5 * 'A fair jaw-cracker dwarf-language must be.' --Samwise Gamgee
7 * [p.285 of _The Lord of the Rings_, II/iii: "The Ring Goes South"]
10 /* This file contains functions for compiling a regular expression. See
11 * also regexec.c which funnily enough, contains functions for executing
12 * a regular expression.
14 * This file is also copied at build time to ext/re/re_comp.c, where
15 * it's built with -DPERL_EXT_RE_BUILD -DPERL_EXT_RE_DEBUG -DPERL_EXT.
16 * This causes the main functions to be compiled under new names and with
17 * debugging support added, which makes "use re 'debug'" work.
20 /* NOTE: this is derived from Henry Spencer's regexp code, and should not
21 * confused with the original package (see point 3 below). Thanks, Henry!
24 /* Additional note: this code is very heavily munged from Henry's version
25 * in places. In some spots I've traded clarity for efficiency, so don't
26 * blame Henry for some of the lack of readability.
29 /* The names of the functions have been changed from regcomp and
30 * regexec to pregcomp and pregexec in order to avoid conflicts
31 * with the POSIX routines of the same names.
34 #ifdef PERL_EXT_RE_BUILD
39 * pregcomp and pregexec -- regsub and regerror are not used in perl
41 * Copyright (c) 1986 by University of Toronto.
42 * Written by Henry Spencer. Not derived from licensed software.
44 * Permission is granted to anyone to use this software for any
45 * purpose on any computer system, and to redistribute it freely,
46 * subject to the following restrictions:
48 * 1. The author is not responsible for the consequences of use of
49 * this software, no matter how awful, even if they arise
52 * 2. The origin of this software must not be misrepresented, either
53 * by explicit claim or by omission.
55 * 3. Altered versions must be plainly marked as such, and must not
56 * be misrepresented as being the original software.
59 **** Alterations to Henry's code are...
61 **** Copyright (C) 1991, 1992, 1993, 1994, 1995, 1996, 1997, 1998, 1999,
62 **** 2000, 2001, 2002, 2003, 2004, 2005, 2006, 2007, 2008
63 **** by Larry Wall and others
65 **** You may distribute under the terms of either the GNU General Public
66 **** License or the Artistic License, as specified in the README file.
69 * Beware that some of this code is subtly aware of the way operator
70 * precedence is structured in regular expressions. Serious changes in
71 * regular-expression syntax might require a total rethink.
74 #define PERL_IN_REGCOMP_C
77 #ifndef PERL_IN_XSUB_RE
82 #ifdef PERL_IN_XSUB_RE
84 EXTERN_C const struct regexp_engine my_reg_engine;
89 #include "dquote_static.c"
90 #include "inline_invlist.c"
91 #include "unicode_constants.h"
93 #define HAS_NONLATIN1_FOLD_CLOSURE(i) \
94 _HAS_NONLATIN1_FOLD_CLOSURE_ONLY_FOR_USE_BY_REGCOMP_DOT_C_AND_REGEXEC_DOT_C(i)
95 #define HAS_NONLATIN1_SIMPLE_FOLD_CLOSURE(i) \
96 _HAS_NONLATIN1_SIMPLE_FOLD_CLOSURE_ONLY_FOR_USE_BY_REGCOMP_DOT_C_AND_REGEXEC_DOT_C(i)
97 #define IS_NON_FINAL_FOLD(c) _IS_NON_FINAL_FOLD_ONLY_FOR_USE_BY_REGCOMP_DOT_C(c)
98 #define IS_IN_SOME_FOLD_L1(c) _IS_IN_SOME_FOLD_ONLY_FOR_USE_BY_REGCOMP_DOT_C(c)
101 #define STATIC static
105 #define MIN(a,b) ((a) < (b) ? (a) : (b))
108 /* this is a chain of data about sub patterns we are processing that
109 need to be handled separately/specially in study_chunk. Its so
110 we can simulate recursion without losing state. */
112 typedef struct scan_frame {
113 regnode *last_regnode; /* last node to process in this frame */
114 regnode *next_regnode; /* next node to process when last is reached */
115 U32 prev_recursed_depth;
116 I32 stopparen; /* what stopparen do we use */
117 U32 is_top_frame; /* what flags do we use? */
119 struct scan_frame *this_prev_frame; /* this previous frame */
120 struct scan_frame *prev_frame; /* previous frame */
121 struct scan_frame *next_frame; /* next frame */
124 /* Certain characters are output as a sequence with the first being a
126 #define isBACKSLASHED_PUNCT(c) \
127 ((c) == '-' || (c) == ']' || (c) == '\\' || (c) == '^')
130 struct RExC_state_t {
131 U32 flags; /* RXf_* are we folding, multilining? */
132 U32 pm_flags; /* PMf_* stuff from the calling PMOP */
133 char *precomp; /* uncompiled string. */
134 REGEXP *rx_sv; /* The SV that is the regexp. */
135 regexp *rx; /* perl core regexp structure */
136 regexp_internal *rxi; /* internal data for regexp object
138 char *start; /* Start of input for compile */
139 char *end; /* End of input for compile */
140 char *parse; /* Input-scan pointer. */
141 SSize_t whilem_seen; /* number of WHILEM in this expr */
142 regnode *emit_start; /* Start of emitted-code area */
143 regnode *emit_bound; /* First regnode outside of the
145 regnode *emit; /* Code-emit pointer; if = &emit_dummy,
146 implies compiling, so don't emit */
147 regnode_ssc emit_dummy; /* placeholder for emit to point to;
148 large enough for the largest
149 non-EXACTish node, so can use it as
151 I32 naughty; /* How bad is this pattern? */
152 I32 sawback; /* Did we see \1, ...? */
154 SSize_t size; /* Code size. */
155 I32 npar; /* Capture buffer count, (OPEN) plus
156 one. ("par" 0 is the whole
158 I32 nestroot; /* root parens we are in - used by
162 regnode **open_parens; /* pointers to open parens */
163 regnode **close_parens; /* pointers to close parens */
164 regnode *opend; /* END node in program */
165 I32 utf8; /* whether the pattern is utf8 or not */
166 I32 orig_utf8; /* whether the pattern was originally in utf8 */
167 /* XXX use this for future optimisation of case
168 * where pattern must be upgraded to utf8. */
169 I32 uni_semantics; /* If a d charset modifier should use unicode
170 rules, even if the pattern is not in
172 HV *paren_names; /* Paren names */
174 regnode **recurse; /* Recurse regops */
175 I32 recurse_count; /* Number of recurse regops */
176 U8 *study_chunk_recursed; /* bitmap of which subs we have moved
178 U32 study_chunk_recursed_bytes; /* bytes in bitmap */
182 I32 override_recoding;
183 I32 in_multi_char_class;
184 struct reg_code_block *code_blocks; /* positions of literal (?{})
186 int num_code_blocks; /* size of code_blocks[] */
187 int code_index; /* next code_blocks[] slot */
188 SSize_t maxlen; /* mininum possible number of chars in string to match */
189 scan_frame *frame_head;
190 scan_frame *frame_last;
193 #ifdef ADD_TO_REGEXEC
194 char *starttry; /* -Dr: where regtry was called. */
195 #define RExC_starttry (pRExC_state->starttry)
197 SV *runtime_code_qr; /* qr with the runtime code blocks */
199 const char *lastparse;
201 AV *paren_name_list; /* idx -> name */
202 U32 study_chunk_recursed_count;
205 #define RExC_lastparse (pRExC_state->lastparse)
206 #define RExC_lastnum (pRExC_state->lastnum)
207 #define RExC_paren_name_list (pRExC_state->paren_name_list)
208 #define RExC_study_chunk_recursed_count (pRExC_state->study_chunk_recursed_count)
209 #define RExC_mysv (pRExC_state->mysv1)
210 #define RExC_mysv1 (pRExC_state->mysv1)
211 #define RExC_mysv2 (pRExC_state->mysv2)
216 #define RExC_flags (pRExC_state->flags)
217 #define RExC_pm_flags (pRExC_state->pm_flags)
218 #define RExC_precomp (pRExC_state->precomp)
219 #define RExC_rx_sv (pRExC_state->rx_sv)
220 #define RExC_rx (pRExC_state->rx)
221 #define RExC_rxi (pRExC_state->rxi)
222 #define RExC_start (pRExC_state->start)
223 #define RExC_end (pRExC_state->end)
224 #define RExC_parse (pRExC_state->parse)
225 #define RExC_whilem_seen (pRExC_state->whilem_seen)
226 #ifdef RE_TRACK_PATTERN_OFFSETS
227 #define RExC_offsets (pRExC_state->rxi->u.offsets) /* I am not like the
230 #define RExC_emit (pRExC_state->emit)
231 #define RExC_emit_dummy (pRExC_state->emit_dummy)
232 #define RExC_emit_start (pRExC_state->emit_start)
233 #define RExC_emit_bound (pRExC_state->emit_bound)
234 #define RExC_sawback (pRExC_state->sawback)
235 #define RExC_seen (pRExC_state->seen)
236 #define RExC_size (pRExC_state->size)
237 #define RExC_maxlen (pRExC_state->maxlen)
238 #define RExC_npar (pRExC_state->npar)
239 #define RExC_nestroot (pRExC_state->nestroot)
240 #define RExC_extralen (pRExC_state->extralen)
241 #define RExC_seen_zerolen (pRExC_state->seen_zerolen)
242 #define RExC_utf8 (pRExC_state->utf8)
243 #define RExC_uni_semantics (pRExC_state->uni_semantics)
244 #define RExC_orig_utf8 (pRExC_state->orig_utf8)
245 #define RExC_open_parens (pRExC_state->open_parens)
246 #define RExC_close_parens (pRExC_state->close_parens)
247 #define RExC_opend (pRExC_state->opend)
248 #define RExC_paren_names (pRExC_state->paren_names)
249 #define RExC_recurse (pRExC_state->recurse)
250 #define RExC_recurse_count (pRExC_state->recurse_count)
251 #define RExC_study_chunk_recursed (pRExC_state->study_chunk_recursed)
252 #define RExC_study_chunk_recursed_bytes \
253 (pRExC_state->study_chunk_recursed_bytes)
254 #define RExC_in_lookbehind (pRExC_state->in_lookbehind)
255 #define RExC_contains_locale (pRExC_state->contains_locale)
256 #define RExC_contains_i (pRExC_state->contains_i)
257 #define RExC_override_recoding (pRExC_state->override_recoding)
258 #define RExC_in_multi_char_class (pRExC_state->in_multi_char_class)
259 #define RExC_frame_head (pRExC_state->frame_head)
260 #define RExC_frame_last (pRExC_state->frame_last)
261 #define RExC_frame_count (pRExC_state->frame_count)
262 #define RExC_strict (pRExC_state->strict)
264 /* Heuristic check on the complexity of the pattern: if TOO_NAUGHTY, we set
265 * a flag to disable back-off on the fixed/floating substrings - if it's
266 * a high complexity pattern we assume the benefit of avoiding a full match
267 * is worth the cost of checking for the substrings even if they rarely help.
269 #define RExC_naughty (pRExC_state->naughty)
270 #define TOO_NAUGHTY (10)
271 #define MARK_NAUGHTY(add) \
272 if (RExC_naughty < TOO_NAUGHTY) \
273 RExC_naughty += (add)
274 #define MARK_NAUGHTY_EXP(exp, add) \
275 if (RExC_naughty < TOO_NAUGHTY) \
276 RExC_naughty += RExC_naughty / (exp) + (add)
278 #define ISMULT1(c) ((c) == '*' || (c) == '+' || (c) == '?')
279 #define ISMULT2(s) ((*s) == '*' || (*s) == '+' || (*s) == '?' || \
280 ((*s) == '{' && regcurly(s)))
283 * Flags to be passed up and down.
285 #define WORST 0 /* Worst case. */
286 #define HASWIDTH 0x01 /* Known to match non-null strings. */
288 /* Simple enough to be STAR/PLUS operand; in an EXACTish node must be a single
289 * character. (There needs to be a case: in the switch statement in regexec.c
290 * for any node marked SIMPLE.) Note that this is not the same thing as
293 #define SPSTART 0x04 /* Starts with * or + */
294 #define POSTPONED 0x08 /* (?1),(?&name), (??{...}) or similar */
295 #define TRYAGAIN 0x10 /* Weeded out a declaration. */
296 #define RESTART_UTF8 0x20 /* Restart, need to calcuate sizes as UTF-8 */
298 #define REG_NODE_NUM(x) ((x) ? (int)((x)-RExC_emit_start) : -1)
300 /* whether trie related optimizations are enabled */
301 #if PERL_ENABLE_EXTENDED_TRIE_OPTIMISATION
302 #define TRIE_STUDY_OPT
303 #define FULL_TRIE_STUDY
309 #define PBYTE(u8str,paren) ((U8*)(u8str))[(paren) >> 3]
310 #define PBITVAL(paren) (1 << ((paren) & 7))
311 #define PAREN_TEST(u8str,paren) ( PBYTE(u8str,paren) & PBITVAL(paren))
312 #define PAREN_SET(u8str,paren) PBYTE(u8str,paren) |= PBITVAL(paren)
313 #define PAREN_UNSET(u8str,paren) PBYTE(u8str,paren) &= (~PBITVAL(paren))
315 #define REQUIRE_UTF8 STMT_START { \
317 *flagp = RESTART_UTF8; \
322 /* This converts the named class defined in regcomp.h to its equivalent class
323 * number defined in handy.h. */
324 #define namedclass_to_classnum(class) ((int) ((class) / 2))
325 #define classnum_to_namedclass(classnum) ((classnum) * 2)
327 #define _invlist_union_complement_2nd(a, b, output) \
328 _invlist_union_maybe_complement_2nd(a, b, TRUE, output)
329 #define _invlist_intersection_complement_2nd(a, b, output) \
330 _invlist_intersection_maybe_complement_2nd(a, b, TRUE, output)
332 /* About scan_data_t.
334 During optimisation we recurse through the regexp program performing
335 various inplace (keyhole style) optimisations. In addition study_chunk
336 and scan_commit populate this data structure with information about
337 what strings MUST appear in the pattern. We look for the longest
338 string that must appear at a fixed location, and we look for the
339 longest string that may appear at a floating location. So for instance
344 Both 'FOO' and 'A' are fixed strings. Both 'B' and 'BAR' are floating
345 strings (because they follow a .* construct). study_chunk will identify
346 both FOO and BAR as being the longest fixed and floating strings respectively.
348 The strings can be composites, for instance
352 will result in a composite fixed substring 'foo'.
354 For each string some basic information is maintained:
356 - offset or min_offset
357 This is the position the string must appear at, or not before.
358 It also implicitly (when combined with minlenp) tells us how many
359 characters must match before the string we are searching for.
360 Likewise when combined with minlenp and the length of the string it
361 tells us how many characters must appear after the string we have
365 Only used for floating strings. This is the rightmost point that
366 the string can appear at. If set to SSize_t_MAX it indicates that the
367 string can occur infinitely far to the right.
370 A pointer to the minimum number of characters of the pattern that the
371 string was found inside. This is important as in the case of positive
372 lookahead or positive lookbehind we can have multiple patterns
377 The minimum length of the pattern overall is 3, the minimum length
378 of the lookahead part is 3, but the minimum length of the part that
379 will actually match is 1. So 'FOO's minimum length is 3, but the
380 minimum length for the F is 1. This is important as the minimum length
381 is used to determine offsets in front of and behind the string being
382 looked for. Since strings can be composites this is the length of the
383 pattern at the time it was committed with a scan_commit. Note that
384 the length is calculated by study_chunk, so that the minimum lengths
385 are not known until the full pattern has been compiled, thus the
386 pointer to the value.
390 In the case of lookbehind the string being searched for can be
391 offset past the start point of the final matching string.
392 If this value was just blithely removed from the min_offset it would
393 invalidate some of the calculations for how many chars must match
394 before or after (as they are derived from min_offset and minlen and
395 the length of the string being searched for).
396 When the final pattern is compiled and the data is moved from the
397 scan_data_t structure into the regexp structure the information
398 about lookbehind is factored in, with the information that would
399 have been lost precalculated in the end_shift field for the
402 The fields pos_min and pos_delta are used to store the minimum offset
403 and the delta to the maximum offset at the current point in the pattern.
407 typedef struct scan_data_t {
408 /*I32 len_min; unused */
409 /*I32 len_delta; unused */
413 SSize_t last_end; /* min value, <0 unless valid. */
414 SSize_t last_start_min;
415 SSize_t last_start_max;
416 SV **longest; /* Either &l_fixed, or &l_float. */
417 SV *longest_fixed; /* longest fixed string found in pattern */
418 SSize_t offset_fixed; /* offset where it starts */
419 SSize_t *minlen_fixed; /* pointer to the minlen relevant to the string */
420 I32 lookbehind_fixed; /* is the position of the string modfied by LB */
421 SV *longest_float; /* longest floating string found in pattern */
422 SSize_t offset_float_min; /* earliest point in string it can appear */
423 SSize_t offset_float_max; /* latest point in string it can appear */
424 SSize_t *minlen_float; /* pointer to the minlen relevant to the string */
425 SSize_t lookbehind_float; /* is the pos of the string modified by LB */
428 SSize_t *last_closep;
429 regnode_ssc *start_class;
433 * Forward declarations for pregcomp()'s friends.
436 static const scan_data_t zero_scan_data =
437 { 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0 ,0};
439 #define SF_BEFORE_EOL (SF_BEFORE_SEOL|SF_BEFORE_MEOL)
440 #define SF_BEFORE_SEOL 0x0001
441 #define SF_BEFORE_MEOL 0x0002
442 #define SF_FIX_BEFORE_EOL (SF_FIX_BEFORE_SEOL|SF_FIX_BEFORE_MEOL)
443 #define SF_FL_BEFORE_EOL (SF_FL_BEFORE_SEOL|SF_FL_BEFORE_MEOL)
445 #define SF_FIX_SHIFT_EOL (+2)
446 #define SF_FL_SHIFT_EOL (+4)
448 #define SF_FIX_BEFORE_SEOL (SF_BEFORE_SEOL << SF_FIX_SHIFT_EOL)
449 #define SF_FIX_BEFORE_MEOL (SF_BEFORE_MEOL << SF_FIX_SHIFT_EOL)
451 #define SF_FL_BEFORE_SEOL (SF_BEFORE_SEOL << SF_FL_SHIFT_EOL)
452 #define SF_FL_BEFORE_MEOL (SF_BEFORE_MEOL << SF_FL_SHIFT_EOL) /* 0x20 */
453 #define SF_IS_INF 0x0040
454 #define SF_HAS_PAR 0x0080
455 #define SF_IN_PAR 0x0100
456 #define SF_HAS_EVAL 0x0200
457 #define SCF_DO_SUBSTR 0x0400
458 #define SCF_DO_STCLASS_AND 0x0800
459 #define SCF_DO_STCLASS_OR 0x1000
460 #define SCF_DO_STCLASS (SCF_DO_STCLASS_AND|SCF_DO_STCLASS_OR)
461 #define SCF_WHILEM_VISITED_POS 0x2000
463 #define SCF_TRIE_RESTUDY 0x4000 /* Do restudy? */
464 #define SCF_SEEN_ACCEPT 0x8000
465 #define SCF_TRIE_DOING_RESTUDY 0x10000
466 #define SCF_IN_DEFINE 0x20000
471 #define UTF cBOOL(RExC_utf8)
473 /* The enums for all these are ordered so things work out correctly */
474 #define LOC (get_regex_charset(RExC_flags) == REGEX_LOCALE_CHARSET)
475 #define DEPENDS_SEMANTICS (get_regex_charset(RExC_flags) \
476 == REGEX_DEPENDS_CHARSET)
477 #define UNI_SEMANTICS (get_regex_charset(RExC_flags) == REGEX_UNICODE_CHARSET)
478 #define AT_LEAST_UNI_SEMANTICS (get_regex_charset(RExC_flags) \
479 >= REGEX_UNICODE_CHARSET)
480 #define ASCII_RESTRICTED (get_regex_charset(RExC_flags) \
481 == REGEX_ASCII_RESTRICTED_CHARSET)
482 #define AT_LEAST_ASCII_RESTRICTED (get_regex_charset(RExC_flags) \
483 >= REGEX_ASCII_RESTRICTED_CHARSET)
484 #define ASCII_FOLD_RESTRICTED (get_regex_charset(RExC_flags) \
485 == REGEX_ASCII_MORE_RESTRICTED_CHARSET)
487 #define FOLD cBOOL(RExC_flags & RXf_PMf_FOLD)
489 /* For programs that want to be strictly Unicode compatible by dying if any
490 * attempt is made to match a non-Unicode code point against a Unicode
492 #define ALWAYS_WARN_SUPER ckDEAD(packWARN(WARN_NON_UNICODE))
494 #define OOB_NAMEDCLASS -1
496 /* There is no code point that is out-of-bounds, so this is problematic. But
497 * its only current use is to initialize a variable that is always set before
499 #define OOB_UNICODE 0xDEADBEEF
501 #define CHR_SVLEN(sv) (UTF ? sv_len_utf8(sv) : SvCUR(sv))
502 #define CHR_DIST(a,b) (UTF ? utf8_distance(a,b) : a - b)
505 /* length of regex to show in messages that don't mark a position within */
506 #define RegexLengthToShowInErrorMessages 127
509 * If MARKER[12] are adjusted, be sure to adjust the constants at the top
510 * of t/op/regmesg.t, the tests in t/op/re_tests, and those in
511 * op/pragma/warn/regcomp.
513 #define MARKER1 "<-- HERE" /* marker as it appears in the description */
514 #define MARKER2 " <-- HERE " /* marker as it appears within the regex */
516 #define REPORT_LOCATION " in regex; marked by " MARKER1 \
517 " in m/%"UTF8f MARKER2 "%"UTF8f"/"
519 #define REPORT_LOCATION_ARGS(offset) \
520 UTF8fARG(UTF, offset, RExC_precomp), \
521 UTF8fARG(UTF, RExC_end - RExC_precomp - offset, RExC_precomp + offset)
523 /* Used to point after bad bytes for an error message, but avoid skipping
524 * past a nul byte. */
525 #define SKIP_IF_CHAR(s) (!*(s) ? 0 : UTF ? UTF8SKIP(s) : 1)
528 * Calls SAVEDESTRUCTOR_X if needed, then calls Perl_croak with the given
529 * arg. Show regex, up to a maximum length. If it's too long, chop and add
532 #define _FAIL(code) STMT_START { \
533 const char *ellipses = ""; \
534 IV len = RExC_end - RExC_precomp; \
537 SAVEFREESV(RExC_rx_sv); \
538 if (len > RegexLengthToShowInErrorMessages) { \
539 /* chop 10 shorter than the max, to ensure meaning of "..." */ \
540 len = RegexLengthToShowInErrorMessages - 10; \
546 #define FAIL(msg) _FAIL( \
547 Perl_croak(aTHX_ "%s in regex m/%"UTF8f"%s/", \
548 msg, UTF8fARG(UTF, len, RExC_precomp), ellipses))
550 #define FAIL2(msg,arg) _FAIL( \
551 Perl_croak(aTHX_ msg " in regex m/%"UTF8f"%s/", \
552 arg, UTF8fARG(UTF, len, RExC_precomp), ellipses))
555 * Simple_vFAIL -- like FAIL, but marks the current location in the scan
557 #define Simple_vFAIL(m) STMT_START { \
559 (RExC_parse > RExC_end ? RExC_end : RExC_parse) - RExC_precomp; \
560 Perl_croak(aTHX_ "%s" REPORT_LOCATION, \
561 m, REPORT_LOCATION_ARGS(offset)); \
565 * Calls SAVEDESTRUCTOR_X if needed, then Simple_vFAIL()
567 #define vFAIL(m) STMT_START { \
569 SAVEFREESV(RExC_rx_sv); \
574 * Like Simple_vFAIL(), but accepts two arguments.
576 #define Simple_vFAIL2(m,a1) STMT_START { \
577 const IV offset = RExC_parse - RExC_precomp; \
578 S_re_croak2(aTHX_ UTF, m, REPORT_LOCATION, a1, \
579 REPORT_LOCATION_ARGS(offset)); \
583 * Calls SAVEDESTRUCTOR_X if needed, then Simple_vFAIL2().
585 #define vFAIL2(m,a1) STMT_START { \
587 SAVEFREESV(RExC_rx_sv); \
588 Simple_vFAIL2(m, a1); \
593 * Like Simple_vFAIL(), but accepts three arguments.
595 #define Simple_vFAIL3(m, a1, a2) STMT_START { \
596 const IV offset = RExC_parse - RExC_precomp; \
597 S_re_croak2(aTHX_ UTF, m, REPORT_LOCATION, a1, a2, \
598 REPORT_LOCATION_ARGS(offset)); \
602 * Calls SAVEDESTRUCTOR_X if needed, then Simple_vFAIL3().
604 #define vFAIL3(m,a1,a2) STMT_START { \
606 SAVEFREESV(RExC_rx_sv); \
607 Simple_vFAIL3(m, a1, a2); \
611 * Like Simple_vFAIL(), but accepts four arguments.
613 #define Simple_vFAIL4(m, a1, a2, a3) STMT_START { \
614 const IV offset = RExC_parse - RExC_precomp; \
615 S_re_croak2(aTHX_ UTF, m, REPORT_LOCATION, a1, a2, a3, \
616 REPORT_LOCATION_ARGS(offset)); \
619 #define vFAIL4(m,a1,a2,a3) STMT_START { \
621 SAVEFREESV(RExC_rx_sv); \
622 Simple_vFAIL4(m, a1, a2, a3); \
625 /* A specialized version of vFAIL2 that works with UTF8f */
626 #define vFAIL2utf8f(m, a1) STMT_START { \
627 const IV offset = RExC_parse - RExC_precomp; \
629 SAVEFREESV(RExC_rx_sv); \
630 S_re_croak2(aTHX_ UTF, m, REPORT_LOCATION, a1, \
631 REPORT_LOCATION_ARGS(offset)); \
634 /* These have asserts in them because of [perl #122671] Many warnings in
635 * regcomp.c can occur twice. If they get output in pass1 and later in that
636 * pass, the pattern has to be converted to UTF-8 and the pass restarted, they
637 * would get output again. So they should be output in pass2, and these
638 * asserts make sure new warnings follow that paradigm. */
640 /* m is not necessarily a "literal string", in this macro */
641 #define reg_warn_non_literal_string(loc, m) STMT_START { \
642 const IV offset = loc - RExC_precomp; \
643 __ASSERT_(PASS2) Perl_warner(aTHX_ packWARN(WARN_REGEXP), "%s" REPORT_LOCATION, \
644 m, REPORT_LOCATION_ARGS(offset)); \
647 #define ckWARNreg(loc,m) STMT_START { \
648 const IV offset = loc - RExC_precomp; \
649 __ASSERT_(PASS2) Perl_ck_warner(aTHX_ packWARN(WARN_REGEXP), m REPORT_LOCATION, \
650 REPORT_LOCATION_ARGS(offset)); \
653 #define vWARN(loc, m) STMT_START { \
654 const IV offset = loc - RExC_precomp; \
655 __ASSERT_(PASS2) Perl_warner(aTHX_ packWARN(WARN_REGEXP), m REPORT_LOCATION, \
656 REPORT_LOCATION_ARGS(offset)); \
659 #define vWARN_dep(loc, m) STMT_START { \
660 const IV offset = loc - RExC_precomp; \
661 __ASSERT_(PASS2) Perl_warner(aTHX_ packWARN(WARN_DEPRECATED), m REPORT_LOCATION, \
662 REPORT_LOCATION_ARGS(offset)); \
665 #define ckWARNdep(loc,m) STMT_START { \
666 const IV offset = loc - RExC_precomp; \
667 __ASSERT_(PASS2) Perl_ck_warner_d(aTHX_ packWARN(WARN_DEPRECATED), \
669 REPORT_LOCATION_ARGS(offset)); \
672 #define ckWARNregdep(loc,m) STMT_START { \
673 const IV offset = loc - RExC_precomp; \
674 __ASSERT_(PASS2) Perl_ck_warner_d(aTHX_ packWARN2(WARN_DEPRECATED, WARN_REGEXP), \
676 REPORT_LOCATION_ARGS(offset)); \
679 #define ckWARN2reg_d(loc,m, a1) STMT_START { \
680 const IV offset = loc - RExC_precomp; \
681 __ASSERT_(PASS2) Perl_ck_warner_d(aTHX_ packWARN(WARN_REGEXP), \
683 a1, REPORT_LOCATION_ARGS(offset)); \
686 #define ckWARN2reg(loc, m, a1) STMT_START { \
687 const IV offset = loc - RExC_precomp; \
688 __ASSERT_(PASS2) Perl_ck_warner(aTHX_ packWARN(WARN_REGEXP), m REPORT_LOCATION, \
689 a1, REPORT_LOCATION_ARGS(offset)); \
692 #define vWARN3(loc, m, a1, a2) STMT_START { \
693 const IV offset = loc - RExC_precomp; \
694 __ASSERT_(PASS2) Perl_warner(aTHX_ packWARN(WARN_REGEXP), m REPORT_LOCATION, \
695 a1, a2, REPORT_LOCATION_ARGS(offset)); \
698 #define ckWARN3reg(loc, m, a1, a2) STMT_START { \
699 const IV offset = loc - RExC_precomp; \
700 __ASSERT_(PASS2) Perl_ck_warner(aTHX_ packWARN(WARN_REGEXP), m REPORT_LOCATION, \
701 a1, a2, REPORT_LOCATION_ARGS(offset)); \
704 #define vWARN4(loc, m, a1, a2, a3) STMT_START { \
705 const IV offset = loc - RExC_precomp; \
706 __ASSERT_(PASS2) Perl_warner(aTHX_ packWARN(WARN_REGEXP), m REPORT_LOCATION, \
707 a1, a2, a3, REPORT_LOCATION_ARGS(offset)); \
710 #define ckWARN4reg(loc, m, a1, a2, a3) STMT_START { \
711 const IV offset = loc - RExC_precomp; \
712 __ASSERT_(PASS2) Perl_ck_warner(aTHX_ packWARN(WARN_REGEXP), m REPORT_LOCATION, \
713 a1, a2, a3, REPORT_LOCATION_ARGS(offset)); \
716 #define vWARN5(loc, m, a1, a2, a3, a4) STMT_START { \
717 const IV offset = loc - RExC_precomp; \
718 __ASSERT_(PASS2) Perl_warner(aTHX_ packWARN(WARN_REGEXP), m REPORT_LOCATION, \
719 a1, a2, a3, a4, REPORT_LOCATION_ARGS(offset)); \
722 /* Macros for recording node offsets. 20001227 mjd@plover.com
723 * Nodes are numbered 1, 2, 3, 4. Node #n's position is recorded in
724 * element 2*n-1 of the array. Element #2n holds the byte length node #n.
725 * Element 0 holds the number n.
726 * Position is 1 indexed.
728 #ifndef RE_TRACK_PATTERN_OFFSETS
729 #define Set_Node_Offset_To_R(node,byte)
730 #define Set_Node_Offset(node,byte)
731 #define Set_Cur_Node_Offset
732 #define Set_Node_Length_To_R(node,len)
733 #define Set_Node_Length(node,len)
734 #define Set_Node_Cur_Length(node,start)
735 #define Node_Offset(n)
736 #define Node_Length(n)
737 #define Set_Node_Offset_Length(node,offset,len)
738 #define ProgLen(ri) ri->u.proglen
739 #define SetProgLen(ri,x) ri->u.proglen = x
741 #define ProgLen(ri) ri->u.offsets[0]
742 #define SetProgLen(ri,x) ri->u.offsets[0] = x
743 #define Set_Node_Offset_To_R(node,byte) STMT_START { \
745 MJD_OFFSET_DEBUG(("** (%d) offset of node %d is %d.\n", \
746 __LINE__, (int)(node), (int)(byte))); \
748 Perl_croak(aTHX_ "value of node is %d in Offset macro", \
751 RExC_offsets[2*(node)-1] = (byte); \
756 #define Set_Node_Offset(node,byte) \
757 Set_Node_Offset_To_R((node)-RExC_emit_start, (byte)-RExC_start)
758 #define Set_Cur_Node_Offset Set_Node_Offset(RExC_emit, RExC_parse)
760 #define Set_Node_Length_To_R(node,len) STMT_START { \
762 MJD_OFFSET_DEBUG(("** (%d) size of node %d is %d.\n", \
763 __LINE__, (int)(node), (int)(len))); \
765 Perl_croak(aTHX_ "value of node is %d in Length macro", \
768 RExC_offsets[2*(node)] = (len); \
773 #define Set_Node_Length(node,len) \
774 Set_Node_Length_To_R((node)-RExC_emit_start, len)
775 #define Set_Node_Cur_Length(node, start) \
776 Set_Node_Length(node, RExC_parse - start)
778 /* Get offsets and lengths */
779 #define Node_Offset(n) (RExC_offsets[2*((n)-RExC_emit_start)-1])
780 #define Node_Length(n) (RExC_offsets[2*((n)-RExC_emit_start)])
782 #define Set_Node_Offset_Length(node,offset,len) STMT_START { \
783 Set_Node_Offset_To_R((node)-RExC_emit_start, (offset)); \
784 Set_Node_Length_To_R((node)-RExC_emit_start, (len)); \
788 #if PERL_ENABLE_EXPERIMENTAL_REGEX_OPTIMISATIONS
789 #define EXPERIMENTAL_INPLACESCAN
790 #endif /*PERL_ENABLE_EXPERIMENTAL_REGEX_OPTIMISATIONS*/
792 #define DEBUG_RExC_seen() \
793 DEBUG_OPTIMISE_MORE_r({ \
794 PerlIO_printf(Perl_debug_log,"RExC_seen: "); \
796 if (RExC_seen & REG_ZERO_LEN_SEEN) \
797 PerlIO_printf(Perl_debug_log,"REG_ZERO_LEN_SEEN "); \
799 if (RExC_seen & REG_LOOKBEHIND_SEEN) \
800 PerlIO_printf(Perl_debug_log,"REG_LOOKBEHIND_SEEN "); \
802 if (RExC_seen & REG_GPOS_SEEN) \
803 PerlIO_printf(Perl_debug_log,"REG_GPOS_SEEN "); \
805 if (RExC_seen & REG_CANY_SEEN) \
806 PerlIO_printf(Perl_debug_log,"REG_CANY_SEEN "); \
808 if (RExC_seen & REG_RECURSE_SEEN) \
809 PerlIO_printf(Perl_debug_log,"REG_RECURSE_SEEN "); \
811 if (RExC_seen & REG_TOP_LEVEL_BRANCHES_SEEN) \
812 PerlIO_printf(Perl_debug_log,"REG_TOP_LEVEL_BRANCHES_SEEN "); \
814 if (RExC_seen & REG_VERBARG_SEEN) \
815 PerlIO_printf(Perl_debug_log,"REG_VERBARG_SEEN "); \
817 if (RExC_seen & REG_CUTGROUP_SEEN) \
818 PerlIO_printf(Perl_debug_log,"REG_CUTGROUP_SEEN "); \
820 if (RExC_seen & REG_RUN_ON_COMMENT_SEEN) \
821 PerlIO_printf(Perl_debug_log,"REG_RUN_ON_COMMENT_SEEN "); \
823 if (RExC_seen & REG_UNFOLDED_MULTI_SEEN) \
824 PerlIO_printf(Perl_debug_log,"REG_UNFOLDED_MULTI_SEEN "); \
826 if (RExC_seen & REG_GOSTART_SEEN) \
827 PerlIO_printf(Perl_debug_log,"REG_GOSTART_SEEN "); \
829 if (RExC_seen & REG_UNBOUNDED_QUANTIFIER_SEEN) \
830 PerlIO_printf(Perl_debug_log,"REG_UNBOUNDED_QUANTIFIER_SEEN "); \
832 PerlIO_printf(Perl_debug_log,"\n"); \
835 #define DEBUG_SHOW_STUDY_FLAG(flags,flag) \
836 if ((flags) & flag) PerlIO_printf(Perl_debug_log, "%s ", #flag)
838 #define DEBUG_SHOW_STUDY_FLAGS(flags,open_str,close_str) \
840 PerlIO_printf(Perl_debug_log, "%s", open_str); \
841 DEBUG_SHOW_STUDY_FLAG(flags,SF_FL_BEFORE_SEOL); \
842 DEBUG_SHOW_STUDY_FLAG(flags,SF_FL_BEFORE_MEOL); \
843 DEBUG_SHOW_STUDY_FLAG(flags,SF_IS_INF); \
844 DEBUG_SHOW_STUDY_FLAG(flags,SF_HAS_PAR); \
845 DEBUG_SHOW_STUDY_FLAG(flags,SF_IN_PAR); \
846 DEBUG_SHOW_STUDY_FLAG(flags,SF_HAS_EVAL); \
847 DEBUG_SHOW_STUDY_FLAG(flags,SCF_DO_SUBSTR); \
848 DEBUG_SHOW_STUDY_FLAG(flags,SCF_DO_STCLASS_AND); \
849 DEBUG_SHOW_STUDY_FLAG(flags,SCF_DO_STCLASS_OR); \
850 DEBUG_SHOW_STUDY_FLAG(flags,SCF_DO_STCLASS); \
851 DEBUG_SHOW_STUDY_FLAG(flags,SCF_WHILEM_VISITED_POS); \
852 DEBUG_SHOW_STUDY_FLAG(flags,SCF_TRIE_RESTUDY); \
853 DEBUG_SHOW_STUDY_FLAG(flags,SCF_SEEN_ACCEPT); \
854 DEBUG_SHOW_STUDY_FLAG(flags,SCF_TRIE_DOING_RESTUDY); \
855 DEBUG_SHOW_STUDY_FLAG(flags,SCF_IN_DEFINE); \
856 PerlIO_printf(Perl_debug_log, "%s", close_str); \
860 #define DEBUG_STUDYDATA(str,data,depth) \
861 DEBUG_OPTIMISE_MORE_r(if(data){ \
862 PerlIO_printf(Perl_debug_log, \
863 "%*s" str "Pos:%"IVdf"/%"IVdf \
865 (int)(depth)*2, "", \
866 (IV)((data)->pos_min), \
867 (IV)((data)->pos_delta), \
868 (UV)((data)->flags) \
870 DEBUG_SHOW_STUDY_FLAGS((data)->flags," [ ","]"); \
871 PerlIO_printf(Perl_debug_log, \
872 " Whilem_c: %"IVdf" Lcp: %"IVdf" %s", \
873 (IV)((data)->whilem_c), \
874 (IV)((data)->last_closep ? *((data)->last_closep) : -1), \
875 is_inf ? "INF " : "" \
877 if ((data)->last_found) \
878 PerlIO_printf(Perl_debug_log, \
879 "Last:'%s' %"IVdf":%"IVdf"/%"IVdf" %sFixed:'%s' @ %"IVdf \
880 " %sFloat: '%s' @ %"IVdf"/%"IVdf"", \
881 SvPVX_const((data)->last_found), \
882 (IV)((data)->last_end), \
883 (IV)((data)->last_start_min), \
884 (IV)((data)->last_start_max), \
885 ((data)->longest && \
886 (data)->longest==&((data)->longest_fixed)) ? "*" : "", \
887 SvPVX_const((data)->longest_fixed), \
888 (IV)((data)->offset_fixed), \
889 ((data)->longest && \
890 (data)->longest==&((data)->longest_float)) ? "*" : "", \
891 SvPVX_const((data)->longest_float), \
892 (IV)((data)->offset_float_min), \
893 (IV)((data)->offset_float_max) \
895 PerlIO_printf(Perl_debug_log,"\n"); \
898 /* is c a control character for which we have a mnemonic? */
899 #define isMNEMONIC_CNTRL(c) _IS_MNEMONIC_CNTRL_ONLY_FOR_USE_BY_REGCOMP_DOT_C(c)
902 S_cntrl_to_mnemonic(const U8 c)
904 /* Returns the mnemonic string that represents character 'c', if one
905 * exists; NULL otherwise. The only ones that exist for the purposes of
906 * this routine are a few control characters */
909 case '\a': return "\\a";
910 case '\b': return "\\b";
911 case ESC_NATIVE: return "\\e";
912 case '\f': return "\\f";
913 case '\n': return "\\n";
914 case '\r': return "\\r";
915 case '\t': return "\\t";
921 /* Mark that we cannot extend a found fixed substring at this point.
922 Update the longest found anchored substring and the longest found
923 floating substrings if needed. */
926 S_scan_commit(pTHX_ const RExC_state_t *pRExC_state, scan_data_t *data,
927 SSize_t *minlenp, int is_inf)
929 const STRLEN l = CHR_SVLEN(data->last_found);
930 const STRLEN old_l = CHR_SVLEN(*data->longest);
931 GET_RE_DEBUG_FLAGS_DECL;
933 PERL_ARGS_ASSERT_SCAN_COMMIT;
935 if ((l >= old_l) && ((l > old_l) || (data->flags & SF_BEFORE_EOL))) {
936 SvSetMagicSV(*data->longest, data->last_found);
937 if (*data->longest == data->longest_fixed) {
938 data->offset_fixed = l ? data->last_start_min : data->pos_min;
939 if (data->flags & SF_BEFORE_EOL)
941 |= ((data->flags & SF_BEFORE_EOL) << SF_FIX_SHIFT_EOL);
943 data->flags &= ~SF_FIX_BEFORE_EOL;
944 data->minlen_fixed=minlenp;
945 data->lookbehind_fixed=0;
947 else { /* *data->longest == data->longest_float */
948 data->offset_float_min = l ? data->last_start_min : data->pos_min;
949 data->offset_float_max = (l
950 ? data->last_start_max
951 : (data->pos_delta > SSize_t_MAX - data->pos_min
953 : data->pos_min + data->pos_delta));
955 || (STRLEN)data->offset_float_max > (STRLEN)SSize_t_MAX)
956 data->offset_float_max = SSize_t_MAX;
957 if (data->flags & SF_BEFORE_EOL)
959 |= ((data->flags & SF_BEFORE_EOL) << SF_FL_SHIFT_EOL);
961 data->flags &= ~SF_FL_BEFORE_EOL;
962 data->minlen_float=minlenp;
963 data->lookbehind_float=0;
966 SvCUR_set(data->last_found, 0);
968 SV * const sv = data->last_found;
969 if (SvUTF8(sv) && SvMAGICAL(sv)) {
970 MAGIC * const mg = mg_find(sv, PERL_MAGIC_utf8);
976 data->flags &= ~SF_BEFORE_EOL;
977 DEBUG_STUDYDATA("commit: ",data,0);
980 /* An SSC is just a regnode_charclass_posix with an extra field: the inversion
981 * list that describes which code points it matches */
984 S_ssc_anything(pTHX_ regnode_ssc *ssc)
986 /* Set the SSC 'ssc' to match an empty string or any code point */
988 PERL_ARGS_ASSERT_SSC_ANYTHING;
990 assert(is_ANYOF_SYNTHETIC(ssc));
992 ssc->invlist = sv_2mortal(_new_invlist(2)); /* mortalize so won't leak */
993 _append_range_to_invlist(ssc->invlist, 0, UV_MAX);
994 ANYOF_FLAGS(ssc) |= SSC_MATCHES_EMPTY_STRING; /* Plus matches empty */
998 S_ssc_is_anything(const regnode_ssc *ssc)
1000 /* Returns TRUE if the SSC 'ssc' can match the empty string and any code
1001 * point; FALSE otherwise. Thus, this is used to see if using 'ssc' buys
1002 * us anything: if the function returns TRUE, 'ssc' hasn't been restricted
1003 * in any way, so there's no point in using it */
1008 PERL_ARGS_ASSERT_SSC_IS_ANYTHING;
1010 assert(is_ANYOF_SYNTHETIC(ssc));
1012 if (! (ANYOF_FLAGS(ssc) & SSC_MATCHES_EMPTY_STRING)) {
1016 /* See if the list consists solely of the range 0 - Infinity */
1017 invlist_iterinit(ssc->invlist);
1018 ret = invlist_iternext(ssc->invlist, &start, &end)
1022 invlist_iterfinish(ssc->invlist);
1028 /* If e.g., both \w and \W are set, matches everything */
1029 if (ANYOF_POSIXL_SSC_TEST_ANY_SET(ssc)) {
1031 for (i = 0; i < ANYOF_POSIXL_MAX; i += 2) {
1032 if (ANYOF_POSIXL_TEST(ssc, i) && ANYOF_POSIXL_TEST(ssc, i+1)) {
1042 S_ssc_init(pTHX_ const RExC_state_t *pRExC_state, regnode_ssc *ssc)
1044 /* Initializes the SSC 'ssc'. This includes setting it to match an empty
1045 * string, any code point, or any posix class under locale */
1047 PERL_ARGS_ASSERT_SSC_INIT;
1049 Zero(ssc, 1, regnode_ssc);
1050 set_ANYOF_SYNTHETIC(ssc);
1051 ARG_SET(ssc, ANYOF_ONLY_HAS_BITMAP);
1054 /* If any portion of the regex is to operate under locale rules that aren't
1055 * fully known at compile time, initialization includes it. The reason
1056 * this isn't done for all regexes is that the optimizer was written under
1057 * the assumption that locale was all-or-nothing. Given the complexity and
1058 * lack of documentation in the optimizer, and that there are inadequate
1059 * test cases for locale, many parts of it may not work properly, it is
1060 * safest to avoid locale unless necessary. */
1061 if (RExC_contains_locale) {
1062 ANYOF_POSIXL_SETALL(ssc);
1065 ANYOF_POSIXL_ZERO(ssc);
1070 S_ssc_is_cp_posixl_init(const RExC_state_t *pRExC_state,
1071 const regnode_ssc *ssc)
1073 /* Returns TRUE if the SSC 'ssc' is in its initial state with regard only
1074 * to the list of code points matched, and locale posix classes; hence does
1075 * not check its flags) */
1080 PERL_ARGS_ASSERT_SSC_IS_CP_POSIXL_INIT;
1082 assert(is_ANYOF_SYNTHETIC(ssc));
1084 invlist_iterinit(ssc->invlist);
1085 ret = invlist_iternext(ssc->invlist, &start, &end)
1089 invlist_iterfinish(ssc->invlist);
1095 if (RExC_contains_locale && ! ANYOF_POSIXL_SSC_TEST_ALL_SET(ssc)) {
1103 S_get_ANYOF_cp_list_for_ssc(pTHX_ const RExC_state_t *pRExC_state,
1104 const regnode_charclass* const node)
1106 /* Returns a mortal inversion list defining which code points are matched
1107 * by 'node', which is of type ANYOF. Handles complementing the result if
1108 * appropriate. If some code points aren't knowable at this time, the
1109 * returned list must, and will, contain every code point that is a
1112 SV* invlist = sv_2mortal(_new_invlist(0));
1113 SV* only_utf8_locale_invlist = NULL;
1115 const U32 n = ARG(node);
1116 bool new_node_has_latin1 = FALSE;
1118 PERL_ARGS_ASSERT_GET_ANYOF_CP_LIST_FOR_SSC;
1120 /* Look at the data structure created by S_set_ANYOF_arg() */
1121 if (n != ANYOF_ONLY_HAS_BITMAP) {
1122 SV * const rv = MUTABLE_SV(RExC_rxi->data->data[n]);
1123 AV * const av = MUTABLE_AV(SvRV(rv));
1124 SV **const ary = AvARRAY(av);
1125 assert(RExC_rxi->data->what[n] == 's');
1127 if (ary[1] && ary[1] != &PL_sv_undef) { /* Has compile-time swash */
1128 invlist = sv_2mortal(invlist_clone(_get_swash_invlist(ary[1])));
1130 else if (ary[0] && ary[0] != &PL_sv_undef) {
1132 /* Here, no compile-time swash, and there are things that won't be
1133 * known until runtime -- we have to assume it could be anything */
1134 return _add_range_to_invlist(invlist, 0, UV_MAX);
1136 else if (ary[3] && ary[3] != &PL_sv_undef) {
1138 /* Here no compile-time swash, and no run-time only data. Use the
1139 * node's inversion list */
1140 invlist = sv_2mortal(invlist_clone(ary[3]));
1143 /* Get the code points valid only under UTF-8 locales */
1144 if ((ANYOF_FLAGS(node) & ANYOF_LOC_FOLD)
1145 && ary[2] && ary[2] != &PL_sv_undef)
1147 only_utf8_locale_invlist = ary[2];
1151 /* An ANYOF node contains a bitmap for the first NUM_ANYOF_CODE_POINTS
1152 * code points, and an inversion list for the others, but if there are code
1153 * points that should match only conditionally on the target string being
1154 * UTF-8, those are placed in the inversion list, and not the bitmap.
1155 * Since there are circumstances under which they could match, they are
1156 * included in the SSC. But if the ANYOF node is to be inverted, we have
1157 * to exclude them here, so that when we invert below, the end result
1158 * actually does include them. (Think about "\xe0" =~ /[^\xc0]/di;). We
1159 * have to do this here before we add the unconditionally matched code
1161 if (ANYOF_FLAGS(node) & ANYOF_INVERT) {
1162 _invlist_intersection_complement_2nd(invlist,
1167 /* Add in the points from the bit map */
1168 for (i = 0; i < NUM_ANYOF_CODE_POINTS; i++) {
1169 if (ANYOF_BITMAP_TEST(node, i)) {
1170 invlist = add_cp_to_invlist(invlist, i);
1171 new_node_has_latin1 = TRUE;
1175 /* If this can match all upper Latin1 code points, have to add them
1177 if (ANYOF_FLAGS(node) & ANYOF_MATCHES_ALL_NON_UTF8_NON_ASCII) {
1178 _invlist_union(invlist, PL_UpperLatin1, &invlist);
1181 /* Similarly for these */
1182 if (ANYOF_FLAGS(node) & ANYOF_MATCHES_ALL_ABOVE_BITMAP) {
1183 _invlist_union_complement_2nd(invlist, PL_InBitmap, &invlist);
1186 if (ANYOF_FLAGS(node) & ANYOF_INVERT) {
1187 _invlist_invert(invlist);
1189 else if (new_node_has_latin1 && ANYOF_FLAGS(node) & ANYOF_LOC_FOLD) {
1191 /* Under /li, any 0-255 could fold to any other 0-255, depending on the
1192 * locale. We can skip this if there are no 0-255 at all. */
1193 _invlist_union(invlist, PL_Latin1, &invlist);
1196 /* Similarly add the UTF-8 locale possible matches. These have to be
1197 * deferred until after the non-UTF-8 locale ones are taken care of just
1198 * above, or it leads to wrong results under ANYOF_INVERT */
1199 if (only_utf8_locale_invlist) {
1200 _invlist_union_maybe_complement_2nd(invlist,
1201 only_utf8_locale_invlist,
1202 ANYOF_FLAGS(node) & ANYOF_INVERT,
1209 /* These two functions currently do the exact same thing */
1210 #define ssc_init_zero ssc_init
1212 #define ssc_add_cp(ssc, cp) ssc_add_range((ssc), (cp), (cp))
1213 #define ssc_match_all_cp(ssc) ssc_add_range(ssc, 0, UV_MAX)
1215 /* 'AND' a given class with another one. Can create false positives. 'ssc'
1216 * should not be inverted. 'and_with->flags & ANYOF_MATCHES_POSIXL' should be
1217 * 0 if 'and_with' is a regnode_charclass instead of a regnode_ssc. */
1220 S_ssc_and(pTHX_ const RExC_state_t *pRExC_state, regnode_ssc *ssc,
1221 const regnode_charclass *and_with)
1223 /* Accumulate into SSC 'ssc' its 'AND' with 'and_with', which is either
1224 * another SSC or a regular ANYOF class. Can create false positives. */
1229 PERL_ARGS_ASSERT_SSC_AND;
1231 assert(is_ANYOF_SYNTHETIC(ssc));
1233 /* 'and_with' is used as-is if it too is an SSC; otherwise have to extract
1234 * the code point inversion list and just the relevant flags */
1235 if (is_ANYOF_SYNTHETIC(and_with)) {
1236 anded_cp_list = ((regnode_ssc *)and_with)->invlist;
1237 anded_flags = ANYOF_FLAGS(and_with);
1239 /* XXX This is a kludge around what appears to be deficiencies in the
1240 * optimizer. If we make S_ssc_anything() add in the WARN_SUPER flag,
1241 * there are paths through the optimizer where it doesn't get weeded
1242 * out when it should. And if we don't make some extra provision for
1243 * it like the code just below, it doesn't get added when it should.
1244 * This solution is to add it only when AND'ing, which is here, and
1245 * only when what is being AND'ed is the pristine, original node
1246 * matching anything. Thus it is like adding it to ssc_anything() but
1247 * only when the result is to be AND'ed. Probably the same solution
1248 * could be adopted for the same problem we have with /l matching,
1249 * which is solved differently in S_ssc_init(), and that would lead to
1250 * fewer false positives than that solution has. But if this solution
1251 * creates bugs, the consequences are only that a warning isn't raised
1252 * that should be; while the consequences for having /l bugs is
1253 * incorrect matches */
1254 if (ssc_is_anything((regnode_ssc *)and_with)) {
1255 anded_flags |= ANYOF_WARN_SUPER;
1259 anded_cp_list = get_ANYOF_cp_list_for_ssc(pRExC_state, and_with);
1260 anded_flags = ANYOF_FLAGS(and_with) & ANYOF_COMMON_FLAGS;
1263 ANYOF_FLAGS(ssc) &= anded_flags;
1265 /* Below, C1 is the list of code points in 'ssc'; P1, its posix classes.
1266 * C2 is the list of code points in 'and-with'; P2, its posix classes.
1267 * 'and_with' may be inverted. When not inverted, we have the situation of
1269 * (C1 | P1) & (C2 | P2)
1270 * = (C1 & (C2 | P2)) | (P1 & (C2 | P2))
1271 * = ((C1 & C2) | (C1 & P2)) | ((P1 & C2) | (P1 & P2))
1272 * <= ((C1 & C2) | P2)) | ( P1 | (P1 & P2))
1273 * <= ((C1 & C2) | P1 | P2)
1274 * Alternatively, the last few steps could be:
1275 * = ((C1 & C2) | (C1 & P2)) | ((P1 & C2) | (P1 & P2))
1276 * <= ((C1 & C2) | C1 ) | ( C2 | (P1 & P2))
1277 * <= (C1 | C2 | (P1 & P2))
1278 * We favor the second approach if either P1 or P2 is non-empty. This is
1279 * because these components are a barrier to doing optimizations, as what
1280 * they match cannot be known until the moment of matching as they are
1281 * dependent on the current locale, 'AND"ing them likely will reduce or
1283 * But we can do better if we know that C1,P1 are in their initial state (a
1284 * frequent occurrence), each matching everything:
1285 * (<everything>) & (C2 | P2) = C2 | P2
1286 * Similarly, if C2,P2 are in their initial state (again a frequent
1287 * occurrence), the result is a no-op
1288 * (C1 | P1) & (<everything>) = C1 | P1
1291 * (C1 | P1) & ~(C2 | P2) = (C1 | P1) & (~C2 & ~P2)
1292 * = (C1 & (~C2 & ~P2)) | (P1 & (~C2 & ~P2))
1293 * <= (C1 & ~C2) | (P1 & ~P2)
1296 if ((ANYOF_FLAGS(and_with) & ANYOF_INVERT)
1297 && ! is_ANYOF_SYNTHETIC(and_with))
1301 ssc_intersection(ssc,
1303 FALSE /* Has already been inverted */
1306 /* If either P1 or P2 is empty, the intersection will be also; can skip
1308 if (! (ANYOF_FLAGS(and_with) & ANYOF_MATCHES_POSIXL)) {
1309 ANYOF_POSIXL_ZERO(ssc);
1311 else if (ANYOF_POSIXL_SSC_TEST_ANY_SET(ssc)) {
1313 /* Note that the Posix class component P from 'and_with' actually
1315 * P = Pa | Pb | ... | Pn
1316 * where each component is one posix class, such as in [\w\s].
1318 * ~P = ~(Pa | Pb | ... | Pn)
1319 * = ~Pa & ~Pb & ... & ~Pn
1320 * <= ~Pa | ~Pb | ... | ~Pn
1321 * The last is something we can easily calculate, but unfortunately
1322 * is likely to have many false positives. We could do better
1323 * in some (but certainly not all) instances if two classes in
1324 * P have known relationships. For example
1325 * :lower: <= :alpha: <= :alnum: <= \w <= :graph: <= :print:
1327 * :lower: & :print: = :lower:
1328 * And similarly for classes that must be disjoint. For example,
1329 * since \s and \w can have no elements in common based on rules in
1330 * the POSIX standard,
1331 * \w & ^\S = nothing
1332 * Unfortunately, some vendor locales do not meet the Posix
1333 * standard, in particular almost everything by Microsoft.
1334 * The loop below just changes e.g., \w into \W and vice versa */
1336 regnode_charclass_posixl temp;
1337 int add = 1; /* To calculate the index of the complement */
1339 ANYOF_POSIXL_ZERO(&temp);
1340 for (i = 0; i < ANYOF_MAX; i++) {
1342 || ! ANYOF_POSIXL_TEST((regnode_charclass_posixl*) and_with, i)
1343 || ! ANYOF_POSIXL_TEST((regnode_charclass_posixl*) and_with, i + 1));
1345 if (ANYOF_POSIXL_TEST((regnode_charclass_posixl*) and_with, i)) {
1346 ANYOF_POSIXL_SET(&temp, i + add);
1348 add = 0 - add; /* 1 goes to -1; -1 goes to 1 */
1350 ANYOF_POSIXL_AND(&temp, ssc);
1352 } /* else ssc already has no posixes */
1353 } /* else: Not inverted. This routine is a no-op if 'and_with' is an SSC
1354 in its initial state */
1355 else if (! is_ANYOF_SYNTHETIC(and_with)
1356 || ! ssc_is_cp_posixl_init(pRExC_state, (regnode_ssc *)and_with))
1358 /* But if 'ssc' is in its initial state, the result is just 'and_with';
1359 * copy it over 'ssc' */
1360 if (ssc_is_cp_posixl_init(pRExC_state, ssc)) {
1361 if (is_ANYOF_SYNTHETIC(and_with)) {
1362 StructCopy(and_with, ssc, regnode_ssc);
1365 ssc->invlist = anded_cp_list;
1366 ANYOF_POSIXL_ZERO(ssc);
1367 if (ANYOF_FLAGS(and_with) & ANYOF_MATCHES_POSIXL) {
1368 ANYOF_POSIXL_OR((regnode_charclass_posixl*) and_with, ssc);
1372 else if (ANYOF_POSIXL_SSC_TEST_ANY_SET(ssc)
1373 || (ANYOF_FLAGS(and_with) & ANYOF_MATCHES_POSIXL))
1375 /* One or the other of P1, P2 is non-empty. */
1376 if (ANYOF_FLAGS(and_with) & ANYOF_MATCHES_POSIXL) {
1377 ANYOF_POSIXL_AND((regnode_charclass_posixl*) and_with, ssc);
1379 ssc_union(ssc, anded_cp_list, FALSE);
1381 else { /* P1 = P2 = empty */
1382 ssc_intersection(ssc, anded_cp_list, FALSE);
1388 S_ssc_or(pTHX_ const RExC_state_t *pRExC_state, regnode_ssc *ssc,
1389 const regnode_charclass *or_with)
1391 /* Accumulate into SSC 'ssc' its 'OR' with 'or_with', which is either
1392 * another SSC or a regular ANYOF class. Can create false positives if
1393 * 'or_with' is to be inverted. */
1398 PERL_ARGS_ASSERT_SSC_OR;
1400 assert(is_ANYOF_SYNTHETIC(ssc));
1402 /* 'or_with' is used as-is if it too is an SSC; otherwise have to extract
1403 * the code point inversion list and just the relevant flags */
1404 if (is_ANYOF_SYNTHETIC(or_with)) {
1405 ored_cp_list = ((regnode_ssc*) or_with)->invlist;
1406 ored_flags = ANYOF_FLAGS(or_with);
1409 ored_cp_list = get_ANYOF_cp_list_for_ssc(pRExC_state, or_with);
1410 ored_flags = ANYOF_FLAGS(or_with) & ANYOF_COMMON_FLAGS;
1413 ANYOF_FLAGS(ssc) |= ored_flags;
1415 /* Below, C1 is the list of code points in 'ssc'; P1, its posix classes.
1416 * C2 is the list of code points in 'or-with'; P2, its posix classes.
1417 * 'or_with' may be inverted. When not inverted, we have the simple
1418 * situation of computing:
1419 * (C1 | P1) | (C2 | P2) = (C1 | C2) | (P1 | P2)
1420 * If P1|P2 yields a situation with both a class and its complement are
1421 * set, like having both \w and \W, this matches all code points, and we
1422 * can delete these from the P component of the ssc going forward. XXX We
1423 * might be able to delete all the P components, but I (khw) am not certain
1424 * about this, and it is better to be safe.
1427 * (C1 | P1) | ~(C2 | P2) = (C1 | P1) | (~C2 & ~P2)
1428 * <= (C1 | P1) | ~C2
1429 * <= (C1 | ~C2) | P1
1430 * (which results in actually simpler code than the non-inverted case)
1433 if ((ANYOF_FLAGS(or_with) & ANYOF_INVERT)
1434 && ! is_ANYOF_SYNTHETIC(or_with))
1436 /* We ignore P2, leaving P1 going forward */
1437 } /* else Not inverted */
1438 else if (ANYOF_FLAGS(or_with) & ANYOF_MATCHES_POSIXL) {
1439 ANYOF_POSIXL_OR((regnode_charclass_posixl*)or_with, ssc);
1440 if (ANYOF_POSIXL_SSC_TEST_ANY_SET(ssc)) {
1442 for (i = 0; i < ANYOF_MAX; i += 2) {
1443 if (ANYOF_POSIXL_TEST(ssc, i) && ANYOF_POSIXL_TEST(ssc, i + 1))
1445 ssc_match_all_cp(ssc);
1446 ANYOF_POSIXL_CLEAR(ssc, i);
1447 ANYOF_POSIXL_CLEAR(ssc, i+1);
1455 FALSE /* Already has been inverted */
1459 PERL_STATIC_INLINE void
1460 S_ssc_union(pTHX_ regnode_ssc *ssc, SV* const invlist, const bool invert2nd)
1462 PERL_ARGS_ASSERT_SSC_UNION;
1464 assert(is_ANYOF_SYNTHETIC(ssc));
1466 _invlist_union_maybe_complement_2nd(ssc->invlist,
1472 PERL_STATIC_INLINE void
1473 S_ssc_intersection(pTHX_ regnode_ssc *ssc,
1475 const bool invert2nd)
1477 PERL_ARGS_ASSERT_SSC_INTERSECTION;
1479 assert(is_ANYOF_SYNTHETIC(ssc));
1481 _invlist_intersection_maybe_complement_2nd(ssc->invlist,
1487 PERL_STATIC_INLINE void
1488 S_ssc_add_range(pTHX_ regnode_ssc *ssc, const UV start, const UV end)
1490 PERL_ARGS_ASSERT_SSC_ADD_RANGE;
1492 assert(is_ANYOF_SYNTHETIC(ssc));
1494 ssc->invlist = _add_range_to_invlist(ssc->invlist, start, end);
1497 PERL_STATIC_INLINE void
1498 S_ssc_cp_and(pTHX_ regnode_ssc *ssc, const UV cp)
1500 /* AND just the single code point 'cp' into the SSC 'ssc' */
1502 SV* cp_list = _new_invlist(2);
1504 PERL_ARGS_ASSERT_SSC_CP_AND;
1506 assert(is_ANYOF_SYNTHETIC(ssc));
1508 cp_list = add_cp_to_invlist(cp_list, cp);
1509 ssc_intersection(ssc, cp_list,
1510 FALSE /* Not inverted */
1512 SvREFCNT_dec_NN(cp_list);
1515 PERL_STATIC_INLINE void
1516 S_ssc_clear_locale(regnode_ssc *ssc)
1518 /* Set the SSC 'ssc' to not match any locale things */
1519 PERL_ARGS_ASSERT_SSC_CLEAR_LOCALE;
1521 assert(is_ANYOF_SYNTHETIC(ssc));
1523 ANYOF_POSIXL_ZERO(ssc);
1524 ANYOF_FLAGS(ssc) &= ~ANYOF_LOCALE_FLAGS;
1527 #define NON_OTHER_COUNT NON_OTHER_COUNT_FOR_USE_ONLY_BY_REGCOMP_DOT_C
1530 S_is_ssc_worth_it(const RExC_state_t * pRExC_state, const regnode_ssc * ssc)
1532 /* The synthetic start class is used to hopefully quickly winnow down
1533 * places where a pattern could start a match in the target string. If it
1534 * doesn't really narrow things down that much, there isn't much point to
1535 * having the overhead of using it. This function uses some very crude
1536 * heuristics to decide if to use the ssc or not.
1538 * It returns TRUE if 'ssc' rules out more than half what it considers to
1539 * be the "likely" possible matches, but of course it doesn't know what the
1540 * actual things being matched are going to be; these are only guesses
1542 * For /l matches, it assumes that the only likely matches are going to be
1543 * in the 0-255 range, uniformly distributed, so half of that is 127
1544 * For /a and /d matches, it assumes that the likely matches will be just
1545 * the ASCII range, so half of that is 63
1546 * For /u and there isn't anything matching above the Latin1 range, it
1547 * assumes that that is the only range likely to be matched, and uses
1548 * half that as the cut-off: 127. If anything matches above Latin1,
1549 * it assumes that all of Unicode could match (uniformly), except for
1550 * non-Unicode code points and things in the General Category "Other"
1551 * (unassigned, private use, surrogates, controls and formats). This
1552 * is a much large number. */
1554 const U32 max_match = (LOC)
1558 : (invlist_highest(ssc->invlist) < 256)
1560 : ((NON_OTHER_COUNT + 1) / 2) - 1;
1561 U32 count = 0; /* Running total of number of code points matched by
1563 UV start, end; /* Start and end points of current range in inversion
1566 PERL_ARGS_ASSERT_IS_SSC_WORTH_IT;
1568 invlist_iterinit(ssc->invlist);
1569 while (invlist_iternext(ssc->invlist, &start, &end)) {
1571 /* /u is the only thing that we expect to match above 255; so if not /u
1572 * and even if there are matches above 255, ignore them. This catches
1573 * things like \d under /d which does match the digits above 255, but
1574 * since the pattern is /d, it is not likely to be expecting them */
1575 if (! UNI_SEMANTICS) {
1579 end = MIN(end, 255);
1581 count += end - start + 1;
1582 if (count > max_match) {
1583 invlist_iterfinish(ssc->invlist);
1593 S_ssc_finalize(pTHX_ RExC_state_t *pRExC_state, regnode_ssc *ssc)
1595 /* The inversion list in the SSC is marked mortal; now we need a more
1596 * permanent copy, which is stored the same way that is done in a regular
1597 * ANYOF node, with the first NUM_ANYOF_CODE_POINTS code points in a bit
1600 SV* invlist = invlist_clone(ssc->invlist);
1602 PERL_ARGS_ASSERT_SSC_FINALIZE;
1604 assert(is_ANYOF_SYNTHETIC(ssc));
1606 /* The code in this file assumes that all but these flags aren't relevant
1607 * to the SSC, except SSC_MATCHES_EMPTY_STRING, which should be cleared
1608 * by the time we reach here */
1609 assert(! (ANYOF_FLAGS(ssc) & ~ANYOF_COMMON_FLAGS));
1611 populate_ANYOF_from_invlist( (regnode *) ssc, &invlist);
1613 set_ANYOF_arg(pRExC_state, (regnode *) ssc, invlist,
1614 NULL, NULL, NULL, FALSE);
1616 /* Make sure is clone-safe */
1617 ssc->invlist = NULL;
1619 if (ANYOF_POSIXL_SSC_TEST_ANY_SET(ssc)) {
1620 ANYOF_FLAGS(ssc) |= ANYOF_MATCHES_POSIXL;
1623 assert(! (ANYOF_FLAGS(ssc) & ANYOF_LOCALE_FLAGS) || RExC_contains_locale);
1626 #define TRIE_LIST_ITEM(state,idx) (trie->states[state].trans.list)[ idx ]
1627 #define TRIE_LIST_CUR(state) ( TRIE_LIST_ITEM( state, 0 ).forid )
1628 #define TRIE_LIST_LEN(state) ( TRIE_LIST_ITEM( state, 0 ).newstate )
1629 #define TRIE_LIST_USED(idx) ( trie->states[state].trans.list \
1630 ? (TRIE_LIST_CUR( idx ) - 1) \
1636 dump_trie(trie,widecharmap,revcharmap)
1637 dump_trie_interim_list(trie,widecharmap,revcharmap,next_alloc)
1638 dump_trie_interim_table(trie,widecharmap,revcharmap,next_alloc)
1640 These routines dump out a trie in a somewhat readable format.
1641 The _interim_ variants are used for debugging the interim
1642 tables that are used to generate the final compressed
1643 representation which is what dump_trie expects.
1645 Part of the reason for their existence is to provide a form
1646 of documentation as to how the different representations function.
1651 Dumps the final compressed table form of the trie to Perl_debug_log.
1652 Used for debugging make_trie().
1656 S_dump_trie(pTHX_ const struct _reg_trie_data *trie, HV *widecharmap,
1657 AV *revcharmap, U32 depth)
1660 SV *sv=sv_newmortal();
1661 int colwidth= widecharmap ? 6 : 4;
1663 GET_RE_DEBUG_FLAGS_DECL;
1665 PERL_ARGS_ASSERT_DUMP_TRIE;
1667 PerlIO_printf( Perl_debug_log, "%*sChar : %-6s%-6s%-4s ",
1668 (int)depth * 2 + 2,"",
1669 "Match","Base","Ofs" );
1671 for( state = 0 ; state < trie->uniquecharcount ; state++ ) {
1672 SV ** const tmp = av_fetch( revcharmap, state, 0);
1674 PerlIO_printf( Perl_debug_log, "%*s",
1676 pv_pretty(sv, SvPV_nolen_const(*tmp), SvCUR(*tmp), colwidth,
1677 PL_colors[0], PL_colors[1],
1678 (SvUTF8(*tmp) ? PERL_PV_ESCAPE_UNI : 0) |
1679 PERL_PV_ESCAPE_FIRSTCHAR
1684 PerlIO_printf( Perl_debug_log, "\n%*sState|-----------------------",
1685 (int)depth * 2 + 2,"");
1687 for( state = 0 ; state < trie->uniquecharcount ; state++ )
1688 PerlIO_printf( Perl_debug_log, "%.*s", colwidth, "--------");
1689 PerlIO_printf( Perl_debug_log, "\n");
1691 for( state = 1 ; state < trie->statecount ; state++ ) {
1692 const U32 base = trie->states[ state ].trans.base;
1694 PerlIO_printf( Perl_debug_log, "%*s#%4"UVXf"|",
1695 (int)depth * 2 + 2,"", (UV)state);
1697 if ( trie->states[ state ].wordnum ) {
1698 PerlIO_printf( Perl_debug_log, " W%4X",
1699 trie->states[ state ].wordnum );
1701 PerlIO_printf( Perl_debug_log, "%6s", "" );
1704 PerlIO_printf( Perl_debug_log, " @%4"UVXf" ", (UV)base );
1709 while( ( base + ofs < trie->uniquecharcount ) ||
1710 ( base + ofs - trie->uniquecharcount < trie->lasttrans
1711 && trie->trans[ base + ofs - trie->uniquecharcount ].check
1715 PerlIO_printf( Perl_debug_log, "+%2"UVXf"[ ", (UV)ofs);
1717 for ( ofs = 0 ; ofs < trie->uniquecharcount ; ofs++ ) {
1718 if ( ( base + ofs >= trie->uniquecharcount )
1719 && ( base + ofs - trie->uniquecharcount
1721 && trie->trans[ base + ofs
1722 - trie->uniquecharcount ].check == state )
1724 PerlIO_printf( Perl_debug_log, "%*"UVXf,
1726 (UV)trie->trans[ base + ofs
1727 - trie->uniquecharcount ].next );
1729 PerlIO_printf( Perl_debug_log, "%*s",colwidth," ." );
1733 PerlIO_printf( Perl_debug_log, "]");
1736 PerlIO_printf( Perl_debug_log, "\n" );
1738 PerlIO_printf(Perl_debug_log, "%*sword_info N:(prev,len)=",
1740 for (word=1; word <= trie->wordcount; word++) {
1741 PerlIO_printf(Perl_debug_log, " %d:(%d,%d)",
1742 (int)word, (int)(trie->wordinfo[word].prev),
1743 (int)(trie->wordinfo[word].len));
1745 PerlIO_printf(Perl_debug_log, "\n" );
1748 Dumps a fully constructed but uncompressed trie in list form.
1749 List tries normally only are used for construction when the number of
1750 possible chars (trie->uniquecharcount) is very high.
1751 Used for debugging make_trie().
1754 S_dump_trie_interim_list(pTHX_ const struct _reg_trie_data *trie,
1755 HV *widecharmap, AV *revcharmap, U32 next_alloc,
1759 SV *sv=sv_newmortal();
1760 int colwidth= widecharmap ? 6 : 4;
1761 GET_RE_DEBUG_FLAGS_DECL;
1763 PERL_ARGS_ASSERT_DUMP_TRIE_INTERIM_LIST;
1765 /* print out the table precompression. */
1766 PerlIO_printf( Perl_debug_log, "%*sState :Word | Transition Data\n%*s%s",
1767 (int)depth * 2 + 2,"", (int)depth * 2 + 2,"",
1768 "------:-----+-----------------\n" );
1770 for( state=1 ; state < next_alloc ; state ++ ) {
1773 PerlIO_printf( Perl_debug_log, "%*s %4"UVXf" :",
1774 (int)depth * 2 + 2,"", (UV)state );
1775 if ( ! trie->states[ state ].wordnum ) {
1776 PerlIO_printf( Perl_debug_log, "%5s| ","");
1778 PerlIO_printf( Perl_debug_log, "W%4x| ",
1779 trie->states[ state ].wordnum
1782 for( charid = 1 ; charid <= TRIE_LIST_USED( state ) ; charid++ ) {
1783 SV ** const tmp = av_fetch( revcharmap,
1784 TRIE_LIST_ITEM(state,charid).forid, 0);
1786 PerlIO_printf( Perl_debug_log, "%*s:%3X=%4"UVXf" | ",
1788 pv_pretty(sv, SvPV_nolen_const(*tmp), SvCUR(*tmp),
1790 PL_colors[0], PL_colors[1],
1791 (SvUTF8(*tmp) ? PERL_PV_ESCAPE_UNI : 0)
1792 | PERL_PV_ESCAPE_FIRSTCHAR
1794 TRIE_LIST_ITEM(state,charid).forid,
1795 (UV)TRIE_LIST_ITEM(state,charid).newstate
1798 PerlIO_printf(Perl_debug_log, "\n%*s| ",
1799 (int)((depth * 2) + 14), "");
1802 PerlIO_printf( Perl_debug_log, "\n");
1807 Dumps a fully constructed but uncompressed trie in table form.
1808 This is the normal DFA style state transition table, with a few
1809 twists to facilitate compression later.
1810 Used for debugging make_trie().
1813 S_dump_trie_interim_table(pTHX_ const struct _reg_trie_data *trie,
1814 HV *widecharmap, AV *revcharmap, U32 next_alloc,
1819 SV *sv=sv_newmortal();
1820 int colwidth= widecharmap ? 6 : 4;
1821 GET_RE_DEBUG_FLAGS_DECL;
1823 PERL_ARGS_ASSERT_DUMP_TRIE_INTERIM_TABLE;
1826 print out the table precompression so that we can do a visual check
1827 that they are identical.
1830 PerlIO_printf( Perl_debug_log, "%*sChar : ",(int)depth * 2 + 2,"" );
1832 for( charid = 0 ; charid < trie->uniquecharcount ; charid++ ) {
1833 SV ** const tmp = av_fetch( revcharmap, charid, 0);
1835 PerlIO_printf( Perl_debug_log, "%*s",
1837 pv_pretty(sv, SvPV_nolen_const(*tmp), SvCUR(*tmp), colwidth,
1838 PL_colors[0], PL_colors[1],
1839 (SvUTF8(*tmp) ? PERL_PV_ESCAPE_UNI : 0) |
1840 PERL_PV_ESCAPE_FIRSTCHAR
1846 PerlIO_printf( Perl_debug_log, "\n%*sState+-",(int)depth * 2 + 2,"" );
1848 for( charid=0 ; charid < trie->uniquecharcount ; charid++ ) {
1849 PerlIO_printf( Perl_debug_log, "%.*s", colwidth,"--------");
1852 PerlIO_printf( Perl_debug_log, "\n" );
1854 for( state=1 ; state < next_alloc ; state += trie->uniquecharcount ) {
1856 PerlIO_printf( Perl_debug_log, "%*s%4"UVXf" : ",
1857 (int)depth * 2 + 2,"",
1858 (UV)TRIE_NODENUM( state ) );
1860 for( charid = 0 ; charid < trie->uniquecharcount ; charid++ ) {
1861 UV v=(UV)SAFE_TRIE_NODENUM( trie->trans[ state + charid ].next );
1863 PerlIO_printf( Perl_debug_log, "%*"UVXf, colwidth, v );
1865 PerlIO_printf( Perl_debug_log, "%*s", colwidth, "." );
1867 if ( ! trie->states[ TRIE_NODENUM( state ) ].wordnum ) {
1868 PerlIO_printf( Perl_debug_log, " (%4"UVXf")\n",
1869 (UV)trie->trans[ state ].check );
1871 PerlIO_printf( Perl_debug_log, " (%4"UVXf") W%4X\n",
1872 (UV)trie->trans[ state ].check,
1873 trie->states[ TRIE_NODENUM( state ) ].wordnum );
1881 /* make_trie(startbranch,first,last,tail,word_count,flags,depth)
1882 startbranch: the first branch in the whole branch sequence
1883 first : start branch of sequence of branch-exact nodes.
1884 May be the same as startbranch
1885 last : Thing following the last branch.
1886 May be the same as tail.
1887 tail : item following the branch sequence
1888 count : words in the sequence
1889 flags : currently the OP() type we will be building one of /EXACT(|F|FA|FU|FU_SS|L|FLU8)/
1890 depth : indent depth
1892 Inplace optimizes a sequence of 2 or more Branch-Exact nodes into a TRIE node.
1894 A trie is an N'ary tree where the branches are determined by digital
1895 decomposition of the key. IE, at the root node you look up the 1st character and
1896 follow that branch repeat until you find the end of the branches. Nodes can be
1897 marked as "accepting" meaning they represent a complete word. Eg:
1901 would convert into the following structure. Numbers represent states, letters
1902 following numbers represent valid transitions on the letter from that state, if
1903 the number is in square brackets it represents an accepting state, otherwise it
1904 will be in parenthesis.
1906 +-h->+-e->[3]-+-r->(8)-+-s->[9]
1910 (1) +-i->(6)-+-s->[7]
1912 +-s->(3)-+-h->(4)-+-e->[5]
1914 Accept Word Mapping: 3=>1 (he),5=>2 (she), 7=>3 (his), 9=>4 (hers)
1916 This shows that when matching against the string 'hers' we will begin at state 1
1917 read 'h' and move to state 2, read 'e' and move to state 3 which is accepting,
1918 then read 'r' and go to state 8 followed by 's' which takes us to state 9 which
1919 is also accepting. Thus we know that we can match both 'he' and 'hers' with a
1920 single traverse. We store a mapping from accepting to state to which word was
1921 matched, and then when we have multiple possibilities we try to complete the
1922 rest of the regex in the order in which they occurred in the alternation.
1924 The only prior NFA like behaviour that would be changed by the TRIE support is
1925 the silent ignoring of duplicate alternations which are of the form:
1927 / (DUPE|DUPE) X? (?{ ... }) Y /x
1929 Thus EVAL blocks following a trie may be called a different number of times with
1930 and without the optimisation. With the optimisations dupes will be silently
1931 ignored. This inconsistent behaviour of EVAL type nodes is well established as
1932 the following demonstrates:
1934 'words'=~/(word|word|word)(?{ print $1 })[xyz]/
1936 which prints out 'word' three times, but
1938 'words'=~/(word|word|word)(?{ print $1 })S/
1940 which doesnt print it out at all. This is due to other optimisations kicking in.
1942 Example of what happens on a structural level:
1944 The regexp /(ac|ad|ab)+/ will produce the following debug output:
1946 1: CURLYM[1] {1,32767}(18)
1957 This would be optimizable with startbranch=5, first=5, last=16, tail=16
1958 and should turn into:
1960 1: CURLYM[1] {1,32767}(18)
1962 [Words:3 Chars Stored:6 Unique Chars:4 States:5 NCP:1]
1970 Cases where tail != last would be like /(?foo|bar)baz/:
1980 which would be optimizable with startbranch=1, first=1, last=7, tail=8
1981 and would end up looking like:
1984 [Words:2 Chars Stored:6 Unique Chars:5 States:7 NCP:1]
1991 d = uvchr_to_utf8_flags(d, uv, 0);
1993 is the recommended Unicode-aware way of saying
1998 #define TRIE_STORE_REVCHAR(val) \
2001 SV *zlopp = newSV(7); /* XXX: optimize me */ \
2002 unsigned char *flrbbbbb = (unsigned char *) SvPVX(zlopp); \
2003 unsigned const char *const kapow = uvchr_to_utf8(flrbbbbb, val); \
2004 SvCUR_set(zlopp, kapow - flrbbbbb); \
2007 av_push(revcharmap, zlopp); \
2009 char ooooff = (char)val; \
2010 av_push(revcharmap, newSVpvn(&ooooff, 1)); \
2014 /* This gets the next character from the input, folding it if not already
2016 #define TRIE_READ_CHAR STMT_START { \
2019 /* if it is UTF then it is either already folded, or does not need \
2021 uvc = valid_utf8_to_uvchr( (const U8*) uc, &len); \
2023 else if (folder == PL_fold_latin1) { \
2024 /* This folder implies Unicode rules, which in the range expressible \
2025 * by not UTF is the lower case, with the two exceptions, one of \
2026 * which should have been taken care of before calling this */ \
2027 assert(*uc != LATIN_SMALL_LETTER_SHARP_S); \
2028 uvc = toLOWER_L1(*uc); \
2029 if (UNLIKELY(uvc == MICRO_SIGN)) uvc = GREEK_SMALL_LETTER_MU; \
2032 /* raw data, will be folded later if needed */ \
2040 #define TRIE_LIST_PUSH(state,fid,ns) STMT_START { \
2041 if ( TRIE_LIST_CUR( state ) >=TRIE_LIST_LEN( state ) ) { \
2042 U32 ging = TRIE_LIST_LEN( state ) *= 2; \
2043 Renew( trie->states[ state ].trans.list, ging, reg_trie_trans_le ); \
2045 TRIE_LIST_ITEM( state, TRIE_LIST_CUR( state ) ).forid = fid; \
2046 TRIE_LIST_ITEM( state, TRIE_LIST_CUR( state ) ).newstate = ns; \
2047 TRIE_LIST_CUR( state )++; \
2050 #define TRIE_LIST_NEW(state) STMT_START { \
2051 Newxz( trie->states[ state ].trans.list, \
2052 4, reg_trie_trans_le ); \
2053 TRIE_LIST_CUR( state ) = 1; \
2054 TRIE_LIST_LEN( state ) = 4; \
2057 #define TRIE_HANDLE_WORD(state) STMT_START { \
2058 U16 dupe= trie->states[ state ].wordnum; \
2059 regnode * const noper_next = regnext( noper ); \
2062 /* store the word for dumping */ \
2064 if (OP(noper) != NOTHING) \
2065 tmp = newSVpvn_utf8(STRING(noper), STR_LEN(noper), UTF); \
2067 tmp = newSVpvn_utf8( "", 0, UTF ); \
2068 av_push( trie_words, tmp ); \
2072 trie->wordinfo[curword].prev = 0; \
2073 trie->wordinfo[curword].len = wordlen; \
2074 trie->wordinfo[curword].accept = state; \
2076 if ( noper_next < tail ) { \
2078 trie->jump = (U16 *) PerlMemShared_calloc( word_count + 1, \
2080 trie->jump[curword] = (U16)(noper_next - convert); \
2082 jumper = noper_next; \
2084 nextbranch= regnext(cur); \
2088 /* It's a dupe. Pre-insert into the wordinfo[].prev */\
2089 /* chain, so that when the bits of chain are later */\
2090 /* linked together, the dups appear in the chain */\
2091 trie->wordinfo[curword].prev = trie->wordinfo[dupe].prev; \
2092 trie->wordinfo[dupe].prev = curword; \
2094 /* we haven't inserted this word yet. */ \
2095 trie->states[ state ].wordnum = curword; \
2100 #define TRIE_TRANS_STATE(state,base,ucharcount,charid,special) \
2101 ( ( base + charid >= ucharcount \
2102 && base + charid < ubound \
2103 && state == trie->trans[ base - ucharcount + charid ].check \
2104 && trie->trans[ base - ucharcount + charid ].next ) \
2105 ? trie->trans[ base - ucharcount + charid ].next \
2106 : ( state==1 ? special : 0 ) \
2110 #define MADE_JUMP_TRIE 2
2111 #define MADE_EXACT_TRIE 4
2114 S_make_trie(pTHX_ RExC_state_t *pRExC_state, regnode *startbranch,
2115 regnode *first, regnode *last, regnode *tail,
2116 U32 word_count, U32 flags, U32 depth)
2118 /* first pass, loop through and scan words */
2119 reg_trie_data *trie;
2120 HV *widecharmap = NULL;
2121 AV *revcharmap = newAV();
2127 regnode *jumper = NULL;
2128 regnode *nextbranch = NULL;
2129 regnode *convert = NULL;
2130 U32 *prev_states; /* temp array mapping each state to previous one */
2131 /* we just use folder as a flag in utf8 */
2132 const U8 * folder = NULL;
2135 const U32 data_slot = add_data( pRExC_state, STR_WITH_LEN("tuuu"));
2136 AV *trie_words = NULL;
2137 /* along with revcharmap, this only used during construction but both are
2138 * useful during debugging so we store them in the struct when debugging.
2141 const U32 data_slot = add_data( pRExC_state, STR_WITH_LEN("tu"));
2142 STRLEN trie_charcount=0;
2144 SV *re_trie_maxbuff;
2145 GET_RE_DEBUG_FLAGS_DECL;
2147 PERL_ARGS_ASSERT_MAKE_TRIE;
2149 PERL_UNUSED_ARG(depth);
2153 case EXACT: case EXACTL: break;
2157 case EXACTFLU8: folder = PL_fold_latin1; break;
2158 case EXACTF: folder = PL_fold; break;
2159 default: Perl_croak( aTHX_ "panic! In trie construction, unknown node type %u %s", (unsigned) flags, PL_reg_name[flags] );
2162 trie = (reg_trie_data *) PerlMemShared_calloc( 1, sizeof(reg_trie_data) );
2164 trie->startstate = 1;
2165 trie->wordcount = word_count;
2166 RExC_rxi->data->data[ data_slot ] = (void*)trie;
2167 trie->charmap = (U16 *) PerlMemShared_calloc( 256, sizeof(U16) );
2168 if (flags == EXACT || flags == EXACTL)
2169 trie->bitmap = (char *) PerlMemShared_calloc( ANYOF_BITMAP_SIZE, 1 );
2170 trie->wordinfo = (reg_trie_wordinfo *) PerlMemShared_calloc(
2171 trie->wordcount+1, sizeof(reg_trie_wordinfo));
2174 trie_words = newAV();
2177 re_trie_maxbuff = get_sv(RE_TRIE_MAXBUF_NAME, 1);
2178 assert(re_trie_maxbuff);
2179 if (!SvIOK(re_trie_maxbuff)) {
2180 sv_setiv(re_trie_maxbuff, RE_TRIE_MAXBUF_INIT);
2182 DEBUG_TRIE_COMPILE_r({
2183 PerlIO_printf( Perl_debug_log,
2184 "%*smake_trie start==%d, first==%d, last==%d, tail==%d depth=%d\n",
2185 (int)depth * 2 + 2, "",
2186 REG_NODE_NUM(startbranch),REG_NODE_NUM(first),
2187 REG_NODE_NUM(last), REG_NODE_NUM(tail), (int)depth);
2190 /* Find the node we are going to overwrite */
2191 if ( first == startbranch && OP( last ) != BRANCH ) {
2192 /* whole branch chain */
2195 /* branch sub-chain */
2196 convert = NEXTOPER( first );
2199 /* -- First loop and Setup --
2201 We first traverse the branches and scan each word to determine if it
2202 contains widechars, and how many unique chars there are, this is
2203 important as we have to build a table with at least as many columns as we
2206 We use an array of integers to represent the character codes 0..255
2207 (trie->charmap) and we use a an HV* to store Unicode characters. We use
2208 the native representation of the character value as the key and IV's for
2211 *TODO* If we keep track of how many times each character is used we can
2212 remap the columns so that the table compression later on is more
2213 efficient in terms of memory by ensuring the most common value is in the
2214 middle and the least common are on the outside. IMO this would be better
2215 than a most to least common mapping as theres a decent chance the most
2216 common letter will share a node with the least common, meaning the node
2217 will not be compressible. With a middle is most common approach the worst
2218 case is when we have the least common nodes twice.
2222 for ( cur = first ; cur < last ; cur = regnext( cur ) ) {
2223 regnode *noper = NEXTOPER( cur );
2224 const U8 *uc = (U8*)STRING( noper );
2225 const U8 *e = uc + STR_LEN( noper );
2227 U32 wordlen = 0; /* required init */
2228 STRLEN minchars = 0;
2229 STRLEN maxchars = 0;
2230 bool set_bit = trie->bitmap ? 1 : 0; /*store the first char in the
2233 if (OP(noper) == NOTHING) {
2234 regnode *noper_next= regnext(noper);
2235 if (noper_next != tail && OP(noper_next) == flags) {
2237 uc= (U8*)STRING(noper);
2238 e= uc + STR_LEN(noper);
2239 trie->minlen= STR_LEN(noper);
2246 if ( set_bit ) { /* bitmap only alloced when !(UTF&&Folding) */
2247 TRIE_BITMAP_SET(trie,*uc); /* store the raw first byte
2248 regardless of encoding */
2249 if (OP( noper ) == EXACTFU_SS) {
2250 /* false positives are ok, so just set this */
2251 TRIE_BITMAP_SET(trie, LATIN_SMALL_LETTER_SHARP_S);
2254 for ( ; uc < e ; uc += len ) { /* Look at each char in the current
2256 TRIE_CHARCOUNT(trie)++;
2259 /* TRIE_READ_CHAR returns the current character, or its fold if /i
2260 * is in effect. Under /i, this character can match itself, or
2261 * anything that folds to it. If not under /i, it can match just
2262 * itself. Most folds are 1-1, for example k, K, and KELVIN SIGN
2263 * all fold to k, and all are single characters. But some folds
2264 * expand to more than one character, so for example LATIN SMALL
2265 * LIGATURE FFI folds to the three character sequence 'ffi'. If
2266 * the string beginning at 'uc' is 'ffi', it could be matched by
2267 * three characters, or just by the one ligature character. (It
2268 * could also be matched by two characters: LATIN SMALL LIGATURE FF
2269 * followed by 'i', or by 'f' followed by LATIN SMALL LIGATURE FI).
2270 * (Of course 'I' and/or 'F' instead of 'i' and 'f' can also
2271 * match.) The trie needs to know the minimum and maximum number
2272 * of characters that could match so that it can use size alone to
2273 * quickly reject many match attempts. The max is simple: it is
2274 * the number of folded characters in this branch (since a fold is
2275 * never shorter than what folds to it. */
2279 /* And the min is equal to the max if not under /i (indicated by
2280 * 'folder' being NULL), or there are no multi-character folds. If
2281 * there is a multi-character fold, the min is incremented just
2282 * once, for the character that folds to the sequence. Each
2283 * character in the sequence needs to be added to the list below of
2284 * characters in the trie, but we count only the first towards the
2285 * min number of characters needed. This is done through the
2286 * variable 'foldlen', which is returned by the macros that look
2287 * for these sequences as the number of bytes the sequence
2288 * occupies. Each time through the loop, we decrement 'foldlen' by
2289 * how many bytes the current char occupies. Only when it reaches
2290 * 0 do we increment 'minchars' or look for another multi-character
2292 if (folder == NULL) {
2295 else if (foldlen > 0) {
2296 foldlen -= (UTF) ? UTF8SKIP(uc) : 1;
2301 /* See if *uc is the beginning of a multi-character fold. If
2302 * so, we decrement the length remaining to look at, to account
2303 * for the current character this iteration. (We can use 'uc'
2304 * instead of the fold returned by TRIE_READ_CHAR because for
2305 * non-UTF, the latin1_safe macro is smart enough to account
2306 * for all the unfolded characters, and because for UTF, the
2307 * string will already have been folded earlier in the
2308 * compilation process */
2310 if ((foldlen = is_MULTI_CHAR_FOLD_utf8_safe(uc, e))) {
2311 foldlen -= UTF8SKIP(uc);
2314 else if ((foldlen = is_MULTI_CHAR_FOLD_latin1_safe(uc, e))) {
2319 /* The current character (and any potential folds) should be added
2320 * to the possible matching characters for this position in this
2324 U8 folded= folder[ (U8) uvc ];
2325 if ( !trie->charmap[ folded ] ) {
2326 trie->charmap[ folded ]=( ++trie->uniquecharcount );
2327 TRIE_STORE_REVCHAR( folded );
2330 if ( !trie->charmap[ uvc ] ) {
2331 trie->charmap[ uvc ]=( ++trie->uniquecharcount );
2332 TRIE_STORE_REVCHAR( uvc );
2335 /* store the codepoint in the bitmap, and its folded
2337 TRIE_BITMAP_SET(trie, uvc);
2339 /* store the folded codepoint */
2340 if ( folder ) TRIE_BITMAP_SET(trie, folder[(U8) uvc ]);
2343 /* store first byte of utf8 representation of
2344 variant codepoints */
2345 if (! UVCHR_IS_INVARIANT(uvc)) {
2346 TRIE_BITMAP_SET(trie, UTF8_TWO_BYTE_HI(uvc));
2349 set_bit = 0; /* We've done our bit :-) */
2353 /* XXX We could come up with the list of code points that fold
2354 * to this using PL_utf8_foldclosures, except not for
2355 * multi-char folds, as there may be multiple combinations
2356 * there that could work, which needs to wait until runtime to
2357 * resolve (The comment about LIGATURE FFI above is such an
2362 widecharmap = newHV();
2364 svpp = hv_fetch( widecharmap, (char*)&uvc, sizeof( UV ), 1 );
2367 Perl_croak( aTHX_ "error creating/fetching widecharmap entry for 0x%"UVXf, uvc );
2369 if ( !SvTRUE( *svpp ) ) {
2370 sv_setiv( *svpp, ++trie->uniquecharcount );
2371 TRIE_STORE_REVCHAR(uvc);
2374 } /* end loop through characters in this branch of the trie */
2376 /* We take the min and max for this branch and combine to find the min
2377 * and max for all branches processed so far */
2378 if( cur == first ) {
2379 trie->minlen = minchars;
2380 trie->maxlen = maxchars;
2381 } else if (minchars < trie->minlen) {
2382 trie->minlen = minchars;
2383 } else if (maxchars > trie->maxlen) {
2384 trie->maxlen = maxchars;
2386 } /* end first pass */
2387 DEBUG_TRIE_COMPILE_r(
2388 PerlIO_printf( Perl_debug_log,
2389 "%*sTRIE(%s): W:%d C:%d Uq:%d Min:%d Max:%d\n",
2390 (int)depth * 2 + 2,"",
2391 ( widecharmap ? "UTF8" : "NATIVE" ), (int)word_count,
2392 (int)TRIE_CHARCOUNT(trie), trie->uniquecharcount,
2393 (int)trie->minlen, (int)trie->maxlen )
2397 We now know what we are dealing with in terms of unique chars and
2398 string sizes so we can calculate how much memory a naive
2399 representation using a flat table will take. If it's over a reasonable
2400 limit (as specified by ${^RE_TRIE_MAXBUF}) we use a more memory
2401 conservative but potentially much slower representation using an array
2404 At the end we convert both representations into the same compressed
2405 form that will be used in regexec.c for matching with. The latter
2406 is a form that cannot be used to construct with but has memory
2407 properties similar to the list form and access properties similar
2408 to the table form making it both suitable for fast searches and
2409 small enough that its feasable to store for the duration of a program.
2411 See the comment in the code where the compressed table is produced
2412 inplace from the flat tabe representation for an explanation of how
2413 the compression works.
2418 Newx(prev_states, TRIE_CHARCOUNT(trie) + 2, U32);
2421 if ( (IV)( ( TRIE_CHARCOUNT(trie) + 1 ) * trie->uniquecharcount + 1)
2422 > SvIV(re_trie_maxbuff) )
2425 Second Pass -- Array Of Lists Representation
2427 Each state will be represented by a list of charid:state records
2428 (reg_trie_trans_le) the first such element holds the CUR and LEN
2429 points of the allocated array. (See defines above).
2431 We build the initial structure using the lists, and then convert
2432 it into the compressed table form which allows faster lookups
2433 (but cant be modified once converted).
2436 STRLEN transcount = 1;
2438 DEBUG_TRIE_COMPILE_MORE_r( PerlIO_printf( Perl_debug_log,
2439 "%*sCompiling trie using list compiler\n",
2440 (int)depth * 2 + 2, ""));
2442 trie->states = (reg_trie_state *)
2443 PerlMemShared_calloc( TRIE_CHARCOUNT(trie) + 2,
2444 sizeof(reg_trie_state) );
2448 for ( cur = first ; cur < last ; cur = regnext( cur ) ) {
2450 regnode *noper = NEXTOPER( cur );
2451 U8 *uc = (U8*)STRING( noper );
2452 const U8 *e = uc + STR_LEN( noper );
2453 U32 state = 1; /* required init */
2454 U16 charid = 0; /* sanity init */
2455 U32 wordlen = 0; /* required init */
2457 if (OP(noper) == NOTHING) {
2458 regnode *noper_next= regnext(noper);
2459 if (noper_next != tail && OP(noper_next) == flags) {
2461 uc= (U8*)STRING(noper);
2462 e= uc + STR_LEN(noper);
2466 if (OP(noper) != NOTHING) {
2467 for ( ; uc < e ; uc += len ) {
2472 charid = trie->charmap[ uvc ];
2474 SV** const svpp = hv_fetch( widecharmap,
2481 charid=(U16)SvIV( *svpp );
2484 /* charid is now 0 if we dont know the char read, or
2485 * nonzero if we do */
2492 if ( !trie->states[ state ].trans.list ) {
2493 TRIE_LIST_NEW( state );
2496 check <= TRIE_LIST_USED( state );
2499 if ( TRIE_LIST_ITEM( state, check ).forid
2502 newstate = TRIE_LIST_ITEM( state, check ).newstate;
2507 newstate = next_alloc++;
2508 prev_states[newstate] = state;
2509 TRIE_LIST_PUSH( state, charid, newstate );
2514 Perl_croak( aTHX_ "panic! In trie construction, no char mapping for %"IVdf, uvc );
2518 TRIE_HANDLE_WORD(state);
2520 } /* end second pass */
2522 /* next alloc is the NEXT state to be allocated */
2523 trie->statecount = next_alloc;
2524 trie->states = (reg_trie_state *)
2525 PerlMemShared_realloc( trie->states,
2527 * sizeof(reg_trie_state) );
2529 /* and now dump it out before we compress it */
2530 DEBUG_TRIE_COMPILE_MORE_r(dump_trie_interim_list(trie, widecharmap,
2531 revcharmap, next_alloc,
2535 trie->trans = (reg_trie_trans *)
2536 PerlMemShared_calloc( transcount, sizeof(reg_trie_trans) );
2543 for( state=1 ; state < next_alloc ; state ++ ) {
2547 DEBUG_TRIE_COMPILE_MORE_r(
2548 PerlIO_printf( Perl_debug_log, "tp: %d zp: %d ",tp,zp)
2552 if (trie->states[state].trans.list) {
2553 U16 minid=TRIE_LIST_ITEM( state, 1).forid;
2557 for( idx = 2 ; idx <= TRIE_LIST_USED( state ) ; idx++ ) {
2558 const U16 forid = TRIE_LIST_ITEM( state, idx).forid;
2559 if ( forid < minid ) {
2561 } else if ( forid > maxid ) {
2565 if ( transcount < tp + maxid - minid + 1) {
2567 trie->trans = (reg_trie_trans *)
2568 PerlMemShared_realloc( trie->trans,
2570 * sizeof(reg_trie_trans) );
2571 Zero( trie->trans + (transcount / 2),
2575 base = trie->uniquecharcount + tp - minid;
2576 if ( maxid == minid ) {
2578 for ( ; zp < tp ; zp++ ) {
2579 if ( ! trie->trans[ zp ].next ) {
2580 base = trie->uniquecharcount + zp - minid;
2581 trie->trans[ zp ].next = TRIE_LIST_ITEM( state,
2583 trie->trans[ zp ].check = state;
2589 trie->trans[ tp ].next = TRIE_LIST_ITEM( state,
2591 trie->trans[ tp ].check = state;
2596 for ( idx=1; idx <= TRIE_LIST_USED( state ) ; idx++ ) {
2597 const U32 tid = base
2598 - trie->uniquecharcount
2599 + TRIE_LIST_ITEM( state, idx ).forid;
2600 trie->trans[ tid ].next = TRIE_LIST_ITEM( state,
2602 trie->trans[ tid ].check = state;
2604 tp += ( maxid - minid + 1 );
2606 Safefree(trie->states[ state ].trans.list);
2609 DEBUG_TRIE_COMPILE_MORE_r(
2610 PerlIO_printf( Perl_debug_log, " base: %d\n",base);
2613 trie->states[ state ].trans.base=base;
2615 trie->lasttrans = tp + 1;
2619 Second Pass -- Flat Table Representation.
2621 we dont use the 0 slot of either trans[] or states[] so we add 1 to
2622 each. We know that we will need Charcount+1 trans at most to store
2623 the data (one row per char at worst case) So we preallocate both
2624 structures assuming worst case.
2626 We then construct the trie using only the .next slots of the entry
2629 We use the .check field of the first entry of the node temporarily
2630 to make compression both faster and easier by keeping track of how
2631 many non zero fields are in the node.
2633 Since trans are numbered from 1 any 0 pointer in the table is a FAIL
2636 There are two terms at use here: state as a TRIE_NODEIDX() which is
2637 a number representing the first entry of the node, and state as a
2638 TRIE_NODENUM() which is the trans number. state 1 is TRIE_NODEIDX(1)
2639 and TRIE_NODENUM(1), state 2 is TRIE_NODEIDX(2) and TRIE_NODENUM(3)
2640 if there are 2 entrys per node. eg:
2648 The table is internally in the right hand, idx form. However as we
2649 also have to deal with the states array which is indexed by nodenum
2650 we have to use TRIE_NODENUM() to convert.
2653 DEBUG_TRIE_COMPILE_MORE_r( PerlIO_printf( Perl_debug_log,
2654 "%*sCompiling trie using table compiler\n",
2655 (int)depth * 2 + 2, ""));
2657 trie->trans = (reg_trie_trans *)
2658 PerlMemShared_calloc( ( TRIE_CHARCOUNT(trie) + 1 )
2659 * trie->uniquecharcount + 1,
2660 sizeof(reg_trie_trans) );
2661 trie->states = (reg_trie_state *)
2662 PerlMemShared_calloc( TRIE_CHARCOUNT(trie) + 2,
2663 sizeof(reg_trie_state) );
2664 next_alloc = trie->uniquecharcount + 1;
2667 for ( cur = first ; cur < last ; cur = regnext( cur ) ) {
2669 regnode *noper = NEXTOPER( cur );
2670 const U8 *uc = (U8*)STRING( noper );
2671 const U8 *e = uc + STR_LEN( noper );
2673 U32 state = 1; /* required init */
2675 U16 charid = 0; /* sanity init */
2676 U32 accept_state = 0; /* sanity init */
2678 U32 wordlen = 0; /* required init */
2680 if (OP(noper) == NOTHING) {
2681 regnode *noper_next= regnext(noper);
2682 if (noper_next != tail && OP(noper_next) == flags) {
2684 uc= (U8*)STRING(noper);
2685 e= uc + STR_LEN(noper);
2689 if ( OP(noper) != NOTHING ) {
2690 for ( ; uc < e ; uc += len ) {
2695 charid = trie->charmap[ uvc ];
2697 SV* const * const svpp = hv_fetch( widecharmap,
2701 charid = svpp ? (U16)SvIV(*svpp) : 0;
2705 if ( !trie->trans[ state + charid ].next ) {
2706 trie->trans[ state + charid ].next = next_alloc;
2707 trie->trans[ state ].check++;
2708 prev_states[TRIE_NODENUM(next_alloc)]
2709 = TRIE_NODENUM(state);
2710 next_alloc += trie->uniquecharcount;
2712 state = trie->trans[ state + charid ].next;
2714 Perl_croak( aTHX_ "panic! In trie construction, no char mapping for %"IVdf, uvc );
2716 /* charid is now 0 if we dont know the char read, or
2717 * nonzero if we do */
2720 accept_state = TRIE_NODENUM( state );
2721 TRIE_HANDLE_WORD(accept_state);
2723 } /* end second pass */
2725 /* and now dump it out before we compress it */
2726 DEBUG_TRIE_COMPILE_MORE_r(dump_trie_interim_table(trie, widecharmap,
2728 next_alloc, depth+1));
2732 * Inplace compress the table.*
2734 For sparse data sets the table constructed by the trie algorithm will
2735 be mostly 0/FAIL transitions or to put it another way mostly empty.
2736 (Note that leaf nodes will not contain any transitions.)
2738 This algorithm compresses the tables by eliminating most such
2739 transitions, at the cost of a modest bit of extra work during lookup:
2741 - Each states[] entry contains a .base field which indicates the
2742 index in the state[] array wheres its transition data is stored.
2744 - If .base is 0 there are no valid transitions from that node.
2746 - If .base is nonzero then charid is added to it to find an entry in
2749 -If trans[states[state].base+charid].check!=state then the
2750 transition is taken to be a 0/Fail transition. Thus if there are fail
2751 transitions at the front of the node then the .base offset will point
2752 somewhere inside the previous nodes data (or maybe even into a node
2753 even earlier), but the .check field determines if the transition is
2757 The following process inplace converts the table to the compressed
2758 table: We first do not compress the root node 1,and mark all its
2759 .check pointers as 1 and set its .base pointer as 1 as well. This
2760 allows us to do a DFA construction from the compressed table later,
2761 and ensures that any .base pointers we calculate later are greater
2764 - We set 'pos' to indicate the first entry of the second node.
2766 - We then iterate over the columns of the node, finding the first and
2767 last used entry at l and m. We then copy l..m into pos..(pos+m-l),
2768 and set the .check pointers accordingly, and advance pos
2769 appropriately and repreat for the next node. Note that when we copy
2770 the next pointers we have to convert them from the original
2771 NODEIDX form to NODENUM form as the former is not valid post
2774 - If a node has no transitions used we mark its base as 0 and do not
2775 advance the pos pointer.
2777 - If a node only has one transition we use a second pointer into the
2778 structure to fill in allocated fail transitions from other states.
2779 This pointer is independent of the main pointer and scans forward
2780 looking for null transitions that are allocated to a state. When it
2781 finds one it writes the single transition into the "hole". If the
2782 pointer doesnt find one the single transition is appended as normal.
2784 - Once compressed we can Renew/realloc the structures to release the
2787 See "Table-Compression Methods" in sec 3.9 of the Red Dragon,
2788 specifically Fig 3.47 and the associated pseudocode.
2792 const U32 laststate = TRIE_NODENUM( next_alloc );
2795 trie->statecount = laststate;
2797 for ( state = 1 ; state < laststate ; state++ ) {
2799 const U32 stateidx = TRIE_NODEIDX( state );
2800 const U32 o_used = trie->trans[ stateidx ].check;
2801 U32 used = trie->trans[ stateidx ].check;
2802 trie->trans[ stateidx ].check = 0;
2805 used && charid < trie->uniquecharcount;
2808 if ( flag || trie->trans[ stateidx + charid ].next ) {
2809 if ( trie->trans[ stateidx + charid ].next ) {
2811 for ( ; zp < pos ; zp++ ) {
2812 if ( ! trie->trans[ zp ].next ) {
2816 trie->states[ state ].trans.base
2818 + trie->uniquecharcount
2820 trie->trans[ zp ].next
2821 = SAFE_TRIE_NODENUM( trie->trans[ stateidx
2823 trie->trans[ zp ].check = state;
2824 if ( ++zp > pos ) pos = zp;
2831 trie->states[ state ].trans.base
2832 = pos + trie->uniquecharcount - charid ;
2834 trie->trans[ pos ].next
2835 = SAFE_TRIE_NODENUM(
2836 trie->trans[ stateidx + charid ].next );
2837 trie->trans[ pos ].check = state;
2842 trie->lasttrans = pos + 1;
2843 trie->states = (reg_trie_state *)
2844 PerlMemShared_realloc( trie->states, laststate
2845 * sizeof(reg_trie_state) );
2846 DEBUG_TRIE_COMPILE_MORE_r(
2847 PerlIO_printf( Perl_debug_log,
2848 "%*sAlloc: %d Orig: %"IVdf" elements, Final:%"IVdf". Savings of %%%5.2f\n",
2849 (int)depth * 2 + 2,"",
2850 (int)( ( TRIE_CHARCOUNT(trie) + 1 ) * trie->uniquecharcount
2854 ( ( next_alloc - pos ) * 100 ) / (double)next_alloc );
2857 } /* end table compress */
2859 DEBUG_TRIE_COMPILE_MORE_r(
2860 PerlIO_printf(Perl_debug_log,
2861 "%*sStatecount:%"UVxf" Lasttrans:%"UVxf"\n",
2862 (int)depth * 2 + 2, "",
2863 (UV)trie->statecount,
2864 (UV)trie->lasttrans)
2866 /* resize the trans array to remove unused space */
2867 trie->trans = (reg_trie_trans *)
2868 PerlMemShared_realloc( trie->trans, trie->lasttrans
2869 * sizeof(reg_trie_trans) );
2871 { /* Modify the program and insert the new TRIE node */
2872 U8 nodetype =(U8)(flags & 0xFF);
2876 regnode *optimize = NULL;
2877 #ifdef RE_TRACK_PATTERN_OFFSETS
2880 U32 mjd_nodelen = 0;
2881 #endif /* RE_TRACK_PATTERN_OFFSETS */
2882 #endif /* DEBUGGING */
2884 This means we convert either the first branch or the first Exact,
2885 depending on whether the thing following (in 'last') is a branch
2886 or not and whther first is the startbranch (ie is it a sub part of
2887 the alternation or is it the whole thing.)
2888 Assuming its a sub part we convert the EXACT otherwise we convert
2889 the whole branch sequence, including the first.
2891 /* Find the node we are going to overwrite */
2892 if ( first != startbranch || OP( last ) == BRANCH ) {
2893 /* branch sub-chain */
2894 NEXT_OFF( first ) = (U16)(last - first);
2895 #ifdef RE_TRACK_PATTERN_OFFSETS
2897 mjd_offset= Node_Offset((convert));
2898 mjd_nodelen= Node_Length((convert));
2901 /* whole branch chain */
2903 #ifdef RE_TRACK_PATTERN_OFFSETS
2906 const regnode *nop = NEXTOPER( convert );
2907 mjd_offset= Node_Offset((nop));
2908 mjd_nodelen= Node_Length((nop));
2912 PerlIO_printf(Perl_debug_log,
2913 "%*sMJD offset:%"UVuf" MJD length:%"UVuf"\n",
2914 (int)depth * 2 + 2, "",
2915 (UV)mjd_offset, (UV)mjd_nodelen)
2918 /* But first we check to see if there is a common prefix we can
2919 split out as an EXACT and put in front of the TRIE node. */
2920 trie->startstate= 1;
2921 if ( trie->bitmap && !widecharmap && !trie->jump ) {
2923 for ( state = 1 ; state < trie->statecount-1 ; state++ ) {
2927 const U32 base = trie->states[ state ].trans.base;
2929 if ( trie->states[state].wordnum )
2932 for ( ofs = 0 ; ofs < trie->uniquecharcount ; ofs++ ) {
2933 if ( ( base + ofs >= trie->uniquecharcount ) &&
2934 ( base + ofs - trie->uniquecharcount < trie->lasttrans ) &&
2935 trie->trans[ base + ofs - trie->uniquecharcount ].check == state )
2937 if ( ++count > 1 ) {
2938 SV **tmp = av_fetch( revcharmap, ofs, 0);
2939 const U8 *ch = (U8*)SvPV_nolen_const( *tmp );
2940 if ( state == 1 ) break;
2942 Zero(trie->bitmap, ANYOF_BITMAP_SIZE, char);
2944 PerlIO_printf(Perl_debug_log,
2945 "%*sNew Start State=%"UVuf" Class: [",
2946 (int)depth * 2 + 2, "",
2949 SV ** const tmp = av_fetch( revcharmap, idx, 0);
2950 const U8 * const ch = (U8*)SvPV_nolen_const( *tmp );
2952 TRIE_BITMAP_SET(trie,*ch);
2954 TRIE_BITMAP_SET(trie, folder[ *ch ]);
2956 PerlIO_printf(Perl_debug_log, "%s", (char*)ch)
2960 TRIE_BITMAP_SET(trie,*ch);
2962 TRIE_BITMAP_SET(trie,folder[ *ch ]);
2963 DEBUG_OPTIMISE_r(PerlIO_printf( Perl_debug_log,"%s", ch));
2969 SV **tmp = av_fetch( revcharmap, idx, 0);
2971 char *ch = SvPV( *tmp, len );
2973 SV *sv=sv_newmortal();
2974 PerlIO_printf( Perl_debug_log,
2975 "%*sPrefix State: %"UVuf" Idx:%"UVuf" Char='%s'\n",
2976 (int)depth * 2 + 2, "",
2978 pv_pretty(sv, SvPV_nolen_const(*tmp), SvCUR(*tmp), 6,
2979 PL_colors[0], PL_colors[1],
2980 (SvUTF8(*tmp) ? PERL_PV_ESCAPE_UNI : 0) |
2981 PERL_PV_ESCAPE_FIRSTCHAR
2986 OP( convert ) = nodetype;
2987 str=STRING(convert);
2990 STR_LEN(convert) += len;
2996 DEBUG_OPTIMISE_r(PerlIO_printf( Perl_debug_log,"]\n"));
3001 trie->prefixlen = (state-1);
3003 regnode *n = convert+NODE_SZ_STR(convert);
3004 NEXT_OFF(convert) = NODE_SZ_STR(convert);
3005 trie->startstate = state;
3006 trie->minlen -= (state - 1);
3007 trie->maxlen -= (state - 1);
3009 /* At least the UNICOS C compiler choked on this
3010 * being argument to DEBUG_r(), so let's just have
3013 #ifdef PERL_EXT_RE_BUILD
3019 regnode *fix = convert;
3020 U32 word = trie->wordcount;
3022 Set_Node_Offset_Length(convert, mjd_offset, state - 1);
3023 while( ++fix < n ) {
3024 Set_Node_Offset_Length(fix, 0, 0);
3027 SV ** const tmp = av_fetch( trie_words, word, 0 );
3029 if ( STR_LEN(convert) <= SvCUR(*tmp) )
3030 sv_chop(*tmp, SvPV_nolen(*tmp) + STR_LEN(convert));
3032 sv_chop(*tmp, SvPV_nolen(*tmp) + SvCUR(*tmp));
3040 NEXT_OFF(convert) = (U16)(tail - convert);
3041 DEBUG_r(optimize= n);
3047 if ( trie->maxlen ) {
3048 NEXT_OFF( convert ) = (U16)(tail - convert);
3049 ARG_SET( convert, data_slot );
3050 /* Store the offset to the first unabsorbed branch in
3051 jump[0], which is otherwise unused by the jump logic.
3052 We use this when dumping a trie and during optimisation. */
3054 trie->jump[0] = (U16)(nextbranch - convert);
3056 /* If the start state is not accepting (meaning there is no empty string/NOTHING)
3057 * and there is a bitmap
3058 * and the first "jump target" node we found leaves enough room
3059 * then convert the TRIE node into a TRIEC node, with the bitmap
3060 * embedded inline in the opcode - this is hypothetically faster.
3062 if ( !trie->states[trie->startstate].wordnum
3064 && ( (char *)jumper - (char *)convert) >= (int)sizeof(struct regnode_charclass) )
3066 OP( convert ) = TRIEC;
3067 Copy(trie->bitmap, ((struct regnode_charclass *)convert)->bitmap, ANYOF_BITMAP_SIZE, char);
3068 PerlMemShared_free(trie->bitmap);
3071 OP( convert ) = TRIE;
3073 /* store the type in the flags */
3074 convert->flags = nodetype;
3078 + regarglen[ OP( convert ) ];
3080 /* XXX We really should free up the resource in trie now,
3081 as we won't use them - (which resources?) dmq */
3083 /* needed for dumping*/
3084 DEBUG_r(if (optimize) {
3085 regnode *opt = convert;
3087 while ( ++opt < optimize) {
3088 Set_Node_Offset_Length(opt,0,0);
3091 Try to clean up some of the debris left after the
3094 while( optimize < jumper ) {
3095 mjd_nodelen += Node_Length((optimize));
3096 OP( optimize ) = OPTIMIZED;
3097 Set_Node_Offset_Length(optimize,0,0);
3100 Set_Node_Offset_Length(convert,mjd_offset,mjd_nodelen);
3102 } /* end node insert */
3104 /* Finish populating the prev field of the wordinfo array. Walk back
3105 * from each accept state until we find another accept state, and if
3106 * so, point the first word's .prev field at the second word. If the
3107 * second already has a .prev field set, stop now. This will be the
3108 * case either if we've already processed that word's accept state,
3109 * or that state had multiple words, and the overspill words were
3110 * already linked up earlier.
3117 for (word=1; word <= trie->wordcount; word++) {
3119 if (trie->wordinfo[word].prev)
3121 state = trie->wordinfo[word].accept;
3123 state = prev_states[state];
3126 prev = trie->states[state].wordnum;
3130 trie->wordinfo[word].prev = prev;
3132 Safefree(prev_states);
3136 /* and now dump out the compressed format */
3137 DEBUG_TRIE_COMPILE_r(dump_trie(trie, widecharmap, revcharmap, depth+1));
3139 RExC_rxi->data->data[ data_slot + 1 ] = (void*)widecharmap;
3141 RExC_rxi->data->data[ data_slot + TRIE_WORDS_OFFSET ] = (void*)trie_words;
3142 RExC_rxi->data->data[ data_slot + 3 ] = (void*)revcharmap;
3144 SvREFCNT_dec_NN(revcharmap);
3148 : trie->startstate>1
3154 S_construct_ahocorasick_from_trie(pTHX_ RExC_state_t *pRExC_state, regnode *source, U32 depth)
3156 /* The Trie is constructed and compressed now so we can build a fail array if
3159 This is basically the Aho-Corasick algorithm. Its from exercise 3.31 and
3161 "Red Dragon" -- Compilers, principles, techniques, and tools. Aho, Sethi,
3165 We find the fail state for each state in the trie, this state is the longest
3166 proper suffix of the current state's 'word' that is also a proper prefix of
3167 another word in our trie. State 1 represents the word '' and is thus the
3168 default fail state. This allows the DFA not to have to restart after its
3169 tried and failed a word at a given point, it simply continues as though it
3170 had been matching the other word in the first place.
3172 'abcdgu'=~/abcdefg|cdgu/
3173 When we get to 'd' we are still matching the first word, we would encounter
3174 'g' which would fail, which would bring us to the state representing 'd' in
3175 the second word where we would try 'g' and succeed, proceeding to match
3178 /* add a fail transition */
3179 const U32 trie_offset = ARG(source);
3180 reg_trie_data *trie=(reg_trie_data *)RExC_rxi->data->data[trie_offset];
3182 const U32 ucharcount = trie->uniquecharcount;
3183 const U32 numstates = trie->statecount;
3184 const U32 ubound = trie->lasttrans + ucharcount;
3188 U32 base = trie->states[ 1 ].trans.base;
3191 const U32 data_slot = add_data( pRExC_state, STR_WITH_LEN("T"));
3193 GET_RE_DEBUG_FLAGS_DECL;
3195 PERL_ARGS_ASSERT_CONSTRUCT_AHOCORASICK_FROM_TRIE;
3196 PERL_UNUSED_CONTEXT;
3198 PERL_UNUSED_ARG(depth);
3201 if ( OP(source) == TRIE ) {
3202 struct regnode_1 *op = (struct regnode_1 *)
3203 PerlMemShared_calloc(1, sizeof(struct regnode_1));
3204 StructCopy(source,op,struct regnode_1);
3205 stclass = (regnode *)op;
3207 struct regnode_charclass *op = (struct regnode_charclass *)
3208 PerlMemShared_calloc(1, sizeof(struct regnode_charclass));
3209 StructCopy(source,op,struct regnode_charclass);
3210 stclass = (regnode *)op;
3212 OP(stclass)+=2; /* convert the TRIE type to its AHO-CORASICK equivalent */
3214 ARG_SET( stclass, data_slot );
3215 aho = (reg_ac_data *) PerlMemShared_calloc( 1, sizeof(reg_ac_data) );
3216 RExC_rxi->data->data[ data_slot ] = (void*)aho;
3217 aho->trie=trie_offset;
3218 aho->states=(reg_trie_state *)PerlMemShared_malloc( numstates * sizeof(reg_trie_state) );
3219 Copy( trie->states, aho->states, numstates, reg_trie_state );
3220 Newxz( q, numstates, U32);
3221 aho->fail = (U32 *) PerlMemShared_calloc( numstates, sizeof(U32) );
3224 /* initialize fail[0..1] to be 1 so that we always have
3225 a valid final fail state */
3226 fail[ 0 ] = fail[ 1 ] = 1;
3228 for ( charid = 0; charid < ucharcount ; charid++ ) {
3229 const U32 newstate = TRIE_TRANS_STATE( 1, base, ucharcount, charid, 0 );
3231 q[ q_write ] = newstate;
3232 /* set to point at the root */
3233 fail[ q[ q_write++ ] ]=1;
3236 while ( q_read < q_write) {
3237 const U32 cur = q[ q_read++ % numstates ];
3238 base = trie->states[ cur ].trans.base;
3240 for ( charid = 0 ; charid < ucharcount ; charid++ ) {
3241 const U32 ch_state = TRIE_TRANS_STATE( cur, base, ucharcount, charid, 1 );
3243 U32 fail_state = cur;
3246 fail_state = fail[ fail_state ];
3247 fail_base = aho->states[ fail_state ].trans.base;
3248 } while ( !TRIE_TRANS_STATE( fail_state, fail_base, ucharcount, charid, 1 ) );
3250 fail_state = TRIE_TRANS_STATE( fail_state, fail_base, ucharcount, charid, 1 );
3251 fail[ ch_state ] = fail_state;
3252 if ( !aho->states[ ch_state ].wordnum && aho->states[ fail_state ].wordnum )
3254 aho->states[ ch_state ].wordnum = aho->states[ fail_state ].wordnum;
3256 q[ q_write++ % numstates] = ch_state;
3260 /* restore fail[0..1] to 0 so that we "fall out" of the AC loop
3261 when we fail in state 1, this allows us to use the
3262 charclass scan to find a valid start char. This is based on the principle
3263 that theres a good chance the string being searched contains lots of stuff
3264 that cant be a start char.
3266 fail[ 0 ] = fail[ 1 ] = 0;
3267 DEBUG_TRIE_COMPILE_r({
3268 PerlIO_printf(Perl_debug_log,
3269 "%*sStclass Failtable (%"UVuf" states): 0",
3270 (int)(depth * 2), "", (UV)numstates
3272 for( q_read=1; q_read<numstates; q_read++ ) {
3273 PerlIO_printf(Perl_debug_log, ", %"UVuf, (UV)fail[q_read]);
3275 PerlIO_printf(Perl_debug_log, "\n");
3278 /*RExC_seen |= REG_TRIEDFA_SEEN;*/
3283 #define DEBUG_PEEP(str,scan,depth) \
3284 DEBUG_OPTIMISE_r({if (scan){ \
3285 regnode *Next = regnext(scan); \
3286 regprop(RExC_rx, RExC_mysv, scan, NULL, pRExC_state); \
3287 PerlIO_printf(Perl_debug_log, "%*s" str ">%3d: %s (%d)", \
3288 (int)depth*2, "", REG_NODE_NUM(scan), SvPV_nolen_const(RExC_mysv),\
3289 Next ? (REG_NODE_NUM(Next)) : 0 ); \
3290 DEBUG_SHOW_STUDY_FLAGS(flags," [ ","]");\
3291 PerlIO_printf(Perl_debug_log, "\n"); \
3294 /* The below joins as many adjacent EXACTish nodes as possible into a single
3295 * one. The regop may be changed if the node(s) contain certain sequences that
3296 * require special handling. The joining is only done if:
3297 * 1) there is room in the current conglomerated node to entirely contain the
3299 * 2) they are the exact same node type
3301 * The adjacent nodes actually may be separated by NOTHING-kind nodes, and
3302 * these get optimized out
3304 * If a node is to match under /i (folded), the number of characters it matches
3305 * can be different than its character length if it contains a multi-character
3306 * fold. *min_subtract is set to the total delta number of characters of the
3309 * And *unfolded_multi_char is set to indicate whether or not the node contains
3310 * an unfolded multi-char fold. This happens when whether the fold is valid or
3311 * not won't be known until runtime; namely for EXACTF nodes that contain LATIN
3312 * SMALL LETTER SHARP S, as only if the target string being matched against
3313 * turns out to be UTF-8 is that fold valid; and also for EXACTFL nodes whose
3314 * folding rules depend on the locale in force at runtime. (Multi-char folds
3315 * whose components are all above the Latin1 range are not run-time locale
3316 * dependent, and have already been folded by the time this function is
3319 * This is as good a place as any to discuss the design of handling these
3320 * multi-character fold sequences. It's been wrong in Perl for a very long
3321 * time. There are three code points in Unicode whose multi-character folds
3322 * were long ago discovered to mess things up. The previous designs for
3323 * dealing with these involved assigning a special node for them. This
3324 * approach doesn't always work, as evidenced by this example:
3325 * "\xDFs" =~ /s\xDF/ui # Used to fail before these patches
3326 * Both sides fold to "sss", but if the pattern is parsed to create a node that
3327 * would match just the \xDF, it won't be able to handle the case where a
3328 * successful match would have to cross the node's boundary. The new approach
3329 * that hopefully generally solves the problem generates an EXACTFU_SS node
3330 * that is "sss" in this case.
3332 * It turns out that there are problems with all multi-character folds, and not
3333 * just these three. Now the code is general, for all such cases. The
3334 * approach taken is:
3335 * 1) This routine examines each EXACTFish node that could contain multi-
3336 * character folded sequences. Since a single character can fold into
3337 * such a sequence, the minimum match length for this node is less than
3338 * the number of characters in the node. This routine returns in
3339 * *min_subtract how many characters to subtract from the the actual
3340 * length of the string to get a real minimum match length; it is 0 if
3341 * there are no multi-char foldeds. This delta is used by the caller to
3342 * adjust the min length of the match, and the delta between min and max,
3343 * so that the optimizer doesn't reject these possibilities based on size
3345 * 2) For the sequence involving the Sharp s (\xDF), the node type EXACTFU_SS
3346 * is used for an EXACTFU node that contains at least one "ss" sequence in
3347 * it. For non-UTF-8 patterns and strings, this is the only case where
3348 * there is a possible fold length change. That means that a regular
3349 * EXACTFU node without UTF-8 involvement doesn't have to concern itself
3350 * with length changes, and so can be processed faster. regexec.c takes
3351 * advantage of this. Generally, an EXACTFish node that is in UTF-8 is
3352 * pre-folded by regcomp.c (except EXACTFL, some of whose folds aren't
3353 * known until runtime). This saves effort in regex matching. However,
3354 * the pre-folding isn't done for non-UTF8 patterns because the fold of
3355 * the MICRO SIGN requires UTF-8, and we don't want to slow things down by
3356 * forcing the pattern into UTF8 unless necessary. Also what EXACTF (and,
3357 * again, EXACTFL) nodes fold to isn't known until runtime. The fold
3358 * possibilities for the non-UTF8 patterns are quite simple, except for
3359 * the sharp s. All the ones that don't involve a UTF-8 target string are
3360 * members of a fold-pair, and arrays are set up for all of them so that
3361 * the other member of the pair can be found quickly. Code elsewhere in
3362 * this file makes sure that in EXACTFU nodes, the sharp s gets folded to
3363 * 'ss', even if the pattern isn't UTF-8. This avoids the issues
3364 * described in the next item.
3365 * 3) A problem remains for unfolded multi-char folds. (These occur when the
3366 * validity of the fold won't be known until runtime, and so must remain
3367 * unfolded for now. This happens for the sharp s in EXACTF and EXACTFA
3368 * nodes when the pattern isn't in UTF-8. (Note, BTW, that there cannot
3369 * be an EXACTF node with a UTF-8 pattern.) They also occur for various
3370 * folds in EXACTFL nodes, regardless of the UTF-ness of the pattern.)
3371 * The reason this is a problem is that the optimizer part of regexec.c
3372 * (probably unwittingly, in Perl_regexec_flags()) makes an assumption
3373 * that a character in the pattern corresponds to at most a single
3374 * character in the target string. (And I do mean character, and not byte
3375 * here, unlike other parts of the documentation that have never been
3376 * updated to account for multibyte Unicode.) sharp s in EXACTF and
3377 * EXACTFL nodes can match the two character string 'ss'; in EXACTFA nodes
3378 * it can match "\x{17F}\x{17F}". These, along with other ones in EXACTFL
3379 * nodes, violate the assumption, and they are the only instances where it
3380 * is violated. I'm reluctant to try to change the assumption, as the
3381 * code involved is impenetrable to me (khw), so instead the code here
3382 * punts. This routine examines EXACTFL nodes, and (when the pattern
3383 * isn't UTF-8) EXACTF and EXACTFA for such unfolded folds, and returns a
3384 * boolean indicating whether or not the node contains such a fold. When
3385 * it is true, the caller sets a flag that later causes the optimizer in
3386 * this file to not set values for the floating and fixed string lengths,
3387 * and thus avoids the optimizer code in regexec.c that makes the invalid
3388 * assumption. Thus, there is no optimization based on string lengths for
3389 * EXACTFL nodes that contain these few folds, nor for non-UTF8-pattern
3390 * EXACTF and EXACTFA nodes that contain the sharp s. (The reason the
3391 * assumption is wrong only in these cases is that all other non-UTF-8
3392 * folds are 1-1; and, for UTF-8 patterns, we pre-fold all other folds to
3393 * their expanded versions. (Again, we can't prefold sharp s to 'ss' in
3394 * EXACTF nodes because we don't know at compile time if it actually
3395 * matches 'ss' or not. For EXACTF nodes it will match iff the target
3396 * string is in UTF-8. This is in contrast to EXACTFU nodes, where it
3397 * always matches; and EXACTFA where it never does. In an EXACTFA node in
3398 * a UTF-8 pattern, sharp s is folded to "\x{17F}\x{17F}, avoiding the
3399 * problem; but in a non-UTF8 pattern, folding it to that above-Latin1
3400 * string would require the pattern to be forced into UTF-8, the overhead
3401 * of which we want to avoid. Similarly the unfolded multi-char folds in
3402 * EXACTFL nodes will match iff the locale at the time of match is a UTF-8
3405 * Similarly, the code that generates tries doesn't currently handle
3406 * not-already-folded multi-char folds, and it looks like a pain to change
3407 * that. Therefore, trie generation of EXACTFA nodes with the sharp s
3408 * doesn't work. Instead, such an EXACTFA is turned into a new regnode,
3409 * EXACTFA_NO_TRIE, which the trie code knows not to handle. Most people
3410 * using /iaa matching will be doing so almost entirely with ASCII
3411 * strings, so this should rarely be encountered in practice */
3413 #define JOIN_EXACT(scan,min_subtract,unfolded_multi_char, flags) \
3414 if (PL_regkind[OP(scan)] == EXACT) \
3415 join_exact(pRExC_state,(scan),(min_subtract),unfolded_multi_char, (flags),NULL,depth+1)
3418 S_join_exact(pTHX_ RExC_state_t *pRExC_state, regnode *scan,
3419 UV *min_subtract, bool *unfolded_multi_char,
3420 U32 flags,regnode *val, U32 depth)
3422 /* Merge several consecutive EXACTish nodes into one. */
3423 regnode *n = regnext(scan);
3425 regnode *next = scan + NODE_SZ_STR(scan);
3429 regnode *stop = scan;
3430 GET_RE_DEBUG_FLAGS_DECL;
3432 PERL_UNUSED_ARG(depth);
3435 PERL_ARGS_ASSERT_JOIN_EXACT;
3436 #ifndef EXPERIMENTAL_INPLACESCAN
3437 PERL_UNUSED_ARG(flags);
3438 PERL_UNUSED_ARG(val);
3440 DEBUG_PEEP("join",scan,depth);
3442 /* Look through the subsequent nodes in the chain. Skip NOTHING, merge
3443 * EXACT ones that are mergeable to the current one. */
3445 && (PL_regkind[OP(n)] == NOTHING
3446 || (stringok && OP(n) == OP(scan)))
3448 && NEXT_OFF(scan) + NEXT_OFF(n) < I16_MAX)
3451 if (OP(n) == TAIL || n > next)
3453 if (PL_regkind[OP(n)] == NOTHING) {
3454 DEBUG_PEEP("skip:",n,depth);
3455 NEXT_OFF(scan) += NEXT_OFF(n);
3456 next = n + NODE_STEP_REGNODE;
3463 else if (stringok) {
3464 const unsigned int oldl = STR_LEN(scan);
3465 regnode * const nnext = regnext(n);
3467 /* XXX I (khw) kind of doubt that this works on platforms (should
3468 * Perl ever run on one) where U8_MAX is above 255 because of lots
3469 * of other assumptions */
3470 /* Don't join if the sum can't fit into a single node */
3471 if (oldl + STR_LEN(n) > U8_MAX)
3474 DEBUG_PEEP("merg",n,depth);
3477 NEXT_OFF(scan) += NEXT_OFF(n);
3478 STR_LEN(scan) += STR_LEN(n);
3479 next = n + NODE_SZ_STR(n);
3480 /* Now we can overwrite *n : */
3481 Move(STRING(n), STRING(scan) + oldl, STR_LEN(n), char);
3489 #ifdef EXPERIMENTAL_INPLACESCAN
3490 if (flags && !NEXT_OFF(n)) {
3491 DEBUG_PEEP("atch", val, depth);
3492 if (reg_off_by_arg[OP(n)]) {
3493 ARG_SET(n, val - n);
3496 NEXT_OFF(n) = val - n;
3504 *unfolded_multi_char = FALSE;
3506 /* Here, all the adjacent mergeable EXACTish nodes have been merged. We
3507 * can now analyze for sequences of problematic code points. (Prior to
3508 * this final joining, sequences could have been split over boundaries, and
3509 * hence missed). The sequences only happen in folding, hence for any
3510 * non-EXACT EXACTish node */
3511 if (OP(scan) != EXACT && OP(scan) != EXACTL) {
3512 U8* s0 = (U8*) STRING(scan);
3514 U8* s_end = s0 + STR_LEN(scan);
3516 int total_count_delta = 0; /* Total delta number of characters that
3517 multi-char folds expand to */
3519 /* One pass is made over the node's string looking for all the
3520 * possibilities. To avoid some tests in the loop, there are two main
3521 * cases, for UTF-8 patterns (which can't have EXACTF nodes) and
3526 if (OP(scan) == EXACTFL) {
3529 /* An EXACTFL node would already have been changed to another
3530 * node type unless there is at least one character in it that
3531 * is problematic; likely a character whose fold definition
3532 * won't be known until runtime, and so has yet to be folded.
3533 * For all but the UTF-8 locale, folds are 1-1 in length, but
3534 * to handle the UTF-8 case, we need to create a temporary
3535 * folded copy using UTF-8 locale rules in order to analyze it.
3536 * This is because our macros that look to see if a sequence is
3537 * a multi-char fold assume everything is folded (otherwise the
3538 * tests in those macros would be too complicated and slow).
3539 * Note that here, the non-problematic folds will have already
3540 * been done, so we can just copy such characters. We actually
3541 * don't completely fold the EXACTFL string. We skip the
3542 * unfolded multi-char folds, as that would just create work
3543 * below to figure out the size they already are */
3545 Newx(folded, UTF8_MAX_FOLD_CHAR_EXPAND * STR_LEN(scan) + 1, U8);
3548 STRLEN s_len = UTF8SKIP(s);
3549 if (! is_PROBLEMATIC_LOCALE_FOLD_utf8(s)) {
3550 Copy(s, d, s_len, U8);
3553 else if (is_FOLDS_TO_MULTI_utf8(s)) {
3554 *unfolded_multi_char = TRUE;
3555 Copy(s, d, s_len, U8);
3558 else if (isASCII(*s)) {
3559 *(d++) = toFOLD(*s);
3563 _to_utf8_fold_flags(s, d, &len, FOLD_FLAGS_FULL);
3569 /* Point the remainder of the routine to look at our temporary
3573 } /* End of creating folded copy of EXACTFL string */
3575 /* Examine the string for a multi-character fold sequence. UTF-8
3576 * patterns have all characters pre-folded by the time this code is
3578 while (s < s_end - 1) /* Can stop 1 before the end, as minimum
3579 length sequence we are looking for is 2 */
3581 int count = 0; /* How many characters in a multi-char fold */
3582 int len = is_MULTI_CHAR_FOLD_utf8_safe(s, s_end);
3583 if (! len) { /* Not a multi-char fold: get next char */
3588 /* Nodes with 'ss' require special handling, except for
3589 * EXACTFA-ish for which there is no multi-char fold to this */
3590 if (len == 2 && *s == 's' && *(s+1) == 's'
3591 && OP(scan) != EXACTFA
3592 && OP(scan) != EXACTFA_NO_TRIE)
3595 if (OP(scan) != EXACTFL) {
3596 OP(scan) = EXACTFU_SS;
3600 else { /* Here is a generic multi-char fold. */
3601 U8* multi_end = s + len;
3603 /* Count how many characters are in it. In the case of
3604 * /aa, no folds which contain ASCII code points are
3605 * allowed, so check for those, and skip if found. */
3606 if (OP(scan) != EXACTFA && OP(scan) != EXACTFA_NO_TRIE) {
3607 count = utf8_length(s, multi_end);
3611 while (s < multi_end) {
3614 goto next_iteration;
3624 /* The delta is how long the sequence is minus 1 (1 is how long
3625 * the character that folds to the sequence is) */
3626 total_count_delta += count - 1;
3630 /* We created a temporary folded copy of the string in EXACTFL
3631 * nodes. Therefore we need to be sure it doesn't go below zero,
3632 * as the real string could be shorter */
3633 if (OP(scan) == EXACTFL) {
3634 int total_chars = utf8_length((U8*) STRING(scan),
3635 (U8*) STRING(scan) + STR_LEN(scan));
3636 if (total_count_delta > total_chars) {
3637 total_count_delta = total_chars;
3641 *min_subtract += total_count_delta;
3644 else if (OP(scan) == EXACTFA) {
3646 /* Non-UTF-8 pattern, EXACTFA node. There can't be a multi-char
3647 * fold to the ASCII range (and there are no existing ones in the
3648 * upper latin1 range). But, as outlined in the comments preceding
3649 * this function, we need to flag any occurrences of the sharp s.
3650 * This character forbids trie formation (because of added
3653 if (*s == LATIN_SMALL_LETTER_SHARP_S) {
3654 OP(scan) = EXACTFA_NO_TRIE;
3655 *unfolded_multi_char = TRUE;
3664 /* Non-UTF-8 pattern, not EXACTFA node. Look for the multi-char
3665 * folds that are all Latin1. As explained in the comments
3666 * preceding this function, we look also for the sharp s in EXACTF
3667 * and EXACTFL nodes; it can be in the final position. Otherwise
3668 * we can stop looking 1 byte earlier because have to find at least
3669 * two characters for a multi-fold */
3670 const U8* upper = (OP(scan) == EXACTF || OP(scan) == EXACTFL)
3675 int len = is_MULTI_CHAR_FOLD_latin1_safe(s, s_end);
3676 if (! len) { /* Not a multi-char fold. */
3677 if (*s == LATIN_SMALL_LETTER_SHARP_S
3678 && (OP(scan) == EXACTF || OP(scan) == EXACTFL))
3680 *unfolded_multi_char = TRUE;
3687 && isALPHA_FOLD_EQ(*s, 's')
3688 && isALPHA_FOLD_EQ(*(s+1), 's'))
3691 /* EXACTF nodes need to know that the minimum length
3692 * changed so that a sharp s in the string can match this
3693 * ss in the pattern, but they remain EXACTF nodes, as they
3694 * won't match this unless the target string is is UTF-8,
3695 * which we don't know until runtime. EXACTFL nodes can't
3696 * transform into EXACTFU nodes */
3697 if (OP(scan) != EXACTF && OP(scan) != EXACTFL) {
3698 OP(scan) = EXACTFU_SS;
3702 *min_subtract += len - 1;
3709 /* Allow dumping but overwriting the collection of skipped
3710 * ops and/or strings with fake optimized ops */
3711 n = scan + NODE_SZ_STR(scan);
3719 DEBUG_OPTIMISE_r(if (merged){DEBUG_PEEP("finl",scan,depth)});
3723 /* REx optimizer. Converts nodes into quicker variants "in place".
3724 Finds fixed substrings. */
3726 /* Stops at toplevel WHILEM as well as at "last". At end *scanp is set
3727 to the position after last scanned or to NULL. */
3729 #define INIT_AND_WITHP \
3730 assert(!and_withp); \
3731 Newx(and_withp,1, regnode_ssc); \
3732 SAVEFREEPV(and_withp)
3736 S_unwind_scan_frames(pTHX_ const void *p)
3738 scan_frame *f= (scan_frame *)p;
3740 scan_frame *n= f->next_frame;
3748 S_study_chunk(pTHX_ RExC_state_t *pRExC_state, regnode **scanp,
3749 SSize_t *minlenp, SSize_t *deltap,
3754 regnode_ssc *and_withp,
3755 U32 flags, U32 depth)
3756 /* scanp: Start here (read-write). */
3757 /* deltap: Write maxlen-minlen here. */
3758 /* last: Stop before this one. */
3759 /* data: string data about the pattern */
3760 /* stopparen: treat close N as END */
3761 /* recursed: which subroutines have we recursed into */
3762 /* and_withp: Valid if flags & SCF_DO_STCLASS_OR */
3764 /* There must be at least this number of characters to match */
3767 regnode *scan = *scanp, *next;
3769 int is_inf = (flags & SCF_DO_SUBSTR) && (data->flags & SF_IS_INF);
3770 int is_inf_internal = 0; /* The studied chunk is infinite */
3771 I32 is_par = OP(scan) == OPEN ? ARG(scan) : 0;
3772 scan_data_t data_fake;
3773 SV *re_trie_maxbuff = NULL;
3774 regnode *first_non_open = scan;
3775 SSize_t stopmin = SSize_t_MAX;
3776 scan_frame *frame = NULL;
3777 GET_RE_DEBUG_FLAGS_DECL;
3779 PERL_ARGS_ASSERT_STUDY_CHUNK;
3783 while (first_non_open && OP(first_non_open) == OPEN)
3784 first_non_open=regnext(first_non_open);
3790 RExC_study_chunk_recursed_count++;
3792 DEBUG_OPTIMISE_MORE_r(
3794 PerlIO_printf(Perl_debug_log,
3795 "%*sstudy_chunk stopparen=%ld recursed_count=%lu depth=%lu recursed_depth=%lu scan=%p last=%p",
3796 (int)(depth*2), "", (long)stopparen,
3797 (unsigned long)RExC_study_chunk_recursed_count,
3798 (unsigned long)depth, (unsigned long)recursed_depth,
3801 if (recursed_depth) {
3804 for ( j = 0 ; j < recursed_depth ; j++ ) {
3805 for ( i = 0 ; i < (U32)RExC_npar ; i++ ) {
3807 PAREN_TEST(RExC_study_chunk_recursed +
3808 ( j * RExC_study_chunk_recursed_bytes), i )
3811 !PAREN_TEST(RExC_study_chunk_recursed +
3812 (( j - 1 ) * RExC_study_chunk_recursed_bytes), i)
3815 PerlIO_printf(Perl_debug_log," %d",(int)i);
3819 if ( j + 1 < recursed_depth ) {
3820 PerlIO_printf(Perl_debug_log, ",");
3824 PerlIO_printf(Perl_debug_log,"\n");
3827 while ( scan && OP(scan) != END && scan < last ){
3828 UV min_subtract = 0; /* How mmany chars to subtract from the minimum
3829 node length to get a real minimum (because
3830 the folded version may be shorter) */
3831 bool unfolded_multi_char = FALSE;
3832 /* Peephole optimizer: */
3833 DEBUG_STUDYDATA("Peep:", data, depth);
3834 DEBUG_PEEP("Peep", scan, depth);
3837 /* The reason we do this here we need to deal with things like /(?:f)(?:o)(?:o)/
3838 * which cant be dealt with by the normal EXACT parsing code, as each (?:..) is handled
3839 * by a different invocation of reg() -- Yves
3841 JOIN_EXACT(scan,&min_subtract, &unfolded_multi_char, 0);
3843 /* Follow the next-chain of the current node and optimize
3844 away all the NOTHINGs from it. */
3845 if (OP(scan) != CURLYX) {
3846 const int max = (reg_off_by_arg[OP(scan)]
3848 /* I32 may be smaller than U16 on CRAYs! */
3849 : (I32_MAX < U16_MAX ? I32_MAX : U16_MAX));
3850 int off = (reg_off_by_arg[OP(scan)] ? ARG(scan) : NEXT_OFF(scan));
3854 /* Skip NOTHING and LONGJMP. */
3855 while ((n = regnext(n))
3856 && ((PL_regkind[OP(n)] == NOTHING && (noff = NEXT_OFF(n)))
3857 || ((OP(n) == LONGJMP) && (noff = ARG(n))))
3858 && off + noff < max)
3860 if (reg_off_by_arg[OP(scan)])
3863 NEXT_OFF(scan) = off;
3866 /* The principal pseudo-switch. Cannot be a switch, since we
3867 look into several different things. */
3868 if ( OP(scan) == DEFINEP ) {
3870 SSize_t deltanext = 0;
3871 SSize_t fake_last_close = 0;
3872 I32 f = SCF_IN_DEFINE;
3874 StructCopy(&zero_scan_data, &data_fake, scan_data_t);
3875 scan = regnext(scan);
3876 assert( OP(scan) == IFTHEN );
3877 DEBUG_PEEP("expect IFTHEN", scan, depth);
3879 data_fake.last_closep= &fake_last_close;
3881 next = regnext(scan);
3882 scan = NEXTOPER(NEXTOPER(scan));
3883 DEBUG_PEEP("scan", scan, depth);
3884 DEBUG_PEEP("next", next, depth);
3886 /* we suppose the run is continuous, last=next...
3887 * NOTE we dont use the return here! */
3888 (void)study_chunk(pRExC_state, &scan, &minlen,
3889 &deltanext, next, &data_fake, stopparen,
3890 recursed_depth, NULL, f, depth+1);
3895 OP(scan) == BRANCH ||
3896 OP(scan) == BRANCHJ ||
3899 next = regnext(scan);
3902 /* The op(next)==code check below is to see if we
3903 * have "BRANCH-BRANCH", "BRANCHJ-BRANCHJ", "IFTHEN-IFTHEN"
3904 * IFTHEN is special as it might not appear in pairs.
3905 * Not sure whether BRANCH-BRANCHJ is possible, regardless
3906 * we dont handle it cleanly. */
3907 if (OP(next) == code || code == IFTHEN) {
3908 /* NOTE - There is similar code to this block below for
3909 * handling TRIE nodes on a re-study. If you change stuff here
3910 * check there too. */
3911 SSize_t max1 = 0, min1 = SSize_t_MAX, num = 0;
3913 regnode * const startbranch=scan;
3915 if (flags & SCF_DO_SUBSTR) {
3916 /* Cannot merge strings after this. */
3917 scan_commit(pRExC_state, data, minlenp, is_inf);
3920 if (flags & SCF_DO_STCLASS)
3921 ssc_init_zero(pRExC_state, &accum);
3923 while (OP(scan) == code) {
3924 SSize_t deltanext, minnext, fake;
3926 regnode_ssc this_class;
3928 DEBUG_PEEP("Branch", scan, depth);
3931 StructCopy(&zero_scan_data, &data_fake, scan_data_t);
3933 data_fake.whilem_c = data->whilem_c;
3934 data_fake.last_closep = data->last_closep;
3937 data_fake.last_closep = &fake;
3939 data_fake.pos_delta = delta;
3940 next = regnext(scan);
3942 scan = NEXTOPER(scan); /* everything */
3943 if (code != BRANCH) /* everything but BRANCH */
3944 scan = NEXTOPER(scan);
3946 if (flags & SCF_DO_STCLASS) {
3947 ssc_init(pRExC_state, &this_class);
3948 data_fake.start_class = &this_class;
3949 f = SCF_DO_STCLASS_AND;
3951 if (flags & SCF_WHILEM_VISITED_POS)
3952 f |= SCF_WHILEM_VISITED_POS;
3954 /* we suppose the run is continuous, last=next...*/
3955 minnext = study_chunk(pRExC_state, &scan, minlenp,
3956 &deltanext, next, &data_fake, stopparen,
3957 recursed_depth, NULL, f,depth+1);
3961 if (deltanext == SSize_t_MAX) {
3962 is_inf = is_inf_internal = 1;
3964 } else if (max1 < minnext + deltanext)
3965 max1 = minnext + deltanext;
3967 if (data_fake.flags & (SF_HAS_PAR|SF_IN_PAR))
3969 if (data_fake.flags & SCF_SEEN_ACCEPT) {
3970 if ( stopmin > minnext)
3971 stopmin = min + min1;
3972 flags &= ~SCF_DO_SUBSTR;
3974 data->flags |= SCF_SEEN_ACCEPT;
3977 if (data_fake.flags & SF_HAS_EVAL)
3978 data->flags |= SF_HAS_EVAL;
3979 data->whilem_c = data_fake.whilem_c;
3981 if (flags & SCF_DO_STCLASS)
3982 ssc_or(pRExC_state, &accum, (regnode_charclass*)&this_class);
3984 if (code == IFTHEN && num < 2) /* Empty ELSE branch */
3986 if (flags & SCF_DO_SUBSTR) {
3987 data->pos_min += min1;
3988 if (data->pos_delta >= SSize_t_MAX - (max1 - min1))
3989 data->pos_delta = SSize_t_MAX;
3991 data->pos_delta += max1 - min1;
3992 if (max1 != min1 || is_inf)
3993 data->longest = &(data->longest_float);
3996 if (delta == SSize_t_MAX
3997 || SSize_t_MAX - delta - (max1 - min1) < 0)
3998 delta = SSize_t_MAX;
4000 delta += max1 - min1;
4001 if (flags & SCF_DO_STCLASS_OR) {
4002 ssc_or(pRExC_state, data->start_class, (regnode_charclass*) &accum);
4004 ssc_and(pRExC_state, data->start_class, (regnode_charclass *) and_withp);
4005 flags &= ~SCF_DO_STCLASS;
4008 else if (flags & SCF_DO_STCLASS_AND) {
4010 ssc_and(pRExC_state, data->start_class, (regnode_charclass *) &accum);
4011 flags &= ~SCF_DO_STCLASS;
4014 /* Switch to OR mode: cache the old value of
4015 * data->start_class */
4017 StructCopy(data->start_class, and_withp, regnode_ssc);
4018 flags &= ~SCF_DO_STCLASS_AND;
4019 StructCopy(&accum, data->start_class, regnode_ssc);
4020 flags |= SCF_DO_STCLASS_OR;
4024 if (PERL_ENABLE_TRIE_OPTIMISATION &&
4025 OP( startbranch ) == BRANCH )
4029 Assuming this was/is a branch we are dealing with: 'scan'
4030 now points at the item that follows the branch sequence,
4031 whatever it is. We now start at the beginning of the
4032 sequence and look for subsequences of
4038 which would be constructed from a pattern like
4041 If we can find such a subsequence we need to turn the first
4042 element into a trie and then add the subsequent branch exact
4043 strings to the trie.
4047 1. patterns where the whole set of branches can be
4050 2. patterns where only a subset can be converted.
4052 In case 1 we can replace the whole set with a single regop
4053 for the trie. In case 2 we need to keep the start and end
4056 'BRANCH EXACT; BRANCH EXACT; BRANCH X'
4057 becomes BRANCH TRIE; BRANCH X;
4059 There is an additional case, that being where there is a
4060 common prefix, which gets split out into an EXACT like node
4061 preceding the TRIE node.
4063 If x(1..n)==tail then we can do a simple trie, if not we make
4064 a "jump" trie, such that when we match the appropriate word
4065 we "jump" to the appropriate tail node. Essentially we turn
4066 a nested if into a case structure of sorts.
4071 if (!re_trie_maxbuff) {
4072 re_trie_maxbuff = get_sv(RE_TRIE_MAXBUF_NAME, 1);
4073 if (!SvIOK(re_trie_maxbuff))
4074 sv_setiv(re_trie_maxbuff, RE_TRIE_MAXBUF_INIT);
4076 if ( SvIV(re_trie_maxbuff)>=0 ) {
4078 regnode *first = (regnode *)NULL;
4079 regnode *last = (regnode *)NULL;
4080 regnode *tail = scan;
4084 /* var tail is used because there may be a TAIL
4085 regop in the way. Ie, the exacts will point to the
4086 thing following the TAIL, but the last branch will
4087 point at the TAIL. So we advance tail. If we
4088 have nested (?:) we may have to move through several
4092 while ( OP( tail ) == TAIL ) {
4093 /* this is the TAIL generated by (?:) */
4094 tail = regnext( tail );
4098 DEBUG_TRIE_COMPILE_r({
4099 regprop(RExC_rx, RExC_mysv, tail, NULL, pRExC_state);
4100 PerlIO_printf( Perl_debug_log, "%*s%s%s\n",
4101 (int)depth * 2 + 2, "",
4102 "Looking for TRIE'able sequences. Tail node is: ",
4103 SvPV_nolen_const( RExC_mysv )
4109 Step through the branches
4110 cur represents each branch,
4111 noper is the first thing to be matched as part
4113 noper_next is the regnext() of that node.
4115 We normally handle a case like this
4116 /FOO[xyz]|BAR[pqr]/ via a "jump trie" but we also
4117 support building with NOJUMPTRIE, which restricts
4118 the trie logic to structures like /FOO|BAR/.
4120 If noper is a trieable nodetype then the branch is
4121 a possible optimization target. If we are building
4122 under NOJUMPTRIE then we require that noper_next is
4123 the same as scan (our current position in the regex
4126 Once we have two or more consecutive such branches
4127 we can create a trie of the EXACT's contents and
4128 stitch it in place into the program.
4130 If the sequence represents all of the branches in
4131 the alternation we replace the entire thing with a
4134 Otherwise when it is a subsequence we need to
4135 stitch it in place and replace only the relevant
4136 branches. This means the first branch has to remain
4137 as it is used by the alternation logic, and its
4138 next pointer, and needs to be repointed at the item
4139 on the branch chain following the last branch we
4140 have optimized away.
4142 This could be either a BRANCH, in which case the
4143 subsequence is internal, or it could be the item
4144 following the branch sequence in which case the
4145 subsequence is at the end (which does not
4146 necessarily mean the first node is the start of the
4149 TRIE_TYPE(X) is a define which maps the optype to a
4153 ----------------+-----------
4157 EXACTFU_SS | EXACTFU
4160 EXACTFLU8 | EXACTFLU8
4164 #define TRIE_TYPE(X) ( ( NOTHING == (X) ) \
4166 : ( EXACT == (X) ) \
4168 : ( EXACTFU == (X) || EXACTFU_SS == (X) ) \
4170 : ( EXACTFA == (X) ) \
4172 : ( EXACTL == (X) ) \
4174 : ( EXACTFLU8 == (X) ) \
4178 /* dont use tail as the end marker for this traverse */
4179 for ( cur = startbranch ; cur != scan ; cur = regnext( cur ) ) {
4180 regnode * const noper = NEXTOPER( cur );
4181 U8 noper_type = OP( noper );
4182 U8 noper_trietype = TRIE_TYPE( noper_type );
4183 #if defined(DEBUGGING) || defined(NOJUMPTRIE)
4184 regnode * const noper_next = regnext( noper );
4185 U8 noper_next_type = (noper_next && noper_next != tail) ? OP(noper_next) : 0;
4186 U8 noper_next_trietype = (noper_next && noper_next != tail) ? TRIE_TYPE( noper_next_type ) :0;
4189 DEBUG_TRIE_COMPILE_r({
4190 regprop(RExC_rx, RExC_mysv, cur, NULL, pRExC_state);
4191 PerlIO_printf( Perl_debug_log, "%*s- %s (%d)",
4192 (int)depth * 2 + 2,"", SvPV_nolen_const( RExC_mysv ), REG_NODE_NUM(cur) );
4194 regprop(RExC_rx, RExC_mysv, noper, NULL, pRExC_state);
4195 PerlIO_printf( Perl_debug_log, " -> %s",
4196 SvPV_nolen_const(RExC_mysv));
4199 regprop(RExC_rx, RExC_mysv, noper_next, NULL, pRExC_state);
4200 PerlIO_printf( Perl_debug_log,"\t=> %s\t",
4201 SvPV_nolen_const(RExC_mysv));
4203 PerlIO_printf( Perl_debug_log, "(First==%d,Last==%d,Cur==%d,tt==%s,nt==%s,nnt==%s)\n",
4204 REG_NODE_NUM(first), REG_NODE_NUM(last), REG_NODE_NUM(cur),
4205 PL_reg_name[trietype], PL_reg_name[noper_trietype], PL_reg_name[noper_next_trietype]
4209 /* Is noper a trieable nodetype that can be merged
4210 * with the current trie (if there is one)? */
4214 ( noper_trietype == NOTHING)
4215 || ( trietype == NOTHING )
4216 || ( trietype == noper_trietype )
4219 && noper_next == tail
4223 /* Handle mergable triable node Either we are
4224 * the first node in a new trieable sequence,
4225 * in which case we do some bookkeeping,
4226 * otherwise we update the end pointer. */
4229 if ( noper_trietype == NOTHING ) {
4230 #if !defined(DEBUGGING) && !defined(NOJUMPTRIE)
4231 regnode * const noper_next = regnext( noper );
4232 U8 noper_next_type = (noper_next && noper_next!=tail) ? OP(noper_next) : 0;
4233 U8 noper_next_trietype = noper_next_type ? TRIE_TYPE( noper_next_type ) :0;
4236 if ( noper_next_trietype ) {
4237 trietype = noper_next_trietype;
4238 } else if (noper_next_type) {
4239 /* a NOTHING regop is 1 regop wide.
4240 * We need at least two for a trie
4241 * so we can't merge this in */
4245 trietype = noper_trietype;
4248 if ( trietype == NOTHING )
4249 trietype = noper_trietype;
4254 } /* end handle mergable triable node */
4256 /* handle unmergable node -
4257 * noper may either be a triable node which can
4258 * not be tried together with the current trie,
4259 * or a non triable node */
4261 /* If last is set and trietype is not
4262 * NOTHING then we have found at least two
4263 * triable branch sequences in a row of a
4264 * similar trietype so we can turn them
4265 * into a trie. If/when we allow NOTHING to
4266 * start a trie sequence this condition
4267 * will be required, and it isn't expensive
4268 * so we leave it in for now. */
4269 if ( trietype && trietype != NOTHING )
4270 make_trie( pRExC_state,
4271 startbranch, first, cur, tail,
4272 count, trietype, depth+1 );
4273 last = NULL; /* note: we clear/update
4274 first, trietype etc below,
4275 so we dont do it here */
4279 && noper_next == tail
4282 /* noper is triable, so we can start a new
4286 trietype = noper_trietype;
4288 /* if we already saw a first but the
4289 * current node is not triable then we have
4290 * to reset the first information. */
4295 } /* end handle unmergable node */
4296 } /* loop over branches */
4297 DEBUG_TRIE_COMPILE_r({
4298 regprop(RExC_rx, RExC_mysv, cur, NULL, pRExC_state);
4299 PerlIO_printf( Perl_debug_log,
4300 "%*s- %s (%d) <SCAN FINISHED>\n",
4302 "", SvPV_nolen_const( RExC_mysv ),REG_NODE_NUM(cur));
4305 if ( last && trietype ) {
4306 if ( trietype != NOTHING ) {
4307 /* the last branch of the sequence was part of
4308 * a trie, so we have to construct it here
4309 * outside of the loop */
4310 made= make_trie( pRExC_state, startbranch,
4311 first, scan, tail, count,
4312 trietype, depth+1 );
4313 #ifdef TRIE_STUDY_OPT
4314 if ( ((made == MADE_EXACT_TRIE &&
4315 startbranch == first)
4316 || ( first_non_open == first )) &&
4318 flags |= SCF_TRIE_RESTUDY;
4319 if ( startbranch == first
4322 RExC_seen &=~REG_TOP_LEVEL_BRANCHES_SEEN;
4327 /* at this point we know whatever we have is a
4328 * NOTHING sequence/branch AND if 'startbranch'
4329 * is 'first' then we can turn the whole thing
4332 if ( startbranch == first ) {
4334 /* the entire thing is a NOTHING sequence,
4335 * something like this: (?:|) So we can
4336 * turn it into a plain NOTHING op. */
4337 DEBUG_TRIE_COMPILE_r({
4338 regprop(RExC_rx, RExC_mysv, cur, NULL, pRExC_state);
4339 PerlIO_printf( Perl_debug_log,
4340 "%*s- %s (%d) <NOTHING BRANCH SEQUENCE>\n", (int)depth * 2 + 2,
4341 "", SvPV_nolen_const( RExC_mysv ),REG_NODE_NUM(cur));
4344 OP(startbranch)= NOTHING;
4345 NEXT_OFF(startbranch)= tail - startbranch;
4346 for ( opt= startbranch + 1; opt < tail ; opt++ )
4350 } /* end if ( last) */
4351 } /* TRIE_MAXBUF is non zero */
4356 else if ( code == BRANCHJ ) { /* single branch is optimized. */
4357 scan = NEXTOPER(NEXTOPER(scan));
4358 } else /* single branch is optimized. */
4359 scan = NEXTOPER(scan);
4361 } else if (OP(scan) == SUSPEND || OP(scan) == GOSUB || OP(scan) == GOSTART) {
4363 regnode *start = NULL;
4364 regnode *end = NULL;
4365 U32 my_recursed_depth= recursed_depth;
4368 if (OP(scan) != SUSPEND) { /* GOSUB/GOSTART */
4369 /* Do setup, note this code has side effects beyond
4370 * the rest of this block. Specifically setting
4371 * RExC_recurse[] must happen at least once during
4373 if (OP(scan) == GOSUB) {
4375 RExC_recurse[ARG2L(scan)] = scan;
4376 start = RExC_open_parens[paren-1];
4377 end = RExC_close_parens[paren-1];
4379 start = RExC_rxi->program + 1;
4382 /* NOTE we MUST always execute the above code, even
4383 * if we do nothing with a GOSUB/GOSTART */
4385 ( flags & SCF_IN_DEFINE )
4388 (is_inf_internal || is_inf || (data && data->flags & SF_IS_INF))
4390 ( (flags & (SCF_DO_STCLASS | SCF_DO_SUBSTR)) == 0 )
4393 /* no need to do anything here if we are in a define. */
4394 /* or we are after some kind of infinite construct
4395 * so we can skip recursing into this item.
4396 * Since it is infinite we will not change the maxlen
4397 * or delta, and if we miss something that might raise
4398 * the minlen it will merely pessimise a little.
4400 * Iow /(?(DEFINE)(?<foo>foo|food))a+(?&foo)/
4401 * might result in a minlen of 1 and not of 4,
4402 * but this doesn't make us mismatch, just try a bit
4403 * harder than we should.
4405 scan= regnext(scan);
4412 !PAREN_TEST(RExC_study_chunk_recursed + ((recursed_depth-1) * RExC_study_chunk_recursed_bytes), paren)
4414 /* it is quite possible that there are more efficient ways
4415 * to do this. We maintain a bitmap per level of recursion
4416 * of which patterns we have entered so we can detect if a
4417 * pattern creates a possible infinite loop. When we
4418 * recurse down a level we copy the previous levels bitmap
4419 * down. When we are at recursion level 0 we zero the top
4420 * level bitmap. It would be nice to implement a different
4421 * more efficient way of doing this. In particular the top
4422 * level bitmap may be unnecessary.
4424 if (!recursed_depth) {
4425 Zero(RExC_study_chunk_recursed, RExC_study_chunk_recursed_bytes, U8);
4427 Copy(RExC_study_chunk_recursed + ((recursed_depth-1) * RExC_study_chunk_recursed_bytes),
4428 RExC_study_chunk_recursed + (recursed_depth * RExC_study_chunk_recursed_bytes),
4429 RExC_study_chunk_recursed_bytes, U8);
4431 /* we havent recursed into this paren yet, so recurse into it */
4432 DEBUG_STUDYDATA("set:", data,depth);
4433 PAREN_SET(RExC_study_chunk_recursed + (recursed_depth * RExC_study_chunk_recursed_bytes), paren);
4434 my_recursed_depth= recursed_depth + 1;
4436 DEBUG_STUDYDATA("inf:", data,depth);
4437 /* some form of infinite recursion, assume infinite length
4439 if (flags & SCF_DO_SUBSTR) {
4440 scan_commit(pRExC_state, data, minlenp, is_inf);
4441 data->longest = &(data->longest_float);
4443 is_inf = is_inf_internal = 1;
4444 if (flags & SCF_DO_STCLASS_OR) /* Allow everything */
4445 ssc_anything(data->start_class);
4446 flags &= ~SCF_DO_STCLASS;
4448 start= NULL; /* reset start so we dont recurse later on. */
4453 end = regnext(scan);
4456 scan_frame *newframe;
4458 if (!RExC_frame_last) {
4459 Newxz(newframe, 1, scan_frame);
4460 SAVEDESTRUCTOR_X(S_unwind_scan_frames, newframe);
4461 RExC_frame_head= newframe;
4463 } else if (!RExC_frame_last->next_frame) {
4464 Newxz(newframe,1,scan_frame);
4465 RExC_frame_last->next_frame= newframe;
4466 newframe->prev_frame= RExC_frame_last;
4469 newframe= RExC_frame_last->next_frame;
4471 RExC_frame_last= newframe;
4473 newframe->next_regnode = regnext(scan);
4474 newframe->last_regnode = last;
4475 newframe->stopparen = stopparen;
4476 newframe->prev_recursed_depth = recursed_depth;
4477 newframe->this_prev_frame= frame;
4479 DEBUG_STUDYDATA("frame-new:",data,depth);
4480 DEBUG_PEEP("fnew", scan, depth);
4487 recursed_depth= my_recursed_depth;
4492 else if (OP(scan) == EXACT || OP(scan) == EXACTL) {
4493 SSize_t l = STR_LEN(scan);
4496 const U8 * const s = (U8*)STRING(scan);
4497 uc = utf8_to_uvchr_buf(s, s + l, NULL);
4498 l = utf8_length(s, s + l);
4500 uc = *((U8*)STRING(scan));
4503 if (flags & SCF_DO_SUBSTR) { /* Update longest substr. */
4504 /* The code below prefers earlier match for fixed
4505 offset, later match for variable offset. */
4506 if (data->last_end == -1) { /* Update the start info. */
4507 data->last_start_min = data->pos_min;
4508 data->last_start_max = is_inf
4509 ? SSize_t_MAX : data->pos_min + data->pos_delta;
4511 sv_catpvn(data->last_found, STRING(scan), STR_LEN(scan));
4513 SvUTF8_on(data->last_found);
4515 SV * const sv = data->last_found;
4516 MAGIC * const mg = SvUTF8(sv) && SvMAGICAL(sv) ?
4517 mg_find(sv, PERL_MAGIC_utf8) : NULL;
4518 if (mg && mg->mg_len >= 0)
4519 mg->mg_len += utf8_length((U8*)STRING(scan),
4520 (U8*)STRING(scan)+STR_LEN(scan));
4522 data->last_end = data->pos_min + l;
4523 data->pos_min += l; /* As in the first entry. */
4524 data->flags &= ~SF_BEFORE_EOL;
4527 /* ANDing the code point leaves at most it, and not in locale, and
4528 * can't match null string */
4529 if (flags & SCF_DO_STCLASS_AND) {
4530 ssc_cp_and(data->start_class, uc);
4531 ANYOF_FLAGS(data->start_class) &= ~SSC_MATCHES_EMPTY_STRING;
4532 ssc_clear_locale(data->start_class);
4534 else if (flags & SCF_DO_STCLASS_OR) {
4535 ssc_add_cp(data->start_class, uc);
4536 ssc_and(pRExC_state, data->start_class, (regnode_charclass *) and_withp);
4538 /* See commit msg 749e076fceedeb708a624933726e7989f2302f6a */
4539 ANYOF_FLAGS(data->start_class) &= ~SSC_MATCHES_EMPTY_STRING;
4541 flags &= ~SCF_DO_STCLASS;
4543 else if (PL_regkind[OP(scan)] == EXACT) {
4544 /* But OP != EXACT!, so is EXACTFish */
4545 SSize_t l = STR_LEN(scan);
4546 const U8 * s = (U8*)STRING(scan);
4548 /* Search for fixed substrings supports EXACT only. */
4549 if (flags & SCF_DO_SUBSTR) {
4551 scan_commit(pRExC_state, data, minlenp, is_inf);
4554 l = utf8_length(s, s + l);
4556 if (unfolded_multi_char) {
4557 RExC_seen |= REG_UNFOLDED_MULTI_SEEN;
4559 min += l - min_subtract;
4561 delta += min_subtract;
4562 if (flags & SCF_DO_SUBSTR) {
4563 data->pos_min += l - min_subtract;
4564 if (data->pos_min < 0) {
4567 data->pos_delta += min_subtract;
4569 data->longest = &(data->longest_float);
4573 if (flags & SCF_DO_STCLASS) {
4574 SV* EXACTF_invlist = _make_exactf_invlist(pRExC_state, scan);
4576 assert(EXACTF_invlist);
4577 if (flags & SCF_DO_STCLASS_AND) {
4578 if (OP(scan) != EXACTFL)
4579 ssc_clear_locale(data->start_class);
4580 ANYOF_FLAGS(data->start_class) &= ~SSC_MATCHES_EMPTY_STRING;
4581 ANYOF_POSIXL_ZERO(data->start_class);
4582 ssc_intersection(data->start_class, EXACTF_invlist, FALSE);
4584 else { /* SCF_DO_STCLASS_OR */
4585 ssc_union(data->start_class, EXACTF_invlist, FALSE);
4586 ssc_and(pRExC_state, data->start_class, (regnode_charclass *) and_withp);
4588 /* See commit msg 749e076fceedeb708a624933726e7989f2302f6a */
4589 ANYOF_FLAGS(data->start_class) &= ~SSC_MATCHES_EMPTY_STRING;
4591 flags &= ~SCF_DO_STCLASS;
4592 SvREFCNT_dec(EXACTF_invlist);
4595 else if (REGNODE_VARIES(OP(scan))) {
4596 SSize_t mincount, maxcount, minnext, deltanext, pos_before = 0;
4597 I32 fl = 0, f = flags;
4598 regnode * const oscan = scan;
4599 regnode_ssc this_class;
4600 regnode_ssc *oclass = NULL;
4601 I32 next_is_eval = 0;
4603 switch (PL_regkind[OP(scan)]) {
4604 case WHILEM: /* End of (?:...)* . */
4605 scan = NEXTOPER(scan);
4608 if (flags & (SCF_DO_SUBSTR | SCF_DO_STCLASS)) {
4609 next = NEXTOPER(scan);
4610 if (OP(next) == EXACT
4611 || OP(next) == EXACTL
4612 || (flags & SCF_DO_STCLASS))
4615 maxcount = REG_INFTY;
4616 next = regnext(scan);
4617 scan = NEXTOPER(scan);
4621 if (flags & SCF_DO_SUBSTR)
4626 if (flags & SCF_DO_STCLASS) {
4628 maxcount = REG_INFTY;
4629 next = regnext(scan);
4630 scan = NEXTOPER(scan);
4633 if (flags & SCF_DO_SUBSTR) {
4634 scan_commit(pRExC_state, data, minlenp, is_inf);
4635 /* Cannot extend fixed substrings */
4636 data->longest = &(data->longest_float);
4638 is_inf = is_inf_internal = 1;
4639 scan = regnext(scan);
4640 goto optimize_curly_tail;
4642 if (stopparen>0 && (OP(scan)==CURLYN || OP(scan)==CURLYM)
4643 && (scan->flags == stopparen))
4648 mincount = ARG1(scan);
4649 maxcount = ARG2(scan);
4651 next = regnext(scan);
4652 if (OP(scan) == CURLYX) {
4653 I32 lp = (data ? *(data->last_closep) : 0);
4654 scan->flags = ((lp <= (I32)U8_MAX) ? (U8)lp : U8_MAX);
4656 scan = NEXTOPER(scan) + EXTRA_STEP_2ARGS;
4657 next_is_eval = (OP(scan) == EVAL);
4659 if (flags & SCF_DO_SUBSTR) {
4661 scan_commit(pRExC_state, data, minlenp, is_inf);
4662 /* Cannot extend fixed substrings */
4663 pos_before = data->pos_min;
4667 data->flags &= ~(SF_HAS_PAR|SF_IN_PAR|SF_HAS_EVAL);
4669 data->flags |= SF_IS_INF;
4671 if (flags & SCF_DO_STCLASS) {
4672 ssc_init(pRExC_state, &this_class);
4673 oclass = data->start_class;
4674 data->start_class = &this_class;
4675 f |= SCF_DO_STCLASS_AND;
4676 f &= ~SCF_DO_STCLASS_OR;
4678 /* Exclude from super-linear cache processing any {n,m}
4679 regops for which the combination of input pos and regex
4680 pos is not enough information to determine if a match
4683 For example, in the regex /foo(bar\s*){4,8}baz/ with the
4684 regex pos at the \s*, the prospects for a match depend not
4685 only on the input position but also on how many (bar\s*)
4686 repeats into the {4,8} we are. */
4687 if ((mincount > 1) || (maxcount > 1 && maxcount != REG_INFTY))
4688 f &= ~SCF_WHILEM_VISITED_POS;
4690 /* This will finish on WHILEM, setting scan, or on NULL: */
4691 minnext = study_chunk(pRExC_state, &scan, minlenp, &deltanext,
4692 last, data, stopparen, recursed_depth, NULL,
4694 ? (f & ~SCF_DO_SUBSTR)
4698 if (flags & SCF_DO_STCLASS)
4699 data->start_class = oclass;
4700 if (mincount == 0 || minnext == 0) {
4701 if (flags & SCF_DO_STCLASS_OR) {
4702 ssc_or(pRExC_state, data->start_class, (regnode_charclass *) &this_class);
4704 else if (flags & SCF_DO_STCLASS_AND) {
4705 /* Switch to OR mode: cache the old value of
4706 * data->start_class */
4708 StructCopy(data->start_class, and_withp, regnode_ssc);
4709 flags &= ~SCF_DO_STCLASS_AND;
4710 StructCopy(&this_class, data->start_class, regnode_ssc);
4711 flags |= SCF_DO_STCLASS_OR;
4712 ANYOF_FLAGS(data->start_class)
4713 |= SSC_MATCHES_EMPTY_STRING;
4715 } else { /* Non-zero len */
4716 if (flags & SCF_DO_STCLASS_OR) {
4717 ssc_or(pRExC_state, data->start_class, (regnode_charclass *) &this_class);
4718 ssc_and(pRExC_state, data->start_class, (regnode_charclass *) and_withp);
4720 else if (flags & SCF_DO_STCLASS_AND)
4721 ssc_and(pRExC_state, data->start_class, (regnode_charclass *) &this_class);
4722 flags &= ~SCF_DO_STCLASS;
4724 if (!scan) /* It was not CURLYX, but CURLY. */
4726 if (!(flags & SCF_TRIE_DOING_RESTUDY)
4727 /* ? quantifier ok, except for (?{ ... }) */
4728 && (next_is_eval || !(mincount == 0 && maxcount == 1))
4729 && (minnext == 0) && (deltanext == 0)
4730 && data && !(data->flags & (SF_HAS_PAR|SF_IN_PAR))
4731 && maxcount <= REG_INFTY/3) /* Complement check for big
4734 /* Fatal warnings may leak the regexp without this: */
4735 SAVEFREESV(RExC_rx_sv);
4736 Perl_ck_warner(aTHX_ packWARN(WARN_REGEXP),
4737 "Quantifier unexpected on zero-length expression "
4738 "in regex m/%"UTF8f"/",
4739 UTF8fARG(UTF, RExC_end - RExC_precomp,
4741 (void)ReREFCNT_inc(RExC_rx_sv);
4744 min += minnext * mincount;
4745 is_inf_internal |= deltanext == SSize_t_MAX
4746 || (maxcount == REG_INFTY && minnext + deltanext > 0);
4747 is_inf |= is_inf_internal;
4749 delta = SSize_t_MAX;
4751 delta += (minnext + deltanext) * maxcount
4752 - minnext * mincount;
4754 /* Try powerful optimization CURLYX => CURLYN. */
4755 if ( OP(oscan) == CURLYX && data
4756 && data->flags & SF_IN_PAR
4757 && !(data->flags & SF_HAS_EVAL)
4758 && !deltanext && minnext == 1 ) {
4759 /* Try to optimize to CURLYN. */
4760 regnode *nxt = NEXTOPER(oscan) + EXTRA_STEP_2ARGS;
4761 regnode * const nxt1 = nxt;
4768 if (!REGNODE_SIMPLE(OP(nxt))
4769 && !(PL_regkind[OP(nxt)] == EXACT
4770 && STR_LEN(nxt) == 1))
4776 if (OP(nxt) != CLOSE)
4778 if (RExC_open_parens) {
4779 RExC_open_parens[ARG(nxt1)-1]=oscan; /*open->CURLYM*/
4780 RExC_close_parens[ARG(nxt1)-1]=nxt+2; /*close->while*/
4782 /* Now we know that nxt2 is the only contents: */
4783 oscan->flags = (U8)ARG(nxt);
4785 OP(nxt1) = NOTHING; /* was OPEN. */
4788 OP(nxt1 + 1) = OPTIMIZED; /* was count. */
4789 NEXT_OFF(nxt1+ 1) = 0; /* just for consistency. */
4790 NEXT_OFF(nxt2) = 0; /* just for consistency with CURLY. */
4791 OP(nxt) = OPTIMIZED; /* was CLOSE. */
4792 OP(nxt + 1) = OPTIMIZED; /* was count. */
4793 NEXT_OFF(nxt+ 1) = 0; /* just for consistency. */
4798 /* Try optimization CURLYX => CURLYM. */
4799 if ( OP(oscan) == CURLYX && data
4800 && !(data->flags & SF_HAS_PAR)
4801 && !(data->flags & SF_HAS_EVAL)
4802 && !deltanext /* atom is fixed width */
4803 && minnext != 0 /* CURLYM can't handle zero width */
4805 /* Nor characters whose fold at run-time may be
4806 * multi-character */
4807 && ! (RExC_seen & REG_UNFOLDED_MULTI_SEEN)
4809 /* XXXX How to optimize if data == 0? */
4810 /* Optimize to a simpler form. */
4811 regnode *nxt = NEXTOPER(oscan) + EXTRA_STEP_2ARGS; /* OPEN */
4815 while ( (nxt2 = regnext(nxt)) /* skip over embedded stuff*/
4816 && (OP(nxt2) != WHILEM))
4818 OP(nxt2) = SUCCEED; /* Whas WHILEM */
4819 /* Need to optimize away parenths. */
4820 if ((data->flags & SF_IN_PAR) && OP(nxt) == CLOSE) {
4821 /* Set the parenth number. */
4822 regnode *nxt1 = NEXTOPER(oscan) + EXTRA_STEP_2ARGS; /* OPEN*/
4824 oscan->flags = (U8)ARG(nxt);
4825 if (RExC_open_parens) {
4826 RExC_open_parens[ARG(nxt1)-1]=oscan; /*open->CURLYM*/
4827 RExC_close_parens[ARG(nxt1)-1]=nxt2+1; /*close->NOTHING*/
4829 OP(nxt1) = OPTIMIZED; /* was OPEN. */
4830 OP(nxt) = OPTIMIZED; /* was CLOSE. */
4833 OP(nxt1 + 1) = OPTIMIZED; /* was count. */
4834 OP(nxt + 1) = OPTIMIZED; /* was count. */
4835 NEXT_OFF(nxt1 + 1) = 0; /* just for consistency. */
4836 NEXT_OFF(nxt + 1) = 0; /* just for consistency. */
4839 while ( nxt1 && (OP(nxt1) != WHILEM)) {
4840 regnode *nnxt = regnext(nxt1);
4842 if (reg_off_by_arg[OP(nxt1)])
4843 ARG_SET(nxt1, nxt2 - nxt1);
4844 else if (nxt2 - nxt1 < U16_MAX)
4845 NEXT_OFF(nxt1) = nxt2 - nxt1;
4847 OP(nxt) = NOTHING; /* Cannot beautify */
4852 /* Optimize again: */
4853 study_chunk(pRExC_state, &nxt1, minlenp, &deltanext, nxt,
4854 NULL, stopparen, recursed_depth, NULL, 0,depth+1);
4859 else if ((OP(oscan) == CURLYX)
4860 && (flags & SCF_WHILEM_VISITED_POS)
4861 /* See the comment on a similar expression above.
4862 However, this time it's not a subexpression
4863 we care about, but the expression itself. */
4864 && (maxcount == REG_INFTY)
4865 && data && ++data->whilem_c < 16) {
4866 /* This stays as CURLYX, we can put the count/of pair. */
4867 /* Find WHILEM (as in regexec.c) */
4868 regnode *nxt = oscan + NEXT_OFF(oscan);
4870 if (OP(PREVOPER(nxt)) == NOTHING) /* LONGJMP */
4872 PREVOPER(nxt)->flags = (U8)(data->whilem_c
4873 | (RExC_whilem_seen << 4)); /* On WHILEM */
4875 if (data && fl & (SF_HAS_PAR|SF_IN_PAR))
4877 if (flags & SCF_DO_SUBSTR) {
4878 SV *last_str = NULL;
4879 STRLEN last_chrs = 0;
4880 int counted = mincount != 0;
4882 if (data->last_end > 0 && mincount != 0) { /* Ends with a
4884 SSize_t b = pos_before >= data->last_start_min
4885 ? pos_before : data->last_start_min;
4887 const char * const s = SvPV_const(data->last_found, l);
4888 SSize_t old = b - data->last_start_min;
4891 old = utf8_hop((U8*)s, old) - (U8*)s;
4893 /* Get the added string: */
4894 last_str = newSVpvn_utf8(s + old, l, UTF);
4895 last_chrs = UTF ? utf8_length((U8*)(s + old),
4896 (U8*)(s + old + l)) : l;
4897 if (deltanext == 0 && pos_before == b) {
4898 /* What was added is a constant string */
4901 SvGROW(last_str, (mincount * l) + 1);
4902 repeatcpy(SvPVX(last_str) + l,
4903 SvPVX_const(last_str), l,
4905 SvCUR_set(last_str, SvCUR(last_str) * mincount);
4906 /* Add additional parts. */
4907 SvCUR_set(data->last_found,
4908 SvCUR(data->last_found) - l);
4909 sv_catsv(data->last_found, last_str);
4911 SV * sv = data->last_found;
4913 SvUTF8(sv) && SvMAGICAL(sv) ?
4914 mg_find(sv, PERL_MAGIC_utf8) : NULL;
4915 if (mg && mg->mg_len >= 0)
4916 mg->mg_len += last_chrs * (mincount-1);
4918 last_chrs *= mincount;
4919 data->last_end += l * (mincount - 1);
4922 /* start offset must point into the last copy */
4923 data->last_start_min += minnext * (mincount - 1);
4924 data->last_start_max =
4927 : data->last_start_max +
4928 (maxcount - 1) * (minnext + data->pos_delta);
4931 /* It is counted once already... */
4932 data->pos_min += minnext * (mincount - counted);
4934 PerlIO_printf(Perl_debug_log, "counted=%"UVuf" deltanext=%"UVuf
4935 " SSize_t_MAX=%"UVuf" minnext=%"UVuf
4936 " maxcount=%"UVuf" mincount=%"UVuf"\n",
4937 (UV)counted, (UV)deltanext, (UV)SSize_t_MAX, (UV)minnext, (UV)maxcount,
4939 if (deltanext != SSize_t_MAX)
4940 PerlIO_printf(Perl_debug_log, "LHS=%"UVuf" RHS=%"UVuf"\n",
4941 (UV)(-counted * deltanext + (minnext + deltanext) * maxcount
4942 - minnext * mincount), (UV)(SSize_t_MAX - data->pos_delta));
4944 if (deltanext == SSize_t_MAX
4945 || -counted * deltanext + (minnext + deltanext) * maxcount - minnext * mincount >= SSize_t_MAX - data->pos_delta)
4946 data->pos_delta = SSize_t_MAX;
4948 data->pos_delta += - counted * deltanext +
4949 (minnext + deltanext) * maxcount - minnext * mincount;
4950 if (mincount != maxcount) {
4951 /* Cannot extend fixed substrings found inside
4953 scan_commit(pRExC_state, data, minlenp, is_inf);
4954 if (mincount && last_str) {
4955 SV * const sv = data->last_found;
4956 MAGIC * const mg = SvUTF8(sv) && SvMAGICAL(sv) ?
4957 mg_find(sv, PERL_MAGIC_utf8) : NULL;
4961 sv_setsv(sv, last_str);
4962 data->last_end = data->pos_min;
4963 data->last_start_min = data->pos_min - last_chrs;
4964 data->last_start_max = is_inf
4966 : data->pos_min + data->pos_delta - last_chrs;
4968 data->longest = &(data->longest_float);
4970 SvREFCNT_dec(last_str);
4972 if (data && (fl & SF_HAS_EVAL))
4973 data->flags |= SF_HAS_EVAL;
4974 optimize_curly_tail:
4975 if (OP(oscan) != CURLYX) {
4976 while (PL_regkind[OP(next = regnext(oscan))] == NOTHING
4978 NEXT_OFF(oscan) += NEXT_OFF(next);
4984 Perl_croak(aTHX_ "panic: unexpected varying REx opcode %d",
4989 if (flags & SCF_DO_SUBSTR) {
4990 /* Cannot expect anything... */
4991 scan_commit(pRExC_state, data, minlenp, is_inf);
4992 data->longest = &(data->longest_float);
4994 is_inf = is_inf_internal = 1;
4995 if (flags & SCF_DO_STCLASS_OR) {
4996 if (OP(scan) == CLUMP) {
4997 /* Actually is any start char, but very few code points
4998 * aren't start characters */
4999 ssc_match_all_cp(data->start_class);
5002 ssc_anything(data->start_class);
5005 flags &= ~SCF_DO_STCLASS;
5009 else if (OP(scan) == LNBREAK) {
5010 if (flags & SCF_DO_STCLASS) {
5011 if (flags & SCF_DO_STCLASS_AND) {
5012 ssc_intersection(data->start_class,
5013 PL_XPosix_ptrs[_CC_VERTSPACE], FALSE);
5014 ssc_clear_locale(data->start_class);
5015 ANYOF_FLAGS(data->start_class)
5016 &= ~SSC_MATCHES_EMPTY_STRING;
5018 else if (flags & SCF_DO_STCLASS_OR) {
5019 ssc_union(data->start_class,
5020 PL_XPosix_ptrs[_CC_VERTSPACE],
5022 ssc_and(pRExC_state, data->start_class, (regnode_charclass *) and_withp);
5024 /* See commit msg for
5025 * 749e076fceedeb708a624933726e7989f2302f6a */
5026 ANYOF_FLAGS(data->start_class)
5027 &= ~SSC_MATCHES_EMPTY_STRING;
5029 flags &= ~SCF_DO_STCLASS;
5032 if (delta != SSize_t_MAX)
5033 delta++; /* Because of the 2 char string cr-lf */
5034 if (flags & SCF_DO_SUBSTR) {
5035 /* Cannot expect anything... */
5036 scan_commit(pRExC_state, data, minlenp, is_inf);
5038 data->pos_delta += 1;
5039 data->longest = &(data->longest_float);
5042 else if (REGNODE_SIMPLE(OP(scan))) {
5044 if (flags & SCF_DO_SUBSTR) {
5045 scan_commit(pRExC_state, data, minlenp, is_inf);
5049 if (flags & SCF_DO_STCLASS) {
5051 SV* my_invlist = NULL;
5054 /* See commit msg 749e076fceedeb708a624933726e7989f2302f6a */
5055 ANYOF_FLAGS(data->start_class) &= ~SSC_MATCHES_EMPTY_STRING;
5057 /* Some of the logic below assumes that switching
5058 locale on will only add false positives. */
5063 Perl_croak(aTHX_ "panic: unexpected simple REx opcode %d",
5068 if (flags & SCF_DO_STCLASS_OR) /* Allow everything */
5069 ssc_match_all_cp(data->start_class);
5074 SV* REG_ANY_invlist = _new_invlist(2);
5075 REG_ANY_invlist = add_cp_to_invlist(REG_ANY_invlist,
5077 if (flags & SCF_DO_STCLASS_OR) {
5078 ssc_union(data->start_class,
5080 TRUE /* TRUE => invert, hence all but \n
5084 else if (flags & SCF_DO_STCLASS_AND) {
5085 ssc_intersection(data->start_class,
5087 TRUE /* TRUE => invert */
5089 ssc_clear_locale(data->start_class);
5091 SvREFCNT_dec_NN(REG_ANY_invlist);
5097 if (flags & SCF_DO_STCLASS_AND)
5098 ssc_and(pRExC_state, data->start_class,
5099 (regnode_charclass *) scan);
5101 ssc_or(pRExC_state, data->start_class,
5102 (regnode_charclass *) scan);
5110 namedclass = classnum_to_namedclass(FLAGS(scan)) + invert;
5111 if (flags & SCF_DO_STCLASS_AND) {
5112 bool was_there = cBOOL(
5113 ANYOF_POSIXL_TEST(data->start_class,
5115 ANYOF_POSIXL_ZERO(data->start_class);
5116 if (was_there) { /* Do an AND */
5117 ANYOF_POSIXL_SET(data->start_class, namedclass);
5119 /* No individual code points can now match */
5120 data->start_class->invlist
5121 = sv_2mortal(_new_invlist(0));
5124 int complement = namedclass + ((invert) ? -1 : 1);
5126 assert(flags & SCF_DO_STCLASS_OR);
5128 /* If the complement of this class was already there,
5129 * the result is that they match all code points,
5130 * (\d + \D == everything). Remove the classes from
5131 * future consideration. Locale is not relevant in
5133 if (ANYOF_POSIXL_TEST(data->start_class, complement)) {
5134 ssc_match_all_cp(data->start_class);
5135 ANYOF_POSIXL_CLEAR(data->start_class, namedclass);
5136 ANYOF_POSIXL_CLEAR(data->start_class, complement);
5138 else { /* The usual case; just add this class to the
5140 ANYOF_POSIXL_SET(data->start_class, namedclass);
5145 case NPOSIXA: /* For these, we always know the exact set of
5150 if (FLAGS(scan) == _CC_ASCII) {
5151 my_invlist = invlist_clone(PL_XPosix_ptrs[_CC_ASCII]);
5154 _invlist_intersection(PL_XPosix_ptrs[FLAGS(scan)],
5155 PL_XPosix_ptrs[_CC_ASCII],
5166 my_invlist = invlist_clone(PL_XPosix_ptrs[FLAGS(scan)]);
5168 /* NPOSIXD matches all upper Latin1 code points unless the
5169 * target string being matched is UTF-8, which is
5170 * unknowable until match time. Since we are going to
5171 * invert, we want to get rid of all of them so that the
5172 * inversion will match all */
5173 if (OP(scan) == NPOSIXD) {
5174 _invlist_subtract(my_invlist, PL_UpperLatin1,
5180 if (flags & SCF_DO_STCLASS_AND) {
5181 ssc_intersection(data->start_class, my_invlist, invert);
5182 ssc_clear_locale(data->start_class);
5185 assert(flags & SCF_DO_STCLASS_OR);
5186 ssc_union(data->start_class, my_invlist, invert);
5188 SvREFCNT_dec(my_invlist);
5190 if (flags & SCF_DO_STCLASS_OR)
5191 ssc_and(pRExC_state, data->start_class, (regnode_charclass *) and_withp);
5192 flags &= ~SCF_DO_STCLASS;
5195 else if (PL_regkind[OP(scan)] == EOL && flags & SCF_DO_SUBSTR) {
5196 data->flags |= (OP(scan) == MEOL
5199 scan_commit(pRExC_state, data, minlenp, is_inf);
5202 else if ( PL_regkind[OP(scan)] == BRANCHJ
5203 /* Lookbehind, or need to calculate parens/evals/stclass: */
5204 && (scan->flags || data || (flags & SCF_DO_STCLASS))
5205 && (OP(scan) == IFMATCH || OP(scan) == UNLESSM))
5207 if ( !PERL_ENABLE_POSITIVE_ASSERTION_STUDY
5208 || OP(scan) == UNLESSM )
5210 /* Negative Lookahead/lookbehind
5211 In this case we can't do fixed string optimisation.
5214 SSize_t deltanext, minnext, fake = 0;
5219 StructCopy(&zero_scan_data, &data_fake, scan_data_t);
5221 data_fake.whilem_c = data->whilem_c;
5222 data_fake.last_closep = data->last_closep;
5225 data_fake.last_closep = &fake;
5226 data_fake.pos_delta = delta;
5227 if ( flags & SCF_DO_STCLASS && !scan->flags
5228 && OP(scan) == IFMATCH ) { /* Lookahead */
5229 ssc_init(pRExC_state, &intrnl);
5230 data_fake.start_class = &intrnl;
5231 f |= SCF_DO_STCLASS_AND;
5233 if (flags & SCF_WHILEM_VISITED_POS)
5234 f |= SCF_WHILEM_VISITED_POS;
5235 next = regnext(scan);
5236 nscan = NEXTOPER(NEXTOPER(scan));
5237 minnext = study_chunk(pRExC_state, &nscan, minlenp, &deltanext,
5238 last, &data_fake, stopparen,
5239 recursed_depth, NULL, f, depth+1);
5242 FAIL("Variable length lookbehind not implemented");
5244 else if (minnext > (I32)U8_MAX) {
5245 FAIL2("Lookbehind longer than %"UVuf" not implemented",
5248 scan->flags = (U8)minnext;
5251 if (data_fake.flags & (SF_HAS_PAR|SF_IN_PAR))
5253 if (data_fake.flags & SF_HAS_EVAL)
5254 data->flags |= SF_HAS_EVAL;
5255 data->whilem_c = data_fake.whilem_c;
5257 if (f & SCF_DO_STCLASS_AND) {
5258 if (flags & SCF_DO_STCLASS_OR) {
5259 /* OR before, AND after: ideally we would recurse with
5260 * data_fake to get the AND applied by study of the
5261 * remainder of the pattern, and then derecurse;
5262 * *** HACK *** for now just treat as "no information".
5263 * See [perl #56690].
5265 ssc_init(pRExC_state, data->start_class);
5267 /* AND before and after: combine and continue. These
5268 * assertions are zero-length, so can match an EMPTY
5270 ssc_and(pRExC_state, data->start_class, (regnode_charclass *) &intrnl);
5271 ANYOF_FLAGS(data->start_class)
5272 |= SSC_MATCHES_EMPTY_STRING;
5276 #if PERL_ENABLE_POSITIVE_ASSERTION_STUDY
5278 /* Positive Lookahead/lookbehind
5279 In this case we can do fixed string optimisation,
5280 but we must be careful about it. Note in the case of
5281 lookbehind the positions will be offset by the minimum
5282 length of the pattern, something we won't know about
5283 until after the recurse.
5285 SSize_t deltanext, fake = 0;
5289 /* We use SAVEFREEPV so that when the full compile
5290 is finished perl will clean up the allocated
5291 minlens when it's all done. This way we don't
5292 have to worry about freeing them when we know
5293 they wont be used, which would be a pain.
5296 Newx( minnextp, 1, SSize_t );
5297 SAVEFREEPV(minnextp);
5300 StructCopy(data, &data_fake, scan_data_t);
5301 if ((flags & SCF_DO_SUBSTR) && data->last_found) {
5304 scan_commit(pRExC_state, &data_fake, minlenp, is_inf);
5305 data_fake.last_found=newSVsv(data->last_found);
5309 data_fake.last_closep = &fake;
5310 data_fake.flags = 0;
5311 data_fake.pos_delta = delta;
5313 data_fake.flags |= SF_IS_INF;
5314 if ( flags & SCF_DO_STCLASS && !scan->flags
5315 && OP(scan) == IFMATCH ) { /* Lookahead */
5316 ssc_init(pRExC_state, &intrnl);
5317 data_fake.start_class = &intrnl;
5318 f |= SCF_DO_STCLASS_AND;
5320 if (flags & SCF_WHILEM_VISITED_POS)
5321 f |= SCF_WHILEM_VISITED_POS;
5322 next = regnext(scan);
5323 nscan = NEXTOPER(NEXTOPER(scan));
5325 *minnextp = study_chunk(pRExC_state, &nscan, minnextp,
5326 &deltanext, last, &data_fake,
5327 stopparen, recursed_depth, NULL,
5331 FAIL("Variable length lookbehind not implemented");
5333 else if (*minnextp > (I32)U8_MAX) {
5334 FAIL2("Lookbehind longer than %"UVuf" not implemented",
5337 scan->flags = (U8)*minnextp;
5342 if (f & SCF_DO_STCLASS_AND) {
5343 ssc_and(pRExC_state, data->start_class, (regnode_charclass *) &intrnl);
5344 ANYOF_FLAGS(data->start_class) |= SSC_MATCHES_EMPTY_STRING;
5347 if (data_fake.flags & (SF_HAS_PAR|SF_IN_PAR))
5349 if (data_fake.flags & SF_HAS_EVAL)
5350 data->flags |= SF_HAS_EVAL;
5351 data->whilem_c = data_fake.whilem_c;
5352 if ((flags & SCF_DO_SUBSTR) && data_fake.last_found) {
5353 if (RExC_rx->minlen<*minnextp)
5354 RExC_rx->minlen=*minnextp;
5355 scan_commit(pRExC_state, &data_fake, minnextp, is_inf);
5356 SvREFCNT_dec_NN(data_fake.last_found);
5358 if ( data_fake.minlen_fixed != minlenp )
5360 data->offset_fixed= data_fake.offset_fixed;
5361 data->minlen_fixed= data_fake.minlen_fixed;
5362 data->lookbehind_fixed+= scan->flags;
5364 if ( data_fake.minlen_float != minlenp )
5366 data->minlen_float= data_fake.minlen_float;
5367 data->offset_float_min=data_fake.offset_float_min;
5368 data->offset_float_max=data_fake.offset_float_max;
5369 data->lookbehind_float+= scan->flags;
5376 else if (OP(scan) == OPEN) {
5377 if (stopparen != (I32)ARG(scan))
5380 else if (OP(scan) == CLOSE) {
5381 if (stopparen == (I32)ARG(scan)) {
5384 if ((I32)ARG(scan) == is_par) {
5385 next = regnext(scan);
5387 if ( next && (OP(next) != WHILEM) && next < last)
5388 is_par = 0; /* Disable optimization */
5391 *(data->last_closep) = ARG(scan);
5393 else if (OP(scan) == EVAL) {
5395 data->flags |= SF_HAS_EVAL;
5397 else if ( PL_regkind[OP(scan)] == ENDLIKE ) {
5398 if (flags & SCF_DO_SUBSTR) {
5399 scan_commit(pRExC_state, data, minlenp, is_inf);
5400 flags &= ~SCF_DO_SUBSTR;
5402 if (data && OP(scan)==ACCEPT) {
5403 data->flags |= SCF_SEEN_ACCEPT;
5408 else if (OP(scan) == LOGICAL && scan->flags == 2) /* Embedded follows */
5410 if (flags & SCF_DO_SUBSTR) {
5411 scan_commit(pRExC_state, data, minlenp, is_inf);
5412 data->longest = &(data->longest_float);
5414 is_inf = is_inf_internal = 1;
5415 if (flags & SCF_DO_STCLASS_OR) /* Allow everything */
5416 ssc_anything(data->start_class);
5417 flags &= ~SCF_DO_STCLASS;
5419 else if (OP(scan) == GPOS) {
5420 if (!(RExC_rx->intflags & PREGf_GPOS_FLOAT) &&
5421 !(delta || is_inf || (data && data->pos_delta)))
5423 if (!(RExC_rx->intflags & PREGf_ANCH) && (flags & SCF_DO_SUBSTR))
5424 RExC_rx->intflags |= PREGf_ANCH_GPOS;
5425 if (RExC_rx->gofs < (STRLEN)min)
5426 RExC_rx->gofs = min;
5428 RExC_rx->intflags |= PREGf_GPOS_FLOAT;
5432 #ifdef TRIE_STUDY_OPT
5433 #ifdef FULL_TRIE_STUDY
5434 else if (PL_regkind[OP(scan)] == TRIE) {
5435 /* NOTE - There is similar code to this block above for handling
5436 BRANCH nodes on the initial study. If you change stuff here
5438 regnode *trie_node= scan;
5439 regnode *tail= regnext(scan);
5440 reg_trie_data *trie = (reg_trie_data*)RExC_rxi->data->data[ ARG(scan) ];
5441 SSize_t max1 = 0, min1 = SSize_t_MAX;
5444 if (flags & SCF_DO_SUBSTR) { /* XXXX Add !SUSPEND? */
5445 /* Cannot merge strings after this. */
5446 scan_commit(pRExC_state, data, minlenp, is_inf);
5448 if (flags & SCF_DO_STCLASS)
5449 ssc_init_zero(pRExC_state, &accum);
5455 const regnode *nextbranch= NULL;
5458 for ( word=1 ; word <= trie->wordcount ; word++)
5460 SSize_t deltanext=0, minnext=0, f = 0, fake;
5461 regnode_ssc this_class;
5463 StructCopy(&zero_scan_data, &data_fake, scan_data_t);
5465 data_fake.whilem_c = data->whilem_c;
5466 data_fake.last_closep = data->last_closep;
5469 data_fake.last_closep = &fake;
5470 data_fake.pos_delta = delta;
5471 if (flags & SCF_DO_STCLASS) {
5472 ssc_init(pRExC_state, &this_class);
5473 data_fake.start_class = &this_class;
5474 f = SCF_DO_STCLASS_AND;
5476 if (flags & SCF_WHILEM_VISITED_POS)
5477 f |= SCF_WHILEM_VISITED_POS;
5479 if (trie->jump[word]) {
5481 nextbranch = trie_node + trie->jump[0];
5482 scan= trie_node + trie->jump[word];
5483 /* We go from the jump point to the branch that follows
5484 it. Note this means we need the vestigal unused
5485 branches even though they arent otherwise used. */
5486 minnext = study_chunk(pRExC_state, &scan, minlenp,
5487 &deltanext, (regnode *)nextbranch, &data_fake,
5488 stopparen, recursed_depth, NULL, f,depth+1);
5490 if (nextbranch && PL_regkind[OP(nextbranch)]==BRANCH)
5491 nextbranch= regnext((regnode*)nextbranch);
5493 if (min1 > (SSize_t)(minnext + trie->minlen))
5494 min1 = minnext + trie->minlen;
5495 if (deltanext == SSize_t_MAX) {
5496 is_inf = is_inf_internal = 1;
5498 } else if (max1 < (SSize_t)(minnext + deltanext + trie->maxlen))
5499 max1 = minnext + deltanext + trie->maxlen;
5501 if (data_fake.flags & (SF_HAS_PAR|SF_IN_PAR))
5503 if (data_fake.flags & SCF_SEEN_ACCEPT) {
5504 if ( stopmin > min + min1)
5505 stopmin = min + min1;
5506 flags &= ~SCF_DO_SUBSTR;
5508 data->flags |= SCF_SEEN_ACCEPT;
5511 if (data_fake.flags & SF_HAS_EVAL)
5512 data->flags |= SF_HAS_EVAL;
5513 data->whilem_c = data_fake.whilem_c;
5515 if (flags & SCF_DO_STCLASS)
5516 ssc_or(pRExC_state, &accum, (regnode_charclass *) &this_class);
5519 if (flags & SCF_DO_SUBSTR) {
5520 data->pos_min += min1;
5521 data->pos_delta += max1 - min1;
5522 if (max1 != min1 || is_inf)
5523 data->longest = &(data->longest_float);
5526 if (delta != SSize_t_MAX)
5527 delta += max1 - min1;
5528 if (flags & SCF_DO_STCLASS_OR) {
5529 ssc_or(pRExC_state, data->start_class, (regnode_charclass *) &accum);
5531 ssc_and(pRExC_state, data->start_class, (regnode_charclass *) and_withp);
5532 flags &= ~SCF_DO_STCLASS;
5535 else if (flags & SCF_DO_STCLASS_AND) {
5537 ssc_and(pRExC_state, data->start_class, (regnode_charclass *) &accum);
5538 flags &= ~SCF_DO_STCLASS;
5541 /* Switch to OR mode: cache the old value of
5542 * data->start_class */
5544 StructCopy(data->start_class, and_withp, regnode_ssc);
5545 flags &= ~SCF_DO_STCLASS_AND;
5546 StructCopy(&accum, data->start_class, regnode_ssc);
5547 flags |= SCF_DO_STCLASS_OR;
5554 else if (PL_regkind[OP(scan)] == TRIE) {
5555 reg_trie_data *trie = (reg_trie_data*)RExC_rxi->data->data[ ARG(scan) ];
5558 min += trie->minlen;
5559 delta += (trie->maxlen - trie->minlen);
5560 flags &= ~SCF_DO_STCLASS; /* xxx */
5561 if (flags & SCF_DO_SUBSTR) {
5562 /* Cannot expect anything... */
5563 scan_commit(pRExC_state, data, minlenp, is_inf);
5564 data->pos_min += trie->minlen;
5565 data->pos_delta += (trie->maxlen - trie->minlen);
5566 if (trie->maxlen != trie->minlen)
5567 data->longest = &(data->longest_float);
5569 if (trie->jump) /* no more substrings -- for now /grr*/
5570 flags &= ~SCF_DO_SUBSTR;
5572 #endif /* old or new */
5573 #endif /* TRIE_STUDY_OPT */
5575 /* Else: zero-length, ignore. */
5576 scan = regnext(scan);
5578 /* If we are exiting a recursion we can unset its recursed bit
5579 * and allow ourselves to enter it again - no danger of an
5580 * infinite loop there.
5581 if (stopparen > -1 && recursed) {
5582 DEBUG_STUDYDATA("unset:", data,depth);
5583 PAREN_UNSET( recursed, stopparen);
5589 DEBUG_STUDYDATA("frame-end:",data,depth);
5590 DEBUG_PEEP("fend", scan, depth);
5592 /* restore previous context */
5593 last = frame->last_regnode;
5594 scan = frame->next_regnode;
5595 stopparen = frame->stopparen;
5596 recursed_depth = frame->prev_recursed_depth;
5598 RExC_frame_last = frame->prev_frame;
5599 frame = frame->this_prev_frame;
5600 goto fake_study_recurse;
5605 DEBUG_STUDYDATA("pre-fin:",data,depth);
5608 *deltap = is_inf_internal ? SSize_t_MAX : delta;
5610 if (flags & SCF_DO_SUBSTR && is_inf)
5611 data->pos_delta = SSize_t_MAX - data->pos_min;
5612 if (is_par > (I32)U8_MAX)
5614 if (is_par && pars==1 && data) {
5615 data->flags |= SF_IN_PAR;
5616 data->flags &= ~SF_HAS_PAR;
5618 else if (pars && data) {
5619 data->flags |= SF_HAS_PAR;
5620 data->flags &= ~SF_IN_PAR;
5622 if (flags & SCF_DO_STCLASS_OR)
5623 ssc_and(pRExC_state, data->start_class, (regnode_charclass *) and_withp);
5624 if (flags & SCF_TRIE_RESTUDY)
5625 data->flags |= SCF_TRIE_RESTUDY;
5627 DEBUG_STUDYDATA("post-fin:",data,depth);
5630 SSize_t final_minlen= min < stopmin ? min : stopmin;
5632 if (!(RExC_seen & REG_UNBOUNDED_QUANTIFIER_SEEN)) {
5633 if (final_minlen > SSize_t_MAX - delta)
5634 RExC_maxlen = SSize_t_MAX;
5635 else if (RExC_maxlen < final_minlen + delta)
5636 RExC_maxlen = final_minlen + delta;
5638 return final_minlen;
5640 NOT_REACHED; /* NOTREACHED */
5644 S_add_data(RExC_state_t* const pRExC_state, const char* const s, const U32 n)
5646 U32 count = RExC_rxi->data ? RExC_rxi->data->count : 0;
5648 PERL_ARGS_ASSERT_ADD_DATA;
5650 Renewc(RExC_rxi->data,
5651 sizeof(*RExC_rxi->data) + sizeof(void*) * (count + n - 1),
5652 char, struct reg_data);
5654 Renew(RExC_rxi->data->what, count + n, U8);
5656 Newx(RExC_rxi->data->what, n, U8);
5657 RExC_rxi->data->count = count + n;
5658 Copy(s, RExC_rxi->data->what + count, n, U8);
5662 /*XXX: todo make this not included in a non debugging perl, but appears to be
5663 * used anyway there, in 'use re' */
5664 #ifndef PERL_IN_XSUB_RE
5666 Perl_reginitcolors(pTHX)
5668 const char * const s = PerlEnv_getenv("PERL_RE_COLORS");
5670 char *t = savepv(s);
5674 t = strchr(t, '\t');
5680 PL_colors[i] = t = (char *)"";
5685 PL_colors[i++] = (char *)"";
5692 #ifdef TRIE_STUDY_OPT
5693 #define CHECK_RESTUDY_GOTO_butfirst(dOsomething) \
5696 (data.flags & SCF_TRIE_RESTUDY) \
5704 #define CHECK_RESTUDY_GOTO_butfirst
5708 * pregcomp - compile a regular expression into internal code
5710 * Decides which engine's compiler to call based on the hint currently in
5714 #ifndef PERL_IN_XSUB_RE
5716 /* return the currently in-scope regex engine (or the default if none) */
5718 regexp_engine const *
5719 Perl_current_re_engine(pTHX)
5721 if (IN_PERL_COMPILETIME) {
5722 HV * const table = GvHV(PL_hintgv);
5725 if (!table || !(PL_hints & HINT_LOCALIZE_HH))
5726 return &PL_core_reg_engine;
5727 ptr = hv_fetchs(table, "regcomp", FALSE);
5728 if ( !(ptr && SvIOK(*ptr) && SvIV(*ptr)))
5729 return &PL_core_reg_engine;
5730 return INT2PTR(regexp_engine*,SvIV(*ptr));
5734 if (!PL_curcop->cop_hints_hash)
5735 return &PL_core_reg_engine;
5736 ptr = cop_hints_fetch_pvs(PL_curcop, "regcomp", 0);
5737 if ( !(ptr && SvIOK(ptr) && SvIV(ptr)))
5738 return &PL_core_reg_engine;
5739 return INT2PTR(regexp_engine*,SvIV(ptr));
5745 Perl_pregcomp(pTHX_ SV * const pattern, const U32 flags)
5747 regexp_engine const *eng = current_re_engine();
5748 GET_RE_DEBUG_FLAGS_DECL;
5750 PERL_ARGS_ASSERT_PREGCOMP;
5752 /* Dispatch a request to compile a regexp to correct regexp engine. */
5754 PerlIO_printf(Perl_debug_log, "Using engine %"UVxf"\n",
5757 return CALLREGCOMP_ENG(eng, pattern, flags);
5761 /* public(ish) entry point for the perl core's own regex compiling code.
5762 * It's actually a wrapper for Perl_re_op_compile that only takes an SV
5763 * pattern rather than a list of OPs, and uses the internal engine rather
5764 * than the current one */
5767 Perl_re_compile(pTHX_ SV * const pattern, U32 rx_flags)
5769 SV *pat = pattern; /* defeat constness! */
5770 PERL_ARGS_ASSERT_RE_COMPILE;
5771 return Perl_re_op_compile(aTHX_ &pat, 1, NULL,
5772 #ifdef PERL_IN_XSUB_RE
5775 &PL_core_reg_engine,
5777 NULL, NULL, rx_flags, 0);
5781 /* upgrade pattern pat_p of length plen_p to UTF8, and if there are code
5782 * blocks, recalculate the indices. Update pat_p and plen_p in-place to
5783 * point to the realloced string and length.
5785 * This is essentially a copy of Perl_bytes_to_utf8() with the code index
5789 S_pat_upgrade_to_utf8(pTHX_ RExC_state_t * const pRExC_state,
5790 char **pat_p, STRLEN *plen_p, int num_code_blocks)
5792 U8 *const src = (U8*)*pat_p;
5797 GET_RE_DEBUG_FLAGS_DECL;
5799 DEBUG_PARSE_r(PerlIO_printf(Perl_debug_log,
5800 "UTF8 mismatch! Converting to utf8 for resizing and compile\n"));
5802 Newx(dst, *plen_p * 2 + 1, U8);
5805 while (s < *plen_p) {
5806 append_utf8_from_native_byte(src[s], &d);
5807 if (n < num_code_blocks) {
5808 if (!do_end && pRExC_state->code_blocks[n].start == s) {
5809 pRExC_state->code_blocks[n].start = d - dst - 1;
5810 assert(*(d - 1) == '(');
5813 else if (do_end && pRExC_state->code_blocks[n].end == s) {
5814 pRExC_state->code_blocks[n].end = d - dst - 1;
5815 assert(*(d - 1) == ')');
5824 *pat_p = (char*) dst;
5826 RExC_orig_utf8 = RExC_utf8 = 1;
5831 /* S_concat_pat(): concatenate a list of args to the pattern string pat,
5832 * while recording any code block indices, and handling overloading,
5833 * nested qr// objects etc. If pat is null, it will allocate a new
5834 * string, or just return the first arg, if there's only one.
5836 * Returns the malloced/updated pat.
5837 * patternp and pat_count is the array of SVs to be concatted;
5838 * oplist is the optional list of ops that generated the SVs;
5839 * recompile_p is a pointer to a boolean that will be set if
5840 * the regex will need to be recompiled.
5841 * delim, if non-null is an SV that will be inserted between each element
5845 S_concat_pat(pTHX_ RExC_state_t * const pRExC_state,
5846 SV *pat, SV ** const patternp, int pat_count,
5847 OP *oplist, bool *recompile_p, SV *delim)
5851 bool use_delim = FALSE;
5852 bool alloced = FALSE;
5854 /* if we know we have at least two args, create an empty string,
5855 * then concatenate args to that. For no args, return an empty string */
5856 if (!pat && pat_count != 1) {
5862 for (svp = patternp; svp < patternp + pat_count; svp++) {
5865 STRLEN orig_patlen = 0;
5867 SV *msv = use_delim ? delim : *svp;
5868 if (!msv) msv = &PL_sv_undef;
5870 /* if we've got a delimiter, we go round the loop twice for each
5871 * svp slot (except the last), using the delimiter the second
5880 if (SvTYPE(msv) == SVt_PVAV) {
5881 /* we've encountered an interpolated array within
5882 * the pattern, e.g. /...@a..../. Expand the list of elements,
5883 * then recursively append elements.
5884 * The code in this block is based on S_pushav() */
5886 AV *const av = (AV*)msv;
5887 const SSize_t maxarg = AvFILL(av) + 1;
5891 assert(oplist->op_type == OP_PADAV
5892 || oplist->op_type == OP_RV2AV);
5893 oplist = OpSIBLING(oplist);
5896 if (SvRMAGICAL(av)) {
5899 Newx(array, maxarg, SV*);
5901 for (i=0; i < maxarg; i++) {
5902 SV ** const svp = av_fetch(av, i, FALSE);
5903 array[i] = svp ? *svp : &PL_sv_undef;
5907 array = AvARRAY(av);
5909 pat = S_concat_pat(aTHX_ pRExC_state, pat,
5910 array, maxarg, NULL, recompile_p,
5912 GvSV((gv_fetchpvs("\"", GV_ADDMULTI, SVt_PV))));
5918 /* we make the assumption here that each op in the list of
5919 * op_siblings maps to one SV pushed onto the stack,
5920 * except for code blocks, with have both an OP_NULL and
5922 * This allows us to match up the list of SVs against the
5923 * list of OPs to find the next code block.
5925 * Note that PUSHMARK PADSV PADSV ..
5927 * PADRANGE PADSV PADSV ..
5928 * so the alignment still works. */
5931 if (oplist->op_type == OP_NULL
5932 && (oplist->op_flags & OPf_SPECIAL))
5934 assert(n < pRExC_state->num_code_blocks);
5935 pRExC_state->code_blocks[n].start = pat ? SvCUR(pat) : 0;
5936 pRExC_state->code_blocks[n].block = oplist;
5937 pRExC_state->code_blocks[n].src_regex = NULL;
5940 oplist = OpSIBLING(oplist); /* skip CONST */
5943 oplist = OpSIBLING(oplist);;
5946 /* apply magic and QR overloading to arg */
5949 if (SvROK(msv) && SvAMAGIC(msv)) {
5950 SV *sv = AMG_CALLunary(msv, regexp_amg);
5954 if (SvTYPE(sv) != SVt_REGEXP)
5955 Perl_croak(aTHX_ "Overloaded qr did not return a REGEXP");
5960 /* try concatenation overload ... */
5961 if (pat && (SvAMAGIC(pat) || SvAMAGIC(msv)) &&
5962 (sv = amagic_call(pat, msv, concat_amg, AMGf_assign)))
5965 /* overloading involved: all bets are off over literal
5966 * code. Pretend we haven't seen it */
5967 pRExC_state->num_code_blocks -= n;
5971 /* ... or failing that, try "" overload */
5972 while (SvAMAGIC(msv)
5973 && (sv = AMG_CALLunary(msv, string_amg))
5977 && SvRV(msv) == SvRV(sv))
5982 if (SvROK(msv) && SvTYPE(SvRV(msv)) == SVt_REGEXP)
5986 /* this is a partially unrolled
5987 * sv_catsv_nomg(pat, msv);
5988 * that allows us to adjust code block indices if
5991 char *dst = SvPV_force_nomg(pat, dlen);
5993 if (SvUTF8(msv) && !SvUTF8(pat)) {
5994 S_pat_upgrade_to_utf8(aTHX_ pRExC_state, &dst, &dlen, n);
5995 sv_setpvn(pat, dst, dlen);
5998 sv_catsv_nomg(pat, msv);
6005 pRExC_state->code_blocks[n-1].end = SvCUR(pat)-1;
6008 /* extract any code blocks within any embedded qr//'s */
6009 if (rx && SvTYPE(rx) == SVt_REGEXP
6010 && RX_ENGINE((REGEXP*)rx)->op_comp)
6013 RXi_GET_DECL(ReANY((REGEXP *)rx), ri);
6014 if (ri->num_code_blocks) {
6016 /* the presence of an embedded qr// with code means
6017 * we should always recompile: the text of the
6018 * qr// may not have changed, but it may be a
6019 * different closure than last time */
6021 Renew(pRExC_state->code_blocks,
6022 pRExC_state->num_code_blocks + ri->num_code_blocks,
6023 struct reg_code_block);
6024 pRExC_state->num_code_blocks += ri->num_code_blocks;
6026 for (i=0; i < ri->num_code_blocks; i++) {
6027 struct reg_code_block *src, *dst;
6028 STRLEN offset = orig_patlen
6029 + ReANY((REGEXP *)rx)->pre_prefix;
6030 assert(n < pRExC_state->num_code_blocks);
6031 src = &ri->code_blocks[i];
6032 dst = &pRExC_state->code_blocks[n];
6033 dst->start = src->start + offset;
6034 dst->end = src->end + offset;
6035 dst->block = src->block;
6036 dst->src_regex = (REGEXP*) SvREFCNT_inc( (SV*)
6045 /* avoid calling magic multiple times on a single element e.g. =~ $qr */
6054 /* see if there are any run-time code blocks in the pattern.
6055 * False positives are allowed */
6058 S_has_runtime_code(pTHX_ RExC_state_t * const pRExC_state,
6059 char *pat, STRLEN plen)
6064 PERL_UNUSED_CONTEXT;
6066 for (s = 0; s < plen; s++) {
6067 if (n < pRExC_state->num_code_blocks
6068 && s == pRExC_state->code_blocks[n].start)
6070 s = pRExC_state->code_blocks[n].end;
6074 /* TODO ideally should handle [..], (#..), /#.../x to reduce false
6076 if (pat[s] == '(' && s+2 <= plen && pat[s+1] == '?' &&
6078 || (s + 2 <= plen && pat[s+2] == '?' && pat[s+3] == '{'))
6085 /* Handle run-time code blocks. We will already have compiled any direct
6086 * or indirect literal code blocks. Now, take the pattern 'pat' and make a
6087 * copy of it, but with any literal code blocks blanked out and
6088 * appropriate chars escaped; then feed it into
6090 * eval "qr'modified_pattern'"
6094 * a\bc(?{"this was literal"})def'ghi\\jkl(?{"this is runtime"})mno
6098 * qr'a\\bc_______________________def\'ghi\\\\jkl(?{"this is runtime"})mno'
6100 * After eval_sv()-ing that, grab any new code blocks from the returned qr
6101 * and merge them with any code blocks of the original regexp.
6103 * If the pat is non-UTF8, while the evalled qr is UTF8, don't merge;
6104 * instead, just save the qr and return FALSE; this tells our caller that
6105 * the original pattern needs upgrading to utf8.
6109 S_compile_runtime_code(pTHX_ RExC_state_t * const pRExC_state,
6110 char *pat, STRLEN plen)
6114 GET_RE_DEBUG_FLAGS_DECL;
6116 if (pRExC_state->runtime_code_qr) {
6117 /* this is the second time we've been called; this should
6118 * only happen if the main pattern got upgraded to utf8
6119 * during compilation; re-use the qr we compiled first time
6120 * round (which should be utf8 too)
6122 qr = pRExC_state->runtime_code_qr;
6123 pRExC_state->runtime_code_qr = NULL;
6124 assert(RExC_utf8 && SvUTF8(qr));
6130 int newlen = plen + 6; /* allow for "qr''x\0" extra chars */
6134 /* determine how many extra chars we need for ' and \ escaping */
6135 for (s = 0; s < plen; s++) {
6136 if (pat[s] == '\'' || pat[s] == '\\')
6140 Newx(newpat, newlen, char);
6142 *p++ = 'q'; *p++ = 'r'; *p++ = '\'';
6144 for (s = 0; s < plen; s++) {
6145 if (n < pRExC_state->num_code_blocks
6146 && s == pRExC_state->code_blocks[n].start)
6148 /* blank out literal code block */
6149 assert(pat[s] == '(');
6150 while (s <= pRExC_state->code_blocks[n].end) {
6158 if (pat[s] == '\'' || pat[s] == '\\')
6163 if (pRExC_state->pm_flags & RXf_PMf_EXTENDED)
6167 PerlIO_printf(Perl_debug_log,
6168 "%sre-parsing pattern for runtime code:%s %s\n",
6169 PL_colors[4],PL_colors[5],newpat);
6172 sv = newSVpvn_flags(newpat, p-newpat-1, RExC_utf8 ? SVf_UTF8 : 0);
6177 PUSHSTACKi(PERLSI_REQUIRE);
6178 /* G_RE_REPARSING causes the toker to collapse \\ into \ when
6179 * parsing qr''; normally only q'' does this. It also alters
6181 eval_sv(sv, G_SCALAR|G_RE_REPARSING);
6182 SvREFCNT_dec_NN(sv);
6187 SV * const errsv = ERRSV;
6188 if (SvTRUE_NN(errsv))
6190 Safefree(pRExC_state->code_blocks);
6191 /* use croak_sv ? */
6192 Perl_croak_nocontext("%"SVf, SVfARG(errsv));
6195 assert(SvROK(qr_ref));
6197 assert(SvTYPE(qr) == SVt_REGEXP && RX_ENGINE((REGEXP*)qr)->op_comp);
6198 /* the leaving below frees the tmp qr_ref.
6199 * Give qr a life of its own */
6207 if (!RExC_utf8 && SvUTF8(qr)) {
6208 /* first time through; the pattern got upgraded; save the
6209 * qr for the next time through */
6210 assert(!pRExC_state->runtime_code_qr);
6211 pRExC_state->runtime_code_qr = qr;
6216 /* extract any code blocks within the returned qr// */
6219 /* merge the main (r1) and run-time (r2) code blocks into one */
6221 RXi_GET_DECL(ReANY((REGEXP *)qr), r2);
6222 struct reg_code_block *new_block, *dst;
6223 RExC_state_t * const r1 = pRExC_state; /* convenient alias */
6226 if (!r2->num_code_blocks) /* we guessed wrong */
6228 SvREFCNT_dec_NN(qr);
6233 r1->num_code_blocks + r2->num_code_blocks,
6234 struct reg_code_block);
6237 while ( i1 < r1->num_code_blocks
6238 || i2 < r2->num_code_blocks)
6240 struct reg_code_block *src;
6243 if (i1 == r1->num_code_blocks) {
6244 src = &r2->code_blocks[i2++];
6247 else if (i2 == r2->num_code_blocks)
6248 src = &r1->code_blocks[i1++];
6249 else if ( r1->code_blocks[i1].start
6250 < r2->code_blocks[i2].start)
6252 src = &r1->code_blocks[i1++];
6253 assert(src->end < r2->code_blocks[i2].start);
6256 assert( r1->code_blocks[i1].start
6257 > r2->code_blocks[i2].start);
6258 src = &r2->code_blocks[i2++];
6260 assert(src->end < r1->code_blocks[i1].start);
6263 assert(pat[src->start] == '(');
6264 assert(pat[src->end] == ')');
6265 dst->start = src->start;
6266 dst->end = src->end;
6267 dst->block = src->block;
6268 dst->src_regex = is_qr ? (REGEXP*) SvREFCNT_inc( (SV*) qr)
6272 r1->num_code_blocks += r2->num_code_blocks;
6273 Safefree(r1->code_blocks);
6274 r1->code_blocks = new_block;
6277 SvREFCNT_dec_NN(qr);
6283 S_setup_longest(pTHX_ RExC_state_t *pRExC_state, SV* sv_longest,
6284 SV** rx_utf8, SV** rx_substr, SSize_t* rx_end_shift,
6285 SSize_t lookbehind, SSize_t offset, SSize_t *minlen,
6286 STRLEN longest_length, bool eol, bool meol)
6288 /* This is the common code for setting up the floating and fixed length
6289 * string data extracted from Perl_re_op_compile() below. Returns a boolean
6290 * as to whether succeeded or not */
6295 if (! (longest_length
6296 || (eol /* Can't have SEOL and MULTI */
6297 && (! meol || (RExC_flags & RXf_PMf_MULTILINE)))
6299 /* See comments for join_exact for why REG_UNFOLDED_MULTI_SEEN */
6300 || (RExC_seen & REG_UNFOLDED_MULTI_SEEN))
6305 /* copy the information about the longest from the reg_scan_data
6306 over to the program. */
6307 if (SvUTF8(sv_longest)) {
6308 *rx_utf8 = sv_longest;
6311 *rx_substr = sv_longest;
6314 /* end_shift is how many chars that must be matched that
6315 follow this item. We calculate it ahead of time as once the
6316 lookbehind offset is added in we lose the ability to correctly
6318 ml = minlen ? *(minlen) : (SSize_t)longest_length;
6319 *rx_end_shift = ml - offset
6320 - longest_length + (SvTAIL(sv_longest) != 0)
6323 t = (eol/* Can't have SEOL and MULTI */
6324 && (! meol || (RExC_flags & RXf_PMf_MULTILINE)));
6325 fbm_compile(sv_longest, t ? FBMcf_TAIL : 0);
6331 * Perl_re_op_compile - the perl internal RE engine's function to compile a
6332 * regular expression into internal code.
6333 * The pattern may be passed either as:
6334 * a list of SVs (patternp plus pat_count)
6335 * a list of OPs (expr)
6336 * If both are passed, the SV list is used, but the OP list indicates
6337 * which SVs are actually pre-compiled code blocks
6339 * The SVs in the list have magic and qr overloading applied to them (and
6340 * the list may be modified in-place with replacement SVs in the latter
6343 * If the pattern hasn't changed from old_re, then old_re will be
6346 * eng is the current engine. If that engine has an op_comp method, then
6347 * handle directly (i.e. we assume that op_comp was us); otherwise, just
6348 * do the initial concatenation of arguments and pass on to the external
6351 * If is_bare_re is not null, set it to a boolean indicating whether the
6352 * arg list reduced (after overloading) to a single bare regex which has
6353 * been returned (i.e. /$qr/).
6355 * orig_rx_flags contains RXf_* flags. See perlreapi.pod for more details.
6357 * pm_flags contains the PMf_* flags, typically based on those from the
6358 * pm_flags field of the related PMOP. Currently we're only interested in
6359 * PMf_HAS_CV, PMf_IS_QR, PMf_USE_RE_EVAL.
6361 * We can't allocate space until we know how big the compiled form will be,
6362 * but we can't compile it (and thus know how big it is) until we've got a
6363 * place to put the code. So we cheat: we compile it twice, once with code
6364 * generation turned off and size counting turned on, and once "for real".
6365 * This also means that we don't allocate space until we are sure that the
6366 * thing really will compile successfully, and we never have to move the
6367 * code and thus invalidate pointers into it. (Note that it has to be in
6368 * one piece because free() must be able to free it all.) [NB: not true in perl]
6370 * Beware that the optimization-preparation code in here knows about some
6371 * of the structure of the compiled regexp. [I'll say.]
6375 Perl_re_op_compile(pTHX_ SV ** const patternp, int pat_count,
6376 OP *expr, const regexp_engine* eng, REGEXP *old_re,
6377 bool *is_bare_re, U32 orig_rx_flags, U32 pm_flags)
6381 regexp_internal *ri;
6389 SV *code_blocksv = NULL;
6390 SV** new_patternp = patternp;
6392 /* these are all flags - maybe they should be turned
6393 * into a single int with different bit masks */
6394 I32 sawlookahead = 0;
6399 regex_charset initial_charset = get_regex_charset(orig_rx_flags);
6401 bool runtime_code = 0;
6403 RExC_state_t RExC_state;
6404 RExC_state_t * const pRExC_state = &RExC_state;
6405 #ifdef TRIE_STUDY_OPT
6407 RExC_state_t copyRExC_state;
6409 GET_RE_DEBUG_FLAGS_DECL;
6411 PERL_ARGS_ASSERT_RE_OP_COMPILE;
6413 DEBUG_r(if (!PL_colorset) reginitcolors());
6415 /* Initialize these here instead of as-needed, as is quick and avoids
6416 * having to test them each time otherwise */
6417 if (! PL_AboveLatin1) {
6418 PL_AboveLatin1 = _new_invlist_C_array(AboveLatin1_invlist);
6419 PL_Latin1 = _new_invlist_C_array(Latin1_invlist);
6420 PL_UpperLatin1 = _new_invlist_C_array(UpperLatin1_invlist);
6421 PL_utf8_foldable = _new_invlist_C_array(_Perl_Any_Folds_invlist);
6422 PL_HasMultiCharFold =
6423 _new_invlist_C_array(_Perl_Folds_To_Multi_Char_invlist);
6425 /* This is calculated here, because the Perl program that generates the
6426 * static global ones doesn't currently have access to
6427 * NUM_ANYOF_CODE_POINTS */
6428 PL_InBitmap = _new_invlist(2);
6429 PL_InBitmap = _add_range_to_invlist(PL_InBitmap, 0,
6430 NUM_ANYOF_CODE_POINTS - 1);
6433 pRExC_state->code_blocks = NULL;
6434 pRExC_state->num_code_blocks = 0;
6437 *is_bare_re = FALSE;
6439 if (expr && (expr->op_type == OP_LIST ||
6440 (expr->op_type == OP_NULL && expr->op_targ == OP_LIST))) {
6441 /* allocate code_blocks if needed */
6445 for (o = cLISTOPx(expr)->op_first; o; o = OpSIBLING(o))
6446 if (o->op_type == OP_NULL && (o->op_flags & OPf_SPECIAL))
6447 ncode++; /* count of DO blocks */
6449 pRExC_state->num_code_blocks = ncode;
6450 Newx(pRExC_state->code_blocks, ncode, struct reg_code_block);
6455 /* compile-time pattern with just OP_CONSTs and DO blocks */
6460 /* find how many CONSTs there are */
6463 if (expr->op_type == OP_CONST)
6466 for (o = cLISTOPx(expr)->op_first; o; o = OpSIBLING(o)) {
6467 if (o->op_type == OP_CONST)
6471 /* fake up an SV array */
6473 assert(!new_patternp);
6474 Newx(new_patternp, n, SV*);
6475 SAVEFREEPV(new_patternp);
6479 if (expr->op_type == OP_CONST)
6480 new_patternp[n] = cSVOPx_sv(expr);
6482 for (o = cLISTOPx(expr)->op_first; o; o = OpSIBLING(o)) {
6483 if (o->op_type == OP_CONST)
6484 new_patternp[n++] = cSVOPo_sv;
6489 DEBUG_PARSE_r(PerlIO_printf(Perl_debug_log,
6490 "Assembling pattern from %d elements%s\n", pat_count,
6491 orig_rx_flags & RXf_SPLIT ? " for split" : ""));
6493 /* set expr to the first arg op */
6495 if (pRExC_state->num_code_blocks
6496 && expr->op_type != OP_CONST)
6498 expr = cLISTOPx(expr)->op_first;
6499 assert( expr->op_type == OP_PUSHMARK
6500 || (expr->op_type == OP_NULL && expr->op_targ == OP_PUSHMARK)
6501 || expr->op_type == OP_PADRANGE);
6502 expr = OpSIBLING(expr);
6505 pat = S_concat_pat(aTHX_ pRExC_state, NULL, new_patternp, pat_count,
6506 expr, &recompile, NULL);
6508 /* handle bare (possibly after overloading) regex: foo =~ $re */
6513 if (SvTYPE(re) == SVt_REGEXP) {
6517 Safefree(pRExC_state->code_blocks);
6518 DEBUG_PARSE_r(PerlIO_printf(Perl_debug_log,
6519 "Precompiled pattern%s\n",
6520 orig_rx_flags & RXf_SPLIT ? " for split" : ""));
6526 exp = SvPV_nomg(pat, plen);
6528 if (!eng->op_comp) {
6529 if ((SvUTF8(pat) && IN_BYTES)
6530 || SvGMAGICAL(pat) || SvAMAGIC(pat))
6532 /* make a temporary copy; either to convert to bytes,
6533 * or to avoid repeating get-magic / overloaded stringify */
6534 pat = newSVpvn_flags(exp, plen, SVs_TEMP |
6535 (IN_BYTES ? 0 : SvUTF8(pat)));
6537 Safefree(pRExC_state->code_blocks);
6538 return CALLREGCOMP_ENG(eng, pat, orig_rx_flags);
6541 /* ignore the utf8ness if the pattern is 0 length */
6542 RExC_utf8 = RExC_orig_utf8 = (plen == 0 || IN_BYTES) ? 0 : SvUTF8(pat);
6543 RExC_uni_semantics = 0;
6544 RExC_contains_locale = 0;
6545 RExC_contains_i = 0;
6546 RExC_strict = cBOOL(pm_flags & RXf_PMf_STRICT);
6547 pRExC_state->runtime_code_qr = NULL;
6548 RExC_frame_head= NULL;
6549 RExC_frame_last= NULL;
6550 RExC_frame_count= 0;
6553 RExC_mysv1= sv_newmortal();
6554 RExC_mysv2= sv_newmortal();
6557 SV *dsv= sv_newmortal();
6558 RE_PV_QUOTED_DECL(s, RExC_utf8, dsv, exp, plen, 60);
6559 PerlIO_printf(Perl_debug_log, "%sCompiling REx%s %s\n",
6560 PL_colors[4],PL_colors[5],s);
6564 /* we jump here if we upgrade the pattern to utf8 and have to
6567 if ((pm_flags & PMf_USE_RE_EVAL)
6568 /* this second condition covers the non-regex literal case,
6569 * i.e. $foo =~ '(?{})'. */
6570 || (IN_PERL_COMPILETIME && (PL_hints & HINT_RE_EVAL))
6572 runtime_code = S_has_runtime_code(aTHX_ pRExC_state, exp, plen);
6574 /* return old regex if pattern hasn't changed */
6575 /* XXX: note in the below we have to check the flags as well as the
6578 * Things get a touch tricky as we have to compare the utf8 flag
6579 * independently from the compile flags. */
6583 && !!RX_UTF8(old_re) == !!RExC_utf8
6584 && ( RX_COMPFLAGS(old_re) == ( orig_rx_flags & RXf_PMf_FLAGCOPYMASK ) )
6585 && RX_PRECOMP(old_re)
6586 && RX_PRELEN(old_re) == plen
6587 && memEQ(RX_PRECOMP(old_re), exp, plen)
6588 && !runtime_code /* with runtime code, always recompile */ )
6590 Safefree(pRExC_state->code_blocks);
6594 rx_flags = orig_rx_flags;
6596 if (rx_flags & PMf_FOLD) {
6597 RExC_contains_i = 1;
6599 if (RExC_utf8 && initial_charset == REGEX_DEPENDS_CHARSET) {
6601 /* Set to use unicode semantics if the pattern is in utf8 and has the
6602 * 'depends' charset specified, as it means unicode when utf8 */
6603 set_regex_charset(&rx_flags, REGEX_UNICODE_CHARSET);
6607 RExC_flags = rx_flags;
6608 RExC_pm_flags = pm_flags;
6611 if (TAINTING_get && TAINT_get)
6612 Perl_croak(aTHX_ "Eval-group in insecure regular expression");
6614 if (!S_compile_runtime_code(aTHX_ pRExC_state, exp, plen)) {
6615 /* whoops, we have a non-utf8 pattern, whilst run-time code
6616 * got compiled as utf8. Try again with a utf8 pattern */
6617 S_pat_upgrade_to_utf8(aTHX_ pRExC_state, &exp, &plen,
6618 pRExC_state->num_code_blocks);
6619 goto redo_first_pass;
6622 assert(!pRExC_state->runtime_code_qr);
6628 RExC_in_lookbehind = 0;
6629 RExC_seen_zerolen = *exp == '^' ? -1 : 0;
6631 RExC_override_recoding = 0;
6632 RExC_in_multi_char_class = 0;
6634 /* First pass: determine size, legality. */
6637 RExC_end = exp + plen;
6642 RExC_emit = (regnode *) &RExC_emit_dummy;
6643 RExC_whilem_seen = 0;
6644 RExC_open_parens = NULL;
6645 RExC_close_parens = NULL;
6647 RExC_paren_names = NULL;
6649 RExC_paren_name_list = NULL;
6651 RExC_recurse = NULL;
6652 RExC_study_chunk_recursed = NULL;
6653 RExC_study_chunk_recursed_bytes= 0;
6654 RExC_recurse_count = 0;
6655 pRExC_state->code_index = 0;
6658 PerlIO_printf(Perl_debug_log, "Starting first pass (sizing)\n");
6660 RExC_lastparse=NULL;
6662 /* reg may croak on us, not giving us a chance to free
6663 pRExC_state->code_blocks. We cannot SAVEFREEPV it now, as we may
6664 need it to survive as long as the regexp (qr/(?{})/).
6665 We must check that code_blocksv is not already set, because we may
6666 have jumped back to restart the sizing pass. */
6667 if (pRExC_state->code_blocks && !code_blocksv) {
6668 code_blocksv = newSV_type(SVt_PV);
6669 SAVEFREESV(code_blocksv);
6670 SvPV_set(code_blocksv, (char *)pRExC_state->code_blocks);
6671 SvLEN_set(code_blocksv, 1); /*sufficient to make sv_clear free it*/
6673 if (reg(pRExC_state, 0, &flags,1) == NULL) {
6674 /* It's possible to write a regexp in ascii that represents Unicode
6675 codepoints outside of the byte range, such as via \x{100}. If we
6676 detect such a sequence we have to convert the entire pattern to utf8
6677 and then recompile, as our sizing calculation will have been based
6678 on 1 byte == 1 character, but we will need to use utf8 to encode
6679 at least some part of the pattern, and therefore must convert the whole
6682 if (flags & RESTART_UTF8) {
6683 S_pat_upgrade_to_utf8(aTHX_ pRExC_state, &exp, &plen,
6684 pRExC_state->num_code_blocks);
6685 goto redo_first_pass;
6687 Perl_croak(aTHX_ "panic: reg returned NULL to re_op_compile for sizing pass, flags=%#"UVxf"", (UV) flags);
6690 SvLEN_set(code_blocksv,0); /* no you can't have it, sv_clear */
6693 PerlIO_printf(Perl_debug_log,
6694 "Required size %"IVdf" nodes\n"
6695 "Starting second pass (creation)\n",
6698 RExC_lastparse=NULL;
6701 /* The first pass could have found things that force Unicode semantics */
6702 if ((RExC_utf8 || RExC_uni_semantics)
6703 && get_regex_charset(rx_flags) == REGEX_DEPENDS_CHARSET)
6705 set_regex_charset(&rx_flags, REGEX_UNICODE_CHARSET);
6708 /* Small enough for pointer-storage convention?
6709 If extralen==0, this means that we will not need long jumps. */
6710 if (RExC_size >= 0x10000L && RExC_extralen)
6711 RExC_size += RExC_extralen;
6714 if (RExC_whilem_seen > 15)
6715 RExC_whilem_seen = 15;
6717 /* Allocate space and zero-initialize. Note, the two step process
6718 of zeroing when in debug mode, thus anything assigned has to
6719 happen after that */
6720 rx = (REGEXP*) newSV_type(SVt_REGEXP);
6722 Newxc(ri, sizeof(regexp_internal) + (unsigned)RExC_size * sizeof(regnode),
6723 char, regexp_internal);
6724 if ( r == NULL || ri == NULL )
6725 FAIL("Regexp out of space");
6727 /* avoid reading uninitialized memory in DEBUGGING code in study_chunk() */
6728 Zero(ri, sizeof(regexp_internal) + (unsigned)RExC_size * sizeof(regnode),
6731 /* bulk initialize base fields with 0. */
6732 Zero(ri, sizeof(regexp_internal), char);
6735 /* non-zero initialization begins here */
6738 r->extflags = rx_flags;
6739 RXp_COMPFLAGS(r) = orig_rx_flags & RXf_PMf_FLAGCOPYMASK;
6741 if (pm_flags & PMf_IS_QR) {
6742 ri->code_blocks = pRExC_state->code_blocks;
6743 ri->num_code_blocks = pRExC_state->num_code_blocks;
6748 for (n = 0; n < pRExC_state->num_code_blocks; n++)
6749 if (pRExC_state->code_blocks[n].src_regex)
6750 SAVEFREESV(pRExC_state->code_blocks[n].src_regex);
6751 SAVEFREEPV(pRExC_state->code_blocks);
6755 bool has_p = ((r->extflags & RXf_PMf_KEEPCOPY) == RXf_PMf_KEEPCOPY);
6756 bool has_charset = (get_regex_charset(r->extflags)
6757 != REGEX_DEPENDS_CHARSET);
6759 /* The caret is output if there are any defaults: if not all the STD
6760 * flags are set, or if no character set specifier is needed */
6762 (((r->extflags & RXf_PMf_STD_PMMOD) != RXf_PMf_STD_PMMOD)
6764 bool has_runon = ((RExC_seen & REG_RUN_ON_COMMENT_SEEN)
6765 == REG_RUN_ON_COMMENT_SEEN);
6766 U16 reganch = (U16)((r->extflags & RXf_PMf_STD_PMMOD)
6767 >> RXf_PMf_STD_PMMOD_SHIFT);
6768 const char *fptr = STD_PAT_MODS; /*"msixn"*/
6770 /* Allocate for the worst case, which is all the std flags are turned
6771 * on. If more precision is desired, we could do a population count of
6772 * the flags set. This could be done with a small lookup table, or by
6773 * shifting, masking and adding, or even, when available, assembly
6774 * language for a machine-language population count.
6775 * We never output a minus, as all those are defaults, so are
6776 * covered by the caret */
6777 const STRLEN wraplen = plen + has_p + has_runon
6778 + has_default /* If needs a caret */
6780 /* If needs a character set specifier */
6781 + ((has_charset) ? MAX_CHARSET_NAME_LENGTH : 0)
6782 + (sizeof(STD_PAT_MODS) - 1)
6783 + (sizeof("(?:)") - 1);
6785 Newx(p, wraplen + 1, char); /* +1 for the ending NUL */
6786 r->xpv_len_u.xpvlenu_pv = p;
6788 SvFLAGS(rx) |= SVf_UTF8;
6791 /* If a default, cover it using the caret */
6793 *p++= DEFAULT_PAT_MOD;
6797 const char* const name = get_regex_charset_name(r->extflags, &len);
6798 Copy(name, p, len, char);
6802 *p++ = KEEPCOPY_PAT_MOD; /*'p'*/
6805 while((ch = *fptr++)) {
6813 Copy(RExC_precomp, p, plen, char);
6814 assert ((RX_WRAPPED(rx) - p) < 16);
6815 r->pre_prefix = p - RX_WRAPPED(rx);
6821 SvCUR_set(rx, p - RX_WRAPPED(rx));
6825 r->nparens = RExC_npar - 1; /* set early to validate backrefs */
6827 /* setup various meta data about recursion, this all requires
6828 * RExC_npar to be correctly set, and a bit later on we clear it */
6829 if (RExC_seen & REG_RECURSE_SEEN) {
6830 Newxz(RExC_open_parens, RExC_npar,regnode *);
6831 SAVEFREEPV(RExC_open_parens);
6832 Newxz(RExC_close_parens,RExC_npar,regnode *);
6833 SAVEFREEPV(RExC_close_parens);
6835 if (RExC_seen & (REG_RECURSE_SEEN | REG_GOSTART_SEEN)) {
6836 /* Note, RExC_npar is 1 + the number of parens in a pattern.
6837 * So its 1 if there are no parens. */
6838 RExC_study_chunk_recursed_bytes= (RExC_npar >> 3) +
6839 ((RExC_npar & 0x07) != 0);
6840 Newx(RExC_study_chunk_recursed,
6841 RExC_study_chunk_recursed_bytes * RExC_npar, U8);
6842 SAVEFREEPV(RExC_study_chunk_recursed);
6845 /* Useful during FAIL. */
6846 #ifdef RE_TRACK_PATTERN_OFFSETS
6847 Newxz(ri->u.offsets, 2*RExC_size+1, U32); /* MJD 20001228 */
6848 DEBUG_OFFSETS_r(PerlIO_printf(Perl_debug_log,
6849 "%s %"UVuf" bytes for offset annotations.\n",
6850 ri->u.offsets ? "Got" : "Couldn't get",
6851 (UV)((2*RExC_size+1) * sizeof(U32))));
6853 SetProgLen(ri,RExC_size);
6858 /* Second pass: emit code. */
6859 RExC_flags = rx_flags; /* don't let top level (?i) bleed */
6860 RExC_pm_flags = pm_flags;
6862 RExC_end = exp + plen;
6865 RExC_emit_start = ri->program;
6866 RExC_emit = ri->program;
6867 RExC_emit_bound = ri->program + RExC_size + 1;
6868 pRExC_state->code_index = 0;
6870 *((char*) RExC_emit++) = (char) REG_MAGIC;
6871 if (reg(pRExC_state, 0, &flags,1) == NULL) {
6873 Perl_croak(aTHX_ "panic: reg returned NULL to re_op_compile for generation pass, flags=%#"UVxf"", (UV) flags);
6875 /* XXXX To minimize changes to RE engine we always allocate
6876 3-units-long substrs field. */
6877 Newx(r->substrs, 1, struct reg_substr_data);
6878 if (RExC_recurse_count) {
6879 Newxz(RExC_recurse,RExC_recurse_count,regnode *);
6880 SAVEFREEPV(RExC_recurse);
6884 r->minlen = minlen = sawlookahead = sawplus = sawopen = sawminmod = 0;
6886 RExC_study_chunk_recursed_count= 0;
6888 Zero(r->substrs, 1, struct reg_substr_data);
6889 if (RExC_study_chunk_recursed) {
6890 Zero(RExC_study_chunk_recursed,
6891 RExC_study_chunk_recursed_bytes * RExC_npar, U8);
6895 #ifdef TRIE_STUDY_OPT
6897 StructCopy(&zero_scan_data, &data, scan_data_t);
6898 copyRExC_state = RExC_state;
6901 DEBUG_OPTIMISE_r(PerlIO_printf(Perl_debug_log,"Restudying\n"));
6903 RExC_state = copyRExC_state;
6904 if (seen & REG_TOP_LEVEL_BRANCHES_SEEN)
6905 RExC_seen |= REG_TOP_LEVEL_BRANCHES_SEEN;
6907 RExC_seen &= ~REG_TOP_LEVEL_BRANCHES_SEEN;
6908 StructCopy(&zero_scan_data, &data, scan_data_t);
6911 StructCopy(&zero_scan_data, &data, scan_data_t);
6914 /* Dig out information for optimizations. */
6915 r->extflags = RExC_flags; /* was pm_op */
6916 /*dmq: removed as part of de-PMOP: pm->op_pmflags = RExC_flags; */
6919 SvUTF8_on(rx); /* Unicode in it? */
6920 ri->regstclass = NULL;
6921 if (RExC_naughty >= TOO_NAUGHTY) /* Probably an expensive pattern. */
6922 r->intflags |= PREGf_NAUGHTY;
6923 scan = ri->program + 1; /* First BRANCH. */
6925 /* testing for BRANCH here tells us whether there is "must appear"
6926 data in the pattern. If there is then we can use it for optimisations */
6927 if (!(RExC_seen & REG_TOP_LEVEL_BRANCHES_SEEN)) { /* Only one top-level choice.
6930 STRLEN longest_float_length, longest_fixed_length;
6931 regnode_ssc ch_class; /* pointed to by data */
6933 SSize_t last_close = 0; /* pointed to by data */
6934 regnode *first= scan;
6935 regnode *first_next= regnext(first);
6937 * Skip introductions and multiplicators >= 1
6938 * so that we can extract the 'meat' of the pattern that must
6939 * match in the large if() sequence following.
6940 * NOTE that EXACT is NOT covered here, as it is normally
6941 * picked up by the optimiser separately.
6943 * This is unfortunate as the optimiser isnt handling lookahead
6944 * properly currently.
6947 while ((OP(first) == OPEN && (sawopen = 1)) ||
6948 /* An OR of *one* alternative - should not happen now. */
6949 (OP(first) == BRANCH && OP(first_next) != BRANCH) ||
6950 /* for now we can't handle lookbehind IFMATCH*/
6951 (OP(first) == IFMATCH && !first->flags && (sawlookahead = 1)) ||
6952 (OP(first) == PLUS) ||
6953 (OP(first) == MINMOD) ||
6954 /* An {n,m} with n>0 */
6955 (PL_regkind[OP(first)] == CURLY && ARG1(first) > 0) ||
6956 (OP(first) == NOTHING && PL_regkind[OP(first_next)] != END ))
6959 * the only op that could be a regnode is PLUS, all the rest
6960 * will be regnode_1 or regnode_2.
6962 * (yves doesn't think this is true)
6964 if (OP(first) == PLUS)
6967 if (OP(first) == MINMOD)
6969 first += regarglen[OP(first)];
6971 first = NEXTOPER(first);
6972 first_next= regnext(first);
6975 /* Starting-point info. */
6977 DEBUG_PEEP("first:",first,0);
6978 /* Ignore EXACT as we deal with it later. */
6979 if (PL_regkind[OP(first)] == EXACT) {
6980 if (OP(first) == EXACT || OP(first) == EXACTL)
6981 NOOP; /* Empty, get anchored substr later. */
6983 ri->regstclass = first;
6986 else if (PL_regkind[OP(first)] == TRIE &&
6987 ((reg_trie_data *)ri->data->data[ ARG(first) ])->minlen>0)
6989 /* this can happen only on restudy */
6990 ri->regstclass = construct_ahocorasick_from_trie(pRExC_state, (regnode *)first, 0);
6993 else if (REGNODE_SIMPLE(OP(first)))
6994 ri->regstclass = first;
6995 else if (PL_regkind[OP(first)] == BOUND ||
6996 PL_regkind[OP(first)] == NBOUND)
6997 ri->regstclass = first;
6998 else if (PL_regkind[OP(first)] == BOL) {
6999 r->intflags |= (OP(first) == MBOL
7002 first = NEXTOPER(first);
7005 else if (OP(first) == GPOS) {
7006 r->intflags |= PREGf_ANCH_GPOS;
7007 first = NEXTOPER(first);
7010 else if ((!sawopen || !RExC_sawback) &&
7012 (OP(first) == STAR &&
7013 PL_regkind[OP(NEXTOPER(first))] == REG_ANY) &&
7014 !(r->intflags & PREGf_ANCH) && !pRExC_state->num_code_blocks)
7016 /* turn .* into ^.* with an implied $*=1 */
7018 (OP(NEXTOPER(first)) == REG_ANY)
7021 r->intflags |= (type | PREGf_IMPLICIT);
7022 first = NEXTOPER(first);
7025 if (sawplus && !sawminmod && !sawlookahead
7026 && (!sawopen || !RExC_sawback)
7027 && !pRExC_state->num_code_blocks) /* May examine pos and $& */
7028 /* x+ must match at the 1st pos of run of x's */
7029 r->intflags |= PREGf_SKIP;
7031 /* Scan is after the zeroth branch, first is atomic matcher. */
7032 #ifdef TRIE_STUDY_OPT
7035 PerlIO_printf(Perl_debug_log, "first at %"IVdf"\n",
7036 (IV)(first - scan + 1))
7040 PerlIO_printf(Perl_debug_log, "first at %"IVdf"\n",
7041 (IV)(first - scan + 1))
7047 * If there's something expensive in the r.e., find the
7048 * longest literal string that must appear and make it the
7049 * regmust. Resolve ties in favor of later strings, since
7050 * the regstart check works with the beginning of the r.e.
7051 * and avoiding duplication strengthens checking. Not a
7052 * strong reason, but sufficient in the absence of others.
7053 * [Now we resolve ties in favor of the earlier string if
7054 * it happens that c_offset_min has been invalidated, since the
7055 * earlier string may buy us something the later one won't.]
7058 data.longest_fixed = newSVpvs("");
7059 data.longest_float = newSVpvs("");
7060 data.last_found = newSVpvs("");
7061 data.longest = &(data.longest_fixed);
7062 ENTER_with_name("study_chunk");
7063 SAVEFREESV(data.longest_fixed);
7064 SAVEFREESV(data.longest_float);
7065 SAVEFREESV(data.last_found);
7067 if (!ri->regstclass) {
7068 ssc_init(pRExC_state, &ch_class);
7069 data.start_class = &ch_class;
7070 stclass_flag = SCF_DO_STCLASS_AND;
7071 } else /* XXXX Check for BOUND? */
7073 data.last_closep = &last_close;
7076 minlen = study_chunk(pRExC_state, &first, &minlen, &fake,
7077 scan + RExC_size, /* Up to end */
7079 SCF_DO_SUBSTR | SCF_WHILEM_VISITED_POS | stclass_flag
7080 | (restudied ? SCF_TRIE_DOING_RESTUDY : 0),
7084 CHECK_RESTUDY_GOTO_butfirst(LEAVE_with_name("study_chunk"));
7087 if ( RExC_npar == 1 && data.longest == &(data.longest_fixed)
7088 && data.last_start_min == 0 && data.last_end > 0
7089 && !RExC_seen_zerolen
7090 && !(RExC_seen & REG_VERBARG_SEEN)
7091 && !(RExC_seen & REG_GPOS_SEEN)
7093 r->extflags |= RXf_CHECK_ALL;
7095 scan_commit(pRExC_state, &data,&minlen,0);
7097 longest_float_length = CHR_SVLEN(data.longest_float);
7099 if (! ((SvCUR(data.longest_fixed) /* ok to leave SvCUR */
7100 && data.offset_fixed == data.offset_float_min
7101 && SvCUR(data.longest_fixed) == SvCUR(data.longest_float)))
7102 && S_setup_longest (aTHX_ pRExC_state,
7106 &(r->float_end_shift),
7107 data.lookbehind_float,
7108 data.offset_float_min,
7110 longest_float_length,
7111 cBOOL(data.flags & SF_FL_BEFORE_EOL),
7112 cBOOL(data.flags & SF_FL_BEFORE_MEOL)))
7114 r->float_min_offset = data.offset_float_min - data.lookbehind_float;
7115 r->float_max_offset = data.offset_float_max;
7116 if (data.offset_float_max < SSize_t_MAX) /* Don't offset infinity */
7117 r->float_max_offset -= data.lookbehind_float;
7118 SvREFCNT_inc_simple_void_NN(data.longest_float);
7121 r->float_substr = r->float_utf8 = NULL;
7122 longest_float_length = 0;
7125 longest_fixed_length = CHR_SVLEN(data.longest_fixed);
7127 if (S_setup_longest (aTHX_ pRExC_state,
7129 &(r->anchored_utf8),
7130 &(r->anchored_substr),
7131 &(r->anchored_end_shift),
7132 data.lookbehind_fixed,
7135 longest_fixed_length,
7136 cBOOL(data.flags & SF_FIX_BEFORE_EOL),
7137 cBOOL(data.flags & SF_FIX_BEFORE_MEOL)))
7139 r->anchored_offset = data.offset_fixed - data.lookbehind_fixed;
7140 SvREFCNT_inc_simple_void_NN(data.longest_fixed);
7143 r->anchored_substr = r->anchored_utf8 = NULL;
7144 longest_fixed_length = 0;
7146 LEAVE_with_name("study_chunk");
7149 && (OP(ri->regstclass) == REG_ANY || OP(ri->regstclass) == SANY))
7150 ri->regstclass = NULL;
7152 if ((!(r->anchored_substr || r->anchored_utf8) || r->anchored_offset)
7154 && ! (ANYOF_FLAGS(data.start_class) & SSC_MATCHES_EMPTY_STRING)
7155 && is_ssc_worth_it(pRExC_state, data.start_class))
7157 const U32 n = add_data(pRExC_state, STR_WITH_LEN("f"));
7159 ssc_finalize(pRExC_state, data.start_class);
7161 Newx(RExC_rxi->data->data[n], 1, regnode_ssc);
7162 StructCopy(data.start_class,
7163 (regnode_ssc*)RExC_rxi->data->data[n],
7165 ri->regstclass = (regnode*)RExC_rxi->data->data[n];
7166 r->intflags &= ~PREGf_SKIP; /* Used in find_byclass(). */
7167 DEBUG_COMPILE_r({ SV *sv = sv_newmortal();
7168 regprop(r, sv, (regnode*)data.start_class, NULL, pRExC_state);
7169 PerlIO_printf(Perl_debug_log,
7170 "synthetic stclass \"%s\".\n",
7171 SvPVX_const(sv));});
7172 data.start_class = NULL;
7175 /* A temporary algorithm prefers floated substr to fixed one to dig
7177 if (longest_fixed_length > longest_float_length) {
7178 r->substrs->check_ix = 0;
7179 r->check_end_shift = r->anchored_end_shift;
7180 r->check_substr = r->anchored_substr;
7181 r->check_utf8 = r->anchored_utf8;
7182 r->check_offset_min = r->check_offset_max = r->anchored_offset;
7183 if (r->intflags & (PREGf_ANCH_SBOL|PREGf_ANCH_GPOS))
7184 r->intflags |= PREGf_NOSCAN;
7187 r->substrs->check_ix = 1;
7188 r->check_end_shift = r->float_end_shift;
7189 r->check_substr = r->float_substr;
7190 r->check_utf8 = r->float_utf8;
7191 r->check_offset_min = r->float_min_offset;
7192 r->check_offset_max = r->float_max_offset;
7194 if ((r->check_substr || r->check_utf8) ) {
7195 r->extflags |= RXf_USE_INTUIT;
7196 if (SvTAIL(r->check_substr ? r->check_substr : r->check_utf8))
7197 r->extflags |= RXf_INTUIT_TAIL;
7199 r->substrs->data[0].max_offset = r->substrs->data[0].min_offset;
7201 /* XXX Unneeded? dmq (shouldn't as this is handled elsewhere)
7202 if ( (STRLEN)minlen < longest_float_length )
7203 minlen= longest_float_length;
7204 if ( (STRLEN)minlen < longest_fixed_length )
7205 minlen= longest_fixed_length;
7209 /* Several toplevels. Best we can is to set minlen. */
7211 regnode_ssc ch_class;
7212 SSize_t last_close = 0;
7214 DEBUG_PARSE_r(PerlIO_printf(Perl_debug_log, "\nMulti Top Level\n"));
7216 scan = ri->program + 1;
7217 ssc_init(pRExC_state, &ch_class);
7218 data.start_class = &ch_class;
7219 data.last_closep = &last_close;
7222 minlen = study_chunk(pRExC_state,
7223 &scan, &minlen, &fake, scan + RExC_size, &data, -1, 0, NULL,
7224 SCF_DO_STCLASS_AND|SCF_WHILEM_VISITED_POS|(restudied
7225 ? SCF_TRIE_DOING_RESTUDY
7229 CHECK_RESTUDY_GOTO_butfirst(NOOP);
7231 r->check_substr = r->check_utf8 = r->anchored_substr = r->anchored_utf8
7232 = r->float_substr = r->float_utf8 = NULL;
7234 if (! (ANYOF_FLAGS(data.start_class) & SSC_MATCHES_EMPTY_STRING)
7235 && is_ssc_worth_it(pRExC_state, data.start_class))
7237 const U32 n = add_data(pRExC_state, STR_WITH_LEN("f"));
7239 ssc_finalize(pRExC_state, data.start_class);
7241 Newx(RExC_rxi->data->data[n], 1, regnode_ssc);
7242 StructCopy(data.start_class,
7243 (regnode_ssc*)RExC_rxi->data->data[n],
7245 ri->regstclass = (regnode*)RExC_rxi->data->data[n];
7246 r->intflags &= ~PREGf_SKIP; /* Used in find_byclass(). */
7247 DEBUG_COMPILE_r({ SV* sv = sv_newmortal();
7248 regprop(r, sv, (regnode*)data.start_class, NULL, pRExC_state);
7249 PerlIO_printf(Perl_debug_log,
7250 "synthetic stclass \"%s\".\n",
7251 SvPVX_const(sv));});
7252 data.start_class = NULL;
7256 if (RExC_seen & REG_UNBOUNDED_QUANTIFIER_SEEN) {
7257 r->extflags |= RXf_UNBOUNDED_QUANTIFIER_SEEN;
7258 r->maxlen = REG_INFTY;
7261 r->maxlen = RExC_maxlen;
7264 /* Guard against an embedded (?=) or (?<=) with a longer minlen than
7265 the "real" pattern. */
7267 PerlIO_printf(Perl_debug_log,"minlen: %"IVdf" r->minlen:%"IVdf" maxlen:%"IVdf"\n",
7268 (IV)minlen, (IV)r->minlen, (IV)RExC_maxlen);
7270 r->minlenret = minlen;
7271 if (r->minlen < minlen)
7274 if (RExC_seen & REG_GPOS_SEEN)
7275 r->intflags |= PREGf_GPOS_SEEN;
7276 if (RExC_seen & REG_LOOKBEHIND_SEEN)
7277 r->extflags |= RXf_NO_INPLACE_SUBST; /* inplace might break the
7279 if (pRExC_state->num_code_blocks)
7280 r->extflags |= RXf_EVAL_SEEN;
7281 if (RExC_seen & REG_CANY_SEEN)
7282 r->intflags |= PREGf_CANY_SEEN;
7283 if (RExC_seen & REG_VERBARG_SEEN)
7285 r->intflags |= PREGf_VERBARG_SEEN;
7286 r->extflags |= RXf_NO_INPLACE_SUBST; /* don't understand this! Yves */
7288 if (RExC_seen & REG_CUTGROUP_SEEN)
7289 r->intflags |= PREGf_CUTGROUP_SEEN;
7290 if (pm_flags & PMf_USE_RE_EVAL)
7291 r->intflags |= PREGf_USE_RE_EVAL;
7292 if (RExC_paren_names)
7293 RXp_PAREN_NAMES(r) = MUTABLE_HV(SvREFCNT_inc(RExC_paren_names));
7295 RXp_PAREN_NAMES(r) = NULL;
7297 /* If we have seen an anchor in our pattern then we set the extflag RXf_IS_ANCHORED
7298 * so it can be used in pp.c */
7299 if (r->intflags & PREGf_ANCH)
7300 r->extflags |= RXf_IS_ANCHORED;
7304 /* this is used to identify "special" patterns that might result
7305 * in Perl NOT calling the regex engine and instead doing the match "itself",
7306 * particularly special cases in split//. By having the regex compiler
7307 * do this pattern matching at a regop level (instead of by inspecting the pattern)
7308 * we avoid weird issues with equivalent patterns resulting in different behavior,
7309 * AND we allow non Perl engines to get the same optimizations by the setting the
7310 * flags appropriately - Yves */
7311 regnode *first = ri->program + 1;
7313 regnode *next = NEXTOPER(first);
7316 if (PL_regkind[fop] == NOTHING && nop == END)
7317 r->extflags |= RXf_NULL;
7318 else if ((fop == MBOL || (fop == SBOL && !first->flags)) && nop == END)
7319 /* when fop is SBOL first->flags will be true only when it was
7320 * produced by parsing /\A/, and not when parsing /^/. This is
7321 * very important for the split code as there we want to
7322 * treat /^/ as /^/m, but we do not want to treat /\A/ as /^/m.
7323 * See rt #122761 for more details. -- Yves */
7324 r->extflags |= RXf_START_ONLY;
7325 else if (fop == PLUS
7326 && PL_regkind[nop] == POSIXD && FLAGS(next) == _CC_SPACE
7327 && OP(regnext(first)) == END)
7328 r->extflags |= RXf_WHITE;
7329 else if ( r->extflags & RXf_SPLIT
7330 && (fop == EXACT || fop == EXACTL)
7331 && STR_LEN(first) == 1
7332 && *(STRING(first)) == ' '
7333 && OP(regnext(first)) == END )
7334 r->extflags |= (RXf_SKIPWHITE|RXf_WHITE);
7338 if (RExC_contains_locale) {
7339 RXp_EXTFLAGS(r) |= RXf_TAINTED;
7343 if (RExC_paren_names) {
7344 ri->name_list_idx = add_data( pRExC_state, STR_WITH_LEN("a"));
7345 ri->data->data[ri->name_list_idx]
7346 = (void*)SvREFCNT_inc(RExC_paren_name_list);
7349 ri->name_list_idx = 0;
7351 if (RExC_recurse_count) {
7352 for ( ; RExC_recurse_count ; RExC_recurse_count-- ) {
7353 const regnode *scan = RExC_recurse[RExC_recurse_count-1];
7354 ARG2L_SET( scan, RExC_open_parens[ARG(scan)-1] - scan );
7357 Newxz(r->offs, RExC_npar, regexp_paren_pair);
7358 /* assume we don't need to swap parens around before we match */
7360 PerlIO_printf(Perl_debug_log,"study_chunk_recursed_count: %lu\n",
7361 (unsigned long)RExC_study_chunk_recursed_count);
7365 PerlIO_printf(Perl_debug_log,"Final program:\n");
7368 #ifdef RE_TRACK_PATTERN_OFFSETS
7369 DEBUG_OFFSETS_r(if (ri->u.offsets) {
7370 const STRLEN len = ri->u.offsets[0];
7372 GET_RE_DEBUG_FLAGS_DECL;
7373 PerlIO_printf(Perl_debug_log,
7374 "Offsets: [%"UVuf"]\n\t", (UV)ri->u.offsets[0]);
7375 for (i = 1; i <= len; i++) {
7376 if (ri->u.offsets[i*2-1] || ri->u.offsets[i*2])
7377 PerlIO_printf(Perl_debug_log, "%"UVuf":%"UVuf"[%"UVuf"] ",
7378 (UV)i, (UV)ri->u.offsets[i*2-1], (UV)ri->u.offsets[i*2]);
7380 PerlIO_printf(Perl_debug_log, "\n");
7385 /* under ithreads the ?pat? PMf_USED flag on the pmop is simulated
7386 * by setting the regexp SV to readonly-only instead. If the
7387 * pattern's been recompiled, the USEDness should remain. */
7388 if (old_re && SvREADONLY(old_re))
7396 Perl_reg_named_buff(pTHX_ REGEXP * const rx, SV * const key, SV * const value,
7399 PERL_ARGS_ASSERT_REG_NAMED_BUFF;
7401 PERL_UNUSED_ARG(value);
7403 if (flags & RXapif_FETCH) {
7404 return reg_named_buff_fetch(rx, key, flags);
7405 } else if (flags & (RXapif_STORE | RXapif_DELETE | RXapif_CLEAR)) {
7406 Perl_croak_no_modify();
7408 } else if (flags & RXapif_EXISTS) {
7409 return reg_named_buff_exists(rx, key, flags)
7412 } else if (flags & RXapif_REGNAMES) {
7413 return reg_named_buff_all(rx, flags);
7414 } else if (flags & (RXapif_SCALAR | RXapif_REGNAMES_COUNT)) {
7415 return reg_named_buff_scalar(rx, flags);
7417 Perl_croak(aTHX_ "panic: Unknown flags %d in named_buff", (int)flags);
7423 Perl_reg_named_buff_iter(pTHX_ REGEXP * const rx, const SV * const lastkey,
7426 PERL_ARGS_ASSERT_REG_NAMED_BUFF_ITER;
7427 PERL_UNUSED_ARG(lastkey);
7429 if (flags & RXapif_FIRSTKEY)
7430 return reg_named_buff_firstkey(rx, flags);
7431 else if (flags & RXapif_NEXTKEY)
7432 return reg_named_buff_nextkey(rx, flags);
7434 Perl_croak(aTHX_ "panic: Unknown flags %d in named_buff_iter",
7441 Perl_reg_named_buff_fetch(pTHX_ REGEXP * const r, SV * const namesv,
7444 AV *retarray = NULL;
7446 struct regexp *const rx = ReANY(r);
7448 PERL_ARGS_ASSERT_REG_NAMED_BUFF_FETCH;
7450 if (flags & RXapif_ALL)
7453 if (rx && RXp_PAREN_NAMES(rx)) {
7454 HE *he_str = hv_fetch_ent( RXp_PAREN_NAMES(rx), namesv, 0, 0 );
7457 SV* sv_dat=HeVAL(he_str);
7458 I32 *nums=(I32*)SvPVX(sv_dat);
7459 for ( i=0; i<SvIVX(sv_dat); i++ ) {
7460 if ((I32)(rx->nparens) >= nums[i]
7461 && rx->offs[nums[i]].start != -1
7462 && rx->offs[nums[i]].end != -1)
7465 CALLREG_NUMBUF_FETCH(r,nums[i],ret);
7470 ret = newSVsv(&PL_sv_undef);
7473 av_push(retarray, ret);
7476 return newRV_noinc(MUTABLE_SV(retarray));
7483 Perl_reg_named_buff_exists(pTHX_ REGEXP * const r, SV * const key,
7486 struct regexp *const rx = ReANY(r);
7488 PERL_ARGS_ASSERT_REG_NAMED_BUFF_EXISTS;
7490 if (rx && RXp_PAREN_NAMES(rx)) {
7491 if (flags & RXapif_ALL) {
7492 return hv_exists_ent(RXp_PAREN_NAMES(rx), key, 0);
7494 SV *sv = CALLREG_NAMED_BUFF_FETCH(r, key, flags);
7496 SvREFCNT_dec_NN(sv);
7508 Perl_reg_named_buff_firstkey(pTHX_ REGEXP * const r, const U32 flags)
7510 struct regexp *const rx = ReANY(r);
7512 PERL_ARGS_ASSERT_REG_NAMED_BUFF_FIRSTKEY;
7514 if ( rx && RXp_PAREN_NAMES(rx) ) {
7515 (void)hv_iterinit(RXp_PAREN_NAMES(rx));
7517 return CALLREG_NAMED_BUFF_NEXTKEY(r, NULL, flags & ~RXapif_FIRSTKEY);
7524 Perl_reg_named_buff_nextkey(pTHX_ REGEXP * const r, const U32 flags)
7526 struct regexp *const rx = ReANY(r);
7527 GET_RE_DEBUG_FLAGS_DECL;
7529 PERL_ARGS_ASSERT_REG_NAMED_BUFF_NEXTKEY;
7531 if (rx && RXp_PAREN_NAMES(rx)) {
7532 HV *hv = RXp_PAREN_NAMES(rx);
7534 while ( (temphe = hv_iternext_flags(hv,0)) ) {
7537 SV* sv_dat = HeVAL(temphe);
7538 I32 *nums = (I32*)SvPVX(sv_dat);
7539 for ( i = 0; i < SvIVX(sv_dat); i++ ) {
7540 if ((I32)(rx->lastparen) >= nums[i] &&
7541 rx->offs[nums[i]].start != -1 &&
7542 rx->offs[nums[i]].end != -1)
7548 if (parno || flags & RXapif_ALL) {
7549 return newSVhek(HeKEY_hek(temphe));
7557 Perl_reg_named_buff_scalar(pTHX_ REGEXP * const r, const U32 flags)
7562 struct regexp *const rx = ReANY(r);
7564 PERL_ARGS_ASSERT_REG_NAMED_BUFF_SCALAR;
7566 if (rx && RXp_PAREN_NAMES(rx)) {
7567 if (flags & (RXapif_ALL | RXapif_REGNAMES_COUNT)) {
7568 return newSViv(HvTOTALKEYS(RXp_PAREN_NAMES(rx)));
7569 } else if (flags & RXapif_ONE) {
7570 ret = CALLREG_NAMED_BUFF_ALL(r, (flags | RXapif_REGNAMES));
7571 av = MUTABLE_AV(SvRV(ret));
7572 length = av_tindex(av);
7573 SvREFCNT_dec_NN(ret);
7574 return newSViv(length + 1);
7576 Perl_croak(aTHX_ "panic: Unknown flags %d in named_buff_scalar",
7581 return &PL_sv_undef;
7585 Perl_reg_named_buff_all(pTHX_ REGEXP * const r, const U32 flags)
7587 struct regexp *const rx = ReANY(r);
7590 PERL_ARGS_ASSERT_REG_NAMED_BUFF_ALL;
7592 if (rx && RXp_PAREN_NAMES(rx)) {
7593 HV *hv= RXp_PAREN_NAMES(rx);
7595 (void)hv_iterinit(hv);
7596 while ( (temphe = hv_iternext_flags(hv,0)) ) {
7599 SV* sv_dat = HeVAL(temphe);
7600 I32 *nums = (I32*)SvPVX(sv_dat);
7601 for ( i = 0; i < SvIVX(sv_dat); i++ ) {
7602 if ((I32)(rx->lastparen) >= nums[i] &&
7603 rx->offs[nums[i]].start != -1 &&
7604 rx->offs[nums[i]].end != -1)
7610 if (parno || flags & RXapif_ALL) {
7611 av_push(av, newSVhek(HeKEY_hek(temphe)));
7616 return newRV_noinc(MUTABLE_SV(av));
7620 Perl_reg_numbered_buff_fetch(pTHX_ REGEXP * const r, const I32 paren,
7623 struct regexp *const rx = ReANY(r);
7629 PERL_ARGS_ASSERT_REG_NUMBERED_BUFF_FETCH;
7631 if ( n == RX_BUFF_IDX_CARET_PREMATCH
7632 || n == RX_BUFF_IDX_CARET_FULLMATCH
7633 || n == RX_BUFF_IDX_CARET_POSTMATCH
7636 bool keepcopy = cBOOL(rx->extflags & RXf_PMf_KEEPCOPY);
7638 /* on something like
7641 * the KEEPCOPY is set on the PMOP rather than the regex */
7642 if (PL_curpm && r == PM_GETRE(PL_curpm))
7643 keepcopy = cBOOL(PL_curpm->op_pmflags & PMf_KEEPCOPY);
7652 if (n == RX_BUFF_IDX_CARET_FULLMATCH)
7653 /* no need to distinguish between them any more */
7654 n = RX_BUFF_IDX_FULLMATCH;
7656 if ((n == RX_BUFF_IDX_PREMATCH || n == RX_BUFF_IDX_CARET_PREMATCH)
7657 && rx->offs[0].start != -1)
7659 /* $`, ${^PREMATCH} */
7660 i = rx->offs[0].start;
7664 if ((n == RX_BUFF_IDX_POSTMATCH || n == RX_BUFF_IDX_CARET_POSTMATCH)
7665 && rx->offs[0].end != -1)
7667 /* $', ${^POSTMATCH} */
7668 s = rx->subbeg - rx->suboffset + rx->offs[0].end;
7669 i = rx->sublen + rx->suboffset - rx->offs[0].end;
7672 if ( 0 <= n && n <= (I32)rx->nparens &&
7673 (s1 = rx->offs[n].start) != -1 &&
7674 (t1 = rx->offs[n].end) != -1)
7676 /* $&, ${^MATCH}, $1 ... */
7678 s = rx->subbeg + s1 - rx->suboffset;
7683 assert(s >= rx->subbeg);
7684 assert((STRLEN)rx->sublen >= (STRLEN)((s - rx->subbeg) + i) );
7686 #ifdef NO_TAINT_SUPPORT
7687 sv_setpvn(sv, s, i);
7689 const int oldtainted = TAINT_get;
7691 sv_setpvn(sv, s, i);
7692 TAINT_set(oldtainted);
7694 if ( (rx->intflags & PREGf_CANY_SEEN)
7695 ? (RXp_MATCH_UTF8(rx)
7696 && (!i || is_utf8_string((U8*)s, i)))
7697 : (RXp_MATCH_UTF8(rx)) )
7704 if (RXp_MATCH_TAINTED(rx)) {
7705 if (SvTYPE(sv) >= SVt_PVMG) {
7706 MAGIC* const mg = SvMAGIC(sv);
7709 SvMAGIC_set(sv, mg->mg_moremagic);
7711 if ((mgt = SvMAGIC(sv))) {
7712 mg->mg_moremagic = mgt;
7713 SvMAGIC_set(sv, mg);
7724 sv_setsv(sv,&PL_sv_undef);
7730 Perl_reg_numbered_buff_store(pTHX_ REGEXP * const rx, const I32 paren,
7731 SV const * const value)
7733 PERL_ARGS_ASSERT_REG_NUMBERED_BUFF_STORE;
7735 PERL_UNUSED_ARG(rx);
7736 PERL_UNUSED_ARG(paren);
7737 PERL_UNUSED_ARG(value);
7740 Perl_croak_no_modify();
7744 Perl_reg_numbered_buff_length(pTHX_ REGEXP * const r, const SV * const sv,
7747 struct regexp *const rx = ReANY(r);
7751 PERL_ARGS_ASSERT_REG_NUMBERED_BUFF_LENGTH;
7753 if ( paren == RX_BUFF_IDX_CARET_PREMATCH
7754 || paren == RX_BUFF_IDX_CARET_FULLMATCH
7755 || paren == RX_BUFF_IDX_CARET_POSTMATCH
7758 bool keepcopy = cBOOL(rx->extflags & RXf_PMf_KEEPCOPY);
7760 /* on something like
7763 * the KEEPCOPY is set on the PMOP rather than the regex */
7764 if (PL_curpm && r == PM_GETRE(PL_curpm))
7765 keepcopy = cBOOL(PL_curpm->op_pmflags & PMf_KEEPCOPY);
7771 /* Some of this code was originally in C<Perl_magic_len> in F<mg.c> */
7773 case RX_BUFF_IDX_CARET_PREMATCH: /* ${^PREMATCH} */
7774 case RX_BUFF_IDX_PREMATCH: /* $` */
7775 if (rx->offs[0].start != -1) {
7776 i = rx->offs[0].start;
7785 case RX_BUFF_IDX_CARET_POSTMATCH: /* ${^POSTMATCH} */
7786 case RX_BUFF_IDX_POSTMATCH: /* $' */
7787 if (rx->offs[0].end != -1) {
7788 i = rx->sublen - rx->offs[0].end;
7790 s1 = rx->offs[0].end;
7797 default: /* $& / ${^MATCH}, $1, $2, ... */
7798 if (paren <= (I32)rx->nparens &&
7799 (s1 = rx->offs[paren].start) != -1 &&
7800 (t1 = rx->offs[paren].end) != -1)
7806 if (ckWARN(WARN_UNINITIALIZED))
7807 report_uninit((const SV *)sv);
7812 if (i > 0 && RXp_MATCH_UTF8(rx)) {
7813 const char * const s = rx->subbeg - rx->suboffset + s1;
7818 if (is_utf8_string_loclen((U8*)s, i, &ep, &el))
7825 Perl_reg_qr_package(pTHX_ REGEXP * const rx)
7827 PERL_ARGS_ASSERT_REG_QR_PACKAGE;
7828 PERL_UNUSED_ARG(rx);
7832 return newSVpvs("Regexp");
7835 /* Scans the name of a named buffer from the pattern.
7836 * If flags is REG_RSN_RETURN_NULL returns null.
7837 * If flags is REG_RSN_RETURN_NAME returns an SV* containing the name
7838 * If flags is REG_RSN_RETURN_DATA returns the data SV* corresponding
7839 * to the parsed name as looked up in the RExC_paren_names hash.
7840 * If there is an error throws a vFAIL().. type exception.
7843 #define REG_RSN_RETURN_NULL 0
7844 #define REG_RSN_RETURN_NAME 1
7845 #define REG_RSN_RETURN_DATA 2
7848 S_reg_scan_name(pTHX_ RExC_state_t *pRExC_state, U32 flags)
7850 char *name_start = RExC_parse;
7852 PERL_ARGS_ASSERT_REG_SCAN_NAME;
7854 assert (RExC_parse <= RExC_end);
7855 if (RExC_parse == RExC_end) NOOP;
7856 else if (isIDFIRST_lazy_if(RExC_parse, UTF)) {
7857 /* skip IDFIRST by using do...while */
7860 RExC_parse += UTF8SKIP(RExC_parse);
7861 } while (isWORDCHAR_utf8((U8*)RExC_parse));
7865 } while (isWORDCHAR(*RExC_parse));
7867 RExC_parse++; /* so the <- from the vFAIL is after the offending
7869 vFAIL("Group name must start with a non-digit word character");
7873 = newSVpvn_flags(name_start, (int)(RExC_parse - name_start),
7874 SVs_TEMP | (UTF ? SVf_UTF8 : 0));
7875 if ( flags == REG_RSN_RETURN_NAME)
7877 else if (flags==REG_RSN_RETURN_DATA) {
7880 if ( ! sv_name ) /* should not happen*/
7881 Perl_croak(aTHX_ "panic: no svname in reg_scan_name");
7882 if (RExC_paren_names)
7883 he_str = hv_fetch_ent( RExC_paren_names, sv_name, 0, 0 );
7885 sv_dat = HeVAL(he_str);
7887 vFAIL("Reference to nonexistent named group");
7891 Perl_croak(aTHX_ "panic: bad flag %lx in reg_scan_name",
7892 (unsigned long) flags);
7894 NOT_REACHED; /* NOTREACHED */
7899 #define DEBUG_PARSE_MSG(funcname) DEBUG_PARSE_r({ \
7901 if (RExC_lastparse!=RExC_parse) { \
7902 PerlIO_printf(Perl_debug_log, "%s", \
7903 Perl_pv_pretty(aTHX_ RExC_mysv1, RExC_parse, \
7904 RExC_end - RExC_parse, 16, \
7906 PERL_PV_ESCAPE_UNI_DETECT | \
7907 PERL_PV_PRETTY_ELLIPSES | \
7908 PERL_PV_PRETTY_LTGT | \
7909 PERL_PV_ESCAPE_RE | \
7910 PERL_PV_PRETTY_EXACTSIZE \
7914 PerlIO_printf(Perl_debug_log,"%16s",""); \
7917 num = RExC_size + 1; \
7919 num=REG_NODE_NUM(RExC_emit); \
7920 if (RExC_lastnum!=num) \
7921 PerlIO_printf(Perl_debug_log,"|%4d",num); \
7923 PerlIO_printf(Perl_debug_log,"|%4s",""); \
7924 PerlIO_printf(Perl_debug_log,"|%*s%-4s", \
7925 (int)((depth*2)), "", \
7929 RExC_lastparse=RExC_parse; \
7934 #define DEBUG_PARSE(funcname) DEBUG_PARSE_r({ \
7935 DEBUG_PARSE_MSG((funcname)); \
7936 PerlIO_printf(Perl_debug_log,"%4s","\n"); \
7938 #define DEBUG_PARSE_FMT(funcname,fmt,args) DEBUG_PARSE_r({ \
7939 DEBUG_PARSE_MSG((funcname)); \
7940 PerlIO_printf(Perl_debug_log,fmt "\n",args); \
7943 /* This section of code defines the inversion list object and its methods. The
7944 * interfaces are highly subject to change, so as much as possible is static to
7945 * this file. An inversion list is here implemented as a malloc'd C UV array
7946 * as an SVt_INVLIST scalar.
7948 * An inversion list for Unicode is an array of code points, sorted by ordinal
7949 * number. The zeroth element is the first code point in the list. The 1th
7950 * element is the first element beyond that not in the list. In other words,
7951 * the first range is
7952 * invlist[0]..(invlist[1]-1)
7953 * The other ranges follow. Thus every element whose index is divisible by two
7954 * marks the beginning of a range that is in the list, and every element not
7955 * divisible by two marks the beginning of a range not in the list. A single
7956 * element inversion list that contains the single code point N generally
7957 * consists of two elements
7960 * (The exception is when N is the highest representable value on the
7961 * machine, in which case the list containing just it would be a single
7962 * element, itself. By extension, if the last range in the list extends to
7963 * infinity, then the first element of that range will be in the inversion list
7964 * at a position that is divisible by two, and is the final element in the
7966 * Taking the complement (inverting) an inversion list is quite simple, if the
7967 * first element is 0, remove it; otherwise add a 0 element at the beginning.
7968 * This implementation reserves an element at the beginning of each inversion
7969 * list to always contain 0; there is an additional flag in the header which
7970 * indicates if the list begins at the 0, or is offset to begin at the next
7973 * More about inversion lists can be found in "Unicode Demystified"
7974 * Chapter 13 by Richard Gillam, published by Addison-Wesley.
7975 * More will be coming when functionality is added later.
7977 * The inversion list data structure is currently implemented as an SV pointing
7978 * to an array of UVs that the SV thinks are bytes. This allows us to have an
7979 * array of UV whose memory management is automatically handled by the existing
7980 * facilities for SV's.
7982 * Some of the methods should always be private to the implementation, and some
7983 * should eventually be made public */
7985 /* The header definitions are in F<inline_invlist.c> */
7987 PERL_STATIC_INLINE UV*
7988 S__invlist_array_init(SV* const invlist, const bool will_have_0)
7990 /* Returns a pointer to the first element in the inversion list's array.
7991 * This is called upon initialization of an inversion list. Where the
7992 * array begins depends on whether the list has the code point U+0000 in it
7993 * or not. The other parameter tells it whether the code that follows this
7994 * call is about to put a 0 in the inversion list or not. The first
7995 * element is either the element reserved for 0, if TRUE, or the element
7996 * after it, if FALSE */
7998 bool* offset = get_invlist_offset_addr(invlist);
7999 UV* zero_addr = (UV *) SvPVX(invlist);
8001 PERL_ARGS_ASSERT__INVLIST_ARRAY_INIT;
8004 assert(! _invlist_len(invlist));
8008 /* 1^1 = 0; 1^0 = 1 */
8009 *offset = 1 ^ will_have_0;
8010 return zero_addr + *offset;
8013 PERL_STATIC_INLINE void
8014 S_invlist_set_len(pTHX_ SV* const invlist, const UV len, const bool offset)
8016 /* Sets the current number of elements stored in the inversion list.
8017 * Updates SvCUR correspondingly */
8018 PERL_UNUSED_CONTEXT;
8019 PERL_ARGS_ASSERT_INVLIST_SET_LEN;
8021 assert(SvTYPE(invlist) == SVt_INVLIST);
8026 : TO_INTERNAL_SIZE(len + offset));
8027 assert(SvLEN(invlist) == 0 || SvCUR(invlist) <= SvLEN(invlist));
8030 #ifndef PERL_IN_XSUB_RE
8032 PERL_STATIC_INLINE IV*
8033 S_get_invlist_previous_index_addr(SV* invlist)
8035 /* Return the address of the IV that is reserved to hold the cached index
8037 PERL_ARGS_ASSERT_GET_INVLIST_PREVIOUS_INDEX_ADDR;
8039 assert(SvTYPE(invlist) == SVt_INVLIST);
8041 return &(((XINVLIST*) SvANY(invlist))->prev_index);
8044 PERL_STATIC_INLINE IV
8045 S_invlist_previous_index(SV* const invlist)
8047 /* Returns cached index of previous search */
8049 PERL_ARGS_ASSERT_INVLIST_PREVIOUS_INDEX;
8051 return *get_invlist_previous_index_addr(invlist);
8054 PERL_STATIC_INLINE void
8055 S_invlist_set_previous_index(SV* const invlist, const IV index)
8057 /* Caches <index> for later retrieval */
8059 PERL_ARGS_ASSERT_INVLIST_SET_PREVIOUS_INDEX;
8061 assert(index == 0 || index < (int) _invlist_len(invlist));
8063 *get_invlist_previous_index_addr(invlist) = index;
8066 PERL_STATIC_INLINE void
8067 S_invlist_trim(SV* const invlist)
8069 PERL_ARGS_ASSERT_INVLIST_TRIM;
8071 assert(SvTYPE(invlist) == SVt_INVLIST);
8073 /* Change the length of the inversion list to how many entries it currently
8075 SvPV_shrink_to_cur((SV *) invlist);
8078 PERL_STATIC_INLINE bool
8079 S_invlist_is_iterating(SV* const invlist)
8081 PERL_ARGS_ASSERT_INVLIST_IS_ITERATING;
8083 return *(get_invlist_iter_addr(invlist)) < (STRLEN) UV_MAX;
8086 #endif /* ifndef PERL_IN_XSUB_RE */
8088 PERL_STATIC_INLINE UV
8089 S_invlist_max(SV* const invlist)
8091 /* Returns the maximum number of elements storable in the inversion list's
8092 * array, without having to realloc() */
8094 PERL_ARGS_ASSERT_INVLIST_MAX;
8096 assert(SvTYPE(invlist) == SVt_INVLIST);
8098 /* Assumes worst case, in which the 0 element is not counted in the
8099 * inversion list, so subtracts 1 for that */
8100 return SvLEN(invlist) == 0 /* This happens under _new_invlist_C_array */
8101 ? FROM_INTERNAL_SIZE(SvCUR(invlist)) - 1
8102 : FROM_INTERNAL_SIZE(SvLEN(invlist)) - 1;
8105 #ifndef PERL_IN_XSUB_RE
8107 Perl__new_invlist(pTHX_ IV initial_size)
8110 /* Return a pointer to a newly constructed inversion list, with enough
8111 * space to store 'initial_size' elements. If that number is negative, a
8112 * system default is used instead */
8116 if (initial_size < 0) {
8120 /* Allocate the initial space */
8121 new_list = newSV_type(SVt_INVLIST);
8123 /* First 1 is in case the zero element isn't in the list; second 1 is for
8125 SvGROW(new_list, TO_INTERNAL_SIZE(initial_size + 1) + 1);
8126 invlist_set_len(new_list, 0, 0);
8128 /* Force iterinit() to be used to get iteration to work */
8129 *get_invlist_iter_addr(new_list) = (STRLEN) UV_MAX;
8131 *get_invlist_previous_index_addr(new_list) = 0;
8137 Perl__new_invlist_C_array(pTHX_ const UV* const list)
8139 /* Return a pointer to a newly constructed inversion list, initialized to
8140 * point to <list>, which has to be in the exact correct inversion list
8141 * form, including internal fields. Thus this is a dangerous routine that
8142 * should not be used in the wrong hands. The passed in 'list' contains
8143 * several header fields at the beginning that are not part of the
8144 * inversion list body proper */
8146 const STRLEN length = (STRLEN) list[0];
8147 const UV version_id = list[1];
8148 const bool offset = cBOOL(list[2]);
8149 #define HEADER_LENGTH 3
8150 /* If any of the above changes in any way, you must change HEADER_LENGTH
8151 * (if appropriate) and regenerate INVLIST_VERSION_ID by running
8152 * perl -E 'say int(rand 2**31-1)'
8154 #define INVLIST_VERSION_ID 148565664 /* This is a combination of a version and
8155 data structure type, so that one being
8156 passed in can be validated to be an
8157 inversion list of the correct vintage.
8160 SV* invlist = newSV_type(SVt_INVLIST);
8162 PERL_ARGS_ASSERT__NEW_INVLIST_C_ARRAY;
8164 if (version_id != INVLIST_VERSION_ID) {
8165 Perl_croak(aTHX_ "panic: Incorrect version for previously generated inversion list");
8168 /* The generated array passed in includes header elements that aren't part
8169 * of the list proper, so start it just after them */
8170 SvPV_set(invlist, (char *) (list + HEADER_LENGTH));
8172 SvLEN_set(invlist, 0); /* Means we own the contents, and the system
8173 shouldn't touch it */
8175 *(get_invlist_offset_addr(invlist)) = offset;
8177 /* The 'length' passed to us is the physical number of elements in the
8178 * inversion list. But if there is an offset the logical number is one
8180 invlist_set_len(invlist, length - offset, offset);
8182 invlist_set_previous_index(invlist, 0);
8184 /* Initialize the iteration pointer. */
8185 invlist_iterfinish(invlist);
8187 SvREADONLY_on(invlist);
8191 #endif /* ifndef PERL_IN_XSUB_RE */
8194 S_invlist_extend(pTHX_ SV* const invlist, const UV new_max)
8196 /* Grow the maximum size of an inversion list */
8198 PERL_ARGS_ASSERT_INVLIST_EXTEND;
8200 assert(SvTYPE(invlist) == SVt_INVLIST);
8202 /* Add one to account for the zero element at the beginning which may not
8203 * be counted by the calling parameters */
8204 SvGROW((SV *)invlist, TO_INTERNAL_SIZE(new_max + 1));
8208 S__append_range_to_invlist(pTHX_ SV* const invlist,
8209 const UV start, const UV end)
8211 /* Subject to change or removal. Append the range from 'start' to 'end' at
8212 * the end of the inversion list. The range must be above any existing
8216 UV max = invlist_max(invlist);
8217 UV len = _invlist_len(invlist);
8220 PERL_ARGS_ASSERT__APPEND_RANGE_TO_INVLIST;
8222 if (len == 0) { /* Empty lists must be initialized */
8223 offset = start != 0;
8224 array = _invlist_array_init(invlist, ! offset);
8227 /* Here, the existing list is non-empty. The current max entry in the
8228 * list is generally the first value not in the set, except when the
8229 * set extends to the end of permissible values, in which case it is
8230 * the first entry in that final set, and so this call is an attempt to
8231 * append out-of-order */
8233 UV final_element = len - 1;
8234 array = invlist_array(invlist);
8235 if (array[final_element] > start
8236 || ELEMENT_RANGE_MATCHES_INVLIST(final_element))
8238 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",
8239 array[final_element], start,
8240 ELEMENT_RANGE_MATCHES_INVLIST(final_element) ? 't' : 'f');
8243 /* Here, it is a legal append. If the new range begins with the first
8244 * value not in the set, it is extending the set, so the new first
8245 * value not in the set is one greater than the newly extended range.
8247 offset = *get_invlist_offset_addr(invlist);
8248 if (array[final_element] == start) {
8249 if (end != UV_MAX) {
8250 array[final_element] = end + 1;
8253 /* But if the end is the maximum representable on the machine,
8254 * just let the range that this would extend to have no end */
8255 invlist_set_len(invlist, len - 1, offset);
8261 /* Here the new range doesn't extend any existing set. Add it */
8263 len += 2; /* Includes an element each for the start and end of range */
8265 /* If wll overflow the existing space, extend, which may cause the array to
8268 invlist_extend(invlist, len);
8270 /* Have to set len here to avoid assert failure in invlist_array() */
8271 invlist_set_len(invlist, len, offset);
8273 array = invlist_array(invlist);
8276 invlist_set_len(invlist, len, offset);
8279 /* The next item on the list starts the range, the one after that is
8280 * one past the new range. */
8281 array[len - 2] = start;
8282 if (end != UV_MAX) {
8283 array[len - 1] = end + 1;
8286 /* But if the end is the maximum representable on the machine, just let
8287 * the range have no end */
8288 invlist_set_len(invlist, len - 1, offset);
8292 #ifndef PERL_IN_XSUB_RE
8295 Perl__invlist_search(SV* const invlist, const UV cp)
8297 /* Searches the inversion list for the entry that contains the input code
8298 * point <cp>. If <cp> is not in the list, -1 is returned. Otherwise, the
8299 * return value is the index into the list's array of the range that
8304 IV high = _invlist_len(invlist);
8305 const IV highest_element = high - 1;
8308 PERL_ARGS_ASSERT__INVLIST_SEARCH;
8310 /* If list is empty, return failure. */
8315 /* (We can't get the array unless we know the list is non-empty) */
8316 array = invlist_array(invlist);
8318 mid = invlist_previous_index(invlist);
8319 assert(mid >=0 && mid <= highest_element);
8321 /* <mid> contains the cache of the result of the previous call to this
8322 * function (0 the first time). See if this call is for the same result,
8323 * or if it is for mid-1. This is under the theory that calls to this
8324 * function will often be for related code points that are near each other.
8325 * And benchmarks show that caching gives better results. We also test
8326 * here if the code point is within the bounds of the list. These tests
8327 * replace others that would have had to be made anyway to make sure that
8328 * the array bounds were not exceeded, and these give us extra information
8329 * at the same time */
8330 if (cp >= array[mid]) {
8331 if (cp >= array[highest_element]) {
8332 return highest_element;
8335 /* Here, array[mid] <= cp < array[highest_element]. This means that
8336 * the final element is not the answer, so can exclude it; it also
8337 * means that <mid> is not the final element, so can refer to 'mid + 1'
8339 if (cp < array[mid + 1]) {
8345 else { /* cp < aray[mid] */
8346 if (cp < array[0]) { /* Fail if outside the array */
8350 if (cp >= array[mid - 1]) {
8355 /* Binary search. What we are looking for is <i> such that
8356 * array[i] <= cp < array[i+1]
8357 * The loop below converges on the i+1. Note that there may not be an
8358 * (i+1)th element in the array, and things work nonetheless */
8359 while (low < high) {
8360 mid = (low + high) / 2;
8361 assert(mid <= highest_element);
8362 if (array[mid] <= cp) { /* cp >= array[mid] */
8365 /* We could do this extra test to exit the loop early.
8366 if (cp < array[low]) {
8371 else { /* cp < array[mid] */
8378 invlist_set_previous_index(invlist, high);
8383 Perl__invlist_populate_swatch(SV* const invlist,
8384 const UV start, const UV end, U8* swatch)
8386 /* populates a swatch of a swash the same way swatch_get() does in utf8.c,
8387 * but is used when the swash has an inversion list. This makes this much
8388 * faster, as it uses a binary search instead of a linear one. This is
8389 * intimately tied to that function, and perhaps should be in utf8.c,
8390 * except it is intimately tied to inversion lists as well. It assumes
8391 * that <swatch> is all 0's on input */
8394 const IV len = _invlist_len(invlist);
8398 PERL_ARGS_ASSERT__INVLIST_POPULATE_SWATCH;
8400 if (len == 0) { /* Empty inversion list */
8404 array = invlist_array(invlist);
8406 /* Find which element it is */
8407 i = _invlist_search(invlist, start);
8409 /* We populate from <start> to <end> */
8410 while (current < end) {
8413 /* The inversion list gives the results for every possible code point
8414 * after the first one in the list. Only those ranges whose index is
8415 * even are ones that the inversion list matches. For the odd ones,
8416 * and if the initial code point is not in the list, we have to skip
8417 * forward to the next element */
8418 if (i == -1 || ! ELEMENT_RANGE_MATCHES_INVLIST(i)) {
8420 if (i >= len) { /* Finished if beyond the end of the array */
8424 if (current >= end) { /* Finished if beyond the end of what we
8426 if (LIKELY(end < UV_MAX)) {
8430 /* We get here when the upper bound is the maximum
8431 * representable on the machine, and we are looking for just
8432 * that code point. Have to special case it */
8434 goto join_end_of_list;
8437 assert(current >= start);
8439 /* The current range ends one below the next one, except don't go past
8442 upper = (i < len && array[i] < end) ? array[i] : end;
8444 /* Here we are in a range that matches. Populate a bit in the 3-bit U8
8445 * for each code point in it */
8446 for (; current < upper; current++) {
8447 const STRLEN offset = (STRLEN)(current - start);
8448 swatch[offset >> 3] |= 1 << (offset & 7);
8453 /* Quit if at the end of the list */
8456 /* But first, have to deal with the highest possible code point on
8457 * the platform. The previous code assumes that <end> is one
8458 * beyond where we want to populate, but that is impossible at the
8459 * platform's infinity, so have to handle it specially */
8460 if (UNLIKELY(end == UV_MAX && ELEMENT_RANGE_MATCHES_INVLIST(len-1)))
8462 const STRLEN offset = (STRLEN)(end - start);
8463 swatch[offset >> 3] |= 1 << (offset & 7);
8468 /* Advance to the next range, which will be for code points not in the
8477 Perl__invlist_union_maybe_complement_2nd(pTHX_ SV* const a, SV* const b,
8478 const bool complement_b, SV** output)
8480 /* Take the union of two inversion lists and point <output> to it. *output
8481 * SHOULD BE DEFINED upon input, and if it points to one of the two lists,
8482 * the reference count to that list will be decremented if not already a
8483 * temporary (mortal); otherwise *output will be made correspondingly
8484 * mortal. The first list, <a>, may be NULL, in which case a copy of the
8485 * second list is returned. If <complement_b> is TRUE, the union is taken
8486 * of the complement (inversion) of <b> instead of b itself.
8488 * The basis for this comes from "Unicode Demystified" Chapter 13 by
8489 * Richard Gillam, published by Addison-Wesley, and explained at some
8490 * length there. The preface says to incorporate its examples into your
8491 * code at your own risk.
8493 * The algorithm is like a merge sort.
8495 * XXX A potential performance improvement is to keep track as we go along
8496 * if only one of the inputs contributes to the result, meaning the other
8497 * is a subset of that one. In that case, we can skip the final copy and
8498 * return the larger of the input lists, but then outside code might need
8499 * to keep track of whether to free the input list or not */
8501 const UV* array_a; /* a's array */
8503 UV len_a; /* length of a's array */
8506 SV* u; /* the resulting union */
8510 UV i_a = 0; /* current index into a's array */
8514 /* running count, as explained in the algorithm source book; items are
8515 * stopped accumulating and are output when the count changes to/from 0.
8516 * The count is incremented when we start a range that's in the set, and
8517 * decremented when we start a range that's not in the set. So its range
8518 * is 0 to 2. Only when the count is zero is something not in the set.
8522 PERL_ARGS_ASSERT__INVLIST_UNION_MAYBE_COMPLEMENT_2ND;
8525 /* If either one is empty, the union is the other one */
8526 if (a == NULL || ((len_a = _invlist_len(a)) == 0)) {
8527 bool make_temp = FALSE; /* Should we mortalize the result? */
8531 if (! (make_temp = cBOOL(SvTEMP(a)))) {
8537 *output = invlist_clone(b);
8539 _invlist_invert(*output);
8541 } /* else *output already = b; */
8544 sv_2mortal(*output);
8548 else if ((len_b = _invlist_len(b)) == 0) {
8549 bool make_temp = FALSE;
8551 if (! (make_temp = cBOOL(SvTEMP(b)))) {
8556 /* The complement of an empty list is a list that has everything in it,
8557 * so the union with <a> includes everything too */
8560 if (! (make_temp = cBOOL(SvTEMP(a)))) {
8564 *output = _new_invlist(1);
8565 _append_range_to_invlist(*output, 0, UV_MAX);
8567 else if (*output != a) {
8568 *output = invlist_clone(a);
8570 /* else *output already = a; */
8573 sv_2mortal(*output);
8578 /* Here both lists exist and are non-empty */
8579 array_a = invlist_array(a);
8580 array_b = invlist_array(b);
8582 /* If are to take the union of 'a' with the complement of b, set it
8583 * up so are looking at b's complement. */
8586 /* To complement, we invert: if the first element is 0, remove it. To
8587 * do this, we just pretend the array starts one later */
8588 if (array_b[0] == 0) {
8594 /* But if the first element is not zero, we pretend the list starts
8595 * at the 0 that is always stored immediately before the array. */
8601 /* Size the union for the worst case: that the sets are completely
8603 u = _new_invlist(len_a + len_b);
8605 /* Will contain U+0000 if either component does */
8606 array_u = _invlist_array_init(u, (len_a > 0 && array_a[0] == 0)
8607 || (len_b > 0 && array_b[0] == 0));
8609 /* Go through each list item by item, stopping when exhausted one of
8611 while (i_a < len_a && i_b < len_b) {
8612 UV cp; /* The element to potentially add to the union's array */
8613 bool cp_in_set; /* is it in the the input list's set or not */
8615 /* We need to take one or the other of the two inputs for the union.
8616 * Since we are merging two sorted lists, we take the smaller of the
8617 * next items. In case of a tie, we take the one that is in its set
8618 * first. If we took one not in the set first, it would decrement the
8619 * count, possibly to 0 which would cause it to be output as ending the
8620 * range, and the next time through we would take the same number, and
8621 * output it again as beginning the next range. By doing it the
8622 * opposite way, there is no possibility that the count will be
8623 * momentarily decremented to 0, and thus the two adjoining ranges will
8624 * be seamlessly merged. (In a tie and both are in the set or both not
8625 * in the set, it doesn't matter which we take first.) */
8626 if (array_a[i_a] < array_b[i_b]
8627 || (array_a[i_a] == array_b[i_b]
8628 && ELEMENT_RANGE_MATCHES_INVLIST(i_a)))
8630 cp_in_set = ELEMENT_RANGE_MATCHES_INVLIST(i_a);
8634 cp_in_set = ELEMENT_RANGE_MATCHES_INVLIST(i_b);
8635 cp = array_b[i_b++];
8638 /* Here, have chosen which of the two inputs to look at. Only output
8639 * if the running count changes to/from 0, which marks the
8640 * beginning/end of a range in that's in the set */
8643 array_u[i_u++] = cp;
8650 array_u[i_u++] = cp;
8655 /* Here, we are finished going through at least one of the lists, which
8656 * means there is something remaining in at most one. We check if the list
8657 * that hasn't been exhausted is positioned such that we are in the middle
8658 * of a range in its set or not. (i_a and i_b point to the element beyond
8659 * the one we care about.) If in the set, we decrement 'count'; if 0, there
8660 * is potentially more to output.
8661 * There are four cases:
8662 * 1) Both weren't in their sets, count is 0, and remains 0. What's left
8663 * in the union is entirely from the non-exhausted set.
8664 * 2) Both were in their sets, count is 2. Nothing further should
8665 * be output, as everything that remains will be in the exhausted
8666 * list's set, hence in the union; decrementing to 1 but not 0 insures
8668 * 3) the exhausted was in its set, non-exhausted isn't, count is 1.
8669 * Nothing further should be output because the union includes
8670 * everything from the exhausted set. Not decrementing ensures that.
8671 * 4) the exhausted wasn't in its set, non-exhausted is, count is 1;
8672 * decrementing to 0 insures that we look at the remainder of the
8673 * non-exhausted set */
8674 if ((i_a != len_a && PREV_RANGE_MATCHES_INVLIST(i_a))
8675 || (i_b != len_b && PREV_RANGE_MATCHES_INVLIST(i_b)))
8680 /* The final length is what we've output so far, plus what else is about to
8681 * be output. (If 'count' is non-zero, then the input list we exhausted
8682 * has everything remaining up to the machine's limit in its set, and hence
8683 * in the union, so there will be no further output. */
8686 /* At most one of the subexpressions will be non-zero */
8687 len_u += (len_a - i_a) + (len_b - i_b);
8690 /* Set result to final length, which can change the pointer to array_u, so
8692 if (len_u != _invlist_len(u)) {
8693 invlist_set_len(u, len_u, *get_invlist_offset_addr(u));
8695 array_u = invlist_array(u);
8698 /* When 'count' is 0, the list that was exhausted (if one was shorter than
8699 * the other) ended with everything above it not in its set. That means
8700 * that the remaining part of the union is precisely the same as the
8701 * non-exhausted list, so can just copy it unchanged. (If both list were
8702 * exhausted at the same time, then the operations below will be both 0.)
8705 IV copy_count; /* At most one will have a non-zero copy count */
8706 if ((copy_count = len_a - i_a) > 0) {
8707 Copy(array_a + i_a, array_u + i_u, copy_count, UV);
8709 else if ((copy_count = len_b - i_b) > 0) {
8710 Copy(array_b + i_b, array_u + i_u, copy_count, UV);
8714 /* We may be removing a reference to one of the inputs. If so, the output
8715 * is made mortal if the input was. (Mortal SVs shouldn't have their ref
8716 * count decremented) */
8717 if (a == *output || b == *output) {
8718 assert(! invlist_is_iterating(*output));
8719 if ((SvTEMP(*output))) {
8723 SvREFCNT_dec_NN(*output);
8733 Perl__invlist_intersection_maybe_complement_2nd(pTHX_ SV* const a, SV* const b,
8734 const bool complement_b, SV** i)
8736 /* Take the intersection of two inversion lists and point <i> to it. *i
8737 * SHOULD BE DEFINED upon input, and if it points to one of the two lists,
8738 * the reference count to that list will be decremented if not already a
8739 * temporary (mortal); otherwise *i will be made correspondingly mortal.
8740 * The first list, <a>, may be NULL, in which case an empty list is
8741 * returned. If <complement_b> is TRUE, the result will be the
8742 * intersection of <a> and the complement (or inversion) of <b> instead of
8745 * The basis for this comes from "Unicode Demystified" Chapter 13 by
8746 * Richard Gillam, published by Addison-Wesley, and explained at some
8747 * length there. The preface says to incorporate its examples into your
8748 * code at your own risk. In fact, it had bugs
8750 * The algorithm is like a merge sort, and is essentially the same as the
8754 const UV* array_a; /* a's array */
8756 UV len_a; /* length of a's array */
8759 SV* r; /* the resulting intersection */
8763 UV i_a = 0; /* current index into a's array */
8767 /* running count, as explained in the algorithm source book; items are
8768 * stopped accumulating and are output when the count changes to/from 2.
8769 * The count is incremented when we start a range that's in the set, and
8770 * decremented when we start a range that's not in the set. So its range
8771 * is 0 to 2. Only when the count is 2 is something in the intersection.
8775 PERL_ARGS_ASSERT__INVLIST_INTERSECTION_MAYBE_COMPLEMENT_2ND;
8778 /* Special case if either one is empty */
8779 len_a = (a == NULL) ? 0 : _invlist_len(a);
8780 if ((len_a == 0) || ((len_b = _invlist_len(b)) == 0)) {
8781 bool make_temp = FALSE;
8783 if (len_a != 0 && complement_b) {
8785 /* Here, 'a' is not empty, therefore from the above 'if', 'b' must
8786 * be empty. Here, also we are using 'b's complement, which hence
8787 * must be every possible code point. Thus the intersection is
8791 if (! (make_temp = cBOOL(SvTEMP(b)))) {
8796 *i = invlist_clone(a);
8798 /* else *i is already 'a' */
8806 /* Here, 'a' or 'b' is empty and not using the complement of 'b'. The
8807 * intersection must be empty */
8809 if (! (make_temp = cBOOL(SvTEMP(a)))) {
8814 if (! (make_temp = cBOOL(SvTEMP(b)))) {
8818 *i = _new_invlist(0);
8826 /* Here both lists exist and are non-empty */
8827 array_a = invlist_array(a);
8828 array_b = invlist_array(b);
8830 /* If are to take the intersection of 'a' with the complement of b, set it
8831 * up so are looking at b's complement. */
8834 /* To complement, we invert: if the first element is 0, remove it. To
8835 * do this, we just pretend the array starts one later */
8836 if (array_b[0] == 0) {
8842 /* But if the first element is not zero, we pretend the list starts
8843 * at the 0 that is always stored immediately before the array. */
8849 /* Size the intersection for the worst case: that the intersection ends up
8850 * fragmenting everything to be completely disjoint */
8851 r= _new_invlist(len_a + len_b);
8853 /* Will contain U+0000 iff both components do */
8854 array_r = _invlist_array_init(r, len_a > 0 && array_a[0] == 0
8855 && len_b > 0 && array_b[0] == 0);
8857 /* Go through each list item by item, stopping when exhausted one of
8859 while (i_a < len_a && i_b < len_b) {
8860 UV cp; /* The element to potentially add to the intersection's
8862 bool cp_in_set; /* Is it in the input list's set or not */
8864 /* We need to take one or the other of the two inputs for the
8865 * intersection. Since we are merging two sorted lists, we take the
8866 * smaller of the next items. In case of a tie, we take the one that
8867 * is not in its set first (a difference from the union algorithm). If
8868 * we took one in the set first, it would increment the count, possibly
8869 * to 2 which would cause it to be output as starting a range in the
8870 * intersection, and the next time through we would take that same
8871 * number, and output it again as ending the set. By doing it the
8872 * opposite of this, there is no possibility that the count will be
8873 * momentarily incremented to 2. (In a tie and both are in the set or
8874 * both not in the set, it doesn't matter which we take first.) */
8875 if (array_a[i_a] < array_b[i_b]
8876 || (array_a[i_a] == array_b[i_b]
8877 && ! ELEMENT_RANGE_MATCHES_INVLIST(i_a)))
8879 cp_in_set = ELEMENT_RANGE_MATCHES_INVLIST(i_a);
8883 cp_in_set = ELEMENT_RANGE_MATCHES_INVLIST(i_b);
8887 /* Here, have chosen which of the two inputs to look at. Only output
8888 * if the running count changes to/from 2, which marks the
8889 * beginning/end of a range that's in the intersection */
8893 array_r[i_r++] = cp;
8898 array_r[i_r++] = cp;
8904 /* Here, we are finished going through at least one of the lists, which
8905 * means there is something remaining in at most one. We check if the list
8906 * that has been exhausted is positioned such that we are in the middle
8907 * of a range in its set or not. (i_a and i_b point to elements 1 beyond
8908 * the ones we care about.) There are four cases:
8909 * 1) Both weren't in their sets, count is 0, and remains 0. There's
8910 * nothing left in the intersection.
8911 * 2) Both were in their sets, count is 2 and perhaps is incremented to
8912 * above 2. What should be output is exactly that which is in the
8913 * non-exhausted set, as everything it has is also in the intersection
8914 * set, and everything it doesn't have can't be in the intersection
8915 * 3) The exhausted was in its set, non-exhausted isn't, count is 1, and
8916 * gets incremented to 2. Like the previous case, the intersection is
8917 * everything that remains in the non-exhausted set.
8918 * 4) the exhausted wasn't in its set, non-exhausted is, count is 1, and
8919 * remains 1. And the intersection has nothing more. */
8920 if ((i_a == len_a && PREV_RANGE_MATCHES_INVLIST(i_a))
8921 || (i_b == len_b && PREV_RANGE_MATCHES_INVLIST(i_b)))
8926 /* The final length is what we've output so far plus what else is in the
8927 * intersection. At most one of the subexpressions below will be non-zero
8931 len_r += (len_a - i_a) + (len_b - i_b);
8934 /* Set result to final length, which can change the pointer to array_r, so
8936 if (len_r != _invlist_len(r)) {
8937 invlist_set_len(r, len_r, *get_invlist_offset_addr(r));
8939 array_r = invlist_array(r);
8942 /* Finish outputting any remaining */
8943 if (count >= 2) { /* At most one will have a non-zero copy count */
8945 if ((copy_count = len_a - i_a) > 0) {
8946 Copy(array_a + i_a, array_r + i_r, copy_count, UV);
8948 else if ((copy_count = len_b - i_b) > 0) {
8949 Copy(array_b + i_b, array_r + i_r, copy_count, UV);
8953 /* We may be removing a reference to one of the inputs. If so, the output
8954 * is made mortal if the input was. (Mortal SVs shouldn't have their ref
8955 * count decremented) */
8956 if (a == *i || b == *i) {
8957 assert(! invlist_is_iterating(*i));
8962 SvREFCNT_dec_NN(*i);
8972 Perl__add_range_to_invlist(pTHX_ SV* invlist, const UV start, const UV end)
8974 /* Add the range from 'start' to 'end' inclusive to the inversion list's
8975 * set. A pointer to the inversion list is returned. This may actually be
8976 * a new list, in which case the passed in one has been destroyed. The
8977 * passed-in inversion list can be NULL, in which case a new one is created
8978 * with just the one range in it */
8983 if (invlist == NULL) {
8984 invlist = _new_invlist(2);
8988 len = _invlist_len(invlist);
8991 /* If comes after the final entry actually in the list, can just append it
8994 || (! ELEMENT_RANGE_MATCHES_INVLIST(len - 1)
8995 && start >= invlist_array(invlist)[len - 1]))
8997 _append_range_to_invlist(invlist, start, end);
9001 /* Here, can't just append things, create and return a new inversion list
9002 * which is the union of this range and the existing inversion list */
9003 range_invlist = _new_invlist(2);
9004 _append_range_to_invlist(range_invlist, start, end);
9006 _invlist_union(invlist, range_invlist, &invlist);
9008 /* The temporary can be freed */
9009 SvREFCNT_dec_NN(range_invlist);
9015 Perl__setup_canned_invlist(pTHX_ const STRLEN size, const UV element0,
9016 UV** other_elements_ptr)
9018 /* Create and return an inversion list whose contents are to be populated
9019 * by the caller. The caller gives the number of elements (in 'size') and
9020 * the very first element ('element0'). This function will set
9021 * '*other_elements_ptr' to an array of UVs, where the remaining elements
9024 * Obviously there is some trust involved that the caller will properly
9025 * fill in the other elements of the array.
9027 * (The first element needs to be passed in, as the underlying code does
9028 * things differently depending on whether it is zero or non-zero) */
9030 SV* invlist = _new_invlist(size);
9033 PERL_ARGS_ASSERT__SETUP_CANNED_INVLIST;
9035 _append_range_to_invlist(invlist, element0, element0);
9036 offset = *get_invlist_offset_addr(invlist);
9038 invlist_set_len(invlist, size, offset);
9039 *other_elements_ptr = invlist_array(invlist) + 1;
9045 PERL_STATIC_INLINE SV*
9046 S_add_cp_to_invlist(pTHX_ SV* invlist, const UV cp) {
9047 return _add_range_to_invlist(invlist, cp, cp);
9050 #ifndef PERL_IN_XSUB_RE
9052 Perl__invlist_invert(pTHX_ SV* const invlist)
9054 /* Complement the input inversion list. This adds a 0 if the list didn't
9055 * have a zero; removes it otherwise. As described above, the data
9056 * structure is set up so that this is very efficient */
9058 PERL_ARGS_ASSERT__INVLIST_INVERT;
9060 assert(! invlist_is_iterating(invlist));
9062 /* The inverse of matching nothing is matching everything */
9063 if (_invlist_len(invlist) == 0) {
9064 _append_range_to_invlist(invlist, 0, UV_MAX);
9068 *get_invlist_offset_addr(invlist) = ! *get_invlist_offset_addr(invlist);
9073 PERL_STATIC_INLINE SV*
9074 S_invlist_clone(pTHX_ SV* const invlist)
9077 /* Return a new inversion list that is a copy of the input one, which is
9078 * unchanged. The new list will not be mortal even if the old one was. */
9080 /* Need to allocate extra space to accommodate Perl's addition of a
9081 * trailing NUL to SvPV's, since it thinks they are always strings */
9082 SV* new_invlist = _new_invlist(_invlist_len(invlist) + 1);
9083 STRLEN physical_length = SvCUR(invlist);
9084 bool offset = *(get_invlist_offset_addr(invlist));
9086 PERL_ARGS_ASSERT_INVLIST_CLONE;
9088 *(get_invlist_offset_addr(new_invlist)) = offset;
9089 invlist_set_len(new_invlist, _invlist_len(invlist), offset);
9090 Copy(SvPVX(invlist), SvPVX(new_invlist), physical_length, char);
9095 PERL_STATIC_INLINE STRLEN*
9096 S_get_invlist_iter_addr(SV* invlist)
9098 /* Return the address of the UV that contains the current iteration
9101 PERL_ARGS_ASSERT_GET_INVLIST_ITER_ADDR;
9103 assert(SvTYPE(invlist) == SVt_INVLIST);
9105 return &(((XINVLIST*) SvANY(invlist))->iterator);
9108 PERL_STATIC_INLINE void
9109 S_invlist_iterinit(SV* invlist) /* Initialize iterator for invlist */
9111 PERL_ARGS_ASSERT_INVLIST_ITERINIT;
9113 *get_invlist_iter_addr(invlist) = 0;
9116 PERL_STATIC_INLINE void
9117 S_invlist_iterfinish(SV* invlist)
9119 /* Terminate iterator for invlist. This is to catch development errors.
9120 * Any iteration that is interrupted before completed should call this
9121 * function. Functions that add code points anywhere else but to the end
9122 * of an inversion list assert that they are not in the middle of an
9123 * iteration. If they were, the addition would make the iteration
9124 * problematical: if the iteration hadn't reached the place where things
9125 * were being added, it would be ok */
9127 PERL_ARGS_ASSERT_INVLIST_ITERFINISH;
9129 *get_invlist_iter_addr(invlist) = (STRLEN) UV_MAX;
9133 S_invlist_iternext(SV* invlist, UV* start, UV* end)
9135 /* An C<invlist_iterinit> call on <invlist> must be used to set this up.
9136 * This call sets in <*start> and <*end>, the next range in <invlist>.
9137 * Returns <TRUE> if successful and the next call will return the next
9138 * range; <FALSE> if was already at the end of the list. If the latter,
9139 * <*start> and <*end> are unchanged, and the next call to this function
9140 * will start over at the beginning of the list */
9142 STRLEN* pos = get_invlist_iter_addr(invlist);
9143 UV len = _invlist_len(invlist);
9146 PERL_ARGS_ASSERT_INVLIST_ITERNEXT;
9149 *pos = (STRLEN) UV_MAX; /* Force iterinit() to be required next time */
9153 array = invlist_array(invlist);
9155 *start = array[(*pos)++];
9161 *end = array[(*pos)++] - 1;
9167 PERL_STATIC_INLINE UV
9168 S_invlist_highest(SV* const invlist)
9170 /* Returns the highest code point that matches an inversion list. This API
9171 * has an ambiguity, as it returns 0 under either the highest is actually
9172 * 0, or if the list is empty. If this distinction matters to you, check
9173 * for emptiness before calling this function */
9175 UV len = _invlist_len(invlist);
9178 PERL_ARGS_ASSERT_INVLIST_HIGHEST;
9184 array = invlist_array(invlist);
9186 /* The last element in the array in the inversion list always starts a
9187 * range that goes to infinity. That range may be for code points that are
9188 * matched in the inversion list, or it may be for ones that aren't
9189 * matched. In the latter case, the highest code point in the set is one
9190 * less than the beginning of this range; otherwise it is the final element
9191 * of this range: infinity */
9192 return (ELEMENT_RANGE_MATCHES_INVLIST(len - 1))
9194 : array[len - 1] - 1;
9197 #ifndef PERL_IN_XSUB_RE
9199 Perl__invlist_contents(pTHX_ SV* const invlist)
9201 /* Get the contents of an inversion list into a string SV so that they can
9202 * be printed out. It uses the format traditionally done for debug tracing
9206 SV* output = newSVpvs("\n");
9208 PERL_ARGS_ASSERT__INVLIST_CONTENTS;
9210 assert(! invlist_is_iterating(invlist));
9212 invlist_iterinit(invlist);
9213 while (invlist_iternext(invlist, &start, &end)) {
9214 if (end == UV_MAX) {
9215 Perl_sv_catpvf(aTHX_ output, "%04"UVXf"\tINFINITY\n", start);
9217 else if (end != start) {
9218 Perl_sv_catpvf(aTHX_ output, "%04"UVXf"\t%04"UVXf"\n",
9222 Perl_sv_catpvf(aTHX_ output, "%04"UVXf"\n", start);
9230 #ifndef PERL_IN_XSUB_RE
9232 Perl__invlist_dump(pTHX_ PerlIO *file, I32 level,
9233 const char * const indent, SV* const invlist)
9235 /* Designed to be called only by do_sv_dump(). Dumps out the ranges of the
9236 * inversion list 'invlist' to 'file' at 'level' Each line is prefixed by
9237 * the string 'indent'. The output looks like this:
9238 [0] 0x000A .. 0x000D
9240 [4] 0x2028 .. 0x2029
9241 [6] 0x3104 .. INFINITY
9242 * This means that the first range of code points matched by the list are
9243 * 0xA through 0xD; the second range contains only the single code point
9244 * 0x85, etc. An inversion list is an array of UVs. Two array elements
9245 * are used to define each range (except if the final range extends to
9246 * infinity, only a single element is needed). The array index of the
9247 * first element for the corresponding range is given in brackets. */
9252 PERL_ARGS_ASSERT__INVLIST_DUMP;
9254 if (invlist_is_iterating(invlist)) {
9255 Perl_dump_indent(aTHX_ level, file,
9256 "%sCan't dump inversion list because is in middle of iterating\n",
9261 invlist_iterinit(invlist);
9262 while (invlist_iternext(invlist, &start, &end)) {
9263 if (end == UV_MAX) {
9264 Perl_dump_indent(aTHX_ level, file,
9265 "%s[%"UVuf"] 0x%04"UVXf" .. INFINITY\n",
9266 indent, (UV)count, start);
9268 else if (end != start) {
9269 Perl_dump_indent(aTHX_ level, file,
9270 "%s[%"UVuf"] 0x%04"UVXf" .. 0x%04"UVXf"\n",
9271 indent, (UV)count, start, end);
9274 Perl_dump_indent(aTHX_ level, file, "%s[%"UVuf"] 0x%04"UVXf"\n",
9275 indent, (UV)count, start);
9282 Perl__load_PL_utf8_foldclosures (pTHX)
9284 assert(! PL_utf8_foldclosures);
9286 /* If the folds haven't been read in, call a fold function
9288 if (! PL_utf8_tofold) {
9289 U8 dummy[UTF8_MAXBYTES_CASE+1];
9291 /* This string is just a short named one above \xff */
9292 to_utf8_fold((U8*) HYPHEN_UTF8, dummy, NULL);
9293 assert(PL_utf8_tofold); /* Verify that worked */
9295 PL_utf8_foldclosures = _swash_inversion_hash(PL_utf8_tofold);
9299 #ifdef PERL_ARGS_ASSERT__INVLISTEQ
9301 S__invlistEQ(pTHX_ SV* const a, SV* const b, const bool complement_b)
9303 /* Return a boolean as to if the two passed in inversion lists are
9304 * identical. The final argument, if TRUE, says to take the complement of
9305 * the second inversion list before doing the comparison */
9307 const UV* array_a = invlist_array(a);
9308 const UV* array_b = invlist_array(b);
9309 UV len_a = _invlist_len(a);
9310 UV len_b = _invlist_len(b);
9312 UV i = 0; /* current index into the arrays */
9313 bool retval = TRUE; /* Assume are identical until proven otherwise */
9315 PERL_ARGS_ASSERT__INVLISTEQ;
9317 /* If are to compare 'a' with the complement of b, set it
9318 * up so are looking at b's complement. */
9321 /* The complement of nothing is everything, so <a> would have to have
9322 * just one element, starting at zero (ending at infinity) */
9324 return (len_a == 1 && array_a[0] == 0);
9326 else if (array_b[0] == 0) {
9328 /* Otherwise, to complement, we invert. Here, the first element is
9329 * 0, just remove it. To do this, we just pretend the array starts
9337 /* But if the first element is not zero, we pretend the list starts
9338 * at the 0 that is always stored immediately before the array. */
9344 /* Make sure that the lengths are the same, as well as the final element
9345 * before looping through the remainder. (Thus we test the length, final,
9346 * and first elements right off the bat) */
9347 if (len_a != len_b || array_a[len_a-1] != array_b[len_a-1]) {
9350 else for (i = 0; i < len_a - 1; i++) {
9351 if (array_a[i] != array_b[i]) {
9362 * As best we can, determine the characters that can match the start of
9363 * the given EXACTF-ish node.
9365 * Returns the invlist as a new SV*; it is the caller's responsibility to
9366 * call SvREFCNT_dec() when done with it.
9369 S__make_exactf_invlist(pTHX_ RExC_state_t *pRExC_state, regnode *node)
9371 const U8 * s = (U8*)STRING(node);
9372 SSize_t bytelen = STR_LEN(node);
9374 /* Start out big enough for 2 separate code points */
9375 SV* invlist = _new_invlist(4);
9377 PERL_ARGS_ASSERT__MAKE_EXACTF_INVLIST;
9382 /* We punt and assume can match anything if the node begins
9383 * with a multi-character fold. Things are complicated. For
9384 * example, /ffi/i could match any of:
9385 * "\N{LATIN SMALL LIGATURE FFI}"
9386 * "\N{LATIN SMALL LIGATURE FF}I"
9387 * "F\N{LATIN SMALL LIGATURE FI}"
9388 * plus several other things; and making sure we have all the
9389 * possibilities is hard. */
9390 if (is_MULTI_CHAR_FOLD_latin1_safe(s, s + bytelen)) {
9391 invlist = _add_range_to_invlist(invlist, 0, UV_MAX);
9394 /* Any Latin1 range character can potentially match any
9395 * other depending on the locale */
9396 if (OP(node) == EXACTFL) {
9397 _invlist_union(invlist, PL_Latin1, &invlist);
9400 /* But otherwise, it matches at least itself. We can
9401 * quickly tell if it has a distinct fold, and if so,
9402 * it matches that as well */
9403 invlist = add_cp_to_invlist(invlist, uc);
9404 if (IS_IN_SOME_FOLD_L1(uc))
9405 invlist = add_cp_to_invlist(invlist, PL_fold_latin1[uc]);
9408 /* Some characters match above-Latin1 ones under /i. This
9409 * is true of EXACTFL ones when the locale is UTF-8 */
9410 if (HAS_NONLATIN1_SIMPLE_FOLD_CLOSURE(uc)
9411 && (! isASCII(uc) || (OP(node) != EXACTFA
9412 && OP(node) != EXACTFA_NO_TRIE)))
9414 add_above_Latin1_folds(pRExC_state, (U8) uc, &invlist);
9418 else { /* Pattern is UTF-8 */
9419 U8 folded[UTF8_MAX_FOLD_CHAR_EXPAND * UTF8_MAXBYTES_CASE + 1] = { '\0' };
9420 STRLEN foldlen = UTF8SKIP(s);
9421 const U8* e = s + bytelen;
9424 uc = utf8_to_uvchr_buf(s, s + bytelen, NULL);
9426 /* The only code points that aren't folded in a UTF EXACTFish
9427 * node are are the problematic ones in EXACTFL nodes */
9428 if (OP(node) == EXACTFL && is_PROBLEMATIC_LOCALE_FOLDEDS_START_cp(uc)) {
9429 /* We need to check for the possibility that this EXACTFL
9430 * node begins with a multi-char fold. Therefore we fold
9431 * the first few characters of it so that we can make that
9436 for (i = 0; i < UTF8_MAX_FOLD_CHAR_EXPAND && s < e; i++) {
9438 *(d++) = (U8) toFOLD(*s);
9443 to_utf8_fold(s, d, &len);
9449 /* And set up so the code below that looks in this folded
9450 * buffer instead of the node's string */
9452 foldlen = UTF8SKIP(folded);
9456 /* When we reach here 's' points to the fold of the first
9457 * character(s) of the node; and 'e' points to far enough along
9458 * the folded string to be just past any possible multi-char
9459 * fold. 'foldlen' is the length in bytes of the first
9462 * Unlike the non-UTF-8 case, the macro for determining if a
9463 * string is a multi-char fold requires all the characters to
9464 * already be folded. This is because of all the complications
9465 * if not. Note that they are folded anyway, except in EXACTFL
9466 * nodes. Like the non-UTF case above, we punt if the node
9467 * begins with a multi-char fold */
9469 if (is_MULTI_CHAR_FOLD_utf8_safe(s, e)) {
9470 invlist = _add_range_to_invlist(invlist, 0, UV_MAX);
9472 else { /* Single char fold */
9474 /* It matches all the things that fold to it, which are
9475 * found in PL_utf8_foldclosures (including itself) */
9476 invlist = add_cp_to_invlist(invlist, uc);
9477 if (! PL_utf8_foldclosures)
9478 _load_PL_utf8_foldclosures();
9479 if ((listp = hv_fetch(PL_utf8_foldclosures,
9480 (char *) s, foldlen, FALSE)))
9482 AV* list = (AV*) *listp;
9484 for (k = 0; k <= av_tindex(list); k++) {
9485 SV** c_p = av_fetch(list, k, FALSE);
9491 /* /aa doesn't allow folds between ASCII and non- */
9492 if ((OP(node) == EXACTFA || OP(node) == EXACTFA_NO_TRIE)
9493 && isASCII(c) != isASCII(uc))
9498 invlist = add_cp_to_invlist(invlist, c);
9507 #undef HEADER_LENGTH
9508 #undef TO_INTERNAL_SIZE
9509 #undef FROM_INTERNAL_SIZE
9510 #undef INVLIST_VERSION_ID
9512 /* End of inversion list object */
9515 S_parse_lparen_question_flags(pTHX_ RExC_state_t *pRExC_state)
9517 /* This parses the flags that are in either the '(?foo)' or '(?foo:bar)'
9518 * constructs, and updates RExC_flags with them. On input, RExC_parse
9519 * should point to the first flag; it is updated on output to point to the
9520 * final ')' or ':'. There needs to be at least one flag, or this will
9523 /* for (?g), (?gc), and (?o) warnings; warning
9524 about (?c) will warn about (?g) -- japhy */
9526 #define WASTED_O 0x01
9527 #define WASTED_G 0x02
9528 #define WASTED_C 0x04
9529 #define WASTED_GC (WASTED_G|WASTED_C)
9530 I32 wastedflags = 0x00;
9531 U32 posflags = 0, negflags = 0;
9532 U32 *flagsp = &posflags;
9533 char has_charset_modifier = '\0';
9535 bool has_use_defaults = FALSE;
9536 const char* const seqstart = RExC_parse - 1; /* Point to the '?' */
9537 int x_mod_count = 0;
9539 PERL_ARGS_ASSERT_PARSE_LPAREN_QUESTION_FLAGS;
9541 /* '^' as an initial flag sets certain defaults */
9542 if (UCHARAT(RExC_parse) == '^') {
9544 has_use_defaults = TRUE;
9545 STD_PMMOD_FLAGS_CLEAR(&RExC_flags);
9546 set_regex_charset(&RExC_flags, (RExC_utf8 || RExC_uni_semantics)
9547 ? REGEX_UNICODE_CHARSET
9548 : REGEX_DEPENDS_CHARSET);
9551 cs = get_regex_charset(RExC_flags);
9552 if (cs == REGEX_DEPENDS_CHARSET
9553 && (RExC_utf8 || RExC_uni_semantics))
9555 cs = REGEX_UNICODE_CHARSET;
9558 while (*RExC_parse) {
9559 /* && strchr("iogcmsx", *RExC_parse) */
9560 /* (?g), (?gc) and (?o) are useless here
9561 and must be globally applied -- japhy */
9562 switch (*RExC_parse) {
9564 /* Code for the imsxn flags */
9565 CASE_STD_PMMOD_FLAGS_PARSE_SET(flagsp, x_mod_count);
9567 case LOCALE_PAT_MOD:
9568 if (has_charset_modifier) {
9569 goto excess_modifier;
9571 else if (flagsp == &negflags) {
9574 cs = REGEX_LOCALE_CHARSET;
9575 has_charset_modifier = LOCALE_PAT_MOD;
9577 case UNICODE_PAT_MOD:
9578 if (has_charset_modifier) {
9579 goto excess_modifier;
9581 else if (flagsp == &negflags) {
9584 cs = REGEX_UNICODE_CHARSET;
9585 has_charset_modifier = UNICODE_PAT_MOD;
9587 case ASCII_RESTRICT_PAT_MOD:
9588 if (flagsp == &negflags) {
9591 if (has_charset_modifier) {
9592 if (cs != REGEX_ASCII_RESTRICTED_CHARSET) {
9593 goto excess_modifier;
9595 /* Doubled modifier implies more restricted */
9596 cs = REGEX_ASCII_MORE_RESTRICTED_CHARSET;
9599 cs = REGEX_ASCII_RESTRICTED_CHARSET;
9601 has_charset_modifier = ASCII_RESTRICT_PAT_MOD;
9603 case DEPENDS_PAT_MOD:
9604 if (has_use_defaults) {
9605 goto fail_modifiers;
9607 else if (flagsp == &negflags) {
9610 else if (has_charset_modifier) {
9611 goto excess_modifier;
9614 /* The dual charset means unicode semantics if the
9615 * pattern (or target, not known until runtime) are
9616 * utf8, or something in the pattern indicates unicode
9618 cs = (RExC_utf8 || RExC_uni_semantics)
9619 ? REGEX_UNICODE_CHARSET
9620 : REGEX_DEPENDS_CHARSET;
9621 has_charset_modifier = DEPENDS_PAT_MOD;
9625 if (has_charset_modifier == ASCII_RESTRICT_PAT_MOD) {
9626 vFAIL2("Regexp modifier \"%c\" may appear a maximum of twice", ASCII_RESTRICT_PAT_MOD);
9628 else if (has_charset_modifier == *(RExC_parse - 1)) {
9629 vFAIL2("Regexp modifier \"%c\" may not appear twice",
9633 vFAIL3("Regexp modifiers \"%c\" and \"%c\" are mutually exclusive", has_charset_modifier, *(RExC_parse - 1));
9635 NOT_REACHED; /*NOTREACHED*/
9638 vFAIL2("Regexp modifier \"%c\" may not appear after the \"-\"",
9640 NOT_REACHED; /*NOTREACHED*/
9641 case ONCE_PAT_MOD: /* 'o' */
9642 case GLOBAL_PAT_MOD: /* 'g' */
9643 if (PASS2 && ckWARN(WARN_REGEXP)) {
9644 const I32 wflagbit = *RExC_parse == 'o'
9647 if (! (wastedflags & wflagbit) ) {
9648 wastedflags |= wflagbit;
9649 /* diag_listed_as: Useless (?-%s) - don't use /%s modifier in regex; marked by <-- HERE in m/%s/ */
9652 "Useless (%s%c) - %suse /%c modifier",
9653 flagsp == &negflags ? "?-" : "?",
9655 flagsp == &negflags ? "don't " : "",
9662 case CONTINUE_PAT_MOD: /* 'c' */
9663 if (PASS2 && ckWARN(WARN_REGEXP)) {
9664 if (! (wastedflags & WASTED_C) ) {
9665 wastedflags |= WASTED_GC;
9666 /* diag_listed_as: Useless (?-%s) - don't use /%s modifier in regex; marked by <-- HERE in m/%s/ */
9669 "Useless (%sc) - %suse /gc modifier",
9670 flagsp == &negflags ? "?-" : "?",
9671 flagsp == &negflags ? "don't " : ""
9676 case KEEPCOPY_PAT_MOD: /* 'p' */
9677 if (flagsp == &negflags) {
9679 ckWARNreg(RExC_parse + 1,"Useless use of (?-p)");
9681 *flagsp |= RXf_PMf_KEEPCOPY;
9685 /* A flag is a default iff it is following a minus, so
9686 * if there is a minus, it means will be trying to
9687 * re-specify a default which is an error */
9688 if (has_use_defaults || flagsp == &negflags) {
9689 goto fail_modifiers;
9692 wastedflags = 0; /* reset so (?g-c) warns twice */
9696 RExC_flags |= posflags;
9697 RExC_flags &= ~negflags;
9698 set_regex_charset(&RExC_flags, cs);
9699 if (RExC_flags & RXf_PMf_FOLD) {
9700 RExC_contains_i = 1;
9703 STD_PMMOD_FLAGS_PARSE_X_WARN(x_mod_count);
9709 RExC_parse += SKIP_IF_CHAR(RExC_parse);
9710 /* diag_listed_as: Sequence (?%s...) not recognized in regex; marked by <-- HERE in m/%s/ */
9711 vFAIL2utf8f("Sequence (%"UTF8f"...) not recognized",
9712 UTF8fARG(UTF, RExC_parse-seqstart, seqstart));
9713 NOT_REACHED; /*NOTREACHED*/
9720 STD_PMMOD_FLAGS_PARSE_X_WARN(x_mod_count);
9725 - reg - regular expression, i.e. main body or parenthesized thing
9727 * Caller must absorb opening parenthesis.
9729 * Combining parenthesis handling with the base level of regular expression
9730 * is a trifle forced, but the need to tie the tails of the branches to what
9731 * follows makes it hard to avoid.
9733 #define REGTAIL(x,y,z) regtail((x),(y),(z),depth+1)
9735 #define REGTAIL_STUDY(x,y,z) regtail_study((x),(y),(z),depth+1)
9737 #define REGTAIL_STUDY(x,y,z) regtail((x),(y),(z),depth+1)
9740 /* Returns NULL, setting *flagp to TRYAGAIN at the end of (?) that only sets
9741 flags. Returns NULL, setting *flagp to RESTART_UTF8 if the sizing scan
9742 needs to be restarted.
9743 Otherwise would only return NULL if regbranch() returns NULL, which
9746 S_reg(pTHX_ RExC_state_t *pRExC_state, I32 paren, I32 *flagp,U32 depth)
9747 /* paren: Parenthesized? 0=top; 1,2=inside '(': changed to letter.
9748 * 2 is like 1, but indicates that nextchar() has been called to advance
9749 * RExC_parse beyond the '('. Things like '(?' are indivisible tokens, and
9750 * this flag alerts us to the need to check for that */
9752 regnode *ret; /* Will be the head of the group. */
9755 regnode *ender = NULL;
9758 U32 oregflags = RExC_flags;
9759 bool have_branch = 0;
9761 I32 freeze_paren = 0;
9762 I32 after_freeze = 0;
9763 I32 num; /* numeric backreferences */
9765 char * parse_start = RExC_parse; /* MJD */
9766 char * const oregcomp_parse = RExC_parse;
9768 GET_RE_DEBUG_FLAGS_DECL;
9770 PERL_ARGS_ASSERT_REG;
9771 DEBUG_PARSE("reg ");
9773 *flagp = 0; /* Tentatively. */
9776 /* Make an OPEN node, if parenthesized. */
9779 /* Under /x, space and comments can be gobbled up between the '(' and
9780 * here (if paren ==2). The forms '(*VERB' and '(?...' disallow such
9781 * intervening space, as the sequence is a token, and a token should be
9783 bool has_intervening_patws = paren == 2 && *(RExC_parse - 1) != '(';
9785 if ( *RExC_parse == '*') { /* (*VERB:ARG) */
9786 char *start_verb = RExC_parse;
9787 STRLEN verb_len = 0;
9788 char *start_arg = NULL;
9789 unsigned char op = 0;
9791 int internal_argval = 0; /* internal_argval is only useful if
9794 if (has_intervening_patws) {
9796 vFAIL("In '(*VERB...)', the '(' and '*' must be adjacent");
9798 while ( *RExC_parse && *RExC_parse != ')' ) {
9799 if ( *RExC_parse == ':' ) {
9800 start_arg = RExC_parse + 1;
9806 verb_len = RExC_parse - start_verb;
9809 while ( *RExC_parse && *RExC_parse != ')' )
9811 if ( *RExC_parse != ')' )
9812 vFAIL("Unterminated verb pattern argument");
9813 if ( RExC_parse == start_arg )
9816 if ( *RExC_parse != ')' )
9817 vFAIL("Unterminated verb pattern");
9820 switch ( *start_verb ) {
9821 case 'A': /* (*ACCEPT) */
9822 if ( memEQs(start_verb,verb_len,"ACCEPT") ) {
9824 internal_argval = RExC_nestroot;
9827 case 'C': /* (*COMMIT) */
9828 if ( memEQs(start_verb,verb_len,"COMMIT") )
9831 case 'F': /* (*FAIL) */
9832 if ( verb_len==1 || memEQs(start_verb,verb_len,"FAIL") ) {
9837 case ':': /* (*:NAME) */
9838 case 'M': /* (*MARK:NAME) */
9839 if ( verb_len==0 || memEQs(start_verb,verb_len,"MARK") ) {
9844 case 'P': /* (*PRUNE) */
9845 if ( memEQs(start_verb,verb_len,"PRUNE") )
9848 case 'S': /* (*SKIP) */
9849 if ( memEQs(start_verb,verb_len,"SKIP") )
9852 case 'T': /* (*THEN) */
9853 /* [19:06] <TimToady> :: is then */
9854 if ( memEQs(start_verb,verb_len,"THEN") ) {
9856 RExC_seen |= REG_CUTGROUP_SEEN;
9861 RExC_parse += UTF ? UTF8SKIP(RExC_parse) : 1;
9863 "Unknown verb pattern '%"UTF8f"'",
9864 UTF8fARG(UTF, verb_len, start_verb));
9867 if ( start_arg && internal_argval ) {
9868 vFAIL3("Verb pattern '%.*s' may not have an argument",
9869 verb_len, start_verb);
9870 } else if ( argok < 0 && !start_arg ) {
9871 vFAIL3("Verb pattern '%.*s' has a mandatory argument",
9872 verb_len, start_verb);
9874 ret = reganode(pRExC_state, op, internal_argval);
9875 if ( ! internal_argval && ! SIZE_ONLY ) {
9877 SV *sv = newSVpvn( start_arg,
9878 RExC_parse - start_arg);
9879 ARG(ret) = add_data( pRExC_state,
9881 RExC_rxi->data->data[ARG(ret)]=(void*)sv;
9888 if (!internal_argval)
9889 RExC_seen |= REG_VERBARG_SEEN;
9890 } else if ( start_arg ) {
9891 vFAIL3("Verb pattern '%.*s' may not have an argument",
9892 verb_len, start_verb);
9894 ret = reg_node(pRExC_state, op);
9896 nextchar(pRExC_state);
9899 else if (*RExC_parse == '?') { /* (?...) */
9900 bool is_logical = 0;
9901 const char * const seqstart = RExC_parse;
9902 const char * endptr;
9903 if (has_intervening_patws) {
9905 vFAIL("In '(?...)', the '(' and '?' must be adjacent");
9909 paren = *RExC_parse++;
9910 ret = NULL; /* For look-ahead/behind. */
9913 case 'P': /* (?P...) variants for those used to PCRE/Python */
9914 paren = *RExC_parse++;
9915 if ( paren == '<') /* (?P<...>) named capture */
9917 else if (paren == '>') { /* (?P>name) named recursion */
9918 goto named_recursion;
9920 else if (paren == '=') { /* (?P=...) named backref */
9921 /* this pretty much dupes the code for \k<NAME> in
9922 * regatom(), if you change this make sure you change that
9924 char* name_start = RExC_parse;
9926 SV *sv_dat = reg_scan_name(pRExC_state,
9927 SIZE_ONLY ? REG_RSN_RETURN_NULL : REG_RSN_RETURN_DATA);
9928 if (RExC_parse == name_start || *RExC_parse != ')')
9929 /* diag_listed_as: Sequence ?P=... not terminated in regex; marked by <-- HERE in m/%s/ */
9930 vFAIL2("Sequence %.3s... not terminated",parse_start);
9933 num = add_data( pRExC_state, STR_WITH_LEN("S"));
9934 RExC_rxi->data->data[num]=(void*)sv_dat;
9935 SvREFCNT_inc_simple_void(sv_dat);
9938 ret = reganode(pRExC_state,
9941 : (ASCII_FOLD_RESTRICTED)
9943 : (AT_LEAST_UNI_SEMANTICS)
9951 Set_Node_Offset(ret, parse_start+1);
9952 Set_Node_Cur_Length(ret, parse_start);
9954 nextchar(pRExC_state);
9958 RExC_parse += SKIP_IF_CHAR(RExC_parse);
9959 /* diag_listed_as: Sequence (?%s...) not recognized in regex; marked by <-- HERE in m/%s/ */
9960 vFAIL3("Sequence (%.*s...) not recognized",
9961 RExC_parse-seqstart, seqstart);
9962 NOT_REACHED; /*NOTREACHED*/
9963 case '<': /* (?<...) */
9964 if (*RExC_parse == '!')
9966 else if (*RExC_parse != '=')
9972 case '\'': /* (?'...') */
9973 name_start= RExC_parse;
9974 svname = reg_scan_name(pRExC_state,
9975 SIZE_ONLY /* reverse test from the others */
9976 ? REG_RSN_RETURN_NAME
9977 : REG_RSN_RETURN_NULL);
9978 if (RExC_parse == name_start || *RExC_parse != paren)
9979 vFAIL2("Sequence (?%c... not terminated",
9980 paren=='>' ? '<' : paren);
9984 if (!svname) /* shouldn't happen */
9986 "panic: reg_scan_name returned NULL");
9987 if (!RExC_paren_names) {
9988 RExC_paren_names= newHV();
9989 sv_2mortal(MUTABLE_SV(RExC_paren_names));
9991 RExC_paren_name_list= newAV();
9992 sv_2mortal(MUTABLE_SV(RExC_paren_name_list));
9995 he_str = hv_fetch_ent( RExC_paren_names, svname, 1, 0 );
9997 sv_dat = HeVAL(he_str);
9999 /* croak baby croak */
10001 "panic: paren_name hash element allocation failed");
10002 } else if ( SvPOK(sv_dat) ) {
10003 /* (?|...) can mean we have dupes so scan to check
10004 its already been stored. Maybe a flag indicating
10005 we are inside such a construct would be useful,
10006 but the arrays are likely to be quite small, so
10007 for now we punt -- dmq */
10008 IV count = SvIV(sv_dat);
10009 I32 *pv = (I32*)SvPVX(sv_dat);
10011 for ( i = 0 ; i < count ; i++ ) {
10012 if ( pv[i] == RExC_npar ) {
10018 pv = (I32*)SvGROW(sv_dat,
10019 SvCUR(sv_dat) + sizeof(I32)+1);
10020 SvCUR_set(sv_dat, SvCUR(sv_dat) + sizeof(I32));
10021 pv[count] = RExC_npar;
10022 SvIV_set(sv_dat, SvIVX(sv_dat) + 1);
10025 (void)SvUPGRADE(sv_dat,SVt_PVNV);
10026 sv_setpvn(sv_dat, (char *)&(RExC_npar),
10029 SvIV_set(sv_dat, 1);
10032 /* Yes this does cause a memory leak in debugging Perls
10034 if (!av_store(RExC_paren_name_list,
10035 RExC_npar, SvREFCNT_inc(svname)))
10036 SvREFCNT_dec_NN(svname);
10039 /*sv_dump(sv_dat);*/
10041 nextchar(pRExC_state);
10043 goto capturing_parens;
10045 RExC_seen |= REG_LOOKBEHIND_SEEN;
10046 RExC_in_lookbehind++;
10049 case '=': /* (?=...) */
10050 RExC_seen_zerolen++;
10052 case '!': /* (?!...) */
10053 RExC_seen_zerolen++;
10054 /* check if we're really just a "FAIL" assertion */
10056 nextchar(pRExC_state);
10057 if (*RExC_parse == ')') {
10058 ret=reg_node(pRExC_state, OPFAIL);
10059 nextchar(pRExC_state);
10063 case '|': /* (?|...) */
10064 /* branch reset, behave like a (?:...) except that
10065 buffers in alternations share the same numbers */
10067 after_freeze = freeze_paren = RExC_npar;
10069 case ':': /* (?:...) */
10070 case '>': /* (?>...) */
10072 case '$': /* (?$...) */
10073 case '@': /* (?@...) */
10074 vFAIL2("Sequence (?%c...) not implemented", (int)paren);
10076 case '0' : /* (?0) */
10077 case 'R' : /* (?R) */
10078 if (*RExC_parse != ')')
10079 FAIL("Sequence (?R) not terminated");
10080 ret = reg_node(pRExC_state, GOSTART);
10081 RExC_seen |= REG_GOSTART_SEEN;
10082 *flagp |= POSTPONED;
10083 nextchar(pRExC_state);
10086 /* named and numeric backreferences */
10087 case '&': /* (?&NAME) */
10088 parse_start = RExC_parse - 1;
10091 SV *sv_dat = reg_scan_name(pRExC_state,
10092 SIZE_ONLY ? REG_RSN_RETURN_NULL : REG_RSN_RETURN_DATA);
10093 num = sv_dat ? *((I32 *)SvPVX(sv_dat)) : 0;
10095 if (RExC_parse == RExC_end || *RExC_parse != ')')
10096 vFAIL("Sequence (?&... not terminated");
10097 goto gen_recurse_regop;
10100 if (!(RExC_parse[0] >= '1' && RExC_parse[0] <= '9')) {
10102 vFAIL("Illegal pattern");
10104 goto parse_recursion;
10106 case '-': /* (?-1) */
10107 if (!(RExC_parse[0] >= '1' && RExC_parse[0] <= '9')) {
10108 RExC_parse--; /* rewind to let it be handled later */
10112 case '1': case '2': case '3': case '4': /* (?1) */
10113 case '5': case '6': case '7': case '8': case '9':
10117 bool is_neg = FALSE;
10119 parse_start = RExC_parse - 1; /* MJD */
10120 if (*RExC_parse == '-') {
10124 unum = grok_atou(RExC_parse, &endptr);
10125 num = (unum > I32_MAX) ? I32_MAX : (I32)unum;
10127 RExC_parse = (char*)endptr;
10129 /* Some limit for num? */
10133 if (*RExC_parse!=')')
10134 vFAIL("Expecting close bracket");
10137 if ( paren == '-' ) {
10139 Diagram of capture buffer numbering.
10140 Top line is the normal capture buffer numbers
10141 Bottom line is the negative indexing as from
10145 /(a(x)y)(a(b(c(?-2)d)e)f)(g(h))/
10149 num = RExC_npar + num;
10152 vFAIL("Reference to nonexistent group");
10154 } else if ( paren == '+' ) {
10155 num = RExC_npar + num - 1;
10158 ret = reg2Lanode(pRExC_state, GOSUB, num, RExC_recurse_count);
10160 if (num > (I32)RExC_rx->nparens) {
10162 vFAIL("Reference to nonexistent group");
10164 RExC_recurse_count++;
10165 DEBUG_OPTIMISE_MORE_r(PerlIO_printf(Perl_debug_log,
10166 "%*s%*s Recurse #%"UVuf" to %"IVdf"\n",
10167 22, "| |", (int)(depth * 2 + 1), "",
10168 (UV)ARG(ret), (IV)ARG2L(ret)));
10170 RExC_seen |= REG_RECURSE_SEEN;
10171 Set_Node_Length(ret, 1 + regarglen[OP(ret)]); /* MJD */
10172 Set_Node_Offset(ret, parse_start); /* MJD */
10174 *flagp |= POSTPONED;
10175 nextchar(pRExC_state);
10180 case '?': /* (??...) */
10182 if (*RExC_parse != '{') {
10183 RExC_parse += SKIP_IF_CHAR(RExC_parse);
10184 /* diag_listed_as: Sequence (?%s...) not recognized in regex; marked by <-- HERE in m/%s/ */
10186 "Sequence (%"UTF8f"...) not recognized",
10187 UTF8fARG(UTF, RExC_parse-seqstart, seqstart));
10188 NOT_REACHED; /*NOTREACHED*/
10190 *flagp |= POSTPONED;
10191 paren = *RExC_parse++;
10193 case '{': /* (?{...}) */
10196 struct reg_code_block *cb;
10198 RExC_seen_zerolen++;
10200 if ( !pRExC_state->num_code_blocks
10201 || pRExC_state->code_index >= pRExC_state->num_code_blocks
10202 || pRExC_state->code_blocks[pRExC_state->code_index].start
10203 != (STRLEN)((RExC_parse -3 - (is_logical ? 1 : 0))
10206 if (RExC_pm_flags & PMf_USE_RE_EVAL)
10207 FAIL("panic: Sequence (?{...}): no code block found\n");
10208 FAIL("Eval-group not allowed at runtime, use re 'eval'");
10210 /* this is a pre-compiled code block (?{...}) */
10211 cb = &pRExC_state->code_blocks[pRExC_state->code_index];
10212 RExC_parse = RExC_start + cb->end;
10215 if (cb->src_regex) {
10216 n = add_data(pRExC_state, STR_WITH_LEN("rl"));
10217 RExC_rxi->data->data[n] =
10218 (void*)SvREFCNT_inc((SV*)cb->src_regex);
10219 RExC_rxi->data->data[n+1] = (void*)o;
10222 n = add_data(pRExC_state,
10223 (RExC_pm_flags & PMf_HAS_CV) ? "L" : "l", 1);
10224 RExC_rxi->data->data[n] = (void*)o;
10227 pRExC_state->code_index++;
10228 nextchar(pRExC_state);
10232 ret = reg_node(pRExC_state, LOGICAL);
10234 eval = reg2Lanode(pRExC_state, EVAL,
10237 /* for later propagation into (??{})
10239 RExC_flags & RXf_PMf_COMPILETIME
10244 REGTAIL(pRExC_state, ret, eval);
10245 /* deal with the length of this later - MJD */
10248 ret = reg2Lanode(pRExC_state, EVAL, n, 0);
10249 Set_Node_Length(ret, RExC_parse - parse_start + 1);
10250 Set_Node_Offset(ret, parse_start);
10253 case '(': /* (?(?{...})...) and (?(?=...)...) */
10256 const int DEFINE_len = sizeof("DEFINE") - 1;
10257 if (RExC_parse[0] == '?') { /* (?(?...)) */
10258 if (RExC_parse[1] == '=' || RExC_parse[1] == '!'
10259 || RExC_parse[1] == '<'
10260 || RExC_parse[1] == '{') { /* Lookahead or eval. */
10264 ret = reg_node(pRExC_state, LOGICAL);
10268 tail = reg(pRExC_state, 1, &flag, depth+1);
10269 if (flag & RESTART_UTF8) {
10270 *flagp = RESTART_UTF8;
10273 REGTAIL(pRExC_state, ret, tail);
10276 /* Fall through to ‘Unknown switch condition’ at the
10277 end of the if/else chain. */
10279 else if ( RExC_parse[0] == '<' /* (?(<NAME>)...) */
10280 || RExC_parse[0] == '\'' ) /* (?('NAME')...) */
10282 char ch = RExC_parse[0] == '<' ? '>' : '\'';
10283 char *name_start= RExC_parse++;
10285 SV *sv_dat=reg_scan_name(pRExC_state,
10286 SIZE_ONLY ? REG_RSN_RETURN_NULL : REG_RSN_RETURN_DATA);
10287 if (RExC_parse == name_start || *RExC_parse != ch)
10288 vFAIL2("Sequence (?(%c... not terminated",
10289 (ch == '>' ? '<' : ch));
10292 num = add_data( pRExC_state, STR_WITH_LEN("S"));
10293 RExC_rxi->data->data[num]=(void*)sv_dat;
10294 SvREFCNT_inc_simple_void(sv_dat);
10296 ret = reganode(pRExC_state,NGROUPP,num);
10297 goto insert_if_check_paren;
10299 else if (RExC_end - RExC_parse >= DEFINE_len
10300 && strnEQ(RExC_parse, "DEFINE", DEFINE_len))
10302 ret = reganode(pRExC_state,DEFINEP,0);
10303 RExC_parse += DEFINE_len;
10305 goto insert_if_check_paren;
10307 else if (RExC_parse[0] == 'R') {
10310 if (RExC_parse[0] >= '1' && RExC_parse[0] <= '9' ) {
10311 parno = grok_atou(RExC_parse, &endptr);
10313 RExC_parse = (char*)endptr;
10314 } else if (RExC_parse[0] == '&') {
10317 sv_dat = reg_scan_name(pRExC_state,
10319 ? REG_RSN_RETURN_NULL
10320 : REG_RSN_RETURN_DATA);
10321 parno = sv_dat ? *((I32 *)SvPVX(sv_dat)) : 0;
10323 ret = reganode(pRExC_state,INSUBP,parno);
10324 goto insert_if_check_paren;
10326 else if (RExC_parse[0] >= '1' && RExC_parse[0] <= '9' ) {
10330 parno = grok_atou(RExC_parse, &endptr);
10332 RExC_parse = (char*)endptr;
10333 ret = reganode(pRExC_state, GROUPP, parno);
10335 insert_if_check_paren:
10336 if (*(tmp = nextchar(pRExC_state)) != ')') {
10337 /* nextchar also skips comments, so undo its work
10338 * and skip over the the next character.
10341 RExC_parse += UTF ? UTF8SKIP(RExC_parse) : 1;
10342 vFAIL("Switch condition not recognized");
10345 REGTAIL(pRExC_state, ret, reganode(pRExC_state, IFTHEN, 0));
10346 br = regbranch(pRExC_state, &flags, 1,depth+1);
10348 if (flags & RESTART_UTF8) {
10349 *flagp = RESTART_UTF8;
10352 FAIL2("panic: regbranch returned NULL, flags=%#"UVxf"",
10355 REGTAIL(pRExC_state, br, reganode(pRExC_state,
10357 c = *nextchar(pRExC_state);
10358 if (flags&HASWIDTH)
10359 *flagp |= HASWIDTH;
10362 vFAIL("(?(DEFINE)....) does not allow branches");
10364 /* Fake one for optimizer. */
10365 lastbr = reganode(pRExC_state, IFTHEN, 0);
10367 if (!regbranch(pRExC_state, &flags, 1,depth+1)) {
10368 if (flags & RESTART_UTF8) {
10369 *flagp = RESTART_UTF8;
10372 FAIL2("panic: regbranch returned NULL, flags=%#"UVxf"",
10375 REGTAIL(pRExC_state, ret, lastbr);
10376 if (flags&HASWIDTH)
10377 *flagp |= HASWIDTH;
10378 c = *nextchar(pRExC_state);
10383 if (RExC_parse>RExC_end)
10384 vFAIL("Switch (?(condition)... not terminated");
10386 vFAIL("Switch (?(condition)... contains too many branches");
10388 ender = reg_node(pRExC_state, TAIL);
10389 REGTAIL(pRExC_state, br, ender);
10391 REGTAIL(pRExC_state, lastbr, ender);
10392 REGTAIL(pRExC_state, NEXTOPER(NEXTOPER(lastbr)), ender);
10395 REGTAIL(pRExC_state, ret, ender);
10396 RExC_size++; /* XXX WHY do we need this?!!
10397 For large programs it seems to be required
10398 but I can't figure out why. -- dmq*/
10401 RExC_parse += UTF ? UTF8SKIP(RExC_parse) : 1;
10402 vFAIL("Unknown switch condition (?(...))");
10404 case '[': /* (?[ ... ]) */
10405 return handle_regex_sets(pRExC_state, NULL, flagp, depth,
10408 RExC_parse--; /* for vFAIL to print correctly */
10409 vFAIL("Sequence (? incomplete");
10411 default: /* e.g., (?i) */
10414 parse_lparen_question_flags(pRExC_state);
10415 if (UCHARAT(RExC_parse) != ':') {
10417 nextchar(pRExC_state);
10422 nextchar(pRExC_state);
10427 else if (!(RExC_flags & RXf_PMf_NOCAPTURE)) { /* (...) */
10432 ret = reganode(pRExC_state, OPEN, parno);
10434 if (!RExC_nestroot)
10435 RExC_nestroot = parno;
10436 if (RExC_seen & REG_RECURSE_SEEN
10437 && !RExC_open_parens[parno-1])
10439 DEBUG_OPTIMISE_MORE_r(PerlIO_printf(Perl_debug_log,
10440 "%*s%*s Setting open paren #%"IVdf" to %d\n",
10441 22, "| |", (int)(depth * 2 + 1), "",
10442 (IV)parno, REG_NODE_NUM(ret)));
10443 RExC_open_parens[parno-1]= ret;
10446 Set_Node_Length(ret, 1); /* MJD */
10447 Set_Node_Offset(ret, RExC_parse); /* MJD */
10450 /* with RXf_PMf_NOCAPTURE treat (...) as (?:...) */
10459 /* Pick up the branches, linking them together. */
10460 parse_start = RExC_parse; /* MJD */
10461 br = regbranch(pRExC_state, &flags, 1,depth+1);
10463 /* branch_len = (paren != 0); */
10466 if (flags & RESTART_UTF8) {
10467 *flagp = RESTART_UTF8;
10470 FAIL2("panic: regbranch returned NULL, flags=%#"UVxf"", (UV) flags);
10472 if (*RExC_parse == '|') {
10473 if (!SIZE_ONLY && RExC_extralen) {
10474 reginsert(pRExC_state, BRANCHJ, br, depth+1);
10477 reginsert(pRExC_state, BRANCH, br, depth+1);
10478 Set_Node_Length(br, paren != 0);
10479 Set_Node_Offset_To_R(br-RExC_emit_start, parse_start-RExC_start);
10483 RExC_extralen += 1; /* For BRANCHJ-BRANCH. */
10485 else if (paren == ':') {
10486 *flagp |= flags&SIMPLE;
10488 if (is_open) { /* Starts with OPEN. */
10489 REGTAIL(pRExC_state, ret, br); /* OPEN -> first. */
10491 else if (paren != '?') /* Not Conditional */
10493 *flagp |= flags & (SPSTART | HASWIDTH | POSTPONED);
10495 while (*RExC_parse == '|') {
10496 if (!SIZE_ONLY && RExC_extralen) {
10497 ender = reganode(pRExC_state, LONGJMP,0);
10499 /* Append to the previous. */
10500 REGTAIL(pRExC_state, NEXTOPER(NEXTOPER(lastbr)), ender);
10503 RExC_extralen += 2; /* Account for LONGJMP. */
10504 nextchar(pRExC_state);
10505 if (freeze_paren) {
10506 if (RExC_npar > after_freeze)
10507 after_freeze = RExC_npar;
10508 RExC_npar = freeze_paren;
10510 br = regbranch(pRExC_state, &flags, 0, depth+1);
10513 if (flags & RESTART_UTF8) {
10514 *flagp = RESTART_UTF8;
10517 FAIL2("panic: regbranch returned NULL, flags=%#"UVxf"", (UV) flags);
10519 REGTAIL(pRExC_state, lastbr, br); /* BRANCH -> BRANCH. */
10521 *flagp |= flags & (SPSTART | HASWIDTH | POSTPONED);
10524 if (have_branch || paren != ':') {
10525 /* Make a closing node, and hook it on the end. */
10528 ender = reg_node(pRExC_state, TAIL);
10531 ender = reganode(pRExC_state, CLOSE, parno);
10532 if (!SIZE_ONLY && RExC_seen & REG_RECURSE_SEEN) {
10533 DEBUG_OPTIMISE_MORE_r(PerlIO_printf(Perl_debug_log,
10534 "%*s%*s Setting close paren #%"IVdf" to %d\n",
10535 22, "| |", (int)(depth * 2 + 1), "", (IV)parno, REG_NODE_NUM(ender)));
10536 RExC_close_parens[parno-1]= ender;
10537 if (RExC_nestroot == parno)
10540 Set_Node_Offset(ender,RExC_parse+1); /* MJD */
10541 Set_Node_Length(ender,1); /* MJD */
10547 *flagp &= ~HASWIDTH;
10550 ender = reg_node(pRExC_state, SUCCEED);
10553 ender = reg_node(pRExC_state, END);
10555 assert(!RExC_opend); /* there can only be one! */
10556 RExC_opend = ender;
10560 DEBUG_PARSE_r(if (!SIZE_ONLY) {
10561 DEBUG_PARSE_MSG("lsbr");
10562 regprop(RExC_rx, RExC_mysv1, lastbr, NULL, pRExC_state);
10563 regprop(RExC_rx, RExC_mysv2, ender, NULL, pRExC_state);
10564 PerlIO_printf(Perl_debug_log, "~ tying lastbr %s (%"IVdf") to ender %s (%"IVdf") offset %"IVdf"\n",
10565 SvPV_nolen_const(RExC_mysv1),
10566 (IV)REG_NODE_NUM(lastbr),
10567 SvPV_nolen_const(RExC_mysv2),
10568 (IV)REG_NODE_NUM(ender),
10569 (IV)(ender - lastbr)
10572 REGTAIL(pRExC_state, lastbr, ender);
10574 if (have_branch && !SIZE_ONLY) {
10575 char is_nothing= 1;
10577 RExC_seen |= REG_TOP_LEVEL_BRANCHES_SEEN;
10579 /* Hook the tails of the branches to the closing node. */
10580 for (br = ret; br; br = regnext(br)) {
10581 const U8 op = PL_regkind[OP(br)];
10582 if (op == BRANCH) {
10583 REGTAIL_STUDY(pRExC_state, NEXTOPER(br), ender);
10584 if ( OP(NEXTOPER(br)) != NOTHING
10585 || regnext(NEXTOPER(br)) != ender)
10588 else if (op == BRANCHJ) {
10589 REGTAIL_STUDY(pRExC_state, NEXTOPER(NEXTOPER(br)), ender);
10590 /* for now we always disable this optimisation * /
10591 if ( OP(NEXTOPER(NEXTOPER(br))) != NOTHING
10592 || regnext(NEXTOPER(NEXTOPER(br))) != ender)
10598 br= PL_regkind[OP(ret)] != BRANCH ? regnext(ret) : ret;
10599 DEBUG_PARSE_r(if (!SIZE_ONLY) {
10600 DEBUG_PARSE_MSG("NADA");
10601 regprop(RExC_rx, RExC_mysv1, ret, NULL, pRExC_state);
10602 regprop(RExC_rx, RExC_mysv2, ender, NULL, pRExC_state);
10603 PerlIO_printf(Perl_debug_log, "~ converting ret %s (%"IVdf") to ender %s (%"IVdf") offset %"IVdf"\n",
10604 SvPV_nolen_const(RExC_mysv1),
10605 (IV)REG_NODE_NUM(ret),
10606 SvPV_nolen_const(RExC_mysv2),
10607 (IV)REG_NODE_NUM(ender),
10612 if (OP(ender) == TAIL) {
10617 for ( opt= br + 1; opt < ender ; opt++ )
10618 OP(opt)= OPTIMIZED;
10619 NEXT_OFF(br)= ender - br;
10627 static const char parens[] = "=!<,>";
10629 if (paren && (p = strchr(parens, paren))) {
10630 U8 node = ((p - parens) % 2) ? UNLESSM : IFMATCH;
10631 int flag = (p - parens) > 1;
10634 node = SUSPEND, flag = 0;
10635 reginsert(pRExC_state, node,ret, depth+1);
10636 Set_Node_Cur_Length(ret, parse_start);
10637 Set_Node_Offset(ret, parse_start + 1);
10639 REGTAIL_STUDY(pRExC_state, ret, reg_node(pRExC_state, TAIL));
10643 /* Check for proper termination. */
10645 /* restore original flags, but keep (?p) */
10646 RExC_flags = oregflags | (RExC_flags & RXf_PMf_KEEPCOPY);
10647 if (RExC_parse >= RExC_end || *nextchar(pRExC_state) != ')') {
10648 RExC_parse = oregcomp_parse;
10649 vFAIL("Unmatched (");
10652 else if (!paren && RExC_parse < RExC_end) {
10653 if (*RExC_parse == ')') {
10655 vFAIL("Unmatched )");
10658 FAIL("Junk on end of regexp"); /* "Can't happen". */
10659 NOT_REACHED; /* NOTREACHED */
10662 if (RExC_in_lookbehind) {
10663 RExC_in_lookbehind--;
10665 if (after_freeze > RExC_npar)
10666 RExC_npar = after_freeze;
10671 - regbranch - one alternative of an | operator
10673 * Implements the concatenation operator.
10675 * Returns NULL, setting *flagp to RESTART_UTF8 if the sizing scan needs to be
10679 S_regbranch(pTHX_ RExC_state_t *pRExC_state, I32 *flagp, I32 first, U32 depth)
10682 regnode *chain = NULL;
10684 I32 flags = 0, c = 0;
10685 GET_RE_DEBUG_FLAGS_DECL;
10687 PERL_ARGS_ASSERT_REGBRANCH;
10689 DEBUG_PARSE("brnc");
10694 if (!SIZE_ONLY && RExC_extralen)
10695 ret = reganode(pRExC_state, BRANCHJ,0);
10697 ret = reg_node(pRExC_state, BRANCH);
10698 Set_Node_Length(ret, 1);
10702 if (!first && SIZE_ONLY)
10703 RExC_extralen += 1; /* BRANCHJ */
10705 *flagp = WORST; /* Tentatively. */
10708 nextchar(pRExC_state);
10709 while (RExC_parse < RExC_end && *RExC_parse != '|' && *RExC_parse != ')') {
10710 flags &= ~TRYAGAIN;
10711 latest = regpiece(pRExC_state, &flags,depth+1);
10712 if (latest == NULL) {
10713 if (flags & TRYAGAIN)
10715 if (flags & RESTART_UTF8) {
10716 *flagp = RESTART_UTF8;
10719 FAIL2("panic: regpiece returned NULL, flags=%#"UVxf"", (UV) flags);
10721 else if (ret == NULL)
10723 *flagp |= flags&(HASWIDTH|POSTPONED);
10724 if (chain == NULL) /* First piece. */
10725 *flagp |= flags&SPSTART;
10727 /* FIXME adding one for every branch after the first is probably
10728 * excessive now we have TRIE support. (hv) */
10730 REGTAIL(pRExC_state, chain, latest);
10735 if (chain == NULL) { /* Loop ran zero times. */
10736 chain = reg_node(pRExC_state, NOTHING);
10741 *flagp |= flags&SIMPLE;
10748 - regpiece - something followed by possible [*+?]
10750 * Note that the branching code sequences used for ? and the general cases
10751 * of * and + are somewhat optimized: they use the same NOTHING node as
10752 * both the endmarker for their branch list and the body of the last branch.
10753 * It might seem that this node could be dispensed with entirely, but the
10754 * endmarker role is not redundant.
10756 * Returns NULL, setting *flagp to TRYAGAIN if regatom() returns NULL with
10758 * Returns NULL, setting *flagp to RESTART_UTF8 if the sizing scan needs to be
10762 S_regpiece(pTHX_ RExC_state_t *pRExC_state, I32 *flagp, U32 depth)
10768 const char * const origparse = RExC_parse;
10770 I32 max = REG_INFTY;
10771 #ifdef RE_TRACK_PATTERN_OFFSETS
10774 const char *maxpos = NULL;
10776 /* Save the original in case we change the emitted regop to a FAIL. */
10777 regnode * const orig_emit = RExC_emit;
10779 GET_RE_DEBUG_FLAGS_DECL;
10781 PERL_ARGS_ASSERT_REGPIECE;
10783 DEBUG_PARSE("piec");
10785 ret = regatom(pRExC_state, &flags,depth+1);
10787 if (flags & (TRYAGAIN|RESTART_UTF8))
10788 *flagp |= flags & (TRYAGAIN|RESTART_UTF8);
10790 FAIL2("panic: regatom returned NULL, flags=%#"UVxf"", (UV) flags);
10796 if (op == '{' && regcurly(RExC_parse)) {
10798 #ifdef RE_TRACK_PATTERN_OFFSETS
10799 parse_start = RExC_parse; /* MJD */
10801 next = RExC_parse + 1;
10802 while (isDIGIT(*next) || *next == ',') {
10803 if (*next == ',') {
10811 if (*next == '}') { /* got one */
10812 const char* endptr;
10816 min = grok_atou(RExC_parse, &endptr);
10817 if (*maxpos == ',')
10820 maxpos = RExC_parse;
10821 max = grok_atou(maxpos, &endptr);
10822 if (!max && *maxpos != '0')
10823 max = REG_INFTY; /* meaning "infinity" */
10824 else if (max >= REG_INFTY)
10825 vFAIL2("Quantifier in {,} bigger than %d", REG_INFTY - 1);
10827 nextchar(pRExC_state);
10828 if (max < min) { /* If can't match, warn and optimize to fail
10832 /* We can't back off the size because we have to reserve
10833 * enough space for all the things we are about to throw
10834 * away, but we can shrink it by the ammount we are about
10835 * to re-use here */
10836 RExC_size = PREVOPER(RExC_size) - regarglen[(U8)OPFAIL];
10839 ckWARNreg(RExC_parse, "Quantifier {n,m} with n > m can't match");
10840 RExC_emit = orig_emit;
10842 ret = reg_node(pRExC_state, OPFAIL);
10845 else if (min == max
10846 && RExC_parse < RExC_end
10847 && (*RExC_parse == '?' || *RExC_parse == '+'))
10850 ckWARN2reg(RExC_parse + 1,
10851 "Useless use of greediness modifier '%c'",
10854 /* Absorb the modifier, so later code doesn't see nor use
10856 nextchar(pRExC_state);
10860 if ((flags&SIMPLE)) {
10861 MARK_NAUGHTY_EXP(2, 2);
10862 reginsert(pRExC_state, CURLY, ret, depth+1);
10863 Set_Node_Offset(ret, parse_start+1); /* MJD */
10864 Set_Node_Cur_Length(ret, parse_start);
10867 regnode * const w = reg_node(pRExC_state, WHILEM);
10870 REGTAIL(pRExC_state, ret, w);
10871 if (!SIZE_ONLY && RExC_extralen) {
10872 reginsert(pRExC_state, LONGJMP,ret, depth+1);
10873 reginsert(pRExC_state, NOTHING,ret, depth+1);
10874 NEXT_OFF(ret) = 3; /* Go over LONGJMP. */
10876 reginsert(pRExC_state, CURLYX,ret, depth+1);
10878 Set_Node_Offset(ret, parse_start+1);
10879 Set_Node_Length(ret,
10880 op == '{' ? (RExC_parse - parse_start) : 1);
10882 if (!SIZE_ONLY && RExC_extralen)
10883 NEXT_OFF(ret) = 3; /* Go over NOTHING to LONGJMP. */
10884 REGTAIL(pRExC_state, ret, reg_node(pRExC_state, NOTHING));
10886 RExC_whilem_seen++, RExC_extralen += 3;
10887 MARK_NAUGHTY_EXP(1, 4); /* compound interest */
10894 *flagp |= HASWIDTH;
10896 ARG1_SET(ret, (U16)min);
10897 ARG2_SET(ret, (U16)max);
10899 if (max == REG_INFTY)
10900 RExC_seen |= REG_UNBOUNDED_QUANTIFIER_SEEN;
10906 if (!ISMULT1(op)) {
10911 #if 0 /* Now runtime fix should be reliable. */
10913 /* if this is reinstated, don't forget to put this back into perldiag:
10915 =item Regexp *+ operand could be empty at {#} in regex m/%s/
10917 (F) The part of the regexp subject to either the * or + quantifier
10918 could match an empty string. The {#} shows in the regular
10919 expression about where the problem was discovered.
10923 if (!(flags&HASWIDTH) && op != '?')
10924 vFAIL("Regexp *+ operand could be empty");
10927 #ifdef RE_TRACK_PATTERN_OFFSETS
10928 parse_start = RExC_parse;
10930 nextchar(pRExC_state);
10932 *flagp = (op != '+') ? (WORST|SPSTART|HASWIDTH) : (WORST|HASWIDTH);
10934 if (op == '*' && (flags&SIMPLE)) {
10935 reginsert(pRExC_state, STAR, ret, depth+1);
10938 RExC_seen |= REG_UNBOUNDED_QUANTIFIER_SEEN;
10940 else if (op == '*') {
10944 else if (op == '+' && (flags&SIMPLE)) {
10945 reginsert(pRExC_state, PLUS, ret, depth+1);
10948 RExC_seen |= REG_UNBOUNDED_QUANTIFIER_SEEN;
10950 else if (op == '+') {
10954 else if (op == '?') {
10959 if (!SIZE_ONLY && !(flags&(HASWIDTH|POSTPONED)) && max > REG_INFTY/3) {
10960 SAVEFREESV(RExC_rx_sv); /* in case of fatal warnings */
10961 ckWARN2reg(RExC_parse,
10962 "%"UTF8f" matches null string many times",
10963 UTF8fARG(UTF, (RExC_parse >= origparse
10964 ? RExC_parse - origparse
10967 (void)ReREFCNT_inc(RExC_rx_sv);
10970 if (RExC_parse < RExC_end && *RExC_parse == '?') {
10971 nextchar(pRExC_state);
10972 reginsert(pRExC_state, MINMOD, ret, depth+1);
10973 REGTAIL(pRExC_state, ret, ret + NODE_STEP_REGNODE);
10976 if (RExC_parse < RExC_end && *RExC_parse == '+') {
10978 nextchar(pRExC_state);
10979 ender = reg_node(pRExC_state, SUCCEED);
10980 REGTAIL(pRExC_state, ret, ender);
10981 reginsert(pRExC_state, SUSPEND, ret, depth+1);
10983 ender = reg_node(pRExC_state, TAIL);
10984 REGTAIL(pRExC_state, ret, ender);
10987 if (RExC_parse < RExC_end && ISMULT2(RExC_parse)) {
10989 vFAIL("Nested quantifiers");
10996 S_grok_bslash_N(pTHX_ RExC_state_t *pRExC_state, regnode** node_p,
10997 UV *valuep, I32 *flagp, U32 depth, SV** substitute_parse
11001 /* This is expected to be called by a parser routine that has recognized '\N'
11002 and needs to handle the rest. RExC_parse is expected to point at the first
11003 char following the N at the time of the call. On successful return,
11004 RExC_parse has been updated to point to just after the sequence identified
11005 by this routine, <*flagp> has been updated, and the non-NULL input pointers
11006 have been set appropriately.
11008 The typical case for this is \N{some character name}. This is usually
11009 called while parsing the input, filling in or ready to fill in an EXACTish
11010 node, and the code point for the character should be returned, so that it
11011 can be added to the node, and parsing continued with the next input
11012 character. But it may be that instead of a single character the \N{}
11013 expands to more than one, a named sequence. In this case any following
11014 quantifier applies to the whole sequence, and it is easier, given the code
11015 structure that calls this, to handle it from a different area of the code.
11016 For this reason, the input parameters can be set so that it returns valid
11017 only on one or the other of these cases.
11019 Another possibility is for the input to be an empty \N{}, which for
11020 backwards compatibility we accept, but generate a NOTHING node which should
11021 later get optimized out. This is handled from the area of code which can
11022 handle a named sequence, so if called with the parameters for the other, it
11025 Still another possibility is for the \N to mean [^\n], and not a single
11026 character or explicit sequence at all. This is determined by context.
11027 Again, this is handled from the area of code which can handle a named
11028 sequence, so if called with the parameters for the other, it also fails.
11030 And the final possibility is for the \N to be called from within a bracketed
11031 character class. In this case the [^\n] meaning makes no sense, and so is
11032 an error. Other anomalous situations are left to the calling code to handle.
11034 For non-single-quoted regexes, the tokenizer has attempted to decide which
11035 of the above applies, and in the case of a named sequence, has converted it
11036 into one of the forms: \N{} (if the sequence is null), or \N{U+c1.c2...},
11037 where c1... are the characters in the sequence. For single-quoted regexes,
11038 the tokenizer passes the \N sequence through unchanged; this code will not
11039 attempt to determine this nor expand those, instead raising a syntax error.
11040 The net effect is that if the beginning of the passed-in pattern isn't '{U+'
11041 or there is no '}', it signals that this \N occurrence means to match a
11042 non-newline. (This mostly was done because of [perl #56444].)
11044 The API is somewhat convoluted due to historical and the above reasons.
11046 The function raises an error (via vFAIL), and doesn't return for various
11047 syntax errors. For other failures, it returns (STRLEN) -1. For successes,
11048 it returns a count of how many characters were accounted for by it. (This
11049 can be 0 for \N{}; 1 for it meaning [^\n]; and otherwise the number of code
11050 points in the sequence. It sets <node_p>, <valuep>, and/or
11051 <substitute_parse> on success.
11053 If <valuep> is non-null, it means the caller can accept an input sequence
11054 consisting of just a single code point; <*valuep> is set to the value of the
11055 only or first code point in the input.
11057 If <substitute_parse> is non-null, it means the caller can accept an input
11058 sequence consisting of one or more code points; <*substitute_parse> is a
11059 newly created mortal SV* in this case, containing \x{} escapes representing
11062 Both <valuep> and <substitute_parse> can be non-NULL.
11064 If <node_p> is non-null, <substitute_parse> must be NULL. This signifies
11065 that the caller can accept any legal sequence other than a single code
11066 point. To wit, <*node_p> is set as follows:
11067 1) \N means not-a-NL: points to a newly created REG_ANY node; return is 1
11068 2) \N{}: points to a new NOTHING node; return is 0
11069 3) otherwise: points to a new EXACT node containing the resolved
11070 string; return is the number of code points in the
11071 string. This will never be 1.
11072 Note that failure is returned for single code point sequences if <valuep> is
11073 null and <node_p> is not.
11076 char * endbrace; /* '}' following the name */
11078 char *endchar; /* Points to '.' or '}' ending cur char in the input
11080 bool has_multiple_chars; /* true if the input stream contains a sequence of
11081 more than one character */
11082 bool in_char_class = substitute_parse != NULL;
11083 STRLEN count = 0; /* Number of characters in this sequence */
11085 GET_RE_DEBUG_FLAGS_DECL;
11087 PERL_ARGS_ASSERT_GROK_BSLASH_N;
11089 GET_RE_DEBUG_FLAGS;
11091 assert(cBOOL(node_p) ^ cBOOL(valuep)); /* Exactly one should be set */
11092 assert(! (node_p && substitute_parse)); /* At most 1 should be set */
11094 /* The [^\n] meaning of \N ignores spaces and comments under the /x
11095 * modifier. The other meaning does not, so use a temporary until we find
11096 * out which we are being called with */
11097 p = (RExC_flags & RXf_PMf_EXTENDED)
11098 ? regpatws(pRExC_state, RExC_parse,
11099 TRUE) /* means recognize comments */
11102 /* Disambiguate between \N meaning a named character versus \N meaning
11103 * [^\n]. The former is assumed when it can't be the latter. */
11104 if (*p != '{' || regcurly(p)) {
11107 /* no bare \N allowed in a charclass */
11108 if (in_char_class) {
11109 vFAIL("\\N in a character class must be a named character: \\N{...}");
11111 return (STRLEN) -1;
11113 RExC_parse--; /* Need to back off so nextchar() doesn't skip the
11115 nextchar(pRExC_state);
11116 *node_p = reg_node(pRExC_state, REG_ANY);
11117 *flagp |= HASWIDTH|SIMPLE;
11119 Set_Node_Length(*node_p, 1); /* MJD */
11123 /* Here, we have decided it should be a named character or sequence */
11125 /* The test above made sure that the next real character is a '{', but
11126 * under the /x modifier, it could be separated by space (or a comment and
11127 * \n) and this is not allowed (for consistency with \x{...} and the
11128 * tokenizer handling of \N{NAME}). */
11129 if (*RExC_parse != '{') {
11130 vFAIL("Missing braces on \\N{}");
11133 RExC_parse++; /* Skip past the '{' */
11135 if (! (endbrace = strchr(RExC_parse, '}')) /* no trailing brace */
11136 || ! (endbrace == RExC_parse /* nothing between the {} */
11137 || (endbrace - RExC_parse >= 2 /* U+ (bad hex is checked... */
11138 && strnEQ(RExC_parse, "U+", 2)))) /* ... below for a better
11141 if (endbrace) RExC_parse = endbrace; /* position msg's '<--HERE' */
11142 vFAIL("\\N{NAME} must be resolved by the lexer");
11145 RExC_uni_semantics = 1; /* Unicode named chars imply Unicode semantics */
11147 if (endbrace == RExC_parse) { /* empty: \N{} */
11149 *node_p = reg_node(pRExC_state,NOTHING);
11151 else if (! in_char_class) {
11152 return (STRLEN) -1;
11154 nextchar(pRExC_state);
11158 RExC_parse += 2; /* Skip past the 'U+' */
11160 endchar = RExC_parse + strcspn(RExC_parse, ".}");
11162 /* Code points are separated by dots. If none, there is only one code
11163 * point, and is terminated by the brace */
11164 has_multiple_chars = (endchar < endbrace);
11166 /* We get the first code point if we want it, and either there is only one,
11167 * or we can accept both cases of one and there is more than one */
11168 if (valuep && (substitute_parse || ! has_multiple_chars)) {
11169 STRLEN length_of_hex = (STRLEN)(endchar - RExC_parse);
11170 I32 grok_hex_flags = PERL_SCAN_ALLOW_UNDERSCORES
11171 | PERL_SCAN_DISALLOW_PREFIX
11173 /* No errors in the first pass (See [perl
11174 * #122671].) We let the code below find the
11175 * errors when there are multiple chars. */
11176 | ((SIZE_ONLY || has_multiple_chars)
11177 ? PERL_SCAN_SILENT_ILLDIGIT
11180 *valuep = grok_hex(RExC_parse, &length_of_hex, &grok_hex_flags, NULL);
11182 /* The tokenizer should have guaranteed validity, but it's possible to
11183 * bypass it by using single quoting, so check. Don't do the check
11184 * here when there are multiple chars; we do it below anyway. */
11185 if (! has_multiple_chars) {
11186 if (length_of_hex == 0
11187 || length_of_hex != (STRLEN)(endchar - RExC_parse) )
11189 RExC_parse += length_of_hex; /* Includes all the valid */
11190 RExC_parse += (RExC_orig_utf8) /* point to after 1st invalid */
11191 ? UTF8SKIP(RExC_parse)
11193 /* Guard against malformed utf8 */
11194 if (RExC_parse >= endchar) {
11195 RExC_parse = endchar;
11197 vFAIL("Invalid hexadecimal number in \\N{U+...}");
11200 RExC_parse = endbrace + 1;
11205 /* Here, we should have already handled the case where a single character
11206 * is expected and found. So it is a failure if we aren't expecting
11207 * multiple chars and got them; or didn't get them but wanted them. We
11208 * fail without advancing the parse, so that the caller can try again with
11209 * different acceptance criteria */
11210 if ((! node_p && ! substitute_parse) || ! has_multiple_chars) {
11212 return (STRLEN) -1;
11216 /* What is done here is to convert this to a sub-pattern of the form
11217 * \x{char1}\x{char2}...
11218 * and then either return it in <*substitute_parse> if non-null; or
11219 * call reg recursively to parse it (enclosing in "(?: ... )" ). That
11220 * way, it retains its atomicness, while not having to worry about
11221 * special handling that some code points may have. toke.c has
11222 * converted the original Unicode values to native, so that we can just
11223 * pass on the hex values unchanged. We do have to set a flag to keep
11224 * recoding from happening in the recursion */
11228 char *orig_end = RExC_end;
11231 if (substitute_parse) {
11232 *substitute_parse = newSVpvs("");
11235 substitute_parse = &dummy;
11236 *substitute_parse = newSVpvs("?:");
11238 *substitute_parse = sv_2mortal(*substitute_parse);
11240 while (RExC_parse < endbrace) {
11242 /* Convert to notation the rest of the code understands */
11243 sv_catpv(*substitute_parse, "\\x{");
11244 sv_catpvn(*substitute_parse, RExC_parse, endchar - RExC_parse);
11245 sv_catpv(*substitute_parse, "}");
11247 /* Point to the beginning of the next character in the sequence. */
11248 RExC_parse = endchar + 1;
11249 endchar = RExC_parse + strcspn(RExC_parse, ".}");
11253 if (! in_char_class) {
11254 sv_catpv(*substitute_parse, ")");
11257 RExC_parse = SvPV(*substitute_parse, len);
11259 /* Don't allow empty number */
11260 if (len < (STRLEN) ((substitute_parse) ? 6 : 8)) {
11261 RExC_parse = endbrace;
11262 vFAIL("Invalid hexadecimal number in \\N{U+...}");
11264 RExC_end = RExC_parse + len;
11266 /* The values are Unicode, and therefore not subject to recoding */
11267 RExC_override_recoding = 1;
11270 if (!(*node_p = reg(pRExC_state, 1, &flags, depth+1))) {
11271 if (flags & RESTART_UTF8) {
11272 *flagp = RESTART_UTF8;
11273 return (STRLEN) -1;
11275 FAIL2("panic: reg returned NULL to grok_bslash_N, flags=%#"UVxf"",
11278 *flagp |= flags&(HASWIDTH|SPSTART|SIMPLE|POSTPONED);
11281 RExC_parse = endbrace;
11282 RExC_end = orig_end;
11283 RExC_override_recoding = 0;
11285 nextchar(pRExC_state);
11295 * It returns the code point in utf8 for the value in *encp.
11296 * value: a code value in the source encoding
11297 * encp: a pointer to an Encode object
11299 * If the result from Encode is not a single character,
11300 * it returns U+FFFD (Replacement character) and sets *encp to NULL.
11303 S_reg_recode(pTHX_ const char value, SV **encp)
11306 SV * const sv = newSVpvn_flags(&value, numlen, SVs_TEMP);
11307 const char * const s = *encp ? sv_recode_to_utf8(sv, *encp) : SvPVX(sv);
11308 const STRLEN newlen = SvCUR(sv);
11309 UV uv = UNICODE_REPLACEMENT;
11311 PERL_ARGS_ASSERT_REG_RECODE;
11315 ? utf8n_to_uvchr((U8*)s, newlen, &numlen, UTF8_ALLOW_DEFAULT)
11318 if (!newlen || numlen != newlen) {
11319 uv = UNICODE_REPLACEMENT;
11325 PERL_STATIC_INLINE U8
11326 S_compute_EXACTish(RExC_state_t *pRExC_state)
11330 PERL_ARGS_ASSERT_COMPUTE_EXACTISH;
11338 op = get_regex_charset(RExC_flags);
11339 if (op >= REGEX_ASCII_RESTRICTED_CHARSET) {
11340 op--; /* /a is same as /u, and map /aa's offset to what /a's would have
11341 been, so there is no hole */
11344 return op + EXACTF;
11347 PERL_STATIC_INLINE void
11348 S_alloc_maybe_populate_EXACT(pTHX_ RExC_state_t *pRExC_state,
11349 regnode *node, I32* flagp, STRLEN len, UV code_point,
11352 /* This knows the details about sizing an EXACTish node, setting flags for
11353 * it (by setting <*flagp>, and potentially populating it with a single
11356 * If <len> (the length in bytes) is non-zero, this function assumes that
11357 * the node has already been populated, and just does the sizing. In this
11358 * case <code_point> should be the final code point that has already been
11359 * placed into the node. This value will be ignored except that under some
11360 * circumstances <*flagp> is set based on it.
11362 * If <len> is zero, the function assumes that the node is to contain only
11363 * the single character given by <code_point> and calculates what <len>
11364 * should be. In pass 1, it sizes the node appropriately. In pass 2, it
11365 * additionally will populate the node's STRING with <code_point> or its
11368 * In both cases <*flagp> is appropriately set
11370 * It knows that under FOLD, the Latin Sharp S and UTF characters above
11371 * 255, must be folded (the former only when the rules indicate it can
11374 * When it does the populating, it looks at the flag 'downgradable'. If
11375 * true with a node that folds, it checks if the single code point
11376 * participates in a fold, and if not downgrades the node to an EXACT.
11377 * This helps the optimizer */
11379 bool len_passed_in = cBOOL(len != 0);
11380 U8 character[UTF8_MAXBYTES_CASE+1];
11382 PERL_ARGS_ASSERT_ALLOC_MAYBE_POPULATE_EXACT;
11384 /* Don't bother to check for downgrading in PASS1, as it doesn't make any
11385 * sizing difference, and is extra work that is thrown away */
11386 if (downgradable && ! PASS2) {
11387 downgradable = FALSE;
11390 if (! len_passed_in) {
11392 if (UVCHR_IS_INVARIANT(code_point)) {
11393 if (LOC || ! FOLD) { /* /l defers folding until runtime */
11394 *character = (U8) code_point;
11396 else { /* Here is /i and not /l. (toFOLD() is defined on just
11397 ASCII, which isn't the same thing as INVARIANT on
11398 EBCDIC, but it works there, as the extra invariants
11399 fold to themselves) */
11400 *character = toFOLD((U8) code_point);
11402 /* We can downgrade to an EXACT node if this character
11403 * isn't a folding one. Note that this assumes that
11404 * nothing above Latin1 folds to some other invariant than
11405 * one of these alphabetics; otherwise we would also have
11407 * && (! HAS_NONLATIN1_FOLD_CLOSURE(code_point)
11408 * || ASCII_FOLD_RESTRICTED))
11410 if (downgradable && PL_fold[code_point] == code_point) {
11416 else if (FOLD && (! LOC
11417 || ! is_PROBLEMATIC_LOCALE_FOLD_cp(code_point)))
11418 { /* Folding, and ok to do so now */
11419 UV folded = _to_uni_fold_flags(
11423 FOLD_FLAGS_FULL | ((ASCII_FOLD_RESTRICTED)
11424 ? FOLD_FLAGS_NOMIX_ASCII
11427 && folded == code_point /* This quickly rules out many
11428 cases, avoiding the
11429 _invlist_contains_cp() overhead
11431 && ! _invlist_contains_cp(PL_utf8_foldable, code_point))
11438 else if (code_point <= MAX_UTF8_TWO_BYTE) {
11440 /* Not folding this cp, and can output it directly */
11441 *character = UTF8_TWO_BYTE_HI(code_point);
11442 *(character + 1) = UTF8_TWO_BYTE_LO(code_point);
11446 uvchr_to_utf8( character, code_point);
11447 len = UTF8SKIP(character);
11449 } /* Else pattern isn't UTF8. */
11451 *character = (U8) code_point;
11453 } /* Else is folded non-UTF8 */
11454 else if (LIKELY(code_point != LATIN_SMALL_LETTER_SHARP_S)) {
11456 /* We don't fold any non-UTF8 except possibly the Sharp s (see
11457 * comments at join_exact()); */
11458 *character = (U8) code_point;
11461 /* Can turn into an EXACT node if we know the fold at compile time,
11462 * and it folds to itself and doesn't particpate in other folds */
11465 && PL_fold_latin1[code_point] == code_point
11466 && (! HAS_NONLATIN1_FOLD_CLOSURE(code_point)
11467 || (isASCII(code_point) && ASCII_FOLD_RESTRICTED)))
11471 } /* else is Sharp s. May need to fold it */
11472 else if (AT_LEAST_UNI_SEMANTICS && ! ASCII_FOLD_RESTRICTED) {
11474 *(character + 1) = 's';
11478 *character = LATIN_SMALL_LETTER_SHARP_S;
11484 RExC_size += STR_SZ(len);
11487 RExC_emit += STR_SZ(len);
11488 STR_LEN(node) = len;
11489 if (! len_passed_in) {
11490 Copy((char *) character, STRING(node), len, char);
11494 *flagp |= HASWIDTH;
11496 /* A single character node is SIMPLE, except for the special-cased SHARP S
11498 if ((len == 1 || (UTF && len == UNISKIP(code_point)))
11499 && (code_point != LATIN_SMALL_LETTER_SHARP_S
11500 || ! FOLD || ! DEPENDS_SEMANTICS))
11505 /* The OP may not be well defined in PASS1 */
11506 if (PASS2 && OP(node) == EXACTFL) {
11507 RExC_contains_locale = 1;
11512 /* Parse backref decimal value, unless it's too big to sensibly be a backref,
11513 * in which case return I32_MAX (rather than possibly 32-bit wrapping) */
11516 S_backref_value(char *p)
11518 const char* endptr;
11519 UV val = grok_atou(p, &endptr);
11520 if (endptr == p || endptr == NULL || val > I32_MAX)
11527 - regatom - the lowest level
11529 Try to identify anything special at the start of the pattern. If there
11530 is, then handle it as required. This may involve generating a single regop,
11531 such as for an assertion; or it may involve recursing, such as to
11532 handle a () structure.
11534 If the string doesn't start with something special then we gobble up
11535 as much literal text as we can.
11537 Once we have been able to handle whatever type of thing started the
11538 sequence, we return.
11540 Note: we have to be careful with escapes, as they can be both literal
11541 and special, and in the case of \10 and friends, context determines which.
11543 A summary of the code structure is:
11545 switch (first_byte) {
11546 cases for each special:
11547 handle this special;
11550 switch (2nd byte) {
11551 cases for each unambiguous special:
11552 handle this special;
11554 cases for each ambigous special/literal:
11556 if (special) handle here
11558 default: // unambiguously literal:
11561 default: // is a literal char
11564 create EXACTish node for literal;
11565 while (more input and node isn't full) {
11566 switch (input_byte) {
11567 cases for each special;
11568 make sure parse pointer is set so that the next call to
11569 regatom will see this special first
11570 goto loopdone; // EXACTish node terminated by prev. char
11572 append char to EXACTISH node;
11574 get next input byte;
11578 return the generated node;
11580 Specifically there are two separate switches for handling
11581 escape sequences, with the one for handling literal escapes requiring
11582 a dummy entry for all of the special escapes that are actually handled
11585 Returns NULL, setting *flagp to TRYAGAIN if reg() returns NULL with
11587 Returns NULL, setting *flagp to RESTART_UTF8 if the sizing scan needs to be
11589 Otherwise does not return NULL.
11593 S_regatom(pTHX_ RExC_state_t *pRExC_state, I32 *flagp, U32 depth)
11595 regnode *ret = NULL;
11597 char *parse_start = RExC_parse;
11602 GET_RE_DEBUG_FLAGS_DECL;
11604 *flagp = WORST; /* Tentatively. */
11606 DEBUG_PARSE("atom");
11608 PERL_ARGS_ASSERT_REGATOM;
11611 switch ((U8)*RExC_parse) {
11613 RExC_seen_zerolen++;
11614 nextchar(pRExC_state);
11615 if (RExC_flags & RXf_PMf_MULTILINE)
11616 ret = reg_node(pRExC_state, MBOL);
11618 ret = reg_node(pRExC_state, SBOL);
11619 Set_Node_Length(ret, 1); /* MJD */
11622 nextchar(pRExC_state);
11624 RExC_seen_zerolen++;
11625 if (RExC_flags & RXf_PMf_MULTILINE)
11626 ret = reg_node(pRExC_state, MEOL);
11628 ret = reg_node(pRExC_state, SEOL);
11629 Set_Node_Length(ret, 1); /* MJD */
11632 nextchar(pRExC_state);
11633 if (RExC_flags & RXf_PMf_SINGLELINE)
11634 ret = reg_node(pRExC_state, SANY);
11636 ret = reg_node(pRExC_state, REG_ANY);
11637 *flagp |= HASWIDTH|SIMPLE;
11639 Set_Node_Length(ret, 1); /* MJD */
11643 char * const oregcomp_parse = ++RExC_parse;
11644 ret = regclass(pRExC_state, flagp,depth+1,
11645 FALSE, /* means parse the whole char class */
11646 TRUE, /* allow multi-char folds */
11647 FALSE, /* don't silence non-portable warnings. */
11648 (bool) RExC_strict,
11650 if (*RExC_parse != ']') {
11651 RExC_parse = oregcomp_parse;
11652 vFAIL("Unmatched [");
11655 if (*flagp & RESTART_UTF8)
11657 FAIL2("panic: regclass returned NULL to regatom, flags=%#"UVxf"",
11660 nextchar(pRExC_state);
11661 Set_Node_Length(ret, RExC_parse - oregcomp_parse + 1); /* MJD */
11665 nextchar(pRExC_state);
11666 ret = reg(pRExC_state, 2, &flags,depth+1);
11668 if (flags & TRYAGAIN) {
11669 if (RExC_parse == RExC_end) {
11670 /* Make parent create an empty node if needed. */
11671 *flagp |= TRYAGAIN;
11676 if (flags & RESTART_UTF8) {
11677 *flagp = RESTART_UTF8;
11680 FAIL2("panic: reg returned NULL to regatom, flags=%#"UVxf"",
11683 *flagp |= flags&(HASWIDTH|SPSTART|SIMPLE|POSTPONED);
11687 if (flags & TRYAGAIN) {
11688 *flagp |= TRYAGAIN;
11691 vFAIL("Internal urp");
11692 /* Supposed to be caught earlier. */
11698 vFAIL("Quantifier follows nothing");
11703 This switch handles escape sequences that resolve to some kind
11704 of special regop and not to literal text. Escape sequnces that
11705 resolve to literal text are handled below in the switch marked
11708 Every entry in this switch *must* have a corresponding entry
11709 in the literal escape switch. However, the opposite is not
11710 required, as the default for this switch is to jump to the
11711 literal text handling code.
11713 switch ((U8)*++RExC_parse) {
11714 /* Special Escapes */
11716 RExC_seen_zerolen++;
11717 ret = reg_node(pRExC_state, SBOL);
11718 /* SBOL is shared with /^/ so we set the flags so we can tell
11719 * /\A/ from /^/ in split. We check ret because first pass we
11720 * have no regop struct to set the flags on. */
11724 goto finish_meta_pat;
11726 ret = reg_node(pRExC_state, GPOS);
11727 RExC_seen |= REG_GPOS_SEEN;
11729 goto finish_meta_pat;
11731 RExC_seen_zerolen++;
11732 ret = reg_node(pRExC_state, KEEPS);
11734 /* XXX:dmq : disabling in-place substitution seems to
11735 * be necessary here to avoid cases of memory corruption, as
11736 * with: C<$_="x" x 80; s/x\K/y/> -- rgs
11738 RExC_seen |= REG_LOOKBEHIND_SEEN;
11739 goto finish_meta_pat;
11741 ret = reg_node(pRExC_state, SEOL);
11743 RExC_seen_zerolen++; /* Do not optimize RE away */
11744 goto finish_meta_pat;
11746 ret = reg_node(pRExC_state, EOS);
11748 RExC_seen_zerolen++; /* Do not optimize RE away */
11749 goto finish_meta_pat;
11751 ret = reg_node(pRExC_state, CANY);
11752 RExC_seen |= REG_CANY_SEEN;
11753 *flagp |= HASWIDTH|SIMPLE;
11755 ckWARNdep(RExC_parse+1, "\\C is deprecated");
11757 goto finish_meta_pat;
11759 ret = reg_node(pRExC_state, CLUMP);
11760 *flagp |= HASWIDTH;
11761 goto finish_meta_pat;
11767 arg = ANYOF_WORDCHAR;
11775 regex_charset charset = get_regex_charset(RExC_flags);
11777 RExC_seen_zerolen++;
11778 RExC_seen |= REG_LOOKBEHIND_SEEN;
11779 op = BOUND + charset;
11781 if (op == BOUNDL) {
11782 RExC_contains_locale = 1;
11785 ret = reg_node(pRExC_state, op);
11787 if (*(RExC_parse + 1) != '{') {
11788 FLAGS(ret) = TRADITIONAL_BOUND;
11789 if (PASS2 && op > BOUNDA) { /* /aa is same as /a */
11795 char name = *RExC_parse;
11798 endbrace = strchr(RExC_parse, '}');
11801 vFAIL2("Missing right brace on \\%c{}", name);
11803 /* XXX Need to decide whether to take spaces or not. Should be
11804 * consistent with \p{}, but that currently is SPACE, which
11805 * means vertical too, which seems wrong
11806 * while (isBLANK(*RExC_parse)) {
11809 if (endbrace == RExC_parse) {
11810 RExC_parse++; /* After the '}' */
11811 vFAIL2("Empty \\%c{}", name);
11813 length = endbrace - RExC_parse;
11814 /*while (isBLANK(*(RExC_parse + length - 1))) {
11817 switch (*RExC_parse) {
11820 && (length != 3 || strnNE(RExC_parse + 1, "cb", 2)))
11822 goto bad_bound_type;
11824 FLAGS(ret) = GCB_BOUND;
11827 if (length != 2 || *(RExC_parse + 1) != 'b') {
11828 goto bad_bound_type;
11830 FLAGS(ret) = SB_BOUND;
11833 if (length != 2 || *(RExC_parse + 1) != 'b') {
11834 goto bad_bound_type;
11836 FLAGS(ret) = WB_BOUND;
11840 RExC_parse = endbrace;
11842 "'%"UTF8f"' is an unknown bound type",
11843 UTF8fARG(UTF, length, endbrace - length));
11844 NOT_REACHED; /*NOTREACHED*/
11846 RExC_parse = endbrace;
11847 RExC_uni_semantics = 1;
11849 if (PASS2 && op >= BOUNDA) { /* /aa is same as /a */
11853 /* Don't have to worry about UTF-8, in this message because
11854 * to get here the contents of the \b must be ASCII */
11855 ckWARN4reg(RExC_parse + 1, /* Include the '}' in msg */
11856 "Using /u for '%.*s' instead of /%s",
11858 endbrace - length + 1,
11859 (charset == REGEX_ASCII_RESTRICTED_CHARSET)
11860 ? ASCII_RESTRICT_PAT_MODS
11861 : ASCII_MORE_RESTRICT_PAT_MODS);
11865 if (PASS2 && invert) {
11866 OP(ret) += NBOUND - BOUND;
11868 goto finish_meta_pat;
11876 if (! DEPENDS_SEMANTICS) {
11880 /* \d doesn't have any matches in the upper Latin1 range, hence /d
11881 * is equivalent to /u. Changing to /u saves some branches at
11884 goto join_posix_op_known;
11887 ret = reg_node(pRExC_state, LNBREAK);
11888 *flagp |= HASWIDTH|SIMPLE;
11889 goto finish_meta_pat;
11897 goto join_posix_op_known;
11903 arg = ANYOF_VERTWS;
11905 goto join_posix_op_known;
11915 op = POSIXD + get_regex_charset(RExC_flags);
11916 if (op > POSIXA) { /* /aa is same as /a */
11919 else if (op == POSIXL) {
11920 RExC_contains_locale = 1;
11923 join_posix_op_known:
11926 op += NPOSIXD - POSIXD;
11929 ret = reg_node(pRExC_state, op);
11931 FLAGS(ret) = namedclass_to_classnum(arg);
11934 *flagp |= HASWIDTH|SIMPLE;
11938 nextchar(pRExC_state);
11939 Set_Node_Length(ret, 2); /* MJD */
11945 char* parse_start = RExC_parse - 2;
11950 ret = regclass(pRExC_state, flagp,depth+1,
11951 TRUE, /* means just parse this element */
11952 FALSE, /* don't allow multi-char folds */
11953 FALSE, /* don't silence non-portable warnings.
11954 It would be a bug if these returned
11956 (bool) RExC_strict,
11958 /* regclass() can only return RESTART_UTF8 if multi-char folds
11961 FAIL2("panic: regclass returned NULL to regatom, flags=%#"UVxf"",
11966 Set_Node_Offset(ret, parse_start + 2);
11967 Set_Node_Cur_Length(ret, parse_start);
11968 nextchar(pRExC_state);
11972 /* Handle \N and \N{NAME} with multiple code points here and not
11973 * below because it can be multicharacter. join_exact() will join
11974 * them up later on. Also this makes sure that things like
11975 * /\N{BLAH}+/ and \N{BLAH} being multi char Just Happen. dmq.
11976 * The options to the grok function call causes it to fail if the
11977 * sequence is just a single code point. We then go treat it as
11978 * just another character in the current EXACT node, and hence it
11979 * gets uniform treatment with all the other characters. The
11980 * special treatment for quantifiers is not needed for such single
11981 * character sequences */
11983 if ((STRLEN) -1 == grok_bslash_N(pRExC_state, &ret, NULL, flagp,
11986 if (*flagp & RESTART_UTF8)
11992 case 'k': /* Handle \k<NAME> and \k'NAME' */
11995 char ch= RExC_parse[1];
11996 if (ch != '<' && ch != '\'' && ch != '{') {
11998 /* diag_listed_as: Sequence \%s... not terminated in regex; marked by <-- HERE in m/%s/ */
11999 vFAIL2("Sequence %.2s... not terminated",parse_start);
12001 /* this pretty much dupes the code for (?P=...) in reg(), if
12002 you change this make sure you change that */
12003 char* name_start = (RExC_parse += 2);
12005 SV *sv_dat = reg_scan_name(pRExC_state,
12006 SIZE_ONLY ? REG_RSN_RETURN_NULL : REG_RSN_RETURN_DATA);
12007 ch= (ch == '<') ? '>' : (ch == '{') ? '}' : '\'';
12008 if (RExC_parse == name_start || *RExC_parse != ch)
12009 /* diag_listed_as: Sequence \%s... not terminated in regex; marked by <-- HERE in m/%s/ */
12010 vFAIL2("Sequence %.3s... not terminated",parse_start);
12013 num = add_data( pRExC_state, STR_WITH_LEN("S"));
12014 RExC_rxi->data->data[num]=(void*)sv_dat;
12015 SvREFCNT_inc_simple_void(sv_dat);
12019 ret = reganode(pRExC_state,
12022 : (ASCII_FOLD_RESTRICTED)
12024 : (AT_LEAST_UNI_SEMANTICS)
12030 *flagp |= HASWIDTH;
12032 /* override incorrect value set in reganode MJD */
12033 Set_Node_Offset(ret, parse_start+1);
12034 Set_Node_Cur_Length(ret, parse_start);
12035 nextchar(pRExC_state);
12041 case '1': case '2': case '3': case '4':
12042 case '5': case '6': case '7': case '8': case '9':
12047 if (*RExC_parse == 'g') {
12051 if (*RExC_parse == '{') {
12055 if (*RExC_parse == '-') {
12059 if (hasbrace && !isDIGIT(*RExC_parse)) {
12060 if (isrel) RExC_parse--;
12062 goto parse_named_seq;
12065 num = S_backref_value(RExC_parse);
12067 vFAIL("Reference to invalid group 0");
12068 else if (num == I32_MAX) {
12069 if (isDIGIT(*RExC_parse))
12070 vFAIL("Reference to nonexistent group");
12072 vFAIL("Unterminated \\g... pattern");
12076 num = RExC_npar - num;
12078 vFAIL("Reference to nonexistent or unclosed group");
12082 num = S_backref_value(RExC_parse);
12083 /* bare \NNN might be backref or octal - if it is larger
12084 * than or equal RExC_npar then it is assumed to be an
12085 * octal escape. Note RExC_npar is +1 from the actual
12086 * number of parens. */
12087 /* Note we do NOT check if num == I32_MAX here, as that is
12088 * handled by the RExC_npar check */
12091 /* any numeric escape < 10 is always a backref */
12093 /* any numeric escape < RExC_npar is a backref */
12094 && num >= RExC_npar
12095 /* cannot be an octal escape if it starts with 8 */
12096 && *RExC_parse != '8'
12097 /* cannot be an octal escape it it starts with 9 */
12098 && *RExC_parse != '9'
12101 /* Probably not a backref, instead likely to be an
12102 * octal character escape, e.g. \35 or \777.
12103 * The above logic should make it obvious why using
12104 * octal escapes in patterns is problematic. - Yves */
12109 /* At this point RExC_parse points at a numeric escape like
12110 * \12 or \88 or something similar, which we should NOT treat
12111 * as an octal escape. It may or may not be a valid backref
12112 * escape. For instance \88888888 is unlikely to be a valid
12115 #ifdef RE_TRACK_PATTERN_OFFSETS
12116 char * const parse_start = RExC_parse - 1; /* MJD */
12118 while (isDIGIT(*RExC_parse))
12121 if (*RExC_parse != '}')
12122 vFAIL("Unterminated \\g{...} pattern");
12126 if (num > (I32)RExC_rx->nparens)
12127 vFAIL("Reference to nonexistent group");
12130 ret = reganode(pRExC_state,
12133 : (ASCII_FOLD_RESTRICTED)
12135 : (AT_LEAST_UNI_SEMANTICS)
12141 *flagp |= HASWIDTH;
12143 /* override incorrect value set in reganode MJD */
12144 Set_Node_Offset(ret, parse_start+1);
12145 Set_Node_Cur_Length(ret, parse_start);
12147 nextchar(pRExC_state);
12152 if (RExC_parse >= RExC_end)
12153 FAIL("Trailing \\");
12156 /* Do not generate "unrecognized" warnings here, we fall
12157 back into the quick-grab loop below */
12164 if (RExC_flags & RXf_PMf_EXTENDED) {
12165 RExC_parse = reg_skipcomment( pRExC_state, RExC_parse );
12166 if (RExC_parse < RExC_end)
12173 parse_start = RExC_parse - 1;
12182 #define MAX_NODE_STRING_SIZE 127
12183 char foldbuf[MAX_NODE_STRING_SIZE+UTF8_MAXBYTES_CASE];
12185 U8 upper_parse = MAX_NODE_STRING_SIZE;
12186 U8 node_type = compute_EXACTish(pRExC_state);
12187 bool next_is_quantifier;
12188 char * oldp = NULL;
12190 /* We can convert EXACTF nodes to EXACTFU if they contain only
12191 * characters that match identically regardless of the target
12192 * string's UTF8ness. The reason to do this is that EXACTF is not
12193 * trie-able, EXACTFU is.
12195 * Similarly, we can convert EXACTFL nodes to EXACTFU if they
12196 * contain only above-Latin1 characters (hence must be in UTF8),
12197 * which don't participate in folds with Latin1-range characters,
12198 * as the latter's folds aren't known until runtime. (We don't
12199 * need to figure this out until pass 2) */
12200 bool maybe_exactfu = PASS2
12201 && (node_type == EXACTF || node_type == EXACTFL);
12203 /* If a folding node contains only code points that don't
12204 * participate in folds, it can be changed into an EXACT node,
12205 * which allows the optimizer more things to look for */
12208 ret = reg_node(pRExC_state, node_type);
12210 /* In pass1, folded, we use a temporary buffer instead of the
12211 * actual node, as the node doesn't exist yet */
12212 s = (SIZE_ONLY && FOLD) ? foldbuf : STRING(ret);
12218 /* We do the EXACTFish to EXACT node only if folding. (And we
12219 * don't need to figure this out until pass 2) */
12220 maybe_exact = FOLD && PASS2;
12222 /* XXX The node can hold up to 255 bytes, yet this only goes to
12223 * 127. I (khw) do not know why. Keeping it somewhat less than
12224 * 255 allows us to not have to worry about overflow due to
12225 * converting to utf8 and fold expansion, but that value is
12226 * 255-UTF8_MAXBYTES_CASE. join_exact() may join adjacent nodes
12227 * split up by this limit into a single one using the real max of
12228 * 255. Even at 127, this breaks under rare circumstances. If
12229 * folding, we do not want to split a node at a character that is a
12230 * non-final in a multi-char fold, as an input string could just
12231 * happen to want to match across the node boundary. The join
12232 * would solve that problem if the join actually happens. But a
12233 * series of more than two nodes in a row each of 127 would cause
12234 * the first join to succeed to get to 254, but then there wouldn't
12235 * be room for the next one, which could at be one of those split
12236 * multi-char folds. I don't know of any fool-proof solution. One
12237 * could back off to end with only a code point that isn't such a
12238 * non-final, but it is possible for there not to be any in the
12240 for (p = RExC_parse - 1;
12241 len < upper_parse && p < RExC_end;
12246 if (RExC_flags & RXf_PMf_EXTENDED)
12247 p = regpatws(pRExC_state, p,
12248 TRUE); /* means recognize comments */
12259 /* Literal Escapes Switch
12261 This switch is meant to handle escape sequences that
12262 resolve to a literal character.
12264 Every escape sequence that represents something
12265 else, like an assertion or a char class, is handled
12266 in the switch marked 'Special Escapes' above in this
12267 routine, but also has an entry here as anything that
12268 isn't explicitly mentioned here will be treated as
12269 an unescaped equivalent literal.
12272 switch ((U8)*++p) {
12273 /* These are all the special escapes. */
12274 case 'A': /* Start assertion */
12275 case 'b': case 'B': /* Word-boundary assertion*/
12276 case 'C': /* Single char !DANGEROUS! */
12277 case 'd': case 'D': /* digit class */
12278 case 'g': case 'G': /* generic-backref, pos assertion */
12279 case 'h': case 'H': /* HORIZWS */
12280 case 'k': case 'K': /* named backref, keep marker */
12281 case 'p': case 'P': /* Unicode property */
12282 case 'R': /* LNBREAK */
12283 case 's': case 'S': /* space class */
12284 case 'v': case 'V': /* VERTWS */
12285 case 'w': case 'W': /* word class */
12286 case 'X': /* eXtended Unicode "combining
12287 character sequence" */
12288 case 'z': case 'Z': /* End of line/string assertion */
12292 /* Anything after here is an escape that resolves to a
12293 literal. (Except digits, which may or may not)
12299 case 'N': /* Handle a single-code point named character. */
12300 /* The options cause it to fail if a multiple code
12301 * point sequence. Handle those in the switch() above
12303 RExC_parse = p + 1;
12304 if ((STRLEN) -1 == grok_bslash_N(pRExC_state, NULL,
12310 if (*flagp & RESTART_UTF8)
12311 FAIL("panic: grok_bslash_N set RESTART_UTF8");
12312 RExC_parse = p = oldp;
12316 if (ender > 0xff) {
12333 ender = ESC_NATIVE;
12343 const char* error_msg;
12345 bool valid = grok_bslash_o(&p,
12348 PASS2, /* out warnings */
12349 (bool) RExC_strict,
12350 TRUE, /* Output warnings
12355 RExC_parse = p; /* going to die anyway; point
12356 to exact spot of failure */
12360 if (IN_ENCODING && ender < 0x100) {
12361 goto recode_encoding;
12363 if (ender > 0xff) {
12370 UV result = UV_MAX; /* initialize to erroneous
12372 const char* error_msg;
12374 bool valid = grok_bslash_x(&p,
12377 PASS2, /* out warnings */
12378 (bool) RExC_strict,
12379 TRUE, /* Silence warnings
12384 RExC_parse = p; /* going to die anyway; point
12385 to exact spot of failure */
12390 if (IN_ENCODING && ender < 0x100) {
12391 goto recode_encoding;
12393 if (ender > 0xff) {
12400 ender = grok_bslash_c(*p++, PASS2);
12402 case '8': case '9': /* must be a backreference */
12404 /* we have an escape like \8 which cannot be an octal escape
12405 * so we exit the loop, and let the outer loop handle this
12406 * escape which may or may not be a legitimate backref. */
12408 case '1': case '2': case '3':case '4':
12409 case '5': case '6': case '7':
12410 /* When we parse backslash escapes there is ambiguity
12411 * between backreferences and octal escapes. Any escape
12412 * from \1 - \9 is a backreference, any multi-digit
12413 * escape which does not start with 0 and which when
12414 * evaluated as decimal could refer to an already
12415 * parsed capture buffer is a back reference. Anything
12418 * Note this implies that \118 could be interpreted as
12419 * 118 OR as "\11" . "8" depending on whether there
12420 * were 118 capture buffers defined already in the
12423 /* NOTE, RExC_npar is 1 more than the actual number of
12424 * parens we have seen so far, hence the < RExC_npar below. */
12426 if ( !isDIGIT(p[1]) || S_backref_value(p) < RExC_npar)
12427 { /* Not to be treated as an octal constant, go
12435 I32 flags = PERL_SCAN_SILENT_ILLDIGIT;
12437 ender = grok_oct(p, &numlen, &flags, NULL);
12438 if (ender > 0xff) {
12442 if (PASS2 /* like \08, \178 */
12445 && isDIGIT(*p) && ckWARN(WARN_REGEXP))
12447 reg_warn_non_literal_string(
12449 form_short_octal_warning(p, numlen));
12452 if (IN_ENCODING && ender < 0x100)
12453 goto recode_encoding;
12456 if (! RExC_override_recoding) {
12457 SV* enc = _get_encoding();
12458 ender = reg_recode((const char)(U8)ender, &enc);
12460 ckWARNreg(p, "Invalid escape in the specified encoding");
12466 FAIL("Trailing \\");
12469 if (!SIZE_ONLY&& isALPHANUMERIC(*p)) {
12470 /* Include any { following the alpha to emphasize
12471 * that it could be part of an escape at some point
12473 int len = (isALPHA(*p) && *(p + 1) == '{') ? 2 : 1;
12474 ckWARN3reg(p + len, "Unrecognized escape \\%.*s passed through", len, p);
12476 goto normal_default;
12477 } /* End of switch on '\' */
12480 /* Currently we don't warn when the lbrace is at the start
12481 * of a construct. This catches it in the middle of a
12482 * literal string, or when its the first thing after
12483 * something like "\b" */
12485 && (len || (p > RExC_start && isALPHA_A(*(p -1)))))
12487 ckWARNregdep(p + 1, "Unescaped left brace in regex is deprecated, passed through");
12490 default: /* A literal character */
12492 if (UTF8_IS_START(*p) && UTF) {
12494 ender = utf8n_to_uvchr((U8*)p, RExC_end - p,
12495 &numlen, UTF8_ALLOW_DEFAULT);
12501 } /* End of switch on the literal */
12503 /* Here, have looked at the literal character and <ender>
12504 * contains its ordinal, <p> points to the character after it
12507 if ( RExC_flags & RXf_PMf_EXTENDED)
12508 p = regpatws(pRExC_state, p,
12509 TRUE); /* means recognize comments */
12511 /* If the next thing is a quantifier, it applies to this
12512 * character only, which means that this character has to be in
12513 * its own node and can't just be appended to the string in an
12514 * existing node, so if there are already other characters in
12515 * the node, close the node with just them, and set up to do
12516 * this character again next time through, when it will be the
12517 * only thing in its new node */
12518 if ((next_is_quantifier = (p < RExC_end && ISMULT2(p))) && len)
12524 if (! FOLD) { /* The simple case, just append the literal */
12526 /* In the sizing pass, we need only the size of the
12527 * character we are appending, hence we can delay getting
12528 * its representation until PASS2. */
12531 const STRLEN unilen = UNISKIP(ender);
12534 /* We have to subtract 1 just below (and again in
12535 * the corresponding PASS2 code) because the loop
12536 * increments <len> each time, as all but this path
12537 * (and one other) through it add a single byte to
12538 * the EXACTish node. But these paths would change
12539 * len to be the correct final value, so cancel out
12540 * the increment that follows */
12546 } else { /* PASS2 */
12549 U8 * new_s = uvchr_to_utf8((U8*)s, ender);
12550 len += (char *) new_s - s - 1;
12551 s = (char *) new_s;
12554 *(s++) = (char) ender;
12558 else if (LOC && is_PROBLEMATIC_LOCALE_FOLD_cp(ender)) {
12560 /* Here are folding under /l, and the code point is
12561 * problematic. First, we know we can't simplify things */
12562 maybe_exact = FALSE;
12563 maybe_exactfu = FALSE;
12565 /* A problematic code point in this context means that its
12566 * fold isn't known until runtime, so we can't fold it now.
12567 * (The non-problematic code points are the above-Latin1
12568 * ones that fold to also all above-Latin1. Their folds
12569 * don't vary no matter what the locale is.) But here we
12570 * have characters whose fold depends on the locale.
12571 * Unlike the non-folding case above, we have to keep track
12572 * of these in the sizing pass, so that we can make sure we
12573 * don't split too-long nodes in the middle of a potential
12574 * multi-char fold. And unlike the regular fold case
12575 * handled in the else clauses below, we don't actually
12576 * fold and don't have special cases to consider. What we
12577 * do for both passes is the PASS2 code for non-folding */
12578 goto not_fold_common;
12580 else /* A regular FOLD code point */
12582 /* See comments for join_exact() as to why we fold this
12583 * non-UTF at compile time */
12584 || (node_type == EXACTFU
12585 && ender == LATIN_SMALL_LETTER_SHARP_S)))
12587 /* Here, are folding and are not UTF-8 encoded; therefore
12588 * the character must be in the range 0-255, and is not /l
12589 * (Not /l because we already handled these under /l in
12590 * is_PROBLEMATIC_LOCALE_FOLD_cp) */
12591 if (IS_IN_SOME_FOLD_L1(ender)) {
12592 maybe_exact = FALSE;
12594 /* See if the character's fold differs between /d and
12595 * /u. This includes the multi-char fold SHARP S to
12598 && (PL_fold[ender] != PL_fold_latin1[ender]
12599 || ender == LATIN_SMALL_LETTER_SHARP_S
12601 && isALPHA_FOLD_EQ(ender, 's')
12602 && isALPHA_FOLD_EQ(*(s-1), 's'))))
12604 maybe_exactfu = FALSE;
12608 /* Even when folding, we store just the input character, as
12609 * we have an array that finds its fold quickly */
12610 *(s++) = (char) ender;
12612 else { /* FOLD and UTF */
12613 /* Unlike the non-fold case, we do actually have to
12614 * calculate the results here in pass 1. This is for two
12615 * reasons, the folded length may be longer than the
12616 * unfolded, and we have to calculate how many EXACTish
12617 * nodes it will take; and we may run out of room in a node
12618 * in the middle of a potential multi-char fold, and have
12619 * to back off accordingly. */
12622 if (isASCII_uni(ender)) {
12623 folded = toFOLD(ender);
12624 *(s)++ = (U8) folded;
12629 folded = _to_uni_fold_flags(
12633 FOLD_FLAGS_FULL | ((ASCII_FOLD_RESTRICTED)
12634 ? FOLD_FLAGS_NOMIX_ASCII
12638 /* The loop increments <len> each time, as all but this
12639 * path (and one other) through it add a single byte to
12640 * the EXACTish node. But this one has changed len to
12641 * be the correct final value, so subtract one to
12642 * cancel out the increment that follows */
12643 len += foldlen - 1;
12645 /* If this node only contains non-folding code points so
12646 * far, see if this new one is also non-folding */
12648 if (folded != ender) {
12649 maybe_exact = FALSE;
12652 /* Here the fold is the original; we have to check
12653 * further to see if anything folds to it */
12654 if (_invlist_contains_cp(PL_utf8_foldable,
12657 maybe_exact = FALSE;
12664 if (next_is_quantifier) {
12666 /* Here, the next input is a quantifier, and to get here,
12667 * the current character is the only one in the node.
12668 * Also, here <len> doesn't include the final byte for this
12674 } /* End of loop through literal characters */
12676 /* Here we have either exhausted the input or ran out of room in
12677 * the node. (If we encountered a character that can't be in the
12678 * node, transfer is made directly to <loopdone>, and so we
12679 * wouldn't have fallen off the end of the loop.) In the latter
12680 * case, we artificially have to split the node into two, because
12681 * we just don't have enough space to hold everything. This
12682 * creates a problem if the final character participates in a
12683 * multi-character fold in the non-final position, as a match that
12684 * should have occurred won't, due to the way nodes are matched,
12685 * and our artificial boundary. So back off until we find a non-
12686 * problematic character -- one that isn't at the beginning or
12687 * middle of such a fold. (Either it doesn't participate in any
12688 * folds, or appears only in the final position of all the folds it
12689 * does participate in.) A better solution with far fewer false
12690 * positives, and that would fill the nodes more completely, would
12691 * be to actually have available all the multi-character folds to
12692 * test against, and to back-off only far enough to be sure that
12693 * this node isn't ending with a partial one. <upper_parse> is set
12694 * further below (if we need to reparse the node) to include just
12695 * up through that final non-problematic character that this code
12696 * identifies, so when it is set to less than the full node, we can
12697 * skip the rest of this */
12698 if (FOLD && p < RExC_end && upper_parse == MAX_NODE_STRING_SIZE) {
12700 const STRLEN full_len = len;
12702 assert(len >= MAX_NODE_STRING_SIZE);
12704 /* Here, <s> points to the final byte of the final character.
12705 * Look backwards through the string until find a non-
12706 * problematic character */
12710 /* This has no multi-char folds to non-UTF characters */
12711 if (ASCII_FOLD_RESTRICTED) {
12715 while (--s >= s0 && IS_NON_FINAL_FOLD(*s)) { }
12719 if (! PL_NonL1NonFinalFold) {
12720 PL_NonL1NonFinalFold = _new_invlist_C_array(
12721 NonL1_Perl_Non_Final_Folds_invlist);
12724 /* Point to the first byte of the final character */
12725 s = (char *) utf8_hop((U8 *) s, -1);
12727 while (s >= s0) { /* Search backwards until find
12728 non-problematic char */
12729 if (UTF8_IS_INVARIANT(*s)) {
12731 /* There are no ascii characters that participate
12732 * in multi-char folds under /aa. In EBCDIC, the
12733 * non-ascii invariants are all control characters,
12734 * so don't ever participate in any folds. */
12735 if (ASCII_FOLD_RESTRICTED
12736 || ! IS_NON_FINAL_FOLD(*s))
12741 else if (UTF8_IS_DOWNGRADEABLE_START(*s)) {
12742 if (! IS_NON_FINAL_FOLD(TWO_BYTE_UTF8_TO_NATIVE(
12748 else if (! _invlist_contains_cp(
12749 PL_NonL1NonFinalFold,
12750 valid_utf8_to_uvchr((U8 *) s, NULL)))
12755 /* Here, the current character is problematic in that
12756 * it does occur in the non-final position of some
12757 * fold, so try the character before it, but have to
12758 * special case the very first byte in the string, so
12759 * we don't read outside the string */
12760 s = (s == s0) ? s -1 : (char *) utf8_hop((U8 *) s, -1);
12761 } /* End of loop backwards through the string */
12763 /* If there were only problematic characters in the string,
12764 * <s> will point to before s0, in which case the length
12765 * should be 0, otherwise include the length of the
12766 * non-problematic character just found */
12767 len = (s < s0) ? 0 : s - s0 + UTF8SKIP(s);
12770 /* Here, have found the final character, if any, that is
12771 * non-problematic as far as ending the node without splitting
12772 * it across a potential multi-char fold. <len> contains the
12773 * number of bytes in the node up-to and including that
12774 * character, or is 0 if there is no such character, meaning
12775 * the whole node contains only problematic characters. In
12776 * this case, give up and just take the node as-is. We can't
12781 /* If the node ends in an 's' we make sure it stays EXACTF,
12782 * as if it turns into an EXACTFU, it could later get
12783 * joined with another 's' that would then wrongly match
12785 if (maybe_exactfu && isALPHA_FOLD_EQ(ender, 's'))
12787 maybe_exactfu = FALSE;
12791 /* Here, the node does contain some characters that aren't
12792 * problematic. If one such is the final character in the
12793 * node, we are done */
12794 if (len == full_len) {
12797 else if (len + ((UTF) ? UTF8SKIP(s) : 1) == full_len) {
12799 /* If the final character is problematic, but the
12800 * penultimate is not, back-off that last character to
12801 * later start a new node with it */
12806 /* Here, the final non-problematic character is earlier
12807 * in the input than the penultimate character. What we do
12808 * is reparse from the beginning, going up only as far as
12809 * this final ok one, thus guaranteeing that the node ends
12810 * in an acceptable character. The reason we reparse is
12811 * that we know how far in the character is, but we don't
12812 * know how to correlate its position with the input parse.
12813 * An alternate implementation would be to build that
12814 * correlation as we go along during the original parse,
12815 * but that would entail extra work for every node, whereas
12816 * this code gets executed only when the string is too
12817 * large for the node, and the final two characters are
12818 * problematic, an infrequent occurrence. Yet another
12819 * possible strategy would be to save the tail of the
12820 * string, and the next time regatom is called, initialize
12821 * with that. The problem with this is that unless you
12822 * back off one more character, you won't be guaranteed
12823 * regatom will get called again, unless regbranch,
12824 * regpiece ... are also changed. If you do back off that
12825 * extra character, so that there is input guaranteed to
12826 * force calling regatom, you can't handle the case where
12827 * just the first character in the node is acceptable. I
12828 * (khw) decided to try this method which doesn't have that
12829 * pitfall; if performance issues are found, we can do a
12830 * combination of the current approach plus that one */
12836 } /* End of verifying node ends with an appropriate char */
12838 loopdone: /* Jumped to when encounters something that shouldn't be
12841 /* I (khw) don't know if you can get here with zero length, but the
12842 * old code handled this situation by creating a zero-length EXACT
12843 * node. Might as well be NOTHING instead */
12849 /* If 'maybe_exact' is still set here, means there are no
12850 * code points in the node that participate in folds;
12851 * similarly for 'maybe_exactfu' and code points that match
12852 * differently depending on UTF8ness of the target string
12853 * (for /u), or depending on locale for /l */
12859 else if (maybe_exactfu) {
12865 alloc_maybe_populate_EXACT(pRExC_state, ret, flagp, len, ender,
12866 FALSE /* Don't look to see if could
12867 be turned into an EXACT
12868 node, as we have already
12873 RExC_parse = p - 1;
12874 Set_Node_Cur_Length(ret, parse_start);
12875 nextchar(pRExC_state);
12877 /* len is STRLEN which is unsigned, need to copy to signed */
12880 vFAIL("Internal disaster");
12883 } /* End of label 'defchar:' */
12885 } /* End of giant switch on input character */
12891 S_regpatws(RExC_state_t *pRExC_state, char *p , const bool recognize_comment )
12893 /* Returns the next non-pattern-white space, non-comment character (the
12894 * latter only if 'recognize_comment is true) in the string p, which is
12895 * ended by RExC_end. See also reg_skipcomment */
12896 const char *e = RExC_end;
12898 PERL_ARGS_ASSERT_REGPATWS;
12902 if ((len = is_PATWS_safe(p, e, UTF))) {
12905 else if (recognize_comment && *p == '#') {
12906 p = reg_skipcomment(pRExC_state, p);
12915 S_populate_ANYOF_from_invlist(pTHX_ regnode *node, SV** invlist_ptr)
12917 /* Uses the inversion list '*invlist_ptr' to populate the ANYOF 'node'. It
12918 * sets up the bitmap and any flags, removing those code points from the
12919 * inversion list, setting it to NULL should it become completely empty */
12921 PERL_ARGS_ASSERT_POPULATE_ANYOF_FROM_INVLIST;
12922 assert(PL_regkind[OP(node)] == ANYOF);
12924 ANYOF_BITMAP_ZERO(node);
12925 if (*invlist_ptr) {
12927 /* This gets set if we actually need to modify things */
12928 bool change_invlist = FALSE;
12932 /* Start looking through *invlist_ptr */
12933 invlist_iterinit(*invlist_ptr);
12934 while (invlist_iternext(*invlist_ptr, &start, &end)) {
12938 if (end == UV_MAX && start <= NUM_ANYOF_CODE_POINTS) {
12939 ANYOF_FLAGS(node) |= ANYOF_MATCHES_ALL_ABOVE_BITMAP;
12941 else if (end >= NUM_ANYOF_CODE_POINTS) {
12942 ANYOF_FLAGS(node) |= ANYOF_HAS_UTF8_NONBITMAP_MATCHES;
12945 /* Quit if are above what we should change */
12946 if (start >= NUM_ANYOF_CODE_POINTS) {
12950 change_invlist = TRUE;
12952 /* Set all the bits in the range, up to the max that we are doing */
12953 high = (end < NUM_ANYOF_CODE_POINTS - 1)
12955 : NUM_ANYOF_CODE_POINTS - 1;
12956 for (i = start; i <= (int) high; i++) {
12957 if (! ANYOF_BITMAP_TEST(node, i)) {
12958 ANYOF_BITMAP_SET(node, i);
12962 invlist_iterfinish(*invlist_ptr);
12964 /* Done with loop; remove any code points that are in the bitmap from
12965 * *invlist_ptr; similarly for code points above the bitmap if we have
12966 * a flag to match all of them anyways */
12967 if (change_invlist) {
12968 _invlist_subtract(*invlist_ptr, PL_InBitmap, invlist_ptr);
12970 if (ANYOF_FLAGS(node) & ANYOF_MATCHES_ALL_ABOVE_BITMAP) {
12971 _invlist_intersection(*invlist_ptr, PL_InBitmap, invlist_ptr);
12974 /* If have completely emptied it, remove it completely */
12975 if (_invlist_len(*invlist_ptr) == 0) {
12976 SvREFCNT_dec_NN(*invlist_ptr);
12977 *invlist_ptr = NULL;
12982 /* Parse POSIX character classes: [[:foo:]], [[=foo=]], [[.foo.]].
12983 Character classes ([:foo:]) can also be negated ([:^foo:]).
12984 Returns a named class id (ANYOF_XXX) if successful, -1 otherwise.
12985 Equivalence classes ([=foo=]) and composites ([.foo.]) are parsed,
12986 but trigger failures because they are currently unimplemented. */
12988 #define POSIXCC_DONE(c) ((c) == ':')
12989 #define POSIXCC_NOTYET(c) ((c) == '=' || (c) == '.')
12990 #define POSIXCC(c) (POSIXCC_DONE(c) || POSIXCC_NOTYET(c))
12992 PERL_STATIC_INLINE I32
12993 S_regpposixcc(pTHX_ RExC_state_t *pRExC_state, I32 value, const bool strict)
12995 I32 namedclass = OOB_NAMEDCLASS;
12997 PERL_ARGS_ASSERT_REGPPOSIXCC;
12999 if (value == '[' && RExC_parse + 1 < RExC_end &&
13000 /* I smell either [: or [= or [. -- POSIX has been here, right? */
13001 POSIXCC(UCHARAT(RExC_parse)))
13003 const char c = UCHARAT(RExC_parse);
13004 char* const s = RExC_parse++;
13006 while (RExC_parse < RExC_end && UCHARAT(RExC_parse) != c)
13008 if (RExC_parse == RExC_end) {
13011 /* Try to give a better location for the error (than the end of
13012 * the string) by looking for the matching ']' */
13014 while (RExC_parse < RExC_end && UCHARAT(RExC_parse) != ']') {
13017 vFAIL2("Unmatched '%c' in POSIX class", c);
13019 /* Grandfather lone [:, [=, [. */
13023 const char* const t = RExC_parse++; /* skip over the c */
13026 if (UCHARAT(RExC_parse) == ']') {
13027 const char *posixcc = s + 1;
13028 RExC_parse++; /* skip over the ending ] */
13031 const I32 complement = *posixcc == '^' ? *posixcc++ : 0;
13032 const I32 skip = t - posixcc;
13034 /* Initially switch on the length of the name. */
13037 if (memEQ(posixcc, "word", 4)) /* this is not POSIX,
13038 this is the Perl \w
13040 namedclass = ANYOF_WORDCHAR;
13043 /* Names all of length 5. */
13044 /* alnum alpha ascii blank cntrl digit graph lower
13045 print punct space upper */
13046 /* Offset 4 gives the best switch position. */
13047 switch (posixcc[4]) {
13049 if (memEQ(posixcc, "alph", 4)) /* alpha */
13050 namedclass = ANYOF_ALPHA;
13053 if (memEQ(posixcc, "spac", 4)) /* space */
13054 namedclass = ANYOF_SPACE;
13057 if (memEQ(posixcc, "grap", 4)) /* graph */
13058 namedclass = ANYOF_GRAPH;
13061 if (memEQ(posixcc, "asci", 4)) /* ascii */
13062 namedclass = ANYOF_ASCII;
13065 if (memEQ(posixcc, "blan", 4)) /* blank */
13066 namedclass = ANYOF_BLANK;
13069 if (memEQ(posixcc, "cntr", 4)) /* cntrl */
13070 namedclass = ANYOF_CNTRL;
13073 if (memEQ(posixcc, "alnu", 4)) /* alnum */
13074 namedclass = ANYOF_ALPHANUMERIC;
13077 if (memEQ(posixcc, "lowe", 4)) /* lower */
13078 namedclass = (FOLD) ? ANYOF_CASED : ANYOF_LOWER;
13079 else if (memEQ(posixcc, "uppe", 4)) /* upper */
13080 namedclass = (FOLD) ? ANYOF_CASED : ANYOF_UPPER;
13083 if (memEQ(posixcc, "digi", 4)) /* digit */
13084 namedclass = ANYOF_DIGIT;
13085 else if (memEQ(posixcc, "prin", 4)) /* print */
13086 namedclass = ANYOF_PRINT;
13087 else if (memEQ(posixcc, "punc", 4)) /* punct */
13088 namedclass = ANYOF_PUNCT;
13093 if (memEQ(posixcc, "xdigit", 6))
13094 namedclass = ANYOF_XDIGIT;
13098 if (namedclass == OOB_NAMEDCLASS)
13100 "POSIX class [:%"UTF8f":] unknown",
13101 UTF8fARG(UTF, t - s - 1, s + 1));
13103 /* The #defines are structured so each complement is +1 to
13104 * the normal one */
13108 assert (posixcc[skip] == ':');
13109 assert (posixcc[skip+1] == ']');
13110 } else if (!SIZE_ONLY) {
13111 /* [[=foo=]] and [[.foo.]] are still future. */
13113 /* adjust RExC_parse so the warning shows after
13114 the class closes */
13115 while (UCHARAT(RExC_parse) && UCHARAT(RExC_parse) != ']')
13117 vFAIL3("POSIX syntax [%c %c] is reserved for future extensions", c, c);
13120 /* Maternal grandfather:
13121 * "[:" ending in ":" but not in ":]" */
13123 vFAIL("Unmatched '[' in POSIX class");
13126 /* Grandfather lone [:, [=, [. */
13136 S_could_it_be_a_POSIX_class(RExC_state_t *pRExC_state)
13138 /* This applies some heuristics at the current parse position (which should
13139 * be at a '[') to see if what follows might be intended to be a [:posix:]
13140 * class. It returns true if it really is a posix class, of course, but it
13141 * also can return true if it thinks that what was intended was a posix
13142 * class that didn't quite make it.
13144 * It will return true for
13146 * [:alphanumerics] (as long as the ] isn't followed immediately by a
13147 * ')' indicating the end of the (?[
13148 * [:any garbage including %^&$ punctuation:]
13150 * This is designed to be called only from S_handle_regex_sets; it could be
13151 * easily adapted to be called from the spot at the beginning of regclass()
13152 * that checks to see in a normal bracketed class if the surrounding []
13153 * have been omitted ([:word:] instead of [[:word:]]). But doing so would
13154 * change long-standing behavior, so I (khw) didn't do that */
13155 char* p = RExC_parse + 1;
13156 char first_char = *p;
13158 PERL_ARGS_ASSERT_COULD_IT_BE_A_POSIX_CLASS;
13160 assert(*(p - 1) == '[');
13162 if (! POSIXCC(first_char)) {
13167 while (p < RExC_end && isWORDCHAR(*p)) p++;
13169 if (p >= RExC_end) {
13173 if (p - RExC_parse > 2 /* Got at least 1 word character */
13174 && (*p == first_char
13175 || (*p == ']' && p + 1 < RExC_end && *(p + 1) != ')')))
13180 p = (char *) memchr(RExC_parse, ']', RExC_end - RExC_parse);
13183 && p - RExC_parse > 2 /* [:] evaluates to colon;
13184 [::] is a bad posix class. */
13185 && first_char == *(p - 1));
13189 S_handle_regex_sets(pTHX_ RExC_state_t *pRExC_state, SV** return_invlist,
13190 I32 *flagp, U32 depth,
13191 char * const oregcomp_parse)
13193 /* Handle the (?[...]) construct to do set operations */
13196 UV start, end; /* End points of code point ranges */
13198 char *save_end, *save_parse;
13203 const bool save_fold = FOLD;
13205 GET_RE_DEBUG_FLAGS_DECL;
13207 PERL_ARGS_ASSERT_HANDLE_REGEX_SETS;
13210 vFAIL("(?[...]) not valid in locale");
13212 RExC_uni_semantics = 1;
13214 /* This will return only an ANYOF regnode, or (unlikely) something smaller
13215 * (such as EXACT). Thus we can skip most everything if just sizing. We
13216 * call regclass to handle '[]' so as to not have to reinvent its parsing
13217 * rules here (throwing away the size it computes each time). And, we exit
13218 * upon an unescaped ']' that isn't one ending a regclass. To do both
13219 * these things, we need to realize that something preceded by a backslash
13220 * is escaped, so we have to keep track of backslashes */
13222 Perl_ck_warner_d(aTHX_
13223 packWARN(WARN_EXPERIMENTAL__REGEX_SETS),
13224 "The regex_sets feature is experimental" REPORT_LOCATION,
13225 UTF8fARG(UTF, (RExC_parse - RExC_precomp), RExC_precomp),
13227 RExC_end - RExC_start - (RExC_parse - RExC_precomp),
13228 RExC_precomp + (RExC_parse - RExC_precomp)));
13231 UV depth = 0; /* how many nested (?[...]) constructs */
13233 while (RExC_parse < RExC_end) {
13234 SV* current = NULL;
13235 RExC_parse = regpatws(pRExC_state, RExC_parse,
13236 TRUE); /* means recognize comments */
13237 switch (*RExC_parse) {
13239 if (RExC_parse[1] == '[') depth++, RExC_parse++;
13244 /* Skip the next byte (which could cause us to end up in
13245 * the middle of a UTF-8 character, but since none of those
13246 * are confusable with anything we currently handle in this
13247 * switch (invariants all), it's safe. We'll just hit the
13248 * default: case next time and keep on incrementing until
13249 * we find one of the invariants we do handle. */
13254 /* If this looks like it is a [:posix:] class, leave the
13255 * parse pointer at the '[' to fool regclass() into
13256 * thinking it is part of a '[[:posix:]]'. That function
13257 * will use strict checking to force a syntax error if it
13258 * doesn't work out to a legitimate class */
13259 bool is_posix_class
13260 = could_it_be_a_POSIX_class(pRExC_state);
13261 if (! is_posix_class) {
13265 /* regclass() can only return RESTART_UTF8 if multi-char
13266 folds are allowed. */
13267 if (!regclass(pRExC_state, flagp,depth+1,
13268 is_posix_class, /* parse the whole char
13269 class only if not a
13271 FALSE, /* don't allow multi-char folds */
13272 TRUE, /* silence non-portable warnings. */
13276 FAIL2("panic: regclass returned NULL to handle_sets, flags=%#"UVxf"",
13279 /* function call leaves parse pointing to the ']', except
13280 * if we faked it */
13281 if (is_posix_class) {
13285 SvREFCNT_dec(current); /* In case it returned something */
13290 if (depth--) break;
13292 if (RExC_parse < RExC_end
13293 && *RExC_parse == ')')
13295 node = reganode(pRExC_state, ANYOF, 0);
13296 RExC_size += ANYOF_SKIP;
13297 nextchar(pRExC_state);
13298 Set_Node_Length(node,
13299 RExC_parse - oregcomp_parse + 1); /* MJD */
13308 FAIL("Syntax error in (?[...])");
13311 /* Pass 2 only after this. Everything in this construct is a
13312 * metacharacter. Operands begin with either a '\' (for an escape
13313 * sequence), or a '[' for a bracketed character class. Any other
13314 * character should be an operator, or parenthesis for grouping. Both
13315 * types of operands are handled by calling regclass() to parse them. It
13316 * is called with a parameter to indicate to return the computed inversion
13317 * list. The parsing here is implemented via a stack. Each entry on the
13318 * stack is a single character representing one of the operators, or the
13319 * '('; or else a pointer to an operand inversion list. */
13321 #define IS_OPERAND(a) (! SvIOK(a))
13323 /* The stack starts empty. It is a syntax error if the first thing parsed
13324 * is a binary operator; everything else is pushed on the stack. When an
13325 * operand is parsed, the top of the stack is examined. If it is a binary
13326 * operator, the item before it should be an operand, and both are replaced
13327 * by the result of doing that operation on the new operand and the one on
13328 * the stack. Thus a sequence of binary operands is reduced to a single
13329 * one before the next one is parsed.
13331 * A unary operator may immediately follow a binary in the input, for
13334 * When an operand is parsed and the top of the stack is a unary operator,
13335 * the operation is performed, and then the stack is rechecked to see if
13336 * this new operand is part of a binary operation; if so, it is handled as
13339 * A '(' is simply pushed on the stack; it is valid only if the stack is
13340 * empty, or the top element of the stack is an operator or another '('
13341 * (for which the parenthesized expression will become an operand). By the
13342 * time the corresponding ')' is parsed everything in between should have
13343 * been parsed and evaluated to a single operand (or else is a syntax
13344 * error), and is handled as a regular operand */
13346 sv_2mortal((SV *)(stack = newAV()));
13348 while (RExC_parse < RExC_end) {
13349 I32 top_index = av_tindex(stack);
13351 SV* current = NULL;
13353 /* Skip white space */
13354 RExC_parse = regpatws(pRExC_state, RExC_parse,
13355 TRUE /* means recognize comments */ );
13356 if (RExC_parse >= RExC_end) {
13357 Perl_croak(aTHX_ "panic: Read past end of '(?[ ])'");
13359 if ((curchar = UCHARAT(RExC_parse)) == ']') {
13366 if (av_tindex(stack) >= 0 /* This makes sure that we can
13367 safely subtract 1 from
13368 RExC_parse in the next clause.
13369 If we have something on the
13370 stack, we have parsed something
13372 && UCHARAT(RExC_parse - 1) == '('
13373 && RExC_parse < RExC_end)
13375 /* If is a '(?', could be an embedded '(?flags:(?[...])'.
13376 * This happens when we have some thing like
13378 * my $thai_or_lao = qr/(?[ \p{Thai} + \p{Lao} ])/;
13380 * qr/(?[ \p{Digit} & $thai_or_lao ])/;
13382 * Here we would be handling the interpolated
13383 * '$thai_or_lao'. We handle this by a recursive call to
13384 * ourselves which returns the inversion list the
13385 * interpolated expression evaluates to. We use the flags
13386 * from the interpolated pattern. */
13387 U32 save_flags = RExC_flags;
13388 const char * const save_parse = ++RExC_parse;
13390 parse_lparen_question_flags(pRExC_state);
13392 if (RExC_parse == save_parse /* Makes sure there was at
13393 least one flag (or this
13394 embedding wasn't compiled)
13396 || RExC_parse >= RExC_end - 4
13397 || UCHARAT(RExC_parse) != ':'
13398 || UCHARAT(++RExC_parse) != '('
13399 || UCHARAT(++RExC_parse) != '?'
13400 || UCHARAT(++RExC_parse) != '[')
13403 /* In combination with the above, this moves the
13404 * pointer to the point just after the first erroneous
13405 * character (or if there are no flags, to where they
13406 * should have been) */
13407 if (RExC_parse >= RExC_end - 4) {
13408 RExC_parse = RExC_end;
13410 else if (RExC_parse != save_parse) {
13411 RExC_parse += (UTF) ? UTF8SKIP(RExC_parse) : 1;
13413 vFAIL("Expecting '(?flags:(?[...'");
13416 (void) handle_regex_sets(pRExC_state, ¤t, flagp,
13417 depth+1, oregcomp_parse);
13419 /* Here, 'current' contains the embedded expression's
13420 * inversion list, and RExC_parse points to the trailing
13421 * ']'; the next character should be the ')' which will be
13422 * paired with the '(' that has been put on the stack, so
13423 * the whole embedded expression reduces to '(operand)' */
13426 RExC_flags = save_flags;
13427 goto handle_operand;
13432 RExC_parse += (UTF) ? UTF8SKIP(RExC_parse) : 1;
13433 vFAIL("Unexpected character");
13436 /* regclass() can only return RESTART_UTF8 if multi-char
13437 folds are allowed. */
13438 if (!regclass(pRExC_state, flagp,depth+1,
13439 TRUE, /* means parse just the next thing */
13440 FALSE, /* don't allow multi-char folds */
13441 FALSE, /* don't silence non-portable warnings. */
13445 FAIL2("panic: regclass returned NULL to handle_sets, flags=%#"UVxf"",
13447 /* regclass() will return with parsing just the \ sequence,
13448 * leaving the parse pointer at the next thing to parse */
13450 goto handle_operand;
13452 case '[': /* Is a bracketed character class */
13454 bool is_posix_class = could_it_be_a_POSIX_class(pRExC_state);
13456 if (! is_posix_class) {
13460 /* regclass() can only return RESTART_UTF8 if multi-char
13461 folds are allowed. */
13462 if(!regclass(pRExC_state, flagp,depth+1,
13463 is_posix_class, /* parse the whole char class
13464 only if not a posix class */
13465 FALSE, /* don't allow multi-char folds */
13466 FALSE, /* don't silence non-portable warnings. */
13470 FAIL2("panic: regclass returned NULL to handle_sets, flags=%#"UVxf"",
13472 /* function call leaves parse pointing to the ']', except if we
13474 if (is_posix_class) {
13478 goto handle_operand;
13487 || ( ! (top_ptr = av_fetch(stack, top_index, FALSE)))
13488 || ! IS_OPERAND(*top_ptr))
13491 vFAIL2("Unexpected binary operator '%c' with no preceding operand", curchar);
13493 av_push(stack, newSVuv(curchar));
13497 av_push(stack, newSVuv(curchar));
13501 if (top_index >= 0) {
13502 top_ptr = av_fetch(stack, top_index, FALSE);
13504 if (IS_OPERAND(*top_ptr)) {
13506 vFAIL("Unexpected '(' with no preceding operator");
13509 av_push(stack, newSVuv(curchar));
13516 || ! (current = av_pop(stack))
13517 || ! IS_OPERAND(current)
13518 || ! (lparen = av_pop(stack))
13519 || IS_OPERAND(lparen)
13520 || SvUV(lparen) != '(')
13522 SvREFCNT_dec(current);
13524 vFAIL("Unexpected ')'");
13527 SvREFCNT_dec_NN(lparen);
13534 /* Here, we have an operand to process, in 'current' */
13536 if (top_index < 0) { /* Just push if stack is empty */
13537 av_push(stack, current);
13540 SV* top = av_pop(stack);
13542 char current_operator;
13544 if (IS_OPERAND(top)) {
13545 SvREFCNT_dec_NN(top);
13546 SvREFCNT_dec_NN(current);
13547 vFAIL("Operand with no preceding operator");
13549 current_operator = (char) SvUV(top);
13550 switch (current_operator) {
13551 case '(': /* Push the '(' back on followed by the new
13553 av_push(stack, top);
13554 av_push(stack, current);
13555 SvREFCNT_inc(top); /* Counters the '_dec' done
13556 just after the 'break', so
13557 it doesn't get wrongly freed
13562 _invlist_invert(current);
13564 /* Unlike binary operators, the top of the stack,
13565 * now that this unary one has been popped off, may
13566 * legally be an operator, and we now have operand
13569 SvREFCNT_dec_NN(top);
13570 goto handle_operand;
13573 prev = av_pop(stack);
13574 _invlist_intersection(prev,
13577 av_push(stack, current);
13582 prev = av_pop(stack);
13583 _invlist_union(prev, current, ¤t);
13584 av_push(stack, current);
13588 prev = av_pop(stack);;
13589 _invlist_subtract(prev, current, ¤t);
13590 av_push(stack, current);
13593 case '^': /* The union minus the intersection */
13599 prev = av_pop(stack);
13600 _invlist_union(prev, current, &u);
13601 _invlist_intersection(prev, current, &i);
13602 /* _invlist_subtract will overwrite current
13603 without freeing what it already contains */
13605 _invlist_subtract(u, i, ¤t);
13606 av_push(stack, current);
13607 SvREFCNT_dec_NN(i);
13608 SvREFCNT_dec_NN(u);
13609 SvREFCNT_dec_NN(element);
13614 Perl_croak(aTHX_ "panic: Unexpected item on '(?[ ])' stack");
13616 SvREFCNT_dec_NN(top);
13617 SvREFCNT_dec(prev);
13621 RExC_parse += (UTF) ? UTF8SKIP(RExC_parse) : 1;
13624 if (av_tindex(stack) < 0 /* Was empty */
13625 || ((final = av_pop(stack)) == NULL)
13626 || ! IS_OPERAND(final)
13627 || av_tindex(stack) >= 0) /* More left on stack */
13629 vFAIL("Incomplete expression within '(?[ ])'");
13632 /* Here, 'final' is the resultant inversion list from evaluating the
13633 * expression. Return it if so requested */
13634 if (return_invlist) {
13635 *return_invlist = final;
13639 /* Otherwise generate a resultant node, based on 'final'. regclass() is
13640 * expecting a string of ranges and individual code points */
13641 invlist_iterinit(final);
13642 result_string = newSVpvs("");
13643 while (invlist_iternext(final, &start, &end)) {
13644 if (start == end) {
13645 Perl_sv_catpvf(aTHX_ result_string, "\\x{%"UVXf"}", start);
13648 Perl_sv_catpvf(aTHX_ result_string, "\\x{%"UVXf"}-\\x{%"UVXf"}",
13653 save_parse = RExC_parse;
13654 RExC_parse = SvPV(result_string, len);
13655 save_end = RExC_end;
13656 RExC_end = RExC_parse + len;
13658 /* We turn off folding around the call, as the class we have constructed
13659 * already has all folding taken into consideration, and we don't want
13660 * regclass() to add to that */
13661 RExC_flags &= ~RXf_PMf_FOLD;
13662 /* regclass() can only return RESTART_UTF8 if multi-char folds are allowed.
13664 node = regclass(pRExC_state, flagp,depth+1,
13665 FALSE, /* means parse the whole char class */
13666 FALSE, /* don't allow multi-char folds */
13667 TRUE, /* silence non-portable warnings. The above may very
13668 well have generated non-portable code points, but
13669 they're valid on this machine */
13670 FALSE, /* similarly, no need for strict */
13674 FAIL2("panic: regclass returned NULL to handle_sets, flags=%#"UVxf,
13677 RExC_flags |= RXf_PMf_FOLD;
13679 RExC_parse = save_parse + 1;
13680 RExC_end = save_end;
13681 SvREFCNT_dec_NN(final);
13682 SvREFCNT_dec_NN(result_string);
13684 nextchar(pRExC_state);
13685 Set_Node_Length(node, RExC_parse - oregcomp_parse + 1); /* MJD */
13691 S_add_above_Latin1_folds(pTHX_ RExC_state_t *pRExC_state, const U8 cp, SV** invlist)
13693 /* This hard-codes the Latin1/above-Latin1 folding rules, so that an
13694 * innocent-looking character class, like /[ks]/i won't have to go out to
13695 * disk to find the possible matches.
13697 * This should be called only for a Latin1-range code points, cp, which is
13698 * known to be involved in a simple fold with other code points above
13699 * Latin1. It would give false results if /aa has been specified.
13700 * Multi-char folds are outside the scope of this, and must be handled
13703 * XXX It would be better to generate these via regen, in case a new
13704 * version of the Unicode standard adds new mappings, though that is not
13705 * really likely, and may be caught by the default: case of the switch
13708 PERL_ARGS_ASSERT_ADD_ABOVE_LATIN1_FOLDS;
13710 assert(HAS_NONLATIN1_SIMPLE_FOLD_CLOSURE(cp));
13716 add_cp_to_invlist(*invlist, KELVIN_SIGN);
13720 *invlist = add_cp_to_invlist(*invlist, LATIN_SMALL_LETTER_LONG_S);
13723 *invlist = add_cp_to_invlist(*invlist, GREEK_CAPITAL_LETTER_MU);
13724 *invlist = add_cp_to_invlist(*invlist, GREEK_SMALL_LETTER_MU);
13726 case LATIN_CAPITAL_LETTER_A_WITH_RING_ABOVE:
13727 case LATIN_SMALL_LETTER_A_WITH_RING_ABOVE:
13728 *invlist = add_cp_to_invlist(*invlist, ANGSTROM_SIGN);
13730 case LATIN_SMALL_LETTER_Y_WITH_DIAERESIS:
13731 *invlist = add_cp_to_invlist(*invlist,
13732 LATIN_CAPITAL_LETTER_Y_WITH_DIAERESIS);
13734 case LATIN_SMALL_LETTER_SHARP_S:
13735 *invlist = add_cp_to_invlist(*invlist, LATIN_CAPITAL_LETTER_SHARP_S);
13738 /* Use deprecated warning to increase the chances of this being
13741 ckWARN2reg_d(RExC_parse, "Perl folding rules are not up-to-date for 0x%02X; please use the perlbug utility to report;", cp);
13748 S_add_multi_match(pTHX_ AV* multi_char_matches, SV* multi_string, const STRLEN cp_count)
13750 /* This adds the string scalar <multi_string> to the array
13751 * <multi_char_matches>. <multi_string> is known to have exactly
13752 * <cp_count> code points in it. This is used when constructing a
13753 * bracketed character class and we find something that needs to match more
13754 * than a single character.
13756 * <multi_char_matches> is actually an array of arrays. Each top-level
13757 * element is an array that contains all the strings known so far that are
13758 * the same length. And that length (in number of code points) is the same
13759 * as the index of the top-level array. Hence, the [2] element is an
13760 * array, each element thereof is a string containing TWO code points;
13761 * while element [3] is for strings of THREE characters, and so on. Since
13762 * this is for multi-char strings there can never be a [0] nor [1] element.
13764 * When we rewrite the character class below, we will do so such that the
13765 * longest strings are written first, so that it prefers the longest
13766 * matching strings first. This is done even if it turns out that any
13767 * quantifier is non-greedy, out of this programmer's (khw) laziness. Tom
13768 * Christiansen has agreed that this is ok. This makes the test for the
13769 * ligature 'ffi' come before the test for 'ff', for example */
13772 AV** this_array_ptr;
13774 PERL_ARGS_ASSERT_ADD_MULTI_MATCH;
13776 if (! multi_char_matches) {
13777 multi_char_matches = newAV();
13780 if (av_exists(multi_char_matches, cp_count)) {
13781 this_array_ptr = (AV**) av_fetch(multi_char_matches, cp_count, FALSE);
13782 this_array = *this_array_ptr;
13785 this_array = newAV();
13786 av_store(multi_char_matches, cp_count,
13789 av_push(this_array, multi_string);
13791 return multi_char_matches;
13794 /* The names of properties whose definitions are not known at compile time are
13795 * stored in this SV, after a constant heading. So if the length has been
13796 * changed since initialization, then there is a run-time definition. */
13797 #define HAS_NONLOCALE_RUNTIME_PROPERTY_DEFINITION \
13798 (SvCUR(listsv) != initial_listsv_len)
13801 S_regclass(pTHX_ RExC_state_t *pRExC_state, I32 *flagp, U32 depth,
13802 const bool stop_at_1, /* Just parse the next thing, don't
13803 look for a full character class */
13804 bool allow_multi_folds,
13805 const bool silence_non_portable, /* Don't output warnings
13809 SV** ret_invlist /* Return an inversion list, not a node */
13812 /* parse a bracketed class specification. Most of these will produce an
13813 * ANYOF node; but something like [a] will produce an EXACT node; [aA], an
13814 * EXACTFish node; [[:ascii:]], a POSIXA node; etc. It is more complex
13815 * under /i with multi-character folds: it will be rewritten following the
13816 * paradigm of this example, where the <multi-fold>s are characters which
13817 * fold to multiple character sequences:
13818 * /[abc\x{multi-fold1}def\x{multi-fold2}ghi]/i
13819 * gets effectively rewritten as:
13820 * /(?:\x{multi-fold1}|\x{multi-fold2}|[abcdefghi]/i
13821 * reg() gets called (recursively) on the rewritten version, and this
13822 * function will return what it constructs. (Actually the <multi-fold>s
13823 * aren't physically removed from the [abcdefghi], it's just that they are
13824 * ignored in the recursion by means of a flag:
13825 * <RExC_in_multi_char_class>.)
13827 * ANYOF nodes contain a bit map for the first NUM_ANYOF_CODE_POINTS
13828 * characters, with the corresponding bit set if that character is in the
13829 * list. For characters above this, a range list or swash is used. There
13830 * are extra bits for \w, etc. in locale ANYOFs, as what these match is not
13831 * determinable at compile time
13833 * Returns NULL, setting *flagp to RESTART_UTF8 if the sizing scan needs
13834 * to be restarted. This can only happen if ret_invlist is non-NULL.
13837 UV prevvalue = OOB_UNICODE, save_prevvalue = OOB_UNICODE;
13839 UV value = OOB_UNICODE, save_value = OOB_UNICODE;
13842 IV namedclass = OOB_NAMEDCLASS;
13843 char *rangebegin = NULL;
13844 bool need_class = 0;
13846 STRLEN initial_listsv_len = 0; /* Kind of a kludge to see if it is more
13847 than just initialized. */
13848 SV* properties = NULL; /* Code points that match \p{} \P{} */
13849 SV* posixes = NULL; /* Code points that match classes like [:word:],
13850 extended beyond the Latin1 range. These have to
13851 be kept separate from other code points for much
13852 of this function because their handling is
13853 different under /i, and for most classes under
13855 SV* nposixes = NULL; /* Similarly for [:^word:]. These are kept
13856 separate for a while from the non-complemented
13857 versions because of complications with /d
13859 SV* simple_posixes = NULL; /* But under some conditions, the classes can be
13860 treated more simply than the general case,
13861 leading to less compilation and execution
13863 UV element_count = 0; /* Number of distinct elements in the class.
13864 Optimizations may be possible if this is tiny */
13865 AV * multi_char_matches = NULL; /* Code points that fold to more than one
13866 character; used under /i */
13868 char * stop_ptr = RExC_end; /* where to stop parsing */
13869 const bool skip_white = cBOOL(ret_invlist); /* ignore unescaped white
13872 /* Unicode properties are stored in a swash; this holds the current one
13873 * being parsed. If this swash is the only above-latin1 component of the
13874 * character class, an optimization is to pass it directly on to the
13875 * execution engine. Otherwise, it is set to NULL to indicate that there
13876 * are other things in the class that have to be dealt with at execution
13878 SV* swash = NULL; /* Code points that match \p{} \P{} */
13880 /* Set if a component of this character class is user-defined; just passed
13881 * on to the engine */
13882 bool has_user_defined_property = FALSE;
13884 /* inversion list of code points this node matches only when the target
13885 * string is in UTF-8. (Because is under /d) */
13886 SV* depends_list = NULL;
13888 /* Inversion list of code points this node matches regardless of things
13889 * like locale, folding, utf8ness of the target string */
13890 SV* cp_list = NULL;
13892 /* Like cp_list, but code points on this list need to be checked for things
13893 * that fold to/from them under /i */
13894 SV* cp_foldable_list = NULL;
13896 /* Like cp_list, but code points on this list are valid only when the
13897 * runtime locale is UTF-8 */
13898 SV* only_utf8_locale_list = NULL;
13900 /* In a range, if one of the endpoints is non-character-set portable,
13901 * meaning that it hard-codes a code point that may mean a different
13902 * charactger in ASCII vs. EBCDIC, as opposed to, say, a literal 'A' or a
13903 * mnemonic '\t' which each mean the same character no matter which
13904 * character set the platform is on. */
13905 unsigned int non_portable_endpoint = 0;
13907 /* Is the range unicode? which means on a platform that isn't 1-1 native
13908 * to Unicode (i.e. non-ASCII), each code point in it should be considered
13909 * to be a Unicode value. */
13910 bool unicode_range = FALSE;
13911 bool invert = FALSE; /* Is this class to be complemented */
13913 bool warn_super = ALWAYS_WARN_SUPER;
13915 regnode * const orig_emit = RExC_emit; /* Save the original RExC_emit in
13916 case we need to change the emitted regop to an EXACT. */
13917 const char * orig_parse = RExC_parse;
13918 const SSize_t orig_size = RExC_size;
13919 bool posixl_matches_all = FALSE; /* Does /l class have both e.g. \W,\w ? */
13920 GET_RE_DEBUG_FLAGS_DECL;
13922 PERL_ARGS_ASSERT_REGCLASS;
13924 PERL_UNUSED_ARG(depth);
13927 DEBUG_PARSE("clas");
13929 /* Assume we are going to generate an ANYOF node. */
13930 ret = reganode(pRExC_state,
13937 RExC_size += ANYOF_SKIP;
13938 listsv = &PL_sv_undef; /* For code scanners: listsv always non-NULL. */
13941 ANYOF_FLAGS(ret) = 0;
13943 RExC_emit += ANYOF_SKIP;
13944 listsv = newSVpvs_flags("# comment\n", SVs_TEMP);
13945 initial_listsv_len = SvCUR(listsv);
13946 SvTEMP_off(listsv); /* Grr, TEMPs and mortals are conflated. */
13950 RExC_parse = regpatws(pRExC_state, RExC_parse,
13951 FALSE /* means don't recognize comments */ );
13954 if (UCHARAT(RExC_parse) == '^') { /* Complement of range. */
13957 allow_multi_folds = FALSE;
13960 RExC_parse = regpatws(pRExC_state, RExC_parse,
13961 FALSE /* means don't recognize comments */ );
13965 /* Check that they didn't say [:posix:] instead of [[:posix:]] */
13966 if (!SIZE_ONLY && RExC_parse < RExC_end && POSIXCC(UCHARAT(RExC_parse))) {
13967 const char *s = RExC_parse;
13968 const char c = *s++;
13973 while (isWORDCHAR(*s))
13975 if (*s && c == *s && s[1] == ']') {
13976 SAVEFREESV(RExC_rx_sv);
13978 "POSIX syntax [%c %c] belongs inside character classes",
13980 (void)ReREFCNT_inc(RExC_rx_sv);
13984 /* If the caller wants us to just parse a single element, accomplish this
13985 * by faking the loop ending condition */
13986 if (stop_at_1 && RExC_end > RExC_parse) {
13987 stop_ptr = RExC_parse + 1;
13990 /* allow 1st char to be ']' (allowing it to be '-' is dealt with later) */
13991 if (UCHARAT(RExC_parse) == ']')
13992 goto charclassloop;
13995 if (RExC_parse >= stop_ptr) {
14000 RExC_parse = regpatws(pRExC_state, RExC_parse,
14001 FALSE /* means don't recognize comments */ );
14004 if (UCHARAT(RExC_parse) == ']') {
14010 namedclass = OOB_NAMEDCLASS; /* initialize as illegal */
14011 save_value = value;
14012 save_prevvalue = prevvalue;
14015 rangebegin = RExC_parse;
14017 non_portable_endpoint = 0;
14020 value = utf8n_to_uvchr((U8*)RExC_parse,
14021 RExC_end - RExC_parse,
14022 &numlen, UTF8_ALLOW_DEFAULT);
14023 RExC_parse += numlen;
14026 value = UCHARAT(RExC_parse++);
14029 && RExC_parse < RExC_end
14030 && POSIXCC(UCHARAT(RExC_parse)))
14032 namedclass = regpposixcc(pRExC_state, value, strict);
14034 else if (value == '\\') {
14035 /* Is a backslash; get the code point of the char after it */
14036 if (UTF && ! UTF8_IS_INVARIANT(UCHARAT(RExC_parse))) {
14037 value = utf8n_to_uvchr((U8*)RExC_parse,
14038 RExC_end - RExC_parse,
14039 &numlen, UTF8_ALLOW_DEFAULT);
14040 RExC_parse += numlen;
14043 value = UCHARAT(RExC_parse++);
14045 /* Some compilers cannot handle switching on 64-bit integer
14046 * values, therefore value cannot be an UV. Yes, this will
14047 * be a problem later if we want switch on Unicode.
14048 * A similar issue a little bit later when switching on
14049 * namedclass. --jhi */
14051 /* If the \ is escaping white space when white space is being
14052 * skipped, it means that that white space is wanted literally, and
14053 * is already in 'value'. Otherwise, need to translate the escape
14054 * into what it signifies. */
14055 if (! skip_white || ! is_PATWS_cp(value)) switch ((I32)value) {
14057 case 'w': namedclass = ANYOF_WORDCHAR; break;
14058 case 'W': namedclass = ANYOF_NWORDCHAR; break;
14059 case 's': namedclass = ANYOF_SPACE; break;
14060 case 'S': namedclass = ANYOF_NSPACE; break;
14061 case 'd': namedclass = ANYOF_DIGIT; break;
14062 case 'D': namedclass = ANYOF_NDIGIT; break;
14063 case 'v': namedclass = ANYOF_VERTWS; break;
14064 case 'V': namedclass = ANYOF_NVERTWS; break;
14065 case 'h': namedclass = ANYOF_HORIZWS; break;
14066 case 'H': namedclass = ANYOF_NHORIZWS; break;
14067 case 'N': /* Handle \N{NAME} in class */
14070 STRLEN cp_count = grok_bslash_N(pRExC_state, NULL, &value,
14071 flagp, depth, &as_text);
14072 if (*flagp & RESTART_UTF8)
14073 FAIL("panic: grok_bslash_N set RESTART_UTF8");
14074 if (cp_count != 1) { /* The typical case drops through */
14075 assert(cp_count != (STRLEN) -1);
14076 if (cp_count == 0) {
14078 RExC_parse++; /* Position after the "}" */
14079 vFAIL("Zero length \\N{}");
14082 ckWARNreg(RExC_parse,
14083 "Ignoring zero length \\N{} in character class");
14086 else { /* cp_count > 1 */
14087 if (! RExC_in_multi_char_class) {
14088 if (invert || range || *RExC_parse == '-') {
14091 vFAIL("\\N{} in inverted character class or as a range end-point is restricted to one character");
14094 ckWARNreg(RExC_parse, "Using just the first character returned by \\N{} in character class");
14099 = add_multi_match(multi_char_matches,
14103 break; /* <value> contains the first code
14104 point. Drop out of the switch to
14107 } /* End of cp_count != 1 */
14109 /* This element should not be processed further in this
14112 value = save_value;
14113 prevvalue = save_prevvalue;
14114 continue; /* Back to top of loop to get next char */
14117 /* Here, is a single code point, and <value> contains it */
14118 unicode_range = TRUE; /* \N{} are Unicode */
14126 /* We will handle any undefined properties ourselves */
14127 U8 swash_init_flags = _CORE_SWASH_INIT_RETURN_IF_UNDEF
14128 /* And we actually would prefer to get
14129 * the straight inversion list of the
14130 * swash, since we will be accessing it
14131 * anyway, to save a little time */
14132 |_CORE_SWASH_INIT_ACCEPT_INVLIST;
14134 if (RExC_parse >= RExC_end)
14135 vFAIL2("Empty \\%c{}", (U8)value);
14136 if (*RExC_parse == '{') {
14137 const U8 c = (U8)value;
14138 e = strchr(RExC_parse++, '}');
14140 vFAIL2("Missing right brace on \\%c{}", c);
14141 while (isSPACE(*RExC_parse))
14143 if (e == RExC_parse)
14144 vFAIL2("Empty \\%c{}", c);
14145 n = e - RExC_parse;
14146 while (isSPACE(*(RExC_parse + n - 1)))
14157 if (UCHARAT(RExC_parse) == '^') {
14160 /* toggle. (The rhs xor gets the single bit that
14161 * differs between P and p; the other xor inverts just
14163 value ^= 'P' ^ 'p';
14165 while (isSPACE(*RExC_parse)) {
14170 /* Try to get the definition of the property into
14171 * <invlist>. If /i is in effect, the effective property
14172 * will have its name be <__NAME_i>. The design is
14173 * discussed in commit
14174 * 2f833f5208e26b208886e51e09e2c072b5eabb46 */
14175 name = savepv(Perl_form(aTHX_
14177 (FOLD) ? "__" : "",
14183 /* Look up the property name, and get its swash and
14184 * inversion list, if the property is found */
14186 SvREFCNT_dec_NN(swash);
14188 swash = _core_swash_init("utf8", name, &PL_sv_undef,
14191 NULL, /* No inversion list */
14194 if (! swash || ! (invlist = _get_swash_invlist(swash))) {
14195 HV* curpkg = (IN_PERL_COMPILETIME)
14197 : CopSTASH(PL_curcop);
14199 SvREFCNT_dec_NN(swash);
14203 /* Here didn't find it. It could be a user-defined
14204 * property that will be available at run-time. If we
14205 * accept only compile-time properties, is an error;
14206 * otherwise add it to the list for run-time look up */
14208 RExC_parse = e + 1;
14210 "Property '%"UTF8f"' is unknown",
14211 UTF8fARG(UTF, n, name));
14214 /* If the property name doesn't already have a package
14215 * name, add the current one to it so that it can be
14216 * referred to outside it. [perl #121777] */
14217 if (curpkg && ! instr(name, "::")) {
14218 char* pkgname = HvNAME(curpkg);
14219 if (strNE(pkgname, "main")) {
14220 char* full_name = Perl_form(aTHX_
14224 n = strlen(full_name);
14226 name = savepvn(full_name, n);
14229 Perl_sv_catpvf(aTHX_ listsv, "%cutf8::%"UTF8f"\n",
14230 (value == 'p' ? '+' : '!'),
14231 UTF8fARG(UTF, n, name));
14232 has_user_defined_property = TRUE;
14234 /* We don't know yet, so have to assume that the
14235 * property could match something in the Latin1 range,
14236 * hence something that isn't utf8. Note that this
14237 * would cause things in <depends_list> to match
14238 * inappropriately, except that any \p{}, including
14239 * this one forces Unicode semantics, which means there
14240 * is no <depends_list> */
14242 |= ANYOF_HAS_NONBITMAP_NON_UTF8_MATCHES;
14246 /* Here, did get the swash and its inversion list. If
14247 * the swash is from a user-defined property, then this
14248 * whole character class should be regarded as such */
14249 if (swash_init_flags
14250 & _CORE_SWASH_INIT_USER_DEFINED_PROPERTY)
14252 has_user_defined_property = TRUE;
14255 /* We warn on matching an above-Unicode code point
14256 * if the match would return true, except don't
14257 * warn for \p{All}, which has exactly one element
14259 (_invlist_contains_cp(invlist, 0x110000)
14260 && (! (_invlist_len(invlist) == 1
14261 && *invlist_array(invlist) == 0)))
14267 /* Invert if asking for the complement */
14268 if (value == 'P') {
14269 _invlist_union_complement_2nd(properties,
14273 /* The swash can't be used as-is, because we've
14274 * inverted things; delay removing it to here after
14275 * have copied its invlist above */
14276 SvREFCNT_dec_NN(swash);
14280 _invlist_union(properties, invlist, &properties);
14285 RExC_parse = e + 1;
14286 namedclass = ANYOF_UNIPROP; /* no official name, but it's
14289 /* \p means they want Unicode semantics */
14290 RExC_uni_semantics = 1;
14293 case 'n': value = '\n'; break;
14294 case 'r': value = '\r'; break;
14295 case 't': value = '\t'; break;
14296 case 'f': value = '\f'; break;
14297 case 'b': value = '\b'; break;
14298 case 'e': value = ESC_NATIVE; break;
14299 case 'a': value = '\a'; break;
14301 RExC_parse--; /* function expects to be pointed at the 'o' */
14303 const char* error_msg;
14304 bool valid = grok_bslash_o(&RExC_parse,
14307 PASS2, /* warnings only in
14310 silence_non_portable,
14316 non_portable_endpoint++;
14317 if (IN_ENCODING && value < 0x100) {
14318 goto recode_encoding;
14322 RExC_parse--; /* function expects to be pointed at the 'x' */
14324 const char* error_msg;
14325 bool valid = grok_bslash_x(&RExC_parse,
14328 PASS2, /* Output warnings */
14330 silence_non_portable,
14336 non_portable_endpoint++;
14337 if (IN_ENCODING && value < 0x100)
14338 goto recode_encoding;
14341 value = grok_bslash_c(*RExC_parse++, PASS2);
14342 non_portable_endpoint++;
14344 case '0': case '1': case '2': case '3': case '4':
14345 case '5': case '6': case '7':
14347 /* Take 1-3 octal digits */
14348 I32 flags = PERL_SCAN_SILENT_ILLDIGIT;
14349 numlen = (strict) ? 4 : 3;
14350 value = grok_oct(--RExC_parse, &numlen, &flags, NULL);
14351 RExC_parse += numlen;
14354 RExC_parse += (UTF) ? UTF8SKIP(RExC_parse) : 1;
14355 vFAIL("Need exactly 3 octal digits");
14357 else if (! SIZE_ONLY /* like \08, \178 */
14359 && RExC_parse < RExC_end
14360 && isDIGIT(*RExC_parse)
14361 && ckWARN(WARN_REGEXP))
14363 SAVEFREESV(RExC_rx_sv);
14364 reg_warn_non_literal_string(
14366 form_short_octal_warning(RExC_parse, numlen));
14367 (void)ReREFCNT_inc(RExC_rx_sv);
14370 non_portable_endpoint++;
14371 if (IN_ENCODING && value < 0x100)
14372 goto recode_encoding;
14376 if (! RExC_override_recoding) {
14377 SV* enc = _get_encoding();
14378 value = reg_recode((const char)(U8)value, &enc);
14381 vFAIL("Invalid escape in the specified encoding");
14384 ckWARNreg(RExC_parse,
14385 "Invalid escape in the specified encoding");
14391 /* Allow \_ to not give an error */
14392 if (!SIZE_ONLY && isWORDCHAR(value) && value != '_') {
14394 vFAIL2("Unrecognized escape \\%c in character class",
14398 SAVEFREESV(RExC_rx_sv);
14399 ckWARN2reg(RExC_parse,
14400 "Unrecognized escape \\%c in character class passed through",
14402 (void)ReREFCNT_inc(RExC_rx_sv);
14406 } /* End of switch on char following backslash */
14407 } /* end of handling backslash escape sequences */
14409 /* Here, we have the current token in 'value' */
14411 if (namedclass > OOB_NAMEDCLASS) { /* this is a named class \blah */
14414 /* a bad range like a-\d, a-[:digit:]. The '-' is taken as a
14415 * literal, as is the character that began the false range, i.e.
14416 * the 'a' in the examples */
14419 const int w = (RExC_parse >= rangebegin)
14420 ? RExC_parse - rangebegin
14424 "False [] range \"%"UTF8f"\"",
14425 UTF8fARG(UTF, w, rangebegin));
14428 SAVEFREESV(RExC_rx_sv); /* in case of fatal warnings */
14429 ckWARN2reg(RExC_parse,
14430 "False [] range \"%"UTF8f"\"",
14431 UTF8fARG(UTF, w, rangebegin));
14432 (void)ReREFCNT_inc(RExC_rx_sv);
14433 cp_list = add_cp_to_invlist(cp_list, '-');
14434 cp_foldable_list = add_cp_to_invlist(cp_foldable_list,
14439 range = 0; /* this was not a true range */
14440 element_count += 2; /* So counts for three values */
14443 classnum = namedclass_to_classnum(namedclass);
14445 if (LOC && namedclass < ANYOF_POSIXL_MAX
14446 #ifndef HAS_ISASCII
14447 && classnum != _CC_ASCII
14450 /* What the Posix classes (like \w, [:space:]) match in locale
14451 * isn't knowable under locale until actual match time. Room
14452 * must be reserved (one time per outer bracketed class) to
14453 * store such classes. The space will contain a bit for each
14454 * named class that is to be matched against. This isn't
14455 * needed for \p{} and pseudo-classes, as they are not affected
14456 * by locale, and hence are dealt with separately */
14457 if (! need_class) {
14460 RExC_size += ANYOF_POSIXL_SKIP - ANYOF_SKIP;
14463 RExC_emit += ANYOF_POSIXL_SKIP - ANYOF_SKIP;
14465 ANYOF_FLAGS(ret) |= ANYOF_MATCHES_POSIXL;
14466 ANYOF_POSIXL_ZERO(ret);
14469 /* Coverity thinks it is possible for this to be negative; both
14470 * jhi and khw think it's not, but be safer */
14471 assert(! (ANYOF_FLAGS(ret) & ANYOF_MATCHES_POSIXL)
14472 || (namedclass + ((namedclass % 2) ? -1 : 1)) >= 0);
14474 /* See if it already matches the complement of this POSIX
14476 if ((ANYOF_FLAGS(ret) & ANYOF_MATCHES_POSIXL)
14477 && ANYOF_POSIXL_TEST(ret, namedclass + ((namedclass % 2)
14481 posixl_matches_all = TRUE;
14482 break; /* No need to continue. Since it matches both
14483 e.g., \w and \W, it matches everything, and the
14484 bracketed class can be optimized into qr/./s */
14487 /* Add this class to those that should be checked at runtime */
14488 ANYOF_POSIXL_SET(ret, namedclass);
14490 /* The above-Latin1 characters are not subject to locale rules.
14491 * Just add them, in the second pass, to the
14492 * unconditionally-matched list */
14494 SV* scratch_list = NULL;
14496 /* Get the list of the above-Latin1 code points this
14498 _invlist_intersection_maybe_complement_2nd(PL_AboveLatin1,
14499 PL_XPosix_ptrs[classnum],
14501 /* Odd numbers are complements, like
14502 * NDIGIT, NASCII, ... */
14503 namedclass % 2 != 0,
14505 /* Checking if 'cp_list' is NULL first saves an extra
14506 * clone. Its reference count will be decremented at the
14507 * next union, etc, or if this is the only instance, at the
14508 * end of the routine */
14510 cp_list = scratch_list;
14513 _invlist_union(cp_list, scratch_list, &cp_list);
14514 SvREFCNT_dec_NN(scratch_list);
14516 continue; /* Go get next character */
14519 else if (! SIZE_ONLY) {
14521 /* Here, not in pass1 (in that pass we skip calculating the
14522 * contents of this class), and is /l, or is a POSIX class for
14523 * which /l doesn't matter (or is a Unicode property, which is
14524 * skipped here). */
14525 if (namedclass >= ANYOF_POSIXL_MAX) { /* If a special class */
14526 if (namedclass != ANYOF_UNIPROP) { /* UNIPROP = \p and \P */
14528 /* Here, should be \h, \H, \v, or \V. None of /d, /i
14529 * nor /l make a difference in what these match,
14530 * therefore we just add what they match to cp_list. */
14531 if (classnum != _CC_VERTSPACE) {
14532 assert( namedclass == ANYOF_HORIZWS
14533 || namedclass == ANYOF_NHORIZWS);
14535 /* It turns out that \h is just a synonym for
14537 classnum = _CC_BLANK;
14540 _invlist_union_maybe_complement_2nd(
14542 PL_XPosix_ptrs[classnum],
14543 namedclass % 2 != 0, /* Complement if odd
14544 (NHORIZWS, NVERTWS)
14549 else if (UNI_SEMANTICS
14550 || classnum == _CC_ASCII
14551 || (DEPENDS_SEMANTICS && (classnum == _CC_DIGIT
14552 || classnum == _CC_XDIGIT)))
14554 /* We usually have to worry about /d and /a affecting what
14555 * POSIX classes match, with special code needed for /d
14556 * because we won't know until runtime what all matches.
14557 * But there is no extra work needed under /u, and
14558 * [:ascii:] is unaffected by /a and /d; and :digit: and
14559 * :xdigit: don't have runtime differences under /d. So we
14560 * can special case these, and avoid some extra work below,
14561 * and at runtime. */
14562 _invlist_union_maybe_complement_2nd(
14564 PL_XPosix_ptrs[classnum],
14565 namedclass % 2 != 0,
14568 else { /* Garden variety class. If is NUPPER, NALPHA, ...
14569 complement and use nposixes */
14570 SV** posixes_ptr = namedclass % 2 == 0
14573 _invlist_union_maybe_complement_2nd(
14575 PL_XPosix_ptrs[classnum],
14576 namedclass % 2 != 0,
14580 } /* end of namedclass \blah */
14583 RExC_parse = regpatws(pRExC_state, RExC_parse,
14584 FALSE /* means don't recognize comments */ );
14587 /* If 'range' is set, 'value' is the ending of a range--check its
14588 * validity. (If value isn't a single code point in the case of a
14589 * range, we should have figured that out above in the code that
14590 * catches false ranges). Later, we will handle each individual code
14591 * point in the range. If 'range' isn't set, this could be the
14592 * beginning of a range, so check for that by looking ahead to see if
14593 * the next real character to be processed is the range indicator--the
14598 /* For unicode ranges, we have to test that the Unicode as opposed
14599 * to the native values are not decreasing. (Above 255, there is
14600 * no difference between native and Unicode) */
14601 if (unicode_range && prevvalue < 255 && value < 255) {
14602 if (NATIVE_TO_LATIN1(prevvalue) > NATIVE_TO_LATIN1(value)) {
14603 goto backwards_range;
14608 if (prevvalue > value) /* b-a */ {
14613 w = RExC_parse - rangebegin;
14615 "Invalid [] range \"%"UTF8f"\"",
14616 UTF8fARG(UTF, w, rangebegin));
14617 NOT_REACHED; /* NOTREACHED */
14621 prevvalue = value; /* save the beginning of the potential range */
14622 if (! stop_at_1 /* Can't be a range if parsing just one thing */
14623 && *RExC_parse == '-')
14625 char* next_char_ptr = RExC_parse + 1;
14626 if (skip_white) { /* Get the next real char after the '-' */
14627 next_char_ptr = regpatws(pRExC_state,
14629 FALSE); /* means don't recognize
14633 /* If the '-' is at the end of the class (just before the ']',
14634 * it is a literal minus; otherwise it is a range */
14635 if (next_char_ptr < RExC_end && *next_char_ptr != ']') {
14636 RExC_parse = next_char_ptr;
14638 /* a bad range like \w-, [:word:]- ? */
14639 if (namedclass > OOB_NAMEDCLASS) {
14640 if (strict || (PASS2 && ckWARN(WARN_REGEXP))) {
14641 const int w = RExC_parse >= rangebegin
14642 ? RExC_parse - rangebegin
14645 vFAIL4("False [] range \"%*.*s\"",
14650 "False [] range \"%*.*s\"",
14655 cp_list = add_cp_to_invlist(cp_list, '-');
14659 range = 1; /* yeah, it's a range! */
14660 continue; /* but do it the next time */
14665 if (namedclass > OOB_NAMEDCLASS) {
14669 /* Here, we have a single value this time through the loop, and
14670 * <prevvalue> is the beginning of the range, if any; or <value> if
14673 /* non-Latin1 code point implies unicode semantics. Must be set in
14674 * pass1 so is there for the whole of pass 2 */
14676 RExC_uni_semantics = 1;
14679 /* Ready to process either the single value, or the completed range.
14680 * For single-valued non-inverted ranges, we consider the possibility
14681 * of multi-char folds. (We made a conscious decision to not do this
14682 * for the other cases because it can often lead to non-intuitive
14683 * results. For example, you have the peculiar case that:
14684 * "s s" =~ /^[^\xDF]+$/i => Y
14685 * "ss" =~ /^[^\xDF]+$/i => N
14687 * See [perl #89750] */
14688 if (FOLD && allow_multi_folds && value == prevvalue) {
14689 if (value == LATIN_SMALL_LETTER_SHARP_S
14690 || (value > 255 && _invlist_contains_cp(PL_HasMultiCharFold,
14693 /* Here <value> is indeed a multi-char fold. Get what it is */
14695 U8 foldbuf[UTF8_MAXBYTES_CASE];
14698 UV folded = _to_uni_fold_flags(
14702 FOLD_FLAGS_FULL | (ASCII_FOLD_RESTRICTED
14703 ? FOLD_FLAGS_NOMIX_ASCII
14707 /* Here, <folded> should be the first character of the
14708 * multi-char fold of <value>, with <foldbuf> containing the
14709 * whole thing. But, if this fold is not allowed (because of
14710 * the flags), <fold> will be the same as <value>, and should
14711 * be processed like any other character, so skip the special
14713 if (folded != value) {
14715 /* Skip if we are recursed, currently parsing the class
14716 * again. Otherwise add this character to the list of
14717 * multi-char folds. */
14718 if (! RExC_in_multi_char_class) {
14719 STRLEN cp_count = utf8_length(foldbuf,
14720 foldbuf + foldlen);
14721 SV* multi_fold = sv_2mortal(newSVpvs(""));
14723 Perl_sv_catpvf(aTHX_ multi_fold, "\\x{%"UVXf"}", value);
14726 = add_multi_match(multi_char_matches,
14732 /* This element should not be processed further in this
14735 value = save_value;
14736 prevvalue = save_prevvalue;
14742 if (strict && PASS2 && ckWARN(WARN_REGEXP)) {
14745 /* If the range starts above 255, everything is portable and
14746 * likely to be so for any forseeable character set, so don't
14748 if (unicode_range && non_portable_endpoint && prevvalue < 256) {
14749 vWARN(RExC_parse, "Both or neither range ends should be Unicode");
14751 else if (prevvalue != value) {
14753 /* Under strict, ranges that stop and/or end in an ASCII
14754 * printable should have each end point be a portable value
14755 * for it (preferably like 'A', but we don't warn if it is
14756 * a (portable) Unicode name or code point), and the range
14757 * must be be all digits or all letters of the same case.
14758 * Otherwise, the range is non-portable and unclear as to
14759 * what it contains */
14760 if ((isPRINT_A(prevvalue) || isPRINT_A(value))
14761 && (non_portable_endpoint
14762 || ! ((isDIGIT_A(prevvalue) && isDIGIT_A(value))
14763 || (isLOWER_A(prevvalue) && isLOWER_A(value))
14764 || (isUPPER_A(prevvalue) && isUPPER_A(value)))))
14766 vWARN(RExC_parse, "Ranges of ASCII printables should be some subset of \"0-9\", \"A-Z\", or \"a-z\"");
14768 else if (prevvalue >= 0x660) { /* ARABIC_INDIC_DIGIT_ZERO */
14770 /* But the nature of Unicode and languages mean we
14771 * can't do the same checks for above-ASCII ranges,
14772 * except in the case of digit ones. These should
14773 * contain only digits from the same group of 10. The
14774 * ASCII case is handled just above. 0x660 is the
14775 * first digit character beyond ASCII. Hence here, the
14776 * range could be a range of digits. Find out. */
14777 IV index_start = _invlist_search(PL_XPosix_ptrs[_CC_DIGIT],
14779 IV index_final = _invlist_search(PL_XPosix_ptrs[_CC_DIGIT],
14782 /* If the range start and final points are in the same
14783 * inversion list element, it means that either both
14784 * are not digits, or both are digits in a consecutive
14785 * sequence of digits. (So far, Unicode has kept all
14786 * such sequences as distinct groups of 10, but assert
14787 * to make sure). If the end points are not in the
14788 * same element, neither should be a digit. */
14789 if (index_start == index_final) {
14790 assert(! ELEMENT_RANGE_MATCHES_INVLIST(index_start)
14791 || invlist_array(PL_XPosix_ptrs[_CC_DIGIT])[index_start+1]
14792 - invlist_array(PL_XPosix_ptrs[_CC_DIGIT])[index_start]
14795 else if ((index_start >= 0
14796 && ELEMENT_RANGE_MATCHES_INVLIST(index_start))
14797 || (index_final >= 0
14798 && ELEMENT_RANGE_MATCHES_INVLIST(index_final)))
14800 vWARN(RExC_parse, "Ranges of digits should be from the same group of 10");
14805 if ((! range || prevvalue == value) && non_portable_endpoint) {
14806 if (isPRINT_A(value)) {
14809 if (isBACKSLASHED_PUNCT(value)) {
14810 literal[d++] = '\\';
14812 literal[d++] = (char) value;
14813 literal[d++] = '\0';
14816 "\"%.*s\" is more clearly written simply as \"%s\"",
14817 (int) (RExC_parse - rangebegin),
14822 else if isMNEMONIC_CNTRL(value) {
14824 "\"%.*s\" is more clearly written simply as \"%s\"",
14825 (int) (RExC_parse - rangebegin),
14827 cntrl_to_mnemonic((char) value)
14833 /* Deal with this element of the class */
14837 cp_foldable_list = _add_range_to_invlist(cp_foldable_list,
14840 /* On non-ASCII platforms, for ranges that span all of 0..255, and
14841 * ones that don't require special handling, we can just add the
14842 * range like we do for ASCII platforms */
14843 if ((UNLIKELY(prevvalue == 0) && value >= 255)
14844 || ! (prevvalue < 256
14846 || (! non_portable_endpoint
14847 && ((isLOWER_A(prevvalue) && isLOWER_A(value))
14848 || (isUPPER_A(prevvalue)
14849 && isUPPER_A(value)))))))
14851 cp_foldable_list = _add_range_to_invlist(cp_foldable_list,
14855 /* Here, requires special handling. This can be because it is
14856 * a range whose code points are considered to be Unicode, and
14857 * so must be individually translated into native, or because
14858 * its a subrange of 'A-Z' or 'a-z' which each aren't
14859 * contiguous in EBCDIC, but we have defined them to include
14860 * only the "expected" upper or lower case ASCII alphabetics.
14861 * Subranges above 255 are the same in native and Unicode, so
14862 * can be added as a range */
14863 U8 start = NATIVE_TO_LATIN1(prevvalue);
14865 U8 end = (value < 256) ? NATIVE_TO_LATIN1(value) : 255;
14866 for (j = start; j <= end; j++) {
14867 cp_foldable_list = add_cp_to_invlist(cp_foldable_list, LATIN1_TO_NATIVE(j));
14870 cp_foldable_list = _add_range_to_invlist(cp_foldable_list,
14877 range = 0; /* this range (if it was one) is done now */
14878 } /* End of loop through all the text within the brackets */
14880 /* If anything in the class expands to more than one character, we have to
14881 * deal with them by building up a substitute parse string, and recursively
14882 * calling reg() on it, instead of proceeding */
14883 if (multi_char_matches) {
14884 SV * substitute_parse = newSVpvn_flags("?:", 2, SVs_TEMP);
14887 char *save_end = RExC_end;
14888 char *save_parse = RExC_parse;
14889 bool first_time = TRUE; /* First multi-char occurrence doesn't get
14894 #if 0 /* Have decided not to deal with multi-char folds in inverted classes,
14895 because too confusing */
14897 sv_catpv(substitute_parse, "(?:");
14901 /* Look at the longest folds first */
14902 for (cp_count = av_tindex(multi_char_matches); cp_count > 0; cp_count--) {
14904 if (av_exists(multi_char_matches, cp_count)) {
14905 AV** this_array_ptr;
14908 this_array_ptr = (AV**) av_fetch(multi_char_matches,
14910 while ((this_sequence = av_pop(*this_array_ptr)) !=
14913 if (! first_time) {
14914 sv_catpv(substitute_parse, "|");
14916 first_time = FALSE;
14918 sv_catpv(substitute_parse, SvPVX(this_sequence));
14923 /* If the character class contains anything else besides these
14924 * multi-character folds, have to include it in recursive parsing */
14925 if (element_count) {
14926 sv_catpv(substitute_parse, "|[");
14927 sv_catpvn(substitute_parse, orig_parse, RExC_parse - orig_parse);
14928 sv_catpv(substitute_parse, "]");
14931 sv_catpv(substitute_parse, ")");
14934 /* This is a way to get the parse to skip forward a whole named
14935 * sequence instead of matching the 2nd character when it fails the
14937 sv_catpv(substitute_parse, "(*THEN)(*SKIP)(*FAIL)|.)");
14941 RExC_parse = SvPV(substitute_parse, len);
14942 RExC_end = RExC_parse + len;
14943 RExC_in_multi_char_class = 1;
14944 RExC_override_recoding = 1;
14945 RExC_emit = (regnode *)orig_emit;
14947 ret = reg(pRExC_state, 1, ®_flags, depth+1);
14949 *flagp |= reg_flags&(HASWIDTH|SIMPLE|SPSTART|POSTPONED|RESTART_UTF8);
14951 RExC_parse = save_parse;
14952 RExC_end = save_end;
14953 RExC_in_multi_char_class = 0;
14954 RExC_override_recoding = 0;
14955 SvREFCNT_dec_NN(multi_char_matches);
14959 /* Here, we've gone through the entire class and dealt with multi-char
14960 * folds. We are now in a position that we can do some checks to see if we
14961 * can optimize this ANYOF node into a simpler one, even in Pass 1.
14962 * Currently we only do two checks:
14963 * 1) is in the unlikely event that the user has specified both, eg. \w and
14964 * \W under /l, then the class matches everything. (This optimization
14965 * is done only to make the optimizer code run later work.)
14966 * 2) if the character class contains only a single element (including a
14967 * single range), we see if there is an equivalent node for it.
14968 * Other checks are possible */
14969 if (! ret_invlist /* Can't optimize if returning the constructed
14971 && (UNLIKELY(posixl_matches_all) || element_count == 1))
14976 if (UNLIKELY(posixl_matches_all)) {
14979 else if (namedclass > OOB_NAMEDCLASS) { /* this is a named class, like
14980 \w or [:digit:] or \p{foo}
14983 /* All named classes are mapped into POSIXish nodes, with its FLAG
14984 * argument giving which class it is */
14985 switch ((I32)namedclass) {
14986 case ANYOF_UNIPROP:
14989 /* These don't depend on the charset modifiers. They always
14990 * match under /u rules */
14991 case ANYOF_NHORIZWS:
14992 case ANYOF_HORIZWS:
14993 namedclass = ANYOF_BLANK + namedclass - ANYOF_HORIZWS;
14996 case ANYOF_NVERTWS:
15001 /* The actual POSIXish node for all the rest depends on the
15002 * charset modifier. The ones in the first set depend only on
15003 * ASCII or, if available on this platform, also locale */
15007 op = (LOC) ? POSIXL : POSIXA;
15013 /* The following don't have any matches in the upper Latin1
15014 * range, hence /d is equivalent to /u for them. Making it /u
15015 * saves some branches at runtime */
15019 case ANYOF_NXDIGIT:
15020 if (! DEPENDS_SEMANTICS) {
15021 goto treat_as_default;
15027 /* The following change to CASED under /i */
15033 namedclass = ANYOF_CASED + (namedclass % 2);
15037 /* The rest have more possibilities depending on the charset.
15038 * We take advantage of the enum ordering of the charset
15039 * modifiers to get the exact node type, */
15042 op = POSIXD + get_regex_charset(RExC_flags);
15043 if (op > POSIXA) { /* /aa is same as /a */
15048 /* The odd numbered ones are the complements of the
15049 * next-lower even number one */
15050 if (namedclass % 2 == 1) {
15054 arg = namedclass_to_classnum(namedclass);
15058 else if (value == prevvalue) {
15060 /* Here, the class consists of just a single code point */
15063 if (! LOC && value == '\n') {
15064 op = REG_ANY; /* Optimize [^\n] */
15065 *flagp |= HASWIDTH|SIMPLE;
15069 else if (value < 256 || UTF) {
15071 /* Optimize a single value into an EXACTish node, but not if it
15072 * would require converting the pattern to UTF-8. */
15073 op = compute_EXACTish(pRExC_state);
15075 } /* Otherwise is a range */
15076 else if (! LOC) { /* locale could vary these */
15077 if (prevvalue == '0') {
15078 if (value == '9') {
15083 else if (AT_LEAST_ASCII_RESTRICTED || ! FOLD) {
15084 /* We can optimize A-Z or a-z, but not if they could match
15085 * something like the KELVIN SIGN under /i (/a means they
15087 if (prevvalue == 'A') {
15090 && ! non_portable_endpoint
15093 arg = (FOLD) ? _CC_ALPHA : _CC_UPPER;
15097 else if (prevvalue == 'a') {
15100 && ! non_portable_endpoint
15103 arg = (FOLD) ? _CC_ALPHA : _CC_LOWER;
15110 /* Here, we have changed <op> away from its initial value iff we found
15111 * an optimization */
15114 /* Throw away this ANYOF regnode, and emit the calculated one,
15115 * which should correspond to the beginning, not current, state of
15117 const char * cur_parse = RExC_parse;
15118 RExC_parse = (char *)orig_parse;
15122 /* To get locale nodes to not use the full ANYOF size would
15123 * require moving the code above that writes the portions
15124 * of it that aren't in other nodes to after this point.
15125 * e.g. ANYOF_POSIXL_SET */
15126 RExC_size = orig_size;
15130 RExC_emit = (regnode *)orig_emit;
15131 if (PL_regkind[op] == POSIXD) {
15132 if (op == POSIXL) {
15133 RExC_contains_locale = 1;
15136 op += NPOSIXD - POSIXD;
15141 ret = reg_node(pRExC_state, op);
15143 if (PL_regkind[op] == POSIXD || PL_regkind[op] == NPOSIXD) {
15147 *flagp |= HASWIDTH|SIMPLE;
15149 else if (PL_regkind[op] == EXACT) {
15150 alloc_maybe_populate_EXACT(pRExC_state, ret, flagp, 0, value,
15151 TRUE /* downgradable to EXACT */
15155 RExC_parse = (char *) cur_parse;
15157 SvREFCNT_dec(posixes);
15158 SvREFCNT_dec(nposixes);
15159 SvREFCNT_dec(simple_posixes);
15160 SvREFCNT_dec(cp_list);
15161 SvREFCNT_dec(cp_foldable_list);
15168 /****** !SIZE_ONLY (Pass 2) AFTER HERE *********/
15170 /* If folding, we calculate all characters that could fold to or from the
15171 * ones already on the list */
15172 if (cp_foldable_list) {
15174 UV start, end; /* End points of code point ranges */
15176 SV* fold_intersection = NULL;
15179 /* Our calculated list will be for Unicode rules. For locale
15180 * matching, we have to keep a separate list that is consulted at
15181 * runtime only when the locale indicates Unicode rules. For
15182 * non-locale, we just use to the general list */
15184 use_list = &only_utf8_locale_list;
15187 use_list = &cp_list;
15190 /* Only the characters in this class that participate in folds need
15191 * be checked. Get the intersection of this class and all the
15192 * possible characters that are foldable. This can quickly narrow
15193 * down a large class */
15194 _invlist_intersection(PL_utf8_foldable, cp_foldable_list,
15195 &fold_intersection);
15197 /* The folds for all the Latin1 characters are hard-coded into this
15198 * program, but we have to go out to disk to get the others. */
15199 if (invlist_highest(cp_foldable_list) >= 256) {
15201 /* This is a hash that for a particular fold gives all
15202 * characters that are involved in it */
15203 if (! PL_utf8_foldclosures) {
15204 _load_PL_utf8_foldclosures();
15208 /* Now look at the foldable characters in this class individually */
15209 invlist_iterinit(fold_intersection);
15210 while (invlist_iternext(fold_intersection, &start, &end)) {
15213 /* Look at every character in the range */
15214 for (j = start; j <= end; j++) {
15215 U8 foldbuf[UTF8_MAXBYTES_CASE+1];
15221 if (IS_IN_SOME_FOLD_L1(j)) {
15223 /* ASCII is always matched; non-ASCII is matched
15224 * only under Unicode rules (which could happen
15225 * under /l if the locale is a UTF-8 one */
15226 if (isASCII(j) || ! DEPENDS_SEMANTICS) {
15227 *use_list = add_cp_to_invlist(*use_list,
15228 PL_fold_latin1[j]);
15232 add_cp_to_invlist(depends_list,
15233 PL_fold_latin1[j]);
15237 if (HAS_NONLATIN1_SIMPLE_FOLD_CLOSURE(j)
15238 && (! isASCII(j) || ! ASCII_FOLD_RESTRICTED))
15240 add_above_Latin1_folds(pRExC_state,
15247 /* Here is an above Latin1 character. We don't have the
15248 * rules hard-coded for it. First, get its fold. This is
15249 * the simple fold, as the multi-character folds have been
15250 * handled earlier and separated out */
15251 _to_uni_fold_flags(j, foldbuf, &foldlen,
15252 (ASCII_FOLD_RESTRICTED)
15253 ? FOLD_FLAGS_NOMIX_ASCII
15256 /* Single character fold of above Latin1. Add everything in
15257 * its fold closure to the list that this node should match.
15258 * The fold closures data structure is a hash with the keys
15259 * being the UTF-8 of every character that is folded to, like
15260 * 'k', and the values each an array of all code points that
15261 * fold to its key. e.g. [ 'k', 'K', KELVIN_SIGN ].
15262 * Multi-character folds are not included */
15263 if ((listp = hv_fetch(PL_utf8_foldclosures,
15264 (char *) foldbuf, foldlen, FALSE)))
15266 AV* list = (AV*) *listp;
15268 for (k = 0; k <= av_tindex(list); k++) {
15269 SV** c_p = av_fetch(list, k, FALSE);
15275 /* /aa doesn't allow folds between ASCII and non- */
15276 if ((ASCII_FOLD_RESTRICTED
15277 && (isASCII(c) != isASCII(j))))
15282 /* Folds under /l which cross the 255/256 boundary
15283 * are added to a separate list. (These are valid
15284 * only when the locale is UTF-8.) */
15285 if (c < 256 && LOC) {
15286 *use_list = add_cp_to_invlist(*use_list, c);
15290 if (isASCII(c) || c > 255 || AT_LEAST_UNI_SEMANTICS)
15292 cp_list = add_cp_to_invlist(cp_list, c);
15295 /* Similarly folds involving non-ascii Latin1
15296 * characters under /d are added to their list */
15297 depends_list = add_cp_to_invlist(depends_list,
15304 SvREFCNT_dec_NN(fold_intersection);
15307 /* Now that we have finished adding all the folds, there is no reason
15308 * to keep the foldable list separate */
15309 _invlist_union(cp_list, cp_foldable_list, &cp_list);
15310 SvREFCNT_dec_NN(cp_foldable_list);
15313 /* And combine the result (if any) with any inversion list from posix
15314 * classes. The lists are kept separate up to now because we don't want to
15315 * fold the classes (folding of those is automatically handled by the swash
15316 * fetching code) */
15317 if (simple_posixes) {
15318 _invlist_union(cp_list, simple_posixes, &cp_list);
15319 SvREFCNT_dec_NN(simple_posixes);
15321 if (posixes || nposixes) {
15322 if (posixes && AT_LEAST_ASCII_RESTRICTED) {
15323 /* Under /a and /aa, nothing above ASCII matches these */
15324 _invlist_intersection(posixes,
15325 PL_XPosix_ptrs[_CC_ASCII],
15329 if (DEPENDS_SEMANTICS) {
15330 /* Under /d, everything in the upper half of the Latin1 range
15331 * matches these complements */
15332 ANYOF_FLAGS(ret) |= ANYOF_MATCHES_ALL_NON_UTF8_NON_ASCII;
15334 else if (AT_LEAST_ASCII_RESTRICTED) {
15335 /* Under /a and /aa, everything above ASCII matches these
15337 _invlist_union_complement_2nd(nposixes,
15338 PL_XPosix_ptrs[_CC_ASCII],
15342 _invlist_union(posixes, nposixes, &posixes);
15343 SvREFCNT_dec_NN(nposixes);
15346 posixes = nposixes;
15349 if (! DEPENDS_SEMANTICS) {
15351 _invlist_union(cp_list, posixes, &cp_list);
15352 SvREFCNT_dec_NN(posixes);
15359 /* Under /d, we put into a separate list the Latin1 things that
15360 * match only when the target string is utf8 */
15361 SV* nonascii_but_latin1_properties = NULL;
15362 _invlist_intersection(posixes, PL_UpperLatin1,
15363 &nonascii_but_latin1_properties);
15364 _invlist_subtract(posixes, nonascii_but_latin1_properties,
15367 _invlist_union(cp_list, posixes, &cp_list);
15368 SvREFCNT_dec_NN(posixes);
15374 if (depends_list) {
15375 _invlist_union(depends_list, nonascii_but_latin1_properties,
15377 SvREFCNT_dec_NN(nonascii_but_latin1_properties);
15380 depends_list = nonascii_but_latin1_properties;
15385 /* And combine the result (if any) with any inversion list from properties.
15386 * The lists are kept separate up to now so that we can distinguish the two
15387 * in regards to matching above-Unicode. A run-time warning is generated
15388 * if a Unicode property is matched against a non-Unicode code point. But,
15389 * we allow user-defined properties to match anything, without any warning,
15390 * and we also suppress the warning if there is a portion of the character
15391 * class that isn't a Unicode property, and which matches above Unicode, \W
15392 * or [\x{110000}] for example.
15393 * (Note that in this case, unlike the Posix one above, there is no
15394 * <depends_list>, because having a Unicode property forces Unicode
15399 /* If it matters to the final outcome, see if a non-property
15400 * component of the class matches above Unicode. If so, the
15401 * warning gets suppressed. This is true even if just a single
15402 * such code point is specified, as though not strictly correct if
15403 * another such code point is matched against, the fact that they
15404 * are using above-Unicode code points indicates they should know
15405 * the issues involved */
15407 warn_super = ! (invert
15408 ^ (invlist_highest(cp_list) > PERL_UNICODE_MAX));
15411 _invlist_union(properties, cp_list, &cp_list);
15412 SvREFCNT_dec_NN(properties);
15415 cp_list = properties;
15419 ANYOF_FLAGS(ret) |= ANYOF_WARN_SUPER;
15423 /* Here, we have calculated what code points should be in the character
15426 * Now we can see about various optimizations. Fold calculation (which we
15427 * did above) needs to take place before inversion. Otherwise /[^k]/i
15428 * would invert to include K, which under /i would match k, which it
15429 * shouldn't. Therefore we can't invert folded locale now, as it won't be
15430 * folded until runtime */
15432 /* If we didn't do folding, it's because some information isn't available
15433 * until runtime; set the run-time fold flag for these. (We don't have to
15434 * worry about properties folding, as that is taken care of by the swash
15435 * fetching). We know to set the flag if we have a non-NULL list for UTF-8
15436 * locales, or the class matches at least one 0-255 range code point */
15438 if (only_utf8_locale_list) {
15439 ANYOF_FLAGS(ret) |= ANYOF_LOC_FOLD;
15441 else if (cp_list) { /* Look to see if there a 0-255 code point is in
15444 invlist_iterinit(cp_list);
15445 if (invlist_iternext(cp_list, &start, &end) && start < 256) {
15446 ANYOF_FLAGS(ret) |= ANYOF_LOC_FOLD;
15448 invlist_iterfinish(cp_list);
15452 /* Optimize inverted simple patterns (e.g. [^a-z]) when everything is known
15453 * at compile time. Besides not inverting folded locale now, we can't
15454 * invert if there are things such as \w, which aren't known until runtime
15458 && ! (ANYOF_FLAGS(ret) & (ANYOF_LOCALE_FLAGS))
15460 && ! HAS_NONLOCALE_RUNTIME_PROPERTY_DEFINITION)
15462 _invlist_invert(cp_list);
15464 /* Any swash can't be used as-is, because we've inverted things */
15466 SvREFCNT_dec_NN(swash);
15470 /* Clear the invert flag since have just done it here */
15477 *ret_invlist = cp_list;
15478 SvREFCNT_dec(swash);
15480 /* Discard the generated node */
15482 RExC_size = orig_size;
15485 RExC_emit = orig_emit;
15490 /* Some character classes are equivalent to other nodes. Such nodes take
15491 * up less room and generally fewer operations to execute than ANYOF nodes.
15492 * Above, we checked for and optimized into some such equivalents for
15493 * certain common classes that are easy to test. Getting to this point in
15494 * the code means that the class didn't get optimized there. Since this
15495 * code is only executed in Pass 2, it is too late to save space--it has
15496 * been allocated in Pass 1, and currently isn't given back. But turning
15497 * things into an EXACTish node can allow the optimizer to join it to any
15498 * adjacent such nodes. And if the class is equivalent to things like /./,
15499 * expensive run-time swashes can be avoided. Now that we have more
15500 * complete information, we can find things necessarily missed by the
15501 * earlier code. I (khw) am not sure how much to look for here. It would
15502 * be easy, but perhaps too slow, to check any candidates against all the
15503 * node types they could possibly match using _invlistEQ(). */
15508 && ! (ANYOF_FLAGS(ret) & (ANYOF_LOCALE_FLAGS))
15509 && ! HAS_NONLOCALE_RUNTIME_PROPERTY_DEFINITION
15511 /* We don't optimize if we are supposed to make sure all non-Unicode
15512 * code points raise a warning, as only ANYOF nodes have this check.
15514 && ! ((ANYOF_FLAGS(ret) & ANYOF_WARN_SUPER) && ALWAYS_WARN_SUPER))
15517 U8 op = END; /* The optimzation node-type */
15518 const char * cur_parse= RExC_parse;
15520 invlist_iterinit(cp_list);
15521 if (! invlist_iternext(cp_list, &start, &end)) {
15523 /* Here, the list is empty. This happens, for example, when a
15524 * Unicode property is the only thing in the character class, and
15525 * it doesn't match anything. (perluniprops.pod notes such
15528 *flagp |= HASWIDTH|SIMPLE;
15530 else if (start == end) { /* The range is a single code point */
15531 if (! invlist_iternext(cp_list, &start, &end)
15533 /* Don't do this optimization if it would require changing
15534 * the pattern to UTF-8 */
15535 && (start < 256 || UTF))
15537 /* Here, the list contains a single code point. Can optimize
15538 * into an EXACTish node */
15549 /* A locale node under folding with one code point can be
15550 * an EXACTFL, as its fold won't be calculated until
15556 /* Here, we are generally folding, but there is only one
15557 * code point to match. If we have to, we use an EXACT
15558 * node, but it would be better for joining with adjacent
15559 * nodes in the optimization pass if we used the same
15560 * EXACTFish node that any such are likely to be. We can
15561 * do this iff the code point doesn't participate in any
15562 * folds. For example, an EXACTF of a colon is the same as
15563 * an EXACT one, since nothing folds to or from a colon. */
15565 if (IS_IN_SOME_FOLD_L1(value)) {
15570 if (_invlist_contains_cp(PL_utf8_foldable, value)) {
15575 /* If we haven't found the node type, above, it means we
15576 * can use the prevailing one */
15578 op = compute_EXACTish(pRExC_state);
15583 else if (start == 0) {
15584 if (end == UV_MAX) {
15586 *flagp |= HASWIDTH|SIMPLE;
15589 else if (end == '\n' - 1
15590 && invlist_iternext(cp_list, &start, &end)
15591 && start == '\n' + 1 && end == UV_MAX)
15594 *flagp |= HASWIDTH|SIMPLE;
15598 invlist_iterfinish(cp_list);
15601 RExC_parse = (char *)orig_parse;
15602 RExC_emit = (regnode *)orig_emit;
15604 ret = reg_node(pRExC_state, op);
15606 RExC_parse = (char *)cur_parse;
15608 if (PL_regkind[op] == EXACT) {
15609 alloc_maybe_populate_EXACT(pRExC_state, ret, flagp, 0, value,
15610 TRUE /* downgradable to EXACT */
15614 SvREFCNT_dec_NN(cp_list);
15619 /* Here, <cp_list> contains all the code points we can determine at
15620 * compile time that match under all conditions. Go through it, and
15621 * for things that belong in the bitmap, put them there, and delete from
15622 * <cp_list>. While we are at it, see if everything above 255 is in the
15623 * list, and if so, set a flag to speed up execution */
15625 populate_ANYOF_from_invlist(ret, &cp_list);
15628 ANYOF_FLAGS(ret) |= ANYOF_INVERT;
15631 /* Here, the bitmap has been populated with all the Latin1 code points that
15632 * always match. Can now add to the overall list those that match only
15633 * when the target string is UTF-8 (<depends_list>). */
15634 if (depends_list) {
15636 _invlist_union(cp_list, depends_list, &cp_list);
15637 SvREFCNT_dec_NN(depends_list);
15640 cp_list = depends_list;
15642 ANYOF_FLAGS(ret) |= ANYOF_HAS_UTF8_NONBITMAP_MATCHES;
15645 /* If there is a swash and more than one element, we can't use the swash in
15646 * the optimization below. */
15647 if (swash && element_count > 1) {
15648 SvREFCNT_dec_NN(swash);
15652 /* Note that the optimization of using 'swash' if it is the only thing in
15653 * the class doesn't have us change swash at all, so it can include things
15654 * that are also in the bitmap; otherwise we have purposely deleted that
15655 * duplicate information */
15656 set_ANYOF_arg(pRExC_state, ret, cp_list,
15657 (HAS_NONLOCALE_RUNTIME_PROPERTY_DEFINITION)
15659 only_utf8_locale_list,
15660 swash, has_user_defined_property);
15662 *flagp |= HASWIDTH|SIMPLE;
15664 if (ANYOF_FLAGS(ret) & ANYOF_LOCALE_FLAGS) {
15665 RExC_contains_locale = 1;
15671 #undef HAS_NONLOCALE_RUNTIME_PROPERTY_DEFINITION
15674 S_set_ANYOF_arg(pTHX_ RExC_state_t* const pRExC_state,
15675 regnode* const node,
15677 SV* const runtime_defns,
15678 SV* const only_utf8_locale_list,
15680 const bool has_user_defined_property)
15682 /* Sets the arg field of an ANYOF-type node 'node', using information about
15683 * the node passed-in. If there is nothing outside the node's bitmap, the
15684 * arg is set to ANYOF_ONLY_HAS_BITMAP. Otherwise, it sets the argument to
15685 * the count returned by add_data(), having allocated and stored an array,
15686 * av, that that count references, as follows:
15687 * av[0] stores the character class description in its textual form.
15688 * This is used later (regexec.c:Perl_regclass_swash()) to
15689 * initialize the appropriate swash, and is also useful for dumping
15690 * the regnode. This is set to &PL_sv_undef if the textual
15691 * description is not needed at run-time (as happens if the other
15692 * elements completely define the class)
15693 * av[1] if &PL_sv_undef, is a placeholder to later contain the swash
15694 * computed from av[0]. But if no further computation need be done,
15695 * the swash is stored here now (and av[0] is &PL_sv_undef).
15696 * av[2] stores the inversion list of code points that match only if the
15697 * current locale is UTF-8
15698 * av[3] stores the cp_list inversion list for use in addition or instead
15699 * of av[0]; used only if cp_list exists and av[1] is &PL_sv_undef.
15700 * (Otherwise everything needed is already in av[0] and av[1])
15701 * av[4] is set if any component of the class is from a user-defined
15702 * property; used only if av[3] exists */
15706 PERL_ARGS_ASSERT_SET_ANYOF_ARG;
15708 if (! cp_list && ! runtime_defns && ! only_utf8_locale_list) {
15709 assert(! (ANYOF_FLAGS(node)
15710 & (ANYOF_HAS_UTF8_NONBITMAP_MATCHES
15711 |ANYOF_HAS_NONBITMAP_NON_UTF8_MATCHES)));
15712 ARG_SET(node, ANYOF_ONLY_HAS_BITMAP);
15715 AV * const av = newAV();
15718 assert(ANYOF_FLAGS(node)
15719 & (ANYOF_HAS_UTF8_NONBITMAP_MATCHES
15720 |ANYOF_HAS_NONBITMAP_NON_UTF8_MATCHES|ANYOF_LOC_FOLD));
15722 av_store(av, 0, (runtime_defns)
15723 ? SvREFCNT_inc(runtime_defns) : &PL_sv_undef);
15726 av_store(av, 1, swash);
15727 SvREFCNT_dec_NN(cp_list);
15730 av_store(av, 1, &PL_sv_undef);
15732 av_store(av, 3, cp_list);
15733 av_store(av, 4, newSVuv(has_user_defined_property));
15737 if (only_utf8_locale_list) {
15738 av_store(av, 2, only_utf8_locale_list);
15741 av_store(av, 2, &PL_sv_undef);
15744 rv = newRV_noinc(MUTABLE_SV(av));
15745 n = add_data(pRExC_state, STR_WITH_LEN("s"));
15746 RExC_rxi->data->data[n] = (void*)rv;
15751 #if !defined(PERL_IN_XSUB_RE) || defined(PLUGGABLE_RE_EXTENSION)
15753 Perl__get_regclass_nonbitmap_data(pTHX_ const regexp *prog,
15754 const regnode* node,
15757 SV** only_utf8_locale_ptr,
15761 /* For internal core use only.
15762 * Returns the swash for the input 'node' in the regex 'prog'.
15763 * If <doinit> is 'true', will attempt to create the swash if not already
15765 * If <listsvp> is non-null, will return the printable contents of the
15766 * swash. This can be used to get debugging information even before the
15767 * swash exists, by calling this function with 'doinit' set to false, in
15768 * which case the components that will be used to eventually create the
15769 * swash are returned (in a printable form).
15770 * If <exclude_list> is not NULL, it is an inversion list of things to
15771 * exclude from what's returned in <listsvp>.
15772 * Tied intimately to how S_set_ANYOF_arg sets up the data structure. Note
15773 * that, in spite of this function's name, the swash it returns may include
15774 * the bitmap data as well */
15777 SV *si = NULL; /* Input swash initialization string */
15778 SV* invlist = NULL;
15780 RXi_GET_DECL(prog,progi);
15781 const struct reg_data * const data = prog ? progi->data : NULL;
15783 PERL_ARGS_ASSERT__GET_REGCLASS_NONBITMAP_DATA;
15785 assert(ANYOF_FLAGS(node)
15786 & (ANYOF_HAS_UTF8_NONBITMAP_MATCHES
15787 |ANYOF_HAS_NONBITMAP_NON_UTF8_MATCHES|ANYOF_LOC_FOLD));
15789 if (data && data->count) {
15790 const U32 n = ARG(node);
15792 if (data->what[n] == 's') {
15793 SV * const rv = MUTABLE_SV(data->data[n]);
15794 AV * const av = MUTABLE_AV(SvRV(rv));
15795 SV **const ary = AvARRAY(av);
15796 U8 swash_init_flags = _CORE_SWASH_INIT_ACCEPT_INVLIST;
15798 si = *ary; /* ary[0] = the string to initialize the swash with */
15800 /* Elements 3 and 4 are either both present or both absent. [3] is
15801 * any inversion list generated at compile time; [4] indicates if
15802 * that inversion list has any user-defined properties in it. */
15803 if (av_tindex(av) >= 2) {
15804 if (only_utf8_locale_ptr
15806 && ary[2] != &PL_sv_undef)
15808 *only_utf8_locale_ptr = ary[2];
15811 assert(only_utf8_locale_ptr);
15812 *only_utf8_locale_ptr = NULL;
15815 if (av_tindex(av) >= 3) {
15817 if (SvUV(ary[4])) {
15818 swash_init_flags |= _CORE_SWASH_INIT_USER_DEFINED_PROPERTY;
15826 /* Element [1] is reserved for the set-up swash. If already there,
15827 * return it; if not, create it and store it there */
15828 if (ary[1] && SvROK(ary[1])) {
15831 else if (doinit && ((si && si != &PL_sv_undef)
15832 || (invlist && invlist != &PL_sv_undef))) {
15834 sw = _core_swash_init("utf8", /* the utf8 package */
15838 0, /* not from tr/// */
15840 &swash_init_flags);
15841 (void)av_store(av, 1, sw);
15846 /* If requested, return a printable version of what this swash matches */
15848 SV* matches_string = newSVpvs("");
15850 /* The swash should be used, if possible, to get the data, as it
15851 * contains the resolved data. But this function can be called at
15852 * compile-time, before everything gets resolved, in which case we
15853 * return the currently best available information, which is the string
15854 * that will eventually be used to do that resolving, 'si' */
15855 if ((! sw || (invlist = _get_swash_invlist(sw)) == NULL)
15856 && (si && si != &PL_sv_undef))
15858 sv_catsv(matches_string, si);
15861 /* Add the inversion list to whatever we have. This may have come from
15862 * the swash, or from an input parameter */
15864 if (exclude_list) {
15865 SV* clone = invlist_clone(invlist);
15866 _invlist_subtract(clone, exclude_list, &clone);
15867 sv_catsv(matches_string, _invlist_contents(clone));
15868 SvREFCNT_dec_NN(clone);
15871 sv_catsv(matches_string, _invlist_contents(invlist));
15874 *listsvp = matches_string;
15879 #endif /* !defined(PERL_IN_XSUB_RE) || defined(PLUGGABLE_RE_EXTENSION) */
15881 /* reg_skipcomment()
15883 Absorbs an /x style # comment from the input stream,
15884 returning a pointer to the first character beyond the comment, or if the
15885 comment terminates the pattern without anything following it, this returns
15886 one past the final character of the pattern (in other words, RExC_end) and
15887 sets the REG_RUN_ON_COMMENT_SEEN flag.
15889 Note it's the callers responsibility to ensure that we are
15890 actually in /x mode
15894 PERL_STATIC_INLINE char*
15895 S_reg_skipcomment(RExC_state_t *pRExC_state, char* p)
15897 PERL_ARGS_ASSERT_REG_SKIPCOMMENT;
15901 while (p < RExC_end) {
15902 if (*(++p) == '\n') {
15907 /* we ran off the end of the pattern without ending the comment, so we have
15908 * to add an \n when wrapping */
15909 RExC_seen |= REG_RUN_ON_COMMENT_SEEN;
15915 Advances the parse position, and optionally absorbs
15916 "whitespace" from the inputstream.
15918 Without /x "whitespace" means (?#...) style comments only,
15919 with /x this means (?#...) and # comments and whitespace proper.
15921 Returns the RExC_parse point from BEFORE the scan occurs.
15923 This is the /x friendly way of saying RExC_parse++.
15927 S_nextchar(pTHX_ RExC_state_t *pRExC_state)
15929 char* const retval = RExC_parse++;
15931 PERL_ARGS_ASSERT_NEXTCHAR;
15934 if (RExC_end - RExC_parse >= 3
15935 && *RExC_parse == '('
15936 && RExC_parse[1] == '?'
15937 && RExC_parse[2] == '#')
15939 while (*RExC_parse != ')') {
15940 if (RExC_parse == RExC_end)
15941 FAIL("Sequence (?#... not terminated");
15947 if (RExC_flags & RXf_PMf_EXTENDED) {
15948 char * p = regpatws(pRExC_state, RExC_parse,
15949 TRUE); /* means recognize comments */
15950 if (p != RExC_parse) {
15960 S_regnode_guts(pTHX_ RExC_state_t *pRExC_state, const U8 op, const STRLEN extra_size, const char* const name)
15962 /* Allocate a regnode for 'op' and returns it, with 'extra_size' extra
15963 * space. In pass1, it aligns and increments RExC_size; in pass2,
15966 regnode * const ret = RExC_emit;
15967 GET_RE_DEBUG_FLAGS_DECL;
15969 PERL_ARGS_ASSERT_REGNODE_GUTS;
15971 assert(extra_size >= regarglen[op]);
15974 SIZE_ALIGN(RExC_size);
15975 RExC_size += 1 + extra_size;
15978 if (RExC_emit >= RExC_emit_bound)
15979 Perl_croak(aTHX_ "panic: reg_node overrun trying to emit %d, %p>=%p",
15980 op, (void*)RExC_emit, (void*)RExC_emit_bound);
15982 NODE_ALIGN_FILL(ret);
15983 #ifndef RE_TRACK_PATTERN_OFFSETS
15984 PERL_UNUSED_ARG(name);
15986 if (RExC_offsets) { /* MJD */
15988 ("%s:%d: (op %s) %s %"UVuf" (len %"UVuf") (max %"UVuf").\n",
15991 (UV)(RExC_emit - RExC_emit_start) > RExC_offsets[0]
15992 ? "Overwriting end of array!\n" : "OK",
15993 (UV)(RExC_emit - RExC_emit_start),
15994 (UV)(RExC_parse - RExC_start),
15995 (UV)RExC_offsets[0]));
15996 Set_Node_Offset(RExC_emit, RExC_parse + (op == END));
16003 - reg_node - emit a node
16005 STATIC regnode * /* Location. */
16006 S_reg_node(pTHX_ RExC_state_t *pRExC_state, U8 op)
16008 regnode * const ret = regnode_guts(pRExC_state, op, regarglen[op], "reg_node");
16010 PERL_ARGS_ASSERT_REG_NODE;
16012 assert(regarglen[op] == 0);
16015 regnode *ptr = ret;
16016 FILL_ADVANCE_NODE(ptr, op);
16023 - reganode - emit a node with an argument
16025 STATIC regnode * /* Location. */
16026 S_reganode(pTHX_ RExC_state_t *pRExC_state, U8 op, U32 arg)
16028 regnode * const ret = regnode_guts(pRExC_state, op, regarglen[op], "reganode");
16030 PERL_ARGS_ASSERT_REGANODE;
16032 assert(regarglen[op] == 1);
16035 regnode *ptr = ret;
16036 FILL_ADVANCE_NODE_ARG(ptr, op, arg);
16043 S_reg2Lanode(pTHX_ RExC_state_t *pRExC_state, const U8 op, const U32 arg1, const I32 arg2)
16045 /* emit a node with U32 and I32 arguments */
16047 regnode * const ret = regnode_guts(pRExC_state, op, regarglen[op], "reg2Lanode");
16049 PERL_ARGS_ASSERT_REG2LANODE;
16051 assert(regarglen[op] == 2);
16054 regnode *ptr = ret;
16055 FILL_ADVANCE_NODE_2L_ARG(ptr, op, arg1, arg2);
16062 - reginsert - insert an operator in front of already-emitted operand
16064 * Means relocating the operand.
16067 S_reginsert(pTHX_ RExC_state_t *pRExC_state, U8 op, regnode *opnd, U32 depth)
16072 const int offset = regarglen[(U8)op];
16073 const int size = NODE_STEP_REGNODE + offset;
16074 GET_RE_DEBUG_FLAGS_DECL;
16076 PERL_ARGS_ASSERT_REGINSERT;
16077 PERL_UNUSED_CONTEXT;
16078 PERL_UNUSED_ARG(depth);
16079 /* (PL_regkind[(U8)op] == CURLY ? EXTRA_STEP_2ARGS : 0); */
16080 DEBUG_PARSE_FMT("inst"," - %s",PL_reg_name[op]);
16089 if (RExC_open_parens) {
16091 /*DEBUG_PARSE_FMT("inst"," - %"IVdf, (IV)RExC_npar);*/
16092 for ( paren=0 ; paren < RExC_npar ; paren++ ) {
16093 if ( RExC_open_parens[paren] >= opnd ) {
16094 /*DEBUG_PARSE_FMT("open"," - %d",size);*/
16095 RExC_open_parens[paren] += size;
16097 /*DEBUG_PARSE_FMT("open"," - %s","ok");*/
16099 if ( RExC_close_parens[paren] >= opnd ) {
16100 /*DEBUG_PARSE_FMT("close"," - %d",size);*/
16101 RExC_close_parens[paren] += size;
16103 /*DEBUG_PARSE_FMT("close"," - %s","ok");*/
16108 while (src > opnd) {
16109 StructCopy(--src, --dst, regnode);
16110 #ifdef RE_TRACK_PATTERN_OFFSETS
16111 if (RExC_offsets) { /* MJD 20010112 */
16113 ("%s(%d): (op %s) %s copy %"UVuf" -> %"UVuf" (max %"UVuf").\n",
16117 (UV)(dst - RExC_emit_start) > RExC_offsets[0]
16118 ? "Overwriting end of array!\n" : "OK",
16119 (UV)(src - RExC_emit_start),
16120 (UV)(dst - RExC_emit_start),
16121 (UV)RExC_offsets[0]));
16122 Set_Node_Offset_To_R(dst-RExC_emit_start, Node_Offset(src));
16123 Set_Node_Length_To_R(dst-RExC_emit_start, Node_Length(src));
16129 place = opnd; /* Op node, where operand used to be. */
16130 #ifdef RE_TRACK_PATTERN_OFFSETS
16131 if (RExC_offsets) { /* MJD */
16133 ("%s(%d): (op %s) %s %"UVuf" <- %"UVuf" (max %"UVuf").\n",
16137 (UV)(place - RExC_emit_start) > RExC_offsets[0]
16138 ? "Overwriting end of array!\n" : "OK",
16139 (UV)(place - RExC_emit_start),
16140 (UV)(RExC_parse - RExC_start),
16141 (UV)RExC_offsets[0]));
16142 Set_Node_Offset(place, RExC_parse);
16143 Set_Node_Length(place, 1);
16146 src = NEXTOPER(place);
16147 FILL_ADVANCE_NODE(place, op);
16148 Zero(src, offset, regnode);
16152 - regtail - set the next-pointer at the end of a node chain of p to val.
16153 - SEE ALSO: regtail_study
16155 /* TODO: All three parms should be const */
16157 S_regtail(pTHX_ RExC_state_t *pRExC_state, regnode *p,
16158 const regnode *val,U32 depth)
16161 GET_RE_DEBUG_FLAGS_DECL;
16163 PERL_ARGS_ASSERT_REGTAIL;
16165 PERL_UNUSED_ARG(depth);
16171 /* Find last node. */
16174 regnode * const temp = regnext(scan);
16176 DEBUG_PARSE_MSG((scan==p ? "tail" : ""));
16177 regprop(RExC_rx, RExC_mysv, scan, NULL, pRExC_state);
16178 PerlIO_printf(Perl_debug_log, "~ %s (%d) %s %s\n",
16179 SvPV_nolen_const(RExC_mysv), REG_NODE_NUM(scan),
16180 (temp == NULL ? "->" : ""),
16181 (temp == NULL ? PL_reg_name[OP(val)] : "")
16189 if (reg_off_by_arg[OP(scan)]) {
16190 ARG_SET(scan, val - scan);
16193 NEXT_OFF(scan) = val - scan;
16199 - regtail_study - set the next-pointer at the end of a node chain of p to val.
16200 - Look for optimizable sequences at the same time.
16201 - currently only looks for EXACT chains.
16203 This is experimental code. The idea is to use this routine to perform
16204 in place optimizations on branches and groups as they are constructed,
16205 with the long term intention of removing optimization from study_chunk so
16206 that it is purely analytical.
16208 Currently only used when in DEBUG mode. The macro REGTAIL_STUDY() is used
16209 to control which is which.
16212 /* TODO: All four parms should be const */
16215 S_regtail_study(pTHX_ RExC_state_t *pRExC_state, regnode *p,
16216 const regnode *val,U32 depth)
16220 #ifdef EXPERIMENTAL_INPLACESCAN
16223 GET_RE_DEBUG_FLAGS_DECL;
16225 PERL_ARGS_ASSERT_REGTAIL_STUDY;
16231 /* Find last node. */
16235 regnode * const temp = regnext(scan);
16236 #ifdef EXPERIMENTAL_INPLACESCAN
16237 if (PL_regkind[OP(scan)] == EXACT) {
16238 bool unfolded_multi_char; /* Unexamined in this routine */
16239 if (join_exact(pRExC_state, scan, &min,
16240 &unfolded_multi_char, 1, val, depth+1))
16245 switch (OP(scan)) {
16249 case EXACTFA_NO_TRIE:
16255 if( exact == PSEUDO )
16257 else if ( exact != OP(scan) )
16266 DEBUG_PARSE_MSG((scan==p ? "tsdy" : ""));
16267 regprop(RExC_rx, RExC_mysv, scan, NULL, pRExC_state);
16268 PerlIO_printf(Perl_debug_log, "~ %s (%d) -> %s\n",
16269 SvPV_nolen_const(RExC_mysv),
16270 REG_NODE_NUM(scan),
16271 PL_reg_name[exact]);
16278 DEBUG_PARSE_MSG("");
16279 regprop(RExC_rx, RExC_mysv, val, NULL, pRExC_state);
16280 PerlIO_printf(Perl_debug_log,
16281 "~ attach to %s (%"IVdf") offset to %"IVdf"\n",
16282 SvPV_nolen_const(RExC_mysv),
16283 (IV)REG_NODE_NUM(val),
16287 if (reg_off_by_arg[OP(scan)]) {
16288 ARG_SET(scan, val - scan);
16291 NEXT_OFF(scan) = val - scan;
16299 - regdump - dump a regexp onto Perl_debug_log in vaguely comprehensible form
16304 S_regdump_intflags(pTHX_ const char *lead, const U32 flags)
16309 ASSUME(REG_INTFLAGS_NAME_SIZE <= sizeof(flags)*8);
16311 for (bit=0; bit<REG_INTFLAGS_NAME_SIZE; bit++) {
16312 if (flags & (1<<bit)) {
16313 if (!set++ && lead)
16314 PerlIO_printf(Perl_debug_log, "%s",lead);
16315 PerlIO_printf(Perl_debug_log, "%s ",PL_reg_intflags_name[bit]);
16320 PerlIO_printf(Perl_debug_log, "\n");
16322 PerlIO_printf(Perl_debug_log, "%s[none-set]\n",lead);
16327 S_regdump_extflags(pTHX_ const char *lead, const U32 flags)
16333 ASSUME(REG_EXTFLAGS_NAME_SIZE <= sizeof(flags)*8);
16335 for (bit=0; bit<REG_EXTFLAGS_NAME_SIZE; bit++) {
16336 if (flags & (1<<bit)) {
16337 if ((1<<bit) & RXf_PMf_CHARSET) { /* Output separately, below */
16340 if (!set++ && lead)
16341 PerlIO_printf(Perl_debug_log, "%s",lead);
16342 PerlIO_printf(Perl_debug_log, "%s ",PL_reg_extflags_name[bit]);
16345 if ((cs = get_regex_charset(flags)) != REGEX_DEPENDS_CHARSET) {
16346 if (!set++ && lead) {
16347 PerlIO_printf(Perl_debug_log, "%s",lead);
16350 case REGEX_UNICODE_CHARSET:
16351 PerlIO_printf(Perl_debug_log, "UNICODE");
16353 case REGEX_LOCALE_CHARSET:
16354 PerlIO_printf(Perl_debug_log, "LOCALE");
16356 case REGEX_ASCII_RESTRICTED_CHARSET:
16357 PerlIO_printf(Perl_debug_log, "ASCII-RESTRICTED");
16359 case REGEX_ASCII_MORE_RESTRICTED_CHARSET:
16360 PerlIO_printf(Perl_debug_log, "ASCII-MORE_RESTRICTED");
16363 PerlIO_printf(Perl_debug_log, "UNKNOWN CHARACTER SET");
16369 PerlIO_printf(Perl_debug_log, "\n");
16371 PerlIO_printf(Perl_debug_log, "%s[none-set]\n",lead);
16377 Perl_regdump(pTHX_ const regexp *r)
16380 SV * const sv = sv_newmortal();
16381 SV *dsv= sv_newmortal();
16382 RXi_GET_DECL(r,ri);
16383 GET_RE_DEBUG_FLAGS_DECL;
16385 PERL_ARGS_ASSERT_REGDUMP;
16387 (void)dumpuntil(r, ri->program, ri->program + 1, NULL, NULL, sv, 0, 0);
16389 /* Header fields of interest. */
16390 if (r->anchored_substr) {
16391 RE_PV_QUOTED_DECL(s, 0, dsv, SvPVX_const(r->anchored_substr),
16392 RE_SV_DUMPLEN(r->anchored_substr), 30);
16393 PerlIO_printf(Perl_debug_log,
16394 "anchored %s%s at %"IVdf" ",
16395 s, RE_SV_TAIL(r->anchored_substr),
16396 (IV)r->anchored_offset);
16397 } else if (r->anchored_utf8) {
16398 RE_PV_QUOTED_DECL(s, 1, dsv, SvPVX_const(r->anchored_utf8),
16399 RE_SV_DUMPLEN(r->anchored_utf8), 30);
16400 PerlIO_printf(Perl_debug_log,
16401 "anchored utf8 %s%s at %"IVdf" ",
16402 s, RE_SV_TAIL(r->anchored_utf8),
16403 (IV)r->anchored_offset);
16405 if (r->float_substr) {
16406 RE_PV_QUOTED_DECL(s, 0, dsv, SvPVX_const(r->float_substr),
16407 RE_SV_DUMPLEN(r->float_substr), 30);
16408 PerlIO_printf(Perl_debug_log,
16409 "floating %s%s at %"IVdf"..%"UVuf" ",
16410 s, RE_SV_TAIL(r->float_substr),
16411 (IV)r->float_min_offset, (UV)r->float_max_offset);
16412 } else if (r->float_utf8) {
16413 RE_PV_QUOTED_DECL(s, 1, dsv, SvPVX_const(r->float_utf8),
16414 RE_SV_DUMPLEN(r->float_utf8), 30);
16415 PerlIO_printf(Perl_debug_log,
16416 "floating utf8 %s%s at %"IVdf"..%"UVuf" ",
16417 s, RE_SV_TAIL(r->float_utf8),
16418 (IV)r->float_min_offset, (UV)r->float_max_offset);
16420 if (r->check_substr || r->check_utf8)
16421 PerlIO_printf(Perl_debug_log,
16423 (r->check_substr == r->float_substr
16424 && r->check_utf8 == r->float_utf8
16425 ? "(checking floating" : "(checking anchored"));
16426 if (r->intflags & PREGf_NOSCAN)
16427 PerlIO_printf(Perl_debug_log, " noscan");
16428 if (r->extflags & RXf_CHECK_ALL)
16429 PerlIO_printf(Perl_debug_log, " isall");
16430 if (r->check_substr || r->check_utf8)
16431 PerlIO_printf(Perl_debug_log, ") ");
16433 if (ri->regstclass) {
16434 regprop(r, sv, ri->regstclass, NULL, NULL);
16435 PerlIO_printf(Perl_debug_log, "stclass %s ", SvPVX_const(sv));
16437 if (r->intflags & PREGf_ANCH) {
16438 PerlIO_printf(Perl_debug_log, "anchored");
16439 if (r->intflags & PREGf_ANCH_MBOL)
16440 PerlIO_printf(Perl_debug_log, "(MBOL)");
16441 if (r->intflags & PREGf_ANCH_SBOL)
16442 PerlIO_printf(Perl_debug_log, "(SBOL)");
16443 if (r->intflags & PREGf_ANCH_GPOS)
16444 PerlIO_printf(Perl_debug_log, "(GPOS)");
16445 PerlIO_putc(Perl_debug_log, ' ');
16447 if (r->intflags & PREGf_GPOS_SEEN)
16448 PerlIO_printf(Perl_debug_log, "GPOS:%"UVuf" ", (UV)r->gofs);
16449 if (r->intflags & PREGf_SKIP)
16450 PerlIO_printf(Perl_debug_log, "plus ");
16451 if (r->intflags & PREGf_IMPLICIT)
16452 PerlIO_printf(Perl_debug_log, "implicit ");
16453 PerlIO_printf(Perl_debug_log, "minlen %"IVdf" ", (IV)r->minlen);
16454 if (r->extflags & RXf_EVAL_SEEN)
16455 PerlIO_printf(Perl_debug_log, "with eval ");
16456 PerlIO_printf(Perl_debug_log, "\n");
16458 regdump_extflags("r->extflags: ",r->extflags);
16459 regdump_intflags("r->intflags: ",r->intflags);
16462 PERL_ARGS_ASSERT_REGDUMP;
16463 PERL_UNUSED_CONTEXT;
16464 PERL_UNUSED_ARG(r);
16465 #endif /* DEBUGGING */
16469 - regprop - printable representation of opcode, with run time support
16473 Perl_regprop(pTHX_ const regexp *prog, SV *sv, const regnode *o, const regmatch_info *reginfo, const RExC_state_t *pRExC_state)
16478 /* Should be synchronized with * ANYOF_ #xdefines in regcomp.h */
16479 static const char * const anyofs[] = {
16480 #if _CC_WORDCHAR != 0 || _CC_DIGIT != 1 || _CC_ALPHA != 2 || _CC_LOWER != 3 \
16481 || _CC_UPPER != 4 || _CC_PUNCT != 5 || _CC_PRINT != 6 \
16482 || _CC_ALPHANUMERIC != 7 || _CC_GRAPH != 8 || _CC_CASED != 9 \
16483 || _CC_SPACE != 10 || _CC_BLANK != 11 || _CC_XDIGIT != 12 \
16484 || _CC_CNTRL != 13 || _CC_ASCII != 14 || _CC_VERTSPACE != 15
16485 #error Need to adjust order of anyofs[]
16520 RXi_GET_DECL(prog,progi);
16521 GET_RE_DEBUG_FLAGS_DECL;
16523 PERL_ARGS_ASSERT_REGPROP;
16525 sv_setpvn(sv, "", 0);
16527 if (OP(o) > REGNODE_MAX) /* regnode.type is unsigned */
16528 /* It would be nice to FAIL() here, but this may be called from
16529 regexec.c, and it would be hard to supply pRExC_state. */
16530 Perl_croak(aTHX_ "Corrupted regexp opcode %d > %d",
16531 (int)OP(o), (int)REGNODE_MAX);
16532 sv_catpv(sv, PL_reg_name[OP(o)]); /* Take off const! */
16534 k = PL_regkind[OP(o)];
16537 sv_catpvs(sv, " ");
16538 /* Using is_utf8_string() (via PERL_PV_UNI_DETECT)
16539 * is a crude hack but it may be the best for now since
16540 * we have no flag "this EXACTish node was UTF-8"
16542 pv_pretty(sv, STRING(o), STR_LEN(o), 60, PL_colors[0], PL_colors[1],
16543 PERL_PV_ESCAPE_UNI_DETECT |
16544 PERL_PV_ESCAPE_NONASCII |
16545 PERL_PV_PRETTY_ELLIPSES |
16546 PERL_PV_PRETTY_LTGT |
16547 PERL_PV_PRETTY_NOCLEAR
16549 } else if (k == TRIE) {
16550 /* print the details of the trie in dumpuntil instead, as
16551 * progi->data isn't available here */
16552 const char op = OP(o);
16553 const U32 n = ARG(o);
16554 const reg_ac_data * const ac = IS_TRIE_AC(op) ?
16555 (reg_ac_data *)progi->data->data[n] :
16557 const reg_trie_data * const trie
16558 = (reg_trie_data*)progi->data->data[!IS_TRIE_AC(op) ? n : ac->trie];
16560 Perl_sv_catpvf(aTHX_ sv, "-%s",PL_reg_name[o->flags]);
16561 DEBUG_TRIE_COMPILE_r(
16562 Perl_sv_catpvf(aTHX_ sv,
16563 "<S:%"UVuf"/%"IVdf" W:%"UVuf" L:%"UVuf"/%"UVuf" C:%"UVuf"/%"UVuf">",
16564 (UV)trie->startstate,
16565 (IV)trie->statecount-1, /* -1 because of the unused 0 element */
16566 (UV)trie->wordcount,
16569 (UV)TRIE_CHARCOUNT(trie),
16570 (UV)trie->uniquecharcount
16573 if ( IS_ANYOF_TRIE(op) || trie->bitmap ) {
16574 sv_catpvs(sv, "[");
16575 (void) put_charclass_bitmap_innards(sv,
16576 (IS_ANYOF_TRIE(op))
16578 : TRIE_BITMAP(trie),
16580 sv_catpvs(sv, "]");
16583 } else if (k == CURLY) {
16584 if (OP(o) == CURLYM || OP(o) == CURLYN || OP(o) == CURLYX)
16585 Perl_sv_catpvf(aTHX_ sv, "[%d]", o->flags); /* Parenth number */
16586 Perl_sv_catpvf(aTHX_ sv, " {%d,%d}", ARG1(o), ARG2(o));
16588 else if (k == WHILEM && o->flags) /* Ordinal/of */
16589 Perl_sv_catpvf(aTHX_ sv, "[%d/%d]", o->flags & 0xf, o->flags>>4);
16590 else if (k == REF || k == OPEN || k == CLOSE
16591 || k == GROUPP || OP(o)==ACCEPT)
16593 AV *name_list= NULL;
16594 Perl_sv_catpvf(aTHX_ sv, "%d", (int)ARG(o)); /* Parenth number */
16595 if ( RXp_PAREN_NAMES(prog) ) {
16596 name_list= MUTABLE_AV(progi->data->data[progi->name_list_idx]);
16597 } else if ( pRExC_state ) {
16598 name_list= RExC_paren_name_list;
16601 if ( k != REF || (OP(o) < NREF)) {
16602 SV **name= av_fetch(name_list, ARG(o), 0 );
16604 Perl_sv_catpvf(aTHX_ sv, " '%"SVf"'", SVfARG(*name));
16607 SV *sv_dat= MUTABLE_SV(progi->data->data[ ARG( o ) ]);
16608 I32 *nums=(I32*)SvPVX(sv_dat);
16609 SV **name= av_fetch(name_list, nums[0], 0 );
16612 for ( n=0; n<SvIVX(sv_dat); n++ ) {
16613 Perl_sv_catpvf(aTHX_ sv, "%s%"IVdf,
16614 (n ? "," : ""), (IV)nums[n]);
16616 Perl_sv_catpvf(aTHX_ sv, " '%"SVf"'", SVfARG(*name));
16620 if ( k == REF && reginfo) {
16621 U32 n = ARG(o); /* which paren pair */
16622 I32 ln = prog->offs[n].start;
16623 if (prog->lastparen < n || ln == -1)
16624 Perl_sv_catpvf(aTHX_ sv, ": FAIL");
16625 else if (ln == prog->offs[n].end)
16626 Perl_sv_catpvf(aTHX_ sv, ": ACCEPT - EMPTY STRING");
16628 const char *s = reginfo->strbeg + ln;
16629 Perl_sv_catpvf(aTHX_ sv, ": ");
16630 Perl_pv_pretty( aTHX_ sv, s, prog->offs[n].end - prog->offs[n].start, 32, 0, 0,
16631 PERL_PV_ESCAPE_UNI_DETECT|PERL_PV_PRETTY_NOCLEAR|PERL_PV_PRETTY_ELLIPSES|PERL_PV_PRETTY_QUOTE );
16634 } else if (k == GOSUB) {
16635 AV *name_list= NULL;
16636 if ( RXp_PAREN_NAMES(prog) ) {
16637 name_list= MUTABLE_AV(progi->data->data[progi->name_list_idx]);
16638 } else if ( pRExC_state ) {
16639 name_list= RExC_paren_name_list;
16642 /* Paren and offset */
16643 Perl_sv_catpvf(aTHX_ sv, "%d[%+d]", (int)ARG(o),(int)ARG2L(o));
16645 SV **name= av_fetch(name_list, ARG(o), 0 );
16647 Perl_sv_catpvf(aTHX_ sv, " '%"SVf"'", SVfARG(*name));
16650 else if (k == VERB) {
16652 Perl_sv_catpvf(aTHX_ sv, ":%"SVf,
16653 SVfARG((MUTABLE_SV(progi->data->data[ ARG( o ) ]))));
16654 } else if (k == LOGICAL)
16655 /* 2: embedded, otherwise 1 */
16656 Perl_sv_catpvf(aTHX_ sv, "[%d]", o->flags);
16657 else if (k == ANYOF) {
16658 const U8 flags = ANYOF_FLAGS(o);
16660 SV* bitmap_invlist; /* Will hold what the bit map contains */
16663 if (OP(o) == ANYOFL)
16664 sv_catpvs(sv, "{loc}");
16665 if (flags & ANYOF_LOC_FOLD)
16666 sv_catpvs(sv, "{i}");
16667 Perl_sv_catpvf(aTHX_ sv, "[%s", PL_colors[0]);
16668 if (flags & ANYOF_INVERT)
16669 sv_catpvs(sv, "^");
16671 /* output what the standard cp 0-NUM_ANYOF_CODE_POINTS-1 bitmap matches
16673 do_sep = put_charclass_bitmap_innards(sv, ANYOF_BITMAP(o),
16676 /* output any special charclass tests (used entirely under use
16678 if (ANYOF_POSIXL_TEST_ANY_SET(o)) {
16680 for (i = 0; i < ANYOF_POSIXL_MAX; i++) {
16681 if (ANYOF_POSIXL_TEST(o,i)) {
16682 sv_catpv(sv, anyofs[i]);
16688 if ((flags & (ANYOF_MATCHES_ALL_ABOVE_BITMAP
16689 |ANYOF_HAS_UTF8_NONBITMAP_MATCHES
16690 |ANYOF_HAS_NONBITMAP_NON_UTF8_MATCHES
16694 Perl_sv_catpvf(aTHX_ sv,"%s][%s",PL_colors[1],PL_colors[0]);
16695 if (flags & ANYOF_INVERT)
16696 /*make sure the invert info is in each */
16697 sv_catpvs(sv, "^");
16700 if (flags & ANYOF_MATCHES_ALL_NON_UTF8_NON_ASCII) {
16701 sv_catpvs(sv, "{non-utf8-latin1-all}");
16704 if (flags & ANYOF_MATCHES_ALL_ABOVE_BITMAP)
16705 sv_catpvs(sv, "{above_bitmap_all}");
16707 if (ARG(o) != ANYOF_ONLY_HAS_BITMAP) {
16708 SV *lv; /* Set if there is something outside the bit map. */
16709 bool byte_output = FALSE; /* If something has been output */
16710 SV *only_utf8_locale;
16712 /* Get the stuff that wasn't in the bitmap. 'bitmap_invlist'
16713 * is used to guarantee that nothing in the bitmap gets
16715 (void) _get_regclass_nonbitmap_data(prog, o, FALSE,
16716 &lv, &only_utf8_locale,
16718 if (lv && lv != &PL_sv_undef) {
16719 char *s = savesvpv(lv);
16720 char * const origs = s;
16722 while (*s && *s != '\n')
16726 const char * const t = ++s;
16728 if (flags & ANYOF_HAS_NONBITMAP_NON_UTF8_MATCHES) {
16729 sv_catpvs(sv, "{outside bitmap}");
16732 sv_catpvs(sv, "{utf8}");
16736 sv_catpvs(sv, " ");
16742 /* Truncate very long output */
16743 if (s - origs > 256) {
16744 Perl_sv_catpvf(aTHX_ sv,
16746 (int) (s - origs - 1),
16752 else if (*s == '\t') {
16766 SvREFCNT_dec_NN(lv);
16769 if ((flags & ANYOF_LOC_FOLD)
16770 && only_utf8_locale
16771 && only_utf8_locale != &PL_sv_undef)
16774 int max_entries = 256;
16776 sv_catpvs(sv, "{utf8 locale}");
16777 invlist_iterinit(only_utf8_locale);
16778 while (invlist_iternext(only_utf8_locale,
16780 put_range(sv, start, end, FALSE);
16782 if (max_entries < 0) {
16783 sv_catpvs(sv, "...");
16787 invlist_iterfinish(only_utf8_locale);
16791 SvREFCNT_dec(bitmap_invlist);
16794 Perl_sv_catpvf(aTHX_ sv, "%s]", PL_colors[1]);
16796 else if (k == POSIXD || k == NPOSIXD) {
16797 U8 index = FLAGS(o) * 2;
16798 if (index < C_ARRAY_LENGTH(anyofs)) {
16799 if (*anyofs[index] != '[') {
16802 sv_catpv(sv, anyofs[index]);
16803 if (*anyofs[index] != '[') {
16808 Perl_sv_catpvf(aTHX_ sv, "[illegal type=%d])", index);
16811 else if (k == BOUND || k == NBOUND) {
16812 /* Must be synced with order of 'bound_type' in regcomp.h */
16813 const char * const bounds[] = {
16814 "", /* Traditional */
16819 sv_catpv(sv, bounds[FLAGS(o)]);
16821 else if (k == BRANCHJ && (OP(o) == UNLESSM || OP(o) == IFMATCH))
16822 Perl_sv_catpvf(aTHX_ sv, "[%d]", -(o->flags));
16823 else if (OP(o) == SBOL)
16824 Perl_sv_catpvf(aTHX_ sv, " /%s/", o->flags ? "\\A" : "^");
16826 PERL_UNUSED_CONTEXT;
16827 PERL_UNUSED_ARG(sv);
16828 PERL_UNUSED_ARG(o);
16829 PERL_UNUSED_ARG(prog);
16830 PERL_UNUSED_ARG(reginfo);
16831 PERL_UNUSED_ARG(pRExC_state);
16832 #endif /* DEBUGGING */
16838 Perl_re_intuit_string(pTHX_ REGEXP * const r)
16839 { /* Assume that RE_INTUIT is set */
16840 struct regexp *const prog = ReANY(r);
16841 GET_RE_DEBUG_FLAGS_DECL;
16843 PERL_ARGS_ASSERT_RE_INTUIT_STRING;
16844 PERL_UNUSED_CONTEXT;
16848 const char * const s = SvPV_nolen_const(RX_UTF8(r)
16849 ? prog->check_utf8 : prog->check_substr);
16851 if (!PL_colorset) reginitcolors();
16852 PerlIO_printf(Perl_debug_log,
16853 "%sUsing REx %ssubstr:%s \"%s%.60s%s%s\"\n",
16855 RX_UTF8(r) ? "utf8 " : "",
16856 PL_colors[5],PL_colors[0],
16859 (strlen(s) > 60 ? "..." : ""));
16862 /* use UTF8 check substring if regexp pattern itself is in UTF8 */
16863 return RX_UTF8(r) ? prog->check_utf8 : prog->check_substr;
16869 handles refcounting and freeing the perl core regexp structure. When
16870 it is necessary to actually free the structure the first thing it
16871 does is call the 'free' method of the regexp_engine associated to
16872 the regexp, allowing the handling of the void *pprivate; member
16873 first. (This routine is not overridable by extensions, which is why
16874 the extensions free is called first.)
16876 See regdupe and regdupe_internal if you change anything here.
16878 #ifndef PERL_IN_XSUB_RE
16880 Perl_pregfree(pTHX_ REGEXP *r)
16886 Perl_pregfree2(pTHX_ REGEXP *rx)
16888 struct regexp *const r = ReANY(rx);
16889 GET_RE_DEBUG_FLAGS_DECL;
16891 PERL_ARGS_ASSERT_PREGFREE2;
16893 if (r->mother_re) {
16894 ReREFCNT_dec(r->mother_re);
16896 CALLREGFREE_PVT(rx); /* free the private data */
16897 SvREFCNT_dec(RXp_PAREN_NAMES(r));
16898 Safefree(r->xpv_len_u.xpvlenu_pv);
16901 SvREFCNT_dec(r->anchored_substr);
16902 SvREFCNT_dec(r->anchored_utf8);
16903 SvREFCNT_dec(r->float_substr);
16904 SvREFCNT_dec(r->float_utf8);
16905 Safefree(r->substrs);
16907 RX_MATCH_COPY_FREE(rx);
16908 #ifdef PERL_ANY_COW
16909 SvREFCNT_dec(r->saved_copy);
16912 SvREFCNT_dec(r->qr_anoncv);
16913 rx->sv_u.svu_rx = 0;
16918 This is a hacky workaround to the structural issue of match results
16919 being stored in the regexp structure which is in turn stored in
16920 PL_curpm/PL_reg_curpm. The problem is that due to qr// the pattern
16921 could be PL_curpm in multiple contexts, and could require multiple
16922 result sets being associated with the pattern simultaneously, such
16923 as when doing a recursive match with (??{$qr})
16925 The solution is to make a lightweight copy of the regexp structure
16926 when a qr// is returned from the code executed by (??{$qr}) this
16927 lightweight copy doesn't actually own any of its data except for
16928 the starp/end and the actual regexp structure itself.
16934 Perl_reg_temp_copy (pTHX_ REGEXP *ret_x, REGEXP *rx)
16936 struct regexp *ret;
16937 struct regexp *const r = ReANY(rx);
16938 const bool islv = ret_x && SvTYPE(ret_x) == SVt_PVLV;
16940 PERL_ARGS_ASSERT_REG_TEMP_COPY;
16943 ret_x = (REGEXP*) newSV_type(SVt_REGEXP);
16945 SvOK_off((SV *)ret_x);
16947 /* For PVLVs, SvANY points to the xpvlv body while sv_u points
16948 to the regexp. (For SVt_REGEXPs, sv_upgrade has already
16949 made both spots point to the same regexp body.) */
16950 REGEXP *temp = (REGEXP *)newSV_type(SVt_REGEXP);
16951 assert(!SvPVX(ret_x));
16952 ret_x->sv_u.svu_rx = temp->sv_any;
16953 temp->sv_any = NULL;
16954 SvFLAGS(temp) = (SvFLAGS(temp) & ~SVTYPEMASK) | SVt_NULL;
16955 SvREFCNT_dec_NN(temp);
16956 /* SvCUR still resides in the xpvlv struct, so the regexp copy-
16957 ing below will not set it. */
16958 SvCUR_set(ret_x, SvCUR(rx));
16961 /* This ensures that SvTHINKFIRST(sv) is true, and hence that
16962 sv_force_normal(sv) is called. */
16964 ret = ReANY(ret_x);
16966 SvFLAGS(ret_x) |= SvUTF8(rx);
16967 /* We share the same string buffer as the original regexp, on which we
16968 hold a reference count, incremented when mother_re is set below.
16969 The string pointer is copied here, being part of the regexp struct.
16971 memcpy(&(ret->xpv_cur), &(r->xpv_cur),
16972 sizeof(regexp) - STRUCT_OFFSET(regexp, xpv_cur));
16974 const I32 npar = r->nparens+1;
16975 Newx(ret->offs, npar, regexp_paren_pair);
16976 Copy(r->offs, ret->offs, npar, regexp_paren_pair);
16979 Newx(ret->substrs, 1, struct reg_substr_data);
16980 StructCopy(r->substrs, ret->substrs, struct reg_substr_data);
16982 SvREFCNT_inc_void(ret->anchored_substr);
16983 SvREFCNT_inc_void(ret->anchored_utf8);
16984 SvREFCNT_inc_void(ret->float_substr);
16985 SvREFCNT_inc_void(ret->float_utf8);
16987 /* check_substr and check_utf8, if non-NULL, point to either their
16988 anchored or float namesakes, and don't hold a second reference. */
16990 RX_MATCH_COPIED_off(ret_x);
16991 #ifdef PERL_ANY_COW
16992 ret->saved_copy = NULL;
16994 ret->mother_re = ReREFCNT_inc(r->mother_re ? r->mother_re : rx);
16995 SvREFCNT_inc_void(ret->qr_anoncv);
17001 /* regfree_internal()
17003 Free the private data in a regexp. This is overloadable by
17004 extensions. Perl takes care of the regexp structure in pregfree(),
17005 this covers the *pprivate pointer which technically perl doesn't
17006 know about, however of course we have to handle the
17007 regexp_internal structure when no extension is in use.
17009 Note this is called before freeing anything in the regexp
17014 Perl_regfree_internal(pTHX_ REGEXP * const rx)
17016 struct regexp *const r = ReANY(rx);
17017 RXi_GET_DECL(r,ri);
17018 GET_RE_DEBUG_FLAGS_DECL;
17020 PERL_ARGS_ASSERT_REGFREE_INTERNAL;
17026 SV *dsv= sv_newmortal();
17027 RE_PV_QUOTED_DECL(s, RX_UTF8(rx),
17028 dsv, RX_PRECOMP(rx), RX_PRELEN(rx), 60);
17029 PerlIO_printf(Perl_debug_log,"%sFreeing REx:%s %s\n",
17030 PL_colors[4],PL_colors[5],s);
17033 #ifdef RE_TRACK_PATTERN_OFFSETS
17035 Safefree(ri->u.offsets); /* 20010421 MJD */
17037 if (ri->code_blocks) {
17039 for (n = 0; n < ri->num_code_blocks; n++)
17040 SvREFCNT_dec(ri->code_blocks[n].src_regex);
17041 Safefree(ri->code_blocks);
17045 int n = ri->data->count;
17048 /* If you add a ->what type here, update the comment in regcomp.h */
17049 switch (ri->data->what[n]) {
17055 SvREFCNT_dec(MUTABLE_SV(ri->data->data[n]));
17058 Safefree(ri->data->data[n]);
17064 { /* Aho Corasick add-on structure for a trie node.
17065 Used in stclass optimization only */
17067 reg_ac_data *aho=(reg_ac_data*)ri->data->data[n];
17068 #ifdef USE_ITHREADS
17072 refcount = --aho->refcount;
17075 PerlMemShared_free(aho->states);
17076 PerlMemShared_free(aho->fail);
17077 /* do this last!!!! */
17078 PerlMemShared_free(ri->data->data[n]);
17079 /* we should only ever get called once, so
17080 * assert as much, and also guard the free
17081 * which /might/ happen twice. At the least
17082 * it will make code anlyzers happy and it
17083 * doesn't cost much. - Yves */
17084 assert(ri->regstclass);
17085 if (ri->regstclass) {
17086 PerlMemShared_free(ri->regstclass);
17087 ri->regstclass = 0;
17094 /* trie structure. */
17096 reg_trie_data *trie=(reg_trie_data*)ri->data->data[n];
17097 #ifdef USE_ITHREADS
17101 refcount = --trie->refcount;
17104 PerlMemShared_free(trie->charmap);
17105 PerlMemShared_free(trie->states);
17106 PerlMemShared_free(trie->trans);
17108 PerlMemShared_free(trie->bitmap);
17110 PerlMemShared_free(trie->jump);
17111 PerlMemShared_free(trie->wordinfo);
17112 /* do this last!!!! */
17113 PerlMemShared_free(ri->data->data[n]);
17118 Perl_croak(aTHX_ "panic: regfree data code '%c'",
17119 ri->data->what[n]);
17122 Safefree(ri->data->what);
17123 Safefree(ri->data);
17129 #define av_dup_inc(s,t) MUTABLE_AV(sv_dup_inc((const SV *)s,t))
17130 #define hv_dup_inc(s,t) MUTABLE_HV(sv_dup_inc((const SV *)s,t))
17131 #define SAVEPVN(p,n) ((p) ? savepvn(p,n) : NULL)
17134 re_dup - duplicate a regexp.
17136 This routine is expected to clone a given regexp structure. It is only
17137 compiled under USE_ITHREADS.
17139 After all of the core data stored in struct regexp is duplicated
17140 the regexp_engine.dupe method is used to copy any private data
17141 stored in the *pprivate pointer. This allows extensions to handle
17142 any duplication it needs to do.
17144 See pregfree() and regfree_internal() if you change anything here.
17146 #if defined(USE_ITHREADS)
17147 #ifndef PERL_IN_XSUB_RE
17149 Perl_re_dup_guts(pTHX_ const REGEXP *sstr, REGEXP *dstr, CLONE_PARAMS *param)
17153 const struct regexp *r = ReANY(sstr);
17154 struct regexp *ret = ReANY(dstr);
17156 PERL_ARGS_ASSERT_RE_DUP_GUTS;
17158 npar = r->nparens+1;
17159 Newx(ret->offs, npar, regexp_paren_pair);
17160 Copy(r->offs, ret->offs, npar, regexp_paren_pair);
17162 if (ret->substrs) {
17163 /* Do it this way to avoid reading from *r after the StructCopy().
17164 That way, if any of the sv_dup_inc()s dislodge *r from the L1
17165 cache, it doesn't matter. */
17166 const bool anchored = r->check_substr
17167 ? r->check_substr == r->anchored_substr
17168 : r->check_utf8 == r->anchored_utf8;
17169 Newx(ret->substrs, 1, struct reg_substr_data);
17170 StructCopy(r->substrs, ret->substrs, struct reg_substr_data);
17172 ret->anchored_substr = sv_dup_inc(ret->anchored_substr, param);
17173 ret->anchored_utf8 = sv_dup_inc(ret->anchored_utf8, param);
17174 ret->float_substr = sv_dup_inc(ret->float_substr, param);
17175 ret->float_utf8 = sv_dup_inc(ret->float_utf8, param);
17177 /* check_substr and check_utf8, if non-NULL, point to either their
17178 anchored or float namesakes, and don't hold a second reference. */
17180 if (ret->check_substr) {
17182 assert(r->check_utf8 == r->anchored_utf8);
17183 ret->check_substr = ret->anchored_substr;
17184 ret->check_utf8 = ret->anchored_utf8;
17186 assert(r->check_substr == r->float_substr);
17187 assert(r->check_utf8 == r->float_utf8);
17188 ret->check_substr = ret->float_substr;
17189 ret->check_utf8 = ret->float_utf8;
17191 } else if (ret->check_utf8) {
17193 ret->check_utf8 = ret->anchored_utf8;
17195 ret->check_utf8 = ret->float_utf8;
17200 RXp_PAREN_NAMES(ret) = hv_dup_inc(RXp_PAREN_NAMES(ret), param);
17201 ret->qr_anoncv = MUTABLE_CV(sv_dup_inc((const SV *)ret->qr_anoncv, param));
17204 RXi_SET(ret,CALLREGDUPE_PVT(dstr,param));
17206 if (RX_MATCH_COPIED(dstr))
17207 ret->subbeg = SAVEPVN(ret->subbeg, ret->sublen);
17209 ret->subbeg = NULL;
17210 #ifdef PERL_ANY_COW
17211 ret->saved_copy = NULL;
17214 /* Whether mother_re be set or no, we need to copy the string. We
17215 cannot refrain from copying it when the storage points directly to
17216 our mother regexp, because that's
17217 1: a buffer in a different thread
17218 2: something we no longer hold a reference on
17219 so we need to copy it locally. */
17220 RX_WRAPPED(dstr) = SAVEPVN(RX_WRAPPED(sstr), SvCUR(sstr)+1);
17221 ret->mother_re = NULL;
17223 #endif /* PERL_IN_XSUB_RE */
17228 This is the internal complement to regdupe() which is used to copy
17229 the structure pointed to by the *pprivate pointer in the regexp.
17230 This is the core version of the extension overridable cloning hook.
17231 The regexp structure being duplicated will be copied by perl prior
17232 to this and will be provided as the regexp *r argument, however
17233 with the /old/ structures pprivate pointer value. Thus this routine
17234 may override any copying normally done by perl.
17236 It returns a pointer to the new regexp_internal structure.
17240 Perl_regdupe_internal(pTHX_ REGEXP * const rx, CLONE_PARAMS *param)
17243 struct regexp *const r = ReANY(rx);
17244 regexp_internal *reti;
17246 RXi_GET_DECL(r,ri);
17248 PERL_ARGS_ASSERT_REGDUPE_INTERNAL;
17252 Newxc(reti, sizeof(regexp_internal) + len*sizeof(regnode),
17253 char, regexp_internal);
17254 Copy(ri->program, reti->program, len+1, regnode);
17256 reti->num_code_blocks = ri->num_code_blocks;
17257 if (ri->code_blocks) {
17259 Newxc(reti->code_blocks, ri->num_code_blocks, struct reg_code_block,
17260 struct reg_code_block);
17261 Copy(ri->code_blocks, reti->code_blocks, ri->num_code_blocks,
17262 struct reg_code_block);
17263 for (n = 0; n < ri->num_code_blocks; n++)
17264 reti->code_blocks[n].src_regex = (REGEXP*)
17265 sv_dup_inc((SV*)(ri->code_blocks[n].src_regex), param);
17268 reti->code_blocks = NULL;
17270 reti->regstclass = NULL;
17273 struct reg_data *d;
17274 const int count = ri->data->count;
17277 Newxc(d, sizeof(struct reg_data) + count*sizeof(void *),
17278 char, struct reg_data);
17279 Newx(d->what, count, U8);
17282 for (i = 0; i < count; i++) {
17283 d->what[i] = ri->data->what[i];
17284 switch (d->what[i]) {
17285 /* see also regcomp.h and regfree_internal() */
17286 case 'a': /* actually an AV, but the dup function is identical. */
17290 case 'u': /* actually an HV, but the dup function is identical. */
17291 d->data[i] = sv_dup_inc((const SV *)ri->data->data[i], param);
17294 /* This is cheating. */
17295 Newx(d->data[i], 1, regnode_ssc);
17296 StructCopy(ri->data->data[i], d->data[i], regnode_ssc);
17297 reti->regstclass = (regnode*)d->data[i];
17300 /* Trie stclasses are readonly and can thus be shared
17301 * without duplication. We free the stclass in pregfree
17302 * when the corresponding reg_ac_data struct is freed.
17304 reti->regstclass= ri->regstclass;
17308 ((reg_trie_data*)ri->data->data[i])->refcount++;
17313 d->data[i] = ri->data->data[i];
17316 Perl_croak(aTHX_ "panic: re_dup unknown data code '%c'",
17317 ri->data->what[i]);
17326 reti->name_list_idx = ri->name_list_idx;
17328 #ifdef RE_TRACK_PATTERN_OFFSETS
17329 if (ri->u.offsets) {
17330 Newx(reti->u.offsets, 2*len+1, U32);
17331 Copy(ri->u.offsets, reti->u.offsets, 2*len+1, U32);
17334 SetProgLen(reti,len);
17337 return (void*)reti;
17340 #endif /* USE_ITHREADS */
17342 #ifndef PERL_IN_XSUB_RE
17345 - regnext - dig the "next" pointer out of a node
17348 Perl_regnext(pTHX_ regnode *p)
17355 if (OP(p) > REGNODE_MAX) { /* regnode.type is unsigned */
17356 Perl_croak(aTHX_ "Corrupted regexp opcode %d > %d",
17357 (int)OP(p), (int)REGNODE_MAX);
17360 offset = (reg_off_by_arg[OP(p)] ? ARG(p) : NEXT_OFF(p));
17369 S_re_croak2(pTHX_ bool utf8, const char* pat1,const char* pat2,...)
17372 STRLEN l1 = strlen(pat1);
17373 STRLEN l2 = strlen(pat2);
17376 const char *message;
17378 PERL_ARGS_ASSERT_RE_CROAK2;
17384 Copy(pat1, buf, l1 , char);
17385 Copy(pat2, buf + l1, l2 , char);
17386 buf[l1 + l2] = '\n';
17387 buf[l1 + l2 + 1] = '\0';
17388 va_start(args, pat2);
17389 msv = vmess(buf, &args);
17391 message = SvPV_const(msv,l1);
17394 Copy(message, buf, l1 , char);
17395 /* l1-1 to avoid \n */
17396 Perl_croak(aTHX_ "%"UTF8f, UTF8fARG(utf8, l1-1, buf));
17402 S_put_code_point(pTHX_ SV *sv, UV c)
17404 PERL_ARGS_ASSERT_PUT_CODE_POINT;
17407 Perl_sv_catpvf(aTHX_ sv, "\\x{%04"UVXf"}", c);
17409 else if (isPRINT(c)) {
17410 const char string = (char) c;
17411 if (isBACKSLASHED_PUNCT(c))
17412 sv_catpvs(sv, "\\");
17413 sv_catpvn(sv, &string, 1);
17416 const char * const mnemonic = cntrl_to_mnemonic((char) c);
17418 Perl_sv_catpvf(aTHX_ sv, "%s", mnemonic);
17421 Perl_sv_catpvf(aTHX_ sv, "\\x{%02X}", (U8) c);
17426 #define MAX_PRINT_A MAX_PRINT_A_FOR_USE_ONLY_BY_REGCOMP_DOT_C
17429 S_put_range(pTHX_ SV *sv, UV start, const UV end, const bool allow_literals)
17431 /* Appends to 'sv' a displayable version of the range of code points from
17432 * 'start' to 'end'. It assumes that only ASCII printables are displayable
17433 * as-is (though some of these will be escaped by put_code_point()). */
17435 const unsigned int min_range_count = 3;
17437 assert(start <= end);
17439 PERL_ARGS_ASSERT_PUT_RANGE;
17441 while (start <= end) {
17443 const char * format;
17445 if (end - start < min_range_count) {
17447 /* Individual chars in short ranges */
17448 for (; start <= end; start++) {
17449 put_code_point(sv, start);
17454 /* If permitted by the input options, and there is a possibility that
17455 * this range contains a printable literal, look to see if there is
17457 if (allow_literals && start <= MAX_PRINT_A) {
17459 /* If the range begin isn't an ASCII printable, effectively split
17460 * the range into two parts:
17461 * 1) the portion before the first such printable,
17463 * and output them separately. */
17464 if (! isPRINT_A(start)) {
17465 UV temp_end = start + 1;
17467 /* There is no point looking beyond the final possible
17468 * printable, in MAX_PRINT_A */
17469 UV max = MIN(end, MAX_PRINT_A);
17471 while (temp_end <= max && ! isPRINT_A(temp_end)) {
17475 /* Here, temp_end points to one beyond the first printable if
17476 * found, or to one beyond 'max' if not. If none found, make
17477 * sure that we use the entire range */
17478 if (temp_end > MAX_PRINT_A) {
17479 temp_end = end + 1;
17482 /* Output the first part of the split range, the part that
17483 * doesn't have printables, with no looking for literals
17484 * (otherwise we would infinitely recurse) */
17485 put_range(sv, start, temp_end - 1, FALSE);
17487 /* The 2nd part of the range (if any) starts here. */
17490 /* We continue instead of dropping down because even if the 2nd
17491 * part is non-empty, it could be so short that we want to
17492 * output it specially, as tested for at the top of this loop.
17497 /* Here, 'start' is a printable ASCII. If it is an alphanumeric,
17498 * output a sub-range of just the digits or letters, then process
17499 * the remaining portion as usual. */
17500 if (isALPHANUMERIC_A(start)) {
17501 UV mask = (isDIGIT_A(start))
17506 UV temp_end = start + 1;
17508 /* Find the end of the sub-range that includes just the
17509 * characters in the same class as the first character in it */
17510 while (temp_end <= end && _generic_isCC_A(temp_end, mask)) {
17515 /* For short ranges, don't duplicate the code above to output
17516 * them; just call recursively */
17517 if (temp_end - start < min_range_count) {
17518 put_range(sv, start, temp_end, FALSE);
17520 else { /* Output as a range */
17521 put_code_point(sv, start);
17522 sv_catpvs(sv, "-");
17523 put_code_point(sv, temp_end);
17525 start = temp_end + 1;
17529 /* We output any other printables as individual characters */
17530 if (isPUNCT_A(start) || isSPACE_A(start)) {
17531 while (start <= end && (isPUNCT_A(start)
17532 || isSPACE_A(start)))
17534 put_code_point(sv, start);
17539 } /* End of looking for literals */
17541 /* Here is not to output as a literal. Some control characters have
17542 * mnemonic names. Split off any of those at the beginning and end of
17543 * the range to print mnemonically. It isn't possible for many of
17544 * these to be in a row, so this won't overwhelm with output */
17545 while (isMNEMONIC_CNTRL(start) && start <= end) {
17546 put_code_point(sv, start);
17549 if (start < end && isMNEMONIC_CNTRL(end)) {
17551 /* Here, the final character in the range has a mnemonic name.
17552 * Work backwards from the end to find the final non-mnemonic */
17553 UV temp_end = end - 1;
17554 while (isMNEMONIC_CNTRL(temp_end)) {
17558 /* And separately output the range that doesn't have mnemonics */
17559 put_range(sv, start, temp_end, FALSE);
17561 /* Then output the mnemonic trailing controls */
17562 start = temp_end + 1;
17563 while (start <= end) {
17564 put_code_point(sv, start);
17570 /* As a final resort, output the range or subrange as hex. */
17572 this_end = (end < NUM_ANYOF_CODE_POINTS)
17574 : NUM_ANYOF_CODE_POINTS - 1;
17575 format = (this_end < 256)
17576 ? "\\x{%02"UVXf"}-\\x{%02"UVXf"}"
17577 : "\\x{%04"UVXf"}-\\x{%04"UVXf"}";
17578 GCC_DIAG_IGNORE(-Wformat-nonliteral);
17579 Perl_sv_catpvf(aTHX_ sv, format, start, this_end);
17586 S_put_charclass_bitmap_innards(pTHX_ SV *sv, char *bitmap, SV** bitmap_invlist)
17588 /* Appends to 'sv' a displayable version of the innards of the bracketed
17589 * character class whose bitmap is 'bitmap'; Returns 'TRUE' if it actually
17590 * output anything, and bitmap_invlist, if not NULL, will point to an
17591 * inversion list of what is in the bit map */
17595 unsigned int punct_count = 0;
17596 SV* invlist = NULL;
17597 SV** invlist_ptr; /* Temporary, in case bitmap_invlist is NULL */
17598 bool allow_literals = TRUE;
17600 PERL_ARGS_ASSERT_PUT_CHARCLASS_BITMAP_INNARDS;
17602 invlist_ptr = (bitmap_invlist) ? bitmap_invlist : &invlist;
17604 /* Worst case is exactly every-other code point is in the list */
17605 *invlist_ptr = _new_invlist(NUM_ANYOF_CODE_POINTS / 2);
17607 /* Convert the bit map to an inversion list, keeping track of how many
17608 * ASCII puncts are set, including an extra amount for the backslashed
17610 for (i = 0; i < NUM_ANYOF_CODE_POINTS; i++) {
17611 if (BITMAP_TEST(bitmap, i)) {
17612 *invlist_ptr = add_cp_to_invlist(*invlist_ptr, i);
17613 if (isPUNCT_A(i)) {
17615 if isBACKSLASHED_PUNCT(i) {
17622 /* Nothing to output */
17623 if (_invlist_len(*invlist_ptr) == 0) {
17624 SvREFCNT_dec(invlist);
17628 /* Generally, it is more readable if printable characters are output as
17629 * literals, but if a range (nearly) spans all of them, it's best to output
17630 * it as a single range. This code will use a single range if all but 2
17631 * printables are in it */
17632 invlist_iterinit(*invlist_ptr);
17633 while (invlist_iternext(*invlist_ptr, &start, &end)) {
17635 /* If range starts beyond final printable, it doesn't have any in it */
17636 if (start > MAX_PRINT_A) {
17640 /* In both ASCII and EBCDIC, a SPACE is the lowest printable. To span
17641 * all but two, the range must start and end no later than 2 from
17643 if (start < ' ' + 2 && end > MAX_PRINT_A - 2) {
17644 if (end > MAX_PRINT_A) {
17650 if (end - start >= MAX_PRINT_A - ' ' - 2) {
17651 allow_literals = FALSE;
17656 invlist_iterfinish(*invlist_ptr);
17658 /* The legibility of the output depends mostly on how many punctuation
17659 * characters are output. There are 32 possible ASCII ones, and some have
17660 * an additional backslash, bringing it to currently 36, so if any more
17661 * than 18 are to be output, we can instead output it as its complement,
17662 * yielding fewer puncts, and making it more legible. But give some weight
17663 * to the fact that outputting it as a complement is less legible than a
17664 * straight output, so don't complement unless we are somewhat over the 18
17666 if (allow_literals && punct_count > 22) {
17667 sv_catpvs(sv, "^");
17669 /* Add everything remaining to the list, so when we invert it just
17670 * below, it will be excluded */
17671 _invlist_union_complement_2nd(*invlist_ptr, PL_InBitmap, invlist_ptr);
17672 _invlist_invert(*invlist_ptr);
17675 /* Here we have figured things out. Output each range */
17676 invlist_iterinit(*invlist_ptr);
17677 while (invlist_iternext(*invlist_ptr, &start, &end)) {
17678 if (start >= NUM_ANYOF_CODE_POINTS) {
17681 put_range(sv, start, end, allow_literals);
17683 invlist_iterfinish(*invlist_ptr);
17688 #define CLEAR_OPTSTART \
17689 if (optstart) STMT_START { \
17690 DEBUG_OPTIMISE_r(PerlIO_printf(Perl_debug_log, \
17691 " (%"IVdf" nodes)\n", (IV)(node - optstart))); \
17695 #define DUMPUNTIL(b,e) \
17697 node=dumpuntil(r,start,(b),(e),last,sv,indent+1,depth+1);
17699 STATIC const regnode *
17700 S_dumpuntil(pTHX_ const regexp *r, const regnode *start, const regnode *node,
17701 const regnode *last, const regnode *plast,
17702 SV* sv, I32 indent, U32 depth)
17704 U8 op = PSEUDO; /* Arbitrary non-END op. */
17705 const regnode *next;
17706 const regnode *optstart= NULL;
17708 RXi_GET_DECL(r,ri);
17709 GET_RE_DEBUG_FLAGS_DECL;
17711 PERL_ARGS_ASSERT_DUMPUNTIL;
17713 #ifdef DEBUG_DUMPUNTIL
17714 PerlIO_printf(Perl_debug_log, "--- %d : %d - %d - %d\n",indent,node-start,
17715 last ? last-start : 0,plast ? plast-start : 0);
17718 if (plast && plast < last)
17721 while (PL_regkind[op] != END && (!last || node < last)) {
17723 /* While that wasn't END last time... */
17726 if (op == CLOSE || op == WHILEM)
17728 next = regnext((regnode *)node);
17731 if (OP(node) == OPTIMIZED) {
17732 if (!optstart && RE_DEBUG_FLAG(RE_DEBUG_COMPILE_OPTIMISE))
17739 regprop(r, sv, node, NULL, NULL);
17740 PerlIO_printf(Perl_debug_log, "%4"IVdf":%*s%s", (IV)(node - start),
17741 (int)(2*indent + 1), "", SvPVX_const(sv));
17743 if (OP(node) != OPTIMIZED) {
17744 if (next == NULL) /* Next ptr. */
17745 PerlIO_printf(Perl_debug_log, " (0)");
17746 else if (PL_regkind[(U8)op] == BRANCH
17747 && PL_regkind[OP(next)] != BRANCH )
17748 PerlIO_printf(Perl_debug_log, " (FAIL)");
17750 PerlIO_printf(Perl_debug_log, " (%"IVdf")", (IV)(next - start));
17751 (void)PerlIO_putc(Perl_debug_log, '\n');
17755 if (PL_regkind[(U8)op] == BRANCHJ) {
17758 const regnode *nnode = (OP(next) == LONGJMP
17759 ? regnext((regnode *)next)
17761 if (last && nnode > last)
17763 DUMPUNTIL(NEXTOPER(NEXTOPER(node)), nnode);
17766 else if (PL_regkind[(U8)op] == BRANCH) {
17768 DUMPUNTIL(NEXTOPER(node), next);
17770 else if ( PL_regkind[(U8)op] == TRIE ) {
17771 const regnode *this_trie = node;
17772 const char op = OP(node);
17773 const U32 n = ARG(node);
17774 const reg_ac_data * const ac = op>=AHOCORASICK ?
17775 (reg_ac_data *)ri->data->data[n] :
17777 const reg_trie_data * const trie =
17778 (reg_trie_data*)ri->data->data[op<AHOCORASICK ? n : ac->trie];
17780 AV *const trie_words
17781 = MUTABLE_AV(ri->data->data[n + TRIE_WORDS_OFFSET]);
17783 const regnode *nextbranch= NULL;
17786 for (word_idx= 0; word_idx < (I32)trie->wordcount; word_idx++) {
17787 SV ** const elem_ptr = av_fetch(trie_words,word_idx,0);
17789 PerlIO_printf(Perl_debug_log, "%*s%s ",
17790 (int)(2*(indent+3)), "",
17792 ? pv_pretty(sv, SvPV_nolen_const(*elem_ptr),
17793 SvCUR(*elem_ptr), 60,
17794 PL_colors[0], PL_colors[1],
17796 ? PERL_PV_ESCAPE_UNI
17798 | PERL_PV_PRETTY_ELLIPSES
17799 | PERL_PV_PRETTY_LTGT
17804 U16 dist= trie->jump[word_idx+1];
17805 PerlIO_printf(Perl_debug_log, "(%"UVuf")\n",
17806 (UV)((dist ? this_trie + dist : next) - start));
17809 nextbranch= this_trie + trie->jump[0];
17810 DUMPUNTIL(this_trie + dist, nextbranch);
17812 if (nextbranch && PL_regkind[OP(nextbranch)]==BRANCH)
17813 nextbranch= regnext((regnode *)nextbranch);
17815 PerlIO_printf(Perl_debug_log, "\n");
17818 if (last && next > last)
17823 else if ( op == CURLY ) { /* "next" might be very big: optimizer */
17824 DUMPUNTIL(NEXTOPER(node) + EXTRA_STEP_2ARGS,
17825 NEXTOPER(node) + EXTRA_STEP_2ARGS + 1);
17827 else if (PL_regkind[(U8)op] == CURLY && op != CURLYX) {
17829 DUMPUNTIL(NEXTOPER(node) + EXTRA_STEP_2ARGS, next);
17831 else if ( op == PLUS || op == STAR) {
17832 DUMPUNTIL(NEXTOPER(node), NEXTOPER(node) + 1);
17834 else if (PL_regkind[(U8)op] == ANYOF) {
17835 /* arglen 1 + class block */
17836 node += 1 + ((ANYOF_FLAGS(node) & ANYOF_MATCHES_POSIXL)
17837 ? ANYOF_POSIXL_SKIP
17839 node = NEXTOPER(node);
17841 else if (PL_regkind[(U8)op] == EXACT) {
17842 /* Literal string, where present. */
17843 node += NODE_SZ_STR(node) - 1;
17844 node = NEXTOPER(node);
17847 node = NEXTOPER(node);
17848 node += regarglen[(U8)op];
17850 if (op == CURLYX || op == OPEN)
17854 #ifdef DEBUG_DUMPUNTIL
17855 PerlIO_printf(Perl_debug_log, "--- %d\n", (int)indent);
17860 #endif /* DEBUGGING */
17864 * c-indentation-style: bsd
17865 * c-basic-offset: 4
17866 * indent-tabs-mode: nil
17869 * ex: set ts=8 sts=4 sw=4 et: