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
75 #undef PERL_IN_XSUB_RE
76 #define PERL_IN_XSUB_RE 1
78 #undef PERL_IN_XSUB_RE
80 #ifndef PERL_IN_XSUB_RE
85 #ifdef PERL_IN_XSUB_RE
87 extern const struct regexp_engine my_reg_engine;
92 #include "dquote_static.c"
93 #include "charclass_invlists.h"
94 #include "inline_invlist.c"
95 #include "unicode_constants.h"
97 #define HAS_NONLATIN1_FOLD_CLOSURE(i) _HAS_NONLATIN1_FOLD_CLOSURE_ONLY_FOR_USE_BY_REGCOMP_DOT_C_AND_REGEXEC_DOT_C(i)
98 #define IS_NON_FINAL_FOLD(c) _IS_NON_FINAL_FOLD_ONLY_FOR_USE_BY_REGCOMP_DOT_C(c)
99 #define IS_IN_SOME_FOLD_L1(c) _IS_IN_SOME_FOLD_ONLY_FOR_USE_BY_REGCOMP_DOT_C(c)
106 # if defined(BUGGY_MSC6)
107 /* MSC 6.00A breaks on op/regexp.t test 85 unless we turn this off */
108 # pragma optimize("a",off)
109 /* But MSC 6.00A is happy with 'w', for aliases only across function calls*/
110 # pragma optimize("w",on )
111 # endif /* BUGGY_MSC6 */
115 #define STATIC static
119 typedef struct RExC_state_t {
120 U32 flags; /* RXf_* are we folding, multilining? */
121 U32 pm_flags; /* PMf_* stuff from the calling PMOP */
122 char *precomp; /* uncompiled string. */
123 REGEXP *rx_sv; /* The SV that is the regexp. */
124 regexp *rx; /* perl core regexp structure */
125 regexp_internal *rxi; /* internal data for regexp object pprivate field */
126 char *start; /* Start of input for compile */
127 char *end; /* End of input for compile */
128 char *parse; /* Input-scan pointer. */
129 I32 whilem_seen; /* number of WHILEM in this expr */
130 regnode *emit_start; /* Start of emitted-code area */
131 regnode *emit_bound; /* First regnode outside of the allocated space */
132 regnode *emit; /* Code-emit pointer; ®dummy = don't = compiling */
133 I32 naughty; /* How bad is this pattern? */
134 I32 sawback; /* Did we see \1, ...? */
136 I32 size; /* Code size. */
137 I32 npar; /* Capture buffer count, (OPEN). */
138 I32 cpar; /* Capture buffer count, (CLOSE). */
139 I32 nestroot; /* root parens we are in - used by accept */
142 regnode **open_parens; /* pointers to open parens */
143 regnode **close_parens; /* pointers to close parens */
144 regnode *opend; /* END node in program */
145 I32 utf8; /* whether the pattern is utf8 or not */
146 I32 orig_utf8; /* whether the pattern was originally in utf8 */
147 /* XXX use this for future optimisation of case
148 * where pattern must be upgraded to utf8. */
149 I32 uni_semantics; /* If a d charset modifier should use unicode
150 rules, even if the pattern is not in
152 HV *paren_names; /* Paren names */
154 regnode **recurse; /* Recurse regops */
155 I32 recurse_count; /* Number of recurse regops */
158 I32 override_recoding;
159 I32 in_multi_char_class;
160 struct reg_code_block *code_blocks; /* positions of literal (?{})
162 int num_code_blocks; /* size of code_blocks[] */
163 int code_index; /* next code_blocks[] slot */
165 char *starttry; /* -Dr: where regtry was called. */
166 #define RExC_starttry (pRExC_state->starttry)
168 SV *runtime_code_qr; /* qr with the runtime code blocks */
170 const char *lastparse;
172 AV *paren_name_list; /* idx -> name */
173 #define RExC_lastparse (pRExC_state->lastparse)
174 #define RExC_lastnum (pRExC_state->lastnum)
175 #define RExC_paren_name_list (pRExC_state->paren_name_list)
179 #define RExC_flags (pRExC_state->flags)
180 #define RExC_pm_flags (pRExC_state->pm_flags)
181 #define RExC_precomp (pRExC_state->precomp)
182 #define RExC_rx_sv (pRExC_state->rx_sv)
183 #define RExC_rx (pRExC_state->rx)
184 #define RExC_rxi (pRExC_state->rxi)
185 #define RExC_start (pRExC_state->start)
186 #define RExC_end (pRExC_state->end)
187 #define RExC_parse (pRExC_state->parse)
188 #define RExC_whilem_seen (pRExC_state->whilem_seen)
189 #ifdef RE_TRACK_PATTERN_OFFSETS
190 #define RExC_offsets (pRExC_state->rxi->u.offsets) /* I am not like the others */
192 #define RExC_emit (pRExC_state->emit)
193 #define RExC_emit_start (pRExC_state->emit_start)
194 #define RExC_emit_bound (pRExC_state->emit_bound)
195 #define RExC_naughty (pRExC_state->naughty)
196 #define RExC_sawback (pRExC_state->sawback)
197 #define RExC_seen (pRExC_state->seen)
198 #define RExC_size (pRExC_state->size)
199 #define RExC_npar (pRExC_state->npar)
200 #define RExC_nestroot (pRExC_state->nestroot)
201 #define RExC_extralen (pRExC_state->extralen)
202 #define RExC_seen_zerolen (pRExC_state->seen_zerolen)
203 #define RExC_utf8 (pRExC_state->utf8)
204 #define RExC_uni_semantics (pRExC_state->uni_semantics)
205 #define RExC_orig_utf8 (pRExC_state->orig_utf8)
206 #define RExC_open_parens (pRExC_state->open_parens)
207 #define RExC_close_parens (pRExC_state->close_parens)
208 #define RExC_opend (pRExC_state->opend)
209 #define RExC_paren_names (pRExC_state->paren_names)
210 #define RExC_recurse (pRExC_state->recurse)
211 #define RExC_recurse_count (pRExC_state->recurse_count)
212 #define RExC_in_lookbehind (pRExC_state->in_lookbehind)
213 #define RExC_contains_locale (pRExC_state->contains_locale)
214 #define RExC_override_recoding (pRExC_state->override_recoding)
215 #define RExC_in_multi_char_class (pRExC_state->in_multi_char_class)
218 #define ISMULT1(c) ((c) == '*' || (c) == '+' || (c) == '?')
219 #define ISMULT2(s) ((*s) == '*' || (*s) == '+' || (*s) == '?' || \
220 ((*s) == '{' && regcurly(s, FALSE)))
223 #undef SPSTART /* dratted cpp namespace... */
226 * Flags to be passed up and down.
228 #define WORST 0 /* Worst case. */
229 #define HASWIDTH 0x01 /* Known to match non-null strings. */
231 /* Simple enough to be STAR/PLUS operand; in an EXACTish node must be a single
232 * character. (There needs to be a case: in the switch statement in regexec.c
233 * for any node marked SIMPLE.) Note that this is not the same thing as
236 #define SPSTART 0x04 /* Starts with * or + */
237 #define POSTPONED 0x08 /* (?1),(?&name), (??{...}) or similar */
238 #define TRYAGAIN 0x10 /* Weeded out a declaration. */
239 #define RESTART_UTF8 0x20 /* Restart, need to calcuate sizes as UTF-8 */
241 #define REG_NODE_NUM(x) ((x) ? (int)((x)-RExC_emit_start) : -1)
243 /* whether trie related optimizations are enabled */
244 #if PERL_ENABLE_EXTENDED_TRIE_OPTIMISATION
245 #define TRIE_STUDY_OPT
246 #define FULL_TRIE_STUDY
252 #define PBYTE(u8str,paren) ((U8*)(u8str))[(paren) >> 3]
253 #define PBITVAL(paren) (1 << ((paren) & 7))
254 #define PAREN_TEST(u8str,paren) ( PBYTE(u8str,paren) & PBITVAL(paren))
255 #define PAREN_SET(u8str,paren) PBYTE(u8str,paren) |= PBITVAL(paren)
256 #define PAREN_UNSET(u8str,paren) PBYTE(u8str,paren) &= (~PBITVAL(paren))
258 #define REQUIRE_UTF8 STMT_START { \
260 *flagp = RESTART_UTF8; \
265 /* This converts the named class defined in regcomp.h to its equivalent class
266 * number defined in handy.h. */
267 #define namedclass_to_classnum(class) ((int) ((class) / 2))
268 #define classnum_to_namedclass(classnum) ((classnum) * 2)
270 /* About scan_data_t.
272 During optimisation we recurse through the regexp program performing
273 various inplace (keyhole style) optimisations. In addition study_chunk
274 and scan_commit populate this data structure with information about
275 what strings MUST appear in the pattern. We look for the longest
276 string that must appear at a fixed location, and we look for the
277 longest string that may appear at a floating location. So for instance
282 Both 'FOO' and 'A' are fixed strings. Both 'B' and 'BAR' are floating
283 strings (because they follow a .* construct). study_chunk will identify
284 both FOO and BAR as being the longest fixed and floating strings respectively.
286 The strings can be composites, for instance
290 will result in a composite fixed substring 'foo'.
292 For each string some basic information is maintained:
294 - offset or min_offset
295 This is the position the string must appear at, or not before.
296 It also implicitly (when combined with minlenp) tells us how many
297 characters must match before the string we are searching for.
298 Likewise when combined with minlenp and the length of the string it
299 tells us how many characters must appear after the string we have
303 Only used for floating strings. This is the rightmost point that
304 the string can appear at. If set to I32 max it indicates that the
305 string can occur infinitely far to the right.
308 A pointer to the minimum number of characters of the pattern that the
309 string was found inside. This is important as in the case of positive
310 lookahead or positive lookbehind we can have multiple patterns
315 The minimum length of the pattern overall is 3, the minimum length
316 of the lookahead part is 3, but the minimum length of the part that
317 will actually match is 1. So 'FOO's minimum length is 3, but the
318 minimum length for the F is 1. This is important as the minimum length
319 is used to determine offsets in front of and behind the string being
320 looked for. Since strings can be composites this is the length of the
321 pattern at the time it was committed with a scan_commit. Note that
322 the length is calculated by study_chunk, so that the minimum lengths
323 are not known until the full pattern has been compiled, thus the
324 pointer to the value.
328 In the case of lookbehind the string being searched for can be
329 offset past the start point of the final matching string.
330 If this value was just blithely removed from the min_offset it would
331 invalidate some of the calculations for how many chars must match
332 before or after (as they are derived from min_offset and minlen and
333 the length of the string being searched for).
334 When the final pattern is compiled and the data is moved from the
335 scan_data_t structure into the regexp structure the information
336 about lookbehind is factored in, with the information that would
337 have been lost precalculated in the end_shift field for the
340 The fields pos_min and pos_delta are used to store the minimum offset
341 and the delta to the maximum offset at the current point in the pattern.
345 typedef struct scan_data_t {
346 /*I32 len_min; unused */
347 /*I32 len_delta; unused */
351 I32 last_end; /* min value, <0 unless valid. */
354 SV **longest; /* Either &l_fixed, or &l_float. */
355 SV *longest_fixed; /* longest fixed string found in pattern */
356 I32 offset_fixed; /* offset where it starts */
357 I32 *minlen_fixed; /* pointer to the minlen relevant to the string */
358 I32 lookbehind_fixed; /* is the position of the string modfied by LB */
359 SV *longest_float; /* longest floating string found in pattern */
360 I32 offset_float_min; /* earliest point in string it can appear */
361 I32 offset_float_max; /* latest point in string it can appear */
362 I32 *minlen_float; /* pointer to the minlen relevant to the string */
363 I32 lookbehind_float; /* is the position of the string modified by LB */
367 struct regnode_charclass_class *start_class;
371 * Forward declarations for pregcomp()'s friends.
374 static const scan_data_t zero_scan_data =
375 { 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0 ,0};
377 #define SF_BEFORE_EOL (SF_BEFORE_SEOL|SF_BEFORE_MEOL)
378 #define SF_BEFORE_SEOL 0x0001
379 #define SF_BEFORE_MEOL 0x0002
380 #define SF_FIX_BEFORE_EOL (SF_FIX_BEFORE_SEOL|SF_FIX_BEFORE_MEOL)
381 #define SF_FL_BEFORE_EOL (SF_FL_BEFORE_SEOL|SF_FL_BEFORE_MEOL)
384 # define SF_FIX_SHIFT_EOL (0+2)
385 # define SF_FL_SHIFT_EOL (0+4)
387 # define SF_FIX_SHIFT_EOL (+2)
388 # define SF_FL_SHIFT_EOL (+4)
391 #define SF_FIX_BEFORE_SEOL (SF_BEFORE_SEOL << SF_FIX_SHIFT_EOL)
392 #define SF_FIX_BEFORE_MEOL (SF_BEFORE_MEOL << SF_FIX_SHIFT_EOL)
394 #define SF_FL_BEFORE_SEOL (SF_BEFORE_SEOL << SF_FL_SHIFT_EOL)
395 #define SF_FL_BEFORE_MEOL (SF_BEFORE_MEOL << SF_FL_SHIFT_EOL) /* 0x20 */
396 #define SF_IS_INF 0x0040
397 #define SF_HAS_PAR 0x0080
398 #define SF_IN_PAR 0x0100
399 #define SF_HAS_EVAL 0x0200
400 #define SCF_DO_SUBSTR 0x0400
401 #define SCF_DO_STCLASS_AND 0x0800
402 #define SCF_DO_STCLASS_OR 0x1000
403 #define SCF_DO_STCLASS (SCF_DO_STCLASS_AND|SCF_DO_STCLASS_OR)
404 #define SCF_WHILEM_VISITED_POS 0x2000
406 #define SCF_TRIE_RESTUDY 0x4000 /* Do restudy? */
407 #define SCF_SEEN_ACCEPT 0x8000
409 #define UTF cBOOL(RExC_utf8)
411 /* The enums for all these are ordered so things work out correctly */
412 #define LOC (get_regex_charset(RExC_flags) == REGEX_LOCALE_CHARSET)
413 #define DEPENDS_SEMANTICS (get_regex_charset(RExC_flags) == REGEX_DEPENDS_CHARSET)
414 #define UNI_SEMANTICS (get_regex_charset(RExC_flags) == REGEX_UNICODE_CHARSET)
415 #define AT_LEAST_UNI_SEMANTICS (get_regex_charset(RExC_flags) >= REGEX_UNICODE_CHARSET)
416 #define ASCII_RESTRICTED (get_regex_charset(RExC_flags) == REGEX_ASCII_RESTRICTED_CHARSET)
417 #define AT_LEAST_ASCII_RESTRICTED (get_regex_charset(RExC_flags) >= REGEX_ASCII_RESTRICTED_CHARSET)
418 #define ASCII_FOLD_RESTRICTED (get_regex_charset(RExC_flags) == REGEX_ASCII_MORE_RESTRICTED_CHARSET)
420 #define FOLD cBOOL(RExC_flags & RXf_PMf_FOLD)
422 #define OOB_NAMEDCLASS -1
424 /* There is no code point that is out-of-bounds, so this is problematic. But
425 * its only current use is to initialize a variable that is always set before
427 #define OOB_UNICODE 0xDEADBEEF
429 #define CHR_SVLEN(sv) (UTF ? sv_len_utf8(sv) : SvCUR(sv))
430 #define CHR_DIST(a,b) (UTF ? utf8_distance(a,b) : a - b)
433 /* length of regex to show in messages that don't mark a position within */
434 #define RegexLengthToShowInErrorMessages 127
437 * If MARKER[12] are adjusted, be sure to adjust the constants at the top
438 * of t/op/regmesg.t, the tests in t/op/re_tests, and those in
439 * op/pragma/warn/regcomp.
441 #define MARKER1 "<-- HERE" /* marker as it appears in the description */
442 #define MARKER2 " <-- HERE " /* marker as it appears within the regex */
444 #define REPORT_LOCATION " in regex; marked by " MARKER1 " in m/%.*s" MARKER2 "%s/"
447 * Calls SAVEDESTRUCTOR_X if needed, then calls Perl_croak with the given
448 * arg. Show regex, up to a maximum length. If it's too long, chop and add
451 #define _FAIL(code) STMT_START { \
452 const char *ellipses = ""; \
453 IV len = RExC_end - RExC_precomp; \
456 SAVEFREESV(RExC_rx_sv); \
457 if (len > RegexLengthToShowInErrorMessages) { \
458 /* chop 10 shorter than the max, to ensure meaning of "..." */ \
459 len = RegexLengthToShowInErrorMessages - 10; \
465 #define FAIL(msg) _FAIL( \
466 Perl_croak(aTHX_ "%s in regex m/%.*s%s/", \
467 msg, (int)len, RExC_precomp, ellipses))
469 #define FAIL2(msg,arg) _FAIL( \
470 Perl_croak(aTHX_ msg " in regex m/%.*s%s/", \
471 arg, (int)len, RExC_precomp, ellipses))
474 * Simple_vFAIL -- like FAIL, but marks the current location in the scan
476 #define Simple_vFAIL(m) STMT_START { \
477 const IV offset = RExC_parse - RExC_precomp; \
478 Perl_croak(aTHX_ "%s" REPORT_LOCATION, \
479 m, (int)offset, RExC_precomp, RExC_precomp + offset); \
483 * Calls SAVEDESTRUCTOR_X if needed, then Simple_vFAIL()
485 #define vFAIL(m) STMT_START { \
487 SAVEFREESV(RExC_rx_sv); \
492 * Like Simple_vFAIL(), but accepts two arguments.
494 #define Simple_vFAIL2(m,a1) STMT_START { \
495 const IV offset = RExC_parse - RExC_precomp; \
496 S_re_croak2(aTHX_ m, REPORT_LOCATION, a1, \
497 (int)offset, RExC_precomp, RExC_precomp + offset); \
501 * Calls SAVEDESTRUCTOR_X if needed, then Simple_vFAIL2().
503 #define vFAIL2(m,a1) STMT_START { \
505 SAVEFREESV(RExC_rx_sv); \
506 Simple_vFAIL2(m, a1); \
511 * Like Simple_vFAIL(), but accepts three arguments.
513 #define Simple_vFAIL3(m, a1, a2) STMT_START { \
514 const IV offset = RExC_parse - RExC_precomp; \
515 S_re_croak2(aTHX_ m, REPORT_LOCATION, a1, a2, \
516 (int)offset, RExC_precomp, RExC_precomp + offset); \
520 * Calls SAVEDESTRUCTOR_X if needed, then Simple_vFAIL3().
522 #define vFAIL3(m,a1,a2) STMT_START { \
524 SAVEFREESV(RExC_rx_sv); \
525 Simple_vFAIL3(m, a1, a2); \
529 * Like Simple_vFAIL(), but accepts four arguments.
531 #define Simple_vFAIL4(m, a1, a2, a3) STMT_START { \
532 const IV offset = RExC_parse - RExC_precomp; \
533 S_re_croak2(aTHX_ m, REPORT_LOCATION, a1, a2, a3, \
534 (int)offset, RExC_precomp, RExC_precomp + offset); \
537 #define vFAIL4(m,a1,a2,a3) STMT_START { \
539 SAVEFREESV(RExC_rx_sv); \
540 Simple_vFAIL4(m, a1, a2, a3); \
543 /* m is not necessarily a "literal string", in this macro */
544 #define reg_warn_non_literal_string(loc, m) STMT_START { \
545 const IV offset = loc - RExC_precomp; \
546 Perl_warner(aTHX_ packWARN(WARN_REGEXP), "%s" REPORT_LOCATION, \
547 m, (int)offset, RExC_precomp, RExC_precomp + offset); \
550 #define ckWARNreg(loc,m) STMT_START { \
551 const IV offset = loc - RExC_precomp; \
552 Perl_ck_warner(aTHX_ packWARN(WARN_REGEXP), m REPORT_LOCATION, \
553 (int)offset, RExC_precomp, RExC_precomp + offset); \
556 #define vWARN_dep(loc, m) STMT_START { \
557 const IV offset = loc - RExC_precomp; \
558 Perl_warner(aTHX_ packWARN(WARN_DEPRECATED), m REPORT_LOCATION, \
559 (int)offset, RExC_precomp, RExC_precomp + offset); \
562 #define ckWARNdep(loc,m) STMT_START { \
563 const IV offset = loc - RExC_precomp; \
564 Perl_ck_warner_d(aTHX_ packWARN(WARN_DEPRECATED), \
566 (int)offset, RExC_precomp, RExC_precomp + offset); \
569 #define ckWARNregdep(loc,m) STMT_START { \
570 const IV offset = loc - RExC_precomp; \
571 Perl_ck_warner_d(aTHX_ packWARN2(WARN_DEPRECATED, WARN_REGEXP), \
573 (int)offset, RExC_precomp, RExC_precomp + offset); \
576 #define ckWARN2regdep(loc,m, a1) STMT_START { \
577 const IV offset = loc - RExC_precomp; \
578 Perl_ck_warner_d(aTHX_ packWARN2(WARN_DEPRECATED, WARN_REGEXP), \
580 a1, (int)offset, RExC_precomp, RExC_precomp + offset); \
583 #define ckWARN2reg(loc, m, a1) STMT_START { \
584 const IV offset = loc - RExC_precomp; \
585 Perl_ck_warner(aTHX_ packWARN(WARN_REGEXP), m REPORT_LOCATION, \
586 a1, (int)offset, RExC_precomp, RExC_precomp + offset); \
589 #define vWARN3(loc, m, a1, a2) STMT_START { \
590 const IV offset = loc - RExC_precomp; \
591 Perl_warner(aTHX_ packWARN(WARN_REGEXP), m REPORT_LOCATION, \
592 a1, a2, (int)offset, RExC_precomp, RExC_precomp + offset); \
595 #define ckWARN3reg(loc, m, a1, a2) STMT_START { \
596 const IV offset = loc - RExC_precomp; \
597 Perl_ck_warner(aTHX_ packWARN(WARN_REGEXP), m REPORT_LOCATION, \
598 a1, a2, (int)offset, RExC_precomp, RExC_precomp + offset); \
601 #define vWARN4(loc, m, a1, a2, a3) STMT_START { \
602 const IV offset = loc - RExC_precomp; \
603 Perl_warner(aTHX_ packWARN(WARN_REGEXP), m REPORT_LOCATION, \
604 a1, a2, a3, (int)offset, RExC_precomp, RExC_precomp + offset); \
607 #define ckWARN4reg(loc, m, a1, a2, a3) STMT_START { \
608 const IV offset = loc - RExC_precomp; \
609 Perl_ck_warner(aTHX_ packWARN(WARN_REGEXP), m REPORT_LOCATION, \
610 a1, a2, a3, (int)offset, RExC_precomp, RExC_precomp + offset); \
613 #define vWARN5(loc, m, a1, a2, a3, a4) STMT_START { \
614 const IV offset = loc - RExC_precomp; \
615 Perl_warner(aTHX_ packWARN(WARN_REGEXP), m REPORT_LOCATION, \
616 a1, a2, a3, a4, (int)offset, RExC_precomp, RExC_precomp + offset); \
620 /* Allow for side effects in s */
621 #define REGC(c,s) STMT_START { \
622 if (!SIZE_ONLY) *(s) = (c); else (void)(s); \
625 /* Macros for recording node offsets. 20001227 mjd@plover.com
626 * Nodes are numbered 1, 2, 3, 4. Node #n's position is recorded in
627 * element 2*n-1 of the array. Element #2n holds the byte length node #n.
628 * Element 0 holds the number n.
629 * Position is 1 indexed.
631 #ifndef RE_TRACK_PATTERN_OFFSETS
632 #define Set_Node_Offset_To_R(node,byte)
633 #define Set_Node_Offset(node,byte)
634 #define Set_Cur_Node_Offset
635 #define Set_Node_Length_To_R(node,len)
636 #define Set_Node_Length(node,len)
637 #define Set_Node_Cur_Length(node)
638 #define Node_Offset(n)
639 #define Node_Length(n)
640 #define Set_Node_Offset_Length(node,offset,len)
641 #define ProgLen(ri) ri->u.proglen
642 #define SetProgLen(ri,x) ri->u.proglen = x
644 #define ProgLen(ri) ri->u.offsets[0]
645 #define SetProgLen(ri,x) ri->u.offsets[0] = x
646 #define Set_Node_Offset_To_R(node,byte) STMT_START { \
648 MJD_OFFSET_DEBUG(("** (%d) offset of node %d is %d.\n", \
649 __LINE__, (int)(node), (int)(byte))); \
651 Perl_croak(aTHX_ "value of node is %d in Offset macro", (int)(node)); \
653 RExC_offsets[2*(node)-1] = (byte); \
658 #define Set_Node_Offset(node,byte) \
659 Set_Node_Offset_To_R((node)-RExC_emit_start, (byte)-RExC_start)
660 #define Set_Cur_Node_Offset Set_Node_Offset(RExC_emit, RExC_parse)
662 #define Set_Node_Length_To_R(node,len) STMT_START { \
664 MJD_OFFSET_DEBUG(("** (%d) size of node %d is %d.\n", \
665 __LINE__, (int)(node), (int)(len))); \
667 Perl_croak(aTHX_ "value of node is %d in Length macro", (int)(node)); \
669 RExC_offsets[2*(node)] = (len); \
674 #define Set_Node_Length(node,len) \
675 Set_Node_Length_To_R((node)-RExC_emit_start, len)
676 #define Set_Cur_Node_Length(len) Set_Node_Length(RExC_emit, len)
677 #define Set_Node_Cur_Length(node) \
678 Set_Node_Length(node, RExC_parse - parse_start)
680 /* Get offsets and lengths */
681 #define Node_Offset(n) (RExC_offsets[2*((n)-RExC_emit_start)-1])
682 #define Node_Length(n) (RExC_offsets[2*((n)-RExC_emit_start)])
684 #define Set_Node_Offset_Length(node,offset,len) STMT_START { \
685 Set_Node_Offset_To_R((node)-RExC_emit_start, (offset)); \
686 Set_Node_Length_To_R((node)-RExC_emit_start, (len)); \
690 #if PERL_ENABLE_EXPERIMENTAL_REGEX_OPTIMISATIONS
691 #define EXPERIMENTAL_INPLACESCAN
692 #endif /*PERL_ENABLE_EXPERIMENTAL_REGEX_OPTIMISATIONS*/
694 #define DEBUG_STUDYDATA(str,data,depth) \
695 DEBUG_OPTIMISE_MORE_r(if(data){ \
696 PerlIO_printf(Perl_debug_log, \
697 "%*s" str "Pos:%"IVdf"/%"IVdf \
698 " Flags: 0x%"UVXf" Whilem_c: %"IVdf" Lcp: %"IVdf" %s", \
699 (int)(depth)*2, "", \
700 (IV)((data)->pos_min), \
701 (IV)((data)->pos_delta), \
702 (UV)((data)->flags), \
703 (IV)((data)->whilem_c), \
704 (IV)((data)->last_closep ? *((data)->last_closep) : -1), \
705 is_inf ? "INF " : "" \
707 if ((data)->last_found) \
708 PerlIO_printf(Perl_debug_log, \
709 "Last:'%s' %"IVdf":%"IVdf"/%"IVdf" %sFixed:'%s' @ %"IVdf \
710 " %sFloat: '%s' @ %"IVdf"/%"IVdf"", \
711 SvPVX_const((data)->last_found), \
712 (IV)((data)->last_end), \
713 (IV)((data)->last_start_min), \
714 (IV)((data)->last_start_max), \
715 ((data)->longest && \
716 (data)->longest==&((data)->longest_fixed)) ? "*" : "", \
717 SvPVX_const((data)->longest_fixed), \
718 (IV)((data)->offset_fixed), \
719 ((data)->longest && \
720 (data)->longest==&((data)->longest_float)) ? "*" : "", \
721 SvPVX_const((data)->longest_float), \
722 (IV)((data)->offset_float_min), \
723 (IV)((data)->offset_float_max) \
725 PerlIO_printf(Perl_debug_log,"\n"); \
728 /* Mark that we cannot extend a found fixed substring at this point.
729 Update the longest found anchored substring and the longest found
730 floating substrings if needed. */
733 S_scan_commit(pTHX_ const RExC_state_t *pRExC_state, scan_data_t *data, I32 *minlenp, int is_inf)
735 const STRLEN l = CHR_SVLEN(data->last_found);
736 const STRLEN old_l = CHR_SVLEN(*data->longest);
737 GET_RE_DEBUG_FLAGS_DECL;
739 PERL_ARGS_ASSERT_SCAN_COMMIT;
741 if ((l >= old_l) && ((l > old_l) || (data->flags & SF_BEFORE_EOL))) {
742 SvSetMagicSV(*data->longest, data->last_found);
743 if (*data->longest == data->longest_fixed) {
744 data->offset_fixed = l ? data->last_start_min : data->pos_min;
745 if (data->flags & SF_BEFORE_EOL)
747 |= ((data->flags & SF_BEFORE_EOL) << SF_FIX_SHIFT_EOL);
749 data->flags &= ~SF_FIX_BEFORE_EOL;
750 data->minlen_fixed=minlenp;
751 data->lookbehind_fixed=0;
753 else { /* *data->longest == data->longest_float */
754 data->offset_float_min = l ? data->last_start_min : data->pos_min;
755 data->offset_float_max = (l
756 ? data->last_start_max
757 : (data->pos_delta == I32_MAX ? I32_MAX : data->pos_min + data->pos_delta));
758 if (is_inf || (U32)data->offset_float_max > (U32)I32_MAX)
759 data->offset_float_max = I32_MAX;
760 if (data->flags & SF_BEFORE_EOL)
762 |= ((data->flags & SF_BEFORE_EOL) << SF_FL_SHIFT_EOL);
764 data->flags &= ~SF_FL_BEFORE_EOL;
765 data->minlen_float=minlenp;
766 data->lookbehind_float=0;
769 SvCUR_set(data->last_found, 0);
771 SV * const sv = data->last_found;
772 if (SvUTF8(sv) && SvMAGICAL(sv)) {
773 MAGIC * const mg = mg_find(sv, PERL_MAGIC_utf8);
779 data->flags &= ~SF_BEFORE_EOL;
780 DEBUG_STUDYDATA("commit: ",data,0);
783 /* These macros set, clear and test whether the synthetic start class ('ssc',
784 * given by the parameter) matches an empty string (EOS). This uses the
785 * 'next_off' field in the node, to save a bit in the flags field. The ssc
786 * stands alone, so there is never a next_off, so this field is otherwise
787 * unused. The EOS information is used only for compilation, but theoretically
788 * it could be passed on to the execution code. This could be used to store
789 * more than one bit of information, but only this one is currently used. */
790 #define SET_SSC_EOS(node) STMT_START { (node)->next_off = TRUE; } STMT_END
791 #define CLEAR_SSC_EOS(node) STMT_START { (node)->next_off = FALSE; } STMT_END
792 #define TEST_SSC_EOS(node) cBOOL((node)->next_off)
794 /* Can match anything (initialization) */
796 S_cl_anything(const RExC_state_t *pRExC_state, struct regnode_charclass_class *cl)
798 PERL_ARGS_ASSERT_CL_ANYTHING;
800 ANYOF_BITMAP_SETALL(cl);
801 cl->flags = ANYOF_UNICODE_ALL;
804 /* If any portion of the regex is to operate under locale rules,
805 * initialization includes it. The reason this isn't done for all regexes
806 * is that the optimizer was written under the assumption that locale was
807 * all-or-nothing. Given the complexity and lack of documentation in the
808 * optimizer, and that there are inadequate test cases for locale, so many
809 * parts of it may not work properly, it is safest to avoid locale unless
811 if (RExC_contains_locale) {
812 ANYOF_CLASS_SETALL(cl); /* /l uses class */
813 cl->flags |= ANYOF_LOCALE|ANYOF_CLASS|ANYOF_LOC_FOLD;
816 ANYOF_CLASS_ZERO(cl); /* Only /l uses class now */
820 /* Can match anything (initialization) */
822 S_cl_is_anything(const struct regnode_charclass_class *cl)
826 PERL_ARGS_ASSERT_CL_IS_ANYTHING;
828 for (value = 0; value < ANYOF_MAX; value += 2)
829 if (ANYOF_CLASS_TEST(cl, value) && ANYOF_CLASS_TEST(cl, value + 1))
831 if (!(cl->flags & ANYOF_UNICODE_ALL))
833 if (!ANYOF_BITMAP_TESTALLSET((const void*)cl))
838 /* Can match anything (initialization) */
840 S_cl_init(const RExC_state_t *pRExC_state, struct regnode_charclass_class *cl)
842 PERL_ARGS_ASSERT_CL_INIT;
844 Zero(cl, 1, struct regnode_charclass_class);
846 cl_anything(pRExC_state, cl);
847 ARG_SET(cl, ANYOF_NONBITMAP_EMPTY);
850 /* These two functions currently do the exact same thing */
851 #define cl_init_zero S_cl_init
853 /* 'AND' a given class with another one. Can create false positives. 'cl'
854 * should not be inverted. 'and_with->flags & ANYOF_CLASS' should be 0 if
855 * 'and_with' is a regnode_charclass instead of a regnode_charclass_class. */
857 S_cl_and(struct regnode_charclass_class *cl,
858 const struct regnode_charclass_class *and_with)
860 PERL_ARGS_ASSERT_CL_AND;
862 assert(PL_regkind[and_with->type] == ANYOF);
864 /* I (khw) am not sure all these restrictions are necessary XXX */
865 if (!(ANYOF_CLASS_TEST_ANY_SET(and_with))
866 && !(ANYOF_CLASS_TEST_ANY_SET(cl))
867 && (and_with->flags & ANYOF_LOCALE) == (cl->flags & ANYOF_LOCALE)
868 && !(and_with->flags & ANYOF_LOC_FOLD)
869 && !(cl->flags & ANYOF_LOC_FOLD)) {
872 if (and_with->flags & ANYOF_INVERT)
873 for (i = 0; i < ANYOF_BITMAP_SIZE; i++)
874 cl->bitmap[i] &= ~and_with->bitmap[i];
876 for (i = 0; i < ANYOF_BITMAP_SIZE; i++)
877 cl->bitmap[i] &= and_with->bitmap[i];
878 } /* XXXX: logic is complicated otherwise, leave it along for a moment. */
880 if (and_with->flags & ANYOF_INVERT) {
882 /* Here, the and'ed node is inverted. Get the AND of the flags that
883 * aren't affected by the inversion. Those that are affected are
884 * handled individually below */
885 U8 affected_flags = cl->flags & ~INVERSION_UNAFFECTED_FLAGS;
886 cl->flags &= (and_with->flags & INVERSION_UNAFFECTED_FLAGS);
887 cl->flags |= affected_flags;
889 /* We currently don't know how to deal with things that aren't in the
890 * bitmap, but we know that the intersection is no greater than what
891 * is already in cl, so let there be false positives that get sorted
892 * out after the synthetic start class succeeds, and the node is
893 * matched for real. */
895 /* The inversion of these two flags indicate that the resulting
896 * intersection doesn't have them */
897 if (and_with->flags & ANYOF_UNICODE_ALL) {
898 cl->flags &= ~ANYOF_UNICODE_ALL;
900 if (and_with->flags & ANYOF_NON_UTF8_LATIN1_ALL) {
901 cl->flags &= ~ANYOF_NON_UTF8_LATIN1_ALL;
904 else { /* and'd node is not inverted */
905 U8 outside_bitmap_but_not_utf8; /* Temp variable */
907 if (! ANYOF_NONBITMAP(and_with)) {
909 /* Here 'and_with' doesn't match anything outside the bitmap
910 * (except possibly ANYOF_UNICODE_ALL), which means the
911 * intersection can't either, except for ANYOF_UNICODE_ALL, in
912 * which case we don't know what the intersection is, but it's no
913 * greater than what cl already has, so can just leave it alone,
914 * with possible false positives */
915 if (! (and_with->flags & ANYOF_UNICODE_ALL)) {
916 ARG_SET(cl, ANYOF_NONBITMAP_EMPTY);
917 cl->flags &= ~ANYOF_NONBITMAP_NON_UTF8;
920 else if (! ANYOF_NONBITMAP(cl)) {
922 /* Here, 'and_with' does match something outside the bitmap, and cl
923 * doesn't have a list of things to match outside the bitmap. If
924 * cl can match all code points above 255, the intersection will
925 * be those above-255 code points that 'and_with' matches. If cl
926 * can't match all Unicode code points, it means that it can't
927 * match anything outside the bitmap (since the 'if' that got us
928 * into this block tested for that), so we leave the bitmap empty.
930 if (cl->flags & ANYOF_UNICODE_ALL) {
931 ARG_SET(cl, ARG(and_with));
933 /* and_with's ARG may match things that don't require UTF8.
934 * And now cl's will too, in spite of this being an 'and'. See
935 * the comments below about the kludge */
936 cl->flags |= and_with->flags & ANYOF_NONBITMAP_NON_UTF8;
940 /* Here, both 'and_with' and cl match something outside the
941 * bitmap. Currently we do not do the intersection, so just match
942 * whatever cl had at the beginning. */
946 /* Take the intersection of the two sets of flags. However, the
947 * ANYOF_NONBITMAP_NON_UTF8 flag is treated as an 'or'. This is a
948 * kludge around the fact that this flag is not treated like the others
949 * which are initialized in cl_anything(). The way the optimizer works
950 * is that the synthetic start class (SSC) is initialized to match
951 * anything, and then the first time a real node is encountered, its
952 * values are AND'd with the SSC's with the result being the values of
953 * the real node. However, there are paths through the optimizer where
954 * the AND never gets called, so those initialized bits are set
955 * inappropriately, which is not usually a big deal, as they just cause
956 * false positives in the SSC, which will just mean a probably
957 * imperceptible slow down in execution. However this bit has a
958 * higher false positive consequence in that it can cause utf8.pm,
959 * utf8_heavy.pl ... to be loaded when not necessary, which is a much
960 * bigger slowdown and also causes significant extra memory to be used.
961 * In order to prevent this, the code now takes a different tack. The
962 * bit isn't set unless some part of the regular expression needs it,
963 * but once set it won't get cleared. This means that these extra
964 * modules won't get loaded unless there was some path through the
965 * pattern that would have required them anyway, and so any false
966 * positives that occur by not ANDing them out when they could be
967 * aren't as severe as they would be if we treated this bit like all
969 outside_bitmap_but_not_utf8 = (cl->flags | and_with->flags)
970 & ANYOF_NONBITMAP_NON_UTF8;
971 cl->flags &= and_with->flags;
972 cl->flags |= outside_bitmap_but_not_utf8;
976 /* 'OR' a given class with another one. Can create false positives. 'cl'
977 * should not be inverted. 'or_with->flags & ANYOF_CLASS' should be 0 if
978 * 'or_with' is a regnode_charclass instead of a regnode_charclass_class. */
980 S_cl_or(const RExC_state_t *pRExC_state, struct regnode_charclass_class *cl, const struct regnode_charclass_class *or_with)
982 PERL_ARGS_ASSERT_CL_OR;
984 if (or_with->flags & ANYOF_INVERT) {
986 /* Here, the or'd node is to be inverted. This means we take the
987 * complement of everything not in the bitmap, but currently we don't
988 * know what that is, so give up and match anything */
989 if (ANYOF_NONBITMAP(or_with)) {
990 cl_anything(pRExC_state, cl);
993 * (B1 | CL1) | (!B2 & !CL2) = (B1 | !B2 & !CL2) | (CL1 | (!B2 & !CL2))
994 * <= (B1 | !B2) | (CL1 | !CL2)
995 * which is wasteful if CL2 is small, but we ignore CL2:
996 * (B1 | CL1) | (!B2 & !CL2) <= (B1 | CL1) | !B2 = (B1 | !B2) | CL1
997 * XXXX Can we handle case-fold? Unclear:
998 * (OK1(i) | OK1(i')) | !(OK1(i) | OK1(i')) =
999 * (OK1(i) | OK1(i')) | (!OK1(i) & !OK1(i'))
1001 else if ( (or_with->flags & ANYOF_LOCALE) == (cl->flags & ANYOF_LOCALE)
1002 && !(or_with->flags & ANYOF_LOC_FOLD)
1003 && !(cl->flags & ANYOF_LOC_FOLD) ) {
1006 for (i = 0; i < ANYOF_BITMAP_SIZE; i++)
1007 cl->bitmap[i] |= ~or_with->bitmap[i];
1008 } /* XXXX: logic is complicated otherwise */
1010 cl_anything(pRExC_state, cl);
1013 /* And, we can just take the union of the flags that aren't affected
1014 * by the inversion */
1015 cl->flags |= or_with->flags & INVERSION_UNAFFECTED_FLAGS;
1017 /* For the remaining flags:
1018 ANYOF_UNICODE_ALL and inverted means to not match anything above
1019 255, which means that the union with cl should just be
1020 what cl has in it, so can ignore this flag
1021 ANYOF_NON_UTF8_LATIN1_ALL and inverted means if not utf8 and ord
1022 is 127-255 to match them, but then invert that, so the
1023 union with cl should just be what cl has in it, so can
1026 } else { /* 'or_with' is not inverted */
1027 /* (B1 | CL1) | (B2 | CL2) = (B1 | B2) | (CL1 | CL2)) */
1028 if ( (or_with->flags & ANYOF_LOCALE) == (cl->flags & ANYOF_LOCALE)
1029 && (!(or_with->flags & ANYOF_LOC_FOLD)
1030 || (cl->flags & ANYOF_LOC_FOLD)) ) {
1033 /* OR char bitmap and class bitmap separately */
1034 for (i = 0; i < ANYOF_BITMAP_SIZE; i++)
1035 cl->bitmap[i] |= or_with->bitmap[i];
1036 if (or_with->flags & ANYOF_CLASS) {
1037 ANYOF_CLASS_OR(or_with, cl);
1040 else { /* XXXX: logic is complicated, leave it along for a moment. */
1041 cl_anything(pRExC_state, cl);
1044 if (ANYOF_NONBITMAP(or_with)) {
1046 /* Use the added node's outside-the-bit-map match if there isn't a
1047 * conflict. If there is a conflict (both nodes match something
1048 * outside the bitmap, but what they match outside is not the same
1049 * pointer, and hence not easily compared until XXX we extend
1050 * inversion lists this far), give up and allow the start class to
1051 * match everything outside the bitmap. If that stuff is all above
1052 * 255, can just set UNICODE_ALL, otherwise caould be anything. */
1053 if (! ANYOF_NONBITMAP(cl)) {
1054 ARG_SET(cl, ARG(or_with));
1056 else if (ARG(cl) != ARG(or_with)) {
1058 if ((or_with->flags & ANYOF_NONBITMAP_NON_UTF8)) {
1059 cl_anything(pRExC_state, cl);
1062 cl->flags |= ANYOF_UNICODE_ALL;
1067 /* Take the union */
1068 cl->flags |= or_with->flags;
1072 #define TRIE_LIST_ITEM(state,idx) (trie->states[state].trans.list)[ idx ]
1073 #define TRIE_LIST_CUR(state) ( TRIE_LIST_ITEM( state, 0 ).forid )
1074 #define TRIE_LIST_LEN(state) ( TRIE_LIST_ITEM( state, 0 ).newstate )
1075 #define TRIE_LIST_USED(idx) ( trie->states[state].trans.list ? (TRIE_LIST_CUR( idx ) - 1) : 0 )
1080 dump_trie(trie,widecharmap,revcharmap)
1081 dump_trie_interim_list(trie,widecharmap,revcharmap,next_alloc)
1082 dump_trie_interim_table(trie,widecharmap,revcharmap,next_alloc)
1084 These routines dump out a trie in a somewhat readable format.
1085 The _interim_ variants are used for debugging the interim
1086 tables that are used to generate the final compressed
1087 representation which is what dump_trie expects.
1089 Part of the reason for their existence is to provide a form
1090 of documentation as to how the different representations function.
1095 Dumps the final compressed table form of the trie to Perl_debug_log.
1096 Used for debugging make_trie().
1100 S_dump_trie(pTHX_ const struct _reg_trie_data *trie, HV *widecharmap,
1101 AV *revcharmap, U32 depth)
1104 SV *sv=sv_newmortal();
1105 int colwidth= widecharmap ? 6 : 4;
1107 GET_RE_DEBUG_FLAGS_DECL;
1109 PERL_ARGS_ASSERT_DUMP_TRIE;
1111 PerlIO_printf( Perl_debug_log, "%*sChar : %-6s%-6s%-4s ",
1112 (int)depth * 2 + 2,"",
1113 "Match","Base","Ofs" );
1115 for( state = 0 ; state < trie->uniquecharcount ; state++ ) {
1116 SV ** const tmp = av_fetch( revcharmap, state, 0);
1118 PerlIO_printf( Perl_debug_log, "%*s",
1120 pv_pretty(sv, SvPV_nolen_const(*tmp), SvCUR(*tmp), colwidth,
1121 PL_colors[0], PL_colors[1],
1122 (SvUTF8(*tmp) ? PERL_PV_ESCAPE_UNI : 0) |
1123 PERL_PV_ESCAPE_FIRSTCHAR
1128 PerlIO_printf( Perl_debug_log, "\n%*sState|-----------------------",
1129 (int)depth * 2 + 2,"");
1131 for( state = 0 ; state < trie->uniquecharcount ; state++ )
1132 PerlIO_printf( Perl_debug_log, "%.*s", colwidth, "--------");
1133 PerlIO_printf( Perl_debug_log, "\n");
1135 for( state = 1 ; state < trie->statecount ; state++ ) {
1136 const U32 base = trie->states[ state ].trans.base;
1138 PerlIO_printf( Perl_debug_log, "%*s#%4"UVXf"|", (int)depth * 2 + 2,"", (UV)state);
1140 if ( trie->states[ state ].wordnum ) {
1141 PerlIO_printf( Perl_debug_log, " W%4X", trie->states[ state ].wordnum );
1143 PerlIO_printf( Perl_debug_log, "%6s", "" );
1146 PerlIO_printf( Perl_debug_log, " @%4"UVXf" ", (UV)base );
1151 while( ( base + ofs < trie->uniquecharcount ) ||
1152 ( base + ofs - trie->uniquecharcount < trie->lasttrans
1153 && trie->trans[ base + ofs - trie->uniquecharcount ].check != state))
1156 PerlIO_printf( Perl_debug_log, "+%2"UVXf"[ ", (UV)ofs);
1158 for ( ofs = 0 ; ofs < trie->uniquecharcount ; ofs++ ) {
1159 if ( ( base + ofs >= trie->uniquecharcount ) &&
1160 ( base + ofs - trie->uniquecharcount < trie->lasttrans ) &&
1161 trie->trans[ base + ofs - trie->uniquecharcount ].check == state )
1163 PerlIO_printf( Perl_debug_log, "%*"UVXf,
1165 (UV)trie->trans[ base + ofs - trie->uniquecharcount ].next );
1167 PerlIO_printf( Perl_debug_log, "%*s",colwidth," ." );
1171 PerlIO_printf( Perl_debug_log, "]");
1174 PerlIO_printf( Perl_debug_log, "\n" );
1176 PerlIO_printf(Perl_debug_log, "%*sword_info N:(prev,len)=", (int)depth*2, "");
1177 for (word=1; word <= trie->wordcount; word++) {
1178 PerlIO_printf(Perl_debug_log, " %d:(%d,%d)",
1179 (int)word, (int)(trie->wordinfo[word].prev),
1180 (int)(trie->wordinfo[word].len));
1182 PerlIO_printf(Perl_debug_log, "\n" );
1185 Dumps a fully constructed but uncompressed trie in list form.
1186 List tries normally only are used for construction when the number of
1187 possible chars (trie->uniquecharcount) is very high.
1188 Used for debugging make_trie().
1191 S_dump_trie_interim_list(pTHX_ const struct _reg_trie_data *trie,
1192 HV *widecharmap, AV *revcharmap, U32 next_alloc,
1196 SV *sv=sv_newmortal();
1197 int colwidth= widecharmap ? 6 : 4;
1198 GET_RE_DEBUG_FLAGS_DECL;
1200 PERL_ARGS_ASSERT_DUMP_TRIE_INTERIM_LIST;
1202 /* print out the table precompression. */
1203 PerlIO_printf( Perl_debug_log, "%*sState :Word | Transition Data\n%*s%s",
1204 (int)depth * 2 + 2,"", (int)depth * 2 + 2,"",
1205 "------:-----+-----------------\n" );
1207 for( state=1 ; state < next_alloc ; state ++ ) {
1210 PerlIO_printf( Perl_debug_log, "%*s %4"UVXf" :",
1211 (int)depth * 2 + 2,"", (UV)state );
1212 if ( ! trie->states[ state ].wordnum ) {
1213 PerlIO_printf( Perl_debug_log, "%5s| ","");
1215 PerlIO_printf( Perl_debug_log, "W%4x| ",
1216 trie->states[ state ].wordnum
1219 for( charid = 1 ; charid <= TRIE_LIST_USED( state ) ; charid++ ) {
1220 SV ** const tmp = av_fetch( revcharmap, TRIE_LIST_ITEM(state,charid).forid, 0);
1222 PerlIO_printf( Perl_debug_log, "%*s:%3X=%4"UVXf" | ",
1224 pv_pretty(sv, SvPV_nolen_const(*tmp), SvCUR(*tmp), colwidth,
1225 PL_colors[0], PL_colors[1],
1226 (SvUTF8(*tmp) ? PERL_PV_ESCAPE_UNI : 0) |
1227 PERL_PV_ESCAPE_FIRSTCHAR
1229 TRIE_LIST_ITEM(state,charid).forid,
1230 (UV)TRIE_LIST_ITEM(state,charid).newstate
1233 PerlIO_printf(Perl_debug_log, "\n%*s| ",
1234 (int)((depth * 2) + 14), "");
1237 PerlIO_printf( Perl_debug_log, "\n");
1242 Dumps a fully constructed but uncompressed trie in table form.
1243 This is the normal DFA style state transition table, with a few
1244 twists to facilitate compression later.
1245 Used for debugging make_trie().
1248 S_dump_trie_interim_table(pTHX_ const struct _reg_trie_data *trie,
1249 HV *widecharmap, AV *revcharmap, U32 next_alloc,
1254 SV *sv=sv_newmortal();
1255 int colwidth= widecharmap ? 6 : 4;
1256 GET_RE_DEBUG_FLAGS_DECL;
1258 PERL_ARGS_ASSERT_DUMP_TRIE_INTERIM_TABLE;
1261 print out the table precompression so that we can do a visual check
1262 that they are identical.
1265 PerlIO_printf( Perl_debug_log, "%*sChar : ",(int)depth * 2 + 2,"" );
1267 for( charid = 0 ; charid < trie->uniquecharcount ; charid++ ) {
1268 SV ** const tmp = av_fetch( revcharmap, charid, 0);
1270 PerlIO_printf( Perl_debug_log, "%*s",
1272 pv_pretty(sv, SvPV_nolen_const(*tmp), SvCUR(*tmp), colwidth,
1273 PL_colors[0], PL_colors[1],
1274 (SvUTF8(*tmp) ? PERL_PV_ESCAPE_UNI : 0) |
1275 PERL_PV_ESCAPE_FIRSTCHAR
1281 PerlIO_printf( Perl_debug_log, "\n%*sState+-",(int)depth * 2 + 2,"" );
1283 for( charid=0 ; charid < trie->uniquecharcount ; charid++ ) {
1284 PerlIO_printf( Perl_debug_log, "%.*s", colwidth,"--------");
1287 PerlIO_printf( Perl_debug_log, "\n" );
1289 for( state=1 ; state < next_alloc ; state += trie->uniquecharcount ) {
1291 PerlIO_printf( Perl_debug_log, "%*s%4"UVXf" : ",
1292 (int)depth * 2 + 2,"",
1293 (UV)TRIE_NODENUM( state ) );
1295 for( charid = 0 ; charid < trie->uniquecharcount ; charid++ ) {
1296 UV v=(UV)SAFE_TRIE_NODENUM( trie->trans[ state + charid ].next );
1298 PerlIO_printf( Perl_debug_log, "%*"UVXf, colwidth, v );
1300 PerlIO_printf( Perl_debug_log, "%*s", colwidth, "." );
1302 if ( ! trie->states[ TRIE_NODENUM( state ) ].wordnum ) {
1303 PerlIO_printf( Perl_debug_log, " (%4"UVXf")\n", (UV)trie->trans[ state ].check );
1305 PerlIO_printf( Perl_debug_log, " (%4"UVXf") W%4X\n", (UV)trie->trans[ state ].check,
1306 trie->states[ TRIE_NODENUM( state ) ].wordnum );
1314 /* make_trie(startbranch,first,last,tail,word_count,flags,depth)
1315 startbranch: the first branch in the whole branch sequence
1316 first : start branch of sequence of branch-exact nodes.
1317 May be the same as startbranch
1318 last : Thing following the last branch.
1319 May be the same as tail.
1320 tail : item following the branch sequence
1321 count : words in the sequence
1322 flags : currently the OP() type we will be building one of /EXACT(|F|Fl)/
1323 depth : indent depth
1325 Inplace optimizes a sequence of 2 or more Branch-Exact nodes into a TRIE node.
1327 A trie is an N'ary tree where the branches are determined by digital
1328 decomposition of the key. IE, at the root node you look up the 1st character and
1329 follow that branch repeat until you find the end of the branches. Nodes can be
1330 marked as "accepting" meaning they represent a complete word. Eg:
1334 would convert into the following structure. Numbers represent states, letters
1335 following numbers represent valid transitions on the letter from that state, if
1336 the number is in square brackets it represents an accepting state, otherwise it
1337 will be in parenthesis.
1339 +-h->+-e->[3]-+-r->(8)-+-s->[9]
1343 (1) +-i->(6)-+-s->[7]
1345 +-s->(3)-+-h->(4)-+-e->[5]
1347 Accept Word Mapping: 3=>1 (he),5=>2 (she), 7=>3 (his), 9=>4 (hers)
1349 This shows that when matching against the string 'hers' we will begin at state 1
1350 read 'h' and move to state 2, read 'e' and move to state 3 which is accepting,
1351 then read 'r' and go to state 8 followed by 's' which takes us to state 9 which
1352 is also accepting. Thus we know that we can match both 'he' and 'hers' with a
1353 single traverse. We store a mapping from accepting to state to which word was
1354 matched, and then when we have multiple possibilities we try to complete the
1355 rest of the regex in the order in which they occured in the alternation.
1357 The only prior NFA like behaviour that would be changed by the TRIE support is
1358 the silent ignoring of duplicate alternations which are of the form:
1360 / (DUPE|DUPE) X? (?{ ... }) Y /x
1362 Thus EVAL blocks following a trie may be called a different number of times with
1363 and without the optimisation. With the optimisations dupes will be silently
1364 ignored. This inconsistent behaviour of EVAL type nodes is well established as
1365 the following demonstrates:
1367 'words'=~/(word|word|word)(?{ print $1 })[xyz]/
1369 which prints out 'word' three times, but
1371 'words'=~/(word|word|word)(?{ print $1 })S/
1373 which doesnt print it out at all. This is due to other optimisations kicking in.
1375 Example of what happens on a structural level:
1377 The regexp /(ac|ad|ab)+/ will produce the following debug output:
1379 1: CURLYM[1] {1,32767}(18)
1390 This would be optimizable with startbranch=5, first=5, last=16, tail=16
1391 and should turn into:
1393 1: CURLYM[1] {1,32767}(18)
1395 [Words:3 Chars Stored:6 Unique Chars:4 States:5 NCP:1]
1403 Cases where tail != last would be like /(?foo|bar)baz/:
1413 which would be optimizable with startbranch=1, first=1, last=7, tail=8
1414 and would end up looking like:
1417 [Words:2 Chars Stored:6 Unique Chars:5 States:7 NCP:1]
1424 d = uvuni_to_utf8_flags(d, uv, 0);
1426 is the recommended Unicode-aware way of saying
1431 #define TRIE_STORE_REVCHAR(val) \
1434 SV *zlopp = newSV(7); /* XXX: optimize me */ \
1435 unsigned char *flrbbbbb = (unsigned char *) SvPVX(zlopp); \
1436 unsigned const char *const kapow = uvuni_to_utf8(flrbbbbb, val); \
1437 SvCUR_set(zlopp, kapow - flrbbbbb); \
1440 av_push(revcharmap, zlopp); \
1442 char ooooff = (char)val; \
1443 av_push(revcharmap, newSVpvn(&ooooff, 1)); \
1447 #define TRIE_READ_CHAR STMT_START { \
1450 /* if it is UTF then it is either already folded, or does not need folding */ \
1451 uvc = utf8n_to_uvuni( (const U8*) uc, UTF8_MAXLEN, &len, uniflags); \
1453 else if (folder == PL_fold_latin1) { \
1454 /* if we use this folder we have to obey unicode rules on latin-1 data */ \
1455 if ( foldlen > 0 ) { \
1456 uvc = utf8n_to_uvuni( (const U8*) scan, UTF8_MAXLEN, &len, uniflags ); \
1462 uvc = _to_fold_latin1( (U8) *uc, foldbuf, &foldlen, 1); \
1463 skiplen = UNISKIP(uvc); \
1464 foldlen -= skiplen; \
1465 scan = foldbuf + skiplen; \
1468 /* raw data, will be folded later if needed */ \
1476 #define TRIE_LIST_PUSH(state,fid,ns) STMT_START { \
1477 if ( TRIE_LIST_CUR( state ) >=TRIE_LIST_LEN( state ) ) { \
1478 U32 ging = TRIE_LIST_LEN( state ) *= 2; \
1479 Renew( trie->states[ state ].trans.list, ging, reg_trie_trans_le ); \
1481 TRIE_LIST_ITEM( state, TRIE_LIST_CUR( state ) ).forid = fid; \
1482 TRIE_LIST_ITEM( state, TRIE_LIST_CUR( state ) ).newstate = ns; \
1483 TRIE_LIST_CUR( state )++; \
1486 #define TRIE_LIST_NEW(state) STMT_START { \
1487 Newxz( trie->states[ state ].trans.list, \
1488 4, reg_trie_trans_le ); \
1489 TRIE_LIST_CUR( state ) = 1; \
1490 TRIE_LIST_LEN( state ) = 4; \
1493 #define TRIE_HANDLE_WORD(state) STMT_START { \
1494 U16 dupe= trie->states[ state ].wordnum; \
1495 regnode * const noper_next = regnext( noper ); \
1498 /* store the word for dumping */ \
1500 if (OP(noper) != NOTHING) \
1501 tmp = newSVpvn_utf8(STRING(noper), STR_LEN(noper), UTF); \
1503 tmp = newSVpvn_utf8( "", 0, UTF ); \
1504 av_push( trie_words, tmp ); \
1508 trie->wordinfo[curword].prev = 0; \
1509 trie->wordinfo[curword].len = wordlen; \
1510 trie->wordinfo[curword].accept = state; \
1512 if ( noper_next < tail ) { \
1514 trie->jump = (U16 *) PerlMemShared_calloc( word_count + 1, sizeof(U16) ); \
1515 trie->jump[curword] = (U16)(noper_next - convert); \
1517 jumper = noper_next; \
1519 nextbranch= regnext(cur); \
1523 /* It's a dupe. Pre-insert into the wordinfo[].prev */\
1524 /* chain, so that when the bits of chain are later */\
1525 /* linked together, the dups appear in the chain */\
1526 trie->wordinfo[curword].prev = trie->wordinfo[dupe].prev; \
1527 trie->wordinfo[dupe].prev = curword; \
1529 /* we haven't inserted this word yet. */ \
1530 trie->states[ state ].wordnum = curword; \
1535 #define TRIE_TRANS_STATE(state,base,ucharcount,charid,special) \
1536 ( ( base + charid >= ucharcount \
1537 && base + charid < ubound \
1538 && state == trie->trans[ base - ucharcount + charid ].check \
1539 && trie->trans[ base - ucharcount + charid ].next ) \
1540 ? trie->trans[ base - ucharcount + charid ].next \
1541 : ( state==1 ? special : 0 ) \
1545 #define MADE_JUMP_TRIE 2
1546 #define MADE_EXACT_TRIE 4
1549 S_make_trie(pTHX_ RExC_state_t *pRExC_state, regnode *startbranch, regnode *first, regnode *last, regnode *tail, U32 word_count, U32 flags, U32 depth)
1552 /* first pass, loop through and scan words */
1553 reg_trie_data *trie;
1554 HV *widecharmap = NULL;
1555 AV *revcharmap = newAV();
1557 const U32 uniflags = UTF8_ALLOW_DEFAULT;
1562 regnode *jumper = NULL;
1563 regnode *nextbranch = NULL;
1564 regnode *convert = NULL;
1565 U32 *prev_states; /* temp array mapping each state to previous one */
1566 /* we just use folder as a flag in utf8 */
1567 const U8 * folder = NULL;
1570 const U32 data_slot = add_data( pRExC_state, 4, "tuuu" );
1571 AV *trie_words = NULL;
1572 /* along with revcharmap, this only used during construction but both are
1573 * useful during debugging so we store them in the struct when debugging.
1576 const U32 data_slot = add_data( pRExC_state, 2, "tu" );
1577 STRLEN trie_charcount=0;
1579 SV *re_trie_maxbuff;
1580 GET_RE_DEBUG_FLAGS_DECL;
1582 PERL_ARGS_ASSERT_MAKE_TRIE;
1584 PERL_UNUSED_ARG(depth);
1591 case EXACTFU_TRICKYFOLD:
1592 case EXACTFU: folder = PL_fold_latin1; break;
1593 case EXACTF: folder = PL_fold; break;
1594 case EXACTFL: folder = PL_fold_locale; break;
1595 default: Perl_croak( aTHX_ "panic! In trie construction, unknown node type %u %s", (unsigned) flags, PL_reg_name[flags] );
1598 trie = (reg_trie_data *) PerlMemShared_calloc( 1, sizeof(reg_trie_data) );
1600 trie->startstate = 1;
1601 trie->wordcount = word_count;
1602 RExC_rxi->data->data[ data_slot ] = (void*)trie;
1603 trie->charmap = (U16 *) PerlMemShared_calloc( 256, sizeof(U16) );
1605 trie->bitmap = (char *) PerlMemShared_calloc( ANYOF_BITMAP_SIZE, 1 );
1606 trie->wordinfo = (reg_trie_wordinfo *) PerlMemShared_calloc(
1607 trie->wordcount+1, sizeof(reg_trie_wordinfo));
1610 trie_words = newAV();
1613 re_trie_maxbuff = get_sv(RE_TRIE_MAXBUF_NAME, 1);
1614 if (!SvIOK(re_trie_maxbuff)) {
1615 sv_setiv(re_trie_maxbuff, RE_TRIE_MAXBUF_INIT);
1617 DEBUG_TRIE_COMPILE_r({
1618 PerlIO_printf( Perl_debug_log,
1619 "%*smake_trie start==%d, first==%d, last==%d, tail==%d depth=%d\n",
1620 (int)depth * 2 + 2, "",
1621 REG_NODE_NUM(startbranch),REG_NODE_NUM(first),
1622 REG_NODE_NUM(last), REG_NODE_NUM(tail),
1626 /* Find the node we are going to overwrite */
1627 if ( first == startbranch && OP( last ) != BRANCH ) {
1628 /* whole branch chain */
1631 /* branch sub-chain */
1632 convert = NEXTOPER( first );
1635 /* -- First loop and Setup --
1637 We first traverse the branches and scan each word to determine if it
1638 contains widechars, and how many unique chars there are, this is
1639 important as we have to build a table with at least as many columns as we
1642 We use an array of integers to represent the character codes 0..255
1643 (trie->charmap) and we use a an HV* to store Unicode characters. We use the
1644 native representation of the character value as the key and IV's for the
1647 *TODO* If we keep track of how many times each character is used we can
1648 remap the columns so that the table compression later on is more
1649 efficient in terms of memory by ensuring the most common value is in the
1650 middle and the least common are on the outside. IMO this would be better
1651 than a most to least common mapping as theres a decent chance the most
1652 common letter will share a node with the least common, meaning the node
1653 will not be compressible. With a middle is most common approach the worst
1654 case is when we have the least common nodes twice.
1658 for ( cur = first ; cur < last ; cur = regnext( cur ) ) {
1659 regnode *noper = NEXTOPER( cur );
1660 const U8 *uc = (U8*)STRING( noper );
1661 const U8 *e = uc + STR_LEN( noper );
1663 U8 foldbuf[ UTF8_MAXBYTES_CASE + 1 ];
1665 const U8 *scan = (U8*)NULL;
1666 U32 wordlen = 0; /* required init */
1668 bool set_bit = trie->bitmap ? 1 : 0; /*store the first char in the bitmap?*/
1670 if (OP(noper) == NOTHING) {
1671 regnode *noper_next= regnext(noper);
1672 if (noper_next != tail && OP(noper_next) == flags) {
1674 uc= (U8*)STRING(noper);
1675 e= uc + STR_LEN(noper);
1676 trie->minlen= STR_LEN(noper);
1683 if ( set_bit ) { /* bitmap only alloced when !(UTF&&Folding) */
1684 TRIE_BITMAP_SET(trie,*uc); /* store the raw first byte
1685 regardless of encoding */
1686 if (OP( noper ) == EXACTFU_SS) {
1687 /* false positives are ok, so just set this */
1688 TRIE_BITMAP_SET(trie,0xDF);
1691 for ( ; uc < e ; uc += len ) {
1692 TRIE_CHARCOUNT(trie)++;
1697 U8 folded= folder[ (U8) uvc ];
1698 if ( !trie->charmap[ folded ] ) {
1699 trie->charmap[ folded ]=( ++trie->uniquecharcount );
1700 TRIE_STORE_REVCHAR( folded );
1703 if ( !trie->charmap[ uvc ] ) {
1704 trie->charmap[ uvc ]=( ++trie->uniquecharcount );
1705 TRIE_STORE_REVCHAR( uvc );
1708 /* store the codepoint in the bitmap, and its folded
1710 TRIE_BITMAP_SET(trie, uvc);
1712 /* store the folded codepoint */
1713 if ( folder ) TRIE_BITMAP_SET(trie, folder[(U8) uvc ]);
1716 /* store first byte of utf8 representation of
1717 variant codepoints */
1718 if (! UNI_IS_INVARIANT(uvc)) {
1719 TRIE_BITMAP_SET(trie, UTF8_TWO_BYTE_HI(uvc));
1722 set_bit = 0; /* We've done our bit :-) */
1727 widecharmap = newHV();
1729 svpp = hv_fetch( widecharmap, (char*)&uvc, sizeof( UV ), 1 );
1732 Perl_croak( aTHX_ "error creating/fetching widecharmap entry for 0x%"UVXf, uvc );
1734 if ( !SvTRUE( *svpp ) ) {
1735 sv_setiv( *svpp, ++trie->uniquecharcount );
1736 TRIE_STORE_REVCHAR(uvc);
1740 if( cur == first ) {
1741 trie->minlen = chars;
1742 trie->maxlen = chars;
1743 } else if (chars < trie->minlen) {
1744 trie->minlen = chars;
1745 } else if (chars > trie->maxlen) {
1746 trie->maxlen = chars;
1748 if (OP( noper ) == EXACTFU_SS) {
1749 /* XXX: workaround - 'ss' could match "\x{DF}" so minlen could be 1 and not 2*/
1750 if (trie->minlen > 1)
1753 if (OP( noper ) == EXACTFU_TRICKYFOLD) {
1754 /* XXX: workround - things like "\x{1FBE}\x{0308}\x{0301}" can match "\x{0390}"
1755 * - We assume that any such sequence might match a 2 byte string */
1756 if (trie->minlen > 2 )
1760 } /* end first pass */
1761 DEBUG_TRIE_COMPILE_r(
1762 PerlIO_printf( Perl_debug_log, "%*sTRIE(%s): W:%d C:%d Uq:%d Min:%d Max:%d\n",
1763 (int)depth * 2 + 2,"",
1764 ( widecharmap ? "UTF8" : "NATIVE" ), (int)word_count,
1765 (int)TRIE_CHARCOUNT(trie), trie->uniquecharcount,
1766 (int)trie->minlen, (int)trie->maxlen )
1770 We now know what we are dealing with in terms of unique chars and
1771 string sizes so we can calculate how much memory a naive
1772 representation using a flat table will take. If it's over a reasonable
1773 limit (as specified by ${^RE_TRIE_MAXBUF}) we use a more memory
1774 conservative but potentially much slower representation using an array
1777 At the end we convert both representations into the same compressed
1778 form that will be used in regexec.c for matching with. The latter
1779 is a form that cannot be used to construct with but has memory
1780 properties similar to the list form and access properties similar
1781 to the table form making it both suitable for fast searches and
1782 small enough that its feasable to store for the duration of a program.
1784 See the comment in the code where the compressed table is produced
1785 inplace from the flat tabe representation for an explanation of how
1786 the compression works.
1791 Newx(prev_states, TRIE_CHARCOUNT(trie) + 2, U32);
1794 if ( (IV)( ( TRIE_CHARCOUNT(trie) + 1 ) * trie->uniquecharcount + 1) > SvIV(re_trie_maxbuff) ) {
1796 Second Pass -- Array Of Lists Representation
1798 Each state will be represented by a list of charid:state records
1799 (reg_trie_trans_le) the first such element holds the CUR and LEN
1800 points of the allocated array. (See defines above).
1802 We build the initial structure using the lists, and then convert
1803 it into the compressed table form which allows faster lookups
1804 (but cant be modified once converted).
1807 STRLEN transcount = 1;
1809 DEBUG_TRIE_COMPILE_MORE_r( PerlIO_printf( Perl_debug_log,
1810 "%*sCompiling trie using list compiler\n",
1811 (int)depth * 2 + 2, ""));
1813 trie->states = (reg_trie_state *)
1814 PerlMemShared_calloc( TRIE_CHARCOUNT(trie) + 2,
1815 sizeof(reg_trie_state) );
1819 for ( cur = first ; cur < last ; cur = regnext( cur ) ) {
1821 regnode *noper = NEXTOPER( cur );
1822 U8 *uc = (U8*)STRING( noper );
1823 const U8 *e = uc + STR_LEN( noper );
1824 U32 state = 1; /* required init */
1825 U16 charid = 0; /* sanity init */
1826 U8 *scan = (U8*)NULL; /* sanity init */
1827 STRLEN foldlen = 0; /* required init */
1828 U32 wordlen = 0; /* required init */
1829 U8 foldbuf[ UTF8_MAXBYTES_CASE + 1 ];
1832 if (OP(noper) == NOTHING) {
1833 regnode *noper_next= regnext(noper);
1834 if (noper_next != tail && OP(noper_next) == flags) {
1836 uc= (U8*)STRING(noper);
1837 e= uc + STR_LEN(noper);
1841 if (OP(noper) != NOTHING) {
1842 for ( ; uc < e ; uc += len ) {
1847 charid = trie->charmap[ uvc ];
1849 SV** const svpp = hv_fetch( widecharmap, (char*)&uvc, sizeof( UV ), 0);
1853 charid=(U16)SvIV( *svpp );
1856 /* charid is now 0 if we dont know the char read, or nonzero if we do */
1863 if ( !trie->states[ state ].trans.list ) {
1864 TRIE_LIST_NEW( state );
1866 for ( check = 1; check <= TRIE_LIST_USED( state ); check++ ) {
1867 if ( TRIE_LIST_ITEM( state, check ).forid == charid ) {
1868 newstate = TRIE_LIST_ITEM( state, check ).newstate;
1873 newstate = next_alloc++;
1874 prev_states[newstate] = state;
1875 TRIE_LIST_PUSH( state, charid, newstate );
1880 Perl_croak( aTHX_ "panic! In trie construction, no char mapping for %"IVdf, uvc );
1884 TRIE_HANDLE_WORD(state);
1886 } /* end second pass */
1888 /* next alloc is the NEXT state to be allocated */
1889 trie->statecount = next_alloc;
1890 trie->states = (reg_trie_state *)
1891 PerlMemShared_realloc( trie->states,
1893 * sizeof(reg_trie_state) );
1895 /* and now dump it out before we compress it */
1896 DEBUG_TRIE_COMPILE_MORE_r(dump_trie_interim_list(trie, widecharmap,
1897 revcharmap, next_alloc,
1901 trie->trans = (reg_trie_trans *)
1902 PerlMemShared_calloc( transcount, sizeof(reg_trie_trans) );
1909 for( state=1 ; state < next_alloc ; state ++ ) {
1913 DEBUG_TRIE_COMPILE_MORE_r(
1914 PerlIO_printf( Perl_debug_log, "tp: %d zp: %d ",tp,zp)
1918 if (trie->states[state].trans.list) {
1919 U16 minid=TRIE_LIST_ITEM( state, 1).forid;
1923 for( idx = 2 ; idx <= TRIE_LIST_USED( state ) ; idx++ ) {
1924 const U16 forid = TRIE_LIST_ITEM( state, idx).forid;
1925 if ( forid < minid ) {
1927 } else if ( forid > maxid ) {
1931 if ( transcount < tp + maxid - minid + 1) {
1933 trie->trans = (reg_trie_trans *)
1934 PerlMemShared_realloc( trie->trans,
1936 * sizeof(reg_trie_trans) );
1937 Zero( trie->trans + (transcount / 2), transcount / 2 , reg_trie_trans );
1939 base = trie->uniquecharcount + tp - minid;
1940 if ( maxid == minid ) {
1942 for ( ; zp < tp ; zp++ ) {
1943 if ( ! trie->trans[ zp ].next ) {
1944 base = trie->uniquecharcount + zp - minid;
1945 trie->trans[ zp ].next = TRIE_LIST_ITEM( state, 1).newstate;
1946 trie->trans[ zp ].check = state;
1952 trie->trans[ tp ].next = TRIE_LIST_ITEM( state, 1).newstate;
1953 trie->trans[ tp ].check = state;
1958 for ( idx=1; idx <= TRIE_LIST_USED( state ) ; idx++ ) {
1959 const U32 tid = base - trie->uniquecharcount + TRIE_LIST_ITEM( state, idx ).forid;
1960 trie->trans[ tid ].next = TRIE_LIST_ITEM( state, idx ).newstate;
1961 trie->trans[ tid ].check = state;
1963 tp += ( maxid - minid + 1 );
1965 Safefree(trie->states[ state ].trans.list);
1968 DEBUG_TRIE_COMPILE_MORE_r(
1969 PerlIO_printf( Perl_debug_log, " base: %d\n",base);
1972 trie->states[ state ].trans.base=base;
1974 trie->lasttrans = tp + 1;
1978 Second Pass -- Flat Table Representation.
1980 we dont use the 0 slot of either trans[] or states[] so we add 1 to each.
1981 We know that we will need Charcount+1 trans at most to store the data
1982 (one row per char at worst case) So we preallocate both structures
1983 assuming worst case.
1985 We then construct the trie using only the .next slots of the entry
1988 We use the .check field of the first entry of the node temporarily to
1989 make compression both faster and easier by keeping track of how many non
1990 zero fields are in the node.
1992 Since trans are numbered from 1 any 0 pointer in the table is a FAIL
1995 There are two terms at use here: state as a TRIE_NODEIDX() which is a
1996 number representing the first entry of the node, and state as a
1997 TRIE_NODENUM() which is the trans number. state 1 is TRIE_NODEIDX(1) and
1998 TRIE_NODENUM(1), state 2 is TRIE_NODEIDX(2) and TRIE_NODENUM(3) if there
1999 are 2 entrys per node. eg:
2007 The table is internally in the right hand, idx form. However as we also
2008 have to deal with the states array which is indexed by nodenum we have to
2009 use TRIE_NODENUM() to convert.
2012 DEBUG_TRIE_COMPILE_MORE_r( PerlIO_printf( Perl_debug_log,
2013 "%*sCompiling trie using table compiler\n",
2014 (int)depth * 2 + 2, ""));
2016 trie->trans = (reg_trie_trans *)
2017 PerlMemShared_calloc( ( TRIE_CHARCOUNT(trie) + 1 )
2018 * trie->uniquecharcount + 1,
2019 sizeof(reg_trie_trans) );
2020 trie->states = (reg_trie_state *)
2021 PerlMemShared_calloc( TRIE_CHARCOUNT(trie) + 2,
2022 sizeof(reg_trie_state) );
2023 next_alloc = trie->uniquecharcount + 1;
2026 for ( cur = first ; cur < last ; cur = regnext( cur ) ) {
2028 regnode *noper = NEXTOPER( cur );
2029 const U8 *uc = (U8*)STRING( noper );
2030 const U8 *e = uc + STR_LEN( noper );
2032 U32 state = 1; /* required init */
2034 U16 charid = 0; /* sanity init */
2035 U32 accept_state = 0; /* sanity init */
2036 U8 *scan = (U8*)NULL; /* sanity init */
2038 STRLEN foldlen = 0; /* required init */
2039 U32 wordlen = 0; /* required init */
2041 U8 foldbuf[ UTF8_MAXBYTES_CASE + 1 ];
2043 if (OP(noper) == NOTHING) {
2044 regnode *noper_next= regnext(noper);
2045 if (noper_next != tail && OP(noper_next) == flags) {
2047 uc= (U8*)STRING(noper);
2048 e= uc + STR_LEN(noper);
2052 if ( OP(noper) != NOTHING ) {
2053 for ( ; uc < e ; uc += len ) {
2058 charid = trie->charmap[ uvc ];
2060 SV* const * const svpp = hv_fetch( widecharmap, (char*)&uvc, sizeof( UV ), 0);
2061 charid = svpp ? (U16)SvIV(*svpp) : 0;
2065 if ( !trie->trans[ state + charid ].next ) {
2066 trie->trans[ state + charid ].next = next_alloc;
2067 trie->trans[ state ].check++;
2068 prev_states[TRIE_NODENUM(next_alloc)]
2069 = TRIE_NODENUM(state);
2070 next_alloc += trie->uniquecharcount;
2072 state = trie->trans[ state + charid ].next;
2074 Perl_croak( aTHX_ "panic! In trie construction, no char mapping for %"IVdf, uvc );
2076 /* charid is now 0 if we dont know the char read, or nonzero if we do */
2079 accept_state = TRIE_NODENUM( state );
2080 TRIE_HANDLE_WORD(accept_state);
2082 } /* end second pass */
2084 /* and now dump it out before we compress it */
2085 DEBUG_TRIE_COMPILE_MORE_r(dump_trie_interim_table(trie, widecharmap,
2087 next_alloc, depth+1));
2091 * Inplace compress the table.*
2093 For sparse data sets the table constructed by the trie algorithm will
2094 be mostly 0/FAIL transitions or to put it another way mostly empty.
2095 (Note that leaf nodes will not contain any transitions.)
2097 This algorithm compresses the tables by eliminating most such
2098 transitions, at the cost of a modest bit of extra work during lookup:
2100 - Each states[] entry contains a .base field which indicates the
2101 index in the state[] array wheres its transition data is stored.
2103 - If .base is 0 there are no valid transitions from that node.
2105 - If .base is nonzero then charid is added to it to find an entry in
2108 -If trans[states[state].base+charid].check!=state then the
2109 transition is taken to be a 0/Fail transition. Thus if there are fail
2110 transitions at the front of the node then the .base offset will point
2111 somewhere inside the previous nodes data (or maybe even into a node
2112 even earlier), but the .check field determines if the transition is
2116 The following process inplace converts the table to the compressed
2117 table: We first do not compress the root node 1,and mark all its
2118 .check pointers as 1 and set its .base pointer as 1 as well. This
2119 allows us to do a DFA construction from the compressed table later,
2120 and ensures that any .base pointers we calculate later are greater
2123 - We set 'pos' to indicate the first entry of the second node.
2125 - We then iterate over the columns of the node, finding the first and
2126 last used entry at l and m. We then copy l..m into pos..(pos+m-l),
2127 and set the .check pointers accordingly, and advance pos
2128 appropriately and repreat for the next node. Note that when we copy
2129 the next pointers we have to convert them from the original
2130 NODEIDX form to NODENUM form as the former is not valid post
2133 - If a node has no transitions used we mark its base as 0 and do not
2134 advance the pos pointer.
2136 - If a node only has one transition we use a second pointer into the
2137 structure to fill in allocated fail transitions from other states.
2138 This pointer is independent of the main pointer and scans forward
2139 looking for null transitions that are allocated to a state. When it
2140 finds one it writes the single transition into the "hole". If the
2141 pointer doesnt find one the single transition is appended as normal.
2143 - Once compressed we can Renew/realloc the structures to release the
2146 See "Table-Compression Methods" in sec 3.9 of the Red Dragon,
2147 specifically Fig 3.47 and the associated pseudocode.
2151 const U32 laststate = TRIE_NODENUM( next_alloc );
2154 trie->statecount = laststate;
2156 for ( state = 1 ; state < laststate ; state++ ) {
2158 const U32 stateidx = TRIE_NODEIDX( state );
2159 const U32 o_used = trie->trans[ stateidx ].check;
2160 U32 used = trie->trans[ stateidx ].check;
2161 trie->trans[ stateidx ].check = 0;
2163 for ( charid = 0 ; used && charid < trie->uniquecharcount ; charid++ ) {
2164 if ( flag || trie->trans[ stateidx + charid ].next ) {
2165 if ( trie->trans[ stateidx + charid ].next ) {
2167 for ( ; zp < pos ; zp++ ) {
2168 if ( ! trie->trans[ zp ].next ) {
2172 trie->states[ state ].trans.base = zp + trie->uniquecharcount - charid ;
2173 trie->trans[ zp ].next = SAFE_TRIE_NODENUM( trie->trans[ stateidx + charid ].next );
2174 trie->trans[ zp ].check = state;
2175 if ( ++zp > pos ) pos = zp;
2182 trie->states[ state ].trans.base = pos + trie->uniquecharcount - charid ;
2184 trie->trans[ pos ].next = SAFE_TRIE_NODENUM( trie->trans[ stateidx + charid ].next );
2185 trie->trans[ pos ].check = state;
2190 trie->lasttrans = pos + 1;
2191 trie->states = (reg_trie_state *)
2192 PerlMemShared_realloc( trie->states, laststate
2193 * sizeof(reg_trie_state) );
2194 DEBUG_TRIE_COMPILE_MORE_r(
2195 PerlIO_printf( Perl_debug_log,
2196 "%*sAlloc: %d Orig: %"IVdf" elements, Final:%"IVdf". Savings of %%%5.2f\n",
2197 (int)depth * 2 + 2,"",
2198 (int)( ( TRIE_CHARCOUNT(trie) + 1 ) * trie->uniquecharcount + 1 ),
2201 ( ( next_alloc - pos ) * 100 ) / (double)next_alloc );
2204 } /* end table compress */
2206 DEBUG_TRIE_COMPILE_MORE_r(
2207 PerlIO_printf(Perl_debug_log, "%*sStatecount:%"UVxf" Lasttrans:%"UVxf"\n",
2208 (int)depth * 2 + 2, "",
2209 (UV)trie->statecount,
2210 (UV)trie->lasttrans)
2212 /* resize the trans array to remove unused space */
2213 trie->trans = (reg_trie_trans *)
2214 PerlMemShared_realloc( trie->trans, trie->lasttrans
2215 * sizeof(reg_trie_trans) );
2217 { /* Modify the program and insert the new TRIE node */
2218 U8 nodetype =(U8)(flags & 0xFF);
2222 regnode *optimize = NULL;
2223 #ifdef RE_TRACK_PATTERN_OFFSETS
2226 U32 mjd_nodelen = 0;
2227 #endif /* RE_TRACK_PATTERN_OFFSETS */
2228 #endif /* DEBUGGING */
2230 This means we convert either the first branch or the first Exact,
2231 depending on whether the thing following (in 'last') is a branch
2232 or not and whther first is the startbranch (ie is it a sub part of
2233 the alternation or is it the whole thing.)
2234 Assuming its a sub part we convert the EXACT otherwise we convert
2235 the whole branch sequence, including the first.
2237 /* Find the node we are going to overwrite */
2238 if ( first != startbranch || OP( last ) == BRANCH ) {
2239 /* branch sub-chain */
2240 NEXT_OFF( first ) = (U16)(last - first);
2241 #ifdef RE_TRACK_PATTERN_OFFSETS
2243 mjd_offset= Node_Offset((convert));
2244 mjd_nodelen= Node_Length((convert));
2247 /* whole branch chain */
2249 #ifdef RE_TRACK_PATTERN_OFFSETS
2252 const regnode *nop = NEXTOPER( convert );
2253 mjd_offset= Node_Offset((nop));
2254 mjd_nodelen= Node_Length((nop));
2258 PerlIO_printf(Perl_debug_log, "%*sMJD offset:%"UVuf" MJD length:%"UVuf"\n",
2259 (int)depth * 2 + 2, "",
2260 (UV)mjd_offset, (UV)mjd_nodelen)
2263 /* But first we check to see if there is a common prefix we can
2264 split out as an EXACT and put in front of the TRIE node. */
2265 trie->startstate= 1;
2266 if ( trie->bitmap && !widecharmap && !trie->jump ) {
2268 for ( state = 1 ; state < trie->statecount-1 ; state++ ) {
2272 const U32 base = trie->states[ state ].trans.base;
2274 if ( trie->states[state].wordnum )
2277 for ( ofs = 0 ; ofs < trie->uniquecharcount ; ofs++ ) {
2278 if ( ( base + ofs >= trie->uniquecharcount ) &&
2279 ( base + ofs - trie->uniquecharcount < trie->lasttrans ) &&
2280 trie->trans[ base + ofs - trie->uniquecharcount ].check == state )
2282 if ( ++count > 1 ) {
2283 SV **tmp = av_fetch( revcharmap, ofs, 0);
2284 const U8 *ch = (U8*)SvPV_nolen_const( *tmp );
2285 if ( state == 1 ) break;
2287 Zero(trie->bitmap, ANYOF_BITMAP_SIZE, char);
2289 PerlIO_printf(Perl_debug_log,
2290 "%*sNew Start State=%"UVuf" Class: [",
2291 (int)depth * 2 + 2, "",
2294 SV ** const tmp = av_fetch( revcharmap, idx, 0);
2295 const U8 * const ch = (U8*)SvPV_nolen_const( *tmp );
2297 TRIE_BITMAP_SET(trie,*ch);
2299 TRIE_BITMAP_SET(trie, folder[ *ch ]);
2301 PerlIO_printf(Perl_debug_log, "%s", (char*)ch)
2305 TRIE_BITMAP_SET(trie,*ch);
2307 TRIE_BITMAP_SET(trie,folder[ *ch ]);
2308 DEBUG_OPTIMISE_r(PerlIO_printf( Perl_debug_log,"%s", ch));
2314 SV **tmp = av_fetch( revcharmap, idx, 0);
2316 char *ch = SvPV( *tmp, len );
2318 SV *sv=sv_newmortal();
2319 PerlIO_printf( Perl_debug_log,
2320 "%*sPrefix State: %"UVuf" Idx:%"UVuf" Char='%s'\n",
2321 (int)depth * 2 + 2, "",
2323 pv_pretty(sv, SvPV_nolen_const(*tmp), SvCUR(*tmp), 6,
2324 PL_colors[0], PL_colors[1],
2325 (SvUTF8(*tmp) ? PERL_PV_ESCAPE_UNI : 0) |
2326 PERL_PV_ESCAPE_FIRSTCHAR
2331 OP( convert ) = nodetype;
2332 str=STRING(convert);
2335 STR_LEN(convert) += len;
2341 DEBUG_OPTIMISE_r(PerlIO_printf( Perl_debug_log,"]\n"));
2346 trie->prefixlen = (state-1);
2348 regnode *n = convert+NODE_SZ_STR(convert);
2349 NEXT_OFF(convert) = NODE_SZ_STR(convert);
2350 trie->startstate = state;
2351 trie->minlen -= (state - 1);
2352 trie->maxlen -= (state - 1);
2354 /* At least the UNICOS C compiler choked on this
2355 * being argument to DEBUG_r(), so let's just have
2358 #ifdef PERL_EXT_RE_BUILD
2364 regnode *fix = convert;
2365 U32 word = trie->wordcount;
2367 Set_Node_Offset_Length(convert, mjd_offset, state - 1);
2368 while( ++fix < n ) {
2369 Set_Node_Offset_Length(fix, 0, 0);
2372 SV ** const tmp = av_fetch( trie_words, word, 0 );
2374 if ( STR_LEN(convert) <= SvCUR(*tmp) )
2375 sv_chop(*tmp, SvPV_nolen(*tmp) + STR_LEN(convert));
2377 sv_chop(*tmp, SvPV_nolen(*tmp) + SvCUR(*tmp));
2385 NEXT_OFF(convert) = (U16)(tail - convert);
2386 DEBUG_r(optimize= n);
2392 if ( trie->maxlen ) {
2393 NEXT_OFF( convert ) = (U16)(tail - convert);
2394 ARG_SET( convert, data_slot );
2395 /* Store the offset to the first unabsorbed branch in
2396 jump[0], which is otherwise unused by the jump logic.
2397 We use this when dumping a trie and during optimisation. */
2399 trie->jump[0] = (U16)(nextbranch - convert);
2401 /* If the start state is not accepting (meaning there is no empty string/NOTHING)
2402 * and there is a bitmap
2403 * and the first "jump target" node we found leaves enough room
2404 * then convert the TRIE node into a TRIEC node, with the bitmap
2405 * embedded inline in the opcode - this is hypothetically faster.
2407 if ( !trie->states[trie->startstate].wordnum
2409 && ( (char *)jumper - (char *)convert) >= (int)sizeof(struct regnode_charclass) )
2411 OP( convert ) = TRIEC;
2412 Copy(trie->bitmap, ((struct regnode_charclass *)convert)->bitmap, ANYOF_BITMAP_SIZE, char);
2413 PerlMemShared_free(trie->bitmap);
2416 OP( convert ) = TRIE;
2418 /* store the type in the flags */
2419 convert->flags = nodetype;
2423 + regarglen[ OP( convert ) ];
2425 /* XXX We really should free up the resource in trie now,
2426 as we won't use them - (which resources?) dmq */
2428 /* needed for dumping*/
2429 DEBUG_r(if (optimize) {
2430 regnode *opt = convert;
2432 while ( ++opt < optimize) {
2433 Set_Node_Offset_Length(opt,0,0);
2436 Try to clean up some of the debris left after the
2439 while( optimize < jumper ) {
2440 mjd_nodelen += Node_Length((optimize));
2441 OP( optimize ) = OPTIMIZED;
2442 Set_Node_Offset_Length(optimize,0,0);
2445 Set_Node_Offset_Length(convert,mjd_offset,mjd_nodelen);
2447 } /* end node insert */
2448 REH_CALL_COMP_NODE_HOOK(pRExC_state->rx, convert);
2450 /* Finish populating the prev field of the wordinfo array. Walk back
2451 * from each accept state until we find another accept state, and if
2452 * so, point the first word's .prev field at the second word. If the
2453 * second already has a .prev field set, stop now. This will be the
2454 * case either if we've already processed that word's accept state,
2455 * or that state had multiple words, and the overspill words were
2456 * already linked up earlier.
2463 for (word=1; word <= trie->wordcount; word++) {
2465 if (trie->wordinfo[word].prev)
2467 state = trie->wordinfo[word].accept;
2469 state = prev_states[state];
2472 prev = trie->states[state].wordnum;
2476 trie->wordinfo[word].prev = prev;
2478 Safefree(prev_states);
2482 /* and now dump out the compressed format */
2483 DEBUG_TRIE_COMPILE_r(dump_trie(trie, widecharmap, revcharmap, depth+1));
2485 RExC_rxi->data->data[ data_slot + 1 ] = (void*)widecharmap;
2487 RExC_rxi->data->data[ data_slot + TRIE_WORDS_OFFSET ] = (void*)trie_words;
2488 RExC_rxi->data->data[ data_slot + 3 ] = (void*)revcharmap;
2490 SvREFCNT_dec_NN(revcharmap);
2494 : trie->startstate>1
2500 S_make_trie_failtable(pTHX_ RExC_state_t *pRExC_state, regnode *source, regnode *stclass, U32 depth)
2502 /* The Trie is constructed and compressed now so we can build a fail array if it's needed
2504 This is basically the Aho-Corasick algorithm. Its from exercise 3.31 and 3.32 in the
2505 "Red Dragon" -- Compilers, principles, techniques, and tools. Aho, Sethi, Ullman 1985/88
2508 We find the fail state for each state in the trie, this state is the longest proper
2509 suffix of the current state's 'word' that is also a proper prefix of another word in our
2510 trie. State 1 represents the word '' and is thus the default fail state. This allows
2511 the DFA not to have to restart after its tried and failed a word at a given point, it
2512 simply continues as though it had been matching the other word in the first place.
2514 'abcdgu'=~/abcdefg|cdgu/
2515 When we get to 'd' we are still matching the first word, we would encounter 'g' which would
2516 fail, which would bring us to the state representing 'd' in the second word where we would
2517 try 'g' and succeed, proceeding to match 'cdgu'.
2519 /* add a fail transition */
2520 const U32 trie_offset = ARG(source);
2521 reg_trie_data *trie=(reg_trie_data *)RExC_rxi->data->data[trie_offset];
2523 const U32 ucharcount = trie->uniquecharcount;
2524 const U32 numstates = trie->statecount;
2525 const U32 ubound = trie->lasttrans + ucharcount;
2529 U32 base = trie->states[ 1 ].trans.base;
2532 const U32 data_slot = add_data( pRExC_state, 1, "T" );
2533 GET_RE_DEBUG_FLAGS_DECL;
2535 PERL_ARGS_ASSERT_MAKE_TRIE_FAILTABLE;
2537 PERL_UNUSED_ARG(depth);
2541 ARG_SET( stclass, data_slot );
2542 aho = (reg_ac_data *) PerlMemShared_calloc( 1, sizeof(reg_ac_data) );
2543 RExC_rxi->data->data[ data_slot ] = (void*)aho;
2544 aho->trie=trie_offset;
2545 aho->states=(reg_trie_state *)PerlMemShared_malloc( numstates * sizeof(reg_trie_state) );
2546 Copy( trie->states, aho->states, numstates, reg_trie_state );
2547 Newxz( q, numstates, U32);
2548 aho->fail = (U32 *) PerlMemShared_calloc( numstates, sizeof(U32) );
2551 /* initialize fail[0..1] to be 1 so that we always have
2552 a valid final fail state */
2553 fail[ 0 ] = fail[ 1 ] = 1;
2555 for ( charid = 0; charid < ucharcount ; charid++ ) {
2556 const U32 newstate = TRIE_TRANS_STATE( 1, base, ucharcount, charid, 0 );
2558 q[ q_write ] = newstate;
2559 /* set to point at the root */
2560 fail[ q[ q_write++ ] ]=1;
2563 while ( q_read < q_write) {
2564 const U32 cur = q[ q_read++ % numstates ];
2565 base = trie->states[ cur ].trans.base;
2567 for ( charid = 0 ; charid < ucharcount ; charid++ ) {
2568 const U32 ch_state = TRIE_TRANS_STATE( cur, base, ucharcount, charid, 1 );
2570 U32 fail_state = cur;
2573 fail_state = fail[ fail_state ];
2574 fail_base = aho->states[ fail_state ].trans.base;
2575 } while ( !TRIE_TRANS_STATE( fail_state, fail_base, ucharcount, charid, 1 ) );
2577 fail_state = TRIE_TRANS_STATE( fail_state, fail_base, ucharcount, charid, 1 );
2578 fail[ ch_state ] = fail_state;
2579 if ( !aho->states[ ch_state ].wordnum && aho->states[ fail_state ].wordnum )
2581 aho->states[ ch_state ].wordnum = aho->states[ fail_state ].wordnum;
2583 q[ q_write++ % numstates] = ch_state;
2587 /* restore fail[0..1] to 0 so that we "fall out" of the AC loop
2588 when we fail in state 1, this allows us to use the
2589 charclass scan to find a valid start char. This is based on the principle
2590 that theres a good chance the string being searched contains lots of stuff
2591 that cant be a start char.
2593 fail[ 0 ] = fail[ 1 ] = 0;
2594 DEBUG_TRIE_COMPILE_r({
2595 PerlIO_printf(Perl_debug_log,
2596 "%*sStclass Failtable (%"UVuf" states): 0",
2597 (int)(depth * 2), "", (UV)numstates
2599 for( q_read=1; q_read<numstates; q_read++ ) {
2600 PerlIO_printf(Perl_debug_log, ", %"UVuf, (UV)fail[q_read]);
2602 PerlIO_printf(Perl_debug_log, "\n");
2605 /*RExC_seen |= REG_SEEN_TRIEDFA;*/
2610 * There are strange code-generation bugs caused on sparc64 by gcc-2.95.2.
2611 * These need to be revisited when a newer toolchain becomes available.
2613 #if defined(__sparc64__) && defined(__GNUC__)
2614 # if __GNUC__ < 2 || (__GNUC__ == 2 && __GNUC_MINOR__ < 96)
2615 # undef SPARC64_GCC_WORKAROUND
2616 # define SPARC64_GCC_WORKAROUND 1
2620 #define DEBUG_PEEP(str,scan,depth) \
2621 DEBUG_OPTIMISE_r({if (scan){ \
2622 SV * const mysv=sv_newmortal(); \
2623 regnode *Next = regnext(scan); \
2624 regprop(RExC_rx, mysv, scan); \
2625 PerlIO_printf(Perl_debug_log, "%*s" str ">%3d: %s (%d)\n", \
2626 (int)depth*2, "", REG_NODE_NUM(scan), SvPV_nolen_const(mysv),\
2627 Next ? (REG_NODE_NUM(Next)) : 0 ); \
2631 /* The below joins as many adjacent EXACTish nodes as possible into a single
2632 * one. The regop may be changed if the node(s) contain certain sequences that
2633 * require special handling. The joining is only done if:
2634 * 1) there is room in the current conglomerated node to entirely contain the
2636 * 2) they are the exact same node type
2638 * The adjacent nodes actually may be separated by NOTHING-kind nodes, and
2639 * these get optimized out
2641 * If a node is to match under /i (folded), the number of characters it matches
2642 * can be different than its character length if it contains a multi-character
2643 * fold. *min_subtract is set to the total delta of the input nodes.
2645 * And *has_exactf_sharp_s is set to indicate whether or not the node is EXACTF
2646 * and contains LATIN SMALL LETTER SHARP S
2648 * This is as good a place as any to discuss the design of handling these
2649 * multi-character fold sequences. It's been wrong in Perl for a very long
2650 * time. There are three code points in Unicode whose multi-character folds
2651 * were long ago discovered to mess things up. The previous designs for
2652 * dealing with these involved assigning a special node for them. This
2653 * approach doesn't work, as evidenced by this example:
2654 * "\xDFs" =~ /s\xDF/ui # Used to fail before these patches
2655 * Both these fold to "sss", but if the pattern is parsed to create a node that
2656 * would match just the \xDF, it won't be able to handle the case where a
2657 * successful match would have to cross the node's boundary. The new approach
2658 * that hopefully generally solves the problem generates an EXACTFU_SS node
2661 * It turns out that there are problems with all multi-character folds, and not
2662 * just these three. Now the code is general, for all such cases, but the
2663 * three still have some special handling. The approach taken is:
2664 * 1) This routine examines each EXACTFish node that could contain multi-
2665 * character fold sequences. It returns in *min_subtract how much to
2666 * subtract from the the actual length of the string to get a real minimum
2667 * match length; it is 0 if there are no multi-char folds. This delta is
2668 * used by the caller to adjust the min length of the match, and the delta
2669 * between min and max, so that the optimizer doesn't reject these
2670 * possibilities based on size constraints.
2671 * 2) Certain of these sequences require special handling by the trie code,
2672 * so, if found, this code changes the joined node type to special ops:
2673 * EXACTFU_TRICKYFOLD and EXACTFU_SS.
2674 * 3) For the sequence involving the Sharp s (\xDF), the node type EXACTFU_SS
2675 * is used for an EXACTFU node that contains at least one "ss" sequence in
2676 * it. For non-UTF-8 patterns and strings, this is the only case where
2677 * there is a possible fold length change. That means that a regular
2678 * EXACTFU node without UTF-8 involvement doesn't have to concern itself
2679 * with length changes, and so can be processed faster. regexec.c takes
2680 * advantage of this. Generally, an EXACTFish node that is in UTF-8 is
2681 * pre-folded by regcomp.c. This saves effort in regex matching.
2682 * However, the pre-folding isn't done for non-UTF8 patterns because the
2683 * fold of the MICRO SIGN requires UTF-8, and we don't want to slow things
2684 * down by forcing the pattern into UTF8 unless necessary. Also what
2685 * EXACTF and EXACTFL nodes fold to isn't known until runtime. The fold
2686 * possibilities for the non-UTF8 patterns are quite simple, except for
2687 * the sharp s. All the ones that don't involve a UTF-8 target string are
2688 * members of a fold-pair, and arrays are set up for all of them so that
2689 * the other member of the pair can be found quickly. Code elsewhere in
2690 * this file makes sure that in EXACTFU nodes, the sharp s gets folded to
2691 * 'ss', even if the pattern isn't UTF-8. This avoids the issues
2692 * described in the next item.
2693 * 4) A problem remains for the sharp s in EXACTF nodes. Whether it matches
2694 * 'ss' or not is not knowable at compile time. It will match iff the
2695 * target string is in UTF-8, unlike the EXACTFU nodes, where it always
2696 * matches; and the EXACTFL and EXACTFA nodes where it never does. Thus
2697 * it can't be folded to "ss" at compile time, unlike EXACTFU does (as
2698 * described in item 3). An assumption that the optimizer part of
2699 * regexec.c (probably unwittingly) makes is that a character in the
2700 * pattern corresponds to at most a single character in the target string.
2701 * (And I do mean character, and not byte here, unlike other parts of the
2702 * documentation that have never been updated to account for multibyte
2703 * Unicode.) This assumption is wrong only in this case, as all other
2704 * cases are either 1-1 folds when no UTF-8 is involved; or is true by
2705 * virtue of having this file pre-fold UTF-8 patterns. I'm
2706 * reluctant to try to change this assumption, so instead the code punts.
2707 * This routine examines EXACTF nodes for the sharp s, and returns a
2708 * boolean indicating whether or not the node is an EXACTF node that
2709 * contains a sharp s. When it is true, the caller sets a flag that later
2710 * causes the optimizer in this file to not set values for the floating
2711 * and fixed string lengths, and thus avoids the optimizer code in
2712 * regexec.c that makes the invalid assumption. Thus, there is no
2713 * optimization based on string lengths for EXACTF nodes that contain the
2714 * sharp s. This only happens for /id rules (which means the pattern
2718 #define JOIN_EXACT(scan,min_subtract,has_exactf_sharp_s, flags) \
2719 if (PL_regkind[OP(scan)] == EXACT) \
2720 join_exact(pRExC_state,(scan),(min_subtract),has_exactf_sharp_s, (flags),NULL,depth+1)
2723 S_join_exact(pTHX_ RExC_state_t *pRExC_state, regnode *scan, UV *min_subtract, bool *has_exactf_sharp_s, U32 flags,regnode *val, U32 depth) {
2724 /* Merge several consecutive EXACTish nodes into one. */
2725 regnode *n = regnext(scan);
2727 regnode *next = scan + NODE_SZ_STR(scan);
2731 regnode *stop = scan;
2732 GET_RE_DEBUG_FLAGS_DECL;
2734 PERL_UNUSED_ARG(depth);
2737 PERL_ARGS_ASSERT_JOIN_EXACT;
2738 #ifndef EXPERIMENTAL_INPLACESCAN
2739 PERL_UNUSED_ARG(flags);
2740 PERL_UNUSED_ARG(val);
2742 DEBUG_PEEP("join",scan,depth);
2744 /* Look through the subsequent nodes in the chain. Skip NOTHING, merge
2745 * EXACT ones that are mergeable to the current one. */
2747 && (PL_regkind[OP(n)] == NOTHING
2748 || (stringok && OP(n) == OP(scan)))
2750 && NEXT_OFF(scan) + NEXT_OFF(n) < I16_MAX)
2753 if (OP(n) == TAIL || n > next)
2755 if (PL_regkind[OP(n)] == NOTHING) {
2756 DEBUG_PEEP("skip:",n,depth);
2757 NEXT_OFF(scan) += NEXT_OFF(n);
2758 next = n + NODE_STEP_REGNODE;
2765 else if (stringok) {
2766 const unsigned int oldl = STR_LEN(scan);
2767 regnode * const nnext = regnext(n);
2769 /* XXX I (khw) kind of doubt that this works on platforms where
2770 * U8_MAX is above 255 because of lots of other assumptions */
2771 /* Don't join if the sum can't fit into a single node */
2772 if (oldl + STR_LEN(n) > U8_MAX)
2775 DEBUG_PEEP("merg",n,depth);
2778 NEXT_OFF(scan) += NEXT_OFF(n);
2779 STR_LEN(scan) += STR_LEN(n);
2780 next = n + NODE_SZ_STR(n);
2781 /* Now we can overwrite *n : */
2782 Move(STRING(n), STRING(scan) + oldl, STR_LEN(n), char);
2790 #ifdef EXPERIMENTAL_INPLACESCAN
2791 if (flags && !NEXT_OFF(n)) {
2792 DEBUG_PEEP("atch", val, depth);
2793 if (reg_off_by_arg[OP(n)]) {
2794 ARG_SET(n, val - n);
2797 NEXT_OFF(n) = val - n;
2805 *has_exactf_sharp_s = FALSE;
2807 /* Here, all the adjacent mergeable EXACTish nodes have been merged. We
2808 * can now analyze for sequences of problematic code points. (Prior to
2809 * this final joining, sequences could have been split over boundaries, and
2810 * hence missed). The sequences only happen in folding, hence for any
2811 * non-EXACT EXACTish node */
2812 if (OP(scan) != EXACT) {
2813 const U8 * const s0 = (U8*) STRING(scan);
2815 const U8 * const s_end = s0 + STR_LEN(scan);
2817 /* One pass is made over the node's string looking for all the
2818 * possibilities. to avoid some tests in the loop, there are two main
2819 * cases, for UTF-8 patterns (which can't have EXACTF nodes) and
2823 /* Examine the string for a multi-character fold sequence. UTF-8
2824 * patterns have all characters pre-folded by the time this code is
2826 while (s < s_end - 1) /* Can stop 1 before the end, as minimum
2827 length sequence we are looking for is 2 */
2830 int len = is_MULTI_CHAR_FOLD_utf8_safe(s, s_end);
2831 if (! len) { /* Not a multi-char fold: get next char */
2836 /* Nodes with 'ss' require special handling, except for EXACTFL
2837 * and EXACTFA for which there is no multi-char fold to this */
2838 if (len == 2 && *s == 's' && *(s+1) == 's'
2839 && OP(scan) != EXACTFL && OP(scan) != EXACTFA)
2842 OP(scan) = EXACTFU_SS;
2845 else if (len == 6 /* len is the same in both ASCII and EBCDIC for these */
2846 && (memEQ(s, GREEK_SMALL_LETTER_IOTA_UTF8
2847 COMBINING_DIAERESIS_UTF8
2848 COMBINING_ACUTE_ACCENT_UTF8,
2850 || memEQ(s, GREEK_SMALL_LETTER_UPSILON_UTF8
2851 COMBINING_DIAERESIS_UTF8
2852 COMBINING_ACUTE_ACCENT_UTF8,
2857 /* These two folds require special handling by trie's, so
2858 * change the node type to indicate this. If EXACTFA and
2859 * EXACTFL were ever to be handled by trie's, this would
2860 * have to be changed. If this node has already been
2861 * changed to EXACTFU_SS in this loop, leave it as is. (I
2862 * (khw) think it doesn't matter in regexec.c for UTF
2863 * patterns, but no need to change it */
2864 if (OP(scan) == EXACTFU) {
2865 OP(scan) = EXACTFU_TRICKYFOLD;
2869 else { /* Here is a generic multi-char fold. */
2870 const U8* multi_end = s + len;
2872 /* Count how many characters in it. In the case of /l and
2873 * /aa, no folds which contain ASCII code points are
2874 * allowed, so check for those, and skip if found. (In
2875 * EXACTFL, no folds are allowed to any Latin1 code point,
2876 * not just ASCII. But there aren't any of these
2877 * currently, nor ever likely, so don't take the time to
2878 * test for them. The code that generates the
2879 * is_MULTI_foo() macros croaks should one actually get put
2880 * into Unicode .) */
2881 if (OP(scan) != EXACTFL && OP(scan) != EXACTFA) {
2882 count = utf8_length(s, multi_end);
2886 while (s < multi_end) {
2889 goto next_iteration;
2899 /* The delta is how long the sequence is minus 1 (1 is how long
2900 * the character that folds to the sequence is) */
2901 *min_subtract += count - 1;
2905 else if (OP(scan) != EXACTFL && OP(scan) != EXACTFA) {
2907 /* Here, the pattern is not UTF-8. Look for the multi-char folds
2908 * that are all ASCII. As in the above case, EXACTFL and EXACTFA
2909 * nodes can't have multi-char folds to this range (and there are
2910 * no existing ones in the upper latin1 range). In the EXACTF
2911 * case we look also for the sharp s, which can be in the final
2912 * position. Otherwise we can stop looking 1 byte earlier because
2913 * have to find at least two characters for a multi-fold */
2914 const U8* upper = (OP(scan) == EXACTF) ? s_end : s_end -1;
2916 /* The below is perhaps overboard, but this allows us to save a
2917 * test each time through the loop at the expense of a mask. This
2918 * is because on both EBCDIC and ASCII machines, 'S' and 's' differ
2919 * by a single bit. On ASCII they are 32 apart; on EBCDIC, they
2920 * are 64. This uses an exclusive 'or' to find that bit and then
2921 * inverts it to form a mask, with just a single 0, in the bit
2922 * position where 'S' and 's' differ. */
2923 const U8 S_or_s_mask = (U8) ~ ('S' ^ 's');
2924 const U8 s_masked = 's' & S_or_s_mask;
2927 int len = is_MULTI_CHAR_FOLD_latin1_safe(s, s_end);
2928 if (! len) { /* Not a multi-char fold. */
2929 if (*s == LATIN_SMALL_LETTER_SHARP_S && OP(scan) == EXACTF)
2931 *has_exactf_sharp_s = TRUE;
2938 && ((*s & S_or_s_mask) == s_masked)
2939 && ((*(s+1) & S_or_s_mask) == s_masked))
2942 /* EXACTF nodes need to know that the minimum length
2943 * changed so that a sharp s in the string can match this
2944 * ss in the pattern, but they remain EXACTF nodes, as they
2945 * won't match this unless the target string is is UTF-8,
2946 * which we don't know until runtime */
2947 if (OP(scan) != EXACTF) {
2948 OP(scan) = EXACTFU_SS;
2952 *min_subtract += len - 1;
2959 /* Allow dumping but overwriting the collection of skipped
2960 * ops and/or strings with fake optimized ops */
2961 n = scan + NODE_SZ_STR(scan);
2969 DEBUG_OPTIMISE_r(if (merged){DEBUG_PEEP("finl",scan,depth)});
2973 /* REx optimizer. Converts nodes into quicker variants "in place".
2974 Finds fixed substrings. */
2976 /* Stops at toplevel WHILEM as well as at "last". At end *scanp is set
2977 to the position after last scanned or to NULL. */
2979 #define INIT_AND_WITHP \
2980 assert(!and_withp); \
2981 Newx(and_withp,1,struct regnode_charclass_class); \
2982 SAVEFREEPV(and_withp)
2984 /* this is a chain of data about sub patterns we are processing that
2985 need to be handled separately/specially in study_chunk. Its so
2986 we can simulate recursion without losing state. */
2988 typedef struct scan_frame {
2989 regnode *last; /* last node to process in this frame */
2990 regnode *next; /* next node to process when last is reached */
2991 struct scan_frame *prev; /*previous frame*/
2992 I32 stop; /* what stopparen do we use */
2996 #define SCAN_COMMIT(s, data, m) scan_commit(s, data, m, is_inf)
2999 S_study_chunk(pTHX_ RExC_state_t *pRExC_state, regnode **scanp,
3000 I32 *minlenp, I32 *deltap,
3005 struct regnode_charclass_class *and_withp,
3006 U32 flags, U32 depth)
3007 /* scanp: Start here (read-write). */
3008 /* deltap: Write maxlen-minlen here. */
3009 /* last: Stop before this one. */
3010 /* data: string data about the pattern */
3011 /* stopparen: treat close N as END */
3012 /* recursed: which subroutines have we recursed into */
3013 /* and_withp: Valid if flags & SCF_DO_STCLASS_OR */
3016 I32 min = 0; /* There must be at least this number of characters to match */
3018 regnode *scan = *scanp, *next;
3020 int is_inf = (flags & SCF_DO_SUBSTR) && (data->flags & SF_IS_INF);
3021 int is_inf_internal = 0; /* The studied chunk is infinite */
3022 I32 is_par = OP(scan) == OPEN ? ARG(scan) : 0;
3023 scan_data_t data_fake;
3024 SV *re_trie_maxbuff = NULL;
3025 regnode *first_non_open = scan;
3026 I32 stopmin = I32_MAX;
3027 scan_frame *frame = NULL;
3028 GET_RE_DEBUG_FLAGS_DECL;
3030 PERL_ARGS_ASSERT_STUDY_CHUNK;
3033 StructCopy(&zero_scan_data, &data_fake, scan_data_t);
3037 while (first_non_open && OP(first_non_open) == OPEN)
3038 first_non_open=regnext(first_non_open);
3043 while ( scan && OP(scan) != END && scan < last ){
3044 UV min_subtract = 0; /* How mmany chars to subtract from the minimum
3045 node length to get a real minimum (because
3046 the folded version may be shorter) */
3047 bool has_exactf_sharp_s = FALSE;
3048 /* Peephole optimizer: */
3049 DEBUG_STUDYDATA("Peep:", data,depth);
3050 DEBUG_PEEP("Peep",scan,depth);
3052 /* Its not clear to khw or hv why this is done here, and not in the
3053 * clauses that deal with EXACT nodes. khw's guess is that it's
3054 * because of a previous design */
3055 JOIN_EXACT(scan,&min_subtract, &has_exactf_sharp_s, 0);
3057 /* Follow the next-chain of the current node and optimize
3058 away all the NOTHINGs from it. */
3059 if (OP(scan) != CURLYX) {
3060 const int max = (reg_off_by_arg[OP(scan)]
3062 /* I32 may be smaller than U16 on CRAYs! */
3063 : (I32_MAX < U16_MAX ? I32_MAX : U16_MAX));
3064 int off = (reg_off_by_arg[OP(scan)] ? ARG(scan) : NEXT_OFF(scan));
3068 /* Skip NOTHING and LONGJMP. */
3069 while ((n = regnext(n))
3070 && ((PL_regkind[OP(n)] == NOTHING && (noff = NEXT_OFF(n)))
3071 || ((OP(n) == LONGJMP) && (noff = ARG(n))))
3072 && off + noff < max)
3074 if (reg_off_by_arg[OP(scan)])
3077 NEXT_OFF(scan) = off;
3082 /* The principal pseudo-switch. Cannot be a switch, since we
3083 look into several different things. */
3084 if (OP(scan) == BRANCH || OP(scan) == BRANCHJ
3085 || OP(scan) == IFTHEN) {
3086 next = regnext(scan);
3088 /* demq: the op(next)==code check is to see if we have "branch-branch" AFAICT */
3090 if (OP(next) == code || code == IFTHEN) {
3091 /* NOTE - There is similar code to this block below for handling
3092 TRIE nodes on a re-study. If you change stuff here check there
3094 I32 max1 = 0, min1 = I32_MAX, num = 0;
3095 struct regnode_charclass_class accum;
3096 regnode * const startbranch=scan;
3098 if (flags & SCF_DO_SUBSTR)
3099 SCAN_COMMIT(pRExC_state, data, minlenp); /* Cannot merge strings after this. */
3100 if (flags & SCF_DO_STCLASS)
3101 cl_init_zero(pRExC_state, &accum);
3103 while (OP(scan) == code) {
3104 I32 deltanext, minnext, f = 0, fake;
3105 struct regnode_charclass_class this_class;
3108 data_fake.flags = 0;
3110 data_fake.whilem_c = data->whilem_c;
3111 data_fake.last_closep = data->last_closep;
3114 data_fake.last_closep = &fake;
3116 data_fake.pos_delta = delta;
3117 next = regnext(scan);
3118 scan = NEXTOPER(scan);
3120 scan = NEXTOPER(scan);
3121 if (flags & SCF_DO_STCLASS) {
3122 cl_init(pRExC_state, &this_class);
3123 data_fake.start_class = &this_class;
3124 f = SCF_DO_STCLASS_AND;
3126 if (flags & SCF_WHILEM_VISITED_POS)
3127 f |= SCF_WHILEM_VISITED_POS;
3129 /* we suppose the run is continuous, last=next...*/
3130 minnext = study_chunk(pRExC_state, &scan, minlenp, &deltanext,
3132 stopparen, recursed, NULL, f,depth+1);
3135 if (deltanext == I32_MAX) {
3136 is_inf = is_inf_internal = 1;
3138 } else if (max1 < minnext + deltanext)
3139 max1 = minnext + deltanext;
3141 if (data_fake.flags & (SF_HAS_PAR|SF_IN_PAR))
3143 if (data_fake.flags & SCF_SEEN_ACCEPT) {
3144 if ( stopmin > minnext)
3145 stopmin = min + min1;
3146 flags &= ~SCF_DO_SUBSTR;
3148 data->flags |= SCF_SEEN_ACCEPT;
3151 if (data_fake.flags & SF_HAS_EVAL)
3152 data->flags |= SF_HAS_EVAL;
3153 data->whilem_c = data_fake.whilem_c;
3155 if (flags & SCF_DO_STCLASS)
3156 cl_or(pRExC_state, &accum, &this_class);
3158 if (code == IFTHEN && num < 2) /* Empty ELSE branch */
3160 if (flags & SCF_DO_SUBSTR) {
3161 data->pos_min += min1;
3162 if (data->pos_delta >= I32_MAX - (max1 - min1))
3163 data->pos_delta = I32_MAX;
3165 data->pos_delta += max1 - min1;
3166 if (max1 != min1 || is_inf)
3167 data->longest = &(data->longest_float);
3170 if (delta == I32_MAX || I32_MAX - delta - (max1 - min1) < 0)
3173 delta += max1 - min1;
3174 if (flags & SCF_DO_STCLASS_OR) {
3175 cl_or(pRExC_state, data->start_class, &accum);
3177 cl_and(data->start_class, and_withp);
3178 flags &= ~SCF_DO_STCLASS;
3181 else if (flags & SCF_DO_STCLASS_AND) {
3183 cl_and(data->start_class, &accum);
3184 flags &= ~SCF_DO_STCLASS;
3187 /* Switch to OR mode: cache the old value of
3188 * data->start_class */
3190 StructCopy(data->start_class, and_withp,
3191 struct regnode_charclass_class);
3192 flags &= ~SCF_DO_STCLASS_AND;
3193 StructCopy(&accum, data->start_class,
3194 struct regnode_charclass_class);
3195 flags |= SCF_DO_STCLASS_OR;
3196 SET_SSC_EOS(data->start_class);
3200 if (PERL_ENABLE_TRIE_OPTIMISATION && OP( startbranch ) == BRANCH ) {
3203 Assuming this was/is a branch we are dealing with: 'scan' now
3204 points at the item that follows the branch sequence, whatever
3205 it is. We now start at the beginning of the sequence and look
3212 which would be constructed from a pattern like /A|LIST|OF|WORDS/
3214 If we can find such a subsequence we need to turn the first
3215 element into a trie and then add the subsequent branch exact
3216 strings to the trie.
3220 1. patterns where the whole set of branches can be converted.
3222 2. patterns where only a subset can be converted.
3224 In case 1 we can replace the whole set with a single regop
3225 for the trie. In case 2 we need to keep the start and end
3228 'BRANCH EXACT; BRANCH EXACT; BRANCH X'
3229 becomes BRANCH TRIE; BRANCH X;
3231 There is an additional case, that being where there is a
3232 common prefix, which gets split out into an EXACT like node
3233 preceding the TRIE node.
3235 If x(1..n)==tail then we can do a simple trie, if not we make
3236 a "jump" trie, such that when we match the appropriate word
3237 we "jump" to the appropriate tail node. Essentially we turn
3238 a nested if into a case structure of sorts.
3243 if (!re_trie_maxbuff) {
3244 re_trie_maxbuff = get_sv(RE_TRIE_MAXBUF_NAME, 1);
3245 if (!SvIOK(re_trie_maxbuff))
3246 sv_setiv(re_trie_maxbuff, RE_TRIE_MAXBUF_INIT);
3248 if ( SvIV(re_trie_maxbuff)>=0 ) {
3250 regnode *first = (regnode *)NULL;
3251 regnode *last = (regnode *)NULL;
3252 regnode *tail = scan;
3257 SV * const mysv = sv_newmortal(); /* for dumping */
3259 /* var tail is used because there may be a TAIL
3260 regop in the way. Ie, the exacts will point to the
3261 thing following the TAIL, but the last branch will
3262 point at the TAIL. So we advance tail. If we
3263 have nested (?:) we may have to move through several
3267 while ( OP( tail ) == TAIL ) {
3268 /* this is the TAIL generated by (?:) */
3269 tail = regnext( tail );
3273 DEBUG_TRIE_COMPILE_r({
3274 regprop(RExC_rx, mysv, tail );
3275 PerlIO_printf( Perl_debug_log, "%*s%s%s\n",
3276 (int)depth * 2 + 2, "",
3277 "Looking for TRIE'able sequences. Tail node is: ",
3278 SvPV_nolen_const( mysv )
3284 Step through the branches
3285 cur represents each branch,
3286 noper is the first thing to be matched as part of that branch
3287 noper_next is the regnext() of that node.
3289 We normally handle a case like this /FOO[xyz]|BAR[pqr]/
3290 via a "jump trie" but we also support building with NOJUMPTRIE,
3291 which restricts the trie logic to structures like /FOO|BAR/.
3293 If noper is a trieable nodetype then the branch is a possible optimization
3294 target. If we are building under NOJUMPTRIE then we require that noper_next
3295 is the same as scan (our current position in the regex program).
3297 Once we have two or more consecutive such branches we can create a
3298 trie of the EXACT's contents and stitch it in place into the program.
3300 If the sequence represents all of the branches in the alternation we
3301 replace the entire thing with a single TRIE node.
3303 Otherwise when it is a subsequence we need to stitch it in place and
3304 replace only the relevant branches. This means the first branch has
3305 to remain as it is used by the alternation logic, and its next pointer,
3306 and needs to be repointed at the item on the branch chain following
3307 the last branch we have optimized away.
3309 This could be either a BRANCH, in which case the subsequence is internal,
3310 or it could be the item following the branch sequence in which case the
3311 subsequence is at the end (which does not necessarily mean the first node
3312 is the start of the alternation).
3314 TRIE_TYPE(X) is a define which maps the optype to a trietype.
3317 ----------------+-----------
3321 EXACTFU_SS | EXACTFU
3322 EXACTFU_TRICKYFOLD | EXACTFU
3327 #define TRIE_TYPE(X) ( ( NOTHING == (X) ) ? NOTHING : \
3328 ( EXACT == (X) ) ? EXACT : \
3329 ( EXACTFU == (X) || EXACTFU_SS == (X) || EXACTFU_TRICKYFOLD == (X) ) ? EXACTFU : \
3332 /* dont use tail as the end marker for this traverse */
3333 for ( cur = startbranch ; cur != scan ; cur = regnext( cur ) ) {
3334 regnode * const noper = NEXTOPER( cur );
3335 U8 noper_type = OP( noper );
3336 U8 noper_trietype = TRIE_TYPE( noper_type );
3337 #if defined(DEBUGGING) || defined(NOJUMPTRIE)
3338 regnode * const noper_next = regnext( noper );
3339 U8 noper_next_type = (noper_next && noper_next != tail) ? OP(noper_next) : 0;
3340 U8 noper_next_trietype = (noper_next && noper_next != tail) ? TRIE_TYPE( noper_next_type ) :0;
3343 DEBUG_TRIE_COMPILE_r({
3344 regprop(RExC_rx, mysv, cur);
3345 PerlIO_printf( Perl_debug_log, "%*s- %s (%d)",
3346 (int)depth * 2 + 2,"", SvPV_nolen_const( mysv ), REG_NODE_NUM(cur) );
3348 regprop(RExC_rx, mysv, noper);
3349 PerlIO_printf( Perl_debug_log, " -> %s",
3350 SvPV_nolen_const(mysv));
3353 regprop(RExC_rx, mysv, noper_next );
3354 PerlIO_printf( Perl_debug_log,"\t=> %s\t",
3355 SvPV_nolen_const(mysv));
3357 PerlIO_printf( Perl_debug_log, "(First==%d,Last==%d,Cur==%d,tt==%s,nt==%s,nnt==%s)\n",
3358 REG_NODE_NUM(first), REG_NODE_NUM(last), REG_NODE_NUM(cur),
3359 PL_reg_name[trietype], PL_reg_name[noper_trietype], PL_reg_name[noper_next_trietype]
3363 /* Is noper a trieable nodetype that can be merged with the
3364 * current trie (if there is one)? */
3368 ( noper_trietype == NOTHING)
3369 || ( trietype == NOTHING )
3370 || ( trietype == noper_trietype )
3373 && noper_next == tail
3377 /* Handle mergable triable node
3378 * Either we are the first node in a new trieable sequence,
3379 * in which case we do some bookkeeping, otherwise we update
3380 * the end pointer. */
3383 if ( noper_trietype == NOTHING ) {
3384 #if !defined(DEBUGGING) && !defined(NOJUMPTRIE)
3385 regnode * const noper_next = regnext( noper );
3386 U8 noper_next_type = (noper_next && noper_next!=tail) ? OP(noper_next) : 0;
3387 U8 noper_next_trietype = noper_next_type ? TRIE_TYPE( noper_next_type ) :0;
3390 if ( noper_next_trietype ) {
3391 trietype = noper_next_trietype;
3392 } else if (noper_next_type) {
3393 /* a NOTHING regop is 1 regop wide. We need at least two
3394 * for a trie so we can't merge this in */
3398 trietype = noper_trietype;
3401 if ( trietype == NOTHING )
3402 trietype = noper_trietype;
3407 } /* end handle mergable triable node */
3409 /* handle unmergable node -
3410 * noper may either be a triable node which can not be tried
3411 * together with the current trie, or a non triable node */
3413 /* If last is set and trietype is not NOTHING then we have found
3414 * at least two triable branch sequences in a row of a similar
3415 * trietype so we can turn them into a trie. If/when we
3416 * allow NOTHING to start a trie sequence this condition will be
3417 * required, and it isn't expensive so we leave it in for now. */
3418 if ( trietype && trietype != NOTHING )
3419 make_trie( pRExC_state,
3420 startbranch, first, cur, tail, count,
3421 trietype, depth+1 );
3422 last = NULL; /* note: we clear/update first, trietype etc below, so we dont do it here */
3426 && noper_next == tail
3429 /* noper is triable, so we can start a new trie sequence */
3432 trietype = noper_trietype;
3434 /* if we already saw a first but the current node is not triable then we have
3435 * to reset the first information. */
3440 } /* end handle unmergable node */
3441 } /* loop over branches */
3442 DEBUG_TRIE_COMPILE_r({
3443 regprop(RExC_rx, mysv, cur);
3444 PerlIO_printf( Perl_debug_log,
3445 "%*s- %s (%d) <SCAN FINISHED>\n", (int)depth * 2 + 2,
3446 "", SvPV_nolen_const( mysv ),REG_NODE_NUM(cur));
3449 if ( last && trietype ) {
3450 if ( trietype != NOTHING ) {
3451 /* the last branch of the sequence was part of a trie,
3452 * so we have to construct it here outside of the loop
3454 made= make_trie( pRExC_state, startbranch, first, scan, tail, count, trietype, depth+1 );
3455 #ifdef TRIE_STUDY_OPT
3456 if ( ((made == MADE_EXACT_TRIE &&
3457 startbranch == first)
3458 || ( first_non_open == first )) &&
3460 flags |= SCF_TRIE_RESTUDY;
3461 if ( startbranch == first
3464 RExC_seen &=~REG_TOP_LEVEL_BRANCHES;
3469 /* at this point we know whatever we have is a NOTHING sequence/branch
3470 * AND if 'startbranch' is 'first' then we can turn the whole thing into a NOTHING
3472 if ( startbranch == first ) {
3474 /* the entire thing is a NOTHING sequence, something like this:
3475 * (?:|) So we can turn it into a plain NOTHING op. */
3476 DEBUG_TRIE_COMPILE_r({
3477 regprop(RExC_rx, mysv, cur);
3478 PerlIO_printf( Perl_debug_log,
3479 "%*s- %s (%d) <NOTHING BRANCH SEQUENCE>\n", (int)depth * 2 + 2,
3480 "", SvPV_nolen_const( mysv ),REG_NODE_NUM(cur));
3483 OP(startbranch)= NOTHING;
3484 NEXT_OFF(startbranch)= tail - startbranch;
3485 for ( opt= startbranch + 1; opt < tail ; opt++ )
3489 } /* end if ( last) */
3490 } /* TRIE_MAXBUF is non zero */
3495 else if ( code == BRANCHJ ) { /* single branch is optimized. */
3496 scan = NEXTOPER(NEXTOPER(scan));
3497 } else /* single branch is optimized. */
3498 scan = NEXTOPER(scan);
3500 } else if (OP(scan) == SUSPEND || OP(scan) == GOSUB || OP(scan) == GOSTART) {
3501 scan_frame *newframe = NULL;
3506 if (OP(scan) != SUSPEND) {
3507 /* set the pointer */
3508 if (OP(scan) == GOSUB) {
3510 RExC_recurse[ARG2L(scan)] = scan;
3511 start = RExC_open_parens[paren-1];
3512 end = RExC_close_parens[paren-1];
3515 start = RExC_rxi->program + 1;
3519 Newxz(recursed, (((RExC_npar)>>3) +1), U8);
3520 SAVEFREEPV(recursed);
3522 if (!PAREN_TEST(recursed,paren+1)) {
3523 PAREN_SET(recursed,paren+1);
3524 Newx(newframe,1,scan_frame);
3526 if (flags & SCF_DO_SUBSTR) {
3527 SCAN_COMMIT(pRExC_state,data,minlenp);
3528 data->longest = &(data->longest_float);
3530 is_inf = is_inf_internal = 1;
3531 if (flags & SCF_DO_STCLASS_OR) /* Allow everything */
3532 cl_anything(pRExC_state, data->start_class);
3533 flags &= ~SCF_DO_STCLASS;
3536 Newx(newframe,1,scan_frame);
3539 end = regnext(scan);
3544 SAVEFREEPV(newframe);
3545 newframe->next = regnext(scan);
3546 newframe->last = last;
3547 newframe->stop = stopparen;
3548 newframe->prev = frame;
3558 else if (OP(scan) == EXACT) {
3559 I32 l = STR_LEN(scan);
3562 const U8 * const s = (U8*)STRING(scan);
3563 uc = utf8_to_uvchr_buf(s, s + l, NULL);
3564 l = utf8_length(s, s + l);
3566 uc = *((U8*)STRING(scan));
3569 if (flags & SCF_DO_SUBSTR) { /* Update longest substr. */
3570 /* The code below prefers earlier match for fixed
3571 offset, later match for variable offset. */
3572 if (data->last_end == -1) { /* Update the start info. */
3573 data->last_start_min = data->pos_min;
3574 data->last_start_max = is_inf
3575 ? I32_MAX : data->pos_min + data->pos_delta;
3577 sv_catpvn(data->last_found, STRING(scan), STR_LEN(scan));
3579 SvUTF8_on(data->last_found);
3581 SV * const sv = data->last_found;
3582 MAGIC * const mg = SvUTF8(sv) && SvMAGICAL(sv) ?
3583 mg_find(sv, PERL_MAGIC_utf8) : NULL;
3584 if (mg && mg->mg_len >= 0)
3585 mg->mg_len += utf8_length((U8*)STRING(scan),
3586 (U8*)STRING(scan)+STR_LEN(scan));
3588 data->last_end = data->pos_min + l;
3589 data->pos_min += l; /* As in the first entry. */
3590 data->flags &= ~SF_BEFORE_EOL;
3592 if (flags & SCF_DO_STCLASS_AND) {
3593 /* Check whether it is compatible with what we know already! */
3597 /* If compatible, we or it in below. It is compatible if is
3598 * in the bitmp and either 1) its bit or its fold is set, or 2)
3599 * it's for a locale. Even if there isn't unicode semantics
3600 * here, at runtime there may be because of matching against a
3601 * utf8 string, so accept a possible false positive for
3602 * latin1-range folds */
3604 (!(data->start_class->flags & ANYOF_LOCALE)
3605 && !ANYOF_BITMAP_TEST(data->start_class, uc)
3606 && (!(data->start_class->flags & ANYOF_LOC_FOLD)
3607 || !ANYOF_BITMAP_TEST(data->start_class, PL_fold_latin1[uc])))
3612 ANYOF_CLASS_ZERO(data->start_class);
3613 ANYOF_BITMAP_ZERO(data->start_class);
3615 ANYOF_BITMAP_SET(data->start_class, uc);
3616 else if (uc >= 0x100) {
3619 /* Some Unicode code points fold to the Latin1 range; as
3620 * XXX temporary code, instead of figuring out if this is
3621 * one, just assume it is and set all the start class bits
3622 * that could be some such above 255 code point's fold
3623 * which will generate fals positives. As the code
3624 * elsewhere that does compute the fold settles down, it
3625 * can be extracted out and re-used here */
3626 for (i = 0; i < 256; i++){
3627 if (HAS_NONLATIN1_FOLD_CLOSURE(i)) {
3628 ANYOF_BITMAP_SET(data->start_class, i);
3632 CLEAR_SSC_EOS(data->start_class);
3634 data->start_class->flags &= ~ANYOF_UNICODE_ALL;
3636 else if (flags & SCF_DO_STCLASS_OR) {
3637 /* false positive possible if the class is case-folded */
3639 ANYOF_BITMAP_SET(data->start_class, uc);
3641 data->start_class->flags |= ANYOF_UNICODE_ALL;
3642 CLEAR_SSC_EOS(data->start_class);
3643 cl_and(data->start_class, and_withp);
3645 flags &= ~SCF_DO_STCLASS;
3647 else if (PL_regkind[OP(scan)] == EXACT) { /* But OP != EXACT! */
3648 I32 l = STR_LEN(scan);
3649 UV uc = *((U8*)STRING(scan));
3651 /* Search for fixed substrings supports EXACT only. */
3652 if (flags & SCF_DO_SUBSTR) {
3654 SCAN_COMMIT(pRExC_state, data, minlenp);
3657 const U8 * const s = (U8 *)STRING(scan);
3658 uc = utf8_to_uvchr_buf(s, s + l, NULL);
3659 l = utf8_length(s, s + l);
3661 if (has_exactf_sharp_s) {
3662 RExC_seen |= REG_SEEN_EXACTF_SHARP_S;
3664 min += l - min_subtract;
3666 delta += min_subtract;
3667 if (flags & SCF_DO_SUBSTR) {
3668 data->pos_min += l - min_subtract;
3669 if (data->pos_min < 0) {
3672 data->pos_delta += min_subtract;
3674 data->longest = &(data->longest_float);
3677 if (flags & SCF_DO_STCLASS_AND) {
3678 /* Check whether it is compatible with what we know already! */
3681 (!(data->start_class->flags & ANYOF_LOCALE)
3682 && !ANYOF_BITMAP_TEST(data->start_class, uc)
3683 && !ANYOF_BITMAP_TEST(data->start_class, PL_fold_latin1[uc])))
3687 ANYOF_CLASS_ZERO(data->start_class);
3688 ANYOF_BITMAP_ZERO(data->start_class);
3690 ANYOF_BITMAP_SET(data->start_class, uc);
3691 CLEAR_SSC_EOS(data->start_class);
3692 if (OP(scan) == EXACTFL) {
3693 /* XXX This set is probably no longer necessary, and
3694 * probably wrong as LOCALE now is on in the initial
3696 data->start_class->flags |= ANYOF_LOCALE|ANYOF_LOC_FOLD;
3700 /* Also set the other member of the fold pair. In case
3701 * that unicode semantics is called for at runtime, use
3702 * the full latin1 fold. (Can't do this for locale,
3703 * because not known until runtime) */
3704 ANYOF_BITMAP_SET(data->start_class, PL_fold_latin1[uc]);
3706 /* All other (EXACTFL handled above) folds except under
3707 * /iaa that include s, S, and sharp_s also may include
3709 if (OP(scan) != EXACTFA) {
3710 if (uc == 's' || uc == 'S') {
3711 ANYOF_BITMAP_SET(data->start_class,
3712 LATIN_SMALL_LETTER_SHARP_S);
3714 else if (uc == LATIN_SMALL_LETTER_SHARP_S) {
3715 ANYOF_BITMAP_SET(data->start_class, 's');
3716 ANYOF_BITMAP_SET(data->start_class, 'S');
3721 else if (uc >= 0x100) {
3723 for (i = 0; i < 256; i++){
3724 if (_HAS_NONLATIN1_FOLD_CLOSURE_ONLY_FOR_USE_BY_REGCOMP_DOT_C_AND_REGEXEC_DOT_C(i)) {
3725 ANYOF_BITMAP_SET(data->start_class, i);
3730 else if (flags & SCF_DO_STCLASS_OR) {
3731 if (data->start_class->flags & ANYOF_LOC_FOLD) {
3732 /* false positive possible if the class is case-folded.
3733 Assume that the locale settings are the same... */
3735 ANYOF_BITMAP_SET(data->start_class, uc);
3736 if (OP(scan) != EXACTFL) {
3738 /* And set the other member of the fold pair, but
3739 * can't do that in locale because not known until
3741 ANYOF_BITMAP_SET(data->start_class,
3742 PL_fold_latin1[uc]);
3744 /* All folds except under /iaa that include s, S,
3745 * and sharp_s also may include the others */
3746 if (OP(scan) != EXACTFA) {
3747 if (uc == 's' || uc == 'S') {
3748 ANYOF_BITMAP_SET(data->start_class,
3749 LATIN_SMALL_LETTER_SHARP_S);
3751 else if (uc == LATIN_SMALL_LETTER_SHARP_S) {
3752 ANYOF_BITMAP_SET(data->start_class, 's');
3753 ANYOF_BITMAP_SET(data->start_class, 'S');
3758 CLEAR_SSC_EOS(data->start_class);
3760 cl_and(data->start_class, and_withp);
3762 flags &= ~SCF_DO_STCLASS;
3764 else if (REGNODE_VARIES(OP(scan))) {
3765 I32 mincount, maxcount, minnext, deltanext, fl = 0;
3766 I32 f = flags, pos_before = 0;
3767 regnode * const oscan = scan;
3768 struct regnode_charclass_class this_class;
3769 struct regnode_charclass_class *oclass = NULL;
3770 I32 next_is_eval = 0;
3772 switch (PL_regkind[OP(scan)]) {
3773 case WHILEM: /* End of (?:...)* . */
3774 scan = NEXTOPER(scan);
3777 if (flags & (SCF_DO_SUBSTR | SCF_DO_STCLASS)) {
3778 next = NEXTOPER(scan);
3779 if (OP(next) == EXACT || (flags & SCF_DO_STCLASS)) {
3781 maxcount = REG_INFTY;
3782 next = regnext(scan);
3783 scan = NEXTOPER(scan);
3787 if (flags & SCF_DO_SUBSTR)
3792 if (flags & SCF_DO_STCLASS) {
3794 maxcount = REG_INFTY;
3795 next = regnext(scan);
3796 scan = NEXTOPER(scan);
3799 is_inf = is_inf_internal = 1;
3800 scan = regnext(scan);
3801 if (flags & SCF_DO_SUBSTR) {
3802 SCAN_COMMIT(pRExC_state, data, minlenp); /* Cannot extend fixed substrings */
3803 data->longest = &(data->longest_float);
3805 goto optimize_curly_tail;
3807 if (stopparen>0 && (OP(scan)==CURLYN || OP(scan)==CURLYM)
3808 && (scan->flags == stopparen))
3813 mincount = ARG1(scan);
3814 maxcount = ARG2(scan);
3816 next = regnext(scan);
3817 if (OP(scan) == CURLYX) {
3818 I32 lp = (data ? *(data->last_closep) : 0);
3819 scan->flags = ((lp <= (I32)U8_MAX) ? (U8)lp : U8_MAX);
3821 scan = NEXTOPER(scan) + EXTRA_STEP_2ARGS;
3822 next_is_eval = (OP(scan) == EVAL);
3824 if (flags & SCF_DO_SUBSTR) {
3825 if (mincount == 0) SCAN_COMMIT(pRExC_state,data,minlenp); /* Cannot extend fixed substrings */
3826 pos_before = data->pos_min;
3830 data->flags &= ~(SF_HAS_PAR|SF_IN_PAR|SF_HAS_EVAL);
3832 data->flags |= SF_IS_INF;
3834 if (flags & SCF_DO_STCLASS) {
3835 cl_init(pRExC_state, &this_class);
3836 oclass = data->start_class;
3837 data->start_class = &this_class;
3838 f |= SCF_DO_STCLASS_AND;
3839 f &= ~SCF_DO_STCLASS_OR;
3841 /* Exclude from super-linear cache processing any {n,m}
3842 regops for which the combination of input pos and regex
3843 pos is not enough information to determine if a match
3846 For example, in the regex /foo(bar\s*){4,8}baz/ with the
3847 regex pos at the \s*, the prospects for a match depend not
3848 only on the input position but also on how many (bar\s*)
3849 repeats into the {4,8} we are. */
3850 if ((mincount > 1) || (maxcount > 1 && maxcount != REG_INFTY))
3851 f &= ~SCF_WHILEM_VISITED_POS;
3853 /* This will finish on WHILEM, setting scan, or on NULL: */
3854 minnext = study_chunk(pRExC_state, &scan, minlenp, &deltanext,
3855 last, data, stopparen, recursed, NULL,
3857 ? (f & ~SCF_DO_SUBSTR) : f),depth+1);
3859 if (flags & SCF_DO_STCLASS)
3860 data->start_class = oclass;
3861 if (mincount == 0 || minnext == 0) {
3862 if (flags & SCF_DO_STCLASS_OR) {
3863 cl_or(pRExC_state, data->start_class, &this_class);
3865 else if (flags & SCF_DO_STCLASS_AND) {
3866 /* Switch to OR mode: cache the old value of
3867 * data->start_class */
3869 StructCopy(data->start_class, and_withp,
3870 struct regnode_charclass_class);
3871 flags &= ~SCF_DO_STCLASS_AND;
3872 StructCopy(&this_class, data->start_class,
3873 struct regnode_charclass_class);
3874 flags |= SCF_DO_STCLASS_OR;
3875 SET_SSC_EOS(data->start_class);
3877 } else { /* Non-zero len */
3878 if (flags & SCF_DO_STCLASS_OR) {
3879 cl_or(pRExC_state, data->start_class, &this_class);
3880 cl_and(data->start_class, and_withp);
3882 else if (flags & SCF_DO_STCLASS_AND)
3883 cl_and(data->start_class, &this_class);
3884 flags &= ~SCF_DO_STCLASS;
3886 if (!scan) /* It was not CURLYX, but CURLY. */
3888 if ( /* ? quantifier ok, except for (?{ ... }) */
3889 (next_is_eval || !(mincount == 0 && maxcount == 1))
3890 && (minnext == 0) && (deltanext == 0)
3891 && data && !(data->flags & (SF_HAS_PAR|SF_IN_PAR))
3892 && maxcount <= REG_INFTY/3) /* Complement check for big count */
3894 /* Fatal warnings may leak the regexp without this: */
3895 SAVEFREESV(RExC_rx_sv);
3896 ckWARNreg(RExC_parse,
3897 "Quantifier unexpected on zero-length expression");
3898 (void)ReREFCNT_inc(RExC_rx_sv);
3901 min += minnext * mincount;
3902 is_inf_internal |= deltanext == I32_MAX
3903 || (maxcount == REG_INFTY && minnext + deltanext > 0);
3904 is_inf |= is_inf_internal;
3908 delta += (minnext + deltanext) * maxcount - minnext * mincount;
3910 /* Try powerful optimization CURLYX => CURLYN. */
3911 if ( OP(oscan) == CURLYX && data
3912 && data->flags & SF_IN_PAR
3913 && !(data->flags & SF_HAS_EVAL)
3914 && !deltanext && minnext == 1 ) {
3915 /* Try to optimize to CURLYN. */
3916 regnode *nxt = NEXTOPER(oscan) + EXTRA_STEP_2ARGS;
3917 regnode * const nxt1 = nxt;
3924 if (!REGNODE_SIMPLE(OP(nxt))
3925 && !(PL_regkind[OP(nxt)] == EXACT
3926 && STR_LEN(nxt) == 1))
3932 if (OP(nxt) != CLOSE)
3934 if (RExC_open_parens) {
3935 RExC_open_parens[ARG(nxt1)-1]=oscan; /*open->CURLYM*/
3936 RExC_close_parens[ARG(nxt1)-1]=nxt+2; /*close->while*/
3938 /* Now we know that nxt2 is the only contents: */
3939 oscan->flags = (U8)ARG(nxt);
3941 OP(nxt1) = NOTHING; /* was OPEN. */
3944 OP(nxt1 + 1) = OPTIMIZED; /* was count. */
3945 NEXT_OFF(nxt1+ 1) = 0; /* just for consistency. */
3946 NEXT_OFF(nxt2) = 0; /* just for consistency with CURLY. */
3947 OP(nxt) = OPTIMIZED; /* was CLOSE. */
3948 OP(nxt + 1) = OPTIMIZED; /* was count. */
3949 NEXT_OFF(nxt+ 1) = 0; /* just for consistency. */
3954 /* Try optimization CURLYX => CURLYM. */
3955 if ( OP(oscan) == CURLYX && data
3956 && !(data->flags & SF_HAS_PAR)
3957 && !(data->flags & SF_HAS_EVAL)
3958 && !deltanext /* atom is fixed width */
3959 && minnext != 0 /* CURLYM can't handle zero width */
3960 && ! (RExC_seen & REG_SEEN_EXACTF_SHARP_S) /* Nor \xDF */
3962 /* XXXX How to optimize if data == 0? */
3963 /* Optimize to a simpler form. */
3964 regnode *nxt = NEXTOPER(oscan) + EXTRA_STEP_2ARGS; /* OPEN */
3968 while ( (nxt2 = regnext(nxt)) /* skip over embedded stuff*/
3969 && (OP(nxt2) != WHILEM))
3971 OP(nxt2) = SUCCEED; /* Whas WHILEM */
3972 /* Need to optimize away parenths. */
3973 if ((data->flags & SF_IN_PAR) && OP(nxt) == CLOSE) {
3974 /* Set the parenth number. */
3975 regnode *nxt1 = NEXTOPER(oscan) + EXTRA_STEP_2ARGS; /* OPEN*/
3977 oscan->flags = (U8)ARG(nxt);
3978 if (RExC_open_parens) {
3979 RExC_open_parens[ARG(nxt1)-1]=oscan; /*open->CURLYM*/
3980 RExC_close_parens[ARG(nxt1)-1]=nxt2+1; /*close->NOTHING*/
3982 OP(nxt1) = OPTIMIZED; /* was OPEN. */
3983 OP(nxt) = OPTIMIZED; /* was CLOSE. */
3986 OP(nxt1 + 1) = OPTIMIZED; /* was count. */
3987 OP(nxt + 1) = OPTIMIZED; /* was count. */
3988 NEXT_OFF(nxt1 + 1) = 0; /* just for consistency. */
3989 NEXT_OFF(nxt + 1) = 0; /* just for consistency. */
3992 while ( nxt1 && (OP(nxt1) != WHILEM)) {
3993 regnode *nnxt = regnext(nxt1);
3995 if (reg_off_by_arg[OP(nxt1)])
3996 ARG_SET(nxt1, nxt2 - nxt1);
3997 else if (nxt2 - nxt1 < U16_MAX)
3998 NEXT_OFF(nxt1) = nxt2 - nxt1;
4000 OP(nxt) = NOTHING; /* Cannot beautify */
4005 /* Optimize again: */
4006 study_chunk(pRExC_state, &nxt1, minlenp, &deltanext, nxt,
4007 NULL, stopparen, recursed, NULL, 0,depth+1);
4012 else if ((OP(oscan) == CURLYX)
4013 && (flags & SCF_WHILEM_VISITED_POS)
4014 /* See the comment on a similar expression above.
4015 However, this time it's not a subexpression
4016 we care about, but the expression itself. */
4017 && (maxcount == REG_INFTY)
4018 && data && ++data->whilem_c < 16) {
4019 /* This stays as CURLYX, we can put the count/of pair. */
4020 /* Find WHILEM (as in regexec.c) */
4021 regnode *nxt = oscan + NEXT_OFF(oscan);
4023 if (OP(PREVOPER(nxt)) == NOTHING) /* LONGJMP */
4025 PREVOPER(nxt)->flags = (U8)(data->whilem_c
4026 | (RExC_whilem_seen << 4)); /* On WHILEM */
4028 if (data && fl & (SF_HAS_PAR|SF_IN_PAR))
4030 if (flags & SCF_DO_SUBSTR) {
4031 SV *last_str = NULL;
4032 int counted = mincount != 0;
4034 if (data->last_end > 0 && mincount != 0) { /* Ends with a string. */
4035 #if defined(SPARC64_GCC_WORKAROUND)
4038 const char *s = NULL;
4041 if (pos_before >= data->last_start_min)
4044 b = data->last_start_min;
4047 s = SvPV_const(data->last_found, l);
4048 old = b - data->last_start_min;
4051 I32 b = pos_before >= data->last_start_min
4052 ? pos_before : data->last_start_min;
4054 const char * const s = SvPV_const(data->last_found, l);
4055 I32 old = b - data->last_start_min;
4059 old = utf8_hop((U8*)s, old) - (U8*)s;
4061 /* Get the added string: */
4062 last_str = newSVpvn_utf8(s + old, l, UTF);
4063 if (deltanext == 0 && pos_before == b) {
4064 /* What was added is a constant string */
4066 SvGROW(last_str, (mincount * l) + 1);
4067 repeatcpy(SvPVX(last_str) + l,
4068 SvPVX_const(last_str), l, mincount - 1);
4069 SvCUR_set(last_str, SvCUR(last_str) * mincount);
4070 /* Add additional parts. */
4071 SvCUR_set(data->last_found,
4072 SvCUR(data->last_found) - l);
4073 sv_catsv(data->last_found, last_str);
4075 SV * sv = data->last_found;
4077 SvUTF8(sv) && SvMAGICAL(sv) ?
4078 mg_find(sv, PERL_MAGIC_utf8) : NULL;
4079 if (mg && mg->mg_len >= 0)
4080 mg->mg_len += CHR_SVLEN(last_str) - l;
4082 data->last_end += l * (mincount - 1);
4085 /* start offset must point into the last copy */
4086 data->last_start_min += minnext * (mincount - 1);
4087 data->last_start_max += is_inf ? I32_MAX
4088 : (maxcount - 1) * (minnext + data->pos_delta);
4091 /* It is counted once already... */
4092 data->pos_min += minnext * (mincount - counted);
4094 PerlIO_printf(Perl_debug_log, "counted=%d deltanext=%d I32_MAX=%d minnext=%d maxcount=%d mincount=%d\n",
4095 counted, deltanext, I32_MAX, minnext, maxcount, mincount);
4096 if (deltanext != I32_MAX)
4097 PerlIO_printf(Perl_debug_log, "LHS=%d RHS=%d\n", -counted * deltanext + (minnext + deltanext) * maxcount - minnext * mincount, I32_MAX - data->pos_delta);
4099 if (deltanext == I32_MAX || -counted * deltanext + (minnext + deltanext) * maxcount - minnext * mincount >= I32_MAX - data->pos_delta)
4100 data->pos_delta = I32_MAX;
4102 data->pos_delta += - counted * deltanext +
4103 (minnext + deltanext) * maxcount - minnext * mincount;
4104 if (mincount != maxcount) {
4105 /* Cannot extend fixed substrings found inside
4107 SCAN_COMMIT(pRExC_state,data,minlenp);
4108 if (mincount && last_str) {
4109 SV * const sv = data->last_found;
4110 MAGIC * const mg = SvUTF8(sv) && SvMAGICAL(sv) ?
4111 mg_find(sv, PERL_MAGIC_utf8) : NULL;
4115 sv_setsv(sv, last_str);
4116 data->last_end = data->pos_min;
4117 data->last_start_min =
4118 data->pos_min - CHR_SVLEN(last_str);
4119 data->last_start_max = is_inf
4121 : data->pos_min + data->pos_delta
4122 - CHR_SVLEN(last_str);
4124 data->longest = &(data->longest_float);
4126 SvREFCNT_dec(last_str);
4128 if (data && (fl & SF_HAS_EVAL))
4129 data->flags |= SF_HAS_EVAL;
4130 optimize_curly_tail:
4131 if (OP(oscan) != CURLYX) {
4132 while (PL_regkind[OP(next = regnext(oscan))] == NOTHING
4134 NEXT_OFF(oscan) += NEXT_OFF(next);
4137 default: /* REF, and CLUMP only? */
4138 if (flags & SCF_DO_SUBSTR) {
4139 SCAN_COMMIT(pRExC_state,data,minlenp); /* Cannot expect anything... */
4140 data->longest = &(data->longest_float);
4142 is_inf = is_inf_internal = 1;
4143 if (flags & SCF_DO_STCLASS_OR)
4144 cl_anything(pRExC_state, data->start_class);
4145 flags &= ~SCF_DO_STCLASS;
4149 else if (OP(scan) == LNBREAK) {
4150 if (flags & SCF_DO_STCLASS) {
4152 CLEAR_SSC_EOS(data->start_class); /* No match on empty */
4153 if (flags & SCF_DO_STCLASS_AND) {
4154 for (value = 0; value < 256; value++)
4155 if (!is_VERTWS_cp(value))
4156 ANYOF_BITMAP_CLEAR(data->start_class, value);
4159 for (value = 0; value < 256; value++)
4160 if (is_VERTWS_cp(value))
4161 ANYOF_BITMAP_SET(data->start_class, value);
4163 if (flags & SCF_DO_STCLASS_OR)
4164 cl_and(data->start_class, and_withp);
4165 flags &= ~SCF_DO_STCLASS;
4168 delta++; /* Because of the 2 char string cr-lf */
4169 if (flags & SCF_DO_SUBSTR) {
4170 SCAN_COMMIT(pRExC_state,data,minlenp); /* Cannot expect anything... */
4172 data->pos_delta += 1;
4173 data->longest = &(data->longest_float);
4176 else if (REGNODE_SIMPLE(OP(scan))) {
4179 if (flags & SCF_DO_SUBSTR) {
4180 SCAN_COMMIT(pRExC_state,data,minlenp);
4184 if (flags & SCF_DO_STCLASS) {
4186 CLEAR_SSC_EOS(data->start_class); /* No match on empty */
4188 /* Some of the logic below assumes that switching
4189 locale on will only add false positives. */
4190 switch (PL_regkind[OP(scan)]) {
4196 Perl_croak(aTHX_ "panic: unexpected simple REx opcode %d", OP(scan));
4199 if (flags & SCF_DO_STCLASS_OR) /* Allow everything */
4200 cl_anything(pRExC_state, data->start_class);
4203 if (OP(scan) == SANY)
4205 if (flags & SCF_DO_STCLASS_OR) { /* Everything but \n */
4206 value = (ANYOF_BITMAP_TEST(data->start_class,'\n')
4207 || ANYOF_CLASS_TEST_ANY_SET(data->start_class));
4208 cl_anything(pRExC_state, data->start_class);
4210 if (flags & SCF_DO_STCLASS_AND || !value)
4211 ANYOF_BITMAP_CLEAR(data->start_class,'\n');
4214 if (flags & SCF_DO_STCLASS_AND)
4215 cl_and(data->start_class,
4216 (struct regnode_charclass_class*)scan);
4218 cl_or(pRExC_state, data->start_class,
4219 (struct regnode_charclass_class*)scan);
4227 classnum = FLAGS(scan);
4228 if (flags & SCF_DO_STCLASS_AND) {
4229 if (!(data->start_class->flags & ANYOF_LOCALE)) {
4230 ANYOF_CLASS_CLEAR(data->start_class, classnum_to_namedclass(classnum) + 1);
4231 for (value = 0; value < loop_max; value++) {
4232 if (! _generic_isCC(UNI_TO_NATIVE(value), classnum)) {
4233 ANYOF_BITMAP_CLEAR(data->start_class, UNI_TO_NATIVE(value));
4239 if (data->start_class->flags & ANYOF_LOCALE) {
4240 ANYOF_CLASS_SET(data->start_class, classnum_to_namedclass(classnum));
4244 /* Even if under locale, set the bits for non-locale
4245 * in case it isn't a true locale-node. This will
4246 * create false positives if it truly is locale */
4247 for (value = 0; value < loop_max; value++) {
4248 if (_generic_isCC(UNI_TO_NATIVE(value), classnum)) {
4249 ANYOF_BITMAP_SET(data->start_class, UNI_TO_NATIVE(value));
4261 classnum = FLAGS(scan);
4262 if (flags & SCF_DO_STCLASS_AND) {
4263 if (!(data->start_class->flags & ANYOF_LOCALE)) {
4264 ANYOF_CLASS_CLEAR(data->start_class, classnum_to_namedclass(classnum));
4265 for (value = 0; value < loop_max; value++) {
4266 if (_generic_isCC(UNI_TO_NATIVE(value), classnum)) {
4267 ANYOF_BITMAP_CLEAR(data->start_class, UNI_TO_NATIVE(value));
4273 if (data->start_class->flags & ANYOF_LOCALE) {
4274 ANYOF_CLASS_SET(data->start_class, classnum_to_namedclass(classnum) + 1);
4278 /* Even if under locale, set the bits for non-locale in
4279 * case it isn't a true locale-node. This will create
4280 * false positives if it truly is locale */
4281 for (value = 0; value < loop_max; value++) {
4282 if (! _generic_isCC(UNI_TO_NATIVE(value), classnum)) {
4283 ANYOF_BITMAP_SET(data->start_class, UNI_TO_NATIVE(value));
4286 if (PL_regkind[OP(scan)] == NPOSIXD) {
4287 data->start_class->flags |= ANYOF_NON_UTF8_LATIN1_ALL;
4293 if (flags & SCF_DO_STCLASS_OR)
4294 cl_and(data->start_class, and_withp);
4295 flags &= ~SCF_DO_STCLASS;
4298 else if (PL_regkind[OP(scan)] == EOL && flags & SCF_DO_SUBSTR) {
4299 data->flags |= (OP(scan) == MEOL
4302 SCAN_COMMIT(pRExC_state, data, minlenp);
4305 else if ( PL_regkind[OP(scan)] == BRANCHJ
4306 /* Lookbehind, or need to calculate parens/evals/stclass: */
4307 && (scan->flags || data || (flags & SCF_DO_STCLASS))
4308 && (OP(scan) == IFMATCH || OP(scan) == UNLESSM)) {
4309 if ( OP(scan) == UNLESSM &&
4311 OP(NEXTOPER(NEXTOPER(scan))) == NOTHING &&
4312 OP(regnext(NEXTOPER(NEXTOPER(scan)))) == SUCCEED
4315 regnode *upto= regnext(scan);
4317 SV * const mysv_val=sv_newmortal();
4318 DEBUG_STUDYDATA("OPFAIL",data,depth);
4320 /*DEBUG_PARSE_MSG("opfail");*/
4321 regprop(RExC_rx, mysv_val, upto);
4322 PerlIO_printf(Perl_debug_log, "~ replace with OPFAIL pointed at %s (%"IVdf") offset %"IVdf"\n",
4323 SvPV_nolen_const(mysv_val),
4324 (IV)REG_NODE_NUM(upto),
4329 NEXT_OFF(scan) = upto - scan;
4330 for (opt= scan + 1; opt < upto ; opt++)
4331 OP(opt) = OPTIMIZED;
4335 if ( !PERL_ENABLE_POSITIVE_ASSERTION_STUDY
4336 || OP(scan) == UNLESSM )
4338 /* Negative Lookahead/lookbehind
4339 In this case we can't do fixed string optimisation.
4342 I32 deltanext, minnext, fake = 0;
4344 struct regnode_charclass_class intrnl;
4347 data_fake.flags = 0;
4349 data_fake.whilem_c = data->whilem_c;
4350 data_fake.last_closep = data->last_closep;
4353 data_fake.last_closep = &fake;
4354 data_fake.pos_delta = delta;
4355 if ( flags & SCF_DO_STCLASS && !scan->flags
4356 && OP(scan) == IFMATCH ) { /* Lookahead */
4357 cl_init(pRExC_state, &intrnl);
4358 data_fake.start_class = &intrnl;
4359 f |= SCF_DO_STCLASS_AND;
4361 if (flags & SCF_WHILEM_VISITED_POS)
4362 f |= SCF_WHILEM_VISITED_POS;
4363 next = regnext(scan);
4364 nscan = NEXTOPER(NEXTOPER(scan));
4365 minnext = study_chunk(pRExC_state, &nscan, minlenp, &deltanext,
4366 last, &data_fake, stopparen, recursed, NULL, f, depth+1);
4369 FAIL("Variable length lookbehind not implemented");
4371 else if (minnext > (I32)U8_MAX) {
4372 FAIL2("Lookbehind longer than %"UVuf" not implemented", (UV)U8_MAX);
4374 scan->flags = (U8)minnext;
4377 if (data_fake.flags & (SF_HAS_PAR|SF_IN_PAR))
4379 if (data_fake.flags & SF_HAS_EVAL)
4380 data->flags |= SF_HAS_EVAL;
4381 data->whilem_c = data_fake.whilem_c;
4383 if (f & SCF_DO_STCLASS_AND) {
4384 if (flags & SCF_DO_STCLASS_OR) {
4385 /* OR before, AND after: ideally we would recurse with
4386 * data_fake to get the AND applied by study of the
4387 * remainder of the pattern, and then derecurse;
4388 * *** HACK *** for now just treat as "no information".
4389 * See [perl #56690].
4391 cl_init(pRExC_state, data->start_class);
4393 /* AND before and after: combine and continue */
4394 const int was = TEST_SSC_EOS(data->start_class);
4396 cl_and(data->start_class, &intrnl);
4398 SET_SSC_EOS(data->start_class);
4402 #if PERL_ENABLE_POSITIVE_ASSERTION_STUDY
4404 /* Positive Lookahead/lookbehind
4405 In this case we can do fixed string optimisation,
4406 but we must be careful about it. Note in the case of
4407 lookbehind the positions will be offset by the minimum
4408 length of the pattern, something we won't know about
4409 until after the recurse.
4411 I32 deltanext, fake = 0;
4413 struct regnode_charclass_class intrnl;
4415 /* We use SAVEFREEPV so that when the full compile
4416 is finished perl will clean up the allocated
4417 minlens when it's all done. This way we don't
4418 have to worry about freeing them when we know
4419 they wont be used, which would be a pain.
4422 Newx( minnextp, 1, I32 );
4423 SAVEFREEPV(minnextp);
4426 StructCopy(data, &data_fake, scan_data_t);
4427 if ((flags & SCF_DO_SUBSTR) && data->last_found) {
4430 SCAN_COMMIT(pRExC_state, &data_fake,minlenp);
4431 data_fake.last_found=newSVsv(data->last_found);
4435 data_fake.last_closep = &fake;
4436 data_fake.flags = 0;
4437 data_fake.pos_delta = delta;
4439 data_fake.flags |= SF_IS_INF;
4440 if ( flags & SCF_DO_STCLASS && !scan->flags
4441 && OP(scan) == IFMATCH ) { /* Lookahead */
4442 cl_init(pRExC_state, &intrnl);
4443 data_fake.start_class = &intrnl;
4444 f |= SCF_DO_STCLASS_AND;
4446 if (flags & SCF_WHILEM_VISITED_POS)
4447 f |= SCF_WHILEM_VISITED_POS;
4448 next = regnext(scan);
4449 nscan = NEXTOPER(NEXTOPER(scan));
4451 *minnextp = study_chunk(pRExC_state, &nscan, minnextp, &deltanext,
4452 last, &data_fake, stopparen, recursed, NULL, f,depth+1);
4455 FAIL("Variable length lookbehind not implemented");
4457 else if (*minnextp > (I32)U8_MAX) {
4458 FAIL2("Lookbehind longer than %"UVuf" not implemented", (UV)U8_MAX);
4460 scan->flags = (U8)*minnextp;
4465 if (f & SCF_DO_STCLASS_AND) {
4466 const int was = TEST_SSC_EOS(data.start_class);
4468 cl_and(data->start_class, &intrnl);
4470 SET_SSC_EOS(data->start_class);
4473 if (data_fake.flags & (SF_HAS_PAR|SF_IN_PAR))
4475 if (data_fake.flags & SF_HAS_EVAL)
4476 data->flags |= SF_HAS_EVAL;
4477 data->whilem_c = data_fake.whilem_c;
4478 if ((flags & SCF_DO_SUBSTR) && data_fake.last_found) {
4479 if (RExC_rx->minlen<*minnextp)
4480 RExC_rx->minlen=*minnextp;
4481 SCAN_COMMIT(pRExC_state, &data_fake, minnextp);
4482 SvREFCNT_dec_NN(data_fake.last_found);
4484 if ( data_fake.minlen_fixed != minlenp )
4486 data->offset_fixed= data_fake.offset_fixed;
4487 data->minlen_fixed= data_fake.minlen_fixed;
4488 data->lookbehind_fixed+= scan->flags;
4490 if ( data_fake.minlen_float != minlenp )
4492 data->minlen_float= data_fake.minlen_float;
4493 data->offset_float_min=data_fake.offset_float_min;
4494 data->offset_float_max=data_fake.offset_float_max;
4495 data->lookbehind_float+= scan->flags;
4502 else if (OP(scan) == OPEN) {
4503 if (stopparen != (I32)ARG(scan))
4506 else if (OP(scan) == CLOSE) {
4507 if (stopparen == (I32)ARG(scan)) {
4510 if ((I32)ARG(scan) == is_par) {
4511 next = regnext(scan);
4513 if ( next && (OP(next) != WHILEM) && next < last)
4514 is_par = 0; /* Disable optimization */
4517 *(data->last_closep) = ARG(scan);
4519 else if (OP(scan) == EVAL) {
4521 data->flags |= SF_HAS_EVAL;
4523 else if ( PL_regkind[OP(scan)] == ENDLIKE ) {
4524 if (flags & SCF_DO_SUBSTR) {
4525 SCAN_COMMIT(pRExC_state,data,minlenp);
4526 flags &= ~SCF_DO_SUBSTR;
4528 if (data && OP(scan)==ACCEPT) {
4529 data->flags |= SCF_SEEN_ACCEPT;
4534 else if (OP(scan) == LOGICAL && scan->flags == 2) /* Embedded follows */
4536 if (flags & SCF_DO_SUBSTR) {
4537 SCAN_COMMIT(pRExC_state,data,minlenp);
4538 data->longest = &(data->longest_float);
4540 is_inf = is_inf_internal = 1;
4541 if (flags & SCF_DO_STCLASS_OR) /* Allow everything */
4542 cl_anything(pRExC_state, data->start_class);
4543 flags &= ~SCF_DO_STCLASS;
4545 else if (OP(scan) == GPOS) {
4546 if (!(RExC_rx->extflags & RXf_GPOS_FLOAT) &&
4547 !(delta || is_inf || (data && data->pos_delta)))
4549 if (!(RExC_rx->extflags & RXf_ANCH) && (flags & SCF_DO_SUBSTR))
4550 RExC_rx->extflags |= RXf_ANCH_GPOS;
4551 if (RExC_rx->gofs < (U32)min)
4552 RExC_rx->gofs = min;
4554 RExC_rx->extflags |= RXf_GPOS_FLOAT;
4558 #ifdef TRIE_STUDY_OPT
4559 #ifdef FULL_TRIE_STUDY
4560 else if (PL_regkind[OP(scan)] == TRIE) {
4561 /* NOTE - There is similar code to this block above for handling
4562 BRANCH nodes on the initial study. If you change stuff here
4564 regnode *trie_node= scan;
4565 regnode *tail= regnext(scan);
4566 reg_trie_data *trie = (reg_trie_data*)RExC_rxi->data->data[ ARG(scan) ];
4567 I32 max1 = 0, min1 = I32_MAX;
4568 struct regnode_charclass_class accum;
4570 if (flags & SCF_DO_SUBSTR) /* XXXX Add !SUSPEND? */
4571 SCAN_COMMIT(pRExC_state, data,minlenp); /* Cannot merge strings after this. */
4572 if (flags & SCF_DO_STCLASS)
4573 cl_init_zero(pRExC_state, &accum);
4579 const regnode *nextbranch= NULL;
4582 for ( word=1 ; word <= trie->wordcount ; word++)
4584 I32 deltanext=0, minnext=0, f = 0, fake;
4585 struct regnode_charclass_class this_class;
4587 data_fake.flags = 0;
4589 data_fake.whilem_c = data->whilem_c;
4590 data_fake.last_closep = data->last_closep;
4593 data_fake.last_closep = &fake;
4594 data_fake.pos_delta = delta;
4595 if (flags & SCF_DO_STCLASS) {
4596 cl_init(pRExC_state, &this_class);
4597 data_fake.start_class = &this_class;
4598 f = SCF_DO_STCLASS_AND;
4600 if (flags & SCF_WHILEM_VISITED_POS)
4601 f |= SCF_WHILEM_VISITED_POS;
4603 if (trie->jump[word]) {
4605 nextbranch = trie_node + trie->jump[0];
4606 scan= trie_node + trie->jump[word];
4607 /* We go from the jump point to the branch that follows
4608 it. Note this means we need the vestigal unused branches
4609 even though they arent otherwise used.
4611 minnext = study_chunk(pRExC_state, &scan, minlenp,
4612 &deltanext, (regnode *)nextbranch, &data_fake,
4613 stopparen, recursed, NULL, f,depth+1);
4615 if (nextbranch && PL_regkind[OP(nextbranch)]==BRANCH)
4616 nextbranch= regnext((regnode*)nextbranch);
4618 if (min1 > (I32)(minnext + trie->minlen))
4619 min1 = minnext + trie->minlen;
4620 if (deltanext == I32_MAX) {
4621 is_inf = is_inf_internal = 1;
4623 } else if (max1 < (I32)(minnext + deltanext + trie->maxlen))
4624 max1 = minnext + deltanext + trie->maxlen;
4626 if (data_fake.flags & (SF_HAS_PAR|SF_IN_PAR))
4628 if (data_fake.flags & SCF_SEEN_ACCEPT) {
4629 if ( stopmin > min + min1)
4630 stopmin = min + min1;
4631 flags &= ~SCF_DO_SUBSTR;
4633 data->flags |= SCF_SEEN_ACCEPT;
4636 if (data_fake.flags & SF_HAS_EVAL)
4637 data->flags |= SF_HAS_EVAL;
4638 data->whilem_c = data_fake.whilem_c;
4640 if (flags & SCF_DO_STCLASS)
4641 cl_or(pRExC_state, &accum, &this_class);
4644 if (flags & SCF_DO_SUBSTR) {
4645 data->pos_min += min1;
4646 data->pos_delta += max1 - min1;
4647 if (max1 != min1 || is_inf)
4648 data->longest = &(data->longest_float);
4651 delta += max1 - min1;
4652 if (flags & SCF_DO_STCLASS_OR) {
4653 cl_or(pRExC_state, data->start_class, &accum);
4655 cl_and(data->start_class, and_withp);
4656 flags &= ~SCF_DO_STCLASS;
4659 else if (flags & SCF_DO_STCLASS_AND) {
4661 cl_and(data->start_class, &accum);
4662 flags &= ~SCF_DO_STCLASS;
4665 /* Switch to OR mode: cache the old value of
4666 * data->start_class */
4668 StructCopy(data->start_class, and_withp,
4669 struct regnode_charclass_class);
4670 flags &= ~SCF_DO_STCLASS_AND;
4671 StructCopy(&accum, data->start_class,
4672 struct regnode_charclass_class);
4673 flags |= SCF_DO_STCLASS_OR;
4674 SET_SSC_EOS(data->start_class);
4681 else if (PL_regkind[OP(scan)] == TRIE) {
4682 reg_trie_data *trie = (reg_trie_data*)RExC_rxi->data->data[ ARG(scan) ];
4685 min += trie->minlen;
4686 delta += (trie->maxlen - trie->minlen);
4687 flags &= ~SCF_DO_STCLASS; /* xxx */
4688 if (flags & SCF_DO_SUBSTR) {
4689 SCAN_COMMIT(pRExC_state,data,minlenp); /* Cannot expect anything... */
4690 data->pos_min += trie->minlen;
4691 data->pos_delta += (trie->maxlen - trie->minlen);
4692 if (trie->maxlen != trie->minlen)
4693 data->longest = &(data->longest_float);
4695 if (trie->jump) /* no more substrings -- for now /grr*/
4696 flags &= ~SCF_DO_SUBSTR;
4698 #endif /* old or new */
4699 #endif /* TRIE_STUDY_OPT */
4701 /* Else: zero-length, ignore. */
4702 scan = regnext(scan);
4707 stopparen = frame->stop;
4708 frame = frame->prev;
4709 goto fake_study_recurse;
4714 DEBUG_STUDYDATA("pre-fin:",data,depth);
4717 *deltap = is_inf_internal ? I32_MAX : delta;
4718 if (flags & SCF_DO_SUBSTR && is_inf)
4719 data->pos_delta = I32_MAX - data->pos_min;
4720 if (is_par > (I32)U8_MAX)
4722 if (is_par && pars==1 && data) {
4723 data->flags |= SF_IN_PAR;
4724 data->flags &= ~SF_HAS_PAR;
4726 else if (pars && data) {
4727 data->flags |= SF_HAS_PAR;
4728 data->flags &= ~SF_IN_PAR;
4730 if (flags & SCF_DO_STCLASS_OR)
4731 cl_and(data->start_class, and_withp);
4732 if (flags & SCF_TRIE_RESTUDY)
4733 data->flags |= SCF_TRIE_RESTUDY;
4735 DEBUG_STUDYDATA("post-fin:",data,depth);
4737 return min < stopmin ? min : stopmin;
4741 S_add_data(RExC_state_t *pRExC_state, U32 n, const char *s)
4743 U32 count = RExC_rxi->data ? RExC_rxi->data->count : 0;
4745 PERL_ARGS_ASSERT_ADD_DATA;
4747 Renewc(RExC_rxi->data,
4748 sizeof(*RExC_rxi->data) + sizeof(void*) * (count + n - 1),
4749 char, struct reg_data);
4751 Renew(RExC_rxi->data->what, count + n, U8);
4753 Newx(RExC_rxi->data->what, n, U8);
4754 RExC_rxi->data->count = count + n;
4755 Copy(s, RExC_rxi->data->what + count, n, U8);
4759 /*XXX: todo make this not included in a non debugging perl */
4760 #ifndef PERL_IN_XSUB_RE
4762 Perl_reginitcolors(pTHX)
4765 const char * const s = PerlEnv_getenv("PERL_RE_COLORS");
4767 char *t = savepv(s);
4771 t = strchr(t, '\t');
4777 PL_colors[i] = t = (char *)"";
4782 PL_colors[i++] = (char *)"";
4789 #ifdef TRIE_STUDY_OPT
4790 #define CHECK_RESTUDY_GOTO_butfirst(dOsomething) \
4793 (data.flags & SCF_TRIE_RESTUDY) \
4801 #define CHECK_RESTUDY_GOTO_butfirst
4805 * pregcomp - compile a regular expression into internal code
4807 * Decides which engine's compiler to call based on the hint currently in
4811 #ifndef PERL_IN_XSUB_RE
4813 /* return the currently in-scope regex engine (or the default if none) */
4815 regexp_engine const *
4816 Perl_current_re_engine(pTHX)
4820 if (IN_PERL_COMPILETIME) {
4821 HV * const table = GvHV(PL_hintgv);
4825 return &reh_regexp_engine;
4826 ptr = hv_fetchs(table, "regcomp", FALSE);
4827 if ( !(ptr && SvIOK(*ptr) && SvIV(*ptr)))
4828 return &reh_regexp_engine;
4829 return INT2PTR(regexp_engine*,SvIV(*ptr));
4833 if (!PL_curcop->cop_hints_hash)
4834 return &reh_regexp_engine;
4835 ptr = cop_hints_fetch_pvs(PL_curcop, "regcomp", 0);
4836 if ( !(ptr && SvIOK(ptr) && SvIV(ptr)))
4837 return &reh_regexp_engine;
4838 return INT2PTR(regexp_engine*,SvIV(ptr));
4844 Perl_pregcomp(pTHX_ SV * const pattern, const U32 flags)
4847 regexp_engine const *eng = current_re_engine();
4848 GET_RE_DEBUG_FLAGS_DECL;
4850 PERL_ARGS_ASSERT_PREGCOMP;
4852 /* Dispatch a request to compile a regexp to correct regexp engine. */
4854 PerlIO_printf(Perl_debug_log, "Using engine %"UVxf"\n",
4857 return CALLREGCOMP_ENG(eng, pattern, flags);
4861 /* public(ish) entry point for the perl core's own regex compiling code.
4862 * It's actually a wrapper for Perl_re_op_compile that only takes an SV
4863 * pattern rather than a list of OPs, and uses the internal engine rather
4864 * than the current one */
4867 Perl_re_compile(pTHX_ SV * const pattern, U32 rx_flags)
4869 SV *pat = pattern; /* defeat constness! */
4870 PERL_ARGS_ASSERT_RE_COMPILE;
4871 return Perl_re_op_compile(aTHX_ &pat, 1, NULL,
4872 #ifdef PERL_IN_XSUB_RE
4877 NULL, NULL, rx_flags, 0);
4881 /* upgrade pattern pat_p of length plen_p to UTF8, and if there are code
4882 * blocks, recalculate the indices. Update pat_p and plen_p in-place to
4883 * point to the realloced string and length.
4885 * This is essentially a copy of Perl_bytes_to_utf8() with the code index
4889 S_pat_upgrade_to_utf8(pTHX_ RExC_state_t * const pRExC_state,
4890 char **pat_p, STRLEN *plen_p, int num_code_blocks)
4892 U8 *const src = (U8*)*pat_p;
4895 STRLEN s = 0, d = 0;
4897 GET_RE_DEBUG_FLAGS_DECL;
4899 DEBUG_PARSE_r(PerlIO_printf(Perl_debug_log,
4900 "UTF8 mismatch! Converting to utf8 for resizing and compile\n"));
4902 Newx(dst, *plen_p * 2 + 1, U8);
4904 while (s < *plen_p) {
4905 const UV uv = NATIVE_TO_ASCII(src[s]);
4906 if (UNI_IS_INVARIANT(uv))
4907 dst[d] = (U8)UTF_TO_NATIVE(uv);
4909 dst[d++] = (U8)UTF8_EIGHT_BIT_HI(uv);
4910 dst[d] = (U8)UTF8_EIGHT_BIT_LO(uv);
4912 if (n < num_code_blocks) {
4913 if (!do_end && pRExC_state->code_blocks[n].start == s) {
4914 pRExC_state->code_blocks[n].start = d;
4915 assert(dst[d] == '(');
4918 else if (do_end && pRExC_state->code_blocks[n].end == s) {
4919 pRExC_state->code_blocks[n].end = d;
4920 assert(dst[d] == ')');
4930 *pat_p = (char*) dst;
4932 RExC_orig_utf8 = RExC_utf8 = 1;
4937 /* S_concat_pat(): concatenate a list of args to the pattern string pat,
4938 * while recording any code block indices, and handling overloading,
4939 * nested qr// objects etc. If pat is null, it will allocate a new
4940 * string, or just return the first arg, if there's only one.
4942 * Returns the malloced/updated pat.
4943 * patternp and pat_count is the array of SVs to be concatted;
4944 * oplist is the optional list of ops that generated the SVs;
4945 * recompile_p is a pointer to a boolean that will be set if
4946 * the regex will need to be recompiled.
4947 * delim, if non-null is an SV that will be inserted between each element
4951 S_concat_pat(pTHX_ RExC_state_t * const pRExC_state,
4952 SV *pat, SV ** const patternp, int pat_count,
4953 OP *oplist, bool *recompile_p, SV *delim)
4957 bool use_delim = FALSE;
4958 bool alloced = FALSE;
4960 /* if we know we have at least two args, create an empty string,
4961 * then concatenate args to that. For no args, return an empty string */
4962 if (!pat && pat_count != 1) {
4963 pat = newSVpvn("", 0);
4968 for (svp = patternp; svp < patternp + pat_count; svp++) {
4971 STRLEN orig_patlen = 0;
4973 SV *msv = use_delim ? delim : *svp;
4975 /* if we've got a delimiter, we go round the loop twice for each
4976 * svp slot (except the last), using the delimiter the second
4985 if (SvTYPE(msv) == SVt_PVAV) {
4986 /* we've encountered an interpolated array within
4987 * the pattern, e.g. /...@a..../. Expand the list of elements,
4988 * then recursively append elements.
4989 * The code in this block is based on S_pushav() */
4991 AV *const av = (AV*)msv;
4992 const I32 maxarg = AvFILL(av) + 1;
4996 assert(oplist->op_type == OP_PADAV
4997 || oplist->op_type == OP_RV2AV);
4998 oplist = oplist->op_sibling;;
5001 if (SvRMAGICAL(av)) {
5004 Newx(array, maxarg, SV*);
5006 for (i=0; i < (U32)maxarg; i++) {
5007 SV ** const svp = av_fetch(av, i, FALSE);
5008 array[i] = svp ? *svp : &PL_sv_undef;
5012 array = AvARRAY(av);
5014 pat = S_concat_pat(aTHX_ pRExC_state, pat,
5015 array, maxarg, NULL, recompile_p,
5017 GvSV((gv_fetchpvs("\"", GV_ADDMULTI, SVt_PV))));
5023 /* we make the assumption here that each op in the list of
5024 * op_siblings maps to one SV pushed onto the stack,
5025 * except for code blocks, with have both an OP_NULL and
5027 * This allows us to match up the list of SVs against the
5028 * list of OPs to find the next code block.
5030 * Note that PUSHMARK PADSV PADSV ..
5032 * PADRANGE PADSV PADSV ..
5033 * so the alignment still works. */
5036 if (oplist->op_type == OP_NULL
5037 && (oplist->op_flags & OPf_SPECIAL))
5039 assert(n < pRExC_state->num_code_blocks);
5040 pRExC_state->code_blocks[n].start = pat ? SvCUR(pat) : 0;
5041 pRExC_state->code_blocks[n].block = oplist;
5042 pRExC_state->code_blocks[n].src_regex = NULL;
5045 oplist = oplist->op_sibling; /* skip CONST */
5048 oplist = oplist->op_sibling;;
5051 /* apply magic and QR overloading to arg */
5054 if (SvROK(msv) && SvAMAGIC(msv)) {
5055 SV *sv = AMG_CALLunary(msv, regexp_amg);
5059 if (SvTYPE(sv) != SVt_REGEXP)
5060 Perl_croak(aTHX_ "Overloaded qr did not return a REGEXP");
5065 /* try concatenation overload ... */
5066 if (pat && (SvAMAGIC(pat) || SvAMAGIC(msv)) &&
5067 (sv = amagic_call(pat, msv, concat_amg, AMGf_assign)))
5070 /* overloading involved: all bets are off over literal
5071 * code. Pretend we haven't seen it */
5072 pRExC_state->num_code_blocks -= n;
5076 /* ... or failing that, try "" overload */
5077 while (SvAMAGIC(msv)
5078 && (sv = AMG_CALLunary(msv, string_amg))
5082 && SvRV(msv) == SvRV(sv))
5087 if (SvROK(msv) && SvTYPE(SvRV(msv)) == SVt_REGEXP)
5091 /* this is a partially unrolled
5092 * sv_catsv_nomg(pat, msv);
5093 * that allows us to adjust code block indices if
5096 char *dst = SvPV_force_nomg(pat, dlen);
5097 const char *src = SvPV_flags_const(msv, slen, 0);
5099 if (SvUTF8(msv) && !SvUTF8(pat)) {
5100 S_pat_upgrade_to_utf8(aTHX_ pRExC_state, &dst, &dlen, n);
5101 sv_setpvn(pat, dst, dlen);
5104 sv_catpvn_nomg(pat, src, slen);
5111 pRExC_state->code_blocks[n-1].end = SvCUR(pat)-1;
5114 /* extract any code blocks within any embedded qr//'s */
5115 if (rx && SvTYPE(rx) == SVt_REGEXP
5116 && RX_ENGINE((REGEXP*)rx)->op_comp)
5119 RXi_GET_DECL(ReANY((REGEXP *)rx), ri);
5120 if (ri->num_code_blocks) {
5122 /* the presence of an embedded qr// with code means
5123 * we should always recompile: the text of the
5124 * qr// may not have changed, but it may be a
5125 * different closure than last time */
5127 Renew(pRExC_state->code_blocks,
5128 pRExC_state->num_code_blocks + ri->num_code_blocks,
5129 struct reg_code_block);
5130 pRExC_state->num_code_blocks += ri->num_code_blocks;
5132 for (i=0; i < ri->num_code_blocks; i++) {
5133 struct reg_code_block *src, *dst;
5134 STRLEN offset = orig_patlen
5135 + ReANY((REGEXP *)rx)->pre_prefix;
5136 assert(n < pRExC_state->num_code_blocks);
5137 src = &ri->code_blocks[i];
5138 dst = &pRExC_state->code_blocks[n];
5139 dst->start = src->start + offset;
5140 dst->end = src->end + offset;
5141 dst->block = src->block;
5142 dst->src_regex = (REGEXP*) SvREFCNT_inc( (SV*)
5151 /* avoid calling magic multiple times on a single element e.g. =~ $qr */
5160 /* see if there are any run-time code blocks in the pattern.
5161 * False positives are allowed */
5164 S_has_runtime_code(pTHX_ RExC_state_t * const pRExC_state,
5165 char *pat, STRLEN plen)
5170 for (s = 0; s < plen; s++) {
5171 if (n < pRExC_state->num_code_blocks
5172 && s == pRExC_state->code_blocks[n].start)
5174 s = pRExC_state->code_blocks[n].end;
5178 /* TODO ideally should handle [..], (#..), /#.../x to reduce false
5180 if (pat[s] == '(' && s+2 <= plen && pat[s+1] == '?' &&
5182 || (s + 2 <= plen && pat[s+2] == '?' && pat[s+3] == '{'))
5189 /* Handle run-time code blocks. We will already have compiled any direct
5190 * or indirect literal code blocks. Now, take the pattern 'pat' and make a
5191 * copy of it, but with any literal code blocks blanked out and
5192 * appropriate chars escaped; then feed it into
5194 * eval "qr'modified_pattern'"
5198 * a\bc(?{"this was literal"})def'ghi\\jkl(?{"this is runtime"})mno
5202 * qr'a\\bc_______________________def\'ghi\\\\jkl(?{"this is runtime"})mno'
5204 * After eval_sv()-ing that, grab any new code blocks from the returned qr
5205 * and merge them with any code blocks of the original regexp.
5207 * If the pat is non-UTF8, while the evalled qr is UTF8, don't merge;
5208 * instead, just save the qr and return FALSE; this tells our caller that
5209 * the original pattern needs upgrading to utf8.
5213 S_compile_runtime_code(pTHX_ RExC_state_t * const pRExC_state,
5214 char *pat, STRLEN plen)
5218 GET_RE_DEBUG_FLAGS_DECL;
5220 if (pRExC_state->runtime_code_qr) {
5221 /* this is the second time we've been called; this should
5222 * only happen if the main pattern got upgraded to utf8
5223 * during compilation; re-use the qr we compiled first time
5224 * round (which should be utf8 too)
5226 qr = pRExC_state->runtime_code_qr;
5227 pRExC_state->runtime_code_qr = NULL;
5228 assert(RExC_utf8 && SvUTF8(qr));
5234 int newlen = plen + 6; /* allow for "qr''x\0" extra chars */
5238 /* determine how many extra chars we need for ' and \ escaping */
5239 for (s = 0; s < plen; s++) {
5240 if (pat[s] == '\'' || pat[s] == '\\')
5244 Newx(newpat, newlen, char);
5246 *p++ = 'q'; *p++ = 'r'; *p++ = '\'';
5248 for (s = 0; s < plen; s++) {
5249 if (n < pRExC_state->num_code_blocks
5250 && s == pRExC_state->code_blocks[n].start)
5252 /* blank out literal code block */
5253 assert(pat[s] == '(');
5254 while (s <= pRExC_state->code_blocks[n].end) {
5262 if (pat[s] == '\'' || pat[s] == '\\')
5267 if (pRExC_state->pm_flags & RXf_PMf_EXTENDED)
5271 PerlIO_printf(Perl_debug_log,
5272 "%sre-parsing pattern for runtime code:%s %s\n",
5273 PL_colors[4],PL_colors[5],newpat);
5276 sv = newSVpvn_flags(newpat, p-newpat-1, RExC_utf8 ? SVf_UTF8 : 0);
5282 PUSHSTACKi(PERLSI_REQUIRE);
5283 /* G_RE_REPARSING causes the toker to collapse \\ into \ when
5284 * parsing qr''; normally only q'' does this. It also alters
5286 eval_sv(sv, G_SCALAR|G_RE_REPARSING);
5287 SvREFCNT_dec_NN(sv);
5292 SV * const errsv = ERRSV;
5293 if (SvTRUE_NN(errsv))
5295 Safefree(pRExC_state->code_blocks);
5296 /* use croak_sv ? */
5297 Perl_croak_nocontext("%s", SvPV_nolen_const(errsv));
5300 assert(SvROK(qr_ref));
5302 assert(SvTYPE(qr) == SVt_REGEXP && RX_ENGINE((REGEXP*)qr)->op_comp);
5303 /* the leaving below frees the tmp qr_ref.
5304 * Give qr a life of its own */
5312 if (!RExC_utf8 && SvUTF8(qr)) {
5313 /* first time through; the pattern got upgraded; save the
5314 * qr for the next time through */
5315 assert(!pRExC_state->runtime_code_qr);
5316 pRExC_state->runtime_code_qr = qr;
5321 /* extract any code blocks within the returned qr// */
5324 /* merge the main (r1) and run-time (r2) code blocks into one */
5326 RXi_GET_DECL(ReANY((REGEXP *)qr), r2);
5327 struct reg_code_block *new_block, *dst;
5328 RExC_state_t * const r1 = pRExC_state; /* convenient alias */
5331 if (!r2->num_code_blocks) /* we guessed wrong */
5333 SvREFCNT_dec_NN(qr);
5338 r1->num_code_blocks + r2->num_code_blocks,
5339 struct reg_code_block);
5342 while ( i1 < r1->num_code_blocks
5343 || i2 < r2->num_code_blocks)
5345 struct reg_code_block *src;
5348 if (i1 == r1->num_code_blocks) {
5349 src = &r2->code_blocks[i2++];
5352 else if (i2 == r2->num_code_blocks)
5353 src = &r1->code_blocks[i1++];
5354 else if ( r1->code_blocks[i1].start
5355 < r2->code_blocks[i2].start)
5357 src = &r1->code_blocks[i1++];
5358 assert(src->end < r2->code_blocks[i2].start);
5361 assert( r1->code_blocks[i1].start
5362 > r2->code_blocks[i2].start);
5363 src = &r2->code_blocks[i2++];
5365 assert(src->end < r1->code_blocks[i1].start);
5368 assert(pat[src->start] == '(');
5369 assert(pat[src->end] == ')');
5370 dst->start = src->start;
5371 dst->end = src->end;
5372 dst->block = src->block;
5373 dst->src_regex = is_qr ? (REGEXP*) SvREFCNT_inc( (SV*) qr)
5377 r1->num_code_blocks += r2->num_code_blocks;
5378 Safefree(r1->code_blocks);
5379 r1->code_blocks = new_block;
5382 SvREFCNT_dec_NN(qr);
5388 S_setup_longest(pTHX_ RExC_state_t *pRExC_state, SV* sv_longest, SV** rx_utf8, SV** rx_substr, I32* rx_end_shift, I32 lookbehind, I32 offset, I32 *minlen, STRLEN longest_length, bool eol, bool meol)
5390 /* This is the common code for setting up the floating and fixed length
5391 * string data extracted from Perl_re_op_compile() below. Returns a boolean
5392 * as to whether succeeded or not */
5396 if (! (longest_length
5397 || (eol /* Can't have SEOL and MULTI */
5398 && (! meol || (RExC_flags & RXf_PMf_MULTILINE)))
5400 /* See comments for join_exact for why REG_SEEN_EXACTF_SHARP_S */
5401 || (RExC_seen & REG_SEEN_EXACTF_SHARP_S))
5406 /* copy the information about the longest from the reg_scan_data
5407 over to the program. */
5408 if (SvUTF8(sv_longest)) {
5409 *rx_utf8 = sv_longest;
5412 *rx_substr = sv_longest;
5415 /* end_shift is how many chars that must be matched that
5416 follow this item. We calculate it ahead of time as once the
5417 lookbehind offset is added in we lose the ability to correctly
5419 ml = minlen ? *(minlen) : (I32)longest_length;
5420 *rx_end_shift = ml - offset
5421 - longest_length + (SvTAIL(sv_longest) != 0)
5424 t = (eol/* Can't have SEOL and MULTI */
5425 && (! meol || (RExC_flags & RXf_PMf_MULTILINE)));
5426 fbm_compile(sv_longest, t ? FBMcf_TAIL : 0);
5432 * Perl_re_op_compile - the perl internal RE engine's function to compile a
5433 * regular expression into internal code.
5434 * The pattern may be passed either as:
5435 * a list of SVs (patternp plus pat_count)
5436 * a list of OPs (expr)
5437 * If both are passed, the SV list is used, but the OP list indicates
5438 * which SVs are actually pre-compiled code blocks
5440 * The SVs in the list have magic and qr overloading applied to them (and
5441 * the list may be modified in-place with replacement SVs in the latter
5444 * If the pattern hasn't changed from old_re, then old_re will be
5447 * eng is the current engine. If that engine has an op_comp method, then
5448 * handle directly (i.e. we assume that op_comp was us); otherwise, just
5449 * do the initial concatenation of arguments and pass on to the external
5452 * If is_bare_re is not null, set it to a boolean indicating whether the
5453 * arg list reduced (after overloading) to a single bare regex which has
5454 * been returned (i.e. /$qr/).
5456 * orig_rx_flags contains RXf_* flags. See perlreapi.pod for more details.
5458 * pm_flags contains the PMf_* flags, typically based on those from the
5459 * pm_flags field of the related PMOP. Currently we're only interested in
5460 * PMf_HAS_CV, PMf_IS_QR, PMf_USE_RE_EVAL.
5462 * We can't allocate space until we know how big the compiled form will be,
5463 * but we can't compile it (and thus know how big it is) until we've got a
5464 * place to put the code. So we cheat: we compile it twice, once with code
5465 * generation turned off and size counting turned on, and once "for real".
5466 * This also means that we don't allocate space until we are sure that the
5467 * thing really will compile successfully, and we never have to move the
5468 * code and thus invalidate pointers into it. (Note that it has to be in
5469 * one piece because free() must be able to free it all.) [NB: not true in perl]
5471 * Beware that the optimization-preparation code in here knows about some
5472 * of the structure of the compiled regexp. [I'll say.]
5476 Perl_re_op_compile(pTHX_ SV ** const patternp, int pat_count,
5477 OP *expr, const regexp_engine* eng, REGEXP *old_re,
5478 bool *is_bare_re, U32 orig_rx_flags, U32 pm_flags)
5483 regexp_internal *ri;
5491 SV *code_blocksv = NULL;
5492 SV** new_patternp = patternp;
5494 /* these are all flags - maybe they should be turned
5495 * into a single int with different bit masks */
5496 I32 sawlookahead = 0;
5499 regex_charset initial_charset = get_regex_charset(orig_rx_flags);
5501 bool runtime_code = 0;
5503 RExC_state_t RExC_state;
5504 RExC_state_t * const pRExC_state = &RExC_state;
5505 #ifdef TRIE_STUDY_OPT
5507 RExC_state_t copyRExC_state;
5509 GET_RE_DEBUG_FLAGS_DECL;
5511 PERL_ARGS_ASSERT_RE_OP_COMPILE;
5513 DEBUG_r(if (!PL_colorset) reginitcolors());
5515 #ifndef PERL_IN_XSUB_RE
5516 /* Initialize these here instead of as-needed, as is quick and avoids
5517 * having to test them each time otherwise */
5518 if (! PL_AboveLatin1) {
5519 PL_AboveLatin1 = _new_invlist_C_array(AboveLatin1_invlist);
5520 PL_ASCII = _new_invlist_C_array(ASCII_invlist);
5521 PL_Latin1 = _new_invlist_C_array(Latin1_invlist);
5523 PL_L1Posix_ptrs[_CC_ALPHANUMERIC]
5524 = _new_invlist_C_array(L1PosixAlnum_invlist);
5525 PL_Posix_ptrs[_CC_ALPHANUMERIC]
5526 = _new_invlist_C_array(PosixAlnum_invlist);
5528 PL_L1Posix_ptrs[_CC_ALPHA]
5529 = _new_invlist_C_array(L1PosixAlpha_invlist);
5530 PL_Posix_ptrs[_CC_ALPHA] = _new_invlist_C_array(PosixAlpha_invlist);
5532 PL_Posix_ptrs[_CC_BLANK] = _new_invlist_C_array(PosixBlank_invlist);
5533 PL_XPosix_ptrs[_CC_BLANK] = _new_invlist_C_array(XPosixBlank_invlist);
5535 /* Cased is the same as Alpha in the ASCII range */
5536 PL_L1Posix_ptrs[_CC_CASED] = _new_invlist_C_array(L1Cased_invlist);
5537 PL_Posix_ptrs[_CC_CASED] = _new_invlist_C_array(PosixAlpha_invlist);
5539 PL_Posix_ptrs[_CC_CNTRL] = _new_invlist_C_array(PosixCntrl_invlist);
5540 PL_XPosix_ptrs[_CC_CNTRL] = _new_invlist_C_array(XPosixCntrl_invlist);
5542 PL_Posix_ptrs[_CC_DIGIT] = _new_invlist_C_array(PosixDigit_invlist);
5543 PL_L1Posix_ptrs[_CC_DIGIT] = _new_invlist_C_array(PosixDigit_invlist);
5545 PL_L1Posix_ptrs[_CC_GRAPH] = _new_invlist_C_array(L1PosixGraph_invlist);
5546 PL_Posix_ptrs[_CC_GRAPH] = _new_invlist_C_array(PosixGraph_invlist);
5548 PL_L1Posix_ptrs[_CC_LOWER] = _new_invlist_C_array(L1PosixLower_invlist);
5549 PL_Posix_ptrs[_CC_LOWER] = _new_invlist_C_array(PosixLower_invlist);
5551 PL_L1Posix_ptrs[_CC_PRINT] = _new_invlist_C_array(L1PosixPrint_invlist);
5552 PL_Posix_ptrs[_CC_PRINT] = _new_invlist_C_array(PosixPrint_invlist);
5554 PL_L1Posix_ptrs[_CC_PUNCT] = _new_invlist_C_array(L1PosixPunct_invlist);
5555 PL_Posix_ptrs[_CC_PUNCT] = _new_invlist_C_array(PosixPunct_invlist);
5557 PL_Posix_ptrs[_CC_SPACE] = _new_invlist_C_array(PerlSpace_invlist);
5558 PL_XPosix_ptrs[_CC_SPACE] = _new_invlist_C_array(XPerlSpace_invlist);
5559 PL_Posix_ptrs[_CC_PSXSPC] = _new_invlist_C_array(PosixSpace_invlist);
5560 PL_XPosix_ptrs[_CC_PSXSPC] = _new_invlist_C_array(XPosixSpace_invlist);
5562 PL_L1Posix_ptrs[_CC_UPPER] = _new_invlist_C_array(L1PosixUpper_invlist);
5563 PL_Posix_ptrs[_CC_UPPER] = _new_invlist_C_array(PosixUpper_invlist);
5565 PL_XPosix_ptrs[_CC_VERTSPACE] = _new_invlist_C_array(VertSpace_invlist);
5567 PL_Posix_ptrs[_CC_WORDCHAR] = _new_invlist_C_array(PosixWord_invlist);
5568 PL_L1Posix_ptrs[_CC_WORDCHAR]
5569 = _new_invlist_C_array(L1PosixWord_invlist);
5571 PL_Posix_ptrs[_CC_XDIGIT] = _new_invlist_C_array(PosixXDigit_invlist);
5572 PL_XPosix_ptrs[_CC_XDIGIT] = _new_invlist_C_array(XPosixXDigit_invlist);
5574 PL_HasMultiCharFold = _new_invlist_C_array(_Perl_Multi_Char_Folds_invlist);
5578 pRExC_state->code_blocks = NULL;
5579 pRExC_state->num_code_blocks = 0;
5582 *is_bare_re = FALSE;
5584 if (expr && (expr->op_type == OP_LIST ||
5585 (expr->op_type == OP_NULL && expr->op_targ == OP_LIST))) {
5586 /* allocate code_blocks if needed */
5590 for (o = cLISTOPx(expr)->op_first; o; o = o->op_sibling)
5591 if (o->op_type == OP_NULL && (o->op_flags & OPf_SPECIAL))
5592 ncode++; /* count of DO blocks */
5594 pRExC_state->num_code_blocks = ncode;
5595 Newx(pRExC_state->code_blocks, ncode, struct reg_code_block);
5600 /* compile-time pattern with just OP_CONSTs and DO blocks */
5605 /* find how many CONSTs there are */
5608 if (expr->op_type == OP_CONST)
5611 for (o = cLISTOPx(expr)->op_first; o; o = o->op_sibling) {
5612 if (o->op_type == OP_CONST)
5616 /* fake up an SV array */
5618 assert(!new_patternp);
5619 Newx(new_patternp, n, SV*);
5620 SAVEFREEPV(new_patternp);
5624 if (expr->op_type == OP_CONST)
5625 new_patternp[n] = cSVOPx_sv(expr);
5627 for (o = cLISTOPx(expr)->op_first; o; o = o->op_sibling) {
5628 if (o->op_type == OP_CONST)
5629 new_patternp[n++] = cSVOPo_sv;
5634 DEBUG_PARSE_r(PerlIO_printf(Perl_debug_log,
5635 "Assembling pattern from %d elements%s\n", pat_count,
5636 orig_rx_flags & RXf_SPLIT ? " for split" : ""));
5638 /* set expr to the first arg op */
5640 if (pRExC_state->num_code_blocks
5641 && expr->op_type != OP_CONST)
5643 expr = cLISTOPx(expr)->op_first;
5644 assert( expr->op_type == OP_PUSHMARK
5645 || (expr->op_type == OP_NULL && expr->op_targ == OP_PUSHMARK)
5646 || expr->op_type == OP_PADRANGE);
5647 expr = expr->op_sibling;
5650 pat = S_concat_pat(aTHX_ pRExC_state, NULL, new_patternp, pat_count,
5651 expr, &recompile, NULL);
5653 /* handle bare (possibly after overloading) regex: foo =~ $re */
5658 if (SvTYPE(re) == SVt_REGEXP) {
5662 Safefree(pRExC_state->code_blocks);
5663 DEBUG_PARSE_r(PerlIO_printf(Perl_debug_log,
5664 "Precompiled pattern%s\n",
5665 orig_rx_flags & RXf_SPLIT ? " for split" : ""));
5671 exp = SvPV_nomg(pat, plen);
5673 if (!eng->op_comp) {
5674 if ((SvUTF8(pat) && IN_BYTES)
5675 || SvGMAGICAL(pat) || SvAMAGIC(pat))
5677 /* make a temporary copy; either to convert to bytes,
5678 * or to avoid repeating get-magic / overloaded stringify */
5679 pat = newSVpvn_flags(exp, plen, SVs_TEMP |
5680 (IN_BYTES ? 0 : SvUTF8(pat)));
5682 Safefree(pRExC_state->code_blocks);
5683 return CALLREGCOMP_ENG(eng, pat, orig_rx_flags);
5686 /* ignore the utf8ness if the pattern is 0 length */
5687 RExC_utf8 = RExC_orig_utf8 = (plen == 0 || IN_BYTES) ? 0 : SvUTF8(pat);
5688 RExC_uni_semantics = 0;
5689 RExC_contains_locale = 0;
5690 pRExC_state->runtime_code_qr = NULL;
5693 SV *dsv= sv_newmortal();
5694 RE_PV_QUOTED_DECL(s, RExC_utf8, dsv, exp, plen, 60);
5695 PerlIO_printf(Perl_debug_log, "%sCompiling REx%s %s\n",
5696 PL_colors[4],PL_colors[5],s);
5700 /* we jump here if we upgrade the pattern to utf8 and have to
5703 if ((pm_flags & PMf_USE_RE_EVAL)
5704 /* this second condition covers the non-regex literal case,
5705 * i.e. $foo =~ '(?{})'. */
5706 || (IN_PERL_COMPILETIME && (PL_hints & HINT_RE_EVAL))
5708 runtime_code = S_has_runtime_code(aTHX_ pRExC_state, exp, plen);
5710 /* return old regex if pattern hasn't changed */
5711 /* XXX: note in the below we have to check the flags as well as the pattern.
5713 * Things get a touch tricky as we have to compare the utf8 flag independently
5714 * from the compile flags.
5719 && !!RX_UTF8(old_re) == !!RExC_utf8
5720 && ( RX_COMPFLAGS(old_re) == ( orig_rx_flags & RXf_PMf_FLAGCOPYMASK ) )
5721 && RX_PRECOMP(old_re)
5722 && RX_PRELEN(old_re) == plen
5723 && memEQ(RX_PRECOMP(old_re), exp, plen)
5724 && !runtime_code /* with runtime code, always recompile */ )
5726 Safefree(pRExC_state->code_blocks);
5730 rx_flags = orig_rx_flags;
5732 if (initial_charset == REGEX_LOCALE_CHARSET) {
5733 RExC_contains_locale = 1;
5735 else if (RExC_utf8 && initial_charset == REGEX_DEPENDS_CHARSET) {
5737 /* Set to use unicode semantics if the pattern is in utf8 and has the
5738 * 'depends' charset specified, as it means unicode when utf8 */
5739 set_regex_charset(&rx_flags, REGEX_UNICODE_CHARSET);
5743 RExC_flags = rx_flags;
5744 RExC_pm_flags = pm_flags;
5747 if (TAINTING_get && TAINT_get)
5748 Perl_croak(aTHX_ "Eval-group in insecure regular expression");
5750 if (!S_compile_runtime_code(aTHX_ pRExC_state, exp, plen)) {
5751 /* whoops, we have a non-utf8 pattern, whilst run-time code
5752 * got compiled as utf8. Try again with a utf8 pattern */
5753 S_pat_upgrade_to_utf8(aTHX_ pRExC_state, &exp, &plen,
5754 pRExC_state->num_code_blocks);
5755 goto redo_first_pass;
5758 assert(!pRExC_state->runtime_code_qr);
5763 RExC_in_lookbehind = 0;
5764 RExC_seen_zerolen = *exp == '^' ? -1 : 0;
5766 RExC_override_recoding = 0;
5767 RExC_in_multi_char_class = 0;
5769 /* First pass: determine size, legality. */
5772 RExC_end = exp + plen;
5777 RExC_emit = &PL_regdummy;
5778 RExC_whilem_seen = 0;
5779 RExC_open_parens = NULL;
5780 RExC_close_parens = NULL;
5782 RExC_paren_names = NULL;
5784 RExC_paren_name_list = NULL;
5786 RExC_recurse = NULL;
5787 RExC_recurse_count = 0;
5788 pRExC_state->code_index = 0;
5790 #if 0 /* REGC() is (currently) a NOP at the first pass.
5791 * Clever compilers notice this and complain. --jhi */
5792 REGC((U8)REG_MAGIC, (char*)RExC_emit);
5795 PerlIO_printf(Perl_debug_log, "Starting first pass (sizing)\n");
5797 RExC_lastparse=NULL;
5799 /* reg may croak on us, not giving us a chance to free
5800 pRExC_state->code_blocks. We cannot SAVEFREEPV it now, as we may
5801 need it to survive as long as the regexp (qr/(?{})/).
5802 We must check that code_blocksv is not already set, because we may
5803 have jumped back to restart the sizing pass. */
5804 if (pRExC_state->code_blocks && !code_blocksv) {
5805 code_blocksv = newSV_type(SVt_PV);
5806 SAVEFREESV(code_blocksv);
5807 SvPV_set(code_blocksv, (char *)pRExC_state->code_blocks);
5808 SvLEN_set(code_blocksv, 1); /*sufficient to make sv_clear free it*/
5810 if (reg(pRExC_state, 0, &flags,1) == NULL) {
5811 /* It's possible to write a regexp in ascii that represents Unicode
5812 codepoints outside of the byte range, such as via \x{100}. If we
5813 detect such a sequence we have to convert the entire pattern to utf8
5814 and then recompile, as our sizing calculation will have been based
5815 on 1 byte == 1 character, but we will need to use utf8 to encode
5816 at least some part of the pattern, and therefore must convert the whole
5819 if (flags & RESTART_UTF8) {
5820 S_pat_upgrade_to_utf8(aTHX_ pRExC_state, &exp, &plen,
5821 pRExC_state->num_code_blocks);
5822 goto redo_first_pass;
5824 Perl_croak(aTHX_ "panic: reg returned NULL to re_op_compile for sizing pass, flags=%#X", flags);
5827 SvLEN_set(code_blocksv,0); /* no you can't have it, sv_clear */
5830 PerlIO_printf(Perl_debug_log,
5831 "Required size %"IVdf" nodes\n"
5832 "Starting second pass (creation)\n",
5835 RExC_lastparse=NULL;
5838 /* The first pass could have found things that force Unicode semantics */
5839 if ((RExC_utf8 || RExC_uni_semantics)
5840 && get_regex_charset(rx_flags) == REGEX_DEPENDS_CHARSET)
5842 set_regex_charset(&rx_flags, REGEX_UNICODE_CHARSET);
5845 /* Small enough for pointer-storage convention?
5846 If extralen==0, this means that we will not need long jumps. */
5847 if (RExC_size >= 0x10000L && RExC_extralen)
5848 RExC_size += RExC_extralen;
5851 if (RExC_whilem_seen > 15)
5852 RExC_whilem_seen = 15;
5854 /* Allocate space and zero-initialize. Note, the two step process
5855 of zeroing when in debug mode, thus anything assigned has to
5856 happen after that */
5857 rx = (REGEXP*) newSV_type(SVt_REGEXP);
5859 Newxc(ri, sizeof(regexp_internal) + (unsigned)RExC_size * sizeof(regnode),
5860 char, regexp_internal);
5861 if ( r == NULL || ri == NULL )
5862 FAIL("Regexp out of space");
5864 /* avoid reading uninitialized memory in DEBUGGING code in study_chunk() */
5865 Zero(ri, sizeof(regexp_internal) + (unsigned)RExC_size * sizeof(regnode), char);
5867 /* bulk initialize base fields with 0. */
5868 Zero(ri, sizeof(regexp_internal), char);
5871 /* non-zero initialization begins here */
5874 r->extflags = rx_flags;
5875 RXp_COMPFLAGS(r) = orig_rx_flags & RXf_PMf_FLAGCOPYMASK;
5877 if (pm_flags & PMf_IS_QR) {
5878 ri->code_blocks = pRExC_state->code_blocks;
5879 ri->num_code_blocks = pRExC_state->num_code_blocks;
5884 for (n = 0; n < pRExC_state->num_code_blocks; n++)
5885 if (pRExC_state->code_blocks[n].src_regex)
5886 SAVEFREESV(pRExC_state->code_blocks[n].src_regex);
5887 SAVEFREEPV(pRExC_state->code_blocks);
5891 bool has_p = ((r->extflags & RXf_PMf_KEEPCOPY) == RXf_PMf_KEEPCOPY);
5892 bool has_charset = (get_regex_charset(r->extflags) != REGEX_DEPENDS_CHARSET);
5894 /* The caret is output if there are any defaults: if not all the STD
5895 * flags are set, or if no character set specifier is needed */
5897 (((r->extflags & RXf_PMf_STD_PMMOD) != RXf_PMf_STD_PMMOD)
5899 bool has_runon = ((RExC_seen & REG_SEEN_RUN_ON_COMMENT)==REG_SEEN_RUN_ON_COMMENT);
5900 U16 reganch = (U16)((r->extflags & RXf_PMf_STD_PMMOD)
5901 >> RXf_PMf_STD_PMMOD_SHIFT);
5902 const char *fptr = STD_PAT_MODS; /*"msix"*/
5904 /* Allocate for the worst case, which is all the std flags are turned
5905 * on. If more precision is desired, we could do a population count of
5906 * the flags set. This could be done with a small lookup table, or by
5907 * shifting, masking and adding, or even, when available, assembly
5908 * language for a machine-language population count.
5909 * We never output a minus, as all those are defaults, so are
5910 * covered by the caret */
5911 const STRLEN wraplen = plen + has_p + has_runon
5912 + has_default /* If needs a caret */
5914 /* If needs a character set specifier */
5915 + ((has_charset) ? MAX_CHARSET_NAME_LENGTH : 0)
5916 + (sizeof(STD_PAT_MODS) - 1)
5917 + (sizeof("(?:)") - 1);
5919 Newx(p, wraplen + 1, char); /* +1 for the ending NUL */
5920 r->xpv_len_u.xpvlenu_pv = p;
5922 SvFLAGS(rx) |= SVf_UTF8;
5925 /* If a default, cover it using the caret */
5927 *p++= DEFAULT_PAT_MOD;
5931 const char* const name = get_regex_charset_name(r->extflags, &len);
5932 Copy(name, p, len, char);
5936 *p++ = KEEPCOPY_PAT_MOD; /*'p'*/
5939 while((ch = *fptr++)) {
5947 Copy(RExC_precomp, p, plen, char);
5948 assert ((RX_WRAPPED(rx) - p) < 16);
5949 r->pre_prefix = p - RX_WRAPPED(rx);
5955 SvCUR_set(rx, p - RX_WRAPPED(rx));
5959 r->nparens = RExC_npar - 1; /* set early to validate backrefs */
5961 if (RExC_seen & REG_SEEN_RECURSE) {
5962 Newxz(RExC_open_parens, RExC_npar,regnode *);
5963 SAVEFREEPV(RExC_open_parens);
5964 Newxz(RExC_close_parens,RExC_npar,regnode *);
5965 SAVEFREEPV(RExC_close_parens);
5968 /* Useful during FAIL. */
5969 #ifdef RE_TRACK_PATTERN_OFFSETS
5970 Newxz(ri->u.offsets, 2*RExC_size+1, U32); /* MJD 20001228 */
5971 DEBUG_OFFSETS_r(PerlIO_printf(Perl_debug_log,
5972 "%s %"UVuf" bytes for offset annotations.\n",
5973 ri->u.offsets ? "Got" : "Couldn't get",
5974 (UV)((2*RExC_size+1) * sizeof(U32))));
5976 SetProgLen(ri,RExC_size);
5980 REH_CALL_COMP_BEGIN_HOOK(pRExC_state->rx);
5982 /* Second pass: emit code. */
5983 RExC_flags = rx_flags; /* don't let top level (?i) bleed */
5984 RExC_pm_flags = pm_flags;
5986 RExC_end = exp + plen;
5989 RExC_emit_start = ri->program;
5990 RExC_emit = ri->program;
5991 RExC_emit_bound = ri->program + RExC_size + 1;
5992 pRExC_state->code_index = 0;
5994 REGC((U8)REG_MAGIC, (char*) RExC_emit++);
5995 if (reg(pRExC_state, 0, &flags,1) == NULL) {
5997 Perl_croak(aTHX_ "panic: reg returned NULL to re_op_compile for generation pass, flags=%#X", flags);
5999 /* XXXX To minimize changes to RE engine we always allocate
6000 3-units-long substrs field. */
6001 Newx(r->substrs, 1, struct reg_substr_data);
6002 if (RExC_recurse_count) {
6003 Newxz(RExC_recurse,RExC_recurse_count,regnode *);
6004 SAVEFREEPV(RExC_recurse);
6008 r->minlen = minlen = sawlookahead = sawplus = sawopen = 0;
6009 Zero(r->substrs, 1, struct reg_substr_data);
6011 #ifdef TRIE_STUDY_OPT
6013 StructCopy(&zero_scan_data, &data, scan_data_t);
6014 copyRExC_state = RExC_state;
6017 DEBUG_OPTIMISE_r(PerlIO_printf(Perl_debug_log,"Restudying\n"));
6019 RExC_state = copyRExC_state;
6020 if (seen & REG_TOP_LEVEL_BRANCHES)
6021 RExC_seen |= REG_TOP_LEVEL_BRANCHES;
6023 RExC_seen &= ~REG_TOP_LEVEL_BRANCHES;
6024 StructCopy(&zero_scan_data, &data, scan_data_t);
6027 StructCopy(&zero_scan_data, &data, scan_data_t);
6030 /* Dig out information for optimizations. */
6031 r->extflags = RExC_flags; /* was pm_op */
6032 /*dmq: removed as part of de-PMOP: pm->op_pmflags = RExC_flags; */
6035 SvUTF8_on(rx); /* Unicode in it? */
6036 ri->regstclass = NULL;
6037 if (RExC_naughty >= 10) /* Probably an expensive pattern. */
6038 r->intflags |= PREGf_NAUGHTY;
6039 scan = ri->program + 1; /* First BRANCH. */
6041 /* testing for BRANCH here tells us whether there is "must appear"
6042 data in the pattern. If there is then we can use it for optimisations */
6043 if (!(RExC_seen & REG_TOP_LEVEL_BRANCHES)) { /* Only one top-level choice. */
6045 STRLEN longest_float_length, longest_fixed_length;
6046 struct regnode_charclass_class ch_class; /* pointed to by data */
6048 I32 last_close = 0; /* pointed to by data */
6049 regnode *first= scan;
6050 regnode *first_next= regnext(first);
6052 * Skip introductions and multiplicators >= 1
6053 * so that we can extract the 'meat' of the pattern that must
6054 * match in the large if() sequence following.
6055 * NOTE that EXACT is NOT covered here, as it is normally
6056 * picked up by the optimiser separately.
6058 * This is unfortunate as the optimiser isnt handling lookahead
6059 * properly currently.
6062 while ((OP(first) == OPEN && (sawopen = 1)) ||
6063 /* An OR of *one* alternative - should not happen now. */
6064 (OP(first) == BRANCH && OP(first_next) != BRANCH) ||
6065 /* for now we can't handle lookbehind IFMATCH*/
6066 (OP(first) == IFMATCH && !first->flags && (sawlookahead = 1)) ||
6067 (OP(first) == PLUS) ||
6068 (OP(first) == MINMOD) ||
6069 /* An {n,m} with n>0 */
6070 (PL_regkind[OP(first)] == CURLY && ARG1(first) > 0) ||
6071 (OP(first) == NOTHING && PL_regkind[OP(first_next)] != END ))
6074 * the only op that could be a regnode is PLUS, all the rest
6075 * will be regnode_1 or regnode_2.
6078 if (OP(first) == PLUS)
6081 first += regarglen[OP(first)];
6083 first = NEXTOPER(first);
6084 first_next= regnext(first);
6087 /* Starting-point info. */
6089 DEBUG_PEEP("first:",first,0);
6090 /* Ignore EXACT as we deal with it later. */
6091 if (PL_regkind[OP(first)] == EXACT) {
6092 if (OP(first) == EXACT)
6093 NOOP; /* Empty, get anchored substr later. */
6095 ri->regstclass = first;
6098 else if (PL_regkind[OP(first)] == TRIE &&
6099 ((reg_trie_data *)ri->data->data[ ARG(first) ])->minlen>0)
6102 /* this can happen only on restudy */
6103 if ( OP(first) == TRIE ) {
6104 struct regnode_1 *trieop = (struct regnode_1 *)
6105 PerlMemShared_calloc(1, sizeof(struct regnode_1));
6106 StructCopy(first,trieop,struct regnode_1);
6107 trie_op=(regnode *)trieop;
6109 struct regnode_charclass *trieop = (struct regnode_charclass *)
6110 PerlMemShared_calloc(1, sizeof(struct regnode_charclass));
6111 StructCopy(first,trieop,struct regnode_charclass);
6112 trie_op=(regnode *)trieop;
6115 make_trie_failtable(pRExC_state, (regnode *)first, trie_op, 0);
6116 ri->regstclass = trie_op;
6119 else if (REGNODE_SIMPLE(OP(first)))
6120 ri->regstclass = first;
6121 else if (PL_regkind[OP(first)] == BOUND ||
6122 PL_regkind[OP(first)] == NBOUND)
6123 ri->regstclass = first;
6124 else if (PL_regkind[OP(first)] == BOL) {
6125 r->extflags |= (OP(first) == MBOL
6127 : (OP(first) == SBOL
6130 first = NEXTOPER(first);
6133 else if (OP(first) == GPOS) {
6134 r->extflags |= RXf_ANCH_GPOS;
6135 first = NEXTOPER(first);
6138 else if ((!sawopen || !RExC_sawback) &&
6139 (OP(first) == STAR &&
6140 PL_regkind[OP(NEXTOPER(first))] == REG_ANY) &&
6141 !(r->extflags & RXf_ANCH) && !pRExC_state->num_code_blocks)
6143 /* turn .* into ^.* with an implied $*=1 */
6145 (OP(NEXTOPER(first)) == REG_ANY)
6148 r->extflags |= type;
6149 r->intflags |= PREGf_IMPLICIT;
6150 first = NEXTOPER(first);
6153 if (sawplus && !sawlookahead && (!sawopen || !RExC_sawback)
6154 && !pRExC_state->num_code_blocks) /* May examine pos and $& */
6155 /* x+ must match at the 1st pos of run of x's */
6156 r->intflags |= PREGf_SKIP;
6158 /* Scan is after the zeroth branch, first is atomic matcher. */
6159 #ifdef TRIE_STUDY_OPT
6162 PerlIO_printf(Perl_debug_log, "first at %"IVdf"\n",
6163 (IV)(first - scan + 1))
6167 PerlIO_printf(Perl_debug_log, "first at %"IVdf"\n",
6168 (IV)(first - scan + 1))
6174 * If there's something expensive in the r.e., find the
6175 * longest literal string that must appear and make it the
6176 * regmust. Resolve ties in favor of later strings, since
6177 * the regstart check works with the beginning of the r.e.
6178 * and avoiding duplication strengthens checking. Not a
6179 * strong reason, but sufficient in the absence of others.
6180 * [Now we resolve ties in favor of the earlier string if
6181 * it happens that c_offset_min has been invalidated, since the
6182 * earlier string may buy us something the later one won't.]
6185 data.longest_fixed = newSVpvs("");
6186 data.longest_float = newSVpvs("");
6187 data.last_found = newSVpvs("");
6188 data.longest = &(data.longest_fixed);
6189 ENTER_with_name("study_chunk");
6190 SAVEFREESV(data.longest_fixed);
6191 SAVEFREESV(data.longest_float);
6192 SAVEFREESV(data.last_found);
6194 if (!ri->regstclass) {
6195 cl_init(pRExC_state, &ch_class);
6196 data.start_class = &ch_class;
6197 stclass_flag = SCF_DO_STCLASS_AND;
6198 } else /* XXXX Check for BOUND? */
6200 data.last_closep = &last_close;
6202 minlen = study_chunk(pRExC_state, &first, &minlen, &fake, scan + RExC_size, /* Up to end */
6203 &data, -1, NULL, NULL,
6204 SCF_DO_SUBSTR | SCF_WHILEM_VISITED_POS | stclass_flag,0);
6207 CHECK_RESTUDY_GOTO_butfirst(LEAVE_with_name("study_chunk"));
6210 if ( RExC_npar == 1 && data.longest == &(data.longest_fixed)
6211 && data.last_start_min == 0 && data.last_end > 0
6212 && !RExC_seen_zerolen
6213 && !(RExC_seen & REG_SEEN_VERBARG)
6214 && (!(RExC_seen & REG_SEEN_GPOS) || (r->extflags & RXf_ANCH_GPOS)))
6215 r->extflags |= RXf_CHECK_ALL;
6216 scan_commit(pRExC_state, &data,&minlen,0);
6218 longest_float_length = CHR_SVLEN(data.longest_float);
6220 if (! ((SvCUR(data.longest_fixed) /* ok to leave SvCUR */
6221 && data.offset_fixed == data.offset_float_min
6222 && SvCUR(data.longest_fixed) == SvCUR(data.longest_float)))
6223 && S_setup_longest (aTHX_ pRExC_state,
6227 &(r->float_end_shift),
6228 data.lookbehind_float,
6229 data.offset_float_min,
6231 longest_float_length,
6232 cBOOL(data.flags & SF_FL_BEFORE_EOL),
6233 cBOOL(data.flags & SF_FL_BEFORE_MEOL)))
6235 r->float_min_offset = data.offset_float_min - data.lookbehind_float;
6236 r->float_max_offset = data.offset_float_max;
6237 if (data.offset_float_max < I32_MAX) /* Don't offset infinity */
6238 r->float_max_offset -= data.lookbehind_float;
6239 SvREFCNT_inc_simple_void_NN(data.longest_float);
6242 r->float_substr = r->float_utf8 = NULL;
6243 longest_float_length = 0;
6246 longest_fixed_length = CHR_SVLEN(data.longest_fixed);
6248 if (S_setup_longest (aTHX_ pRExC_state,
6250 &(r->anchored_utf8),
6251 &(r->anchored_substr),
6252 &(r->anchored_end_shift),
6253 data.lookbehind_fixed,
6256 longest_fixed_length,
6257 cBOOL(data.flags & SF_FIX_BEFORE_EOL),
6258 cBOOL(data.flags & SF_FIX_BEFORE_MEOL)))
6260 r->anchored_offset = data.offset_fixed - data.lookbehind_fixed;
6261 SvREFCNT_inc_simple_void_NN(data.longest_fixed);
6264 r->anchored_substr = r->anchored_utf8 = NULL;
6265 longest_fixed_length = 0;
6267 LEAVE_with_name("study_chunk");
6270 && (OP(ri->regstclass) == REG_ANY || OP(ri->regstclass) == SANY))
6271 ri->regstclass = NULL;
6273 if ((!(r->anchored_substr || r->anchored_utf8) || r->anchored_offset)
6275 && ! TEST_SSC_EOS(data.start_class)
6276 && !cl_is_anything(data.start_class))
6278 const U32 n = add_data(pRExC_state, 1, "f");
6279 OP(data.start_class) = ANYOF_SYNTHETIC;
6281 Newx(RExC_rxi->data->data[n], 1,
6282 struct regnode_charclass_class);
6283 StructCopy(data.start_class,
6284 (struct regnode_charclass_class*)RExC_rxi->data->data[n],
6285 struct regnode_charclass_class);
6286 ri->regstclass = (regnode*)RExC_rxi->data->data[n];
6287 r->intflags &= ~PREGf_SKIP; /* Used in find_byclass(). */
6288 DEBUG_COMPILE_r({ SV *sv = sv_newmortal();
6289 regprop(r, sv, (regnode*)data.start_class);
6290 PerlIO_printf(Perl_debug_log,
6291 "synthetic stclass \"%s\".\n",
6292 SvPVX_const(sv));});
6295 /* A temporary algorithm prefers floated substr to fixed one to dig more info. */
6296 if (longest_fixed_length > longest_float_length) {
6297 r->check_end_shift = r->anchored_end_shift;
6298 r->check_substr = r->anchored_substr;
6299 r->check_utf8 = r->anchored_utf8;
6300 r->check_offset_min = r->check_offset_max = r->anchored_offset;
6301 if (r->extflags & RXf_ANCH_SINGLE)
6302 r->extflags |= RXf_NOSCAN;
6305 r->check_end_shift = r->float_end_shift;
6306 r->check_substr = r->float_substr;
6307 r->check_utf8 = r->float_utf8;
6308 r->check_offset_min = r->float_min_offset;
6309 r->check_offset_max = r->float_max_offset;
6311 /* XXXX Currently intuiting is not compatible with ANCH_GPOS.
6312 This should be changed ASAP! */
6313 if ((r->check_substr || r->check_utf8) && !(r->extflags & RXf_ANCH_GPOS)) {
6314 r->extflags |= RXf_USE_INTUIT;
6315 if (SvTAIL(r->check_substr ? r->check_substr : r->check_utf8))
6316 r->extflags |= RXf_INTUIT_TAIL;
6318 /* XXX Unneeded? dmq (shouldn't as this is handled elsewhere)
6319 if ( (STRLEN)minlen < longest_float_length )
6320 minlen= longest_float_length;
6321 if ( (STRLEN)minlen < longest_fixed_length )
6322 minlen= longest_fixed_length;
6326 /* Several toplevels. Best we can is to set minlen. */
6328 struct regnode_charclass_class ch_class;
6331 DEBUG_PARSE_r(PerlIO_printf(Perl_debug_log, "\nMulti Top Level\n"));
6333 scan = ri->program + 1;
6334 cl_init(pRExC_state, &ch_class);
6335 data.start_class = &ch_class;
6336 data.last_closep = &last_close;
6339 minlen = study_chunk(pRExC_state, &scan, &minlen, &fake, scan + RExC_size,
6340 &data, -1, NULL, NULL, SCF_DO_STCLASS_AND|SCF_WHILEM_VISITED_POS,0);
6342 CHECK_RESTUDY_GOTO_butfirst(NOOP);
6344 r->check_substr = r->check_utf8 = r->anchored_substr = r->anchored_utf8
6345 = r->float_substr = r->float_utf8 = NULL;
6347 if (! TEST_SSC_EOS(data.start_class)
6348 && !cl_is_anything(data.start_class))
6350 const U32 n = add_data(pRExC_state, 1, "f");
6351 OP(data.start_class) = ANYOF_SYNTHETIC;
6353 Newx(RExC_rxi->data->data[n], 1,
6354 struct regnode_charclass_class);
6355 StructCopy(data.start_class,
6356 (struct regnode_charclass_class*)RExC_rxi->data->data[n],
6357 struct regnode_charclass_class);
6358 ri->regstclass = (regnode*)RExC_rxi->data->data[n];
6359 r->intflags &= ~PREGf_SKIP; /* Used in find_byclass(). */
6360 DEBUG_COMPILE_r({ SV* sv = sv_newmortal();
6361 regprop(r, sv, (regnode*)data.start_class);
6362 PerlIO_printf(Perl_debug_log,
6363 "synthetic stclass \"%s\".\n",
6364 SvPVX_const(sv));});
6368 /* Guard against an embedded (?=) or (?<=) with a longer minlen than
6369 the "real" pattern. */
6371 PerlIO_printf(Perl_debug_log,"minlen: %"IVdf" r->minlen:%"IVdf"\n",
6372 (IV)minlen, (IV)r->minlen);
6374 r->minlenret = minlen;
6375 if (r->minlen < minlen)
6378 if (RExC_seen & REG_SEEN_GPOS)
6379 r->extflags |= RXf_GPOS_SEEN;
6380 if (RExC_seen & REG_SEEN_LOOKBEHIND)
6381 r->extflags |= RXf_NO_INPLACE_SUBST; /* inplace might break the lookbehind */
6382 if (pRExC_state->num_code_blocks)
6383 r->extflags |= RXf_EVAL_SEEN;
6384 if (RExC_seen & REG_SEEN_CANY)
6385 r->extflags |= RXf_CANY_SEEN;
6386 if (RExC_seen & REG_SEEN_VERBARG)
6388 r->intflags |= PREGf_VERBARG_SEEN;
6389 r->extflags |= RXf_NO_INPLACE_SUBST; /* don't understand this! Yves */
6391 if (RExC_seen & REG_SEEN_CUTGROUP)
6392 r->intflags |= PREGf_CUTGROUP_SEEN;
6393 if (pm_flags & PMf_USE_RE_EVAL)
6394 r->intflags |= PREGf_USE_RE_EVAL;
6395 if (RExC_paren_names)
6396 RXp_PAREN_NAMES(r) = MUTABLE_HV(SvREFCNT_inc(RExC_paren_names));
6398 RXp_PAREN_NAMES(r) = NULL;
6401 regnode *first = ri->program + 1;
6403 regnode *next = NEXTOPER(first);
6406 if (PL_regkind[fop] == NOTHING && nop == END)
6407 r->extflags |= RXf_NULL;
6408 else if (PL_regkind[fop] == BOL && nop == END)
6409 r->extflags |= RXf_START_ONLY;
6410 else if (fop == PLUS && PL_regkind[nop] == POSIXD && FLAGS(next) == _CC_SPACE && OP(regnext(first)) == END)
6411 r->extflags |= RXf_WHITE;
6412 else if ( r->extflags & RXf_SPLIT && fop == EXACT && STR_LEN(first) == 1 && *(STRING(first)) == ' ' && OP(regnext(first)) == END )
6413 r->extflags |= (RXf_SKIPWHITE|RXf_WHITE);
6417 if (RExC_paren_names) {
6418 ri->name_list_idx = add_data( pRExC_state, 1, "a" );
6419 ri->data->data[ri->name_list_idx] = (void*)SvREFCNT_inc(RExC_paren_name_list);
6422 ri->name_list_idx = 0;
6424 if (RExC_recurse_count) {
6425 for ( ; RExC_recurse_count ; RExC_recurse_count-- ) {
6426 const regnode *scan = RExC_recurse[RExC_recurse_count-1];
6427 ARG2L_SET( scan, RExC_open_parens[ARG(scan)-1] - scan );
6430 Newxz(r->offs, RExC_npar, regexp_paren_pair);
6431 /* assume we don't need to swap parens around before we match */
6434 PerlIO_printf(Perl_debug_log,"Final program:\n");
6437 #ifdef RE_TRACK_PATTERN_OFFSETS
6438 DEBUG_OFFSETS_r(if (ri->u.offsets) {
6439 const U32 len = ri->u.offsets[0];
6441 GET_RE_DEBUG_FLAGS_DECL;
6442 PerlIO_printf(Perl_debug_log, "Offsets: [%"UVuf"]\n\t", (UV)ri->u.offsets[0]);
6443 for (i = 1; i <= len; i++) {
6444 if (ri->u.offsets[i*2-1] || ri->u.offsets[i*2])
6445 PerlIO_printf(Perl_debug_log, "%"UVuf":%"UVuf"[%"UVuf"] ",
6446 (UV)i, (UV)ri->u.offsets[i*2-1], (UV)ri->u.offsets[i*2]);
6448 PerlIO_printf(Perl_debug_log, "\n");
6453 /* under ithreads the ?pat? PMf_USED flag on the pmop is simulated
6454 * by setting the regexp SV to readonly-only instead. If the
6455 * pattern's been recompiled, the USEDness should remain. */
6456 if (old_re && SvREADONLY(old_re))
6464 Perl_reg_named_buff(pTHX_ REGEXP * const rx, SV * const key, SV * const value,
6467 PERL_ARGS_ASSERT_REG_NAMED_BUFF;
6469 PERL_UNUSED_ARG(value);
6471 if (flags & RXapif_FETCH) {
6472 return reg_named_buff_fetch(rx, key, flags);
6473 } else if (flags & (RXapif_STORE | RXapif_DELETE | RXapif_CLEAR)) {
6474 Perl_croak_no_modify();
6476 } else if (flags & RXapif_EXISTS) {
6477 return reg_named_buff_exists(rx, key, flags)
6480 } else if (flags & RXapif_REGNAMES) {
6481 return reg_named_buff_all(rx, flags);
6482 } else if (flags & (RXapif_SCALAR | RXapif_REGNAMES_COUNT)) {
6483 return reg_named_buff_scalar(rx, flags);
6485 Perl_croak(aTHX_ "panic: Unknown flags %d in named_buff", (int)flags);
6491 Perl_reg_named_buff_iter(pTHX_ REGEXP * const rx, const SV * const lastkey,
6494 PERL_ARGS_ASSERT_REG_NAMED_BUFF_ITER;
6495 PERL_UNUSED_ARG(lastkey);
6497 if (flags & RXapif_FIRSTKEY)
6498 return reg_named_buff_firstkey(rx, flags);
6499 else if (flags & RXapif_NEXTKEY)
6500 return reg_named_buff_nextkey(rx, flags);
6502 Perl_croak(aTHX_ "panic: Unknown flags %d in named_buff_iter", (int)flags);
6508 Perl_reg_named_buff_fetch(pTHX_ REGEXP * const r, SV * const namesv,
6511 AV *retarray = NULL;
6513 struct regexp *const rx = ReANY(r);
6515 PERL_ARGS_ASSERT_REG_NAMED_BUFF_FETCH;
6517 if (flags & RXapif_ALL)
6520 if (rx && RXp_PAREN_NAMES(rx)) {
6521 HE *he_str = hv_fetch_ent( RXp_PAREN_NAMES(rx), namesv, 0, 0 );
6524 SV* sv_dat=HeVAL(he_str);
6525 I32 *nums=(I32*)SvPVX(sv_dat);
6526 for ( i=0; i<SvIVX(sv_dat); i++ ) {
6527 if ((I32)(rx->nparens) >= nums[i]
6528 && rx->offs[nums[i]].start != -1
6529 && rx->offs[nums[i]].end != -1)
6532 CALLREG_NUMBUF_FETCH(r,nums[i],ret);
6537 ret = newSVsv(&PL_sv_undef);
6540 av_push(retarray, ret);
6543 return newRV_noinc(MUTABLE_SV(retarray));
6550 Perl_reg_named_buff_exists(pTHX_ REGEXP * const r, SV * const key,
6553 struct regexp *const rx = ReANY(r);
6555 PERL_ARGS_ASSERT_REG_NAMED_BUFF_EXISTS;
6557 if (rx && RXp_PAREN_NAMES(rx)) {
6558 if (flags & RXapif_ALL) {
6559 return hv_exists_ent(RXp_PAREN_NAMES(rx), key, 0);
6561 SV *sv = CALLREG_NAMED_BUFF_FETCH(r, key, flags);
6563 SvREFCNT_dec_NN(sv);
6575 Perl_reg_named_buff_firstkey(pTHX_ REGEXP * const r, const U32 flags)
6577 struct regexp *const rx = ReANY(r);
6579 PERL_ARGS_ASSERT_REG_NAMED_BUFF_FIRSTKEY;
6581 if ( rx && RXp_PAREN_NAMES(rx) ) {
6582 (void)hv_iterinit(RXp_PAREN_NAMES(rx));
6584 return CALLREG_NAMED_BUFF_NEXTKEY(r, NULL, flags & ~RXapif_FIRSTKEY);
6591 Perl_reg_named_buff_nextkey(pTHX_ REGEXP * const r, const U32 flags)
6593 struct regexp *const rx = ReANY(r);
6594 GET_RE_DEBUG_FLAGS_DECL;
6596 PERL_ARGS_ASSERT_REG_NAMED_BUFF_NEXTKEY;
6598 if (rx && RXp_PAREN_NAMES(rx)) {
6599 HV *hv = RXp_PAREN_NAMES(rx);
6601 while ( (temphe = hv_iternext_flags(hv,0)) ) {
6604 SV* sv_dat = HeVAL(temphe);
6605 I32 *nums = (I32*)SvPVX(sv_dat);
6606 for ( i = 0; i < SvIVX(sv_dat); i++ ) {
6607 if ((I32)(rx->lastparen) >= nums[i] &&
6608 rx->offs[nums[i]].start != -1 &&
6609 rx->offs[nums[i]].end != -1)
6615 if (parno || flags & RXapif_ALL) {
6616 return newSVhek(HeKEY_hek(temphe));
6624 Perl_reg_named_buff_scalar(pTHX_ REGEXP * const r, const U32 flags)
6629 struct regexp *const rx = ReANY(r);
6631 PERL_ARGS_ASSERT_REG_NAMED_BUFF_SCALAR;
6633 if (rx && RXp_PAREN_NAMES(rx)) {
6634 if (flags & (RXapif_ALL | RXapif_REGNAMES_COUNT)) {
6635 return newSViv(HvTOTALKEYS(RXp_PAREN_NAMES(rx)));
6636 } else if (flags & RXapif_ONE) {
6637 ret = CALLREG_NAMED_BUFF_ALL(r, (flags | RXapif_REGNAMES));
6638 av = MUTABLE_AV(SvRV(ret));
6639 length = av_len(av);
6640 SvREFCNT_dec_NN(ret);
6641 return newSViv(length + 1);
6643 Perl_croak(aTHX_ "panic: Unknown flags %d in named_buff_scalar", (int)flags);
6647 return &PL_sv_undef;
6651 Perl_reg_named_buff_all(pTHX_ REGEXP * const r, const U32 flags)
6653 struct regexp *const rx = ReANY(r);
6656 PERL_ARGS_ASSERT_REG_NAMED_BUFF_ALL;
6658 if (rx && RXp_PAREN_NAMES(rx)) {
6659 HV *hv= RXp_PAREN_NAMES(rx);
6661 (void)hv_iterinit(hv);
6662 while ( (temphe = hv_iternext_flags(hv,0)) ) {
6665 SV* sv_dat = HeVAL(temphe);
6666 I32 *nums = (I32*)SvPVX(sv_dat);
6667 for ( i = 0; i < SvIVX(sv_dat); i++ ) {
6668 if ((I32)(rx->lastparen) >= nums[i] &&
6669 rx->offs[nums[i]].start != -1 &&
6670 rx->offs[nums[i]].end != -1)
6676 if (parno || flags & RXapif_ALL) {
6677 av_push(av, newSVhek(HeKEY_hek(temphe)));
6682 return newRV_noinc(MUTABLE_SV(av));
6686 Perl_reg_numbered_buff_fetch(pTHX_ REGEXP * const r, const I32 paren,
6689 struct regexp *const rx = ReANY(r);
6695 PERL_ARGS_ASSERT_REG_NUMBERED_BUFF_FETCH;
6697 if ( ( n == RX_BUFF_IDX_CARET_PREMATCH
6698 || n == RX_BUFF_IDX_CARET_FULLMATCH
6699 || n == RX_BUFF_IDX_CARET_POSTMATCH
6701 && !(rx->extflags & RXf_PMf_KEEPCOPY)
6708 if (n == RX_BUFF_IDX_CARET_FULLMATCH)
6709 /* no need to distinguish between them any more */
6710 n = RX_BUFF_IDX_FULLMATCH;
6712 if ((n == RX_BUFF_IDX_PREMATCH || n == RX_BUFF_IDX_CARET_PREMATCH)
6713 && rx->offs[0].start != -1)
6715 /* $`, ${^PREMATCH} */
6716 i = rx->offs[0].start;
6720 if ((n == RX_BUFF_IDX_POSTMATCH || n == RX_BUFF_IDX_CARET_POSTMATCH)
6721 && rx->offs[0].end != -1)
6723 /* $', ${^POSTMATCH} */
6724 s = rx->subbeg - rx->suboffset + rx->offs[0].end;
6725 i = rx->sublen + rx->suboffset - rx->offs[0].end;
6728 if ( 0 <= n && n <= (I32)rx->nparens &&
6729 (s1 = rx->offs[n].start) != -1 &&
6730 (t1 = rx->offs[n].end) != -1)
6732 /* $&, ${^MATCH}, $1 ... */
6734 s = rx->subbeg + s1 - rx->suboffset;
6739 assert(s >= rx->subbeg);
6740 assert(rx->sublen >= (s - rx->subbeg) + i );
6742 #if NO_TAINT_SUPPORT
6743 sv_setpvn(sv, s, i);
6745 const int oldtainted = TAINT_get;
6747 sv_setpvn(sv, s, i);
6748 TAINT_set(oldtainted);
6750 if ( (rx->extflags & RXf_CANY_SEEN)
6751 ? (RXp_MATCH_UTF8(rx)
6752 && (!i || is_utf8_string((U8*)s, i)))
6753 : (RXp_MATCH_UTF8(rx)) )
6760 if (RXp_MATCH_TAINTED(rx)) {
6761 if (SvTYPE(sv) >= SVt_PVMG) {
6762 MAGIC* const mg = SvMAGIC(sv);
6765 SvMAGIC_set(sv, mg->mg_moremagic);
6767 if ((mgt = SvMAGIC(sv))) {
6768 mg->mg_moremagic = mgt;
6769 SvMAGIC_set(sv, mg);
6780 sv_setsv(sv,&PL_sv_undef);
6786 Perl_reg_numbered_buff_store(pTHX_ REGEXP * const rx, const I32 paren,
6787 SV const * const value)
6789 PERL_ARGS_ASSERT_REG_NUMBERED_BUFF_STORE;
6791 PERL_UNUSED_ARG(rx);
6792 PERL_UNUSED_ARG(paren);
6793 PERL_UNUSED_ARG(value);
6796 Perl_croak_no_modify();
6800 Perl_reg_numbered_buff_length(pTHX_ REGEXP * const r, const SV * const sv,
6803 struct regexp *const rx = ReANY(r);
6807 PERL_ARGS_ASSERT_REG_NUMBERED_BUFF_LENGTH;
6809 /* Some of this code was originally in C<Perl_magic_len> in F<mg.c> */
6811 case RX_BUFF_IDX_CARET_PREMATCH: /* ${^PREMATCH} */
6812 if (!(rx->extflags & RXf_PMf_KEEPCOPY))
6816 case RX_BUFF_IDX_PREMATCH: /* $` */
6817 if (rx->offs[0].start != -1) {
6818 i = rx->offs[0].start;
6827 case RX_BUFF_IDX_CARET_POSTMATCH: /* ${^POSTMATCH} */
6828 if (!(rx->extflags & RXf_PMf_KEEPCOPY))
6830 case RX_BUFF_IDX_POSTMATCH: /* $' */
6831 if (rx->offs[0].end != -1) {
6832 i = rx->sublen - rx->offs[0].end;
6834 s1 = rx->offs[0].end;
6841 case RX_BUFF_IDX_CARET_FULLMATCH: /* ${^MATCH} */
6842 if (!(rx->extflags & RXf_PMf_KEEPCOPY))
6846 /* $& / ${^MATCH}, $1, $2, ... */
6848 if (paren <= (I32)rx->nparens &&
6849 (s1 = rx->offs[paren].start) != -1 &&
6850 (t1 = rx->offs[paren].end) != -1)
6856 if (ckWARN(WARN_UNINITIALIZED))
6857 report_uninit((const SV *)sv);
6862 if (i > 0 && RXp_MATCH_UTF8(rx)) {
6863 const char * const s = rx->subbeg - rx->suboffset + s1;
6868 if (is_utf8_string_loclen((U8*)s, i, &ep, &el))
6875 Perl_reg_qr_package(pTHX_ REGEXP * const rx)
6877 PERL_ARGS_ASSERT_REG_QR_PACKAGE;
6878 PERL_UNUSED_ARG(rx);
6882 return newSVpvs("Regexp");
6885 /* Scans the name of a named buffer from the pattern.
6886 * If flags is REG_RSN_RETURN_NULL returns null.
6887 * If flags is REG_RSN_RETURN_NAME returns an SV* containing the name
6888 * If flags is REG_RSN_RETURN_DATA returns the data SV* corresponding
6889 * to the parsed name as looked up in the RExC_paren_names hash.
6890 * If there is an error throws a vFAIL().. type exception.
6893 #define REG_RSN_RETURN_NULL 0
6894 #define REG_RSN_RETURN_NAME 1
6895 #define REG_RSN_RETURN_DATA 2
6898 S_reg_scan_name(pTHX_ RExC_state_t *pRExC_state, U32 flags)
6900 char *name_start = RExC_parse;
6902 PERL_ARGS_ASSERT_REG_SCAN_NAME;
6904 if (isIDFIRST_lazy_if(RExC_parse, UTF)) {
6905 /* skip IDFIRST by using do...while */
6908 RExC_parse += UTF8SKIP(RExC_parse);
6909 } while (isWORDCHAR_utf8((U8*)RExC_parse));
6913 } while (isWORDCHAR(*RExC_parse));
6915 RExC_parse++; /* so the <- from the vFAIL is after the offending character */
6916 vFAIL("Group name must start with a non-digit word character");
6920 = newSVpvn_flags(name_start, (int)(RExC_parse - name_start),
6921 SVs_TEMP | (UTF ? SVf_UTF8 : 0));
6922 if ( flags == REG_RSN_RETURN_NAME)
6924 else if (flags==REG_RSN_RETURN_DATA) {
6927 if ( ! sv_name ) /* should not happen*/
6928 Perl_croak(aTHX_ "panic: no svname in reg_scan_name");
6929 if (RExC_paren_names)
6930 he_str = hv_fetch_ent( RExC_paren_names, sv_name, 0, 0 );
6932 sv_dat = HeVAL(he_str);
6934 vFAIL("Reference to nonexistent named group");
6938 Perl_croak(aTHX_ "panic: bad flag %lx in reg_scan_name",
6939 (unsigned long) flags);
6941 assert(0); /* NOT REACHED */
6946 #define DEBUG_PARSE_MSG(funcname) DEBUG_PARSE_r({ \
6947 int rem=(int)(RExC_end - RExC_parse); \
6956 if (RExC_lastparse!=RExC_parse) \
6957 PerlIO_printf(Perl_debug_log," >%.*s%-*s", \
6960 iscut ? "..." : "<" \
6963 PerlIO_printf(Perl_debug_log,"%16s",""); \
6966 num = RExC_size + 1; \
6968 num=REG_NODE_NUM(RExC_emit); \
6969 if (RExC_lastnum!=num) \
6970 PerlIO_printf(Perl_debug_log,"|%4d",num); \
6972 PerlIO_printf(Perl_debug_log,"|%4s",""); \
6973 PerlIO_printf(Perl_debug_log,"|%*s%-4s", \
6974 (int)((depth*2)), "", \
6978 RExC_lastparse=RExC_parse; \
6983 #define DEBUG_PARSE(funcname) DEBUG_PARSE_r({ \
6984 DEBUG_PARSE_MSG((funcname)); \
6985 PerlIO_printf(Perl_debug_log,"%4s","\n"); \
6987 #define DEBUG_PARSE_FMT(funcname,fmt,args) DEBUG_PARSE_r({ \
6988 DEBUG_PARSE_MSG((funcname)); \
6989 PerlIO_printf(Perl_debug_log,fmt "\n",args); \
6992 /* This section of code defines the inversion list object and its methods. The
6993 * interfaces are highly subject to change, so as much as possible is static to
6994 * this file. An inversion list is here implemented as a malloc'd C UV array
6995 * with some added info that is placed as UVs at the beginning in a header
6996 * portion. An inversion list for Unicode is an array of code points, sorted
6997 * by ordinal number. The zeroth element is the first code point in the list.
6998 * The 1th element is the first element beyond that not in the list. In other
6999 * words, the first range is
7000 * invlist[0]..(invlist[1]-1)
7001 * The other ranges follow. Thus every element whose index is divisible by two
7002 * marks the beginning of a range that is in the list, and every element not
7003 * divisible by two marks the beginning of a range not in the list. A single
7004 * element inversion list that contains the single code point N generally
7005 * consists of two elements
7008 * (The exception is when N is the highest representable value on the
7009 * machine, in which case the list containing just it would be a single
7010 * element, itself. By extension, if the last range in the list extends to
7011 * infinity, then the first element of that range will be in the inversion list
7012 * at a position that is divisible by two, and is the final element in the
7014 * Taking the complement (inverting) an inversion list is quite simple, if the
7015 * first element is 0, remove it; otherwise add a 0 element at the beginning.
7016 * This implementation reserves an element at the beginning of each inversion
7017 * list to contain 0 when the list contains 0, and contains 1 otherwise. The
7018 * actual beginning of the list is either that element if 0, or the next one if
7021 * More about inversion lists can be found in "Unicode Demystified"
7022 * Chapter 13 by Richard Gillam, published by Addison-Wesley.
7023 * More will be coming when functionality is added later.
7025 * The inversion list data structure is currently implemented as an SV pointing
7026 * to an array of UVs that the SV thinks are bytes. This allows us to have an
7027 * array of UV whose memory management is automatically handled by the existing
7028 * facilities for SV's.
7030 * Some of the methods should always be private to the implementation, and some
7031 * should eventually be made public */
7033 /* The header definitions are in F<inline_invlist.c> */
7034 #define TO_INTERNAL_SIZE(x) (((x) + HEADER_LENGTH) * sizeof(UV))
7035 #define FROM_INTERNAL_SIZE(x) (((x)/ sizeof(UV)) - HEADER_LENGTH)
7037 #define INVLIST_INITIAL_LEN 10
7039 PERL_STATIC_INLINE UV*
7040 S__invlist_array_init(pTHX_ SV* const invlist, const bool will_have_0)
7042 /* Returns a pointer to the first element in the inversion list's array.
7043 * This is called upon initialization of an inversion list. Where the
7044 * array begins depends on whether the list has the code point U+0000
7045 * in it or not. The other parameter tells it whether the code that
7046 * follows this call is about to put a 0 in the inversion list or not.
7047 * The first element is either the element with 0, if 0, or the next one,
7050 UV* zero = get_invlist_zero_addr(invlist);
7052 PERL_ARGS_ASSERT__INVLIST_ARRAY_INIT;
7055 assert(! *_get_invlist_len_addr(invlist));
7057 /* 1^1 = 0; 1^0 = 1 */
7058 *zero = 1 ^ will_have_0;
7059 return zero + *zero;
7062 PERL_STATIC_INLINE UV*
7063 S_invlist_array(pTHX_ SV* const invlist)
7065 /* Returns the pointer to the inversion list's array. Every time the
7066 * length changes, this needs to be called in case malloc or realloc moved
7069 PERL_ARGS_ASSERT_INVLIST_ARRAY;
7071 /* Must not be empty. If these fail, you probably didn't check for <len>
7072 * being non-zero before trying to get the array */
7073 assert(*_get_invlist_len_addr(invlist));
7074 assert(*get_invlist_zero_addr(invlist) == 0
7075 || *get_invlist_zero_addr(invlist) == 1);
7077 /* The array begins either at the element reserved for zero if the
7078 * list contains 0 (that element will be set to 0), or otherwise the next
7079 * element (in which case the reserved element will be set to 1). */
7080 return (UV *) (get_invlist_zero_addr(invlist)
7081 + *get_invlist_zero_addr(invlist));
7084 PERL_STATIC_INLINE void
7085 S_invlist_set_len(pTHX_ SV* const invlist, const UV len)
7087 /* Sets the current number of elements stored in the inversion list */
7089 PERL_ARGS_ASSERT_INVLIST_SET_LEN;
7091 *_get_invlist_len_addr(invlist) = len;
7093 assert(len <= SvLEN(invlist));
7095 SvCUR_set(invlist, TO_INTERNAL_SIZE(len));
7096 /* If the list contains U+0000, that element is part of the header,
7097 * and should not be counted as part of the array. It will contain
7098 * 0 in that case, and 1 otherwise. So we could flop 0=>1, 1=>0 and
7100 * SvCUR_set(invlist,
7101 * TO_INTERNAL_SIZE(len
7102 * - (*get_invlist_zero_addr(inv_list) ^ 1)));
7103 * But, this is only valid if len is not 0. The consequences of not doing
7104 * this is that the memory allocation code may think that 1 more UV is
7105 * being used than actually is, and so might do an unnecessary grow. That
7106 * seems worth not bothering to make this the precise amount.
7108 * Note that when inverting, SvCUR shouldn't change */
7111 PERL_STATIC_INLINE IV*
7112 S_get_invlist_previous_index_addr(pTHX_ SV* invlist)
7114 /* Return the address of the UV that is reserved to hold the cached index
7117 PERL_ARGS_ASSERT_GET_INVLIST_PREVIOUS_INDEX_ADDR;
7119 return (IV *) (SvPVX(invlist) + (INVLIST_PREVIOUS_INDEX_OFFSET * sizeof (UV)));
7122 PERL_STATIC_INLINE IV
7123 S_invlist_previous_index(pTHX_ SV* const invlist)
7125 /* Returns cached index of previous search */
7127 PERL_ARGS_ASSERT_INVLIST_PREVIOUS_INDEX;
7129 return *get_invlist_previous_index_addr(invlist);
7132 PERL_STATIC_INLINE void
7133 S_invlist_set_previous_index(pTHX_ SV* const invlist, const IV index)
7135 /* Caches <index> for later retrieval */
7137 PERL_ARGS_ASSERT_INVLIST_SET_PREVIOUS_INDEX;
7139 assert(index == 0 || index < (int) _invlist_len(invlist));
7141 *get_invlist_previous_index_addr(invlist) = index;
7144 PERL_STATIC_INLINE UV
7145 S_invlist_max(pTHX_ SV* const invlist)
7147 /* Returns the maximum number of elements storable in the inversion list's
7148 * array, without having to realloc() */
7150 PERL_ARGS_ASSERT_INVLIST_MAX;
7152 return SvLEN(invlist) == 0 /* This happens under _new_invlist_C_array */
7153 ? _invlist_len(invlist)
7154 : FROM_INTERNAL_SIZE(SvLEN(invlist));
7157 PERL_STATIC_INLINE UV*
7158 S_get_invlist_zero_addr(pTHX_ SV* invlist)
7160 /* Return the address of the UV that is reserved to hold 0 if the inversion
7161 * list contains 0. This has to be the last element of the heading, as the
7162 * list proper starts with either it if 0, or the next element if not.
7163 * (But we force it to contain either 0 or 1) */
7165 PERL_ARGS_ASSERT_GET_INVLIST_ZERO_ADDR;
7167 return (UV *) (SvPVX(invlist) + (INVLIST_ZERO_OFFSET * sizeof (UV)));
7170 #ifndef PERL_IN_XSUB_RE
7172 Perl__new_invlist(pTHX_ IV initial_size)
7175 /* Return a pointer to a newly constructed inversion list, with enough
7176 * space to store 'initial_size' elements. If that number is negative, a
7177 * system default is used instead */
7181 if (initial_size < 0) {
7182 initial_size = INVLIST_INITIAL_LEN;
7185 /* Allocate the initial space */
7186 new_list = newSV(TO_INTERNAL_SIZE(initial_size));
7187 invlist_set_len(new_list, 0);
7189 /* Force iterinit() to be used to get iteration to work */
7190 *get_invlist_iter_addr(new_list) = UV_MAX;
7192 /* This should force a segfault if a method doesn't initialize this
7194 *get_invlist_zero_addr(new_list) = UV_MAX;
7196 *get_invlist_previous_index_addr(new_list) = 0;
7197 *get_invlist_version_id_addr(new_list) = INVLIST_VERSION_ID;
7198 #if HEADER_LENGTH != 5
7199 # error Need to regenerate INVLIST_VERSION_ID by running perl -E 'say int(rand 2**31-1)', and then changing the #if to the new length
7207 S__new_invlist_C_array(pTHX_ UV* list)
7209 /* Return a pointer to a newly constructed inversion list, initialized to
7210 * point to <list>, which has to be in the exact correct inversion list
7211 * form, including internal fields. Thus this is a dangerous routine that
7212 * should not be used in the wrong hands */
7214 SV* invlist = newSV_type(SVt_PV);
7216 PERL_ARGS_ASSERT__NEW_INVLIST_C_ARRAY;
7218 SvPV_set(invlist, (char *) list);
7219 SvLEN_set(invlist, 0); /* Means we own the contents, and the system
7220 shouldn't touch it */
7221 SvCUR_set(invlist, TO_INTERNAL_SIZE(_invlist_len(invlist)));
7223 if (*get_invlist_version_id_addr(invlist) != INVLIST_VERSION_ID) {
7224 Perl_croak(aTHX_ "panic: Incorrect version for previously generated inversion list");
7227 /* Initialize the iteration pointer.
7228 * XXX This could be done at compile time in charclass_invlists.h, but I
7229 * (khw) am not confident that the suffixes for specifying the C constant
7230 * UV_MAX are portable, e.g. 'ull' on a 32 bit machine that is configured
7231 * to use 64 bits; might need a Configure probe */
7232 invlist_iterfinish(invlist);
7238 S_invlist_extend(pTHX_ SV* const invlist, const UV new_max)
7240 /* Grow the maximum size of an inversion list */
7242 PERL_ARGS_ASSERT_INVLIST_EXTEND;
7244 SvGROW((SV *)invlist, TO_INTERNAL_SIZE(new_max));
7247 PERL_STATIC_INLINE void
7248 S_invlist_trim(pTHX_ SV* const invlist)
7250 PERL_ARGS_ASSERT_INVLIST_TRIM;
7252 /* Change the length of the inversion list to how many entries it currently
7255 SvPV_shrink_to_cur((SV *) invlist);
7258 #define _invlist_union_complement_2nd(a, b, output) _invlist_union_maybe_complement_2nd(a, b, TRUE, output)
7261 S__append_range_to_invlist(pTHX_ SV* const invlist, const UV start, const UV end)
7263 /* Subject to change or removal. Append the range from 'start' to 'end' at
7264 * the end of the inversion list. The range must be above any existing
7268 UV max = invlist_max(invlist);
7269 UV len = _invlist_len(invlist);
7271 PERL_ARGS_ASSERT__APPEND_RANGE_TO_INVLIST;
7273 if (len == 0) { /* Empty lists must be initialized */
7274 array = _invlist_array_init(invlist, start == 0);
7277 /* Here, the existing list is non-empty. The current max entry in the
7278 * list is generally the first value not in the set, except when the
7279 * set extends to the end of permissible values, in which case it is
7280 * the first entry in that final set, and so this call is an attempt to
7281 * append out-of-order */
7283 UV final_element = len - 1;
7284 array = invlist_array(invlist);
7285 if (array[final_element] > start
7286 || ELEMENT_RANGE_MATCHES_INVLIST(final_element))
7288 Perl_croak(aTHX_ "panic: attempting to append to an inversion list, but wasn't at the end of the list, final=%"UVuf", start=%"UVuf", match=%c",
7289 array[final_element], start,
7290 ELEMENT_RANGE_MATCHES_INVLIST(final_element) ? 't' : 'f');
7293 /* Here, it is a legal append. If the new range begins with the first
7294 * value not in the set, it is extending the set, so the new first
7295 * value not in the set is one greater than the newly extended range.
7297 if (array[final_element] == start) {
7298 if (end != UV_MAX) {
7299 array[final_element] = end + 1;
7302 /* But if the end is the maximum representable on the machine,
7303 * just let the range that this would extend to have no end */
7304 invlist_set_len(invlist, len - 1);
7310 /* Here the new range doesn't extend any existing set. Add it */
7312 len += 2; /* Includes an element each for the start and end of range */
7314 /* If overflows the existing space, extend, which may cause the array to be
7317 invlist_extend(invlist, len);
7318 invlist_set_len(invlist, len); /* Have to set len here to avoid assert
7319 failure in invlist_array() */
7320 array = invlist_array(invlist);
7323 invlist_set_len(invlist, len);
7326 /* The next item on the list starts the range, the one after that is
7327 * one past the new range. */
7328 array[len - 2] = start;
7329 if (end != UV_MAX) {
7330 array[len - 1] = end + 1;
7333 /* But if the end is the maximum representable on the machine, just let
7334 * the range have no end */
7335 invlist_set_len(invlist, len - 1);
7339 #ifndef PERL_IN_XSUB_RE
7342 Perl__invlist_search(pTHX_ SV* const invlist, const UV cp)
7344 /* Searches the inversion list for the entry that contains the input code
7345 * point <cp>. If <cp> is not in the list, -1 is returned. Otherwise, the
7346 * return value is the index into the list's array of the range that
7351 IV high = _invlist_len(invlist);
7352 const IV highest_element = high - 1;
7355 PERL_ARGS_ASSERT__INVLIST_SEARCH;
7357 /* If list is empty, return failure. */
7362 /* (We can't get the array unless we know the list is non-empty) */
7363 array = invlist_array(invlist);
7365 mid = invlist_previous_index(invlist);
7366 assert(mid >=0 && mid <= highest_element);
7368 /* <mid> contains the cache of the result of the previous call to this
7369 * function (0 the first time). See if this call is for the same result,
7370 * or if it is for mid-1. This is under the theory that calls to this
7371 * function will often be for related code points that are near each other.
7372 * And benchmarks show that caching gives better results. We also test
7373 * here if the code point is within the bounds of the list. These tests
7374 * replace others that would have had to be made anyway to make sure that
7375 * the array bounds were not exceeded, and these give us extra information
7376 * at the same time */
7377 if (cp >= array[mid]) {
7378 if (cp >= array[highest_element]) {
7379 return highest_element;
7382 /* Here, array[mid] <= cp < array[highest_element]. This means that
7383 * the final element is not the answer, so can exclude it; it also
7384 * means that <mid> is not the final element, so can refer to 'mid + 1'
7386 if (cp < array[mid + 1]) {
7392 else { /* cp < aray[mid] */
7393 if (cp < array[0]) { /* Fail if outside the array */
7397 if (cp >= array[mid - 1]) {
7402 /* Binary search. What we are looking for is <i> such that
7403 * array[i] <= cp < array[i+1]
7404 * The loop below converges on the i+1. Note that there may not be an
7405 * (i+1)th element in the array, and things work nonetheless */
7406 while (low < high) {
7407 mid = (low + high) / 2;
7408 assert(mid <= highest_element);
7409 if (array[mid] <= cp) { /* cp >= array[mid] */
7412 /* We could do this extra test to exit the loop early.
7413 if (cp < array[low]) {
7418 else { /* cp < array[mid] */
7425 invlist_set_previous_index(invlist, high);
7430 Perl__invlist_populate_swatch(pTHX_ SV* const invlist, const UV start, const UV end, U8* swatch)
7432 /* populates a swatch of a swash the same way swatch_get() does in utf8.c,
7433 * but is used when the swash has an inversion list. This makes this much
7434 * faster, as it uses a binary search instead of a linear one. This is
7435 * intimately tied to that function, and perhaps should be in utf8.c,
7436 * except it is intimately tied to inversion lists as well. It assumes
7437 * that <swatch> is all 0's on input */
7440 const IV len = _invlist_len(invlist);
7444 PERL_ARGS_ASSERT__INVLIST_POPULATE_SWATCH;
7446 if (len == 0) { /* Empty inversion list */
7450 array = invlist_array(invlist);
7452 /* Find which element it is */
7453 i = _invlist_search(invlist, start);
7455 /* We populate from <start> to <end> */
7456 while (current < end) {
7459 /* The inversion list gives the results for every possible code point
7460 * after the first one in the list. Only those ranges whose index is
7461 * even are ones that the inversion list matches. For the odd ones,
7462 * and if the initial code point is not in the list, we have to skip
7463 * forward to the next element */
7464 if (i == -1 || ! ELEMENT_RANGE_MATCHES_INVLIST(i)) {
7466 if (i >= len) { /* Finished if beyond the end of the array */
7470 if (current >= end) { /* Finished if beyond the end of what we
7472 if (LIKELY(end < UV_MAX)) {
7476 /* We get here when the upper bound is the maximum
7477 * representable on the machine, and we are looking for just
7478 * that code point. Have to special case it */
7480 goto join_end_of_list;
7483 assert(current >= start);
7485 /* The current range ends one below the next one, except don't go past
7488 upper = (i < len && array[i] < end) ? array[i] : end;
7490 /* Here we are in a range that matches. Populate a bit in the 3-bit U8
7491 * for each code point in it */
7492 for (; current < upper; current++) {
7493 const STRLEN offset = (STRLEN)(current - start);
7494 swatch[offset >> 3] |= 1 << (offset & 7);
7499 /* Quit if at the end of the list */
7502 /* But first, have to deal with the highest possible code point on
7503 * the platform. The previous code assumes that <end> is one
7504 * beyond where we want to populate, but that is impossible at the
7505 * platform's infinity, so have to handle it specially */
7506 if (UNLIKELY(end == UV_MAX && ELEMENT_RANGE_MATCHES_INVLIST(len-1)))
7508 const STRLEN offset = (STRLEN)(end - start);
7509 swatch[offset >> 3] |= 1 << (offset & 7);
7514 /* Advance to the next range, which will be for code points not in the
7523 Perl__invlist_union_maybe_complement_2nd(pTHX_ SV* const a, SV* const b, bool complement_b, SV** output)
7525 /* Take the union of two inversion lists and point <output> to it. *output
7526 * SHOULD BE DEFINED upon input, and if it points to one of the two lists,
7527 * the reference count to that list will be decremented. The first list,
7528 * <a>, may be NULL, in which case a copy of the second list is returned.
7529 * If <complement_b> is TRUE, the union is taken of the complement
7530 * (inversion) of <b> instead of b itself.
7532 * The basis for this comes from "Unicode Demystified" Chapter 13 by
7533 * Richard Gillam, published by Addison-Wesley, and explained at some
7534 * length there. The preface says to incorporate its examples into your
7535 * code at your own risk.
7537 * The algorithm is like a merge sort.
7539 * XXX A potential performance improvement is to keep track as we go along
7540 * if only one of the inputs contributes to the result, meaning the other
7541 * is a subset of that one. In that case, we can skip the final copy and
7542 * return the larger of the input lists, but then outside code might need
7543 * to keep track of whether to free the input list or not */
7545 UV* array_a; /* a's array */
7547 UV len_a; /* length of a's array */
7550 SV* u; /* the resulting union */
7554 UV i_a = 0; /* current index into a's array */
7558 /* running count, as explained in the algorithm source book; items are
7559 * stopped accumulating and are output when the count changes to/from 0.
7560 * The count is incremented when we start a range that's in the set, and
7561 * decremented when we start a range that's not in the set. So its range
7562 * is 0 to 2. Only when the count is zero is something not in the set.
7566 PERL_ARGS_ASSERT__INVLIST_UNION_MAYBE_COMPLEMENT_2ND;
7569 /* If either one is empty, the union is the other one */
7570 if (a == NULL || ((len_a = _invlist_len(a)) == 0)) {
7577 *output = invlist_clone(b);
7579 _invlist_invert(*output);
7581 } /* else *output already = b; */
7584 else if ((len_b = _invlist_len(b)) == 0) {
7589 /* The complement of an empty list is a list that has everything in it,
7590 * so the union with <a> includes everything too */
7595 *output = _new_invlist(1);
7596 _append_range_to_invlist(*output, 0, UV_MAX);
7598 else if (*output != a) {
7599 *output = invlist_clone(a);
7601 /* else *output already = a; */
7605 /* Here both lists exist and are non-empty */
7606 array_a = invlist_array(a);
7607 array_b = invlist_array(b);
7609 /* If are to take the union of 'a' with the complement of b, set it
7610 * up so are looking at b's complement. */
7613 /* To complement, we invert: if the first element is 0, remove it. To
7614 * do this, we just pretend the array starts one later, and clear the
7615 * flag as we don't have to do anything else later */
7616 if (array_b[0] == 0) {
7619 complement_b = FALSE;
7623 /* But if the first element is not zero, we unshift a 0 before the
7624 * array. The data structure reserves a space for that 0 (which
7625 * should be a '1' right now), so physical shifting is unneeded,
7626 * but temporarily change that element to 0. Before exiting the
7627 * routine, we must restore the element to '1' */
7634 /* Size the union for the worst case: that the sets are completely
7636 u = _new_invlist(len_a + len_b);
7638 /* Will contain U+0000 if either component does */
7639 array_u = _invlist_array_init(u, (len_a > 0 && array_a[0] == 0)
7640 || (len_b > 0 && array_b[0] == 0));
7642 /* Go through each list item by item, stopping when exhausted one of
7644 while (i_a < len_a && i_b < len_b) {
7645 UV cp; /* The element to potentially add to the union's array */
7646 bool cp_in_set; /* is it in the the input list's set or not */
7648 /* We need to take one or the other of the two inputs for the union.
7649 * Since we are merging two sorted lists, we take the smaller of the
7650 * next items. In case of a tie, we take the one that is in its set
7651 * first. If we took one not in the set first, it would decrement the
7652 * count, possibly to 0 which would cause it to be output as ending the
7653 * range, and the next time through we would take the same number, and
7654 * output it again as beginning the next range. By doing it the
7655 * opposite way, there is no possibility that the count will be
7656 * momentarily decremented to 0, and thus the two adjoining ranges will
7657 * be seamlessly merged. (In a tie and both are in the set or both not
7658 * in the set, it doesn't matter which we take first.) */
7659 if (array_a[i_a] < array_b[i_b]
7660 || (array_a[i_a] == array_b[i_b]
7661 && ELEMENT_RANGE_MATCHES_INVLIST(i_a)))
7663 cp_in_set = ELEMENT_RANGE_MATCHES_INVLIST(i_a);
7667 cp_in_set = ELEMENT_RANGE_MATCHES_INVLIST(i_b);
7668 cp = array_b[i_b++];
7671 /* Here, have chosen which of the two inputs to look at. Only output
7672 * if the running count changes to/from 0, which marks the
7673 * beginning/end of a range in that's in the set */
7676 array_u[i_u++] = cp;
7683 array_u[i_u++] = cp;
7688 /* Here, we are finished going through at least one of the lists, which
7689 * means there is something remaining in at most one. We check if the list
7690 * that hasn't been exhausted is positioned such that we are in the middle
7691 * of a range in its set or not. (i_a and i_b point to the element beyond
7692 * the one we care about.) If in the set, we decrement 'count'; if 0, there
7693 * is potentially more to output.
7694 * There are four cases:
7695 * 1) Both weren't in their sets, count is 0, and remains 0. What's left
7696 * in the union is entirely from the non-exhausted set.
7697 * 2) Both were in their sets, count is 2. Nothing further should
7698 * be output, as everything that remains will be in the exhausted
7699 * list's set, hence in the union; decrementing to 1 but not 0 insures
7701 * 3) the exhausted was in its set, non-exhausted isn't, count is 1.
7702 * Nothing further should be output because the union includes
7703 * everything from the exhausted set. Not decrementing ensures that.
7704 * 4) the exhausted wasn't in its set, non-exhausted is, count is 1;
7705 * decrementing to 0 insures that we look at the remainder of the
7706 * non-exhausted set */
7707 if ((i_a != len_a && PREV_RANGE_MATCHES_INVLIST(i_a))
7708 || (i_b != len_b && PREV_RANGE_MATCHES_INVLIST(i_b)))
7713 /* The final length is what we've output so far, plus what else is about to
7714 * be output. (If 'count' is non-zero, then the input list we exhausted
7715 * has everything remaining up to the machine's limit in its set, and hence
7716 * in the union, so there will be no further output. */
7719 /* At most one of the subexpressions will be non-zero */
7720 len_u += (len_a - i_a) + (len_b - i_b);
7723 /* Set result to final length, which can change the pointer to array_u, so
7725 if (len_u != _invlist_len(u)) {
7726 invlist_set_len(u, len_u);
7728 array_u = invlist_array(u);
7731 /* When 'count' is 0, the list that was exhausted (if one was shorter than
7732 * the other) ended with everything above it not in its set. That means
7733 * that the remaining part of the union is precisely the same as the
7734 * non-exhausted list, so can just copy it unchanged. (If both list were
7735 * exhausted at the same time, then the operations below will be both 0.)
7738 IV copy_count; /* At most one will have a non-zero copy count */
7739 if ((copy_count = len_a - i_a) > 0) {
7740 Copy(array_a + i_a, array_u + i_u, copy_count, UV);
7742 else if ((copy_count = len_b - i_b) > 0) {
7743 Copy(array_b + i_b, array_u + i_u, copy_count, UV);
7747 /* If we've changed b, restore it */
7752 /* We may be removing a reference to one of the inputs */
7753 if (a == *output || b == *output) {
7754 assert(! invlist_is_iterating(*output));
7755 SvREFCNT_dec_NN(*output);
7763 Perl__invlist_intersection_maybe_complement_2nd(pTHX_ SV* const a, SV* const b, bool complement_b, SV** i)
7765 /* Take the intersection of two inversion lists and point <i> to it. *i
7766 * SHOULD BE DEFINED upon input, and if it points to one of the two lists,
7767 * the reference count to that list will be decremented.
7768 * If <complement_b> is TRUE, the result will be the intersection of <a>
7769 * and the complement (or inversion) of <b> instead of <b> directly.
7771 * The basis for this comes from "Unicode Demystified" Chapter 13 by
7772 * Richard Gillam, published by Addison-Wesley, and explained at some
7773 * length there. The preface says to incorporate its examples into your
7774 * code at your own risk. In fact, it had bugs
7776 * The algorithm is like a merge sort, and is essentially the same as the
7780 UV* array_a; /* a's array */
7782 UV len_a; /* length of a's array */
7785 SV* r; /* the resulting intersection */
7789 UV i_a = 0; /* current index into a's array */
7793 /* running count, as explained in the algorithm source book; items are
7794 * stopped accumulating and are output when the count changes to/from 2.
7795 * The count is incremented when we start a range that's in the set, and
7796 * decremented when we start a range that's not in the set. So its range
7797 * is 0 to 2. Only when the count is 2 is something in the intersection.
7801 PERL_ARGS_ASSERT__INVLIST_INTERSECTION_MAYBE_COMPLEMENT_2ND;
7804 /* Special case if either one is empty */
7805 len_a = _invlist_len(a);
7806 if ((len_a == 0) || ((len_b = _invlist_len(b)) == 0)) {
7808 if (len_a != 0 && complement_b) {
7810 /* Here, 'a' is not empty, therefore from the above 'if', 'b' must
7811 * be empty. Here, also we are using 'b's complement, which hence
7812 * must be every possible code point. Thus the intersection is
7815 *i = invlist_clone(a);
7821 /* else *i is already 'a' */
7825 /* Here, 'a' or 'b' is empty and not using the complement of 'b'. The
7826 * intersection must be empty */
7833 *i = _new_invlist(0);
7837 /* Here both lists exist and are non-empty */
7838 array_a = invlist_array(a);
7839 array_b = invlist_array(b);
7841 /* If are to take the intersection of 'a' with the complement of b, set it
7842 * up so are looking at b's complement. */
7845 /* To complement, we invert: if the first element is 0, remove it. To
7846 * do this, we just pretend the array starts one later, and clear the
7847 * flag as we don't have to do anything else later */
7848 if (array_b[0] == 0) {
7851 complement_b = FALSE;
7855 /* But if the first element is not zero, we unshift a 0 before the
7856 * array. The data structure reserves a space for that 0 (which
7857 * should be a '1' right now), so physical shifting is unneeded,
7858 * but temporarily change that element to 0. Before exiting the
7859 * routine, we must restore the element to '1' */
7866 /* Size the intersection for the worst case: that the intersection ends up
7867 * fragmenting everything to be completely disjoint */
7868 r= _new_invlist(len_a + len_b);
7870 /* Will contain U+0000 iff both components do */
7871 array_r = _invlist_array_init(r, len_a > 0 && array_a[0] == 0
7872 && len_b > 0 && array_b[0] == 0);
7874 /* Go through each list item by item, stopping when exhausted one of
7876 while (i_a < len_a && i_b < len_b) {
7877 UV cp; /* The element to potentially add to the intersection's
7879 bool cp_in_set; /* Is it in the input list's set or not */
7881 /* We need to take one or the other of the two inputs for the
7882 * intersection. Since we are merging two sorted lists, we take the
7883 * smaller of the next items. In case of a tie, we take the one that
7884 * is not in its set first (a difference from the union algorithm). If
7885 * we took one in the set first, it would increment the count, possibly
7886 * to 2 which would cause it to be output as starting a range in the
7887 * intersection, and the next time through we would take that same
7888 * number, and output it again as ending the set. By doing it the
7889 * opposite of this, there is no possibility that the count will be
7890 * momentarily incremented to 2. (In a tie and both are in the set or
7891 * both not in the set, it doesn't matter which we take first.) */
7892 if (array_a[i_a] < array_b[i_b]
7893 || (array_a[i_a] == array_b[i_b]
7894 && ! ELEMENT_RANGE_MATCHES_INVLIST(i_a)))
7896 cp_in_set = ELEMENT_RANGE_MATCHES_INVLIST(i_a);
7900 cp_in_set = ELEMENT_RANGE_MATCHES_INVLIST(i_b);
7904 /* Here, have chosen which of the two inputs to look at. Only output
7905 * if the running count changes to/from 2, which marks the
7906 * beginning/end of a range that's in the intersection */
7910 array_r[i_r++] = cp;
7915 array_r[i_r++] = cp;
7921 /* Here, we are finished going through at least one of the lists, which
7922 * means there is something remaining in at most one. We check if the list
7923 * that has been exhausted is positioned such that we are in the middle
7924 * of a range in its set or not. (i_a and i_b point to elements 1 beyond
7925 * the ones we care about.) There are four cases:
7926 * 1) Both weren't in their sets, count is 0, and remains 0. There's
7927 * nothing left in the intersection.
7928 * 2) Both were in their sets, count is 2 and perhaps is incremented to
7929 * above 2. What should be output is exactly that which is in the
7930 * non-exhausted set, as everything it has is also in the intersection
7931 * set, and everything it doesn't have can't be in the intersection
7932 * 3) The exhausted was in its set, non-exhausted isn't, count is 1, and
7933 * gets incremented to 2. Like the previous case, the intersection is
7934 * everything that remains in the non-exhausted set.
7935 * 4) the exhausted wasn't in its set, non-exhausted is, count is 1, and
7936 * remains 1. And the intersection has nothing more. */
7937 if ((i_a == len_a && PREV_RANGE_MATCHES_INVLIST(i_a))
7938 || (i_b == len_b && PREV_RANGE_MATCHES_INVLIST(i_b)))
7943 /* The final length is what we've output so far plus what else is in the
7944 * intersection. At most one of the subexpressions below will be non-zero */
7947 len_r += (len_a - i_a) + (len_b - i_b);
7950 /* Set result to final length, which can change the pointer to array_r, so
7952 if (len_r != _invlist_len(r)) {
7953 invlist_set_len(r, len_r);
7955 array_r = invlist_array(r);
7958 /* Finish outputting any remaining */
7959 if (count >= 2) { /* At most one will have a non-zero copy count */
7961 if ((copy_count = len_a - i_a) > 0) {
7962 Copy(array_a + i_a, array_r + i_r, copy_count, UV);
7964 else if ((copy_count = len_b - i_b) > 0) {
7965 Copy(array_b + i_b, array_r + i_r, copy_count, UV);
7969 /* If we've changed b, restore it */
7974 /* We may be removing a reference to one of the inputs */
7975 if (a == *i || b == *i) {
7976 assert(! invlist_is_iterating(*i));
7977 SvREFCNT_dec_NN(*i);
7985 Perl__add_range_to_invlist(pTHX_ SV* invlist, const UV start, const UV end)
7987 /* Add the range from 'start' to 'end' inclusive to the inversion list's
7988 * set. A pointer to the inversion list is returned. This may actually be
7989 * a new list, in which case the passed in one has been destroyed. The
7990 * passed in inversion list can be NULL, in which case a new one is created
7991 * with just the one range in it */
7996 if (invlist == NULL) {
7997 invlist = _new_invlist(2);
8001 len = _invlist_len(invlist);
8004 /* If comes after the final entry actually in the list, can just append it
8007 || (! ELEMENT_RANGE_MATCHES_INVLIST(len - 1)
8008 && start >= invlist_array(invlist)[len - 1]))
8010 _append_range_to_invlist(invlist, start, end);
8014 /* Here, can't just append things, create and return a new inversion list
8015 * which is the union of this range and the existing inversion list */
8016 range_invlist = _new_invlist(2);
8017 _append_range_to_invlist(range_invlist, start, end);
8019 _invlist_union(invlist, range_invlist, &invlist);
8021 /* The temporary can be freed */
8022 SvREFCNT_dec_NN(range_invlist);
8029 PERL_STATIC_INLINE SV*
8030 S_add_cp_to_invlist(pTHX_ SV* invlist, const UV cp) {
8031 return _add_range_to_invlist(invlist, cp, cp);
8034 #ifndef PERL_IN_XSUB_RE
8036 Perl__invlist_invert(pTHX_ SV* const invlist)
8038 /* Complement the input inversion list. This adds a 0 if the list didn't
8039 * have a zero; removes it otherwise. As described above, the data
8040 * structure is set up so that this is very efficient */
8042 UV* len_pos = _get_invlist_len_addr(invlist);
8044 PERL_ARGS_ASSERT__INVLIST_INVERT;
8046 assert(! invlist_is_iterating(invlist));
8048 /* The inverse of matching nothing is matching everything */
8049 if (*len_pos == 0) {
8050 _append_range_to_invlist(invlist, 0, UV_MAX);
8054 /* The exclusive or complents 0 to 1; and 1 to 0. If the result is 1, the
8055 * zero element was a 0, so it is being removed, so the length decrements
8056 * by 1; and vice-versa. SvCUR is unaffected */
8057 if (*get_invlist_zero_addr(invlist) ^= 1) {
8066 Perl__invlist_invert_prop(pTHX_ SV* const invlist)
8068 /* Complement the input inversion list (which must be a Unicode property,
8069 * all of which don't match above the Unicode maximum code point.) And
8070 * Perl has chosen to not have the inversion match above that either. This
8071 * adds a 0x110000 if the list didn't end with it, and removes it if it did
8077 PERL_ARGS_ASSERT__INVLIST_INVERT_PROP;
8079 _invlist_invert(invlist);
8081 len = _invlist_len(invlist);
8083 if (len != 0) { /* If empty do nothing */
8084 array = invlist_array(invlist);
8085 if (array[len - 1] != PERL_UNICODE_MAX + 1) {
8086 /* Add 0x110000. First, grow if necessary */
8088 if (invlist_max(invlist) < len) {
8089 invlist_extend(invlist, len);
8090 array = invlist_array(invlist);
8092 invlist_set_len(invlist, len);
8093 array[len - 1] = PERL_UNICODE_MAX + 1;
8095 else { /* Remove the 0x110000 */
8096 invlist_set_len(invlist, len - 1);
8104 PERL_STATIC_INLINE SV*
8105 S_invlist_clone(pTHX_ SV* const invlist)
8108 /* Return a new inversion list that is a copy of the input one, which is
8111 /* Need to allocate extra space to accommodate Perl's addition of a
8112 * trailing NUL to SvPV's, since it thinks they are always strings */
8113 SV* new_invlist = _new_invlist(_invlist_len(invlist) + 1);
8114 STRLEN length = SvCUR(invlist);
8116 PERL_ARGS_ASSERT_INVLIST_CLONE;
8118 SvCUR_set(new_invlist, length); /* This isn't done automatically */
8119 Copy(SvPVX(invlist), SvPVX(new_invlist), length, char);
8124 PERL_STATIC_INLINE UV*
8125 S_get_invlist_iter_addr(pTHX_ SV* invlist)
8127 /* Return the address of the UV that contains the current iteration
8130 PERL_ARGS_ASSERT_GET_INVLIST_ITER_ADDR;
8132 return (UV *) (SvPVX(invlist) + (INVLIST_ITER_OFFSET * sizeof (UV)));
8135 PERL_STATIC_INLINE UV*
8136 S_get_invlist_version_id_addr(pTHX_ SV* invlist)
8138 /* Return the address of the UV that contains the version id. */
8140 PERL_ARGS_ASSERT_GET_INVLIST_VERSION_ID_ADDR;
8142 return (UV *) (SvPVX(invlist) + (INVLIST_VERSION_ID_OFFSET * sizeof (UV)));
8145 PERL_STATIC_INLINE void
8146 S_invlist_iterinit(pTHX_ SV* invlist) /* Initialize iterator for invlist */
8148 PERL_ARGS_ASSERT_INVLIST_ITERINIT;
8150 *get_invlist_iter_addr(invlist) = 0;
8153 PERL_STATIC_INLINE void
8154 S_invlist_iterfinish(pTHX_ SV* invlist)
8156 /* Terminate iterator for invlist. This is to catch development errors.
8157 * Any iteration that is interrupted before completed should call this
8158 * function. Functions that add code points anywhere else but to the end
8159 * of an inversion list assert that they are not in the middle of an
8160 * iteration. If they were, the addition would make the iteration
8161 * problematical: if the iteration hadn't reached the place where things
8162 * were being added, it would be ok */
8164 PERL_ARGS_ASSERT_INVLIST_ITERFINISH;
8166 *get_invlist_iter_addr(invlist) = UV_MAX;
8170 S_invlist_iternext(pTHX_ SV* invlist, UV* start, UV* end)
8172 /* An C<invlist_iterinit> call on <invlist> must be used to set this up.
8173 * This call sets in <*start> and <*end>, the next range in <invlist>.
8174 * Returns <TRUE> if successful and the next call will return the next
8175 * range; <FALSE> if was already at the end of the list. If the latter,
8176 * <*start> and <*end> are unchanged, and the next call to this function
8177 * will start over at the beginning of the list */
8179 UV* pos = get_invlist_iter_addr(invlist);
8180 UV len = _invlist_len(invlist);
8183 PERL_ARGS_ASSERT_INVLIST_ITERNEXT;
8186 *pos = UV_MAX; /* Force iterinit() to be required next time */
8190 array = invlist_array(invlist);
8192 *start = array[(*pos)++];
8198 *end = array[(*pos)++] - 1;
8204 PERL_STATIC_INLINE bool
8205 S_invlist_is_iterating(pTHX_ SV* const invlist)
8207 PERL_ARGS_ASSERT_INVLIST_IS_ITERATING;
8209 return *(get_invlist_iter_addr(invlist)) < UV_MAX;
8212 PERL_STATIC_INLINE UV
8213 S_invlist_highest(pTHX_ SV* const invlist)
8215 /* Returns the highest code point that matches an inversion list. This API
8216 * has an ambiguity, as it returns 0 under either the highest is actually
8217 * 0, or if the list is empty. If this distinction matters to you, check
8218 * for emptiness before calling this function */
8220 UV len = _invlist_len(invlist);
8223 PERL_ARGS_ASSERT_INVLIST_HIGHEST;
8229 array = invlist_array(invlist);
8231 /* The last element in the array in the inversion list always starts a
8232 * range that goes to infinity. That range may be for code points that are
8233 * matched in the inversion list, or it may be for ones that aren't
8234 * matched. In the latter case, the highest code point in the set is one
8235 * less than the beginning of this range; otherwise it is the final element
8236 * of this range: infinity */
8237 return (ELEMENT_RANGE_MATCHES_INVLIST(len - 1))
8239 : array[len - 1] - 1;
8242 #ifndef PERL_IN_XSUB_RE
8244 Perl__invlist_contents(pTHX_ SV* const invlist)
8246 /* Get the contents of an inversion list into a string SV so that they can
8247 * be printed out. It uses the format traditionally done for debug tracing
8251 SV* output = newSVpvs("\n");
8253 PERL_ARGS_ASSERT__INVLIST_CONTENTS;
8255 assert(! invlist_is_iterating(invlist));
8257 invlist_iterinit(invlist);
8258 while (invlist_iternext(invlist, &start, &end)) {
8259 if (end == UV_MAX) {
8260 Perl_sv_catpvf(aTHX_ output, "%04"UVXf"\tINFINITY\n", start);
8262 else if (end != start) {
8263 Perl_sv_catpvf(aTHX_ output, "%04"UVXf"\t%04"UVXf"\n",
8267 Perl_sv_catpvf(aTHX_ output, "%04"UVXf"\n", start);
8275 #ifdef PERL_ARGS_ASSERT__INVLIST_DUMP
8277 Perl__invlist_dump(pTHX_ SV* const invlist, const char * const header)
8279 /* Dumps out the ranges in an inversion list. The string 'header'
8280 * if present is output on a line before the first range */
8284 PERL_ARGS_ASSERT__INVLIST_DUMP;
8286 if (header && strlen(header)) {
8287 PerlIO_printf(Perl_debug_log, "%s\n", header);
8289 if (invlist_is_iterating(invlist)) {
8290 PerlIO_printf(Perl_debug_log, "Can't dump because is in middle of iterating\n");
8294 invlist_iterinit(invlist);
8295 while (invlist_iternext(invlist, &start, &end)) {
8296 if (end == UV_MAX) {
8297 PerlIO_printf(Perl_debug_log, "0x%04"UVXf" .. INFINITY\n", start);
8299 else if (end != start) {
8300 PerlIO_printf(Perl_debug_log, "0x%04"UVXf" .. 0x%04"UVXf"\n",
8304 PerlIO_printf(Perl_debug_log, "0x%04"UVXf"\n", start);
8312 S__invlistEQ(pTHX_ SV* const a, SV* const b, bool complement_b)
8314 /* Return a boolean as to if the two passed in inversion lists are
8315 * identical. The final argument, if TRUE, says to take the complement of
8316 * the second inversion list before doing the comparison */
8318 UV* array_a = invlist_array(a);
8319 UV* array_b = invlist_array(b);
8320 UV len_a = _invlist_len(a);
8321 UV len_b = _invlist_len(b);
8323 UV i = 0; /* current index into the arrays */
8324 bool retval = TRUE; /* Assume are identical until proven otherwise */
8326 PERL_ARGS_ASSERT__INVLISTEQ;
8328 /* If are to compare 'a' with the complement of b, set it
8329 * up so are looking at b's complement. */
8332 /* The complement of nothing is everything, so <a> would have to have
8333 * just one element, starting at zero (ending at infinity) */
8335 return (len_a == 1 && array_a[0] == 0);
8337 else if (array_b[0] == 0) {
8339 /* Otherwise, to complement, we invert. Here, the first element is
8340 * 0, just remove it. To do this, we just pretend the array starts
8341 * one later, and clear the flag as we don't have to do anything
8346 complement_b = FALSE;
8350 /* But if the first element is not zero, we unshift a 0 before the
8351 * array. The data structure reserves a space for that 0 (which
8352 * should be a '1' right now), so physical shifting is unneeded,
8353 * but temporarily change that element to 0. Before exiting the
8354 * routine, we must restore the element to '1' */
8361 /* Make sure that the lengths are the same, as well as the final element
8362 * before looping through the remainder. (Thus we test the length, final,
8363 * and first elements right off the bat) */
8364 if (len_a != len_b || array_a[len_a-1] != array_b[len_a-1]) {
8367 else for (i = 0; i < len_a - 1; i++) {
8368 if (array_a[i] != array_b[i]) {
8381 #undef HEADER_LENGTH
8382 #undef INVLIST_INITIAL_LENGTH
8383 #undef TO_INTERNAL_SIZE
8384 #undef FROM_INTERNAL_SIZE
8385 #undef INVLIST_LEN_OFFSET
8386 #undef INVLIST_ZERO_OFFSET
8387 #undef INVLIST_ITER_OFFSET
8388 #undef INVLIST_VERSION_ID
8389 #undef INVLIST_PREVIOUS_INDEX_OFFSET
8391 /* End of inversion list object */
8394 S_parse_lparen_question_flags(pTHX_ struct RExC_state_t *pRExC_state)
8396 /* This parses the flags that are in either the '(?foo)' or '(?foo:bar)'
8397 * constructs, and updates RExC_flags with them. On input, RExC_parse
8398 * should point to the first flag; it is updated on output to point to the
8399 * final ')' or ':'. There needs to be at least one flag, or this will
8402 /* for (?g), (?gc), and (?o) warnings; warning
8403 about (?c) will warn about (?g) -- japhy */
8405 #define WASTED_O 0x01
8406 #define WASTED_G 0x02
8407 #define WASTED_C 0x04
8408 #define WASTED_GC (0x02|0x04)
8409 I32 wastedflags = 0x00;
8410 U32 posflags = 0, negflags = 0;
8411 U32 *flagsp = &posflags;
8412 char has_charset_modifier = '\0';
8414 bool has_use_defaults = FALSE;
8415 const char* const seqstart = RExC_parse - 1; /* Point to the '?' */
8417 PERL_ARGS_ASSERT_PARSE_LPAREN_QUESTION_FLAGS;
8419 /* '^' as an initial flag sets certain defaults */
8420 if (UCHARAT(RExC_parse) == '^') {
8422 has_use_defaults = TRUE;
8423 STD_PMMOD_FLAGS_CLEAR(&RExC_flags);
8424 set_regex_charset(&RExC_flags, (RExC_utf8 || RExC_uni_semantics)
8425 ? REGEX_UNICODE_CHARSET
8426 : REGEX_DEPENDS_CHARSET);
8429 cs = get_regex_charset(RExC_flags);
8430 if (cs == REGEX_DEPENDS_CHARSET
8431 && (RExC_utf8 || RExC_uni_semantics))
8433 cs = REGEX_UNICODE_CHARSET;
8436 while (*RExC_parse) {
8437 /* && strchr("iogcmsx", *RExC_parse) */
8438 /* (?g), (?gc) and (?o) are useless here
8439 and must be globally applied -- japhy */
8440 switch (*RExC_parse) {
8442 /* Code for the imsx flags */
8443 CASE_STD_PMMOD_FLAGS_PARSE_SET(flagsp);
8445 case LOCALE_PAT_MOD:
8446 if (has_charset_modifier) {
8447 goto excess_modifier;
8449 else if (flagsp == &negflags) {
8452 cs = REGEX_LOCALE_CHARSET;
8453 has_charset_modifier = LOCALE_PAT_MOD;
8454 RExC_contains_locale = 1;
8456 case UNICODE_PAT_MOD:
8457 if (has_charset_modifier) {
8458 goto excess_modifier;
8460 else if (flagsp == &negflags) {
8463 cs = REGEX_UNICODE_CHARSET;
8464 has_charset_modifier = UNICODE_PAT_MOD;
8466 case ASCII_RESTRICT_PAT_MOD:
8467 if (flagsp == &negflags) {
8470 if (has_charset_modifier) {
8471 if (cs != REGEX_ASCII_RESTRICTED_CHARSET) {
8472 goto excess_modifier;
8474 /* Doubled modifier implies more restricted */
8475 cs = REGEX_ASCII_MORE_RESTRICTED_CHARSET;
8478 cs = REGEX_ASCII_RESTRICTED_CHARSET;
8480 has_charset_modifier = ASCII_RESTRICT_PAT_MOD;
8482 case DEPENDS_PAT_MOD:
8483 if (has_use_defaults) {
8484 goto fail_modifiers;
8486 else if (flagsp == &negflags) {
8489 else if (has_charset_modifier) {
8490 goto excess_modifier;
8493 /* The dual charset means unicode semantics if the
8494 * pattern (or target, not known until runtime) are
8495 * utf8, or something in the pattern indicates unicode
8497 cs = (RExC_utf8 || RExC_uni_semantics)
8498 ? REGEX_UNICODE_CHARSET
8499 : REGEX_DEPENDS_CHARSET;
8500 has_charset_modifier = DEPENDS_PAT_MOD;
8504 if (has_charset_modifier == ASCII_RESTRICT_PAT_MOD) {
8505 vFAIL2("Regexp modifier \"%c\" may appear a maximum of twice", ASCII_RESTRICT_PAT_MOD);
8507 else if (has_charset_modifier == *(RExC_parse - 1)) {
8508 vFAIL2("Regexp modifier \"%c\" may not appear twice", *(RExC_parse - 1));
8511 vFAIL3("Regexp modifiers \"%c\" and \"%c\" are mutually exclusive", has_charset_modifier, *(RExC_parse - 1));
8516 vFAIL2("Regexp modifier \"%c\" may not appear after the \"-\"", *(RExC_parse - 1));
8518 case ONCE_PAT_MOD: /* 'o' */
8519 case GLOBAL_PAT_MOD: /* 'g' */
8520 if (SIZE_ONLY && ckWARN(WARN_REGEXP)) {
8521 const I32 wflagbit = *RExC_parse == 'o' ? WASTED_O : WASTED_G;
8522 if (! (wastedflags & wflagbit) ) {
8523 wastedflags |= wflagbit;
8526 "Useless (%s%c) - %suse /%c modifier",
8527 flagsp == &negflags ? "?-" : "?",
8529 flagsp == &negflags ? "don't " : "",
8536 case CONTINUE_PAT_MOD: /* 'c' */
8537 if (SIZE_ONLY && ckWARN(WARN_REGEXP)) {
8538 if (! (wastedflags & WASTED_C) ) {
8539 wastedflags |= WASTED_GC;
8542 "Useless (%sc) - %suse /gc modifier",
8543 flagsp == &negflags ? "?-" : "?",
8544 flagsp == &negflags ? "don't " : ""
8549 case KEEPCOPY_PAT_MOD: /* 'p' */
8550 if (flagsp == &negflags) {
8552 ckWARNreg(RExC_parse + 1,"Useless use of (?-p)");
8554 *flagsp |= RXf_PMf_KEEPCOPY;
8558 /* A flag is a default iff it is following a minus, so
8559 * if there is a minus, it means will be trying to
8560 * re-specify a default which is an error */
8561 if (has_use_defaults || flagsp == &negflags) {
8562 goto fail_modifiers;
8565 wastedflags = 0; /* reset so (?g-c) warns twice */
8569 RExC_flags |= posflags;
8570 RExC_flags &= ~negflags;
8571 set_regex_charset(&RExC_flags, cs);
8577 vFAIL3("Sequence (%.*s...) not recognized",
8578 RExC_parse-seqstart, seqstart);
8587 - reg - regular expression, i.e. main body or parenthesized thing
8589 * Caller must absorb opening parenthesis.
8591 * Combining parenthesis handling with the base level of regular expression
8592 * is a trifle forced, but the need to tie the tails of the branches to what
8593 * follows makes it hard to avoid.
8595 #define REGTAIL(x,y,z) regtail((x),(y),(z),depth+1)
8597 #define REGTAIL_STUDY(x,y,z) regtail_study((x),(y),(z),depth+1)
8599 #define REGTAIL_STUDY(x,y,z) regtail((x),(y),(z),depth+1)
8602 /* Returns NULL, setting *flagp to TRYAGAIN at the end of (?) that only sets
8603 flags. Returns NULL, setting *flagp to RESTART_UTF8 if the sizing scan
8604 needs to be restarted.
8605 Otherwise would only return NULL if regbranch() returns NULL, which
8608 S_reg(pTHX_ RExC_state_t *pRExC_state, I32 paren, I32 *flagp,U32 depth)
8609 /* paren: Parenthesized? 0=top; 1,2=inside '(': changed to letter.
8610 * 2 is like 1, but indicates that nextchar() has been called to advance
8611 * RExC_parse beyond the '('. Things like '(?' are indivisible tokens, and
8612 * this flag alerts us to the need to check for that */
8615 regnode *ret; /* Will be the head of the group. */
8618 regnode *ender = NULL;
8621 U32 oregflags = RExC_flags;
8622 bool have_branch = 0;
8624 I32 freeze_paren = 0;
8625 I32 after_freeze = 0;
8627 char * parse_start = RExC_parse; /* MJD */
8628 char * const oregcomp_parse = RExC_parse;
8630 GET_RE_DEBUG_FLAGS_DECL;
8632 PERL_ARGS_ASSERT_REG;
8633 DEBUG_PARSE("reg ");
8635 *flagp = 0; /* Tentatively. */
8638 /* Make an OPEN node, if parenthesized. */
8641 /* Under /x, space and comments can be gobbled up between the '(' and
8642 * here (if paren ==2). The forms '(*VERB' and '(?...' disallow such
8643 * intervening space, as the sequence is a token, and a token should be
8645 bool has_intervening_patws = paren == 2 && *(RExC_parse - 1) != '(';
8647 if ( *RExC_parse == '*') { /* (*VERB:ARG) */
8648 char *start_verb = RExC_parse;
8649 STRLEN verb_len = 0;
8650 char *start_arg = NULL;
8651 unsigned char op = 0;
8653 int internal_argval = 0; /* internal_argval is only useful if !argok */
8655 if (has_intervening_patws && SIZE_ONLY) {
8656 ckWARNregdep(RExC_parse + 1, "In '(*VERB...)', splitting the initial '(*' is deprecated");
8658 while ( *RExC_parse && *RExC_parse != ')' ) {
8659 if ( *RExC_parse == ':' ) {
8660 start_arg = RExC_parse + 1;
8666 verb_len = RExC_parse - start_verb;
8669 while ( *RExC_parse && *RExC_parse != ')' )
8671 if ( *RExC_parse != ')' )
8672 vFAIL("Unterminated verb pattern argument");
8673 if ( RExC_parse == start_arg )
8676 if ( *RExC_parse != ')' )
8677 vFAIL("Unterminated verb pattern");
8680 switch ( *start_verb ) {
8681 case 'A': /* (*ACCEPT) */
8682 if ( memEQs(start_verb,verb_len,"ACCEPT") ) {
8684 internal_argval = RExC_nestroot;
8687 case 'C': /* (*COMMIT) */
8688 if ( memEQs(start_verb,verb_len,"COMMIT") )
8691 case 'F': /* (*FAIL) */
8692 if ( verb_len==1 || memEQs(start_verb,verb_len,"FAIL") ) {
8697 case ':': /* (*:NAME) */
8698 case 'M': /* (*MARK:NAME) */
8699 if ( verb_len==0 || memEQs(start_verb,verb_len,"MARK") ) {
8704 case 'P': /* (*PRUNE) */
8705 if ( memEQs(start_verb,verb_len,"PRUNE") )
8708 case 'S': /* (*SKIP) */
8709 if ( memEQs(start_verb,verb_len,"SKIP") )
8712 case 'T': /* (*THEN) */
8713 /* [19:06] <TimToady> :: is then */
8714 if ( memEQs(start_verb,verb_len,"THEN") ) {
8716 RExC_seen |= REG_SEEN_CUTGROUP;
8722 vFAIL3("Unknown verb pattern '%.*s'",
8723 verb_len, start_verb);
8726 if ( start_arg && internal_argval ) {
8727 vFAIL3("Verb pattern '%.*s' may not have an argument",
8728 verb_len, start_verb);
8729 } else if ( argok < 0 && !start_arg ) {
8730 vFAIL3("Verb pattern '%.*s' has a mandatory argument",
8731 verb_len, start_verb);
8733 ret = reganode(pRExC_state, op, internal_argval);
8734 if ( ! internal_argval && ! SIZE_ONLY ) {
8736 SV *sv = newSVpvn( start_arg, RExC_parse - start_arg);
8737 ARG(ret) = add_data( pRExC_state, 1, "S" );
8738 RExC_rxi->data->data[ARG(ret)]=(void*)sv;
8745 if (!internal_argval)
8746 RExC_seen |= REG_SEEN_VERBARG;
8747 } else if ( start_arg ) {
8748 vFAIL3("Verb pattern '%.*s' may not have an argument",
8749 verb_len, start_verb);
8751 ret = reg_node(pRExC_state, op);
8753 nextchar(pRExC_state);
8756 if (*RExC_parse == '?') { /* (?...) */
8757 bool is_logical = 0;
8758 const char * const seqstart = RExC_parse;
8759 if (has_intervening_patws && SIZE_ONLY) {
8760 ckWARNregdep(RExC_parse + 1, "In '(?...)', splitting the initial '(?' is deprecated");
8764 paren = *RExC_parse++;
8765 ret = NULL; /* For look-ahead/behind. */
8768 case 'P': /* (?P...) variants for those used to PCRE/Python */
8769 paren = *RExC_parse++;
8770 if ( paren == '<') /* (?P<...>) named capture */
8772 else if (paren == '>') { /* (?P>name) named recursion */
8773 goto named_recursion;
8775 else if (paren == '=') { /* (?P=...) named backref */
8776 /* this pretty much dupes the code for \k<NAME> in regatom(), if
8777 you change this make sure you change that */
8778 char* name_start = RExC_parse;
8780 SV *sv_dat = reg_scan_name(pRExC_state,
8781 SIZE_ONLY ? REG_RSN_RETURN_NULL : REG_RSN_RETURN_DATA);
8782 if (RExC_parse == name_start || *RExC_parse != ')')
8783 vFAIL2("Sequence %.3s... not terminated",parse_start);
8786 num = add_data( pRExC_state, 1, "S" );
8787 RExC_rxi->data->data[num]=(void*)sv_dat;
8788 SvREFCNT_inc_simple_void(sv_dat);
8791 ret = reganode(pRExC_state,
8794 : (ASCII_FOLD_RESTRICTED)
8796 : (AT_LEAST_UNI_SEMANTICS)
8804 Set_Node_Offset(ret, parse_start+1);
8805 Set_Node_Cur_Length(ret); /* MJD */
8807 nextchar(pRExC_state);
8811 vFAIL3("Sequence (%.*s...) not recognized", RExC_parse-seqstart, seqstart);
8813 case '<': /* (?<...) */
8814 if (*RExC_parse == '!')
8816 else if (*RExC_parse != '=')
8822 case '\'': /* (?'...') */
8823 name_start= RExC_parse;
8824 svname = reg_scan_name(pRExC_state,
8825 SIZE_ONLY ? /* reverse test from the others */
8826 REG_RSN_RETURN_NAME :
8827 REG_RSN_RETURN_NULL);
8828 if (RExC_parse == name_start) {
8830 vFAIL3("Sequence (%.*s...) not recognized", RExC_parse-seqstart, seqstart);
8833 if (*RExC_parse != paren)
8834 vFAIL2("Sequence (?%c... not terminated",
8835 paren=='>' ? '<' : paren);
8839 if (!svname) /* shouldn't happen */
8841 "panic: reg_scan_name returned NULL");
8842 if (!RExC_paren_names) {
8843 RExC_paren_names= newHV();
8844 sv_2mortal(MUTABLE_SV(RExC_paren_names));
8846 RExC_paren_name_list= newAV();
8847 sv_2mortal(MUTABLE_SV(RExC_paren_name_list));
8850 he_str = hv_fetch_ent( RExC_paren_names, svname, 1, 0 );
8852 sv_dat = HeVAL(he_str);
8854 /* croak baby croak */
8856 "panic: paren_name hash element allocation failed");
8857 } else if ( SvPOK(sv_dat) ) {
8858 /* (?|...) can mean we have dupes so scan to check
8859 its already been stored. Maybe a flag indicating
8860 we are inside such a construct would be useful,
8861 but the arrays are likely to be quite small, so
8862 for now we punt -- dmq */
8863 IV count = SvIV(sv_dat);
8864 I32 *pv = (I32*)SvPVX(sv_dat);
8866 for ( i = 0 ; i < count ; i++ ) {
8867 if ( pv[i] == RExC_npar ) {
8873 pv = (I32*)SvGROW(sv_dat, SvCUR(sv_dat) + sizeof(I32)+1);
8874 SvCUR_set(sv_dat, SvCUR(sv_dat) + sizeof(I32));
8875 pv[count] = RExC_npar;
8876 SvIV_set(sv_dat, SvIVX(sv_dat) + 1);
8879 (void)SvUPGRADE(sv_dat,SVt_PVNV);
8880 sv_setpvn(sv_dat, (char *)&(RExC_npar), sizeof(I32));
8882 SvIV_set(sv_dat, 1);
8885 /* Yes this does cause a memory leak in debugging Perls */
8886 if (!av_store(RExC_paren_name_list, RExC_npar, SvREFCNT_inc(svname)))
8887 SvREFCNT_dec_NN(svname);
8890 /*sv_dump(sv_dat);*/
8892 nextchar(pRExC_state);
8894 goto capturing_parens;
8896 RExC_seen |= REG_SEEN_LOOKBEHIND;
8897 RExC_in_lookbehind++;
8899 case '=': /* (?=...) */
8900 RExC_seen_zerolen++;
8902 case '!': /* (?!...) */
8903 RExC_seen_zerolen++;
8904 if (*RExC_parse == ')') {
8905 ret=reg_node(pRExC_state, OPFAIL);
8906 nextchar(pRExC_state);
8910 case '|': /* (?|...) */
8911 /* branch reset, behave like a (?:...) except that
8912 buffers in alternations share the same numbers */
8914 after_freeze = freeze_paren = RExC_npar;
8916 case ':': /* (?:...) */
8917 case '>': /* (?>...) */
8919 case '$': /* (?$...) */
8920 case '@': /* (?@...) */
8921 vFAIL2("Sequence (?%c...) not implemented", (int)paren);
8923 case '#': /* (?#...) */
8924 /* XXX As soon as we disallow separating the '?' and '*' (by
8925 * spaces or (?#...) comment), it is believed that this case
8926 * will be unreachable and can be removed. See
8928 while (*RExC_parse && *RExC_parse != ')')
8930 if (*RExC_parse != ')')
8931 FAIL("Sequence (?#... not terminated");
8932 nextchar(pRExC_state);
8935 case '0' : /* (?0) */
8936 case 'R' : /* (?R) */
8937 if (*RExC_parse != ')')
8938 FAIL("Sequence (?R) not terminated");
8939 ret = reg_node(pRExC_state, GOSTART);
8940 *flagp |= POSTPONED;
8941 nextchar(pRExC_state);
8944 { /* named and numeric backreferences */
8946 case '&': /* (?&NAME) */
8947 parse_start = RExC_parse - 1;
8950 SV *sv_dat = reg_scan_name(pRExC_state,
8951 SIZE_ONLY ? REG_RSN_RETURN_NULL : REG_RSN_RETURN_DATA);
8952 num = sv_dat ? *((I32 *)SvPVX(sv_dat)) : 0;
8954 goto gen_recurse_regop;
8955 assert(0); /* NOT REACHED */
8957 if (!(RExC_parse[0] >= '1' && RExC_parse[0] <= '9')) {
8959 vFAIL("Illegal pattern");
8961 goto parse_recursion;
8963 case '-': /* (?-1) */
8964 if (!(RExC_parse[0] >= '1' && RExC_parse[0] <= '9')) {
8965 RExC_parse--; /* rewind to let it be handled later */
8969 case '1': case '2': case '3': case '4': /* (?1) */
8970 case '5': case '6': case '7': case '8': case '9':
8973 num = atoi(RExC_parse);
8974 parse_start = RExC_parse - 1; /* MJD */
8975 if (*RExC_parse == '-')
8977 while (isDIGIT(*RExC_parse))
8979 if (*RExC_parse!=')')
8980 vFAIL("Expecting close bracket");
8983 if ( paren == '-' ) {
8985 Diagram of capture buffer numbering.
8986 Top line is the normal capture buffer numbers
8987 Bottom line is the negative indexing as from
8991 /(a(x)y)(a(b(c(?-2)d)e)f)(g(h))/
8995 num = RExC_npar + num;
8998 vFAIL("Reference to nonexistent group");
9000 } else if ( paren == '+' ) {
9001 num = RExC_npar + num - 1;
9004 ret = reganode(pRExC_state, GOSUB, num);
9006 if (num > (I32)RExC_rx->nparens) {
9008 vFAIL("Reference to nonexistent group");
9010 ARG2L_SET( ret, RExC_recurse_count++);
9012 DEBUG_OPTIMISE_MORE_r(PerlIO_printf(Perl_debug_log,
9013 "Recurse #%"UVuf" to %"IVdf"\n", (UV)ARG(ret), (IV)ARG2L(ret)));
9017 RExC_seen |= REG_SEEN_RECURSE;
9018 Set_Node_Length(ret, 1 + regarglen[OP(ret)]); /* MJD */
9019 Set_Node_Offset(ret, parse_start); /* MJD */
9021 *flagp |= POSTPONED;
9022 nextchar(pRExC_state);
9024 } /* named and numeric backreferences */
9025 assert(0); /* NOT REACHED */
9027 case '?': /* (??...) */
9029 if (*RExC_parse != '{') {
9031 vFAIL3("Sequence (%.*s...) not recognized", RExC_parse-seqstart, seqstart);
9034 *flagp |= POSTPONED;
9035 paren = *RExC_parse++;
9037 case '{': /* (?{...}) */
9040 struct reg_code_block *cb;
9042 RExC_seen_zerolen++;
9044 if ( !pRExC_state->num_code_blocks
9045 || pRExC_state->code_index >= pRExC_state->num_code_blocks
9046 || pRExC_state->code_blocks[pRExC_state->code_index].start
9047 != (STRLEN)((RExC_parse -3 - (is_logical ? 1 : 0))
9050 if (RExC_pm_flags & PMf_USE_RE_EVAL)
9051 FAIL("panic: Sequence (?{...}): no code block found\n");
9052 FAIL("Eval-group not allowed at runtime, use re 'eval'");
9054 /* this is a pre-compiled code block (?{...}) */
9055 cb = &pRExC_state->code_blocks[pRExC_state->code_index];
9056 RExC_parse = RExC_start + cb->end;
9059 if (cb->src_regex) {
9060 n = add_data(pRExC_state, 2, "rl");
9061 RExC_rxi->data->data[n] =
9062 (void*)SvREFCNT_inc((SV*)cb->src_regex);
9063 RExC_rxi->data->data[n+1] = (void*)o;
9066 n = add_data(pRExC_state, 1,
9067 (RExC_pm_flags & PMf_HAS_CV) ? "L" : "l");
9068 RExC_rxi->data->data[n] = (void*)o;
9071 pRExC_state->code_index++;
9072 nextchar(pRExC_state);
9076 ret = reg_node(pRExC_state, LOGICAL);
9077 eval = reganode(pRExC_state, EVAL, n);
9080 /* for later propagation into (??{}) return value */
9081 eval->flags = (U8) (RExC_flags & RXf_PMf_COMPILETIME);
9083 REGTAIL(pRExC_state, ret, eval);
9084 /* deal with the length of this later - MJD */
9087 ret = reganode(pRExC_state, EVAL, n);
9088 Set_Node_Length(ret, RExC_parse - parse_start + 1);
9089 Set_Node_Offset(ret, parse_start);
9092 case '(': /* (?(?{...})...) and (?(?=...)...) */
9095 if (RExC_parse[0] == '?') { /* (?(?...)) */
9096 if (RExC_parse[1] == '=' || RExC_parse[1] == '!'
9097 || RExC_parse[1] == '<'
9098 || RExC_parse[1] == '{') { /* Lookahead or eval. */
9102 ret = reg_node(pRExC_state, LOGICAL);
9106 tail = reg(pRExC_state, 1, &flag, depth+1);
9107 if (flag & RESTART_UTF8) {
9108 *flagp = RESTART_UTF8;
9111 REGTAIL(pRExC_state, ret, tail);
9115 else if ( RExC_parse[0] == '<' /* (?(<NAME>)...) */
9116 || RExC_parse[0] == '\'' ) /* (?('NAME')...) */
9118 char ch = RExC_parse[0] == '<' ? '>' : '\'';
9119 char *name_start= RExC_parse++;
9121 SV *sv_dat=reg_scan_name(pRExC_state,
9122 SIZE_ONLY ? REG_RSN_RETURN_NULL : REG_RSN_RETURN_DATA);
9123 if (RExC_parse == name_start || *RExC_parse != ch)
9124 vFAIL2("Sequence (?(%c... not terminated",
9125 (ch == '>' ? '<' : ch));
9128 num = add_data( pRExC_state, 1, "S" );
9129 RExC_rxi->data->data[num]=(void*)sv_dat;
9130 SvREFCNT_inc_simple_void(sv_dat);
9132 ret = reganode(pRExC_state,NGROUPP,num);
9133 goto insert_if_check_paren;
9135 else if (RExC_parse[0] == 'D' &&
9136 RExC_parse[1] == 'E' &&
9137 RExC_parse[2] == 'F' &&
9138 RExC_parse[3] == 'I' &&
9139 RExC_parse[4] == 'N' &&
9140 RExC_parse[5] == 'E')
9142 ret = reganode(pRExC_state,DEFINEP,0);
9145 goto insert_if_check_paren;
9147 else if (RExC_parse[0] == 'R') {
9150 if (RExC_parse[0] >= '1' && RExC_parse[0] <= '9' ) {
9151 parno = atoi(RExC_parse++);
9152 while (isDIGIT(*RExC_parse))
9154 } else if (RExC_parse[0] == '&') {
9157 sv_dat = reg_scan_name(pRExC_state,
9158 SIZE_ONLY ? REG_RSN_RETURN_NULL : REG_RSN_RETURN_DATA);
9159 parno = sv_dat ? *((I32 *)SvPVX(sv_dat)) : 0;
9161 ret = reganode(pRExC_state,INSUBP,parno);
9162 goto insert_if_check_paren;
9164 else if (RExC_parse[0] >= '1' && RExC_parse[0] <= '9' ) {
9167 parno = atoi(RExC_parse++);
9169 while (isDIGIT(*RExC_parse))
9171 ret = reganode(pRExC_state, GROUPP, parno);
9173 insert_if_check_paren:
9174 if ((c = *nextchar(pRExC_state)) != ')')
9175 vFAIL("Switch condition not recognized");
9177 REGTAIL(pRExC_state, ret, reganode(pRExC_state, IFTHEN, 0));
9178 br = regbranch(pRExC_state, &flags, 1,depth+1);
9180 if (flags & RESTART_UTF8) {
9181 *flagp = RESTART_UTF8;
9184 FAIL2("panic: regbranch returned NULL, flags=%#X",
9187 REGTAIL(pRExC_state, br, reganode(pRExC_state, LONGJMP, 0));
9188 c = *nextchar(pRExC_state);
9193 vFAIL("(?(DEFINE)....) does not allow branches");
9194 lastbr = reganode(pRExC_state, IFTHEN, 0); /* Fake one for optimizer. */
9195 if (!regbranch(pRExC_state, &flags, 1,depth+1)) {
9196 if (flags & RESTART_UTF8) {
9197 *flagp = RESTART_UTF8;
9200 FAIL2("panic: regbranch returned NULL, flags=%#X",
9203 REGTAIL(pRExC_state, ret, lastbr);
9206 c = *nextchar(pRExC_state);
9211 vFAIL("Switch (?(condition)... contains too many branches");
9212 ender = reg_node(pRExC_state, TAIL);
9213 REGTAIL(pRExC_state, br, ender);
9215 REGTAIL(pRExC_state, lastbr, ender);
9216 REGTAIL(pRExC_state, NEXTOPER(NEXTOPER(lastbr)), ender);
9219 REGTAIL(pRExC_state, ret, ender);
9220 RExC_size++; /* XXX WHY do we need this?!!
9221 For large programs it seems to be required
9222 but I can't figure out why. -- dmq*/
9226 vFAIL2("Unknown switch condition (?(%.2s", RExC_parse);
9229 case '[': /* (?[ ... ]) */
9230 return handle_regex_sets(pRExC_state, NULL, flagp, depth,
9233 RExC_parse--; /* for vFAIL to print correctly */
9234 vFAIL("Sequence (? incomplete");
9236 default: /* e.g., (?i) */
9239 parse_lparen_question_flags(pRExC_state);
9240 if (UCHARAT(RExC_parse) != ':') {
9241 nextchar(pRExC_state);
9246 nextchar(pRExC_state);
9256 ret = reganode(pRExC_state, OPEN, parno);
9259 RExC_nestroot = parno;
9260 if (RExC_seen & REG_SEEN_RECURSE
9261 && !RExC_open_parens[parno-1])
9263 DEBUG_OPTIMISE_MORE_r(PerlIO_printf(Perl_debug_log,
9264 "Setting open paren #%"IVdf" to %d\n",
9265 (IV)parno, REG_NODE_NUM(ret)));
9266 RExC_open_parens[parno-1]= ret;
9269 Set_Node_Length(ret, 1); /* MJD */
9270 Set_Node_Offset(ret, RExC_parse); /* MJD */
9278 /* Pick up the branches, linking them together. */
9279 parse_start = RExC_parse; /* MJD */
9280 br = regbranch(pRExC_state, &flags, 1,depth+1);
9282 /* branch_len = (paren != 0); */
9285 if (flags & RESTART_UTF8) {
9286 *flagp = RESTART_UTF8;
9289 FAIL2("panic: regbranch returned NULL, flags=%#X", flags);
9291 if (*RExC_parse == '|') {
9292 if (!SIZE_ONLY && RExC_extralen) {
9293 reginsert(pRExC_state, BRANCHJ, br, depth+1);
9296 reginsert(pRExC_state, BRANCH, br, depth+1);
9297 Set_Node_Length(br, paren != 0);
9298 Set_Node_Offset_To_R(br-RExC_emit_start, parse_start-RExC_start);
9302 RExC_extralen += 1; /* For BRANCHJ-BRANCH. */
9304 else if (paren == ':') {
9305 *flagp |= flags&SIMPLE;
9307 if (is_open) { /* Starts with OPEN. */
9308 REGTAIL(pRExC_state, ret, br); /* OPEN -> first. */
9310 else if (paren != '?') /* Not Conditional */
9312 *flagp |= flags & (SPSTART | HASWIDTH | POSTPONED);
9314 while (*RExC_parse == '|') {
9315 if (!SIZE_ONLY && RExC_extralen) {
9316 ender = reganode(pRExC_state, LONGJMP,0);
9317 REGTAIL(pRExC_state, NEXTOPER(NEXTOPER(lastbr)), ender); /* Append to the previous. */
9320 RExC_extralen += 2; /* Account for LONGJMP. */
9321 nextchar(pRExC_state);
9323 if (RExC_npar > after_freeze)
9324 after_freeze = RExC_npar;
9325 RExC_npar = freeze_paren;
9327 br = regbranch(pRExC_state, &flags, 0, depth+1);
9330 if (flags & RESTART_UTF8) {
9331 *flagp = RESTART_UTF8;
9334 FAIL2("panic: regbranch returned NULL, flags=%#X", flags);
9336 REGTAIL(pRExC_state, lastbr, br); /* BRANCH -> BRANCH. */
9338 *flagp |= flags & (SPSTART | HASWIDTH | POSTPONED);
9341 if (have_branch || paren != ':') {
9342 /* Make a closing node, and hook it on the end. */
9345 ender = reg_node(pRExC_state, TAIL);
9348 ender = reganode(pRExC_state, CLOSE, parno);
9349 if (!SIZE_ONLY && RExC_seen & REG_SEEN_RECURSE) {
9350 DEBUG_OPTIMISE_MORE_r(PerlIO_printf(Perl_debug_log,
9351 "Setting close paren #%"IVdf" to %d\n",
9352 (IV)parno, REG_NODE_NUM(ender)));
9353 RExC_close_parens[parno-1]= ender;
9354 if (RExC_nestroot == parno)
9357 Set_Node_Offset(ender,RExC_parse+1); /* MJD */
9358 Set_Node_Length(ender,1); /* MJD */
9364 *flagp &= ~HASWIDTH;
9367 ender = reg_node(pRExC_state, SUCCEED);
9370 ender = reg_node(pRExC_state, END);
9372 assert(!RExC_opend); /* there can only be one! */
9377 DEBUG_PARSE_r(if (!SIZE_ONLY) {
9378 SV * const mysv_val1=sv_newmortal();
9379 SV * const mysv_val2=sv_newmortal();
9380 DEBUG_PARSE_MSG("lsbr");
9381 regprop(RExC_rx, mysv_val1, lastbr);
9382 regprop(RExC_rx, mysv_val2, ender);
9383 PerlIO_printf(Perl_debug_log, "~ tying lastbr %s (%"IVdf") to ender %s (%"IVdf") offset %"IVdf"\n",
9384 SvPV_nolen_const(mysv_val1),
9385 (IV)REG_NODE_NUM(lastbr),
9386 SvPV_nolen_const(mysv_val2),
9387 (IV)REG_NODE_NUM(ender),
9388 (IV)(ender - lastbr)
9391 REGTAIL(pRExC_state, lastbr, ender);
9393 if (have_branch && !SIZE_ONLY) {
9396 RExC_seen |= REG_TOP_LEVEL_BRANCHES;
9398 /* Hook the tails of the branches to the closing node. */
9399 for (br = ret; br; br = regnext(br)) {
9400 const U8 op = PL_regkind[OP(br)];
9402 REGTAIL_STUDY(pRExC_state, NEXTOPER(br), ender);
9403 if (OP(NEXTOPER(br)) != NOTHING || regnext(NEXTOPER(br)) != ender)
9406 else if (op == BRANCHJ) {
9407 REGTAIL_STUDY(pRExC_state, NEXTOPER(NEXTOPER(br)), ender);
9408 /* for now we always disable this optimisation * /
9409 if (OP(NEXTOPER(NEXTOPER(br))) != NOTHING || regnext(NEXTOPER(NEXTOPER(br))) != ender)
9415 br= PL_regkind[OP(ret)] != BRANCH ? regnext(ret) : ret;
9416 DEBUG_PARSE_r(if (!SIZE_ONLY) {
9417 SV * const mysv_val1=sv_newmortal();
9418 SV * const mysv_val2=sv_newmortal();
9419 DEBUG_PARSE_MSG("NADA");
9420 regprop(RExC_rx, mysv_val1, ret);
9421 regprop(RExC_rx, mysv_val2, ender);
9422 PerlIO_printf(Perl_debug_log, "~ converting ret %s (%"IVdf") to ender %s (%"IVdf") offset %"IVdf"\n",
9423 SvPV_nolen_const(mysv_val1),
9424 (IV)REG_NODE_NUM(ret),
9425 SvPV_nolen_const(mysv_val2),
9426 (IV)REG_NODE_NUM(ender),
9431 if (OP(ender) == TAIL) {
9436 for ( opt= br + 1; opt < ender ; opt++ )
9438 NEXT_OFF(br)= ender - br;
9446 static const char parens[] = "=!<,>";
9448 if (paren && (p = strchr(parens, paren))) {
9449 U8 node = ((p - parens) % 2) ? UNLESSM : IFMATCH;
9450 int flag = (p - parens) > 1;
9453 node = SUSPEND, flag = 0;
9454 reginsert(pRExC_state, node,ret, depth+1);
9455 Set_Node_Cur_Length(ret);
9456 Set_Node_Offset(ret, parse_start + 1);
9458 REGTAIL_STUDY(pRExC_state, ret, reg_node(pRExC_state, TAIL));
9462 /* Check for proper termination. */
9464 /* restore original flags, but keep (?p) */
9465 RExC_flags = oregflags | (RExC_flags & RXf_PMf_KEEPCOPY);
9466 if (RExC_parse >= RExC_end || *nextchar(pRExC_state) != ')') {
9467 RExC_parse = oregcomp_parse;
9468 vFAIL("Unmatched (");
9471 else if (!paren && RExC_parse < RExC_end) {
9472 if (*RExC_parse == ')') {
9474 vFAIL("Unmatched )");
9477 FAIL("Junk on end of regexp"); /* "Can't happen". */
9478 assert(0); /* NOTREACHED */
9481 if (RExC_in_lookbehind) {
9482 RExC_in_lookbehind--;
9484 if (after_freeze > RExC_npar)
9485 RExC_npar = after_freeze;
9490 - regbranch - one alternative of an | operator
9492 * Implements the concatenation operator.
9494 * Returns NULL, setting *flagp to RESTART_UTF8 if the sizing scan needs to be
9498 S_regbranch(pTHX_ RExC_state_t *pRExC_state, I32 *flagp, I32 first, U32 depth)
9502 regnode *chain = NULL;
9504 I32 flags = 0, c = 0;
9505 GET_RE_DEBUG_FLAGS_DECL;
9507 PERL_ARGS_ASSERT_REGBRANCH;
9509 DEBUG_PARSE("brnc");
9514 if (!SIZE_ONLY && RExC_extralen)
9515 ret = reganode(pRExC_state, BRANCHJ,0);
9517 ret = reg_node(pRExC_state, BRANCH);
9518 Set_Node_Length(ret, 1);
9522 if (!first && SIZE_ONLY)
9523 RExC_extralen += 1; /* BRANCHJ */
9525 *flagp = WORST; /* Tentatively. */
9528 nextchar(pRExC_state);
9529 while (RExC_parse < RExC_end && *RExC_parse != '|' && *RExC_parse != ')') {
9531 latest = regpiece(pRExC_state, &flags,depth+1);
9532 if (latest == NULL) {
9533 if (flags & TRYAGAIN)
9535 if (flags & RESTART_UTF8) {
9536 *flagp = RESTART_UTF8;
9539 FAIL2("panic: regpiece returned NULL, flags=%#X", flags);
9541 else if (ret == NULL)
9543 *flagp |= flags&(HASWIDTH|POSTPONED);
9544 if (chain == NULL) /* First piece. */
9545 *flagp |= flags&SPSTART;
9548 REGTAIL(pRExC_state, chain, latest);
9553 if (chain == NULL) { /* Loop ran zero times. */
9554 chain = reg_node(pRExC_state, NOTHING);
9559 *flagp |= flags&SIMPLE;
9566 - regpiece - something followed by possible [*+?]
9568 * Note that the branching code sequences used for ? and the general cases
9569 * of * and + are somewhat optimized: they use the same NOTHING node as
9570 * both the endmarker for their branch list and the body of the last branch.
9571 * It might seem that this node could be dispensed with entirely, but the
9572 * endmarker role is not redundant.
9574 * Returns NULL, setting *flagp to TRYAGAIN if regatom() returns NULL with
9576 * Returns NULL, setting *flagp to RESTART_UTF8 if the sizing scan needs to be
9580 S_regpiece(pTHX_ RExC_state_t *pRExC_state, I32 *flagp, U32 depth)
9587 const char * const origparse = RExC_parse;
9589 I32 max = REG_INFTY;
9590 #ifdef RE_TRACK_PATTERN_OFFSETS
9593 const char *maxpos = NULL;
9595 /* Save the original in case we change the emitted regop to a FAIL. */
9596 regnode * const orig_emit = RExC_emit;
9598 GET_RE_DEBUG_FLAGS_DECL;
9600 PERL_ARGS_ASSERT_REGPIECE;
9602 DEBUG_PARSE("piec");
9604 ret = regatom(pRExC_state, &flags,depth+1);
9606 if (flags & (TRYAGAIN|RESTART_UTF8))
9607 *flagp |= flags & (TRYAGAIN|RESTART_UTF8);
9609 FAIL2("panic: regatom returned NULL, flags=%#X", flags);
9615 if (op == '{' && regcurly(RExC_parse, FALSE)) {
9617 #ifdef RE_TRACK_PATTERN_OFFSETS
9618 parse_start = RExC_parse; /* MJD */
9620 next = RExC_parse + 1;
9621 while (isDIGIT(*next) || *next == ',') {
9630 if (*next == '}') { /* got one */
9634 min = atoi(RExC_parse);
9638 maxpos = RExC_parse;
9640 if (!max && *maxpos != '0')
9641 max = REG_INFTY; /* meaning "infinity" */
9642 else if (max >= REG_INFTY)
9643 vFAIL2("Quantifier in {,} bigger than %d", REG_INFTY - 1);
9645 nextchar(pRExC_state);
9646 if (max < min) { /* If can't match, warn and optimize to fail
9649 ckWARNreg(RExC_parse, "Quantifier {n,m} with n > m can't match");
9651 /* We can't back off the size because we have to reserve
9652 * enough space for all the things we are about to throw
9653 * away, but we can shrink it by the ammount we are about
9655 RExC_size = PREVOPER(RExC_size) - regarglen[(U8)OPFAIL];
9658 RExC_emit = orig_emit;
9660 ret = reg_node(pRExC_state, OPFAIL);
9663 else if (max == 0) { /* replace {0} with a nothing node */
9665 RExC_size = PREVOPER(RExC_size) - regarglen[(U8)NOTHING];
9668 RExC_emit = orig_emit;
9670 ret = reg_node(pRExC_state, NOTHING);
9675 if ((flags&SIMPLE)) {
9676 RExC_naughty += 2 + RExC_naughty / 2;
9677 reginsert(pRExC_state, CURLY, ret, depth+1);
9678 Set_Node_Offset(ret, parse_start+1); /* MJD */
9679 Set_Node_Cur_Length(ret);
9682 regnode * const w = reg_node(pRExC_state, WHILEM);
9685 REGTAIL(pRExC_state, ret, w);
9686 if (!SIZE_ONLY && RExC_extralen) {
9687 reginsert(pRExC_state, LONGJMP,ret, depth+1);
9688 reginsert(pRExC_state, NOTHING,ret, depth+1);
9689 NEXT_OFF(ret) = 3; /* Go over LONGJMP. */
9691 reginsert(pRExC_state, CURLYX,ret, depth+1);
9693 Set_Node_Offset(ret, parse_start+1);
9694 Set_Node_Length(ret,
9695 op == '{' ? (RExC_parse - parse_start) : 1);
9697 if (!SIZE_ONLY && RExC_extralen)
9698 NEXT_OFF(ret) = 3; /* Go over NOTHING to LONGJMP. */
9699 REGTAIL(pRExC_state, ret, reg_node(pRExC_state, NOTHING));
9701 RExC_whilem_seen++, RExC_extralen += 3;
9702 RExC_naughty += 4 + RExC_naughty; /* compound interest */
9711 ARG1_SET(ret, (U16)min);
9712 ARG2_SET(ret, (U16)max);
9724 #if 0 /* Now runtime fix should be reliable. */
9726 /* if this is reinstated, don't forget to put this back into perldiag:
9728 =item Regexp *+ operand could be empty at {#} in regex m/%s/
9730 (F) The part of the regexp subject to either the * or + quantifier
9731 could match an empty string. The {#} shows in the regular
9732 expression about where the problem was discovered.
9736 if (!(flags&HASWIDTH) && op != '?')
9737 vFAIL("Regexp *+ operand could be empty");
9740 #ifdef RE_TRACK_PATTERN_OFFSETS
9741 parse_start = RExC_parse;
9743 nextchar(pRExC_state);
9745 *flagp = (op != '+') ? (WORST|SPSTART|HASWIDTH) : (WORST|HASWIDTH);
9747 if (op == '*' && (flags&SIMPLE)) {
9748 reginsert(pRExC_state, STAR, ret, depth+1);
9752 else if (op == '*') {
9756 else if (op == '+' && (flags&SIMPLE)) {
9757 reginsert(pRExC_state, PLUS, ret, depth+1);
9761 else if (op == '+') {
9765 else if (op == '?') {
9770 if (!SIZE_ONLY && !(flags&(HASWIDTH|POSTPONED)) && max > REG_INFTY/3) {
9771 SAVEFREESV(RExC_rx_sv); /* in case of fatal warnings */
9772 ckWARN3reg(RExC_parse,
9773 "%.*s matches null string many times",
9774 (int)(RExC_parse >= origparse ? RExC_parse - origparse : 0),
9776 (void)ReREFCNT_inc(RExC_rx_sv);
9779 if (RExC_parse < RExC_end && *RExC_parse == '?') {
9780 nextchar(pRExC_state);
9781 reginsert(pRExC_state, MINMOD, ret, depth+1);
9782 REGTAIL(pRExC_state, ret, ret + NODE_STEP_REGNODE);
9784 #ifndef REG_ALLOW_MINMOD_SUSPEND
9787 if (RExC_parse < RExC_end && *RExC_parse == '+') {
9789 nextchar(pRExC_state);
9790 ender = reg_node(pRExC_state, SUCCEED);
9791 REGTAIL(pRExC_state, ret, ender);
9792 reginsert(pRExC_state, SUSPEND, ret, depth+1);
9794 ender = reg_node(pRExC_state, TAIL);
9795 REGTAIL(pRExC_state, ret, ender);
9799 if (RExC_parse < RExC_end && ISMULT2(RExC_parse)) {
9801 vFAIL("Nested quantifiers");
9808 S_grok_bslash_N(pTHX_ RExC_state_t *pRExC_state, regnode** node_p, UV *valuep, I32 *flagp, U32 depth, bool in_char_class,
9809 const bool strict /* Apply stricter parsing rules? */
9813 /* This is expected to be called by a parser routine that has recognized '\N'
9814 and needs to handle the rest. RExC_parse is expected to point at the first
9815 char following the N at the time of the call. On successful return,
9816 RExC_parse has been updated to point to just after the sequence identified
9817 by this routine, and <*flagp> has been updated.
9819 The \N may be inside (indicated by the boolean <in_char_class>) or outside a
9822 \N may begin either a named sequence, or if outside a character class, mean
9823 to match a non-newline. For non single-quoted regexes, the tokenizer has
9824 attempted to decide which, and in the case of a named sequence, converted it
9825 into one of the forms: \N{} (if the sequence is null), or \N{U+c1.c2...},
9826 where c1... are the characters in the sequence. For single-quoted regexes,
9827 the tokenizer passes the \N sequence through unchanged; this code will not
9828 attempt to determine this nor expand those, instead raising a syntax error.
9829 The net effect is that if the beginning of the passed-in pattern isn't '{U+'
9830 or there is no '}', it signals that this \N occurrence means to match a
9833 Only the \N{U+...} form should occur in a character class, for the same
9834 reason that '.' inside a character class means to just match a period: it
9835 just doesn't make sense.
9837 The function raises an error (via vFAIL), and doesn't return for various
9838 syntax errors. Otherwise it returns TRUE and sets <node_p> or <valuep> on
9839 success; it returns FALSE otherwise. Returns FALSE, setting *flagp to
9840 RESTART_UTF8 if the sizing scan needs to be restarted. Such a restart is
9841 only possible if node_p is non-NULL.
9844 If <valuep> is non-null, it means the caller can accept an input sequence
9845 consisting of a just a single code point; <*valuep> is set to that value
9846 if the input is such.
9848 If <node_p> is non-null it signifies that the caller can accept any other
9849 legal sequence (i.e., one that isn't just a single code point). <*node_p>
9851 1) \N means not-a-NL: points to a newly created REG_ANY node;
9852 2) \N{}: points to a new NOTHING node;
9853 3) otherwise: points to a new EXACT node containing the resolved
9855 Note that FALSE is returned for single code point sequences if <valuep> is
9859 char * endbrace; /* '}' following the name */
9861 char *endchar; /* Points to '.' or '}' ending cur char in the input
9863 bool has_multiple_chars; /* true if the input stream contains a sequence of
9864 more than one character */
9866 GET_RE_DEBUG_FLAGS_DECL;
9868 PERL_ARGS_ASSERT_GROK_BSLASH_N;
9872 assert(cBOOL(node_p) ^ cBOOL(valuep)); /* Exactly one should be set */
9874 /* The [^\n] meaning of \N ignores spaces and comments under the /x
9875 * modifier. The other meaning does not */
9876 p = (RExC_flags & RXf_PMf_EXTENDED)
9877 ? regwhite( pRExC_state, RExC_parse )
9880 /* Disambiguate between \N meaning a named character versus \N meaning
9881 * [^\n]. The former is assumed when it can't be the latter. */
9882 if (*p != '{' || regcurly(p, FALSE)) {
9885 /* no bare \N in a charclass */
9886 if (in_char_class) {
9887 vFAIL("\\N in a character class must be a named character: \\N{...}");
9891 nextchar(pRExC_state);
9892 *node_p = reg_node(pRExC_state, REG_ANY);
9893 *flagp |= HASWIDTH|SIMPLE;
9896 Set_Node_Length(*node_p, 1); /* MJD */
9900 /* Here, we have decided it should be a named character or sequence */
9902 /* The test above made sure that the next real character is a '{', but
9903 * under the /x modifier, it could be separated by space (or a comment and
9904 * \n) and this is not allowed (for consistency with \x{...} and the
9905 * tokenizer handling of \N{NAME}). */
9906 if (*RExC_parse != '{') {
9907 vFAIL("Missing braces on \\N{}");
9910 RExC_parse++; /* Skip past the '{' */
9912 if (! (endbrace = strchr(RExC_parse, '}')) /* no trailing brace */
9913 || ! (endbrace == RExC_parse /* nothing between the {} */
9914 || (endbrace - RExC_parse >= 2 /* U+ (bad hex is checked below */
9915 && strnEQ(RExC_parse, "U+", 2)))) /* for a better error msg) */
9917 if (endbrace) RExC_parse = endbrace; /* position msg's '<--HERE' */
9918 vFAIL("\\N{NAME} must be resolved by the lexer");
9921 if (endbrace == RExC_parse) { /* empty: \N{} */
9924 *node_p = reg_node(pRExC_state,NOTHING);
9926 else if (in_char_class) {
9927 if (SIZE_ONLY && in_char_class) {
9929 RExC_parse++; /* Position after the "}" */
9930 vFAIL("Zero length \\N{}");
9933 ckWARNreg(RExC_parse,
9934 "Ignoring zero length \\N{} in character class");
9942 nextchar(pRExC_state);
9946 RExC_uni_semantics = 1; /* Unicode named chars imply Unicode semantics */
9947 RExC_parse += 2; /* Skip past the 'U+' */
9949 endchar = RExC_parse + strcspn(RExC_parse, ".}");
9951 /* Code points are separated by dots. If none, there is only one code
9952 * point, and is terminated by the brace */
9953 has_multiple_chars = (endchar < endbrace);
9955 if (valuep && (! has_multiple_chars || in_char_class)) {
9956 /* We only pay attention to the first char of
9957 multichar strings being returned in char classes. I kinda wonder
9958 if this makes sense as it does change the behaviour
9959 from earlier versions, OTOH that behaviour was broken
9960 as well. XXX Solution is to recharacterize as
9961 [rest-of-class]|multi1|multi2... */
9963 STRLEN length_of_hex = (STRLEN)(endchar - RExC_parse);
9964 I32 grok_hex_flags = PERL_SCAN_ALLOW_UNDERSCORES
9965 | PERL_SCAN_DISALLOW_PREFIX
9966 | (SIZE_ONLY ? PERL_SCAN_SILENT_ILLDIGIT : 0);
9968 *valuep = grok_hex(RExC_parse, &length_of_hex, &grok_hex_flags, NULL);
9970 /* The tokenizer should have guaranteed validity, but it's possible to
9971 * bypass it by using single quoting, so check */
9972 if (length_of_hex == 0
9973 || length_of_hex != (STRLEN)(endchar - RExC_parse) )
9975 RExC_parse += length_of_hex; /* Includes all the valid */
9976 RExC_parse += (RExC_orig_utf8) /* point to after 1st invalid */
9977 ? UTF8SKIP(RExC_parse)
9979 /* Guard against malformed utf8 */
9980 if (RExC_parse >= endchar) {
9981 RExC_parse = endchar;
9983 vFAIL("Invalid hexadecimal number in \\N{U+...}");
9986 if (in_char_class && has_multiple_chars) {
9988 RExC_parse = endbrace;
9989 vFAIL("\\N{} in character class restricted to one character");
9992 ckWARNreg(endchar, "Using just the first character returned by \\N{} in character class");
9996 RExC_parse = endbrace + 1;
9998 else if (! node_p || ! has_multiple_chars) {
10000 /* Here, the input is legal, but not according to the caller's
10001 * options. We fail without advancing the parse, so that the
10002 * caller can try again */
10008 /* What is done here is to convert this to a sub-pattern of the form
10009 * (?:\x{char1}\x{char2}...)
10010 * and then call reg recursively. That way, it retains its atomicness,
10011 * while not having to worry about special handling that some code
10012 * points may have. toke.c has converted the original Unicode values
10013 * to native, so that we can just pass on the hex values unchanged. We
10014 * do have to set a flag to keep recoding from happening in the
10017 SV * substitute_parse = newSVpvn_flags("?:", 2, SVf_UTF8|SVs_TEMP);
10019 char *orig_end = RExC_end;
10022 while (RExC_parse < endbrace) {
10024 /* Convert to notation the rest of the code understands */
10025 sv_catpv(substitute_parse, "\\x{");
10026 sv_catpvn(substitute_parse, RExC_parse, endchar - RExC_parse);
10027 sv_catpv(substitute_parse, "}");
10029 /* Point to the beginning of the next character in the sequence. */
10030 RExC_parse = endchar + 1;
10031 endchar = RExC_parse + strcspn(RExC_parse, ".}");
10033 sv_catpv(substitute_parse, ")");
10035 RExC_parse = SvPV(substitute_parse, len);
10037 /* Don't allow empty number */
10039 vFAIL("Invalid hexadecimal number in \\N{U+...}");
10041 RExC_end = RExC_parse + len;
10043 /* The values are Unicode, and therefore not subject to recoding */
10044 RExC_override_recoding = 1;
10046 if (!(*node_p = reg(pRExC_state, 1, &flags, depth+1))) {
10047 if (flags & RESTART_UTF8) {
10048 *flagp = RESTART_UTF8;
10051 FAIL2("panic: reg returned NULL to grok_bslash_N, flags=%#X",
10054 *flagp |= flags&(HASWIDTH|SPSTART|SIMPLE|POSTPONED);
10056 RExC_parse = endbrace;
10057 RExC_end = orig_end;
10058 RExC_override_recoding = 0;
10060 nextchar(pRExC_state);
10070 * It returns the code point in utf8 for the value in *encp.
10071 * value: a code value in the source encoding
10072 * encp: a pointer to an Encode object
10074 * If the result from Encode is not a single character,
10075 * it returns U+FFFD (Replacement character) and sets *encp to NULL.
10078 S_reg_recode(pTHX_ const char value, SV **encp)
10081 SV * const sv = newSVpvn_flags(&value, numlen, SVs_TEMP);
10082 const char * const s = *encp ? sv_recode_to_utf8(sv, *encp) : SvPVX(sv);
10083 const STRLEN newlen = SvCUR(sv);
10084 UV uv = UNICODE_REPLACEMENT;
10086 PERL_ARGS_ASSERT_REG_RECODE;
10090 ? utf8n_to_uvchr((U8*)s, newlen, &numlen, UTF8_ALLOW_DEFAULT)
10093 if (!newlen || numlen != newlen) {
10094 uv = UNICODE_REPLACEMENT;
10100 PERL_STATIC_INLINE U8
10101 S_compute_EXACTish(pTHX_ RExC_state_t *pRExC_state)
10105 PERL_ARGS_ASSERT_COMPUTE_EXACTISH;
10111 op = get_regex_charset(RExC_flags);
10112 if (op >= REGEX_ASCII_RESTRICTED_CHARSET) {
10113 op--; /* /a is same as /u, and map /aa's offset to what /a's would have
10114 been, so there is no hole */
10117 return op + EXACTF;
10120 PERL_STATIC_INLINE void
10121 S_alloc_maybe_populate_EXACT(pTHX_ RExC_state_t *pRExC_state, regnode *node, I32* flagp, STRLEN len, UV code_point)
10123 /* This knows the details about sizing an EXACTish node, setting flags for
10124 * it (by setting <*flagp>, and potentially populating it with a single
10127 * If <len> (the length in bytes) is non-zero, this function assumes that
10128 * the node has already been populated, and just does the sizing. In this
10129 * case <code_point> should be the final code point that has already been
10130 * placed into the node. This value will be ignored except that under some
10131 * circumstances <*flagp> is set based on it.
10133 * If <len> is zero, the function assumes that the node is to contain only
10134 * the single character given by <code_point> and calculates what <len>
10135 * should be. In pass 1, it sizes the node appropriately. In pass 2, it
10136 * additionally will populate the node's STRING with <code_point>, if <len>
10137 * is 0. In both cases <*flagp> is appropriately set
10139 * It knows that under FOLD, the Latin Sharp S and UTF characters above
10140 * 255, must be folded (the former only when the rules indicate it can
10143 bool len_passed_in = cBOOL(len != 0);
10144 U8 character[UTF8_MAXBYTES_CASE+1];
10146 PERL_ARGS_ASSERT_ALLOC_MAYBE_POPULATE_EXACT;
10148 if (! len_passed_in) {
10150 if (FOLD && (! LOC || code_point > 255)) {
10151 _to_uni_fold_flags(NATIVE_TO_UNI(code_point),
10154 FOLD_FLAGS_FULL | ((LOC)
10155 ? FOLD_FLAGS_LOCALE
10156 : (ASCII_FOLD_RESTRICTED)
10157 ? FOLD_FLAGS_NOMIX_ASCII
10161 uvchr_to_utf8( character, code_point);
10162 len = UTF8SKIP(character);
10166 || code_point != LATIN_SMALL_LETTER_SHARP_S
10167 || ASCII_FOLD_RESTRICTED
10168 || ! AT_LEAST_UNI_SEMANTICS)
10170 *character = (U8) code_point;
10175 *(character + 1) = 's';
10181 RExC_size += STR_SZ(len);
10184 RExC_emit += STR_SZ(len);
10185 STR_LEN(node) = len;
10186 if (! len_passed_in) {
10187 Copy((char *) character, STRING(node), len, char);
10191 *flagp |= HASWIDTH;
10193 /* A single character node is SIMPLE, except for the special-cased SHARP S
10195 if ((len == 1 || (UTF && len == UNISKIP(code_point)))
10196 && (code_point != LATIN_SMALL_LETTER_SHARP_S
10197 || ! FOLD || ! DEPENDS_SEMANTICS))
10204 - regatom - the lowest level
10206 Try to identify anything special at the start of the pattern. If there
10207 is, then handle it as required. This may involve generating a single regop,
10208 such as for an assertion; or it may involve recursing, such as to
10209 handle a () structure.
10211 If the string doesn't start with something special then we gobble up
10212 as much literal text as we can.
10214 Once we have been able to handle whatever type of thing started the
10215 sequence, we return.
10217 Note: we have to be careful with escapes, as they can be both literal
10218 and special, and in the case of \10 and friends, context determines which.
10220 A summary of the code structure is:
10222 switch (first_byte) {
10223 cases for each special:
10224 handle this special;
10227 switch (2nd byte) {
10228 cases for each unambiguous special:
10229 handle this special;
10231 cases for each ambigous special/literal:
10233 if (special) handle here
10235 default: // unambiguously literal:
10238 default: // is a literal char
10241 create EXACTish node for literal;
10242 while (more input and node isn't full) {
10243 switch (input_byte) {
10244 cases for each special;
10245 make sure parse pointer is set so that the next call to
10246 regatom will see this special first
10247 goto loopdone; // EXACTish node terminated by prev. char
10249 append char to EXACTISH node;
10251 get next input byte;
10255 return the generated node;
10257 Specifically there are two separate switches for handling
10258 escape sequences, with the one for handling literal escapes requiring
10259 a dummy entry for all of the special escapes that are actually handled
10262 Returns NULL, setting *flagp to TRYAGAIN if reg() returns NULL with
10264 Returns NULL, setting *flagp to RESTART_UTF8 if the sizing scan needs to be
10266 Otherwise does not return NULL.
10270 S_regatom(pTHX_ RExC_state_t *pRExC_state, I32 *flagp, U32 depth)
10273 regnode *ret = NULL;
10275 char *parse_start = RExC_parse;
10279 GET_RE_DEBUG_FLAGS_DECL;
10281 *flagp = WORST; /* Tentatively. */
10283 DEBUG_PARSE("atom");
10285 PERL_ARGS_ASSERT_REGATOM;
10288 switch ((U8)*RExC_parse) {
10290 RExC_seen_zerolen++;
10291 nextchar(pRExC_state);
10292 if (RExC_flags & RXf_PMf_MULTILINE)
10293 ret = reg_node(pRExC_state, MBOL);
10294 else if (RExC_flags & RXf_PMf_SINGLELINE)
10295 ret = reg_node(pRExC_state, SBOL);
10297 ret = reg_node(pRExC_state, BOL);
10298 Set_Node_Length(ret, 1); /* MJD */
10301 nextchar(pRExC_state);
10303 RExC_seen_zerolen++;
10304 if (RExC_flags & RXf_PMf_MULTILINE)
10305 ret = reg_node(pRExC_state, MEOL);
10306 else if (RExC_flags & RXf_PMf_SINGLELINE)
10307 ret = reg_node(pRExC_state, SEOL);
10309 ret = reg_node(pRExC_state, EOL);
10310 Set_Node_Length(ret, 1); /* MJD */
10313 nextchar(pRExC_state);
10314 if (RExC_flags & RXf_PMf_SINGLELINE)
10315 ret = reg_node(pRExC_state, SANY);
10317 ret = reg_node(pRExC_state, REG_ANY);
10318 *flagp |= HASWIDTH|SIMPLE;
10320 Set_Node_Length(ret, 1); /* MJD */
10324 char * const oregcomp_parse = ++RExC_parse;
10325 ret = regclass(pRExC_state, flagp,depth+1,
10326 FALSE, /* means parse the whole char class */
10327 TRUE, /* allow multi-char folds */
10328 FALSE, /* don't silence non-portable warnings. */
10330 if (*RExC_parse != ']') {
10331 RExC_parse = oregcomp_parse;
10332 vFAIL("Unmatched [");
10335 if (*flagp & RESTART_UTF8)
10337 FAIL2("panic: regclass returned NULL to regatom, flags=%#X",
10340 nextchar(pRExC_state);
10341 Set_Node_Length(ret, RExC_parse - oregcomp_parse + 1); /* MJD */
10345 nextchar(pRExC_state);
10346 ret = reg(pRExC_state, 2, &flags,depth+1);
10348 if (flags & TRYAGAIN) {
10349 if (RExC_parse == RExC_end) {
10350 /* Make parent create an empty node if needed. */
10351 *flagp |= TRYAGAIN;
10356 if (flags & RESTART_UTF8) {
10357 *flagp = RESTART_UTF8;
10360 FAIL2("panic: reg returned NULL to regatom, flags=%#X", flags);
10362 *flagp |= flags&(HASWIDTH|SPSTART|SIMPLE|POSTPONED);
10366 if (flags & TRYAGAIN) {
10367 *flagp |= TRYAGAIN;
10370 vFAIL("Internal urp");
10371 /* Supposed to be caught earlier. */
10374 if (!regcurly(RExC_parse, FALSE)) {
10383 vFAIL("Quantifier follows nothing");
10388 This switch handles escape sequences that resolve to some kind
10389 of special regop and not to literal text. Escape sequnces that
10390 resolve to literal text are handled below in the switch marked
10393 Every entry in this switch *must* have a corresponding entry
10394 in the literal escape switch. However, the opposite is not
10395 required, as the default for this switch is to jump to the
10396 literal text handling code.
10398 switch ((U8)*++RExC_parse) {
10400 /* Special Escapes */
10402 RExC_seen_zerolen++;
10403 ret = reg_node(pRExC_state, SBOL);
10405 goto finish_meta_pat;
10407 ret = reg_node(pRExC_state, GPOS);
10408 RExC_seen |= REG_SEEN_GPOS;
10410 goto finish_meta_pat;
10412 RExC_seen_zerolen++;
10413 ret = reg_node(pRExC_state, KEEPS);
10415 /* XXX:dmq : disabling in-place substitution seems to
10416 * be necessary here to avoid cases of memory corruption, as
10417 * with: C<$_="x" x 80; s/x\K/y/> -- rgs
10419 RExC_seen |= REG_SEEN_LOOKBEHIND;
10420 goto finish_meta_pat;
10422 ret = reg_node(pRExC_state, SEOL);
10424 RExC_seen_zerolen++; /* Do not optimize RE away */
10425 goto finish_meta_pat;
10427 ret = reg_node(pRExC_state, EOS);
10429 RExC_seen_zerolen++; /* Do not optimize RE away */
10430 goto finish_meta_pat;
10432 ret = reg_node(pRExC_state, CANY);
10433 RExC_seen |= REG_SEEN_CANY;
10434 *flagp |= HASWIDTH|SIMPLE;
10435 goto finish_meta_pat;
10437 ret = reg_node(pRExC_state, CLUMP);
10438 *flagp |= HASWIDTH;
10439 goto finish_meta_pat;
10445 arg = ANYOF_WORDCHAR;
10449 RExC_seen_zerolen++;
10450 RExC_seen |= REG_SEEN_LOOKBEHIND;
10451 op = BOUND + get_regex_charset(RExC_flags);
10452 if (op > BOUNDA) { /* /aa is same as /a */
10455 ret = reg_node(pRExC_state, op);
10456 FLAGS(ret) = get_regex_charset(RExC_flags);
10458 if (! SIZE_ONLY && (U8) *(RExC_parse + 1) == '{') {
10459 ckWARNdep(RExC_parse, "\"\\b{\" is deprecated; use \"\\b\\{\" or \"\\b[{]\" instead");
10461 goto finish_meta_pat;
10463 RExC_seen_zerolen++;
10464 RExC_seen |= REG_SEEN_LOOKBEHIND;
10465 op = NBOUND + get_regex_charset(RExC_flags);
10466 if (op > NBOUNDA) { /* /aa is same as /a */
10469 ret = reg_node(pRExC_state, op);
10470 FLAGS(ret) = get_regex_charset(RExC_flags);
10472 if (! SIZE_ONLY && (U8) *(RExC_parse + 1) == '{') {
10473 ckWARNdep(RExC_parse, "\"\\B{\" is deprecated; use \"\\B\\{\" or \"\\B[{]\" instead");
10475 goto finish_meta_pat;
10485 ret = reg_node(pRExC_state, LNBREAK);
10486 *flagp |= HASWIDTH|SIMPLE;
10487 goto finish_meta_pat;
10495 goto join_posix_op_known;
10501 arg = ANYOF_VERTWS;
10503 goto join_posix_op_known;
10513 op = POSIXD + get_regex_charset(RExC_flags);
10514 if (op > POSIXA) { /* /aa is same as /a */
10518 join_posix_op_known:
10521 op += NPOSIXD - POSIXD;
10524 ret = reg_node(pRExC_state, op);
10526 FLAGS(ret) = namedclass_to_classnum(arg);
10529 *flagp |= HASWIDTH|SIMPLE;
10533 nextchar(pRExC_state);
10534 Set_Node_Length(ret, 2); /* MJD */
10540 char* parse_start = RExC_parse - 2;
10545 ret = regclass(pRExC_state, flagp,depth+1,
10546 TRUE, /* means just parse this element */
10547 FALSE, /* don't allow multi-char folds */
10548 FALSE, /* don't silence non-portable warnings.
10549 It would be a bug if these returned
10552 /* regclass() can only return RESTART_UTF8 if multi-char folds
10555 FAIL2("panic: regclass returned NULL to regatom, flags=%#X",
10560 Set_Node_Offset(ret, parse_start + 2);
10561 Set_Node_Cur_Length(ret);
10562 nextchar(pRExC_state);
10566 /* Handle \N and \N{NAME} with multiple code points here and not
10567 * below because it can be multicharacter. join_exact() will join
10568 * them up later on. Also this makes sure that things like
10569 * /\N{BLAH}+/ and \N{BLAH} being multi char Just Happen. dmq.
10570 * The options to the grok function call causes it to fail if the
10571 * sequence is just a single code point. We then go treat it as
10572 * just another character in the current EXACT node, and hence it
10573 * gets uniform treatment with all the other characters. The
10574 * special treatment for quantifiers is not needed for such single
10575 * character sequences */
10577 if (! grok_bslash_N(pRExC_state, &ret, NULL, flagp, depth, FALSE,
10578 FALSE /* not strict */ )) {
10579 if (*flagp & RESTART_UTF8)
10585 case 'k': /* Handle \k<NAME> and \k'NAME' */
10588 char ch= RExC_parse[1];
10589 if (ch != '<' && ch != '\'' && ch != '{') {
10591 vFAIL2("Sequence %.2s... not terminated",parse_start);
10593 /* this pretty much dupes the code for (?P=...) in reg(), if
10594 you change this make sure you change that */
10595 char* name_start = (RExC_parse += 2);
10597 SV *sv_dat = reg_scan_name(pRExC_state,
10598 SIZE_ONLY ? REG_RSN_RETURN_NULL : REG_RSN_RETURN_DATA);
10599 ch= (ch == '<') ? '>' : (ch == '{') ? '}' : '\'';
10600 if (RExC_parse == name_start || *RExC_parse != ch)
10601 vFAIL2("Sequence %.3s... not terminated",parse_start);
10604 num = add_data( pRExC_state, 1, "S" );
10605 RExC_rxi->data->data[num]=(void*)sv_dat;
10606 SvREFCNT_inc_simple_void(sv_dat);
10610 ret = reganode(pRExC_state,
10613 : (ASCII_FOLD_RESTRICTED)
10615 : (AT_LEAST_UNI_SEMANTICS)
10621 *flagp |= HASWIDTH;
10623 /* override incorrect value set in reganode MJD */
10624 Set_Node_Offset(ret, parse_start+1);
10625 Set_Node_Cur_Length(ret); /* MJD */
10626 nextchar(pRExC_state);
10632 case '1': case '2': case '3': case '4':
10633 case '5': case '6': case '7': case '8': case '9':
10636 bool isg = *RExC_parse == 'g';
10641 if (*RExC_parse == '{') {
10645 if (*RExC_parse == '-') {
10649 if (hasbrace && !isDIGIT(*RExC_parse)) {
10650 if (isrel) RExC_parse--;
10652 goto parse_named_seq;
10654 num = atoi(RExC_parse);
10655 if (isg && num == 0)
10656 vFAIL("Reference to invalid group 0");
10658 num = RExC_npar - num;
10660 vFAIL("Reference to nonexistent or unclosed group");
10662 if (!isg && num > 9 && num >= RExC_npar)
10663 /* Probably a character specified in octal, e.g. \35 */
10666 char * const parse_start = RExC_parse - 1; /* MJD */
10667 while (isDIGIT(*RExC_parse))
10669 if (parse_start == RExC_parse - 1)
10670 vFAIL("Unterminated \\g... pattern");
10672 if (*RExC_parse != '}')
10673 vFAIL("Unterminated \\g{...} pattern");
10677 if (num > (I32)RExC_rx->nparens)
10678 vFAIL("Reference to nonexistent group");
10681 ret = reganode(pRExC_state,
10684 : (ASCII_FOLD_RESTRICTED)
10686 : (AT_LEAST_UNI_SEMANTICS)
10692 *flagp |= HASWIDTH;
10694 /* override incorrect value set in reganode MJD */
10695 Set_Node_Offset(ret, parse_start+1);
10696 Set_Node_Cur_Length(ret); /* MJD */
10698 nextchar(pRExC_state);
10703 if (RExC_parse >= RExC_end)
10704 FAIL("Trailing \\");
10707 /* Do not generate "unrecognized" warnings here, we fall
10708 back into the quick-grab loop below */
10715 if (RExC_flags & RXf_PMf_EXTENDED) {
10716 if ( reg_skipcomment( pRExC_state ) )
10723 parse_start = RExC_parse - 1;
10732 #define MAX_NODE_STRING_SIZE 127
10733 char foldbuf[MAX_NODE_STRING_SIZE+UTF8_MAXBYTES_CASE];
10735 U8 upper_parse = MAX_NODE_STRING_SIZE;
10738 bool next_is_quantifier;
10739 char * oldp = NULL;
10741 /* If a folding node contains only code points that don't
10742 * participate in folds, it can be changed into an EXACT node,
10743 * which allows the optimizer more things to look for */
10747 node_type = compute_EXACTish(pRExC_state);
10748 ret = reg_node(pRExC_state, node_type);
10750 /* In pass1, folded, we use a temporary buffer instead of the
10751 * actual node, as the node doesn't exist yet */
10752 s = (SIZE_ONLY && FOLD) ? foldbuf : STRING(ret);
10758 /* We do the EXACTFish to EXACT node only if folding, and not if in
10759 * locale, as whether a character folds or not isn't known until
10761 maybe_exact = FOLD && ! LOC;
10763 /* XXX The node can hold up to 255 bytes, yet this only goes to
10764 * 127. I (khw) do not know why. Keeping it somewhat less than
10765 * 255 allows us to not have to worry about overflow due to
10766 * converting to utf8 and fold expansion, but that value is
10767 * 255-UTF8_MAXBYTES_CASE. join_exact() may join adjacent nodes
10768 * split up by this limit into a single one using the real max of
10769 * 255. Even at 127, this breaks under rare circumstances. If
10770 * folding, we do not want to split a node at a character that is a
10771 * non-final in a multi-char fold, as an input string could just
10772 * happen to want to match across the node boundary. The join
10773 * would solve that problem if the join actually happens. But a
10774 * series of more than two nodes in a row each of 127 would cause
10775 * the first join to succeed to get to 254, but then there wouldn't
10776 * be room for the next one, which could at be one of those split
10777 * multi-char folds. I don't know of any fool-proof solution. One
10778 * could back off to end with only a code point that isn't such a
10779 * non-final, but it is possible for there not to be any in the
10781 for (p = RExC_parse - 1;
10782 len < upper_parse && p < RExC_end;
10787 if (RExC_flags & RXf_PMf_EXTENDED)
10788 p = regwhite( pRExC_state, p );
10799 /* Literal Escapes Switch
10801 This switch is meant to handle escape sequences that
10802 resolve to a literal character.
10804 Every escape sequence that represents something
10805 else, like an assertion or a char class, is handled
10806 in the switch marked 'Special Escapes' above in this
10807 routine, but also has an entry here as anything that
10808 isn't explicitly mentioned here will be treated as
10809 an unescaped equivalent literal.
10812 switch ((U8)*++p) {
10813 /* These are all the special escapes. */
10814 case 'A': /* Start assertion */
10815 case 'b': case 'B': /* Word-boundary assertion*/
10816 case 'C': /* Single char !DANGEROUS! */
10817 case 'd': case 'D': /* digit class */
10818 case 'g': case 'G': /* generic-backref, pos assertion */
10819 case 'h': case 'H': /* HORIZWS */
10820 case 'k': case 'K': /* named backref, keep marker */
10821 case 'p': case 'P': /* Unicode property */
10822 case 'R': /* LNBREAK */
10823 case 's': case 'S': /* space class */
10824 case 'v': case 'V': /* VERTWS */
10825 case 'w': case 'W': /* word class */
10826 case 'X': /* eXtended Unicode "combining character sequence" */
10827 case 'z': case 'Z': /* End of line/string assertion */
10831 /* Anything after here is an escape that resolves to a
10832 literal. (Except digits, which may or may not)
10838 case 'N': /* Handle a single-code point named character. */
10839 /* The options cause it to fail if a multiple code
10840 * point sequence. Handle those in the switch() above
10842 RExC_parse = p + 1;
10843 if (! grok_bslash_N(pRExC_state, NULL, &ender,
10844 flagp, depth, FALSE,
10845 FALSE /* not strict */ ))
10847 if (*flagp & RESTART_UTF8)
10848 FAIL("panic: grok_bslash_N set RESTART_UTF8");
10849 RExC_parse = p = oldp;
10853 if (ender > 0xff) {
10870 ender = ASCII_TO_NATIVE('\033');
10874 ender = ASCII_TO_NATIVE('\007');
10880 const char* error_msg;
10882 bool valid = grok_bslash_o(&p,
10885 TRUE, /* out warnings */
10886 FALSE, /* not strict */
10887 TRUE, /* Output warnings
10892 RExC_parse = p; /* going to die anyway; point
10893 to exact spot of failure */
10897 if (PL_encoding && ender < 0x100) {
10898 goto recode_encoding;
10900 if (ender > 0xff) {
10907 UV result = UV_MAX; /* initialize to erroneous
10909 const char* error_msg;
10911 bool valid = grok_bslash_x(&p,
10914 TRUE, /* out warnings */
10915 FALSE, /* not strict */
10916 TRUE, /* Output warnings
10921 RExC_parse = p; /* going to die anyway; point
10922 to exact spot of failure */
10927 if (PL_encoding && ender < 0x100) {
10928 goto recode_encoding;
10930 if (ender > 0xff) {
10937 ender = grok_bslash_c(*p++, UTF, SIZE_ONLY);
10939 case '0': case '1': case '2': case '3':case '4':
10940 case '5': case '6': case '7':
10942 (isDIGIT(p[1]) && atoi(p) >= RExC_npar))
10944 I32 flags = PERL_SCAN_SILENT_ILLDIGIT;
10946 ender = grok_oct(p, &numlen, &flags, NULL);
10947 if (ender > 0xff) {
10951 if (SIZE_ONLY /* like \08, \178 */
10954 && isDIGIT(*p) && ckWARN(WARN_REGEXP))
10956 reg_warn_non_literal_string(
10958 form_short_octal_warning(p, numlen));
10961 else { /* Not to be treated as an octal constant, go
10966 if (PL_encoding && ender < 0x100)
10967 goto recode_encoding;
10970 if (! RExC_override_recoding) {
10971 SV* enc = PL_encoding;
10972 ender = reg_recode((const char)(U8)ender, &enc);
10973 if (!enc && SIZE_ONLY)
10974 ckWARNreg(p, "Invalid escape in the specified encoding");
10980 FAIL("Trailing \\");
10983 if (!SIZE_ONLY&& isALPHANUMERIC(*p)) {
10984 /* Include any { following the alpha to emphasize
10985 * that it could be part of an escape at some point
10987 int len = (isALPHA(*p) && *(p + 1) == '{') ? 2 : 1;
10988 ckWARN3reg(p + len, "Unrecognized escape \\%.*s passed through", len, p);
10990 goto normal_default;
10991 } /* End of switch on '\' */
10993 default: /* A literal character */
10996 && RExC_flags & RXf_PMf_EXTENDED
10997 && ckWARN(WARN_DEPRECATED)
10998 && is_PATWS_non_low(p, UTF))
11000 vWARN_dep(p + ((UTF) ? UTF8SKIP(p) : 1),
11001 "Escape literal pattern white space under /x");
11005 if (UTF8_IS_START(*p) && UTF) {
11007 ender = utf8n_to_uvchr((U8*)p, RExC_end - p,
11008 &numlen, UTF8_ALLOW_DEFAULT);
11014 } /* End of switch on the literal */
11016 /* Here, have looked at the literal character and <ender>
11017 * contains its ordinal, <p> points to the character after it
11020 if ( RExC_flags & RXf_PMf_EXTENDED)
11021 p = regwhite( pRExC_state, p );
11023 /* If the next thing is a quantifier, it applies to this
11024 * character only, which means that this character has to be in
11025 * its own node and can't just be appended to the string in an
11026 * existing node, so if there are already other characters in
11027 * the node, close the node with just them, and set up to do
11028 * this character again next time through, when it will be the
11029 * only thing in its new node */
11030 if ((next_is_quantifier = (p < RExC_end && ISMULT2(p))) && len)
11038 /* See comments for join_exact() as to why we fold
11039 * this non-UTF at compile time */
11040 || (node_type == EXACTFU
11041 && ender == LATIN_SMALL_LETTER_SHARP_S))
11045 /* Prime the casefolded buffer. Locale rules, which
11046 * apply only to code points < 256, aren't known until
11047 * execution, so for them, just output the original
11048 * character using utf8. If we start to fold non-UTF
11049 * patterns, be sure to update join_exact() */
11050 if (LOC && ender < 256) {
11051 if (UNI_IS_INVARIANT(ender)) {
11055 *s = UTF8_TWO_BYTE_HI(ender);
11056 *(s + 1) = UTF8_TWO_BYTE_LO(ender);
11061 UV folded = _to_uni_fold_flags(
11066 | ((LOC) ? FOLD_FLAGS_LOCALE
11067 : (ASCII_FOLD_RESTRICTED)
11068 ? FOLD_FLAGS_NOMIX_ASCII
11072 /* If this node only contains non-folding code
11073 * points so far, see if this new one is also
11076 if (folded != ender) {
11077 maybe_exact = FALSE;
11080 /* Here the fold is the original; we have
11081 * to check further to see if anything
11083 if (! PL_utf8_foldable) {
11084 SV* swash = swash_init("utf8",
11086 &PL_sv_undef, 1, 0);
11088 _get_swash_invlist(swash);
11089 SvREFCNT_dec_NN(swash);
11091 if (_invlist_contains_cp(PL_utf8_foldable,
11094 maybe_exact = FALSE;
11102 /* The loop increments <len> each time, as all but this
11103 * path (and the one just below for UTF) through it add
11104 * a single byte to the EXACTish node. But this one
11105 * has changed len to be the correct final value, so
11106 * subtract one to cancel out the increment that
11108 len += foldlen - 1;
11111 *(s++) = (char) ender;
11112 maybe_exact &= ! IS_IN_SOME_FOLD_L1(ender);
11116 const STRLEN unilen = reguni(pRExC_state, ender, s);
11122 /* See comment just above for - 1 */
11126 REGC((char)ender, s++);
11129 if (next_is_quantifier) {
11131 /* Here, the next input is a quantifier, and to get here,
11132 * the current character is the only one in the node.
11133 * Also, here <len> doesn't include the final byte for this
11139 } /* End of loop through literal characters */
11141 /* Here we have either exhausted the input or ran out of room in
11142 * the node. (If we encountered a character that can't be in the
11143 * node, transfer is made directly to <loopdone>, and so we
11144 * wouldn't have fallen off the end of the loop.) In the latter
11145 * case, we artificially have to split the node into two, because
11146 * we just don't have enough space to hold everything. This
11147 * creates a problem if the final character participates in a
11148 * multi-character fold in the non-final position, as a match that
11149 * should have occurred won't, due to the way nodes are matched,
11150 * and our artificial boundary. So back off until we find a non-
11151 * problematic character -- one that isn't at the beginning or
11152 * middle of such a fold. (Either it doesn't participate in any
11153 * folds, or appears only in the final position of all the folds it
11154 * does participate in.) A better solution with far fewer false
11155 * positives, and that would fill the nodes more completely, would
11156 * be to actually have available all the multi-character folds to
11157 * test against, and to back-off only far enough to be sure that
11158 * this node isn't ending with a partial one. <upper_parse> is set
11159 * further below (if we need to reparse the node) to include just
11160 * up through that final non-problematic character that this code
11161 * identifies, so when it is set to less than the full node, we can
11162 * skip the rest of this */
11163 if (FOLD && p < RExC_end && upper_parse == MAX_NODE_STRING_SIZE) {
11165 const STRLEN full_len = len;
11167 assert(len >= MAX_NODE_STRING_SIZE);
11169 /* Here, <s> points to the final byte of the final character.
11170 * Look backwards through the string until find a non-
11171 * problematic character */
11175 /* These two have no multi-char folds to non-UTF characters
11177 if (ASCII_FOLD_RESTRICTED || LOC) {
11181 while (--s >= s0 && IS_NON_FINAL_FOLD(*s)) { }
11185 if (! PL_NonL1NonFinalFold) {
11186 PL_NonL1NonFinalFold = _new_invlist_C_array(
11187 NonL1_Perl_Non_Final_Folds_invlist);
11190 /* Point to the first byte of the final character */
11191 s = (char *) utf8_hop((U8 *) s, -1);
11193 while (s >= s0) { /* Search backwards until find
11194 non-problematic char */
11195 if (UTF8_IS_INVARIANT(*s)) {
11197 /* There are no ascii characters that participate
11198 * in multi-char folds under /aa. In EBCDIC, the
11199 * non-ascii invariants are all control characters,
11200 * so don't ever participate in any folds. */
11201 if (ASCII_FOLD_RESTRICTED
11202 || ! IS_NON_FINAL_FOLD(*s))
11207 else if (UTF8_IS_DOWNGRADEABLE_START(*s)) {
11209 /* No Latin1 characters participate in multi-char
11210 * folds under /l */
11212 || ! IS_NON_FINAL_FOLD(TWO_BYTE_UTF8_TO_UNI(
11218 else if (! _invlist_contains_cp(
11219 PL_NonL1NonFinalFold,
11220 valid_utf8_to_uvchr((U8 *) s, NULL)))
11225 /* Here, the current character is problematic in that
11226 * it does occur in the non-final position of some
11227 * fold, so try the character before it, but have to
11228 * special case the very first byte in the string, so
11229 * we don't read outside the string */
11230 s = (s == s0) ? s -1 : (char *) utf8_hop((U8 *) s, -1);
11231 } /* End of loop backwards through the string */
11233 /* If there were only problematic characters in the string,
11234 * <s> will point to before s0, in which case the length
11235 * should be 0, otherwise include the length of the
11236 * non-problematic character just found */
11237 len = (s < s0) ? 0 : s - s0 + UTF8SKIP(s);
11240 /* Here, have found the final character, if any, that is
11241 * non-problematic as far as ending the node without splitting
11242 * it across a potential multi-char fold. <len> contains the
11243 * number of bytes in the node up-to and including that
11244 * character, or is 0 if there is no such character, meaning
11245 * the whole node contains only problematic characters. In
11246 * this case, give up and just take the node as-is. We can't
11252 /* Here, the node does contain some characters that aren't
11253 * problematic. If one such is the final character in the
11254 * node, we are done */
11255 if (len == full_len) {
11258 else if (len + ((UTF) ? UTF8SKIP(s) : 1) == full_len) {
11260 /* If the final character is problematic, but the
11261 * penultimate is not, back-off that last character to
11262 * later start a new node with it */
11267 /* Here, the final non-problematic character is earlier
11268 * in the input than the penultimate character. What we do
11269 * is reparse from the beginning, going up only as far as
11270 * this final ok one, thus guaranteeing that the node ends
11271 * in an acceptable character. The reason we reparse is
11272 * that we know how far in the character is, but we don't
11273 * know how to correlate its position with the input parse.
11274 * An alternate implementation would be to build that
11275 * correlation as we go along during the original parse,
11276 * but that would entail extra work for every node, whereas
11277 * this code gets executed only when the string is too
11278 * large for the node, and the final two characters are
11279 * problematic, an infrequent occurrence. Yet another
11280 * possible strategy would be to save the tail of the
11281 * string, and the next time regatom is called, initialize
11282 * with that. The problem with this is that unless you
11283 * back off one more character, you won't be guaranteed
11284 * regatom will get called again, unless regbranch,
11285 * regpiece ... are also changed. If you do back off that
11286 * extra character, so that there is input guaranteed to
11287 * force calling regatom, you can't handle the case where
11288 * just the first character in the node is acceptable. I
11289 * (khw) decided to try this method which doesn't have that
11290 * pitfall; if performance issues are found, we can do a
11291 * combination of the current approach plus that one */
11297 } /* End of verifying node ends with an appropriate char */
11299 loopdone: /* Jumped to when encounters something that shouldn't be in
11302 /* If 'maybe_exact' is still set here, means there are no
11303 * code points in the node that participate in folds */
11304 if (FOLD && maybe_exact) {
11308 /* I (khw) don't know if you can get here with zero length, but the
11309 * old code handled this situation by creating a zero-length EXACT
11310 * node. Might as well be NOTHING instead */
11315 alloc_maybe_populate_EXACT(pRExC_state, ret, flagp, len, ender);
11318 RExC_parse = p - 1;
11319 Set_Node_Cur_Length(ret); /* MJD */
11320 nextchar(pRExC_state);
11322 /* len is STRLEN which is unsigned, need to copy to signed */
11325 vFAIL("Internal disaster");
11328 } /* End of label 'defchar:' */
11330 } /* End of giant switch on input character */
11336 S_regwhite( RExC_state_t *pRExC_state, char *p )
11338 const char *e = RExC_end;
11340 PERL_ARGS_ASSERT_REGWHITE;
11345 else if (*p == '#') {
11348 if (*p++ == '\n') {
11354 RExC_seen |= REG_SEEN_RUN_ON_COMMENT;
11363 S_regpatws( RExC_state_t *pRExC_state, char *p , const bool recognize_comment )
11365 /* Returns the next non-pattern-white space, non-comment character (the
11366 * latter only if 'recognize_comment is true) in the string p, which is
11367 * ended by RExC_end. If there is no line break ending a comment,
11368 * RExC_seen has added the REG_SEEN_RUN_ON_COMMENT flag; */
11369 const char *e = RExC_end;
11371 PERL_ARGS_ASSERT_REGPATWS;
11375 if ((len = is_PATWS_safe(p, e, UTF))) {
11378 else if (recognize_comment && *p == '#') {
11382 if (is_LNBREAK_safe(p, e, UTF)) {
11388 RExC_seen |= REG_SEEN_RUN_ON_COMMENT;
11396 /* Parse POSIX character classes: [[:foo:]], [[=foo=]], [[.foo.]].
11397 Character classes ([:foo:]) can also be negated ([:^foo:]).
11398 Returns a named class id (ANYOF_XXX) if successful, -1 otherwise.
11399 Equivalence classes ([=foo=]) and composites ([.foo.]) are parsed,
11400 but trigger failures because they are currently unimplemented. */
11402 #define POSIXCC_DONE(c) ((c) == ':')
11403 #define POSIXCC_NOTYET(c) ((c) == '=' || (c) == '.')
11404 #define POSIXCC(c) (POSIXCC_DONE(c) || POSIXCC_NOTYET(c))
11406 PERL_STATIC_INLINE I32
11407 S_regpposixcc(pTHX_ RExC_state_t *pRExC_state, I32 value, const bool strict)
11410 I32 namedclass = OOB_NAMEDCLASS;
11412 PERL_ARGS_ASSERT_REGPPOSIXCC;
11414 if (value == '[' && RExC_parse + 1 < RExC_end &&
11415 /* I smell either [: or [= or [. -- POSIX has been here, right? */
11416 POSIXCC(UCHARAT(RExC_parse)))
11418 const char c = UCHARAT(RExC_parse);
11419 char* const s = RExC_parse++;
11421 while (RExC_parse < RExC_end && UCHARAT(RExC_parse) != c)
11423 if (RExC_parse == RExC_end) {
11426 /* Try to give a better location for the error (than the end of
11427 * the string) by looking for the matching ']' */
11429 while (RExC_parse < RExC_end && UCHARAT(RExC_parse) != ']') {
11432 vFAIL2("Unmatched '%c' in POSIX class", c);
11434 /* Grandfather lone [:, [=, [. */
11438 const char* const t = RExC_parse++; /* skip over the c */
11441 if (UCHARAT(RExC_parse) == ']') {
11442 const char *posixcc = s + 1;
11443 RExC_parse++; /* skip over the ending ] */
11446 const I32 complement = *posixcc == '^' ? *posixcc++ : 0;
11447 const I32 skip = t - posixcc;
11449 /* Initially switch on the length of the name. */
11452 if (memEQ(posixcc, "word", 4)) /* this is not POSIX,
11453 this is the Perl \w
11455 namedclass = ANYOF_WORDCHAR;
11458 /* Names all of length 5. */
11459 /* alnum alpha ascii blank cntrl digit graph lower
11460 print punct space upper */
11461 /* Offset 4 gives the best switch position. */
11462 switch (posixcc[4]) {
11464 if (memEQ(posixcc, "alph", 4)) /* alpha */
11465 namedclass = ANYOF_ALPHA;
11468 if (memEQ(posixcc, "spac", 4)) /* space */
11469 namedclass = ANYOF_PSXSPC;
11472 if (memEQ(posixcc, "grap", 4)) /* graph */
11473 namedclass = ANYOF_GRAPH;
11476 if (memEQ(posixcc, "asci", 4)) /* ascii */
11477 namedclass = ANYOF_ASCII;
11480 if (memEQ(posixcc, "blan", 4)) /* blank */
11481 namedclass = ANYOF_BLANK;
11484 if (memEQ(posixcc, "cntr", 4)) /* cntrl */
11485 namedclass = ANYOF_CNTRL;
11488 if (memEQ(posixcc, "alnu", 4)) /* alnum */
11489 namedclass = ANYOF_ALPHANUMERIC;
11492 if (memEQ(posixcc, "lowe", 4)) /* lower */
11493 namedclass = (FOLD) ? ANYOF_CASED : ANYOF_LOWER;
11494 else if (memEQ(posixcc, "uppe", 4)) /* upper */
11495 namedclass = (FOLD) ? ANYOF_CASED : ANYOF_UPPER;
11498 if (memEQ(posixcc, "digi", 4)) /* digit */
11499 namedclass = ANYOF_DIGIT;
11500 else if (memEQ(posixcc, "prin", 4)) /* print */
11501 namedclass = ANYOF_PRINT;
11502 else if (memEQ(posixcc, "punc", 4)) /* punct */
11503 namedclass = ANYOF_PUNCT;
11508 if (memEQ(posixcc, "xdigit", 6))
11509 namedclass = ANYOF_XDIGIT;
11513 if (namedclass == OOB_NAMEDCLASS)
11514 Simple_vFAIL3("POSIX class [:%.*s:] unknown",
11517 /* The #defines are structured so each complement is +1 to
11518 * the normal one */
11522 assert (posixcc[skip] == ':');
11523 assert (posixcc[skip+1] == ']');
11524 } else if (!SIZE_ONLY) {
11525 /* [[=foo=]] and [[.foo.]] are still future. */
11527 /* adjust RExC_parse so the warning shows after
11528 the class closes */
11529 while (UCHARAT(RExC_parse) && UCHARAT(RExC_parse) != ']')
11531 vFAIL3("POSIX syntax [%c %c] is reserved for future extensions", c, c);
11534 /* Maternal grandfather:
11535 * "[:" ending in ":" but not in ":]" */
11537 vFAIL("Unmatched '[' in POSIX class");
11540 /* Grandfather lone [:, [=, [. */
11550 S_could_it_be_a_POSIX_class(pTHX_ RExC_state_t *pRExC_state)
11552 /* This applies some heuristics at the current parse position (which should
11553 * be at a '[') to see if what follows might be intended to be a [:posix:]
11554 * class. It returns true if it really is a posix class, of course, but it
11555 * also can return true if it thinks that what was intended was a posix
11556 * class that didn't quite make it.
11558 * It will return true for
11560 * [:alphanumerics] (as long as the ] isn't followed immediately by a
11561 * ')' indicating the end of the (?[
11562 * [:any garbage including %^&$ punctuation:]
11564 * This is designed to be called only from S_handle_regex_sets; it could be
11565 * easily adapted to be called from the spot at the beginning of regclass()
11566 * that checks to see in a normal bracketed class if the surrounding []
11567 * have been omitted ([:word:] instead of [[:word:]]). But doing so would
11568 * change long-standing behavior, so I (khw) didn't do that */
11569 char* p = RExC_parse + 1;
11570 char first_char = *p;
11572 PERL_ARGS_ASSERT_COULD_IT_BE_A_POSIX_CLASS;
11574 assert(*(p - 1) == '[');
11576 if (! POSIXCC(first_char)) {
11581 while (p < RExC_end && isWORDCHAR(*p)) p++;
11583 if (p >= RExC_end) {
11587 if (p - RExC_parse > 2 /* Got at least 1 word character */
11588 && (*p == first_char
11589 || (*p == ']' && p + 1 < RExC_end && *(p + 1) != ')')))
11594 p = (char *) memchr(RExC_parse, ']', RExC_end - RExC_parse);
11597 && p - RExC_parse > 2 /* [:] evaluates to colon;
11598 [::] is a bad posix class. */
11599 && first_char == *(p - 1));
11603 S_handle_regex_sets(pTHX_ RExC_state_t *pRExC_state, SV** return_invlist, I32 *flagp, U32 depth,
11604 char * const oregcomp_parse)
11606 /* Handle the (?[...]) construct to do set operations */
11609 UV start, end; /* End points of code point ranges */
11611 char *save_end, *save_parse;
11616 const bool save_fold = FOLD;
11618 GET_RE_DEBUG_FLAGS_DECL;
11620 PERL_ARGS_ASSERT_HANDLE_REGEX_SETS;
11623 vFAIL("(?[...]) not valid in locale");
11625 RExC_uni_semantics = 1;
11627 /* This will return only an ANYOF regnode, or (unlikely) something smaller
11628 * (such as EXACT). Thus we can skip most everything if just sizing. We
11629 * call regclass to handle '[]' so as to not have to reinvent its parsing
11630 * rules here (throwing away the size it computes each time). And, we exit
11631 * upon an unescaped ']' that isn't one ending a regclass. To do both
11632 * these things, we need to realize that something preceded by a backslash
11633 * is escaped, so we have to keep track of backslashes */
11636 Perl_ck_warner_d(aTHX_
11637 packWARN(WARN_EXPERIMENTAL__REGEX_SETS),
11638 "The regex_sets feature is experimental" REPORT_LOCATION,
11639 (int) (RExC_parse - RExC_precomp) , RExC_precomp, RExC_parse);
11641 while (RExC_parse < RExC_end) {
11642 SV* current = NULL;
11643 RExC_parse = regpatws(pRExC_state, RExC_parse,
11644 TRUE); /* means recognize comments */
11645 switch (*RExC_parse) {
11649 /* Skip the next byte (which could cause us to end up in
11650 * the middle of a UTF-8 character, but since none of those
11651 * are confusable with anything we currently handle in this
11652 * switch (invariants all), it's safe. We'll just hit the
11653 * default: case next time and keep on incrementing until
11654 * we find one of the invariants we do handle. */
11659 /* If this looks like it is a [:posix:] class, leave the
11660 * parse pointer at the '[' to fool regclass() into
11661 * thinking it is part of a '[[:posix:]]'. That function
11662 * will use strict checking to force a syntax error if it
11663 * doesn't work out to a legitimate class */
11664 bool is_posix_class
11665 = could_it_be_a_POSIX_class(pRExC_state);
11666 if (! is_posix_class) {
11670 /* regclass() can only return RESTART_UTF8 if multi-char
11671 folds are allowed. */
11672 if (!regclass(pRExC_state, flagp,depth+1,
11673 is_posix_class, /* parse the whole char
11674 class only if not a
11676 FALSE, /* don't allow multi-char folds */
11677 TRUE, /* silence non-portable warnings. */
11679 FAIL2("panic: regclass returned NULL to handle_sets, flags=%#X",
11682 /* function call leaves parse pointing to the ']', except
11683 * if we faked it */
11684 if (is_posix_class) {
11688 SvREFCNT_dec(current); /* In case it returned something */
11694 if (RExC_parse < RExC_end
11695 && *RExC_parse == ')')
11697 node = reganode(pRExC_state, ANYOF, 0);
11698 RExC_size += ANYOF_SKIP;
11699 nextchar(pRExC_state);
11700 Set_Node_Length(node,
11701 RExC_parse - oregcomp_parse + 1); /* MJD */
11710 FAIL("Syntax error in (?[...])");
11713 /* Pass 2 only after this. Everything in this construct is a
11714 * metacharacter. Operands begin with either a '\' (for an escape
11715 * sequence), or a '[' for a bracketed character class. Any other
11716 * character should be an operator, or parenthesis for grouping. Both
11717 * types of operands are handled by calling regclass() to parse them. It
11718 * is called with a parameter to indicate to return the computed inversion
11719 * list. The parsing here is implemented via a stack. Each entry on the
11720 * stack is a single character representing one of the operators, or the
11721 * '('; or else a pointer to an operand inversion list. */
11723 #define IS_OPERAND(a) (! SvIOK(a))
11725 /* The stack starts empty. It is a syntax error if the first thing parsed
11726 * is a binary operator; everything else is pushed on the stack. When an
11727 * operand is parsed, the top of the stack is examined. If it is a binary
11728 * operator, the item before it should be an operand, and both are replaced
11729 * by the result of doing that operation on the new operand and the one on
11730 * the stack. Thus a sequence of binary operands is reduced to a single
11731 * one before the next one is parsed.
11733 * A unary operator may immediately follow a binary in the input, for
11736 * When an operand is parsed and the top of the stack is a unary operator,
11737 * the operation is performed, and then the stack is rechecked to see if
11738 * this new operand is part of a binary operation; if so, it is handled as
11741 * A '(' is simply pushed on the stack; it is valid only if the stack is
11742 * empty, or the top element of the stack is an operator or another '('
11743 * (for which the parenthesized expression will become an operand). By the
11744 * time the corresponding ')' is parsed everything in between should have
11745 * been parsed and evaluated to a single operand (or else is a syntax
11746 * error), and is handled as a regular operand */
11750 while (RExC_parse < RExC_end) {
11751 I32 top_index = av_tindex(stack);
11753 SV* current = NULL;
11755 /* Skip white space */
11756 RExC_parse = regpatws(pRExC_state, RExC_parse,
11757 TRUE); /* means recognize comments */
11758 if (RExC_parse >= RExC_end) {
11759 Perl_croak(aTHX_ "panic: Read past end of '(?[ ])'");
11761 if ((curchar = UCHARAT(RExC_parse)) == ']') {
11768 if (av_tindex(stack) >= 0 /* This makes sure that we can
11769 safely subtract 1 from
11770 RExC_parse in the next clause.
11771 If we have something on the
11772 stack, we have parsed something
11774 && UCHARAT(RExC_parse - 1) == '('
11775 && RExC_parse < RExC_end)
11777 /* If is a '(?', could be an embedded '(?flags:(?[...])'.
11778 * This happens when we have some thing like
11780 * my $thai_or_lao = qr/(?[ \p{Thai} + \p{Lao} ])/;
11782 * qr/(?[ \p{Digit} & $thai_or_lao ])/;
11784 * Here we would be handling the interpolated
11785 * '$thai_or_lao'. We handle this by a recursive call to
11786 * ourselves which returns the inversion list the
11787 * interpolated expression evaluates to. We use the flags
11788 * from the interpolated pattern. */
11789 U32 save_flags = RExC_flags;
11790 const char * const save_parse = ++RExC_parse;
11792 parse_lparen_question_flags(pRExC_state);
11794 if (RExC_parse == save_parse /* Makes sure there was at
11795 least one flag (or this
11796 embedding wasn't compiled)
11798 || RExC_parse >= RExC_end - 4
11799 || UCHARAT(RExC_parse) != ':'
11800 || UCHARAT(++RExC_parse) != '('
11801 || UCHARAT(++RExC_parse) != '?'
11802 || UCHARAT(++RExC_parse) != '[')
11805 /* In combination with the above, this moves the
11806 * pointer to the point just after the first erroneous
11807 * character (or if there are no flags, to where they
11808 * should have been) */
11809 if (RExC_parse >= RExC_end - 4) {
11810 RExC_parse = RExC_end;
11812 else if (RExC_parse != save_parse) {
11813 RExC_parse += (UTF) ? UTF8SKIP(RExC_parse) : 1;
11815 vFAIL("Expecting '(?flags:(?[...'");
11818 (void) handle_regex_sets(pRExC_state, ¤t, flagp,
11819 depth+1, oregcomp_parse);
11821 /* Here, 'current' contains the embedded expression's
11822 * inversion list, and RExC_parse points to the trailing
11823 * ']'; the next character should be the ')' which will be
11824 * paired with the '(' that has been put on the stack, so
11825 * the whole embedded expression reduces to '(operand)' */
11828 RExC_flags = save_flags;
11829 goto handle_operand;
11834 RExC_parse += (UTF) ? UTF8SKIP(RExC_parse) : 1;
11835 vFAIL("Unexpected character");
11838 /* regclass() can only return RESTART_UTF8 if multi-char
11839 folds are allowed. */
11840 if (!regclass(pRExC_state, flagp,depth+1,
11841 TRUE, /* means parse just the next thing */
11842 FALSE, /* don't allow multi-char folds */
11843 FALSE, /* don't silence non-portable warnings. */
11845 FAIL2("panic: regclass returned NULL to handle_sets, flags=%#X",
11847 /* regclass() will return with parsing just the \ sequence,
11848 * leaving the parse pointer at the next thing to parse */
11850 goto handle_operand;
11852 case '[': /* Is a bracketed character class */
11854 bool is_posix_class = could_it_be_a_POSIX_class(pRExC_state);
11856 if (! is_posix_class) {
11860 /* regclass() can only return RESTART_UTF8 if multi-char
11861 folds are allowed. */
11862 if(!regclass(pRExC_state, flagp,depth+1,
11863 is_posix_class, /* parse the whole char class
11864 only if not a posix class */
11865 FALSE, /* don't allow multi-char folds */
11866 FALSE, /* don't silence non-portable warnings. */
11868 FAIL2("panic: regclass returned NULL to handle_sets, flags=%#X",
11870 /* function call leaves parse pointing to the ']', except if we
11872 if (is_posix_class) {
11876 goto handle_operand;
11885 || ( ! (top_ptr = av_fetch(stack, top_index, FALSE)))
11886 || ! IS_OPERAND(*top_ptr))
11889 vFAIL2("Unexpected binary operator '%c' with no preceding operand", curchar);
11891 av_push(stack, newSVuv(curchar));
11895 av_push(stack, newSVuv(curchar));
11899 if (top_index >= 0) {
11900 top_ptr = av_fetch(stack, top_index, FALSE);
11902 if (IS_OPERAND(*top_ptr)) {
11904 vFAIL("Unexpected '(' with no preceding operator");
11907 av_push(stack, newSVuv(curchar));
11914 || ! (current = av_pop(stack))
11915 || ! IS_OPERAND(current)
11916 || ! (lparen = av_pop(stack))
11917 || IS_OPERAND(lparen)
11918 || SvUV(lparen) != '(')
11921 vFAIL("Unexpected ')'");
11924 SvREFCNT_dec_NN(lparen);
11931 /* Here, we have an operand to process, in 'current' */
11933 if (top_index < 0) { /* Just push if stack is empty */
11934 av_push(stack, current);
11937 SV* top = av_pop(stack);
11938 char current_operator;
11940 if (IS_OPERAND(top)) {
11941 vFAIL("Operand with no preceding operator");
11943 current_operator = (char) SvUV(top);
11944 switch (current_operator) {
11945 case '(': /* Push the '(' back on followed by the new
11947 av_push(stack, top);
11948 av_push(stack, current);
11949 SvREFCNT_inc(top); /* Counters the '_dec' done
11950 just after the 'break', so
11951 it doesn't get wrongly freed
11956 _invlist_invert(current);
11958 /* Unlike binary operators, the top of the stack,
11959 * now that this unary one has been popped off, may
11960 * legally be an operator, and we now have operand
11963 SvREFCNT_dec_NN(top);
11964 goto handle_operand;
11967 _invlist_intersection(av_pop(stack),
11970 av_push(stack, current);
11975 _invlist_union(av_pop(stack), current, ¤t);
11976 av_push(stack, current);
11980 _invlist_subtract(av_pop(stack), current, ¤t);
11981 av_push(stack, current);
11984 case '^': /* The union minus the intersection */
11990 element = av_pop(stack);
11991 _invlist_union(element, current, &u);
11992 _invlist_intersection(element, current, &i);
11993 _invlist_subtract(u, i, ¤t);
11994 av_push(stack, current);
11995 SvREFCNT_dec_NN(i);
11996 SvREFCNT_dec_NN(u);
11997 SvREFCNT_dec_NN(element);
12002 Perl_croak(aTHX_ "panic: Unexpected item on '(?[ ])' stack");
12004 SvREFCNT_dec_NN(top);
12008 RExC_parse += (UTF) ? UTF8SKIP(RExC_parse) : 1;
12011 if (av_tindex(stack) < 0 /* Was empty */
12012 || ((final = av_pop(stack)) == NULL)
12013 || ! IS_OPERAND(final)
12014 || av_tindex(stack) >= 0) /* More left on stack */
12016 vFAIL("Incomplete expression within '(?[ ])'");
12019 /* Here, 'final' is the resultant inversion list from evaluating the
12020 * expression. Return it if so requested */
12021 if (return_invlist) {
12022 *return_invlist = final;
12026 /* Otherwise generate a resultant node, based on 'final'. regclass() is
12027 * expecting a string of ranges and individual code points */
12028 invlist_iterinit(final);
12029 result_string = newSVpvs("");
12030 while (invlist_iternext(final, &start, &end)) {
12031 if (start == end) {
12032 Perl_sv_catpvf(aTHX_ result_string, "\\x{%"UVXf"}", start);
12035 Perl_sv_catpvf(aTHX_ result_string, "\\x{%"UVXf"}-\\x{%"UVXf"}",
12040 save_parse = RExC_parse;
12041 RExC_parse = SvPV(result_string, len);
12042 save_end = RExC_end;
12043 RExC_end = RExC_parse + len;
12045 /* We turn off folding around the call, as the class we have constructed
12046 * already has all folding taken into consideration, and we don't want
12047 * regclass() to add to that */
12048 RExC_flags &= ~RXf_PMf_FOLD;
12049 /* regclass() can only return RESTART_UTF8 if multi-char folds are allowed.
12051 node = regclass(pRExC_state, flagp,depth+1,
12052 FALSE, /* means parse the whole char class */
12053 FALSE, /* don't allow multi-char folds */
12054 TRUE, /* silence non-portable warnings. The above may very
12055 well have generated non-portable code points, but
12056 they're valid on this machine */
12059 FAIL2("panic: regclass returned NULL to handle_sets, flags=%#"UVxf,
12062 RExC_flags |= RXf_PMf_FOLD;
12064 RExC_parse = save_parse + 1;
12065 RExC_end = save_end;
12066 SvREFCNT_dec_NN(final);
12067 SvREFCNT_dec_NN(result_string);
12068 SvREFCNT_dec_NN(stack);
12070 nextchar(pRExC_state);
12071 Set_Node_Length(node, RExC_parse - oregcomp_parse + 1); /* MJD */
12076 /* The names of properties whose definitions are not known at compile time are
12077 * stored in this SV, after a constant heading. So if the length has been
12078 * changed since initialization, then there is a run-time definition. */
12079 #define HAS_NONLOCALE_RUNTIME_PROPERTY_DEFINITION (SvCUR(listsv) != initial_listsv_len)
12082 S_regclass(pTHX_ RExC_state_t *pRExC_state, I32 *flagp, U32 depth,
12083 const bool stop_at_1, /* Just parse the next thing, don't
12084 look for a full character class */
12085 bool allow_multi_folds,
12086 const bool silence_non_portable, /* Don't output warnings
12089 SV** ret_invlist) /* Return an inversion list, not a node */
12091 /* parse a bracketed class specification. Most of these will produce an
12092 * ANYOF node; but something like [a] will produce an EXACT node; [aA], an
12093 * EXACTFish node; [[:ascii:]], a POSIXA node; etc. It is more complex
12094 * under /i with multi-character folds: it will be rewritten following the
12095 * paradigm of this example, where the <multi-fold>s are characters which
12096 * fold to multiple character sequences:
12097 * /[abc\x{multi-fold1}def\x{multi-fold2}ghi]/i
12098 * gets effectively rewritten as:
12099 * /(?:\x{multi-fold1}|\x{multi-fold2}|[abcdefghi]/i
12100 * reg() gets called (recursively) on the rewritten version, and this
12101 * function will return what it constructs. (Actually the <multi-fold>s
12102 * aren't physically removed from the [abcdefghi], it's just that they are
12103 * ignored in the recursion by means of a flag:
12104 * <RExC_in_multi_char_class>.)
12106 * ANYOF nodes contain a bit map for the first 256 characters, with the
12107 * corresponding bit set if that character is in the list. For characters
12108 * above 255, a range list or swash is used. There are extra bits for \w,
12109 * etc. in locale ANYOFs, as what these match is not determinable at
12112 * Returns NULL, setting *flagp to RESTART_UTF8 if the sizing scan needs
12113 * to be restarted. This can only happen if ret_invlist is non-NULL.
12117 UV prevvalue = OOB_UNICODE, save_prevvalue = OOB_UNICODE;
12119 UV value = OOB_UNICODE, save_value = OOB_UNICODE;
12122 IV namedclass = OOB_NAMEDCLASS;
12123 char *rangebegin = NULL;
12124 bool need_class = 0;
12126 STRLEN initial_listsv_len = 0; /* Kind of a kludge to see if it is more
12127 than just initialized. */
12128 SV* properties = NULL; /* Code points that match \p{} \P{} */
12129 SV* posixes = NULL; /* Code points that match classes like, [:word:],
12130 extended beyond the Latin1 range */
12131 UV element_count = 0; /* Number of distinct elements in the class.
12132 Optimizations may be possible if this is tiny */
12133 AV * multi_char_matches = NULL; /* Code points that fold to more than one
12134 character; used under /i */
12136 char * stop_ptr = RExC_end; /* where to stop parsing */
12137 const bool skip_white = cBOOL(ret_invlist); /* ignore unescaped white
12139 const bool strict = cBOOL(ret_invlist); /* Apply strict parsing rules? */
12141 /* Unicode properties are stored in a swash; this holds the current one
12142 * being parsed. If this swash is the only above-latin1 component of the
12143 * character class, an optimization is to pass it directly on to the
12144 * execution engine. Otherwise, it is set to NULL to indicate that there
12145 * are other things in the class that have to be dealt with at execution
12147 SV* swash = NULL; /* Code points that match \p{} \P{} */
12149 /* Set if a component of this character class is user-defined; just passed
12150 * on to the engine */
12151 bool has_user_defined_property = FALSE;
12153 /* inversion list of code points this node matches only when the target
12154 * string is in UTF-8. (Because is under /d) */
12155 SV* depends_list = NULL;
12157 /* inversion list of code points this node matches. For much of the
12158 * function, it includes only those that match regardless of the utf8ness
12159 * of the target string */
12160 SV* cp_list = NULL;
12163 /* In a range, counts how many 0-2 of the ends of it came from literals,
12164 * not escapes. Thus we can tell if 'A' was input vs \x{C1} */
12165 UV literal_endpoint = 0;
12167 bool invert = FALSE; /* Is this class to be complemented */
12169 /* Is there any thing like \W or [:^digit:] that matches above the legal
12170 * Unicode range? */
12171 bool runtime_posix_matches_above_Unicode = FALSE;
12173 regnode * const orig_emit = RExC_emit; /* Save the original RExC_emit in
12174 case we need to change the emitted regop to an EXACT. */
12175 const char * orig_parse = RExC_parse;
12176 const I32 orig_size = RExC_size;
12177 GET_RE_DEBUG_FLAGS_DECL;
12179 PERL_ARGS_ASSERT_REGCLASS;
12181 PERL_UNUSED_ARG(depth);
12184 DEBUG_PARSE("clas");
12186 /* Assume we are going to generate an ANYOF node. */
12187 ret = reganode(pRExC_state, ANYOF, 0);
12190 RExC_size += ANYOF_SKIP;
12191 listsv = &PL_sv_undef; /* For code scanners: listsv always non-NULL. */
12194 ANYOF_FLAGS(ret) = 0;
12196 RExC_emit += ANYOF_SKIP;
12198 ANYOF_FLAGS(ret) |= ANYOF_LOCALE;
12200 listsv = newSVpvs_flags("# comment\n", SVs_TEMP);
12201 initial_listsv_len = SvCUR(listsv);
12202 SvTEMP_off(listsv); /* Grr, TEMPs and mortals are conflated. */
12206 RExC_parse = regpatws(pRExC_state, RExC_parse,
12207 FALSE /* means don't recognize comments */);
12210 if (UCHARAT(RExC_parse) == '^') { /* Complement of range. */
12213 allow_multi_folds = FALSE;
12216 RExC_parse = regpatws(pRExC_state, RExC_parse,
12217 FALSE /* means don't recognize comments */);
12221 /* Check that they didn't say [:posix:] instead of [[:posix:]] */
12222 if (!SIZE_ONLY && RExC_parse < RExC_end && POSIXCC(UCHARAT(RExC_parse))) {
12223 const char *s = RExC_parse;
12224 const char c = *s++;
12226 while (isWORDCHAR(*s))
12228 if (*s && c == *s && s[1] == ']') {
12229 SAVEFREESV(RExC_rx_sv);
12231 "POSIX syntax [%c %c] belongs inside character classes",
12233 (void)ReREFCNT_inc(RExC_rx_sv);
12237 /* If the caller wants us to just parse a single element, accomplish this
12238 * by faking the loop ending condition */
12239 if (stop_at_1 && RExC_end > RExC_parse) {
12240 stop_ptr = RExC_parse + 1;
12243 /* allow 1st char to be ']' (allowing it to be '-' is dealt with later) */
12244 if (UCHARAT(RExC_parse) == ']')
12245 goto charclassloop;
12249 if (RExC_parse >= stop_ptr) {
12254 RExC_parse = regpatws(pRExC_state, RExC_parse,
12255 FALSE /* means don't recognize comments */);
12258 if (UCHARAT(RExC_parse) == ']') {
12264 namedclass = OOB_NAMEDCLASS; /* initialize as illegal */
12265 save_value = value;
12266 save_prevvalue = prevvalue;
12269 rangebegin = RExC_parse;
12273 value = utf8n_to_uvchr((U8*)RExC_parse,
12274 RExC_end - RExC_parse,
12275 &numlen, UTF8_ALLOW_DEFAULT);
12276 RExC_parse += numlen;
12279 value = UCHARAT(RExC_parse++);
12282 && RExC_parse < RExC_end
12283 && POSIXCC(UCHARAT(RExC_parse)))
12285 namedclass = regpposixcc(pRExC_state, value, strict);
12287 else if (value == '\\') {
12289 value = utf8n_to_uvchr((U8*)RExC_parse,
12290 RExC_end - RExC_parse,
12291 &numlen, UTF8_ALLOW_DEFAULT);
12292 RExC_parse += numlen;
12295 value = UCHARAT(RExC_parse++);
12297 /* Some compilers cannot handle switching on 64-bit integer
12298 * values, therefore value cannot be an UV. Yes, this will
12299 * be a problem later if we want switch on Unicode.
12300 * A similar issue a little bit later when switching on
12301 * namedclass. --jhi */
12303 /* If the \ is escaping white space when white space is being
12304 * skipped, it means that that white space is wanted literally, and
12305 * is already in 'value'. Otherwise, need to translate the escape
12306 * into what it signifies. */
12307 if (! skip_white || ! is_PATWS_cp(value)) switch ((I32)value) {
12309 case 'w': namedclass = ANYOF_WORDCHAR; break;
12310 case 'W': namedclass = ANYOF_NWORDCHAR; break;
12311 case 's': namedclass = ANYOF_SPACE; break;
12312 case 'S': namedclass = ANYOF_NSPACE; break;
12313 case 'd': namedclass = ANYOF_DIGIT; break;
12314 case 'D': namedclass = ANYOF_NDIGIT; break;
12315 case 'v': namedclass = ANYOF_VERTWS; break;
12316 case 'V': namedclass = ANYOF_NVERTWS; break;
12317 case 'h': namedclass = ANYOF_HORIZWS; break;
12318 case 'H': namedclass = ANYOF_NHORIZWS; break;
12319 case 'N': /* Handle \N{NAME} in class */
12321 /* We only pay attention to the first char of
12322 multichar strings being returned. I kinda wonder
12323 if this makes sense as it does change the behaviour
12324 from earlier versions, OTOH that behaviour was broken
12326 if (! grok_bslash_N(pRExC_state, NULL, &value, flagp, depth,
12327 TRUE, /* => charclass */
12330 if (*flagp & RESTART_UTF8)
12331 FAIL("panic: grok_bslash_N set RESTART_UTF8");
12341 /* We will handle any undefined properties ourselves */
12342 U8 swash_init_flags = _CORE_SWASH_INIT_RETURN_IF_UNDEF;
12344 if (RExC_parse >= RExC_end)
12345 vFAIL2("Empty \\%c{}", (U8)value);
12346 if (*RExC_parse == '{') {
12347 const U8 c = (U8)value;
12348 e = strchr(RExC_parse++, '}');
12350 vFAIL2("Missing right brace on \\%c{}", c);
12351 while (isSPACE(UCHARAT(RExC_parse)))
12353 if (e == RExC_parse)
12354 vFAIL2("Empty \\%c{}", c);
12355 n = e - RExC_parse;
12356 while (isSPACE(UCHARAT(RExC_parse + n - 1)))
12367 if (UCHARAT(RExC_parse) == '^') {
12370 /* toggle. (The rhs xor gets the single bit that
12371 * differs between P and p; the other xor inverts just
12373 value ^= 'P' ^ 'p';
12375 while (isSPACE(UCHARAT(RExC_parse))) {
12380 /* Try to get the definition of the property into
12381 * <invlist>. If /i is in effect, the effective property
12382 * will have its name be <__NAME_i>. The design is
12383 * discussed in commit
12384 * 2f833f5208e26b208886e51e09e2c072b5eabb46 */
12385 Newx(name, n + sizeof("_i__\n"), char);
12387 sprintf(name, "%s%.*s%s\n",
12388 (FOLD) ? "__" : "",
12394 /* Look up the property name, and get its swash and
12395 * inversion list, if the property is found */
12397 SvREFCNT_dec_NN(swash);
12399 swash = _core_swash_init("utf8", name, &PL_sv_undef,
12402 NULL, /* No inversion list */
12405 if (! swash || ! (invlist = _get_swash_invlist(swash))) {
12407 SvREFCNT_dec_NN(swash);
12411 /* Here didn't find it. It could be a user-defined
12412 * property that will be available at run-time. If we
12413 * accept only compile-time properties, is an error;
12414 * otherwise add it to the list for run-time look up */
12416 RExC_parse = e + 1;
12417 vFAIL3("Property '%.*s' is unknown", (int) n, name);
12419 Perl_sv_catpvf(aTHX_ listsv, "%cutf8::%s\n",
12420 (value == 'p' ? '+' : '!'),
12422 has_user_defined_property = TRUE;
12424 /* We don't know yet, so have to assume that the
12425 * property could match something in the Latin1 range,
12426 * hence something that isn't utf8. Note that this
12427 * would cause things in <depends_list> to match
12428 * inappropriately, except that any \p{}, including
12429 * this one forces Unicode semantics, which means there
12430 * is <no depends_list> */
12431 ANYOF_FLAGS(ret) |= ANYOF_NONBITMAP_NON_UTF8;
12435 /* Here, did get the swash and its inversion list. If
12436 * the swash is from a user-defined property, then this
12437 * whole character class should be regarded as such */
12438 has_user_defined_property =
12440 & _CORE_SWASH_INIT_USER_DEFINED_PROPERTY);
12442 /* Invert if asking for the complement */
12443 if (value == 'P') {
12444 _invlist_union_complement_2nd(properties,
12448 /* The swash can't be used as-is, because we've
12449 * inverted things; delay removing it to here after
12450 * have copied its invlist above */
12451 SvREFCNT_dec_NN(swash);
12455 _invlist_union(properties, invlist, &properties);
12460 RExC_parse = e + 1;
12461 namedclass = ANYOF_UNIPROP; /* no official name, but it's
12464 /* \p means they want Unicode semantics */
12465 RExC_uni_semantics = 1;
12468 case 'n': value = '\n'; break;
12469 case 'r': value = '\r'; break;
12470 case 't': value = '\t'; break;
12471 case 'f': value = '\f'; break;
12472 case 'b': value = '\b'; break;
12473 case 'e': value = ASCII_TO_NATIVE('\033');break;
12474 case 'a': value = ASCII_TO_NATIVE('\007');break;
12476 RExC_parse--; /* function expects to be pointed at the 'o' */
12478 const char* error_msg;
12479 bool valid = grok_bslash_o(&RExC_parse,
12482 SIZE_ONLY, /* warnings in pass
12485 silence_non_portable,
12491 if (PL_encoding && value < 0x100) {
12492 goto recode_encoding;
12496 RExC_parse--; /* function expects to be pointed at the 'x' */
12498 const char* error_msg;
12499 bool valid = grok_bslash_x(&RExC_parse,
12502 TRUE, /* Output warnings */
12504 silence_non_portable,
12510 if (PL_encoding && value < 0x100)
12511 goto recode_encoding;
12514 value = grok_bslash_c(*RExC_parse++, UTF, SIZE_ONLY);
12516 case '0': case '1': case '2': case '3': case '4':
12517 case '5': case '6': case '7':
12519 /* Take 1-3 octal digits */
12520 I32 flags = PERL_SCAN_SILENT_ILLDIGIT;
12521 numlen = (strict) ? 4 : 3;
12522 value = grok_oct(--RExC_parse, &numlen, &flags, NULL);
12523 RExC_parse += numlen;
12526 RExC_parse += (UTF) ? UTF8SKIP(RExC_parse) : 1;
12527 vFAIL("Need exactly 3 octal digits");
12529 else if (! SIZE_ONLY /* like \08, \178 */
12531 && RExC_parse < RExC_end
12532 && isDIGIT(*RExC_parse)
12533 && ckWARN(WARN_REGEXP))
12535 SAVEFREESV(RExC_rx_sv);
12536 reg_warn_non_literal_string(
12538 form_short_octal_warning(RExC_parse, numlen));
12539 (void)ReREFCNT_inc(RExC_rx_sv);
12542 if (PL_encoding && value < 0x100)
12543 goto recode_encoding;
12547 if (! RExC_override_recoding) {
12548 SV* enc = PL_encoding;
12549 value = reg_recode((const char)(U8)value, &enc);
12552 vFAIL("Invalid escape in the specified encoding");
12554 else if (SIZE_ONLY) {
12555 ckWARNreg(RExC_parse,
12556 "Invalid escape in the specified encoding");
12562 /* Allow \_ to not give an error */
12563 if (!SIZE_ONLY && isWORDCHAR(value) && value != '_') {
12565 vFAIL2("Unrecognized escape \\%c in character class",
12569 SAVEFREESV(RExC_rx_sv);
12570 ckWARN2reg(RExC_parse,
12571 "Unrecognized escape \\%c in character class passed through",
12573 (void)ReREFCNT_inc(RExC_rx_sv);
12577 } /* End of switch on char following backslash */
12578 } /* end of handling backslash escape sequences */
12581 literal_endpoint++;
12584 /* Here, we have the current token in 'value' */
12586 /* What matches in a locale is not known until runtime. This includes
12587 * what the Posix classes (like \w, [:space:]) match. Room must be
12588 * reserved (one time per class) to store such classes, either if Perl
12589 * is compiled so that locale nodes always should have this space, or
12590 * if there is such class info to be stored. The space will contain a
12591 * bit for each named class that is to be matched against. This isn't
12592 * needed for \p{} and pseudo-classes, as they are not affected by
12593 * locale, and hence are dealt with separately */
12596 && (ANYOF_LOCALE == ANYOF_CLASS
12597 || (namedclass > OOB_NAMEDCLASS && namedclass < ANYOF_MAX)))
12601 RExC_size += ANYOF_CLASS_SKIP - ANYOF_SKIP;
12604 RExC_emit += ANYOF_CLASS_SKIP - ANYOF_SKIP;
12605 ANYOF_CLASS_ZERO(ret);
12607 ANYOF_FLAGS(ret) |= ANYOF_CLASS;
12610 if (namedclass > OOB_NAMEDCLASS) { /* this is a named class \blah */
12612 /* a bad range like a-\d, a-[:digit:]. The '-' is taken as a
12613 * literal, as is the character that began the false range, i.e.
12614 * the 'a' in the examples */
12617 const int w = (RExC_parse >= rangebegin)
12618 ? RExC_parse - rangebegin
12621 vFAIL4("False [] range \"%*.*s\"", w, w, rangebegin);
12624 SAVEFREESV(RExC_rx_sv); /* in case of fatal warnings */
12625 ckWARN4reg(RExC_parse,
12626 "False [] range \"%*.*s\"",
12628 (void)ReREFCNT_inc(RExC_rx_sv);
12629 cp_list = add_cp_to_invlist(cp_list, '-');
12630 cp_list = add_cp_to_invlist(cp_list, prevvalue);
12634 range = 0; /* this was not a true range */
12635 element_count += 2; /* So counts for three values */
12639 U8 classnum = namedclass_to_classnum(namedclass);
12640 if (namedclass >= ANYOF_MAX) { /* If a special class */
12641 if (namedclass != ANYOF_UNIPROP) { /* UNIPROP = \p and \P */
12643 /* Here, should be \h, \H, \v, or \V. Neither /d nor
12644 * /l make a difference in what these match. There
12645 * would be problems if these characters had folds
12646 * other than themselves, as cp_list is subject to
12648 if (classnum != _CC_VERTSPACE) {
12649 assert( namedclass == ANYOF_HORIZWS
12650 || namedclass == ANYOF_NHORIZWS);
12652 /* It turns out that \h is just a synonym for
12654 classnum = _CC_BLANK;
12657 _invlist_union_maybe_complement_2nd(
12659 PL_XPosix_ptrs[classnum],
12660 cBOOL(namedclass % 2), /* Complement if odd
12661 (NHORIZWS, NVERTWS)
12666 else if (classnum == _CC_ASCII) {
12669 ANYOF_CLASS_SET(ret, namedclass);
12672 #endif /* Not isascii(); just use the hard-coded definition for it */
12673 _invlist_union_maybe_complement_2nd(
12676 cBOOL(namedclass % 2), /* Complement if odd
12680 else { /* Garden variety class */
12682 /* The ascii range inversion list */
12683 SV* ascii_source = PL_Posix_ptrs[classnum];
12685 /* The full Latin1 range inversion list */
12686 SV* l1_source = PL_L1Posix_ptrs[classnum];
12688 /* This code is structured into two major clauses. The
12689 * first is for classes whose complete definitions may not
12690 * already be known. It not, the Latin1 definition
12691 * (guaranteed to already known) is used plus code is
12692 * generated to load the rest at run-time (only if needed).
12693 * If the complete definition is known, it drops down to
12694 * the second clause, where the complete definition is
12697 if (classnum < _FIRST_NON_SWASH_CC) {
12699 /* Here, the class has a swash, which may or not
12700 * already be loaded */
12702 /* The name of the property to use to match the full
12703 * eXtended Unicode range swash for this character
12705 const char *Xname = swash_property_names[classnum];
12707 /* If returning the inversion list, we can't defer
12708 * getting this until runtime */
12709 if (ret_invlist && ! PL_utf8_swash_ptrs[classnum]) {
12710 PL_utf8_swash_ptrs[classnum] =
12711 _core_swash_init("utf8", Xname, &PL_sv_undef,
12714 NULL, /* No inversion list */
12715 NULL /* No flags */
12717 assert(PL_utf8_swash_ptrs[classnum]);
12719 if ( ! PL_utf8_swash_ptrs[classnum]) {
12720 if (namedclass % 2 == 0) { /* A non-complemented
12722 /* If not /a matching, there are code points we
12723 * don't know at compile time. Arrange for the
12724 * unknown matches to be loaded at run-time, if
12726 if (! AT_LEAST_ASCII_RESTRICTED) {
12727 Perl_sv_catpvf(aTHX_ listsv, "+utf8::%s\n",
12730 if (LOC) { /* Under locale, set run-time
12732 ANYOF_CLASS_SET(ret, namedclass);
12735 /* Add the current class's code points to
12736 * the running total */
12737 _invlist_union(posixes,
12738 (AT_LEAST_ASCII_RESTRICTED)
12744 else { /* A complemented class */
12745 if (AT_LEAST_ASCII_RESTRICTED) {
12746 /* Under /a should match everything above
12747 * ASCII, plus the complement of the set's
12749 _invlist_union_complement_2nd(posixes,
12754 /* Arrange for the unknown matches to be
12755 * loaded at run-time, if needed */
12756 Perl_sv_catpvf(aTHX_ listsv, "!utf8::%s\n",
12758 runtime_posix_matches_above_Unicode = TRUE;
12760 ANYOF_CLASS_SET(ret, namedclass);
12764 /* We want to match everything in
12765 * Latin1, except those things that
12766 * l1_source matches */
12767 SV* scratch_list = NULL;
12768 _invlist_subtract(PL_Latin1, l1_source,
12771 /* Add the list from this class to the
12774 posixes = scratch_list;
12777 _invlist_union(posixes,
12780 SvREFCNT_dec_NN(scratch_list);
12782 if (DEPENDS_SEMANTICS) {
12784 |= ANYOF_NON_UTF8_LATIN1_ALL;
12789 goto namedclass_done;
12792 /* Here, there is a swash loaded for the class. If no
12793 * inversion list for it yet, get it */
12794 if (! PL_XPosix_ptrs[classnum]) {
12795 PL_XPosix_ptrs[classnum]
12796 = _swash_to_invlist(PL_utf8_swash_ptrs[classnum]);
12800 /* Here there is an inversion list already loaded for the
12803 if (namedclass % 2 == 0) { /* A non-complemented class,
12804 like ANYOF_PUNCT */
12806 /* For non-locale, just add it to any existing list
12808 _invlist_union(posixes,
12809 (AT_LEAST_ASCII_RESTRICTED)
12811 : PL_XPosix_ptrs[classnum],
12814 else { /* Locale */
12815 SV* scratch_list = NULL;
12817 /* For above Latin1 code points, we use the full
12819 _invlist_intersection(PL_AboveLatin1,
12820 PL_XPosix_ptrs[classnum],
12822 /* And set the output to it, adding instead if
12823 * there already is an output. Checking if
12824 * 'posixes' is NULL first saves an extra clone.
12825 * Its reference count will be decremented at the
12826 * next union, etc, or if this is the only
12827 * instance, at the end of the routine */
12829 posixes = scratch_list;
12832 _invlist_union(posixes, scratch_list, &posixes);
12833 SvREFCNT_dec_NN(scratch_list);
12836 #ifndef HAS_ISBLANK
12837 if (namedclass != ANYOF_BLANK) {
12839 /* Set this class in the node for runtime
12841 ANYOF_CLASS_SET(ret, namedclass);
12842 #ifndef HAS_ISBLANK
12845 /* No isblank(), use the hard-coded ASCII-range
12846 * blanks, adding them to the running total. */
12848 _invlist_union(posixes, ascii_source, &posixes);
12853 else { /* A complemented class, like ANYOF_NPUNCT */
12855 _invlist_union_complement_2nd(
12857 (AT_LEAST_ASCII_RESTRICTED)
12859 : PL_XPosix_ptrs[classnum],
12861 /* Under /d, everything in the upper half of the
12862 * Latin1 range matches this complement */
12863 if (DEPENDS_SEMANTICS) {
12864 ANYOF_FLAGS(ret) |= ANYOF_NON_UTF8_LATIN1_ALL;
12867 else { /* Locale */
12868 SV* scratch_list = NULL;
12869 _invlist_subtract(PL_AboveLatin1,
12870 PL_XPosix_ptrs[classnum],
12873 posixes = scratch_list;
12876 _invlist_union(posixes, scratch_list, &posixes);
12877 SvREFCNT_dec_NN(scratch_list);
12879 #ifndef HAS_ISBLANK
12880 if (namedclass != ANYOF_NBLANK) {
12882 ANYOF_CLASS_SET(ret, namedclass);
12883 #ifndef HAS_ISBLANK
12886 /* Get the list of all code points in Latin1
12887 * that are not ASCII blanks, and add them to
12888 * the running total */
12889 _invlist_subtract(PL_Latin1, ascii_source,
12891 _invlist_union(posixes, scratch_list, &posixes);
12892 SvREFCNT_dec_NN(scratch_list);
12899 continue; /* Go get next character */
12901 } /* end of namedclass \blah */
12903 /* Here, we have a single value. If 'range' is set, it is the ending
12904 * of a range--check its validity. Later, we will handle each
12905 * individual code point in the range. If 'range' isn't set, this
12906 * could be the beginning of a range, so check for that by looking
12907 * ahead to see if the next real character to be processed is the range
12908 * indicator--the minus sign */
12911 RExC_parse = regpatws(pRExC_state, RExC_parse,
12912 FALSE /* means don't recognize comments */);
12916 if (prevvalue > value) /* b-a */ {
12917 const int w = RExC_parse - rangebegin;
12918 Simple_vFAIL4("Invalid [] range \"%*.*s\"", w, w, rangebegin);
12919 range = 0; /* not a valid range */
12923 prevvalue = value; /* save the beginning of the potential range */
12924 if (! stop_at_1 /* Can't be a range if parsing just one thing */
12925 && *RExC_parse == '-')
12927 char* next_char_ptr = RExC_parse + 1;
12928 if (skip_white) { /* Get the next real char after the '-' */
12929 next_char_ptr = regpatws(pRExC_state,
12931 FALSE); /* means don't recognize
12935 /* If the '-' is at the end of the class (just before the ']',
12936 * it is a literal minus; otherwise it is a range */
12937 if (next_char_ptr < RExC_end && *next_char_ptr != ']') {
12938 RExC_parse = next_char_ptr;
12940 /* a bad range like \w-, [:word:]- ? */
12941 if (namedclass > OOB_NAMEDCLASS) {
12942 if (strict || ckWARN(WARN_REGEXP)) {
12944 RExC_parse >= rangebegin ?
12945 RExC_parse - rangebegin : 0;
12947 vFAIL4("False [] range \"%*.*s\"",
12952 "False [] range \"%*.*s\"",
12957 cp_list = add_cp_to_invlist(cp_list, '-');
12961 range = 1; /* yeah, it's a range! */
12962 continue; /* but do it the next time */
12967 /* Here, <prevvalue> is the beginning of the range, if any; or <value>
12970 /* non-Latin1 code point implies unicode semantics. Must be set in
12971 * pass1 so is there for the whole of pass 2 */
12973 RExC_uni_semantics = 1;
12976 /* Ready to process either the single value, or the completed range.
12977 * For single-valued non-inverted ranges, we consider the possibility
12978 * of multi-char folds. (We made a conscious decision to not do this
12979 * for the other cases because it can often lead to non-intuitive
12980 * results. For example, you have the peculiar case that:
12981 * "s s" =~ /^[^\xDF]+$/i => Y
12982 * "ss" =~ /^[^\xDF]+$/i => N
12984 * See [perl #89750] */
12985 if (FOLD && allow_multi_folds && value == prevvalue) {
12986 if (value == LATIN_SMALL_LETTER_SHARP_S
12987 || (value > 255 && _invlist_contains_cp(PL_HasMultiCharFold,
12990 /* Here <value> is indeed a multi-char fold. Get what it is */
12992 U8 foldbuf[UTF8_MAXBYTES_CASE];
12995 UV folded = _to_uni_fold_flags(
13000 | ((LOC) ? FOLD_FLAGS_LOCALE
13001 : (ASCII_FOLD_RESTRICTED)
13002 ? FOLD_FLAGS_NOMIX_ASCII
13006 /* Here, <folded> should be the first character of the
13007 * multi-char fold of <value>, with <foldbuf> containing the
13008 * whole thing. But, if this fold is not allowed (because of
13009 * the flags), <fold> will be the same as <value>, and should
13010 * be processed like any other character, so skip the special
13012 if (folded != value) {
13014 /* Skip if we are recursed, currently parsing the class
13015 * again. Otherwise add this character to the list of
13016 * multi-char folds. */
13017 if (! RExC_in_multi_char_class) {
13018 AV** this_array_ptr;
13020 STRLEN cp_count = utf8_length(foldbuf,
13021 foldbuf + foldlen);
13022 SV* multi_fold = sv_2mortal(newSVpvn("", 0));
13024 Perl_sv_catpvf(aTHX_ multi_fold, "\\x{%"UVXf"}", value);
13027 if (! multi_char_matches) {
13028 multi_char_matches = newAV();
13031 /* <multi_char_matches> is actually an array of arrays.
13032 * There will be one or two top-level elements: [2],
13033 * and/or [3]. The [2] element is an array, each
13034 * element thereof is a character which folds to two
13035 * characters; likewise for [3]. (Unicode guarantees a
13036 * maximum of 3 characters in any fold.) When we
13037 * rewrite the character class below, we will do so
13038 * such that the longest folds are written first, so
13039 * that it prefers the longest matching strings first.
13040 * This is done even if it turns out that any
13041 * quantifier is non-greedy, out of programmer
13042 * laziness. Tom Christiansen has agreed that this is
13043 * ok. This makes the test for the ligature 'ffi' come
13044 * before the test for 'ff' */
13045 if (av_exists(multi_char_matches, cp_count)) {
13046 this_array_ptr = (AV**) av_fetch(multi_char_matches,
13048 this_array = *this_array_ptr;
13051 this_array = newAV();
13052 av_store(multi_char_matches, cp_count,
13055 av_push(this_array, multi_fold);
13058 /* This element should not be processed further in this
13061 value = save_value;
13062 prevvalue = save_prevvalue;
13068 /* Deal with this element of the class */
13071 cp_list = _add_range_to_invlist(cp_list, prevvalue, value);
13073 SV* this_range = _new_invlist(1);
13074 _append_range_to_invlist(this_range, prevvalue, value);
13076 /* In EBCDIC, the ranges 'A-Z' and 'a-z' are each not contiguous.
13077 * If this range was specified using something like 'i-j', we want
13078 * to include only the 'i' and the 'j', and not anything in
13079 * between, so exclude non-ASCII, non-alphabetics from it.
13080 * However, if the range was specified with something like
13081 * [\x89-\x91] or [\x89-j], all code points within it should be
13082 * included. literal_endpoint==2 means both ends of the range used
13083 * a literal character, not \x{foo} */
13084 if (literal_endpoint == 2
13085 && (prevvalue >= 'a' && value <= 'z')
13086 || (prevvalue >= 'A' && value <= 'Z'))
13088 _invlist_intersection(this_range, PL_Posix_ptrs[_CC_ALPHA],
13091 _invlist_union(cp_list, this_range, &cp_list);
13092 literal_endpoint = 0;
13096 range = 0; /* this range (if it was one) is done now */
13097 } /* End of loop through all the text within the brackets */
13099 /* If anything in the class expands to more than one character, we have to
13100 * deal with them by building up a substitute parse string, and recursively
13101 * calling reg() on it, instead of proceeding */
13102 if (multi_char_matches) {
13103 SV * substitute_parse = newSVpvn_flags("?:", 2, SVs_TEMP);
13106 char *save_end = RExC_end;
13107 char *save_parse = RExC_parse;
13108 bool first_time = TRUE; /* First multi-char occurrence doesn't get
13113 #if 0 /* Have decided not to deal with multi-char folds in inverted classes,
13114 because too confusing */
13116 sv_catpv(substitute_parse, "(?:");
13120 /* Look at the longest folds first */
13121 for (cp_count = av_len(multi_char_matches); cp_count > 0; cp_count--) {
13123 if (av_exists(multi_char_matches, cp_count)) {
13124 AV** this_array_ptr;
13127 this_array_ptr = (AV**) av_fetch(multi_char_matches,
13129 while ((this_sequence = av_pop(*this_array_ptr)) !=
13132 if (! first_time) {
13133 sv_catpv(substitute_parse, "|");
13135 first_time = FALSE;
13137 sv_catpv(substitute_parse, SvPVX(this_sequence));
13142 /* If the character class contains anything else besides these
13143 * multi-character folds, have to include it in recursive parsing */
13144 if (element_count) {
13145 sv_catpv(substitute_parse, "|[");
13146 sv_catpvn(substitute_parse, orig_parse, RExC_parse - orig_parse);
13147 sv_catpv(substitute_parse, "]");
13150 sv_catpv(substitute_parse, ")");
13153 /* This is a way to get the parse to skip forward a whole named
13154 * sequence instead of matching the 2nd character when it fails the
13156 sv_catpv(substitute_parse, "(*THEN)(*SKIP)(*FAIL)|.)");
13160 RExC_parse = SvPV(substitute_parse, len);
13161 RExC_end = RExC_parse + len;
13162 RExC_in_multi_char_class = 1;
13163 RExC_emit = (regnode *)orig_emit;
13165 ret = reg(pRExC_state, 1, ®_flags, depth+1);
13167 *flagp |= reg_flags&(HASWIDTH|SIMPLE|SPSTART|POSTPONED|RESTART_UTF8);
13169 RExC_parse = save_parse;
13170 RExC_end = save_end;
13171 RExC_in_multi_char_class = 0;
13172 SvREFCNT_dec_NN(multi_char_matches);
13176 /* If the character class contains only a single element, it may be
13177 * optimizable into another node type which is smaller and runs faster.
13178 * Check if this is the case for this class */
13179 if (element_count == 1 && ! ret_invlist) {
13183 if (namedclass > OOB_NAMEDCLASS) { /* this is a named class, like \w or
13184 [:digit:] or \p{foo} */
13186 /* All named classes are mapped into POSIXish nodes, with its FLAG
13187 * argument giving which class it is */
13188 switch ((I32)namedclass) {
13189 case ANYOF_UNIPROP:
13192 /* These don't depend on the charset modifiers. They always
13193 * match under /u rules */
13194 case ANYOF_NHORIZWS:
13195 case ANYOF_HORIZWS:
13196 namedclass = ANYOF_BLANK + namedclass - ANYOF_HORIZWS;
13199 case ANYOF_NVERTWS:
13204 /* The actual POSIXish node for all the rest depends on the
13205 * charset modifier. The ones in the first set depend only on
13206 * ASCII or, if available on this platform, locale */
13210 op = (LOC) ? POSIXL : POSIXA;
13221 /* under /a could be alpha */
13223 if (ASCII_RESTRICTED) {
13224 namedclass = ANYOF_ALPHA + (namedclass % 2);
13232 /* The rest have more possibilities depending on the charset.
13233 * We take advantage of the enum ordering of the charset
13234 * modifiers to get the exact node type, */
13236 op = POSIXD + get_regex_charset(RExC_flags);
13237 if (op > POSIXA) { /* /aa is same as /a */
13240 #ifndef HAS_ISBLANK
13242 && (namedclass == ANYOF_BLANK
13243 || namedclass == ANYOF_NBLANK))
13250 /* The odd numbered ones are the complements of the
13251 * next-lower even number one */
13252 if (namedclass % 2 == 1) {
13256 arg = namedclass_to_classnum(namedclass);
13260 else if (value == prevvalue) {
13262 /* Here, the class consists of just a single code point */
13265 if (! LOC && value == '\n') {
13266 op = REG_ANY; /* Optimize [^\n] */
13267 *flagp |= HASWIDTH|SIMPLE;
13271 else if (value < 256 || UTF) {
13273 /* Optimize a single value into an EXACTish node, but not if it
13274 * would require converting the pattern to UTF-8. */
13275 op = compute_EXACTish(pRExC_state);
13277 } /* Otherwise is a range */
13278 else if (! LOC) { /* locale could vary these */
13279 if (prevvalue == '0') {
13280 if (value == '9') {
13287 /* Here, we have changed <op> away from its initial value iff we found
13288 * an optimization */
13291 /* Throw away this ANYOF regnode, and emit the calculated one,
13292 * which should correspond to the beginning, not current, state of
13294 const char * cur_parse = RExC_parse;
13295 RExC_parse = (char *)orig_parse;
13299 /* To get locale nodes to not use the full ANYOF size would
13300 * require moving the code above that writes the portions
13301 * of it that aren't in other nodes to after this point.
13302 * e.g. ANYOF_CLASS_SET */
13303 RExC_size = orig_size;
13307 RExC_emit = (regnode *)orig_emit;
13308 if (PL_regkind[op] == POSIXD) {
13310 op += NPOSIXD - POSIXD;
13315 ret = reg_node(pRExC_state, op);
13317 if (PL_regkind[op] == POSIXD || PL_regkind[op] == NPOSIXD) {
13321 *flagp |= HASWIDTH|SIMPLE;
13323 else if (PL_regkind[op] == EXACT) {
13324 alloc_maybe_populate_EXACT(pRExC_state, ret, flagp, 0, value);
13327 RExC_parse = (char *) cur_parse;
13329 SvREFCNT_dec(posixes);
13330 SvREFCNT_dec(cp_list);
13337 /****** !SIZE_ONLY (Pass 2) AFTER HERE *********/
13339 /* If folding, we calculate all characters that could fold to or from the
13340 * ones already on the list */
13341 if (FOLD && cp_list) {
13342 UV start, end; /* End points of code point ranges */
13344 SV* fold_intersection = NULL;
13346 /* If the highest code point is within Latin1, we can use the
13347 * compiled-in Alphas list, and not have to go out to disk. This
13348 * yields two false positives, the masculine and feminine ordinal
13349 * indicators, which are weeded out below using the
13350 * IS_IN_SOME_FOLD_L1() macro */
13351 if (invlist_highest(cp_list) < 256) {
13352 _invlist_intersection(PL_L1Posix_ptrs[_CC_ALPHA], cp_list,
13353 &fold_intersection);
13357 /* Here, there are non-Latin1 code points, so we will have to go
13358 * fetch the list of all the characters that participate in folds
13360 if (! PL_utf8_foldable) {
13361 SV* swash = swash_init("utf8", "_Perl_Any_Folds",
13362 &PL_sv_undef, 1, 0);
13363 PL_utf8_foldable = _get_swash_invlist(swash);
13364 SvREFCNT_dec_NN(swash);
13367 /* This is a hash that for a particular fold gives all characters
13368 * that are involved in it */
13369 if (! PL_utf8_foldclosures) {
13371 /* If we were unable to find any folds, then we likely won't be
13372 * able to find the closures. So just create an empty list.
13373 * Folding will effectively be restricted to the non-Unicode
13374 * rules hard-coded into Perl. (This case happens legitimately
13375 * during compilation of Perl itself before the Unicode tables
13376 * are generated) */
13377 if (_invlist_len(PL_utf8_foldable) == 0) {
13378 PL_utf8_foldclosures = newHV();
13381 /* If the folds haven't been read in, call a fold function
13383 if (! PL_utf8_tofold) {
13384 U8 dummy[UTF8_MAXBYTES+1];
13386 /* This string is just a short named one above \xff */
13387 to_utf8_fold((U8*) HYPHEN_UTF8, dummy, NULL);
13388 assert(PL_utf8_tofold); /* Verify that worked */
13390 PL_utf8_foldclosures =
13391 _swash_inversion_hash(PL_utf8_tofold);
13395 /* Only the characters in this class that participate in folds need
13396 * be checked. Get the intersection of this class and all the
13397 * possible characters that are foldable. This can quickly narrow
13398 * down a large class */
13399 _invlist_intersection(PL_utf8_foldable, cp_list,
13400 &fold_intersection);
13403 /* Now look at the foldable characters in this class individually */
13404 invlist_iterinit(fold_intersection);
13405 while (invlist_iternext(fold_intersection, &start, &end)) {
13408 /* Locale folding for Latin1 characters is deferred until runtime */
13409 if (LOC && start < 256) {
13413 /* Look at every character in the range */
13414 for (j = start; j <= end; j++) {
13416 U8 foldbuf[UTF8_MAXBYTES_CASE+1];
13422 /* We have the latin1 folding rules hard-coded here so that
13423 * an innocent-looking character class, like /[ks]/i won't
13424 * have to go out to disk to find the possible matches.
13425 * XXX It would be better to generate these via regen, in
13426 * case a new version of the Unicode standard adds new
13427 * mappings, though that is not really likely, and may be
13428 * caught by the default: case of the switch below. */
13430 if (IS_IN_SOME_FOLD_L1(j)) {
13432 /* ASCII is always matched; non-ASCII is matched only
13433 * under Unicode rules */
13434 if (isASCII(j) || AT_LEAST_UNI_SEMANTICS) {
13436 add_cp_to_invlist(cp_list, PL_fold_latin1[j]);
13440 add_cp_to_invlist(depends_list, PL_fold_latin1[j]);
13444 if (HAS_NONLATIN1_FOLD_CLOSURE(j)
13445 && (! isASCII(j) || ! ASCII_FOLD_RESTRICTED))
13447 /* Certain Latin1 characters have matches outside
13448 * Latin1. To get here, <j> is one of those
13449 * characters. None of these matches is valid for
13450 * ASCII characters under /aa, which is why the 'if'
13451 * just above excludes those. These matches only
13452 * happen when the target string is utf8. The code
13453 * below adds the single fold closures for <j> to the
13454 * inversion list. */
13459 add_cp_to_invlist(cp_list, KELVIN_SIGN);
13463 cp_list = add_cp_to_invlist(cp_list,
13464 LATIN_SMALL_LETTER_LONG_S);
13467 cp_list = add_cp_to_invlist(cp_list,
13468 GREEK_CAPITAL_LETTER_MU);
13469 cp_list = add_cp_to_invlist(cp_list,
13470 GREEK_SMALL_LETTER_MU);
13472 case LATIN_CAPITAL_LETTER_A_WITH_RING_ABOVE:
13473 case LATIN_SMALL_LETTER_A_WITH_RING_ABOVE:
13475 add_cp_to_invlist(cp_list, ANGSTROM_SIGN);
13477 case LATIN_SMALL_LETTER_Y_WITH_DIAERESIS:
13478 cp_list = add_cp_to_invlist(cp_list,
13479 LATIN_CAPITAL_LETTER_Y_WITH_DIAERESIS);
13481 case LATIN_SMALL_LETTER_SHARP_S:
13482 cp_list = add_cp_to_invlist(cp_list,
13483 LATIN_CAPITAL_LETTER_SHARP_S);
13485 case 'F': case 'f':
13486 case 'I': case 'i':
13487 case 'L': case 'l':
13488 case 'T': case 't':
13489 case 'A': case 'a':
13490 case 'H': case 'h':
13491 case 'J': case 'j':
13492 case 'N': case 'n':
13493 case 'W': case 'w':
13494 case 'Y': case 'y':
13495 /* These all are targets of multi-character
13496 * folds from code points that require UTF8 to
13497 * express, so they can't match unless the
13498 * target string is in UTF-8, so no action here
13499 * is necessary, as regexec.c properly handles
13500 * the general case for UTF-8 matching and
13501 * multi-char folds */
13504 /* Use deprecated warning to increase the
13505 * chances of this being output */
13506 ckWARN2regdep(RExC_parse, "Perl folding rules are not up-to-date for 0x%"UVXf"; please use the perlbug utility to report;", j);
13513 /* Here is an above Latin1 character. We don't have the rules
13514 * hard-coded for it. First, get its fold. This is the simple
13515 * fold, as the multi-character folds have been handled earlier
13516 * and separated out */
13517 _to_uni_fold_flags(j, foldbuf, &foldlen,
13519 ? FOLD_FLAGS_LOCALE
13520 : (ASCII_FOLD_RESTRICTED)
13521 ? FOLD_FLAGS_NOMIX_ASCII
13524 /* Single character fold of above Latin1. Add everything in
13525 * its fold closure to the list that this node should match.
13526 * The fold closures data structure is a hash with the keys
13527 * being the UTF-8 of every character that is folded to, like
13528 * 'k', and the values each an array of all code points that
13529 * fold to its key. e.g. [ 'k', 'K', KELVIN_SIGN ].
13530 * Multi-character folds are not included */
13531 if ((listp = hv_fetch(PL_utf8_foldclosures,
13532 (char *) foldbuf, foldlen, FALSE)))
13534 AV* list = (AV*) *listp;
13536 for (k = 0; k <= av_len(list); k++) {
13537 SV** c_p = av_fetch(list, k, FALSE);
13540 Perl_croak(aTHX_ "panic: invalid PL_utf8_foldclosures structure");
13544 /* /aa doesn't allow folds between ASCII and non-; /l
13545 * doesn't allow them between above and below 256 */
13546 if ((ASCII_FOLD_RESTRICTED
13547 && (isASCII(c) != isASCII(j)))
13548 || (LOC && ((c < 256) != (j < 256))))
13553 /* Folds involving non-ascii Latin1 characters
13554 * under /d are added to a separate list */
13555 if (isASCII(c) || c > 255 || AT_LEAST_UNI_SEMANTICS)
13557 cp_list = add_cp_to_invlist(cp_list, c);
13560 depends_list = add_cp_to_invlist(depends_list, c);
13566 SvREFCNT_dec_NN(fold_intersection);
13569 /* And combine the result (if any) with any inversion list from posix
13570 * classes. The lists are kept separate up to now because we don't want to
13571 * fold the classes (folding of those is automatically handled by the swash
13572 * fetching code) */
13574 if (! DEPENDS_SEMANTICS) {
13576 _invlist_union(cp_list, posixes, &cp_list);
13577 SvREFCNT_dec_NN(posixes);
13584 /* Under /d, we put into a separate list the Latin1 things that
13585 * match only when the target string is utf8 */
13586 SV* nonascii_but_latin1_properties = NULL;
13587 _invlist_intersection(posixes, PL_Latin1,
13588 &nonascii_but_latin1_properties);
13589 _invlist_subtract(nonascii_but_latin1_properties, PL_ASCII,
13590 &nonascii_but_latin1_properties);
13591 _invlist_subtract(posixes, nonascii_but_latin1_properties,
13594 _invlist_union(cp_list, posixes, &cp_list);
13595 SvREFCNT_dec_NN(posixes);
13601 if (depends_list) {
13602 _invlist_union(depends_list, nonascii_but_latin1_properties,
13604 SvREFCNT_dec_NN(nonascii_but_latin1_properties);
13607 depends_list = nonascii_but_latin1_properties;
13612 /* And combine the result (if any) with any inversion list from properties.
13613 * The lists are kept separate up to now so that we can distinguish the two
13614 * in regards to matching above-Unicode. A run-time warning is generated
13615 * if a Unicode property is matched against a non-Unicode code point. But,
13616 * we allow user-defined properties to match anything, without any warning,
13617 * and we also suppress the warning if there is a portion of the character
13618 * class that isn't a Unicode property, and which matches above Unicode, \W
13619 * or [\x{110000}] for example.
13620 * (Note that in this case, unlike the Posix one above, there is no
13621 * <depends_list>, because having a Unicode property forces Unicode
13624 bool warn_super = ! has_user_defined_property;
13627 /* If it matters to the final outcome, see if a non-property
13628 * component of the class matches above Unicode. If so, the
13629 * warning gets suppressed. This is true even if just a single
13630 * such code point is specified, as though not strictly correct if
13631 * another such code point is matched against, the fact that they
13632 * are using above-Unicode code points indicates they should know
13633 * the issues involved */
13635 bool non_prop_matches_above_Unicode =
13636 runtime_posix_matches_above_Unicode
13637 | (invlist_highest(cp_list) > PERL_UNICODE_MAX);
13639 non_prop_matches_above_Unicode =
13640 ! non_prop_matches_above_Unicode;
13642 warn_super = ! non_prop_matches_above_Unicode;
13645 _invlist_union(properties, cp_list, &cp_list);
13646 SvREFCNT_dec_NN(properties);
13649 cp_list = properties;
13653 OP(ret) = ANYOF_WARN_SUPER;
13657 /* Here, we have calculated what code points should be in the character
13660 * Now we can see about various optimizations. Fold calculation (which we
13661 * did above) needs to take place before inversion. Otherwise /[^k]/i
13662 * would invert to include K, which under /i would match k, which it
13663 * shouldn't. Therefore we can't invert folded locale now, as it won't be
13664 * folded until runtime */
13666 /* Optimize inverted simple patterns (e.g. [^a-z]) when everything is known
13667 * at compile time. Besides not inverting folded locale now, we can't
13668 * invert if there are things such as \w, which aren't known until runtime
13671 && ! (LOC && (FOLD || (ANYOF_FLAGS(ret) & ANYOF_CLASS)))
13673 && ! HAS_NONLOCALE_RUNTIME_PROPERTY_DEFINITION)
13675 _invlist_invert(cp_list);
13677 /* Any swash can't be used as-is, because we've inverted things */
13679 SvREFCNT_dec_NN(swash);
13683 /* Clear the invert flag since have just done it here */
13688 *ret_invlist = cp_list;
13690 /* Discard the generated node */
13692 RExC_size = orig_size;
13695 RExC_emit = orig_emit;
13700 /* If we didn't do folding, it's because some information isn't available
13701 * until runtime; set the run-time fold flag for these. (We don't have to
13702 * worry about properties folding, as that is taken care of by the swash
13706 ANYOF_FLAGS(ret) |= ANYOF_LOC_FOLD;
13709 /* Some character classes are equivalent to other nodes. Such nodes take
13710 * up less room and generally fewer operations to execute than ANYOF nodes.
13711 * Above, we checked for and optimized into some such equivalents for
13712 * certain common classes that are easy to test. Getting to this point in
13713 * the code means that the class didn't get optimized there. Since this
13714 * code is only executed in Pass 2, it is too late to save space--it has
13715 * been allocated in Pass 1, and currently isn't given back. But turning
13716 * things into an EXACTish node can allow the optimizer to join it to any
13717 * adjacent such nodes. And if the class is equivalent to things like /./,
13718 * expensive run-time swashes can be avoided. Now that we have more
13719 * complete information, we can find things necessarily missed by the
13720 * earlier code. I (khw) am not sure how much to look for here. It would
13721 * be easy, but perhaps too slow, to check any candidates against all the
13722 * node types they could possibly match using _invlistEQ(). */
13727 && ! (ANYOF_FLAGS(ret) & ANYOF_CLASS)
13728 && ! HAS_NONLOCALE_RUNTIME_PROPERTY_DEFINITION)
13731 U8 op = END; /* The optimzation node-type */
13732 const char * cur_parse= RExC_parse;
13734 invlist_iterinit(cp_list);
13735 if (! invlist_iternext(cp_list, &start, &end)) {
13737 /* Here, the list is empty. This happens, for example, when a
13738 * Unicode property is the only thing in the character class, and
13739 * it doesn't match anything. (perluniprops.pod notes such
13742 *flagp |= HASWIDTH|SIMPLE;
13744 else if (start == end) { /* The range is a single code point */
13745 if (! invlist_iternext(cp_list, &start, &end)
13747 /* Don't do this optimization if it would require changing
13748 * the pattern to UTF-8 */
13749 && (start < 256 || UTF))
13751 /* Here, the list contains a single code point. Can optimize
13752 * into an EXACT node */
13761 /* A locale node under folding with one code point can be
13762 * an EXACTFL, as its fold won't be calculated until
13768 /* Here, we are generally folding, but there is only one
13769 * code point to match. If we have to, we use an EXACT
13770 * node, but it would be better for joining with adjacent
13771 * nodes in the optimization pass if we used the same
13772 * EXACTFish node that any such are likely to be. We can
13773 * do this iff the code point doesn't participate in any
13774 * folds. For example, an EXACTF of a colon is the same as
13775 * an EXACT one, since nothing folds to or from a colon. */
13777 if (IS_IN_SOME_FOLD_L1(value)) {
13782 if (! PL_utf8_foldable) {
13783 SV* swash = swash_init("utf8", "_Perl_Any_Folds",
13784 &PL_sv_undef, 1, 0);
13785 PL_utf8_foldable = _get_swash_invlist(swash);
13786 SvREFCNT_dec_NN(swash);
13788 if (_invlist_contains_cp(PL_utf8_foldable, value)) {
13793 /* If we haven't found the node type, above, it means we
13794 * can use the prevailing one */
13796 op = compute_EXACTish(pRExC_state);
13801 else if (start == 0) {
13802 if (end == UV_MAX) {
13804 *flagp |= HASWIDTH|SIMPLE;
13807 else if (end == '\n' - 1
13808 && invlist_iternext(cp_list, &start, &end)
13809 && start == '\n' + 1 && end == UV_MAX)
13812 *flagp |= HASWIDTH|SIMPLE;
13816 invlist_iterfinish(cp_list);
13819 RExC_parse = (char *)orig_parse;
13820 RExC_emit = (regnode *)orig_emit;
13822 ret = reg_node(pRExC_state, op);
13824 RExC_parse = (char *)cur_parse;
13826 if (PL_regkind[op] == EXACT) {
13827 alloc_maybe_populate_EXACT(pRExC_state, ret, flagp, 0, value);
13830 SvREFCNT_dec_NN(cp_list);
13835 /* Here, <cp_list> contains all the code points we can determine at
13836 * compile time that match under all conditions. Go through it, and
13837 * for things that belong in the bitmap, put them there, and delete from
13838 * <cp_list>. While we are at it, see if everything above 255 is in the
13839 * list, and if so, set a flag to speed up execution */
13840 ANYOF_BITMAP_ZERO(ret);
13843 /* This gets set if we actually need to modify things */
13844 bool change_invlist = FALSE;
13848 /* Start looking through <cp_list> */
13849 invlist_iterinit(cp_list);
13850 while (invlist_iternext(cp_list, &start, &end)) {
13854 if (end == UV_MAX && start <= 256) {
13855 ANYOF_FLAGS(ret) |= ANYOF_UNICODE_ALL;
13858 /* Quit if are above what we should change */
13863 change_invlist = TRUE;
13865 /* Set all the bits in the range, up to the max that we are doing */
13866 high = (end < 255) ? end : 255;
13867 for (i = start; i <= (int) high; i++) {
13868 if (! ANYOF_BITMAP_TEST(ret, i)) {
13869 ANYOF_BITMAP_SET(ret, i);
13875 invlist_iterfinish(cp_list);
13877 /* Done with loop; remove any code points that are in the bitmap from
13879 if (change_invlist) {
13880 _invlist_subtract(cp_list, PL_Latin1, &cp_list);
13883 /* If have completely emptied it, remove it completely */
13884 if (_invlist_len(cp_list) == 0) {
13885 SvREFCNT_dec_NN(cp_list);
13891 ANYOF_FLAGS(ret) |= ANYOF_INVERT;
13894 /* Here, the bitmap has been populated with all the Latin1 code points that
13895 * always match. Can now add to the overall list those that match only
13896 * when the target string is UTF-8 (<depends_list>). */
13897 if (depends_list) {
13899 _invlist_union(cp_list, depends_list, &cp_list);
13900 SvREFCNT_dec_NN(depends_list);
13903 cp_list = depends_list;
13907 /* If there is a swash and more than one element, we can't use the swash in
13908 * the optimization below. */
13909 if (swash && element_count > 1) {
13910 SvREFCNT_dec_NN(swash);
13915 && ! HAS_NONLOCALE_RUNTIME_PROPERTY_DEFINITION)
13917 ARG_SET(ret, ANYOF_NONBITMAP_EMPTY);
13920 /* av[0] stores the character class description in its textual form:
13921 * used later (regexec.c:Perl_regclass_swash()) to initialize the
13922 * appropriate swash, and is also useful for dumping the regnode.
13923 * av[1] if NULL, is a placeholder to later contain the swash computed
13924 * from av[0]. But if no further computation need be done, the
13925 * swash is stored there now.
13926 * av[2] stores the cp_list inversion list for use in addition or
13927 * instead of av[0]; used only if av[1] is NULL
13928 * av[3] is set if any component of the class is from a user-defined
13929 * property; used only if av[1] is NULL */
13930 AV * const av = newAV();
13933 av_store(av, 0, (HAS_NONLOCALE_RUNTIME_PROPERTY_DEFINITION)
13934 ? SvREFCNT_inc(listsv) : &PL_sv_undef);
13936 av_store(av, 1, swash);
13937 SvREFCNT_dec_NN(cp_list);
13940 av_store(av, 1, NULL);
13942 av_store(av, 2, cp_list);
13943 av_store(av, 3, newSVuv(has_user_defined_property));
13947 rv = newRV_noinc(MUTABLE_SV(av));
13948 n = add_data(pRExC_state, 1, "s");
13949 RExC_rxi->data->data[n] = (void*)rv;
13953 *flagp |= HASWIDTH|SIMPLE;
13956 #undef HAS_NONLOCALE_RUNTIME_PROPERTY_DEFINITION
13959 /* reg_skipcomment()
13961 Absorbs an /x style # comments from the input stream.
13962 Returns true if there is more text remaining in the stream.
13963 Will set the REG_SEEN_RUN_ON_COMMENT flag if the comment
13964 terminates the pattern without including a newline.
13966 Note its the callers responsibility to ensure that we are
13967 actually in /x mode
13972 S_reg_skipcomment(pTHX_ RExC_state_t *pRExC_state)
13976 PERL_ARGS_ASSERT_REG_SKIPCOMMENT;
13978 while (RExC_parse < RExC_end)
13979 if (*RExC_parse++ == '\n') {
13984 /* we ran off the end of the pattern without ending
13985 the comment, so we have to add an \n when wrapping */
13986 RExC_seen |= REG_SEEN_RUN_ON_COMMENT;
13994 Advances the parse position, and optionally absorbs
13995 "whitespace" from the inputstream.
13997 Without /x "whitespace" means (?#...) style comments only,
13998 with /x this means (?#...) and # comments and whitespace proper.
14000 Returns the RExC_parse point from BEFORE the scan occurs.
14002 This is the /x friendly way of saying RExC_parse++.
14006 S_nextchar(pTHX_ RExC_state_t *pRExC_state)
14008 char* const retval = RExC_parse++;
14010 PERL_ARGS_ASSERT_NEXTCHAR;
14013 if (RExC_end - RExC_parse >= 3
14014 && *RExC_parse == '('
14015 && RExC_parse[1] == '?'
14016 && RExC_parse[2] == '#')
14018 while (*RExC_parse != ')') {
14019 if (RExC_parse == RExC_end)
14020 FAIL("Sequence (?#... not terminated");
14026 if (RExC_flags & RXf_PMf_EXTENDED) {
14027 if (isSPACE(*RExC_parse)) {
14031 else if (*RExC_parse == '#') {
14032 if ( reg_skipcomment( pRExC_state ) )
14041 - reg_node - emit a node
14043 STATIC regnode * /* Location. */
14044 S_reg_node(pTHX_ RExC_state_t *pRExC_state, U8 op)
14048 regnode * const ret = RExC_emit;
14049 GET_RE_DEBUG_FLAGS_DECL;
14051 PERL_ARGS_ASSERT_REG_NODE;
14054 SIZE_ALIGN(RExC_size);
14058 if (RExC_emit >= RExC_emit_bound)
14059 Perl_croak(aTHX_ "panic: reg_node overrun trying to emit %d, %p>=%p",
14060 op, RExC_emit, RExC_emit_bound);
14062 NODE_ALIGN_FILL(ret);
14064 FILL_ADVANCE_NODE(ptr, op);
14065 REH_CALL_COMP_NODE_HOOK(pRExC_state->rx, (ptr) - 1);
14066 #ifdef RE_TRACK_PATTERN_OFFSETS
14067 if (RExC_offsets) { /* MJD */
14068 MJD_OFFSET_DEBUG(("%s:%d: (op %s) %s %"UVuf" (len %"UVuf") (max %"UVuf").\n",
14069 "reg_node", __LINE__,
14071 (UV)(RExC_emit - RExC_emit_start) > RExC_offsets[0]
14072 ? "Overwriting end of array!\n" : "OK",
14073 (UV)(RExC_emit - RExC_emit_start),
14074 (UV)(RExC_parse - RExC_start),
14075 (UV)RExC_offsets[0]));
14076 Set_Node_Offset(RExC_emit, RExC_parse + (op == END));
14084 - reganode - emit a node with an argument
14086 STATIC regnode * /* Location. */
14087 S_reganode(pTHX_ RExC_state_t *pRExC_state, U8 op, U32 arg)
14091 regnode * const ret = RExC_emit;
14092 GET_RE_DEBUG_FLAGS_DECL;
14094 PERL_ARGS_ASSERT_REGANODE;
14097 SIZE_ALIGN(RExC_size);
14102 assert(2==regarglen[op]+1);
14104 Anything larger than this has to allocate the extra amount.
14105 If we changed this to be:
14107 RExC_size += (1 + regarglen[op]);
14109 then it wouldn't matter. Its not clear what side effect
14110 might come from that so its not done so far.
14115 if (RExC_emit >= RExC_emit_bound)
14116 Perl_croak(aTHX_ "panic: reg_node overrun trying to emit %d, %p>=%p",
14117 op, RExC_emit, RExC_emit_bound);
14119 NODE_ALIGN_FILL(ret);
14121 FILL_ADVANCE_NODE_ARG(ptr, op, arg);
14122 REH_CALL_COMP_NODE_HOOK(pRExC_state->rx, (ptr) - 2);
14123 #ifdef RE_TRACK_PATTERN_OFFSETS
14124 if (RExC_offsets) { /* MJD */
14125 MJD_OFFSET_DEBUG(("%s(%d): (op %s) %s %"UVuf" <- %"UVuf" (max %"UVuf").\n",
14129 (UV)(RExC_emit - RExC_emit_start) > RExC_offsets[0] ?
14130 "Overwriting end of array!\n" : "OK",
14131 (UV)(RExC_emit - RExC_emit_start),
14132 (UV)(RExC_parse - RExC_start),
14133 (UV)RExC_offsets[0]));
14134 Set_Cur_Node_Offset;
14142 - reguni - emit (if appropriate) a Unicode character
14145 S_reguni(pTHX_ const RExC_state_t *pRExC_state, UV uv, char* s)
14149 PERL_ARGS_ASSERT_REGUNI;
14151 return SIZE_ONLY ? UNISKIP(uv) : (uvchr_to_utf8((U8*)s, uv) - (U8*)s);
14155 - reginsert - insert an operator in front of already-emitted operand
14157 * Means relocating the operand.
14160 S_reginsert(pTHX_ RExC_state_t *pRExC_state, U8 op, regnode *opnd, U32 depth)
14166 const int offset = regarglen[(U8)op];
14167 const int size = NODE_STEP_REGNODE + offset;
14168 GET_RE_DEBUG_FLAGS_DECL;
14170 PERL_ARGS_ASSERT_REGINSERT;
14171 PERL_UNUSED_ARG(depth);
14172 /* (PL_regkind[(U8)op] == CURLY ? EXTRA_STEP_2ARGS : 0); */
14173 DEBUG_PARSE_FMT("inst"," - %s",PL_reg_name[op]);
14182 if (RExC_open_parens) {
14184 /*DEBUG_PARSE_FMT("inst"," - %"IVdf, (IV)RExC_npar);*/
14185 for ( paren=0 ; paren < RExC_npar ; paren++ ) {
14186 if ( RExC_open_parens[paren] >= opnd ) {
14187 /*DEBUG_PARSE_FMT("open"," - %d",size);*/
14188 RExC_open_parens[paren] += size;
14190 /*DEBUG_PARSE_FMT("open"," - %s","ok");*/
14192 if ( RExC_close_parens[paren] >= opnd ) {
14193 /*DEBUG_PARSE_FMT("close"," - %d",size);*/
14194 RExC_close_parens[paren] += size;
14196 /*DEBUG_PARSE_FMT("close"," - %s","ok");*/
14201 while (src > opnd) {
14202 StructCopy(--src, --dst, regnode);
14203 #ifdef RE_TRACK_PATTERN_OFFSETS
14204 if (RExC_offsets) { /* MJD 20010112 */
14205 MJD_OFFSET_DEBUG(("%s(%d): (op %s) %s copy %"UVuf" -> %"UVuf" (max %"UVuf").\n",
14209 (UV)(dst - RExC_emit_start) > RExC_offsets[0]
14210 ? "Overwriting end of array!\n" : "OK",
14211 (UV)(src - RExC_emit_start),
14212 (UV)(dst - RExC_emit_start),
14213 (UV)RExC_offsets[0]));
14214 Set_Node_Offset_To_R(dst-RExC_emit_start, Node_Offset(src));
14215 Set_Node_Length_To_R(dst-RExC_emit_start, Node_Length(src));
14221 place = opnd; /* Op node, where operand used to be. */
14222 #ifdef RE_TRACK_PATTERN_OFFSETS
14223 if (RExC_offsets) { /* MJD */
14224 MJD_OFFSET_DEBUG(("%s(%d): (op %s) %s %"UVuf" <- %"UVuf" (max %"UVuf").\n",
14228 (UV)(place - RExC_emit_start) > RExC_offsets[0]
14229 ? "Overwriting end of array!\n" : "OK",
14230 (UV)(place - RExC_emit_start),
14231 (UV)(RExC_parse - RExC_start),
14232 (UV)RExC_offsets[0]));
14233 Set_Node_Offset(place, RExC_parse);
14234 Set_Node_Length(place, 1);
14237 src = NEXTOPER(place);
14238 FILL_ADVANCE_NODE(place, op);
14239 REH_CALL_COMP_NODE_HOOK(pRExC_state->rx, (place) - 1);
14240 Zero(src, offset, regnode);
14244 - regtail - set the next-pointer at the end of a node chain of p to val.
14245 - SEE ALSO: regtail_study
14247 /* TODO: All three parms should be const */
14249 S_regtail(pTHX_ RExC_state_t *pRExC_state, regnode *p, const regnode *val,U32 depth)
14253 GET_RE_DEBUG_FLAGS_DECL;
14255 PERL_ARGS_ASSERT_REGTAIL;
14257 PERL_UNUSED_ARG(depth);
14263 /* Find last node. */
14266 regnode * const temp = regnext(scan);
14268 SV * const mysv=sv_newmortal();
14269 DEBUG_PARSE_MSG((scan==p ? "tail" : ""));
14270 regprop(RExC_rx, mysv, scan);
14271 PerlIO_printf(Perl_debug_log, "~ %s (%d) %s %s\n",
14272 SvPV_nolen_const(mysv), REG_NODE_NUM(scan),
14273 (temp == NULL ? "->" : ""),
14274 (temp == NULL ? PL_reg_name[OP(val)] : "")
14282 if (reg_off_by_arg[OP(scan)]) {
14283 ARG_SET(scan, val - scan);
14286 NEXT_OFF(scan) = val - scan;
14292 - regtail_study - set the next-pointer at the end of a node chain of p to val.
14293 - Look for optimizable sequences at the same time.
14294 - currently only looks for EXACT chains.
14296 This is experimental code. The idea is to use this routine to perform
14297 in place optimizations on branches and groups as they are constructed,
14298 with the long term intention of removing optimization from study_chunk so
14299 that it is purely analytical.
14301 Currently only used when in DEBUG mode. The macro REGTAIL_STUDY() is used
14302 to control which is which.
14305 /* TODO: All four parms should be const */
14308 S_regtail_study(pTHX_ RExC_state_t *pRExC_state, regnode *p, const regnode *val,U32 depth)
14313 #ifdef EXPERIMENTAL_INPLACESCAN
14316 GET_RE_DEBUG_FLAGS_DECL;
14318 PERL_ARGS_ASSERT_REGTAIL_STUDY;
14324 /* Find last node. */
14328 regnode * const temp = regnext(scan);
14329 #ifdef EXPERIMENTAL_INPLACESCAN
14330 if (PL_regkind[OP(scan)] == EXACT) {
14331 bool has_exactf_sharp_s; /* Unexamined in this routine */
14332 if (join_exact(pRExC_state,scan,&min, &has_exactf_sharp_s, 1,val,depth+1))
14337 switch (OP(scan)) {
14343 case EXACTFU_TRICKYFOLD:
14345 if( exact == PSEUDO )
14347 else if ( exact != OP(scan) )
14356 SV * const mysv=sv_newmortal();
14357 DEBUG_PARSE_MSG((scan==p ? "tsdy" : ""));
14358 regprop(RExC_rx, mysv, scan);
14359 PerlIO_printf(Perl_debug_log, "~ %s (%d) -> %s\n",
14360 SvPV_nolen_const(mysv),
14361 REG_NODE_NUM(scan),
14362 PL_reg_name[exact]);
14369 SV * const mysv_val=sv_newmortal();
14370 DEBUG_PARSE_MSG("");
14371 regprop(RExC_rx, mysv_val, val);
14372 PerlIO_printf(Perl_debug_log, "~ attach to %s (%"IVdf") offset to %"IVdf"\n",
14373 SvPV_nolen_const(mysv_val),
14374 (IV)REG_NODE_NUM(val),
14378 if (reg_off_by_arg[OP(scan)]) {
14379 ARG_SET(scan, val - scan);
14382 NEXT_OFF(scan) = val - scan;
14390 - regdump - dump a regexp onto Perl_debug_log in vaguely comprehensible form
14394 S_regdump_extflags(pTHX_ const char *lead, const U32 flags)
14400 for (bit=0; bit<32; bit++) {
14401 if (flags & (1<<bit)) {
14402 if ((1<<bit) & RXf_PMf_CHARSET) { /* Output separately, below */
14405 if (!set++ && lead)
14406 PerlIO_printf(Perl_debug_log, "%s",lead);
14407 PerlIO_printf(Perl_debug_log, "%s ",PL_reg_extflags_name[bit]);
14410 if ((cs = get_regex_charset(flags)) != REGEX_DEPENDS_CHARSET) {
14411 if (!set++ && lead) {
14412 PerlIO_printf(Perl_debug_log, "%s",lead);
14415 case REGEX_UNICODE_CHARSET:
14416 PerlIO_printf(Perl_debug_log, "UNICODE");
14418 case REGEX_LOCALE_CHARSET:
14419 PerlIO_printf(Perl_debug_log, "LOCALE");
14421 case REGEX_ASCII_RESTRICTED_CHARSET:
14422 PerlIO_printf(Perl_debug_log, "ASCII-RESTRICTED");
14424 case REGEX_ASCII_MORE_RESTRICTED_CHARSET:
14425 PerlIO_printf(Perl_debug_log, "ASCII-MORE_RESTRICTED");
14428 PerlIO_printf(Perl_debug_log, "UNKNOWN CHARACTER SET");
14434 PerlIO_printf(Perl_debug_log, "\n");
14436 PerlIO_printf(Perl_debug_log, "%s[none-set]\n",lead);
14442 Perl_regdump(pTHX_ const regexp *r)
14446 SV * const sv = sv_newmortal();
14447 SV *dsv= sv_newmortal();
14448 RXi_GET_DECL(r,ri);
14449 GET_RE_DEBUG_FLAGS_DECL;
14451 PERL_ARGS_ASSERT_REGDUMP;
14453 (void)dumpuntil(r, ri->program, ri->program + 1, NULL, NULL, sv, 0, 0);
14455 /* Header fields of interest. */
14456 if (r->anchored_substr) {
14457 RE_PV_QUOTED_DECL(s, 0, dsv, SvPVX_const(r->anchored_substr),
14458 RE_SV_DUMPLEN(r->anchored_substr), 30);
14459 PerlIO_printf(Perl_debug_log,
14460 "anchored %s%s at %"IVdf" ",
14461 s, RE_SV_TAIL(r->anchored_substr),
14462 (IV)r->anchored_offset);
14463 } else if (r->anchored_utf8) {
14464 RE_PV_QUOTED_DECL(s, 1, dsv, SvPVX_const(r->anchored_utf8),
14465 RE_SV_DUMPLEN(r->anchored_utf8), 30);
14466 PerlIO_printf(Perl_debug_log,
14467 "anchored utf8 %s%s at %"IVdf" ",
14468 s, RE_SV_TAIL(r->anchored_utf8),
14469 (IV)r->anchored_offset);
14471 if (r->float_substr) {
14472 RE_PV_QUOTED_DECL(s, 0, dsv, SvPVX_const(r->float_substr),
14473 RE_SV_DUMPLEN(r->float_substr), 30);
14474 PerlIO_printf(Perl_debug_log,
14475 "floating %s%s at %"IVdf"..%"UVuf" ",
14476 s, RE_SV_TAIL(r->float_substr),
14477 (IV)r->float_min_offset, (UV)r->float_max_offset);
14478 } else if (r->float_utf8) {
14479 RE_PV_QUOTED_DECL(s, 1, dsv, SvPVX_const(r->float_utf8),
14480 RE_SV_DUMPLEN(r->float_utf8), 30);
14481 PerlIO_printf(Perl_debug_log,
14482 "floating utf8 %s%s at %"IVdf"..%"UVuf" ",
14483 s, RE_SV_TAIL(r->float_utf8),
14484 (IV)r->float_min_offset, (UV)r->float_max_offset);
14486 if (r->check_substr || r->check_utf8)
14487 PerlIO_printf(Perl_debug_log,
14489 (r->check_substr == r->float_substr
14490 && r->check_utf8 == r->float_utf8
14491 ? "(checking floating" : "(checking anchored"));
14492 if (r->extflags & RXf_NOSCAN)
14493 PerlIO_printf(Perl_debug_log, " noscan");
14494 if (r->extflags & RXf_CHECK_ALL)
14495 PerlIO_printf(Perl_debug_log, " isall");
14496 if (r->check_substr || r->check_utf8)
14497 PerlIO_printf(Perl_debug_log, ") ");
14499 if (ri->regstclass) {
14500 regprop(r, sv, ri->regstclass);
14501 PerlIO_printf(Perl_debug_log, "stclass %s ", SvPVX_const(sv));
14503 if (r->extflags & RXf_ANCH) {
14504 PerlIO_printf(Perl_debug_log, "anchored");
14505 if (r->extflags & RXf_ANCH_BOL)
14506 PerlIO_printf(Perl_debug_log, "(BOL)");
14507 if (r->extflags & RXf_ANCH_MBOL)
14508 PerlIO_printf(Perl_debug_log, "(MBOL)");
14509 if (r->extflags & RXf_ANCH_SBOL)
14510 PerlIO_printf(Perl_debug_log, "(SBOL)");
14511 if (r->extflags & RXf_ANCH_GPOS)
14512 PerlIO_printf(Perl_debug_log, "(GPOS)");
14513 PerlIO_putc(Perl_debug_log, ' ');
14515 if (r->extflags & RXf_GPOS_SEEN)
14516 PerlIO_printf(Perl_debug_log, "GPOS:%"UVuf" ", (UV)r->gofs);
14517 if (r->intflags & PREGf_SKIP)
14518 PerlIO_printf(Perl_debug_log, "plus ");
14519 if (r->intflags & PREGf_IMPLICIT)
14520 PerlIO_printf(Perl_debug_log, "implicit ");
14521 PerlIO_printf(Perl_debug_log, "minlen %"IVdf" ", (IV)r->minlen);
14522 if (r->extflags & RXf_EVAL_SEEN)
14523 PerlIO_printf(Perl_debug_log, "with eval ");
14524 PerlIO_printf(Perl_debug_log, "\n");
14525 DEBUG_FLAGS_r(regdump_extflags("r->extflags: ",r->extflags));
14527 PERL_ARGS_ASSERT_REGDUMP;
14528 PERL_UNUSED_CONTEXT;
14529 PERL_UNUSED_ARG(r);
14530 #endif /* DEBUGGING */
14534 - regprop - printable representation of opcode
14536 #define EMIT_ANYOF_TEST_SEPARATOR(do_sep,sv,flags) \
14539 Perl_sv_catpvf(aTHX_ sv,"%s][%s",PL_colors[1],PL_colors[0]); \
14540 if (flags & ANYOF_INVERT) \
14541 /*make sure the invert info is in each */ \
14542 sv_catpvs(sv, "^"); \
14548 Perl_regprop(pTHX_ const regexp *prog, SV *sv, const regnode *o)
14554 /* Should be synchronized with * ANYOF_ #xdefines in regcomp.h */
14555 static const char * const anyofs[] = {
14556 #if _CC_WORDCHAR != 0 || _CC_DIGIT != 1 || _CC_ALPHA != 2 || _CC_LOWER != 3 \
14557 || _CC_UPPER != 4 || _CC_PUNCT != 5 || _CC_PRINT != 6 \
14558 || _CC_ALPHANUMERIC != 7 || _CC_GRAPH != 8 || _CC_CASED != 9 \
14559 || _CC_SPACE != 10 || _CC_BLANK != 11 || _CC_XDIGIT != 12 \
14560 || _CC_PSXSPC != 13 || _CC_CNTRL != 14 || _CC_ASCII != 15 \
14561 || _CC_VERTSPACE != 16
14562 #error Need to adjust order of anyofs[]
14599 RXi_GET_DECL(prog,progi);
14600 GET_RE_DEBUG_FLAGS_DECL;
14602 PERL_ARGS_ASSERT_REGPROP;
14606 if (OP(o) > REGNODE_MAX) /* regnode.type is unsigned */
14607 /* It would be nice to FAIL() here, but this may be called from
14608 regexec.c, and it would be hard to supply pRExC_state. */
14609 Perl_croak(aTHX_ "Corrupted regexp opcode %d > %d", (int)OP(o), (int)REGNODE_MAX);
14610 sv_catpv(sv, PL_reg_name[OP(o)]); /* Take off const! */
14612 k = PL_regkind[OP(o)];
14615 sv_catpvs(sv, " ");
14616 /* Using is_utf8_string() (via PERL_PV_UNI_DETECT)
14617 * is a crude hack but it may be the best for now since
14618 * we have no flag "this EXACTish node was UTF-8"
14620 pv_pretty(sv, STRING(o), STR_LEN(o), 60, PL_colors[0], PL_colors[1],
14621 PERL_PV_ESCAPE_UNI_DETECT |
14622 PERL_PV_ESCAPE_NONASCII |
14623 PERL_PV_PRETTY_ELLIPSES |
14624 PERL_PV_PRETTY_LTGT |
14625 PERL_PV_PRETTY_NOCLEAR
14627 } else if (k == TRIE) {
14628 /* print the details of the trie in dumpuntil instead, as
14629 * progi->data isn't available here */
14630 const char op = OP(o);
14631 const U32 n = ARG(o);
14632 const reg_ac_data * const ac = IS_TRIE_AC(op) ?
14633 (reg_ac_data *)progi->data->data[n] :
14635 const reg_trie_data * const trie
14636 = (reg_trie_data*)progi->data->data[!IS_TRIE_AC(op) ? n : ac->trie];
14638 Perl_sv_catpvf(aTHX_ sv, "-%s",PL_reg_name[o->flags]);
14639 DEBUG_TRIE_COMPILE_r(
14640 Perl_sv_catpvf(aTHX_ sv,
14641 "<S:%"UVuf"/%"IVdf" W:%"UVuf" L:%"UVuf"/%"UVuf" C:%"UVuf"/%"UVuf">",
14642 (UV)trie->startstate,
14643 (IV)trie->statecount-1, /* -1 because of the unused 0 element */
14644 (UV)trie->wordcount,
14647 (UV)TRIE_CHARCOUNT(trie),
14648 (UV)trie->uniquecharcount
14651 if ( IS_ANYOF_TRIE(op) || trie->bitmap ) {
14653 int rangestart = -1;
14654 U8* bitmap = IS_ANYOF_TRIE(op) ? (U8*)ANYOF_BITMAP(o) : (U8*)TRIE_BITMAP(trie);
14655 sv_catpvs(sv, "[");
14656 for (i = 0; i <= 256; i++) {
14657 if (i < 256 && BITMAP_TEST(bitmap,i)) {
14658 if (rangestart == -1)
14660 } else if (rangestart != -1) {
14661 if (i <= rangestart + 3)
14662 for (; rangestart < i; rangestart++)
14663 put_byte(sv, rangestart);
14665 put_byte(sv, rangestart);
14666 sv_catpvs(sv, "-");
14667 put_byte(sv, i - 1);
14672 sv_catpvs(sv, "]");
14675 } else if (k == CURLY) {
14676 if (OP(o) == CURLYM || OP(o) == CURLYN || OP(o) == CURLYX)
14677 Perl_sv_catpvf(aTHX_ sv, "[%d]", o->flags); /* Parenth number */
14678 Perl_sv_catpvf(aTHX_ sv, " {%d,%d}", ARG1(o), ARG2(o));
14680 else if (k == WHILEM && o->flags) /* Ordinal/of */
14681 Perl_sv_catpvf(aTHX_ sv, "[%d/%d]", o->flags & 0xf, o->flags>>4);
14682 else if (k == REF || k == OPEN || k == CLOSE || k == GROUPP || OP(o)==ACCEPT) {
14683 Perl_sv_catpvf(aTHX_ sv, "%d", (int)ARG(o)); /* Parenth number */
14684 if ( RXp_PAREN_NAMES(prog) ) {
14685 if ( k != REF || (OP(o) < NREF)) {
14686 AV *list= MUTABLE_AV(progi->data->data[progi->name_list_idx]);
14687 SV **name= av_fetch(list, ARG(o), 0 );
14689 Perl_sv_catpvf(aTHX_ sv, " '%"SVf"'", SVfARG(*name));
14692 AV *list= MUTABLE_AV(progi->data->data[ progi->name_list_idx ]);
14693 SV *sv_dat= MUTABLE_SV(progi->data->data[ ARG( o ) ]);
14694 I32 *nums=(I32*)SvPVX(sv_dat);
14695 SV **name= av_fetch(list, nums[0], 0 );
14698 for ( n=0; n<SvIVX(sv_dat); n++ ) {
14699 Perl_sv_catpvf(aTHX_ sv, "%s%"IVdf,
14700 (n ? "," : ""), (IV)nums[n]);
14702 Perl_sv_catpvf(aTHX_ sv, " '%"SVf"'", SVfARG(*name));
14706 } else if (k == GOSUB)
14707 Perl_sv_catpvf(aTHX_ sv, "%d[%+d]", (int)ARG(o),(int)ARG2L(o)); /* Paren and offset */
14708 else if (k == VERB) {
14710 Perl_sv_catpvf(aTHX_ sv, ":%"SVf,
14711 SVfARG((MUTABLE_SV(progi->data->data[ ARG( o ) ]))));
14712 } else if (k == LOGICAL)
14713 Perl_sv_catpvf(aTHX_ sv, "[%d]", o->flags); /* 2: embedded, otherwise 1 */
14714 else if (k == ANYOF) {
14715 int i, rangestart = -1;
14716 const U8 flags = ANYOF_FLAGS(o);
14720 if (flags & ANYOF_LOCALE)
14721 sv_catpvs(sv, "{loc}");
14722 if (flags & ANYOF_LOC_FOLD)
14723 sv_catpvs(sv, "{i}");
14724 Perl_sv_catpvf(aTHX_ sv, "[%s", PL_colors[0]);
14725 if (flags & ANYOF_INVERT)
14726 sv_catpvs(sv, "^");
14728 /* output what the standard cp 0-255 bitmap matches */
14729 for (i = 0; i <= 256; i++) {
14730 if (i < 256 && ANYOF_BITMAP_TEST(o,i)) {
14731 if (rangestart == -1)
14733 } else if (rangestart != -1) {
14734 if (i <= rangestart + 3)
14735 for (; rangestart < i; rangestart++)
14736 put_byte(sv, rangestart);
14738 put_byte(sv, rangestart);
14739 sv_catpvs(sv, "-");
14740 put_byte(sv, i - 1);
14747 EMIT_ANYOF_TEST_SEPARATOR(do_sep,sv,flags);
14748 /* output any special charclass tests (used entirely under use locale) */
14749 if (ANYOF_CLASS_TEST_ANY_SET(o))
14750 for (i = 0; i < (int)(sizeof(anyofs)/sizeof(char*)); i++)
14751 if (ANYOF_CLASS_TEST(o,i)) {
14752 sv_catpv(sv, anyofs[i]);
14756 EMIT_ANYOF_TEST_SEPARATOR(do_sep,sv,flags);
14758 if (flags & ANYOF_NON_UTF8_LATIN1_ALL) {
14759 sv_catpvs(sv, "{non-utf8-latin1-all}");
14762 /* output information about the unicode matching */
14763 if (flags & ANYOF_UNICODE_ALL)
14764 sv_catpvs(sv, "{unicode_all}");
14765 else if (ANYOF_NONBITMAP(o))
14766 sv_catpvs(sv, "{unicode}");
14767 if (flags & ANYOF_NONBITMAP_NON_UTF8)
14768 sv_catpvs(sv, "{outside bitmap}");
14770 if (ANYOF_NONBITMAP(o)) {
14771 SV *lv; /* Set if there is something outside the bit map */
14772 SV * const sw = regclass_swash(prog, o, FALSE, &lv, NULL);
14773 bool byte_output = FALSE; /* If something in the bitmap has been
14776 if (lv && lv != &PL_sv_undef) {
14778 U8 s[UTF8_MAXBYTES_CASE+1];
14780 for (i = 0; i <= 256; i++) { /* Look at chars in bitmap */
14781 uvchr_to_utf8(s, i);
14784 && ! ANYOF_BITMAP_TEST(o, i) /* Don't duplicate
14788 && swash_fetch(sw, s, TRUE))
14790 if (rangestart == -1)
14792 } else if (rangestart != -1) {
14793 byte_output = TRUE;
14794 if (i <= rangestart + 3)
14795 for (; rangestart < i; rangestart++) {
14796 put_byte(sv, rangestart);
14799 put_byte(sv, rangestart);
14800 sv_catpvs(sv, "-");
14809 char *s = savesvpv(lv);
14810 char * const origs = s;
14812 while (*s && *s != '\n')
14816 const char * const t = ++s;
14819 sv_catpvs(sv, " ");
14825 /* Truncate very long output */
14826 if (s - origs > 256) {
14827 Perl_sv_catpvf(aTHX_ sv,
14829 (int) (s - origs - 1),
14835 else if (*s == '\t') {
14850 SvREFCNT_dec_NN(lv);
14854 Perl_sv_catpvf(aTHX_ sv, "%s]", PL_colors[1]);
14856 else if (k == POSIXD || k == NPOSIXD) {
14857 U8 index = FLAGS(o) * 2;
14858 if (index > (sizeof(anyofs) / sizeof(anyofs[0]))) {
14859 Perl_sv_catpvf(aTHX_ sv, "[illegal type=%d])", index);
14862 sv_catpv(sv, anyofs[index]);
14865 else if (k == BRANCHJ && (OP(o) == UNLESSM || OP(o) == IFMATCH))
14866 Perl_sv_catpvf(aTHX_ sv, "[%d]", -(o->flags));
14868 PERL_UNUSED_CONTEXT;
14869 PERL_UNUSED_ARG(sv);
14870 PERL_UNUSED_ARG(o);
14871 PERL_UNUSED_ARG(prog);
14872 #endif /* DEBUGGING */
14876 Perl_re_intuit_string(pTHX_ REGEXP * const r)
14877 { /* Assume that RE_INTUIT is set */
14879 struct regexp *const prog = ReANY(r);
14880 GET_RE_DEBUG_FLAGS_DECL;
14882 PERL_ARGS_ASSERT_RE_INTUIT_STRING;
14883 PERL_UNUSED_CONTEXT;
14887 const char * const s = SvPV_nolen_const(prog->check_substr
14888 ? prog->check_substr : prog->check_utf8);
14890 if (!PL_colorset) reginitcolors();
14891 PerlIO_printf(Perl_debug_log,
14892 "%sUsing REx %ssubstr:%s \"%s%.60s%s%s\"\n",
14894 prog->check_substr ? "" : "utf8 ",
14895 PL_colors[5],PL_colors[0],
14898 (strlen(s) > 60 ? "..." : ""));
14901 return prog->check_substr ? prog->check_substr : prog->check_utf8;
14907 handles refcounting and freeing the perl core regexp structure. When
14908 it is necessary to actually free the structure the first thing it
14909 does is call the 'free' method of the regexp_engine associated to
14910 the regexp, allowing the handling of the void *pprivate; member
14911 first. (This routine is not overridable by extensions, which is why
14912 the extensions free is called first.)
14914 See regdupe and regdupe_internal if you change anything here.
14916 #ifndef PERL_IN_XSUB_RE
14918 Perl_pregfree(pTHX_ REGEXP *r)
14924 Perl_pregfree2(pTHX_ REGEXP *rx)
14927 struct regexp *const r = ReANY(rx);
14928 GET_RE_DEBUG_FLAGS_DECL;
14930 PERL_ARGS_ASSERT_PREGFREE2;
14932 if (r->mother_re) {
14933 ReREFCNT_dec(r->mother_re);
14935 CALLREGFREE_PVT(rx); /* free the private data */
14936 SvREFCNT_dec(RXp_PAREN_NAMES(r));
14937 Safefree(r->xpv_len_u.xpvlenu_pv);
14940 SvREFCNT_dec(r->anchored_substr);
14941 SvREFCNT_dec(r->anchored_utf8);
14942 SvREFCNT_dec(r->float_substr);
14943 SvREFCNT_dec(r->float_utf8);
14944 Safefree(r->substrs);
14946 RX_MATCH_COPY_FREE(rx);
14947 #ifdef PERL_ANY_COW
14948 SvREFCNT_dec(r->saved_copy);
14951 SvREFCNT_dec(r->qr_anoncv);
14952 rx->sv_u.svu_rx = 0;
14957 This is a hacky workaround to the structural issue of match results
14958 being stored in the regexp structure which is in turn stored in
14959 PL_curpm/PL_reg_curpm. The problem is that due to qr// the pattern
14960 could be PL_curpm in multiple contexts, and could require multiple
14961 result sets being associated with the pattern simultaneously, such
14962 as when doing a recursive match with (??{$qr})
14964 The solution is to make a lightweight copy of the regexp structure
14965 when a qr// is returned from the code executed by (??{$qr}) this
14966 lightweight copy doesn't actually own any of its data except for
14967 the starp/end and the actual regexp structure itself.
14973 Perl_reg_temp_copy (pTHX_ REGEXP *ret_x, REGEXP *rx)
14975 struct regexp *ret;
14976 struct regexp *const r = ReANY(rx);
14977 const bool islv = ret_x && SvTYPE(ret_x) == SVt_PVLV;
14979 PERL_ARGS_ASSERT_REG_TEMP_COPY;
14982 ret_x = (REGEXP*) newSV_type(SVt_REGEXP);
14984 SvOK_off((SV *)ret_x);
14986 /* For PVLVs, SvANY points to the xpvlv body while sv_u points
14987 to the regexp. (For SVt_REGEXPs, sv_upgrade has already
14988 made both spots point to the same regexp body.) */
14989 REGEXP *temp = (REGEXP *)newSV_type(SVt_REGEXP);
14990 assert(!SvPVX(ret_x));
14991 ret_x->sv_u.svu_rx = temp->sv_any;
14992 temp->sv_any = NULL;
14993 SvFLAGS(temp) = (SvFLAGS(temp) & ~SVTYPEMASK) | SVt_NULL;
14994 SvREFCNT_dec_NN(temp);
14995 /* SvCUR still resides in the xpvlv struct, so the regexp copy-
14996 ing below will not set it. */
14997 SvCUR_set(ret_x, SvCUR(rx));
15000 /* This ensures that SvTHINKFIRST(sv) is true, and hence that
15001 sv_force_normal(sv) is called. */
15003 ret = ReANY(ret_x);
15005 SvFLAGS(ret_x) |= SvUTF8(rx);
15006 /* We share the same string buffer as the original regexp, on which we
15007 hold a reference count, incremented when mother_re is set below.
15008 The string pointer is copied here, being part of the regexp struct.
15010 memcpy(&(ret->xpv_cur), &(r->xpv_cur),
15011 sizeof(regexp) - STRUCT_OFFSET(regexp, xpv_cur));
15013 const I32 npar = r->nparens+1;
15014 Newx(ret->offs, npar, regexp_paren_pair);
15015 Copy(r->offs, ret->offs, npar, regexp_paren_pair);
15018 Newx(ret->substrs, 1, struct reg_substr_data);
15019 StructCopy(r->substrs, ret->substrs, struct reg_substr_data);
15021 SvREFCNT_inc_void(ret->anchored_substr);
15022 SvREFCNT_inc_void(ret->anchored_utf8);
15023 SvREFCNT_inc_void(ret->float_substr);
15024 SvREFCNT_inc_void(ret->float_utf8);
15026 /* check_substr and check_utf8, if non-NULL, point to either their
15027 anchored or float namesakes, and don't hold a second reference. */
15029 RX_MATCH_COPIED_off(ret_x);
15030 #ifdef PERL_ANY_COW
15031 ret->saved_copy = NULL;
15033 ret->mother_re = ReREFCNT_inc(r->mother_re ? r->mother_re : rx);
15034 SvREFCNT_inc_void(ret->qr_anoncv);
15040 /* regfree_internal()
15042 Free the private data in a regexp. This is overloadable by
15043 extensions. Perl takes care of the regexp structure in pregfree(),
15044 this covers the *pprivate pointer which technically perl doesn't
15045 know about, however of course we have to handle the
15046 regexp_internal structure when no extension is in use.
15048 Note this is called before freeing anything in the regexp
15053 Perl_regfree_internal(pTHX_ REGEXP * const rx)
15056 struct regexp *const r = ReANY(rx);
15057 RXi_GET_DECL(r,ri);
15058 GET_RE_DEBUG_FLAGS_DECL;
15060 PERL_ARGS_ASSERT_REGFREE_INTERNAL;
15066 SV *dsv= sv_newmortal();
15067 RE_PV_QUOTED_DECL(s, RX_UTF8(rx),
15068 dsv, RX_PRECOMP(rx), RX_PRELEN(rx), 60);
15069 PerlIO_printf(Perl_debug_log,"%sFreeing REx:%s %s\n",
15070 PL_colors[4],PL_colors[5],s);
15073 #ifdef RE_TRACK_PATTERN_OFFSETS
15075 Safefree(ri->u.offsets); /* 20010421 MJD */
15077 if (ri->code_blocks) {
15079 for (n = 0; n < ri->num_code_blocks; n++)
15080 SvREFCNT_dec(ri->code_blocks[n].src_regex);
15081 Safefree(ri->code_blocks);
15085 int n = ri->data->count;
15088 /* If you add a ->what type here, update the comment in regcomp.h */
15089 switch (ri->data->what[n]) {
15095 SvREFCNT_dec(MUTABLE_SV(ri->data->data[n]));
15098 Safefree(ri->data->data[n]);
15104 { /* Aho Corasick add-on structure for a trie node.
15105 Used in stclass optimization only */
15107 reg_ac_data *aho=(reg_ac_data*)ri->data->data[n];
15109 refcount = --aho->refcount;
15112 PerlMemShared_free(aho->states);
15113 PerlMemShared_free(aho->fail);
15114 /* do this last!!!! */
15115 PerlMemShared_free(ri->data->data[n]);
15116 PerlMemShared_free(ri->regstclass);
15122 /* trie structure. */
15124 reg_trie_data *trie=(reg_trie_data*)ri->data->data[n];
15126 refcount = --trie->refcount;
15129 PerlMemShared_free(trie->charmap);
15130 PerlMemShared_free(trie->states);
15131 PerlMemShared_free(trie->trans);
15133 PerlMemShared_free(trie->bitmap);
15135 PerlMemShared_free(trie->jump);
15136 PerlMemShared_free(trie->wordinfo);
15137 /* do this last!!!! */
15138 PerlMemShared_free(ri->data->data[n]);
15143 Perl_croak(aTHX_ "panic: regfree data code '%c'", ri->data->what[n]);
15146 Safefree(ri->data->what);
15147 Safefree(ri->data);
15153 #define av_dup_inc(s,t) MUTABLE_AV(sv_dup_inc((const SV *)s,t))
15154 #define hv_dup_inc(s,t) MUTABLE_HV(sv_dup_inc((const SV *)s,t))
15155 #define SAVEPVN(p,n) ((p) ? savepvn(p,n) : NULL)
15158 re_dup - duplicate a regexp.
15160 This routine is expected to clone a given regexp structure. It is only
15161 compiled under USE_ITHREADS.
15163 After all of the core data stored in struct regexp is duplicated
15164 the regexp_engine.dupe method is used to copy any private data
15165 stored in the *pprivate pointer. This allows extensions to handle
15166 any duplication it needs to do.
15168 See pregfree() and regfree_internal() if you change anything here.
15170 #if defined(USE_ITHREADS)
15171 #ifndef PERL_IN_XSUB_RE
15173 Perl_re_dup_guts(pTHX_ const REGEXP *sstr, REGEXP *dstr, CLONE_PARAMS *param)
15177 const struct regexp *r = ReANY(sstr);
15178 struct regexp *ret = ReANY(dstr);
15180 PERL_ARGS_ASSERT_RE_DUP_GUTS;
15182 npar = r->nparens+1;
15183 Newx(ret->offs, npar, regexp_paren_pair);
15184 Copy(r->offs, ret->offs, npar, regexp_paren_pair);
15186 if (ret->substrs) {
15187 /* Do it this way to avoid reading from *r after the StructCopy().
15188 That way, if any of the sv_dup_inc()s dislodge *r from the L1
15189 cache, it doesn't matter. */
15190 const bool anchored = r->check_substr
15191 ? r->check_substr == r->anchored_substr
15192 : r->check_utf8 == r->anchored_utf8;
15193 Newx(ret->substrs, 1, struct reg_substr_data);
15194 StructCopy(r->substrs, ret->substrs, struct reg_substr_data);
15196 ret->anchored_substr = sv_dup_inc(ret->anchored_substr, param);
15197 ret->anchored_utf8 = sv_dup_inc(ret->anchored_utf8, param);
15198 ret->float_substr = sv_dup_inc(ret->float_substr, param);
15199 ret->float_utf8 = sv_dup_inc(ret->float_utf8, param);
15201 /* check_substr and check_utf8, if non-NULL, point to either their
15202 anchored or float namesakes, and don't hold a second reference. */
15204 if (ret->check_substr) {
15206 assert(r->check_utf8 == r->anchored_utf8);
15207 ret->check_substr = ret->anchored_substr;
15208 ret->check_utf8 = ret->anchored_utf8;
15210 assert(r->check_substr == r->float_substr);
15211 assert(r->check_utf8 == r->float_utf8);
15212 ret->check_substr = ret->float_substr;
15213 ret->check_utf8 = ret->float_utf8;
15215 } else if (ret->check_utf8) {
15217 ret->check_utf8 = ret->anchored_utf8;
15219 ret->check_utf8 = ret->float_utf8;
15224 RXp_PAREN_NAMES(ret) = hv_dup_inc(RXp_PAREN_NAMES(ret), param);
15225 ret->qr_anoncv = MUTABLE_CV(sv_dup_inc((const SV *)ret->qr_anoncv, param));
15228 RXi_SET(ret,CALLREGDUPE_PVT(dstr,param));
15230 if (RX_MATCH_COPIED(dstr))
15231 ret->subbeg = SAVEPVN(ret->subbeg, ret->sublen);
15233 ret->subbeg = NULL;
15234 #ifdef PERL_ANY_COW
15235 ret->saved_copy = NULL;
15238 /* Whether mother_re be set or no, we need to copy the string. We
15239 cannot refrain from copying it when the storage points directly to
15240 our mother regexp, because that's
15241 1: a buffer in a different thread
15242 2: something we no longer hold a reference on
15243 so we need to copy it locally. */
15244 RX_WRAPPED(dstr) = SAVEPVN(RX_WRAPPED(sstr), SvCUR(sstr)+1);
15245 ret->mother_re = NULL;
15248 #endif /* PERL_IN_XSUB_RE */
15253 This is the internal complement to regdupe() which is used to copy
15254 the structure pointed to by the *pprivate pointer in the regexp.
15255 This is the core version of the extension overridable cloning hook.
15256 The regexp structure being duplicated will be copied by perl prior
15257 to this and will be provided as the regexp *r argument, however
15258 with the /old/ structures pprivate pointer value. Thus this routine
15259 may override any copying normally done by perl.
15261 It returns a pointer to the new regexp_internal structure.
15265 Perl_regdupe_internal(pTHX_ REGEXP * const rx, CLONE_PARAMS *param)
15268 struct regexp *const r = ReANY(rx);
15269 regexp_internal *reti;
15271 RXi_GET_DECL(r,ri);
15273 PERL_ARGS_ASSERT_REGDUPE_INTERNAL;
15277 Newxc(reti, sizeof(regexp_internal) + len*sizeof(regnode), char, regexp_internal);
15278 Copy(ri->program, reti->program, len+1, regnode);
15280 reti->num_code_blocks = ri->num_code_blocks;
15281 if (ri->code_blocks) {
15283 Newxc(reti->code_blocks, ri->num_code_blocks, struct reg_code_block,
15284 struct reg_code_block);
15285 Copy(ri->code_blocks, reti->code_blocks, ri->num_code_blocks,
15286 struct reg_code_block);
15287 for (n = 0; n < ri->num_code_blocks; n++)
15288 reti->code_blocks[n].src_regex = (REGEXP*)
15289 sv_dup_inc((SV*)(ri->code_blocks[n].src_regex), param);
15292 reti->code_blocks = NULL;
15294 reti->regstclass = NULL;
15297 struct reg_data *d;
15298 const int count = ri->data->count;
15301 Newxc(d, sizeof(struct reg_data) + count*sizeof(void *),
15302 char, struct reg_data);
15303 Newx(d->what, count, U8);
15306 for (i = 0; i < count; i++) {
15307 d->what[i] = ri->data->what[i];
15308 switch (d->what[i]) {
15309 /* see also regcomp.h and regfree_internal() */
15310 case 'a': /* actually an AV, but the dup function is identical. */
15314 case 'u': /* actually an HV, but the dup function is identical. */
15315 d->data[i] = sv_dup_inc((const SV *)ri->data->data[i], param);
15318 /* This is cheating. */
15319 Newx(d->data[i], 1, struct regnode_charclass_class);
15320 StructCopy(ri->data->data[i], d->data[i],
15321 struct regnode_charclass_class);
15322 reti->regstclass = (regnode*)d->data[i];
15325 /* Trie stclasses are readonly and can thus be shared
15326 * without duplication. We free the stclass in pregfree
15327 * when the corresponding reg_ac_data struct is freed.
15329 reti->regstclass= ri->regstclass;
15333 ((reg_trie_data*)ri->data->data[i])->refcount++;
15338 d->data[i] = ri->data->data[i];
15341 Perl_croak(aTHX_ "panic: re_dup unknown data code '%c'", ri->data->what[i]);
15350 reti->name_list_idx = ri->name_list_idx;
15352 #ifdef RE_TRACK_PATTERN_OFFSETS
15353 if (ri->u.offsets) {
15354 Newx(reti->u.offsets, 2*len+1, U32);
15355 Copy(ri->u.offsets, reti->u.offsets, 2*len+1, U32);
15358 SetProgLen(reti,len);
15361 return (void*)reti;
15364 #endif /* USE_ITHREADS */
15366 #ifndef PERL_IN_XSUB_RE
15369 - regnext - dig the "next" pointer out of a node
15372 Perl_regnext(pTHX_ regnode *p)
15380 if (OP(p) > REGNODE_MAX) { /* regnode.type is unsigned */
15381 Perl_croak(aTHX_ "Corrupted regexp opcode %d > %d", (int)OP(p), (int)REGNODE_MAX);
15384 offset = (reg_off_by_arg[OP(p)] ? ARG(p) : NEXT_OFF(p));
15393 S_re_croak2(pTHX_ const char* pat1,const char* pat2,...)
15396 STRLEN l1 = strlen(pat1);
15397 STRLEN l2 = strlen(pat2);
15400 const char *message;
15402 PERL_ARGS_ASSERT_RE_CROAK2;
15408 Copy(pat1, buf, l1 , char);
15409 Copy(pat2, buf + l1, l2 , char);
15410 buf[l1 + l2] = '\n';
15411 buf[l1 + l2 + 1] = '\0';
15413 /* ANSI variant takes additional second argument */
15414 va_start(args, pat2);
15418 msv = vmess(buf, &args);
15420 message = SvPV_const(msv,l1);
15423 Copy(message, buf, l1 , char);
15424 buf[l1-1] = '\0'; /* Overwrite \n */
15425 Perl_croak(aTHX_ "%s", buf);
15428 /* XXX Here's a total kludge. But we need to re-enter for swash routines. */
15430 #ifndef PERL_IN_XSUB_RE
15432 Perl_save_re_context(pTHX)
15436 struct re_save_state *state;
15438 SAVEVPTR(PL_curcop);
15439 SSGROW(SAVESTACK_ALLOC_FOR_RE_SAVE_STATE + 1);
15441 state = (struct re_save_state *)(PL_savestack + PL_savestack_ix);
15442 PL_savestack_ix += SAVESTACK_ALLOC_FOR_RE_SAVE_STATE;
15443 SSPUSHUV(SAVEt_RE_STATE);
15445 Copy(&PL_reg_state, state, 1, struct re_save_state);
15447 PL_reg_oldsaved = NULL;
15448 PL_reg_oldsavedlen = 0;
15449 PL_reg_oldsavedoffset = 0;
15450 PL_reg_oldsavedcoffset = 0;
15451 PL_reg_maxiter = 0;
15452 PL_reg_leftiter = 0;
15453 PL_reg_poscache = NULL;
15454 PL_reg_poscache_size = 0;
15455 #ifdef PERL_ANY_COW
15459 /* Save $1..$n (#18107: UTF-8 s/(\w+)/uc($1)/e); AMS 20021106. */
15461 const REGEXP * const rx = PM_GETRE(PL_curpm);
15464 for (i = 1; i <= RX_NPARENS(rx); i++) {
15465 char digits[TYPE_CHARS(long)];
15466 const STRLEN len = my_snprintf(digits, sizeof(digits), "%lu", (long)i);
15467 GV *const *const gvp
15468 = (GV**)hv_fetch(PL_defstash, digits, len, 0);
15471 GV * const gv = *gvp;
15472 if (SvTYPE(gv) == SVt_PVGV && GvSV(gv))
15484 S_put_byte(pTHX_ SV *sv, int c)
15486 PERL_ARGS_ASSERT_PUT_BYTE;
15488 /* Our definition of isPRINT() ignores locales, so only bytes that are
15489 not part of UTF-8 are considered printable. I assume that the same
15490 holds for UTF-EBCDIC.
15491 Also, code point 255 is not printable in either (it's E0 in EBCDIC,
15492 which Wikipedia says:
15494 EO, or Eight Ones, is an 8-bit EBCDIC character code represented as all
15495 ones (binary 1111 1111, hexadecimal FF). It is similar, but not
15496 identical, to the ASCII delete (DEL) or rubout control character. ...
15497 it is typically mapped to hexadecimal code 9F, in order to provide a
15498 unique character mapping in both directions)
15500 So the old condition can be simplified to !isPRINT(c) */
15503 Perl_sv_catpvf(aTHX_ sv, "\\x%02x", c);
15506 Perl_sv_catpvf(aTHX_ sv, "\\x{%x}", c);
15510 const char string = c;
15511 if (c == '-' || c == ']' || c == '\\' || c == '^')
15512 sv_catpvs(sv, "\\");
15513 sv_catpvn(sv, &string, 1);
15518 #define CLEAR_OPTSTART \
15519 if (optstart) STMT_START { \
15520 DEBUG_OPTIMISE_r(PerlIO_printf(Perl_debug_log, " (%"IVdf" nodes)\n", (IV)(node - optstart))); \
15524 #define DUMPUNTIL(b,e) CLEAR_OPTSTART; node=dumpuntil(r,start,(b),(e),last,sv,indent+1,depth+1);
15526 STATIC const regnode *
15527 S_dumpuntil(pTHX_ const regexp *r, const regnode *start, const regnode *node,
15528 const regnode *last, const regnode *plast,
15529 SV* sv, I32 indent, U32 depth)
15532 U8 op = PSEUDO; /* Arbitrary non-END op. */
15533 const regnode *next;
15534 const regnode *optstart= NULL;
15536 RXi_GET_DECL(r,ri);
15537 GET_RE_DEBUG_FLAGS_DECL;
15539 PERL_ARGS_ASSERT_DUMPUNTIL;
15541 #ifdef DEBUG_DUMPUNTIL
15542 PerlIO_printf(Perl_debug_log, "--- %d : %d - %d - %d\n",indent,node-start,
15543 last ? last-start : 0,plast ? plast-start : 0);
15546 if (plast && plast < last)
15549 while (PL_regkind[op] != END && (!last || node < last)) {
15550 /* While that wasn't END last time... */
15553 if (op == CLOSE || op == WHILEM)
15555 next = regnext((regnode *)node);
15558 if (OP(node) == OPTIMIZED) {
15559 if (!optstart && RE_DEBUG_FLAG(RE_DEBUG_COMPILE_OPTIMISE))
15566 regprop(r, sv, node);
15567 PerlIO_printf(Perl_debug_log, "%4"IVdf":%*s%s", (IV)(node - start),
15568 (int)(2*indent + 1), "", SvPVX_const(sv));
15570 if (OP(node) != OPTIMIZED) {
15571 if (next == NULL) /* Next ptr. */
15572 PerlIO_printf(Perl_debug_log, " (0)");
15573 else if (PL_regkind[(U8)op] == BRANCH && PL_regkind[OP(next)] != BRANCH )
15574 PerlIO_printf(Perl_debug_log, " (FAIL)");
15576 PerlIO_printf(Perl_debug_log, " (%"IVdf")", (IV)(next - start));
15577 (void)PerlIO_putc(Perl_debug_log, '\n');
15581 if (PL_regkind[(U8)op] == BRANCHJ) {
15584 const regnode *nnode = (OP(next) == LONGJMP
15585 ? regnext((regnode *)next)
15587 if (last && nnode > last)
15589 DUMPUNTIL(NEXTOPER(NEXTOPER(node)), nnode);
15592 else if (PL_regkind[(U8)op] == BRANCH) {
15594 DUMPUNTIL(NEXTOPER(node), next);
15596 else if ( PL_regkind[(U8)op] == TRIE ) {
15597 const regnode *this_trie = node;
15598 const char op = OP(node);
15599 const U32 n = ARG(node);
15600 const reg_ac_data * const ac = op>=AHOCORASICK ?
15601 (reg_ac_data *)ri->data->data[n] :
15603 const reg_trie_data * const trie =
15604 (reg_trie_data*)ri->data->data[op<AHOCORASICK ? n : ac->trie];
15606 AV *const trie_words = MUTABLE_AV(ri->data->data[n + TRIE_WORDS_OFFSET]);
15608 const regnode *nextbranch= NULL;
15611 for (word_idx= 0; word_idx < (I32)trie->wordcount; word_idx++) {
15612 SV ** const elem_ptr = av_fetch(trie_words,word_idx,0);
15614 PerlIO_printf(Perl_debug_log, "%*s%s ",
15615 (int)(2*(indent+3)), "",
15616 elem_ptr ? pv_pretty(sv, SvPV_nolen_const(*elem_ptr), SvCUR(*elem_ptr), 60,
15617 PL_colors[0], PL_colors[1],
15618 (SvUTF8(*elem_ptr) ? PERL_PV_ESCAPE_UNI : 0) |
15619 PERL_PV_PRETTY_ELLIPSES |
15620 PERL_PV_PRETTY_LTGT
15625 U16 dist= trie->jump[word_idx+1];
15626 PerlIO_printf(Perl_debug_log, "(%"UVuf")\n",
15627 (UV)((dist ? this_trie + dist : next) - start));
15630 nextbranch= this_trie + trie->jump[0];
15631 DUMPUNTIL(this_trie + dist, nextbranch);
15633 if (nextbranch && PL_regkind[OP(nextbranch)]==BRANCH)
15634 nextbranch= regnext((regnode *)nextbranch);
15636 PerlIO_printf(Perl_debug_log, "\n");
15639 if (last && next > last)
15644 else if ( op == CURLY ) { /* "next" might be very big: optimizer */
15645 DUMPUNTIL(NEXTOPER(node) + EXTRA_STEP_2ARGS,
15646 NEXTOPER(node) + EXTRA_STEP_2ARGS + 1);
15648 else if (PL_regkind[(U8)op] == CURLY && op != CURLYX) {
15650 DUMPUNTIL(NEXTOPER(node) + EXTRA_STEP_2ARGS, next);
15652 else if ( op == PLUS || op == STAR) {
15653 DUMPUNTIL(NEXTOPER(node), NEXTOPER(node) + 1);
15655 else if (PL_regkind[(U8)op] == ANYOF) {
15656 /* arglen 1 + class block */
15657 node += 1 + ((ANYOF_FLAGS(node) & ANYOF_CLASS)
15658 ? ANYOF_CLASS_SKIP : ANYOF_SKIP);
15659 node = NEXTOPER(node);
15661 else if (PL_regkind[(U8)op] == EXACT) {
15662 /* Literal string, where present. */
15663 node += NODE_SZ_STR(node) - 1;
15664 node = NEXTOPER(node);
15667 node = NEXTOPER(node);
15668 node += regarglen[(U8)op];
15670 if (op == CURLYX || op == OPEN)
15674 #ifdef DEBUG_DUMPUNTIL
15675 PerlIO_printf(Perl_debug_log, "--- %d\n", (int)indent);
15680 #endif /* DEBUGGING */
15684 * c-indentation-style: bsd
15685 * c-basic-offset: 4
15686 * indent-tabs-mode: nil
15689 * ex: set ts=8 sts=4 sw=4 et: