5 * 'A fair jaw-cracker dwarf-language must be.' --Samwise Gamgee
7 * [p.285 of _The Lord of the Rings_, II/iii: "The Ring Goes South"]
10 /* This file contains functions for compiling a regular expression. See
11 * also regexec.c which funnily enough, contains functions for executing
12 * a regular expression.
14 * This file is also copied at build time to ext/re/re_comp.c, where
15 * it's built with -DPERL_EXT_RE_BUILD -DPERL_EXT_RE_DEBUG -DPERL_EXT.
16 * This causes the main functions to be compiled under new names and with
17 * debugging support added, which makes "use re 'debug'" work.
20 /* NOTE: this is derived from Henry Spencer's regexp code, and should not
21 * confused with the original package (see point 3 below). Thanks, Henry!
24 /* Additional note: this code is very heavily munged from Henry's version
25 * in places. In some spots I've traded clarity for efficiency, so don't
26 * blame Henry for some of the lack of readability.
29 /* The names of the functions have been changed from regcomp and
30 * regexec to pregcomp and pregexec in order to avoid conflicts
31 * with the POSIX routines of the same names.
34 #ifdef PERL_EXT_RE_BUILD
39 * pregcomp and pregexec -- regsub and regerror are not used in perl
41 * Copyright (c) 1986 by University of Toronto.
42 * Written by Henry Spencer. Not derived from licensed software.
44 * Permission is granted to anyone to use this software for any
45 * purpose on any computer system, and to redistribute it freely,
46 * subject to the following restrictions:
48 * 1. The author is not responsible for the consequences of use of
49 * this software, no matter how awful, even if they arise
52 * 2. The origin of this software must not be misrepresented, either
53 * by explicit claim or by omission.
55 * 3. Altered versions must be plainly marked as such, and must not
56 * be misrepresented as being the original software.
59 **** Alterations to Henry's code are...
61 **** Copyright (C) 1991, 1992, 1993, 1994, 1995, 1996, 1997, 1998, 1999,
62 **** 2000, 2001, 2002, 2003, 2004, 2005, 2006, 2007, 2008
63 **** by Larry Wall and others
65 **** You may distribute under the terms of either the GNU General Public
66 **** License or the Artistic License, as specified in the README file.
69 * Beware that some of this code is subtly aware of the way operator
70 * precedence is structured in regular expressions. Serious changes in
71 * regular-expression syntax might require a total rethink.
74 #define PERL_IN_REGCOMP_C
77 #ifndef PERL_IN_XSUB_RE
82 #ifdef PERL_IN_XSUB_RE
93 # if defined(BUGGY_MSC6)
94 /* MSC 6.00A breaks on op/regexp.t test 85 unless we turn this off */
95 # pragma optimize("a",off)
96 /* But MSC 6.00A is happy with 'w', for aliases only across function calls*/
97 # pragma optimize("w",on )
98 # endif /* BUGGY_MSC6 */
102 #define STATIC static
105 typedef struct RExC_state_t {
106 U32 flags; /* are we folding, multilining? */
107 char *precomp; /* uncompiled string. */
108 REGEXP *rx_sv; /* The SV that is the regexp. */
109 regexp *rx; /* perl core regexp structure */
110 regexp_internal *rxi; /* internal data for regexp object pprivate field */
111 char *start; /* Start of input for compile */
112 char *end; /* End of input for compile */
113 char *parse; /* Input-scan pointer. */
114 I32 whilem_seen; /* number of WHILEM in this expr */
115 regnode *emit_start; /* Start of emitted-code area */
116 regnode *emit_bound; /* First regnode outside of the allocated space */
117 regnode *emit; /* Code-emit pointer; ®dummy = don't = compiling */
118 I32 naughty; /* How bad is this pattern? */
119 I32 sawback; /* Did we see \1, ...? */
121 I32 size; /* Code size. */
122 I32 npar; /* Capture buffer count, (OPEN). */
123 I32 cpar; /* Capture buffer count, (CLOSE). */
124 I32 nestroot; /* root parens we are in - used by accept */
128 regnode **open_parens; /* pointers to open parens */
129 regnode **close_parens; /* pointers to close parens */
130 regnode *opend; /* END node in program */
131 I32 utf8; /* whether the pattern is utf8 or not */
132 I32 orig_utf8; /* whether the pattern was originally in utf8 */
133 /* XXX use this for future optimisation of case
134 * where pattern must be upgraded to utf8. */
135 HV *paren_names; /* Paren names */
137 regnode **recurse; /* Recurse regops */
138 I32 recurse_count; /* Number of recurse regops */
140 char *starttry; /* -Dr: where regtry was called. */
141 #define RExC_starttry (pRExC_state->starttry)
144 const char *lastparse;
146 AV *paren_name_list; /* idx -> name */
147 #define RExC_lastparse (pRExC_state->lastparse)
148 #define RExC_lastnum (pRExC_state->lastnum)
149 #define RExC_paren_name_list (pRExC_state->paren_name_list)
153 #define RExC_flags (pRExC_state->flags)
154 #define RExC_precomp (pRExC_state->precomp)
155 #define RExC_rx_sv (pRExC_state->rx_sv)
156 #define RExC_rx (pRExC_state->rx)
157 #define RExC_rxi (pRExC_state->rxi)
158 #define RExC_start (pRExC_state->start)
159 #define RExC_end (pRExC_state->end)
160 #define RExC_parse (pRExC_state->parse)
161 #define RExC_whilem_seen (pRExC_state->whilem_seen)
162 #ifdef RE_TRACK_PATTERN_OFFSETS
163 #define RExC_offsets (pRExC_state->rxi->u.offsets) /* I am not like the others */
165 #define RExC_emit (pRExC_state->emit)
166 #define RExC_emit_start (pRExC_state->emit_start)
167 #define RExC_emit_bound (pRExC_state->emit_bound)
168 #define RExC_naughty (pRExC_state->naughty)
169 #define RExC_sawback (pRExC_state->sawback)
170 #define RExC_seen (pRExC_state->seen)
171 #define RExC_size (pRExC_state->size)
172 #define RExC_npar (pRExC_state->npar)
173 #define RExC_nestroot (pRExC_state->nestroot)
174 #define RExC_extralen (pRExC_state->extralen)
175 #define RExC_seen_zerolen (pRExC_state->seen_zerolen)
176 #define RExC_seen_evals (pRExC_state->seen_evals)
177 #define RExC_utf8 (pRExC_state->utf8)
178 #define RExC_orig_utf8 (pRExC_state->orig_utf8)
179 #define RExC_open_parens (pRExC_state->open_parens)
180 #define RExC_close_parens (pRExC_state->close_parens)
181 #define RExC_opend (pRExC_state->opend)
182 #define RExC_paren_names (pRExC_state->paren_names)
183 #define RExC_recurse (pRExC_state->recurse)
184 #define RExC_recurse_count (pRExC_state->recurse_count)
187 #define ISMULT1(c) ((c) == '*' || (c) == '+' || (c) == '?')
188 #define ISMULT2(s) ((*s) == '*' || (*s) == '+' || (*s) == '?' || \
189 ((*s) == '{' && regcurly(s)))
192 #undef SPSTART /* dratted cpp namespace... */
195 * Flags to be passed up and down.
197 #define WORST 0 /* Worst case. */
198 #define HASWIDTH 0x01 /* Known to match non-null strings. */
200 /* Simple enough to be STAR/PLUS operand, in an EXACT node must be a single
201 * character, and if utf8, must be invariant. */
203 #define SPSTART 0x04 /* Starts with * or +. */
204 #define TRYAGAIN 0x08 /* Weeded out a declaration. */
205 #define POSTPONED 0x10 /* (?1),(?&name), (??{...}) or similar */
207 #define REG_NODE_NUM(x) ((x) ? (int)((x)-RExC_emit_start) : -1)
209 /* whether trie related optimizations are enabled */
210 #if PERL_ENABLE_EXTENDED_TRIE_OPTIMISATION
211 #define TRIE_STUDY_OPT
212 #define FULL_TRIE_STUDY
218 #define PBYTE(u8str,paren) ((U8*)(u8str))[(paren) >> 3]
219 #define PBITVAL(paren) (1 << ((paren) & 7))
220 #define PAREN_TEST(u8str,paren) ( PBYTE(u8str,paren) & PBITVAL(paren))
221 #define PAREN_SET(u8str,paren) PBYTE(u8str,paren) |= PBITVAL(paren)
222 #define PAREN_UNSET(u8str,paren) PBYTE(u8str,paren) &= (~PBITVAL(paren))
224 /* If not already in utf8, do a longjmp back to the beginning */
225 #define UTF8_LONGJMP 42 /* Choose a value not likely to ever conflict */
226 #define REQUIRE_UTF8 STMT_START { \
227 if (! UTF) JMPENV_JUMP(UTF8_LONGJMP); \
230 /* About scan_data_t.
232 During optimisation we recurse through the regexp program performing
233 various inplace (keyhole style) optimisations. In addition study_chunk
234 and scan_commit populate this data structure with information about
235 what strings MUST appear in the pattern. We look for the longest
236 string that must appear for at a fixed location, and we look for the
237 longest string that may appear at a floating location. So for instance
242 Both 'FOO' and 'A' are fixed strings. Both 'B' and 'BAR' are floating
243 strings (because they follow a .* construct). study_chunk will identify
244 both FOO and BAR as being the longest fixed and floating strings respectively.
246 The strings can be composites, for instance
250 will result in a composite fixed substring 'foo'.
252 For each string some basic information is maintained:
254 - offset or min_offset
255 This is the position the string must appear at, or not before.
256 It also implicitly (when combined with minlenp) tells us how many
257 character must match before the string we are searching.
258 Likewise when combined with minlenp and the length of the string
259 tells us how many characters must appear after the string we have
263 Only used for floating strings. This is the rightmost point that
264 the string can appear at. Ifset to I32 max it indicates that the
265 string can occur infinitely far to the right.
268 A pointer to the minimum length of the pattern that the string
269 was found inside. This is important as in the case of positive
270 lookahead or positive lookbehind we can have multiple patterns
275 The minimum length of the pattern overall is 3, the minimum length
276 of the lookahead part is 3, but the minimum length of the part that
277 will actually match is 1. So 'FOO's minimum length is 3, but the
278 minimum length for the F is 1. This is important as the minimum length
279 is used to determine offsets in front of and behind the string being
280 looked for. Since strings can be composites this is the length of the
281 pattern at the time it was commited with a scan_commit. Note that
282 the length is calculated by study_chunk, so that the minimum lengths
283 are not known until the full pattern has been compiled, thus the
284 pointer to the value.
288 In the case of lookbehind the string being searched for can be
289 offset past the start point of the final matching string.
290 If this value was just blithely removed from the min_offset it would
291 invalidate some of the calculations for how many chars must match
292 before or after (as they are derived from min_offset and minlen and
293 the length of the string being searched for).
294 When the final pattern is compiled and the data is moved from the
295 scan_data_t structure into the regexp structure the information
296 about lookbehind is factored in, with the information that would
297 have been lost precalculated in the end_shift field for the
300 The fields pos_min and pos_delta are used to store the minimum offset
301 and the delta to the maximum offset at the current point in the pattern.
305 typedef struct scan_data_t {
306 /*I32 len_min; unused */
307 /*I32 len_delta; unused */
311 I32 last_end; /* min value, <0 unless valid. */
314 SV **longest; /* Either &l_fixed, or &l_float. */
315 SV *longest_fixed; /* longest fixed string found in pattern */
316 I32 offset_fixed; /* offset where it starts */
317 I32 *minlen_fixed; /* pointer to the minlen relevent to the string */
318 I32 lookbehind_fixed; /* is the position of the string modfied by LB */
319 SV *longest_float; /* longest floating string found in pattern */
320 I32 offset_float_min; /* earliest point in string it can appear */
321 I32 offset_float_max; /* latest point in string it can appear */
322 I32 *minlen_float; /* pointer to the minlen relevent to the string */
323 I32 lookbehind_float; /* is the position of the string modified by LB */
327 struct regnode_charclass_class *start_class;
331 * Forward declarations for pregcomp()'s friends.
334 static const scan_data_t zero_scan_data =
335 { 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0 ,0};
337 #define SF_BEFORE_EOL (SF_BEFORE_SEOL|SF_BEFORE_MEOL)
338 #define SF_BEFORE_SEOL 0x0001
339 #define SF_BEFORE_MEOL 0x0002
340 #define SF_FIX_BEFORE_EOL (SF_FIX_BEFORE_SEOL|SF_FIX_BEFORE_MEOL)
341 #define SF_FL_BEFORE_EOL (SF_FL_BEFORE_SEOL|SF_FL_BEFORE_MEOL)
344 # define SF_FIX_SHIFT_EOL (0+2)
345 # define SF_FL_SHIFT_EOL (0+4)
347 # define SF_FIX_SHIFT_EOL (+2)
348 # define SF_FL_SHIFT_EOL (+4)
351 #define SF_FIX_BEFORE_SEOL (SF_BEFORE_SEOL << SF_FIX_SHIFT_EOL)
352 #define SF_FIX_BEFORE_MEOL (SF_BEFORE_MEOL << SF_FIX_SHIFT_EOL)
354 #define SF_FL_BEFORE_SEOL (SF_BEFORE_SEOL << SF_FL_SHIFT_EOL)
355 #define SF_FL_BEFORE_MEOL (SF_BEFORE_MEOL << SF_FL_SHIFT_EOL) /* 0x20 */
356 #define SF_IS_INF 0x0040
357 #define SF_HAS_PAR 0x0080
358 #define SF_IN_PAR 0x0100
359 #define SF_HAS_EVAL 0x0200
360 #define SCF_DO_SUBSTR 0x0400
361 #define SCF_DO_STCLASS_AND 0x0800
362 #define SCF_DO_STCLASS_OR 0x1000
363 #define SCF_DO_STCLASS (SCF_DO_STCLASS_AND|SCF_DO_STCLASS_OR)
364 #define SCF_WHILEM_VISITED_POS 0x2000
366 #define SCF_TRIE_RESTUDY 0x4000 /* Do restudy? */
367 #define SCF_SEEN_ACCEPT 0x8000
369 #define UTF (RExC_utf8 != 0)
370 #define LOC ((RExC_flags & RXf_PMf_LOCALE) != 0)
371 #define FOLD ((RExC_flags & RXf_PMf_FOLD) != 0)
373 #define OOB_UNICODE 12345678
374 #define OOB_NAMEDCLASS -1
376 #define CHR_SVLEN(sv) (UTF ? sv_len_utf8(sv) : SvCUR(sv))
377 #define CHR_DIST(a,b) (UTF ? utf8_distance(a,b) : a - b)
380 /* length of regex to show in messages that don't mark a position within */
381 #define RegexLengthToShowInErrorMessages 127
384 * If MARKER[12] are adjusted, be sure to adjust the constants at the top
385 * of t/op/regmesg.t, the tests in t/op/re_tests, and those in
386 * op/pragma/warn/regcomp.
388 #define MARKER1 "<-- HERE" /* marker as it appears in the description */
389 #define MARKER2 " <-- HERE " /* marker as it appears within the regex */
391 #define REPORT_LOCATION " in regex; marked by " MARKER1 " in m/%.*s" MARKER2 "%s/"
394 * Calls SAVEDESTRUCTOR_X if needed, then calls Perl_croak with the given
395 * arg. Show regex, up to a maximum length. If it's too long, chop and add
398 #define _FAIL(code) STMT_START { \
399 const char *ellipses = ""; \
400 IV len = RExC_end - RExC_precomp; \
403 SAVEDESTRUCTOR_X(clear_re,(void*)RExC_rx_sv); \
404 if (len > RegexLengthToShowInErrorMessages) { \
405 /* chop 10 shorter than the max, to ensure meaning of "..." */ \
406 len = RegexLengthToShowInErrorMessages - 10; \
412 #define FAIL(msg) _FAIL( \
413 Perl_croak(aTHX_ "%s in regex m/%.*s%s/", \
414 msg, (int)len, RExC_precomp, ellipses))
416 #define FAIL2(msg,arg) _FAIL( \
417 Perl_croak(aTHX_ msg " in regex m/%.*s%s/", \
418 arg, (int)len, RExC_precomp, ellipses))
421 * Simple_vFAIL -- like FAIL, but marks the current location in the scan
423 #define Simple_vFAIL(m) STMT_START { \
424 const IV offset = RExC_parse - RExC_precomp; \
425 Perl_croak(aTHX_ "%s" REPORT_LOCATION, \
426 m, (int)offset, RExC_precomp, RExC_precomp + offset); \
430 * Calls SAVEDESTRUCTOR_X if needed, then Simple_vFAIL()
432 #define vFAIL(m) STMT_START { \
434 SAVEDESTRUCTOR_X(clear_re,(void*)RExC_rx_sv); \
439 * Like Simple_vFAIL(), but accepts two arguments.
441 #define Simple_vFAIL2(m,a1) STMT_START { \
442 const IV offset = RExC_parse - RExC_precomp; \
443 S_re_croak2(aTHX_ m, REPORT_LOCATION, a1, \
444 (int)offset, RExC_precomp, RExC_precomp + offset); \
448 * Calls SAVEDESTRUCTOR_X if needed, then Simple_vFAIL2().
450 #define vFAIL2(m,a1) STMT_START { \
452 SAVEDESTRUCTOR_X(clear_re,(void*)RExC_rx_sv); \
453 Simple_vFAIL2(m, a1); \
458 * Like Simple_vFAIL(), but accepts three arguments.
460 #define Simple_vFAIL3(m, a1, a2) STMT_START { \
461 const IV offset = RExC_parse - RExC_precomp; \
462 S_re_croak2(aTHX_ m, REPORT_LOCATION, a1, a2, \
463 (int)offset, RExC_precomp, RExC_precomp + offset); \
467 * Calls SAVEDESTRUCTOR_X if needed, then Simple_vFAIL3().
469 #define vFAIL3(m,a1,a2) STMT_START { \
471 SAVEDESTRUCTOR_X(clear_re,(void*)RExC_rx_sv); \
472 Simple_vFAIL3(m, a1, a2); \
476 * Like Simple_vFAIL(), but accepts four arguments.
478 #define Simple_vFAIL4(m, a1, a2, a3) STMT_START { \
479 const IV offset = RExC_parse - RExC_precomp; \
480 S_re_croak2(aTHX_ m, REPORT_LOCATION, a1, a2, a3, \
481 (int)offset, RExC_precomp, RExC_precomp + offset); \
484 #define ckWARNreg(loc,m) STMT_START { \
485 const IV offset = loc - RExC_precomp; \
486 Perl_ck_warner(aTHX_ packWARN(WARN_REGEXP), m REPORT_LOCATION, \
487 (int)offset, RExC_precomp, RExC_precomp + offset); \
490 #define ckWARNregdep(loc,m) STMT_START { \
491 const IV offset = loc - RExC_precomp; \
492 Perl_ck_warner_d(aTHX_ packWARN2(WARN_DEPRECATED, WARN_REGEXP), \
494 (int)offset, RExC_precomp, RExC_precomp + offset); \
497 #define ckWARN2reg(loc, m, a1) STMT_START { \
498 const IV offset = loc - RExC_precomp; \
499 Perl_ck_warner(aTHX_ packWARN(WARN_REGEXP), m REPORT_LOCATION, \
500 a1, (int)offset, RExC_precomp, RExC_precomp + offset); \
503 #define vWARN3(loc, m, a1, a2) STMT_START { \
504 const IV offset = loc - RExC_precomp; \
505 Perl_warner(aTHX_ packWARN(WARN_REGEXP), m REPORT_LOCATION, \
506 a1, a2, (int)offset, RExC_precomp, RExC_precomp + offset); \
509 #define ckWARN3reg(loc, m, a1, a2) STMT_START { \
510 const IV offset = loc - RExC_precomp; \
511 Perl_ck_warner(aTHX_ packWARN(WARN_REGEXP), m REPORT_LOCATION, \
512 a1, a2, (int)offset, RExC_precomp, RExC_precomp + offset); \
515 #define vWARN4(loc, m, a1, a2, a3) STMT_START { \
516 const IV offset = loc - RExC_precomp; \
517 Perl_warner(aTHX_ packWARN(WARN_REGEXP), m REPORT_LOCATION, \
518 a1, a2, a3, (int)offset, RExC_precomp, RExC_precomp + offset); \
521 #define ckWARN4reg(loc, m, a1, a2, a3) STMT_START { \
522 const IV offset = loc - RExC_precomp; \
523 Perl_ck_warner(aTHX_ packWARN(WARN_REGEXP), m REPORT_LOCATION, \
524 a1, a2, a3, (int)offset, RExC_precomp, RExC_precomp + offset); \
527 #define vWARN5(loc, m, a1, a2, a3, a4) STMT_START { \
528 const IV offset = loc - RExC_precomp; \
529 Perl_warner(aTHX_ packWARN(WARN_REGEXP), m REPORT_LOCATION, \
530 a1, a2, a3, a4, (int)offset, RExC_precomp, RExC_precomp + offset); \
534 /* Allow for side effects in s */
535 #define REGC(c,s) STMT_START { \
536 if (!SIZE_ONLY) *(s) = (c); else (void)(s); \
539 /* Macros for recording node offsets. 20001227 mjd@plover.com
540 * Nodes are numbered 1, 2, 3, 4. Node #n's position is recorded in
541 * element 2*n-1 of the array. Element #2n holds the byte length node #n.
542 * Element 0 holds the number n.
543 * Position is 1 indexed.
545 #ifndef RE_TRACK_PATTERN_OFFSETS
546 #define Set_Node_Offset_To_R(node,byte)
547 #define Set_Node_Offset(node,byte)
548 #define Set_Cur_Node_Offset
549 #define Set_Node_Length_To_R(node,len)
550 #define Set_Node_Length(node,len)
551 #define Set_Node_Cur_Length(node)
552 #define Node_Offset(n)
553 #define Node_Length(n)
554 #define Set_Node_Offset_Length(node,offset,len)
555 #define ProgLen(ri) ri->u.proglen
556 #define SetProgLen(ri,x) ri->u.proglen = x
558 #define ProgLen(ri) ri->u.offsets[0]
559 #define SetProgLen(ri,x) ri->u.offsets[0] = x
560 #define Set_Node_Offset_To_R(node,byte) STMT_START { \
562 MJD_OFFSET_DEBUG(("** (%d) offset of node %d is %d.\n", \
563 __LINE__, (int)(node), (int)(byte))); \
565 Perl_croak(aTHX_ "value of node is %d in Offset macro", (int)(node)); \
567 RExC_offsets[2*(node)-1] = (byte); \
572 #define Set_Node_Offset(node,byte) \
573 Set_Node_Offset_To_R((node)-RExC_emit_start, (byte)-RExC_start)
574 #define Set_Cur_Node_Offset Set_Node_Offset(RExC_emit, RExC_parse)
576 #define Set_Node_Length_To_R(node,len) STMT_START { \
578 MJD_OFFSET_DEBUG(("** (%d) size of node %d is %d.\n", \
579 __LINE__, (int)(node), (int)(len))); \
581 Perl_croak(aTHX_ "value of node is %d in Length macro", (int)(node)); \
583 RExC_offsets[2*(node)] = (len); \
588 #define Set_Node_Length(node,len) \
589 Set_Node_Length_To_R((node)-RExC_emit_start, len)
590 #define Set_Cur_Node_Length(len) Set_Node_Length(RExC_emit, len)
591 #define Set_Node_Cur_Length(node) \
592 Set_Node_Length(node, RExC_parse - parse_start)
594 /* Get offsets and lengths */
595 #define Node_Offset(n) (RExC_offsets[2*((n)-RExC_emit_start)-1])
596 #define Node_Length(n) (RExC_offsets[2*((n)-RExC_emit_start)])
598 #define Set_Node_Offset_Length(node,offset,len) STMT_START { \
599 Set_Node_Offset_To_R((node)-RExC_emit_start, (offset)); \
600 Set_Node_Length_To_R((node)-RExC_emit_start, (len)); \
604 #if PERL_ENABLE_EXPERIMENTAL_REGEX_OPTIMISATIONS
605 #define EXPERIMENTAL_INPLACESCAN
606 #endif /*RE_TRACK_PATTERN_OFFSETS*/
608 #define DEBUG_STUDYDATA(str,data,depth) \
609 DEBUG_OPTIMISE_MORE_r(if(data){ \
610 PerlIO_printf(Perl_debug_log, \
611 "%*s" str "Pos:%"IVdf"/%"IVdf \
612 " Flags: 0x%"UVXf" Whilem_c: %"IVdf" Lcp: %"IVdf" %s", \
613 (int)(depth)*2, "", \
614 (IV)((data)->pos_min), \
615 (IV)((data)->pos_delta), \
616 (UV)((data)->flags), \
617 (IV)((data)->whilem_c), \
618 (IV)((data)->last_closep ? *((data)->last_closep) : -1), \
619 is_inf ? "INF " : "" \
621 if ((data)->last_found) \
622 PerlIO_printf(Perl_debug_log, \
623 "Last:'%s' %"IVdf":%"IVdf"/%"IVdf" %sFixed:'%s' @ %"IVdf \
624 " %sFloat: '%s' @ %"IVdf"/%"IVdf"", \
625 SvPVX_const((data)->last_found), \
626 (IV)((data)->last_end), \
627 (IV)((data)->last_start_min), \
628 (IV)((data)->last_start_max), \
629 ((data)->longest && \
630 (data)->longest==&((data)->longest_fixed)) ? "*" : "", \
631 SvPVX_const((data)->longest_fixed), \
632 (IV)((data)->offset_fixed), \
633 ((data)->longest && \
634 (data)->longest==&((data)->longest_float)) ? "*" : "", \
635 SvPVX_const((data)->longest_float), \
636 (IV)((data)->offset_float_min), \
637 (IV)((data)->offset_float_max) \
639 PerlIO_printf(Perl_debug_log,"\n"); \
642 static void clear_re(pTHX_ void *r);
644 /* Mark that we cannot extend a found fixed substring at this point.
645 Update the longest found anchored substring and the longest found
646 floating substrings if needed. */
649 S_scan_commit(pTHX_ const RExC_state_t *pRExC_state, scan_data_t *data, I32 *minlenp, int is_inf)
651 const STRLEN l = CHR_SVLEN(data->last_found);
652 const STRLEN old_l = CHR_SVLEN(*data->longest);
653 GET_RE_DEBUG_FLAGS_DECL;
655 PERL_ARGS_ASSERT_SCAN_COMMIT;
657 if ((l >= old_l) && ((l > old_l) || (data->flags & SF_BEFORE_EOL))) {
658 SvSetMagicSV(*data->longest, data->last_found);
659 if (*data->longest == data->longest_fixed) {
660 data->offset_fixed = l ? data->last_start_min : data->pos_min;
661 if (data->flags & SF_BEFORE_EOL)
663 |= ((data->flags & SF_BEFORE_EOL) << SF_FIX_SHIFT_EOL);
665 data->flags &= ~SF_FIX_BEFORE_EOL;
666 data->minlen_fixed=minlenp;
667 data->lookbehind_fixed=0;
669 else { /* *data->longest == data->longest_float */
670 data->offset_float_min = l ? data->last_start_min : data->pos_min;
671 data->offset_float_max = (l
672 ? data->last_start_max
673 : data->pos_min + data->pos_delta);
674 if (is_inf || (U32)data->offset_float_max > (U32)I32_MAX)
675 data->offset_float_max = I32_MAX;
676 if (data->flags & SF_BEFORE_EOL)
678 |= ((data->flags & SF_BEFORE_EOL) << SF_FL_SHIFT_EOL);
680 data->flags &= ~SF_FL_BEFORE_EOL;
681 data->minlen_float=minlenp;
682 data->lookbehind_float=0;
685 SvCUR_set(data->last_found, 0);
687 SV * const sv = data->last_found;
688 if (SvUTF8(sv) && SvMAGICAL(sv)) {
689 MAGIC * const mg = mg_find(sv, PERL_MAGIC_utf8);
695 data->flags &= ~SF_BEFORE_EOL;
696 DEBUG_STUDYDATA("commit: ",data,0);
699 /* Can match anything (initialization) */
701 S_cl_anything(const RExC_state_t *pRExC_state, struct regnode_charclass_class *cl)
703 PERL_ARGS_ASSERT_CL_ANYTHING;
705 ANYOF_CLASS_ZERO(cl);
706 ANYOF_BITMAP_SETALL(cl);
707 cl->flags = ANYOF_EOS|ANYOF_UNICODE_ALL;
709 cl->flags |= ANYOF_LOCALE;
712 /* Can match anything (initialization) */
714 S_cl_is_anything(const struct regnode_charclass_class *cl)
718 PERL_ARGS_ASSERT_CL_IS_ANYTHING;
720 for (value = 0; value <= ANYOF_MAX; value += 2)
721 if (ANYOF_CLASS_TEST(cl, value) && ANYOF_CLASS_TEST(cl, value + 1))
723 if (!(cl->flags & ANYOF_UNICODE_ALL))
725 if (!ANYOF_BITMAP_TESTALLSET((const void*)cl))
730 /* Can match anything (initialization) */
732 S_cl_init(const RExC_state_t *pRExC_state, struct regnode_charclass_class *cl)
734 PERL_ARGS_ASSERT_CL_INIT;
736 Zero(cl, 1, struct regnode_charclass_class);
738 cl_anything(pRExC_state, cl);
742 S_cl_init_zero(const RExC_state_t *pRExC_state, struct regnode_charclass_class *cl)
744 PERL_ARGS_ASSERT_CL_INIT_ZERO;
746 Zero(cl, 1, struct regnode_charclass_class);
748 cl_anything(pRExC_state, cl);
750 cl->flags |= ANYOF_LOCALE;
753 /* 'And' a given class with another one. Can create false positives */
754 /* We assume that cl is not inverted */
756 S_cl_and(struct regnode_charclass_class *cl,
757 const struct regnode_charclass_class *and_with)
759 PERL_ARGS_ASSERT_CL_AND;
761 assert(and_with->type == ANYOF);
762 if (!(and_with->flags & ANYOF_CLASS)
763 && !(cl->flags & ANYOF_CLASS)
764 && (and_with->flags & ANYOF_LOCALE) == (cl->flags & ANYOF_LOCALE)
765 && !(and_with->flags & ANYOF_FOLD)
766 && !(cl->flags & ANYOF_FOLD)) {
769 if (and_with->flags & ANYOF_INVERT)
770 for (i = 0; i < ANYOF_BITMAP_SIZE; i++)
771 cl->bitmap[i] &= ~and_with->bitmap[i];
773 for (i = 0; i < ANYOF_BITMAP_SIZE; i++)
774 cl->bitmap[i] &= and_with->bitmap[i];
775 } /* XXXX: logic is complicated otherwise, leave it along for a moment. */
776 if (!(and_with->flags & ANYOF_EOS))
777 cl->flags &= ~ANYOF_EOS;
779 if (cl->flags & ANYOF_UNICODE_ALL && and_with->flags & ANYOF_UNICODE &&
780 !(and_with->flags & ANYOF_INVERT)) {
781 cl->flags &= ~ANYOF_UNICODE_ALL;
782 cl->flags |= ANYOF_UNICODE;
783 ARG_SET(cl, ARG(and_with));
785 if (!(and_with->flags & ANYOF_UNICODE_ALL) &&
786 !(and_with->flags & ANYOF_INVERT))
787 cl->flags &= ~ANYOF_UNICODE_ALL;
788 if (!(and_with->flags & (ANYOF_UNICODE|ANYOF_UNICODE_ALL)) &&
789 !(and_with->flags & ANYOF_INVERT))
790 cl->flags &= ~ANYOF_UNICODE;
793 /* 'OR' a given class with another one. Can create false positives */
794 /* We assume that cl is not inverted */
796 S_cl_or(const RExC_state_t *pRExC_state, struct regnode_charclass_class *cl, const struct regnode_charclass_class *or_with)
798 PERL_ARGS_ASSERT_CL_OR;
800 if (or_with->flags & ANYOF_INVERT) {
802 * (B1 | CL1) | (!B2 & !CL2) = (B1 | !B2 & !CL2) | (CL1 | (!B2 & !CL2))
803 * <= (B1 | !B2) | (CL1 | !CL2)
804 * which is wasteful if CL2 is small, but we ignore CL2:
805 * (B1 | CL1) | (!B2 & !CL2) <= (B1 | CL1) | !B2 = (B1 | !B2) | CL1
806 * XXXX Can we handle case-fold? Unclear:
807 * (OK1(i) | OK1(i')) | !(OK1(i) | OK1(i')) =
808 * (OK1(i) | OK1(i')) | (!OK1(i) & !OK1(i'))
810 if ( (or_with->flags & ANYOF_LOCALE) == (cl->flags & ANYOF_LOCALE)
811 && !(or_with->flags & ANYOF_FOLD)
812 && !(cl->flags & ANYOF_FOLD) ) {
815 for (i = 0; i < ANYOF_BITMAP_SIZE; i++)
816 cl->bitmap[i] |= ~or_with->bitmap[i];
817 } /* XXXX: logic is complicated otherwise */
819 cl_anything(pRExC_state, cl);
822 /* (B1 | CL1) | (B2 | CL2) = (B1 | B2) | (CL1 | CL2)) */
823 if ( (or_with->flags & ANYOF_LOCALE) == (cl->flags & ANYOF_LOCALE)
824 && (!(or_with->flags & ANYOF_FOLD)
825 || (cl->flags & ANYOF_FOLD)) ) {
828 /* OR char bitmap and class bitmap separately */
829 for (i = 0; i < ANYOF_BITMAP_SIZE; i++)
830 cl->bitmap[i] |= or_with->bitmap[i];
831 if (or_with->flags & ANYOF_CLASS) {
832 for (i = 0; i < ANYOF_CLASSBITMAP_SIZE; i++)
833 cl->classflags[i] |= or_with->classflags[i];
834 cl->flags |= ANYOF_CLASS;
837 else { /* XXXX: logic is complicated, leave it along for a moment. */
838 cl_anything(pRExC_state, cl);
841 if (or_with->flags & ANYOF_EOS)
842 cl->flags |= ANYOF_EOS;
844 if (cl->flags & ANYOF_UNICODE && or_with->flags & ANYOF_UNICODE &&
845 ARG(cl) != ARG(or_with)) {
846 cl->flags |= ANYOF_UNICODE_ALL;
847 cl->flags &= ~ANYOF_UNICODE;
849 if (or_with->flags & ANYOF_UNICODE_ALL) {
850 cl->flags |= ANYOF_UNICODE_ALL;
851 cl->flags &= ~ANYOF_UNICODE;
855 #define TRIE_LIST_ITEM(state,idx) (trie->states[state].trans.list)[ idx ]
856 #define TRIE_LIST_CUR(state) ( TRIE_LIST_ITEM( state, 0 ).forid )
857 #define TRIE_LIST_LEN(state) ( TRIE_LIST_ITEM( state, 0 ).newstate )
858 #define TRIE_LIST_USED(idx) ( trie->states[state].trans.list ? (TRIE_LIST_CUR( idx ) - 1) : 0 )
863 dump_trie(trie,widecharmap,revcharmap)
864 dump_trie_interim_list(trie,widecharmap,revcharmap,next_alloc)
865 dump_trie_interim_table(trie,widecharmap,revcharmap,next_alloc)
867 These routines dump out a trie in a somewhat readable format.
868 The _interim_ variants are used for debugging the interim
869 tables that are used to generate the final compressed
870 representation which is what dump_trie expects.
872 Part of the reason for their existance is to provide a form
873 of documentation as to how the different representations function.
878 Dumps the final compressed table form of the trie to Perl_debug_log.
879 Used for debugging make_trie().
883 S_dump_trie(pTHX_ const struct _reg_trie_data *trie, HV *widecharmap,
884 AV *revcharmap, U32 depth)
887 SV *sv=sv_newmortal();
888 int colwidth= widecharmap ? 6 : 4;
890 GET_RE_DEBUG_FLAGS_DECL;
892 PERL_ARGS_ASSERT_DUMP_TRIE;
894 PerlIO_printf( Perl_debug_log, "%*sChar : %-6s%-6s%-4s ",
895 (int)depth * 2 + 2,"",
896 "Match","Base","Ofs" );
898 for( state = 0 ; state < trie->uniquecharcount ; state++ ) {
899 SV ** const tmp = av_fetch( revcharmap, state, 0);
901 PerlIO_printf( Perl_debug_log, "%*s",
903 pv_pretty(sv, SvPV_nolen_const(*tmp), SvCUR(*tmp), colwidth,
904 PL_colors[0], PL_colors[1],
905 (SvUTF8(*tmp) ? PERL_PV_ESCAPE_UNI : 0) |
906 PERL_PV_ESCAPE_FIRSTCHAR
911 PerlIO_printf( Perl_debug_log, "\n%*sState|-----------------------",
912 (int)depth * 2 + 2,"");
914 for( state = 0 ; state < trie->uniquecharcount ; state++ )
915 PerlIO_printf( Perl_debug_log, "%.*s", colwidth, "--------");
916 PerlIO_printf( Perl_debug_log, "\n");
918 for( state = 1 ; state < trie->statecount ; state++ ) {
919 const U32 base = trie->states[ state ].trans.base;
921 PerlIO_printf( Perl_debug_log, "%*s#%4"UVXf"|", (int)depth * 2 + 2,"", (UV)state);
923 if ( trie->states[ state ].wordnum ) {
924 PerlIO_printf( Perl_debug_log, " W%4X", trie->states[ state ].wordnum );
926 PerlIO_printf( Perl_debug_log, "%6s", "" );
929 PerlIO_printf( Perl_debug_log, " @%4"UVXf" ", (UV)base );
934 while( ( base + ofs < trie->uniquecharcount ) ||
935 ( base + ofs - trie->uniquecharcount < trie->lasttrans
936 && trie->trans[ base + ofs - trie->uniquecharcount ].check != state))
939 PerlIO_printf( Perl_debug_log, "+%2"UVXf"[ ", (UV)ofs);
941 for ( ofs = 0 ; ofs < trie->uniquecharcount ; ofs++ ) {
942 if ( ( base + ofs >= trie->uniquecharcount ) &&
943 ( base + ofs - trie->uniquecharcount < trie->lasttrans ) &&
944 trie->trans[ base + ofs - trie->uniquecharcount ].check == state )
946 PerlIO_printf( Perl_debug_log, "%*"UVXf,
948 (UV)trie->trans[ base + ofs - trie->uniquecharcount ].next );
950 PerlIO_printf( Perl_debug_log, "%*s",colwidth," ." );
954 PerlIO_printf( Perl_debug_log, "]");
957 PerlIO_printf( Perl_debug_log, "\n" );
959 PerlIO_printf(Perl_debug_log, "%*sword_info N:(prev,len)=", (int)depth*2, "");
960 for (word=1; word <= trie->wordcount; word++) {
961 PerlIO_printf(Perl_debug_log, " %d:(%d,%d)",
962 (int)word, (int)(trie->wordinfo[word].prev),
963 (int)(trie->wordinfo[word].len));
965 PerlIO_printf(Perl_debug_log, "\n" );
968 Dumps a fully constructed but uncompressed trie in list form.
969 List tries normally only are used for construction when the number of
970 possible chars (trie->uniquecharcount) is very high.
971 Used for debugging make_trie().
974 S_dump_trie_interim_list(pTHX_ const struct _reg_trie_data *trie,
975 HV *widecharmap, AV *revcharmap, U32 next_alloc,
979 SV *sv=sv_newmortal();
980 int colwidth= widecharmap ? 6 : 4;
981 GET_RE_DEBUG_FLAGS_DECL;
983 PERL_ARGS_ASSERT_DUMP_TRIE_INTERIM_LIST;
985 /* print out the table precompression. */
986 PerlIO_printf( Perl_debug_log, "%*sState :Word | Transition Data\n%*s%s",
987 (int)depth * 2 + 2,"", (int)depth * 2 + 2,"",
988 "------:-----+-----------------\n" );
990 for( state=1 ; state < next_alloc ; state ++ ) {
993 PerlIO_printf( Perl_debug_log, "%*s %4"UVXf" :",
994 (int)depth * 2 + 2,"", (UV)state );
995 if ( ! trie->states[ state ].wordnum ) {
996 PerlIO_printf( Perl_debug_log, "%5s| ","");
998 PerlIO_printf( Perl_debug_log, "W%4x| ",
999 trie->states[ state ].wordnum
1002 for( charid = 1 ; charid <= TRIE_LIST_USED( state ) ; charid++ ) {
1003 SV ** const tmp = av_fetch( revcharmap, TRIE_LIST_ITEM(state,charid).forid, 0);
1005 PerlIO_printf( Perl_debug_log, "%*s:%3X=%4"UVXf" | ",
1007 pv_pretty(sv, SvPV_nolen_const(*tmp), SvCUR(*tmp), colwidth,
1008 PL_colors[0], PL_colors[1],
1009 (SvUTF8(*tmp) ? PERL_PV_ESCAPE_UNI : 0) |
1010 PERL_PV_ESCAPE_FIRSTCHAR
1012 TRIE_LIST_ITEM(state,charid).forid,
1013 (UV)TRIE_LIST_ITEM(state,charid).newstate
1016 PerlIO_printf(Perl_debug_log, "\n%*s| ",
1017 (int)((depth * 2) + 14), "");
1020 PerlIO_printf( Perl_debug_log, "\n");
1025 Dumps a fully constructed but uncompressed trie in table form.
1026 This is the normal DFA style state transition table, with a few
1027 twists to facilitate compression later.
1028 Used for debugging make_trie().
1031 S_dump_trie_interim_table(pTHX_ const struct _reg_trie_data *trie,
1032 HV *widecharmap, AV *revcharmap, U32 next_alloc,
1037 SV *sv=sv_newmortal();
1038 int colwidth= widecharmap ? 6 : 4;
1039 GET_RE_DEBUG_FLAGS_DECL;
1041 PERL_ARGS_ASSERT_DUMP_TRIE_INTERIM_TABLE;
1044 print out the table precompression so that we can do a visual check
1045 that they are identical.
1048 PerlIO_printf( Perl_debug_log, "%*sChar : ",(int)depth * 2 + 2,"" );
1050 for( charid = 0 ; charid < trie->uniquecharcount ; charid++ ) {
1051 SV ** const tmp = av_fetch( revcharmap, charid, 0);
1053 PerlIO_printf( Perl_debug_log, "%*s",
1055 pv_pretty(sv, SvPV_nolen_const(*tmp), SvCUR(*tmp), colwidth,
1056 PL_colors[0], PL_colors[1],
1057 (SvUTF8(*tmp) ? PERL_PV_ESCAPE_UNI : 0) |
1058 PERL_PV_ESCAPE_FIRSTCHAR
1064 PerlIO_printf( Perl_debug_log, "\n%*sState+-",(int)depth * 2 + 2,"" );
1066 for( charid=0 ; charid < trie->uniquecharcount ; charid++ ) {
1067 PerlIO_printf( Perl_debug_log, "%.*s", colwidth,"--------");
1070 PerlIO_printf( Perl_debug_log, "\n" );
1072 for( state=1 ; state < next_alloc ; state += trie->uniquecharcount ) {
1074 PerlIO_printf( Perl_debug_log, "%*s%4"UVXf" : ",
1075 (int)depth * 2 + 2,"",
1076 (UV)TRIE_NODENUM( state ) );
1078 for( charid = 0 ; charid < trie->uniquecharcount ; charid++ ) {
1079 UV v=(UV)SAFE_TRIE_NODENUM( trie->trans[ state + charid ].next );
1081 PerlIO_printf( Perl_debug_log, "%*"UVXf, colwidth, v );
1083 PerlIO_printf( Perl_debug_log, "%*s", colwidth, "." );
1085 if ( ! trie->states[ TRIE_NODENUM( state ) ].wordnum ) {
1086 PerlIO_printf( Perl_debug_log, " (%4"UVXf")\n", (UV)trie->trans[ state ].check );
1088 PerlIO_printf( Perl_debug_log, " (%4"UVXf") W%4X\n", (UV)trie->trans[ state ].check,
1089 trie->states[ TRIE_NODENUM( state ) ].wordnum );
1097 /* make_trie(startbranch,first,last,tail,word_count,flags,depth)
1098 startbranch: the first branch in the whole branch sequence
1099 first : start branch of sequence of branch-exact nodes.
1100 May be the same as startbranch
1101 last : Thing following the last branch.
1102 May be the same as tail.
1103 tail : item following the branch sequence
1104 count : words in the sequence
1105 flags : currently the OP() type we will be building one of /EXACT(|F|Fl)/
1106 depth : indent depth
1108 Inplace optimizes a sequence of 2 or more Branch-Exact nodes into a TRIE node.
1110 A trie is an N'ary tree where the branches are determined by digital
1111 decomposition of the key. IE, at the root node you look up the 1st character and
1112 follow that branch repeat until you find the end of the branches. Nodes can be
1113 marked as "accepting" meaning they represent a complete word. Eg:
1117 would convert into the following structure. Numbers represent states, letters
1118 following numbers represent valid transitions on the letter from that state, if
1119 the number is in square brackets it represents an accepting state, otherwise it
1120 will be in parenthesis.
1122 +-h->+-e->[3]-+-r->(8)-+-s->[9]
1126 (1) +-i->(6)-+-s->[7]
1128 +-s->(3)-+-h->(4)-+-e->[5]
1130 Accept Word Mapping: 3=>1 (he),5=>2 (she), 7=>3 (his), 9=>4 (hers)
1132 This shows that when matching against the string 'hers' we will begin at state 1
1133 read 'h' and move to state 2, read 'e' and move to state 3 which is accepting,
1134 then read 'r' and go to state 8 followed by 's' which takes us to state 9 which
1135 is also accepting. Thus we know that we can match both 'he' and 'hers' with a
1136 single traverse. We store a mapping from accepting to state to which word was
1137 matched, and then when we have multiple possibilities we try to complete the
1138 rest of the regex in the order in which they occured in the alternation.
1140 The only prior NFA like behaviour that would be changed by the TRIE support is
1141 the silent ignoring of duplicate alternations which are of the form:
1143 / (DUPE|DUPE) X? (?{ ... }) Y /x
1145 Thus EVAL blocks follwing a trie may be called a different number of times with
1146 and without the optimisation. With the optimisations dupes will be silently
1147 ignored. This inconsistant behaviour of EVAL type nodes is well established as
1148 the following demonstrates:
1150 'words'=~/(word|word|word)(?{ print $1 })[xyz]/
1152 which prints out 'word' three times, but
1154 'words'=~/(word|word|word)(?{ print $1 })S/
1156 which doesnt print it out at all. This is due to other optimisations kicking in.
1158 Example of what happens on a structural level:
1160 The regexp /(ac|ad|ab)+/ will produce the folowing debug output:
1162 1: CURLYM[1] {1,32767}(18)
1173 This would be optimizable with startbranch=5, first=5, last=16, tail=16
1174 and should turn into:
1176 1: CURLYM[1] {1,32767}(18)
1178 [Words:3 Chars Stored:6 Unique Chars:4 States:5 NCP:1]
1186 Cases where tail != last would be like /(?foo|bar)baz/:
1196 which would be optimizable with startbranch=1, first=1, last=7, tail=8
1197 and would end up looking like:
1200 [Words:2 Chars Stored:6 Unique Chars:5 States:7 NCP:1]
1207 d = uvuni_to_utf8_flags(d, uv, 0);
1209 is the recommended Unicode-aware way of saying
1214 #define TRIE_STORE_REVCHAR \
1217 SV *zlopp = newSV(2); \
1218 unsigned char *flrbbbbb = (unsigned char *) SvPVX(zlopp); \
1219 unsigned const char *const kapow = uvuni_to_utf8(flrbbbbb, uvc & 0xFF); \
1220 SvCUR_set(zlopp, kapow - flrbbbbb); \
1223 av_push(revcharmap, zlopp); \
1225 char ooooff = (char)uvc; \
1226 av_push(revcharmap, newSVpvn(&ooooff, 1)); \
1230 #define TRIE_READ_CHAR STMT_START { \
1234 if ( foldlen > 0 ) { \
1235 uvc = utf8n_to_uvuni( scan, UTF8_MAXLEN, &len, uniflags ); \
1240 uvc = utf8n_to_uvuni( (const U8*)uc, UTF8_MAXLEN, &len, uniflags);\
1241 uvc = to_uni_fold( uvc, foldbuf, &foldlen ); \
1242 foldlen -= UNISKIP( uvc ); \
1243 scan = foldbuf + UNISKIP( uvc ); \
1246 uvc = utf8n_to_uvuni( (const U8*)uc, UTF8_MAXLEN, &len, uniflags);\
1256 #define TRIE_LIST_PUSH(state,fid,ns) STMT_START { \
1257 if ( TRIE_LIST_CUR( state ) >=TRIE_LIST_LEN( state ) ) { \
1258 U32 ging = TRIE_LIST_LEN( state ) *= 2; \
1259 Renew( trie->states[ state ].trans.list, ging, reg_trie_trans_le ); \
1261 TRIE_LIST_ITEM( state, TRIE_LIST_CUR( state ) ).forid = fid; \
1262 TRIE_LIST_ITEM( state, TRIE_LIST_CUR( state ) ).newstate = ns; \
1263 TRIE_LIST_CUR( state )++; \
1266 #define TRIE_LIST_NEW(state) STMT_START { \
1267 Newxz( trie->states[ state ].trans.list, \
1268 4, reg_trie_trans_le ); \
1269 TRIE_LIST_CUR( state ) = 1; \
1270 TRIE_LIST_LEN( state ) = 4; \
1273 #define TRIE_HANDLE_WORD(state) STMT_START { \
1274 U16 dupe= trie->states[ state ].wordnum; \
1275 regnode * const noper_next = regnext( noper ); \
1278 /* store the word for dumping */ \
1280 if (OP(noper) != NOTHING) \
1281 tmp = newSVpvn_utf8(STRING(noper), STR_LEN(noper), UTF); \
1283 tmp = newSVpvn_utf8( "", 0, UTF ); \
1284 av_push( trie_words, tmp ); \
1288 trie->wordinfo[curword].prev = 0; \
1289 trie->wordinfo[curword].len = wordlen; \
1290 trie->wordinfo[curword].accept = state; \
1292 if ( noper_next < tail ) { \
1294 trie->jump = (U16 *) PerlMemShared_calloc( word_count + 1, sizeof(U16) ); \
1295 trie->jump[curword] = (U16)(noper_next - convert); \
1297 jumper = noper_next; \
1299 nextbranch= regnext(cur); \
1303 /* It's a dupe. Pre-insert into the wordinfo[].prev */\
1304 /* chain, so that when the bits of chain are later */\
1305 /* linked together, the dups appear in the chain */\
1306 trie->wordinfo[curword].prev = trie->wordinfo[dupe].prev; \
1307 trie->wordinfo[dupe].prev = curword; \
1309 /* we haven't inserted this word yet. */ \
1310 trie->states[ state ].wordnum = curword; \
1315 #define TRIE_TRANS_STATE(state,base,ucharcount,charid,special) \
1316 ( ( base + charid >= ucharcount \
1317 && base + charid < ubound \
1318 && state == trie->trans[ base - ucharcount + charid ].check \
1319 && trie->trans[ base - ucharcount + charid ].next ) \
1320 ? trie->trans[ base - ucharcount + charid ].next \
1321 : ( state==1 ? special : 0 ) \
1325 #define MADE_JUMP_TRIE 2
1326 #define MADE_EXACT_TRIE 4
1329 S_make_trie(pTHX_ RExC_state_t *pRExC_state, regnode *startbranch, regnode *first, regnode *last, regnode *tail, U32 word_count, U32 flags, U32 depth)
1332 /* first pass, loop through and scan words */
1333 reg_trie_data *trie;
1334 HV *widecharmap = NULL;
1335 AV *revcharmap = newAV();
1337 const U32 uniflags = UTF8_ALLOW_DEFAULT;
1342 regnode *jumper = NULL;
1343 regnode *nextbranch = NULL;
1344 regnode *convert = NULL;
1345 U32 *prev_states; /* temp array mapping each state to previous one */
1346 /* we just use folder as a flag in utf8 */
1347 const U8 * const folder = ( flags == EXACTF
1349 : ( flags == EXACTFL
1356 const U32 data_slot = add_data( pRExC_state, 4, "tuuu" );
1357 AV *trie_words = NULL;
1358 /* along with revcharmap, this only used during construction but both are
1359 * useful during debugging so we store them in the struct when debugging.
1362 const U32 data_slot = add_data( pRExC_state, 2, "tu" );
1363 STRLEN trie_charcount=0;
1365 SV *re_trie_maxbuff;
1366 GET_RE_DEBUG_FLAGS_DECL;
1368 PERL_ARGS_ASSERT_MAKE_TRIE;
1370 PERL_UNUSED_ARG(depth);
1373 trie = (reg_trie_data *) PerlMemShared_calloc( 1, sizeof(reg_trie_data) );
1375 trie->startstate = 1;
1376 trie->wordcount = word_count;
1377 RExC_rxi->data->data[ data_slot ] = (void*)trie;
1378 trie->charmap = (U16 *) PerlMemShared_calloc( 256, sizeof(U16) );
1379 if (!(UTF && folder))
1380 trie->bitmap = (char *) PerlMemShared_calloc( ANYOF_BITMAP_SIZE, 1 );
1381 trie->wordinfo = (reg_trie_wordinfo *) PerlMemShared_calloc(
1382 trie->wordcount+1, sizeof(reg_trie_wordinfo));
1385 trie_words = newAV();
1388 re_trie_maxbuff = get_sv(RE_TRIE_MAXBUF_NAME, 1);
1389 if (!SvIOK(re_trie_maxbuff)) {
1390 sv_setiv(re_trie_maxbuff, RE_TRIE_MAXBUF_INIT);
1393 PerlIO_printf( Perl_debug_log,
1394 "%*smake_trie start==%d, first==%d, last==%d, tail==%d depth=%d\n",
1395 (int)depth * 2 + 2, "",
1396 REG_NODE_NUM(startbranch),REG_NODE_NUM(first),
1397 REG_NODE_NUM(last), REG_NODE_NUM(tail),
1401 /* Find the node we are going to overwrite */
1402 if ( first == startbranch && OP( last ) != BRANCH ) {
1403 /* whole branch chain */
1406 /* branch sub-chain */
1407 convert = NEXTOPER( first );
1410 /* -- First loop and Setup --
1412 We first traverse the branches and scan each word to determine if it
1413 contains widechars, and how many unique chars there are, this is
1414 important as we have to build a table with at least as many columns as we
1417 We use an array of integers to represent the character codes 0..255
1418 (trie->charmap) and we use a an HV* to store Unicode characters. We use the
1419 native representation of the character value as the key and IV's for the
1422 *TODO* If we keep track of how many times each character is used we can
1423 remap the columns so that the table compression later on is more
1424 efficient in terms of memory by ensuring most common value is in the
1425 middle and the least common are on the outside. IMO this would be better
1426 than a most to least common mapping as theres a decent chance the most
1427 common letter will share a node with the least common, meaning the node
1428 will not be compressable. With a middle is most common approach the worst
1429 case is when we have the least common nodes twice.
1433 for ( cur = first ; cur < last ; cur = regnext( cur ) ) {
1434 regnode * const noper = NEXTOPER( cur );
1435 const U8 *uc = (U8*)STRING( noper );
1436 const U8 * const e = uc + STR_LEN( noper );
1438 U8 foldbuf[ UTF8_MAXBYTES_CASE + 1 ];
1439 const U8 *scan = (U8*)NULL;
1440 U32 wordlen = 0; /* required init */
1442 bool set_bit = trie->bitmap ? 1 : 0; /*store the first char in the bitmap?*/
1444 if (OP(noper) == NOTHING) {
1448 if ( set_bit ) /* bitmap only alloced when !(UTF&&Folding) */
1449 TRIE_BITMAP_SET(trie,*uc); /* store the raw first byte
1450 regardless of encoding */
1452 for ( ; uc < e ; uc += len ) {
1453 TRIE_CHARCOUNT(trie)++;
1457 if ( !trie->charmap[ uvc ] ) {
1458 trie->charmap[ uvc ]=( ++trie->uniquecharcount );
1460 trie->charmap[ folder[ uvc ] ] = trie->charmap[ uvc ];
1464 /* store the codepoint in the bitmap, and if its ascii
1465 also store its folded equivelent. */
1466 TRIE_BITMAP_SET(trie,uvc);
1468 /* store the folded codepoint */
1469 if ( folder ) TRIE_BITMAP_SET(trie,folder[ uvc ]);
1472 /* store first byte of utf8 representation of
1473 codepoints in the 127 < uvc < 256 range */
1474 if (127 < uvc && uvc < 192) {
1475 TRIE_BITMAP_SET(trie,194);
1476 } else if (191 < uvc ) {
1477 TRIE_BITMAP_SET(trie,195);
1478 /* && uvc < 256 -- we know uvc is < 256 already */
1481 set_bit = 0; /* We've done our bit :-) */
1486 widecharmap = newHV();
1488 svpp = hv_fetch( widecharmap, (char*)&uvc, sizeof( UV ), 1 );
1491 Perl_croak( aTHX_ "error creating/fetching widecharmap entry for 0x%"UVXf, uvc );
1493 if ( !SvTRUE( *svpp ) ) {
1494 sv_setiv( *svpp, ++trie->uniquecharcount );
1499 if( cur == first ) {
1502 } else if (chars < trie->minlen) {
1504 } else if (chars > trie->maxlen) {
1508 } /* end first pass */
1509 DEBUG_TRIE_COMPILE_r(
1510 PerlIO_printf( Perl_debug_log, "%*sTRIE(%s): W:%d C:%d Uq:%d Min:%d Max:%d\n",
1511 (int)depth * 2 + 2,"",
1512 ( widecharmap ? "UTF8" : "NATIVE" ), (int)word_count,
1513 (int)TRIE_CHARCOUNT(trie), trie->uniquecharcount,
1514 (int)trie->minlen, (int)trie->maxlen )
1518 We now know what we are dealing with in terms of unique chars and
1519 string sizes so we can calculate how much memory a naive
1520 representation using a flat table will take. If it's over a reasonable
1521 limit (as specified by ${^RE_TRIE_MAXBUF}) we use a more memory
1522 conservative but potentially much slower representation using an array
1525 At the end we convert both representations into the same compressed
1526 form that will be used in regexec.c for matching with. The latter
1527 is a form that cannot be used to construct with but has memory
1528 properties similar to the list form and access properties similar
1529 to the table form making it both suitable for fast searches and
1530 small enough that its feasable to store for the duration of a program.
1532 See the comment in the code where the compressed table is produced
1533 inplace from the flat tabe representation for an explanation of how
1534 the compression works.
1539 Newx(prev_states, TRIE_CHARCOUNT(trie) + 2, U32);
1542 if ( (IV)( ( TRIE_CHARCOUNT(trie) + 1 ) * trie->uniquecharcount + 1) > SvIV(re_trie_maxbuff) ) {
1544 Second Pass -- Array Of Lists Representation
1546 Each state will be represented by a list of charid:state records
1547 (reg_trie_trans_le) the first such element holds the CUR and LEN
1548 points of the allocated array. (See defines above).
1550 We build the initial structure using the lists, and then convert
1551 it into the compressed table form which allows faster lookups
1552 (but cant be modified once converted).
1555 STRLEN transcount = 1;
1557 DEBUG_TRIE_COMPILE_MORE_r( PerlIO_printf( Perl_debug_log,
1558 "%*sCompiling trie using list compiler\n",
1559 (int)depth * 2 + 2, ""));
1561 trie->states = (reg_trie_state *)
1562 PerlMemShared_calloc( TRIE_CHARCOUNT(trie) + 2,
1563 sizeof(reg_trie_state) );
1567 for ( cur = first ; cur < last ; cur = regnext( cur ) ) {
1569 regnode * const noper = NEXTOPER( cur );
1570 U8 *uc = (U8*)STRING( noper );
1571 const U8 * const e = uc + STR_LEN( noper );
1572 U32 state = 1; /* required init */
1573 U16 charid = 0; /* sanity init */
1574 U8 *scan = (U8*)NULL; /* sanity init */
1575 STRLEN foldlen = 0; /* required init */
1576 U32 wordlen = 0; /* required init */
1577 U8 foldbuf[ UTF8_MAXBYTES_CASE + 1 ];
1579 if (OP(noper) != NOTHING) {
1580 for ( ; uc < e ; uc += len ) {
1585 charid = trie->charmap[ uvc ];
1587 SV** const svpp = hv_fetch( widecharmap, (char*)&uvc, sizeof( UV ), 0);
1591 charid=(U16)SvIV( *svpp );
1594 /* charid is now 0 if we dont know the char read, or nonzero if we do */
1601 if ( !trie->states[ state ].trans.list ) {
1602 TRIE_LIST_NEW( state );
1604 for ( check = 1; check <= TRIE_LIST_USED( state ); check++ ) {
1605 if ( TRIE_LIST_ITEM( state, check ).forid == charid ) {
1606 newstate = TRIE_LIST_ITEM( state, check ).newstate;
1611 newstate = next_alloc++;
1612 prev_states[newstate] = state;
1613 TRIE_LIST_PUSH( state, charid, newstate );
1618 Perl_croak( aTHX_ "panic! In trie construction, no char mapping for %"IVdf, uvc );
1622 TRIE_HANDLE_WORD(state);
1624 } /* end second pass */
1626 /* next alloc is the NEXT state to be allocated */
1627 trie->statecount = next_alloc;
1628 trie->states = (reg_trie_state *)
1629 PerlMemShared_realloc( trie->states,
1631 * sizeof(reg_trie_state) );
1633 /* and now dump it out before we compress it */
1634 DEBUG_TRIE_COMPILE_MORE_r(dump_trie_interim_list(trie, widecharmap,
1635 revcharmap, next_alloc,
1639 trie->trans = (reg_trie_trans *)
1640 PerlMemShared_calloc( transcount, sizeof(reg_trie_trans) );
1647 for( state=1 ; state < next_alloc ; state ++ ) {
1651 DEBUG_TRIE_COMPILE_MORE_r(
1652 PerlIO_printf( Perl_debug_log, "tp: %d zp: %d ",tp,zp)
1656 if (trie->states[state].trans.list) {
1657 U16 minid=TRIE_LIST_ITEM( state, 1).forid;
1661 for( idx = 2 ; idx <= TRIE_LIST_USED( state ) ; idx++ ) {
1662 const U16 forid = TRIE_LIST_ITEM( state, idx).forid;
1663 if ( forid < minid ) {
1665 } else if ( forid > maxid ) {
1669 if ( transcount < tp + maxid - minid + 1) {
1671 trie->trans = (reg_trie_trans *)
1672 PerlMemShared_realloc( trie->trans,
1674 * sizeof(reg_trie_trans) );
1675 Zero( trie->trans + (transcount / 2), transcount / 2 , reg_trie_trans );
1677 base = trie->uniquecharcount + tp - minid;
1678 if ( maxid == minid ) {
1680 for ( ; zp < tp ; zp++ ) {
1681 if ( ! trie->trans[ zp ].next ) {
1682 base = trie->uniquecharcount + zp - minid;
1683 trie->trans[ zp ].next = TRIE_LIST_ITEM( state, 1).newstate;
1684 trie->trans[ zp ].check = state;
1690 trie->trans[ tp ].next = TRIE_LIST_ITEM( state, 1).newstate;
1691 trie->trans[ tp ].check = state;
1696 for ( idx=1; idx <= TRIE_LIST_USED( state ) ; idx++ ) {
1697 const U32 tid = base - trie->uniquecharcount + TRIE_LIST_ITEM( state, idx ).forid;
1698 trie->trans[ tid ].next = TRIE_LIST_ITEM( state, idx ).newstate;
1699 trie->trans[ tid ].check = state;
1701 tp += ( maxid - minid + 1 );
1703 Safefree(trie->states[ state ].trans.list);
1706 DEBUG_TRIE_COMPILE_MORE_r(
1707 PerlIO_printf( Perl_debug_log, " base: %d\n",base);
1710 trie->states[ state ].trans.base=base;
1712 trie->lasttrans = tp + 1;
1716 Second Pass -- Flat Table Representation.
1718 we dont use the 0 slot of either trans[] or states[] so we add 1 to each.
1719 We know that we will need Charcount+1 trans at most to store the data
1720 (one row per char at worst case) So we preallocate both structures
1721 assuming worst case.
1723 We then construct the trie using only the .next slots of the entry
1726 We use the .check field of the first entry of the node temporarily to
1727 make compression both faster and easier by keeping track of how many non
1728 zero fields are in the node.
1730 Since trans are numbered from 1 any 0 pointer in the table is a FAIL
1733 There are two terms at use here: state as a TRIE_NODEIDX() which is a
1734 number representing the first entry of the node, and state as a
1735 TRIE_NODENUM() which is the trans number. state 1 is TRIE_NODEIDX(1) and
1736 TRIE_NODENUM(1), state 2 is TRIE_NODEIDX(2) and TRIE_NODENUM(3) if there
1737 are 2 entrys per node. eg:
1745 The table is internally in the right hand, idx form. However as we also
1746 have to deal with the states array which is indexed by nodenum we have to
1747 use TRIE_NODENUM() to convert.
1750 DEBUG_TRIE_COMPILE_MORE_r( PerlIO_printf( Perl_debug_log,
1751 "%*sCompiling trie using table compiler\n",
1752 (int)depth * 2 + 2, ""));
1754 trie->trans = (reg_trie_trans *)
1755 PerlMemShared_calloc( ( TRIE_CHARCOUNT(trie) + 1 )
1756 * trie->uniquecharcount + 1,
1757 sizeof(reg_trie_trans) );
1758 trie->states = (reg_trie_state *)
1759 PerlMemShared_calloc( TRIE_CHARCOUNT(trie) + 2,
1760 sizeof(reg_trie_state) );
1761 next_alloc = trie->uniquecharcount + 1;
1764 for ( cur = first ; cur < last ; cur = regnext( cur ) ) {
1766 regnode * const noper = NEXTOPER( cur );
1767 const U8 *uc = (U8*)STRING( noper );
1768 const U8 * const e = uc + STR_LEN( noper );
1770 U32 state = 1; /* required init */
1772 U16 charid = 0; /* sanity init */
1773 U32 accept_state = 0; /* sanity init */
1774 U8 *scan = (U8*)NULL; /* sanity init */
1776 STRLEN foldlen = 0; /* required init */
1777 U32 wordlen = 0; /* required init */
1778 U8 foldbuf[ UTF8_MAXBYTES_CASE + 1 ];
1780 if ( OP(noper) != NOTHING ) {
1781 for ( ; uc < e ; uc += len ) {
1786 charid = trie->charmap[ uvc ];
1788 SV* const * const svpp = hv_fetch( widecharmap, (char*)&uvc, sizeof( UV ), 0);
1789 charid = svpp ? (U16)SvIV(*svpp) : 0;
1793 if ( !trie->trans[ state + charid ].next ) {
1794 trie->trans[ state + charid ].next = next_alloc;
1795 trie->trans[ state ].check++;
1796 prev_states[TRIE_NODENUM(next_alloc)]
1797 = TRIE_NODENUM(state);
1798 next_alloc += trie->uniquecharcount;
1800 state = trie->trans[ state + charid ].next;
1802 Perl_croak( aTHX_ "panic! In trie construction, no char mapping for %"IVdf, uvc );
1804 /* charid is now 0 if we dont know the char read, or nonzero if we do */
1807 accept_state = TRIE_NODENUM( state );
1808 TRIE_HANDLE_WORD(accept_state);
1810 } /* end second pass */
1812 /* and now dump it out before we compress it */
1813 DEBUG_TRIE_COMPILE_MORE_r(dump_trie_interim_table(trie, widecharmap,
1815 next_alloc, depth+1));
1819 * Inplace compress the table.*
1821 For sparse data sets the table constructed by the trie algorithm will
1822 be mostly 0/FAIL transitions or to put it another way mostly empty.
1823 (Note that leaf nodes will not contain any transitions.)
1825 This algorithm compresses the tables by eliminating most such
1826 transitions, at the cost of a modest bit of extra work during lookup:
1828 - Each states[] entry contains a .base field which indicates the
1829 index in the state[] array wheres its transition data is stored.
1831 - If .base is 0 there are no valid transitions from that node.
1833 - If .base is nonzero then charid is added to it to find an entry in
1836 -If trans[states[state].base+charid].check!=state then the
1837 transition is taken to be a 0/Fail transition. Thus if there are fail
1838 transitions at the front of the node then the .base offset will point
1839 somewhere inside the previous nodes data (or maybe even into a node
1840 even earlier), but the .check field determines if the transition is
1844 The following process inplace converts the table to the compressed
1845 table: We first do not compress the root node 1,and mark its all its
1846 .check pointers as 1 and set its .base pointer as 1 as well. This
1847 allows to do a DFA construction from the compressed table later, and
1848 ensures that any .base pointers we calculate later are greater than
1851 - We set 'pos' to indicate the first entry of the second node.
1853 - We then iterate over the columns of the node, finding the first and
1854 last used entry at l and m. We then copy l..m into pos..(pos+m-l),
1855 and set the .check pointers accordingly, and advance pos
1856 appropriately and repreat for the next node. Note that when we copy
1857 the next pointers we have to convert them from the original
1858 NODEIDX form to NODENUM form as the former is not valid post
1861 - If a node has no transitions used we mark its base as 0 and do not
1862 advance the pos pointer.
1864 - If a node only has one transition we use a second pointer into the
1865 structure to fill in allocated fail transitions from other states.
1866 This pointer is independent of the main pointer and scans forward
1867 looking for null transitions that are allocated to a state. When it
1868 finds one it writes the single transition into the "hole". If the
1869 pointer doesnt find one the single transition is appended as normal.
1871 - Once compressed we can Renew/realloc the structures to release the
1874 See "Table-Compression Methods" in sec 3.9 of the Red Dragon,
1875 specifically Fig 3.47 and the associated pseudocode.
1879 const U32 laststate = TRIE_NODENUM( next_alloc );
1882 trie->statecount = laststate;
1884 for ( state = 1 ; state < laststate ; state++ ) {
1886 const U32 stateidx = TRIE_NODEIDX( state );
1887 const U32 o_used = trie->trans[ stateidx ].check;
1888 U32 used = trie->trans[ stateidx ].check;
1889 trie->trans[ stateidx ].check = 0;
1891 for ( charid = 0 ; used && charid < trie->uniquecharcount ; charid++ ) {
1892 if ( flag || trie->trans[ stateidx + charid ].next ) {
1893 if ( trie->trans[ stateidx + charid ].next ) {
1895 for ( ; zp < pos ; zp++ ) {
1896 if ( ! trie->trans[ zp ].next ) {
1900 trie->states[ state ].trans.base = zp + trie->uniquecharcount - charid ;
1901 trie->trans[ zp ].next = SAFE_TRIE_NODENUM( trie->trans[ stateidx + charid ].next );
1902 trie->trans[ zp ].check = state;
1903 if ( ++zp > pos ) pos = zp;
1910 trie->states[ state ].trans.base = pos + trie->uniquecharcount - charid ;
1912 trie->trans[ pos ].next = SAFE_TRIE_NODENUM( trie->trans[ stateidx + charid ].next );
1913 trie->trans[ pos ].check = state;
1918 trie->lasttrans = pos + 1;
1919 trie->states = (reg_trie_state *)
1920 PerlMemShared_realloc( trie->states, laststate
1921 * sizeof(reg_trie_state) );
1922 DEBUG_TRIE_COMPILE_MORE_r(
1923 PerlIO_printf( Perl_debug_log,
1924 "%*sAlloc: %d Orig: %"IVdf" elements, Final:%"IVdf". Savings of %%%5.2f\n",
1925 (int)depth * 2 + 2,"",
1926 (int)( ( TRIE_CHARCOUNT(trie) + 1 ) * trie->uniquecharcount + 1 ),
1929 ( ( next_alloc - pos ) * 100 ) / (double)next_alloc );
1932 } /* end table compress */
1934 DEBUG_TRIE_COMPILE_MORE_r(
1935 PerlIO_printf(Perl_debug_log, "%*sStatecount:%"UVxf" Lasttrans:%"UVxf"\n",
1936 (int)depth * 2 + 2, "",
1937 (UV)trie->statecount,
1938 (UV)trie->lasttrans)
1940 /* resize the trans array to remove unused space */
1941 trie->trans = (reg_trie_trans *)
1942 PerlMemShared_realloc( trie->trans, trie->lasttrans
1943 * sizeof(reg_trie_trans) );
1945 { /* Modify the program and insert the new TRIE node*/
1946 U8 nodetype =(U8)(flags & 0xFF);
1950 regnode *optimize = NULL;
1951 #ifdef RE_TRACK_PATTERN_OFFSETS
1954 U32 mjd_nodelen = 0;
1955 #endif /* RE_TRACK_PATTERN_OFFSETS */
1956 #endif /* DEBUGGING */
1958 This means we convert either the first branch or the first Exact,
1959 depending on whether the thing following (in 'last') is a branch
1960 or not and whther first is the startbranch (ie is it a sub part of
1961 the alternation or is it the whole thing.)
1962 Assuming its a sub part we conver the EXACT otherwise we convert
1963 the whole branch sequence, including the first.
1965 /* Find the node we are going to overwrite */
1966 if ( first != startbranch || OP( last ) == BRANCH ) {
1967 /* branch sub-chain */
1968 NEXT_OFF( first ) = (U16)(last - first);
1969 #ifdef RE_TRACK_PATTERN_OFFSETS
1971 mjd_offset= Node_Offset((convert));
1972 mjd_nodelen= Node_Length((convert));
1975 /* whole branch chain */
1977 #ifdef RE_TRACK_PATTERN_OFFSETS
1980 const regnode *nop = NEXTOPER( convert );
1981 mjd_offset= Node_Offset((nop));
1982 mjd_nodelen= Node_Length((nop));
1986 PerlIO_printf(Perl_debug_log, "%*sMJD offset:%"UVuf" MJD length:%"UVuf"\n",
1987 (int)depth * 2 + 2, "",
1988 (UV)mjd_offset, (UV)mjd_nodelen)
1991 /* But first we check to see if there is a common prefix we can
1992 split out as an EXACT and put in front of the TRIE node. */
1993 trie->startstate= 1;
1994 if ( trie->bitmap && !widecharmap && !trie->jump ) {
1996 for ( state = 1 ; state < trie->statecount-1 ; state++ ) {
2000 const U32 base = trie->states[ state ].trans.base;
2002 if ( trie->states[state].wordnum )
2005 for ( ofs = 0 ; ofs < trie->uniquecharcount ; ofs++ ) {
2006 if ( ( base + ofs >= trie->uniquecharcount ) &&
2007 ( base + ofs - trie->uniquecharcount < trie->lasttrans ) &&
2008 trie->trans[ base + ofs - trie->uniquecharcount ].check == state )
2010 if ( ++count > 1 ) {
2011 SV **tmp = av_fetch( revcharmap, ofs, 0);
2012 const U8 *ch = (U8*)SvPV_nolen_const( *tmp );
2013 if ( state == 1 ) break;
2015 Zero(trie->bitmap, ANYOF_BITMAP_SIZE, char);
2017 PerlIO_printf(Perl_debug_log,
2018 "%*sNew Start State=%"UVuf" Class: [",
2019 (int)depth * 2 + 2, "",
2022 SV ** const tmp = av_fetch( revcharmap, idx, 0);
2023 const U8 * const ch = (U8*)SvPV_nolen_const( *tmp );
2025 TRIE_BITMAP_SET(trie,*ch);
2027 TRIE_BITMAP_SET(trie, folder[ *ch ]);
2029 PerlIO_printf(Perl_debug_log, "%s", (char*)ch)
2033 TRIE_BITMAP_SET(trie,*ch);
2035 TRIE_BITMAP_SET(trie,folder[ *ch ]);
2036 DEBUG_OPTIMISE_r(PerlIO_printf( Perl_debug_log,"%s", ch));
2042 SV **tmp = av_fetch( revcharmap, idx, 0);
2044 char *ch = SvPV( *tmp, len );
2046 SV *sv=sv_newmortal();
2047 PerlIO_printf( Perl_debug_log,
2048 "%*sPrefix State: %"UVuf" Idx:%"UVuf" Char='%s'\n",
2049 (int)depth * 2 + 2, "",
2051 pv_pretty(sv, SvPV_nolen_const(*tmp), SvCUR(*tmp), 6,
2052 PL_colors[0], PL_colors[1],
2053 (SvUTF8(*tmp) ? PERL_PV_ESCAPE_UNI : 0) |
2054 PERL_PV_ESCAPE_FIRSTCHAR
2059 OP( convert ) = nodetype;
2060 str=STRING(convert);
2063 STR_LEN(convert) += len;
2069 DEBUG_OPTIMISE_r(PerlIO_printf( Perl_debug_log,"]\n"));
2074 trie->prefixlen = (state-1);
2076 regnode *n = convert+NODE_SZ_STR(convert);
2077 NEXT_OFF(convert) = NODE_SZ_STR(convert);
2078 trie->startstate = state;
2079 trie->minlen -= (state - 1);
2080 trie->maxlen -= (state - 1);
2082 /* At least the UNICOS C compiler choked on this
2083 * being argument to DEBUG_r(), so let's just have
2086 #ifdef PERL_EXT_RE_BUILD
2092 regnode *fix = convert;
2093 U32 word = trie->wordcount;
2095 Set_Node_Offset_Length(convert, mjd_offset, state - 1);
2096 while( ++fix < n ) {
2097 Set_Node_Offset_Length(fix, 0, 0);
2100 SV ** const tmp = av_fetch( trie_words, word, 0 );
2102 if ( STR_LEN(convert) <= SvCUR(*tmp) )
2103 sv_chop(*tmp, SvPV_nolen(*tmp) + STR_LEN(convert));
2105 sv_chop(*tmp, SvPV_nolen(*tmp) + SvCUR(*tmp));
2113 NEXT_OFF(convert) = (U16)(tail - convert);
2114 DEBUG_r(optimize= n);
2120 if ( trie->maxlen ) {
2121 NEXT_OFF( convert ) = (U16)(tail - convert);
2122 ARG_SET( convert, data_slot );
2123 /* Store the offset to the first unabsorbed branch in
2124 jump[0], which is otherwise unused by the jump logic.
2125 We use this when dumping a trie and during optimisation. */
2127 trie->jump[0] = (U16)(nextbranch - convert);
2130 if ( !trie->states[trie->startstate].wordnum && trie->bitmap &&
2131 ( (char *)jumper - (char *)convert) >= (int)sizeof(struct regnode_charclass) )
2133 OP( convert ) = TRIEC;
2134 Copy(trie->bitmap, ((struct regnode_charclass *)convert)->bitmap, ANYOF_BITMAP_SIZE, char);
2135 PerlMemShared_free(trie->bitmap);
2138 OP( convert ) = TRIE;
2140 /* store the type in the flags */
2141 convert->flags = nodetype;
2145 + regarglen[ OP( convert ) ];
2147 /* XXX We really should free up the resource in trie now,
2148 as we won't use them - (which resources?) dmq */
2150 /* needed for dumping*/
2151 DEBUG_r(if (optimize) {
2152 regnode *opt = convert;
2154 while ( ++opt < optimize) {
2155 Set_Node_Offset_Length(opt,0,0);
2158 Try to clean up some of the debris left after the
2161 while( optimize < jumper ) {
2162 mjd_nodelen += Node_Length((optimize));
2163 OP( optimize ) = OPTIMIZED;
2164 Set_Node_Offset_Length(optimize,0,0);
2167 Set_Node_Offset_Length(convert,mjd_offset,mjd_nodelen);
2169 } /* end node insert */
2171 /* Finish populating the prev field of the wordinfo array. Walk back
2172 * from each accept state until we find another accept state, and if
2173 * so, point the first word's .prev field at the second word. If the
2174 * second already has a .prev field set, stop now. This will be the
2175 * case either if we've already processed that word's accept state,
2176 * or that that state had multiple words, and the overspill words
2177 * were already linked up earlier.
2184 for (word=1; word <= trie->wordcount; word++) {
2186 if (trie->wordinfo[word].prev)
2188 state = trie->wordinfo[word].accept;
2190 state = prev_states[state];
2193 prev = trie->states[state].wordnum;
2197 trie->wordinfo[word].prev = prev;
2199 Safefree(prev_states);
2203 /* and now dump out the compressed format */
2204 DEBUG_TRIE_COMPILE_r(dump_trie(trie, widecharmap, revcharmap, depth+1));
2206 RExC_rxi->data->data[ data_slot + 1 ] = (void*)widecharmap;
2208 RExC_rxi->data->data[ data_slot + TRIE_WORDS_OFFSET ] = (void*)trie_words;
2209 RExC_rxi->data->data[ data_slot + 3 ] = (void*)revcharmap;
2211 SvREFCNT_dec(revcharmap);
2215 : trie->startstate>1
2221 S_make_trie_failtable(pTHX_ RExC_state_t *pRExC_state, regnode *source, regnode *stclass, U32 depth)
2223 /* The Trie is constructed and compressed now so we can build a fail array now if its needed
2225 This is basically the Aho-Corasick algorithm. Its from exercise 3.31 and 3.32 in the
2226 "Red Dragon" -- Compilers, principles, techniques, and tools. Aho, Sethi, Ullman 1985/88
2229 We find the fail state for each state in the trie, this state is the longest proper
2230 suffix of the current states 'word' that is also a proper prefix of another word in our
2231 trie. State 1 represents the word '' and is the thus the default fail state. This allows
2232 the DFA not to have to restart after its tried and failed a word at a given point, it
2233 simply continues as though it had been matching the other word in the first place.
2235 'abcdgu'=~/abcdefg|cdgu/
2236 When we get to 'd' we are still matching the first word, we would encounter 'g' which would
2237 fail, which would bring use to the state representing 'd' in the second word where we would
2238 try 'g' and succeed, prodceding to match 'cdgu'.
2240 /* add a fail transition */
2241 const U32 trie_offset = ARG(source);
2242 reg_trie_data *trie=(reg_trie_data *)RExC_rxi->data->data[trie_offset];
2244 const U32 ucharcount = trie->uniquecharcount;
2245 const U32 numstates = trie->statecount;
2246 const U32 ubound = trie->lasttrans + ucharcount;
2250 U32 base = trie->states[ 1 ].trans.base;
2253 const U32 data_slot = add_data( pRExC_state, 1, "T" );
2254 GET_RE_DEBUG_FLAGS_DECL;
2256 PERL_ARGS_ASSERT_MAKE_TRIE_FAILTABLE;
2258 PERL_UNUSED_ARG(depth);
2262 ARG_SET( stclass, data_slot );
2263 aho = (reg_ac_data *) PerlMemShared_calloc( 1, sizeof(reg_ac_data) );
2264 RExC_rxi->data->data[ data_slot ] = (void*)aho;
2265 aho->trie=trie_offset;
2266 aho->states=(reg_trie_state *)PerlMemShared_malloc( numstates * sizeof(reg_trie_state) );
2267 Copy( trie->states, aho->states, numstates, reg_trie_state );
2268 Newxz( q, numstates, U32);
2269 aho->fail = (U32 *) PerlMemShared_calloc( numstates, sizeof(U32) );
2272 /* initialize fail[0..1] to be 1 so that we always have
2273 a valid final fail state */
2274 fail[ 0 ] = fail[ 1 ] = 1;
2276 for ( charid = 0; charid < ucharcount ; charid++ ) {
2277 const U32 newstate = TRIE_TRANS_STATE( 1, base, ucharcount, charid, 0 );
2279 q[ q_write ] = newstate;
2280 /* set to point at the root */
2281 fail[ q[ q_write++ ] ]=1;
2284 while ( q_read < q_write) {
2285 const U32 cur = q[ q_read++ % numstates ];
2286 base = trie->states[ cur ].trans.base;
2288 for ( charid = 0 ; charid < ucharcount ; charid++ ) {
2289 const U32 ch_state = TRIE_TRANS_STATE( cur, base, ucharcount, charid, 1 );
2291 U32 fail_state = cur;
2294 fail_state = fail[ fail_state ];
2295 fail_base = aho->states[ fail_state ].trans.base;
2296 } while ( !TRIE_TRANS_STATE( fail_state, fail_base, ucharcount, charid, 1 ) );
2298 fail_state = TRIE_TRANS_STATE( fail_state, fail_base, ucharcount, charid, 1 );
2299 fail[ ch_state ] = fail_state;
2300 if ( !aho->states[ ch_state ].wordnum && aho->states[ fail_state ].wordnum )
2302 aho->states[ ch_state ].wordnum = aho->states[ fail_state ].wordnum;
2304 q[ q_write++ % numstates] = ch_state;
2308 /* restore fail[0..1] to 0 so that we "fall out" of the AC loop
2309 when we fail in state 1, this allows us to use the
2310 charclass scan to find a valid start char. This is based on the principle
2311 that theres a good chance the string being searched contains lots of stuff
2312 that cant be a start char.
2314 fail[ 0 ] = fail[ 1 ] = 0;
2315 DEBUG_TRIE_COMPILE_r({
2316 PerlIO_printf(Perl_debug_log,
2317 "%*sStclass Failtable (%"UVuf" states): 0",
2318 (int)(depth * 2), "", (UV)numstates
2320 for( q_read=1; q_read<numstates; q_read++ ) {
2321 PerlIO_printf(Perl_debug_log, ", %"UVuf, (UV)fail[q_read]);
2323 PerlIO_printf(Perl_debug_log, "\n");
2326 /*RExC_seen |= REG_SEEN_TRIEDFA;*/
2331 * There are strange code-generation bugs caused on sparc64 by gcc-2.95.2.
2332 * These need to be revisited when a newer toolchain becomes available.
2334 #if defined(__sparc64__) && defined(__GNUC__)
2335 # if __GNUC__ < 2 || (__GNUC__ == 2 && __GNUC_MINOR__ < 96)
2336 # undef SPARC64_GCC_WORKAROUND
2337 # define SPARC64_GCC_WORKAROUND 1
2341 #define DEBUG_PEEP(str,scan,depth) \
2342 DEBUG_OPTIMISE_r({if (scan){ \
2343 SV * const mysv=sv_newmortal(); \
2344 regnode *Next = regnext(scan); \
2345 regprop(RExC_rx, mysv, scan); \
2346 PerlIO_printf(Perl_debug_log, "%*s" str ">%3d: %s (%d)\n", \
2347 (int)depth*2, "", REG_NODE_NUM(scan), SvPV_nolen_const(mysv),\
2348 Next ? (REG_NODE_NUM(Next)) : 0 ); \
2355 #define JOIN_EXACT(scan,min,flags) \
2356 if (PL_regkind[OP(scan)] == EXACT) \
2357 join_exact(pRExC_state,(scan),(min),(flags),NULL,depth+1)
2360 S_join_exact(pTHX_ RExC_state_t *pRExC_state, regnode *scan, I32 *min, U32 flags,regnode *val, U32 depth) {
2361 /* Merge several consecutive EXACTish nodes into one. */
2362 regnode *n = regnext(scan);
2364 regnode *next = scan + NODE_SZ_STR(scan);
2368 regnode *stop = scan;
2369 GET_RE_DEBUG_FLAGS_DECL;
2371 PERL_UNUSED_ARG(depth);
2374 PERL_ARGS_ASSERT_JOIN_EXACT;
2375 #ifndef EXPERIMENTAL_INPLACESCAN
2376 PERL_UNUSED_ARG(flags);
2377 PERL_UNUSED_ARG(val);
2379 DEBUG_PEEP("join",scan,depth);
2381 /* Skip NOTHING, merge EXACT*. */
2383 ( PL_regkind[OP(n)] == NOTHING ||
2384 (stringok && (OP(n) == OP(scan))))
2386 && NEXT_OFF(scan) + NEXT_OFF(n) < I16_MAX) {
2388 if (OP(n) == TAIL || n > next)
2390 if (PL_regkind[OP(n)] == NOTHING) {
2391 DEBUG_PEEP("skip:",n,depth);
2392 NEXT_OFF(scan) += NEXT_OFF(n);
2393 next = n + NODE_STEP_REGNODE;
2400 else if (stringok) {
2401 const unsigned int oldl = STR_LEN(scan);
2402 regnode * const nnext = regnext(n);
2404 DEBUG_PEEP("merg",n,depth);
2407 if (oldl + STR_LEN(n) > U8_MAX)
2409 NEXT_OFF(scan) += NEXT_OFF(n);
2410 STR_LEN(scan) += STR_LEN(n);
2411 next = n + NODE_SZ_STR(n);
2412 /* Now we can overwrite *n : */
2413 Move(STRING(n), STRING(scan) + oldl, STR_LEN(n), char);
2421 #ifdef EXPERIMENTAL_INPLACESCAN
2422 if (flags && !NEXT_OFF(n)) {
2423 DEBUG_PEEP("atch", val, depth);
2424 if (reg_off_by_arg[OP(n)]) {
2425 ARG_SET(n, val - n);
2428 NEXT_OFF(n) = val - n;
2435 if (UTF && ( OP(scan) == EXACTF ) && ( STR_LEN(scan) >= 6 ) ) {
2437 Two problematic code points in Unicode casefolding of EXACT nodes:
2439 U+0390 - GREEK SMALL LETTER IOTA WITH DIALYTIKA AND TONOS
2440 U+03B0 - GREEK SMALL LETTER UPSILON WITH DIALYTIKA AND TONOS
2446 U+03B9 U+0308 U+0301 0xCE 0xB9 0xCC 0x88 0xCC 0x81
2447 U+03C5 U+0308 U+0301 0xCF 0x85 0xCC 0x88 0xCC 0x81
2449 This means that in case-insensitive matching (or "loose matching",
2450 as Unicode calls it), an EXACTF of length six (the UTF-8 encoded byte
2451 length of the above casefolded versions) can match a target string
2452 of length two (the byte length of UTF-8 encoded U+0390 or U+03B0).
2453 This would rather mess up the minimum length computation.
2455 What we'll do is to look for the tail four bytes, and then peek
2456 at the preceding two bytes to see whether we need to decrease
2457 the minimum length by four (six minus two).
2459 Thanks to the design of UTF-8, there cannot be false matches:
2460 A sequence of valid UTF-8 bytes cannot be a subsequence of
2461 another valid sequence of UTF-8 bytes.
2464 char * const s0 = STRING(scan), *s, *t;
2465 char * const s1 = s0 + STR_LEN(scan) - 1;
2466 char * const s2 = s1 - 4;
2467 #ifdef EBCDIC /* RD tunifold greek 0390 and 03B0 */
2468 const char t0[] = "\xaf\x49\xaf\x42";
2470 const char t0[] = "\xcc\x88\xcc\x81";
2472 const char * const t1 = t0 + 3;
2475 s < s2 && (t = ninstr(s, s1, t0, t1));
2478 if (((U8)t[-1] == 0x68 && (U8)t[-2] == 0xB4) ||
2479 ((U8)t[-1] == 0x46 && (U8)t[-2] == 0xB5))
2481 if (((U8)t[-1] == 0xB9 && (U8)t[-2] == 0xCE) ||
2482 ((U8)t[-1] == 0x85 && (U8)t[-2] == 0xCF))
2490 n = scan + NODE_SZ_STR(scan);
2492 if (PL_regkind[OP(n)] != NOTHING || OP(n) == NOTHING) {
2499 DEBUG_OPTIMISE_r(if (merged){DEBUG_PEEP("finl",scan,depth)});
2503 /* REx optimizer. Converts nodes into quickier variants "in place".
2504 Finds fixed substrings. */
2506 /* Stops at toplevel WHILEM as well as at "last". At end *scanp is set
2507 to the position after last scanned or to NULL. */
2509 #define INIT_AND_WITHP \
2510 assert(!and_withp); \
2511 Newx(and_withp,1,struct regnode_charclass_class); \
2512 SAVEFREEPV(and_withp)
2514 /* this is a chain of data about sub patterns we are processing that
2515 need to be handled seperately/specially in study_chunk. Its so
2516 we can simulate recursion without losing state. */
2518 typedef struct scan_frame {
2519 regnode *last; /* last node to process in this frame */
2520 regnode *next; /* next node to process when last is reached */
2521 struct scan_frame *prev; /*previous frame*/
2522 I32 stop; /* what stopparen do we use */
2526 #define SCAN_COMMIT(s, data, m) scan_commit(s, data, m, is_inf)
2528 #define CASE_SYNST_FNC(nAmE) \
2530 if (flags & SCF_DO_STCLASS_AND) { \
2531 for (value = 0; value < 256; value++) \
2532 if (!is_ ## nAmE ## _cp(value)) \
2533 ANYOF_BITMAP_CLEAR(data->start_class, value); \
2536 for (value = 0; value < 256; value++) \
2537 if (is_ ## nAmE ## _cp(value)) \
2538 ANYOF_BITMAP_SET(data->start_class, value); \
2542 if (flags & SCF_DO_STCLASS_AND) { \
2543 for (value = 0; value < 256; value++) \
2544 if (is_ ## nAmE ## _cp(value)) \
2545 ANYOF_BITMAP_CLEAR(data->start_class, value); \
2548 for (value = 0; value < 256; value++) \
2549 if (!is_ ## nAmE ## _cp(value)) \
2550 ANYOF_BITMAP_SET(data->start_class, value); \
2557 S_study_chunk(pTHX_ RExC_state_t *pRExC_state, regnode **scanp,
2558 I32 *minlenp, I32 *deltap,
2563 struct regnode_charclass_class *and_withp,
2564 U32 flags, U32 depth)
2565 /* scanp: Start here (read-write). */
2566 /* deltap: Write maxlen-minlen here. */
2567 /* last: Stop before this one. */
2568 /* data: string data about the pattern */
2569 /* stopparen: treat close N as END */
2570 /* recursed: which subroutines have we recursed into */
2571 /* and_withp: Valid if flags & SCF_DO_STCLASS_OR */
2574 I32 min = 0, pars = 0, code;
2575 regnode *scan = *scanp, *next;
2577 int is_inf = (flags & SCF_DO_SUBSTR) && (data->flags & SF_IS_INF);
2578 int is_inf_internal = 0; /* The studied chunk is infinite */
2579 I32 is_par = OP(scan) == OPEN ? ARG(scan) : 0;
2580 scan_data_t data_fake;
2581 SV *re_trie_maxbuff = NULL;
2582 regnode *first_non_open = scan;
2583 I32 stopmin = I32_MAX;
2584 scan_frame *frame = NULL;
2585 GET_RE_DEBUG_FLAGS_DECL;
2587 PERL_ARGS_ASSERT_STUDY_CHUNK;
2590 StructCopy(&zero_scan_data, &data_fake, scan_data_t);
2594 while (first_non_open && OP(first_non_open) == OPEN)
2595 first_non_open=regnext(first_non_open);
2600 while ( scan && OP(scan) != END && scan < last ){
2601 /* Peephole optimizer: */
2602 DEBUG_STUDYDATA("Peep:", data,depth);
2603 DEBUG_PEEP("Peep",scan,depth);
2604 JOIN_EXACT(scan,&min,0);
2606 /* Follow the next-chain of the current node and optimize
2607 away all the NOTHINGs from it. */
2608 if (OP(scan) != CURLYX) {
2609 const int max = (reg_off_by_arg[OP(scan)]
2611 /* I32 may be smaller than U16 on CRAYs! */
2612 : (I32_MAX < U16_MAX ? I32_MAX : U16_MAX));
2613 int off = (reg_off_by_arg[OP(scan)] ? ARG(scan) : NEXT_OFF(scan));
2617 /* Skip NOTHING and LONGJMP. */
2618 while ((n = regnext(n))
2619 && ((PL_regkind[OP(n)] == NOTHING && (noff = NEXT_OFF(n)))
2620 || ((OP(n) == LONGJMP) && (noff = ARG(n))))
2621 && off + noff < max)
2623 if (reg_off_by_arg[OP(scan)])
2626 NEXT_OFF(scan) = off;
2631 /* The principal pseudo-switch. Cannot be a switch, since we
2632 look into several different things. */
2633 if (OP(scan) == BRANCH || OP(scan) == BRANCHJ
2634 || OP(scan) == IFTHEN) {
2635 next = regnext(scan);
2637 /* demq: the op(next)==code check is to see if we have "branch-branch" AFAICT */
2639 if (OP(next) == code || code == IFTHEN) {
2640 /* NOTE - There is similar code to this block below for handling
2641 TRIE nodes on a re-study. If you change stuff here check there
2643 I32 max1 = 0, min1 = I32_MAX, num = 0;
2644 struct regnode_charclass_class accum;
2645 regnode * const startbranch=scan;
2647 if (flags & SCF_DO_SUBSTR)
2648 SCAN_COMMIT(pRExC_state, data, minlenp); /* Cannot merge strings after this. */
2649 if (flags & SCF_DO_STCLASS)
2650 cl_init_zero(pRExC_state, &accum);
2652 while (OP(scan) == code) {
2653 I32 deltanext, minnext, f = 0, fake;
2654 struct regnode_charclass_class this_class;
2657 data_fake.flags = 0;
2659 data_fake.whilem_c = data->whilem_c;
2660 data_fake.last_closep = data->last_closep;
2663 data_fake.last_closep = &fake;
2665 data_fake.pos_delta = delta;
2666 next = regnext(scan);
2667 scan = NEXTOPER(scan);
2669 scan = NEXTOPER(scan);
2670 if (flags & SCF_DO_STCLASS) {
2671 cl_init(pRExC_state, &this_class);
2672 data_fake.start_class = &this_class;
2673 f = SCF_DO_STCLASS_AND;
2675 if (flags & SCF_WHILEM_VISITED_POS)
2676 f |= SCF_WHILEM_VISITED_POS;
2678 /* we suppose the run is continuous, last=next...*/
2679 minnext = study_chunk(pRExC_state, &scan, minlenp, &deltanext,
2681 stopparen, recursed, NULL, f,depth+1);
2684 if (max1 < minnext + deltanext)
2685 max1 = minnext + deltanext;
2686 if (deltanext == I32_MAX)
2687 is_inf = is_inf_internal = 1;
2689 if (data_fake.flags & (SF_HAS_PAR|SF_IN_PAR))
2691 if (data_fake.flags & SCF_SEEN_ACCEPT) {
2692 if ( stopmin > minnext)
2693 stopmin = min + min1;
2694 flags &= ~SCF_DO_SUBSTR;
2696 data->flags |= SCF_SEEN_ACCEPT;
2699 if (data_fake.flags & SF_HAS_EVAL)
2700 data->flags |= SF_HAS_EVAL;
2701 data->whilem_c = data_fake.whilem_c;
2703 if (flags & SCF_DO_STCLASS)
2704 cl_or(pRExC_state, &accum, &this_class);
2706 if (code == IFTHEN && num < 2) /* Empty ELSE branch */
2708 if (flags & SCF_DO_SUBSTR) {
2709 data->pos_min += min1;
2710 data->pos_delta += max1 - min1;
2711 if (max1 != min1 || is_inf)
2712 data->longest = &(data->longest_float);
2715 delta += max1 - min1;
2716 if (flags & SCF_DO_STCLASS_OR) {
2717 cl_or(pRExC_state, data->start_class, &accum);
2719 cl_and(data->start_class, and_withp);
2720 flags &= ~SCF_DO_STCLASS;
2723 else if (flags & SCF_DO_STCLASS_AND) {
2725 cl_and(data->start_class, &accum);
2726 flags &= ~SCF_DO_STCLASS;
2729 /* Switch to OR mode: cache the old value of
2730 * data->start_class */
2732 StructCopy(data->start_class, and_withp,
2733 struct regnode_charclass_class);
2734 flags &= ~SCF_DO_STCLASS_AND;
2735 StructCopy(&accum, data->start_class,
2736 struct regnode_charclass_class);
2737 flags |= SCF_DO_STCLASS_OR;
2738 data->start_class->flags |= ANYOF_EOS;
2742 if (PERL_ENABLE_TRIE_OPTIMISATION && OP( startbranch ) == BRANCH ) {
2745 Assuming this was/is a branch we are dealing with: 'scan' now
2746 points at the item that follows the branch sequence, whatever
2747 it is. We now start at the beginning of the sequence and look
2754 which would be constructed from a pattern like /A|LIST|OF|WORDS/
2756 If we can find such a subseqence we need to turn the first
2757 element into a trie and then add the subsequent branch exact
2758 strings to the trie.
2762 1. patterns where the whole set of branch can be converted.
2764 2. patterns where only a subset can be converted.
2766 In case 1 we can replace the whole set with a single regop
2767 for the trie. In case 2 we need to keep the start and end
2770 'BRANCH EXACT; BRANCH EXACT; BRANCH X'
2771 becomes BRANCH TRIE; BRANCH X;
2773 There is an additional case, that being where there is a
2774 common prefix, which gets split out into an EXACT like node
2775 preceding the TRIE node.
2777 If x(1..n)==tail then we can do a simple trie, if not we make
2778 a "jump" trie, such that when we match the appropriate word
2779 we "jump" to the appopriate tail node. Essentailly we turn
2780 a nested if into a case structure of sorts.
2785 if (!re_trie_maxbuff) {
2786 re_trie_maxbuff = get_sv(RE_TRIE_MAXBUF_NAME, 1);
2787 if (!SvIOK(re_trie_maxbuff))
2788 sv_setiv(re_trie_maxbuff, RE_TRIE_MAXBUF_INIT);
2790 if ( SvIV(re_trie_maxbuff)>=0 ) {
2792 regnode *first = (regnode *)NULL;
2793 regnode *last = (regnode *)NULL;
2794 regnode *tail = scan;
2799 SV * const mysv = sv_newmortal(); /* for dumping */
2801 /* var tail is used because there may be a TAIL
2802 regop in the way. Ie, the exacts will point to the
2803 thing following the TAIL, but the last branch will
2804 point at the TAIL. So we advance tail. If we
2805 have nested (?:) we may have to move through several
2809 while ( OP( tail ) == TAIL ) {
2810 /* this is the TAIL generated by (?:) */
2811 tail = regnext( tail );
2816 regprop(RExC_rx, mysv, tail );
2817 PerlIO_printf( Perl_debug_log, "%*s%s%s\n",
2818 (int)depth * 2 + 2, "",
2819 "Looking for TRIE'able sequences. Tail node is: ",
2820 SvPV_nolen_const( mysv )
2826 step through the branches, cur represents each
2827 branch, noper is the first thing to be matched
2828 as part of that branch and noper_next is the
2829 regnext() of that node. if noper is an EXACT
2830 and noper_next is the same as scan (our current
2831 position in the regex) then the EXACT branch is
2832 a possible optimization target. Once we have
2833 two or more consequetive such branches we can
2834 create a trie of the EXACT's contents and stich
2835 it in place. If the sequence represents all of
2836 the branches we eliminate the whole thing and
2837 replace it with a single TRIE. If it is a
2838 subsequence then we need to stitch it in. This
2839 means the first branch has to remain, and needs
2840 to be repointed at the item on the branch chain
2841 following the last branch optimized. This could
2842 be either a BRANCH, in which case the
2843 subsequence is internal, or it could be the
2844 item following the branch sequence in which
2845 case the subsequence is at the end.
2849 /* dont use tail as the end marker for this traverse */
2850 for ( cur = startbranch ; cur != scan ; cur = regnext( cur ) ) {
2851 regnode * const noper = NEXTOPER( cur );
2852 #if defined(DEBUGGING) || defined(NOJUMPTRIE)
2853 regnode * const noper_next = regnext( noper );
2857 regprop(RExC_rx, mysv, cur);
2858 PerlIO_printf( Perl_debug_log, "%*s- %s (%d)",
2859 (int)depth * 2 + 2,"", SvPV_nolen_const( mysv ), REG_NODE_NUM(cur) );
2861 regprop(RExC_rx, mysv, noper);
2862 PerlIO_printf( Perl_debug_log, " -> %s",
2863 SvPV_nolen_const(mysv));
2866 regprop(RExC_rx, mysv, noper_next );
2867 PerlIO_printf( Perl_debug_log,"\t=> %s\t",
2868 SvPV_nolen_const(mysv));
2870 PerlIO_printf( Perl_debug_log, "(First==%d,Last==%d,Cur==%d)\n",
2871 REG_NODE_NUM(first), REG_NODE_NUM(last), REG_NODE_NUM(cur) );
2873 if ( (((first && optype!=NOTHING) ? OP( noper ) == optype
2874 : PL_regkind[ OP( noper ) ] == EXACT )
2875 || OP(noper) == NOTHING )
2877 && noper_next == tail
2882 if ( !first || optype == NOTHING ) {
2883 if (!first) first = cur;
2884 optype = OP( noper );
2890 Currently we do not believe that the trie logic can
2891 handle case insensitive matching properly when the
2892 pattern is not unicode (thus forcing unicode semantics).
2894 If/when this is fixed the following define can be swapped
2895 in below to fully enable trie logic.
2897 #define TRIE_TYPE_IS_SAFE 1
2900 #define TRIE_TYPE_IS_SAFE (UTF || optype==EXACT)
2902 if ( last && TRIE_TYPE_IS_SAFE ) {
2903 make_trie( pRExC_state,
2904 startbranch, first, cur, tail, count,
2907 if ( PL_regkind[ OP( noper ) ] == EXACT
2909 && noper_next == tail
2914 optype = OP( noper );
2924 regprop(RExC_rx, mysv, cur);
2925 PerlIO_printf( Perl_debug_log,
2926 "%*s- %s (%d) <SCAN FINISHED>\n", (int)depth * 2 + 2,
2927 "", SvPV_nolen_const( mysv ),REG_NODE_NUM(cur));
2931 if ( last && TRIE_TYPE_IS_SAFE ) {
2932 made= make_trie( pRExC_state, startbranch, first, scan, tail, count, optype, depth+1 );
2933 #ifdef TRIE_STUDY_OPT
2934 if ( ((made == MADE_EXACT_TRIE &&
2935 startbranch == first)
2936 || ( first_non_open == first )) &&
2938 flags |= SCF_TRIE_RESTUDY;
2939 if ( startbranch == first
2942 RExC_seen &=~REG_TOP_LEVEL_BRANCHES;
2952 else if ( code == BRANCHJ ) { /* single branch is optimized. */
2953 scan = NEXTOPER(NEXTOPER(scan));
2954 } else /* single branch is optimized. */
2955 scan = NEXTOPER(scan);
2957 } else if (OP(scan) == SUSPEND || OP(scan) == GOSUB || OP(scan) == GOSTART) {
2958 scan_frame *newframe = NULL;
2963 if (OP(scan) != SUSPEND) {
2964 /* set the pointer */
2965 if (OP(scan) == GOSUB) {
2967 RExC_recurse[ARG2L(scan)] = scan;
2968 start = RExC_open_parens[paren-1];
2969 end = RExC_close_parens[paren-1];
2972 start = RExC_rxi->program + 1;
2976 Newxz(recursed, (((RExC_npar)>>3) +1), U8);
2977 SAVEFREEPV(recursed);
2979 if (!PAREN_TEST(recursed,paren+1)) {
2980 PAREN_SET(recursed,paren+1);
2981 Newx(newframe,1,scan_frame);
2983 if (flags & SCF_DO_SUBSTR) {
2984 SCAN_COMMIT(pRExC_state,data,minlenp);
2985 data->longest = &(data->longest_float);
2987 is_inf = is_inf_internal = 1;
2988 if (flags & SCF_DO_STCLASS_OR) /* Allow everything */
2989 cl_anything(pRExC_state, data->start_class);
2990 flags &= ~SCF_DO_STCLASS;
2993 Newx(newframe,1,scan_frame);
2996 end = regnext(scan);
3001 SAVEFREEPV(newframe);
3002 newframe->next = regnext(scan);
3003 newframe->last = last;
3004 newframe->stop = stopparen;
3005 newframe->prev = frame;
3015 else if (OP(scan) == EXACT) {
3016 I32 l = STR_LEN(scan);
3019 const U8 * const s = (U8*)STRING(scan);
3020 l = utf8_length(s, s + l);
3021 uc = utf8_to_uvchr(s, NULL);
3023 uc = *((U8*)STRING(scan));
3026 if (flags & SCF_DO_SUBSTR) { /* Update longest substr. */
3027 /* The code below prefers earlier match for fixed
3028 offset, later match for variable offset. */
3029 if (data->last_end == -1) { /* Update the start info. */
3030 data->last_start_min = data->pos_min;
3031 data->last_start_max = is_inf
3032 ? I32_MAX : data->pos_min + data->pos_delta;
3034 sv_catpvn(data->last_found, STRING(scan), STR_LEN(scan));
3036 SvUTF8_on(data->last_found);
3038 SV * const sv = data->last_found;
3039 MAGIC * const mg = SvUTF8(sv) && SvMAGICAL(sv) ?
3040 mg_find(sv, PERL_MAGIC_utf8) : NULL;
3041 if (mg && mg->mg_len >= 0)
3042 mg->mg_len += utf8_length((U8*)STRING(scan),
3043 (U8*)STRING(scan)+STR_LEN(scan));
3045 data->last_end = data->pos_min + l;
3046 data->pos_min += l; /* As in the first entry. */
3047 data->flags &= ~SF_BEFORE_EOL;
3049 if (flags & SCF_DO_STCLASS_AND) {
3050 /* Check whether it is compatible with what we know already! */
3054 (!(data->start_class->flags & (ANYOF_CLASS | ANYOF_LOCALE))
3055 && !ANYOF_BITMAP_TEST(data->start_class, uc)
3056 && (!(data->start_class->flags & ANYOF_FOLD)
3057 || !ANYOF_BITMAP_TEST(data->start_class, PL_fold[uc])))
3060 ANYOF_CLASS_ZERO(data->start_class);
3061 ANYOF_BITMAP_ZERO(data->start_class);
3063 ANYOF_BITMAP_SET(data->start_class, uc);
3064 data->start_class->flags &= ~ANYOF_EOS;
3066 data->start_class->flags &= ~ANYOF_UNICODE_ALL;
3068 else if (flags & SCF_DO_STCLASS_OR) {
3069 /* false positive possible if the class is case-folded */
3071 ANYOF_BITMAP_SET(data->start_class, uc);
3073 data->start_class->flags |= ANYOF_UNICODE_ALL;
3074 data->start_class->flags &= ~ANYOF_EOS;
3075 cl_and(data->start_class, and_withp);
3077 flags &= ~SCF_DO_STCLASS;
3079 else if (PL_regkind[OP(scan)] == EXACT) { /* But OP != EXACT! */
3080 I32 l = STR_LEN(scan);
3081 UV uc = *((U8*)STRING(scan));
3083 /* Search for fixed substrings supports EXACT only. */
3084 if (flags & SCF_DO_SUBSTR) {
3086 SCAN_COMMIT(pRExC_state, data, minlenp);
3089 const U8 * const s = (U8 *)STRING(scan);
3090 l = utf8_length(s, s + l);
3091 uc = utf8_to_uvchr(s, NULL);
3094 if (flags & SCF_DO_SUBSTR)
3096 if (flags & SCF_DO_STCLASS_AND) {
3097 /* Check whether it is compatible with what we know already! */
3101 (!(data->start_class->flags & (ANYOF_CLASS | ANYOF_LOCALE))
3102 && !ANYOF_BITMAP_TEST(data->start_class, uc)
3103 && !ANYOF_BITMAP_TEST(data->start_class, PL_fold[uc])))
3105 ANYOF_CLASS_ZERO(data->start_class);
3106 ANYOF_BITMAP_ZERO(data->start_class);
3108 ANYOF_BITMAP_SET(data->start_class, uc);
3109 data->start_class->flags &= ~ANYOF_EOS;
3110 data->start_class->flags |= ANYOF_FOLD;
3111 if (OP(scan) == EXACTFL)
3112 data->start_class->flags |= ANYOF_LOCALE;
3115 else if (flags & SCF_DO_STCLASS_OR) {
3116 if (data->start_class->flags & ANYOF_FOLD) {
3117 /* false positive possible if the class is case-folded.
3118 Assume that the locale settings are the same... */
3120 ANYOF_BITMAP_SET(data->start_class, uc);
3121 data->start_class->flags &= ~ANYOF_EOS;
3123 cl_and(data->start_class, and_withp);
3125 flags &= ~SCF_DO_STCLASS;
3127 else if (REGNODE_VARIES(OP(scan))) {
3128 I32 mincount, maxcount, minnext, deltanext, fl = 0;
3129 I32 f = flags, pos_before = 0;
3130 regnode * const oscan = scan;
3131 struct regnode_charclass_class this_class;
3132 struct regnode_charclass_class *oclass = NULL;
3133 I32 next_is_eval = 0;
3135 switch (PL_regkind[OP(scan)]) {
3136 case WHILEM: /* End of (?:...)* . */
3137 scan = NEXTOPER(scan);
3140 if (flags & (SCF_DO_SUBSTR | SCF_DO_STCLASS)) {
3141 next = NEXTOPER(scan);
3142 if (OP(next) == EXACT || (flags & SCF_DO_STCLASS)) {
3144 maxcount = REG_INFTY;
3145 next = regnext(scan);
3146 scan = NEXTOPER(scan);
3150 if (flags & SCF_DO_SUBSTR)
3155 if (flags & SCF_DO_STCLASS) {
3157 maxcount = REG_INFTY;
3158 next = regnext(scan);
3159 scan = NEXTOPER(scan);
3162 is_inf = is_inf_internal = 1;
3163 scan = regnext(scan);
3164 if (flags & SCF_DO_SUBSTR) {
3165 SCAN_COMMIT(pRExC_state, data, minlenp); /* Cannot extend fixed substrings */
3166 data->longest = &(data->longest_float);
3168 goto optimize_curly_tail;
3170 if (stopparen>0 && (OP(scan)==CURLYN || OP(scan)==CURLYM)
3171 && (scan->flags == stopparen))
3176 mincount = ARG1(scan);
3177 maxcount = ARG2(scan);
3179 next = regnext(scan);
3180 if (OP(scan) == CURLYX) {
3181 I32 lp = (data ? *(data->last_closep) : 0);
3182 scan->flags = ((lp <= (I32)U8_MAX) ? (U8)lp : U8_MAX);
3184 scan = NEXTOPER(scan) + EXTRA_STEP_2ARGS;
3185 next_is_eval = (OP(scan) == EVAL);
3187 if (flags & SCF_DO_SUBSTR) {
3188 if (mincount == 0) SCAN_COMMIT(pRExC_state,data,minlenp); /* Cannot extend fixed substrings */
3189 pos_before = data->pos_min;
3193 data->flags &= ~(SF_HAS_PAR|SF_IN_PAR|SF_HAS_EVAL);
3195 data->flags |= SF_IS_INF;
3197 if (flags & SCF_DO_STCLASS) {
3198 cl_init(pRExC_state, &this_class);
3199 oclass = data->start_class;
3200 data->start_class = &this_class;
3201 f |= SCF_DO_STCLASS_AND;
3202 f &= ~SCF_DO_STCLASS_OR;
3204 /* These are the cases when once a subexpression
3205 fails at a particular position, it cannot succeed
3206 even after backtracking at the enclosing scope.
3208 XXXX what if minimal match and we are at the
3209 initial run of {n,m}? */
3210 if ((mincount != maxcount - 1) && (maxcount != REG_INFTY))
3211 f &= ~SCF_WHILEM_VISITED_POS;
3213 /* This will finish on WHILEM, setting scan, or on NULL: */
3214 minnext = study_chunk(pRExC_state, &scan, minlenp, &deltanext,
3215 last, data, stopparen, recursed, NULL,
3217 ? (f & ~SCF_DO_SUBSTR) : f),depth+1);
3219 if (flags & SCF_DO_STCLASS)
3220 data->start_class = oclass;
3221 if (mincount == 0 || minnext == 0) {
3222 if (flags & SCF_DO_STCLASS_OR) {
3223 cl_or(pRExC_state, data->start_class, &this_class);
3225 else if (flags & SCF_DO_STCLASS_AND) {
3226 /* Switch to OR mode: cache the old value of
3227 * data->start_class */
3229 StructCopy(data->start_class, and_withp,
3230 struct regnode_charclass_class);
3231 flags &= ~SCF_DO_STCLASS_AND;
3232 StructCopy(&this_class, data->start_class,
3233 struct regnode_charclass_class);
3234 flags |= SCF_DO_STCLASS_OR;
3235 data->start_class->flags |= ANYOF_EOS;
3237 } else { /* Non-zero len */
3238 if (flags & SCF_DO_STCLASS_OR) {
3239 cl_or(pRExC_state, data->start_class, &this_class);
3240 cl_and(data->start_class, and_withp);
3242 else if (flags & SCF_DO_STCLASS_AND)
3243 cl_and(data->start_class, &this_class);
3244 flags &= ~SCF_DO_STCLASS;
3246 if (!scan) /* It was not CURLYX, but CURLY. */
3248 if ( /* ? quantifier ok, except for (?{ ... }) */
3249 (next_is_eval || !(mincount == 0 && maxcount == 1))
3250 && (minnext == 0) && (deltanext == 0)
3251 && data && !(data->flags & (SF_HAS_PAR|SF_IN_PAR))
3252 && maxcount <= REG_INFTY/3) /* Complement check for big count */
3254 ckWARNreg(RExC_parse,
3255 "Quantifier unexpected on zero-length expression");
3258 min += minnext * mincount;
3259 is_inf_internal |= ((maxcount == REG_INFTY
3260 && (minnext + deltanext) > 0)
3261 || deltanext == I32_MAX);
3262 is_inf |= is_inf_internal;
3263 delta += (minnext + deltanext) * maxcount - minnext * mincount;
3265 /* Try powerful optimization CURLYX => CURLYN. */
3266 if ( OP(oscan) == CURLYX && data
3267 && data->flags & SF_IN_PAR
3268 && !(data->flags & SF_HAS_EVAL)
3269 && !deltanext && minnext == 1 ) {
3270 /* Try to optimize to CURLYN. */
3271 regnode *nxt = NEXTOPER(oscan) + EXTRA_STEP_2ARGS;
3272 regnode * const nxt1 = nxt;
3279 if (!REGNODE_SIMPLE(OP(nxt))
3280 && !(PL_regkind[OP(nxt)] == EXACT
3281 && STR_LEN(nxt) == 1))
3287 if (OP(nxt) != CLOSE)
3289 if (RExC_open_parens) {
3290 RExC_open_parens[ARG(nxt1)-1]=oscan; /*open->CURLYM*/
3291 RExC_close_parens[ARG(nxt1)-1]=nxt+2; /*close->while*/
3293 /* Now we know that nxt2 is the only contents: */
3294 oscan->flags = (U8)ARG(nxt);
3296 OP(nxt1) = NOTHING; /* was OPEN. */
3299 OP(nxt1 + 1) = OPTIMIZED; /* was count. */
3300 NEXT_OFF(nxt1+ 1) = 0; /* just for consistency. */
3301 NEXT_OFF(nxt2) = 0; /* just for consistency with CURLY. */
3302 OP(nxt) = OPTIMIZED; /* was CLOSE. */
3303 OP(nxt + 1) = OPTIMIZED; /* was count. */
3304 NEXT_OFF(nxt+ 1) = 0; /* just for consistency. */
3309 /* Try optimization CURLYX => CURLYM. */
3310 if ( OP(oscan) == CURLYX && data
3311 && !(data->flags & SF_HAS_PAR)
3312 && !(data->flags & SF_HAS_EVAL)
3313 && !deltanext /* atom is fixed width */
3314 && minnext != 0 /* CURLYM can't handle zero width */
3316 /* XXXX How to optimize if data == 0? */
3317 /* Optimize to a simpler form. */
3318 regnode *nxt = NEXTOPER(oscan) + EXTRA_STEP_2ARGS; /* OPEN */
3322 while ( (nxt2 = regnext(nxt)) /* skip over embedded stuff*/
3323 && (OP(nxt2) != WHILEM))
3325 OP(nxt2) = SUCCEED; /* Whas WHILEM */
3326 /* Need to optimize away parenths. */
3327 if ((data->flags & SF_IN_PAR) && OP(nxt) == CLOSE) {
3328 /* Set the parenth number. */
3329 regnode *nxt1 = NEXTOPER(oscan) + EXTRA_STEP_2ARGS; /* OPEN*/
3331 oscan->flags = (U8)ARG(nxt);
3332 if (RExC_open_parens) {
3333 RExC_open_parens[ARG(nxt1)-1]=oscan; /*open->CURLYM*/
3334 RExC_close_parens[ARG(nxt1)-1]=nxt2+1; /*close->NOTHING*/
3336 OP(nxt1) = OPTIMIZED; /* was OPEN. */
3337 OP(nxt) = OPTIMIZED; /* was CLOSE. */
3340 OP(nxt1 + 1) = OPTIMIZED; /* was count. */
3341 OP(nxt + 1) = OPTIMIZED; /* was count. */
3342 NEXT_OFF(nxt1 + 1) = 0; /* just for consistancy. */
3343 NEXT_OFF(nxt + 1) = 0; /* just for consistancy. */
3346 while ( nxt1 && (OP(nxt1) != WHILEM)) {
3347 regnode *nnxt = regnext(nxt1);
3350 if (reg_off_by_arg[OP(nxt1)])
3351 ARG_SET(nxt1, nxt2 - nxt1);
3352 else if (nxt2 - nxt1 < U16_MAX)
3353 NEXT_OFF(nxt1) = nxt2 - nxt1;
3355 OP(nxt) = NOTHING; /* Cannot beautify */
3360 /* Optimize again: */
3361 study_chunk(pRExC_state, &nxt1, minlenp, &deltanext, nxt,
3362 NULL, stopparen, recursed, NULL, 0,depth+1);
3367 else if ((OP(oscan) == CURLYX)
3368 && (flags & SCF_WHILEM_VISITED_POS)
3369 /* See the comment on a similar expression above.
3370 However, this time it not a subexpression
3371 we care about, but the expression itself. */
3372 && (maxcount == REG_INFTY)
3373 && data && ++data->whilem_c < 16) {
3374 /* This stays as CURLYX, we can put the count/of pair. */
3375 /* Find WHILEM (as in regexec.c) */
3376 regnode *nxt = oscan + NEXT_OFF(oscan);
3378 if (OP(PREVOPER(nxt)) == NOTHING) /* LONGJMP */
3380 PREVOPER(nxt)->flags = (U8)(data->whilem_c
3381 | (RExC_whilem_seen << 4)); /* On WHILEM */
3383 if (data && fl & (SF_HAS_PAR|SF_IN_PAR))
3385 if (flags & SCF_DO_SUBSTR) {
3386 SV *last_str = NULL;
3387 int counted = mincount != 0;
3389 if (data->last_end > 0 && mincount != 0) { /* Ends with a string. */
3390 #if defined(SPARC64_GCC_WORKAROUND)
3393 const char *s = NULL;
3396 if (pos_before >= data->last_start_min)
3399 b = data->last_start_min;
3402 s = SvPV_const(data->last_found, l);
3403 old = b - data->last_start_min;
3406 I32 b = pos_before >= data->last_start_min
3407 ? pos_before : data->last_start_min;
3409 const char * const s = SvPV_const(data->last_found, l);
3410 I32 old = b - data->last_start_min;
3414 old = utf8_hop((U8*)s, old) - (U8*)s;
3417 /* Get the added string: */
3418 last_str = newSVpvn_utf8(s + old, l, UTF);
3419 if (deltanext == 0 && pos_before == b) {
3420 /* What was added is a constant string */
3422 SvGROW(last_str, (mincount * l) + 1);
3423 repeatcpy(SvPVX(last_str) + l,
3424 SvPVX_const(last_str), l, mincount - 1);
3425 SvCUR_set(last_str, SvCUR(last_str) * mincount);
3426 /* Add additional parts. */
3427 SvCUR_set(data->last_found,
3428 SvCUR(data->last_found) - l);
3429 sv_catsv(data->last_found, last_str);
3431 SV * sv = data->last_found;
3433 SvUTF8(sv) && SvMAGICAL(sv) ?
3434 mg_find(sv, PERL_MAGIC_utf8) : NULL;
3435 if (mg && mg->mg_len >= 0)
3436 mg->mg_len += CHR_SVLEN(last_str) - l;
3438 data->last_end += l * (mincount - 1);
3441 /* start offset must point into the last copy */
3442 data->last_start_min += minnext * (mincount - 1);
3443 data->last_start_max += is_inf ? I32_MAX
3444 : (maxcount - 1) * (minnext + data->pos_delta);
3447 /* It is counted once already... */
3448 data->pos_min += minnext * (mincount - counted);
3449 data->pos_delta += - counted * deltanext +
3450 (minnext + deltanext) * maxcount - minnext * mincount;
3451 if (mincount != maxcount) {
3452 /* Cannot extend fixed substrings found inside
3454 SCAN_COMMIT(pRExC_state,data,minlenp);
3455 if (mincount && last_str) {
3456 SV * const sv = data->last_found;
3457 MAGIC * const mg = SvUTF8(sv) && SvMAGICAL(sv) ?
3458 mg_find(sv, PERL_MAGIC_utf8) : NULL;
3462 sv_setsv(sv, last_str);
3463 data->last_end = data->pos_min;
3464 data->last_start_min =
3465 data->pos_min - CHR_SVLEN(last_str);
3466 data->last_start_max = is_inf
3468 : data->pos_min + data->pos_delta
3469 - CHR_SVLEN(last_str);
3471 data->longest = &(data->longest_float);
3473 SvREFCNT_dec(last_str);
3475 if (data && (fl & SF_HAS_EVAL))
3476 data->flags |= SF_HAS_EVAL;
3477 optimize_curly_tail:
3478 if (OP(oscan) != CURLYX) {
3479 while (PL_regkind[OP(next = regnext(oscan))] == NOTHING
3481 NEXT_OFF(oscan) += NEXT_OFF(next);
3484 default: /* REF and CLUMP only? */
3485 if (flags & SCF_DO_SUBSTR) {
3486 SCAN_COMMIT(pRExC_state,data,minlenp); /* Cannot expect anything... */
3487 data->longest = &(data->longest_float);
3489 is_inf = is_inf_internal = 1;
3490 if (flags & SCF_DO_STCLASS_OR)
3491 cl_anything(pRExC_state, data->start_class);
3492 flags &= ~SCF_DO_STCLASS;
3496 else if (OP(scan) == LNBREAK) {
3497 if (flags & SCF_DO_STCLASS) {
3499 data->start_class->flags &= ~ANYOF_EOS; /* No match on empty */
3500 if (flags & SCF_DO_STCLASS_AND) {
3501 for (value = 0; value < 256; value++)
3502 if (!is_VERTWS_cp(value))
3503 ANYOF_BITMAP_CLEAR(data->start_class, value);
3506 for (value = 0; value < 256; value++)
3507 if (is_VERTWS_cp(value))
3508 ANYOF_BITMAP_SET(data->start_class, value);
3510 if (flags & SCF_DO_STCLASS_OR)
3511 cl_and(data->start_class, and_withp);
3512 flags &= ~SCF_DO_STCLASS;
3516 if (flags & SCF_DO_SUBSTR) {
3517 SCAN_COMMIT(pRExC_state,data,minlenp); /* Cannot expect anything... */
3519 data->pos_delta += 1;
3520 data->longest = &(data->longest_float);
3524 else if (OP(scan) == FOLDCHAR) {
3525 int d = ARG(scan)==0xDF ? 1 : 2;
3526 flags &= ~SCF_DO_STCLASS;
3529 if (flags & SCF_DO_SUBSTR) {
3530 SCAN_COMMIT(pRExC_state,data,minlenp); /* Cannot expect anything... */
3532 data->pos_delta += d;
3533 data->longest = &(data->longest_float);
3536 else if (REGNODE_SIMPLE(OP(scan))) {
3539 if (flags & SCF_DO_SUBSTR) {
3540 SCAN_COMMIT(pRExC_state,data,minlenp);
3544 if (flags & SCF_DO_STCLASS) {
3545 data->start_class->flags &= ~ANYOF_EOS; /* No match on empty */
3547 /* Some of the logic below assumes that switching
3548 locale on will only add false positives. */
3549 switch (PL_regkind[OP(scan)]) {
3553 /* Perl_croak(aTHX_ "panic: unexpected simple REx opcode %d", OP(scan)); */
3554 if (flags & SCF_DO_STCLASS_OR) /* Allow everything */
3555 cl_anything(pRExC_state, data->start_class);
3558 if (OP(scan) == SANY)
3560 if (flags & SCF_DO_STCLASS_OR) { /* Everything but \n */
3561 value = (ANYOF_BITMAP_TEST(data->start_class,'\n')
3562 || (data->start_class->flags & ANYOF_CLASS));
3563 cl_anything(pRExC_state, data->start_class);
3565 if (flags & SCF_DO_STCLASS_AND || !value)
3566 ANYOF_BITMAP_CLEAR(data->start_class,'\n');
3569 if (flags & SCF_DO_STCLASS_AND)
3570 cl_and(data->start_class,
3571 (struct regnode_charclass_class*)scan);
3573 cl_or(pRExC_state, data->start_class,
3574 (struct regnode_charclass_class*)scan);
3577 if (flags & SCF_DO_STCLASS_AND) {
3578 if (!(data->start_class->flags & ANYOF_LOCALE)) {
3579 ANYOF_CLASS_CLEAR(data->start_class,ANYOF_NALNUM);
3580 for (value = 0; value < 256; value++)
3581 if (!isALNUM(value))
3582 ANYOF_BITMAP_CLEAR(data->start_class, value);
3586 if (data->start_class->flags & ANYOF_LOCALE)
3587 ANYOF_CLASS_SET(data->start_class,ANYOF_ALNUM);
3589 for (value = 0; value < 256; value++)
3591 ANYOF_BITMAP_SET(data->start_class, value);
3596 if (flags & SCF_DO_STCLASS_AND) {
3597 if (data->start_class->flags & ANYOF_LOCALE)
3598 ANYOF_CLASS_CLEAR(data->start_class,ANYOF_NALNUM);
3601 ANYOF_CLASS_SET(data->start_class,ANYOF_ALNUM);
3602 data->start_class->flags |= ANYOF_LOCALE;
3606 if (flags & SCF_DO_STCLASS_AND) {
3607 if (!(data->start_class->flags & ANYOF_LOCALE)) {
3608 ANYOF_CLASS_CLEAR(data->start_class,ANYOF_ALNUM);
3609 for (value = 0; value < 256; value++)
3611 ANYOF_BITMAP_CLEAR(data->start_class, value);
3615 if (data->start_class->flags & ANYOF_LOCALE)
3616 ANYOF_CLASS_SET(data->start_class,ANYOF_NALNUM);
3618 for (value = 0; value < 256; value++)
3619 if (!isALNUM(value))
3620 ANYOF_BITMAP_SET(data->start_class, value);
3625 if (flags & SCF_DO_STCLASS_AND) {
3626 if (data->start_class->flags & ANYOF_LOCALE)
3627 ANYOF_CLASS_CLEAR(data->start_class,ANYOF_ALNUM);
3630 data->start_class->flags |= ANYOF_LOCALE;
3631 ANYOF_CLASS_SET(data->start_class,ANYOF_NALNUM);
3635 if (flags & SCF_DO_STCLASS_AND) {
3636 if (!(data->start_class->flags & ANYOF_LOCALE)) {
3637 ANYOF_CLASS_CLEAR(data->start_class,ANYOF_NSPACE);
3638 for (value = 0; value < 256; value++)
3639 if (!isSPACE(value))
3640 ANYOF_BITMAP_CLEAR(data->start_class, value);
3644 if (data->start_class->flags & ANYOF_LOCALE)
3645 ANYOF_CLASS_SET(data->start_class,ANYOF_SPACE);
3647 for (value = 0; value < 256; value++)
3649 ANYOF_BITMAP_SET(data->start_class, value);
3654 if (flags & SCF_DO_STCLASS_AND) {
3655 if (data->start_class->flags & ANYOF_LOCALE)
3656 ANYOF_CLASS_CLEAR(data->start_class,ANYOF_NSPACE);
3659 data->start_class->flags |= ANYOF_LOCALE;
3660 ANYOF_CLASS_SET(data->start_class,ANYOF_SPACE);
3664 if (flags & SCF_DO_STCLASS_AND) {
3665 if (!(data->start_class->flags & ANYOF_LOCALE)) {
3666 ANYOF_CLASS_CLEAR(data->start_class,ANYOF_SPACE);
3667 for (value = 0; value < 256; value++)
3669 ANYOF_BITMAP_CLEAR(data->start_class, value);
3673 if (data->start_class->flags & ANYOF_LOCALE)
3674 ANYOF_CLASS_SET(data->start_class,ANYOF_NSPACE);
3676 for (value = 0; value < 256; value++)
3677 if (!isSPACE(value))
3678 ANYOF_BITMAP_SET(data->start_class, value);
3683 if (flags & SCF_DO_STCLASS_AND) {
3684 if (data->start_class->flags & ANYOF_LOCALE) {
3685 ANYOF_CLASS_CLEAR(data->start_class,ANYOF_SPACE);
3686 for (value = 0; value < 256; value++)
3687 if (!isSPACE(value))
3688 ANYOF_BITMAP_CLEAR(data->start_class, value);
3692 data->start_class->flags |= ANYOF_LOCALE;
3693 ANYOF_CLASS_SET(data->start_class,ANYOF_NSPACE);
3697 if (flags & SCF_DO_STCLASS_AND) {
3698 ANYOF_CLASS_CLEAR(data->start_class,ANYOF_NDIGIT);
3699 for (value = 0; value < 256; value++)
3700 if (!isDIGIT(value))
3701 ANYOF_BITMAP_CLEAR(data->start_class, value);
3704 if (data->start_class->flags & ANYOF_LOCALE)
3705 ANYOF_CLASS_SET(data->start_class,ANYOF_DIGIT);
3707 for (value = 0; value < 256; value++)
3709 ANYOF_BITMAP_SET(data->start_class, value);
3714 if (flags & SCF_DO_STCLASS_AND) {
3715 ANYOF_CLASS_CLEAR(data->start_class,ANYOF_DIGIT);
3716 for (value = 0; value < 256; value++)
3718 ANYOF_BITMAP_CLEAR(data->start_class, value);
3721 if (data->start_class->flags & ANYOF_LOCALE)
3722 ANYOF_CLASS_SET(data->start_class,ANYOF_NDIGIT);
3724 for (value = 0; value < 256; value++)
3725 if (!isDIGIT(value))
3726 ANYOF_BITMAP_SET(data->start_class, value);
3730 CASE_SYNST_FNC(VERTWS);
3731 CASE_SYNST_FNC(HORIZWS);
3734 if (flags & SCF_DO_STCLASS_OR)
3735 cl_and(data->start_class, and_withp);
3736 flags &= ~SCF_DO_STCLASS;
3739 else if (PL_regkind[OP(scan)] == EOL && flags & SCF_DO_SUBSTR) {
3740 data->flags |= (OP(scan) == MEOL
3744 else if ( PL_regkind[OP(scan)] == BRANCHJ
3745 /* Lookbehind, or need to calculate parens/evals/stclass: */
3746 && (scan->flags || data || (flags & SCF_DO_STCLASS))
3747 && (OP(scan) == IFMATCH || OP(scan) == UNLESSM)) {
3748 if ( !PERL_ENABLE_POSITIVE_ASSERTION_STUDY
3749 || OP(scan) == UNLESSM )
3751 /* Negative Lookahead/lookbehind
3752 In this case we can't do fixed string optimisation.
3755 I32 deltanext, minnext, fake = 0;
3757 struct regnode_charclass_class intrnl;
3760 data_fake.flags = 0;
3762 data_fake.whilem_c = data->whilem_c;
3763 data_fake.last_closep = data->last_closep;
3766 data_fake.last_closep = &fake;
3767 data_fake.pos_delta = delta;
3768 if ( flags & SCF_DO_STCLASS && !scan->flags
3769 && OP(scan) == IFMATCH ) { /* Lookahead */
3770 cl_init(pRExC_state, &intrnl);
3771 data_fake.start_class = &intrnl;
3772 f |= SCF_DO_STCLASS_AND;
3774 if (flags & SCF_WHILEM_VISITED_POS)
3775 f |= SCF_WHILEM_VISITED_POS;
3776 next = regnext(scan);
3777 nscan = NEXTOPER(NEXTOPER(scan));
3778 minnext = study_chunk(pRExC_state, &nscan, minlenp, &deltanext,
3779 last, &data_fake, stopparen, recursed, NULL, f, depth+1);
3782 FAIL("Variable length lookbehind not implemented");
3784 else if (minnext > (I32)U8_MAX) {
3785 FAIL2("Lookbehind longer than %"UVuf" not implemented", (UV)U8_MAX);
3787 scan->flags = (U8)minnext;
3790 if (data_fake.flags & (SF_HAS_PAR|SF_IN_PAR))
3792 if (data_fake.flags & SF_HAS_EVAL)
3793 data->flags |= SF_HAS_EVAL;
3794 data->whilem_c = data_fake.whilem_c;
3796 if (f & SCF_DO_STCLASS_AND) {
3797 if (flags & SCF_DO_STCLASS_OR) {
3798 /* OR before, AND after: ideally we would recurse with
3799 * data_fake to get the AND applied by study of the
3800 * remainder of the pattern, and then derecurse;
3801 * *** HACK *** for now just treat as "no information".
3802 * See [perl #56690].
3804 cl_init(pRExC_state, data->start_class);
3806 /* AND before and after: combine and continue */
3807 const int was = (data->start_class->flags & ANYOF_EOS);
3809 cl_and(data->start_class, &intrnl);
3811 data->start_class->flags |= ANYOF_EOS;
3815 #if PERL_ENABLE_POSITIVE_ASSERTION_STUDY
3817 /* Positive Lookahead/lookbehind
3818 In this case we can do fixed string optimisation,
3819 but we must be careful about it. Note in the case of
3820 lookbehind the positions will be offset by the minimum
3821 length of the pattern, something we won't know about
3822 until after the recurse.
3824 I32 deltanext, fake = 0;
3826 struct regnode_charclass_class intrnl;
3828 /* We use SAVEFREEPV so that when the full compile
3829 is finished perl will clean up the allocated
3830 minlens when its all done. This was we don't
3831 have to worry about freeing them when we know
3832 they wont be used, which would be a pain.
3835 Newx( minnextp, 1, I32 );
3836 SAVEFREEPV(minnextp);
3839 StructCopy(data, &data_fake, scan_data_t);
3840 if ((flags & SCF_DO_SUBSTR) && data->last_found) {
3843 SCAN_COMMIT(pRExC_state, &data_fake,minlenp);
3844 data_fake.last_found=newSVsv(data->last_found);
3848 data_fake.last_closep = &fake;
3849 data_fake.flags = 0;
3850 data_fake.pos_delta = delta;
3852 data_fake.flags |= SF_IS_INF;
3853 if ( flags & SCF_DO_STCLASS && !scan->flags
3854 && OP(scan) == IFMATCH ) { /* Lookahead */
3855 cl_init(pRExC_state, &intrnl);
3856 data_fake.start_class = &intrnl;
3857 f |= SCF_DO_STCLASS_AND;
3859 if (flags & SCF_WHILEM_VISITED_POS)
3860 f |= SCF_WHILEM_VISITED_POS;
3861 next = regnext(scan);
3862 nscan = NEXTOPER(NEXTOPER(scan));
3864 *minnextp = study_chunk(pRExC_state, &nscan, minnextp, &deltanext,
3865 last, &data_fake, stopparen, recursed, NULL, f,depth+1);
3868 FAIL("Variable length lookbehind not implemented");
3870 else if (*minnextp > (I32)U8_MAX) {
3871 FAIL2("Lookbehind longer than %"UVuf" not implemented", (UV)U8_MAX);
3873 scan->flags = (U8)*minnextp;
3878 if (f & SCF_DO_STCLASS_AND) {
3879 const int was = (data->start_class->flags & ANYOF_EOS);
3881 cl_and(data->start_class, &intrnl);
3883 data->start_class->flags |= ANYOF_EOS;
3886 if (data_fake.flags & (SF_HAS_PAR|SF_IN_PAR))
3888 if (data_fake.flags & SF_HAS_EVAL)
3889 data->flags |= SF_HAS_EVAL;
3890 data->whilem_c = data_fake.whilem_c;
3891 if ((flags & SCF_DO_SUBSTR) && data_fake.last_found) {
3892 if (RExC_rx->minlen<*minnextp)
3893 RExC_rx->minlen=*minnextp;
3894 SCAN_COMMIT(pRExC_state, &data_fake, minnextp);
3895 SvREFCNT_dec(data_fake.last_found);
3897 if ( data_fake.minlen_fixed != minlenp )
3899 data->offset_fixed= data_fake.offset_fixed;
3900 data->minlen_fixed= data_fake.minlen_fixed;
3901 data->lookbehind_fixed+= scan->flags;
3903 if ( data_fake.minlen_float != minlenp )
3905 data->minlen_float= data_fake.minlen_float;
3906 data->offset_float_min=data_fake.offset_float_min;
3907 data->offset_float_max=data_fake.offset_float_max;
3908 data->lookbehind_float+= scan->flags;
3917 else if (OP(scan) == OPEN) {
3918 if (stopparen != (I32)ARG(scan))
3921 else if (OP(scan) == CLOSE) {
3922 if (stopparen == (I32)ARG(scan)) {
3925 if ((I32)ARG(scan) == is_par) {
3926 next = regnext(scan);
3928 if ( next && (OP(next) != WHILEM) && next < last)
3929 is_par = 0; /* Disable optimization */
3932 *(data->last_closep) = ARG(scan);
3934 else if (OP(scan) == EVAL) {
3936 data->flags |= SF_HAS_EVAL;
3938 else if ( PL_regkind[OP(scan)] == ENDLIKE ) {
3939 if (flags & SCF_DO_SUBSTR) {
3940 SCAN_COMMIT(pRExC_state,data,minlenp);
3941 flags &= ~SCF_DO_SUBSTR;
3943 if (data && OP(scan)==ACCEPT) {
3944 data->flags |= SCF_SEEN_ACCEPT;
3949 else if (OP(scan) == LOGICAL && scan->flags == 2) /* Embedded follows */
3951 if (flags & SCF_DO_SUBSTR) {
3952 SCAN_COMMIT(pRExC_state,data,minlenp);
3953 data->longest = &(data->longest_float);
3955 is_inf = is_inf_internal = 1;
3956 if (flags & SCF_DO_STCLASS_OR) /* Allow everything */
3957 cl_anything(pRExC_state, data->start_class);
3958 flags &= ~SCF_DO_STCLASS;
3960 else if (OP(scan) == GPOS) {
3961 if (!(RExC_rx->extflags & RXf_GPOS_FLOAT) &&
3962 !(delta || is_inf || (data && data->pos_delta)))
3964 if (!(RExC_rx->extflags & RXf_ANCH) && (flags & SCF_DO_SUBSTR))
3965 RExC_rx->extflags |= RXf_ANCH_GPOS;
3966 if (RExC_rx->gofs < (U32)min)
3967 RExC_rx->gofs = min;
3969 RExC_rx->extflags |= RXf_GPOS_FLOAT;
3973 #ifdef TRIE_STUDY_OPT
3974 #ifdef FULL_TRIE_STUDY
3975 else if (PL_regkind[OP(scan)] == TRIE) {
3976 /* NOTE - There is similar code to this block above for handling
3977 BRANCH nodes on the initial study. If you change stuff here
3979 regnode *trie_node= scan;
3980 regnode *tail= regnext(scan);
3981 reg_trie_data *trie = (reg_trie_data*)RExC_rxi->data->data[ ARG(scan) ];
3982 I32 max1 = 0, min1 = I32_MAX;
3983 struct regnode_charclass_class accum;
3985 if (flags & SCF_DO_SUBSTR) /* XXXX Add !SUSPEND? */
3986 SCAN_COMMIT(pRExC_state, data,minlenp); /* Cannot merge strings after this. */
3987 if (flags & SCF_DO_STCLASS)
3988 cl_init_zero(pRExC_state, &accum);
3994 const regnode *nextbranch= NULL;
3997 for ( word=1 ; word <= trie->wordcount ; word++)
3999 I32 deltanext=0, minnext=0, f = 0, fake;
4000 struct regnode_charclass_class this_class;
4002 data_fake.flags = 0;
4004 data_fake.whilem_c = data->whilem_c;
4005 data_fake.last_closep = data->last_closep;
4008 data_fake.last_closep = &fake;
4009 data_fake.pos_delta = delta;
4010 if (flags & SCF_DO_STCLASS) {
4011 cl_init(pRExC_state, &this_class);
4012 data_fake.start_class = &this_class;
4013 f = SCF_DO_STCLASS_AND;
4015 if (flags & SCF_WHILEM_VISITED_POS)
4016 f |= SCF_WHILEM_VISITED_POS;
4018 if (trie->jump[word]) {
4020 nextbranch = trie_node + trie->jump[0];
4021 scan= trie_node + trie->jump[word];
4022 /* We go from the jump point to the branch that follows
4023 it. Note this means we need the vestigal unused branches
4024 even though they arent otherwise used.
4026 minnext = study_chunk(pRExC_state, &scan, minlenp,
4027 &deltanext, (regnode *)nextbranch, &data_fake,
4028 stopparen, recursed, NULL, f,depth+1);
4030 if (nextbranch && PL_regkind[OP(nextbranch)]==BRANCH)
4031 nextbranch= regnext((regnode*)nextbranch);
4033 if (min1 > (I32)(minnext + trie->minlen))
4034 min1 = minnext + trie->minlen;
4035 if (max1 < (I32)(minnext + deltanext + trie->maxlen))
4036 max1 = minnext + deltanext + trie->maxlen;
4037 if (deltanext == I32_MAX)
4038 is_inf = is_inf_internal = 1;
4040 if (data_fake.flags & (SF_HAS_PAR|SF_IN_PAR))
4042 if (data_fake.flags & SCF_SEEN_ACCEPT) {
4043 if ( stopmin > min + min1)
4044 stopmin = min + min1;
4045 flags &= ~SCF_DO_SUBSTR;
4047 data->flags |= SCF_SEEN_ACCEPT;
4050 if (data_fake.flags & SF_HAS_EVAL)
4051 data->flags |= SF_HAS_EVAL;
4052 data->whilem_c = data_fake.whilem_c;
4054 if (flags & SCF_DO_STCLASS)
4055 cl_or(pRExC_state, &accum, &this_class);
4058 if (flags & SCF_DO_SUBSTR) {
4059 data->pos_min += min1;
4060 data->pos_delta += max1 - min1;
4061 if (max1 != min1 || is_inf)
4062 data->longest = &(data->longest_float);
4065 delta += max1 - min1;
4066 if (flags & SCF_DO_STCLASS_OR) {
4067 cl_or(pRExC_state, data->start_class, &accum);
4069 cl_and(data->start_class, and_withp);
4070 flags &= ~SCF_DO_STCLASS;
4073 else if (flags & SCF_DO_STCLASS_AND) {
4075 cl_and(data->start_class, &accum);
4076 flags &= ~SCF_DO_STCLASS;
4079 /* Switch to OR mode: cache the old value of
4080 * data->start_class */
4082 StructCopy(data->start_class, and_withp,
4083 struct regnode_charclass_class);
4084 flags &= ~SCF_DO_STCLASS_AND;
4085 StructCopy(&accum, data->start_class,
4086 struct regnode_charclass_class);
4087 flags |= SCF_DO_STCLASS_OR;
4088 data->start_class->flags |= ANYOF_EOS;
4095 else if (PL_regkind[OP(scan)] == TRIE) {
4096 reg_trie_data *trie = (reg_trie_data*)RExC_rxi->data->data[ ARG(scan) ];
4099 min += trie->minlen;
4100 delta += (trie->maxlen - trie->minlen);
4101 flags &= ~SCF_DO_STCLASS; /* xxx */
4102 if (flags & SCF_DO_SUBSTR) {
4103 SCAN_COMMIT(pRExC_state,data,minlenp); /* Cannot expect anything... */
4104 data->pos_min += trie->minlen;
4105 data->pos_delta += (trie->maxlen - trie->minlen);
4106 if (trie->maxlen != trie->minlen)
4107 data->longest = &(data->longest_float);
4109 if (trie->jump) /* no more substrings -- for now /grr*/
4110 flags &= ~SCF_DO_SUBSTR;
4112 #endif /* old or new */
4113 #endif /* TRIE_STUDY_OPT */
4115 /* Else: zero-length, ignore. */
4116 scan = regnext(scan);
4121 stopparen = frame->stop;
4122 frame = frame->prev;
4123 goto fake_study_recurse;
4128 DEBUG_STUDYDATA("pre-fin:",data,depth);
4131 *deltap = is_inf_internal ? I32_MAX : delta;
4132 if (flags & SCF_DO_SUBSTR && is_inf)
4133 data->pos_delta = I32_MAX - data->pos_min;
4134 if (is_par > (I32)U8_MAX)
4136 if (is_par && pars==1 && data) {
4137 data->flags |= SF_IN_PAR;
4138 data->flags &= ~SF_HAS_PAR;
4140 else if (pars && data) {
4141 data->flags |= SF_HAS_PAR;
4142 data->flags &= ~SF_IN_PAR;
4144 if (flags & SCF_DO_STCLASS_OR)
4145 cl_and(data->start_class, and_withp);
4146 if (flags & SCF_TRIE_RESTUDY)
4147 data->flags |= SCF_TRIE_RESTUDY;
4149 DEBUG_STUDYDATA("post-fin:",data,depth);
4151 return min < stopmin ? min : stopmin;
4155 S_add_data(RExC_state_t *pRExC_state, U32 n, const char *s)
4157 U32 count = RExC_rxi->data ? RExC_rxi->data->count : 0;
4159 PERL_ARGS_ASSERT_ADD_DATA;
4161 Renewc(RExC_rxi->data,
4162 sizeof(*RExC_rxi->data) + sizeof(void*) * (count + n - 1),
4163 char, struct reg_data);
4165 Renew(RExC_rxi->data->what, count + n, U8);
4167 Newx(RExC_rxi->data->what, n, U8);
4168 RExC_rxi->data->count = count + n;
4169 Copy(s, RExC_rxi->data->what + count, n, U8);
4173 /*XXX: todo make this not included in a non debugging perl */
4174 #ifndef PERL_IN_XSUB_RE
4176 Perl_reginitcolors(pTHX)
4179 const char * const s = PerlEnv_getenv("PERL_RE_COLORS");
4181 char *t = savepv(s);
4185 t = strchr(t, '\t');
4191 PL_colors[i] = t = (char *)"";
4196 PL_colors[i++] = (char *)"";
4203 #ifdef TRIE_STUDY_OPT
4204 #define CHECK_RESTUDY_GOTO \
4206 (data.flags & SCF_TRIE_RESTUDY) \
4210 #define CHECK_RESTUDY_GOTO
4214 - pregcomp - compile a regular expression into internal code
4216 * We can't allocate space until we know how big the compiled form will be,
4217 * but we can't compile it (and thus know how big it is) until we've got a
4218 * place to put the code. So we cheat: we compile it twice, once with code
4219 * generation turned off and size counting turned on, and once "for real".
4220 * This also means that we don't allocate space until we are sure that the
4221 * thing really will compile successfully, and we never have to move the
4222 * code and thus invalidate pointers into it. (Note that it has to be in
4223 * one piece because free() must be able to free it all.) [NB: not true in perl]
4225 * Beware that the optimization-preparation code in here knows about some
4226 * of the structure of the compiled regexp. [I'll say.]
4231 #ifndef PERL_IN_XSUB_RE
4232 #define RE_ENGINE_PTR &PL_core_reg_engine
4234 extern const struct regexp_engine my_reg_engine;
4235 #define RE_ENGINE_PTR &my_reg_engine
4238 #ifndef PERL_IN_XSUB_RE
4240 Perl_pregcomp(pTHX_ SV * const pattern, const U32 flags)
4243 HV * const table = GvHV(PL_hintgv);
4245 PERL_ARGS_ASSERT_PREGCOMP;
4247 /* Dispatch a request to compile a regexp to correct
4250 SV **ptr= hv_fetchs(table, "regcomp", FALSE);
4251 GET_RE_DEBUG_FLAGS_DECL;
4252 if (ptr && SvIOK(*ptr) && SvIV(*ptr)) {
4253 const regexp_engine *eng=INT2PTR(regexp_engine*,SvIV(*ptr));
4255 PerlIO_printf(Perl_debug_log, "Using engine %"UVxf"\n",
4258 return CALLREGCOMP_ENG(eng, pattern, flags);
4261 return Perl_re_compile(aTHX_ pattern, flags);
4266 Perl_re_compile(pTHX_ SV * const pattern, U32 pm_flags)
4271 register regexp_internal *ri;
4273 char *exp = SvPV(pattern, plen);
4274 char* xend = exp + plen;
4283 RExC_state_t RExC_state;
4284 RExC_state_t * const pRExC_state = &RExC_state;
4285 #ifdef TRIE_STUDY_OPT
4287 RExC_state_t copyRExC_state;
4289 GET_RE_DEBUG_FLAGS_DECL;
4291 PERL_ARGS_ASSERT_RE_COMPILE;
4293 DEBUG_r(if (!PL_colorset) reginitcolors());
4295 RExC_utf8 = RExC_orig_utf8 = SvUTF8(pattern);
4298 SV *dsv= sv_newmortal();
4299 RE_PV_QUOTED_DECL(s, RExC_utf8,
4300 dsv, exp, plen, 60);
4301 PerlIO_printf(Perl_debug_log, "%sCompiling REx%s %s\n",
4302 PL_colors[4],PL_colors[5],s);
4305 /* Longjmp back to here if have to switch in midstream to utf8 */
4306 if (! RExC_orig_utf8) {
4307 JMPENV_PUSH(jump_ret);
4310 if (jump_ret != 0) {
4313 /* Here, we longjmped back. If the cause was other than changing to
4314 * utf8, pop our own setjmp, and longjmp to the correct handler */
4315 if (jump_ret != UTF8_LONGJMP) {
4317 JMPENV_JUMP(jump_ret);
4322 /* It's possible to write a regexp in ascii that represents Unicode
4323 codepoints outside of the byte range, such as via \x{100}. If we
4324 detect such a sequence we have to convert the entire pattern to utf8
4325 and then recompile, as our sizing calculation will have been based
4326 on 1 byte == 1 character, but we will need to use utf8 to encode
4327 at least some part of the pattern, and therefore must convert the whole
4330 DEBUG_PARSE_r(PerlIO_printf(Perl_debug_log,
4331 "UTF8 mismatch! Converting to utf8 for resizing and compile\n"));
4332 exp = (char*)Perl_bytes_to_utf8(aTHX_ (U8*)SvPV(pattern, plen), &len);
4334 RExC_orig_utf8 = RExC_utf8 = 1;
4339 RExC_flags = pm_flags;
4343 RExC_seen_zerolen = *exp == '^' ? -1 : 0;
4344 RExC_seen_evals = 0;
4347 /* First pass: determine size, legality. */
4355 RExC_emit = &PL_regdummy;
4356 RExC_whilem_seen = 0;
4357 RExC_open_parens = NULL;
4358 RExC_close_parens = NULL;
4360 RExC_paren_names = NULL;
4362 RExC_paren_name_list = NULL;
4364 RExC_recurse = NULL;
4365 RExC_recurse_count = 0;
4367 #if 0 /* REGC() is (currently) a NOP at the first pass.
4368 * Clever compilers notice this and complain. --jhi */
4369 REGC((U8)REG_MAGIC, (char*)RExC_emit);
4371 DEBUG_PARSE_r(PerlIO_printf(Perl_debug_log, "Starting first pass (sizing)\n"));
4372 if (reg(pRExC_state, 0, &flags,1) == NULL) {
4373 RExC_precomp = NULL;
4377 /* Here, finished first pass. Get rid of our setjmp, which we added for
4378 * efficiency only if the passed-in string wasn't in utf8, as shown by
4379 * RExC_orig_utf8. But if the first pass was redone, that variable will be
4380 * 1 here even though the original string wasn't utf8, but in this case
4381 * there will have been a long jump */
4382 if (jump_ret == UTF8_LONGJMP || ! RExC_orig_utf8) {
4386 PerlIO_printf(Perl_debug_log,
4387 "Required size %"IVdf" nodes\n"
4388 "Starting second pass (creation)\n",
4391 RExC_lastparse=NULL;
4393 /* Small enough for pointer-storage convention?
4394 If extralen==0, this means that we will not need long jumps. */
4395 if (RExC_size >= 0x10000L && RExC_extralen)
4396 RExC_size += RExC_extralen;
4399 if (RExC_whilem_seen > 15)
4400 RExC_whilem_seen = 15;
4402 /* Allocate space and zero-initialize. Note, the two step process
4403 of zeroing when in debug mode, thus anything assigned has to
4404 happen after that */
4405 rx = (REGEXP*) newSV_type(SVt_REGEXP);
4406 r = (struct regexp*)SvANY(rx);
4407 Newxc(ri, sizeof(regexp_internal) + (unsigned)RExC_size * sizeof(regnode),
4408 char, regexp_internal);
4409 if ( r == NULL || ri == NULL )
4410 FAIL("Regexp out of space");
4412 /* avoid reading uninitialized memory in DEBUGGING code in study_chunk() */
4413 Zero(ri, sizeof(regexp_internal) + (unsigned)RExC_size * sizeof(regnode), char);
4415 /* bulk initialize base fields with 0. */
4416 Zero(ri, sizeof(regexp_internal), char);
4419 /* non-zero initialization begins here */
4421 r->engine= RE_ENGINE_PTR;
4422 r->extflags = pm_flags;
4424 bool has_p = ((r->extflags & RXf_PMf_KEEPCOPY) == RXf_PMf_KEEPCOPY);
4425 bool has_minus = ((r->extflags & RXf_PMf_STD_PMMOD) != RXf_PMf_STD_PMMOD);
4426 bool has_runon = ((RExC_seen & REG_SEEN_RUN_ON_COMMENT)==REG_SEEN_RUN_ON_COMMENT);
4427 U16 reganch = (U16)((r->extflags & RXf_PMf_STD_PMMOD)
4428 >> RXf_PMf_STD_PMMOD_SHIFT);
4429 const char *fptr = STD_PAT_MODS; /*"msix"*/
4431 const STRLEN wraplen = plen + has_minus + has_p + has_runon
4432 + (sizeof(STD_PAT_MODS) - 1)
4433 + (sizeof("(?:)") - 1);
4435 p = sv_grow(MUTABLE_SV(rx), wraplen + 1);
4436 SvCUR_set(rx, wraplen);
4438 SvFLAGS(rx) |= SvUTF8(pattern);
4441 *p++ = KEEPCOPY_PAT_MOD; /*'p'*/
4443 char *r = p + (sizeof(STD_PAT_MODS) - 1) + has_minus - 1;
4444 char *colon = r + 1;
4447 while((ch = *fptr++)) {
4461 Copy(RExC_precomp, p, plen, char);
4462 assert ((RX_WRAPPED(rx) - p) < 16);
4463 r->pre_prefix = p - RX_WRAPPED(rx);
4472 r->nparens = RExC_npar - 1; /* set early to validate backrefs */
4474 if (RExC_seen & REG_SEEN_RECURSE) {
4475 Newxz(RExC_open_parens, RExC_npar,regnode *);
4476 SAVEFREEPV(RExC_open_parens);
4477 Newxz(RExC_close_parens,RExC_npar,regnode *);
4478 SAVEFREEPV(RExC_close_parens);
4481 /* Useful during FAIL. */
4482 #ifdef RE_TRACK_PATTERN_OFFSETS
4483 Newxz(ri->u.offsets, 2*RExC_size+1, U32); /* MJD 20001228 */
4484 DEBUG_OFFSETS_r(PerlIO_printf(Perl_debug_log,
4485 "%s %"UVuf" bytes for offset annotations.\n",
4486 ri->u.offsets ? "Got" : "Couldn't get",
4487 (UV)((2*RExC_size+1) * sizeof(U32))));
4489 SetProgLen(ri,RExC_size);
4494 /* Second pass: emit code. */
4495 RExC_flags = pm_flags; /* don't let top level (?i) bleed */
4500 RExC_emit_start = ri->program;
4501 RExC_emit = ri->program;
4502 RExC_emit_bound = ri->program + RExC_size + 1;
4504 /* Store the count of eval-groups for security checks: */
4505 RExC_rx->seen_evals = RExC_seen_evals;
4506 REGC((U8)REG_MAGIC, (char*) RExC_emit++);
4507 if (reg(pRExC_state, 0, &flags,1) == NULL) {
4511 /* XXXX To minimize changes to RE engine we always allocate
4512 3-units-long substrs field. */
4513 Newx(r->substrs, 1, struct reg_substr_data);
4514 if (RExC_recurse_count) {
4515 Newxz(RExC_recurse,RExC_recurse_count,regnode *);
4516 SAVEFREEPV(RExC_recurse);
4520 r->minlen = minlen = sawplus = sawopen = 0;
4521 Zero(r->substrs, 1, struct reg_substr_data);
4523 #ifdef TRIE_STUDY_OPT
4525 StructCopy(&zero_scan_data, &data, scan_data_t);
4526 copyRExC_state = RExC_state;
4529 DEBUG_OPTIMISE_r(PerlIO_printf(Perl_debug_log,"Restudying\n"));
4531 RExC_state = copyRExC_state;
4532 if (seen & REG_TOP_LEVEL_BRANCHES)
4533 RExC_seen |= REG_TOP_LEVEL_BRANCHES;
4535 RExC_seen &= ~REG_TOP_LEVEL_BRANCHES;
4536 if (data.last_found) {
4537 SvREFCNT_dec(data.longest_fixed);
4538 SvREFCNT_dec(data.longest_float);
4539 SvREFCNT_dec(data.last_found);
4541 StructCopy(&zero_scan_data, &data, scan_data_t);
4544 StructCopy(&zero_scan_data, &data, scan_data_t);
4547 /* Dig out information for optimizations. */
4548 r->extflags = RExC_flags; /* was pm_op */
4549 /*dmq: removed as part of de-PMOP: pm->op_pmflags = RExC_flags; */
4552 SvUTF8_on(rx); /* Unicode in it? */
4553 ri->regstclass = NULL;
4554 if (RExC_naughty >= 10) /* Probably an expensive pattern. */
4555 r->intflags |= PREGf_NAUGHTY;
4556 scan = ri->program + 1; /* First BRANCH. */
4558 /* testing for BRANCH here tells us whether there is "must appear"
4559 data in the pattern. If there is then we can use it for optimisations */
4560 if (!(RExC_seen & REG_TOP_LEVEL_BRANCHES)) { /* Only one top-level choice. */
4562 STRLEN longest_float_length, longest_fixed_length;
4563 struct regnode_charclass_class ch_class; /* pointed to by data */
4565 I32 last_close = 0; /* pointed to by data */
4566 regnode *first= scan;
4567 regnode *first_next= regnext(first);
4570 * Skip introductions and multiplicators >= 1
4571 * so that we can extract the 'meat' of the pattern that must
4572 * match in the large if() sequence following.
4573 * NOTE that EXACT is NOT covered here, as it is normally
4574 * picked up by the optimiser separately.
4576 * This is unfortunate as the optimiser isnt handling lookahead
4577 * properly currently.
4580 while ((OP(first) == OPEN && (sawopen = 1)) ||
4581 /* An OR of *one* alternative - should not happen now. */
4582 (OP(first) == BRANCH && OP(first_next) != BRANCH) ||
4583 /* for now we can't handle lookbehind IFMATCH*/
4584 (OP(first) == IFMATCH && !first->flags) ||
4585 (OP(first) == PLUS) ||
4586 (OP(first) == MINMOD) ||
4587 /* An {n,m} with n>0 */
4588 (PL_regkind[OP(first)] == CURLY && ARG1(first) > 0) ||
4589 (OP(first) == NOTHING && PL_regkind[OP(first_next)] != END ))
4592 * the only op that could be a regnode is PLUS, all the rest
4593 * will be regnode_1 or regnode_2.
4596 if (OP(first) == PLUS)
4599 first += regarglen[OP(first)];
4601 first = NEXTOPER(first);
4602 first_next= regnext(first);
4605 /* Starting-point info. */
4607 DEBUG_PEEP("first:",first,0);
4608 /* Ignore EXACT as we deal with it later. */
4609 if (PL_regkind[OP(first)] == EXACT) {
4610 if (OP(first) == EXACT)
4611 NOOP; /* Empty, get anchored substr later. */
4612 else if ((OP(first) == EXACTF || OP(first) == EXACTFL))
4613 ri->regstclass = first;
4616 else if (PL_regkind[OP(first)] == TRIE &&
4617 ((reg_trie_data *)ri->data->data[ ARG(first) ])->minlen>0)
4620 /* this can happen only on restudy */
4621 if ( OP(first) == TRIE ) {
4622 struct regnode_1 *trieop = (struct regnode_1 *)
4623 PerlMemShared_calloc(1, sizeof(struct regnode_1));
4624 StructCopy(first,trieop,struct regnode_1);
4625 trie_op=(regnode *)trieop;
4627 struct regnode_charclass *trieop = (struct regnode_charclass *)
4628 PerlMemShared_calloc(1, sizeof(struct regnode_charclass));
4629 StructCopy(first,trieop,struct regnode_charclass);
4630 trie_op=(regnode *)trieop;
4633 make_trie_failtable(pRExC_state, (regnode *)first, trie_op, 0);
4634 ri->regstclass = trie_op;
4637 else if (REGNODE_SIMPLE(OP(first)))
4638 ri->regstclass = first;
4639 else if (PL_regkind[OP(first)] == BOUND ||
4640 PL_regkind[OP(first)] == NBOUND)
4641 ri->regstclass = first;
4642 else if (PL_regkind[OP(first)] == BOL) {
4643 r->extflags |= (OP(first) == MBOL
4645 : (OP(first) == SBOL
4648 first = NEXTOPER(first);
4651 else if (OP(first) == GPOS) {
4652 r->extflags |= RXf_ANCH_GPOS;
4653 first = NEXTOPER(first);
4656 else if ((!sawopen || !RExC_sawback) &&
4657 (OP(first) == STAR &&
4658 PL_regkind[OP(NEXTOPER(first))] == REG_ANY) &&
4659 !(r->extflags & RXf_ANCH) && !(RExC_seen & REG_SEEN_EVAL))
4661 /* turn .* into ^.* with an implied $*=1 */
4663 (OP(NEXTOPER(first)) == REG_ANY)
4666 r->extflags |= type;
4667 r->intflags |= PREGf_IMPLICIT;
4668 first = NEXTOPER(first);
4671 if (sawplus && (!sawopen || !RExC_sawback)
4672 && !(RExC_seen & REG_SEEN_EVAL)) /* May examine pos and $& */
4673 /* x+ must match at the 1st pos of run of x's */
4674 r->intflags |= PREGf_SKIP;
4676 /* Scan is after the zeroth branch, first is atomic matcher. */
4677 #ifdef TRIE_STUDY_OPT
4680 PerlIO_printf(Perl_debug_log, "first at %"IVdf"\n",
4681 (IV)(first - scan + 1))
4685 PerlIO_printf(Perl_debug_log, "first at %"IVdf"\n",
4686 (IV)(first - scan + 1))
4692 * If there's something expensive in the r.e., find the
4693 * longest literal string that must appear and make it the
4694 * regmust. Resolve ties in favor of later strings, since
4695 * the regstart check works with the beginning of the r.e.
4696 * and avoiding duplication strengthens checking. Not a
4697 * strong reason, but sufficient in the absence of others.
4698 * [Now we resolve ties in favor of the earlier string if
4699 * it happens that c_offset_min has been invalidated, since the
4700 * earlier string may buy us something the later one won't.]
4703 data.longest_fixed = newSVpvs("");
4704 data.longest_float = newSVpvs("");
4705 data.last_found = newSVpvs("");
4706 data.longest = &(data.longest_fixed);
4708 if (!ri->regstclass) {
4709 cl_init(pRExC_state, &ch_class);
4710 data.start_class = &ch_class;
4711 stclass_flag = SCF_DO_STCLASS_AND;
4712 } else /* XXXX Check for BOUND? */
4714 data.last_closep = &last_close;
4716 minlen = study_chunk(pRExC_state, &first, &minlen, &fake, scan + RExC_size, /* Up to end */
4717 &data, -1, NULL, NULL,
4718 SCF_DO_SUBSTR | SCF_WHILEM_VISITED_POS | stclass_flag,0);
4724 if ( RExC_npar == 1 && data.longest == &(data.longest_fixed)
4725 && data.last_start_min == 0 && data.last_end > 0
4726 && !RExC_seen_zerolen
4727 && !(RExC_seen & REG_SEEN_VERBARG)
4728 && (!(RExC_seen & REG_SEEN_GPOS) || (r->extflags & RXf_ANCH_GPOS)))
4729 r->extflags |= RXf_CHECK_ALL;
4730 scan_commit(pRExC_state, &data,&minlen,0);
4731 SvREFCNT_dec(data.last_found);
4733 /* Note that code very similar to this but for anchored string
4734 follows immediately below, changes may need to be made to both.
4737 longest_float_length = CHR_SVLEN(data.longest_float);
4738 if (longest_float_length
4739 || (data.flags & SF_FL_BEFORE_EOL
4740 && (!(data.flags & SF_FL_BEFORE_MEOL)
4741 || (RExC_flags & RXf_PMf_MULTILINE))))
4745 if (SvCUR(data.longest_fixed) /* ok to leave SvCUR */
4746 && data.offset_fixed == data.offset_float_min
4747 && SvCUR(data.longest_fixed) == SvCUR(data.longest_float))
4748 goto remove_float; /* As in (a)+. */
4750 /* copy the information about the longest float from the reg_scan_data
4751 over to the program. */
4752 if (SvUTF8(data.longest_float)) {
4753 r->float_utf8 = data.longest_float;
4754 r->float_substr = NULL;
4756 r->float_substr = data.longest_float;
4757 r->float_utf8 = NULL;
4759 /* float_end_shift is how many chars that must be matched that
4760 follow this item. We calculate it ahead of time as once the
4761 lookbehind offset is added in we lose the ability to correctly
4763 ml = data.minlen_float ? *(data.minlen_float)
4764 : (I32)longest_float_length;
4765 r->float_end_shift = ml - data.offset_float_min
4766 - longest_float_length + (SvTAIL(data.longest_float) != 0)
4767 + data.lookbehind_float;
4768 r->float_min_offset = data.offset_float_min - data.lookbehind_float;
4769 r->float_max_offset = data.offset_float_max;
4770 if (data.offset_float_max < I32_MAX) /* Don't offset infinity */
4771 r->float_max_offset -= data.lookbehind_float;
4773 t = (data.flags & SF_FL_BEFORE_EOL /* Can't have SEOL and MULTI */
4774 && (!(data.flags & SF_FL_BEFORE_MEOL)
4775 || (RExC_flags & RXf_PMf_MULTILINE)));
4776 fbm_compile(data.longest_float, t ? FBMcf_TAIL : 0);
4780 r->float_substr = r->float_utf8 = NULL;
4781 SvREFCNT_dec(data.longest_float);
4782 longest_float_length = 0;
4785 /* Note that code very similar to this but for floating string
4786 is immediately above, changes may need to be made to both.
4789 longest_fixed_length = CHR_SVLEN(data.longest_fixed);
4790 if (longest_fixed_length
4791 || (data.flags & SF_FIX_BEFORE_EOL /* Cannot have SEOL and MULTI */
4792 && (!(data.flags & SF_FIX_BEFORE_MEOL)
4793 || (RExC_flags & RXf_PMf_MULTILINE))))
4797 /* copy the information about the longest fixed
4798 from the reg_scan_data over to the program. */
4799 if (SvUTF8(data.longest_fixed)) {
4800 r->anchored_utf8 = data.longest_fixed;
4801 r->anchored_substr = NULL;
4803 r->anchored_substr = data.longest_fixed;
4804 r->anchored_utf8 = NULL;
4806 /* fixed_end_shift is how many chars that must be matched that
4807 follow this item. We calculate it ahead of time as once the
4808 lookbehind offset is added in we lose the ability to correctly
4810 ml = data.minlen_fixed ? *(data.minlen_fixed)
4811 : (I32)longest_fixed_length;
4812 r->anchored_end_shift = ml - data.offset_fixed
4813 - longest_fixed_length + (SvTAIL(data.longest_fixed) != 0)
4814 + data.lookbehind_fixed;
4815 r->anchored_offset = data.offset_fixed - data.lookbehind_fixed;
4817 t = (data.flags & SF_FIX_BEFORE_EOL /* Can't have SEOL and MULTI */
4818 && (!(data.flags & SF_FIX_BEFORE_MEOL)
4819 || (RExC_flags & RXf_PMf_MULTILINE)));
4820 fbm_compile(data.longest_fixed, t ? FBMcf_TAIL : 0);
4823 r->anchored_substr = r->anchored_utf8 = NULL;
4824 SvREFCNT_dec(data.longest_fixed);
4825 longest_fixed_length = 0;
4828 && (OP(ri->regstclass) == REG_ANY || OP(ri->regstclass) == SANY))
4829 ri->regstclass = NULL;
4830 if ((!(r->anchored_substr || r->anchored_utf8) || r->anchored_offset)
4832 && !(data.start_class->flags & ANYOF_EOS)
4833 && !cl_is_anything(data.start_class))
4835 const U32 n = add_data(pRExC_state, 1, "f");
4837 Newx(RExC_rxi->data->data[n], 1,
4838 struct regnode_charclass_class);
4839 StructCopy(data.start_class,
4840 (struct regnode_charclass_class*)RExC_rxi->data->data[n],
4841 struct regnode_charclass_class);
4842 ri->regstclass = (regnode*)RExC_rxi->data->data[n];
4843 r->intflags &= ~PREGf_SKIP; /* Used in find_byclass(). */
4844 DEBUG_COMPILE_r({ SV *sv = sv_newmortal();
4845 regprop(r, sv, (regnode*)data.start_class);
4846 PerlIO_printf(Perl_debug_log,
4847 "synthetic stclass \"%s\".\n",
4848 SvPVX_const(sv));});
4851 /* A temporary algorithm prefers floated substr to fixed one to dig more info. */
4852 if (longest_fixed_length > longest_float_length) {
4853 r->check_end_shift = r->anchored_end_shift;
4854 r->check_substr = r->anchored_substr;
4855 r->check_utf8 = r->anchored_utf8;
4856 r->check_offset_min = r->check_offset_max = r->anchored_offset;
4857 if (r->extflags & RXf_ANCH_SINGLE)
4858 r->extflags |= RXf_NOSCAN;
4861 r->check_end_shift = r->float_end_shift;
4862 r->check_substr = r->float_substr;
4863 r->check_utf8 = r->float_utf8;
4864 r->check_offset_min = r->float_min_offset;
4865 r->check_offset_max = r->float_max_offset;
4867 /* XXXX Currently intuiting is not compatible with ANCH_GPOS.
4868 This should be changed ASAP! */
4869 if ((r->check_substr || r->check_utf8) && !(r->extflags & RXf_ANCH_GPOS)) {
4870 r->extflags |= RXf_USE_INTUIT;
4871 if (SvTAIL(r->check_substr ? r->check_substr : r->check_utf8))
4872 r->extflags |= RXf_INTUIT_TAIL;
4874 /* XXX Unneeded? dmq (shouldn't as this is handled elsewhere)
4875 if ( (STRLEN)minlen < longest_float_length )
4876 minlen= longest_float_length;
4877 if ( (STRLEN)minlen < longest_fixed_length )
4878 minlen= longest_fixed_length;
4882 /* Several toplevels. Best we can is to set minlen. */
4884 struct regnode_charclass_class ch_class;
4887 DEBUG_PARSE_r(PerlIO_printf(Perl_debug_log, "\nMulti Top Level\n"));
4889 scan = ri->program + 1;
4890 cl_init(pRExC_state, &ch_class);
4891 data.start_class = &ch_class;
4892 data.last_closep = &last_close;
4895 minlen = study_chunk(pRExC_state, &scan, &minlen, &fake, scan + RExC_size,
4896 &data, -1, NULL, NULL, SCF_DO_STCLASS_AND|SCF_WHILEM_VISITED_POS,0);
4900 r->check_substr = r->check_utf8 = r->anchored_substr = r->anchored_utf8
4901 = r->float_substr = r->float_utf8 = NULL;
4902 if (!(data.start_class->flags & ANYOF_EOS)
4903 && !cl_is_anything(data.start_class))
4905 const U32 n = add_data(pRExC_state, 1, "f");
4907 Newx(RExC_rxi->data->data[n], 1,
4908 struct regnode_charclass_class);
4909 StructCopy(data.start_class,
4910 (struct regnode_charclass_class*)RExC_rxi->data->data[n],
4911 struct regnode_charclass_class);
4912 ri->regstclass = (regnode*)RExC_rxi->data->data[n];
4913 r->intflags &= ~PREGf_SKIP; /* Used in find_byclass(). */
4914 DEBUG_COMPILE_r({ SV* sv = sv_newmortal();
4915 regprop(r, sv, (regnode*)data.start_class);
4916 PerlIO_printf(Perl_debug_log,
4917 "synthetic stclass \"%s\".\n",
4918 SvPVX_const(sv));});
4922 /* Guard against an embedded (?=) or (?<=) with a longer minlen than
4923 the "real" pattern. */
4925 PerlIO_printf(Perl_debug_log,"minlen: %"IVdf" r->minlen:%"IVdf"\n",
4926 (IV)minlen, (IV)r->minlen);
4928 r->minlenret = minlen;
4929 if (r->minlen < minlen)
4932 if (RExC_seen & REG_SEEN_GPOS)
4933 r->extflags |= RXf_GPOS_SEEN;
4934 if (RExC_seen & REG_SEEN_LOOKBEHIND)
4935 r->extflags |= RXf_LOOKBEHIND_SEEN;
4936 if (RExC_seen & REG_SEEN_EVAL)
4937 r->extflags |= RXf_EVAL_SEEN;
4938 if (RExC_seen & REG_SEEN_CANY)
4939 r->extflags |= RXf_CANY_SEEN;
4940 if (RExC_seen & REG_SEEN_VERBARG)
4941 r->intflags |= PREGf_VERBARG_SEEN;
4942 if (RExC_seen & REG_SEEN_CUTGROUP)
4943 r->intflags |= PREGf_CUTGROUP_SEEN;
4944 if (RExC_paren_names)
4945 RXp_PAREN_NAMES(r) = MUTABLE_HV(SvREFCNT_inc(RExC_paren_names));
4947 RXp_PAREN_NAMES(r) = NULL;
4949 #ifdef STUPID_PATTERN_CHECKS
4950 if (RX_PRELEN(rx) == 0)
4951 r->extflags |= RXf_NULL;
4952 if (r->extflags & RXf_SPLIT && RX_PRELEN(rx) == 1 && RX_PRECOMP(rx)[0] == ' ')
4953 /* XXX: this should happen BEFORE we compile */
4954 r->extflags |= (RXf_SKIPWHITE|RXf_WHITE);
4955 else if (RX_PRELEN(rx) == 3 && memEQ("\\s+", RX_PRECOMP(rx), 3))
4956 r->extflags |= RXf_WHITE;
4957 else if (RX_PRELEN(rx) == 1 && RXp_PRECOMP(rx)[0] == '^')
4958 r->extflags |= RXf_START_ONLY;
4960 if (r->extflags & RXf_SPLIT && RX_PRELEN(rx) == 1 && RX_PRECOMP(rx)[0] == ' ')
4961 /* XXX: this should happen BEFORE we compile */
4962 r->extflags |= (RXf_SKIPWHITE|RXf_WHITE);
4964 regnode *first = ri->program + 1;
4966 U8 nop = OP(NEXTOPER(first));
4968 if (PL_regkind[fop] == NOTHING && nop == END)
4969 r->extflags |= RXf_NULL;
4970 else if (PL_regkind[fop] == BOL && nop == END)
4971 r->extflags |= RXf_START_ONLY;
4972 else if (fop == PLUS && nop ==SPACE && OP(regnext(first))==END)
4973 r->extflags |= RXf_WHITE;
4977 if (RExC_paren_names) {
4978 ri->name_list_idx = add_data( pRExC_state, 1, "a" );
4979 ri->data->data[ri->name_list_idx] = (void*)SvREFCNT_inc(RExC_paren_name_list);
4982 ri->name_list_idx = 0;
4984 if (RExC_recurse_count) {
4985 for ( ; RExC_recurse_count ; RExC_recurse_count-- ) {
4986 const regnode *scan = RExC_recurse[RExC_recurse_count-1];
4987 ARG2L_SET( scan, RExC_open_parens[ARG(scan)-1] - scan );
4990 Newxz(r->offs, RExC_npar, regexp_paren_pair);
4991 /* assume we don't need to swap parens around before we match */
4994 PerlIO_printf(Perl_debug_log,"Final program:\n");
4997 #ifdef RE_TRACK_PATTERN_OFFSETS
4998 DEBUG_OFFSETS_r(if (ri->u.offsets) {
4999 const U32 len = ri->u.offsets[0];
5001 GET_RE_DEBUG_FLAGS_DECL;
5002 PerlIO_printf(Perl_debug_log, "Offsets: [%"UVuf"]\n\t", (UV)ri->u.offsets[0]);
5003 for (i = 1; i <= len; i++) {
5004 if (ri->u.offsets[i*2-1] || ri->u.offsets[i*2])
5005 PerlIO_printf(Perl_debug_log, "%"UVuf":%"UVuf"[%"UVuf"] ",
5006 (UV)i, (UV)ri->u.offsets[i*2-1], (UV)ri->u.offsets[i*2]);
5008 PerlIO_printf(Perl_debug_log, "\n");
5014 #undef RE_ENGINE_PTR
5018 Perl_reg_named_buff(pTHX_ REGEXP * const rx, SV * const key, SV * const value,
5021 PERL_ARGS_ASSERT_REG_NAMED_BUFF;
5023 PERL_UNUSED_ARG(value);
5025 if (flags & RXapif_FETCH) {
5026 return reg_named_buff_fetch(rx, key, flags);
5027 } else if (flags & (RXapif_STORE | RXapif_DELETE | RXapif_CLEAR)) {
5028 Perl_croak_no_modify(aTHX);
5030 } else if (flags & RXapif_EXISTS) {
5031 return reg_named_buff_exists(rx, key, flags)
5034 } else if (flags & RXapif_REGNAMES) {
5035 return reg_named_buff_all(rx, flags);
5036 } else if (flags & (RXapif_SCALAR | RXapif_REGNAMES_COUNT)) {
5037 return reg_named_buff_scalar(rx, flags);
5039 Perl_croak(aTHX_ "panic: Unknown flags %d in named_buff", (int)flags);
5045 Perl_reg_named_buff_iter(pTHX_ REGEXP * const rx, const SV * const lastkey,
5048 PERL_ARGS_ASSERT_REG_NAMED_BUFF_ITER;
5049 PERL_UNUSED_ARG(lastkey);
5051 if (flags & RXapif_FIRSTKEY)
5052 return reg_named_buff_firstkey(rx, flags);
5053 else if (flags & RXapif_NEXTKEY)
5054 return reg_named_buff_nextkey(rx, flags);
5056 Perl_croak(aTHX_ "panic: Unknown flags %d in named_buff_iter", (int)flags);
5062 Perl_reg_named_buff_fetch(pTHX_ REGEXP * const r, SV * const namesv,
5065 AV *retarray = NULL;
5067 struct regexp *const rx = (struct regexp *)SvANY(r);
5069 PERL_ARGS_ASSERT_REG_NAMED_BUFF_FETCH;
5071 if (flags & RXapif_ALL)
5074 if (rx && RXp_PAREN_NAMES(rx)) {
5075 HE *he_str = hv_fetch_ent( RXp_PAREN_NAMES(rx), namesv, 0, 0 );
5078 SV* sv_dat=HeVAL(he_str);
5079 I32 *nums=(I32*)SvPVX(sv_dat);
5080 for ( i=0; i<SvIVX(sv_dat); i++ ) {
5081 if ((I32)(rx->nparens) >= nums[i]
5082 && rx->offs[nums[i]].start != -1
5083 && rx->offs[nums[i]].end != -1)
5086 CALLREG_NUMBUF_FETCH(r,nums[i],ret);
5090 ret = newSVsv(&PL_sv_undef);
5093 av_push(retarray, ret);
5096 return newRV_noinc(MUTABLE_SV(retarray));
5103 Perl_reg_named_buff_exists(pTHX_ REGEXP * const r, SV * const key,
5106 struct regexp *const rx = (struct regexp *)SvANY(r);
5108 PERL_ARGS_ASSERT_REG_NAMED_BUFF_EXISTS;
5110 if (rx && RXp_PAREN_NAMES(rx)) {
5111 if (flags & RXapif_ALL) {
5112 return hv_exists_ent(RXp_PAREN_NAMES(rx), key, 0);
5114 SV *sv = CALLREG_NAMED_BUFF_FETCH(r, key, flags);
5128 Perl_reg_named_buff_firstkey(pTHX_ REGEXP * const r, const U32 flags)
5130 struct regexp *const rx = (struct regexp *)SvANY(r);
5132 PERL_ARGS_ASSERT_REG_NAMED_BUFF_FIRSTKEY;
5134 if ( rx && RXp_PAREN_NAMES(rx) ) {
5135 (void)hv_iterinit(RXp_PAREN_NAMES(rx));
5137 return CALLREG_NAMED_BUFF_NEXTKEY(r, NULL, flags & ~RXapif_FIRSTKEY);
5144 Perl_reg_named_buff_nextkey(pTHX_ REGEXP * const r, const U32 flags)
5146 struct regexp *const rx = (struct regexp *)SvANY(r);
5147 GET_RE_DEBUG_FLAGS_DECL;
5149 PERL_ARGS_ASSERT_REG_NAMED_BUFF_NEXTKEY;
5151 if (rx && RXp_PAREN_NAMES(rx)) {
5152 HV *hv = RXp_PAREN_NAMES(rx);
5154 while ( (temphe = hv_iternext_flags(hv,0)) ) {
5157 SV* sv_dat = HeVAL(temphe);
5158 I32 *nums = (I32*)SvPVX(sv_dat);
5159 for ( i = 0; i < SvIVX(sv_dat); i++ ) {
5160 if ((I32)(rx->lastparen) >= nums[i] &&
5161 rx->offs[nums[i]].start != -1 &&
5162 rx->offs[nums[i]].end != -1)
5168 if (parno || flags & RXapif_ALL) {
5169 return newSVhek(HeKEY_hek(temphe));
5177 Perl_reg_named_buff_scalar(pTHX_ REGEXP * const r, const U32 flags)
5182 struct regexp *const rx = (struct regexp *)SvANY(r);
5184 PERL_ARGS_ASSERT_REG_NAMED_BUFF_SCALAR;
5186 if (rx && RXp_PAREN_NAMES(rx)) {
5187 if (flags & (RXapif_ALL | RXapif_REGNAMES_COUNT)) {
5188 return newSViv(HvTOTALKEYS(RXp_PAREN_NAMES(rx)));
5189 } else if (flags & RXapif_ONE) {
5190 ret = CALLREG_NAMED_BUFF_ALL(r, (flags | RXapif_REGNAMES));
5191 av = MUTABLE_AV(SvRV(ret));
5192 length = av_len(av);
5194 return newSViv(length + 1);
5196 Perl_croak(aTHX_ "panic: Unknown flags %d in named_buff_scalar", (int)flags);
5200 return &PL_sv_undef;
5204 Perl_reg_named_buff_all(pTHX_ REGEXP * const r, const U32 flags)
5206 struct regexp *const rx = (struct regexp *)SvANY(r);
5209 PERL_ARGS_ASSERT_REG_NAMED_BUFF_ALL;
5211 if (rx && RXp_PAREN_NAMES(rx)) {
5212 HV *hv= RXp_PAREN_NAMES(rx);
5214 (void)hv_iterinit(hv);
5215 while ( (temphe = hv_iternext_flags(hv,0)) ) {
5218 SV* sv_dat = HeVAL(temphe);
5219 I32 *nums = (I32*)SvPVX(sv_dat);
5220 for ( i = 0; i < SvIVX(sv_dat); i++ ) {
5221 if ((I32)(rx->lastparen) >= nums[i] &&
5222 rx->offs[nums[i]].start != -1 &&
5223 rx->offs[nums[i]].end != -1)
5229 if (parno || flags & RXapif_ALL) {
5230 av_push(av, newSVhek(HeKEY_hek(temphe)));
5235 return newRV_noinc(MUTABLE_SV(av));
5239 Perl_reg_numbered_buff_fetch(pTHX_ REGEXP * const r, const I32 paren,
5242 struct regexp *const rx = (struct regexp *)SvANY(r);
5247 PERL_ARGS_ASSERT_REG_NUMBERED_BUFF_FETCH;
5250 sv_setsv(sv,&PL_sv_undef);
5254 if (paren == RX_BUFF_IDX_PREMATCH && rx->offs[0].start != -1) {
5256 i = rx->offs[0].start;
5260 if (paren == RX_BUFF_IDX_POSTMATCH && rx->offs[0].end != -1) {
5262 s = rx->subbeg + rx->offs[0].end;
5263 i = rx->sublen - rx->offs[0].end;
5266 if ( 0 <= paren && paren <= (I32)rx->nparens &&
5267 (s1 = rx->offs[paren].start) != -1 &&
5268 (t1 = rx->offs[paren].end) != -1)
5272 s = rx->subbeg + s1;
5274 sv_setsv(sv,&PL_sv_undef);
5277 assert(rx->sublen >= (s - rx->subbeg) + i );
5279 const int oldtainted = PL_tainted;
5281 sv_setpvn(sv, s, i);
5282 PL_tainted = oldtainted;
5283 if ( (rx->extflags & RXf_CANY_SEEN)
5284 ? (RXp_MATCH_UTF8(rx)
5285 && (!i || is_utf8_string((U8*)s, i)))
5286 : (RXp_MATCH_UTF8(rx)) )
5293 if (RXp_MATCH_TAINTED(rx)) {
5294 if (SvTYPE(sv) >= SVt_PVMG) {
5295 MAGIC* const mg = SvMAGIC(sv);
5298 SvMAGIC_set(sv, mg->mg_moremagic);
5300 if ((mgt = SvMAGIC(sv))) {
5301 mg->mg_moremagic = mgt;
5302 SvMAGIC_set(sv, mg);
5312 sv_setsv(sv,&PL_sv_undef);
5318 Perl_reg_numbered_buff_store(pTHX_ REGEXP * const rx, const I32 paren,
5319 SV const * const value)
5321 PERL_ARGS_ASSERT_REG_NUMBERED_BUFF_STORE;
5323 PERL_UNUSED_ARG(rx);
5324 PERL_UNUSED_ARG(paren);
5325 PERL_UNUSED_ARG(value);
5328 Perl_croak_no_modify(aTHX);
5332 Perl_reg_numbered_buff_length(pTHX_ REGEXP * const r, const SV * const sv,
5335 struct regexp *const rx = (struct regexp *)SvANY(r);
5339 PERL_ARGS_ASSERT_REG_NUMBERED_BUFF_LENGTH;
5341 /* Some of this code was originally in C<Perl_magic_len> in F<mg.c> */
5343 /* $` / ${^PREMATCH} */
5344 case RX_BUFF_IDX_PREMATCH:
5345 if (rx->offs[0].start != -1) {
5346 i = rx->offs[0].start;
5354 /* $' / ${^POSTMATCH} */
5355 case RX_BUFF_IDX_POSTMATCH:
5356 if (rx->offs[0].end != -1) {
5357 i = rx->sublen - rx->offs[0].end;
5359 s1 = rx->offs[0].end;
5365 /* $& / ${^MATCH}, $1, $2, ... */
5367 if (paren <= (I32)rx->nparens &&
5368 (s1 = rx->offs[paren].start) != -1 &&
5369 (t1 = rx->offs[paren].end) != -1)
5374 if (ckWARN(WARN_UNINITIALIZED))
5375 report_uninit((const SV *)sv);
5380 if (i > 0 && RXp_MATCH_UTF8(rx)) {
5381 const char * const s = rx->subbeg + s1;
5386 if (is_utf8_string_loclen((U8*)s, i, &ep, &el))
5393 Perl_reg_qr_package(pTHX_ REGEXP * const rx)
5395 PERL_ARGS_ASSERT_REG_QR_PACKAGE;
5396 PERL_UNUSED_ARG(rx);
5400 return newSVpvs("Regexp");
5403 /* Scans the name of a named buffer from the pattern.
5404 * If flags is REG_RSN_RETURN_NULL returns null.
5405 * If flags is REG_RSN_RETURN_NAME returns an SV* containing the name
5406 * If flags is REG_RSN_RETURN_DATA returns the data SV* corresponding
5407 * to the parsed name as looked up in the RExC_paren_names hash.
5408 * If there is an error throws a vFAIL().. type exception.
5411 #define REG_RSN_RETURN_NULL 0
5412 #define REG_RSN_RETURN_NAME 1
5413 #define REG_RSN_RETURN_DATA 2
5416 S_reg_scan_name(pTHX_ RExC_state_t *pRExC_state, U32 flags)
5418 char *name_start = RExC_parse;
5420 PERL_ARGS_ASSERT_REG_SCAN_NAME;
5422 if (isIDFIRST_lazy_if(RExC_parse, UTF)) {
5423 /* skip IDFIRST by using do...while */
5426 RExC_parse += UTF8SKIP(RExC_parse);
5427 } while (isALNUM_utf8((U8*)RExC_parse));
5431 } while (isALNUM(*RExC_parse));
5436 = newSVpvn_flags(name_start, (int)(RExC_parse - name_start),
5437 SVs_TEMP | (UTF ? SVf_UTF8 : 0));
5438 if ( flags == REG_RSN_RETURN_NAME)
5440 else if (flags==REG_RSN_RETURN_DATA) {
5443 if ( ! sv_name ) /* should not happen*/
5444 Perl_croak(aTHX_ "panic: no svname in reg_scan_name");
5445 if (RExC_paren_names)
5446 he_str = hv_fetch_ent( RExC_paren_names, sv_name, 0, 0 );
5448 sv_dat = HeVAL(he_str);
5450 vFAIL("Reference to nonexistent named group");
5454 Perl_croak(aTHX_ "panic: bad flag in reg_scan_name");
5461 #define DEBUG_PARSE_MSG(funcname) DEBUG_PARSE_r({ \
5462 int rem=(int)(RExC_end - RExC_parse); \
5471 if (RExC_lastparse!=RExC_parse) \
5472 PerlIO_printf(Perl_debug_log," >%.*s%-*s", \
5475 iscut ? "..." : "<" \
5478 PerlIO_printf(Perl_debug_log,"%16s",""); \
5481 num = RExC_size + 1; \
5483 num=REG_NODE_NUM(RExC_emit); \
5484 if (RExC_lastnum!=num) \
5485 PerlIO_printf(Perl_debug_log,"|%4d",num); \
5487 PerlIO_printf(Perl_debug_log,"|%4s",""); \
5488 PerlIO_printf(Perl_debug_log,"|%*s%-4s", \
5489 (int)((depth*2)), "", \
5493 RExC_lastparse=RExC_parse; \
5498 #define DEBUG_PARSE(funcname) DEBUG_PARSE_r({ \
5499 DEBUG_PARSE_MSG((funcname)); \
5500 PerlIO_printf(Perl_debug_log,"%4s","\n"); \
5502 #define DEBUG_PARSE_FMT(funcname,fmt,args) DEBUG_PARSE_r({ \
5503 DEBUG_PARSE_MSG((funcname)); \
5504 PerlIO_printf(Perl_debug_log,fmt "\n",args); \
5507 - reg - regular expression, i.e. main body or parenthesized thing
5509 * Caller must absorb opening parenthesis.
5511 * Combining parenthesis handling with the base level of regular expression
5512 * is a trifle forced, but the need to tie the tails of the branches to what
5513 * follows makes it hard to avoid.
5515 #define REGTAIL(x,y,z) regtail((x),(y),(z),depth+1)
5517 #define REGTAIL_STUDY(x,y,z) regtail_study((x),(y),(z),depth+1)
5519 #define REGTAIL_STUDY(x,y,z) regtail((x),(y),(z),depth+1)
5523 S_reg(pTHX_ RExC_state_t *pRExC_state, I32 paren, I32 *flagp,U32 depth)
5524 /* paren: Parenthesized? 0=top, 1=(, inside: changed to letter. */
5527 register regnode *ret; /* Will be the head of the group. */
5528 register regnode *br;
5529 register regnode *lastbr;
5530 register regnode *ender = NULL;
5531 register I32 parno = 0;
5533 U32 oregflags = RExC_flags;
5534 bool have_branch = 0;
5536 I32 freeze_paren = 0;
5537 I32 after_freeze = 0;
5539 /* for (?g), (?gc), and (?o) warnings; warning
5540 about (?c) will warn about (?g) -- japhy */
5542 #define WASTED_O 0x01
5543 #define WASTED_G 0x02
5544 #define WASTED_C 0x04
5545 #define WASTED_GC (0x02|0x04)
5546 I32 wastedflags = 0x00;
5548 char * parse_start = RExC_parse; /* MJD */
5549 char * const oregcomp_parse = RExC_parse;
5551 GET_RE_DEBUG_FLAGS_DECL;
5553 PERL_ARGS_ASSERT_REG;
5554 DEBUG_PARSE("reg ");
5556 *flagp = 0; /* Tentatively. */
5559 /* Make an OPEN node, if parenthesized. */
5561 if ( *RExC_parse == '*') { /* (*VERB:ARG) */
5562 char *start_verb = RExC_parse;
5563 STRLEN verb_len = 0;
5564 char *start_arg = NULL;
5565 unsigned char op = 0;
5567 int internal_argval = 0; /* internal_argval is only useful if !argok */
5568 while ( *RExC_parse && *RExC_parse != ')' ) {
5569 if ( *RExC_parse == ':' ) {
5570 start_arg = RExC_parse + 1;
5576 verb_len = RExC_parse - start_verb;
5579 while ( *RExC_parse && *RExC_parse != ')' )
5581 if ( *RExC_parse != ')' )
5582 vFAIL("Unterminated verb pattern argument");
5583 if ( RExC_parse == start_arg )
5586 if ( *RExC_parse != ')' )
5587 vFAIL("Unterminated verb pattern");
5590 switch ( *start_verb ) {
5591 case 'A': /* (*ACCEPT) */
5592 if ( memEQs(start_verb,verb_len,"ACCEPT") ) {
5594 internal_argval = RExC_nestroot;
5597 case 'C': /* (*COMMIT) */
5598 if ( memEQs(start_verb,verb_len,"COMMIT") )
5601 case 'F': /* (*FAIL) */
5602 if ( verb_len==1 || memEQs(start_verb,verb_len,"FAIL") ) {
5607 case ':': /* (*:NAME) */
5608 case 'M': /* (*MARK:NAME) */
5609 if ( verb_len==0 || memEQs(start_verb,verb_len,"MARK") ) {
5614 case 'P': /* (*PRUNE) */
5615 if ( memEQs(start_verb,verb_len,"PRUNE") )
5618 case 'S': /* (*SKIP) */
5619 if ( memEQs(start_verb,verb_len,"SKIP") )
5622 case 'T': /* (*THEN) */
5623 /* [19:06] <TimToady> :: is then */
5624 if ( memEQs(start_verb,verb_len,"THEN") ) {
5626 RExC_seen |= REG_SEEN_CUTGROUP;
5632 vFAIL3("Unknown verb pattern '%.*s'",
5633 verb_len, start_verb);
5636 if ( start_arg && internal_argval ) {
5637 vFAIL3("Verb pattern '%.*s' may not have an argument",
5638 verb_len, start_verb);
5639 } else if ( argok < 0 && !start_arg ) {
5640 vFAIL3("Verb pattern '%.*s' has a mandatory argument",
5641 verb_len, start_verb);
5643 ret = reganode(pRExC_state, op, internal_argval);
5644 if ( ! internal_argval && ! SIZE_ONLY ) {
5646 SV *sv = newSVpvn( start_arg, RExC_parse - start_arg);
5647 ARG(ret) = add_data( pRExC_state, 1, "S" );
5648 RExC_rxi->data->data[ARG(ret)]=(void*)sv;
5655 if (!internal_argval)
5656 RExC_seen |= REG_SEEN_VERBARG;
5657 } else if ( start_arg ) {
5658 vFAIL3("Verb pattern '%.*s' may not have an argument",
5659 verb_len, start_verb);
5661 ret = reg_node(pRExC_state, op);
5663 nextchar(pRExC_state);
5666 if (*RExC_parse == '?') { /* (?...) */
5667 bool is_logical = 0;
5668 const char * const seqstart = RExC_parse;
5671 paren = *RExC_parse++;
5672 ret = NULL; /* For look-ahead/behind. */
5675 case 'P': /* (?P...) variants for those used to PCRE/Python */
5676 paren = *RExC_parse++;
5677 if ( paren == '<') /* (?P<...>) named capture */
5679 else if (paren == '>') { /* (?P>name) named recursion */
5680 goto named_recursion;
5682 else if (paren == '=') { /* (?P=...) named backref */
5683 /* this pretty much dupes the code for \k<NAME> in regatom(), if
5684 you change this make sure you change that */
5685 char* name_start = RExC_parse;
5687 SV *sv_dat = reg_scan_name(pRExC_state,
5688 SIZE_ONLY ? REG_RSN_RETURN_NULL : REG_RSN_RETURN_DATA);
5689 if (RExC_parse == name_start || *RExC_parse != ')')
5690 vFAIL2("Sequence %.3s... not terminated",parse_start);
5693 num = add_data( pRExC_state, 1, "S" );
5694 RExC_rxi->data->data[num]=(void*)sv_dat;
5695 SvREFCNT_inc_simple_void(sv_dat);
5698 ret = reganode(pRExC_state,
5699 (U8)(FOLD ? (LOC ? NREFFL : NREFF) : NREF),
5703 Set_Node_Offset(ret, parse_start+1);
5704 Set_Node_Cur_Length(ret); /* MJD */
5706 nextchar(pRExC_state);
5710 vFAIL3("Sequence (%.*s...) not recognized", RExC_parse-seqstart, seqstart);
5712 case '<': /* (?<...) */
5713 if (*RExC_parse == '!')
5715 else if (*RExC_parse != '=')
5721 case '\'': /* (?'...') */
5722 name_start= RExC_parse;
5723 svname = reg_scan_name(pRExC_state,
5724 SIZE_ONLY ? /* reverse test from the others */
5725 REG_RSN_RETURN_NAME :
5726 REG_RSN_RETURN_NULL);
5727 if (RExC_parse == name_start) {
5729 vFAIL3("Sequence (%.*s...) not recognized", RExC_parse-seqstart, seqstart);
5732 if (*RExC_parse != paren)
5733 vFAIL2("Sequence (?%c... not terminated",
5734 paren=='>' ? '<' : paren);
5738 if (!svname) /* shouldnt happen */
5740 "panic: reg_scan_name returned NULL");
5741 if (!RExC_paren_names) {
5742 RExC_paren_names= newHV();
5743 sv_2mortal(MUTABLE_SV(RExC_paren_names));
5745 RExC_paren_name_list= newAV();
5746 sv_2mortal(MUTABLE_SV(RExC_paren_name_list));
5749 he_str = hv_fetch_ent( RExC_paren_names, svname, 1, 0 );
5751 sv_dat = HeVAL(he_str);
5753 /* croak baby croak */
5755 "panic: paren_name hash element allocation failed");
5756 } else if ( SvPOK(sv_dat) ) {
5757 /* (?|...) can mean we have dupes so scan to check
5758 its already been stored. Maybe a flag indicating
5759 we are inside such a construct would be useful,
5760 but the arrays are likely to be quite small, so
5761 for now we punt -- dmq */
5762 IV count = SvIV(sv_dat);
5763 I32 *pv = (I32*)SvPVX(sv_dat);
5765 for ( i = 0 ; i < count ; i++ ) {
5766 if ( pv[i] == RExC_npar ) {
5772 pv = (I32*)SvGROW(sv_dat, SvCUR(sv_dat) + sizeof(I32)+1);
5773 SvCUR_set(sv_dat, SvCUR(sv_dat) + sizeof(I32));
5774 pv[count] = RExC_npar;
5775 SvIV_set(sv_dat, SvIVX(sv_dat) + 1);
5778 (void)SvUPGRADE(sv_dat,SVt_PVNV);
5779 sv_setpvn(sv_dat, (char *)&(RExC_npar), sizeof(I32));
5781 SvIV_set(sv_dat, 1);
5784 if (!av_store(RExC_paren_name_list, RExC_npar, SvREFCNT_inc(svname)))
5785 SvREFCNT_dec(svname);
5788 /*sv_dump(sv_dat);*/
5790 nextchar(pRExC_state);
5792 goto capturing_parens;
5794 RExC_seen |= REG_SEEN_LOOKBEHIND;
5796 case '=': /* (?=...) */
5797 RExC_seen_zerolen++;
5799 case '!': /* (?!...) */
5800 RExC_seen_zerolen++;
5801 if (*RExC_parse == ')') {
5802 ret=reg_node(pRExC_state, OPFAIL);
5803 nextchar(pRExC_state);
5807 case '|': /* (?|...) */
5808 /* branch reset, behave like a (?:...) except that
5809 buffers in alternations share the same numbers */
5811 after_freeze = freeze_paren = RExC_npar;
5813 case ':': /* (?:...) */
5814 case '>': /* (?>...) */
5816 case '$': /* (?$...) */
5817 case '@': /* (?@...) */
5818 vFAIL2("Sequence (?%c...) not implemented", (int)paren);
5820 case '#': /* (?#...) */
5821 while (*RExC_parse && *RExC_parse != ')')
5823 if (*RExC_parse != ')')
5824 FAIL("Sequence (?#... not terminated");
5825 nextchar(pRExC_state);
5828 case '0' : /* (?0) */
5829 case 'R' : /* (?R) */
5830 if (*RExC_parse != ')')
5831 FAIL("Sequence (?R) not terminated");
5832 ret = reg_node(pRExC_state, GOSTART);
5833 *flagp |= POSTPONED;
5834 nextchar(pRExC_state);
5837 { /* named and numeric backreferences */
5839 case '&': /* (?&NAME) */
5840 parse_start = RExC_parse - 1;
5843 SV *sv_dat = reg_scan_name(pRExC_state,
5844 SIZE_ONLY ? REG_RSN_RETURN_NULL : REG_RSN_RETURN_DATA);
5845 num = sv_dat ? *((I32 *)SvPVX(sv_dat)) : 0;
5847 goto gen_recurse_regop;
5850 if (!(RExC_parse[0] >= '1' && RExC_parse[0] <= '9')) {
5852 vFAIL("Illegal pattern");
5854 goto parse_recursion;
5856 case '-': /* (?-1) */
5857 if (!(RExC_parse[0] >= '1' && RExC_parse[0] <= '9')) {
5858 RExC_parse--; /* rewind to let it be handled later */
5862 case '1': case '2': case '3': case '4': /* (?1) */
5863 case '5': case '6': case '7': case '8': case '9':
5866 num = atoi(RExC_parse);
5867 parse_start = RExC_parse - 1; /* MJD */
5868 if (*RExC_parse == '-')
5870 while (isDIGIT(*RExC_parse))
5872 if (*RExC_parse!=')')
5873 vFAIL("Expecting close bracket");
5876 if ( paren == '-' ) {
5878 Diagram of capture buffer numbering.
5879 Top line is the normal capture buffer numbers
5880 Botton line is the negative indexing as from
5884 /(a(x)y)(a(b(c(?-2)d)e)f)(g(h))/
5888 num = RExC_npar + num;
5891 vFAIL("Reference to nonexistent group");
5893 } else if ( paren == '+' ) {
5894 num = RExC_npar + num - 1;
5897 ret = reganode(pRExC_state, GOSUB, num);
5899 if (num > (I32)RExC_rx->nparens) {
5901 vFAIL("Reference to nonexistent group");
5903 ARG2L_SET( ret, RExC_recurse_count++);
5905 DEBUG_OPTIMISE_MORE_r(PerlIO_printf(Perl_debug_log,
5906 "Recurse #%"UVuf" to %"IVdf"\n", (UV)ARG(ret), (IV)ARG2L(ret)));
5910 RExC_seen |= REG_SEEN_RECURSE;
5911 Set_Node_Length(ret, 1 + regarglen[OP(ret)]); /* MJD */
5912 Set_Node_Offset(ret, parse_start); /* MJD */
5914 *flagp |= POSTPONED;
5915 nextchar(pRExC_state);
5917 } /* named and numeric backreferences */
5920 case '?': /* (??...) */
5922 if (*RExC_parse != '{') {
5924 vFAIL3("Sequence (%.*s...) not recognized", RExC_parse-seqstart, seqstart);
5927 *flagp |= POSTPONED;
5928 paren = *RExC_parse++;
5930 case '{': /* (?{...}) */
5935 char *s = RExC_parse;
5937 RExC_seen_zerolen++;
5938 RExC_seen |= REG_SEEN_EVAL;
5939 while (count && (c = *RExC_parse)) {
5950 if (*RExC_parse != ')') {
5952 vFAIL("Sequence (?{...}) not terminated or not {}-balanced");
5956 OP_4tree *sop, *rop;
5957 SV * const sv = newSVpvn(s, RExC_parse - 1 - s);
5960 Perl_save_re_context(aTHX);
5961 rop = sv_compile_2op(sv, &sop, "re", &pad);
5962 sop->op_private |= OPpREFCOUNTED;
5963 /* re_dup will OpREFCNT_inc */
5964 OpREFCNT_set(sop, 1);
5967 n = add_data(pRExC_state, 3, "nop");
5968 RExC_rxi->data->data[n] = (void*)rop;
5969 RExC_rxi->data->data[n+1] = (void*)sop;
5970 RExC_rxi->data->data[n+2] = (void*)pad;
5973 else { /* First pass */
5974 if (PL_reginterp_cnt < ++RExC_seen_evals
5976 /* No compiled RE interpolated, has runtime
5977 components ===> unsafe. */
5978 FAIL("Eval-group not allowed at runtime, use re 'eval'");
5979 if (PL_tainting && PL_tainted)
5980 FAIL("Eval-group in insecure regular expression");
5981 #if PERL_VERSION > 8
5982 if (IN_PERL_COMPILETIME)
5987 nextchar(pRExC_state);
5989 ret = reg_node(pRExC_state, LOGICAL);
5992 REGTAIL(pRExC_state, ret, reganode(pRExC_state, EVAL, n));
5993 /* deal with the length of this later - MJD */
5996 ret = reganode(pRExC_state, EVAL, n);
5997 Set_Node_Length(ret, RExC_parse - parse_start + 1);
5998 Set_Node_Offset(ret, parse_start);
6001 case '(': /* (?(?{...})...) and (?(?=...)...) */
6004 if (RExC_parse[0] == '?') { /* (?(?...)) */
6005 if (RExC_parse[1] == '=' || RExC_parse[1] == '!'
6006 || RExC_parse[1] == '<'
6007 || RExC_parse[1] == '{') { /* Lookahead or eval. */
6010 ret = reg_node(pRExC_state, LOGICAL);
6013 REGTAIL(pRExC_state, ret, reg(pRExC_state, 1, &flag,depth+1));
6017 else if ( RExC_parse[0] == '<' /* (?(<NAME>)...) */
6018 || RExC_parse[0] == '\'' ) /* (?('NAME')...) */
6020 char ch = RExC_parse[0] == '<' ? '>' : '\'';
6021 char *name_start= RExC_parse++;
6023 SV *sv_dat=reg_scan_name(pRExC_state,
6024 SIZE_ONLY ? REG_RSN_RETURN_NULL : REG_RSN_RETURN_DATA);
6025 if (RExC_parse == name_start || *RExC_parse != ch)
6026 vFAIL2("Sequence (?(%c... not terminated",
6027 (ch == '>' ? '<' : ch));
6030 num = add_data( pRExC_state, 1, "S" );
6031 RExC_rxi->data->data[num]=(void*)sv_dat;
6032 SvREFCNT_inc_simple_void(sv_dat);
6034 ret = reganode(pRExC_state,NGROUPP,num);
6035 goto insert_if_check_paren;
6037 else if (RExC_parse[0] == 'D' &&
6038 RExC_parse[1] == 'E' &&
6039 RExC_parse[2] == 'F' &&
6040 RExC_parse[3] == 'I' &&
6041 RExC_parse[4] == 'N' &&
6042 RExC_parse[5] == 'E')
6044 ret = reganode(pRExC_state,DEFINEP,0);
6047 goto insert_if_check_paren;
6049 else if (RExC_parse[0] == 'R') {
6052 if (RExC_parse[0] >= '1' && RExC_parse[0] <= '9' ) {
6053 parno = atoi(RExC_parse++);
6054 while (isDIGIT(*RExC_parse))
6056 } else if (RExC_parse[0] == '&') {
6059 sv_dat = reg_scan_name(pRExC_state,
6060 SIZE_ONLY ? REG_RSN_RETURN_NULL : REG_RSN_RETURN_DATA);
6061 parno = sv_dat ? *((I32 *)SvPVX(sv_dat)) : 0;
6063 ret = reganode(pRExC_state,INSUBP,parno);
6064 goto insert_if_check_paren;
6066 else if (RExC_parse[0] >= '1' && RExC_parse[0] <= '9' ) {
6069 parno = atoi(RExC_parse++);
6071 while (isDIGIT(*RExC_parse))
6073 ret = reganode(pRExC_state, GROUPP, parno);
6075 insert_if_check_paren:
6076 if ((c = *nextchar(pRExC_state)) != ')')
6077 vFAIL("Switch condition not recognized");
6079 REGTAIL(pRExC_state, ret, reganode(pRExC_state, IFTHEN, 0));
6080 br = regbranch(pRExC_state, &flags, 1,depth+1);
6082 br = reganode(pRExC_state, LONGJMP, 0);
6084 REGTAIL(pRExC_state, br, reganode(pRExC_state, LONGJMP, 0));
6085 c = *nextchar(pRExC_state);
6090 vFAIL("(?(DEFINE)....) does not allow branches");
6091 lastbr = reganode(pRExC_state, IFTHEN, 0); /* Fake one for optimizer. */
6092 regbranch(pRExC_state, &flags, 1,depth+1);
6093 REGTAIL(pRExC_state, ret, lastbr);
6096 c = *nextchar(pRExC_state);
6101 vFAIL("Switch (?(condition)... contains too many branches");
6102 ender = reg_node(pRExC_state, TAIL);
6103 REGTAIL(pRExC_state, br, ender);
6105 REGTAIL(pRExC_state, lastbr, ender);
6106 REGTAIL(pRExC_state, NEXTOPER(NEXTOPER(lastbr)), ender);
6109 REGTAIL(pRExC_state, ret, ender);
6110 RExC_size++; /* XXX WHY do we need this?!!
6111 For large programs it seems to be required
6112 but I can't figure out why. -- dmq*/
6116 vFAIL2("Unknown switch condition (?(%.2s", RExC_parse);
6120 RExC_parse--; /* for vFAIL to print correctly */
6121 vFAIL("Sequence (? incomplete");
6125 parse_flags: /* (?i) */
6127 U32 posflags = 0, negflags = 0;
6128 U32 *flagsp = &posflags;
6130 while (*RExC_parse) {
6131 /* && strchr("iogcmsx", *RExC_parse) */
6132 /* (?g), (?gc) and (?o) are useless here
6133 and must be globally applied -- japhy */
6134 switch (*RExC_parse) {
6135 CASE_STD_PMMOD_FLAGS_PARSE_SET(flagsp);
6136 case ONCE_PAT_MOD: /* 'o' */
6137 case GLOBAL_PAT_MOD: /* 'g' */
6138 if (SIZE_ONLY && ckWARN(WARN_REGEXP)) {
6139 const I32 wflagbit = *RExC_parse == 'o' ? WASTED_O : WASTED_G;
6140 if (! (wastedflags & wflagbit) ) {
6141 wastedflags |= wflagbit;
6144 "Useless (%s%c) - %suse /%c modifier",
6145 flagsp == &negflags ? "?-" : "?",
6147 flagsp == &negflags ? "don't " : "",
6154 case CONTINUE_PAT_MOD: /* 'c' */
6155 if (SIZE_ONLY && ckWARN(WARN_REGEXP)) {
6156 if (! (wastedflags & WASTED_C) ) {
6157 wastedflags |= WASTED_GC;
6160 "Useless (%sc) - %suse /gc modifier",
6161 flagsp == &negflags ? "?-" : "?",
6162 flagsp == &negflags ? "don't " : ""
6167 case KEEPCOPY_PAT_MOD: /* 'p' */
6168 if (flagsp == &negflags) {
6170 ckWARNreg(RExC_parse + 1,"Useless use of (?-p)");
6172 *flagsp |= RXf_PMf_KEEPCOPY;
6176 if (flagsp == &negflags) {
6178 vFAIL3("Sequence (%.*s...) not recognized", RExC_parse-seqstart, seqstart);
6182 wastedflags = 0; /* reset so (?g-c) warns twice */
6188 RExC_flags |= posflags;
6189 RExC_flags &= ~negflags;
6191 oregflags |= posflags;
6192 oregflags &= ~negflags;
6194 nextchar(pRExC_state);
6205 vFAIL3("Sequence (%.*s...) not recognized", RExC_parse-seqstart, seqstart);
6210 }} /* one for the default block, one for the switch */
6217 ret = reganode(pRExC_state, OPEN, parno);
6220 RExC_nestroot = parno;
6221 if (RExC_seen & REG_SEEN_RECURSE
6222 && !RExC_open_parens[parno-1])
6224 DEBUG_OPTIMISE_MORE_r(PerlIO_printf(Perl_debug_log,
6225 "Setting open paren #%"IVdf" to %d\n",
6226 (IV)parno, REG_NODE_NUM(ret)));
6227 RExC_open_parens[parno-1]= ret;
6230 Set_Node_Length(ret, 1); /* MJD */
6231 Set_Node_Offset(ret, RExC_parse); /* MJD */
6239 /* Pick up the branches, linking them together. */
6240 parse_start = RExC_parse; /* MJD */
6241 br = regbranch(pRExC_state, &flags, 1,depth+1);
6244 if (RExC_npar > after_freeze)
6245 after_freeze = RExC_npar;
6246 RExC_npar = freeze_paren;
6249 /* branch_len = (paren != 0); */
6253 if (*RExC_parse == '|') {
6254 if (!SIZE_ONLY && RExC_extralen) {
6255 reginsert(pRExC_state, BRANCHJ, br, depth+1);
6258 reginsert(pRExC_state, BRANCH, br, depth+1);
6259 Set_Node_Length(br, paren != 0);
6260 Set_Node_Offset_To_R(br-RExC_emit_start, parse_start-RExC_start);
6264 RExC_extralen += 1; /* For BRANCHJ-BRANCH. */
6266 else if (paren == ':') {
6267 *flagp |= flags&SIMPLE;
6269 if (is_open) { /* Starts with OPEN. */
6270 REGTAIL(pRExC_state, ret, br); /* OPEN -> first. */
6272 else if (paren != '?') /* Not Conditional */
6274 *flagp |= flags & (SPSTART | HASWIDTH | POSTPONED);
6276 while (*RExC_parse == '|') {
6277 if (!SIZE_ONLY && RExC_extralen) {
6278 ender = reganode(pRExC_state, LONGJMP,0);
6279 REGTAIL(pRExC_state, NEXTOPER(NEXTOPER(lastbr)), ender); /* Append to the previous. */
6282 RExC_extralen += 2; /* Account for LONGJMP. */
6283 nextchar(pRExC_state);
6285 if (RExC_npar > after_freeze)
6286 after_freeze = RExC_npar;
6287 RExC_npar = freeze_paren;
6289 br = regbranch(pRExC_state, &flags, 0, depth+1);
6293 REGTAIL(pRExC_state, lastbr, br); /* BRANCH -> BRANCH. */
6295 *flagp |= flags & (SPSTART | HASWIDTH | POSTPONED);
6298 if (have_branch || paren != ':') {
6299 /* Make a closing node, and hook it on the end. */
6302 ender = reg_node(pRExC_state, TAIL);
6305 ender = reganode(pRExC_state, CLOSE, parno);
6306 if (!SIZE_ONLY && RExC_seen & REG_SEEN_RECURSE) {
6307 DEBUG_OPTIMISE_MORE_r(PerlIO_printf(Perl_debug_log,
6308 "Setting close paren #%"IVdf" to %d\n",
6309 (IV)parno, REG_NODE_NUM(ender)));
6310 RExC_close_parens[parno-1]= ender;
6311 if (RExC_nestroot == parno)
6314 Set_Node_Offset(ender,RExC_parse+1); /* MJD */
6315 Set_Node_Length(ender,1); /* MJD */
6321 *flagp &= ~HASWIDTH;
6324 ender = reg_node(pRExC_state, SUCCEED);
6327 ender = reg_node(pRExC_state, END);
6329 assert(!RExC_opend); /* there can only be one! */
6334 REGTAIL(pRExC_state, lastbr, ender);
6336 if (have_branch && !SIZE_ONLY) {
6338 RExC_seen |= REG_TOP_LEVEL_BRANCHES;
6340 /* Hook the tails of the branches to the closing node. */
6341 for (br = ret; br; br = regnext(br)) {
6342 const U8 op = PL_regkind[OP(br)];
6344 REGTAIL_STUDY(pRExC_state, NEXTOPER(br), ender);
6346 else if (op == BRANCHJ) {
6347 REGTAIL_STUDY(pRExC_state, NEXTOPER(NEXTOPER(br)), ender);
6355 static const char parens[] = "=!<,>";
6357 if (paren && (p = strchr(parens, paren))) {
6358 U8 node = ((p - parens) % 2) ? UNLESSM : IFMATCH;
6359 int flag = (p - parens) > 1;
6362 node = SUSPEND, flag = 0;
6363 reginsert(pRExC_state, node,ret, depth+1);
6364 Set_Node_Cur_Length(ret);
6365 Set_Node_Offset(ret, parse_start + 1);
6367 REGTAIL_STUDY(pRExC_state, ret, reg_node(pRExC_state, TAIL));
6371 /* Check for proper termination. */
6373 RExC_flags = oregflags;
6374 if (RExC_parse >= RExC_end || *nextchar(pRExC_state) != ')') {
6375 RExC_parse = oregcomp_parse;
6376 vFAIL("Unmatched (");
6379 else if (!paren && RExC_parse < RExC_end) {
6380 if (*RExC_parse == ')') {
6382 vFAIL("Unmatched )");
6385 FAIL("Junk on end of regexp"); /* "Can't happen". */
6389 RExC_npar = after_freeze;
6394 - regbranch - one alternative of an | operator
6396 * Implements the concatenation operator.
6399 S_regbranch(pTHX_ RExC_state_t *pRExC_state, I32 *flagp, I32 first, U32 depth)
6402 register regnode *ret;
6403 register regnode *chain = NULL;
6404 register regnode *latest;
6405 I32 flags = 0, c = 0;
6406 GET_RE_DEBUG_FLAGS_DECL;
6408 PERL_ARGS_ASSERT_REGBRANCH;
6410 DEBUG_PARSE("brnc");
6415 if (!SIZE_ONLY && RExC_extralen)
6416 ret = reganode(pRExC_state, BRANCHJ,0);
6418 ret = reg_node(pRExC_state, BRANCH);
6419 Set_Node_Length(ret, 1);
6423 if (!first && SIZE_ONLY)
6424 RExC_extralen += 1; /* BRANCHJ */
6426 *flagp = WORST; /* Tentatively. */
6429 nextchar(pRExC_state);
6430 while (RExC_parse < RExC_end && *RExC_parse != '|' && *RExC_parse != ')') {
6432 latest = regpiece(pRExC_state, &flags,depth+1);
6433 if (latest == NULL) {
6434 if (flags & TRYAGAIN)
6438 else if (ret == NULL)
6440 *flagp |= flags&(HASWIDTH|POSTPONED);
6441 if (chain == NULL) /* First piece. */
6442 *flagp |= flags&SPSTART;
6445 REGTAIL(pRExC_state, chain, latest);
6450 if (chain == NULL) { /* Loop ran zero times. */
6451 chain = reg_node(pRExC_state, NOTHING);
6456 *flagp |= flags&SIMPLE;
6463 - regpiece - something followed by possible [*+?]
6465 * Note that the branching code sequences used for ? and the general cases
6466 * of * and + are somewhat optimized: they use the same NOTHING node as
6467 * both the endmarker for their branch list and the body of the last branch.
6468 * It might seem that this node could be dispensed with entirely, but the
6469 * endmarker role is not redundant.
6472 S_regpiece(pTHX_ RExC_state_t *pRExC_state, I32 *flagp, U32 depth)
6475 register regnode *ret;
6477 register char *next;
6479 const char * const origparse = RExC_parse;
6481 I32 max = REG_INFTY;
6483 const char *maxpos = NULL;
6484 GET_RE_DEBUG_FLAGS_DECL;
6486 PERL_ARGS_ASSERT_REGPIECE;
6488 DEBUG_PARSE("piec");
6490 ret = regatom(pRExC_state, &flags,depth+1);
6492 if (flags & TRYAGAIN)
6499 if (op == '{' && regcurly(RExC_parse)) {
6501 parse_start = RExC_parse; /* MJD */
6502 next = RExC_parse + 1;
6503 while (isDIGIT(*next) || *next == ',') {
6512 if (*next == '}') { /* got one */
6516 min = atoi(RExC_parse);
6520 maxpos = RExC_parse;
6522 if (!max && *maxpos != '0')
6523 max = REG_INFTY; /* meaning "infinity" */
6524 else if (max >= REG_INFTY)
6525 vFAIL2("Quantifier in {,} bigger than %d", REG_INFTY - 1);
6527 nextchar(pRExC_state);
6530 if ((flags&SIMPLE)) {
6531 RExC_naughty += 2 + RExC_naughty / 2;
6532 reginsert(pRExC_state, CURLY, ret, depth+1);
6533 Set_Node_Offset(ret, parse_start+1); /* MJD */
6534 Set_Node_Cur_Length(ret);
6537 regnode * const w = reg_node(pRExC_state, WHILEM);
6540 REGTAIL(pRExC_state, ret, w);
6541 if (!SIZE_ONLY && RExC_extralen) {
6542 reginsert(pRExC_state, LONGJMP,ret, depth+1);
6543 reginsert(pRExC_state, NOTHING,ret, depth+1);
6544 NEXT_OFF(ret) = 3; /* Go over LONGJMP. */
6546 reginsert(pRExC_state, CURLYX,ret, depth+1);
6548 Set_Node_Offset(ret, parse_start+1);
6549 Set_Node_Length(ret,
6550 op == '{' ? (RExC_parse - parse_start) : 1);
6552 if (!SIZE_ONLY && RExC_extralen)
6553 NEXT_OFF(ret) = 3; /* Go over NOTHING to LONGJMP. */
6554 REGTAIL(pRExC_state, ret, reg_node(pRExC_state, NOTHING));
6556 RExC_whilem_seen++, RExC_extralen += 3;
6557 RExC_naughty += 4 + RExC_naughty; /* compound interest */
6566 vFAIL("Can't do {n,m} with n > m");
6568 ARG1_SET(ret, (U16)min);
6569 ARG2_SET(ret, (U16)max);
6581 #if 0 /* Now runtime fix should be reliable. */
6583 /* if this is reinstated, don't forget to put this back into perldiag:
6585 =item Regexp *+ operand could be empty at {#} in regex m/%s/
6587 (F) The part of the regexp subject to either the * or + quantifier
6588 could match an empty string. The {#} shows in the regular
6589 expression about where the problem was discovered.
6593 if (!(flags&HASWIDTH) && op != '?')
6594 vFAIL("Regexp *+ operand could be empty");
6597 parse_start = RExC_parse;
6598 nextchar(pRExC_state);
6600 *flagp = (op != '+') ? (WORST|SPSTART|HASWIDTH) : (WORST|HASWIDTH);
6602 if (op == '*' && (flags&SIMPLE)) {
6603 reginsert(pRExC_state, STAR, ret, depth+1);
6607 else if (op == '*') {
6611 else if (op == '+' && (flags&SIMPLE)) {
6612 reginsert(pRExC_state, PLUS, ret, depth+1);
6616 else if (op == '+') {
6620 else if (op == '?') {
6625 if (!SIZE_ONLY && !(flags&(HASWIDTH|POSTPONED)) && max > REG_INFTY/3) {
6626 ckWARN3reg(RExC_parse,
6627 "%.*s matches null string many times",
6628 (int)(RExC_parse >= origparse ? RExC_parse - origparse : 0),
6632 if (RExC_parse < RExC_end && *RExC_parse == '?') {
6633 nextchar(pRExC_state);
6634 reginsert(pRExC_state, MINMOD, ret, depth+1);
6635 REGTAIL(pRExC_state, ret, ret + NODE_STEP_REGNODE);
6637 #ifndef REG_ALLOW_MINMOD_SUSPEND
6640 if (RExC_parse < RExC_end && *RExC_parse == '+') {
6642 nextchar(pRExC_state);
6643 ender = reg_node(pRExC_state, SUCCEED);
6644 REGTAIL(pRExC_state, ret, ender);
6645 reginsert(pRExC_state, SUSPEND, ret, depth+1);
6647 ender = reg_node(pRExC_state, TAIL);
6648 REGTAIL(pRExC_state, ret, ender);
6652 if (RExC_parse < RExC_end && ISMULT2(RExC_parse)) {
6654 vFAIL("Nested quantifiers");
6661 /* reg_namedseq(pRExC_state,UVp)
6663 This is expected to be called by a parser routine that has
6664 recognized '\N' and needs to handle the rest. RExC_parse is
6665 expected to point at the first char following the N at the time
6668 The \N may be inside (indicated by valuep not being NULL) or outside a
6671 \N may begin either a named sequence, or if outside a character class, mean
6672 to match a non-newline. For non single-quoted regexes, the tokenizer has
6673 attempted to decide which, and in the case of a named sequence converted it
6674 into one of the forms: \N{} (if the sequence is null), or \N{U+c1.c2...},
6675 where c1... are the characters in the sequence. For single-quoted regexes,
6676 the tokenizer passes the \N sequence through unchanged; this code will not
6677 attempt to determine this nor expand those. The net effect is that if the
6678 beginning of the passed-in pattern isn't '{U+' or there is no '}', it
6679 signals that this \N occurrence means to match a non-newline.
6681 Only the \N{U+...} form should occur in a character class, for the same
6682 reason that '.' inside a character class means to just match a period: it
6683 just doesn't make sense.
6685 If valuep is non-null then it is assumed that we are parsing inside
6686 of a charclass definition and the first codepoint in the resolved
6687 string is returned via *valuep and the routine will return NULL.
6688 In this mode if a multichar string is returned from the charnames
6689 handler, a warning will be issued, and only the first char in the
6690 sequence will be examined. If the string returned is zero length
6691 then the value of *valuep is undefined and NON-NULL will
6692 be returned to indicate failure. (This will NOT be a valid pointer
6695 If valuep is null then it is assumed that we are parsing normal text and a
6696 new EXACT node is inserted into the program containing the resolved string,
6697 and a pointer to the new node is returned. But if the string is zero length
6698 a NOTHING node is emitted instead.
6700 On success RExC_parse is set to the char following the endbrace.
6701 Parsing failures will generate a fatal error via vFAIL(...)
6704 S_reg_namedseq(pTHX_ RExC_state_t *pRExC_state, UV *valuep, I32 *flagp)
6706 char * endbrace; /* '}' following the name */
6707 regnode *ret = NULL;
6709 char* parse_start = RExC_parse - 2; /* points to the '\N' */
6713 GET_RE_DEBUG_FLAGS_DECL;
6715 PERL_ARGS_ASSERT_REG_NAMEDSEQ;
6719 /* The [^\n] meaning of \N ignores spaces and comments under the /x
6720 * modifier. The other meaning does not */
6721 p = (RExC_flags & RXf_PMf_EXTENDED)
6722 ? regwhite( pRExC_state, RExC_parse )
6725 /* Disambiguate between \N meaning a named character versus \N meaning
6726 * [^\n]. The former is assumed when it can't be the latter. */
6727 if (*p != '{' || regcurly(p)) {
6730 /* no bare \N in a charclass */
6731 vFAIL("\\N in a character class must be a named character: \\N{...}");
6733 nextchar(pRExC_state);
6734 ret = reg_node(pRExC_state, REG_ANY);
6735 *flagp |= HASWIDTH|SIMPLE;
6738 Set_Node_Length(ret, 1); /* MJD */
6742 /* Here, we have decided it should be a named sequence */
6744 /* The test above made sure that the next real character is a '{', but
6745 * under the /x modifier, it could be separated by space (or a comment and
6746 * \n) and this is not allowed (for consistency with \x{...} and the
6747 * tokenizer handling of \N{NAME}). */
6748 if (*RExC_parse != '{') {
6749 vFAIL("Missing braces on \\N{}");
6752 RExC_parse++; /* Skip past the '{' */
6754 if (! (endbrace = strchr(RExC_parse, '}')) /* no trailing brace */
6755 || ! (endbrace == RExC_parse /* nothing between the {} */
6756 || (endbrace - RExC_parse >= 2 /* U+ (bad hex is checked below */
6757 && strnEQ(RExC_parse, "U+", 2)))) /* for a better error msg) */
6759 if (endbrace) RExC_parse = endbrace; /* position msg's '<--HERE' */
6760 vFAIL("\\N{NAME} must be resolved by the lexer");
6763 if (endbrace == RExC_parse) { /* empty: \N{} */
6765 RExC_parse = endbrace + 1;
6766 return reg_node(pRExC_state,NOTHING);
6770 ckWARNreg(RExC_parse,
6771 "Ignoring zero length \\N{} in character class"
6773 RExC_parse = endbrace + 1;
6776 return (regnode *) &RExC_parse; /* Invalid regnode pointer */
6779 REQUIRE_UTF8; /* named sequences imply Unicode semantics */
6780 RExC_parse += 2; /* Skip past the 'U+' */
6782 if (valuep) { /* In a bracketed char class */
6783 /* We only pay attention to the first char of
6784 multichar strings being returned. I kinda wonder
6785 if this makes sense as it does change the behaviour
6786 from earlier versions, OTOH that behaviour was broken
6787 as well. XXX Solution is to recharacterize as
6788 [rest-of-class]|multi1|multi2... */
6790 STRLEN length_of_hex;
6791 I32 flags = PERL_SCAN_ALLOW_UNDERSCORES
6792 | PERL_SCAN_DISALLOW_PREFIX
6793 | (SIZE_ONLY ? PERL_SCAN_SILENT_ILLDIGIT : 0);
6795 char * endchar = RExC_parse + strcspn(RExC_parse, ".}");
6796 if (endchar < endbrace) {
6797 ckWARNreg(endchar, "Using just the first character returned by \\N{} in character class");
6800 length_of_hex = (STRLEN)(endchar - RExC_parse);
6801 *valuep = grok_hex(RExC_parse, &length_of_hex, &flags, NULL);
6803 /* The tokenizer should have guaranteed validity, but it's possible to
6804 * bypass it by using single quoting, so check */
6805 if (length_of_hex == 0
6806 || length_of_hex != (STRLEN)(endchar - RExC_parse) )
6808 RExC_parse += length_of_hex; /* Includes all the valid */
6809 RExC_parse += (RExC_orig_utf8) /* point to after 1st invalid */
6810 ? UTF8SKIP(RExC_parse)
6812 /* Guard against malformed utf8 */
6813 if (RExC_parse >= endchar) RExC_parse = endchar;
6814 vFAIL("Invalid hexadecimal number in \\N{U+...}");
6817 RExC_parse = endbrace + 1;
6818 if (endchar == endbrace) return NULL;
6820 ret = (regnode *) &RExC_parse; /* Invalid regnode pointer */
6822 else { /* Not a char class */
6823 char *s; /* String to put in generated EXACT node */
6824 STRLEN len = 0; /* Its current byte length */
6825 char *endchar; /* Points to '.' or '}' ending cur char in the input
6828 ret = reg_node(pRExC_state,
6829 (U8)(FOLD ? (LOC ? EXACTFL : EXACTF) : EXACT));
6832 /* Exact nodes can hold only a U8 length's of text = 255. Loop through
6833 * the input which is of the form now 'c1.c2.c3...}' until find the
6834 * ending brace or exceed length 255. The characters that exceed this
6835 * limit are dropped. The limit could be relaxed should it become
6836 * desirable by reparsing this as (?:\N{NAME}), so could generate
6837 * multiple EXACT nodes, as is done for just regular input. But this
6838 * is primarily a named character, and not intended to be a huge long
6839 * string, so 255 bytes should be good enough */
6841 STRLEN length_of_hex;
6842 I32 grok_flags = PERL_SCAN_ALLOW_UNDERSCORES
6843 | PERL_SCAN_DISALLOW_PREFIX
6844 | (SIZE_ONLY ? PERL_SCAN_SILENT_ILLDIGIT : 0);
6845 UV cp; /* Ord of current character */
6847 /* Code points are separated by dots. If none, there is only one
6848 * code point, and is terminated by the brace */
6849 endchar = RExC_parse + strcspn(RExC_parse, ".}");
6851 /* The values are Unicode even on EBCDIC machines */
6852 length_of_hex = (STRLEN)(endchar - RExC_parse);
6853 cp = grok_hex(RExC_parse, &length_of_hex, &grok_flags, NULL);
6854 if ( length_of_hex == 0
6855 || length_of_hex != (STRLEN)(endchar - RExC_parse) )
6857 RExC_parse += length_of_hex; /* Includes all the valid */
6858 RExC_parse += (RExC_orig_utf8) /* point to after 1st invalid */
6859 ? UTF8SKIP(RExC_parse)
6861 /* Guard against malformed utf8 */
6862 if (RExC_parse >= endchar) RExC_parse = endchar;
6863 vFAIL("Invalid hexadecimal number in \\N{U+...}");
6866 if (! FOLD) { /* Not folding, just append to the string */
6869 /* Quit before adding this character if would exceed limit */
6870 if (len + UNISKIP(cp) > U8_MAX) break;
6872 unilen = reguni(pRExC_state, cp, s);
6877 } else { /* Folding, output the folded equivalent */
6878 STRLEN foldlen,numlen;
6879 U8 tmpbuf[UTF8_MAXBYTES_CASE+1], *foldbuf;
6880 cp = toFOLD_uni(cp, tmpbuf, &foldlen);
6882 /* Quit before exceeding size limit */
6883 if (len + foldlen > U8_MAX) break;
6885 for (foldbuf = tmpbuf;
6889 cp = utf8_to_uvchr(foldbuf, &numlen);
6891 const STRLEN unilen = reguni(pRExC_state, cp, s);
6894 /* In EBCDIC the numlen and unilen can differ. */
6896 if (numlen >= foldlen)
6900 break; /* "Can't happen." */
6904 /* Point to the beginning of the next character in the sequence. */
6905 RExC_parse = endchar + 1;
6907 /* Quit if no more characters */
6908 if (RExC_parse >= endbrace) break;
6913 if (RExC_parse < endbrace) {
6914 ckWARNreg(RExC_parse - 1,
6915 "Using just the first characters returned by \\N{}");
6918 RExC_size += STR_SZ(len);
6921 RExC_emit += STR_SZ(len);
6924 RExC_parse = endbrace + 1;
6926 *flagp |= HASWIDTH; /* Not SIMPLE, as that causes the engine to fail
6927 with malformed in t/re/pat_advanced.t */
6929 Set_Node_Cur_Length(ret); /* MJD */
6930 nextchar(pRExC_state);
6940 * It returns the code point in utf8 for the value in *encp.
6941 * value: a code value in the source encoding
6942 * encp: a pointer to an Encode object
6944 * If the result from Encode is not a single character,
6945 * it returns U+FFFD (Replacement character) and sets *encp to NULL.
6948 S_reg_recode(pTHX_ const char value, SV **encp)
6951 SV * const sv = newSVpvn_flags(&value, numlen, SVs_TEMP);
6952 const char * const s = *encp ? sv_recode_to_utf8(sv, *encp) : SvPVX(sv);
6953 const STRLEN newlen = SvCUR(sv);
6954 UV uv = UNICODE_REPLACEMENT;
6956 PERL_ARGS_ASSERT_REG_RECODE;
6960 ? utf8n_to_uvchr((U8*)s, newlen, &numlen, UTF8_ALLOW_DEFAULT)
6963 if (!newlen || numlen != newlen) {
6964 uv = UNICODE_REPLACEMENT;
6972 - regatom - the lowest level
6974 Try to identify anything special at the start of the pattern. If there
6975 is, then handle it as required. This may involve generating a single regop,
6976 such as for an assertion; or it may involve recursing, such as to
6977 handle a () structure.
6979 If the string doesn't start with something special then we gobble up
6980 as much literal text as we can.
6982 Once we have been able to handle whatever type of thing started the
6983 sequence, we return.
6985 Note: we have to be careful with escapes, as they can be both literal
6986 and special, and in the case of \10 and friends can either, depending
6987 on context. Specifically there are two seperate switches for handling
6988 escape sequences, with the one for handling literal escapes requiring
6989 a dummy entry for all of the special escapes that are actually handled
6994 S_regatom(pTHX_ RExC_state_t *pRExC_state, I32 *flagp, U32 depth)
6997 register regnode *ret = NULL;
6999 char *parse_start = RExC_parse;
7000 GET_RE_DEBUG_FLAGS_DECL;
7001 DEBUG_PARSE("atom");
7002 *flagp = WORST; /* Tentatively. */
7004 PERL_ARGS_ASSERT_REGATOM;
7007 switch ((U8)*RExC_parse) {
7009 RExC_seen_zerolen++;
7010 nextchar(pRExC_state);
7011 if (RExC_flags & RXf_PMf_MULTILINE)
7012 ret = reg_node(pRExC_state, MBOL);
7013 else if (RExC_flags & RXf_PMf_SINGLELINE)
7014 ret = reg_node(pRExC_state, SBOL);
7016 ret = reg_node(pRExC_state, BOL);
7017 Set_Node_Length(ret, 1); /* MJD */
7020 nextchar(pRExC_state);
7022 RExC_seen_zerolen++;
7023 if (RExC_flags & RXf_PMf_MULTILINE)
7024 ret = reg_node(pRExC_state, MEOL);
7025 else if (RExC_flags & RXf_PMf_SINGLELINE)
7026 ret = reg_node(pRExC_state, SEOL);
7028 ret = reg_node(pRExC_state, EOL);
7029 Set_Node_Length(ret, 1); /* MJD */
7032 nextchar(pRExC_state);
7033 if (RExC_flags & RXf_PMf_SINGLELINE)
7034 ret = reg_node(pRExC_state, SANY);
7036 ret = reg_node(pRExC_state, REG_ANY);
7037 *flagp |= HASWIDTH|SIMPLE;
7039 Set_Node_Length(ret, 1); /* MJD */
7043 char * const oregcomp_parse = ++RExC_parse;
7044 ret = regclass(pRExC_state,depth+1);
7045 if (*RExC_parse != ']') {
7046 RExC_parse = oregcomp_parse;
7047 vFAIL("Unmatched [");
7049 nextchar(pRExC_state);
7050 *flagp |= HASWIDTH|SIMPLE;
7051 Set_Node_Length(ret, RExC_parse - oregcomp_parse + 1); /* MJD */
7055 nextchar(pRExC_state);
7056 ret = reg(pRExC_state, 1, &flags,depth+1);
7058 if (flags & TRYAGAIN) {
7059 if (RExC_parse == RExC_end) {
7060 /* Make parent create an empty node if needed. */
7068 *flagp |= flags&(HASWIDTH|SPSTART|SIMPLE|POSTPONED);
7072 if (flags & TRYAGAIN) {
7076 vFAIL("Internal urp");
7077 /* Supposed to be caught earlier. */
7080 if (!regcurly(RExC_parse)) {
7089 vFAIL("Quantifier follows nothing");
7097 len=0; /* silence a spurious compiler warning */
7098 if ((cp = what_len_TRICKYFOLD_safe(RExC_parse,RExC_end,UTF,len))) {
7099 *flagp |= HASWIDTH; /* could be SIMPLE too, but needs a handler in regexec.regrepeat */
7100 RExC_parse+=len-1; /* we get one from nextchar() as well. :-( */
7101 ret = reganode(pRExC_state, FOLDCHAR, cp);
7102 Set_Node_Length(ret, 1); /* MJD */
7103 nextchar(pRExC_state); /* kill whitespace under /x */
7111 This switch handles escape sequences that resolve to some kind
7112 of special regop and not to literal text. Escape sequnces that
7113 resolve to literal text are handled below in the switch marked
7116 Every entry in this switch *must* have a corresponding entry
7117 in the literal escape switch. However, the opposite is not
7118 required, as the default for this switch is to jump to the
7119 literal text handling code.
7121 switch ((U8)*++RExC_parse) {
7126 /* Special Escapes */
7128 RExC_seen_zerolen++;
7129 ret = reg_node(pRExC_state, SBOL);
7131 goto finish_meta_pat;
7133 ret = reg_node(pRExC_state, GPOS);
7134 RExC_seen |= REG_SEEN_GPOS;
7136 goto finish_meta_pat;
7138 RExC_seen_zerolen++;
7139 ret = reg_node(pRExC_state, KEEPS);
7141 /* XXX:dmq : disabling in-place substitution seems to
7142 * be necessary here to avoid cases of memory corruption, as
7143 * with: C<$_="x" x 80; s/x\K/y/> -- rgs
7145 RExC_seen |= REG_SEEN_LOOKBEHIND;
7146 goto finish_meta_pat;
7148 ret = reg_node(pRExC_state, SEOL);
7150 RExC_seen_zerolen++; /* Do not optimize RE away */
7151 goto finish_meta_pat;
7153 ret = reg_node(pRExC_state, EOS);
7155 RExC_seen_zerolen++; /* Do not optimize RE away */
7156 goto finish_meta_pat;
7158 ret = reg_node(pRExC_state, CANY);
7159 RExC_seen |= REG_SEEN_CANY;
7160 *flagp |= HASWIDTH|SIMPLE;
7161 goto finish_meta_pat;
7163 ret = reg_node(pRExC_state, CLUMP);
7165 goto finish_meta_pat;
7167 ret = reg_node(pRExC_state, (U8)(LOC ? ALNUML : ALNUM));
7168 *flagp |= HASWIDTH|SIMPLE;
7169 goto finish_meta_pat;
7171 ret = reg_node(pRExC_state, (U8)(LOC ? NALNUML : NALNUM));
7172 *flagp |= HASWIDTH|SIMPLE;
7173 goto finish_meta_pat;
7175 RExC_seen_zerolen++;
7176 RExC_seen |= REG_SEEN_LOOKBEHIND;
7177 ret = reg_node(pRExC_state, (U8)(LOC ? BOUNDL : BOUND));
7179 goto finish_meta_pat;
7181 RExC_seen_zerolen++;
7182 RExC_seen |= REG_SEEN_LOOKBEHIND;
7183 ret = reg_node(pRExC_state, (U8)(LOC ? NBOUNDL : NBOUND));
7185 goto finish_meta_pat;
7187 ret = reg_node(pRExC_state, (U8)(LOC ? SPACEL : SPACE));
7188 *flagp |= HASWIDTH|SIMPLE;
7189 goto finish_meta_pat;
7191 ret = reg_node(pRExC_state, (U8)(LOC ? NSPACEL : NSPACE));
7192 *flagp |= HASWIDTH|SIMPLE;
7193 goto finish_meta_pat;
7195 ret = reg_node(pRExC_state, DIGIT);
7196 *flagp |= HASWIDTH|SIMPLE;
7197 goto finish_meta_pat;
7199 ret = reg_node(pRExC_state, NDIGIT);
7200 *flagp |= HASWIDTH|SIMPLE;
7201 goto finish_meta_pat;
7203 ret = reg_node(pRExC_state, LNBREAK);
7204 *flagp |= HASWIDTH|SIMPLE;
7205 goto finish_meta_pat;
7207 ret = reg_node(pRExC_state, HORIZWS);
7208 *flagp |= HASWIDTH|SIMPLE;
7209 goto finish_meta_pat;
7211 ret = reg_node(pRExC_state, NHORIZWS);
7212 *flagp |= HASWIDTH|SIMPLE;
7213 goto finish_meta_pat;
7215 ret = reg_node(pRExC_state, VERTWS);
7216 *flagp |= HASWIDTH|SIMPLE;
7217 goto finish_meta_pat;
7219 ret = reg_node(pRExC_state, NVERTWS);
7220 *flagp |= HASWIDTH|SIMPLE;
7222 nextchar(pRExC_state);
7223 Set_Node_Length(ret, 2); /* MJD */
7228 char* const oldregxend = RExC_end;
7230 char* parse_start = RExC_parse - 2;
7233 if (RExC_parse[1] == '{') {
7234 /* a lovely hack--pretend we saw [\pX] instead */
7235 RExC_end = strchr(RExC_parse, '}');
7237 const U8 c = (U8)*RExC_parse;
7239 RExC_end = oldregxend;
7240 vFAIL2("Missing right brace on \\%c{}", c);
7245 RExC_end = RExC_parse + 2;
7246 if (RExC_end > oldregxend)
7247 RExC_end = oldregxend;
7251 ret = regclass(pRExC_state,depth+1);
7253 RExC_end = oldregxend;
7256 Set_Node_Offset(ret, parse_start + 2);
7257 Set_Node_Cur_Length(ret);
7258 nextchar(pRExC_state);
7259 *flagp |= HASWIDTH|SIMPLE;
7263 /* Handle \N and \N{NAME} here and not below because it can be
7264 multicharacter. join_exact() will join them up later on.
7265 Also this makes sure that things like /\N{BLAH}+/ and
7266 \N{BLAH} being multi char Just Happen. dmq*/
7268 ret= reg_namedseq(pRExC_state, NULL, flagp);
7270 case 'k': /* Handle \k<NAME> and \k'NAME' */
7273 char ch= RExC_parse[1];
7274 if (ch != '<' && ch != '\'' && ch != '{') {
7276 vFAIL2("Sequence %.2s... not terminated",parse_start);
7278 /* this pretty much dupes the code for (?P=...) in reg(), if
7279 you change this make sure you change that */
7280 char* name_start = (RExC_parse += 2);
7282 SV *sv_dat = reg_scan_name(pRExC_state,
7283 SIZE_ONLY ? REG_RSN_RETURN_NULL : REG_RSN_RETURN_DATA);
7284 ch= (ch == '<') ? '>' : (ch == '{') ? '}' : '\'';
7285 if (RExC_parse == name_start || *RExC_parse != ch)
7286 vFAIL2("Sequence %.3s... not terminated",parse_start);
7289 num = add_data( pRExC_state, 1, "S" );
7290 RExC_rxi->data->data[num]=(void*)sv_dat;
7291 SvREFCNT_inc_simple_void(sv_dat);
7295 ret = reganode(pRExC_state,
7296 (U8)(FOLD ? (LOC ? NREFFL : NREFF) : NREF),
7300 /* override incorrect value set in reganode MJD */
7301 Set_Node_Offset(ret, parse_start+1);
7302 Set_Node_Cur_Length(ret); /* MJD */
7303 nextchar(pRExC_state);
7309 case '1': case '2': case '3': case '4':
7310 case '5': case '6': case '7': case '8': case '9':
7313 bool isg = *RExC_parse == 'g';
7318 if (*RExC_parse == '{') {
7322 if (*RExC_parse == '-') {
7326 if (hasbrace && !isDIGIT(*RExC_parse)) {
7327 if (isrel) RExC_parse--;
7329 goto parse_named_seq;
7331 num = atoi(RExC_parse);
7332 if (isg && num == 0)
7333 vFAIL("Reference to invalid group 0");
7335 num = RExC_npar - num;
7337 vFAIL("Reference to nonexistent or unclosed group");
7339 if (!isg && num > 9 && num >= RExC_npar)
7342 char * const parse_start = RExC_parse - 1; /* MJD */
7343 while (isDIGIT(*RExC_parse))
7345 if (parse_start == RExC_parse - 1)
7346 vFAIL("Unterminated \\g... pattern");
7348 if (*RExC_parse != '}')
7349 vFAIL("Unterminated \\g{...} pattern");
7353 if (num > (I32)RExC_rx->nparens)
7354 vFAIL("Reference to nonexistent group");
7357 ret = reganode(pRExC_state,
7358 (U8)(FOLD ? (LOC ? REFFL : REFF) : REF),
7362 /* override incorrect value set in reganode MJD */
7363 Set_Node_Offset(ret, parse_start+1);
7364 Set_Node_Cur_Length(ret); /* MJD */
7366 nextchar(pRExC_state);
7371 if (RExC_parse >= RExC_end)
7372 FAIL("Trailing \\");
7375 /* Do not generate "unrecognized" warnings here, we fall
7376 back into the quick-grab loop below */
7383 if (RExC_flags & RXf_PMf_EXTENDED) {
7384 if ( reg_skipcomment( pRExC_state ) )
7391 register STRLEN len;
7396 U8 tmpbuf[UTF8_MAXBYTES_CASE+1], *foldbuf;
7398 parse_start = RExC_parse - 1;
7404 ret = reg_node(pRExC_state,
7405 (U8)(FOLD ? (LOC ? EXACTFL : EXACTF) : EXACT));
7407 for (len = 0, p = RExC_parse - 1;
7408 len < 127 && p < RExC_end;
7411 char * const oldp = p;
7413 if (RExC_flags & RXf_PMf_EXTENDED)
7414 p = regwhite( pRExC_state, p );
7419 if (LOC || !FOLD || !is_TRICKYFOLD_safe(p,RExC_end,UTF))
7420 goto normal_default;
7430 /* Literal Escapes Switch
7432 This switch is meant to handle escape sequences that
7433 resolve to a literal character.
7435 Every escape sequence that represents something
7436 else, like an assertion or a char class, is handled
7437 in the switch marked 'Special Escapes' above in this
7438 routine, but also has an entry here as anything that
7439 isn't explicitly mentioned here will be treated as
7440 an unescaped equivalent literal.
7444 /* These are all the special escapes. */
7448 if (LOC || !FOLD || !is_TRICKYFOLD_safe(p,RExC_end,UTF))
7449 goto normal_default;
7450 case 'A': /* Start assertion */
7451 case 'b': case 'B': /* Word-boundary assertion*/
7452 case 'C': /* Single char !DANGEROUS! */
7453 case 'd': case 'D': /* digit class */
7454 case 'g': case 'G': /* generic-backref, pos assertion */
7455 case 'h': case 'H': /* HORIZWS */
7456 case 'k': case 'K': /* named backref, keep marker */
7457 case 'N': /* named char sequence */
7458 case 'p': case 'P': /* Unicode property */
7459 case 'R': /* LNBREAK */
7460 case 's': case 'S': /* space class */
7461 case 'v': case 'V': /* VERTWS */
7462 case 'w': case 'W': /* word class */
7463 case 'X': /* eXtended Unicode "combining character sequence" */
7464 case 'z': case 'Z': /* End of line/string assertion */
7468 /* Anything after here is an escape that resolves to a
7469 literal. (Except digits, which may or may not)
7488 ender = ASCII_TO_NATIVE('\033');
7492 ender = ASCII_TO_NATIVE('\007');
7497 STRLEN brace_len = len;
7499 const char* error_msg;
7501 bool valid = grok_bslash_o(p,
7508 RExC_parse = p; /* going to die anyway; point
7509 to exact spot of failure */
7516 if (PL_encoding && ender < 0x100) {
7517 goto recode_encoding;
7526 char* const e = strchr(p, '}');
7530 vFAIL("Missing right brace on \\x{}");
7533 I32 flags = PERL_SCAN_ALLOW_UNDERSCORES
7534 | PERL_SCAN_DISALLOW_PREFIX;
7535 STRLEN numlen = e - p - 1;
7536 ender = grok_hex(p + 1, &numlen, &flags, NULL);
7543 I32 flags = PERL_SCAN_DISALLOW_PREFIX;
7545 ender = grok_hex(p, &numlen, &flags, NULL);
7548 if (PL_encoding && ender < 0x100)
7549 goto recode_encoding;
7553 ender = grok_bslash_c(*p++, SIZE_ONLY);
7555 case '0': case '1': case '2': case '3':case '4':
7556 case '5': case '6': case '7': case '8':case '9':
7558 (isOCTAL(p[1]) && atoi(p) >= RExC_npar))
7560 I32 flags = PERL_SCAN_SILENT_ILLDIGIT;
7562 ender = grok_oct(p, &numlen, &flags, NULL);
7572 if (PL_encoding && ender < 0x100)
7573 goto recode_encoding;
7577 SV* enc = PL_encoding;
7578 ender = reg_recode((const char)(U8)ender, &enc);
7579 if (!enc && SIZE_ONLY)
7580 ckWARNreg(p, "Invalid escape in the specified encoding");
7586 FAIL("Trailing \\");
7589 if (!SIZE_ONLY&& isALPHA(*p))
7590 ckWARN2reg(p + 1, "Unrecognized escape \\%c passed through", UCHARAT(p));
7591 goto normal_default;
7596 if (UTF8_IS_START(*p) && UTF) {
7598 ender = utf8n_to_uvchr((U8*)p, RExC_end - p,
7599 &numlen, UTF8_ALLOW_DEFAULT);
7606 if ( RExC_flags & RXf_PMf_EXTENDED)
7607 p = regwhite( pRExC_state, p );
7609 /* Prime the casefolded buffer. */
7610 ender = toFOLD_uni(ender, tmpbuf, &foldlen);
7612 if (p < RExC_end && ISMULT2(p)) { /* Back off on ?+*. */
7617 /* Emit all the Unicode characters. */
7619 for (foldbuf = tmpbuf;
7621 foldlen -= numlen) {
7622 ender = utf8_to_uvchr(foldbuf, &numlen);
7624 const STRLEN unilen = reguni(pRExC_state, ender, s);
7627 /* In EBCDIC the numlen
7628 * and unilen can differ. */
7630 if (numlen >= foldlen)
7634 break; /* "Can't happen." */
7638 const STRLEN unilen = reguni(pRExC_state, ender, s);
7647 REGC((char)ender, s++);
7653 /* Emit all the Unicode characters. */
7655 for (foldbuf = tmpbuf;
7657 foldlen -= numlen) {
7658 ender = utf8_to_uvchr(foldbuf, &numlen);
7660 const STRLEN unilen = reguni(pRExC_state, ender, s);
7663 /* In EBCDIC the numlen
7664 * and unilen can differ. */
7666 if (numlen >= foldlen)
7674 const STRLEN unilen = reguni(pRExC_state, ender, s);
7683 REGC((char)ender, s++);
7687 Set_Node_Cur_Length(ret); /* MJD */
7688 nextchar(pRExC_state);
7690 /* len is STRLEN which is unsigned, need to copy to signed */
7693 vFAIL("Internal disaster");
7697 if (len == 1 && UNI_IS_INVARIANT(ender))
7701 RExC_size += STR_SZ(len);
7704 RExC_emit += STR_SZ(len);
7714 S_regwhite( RExC_state_t *pRExC_state, char *p )
7716 const char *e = RExC_end;
7718 PERL_ARGS_ASSERT_REGWHITE;
7723 else if (*p == '#') {
7732 RExC_seen |= REG_SEEN_RUN_ON_COMMENT;
7740 /* Parse POSIX character classes: [[:foo:]], [[=foo=]], [[.foo.]].
7741 Character classes ([:foo:]) can also be negated ([:^foo:]).
7742 Returns a named class id (ANYOF_XXX) if successful, -1 otherwise.
7743 Equivalence classes ([=foo=]) and composites ([.foo.]) are parsed,
7744 but trigger failures because they are currently unimplemented. */
7746 #define POSIXCC_DONE(c) ((c) == ':')
7747 #define POSIXCC_NOTYET(c) ((c) == '=' || (c) == '.')
7748 #define POSIXCC(c) (POSIXCC_DONE(c) || POSIXCC_NOTYET(c))
7751 S_regpposixcc(pTHX_ RExC_state_t *pRExC_state, I32 value)
7754 I32 namedclass = OOB_NAMEDCLASS;
7756 PERL_ARGS_ASSERT_REGPPOSIXCC;
7758 if (value == '[' && RExC_parse + 1 < RExC_end &&
7759 /* I smell either [: or [= or [. -- POSIX has been here, right? */
7760 POSIXCC(UCHARAT(RExC_parse))) {
7761 const char c = UCHARAT(RExC_parse);
7762 char* const s = RExC_parse++;
7764 while (RExC_parse < RExC_end && UCHARAT(RExC_parse) != c)
7766 if (RExC_parse == RExC_end)
7767 /* Grandfather lone [:, [=, [. */
7770 const char* const t = RExC_parse++; /* skip over the c */
7773 if (UCHARAT(RExC_parse) == ']') {
7774 const char *posixcc = s + 1;
7775 RExC_parse++; /* skip over the ending ] */
7778 const I32 complement = *posixcc == '^' ? *posixcc++ : 0;
7779 const I32 skip = t - posixcc;
7781 /* Initially switch on the length of the name. */
7784 if (memEQ(posixcc, "word", 4)) /* this is not POSIX, this is the Perl \w */
7785 namedclass = complement ? ANYOF_NALNUM : ANYOF_ALNUM;
7788 /* Names all of length 5. */
7789 /* alnum alpha ascii blank cntrl digit graph lower
7790 print punct space upper */
7791 /* Offset 4 gives the best switch position. */
7792 switch (posixcc[4]) {
7794 if (memEQ(posixcc, "alph", 4)) /* alpha */
7795 namedclass = complement ? ANYOF_NALPHA : ANYOF_ALPHA;
7798 if (memEQ(posixcc, "spac", 4)) /* space */
7799 namedclass = complement ? ANYOF_NPSXSPC : ANYOF_PSXSPC;
7802 if (memEQ(posixcc, "grap", 4)) /* graph */
7803 namedclass = complement ? ANYOF_NGRAPH : ANYOF_GRAPH;
7806 if (memEQ(posixcc, "asci", 4)) /* ascii */
7807 namedclass = complement ? ANYOF_NASCII : ANYOF_ASCII;
7810 if (memEQ(posixcc, "blan", 4)) /* blank */
7811 namedclass = complement ? ANYOF_NBLANK : ANYOF_BLANK;
7814 if (memEQ(posixcc, "cntr", 4)) /* cntrl */
7815 namedclass = complement ? ANYOF_NCNTRL : ANYOF_CNTRL;
7818 if (memEQ(posixcc, "alnu", 4)) /* alnum */
7819 namedclass = complement ? ANYOF_NALNUMC : ANYOF_ALNUMC;
7822 if (memEQ(posixcc, "lowe", 4)) /* lower */
7823 namedclass = complement ? ANYOF_NLOWER : ANYOF_LOWER;
7824 else if (memEQ(posixcc, "uppe", 4)) /* upper */
7825 namedclass = complement ? ANYOF_NUPPER : ANYOF_UPPER;
7828 if (memEQ(posixcc, "digi", 4)) /* digit */
7829 namedclass = complement ? ANYOF_NDIGIT : ANYOF_DIGIT;
7830 else if (memEQ(posixcc, "prin", 4)) /* print */
7831 namedclass = complement ? ANYOF_NPRINT : ANYOF_PRINT;
7832 else if (memEQ(posixcc, "punc", 4)) /* punct */
7833 namedclass = complement ? ANYOF_NPUNCT : ANYOF_PUNCT;
7838 if (memEQ(posixcc, "xdigit", 6))
7839 namedclass = complement ? ANYOF_NXDIGIT : ANYOF_XDIGIT;
7843 if (namedclass == OOB_NAMEDCLASS)
7844 Simple_vFAIL3("POSIX class [:%.*s:] unknown",
7846 assert (posixcc[skip] == ':');
7847 assert (posixcc[skip+1] == ']');
7848 } else if (!SIZE_ONLY) {
7849 /* [[=foo=]] and [[.foo.]] are still future. */
7851 /* adjust RExC_parse so the warning shows after
7853 while (UCHARAT(RExC_parse) && UCHARAT(RExC_parse) != ']')
7855 Simple_vFAIL3("POSIX syntax [%c %c] is reserved for future extensions", c, c);
7858 /* Maternal grandfather:
7859 * "[:" ending in ":" but not in ":]" */
7869 S_checkposixcc(pTHX_ RExC_state_t *pRExC_state)
7873 PERL_ARGS_ASSERT_CHECKPOSIXCC;
7875 if (POSIXCC(UCHARAT(RExC_parse))) {
7876 const char *s = RExC_parse;
7877 const char c = *s++;
7881 if (*s && c == *s && s[1] == ']') {
7883 "POSIX syntax [%c %c] belongs inside character classes",
7886 /* [[=foo=]] and [[.foo.]] are still future. */
7887 if (POSIXCC_NOTYET(c)) {
7888 /* adjust RExC_parse so the error shows after
7890 while (UCHARAT(RExC_parse) && UCHARAT(RExC_parse++) != ']')
7892 Simple_vFAIL3("POSIX syntax [%c %c] is reserved for future extensions", c, c);
7899 #define _C_C_T_(NAME,TEST,WORD) \
7902 ANYOF_CLASS_SET(ret, ANYOF_##NAME); \
7904 for (value = 0; value < 256; value++) \
7906 ANYOF_BITMAP_SET(ret, value); \
7911 case ANYOF_N##NAME: \
7913 ANYOF_CLASS_SET(ret, ANYOF_N##NAME); \
7915 for (value = 0; value < 256; value++) \
7917 ANYOF_BITMAP_SET(ret, value); \
7923 #define _C_C_T_NOLOC_(NAME,TEST,WORD) \
7925 for (value = 0; value < 256; value++) \
7927 ANYOF_BITMAP_SET(ret, value); \
7931 case ANYOF_N##NAME: \
7932 for (value = 0; value < 256; value++) \
7934 ANYOF_BITMAP_SET(ret, value); \
7940 We dont use PERL_LEGACY_UNICODE_CHARCLASS_MAPPINGS as the direct test
7941 so that it is possible to override the option here without having to
7942 rebuild the entire core. as we are required to do if we change regcomp.h
7943 which is where PERL_LEGACY_UNICODE_CHARCLASS_MAPPINGS is defined.
7945 #if PERL_LEGACY_UNICODE_CHARCLASS_MAPPINGS
7946 #define BROKEN_UNICODE_CHARCLASS_MAPPINGS
7949 #ifdef BROKEN_UNICODE_CHARCLASS_MAPPINGS
7950 #define POSIX_CC_UNI_NAME(CCNAME) CCNAME
7952 #define POSIX_CC_UNI_NAME(CCNAME) "Posix" CCNAME
7956 parse a class specification and produce either an ANYOF node that
7957 matches the pattern or if the pattern matches a single char only and
7958 that char is < 256 and we are case insensitive then we produce an
7963 S_regclass(pTHX_ RExC_state_t *pRExC_state, U32 depth)
7966 register UV nextvalue;
7967 register IV prevvalue = OOB_UNICODE;
7968 register IV range = 0;
7969 UV value = 0; /* XXX:dmq: needs to be referenceable (unfortunately) */
7970 register regnode *ret;
7973 char *rangebegin = NULL;
7974 bool need_class = 0;
7977 bool optimize_invert = TRUE;
7978 AV* unicode_alternate = NULL;
7980 UV literal_endpoint = 0;
7982 UV stored = 0; /* number of chars stored in the class */
7984 regnode * const orig_emit = RExC_emit; /* Save the original RExC_emit in
7985 case we need to change the emitted regop to an EXACT. */
7986 const char * orig_parse = RExC_parse;
7987 GET_RE_DEBUG_FLAGS_DECL;
7989 PERL_ARGS_ASSERT_REGCLASS;
7991 PERL_UNUSED_ARG(depth);
7994 DEBUG_PARSE("clas");
7996 /* Assume we are going to generate an ANYOF node. */
7997 ret = reganode(pRExC_state, ANYOF, 0);
8000 ANYOF_FLAGS(ret) = 0;
8002 if (UCHARAT(RExC_parse) == '^') { /* Complement of range. */
8006 ANYOF_FLAGS(ret) |= ANYOF_INVERT;
8010 RExC_size += ANYOF_SKIP;
8011 listsv = &PL_sv_undef; /* For code scanners: listsv always non-NULL. */
8014 RExC_emit += ANYOF_SKIP;
8016 ANYOF_FLAGS(ret) |= ANYOF_FOLD;
8018 ANYOF_FLAGS(ret) |= ANYOF_LOCALE;
8019 ANYOF_BITMAP_ZERO(ret);
8020 listsv = newSVpvs("# comment\n");
8023 nextvalue = RExC_parse < RExC_end ? UCHARAT(RExC_parse) : 0;
8025 if (!SIZE_ONLY && POSIXCC(nextvalue))
8026 checkposixcc(pRExC_state);
8028 /* allow 1st char to be ] (allowing it to be - is dealt with later) */
8029 if (UCHARAT(RExC_parse) == ']')
8033 while (RExC_parse < RExC_end && UCHARAT(RExC_parse) != ']') {
8037 namedclass = OOB_NAMEDCLASS; /* initialize as illegal */
8040 rangebegin = RExC_parse;
8042 value = utf8n_to_uvchr((U8*)RExC_parse,
8043 RExC_end - RExC_parse,
8044 &numlen, UTF8_ALLOW_DEFAULT);
8045 RExC_parse += numlen;
8048 value = UCHARAT(RExC_parse++);
8050 nextvalue = RExC_parse < RExC_end ? UCHARAT(RExC_parse) : 0;
8051 if (value == '[' && POSIXCC(nextvalue))
8052 namedclass = regpposixcc(pRExC_state, value);
8053 else if (value == '\\') {
8055 value = utf8n_to_uvchr((U8*)RExC_parse,
8056 RExC_end - RExC_parse,
8057 &numlen, UTF8_ALLOW_DEFAULT);
8058 RExC_parse += numlen;
8061 value = UCHARAT(RExC_parse++);
8062 /* Some compilers cannot handle switching on 64-bit integer
8063 * values, therefore value cannot be an UV. Yes, this will
8064 * be a problem later if we want switch on Unicode.
8065 * A similar issue a little bit later when switching on
8066 * namedclass. --jhi */
8067 switch ((I32)value) {
8068 case 'w': namedclass = ANYOF_ALNUM; break;
8069 case 'W': namedclass = ANYOF_NALNUM; break;
8070 case 's': namedclass = ANYOF_SPACE; break;
8071 case 'S': namedclass = ANYOF_NSPACE; break;
8072 case 'd': namedclass = ANYOF_DIGIT; break;
8073 case 'D': namedclass = ANYOF_NDIGIT; break;
8074 case 'v': namedclass = ANYOF_VERTWS; break;
8075 case 'V': namedclass = ANYOF_NVERTWS; break;
8076 case 'h': namedclass = ANYOF_HORIZWS; break;
8077 case 'H': namedclass = ANYOF_NHORIZWS; break;
8078 case 'N': /* Handle \N{NAME} in class */
8080 /* We only pay attention to the first char of
8081 multichar strings being returned. I kinda wonder
8082 if this makes sense as it does change the behaviour
8083 from earlier versions, OTOH that behaviour was broken
8085 UV v; /* value is register so we cant & it /grrr */
8086 if (reg_namedseq(pRExC_state, &v, NULL)) {
8096 if (RExC_parse >= RExC_end)
8097 vFAIL2("Empty \\%c{}", (U8)value);
8098 if (*RExC_parse == '{') {
8099 const U8 c = (U8)value;
8100 e = strchr(RExC_parse++, '}');
8102 vFAIL2("Missing right brace on \\%c{}", c);
8103 while (isSPACE(UCHARAT(RExC_parse)))
8105 if (e == RExC_parse)
8106 vFAIL2("Empty \\%c{}", c);
8108 while (isSPACE(UCHARAT(RExC_parse + n - 1)))
8116 if (UCHARAT(RExC_parse) == '^') {
8119 value = value == 'p' ? 'P' : 'p'; /* toggle */
8120 while (isSPACE(UCHARAT(RExC_parse))) {
8125 Perl_sv_catpvf(aTHX_ listsv, "%cutf8::%.*s\n",
8126 (value=='p' ? '+' : '!'), (int)n, RExC_parse);
8129 ANYOF_FLAGS(ret) |= ANYOF_UNICODE;
8130 namedclass = ANYOF_MAX; /* no official name, but it's named */
8133 case 'n': value = '\n'; break;
8134 case 'r': value = '\r'; break;
8135 case 't': value = '\t'; break;
8136 case 'f': value = '\f'; break;
8137 case 'b': value = '\b'; break;
8138 case 'e': value = ASCII_TO_NATIVE('\033');break;
8139 case 'a': value = ASCII_TO_NATIVE('\007');break;
8141 RExC_parse--; /* function expects to be pointed at the 'o' */
8143 const char* error_msg;
8144 bool valid = grok_bslash_o(RExC_parse,
8149 RExC_parse += numlen;
8154 if (PL_encoding && value < 0x100) {
8155 goto recode_encoding;
8159 if (*RExC_parse == '{') {
8160 I32 flags = PERL_SCAN_ALLOW_UNDERSCORES
8161 | PERL_SCAN_DISALLOW_PREFIX;
8162 char * const e = strchr(RExC_parse++, '}');
8164 vFAIL("Missing right brace on \\x{}");
8166 numlen = e - RExC_parse;
8167 value = grok_hex(RExC_parse, &numlen, &flags, NULL);
8171 I32 flags = PERL_SCAN_DISALLOW_PREFIX;
8173 value = grok_hex(RExC_parse, &numlen, &flags, NULL);
8174 RExC_parse += numlen;
8176 if (PL_encoding && value < 0x100)
8177 goto recode_encoding;
8180 value = grok_bslash_c(*RExC_parse++, SIZE_ONLY);
8182 case '0': case '1': case '2': case '3': case '4':
8183 case '5': case '6': case '7':
8185 /* Take 1-3 octal digits */
8186 I32 flags = PERL_SCAN_SILENT_ILLDIGIT;
8188 value = grok_oct(--RExC_parse, &numlen, &flags, NULL);
8189 RExC_parse += numlen;
8190 if (PL_encoding && value < 0x100)
8191 goto recode_encoding;
8196 SV* enc = PL_encoding;
8197 value = reg_recode((const char)(U8)value, &enc);
8198 if (!enc && SIZE_ONLY)
8199 ckWARNreg(RExC_parse,
8200 "Invalid escape in the specified encoding");
8204 /* Allow \_ to not give an error */
8205 if (!SIZE_ONLY && isALNUM(value) && value != '_') {
8206 ckWARN2reg(RExC_parse,
8207 "Unrecognized escape \\%c in character class passed through",
8212 } /* end of \blah */
8218 if (namedclass > OOB_NAMEDCLASS) { /* this is a named class \blah */
8220 if (!SIZE_ONLY && !need_class)
8221 ANYOF_CLASS_ZERO(ret);
8225 /* a bad range like a-\d, a-[:digit:] ? */
8229 RExC_parse >= rangebegin ?
8230 RExC_parse - rangebegin : 0;
8231 ckWARN4reg(RExC_parse,
8232 "False [] range \"%*.*s\"",
8235 if (prevvalue < 256) {
8236 ANYOF_BITMAP_SET(ret, prevvalue);
8237 ANYOF_BITMAP_SET(ret, '-');
8240 ANYOF_FLAGS(ret) |= ANYOF_UNICODE;
8241 Perl_sv_catpvf(aTHX_ listsv,
8242 "%04"UVxf"\n%04"UVxf"\n", (UV)prevvalue, (UV) '-');
8246 range = 0; /* this was not a true range */
8252 const char *what = NULL;
8255 if (namedclass > OOB_NAMEDCLASS)
8256 optimize_invert = FALSE;
8257 /* Possible truncation here but in some 64-bit environments
8258 * the compiler gets heartburn about switch on 64-bit values.
8259 * A similar issue a little earlier when switching on value.
8261 switch ((I32)namedclass) {
8263 case _C_C_T_(ALNUMC, isALNUMC(value), POSIX_CC_UNI_NAME("Alnum"));
8264 case _C_C_T_(ALPHA, isALPHA(value), POSIX_CC_UNI_NAME("Alpha"));
8265 case _C_C_T_(BLANK, isBLANK(value), POSIX_CC_UNI_NAME("Blank"));
8266 case _C_C_T_(CNTRL, isCNTRL(value), POSIX_CC_UNI_NAME("Cntrl"));
8267 case _C_C_T_(GRAPH, isGRAPH(value), POSIX_CC_UNI_NAME("Graph"));
8268 case _C_C_T_(LOWER, isLOWER(value), POSIX_CC_UNI_NAME("Lower"));
8269 case _C_C_T_(PRINT, isPRINT(value), POSIX_CC_UNI_NAME("Print"));
8270 case _C_C_T_(PSXSPC, isPSXSPC(value), POSIX_CC_UNI_NAME("Space"));
8271 case _C_C_T_(PUNCT, isPUNCT(value), POSIX_CC_UNI_NAME("Punct"));
8272 case _C_C_T_(UPPER, isUPPER(value), POSIX_CC_UNI_NAME("Upper"));
8273 #ifdef BROKEN_UNICODE_CHARCLASS_MAPPINGS
8274 case _C_C_T_(ALNUM, isALNUM(value), "Word");
8275 case _C_C_T_(SPACE, isSPACE(value), "SpacePerl");
8277 case _C_C_T_(SPACE, isSPACE(value), "PerlSpace");
8278 case _C_C_T_(ALNUM, isALNUM(value), "PerlWord");
8280 case _C_C_T_(XDIGIT, isXDIGIT(value), "XDigit");
8281 case _C_C_T_NOLOC_(VERTWS, is_VERTWS_latin1(&value), "VertSpace");
8282 case _C_C_T_NOLOC_(HORIZWS, is_HORIZWS_latin1(&value), "HorizSpace");
8285 ANYOF_CLASS_SET(ret, ANYOF_ASCII);
8288 for (value = 0; value < 128; value++)
8289 ANYOF_BITMAP_SET(ret, value);
8291 for (value = 0; value < 256; value++) {
8293 ANYOF_BITMAP_SET(ret, value);
8302 ANYOF_CLASS_SET(ret, ANYOF_NASCII);
8305 for (value = 128; value < 256; value++)
8306 ANYOF_BITMAP_SET(ret, value);
8308 for (value = 0; value < 256; value++) {
8309 if (!isASCII(value))
8310 ANYOF_BITMAP_SET(ret, value);
8319 ANYOF_CLASS_SET(ret, ANYOF_DIGIT);
8321 /* consecutive digits assumed */
8322 for (value = '0'; value <= '9'; value++)
8323 ANYOF_BITMAP_SET(ret, value);
8326 what = POSIX_CC_UNI_NAME("Digit");
8330 ANYOF_CLASS_SET(ret, ANYOF_NDIGIT);
8332 /* consecutive digits assumed */
8333 for (value = 0; value < '0'; value++)
8334 ANYOF_BITMAP_SET(ret, value);
8335 for (value = '9' + 1; value < 256; value++)
8336 ANYOF_BITMAP_SET(ret, value);
8339 what = POSIX_CC_UNI_NAME("Digit");
8342 /* this is to handle \p and \P */
8345 vFAIL("Invalid [::] class");
8349 /* Strings such as "+utf8::isWord\n" */
8350 Perl_sv_catpvf(aTHX_ listsv, "%cutf8::Is%s\n", yesno, what);
8353 ANYOF_FLAGS(ret) |= ANYOF_CLASS;
8356 } /* end of namedclass \blah */
8359 if (prevvalue > (IV)value) /* b-a */ {
8360 const int w = RExC_parse - rangebegin;
8361 Simple_vFAIL4("Invalid [] range \"%*.*s\"", w, w, rangebegin);
8362 range = 0; /* not a valid range */
8366 prevvalue = value; /* save the beginning of the range */
8367 if (*RExC_parse == '-' && RExC_parse+1 < RExC_end &&
8368 RExC_parse[1] != ']') {
8371 /* a bad range like \w-, [:word:]- ? */
8372 if (namedclass > OOB_NAMEDCLASS) {
8373 if (ckWARN(WARN_REGEXP)) {
8375 RExC_parse >= rangebegin ?
8376 RExC_parse - rangebegin : 0;
8378 "False [] range \"%*.*s\"",
8382 ANYOF_BITMAP_SET(ret, '-');
8384 range = 1; /* yeah, it's a range! */
8385 continue; /* but do it the next time */
8389 /* now is the next time */
8390 /*stored += (value - prevvalue + 1);*/
8392 if (prevvalue < 256) {
8393 const IV ceilvalue = value < 256 ? value : 255;
8396 /* In EBCDIC [\x89-\x91] should include
8397 * the \x8e but [i-j] should not. */
8398 if (literal_endpoint == 2 &&
8399 ((isLOWER(prevvalue) && isLOWER(ceilvalue)) ||
8400 (isUPPER(prevvalue) && isUPPER(ceilvalue))))
8402 if (isLOWER(prevvalue)) {
8403 for (i = prevvalue; i <= ceilvalue; i++)
8404 if (isLOWER(i) && !ANYOF_BITMAP_TEST(ret,i)) {
8406 ANYOF_BITMAP_SET(ret, i);
8409 for (i = prevvalue; i <= ceilvalue; i++)
8410 if (isUPPER(i) && !ANYOF_BITMAP_TEST(ret,i)) {
8412 ANYOF_BITMAP_SET(ret, i);
8418 for (i = prevvalue; i <= ceilvalue; i++) {
8419 if (!ANYOF_BITMAP_TEST(ret,i)) {
8421 ANYOF_BITMAP_SET(ret, i);
8425 if (value > 255 || UTF) {
8426 const UV prevnatvalue = NATIVE_TO_UNI(prevvalue);
8427 const UV natvalue = NATIVE_TO_UNI(value);
8428 stored+=2; /* can't optimize this class */
8429 ANYOF_FLAGS(ret) |= ANYOF_UNICODE;
8430 if (prevnatvalue < natvalue) { /* what about > ? */
8431 Perl_sv_catpvf(aTHX_ listsv, "%04"UVxf"\t%04"UVxf"\n",
8432 prevnatvalue, natvalue);
8434 else if (prevnatvalue == natvalue) {
8435 Perl_sv_catpvf(aTHX_ listsv, "%04"UVxf"\n", natvalue);
8437 U8 foldbuf[UTF8_MAXBYTES_CASE+1];
8439 const UV f = to_uni_fold(natvalue, foldbuf, &foldlen);
8441 #ifdef EBCDIC /* RD t/uni/fold ff and 6b */
8442 if (RExC_precomp[0] == ':' &&
8443 RExC_precomp[1] == '[' &&
8444 (f == 0xDF || f == 0x92)) {
8445 f = NATIVE_TO_UNI(f);
8448 /* If folding and foldable and a single
8449 * character, insert also the folded version
8450 * to the charclass. */
8452 #ifdef EBCDIC /* RD tunifold ligatures s,t fb05, fb06 */
8453 if ((RExC_precomp[0] == ':' &&
8454 RExC_precomp[1] == '[' &&
8456 (value == 0xFB05 || value == 0xFB06))) ?
8457 foldlen == ((STRLEN)UNISKIP(f) - 1) :
8458 foldlen == (STRLEN)UNISKIP(f) )
8460 if (foldlen == (STRLEN)UNISKIP(f))
8462 Perl_sv_catpvf(aTHX_ listsv,
8465 /* Any multicharacter foldings
8466 * require the following transform:
8467 * [ABCDEF] -> (?:[ABCabcDEFd]|pq|rst)
8468 * where E folds into "pq" and F folds
8469 * into "rst", all other characters
8470 * fold to single characters. We save
8471 * away these multicharacter foldings,
8472 * to be later saved as part of the
8473 * additional "s" data. */
8476 if (!unicode_alternate)
8477 unicode_alternate = newAV();
8478 sv = newSVpvn_utf8((char*)foldbuf, foldlen,
8480 av_push(unicode_alternate, sv);
8484 /* If folding and the value is one of the Greek
8485 * sigmas insert a few more sigmas to make the
8486 * folding rules of the sigmas to work right.
8487 * Note that not all the possible combinations
8488 * are handled here: some of them are handled
8489 * by the standard folding rules, and some of
8490 * them (literal or EXACTF cases) are handled
8491 * during runtime in regexec.c:S_find_byclass(). */
8492 if (value == UNICODE_GREEK_SMALL_LETTER_FINAL_SIGMA) {
8493 Perl_sv_catpvf(aTHX_ listsv, "%04"UVxf"\n",
8494 (UV)UNICODE_GREEK_CAPITAL_LETTER_SIGMA);
8495 Perl_sv_catpvf(aTHX_ listsv, "%04"UVxf"\n",
8496 (UV)UNICODE_GREEK_SMALL_LETTER_SIGMA);
8498 else if (value == UNICODE_GREEK_CAPITAL_LETTER_SIGMA)
8499 Perl_sv_catpvf(aTHX_ listsv, "%04"UVxf"\n",
8500 (UV)UNICODE_GREEK_SMALL_LETTER_SIGMA);
8505 literal_endpoint = 0;
8509 range = 0; /* this range (if it was one) is done now */
8513 ANYOF_FLAGS(ret) |= ANYOF_LARGE;
8515 RExC_size += ANYOF_CLASS_ADD_SKIP;
8517 RExC_emit += ANYOF_CLASS_ADD_SKIP;
8523 /****** !SIZE_ONLY AFTER HERE *********/
8525 if( stored == 1 && (value < 128 || (value < 256 && !UTF))
8526 && !( ANYOF_FLAGS(ret) & ( ANYOF_FLAGS_ALL ^ ANYOF_FOLD ) )
8528 /* optimize single char class to an EXACT node
8529 but *only* when its not a UTF/high char */
8530 const char * cur_parse= RExC_parse;
8531 RExC_emit = (regnode *)orig_emit;
8532 RExC_parse = (char *)orig_parse;
8533 ret = reg_node(pRExC_state,
8534 (U8)((ANYOF_FLAGS(ret) & ANYOF_FOLD) ? EXACTF : EXACT));
8535 RExC_parse = (char *)cur_parse;
8536 *STRING(ret)= (char)value;
8538 RExC_emit += STR_SZ(1);
8539 SvREFCNT_dec(listsv);
8542 /* optimize case-insensitive simple patterns (e.g. /[a-z]/i) */
8543 if ( /* If the only flag is folding (plus possibly inversion). */
8544 ((ANYOF_FLAGS(ret) & (ANYOF_FLAGS_ALL ^ ANYOF_INVERT)) == ANYOF_FOLD)
8546 for (value = 0; value < 256; ++value) {
8547 if (ANYOF_BITMAP_TEST(ret, value)) {
8548 UV fold = PL_fold[value];
8551 ANYOF_BITMAP_SET(ret, fold);
8554 ANYOF_FLAGS(ret) &= ~ANYOF_FOLD;
8557 /* optimize inverted simple patterns (e.g. [^a-z]) */
8558 if (optimize_invert &&
8559 /* If the only flag is inversion. */
8560 (ANYOF_FLAGS(ret) & ANYOF_FLAGS_ALL) == ANYOF_INVERT) {
8561 for (value = 0; value < ANYOF_BITMAP_SIZE; ++value)
8562 ANYOF_BITMAP(ret)[value] ^= ANYOF_FLAGS_ALL;
8563 ANYOF_FLAGS(ret) = ANYOF_UNICODE_ALL;
8566 AV * const av = newAV();
8568 /* The 0th element stores the character class description
8569 * in its textual form: used later (regexec.c:Perl_regclass_swash())
8570 * to initialize the appropriate swash (which gets stored in
8571 * the 1st element), and also useful for dumping the regnode.
8572 * The 2nd element stores the multicharacter foldings,
8573 * used later (regexec.c:S_reginclass()). */
8574 av_store(av, 0, listsv);
8575 av_store(av, 1, NULL);
8576 av_store(av, 2, MUTABLE_SV(unicode_alternate));
8577 rv = newRV_noinc(MUTABLE_SV(av));
8578 n = add_data(pRExC_state, 1, "s");
8579 RExC_rxi->data->data[n] = (void*)rv;
8587 /* reg_skipcomment()
8589 Absorbs an /x style # comments from the input stream.
8590 Returns true if there is more text remaining in the stream.
8591 Will set the REG_SEEN_RUN_ON_COMMENT flag if the comment
8592 terminates the pattern without including a newline.
8594 Note its the callers responsibility to ensure that we are
8600 S_reg_skipcomment(pTHX_ RExC_state_t *pRExC_state)
8604 PERL_ARGS_ASSERT_REG_SKIPCOMMENT;
8606 while (RExC_parse < RExC_end)
8607 if (*RExC_parse++ == '\n') {
8612 /* we ran off the end of the pattern without ending
8613 the comment, so we have to add an \n when wrapping */
8614 RExC_seen |= REG_SEEN_RUN_ON_COMMENT;
8622 Advance that parse position, and optionall absorbs
8623 "whitespace" from the inputstream.
8625 Without /x "whitespace" means (?#...) style comments only,
8626 with /x this means (?#...) and # comments and whitespace proper.
8628 Returns the RExC_parse point from BEFORE the scan occurs.
8630 This is the /x friendly way of saying RExC_parse++.
8634 S_nextchar(pTHX_ RExC_state_t *pRExC_state)
8636 char* const retval = RExC_parse++;
8638 PERL_ARGS_ASSERT_NEXTCHAR;
8641 if (*RExC_parse == '(' && RExC_parse[1] == '?' &&
8642 RExC_parse[2] == '#') {
8643 while (*RExC_parse != ')') {
8644 if (RExC_parse == RExC_end)
8645 FAIL("Sequence (?#... not terminated");
8651 if (RExC_flags & RXf_PMf_EXTENDED) {
8652 if (isSPACE(*RExC_parse)) {
8656 else if (*RExC_parse == '#') {
8657 if ( reg_skipcomment( pRExC_state ) )
8666 - reg_node - emit a node
8668 STATIC regnode * /* Location. */
8669 S_reg_node(pTHX_ RExC_state_t *pRExC_state, U8 op)
8672 register regnode *ptr;
8673 regnode * const ret = RExC_emit;
8674 GET_RE_DEBUG_FLAGS_DECL;
8676 PERL_ARGS_ASSERT_REG_NODE;
8679 SIZE_ALIGN(RExC_size);
8683 if (RExC_emit >= RExC_emit_bound)
8684 Perl_croak(aTHX_ "panic: reg_node overrun trying to emit %d", op);
8686 NODE_ALIGN_FILL(ret);
8688 FILL_ADVANCE_NODE(ptr, op);
8689 #ifdef RE_TRACK_PATTERN_OFFSETS
8690 if (RExC_offsets) { /* MJD */
8691 MJD_OFFSET_DEBUG(("%s:%d: (op %s) %s %"UVuf" (len %"UVuf") (max %"UVuf").\n",
8692 "reg_node", __LINE__,
8694 (UV)(RExC_emit - RExC_emit_start) > RExC_offsets[0]
8695 ? "Overwriting end of array!\n" : "OK",
8696 (UV)(RExC_emit - RExC_emit_start),
8697 (UV)(RExC_parse - RExC_start),
8698 (UV)RExC_offsets[0]));
8699 Set_Node_Offset(RExC_emit, RExC_parse + (op == END));
8707 - reganode - emit a node with an argument
8709 STATIC regnode * /* Location. */
8710 S_reganode(pTHX_ RExC_state_t *pRExC_state, U8 op, U32 arg)
8713 register regnode *ptr;
8714 regnode * const ret = RExC_emit;
8715 GET_RE_DEBUG_FLAGS_DECL;
8717 PERL_ARGS_ASSERT_REGANODE;
8720 SIZE_ALIGN(RExC_size);
8725 assert(2==regarglen[op]+1);
8727 Anything larger than this has to allocate the extra amount.
8728 If we changed this to be:
8730 RExC_size += (1 + regarglen[op]);
8732 then it wouldn't matter. Its not clear what side effect
8733 might come from that so its not done so far.
8738 if (RExC_emit >= RExC_emit_bound)
8739 Perl_croak(aTHX_ "panic: reg_node overrun trying to emit %d", op);
8741 NODE_ALIGN_FILL(ret);
8743 FILL_ADVANCE_NODE_ARG(ptr, op, arg);
8744 #ifdef RE_TRACK_PATTERN_OFFSETS
8745 if (RExC_offsets) { /* MJD */
8746 MJD_OFFSET_DEBUG(("%s(%d): (op %s) %s %"UVuf" <- %"UVuf" (max %"UVuf").\n",
8750 (UV)(RExC_emit - RExC_emit_start) > RExC_offsets[0] ?
8751 "Overwriting end of array!\n" : "OK",
8752 (UV)(RExC_emit - RExC_emit_start),
8753 (UV)(RExC_parse - RExC_start),
8754 (UV)RExC_offsets[0]));
8755 Set_Cur_Node_Offset;
8763 - reguni - emit (if appropriate) a Unicode character
8766 S_reguni(pTHX_ const RExC_state_t *pRExC_state, UV uv, char* s)
8770 PERL_ARGS_ASSERT_REGUNI;
8772 return SIZE_ONLY ? UNISKIP(uv) : (uvchr_to_utf8((U8*)s, uv) - (U8*)s);
8776 - reginsert - insert an operator in front of already-emitted operand
8778 * Means relocating the operand.
8781 S_reginsert(pTHX_ RExC_state_t *pRExC_state, U8 op, regnode *opnd, U32 depth)
8784 register regnode *src;
8785 register regnode *dst;
8786 register regnode *place;
8787 const int offset = regarglen[(U8)op];
8788 const int size = NODE_STEP_REGNODE + offset;
8789 GET_RE_DEBUG_FLAGS_DECL;
8791 PERL_ARGS_ASSERT_REGINSERT;
8792 PERL_UNUSED_ARG(depth);
8793 /* (PL_regkind[(U8)op] == CURLY ? EXTRA_STEP_2ARGS : 0); */
8794 DEBUG_PARSE_FMT("inst"," - %s",PL_reg_name[op]);
8803 if (RExC_open_parens) {
8805 /*DEBUG_PARSE_FMT("inst"," - %"IVdf, (IV)RExC_npar);*/
8806 for ( paren=0 ; paren < RExC_npar ; paren++ ) {
8807 if ( RExC_open_parens[paren] >= opnd ) {
8808 /*DEBUG_PARSE_FMT("open"," - %d",size);*/
8809 RExC_open_parens[paren] += size;
8811 /*DEBUG_PARSE_FMT("open"," - %s","ok");*/
8813 if ( RExC_close_parens[paren] >= opnd ) {
8814 /*DEBUG_PARSE_FMT("close"," - %d",size);*/
8815 RExC_close_parens[paren] += size;
8817 /*DEBUG_PARSE_FMT("close"," - %s","ok");*/
8822 while (src > opnd) {
8823 StructCopy(--src, --dst, regnode);
8824 #ifdef RE_TRACK_PATTERN_OFFSETS
8825 if (RExC_offsets) { /* MJD 20010112 */
8826 MJD_OFFSET_DEBUG(("%s(%d): (op %s) %s copy %"UVuf" -> %"UVuf" (max %"UVuf").\n",
8830 (UV)(dst - RExC_emit_start) > RExC_offsets[0]
8831 ? "Overwriting end of array!\n" : "OK",
8832 (UV)(src - RExC_emit_start),
8833 (UV)(dst - RExC_emit_start),
8834 (UV)RExC_offsets[0]));
8835 Set_Node_Offset_To_R(dst-RExC_emit_start, Node_Offset(src));
8836 Set_Node_Length_To_R(dst-RExC_emit_start, Node_Length(src));
8842 place = opnd; /* Op node, where operand used to be. */
8843 #ifdef RE_TRACK_PATTERN_OFFSETS
8844 if (RExC_offsets) { /* MJD */
8845 MJD_OFFSET_DEBUG(("%s(%d): (op %s) %s %"UVuf" <- %"UVuf" (max %"UVuf").\n",
8849 (UV)(place - RExC_emit_start) > RExC_offsets[0]
8850 ? "Overwriting end of array!\n" : "OK",
8851 (UV)(place - RExC_emit_start),
8852 (UV)(RExC_parse - RExC_start),
8853 (UV)RExC_offsets[0]));
8854 Set_Node_Offset(place, RExC_parse);
8855 Set_Node_Length(place, 1);
8858 src = NEXTOPER(place);
8859 FILL_ADVANCE_NODE(place, op);
8860 Zero(src, offset, regnode);
8864 - regtail - set the next-pointer at the end of a node chain of p to val.
8865 - SEE ALSO: regtail_study
8867 /* TODO: All three parms should be const */
8869 S_regtail(pTHX_ RExC_state_t *pRExC_state, regnode *p, const regnode *val,U32 depth)
8872 register regnode *scan;
8873 GET_RE_DEBUG_FLAGS_DECL;
8875 PERL_ARGS_ASSERT_REGTAIL;
8877 PERL_UNUSED_ARG(depth);
8883 /* Find last node. */
8886 regnode * const temp = regnext(scan);
8888 SV * const mysv=sv_newmortal();
8889 DEBUG_PARSE_MSG((scan==p ? "tail" : ""));
8890 regprop(RExC_rx, mysv, scan);
8891 PerlIO_printf(Perl_debug_log, "~ %s (%d) %s %s\n",
8892 SvPV_nolen_const(mysv), REG_NODE_NUM(scan),
8893 (temp == NULL ? "->" : ""),
8894 (temp == NULL ? PL_reg_name[OP(val)] : "")
8902 if (reg_off_by_arg[OP(scan)]) {
8903 ARG_SET(scan, val - scan);
8906 NEXT_OFF(scan) = val - scan;
8912 - regtail_study - set the next-pointer at the end of a node chain of p to val.
8913 - Look for optimizable sequences at the same time.
8914 - currently only looks for EXACT chains.
8916 This is expermental code. The idea is to use this routine to perform
8917 in place optimizations on branches and groups as they are constructed,
8918 with the long term intention of removing optimization from study_chunk so
8919 that it is purely analytical.
8921 Currently only used when in DEBUG mode. The macro REGTAIL_STUDY() is used
8922 to control which is which.
8925 /* TODO: All four parms should be const */
8928 S_regtail_study(pTHX_ RExC_state_t *pRExC_state, regnode *p, const regnode *val,U32 depth)
8931 register regnode *scan;
8933 #ifdef EXPERIMENTAL_INPLACESCAN
8936 GET_RE_DEBUG_FLAGS_DECL;
8938 PERL_ARGS_ASSERT_REGTAIL_STUDY;
8944 /* Find last node. */
8948 regnode * const temp = regnext(scan);
8949 #ifdef EXPERIMENTAL_INPLACESCAN
8950 if (PL_regkind[OP(scan)] == EXACT)
8951 if (join_exact(pRExC_state,scan,&min,1,val,depth+1))
8959 if( exact == PSEUDO )
8961 else if ( exact != OP(scan) )
8970 SV * const mysv=sv_newmortal();
8971 DEBUG_PARSE_MSG((scan==p ? "tsdy" : ""));
8972 regprop(RExC_rx, mysv, scan);
8973 PerlIO_printf(Perl_debug_log, "~ %s (%d) -> %s\n",
8974 SvPV_nolen_const(mysv),
8976 PL_reg_name[exact]);
8983 SV * const mysv_val=sv_newmortal();
8984 DEBUG_PARSE_MSG("");
8985 regprop(RExC_rx, mysv_val, val);
8986 PerlIO_printf(Perl_debug_log, "~ attach to %s (%"IVdf") offset to %"IVdf"\n",
8987 SvPV_nolen_const(mysv_val),
8988 (IV)REG_NODE_NUM(val),
8992 if (reg_off_by_arg[OP(scan)]) {
8993 ARG_SET(scan, val - scan);
8996 NEXT_OFF(scan) = val - scan;
9004 - regcurly - a little FSA that accepts {\d+,?\d*}
9006 #ifndef PERL_IN_XSUB_RE
9008 Perl_regcurly(register const char *s)
9010 PERL_ARGS_ASSERT_REGCURLY;
9029 - regdump - dump a regexp onto Perl_debug_log in vaguely comprehensible form
9033 S_regdump_extflags(pTHX_ const char *lead, const U32 flags)
9038 for (bit=0; bit<32; bit++) {
9039 if (flags & (1<<bit)) {
9041 PerlIO_printf(Perl_debug_log, "%s",lead);
9042 PerlIO_printf(Perl_debug_log, "%s ",PL_reg_extflags_name[bit]);
9047 PerlIO_printf(Perl_debug_log, "\n");
9049 PerlIO_printf(Perl_debug_log, "%s[none-set]\n",lead);
9055 Perl_regdump(pTHX_ const regexp *r)
9059 SV * const sv = sv_newmortal();
9060 SV *dsv= sv_newmortal();
9062 GET_RE_DEBUG_FLAGS_DECL;
9064 PERL_ARGS_ASSERT_REGDUMP;
9066 (void)dumpuntil(r, ri->program, ri->program + 1, NULL, NULL, sv, 0, 0);
9068 /* Header fields of interest. */
9069 if (r->anchored_substr) {
9070 RE_PV_QUOTED_DECL(s, 0, dsv, SvPVX_const(r->anchored_substr),
9071 RE_SV_DUMPLEN(r->anchored_substr), 30);
9072 PerlIO_printf(Perl_debug_log,
9073 "anchored %s%s at %"IVdf" ",
9074 s, RE_SV_TAIL(r->anchored_substr),
9075 (IV)r->anchored_offset);
9076 } else if (r->anchored_utf8) {
9077 RE_PV_QUOTED_DECL(s, 1, dsv, SvPVX_const(r->anchored_utf8),
9078 RE_SV_DUMPLEN(r->anchored_utf8), 30);
9079 PerlIO_printf(Perl_debug_log,
9080 "anchored utf8 %s%s at %"IVdf" ",
9081 s, RE_SV_TAIL(r->anchored_utf8),
9082 (IV)r->anchored_offset);
9084 if (r->float_substr) {
9085 RE_PV_QUOTED_DECL(s, 0, dsv, SvPVX_const(r->float_substr),
9086 RE_SV_DUMPLEN(r->float_substr), 30);
9087 PerlIO_printf(Perl_debug_log,
9088 "floating %s%s at %"IVdf"..%"UVuf" ",
9089 s, RE_SV_TAIL(r->float_substr),
9090 (IV)r->float_min_offset, (UV)r->float_max_offset);
9091 } else if (r->float_utf8) {
9092 RE_PV_QUOTED_DECL(s, 1, dsv, SvPVX_const(r->float_utf8),
9093 RE_SV_DUMPLEN(r->float_utf8), 30);
9094 PerlIO_printf(Perl_debug_log,
9095 "floating utf8 %s%s at %"IVdf"..%"UVuf" ",
9096 s, RE_SV_TAIL(r->float_utf8),
9097 (IV)r->float_min_offset, (UV)r->float_max_offset);
9099 if (r->check_substr || r->check_utf8)
9100 PerlIO_printf(Perl_debug_log,
9102 (r->check_substr == r->float_substr
9103 && r->check_utf8 == r->float_utf8
9104 ? "(checking floating" : "(checking anchored"));
9105 if (r->extflags & RXf_NOSCAN)
9106 PerlIO_printf(Perl_debug_log, " noscan");
9107 if (r->extflags & RXf_CHECK_ALL)
9108 PerlIO_printf(Perl_debug_log, " isall");
9109 if (r->check_substr || r->check_utf8)
9110 PerlIO_printf(Perl_debug_log, ") ");
9112 if (ri->regstclass) {
9113 regprop(r, sv, ri->regstclass);
9114 PerlIO_printf(Perl_debug_log, "stclass %s ", SvPVX_const(sv));
9116 if (r->extflags & RXf_ANCH) {
9117 PerlIO_printf(Perl_debug_log, "anchored");
9118 if (r->extflags & RXf_ANCH_BOL)
9119 PerlIO_printf(Perl_debug_log, "(BOL)");
9120 if (r->extflags & RXf_ANCH_MBOL)
9121 PerlIO_printf(Perl_debug_log, "(MBOL)");
9122 if (r->extflags & RXf_ANCH_SBOL)
9123 PerlIO_printf(Perl_debug_log, "(SBOL)");
9124 if (r->extflags & RXf_ANCH_GPOS)
9125 PerlIO_printf(Perl_debug_log, "(GPOS)");
9126 PerlIO_putc(Perl_debug_log, ' ');
9128 if (r->extflags & RXf_GPOS_SEEN)
9129 PerlIO_printf(Perl_debug_log, "GPOS:%"UVuf" ", (UV)r->gofs);
9130 if (r->intflags & PREGf_SKIP)
9131 PerlIO_printf(Perl_debug_log, "plus ");
9132 if (r->intflags & PREGf_IMPLICIT)
9133 PerlIO_printf(Perl_debug_log, "implicit ");
9134 PerlIO_printf(Perl_debug_log, "minlen %"IVdf" ", (IV)r->minlen);
9135 if (r->extflags & RXf_EVAL_SEEN)
9136 PerlIO_printf(Perl_debug_log, "with eval ");
9137 PerlIO_printf(Perl_debug_log, "\n");
9138 DEBUG_FLAGS_r(regdump_extflags("r->extflags: ",r->extflags));
9140 PERL_ARGS_ASSERT_REGDUMP;
9141 PERL_UNUSED_CONTEXT;
9143 #endif /* DEBUGGING */
9147 - regprop - printable representation of opcode
9149 #define EMIT_ANYOF_TEST_SEPARATOR(do_sep,sv,flags) \
9152 Perl_sv_catpvf(aTHX_ sv,"%s][%s",PL_colors[1],PL_colors[0]); \
9153 if (flags & ANYOF_INVERT) \
9154 /*make sure the invert info is in each */ \
9155 sv_catpvs(sv, "^"); \
9161 Perl_regprop(pTHX_ const regexp *prog, SV *sv, const regnode *o)
9166 RXi_GET_DECL(prog,progi);
9167 GET_RE_DEBUG_FLAGS_DECL;
9169 PERL_ARGS_ASSERT_REGPROP;
9173 if (OP(o) > REGNODE_MAX) /* regnode.type is unsigned */
9174 /* It would be nice to FAIL() here, but this may be called from
9175 regexec.c, and it would be hard to supply pRExC_state. */
9176 Perl_croak(aTHX_ "Corrupted regexp opcode %d > %d", (int)OP(o), (int)REGNODE_MAX);
9177 sv_catpv(sv, PL_reg_name[OP(o)]); /* Take off const! */
9179 k = PL_regkind[OP(o)];
9183 /* Using is_utf8_string() (via PERL_PV_UNI_DETECT)
9184 * is a crude hack but it may be the best for now since
9185 * we have no flag "this EXACTish node was UTF-8"
9187 pv_pretty(sv, STRING(o), STR_LEN(o), 60, PL_colors[0], PL_colors[1],
9188 PERL_PV_ESCAPE_UNI_DETECT |
9189 PERL_PV_PRETTY_ELLIPSES |
9190 PERL_PV_PRETTY_LTGT |
9191 PERL_PV_PRETTY_NOCLEAR
9193 } else if (k == TRIE) {
9194 /* print the details of the trie in dumpuntil instead, as
9195 * progi->data isn't available here */
9196 const char op = OP(o);
9197 const U32 n = ARG(o);
9198 const reg_ac_data * const ac = IS_TRIE_AC(op) ?
9199 (reg_ac_data *)progi->data->data[n] :
9201 const reg_trie_data * const trie
9202 = (reg_trie_data*)progi->data->data[!IS_TRIE_AC(op) ? n : ac->trie];
9204 Perl_sv_catpvf(aTHX_ sv, "-%s",PL_reg_name[o->flags]);
9205 DEBUG_TRIE_COMPILE_r(
9206 Perl_sv_catpvf(aTHX_ sv,
9207 "<S:%"UVuf"/%"IVdf" W:%"UVuf" L:%"UVuf"/%"UVuf" C:%"UVuf"/%"UVuf">",
9208 (UV)trie->startstate,
9209 (IV)trie->statecount-1, /* -1 because of the unused 0 element */
9210 (UV)trie->wordcount,
9213 (UV)TRIE_CHARCOUNT(trie),
9214 (UV)trie->uniquecharcount
9217 if ( IS_ANYOF_TRIE(op) || trie->bitmap ) {
9219 int rangestart = -1;
9220 U8* bitmap = IS_ANYOF_TRIE(op) ? (U8*)ANYOF_BITMAP(o) : (U8*)TRIE_BITMAP(trie);
9222 for (i = 0; i <= 256; i++) {
9223 if (i < 256 && BITMAP_TEST(bitmap,i)) {
9224 if (rangestart == -1)
9226 } else if (rangestart != -1) {
9227 if (i <= rangestart + 3)
9228 for (; rangestart < i; rangestart++)
9229 put_byte(sv, rangestart);
9231 put_byte(sv, rangestart);
9233 put_byte(sv, i - 1);
9241 } else if (k == CURLY) {
9242 if (OP(o) == CURLYM || OP(o) == CURLYN || OP(o) == CURLYX)
9243 Perl_sv_catpvf(aTHX_ sv, "[%d]", o->flags); /* Parenth number */
9244 Perl_sv_catpvf(aTHX_ sv, " {%d,%d}", ARG1(o), ARG2(o));
9246 else if (k == WHILEM && o->flags) /* Ordinal/of */
9247 Perl_sv_catpvf(aTHX_ sv, "[%d/%d]", o->flags & 0xf, o->flags>>4);
9248 else if (k == REF || k == OPEN || k == CLOSE || k == GROUPP || OP(o)==ACCEPT) {
9249 Perl_sv_catpvf(aTHX_ sv, "%d", (int)ARG(o)); /* Parenth number */
9250 if ( RXp_PAREN_NAMES(prog) ) {
9251 if ( k != REF || OP(o) < NREF) {
9252 AV *list= MUTABLE_AV(progi->data->data[progi->name_list_idx]);
9253 SV **name= av_fetch(list, ARG(o), 0 );
9255 Perl_sv_catpvf(aTHX_ sv, " '%"SVf"'", SVfARG(*name));
9258 AV *list= MUTABLE_AV(progi->data->data[ progi->name_list_idx ]);
9259 SV *sv_dat= MUTABLE_SV(progi->data->data[ ARG( o ) ]);
9260 I32 *nums=(I32*)SvPVX(sv_dat);
9261 SV **name= av_fetch(list, nums[0], 0 );
9264 for ( n=0; n<SvIVX(sv_dat); n++ ) {
9265 Perl_sv_catpvf(aTHX_ sv, "%s%"IVdf,
9266 (n ? "," : ""), (IV)nums[n]);
9268 Perl_sv_catpvf(aTHX_ sv, " '%"SVf"'", SVfARG(*name));
9272 } else if (k == GOSUB)
9273 Perl_sv_catpvf(aTHX_ sv, "%d[%+d]", (int)ARG(o),(int)ARG2L(o)); /* Paren and offset */
9274 else if (k == VERB) {
9276 Perl_sv_catpvf(aTHX_ sv, ":%"SVf,
9277 SVfARG((MUTABLE_SV(progi->data->data[ ARG( o ) ]))));
9278 } else if (k == LOGICAL)
9279 Perl_sv_catpvf(aTHX_ sv, "[%d]", o->flags); /* 2: embedded, otherwise 1 */
9280 else if (k == FOLDCHAR)
9281 Perl_sv_catpvf(aTHX_ sv, "[0x%"UVXf"]", PTR2UV(ARG(o)) );
9282 else if (k == ANYOF) {
9283 int i, rangestart = -1;
9284 const U8 flags = ANYOF_FLAGS(o);
9287 /* Should be synchronized with * ANYOF_ #xdefines in regcomp.h */
9288 static const char * const anyofs[] = {
9321 if (flags & ANYOF_LOCALE)
9322 sv_catpvs(sv, "{loc}");
9323 if (flags & ANYOF_FOLD)
9324 sv_catpvs(sv, "{i}");
9325 Perl_sv_catpvf(aTHX_ sv, "[%s", PL_colors[0]);
9326 if (flags & ANYOF_INVERT)
9329 /* output what the standard cp 0-255 bitmap matches */
9330 for (i = 0; i <= 256; i++) {
9331 if (i < 256 && ANYOF_BITMAP_TEST(o,i)) {
9332 if (rangestart == -1)
9334 } else if (rangestart != -1) {
9335 if (i <= rangestart + 3)
9336 for (; rangestart < i; rangestart++)
9337 put_byte(sv, rangestart);
9339 put_byte(sv, rangestart);
9341 put_byte(sv, i - 1);
9348 EMIT_ANYOF_TEST_SEPARATOR(do_sep,sv,flags);
9349 /* output any special charclass tests (used mostly under use locale) */
9350 if (o->flags & ANYOF_CLASS)
9351 for (i = 0; i < (int)(sizeof(anyofs)/sizeof(char*)); i++)
9352 if (ANYOF_CLASS_TEST(o,i)) {
9353 sv_catpv(sv, anyofs[i]);
9357 EMIT_ANYOF_TEST_SEPARATOR(do_sep,sv,flags);
9359 /* output information about the unicode matching */
9360 if (flags & ANYOF_UNICODE)
9361 sv_catpvs(sv, "{unicode}");
9362 else if (flags & ANYOF_UNICODE_ALL)
9363 sv_catpvs(sv, "{unicode_all}");
9367 SV * const sw = regclass_swash(prog, o, FALSE, &lv, 0);
9371 U8 s[UTF8_MAXBYTES_CASE+1];
9373 for (i = 0; i <= 256; i++) { /* just the first 256 */
9374 uvchr_to_utf8(s, i);
9376 if (i < 256 && swash_fetch(sw, s, TRUE)) {
9377 if (rangestart == -1)
9379 } else if (rangestart != -1) {
9380 if (i <= rangestart + 3)
9381 for (; rangestart < i; rangestart++) {
9382 const U8 * const e = uvchr_to_utf8(s,rangestart);
9384 for(p = s; p < e; p++)
9388 const U8 *e = uvchr_to_utf8(s,rangestart);
9390 for (p = s; p < e; p++)
9393 e = uvchr_to_utf8(s, i-1);
9394 for (p = s; p < e; p++)
9401 sv_catpvs(sv, "..."); /* et cetera */
9405 char *s = savesvpv(lv);
9406 char * const origs = s;
9408 while (*s && *s != '\n')
9412 const char * const t = ++s;
9430 Perl_sv_catpvf(aTHX_ sv, "%s]", PL_colors[1]);
9432 else if (k == BRANCHJ && (OP(o) == UNLESSM || OP(o) == IFMATCH))
9433 Perl_sv_catpvf(aTHX_ sv, "[%d]", -(o->flags));
9435 PERL_UNUSED_CONTEXT;
9436 PERL_UNUSED_ARG(sv);
9438 PERL_UNUSED_ARG(prog);
9439 #endif /* DEBUGGING */
9443 Perl_re_intuit_string(pTHX_ REGEXP * const r)
9444 { /* Assume that RE_INTUIT is set */
9446 struct regexp *const prog = (struct regexp *)SvANY(r);
9447 GET_RE_DEBUG_FLAGS_DECL;
9449 PERL_ARGS_ASSERT_RE_INTUIT_STRING;
9450 PERL_UNUSED_CONTEXT;
9454 const char * const s = SvPV_nolen_const(prog->check_substr
9455 ? prog->check_substr : prog->check_utf8);
9457 if (!PL_colorset) reginitcolors();
9458 PerlIO_printf(Perl_debug_log,
9459 "%sUsing REx %ssubstr:%s \"%s%.60s%s%s\"\n",
9461 prog->check_substr ? "" : "utf8 ",
9462 PL_colors[5],PL_colors[0],
9465 (strlen(s) > 60 ? "..." : ""));
9468 return prog->check_substr ? prog->check_substr : prog->check_utf8;
9474 handles refcounting and freeing the perl core regexp structure. When
9475 it is necessary to actually free the structure the first thing it
9476 does is call the 'free' method of the regexp_engine associated to to
9477 the regexp, allowing the handling of the void *pprivate; member
9478 first. (This routine is not overridable by extensions, which is why
9479 the extensions free is called first.)
9481 See regdupe and regdupe_internal if you change anything here.
9483 #ifndef PERL_IN_XSUB_RE
9485 Perl_pregfree(pTHX_ REGEXP *r)
9491 Perl_pregfree2(pTHX_ REGEXP *rx)
9494 struct regexp *const r = (struct regexp *)SvANY(rx);
9495 GET_RE_DEBUG_FLAGS_DECL;
9497 PERL_ARGS_ASSERT_PREGFREE2;
9500 ReREFCNT_dec(r->mother_re);
9502 CALLREGFREE_PVT(rx); /* free the private data */
9503 SvREFCNT_dec(RXp_PAREN_NAMES(r));
9506 SvREFCNT_dec(r->anchored_substr);
9507 SvREFCNT_dec(r->anchored_utf8);
9508 SvREFCNT_dec(r->float_substr);
9509 SvREFCNT_dec(r->float_utf8);
9510 Safefree(r->substrs);
9512 RX_MATCH_COPY_FREE(rx);
9513 #ifdef PERL_OLD_COPY_ON_WRITE
9514 SvREFCNT_dec(r->saved_copy);
9521 This is a hacky workaround to the structural issue of match results
9522 being stored in the regexp structure which is in turn stored in
9523 PL_curpm/PL_reg_curpm. The problem is that due to qr// the pattern
9524 could be PL_curpm in multiple contexts, and could require multiple
9525 result sets being associated with the pattern simultaneously, such
9526 as when doing a recursive match with (??{$qr})
9528 The solution is to make a lightweight copy of the regexp structure
9529 when a qr// is returned from the code executed by (??{$qr}) this
9530 lightweight copy doesnt actually own any of its data except for
9531 the starp/end and the actual regexp structure itself.
9537 Perl_reg_temp_copy (pTHX_ REGEXP *ret_x, REGEXP *rx)
9540 struct regexp *const r = (struct regexp *)SvANY(rx);
9541 register const I32 npar = r->nparens+1;
9543 PERL_ARGS_ASSERT_REG_TEMP_COPY;
9546 ret_x = (REGEXP*) newSV_type(SVt_REGEXP);
9547 ret = (struct regexp *)SvANY(ret_x);
9549 (void)ReREFCNT_inc(rx);
9550 /* We can take advantage of the existing "copied buffer" mechanism in SVs
9551 by pointing directly at the buffer, but flagging that the allocated
9552 space in the copy is zero. As we've just done a struct copy, it's now
9553 a case of zero-ing that, rather than copying the current length. */
9554 SvPV_set(ret_x, RX_WRAPPED(rx));
9555 SvFLAGS(ret_x) |= SvFLAGS(rx) & (SVf_POK|SVp_POK|SVf_UTF8);
9556 memcpy(&(ret->xpv_cur), &(r->xpv_cur),
9557 sizeof(regexp) - STRUCT_OFFSET(regexp, xpv_cur));
9558 SvLEN_set(ret_x, 0);
9559 SvSTASH_set(ret_x, NULL);
9560 SvMAGIC_set(ret_x, NULL);
9561 Newx(ret->offs, npar, regexp_paren_pair);
9562 Copy(r->offs, ret->offs, npar, regexp_paren_pair);
9564 Newx(ret->substrs, 1, struct reg_substr_data);
9565 StructCopy(r->substrs, ret->substrs, struct reg_substr_data);
9567 SvREFCNT_inc_void(ret->anchored_substr);
9568 SvREFCNT_inc_void(ret->anchored_utf8);
9569 SvREFCNT_inc_void(ret->float_substr);
9570 SvREFCNT_inc_void(ret->float_utf8);
9572 /* check_substr and check_utf8, if non-NULL, point to either their
9573 anchored or float namesakes, and don't hold a second reference. */
9575 RX_MATCH_COPIED_off(ret_x);
9576 #ifdef PERL_OLD_COPY_ON_WRITE
9577 ret->saved_copy = NULL;
9579 ret->mother_re = rx;
9585 /* regfree_internal()
9587 Free the private data in a regexp. This is overloadable by
9588 extensions. Perl takes care of the regexp structure in pregfree(),
9589 this covers the *pprivate pointer which technically perldoesnt
9590 know about, however of course we have to handle the
9591 regexp_internal structure when no extension is in use.
9593 Note this is called before freeing anything in the regexp
9598 Perl_regfree_internal(pTHX_ REGEXP * const rx)
9601 struct regexp *const r = (struct regexp *)SvANY(rx);
9603 GET_RE_DEBUG_FLAGS_DECL;
9605 PERL_ARGS_ASSERT_REGFREE_INTERNAL;
9611 SV *dsv= sv_newmortal();
9612 RE_PV_QUOTED_DECL(s, RX_UTF8(rx),
9613 dsv, RX_PRECOMP(rx), RX_PRELEN(rx), 60);
9614 PerlIO_printf(Perl_debug_log,"%sFreeing REx:%s %s\n",
9615 PL_colors[4],PL_colors[5],s);
9618 #ifdef RE_TRACK_PATTERN_OFFSETS
9620 Safefree(ri->u.offsets); /* 20010421 MJD */
9623 int n = ri->data->count;
9624 PAD* new_comppad = NULL;
9629 /* If you add a ->what type here, update the comment in regcomp.h */
9630 switch (ri->data->what[n]) {
9635 SvREFCNT_dec(MUTABLE_SV(ri->data->data[n]));
9638 Safefree(ri->data->data[n]);
9641 new_comppad = MUTABLE_AV(ri->data->data[n]);
9644 if (new_comppad == NULL)
9645 Perl_croak(aTHX_ "panic: pregfree comppad");
9646 PAD_SAVE_LOCAL(old_comppad,
9647 /* Watch out for global destruction's random ordering. */
9648 (SvTYPE(new_comppad) == SVt_PVAV) ? new_comppad : NULL
9651 refcnt = OpREFCNT_dec((OP_4tree*)ri->data->data[n]);
9654 op_free((OP_4tree*)ri->data->data[n]);
9656 PAD_RESTORE_LOCAL(old_comppad);
9657 SvREFCNT_dec(MUTABLE_SV(new_comppad));
9663 { /* Aho Corasick add-on structure for a trie node.
9664 Used in stclass optimization only */
9666 reg_ac_data *aho=(reg_ac_data*)ri->data->data[n];
9668 refcount = --aho->refcount;
9671 PerlMemShared_free(aho->states);
9672 PerlMemShared_free(aho->fail);
9673 /* do this last!!!! */
9674 PerlMemShared_free(ri->data->data[n]);
9675 PerlMemShared_free(ri->regstclass);
9681 /* trie structure. */
9683 reg_trie_data *trie=(reg_trie_data*)ri->data->data[n];
9685 refcount = --trie->refcount;
9688 PerlMemShared_free(trie->charmap);
9689 PerlMemShared_free(trie->states);
9690 PerlMemShared_free(trie->trans);
9692 PerlMemShared_free(trie->bitmap);
9694 PerlMemShared_free(trie->jump);
9695 PerlMemShared_free(trie->wordinfo);
9696 /* do this last!!!! */
9697 PerlMemShared_free(ri->data->data[n]);
9702 Perl_croak(aTHX_ "panic: regfree data code '%c'", ri->data->what[n]);
9705 Safefree(ri->data->what);
9712 #define av_dup_inc(s,t) MUTABLE_AV(sv_dup_inc((const SV *)s,t))
9713 #define hv_dup_inc(s,t) MUTABLE_HV(sv_dup_inc((const SV *)s,t))
9714 #define SAVEPVN(p,n) ((p) ? savepvn(p,n) : NULL)
9717 re_dup - duplicate a regexp.
9719 This routine is expected to clone a given regexp structure. It is only
9720 compiled under USE_ITHREADS.
9722 After all of the core data stored in struct regexp is duplicated
9723 the regexp_engine.dupe method is used to copy any private data
9724 stored in the *pprivate pointer. This allows extensions to handle
9725 any duplication it needs to do.
9727 See pregfree() and regfree_internal() if you change anything here.
9729 #if defined(USE_ITHREADS)
9730 #ifndef PERL_IN_XSUB_RE
9732 Perl_re_dup_guts(pTHX_ const REGEXP *sstr, REGEXP *dstr, CLONE_PARAMS *param)
9736 const struct regexp *r = (const struct regexp *)SvANY(sstr);
9737 struct regexp *ret = (struct regexp *)SvANY(dstr);
9739 PERL_ARGS_ASSERT_RE_DUP_GUTS;
9741 npar = r->nparens+1;
9742 Newx(ret->offs, npar, regexp_paren_pair);
9743 Copy(r->offs, ret->offs, npar, regexp_paren_pair);
9745 /* no need to copy these */
9746 Newx(ret->swap, npar, regexp_paren_pair);
9750 /* Do it this way to avoid reading from *r after the StructCopy().
9751 That way, if any of the sv_dup_inc()s dislodge *r from the L1
9752 cache, it doesn't matter. */
9753 const bool anchored = r->check_substr
9754 ? r->check_substr == r->anchored_substr
9755 : r->check_utf8 == r->anchored_utf8;
9756 Newx(ret->substrs, 1, struct reg_substr_data);
9757 StructCopy(r->substrs, ret->substrs, struct reg_substr_data);
9759 ret->anchored_substr = sv_dup_inc(ret->anchored_substr, param);
9760 ret->anchored_utf8 = sv_dup_inc(ret->anchored_utf8, param);
9761 ret->float_substr = sv_dup_inc(ret->float_substr, param);
9762 ret->float_utf8 = sv_dup_inc(ret->float_utf8, param);
9764 /* check_substr and check_utf8, if non-NULL, point to either their
9765 anchored or float namesakes, and don't hold a second reference. */
9767 if (ret->check_substr) {
9769 assert(r->check_utf8 == r->anchored_utf8);
9770 ret->check_substr = ret->anchored_substr;
9771 ret->check_utf8 = ret->anchored_utf8;
9773 assert(r->check_substr == r->float_substr);
9774 assert(r->check_utf8 == r->float_utf8);
9775 ret->check_substr = ret->float_substr;
9776 ret->check_utf8 = ret->float_utf8;
9778 } else if (ret->check_utf8) {
9780 ret->check_utf8 = ret->anchored_utf8;
9782 ret->check_utf8 = ret->float_utf8;
9787 RXp_PAREN_NAMES(ret) = hv_dup_inc(RXp_PAREN_NAMES(ret), param);
9790 RXi_SET(ret,CALLREGDUPE_PVT(dstr,param));
9792 if (RX_MATCH_COPIED(dstr))
9793 ret->subbeg = SAVEPVN(ret->subbeg, ret->sublen);
9796 #ifdef PERL_OLD_COPY_ON_WRITE
9797 ret->saved_copy = NULL;
9800 if (ret->mother_re) {
9801 if (SvPVX_const(dstr) == SvPVX_const(ret->mother_re)) {
9802 /* Our storage points directly to our mother regexp, but that's
9803 1: a buffer in a different thread
9804 2: something we no longer hold a reference on
9805 so we need to copy it locally. */
9806 /* Note we need to sue SvCUR() on our mother_re, because it, in
9807 turn, may well be pointing to its own mother_re. */
9808 SvPV_set(dstr, SAVEPVN(SvPVX_const(ret->mother_re),
9809 SvCUR(ret->mother_re)+1));
9810 SvLEN_set(dstr, SvCUR(ret->mother_re)+1);
9812 ret->mother_re = NULL;
9816 #endif /* PERL_IN_XSUB_RE */
9821 This is the internal complement to regdupe() which is used to copy
9822 the structure pointed to by the *pprivate pointer in the regexp.
9823 This is the core version of the extension overridable cloning hook.
9824 The regexp structure being duplicated will be copied by perl prior
9825 to this and will be provided as the regexp *r argument, however
9826 with the /old/ structures pprivate pointer value. Thus this routine
9827 may override any copying normally done by perl.
9829 It returns a pointer to the new regexp_internal structure.
9833 Perl_regdupe_internal(pTHX_ REGEXP * const rx, CLONE_PARAMS *param)
9836 struct regexp *const r = (struct regexp *)SvANY(rx);
9837 regexp_internal *reti;
9841 PERL_ARGS_ASSERT_REGDUPE_INTERNAL;
9843 npar = r->nparens+1;
9846 Newxc(reti, sizeof(regexp_internal) + len*sizeof(regnode), char, regexp_internal);
9847 Copy(ri->program, reti->program, len+1, regnode);
9850 reti->regstclass = NULL;
9854 const int count = ri->data->count;
9857 Newxc(d, sizeof(struct reg_data) + count*sizeof(void *),
9858 char, struct reg_data);
9859 Newx(d->what, count, U8);
9862 for (i = 0; i < count; i++) {
9863 d->what[i] = ri->data->what[i];
9864 switch (d->what[i]) {
9865 /* legal options are one of: sSfpontTua
9866 see also regcomp.h and pregfree() */
9867 case 'a': /* actually an AV, but the dup function is identical. */
9870 case 'p': /* actually an AV, but the dup function is identical. */
9871 case 'u': /* actually an HV, but the dup function is identical. */
9872 d->data[i] = sv_dup_inc((const SV *)ri->data->data[i], param);
9875 /* This is cheating. */
9876 Newx(d->data[i], 1, struct regnode_charclass_class);
9877 StructCopy(ri->data->data[i], d->data[i],
9878 struct regnode_charclass_class);
9879 reti->regstclass = (regnode*)d->data[i];
9882 /* Compiled op trees are readonly and in shared memory,
9883 and can thus be shared without duplication. */
9885 d->data[i] = (void*)OpREFCNT_inc((OP*)ri->data->data[i]);
9889 /* Trie stclasses are readonly and can thus be shared
9890 * without duplication. We free the stclass in pregfree
9891 * when the corresponding reg_ac_data struct is freed.
9893 reti->regstclass= ri->regstclass;
9897 ((reg_trie_data*)ri->data->data[i])->refcount++;
9901 d->data[i] = ri->data->data[i];
9904 Perl_croak(aTHX_ "panic: re_dup unknown data code '%c'", ri->data->what[i]);
9913 reti->name_list_idx = ri->name_list_idx;
9915 #ifdef RE_TRACK_PATTERN_OFFSETS
9916 if (ri->u.offsets) {
9917 Newx(reti->u.offsets, 2*len+1, U32);
9918 Copy(ri->u.offsets, reti->u.offsets, 2*len+1, U32);
9921 SetProgLen(reti,len);
9927 #endif /* USE_ITHREADS */
9929 #ifndef PERL_IN_XSUB_RE
9932 - regnext - dig the "next" pointer out of a node
9935 Perl_regnext(pTHX_ register regnode *p)
9938 register I32 offset;
9943 if (OP(p) > REGNODE_MAX) { /* regnode.type is unsigned */
9944 Perl_croak(aTHX_ "Corrupted regexp opcode %d > %d", (int)OP(p), (int)REGNODE_MAX);
9947 offset = (reg_off_by_arg[OP(p)] ? ARG(p) : NEXT_OFF(p));
9956 S_re_croak2(pTHX_ const char* pat1,const char* pat2,...)
9959 STRLEN l1 = strlen(pat1);
9960 STRLEN l2 = strlen(pat2);
9963 const char *message;
9965 PERL_ARGS_ASSERT_RE_CROAK2;
9971 Copy(pat1, buf, l1 , char);
9972 Copy(pat2, buf + l1, l2 , char);
9973 buf[l1 + l2] = '\n';
9974 buf[l1 + l2 + 1] = '\0';
9976 /* ANSI variant takes additional second argument */
9977 va_start(args, pat2);
9981 msv = vmess(buf, &args);
9983 message = SvPV_const(msv,l1);
9986 Copy(message, buf, l1 , char);
9987 buf[l1-1] = '\0'; /* Overwrite \n */
9988 Perl_croak(aTHX_ "%s", buf);
9991 /* XXX Here's a total kludge. But we need to re-enter for swash routines. */
9993 #ifndef PERL_IN_XSUB_RE
9995 Perl_save_re_context(pTHX)
9999 struct re_save_state *state;
10001 SAVEVPTR(PL_curcop);
10002 SSGROW(SAVESTACK_ALLOC_FOR_RE_SAVE_STATE + 1);
10004 state = (struct re_save_state *)(PL_savestack + PL_savestack_ix);
10005 PL_savestack_ix += SAVESTACK_ALLOC_FOR_RE_SAVE_STATE;
10006 SSPUSHUV(SAVEt_RE_STATE);
10008 Copy(&PL_reg_state, state, 1, struct re_save_state);
10010 PL_reg_start_tmp = 0;
10011 PL_reg_start_tmpl = 0;
10012 PL_reg_oldsaved = NULL;
10013 PL_reg_oldsavedlen = 0;
10014 PL_reg_maxiter = 0;
10015 PL_reg_leftiter = 0;
10016 PL_reg_poscache = NULL;
10017 PL_reg_poscache_size = 0;
10018 #ifdef PERL_OLD_COPY_ON_WRITE
10022 /* Save $1..$n (#18107: UTF-8 s/(\w+)/uc($1)/e); AMS 20021106. */
10024 const REGEXP * const rx = PM_GETRE(PL_curpm);
10027 for (i = 1; i <= RX_NPARENS(rx); i++) {
10028 char digits[TYPE_CHARS(long)];
10029 const STRLEN len = my_snprintf(digits, sizeof(digits), "%lu", (long)i);
10030 GV *const *const gvp
10031 = (GV**)hv_fetch(PL_defstash, digits, len, 0);
10034 GV * const gv = *gvp;
10035 if (SvTYPE(gv) == SVt_PVGV && GvSV(gv))
10045 clear_re(pTHX_ void *r)
10048 ReREFCNT_dec((REGEXP *)r);
10054 S_put_byte(pTHX_ SV *sv, int c)
10056 PERL_ARGS_ASSERT_PUT_BYTE;
10058 /* Our definition of isPRINT() ignores locales, so only bytes that are
10059 not part of UTF-8 are considered printable. I assume that the same
10060 holds for UTF-EBCDIC.
10061 Also, code point 255 is not printable in either (it's E0 in EBCDIC,
10062 which Wikipedia says:
10064 EO, or Eight Ones, is an 8-bit EBCDIC character code represented as all
10065 ones (binary 1111 1111, hexadecimal FF). It is similar, but not
10066 identical, to the ASCII delete (DEL) or rubout control character.
10067 ) So the old condition can be simplified to !isPRINT(c) */
10069 Perl_sv_catpvf(aTHX_ sv, "\\%o", c);
10071 const char string = c;
10072 if (c == '-' || c == ']' || c == '\\' || c == '^')
10073 sv_catpvs(sv, "\\");
10074 sv_catpvn(sv, &string, 1);
10079 #define CLEAR_OPTSTART \
10080 if (optstart) STMT_START { \
10081 DEBUG_OPTIMISE_r(PerlIO_printf(Perl_debug_log, " (%"IVdf" nodes)\n", (IV)(node - optstart))); \
10085 #define DUMPUNTIL(b,e) CLEAR_OPTSTART; node=dumpuntil(r,start,(b),(e),last,sv,indent+1,depth+1);
10087 STATIC const regnode *
10088 S_dumpuntil(pTHX_ const regexp *r, const regnode *start, const regnode *node,
10089 const regnode *last, const regnode *plast,
10090 SV* sv, I32 indent, U32 depth)
10093 register U8 op = PSEUDO; /* Arbitrary non-END op. */
10094 register const regnode *next;
10095 const regnode *optstart= NULL;
10097 RXi_GET_DECL(r,ri);
10098 GET_RE_DEBUG_FLAGS_DECL;
10100 PERL_ARGS_ASSERT_DUMPUNTIL;
10102 #ifdef DEBUG_DUMPUNTIL
10103 PerlIO_printf(Perl_debug_log, "--- %d : %d - %d - %d\n",indent,node-start,
10104 last ? last-start : 0,plast ? plast-start : 0);
10107 if (plast && plast < last)
10110 while (PL_regkind[op] != END && (!last || node < last)) {
10111 /* While that wasn't END last time... */
10114 if (op == CLOSE || op == WHILEM)
10116 next = regnext((regnode *)node);
10119 if (OP(node) == OPTIMIZED) {
10120 if (!optstart && RE_DEBUG_FLAG(RE_DEBUG_COMPILE_OPTIMISE))
10127 regprop(r, sv, node);
10128 PerlIO_printf(Perl_debug_log, "%4"IVdf":%*s%s", (IV)(node - start),
10129 (int)(2*indent + 1), "", SvPVX_const(sv));
10131 if (OP(node) != OPTIMIZED) {
10132 if (next == NULL) /* Next ptr. */
10133 PerlIO_printf(Perl_debug_log, " (0)");
10134 else if (PL_regkind[(U8)op] == BRANCH && PL_regkind[OP(next)] != BRANCH )
10135 PerlIO_printf(Perl_debug_log, " (FAIL)");
10137 PerlIO_printf(Perl_debug_log, " (%"IVdf")", (IV)(next - start));
10138 (void)PerlIO_putc(Perl_debug_log, '\n');
10142 if (PL_regkind[(U8)op] == BRANCHJ) {
10145 register const regnode *nnode = (OP(next) == LONGJMP
10146 ? regnext((regnode *)next)
10148 if (last && nnode > last)
10150 DUMPUNTIL(NEXTOPER(NEXTOPER(node)), nnode);
10153 else if (PL_regkind[(U8)op] == BRANCH) {
10155 DUMPUNTIL(NEXTOPER(node), next);
10157 else if ( PL_regkind[(U8)op] == TRIE ) {
10158 const regnode *this_trie = node;
10159 const char op = OP(node);
10160 const U32 n = ARG(node);
10161 const reg_ac_data * const ac = op>=AHOCORASICK ?
10162 (reg_ac_data *)ri->data->data[n] :
10164 const reg_trie_data * const trie =
10165 (reg_trie_data*)ri->data->data[op<AHOCORASICK ? n : ac->trie];
10167 AV *const trie_words = MUTABLE_AV(ri->data->data[n + TRIE_WORDS_OFFSET]);
10169 const regnode *nextbranch= NULL;
10172 for (word_idx= 0; word_idx < (I32)trie->wordcount; word_idx++) {
10173 SV ** const elem_ptr = av_fetch(trie_words,word_idx,0);
10175 PerlIO_printf(Perl_debug_log, "%*s%s ",
10176 (int)(2*(indent+3)), "",
10177 elem_ptr ? pv_pretty(sv, SvPV_nolen_const(*elem_ptr), SvCUR(*elem_ptr), 60,
10178 PL_colors[0], PL_colors[1],
10179 (SvUTF8(*elem_ptr) ? PERL_PV_ESCAPE_UNI : 0) |
10180 PERL_PV_PRETTY_ELLIPSES |
10181 PERL_PV_PRETTY_LTGT
10186 U16 dist= trie->jump[word_idx+1];
10187 PerlIO_printf(Perl_debug_log, "(%"UVuf")\n",
10188 (UV)((dist ? this_trie + dist : next) - start));
10191 nextbranch= this_trie + trie->jump[0];
10192 DUMPUNTIL(this_trie + dist, nextbranch);
10194 if (nextbranch && PL_regkind[OP(nextbranch)]==BRANCH)
10195 nextbranch= regnext((regnode *)nextbranch);
10197 PerlIO_printf(Perl_debug_log, "\n");
10200 if (last && next > last)
10205 else if ( op == CURLY ) { /* "next" might be very big: optimizer */
10206 DUMPUNTIL(NEXTOPER(node) + EXTRA_STEP_2ARGS,
10207 NEXTOPER(node) + EXTRA_STEP_2ARGS + 1);
10209 else if (PL_regkind[(U8)op] == CURLY && op != CURLYX) {
10211 DUMPUNTIL(NEXTOPER(node) + EXTRA_STEP_2ARGS, next);
10213 else if ( op == PLUS || op == STAR) {
10214 DUMPUNTIL(NEXTOPER(node), NEXTOPER(node) + 1);
10216 else if (op == ANYOF) {
10217 /* arglen 1 + class block */
10218 node += 1 + ((ANYOF_FLAGS(node) & ANYOF_LARGE)
10219 ? ANYOF_CLASS_SKIP : ANYOF_SKIP);
10220 node = NEXTOPER(node);
10222 else if (PL_regkind[(U8)op] == EXACT) {
10223 /* Literal string, where present. */
10224 node += NODE_SZ_STR(node) - 1;
10225 node = NEXTOPER(node);
10228 node = NEXTOPER(node);
10229 node += regarglen[(U8)op];
10231 if (op == CURLYX || op == OPEN)
10235 #ifdef DEBUG_DUMPUNTIL
10236 PerlIO_printf(Perl_debug_log, "--- %d\n", (int)indent);
10241 #endif /* DEBUGGING */
10245 * c-indentation-style: bsd
10246 * c-basic-offset: 4
10247 * indent-tabs-mode: t
10250 * ex: set ts=8 sts=4 sw=4 noet:
projects / perl / modules / re-engine-Hooks.git / blob