regexp.new */
4844  rb_define_const(rb_cRegexp, "NOENCODING", INT2FIX(ARG_ENCODING_NONE));
4845 
4846  rb_global_variable(&reg_cache);
4847 
4848  rb_cMatch = rb_define_class("MatchData", rb_cObject);
4849  rb_define_alloc_func(rb_cMatch, match_alloc);
4851  rb_undef_method(CLASS_OF(rb_cMatch), "allocate");
4852 
4853  rb_define_method(rb_cMatch, "initialize_copy", match_init_copy, 1);
4854  rb_define_method(rb_cMatch, "regexp", match_regexp, 0);
4855  rb_define_method(rb_cMatch, "names", match_names, 0);
4856  rb_define_method(rb_cMatch, "size", match_size, 0);
4857  rb_define_method(rb_cMatch, "length", match_size, 0);
4858  rb_define_method(rb_cMatch, "offset", match_offset, 1);
4859  rb_define_method(rb_cMatch, "byteoffset", match_byteoffset, 1);
4860  rb_define_method(rb_cMatch, "bytebegin", match_bytebegin, 1);
4861  rb_define_method(rb_cMatch, "byteend", match_byteend, 1);
4862  rb_define_method(rb_cMatch, "begin", match_begin, 1);
4863  rb_define_method(rb_cMatch, "end", match_end, 1);
4864  rb_define_method(rb_cMatch, "match", match_nth, 1);
4865  rb_define_method(rb_cMatch, "match_length", match_nth_length, 1);
4866  rb_define_method(rb_cMatch, "to_a", match_to_a, 0);
4867  rb_define_method(rb_cMatch, "[]", match_aref, -1);
4868  rb_define_method(rb_cMatch, "captures", match_captures, 0);
4869  rb_define_alias(rb_cMatch, "deconstruct", "captures");
4870  rb_define_method(rb_cMatch, "named_captures", match_named_captures, -1);
4871  rb_define_method(rb_cMatch, "deconstruct_keys", match_deconstruct_keys, 1);
4872  rb_define_method(rb_cMatch, "values_at", match_values_at, -1);
4873  rb_define_method(rb_cMatch, "pre_match", rb_reg_match_pre, 0);
4874  rb_define_method(rb_cMatch, "post_match", rb_reg_match_post, 0);
4875  rb_define_method(rb_cMatch, "to_s", match_to_s, 0);
4876  rb_define_method(rb_cMatch, "inspect", match_inspect, 0);
4877  rb_define_method(rb_cMatch, "string", match_string, 0);
4878  rb_define_method(rb_cMatch, "hash", match_hash, 0);
4879  rb_define_method(rb_cMatch, "eql?", match_equal, 1);
4880  rb_define_method(rb_cMatch, "==", match_equal, 1);
4881 }
#define rb_define_singleton_method(klass, mid, func, arity)
Defines klass.mid.
Definition: cxxanyargs.hpp:685
static bool rb_enc_isprint(OnigCodePoint c, rb_encoding *enc)
Identical to rb_isprint(), except it additionally takes an encoding.
Definition: ctype.h:180
static bool rb_enc_isspace(OnigCodePoint c, rb_encoding *enc)
Identical to rb_isspace(), except it additionally takes an encoding.
Definition: ctype.h:194
VALUE rb_define_class(const char *name, VALUE super)
Defines a top-level class.
Definition: class.c:980
VALUE rb_define_class_under(VALUE outer, const char *name, VALUE super)
Defines a class under the namespace of outer.
Definition: class.c:1012
void rb_define_alias(VALUE klass, const char *name1, const char *name2)
Defines an alias of a method.
Definition: class.c:2345
void rb_undef_method(VALUE klass, const char *name)
Defines an undef of a method.
Definition: class.c:2166
int rb_scan_args(int argc, const VALUE *argv, const char *fmt,...)
Retrieves argument from argc and argv to given VALUE references according to the format string.
Definition: class.c:2635
void rb_define_method(VALUE klass, const char *name, VALUE(*func)(ANYARGS), int argc)
Defines a method.
Definition: class.c:2142
int rb_block_given_p(void)
Determines if the current method is given a block.
Definition: eval.c:916
int rb_get_kwargs(VALUE keyword_hash, const ID *table, int required, int optional, VALUE *values)
Keyword argument deconstructor.
Definition: class.c:2424
#define rb_str_new2
Old name of rb_str_new_cstr.
Definition: string.h:1675
#define ENC_CODERANGE_7BIT
Old name of RUBY_ENC_CODERANGE_7BIT.
Definition: coderange.h:180
#define REALLOC_N
Old name of RB_REALLOC_N.
Definition: memory.h:398
#define OBJ_INIT_COPY(obj, orig)
Old name of RB_OBJ_INIT_COPY.
Definition: object.h:41
#define ISSPACE
Old name of rb_isspace.
Definition: ctype.h:88
#define T_STRING
Old name of RUBY_T_STRING.
Definition: value_type.h:78
#define ENC_CODERANGE_CLEAN_P(cr)
Old name of RB_ENC_CODERANGE_CLEAN_P.
Definition: coderange.h:183
#define Qundef
Old name of RUBY_Qundef.
#define INT2FIX
Old name of RB_INT2FIX.
Definition: long.h:48
#define rb_str_buf_new2
Old name of rb_str_buf_new_cstr.
Definition: string.h:1679
#define ENC_CODERANGE(obj)
Old name of RB_ENC_CODERANGE.
Definition: coderange.h:184
#define CLASS_OF
Old name of rb_class_of.
Definition: globals.h:203
#define ENC_CODERANGE_UNKNOWN
Old name of RUBY_ENC_CODERANGE_UNKNOWN.
Definition: coderange.h:179
#define ENCODING_GET(obj)
Old name of RB_ENCODING_GET.
Definition: encoding.h:109
#define LONG2FIX
Old name of RB_INT2FIX.
Definition: long.h:49
#define FIX2INT
Old name of RB_FIX2INT.
Definition: int.h:41
#define NUM2DBL
Old name of rb_num2dbl.
Definition: double.h:27
#define rb_str_new3
Old name of rb_str_new_shared.
Definition: string.h:1676
#define MBCLEN_CHARFOUND_LEN(ret)
Old name of ONIGENC_MBCLEN_CHARFOUND_LEN.
Definition: encoding.h:517
#define FL_TEST_RAW
Old name of RB_FL_TEST_RAW.
Definition: fl_type.h:132
#define FL_SET
Old name of RB_FL_SET.
Definition: fl_type.h:129
#define LONG2NUM
Old name of RB_LONG2NUM.
Definition: long.h:50
#define rb_exc_new3
Old name of rb_exc_new_str.
Definition: error.h:38
#define MBCLEN_INVALID_P(ret)
Old name of ONIGENC_MBCLEN_INVALID_P.
Definition: encoding.h:518
#define Qtrue
Old name of RUBY_Qtrue.
#define ST2FIX
Old name of RB_ST2FIX.
Definition: st_data_t.h:33
#define MBCLEN_NEEDMORE_P(ret)
Old name of ONIGENC_MBCLEN_NEEDMORE_P.
Definition: encoding.h:519
#define NUM2INT
Old name of RB_NUM2INT.
Definition: int.h:44
#define INT2NUM
Old name of RB_INT2NUM.
Definition: int.h:43
#define Qnil
Old name of RUBY_Qnil.
#define Qfalse
Old name of RUBY_Qfalse.
#define ENC_CODERANGE_BROKEN
Old name of RUBY_ENC_CODERANGE_BROKEN.
Definition: coderange.h:182
#define T_ARRAY
Old name of RUBY_T_ARRAY.
Definition: value_type.h:56
#define scan_hex(s, l, e)
Old name of ruby_scan_hex.
Definition: util.h:108
#define NIL_P
Old name of RB_NIL_P.
#define MBCLEN_CHARFOUND_P(ret)
Old name of ONIGENC_MBCLEN_CHARFOUND_P.
Definition: encoding.h:516
#define FL_WB_PROTECTED
Old name of RUBY_FL_WB_PROTECTED.
Definition: fl_type.h:59
#define T_SYMBOL
Old name of RUBY_T_SYMBOL.
Definition: value_type.h:80
#define DBL2NUM
Old name of rb_float_new.
Definition: double.h:29
#define T_MATCH
Old name of RUBY_T_MATCH.
Definition: value_type.h:69
#define FL_TEST
Old name of RB_FL_TEST.
Definition: fl_type.h:131
#define NUM2LONG
Old name of RB_NUM2LONG.
Definition: long.h:51
#define FL_UNSET
Old name of RB_FL_UNSET.
Definition: fl_type.h:133
#define FIXNUM_P
Old name of RB_FIXNUM_P.
#define scan_oct(s, l, e)
Old name of ruby_scan_oct.
Definition: util.h:85
#define rb_ary_new2
Old name of rb_ary_new_capa.
Definition: array.h:657
#define FL_SET_RAW
Old name of RB_FL_SET_RAW.
Definition: fl_type.h:130
#define rb_str_new4
Old name of rb_str_new_frozen.
Definition: string.h:1677
#define SYMBOL_P
Old name of RB_SYMBOL_P.
Definition: value_type.h:88
#define T_REGEXP
Old name of RUBY_T_REGEXP.
Definition: value_type.h:77
void rb_category_warn(rb_warning_category_t category, const char *fmt,...)
Identical to rb_category_warning(), except it reports unless $VERBOSE is nil.
Definition: error.c:476
void rb_raise(VALUE exc, const char *fmt,...)
Exception entry point.
Definition: error.c:3635
void rb_exc_raise(VALUE mesg)
Raises an exception in the current thread.
Definition: eval.c:676
void rb_bug(const char *fmt,...)
Interpreter panic switch.
Definition: error.c:1089
VALUE rb_eStandardError
StandardError exception.
Definition: error.c:1405
void rb_set_errinfo(VALUE err)
Sets the current exception ($!) to the given value.
Definition: eval.c:1926
VALUE rb_eRegexpError
RegexpError exception.
Definition: re.c:32
#define ruby_verbose
This variable controls whether the interpreter is in debug mode.
Definition: error.h:475
VALUE rb_eTypeError
TypeError exception.
Definition: error.c:1408
VALUE rb_eEncCompatError
Encoding::CompatibilityError exception.
Definition: error.c:1415
VALUE rb_eRuntimeError
RuntimeError exception.
Definition: error.c:1406
void rb_warn(const char *fmt,...)
Identical to rb_warning(), except it reports unless $VERBOSE is nil.
Definition: error.c:466
VALUE rb_eArgError
ArgumentError exception.
Definition: error.c:1409
VALUE rb_eIndexError
IndexError exception.
Definition: error.c:1410
@ RB_WARN_CATEGORY_DEPRECATED
Warning is for deprecated features.
Definition: error.h:48
VALUE rb_check_convert_type(VALUE val, int type, const char *name, const char *mid)
Identical to rb_convert_type(), except it returns RUBY_Qnil instead of raising exceptions,...
Definition: object.c:3115
VALUE rb_any_to_s(VALUE obj)
Generates a textual representation of the given object.
Definition: object.c:669
VALUE rb_class_new_instance(int argc, const VALUE *argv, VALUE klass)
Allocates, then initialises an instance of the given class.
Definition: object.c:2134
VALUE rb_cMatch
MatchData class.
Definition: re.c:967
VALUE rb_obj_hide(VALUE obj)
Make the object invisible from Ruby code.
Definition: object.c:104
VALUE rb_class_new_instance_pass_kw(int argc, const VALUE *argv, VALUE klass)
Identical to rb_class_new_instance(), except it passes the passed keywords if any to the #initialize ...
Definition: object.c:2111
VALUE rb_cRegexp
Regexp class.
Definition: re.c:2640
VALUE rb_obj_class(VALUE obj)
Queries the class of an object.
Definition: object.c:247
VALUE rb_obj_freeze(VALUE obj)
Just calls rb_obj_freeze_inline() inside.
Definition: object.c:1260
#define RB_OBJ_WRITE(old, slot, young)
Declaration of a "back" pointer.
Definition: gc.h:603
Encoding relates APIs.
rb_encoding * rb_default_external_encoding(void)
Queries the "default external" encoding.
Definition: encoding.c:1589
int rb_enc_dummy_p(rb_encoding *enc)
Queries if the passed encoding is dummy.
Definition: encoding.c:197
int rb_enc_precise_mbclen(const char *p, const char *e, rb_encoding *enc)
Queries the number of bytes of the character at the passed pointer.
Definition: encoding.c:1191
VALUE rb_enc_associate(VALUE obj, rb_encoding *enc)
Identical to rb_enc_associate_index(), except it takes an encoding itself instead of its index.
Definition: encoding.c:1022
rb_encoding * rb_usascii_encoding(void)
Queries the encoding that represents US-ASCII.
Definition: encoding.c:1487
static char * rb_enc_left_char_head(const char *s, const char *p, const char *e, rb_encoding *enc)
Queries the left boundary of a character.
Definition: encoding.h:683
void rb_enc_copy(VALUE dst, VALUE src)
Destructively copies the encoding of the latter object to that of former one.
Definition: encoding.c:1149
int rb_utf8_encindex(void)
Identical to rb_utf8_encoding(), except it returns the encoding's index instead of the encoding itsel...
Definition: encoding.c:1481
int rb_ascii8bit_encindex(void)
Identical to rb_ascii8bit_encoding(), except it returns the encoding's index instead of the encoding ...
Definition: encoding.c:1469
int rb_enc_unicode_p(rb_encoding *enc)
Queries if the passed encoding is either one of UTF-8/16/32.
Definition: encoding.c:638
rb_encoding * rb_default_internal_encoding(void)
Queries the "default internal" encoding.
Definition: encoding.c:1676
rb_encoding * rb_ascii8bit_encoding(void)
Queries the encoding that represents ASCII-8BIT a.k.a.
Definition: encoding.c:1463
VALUE rb_enc_from_encoding(rb_encoding *enc)
Queries the Ruby-level counterpart instance of rb_cEncoding that corresponds to the passed encoding.
Definition: encoding.c:182
static bool rb_enc_asciicompat(rb_encoding *enc)
Queries if the passed encoding is in some sense compatible with ASCII.
Definition: encoding.h:768
rb_encoding * rb_utf8_encoding(void)
Queries the encoding that represents UTF-8.
Definition: encoding.c:1475
static int rb_enc_mbcput(unsigned int c, void *buf, rb_encoding *enc)
Identical to rb_enc_uint_chr(), except it writes back to the passed buffer instead of allocating one.
Definition: encoding.h:643
int rb_char_to_option_kcode(int c, int *option, int *kcode)
Converts a character option to its encoding.
Definition: re.c:333
static int rb_enc_mbmaxlen(rb_encoding *enc)
Queries the maximum number of bytes that the passed encoding needs to represent a character.
Definition: encoding.h:447
rb_encoding * rb_enc_get(VALUE obj)
Identical to rb_enc_get_index(), except the return type.
Definition: encoding.c:1028
static OnigCodePoint rb_enc_mbc_to_codepoint(const char *p, const char *e, rb_encoding *enc)
Identical to rb_enc_codepoint(), except it assumes the passed character is not broken.
Definition: encoding.h:591
static const char * rb_enc_name(rb_encoding *enc)
Queries the (canonical) name of the passed encoding.
Definition: encoding.h:417
static int rb_enc_mbminlen(rb_encoding *enc)
Queries the minimum number of bytes that the passed encoding needs to represent a character.
Definition: encoding.h:432
int rb_enc_ascget(const char *p, const char *e, int *len, rb_encoding *enc)
Queries the code point of character pointed by the passed pointer.
Definition: encoding.c:1203
VALUE rb_enc_reg_new(const char *ptr, long len, rb_encoding *enc, int opts)
Identical to rb_reg_new(), except it additionally takes an encoding.
Definition: re.c:3442
int rb_enc_str_coderange(VALUE str)
Scans the passed string to collect its code range.
Definition: string.c:900
long rb_memsearch(const void *x, long m, const void *y, long n, rb_encoding *enc)
Looks for the passed string in the passed buffer.
Definition: re.c:252
long rb_enc_strlen(const char *head, const char *tail, rb_encoding *enc)
Counts the number of characters of the passed string, according to the passed encoding.
Definition: string.c:2260
VALUE rb_enc_str_buf_cat(VALUE str, const char *ptr, long len, rb_encoding *enc)
Identical to rb_str_cat(), except it additionally takes an encoding.
Definition: string.c:3612
VALUE rb_enc_str_new(const char *ptr, long len, rb_encoding *enc)
Identical to rb_str_new(), except it additionally takes an encoding.
Definition: string.c:1068
int rb_enc_str_asciionly_p(VALUE str)
Queries if the passed string is "ASCII only".
Definition: string.c:919
VALUE rb_obj_encoding(VALUE obj)
Identical to rb_enc_get_index(), except the return type.
Definition: encoding.c:1163
long rb_str_coderange_scan_restartable(const char *str, const char *end, rb_encoding *enc, int *cr)
Scans the passed string until it finds something odd.
Definition: string.c:784
VALUE rb_str_encode(VALUE str, VALUE to, int ecflags, VALUE ecopts)
Converts the contents of the passed string from its encoding to the passed one.
Definition: transcode.c:2914
void rb_memerror(void)
Triggers out-of-memory error.
Definition: gc.c:4188
#define RGENGC_WB_PROTECTED_MATCH
This is a compile-time flag to enable/disable write barrier for struct RMatch.
Definition: gc.h:512
void rb_gc(void)
Triggers a GC process.
Definition: gc.c:3404
void rb_global_variable(VALUE *)
An alias for rb_gc_register_address().
Definition: gc.c:2876
#define RGENGC_WB_PROTECTED_REGEXP
This is a compile-time flag to enable/disable write barrier for struct RRegexp.
Definition: gc.h:501
VALUE rb_check_array_type(VALUE obj)
Try converting an object to its array representation using its to_ary method, if any.
Definition: array.c:1014
VALUE rb_ary_new_capa(long capa)
Identical to rb_ary_new(), except it additionally specifies how many rooms of objects it should alloc...
Definition: array.c:741
VALUE rb_ary_resize(VALUE ary, long len)
Expands or shrinks the passed array to the passed length.
Definition: array.c:2296
VALUE rb_ary_push(VALUE ary, VALUE elem)
Special case of rb_ary_cat() that it adds only one element.
Definition: array.c:1384
VALUE rb_ary_entry(VALUE ary, long off)
Queries an element of an array.
Definition: array.c:1737
VALUE rb_assoc_new(VALUE car, VALUE cdr)
Identical to rb_ary_new_from_values(), except it expects exactly two parameters.
Definition: array.c:1001
void rb_ary_store(VALUE ary, long key, VALUE val)
Destructively stores the passed value to the passed array's passed index.
Definition: array.c:1207
int rb_uv_to_utf8(char buf[6], unsigned long uv)
Encodes a Unicode codepoint into its UTF-8 representation.
Definition: pack.c:1601
static int rb_check_arity(int argc, int min, int max)
Ensures that the passed integer is in the passed range.
Definition: error.h:284
VALUE rb_hash_aset(VALUE hash, VALUE key, VALUE val)
Inserts or replaces ("upsert"s) the objects into the given hash table.
Definition: hash.c:2893
VALUE rb_hash_new(void)
Creates a new, empty hash object.
Definition: hash.c:1475
VALUE rb_backref_get(void)
Queries the last match, or Regexp.last_match, or the $~.
Definition: vm.c:1825
VALUE rb_lastline_get(void)
Queries the last line, or the $_.
Definition: vm.c:1837
void rb_backref_set(VALUE md)
Updates $~.
Definition: vm.c:1831
VALUE rb_range_beg_len(VALUE range, long *begp, long *lenp, long len, int err)
Deconstructs a numerical range.
Definition: range.c:1842
int rb_reg_backref_number(VALUE match, VALUE backref)
Queries the index of the given named capture.
Definition: re.c:1235
int rb_reg_options(VALUE re)
Queries the options of the passed regular expression.
Definition: re.c:4198
VALUE rb_reg_last_match(VALUE md)
This just returns the argument, stringified.
Definition: re.c:1930
VALUE rb_reg_match(VALUE re, VALUE str)
This is the match operator.
Definition: re.c:3695
void rb_match_busy(VALUE md)
Asserts that the given MatchData is "occupied".
Definition: re.c:1489
VALUE rb_reg_nth_match(int n, VALUE md)
Queries the nth captured substring.
Definition: re.c:1905
VALUE rb_reg_match_post(VALUE md)
The portion of the original string after the given match.
Definition: re.c:1987
VALUE rb_reg_nth_defined(int n, VALUE md)
Identical to rb_reg_nth_match(), except it just returns Boolean.
Definition: re.c:1888
VALUE rb_reg_match_pre(VALUE md)
The portion of the original string before the given match.
Definition: re.c:1954
VALUE rb_reg_new_str(VALUE src, int opts)
Identical to rb_reg_new(), except it takes the expression in Ruby's string instead of C's.
Definition: re.c:3402
VALUE rb_reg_match_last(VALUE md)
The portion of the original string that captured at the very last.
Definition: re.c:2020
VALUE rb_reg_match2(VALUE re)
Identical to rb_reg_match(), except it matches against rb_lastline_get() (or, the $_).
Definition: re.c:3750
VALUE rb_reg_new(const char *src, long len, int opts)
Creates a new Regular expression.
Definition: re.c:3456
int rb_memcicmp(const void *s1, const void *s2, long n)
Identical to st_locale_insensitive_strcasecmp(), except it is timing safe and returns something diffe...
Definition: re.c:95
#define rb_hash_uint(h, i)
Just another name of st_hash_uint.
Definition: string.h:942
#define rb_hash_end(h)
Just another name of st_hash_end.
Definition: string.h:945
VALUE rb_str_append(VALUE dst, VALUE src)
Identical to rb_str_buf_append(), except it converts the right hand side before concatenating.
Definition: string.c:3677
long rb_str_offset(VALUE str, long pos)
"Inverse" of rb_str_sublen().
Definition: string.c:2958
VALUE rb_str_subseq(VALUE str, long beg, long len)
Identical to rb_str_substr(), except the numbers are interpreted as byte offsets instead of character...
Definition: string.c:3055
st_index_t rb_memhash(const void *ptr, long len)
This is a universal hash function.
Definition: random.c:1752
VALUE rb_str_buf_cat(VALUE, const char *, long)
Just another name of rb_str_cat.
VALUE rb_str_dup(VALUE str)
Duplicates a string.
Definition: string.c:1927
st_index_t rb_str_hash(VALUE str)
Calculates a hash value of a string.
Definition: string.c:4032
VALUE rb_str_buf_cat2(VALUE, const char *)
Just another name of rb_str_cat_cstr.
char * rb_str_subpos(VALUE str, long beg, long *len)
Identical to rb_str_substr(), except it returns a C's string instead of Ruby's.
Definition: string.c:3063
VALUE rb_str_buf_append(VALUE dst, VALUE src)
Identical to rb_str_cat_cstr(), except it takes Ruby's string instead of C's.
Definition: string.c:3643
long rb_str_sublen(VALUE str, long pos)
Byte offset to character offset conversion.
Definition: string.c:3005
VALUE rb_str_equal(VALUE str1, VALUE str2)
Equality of two strings.
Definition: string.c:4148
st_index_t rb_hash_start(st_index_t i)
Starts a series of hashing.
Definition: random.c:1746
VALUE rb_str_inspect(VALUE str)
Generates a "readable" version of the receiver.
Definition: string.c:7198
VALUE rb_str_buf_cat_ascii(VALUE dst, const char *src)
Identical to rb_str_cat_cstr(), except it additionally assumes the source string be a NUL terminated ...
Definition: string.c:3619
VALUE rb_str_new(const char *ptr, long len)
Allocates an instance of rb_cString.
Definition: string.c:1050
VALUE rb_check_string_type(VALUE obj)
Try converting an object to its stringised representation using its to_str method,...
Definition: string.c:2854
VALUE rb_str_resize(VALUE str, long len)
Overwrites the length of the string.
Definition: string.c:3317
VALUE rb_str_buf_new(long capa)
Allocates a "string buffer".
Definition: string.c:1643
VALUE rb_str_length(VALUE)
Identical to rb_str_strlen(), except it returns the value in rb_cInteger.
Definition: string.c:2363
VALUE rb_str_intern(VALUE str)
Identical to rb_to_symbol(), except it assumes the receiver being an instance of RString.
Definition: symbol.c:878
VALUE rb_class_path(VALUE mod)
Identical to rb_mod_name(), except it returns #<Class: ...> style inspection for anonymous modules.
Definition: variable.c:293
void rb_define_alloc_func(VALUE klass, rb_alloc_func_t func)
Sets the allocator function of a class.
static ID rb_intern_const(const char *str)
This is a "tiny optimisation" over rb_intern().
Definition: symbol.h:276
VALUE rb_sym2str(VALUE id)
Identical to rb_id2str(), except it takes an instance of rb_cSymbol rather than an ID.
Definition: symbol.c:970
void rb_define_virtual_variable(const char *name, rb_gvar_getter_t *getter, rb_gvar_setter_t *setter)
Defines a global variable that is purely function-backended.
Definition: variable.c:738
void rb_define_const(VALUE klass, const char *name, VALUE val)
Defines a Ruby level constant under a namespace.
Definition: variable.c:3714
char * ptr
Pointer to the underlying memory region, of at least capa bytes.
Definition: io.h:2
int len
Length of the buffer.
Definition: io.h:8
regex_t * rb_reg_prepare_re(VALUE re, VALUE str)
Exercises various checks and preprocesses so that the given regular expression can be applied to the ...
Definition: re.c:1635
long rb_reg_search(VALUE re, VALUE str, long pos, int dir)
Runs the passed regular expression over the passed string.
Definition: re.c:1844
long rb_reg_adjust_startpos(VALUE re, VALUE str, long pos, int dir)
Tell us if this is a wrong idea, but it seems this function has no usage at all.
Definition: re.c:1739
OnigPosition rb_reg_onig_match(VALUE re, VALUE str, OnigPosition(*match)(regex_t *reg, VALUE str, struct re_registers *regs, void *args), void *args, struct re_registers *regs)
Runs a regular expression match using function match.
Definition: re.c:1703
VALUE rb_reg_regcomp(VALUE str)
Creates a new instance of rb_cRegexp.
Definition: re.c:3479
VALUE rb_reg_quote(VALUE str)
Escapes any characters that would have special meaning in a regular expression.
Definition: re.c:4078
VALUE rb_reg_regsub(VALUE repl, VALUE src, struct re_registers *regs, VALUE rexp)
Substitution.
Definition: re.c:4442
int rb_reg_region_copy(struct re_registers *dst, const struct re_registers *src)
Duplicates a match data.
Definition: re.c:984
unsigned long ruby_scan_hex(const char *str, size_t len, size_t *ret)
Interprets the passed string a hexadecimal unsigned integer.
Definition: util.c:61
unsigned long ruby_scan_oct(const char *str, size_t len, size_t *consumed)
Interprets the passed string as an octal unsigned integer.
Definition: util.c:43
VALUE rb_sprintf(const char *fmt,...)
Ruby's extended sprintf(3).
Definition: sprintf.c:1217
VALUE rb_str_catf(VALUE dst, const char *fmt,...)
Identical to rb_sprintf(), except it renders the output to the specified object rather than creating ...
Definition: sprintf.c:1240
VALUE rb_yield(VALUE val)
Yields the block.
Definition: vm_eval.c:1354
#define MEMCPY(p1, p2, type, n)
Handy macro to call memcpy.
Definition: memory.h:367
#define ALLOCA_N(type, n)
Definition: memory.h:287
#define MEMZERO(p, type, n)
Handy macro to erase a region of memory.
Definition: memory.h:355
#define RB_GC_GUARD(v)
Prevents premature destruction of local objects.
Definition: memory.h:162
#define RARRAY_LEN
Just another name of rb_array_len.
Definition: rarray.h:51
#define RARRAY_AREF(a, i)
Definition: rarray.h:403
#define RBASIC(obj)
Convenient casting macro.
Definition: rbasic.h:40
#define RMATCH(obj)
Convenient casting macro.
Definition: rmatch.h:37
static struct re_registers * RMATCH_REGS(VALUE match)
Queries the raw re_registers.
Definition: rmatch.h:138
#define RREGEXP(obj)
Convenient casting macro.
Definition: rregexp.h:37
static VALUE RREGEXP_SRC(VALUE rexp)
Convenient getter function.
Definition: rregexp.h:103
static char * RREGEXP_SRC_PTR(VALUE rexp)
Convenient getter function.
Definition: rregexp.h:125
#define RREGEXP_PTR(obj)
Convenient accessor macro.
Definition: rregexp.h:45
static long RREGEXP_SRC_LEN(VALUE rexp)
Convenient getter function.
Definition: rregexp.h:144
#define StringValue(v)
Ensures that the parameter object is a String.
Definition: rstring.h:66
static char * RSTRING_END(VALUE str)
Queries the end of the contents pointer of the string.
Definition: rstring.h:442
static char * RSTRING_PTR(VALUE str)
Queries the contents pointer of the string.
Definition: rstring.h:416
#define RSTRING_GETMEM(str, ptrvar, lenvar)
Convenient macro to obtain the contents and length at once.
Definition: rstring.h:488
static long RSTRING_LEN(VALUE str)
Queries the length of the string.
Definition: rstring.h:367
VALUE rb_str_to_str(VALUE obj)
Identical to rb_check_string_type(), except it raises exceptions in case of conversion failures.
Definition: string.c:1715
#define StringValueCStr(v)
Identical to StringValuePtr, except it additionally checks for the contents for viability as a C stri...
Definition: rstring.h:89
#define RTEST
This is an old name of RB_TEST.
#define _(args)
This was a transition path from K&R to ANSI.
Definition: stdarg.h:35
MEMO.
Definition: imemo.h:109
VALUE flags
Per-object flags.
Definition: rbasic.h:75
Regular expression execution context.
Definition: rmatch.h:96
VALUE regexp
The expression of this match.
Definition: rmatch.h:109
VALUE str
The target string that the match was made against.
Definition: rmatch.h:104
Ruby's regular expression.
Definition: rregexp.h:60
struct RBasic basic
Basic part, including flags and class.
Definition: rregexp.h:63
const VALUE src
Source code of this expression.
Definition: rregexp.h:74
unsigned long usecnt
Reference count.
Definition: rregexp.h:90
struct re_pattern_buffer * ptr
The pattern buffer.
Definition: rregexp.h:71
Definition: re.c:994
Represents a match.
Definition: rmatch.h:71
struct rmatch_offset * char_offset
Capture group offsets, in C array.
Definition: rmatch.h:79
int char_offset_num_allocated
Number of rmatch_offset that ::rmatch::char_offset holds.
Definition: rmatch.h:82
struct re_registers regs
"Registers" of a match.
Definition: rmatch.h:76
Represents the region of a capture group.
Definition: rmatch.h:65
long beg
Beginning of a group.
Definition: rmatch.h:66
long end
End of a group.
Definition: rmatch.h:67
uintptr_t ID
Type that represents a Ruby identifier such as a variable name.
Definition: value.h:52
#define SIZEOF_VALUE
Identical to sizeof(VALUE), except it is a macro that can also be used inside of preprocessor directi...
Definition: value.h:69
uintptr_t VALUE
Type that represents a Ruby object.
Definition: value.h:40
static void Check_Type(VALUE v, enum ruby_value_type t)
Identical to RB_TYPE_P(), except it raises exceptions on predication failure.
Definition: value_type.h:433
static bool RB_TYPE_P(VALUE obj, enum ruby_value_type t)
Queries if the given object is of given type.
Definition: value_type.h:376