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 typedef struct regexp_paren_pair {
59 The regexp/REGEXP struct, see L<perlreapi> for further documentation
60 on the individual fields. The struct is ordered so that the most
61 commonly used fields are placed at the start.
63 Any patch that adds items to this struct will need to include
64 changes to F<sv.c> (C<Perl_re_dup()>) and F<regcomp.c>
65 (C<pregfree()>). This involves freeing or cloning items in the
66 regexp's data array based on the data item's type.
69 #define _REGEXP_COMMON \
70 /* what engine created this regexp? */ \
71 const struct regexp_engine* engine; \
72 REGEXP *mother_re; /* what re is this a lightweight copy of? */ \
73 HV *paren_names; /* Optional hash of paren names */ \
74 /* Information about the match that the perl core uses to */ \
76 U32 extflags; /* Flags used both externally and internally */ \
77 I32 minlen; /* mininum possible length of string to match */\
78 I32 minlenret; /* mininum possible length of $& */ \
79 U32 gofs; /* chars left of pos that we search from */ \
80 /* substring data about strings that must appear in the */ \
81 /* final match, used for optimisations */ \
82 struct reg_substr_data *substrs; \
83 U32 nparens; /* number of capture buffers */ \
84 /* private engine specific data */ \
85 U32 intflags; /* Engine Specific Internal flags */ \
86 void *pprivate; /* Data private to the regex engine which */ \
87 /* created this object. */ \
88 /* Data about the last/current match. These are modified */ \
89 /* during matching */ \
90 U32 lastparen; /* last open paren matched */ \
91 U32 lastcloseparen; /* last close paren matched */ \
92 regexp_paren_pair *swap; /* Unused: 5.10.1 and later */ \
93 /* Array of offsets for (@-) and (@+) */ \
94 regexp_paren_pair *offs; \
95 /* saved or original string so \digit works forever. */ \
97 SV_SAVED_COPY /* If non-NULL, SV which is COW from original */\
98 I32 sublen; /* Length of string pointed by subbeg */ \
99 /* Information about the match that isn't often used */ \
100 /* offset from wrapped to the start of precomp */ \
101 PERL_BITFIELD32 pre_prefix:4; \
102 /* number of eval groups in the pattern - for security checks */\
103 PERL_BITFIELD32 seen_evals:28
105 typedef struct regexp {
110 #define RXp_PAREN_NAMES(rx) ((rx)->paren_names)
112 /* used for high speed searches */
113 typedef struct re_scream_pos_data_s
115 char **scream_olds; /* match pos */
116 I32 *scream_pos; /* Internal iterator of scream. */
117 } re_scream_pos_data;
119 /* regexp_engine structure. This is the dispatch table for regexes.
120 * Any regex engine implementation must be able to build one of these.
122 typedef struct regexp_engine {
123 REGEXP* (*comp) (pTHX_ SV * const pattern, U32 flags);
124 I32 (*exec) (pTHX_ REGEXP * const rx, char* stringarg, char* strend,
125 char* strbeg, I32 minend, SV* screamer,
126 void* data, U32 flags);
127 char* (*intuit) (pTHX_ REGEXP * const rx, SV *sv, char *strpos,
128 char *strend, const U32 flags,
129 re_scream_pos_data *data);
130 SV* (*checkstr) (pTHX_ REGEXP * const rx);
131 void (*free) (pTHX_ REGEXP * const rx);
132 void (*numbered_buff_FETCH) (pTHX_ REGEXP * const rx, const I32 paren,
134 void (*numbered_buff_STORE) (pTHX_ REGEXP * const rx, const I32 paren,
135 SV const * const value);
136 I32 (*numbered_buff_LENGTH) (pTHX_ REGEXP * const rx, const SV * const sv,
138 SV* (*named_buff) (pTHX_ REGEXP * const rx, SV * const key,
139 SV * const value, const U32 flags);
140 SV* (*named_buff_iter) (pTHX_ REGEXP * const rx, const SV * const lastkey,
142 SV* (*qr_package)(pTHX_ REGEXP * const rx);
144 void* (*dupe) (pTHX_ REGEXP * const rx, CLONE_PARAMS *param);
149 These are passed to the numbered capture variable callbacks as the
150 paren name. >= 1 is reserved for actual numbered captures, i.e. $1,
153 #define RX_BUFF_IDX_PREMATCH -2 /* $` / ${^PREMATCH} */
154 #define RX_BUFF_IDX_POSTMATCH -1 /* $' / ${^POSTMATCH} */
155 #define RX_BUFF_IDX_FULLMATCH 0 /* $& / ${^MATCH} */
158 Flags that are passed to the named_buff and named_buff_iter
159 callbacks above. Those routines are called from universal.c via the
160 Tie::Hash::NamedCapture interface for %+ and %- and the re::
161 functions in the same file.
164 /* The Tie::Hash::NamedCapture operation this is part of, if any */
165 #define RXapif_FETCH 0x0001
166 #define RXapif_STORE 0x0002
167 #define RXapif_DELETE 0x0004
168 #define RXapif_CLEAR 0x0008
169 #define RXapif_EXISTS 0x0010
170 #define RXapif_SCALAR 0x0020
171 #define RXapif_FIRSTKEY 0x0040
172 #define RXapif_NEXTKEY 0x0080
174 /* Whether %+ or %- is being operated on */
175 #define RXapif_ONE 0x0100 /* %+ */
176 #define RXapif_ALL 0x0200 /* %- */
178 /* Whether this is being called from a re:: function */
179 #define RXapif_REGNAME 0x0400
180 #define RXapif_REGNAMES 0x0800
181 #define RXapif_REGNAMES_COUNT 0x1000
184 =head1 REGEXP Functions
186 =for apidoc Am|REGEXP *|SvRX|SV *sv
188 Convenience macro to get the REGEXP from a SV. This is approximately
189 equivalent to the following snippet:
194 (tmpsv = (SV*)SvRV(sv)) &&
195 SvTYPE(tmpsv) == SVt_PVMG &&
196 (tmpmg = mg_find(tmpsv, PERL_MAGIC_qr)))
198 return (REGEXP *)tmpmg->mg_obj;
201 NULL will be returned if a REGEXP* is not found.
203 =for apidoc Am|bool|SvRXOK|SV* sv
205 Returns a boolean indicating whether the SV contains qr magic
208 If you want to do something with the REGEXP* later use SvRX instead
214 #define SvRX(sv) (Perl_get_re_arg(aTHX_ sv))
215 #define SvRXOK(sv) (Perl_get_re_arg(aTHX_ sv) ? TRUE : FALSE)
218 /* Flags stored in regexp->extflags
219 * These are used by code external to the regexp engine
221 * Note that the flags whose names start with RXf_PMf_ are defined in
222 * op_reg_common.h, being copied from the parallel flags of op_pmflags
224 * NOTE: if you modify any RXf flags you should run regen.pl or
225 * regen/regcomp.pl so that regnodes.h is updated with the changes.
229 #include "op_reg_common.h"
231 #define RXf_PMf_STD_PMMOD (RXf_PMf_MULTILINE|RXf_PMf_SINGLELINE|RXf_PMf_FOLD|RXf_PMf_EXTENDED)
233 #define CASE_STD_PMMOD_FLAGS_PARSE_SET(pmfl) \
234 case IGNORE_PAT_MOD: *(pmfl) |= RXf_PMf_FOLD; break; \
235 case MULTILINE_PAT_MOD: *(pmfl) |= RXf_PMf_MULTILINE; break; \
236 case SINGLE_PAT_MOD: *(pmfl) |= RXf_PMf_SINGLELINE; break; \
237 case XTENDED_PAT_MOD: *(pmfl) |= RXf_PMf_EXTENDED; break
239 /* Note, includes locale, unicode */
240 #define STD_PMMOD_FLAGS_CLEAR(pmfl) \
241 *(pmfl) &= ~(RXf_PMf_FOLD|RXf_PMf_MULTILINE|RXf_PMf_SINGLELINE|RXf_PMf_EXTENDED|RXf_PMf_LOCALE|RXf_PMf_UNICODE)
243 /* chars and strings used as regex pattern modifiers
244 * Singlular is a 'c'har, plural is a "string"
246 * NOTE, KEEPCOPY was originally 'k', but was changed to 'p' for preserve
247 * for compatibility reasons with Regexp::Common which highjacked (?k:...)
248 * for its own uses. So 'k' is out as well.
250 #define DEFAULT_PAT_MOD '^' /* Short for all the default modifiers */
251 #define EXEC_PAT_MOD 'e'
252 #define KEEPCOPY_PAT_MOD 'p'
253 #define ONCE_PAT_MOD 'o'
254 #define GLOBAL_PAT_MOD 'g'
255 #define CONTINUE_PAT_MOD 'c'
256 #define MULTILINE_PAT_MOD 'm'
257 #define SINGLE_PAT_MOD 's'
258 #define IGNORE_PAT_MOD 'i'
259 #define XTENDED_PAT_MOD 'x'
260 #define NONDESTRUCT_PAT_MOD 'r'
261 #define LOCALE_PAT_MOD 'l'
262 #define UNICODE_PAT_MOD 'u'
263 #define DUAL_PAT_MOD 'd'
265 #define ONCE_PAT_MODS "o"
266 #define KEEPCOPY_PAT_MODS "p"
267 #define EXEC_PAT_MODS "e"
268 #define LOOP_PAT_MODS "gc"
269 #define NONDESTRUCT_PAT_MODS "r"
270 #define LOCALE_PAT_MODS "l"
271 #define UNICODE_PAT_MODS "u"
272 #define DUAL_PAT_MODS "d"
274 /* This string is expected by regcomp.c to be ordered so that the first
275 * character is the flag in bit RXf_PMf_STD_PMMOD_SHIFT of extflags; the next
276 * character is bit +1, etc. */
277 #define STD_PAT_MODS "msix"
279 /* This string is expected by XS_re_regexp_pattern() in universal.c to be ordered
280 * so that the first character is the flag in bit RXf_PMf_STD_PMMOD_SHIFT of
281 * extflags; the next character is in bit +1, etc. */
282 #define INT_PAT_MODS STD_PAT_MODS KEEPCOPY_PAT_MODS
284 #define EXT_PAT_MODS ONCE_PAT_MODS KEEPCOPY_PAT_MODS
285 #define QR_PAT_MODS STD_PAT_MODS EXT_PAT_MODS
286 #define M_PAT_MODS QR_PAT_MODS LOOP_PAT_MODS
287 #define S_PAT_MODS M_PAT_MODS EXEC_PAT_MODS NONDESTRUCT_PAT_MODS
290 * NOTE: if you modify any RXf flags you should run regen.pl or
291 * regen/regcomp.pl so that regnodes.h is updated with the changes.
295 /* Leave some space, so future bit allocations can go either in the shared or
296 * unshared area without affecting binary compatibility */
297 #define RXf_BASE_SHIFT (_RXf_PMf_SHIFT_NEXT+2)
299 /* Anchor and GPOS related stuff */
300 #define RXf_ANCH_BOL (1<<(RXf_BASE_SHIFT+0))
301 #define RXf_ANCH_MBOL (1<<(RXf_BASE_SHIFT+1))
302 #define RXf_ANCH_SBOL (1<<(RXf_BASE_SHIFT+2))
303 #define RXf_ANCH_GPOS (1<<(RXf_BASE_SHIFT+3))
304 #define RXf_GPOS_SEEN (1<<(RXf_BASE_SHIFT+4))
305 #define RXf_GPOS_FLOAT (1<<(RXf_BASE_SHIFT+5))
307 #define RXf_ANCH (RXf_ANCH_BOL|RXf_ANCH_MBOL|RXf_ANCH_GPOS|RXf_ANCH_SBOL)
308 #define RXf_GPOS_CHECK (RXf_GPOS_SEEN|RXf_ANCH_GPOS)
309 #define RXf_ANCH_SINGLE (RXf_ANCH_SBOL|RXf_ANCH_GPOS)
311 /* What we have seen */
312 #define RXf_LOOKBEHIND_SEEN (1<<(RXf_BASE_SHIFT+6))
313 #define RXf_EVAL_SEEN (1<<(RXf_BASE_SHIFT+7))
314 #define RXf_CANY_SEEN (1<<(RXf_BASE_SHIFT+8))
317 #define RXf_NOSCAN (1<<(RXf_BASE_SHIFT+9))
318 #define RXf_CHECK_ALL (1<<(RXf_BASE_SHIFT+10))
321 #define RXf_MATCH_UTF8 (1<<(RXf_BASE_SHIFT+11))
324 #define RXf_USE_INTUIT_NOML (1<<(RXf_BASE_SHIFT+12))
325 #define RXf_USE_INTUIT_ML (1<<(RXf_BASE_SHIFT+13))
326 #define RXf_INTUIT_TAIL (1<<(RXf_BASE_SHIFT+14))
329 Set in Perl_pmruntime if op_flags & OPf_SPECIAL, i.e. split. Will
330 be used by regex engines to check whether they should set
333 #define RXf_SPLIT (1<<(RXf_BASE_SHIFT+15))
335 #define RXf_USE_INTUIT (RXf_USE_INTUIT_NOML|RXf_USE_INTUIT_ML)
337 /* Copy and tainted info */
338 #define RXf_COPY_DONE (1<<(RXf_BASE_SHIFT+16))
339 #define RXf_TAINTED_SEEN (1<<(RXf_BASE_SHIFT+17))
340 #define RXf_TAINTED (1<<(RXf_BASE_SHIFT+18)) /* this pattern is tainted */
342 /* Flags indicating special patterns */
343 #define RXf_START_ONLY (1<<(RXf_BASE_SHIFT+19)) /* Pattern is /^/ */
344 #define RXf_SKIPWHITE (1<<(RXf_BASE_SHIFT+20)) /* Pattern is for a split / / */
345 #define RXf_WHITE (1<<(RXf_BASE_SHIFT+21)) /* Pattern is /\s+/ */
346 #define RXf_NULL (1<<(RXf_BASE_SHIFT+22)) /* Pattern is // */
347 #if RXf_BASE_SHIFT+22 > 31
348 # error Too many RXf_PMf bits used. See regnodes.h for any spare in middle
352 * NOTE: if you modify any RXf flags you should run regen.pl or
353 * regen/regcomp.pl so that regnodes.h is updated with the changes.
357 #define RX_HAS_CUTGROUP(prog) ((prog)->intflags & PREGf_CUTGROUP_SEEN)
358 #define RXp_MATCH_TAINTED(prog) (RXp_EXTFLAGS(prog) & RXf_TAINTED_SEEN)
359 #define RX_MATCH_TAINTED(prog) (RX_EXTFLAGS(prog) & RXf_TAINTED_SEEN)
360 #define RX_MATCH_TAINTED_on(prog) (RX_EXTFLAGS(prog) |= RXf_TAINTED_SEEN)
361 #define RX_MATCH_TAINTED_off(prog) (RX_EXTFLAGS(prog) &= ~RXf_TAINTED_SEEN)
362 #define RX_MATCH_TAINTED_set(prog, t) ((t) \
363 ? RX_MATCH_TAINTED_on(prog) \
364 : RX_MATCH_TAINTED_off(prog))
366 #define RXp_MATCH_COPIED(prog) (RXp_EXTFLAGS(prog) & RXf_COPY_DONE)
367 #define RX_MATCH_COPIED(prog) (RX_EXTFLAGS(prog) & RXf_COPY_DONE)
368 #define RXp_MATCH_COPIED_on(prog) (RXp_EXTFLAGS(prog) |= RXf_COPY_DONE)
369 #define RX_MATCH_COPIED_on(prog) (RX_EXTFLAGS(prog) |= RXf_COPY_DONE)
370 #define RXp_MATCH_COPIED_off(prog) (RXp_EXTFLAGS(prog) &= ~RXf_COPY_DONE)
371 #define RX_MATCH_COPIED_off(prog) (RX_EXTFLAGS(prog) &= ~RXf_COPY_DONE)
372 #define RX_MATCH_COPIED_set(prog,t) ((t) \
373 ? RX_MATCH_COPIED_on(prog) \
374 : RX_MATCH_COPIED_off(prog))
376 #define RXp_EXTFLAGS(rx) ((rx)->extflags)
378 /* For source compatibility. We used to store these explicitly. */
379 #define RX_PRECOMP(prog) (RX_WRAPPED(prog) + ((struct regexp *)SvANY(prog))->pre_prefix)
380 #define RX_PRECOMP_const(prog) (RX_WRAPPED_const(prog) + ((struct regexp *)SvANY(prog))->pre_prefix)
381 /* FIXME? Are we hardcoding too much here and constraining plugin extension
382 writers? Specifically, the value 1 assumes that the wrapped version always
383 has exactly one character at the end, a ')'. Will that always be true? */
384 #define RX_PRELEN(prog) (RX_WRAPLEN(prog) - ((struct regexp *)SvANY(prog))->pre_prefix - 1)
385 #define RX_WRAPPED(prog) SvPVX(prog)
386 #define RX_WRAPPED_const(prog) SvPVX_const(prog)
387 #define RX_WRAPLEN(prog) SvCUR(prog)
388 #define RX_CHECK_SUBSTR(prog) (((struct regexp *)SvANY(prog))->check_substr)
389 #define RX_REFCNT(prog) SvREFCNT(prog)
390 #if defined(__GNUC__) && !defined(PERL_GCC_BRACE_GROUPS_FORBIDDEN)
391 # define RX_EXTFLAGS(prog) \
393 const REGEXP *const _rx_extflags = (prog); \
394 assert(SvTYPE(_rx_extflags) == SVt_REGEXP); \
395 &RXp_EXTFLAGS(SvANY(_rx_extflags)); \
397 # define RX_ENGINE(prog) \
399 const REGEXP *const _rx_engine = (prog); \
400 assert(SvTYPE(_rx_engine) == SVt_REGEXP); \
401 &SvANY(_rx_engine)->engine; \
403 # define RX_SUBBEG(prog) \
405 const REGEXP *const _rx_subbeg = (prog); \
406 assert(SvTYPE(_rx_subbeg) == SVt_REGEXP); \
407 &SvANY(_rx_subbeg)->subbeg; \
409 # define RX_OFFS(prog) \
411 const REGEXP *const _rx_offs = (prog); \
412 assert(SvTYPE(_rx_offs) == SVt_REGEXP); \
413 &SvANY(_rx_offs)->offs; \
415 # define RX_NPARENS(prog) \
417 const REGEXP *const _rx_nparens = (prog); \
418 assert(SvTYPE(_rx_nparens) == SVt_REGEXP); \
419 &SvANY(_rx_nparens)->nparens; \
422 # define RX_EXTFLAGS(prog) RXp_EXTFLAGS((struct regexp *)SvANY(prog))
423 # define RX_ENGINE(prog) (((struct regexp *)SvANY(prog))->engine)
424 # define RX_SUBBEG(prog) (((struct regexp *)SvANY(prog))->subbeg)
425 # define RX_OFFS(prog) (((struct regexp *)SvANY(prog))->offs)
426 # define RX_NPARENS(prog) (((struct regexp *)SvANY(prog))->nparens)
428 #define RX_SUBLEN(prog) (((struct regexp *)SvANY(prog))->sublen)
429 #define RX_MINLEN(prog) (((struct regexp *)SvANY(prog))->minlen)
430 #define RX_MINLENRET(prog) (((struct regexp *)SvANY(prog))->minlenret)
431 #define RX_GOFS(prog) (((struct regexp *)SvANY(prog))->gofs)
432 #define RX_LASTPAREN(prog) (((struct regexp *)SvANY(prog))->lastparen)
433 #define RX_LASTCLOSEPAREN(prog) (((struct regexp *)SvANY(prog))->lastcloseparen)
434 #define RX_SEEN_EVALS(prog) (((struct regexp *)SvANY(prog))->seen_evals)
435 #define RX_SAVED_COPY(prog) (((struct regexp *)SvANY(prog))->saved_copy)
437 #endif /* PLUGGABLE_RE_EXTENSION */
439 /* Stuff that needs to be included in the plugable extension goes below here */
441 #ifdef PERL_OLD_COPY_ON_WRITE
442 #define RX_MATCH_COPY_FREE(rx) \
443 STMT_START {if (RX_SAVED_COPY(rx)) { \
444 SV_CHECK_THINKFIRST_COW_DROP(RX_SAVED_COPY(rx)); \
446 if (RX_MATCH_COPIED(rx)) { \
447 Safefree(RX_SUBBEG(rx)); \
448 RX_MATCH_COPIED_off(rx); \
451 #define RX_MATCH_COPY_FREE(rx) \
452 STMT_START {if (RX_MATCH_COPIED(rx)) { \
453 Safefree(RX_SUBBEG(rx)); \
454 RX_MATCH_COPIED_off(rx); \
458 #define RXp_MATCH_UTF8(prog) (RXp_EXTFLAGS(prog) & RXf_MATCH_UTF8)
459 #define RX_MATCH_UTF8(prog) (RX_EXTFLAGS(prog) & RXf_MATCH_UTF8)
460 #define RX_MATCH_UTF8_on(prog) (RX_EXTFLAGS(prog) |= RXf_MATCH_UTF8)
461 #define RX_MATCH_UTF8_off(prog) (RX_EXTFLAGS(prog) &= ~RXf_MATCH_UTF8)
462 #define RX_MATCH_UTF8_set(prog, t) ((t) \
463 ? (RX_MATCH_UTF8_on(prog), (PL_reg_match_utf8 = 1)) \
464 : (RX_MATCH_UTF8_off(prog), (PL_reg_match_utf8 = 0)))
466 /* Whether the pattern stored at RX_WRAPPED is in UTF-8 */
467 #define RX_UTF8(prog) SvUTF8(prog)
469 #define REXEC_COPY_STR 0x01 /* Need to copy the string. */
470 #define REXEC_CHECKED 0x02 /* check_substr already checked. */
471 #define REXEC_SCREAM 0x04 /* use scream table. */
472 #define REXEC_IGNOREPOS 0x08 /* \G matches at start. */
473 #define REXEC_NOT_FIRST 0x10 /* This is another iteration of //g. */
475 #if defined(__GNUC__) && !defined(PERL_GCC_BRACE_GROUPS_FORBIDDEN)
476 # define ReREFCNT_inc(re) \
478 /* This is here to generate a casting warning if incorrect. */ \
479 REGEXP *const _rerefcnt_inc = (re); \
480 assert(SvTYPE(_rerefcnt_inc) == SVt_REGEXP); \
481 SvREFCNT_inc(_rerefcnt_inc); \
484 # define ReREFCNT_dec(re) \
486 /* This is here to generate a casting warning if incorrect. */ \
487 REGEXP *const _rerefcnt_dec = (re); \
488 SvREFCNT_dec(_rerefcnt_dec); \
491 # define ReREFCNT_dec(re) SvREFCNT_dec(re)
492 # define ReREFCNT_inc(re) ((REGEXP *) SvREFCNT_inc(re))
495 /* FIXME for plugins. */
497 #define FBMcf_TAIL_DOLLAR 1
498 #define FBMcf_TAIL_DOLLARM 2
499 #define FBMcf_TAIL_Z 4
500 #define FBMcf_TAIL_z 8
501 #define FBMcf_TAIL (FBMcf_TAIL_DOLLAR|FBMcf_TAIL_DOLLARM|FBMcf_TAIL_Z|FBMcf_TAIL_z)
503 #define FBMrf_MULTILINE 1
505 /* some basic information about the current match that is created by
506 * Perl_regexec_flags and then passed to regtry(), regmatch() etc */
518 /* structures for holding and saving the state maintained by regmatch() */
520 #ifndef MAX_RECURSE_EVAL_NOCHANGE_DEPTH
521 #define MAX_RECURSE_EVAL_NOCHANGE_DEPTH 1000
524 typedef I32 CHECKPOINT;
526 typedef struct regmatch_state {
527 int resume_state; /* where to jump to on return */
528 char *locinput; /* where to backtrack in string on failure */
532 /* this is a fake union member that matches the first element
533 * of each member that needs to store positive backtrack
536 struct regmatch_state *prev_yes_state;
539 /* branchlike members */
540 /* this is a fake union member that matches the first elements
541 * of each member that needs to behave like a branch */
543 /* this first element must match u.yes */
544 struct regmatch_state *prev_yes_state;
551 /* the first elements must match u.branchlike */
552 struct regmatch_state *prev_yes_state;
556 regnode *next_branch; /* next branch node */
560 /* the first elements must match u.branchlike */
561 struct regmatch_state *prev_yes_state;
565 U32 accepted; /* how many accepting states left */
566 U16 *jump; /* positive offsets from me */
567 regnode *B; /* node following the trie */
568 regnode *me; /* Which node am I - needed for jump tries*/
569 U8 *firstpos;/* pos in string of first trie match */
570 U32 firstchars;/* len in chars of firstpos from start */
571 U16 nextword;/* next word to try */
572 U16 topword; /* longest accepted word */
573 bool longfold;/* saw a fold with a 1->n char mapping */
576 /* special types - these members are used to store state for special
577 regops like eval, if/then, lookaround and the markpoint state */
579 /* this first element must match u.yes */
580 struct regmatch_state *prev_yes_state;
581 struct regmatch_state *prev_eval;
582 struct regmatch_state *prev_curlyx;
584 U32 toggle_reg_flags; /* what bits in PL_reg_flags to
585 flip when transitioning between
586 inner and outer rexen */
587 CHECKPOINT cp; /* remember current savestack indexes */
589 U32 close_paren; /* which close bracket is our end */
590 regnode *B; /* the node following us */
594 /* this first element must match u.yes */
595 struct regmatch_state *prev_yes_state;
597 I32 logical; /* saved copy of 'logical' var */
598 regnode *me; /* the IFMATCH/SUSPEND/UNLESSM node */
599 } ifmatch; /* and SUSPEND/UNLESSM */
602 /* this first element must match u.yes */
603 struct regmatch_state *prev_yes_state;
604 struct regmatch_state *prev_mark;
613 /* quantifiers - these members are used for storing state for
614 for the regops used to implement quantifiers */
616 /* this first element must match u.yes */
617 struct regmatch_state *prev_yes_state;
618 struct regmatch_state *prev_curlyx; /* previous cur_curlyx */
619 regnode *me; /* the CURLYX node */
620 regnode *B; /* the B node in /A*B/ */
621 CHECKPOINT cp; /* remember current savestack index */
623 int parenfloor;/* how far back to strip paren data */
625 /* these two are modified by WHILEM */
626 int count; /* how many instances of A we've matched */
627 char *lastloc;/* where previous A matched (0-len detect) */
631 /* this first element must match u.yes */
632 struct regmatch_state *prev_yes_state;
633 struct regmatch_state *save_curlyx;
634 CHECKPOINT cp; /* remember current savestack indexes */
636 char *save_lastloc; /* previous curlyx.lastloc */
642 /* this first element must match u.yes */
643 struct regmatch_state *prev_yes_state;
644 I32 c1, c2; /* case fold search */
646 I32 alen; /* length of first-matched A string */
649 regnode *A, *B; /* the nodes corresponding to /A*B/ */
650 regnode *me; /* the curlym node */
656 I32 c1, c2; /* case fold search */
657 char *maxpos; /* highest possible point in string to match */
658 char *oldloc; /* the previous locinput */
660 int min, max; /* {m,n} */
661 regnode *A, *B; /* the nodes corresponding to /A*B/ */
662 } curly; /* and CURLYN/PLUS/STAR */
667 /* how many regmatch_state structs to allocate as a single slab.
668 * We do it in 4K blocks for efficiency. The "3" is 2 for the next/prev
669 * pointers, plus 1 for any mythical malloc overhead. */
671 #define PERL_REGMATCH_SLAB_SLOTS \
672 ((4096 - 3 * sizeof (void*)) / sizeof(regmatch_state))
674 typedef struct regmatch_slab {
675 regmatch_state states[PERL_REGMATCH_SLAB_SLOTS];
676 struct regmatch_slab *prev, *next;
679 #define PL_reg_flags PL_reg_state.re_state_reg_flags
680 #define PL_bostr PL_reg_state.re_state_bostr
681 #define PL_reginput PL_reg_state.re_state_reginput
682 #define PL_regeol PL_reg_state.re_state_regeol
683 #define PL_regoffs PL_reg_state.re_state_regoffs
684 #define PL_reglastparen PL_reg_state.re_state_reglastparen
685 #define PL_reglastcloseparen PL_reg_state.re_state_reglastcloseparen
686 #define PL_reg_start_tmp PL_reg_state.re_state_reg_start_tmp
687 #define PL_reg_start_tmpl PL_reg_state.re_state_reg_start_tmpl
688 #define PL_reg_eval_set PL_reg_state.re_state_reg_eval_set
689 #define PL_reg_match_utf8 PL_reg_state.re_state_reg_match_utf8
690 #define PL_reg_magic PL_reg_state.re_state_reg_magic
691 #define PL_reg_oldpos PL_reg_state.re_state_reg_oldpos
692 #define PL_reg_oldcurpm PL_reg_state.re_state_reg_oldcurpm
693 #define PL_reg_curpm PL_reg_state.re_state_reg_curpm
694 #define PL_reg_oldsaved PL_reg_state.re_state_reg_oldsaved
695 #define PL_reg_oldsavedlen PL_reg_state.re_state_reg_oldsavedlen
696 #define PL_reg_maxiter PL_reg_state.re_state_reg_maxiter
697 #define PL_reg_leftiter PL_reg_state.re_state_reg_leftiter
698 #define PL_reg_poscache PL_reg_state.re_state_reg_poscache
699 #define PL_reg_poscache_size PL_reg_state.re_state_reg_poscache_size
700 #define PL_regsize PL_reg_state.re_state_regsize
701 #define PL_reg_starttry PL_reg_state.re_state_reg_starttry
702 #define PL_nrs PL_reg_state.re_state_nrs
704 struct re_save_state {
705 U32 re_state_reg_flags; /* from regexec.c */
706 U32 re_state_reg_start_tmpl; /* from regexec.c */
707 I32 re_state_reg_eval_set; /* from regexec.c */
708 bool re_state_reg_match_utf8; /* from regexec.c */
709 char *re_state_bostr;
710 char *re_state_reginput; /* String-input pointer. */
711 char *re_state_regeol; /* End of input, for $ check. */
712 regexp_paren_pair *re_state_regoffs; /* Pointer to start/end pairs */
713 U32 *re_state_reglastparen; /* Similarly for lastparen. */
714 U32 *re_state_reglastcloseparen; /* Similarly for lastcloseparen. */
715 char **re_state_reg_start_tmp; /* from regexec.c */
716 MAGIC *re_state_reg_magic; /* from regexec.c */
717 PMOP *re_state_reg_oldcurpm; /* from regexec.c */
718 PMOP *re_state_reg_curpm; /* from regexec.c */
719 char *re_state_reg_oldsaved; /* old saved substr during match */
720 STRLEN re_state_reg_oldsavedlen; /* old length of saved substr during match */
721 STRLEN re_state_reg_poscache_size; /* size of pos cache of WHILEM */
722 I32 re_state_reg_oldpos; /* from regexec.c */
723 I32 re_state_reg_maxiter; /* max wait until caching pos */
724 I32 re_state_reg_leftiter; /* wait until caching pos */
725 U32 re_state_regsize; /* from regexec.c */
726 char *re_state_reg_poscache; /* cache of pos of WHILEM */
727 char *re_state_reg_starttry; /* from regexec.c */
728 #ifdef PERL_OLD_COPY_ON_WRITE
729 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 */
733 #define SAVESTACK_ALLOC_FOR_RE_SAVE_STATE \
734 (1 + ((sizeof(struct re_save_state) - 1) / sizeof(*PL_savestack)))
738 * c-indentation-style: bsd
740 * indent-tabs-mode: t
743 * ex: set ts=8 sts=4 sw=4 noet: