3 * Copyright (C) 1993, 1994, 1996, 1997, 1999, 2000, 2001, 2003,
4 * 2005, 2006, 2007, 2008 by Larry Wall and others
6 * You may distribute under the terms of either the GNU General Public
7 * License or the Artistic License, as specified in the README file.
12 * Definitions etc. for regexp(3) routines.
14 * Caveat: this is V8 regexp(3) [actually, a reimplementation thereof],
15 * not the System V one.
17 #ifndef PLUGGABLE_RE_EXTENSION
18 /* we don't want to include this stuff if we are inside of
19 an external regex engine based on the core one - like re 'debug'*/
27 typedef struct regnode regnode;
29 struct reg_substr_data;
36 struct reg_substr_datum {
39 SV *substr; /* non-utf8 variant */
40 SV *utf8_substr; /* utf8 variant */
43 struct reg_substr_data {
44 struct reg_substr_datum data[3]; /* Actual array */
47 #ifdef PERL_OLD_COPY_ON_WRITE
48 #define SV_SAVED_COPY SV *saved_copy; /* If non-NULL, SV which is COW from original */
53 /* offsets within a string of a particular /(.)/ capture */
55 typedef struct regexp_paren_pair {
58 /* 'start_tmp' records a new opening position before the matching end
59 * has been found, so that the old start and end values are still
61 * "abc" =~ /(.(?{print "[$1]"}))+/
63 * This field is not part of the API. */
67 #if defined(PERL_IN_REGCOMP_C) || defined(PERL_IN_UTF8_C)
68 #define _invlist_union(a, b, output) _invlist_union_maybe_complement_2nd(a, b, FALSE, output)
69 #define _invlist_intersection(a, b, output) _invlist_intersection_maybe_complement_2nd(a, b, FALSE, output)
71 /* Subtracting b from a leaves in a everything that was there that isn't in b,
72 * that is the intersection of a with b's complement */
73 #define _invlist_subtract(a, b, output) _invlist_intersection_maybe_complement_2nd(a, b, TRUE, output)
76 /* record the position of a (?{...}) within a pattern */
78 struct reg_code_block {
87 The regexp/REGEXP struct, see L<perlreapi> for further documentation
88 on the individual fields. The struct is ordered so that the most
89 commonly used fields are placed at the start.
91 Any patch that adds items to this struct will need to include
92 changes to F<sv.c> (C<Perl_re_dup()>) and F<regcomp.c>
93 (C<pregfree()>). This involves freeing or cloning items in the
94 regexp's data array based on the data item's type.
97 #define _REGEXP_COMMON \
98 /* what engine created this regexp? */ \
99 const struct regexp_engine* engine; \
100 REGEXP *mother_re; /* what re is this a lightweight copy of? */ \
101 HV *paren_names; /* Optional hash of paren names */ \
102 /* Information about the match that the perl core uses to */ \
103 /* manage things */ \
104 U32 extflags; /* Flags used both externally and internally */ \
105 I32 minlen; /* mininum possible length of string to match */\
106 I32 minlenret; /* mininum possible length of $& */ \
107 U32 gofs; /* chars left of pos that we search from */ \
108 /* substring data about strings that must appear in the */ \
109 /* final match, used for optimisations */ \
110 struct reg_substr_data *substrs; \
111 U32 nparens; /* number of capture buffers */ \
112 /* private engine specific data */ \
113 U32 intflags; /* Engine Specific Internal flags */ \
114 void *pprivate; /* Data private to the regex engine which */ \
115 /* created this object. */ \
116 /* Data about the last/current match. These are modified */ \
117 /* during matching */ \
118 U32 lastparen; /* last open paren matched */ \
119 U32 lastcloseparen; /* last close paren matched */ \
120 regexp_paren_pair *swap; /* Unused: 5.10.1 and later */ \
121 /* Array of offsets for (@-) and (@+) */ \
122 regexp_paren_pair *offs; \
123 /* saved or original string so \digit works forever. */ \
125 SV_SAVED_COPY /* If non-NULL, SV which is COW from original */\
126 I32 sublen; /* Length of string pointed by subbeg */ \
127 /* Information about the match that isn't often used */ \
128 /* offset from wrapped to the start of precomp */ \
129 PERL_BITFIELD32 pre_prefix:4; \
130 CV *qr_anoncv /* the anon sub wrapped round qr/(?{..})/ */
132 typedef struct regexp {
137 #define RXp_PAREN_NAMES(rx) ((rx)->paren_names)
139 /* used for high speed searches */
140 typedef struct re_scream_pos_data_s
142 char **scream_olds; /* match pos */
143 I32 *scream_pos; /* Internal iterator of scream. */
144 } re_scream_pos_data;
146 /* regexp_engine structure. This is the dispatch table for regexes.
147 * Any regex engine implementation must be able to build one of these.
149 typedef struct regexp_engine {
150 REGEXP* (*comp) (pTHX_ SV * const pattern, U32 flags);
151 I32 (*exec) (pTHX_ REGEXP * const rx, char* stringarg, char* strend,
152 char* strbeg, I32 minend, SV* screamer,
153 void* data, U32 flags);
154 char* (*intuit) (pTHX_ REGEXP * const rx, SV *sv, char *strpos,
155 char *strend, const U32 flags,
156 re_scream_pos_data *data);
157 SV* (*checkstr) (pTHX_ REGEXP * const rx);
158 void (*free) (pTHX_ REGEXP * const rx);
159 void (*numbered_buff_FETCH) (pTHX_ REGEXP * const rx, const I32 paren,
161 void (*numbered_buff_STORE) (pTHX_ REGEXP * const rx, const I32 paren,
162 SV const * const value);
163 I32 (*numbered_buff_LENGTH) (pTHX_ REGEXP * const rx, const SV * const sv,
165 SV* (*named_buff) (pTHX_ REGEXP * const rx, SV * const key,
166 SV * const value, const U32 flags);
167 SV* (*named_buff_iter) (pTHX_ REGEXP * const rx, const SV * const lastkey,
169 SV* (*qr_package)(pTHX_ REGEXP * const rx);
171 void* (*dupe) (pTHX_ REGEXP * const rx, CLONE_PARAMS *param);
173 REGEXP* (*op_comp) (pTHX_ SV ** const patternp, int pat_count,
174 OP *expr, const struct regexp_engine* eng,
176 bool *is_bare_re, U32 orig_rx_flags, U32 pm_flags);
180 These are passed to the numbered capture variable callbacks as the
181 paren name. >= 1 is reserved for actual numbered captures, i.e. $1,
184 #define RX_BUFF_IDX_PREMATCH -2 /* $` / ${^PREMATCH} */
185 #define RX_BUFF_IDX_POSTMATCH -1 /* $' / ${^POSTMATCH} */
186 #define RX_BUFF_IDX_FULLMATCH 0 /* $& / ${^MATCH} */
189 Flags that are passed to the named_buff and named_buff_iter
190 callbacks above. Those routines are called from universal.c via the
191 Tie::Hash::NamedCapture interface for %+ and %- and the re::
192 functions in the same file.
195 /* The Tie::Hash::NamedCapture operation this is part of, if any */
196 #define RXapif_FETCH 0x0001
197 #define RXapif_STORE 0x0002
198 #define RXapif_DELETE 0x0004
199 #define RXapif_CLEAR 0x0008
200 #define RXapif_EXISTS 0x0010
201 #define RXapif_SCALAR 0x0020
202 #define RXapif_FIRSTKEY 0x0040
203 #define RXapif_NEXTKEY 0x0080
205 /* Whether %+ or %- is being operated on */
206 #define RXapif_ONE 0x0100 /* %+ */
207 #define RXapif_ALL 0x0200 /* %- */
209 /* Whether this is being called from a re:: function */
210 #define RXapif_REGNAME 0x0400
211 #define RXapif_REGNAMES 0x0800
212 #define RXapif_REGNAMES_COUNT 0x1000
215 =head1 REGEXP Functions
217 =for apidoc Am|REGEXP *|SvRX|SV *sv
219 Convenience macro to get the REGEXP from a SV. This is approximately
220 equivalent to the following snippet:
225 sv = MUTABLE_SV(SvRV(sv));
226 if (SvTYPE(sv) == SVt_REGEXP)
229 NULL will be returned if a REGEXP* is not found.
231 =for apidoc Am|bool|SvRXOK|SV* sv
233 Returns a boolean indicating whether the SV (or the one it references)
236 If you want to do something with the REGEXP* later use SvRX instead
242 #define SvRX(sv) (Perl_get_re_arg(aTHX_ sv))
243 #define SvRXOK(sv) (Perl_get_re_arg(aTHX_ sv) ? TRUE : FALSE)
246 /* Flags stored in regexp->extflags
247 * These are used by code external to the regexp engine
249 * Note that the flags whose names start with RXf_PMf_ are defined in
250 * op_reg_common.h, being copied from the parallel flags of op_pmflags
252 * NOTE: if you modify any RXf flags you should run regen.pl or
253 * regen/regcomp.pl so that regnodes.h is updated with the changes.
257 #include "op_reg_common.h"
259 #define RXf_PMf_STD_PMMOD (RXf_PMf_MULTILINE|RXf_PMf_SINGLELINE|RXf_PMf_FOLD|RXf_PMf_EXTENDED)
261 #define CASE_STD_PMMOD_FLAGS_PARSE_SET(pmfl) \
262 case IGNORE_PAT_MOD: *(pmfl) |= RXf_PMf_FOLD; break; \
263 case MULTILINE_PAT_MOD: *(pmfl) |= RXf_PMf_MULTILINE; break; \
264 case SINGLE_PAT_MOD: *(pmfl) |= RXf_PMf_SINGLELINE; break; \
265 case XTENDED_PAT_MOD: *(pmfl) |= RXf_PMf_EXTENDED; break
267 /* Note, includes charset ones, assumes 0 is the default for them */
268 #define STD_PMMOD_FLAGS_CLEAR(pmfl) \
269 *(pmfl) &= ~(RXf_PMf_FOLD|RXf_PMf_MULTILINE|RXf_PMf_SINGLELINE|RXf_PMf_EXTENDED|RXf_PMf_CHARSET)
271 /* chars and strings used as regex pattern modifiers
272 * Singular is a 'c'har, plural is a "string"
274 * NOTE, KEEPCOPY was originally 'k', but was changed to 'p' for preserve
275 * for compatibility reasons with Regexp::Common which highjacked (?k:...)
276 * for its own uses. So 'k' is out as well.
278 #define DEFAULT_PAT_MOD '^' /* Short for all the default modifiers */
279 #define EXEC_PAT_MOD 'e'
280 #define KEEPCOPY_PAT_MOD 'p'
281 #define ONCE_PAT_MOD 'o'
282 #define GLOBAL_PAT_MOD 'g'
283 #define CONTINUE_PAT_MOD 'c'
284 #define MULTILINE_PAT_MOD 'm'
285 #define SINGLE_PAT_MOD 's'
286 #define IGNORE_PAT_MOD 'i'
287 #define XTENDED_PAT_MOD 'x'
288 #define NONDESTRUCT_PAT_MOD 'r'
289 #define LOCALE_PAT_MOD 'l'
290 #define UNICODE_PAT_MOD 'u'
291 #define DEPENDS_PAT_MOD 'd'
292 #define ASCII_RESTRICT_PAT_MOD 'a'
294 #define ONCE_PAT_MODS "o"
295 #define KEEPCOPY_PAT_MODS "p"
296 #define EXEC_PAT_MODS "e"
297 #define LOOP_PAT_MODS "gc"
298 #define NONDESTRUCT_PAT_MODS "r"
299 #define LOCALE_PAT_MODS "l"
300 #define UNICODE_PAT_MODS "u"
301 #define DEPENDS_PAT_MODS "d"
302 #define ASCII_RESTRICT_PAT_MODS "a"
303 #define ASCII_MORE_RESTRICT_PAT_MODS "aa"
305 /* This string is expected by regcomp.c to be ordered so that the first
306 * character is the flag in bit RXf_PMf_STD_PMMOD_SHIFT of extflags; the next
307 * character is bit +1, etc. */
308 #define STD_PAT_MODS "msix"
310 #define CHARSET_PAT_MODS ASCII_RESTRICT_PAT_MODS DEPENDS_PAT_MODS LOCALE_PAT_MODS UNICODE_PAT_MODS
312 /* This string is expected by XS_re_regexp_pattern() in universal.c to be ordered
313 * so that the first character is the flag in bit RXf_PMf_STD_PMMOD_SHIFT of
314 * extflags; the next character is in bit +1, etc. */
315 #define INT_PAT_MODS STD_PAT_MODS KEEPCOPY_PAT_MODS
317 #define EXT_PAT_MODS ONCE_PAT_MODS KEEPCOPY_PAT_MODS
318 #define QR_PAT_MODS STD_PAT_MODS EXT_PAT_MODS CHARSET_PAT_MODS
319 #define M_PAT_MODS QR_PAT_MODS LOOP_PAT_MODS
320 #define S_PAT_MODS M_PAT_MODS EXEC_PAT_MODS NONDESTRUCT_PAT_MODS
323 * NOTE: if you modify any RXf flags you should run regen.pl or
324 * regen/regcomp.pl so that regnodes.h is updated with the changes.
328 /* Leave some space, so future bit allocations can go either in the shared or
329 * unshared area without affecting binary compatibility */
330 #define RXf_BASE_SHIFT (_RXf_PMf_SHIFT_NEXT+1)
332 /* Manually decorate this function with gcc-style attributes just to
333 * avoid having to restructure the header files and their called order,
334 * as proto.h would have to be included before this file, and isn't */
336 PERL_STATIC_INLINE const char *
337 get_regex_charset_name(const U32 flags, STRLEN* const lenp)
338 __attribute__warn_unused_result__;
340 #define MAX_CHARSET_NAME_LENGTH 2
342 PERL_STATIC_INLINE const char *
343 get_regex_charset_name(const U32 flags, STRLEN* const lenp)
345 /* Returns a string that corresponds to the name of the regex character set
346 * given by 'flags', and *lenp is set the length of that string, which
347 * cannot exceed MAX_CHARSET_NAME_LENGTH characters */
350 switch (get_regex_charset(flags)) {
351 case REGEX_DEPENDS_CHARSET: return DEPENDS_PAT_MODS;
352 case REGEX_LOCALE_CHARSET: return LOCALE_PAT_MODS;
353 case REGEX_UNICODE_CHARSET: return UNICODE_PAT_MODS;
354 case REGEX_ASCII_RESTRICTED_CHARSET: return ASCII_RESTRICT_PAT_MODS;
355 case REGEX_ASCII_MORE_RESTRICTED_CHARSET:
357 return ASCII_MORE_RESTRICT_PAT_MODS;
359 return "?"; /* Unknown */
363 /* Anchor and GPOS related stuff */
364 #define RXf_ANCH_BOL (1<<(RXf_BASE_SHIFT+0))
365 #define RXf_ANCH_MBOL (1<<(RXf_BASE_SHIFT+1))
366 #define RXf_ANCH_SBOL (1<<(RXf_BASE_SHIFT+2))
367 #define RXf_ANCH_GPOS (1<<(RXf_BASE_SHIFT+3))
368 #define RXf_GPOS_SEEN (1<<(RXf_BASE_SHIFT+4))
369 #define RXf_GPOS_FLOAT (1<<(RXf_BASE_SHIFT+5))
371 #define RXf_ANCH (RXf_ANCH_BOL|RXf_ANCH_MBOL|RXf_ANCH_GPOS|RXf_ANCH_SBOL)
372 #define RXf_GPOS_CHECK (RXf_GPOS_SEEN|RXf_ANCH_GPOS)
373 #define RXf_ANCH_SINGLE (RXf_ANCH_SBOL|RXf_ANCH_GPOS)
375 /* What we have seen */
376 #define RXf_LOOKBEHIND_SEEN (1<<(RXf_BASE_SHIFT+6))
377 #define RXf_EVAL_SEEN (1<<(RXf_BASE_SHIFT+7))
378 #define RXf_CANY_SEEN (1<<(RXf_BASE_SHIFT+8))
381 #define RXf_NOSCAN (1<<(RXf_BASE_SHIFT+9))
382 #define RXf_CHECK_ALL (1<<(RXf_BASE_SHIFT+10))
385 #define RXf_MATCH_UTF8 (1<<(RXf_BASE_SHIFT+11))
388 #define RXf_USE_INTUIT_NOML (1<<(RXf_BASE_SHIFT+12))
389 #define RXf_USE_INTUIT_ML (1<<(RXf_BASE_SHIFT+13))
390 #define RXf_INTUIT_TAIL (1<<(RXf_BASE_SHIFT+14))
393 Set in Perl_pmruntime if op_flags & OPf_SPECIAL, i.e. split. Will
394 be used by regex engines to check whether they should set
397 #define RXf_SPLIT (1<<(RXf_BASE_SHIFT+15))
399 #define RXf_USE_INTUIT (RXf_USE_INTUIT_NOML|RXf_USE_INTUIT_ML)
401 /* Copy and tainted info */
402 #define RXf_COPY_DONE (1<<(RXf_BASE_SHIFT+16))
404 /* during execution: pattern temporarily tainted by executing locale ops;
405 * post-execution: $1 et al are tainted */
406 #define RXf_TAINTED_SEEN (1<<(RXf_BASE_SHIFT+17))
407 /* this pattern was tainted during compilation */
408 #define RXf_TAINTED (1<<(RXf_BASE_SHIFT+18))
410 /* Flags indicating special patterns */
411 #define RXf_START_ONLY (1<<(RXf_BASE_SHIFT+19)) /* Pattern is /^/ */
412 #define RXf_SKIPWHITE (1<<(RXf_BASE_SHIFT+20)) /* Pattern is for a split / / */
413 #define RXf_WHITE (1<<(RXf_BASE_SHIFT+21)) /* Pattern is /\s+/ */
414 #define RXf_NULL (1U<<(RXf_BASE_SHIFT+22)) /* Pattern is // */
415 #if RXf_BASE_SHIFT+22 > 31
416 # error Too many RXf_PMf bits used. See regnodes.h for any spare in middle
420 * NOTE: if you modify any RXf flags you should run regen.pl or
421 * regen/regcomp.pl so that regnodes.h is updated with the changes.
425 #define RX_HAS_CUTGROUP(prog) ((prog)->intflags & PREGf_CUTGROUP_SEEN)
426 #define RXp_MATCH_TAINTED(prog) (RXp_EXTFLAGS(prog) & RXf_TAINTED_SEEN)
427 #define RX_MATCH_TAINTED(prog) (RX_EXTFLAGS(prog) & RXf_TAINTED_SEEN)
428 #define RX_MATCH_TAINTED_on(prog) (RX_EXTFLAGS(prog) |= RXf_TAINTED_SEEN)
429 #define RX_MATCH_TAINTED_off(prog) (RX_EXTFLAGS(prog) &= ~RXf_TAINTED_SEEN)
430 #define RX_MATCH_TAINTED_set(prog, t) ((t) \
431 ? RX_MATCH_TAINTED_on(prog) \
432 : RX_MATCH_TAINTED_off(prog))
434 #define RXp_MATCH_COPIED(prog) (RXp_EXTFLAGS(prog) & RXf_COPY_DONE)
435 #define RX_MATCH_COPIED(prog) (RX_EXTFLAGS(prog) & RXf_COPY_DONE)
436 #define RXp_MATCH_COPIED_on(prog) (RXp_EXTFLAGS(prog) |= RXf_COPY_DONE)
437 #define RX_MATCH_COPIED_on(prog) (RX_EXTFLAGS(prog) |= RXf_COPY_DONE)
438 #define RXp_MATCH_COPIED_off(prog) (RXp_EXTFLAGS(prog) &= ~RXf_COPY_DONE)
439 #define RX_MATCH_COPIED_off(prog) (RX_EXTFLAGS(prog) &= ~RXf_COPY_DONE)
440 #define RX_MATCH_COPIED_set(prog,t) ((t) \
441 ? RX_MATCH_COPIED_on(prog) \
442 : RX_MATCH_COPIED_off(prog))
444 #define RXp_EXTFLAGS(rx) ((rx)->extflags)
446 /* For source compatibility. We used to store these explicitly. */
447 #define RX_PRECOMP(prog) (RX_WRAPPED(prog) + ((struct regexp *)SvANY(prog))->pre_prefix)
448 #define RX_PRECOMP_const(prog) (RX_WRAPPED_const(prog) + ((struct regexp *)SvANY(prog))->pre_prefix)
449 /* FIXME? Are we hardcoding too much here and constraining plugin extension
450 writers? Specifically, the value 1 assumes that the wrapped version always
451 has exactly one character at the end, a ')'. Will that always be true? */
452 #define RX_PRELEN(prog) (RX_WRAPLEN(prog) - ((struct regexp *)SvANY(prog))->pre_prefix - 1)
453 #define RX_WRAPPED(prog) SvPVX(prog)
454 #define RX_WRAPPED_const(prog) SvPVX_const(prog)
455 #define RX_WRAPLEN(prog) SvCUR(prog)
456 #define RX_CHECK_SUBSTR(prog) (((struct regexp *)SvANY(prog))->check_substr)
457 #define RX_REFCNT(prog) SvREFCNT(prog)
458 #if defined(__GNUC__) && !defined(PERL_GCC_BRACE_GROUPS_FORBIDDEN)
459 # define RX_EXTFLAGS(prog) \
461 const REGEXP *const _rx_extflags = (prog); \
462 assert(SvTYPE(_rx_extflags) == SVt_REGEXP); \
463 &RXp_EXTFLAGS(SvANY(_rx_extflags)); \
465 # define RX_ENGINE(prog) \
467 const REGEXP *const _rx_engine = (prog); \
468 assert(SvTYPE(_rx_engine) == SVt_REGEXP); \
469 &SvANY(_rx_engine)->engine; \
471 # define RX_SUBBEG(prog) \
473 const REGEXP *const _rx_subbeg = (prog); \
474 assert(SvTYPE(_rx_subbeg) == SVt_REGEXP); \
475 &SvANY(_rx_subbeg)->subbeg; \
477 # define RX_OFFS(prog) \
479 const REGEXP *const _rx_offs = (prog); \
480 assert(SvTYPE(_rx_offs) == SVt_REGEXP); \
481 &SvANY(_rx_offs)->offs; \
483 # define RX_NPARENS(prog) \
485 const REGEXP *const _rx_nparens = (prog); \
486 assert(SvTYPE(_rx_nparens) == SVt_REGEXP); \
487 &SvANY(_rx_nparens)->nparens; \
490 # define RX_EXTFLAGS(prog) RXp_EXTFLAGS((struct regexp *)SvANY(prog))
491 # define RX_ENGINE(prog) (((struct regexp *)SvANY(prog))->engine)
492 # define RX_SUBBEG(prog) (((struct regexp *)SvANY(prog))->subbeg)
493 # define RX_OFFS(prog) (((struct regexp *)SvANY(prog))->offs)
494 # define RX_NPARENS(prog) (((struct regexp *)SvANY(prog))->nparens)
496 #define RX_SUBLEN(prog) (((struct regexp *)SvANY(prog))->sublen)
497 #define RX_MINLEN(prog) (((struct regexp *)SvANY(prog))->minlen)
498 #define RX_MINLENRET(prog) (((struct regexp *)SvANY(prog))->minlenret)
499 #define RX_GOFS(prog) (((struct regexp *)SvANY(prog))->gofs)
500 #define RX_LASTPAREN(prog) (((struct regexp *)SvANY(prog))->lastparen)
501 #define RX_LASTCLOSEPAREN(prog) (((struct regexp *)SvANY(prog))->lastcloseparen)
502 #define RX_SAVED_COPY(prog) (((struct regexp *)SvANY(prog))->saved_copy)
504 #endif /* PLUGGABLE_RE_EXTENSION */
506 /* Stuff that needs to be included in the pluggable extension goes below here */
508 #ifdef PERL_OLD_COPY_ON_WRITE
509 #define RX_MATCH_COPY_FREE(rx) \
510 STMT_START {if (RX_SAVED_COPY(rx)) { \
511 SV_CHECK_THINKFIRST_COW_DROP(RX_SAVED_COPY(rx)); \
513 if (RX_MATCH_COPIED(rx)) { \
514 Safefree(RX_SUBBEG(rx)); \
515 RX_MATCH_COPIED_off(rx); \
518 #define RX_MATCH_COPY_FREE(rx) \
519 STMT_START {if (RX_MATCH_COPIED(rx)) { \
520 Safefree(RX_SUBBEG(rx)); \
521 RX_MATCH_COPIED_off(rx); \
525 #define RXp_MATCH_UTF8(prog) (RXp_EXTFLAGS(prog) & RXf_MATCH_UTF8)
526 #define RX_MATCH_UTF8(prog) (RX_EXTFLAGS(prog) & RXf_MATCH_UTF8)
527 #define RX_MATCH_UTF8_on(prog) (RX_EXTFLAGS(prog) |= RXf_MATCH_UTF8)
528 #define RX_MATCH_UTF8_off(prog) (RX_EXTFLAGS(prog) &= ~RXf_MATCH_UTF8)
529 #define RX_MATCH_UTF8_set(prog, t) ((t) \
530 ? (RX_MATCH_UTF8_on(prog), (PL_reg_match_utf8 = 1)) \
531 : (RX_MATCH_UTF8_off(prog), (PL_reg_match_utf8 = 0)))
533 /* Whether the pattern stored at RX_WRAPPED is in UTF-8 */
534 #define RX_UTF8(prog) SvUTF8(prog)
536 #define REXEC_COPY_STR 0x01 /* Need to copy the string. */
537 #define REXEC_CHECKED 0x02 /* check_substr already checked. */
538 #define REXEC_SCREAM 0x04 /* use scream table. */
539 #define REXEC_IGNOREPOS 0x08 /* \G matches at start. */
540 #define REXEC_NOT_FIRST 0x10 /* This is another iteration of //g. */
542 #if defined(__GNUC__) && !defined(PERL_GCC_BRACE_GROUPS_FORBIDDEN)
543 # define ReREFCNT_inc(re) \
545 /* This is here to generate a casting warning if incorrect. */ \
546 REGEXP *const _rerefcnt_inc = (re); \
547 assert(SvTYPE(_rerefcnt_inc) == SVt_REGEXP); \
548 SvREFCNT_inc(_rerefcnt_inc); \
551 # define ReREFCNT_dec(re) \
553 /* This is here to generate a casting warning if incorrect. */ \
554 REGEXP *const _rerefcnt_dec = (re); \
555 SvREFCNT_dec(_rerefcnt_dec); \
558 # define ReREFCNT_dec(re) SvREFCNT_dec(re)
559 # define ReREFCNT_inc(re) ((REGEXP *) SvREFCNT_inc(re))
562 /* FIXME for plugins. */
564 #define FBMcf_TAIL_DOLLAR 1
565 #define FBMcf_TAIL_DOLLARM 2
566 #define FBMcf_TAIL_Z 4
567 #define FBMcf_TAIL_z 8
568 #define FBMcf_TAIL (FBMcf_TAIL_DOLLAR|FBMcf_TAIL_DOLLARM|FBMcf_TAIL_Z|FBMcf_TAIL_z)
570 #define FBMrf_MULTILINE 1
572 /* some basic information about the current match that is created by
573 * Perl_regexec_flags and then passed to regtry(), regmatch() etc */
585 /* structures for holding and saving the state maintained by regmatch() */
587 #ifndef MAX_RECURSE_EVAL_NOCHANGE_DEPTH
588 #define MAX_RECURSE_EVAL_NOCHANGE_DEPTH 1000
591 typedef I32 CHECKPOINT;
593 typedef struct regmatch_state {
594 int resume_state; /* where to jump to on return */
595 char *locinput; /* where to backtrack in string on failure */
599 /* this is a fake union member that matches the first element
600 * of each member that needs to store positive backtrack
603 struct regmatch_state *prev_yes_state;
606 /* branchlike members */
607 /* this is a fake union member that matches the first elements
608 * of each member that needs to behave like a branch */
610 /* this first element must match u.yes */
611 struct regmatch_state *prev_yes_state;
619 /* the first elements must match u.branchlike */
620 struct regmatch_state *prev_yes_state;
625 regnode *next_branch; /* next branch node */
629 /* the first elements must match u.branchlike */
630 struct regmatch_state *prev_yes_state;
635 U32 accepted; /* how many accepting states left */
636 bool longfold;/* saw a fold with a 1->n char mapping */
637 U16 *jump; /* positive offsets from me */
638 regnode *me; /* Which node am I - needed for jump tries*/
639 U8 *firstpos;/* pos in string of first trie match */
640 U32 firstchars;/* len in chars of firstpos from start */
641 U16 nextword;/* next word to try */
642 U16 topword; /* longest accepted word */
645 /* special types - these members are used to store state for special
646 regops like eval, if/then, lookaround and the markpoint state */
648 /* this first element must match u.yes */
649 struct regmatch_state *prev_yes_state;
650 struct regmatch_state *prev_eval;
651 struct regmatch_state *prev_curlyx;
653 U32 toggle_reg_flags; /* what bits in PL_reg_flags to
654 flip when transitioning between
655 inner and outer rexen */
656 CHECKPOINT cp; /* remember current savestack indexes */
658 U32 close_paren; /* which close bracket is our end */
659 regnode *B; /* the node following us */
663 /* this first element must match u.yes */
664 struct regmatch_state *prev_yes_state;
666 I32 logical; /* saved copy of 'logical' var */
667 regnode *me; /* the IFMATCH/SUSPEND/UNLESSM node */
668 } ifmatch; /* and SUSPEND/UNLESSM */
671 /* this first element must match u.yes */
672 struct regmatch_state *prev_yes_state;
673 struct regmatch_state *prev_mark;
682 /* quantifiers - these members are used for storing state for
683 for the regops used to implement quantifiers */
685 /* this first element must match u.yes */
686 struct regmatch_state *prev_yes_state;
687 struct regmatch_state *prev_curlyx; /* previous cur_curlyx */
688 regnode *me; /* the CURLYX node */
689 regnode *B; /* the B node in /A*B/ */
690 CHECKPOINT cp; /* remember current savestack index */
692 int parenfloor;/* how far back to strip paren data */
694 /* these two are modified by WHILEM */
695 int count; /* how many instances of A we've matched */
696 char *lastloc;/* where previous A matched (0-len detect) */
700 /* this first element must match u.yes */
701 struct regmatch_state *prev_yes_state;
702 struct regmatch_state *save_curlyx;
703 CHECKPOINT cp; /* remember current savestack indexes */
705 char *save_lastloc; /* previous curlyx.lastloc */
711 /* this first element must match u.yes */
712 struct regmatch_state *prev_yes_state;
713 I32 c1, c2; /* case fold search */
717 I32 alen; /* length of first-matched A string */
720 regnode *A, *B; /* the nodes corresponding to /A*B/ */
721 regnode *me; /* the curlym node */
729 I32 c1, c2; /* case fold search */
730 char *maxpos; /* highest possible point in string to match */
731 char *oldloc; /* the previous locinput */
733 int min, max; /* {m,n} */
734 regnode *A, *B; /* the nodes corresponding to /A*B/ */
735 } curly; /* and CURLYN/PLUS/STAR */
740 /* how many regmatch_state structs to allocate as a single slab.
741 * We do it in 4K blocks for efficiency. The "3" is 2 for the next/prev
742 * pointers, plus 1 for any mythical malloc overhead. */
744 #define PERL_REGMATCH_SLAB_SLOTS \
745 ((4096 - 3 * sizeof (void*)) / sizeof(regmatch_state))
747 typedef struct regmatch_slab {
748 regmatch_state states[PERL_REGMATCH_SLAB_SLOTS];
749 struct regmatch_slab *prev, *next;
752 #define PL_reg_flags PL_reg_state.re_state_reg_flags
753 #define PL_bostr PL_reg_state.re_state_bostr
754 #define PL_reginput PL_reg_state.re_state_reginput
755 #define PL_regeol PL_reg_state.re_state_regeol
756 #define PL_reg_match_utf8 PL_reg_state.re_state_reg_match_utf8
757 #define PL_reg_magic PL_reg_state.re_state_reg_magic
758 #define PL_reg_oldpos PL_reg_state.re_state_reg_oldpos
759 #define PL_reg_oldcurpm PL_reg_state.re_state_reg_oldcurpm
760 #define PL_reg_curpm PL_reg_state.re_state_reg_curpm
761 #define PL_reg_oldsaved PL_reg_state.re_state_reg_oldsaved
762 #define PL_reg_oldsavedlen PL_reg_state.re_state_reg_oldsavedlen
763 #define PL_reg_maxiter PL_reg_state.re_state_reg_maxiter
764 #define PL_reg_leftiter PL_reg_state.re_state_reg_leftiter
765 #define PL_reg_poscache PL_reg_state.re_state_reg_poscache
766 #define PL_reg_poscache_size PL_reg_state.re_state_reg_poscache_size
767 #define PL_regsize PL_reg_state.re_state_regsize
768 #define PL_reg_starttry PL_reg_state.re_state_reg_starttry
769 #define PL_nrs PL_reg_state.re_state_nrs
771 struct re_save_state {
772 U32 re_state_reg_flags; /* from regexec.c */
773 bool re_state_eval_setup_done; /* from regexec.c */
774 bool re_state_reg_match_utf8; /* from regexec.c */
775 bool re_reparsing; /* runtime (?{}) fed back into parser */
776 char *re_state_bostr;
777 char *re_state_reginput; /* String-input pointer. */
778 char *re_state_regeol; /* End of input, for $ check. */
779 MAGIC *re_state_reg_magic; /* from regexec.c */
780 PMOP *re_state_reg_oldcurpm; /* from regexec.c */
781 PMOP *re_state_reg_curpm; /* from regexec.c */
782 char *re_state_reg_oldsaved; /* old saved substr during match */
783 STRLEN re_state_reg_oldsavedlen; /* old length of saved substr during match */
784 STRLEN re_state_reg_poscache_size; /* size of pos cache of WHILEM */
785 I32 re_state_reg_oldpos; /* from regexec.c */
786 I32 re_state_reg_maxiter; /* max wait until caching pos */
787 I32 re_state_reg_leftiter; /* wait until caching pos */
788 U32 re_state_regsize; /* from regexec.c */
789 char *re_state_reg_poscache; /* cache of pos of WHILEM */
790 char *re_state_reg_starttry; /* from regexec.c */
791 #ifdef PERL_OLD_COPY_ON_WRITE
792 SV *re_state_nrs; /* was placeholder: unused since 5.8.0 (5.7.2 patch #12027 for bug ID 20010815.012). Used to save rx->saved_copy */
796 #define SAVESTACK_ALLOC_FOR_RE_SAVE_STATE \
797 (1 + ((sizeof(struct re_save_state) - 1) / sizeof(*PL_savestack)))
801 * c-indentation-style: bsd
803 * indent-tabs-mode: nil
806 * ex: set ts=8 sts=4 sw=4 et: