5 * "A fair jaw-cracker dwarf-language must be." --Samwise Gamgee
8 /* This file contains functions for compiling a regular expression. See
9 * also regexec.c which funnily enough, contains functions for executing
10 * a regular expression.
12 * This file is also copied at build time to ext/re/re_comp.c, where
13 * it's built with -DPERL_EXT_RE_BUILD -DPERL_EXT_RE_DEBUG -DPERL_EXT.
14 * This causes the main functions to be compiled under new names and with
15 * debugging support added, which makes "use re 'debug'" work.
18 /* NOTE: this is derived from Henry Spencer's regexp code, and should not
19 * confused with the original package (see point 3 below). Thanks, Henry!
22 /* Additional note: this code is very heavily munged from Henry's version
23 * in places. In some spots I've traded clarity for efficiency, so don't
24 * blame Henry for some of the lack of readability.
27 /* The names of the functions have been changed from regcomp and
28 * regexec to pregcomp and pregexec in order to avoid conflicts
29 * with the POSIX routines of the same names.
32 #ifdef PERL_EXT_RE_BUILD
37 * pregcomp and pregexec -- regsub and regerror are not used in perl
39 * Copyright (c) 1986 by University of Toronto.
40 * Written by Henry Spencer. Not derived from licensed software.
42 * Permission is granted to anyone to use this software for any
43 * purpose on any computer system, and to redistribute it freely,
44 * subject to the following restrictions:
46 * 1. The author is not responsible for the consequences of use of
47 * this software, no matter how awful, even if they arise
50 * 2. The origin of this software must not be misrepresented, either
51 * by explicit claim or by omission.
53 * 3. Altered versions must be plainly marked as such, and must not
54 * be misrepresented as being the original software.
57 **** Alterations to Henry's code are...
59 **** Copyright (C) 1991, 1992, 1993, 1994, 1995, 1996, 1997, 1998, 1999,
60 **** 2000, 2001, 2002, 2003, 2004, 2005, 2006, by Larry Wall and others
62 **** You may distribute under the terms of either the GNU General Public
63 **** License or the Artistic License, as specified in the README file.
66 * Beware that some of this code is subtly aware of the way operator
67 * precedence is structured in regular expressions. Serious changes in
68 * regular-expression syntax might require a total rethink.
71 #define PERL_IN_REGCOMP_C
74 #ifndef PERL_IN_XSUB_RE
79 #ifdef PERL_IN_XSUB_RE
90 # if defined(BUGGY_MSC6)
91 /* MSC 6.00A breaks on op/regexp.t test 85 unless we turn this off */
92 # pragma optimize("a",off)
93 /* But MSC 6.00A is happy with 'w', for aliases only across function calls*/
94 # pragma optimize("w",on )
95 # endif /* BUGGY_MSC6 */
102 typedef struct RExC_state_t {
103 U32 flags; /* are we folding, multilining? */
104 char *precomp; /* uncompiled string. */
106 char *start; /* Start of input for compile */
107 char *end; /* End of input for compile */
108 char *parse; /* Input-scan pointer. */
109 I32 whilem_seen; /* number of WHILEM in this expr */
110 regnode *emit_start; /* Start of emitted-code area */
111 regnode *emit; /* Code-emit pointer; ®dummy = don't = compiling */
112 I32 naughty; /* How bad is this pattern? */
113 I32 sawback; /* Did we see \1, ...? */
115 I32 size; /* Code size. */
116 I32 npar; /* () count. */
121 HV *charnames; /* cache of named sequences */
123 char *starttry; /* -Dr: where regtry was called. */
124 #define RExC_starttry (pRExC_state->starttry)
127 const char *lastparse;
129 #define RExC_lastparse (pRExC_state->lastparse)
130 #define RExC_lastnum (pRExC_state->lastnum)
134 #define RExC_flags (pRExC_state->flags)
135 #define RExC_precomp (pRExC_state->precomp)
136 #define RExC_rx (pRExC_state->rx)
137 #define RExC_start (pRExC_state->start)
138 #define RExC_end (pRExC_state->end)
139 #define RExC_parse (pRExC_state->parse)
140 #define RExC_whilem_seen (pRExC_state->whilem_seen)
141 #define RExC_offsets (pRExC_state->rx->offsets) /* I am not like the others */
142 #define RExC_emit (pRExC_state->emit)
143 #define RExC_emit_start (pRExC_state->emit_start)
144 #define RExC_naughty (pRExC_state->naughty)
145 #define RExC_sawback (pRExC_state->sawback)
146 #define RExC_seen (pRExC_state->seen)
147 #define RExC_size (pRExC_state->size)
148 #define RExC_npar (pRExC_state->npar)
149 #define RExC_extralen (pRExC_state->extralen)
150 #define RExC_seen_zerolen (pRExC_state->seen_zerolen)
151 #define RExC_seen_evals (pRExC_state->seen_evals)
152 #define RExC_utf8 (pRExC_state->utf8)
153 #define RExC_charnames (pRExC_state->charnames)
155 #define ISMULT1(c) ((c) == '*' || (c) == '+' || (c) == '?')
156 #define ISMULT2(s) ((*s) == '*' || (*s) == '+' || (*s) == '?' || \
157 ((*s) == '{' && regcurly(s)))
160 #undef SPSTART /* dratted cpp namespace... */
163 * Flags to be passed up and down.
165 #define WORST 0 /* Worst case. */
166 #define HASWIDTH 0x1 /* Known to match non-null strings. */
167 #define SIMPLE 0x2 /* Simple enough to be STAR/PLUS operand. */
168 #define SPSTART 0x4 /* Starts with * or +. */
169 #define TRYAGAIN 0x8 /* Weeded out a declaration. */
171 #define REG_NODE_NUM(x) ((x) ? (int)((x)-RExC_emit_start) : -1)
173 /* whether trie related optimizations are enabled */
174 #if PERL_ENABLE_EXTENDED_TRIE_OPTIMISATION
175 #define TRIE_STUDY_OPT
176 #define FULL_TRIE_STUDY
181 /* About scan_data_t.
183 During optimisation we recurse through the regexp program performing
184 various inplace (keyhole style) optimisations. In addition study_chunk
185 and scan_commit populate this data structure with information about
186 what strings MUST appear in the pattern. We look for the longest
187 string that must appear for at a fixed location, and we look for the
188 longest string that may appear at a floating location. So for instance
193 Both 'FOO' and 'A' are fixed strings. Both 'B' and 'BAR' are floating
194 strings (because they follow a .* construct). study_chunk will identify
195 both FOO and BAR as being the longest fixed and floating strings respectively.
197 The strings can be composites, for instance
201 will result in a composite fixed substring 'foo'.
203 For each string some basic information is maintained:
205 - offset or min_offset
206 This is the position the string must appear at, or not before.
207 It also implicitly (when combined with minlenp) tells us how many
208 character must match before the string we are searching.
209 Likewise when combined with minlenp and the length of the string
210 tells us how many characters must appear after the string we have
214 Only used for floating strings. This is the rightmost point that
215 the string can appear at. Ifset to I32 max it indicates that the
216 string can occur infinitely far to the right.
219 A pointer to the minimum length of the pattern that the string
220 was found inside. This is important as in the case of positive
221 lookahead or positive lookbehind we can have multiple patterns
226 The minimum length of the pattern overall is 3, the minimum length
227 of the lookahead part is 3, but the minimum length of the part that
228 will actually match is 1. So 'FOO's minimum length is 3, but the
229 minimum length for the F is 1. This is important as the minimum length
230 is used to determine offsets in front of and behind the string being
231 looked for. Since strings can be composites this is the length of the
232 pattern at the time it was commited with a scan_commit. Note that
233 the length is calculated by study_chunk, so that the minimum lengths
234 are not known until the full pattern has been compiled, thus the
235 pointer to the value.
239 In the case of lookbehind the string being searched for can be
240 offset past the start point of the final matching string.
241 If this value was just blithely removed from the min_offset it would
242 invalidate some of the calculations for how many chars must match
243 before or after (as they are derived from min_offset and minlen and
244 the length of the string being searched for).
245 When the final pattern is compiled and the data is moved from the
246 scan_data_t structure into the regexp structure the information
247 about lookbehind is factored in, with the information that would
248 have been lost precalculated in the end_shift field for the
251 The fields pos_min and pos_delta are used to store the minimum offset
252 and the delta to the maximum offset at the current point in the pattern.
256 typedef struct scan_data_t {
257 /*I32 len_min; unused */
258 /*I32 len_delta; unused */
262 I32 last_end; /* min value, <0 unless valid. */
265 SV **longest; /* Either &l_fixed, or &l_float. */
266 SV *longest_fixed; /* longest fixed string found in pattern */
267 I32 offset_fixed; /* offset where it starts */
268 I32 *minlen_fixed; /* pointer to the minlen relevent to the string */
269 I32 lookbehind_fixed; /* is the position of the string modfied by LB */
270 SV *longest_float; /* longest floating string found in pattern */
271 I32 offset_float_min; /* earliest point in string it can appear */
272 I32 offset_float_max; /* latest point in string it can appear */
273 I32 *minlen_float; /* pointer to the minlen relevent to the string */
274 I32 lookbehind_float; /* is the position of the string modified by LB */
278 struct regnode_charclass_class *start_class;
282 * Forward declarations for pregcomp()'s friends.
285 static const scan_data_t zero_scan_data =
286 { 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0 ,0};
288 #define SF_BEFORE_EOL (SF_BEFORE_SEOL|SF_BEFORE_MEOL)
289 #define SF_BEFORE_SEOL 0x0001
290 #define SF_BEFORE_MEOL 0x0002
291 #define SF_FIX_BEFORE_EOL (SF_FIX_BEFORE_SEOL|SF_FIX_BEFORE_MEOL)
292 #define SF_FL_BEFORE_EOL (SF_FL_BEFORE_SEOL|SF_FL_BEFORE_MEOL)
295 # define SF_FIX_SHIFT_EOL (0+2)
296 # define SF_FL_SHIFT_EOL (0+4)
298 # define SF_FIX_SHIFT_EOL (+2)
299 # define SF_FL_SHIFT_EOL (+4)
302 #define SF_FIX_BEFORE_SEOL (SF_BEFORE_SEOL << SF_FIX_SHIFT_EOL)
303 #define SF_FIX_BEFORE_MEOL (SF_BEFORE_MEOL << SF_FIX_SHIFT_EOL)
305 #define SF_FL_BEFORE_SEOL (SF_BEFORE_SEOL << SF_FL_SHIFT_EOL)
306 #define SF_FL_BEFORE_MEOL (SF_BEFORE_MEOL << SF_FL_SHIFT_EOL) /* 0x20 */
307 #define SF_IS_INF 0x0040
308 #define SF_HAS_PAR 0x0080
309 #define SF_IN_PAR 0x0100
310 #define SF_HAS_EVAL 0x0200
311 #define SCF_DO_SUBSTR 0x0400
312 #define SCF_DO_STCLASS_AND 0x0800
313 #define SCF_DO_STCLASS_OR 0x1000
314 #define SCF_DO_STCLASS (SCF_DO_STCLASS_AND|SCF_DO_STCLASS_OR)
315 #define SCF_WHILEM_VISITED_POS 0x2000
317 #define SCF_TRIE_RESTUDY 0x4000 /* Do restudy? */
320 #define UTF (RExC_utf8 != 0)
321 #define LOC ((RExC_flags & PMf_LOCALE) != 0)
322 #define FOLD ((RExC_flags & PMf_FOLD) != 0)
324 #define OOB_UNICODE 12345678
325 #define OOB_NAMEDCLASS -1
327 #define CHR_SVLEN(sv) (UTF ? sv_len_utf8(sv) : SvCUR(sv))
328 #define CHR_DIST(a,b) (UTF ? utf8_distance(a,b) : a - b)
331 /* length of regex to show in messages that don't mark a position within */
332 #define RegexLengthToShowInErrorMessages 127
335 * If MARKER[12] are adjusted, be sure to adjust the constants at the top
336 * of t/op/regmesg.t, the tests in t/op/re_tests, and those in
337 * op/pragma/warn/regcomp.
339 #define MARKER1 "<-- HERE" /* marker as it appears in the description */
340 #define MARKER2 " <-- HERE " /* marker as it appears within the regex */
342 #define REPORT_LOCATION " in regex; marked by " MARKER1 " in m/%.*s" MARKER2 "%s/"
345 * Calls SAVEDESTRUCTOR_X if needed, then calls Perl_croak with the given
346 * arg. Show regex, up to a maximum length. If it's too long, chop and add
349 #define FAIL(msg) STMT_START { \
350 const char *ellipses = ""; \
351 IV len = RExC_end - RExC_precomp; \
354 SAVEDESTRUCTOR_X(clear_re,(void*)RExC_rx); \
355 if (len > RegexLengthToShowInErrorMessages) { \
356 /* chop 10 shorter than the max, to ensure meaning of "..." */ \
357 len = RegexLengthToShowInErrorMessages - 10; \
360 Perl_croak(aTHX_ "%s in regex m/%.*s%s/", \
361 msg, (int)len, RExC_precomp, ellipses); \
365 * Simple_vFAIL -- like FAIL, but marks the current location in the scan
367 #define Simple_vFAIL(m) STMT_START { \
368 const IV offset = RExC_parse - RExC_precomp; \
369 Perl_croak(aTHX_ "%s" REPORT_LOCATION, \
370 m, (int)offset, RExC_precomp, RExC_precomp + offset); \
374 * Calls SAVEDESTRUCTOR_X if needed, then Simple_vFAIL()
376 #define vFAIL(m) STMT_START { \
378 SAVEDESTRUCTOR_X(clear_re,(void*)RExC_rx); \
383 * Like Simple_vFAIL(), but accepts two arguments.
385 #define Simple_vFAIL2(m,a1) STMT_START { \
386 const IV offset = RExC_parse - RExC_precomp; \
387 S_re_croak2(aTHX_ m, REPORT_LOCATION, a1, \
388 (int)offset, RExC_precomp, RExC_precomp + offset); \
392 * Calls SAVEDESTRUCTOR_X if needed, then Simple_vFAIL2().
394 #define vFAIL2(m,a1) STMT_START { \
396 SAVEDESTRUCTOR_X(clear_re,(void*)RExC_rx); \
397 Simple_vFAIL2(m, a1); \
402 * Like Simple_vFAIL(), but accepts three arguments.
404 #define Simple_vFAIL3(m, a1, a2) STMT_START { \
405 const IV offset = RExC_parse - RExC_precomp; \
406 S_re_croak2(aTHX_ m, REPORT_LOCATION, a1, a2, \
407 (int)offset, RExC_precomp, RExC_precomp + offset); \
411 * Calls SAVEDESTRUCTOR_X if needed, then Simple_vFAIL3().
413 #define vFAIL3(m,a1,a2) STMT_START { \
415 SAVEDESTRUCTOR_X(clear_re,(void*)RExC_rx); \
416 Simple_vFAIL3(m, a1, a2); \
420 * Like Simple_vFAIL(), but accepts four arguments.
422 #define Simple_vFAIL4(m, a1, a2, a3) STMT_START { \
423 const IV offset = RExC_parse - RExC_precomp; \
424 S_re_croak2(aTHX_ m, REPORT_LOCATION, a1, a2, a3, \
425 (int)offset, RExC_precomp, RExC_precomp + offset); \
428 #define vWARN(loc,m) STMT_START { \
429 const IV offset = loc - RExC_precomp; \
430 Perl_warner(aTHX_ packWARN(WARN_REGEXP), "%s" REPORT_LOCATION, \
431 m, (int)offset, RExC_precomp, RExC_precomp + offset); \
434 #define vWARNdep(loc,m) STMT_START { \
435 const IV offset = loc - RExC_precomp; \
436 Perl_warner(aTHX_ packWARN2(WARN_DEPRECATED, WARN_REGEXP), \
437 "%s" REPORT_LOCATION, \
438 m, (int)offset, RExC_precomp, RExC_precomp + offset); \
442 #define vWARN2(loc, m, a1) STMT_START { \
443 const IV offset = loc - RExC_precomp; \
444 Perl_warner(aTHX_ packWARN(WARN_REGEXP), m REPORT_LOCATION, \
445 a1, (int)offset, RExC_precomp, RExC_precomp + offset); \
448 #define vWARN3(loc, m, a1, a2) STMT_START { \
449 const IV offset = loc - RExC_precomp; \
450 Perl_warner(aTHX_ packWARN(WARN_REGEXP), m REPORT_LOCATION, \
451 a1, a2, (int)offset, RExC_precomp, RExC_precomp + offset); \
454 #define vWARN4(loc, m, a1, a2, a3) STMT_START { \
455 const IV offset = loc - RExC_precomp; \
456 Perl_warner(aTHX_ packWARN(WARN_REGEXP), m REPORT_LOCATION, \
457 a1, a2, a3, (int)offset, RExC_precomp, RExC_precomp + offset); \
460 #define vWARN5(loc, m, a1, a2, a3, a4) STMT_START { \
461 const IV offset = loc - RExC_precomp; \
462 Perl_warner(aTHX_ packWARN(WARN_REGEXP), m REPORT_LOCATION, \
463 a1, a2, a3, a4, (int)offset, RExC_precomp, RExC_precomp + offset); \
467 /* Allow for side effects in s */
468 #define REGC(c,s) STMT_START { \
469 if (!SIZE_ONLY) *(s) = (c); else (void)(s); \
472 /* Macros for recording node offsets. 20001227 mjd@plover.com
473 * Nodes are numbered 1, 2, 3, 4. Node #n's position is recorded in
474 * element 2*n-1 of the array. Element #2n holds the byte length node #n.
475 * Element 0 holds the number n.
476 * Position is 1 indexed.
479 #define Set_Node_Offset_To_R(node,byte) STMT_START { \
481 MJD_OFFSET_DEBUG(("** (%d) offset of node %d is %d.\n", \
482 __LINE__, (node), (int)(byte))); \
484 Perl_croak(aTHX_ "value of node is %d in Offset macro", (int)(node)); \
486 RExC_offsets[2*(node)-1] = (byte); \
491 #define Set_Node_Offset(node,byte) \
492 Set_Node_Offset_To_R((node)-RExC_emit_start, (byte)-RExC_start)
493 #define Set_Cur_Node_Offset Set_Node_Offset(RExC_emit, RExC_parse)
495 #define Set_Node_Length_To_R(node,len) STMT_START { \
497 MJD_OFFSET_DEBUG(("** (%d) size of node %d is %d.\n", \
498 __LINE__, (int)(node), (int)(len))); \
500 Perl_croak(aTHX_ "value of node is %d in Length macro", (int)(node)); \
502 RExC_offsets[2*(node)] = (len); \
507 #define Set_Node_Length(node,len) \
508 Set_Node_Length_To_R((node)-RExC_emit_start, len)
509 #define Set_Cur_Node_Length(len) Set_Node_Length(RExC_emit, len)
510 #define Set_Node_Cur_Length(node) \
511 Set_Node_Length(node, RExC_parse - parse_start)
513 /* Get offsets and lengths */
514 #define Node_Offset(n) (RExC_offsets[2*((n)-RExC_emit_start)-1])
515 #define Node_Length(n) (RExC_offsets[2*((n)-RExC_emit_start)])
517 #define Set_Node_Offset_Length(node,offset,len) STMT_START { \
518 Set_Node_Offset_To_R((node)-RExC_emit_start, (offset)); \
519 Set_Node_Length_To_R((node)-RExC_emit_start, (len)); \
523 #if PERL_ENABLE_EXPERIMENTAL_REGEX_OPTIMISATIONS
524 #define EXPERIMENTAL_INPLACESCAN
527 #define DEBUG_STUDYDATA(data,depth) \
528 DEBUG_OPTIMISE_MORE_r(if(data){ \
529 PerlIO_printf(Perl_debug_log, \
530 "%*s"/* Len:%"IVdf"/%"IVdf" */" Pos:%"IVdf"/%"IVdf \
531 " Flags: %"IVdf" Whilem_c: %"IVdf" Lcp: %"IVdf" ", \
532 (int)(depth)*2, "", \
533 (IV)((data)->pos_min), \
534 (IV)((data)->pos_delta), \
535 (IV)((data)->flags), \
536 (IV)((data)->whilem_c), \
537 (IV)((data)->last_closep ? *((data)->last_closep) : -1) \
539 if ((data)->last_found) \
540 PerlIO_printf(Perl_debug_log, \
541 "Last:'%s' %"IVdf":%"IVdf"/%"IVdf" %sFixed:'%s' @ %"IVdf \
542 " %sFloat: '%s' @ %"IVdf"/%"IVdf"", \
543 SvPVX_const((data)->last_found), \
544 (IV)((data)->last_end), \
545 (IV)((data)->last_start_min), \
546 (IV)((data)->last_start_max), \
547 ((data)->longest && \
548 (data)->longest==&((data)->longest_fixed)) ? "*" : "", \
549 SvPVX_const((data)->longest_fixed), \
550 (IV)((data)->offset_fixed), \
551 ((data)->longest && \
552 (data)->longest==&((data)->longest_float)) ? "*" : "", \
553 SvPVX_const((data)->longest_float), \
554 (IV)((data)->offset_float_min), \
555 (IV)((data)->offset_float_max) \
557 PerlIO_printf(Perl_debug_log,"\n"); \
560 static void clear_re(pTHX_ void *r);
562 /* Mark that we cannot extend a found fixed substring at this point.
563 Update the longest found anchored substring and the longest found
564 floating substrings if needed. */
567 S_scan_commit(pTHX_ const RExC_state_t *pRExC_state, scan_data_t *data, I32 *minlenp)
569 const STRLEN l = CHR_SVLEN(data->last_found);
570 const STRLEN old_l = CHR_SVLEN(*data->longest);
571 GET_RE_DEBUG_FLAGS_DECL;
573 if ((l >= old_l) && ((l > old_l) || (data->flags & SF_BEFORE_EOL))) {
574 SvSetMagicSV(*data->longest, data->last_found);
575 if (*data->longest == data->longest_fixed) {
576 data->offset_fixed = l ? data->last_start_min : data->pos_min;
577 if (data->flags & SF_BEFORE_EOL)
579 |= ((data->flags & SF_BEFORE_EOL) << SF_FIX_SHIFT_EOL);
581 data->flags &= ~SF_FIX_BEFORE_EOL;
582 data->minlen_fixed=minlenp;
583 data->lookbehind_fixed=0;
586 data->offset_float_min = l ? data->last_start_min : data->pos_min;
587 data->offset_float_max = (l
588 ? data->last_start_max
589 : data->pos_min + data->pos_delta);
590 if ((U32)data->offset_float_max > (U32)I32_MAX)
591 data->offset_float_max = I32_MAX;
592 if (data->flags & SF_BEFORE_EOL)
594 |= ((data->flags & SF_BEFORE_EOL) << SF_FL_SHIFT_EOL);
596 data->flags &= ~SF_FL_BEFORE_EOL;
597 data->minlen_float=minlenp;
598 data->lookbehind_float=0;
601 SvCUR_set(data->last_found, 0);
603 SV * const sv = data->last_found;
604 if (SvUTF8(sv) && SvMAGICAL(sv)) {
605 MAGIC * const mg = mg_find(sv, PERL_MAGIC_utf8);
611 data->flags &= ~SF_BEFORE_EOL;
612 DEBUG_STUDYDATA(data,0);
615 /* Can match anything (initialization) */
617 S_cl_anything(const RExC_state_t *pRExC_state, struct regnode_charclass_class *cl)
619 ANYOF_CLASS_ZERO(cl);
620 ANYOF_BITMAP_SETALL(cl);
621 cl->flags = ANYOF_EOS|ANYOF_UNICODE_ALL;
623 cl->flags |= ANYOF_LOCALE;
626 /* Can match anything (initialization) */
628 S_cl_is_anything(const struct regnode_charclass_class *cl)
632 for (value = 0; value <= ANYOF_MAX; value += 2)
633 if (ANYOF_CLASS_TEST(cl, value) && ANYOF_CLASS_TEST(cl, value + 1))
635 if (!(cl->flags & ANYOF_UNICODE_ALL))
637 if (!ANYOF_BITMAP_TESTALLSET((const void*)cl))
642 /* Can match anything (initialization) */
644 S_cl_init(const RExC_state_t *pRExC_state, struct regnode_charclass_class *cl)
646 Zero(cl, 1, struct regnode_charclass_class);
648 cl_anything(pRExC_state, cl);
652 S_cl_init_zero(const RExC_state_t *pRExC_state, struct regnode_charclass_class *cl)
654 Zero(cl, 1, struct regnode_charclass_class);
656 cl_anything(pRExC_state, cl);
658 cl->flags |= ANYOF_LOCALE;
661 /* 'And' a given class with another one. Can create false positives */
662 /* We assume that cl is not inverted */
664 S_cl_and(struct regnode_charclass_class *cl,
665 const struct regnode_charclass_class *and_with)
667 if (!(and_with->flags & ANYOF_CLASS)
668 && !(cl->flags & ANYOF_CLASS)
669 && (and_with->flags & ANYOF_LOCALE) == (cl->flags & ANYOF_LOCALE)
670 && !(and_with->flags & ANYOF_FOLD)
671 && !(cl->flags & ANYOF_FOLD)) {
674 if (and_with->flags & ANYOF_INVERT)
675 for (i = 0; i < ANYOF_BITMAP_SIZE; i++)
676 cl->bitmap[i] &= ~and_with->bitmap[i];
678 for (i = 0; i < ANYOF_BITMAP_SIZE; i++)
679 cl->bitmap[i] &= and_with->bitmap[i];
680 } /* XXXX: logic is complicated otherwise, leave it along for a moment. */
681 if (!(and_with->flags & ANYOF_EOS))
682 cl->flags &= ~ANYOF_EOS;
684 if (cl->flags & ANYOF_UNICODE_ALL && and_with->flags & ANYOF_UNICODE &&
685 !(and_with->flags & ANYOF_INVERT)) {
686 cl->flags &= ~ANYOF_UNICODE_ALL;
687 cl->flags |= ANYOF_UNICODE;
688 ARG_SET(cl, ARG(and_with));
690 if (!(and_with->flags & ANYOF_UNICODE_ALL) &&
691 !(and_with->flags & ANYOF_INVERT))
692 cl->flags &= ~ANYOF_UNICODE_ALL;
693 if (!(and_with->flags & (ANYOF_UNICODE|ANYOF_UNICODE_ALL)) &&
694 !(and_with->flags & ANYOF_INVERT))
695 cl->flags &= ~ANYOF_UNICODE;
698 /* 'OR' a given class with another one. Can create false positives */
699 /* We assume that cl is not inverted */
701 S_cl_or(const RExC_state_t *pRExC_state, struct regnode_charclass_class *cl, const struct regnode_charclass_class *or_with)
703 if (or_with->flags & ANYOF_INVERT) {
705 * (B1 | CL1) | (!B2 & !CL2) = (B1 | !B2 & !CL2) | (CL1 | (!B2 & !CL2))
706 * <= (B1 | !B2) | (CL1 | !CL2)
707 * which is wasteful if CL2 is small, but we ignore CL2:
708 * (B1 | CL1) | (!B2 & !CL2) <= (B1 | CL1) | !B2 = (B1 | !B2) | CL1
709 * XXXX Can we handle case-fold? Unclear:
710 * (OK1(i) | OK1(i')) | !(OK1(i) | OK1(i')) =
711 * (OK1(i) | OK1(i')) | (!OK1(i) & !OK1(i'))
713 if ( (or_with->flags & ANYOF_LOCALE) == (cl->flags & ANYOF_LOCALE)
714 && !(or_with->flags & ANYOF_FOLD)
715 && !(cl->flags & ANYOF_FOLD) ) {
718 for (i = 0; i < ANYOF_BITMAP_SIZE; i++)
719 cl->bitmap[i] |= ~or_with->bitmap[i];
720 } /* XXXX: logic is complicated otherwise */
722 cl_anything(pRExC_state, cl);
725 /* (B1 | CL1) | (B2 | CL2) = (B1 | B2) | (CL1 | CL2)) */
726 if ( (or_with->flags & ANYOF_LOCALE) == (cl->flags & ANYOF_LOCALE)
727 && (!(or_with->flags & ANYOF_FOLD)
728 || (cl->flags & ANYOF_FOLD)) ) {
731 /* OR char bitmap and class bitmap separately */
732 for (i = 0; i < ANYOF_BITMAP_SIZE; i++)
733 cl->bitmap[i] |= or_with->bitmap[i];
734 if (or_with->flags & ANYOF_CLASS) {
735 for (i = 0; i < ANYOF_CLASSBITMAP_SIZE; i++)
736 cl->classflags[i] |= or_with->classflags[i];
737 cl->flags |= ANYOF_CLASS;
740 else { /* XXXX: logic is complicated, leave it along for a moment. */
741 cl_anything(pRExC_state, cl);
744 if (or_with->flags & ANYOF_EOS)
745 cl->flags |= ANYOF_EOS;
747 if (cl->flags & ANYOF_UNICODE && or_with->flags & ANYOF_UNICODE &&
748 ARG(cl) != ARG(or_with)) {
749 cl->flags |= ANYOF_UNICODE_ALL;
750 cl->flags &= ~ANYOF_UNICODE;
752 if (or_with->flags & ANYOF_UNICODE_ALL) {
753 cl->flags |= ANYOF_UNICODE_ALL;
754 cl->flags &= ~ANYOF_UNICODE;
758 #define TRIE_LIST_ITEM(state,idx) (trie->states[state].trans.list)[ idx ]
759 #define TRIE_LIST_CUR(state) ( TRIE_LIST_ITEM( state, 0 ).forid )
760 #define TRIE_LIST_LEN(state) ( TRIE_LIST_ITEM( state, 0 ).newstate )
761 #define TRIE_LIST_USED(idx) ( trie->states[state].trans.list ? (TRIE_LIST_CUR( idx ) - 1) : 0 )
767 dump_trie_interim_list(trie,next_alloc)
768 dump_trie_interim_table(trie,next_alloc)
770 These routines dump out a trie in a somewhat readable format.
771 The _interim_ variants are used for debugging the interim
772 tables that are used to generate the final compressed
773 representation which is what dump_trie expects.
775 Part of the reason for their existance is to provide a form
776 of documentation as to how the different representations function.
782 Dumps the final compressed table form of the trie to Perl_debug_log.
783 Used for debugging make_trie().
787 S_dump_trie(pTHX_ const struct _reg_trie_data *trie,U32 depth)
790 SV *sv=sv_newmortal();
791 int colwidth= trie->widecharmap ? 6 : 4;
792 GET_RE_DEBUG_FLAGS_DECL;
795 PerlIO_printf( Perl_debug_log, "%*sChar : %-6s%-6s%-4s ",
796 (int)depth * 2 + 2,"",
797 "Match","Base","Ofs" );
799 for( state = 0 ; state < trie->uniquecharcount ; state++ ) {
800 SV ** const tmp = av_fetch( trie->revcharmap, state, 0);
802 PerlIO_printf( Perl_debug_log, "%*s",
804 pv_pretty(sv, SvPV_nolen_const(*tmp), SvCUR(*tmp), colwidth,
805 PL_colors[0], PL_colors[1],
806 (SvUTF8(*tmp) ? PERL_PV_ESCAPE_UNI : 0) |
807 PERL_PV_ESCAPE_FIRSTCHAR
812 PerlIO_printf( Perl_debug_log, "\n%*sState|-----------------------",
813 (int)depth * 2 + 2,"");
815 for( state = 0 ; state < trie->uniquecharcount ; state++ )
816 PerlIO_printf( Perl_debug_log, "%.*s", colwidth, "--------");
817 PerlIO_printf( Perl_debug_log, "\n");
819 for( state = 1 ; state < trie->laststate ; state++ ) {
820 const U32 base = trie->states[ state ].trans.base;
822 PerlIO_printf( Perl_debug_log, "%*s#%4"UVXf"|", (int)depth * 2 + 2,"", (UV)state);
824 if ( trie->states[ state ].wordnum ) {
825 PerlIO_printf( Perl_debug_log, " W%4X", trie->states[ state ].wordnum );
827 PerlIO_printf( Perl_debug_log, "%6s", "" );
830 PerlIO_printf( Perl_debug_log, " @%4"UVXf" ", (UV)base );
835 while( ( base + ofs < trie->uniquecharcount ) ||
836 ( base + ofs - trie->uniquecharcount < trie->lasttrans
837 && trie->trans[ base + ofs - trie->uniquecharcount ].check != state))
840 PerlIO_printf( Perl_debug_log, "+%2"UVXf"[ ", (UV)ofs);
842 for ( ofs = 0 ; ofs < trie->uniquecharcount ; ofs++ ) {
843 if ( ( base + ofs >= trie->uniquecharcount ) &&
844 ( base + ofs - trie->uniquecharcount < trie->lasttrans ) &&
845 trie->trans[ base + ofs - trie->uniquecharcount ].check == state )
847 PerlIO_printf( Perl_debug_log, "%*"UVXf,
849 (UV)trie->trans[ base + ofs - trie->uniquecharcount ].next );
851 PerlIO_printf( Perl_debug_log, "%*s",colwidth," ." );
855 PerlIO_printf( Perl_debug_log, "]");
858 PerlIO_printf( Perl_debug_log, "\n" );
862 dump_trie_interim_list(trie,next_alloc)
863 Dumps a fully constructed but uncompressed trie in list form.
864 List tries normally only are used for construction when the number of
865 possible chars (trie->uniquecharcount) is very high.
866 Used for debugging make_trie().
869 S_dump_trie_interim_list(pTHX_ const struct _reg_trie_data *trie, U32 next_alloc,U32 depth)
872 SV *sv=sv_newmortal();
873 int colwidth= trie->widecharmap ? 6 : 4;
874 GET_RE_DEBUG_FLAGS_DECL;
875 /* print out the table precompression. */
876 PerlIO_printf( Perl_debug_log, "%*sState :Word | Transition Data\n%*s%s",
877 (int)depth * 2 + 2,"", (int)depth * 2 + 2,"",
878 "------:-----+-----------------\n" );
880 for( state=1 ; state < next_alloc ; state ++ ) {
883 PerlIO_printf( Perl_debug_log, "%*s %4"UVXf" :",
884 (int)depth * 2 + 2,"", (UV)state );
885 if ( ! trie->states[ state ].wordnum ) {
886 PerlIO_printf( Perl_debug_log, "%5s| ","");
888 PerlIO_printf( Perl_debug_log, "W%4x| ",
889 trie->states[ state ].wordnum
892 for( charid = 1 ; charid <= TRIE_LIST_USED( state ) ; charid++ ) {
893 SV ** const tmp = av_fetch( trie->revcharmap, TRIE_LIST_ITEM(state,charid).forid, 0);
895 PerlIO_printf( Perl_debug_log, "%*s:%3X=%4"UVXf" | ",
897 pv_pretty(sv, SvPV_nolen_const(*tmp), SvCUR(*tmp), colwidth,
898 PL_colors[0], PL_colors[1],
899 (SvUTF8(*tmp) ? PERL_PV_ESCAPE_UNI : 0) |
900 PERL_PV_ESCAPE_FIRSTCHAR
902 TRIE_LIST_ITEM(state,charid).forid,
903 (UV)TRIE_LIST_ITEM(state,charid).newstate
907 PerlIO_printf( Perl_debug_log, "\n");
912 dump_trie_interim_table(trie,next_alloc)
913 Dumps a fully constructed but uncompressed trie in table form.
914 This is the normal DFA style state transition table, with a few
915 twists to facilitate compression later.
916 Used for debugging make_trie().
919 S_dump_trie_interim_table(pTHX_ const struct _reg_trie_data *trie, U32 next_alloc, U32 depth)
923 SV *sv=sv_newmortal();
924 int colwidth= trie->widecharmap ? 6 : 4;
925 GET_RE_DEBUG_FLAGS_DECL;
928 print out the table precompression so that we can do a visual check
929 that they are identical.
932 PerlIO_printf( Perl_debug_log, "%*sChar : ",(int)depth * 2 + 2,"" );
934 for( charid = 0 ; charid < trie->uniquecharcount ; charid++ ) {
935 SV ** const tmp = av_fetch( trie->revcharmap, charid, 0);
937 PerlIO_printf( Perl_debug_log, "%*s",
939 pv_pretty(sv, SvPV_nolen_const(*tmp), SvCUR(*tmp), colwidth,
940 PL_colors[0], PL_colors[1],
941 (SvUTF8(*tmp) ? PERL_PV_ESCAPE_UNI : 0) |
942 PERL_PV_ESCAPE_FIRSTCHAR
948 PerlIO_printf( Perl_debug_log, "\n%*sState+-",(int)depth * 2 + 2,"" );
950 for( charid=0 ; charid < trie->uniquecharcount ; charid++ ) {
951 PerlIO_printf( Perl_debug_log, "%.*s", colwidth,"--------");
954 PerlIO_printf( Perl_debug_log, "\n" );
956 for( state=1 ; state < next_alloc ; state += trie->uniquecharcount ) {
958 PerlIO_printf( Perl_debug_log, "%*s%4"UVXf" : ",
959 (int)depth * 2 + 2,"",
960 (UV)TRIE_NODENUM( state ) );
962 for( charid = 0 ; charid < trie->uniquecharcount ; charid++ ) {
963 UV v=(UV)SAFE_TRIE_NODENUM( trie->trans[ state + charid ].next );
965 PerlIO_printf( Perl_debug_log, "%*"UVXf, colwidth, v );
967 PerlIO_printf( Perl_debug_log, "%*s", colwidth, "." );
969 if ( ! trie->states[ TRIE_NODENUM( state ) ].wordnum ) {
970 PerlIO_printf( Perl_debug_log, " (%4"UVXf")\n", (UV)trie->trans[ state ].check );
972 PerlIO_printf( Perl_debug_log, " (%4"UVXf") W%4X\n", (UV)trie->trans[ state ].check,
973 trie->states[ TRIE_NODENUM( state ) ].wordnum );
980 /* make_trie(startbranch,first,last,tail,word_count,flags,depth)
981 startbranch: the first branch in the whole branch sequence
982 first : start branch of sequence of branch-exact nodes.
983 May be the same as startbranch
984 last : Thing following the last branch.
985 May be the same as tail.
986 tail : item following the branch sequence
987 count : words in the sequence
988 flags : currently the OP() type we will be building one of /EXACT(|F|Fl)/
991 Inplace optimizes a sequence of 2 or more Branch-Exact nodes into a TRIE node.
993 A trie is an N'ary tree where the branches are determined by digital
994 decomposition of the key. IE, at the root node you look up the 1st character and
995 follow that branch repeat until you find the end of the branches. Nodes can be
996 marked as "accepting" meaning they represent a complete word. Eg:
1000 would convert into the following structure. Numbers represent states, letters
1001 following numbers represent valid transitions on the letter from that state, if
1002 the number is in square brackets it represents an accepting state, otherwise it
1003 will be in parenthesis.
1005 +-h->+-e->[3]-+-r->(8)-+-s->[9]
1009 (1) +-i->(6)-+-s->[7]
1011 +-s->(3)-+-h->(4)-+-e->[5]
1013 Accept Word Mapping: 3=>1 (he),5=>2 (she), 7=>3 (his), 9=>4 (hers)
1015 This shows that when matching against the string 'hers' we will begin at state 1
1016 read 'h' and move to state 2, read 'e' and move to state 3 which is accepting,
1017 then read 'r' and go to state 8 followed by 's' which takes us to state 9 which
1018 is also accepting. Thus we know that we can match both 'he' and 'hers' with a
1019 single traverse. We store a mapping from accepting to state to which word was
1020 matched, and then when we have multiple possibilities we try to complete the
1021 rest of the regex in the order in which they occured in the alternation.
1023 The only prior NFA like behaviour that would be changed by the TRIE support is
1024 the silent ignoring of duplicate alternations which are of the form:
1026 / (DUPE|DUPE) X? (?{ ... }) Y /x
1028 Thus EVAL blocks follwing a trie may be called a different number of times with
1029 and without the optimisation. With the optimisations dupes will be silently
1030 ignored. This inconsistant behaviour of EVAL type nodes is well established as
1031 the following demonstrates:
1033 'words'=~/(word|word|word)(?{ print $1 })[xyz]/
1035 which prints out 'word' three times, but
1037 'words'=~/(word|word|word)(?{ print $1 })S/
1039 which doesnt print it out at all. This is due to other optimisations kicking in.
1041 Example of what happens on a structural level:
1043 The regexp /(ac|ad|ab)+/ will produce the folowing debug output:
1045 1: CURLYM[1] {1,32767}(18)
1056 This would be optimizable with startbranch=5, first=5, last=16, tail=16
1057 and should turn into:
1059 1: CURLYM[1] {1,32767}(18)
1061 [Words:3 Chars Stored:6 Unique Chars:4 States:5 NCP:1]
1069 Cases where tail != last would be like /(?foo|bar)baz/:
1079 which would be optimizable with startbranch=1, first=1, last=7, tail=8
1080 and would end up looking like:
1083 [Words:2 Chars Stored:6 Unique Chars:5 States:7 NCP:1]
1090 d = uvuni_to_utf8_flags(d, uv, 0);
1092 is the recommended Unicode-aware way of saying
1097 #define TRIE_STORE_REVCHAR \
1099 SV *tmp = Perl_newSVpvf_nocontext( "%c", (int)uvc ); \
1100 if (UTF) SvUTF8_on(tmp); \
1101 av_push( TRIE_REVCHARMAP(trie), tmp ); \
1104 #define TRIE_READ_CHAR STMT_START { \
1108 if ( foldlen > 0 ) { \
1109 uvc = utf8n_to_uvuni( scan, UTF8_MAXLEN, &len, uniflags ); \
1114 uvc = utf8n_to_uvuni( (const U8*)uc, UTF8_MAXLEN, &len, uniflags);\
1115 uvc = to_uni_fold( uvc, foldbuf, &foldlen ); \
1116 foldlen -= UNISKIP( uvc ); \
1117 scan = foldbuf + UNISKIP( uvc ); \
1120 uvc = utf8n_to_uvuni( (const U8*)uc, UTF8_MAXLEN, &len, uniflags);\
1130 #define TRIE_LIST_PUSH(state,fid,ns) STMT_START { \
1131 if ( TRIE_LIST_CUR( state ) >=TRIE_LIST_LEN( state ) ) { \
1132 TRIE_LIST_LEN( state ) *= 2; \
1133 Renew( trie->states[ state ].trans.list, \
1134 TRIE_LIST_LEN( state ), reg_trie_trans_le ); \
1136 TRIE_LIST_ITEM( state, TRIE_LIST_CUR( state ) ).forid = fid; \
1137 TRIE_LIST_ITEM( state, TRIE_LIST_CUR( state ) ).newstate = ns; \
1138 TRIE_LIST_CUR( state )++; \
1141 #define TRIE_LIST_NEW(state) STMT_START { \
1142 Newxz( trie->states[ state ].trans.list, \
1143 4, reg_trie_trans_le ); \
1144 TRIE_LIST_CUR( state ) = 1; \
1145 TRIE_LIST_LEN( state ) = 4; \
1148 #define TRIE_HANDLE_WORD(state) STMT_START { \
1149 U16 dupe= trie->states[ state ].wordnum; \
1150 regnode * const noper_next = regnext( noper ); \
1152 if (trie->wordlen) \
1153 trie->wordlen[ curword ] = wordlen; \
1155 /* store the word for dumping */ \
1157 if (OP(noper) != NOTHING) \
1158 tmp = newSVpvn(STRING(noper), STR_LEN(noper)); \
1160 tmp = newSVpvn( "", 0 ); \
1161 if ( UTF ) SvUTF8_on( tmp ); \
1162 av_push( trie->words, tmp ); \
1167 if ( noper_next < tail ) { \
1169 Newxz( trie->jump, word_count + 1, U16); \
1170 trie->jump[curword] = (U16)(tail - noper_next); \
1172 jumper = noper_next; \
1174 nextbranch= regnext(cur); \
1178 /* So it's a dupe. This means we need to maintain a */\
1179 /* linked-list from the first to the next. */\
1180 /* we only allocate the nextword buffer when there */\
1181 /* a dupe, so first time we have to do the allocation */\
1182 if (!trie->nextword) \
1183 Newxz( trie->nextword, word_count + 1, U16); \
1184 while ( trie->nextword[dupe] ) \
1185 dupe= trie->nextword[dupe]; \
1186 trie->nextword[dupe]= curword; \
1188 /* we haven't inserted this word yet. */ \
1189 trie->states[ state ].wordnum = curword; \
1194 #define TRIE_TRANS_STATE(state,base,ucharcount,charid,special) \
1195 ( ( base + charid >= ucharcount \
1196 && base + charid < ubound \
1197 && state == trie->trans[ base - ucharcount + charid ].check \
1198 && trie->trans[ base - ucharcount + charid ].next ) \
1199 ? trie->trans[ base - ucharcount + charid ].next \
1200 : ( state==1 ? special : 0 ) \
1204 #define MADE_JUMP_TRIE 2
1205 #define MADE_EXACT_TRIE 4
1208 S_make_trie(pTHX_ RExC_state_t *pRExC_state, regnode *startbranch, regnode *first, regnode *last, regnode *tail, U32 word_count, U32 flags, U32 depth)
1211 /* first pass, loop through and scan words */
1212 reg_trie_data *trie;
1214 const U32 uniflags = UTF8_ALLOW_DEFAULT;
1219 regnode *jumper = NULL;
1220 regnode *nextbranch = NULL;
1221 /* we just use folder as a flag in utf8 */
1222 const U8 * const folder = ( flags == EXACTF
1224 : ( flags == EXACTFL
1230 const U32 data_slot = add_data( pRExC_state, 1, "t" );
1231 SV *re_trie_maxbuff;
1233 /* these are only used during construction but are useful during
1234 * debugging so we store them in the struct when debugging.
1236 STRLEN trie_charcount=0;
1237 AV *trie_revcharmap;
1239 GET_RE_DEBUG_FLAGS_DECL;
1241 PERL_UNUSED_ARG(depth);
1244 Newxz( trie, 1, reg_trie_data );
1246 trie->startstate = 1;
1247 trie->wordcount = word_count;
1248 RExC_rx->data->data[ data_slot ] = (void*)trie;
1249 Newxz( trie->charmap, 256, U16 );
1250 if (!(UTF && folder))
1251 Newxz( trie->bitmap, ANYOF_BITMAP_SIZE, char );
1253 trie->words = newAV();
1255 TRIE_REVCHARMAP(trie) = newAV();
1257 re_trie_maxbuff = get_sv(RE_TRIE_MAXBUF_NAME, 1);
1258 if (!SvIOK(re_trie_maxbuff)) {
1259 sv_setiv(re_trie_maxbuff, RE_TRIE_MAXBUF_INIT);
1262 PerlIO_printf( Perl_debug_log,
1263 "%*smake_trie start==%d, first==%d, last==%d, tail==%d depth=%d\n",
1264 (int)depth * 2 + 2, "",
1265 REG_NODE_NUM(startbranch),REG_NODE_NUM(first),
1266 REG_NODE_NUM(last), REG_NODE_NUM(tail),
1269 /* -- First loop and Setup --
1271 We first traverse the branches and scan each word to determine if it
1272 contains widechars, and how many unique chars there are, this is
1273 important as we have to build a table with at least as many columns as we
1276 We use an array of integers to represent the character codes 0..255
1277 (trie->charmap) and we use a an HV* to store unicode characters. We use the
1278 native representation of the character value as the key and IV's for the
1281 *TODO* If we keep track of how many times each character is used we can
1282 remap the columns so that the table compression later on is more
1283 efficient in terms of memory by ensuring most common value is in the
1284 middle and the least common are on the outside. IMO this would be better
1285 than a most to least common mapping as theres a decent chance the most
1286 common letter will share a node with the least common, meaning the node
1287 will not be compressable. With a middle is most common approach the worst
1288 case is when we have the least common nodes twice.
1292 for ( cur = first ; cur < last ; cur = regnext( cur ) ) {
1293 regnode * const noper = NEXTOPER( cur );
1294 const U8 *uc = (U8*)STRING( noper );
1295 const U8 * const e = uc + STR_LEN( noper );
1297 U8 foldbuf[ UTF8_MAXBYTES_CASE + 1 ];
1298 const U8 *scan = (U8*)NULL;
1299 U32 wordlen = 0; /* required init */
1302 if (OP(noper) == NOTHING) {
1307 TRIE_BITMAP_SET(trie,*uc);
1308 if ( folder ) TRIE_BITMAP_SET(trie,folder[ *uc ]);
1310 for ( ; uc < e ; uc += len ) {
1311 TRIE_CHARCOUNT(trie)++;
1315 if ( !trie->charmap[ uvc ] ) {
1316 trie->charmap[ uvc ]=( ++trie->uniquecharcount );
1318 trie->charmap[ folder[ uvc ] ] = trie->charmap[ uvc ];
1323 if ( !trie->widecharmap )
1324 trie->widecharmap = newHV();
1326 svpp = hv_fetch( trie->widecharmap, (char*)&uvc, sizeof( UV ), 1 );
1329 Perl_croak( aTHX_ "error creating/fetching widecharmap entry for 0x%"UVXf, uvc );
1331 if ( !SvTRUE( *svpp ) ) {
1332 sv_setiv( *svpp, ++trie->uniquecharcount );
1337 if( cur == first ) {
1340 } else if (chars < trie->minlen) {
1342 } else if (chars > trie->maxlen) {
1346 } /* end first pass */
1347 DEBUG_TRIE_COMPILE_r(
1348 PerlIO_printf( Perl_debug_log, "%*sTRIE(%s): W:%d C:%d Uq:%d Min:%d Max:%d\n",
1349 (int)depth * 2 + 2,"",
1350 ( trie->widecharmap ? "UTF8" : "NATIVE" ), (int)word_count,
1351 (int)TRIE_CHARCOUNT(trie), trie->uniquecharcount,
1352 (int)trie->minlen, (int)trie->maxlen )
1354 Newxz( trie->wordlen, word_count, U32 );
1357 We now know what we are dealing with in terms of unique chars and
1358 string sizes so we can calculate how much memory a naive
1359 representation using a flat table will take. If it's over a reasonable
1360 limit (as specified by ${^RE_TRIE_MAXBUF}) we use a more memory
1361 conservative but potentially much slower representation using an array
1364 At the end we convert both representations into the same compressed
1365 form that will be used in regexec.c for matching with. The latter
1366 is a form that cannot be used to construct with but has memory
1367 properties similar to the list form and access properties similar
1368 to the table form making it both suitable for fast searches and
1369 small enough that its feasable to store for the duration of a program.
1371 See the comment in the code where the compressed table is produced
1372 inplace from the flat tabe representation for an explanation of how
1373 the compression works.
1378 if ( (IV)( ( TRIE_CHARCOUNT(trie) + 1 ) * trie->uniquecharcount + 1) > SvIV(re_trie_maxbuff) ) {
1380 Second Pass -- Array Of Lists Representation
1382 Each state will be represented by a list of charid:state records
1383 (reg_trie_trans_le) the first such element holds the CUR and LEN
1384 points of the allocated array. (See defines above).
1386 We build the initial structure using the lists, and then convert
1387 it into the compressed table form which allows faster lookups
1388 (but cant be modified once converted).
1391 STRLEN transcount = 1;
1393 Newxz( trie->states, TRIE_CHARCOUNT(trie) + 2, reg_trie_state );
1397 for ( cur = first ; cur < last ; cur = regnext( cur ) ) {
1399 regnode * const noper = NEXTOPER( cur );
1400 U8 *uc = (U8*)STRING( noper );
1401 const U8 * const e = uc + STR_LEN( noper );
1402 U32 state = 1; /* required init */
1403 U16 charid = 0; /* sanity init */
1404 U8 *scan = (U8*)NULL; /* sanity init */
1405 STRLEN foldlen = 0; /* required init */
1406 U32 wordlen = 0; /* required init */
1407 U8 foldbuf[ UTF8_MAXBYTES_CASE + 1 ];
1409 if (OP(noper) != NOTHING) {
1410 for ( ; uc < e ; uc += len ) {
1415 charid = trie->charmap[ uvc ];
1417 SV** const svpp = hv_fetch( trie->widecharmap, (char*)&uvc, sizeof( UV ), 0);
1421 charid=(U16)SvIV( *svpp );
1424 /* charid is now 0 if we dont know the char read, or nonzero if we do */
1431 if ( !trie->states[ state ].trans.list ) {
1432 TRIE_LIST_NEW( state );
1434 for ( check = 1; check <= TRIE_LIST_USED( state ); check++ ) {
1435 if ( TRIE_LIST_ITEM( state, check ).forid == charid ) {
1436 newstate = TRIE_LIST_ITEM( state, check ).newstate;
1441 newstate = next_alloc++;
1442 TRIE_LIST_PUSH( state, charid, newstate );
1447 Perl_croak( aTHX_ "panic! In trie construction, no char mapping for %"IVdf, uvc );
1451 TRIE_HANDLE_WORD(state);
1453 } /* end second pass */
1455 trie->laststate = next_alloc;
1456 Renew( trie->states, next_alloc, reg_trie_state );
1458 /* and now dump it out before we compress it */
1459 DEBUG_TRIE_COMPILE_MORE_r(
1460 dump_trie_interim_list(trie,next_alloc,depth+1)
1463 Newxz( trie->trans, transcount ,reg_trie_trans );
1470 for( state=1 ; state < next_alloc ; state ++ ) {
1474 DEBUG_TRIE_COMPILE_MORE_r(
1475 PerlIO_printf( Perl_debug_log, "tp: %d zp: %d ",tp,zp)
1479 if (trie->states[state].trans.list) {
1480 U16 minid=TRIE_LIST_ITEM( state, 1).forid;
1484 for( idx = 2 ; idx <= TRIE_LIST_USED( state ) ; idx++ ) {
1485 const U16 forid = TRIE_LIST_ITEM( state, idx).forid;
1486 if ( forid < minid ) {
1488 } else if ( forid > maxid ) {
1492 if ( transcount < tp + maxid - minid + 1) {
1494 Renew( trie->trans, transcount, reg_trie_trans );
1495 Zero( trie->trans + (transcount / 2), transcount / 2 , reg_trie_trans );
1497 base = trie->uniquecharcount + tp - minid;
1498 if ( maxid == minid ) {
1500 for ( ; zp < tp ; zp++ ) {
1501 if ( ! trie->trans[ zp ].next ) {
1502 base = trie->uniquecharcount + zp - minid;
1503 trie->trans[ zp ].next = TRIE_LIST_ITEM( state, 1).newstate;
1504 trie->trans[ zp ].check = state;
1510 trie->trans[ tp ].next = TRIE_LIST_ITEM( state, 1).newstate;
1511 trie->trans[ tp ].check = state;
1516 for ( idx=1; idx <= TRIE_LIST_USED( state ) ; idx++ ) {
1517 const U32 tid = base - trie->uniquecharcount + TRIE_LIST_ITEM( state, idx ).forid;
1518 trie->trans[ tid ].next = TRIE_LIST_ITEM( state, idx ).newstate;
1519 trie->trans[ tid ].check = state;
1521 tp += ( maxid - minid + 1 );
1523 Safefree(trie->states[ state ].trans.list);
1526 DEBUG_TRIE_COMPILE_MORE_r(
1527 PerlIO_printf( Perl_debug_log, " base: %d\n",base);
1530 trie->states[ state ].trans.base=base;
1532 trie->lasttrans = tp + 1;
1536 Second Pass -- Flat Table Representation.
1538 we dont use the 0 slot of either trans[] or states[] so we add 1 to each.
1539 We know that we will need Charcount+1 trans at most to store the data
1540 (one row per char at worst case) So we preallocate both structures
1541 assuming worst case.
1543 We then construct the trie using only the .next slots of the entry
1546 We use the .check field of the first entry of the node temporarily to
1547 make compression both faster and easier by keeping track of how many non
1548 zero fields are in the node.
1550 Since trans are numbered from 1 any 0 pointer in the table is a FAIL
1553 There are two terms at use here: state as a TRIE_NODEIDX() which is a
1554 number representing the first entry of the node, and state as a
1555 TRIE_NODENUM() which is the trans number. state 1 is TRIE_NODEIDX(1) and
1556 TRIE_NODENUM(1), state 2 is TRIE_NODEIDX(2) and TRIE_NODENUM(3) if there
1557 are 2 entrys per node. eg:
1565 The table is internally in the right hand, idx form. However as we also
1566 have to deal with the states array which is indexed by nodenum we have to
1567 use TRIE_NODENUM() to convert.
1572 Newxz( trie->trans, ( TRIE_CHARCOUNT(trie) + 1 ) * trie->uniquecharcount + 1,
1574 Newxz( trie->states, TRIE_CHARCOUNT(trie) + 2, reg_trie_state );
1575 next_alloc = trie->uniquecharcount + 1;
1578 for ( cur = first ; cur < last ; cur = regnext( cur ) ) {
1580 regnode * const noper = NEXTOPER( cur );
1581 const U8 *uc = (U8*)STRING( noper );
1582 const U8 * const e = uc + STR_LEN( noper );
1584 U32 state = 1; /* required init */
1586 U16 charid = 0; /* sanity init */
1587 U32 accept_state = 0; /* sanity init */
1588 U8 *scan = (U8*)NULL; /* sanity init */
1590 STRLEN foldlen = 0; /* required init */
1591 U32 wordlen = 0; /* required init */
1592 U8 foldbuf[ UTF8_MAXBYTES_CASE + 1 ];
1594 if ( OP(noper) != NOTHING ) {
1595 for ( ; uc < e ; uc += len ) {
1600 charid = trie->charmap[ uvc ];
1602 SV* const * const svpp = hv_fetch( trie->widecharmap, (char*)&uvc, sizeof( UV ), 0);
1603 charid = svpp ? (U16)SvIV(*svpp) : 0;
1607 if ( !trie->trans[ state + charid ].next ) {
1608 trie->trans[ state + charid ].next = next_alloc;
1609 trie->trans[ state ].check++;
1610 next_alloc += trie->uniquecharcount;
1612 state = trie->trans[ state + charid ].next;
1614 Perl_croak( aTHX_ "panic! In trie construction, no char mapping for %"IVdf, uvc );
1616 /* charid is now 0 if we dont know the char read, or nonzero if we do */
1619 accept_state = TRIE_NODENUM( state );
1620 TRIE_HANDLE_WORD(accept_state);
1622 } /* end second pass */
1624 /* and now dump it out before we compress it */
1625 DEBUG_TRIE_COMPILE_MORE_r(
1626 dump_trie_interim_table(trie,next_alloc,depth+1)
1631 * Inplace compress the table.*
1633 For sparse data sets the table constructed by the trie algorithm will
1634 be mostly 0/FAIL transitions or to put it another way mostly empty.
1635 (Note that leaf nodes will not contain any transitions.)
1637 This algorithm compresses the tables by eliminating most such
1638 transitions, at the cost of a modest bit of extra work during lookup:
1640 - Each states[] entry contains a .base field which indicates the
1641 index in the state[] array wheres its transition data is stored.
1643 - If .base is 0 there are no valid transitions from that node.
1645 - If .base is nonzero then charid is added to it to find an entry in
1648 -If trans[states[state].base+charid].check!=state then the
1649 transition is taken to be a 0/Fail transition. Thus if there are fail
1650 transitions at the front of the node then the .base offset will point
1651 somewhere inside the previous nodes data (or maybe even into a node
1652 even earlier), but the .check field determines if the transition is
1656 The following process inplace converts the table to the compressed
1657 table: We first do not compress the root node 1,and mark its all its
1658 .check pointers as 1 and set its .base pointer as 1 as well. This
1659 allows to do a DFA construction from the compressed table later, and
1660 ensures that any .base pointers we calculate later are greater than
1663 - We set 'pos' to indicate the first entry of the second node.
1665 - We then iterate over the columns of the node, finding the first and
1666 last used entry at l and m. We then copy l..m into pos..(pos+m-l),
1667 and set the .check pointers accordingly, and advance pos
1668 appropriately and repreat for the next node. Note that when we copy
1669 the next pointers we have to convert them from the original
1670 NODEIDX form to NODENUM form as the former is not valid post
1673 - If a node has no transitions used we mark its base as 0 and do not
1674 advance the pos pointer.
1676 - If a node only has one transition we use a second pointer into the
1677 structure to fill in allocated fail transitions from other states.
1678 This pointer is independent of the main pointer and scans forward
1679 looking for null transitions that are allocated to a state. When it
1680 finds one it writes the single transition into the "hole". If the
1681 pointer doesnt find one the single transition is appended as normal.
1683 - Once compressed we can Renew/realloc the structures to release the
1686 See "Table-Compression Methods" in sec 3.9 of the Red Dragon,
1687 specifically Fig 3.47 and the associated pseudocode.
1691 const U32 laststate = TRIE_NODENUM( next_alloc );
1694 trie->laststate = laststate;
1696 for ( state = 1 ; state < laststate ; state++ ) {
1698 const U32 stateidx = TRIE_NODEIDX( state );
1699 const U32 o_used = trie->trans[ stateidx ].check;
1700 U32 used = trie->trans[ stateidx ].check;
1701 trie->trans[ stateidx ].check = 0;
1703 for ( charid = 0 ; used && charid < trie->uniquecharcount ; charid++ ) {
1704 if ( flag || trie->trans[ stateidx + charid ].next ) {
1705 if ( trie->trans[ stateidx + charid ].next ) {
1707 for ( ; zp < pos ; zp++ ) {
1708 if ( ! trie->trans[ zp ].next ) {
1712 trie->states[ state ].trans.base = zp + trie->uniquecharcount - charid ;
1713 trie->trans[ zp ].next = SAFE_TRIE_NODENUM( trie->trans[ stateidx + charid ].next );
1714 trie->trans[ zp ].check = state;
1715 if ( ++zp > pos ) pos = zp;
1722 trie->states[ state ].trans.base = pos + trie->uniquecharcount - charid ;
1724 trie->trans[ pos ].next = SAFE_TRIE_NODENUM( trie->trans[ stateidx + charid ].next );
1725 trie->trans[ pos ].check = state;
1730 trie->lasttrans = pos + 1;
1731 Renew( trie->states, laststate + 1, reg_trie_state);
1732 DEBUG_TRIE_COMPILE_MORE_r(
1733 PerlIO_printf( Perl_debug_log,
1734 "%*sAlloc: %d Orig: %"IVdf" elements, Final:%"IVdf". Savings of %%%5.2f\n",
1735 (int)depth * 2 + 2,"",
1736 (int)( ( TRIE_CHARCOUNT(trie) + 1 ) * trie->uniquecharcount + 1 ),
1739 ( ( next_alloc - pos ) * 100 ) / (double)next_alloc );
1742 } /* end table compress */
1744 /* resize the trans array to remove unused space */
1745 Renew( trie->trans, trie->lasttrans, reg_trie_trans);
1747 /* and now dump out the compressed format */
1748 DEBUG_TRIE_COMPILE_r(
1749 dump_trie(trie,depth+1)
1752 { /* Modify the program and insert the new TRIE node*/
1754 U8 nodetype =(U8)(flags & 0xFF);
1760 U32 mjd_nodelen = 0;
1763 This means we convert either the first branch or the first Exact,
1764 depending on whether the thing following (in 'last') is a branch
1765 or not and whther first is the startbranch (ie is it a sub part of
1766 the alternation or is it the whole thing.)
1767 Assuming its a sub part we conver the EXACT otherwise we convert
1768 the whole branch sequence, including the first.
1770 /* Find the node we are going to overwrite */
1771 if ( first == startbranch && OP( last ) != BRANCH ) {
1772 /* whole branch chain */
1775 const regnode *nop = NEXTOPER( convert );
1776 mjd_offset= Node_Offset((nop));
1777 mjd_nodelen= Node_Length((nop));
1780 /* branch sub-chain */
1781 convert = NEXTOPER( first );
1782 NEXT_OFF( first ) = (U16)(last - first);
1784 mjd_offset= Node_Offset((convert));
1785 mjd_nodelen= Node_Length((convert));
1789 PerlIO_printf(Perl_debug_log, "%*sMJD offset:%"UVuf" MJD length:%"UVuf"\n",
1790 (int)depth * 2 + 2, "",
1791 (UV)mjd_offset, (UV)mjd_nodelen)
1794 /* But first we check to see if there is a common prefix we can
1795 split out as an EXACT and put in front of the TRIE node. */
1796 trie->startstate= 1;
1797 if ( trie->bitmap && !trie->widecharmap && !trie->jump ) {
1800 PerlIO_printf(Perl_debug_log, "%*sLaststate:%"UVuf"\n",
1801 (int)depth * 2 + 2, "",
1802 (UV)trie->laststate)
1804 for ( state = 1 ; state < trie->laststate-1 ; state++ ) {
1808 const U32 base = trie->states[ state ].trans.base;
1810 if ( trie->states[state].wordnum )
1813 for ( ofs = 0 ; ofs < trie->uniquecharcount ; ofs++ ) {
1814 if ( ( base + ofs >= trie->uniquecharcount ) &&
1815 ( base + ofs - trie->uniquecharcount < trie->lasttrans ) &&
1816 trie->trans[ base + ofs - trie->uniquecharcount ].check == state )
1818 if ( ++count > 1 ) {
1819 SV **tmp = av_fetch( TRIE_REVCHARMAP(trie), ofs, 0);
1820 const U8 *ch = (U8*)SvPV_nolen_const( *tmp );
1821 if ( state == 1 ) break;
1823 Zero(trie->bitmap, ANYOF_BITMAP_SIZE, char);
1825 PerlIO_printf(Perl_debug_log,
1826 "%*sNew Start State=%"UVuf" Class: [",
1827 (int)depth * 2 + 2, "",
1830 SV ** const tmp = av_fetch( TRIE_REVCHARMAP(trie), idx, 0);
1831 const U8 * const ch = (U8*)SvPV_nolen_const( *tmp );
1833 TRIE_BITMAP_SET(trie,*ch);
1835 TRIE_BITMAP_SET(trie, folder[ *ch ]);
1837 PerlIO_printf(Perl_debug_log, (char*)ch)
1841 TRIE_BITMAP_SET(trie,*ch);
1843 TRIE_BITMAP_SET(trie,folder[ *ch ]);
1844 DEBUG_OPTIMISE_r(PerlIO_printf( Perl_debug_log,"%s", ch));
1850 SV **tmp = av_fetch( TRIE_REVCHARMAP(trie), idx, 0);
1851 const char *ch = SvPV_nolen_const( *tmp );
1853 PerlIO_printf( Perl_debug_log,
1854 "%*sPrefix State: %"UVuf" Idx:%"UVuf" Char='%s'\n",
1855 (int)depth * 2 + 2, "",
1856 (UV)state, (UV)idx, ch)
1859 OP( convert ) = nodetype;
1860 str=STRING(convert);
1869 DEBUG_OPTIMISE_r(PerlIO_printf( Perl_debug_log,"]\n"));
1875 regnode *n = convert+NODE_SZ_STR(convert);
1876 NEXT_OFF(convert) = NODE_SZ_STR(convert);
1877 trie->startstate = state;
1878 trie->minlen -= (state - 1);
1879 trie->maxlen -= (state - 1);
1881 regnode *fix = convert;
1883 Set_Node_Offset_Length(convert, mjd_offset, state - 1);
1884 while( ++fix < n ) {
1885 Set_Node_Offset_Length(fix, 0, 0);
1891 NEXT_OFF(convert) = (U16)(tail - convert);
1892 DEBUG_r(optimize= n);
1898 if ( trie->maxlen ) {
1899 NEXT_OFF( convert ) = (U16)(tail - convert);
1900 ARG_SET( convert, data_slot );
1901 /* Store the offset to the first unabsorbed branch in
1902 jump[0], which is otherwise unused by the jump logic.
1903 We use this when dumping a trie and during optimisation. */
1905 trie->jump[0] = (U16)(tail - nextbranch);
1908 if ( !trie->states[trie->startstate].wordnum && trie->bitmap &&
1909 ( (char *)jumper - (char *)convert) >= (int)sizeof(struct regnode_charclass) )
1911 OP( convert ) = TRIEC;
1912 Copy(trie->bitmap, ((struct regnode_charclass *)convert)->bitmap, ANYOF_BITMAP_SIZE, char);
1913 Safefree(trie->bitmap);
1916 OP( convert ) = TRIE;
1918 /* store the type in the flags */
1919 convert->flags = nodetype;
1923 + regarglen[ OP( convert ) ];
1925 /* XXX We really should free up the resource in trie now,
1926 as we won't use them - (which resources?) dmq */
1928 /* needed for dumping*/
1930 regnode *opt = convert;
1931 while (++opt<optimize) {
1932 Set_Node_Offset_Length(opt,0,0);
1935 Try to clean up some of the debris left after the
1938 while( optimize < jumper ) {
1939 mjd_nodelen += Node_Length((optimize));
1940 OP( optimize ) = OPTIMIZED;
1941 Set_Node_Offset_Length(optimize,0,0);
1944 Set_Node_Offset_Length(convert,mjd_offset,mjd_nodelen);
1946 } /* end node insert */
1948 SvREFCNT_dec(TRIE_REVCHARMAP(trie));
1952 : trie->startstate>1
1958 S_make_trie_failtable(pTHX_ RExC_state_t *pRExC_state, regnode *source, regnode *stclass, U32 depth)
1960 /* The Trie is constructed and compressed now so we can build a fail array now if its needed
1962 This is basically the Aho-Corasick algorithm. Its from exercise 3.31 and 3.32 in the
1963 "Red Dragon" -- Compilers, principles, techniques, and tools. Aho, Sethi, Ullman 1985/88
1966 We find the fail state for each state in the trie, this state is the longest proper
1967 suffix of the current states 'word' that is also a proper prefix of another word in our
1968 trie. State 1 represents the word '' and is the thus the default fail state. This allows
1969 the DFA not to have to restart after its tried and failed a word at a given point, it
1970 simply continues as though it had been matching the other word in the first place.
1972 'abcdgu'=~/abcdefg|cdgu/
1973 When we get to 'd' we are still matching the first word, we would encounter 'g' which would
1974 fail, which would bring use to the state representing 'd' in the second word where we would
1975 try 'g' and succeed, prodceding to match 'cdgu'.
1977 /* add a fail transition */
1978 reg_trie_data *trie=(reg_trie_data *)RExC_rx->data->data[ARG(source)];
1980 const U32 ucharcount = trie->uniquecharcount;
1981 const U32 numstates = trie->laststate;
1982 const U32 ubound = trie->lasttrans + ucharcount;
1986 U32 base = trie->states[ 1 ].trans.base;
1989 const U32 data_slot = add_data( pRExC_state, 1, "T" );
1990 GET_RE_DEBUG_FLAGS_DECL;
1992 PERL_UNUSED_ARG(depth);
1996 ARG_SET( stclass, data_slot );
1997 Newxz( aho, 1, reg_ac_data );
1998 RExC_rx->data->data[ data_slot ] = (void*)aho;
2000 aho->states=(reg_trie_state *)savepvn((const char*)trie->states,
2001 (trie->laststate+1)*sizeof(reg_trie_state));
2002 Newxz( q, numstates, U32);
2003 Newxz( aho->fail, numstates, U32 );
2006 /* initialize fail[0..1] to be 1 so that we always have
2007 a valid final fail state */
2008 fail[ 0 ] = fail[ 1 ] = 1;
2010 for ( charid = 0; charid < ucharcount ; charid++ ) {
2011 const U32 newstate = TRIE_TRANS_STATE( 1, base, ucharcount, charid, 0 );
2013 q[ q_write ] = newstate;
2014 /* set to point at the root */
2015 fail[ q[ q_write++ ] ]=1;
2018 while ( q_read < q_write) {
2019 const U32 cur = q[ q_read++ % numstates ];
2020 base = trie->states[ cur ].trans.base;
2022 for ( charid = 0 ; charid < ucharcount ; charid++ ) {
2023 const U32 ch_state = TRIE_TRANS_STATE( cur, base, ucharcount, charid, 1 );
2025 U32 fail_state = cur;
2028 fail_state = fail[ fail_state ];
2029 fail_base = aho->states[ fail_state ].trans.base;
2030 } while ( !TRIE_TRANS_STATE( fail_state, fail_base, ucharcount, charid, 1 ) );
2032 fail_state = TRIE_TRANS_STATE( fail_state, fail_base, ucharcount, charid, 1 );
2033 fail[ ch_state ] = fail_state;
2034 if ( !aho->states[ ch_state ].wordnum && aho->states[ fail_state ].wordnum )
2036 aho->states[ ch_state ].wordnum = aho->states[ fail_state ].wordnum;
2038 q[ q_write++ % numstates] = ch_state;
2042 /* restore fail[0..1] to 0 so that we "fall out" of the AC loop
2043 when we fail in state 1, this allows us to use the
2044 charclass scan to find a valid start char. This is based on the principle
2045 that theres a good chance the string being searched contains lots of stuff
2046 that cant be a start char.
2048 fail[ 0 ] = fail[ 1 ] = 0;
2049 DEBUG_TRIE_COMPILE_r({
2050 PerlIO_printf(Perl_debug_log, "%*sStclass Failtable: 0", (int)(depth * 2), "");
2051 for( q_read=1; q_read<numstates; q_read++ ) {
2052 PerlIO_printf(Perl_debug_log, ", %"UVuf, (UV)fail[q_read]);
2054 PerlIO_printf(Perl_debug_log, "\n");
2057 /*RExC_seen |= REG_SEEN_TRIEDFA;*/
2062 * There are strange code-generation bugs caused on sparc64 by gcc-2.95.2.
2063 * These need to be revisited when a newer toolchain becomes available.
2065 #if defined(__sparc64__) && defined(__GNUC__)
2066 # if __GNUC__ < 2 || (__GNUC__ == 2 && __GNUC_MINOR__ < 96)
2067 # undef SPARC64_GCC_WORKAROUND
2068 # define SPARC64_GCC_WORKAROUND 1
2072 #define DEBUG_PEEP(str,scan,depth) \
2073 DEBUG_OPTIMISE_r({ \
2074 SV * const mysv=sv_newmortal(); \
2075 regnode *Next = regnext(scan); \
2076 regprop(RExC_rx, mysv, scan); \
2077 PerlIO_printf(Perl_debug_log, "%*s" str ">%3d: %s [%d]\n", \
2078 (int)depth*2, "", REG_NODE_NUM(scan), SvPV_nolen_const(mysv),\
2079 Next ? (REG_NODE_NUM(Next)) : 0 ); \
2086 #define JOIN_EXACT(scan,min,flags) \
2087 if (PL_regkind[OP(scan)] == EXACT) \
2088 join_exact(pRExC_state,(scan),(min),(flags),NULL,depth+1)
2091 S_join_exact(pTHX_ RExC_state_t *pRExC_state, regnode *scan, I32 *min, U32 flags,regnode *val, U32 depth) {
2092 /* Merge several consecutive EXACTish nodes into one. */
2093 regnode *n = regnext(scan);
2095 regnode *next = scan + NODE_SZ_STR(scan);
2099 regnode *stop = scan;
2100 GET_RE_DEBUG_FLAGS_DECL;
2102 PERL_UNUSED_ARG(depth);
2104 #ifndef EXPERIMENTAL_INPLACESCAN
2105 PERL_UNUSED_ARG(flags);
2106 PERL_UNUSED_ARG(val);
2108 DEBUG_PEEP("join",scan,depth);
2110 /* Skip NOTHING, merge EXACT*. */
2112 ( PL_regkind[OP(n)] == NOTHING ||
2113 (stringok && (OP(n) == OP(scan))))
2115 && NEXT_OFF(scan) + NEXT_OFF(n) < I16_MAX) {
2117 if (OP(n) == TAIL || n > next)
2119 if (PL_regkind[OP(n)] == NOTHING) {
2120 DEBUG_PEEP("skip:",n,depth);
2121 NEXT_OFF(scan) += NEXT_OFF(n);
2122 next = n + NODE_STEP_REGNODE;
2129 else if (stringok) {
2130 const unsigned int oldl = STR_LEN(scan);
2131 regnode * const nnext = regnext(n);
2133 DEBUG_PEEP("merg",n,depth);
2136 if (oldl + STR_LEN(n) > U8_MAX)
2138 NEXT_OFF(scan) += NEXT_OFF(n);
2139 STR_LEN(scan) += STR_LEN(n);
2140 next = n + NODE_SZ_STR(n);
2141 /* Now we can overwrite *n : */
2142 Move(STRING(n), STRING(scan) + oldl, STR_LEN(n), char);
2150 #ifdef EXPERIMENTAL_INPLACESCAN
2151 if (flags && !NEXT_OFF(n)) {
2152 DEBUG_PEEP("atch", val, depth);
2153 if (reg_off_by_arg[OP(n)]) {
2154 ARG_SET(n, val - n);
2157 NEXT_OFF(n) = val - n;
2164 if (UTF && ( OP(scan) == EXACTF ) && ( STR_LEN(scan) >= 6 ) ) {
2166 Two problematic code points in Unicode casefolding of EXACT nodes:
2168 U+0390 - GREEK SMALL LETTER IOTA WITH DIALYTIKA AND TONOS
2169 U+03B0 - GREEK SMALL LETTER UPSILON WITH DIALYTIKA AND TONOS
2175 U+03B9 U+0308 U+0301 0xCE 0xB9 0xCC 0x88 0xCC 0x81
2176 U+03C5 U+0308 U+0301 0xCF 0x85 0xCC 0x88 0xCC 0x81
2178 This means that in case-insensitive matching (or "loose matching",
2179 as Unicode calls it), an EXACTF of length six (the UTF-8 encoded byte
2180 length of the above casefolded versions) can match a target string
2181 of length two (the byte length of UTF-8 encoded U+0390 or U+03B0).
2182 This would rather mess up the minimum length computation.
2184 What we'll do is to look for the tail four bytes, and then peek
2185 at the preceding two bytes to see whether we need to decrease
2186 the minimum length by four (six minus two).
2188 Thanks to the design of UTF-8, there cannot be false matches:
2189 A sequence of valid UTF-8 bytes cannot be a subsequence of
2190 another valid sequence of UTF-8 bytes.
2193 char * const s0 = STRING(scan), *s, *t;
2194 char * const s1 = s0 + STR_LEN(scan) - 1;
2195 char * const s2 = s1 - 4;
2196 #ifdef EBCDIC /* RD tunifold greek 0390 and 03B0 */
2197 const char t0[] = "\xaf\x49\xaf\x42";
2199 const char t0[] = "\xcc\x88\xcc\x81";
2201 const char * const t1 = t0 + 3;
2204 s < s2 && (t = ninstr(s, s1, t0, t1));
2207 if (((U8)t[-1] == 0x68 && (U8)t[-2] == 0xB4) ||
2208 ((U8)t[-1] == 0x46 && (U8)t[-2] == 0xB5))
2210 if (((U8)t[-1] == 0xB9 && (U8)t[-2] == 0xCE) ||
2211 ((U8)t[-1] == 0x85 && (U8)t[-2] == 0xCF))
2219 n = scan + NODE_SZ_STR(scan);
2221 if (PL_regkind[OP(n)] != NOTHING || OP(n) == NOTHING) {
2228 DEBUG_OPTIMISE_r(if (merged){DEBUG_PEEP("finl",scan,depth)});
2232 /* REx optimizer. Converts nodes into quickier variants "in place".
2233 Finds fixed substrings. */
2235 /* Stops at toplevel WHILEM as well as at "last". At end *scanp is set
2236 to the position after last scanned or to NULL. */
2241 S_study_chunk(pTHX_ RExC_state_t *pRExC_state, regnode **scanp,
2242 I32 *minlenp, I32 *deltap,
2243 regnode *last, scan_data_t *data, U32 flags, U32 depth)
2244 /* scanp: Start here (read-write). */
2245 /* deltap: Write maxlen-minlen here. */
2246 /* last: Stop before this one. */
2249 I32 min = 0, pars = 0, code;
2250 regnode *scan = *scanp, *next;
2252 int is_inf = (flags & SCF_DO_SUBSTR) && (data->flags & SF_IS_INF);
2253 int is_inf_internal = 0; /* The studied chunk is infinite */
2254 I32 is_par = OP(scan) == OPEN ? ARG(scan) : 0;
2255 scan_data_t data_fake;
2256 struct regnode_charclass_class and_with; /* Valid if flags & SCF_DO_STCLASS_OR */
2257 SV *re_trie_maxbuff = NULL;
2258 regnode *first_non_open = scan;
2261 GET_RE_DEBUG_FLAGS_DECL;
2263 StructCopy(&zero_scan_data, &data_fake, scan_data_t);
2266 while (first_non_open && OP(first_non_open) == OPEN)
2267 first_non_open=regnext(first_non_open);
2271 while (scan && OP(scan) != END && scan < last) {
2272 /* Peephole optimizer: */
2273 DEBUG_STUDYDATA(data,depth);
2274 DEBUG_PEEP("Peep",scan,depth);
2275 JOIN_EXACT(scan,&min,0);
2277 /* Follow the next-chain of the current node and optimize
2278 away all the NOTHINGs from it. */
2279 if (OP(scan) != CURLYX) {
2280 const int max = (reg_off_by_arg[OP(scan)]
2282 /* I32 may be smaller than U16 on CRAYs! */
2283 : (I32_MAX < U16_MAX ? I32_MAX : U16_MAX));
2284 int off = (reg_off_by_arg[OP(scan)] ? ARG(scan) : NEXT_OFF(scan));
2288 /* Skip NOTHING and LONGJMP. */
2289 while ((n = regnext(n))
2290 && ((PL_regkind[OP(n)] == NOTHING && (noff = NEXT_OFF(n)))
2291 || ((OP(n) == LONGJMP) && (noff = ARG(n))))
2292 && off + noff < max)
2294 if (reg_off_by_arg[OP(scan)])
2297 NEXT_OFF(scan) = off;
2302 /* The principal pseudo-switch. Cannot be a switch, since we
2303 look into several different things. */
2304 if (OP(scan) == BRANCH || OP(scan) == BRANCHJ
2305 || OP(scan) == IFTHEN || OP(scan) == SUSPEND) {
2306 next = regnext(scan);
2308 /* demq: the op(next)==code check is to see if we have "branch-branch" AFAICT */
2310 if (OP(next) == code || code == IFTHEN || code == SUSPEND) {
2311 /* NOTE - There is similar code to this block below for handling
2312 TRIE nodes on a re-study. If you change stuff here check there
2314 I32 max1 = 0, min1 = I32_MAX, num = 0;
2315 struct regnode_charclass_class accum;
2316 regnode * const startbranch=scan;
2318 if (flags & SCF_DO_SUBSTR) /* XXXX Add !SUSPEND? */
2319 scan_commit(pRExC_state, data, minlenp); /* Cannot merge strings after this. */
2320 if (flags & SCF_DO_STCLASS)
2321 cl_init_zero(pRExC_state, &accum);
2323 while (OP(scan) == code) {
2324 I32 deltanext, minnext, f = 0, fake;
2325 struct regnode_charclass_class this_class;
2328 data_fake.flags = 0;
2330 data_fake.whilem_c = data->whilem_c;
2331 data_fake.last_closep = data->last_closep;
2334 data_fake.last_closep = &fake;
2335 next = regnext(scan);
2336 scan = NEXTOPER(scan);
2338 scan = NEXTOPER(scan);
2339 if (flags & SCF_DO_STCLASS) {
2340 cl_init(pRExC_state, &this_class);
2341 data_fake.start_class = &this_class;
2342 f = SCF_DO_STCLASS_AND;
2344 if (flags & SCF_WHILEM_VISITED_POS)
2345 f |= SCF_WHILEM_VISITED_POS;
2347 /* we suppose the run is continuous, last=next...*/
2348 minnext = study_chunk(pRExC_state, &scan, minlenp, &deltanext,
2349 next, &data_fake, f,depth+1);
2352 if (max1 < minnext + deltanext)
2353 max1 = minnext + deltanext;
2354 if (deltanext == I32_MAX)
2355 is_inf = is_inf_internal = 1;
2357 if (data_fake.flags & (SF_HAS_PAR|SF_IN_PAR))
2360 if (data_fake.flags & SF_HAS_EVAL)
2361 data->flags |= SF_HAS_EVAL;
2362 data->whilem_c = data_fake.whilem_c;
2364 if (flags & SCF_DO_STCLASS)
2365 cl_or(pRExC_state, &accum, &this_class);
2366 if (code == SUSPEND)
2369 if (code == IFTHEN && num < 2) /* Empty ELSE branch */
2371 if (flags & SCF_DO_SUBSTR) {
2372 data->pos_min += min1;
2373 data->pos_delta += max1 - min1;
2374 if (max1 != min1 || is_inf)
2375 data->longest = &(data->longest_float);
2378 delta += max1 - min1;
2379 if (flags & SCF_DO_STCLASS_OR) {
2380 cl_or(pRExC_state, data->start_class, &accum);
2382 cl_and(data->start_class, &and_with);
2383 flags &= ~SCF_DO_STCLASS;
2386 else if (flags & SCF_DO_STCLASS_AND) {
2388 cl_and(data->start_class, &accum);
2389 flags &= ~SCF_DO_STCLASS;
2392 /* Switch to OR mode: cache the old value of
2393 * data->start_class */
2394 StructCopy(data->start_class, &and_with,
2395 struct regnode_charclass_class);
2396 flags &= ~SCF_DO_STCLASS_AND;
2397 StructCopy(&accum, data->start_class,
2398 struct regnode_charclass_class);
2399 flags |= SCF_DO_STCLASS_OR;
2400 data->start_class->flags |= ANYOF_EOS;
2404 if (PERL_ENABLE_TRIE_OPTIMISATION && OP( startbranch ) == BRANCH ) {
2407 Assuming this was/is a branch we are dealing with: 'scan' now
2408 points at the item that follows the branch sequence, whatever
2409 it is. We now start at the beginning of the sequence and look
2416 which would be constructed from a pattern like /A|LIST|OF|WORDS/
2418 If we can find such a subseqence we need to turn the first
2419 element into a trie and then add the subsequent branch exact
2420 strings to the trie.
2424 1. patterns where the whole set of branch can be converted.
2426 2. patterns where only a subset can be converted.
2428 In case 1 we can replace the whole set with a single regop
2429 for the trie. In case 2 we need to keep the start and end
2432 'BRANCH EXACT; BRANCH EXACT; BRANCH X'
2433 becomes BRANCH TRIE; BRANCH X;
2435 There is an additional case, that being where there is a
2436 common prefix, which gets split out into an EXACT like node
2437 preceding the TRIE node.
2439 If x(1..n)==tail then we can do a simple trie, if not we make
2440 a "jump" trie, such that when we match the appropriate word
2441 we "jump" to the appopriate tail node. Essentailly we turn
2442 a nested if into a case structure of sorts.
2447 if (!re_trie_maxbuff) {
2448 re_trie_maxbuff = get_sv(RE_TRIE_MAXBUF_NAME, 1);
2449 if (!SvIOK(re_trie_maxbuff))
2450 sv_setiv(re_trie_maxbuff, RE_TRIE_MAXBUF_INIT);
2452 if ( SvIV(re_trie_maxbuff)>=0 ) {
2454 regnode *first = (regnode *)NULL;
2455 regnode *last = (regnode *)NULL;
2456 regnode *tail = scan;
2461 SV * const mysv = sv_newmortal(); /* for dumping */
2463 /* var tail is used because there may be a TAIL
2464 regop in the way. Ie, the exacts will point to the
2465 thing following the TAIL, but the last branch will
2466 point at the TAIL. So we advance tail. If we
2467 have nested (?:) we may have to move through several
2471 while ( OP( tail ) == TAIL ) {
2472 /* this is the TAIL generated by (?:) */
2473 tail = regnext( tail );
2478 regprop(RExC_rx, mysv, tail );
2479 PerlIO_printf( Perl_debug_log, "%*s%s%s\n",
2480 (int)depth * 2 + 2, "",
2481 "Looking for TRIE'able sequences. Tail node is: ",
2482 SvPV_nolen_const( mysv )
2488 step through the branches, cur represents each
2489 branch, noper is the first thing to be matched
2490 as part of that branch and noper_next is the
2491 regnext() of that node. if noper is an EXACT
2492 and noper_next is the same as scan (our current
2493 position in the regex) then the EXACT branch is
2494 a possible optimization target. Once we have
2495 two or more consequetive such branches we can
2496 create a trie of the EXACT's contents and stich
2497 it in place. If the sequence represents all of
2498 the branches we eliminate the whole thing and
2499 replace it with a single TRIE. If it is a
2500 subsequence then we need to stitch it in. This
2501 means the first branch has to remain, and needs
2502 to be repointed at the item on the branch chain
2503 following the last branch optimized. This could
2504 be either a BRANCH, in which case the
2505 subsequence is internal, or it could be the
2506 item following the branch sequence in which
2507 case the subsequence is at the end.
2511 /* dont use tail as the end marker for this traverse */
2512 for ( cur = startbranch ; cur != scan ; cur = regnext( cur ) ) {
2513 regnode * const noper = NEXTOPER( cur );
2514 #if defined(DEBUGGING) || defined(NOJUMPTRIE)
2515 regnode * const noper_next = regnext( noper );
2519 regprop(RExC_rx, mysv, cur);
2520 PerlIO_printf( Perl_debug_log, "%*s- %s (%d)",
2521 (int)depth * 2 + 2,"", SvPV_nolen_const( mysv ), REG_NODE_NUM(cur) );
2523 regprop(RExC_rx, mysv, noper);
2524 PerlIO_printf( Perl_debug_log, " -> %s",
2525 SvPV_nolen_const(mysv));
2528 regprop(RExC_rx, mysv, noper_next );
2529 PerlIO_printf( Perl_debug_log,"\t=> %s\t",
2530 SvPV_nolen_const(mysv));
2532 PerlIO_printf( Perl_debug_log, "(First==%d,Last==%d,Cur==%d)\n",
2533 REG_NODE_NUM(first), REG_NODE_NUM(last), REG_NODE_NUM(cur) );
2535 if ( (((first && optype!=NOTHING) ? OP( noper ) == optype
2536 : PL_regkind[ OP( noper ) ] == EXACT )
2537 || OP(noper) == NOTHING )
2539 && noper_next == tail
2544 if ( !first || optype == NOTHING ) {
2545 if (!first) first = cur;
2546 optype = OP( noper );
2552 make_trie( pRExC_state,
2553 startbranch, first, cur, tail, count,
2556 if ( PL_regkind[ OP( noper ) ] == EXACT
2558 && noper_next == tail
2563 optype = OP( noper );
2573 regprop(RExC_rx, mysv, cur);
2574 PerlIO_printf( Perl_debug_log,
2575 "%*s- %s (%d) <SCAN FINISHED>\n", (int)depth * 2 + 2,
2576 "", SvPV_nolen_const( mysv ),REG_NODE_NUM(cur));
2580 made= make_trie( pRExC_state, startbranch, first, scan, tail, count, optype, depth+1 );
2581 #ifdef TRIE_STUDY_OPT
2582 if ( ((made == MADE_EXACT_TRIE &&
2583 startbranch == first)
2584 || ( first_non_open == first )) &&
2586 flags |= SCF_TRIE_RESTUDY;
2594 else if ( code == BRANCHJ ) { /* single branch is optimized. */
2595 scan = NEXTOPER(NEXTOPER(scan));
2596 } else /* single branch is optimized. */
2597 scan = NEXTOPER(scan);
2600 else if (OP(scan) == EXACT) {
2601 I32 l = STR_LEN(scan);
2604 const U8 * const s = (U8*)STRING(scan);
2605 l = utf8_length(s, s + l);
2606 uc = utf8_to_uvchr(s, NULL);
2608 uc = *((U8*)STRING(scan));
2611 if (flags & SCF_DO_SUBSTR) { /* Update longest substr. */
2612 /* The code below prefers earlier match for fixed
2613 offset, later match for variable offset. */
2614 if (data->last_end == -1) { /* Update the start info. */
2615 data->last_start_min = data->pos_min;
2616 data->last_start_max = is_inf
2617 ? I32_MAX : data->pos_min + data->pos_delta;
2619 sv_catpvn(data->last_found, STRING(scan), STR_LEN(scan));
2621 SvUTF8_on(data->last_found);
2623 SV * const sv = data->last_found;
2624 MAGIC * const mg = SvUTF8(sv) && SvMAGICAL(sv) ?
2625 mg_find(sv, PERL_MAGIC_utf8) : NULL;
2626 if (mg && mg->mg_len >= 0)
2627 mg->mg_len += utf8_length((U8*)STRING(scan),
2628 (U8*)STRING(scan)+STR_LEN(scan));
2630 data->last_end = data->pos_min + l;
2631 data->pos_min += l; /* As in the first entry. */
2632 data->flags &= ~SF_BEFORE_EOL;
2634 if (flags & SCF_DO_STCLASS_AND) {
2635 /* Check whether it is compatible with what we know already! */
2639 (!(data->start_class->flags & (ANYOF_CLASS | ANYOF_LOCALE))
2640 && !ANYOF_BITMAP_TEST(data->start_class, uc)
2641 && (!(data->start_class->flags & ANYOF_FOLD)
2642 || !ANYOF_BITMAP_TEST(data->start_class, PL_fold[uc])))
2645 ANYOF_CLASS_ZERO(data->start_class);
2646 ANYOF_BITMAP_ZERO(data->start_class);
2648 ANYOF_BITMAP_SET(data->start_class, uc);
2649 data->start_class->flags &= ~ANYOF_EOS;
2651 data->start_class->flags &= ~ANYOF_UNICODE_ALL;
2653 else if (flags & SCF_DO_STCLASS_OR) {
2654 /* false positive possible if the class is case-folded */
2656 ANYOF_BITMAP_SET(data->start_class, uc);
2658 data->start_class->flags |= ANYOF_UNICODE_ALL;
2659 data->start_class->flags &= ~ANYOF_EOS;
2660 cl_and(data->start_class, &and_with);
2662 flags &= ~SCF_DO_STCLASS;
2664 else if (PL_regkind[OP(scan)] == EXACT) { /* But OP != EXACT! */
2665 I32 l = STR_LEN(scan);
2666 UV uc = *((U8*)STRING(scan));
2668 /* Search for fixed substrings supports EXACT only. */
2669 if (flags & SCF_DO_SUBSTR) {
2671 scan_commit(pRExC_state, data, minlenp);
2674 const U8 * const s = (U8 *)STRING(scan);
2675 l = utf8_length(s, s + l);
2676 uc = utf8_to_uvchr(s, NULL);
2679 if (flags & SCF_DO_SUBSTR)
2681 if (flags & SCF_DO_STCLASS_AND) {
2682 /* Check whether it is compatible with what we know already! */
2686 (!(data->start_class->flags & (ANYOF_CLASS | ANYOF_LOCALE))
2687 && !ANYOF_BITMAP_TEST(data->start_class, uc)
2688 && !ANYOF_BITMAP_TEST(data->start_class, PL_fold[uc])))
2690 ANYOF_CLASS_ZERO(data->start_class);
2691 ANYOF_BITMAP_ZERO(data->start_class);
2693 ANYOF_BITMAP_SET(data->start_class, uc);
2694 data->start_class->flags &= ~ANYOF_EOS;
2695 data->start_class->flags |= ANYOF_FOLD;
2696 if (OP(scan) == EXACTFL)
2697 data->start_class->flags |= ANYOF_LOCALE;
2700 else if (flags & SCF_DO_STCLASS_OR) {
2701 if (data->start_class->flags & ANYOF_FOLD) {
2702 /* false positive possible if the class is case-folded.
2703 Assume that the locale settings are the same... */
2705 ANYOF_BITMAP_SET(data->start_class, uc);
2706 data->start_class->flags &= ~ANYOF_EOS;
2708 cl_and(data->start_class, &and_with);
2710 flags &= ~SCF_DO_STCLASS;
2712 else if (strchr((const char*)PL_varies,OP(scan))) {
2713 I32 mincount, maxcount, minnext, deltanext, fl = 0;
2714 I32 f = flags, pos_before = 0;
2715 regnode * const oscan = scan;
2716 struct regnode_charclass_class this_class;
2717 struct regnode_charclass_class *oclass = NULL;
2718 I32 next_is_eval = 0;
2720 switch (PL_regkind[OP(scan)]) {
2721 case WHILEM: /* End of (?:...)* . */
2722 scan = NEXTOPER(scan);
2725 if (flags & (SCF_DO_SUBSTR | SCF_DO_STCLASS)) {
2726 next = NEXTOPER(scan);
2727 if (OP(next) == EXACT || (flags & SCF_DO_STCLASS)) {
2729 maxcount = REG_INFTY;
2730 next = regnext(scan);
2731 scan = NEXTOPER(scan);
2735 if (flags & SCF_DO_SUBSTR)
2740 if (flags & SCF_DO_STCLASS) {
2742 maxcount = REG_INFTY;
2743 next = regnext(scan);
2744 scan = NEXTOPER(scan);
2747 is_inf = is_inf_internal = 1;
2748 scan = regnext(scan);
2749 if (flags & SCF_DO_SUBSTR) {
2750 scan_commit(pRExC_state, data, minlenp); /* Cannot extend fixed substrings */
2751 data->longest = &(data->longest_float);
2753 goto optimize_curly_tail;
2755 mincount = ARG1(scan);
2756 maxcount = ARG2(scan);
2757 next = regnext(scan);
2758 if (OP(scan) == CURLYX) {
2759 I32 lp = (data ? *(data->last_closep) : 0);
2760 scan->flags = ((lp <= (I32)U8_MAX) ? (U8)lp : U8_MAX);
2762 scan = NEXTOPER(scan) + EXTRA_STEP_2ARGS;
2763 next_is_eval = (OP(scan) == EVAL);
2765 if (flags & SCF_DO_SUBSTR) {
2766 if (mincount == 0) scan_commit(pRExC_state,data,minlenp); /* Cannot extend fixed substrings */
2767 pos_before = data->pos_min;
2771 data->flags &= ~(SF_HAS_PAR|SF_IN_PAR|SF_HAS_EVAL);
2773 data->flags |= SF_IS_INF;
2775 if (flags & SCF_DO_STCLASS) {
2776 cl_init(pRExC_state, &this_class);
2777 oclass = data->start_class;
2778 data->start_class = &this_class;
2779 f |= SCF_DO_STCLASS_AND;
2780 f &= ~SCF_DO_STCLASS_OR;
2782 /* These are the cases when once a subexpression
2783 fails at a particular position, it cannot succeed
2784 even after backtracking at the enclosing scope.
2786 XXXX what if minimal match and we are at the
2787 initial run of {n,m}? */
2788 if ((mincount != maxcount - 1) && (maxcount != REG_INFTY))
2789 f &= ~SCF_WHILEM_VISITED_POS;
2791 /* This will finish on WHILEM, setting scan, or on NULL: */
2792 minnext = study_chunk(pRExC_state, &scan, minlenp, &deltanext, last, data,
2794 ? (f & ~SCF_DO_SUBSTR) : f),depth+1);
2796 if (flags & SCF_DO_STCLASS)
2797 data->start_class = oclass;
2798 if (mincount == 0 || minnext == 0) {
2799 if (flags & SCF_DO_STCLASS_OR) {
2800 cl_or(pRExC_state, data->start_class, &this_class);
2802 else if (flags & SCF_DO_STCLASS_AND) {
2803 /* Switch to OR mode: cache the old value of
2804 * data->start_class */
2805 StructCopy(data->start_class, &and_with,
2806 struct regnode_charclass_class);
2807 flags &= ~SCF_DO_STCLASS_AND;
2808 StructCopy(&this_class, data->start_class,
2809 struct regnode_charclass_class);
2810 flags |= SCF_DO_STCLASS_OR;
2811 data->start_class->flags |= ANYOF_EOS;
2813 } else { /* Non-zero len */
2814 if (flags & SCF_DO_STCLASS_OR) {
2815 cl_or(pRExC_state, data->start_class, &this_class);
2816 cl_and(data->start_class, &and_with);
2818 else if (flags & SCF_DO_STCLASS_AND)
2819 cl_and(data->start_class, &this_class);
2820 flags &= ~SCF_DO_STCLASS;
2822 if (!scan) /* It was not CURLYX, but CURLY. */
2824 if ( /* ? quantifier ok, except for (?{ ... }) */
2825 (next_is_eval || !(mincount == 0 && maxcount == 1))
2826 && (minnext == 0) && (deltanext == 0)
2827 && data && !(data->flags & (SF_HAS_PAR|SF_IN_PAR))
2828 && maxcount <= REG_INFTY/3 /* Complement check for big count */
2829 && ckWARN(WARN_REGEXP))
2832 "Quantifier unexpected on zero-length expression");
2835 min += minnext * mincount;
2836 is_inf_internal |= ((maxcount == REG_INFTY
2837 && (minnext + deltanext) > 0)
2838 || deltanext == I32_MAX);
2839 is_inf |= is_inf_internal;
2840 delta += (minnext + deltanext) * maxcount - minnext * mincount;
2842 /* Try powerful optimization CURLYX => CURLYN. */
2843 if ( OP(oscan) == CURLYX && data
2844 && data->flags & SF_IN_PAR
2845 && !(data->flags & SF_HAS_EVAL)
2846 && !deltanext && minnext == 1 ) {
2847 /* Try to optimize to CURLYN. */
2848 regnode *nxt = NEXTOPER(oscan) + EXTRA_STEP_2ARGS;
2849 regnode * const nxt1 = nxt;
2856 if (!strchr((const char*)PL_simple,OP(nxt))
2857 && !(PL_regkind[OP(nxt)] == EXACT
2858 && STR_LEN(nxt) == 1))
2864 if (OP(nxt) != CLOSE)
2866 /* Now we know that nxt2 is the only contents: */
2867 oscan->flags = (U8)ARG(nxt);
2869 OP(nxt1) = NOTHING; /* was OPEN. */
2871 OP(nxt1 + 1) = OPTIMIZED; /* was count. */
2872 NEXT_OFF(nxt1+ 1) = 0; /* just for consistancy. */
2873 NEXT_OFF(nxt2) = 0; /* just for consistancy with CURLY. */
2874 OP(nxt) = OPTIMIZED; /* was CLOSE. */
2875 OP(nxt + 1) = OPTIMIZED; /* was count. */
2876 NEXT_OFF(nxt+ 1) = 0; /* just for consistancy. */
2881 /* Try optimization CURLYX => CURLYM. */
2882 if ( OP(oscan) == CURLYX && data
2883 && !(data->flags & SF_HAS_PAR)
2884 && !(data->flags & SF_HAS_EVAL)
2885 && !deltanext /* atom is fixed width */
2886 && minnext != 0 /* CURLYM can't handle zero width */
2888 /* XXXX How to optimize if data == 0? */
2889 /* Optimize to a simpler form. */
2890 regnode *nxt = NEXTOPER(oscan) + EXTRA_STEP_2ARGS; /* OPEN */
2894 while ( (nxt2 = regnext(nxt)) /* skip over embedded stuff*/
2895 && (OP(nxt2) != WHILEM))
2897 OP(nxt2) = SUCCEED; /* Whas WHILEM */
2898 /* Need to optimize away parenths. */
2899 if (data->flags & SF_IN_PAR) {
2900 /* Set the parenth number. */
2901 regnode *nxt1 = NEXTOPER(oscan) + EXTRA_STEP_2ARGS; /* OPEN*/
2903 if (OP(nxt) != CLOSE)
2904 FAIL("Panic opt close");
2905 oscan->flags = (U8)ARG(nxt);
2906 OP(nxt1) = OPTIMIZED; /* was OPEN. */
2907 OP(nxt) = OPTIMIZED; /* was CLOSE. */
2909 OP(nxt1 + 1) = OPTIMIZED; /* was count. */
2910 OP(nxt + 1) = OPTIMIZED; /* was count. */
2911 NEXT_OFF(nxt1 + 1) = 0; /* just for consistancy. */
2912 NEXT_OFF(nxt + 1) = 0; /* just for consistancy. */
2915 while ( nxt1 && (OP(nxt1) != WHILEM)) {
2916 regnode *nnxt = regnext(nxt1);
2919 if (reg_off_by_arg[OP(nxt1)])
2920 ARG_SET(nxt1, nxt2 - nxt1);
2921 else if (nxt2 - nxt1 < U16_MAX)
2922 NEXT_OFF(nxt1) = nxt2 - nxt1;
2924 OP(nxt) = NOTHING; /* Cannot beautify */
2929 /* Optimize again: */
2930 study_chunk(pRExC_state, &nxt1, minlenp, &deltanext, nxt,
2936 else if ((OP(oscan) == CURLYX)
2937 && (flags & SCF_WHILEM_VISITED_POS)
2938 /* See the comment on a similar expression above.
2939 However, this time it not a subexpression
2940 we care about, but the expression itself. */
2941 && (maxcount == REG_INFTY)
2942 && data && ++data->whilem_c < 16) {
2943 /* This stays as CURLYX, we can put the count/of pair. */
2944 /* Find WHILEM (as in regexec.c) */
2945 regnode *nxt = oscan + NEXT_OFF(oscan);
2947 if (OP(PREVOPER(nxt)) == NOTHING) /* LONGJMP */
2949 PREVOPER(nxt)->flags = (U8)(data->whilem_c
2950 | (RExC_whilem_seen << 4)); /* On WHILEM */
2952 if (data && fl & (SF_HAS_PAR|SF_IN_PAR))
2954 if (flags & SCF_DO_SUBSTR) {
2955 SV *last_str = NULL;
2956 int counted = mincount != 0;
2958 if (data->last_end > 0 && mincount != 0) { /* Ends with a string. */
2959 #if defined(SPARC64_GCC_WORKAROUND)
2962 const char *s = NULL;
2965 if (pos_before >= data->last_start_min)
2968 b = data->last_start_min;
2971 s = SvPV_const(data->last_found, l);
2972 old = b - data->last_start_min;
2975 I32 b = pos_before >= data->last_start_min
2976 ? pos_before : data->last_start_min;
2978 const char * const s = SvPV_const(data->last_found, l);
2979 I32 old = b - data->last_start_min;
2983 old = utf8_hop((U8*)s, old) - (U8*)s;
2986 /* Get the added string: */
2987 last_str = newSVpvn(s + old, l);
2989 SvUTF8_on(last_str);
2990 if (deltanext == 0 && pos_before == b) {
2991 /* What was added is a constant string */
2993 SvGROW(last_str, (mincount * l) + 1);
2994 repeatcpy(SvPVX(last_str) + l,
2995 SvPVX_const(last_str), l, mincount - 1);
2996 SvCUR_set(last_str, SvCUR(last_str) * mincount);
2997 /* Add additional parts. */
2998 SvCUR_set(data->last_found,
2999 SvCUR(data->last_found) - l);
3000 sv_catsv(data->last_found, last_str);
3002 SV * sv = data->last_found;
3004 SvUTF8(sv) && SvMAGICAL(sv) ?
3005 mg_find(sv, PERL_MAGIC_utf8) : NULL;
3006 if (mg && mg->mg_len >= 0)
3007 mg->mg_len += CHR_SVLEN(last_str);
3009 data->last_end += l * (mincount - 1);
3012 /* start offset must point into the last copy */
3013 data->last_start_min += minnext * (mincount - 1);
3014 data->last_start_max += is_inf ? I32_MAX
3015 : (maxcount - 1) * (minnext + data->pos_delta);
3018 /* It is counted once already... */
3019 data->pos_min += minnext * (mincount - counted);
3020 data->pos_delta += - counted * deltanext +
3021 (minnext + deltanext) * maxcount - minnext * mincount;
3022 if (mincount != maxcount) {
3023 /* Cannot extend fixed substrings found inside
3025 scan_commit(pRExC_state,data,minlenp);
3026 if (mincount && last_str) {
3027 SV * const sv = data->last_found;
3028 MAGIC * const mg = SvUTF8(sv) && SvMAGICAL(sv) ?
3029 mg_find(sv, PERL_MAGIC_utf8) : NULL;
3033 sv_setsv(sv, last_str);
3034 data->last_end = data->pos_min;
3035 data->last_start_min =
3036 data->pos_min - CHR_SVLEN(last_str);
3037 data->last_start_max = is_inf
3039 : data->pos_min + data->pos_delta
3040 - CHR_SVLEN(last_str);
3042 data->longest = &(data->longest_float);
3044 SvREFCNT_dec(last_str);
3046 if (data && (fl & SF_HAS_EVAL))
3047 data->flags |= SF_HAS_EVAL;
3048 optimize_curly_tail:
3049 if (OP(oscan) != CURLYX) {
3050 while (PL_regkind[OP(next = regnext(oscan))] == NOTHING
3052 NEXT_OFF(oscan) += NEXT_OFF(next);
3055 default: /* REF and CLUMP only? */
3056 if (flags & SCF_DO_SUBSTR) {
3057 scan_commit(pRExC_state,data,minlenp); /* Cannot expect anything... */
3058 data->longest = &(data->longest_float);
3060 is_inf = is_inf_internal = 1;
3061 if (flags & SCF_DO_STCLASS_OR)
3062 cl_anything(pRExC_state, data->start_class);
3063 flags &= ~SCF_DO_STCLASS;
3067 else if (strchr((const char*)PL_simple,OP(scan))) {
3070 if (flags & SCF_DO_SUBSTR) {
3071 scan_commit(pRExC_state,data,minlenp);
3075 if (flags & SCF_DO_STCLASS) {
3076 data->start_class->flags &= ~ANYOF_EOS; /* No match on empty */
3078 /* Some of the logic below assumes that switching
3079 locale on will only add false positives. */
3080 switch (PL_regkind[OP(scan)]) {
3084 /* Perl_croak(aTHX_ "panic: unexpected simple REx opcode %d", OP(scan)); */
3085 if (flags & SCF_DO_STCLASS_OR) /* Allow everything */
3086 cl_anything(pRExC_state, data->start_class);
3089 if (OP(scan) == SANY)
3091 if (flags & SCF_DO_STCLASS_OR) { /* Everything but \n */
3092 value = (ANYOF_BITMAP_TEST(data->start_class,'\n')
3093 || (data->start_class->flags & ANYOF_CLASS));
3094 cl_anything(pRExC_state, data->start_class);
3096 if (flags & SCF_DO_STCLASS_AND || !value)
3097 ANYOF_BITMAP_CLEAR(data->start_class,'\n');
3100 if (flags & SCF_DO_STCLASS_AND)
3101 cl_and(data->start_class,
3102 (struct regnode_charclass_class*)scan);
3104 cl_or(pRExC_state, data->start_class,
3105 (struct regnode_charclass_class*)scan);
3108 if (flags & SCF_DO_STCLASS_AND) {
3109 if (!(data->start_class->flags & ANYOF_LOCALE)) {
3110 ANYOF_CLASS_CLEAR(data->start_class,ANYOF_NALNUM);
3111 for (value = 0; value < 256; value++)
3112 if (!isALNUM(value))
3113 ANYOF_BITMAP_CLEAR(data->start_class, value);
3117 if (data->start_class->flags & ANYOF_LOCALE)
3118 ANYOF_CLASS_SET(data->start_class,ANYOF_ALNUM);
3120 for (value = 0; value < 256; value++)
3122 ANYOF_BITMAP_SET(data->start_class, value);
3127 if (flags & SCF_DO_STCLASS_AND) {
3128 if (data->start_class->flags & ANYOF_LOCALE)
3129 ANYOF_CLASS_CLEAR(data->start_class,ANYOF_NALNUM);
3132 ANYOF_CLASS_SET(data->start_class,ANYOF_ALNUM);
3133 data->start_class->flags |= ANYOF_LOCALE;
3137 if (flags & SCF_DO_STCLASS_AND) {
3138 if (!(data->start_class->flags & ANYOF_LOCALE)) {
3139 ANYOF_CLASS_CLEAR(data->start_class,ANYOF_ALNUM);
3140 for (value = 0; value < 256; value++)
3142 ANYOF_BITMAP_CLEAR(data->start_class, value);
3146 if (data->start_class->flags & ANYOF_LOCALE)
3147 ANYOF_CLASS_SET(data->start_class,ANYOF_NALNUM);
3149 for (value = 0; value < 256; value++)
3150 if (!isALNUM(value))
3151 ANYOF_BITMAP_SET(data->start_class, value);
3156 if (flags & SCF_DO_STCLASS_AND) {
3157 if (data->start_class->flags & ANYOF_LOCALE)
3158 ANYOF_CLASS_CLEAR(data->start_class,ANYOF_ALNUM);
3161 data->start_class->flags |= ANYOF_LOCALE;
3162 ANYOF_CLASS_SET(data->start_class,ANYOF_NALNUM);
3166 if (flags & SCF_DO_STCLASS_AND) {
3167 if (!(data->start_class->flags & ANYOF_LOCALE)) {
3168 ANYOF_CLASS_CLEAR(data->start_class,ANYOF_NSPACE);
3169 for (value = 0; value < 256; value++)
3170 if (!isSPACE(value))
3171 ANYOF_BITMAP_CLEAR(data->start_class, value);
3175 if (data->start_class->flags & ANYOF_LOCALE)
3176 ANYOF_CLASS_SET(data->start_class,ANYOF_SPACE);
3178 for (value = 0; value < 256; value++)
3180 ANYOF_BITMAP_SET(data->start_class, value);
3185 if (flags & SCF_DO_STCLASS_AND) {
3186 if (data->start_class->flags & ANYOF_LOCALE)
3187 ANYOF_CLASS_CLEAR(data->start_class,ANYOF_NSPACE);
3190 data->start_class->flags |= ANYOF_LOCALE;
3191 ANYOF_CLASS_SET(data->start_class,ANYOF_SPACE);
3195 if (flags & SCF_DO_STCLASS_AND) {
3196 if (!(data->start_class->flags & ANYOF_LOCALE)) {
3197 ANYOF_CLASS_CLEAR(data->start_class,ANYOF_SPACE);
3198 for (value = 0; value < 256; value++)
3200 ANYOF_BITMAP_CLEAR(data->start_class, value);
3204 if (data->start_class->flags & ANYOF_LOCALE)
3205 ANYOF_CLASS_SET(data->start_class,ANYOF_NSPACE);
3207 for (value = 0; value < 256; value++)
3208 if (!isSPACE(value))
3209 ANYOF_BITMAP_SET(data->start_class, value);
3214 if (flags & SCF_DO_STCLASS_AND) {
3215 if (data->start_class->flags & ANYOF_LOCALE) {
3216 ANYOF_CLASS_CLEAR(data->start_class,ANYOF_SPACE);
3217 for (value = 0; value < 256; value++)
3218 if (!isSPACE(value))
3219 ANYOF_BITMAP_CLEAR(data->start_class, value);
3223 data->start_class->flags |= ANYOF_LOCALE;
3224 ANYOF_CLASS_SET(data->start_class,ANYOF_NSPACE);
3228 if (flags & SCF_DO_STCLASS_AND) {
3229 ANYOF_CLASS_CLEAR(data->start_class,ANYOF_NDIGIT);
3230 for (value = 0; value < 256; value++)
3231 if (!isDIGIT(value))
3232 ANYOF_BITMAP_CLEAR(data->start_class, value);
3235 if (data->start_class->flags & ANYOF_LOCALE)
3236 ANYOF_CLASS_SET(data->start_class,ANYOF_DIGIT);
3238 for (value = 0; value < 256; value++)
3240 ANYOF_BITMAP_SET(data->start_class, value);
3245 if (flags & SCF_DO_STCLASS_AND) {
3246 ANYOF_CLASS_CLEAR(data->start_class,ANYOF_DIGIT);
3247 for (value = 0; value < 256; value++)
3249 ANYOF_BITMAP_CLEAR(data->start_class, value);
3252 if (data->start_class->flags & ANYOF_LOCALE)
3253 ANYOF_CLASS_SET(data->start_class,ANYOF_NDIGIT);
3255 for (value = 0; value < 256; value++)
3256 if (!isDIGIT(value))
3257 ANYOF_BITMAP_SET(data->start_class, value);
3262 if (flags & SCF_DO_STCLASS_OR)
3263 cl_and(data->start_class, &and_with);
3264 flags &= ~SCF_DO_STCLASS;
3267 else if (PL_regkind[OP(scan)] == EOL && flags & SCF_DO_SUBSTR) {
3268 data->flags |= (OP(scan) == MEOL
3272 else if ( PL_regkind[OP(scan)] == BRANCHJ
3273 /* Lookbehind, or need to calculate parens/evals/stclass: */
3274 && (scan->flags || data || (flags & SCF_DO_STCLASS))
3275 && (OP(scan) == IFMATCH || OP(scan) == UNLESSM)) {
3276 if ( !PERL_ENABLE_POSITIVE_ASSERTION_STUDY
3277 || OP(scan) == UNLESSM )
3279 /* Negative Lookahead/lookbehind
3280 In this case we can't do fixed string optimisation.
3283 I32 deltanext, minnext, fake = 0;
3285 struct regnode_charclass_class intrnl;
3288 data_fake.flags = 0;
3290 data_fake.whilem_c = data->whilem_c;
3291 data_fake.last_closep = data->last_closep;
3294 data_fake.last_closep = &fake;
3295 if ( flags & SCF_DO_STCLASS && !scan->flags
3296 && OP(scan) == IFMATCH ) { /* Lookahead */
3297 cl_init(pRExC_state, &intrnl);
3298 data_fake.start_class = &intrnl;
3299 f |= SCF_DO_STCLASS_AND;
3301 if (flags & SCF_WHILEM_VISITED_POS)
3302 f |= SCF_WHILEM_VISITED_POS;
3303 next = regnext(scan);
3304 nscan = NEXTOPER(NEXTOPER(scan));
3305 minnext = study_chunk(pRExC_state, &nscan, minlenp, &deltanext, last, &data_fake, f,depth+1);
3308 vFAIL("Variable length lookbehind not implemented");
3310 else if (minnext > (I32)U8_MAX) {
3311 vFAIL2("Lookbehind longer than %"UVuf" not implemented", (UV)U8_MAX);
3313 scan->flags = (U8)minnext;
3316 if (data_fake.flags & (SF_HAS_PAR|SF_IN_PAR))
3318 if (data_fake.flags & SF_HAS_EVAL)
3319 data->flags |= SF_HAS_EVAL;
3320 data->whilem_c = data_fake.whilem_c;
3322 if (f & SCF_DO_STCLASS_AND) {
3323 const int was = (data->start_class->flags & ANYOF_EOS);
3325 cl_and(data->start_class, &intrnl);
3327 data->start_class->flags |= ANYOF_EOS;
3330 #if PERL_ENABLE_POSITIVE_ASSERTION_STUDY
3332 /* Positive Lookahead/lookbehind
3333 In this case we can do fixed string optimisation,
3334 but we must be careful about it. Note in the case of
3335 lookbehind the positions will be offset by the minimum
3336 length of the pattern, something we won't know about
3337 until after the recurse.
3339 I32 deltanext, fake = 0;
3341 struct regnode_charclass_class intrnl;
3343 /* We use SAVEFREEPV so that when the full compile
3344 is finished perl will clean up the allocated
3345 minlens when its all done. This was we don't
3346 have to worry about freeing them when we know
3347 they wont be used, which would be a pain.
3350 Newx( minnextp, 1, I32 );
3351 SAVEFREEPV(minnextp);
3354 StructCopy(data, &data_fake, scan_data_t);
3355 if ((flags & SCF_DO_SUBSTR) && data->last_found) {
3358 scan_commit(pRExC_state, &data_fake,minlenp);
3359 data_fake.last_found=newSVsv(data->last_found);
3363 data_fake.last_closep = &fake;
3364 data_fake.flags = 0;
3366 data_fake.flags |= SF_IS_INF;
3367 if ( flags & SCF_DO_STCLASS && !scan->flags
3368 && OP(scan) == IFMATCH ) { /* Lookahead */
3369 cl_init(pRExC_state, &intrnl);
3370 data_fake.start_class = &intrnl;
3371 f |= SCF_DO_STCLASS_AND;
3373 if (flags & SCF_WHILEM_VISITED_POS)
3374 f |= SCF_WHILEM_VISITED_POS;
3375 next = regnext(scan);
3376 nscan = NEXTOPER(NEXTOPER(scan));
3378 *minnextp = study_chunk(pRExC_state, &nscan, minnextp, &deltanext, last, &data_fake, f,depth+1);
3381 vFAIL("Variable length lookbehind not implemented");
3383 else if (*minnextp > (I32)U8_MAX) {
3384 vFAIL2("Lookbehind longer than %"UVuf" not implemented", (UV)U8_MAX);
3386 scan->flags = (U8)*minnextp;
3392 if (f & SCF_DO_STCLASS_AND) {
3393 const int was = (data->start_class->flags & ANYOF_EOS);
3395 cl_and(data->start_class, &intrnl);
3397 data->start_class->flags |= ANYOF_EOS;
3400 if (data_fake.flags & (SF_HAS_PAR|SF_IN_PAR))
3402 if (data_fake.flags & SF_HAS_EVAL)
3403 data->flags |= SF_HAS_EVAL;
3404 data->whilem_c = data_fake.whilem_c;
3405 if ((flags & SCF_DO_SUBSTR) && data_fake.last_found) {
3406 if (RExC_rx->minlen<*minnextp)
3407 RExC_rx->minlen=*minnextp;
3408 scan_commit(pRExC_state, &data_fake, minnextp);
3409 SvREFCNT_dec(data_fake.last_found);
3411 if ( data_fake.minlen_fixed != minlenp )
3413 data->offset_fixed= data_fake.offset_fixed;
3414 data->minlen_fixed= data_fake.minlen_fixed;
3415 data->lookbehind_fixed+= scan->flags;
3417 if ( data_fake.minlen_float != minlenp )
3419 data->minlen_float= data_fake.minlen_float;
3420 data->offset_float_min=data_fake.offset_float_min;
3421 data->offset_float_max=data_fake.offset_float_max;
3422 data->lookbehind_float+= scan->flags;
3431 else if (OP(scan) == OPEN) {
3434 else if (OP(scan) == CLOSE) {
3435 if ((I32)ARG(scan) == is_par) {
3436 next = regnext(scan);
3438 if ( next && (OP(next) != WHILEM) && next < last)
3439 is_par = 0; /* Disable optimization */
3442 *(data->last_closep) = ARG(scan);
3444 else if (OP(scan) == EVAL) {
3446 data->flags |= SF_HAS_EVAL;
3448 else if (OP(scan) == LOGICAL && scan->flags == 2) { /* Embedded follows */
3449 if (flags & SCF_DO_SUBSTR) {
3450 scan_commit(pRExC_state,data,minlenp);
3451 data->longest = &(data->longest_float);
3453 is_inf = is_inf_internal = 1;
3454 if (flags & SCF_DO_STCLASS_OR) /* Allow everything */
3455 cl_anything(pRExC_state, data->start_class);
3456 flags &= ~SCF_DO_STCLASS;
3458 #ifdef TRIE_STUDY_OPT
3459 #ifdef FULL_TRIE_STUDY
3460 else if (PL_regkind[OP(scan)] == TRIE) {
3461 /* NOTE - There is similar code to this block above for handling
3462 BRANCH nodes on the initial study. If you change stuff here
3464 regnode *tail= regnext(scan);
3465 reg_trie_data *trie = (reg_trie_data*)RExC_rx->data->data[ ARG(scan) ];
3466 I32 max1 = 0, min1 = I32_MAX;
3467 struct regnode_charclass_class accum;
3469 if (flags & SCF_DO_SUBSTR) /* XXXX Add !SUSPEND? */
3470 scan_commit(pRExC_state, data,minlenp); /* Cannot merge strings after this. */
3471 if (flags & SCF_DO_STCLASS)
3472 cl_init_zero(pRExC_state, &accum);
3478 const regnode *nextbranch= NULL;
3481 for ( word=1 ; word <= trie->wordcount ; word++)
3483 I32 deltanext=0, minnext=0, f = 0, fake;
3484 struct regnode_charclass_class this_class;
3486 data_fake.flags = 0;
3488 data_fake.whilem_c = data->whilem_c;
3489 data_fake.last_closep = data->last_closep;
3492 data_fake.last_closep = &fake;
3494 if (flags & SCF_DO_STCLASS) {
3495 cl_init(pRExC_state, &this_class);
3496 data_fake.start_class = &this_class;
3497 f = SCF_DO_STCLASS_AND;
3499 if (flags & SCF_WHILEM_VISITED_POS)
3500 f |= SCF_WHILEM_VISITED_POS;
3502 if (trie->jump[word]) {
3504 nextbranch = tail - trie->jump[0];
3505 scan= tail - trie->jump[word];
3506 /* We go from the jump point to the branch that follows
3507 it. Note this means we need the vestigal unused branches
3508 even though they arent otherwise used.
3510 minnext = study_chunk(pRExC_state, &scan, minlenp, &deltanext,
3511 (regnode *)nextbranch, &data_fake, f,depth+1);
3513 if (nextbranch && PL_regkind[OP(nextbranch)]==BRANCH)
3514 nextbranch= regnext((regnode*)nextbranch);
3516 if (min1 > (I32)(minnext + trie->minlen))
3517 min1 = minnext + trie->minlen;
3518 if (max1 < (I32)(minnext + deltanext + trie->maxlen))
3519 max1 = minnext + deltanext + trie->maxlen;
3520 if (deltanext == I32_MAX)
3521 is_inf = is_inf_internal = 1;
3523 if (data_fake.flags & (SF_HAS_PAR|SF_IN_PAR))
3527 if (data_fake.flags & SF_HAS_EVAL)
3528 data->flags |= SF_HAS_EVAL;
3529 data->whilem_c = data_fake.whilem_c;
3531 if (flags & SCF_DO_STCLASS)
3532 cl_or(pRExC_state, &accum, &this_class);
3535 if (flags & SCF_DO_SUBSTR) {
3536 data->pos_min += min1;
3537 data->pos_delta += max1 - min1;
3538 if (max1 != min1 || is_inf)
3539 data->longest = &(data->longest_float);
3542 delta += max1 - min1;
3543 if (flags & SCF_DO_STCLASS_OR) {
3544 cl_or(pRExC_state, data->start_class, &accum);
3546 cl_and(data->start_class, &and_with);
3547 flags &= ~SCF_DO_STCLASS;
3550 else if (flags & SCF_DO_STCLASS_AND) {
3552 cl_and(data->start_class, &accum);
3553 flags &= ~SCF_DO_STCLASS;
3556 /* Switch to OR mode: cache the old value of
3557 * data->start_class */
3558 StructCopy(data->start_class, &and_with,
3559 struct regnode_charclass_class);
3560 flags &= ~SCF_DO_STCLASS_AND;
3561 StructCopy(&accum, data->start_class,
3562 struct regnode_charclass_class);
3563 flags |= SCF_DO_STCLASS_OR;
3564 data->start_class->flags |= ANYOF_EOS;
3571 else if (PL_regkind[OP(scan)] == TRIE) {
3572 reg_trie_data *trie = (reg_trie_data*)RExC_rx->data->data[ ARG(scan) ];
3575 min += trie->minlen;
3576 delta += (trie->maxlen - trie->minlen);
3577 flags &= ~SCF_DO_STCLASS; /* xxx */
3578 if (flags & SCF_DO_SUBSTR) {
3579 scan_commit(pRExC_state,data,minlenp); /* Cannot expect anything... */
3580 data->pos_min += trie->minlen;
3581 data->pos_delta += (trie->maxlen - trie->minlen);
3582 if (trie->maxlen != trie->minlen)
3583 data->longest = &(data->longest_float);
3585 if (trie->jump) /* no more substrings -- for now /grr*/
3586 flags &= ~SCF_DO_SUBSTR;
3588 #endif /* old or new */
3589 #endif /* TRIE_STUDY_OPT */
3590 /* Else: zero-length, ignore. */
3591 scan = regnext(scan);
3596 *deltap = is_inf_internal ? I32_MAX : delta;
3597 if (flags & SCF_DO_SUBSTR && is_inf)
3598 data->pos_delta = I32_MAX - data->pos_min;
3599 if (is_par > (I32)U8_MAX)
3601 if (is_par && pars==1 && data) {
3602 data->flags |= SF_IN_PAR;
3603 data->flags &= ~SF_HAS_PAR;
3605 else if (pars && data) {
3606 data->flags |= SF_HAS_PAR;
3607 data->flags &= ~SF_IN_PAR;
3609 if (flags & SCF_DO_STCLASS_OR)
3610 cl_and(data->start_class, &and_with);
3611 if (flags & SCF_TRIE_RESTUDY)
3612 data->flags |= SCF_TRIE_RESTUDY;
3614 DEBUG_STUDYDATA(data,depth);
3620 S_add_data(RExC_state_t *pRExC_state, I32 n, const char *s)
3622 if (RExC_rx->data) {
3623 Renewc(RExC_rx->data,
3624 sizeof(*RExC_rx->data) + sizeof(void*) * (RExC_rx->data->count + n - 1),
3625 char, struct reg_data);
3626 Renew(RExC_rx->data->what, RExC_rx->data->count + n, U8);
3627 RExC_rx->data->count += n;
3630 Newxc(RExC_rx->data, sizeof(*RExC_rx->data) + sizeof(void*) * (n - 1),
3631 char, struct reg_data);
3632 Newx(RExC_rx->data->what, n, U8);
3633 RExC_rx->data->count = n;
3635 Copy(s, RExC_rx->data->what + RExC_rx->data->count - n, n, U8);
3636 return RExC_rx->data->count - n;
3639 #ifndef PERL_IN_XSUB_RE
3641 Perl_reginitcolors(pTHX)
3644 const char * const s = PerlEnv_getenv("PERL_RE_COLORS");
3646 char *t = savepv(s);
3650 t = strchr(t, '\t');
3656 PL_colors[i] = t = (char *)"";
3661 PL_colors[i++] = (char *)"";
3668 #ifdef TRIE_STUDY_OPT
3669 #define CHECK_RESTUDY_GOTO \
3671 (data.flags & SCF_TRIE_RESTUDY) \
3675 #define CHECK_RESTUDY_GOTO
3679 - pregcomp - compile a regular expression into internal code
3681 * We can't allocate space until we know how big the compiled form will be,
3682 * but we can't compile it (and thus know how big it is) until we've got a
3683 * place to put the code. So we cheat: we compile it twice, once with code
3684 * generation turned off and size counting turned on, and once "for real".
3685 * This also means that we don't allocate space until we are sure that the
3686 * thing really will compile successfully, and we never have to move the
3687 * code and thus invalidate pointers into it. (Note that it has to be in
3688 * one piece because free() must be able to free it all.) [NB: not true in perl]
3690 * Beware that the optimization-preparation code in here knows about some
3691 * of the structure of the compiled regexp. [I'll say.]
3693 #ifndef PERL_IN_XSUB_RE
3694 #define CORE_ONLY_BLOCK(c) {c}{
3695 #define RE_ENGINE_PTR &PL_core_reg_engine
3697 #define CORE_ONLY_BLOCK(c) {
3698 extern const struct regexp_engine my_reg_engine;
3699 #define RE_ENGINE_PTR &my_reg_engine
3704 Perl_pregcomp(pTHX_ char *exp, char *xend, PMOP *pm)
3707 GET_RE_DEBUG_FLAGS_DECL;
3708 DEBUG_r(if (!PL_colorset) reginitcolors());
3710 /* Dispatch a request to compile a regexp to correct
3712 HV * const table = GvHV(PL_hintgv);
3714 SV **ptr= hv_fetchs(table, "regcomp", FALSE);
3715 if (ptr && SvIOK(*ptr)) {
3716 const regexp_engine *eng=INT2PTR(regexp_engine*,SvIV(*ptr));
3718 PerlIO_printf(Perl_debug_log, "Using engine %"UVxf"\n",
3721 return CALLREGCOMP_ENG(eng, exp, xend, pm);
3732 RExC_state_t RExC_state;
3733 RExC_state_t * const pRExC_state = &RExC_state;
3734 #ifdef TRIE_STUDY_OPT
3736 RExC_state_t copyRExC_state;
3739 FAIL("NULL regexp argument");
3741 RExC_utf8 = pm->op_pmdynflags & PMdf_CMP_UTF8;
3745 SV *dsv= sv_newmortal();
3746 RE_PV_QUOTED_DECL(s, RExC_utf8,
3747 dsv, RExC_precomp, (xend - exp), 60);
3748 PerlIO_printf(Perl_debug_log, "%sCompiling REx%s %s\n",
3749 PL_colors[4],PL_colors[5],s);
3751 RExC_flags = pm->op_pmflags;
3755 RExC_seen_zerolen = *exp == '^' ? -1 : 0;
3756 RExC_seen_evals = 0;
3759 /* First pass: determine size, legality. */
3766 RExC_emit = &PL_regdummy;
3767 RExC_whilem_seen = 0;
3768 RExC_charnames = NULL;
3770 #if 0 /* REGC() is (currently) a NOP at the first pass.
3771 * Clever compilers notice this and complain. --jhi */
3772 REGC((U8)REG_MAGIC, (char*)RExC_emit);
3774 DEBUG_PARSE_r(PerlIO_printf(Perl_debug_log, "Starting first pass (sizing)\n"));
3775 if (reg(pRExC_state, 0, &flags,1) == NULL) {
3776 RExC_precomp = NULL;
3779 DEBUG_PARSE_r(PerlIO_printf(Perl_debug_log, "Required "));
3780 DEBUG_COMPILE_r(PerlIO_printf(Perl_debug_log, "size %"IVdf" nodes ", (IV)RExC_size));
3781 DEBUG_PARSE_r(PerlIO_printf(Perl_debug_log, "\nStarting second pass (creation)\n"));
3784 RExC_lastparse=NULL;
3788 /* Small enough for pointer-storage convention?
3789 If extralen==0, this means that we will not need long jumps. */
3790 if (RExC_size >= 0x10000L && RExC_extralen)
3791 RExC_size += RExC_extralen;
3794 if (RExC_whilem_seen > 15)
3795 RExC_whilem_seen = 15;
3797 /* Allocate space and zero-initialize. Note, the two step process
3798 of zeroing when in debug mode, thus anything assigned has to
3799 happen after that */
3800 Newxc(r, sizeof(regexp) + (unsigned)RExC_size * sizeof(regnode),
3803 FAIL("Regexp out of space");
3805 /* avoid reading uninitialized memory in DEBUGGING code in study_chunk() */
3806 Zero(r, sizeof(regexp) + (unsigned)RExC_size * sizeof(regnode), char);
3808 /* initialization begins here */
3809 r->engine= RE_ENGINE_PTR;
3811 r->prelen = xend - exp;
3812 r->precomp = savepvn(RExC_precomp, r->prelen);
3814 #ifdef PERL_OLD_COPY_ON_WRITE
3815 r->saved_copy = NULL;
3817 r->reganch = pm->op_pmflags & PMf_COMPILETIME;
3818 r->nparens = RExC_npar - 1; /* set early to validate backrefs */
3819 r->lastparen = 0; /* mg.c reads this. */
3821 r->substrs = 0; /* Useful during FAIL. */
3822 r->startp = 0; /* Useful during FAIL. */
3823 r->endp = 0; /* Useful during FAIL. */
3825 Newxz(r->offsets, 2*RExC_size+1, U32); /* MJD 20001228 */
3827 r->offsets[0] = RExC_size;
3829 DEBUG_OFFSETS_r(PerlIO_printf(Perl_debug_log,
3830 "%s %"UVuf" bytes for offset annotations.\n",
3831 r->offsets ? "Got" : "Couldn't get",
3832 (UV)((2*RExC_size+1) * sizeof(U32))));
3836 /* Second pass: emit code. */
3837 RExC_flags = pm->op_pmflags; /* don't let top level (?i) bleed */
3842 RExC_emit_start = r->program;
3843 RExC_emit = r->program;
3844 /* Store the count of eval-groups for security checks: */
3845 RExC_emit->next_off = (RExC_seen_evals > (I32)U16_MAX) ? U16_MAX : (U16)RExC_seen_evals;
3846 REGC((U8)REG_MAGIC, (char*) RExC_emit++);
3848 if (reg(pRExC_state, 0, &flags,1) == NULL)
3850 /* XXXX To minimize changes to RE engine we always allocate
3851 3-units-long substrs field. */
3852 Newx(r->substrs, 1, struct reg_substr_data);
3855 r->minlen = minlen = sawplus = sawopen = 0;
3856 Zero(r->substrs, 1, struct reg_substr_data);
3857 StructCopy(&zero_scan_data, &data, scan_data_t);
3859 #ifdef TRIE_STUDY_OPT
3861 DEBUG_OPTIMISE_r(PerlIO_printf(Perl_debug_log,"Restudying\n"));
3862 RExC_state=copyRExC_state;
3863 if (data.last_found) {
3864 SvREFCNT_dec(data.longest_fixed);
3865 SvREFCNT_dec(data.longest_float);
3866 SvREFCNT_dec(data.last_found);
3869 copyRExC_state=RExC_state;
3873 /* Dig out information for optimizations. */
3874 r->reganch = pm->op_pmflags & PMf_COMPILETIME; /* Again? */
3875 pm->op_pmflags = RExC_flags;
3877 r->reganch |= ROPT_UTF8; /* Unicode in it? */
3878 r->regstclass = NULL;
3879 if (RExC_naughty >= 10) /* Probably an expensive pattern. */
3880 r->reganch |= ROPT_NAUGHTY;
3881 scan = r->program + 1; /* First BRANCH. */
3883 /* testing for BRANCH here tells us whether there is "must appear"
3884 data in the pattern. If there is then we can use it for optimisations */
3885 if (OP(scan) != BRANCH) { /* Only one top-level choice. */
3887 STRLEN longest_float_length, longest_fixed_length;
3888 struct regnode_charclass_class ch_class; /* pointed to by data */
3890 I32 last_close = 0; /* pointed to by data */
3893 /* Skip introductions and multiplicators >= 1. */
3894 while ((OP(first) == OPEN && (sawopen = 1)) ||
3895 /* An OR of *one* alternative - should not happen now. */
3896 (OP(first) == BRANCH && OP(regnext(first)) != BRANCH) ||
3897 /* for now we can't handle lookbehind IFMATCH*/
3898 (OP(first) == IFMATCH && !first->flags) ||
3899 (OP(first) == PLUS) ||
3900 (OP(first) == MINMOD) ||
3901 /* An {n,m} with n>0 */
3902 (PL_regkind[OP(first)] == CURLY && ARG1(first) > 0) )
3905 if (OP(first) == PLUS)
3908 first += regarglen[OP(first)];
3909 if (OP(first) == IFMATCH) {
3910 first = NEXTOPER(first);
3911 first += EXTRA_STEP_2ARGS;
3912 } else /* XXX possible optimisation for /(?=)/ */
3913 first = NEXTOPER(first);
3916 /* Starting-point info. */
3918 DEBUG_PEEP("first:",first,0);
3919 /* Ignore EXACT as we deal with it later. */
3920 if (PL_regkind[OP(first)] == EXACT) {
3921 if (OP(first) == EXACT)
3922 NOOP; /* Empty, get anchored substr later. */
3923 else if ((OP(first) == EXACTF || OP(first) == EXACTFL))
3924 r->regstclass = first;
3927 else if (PL_regkind[OP(first)] == TRIE &&
3928 ((reg_trie_data *)r->data->data[ ARG(first) ])->minlen>0)
3931 /* this can happen only on restudy */
3932 if ( OP(first) == TRIE ) {
3933 struct regnode_1 *trieop;
3934 Newxz(trieop,1,struct regnode_1);
3935 StructCopy(first,trieop,struct regnode_1);
3936 trie_op=(regnode *)trieop;
3938 struct regnode_charclass *trieop;
3939 Newxz(trieop,1,struct regnode_charclass);
3940 StructCopy(first,trieop,struct regnode_charclass);
3941 trie_op=(regnode *)trieop;
3944 make_trie_failtable(pRExC_state, (regnode *)first, trie_op, 0);
3945 r->regstclass = trie_op;
3948 else if (strchr((const char*)PL_simple,OP(first)))
3949 r->regstclass = first;
3950 else if (PL_regkind[OP(first)] == BOUND ||
3951 PL_regkind[OP(first)] == NBOUND)
3952 r->regstclass = first;
3953 else if (PL_regkind[OP(first)] == BOL) {
3954 r->reganch |= (OP(first) == MBOL
3956 : (OP(first) == SBOL
3959 first = NEXTOPER(first);
3962 else if (OP(first) == GPOS) {
3963 r->reganch |= ROPT_ANCH_GPOS;
3964 first = NEXTOPER(first);
3967 else if (!sawopen && (OP(first) == STAR &&
3968 PL_regkind[OP(NEXTOPER(first))] == REG_ANY) &&
3969 !(r->reganch & ROPT_ANCH) )
3971 /* turn .* into ^.* with an implied $*=1 */
3973 (OP(NEXTOPER(first)) == REG_ANY)
3976 r->reganch |= type | ROPT_IMPLICIT;
3977 first = NEXTOPER(first);
3980 if (sawplus && (!sawopen || !RExC_sawback)
3981 && !(RExC_seen & REG_SEEN_EVAL)) /* May examine pos and $& */
3982 /* x+ must match at the 1st pos of run of x's */
3983 r->reganch |= ROPT_SKIP;
3985 /* Scan is after the zeroth branch, first is atomic matcher. */
3986 #ifdef TRIE_STUDY_OPT
3989 PerlIO_printf(Perl_debug_log, "first at %"IVdf"\n",
3990 (IV)(first - scan + 1))
3994 PerlIO_printf(Perl_debug_log, "first at %"IVdf"\n",
3995 (IV)(first - scan + 1))
4001 * If there's something expensive in the r.e., find the
4002 * longest literal string that must appear and make it the
4003 * regmust. Resolve ties in favor of later strings, since
4004 * the regstart check works with the beginning of the r.e.
4005 * and avoiding duplication strengthens checking. Not a
4006 * strong reason, but sufficient in the absence of others.
4007 * [Now we resolve ties in favor of the earlier string if
4008 * it happens that c_offset_min has been invalidated, since the
4009 * earlier string may buy us something the later one won't.]
4013 data.longest_fixed = newSVpvs("");
4014 data.longest_float = newSVpvs("");
4015 data.last_found = newSVpvs("");
4016 data.longest = &(data.longest_fixed);
4018 if (!r->regstclass) {
4019 cl_init(pRExC_state, &ch_class);
4020 data.start_class = &ch_class;
4021 stclass_flag = SCF_DO_STCLASS_AND;
4022 } else /* XXXX Check for BOUND? */
4024 data.last_closep = &last_close;
4026 minlen = study_chunk(pRExC_state, &first, &minlen, &fake, scan + RExC_size, /* Up to end */
4027 &data, SCF_DO_SUBSTR | SCF_WHILEM_VISITED_POS | stclass_flag,0);
4033 if ( RExC_npar == 1 && data.longest == &(data.longest_fixed)
4034 && data.last_start_min == 0 && data.last_end > 0
4035 && !RExC_seen_zerolen
4036 && (!(RExC_seen & REG_SEEN_GPOS) || (r->reganch & ROPT_ANCH_GPOS)))
4037 r->reganch |= ROPT_CHECK_ALL;
4038 scan_commit(pRExC_state, &data,&minlen);
4039 SvREFCNT_dec(data.last_found);
4041 /* Note that code very similar to this but for anchored string
4042 follows immediately below, changes may need to be made to both.
4045 longest_float_length = CHR_SVLEN(data.longest_float);
4046 if (longest_float_length
4047 || (data.flags & SF_FL_BEFORE_EOL
4048 && (!(data.flags & SF_FL_BEFORE_MEOL)
4049 || (RExC_flags & PMf_MULTILINE))))
4053 if (SvCUR(data.longest_fixed) /* ok to leave SvCUR */
4054 && data.offset_fixed == data.offset_float_min
4055 && SvCUR(data.longest_fixed) == SvCUR(data.longest_float))
4056 goto remove_float; /* As in (a)+. */
4058 /* copy the information about the longest float from the reg_scan_data
4059 over to the program. */
4060 if (SvUTF8(data.longest_float)) {
4061 r->float_utf8 = data.longest_float;
4062 r->float_substr = NULL;
4064 r->float_substr = data.longest_float;
4065 r->float_utf8 = NULL;
4067 /* float_end_shift is how many chars that must be matched that
4068 follow this item. We calculate it ahead of time as once the
4069 lookbehind offset is added in we lose the ability to correctly
4071 ml = data.minlen_float ? *(data.minlen_float)
4072 : (I32)longest_float_length;
4073 r->float_end_shift = ml - data.offset_float_min
4074 - longest_float_length + (SvTAIL(data.longest_float) != 0)
4075 + data.lookbehind_float;
4076 r->float_min_offset = data.offset_float_min - data.lookbehind_float;
4077 r->float_max_offset = data.offset_float_max;
4078 if (data.offset_float_max < I32_MAX) /* Don't offset infinity */
4079 r->float_max_offset -= data.lookbehind_float;
4081 t = (data.flags & SF_FL_BEFORE_EOL /* Can't have SEOL and MULTI */
4082 && (!(data.flags & SF_FL_BEFORE_MEOL)
4083 || (RExC_flags & PMf_MULTILINE)));
4084 fbm_compile(data.longest_float, t ? FBMcf_TAIL : 0);
4088 r->float_substr = r->float_utf8 = NULL;
4089 SvREFCNT_dec(data.longest_float);
4090 longest_float_length = 0;
4093 /* Note that code very similar to this but for floating string
4094 is immediately above, changes may need to be made to both.
4097 longest_fixed_length = CHR_SVLEN(data.longest_fixed);
4098 if (longest_fixed_length
4099 || (data.flags & SF_FIX_BEFORE_EOL /* Cannot have SEOL and MULTI */
4100 && (!(data.flags & SF_FIX_BEFORE_MEOL)
4101 || (RExC_flags & PMf_MULTILINE))))
4105 /* copy the information about the longest fixed
4106 from the reg_scan_data over to the program. */
4107 if (SvUTF8(data.longest_fixed)) {
4108 r->anchored_utf8 = data.longest_fixed;
4109 r->anchored_substr = NULL;
4111 r->anchored_substr = data.longest_fixed;
4112 r->anchored_utf8 = NULL;
4114 /* fixed_end_shift is how many chars that must be matched that
4115 follow this item. We calculate it ahead of time as once the
4116 lookbehind offset is added in we lose the ability to correctly
4118 ml = data.minlen_fixed ? *(data.minlen_fixed)
4119 : (I32)longest_fixed_length;
4120 r->anchored_end_shift = ml - data.offset_fixed
4121 - longest_fixed_length + (SvTAIL(data.longest_fixed) != 0)
4122 + data.lookbehind_fixed;
4123 r->anchored_offset = data.offset_fixed - data.lookbehind_fixed;
4125 t = (data.flags & SF_FIX_BEFORE_EOL /* Can't have SEOL and MULTI */
4126 && (!(data.flags & SF_FIX_BEFORE_MEOL)
4127 || (RExC_flags & PMf_MULTILINE)));
4128 fbm_compile(data.longest_fixed, t ? FBMcf_TAIL : 0);
4131 r->anchored_substr = r->anchored_utf8 = NULL;
4132 SvREFCNT_dec(data.longest_fixed);
4133 longest_fixed_length = 0;
4136 && (OP(r->regstclass) == REG_ANY || OP(r->regstclass) == SANY))
4137 r->regstclass = NULL;
4138 if ((!(r->anchored_substr || r->anchored_utf8) || r->anchored_offset)
4140 && !(data.start_class->flags & ANYOF_EOS)
4141 && !cl_is_anything(data.start_class))
4143 const I32 n = add_data(pRExC_state, 1, "f");
4145 Newx(RExC_rx->data->data[n], 1,
4146 struct regnode_charclass_class);
4147 StructCopy(data.start_class,
4148 (struct regnode_charclass_class*)RExC_rx->data->data[n],
4149 struct regnode_charclass_class);
4150 r->regstclass = (regnode*)RExC_rx->data->data[n];
4151 r->reganch &= ~ROPT_SKIP; /* Used in find_byclass(). */
4152 DEBUG_COMPILE_r({ SV *sv = sv_newmortal();
4153 regprop(r, sv, (regnode*)data.start_class);
4154 PerlIO_printf(Perl_debug_log,
4155 "synthetic stclass \"%s\".\n",
4156 SvPVX_const(sv));});
4159 /* A temporary algorithm prefers floated substr to fixed one to dig more info. */
4160 if (longest_fixed_length > longest_float_length) {
4161 r->check_end_shift = r->anchored_end_shift;
4162 r->check_substr = r->anchored_substr;
4163 r->check_utf8 = r->anchored_utf8;
4164 r->check_offset_min = r->check_offset_max = r->anchored_offset;
4165 if (r->reganch & ROPT_ANCH_SINGLE)
4166 r->reganch |= ROPT_NOSCAN;
4169 r->check_end_shift = r->float_end_shift;
4170 r->check_substr = r->float_substr;
4171 r->check_utf8 = r->float_utf8;
4172 r->check_offset_min = r->float_min_offset;
4173 r->check_offset_max = r->float_max_offset;
4175 /* XXXX Currently intuiting is not compatible with ANCH_GPOS.
4176 This should be changed ASAP! */
4177 if ((r->check_substr || r->check_utf8) && !(r->reganch & ROPT_ANCH_GPOS)) {
4178 r->reganch |= RE_USE_INTUIT;
4179 if (SvTAIL(r->check_substr ? r->check_substr : r->check_utf8))
4180 r->reganch |= RE_INTUIT_TAIL;
4182 /* XXX Unneeded? dmq (shouldn't as this is handled elsewhere)
4183 if ( (STRLEN)minlen < longest_float_length )
4184 minlen= longest_float_length;
4185 if ( (STRLEN)minlen < longest_fixed_length )
4186 minlen= longest_fixed_length;
4190 /* Several toplevels. Best we can is to set minlen. */
4192 struct regnode_charclass_class ch_class;
4195 DEBUG_COMPILE_r(PerlIO_printf(Perl_debug_log, "\n"));
4197 scan = r->program + 1;
4198 cl_init(pRExC_state, &ch_class);
4199 data.start_class = &ch_class;
4200 data.last_closep = &last_close;
4202 minlen = study_chunk(pRExC_state, &scan, &minlen, &fake, scan + RExC_size,
4203 &data, SCF_DO_STCLASS_AND|SCF_WHILEM_VISITED_POS,0);
4207 r->check_substr = r->check_utf8 = r->anchored_substr = r->anchored_utf8
4208 = r->float_substr = r->float_utf8 = NULL;
4209 if (!(data.start_class->flags & ANYOF_EOS)
4210 && !cl_is_anything(data.start_class))
4212 const I32 n = add_data(pRExC_state, 1, "f");
4214 Newx(RExC_rx->data->data[n], 1,
4215 struct regnode_charclass_class);
4216 StructCopy(data.start_class,
4217 (struct regnode_charclass_class*)RExC_rx->data->data[n],
4218 struct regnode_charclass_class);
4219 r->regstclass = (regnode*)RExC_rx->data->data[n];
4220 r->reganch &= ~ROPT_SKIP; /* Used in find_byclass(). */
4221 DEBUG_COMPILE_r({ SV* sv = sv_newmortal();
4222 regprop(r, sv, (regnode*)data.start_class);
4223 PerlIO_printf(Perl_debug_log,
4224 "synthetic stclass \"%s\".\n",
4225 SvPVX_const(sv));});
4229 /* Guard against an embedded (?=) or (?<=) with a longer minlen than
4230 the "real" pattern. */
4231 if (r->minlen < minlen)
4234 if (RExC_seen & REG_SEEN_GPOS)
4235 r->reganch |= ROPT_GPOS_SEEN;
4236 if (RExC_seen & REG_SEEN_LOOKBEHIND)
4237 r->reganch |= ROPT_LOOKBEHIND_SEEN;
4238 if (RExC_seen & REG_SEEN_EVAL)
4239 r->reganch |= ROPT_EVAL_SEEN;
4240 if (RExC_seen & REG_SEEN_CANY)
4241 r->reganch |= ROPT_CANY_SEEN;
4242 Newxz(r->startp, RExC_npar, I32);
4243 Newxz(r->endp, RExC_npar, I32);
4247 SvREFCNT_dec((SV*)(RExC_charnames));
4249 DEBUG_r( RX_DEBUG_on(r) );
4251 PerlIO_printf(Perl_debug_log,"Final program:\n");
4254 DEBUG_OFFSETS_r(if (r->offsets) {
4255 const U32 len = r->offsets[0];
4257 GET_RE_DEBUG_FLAGS_DECL;
4258 PerlIO_printf(Perl_debug_log, "Offsets: [%"UVuf"]\n\t", (UV)r->offsets[0]);
4259 for (i = 1; i <= len; i++) {
4260 if (r->offsets[i*2-1] || r->offsets[i*2])
4261 PerlIO_printf(Perl_debug_log, "%"UVuf":%"UVuf"[%"UVuf"] ",
4262 (UV)i, (UV)r->offsets[i*2-1], (UV)r->offsets[i*2]);
4264 PerlIO_printf(Perl_debug_log, "\n");
4270 #undef CORE_ONLY_BLOCK
4272 #undef RE_ENGINE_PTR
4274 #define DEBUG_PARSE_MSG(funcname) DEBUG_PARSE_r({ \
4275 int rem=(int)(RExC_end - RExC_parse); \
4284 if (RExC_lastparse!=RExC_parse) \
4285 PerlIO_printf(Perl_debug_log," >%.*s%-*s", \
4288 iscut ? "..." : "<" \
4291 PerlIO_printf(Perl_debug_log,"%16s",""); \
4296 num=REG_NODE_NUM(RExC_emit); \
4297 if (RExC_lastnum!=num) \
4298 PerlIO_printf(Perl_debug_log,"|%4d",num); \
4300 PerlIO_printf(Perl_debug_log,"|%4s",""); \
4301 PerlIO_printf(Perl_debug_log,"|%*s%-4s", \
4302 (int)((depth*2)), "", \
4306 RExC_lastparse=RExC_parse; \
4311 #define DEBUG_PARSE(funcname) DEBUG_PARSE_r({ \
4312 DEBUG_PARSE_MSG((funcname)); \
4313 PerlIO_printf(Perl_debug_log,"%4s","\n"); \
4316 - reg - regular expression, i.e. main body or parenthesized thing
4318 * Caller must absorb opening parenthesis.
4320 * Combining parenthesis handling with the base level of regular expression
4321 * is a trifle forced, but the need to tie the tails of the branches to what
4322 * follows makes it hard to avoid.
4324 #define REGTAIL(x,y,z) regtail((x),(y),(z),depth+1)
4326 #define REGTAIL_STUDY(x,y,z) regtail_study((x),(y),(z),depth+1)
4328 #define REGTAIL_STUDY(x,y,z) regtail((x),(y),(z),depth+1)
4332 S_reg(pTHX_ RExC_state_t *pRExC_state, I32 paren, I32 *flagp,U32 depth)
4333 /* paren: Parenthesized? 0=top, 1=(, inside: changed to letter. */
4336 register regnode *ret; /* Will be the head of the group. */
4337 register regnode *br;
4338 register regnode *lastbr;
4339 register regnode *ender = NULL;
4340 register I32 parno = 0;
4342 const I32 oregflags = RExC_flags;
4343 bool have_branch = 0;
4346 /* for (?g), (?gc), and (?o) warnings; warning
4347 about (?c) will warn about (?g) -- japhy */
4349 #define WASTED_O 0x01
4350 #define WASTED_G 0x02
4351 #define WASTED_C 0x04
4352 #define WASTED_GC (0x02|0x04)
4353 I32 wastedflags = 0x00;
4355 char * parse_start = RExC_parse; /* MJD */
4356 char * const oregcomp_parse = RExC_parse;
4358 GET_RE_DEBUG_FLAGS_DECL;
4359 DEBUG_PARSE("reg ");
4362 *flagp = 0; /* Tentatively. */
4365 /* Make an OPEN node, if parenthesized. */
4367 if (*RExC_parse == '?') { /* (?...) */
4368 U32 posflags = 0, negflags = 0;
4369 U32 *flagsp = &posflags;
4370 bool is_logical = 0;
4371 const char * const seqstart = RExC_parse;
4374 paren = *RExC_parse++;
4375 ret = NULL; /* For look-ahead/behind. */
4377 case '<': /* (?<...) */
4378 RExC_seen |= REG_SEEN_LOOKBEHIND;
4379 if (*RExC_parse == '!')
4381 if (*RExC_parse != '=' && *RExC_parse != '!')
4384 case '=': /* (?=...) */
4385 case '!': /* (?!...) */
4386 RExC_seen_zerolen++;
4387 case ':': /* (?:...) */
4388 case '>': /* (?>...) */
4390 case '$': /* (?$...) */
4391 case '@': /* (?@...) */
4392 vFAIL2("Sequence (?%c...) not implemented", (int)paren);
4394 case '#': /* (?#...) */
4395 while (*RExC_parse && *RExC_parse != ')')
4397 if (*RExC_parse != ')')
4398 FAIL("Sequence (?#... not terminated");
4399 nextchar(pRExC_state);
4402 case 'p': /* (?p...) */
4403 if (SIZE_ONLY && ckWARN2(WARN_DEPRECATED, WARN_REGEXP))
4404 vWARNdep(RExC_parse, "(?p{}) is deprecated - use (??{})");
4406 case '?': /* (??...) */
4408 if (*RExC_parse != '{')
4410 paren = *RExC_parse++;
4412 case '{': /* (?{...}) */
4414 I32 count = 1, n = 0;
4416 char *s = RExC_parse;
4418 RExC_seen_zerolen++;
4419 RExC_seen |= REG_SEEN_EVAL;
4420 while (count && (c = *RExC_parse)) {
4431 if (*RExC_parse != ')') {
4433 vFAIL("Sequence (?{...}) not terminated or not {}-balanced");
4437 OP_4tree *sop, *rop;
4438 SV * const sv = newSVpvn(s, RExC_parse - 1 - s);
4441 Perl_save_re_context(aTHX);
4442 rop = sv_compile_2op(sv, &sop, "re", &pad);
4443 sop->op_private |= OPpREFCOUNTED;
4444 /* re_dup will OpREFCNT_inc */
4445 OpREFCNT_set(sop, 1);
4448 n = add_data(pRExC_state, 3, "nop");
4449 RExC_rx->data->data[n] = (void*)rop;
4450 RExC_rx->data->data[n+1] = (void*)sop;
4451 RExC_rx->data->data[n+2] = (void*)pad;
4454 else { /* First pass */
4455 if (PL_reginterp_cnt < ++RExC_seen_evals
4457 /* No compiled RE interpolated, has runtime
4458 components ===> unsafe. */
4459 FAIL("Eval-group not allowed at runtime, use re 'eval'");
4460 if (PL_tainting && PL_tainted)
4461 FAIL("Eval-group in insecure regular expression");
4462 #if PERL_VERSION > 8
4463 if (IN_PERL_COMPILETIME)
4468 nextchar(pRExC_state);
4470 ret = reg_node(pRExC_state, LOGICAL);
4473 REGTAIL(pRExC_state, ret, reganode(pRExC_state, EVAL, n));
4474 /* deal with the length of this later - MJD */
4477 ret = reganode(pRExC_state, EVAL, n);
4478 Set_Node_Length(ret, RExC_parse - parse_start + 1);
4479 Set_Node_Offset(ret, parse_start);
4482 case '(': /* (?(?{...})...) and (?(?=...)...) */
4484 if (RExC_parse[0] == '?') { /* (?(?...)) */
4485 if (RExC_parse[1] == '=' || RExC_parse[1] == '!'
4486 || RExC_parse[1] == '<'
4487 || RExC_parse[1] == '{') { /* Lookahead or eval. */
4490 ret = reg_node(pRExC_state, LOGICAL);
4493 REGTAIL(pRExC_state, ret, reg(pRExC_state, 1, &flag,depth+1));
4497 else if (RExC_parse[0] >= '1' && RExC_parse[0] <= '9' ) {
4500 parno = atoi(RExC_parse++);
4502 while (isDIGIT(*RExC_parse))
4504 ret = reganode(pRExC_state, GROUPP, parno);
4506 if ((c = *nextchar(pRExC_state)) != ')')
4507 vFAIL("Switch condition not recognized");
4509 REGTAIL(pRExC_state, ret, reganode(pRExC_state, IFTHEN, 0));
4510 br = regbranch(pRExC_state, &flags, 1,depth+1);
4512 br = reganode(pRExC_state, LONGJMP, 0);
4514 REGTAIL(pRExC_state, br, reganode(pRExC_state, LONGJMP, 0));
4515 c = *nextchar(pRExC_state);
4519 lastbr = reganode(pRExC_state, IFTHEN, 0); /* Fake one for optimizer. */
4520 regbranch(pRExC_state, &flags, 1,depth+1);
4521 REGTAIL(pRExC_state, ret, lastbr);
4524 c = *nextchar(pRExC_state);
4529 vFAIL("Switch (?(condition)... contains too many branches");
4530 ender = reg_node(pRExC_state, TAIL);
4531 REGTAIL(pRExC_state, br, ender);
4533 REGTAIL(pRExC_state, lastbr, ender);
4534 REGTAIL(pRExC_state, NEXTOPER(NEXTOPER(lastbr)), ender);
4537 REGTAIL(pRExC_state, ret, ender);
4541 vFAIL2("Unknown switch condition (?(%.2s", RExC_parse);
4545 RExC_parse--; /* for vFAIL to print correctly */
4546 vFAIL("Sequence (? incomplete");
4550 parse_flags: /* (?i) */
4551 while (*RExC_parse && strchr("iogcmsx", *RExC_parse)) {
4552 /* (?g), (?gc) and (?o) are useless here
4553 and must be globally applied -- japhy */
4555 if (*RExC_parse == 'o' || *RExC_parse == 'g') {
4556 if (SIZE_ONLY && ckWARN(WARN_REGEXP)) {
4557 const I32 wflagbit = *RExC_parse == 'o' ? WASTED_O : WASTED_G;
4558 if (! (wastedflags & wflagbit) ) {
4559 wastedflags |= wflagbit;
4562 "Useless (%s%c) - %suse /%c modifier",
4563 flagsp == &negflags ? "?-" : "?",
4565 flagsp == &negflags ? "don't " : "",
4571 else if (*RExC_parse == 'c') {
4572 if (SIZE_ONLY && ckWARN(WARN_REGEXP)) {
4573 if (! (wastedflags & WASTED_C) ) {
4574 wastedflags |= WASTED_GC;
4577 "Useless (%sc) - %suse /gc modifier",
4578 flagsp == &negflags ? "?-" : "?",
4579 flagsp == &negflags ? "don't " : ""
4584 else { pmflag(flagsp, *RExC_parse); }
4588 if (*RExC_parse == '-') {
4590 wastedflags = 0; /* reset so (?g-c) warns twice */
4594 RExC_flags |= posflags;
4595 RExC_flags &= ~negflags;
4596 if (*RExC_parse == ':') {
4602 if (*RExC_parse != ')') {
4604 vFAIL3("Sequence (%.*s...) not recognized", RExC_parse-seqstart, seqstart);
4606 nextchar(pRExC_state);
4614 ret = reganode(pRExC_state, OPEN, parno);
4615 Set_Node_Length(ret, 1); /* MJD */
4616 Set_Node_Offset(ret, RExC_parse); /* MJD */
4623 /* Pick up the branches, linking them together. */
4624 parse_start = RExC_parse; /* MJD */
4625 br = regbranch(pRExC_state, &flags, 1,depth+1);
4626 /* branch_len = (paren != 0); */
4630 if (*RExC_parse == '|') {
4631 if (!SIZE_ONLY && RExC_extralen) {
4632 reginsert(pRExC_state, BRANCHJ, br);
4635 reginsert(pRExC_state, BRANCH, br);
4636 Set_Node_Length(br, paren != 0);
4637 Set_Node_Offset_To_R(br-RExC_emit_start, parse_start-RExC_start);
4641 RExC_extralen += 1; /* For BRANCHJ-BRANCH. */
4643 else if (paren == ':') {
4644 *flagp |= flags&SIMPLE;
4646 if (is_open) { /* Starts with OPEN. */
4647 REGTAIL(pRExC_state, ret, br); /* OPEN -> first. */
4649 else if (paren != '?') /* Not Conditional */
4651 *flagp |= flags & (SPSTART | HASWIDTH);
4653 while (*RExC_parse == '|') {
4654 if (!SIZE_ONLY && RExC_extralen) {
4655 ender = reganode(pRExC_state, LONGJMP,0);
4656 REGTAIL(pRExC_state, NEXTOPER(NEXTOPER(lastbr)), ender); /* Append to the previous. */
4659 RExC_extralen += 2; /* Account for LONGJMP. */
4660 nextchar(pRExC_state);
4661 br = regbranch(pRExC_state, &flags, 0, depth+1);
4665 REGTAIL(pRExC_state, lastbr, br); /* BRANCH -> BRANCH. */
4669 *flagp |= flags&SPSTART;
4672 if (have_branch || paren != ':') {
4673 /* Make a closing node, and hook it on the end. */
4676 ender = reg_node(pRExC_state, TAIL);
4679 ender = reganode(pRExC_state, CLOSE, parno);
4680 Set_Node_Offset(ender,RExC_parse+1); /* MJD */
4681 Set_Node_Length(ender,1); /* MJD */
4687 *flagp &= ~HASWIDTH;
4690 ender = reg_node(pRExC_state, SUCCEED);
4693 ender = reg_node(pRExC_state, END);
4696 REGTAIL_STUDY(pRExC_state, lastbr, ender);
4698 if (have_branch && !SIZE_ONLY) {
4699 /* Hook the tails of the branches to the closing node. */
4700 for (br = ret; br; br = regnext(br)) {
4701 const U8 op = PL_regkind[OP(br)];
4703 REGTAIL_STUDY(pRExC_state, NEXTOPER(br), ender);
4705 else if (op == BRANCHJ) {
4706 REGTAIL_STUDY(pRExC_state, NEXTOPER(NEXTOPER(br)), ender);
4714 static const char parens[] = "=!<,>";
4716 if (paren && (p = strchr(parens, paren))) {
4717 U8 node = ((p - parens) % 2) ? UNLESSM : IFMATCH;
4718 int flag = (p - parens) > 1;
4721 node = SUSPEND, flag = 0;
4722 reginsert(pRExC_state, node,ret);
4723 Set_Node_Cur_Length(ret);
4724 Set_Node_Offset(ret, parse_start + 1);
4726 REGTAIL_STUDY(pRExC_state, ret, reg_node(pRExC_state, TAIL));
4730 /* Check for proper termination. */
4732 RExC_flags = oregflags;
4733 if (RExC_parse >= RExC_end || *nextchar(pRExC_state) != ')') {
4734 RExC_parse = oregcomp_parse;
4735 vFAIL("Unmatched (");
4738 else if (!paren && RExC_parse < RExC_end) {
4739 if (*RExC_parse == ')') {
4741 vFAIL("Unmatched )");
4744 FAIL("Junk on end of regexp"); /* "Can't happen". */
4752 - regbranch - one alternative of an | operator
4754 * Implements the concatenation operator.
4757 S_regbranch(pTHX_ RExC_state_t *pRExC_state, I32 *flagp, I32 first, U32 depth)
4760 register regnode *ret;
4761 register regnode *chain = NULL;
4762 register regnode *latest;
4763 I32 flags = 0, c = 0;
4764 GET_RE_DEBUG_FLAGS_DECL;
4765 DEBUG_PARSE("brnc");
4769 if (!SIZE_ONLY && RExC_extralen)
4770 ret = reganode(pRExC_state, BRANCHJ,0);
4772 ret = reg_node(pRExC_state, BRANCH);
4773 Set_Node_Length(ret, 1);
4777 if (!first && SIZE_ONLY)
4778 RExC_extralen += 1; /* BRANCHJ */
4780 *flagp = WORST; /* Tentatively. */
4783 nextchar(pRExC_state);
4784 while (RExC_parse < RExC_end && *RExC_parse != '|' && *RExC_parse != ')') {
4786 latest = regpiece(pRExC_state, &flags,depth+1);
4787 if (latest == NULL) {
4788 if (flags & TRYAGAIN)
4792 else if (ret == NULL)
4794 *flagp |= flags&HASWIDTH;
4795 if (chain == NULL) /* First piece. */
4796 *flagp |= flags&SPSTART;
4799 REGTAIL(pRExC_state, chain, latest);
4804 if (chain == NULL) { /* Loop ran zero times. */
4805 chain = reg_node(pRExC_state, NOTHING);
4810 *flagp |= flags&SIMPLE;
4817 - regpiece - something followed by possible [*+?]
4819 * Note that the branching code sequences used for ? and the general cases
4820 * of * and + are somewhat optimized: they use the same NOTHING node as
4821 * both the endmarker for their branch list and the body of the last branch.
4822 * It might seem that this node could be dispensed with entirely, but the
4823 * endmarker role is not redundant.
4826 S_regpiece(pTHX_ RExC_state_t *pRExC_state, I32 *flagp, U32 depth)
4829 register regnode *ret;
4831 register char *next;
4833 const char * const origparse = RExC_parse;
4835 I32 max = REG_INFTY;
4837 const char *maxpos = NULL;
4838 GET_RE_DEBUG_FLAGS_DECL;
4839 DEBUG_PARSE("piec");
4841 ret = regatom(pRExC_state, &flags,depth+1);
4843 if (flags & TRYAGAIN)
4850 if (op == '{' && regcurly(RExC_parse)) {
4852 parse_start = RExC_parse; /* MJD */
4853 next = RExC_parse + 1;
4854 while (isDIGIT(*next) || *next == ',') {
4863 if (*next == '}') { /* got one */
4867 min = atoi(RExC_parse);
4871 maxpos = RExC_parse;
4873 if (!max && *maxpos != '0')
4874 max = REG_INFTY; /* meaning "infinity" */
4875 else if (max >= REG_INFTY)
4876 vFAIL2("Quantifier in {,} bigger than %d", REG_INFTY - 1);
4878 nextchar(pRExC_state);
4881 if ((flags&SIMPLE)) {
4882 RExC_naughty += 2 + RExC_naughty / 2;
4883 reginsert(pRExC_state, CURLY, ret);
4884 Set_Node_Offset(ret, parse_start+1); /* MJD */
4885 Set_Node_Cur_Length(ret);
4888 regnode * const w = reg_node(pRExC_state, WHILEM);
4891 REGTAIL(pRExC_state, ret, w);
4892 if (!SIZE_ONLY && RExC_extralen) {
4893 reginsert(pRExC_state, LONGJMP,ret);
4894 reginsert(pRExC_state, NOTHING,ret);
4895 NEXT_OFF(ret) = 3; /* Go over LONGJMP. */
4897 reginsert(pRExC_state, CURLYX,ret);
4899 Set_Node_Offset(ret, parse_start+1);
4900 Set_Node_Length(ret,
4901 op == '{' ? (RExC_parse - parse_start) : 1);
4903 if (!SIZE_ONLY && RExC_extralen)
4904 NEXT_OFF(ret) = 3; /* Go over NOTHING to LONGJMP. */
4905 REGTAIL(pRExC_state, ret, reg_node(pRExC_state, NOTHING));
4907 RExC_whilem_seen++, RExC_extralen += 3;
4908 RExC_naughty += 4 + RExC_naughty; /* compound interest */
4916 if (max && max < min)
4917 vFAIL("Can't do {n,m} with n > m");
4919 ARG1_SET(ret, (U16)min);
4920 ARG2_SET(ret, (U16)max);
4932 #if 0 /* Now runtime fix should be reliable. */
4934 /* if this is reinstated, don't forget to put this back into perldiag:
4936 =item Regexp *+ operand could be empty at {#} in regex m/%s/
4938 (F) The part of the regexp subject to either the * or + quantifier
4939 could match an empty string. The {#} shows in the regular
4940 expression about where the problem was discovered.
4944 if (!(flags&HASWIDTH) && op != '?')
4945 vFAIL("Regexp *+ operand could be empty");
4948 parse_start = RExC_parse;
4949 nextchar(pRExC_state);
4951 *flagp = (op != '+') ? (WORST|SPSTART|HASWIDTH) : (WORST|HASWIDTH);
4953 if (op == '*' && (flags&SIMPLE)) {
4954 reginsert(pRExC_state, STAR, ret);
4958 else if (op == '*') {
4962 else if (op == '+' && (flags&SIMPLE)) {
4963 reginsert(pRExC_state, PLUS, ret);
4967 else if (op == '+') {
4971 else if (op == '?') {
4976 if (!SIZE_ONLY && !(flags&HASWIDTH) && max > REG_INFTY/3 && ckWARN(WARN_REGEXP)) {
4978 "%.*s matches null string many times",
4979 (int)(RExC_parse >= origparse ? RExC_parse - origparse : 0),
4983 if (*RExC_parse == '?') {
4984 nextchar(pRExC_state);
4985 reginsert(pRExC_state, MINMOD, ret);
4986 REGTAIL(pRExC_state, ret, ret + NODE_STEP_REGNODE);
4988 if (ISMULT2(RExC_parse)) {
4990 vFAIL("Nested quantifiers");
4997 /* reg_namedseq(pRExC_state,UVp)
4999 This is expected to be called by a parser routine that has
5000 recognized'\N' and needs to handle the rest. RExC_parse is
5001 expected to point at the first char following the N at the time
5004 If valuep is non-null then it is assumed that we are parsing inside
5005 of a charclass definition and the first codepoint in the resolved
5006 string is returned via *valuep and the routine will return NULL.
5007 In this mode if a multichar string is returned from the charnames
5008 handler a warning will be issued, and only the first char in the
5009 sequence will be examined. If the string returned is zero length
5010 then the value of *valuep is undefined and NON-NULL will
5011 be returned to indicate failure. (This will NOT be a valid pointer
5014 If value is null then it is assumed that we are parsing normal text
5015 and inserts a new EXACT node into the program containing the resolved
5016 string and returns a pointer to the new node. If the string is
5017 zerolength a NOTHING node is emitted.
5019 On success RExC_parse is set to the char following the endbrace.
5020 Parsing failures will generate a fatal errorvia vFAIL(...)
5022 NOTE: We cache all results from the charnames handler locally in
5023 the RExC_charnames hash (created on first use) to prevent a charnames
5024 handler from playing silly-buggers and returning a short string and
5025 then a long string for a given pattern. Since the regexp program
5026 size is calculated during an initial parse this would result
5027 in a buffer overrun so we cache to prevent the charname result from
5028 changing during the course of the parse.
5032 S_reg_namedseq(pTHX_ RExC_state_t *pRExC_state, UV *valuep)
5034 char * name; /* start of the content of the name */
5035 char * endbrace; /* endbrace following the name */
5038 STRLEN len; /* this has various purposes throughout the code */
5039 bool cached = 0; /* if this is true then we shouldn't refcount dev sv_str */
5040 regnode *ret = NULL;
5042 if (*RExC_parse != '{') {
5043 vFAIL("Missing braces on \\N{}");
5045 name = RExC_parse+1;
5046 endbrace = strchr(RExC_parse, '}');
5049 vFAIL("Missing right brace on \\N{}");
5051 RExC_parse = endbrace + 1;
5054 /* RExC_parse points at the beginning brace,
5055 endbrace points at the last */
5056 if ( name[0]=='U' && name[1]=='+' ) {
5057 /* its a "unicode hex" notation {U+89AB} */
5058 I32 fl = PERL_SCAN_ALLOW_UNDERSCORES
5059 | PERL_SCAN_DISALLOW_PREFIX
5060 | (SIZE_ONLY ? PERL_SCAN_SILENT_ILLDIGIT : 0);
5062 len = (STRLEN)(endbrace - name - 2);
5063 cp = grok_hex(name + 2, &len, &fl, NULL);
5064 if ( len != (STRLEN)(endbrace - name - 2) ) {
5073 sv_str= Perl_newSVpvf_nocontext("%c",(int)cp);
5075 /* fetch the charnames handler for this scope */
5076 HV * const table = GvHV(PL_hintgv);
5078 hv_fetchs(table, "charnames", FALSE) :
5080 SV *cv= cvp ? *cvp : NULL;
5083 /* create an SV with the name as argument */
5084 sv_name = newSVpvn(name, endbrace - name);
5086 if (!table || !(PL_hints & HINT_LOCALIZE_HH)) {
5087 vFAIL2("Constant(\\N{%s}) unknown: "
5088 "(possibly a missing \"use charnames ...\")",
5091 if (!cvp || !SvOK(*cvp)) { /* when $^H{charnames} = undef; */
5092 vFAIL2("Constant(\\N{%s}): "
5093 "$^H{charnames} is not defined",SvPVX(sv_name));
5098 if (!RExC_charnames) {
5099 /* make sure our cache is allocated */
5100 RExC_charnames = newHV();
5102 /* see if we have looked this one up before */
5103 he_str = hv_fetch_ent( RExC_charnames, sv_name, 0, 0 );
5105 sv_str = HeVAL(he_str);
5118 count= call_sv(cv, G_SCALAR);
5120 if (count == 1) { /* XXXX is this right? dmq */
5122 SvREFCNT_inc_simple_void(sv_str);
5130 if ( !sv_str || !SvOK(sv_str) ) {
5131 vFAIL2("Constant(\\N{%s}): Call to &{$^H{charnames}} "
5132 "did not return a defined value",SvPVX(sv_name));
5134 if (hv_store_ent( RExC_charnames, sv_name, sv_str, 0))
5139 char *p = SvPV(sv_str, len);
5142 if ( SvUTF8(sv_str) ) {
5143 *valuep = utf8_to_uvchr((U8*)p, &numlen);
5147 We have to turn on utf8 for high bit chars otherwise
5148 we get failures with
5150 "ss" =~ /[\N{LATIN SMALL LETTER SHARP S}]/i
5151 "SS" =~ /[\N{LATIN SMALL LETTER SHARP S}]/i
5153 This is different from what \x{} would do with the same
5154 codepoint, where the condition is > 0xFF.
5161 /* warn if we havent used the whole string? */
5163 if (numlen<len && SIZE_ONLY && ckWARN(WARN_REGEXP)) {
5165 "Ignoring excess chars from \\N{%s} in character class",
5169 } else if (SIZE_ONLY && ckWARN(WARN_REGEXP)) {
5171 "Ignoring zero length \\N{%s} in character class",
5176 SvREFCNT_dec(sv_name);
5178 SvREFCNT_dec(sv_str);
5179 return len ? NULL : (regnode *)&len;
5180 } else if(SvCUR(sv_str)) {
5185 char * parse_start = name-3; /* needed for the offsets */
5186 GET_RE_DEBUG_FLAGS_DECL; /* needed for the offsets */
5188 ret = reg_node(pRExC_state,
5189 (U8)(FOLD ? (LOC ? EXACTFL : EXACTF) : EXACT));
5192 if ( RExC_utf8 && !SvUTF8(sv_str) ) {
5193 sv_utf8_upgrade(sv_str);
5194 } else if ( !RExC_utf8 && SvUTF8(sv_str) ) {
5198 p = SvPV(sv_str, len);
5200 /* len is the length written, charlen is the size the char read */
5201 for ( len = 0; p < pend; p += charlen ) {
5203 UV uvc = utf8_to_uvchr((U8*)p, &charlen);
5205 STRLEN foldlen,numlen;
5206 U8 tmpbuf[UTF8_MAXBYTES_CASE+1], *foldbuf;
5207 uvc = toFOLD_uni(uvc, tmpbuf, &foldlen);
5208 /* Emit all the Unicode characters. */
5210 for (foldbuf = tmpbuf;
5214 uvc = utf8_to_uvchr(foldbuf, &numlen);
5216 const STRLEN unilen = reguni(pRExC_state, uvc, s);
5219 /* In EBCDIC the numlen
5220 * and unilen can differ. */
5222 if (numlen >= foldlen)
5226 break; /* "Can't happen." */
5229 const STRLEN unilen = reguni(pRExC_state, uvc, s);
5241 RExC_size += STR_SZ(len);
5244 RExC_emit += STR_SZ(len);
5246 Set_Node_Cur_Length(ret); /* MJD */
5248 nextchar(pRExC_state);
5250 ret = reg_node(pRExC_state,NOTHING);
5253 SvREFCNT_dec(sv_str);
5256 SvREFCNT_dec(sv_name);
5265 - regatom - the lowest level
5267 * Optimization: gobbles an entire sequence of ordinary characters so that
5268 * it can turn them into a single node, which is smaller to store and
5269 * faster to run. Backslashed characters are exceptions, each becoming a
5270 * separate node; the code is simpler that way and it's not worth fixing.
5272 * [Yes, it is worth fixing, some scripts can run twice the speed.]
5273 * [It looks like its ok, as in S_study_chunk we merge adjacent EXACT nodes]
5276 S_regatom(pTHX_ RExC_state_t *pRExC_state, I32 *flagp, U32 depth)
5279 register regnode *ret = NULL;
5281 char *parse_start = RExC_parse;
5282 GET_RE_DEBUG_FLAGS_DECL;
5283 DEBUG_PARSE("atom");
5284 *flagp = WORST; /* Tentatively. */
5287 switch (*RExC_parse) {
5289 RExC_seen_zerolen++;
5290 nextchar(pRExC_state);
5291 if (RExC_flags & PMf_MULTILINE)
5292 ret = reg_node(pRExC_state, MBOL);
5293 else if (RExC_flags & PMf_SINGLELINE)
5294 ret = reg_node(pRExC_state, SBOL);
5296 ret = reg_node(pRExC_state, BOL);
5297 Set_Node_Length(ret, 1); /* MJD */
5300 nextchar(pRExC_state);
5302 RExC_seen_zerolen++;
5303 if (RExC_flags & PMf_MULTILINE)
5304 ret = reg_node(pRExC_state, MEOL);
5305 else if (RExC_flags & PMf_SINGLELINE)
5306 ret = reg_node(pRExC_state, SEOL);
5308 ret = reg_node(pRExC_state, EOL);
5309 Set_Node_Length(ret, 1); /* MJD */
5312 nextchar(pRExC_state);
5313 if (RExC_flags & PMf_SINGLELINE)
5314 ret = reg_node(pRExC_state, SANY);
5316 ret = reg_node(pRExC_state, REG_ANY);
5317 *flagp |= HASWIDTH|SIMPLE;
5319 Set_Node_Length(ret, 1); /* MJD */
5323 char * const oregcomp_parse = ++RExC_parse;
5324 ret = regclass(pRExC_state,depth+1);
5325 if (*RExC_parse != ']') {
5326 RExC_parse = oregcomp_parse;
5327 vFAIL("Unmatched [");
5329 nextchar(pRExC_state);
5330 *flagp |= HASWIDTH|SIMPLE;
5331 Set_Node_Length(ret, RExC_parse - oregcomp_parse + 1); /* MJD */
5335 nextchar(pRExC_state);
5336 ret = reg(pRExC_state, 1, &flags,depth+1);
5338 if (flags & TRYAGAIN) {
5339 if (RExC_parse == RExC_end) {
5340 /* Make parent create an empty node if needed. */
5348 *flagp |= flags&(HASWIDTH|SPSTART|SIMPLE);
5352 if (flags & TRYAGAIN) {
5356 vFAIL("Internal urp");
5357 /* Supposed to be caught earlier. */
5360 if (!regcurly(RExC_parse)) {
5369 vFAIL("Quantifier follows nothing");
5372 switch (*++RExC_parse) {
5374 RExC_seen_zerolen++;
5375 ret = reg_node(pRExC_state, SBOL);
5377 nextchar(pRExC_state);
5378 Set_Node_Length(ret, 2); /* MJD */
5381 ret = reg_node(pRExC_state, GPOS);
5382 RExC_seen |= REG_SEEN_GPOS;
5384 nextchar(pRExC_state);
5385 Set_Node_Length(ret, 2); /* MJD */
5388 ret = reg_node(pRExC_state, SEOL);
5390 RExC_seen_zerolen++; /* Do not optimize RE away */
5391 nextchar(pRExC_state);
5394 ret = reg_node(pRExC_state, EOS);
5396 RExC_seen_zerolen++; /* Do not optimize RE away */
5397 nextchar(pRExC_state);
5398 Set_Node_Length(ret, 2); /* MJD */
5401 ret = reg_node(pRExC_state, CANY);
5402 RExC_seen |= REG_SEEN_CANY;
5403 *flagp |= HASWIDTH|SIMPLE;
5404 nextchar(pRExC_state);
5405 Set_Node_Length(ret, 2); /* MJD */
5408 ret = reg_node(pRExC_state, CLUMP);
5410 nextchar(pRExC_state);
5411 Set_Node_Length(ret, 2); /* MJD */
5414 ret = reg_node(pRExC_state, (U8)(LOC ? ALNUML : ALNUM));
5415 *flagp |= HASWIDTH|SIMPLE;
5416 nextchar(pRExC_state);
5417 Set_Node_Length(ret, 2); /* MJD */
5420 ret = reg_node(pRExC_state, (U8)(LOC ? NALNUML : NALNUM));
5421 *flagp |= HASWIDTH|SIMPLE;
5422 nextchar(pRExC_state);
5423 Set_Node_Length(ret, 2); /* MJD */
5426 RExC_seen_zerolen++;
5427 RExC_seen |= REG_SEEN_LOOKBEHIND;
5428 ret = reg_node(pRExC_state, (U8)(LOC ? BOUNDL : BOUND));
5430 nextchar(pRExC_state);
5431 Set_Node_Length(ret, 2); /* MJD */
5434 RExC_seen_zerolen++;
5435 RExC_seen |= REG_SEEN_LOOKBEHIND;
5436 ret = reg_node(pRExC_state, (U8)(LOC ? NBOUNDL : NBOUND));
5438 nextchar(pRExC_state);
5439 Set_Node_Length(ret, 2); /* MJD */
5442 ret = reg_node(pRExC_state, (U8)(LOC ? SPACEL : SPACE));
5443 *flagp |= HASWIDTH|SIMPLE;
5444 nextchar(pRExC_state);
5445 Set_Node_Length(ret, 2); /* MJD */
5448 ret = reg_node(pRExC_state, (U8)(LOC ? NSPACEL : NSPACE));
5449 *flagp |= HASWIDTH|SIMPLE;
5450 nextchar(pRExC_state);
5451 Set_Node_Length(ret, 2); /* MJD */
5454 ret = reg_node(pRExC_state, DIGIT);
5455 *flagp |= HASWIDTH|SIMPLE;
5456 nextchar(pRExC_state);
5457 Set_Node_Length(ret, 2); /* MJD */
5460 ret = reg_node(pRExC_state, NDIGIT);
5461 *flagp |= HASWIDTH|SIMPLE;
5462 nextchar(pRExC_state);
5463 Set_Node_Length(ret, 2); /* MJD */
5468 char* const oldregxend = RExC_end;
5469 char* parse_start = RExC_parse - 2;
5471 if (RExC_parse[1] == '{') {
5472 /* a lovely hack--pretend we saw [\pX] instead */
5473 RExC_end = strchr(RExC_parse, '}');
5475 const U8 c = (U8)*RExC_parse;
5477 RExC_end = oldregxend;
5478 vFAIL2("Missing right brace on \\%c{}", c);
5483 RExC_end = RExC_parse + 2;
5484 if (RExC_end > oldregxend)
5485 RExC_end = oldregxend;
5489 ret = regclass(pRExC_state,depth+1);
5491 RExC_end = oldregxend;
5494 Set_Node_Offset(ret, parse_start + 2);
5495 Set_Node_Cur_Length(ret);
5496 nextchar(pRExC_state);
5497 *flagp |= HASWIDTH|SIMPLE;
5501 /* Handle \N{NAME} here and not below because it can be
5502 multicharacter. join_exact() will join them up later on.
5503 Also this makes sure that things like /\N{BLAH}+/ and
5504 \N{BLAH} being multi char Just Happen. dmq*/
5506 ret= reg_namedseq(pRExC_state, NULL);
5518 case '1': case '2': case '3': case '4':
5519 case '5': case '6': case '7': case '8': case '9':
5521 const I32 num = atoi(RExC_parse);
5523 if (num > 9 && num >= RExC_npar)
5526 char * const parse_start = RExC_parse - 1; /* MJD */
5527 while (isDIGIT(*RExC_parse))
5530 if (!SIZE_ONLY && num > (I32)RExC_rx->nparens)
5531 vFAIL("Reference to nonexistent group");
5533 ret = reganode(pRExC_state,
5534 (U8)(FOLD ? (LOC ? REFFL : REFF) : REF),
5538 /* override incorrect value set in reganode MJD */
5539 Set_Node_Offset(ret, parse_start+1);
5540 Set_Node_Cur_Length(ret); /* MJD */
5542 nextchar(pRExC_state);
5547 if (RExC_parse >= RExC_end)
5548 FAIL("Trailing \\");
5551 /* Do not generate "unrecognized" warnings here, we fall
5552 back into the quick-grab loop below */
5559 if (RExC_flags & PMf_EXTENDED) {
5560 while (RExC_parse < RExC_end && *RExC_parse != '\n')
5562 if (RExC_parse < RExC_end)
5568 register STRLEN len;
5573 U8 tmpbuf[UTF8_MAXBYTES_CASE+1], *foldbuf;
5575 parse_start = RExC_parse - 1;
5581 ret = reg_node(pRExC_state,
5582 (U8)(FOLD ? (LOC ? EXACTFL : EXACTF) : EXACT));
5584 for (len = 0, p = RExC_parse - 1;
5585 len < 127 && p < RExC_end;
5588 char * const oldp = p;
5590 if (RExC_flags & PMf_EXTENDED)
5591 p = regwhite(p, RExC_end);
5639 ender = ASCII_TO_NATIVE('\033');
5643 ender = ASCII_TO_NATIVE('\007');
5648 char* const e = strchr(p, '}');
5652 vFAIL("Missing right brace on \\x{}");
5655 I32 flags = PERL_SCAN_ALLOW_UNDERSCORES
5656 | PERL_SCAN_DISALLOW_PREFIX;
5657 STRLEN numlen = e - p - 1;
5658 ender = grok_hex(p + 1, &numlen, &flags, NULL);
5665 I32 flags = PERL_SCAN_DISALLOW_PREFIX;
5667 ender = grok_hex(p, &numlen, &flags, NULL);
5673 ender = UCHARAT(p++);
5674 ender = toCTRL(ender);
5676 case '0': case '1': case '2': case '3':case '4':
5677 case '5': case '6': case '7': case '8':case '9':
5679 (isDIGIT(p[1]) && atoi(p) >= RExC_npar) ) {
5682 ender = grok_oct(p, &numlen, &flags, NULL);
5692 FAIL("Trailing \\");
5695 if (!SIZE_ONLY&& isALPHA(*p) && ckWARN(WARN_REGEXP))
5696 vWARN2(p + 1, "Unrecognized escape \\%c passed through", UCHARAT(p));
5697 goto normal_default;
5702 if (UTF8_IS_START(*p) && UTF) {
5704 ender = utf8n_to_uvchr((U8*)p, RExC_end - p,
5705 &numlen, UTF8_ALLOW_DEFAULT);
5712 if (RExC_flags & PMf_EXTENDED)
5713 p = regwhite(p, RExC_end);
5715 /* Prime the casefolded buffer. */
5716 ender = toFOLD_uni(ender, tmpbuf, &foldlen);
5718 if (ISMULT2(p)) { /* Back off on ?+*. */
5723 /* Emit all the Unicode characters. */
5725 for (foldbuf = tmpbuf;
5727 foldlen -= numlen) {
5728 ender = utf8_to_uvchr(foldbuf, &numlen);
5730 const STRLEN unilen = reguni(pRExC_state, ender, s);
5733 /* In EBCDIC the numlen
5734 * and unilen can differ. */
5736 if (numlen >= foldlen)
5740 break; /* "Can't happen." */
5744 const STRLEN unilen = reguni(pRExC_state, ender, s);
5753 REGC((char)ender, s++);
5759 /* Emit all the Unicode characters. */
5761 for (foldbuf = tmpbuf;
5763 foldlen -= numlen) {
5764 ender = utf8_to_uvchr(foldbuf, &numlen);
5766 const STRLEN unilen = reguni(pRExC_state, ender, s);
5769 /* In EBCDIC the numlen
5770 * and unilen can differ. */
5772 if (numlen >= foldlen)
5780 const STRLEN unilen = reguni(pRExC_state, ender, s);
5789 REGC((char)ender, s++);
5793 Set_Node_Cur_Length(ret); /* MJD */
5794 nextchar(pRExC_state);
5796 /* len is STRLEN which is unsigned, need to copy to signed */
5799 vFAIL("Internal disaster");
5803 if (len == 1 && UNI_IS_INVARIANT(ender))
5807 RExC_size += STR_SZ(len);
5810 RExC_emit += STR_SZ(len);
5816 /* If the encoding pragma is in effect recode the text of
5817 * any EXACT-kind nodes. */
5818 if (ret && PL_encoding && PL_regkind[OP(ret)] == EXACT) {
5819 const STRLEN oldlen = STR_LEN(ret);
5820 SV * const sv = sv_2mortal(newSVpvn(STRING(ret), oldlen));
5824 if (sv_utf8_downgrade(sv, TRUE)) {
5825 const char * const s = sv_recode_to_utf8(sv, PL_encoding);
5826 const STRLEN newlen = SvCUR(sv);
5831 GET_RE_DEBUG_FLAGS_DECL;
5832 DEBUG_COMPILE_r(PerlIO_printf(Perl_debug_log, "recode %*s to %*s\n",
5833 (int)oldlen, STRING(ret),
5835 Copy(s, STRING(ret), newlen, char);
5836 STR_LEN(ret) += newlen - oldlen;
5837 RExC_emit += STR_SZ(newlen) - STR_SZ(oldlen);
5839 RExC_size += STR_SZ(newlen) - STR_SZ(oldlen);
5847 S_regwhite(char *p, const char *e)
5852 else if (*p == '#') {
5855 } while (p < e && *p != '\n');
5863 /* Parse POSIX character classes: [[:foo:]], [[=foo=]], [[.foo.]].
5864 Character classes ([:foo:]) can also be negated ([:^foo:]).
5865 Returns a named class id (ANYOF_XXX) if successful, -1 otherwise.
5866 Equivalence classes ([=foo=]) and composites ([.foo.]) are parsed,
5867 but trigger failures because they are currently unimplemented. */
5869 #define POSIXCC_DONE(c) ((c) == ':')
5870 #define POSIXCC_NOTYET(c) ((c) == '=' || (c) == '.')
5871 #define POSIXCC(c) (POSIXCC_DONE(c) || POSIXCC_NOTYET(c))
5874 S_regpposixcc(pTHX_ RExC_state_t *pRExC_state, I32 value)
5877 I32 namedclass = OOB_NAMEDCLASS;
5879 if (value == '[' && RExC_parse + 1 < RExC_end &&
5880 /* I smell either [: or [= or [. -- POSIX has been here, right? */
5881 POSIXCC(UCHARAT(RExC_parse))) {
5882 const char c = UCHARAT(RExC_parse);
5883 char* const s = RExC_parse++;
5885 while (RExC_parse < RExC_end && UCHARAT(RExC_parse) != c)
5887 if (RExC_parse == RExC_end)
5888 /* Grandfather lone [:, [=, [. */
5891 const char* const t = RExC_parse++; /* skip over the c */
5894 if (UCHARAT(RExC_parse) == ']') {
5895 const char *posixcc = s + 1;
5896 RExC_parse++; /* skip over the ending ] */
5899 const I32 complement = *posixcc == '^' ? *posixcc++ : 0;
5900 const I32 skip = t - posixcc;
5902 /* Initially switch on the length of the name. */
5905 if (memEQ(posixcc, "word", 4)) /* this is not POSIX, this is the Perl \w */
5906 namedclass = complement ? ANYOF_NALNUM : ANYOF_ALNUM;
5909 /* Names all of length 5. */
5910 /* alnum alpha ascii blank cntrl digit graph lower
5911 print punct space upper */
5912 /* Offset 4 gives the best switch position. */
5913 switch (posixcc[4]) {
5915 if (memEQ(posixcc, "alph", 4)) /* alpha */
5916 namedclass = complement ? ANYOF_NALPHA : ANYOF_ALPHA;
5919 if (memEQ(posixcc, "spac", 4)) /* space */
5920 namedclass = complement ? ANYOF_NPSXSPC : ANYOF_PSXSPC;
5923 if (memEQ(posixcc, "grap", 4)) /* graph */
5924 namedclass = complement ? ANYOF_NGRAPH : ANYOF_GRAPH;
5927 if (memEQ(posixcc, "asci", 4)) /* ascii */
5928 namedclass = complement ? ANYOF_NASCII : ANYOF_ASCII;
5931 if (memEQ(posixcc, "blan", 4)) /* blank */
5932 namedclass = complement ? ANYOF_NBLANK : ANYOF_BLANK;
5935 if (memEQ(posixcc, "cntr", 4)) /* cntrl */
5936 namedclass = complement ? ANYOF_NCNTRL : ANYOF_CNTRL;
5939 if (memEQ(posixcc, "alnu", 4)) /* alnum */
5940 namedclass = complement ? ANYOF_NALNUMC : ANYOF_ALNUMC;
5943 if (memEQ(posixcc, "lowe", 4)) /* lower */
5944 namedclass = complement ? ANYOF_NLOWER : ANYOF_LOWER;
5945 else if (memEQ(posixcc, "uppe", 4)) /* upper */
5946 namedclass = complement ? ANYOF_NUPPER : ANYOF_UPPER;
5949 if (memEQ(posixcc, "digi", 4)) /* digit */
5950 namedclass = complement ? ANYOF_NDIGIT : ANYOF_DIGIT;
5951 else if (memEQ(posixcc, "prin", 4)) /* print */
5952 namedclass = complement ? ANYOF_NPRINT : ANYOF_PRINT;
5953 else if (memEQ(posixcc, "punc", 4)) /* punct */
5954 namedclass = complement ? ANYOF_NPUNCT : ANYOF_PUNCT;
5959 if (memEQ(posixcc, "xdigit", 6))
5960 namedclass = complement ? ANYOF_NXDIGIT : ANYOF_XDIGIT;
5964 if (namedclass == OOB_NAMEDCLASS)
5965 Simple_vFAIL3("POSIX class [:%.*s:] unknown",
5967 assert (posixcc[skip] == ':');
5968 assert (posixcc[skip+1] == ']');
5969 } else if (!SIZE_ONLY) {
5970 /* [[=foo=]] and [[.foo.]] are still future. */
5972 /* adjust RExC_parse so the warning shows after
5974 while (UCHARAT(RExC_parse) && UCHARAT(RExC_parse) != ']')
5976 Simple_vFAIL3("POSIX syntax [%c %c] is reserved for future extensions", c, c);
5979 /* Maternal grandfather:
5980 * "[:" ending in ":" but not in ":]" */
5990 S_checkposixcc(pTHX_ RExC_state_t *pRExC_state)
5993 if (POSIXCC(UCHARAT(RExC_parse))) {
5994 const char *s = RExC_parse;
5995 const char c = *s++;
5999 if (*s && c == *s && s[1] == ']') {
6000 if (ckWARN(WARN_REGEXP))
6002 "POSIX syntax [%c %c] belongs inside character classes",
6005 /* [[=foo=]] and [[.foo.]] are still future. */
6006 if (POSIXCC_NOTYET(c)) {
6007 /* adjust RExC_parse so the error shows after
6009 while (UCHARAT(RExC_parse) && UCHARAT(RExC_parse++) != ']')
6011 Simple_vFAIL3("POSIX syntax [%c %c] is reserved for future extensions", c, c);
6019 parse a class specification and produce either an ANYOF node that
6020 matches the pattern. If the pattern matches a single char only and
6021 that char is < 256 then we produce an EXACT node instead.
6024 S_regclass(pTHX_ RExC_state_t *pRExC_state, U32 depth)
6027 register UV value = 0;
6028 register UV nextvalue;
6029 register IV prevvalue = OOB_UNICODE;
6030 register IV range = 0;
6031 register regnode *ret;
6034 char *rangebegin = NULL;
6035 bool need_class = 0;
6038 bool optimize_invert = TRUE;
6039 AV* unicode_alternate = NULL;
6041 UV literal_endpoint = 0;
6043 UV stored = 0; /* number of chars stored in the class */
6045 regnode * const orig_emit = RExC_emit; /* Save the original RExC_emit in
6046 case we need to change the emitted regop to an EXACT. */
6047 const char * orig_parse = RExC_parse;
6048 GET_RE_DEBUG_FLAGS_DECL;
6050 PERL_UNUSED_ARG(depth);
6053 DEBUG_PARSE("clas");
6055 /* Assume we are going to generate an ANYOF node. */
6056 ret = reganode(pRExC_state, ANYOF, 0);
6059 ANYOF_FLAGS(ret) = 0;
6061 if (UCHARAT(RExC_parse) == '^') { /* Complement of range. */
6065 ANYOF_FLAGS(ret) |= ANYOF_INVERT;
6069 RExC_size += ANYOF_SKIP;
6070 listsv = &PL_sv_undef; /* For code scanners: listsv always non-NULL. */
6073 RExC_emit += ANYOF_SKIP;
6075 ANYOF_FLAGS(ret) |= ANYOF_FOLD;
6077 ANYOF_FLAGS(ret) |= ANYOF_LOCALE;
6078 ANYOF_BITMAP_ZERO(ret);
6079 listsv = newSVpvs("# comment\n");
6082 nextvalue = RExC_parse < RExC_end ? UCHARAT(RExC_parse) : 0;
6084 if (!SIZE_ONLY && POSIXCC(nextvalue))
6085 checkposixcc(pRExC_state);
6087 /* allow 1st char to be ] (allowing it to be - is dealt with later) */
6088 if (UCHARAT(RExC_parse) == ']')
6092 while (RExC_parse < RExC_end && UCHARAT(RExC_parse) != ']') {
6096 namedclass = OOB_NAMEDCLASS; /* initialize as illegal */
6099 rangebegin = RExC_parse;
6101 value = utf8n_to_uvchr((U8*)RExC_parse,
6102 RExC_end - RExC_parse,
6103 &numlen, UTF8_ALLOW_DEFAULT);
6104 RExC_parse += numlen;
6107 value = UCHARAT(RExC_parse++);
6109 nextvalue = RExC_parse < RExC_end ? UCHARAT(RExC_parse) : 0;
6110 if (value == '[' && POSIXCC(nextvalue))
6111 namedclass = regpposixcc(pRExC_state, value);
6112 else if (value == '\\') {
6114 value = utf8n_to_uvchr((U8*)RExC_parse,
6115 RExC_end - RExC_parse,
6116 &numlen, UTF8_ALLOW_DEFAULT);
6117 RExC_parse += numlen;
6120 value = UCHARAT(RExC_parse++);
6121 /* Some compilers cannot handle switching on 64-bit integer
6122 * values, therefore value cannot be an UV. Yes, this will
6123 * be a problem later if we want switch on Unicode.
6124 * A similar issue a little bit later when switching on
6125 * namedclass. --jhi */
6126 switch ((I32)value) {
6127 case 'w': namedclass = ANYOF_ALNUM; break;
6128 case 'W': namedclass = ANYOF_NALNUM; break;
6129 case 's': namedclass = ANYOF_SPACE; break;
6130 case 'S': namedclass = ANYOF_NSPACE; break;
6131 case 'd': namedclass = ANYOF_DIGIT; break;
6132 case 'D': namedclass = ANYOF_NDIGIT; break;
6133 case 'N': /* Handle \N{NAME} in class */
6135 /* We only pay attention to the first char of
6136 multichar strings being returned. I kinda wonder
6137 if this makes sense as it does change the behaviour
6138 from earlier versions, OTOH that behaviour was broken
6140 UV v; /* value is register so we cant & it /grrr */
6141 if (reg_namedseq(pRExC_state, &v)) {
6151 if (RExC_parse >= RExC_end)
6152 vFAIL2("Empty \\%c{}", (U8)value);
6153 if (*RExC_parse == '{') {
6154 const U8 c = (U8)value;
6155 e = strchr(RExC_parse++, '}');
6157 vFAIL2("Missing right brace on \\%c{}", c);
6158 while (isSPACE(UCHARAT(RExC_parse)))
6160 if (e == RExC_parse)
6161 vFAIL2("Empty \\%c{}", c);
6163 while (isSPACE(UCHARAT(RExC_parse + n - 1)))
6171 if (UCHARAT(RExC_parse) == '^') {
6174 value = value == 'p' ? 'P' : 'p'; /* toggle */
6175 while (isSPACE(UCHARAT(RExC_parse))) {
6180 Perl_sv_catpvf(aTHX_ listsv, "%cutf8::%.*s\n",
6181 (value=='p' ? '+' : '!'), (int)n, RExC_parse);
6184 ANYOF_FLAGS(ret) |= ANYOF_UNICODE;
6185 namedclass = ANYOF_MAX; /* no official name, but it's named */
6188 case 'n': value = '\n'; break;
6189 case 'r': value = '\r'; break;
6190 case 't': value = '\t'; break;
6191 case 'f': value = '\f'; break;
6192 case 'b': value = '\b'; break;
6193 case 'e': value = ASCII_TO_NATIVE('\033');break;
6194 case 'a': value = ASCII_TO_NATIVE('\007');break;
6196 if (*RExC_parse == '{') {
6197 I32 flags = PERL_SCAN_ALLOW_UNDERSCORES
6198 | PERL_SCAN_DISALLOW_PREFIX;
6199 char * const e = strchr(RExC_parse++, '}');
6201 vFAIL("Missing right brace on \\x{}");
6203 numlen = e - RExC_parse;
6204 value = grok_hex(RExC_parse, &numlen, &flags, NULL);
6208 I32 flags = PERL_SCAN_DISALLOW_PREFIX;
6210 value = grok_hex(RExC_parse, &numlen, &flags, NULL);
6211 RExC_parse += numlen;
6215 value = UCHARAT(RExC_parse++);
6216 value = toCTRL(value);
6218 case '0': case '1': case '2': case '3': case '4':
6219 case '5': case '6': case '7': case '8': case '9':
6223 value = grok_oct(--RExC_parse, &numlen, &flags, NULL);
6224 RExC_parse += numlen;
6228 if (!SIZE_ONLY && isALPHA(value) && ckWARN(WARN_REGEXP))
6230 "Unrecognized escape \\%c in character class passed through",
6234 } /* end of \blah */
6240 if (namedclass > OOB_NAMEDCLASS) { /* this is a named class \blah */
6242 if (!SIZE_ONLY && !need_class)
6243 ANYOF_CLASS_ZERO(ret);
6247 /* a bad range like a-\d, a-[:digit:] ? */
6250 if (ckWARN(WARN_REGEXP)) {
6252 RExC_parse >= rangebegin ?
6253 RExC_parse - rangebegin : 0;
6255 "False [] range \"%*.*s\"",
6258 if (prevvalue < 256) {
6259 ANYOF_BITMAP_SET(ret, prevvalue);
6260 ANYOF_BITMAP_SET(ret, '-');
6263 ANYOF_FLAGS(ret) |= ANYOF_UNICODE;
6264 Perl_sv_catpvf(aTHX_ listsv,
6265 "%04"UVxf"\n%04"UVxf"\n", (UV)prevvalue, (UV) '-');
6269 range = 0; /* this was not a true range */
6273 const char *what = NULL;
6276 if (namedclass > OOB_NAMEDCLASS)
6277 optimize_invert = FALSE;
6278 /* Possible truncation here but in some 64-bit environments
6279 * the compiler gets heartburn about switch on 64-bit values.
6280 * A similar issue a little earlier when switching on value.
6282 switch ((I32)namedclass) {
6285 ANYOF_CLASS_SET(ret, ANYOF_ALNUM);
6287 for (value = 0; value < 256; value++)
6289 ANYOF_BITMAP_SET(ret, value);
6296 ANYOF_CLASS_SET(ret, ANYOF_NALNUM);
6298 for (value = 0; value < 256; value++)
6299 if (!isALNUM(value))
6300 ANYOF_BITMAP_SET(ret, value);
6307 ANYOF_CLASS_SET(ret, ANYOF_ALNUMC);
6309 for (value = 0; value < 256; value++)
6310 if (isALNUMC(value))
6311 ANYOF_BITMAP_SET(ret, value);
6318 ANYOF_CLASS_SET(ret, ANYOF_NALNUMC);
6320 for (value = 0; value < 256; value++)
6321 if (!isALNUMC(value))
6322 ANYOF_BITMAP_SET(ret, value);
6329 ANYOF_CLASS_SET(ret, ANYOF_ALPHA);
6331 for (value = 0; value < 256; value++)
6333 ANYOF_BITMAP_SET(ret, value);
6340 ANYOF_CLASS_SET(ret, ANYOF_NALPHA);
6342 for (value = 0; value < 256; value++)
6343 if (!isALPHA(value))
6344 ANYOF_BITMAP_SET(ret, value);
6351 ANYOF_CLASS_SET(ret, ANYOF_ASCII);
6354 for (value = 0; value < 128; value++)
6355 ANYOF_BITMAP_SET(ret, value);
6357 for (value = 0; value < 256; value++) {
6359 ANYOF_BITMAP_SET(ret, value);
6368 ANYOF_CLASS_SET(ret, ANYOF_NASCII);
6371 for (value = 128; value < 256; value++)
6372 ANYOF_BITMAP_SET(ret, value);
6374 for (value = 0; value < 256; value++) {
6375 if (!isASCII(value))
6376 ANYOF_BITMAP_SET(ret, value);
6385 ANYOF_CLASS_SET(ret, ANYOF_BLANK);
6387 for (value = 0; value < 256; value++)
6389 ANYOF_BITMAP_SET(ret, value);
6396 ANYOF_CLASS_SET(ret, ANYOF_NBLANK);
6398 for (value = 0; value < 256; value++)
6399 if (!isBLANK(value))
6400 ANYOF_BITMAP_SET(ret, value);
6407 ANYOF_CLASS_SET(ret, ANYOF_CNTRL);
6409 for (value = 0; value < 256; value++)
6411 ANYOF_BITMAP_SET(ret, value);
6418 ANYOF_CLASS_SET(ret, ANYOF_NCNTRL);
6420 for (value = 0; value < 256; value++)
6421 if (!isCNTRL(value))
6422 ANYOF_BITMAP_SET(ret, value);
6429 ANYOF_CLASS_SET(ret, ANYOF_DIGIT);
6431 /* consecutive digits assumed */
6432 for (value = '0'; value <= '9'; value++)
6433 ANYOF_BITMAP_SET(ret, value);
6440 ANYOF_CLASS_SET(ret, ANYOF_NDIGIT);
6442 /* consecutive digits assumed */
6443 for (value = 0; value < '0'; value++)
6444 ANYOF_BITMAP_SET(ret, value);
6445 for (value = '9' + 1; value < 256; value++)
6446 ANYOF_BITMAP_SET(ret, value);
6453 ANYOF_CLASS_SET(ret, ANYOF_GRAPH);
6455 for (value = 0; value < 256; value++)
6457 ANYOF_BITMAP_SET(ret, value);
6464 ANYOF_CLASS_SET(ret, ANYOF_NGRAPH);
6466 for (value = 0; value < 256; value++)
6467 if (!isGRAPH(value))
6468 ANYOF_BITMAP_SET(ret, value);
6475 ANYOF_CLASS_SET(ret, ANYOF_LOWER);
6477 for (value = 0; value < 256; value++)
6479 ANYOF_BITMAP_SET(ret, value);
6486 ANYOF_CLASS_SET(ret, ANYOF_NLOWER);
6488 for (value = 0; value < 256; value++)
6489 if (!isLOWER(value))
6490 ANYOF_BITMAP_SET(ret, value);
6497 ANYOF_CLASS_SET(ret, ANYOF_PRINT);
6499 for (value = 0; value < 256; value++)
6501 ANYOF_BITMAP_SET(ret, value);
6508 ANYOF_CLASS_SET(ret, ANYOF_NPRINT);
6510 for (value = 0; value < 256; value++)
6511 if (!isPRINT(value))
6512 ANYOF_BITMAP_SET(ret, value);
6519 ANYOF_CLASS_SET(ret, ANYOF_PSXSPC);
6521 for (value = 0; value < 256; value++)
6522 if (isPSXSPC(value))
6523 ANYOF_BITMAP_SET(ret, value);
6530 ANYOF_CLASS_SET(ret, ANYOF_NPSXSPC);
6532 for (value = 0; value < 256; value++)
6533 if (!isPSXSPC(value))
6534 ANYOF_BITMAP_SET(ret, value);
6541 ANYOF_CLASS_SET(ret, ANYOF_PUNCT);
6543 for (value = 0; value < 256; value++)
6545 ANYOF_BITMAP_SET(ret, value);
6552 ANYOF_CLASS_SET(ret, ANYOF_NPUNCT);
6554 for (value = 0; value < 256; value++)
6555 if (!isPUNCT(value))
6556 ANYOF_BITMAP_SET(ret, value);
6563 ANYOF_CLASS_SET(ret, ANYOF_SPACE);
6565 for (value = 0; value < 256; value++)
6567 ANYOF_BITMAP_SET(ret, value);
6574 ANYOF_CLASS_SET(ret, ANYOF_NSPACE);
6576 for (value = 0; value < 256; value++)
6577 if (!isSPACE(value))
6578 ANYOF_BITMAP_SET(ret, value);
6585 ANYOF_CLASS_SET(ret, ANYOF_UPPER);
6587 for (value = 0; value < 256; value++)
6589 ANYOF_BITMAP_SET(ret, value);
6596 ANYOF_CLASS_SET(ret, ANYOF_NUPPER);
6598 for (value = 0; value < 256; value++)
6599 if (!isUPPER(value))
6600 ANYOF_BITMAP_SET(ret, value);
6607 ANYOF_CLASS_SET(ret, ANYOF_XDIGIT);
6609 for (value = 0; value < 256; value++)
6610 if (isXDIGIT(value))
6611 ANYOF_BITMAP_SET(ret, value);
6618 ANYOF_CLASS_SET(ret, ANYOF_NXDIGIT);
6620 for (value = 0; value < 256; value++)
6621 if (!isXDIGIT(value))
6622 ANYOF_BITMAP_SET(ret, value);
6628 /* this is to handle \p and \P */
6631 vFAIL("Invalid [::] class");
6635 /* Strings such as "+utf8::isWord\n" */
6636 Perl_sv_catpvf(aTHX_ listsv, "%cutf8::Is%s\n", yesno, what);
6639 ANYOF_FLAGS(ret) |= ANYOF_CLASS;
6642 } /* end of namedclass \blah */
6645 if (prevvalue > (IV)value) /* b-a */ {
6646 const int w = RExC_parse - rangebegin;
6647 Simple_vFAIL4("Invalid [] range \"%*.*s\"", w, w, rangebegin);
6648 range = 0; /* not a valid range */
6652 prevvalue = value; /* save the beginning of the range */
6653 if (*RExC_parse == '-' && RExC_parse+1 < RExC_end &&
6654 RExC_parse[1] != ']') {
6657 /* a bad range like \w-, [:word:]- ? */
6658 if (namedclass > OOB_NAMEDCLASS) {
6659 if (ckWARN(WARN_REGEXP)) {
6661 RExC_parse >= rangebegin ?
6662 RExC_parse - rangebegin : 0;
6664 "False [] range \"%*.*s\"",
6668 ANYOF_BITMAP_SET(ret, '-');
6670 range = 1; /* yeah, it's a range! */
6671 continue; /* but do it the next time */
6675 /* now is the next time */
6676 /*stored += (value - prevvalue + 1);*/
6678 if (prevvalue < 256) {
6679 const IV ceilvalue = value < 256 ? value : 255;
6682 /* In EBCDIC [\x89-\x91] should include
6683 * the \x8e but [i-j] should not. */
6684 if (literal_endpoint == 2 &&
6685 ((isLOWER(prevvalue) && isLOWER(ceilvalue)) ||
6686 (isUPPER(prevvalue) && isUPPER(ceilvalue))))
6688 if (isLOWER(prevvalue)) {
6689 for (i = prevvalue; i <= ceilvalue; i++)
6691 ANYOF_BITMAP_SET(ret, i);
6693 for (i = prevvalue; i <= ceilvalue; i++)
6695 ANYOF_BITMAP_SET(ret, i);
6700 for (i = prevvalue; i <= ceilvalue; i++) {
6701 if (!ANYOF_BITMAP_TEST(ret,i)) {
6703 ANYOF_BITMAP_SET(ret, i);
6707 if (value > 255 || UTF) {
6708 const UV prevnatvalue = NATIVE_TO_UNI(prevvalue);
6709 const UV natvalue = NATIVE_TO_UNI(value);
6710 stored+=2; /* can't optimize this class */
6711 ANYOF_FLAGS(ret) |= ANYOF_UNICODE;
6712 if (prevnatvalue < natvalue) { /* what about > ? */
6713 Perl_sv_catpvf(aTHX_ listsv, "%04"UVxf"\t%04"UVxf"\n",
6714 prevnatvalue, natvalue);
6716 else if (prevnatvalue == natvalue) {
6717 Perl_sv_catpvf(aTHX_ listsv, "%04"UVxf"\n", natvalue);
6719 U8 foldbuf[UTF8_MAXBYTES_CASE+1];
6721 const UV f = to_uni_fold(natvalue, foldbuf, &foldlen);
6723 #ifdef EBCDIC /* RD t/uni/fold ff and 6b */
6724 if (RExC_precomp[0] == ':' &&
6725 RExC_precomp[1] == '[' &&
6726 (f == 0xDF || f == 0x92)) {
6727 f = NATIVE_TO_UNI(f);
6730 /* If folding and foldable and a single
6731 * character, insert also the folded version
6732 * to the charclass. */
6734 #ifdef EBCDIC /* RD tunifold ligatures s,t fb05, fb06 */
6735 if ((RExC_precomp[0] == ':' &&
6736 RExC_precomp[1] == '[' &&
6738 (value == 0xFB05 || value == 0xFB06))) ?
6739 foldlen == ((STRLEN)UNISKIP(f) - 1) :
6740 foldlen == (STRLEN)UNISKIP(f) )
6742 if (foldlen == (STRLEN)UNISKIP(f))
6744 Perl_sv_catpvf(aTHX_ listsv,
6747 /* Any multicharacter foldings
6748 * require the following transform:
6749 * [ABCDEF] -> (?:[ABCabcDEFd]|pq|rst)
6750 * where E folds into "pq" and F folds
6751 * into "rst", all other characters
6752 * fold to single characters. We save
6753 * away these multicharacter foldings,
6754 * to be later saved as part of the
6755 * additional "s" data. */
6758 if (!unicode_alternate)
6759 unicode_alternate = newAV();
6760 sv = newSVpvn((char*)foldbuf, foldlen);
6762 av_push(unicode_alternate, sv);
6766 /* If folding and the value is one of the Greek
6767 * sigmas insert a few more sigmas to make the
6768 * folding rules of the sigmas to work right.
6769 * Note that not all the possible combinations
6770 * are handled here: some of them are handled
6771 * by the standard folding rules, and some of
6772 * them (literal or EXACTF cases) are handled
6773 * during runtime in regexec.c:S_find_byclass(). */
6774 if (value == UNICODE_GREEK_SMALL_LETTER_FINAL_SIGMA) {
6775 Perl_sv_catpvf(aTHX_ listsv, "%04"UVxf"\n",
6776 (UV)UNICODE_GREEK_CAPITAL_LETTER_SIGMA);
6777 Perl_sv_catpvf(aTHX_ listsv, "%04"UVxf"\n",
6778 (UV)UNICODE_GREEK_SMALL_LETTER_SIGMA);
6780 else if (value == UNICODE_GREEK_CAPITAL_LETTER_SIGMA)
6781 Perl_sv_catpvf(aTHX_ listsv, "%04"UVxf"\n",
6782 (UV)UNICODE_GREEK_SMALL_LETTER_SIGMA);
6787 literal_endpoint = 0;
6791 range = 0; /* this range (if it was one) is done now */
6795 ANYOF_FLAGS(ret) |= ANYOF_LARGE;
6797 RExC_size += ANYOF_CLASS_ADD_SKIP;
6799 RExC_emit += ANYOF_CLASS_ADD_SKIP;
6805 /****** !SIZE_ONLY AFTER HERE *********/
6807 if( stored == 1 && value < 256
6808 && !( ANYOF_FLAGS(ret) & ( ANYOF_FLAGS_ALL ^ ANYOF_FOLD ) )
6810 /* optimize single char class to an EXACT node
6811 but *only* when its not a UTF/high char */
6812 const char * cur_parse= RExC_parse;
6813 RExC_emit = (regnode *)orig_emit;
6814 RExC_parse = (char *)orig_parse;
6815 ret = reg_node(pRExC_state,
6816 (U8)((ANYOF_FLAGS(ret) & ANYOF_FOLD) ? EXACTF : EXACT));
6817 RExC_parse = (char *)cur_parse;
6818 *STRING(ret)= (char)value;
6820 RExC_emit += STR_SZ(1);
6823 /* optimize case-insensitive simple patterns (e.g. /[a-z]/i) */
6824 if ( /* If the only flag is folding (plus possibly inversion). */
6825 ((ANYOF_FLAGS(ret) & (ANYOF_FLAGS_ALL ^ ANYOF_INVERT)) == ANYOF_FOLD)
6827 for (value = 0; value < 256; ++value) {
6828 if (ANYOF_BITMAP_TEST(ret, value)) {
6829 UV fold = PL_fold[value];
6832 ANYOF_BITMAP_SET(ret, fold);
6835 ANYOF_FLAGS(ret) &= ~ANYOF_FOLD;
6838 /* optimize inverted simple patterns (e.g. [^a-z]) */
6839 if (optimize_invert &&
6840 /* If the only flag is inversion. */
6841 (ANYOF_FLAGS(ret) & ANYOF_FLAGS_ALL) == ANYOF_INVERT) {
6842 for (value = 0; value < ANYOF_BITMAP_SIZE; ++value)
6843 ANYOF_BITMAP(ret)[value] ^= ANYOF_FLAGS_ALL;
6844 ANYOF_FLAGS(ret) = ANYOF_UNICODE_ALL;
6847 AV * const av = newAV();
6849 /* The 0th element stores the character class description
6850 * in its textual form: used later (regexec.c:Perl_regclass_swash())
6851 * to initialize the appropriate swash (which gets stored in
6852 * the 1st element), and also useful for dumping the regnode.
6853 * The 2nd element stores the multicharacter foldings,
6854 * used later (regexec.c:S_reginclass()). */
6855 av_store(av, 0, listsv);
6856 av_store(av, 1, NULL);
6857 av_store(av, 2, (SV*)unicode_alternate);
6858 rv = newRV_noinc((SV*)av);
6859 n = add_data(pRExC_state, 1, "s");
6860 RExC_rx->data->data[n] = (void*)rv;
6867 S_nextchar(pTHX_ RExC_state_t *pRExC_state)
6869 char* const retval = RExC_parse++;
6872 if (*RExC_parse == '(' && RExC_parse[1] == '?' &&
6873 RExC_parse[2] == '#') {
6874 while (*RExC_parse != ')') {
6875 if (RExC_parse == RExC_end)
6876 FAIL("Sequence (?#... not terminated");
6882 if (RExC_flags & PMf_EXTENDED) {
6883 if (isSPACE(*RExC_parse)) {
6887 else if (*RExC_parse == '#') {
6888 while (RExC_parse < RExC_end)
6889 if (*RExC_parse++ == '\n') break;
6898 - reg_node - emit a node
6900 STATIC regnode * /* Location. */
6901 S_reg_node(pTHX_ RExC_state_t *pRExC_state, U8 op)
6904 register regnode *ptr;
6905 regnode * const ret = RExC_emit;
6906 GET_RE_DEBUG_FLAGS_DECL;
6909 SIZE_ALIGN(RExC_size);
6913 NODE_ALIGN_FILL(ret);
6915 FILL_ADVANCE_NODE(ptr, op);
6916 if (RExC_offsets) { /* MJD */
6917 MJD_OFFSET_DEBUG(("%s:%d: (op %s) %s %"UVuf" (len %"UVuf") (max %"UVuf").\n",
6918 "reg_node", __LINE__,
6920 (UV)(RExC_emit - RExC_emit_start) > RExC_offsets[0]
6921 ? "Overwriting end of array!\n" : "OK",
6922 (UV)(RExC_emit - RExC_emit_start),
6923 (UV)(RExC_parse - RExC_start),
6924 (UV)RExC_offsets[0]));
6925 Set_Node_Offset(RExC_emit, RExC_parse + (op == END));
6934 - reganode - emit a node with an argument
6936 STATIC regnode * /* Location. */
6937 S_reganode(pTHX_ RExC_state_t *pRExC_state, U8 op, U32 arg)
6940 register regnode *ptr;
6941 regnode * const ret = RExC_emit;
6942 GET_RE_DEBUG_FLAGS_DECL;
6945 SIZE_ALIGN(RExC_size);
6950 NODE_ALIGN_FILL(ret);
6952 FILL_ADVANCE_NODE_ARG(ptr, op, arg);
6953 if (RExC_offsets) { /* MJD */
6954 MJD_OFFSET_DEBUG(("%s(%d): (op %s) %s %"UVuf" <- %"UVuf" (max %"UVuf").\n",
6958 (UV)(RExC_emit - RExC_emit_start) > RExC_offsets[0] ?
6959 "Overwriting end of array!\n" : "OK",
6960 (UV)(RExC_emit - RExC_emit_start),
6961 (UV)(RExC_parse - RExC_start),
6962 (UV)RExC_offsets[0]));
6963 Set_Cur_Node_Offset;
6972 - reguni - emit (if appropriate) a Unicode character
6975 S_reguni(pTHX_ const RExC_state_t *pRExC_state, UV uv, char* s)
6978 return SIZE_ONLY ? UNISKIP(uv) : (uvchr_to_utf8((U8*)s, uv) - (U8*)s);
6982 - reginsert - insert an operator in front of already-emitted operand
6984 * Means relocating the operand.
6987 S_reginsert(pTHX_ RExC_state_t *pRExC_state, U8 op, regnode *opnd)
6990 register regnode *src;
6991 register regnode *dst;
6992 register regnode *place;
6993 const int offset = regarglen[(U8)op];
6994 GET_RE_DEBUG_FLAGS_DECL;
6995 /* (PL_regkind[(U8)op] == CURLY ? EXTRA_STEP_2ARGS : 0); */
6998 RExC_size += NODE_STEP_REGNODE + offset;
7003 RExC_emit += NODE_STEP_REGNODE + offset;
7005 while (src > opnd) {
7006 StructCopy(--src, --dst, regnode);
7007 if (RExC_offsets) { /* MJD 20010112 */
7008 MJD_OFFSET_DEBUG(("%s(%d): (op %s) %s copy %"UVuf" -> %"UVuf" (max %"UVuf").\n",
7012 (UV)(dst - RExC_emit_start) > RExC_offsets[0]
7013 ? "Overwriting end of array!\n" : "OK",
7014 (UV)(src - RExC_emit_start),
7015 (UV)(dst - RExC_emit_start),
7016 (UV)RExC_offsets[0]));
7017 Set_Node_Offset_To_R(dst-RExC_emit_start, Node_Offset(src));
7018 Set_Node_Length_To_R(dst-RExC_emit_start, Node_Length(src));
7023 place = opnd; /* Op node, where operand used to be. */
7024 if (RExC_offsets) { /* MJD */
7025 MJD_OFFSET_DEBUG(("%s(%d): (op %s) %s %"UVuf" <- %"UVuf" (max %"UVuf").\n",
7029 (UV)(place - RExC_emit_start) > RExC_offsets[0]
7030 ? "Overwriting end of array!\n" : "OK",
7031 (UV)(place - RExC_emit_start),
7032 (UV)(RExC_parse - RExC_start),
7033 (UV)RExC_offsets[0]));
7034 Set_Node_Offset(place, RExC_parse);
7035 Set_Node_Length(place, 1);
7037 src = NEXTOPER(place);
7038 FILL_ADVANCE_NODE(place, op);
7039 Zero(src, offset, regnode);
7043 - regtail - set the next-pointer at the end of a node chain of p to val.
7044 - SEE ALSO: regtail_study
7046 /* TODO: All three parms should be const */
7048 S_regtail(pTHX_ RExC_state_t *pRExC_state, regnode *p, const regnode *val,U32 depth)
7051 register regnode *scan;
7052 GET_RE_DEBUG_FLAGS_DECL;
7054 PERL_UNUSED_ARG(depth);
7060 /* Find last node. */
7063 regnode * const temp = regnext(scan);
7065 SV * const mysv=sv_newmortal();
7066 DEBUG_PARSE_MSG((scan==p ? "tail" : ""));
7067 regprop(RExC_rx, mysv, scan);
7068 PerlIO_printf(Perl_debug_log, "~ %s (%d)\n",
7069 SvPV_nolen_const(mysv), REG_NODE_NUM(scan));
7076 if (reg_off_by_arg[OP(scan)]) {
7077 ARG_SET(scan, val - scan);
7080 NEXT_OFF(scan) = val - scan;
7086 - regtail_study - set the next-pointer at the end of a node chain of p to val.
7087 - Look for optimizable sequences at the same time.
7088 - currently only looks for EXACT chains.
7090 This is expermental code. The idea is to use this routine to perform
7091 in place optimizations on branches and groups as they are constructed,
7092 with the long term intention of removing optimization from study_chunk so
7093 that it is purely analytical.
7095 Currently only used when in DEBUG mode. The macro REGTAIL_STUDY() is used
7096 to control which is which.
7099 /* TODO: All four parms should be const */
7102 S_regtail_study(pTHX_ RExC_state_t *pRExC_state, regnode *p, const regnode *val,U32 depth)
7105 register regnode *scan;
7107 #ifdef EXPERIMENTAL_INPLACESCAN
7111 GET_RE_DEBUG_FLAGS_DECL;
7117 /* Find last node. */
7121 regnode * const temp = regnext(scan);
7122 #ifdef EXPERIMENTAL_INPLACESCAN
7123 if (PL_regkind[OP(scan)] == EXACT)
7124 if (join_exact(pRExC_state,scan,&min,1,val,depth+1))
7132 if( exact == PSEUDO )
7134 else if ( exact != OP(scan) )
7143 SV * const mysv=sv_newmortal();
7144 DEBUG_PARSE_MSG((scan==p ? "tsdy" : ""));
7145 regprop(RExC_rx, mysv, scan);
7146 PerlIO_printf(Perl_debug_log, "~ %s (%s) (%d)\n",
7147 SvPV_nolen_const(mysv),
7149 REG_NODE_NUM(scan));
7156 SV * const mysv_val=sv_newmortal();
7157 DEBUG_PARSE_MSG("");
7158 regprop(RExC_rx, mysv_val, val);
7159 PerlIO_printf(Perl_debug_log, "~ attach to %s (%d) offset to %d\n",
7160 SvPV_nolen_const(mysv_val),
7165 if (reg_off_by_arg[OP(scan)]) {
7166 ARG_SET(scan, val - scan);
7169 NEXT_OFF(scan) = val - scan;
7177 - regcurly - a little FSA that accepts {\d+,?\d*}
7180 S_regcurly(register const char *s)
7199 - regdump - dump a regexp onto Perl_debug_log in vaguely comprehensible form
7202 Perl_regdump(pTHX_ const regexp *r)
7206 SV * const sv = sv_newmortal();
7207 SV *dsv= sv_newmortal();
7209 (void)dumpuntil(r, r->program, r->program + 1, NULL, NULL, sv, 0, 0);
7211 /* Header fields of interest. */
7212 if (r->anchored_substr) {
7213 RE_PV_QUOTED_DECL(s, 0, dsv, SvPVX_const(r->anchored_substr),
7214 RE_SV_DUMPLEN(r->anchored_substr), 30);
7215 PerlIO_printf(Perl_debug_log,
7216 "anchored %s%s at %"IVdf" ",
7217 s, RE_SV_TAIL(r->anchored_substr),
7218 (IV)r->anchored_offset);
7219 } else if (r->anchored_utf8) {
7220 RE_PV_QUOTED_DECL(s, 1, dsv, SvPVX_const(r->anchored_utf8),
7221 RE_SV_DUMPLEN(r->anchored_utf8), 30);
7222 PerlIO_printf(Perl_debug_log,
7223 "anchored utf8 %s%s at %"IVdf" ",
7224 s, RE_SV_TAIL(r->anchored_utf8),
7225 (IV)r->anchored_offset);
7227 if (r->float_substr) {
7228 RE_PV_QUOTED_DECL(s, 0, dsv, SvPVX_const(r->float_substr),
7229 RE_SV_DUMPLEN(r->float_substr), 30);
7230 PerlIO_printf(Perl_debug_log,
7231 "floating %s%s at %"IVdf"..%"UVuf" ",
7232 s, RE_SV_TAIL(r->float_substr),
7233 (IV)r->float_min_offset, (UV)r->float_max_offset);
7234 } else if (r->float_utf8) {
7235 RE_PV_QUOTED_DECL(s, 1, dsv, SvPVX_const(r->float_utf8),
7236 RE_SV_DUMPLEN(r->float_utf8), 30);
7237 PerlIO_printf(Perl_debug_log,
7238 "floating utf8 %s%s at %"IVdf"..%"UVuf" ",
7239 s, RE_SV_TAIL(r->float_utf8),
7240 (IV)r->float_min_offset, (UV)r->float_max_offset);
7242 if (r->check_substr || r->check_utf8)
7243 PerlIO_printf(Perl_debug_log,
7245 (r->check_substr == r->float_substr
7246 && r->check_utf8 == r->float_utf8
7247 ? "(checking floating" : "(checking anchored"));
7248 if (r->reganch & ROPT_NOSCAN)
7249 PerlIO_printf(Perl_debug_log, " noscan");
7250 if (r->reganch & ROPT_CHECK_ALL)
7251 PerlIO_printf(Perl_debug_log, " isall");
7252 if (r->check_substr || r->check_utf8)
7253 PerlIO_printf(Perl_debug_log, ") ");
7255 if (r->regstclass) {
7256 regprop(r, sv, r->regstclass);
7257 PerlIO_printf(Perl_debug_log, "stclass %s ", SvPVX_const(sv));
7259 if (r->reganch & ROPT_ANCH) {
7260 PerlIO_printf(Perl_debug_log, "anchored");
7261 if (r->reganch & ROPT_ANCH_BOL)
7262 PerlIO_printf(Perl_debug_log, "(BOL)");
7263 if (r->reganch & ROPT_ANCH_MBOL)
7264 PerlIO_printf(Perl_debug_log, "(MBOL)");
7265 if (r->reganch & ROPT_ANCH_SBOL)
7266 PerlIO_printf(Perl_debug_log, "(SBOL)");
7267 if (r->reganch & ROPT_ANCH_GPOS)
7268 PerlIO_printf(Perl_debug_log, "(GPOS)");
7269 PerlIO_putc(Perl_debug_log, ' ');
7271 if (r->reganch & ROPT_GPOS_SEEN)
7272 PerlIO_printf(Perl_debug_log, "GPOS ");
7273 if (r->reganch & ROPT_SKIP)
7274 PerlIO_printf(Perl_debug_log, "plus ");
7275 if (r->reganch & ROPT_IMPLICIT)
7276 PerlIO_printf(Perl_debug_log, "implicit ");
7277 PerlIO_printf(Perl_debug_log, "minlen %ld ", (long) r->minlen);
7278 if (r->reganch & ROPT_EVAL_SEEN)
7279 PerlIO_printf(Perl_debug_log, "with eval ");
7280 PerlIO_printf(Perl_debug_log, "\n");
7282 PERL_UNUSED_CONTEXT;
7284 #endif /* DEBUGGING */
7288 - regprop - printable representation of opcode
7291 Perl_regprop(pTHX_ const regexp *prog, SV *sv, const regnode *o)
7296 GET_RE_DEBUG_FLAGS_DECL;
7298 sv_setpvn(sv, "", 0);
7299 if (OP(o) > REGNODE_MAX) /* regnode.type is unsigned */
7300 /* It would be nice to FAIL() here, but this may be called from
7301 regexec.c, and it would be hard to supply pRExC_state. */
7302 Perl_croak(aTHX_ "Corrupted regexp opcode %d > %d", (int)OP(o), (int)REGNODE_MAX);
7303 sv_catpv(sv, reg_name[OP(o)]); /* Take off const! */
7305 k = PL_regkind[OP(o)];
7308 SV * const dsv = sv_2mortal(newSVpvs(""));
7309 /* Using is_utf8_string() (via PERL_PV_UNI_DETECT)
7310 * is a crude hack but it may be the best for now since
7311 * we have no flag "this EXACTish node was UTF-8"
7313 const char * const s =
7314 pv_pretty(dsv, STRING(o), STR_LEN(o), 60,
7315 PL_colors[0], PL_colors[1],
7316 PERL_PV_ESCAPE_UNI_DETECT |
7317 PERL_PV_PRETTY_ELIPSES |
7320 Perl_sv_catpvf(aTHX_ sv, " %s", s );
7321 } else if (k == TRIE) {
7322 /* print the details of the trie in dumpuntil instead, as
7323 * prog->data isn't available here */
7324 const char op = OP(o);
7325 const I32 n = ARG(o);
7326 const reg_ac_data * const ac = IS_TRIE_AC(op) ?
7327 (reg_ac_data *)prog->data->data[n] :
7329 const reg_trie_data * const trie = !IS_TRIE_AC(op) ?
7330 (reg_trie_data*)prog->data->data[n] :
7333 Perl_sv_catpvf(aTHX_ sv, "-%s",reg_name[o->flags]);
7334 DEBUG_TRIE_COMPILE_r(
7335 Perl_sv_catpvf(aTHX_ sv,
7336 "<S:%"UVuf"/%"IVdf" W:%"UVuf" L:%"UVuf"/%"UVuf" C:%"UVuf"/%"UVuf">",
7337 (UV)trie->startstate,
7338 (IV)trie->laststate-1,
7339 (UV)trie->wordcount,
7342 (UV)TRIE_CHARCOUNT(trie),
7343 (UV)trie->uniquecharcount
7346 if ( IS_ANYOF_TRIE(op) || trie->bitmap ) {
7348 int rangestart = -1;
7349 U8* bitmap = IS_ANYOF_TRIE(op) ? (U8*)ANYOF_BITMAP(o) : (U8*)TRIE_BITMAP(trie);
7350 Perl_sv_catpvf(aTHX_ sv, "[");
7351 for (i = 0; i <= 256; i++) {
7352 if (i < 256 && BITMAP_TEST(bitmap,i)) {
7353 if (rangestart == -1)
7355 } else if (rangestart != -1) {
7356 if (i <= rangestart + 3)
7357 for (; rangestart < i; rangestart++)
7358 put_byte(sv, rangestart);
7360 put_byte(sv, rangestart);
7362 put_byte(sv, i - 1);
7367 Perl_sv_catpvf(aTHX_ sv, "]");
7370 } else if (k == CURLY) {
7371 if (OP(o) == CURLYM || OP(o) == CURLYN || OP(o) == CURLYX)
7372 Perl_sv_catpvf(aTHX_ sv, "[%d]", o->flags); /* Parenth number */
7373 Perl_sv_catpvf(aTHX_ sv, " {%d,%d}", ARG1(o), ARG2(o));
7375 else if (k == WHILEM && o->flags) /* Ordinal/of */
7376 Perl_sv_catpvf(aTHX_ sv, "[%d/%d]", o->flags & 0xf, o->flags>>4);
7377 else if (k == REF || k == OPEN || k == CLOSE || k == GROUPP )
7378 Perl_sv_catpvf(aTHX_ sv, "%d", (int)ARG(o)); /* Parenth number */
7379 else if (k == LOGICAL)
7380 Perl_sv_catpvf(aTHX_ sv, "[%d]", o->flags); /* 2: embedded, otherwise 1 */
7381 else if (k == ANYOF) {
7382 int i, rangestart = -1;
7383 const U8 flags = ANYOF_FLAGS(o);
7385 /* Should be synchronized with * ANYOF_ #xdefines in regcomp.h */
7386 static const char * const anyofs[] = {
7419 if (flags & ANYOF_LOCALE)
7420 sv_catpvs(sv, "{loc}");
7421 if (flags & ANYOF_FOLD)
7422 sv_catpvs(sv, "{i}");
7423 Perl_sv_catpvf(aTHX_ sv, "[%s", PL_colors[0]);
7424 if (flags & ANYOF_INVERT)
7426 for (i = 0; i <= 256; i++) {
7427 if (i < 256 && ANYOF_BITMAP_TEST(o,i)) {
7428 if (rangestart == -1)
7430 } else if (rangestart != -1) {
7431 if (i <= rangestart + 3)
7432 for (; rangestart < i; rangestart++)
7433 put_byte(sv, rangestart);
7435 put_byte(sv, rangestart);
7437 put_byte(sv, i - 1);
7443 if (o->flags & ANYOF_CLASS)
7444 for (i = 0; i < (int)(sizeof(anyofs)/sizeof(char*)); i++)
7445 if (ANYOF_CLASS_TEST(o,i))
7446 sv_catpv(sv, anyofs[i]);
7448 if (flags & ANYOF_UNICODE)
7449 sv_catpvs(sv, "{unicode}");
7450 else if (flags & ANYOF_UNICODE_ALL)
7451 sv_catpvs(sv, "{unicode_all}");
7455 SV * const sw = regclass_swash(prog, o, FALSE, &lv, 0);
7459 U8 s[UTF8_MAXBYTES_CASE+1];
7461 for (i = 0; i <= 256; i++) { /* just the first 256 */
7462 uvchr_to_utf8(s, i);
7464 if (i < 256 && swash_fetch(sw, s, TRUE)) {
7465 if (rangestart == -1)
7467 } else if (rangestart != -1) {
7468 if (i <= rangestart + 3)
7469 for (; rangestart < i; rangestart++) {
7470 const U8 * const e = uvchr_to_utf8(s,rangestart);
7472 for(p = s; p < e; p++)
7476 const U8 *e = uvchr_to_utf8(s,rangestart);
7478 for (p = s; p < e; p++)
7481 e = uvchr_to_utf8(s, i-1);
7482 for (p = s; p < e; p++)
7489 sv_catpvs(sv, "..."); /* et cetera */
7493 char *s = savesvpv(lv);
7494 char * const origs = s;
7496 while (*s && *s != '\n')
7500 const char * const t = ++s;
7518 Perl_sv_catpvf(aTHX_ sv, "%s]", PL_colors[1]);
7520 else if (k == BRANCHJ && (OP(o) == UNLESSM || OP(o) == IFMATCH))
7521 Perl_sv_catpvf(aTHX_ sv, "[%d]", -(o->flags));
7523 PERL_UNUSED_CONTEXT;
7524 PERL_UNUSED_ARG(sv);
7526 PERL_UNUSED_ARG(prog);
7527 #endif /* DEBUGGING */
7531 Perl_re_intuit_string(pTHX_ regexp *prog)
7532 { /* Assume that RE_INTUIT is set */
7534 GET_RE_DEBUG_FLAGS_DECL;
7535 PERL_UNUSED_CONTEXT;
7539 const char * const s = SvPV_nolen_const(prog->check_substr
7540 ? prog->check_substr : prog->check_utf8);
7542 if (!PL_colorset) reginitcolors();
7543 PerlIO_printf(Perl_debug_log,
7544 "%sUsing REx %ssubstr:%s \"%s%.60s%s%s\"\n",
7546 prog->check_substr ? "" : "utf8 ",
7547 PL_colors[5],PL_colors[0],
7550 (strlen(s) > 60 ? "..." : ""));
7553 return prog->check_substr ? prog->check_substr : prog->check_utf8;
7557 pregfree - free a regexp
7559 See regdupe below if you change anything here.
7563 Perl_pregfree(pTHX_ struct regexp *r)
7567 GET_RE_DEBUG_FLAGS_DECL;
7569 if (!r || (--r->refcnt > 0))
7575 SV *dsv= sv_newmortal();
7576 RE_PV_QUOTED_DECL(s, (r->reganch & ROPT_UTF8),
7577 dsv, r->precomp, r->prelen, 60);
7578 PerlIO_printf(Perl_debug_log,"%sFreeing REx:%s %s\n",
7579 PL_colors[4],PL_colors[5],s);
7583 /* gcov results gave these as non-null 100% of the time, so there's no
7584 optimisation in checking them before calling Safefree */
7585 Safefree(r->precomp);
7586 Safefree(r->offsets); /* 20010421 MJD */
7587 RX_MATCH_COPY_FREE(r);
7588 #ifdef PERL_OLD_COPY_ON_WRITE
7590 SvREFCNT_dec(r->saved_copy);
7593 if (r->anchored_substr)
7594 SvREFCNT_dec(r->anchored_substr);
7595 if (r->anchored_utf8)
7596 SvREFCNT_dec(r->anchored_utf8);
7597 if (r->float_substr)
7598 SvREFCNT_dec(r->float_substr);
7600 SvREFCNT_dec(r->float_utf8);
7601 Safefree(r->substrs);
7604 int n = r->data->count;
7605 PAD* new_comppad = NULL;
7610 /* If you add a ->what type here, update the comment in regcomp.h */
7611 switch (r->data->what[n]) {
7613 SvREFCNT_dec((SV*)r->data->data[n]);
7616 Safefree(r->data->data[n]);
7619 new_comppad = (AV*)r->data->data[n];
7622 if (new_comppad == NULL)
7623 Perl_croak(aTHX_ "panic: pregfree comppad");
7624 PAD_SAVE_LOCAL(old_comppad,
7625 /* Watch out for global destruction's random ordering. */
7626 (SvTYPE(new_comppad) == SVt_PVAV) ? new_comppad : NULL
7629 refcnt = OpREFCNT_dec((OP_4tree*)r->data->data[n]);
7632 op_free((OP_4tree*)r->data->data[n]);
7634 PAD_RESTORE_LOCAL(old_comppad);
7635 SvREFCNT_dec((SV*)new_comppad);
7641 { /* Aho Corasick add-on structure for a trie node.
7642 Used in stclass optimization only */
7644 reg_ac_data *aho=(reg_ac_data*)r->data->data[n];
7646 refcount = --aho->refcount;
7649 Safefree(aho->states);
7650 Safefree(aho->fail);
7651 aho->trie=NULL; /* not necessary to free this as it is
7652 handled by the 't' case */
7653 Safefree(r->data->data[n]); /* do this last!!!! */
7654 Safefree(r->regstclass);
7660 /* trie structure. */
7662 reg_trie_data *trie=(reg_trie_data*)r->data->data[n];
7664 refcount = --trie->refcount;
7667 Safefree(trie->charmap);
7668 if (trie->widecharmap)
7669 SvREFCNT_dec((SV*)trie->widecharmap);
7670 Safefree(trie->states);
7671 Safefree(trie->trans);
7673 Safefree(trie->bitmap);
7675 Safefree(trie->wordlen);
7677 Safefree(trie->jump);
7679 Safefree(trie->nextword);
7683 SvREFCNT_dec((SV*)trie->words);
7684 if (trie->revcharmap)
7685 SvREFCNT_dec((SV*)trie->revcharmap);
7688 Safefree(r->data->data[n]); /* do this last!!!! */
7693 Perl_croak(aTHX_ "panic: regfree data code '%c'", r->data->what[n]);
7696 Safefree(r->data->what);
7699 Safefree(r->startp);
7704 #define sv_dup_inc(s,t) SvREFCNT_inc(sv_dup(s,t))
7705 #define av_dup_inc(s,t) (AV*)SvREFCNT_inc(sv_dup((SV*)s,t))
7706 #define SAVEPVN(p,n) ((p) ? savepvn(p,n) : NULL)
7709 regdupe - duplicate a regexp.
7711 This routine is called by sv.c's re_dup and is expected to clone a
7712 given regexp structure. It is a no-op when not under USE_ITHREADS.
7713 (Originally this *was* re_dup() for change history see sv.c)
7715 See pregfree() above if you change anything here.
7717 #if defined(USE_ITHREADS)
7719 Perl_regdupe(pTHX_ const regexp *r, CLONE_PARAMS *param)
7724 struct reg_substr_datum *s;
7727 return (REGEXP *)NULL;
7729 if ((ret = (REGEXP *)ptr_table_fetch(PL_ptr_table, r)))
7732 len = r->offsets[0];
7733 npar = r->nparens+1;
7735 Newxc(ret, sizeof(regexp) + (len+1)*sizeof(regnode), char, regexp);
7736 Copy(r->program, ret->program, len+1, regnode);
7738 Newx(ret->startp, npar, I32);
7739 Copy(r->startp, ret->startp, npar, I32);
7740 Newx(ret->endp, npar, I32);
7741 Copy(r->startp, ret->startp, npar, I32);
7743 Newx(ret->substrs, 1, struct reg_substr_data);
7744 for (s = ret->substrs->data, i = 0; i < 3; i++, s++) {
7745 s->min_offset = r->substrs->data[i].min_offset;
7746 s->max_offset = r->substrs->data[i].max_offset;
7747 s->end_shift = r->substrs->data[i].end_shift;
7748 s->substr = sv_dup_inc(r->substrs->data[i].substr, param);
7749 s->utf8_substr = sv_dup_inc(r->substrs->data[i].utf8_substr, param);
7752 ret->regstclass = NULL;
7755 const int count = r->data->count;
7758 Newxc(d, sizeof(struct reg_data) + count*sizeof(void *),
7759 char, struct reg_data);
7760 Newx(d->what, count, U8);
7763 for (i = 0; i < count; i++) {
7764 d->what[i] = r->data->what[i];
7765 switch (d->what[i]) {
7766 /* legal options are one of: sfpont
7767 see also regcomp.h and pregfree() */
7769 d->data[i] = sv_dup_inc((SV *)r->data->data[i], param);
7772 d->data[i] = av_dup_inc((AV *)r->data->data[i], param);
7775 /* This is cheating. */
7776 Newx(d->data[i], 1, struct regnode_charclass_class);
7777 StructCopy(r->data->data[i], d->data[i],
7778 struct regnode_charclass_class);
7779 ret->regstclass = (regnode*)d->data[i];
7782 /* Compiled op trees are readonly, and can thus be
7783 shared without duplication. */
7785 d->data[i] = (void*)OpREFCNT_inc((OP*)r->data->data[i]);
7789 d->data[i] = r->data->data[i];
7792 d->data[i] = r->data->data[i];
7794 ((reg_trie_data*)d->data[i])->refcount++;
7798 d->data[i] = r->data->data[i];
7800 ((reg_ac_data*)d->data[i])->refcount++;
7802 /* Trie stclasses are readonly and can thus be shared
7803 * without duplication. We free the stclass in pregfree
7804 * when the corresponding reg_ac_data struct is freed.
7806 ret->regstclass= r->regstclass;
7809 Perl_croak(aTHX_ "panic: re_dup unknown data code '%c'", r->data->what[i]);
7818 Newx(ret->offsets, 2*len+1, U32);
7819 Copy(r->offsets, ret->offsets, 2*len+1, U32);
7821 ret->precomp = SAVEPVN(r->precomp, r->prelen);
7822 ret->refcnt = r->refcnt;
7823 ret->minlen = r->minlen;
7824 ret->prelen = r->prelen;
7825 ret->nparens = r->nparens;
7826 ret->lastparen = r->lastparen;
7827 ret->lastcloseparen = r->lastcloseparen;
7828 ret->reganch = r->reganch;
7830 ret->sublen = r->sublen;
7832 ret->engine = r->engine;
7834 if (RX_MATCH_COPIED(ret))
7835 ret->subbeg = SAVEPVN(r->subbeg, r->sublen);
7838 #ifdef PERL_OLD_COPY_ON_WRITE
7839 ret->saved_copy = NULL;
7842 ptr_table_store(PL_ptr_table, r, ret);
7847 #ifndef PERL_IN_XSUB_RE
7849 - regnext - dig the "next" pointer out of a node
7852 Perl_regnext(pTHX_ register regnode *p)
7855 register I32 offset;
7857 if (p == &PL_regdummy)
7860 offset = (reg_off_by_arg[OP(p)] ? ARG(p) : NEXT_OFF(p));
7869 S_re_croak2(pTHX_ const char* pat1,const char* pat2,...)
7872 STRLEN l1 = strlen(pat1);
7873 STRLEN l2 = strlen(pat2);
7876 const char *message;
7882 Copy(pat1, buf, l1 , char);
7883 Copy(pat2, buf + l1, l2 , char);
7884 buf[l1 + l2] = '\n';
7885 buf[l1 + l2 + 1] = '\0';
7887 /* ANSI variant takes additional second argument */
7888 va_start(args, pat2);
7892 msv = vmess(buf, &args);
7894 message = SvPV_const(msv,l1);
7897 Copy(message, buf, l1 , char);
7898 buf[l1-1] = '\0'; /* Overwrite \n */
7899 Perl_croak(aTHX_ "%s", buf);
7902 /* XXX Here's a total kludge. But we need to re-enter for swash routines. */
7904 #ifndef PERL_IN_XSUB_RE
7906 Perl_save_re_context(pTHX)
7910 struct re_save_state *state;
7912 SAVEVPTR(PL_curcop);
7913 SSGROW(SAVESTACK_ALLOC_FOR_RE_SAVE_STATE + 1);
7915 state = (struct re_save_state *)(PL_savestack + PL_savestack_ix);
7916 PL_savestack_ix += SAVESTACK_ALLOC_FOR_RE_SAVE_STATE;
7917 SSPUSHINT(SAVEt_RE_STATE);
7919 Copy(&PL_reg_state, state, 1, struct re_save_state);
7921 PL_reg_start_tmp = 0;
7922 PL_reg_start_tmpl = 0;
7923 PL_reg_oldsaved = NULL;
7924 PL_reg_oldsavedlen = 0;
7926 PL_reg_leftiter = 0;
7927 PL_reg_poscache = NULL;
7928 PL_reg_poscache_size = 0;
7929 #ifdef PERL_OLD_COPY_ON_WRITE
7933 /* Save $1..$n (#18107: UTF-8 s/(\w+)/uc($1)/e); AMS 20021106. */
7935 const REGEXP * const rx = PM_GETRE(PL_curpm);
7938 for (i = 1; i <= rx->nparens; i++) {
7939 char digits[TYPE_CHARS(long)];
7940 const STRLEN len = my_snprintf(digits, sizeof(digits), "%lu", (long)i);
7941 GV *const *const gvp
7942 = (GV**)hv_fetch(PL_defstash, digits, len, 0);
7945 GV * const gv = *gvp;
7946 if (SvTYPE(gv) == SVt_PVGV && GvSV(gv))
7956 clear_re(pTHX_ void *r)
7959 ReREFCNT_dec((regexp *)r);
7965 S_put_byte(pTHX_ SV *sv, int c)
7967 if (isCNTRL(c) || c == 255 || !isPRINT(c))
7968 Perl_sv_catpvf(aTHX_ sv, "\\%o", c);
7969 else if (c == '-' || c == ']' || c == '\\' || c == '^')
7970 Perl_sv_catpvf(aTHX_ sv, "\\%c", c);
7972 Perl_sv_catpvf(aTHX_ sv, "%c", c);
7976 #define CLEAR_OPTSTART \
7977 if (optstart) STMT_START { \
7978 DEBUG_OPTIMISE_r(PerlIO_printf(Perl_debug_log, " (%d nodes)\n", node - optstart)); \
7982 #define DUMPUNTIL(b,e) CLEAR_OPTSTART; node=dumpuntil(r,start,(b),(e),last,sv,indent+1,depth+1);
7984 STATIC const regnode *
7985 S_dumpuntil(pTHX_ const regexp *r, const regnode *start, const regnode *node,
7986 const regnode *last, const regnode *plast,
7987 SV* sv, I32 indent, U32 depth)
7990 register U8 op = PSEUDO; /* Arbitrary non-END op. */
7991 register const regnode *next;
7992 const regnode *optstart= NULL;
7993 GET_RE_DEBUG_FLAGS_DECL;
7995 #ifdef DEBUG_DUMPUNTIL
7996 PerlIO_printf(Perl_debug_log, "--- %d : %d - %d - %d\n",indent,node-start,
7997 last ? last-start : 0,plast ? plast-start : 0);
8000 if (plast && plast < last)
8003 while (PL_regkind[op] != END && (!last || node < last)) {
8004 /* While that wasn't END last time... */
8010 next = regnext((regnode *)node);
8013 if (OP(node) == OPTIMIZED) {
8014 if (!optstart && RE_DEBUG_FLAG(RE_DEBUG_COMPILE_OPTIMISE))
8021 regprop(r, sv, node);
8022 PerlIO_printf(Perl_debug_log, "%4"IVdf":%*s%s", (IV)(node - start),
8023 (int)(2*indent + 1), "", SvPVX_const(sv));
8025 if (OP(node) != OPTIMIZED) {
8026 if (next == NULL) /* Next ptr. */
8027 PerlIO_printf(Perl_debug_log, "(0)");
8028 else if (PL_regkind[(U8)op] == BRANCH && PL_regkind[OP(next)] != BRANCH )
8029 PerlIO_printf(Perl_debug_log, "(FAIL)");
8031 PerlIO_printf(Perl_debug_log, "(%"IVdf")", (IV)(next - start));
8033 /*if (PL_regkind[(U8)op] != TRIE)*/
8034 (void)PerlIO_putc(Perl_debug_log, '\n');
8038 if (PL_regkind[(U8)op] == BRANCHJ) {
8041 register const regnode *nnode = (OP(next) == LONGJMP
8042 ? regnext((regnode *)next)
8044 if (last && nnode > last)
8046 DUMPUNTIL(NEXTOPER(NEXTOPER(node)), nnode);
8049 else if (PL_regkind[(U8)op] == BRANCH) {
8051 DUMPUNTIL(NEXTOPER(node), next);
8053 else if ( PL_regkind[(U8)op] == TRIE ) {
8054 const char op = OP(node);
8055 const I32 n = ARG(node);
8056 const reg_ac_data * const ac = op>=AHOCORASICK ?
8057 (reg_ac_data *)r->data->data[n] :
8059 const reg_trie_data * const trie = op<AHOCORASICK ?
8060 (reg_trie_data*)r->data->data[n] :
8062 const regnode *nextbranch= NULL;
8064 sv_setpvn(sv, "", 0);
8065 for (word_idx= 0; word_idx < (I32)trie->wordcount; word_idx++) {
8066 SV ** const elem_ptr = av_fetch(trie->words,word_idx,0);
8068 PerlIO_printf(Perl_debug_log, "%*s%s ",
8069 (int)(2*(indent+3)), "",
8070 elem_ptr ? pv_pretty(sv, SvPV_nolen_const(*elem_ptr), SvCUR(*elem_ptr), 60,
8071 PL_colors[0], PL_colors[1],
8072 (SvUTF8(*elem_ptr) ? PERL_PV_ESCAPE_UNI : 0) |
8073 PERL_PV_PRETTY_ELIPSES |
8079 U16 dist= trie->jump[word_idx+1];
8080 PerlIO_printf(Perl_debug_log, "(%u)\n",(next - dist) - start);
8083 nextbranch= next - trie->jump[0];
8084 DUMPUNTIL(next - dist, nextbranch);
8086 if (nextbranch && PL_regkind[OP(nextbranch)]==BRANCH)
8087 nextbranch= regnext((regnode *)nextbranch);
8089 PerlIO_printf(Perl_debug_log, "\n");
8092 if (last && next > last)
8097 else if ( op == CURLY ) { /* "next" might be very big: optimizer */
8098 DUMPUNTIL(NEXTOPER(node) + EXTRA_STEP_2ARGS,
8099 NEXTOPER(node) + EXTRA_STEP_2ARGS + 1);
8101 else if (PL_regkind[(U8)op] == CURLY && op != CURLYX) {
8103 DUMPUNTIL(NEXTOPER(node) + EXTRA_STEP_2ARGS, next);
8105 else if ( op == PLUS || op == STAR) {
8106 DUMPUNTIL(NEXTOPER(node), NEXTOPER(node) + 1);
8108 else if (op == ANYOF) {
8109 /* arglen 1 + class block */
8110 node += 1 + ((ANYOF_FLAGS(node) & ANYOF_LARGE)
8111 ? ANYOF_CLASS_SKIP : ANYOF_SKIP);
8112 node = NEXTOPER(node);
8114 else if (PL_regkind[(U8)op] == EXACT) {
8115 /* Literal string, where present. */
8116 node += NODE_SZ_STR(node) - 1;
8117 node = NEXTOPER(node);
8120 node = NEXTOPER(node);
8121 node += regarglen[(U8)op];
8123 if (op == CURLYX || op == OPEN)
8125 else if (op == WHILEM)
8129 #ifdef DEBUG_DUMPUNTIL
8130 PerlIO_printf(Perl_debug_log, "--- %d\n",indent);
8135 #endif /* DEBUGGING */
8139 * c-indentation-style: bsd
8141 * indent-tabs-mode: t
8144 * ex: set ts=8 sts=4 sw=4 noet: