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
88 #include "dquote_static.c"
95 # if defined(BUGGY_MSC6)
96 /* MSC 6.00A breaks on op/regexp.t test 85 unless we turn this off */
97 # pragma optimize("a",off)
98 /* But MSC 6.00A is happy with 'w', for aliases only across function calls*/
99 # pragma optimize("w",on )
100 # endif /* BUGGY_MSC6 */
104 #define STATIC static
107 typedef struct RExC_state_t {
108 U32 flags; /* are we folding, multilining? */
109 char *precomp; /* uncompiled string. */
110 REGEXP *rx_sv; /* The SV that is the regexp. */
111 regexp *rx; /* perl core regexp structure */
112 regexp_internal *rxi; /* internal data for regexp object pprivate field */
113 char *start; /* Start of input for compile */
114 char *end; /* End of input for compile */
115 char *parse; /* Input-scan pointer. */
116 I32 whilem_seen; /* number of WHILEM in this expr */
117 regnode *emit_start; /* Start of emitted-code area */
118 regnode *emit_bound; /* First regnode outside of the allocated space */
119 regnode *emit; /* Code-emit pointer; ®dummy = don't = compiling */
120 I32 naughty; /* How bad is this pattern? */
121 I32 sawback; /* Did we see \1, ...? */
123 I32 size; /* Code size. */
124 I32 npar; /* Capture buffer count, (OPEN). */
125 I32 cpar; /* Capture buffer count, (CLOSE). */
126 I32 nestroot; /* root parens we are in - used by accept */
130 regnode **open_parens; /* pointers to open parens */
131 regnode **close_parens; /* pointers to close parens */
132 regnode *opend; /* END node in program */
133 I32 utf8; /* whether the pattern is utf8 or not */
134 I32 orig_utf8; /* whether the pattern was originally in utf8 */
135 /* XXX use this for future optimisation of case
136 * where pattern must be upgraded to utf8. */
137 HV *paren_names; /* Paren names */
139 regnode **recurse; /* Recurse regops */
140 I32 recurse_count; /* Number of recurse regops */
143 char *starttry; /* -Dr: where regtry was called. */
144 #define RExC_starttry (pRExC_state->starttry)
147 const char *lastparse;
149 AV *paren_name_list; /* idx -> name */
150 #define RExC_lastparse (pRExC_state->lastparse)
151 #define RExC_lastnum (pRExC_state->lastnum)
152 #define RExC_paren_name_list (pRExC_state->paren_name_list)
156 #define RExC_flags (pRExC_state->flags)
157 #define RExC_precomp (pRExC_state->precomp)
158 #define RExC_rx_sv (pRExC_state->rx_sv)
159 #define RExC_rx (pRExC_state->rx)
160 #define RExC_rxi (pRExC_state->rxi)
161 #define RExC_start (pRExC_state->start)
162 #define RExC_end (pRExC_state->end)
163 #define RExC_parse (pRExC_state->parse)
164 #define RExC_whilem_seen (pRExC_state->whilem_seen)
165 #ifdef RE_TRACK_PATTERN_OFFSETS
166 #define RExC_offsets (pRExC_state->rxi->u.offsets) /* I am not like the others */
168 #define RExC_emit (pRExC_state->emit)
169 #define RExC_emit_start (pRExC_state->emit_start)
170 #define RExC_emit_bound (pRExC_state->emit_bound)
171 #define RExC_naughty (pRExC_state->naughty)
172 #define RExC_sawback (pRExC_state->sawback)
173 #define RExC_seen (pRExC_state->seen)
174 #define RExC_size (pRExC_state->size)
175 #define RExC_npar (pRExC_state->npar)
176 #define RExC_nestroot (pRExC_state->nestroot)
177 #define RExC_extralen (pRExC_state->extralen)
178 #define RExC_seen_zerolen (pRExC_state->seen_zerolen)
179 #define RExC_seen_evals (pRExC_state->seen_evals)
180 #define RExC_utf8 (pRExC_state->utf8)
181 #define RExC_orig_utf8 (pRExC_state->orig_utf8)
182 #define RExC_open_parens (pRExC_state->open_parens)
183 #define RExC_close_parens (pRExC_state->close_parens)
184 #define RExC_opend (pRExC_state->opend)
185 #define RExC_paren_names (pRExC_state->paren_names)
186 #define RExC_recurse (pRExC_state->recurse)
187 #define RExC_recurse_count (pRExC_state->recurse_count)
188 #define RExC_in_lookbehind (pRExC_state->in_lookbehind)
191 #define ISMULT1(c) ((c) == '*' || (c) == '+' || (c) == '?')
192 #define ISMULT2(s) ((*s) == '*' || (*s) == '+' || (*s) == '?' || \
193 ((*s) == '{' && regcurly(s)))
196 #undef SPSTART /* dratted cpp namespace... */
199 * Flags to be passed up and down.
201 #define WORST 0 /* Worst case. */
202 #define HASWIDTH 0x01 /* Known to match non-null strings. */
204 /* Simple enough to be STAR/PLUS operand, in an EXACT node must be a single
205 * character, and if utf8, must be invariant. Note that this is not the same thing as REGNODE_SIMPLE */
207 #define SPSTART 0x04 /* Starts with * or +. */
208 #define TRYAGAIN 0x08 /* Weeded out a declaration. */
209 #define POSTPONED 0x10 /* (?1),(?&name), (??{...}) or similar */
211 #define REG_NODE_NUM(x) ((x) ? (int)((x)-RExC_emit_start) : -1)
213 /* whether trie related optimizations are enabled */
214 #if PERL_ENABLE_EXTENDED_TRIE_OPTIMISATION
215 #define TRIE_STUDY_OPT
216 #define FULL_TRIE_STUDY
222 #define PBYTE(u8str,paren) ((U8*)(u8str))[(paren) >> 3]
223 #define PBITVAL(paren) (1 << ((paren) & 7))
224 #define PAREN_TEST(u8str,paren) ( PBYTE(u8str,paren) & PBITVAL(paren))
225 #define PAREN_SET(u8str,paren) PBYTE(u8str,paren) |= PBITVAL(paren)
226 #define PAREN_UNSET(u8str,paren) PBYTE(u8str,paren) &= (~PBITVAL(paren))
228 /* If not already in utf8, do a longjmp back to the beginning */
229 #define UTF8_LONGJMP 42 /* Choose a value not likely to ever conflict */
230 #define REQUIRE_UTF8 STMT_START { \
231 if (! UTF) JMPENV_JUMP(UTF8_LONGJMP); \
234 /* About scan_data_t.
236 During optimisation we recurse through the regexp program performing
237 various inplace (keyhole style) optimisations. In addition study_chunk
238 and scan_commit populate this data structure with information about
239 what strings MUST appear in the pattern. We look for the longest
240 string that must appear at a fixed location, and we look for the
241 longest string that may appear at a floating location. So for instance
246 Both 'FOO' and 'A' are fixed strings. Both 'B' and 'BAR' are floating
247 strings (because they follow a .* construct). study_chunk will identify
248 both FOO and BAR as being the longest fixed and floating strings respectively.
250 The strings can be composites, for instance
254 will result in a composite fixed substring 'foo'.
256 For each string some basic information is maintained:
258 - offset or min_offset
259 This is the position the string must appear at, or not before.
260 It also implicitly (when combined with minlenp) tells us how many
261 characters must match before the string we are searching for.
262 Likewise when combined with minlenp and the length of the string it
263 tells us how many characters must appear after the string we have
267 Only used for floating strings. This is the rightmost point that
268 the string can appear at. If set to I32 max it indicates that the
269 string can occur infinitely far to the right.
272 A pointer to the minimum length of the pattern that the string
273 was found inside. This is important as in the case of positive
274 lookahead or positive lookbehind we can have multiple patterns
279 The minimum length of the pattern overall is 3, the minimum length
280 of the lookahead part is 3, but the minimum length of the part that
281 will actually match is 1. So 'FOO's minimum length is 3, but the
282 minimum length for the F is 1. This is important as the minimum length
283 is used to determine offsets in front of and behind the string being
284 looked for. Since strings can be composites this is the length of the
285 pattern at the time it was committed with a scan_commit. Note that
286 the length is calculated by study_chunk, so that the minimum lengths
287 are not known until the full pattern has been compiled, thus the
288 pointer to the value.
292 In the case of lookbehind the string being searched for can be
293 offset past the start point of the final matching string.
294 If this value was just blithely removed from the min_offset it would
295 invalidate some of the calculations for how many chars must match
296 before or after (as they are derived from min_offset and minlen and
297 the length of the string being searched for).
298 When the final pattern is compiled and the data is moved from the
299 scan_data_t structure into the regexp structure the information
300 about lookbehind is factored in, with the information that would
301 have been lost precalculated in the end_shift field for the
304 The fields pos_min and pos_delta are used to store the minimum offset
305 and the delta to the maximum offset at the current point in the pattern.
309 typedef struct scan_data_t {
310 /*I32 len_min; unused */
311 /*I32 len_delta; unused */
315 I32 last_end; /* min value, <0 unless valid. */
318 SV **longest; /* Either &l_fixed, or &l_float. */
319 SV *longest_fixed; /* longest fixed string found in pattern */
320 I32 offset_fixed; /* offset where it starts */
321 I32 *minlen_fixed; /* pointer to the minlen relevant to the string */
322 I32 lookbehind_fixed; /* is the position of the string modfied by LB */
323 SV *longest_float; /* longest floating string found in pattern */
324 I32 offset_float_min; /* earliest point in string it can appear */
325 I32 offset_float_max; /* latest point in string it can appear */
326 I32 *minlen_float; /* pointer to the minlen relevant to the string */
327 I32 lookbehind_float; /* is the position of the string modified by LB */
331 struct regnode_charclass_class *start_class;
335 * Forward declarations for pregcomp()'s friends.
338 static const scan_data_t zero_scan_data =
339 { 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0 ,0};
341 #define SF_BEFORE_EOL (SF_BEFORE_SEOL|SF_BEFORE_MEOL)
342 #define SF_BEFORE_SEOL 0x0001
343 #define SF_BEFORE_MEOL 0x0002
344 #define SF_FIX_BEFORE_EOL (SF_FIX_BEFORE_SEOL|SF_FIX_BEFORE_MEOL)
345 #define SF_FL_BEFORE_EOL (SF_FL_BEFORE_SEOL|SF_FL_BEFORE_MEOL)
348 # define SF_FIX_SHIFT_EOL (0+2)
349 # define SF_FL_SHIFT_EOL (0+4)
351 # define SF_FIX_SHIFT_EOL (+2)
352 # define SF_FL_SHIFT_EOL (+4)
355 #define SF_FIX_BEFORE_SEOL (SF_BEFORE_SEOL << SF_FIX_SHIFT_EOL)
356 #define SF_FIX_BEFORE_MEOL (SF_BEFORE_MEOL << SF_FIX_SHIFT_EOL)
358 #define SF_FL_BEFORE_SEOL (SF_BEFORE_SEOL << SF_FL_SHIFT_EOL)
359 #define SF_FL_BEFORE_MEOL (SF_BEFORE_MEOL << SF_FL_SHIFT_EOL) /* 0x20 */
360 #define SF_IS_INF 0x0040
361 #define SF_HAS_PAR 0x0080
362 #define SF_IN_PAR 0x0100
363 #define SF_HAS_EVAL 0x0200
364 #define SCF_DO_SUBSTR 0x0400
365 #define SCF_DO_STCLASS_AND 0x0800
366 #define SCF_DO_STCLASS_OR 0x1000
367 #define SCF_DO_STCLASS (SCF_DO_STCLASS_AND|SCF_DO_STCLASS_OR)
368 #define SCF_WHILEM_VISITED_POS 0x2000
370 #define SCF_TRIE_RESTUDY 0x4000 /* Do restudy? */
371 #define SCF_SEEN_ACCEPT 0x8000
373 #define UTF cBOOL(RExC_utf8)
374 #define LOC (get_regex_charset(RExC_flags) == REGEX_LOCALE_CHARSET)
375 #define UNI_SEMANTICS (get_regex_charset(RExC_flags) == REGEX_UNICODE_CHARSET)
376 #define DEPENDS_SEMANTICS (get_regex_charset(RExC_flags) == REGEX_DEPENDS_CHARSET)
377 #define AT_LEAST_UNI_SEMANTICS (get_regex_charset(RExC_flags) >= REGEX_UNICODE_CHARSET)
378 #define ASCII_RESTRICTED (get_regex_charset(RExC_flags) == REGEX_ASCII_RESTRICTED_CHARSET)
380 #define FOLD cBOOL(RExC_flags & RXf_PMf_FOLD)
382 #define OOB_UNICODE 12345678
383 #define OOB_NAMEDCLASS -1
385 #define CHR_SVLEN(sv) (UTF ? sv_len_utf8(sv) : SvCUR(sv))
386 #define CHR_DIST(a,b) (UTF ? utf8_distance(a,b) : a - b)
389 /* length of regex to show in messages that don't mark a position within */
390 #define RegexLengthToShowInErrorMessages 127
393 * If MARKER[12] are adjusted, be sure to adjust the constants at the top
394 * of t/op/regmesg.t, the tests in t/op/re_tests, and those in
395 * op/pragma/warn/regcomp.
397 #define MARKER1 "<-- HERE" /* marker as it appears in the description */
398 #define MARKER2 " <-- HERE " /* marker as it appears within the regex */
400 #define REPORT_LOCATION " in regex; marked by " MARKER1 " in m/%.*s" MARKER2 "%s/"
403 * Calls SAVEDESTRUCTOR_X if needed, then calls Perl_croak with the given
404 * arg. Show regex, up to a maximum length. If it's too long, chop and add
407 #define _FAIL(code) STMT_START { \
408 const char *ellipses = ""; \
409 IV len = RExC_end - RExC_precomp; \
412 SAVEDESTRUCTOR_X(clear_re,(void*)RExC_rx_sv); \
413 if (len > RegexLengthToShowInErrorMessages) { \
414 /* chop 10 shorter than the max, to ensure meaning of "..." */ \
415 len = RegexLengthToShowInErrorMessages - 10; \
421 #define FAIL(msg) _FAIL( \
422 Perl_croak(aTHX_ "%s in regex m/%.*s%s/", \
423 msg, (int)len, RExC_precomp, ellipses))
425 #define FAIL2(msg,arg) _FAIL( \
426 Perl_croak(aTHX_ msg " in regex m/%.*s%s/", \
427 arg, (int)len, RExC_precomp, ellipses))
430 * Simple_vFAIL -- like FAIL, but marks the current location in the scan
432 #define Simple_vFAIL(m) STMT_START { \
433 const IV offset = RExC_parse - RExC_precomp; \
434 Perl_croak(aTHX_ "%s" REPORT_LOCATION, \
435 m, (int)offset, RExC_precomp, RExC_precomp + offset); \
439 * Calls SAVEDESTRUCTOR_X if needed, then Simple_vFAIL()
441 #define vFAIL(m) STMT_START { \
443 SAVEDESTRUCTOR_X(clear_re,(void*)RExC_rx_sv); \
448 * Like Simple_vFAIL(), but accepts two arguments.
450 #define Simple_vFAIL2(m,a1) STMT_START { \
451 const IV offset = RExC_parse - RExC_precomp; \
452 S_re_croak2(aTHX_ m, REPORT_LOCATION, a1, \
453 (int)offset, RExC_precomp, RExC_precomp + offset); \
457 * Calls SAVEDESTRUCTOR_X if needed, then Simple_vFAIL2().
459 #define vFAIL2(m,a1) STMT_START { \
461 SAVEDESTRUCTOR_X(clear_re,(void*)RExC_rx_sv); \
462 Simple_vFAIL2(m, a1); \
467 * Like Simple_vFAIL(), but accepts three arguments.
469 #define Simple_vFAIL3(m, a1, a2) STMT_START { \
470 const IV offset = RExC_parse - RExC_precomp; \
471 S_re_croak2(aTHX_ m, REPORT_LOCATION, a1, a2, \
472 (int)offset, RExC_precomp, RExC_precomp + offset); \
476 * Calls SAVEDESTRUCTOR_X if needed, then Simple_vFAIL3().
478 #define vFAIL3(m,a1,a2) STMT_START { \
480 SAVEDESTRUCTOR_X(clear_re,(void*)RExC_rx_sv); \
481 Simple_vFAIL3(m, a1, a2); \
485 * Like Simple_vFAIL(), but accepts four arguments.
487 #define Simple_vFAIL4(m, a1, a2, a3) STMT_START { \
488 const IV offset = RExC_parse - RExC_precomp; \
489 S_re_croak2(aTHX_ m, REPORT_LOCATION, a1, a2, a3, \
490 (int)offset, RExC_precomp, RExC_precomp + offset); \
493 #define ckWARNreg(loc,m) STMT_START { \
494 const IV offset = loc - RExC_precomp; \
495 Perl_ck_warner(aTHX_ packWARN(WARN_REGEXP), m REPORT_LOCATION, \
496 (int)offset, RExC_precomp, RExC_precomp + offset); \
499 #define ckWARNregdep(loc,m) STMT_START { \
500 const IV offset = loc - RExC_precomp; \
501 Perl_ck_warner_d(aTHX_ packWARN2(WARN_DEPRECATED, WARN_REGEXP), \
503 (int)offset, RExC_precomp, RExC_precomp + offset); \
506 #define ckWARN2reg(loc, m, a1) STMT_START { \
507 const IV offset = loc - RExC_precomp; \
508 Perl_ck_warner(aTHX_ packWARN(WARN_REGEXP), m REPORT_LOCATION, \
509 a1, (int)offset, RExC_precomp, RExC_precomp + offset); \
512 #define vWARN3(loc, m, a1, a2) STMT_START { \
513 const IV offset = loc - RExC_precomp; \
514 Perl_warner(aTHX_ packWARN(WARN_REGEXP), m REPORT_LOCATION, \
515 a1, a2, (int)offset, RExC_precomp, RExC_precomp + offset); \
518 #define ckWARN3reg(loc, m, a1, a2) STMT_START { \
519 const IV offset = loc - RExC_precomp; \
520 Perl_ck_warner(aTHX_ packWARN(WARN_REGEXP), m REPORT_LOCATION, \
521 a1, a2, (int)offset, RExC_precomp, RExC_precomp + offset); \
524 #define vWARN4(loc, m, a1, a2, a3) STMT_START { \
525 const IV offset = loc - RExC_precomp; \
526 Perl_warner(aTHX_ packWARN(WARN_REGEXP), m REPORT_LOCATION, \
527 a1, a2, a3, (int)offset, RExC_precomp, RExC_precomp + offset); \
530 #define ckWARN4reg(loc, m, a1, a2, a3) STMT_START { \
531 const IV offset = loc - RExC_precomp; \
532 Perl_ck_warner(aTHX_ packWARN(WARN_REGEXP), m REPORT_LOCATION, \
533 a1, a2, a3, (int)offset, RExC_precomp, RExC_precomp + offset); \
536 #define vWARN5(loc, m, a1, a2, a3, a4) STMT_START { \
537 const IV offset = loc - RExC_precomp; \
538 Perl_warner(aTHX_ packWARN(WARN_REGEXP), m REPORT_LOCATION, \
539 a1, a2, a3, a4, (int)offset, RExC_precomp, RExC_precomp + offset); \
543 /* Allow for side effects in s */
544 #define REGC(c,s) STMT_START { \
545 if (!SIZE_ONLY) *(s) = (c); else (void)(s); \
548 /* Macros for recording node offsets. 20001227 mjd@plover.com
549 * Nodes are numbered 1, 2, 3, 4. Node #n's position is recorded in
550 * element 2*n-1 of the array. Element #2n holds the byte length node #n.
551 * Element 0 holds the number n.
552 * Position is 1 indexed.
554 #ifndef RE_TRACK_PATTERN_OFFSETS
555 #define Set_Node_Offset_To_R(node,byte)
556 #define Set_Node_Offset(node,byte)
557 #define Set_Cur_Node_Offset
558 #define Set_Node_Length_To_R(node,len)
559 #define Set_Node_Length(node,len)
560 #define Set_Node_Cur_Length(node)
561 #define Node_Offset(n)
562 #define Node_Length(n)
563 #define Set_Node_Offset_Length(node,offset,len)
564 #define ProgLen(ri) ri->u.proglen
565 #define SetProgLen(ri,x) ri->u.proglen = x
567 #define ProgLen(ri) ri->u.offsets[0]
568 #define SetProgLen(ri,x) ri->u.offsets[0] = x
569 #define Set_Node_Offset_To_R(node,byte) STMT_START { \
571 MJD_OFFSET_DEBUG(("** (%d) offset of node %d is %d.\n", \
572 __LINE__, (int)(node), (int)(byte))); \
574 Perl_croak(aTHX_ "value of node is %d in Offset macro", (int)(node)); \
576 RExC_offsets[2*(node)-1] = (byte); \
581 #define Set_Node_Offset(node,byte) \
582 Set_Node_Offset_To_R((node)-RExC_emit_start, (byte)-RExC_start)
583 #define Set_Cur_Node_Offset Set_Node_Offset(RExC_emit, RExC_parse)
585 #define Set_Node_Length_To_R(node,len) STMT_START { \
587 MJD_OFFSET_DEBUG(("** (%d) size of node %d is %d.\n", \
588 __LINE__, (int)(node), (int)(len))); \
590 Perl_croak(aTHX_ "value of node is %d in Length macro", (int)(node)); \
592 RExC_offsets[2*(node)] = (len); \
597 #define Set_Node_Length(node,len) \
598 Set_Node_Length_To_R((node)-RExC_emit_start, len)
599 #define Set_Cur_Node_Length(len) Set_Node_Length(RExC_emit, len)
600 #define Set_Node_Cur_Length(node) \
601 Set_Node_Length(node, RExC_parse - parse_start)
603 /* Get offsets and lengths */
604 #define Node_Offset(n) (RExC_offsets[2*((n)-RExC_emit_start)-1])
605 #define Node_Length(n) (RExC_offsets[2*((n)-RExC_emit_start)])
607 #define Set_Node_Offset_Length(node,offset,len) STMT_START { \
608 Set_Node_Offset_To_R((node)-RExC_emit_start, (offset)); \
609 Set_Node_Length_To_R((node)-RExC_emit_start, (len)); \
613 #if PERL_ENABLE_EXPERIMENTAL_REGEX_OPTIMISATIONS
614 #define EXPERIMENTAL_INPLACESCAN
615 #endif /*PERL_ENABLE_EXPERIMENTAL_REGEX_OPTIMISATIONS*/
617 #define DEBUG_STUDYDATA(str,data,depth) \
618 DEBUG_OPTIMISE_MORE_r(if(data){ \
619 PerlIO_printf(Perl_debug_log, \
620 "%*s" str "Pos:%"IVdf"/%"IVdf \
621 " Flags: 0x%"UVXf" Whilem_c: %"IVdf" Lcp: %"IVdf" %s", \
622 (int)(depth)*2, "", \
623 (IV)((data)->pos_min), \
624 (IV)((data)->pos_delta), \
625 (UV)((data)->flags), \
626 (IV)((data)->whilem_c), \
627 (IV)((data)->last_closep ? *((data)->last_closep) : -1), \
628 is_inf ? "INF " : "" \
630 if ((data)->last_found) \
631 PerlIO_printf(Perl_debug_log, \
632 "Last:'%s' %"IVdf":%"IVdf"/%"IVdf" %sFixed:'%s' @ %"IVdf \
633 " %sFloat: '%s' @ %"IVdf"/%"IVdf"", \
634 SvPVX_const((data)->last_found), \
635 (IV)((data)->last_end), \
636 (IV)((data)->last_start_min), \
637 (IV)((data)->last_start_max), \
638 ((data)->longest && \
639 (data)->longest==&((data)->longest_fixed)) ? "*" : "", \
640 SvPVX_const((data)->longest_fixed), \
641 (IV)((data)->offset_fixed), \
642 ((data)->longest && \
643 (data)->longest==&((data)->longest_float)) ? "*" : "", \
644 SvPVX_const((data)->longest_float), \
645 (IV)((data)->offset_float_min), \
646 (IV)((data)->offset_float_max) \
648 PerlIO_printf(Perl_debug_log,"\n"); \
651 static void clear_re(pTHX_ void *r);
653 /* Mark that we cannot extend a found fixed substring at this point.
654 Update the longest found anchored substring and the longest found
655 floating substrings if needed. */
658 S_scan_commit(pTHX_ const RExC_state_t *pRExC_state, scan_data_t *data, I32 *minlenp, int is_inf)
660 const STRLEN l = CHR_SVLEN(data->last_found);
661 const STRLEN old_l = CHR_SVLEN(*data->longest);
662 GET_RE_DEBUG_FLAGS_DECL;
664 PERL_ARGS_ASSERT_SCAN_COMMIT;
666 if ((l >= old_l) && ((l > old_l) || (data->flags & SF_BEFORE_EOL))) {
667 SvSetMagicSV(*data->longest, data->last_found);
668 if (*data->longest == data->longest_fixed) {
669 data->offset_fixed = l ? data->last_start_min : data->pos_min;
670 if (data->flags & SF_BEFORE_EOL)
672 |= ((data->flags & SF_BEFORE_EOL) << SF_FIX_SHIFT_EOL);
674 data->flags &= ~SF_FIX_BEFORE_EOL;
675 data->minlen_fixed=minlenp;
676 data->lookbehind_fixed=0;
678 else { /* *data->longest == data->longest_float */
679 data->offset_float_min = l ? data->last_start_min : data->pos_min;
680 data->offset_float_max = (l
681 ? data->last_start_max
682 : data->pos_min + data->pos_delta);
683 if (is_inf || (U32)data->offset_float_max > (U32)I32_MAX)
684 data->offset_float_max = I32_MAX;
685 if (data->flags & SF_BEFORE_EOL)
687 |= ((data->flags & SF_BEFORE_EOL) << SF_FL_SHIFT_EOL);
689 data->flags &= ~SF_FL_BEFORE_EOL;
690 data->minlen_float=minlenp;
691 data->lookbehind_float=0;
694 SvCUR_set(data->last_found, 0);
696 SV * const sv = data->last_found;
697 if (SvUTF8(sv) && SvMAGICAL(sv)) {
698 MAGIC * const mg = mg_find(sv, PERL_MAGIC_utf8);
704 data->flags &= ~SF_BEFORE_EOL;
705 DEBUG_STUDYDATA("commit: ",data,0);
708 /* Can match anything (initialization) */
710 S_cl_anything(const RExC_state_t *pRExC_state, struct regnode_charclass_class *cl)
712 PERL_ARGS_ASSERT_CL_ANYTHING;
714 ANYOF_CLASS_ZERO(cl);
715 ANYOF_BITMAP_SETALL(cl);
716 cl->flags = ANYOF_EOS|ANYOF_UNICODE_ALL|ANYOF_LOC_NONBITMAP_FOLD|ANYOF_NON_UTF8_LATIN1_ALL;
718 cl->flags |= ANYOF_LOCALE;
721 /* Can match anything (initialization) */
723 S_cl_is_anything(const struct regnode_charclass_class *cl)
727 PERL_ARGS_ASSERT_CL_IS_ANYTHING;
729 for (value = 0; value <= ANYOF_MAX; value += 2)
730 if (ANYOF_CLASS_TEST(cl, value) && ANYOF_CLASS_TEST(cl, value + 1))
732 if (!(cl->flags & ANYOF_UNICODE_ALL))
734 if (!ANYOF_BITMAP_TESTALLSET((const void*)cl))
739 /* Can match anything (initialization) */
741 S_cl_init(const RExC_state_t *pRExC_state, struct regnode_charclass_class *cl)
743 PERL_ARGS_ASSERT_CL_INIT;
745 Zero(cl, 1, struct regnode_charclass_class);
747 cl_anything(pRExC_state, cl);
751 S_cl_init_zero(const RExC_state_t *pRExC_state, struct regnode_charclass_class *cl)
753 PERL_ARGS_ASSERT_CL_INIT_ZERO;
755 Zero(cl, 1, struct regnode_charclass_class);
757 cl_anything(pRExC_state, cl);
759 cl->flags |= ANYOF_LOCALE;
762 /* 'And' a given class with another one. Can create false positives */
763 /* We assume that cl is not inverted */
765 S_cl_and(struct regnode_charclass_class *cl,
766 const struct regnode_charclass_class *and_with)
768 PERL_ARGS_ASSERT_CL_AND;
770 assert(and_with->type == ANYOF);
772 if (!(ANYOF_CLASS_TEST_ANY_SET(and_with))
773 && !(ANYOF_CLASS_TEST_ANY_SET(cl))
774 && (and_with->flags & ANYOF_LOCALE) == (cl->flags & ANYOF_LOCALE)
775 && !(and_with->flags & ANYOF_LOC_NONBITMAP_FOLD)
776 && !(cl->flags & ANYOF_LOC_NONBITMAP_FOLD)) {
779 if (and_with->flags & ANYOF_INVERT)
780 for (i = 0; i < ANYOF_BITMAP_SIZE; i++)
781 cl->bitmap[i] &= ~and_with->bitmap[i];
783 for (i = 0; i < ANYOF_BITMAP_SIZE; i++)
784 cl->bitmap[i] &= and_with->bitmap[i];
785 } /* XXXX: logic is complicated otherwise, leave it along for a moment. */
786 if (!(and_with->flags & ANYOF_EOS))
787 cl->flags &= ~ANYOF_EOS;
789 if (!(and_with->flags & ANYOF_LOC_NONBITMAP_FOLD))
790 cl->flags &= ~ANYOF_LOC_NONBITMAP_FOLD;
791 if (!(and_with->flags & ANYOF_NON_UTF8_LATIN1_ALL))
792 cl->flags &= ~ANYOF_NON_UTF8_LATIN1_ALL;
794 if (cl->flags & ANYOF_UNICODE_ALL && and_with->flags & ANYOF_NONBITMAP &&
795 !(and_with->flags & ANYOF_INVERT)) {
796 cl->flags &= ~ANYOF_UNICODE_ALL;
797 cl->flags |= and_with->flags & ANYOF_NONBITMAP; /* field is 2 bits; use
800 ARG_SET(cl, ARG(and_with));
802 if (!(and_with->flags & ANYOF_UNICODE_ALL) &&
803 !(and_with->flags & ANYOF_INVERT))
804 cl->flags &= ~ANYOF_UNICODE_ALL;
805 if (!(and_with->flags & (ANYOF_NONBITMAP|ANYOF_UNICODE_ALL)) &&
806 !(and_with->flags & ANYOF_INVERT))
807 cl->flags &= ~ANYOF_NONBITMAP;
810 /* 'OR' a given class with another one. Can create false positives */
811 /* We assume that cl is not inverted */
813 S_cl_or(const RExC_state_t *pRExC_state, struct regnode_charclass_class *cl, const struct regnode_charclass_class *or_with)
815 PERL_ARGS_ASSERT_CL_OR;
817 if (or_with->flags & ANYOF_INVERT) {
819 * (B1 | CL1) | (!B2 & !CL2) = (B1 | !B2 & !CL2) | (CL1 | (!B2 & !CL2))
820 * <= (B1 | !B2) | (CL1 | !CL2)
821 * which is wasteful if CL2 is small, but we ignore CL2:
822 * (B1 | CL1) | (!B2 & !CL2) <= (B1 | CL1) | !B2 = (B1 | !B2) | CL1
823 * XXXX Can we handle case-fold? Unclear:
824 * (OK1(i) | OK1(i')) | !(OK1(i) | OK1(i')) =
825 * (OK1(i) | OK1(i')) | (!OK1(i) & !OK1(i'))
827 if ( (or_with->flags & ANYOF_LOCALE) == (cl->flags & ANYOF_LOCALE)
828 && !(or_with->flags & ANYOF_LOC_NONBITMAP_FOLD)
829 && !(cl->flags & ANYOF_LOC_NONBITMAP_FOLD) ) {
832 for (i = 0; i < ANYOF_BITMAP_SIZE; i++)
833 cl->bitmap[i] |= ~or_with->bitmap[i];
834 } /* XXXX: logic is complicated otherwise */
836 cl_anything(pRExC_state, cl);
839 /* (B1 | CL1) | (B2 | CL2) = (B1 | B2) | (CL1 | CL2)) */
840 if ( (or_with->flags & ANYOF_LOCALE) == (cl->flags & ANYOF_LOCALE)
841 && (!(or_with->flags & ANYOF_LOC_NONBITMAP_FOLD)
842 || (cl->flags & ANYOF_LOC_NONBITMAP_FOLD)) ) {
845 /* OR char bitmap and class bitmap separately */
846 for (i = 0; i < ANYOF_BITMAP_SIZE; i++)
847 cl->bitmap[i] |= or_with->bitmap[i];
848 if (ANYOF_CLASS_TEST_ANY_SET(or_with)) {
849 for (i = 0; i < ANYOF_CLASSBITMAP_SIZE; i++)
850 cl->classflags[i] |= or_with->classflags[i];
851 cl->flags |= ANYOF_CLASS;
854 else { /* XXXX: logic is complicated, leave it along for a moment. */
855 cl_anything(pRExC_state, cl);
858 if (or_with->flags & ANYOF_EOS)
859 cl->flags |= ANYOF_EOS;
860 if (!(or_with->flags & ANYOF_NON_UTF8_LATIN1_ALL))
861 cl->flags |= ANYOF_NON_UTF8_LATIN1_ALL;
863 if (or_with->flags & ANYOF_LOC_NONBITMAP_FOLD)
864 cl->flags |= ANYOF_LOC_NONBITMAP_FOLD;
866 /* If both nodes match something outside the bitmap, but what they match
867 * outside is not the same pointer, and hence not easily compared, give up
868 * and allow the start class to match everything outside the bitmap */
869 if (cl->flags & ANYOF_NONBITMAP && or_with->flags & ANYOF_NONBITMAP &&
870 ARG(cl) != ARG(or_with)) {
871 cl->flags |= ANYOF_UNICODE_ALL;
874 if (or_with->flags & ANYOF_UNICODE_ALL) {
875 cl->flags |= ANYOF_UNICODE_ALL;
879 #define TRIE_LIST_ITEM(state,idx) (trie->states[state].trans.list)[ idx ]
880 #define TRIE_LIST_CUR(state) ( TRIE_LIST_ITEM( state, 0 ).forid )
881 #define TRIE_LIST_LEN(state) ( TRIE_LIST_ITEM( state, 0 ).newstate )
882 #define TRIE_LIST_USED(idx) ( trie->states[state].trans.list ? (TRIE_LIST_CUR( idx ) - 1) : 0 )
887 dump_trie(trie,widecharmap,revcharmap)
888 dump_trie_interim_list(trie,widecharmap,revcharmap,next_alloc)
889 dump_trie_interim_table(trie,widecharmap,revcharmap,next_alloc)
891 These routines dump out a trie in a somewhat readable format.
892 The _interim_ variants are used for debugging the interim
893 tables that are used to generate the final compressed
894 representation which is what dump_trie expects.
896 Part of the reason for their existence is to provide a form
897 of documentation as to how the different representations function.
902 Dumps the final compressed table form of the trie to Perl_debug_log.
903 Used for debugging make_trie().
907 S_dump_trie(pTHX_ const struct _reg_trie_data *trie, HV *widecharmap,
908 AV *revcharmap, U32 depth)
911 SV *sv=sv_newmortal();
912 int colwidth= widecharmap ? 6 : 4;
914 GET_RE_DEBUG_FLAGS_DECL;
916 PERL_ARGS_ASSERT_DUMP_TRIE;
918 PerlIO_printf( Perl_debug_log, "%*sChar : %-6s%-6s%-4s ",
919 (int)depth * 2 + 2,"",
920 "Match","Base","Ofs" );
922 for( state = 0 ; state < trie->uniquecharcount ; state++ ) {
923 SV ** const tmp = av_fetch( revcharmap, state, 0);
925 PerlIO_printf( Perl_debug_log, "%*s",
927 pv_pretty(sv, SvPV_nolen_const(*tmp), SvCUR(*tmp), colwidth,
928 PL_colors[0], PL_colors[1],
929 (SvUTF8(*tmp) ? PERL_PV_ESCAPE_UNI : 0) |
930 PERL_PV_ESCAPE_FIRSTCHAR
935 PerlIO_printf( Perl_debug_log, "\n%*sState|-----------------------",
936 (int)depth * 2 + 2,"");
938 for( state = 0 ; state < trie->uniquecharcount ; state++ )
939 PerlIO_printf( Perl_debug_log, "%.*s", colwidth, "--------");
940 PerlIO_printf( Perl_debug_log, "\n");
942 for( state = 1 ; state < trie->statecount ; state++ ) {
943 const U32 base = trie->states[ state ].trans.base;
945 PerlIO_printf( Perl_debug_log, "%*s#%4"UVXf"|", (int)depth * 2 + 2,"", (UV)state);
947 if ( trie->states[ state ].wordnum ) {
948 PerlIO_printf( Perl_debug_log, " W%4X", trie->states[ state ].wordnum );
950 PerlIO_printf( Perl_debug_log, "%6s", "" );
953 PerlIO_printf( Perl_debug_log, " @%4"UVXf" ", (UV)base );
958 while( ( base + ofs < trie->uniquecharcount ) ||
959 ( base + ofs - trie->uniquecharcount < trie->lasttrans
960 && trie->trans[ base + ofs - trie->uniquecharcount ].check != state))
963 PerlIO_printf( Perl_debug_log, "+%2"UVXf"[ ", (UV)ofs);
965 for ( ofs = 0 ; ofs < trie->uniquecharcount ; ofs++ ) {
966 if ( ( base + ofs >= trie->uniquecharcount ) &&
967 ( base + ofs - trie->uniquecharcount < trie->lasttrans ) &&
968 trie->trans[ base + ofs - trie->uniquecharcount ].check == state )
970 PerlIO_printf( Perl_debug_log, "%*"UVXf,
972 (UV)trie->trans[ base + ofs - trie->uniquecharcount ].next );
974 PerlIO_printf( Perl_debug_log, "%*s",colwidth," ." );
978 PerlIO_printf( Perl_debug_log, "]");
981 PerlIO_printf( Perl_debug_log, "\n" );
983 PerlIO_printf(Perl_debug_log, "%*sword_info N:(prev,len)=", (int)depth*2, "");
984 for (word=1; word <= trie->wordcount; word++) {
985 PerlIO_printf(Perl_debug_log, " %d:(%d,%d)",
986 (int)word, (int)(trie->wordinfo[word].prev),
987 (int)(trie->wordinfo[word].len));
989 PerlIO_printf(Perl_debug_log, "\n" );
992 Dumps a fully constructed but uncompressed trie in list form.
993 List tries normally only are used for construction when the number of
994 possible chars (trie->uniquecharcount) is very high.
995 Used for debugging make_trie().
998 S_dump_trie_interim_list(pTHX_ const struct _reg_trie_data *trie,
999 HV *widecharmap, AV *revcharmap, U32 next_alloc,
1003 SV *sv=sv_newmortal();
1004 int colwidth= widecharmap ? 6 : 4;
1005 GET_RE_DEBUG_FLAGS_DECL;
1007 PERL_ARGS_ASSERT_DUMP_TRIE_INTERIM_LIST;
1009 /* print out the table precompression. */
1010 PerlIO_printf( Perl_debug_log, "%*sState :Word | Transition Data\n%*s%s",
1011 (int)depth * 2 + 2,"", (int)depth * 2 + 2,"",
1012 "------:-----+-----------------\n" );
1014 for( state=1 ; state < next_alloc ; state ++ ) {
1017 PerlIO_printf( Perl_debug_log, "%*s %4"UVXf" :",
1018 (int)depth * 2 + 2,"", (UV)state );
1019 if ( ! trie->states[ state ].wordnum ) {
1020 PerlIO_printf( Perl_debug_log, "%5s| ","");
1022 PerlIO_printf( Perl_debug_log, "W%4x| ",
1023 trie->states[ state ].wordnum
1026 for( charid = 1 ; charid <= TRIE_LIST_USED( state ) ; charid++ ) {
1027 SV ** const tmp = av_fetch( revcharmap, TRIE_LIST_ITEM(state,charid).forid, 0);
1029 PerlIO_printf( Perl_debug_log, "%*s:%3X=%4"UVXf" | ",
1031 pv_pretty(sv, SvPV_nolen_const(*tmp), SvCUR(*tmp), colwidth,
1032 PL_colors[0], PL_colors[1],
1033 (SvUTF8(*tmp) ? PERL_PV_ESCAPE_UNI : 0) |
1034 PERL_PV_ESCAPE_FIRSTCHAR
1036 TRIE_LIST_ITEM(state,charid).forid,
1037 (UV)TRIE_LIST_ITEM(state,charid).newstate
1040 PerlIO_printf(Perl_debug_log, "\n%*s| ",
1041 (int)((depth * 2) + 14), "");
1044 PerlIO_printf( Perl_debug_log, "\n");
1049 Dumps a fully constructed but uncompressed trie in table form.
1050 This is the normal DFA style state transition table, with a few
1051 twists to facilitate compression later.
1052 Used for debugging make_trie().
1055 S_dump_trie_interim_table(pTHX_ const struct _reg_trie_data *trie,
1056 HV *widecharmap, AV *revcharmap, U32 next_alloc,
1061 SV *sv=sv_newmortal();
1062 int colwidth= widecharmap ? 6 : 4;
1063 GET_RE_DEBUG_FLAGS_DECL;
1065 PERL_ARGS_ASSERT_DUMP_TRIE_INTERIM_TABLE;
1068 print out the table precompression so that we can do a visual check
1069 that they are identical.
1072 PerlIO_printf( Perl_debug_log, "%*sChar : ",(int)depth * 2 + 2,"" );
1074 for( charid = 0 ; charid < trie->uniquecharcount ; charid++ ) {
1075 SV ** const tmp = av_fetch( revcharmap, charid, 0);
1077 PerlIO_printf( Perl_debug_log, "%*s",
1079 pv_pretty(sv, SvPV_nolen_const(*tmp), SvCUR(*tmp), colwidth,
1080 PL_colors[0], PL_colors[1],
1081 (SvUTF8(*tmp) ? PERL_PV_ESCAPE_UNI : 0) |
1082 PERL_PV_ESCAPE_FIRSTCHAR
1088 PerlIO_printf( Perl_debug_log, "\n%*sState+-",(int)depth * 2 + 2,"" );
1090 for( charid=0 ; charid < trie->uniquecharcount ; charid++ ) {
1091 PerlIO_printf( Perl_debug_log, "%.*s", colwidth,"--------");
1094 PerlIO_printf( Perl_debug_log, "\n" );
1096 for( state=1 ; state < next_alloc ; state += trie->uniquecharcount ) {
1098 PerlIO_printf( Perl_debug_log, "%*s%4"UVXf" : ",
1099 (int)depth * 2 + 2,"",
1100 (UV)TRIE_NODENUM( state ) );
1102 for( charid = 0 ; charid < trie->uniquecharcount ; charid++ ) {
1103 UV v=(UV)SAFE_TRIE_NODENUM( trie->trans[ state + charid ].next );
1105 PerlIO_printf( Perl_debug_log, "%*"UVXf, colwidth, v );
1107 PerlIO_printf( Perl_debug_log, "%*s", colwidth, "." );
1109 if ( ! trie->states[ TRIE_NODENUM( state ) ].wordnum ) {
1110 PerlIO_printf( Perl_debug_log, " (%4"UVXf")\n", (UV)trie->trans[ state ].check );
1112 PerlIO_printf( Perl_debug_log, " (%4"UVXf") W%4X\n", (UV)trie->trans[ state ].check,
1113 trie->states[ TRIE_NODENUM( state ) ].wordnum );
1121 /* make_trie(startbranch,first,last,tail,word_count,flags,depth)
1122 startbranch: the first branch in the whole branch sequence
1123 first : start branch of sequence of branch-exact nodes.
1124 May be the same as startbranch
1125 last : Thing following the last branch.
1126 May be the same as tail.
1127 tail : item following the branch sequence
1128 count : words in the sequence
1129 flags : currently the OP() type we will be building one of /EXACT(|F|Fl)/
1130 depth : indent depth
1132 Inplace optimizes a sequence of 2 or more Branch-Exact nodes into a TRIE node.
1134 A trie is an N'ary tree where the branches are determined by digital
1135 decomposition of the key. IE, at the root node you look up the 1st character and
1136 follow that branch repeat until you find the end of the branches. Nodes can be
1137 marked as "accepting" meaning they represent a complete word. Eg:
1141 would convert into the following structure. Numbers represent states, letters
1142 following numbers represent valid transitions on the letter from that state, if
1143 the number is in square brackets it represents an accepting state, otherwise it
1144 will be in parenthesis.
1146 +-h->+-e->[3]-+-r->(8)-+-s->[9]
1150 (1) +-i->(6)-+-s->[7]
1152 +-s->(3)-+-h->(4)-+-e->[5]
1154 Accept Word Mapping: 3=>1 (he),5=>2 (she), 7=>3 (his), 9=>4 (hers)
1156 This shows that when matching against the string 'hers' we will begin at state 1
1157 read 'h' and move to state 2, read 'e' and move to state 3 which is accepting,
1158 then read 'r' and go to state 8 followed by 's' which takes us to state 9 which
1159 is also accepting. Thus we know that we can match both 'he' and 'hers' with a
1160 single traverse. We store a mapping from accepting to state to which word was
1161 matched, and then when we have multiple possibilities we try to complete the
1162 rest of the regex in the order in which they occured in the alternation.
1164 The only prior NFA like behaviour that would be changed by the TRIE support is
1165 the silent ignoring of duplicate alternations which are of the form:
1167 / (DUPE|DUPE) X? (?{ ... }) Y /x
1169 Thus EVAL blocks following a trie may be called a different number of times with
1170 and without the optimisation. With the optimisations dupes will be silently
1171 ignored. This inconsistent behaviour of EVAL type nodes is well established as
1172 the following demonstrates:
1174 'words'=~/(word|word|word)(?{ print $1 })[xyz]/
1176 which prints out 'word' three times, but
1178 'words'=~/(word|word|word)(?{ print $1 })S/
1180 which doesnt print it out at all. This is due to other optimisations kicking in.
1182 Example of what happens on a structural level:
1184 The regexp /(ac|ad|ab)+/ will produce the following debug output:
1186 1: CURLYM[1] {1,32767}(18)
1197 This would be optimizable with startbranch=5, first=5, last=16, tail=16
1198 and should turn into:
1200 1: CURLYM[1] {1,32767}(18)
1202 [Words:3 Chars Stored:6 Unique Chars:4 States:5 NCP:1]
1210 Cases where tail != last would be like /(?foo|bar)baz/:
1220 which would be optimizable with startbranch=1, first=1, last=7, tail=8
1221 and would end up looking like:
1224 [Words:2 Chars Stored:6 Unique Chars:5 States:7 NCP:1]
1231 d = uvuni_to_utf8_flags(d, uv, 0);
1233 is the recommended Unicode-aware way of saying
1238 #define TRIE_STORE_REVCHAR \
1241 SV *zlopp = newSV(2); \
1242 unsigned char *flrbbbbb = (unsigned char *) SvPVX(zlopp); \
1243 unsigned const char *const kapow = uvuni_to_utf8(flrbbbbb, uvc & 0xFF); \
1244 SvCUR_set(zlopp, kapow - flrbbbbb); \
1247 av_push(revcharmap, zlopp); \
1249 char ooooff = (char)uvc; \
1250 av_push(revcharmap, newSVpvn(&ooooff, 1)); \
1254 #define TRIE_READ_CHAR STMT_START { \
1258 if ( foldlen > 0 ) { \
1259 uvc = utf8n_to_uvuni( scan, UTF8_MAXLEN, &len, uniflags ); \
1264 uvc = utf8n_to_uvuni( (const U8*)uc, UTF8_MAXLEN, &len, uniflags);\
1265 uvc = to_uni_fold( uvc, foldbuf, &foldlen ); \
1266 foldlen -= UNISKIP( uvc ); \
1267 scan = foldbuf + UNISKIP( uvc ); \
1270 uvc = utf8n_to_uvuni( (const U8*)uc, UTF8_MAXLEN, &len, uniflags);\
1280 #define TRIE_LIST_PUSH(state,fid,ns) STMT_START { \
1281 if ( TRIE_LIST_CUR( state ) >=TRIE_LIST_LEN( state ) ) { \
1282 U32 ging = TRIE_LIST_LEN( state ) *= 2; \
1283 Renew( trie->states[ state ].trans.list, ging, reg_trie_trans_le ); \
1285 TRIE_LIST_ITEM( state, TRIE_LIST_CUR( state ) ).forid = fid; \
1286 TRIE_LIST_ITEM( state, TRIE_LIST_CUR( state ) ).newstate = ns; \
1287 TRIE_LIST_CUR( state )++; \
1290 #define TRIE_LIST_NEW(state) STMT_START { \
1291 Newxz( trie->states[ state ].trans.list, \
1292 4, reg_trie_trans_le ); \
1293 TRIE_LIST_CUR( state ) = 1; \
1294 TRIE_LIST_LEN( state ) = 4; \
1297 #define TRIE_HANDLE_WORD(state) STMT_START { \
1298 U16 dupe= trie->states[ state ].wordnum; \
1299 regnode * const noper_next = regnext( noper ); \
1302 /* store the word for dumping */ \
1304 if (OP(noper) != NOTHING) \
1305 tmp = newSVpvn_utf8(STRING(noper), STR_LEN(noper), UTF); \
1307 tmp = newSVpvn_utf8( "", 0, UTF ); \
1308 av_push( trie_words, tmp ); \
1312 trie->wordinfo[curword].prev = 0; \
1313 trie->wordinfo[curword].len = wordlen; \
1314 trie->wordinfo[curword].accept = state; \
1316 if ( noper_next < tail ) { \
1318 trie->jump = (U16 *) PerlMemShared_calloc( word_count + 1, sizeof(U16) ); \
1319 trie->jump[curword] = (U16)(noper_next - convert); \
1321 jumper = noper_next; \
1323 nextbranch= regnext(cur); \
1327 /* It's a dupe. Pre-insert into the wordinfo[].prev */\
1328 /* chain, so that when the bits of chain are later */\
1329 /* linked together, the dups appear in the chain */\
1330 trie->wordinfo[curword].prev = trie->wordinfo[dupe].prev; \
1331 trie->wordinfo[dupe].prev = curword; \
1333 /* we haven't inserted this word yet. */ \
1334 trie->states[ state ].wordnum = curword; \
1339 #define TRIE_TRANS_STATE(state,base,ucharcount,charid,special) \
1340 ( ( base + charid >= ucharcount \
1341 && base + charid < ubound \
1342 && state == trie->trans[ base - ucharcount + charid ].check \
1343 && trie->trans[ base - ucharcount + charid ].next ) \
1344 ? trie->trans[ base - ucharcount + charid ].next \
1345 : ( state==1 ? special : 0 ) \
1349 #define MADE_JUMP_TRIE 2
1350 #define MADE_EXACT_TRIE 4
1353 S_make_trie(pTHX_ RExC_state_t *pRExC_state, regnode *startbranch, regnode *first, regnode *last, regnode *tail, U32 word_count, U32 flags, U32 depth)
1356 /* first pass, loop through and scan words */
1357 reg_trie_data *trie;
1358 HV *widecharmap = NULL;
1359 AV *revcharmap = newAV();
1361 const U32 uniflags = UTF8_ALLOW_DEFAULT;
1366 regnode *jumper = NULL;
1367 regnode *nextbranch = NULL;
1368 regnode *convert = NULL;
1369 U32 *prev_states; /* temp array mapping each state to previous one */
1370 /* we just use folder as a flag in utf8 */
1371 const U8 * folder = NULL;
1374 const U32 data_slot = add_data( pRExC_state, 4, "tuuu" );
1375 AV *trie_words = NULL;
1376 /* along with revcharmap, this only used during construction but both are
1377 * useful during debugging so we store them in the struct when debugging.
1380 const U32 data_slot = add_data( pRExC_state, 2, "tu" );
1381 STRLEN trie_charcount=0;
1383 SV *re_trie_maxbuff;
1384 GET_RE_DEBUG_FLAGS_DECL;
1386 PERL_ARGS_ASSERT_MAKE_TRIE;
1388 PERL_UNUSED_ARG(depth);
1392 case EXACTFU: folder = PL_fold_latin1; break;
1393 case EXACTF: folder = PL_fold; break;
1394 case EXACTFL: folder = PL_fold_locale; break;
1397 trie = (reg_trie_data *) PerlMemShared_calloc( 1, sizeof(reg_trie_data) );
1399 trie->startstate = 1;
1400 trie->wordcount = word_count;
1401 RExC_rxi->data->data[ data_slot ] = (void*)trie;
1402 trie->charmap = (U16 *) PerlMemShared_calloc( 256, sizeof(U16) );
1403 if (!(UTF && folder))
1404 trie->bitmap = (char *) PerlMemShared_calloc( ANYOF_BITMAP_SIZE, 1 );
1405 trie->wordinfo = (reg_trie_wordinfo *) PerlMemShared_calloc(
1406 trie->wordcount+1, sizeof(reg_trie_wordinfo));
1409 trie_words = newAV();
1412 re_trie_maxbuff = get_sv(RE_TRIE_MAXBUF_NAME, 1);
1413 if (!SvIOK(re_trie_maxbuff)) {
1414 sv_setiv(re_trie_maxbuff, RE_TRIE_MAXBUF_INIT);
1417 PerlIO_printf( Perl_debug_log,
1418 "%*smake_trie start==%d, first==%d, last==%d, tail==%d depth=%d\n",
1419 (int)depth * 2 + 2, "",
1420 REG_NODE_NUM(startbranch),REG_NODE_NUM(first),
1421 REG_NODE_NUM(last), REG_NODE_NUM(tail),
1425 /* Find the node we are going to overwrite */
1426 if ( first == startbranch && OP( last ) != BRANCH ) {
1427 /* whole branch chain */
1430 /* branch sub-chain */
1431 convert = NEXTOPER( first );
1434 /* -- First loop and Setup --
1436 We first traverse the branches and scan each word to determine if it
1437 contains widechars, and how many unique chars there are, this is
1438 important as we have to build a table with at least as many columns as we
1441 We use an array of integers to represent the character codes 0..255
1442 (trie->charmap) and we use a an HV* to store Unicode characters. We use the
1443 native representation of the character value as the key and IV's for the
1446 *TODO* If we keep track of how many times each character is used we can
1447 remap the columns so that the table compression later on is more
1448 efficient in terms of memory by ensuring the most common value is in the
1449 middle and the least common are on the outside. IMO this would be better
1450 than a most to least common mapping as theres a decent chance the most
1451 common letter will share a node with the least common, meaning the node
1452 will not be compressible. With a middle is most common approach the worst
1453 case is when we have the least common nodes twice.
1457 for ( cur = first ; cur < last ; cur = regnext( cur ) ) {
1458 regnode * const noper = NEXTOPER( cur );
1459 const U8 *uc = (U8*)STRING( noper );
1460 const U8 * const e = uc + STR_LEN( noper );
1462 U8 foldbuf[ UTF8_MAXBYTES_CASE + 1 ];
1463 const U8 *scan = (U8*)NULL;
1464 U32 wordlen = 0; /* required init */
1466 bool set_bit = trie->bitmap ? 1 : 0; /*store the first char in the bitmap?*/
1468 if (OP(noper) == NOTHING) {
1472 if ( set_bit ) /* bitmap only alloced when !(UTF&&Folding) */
1473 TRIE_BITMAP_SET(trie,*uc); /* store the raw first byte
1474 regardless of encoding */
1476 for ( ; uc < e ; uc += len ) {
1477 TRIE_CHARCOUNT(trie)++;
1481 if ( !trie->charmap[ uvc ] ) {
1482 trie->charmap[ uvc ]=( ++trie->uniquecharcount );
1484 trie->charmap[ folder[ uvc ] ] = trie->charmap[ uvc ];
1488 /* store the codepoint in the bitmap, and its folded
1490 TRIE_BITMAP_SET(trie,uvc);
1492 /* store the folded codepoint */
1493 if ( folder ) TRIE_BITMAP_SET(trie,folder[ uvc ]);
1496 /* store first byte of utf8 representation of
1497 variant codepoints */
1498 if (! UNI_IS_INVARIANT(uvc)) {
1499 TRIE_BITMAP_SET(trie, UTF8_TWO_BYTE_HI(uvc));
1502 set_bit = 0; /* We've done our bit :-) */
1507 widecharmap = newHV();
1509 svpp = hv_fetch( widecharmap, (char*)&uvc, sizeof( UV ), 1 );
1512 Perl_croak( aTHX_ "error creating/fetching widecharmap entry for 0x%"UVXf, uvc );
1514 if ( !SvTRUE( *svpp ) ) {
1515 sv_setiv( *svpp, ++trie->uniquecharcount );
1520 if( cur == first ) {
1523 } else if (chars < trie->minlen) {
1525 } else if (chars > trie->maxlen) {
1529 } /* end first pass */
1530 DEBUG_TRIE_COMPILE_r(
1531 PerlIO_printf( Perl_debug_log, "%*sTRIE(%s): W:%d C:%d Uq:%d Min:%d Max:%d\n",
1532 (int)depth * 2 + 2,"",
1533 ( widecharmap ? "UTF8" : "NATIVE" ), (int)word_count,
1534 (int)TRIE_CHARCOUNT(trie), trie->uniquecharcount,
1535 (int)trie->minlen, (int)trie->maxlen )
1539 We now know what we are dealing with in terms of unique chars and
1540 string sizes so we can calculate how much memory a naive
1541 representation using a flat table will take. If it's over a reasonable
1542 limit (as specified by ${^RE_TRIE_MAXBUF}) we use a more memory
1543 conservative but potentially much slower representation using an array
1546 At the end we convert both representations into the same compressed
1547 form that will be used in regexec.c for matching with. The latter
1548 is a form that cannot be used to construct with but has memory
1549 properties similar to the list form and access properties similar
1550 to the table form making it both suitable for fast searches and
1551 small enough that its feasable to store for the duration of a program.
1553 See the comment in the code where the compressed table is produced
1554 inplace from the flat tabe representation for an explanation of how
1555 the compression works.
1560 Newx(prev_states, TRIE_CHARCOUNT(trie) + 2, U32);
1563 if ( (IV)( ( TRIE_CHARCOUNT(trie) + 1 ) * trie->uniquecharcount + 1) > SvIV(re_trie_maxbuff) ) {
1565 Second Pass -- Array Of Lists Representation
1567 Each state will be represented by a list of charid:state records
1568 (reg_trie_trans_le) the first such element holds the CUR and LEN
1569 points of the allocated array. (See defines above).
1571 We build the initial structure using the lists, and then convert
1572 it into the compressed table form which allows faster lookups
1573 (but cant be modified once converted).
1576 STRLEN transcount = 1;
1578 DEBUG_TRIE_COMPILE_MORE_r( PerlIO_printf( Perl_debug_log,
1579 "%*sCompiling trie using list compiler\n",
1580 (int)depth * 2 + 2, ""));
1582 trie->states = (reg_trie_state *)
1583 PerlMemShared_calloc( TRIE_CHARCOUNT(trie) + 2,
1584 sizeof(reg_trie_state) );
1588 for ( cur = first ; cur < last ; cur = regnext( cur ) ) {
1590 regnode * const noper = NEXTOPER( cur );
1591 U8 *uc = (U8*)STRING( noper );
1592 const U8 * const e = uc + STR_LEN( noper );
1593 U32 state = 1; /* required init */
1594 U16 charid = 0; /* sanity init */
1595 U8 *scan = (U8*)NULL; /* sanity init */
1596 STRLEN foldlen = 0; /* required init */
1597 U32 wordlen = 0; /* required init */
1598 U8 foldbuf[ UTF8_MAXBYTES_CASE + 1 ];
1600 if (OP(noper) != NOTHING) {
1601 for ( ; uc < e ; uc += len ) {
1606 charid = trie->charmap[ uvc ];
1608 SV** const svpp = hv_fetch( widecharmap, (char*)&uvc, sizeof( UV ), 0);
1612 charid=(U16)SvIV( *svpp );
1615 /* charid is now 0 if we dont know the char read, or nonzero if we do */
1622 if ( !trie->states[ state ].trans.list ) {
1623 TRIE_LIST_NEW( state );
1625 for ( check = 1; check <= TRIE_LIST_USED( state ); check++ ) {
1626 if ( TRIE_LIST_ITEM( state, check ).forid == charid ) {
1627 newstate = TRIE_LIST_ITEM( state, check ).newstate;
1632 newstate = next_alloc++;
1633 prev_states[newstate] = state;
1634 TRIE_LIST_PUSH( state, charid, newstate );
1639 Perl_croak( aTHX_ "panic! In trie construction, no char mapping for %"IVdf, uvc );
1643 TRIE_HANDLE_WORD(state);
1645 } /* end second pass */
1647 /* next alloc is the NEXT state to be allocated */
1648 trie->statecount = next_alloc;
1649 trie->states = (reg_trie_state *)
1650 PerlMemShared_realloc( trie->states,
1652 * sizeof(reg_trie_state) );
1654 /* and now dump it out before we compress it */
1655 DEBUG_TRIE_COMPILE_MORE_r(dump_trie_interim_list(trie, widecharmap,
1656 revcharmap, next_alloc,
1660 trie->trans = (reg_trie_trans *)
1661 PerlMemShared_calloc( transcount, sizeof(reg_trie_trans) );
1668 for( state=1 ; state < next_alloc ; state ++ ) {
1672 DEBUG_TRIE_COMPILE_MORE_r(
1673 PerlIO_printf( Perl_debug_log, "tp: %d zp: %d ",tp,zp)
1677 if (trie->states[state].trans.list) {
1678 U16 minid=TRIE_LIST_ITEM( state, 1).forid;
1682 for( idx = 2 ; idx <= TRIE_LIST_USED( state ) ; idx++ ) {
1683 const U16 forid = TRIE_LIST_ITEM( state, idx).forid;
1684 if ( forid < minid ) {
1686 } else if ( forid > maxid ) {
1690 if ( transcount < tp + maxid - minid + 1) {
1692 trie->trans = (reg_trie_trans *)
1693 PerlMemShared_realloc( trie->trans,
1695 * sizeof(reg_trie_trans) );
1696 Zero( trie->trans + (transcount / 2), transcount / 2 , reg_trie_trans );
1698 base = trie->uniquecharcount + tp - minid;
1699 if ( maxid == minid ) {
1701 for ( ; zp < tp ; zp++ ) {
1702 if ( ! trie->trans[ zp ].next ) {
1703 base = trie->uniquecharcount + zp - minid;
1704 trie->trans[ zp ].next = TRIE_LIST_ITEM( state, 1).newstate;
1705 trie->trans[ zp ].check = state;
1711 trie->trans[ tp ].next = TRIE_LIST_ITEM( state, 1).newstate;
1712 trie->trans[ tp ].check = state;
1717 for ( idx=1; idx <= TRIE_LIST_USED( state ) ; idx++ ) {
1718 const U32 tid = base - trie->uniquecharcount + TRIE_LIST_ITEM( state, idx ).forid;
1719 trie->trans[ tid ].next = TRIE_LIST_ITEM( state, idx ).newstate;
1720 trie->trans[ tid ].check = state;
1722 tp += ( maxid - minid + 1 );
1724 Safefree(trie->states[ state ].trans.list);
1727 DEBUG_TRIE_COMPILE_MORE_r(
1728 PerlIO_printf( Perl_debug_log, " base: %d\n",base);
1731 trie->states[ state ].trans.base=base;
1733 trie->lasttrans = tp + 1;
1737 Second Pass -- Flat Table Representation.
1739 we dont use the 0 slot of either trans[] or states[] so we add 1 to each.
1740 We know that we will need Charcount+1 trans at most to store the data
1741 (one row per char at worst case) So we preallocate both structures
1742 assuming worst case.
1744 We then construct the trie using only the .next slots of the entry
1747 We use the .check field of the first entry of the node temporarily to
1748 make compression both faster and easier by keeping track of how many non
1749 zero fields are in the node.
1751 Since trans are numbered from 1 any 0 pointer in the table is a FAIL
1754 There are two terms at use here: state as a TRIE_NODEIDX() which is a
1755 number representing the first entry of the node, and state as a
1756 TRIE_NODENUM() which is the trans number. state 1 is TRIE_NODEIDX(1) and
1757 TRIE_NODENUM(1), state 2 is TRIE_NODEIDX(2) and TRIE_NODENUM(3) if there
1758 are 2 entrys per node. eg:
1766 The table is internally in the right hand, idx form. However as we also
1767 have to deal with the states array which is indexed by nodenum we have to
1768 use TRIE_NODENUM() to convert.
1771 DEBUG_TRIE_COMPILE_MORE_r( PerlIO_printf( Perl_debug_log,
1772 "%*sCompiling trie using table compiler\n",
1773 (int)depth * 2 + 2, ""));
1775 trie->trans = (reg_trie_trans *)
1776 PerlMemShared_calloc( ( TRIE_CHARCOUNT(trie) + 1 )
1777 * trie->uniquecharcount + 1,
1778 sizeof(reg_trie_trans) );
1779 trie->states = (reg_trie_state *)
1780 PerlMemShared_calloc( TRIE_CHARCOUNT(trie) + 2,
1781 sizeof(reg_trie_state) );
1782 next_alloc = trie->uniquecharcount + 1;
1785 for ( cur = first ; cur < last ; cur = regnext( cur ) ) {
1787 regnode * const noper = NEXTOPER( cur );
1788 const U8 *uc = (U8*)STRING( noper );
1789 const U8 * const e = uc + STR_LEN( noper );
1791 U32 state = 1; /* required init */
1793 U16 charid = 0; /* sanity init */
1794 U32 accept_state = 0; /* sanity init */
1795 U8 *scan = (U8*)NULL; /* sanity init */
1797 STRLEN foldlen = 0; /* required init */
1798 U32 wordlen = 0; /* required init */
1799 U8 foldbuf[ UTF8_MAXBYTES_CASE + 1 ];
1801 if ( OP(noper) != NOTHING ) {
1802 for ( ; uc < e ; uc += len ) {
1807 charid = trie->charmap[ uvc ];
1809 SV* const * const svpp = hv_fetch( widecharmap, (char*)&uvc, sizeof( UV ), 0);
1810 charid = svpp ? (U16)SvIV(*svpp) : 0;
1814 if ( !trie->trans[ state + charid ].next ) {
1815 trie->trans[ state + charid ].next = next_alloc;
1816 trie->trans[ state ].check++;
1817 prev_states[TRIE_NODENUM(next_alloc)]
1818 = TRIE_NODENUM(state);
1819 next_alloc += trie->uniquecharcount;
1821 state = trie->trans[ state + charid ].next;
1823 Perl_croak( aTHX_ "panic! In trie construction, no char mapping for %"IVdf, uvc );
1825 /* charid is now 0 if we dont know the char read, or nonzero if we do */
1828 accept_state = TRIE_NODENUM( state );
1829 TRIE_HANDLE_WORD(accept_state);
1831 } /* end second pass */
1833 /* and now dump it out before we compress it */
1834 DEBUG_TRIE_COMPILE_MORE_r(dump_trie_interim_table(trie, widecharmap,
1836 next_alloc, depth+1));
1840 * Inplace compress the table.*
1842 For sparse data sets the table constructed by the trie algorithm will
1843 be mostly 0/FAIL transitions or to put it another way mostly empty.
1844 (Note that leaf nodes will not contain any transitions.)
1846 This algorithm compresses the tables by eliminating most such
1847 transitions, at the cost of a modest bit of extra work during lookup:
1849 - Each states[] entry contains a .base field which indicates the
1850 index in the state[] array wheres its transition data is stored.
1852 - If .base is 0 there are no valid transitions from that node.
1854 - If .base is nonzero then charid is added to it to find an entry in
1857 -If trans[states[state].base+charid].check!=state then the
1858 transition is taken to be a 0/Fail transition. Thus if there are fail
1859 transitions at the front of the node then the .base offset will point
1860 somewhere inside the previous nodes data (or maybe even into a node
1861 even earlier), but the .check field determines if the transition is
1865 The following process inplace converts the table to the compressed
1866 table: We first do not compress the root node 1,and mark all its
1867 .check pointers as 1 and set its .base pointer as 1 as well. This
1868 allows us to do a DFA construction from the compressed table later,
1869 and ensures that any .base pointers we calculate later are greater
1872 - We set 'pos' to indicate the first entry of the second node.
1874 - We then iterate over the columns of the node, finding the first and
1875 last used entry at l and m. We then copy l..m into pos..(pos+m-l),
1876 and set the .check pointers accordingly, and advance pos
1877 appropriately and repreat for the next node. Note that when we copy
1878 the next pointers we have to convert them from the original
1879 NODEIDX form to NODENUM form as the former is not valid post
1882 - If a node has no transitions used we mark its base as 0 and do not
1883 advance the pos pointer.
1885 - If a node only has one transition we use a second pointer into the
1886 structure to fill in allocated fail transitions from other states.
1887 This pointer is independent of the main pointer and scans forward
1888 looking for null transitions that are allocated to a state. When it
1889 finds one it writes the single transition into the "hole". If the
1890 pointer doesnt find one the single transition is appended as normal.
1892 - Once compressed we can Renew/realloc the structures to release the
1895 See "Table-Compression Methods" in sec 3.9 of the Red Dragon,
1896 specifically Fig 3.47 and the associated pseudocode.
1900 const U32 laststate = TRIE_NODENUM( next_alloc );
1903 trie->statecount = laststate;
1905 for ( state = 1 ; state < laststate ; state++ ) {
1907 const U32 stateidx = TRIE_NODEIDX( state );
1908 const U32 o_used = trie->trans[ stateidx ].check;
1909 U32 used = trie->trans[ stateidx ].check;
1910 trie->trans[ stateidx ].check = 0;
1912 for ( charid = 0 ; used && charid < trie->uniquecharcount ; charid++ ) {
1913 if ( flag || trie->trans[ stateidx + charid ].next ) {
1914 if ( trie->trans[ stateidx + charid ].next ) {
1916 for ( ; zp < pos ; zp++ ) {
1917 if ( ! trie->trans[ zp ].next ) {
1921 trie->states[ state ].trans.base = zp + trie->uniquecharcount - charid ;
1922 trie->trans[ zp ].next = SAFE_TRIE_NODENUM( trie->trans[ stateidx + charid ].next );
1923 trie->trans[ zp ].check = state;
1924 if ( ++zp > pos ) pos = zp;
1931 trie->states[ state ].trans.base = pos + trie->uniquecharcount - charid ;
1933 trie->trans[ pos ].next = SAFE_TRIE_NODENUM( trie->trans[ stateidx + charid ].next );
1934 trie->trans[ pos ].check = state;
1939 trie->lasttrans = pos + 1;
1940 trie->states = (reg_trie_state *)
1941 PerlMemShared_realloc( trie->states, laststate
1942 * sizeof(reg_trie_state) );
1943 DEBUG_TRIE_COMPILE_MORE_r(
1944 PerlIO_printf( Perl_debug_log,
1945 "%*sAlloc: %d Orig: %"IVdf" elements, Final:%"IVdf". Savings of %%%5.2f\n",
1946 (int)depth * 2 + 2,"",
1947 (int)( ( TRIE_CHARCOUNT(trie) + 1 ) * trie->uniquecharcount + 1 ),
1950 ( ( next_alloc - pos ) * 100 ) / (double)next_alloc );
1953 } /* end table compress */
1955 DEBUG_TRIE_COMPILE_MORE_r(
1956 PerlIO_printf(Perl_debug_log, "%*sStatecount:%"UVxf" Lasttrans:%"UVxf"\n",
1957 (int)depth * 2 + 2, "",
1958 (UV)trie->statecount,
1959 (UV)trie->lasttrans)
1961 /* resize the trans array to remove unused space */
1962 trie->trans = (reg_trie_trans *)
1963 PerlMemShared_realloc( trie->trans, trie->lasttrans
1964 * sizeof(reg_trie_trans) );
1966 { /* Modify the program and insert the new TRIE node */
1967 U8 nodetype =(U8)(flags & 0xFF);
1971 regnode *optimize = NULL;
1972 #ifdef RE_TRACK_PATTERN_OFFSETS
1975 U32 mjd_nodelen = 0;
1976 #endif /* RE_TRACK_PATTERN_OFFSETS */
1977 #endif /* DEBUGGING */
1979 This means we convert either the first branch or the first Exact,
1980 depending on whether the thing following (in 'last') is a branch
1981 or not and whther first is the startbranch (ie is it a sub part of
1982 the alternation or is it the whole thing.)
1983 Assuming its a sub part we convert the EXACT otherwise we convert
1984 the whole branch sequence, including the first.
1986 /* Find the node we are going to overwrite */
1987 if ( first != startbranch || OP( last ) == BRANCH ) {
1988 /* branch sub-chain */
1989 NEXT_OFF( first ) = (U16)(last - first);
1990 #ifdef RE_TRACK_PATTERN_OFFSETS
1992 mjd_offset= Node_Offset((convert));
1993 mjd_nodelen= Node_Length((convert));
1996 /* whole branch chain */
1998 #ifdef RE_TRACK_PATTERN_OFFSETS
2001 const regnode *nop = NEXTOPER( convert );
2002 mjd_offset= Node_Offset((nop));
2003 mjd_nodelen= Node_Length((nop));
2007 PerlIO_printf(Perl_debug_log, "%*sMJD offset:%"UVuf" MJD length:%"UVuf"\n",
2008 (int)depth * 2 + 2, "",
2009 (UV)mjd_offset, (UV)mjd_nodelen)
2012 /* But first we check to see if there is a common prefix we can
2013 split out as an EXACT and put in front of the TRIE node. */
2014 trie->startstate= 1;
2015 if ( trie->bitmap && !widecharmap && !trie->jump ) {
2017 for ( state = 1 ; state < trie->statecount-1 ; state++ ) {
2021 const U32 base = trie->states[ state ].trans.base;
2023 if ( trie->states[state].wordnum )
2026 for ( ofs = 0 ; ofs < trie->uniquecharcount ; ofs++ ) {
2027 if ( ( base + ofs >= trie->uniquecharcount ) &&
2028 ( base + ofs - trie->uniquecharcount < trie->lasttrans ) &&
2029 trie->trans[ base + ofs - trie->uniquecharcount ].check == state )
2031 if ( ++count > 1 ) {
2032 SV **tmp = av_fetch( revcharmap, ofs, 0);
2033 const U8 *ch = (U8*)SvPV_nolen_const( *tmp );
2034 if ( state == 1 ) break;
2036 Zero(trie->bitmap, ANYOF_BITMAP_SIZE, char);
2038 PerlIO_printf(Perl_debug_log,
2039 "%*sNew Start State=%"UVuf" Class: [",
2040 (int)depth * 2 + 2, "",
2043 SV ** const tmp = av_fetch( revcharmap, idx, 0);
2044 const U8 * const ch = (U8*)SvPV_nolen_const( *tmp );
2046 TRIE_BITMAP_SET(trie,*ch);
2048 TRIE_BITMAP_SET(trie, folder[ *ch ]);
2050 PerlIO_printf(Perl_debug_log, "%s", (char*)ch)
2054 TRIE_BITMAP_SET(trie,*ch);
2056 TRIE_BITMAP_SET(trie,folder[ *ch ]);
2057 DEBUG_OPTIMISE_r(PerlIO_printf( Perl_debug_log,"%s", ch));
2063 SV **tmp = av_fetch( revcharmap, idx, 0);
2065 char *ch = SvPV( *tmp, len );
2067 SV *sv=sv_newmortal();
2068 PerlIO_printf( Perl_debug_log,
2069 "%*sPrefix State: %"UVuf" Idx:%"UVuf" Char='%s'\n",
2070 (int)depth * 2 + 2, "",
2072 pv_pretty(sv, SvPV_nolen_const(*tmp), SvCUR(*tmp), 6,
2073 PL_colors[0], PL_colors[1],
2074 (SvUTF8(*tmp) ? PERL_PV_ESCAPE_UNI : 0) |
2075 PERL_PV_ESCAPE_FIRSTCHAR
2080 OP( convert ) = nodetype;
2081 str=STRING(convert);
2084 STR_LEN(convert) += len;
2090 DEBUG_OPTIMISE_r(PerlIO_printf( Perl_debug_log,"]\n"));
2095 trie->prefixlen = (state-1);
2097 regnode *n = convert+NODE_SZ_STR(convert);
2098 NEXT_OFF(convert) = NODE_SZ_STR(convert);
2099 trie->startstate = state;
2100 trie->minlen -= (state - 1);
2101 trie->maxlen -= (state - 1);
2103 /* At least the UNICOS C compiler choked on this
2104 * being argument to DEBUG_r(), so let's just have
2107 #ifdef PERL_EXT_RE_BUILD
2113 regnode *fix = convert;
2114 U32 word = trie->wordcount;
2116 Set_Node_Offset_Length(convert, mjd_offset, state - 1);
2117 while( ++fix < n ) {
2118 Set_Node_Offset_Length(fix, 0, 0);
2121 SV ** const tmp = av_fetch( trie_words, word, 0 );
2123 if ( STR_LEN(convert) <= SvCUR(*tmp) )
2124 sv_chop(*tmp, SvPV_nolen(*tmp) + STR_LEN(convert));
2126 sv_chop(*tmp, SvPV_nolen(*tmp) + SvCUR(*tmp));
2134 NEXT_OFF(convert) = (U16)(tail - convert);
2135 DEBUG_r(optimize= n);
2141 if ( trie->maxlen ) {
2142 NEXT_OFF( convert ) = (U16)(tail - convert);
2143 ARG_SET( convert, data_slot );
2144 /* Store the offset to the first unabsorbed branch in
2145 jump[0], which is otherwise unused by the jump logic.
2146 We use this when dumping a trie and during optimisation. */
2148 trie->jump[0] = (U16)(nextbranch - convert);
2150 /* If the start state is not accepting (meaning there is no empty string/NOTHING)
2151 * and there is a bitmap
2152 * and the first "jump target" node we found leaves enough room
2153 * then convert the TRIE node into a TRIEC node, with the bitmap
2154 * embedded inline in the opcode - this is hypothetically faster.
2156 if ( !trie->states[trie->startstate].wordnum
2158 && ( (char *)jumper - (char *)convert) >= (int)sizeof(struct regnode_charclass) )
2160 OP( convert ) = TRIEC;
2161 Copy(trie->bitmap, ((struct regnode_charclass *)convert)->bitmap, ANYOF_BITMAP_SIZE, char);
2162 PerlMemShared_free(trie->bitmap);
2165 OP( convert ) = TRIE;
2167 /* store the type in the flags */
2168 convert->flags = nodetype;
2172 + regarglen[ OP( convert ) ];
2174 /* XXX We really should free up the resource in trie now,
2175 as we won't use them - (which resources?) dmq */
2177 /* needed for dumping*/
2178 DEBUG_r(if (optimize) {
2179 regnode *opt = convert;
2181 while ( ++opt < optimize) {
2182 Set_Node_Offset_Length(opt,0,0);
2185 Try to clean up some of the debris left after the
2188 while( optimize < jumper ) {
2189 mjd_nodelen += Node_Length((optimize));
2190 OP( optimize ) = OPTIMIZED;
2191 Set_Node_Offset_Length(optimize,0,0);
2194 Set_Node_Offset_Length(convert,mjd_offset,mjd_nodelen);
2196 } /* end node insert */
2197 REH_CALL_COMP_NODE_HOOK(pRExC_state->rx, convert);
2199 /* Finish populating the prev field of the wordinfo array. Walk back
2200 * from each accept state until we find another accept state, and if
2201 * so, point the first word's .prev field at the second word. If the
2202 * second already has a .prev field set, stop now. This will be the
2203 * case either if we've already processed that word's accept state,
2204 * or that state had multiple words, and the overspill words were
2205 * already linked up earlier.
2212 for (word=1; word <= trie->wordcount; word++) {
2214 if (trie->wordinfo[word].prev)
2216 state = trie->wordinfo[word].accept;
2218 state = prev_states[state];
2221 prev = trie->states[state].wordnum;
2225 trie->wordinfo[word].prev = prev;
2227 Safefree(prev_states);
2231 /* and now dump out the compressed format */
2232 DEBUG_TRIE_COMPILE_r(dump_trie(trie, widecharmap, revcharmap, depth+1));
2234 RExC_rxi->data->data[ data_slot + 1 ] = (void*)widecharmap;
2236 RExC_rxi->data->data[ data_slot + TRIE_WORDS_OFFSET ] = (void*)trie_words;
2237 RExC_rxi->data->data[ data_slot + 3 ] = (void*)revcharmap;
2239 SvREFCNT_dec(revcharmap);
2243 : trie->startstate>1
2249 S_make_trie_failtable(pTHX_ RExC_state_t *pRExC_state, regnode *source, regnode *stclass, U32 depth)
2251 /* The Trie is constructed and compressed now so we can build a fail array if it's needed
2253 This is basically the Aho-Corasick algorithm. Its from exercise 3.31 and 3.32 in the
2254 "Red Dragon" -- Compilers, principles, techniques, and tools. Aho, Sethi, Ullman 1985/88
2257 We find the fail state for each state in the trie, this state is the longest proper
2258 suffix of the current state's 'word' that is also a proper prefix of another word in our
2259 trie. State 1 represents the word '' and is thus the default fail state. This allows
2260 the DFA not to have to restart after its tried and failed a word at a given point, it
2261 simply continues as though it had been matching the other word in the first place.
2263 'abcdgu'=~/abcdefg|cdgu/
2264 When we get to 'd' we are still matching the first word, we would encounter 'g' which would
2265 fail, which would bring us to the state representing 'd' in the second word where we would
2266 try 'g' and succeed, proceeding to match 'cdgu'.
2268 /* add a fail transition */
2269 const U32 trie_offset = ARG(source);
2270 reg_trie_data *trie=(reg_trie_data *)RExC_rxi->data->data[trie_offset];
2272 const U32 ucharcount = trie->uniquecharcount;
2273 const U32 numstates = trie->statecount;
2274 const U32 ubound = trie->lasttrans + ucharcount;
2278 U32 base = trie->states[ 1 ].trans.base;
2281 const U32 data_slot = add_data( pRExC_state, 1, "T" );
2282 GET_RE_DEBUG_FLAGS_DECL;
2284 PERL_ARGS_ASSERT_MAKE_TRIE_FAILTABLE;
2286 PERL_UNUSED_ARG(depth);
2290 ARG_SET( stclass, data_slot );
2291 aho = (reg_ac_data *) PerlMemShared_calloc( 1, sizeof(reg_ac_data) );
2292 RExC_rxi->data->data[ data_slot ] = (void*)aho;
2293 aho->trie=trie_offset;
2294 aho->states=(reg_trie_state *)PerlMemShared_malloc( numstates * sizeof(reg_trie_state) );
2295 Copy( trie->states, aho->states, numstates, reg_trie_state );
2296 Newxz( q, numstates, U32);
2297 aho->fail = (U32 *) PerlMemShared_calloc( numstates, sizeof(U32) );
2300 /* initialize fail[0..1] to be 1 so that we always have
2301 a valid final fail state */
2302 fail[ 0 ] = fail[ 1 ] = 1;
2304 for ( charid = 0; charid < ucharcount ; charid++ ) {
2305 const U32 newstate = TRIE_TRANS_STATE( 1, base, ucharcount, charid, 0 );
2307 q[ q_write ] = newstate;
2308 /* set to point at the root */
2309 fail[ q[ q_write++ ] ]=1;
2312 while ( q_read < q_write) {
2313 const U32 cur = q[ q_read++ % numstates ];
2314 base = trie->states[ cur ].trans.base;
2316 for ( charid = 0 ; charid < ucharcount ; charid++ ) {
2317 const U32 ch_state = TRIE_TRANS_STATE( cur, base, ucharcount, charid, 1 );
2319 U32 fail_state = cur;
2322 fail_state = fail[ fail_state ];
2323 fail_base = aho->states[ fail_state ].trans.base;
2324 } while ( !TRIE_TRANS_STATE( fail_state, fail_base, ucharcount, charid, 1 ) );
2326 fail_state = TRIE_TRANS_STATE( fail_state, fail_base, ucharcount, charid, 1 );
2327 fail[ ch_state ] = fail_state;
2328 if ( !aho->states[ ch_state ].wordnum && aho->states[ fail_state ].wordnum )
2330 aho->states[ ch_state ].wordnum = aho->states[ fail_state ].wordnum;
2332 q[ q_write++ % numstates] = ch_state;
2336 /* restore fail[0..1] to 0 so that we "fall out" of the AC loop
2337 when we fail in state 1, this allows us to use the
2338 charclass scan to find a valid start char. This is based on the principle
2339 that theres a good chance the string being searched contains lots of stuff
2340 that cant be a start char.
2342 fail[ 0 ] = fail[ 1 ] = 0;
2343 DEBUG_TRIE_COMPILE_r({
2344 PerlIO_printf(Perl_debug_log,
2345 "%*sStclass Failtable (%"UVuf" states): 0",
2346 (int)(depth * 2), "", (UV)numstates
2348 for( q_read=1; q_read<numstates; q_read++ ) {
2349 PerlIO_printf(Perl_debug_log, ", %"UVuf, (UV)fail[q_read]);
2351 PerlIO_printf(Perl_debug_log, "\n");
2354 /*RExC_seen |= REG_SEEN_TRIEDFA;*/
2359 * There are strange code-generation bugs caused on sparc64 by gcc-2.95.2.
2360 * These need to be revisited when a newer toolchain becomes available.
2362 #if defined(__sparc64__) && defined(__GNUC__)
2363 # if __GNUC__ < 2 || (__GNUC__ == 2 && __GNUC_MINOR__ < 96)
2364 # undef SPARC64_GCC_WORKAROUND
2365 # define SPARC64_GCC_WORKAROUND 1
2369 #define DEBUG_PEEP(str,scan,depth) \
2370 DEBUG_OPTIMISE_r({if (scan){ \
2371 SV * const mysv=sv_newmortal(); \
2372 regnode *Next = regnext(scan); \
2373 regprop(RExC_rx, mysv, scan); \
2374 PerlIO_printf(Perl_debug_log, "%*s" str ">%3d: %s (%d)\n", \
2375 (int)depth*2, "", REG_NODE_NUM(scan), SvPV_nolen_const(mysv),\
2376 Next ? (REG_NODE_NUM(Next)) : 0 ); \
2383 #define JOIN_EXACT(scan,min,flags) \
2384 if (PL_regkind[OP(scan)] == EXACT) \
2385 join_exact(pRExC_state,(scan),(min),(flags),NULL,depth+1)
2388 S_join_exact(pTHX_ RExC_state_t *pRExC_state, regnode *scan, I32 *min, U32 flags,regnode *val, U32 depth) {
2389 /* Merge several consecutive EXACTish nodes into one. */
2390 regnode *n = regnext(scan);
2392 regnode *next = scan + NODE_SZ_STR(scan);
2396 regnode *stop = scan;
2397 GET_RE_DEBUG_FLAGS_DECL;
2399 PERL_UNUSED_ARG(depth);
2402 PERL_ARGS_ASSERT_JOIN_EXACT;
2403 #ifndef EXPERIMENTAL_INPLACESCAN
2404 PERL_UNUSED_ARG(flags);
2405 PERL_UNUSED_ARG(val);
2407 DEBUG_PEEP("join",scan,depth);
2409 /* Skip NOTHING, merge EXACT*. */
2411 ( PL_regkind[OP(n)] == NOTHING ||
2412 (stringok && (OP(n) == OP(scan))))
2414 && NEXT_OFF(scan) + NEXT_OFF(n) < I16_MAX) {
2416 if (OP(n) == TAIL || n > next)
2418 if (PL_regkind[OP(n)] == NOTHING) {
2419 DEBUG_PEEP("skip:",n,depth);
2420 NEXT_OFF(scan) += NEXT_OFF(n);
2421 next = n + NODE_STEP_REGNODE;
2428 else if (stringok) {
2429 const unsigned int oldl = STR_LEN(scan);
2430 regnode * const nnext = regnext(n);
2432 DEBUG_PEEP("merg",n,depth);
2435 if (oldl + STR_LEN(n) > U8_MAX)
2437 NEXT_OFF(scan) += NEXT_OFF(n);
2438 STR_LEN(scan) += STR_LEN(n);
2439 next = n + NODE_SZ_STR(n);
2440 /* Now we can overwrite *n : */
2441 Move(STRING(n), STRING(scan) + oldl, STR_LEN(n), char);
2449 #ifdef EXPERIMENTAL_INPLACESCAN
2450 if (flags && !NEXT_OFF(n)) {
2451 DEBUG_PEEP("atch", val, depth);
2452 if (reg_off_by_arg[OP(n)]) {
2453 ARG_SET(n, val - n);
2456 NEXT_OFF(n) = val - n;
2462 #define GREEK_SMALL_LETTER_IOTA_WITH_DIALYTIKA_AND_TONOS 0x0390
2463 #define IOTA_D_T GREEK_SMALL_LETTER_IOTA_WITH_DIALYTIKA_AND_TONOS
2464 #define GREEK_SMALL_LETTER_UPSILON_WITH_DIALYTIKA_AND_TONOS 0x03B0
2465 #define UPSILON_D_T GREEK_SMALL_LETTER_UPSILON_WITH_DIALYTIKA_AND_TONOS
2468 && ( OP(scan) == EXACTF || OP(scan) == EXACTFU)
2469 && ( STR_LEN(scan) >= 6 ) )
2472 Two problematic code points in Unicode casefolding of EXACT nodes:
2474 U+0390 - GREEK SMALL LETTER IOTA WITH DIALYTIKA AND TONOS
2475 U+03B0 - GREEK SMALL LETTER UPSILON WITH DIALYTIKA AND TONOS
2481 U+03B9 U+0308 U+0301 0xCE 0xB9 0xCC 0x88 0xCC 0x81
2482 U+03C5 U+0308 U+0301 0xCF 0x85 0xCC 0x88 0xCC 0x81
2484 This means that in case-insensitive matching (or "loose matching",
2485 as Unicode calls it), an EXACTF of length six (the UTF-8 encoded byte
2486 length of the above casefolded versions) can match a target string
2487 of length two (the byte length of UTF-8 encoded U+0390 or U+03B0).
2488 This would rather mess up the minimum length computation.
2490 What we'll do is to look for the tail four bytes, and then peek
2491 at the preceding two bytes to see whether we need to decrease
2492 the minimum length by four (six minus two).
2494 Thanks to the design of UTF-8, there cannot be false matches:
2495 A sequence of valid UTF-8 bytes cannot be a subsequence of
2496 another valid sequence of UTF-8 bytes.
2499 char * const s0 = STRING(scan), *s, *t;
2500 char * const s1 = s0 + STR_LEN(scan) - 1;
2501 char * const s2 = s1 - 4;
2502 #ifdef EBCDIC /* RD tunifold greek 0390 and 03B0 */
2503 const char t0[] = "\xaf\x49\xaf\x42";
2505 const char t0[] = "\xcc\x88\xcc\x81";
2507 const char * const t1 = t0 + 3;
2510 s < s2 && (t = ninstr(s, s1, t0, t1));
2513 if (((U8)t[-1] == 0x68 && (U8)t[-2] == 0xB4) ||
2514 ((U8)t[-1] == 0x46 && (U8)t[-2] == 0xB5))
2516 if (((U8)t[-1] == 0xB9 && (U8)t[-2] == 0xCE) ||
2517 ((U8)t[-1] == 0x85 && (U8)t[-2] == 0xCF))
2525 n = scan + NODE_SZ_STR(scan);
2527 if (PL_regkind[OP(n)] != NOTHING || OP(n) == NOTHING) {
2534 DEBUG_OPTIMISE_r(if (merged){DEBUG_PEEP("finl",scan,depth)});
2538 /* REx optimizer. Converts nodes into quicker variants "in place".
2539 Finds fixed substrings. */
2541 /* Stops at toplevel WHILEM as well as at "last". At end *scanp is set
2542 to the position after last scanned or to NULL. */
2544 #define INIT_AND_WITHP \
2545 assert(!and_withp); \
2546 Newx(and_withp,1,struct regnode_charclass_class); \
2547 SAVEFREEPV(and_withp)
2549 /* this is a chain of data about sub patterns we are processing that
2550 need to be handled separately/specially in study_chunk. Its so
2551 we can simulate recursion without losing state. */
2553 typedef struct scan_frame {
2554 regnode *last; /* last node to process in this frame */
2555 regnode *next; /* next node to process when last is reached */
2556 struct scan_frame *prev; /*previous frame*/
2557 I32 stop; /* what stopparen do we use */
2561 #define SCAN_COMMIT(s, data, m) scan_commit(s, data, m, is_inf)
2563 #define CASE_SYNST_FNC(nAmE) \
2565 if (flags & SCF_DO_STCLASS_AND) { \
2566 for (value = 0; value < 256; value++) \
2567 if (!is_ ## nAmE ## _cp(value)) \
2568 ANYOF_BITMAP_CLEAR(data->start_class, value); \
2571 for (value = 0; value < 256; value++) \
2572 if (is_ ## nAmE ## _cp(value)) \
2573 ANYOF_BITMAP_SET(data->start_class, value); \
2577 if (flags & SCF_DO_STCLASS_AND) { \
2578 for (value = 0; value < 256; value++) \
2579 if (is_ ## nAmE ## _cp(value)) \
2580 ANYOF_BITMAP_CLEAR(data->start_class, value); \
2583 for (value = 0; value < 256; value++) \
2584 if (!is_ ## nAmE ## _cp(value)) \
2585 ANYOF_BITMAP_SET(data->start_class, value); \
2592 S_study_chunk(pTHX_ RExC_state_t *pRExC_state, regnode **scanp,
2593 I32 *minlenp, I32 *deltap,
2598 struct regnode_charclass_class *and_withp,
2599 U32 flags, U32 depth)
2600 /* scanp: Start here (read-write). */
2601 /* deltap: Write maxlen-minlen here. */
2602 /* last: Stop before this one. */
2603 /* data: string data about the pattern */
2604 /* stopparen: treat close N as END */
2605 /* recursed: which subroutines have we recursed into */
2606 /* and_withp: Valid if flags & SCF_DO_STCLASS_OR */
2609 I32 min = 0, pars = 0, code;
2610 regnode *scan = *scanp, *next;
2612 int is_inf = (flags & SCF_DO_SUBSTR) && (data->flags & SF_IS_INF);
2613 int is_inf_internal = 0; /* The studied chunk is infinite */
2614 I32 is_par = OP(scan) == OPEN ? ARG(scan) : 0;
2615 scan_data_t data_fake;
2616 SV *re_trie_maxbuff = NULL;
2617 regnode *first_non_open = scan;
2618 I32 stopmin = I32_MAX;
2619 scan_frame *frame = NULL;
2620 GET_RE_DEBUG_FLAGS_DECL;
2622 PERL_ARGS_ASSERT_STUDY_CHUNK;
2625 StructCopy(&zero_scan_data, &data_fake, scan_data_t);
2629 while (first_non_open && OP(first_non_open) == OPEN)
2630 first_non_open=regnext(first_non_open);
2635 while ( scan && OP(scan) != END && scan < last ){
2636 /* Peephole optimizer: */
2637 DEBUG_STUDYDATA("Peep:", data,depth);
2638 DEBUG_PEEP("Peep",scan,depth);
2639 JOIN_EXACT(scan,&min,0);
2641 /* Follow the next-chain of the current node and optimize
2642 away all the NOTHINGs from it. */
2643 if (OP(scan) != CURLYX) {
2644 const int max = (reg_off_by_arg[OP(scan)]
2646 /* I32 may be smaller than U16 on CRAYs! */
2647 : (I32_MAX < U16_MAX ? I32_MAX : U16_MAX));
2648 int off = (reg_off_by_arg[OP(scan)] ? ARG(scan) : NEXT_OFF(scan));
2652 /* Skip NOTHING and LONGJMP. */
2653 while ((n = regnext(n))
2654 && ((PL_regkind[OP(n)] == NOTHING && (noff = NEXT_OFF(n)))
2655 || ((OP(n) == LONGJMP) && (noff = ARG(n))))
2656 && off + noff < max)
2658 if (reg_off_by_arg[OP(scan)])
2661 NEXT_OFF(scan) = off;
2666 /* The principal pseudo-switch. Cannot be a switch, since we
2667 look into several different things. */
2668 if (OP(scan) == BRANCH || OP(scan) == BRANCHJ
2669 || OP(scan) == IFTHEN) {
2670 next = regnext(scan);
2672 /* demq: the op(next)==code check is to see if we have "branch-branch" AFAICT */
2674 if (OP(next) == code || code == IFTHEN) {
2675 /* NOTE - There is similar code to this block below for handling
2676 TRIE nodes on a re-study. If you change stuff here check there
2678 I32 max1 = 0, min1 = I32_MAX, num = 0;
2679 struct regnode_charclass_class accum;
2680 regnode * const startbranch=scan;
2682 if (flags & SCF_DO_SUBSTR)
2683 SCAN_COMMIT(pRExC_state, data, minlenp); /* Cannot merge strings after this. */
2684 if (flags & SCF_DO_STCLASS)
2685 cl_init_zero(pRExC_state, &accum);
2687 while (OP(scan) == code) {
2688 I32 deltanext, minnext, f = 0, fake;
2689 struct regnode_charclass_class this_class;
2692 data_fake.flags = 0;
2694 data_fake.whilem_c = data->whilem_c;
2695 data_fake.last_closep = data->last_closep;
2698 data_fake.last_closep = &fake;
2700 data_fake.pos_delta = delta;
2701 next = regnext(scan);
2702 scan = NEXTOPER(scan);
2704 scan = NEXTOPER(scan);
2705 if (flags & SCF_DO_STCLASS) {
2706 cl_init(pRExC_state, &this_class);
2707 data_fake.start_class = &this_class;
2708 f = SCF_DO_STCLASS_AND;
2710 if (flags & SCF_WHILEM_VISITED_POS)
2711 f |= SCF_WHILEM_VISITED_POS;
2713 /* we suppose the run is continuous, last=next...*/
2714 minnext = study_chunk(pRExC_state, &scan, minlenp, &deltanext,
2716 stopparen, recursed, NULL, f,depth+1);
2719 if (max1 < minnext + deltanext)
2720 max1 = minnext + deltanext;
2721 if (deltanext == I32_MAX)
2722 is_inf = is_inf_internal = 1;
2724 if (data_fake.flags & (SF_HAS_PAR|SF_IN_PAR))
2726 if (data_fake.flags & SCF_SEEN_ACCEPT) {
2727 if ( stopmin > minnext)
2728 stopmin = min + min1;
2729 flags &= ~SCF_DO_SUBSTR;
2731 data->flags |= SCF_SEEN_ACCEPT;
2734 if (data_fake.flags & SF_HAS_EVAL)
2735 data->flags |= SF_HAS_EVAL;
2736 data->whilem_c = data_fake.whilem_c;
2738 if (flags & SCF_DO_STCLASS)
2739 cl_or(pRExC_state, &accum, &this_class);
2741 if (code == IFTHEN && num < 2) /* Empty ELSE branch */
2743 if (flags & SCF_DO_SUBSTR) {
2744 data->pos_min += min1;
2745 data->pos_delta += max1 - min1;
2746 if (max1 != min1 || is_inf)
2747 data->longest = &(data->longest_float);
2750 delta += max1 - min1;
2751 if (flags & SCF_DO_STCLASS_OR) {
2752 cl_or(pRExC_state, data->start_class, &accum);
2754 cl_and(data->start_class, and_withp);
2755 flags &= ~SCF_DO_STCLASS;
2758 else if (flags & SCF_DO_STCLASS_AND) {
2760 cl_and(data->start_class, &accum);
2761 flags &= ~SCF_DO_STCLASS;
2764 /* Switch to OR mode: cache the old value of
2765 * data->start_class */
2767 StructCopy(data->start_class, and_withp,
2768 struct regnode_charclass_class);
2769 flags &= ~SCF_DO_STCLASS_AND;
2770 StructCopy(&accum, data->start_class,
2771 struct regnode_charclass_class);
2772 flags |= SCF_DO_STCLASS_OR;
2773 data->start_class->flags |= ANYOF_EOS;
2777 if (PERL_ENABLE_TRIE_OPTIMISATION && OP( startbranch ) == BRANCH ) {
2780 Assuming this was/is a branch we are dealing with: 'scan' now
2781 points at the item that follows the branch sequence, whatever
2782 it is. We now start at the beginning of the sequence and look
2789 which would be constructed from a pattern like /A|LIST|OF|WORDS/
2791 If we can find such a subsequence we need to turn the first
2792 element into a trie and then add the subsequent branch exact
2793 strings to the trie.
2797 1. patterns where the whole set of branches can be converted.
2799 2. patterns where only a subset can be converted.
2801 In case 1 we can replace the whole set with a single regop
2802 for the trie. In case 2 we need to keep the start and end
2805 'BRANCH EXACT; BRANCH EXACT; BRANCH X'
2806 becomes BRANCH TRIE; BRANCH X;
2808 There is an additional case, that being where there is a
2809 common prefix, which gets split out into an EXACT like node
2810 preceding the TRIE node.
2812 If x(1..n)==tail then we can do a simple trie, if not we make
2813 a "jump" trie, such that when we match the appropriate word
2814 we "jump" to the appropriate tail node. Essentially we turn
2815 a nested if into a case structure of sorts.
2820 if (!re_trie_maxbuff) {
2821 re_trie_maxbuff = get_sv(RE_TRIE_MAXBUF_NAME, 1);
2822 if (!SvIOK(re_trie_maxbuff))
2823 sv_setiv(re_trie_maxbuff, RE_TRIE_MAXBUF_INIT);
2825 if ( SvIV(re_trie_maxbuff)>=0 ) {
2827 regnode *first = (regnode *)NULL;
2828 regnode *last = (regnode *)NULL;
2829 regnode *tail = scan;
2834 SV * const mysv = sv_newmortal(); /* for dumping */
2836 /* var tail is used because there may be a TAIL
2837 regop in the way. Ie, the exacts will point to the
2838 thing following the TAIL, but the last branch will
2839 point at the TAIL. So we advance tail. If we
2840 have nested (?:) we may have to move through several
2844 while ( OP( tail ) == TAIL ) {
2845 /* this is the TAIL generated by (?:) */
2846 tail = regnext( tail );
2851 regprop(RExC_rx, mysv, tail );
2852 PerlIO_printf( Perl_debug_log, "%*s%s%s\n",
2853 (int)depth * 2 + 2, "",
2854 "Looking for TRIE'able sequences. Tail node is: ",
2855 SvPV_nolen_const( mysv )
2861 step through the branches, cur represents each
2862 branch, noper is the first thing to be matched
2863 as part of that branch and noper_next is the
2864 regnext() of that node. if noper is an EXACT
2865 and noper_next is the same as scan (our current
2866 position in the regex) then the EXACT branch is
2867 a possible optimization target. Once we have
2868 two or more consecutive such branches we can
2869 create a trie of the EXACT's contents and stich
2870 it in place. If the sequence represents all of
2871 the branches we eliminate the whole thing and
2872 replace it with a single TRIE. If it is a
2873 subsequence then we need to stitch it in. This
2874 means the first branch has to remain, and needs
2875 to be repointed at the item on the branch chain
2876 following the last branch optimized. This could
2877 be either a BRANCH, in which case the
2878 subsequence is internal, or it could be the
2879 item following the branch sequence in which
2880 case the subsequence is at the end.
2884 /* dont use tail as the end marker for this traverse */
2885 for ( cur = startbranch ; cur != scan ; cur = regnext( cur ) ) {
2886 regnode * const noper = NEXTOPER( cur );
2887 #if defined(DEBUGGING) || defined(NOJUMPTRIE)
2888 regnode * const noper_next = regnext( noper );
2892 regprop(RExC_rx, mysv, cur);
2893 PerlIO_printf( Perl_debug_log, "%*s- %s (%d)",
2894 (int)depth * 2 + 2,"", SvPV_nolen_const( mysv ), REG_NODE_NUM(cur) );
2896 regprop(RExC_rx, mysv, noper);
2897 PerlIO_printf( Perl_debug_log, " -> %s",
2898 SvPV_nolen_const(mysv));
2901 regprop(RExC_rx, mysv, noper_next );
2902 PerlIO_printf( Perl_debug_log,"\t=> %s\t",
2903 SvPV_nolen_const(mysv));
2905 PerlIO_printf( Perl_debug_log, "(First==%d,Last==%d,Cur==%d)\n",
2906 REG_NODE_NUM(first), REG_NODE_NUM(last), REG_NODE_NUM(cur) );
2908 if ( (((first && optype!=NOTHING) ? OP( noper ) == optype
2909 : PL_regkind[ OP( noper ) ] == EXACT )
2910 || OP(noper) == NOTHING )
2912 && noper_next == tail
2917 if ( !first || optype == NOTHING ) {
2918 if (!first) first = cur;
2919 optype = OP( noper );
2925 Currently we do not believe that the trie logic can
2926 handle case insensitive matching properly when the
2927 pattern is not unicode (thus forcing unicode semantics).
2929 If/when this is fixed the following define can be swapped
2930 in below to fully enable trie logic.
2932 #define TRIE_TYPE_IS_SAFE 1
2935 #define TRIE_TYPE_IS_SAFE (UTF || optype==EXACT)
2937 if ( last && TRIE_TYPE_IS_SAFE ) {
2938 make_trie( pRExC_state,
2939 startbranch, first, cur, tail, count,
2942 if ( PL_regkind[ OP( noper ) ] == EXACT
2944 && noper_next == tail
2949 optype = OP( noper );
2959 regprop(RExC_rx, mysv, cur);
2960 PerlIO_printf( Perl_debug_log,
2961 "%*s- %s (%d) <SCAN FINISHED>\n", (int)depth * 2 + 2,
2962 "", SvPV_nolen_const( mysv ),REG_NODE_NUM(cur));
2966 if ( last && TRIE_TYPE_IS_SAFE ) {
2967 made= make_trie( pRExC_state, startbranch, first, scan, tail, count, optype, depth+1 );
2968 #ifdef TRIE_STUDY_OPT
2969 if ( ((made == MADE_EXACT_TRIE &&
2970 startbranch == first)
2971 || ( first_non_open == first )) &&
2973 flags |= SCF_TRIE_RESTUDY;
2974 if ( startbranch == first
2977 RExC_seen &=~REG_TOP_LEVEL_BRANCHES;
2987 else if ( code == BRANCHJ ) { /* single branch is optimized. */
2988 scan = NEXTOPER(NEXTOPER(scan));
2989 } else /* single branch is optimized. */
2990 scan = NEXTOPER(scan);
2992 } else if (OP(scan) == SUSPEND || OP(scan) == GOSUB || OP(scan) == GOSTART) {
2993 scan_frame *newframe = NULL;
2998 if (OP(scan) != SUSPEND) {
2999 /* set the pointer */
3000 if (OP(scan) == GOSUB) {
3002 RExC_recurse[ARG2L(scan)] = scan;
3003 start = RExC_open_parens[paren-1];
3004 end = RExC_close_parens[paren-1];
3007 start = RExC_rxi->program + 1;
3011 Newxz(recursed, (((RExC_npar)>>3) +1), U8);
3012 SAVEFREEPV(recursed);
3014 if (!PAREN_TEST(recursed,paren+1)) {
3015 PAREN_SET(recursed,paren+1);
3016 Newx(newframe,1,scan_frame);
3018 if (flags & SCF_DO_SUBSTR) {
3019 SCAN_COMMIT(pRExC_state,data,minlenp);
3020 data->longest = &(data->longest_float);
3022 is_inf = is_inf_internal = 1;
3023 if (flags & SCF_DO_STCLASS_OR) /* Allow everything */
3024 cl_anything(pRExC_state, data->start_class);
3025 flags &= ~SCF_DO_STCLASS;
3028 Newx(newframe,1,scan_frame);
3031 end = regnext(scan);
3036 SAVEFREEPV(newframe);
3037 newframe->next = regnext(scan);
3038 newframe->last = last;
3039 newframe->stop = stopparen;
3040 newframe->prev = frame;
3050 else if (OP(scan) == EXACT) {
3051 I32 l = STR_LEN(scan);
3054 const U8 * const s = (U8*)STRING(scan);
3055 l = utf8_length(s, s + l);
3056 uc = utf8_to_uvchr(s, NULL);
3058 uc = *((U8*)STRING(scan));
3061 if (flags & SCF_DO_SUBSTR) { /* Update longest substr. */
3062 /* The code below prefers earlier match for fixed
3063 offset, later match for variable offset. */
3064 if (data->last_end == -1) { /* Update the start info. */
3065 data->last_start_min = data->pos_min;
3066 data->last_start_max = is_inf
3067 ? I32_MAX : data->pos_min + data->pos_delta;
3069 sv_catpvn(data->last_found, STRING(scan), STR_LEN(scan));
3071 SvUTF8_on(data->last_found);
3073 SV * const sv = data->last_found;
3074 MAGIC * const mg = SvUTF8(sv) && SvMAGICAL(sv) ?
3075 mg_find(sv, PERL_MAGIC_utf8) : NULL;
3076 if (mg && mg->mg_len >= 0)
3077 mg->mg_len += utf8_length((U8*)STRING(scan),
3078 (U8*)STRING(scan)+STR_LEN(scan));
3080 data->last_end = data->pos_min + l;
3081 data->pos_min += l; /* As in the first entry. */
3082 data->flags &= ~SF_BEFORE_EOL;
3084 if (flags & SCF_DO_STCLASS_AND) {
3085 /* Check whether it is compatible with what we know already! */
3089 /* If compatible, we or it in below. It is compatible if is
3090 * in the bitmp and either 1) its bit or its fold is set, or 2)
3091 * it's for a locale. Even if there isn't unicode semantics
3092 * here, at runtime there may be because of matching against a
3093 * utf8 string, so accept a possible false positive for
3094 * latin1-range folds */
3096 (!(data->start_class->flags & (ANYOF_CLASS | ANYOF_LOCALE))
3097 && !ANYOF_BITMAP_TEST(data->start_class, uc)
3098 && (!(data->start_class->flags & ANYOF_LOC_NONBITMAP_FOLD)
3099 || !ANYOF_BITMAP_TEST(data->start_class, PL_fold_latin1[uc])))
3102 ANYOF_CLASS_ZERO(data->start_class);
3103 ANYOF_BITMAP_ZERO(data->start_class);
3105 ANYOF_BITMAP_SET(data->start_class, uc);
3106 data->start_class->flags &= ~ANYOF_EOS;
3108 data->start_class->flags &= ~ANYOF_UNICODE_ALL;
3110 else if (flags & SCF_DO_STCLASS_OR) {
3111 /* false positive possible if the class is case-folded */
3113 ANYOF_BITMAP_SET(data->start_class, uc);
3115 data->start_class->flags |= ANYOF_UNICODE_ALL;
3116 data->start_class->flags &= ~ANYOF_EOS;
3117 cl_and(data->start_class, and_withp);
3119 flags &= ~SCF_DO_STCLASS;
3121 else if (PL_regkind[OP(scan)] == EXACT) { /* But OP != EXACT! */
3122 I32 l = STR_LEN(scan);
3123 UV uc = *((U8*)STRING(scan));
3125 /* Search for fixed substrings supports EXACT only. */
3126 if (flags & SCF_DO_SUBSTR) {
3128 SCAN_COMMIT(pRExC_state, data, minlenp);
3131 const U8 * const s = (U8 *)STRING(scan);
3132 l = utf8_length(s, s + l);
3133 uc = utf8_to_uvchr(s, NULL);
3136 if (flags & SCF_DO_SUBSTR)
3138 if (flags & SCF_DO_STCLASS_AND) {
3139 /* Check whether it is compatible with what we know already! */
3142 (!(data->start_class->flags & (ANYOF_CLASS | ANYOF_LOCALE))
3143 && !ANYOF_BITMAP_TEST(data->start_class, uc)
3144 && !ANYOF_BITMAP_TEST(data->start_class, PL_fold_latin1[uc])))
3148 ANYOF_CLASS_ZERO(data->start_class);
3149 ANYOF_BITMAP_ZERO(data->start_class);
3151 ANYOF_BITMAP_SET(data->start_class, uc);
3152 data->start_class->flags &= ~ANYOF_EOS;
3153 data->start_class->flags |= ANYOF_LOC_NONBITMAP_FOLD;
3154 if (OP(scan) == EXACTFL) {
3155 data->start_class->flags |= ANYOF_LOCALE;
3159 /* Also set the other member of the fold pair. In case
3160 * that unicode semantics is called for at runtime, use
3161 * the full latin1 fold. (Can't do this for locale,
3162 * because not known until runtime */
3163 ANYOF_BITMAP_SET(data->start_class, PL_fold_latin1[uc]);
3167 else if (flags & SCF_DO_STCLASS_OR) {
3168 if (data->start_class->flags & ANYOF_LOC_NONBITMAP_FOLD) {
3169 /* false positive possible if the class is case-folded.
3170 Assume that the locale settings are the same... */
3172 ANYOF_BITMAP_SET(data->start_class, uc);
3173 if (OP(scan) != EXACTFL) {
3175 /* And set the other member of the fold pair, but
3176 * can't do that in locale because not known until
3178 ANYOF_BITMAP_SET(data->start_class,
3179 PL_fold_latin1[uc]);
3182 data->start_class->flags &= ~ANYOF_EOS;
3184 cl_and(data->start_class, and_withp);
3186 flags &= ~SCF_DO_STCLASS;
3188 else if (REGNODE_VARIES(OP(scan))) {
3189 I32 mincount, maxcount, minnext, deltanext, fl = 0;
3190 I32 f = flags, pos_before = 0;
3191 regnode * const oscan = scan;
3192 struct regnode_charclass_class this_class;
3193 struct regnode_charclass_class *oclass = NULL;
3194 I32 next_is_eval = 0;
3196 switch (PL_regkind[OP(scan)]) {
3197 case WHILEM: /* End of (?:...)* . */
3198 scan = NEXTOPER(scan);
3201 if (flags & (SCF_DO_SUBSTR | SCF_DO_STCLASS)) {
3202 next = NEXTOPER(scan);
3203 if (OP(next) == EXACT || (flags & SCF_DO_STCLASS)) {
3205 maxcount = REG_INFTY;
3206 next = regnext(scan);
3207 scan = NEXTOPER(scan);
3211 if (flags & SCF_DO_SUBSTR)
3216 if (flags & SCF_DO_STCLASS) {
3218 maxcount = REG_INFTY;
3219 next = regnext(scan);
3220 scan = NEXTOPER(scan);
3223 is_inf = is_inf_internal = 1;
3224 scan = regnext(scan);
3225 if (flags & SCF_DO_SUBSTR) {
3226 SCAN_COMMIT(pRExC_state, data, minlenp); /* Cannot extend fixed substrings */
3227 data->longest = &(data->longest_float);
3229 goto optimize_curly_tail;
3231 if (stopparen>0 && (OP(scan)==CURLYN || OP(scan)==CURLYM)
3232 && (scan->flags == stopparen))
3237 mincount = ARG1(scan);
3238 maxcount = ARG2(scan);
3240 next = regnext(scan);
3241 if (OP(scan) == CURLYX) {
3242 I32 lp = (data ? *(data->last_closep) : 0);
3243 scan->flags = ((lp <= (I32)U8_MAX) ? (U8)lp : U8_MAX);
3245 scan = NEXTOPER(scan) + EXTRA_STEP_2ARGS;
3246 next_is_eval = (OP(scan) == EVAL);
3248 if (flags & SCF_DO_SUBSTR) {
3249 if (mincount == 0) SCAN_COMMIT(pRExC_state,data,minlenp); /* Cannot extend fixed substrings */
3250 pos_before = data->pos_min;
3254 data->flags &= ~(SF_HAS_PAR|SF_IN_PAR|SF_HAS_EVAL);
3256 data->flags |= SF_IS_INF;
3258 if (flags & SCF_DO_STCLASS) {
3259 cl_init(pRExC_state, &this_class);
3260 oclass = data->start_class;
3261 data->start_class = &this_class;
3262 f |= SCF_DO_STCLASS_AND;
3263 f &= ~SCF_DO_STCLASS_OR;
3265 /* Exclude from super-linear cache processing any {n,m}
3266 regops for which the combination of input pos and regex
3267 pos is not enough information to determine if a match
3270 For example, in the regex /foo(bar\s*){4,8}baz/ with the
3271 regex pos at the \s*, the prospects for a match depend not
3272 only on the input position but also on how many (bar\s*)
3273 repeats into the {4,8} we are. */
3274 if ((mincount > 1) || (maxcount > 1 && maxcount != REG_INFTY))
3275 f &= ~SCF_WHILEM_VISITED_POS;
3277 /* This will finish on WHILEM, setting scan, or on NULL: */
3278 minnext = study_chunk(pRExC_state, &scan, minlenp, &deltanext,
3279 last, data, stopparen, recursed, NULL,
3281 ? (f & ~SCF_DO_SUBSTR) : f),depth+1);
3283 if (flags & SCF_DO_STCLASS)
3284 data->start_class = oclass;
3285 if (mincount == 0 || minnext == 0) {
3286 if (flags & SCF_DO_STCLASS_OR) {
3287 cl_or(pRExC_state, data->start_class, &this_class);
3289 else if (flags & SCF_DO_STCLASS_AND) {
3290 /* Switch to OR mode: cache the old value of
3291 * data->start_class */
3293 StructCopy(data->start_class, and_withp,
3294 struct regnode_charclass_class);
3295 flags &= ~SCF_DO_STCLASS_AND;
3296 StructCopy(&this_class, data->start_class,
3297 struct regnode_charclass_class);
3298 flags |= SCF_DO_STCLASS_OR;
3299 data->start_class->flags |= ANYOF_EOS;
3301 } else { /* Non-zero len */
3302 if (flags & SCF_DO_STCLASS_OR) {
3303 cl_or(pRExC_state, data->start_class, &this_class);
3304 cl_and(data->start_class, and_withp);
3306 else if (flags & SCF_DO_STCLASS_AND)
3307 cl_and(data->start_class, &this_class);
3308 flags &= ~SCF_DO_STCLASS;
3310 if (!scan) /* It was not CURLYX, but CURLY. */
3312 if ( /* ? quantifier ok, except for (?{ ... }) */
3313 (next_is_eval || !(mincount == 0 && maxcount == 1))
3314 && (minnext == 0) && (deltanext == 0)
3315 && data && !(data->flags & (SF_HAS_PAR|SF_IN_PAR))
3316 && maxcount <= REG_INFTY/3) /* Complement check for big count */
3318 ckWARNreg(RExC_parse,
3319 "Quantifier unexpected on zero-length expression");
3322 min += minnext * mincount;
3323 is_inf_internal |= ((maxcount == REG_INFTY
3324 && (minnext + deltanext) > 0)
3325 || deltanext == I32_MAX);
3326 is_inf |= is_inf_internal;
3327 delta += (minnext + deltanext) * maxcount - minnext * mincount;
3329 /* Try powerful optimization CURLYX => CURLYN. */
3330 if ( OP(oscan) == CURLYX && data
3331 && data->flags & SF_IN_PAR
3332 && !(data->flags & SF_HAS_EVAL)
3333 && !deltanext && minnext == 1 ) {
3334 /* Try to optimize to CURLYN. */
3335 regnode *nxt = NEXTOPER(oscan) + EXTRA_STEP_2ARGS;
3336 regnode * const nxt1 = nxt;
3343 if (!REGNODE_SIMPLE(OP(nxt))
3344 && !(PL_regkind[OP(nxt)] == EXACT
3345 && STR_LEN(nxt) == 1))
3351 if (OP(nxt) != CLOSE)
3353 if (RExC_open_parens) {
3354 RExC_open_parens[ARG(nxt1)-1]=oscan; /*open->CURLYM*/
3355 RExC_close_parens[ARG(nxt1)-1]=nxt+2; /*close->while*/
3357 /* Now we know that nxt2 is the only contents: */
3358 oscan->flags = (U8)ARG(nxt);
3360 OP(nxt1) = NOTHING; /* was OPEN. */
3363 OP(nxt1 + 1) = OPTIMIZED; /* was count. */
3364 NEXT_OFF(nxt1+ 1) = 0; /* just for consistency. */
3365 NEXT_OFF(nxt2) = 0; /* just for consistency with CURLY. */
3366 OP(nxt) = OPTIMIZED; /* was CLOSE. */
3367 OP(nxt + 1) = OPTIMIZED; /* was count. */
3368 NEXT_OFF(nxt+ 1) = 0; /* just for consistency. */
3373 /* Try optimization CURLYX => CURLYM. */
3374 if ( OP(oscan) == CURLYX && data
3375 && !(data->flags & SF_HAS_PAR)
3376 && !(data->flags & SF_HAS_EVAL)
3377 && !deltanext /* atom is fixed width */
3378 && minnext != 0 /* CURLYM can't handle zero width */
3380 /* XXXX How to optimize if data == 0? */
3381 /* Optimize to a simpler form. */
3382 regnode *nxt = NEXTOPER(oscan) + EXTRA_STEP_2ARGS; /* OPEN */
3386 while ( (nxt2 = regnext(nxt)) /* skip over embedded stuff*/
3387 && (OP(nxt2) != WHILEM))
3389 OP(nxt2) = SUCCEED; /* Whas WHILEM */
3390 /* Need to optimize away parenths. */
3391 if ((data->flags & SF_IN_PAR) && OP(nxt) == CLOSE) {
3392 /* Set the parenth number. */
3393 regnode *nxt1 = NEXTOPER(oscan) + EXTRA_STEP_2ARGS; /* OPEN*/
3395 oscan->flags = (U8)ARG(nxt);
3396 if (RExC_open_parens) {
3397 RExC_open_parens[ARG(nxt1)-1]=oscan; /*open->CURLYM*/
3398 RExC_close_parens[ARG(nxt1)-1]=nxt2+1; /*close->NOTHING*/
3400 OP(nxt1) = OPTIMIZED; /* was OPEN. */
3401 OP(nxt) = OPTIMIZED; /* was CLOSE. */
3404 OP(nxt1 + 1) = OPTIMIZED; /* was count. */
3405 OP(nxt + 1) = OPTIMIZED; /* was count. */
3406 NEXT_OFF(nxt1 + 1) = 0; /* just for consistency. */
3407 NEXT_OFF(nxt + 1) = 0; /* just for consistency. */
3410 while ( nxt1 && (OP(nxt1) != WHILEM)) {
3411 regnode *nnxt = regnext(nxt1);
3413 if (reg_off_by_arg[OP(nxt1)])
3414 ARG_SET(nxt1, nxt2 - nxt1);
3415 else if (nxt2 - nxt1 < U16_MAX)
3416 NEXT_OFF(nxt1) = nxt2 - nxt1;
3418 OP(nxt) = NOTHING; /* Cannot beautify */
3423 /* Optimize again: */
3424 study_chunk(pRExC_state, &nxt1, minlenp, &deltanext, nxt,
3425 NULL, stopparen, recursed, NULL, 0,depth+1);
3430 else if ((OP(oscan) == CURLYX)
3431 && (flags & SCF_WHILEM_VISITED_POS)
3432 /* See the comment on a similar expression above.
3433 However, this time it's not a subexpression
3434 we care about, but the expression itself. */
3435 && (maxcount == REG_INFTY)
3436 && data && ++data->whilem_c < 16) {
3437 /* This stays as CURLYX, we can put the count/of pair. */
3438 /* Find WHILEM (as in regexec.c) */
3439 regnode *nxt = oscan + NEXT_OFF(oscan);
3441 if (OP(PREVOPER(nxt)) == NOTHING) /* LONGJMP */
3443 PREVOPER(nxt)->flags = (U8)(data->whilem_c
3444 | (RExC_whilem_seen << 4)); /* On WHILEM */
3446 if (data && fl & (SF_HAS_PAR|SF_IN_PAR))
3448 if (flags & SCF_DO_SUBSTR) {
3449 SV *last_str = NULL;
3450 int counted = mincount != 0;
3452 if (data->last_end > 0 && mincount != 0) { /* Ends with a string. */
3453 #if defined(SPARC64_GCC_WORKAROUND)
3456 const char *s = NULL;
3459 if (pos_before >= data->last_start_min)
3462 b = data->last_start_min;
3465 s = SvPV_const(data->last_found, l);
3466 old = b - data->last_start_min;
3469 I32 b = pos_before >= data->last_start_min
3470 ? pos_before : data->last_start_min;
3472 const char * const s = SvPV_const(data->last_found, l);
3473 I32 old = b - data->last_start_min;
3477 old = utf8_hop((U8*)s, old) - (U8*)s;
3479 /* Get the added string: */
3480 last_str = newSVpvn_utf8(s + old, l, UTF);
3481 if (deltanext == 0 && pos_before == b) {
3482 /* What was added is a constant string */
3484 SvGROW(last_str, (mincount * l) + 1);
3485 repeatcpy(SvPVX(last_str) + l,
3486 SvPVX_const(last_str), l, mincount - 1);
3487 SvCUR_set(last_str, SvCUR(last_str) * mincount);
3488 /* Add additional parts. */
3489 SvCUR_set(data->last_found,
3490 SvCUR(data->last_found) - l);
3491 sv_catsv(data->last_found, last_str);
3493 SV * sv = data->last_found;
3495 SvUTF8(sv) && SvMAGICAL(sv) ?
3496 mg_find(sv, PERL_MAGIC_utf8) : NULL;
3497 if (mg && mg->mg_len >= 0)
3498 mg->mg_len += CHR_SVLEN(last_str) - l;
3500 data->last_end += l * (mincount - 1);
3503 /* start offset must point into the last copy */
3504 data->last_start_min += minnext * (mincount - 1);
3505 data->last_start_max += is_inf ? I32_MAX
3506 : (maxcount - 1) * (minnext + data->pos_delta);
3509 /* It is counted once already... */
3510 data->pos_min += minnext * (mincount - counted);
3511 data->pos_delta += - counted * deltanext +
3512 (minnext + deltanext) * maxcount - minnext * mincount;
3513 if (mincount != maxcount) {
3514 /* Cannot extend fixed substrings found inside
3516 SCAN_COMMIT(pRExC_state,data,minlenp);
3517 if (mincount && last_str) {
3518 SV * const sv = data->last_found;
3519 MAGIC * const mg = SvUTF8(sv) && SvMAGICAL(sv) ?
3520 mg_find(sv, PERL_MAGIC_utf8) : NULL;
3524 sv_setsv(sv, last_str);
3525 data->last_end = data->pos_min;
3526 data->last_start_min =
3527 data->pos_min - CHR_SVLEN(last_str);
3528 data->last_start_max = is_inf
3530 : data->pos_min + data->pos_delta
3531 - CHR_SVLEN(last_str);
3533 data->longest = &(data->longest_float);
3535 SvREFCNT_dec(last_str);
3537 if (data && (fl & SF_HAS_EVAL))
3538 data->flags |= SF_HAS_EVAL;
3539 optimize_curly_tail:
3540 if (OP(oscan) != CURLYX) {
3541 while (PL_regkind[OP(next = regnext(oscan))] == NOTHING
3543 NEXT_OFF(oscan) += NEXT_OFF(next);
3546 default: /* REF, ANYOFV, and CLUMP only? */
3547 if (flags & SCF_DO_SUBSTR) {
3548 SCAN_COMMIT(pRExC_state,data,minlenp); /* Cannot expect anything... */
3549 data->longest = &(data->longest_float);
3551 is_inf = is_inf_internal = 1;
3552 if (flags & SCF_DO_STCLASS_OR)
3553 cl_anything(pRExC_state, data->start_class);
3554 flags &= ~SCF_DO_STCLASS;
3558 else if (OP(scan) == LNBREAK) {
3559 if (flags & SCF_DO_STCLASS) {
3561 data->start_class->flags &= ~ANYOF_EOS; /* No match on empty */
3562 if (flags & SCF_DO_STCLASS_AND) {
3563 for (value = 0; value < 256; value++)
3564 if (!is_VERTWS_cp(value))
3565 ANYOF_BITMAP_CLEAR(data->start_class, value);
3568 for (value = 0; value < 256; value++)
3569 if (is_VERTWS_cp(value))
3570 ANYOF_BITMAP_SET(data->start_class, value);
3572 if (flags & SCF_DO_STCLASS_OR)
3573 cl_and(data->start_class, and_withp);
3574 flags &= ~SCF_DO_STCLASS;
3578 if (flags & SCF_DO_SUBSTR) {
3579 SCAN_COMMIT(pRExC_state,data,minlenp); /* Cannot expect anything... */
3581 data->pos_delta += 1;
3582 data->longest = &(data->longest_float);
3585 else if (OP(scan) == FOLDCHAR) {
3586 int d = ARG(scan) == LATIN_SMALL_LETTER_SHARP_S ? 1 : 2;
3587 flags &= ~SCF_DO_STCLASS;
3590 if (flags & SCF_DO_SUBSTR) {
3591 SCAN_COMMIT(pRExC_state,data,minlenp); /* Cannot expect anything... */
3593 data->pos_delta += d;
3594 data->longest = &(data->longest_float);
3597 else if (REGNODE_SIMPLE(OP(scan))) {
3600 if (flags & SCF_DO_SUBSTR) {
3601 SCAN_COMMIT(pRExC_state,data,minlenp);
3605 if (flags & SCF_DO_STCLASS) {
3606 data->start_class->flags &= ~ANYOF_EOS; /* No match on empty */
3608 /* Some of the logic below assumes that switching
3609 locale on will only add false positives. */
3610 switch (PL_regkind[OP(scan)]) {
3614 /* Perl_croak(aTHX_ "panic: unexpected simple REx opcode %d", OP(scan)); */
3615 if (flags & SCF_DO_STCLASS_OR) /* Allow everything */
3616 cl_anything(pRExC_state, data->start_class);
3619 if (OP(scan) == SANY)
3621 if (flags & SCF_DO_STCLASS_OR) { /* Everything but \n */
3622 value = (ANYOF_BITMAP_TEST(data->start_class,'\n')
3623 || ANYOF_CLASS_TEST_ANY_SET(data->start_class));
3624 cl_anything(pRExC_state, data->start_class);
3626 if (flags & SCF_DO_STCLASS_AND || !value)
3627 ANYOF_BITMAP_CLEAR(data->start_class,'\n');
3630 if (flags & SCF_DO_STCLASS_AND)
3631 cl_and(data->start_class,
3632 (struct regnode_charclass_class*)scan);
3634 cl_or(pRExC_state, data->start_class,
3635 (struct regnode_charclass_class*)scan);
3638 if (flags & SCF_DO_STCLASS_AND) {
3639 if (!(data->start_class->flags & ANYOF_LOCALE)) {
3640 ANYOF_CLASS_CLEAR(data->start_class,ANYOF_NALNUM);
3641 if (OP(scan) == ALNUMU) {
3642 for (value = 0; value < 256; value++) {
3643 if (!isWORDCHAR_L1(value)) {
3644 ANYOF_BITMAP_CLEAR(data->start_class, value);
3648 for (value = 0; value < 256; value++) {
3649 if (!isALNUM(value)) {
3650 ANYOF_BITMAP_CLEAR(data->start_class, value);
3657 if (data->start_class->flags & ANYOF_LOCALE)
3658 ANYOF_CLASS_SET(data->start_class,ANYOF_ALNUM);
3659 else if (OP(scan) == ALNUMU) {
3660 for (value = 0; value < 256; value++) {
3661 if (isWORDCHAR_L1(value)) {
3662 ANYOF_BITMAP_SET(data->start_class, value);
3666 for (value = 0; value < 256; value++) {
3667 if (isALNUM(value)) {
3668 ANYOF_BITMAP_SET(data->start_class, value);
3675 if (flags & SCF_DO_STCLASS_AND) {
3676 if (!(data->start_class->flags & ANYOF_LOCALE)) {
3677 ANYOF_CLASS_CLEAR(data->start_class,ANYOF_ALNUM);
3678 if (OP(scan) == NALNUMU) {
3679 for (value = 0; value < 256; value++) {
3680 if (isWORDCHAR_L1(value)) {
3681 ANYOF_BITMAP_CLEAR(data->start_class, value);
3685 for (value = 0; value < 256; value++) {
3686 if (isALNUM(value)) {
3687 ANYOF_BITMAP_CLEAR(data->start_class, value);
3694 if (data->start_class->flags & ANYOF_LOCALE)
3695 ANYOF_CLASS_SET(data->start_class,ANYOF_NALNUM);
3697 if (OP(scan) == NALNUMU) {
3698 for (value = 0; value < 256; value++) {
3699 if (! isWORDCHAR_L1(value)) {
3700 ANYOF_BITMAP_SET(data->start_class, value);
3704 for (value = 0; value < 256; value++) {
3705 if (! isALNUM(value)) {
3706 ANYOF_BITMAP_SET(data->start_class, value);
3714 if (flags & SCF_DO_STCLASS_AND) {
3715 if (!(data->start_class->flags & ANYOF_LOCALE)) {
3716 ANYOF_CLASS_CLEAR(data->start_class,ANYOF_NSPACE);
3717 if (OP(scan) == SPACEU) {
3718 for (value = 0; value < 256; value++) {
3719 if (!isSPACE_L1(value)) {
3720 ANYOF_BITMAP_CLEAR(data->start_class, value);
3724 for (value = 0; value < 256; value++) {
3725 if (!isSPACE(value)) {
3726 ANYOF_BITMAP_CLEAR(data->start_class, value);
3733 if (data->start_class->flags & ANYOF_LOCALE) {
3734 ANYOF_CLASS_SET(data->start_class,ANYOF_SPACE);
3736 else if (OP(scan) == SPACEU) {
3737 for (value = 0; value < 256; value++) {
3738 if (isSPACE_L1(value)) {
3739 ANYOF_BITMAP_SET(data->start_class, value);
3743 for (value = 0; value < 256; value++) {
3744 if (isSPACE(value)) {
3745 ANYOF_BITMAP_SET(data->start_class, value);
3752 if (flags & SCF_DO_STCLASS_AND) {
3753 if (!(data->start_class->flags & ANYOF_LOCALE)) {
3754 ANYOF_CLASS_CLEAR(data->start_class,ANYOF_SPACE);
3755 if (OP(scan) == NSPACEU) {
3756 for (value = 0; value < 256; value++) {
3757 if (isSPACE_L1(value)) {
3758 ANYOF_BITMAP_CLEAR(data->start_class, value);
3762 for (value = 0; value < 256; value++) {
3763 if (isSPACE(value)) {
3764 ANYOF_BITMAP_CLEAR(data->start_class, value);
3771 if (data->start_class->flags & ANYOF_LOCALE)
3772 ANYOF_CLASS_SET(data->start_class,ANYOF_NSPACE);
3773 else if (OP(scan) == NSPACEU) {
3774 for (value = 0; value < 256; value++) {
3775 if (!isSPACE_L1(value)) {
3776 ANYOF_BITMAP_SET(data->start_class, value);
3781 for (value = 0; value < 256; value++) {
3782 if (!isSPACE(value)) {
3783 ANYOF_BITMAP_SET(data->start_class, value);
3790 if (flags & SCF_DO_STCLASS_AND) {
3791 ANYOF_CLASS_CLEAR(data->start_class,ANYOF_NDIGIT);
3792 for (value = 0; value < 256; value++)
3793 if (!isDIGIT(value))
3794 ANYOF_BITMAP_CLEAR(data->start_class, value);
3797 if (data->start_class->flags & ANYOF_LOCALE)
3798 ANYOF_CLASS_SET(data->start_class,ANYOF_DIGIT);
3800 for (value = 0; value < 256; value++)
3802 ANYOF_BITMAP_SET(data->start_class, value);
3807 if (flags & SCF_DO_STCLASS_AND) {
3808 ANYOF_CLASS_CLEAR(data->start_class,ANYOF_DIGIT);
3809 for (value = 0; value < 256; value++)
3811 ANYOF_BITMAP_CLEAR(data->start_class, value);
3814 if (data->start_class->flags & ANYOF_LOCALE)
3815 ANYOF_CLASS_SET(data->start_class,ANYOF_NDIGIT);
3817 for (value = 0; value < 256; value++)
3818 if (!isDIGIT(value))
3819 ANYOF_BITMAP_SET(data->start_class, value);
3823 CASE_SYNST_FNC(VERTWS);
3824 CASE_SYNST_FNC(HORIZWS);
3827 if (flags & SCF_DO_STCLASS_OR)
3828 cl_and(data->start_class, and_withp);
3829 flags &= ~SCF_DO_STCLASS;
3832 else if (PL_regkind[OP(scan)] == EOL && flags & SCF_DO_SUBSTR) {
3833 data->flags |= (OP(scan) == MEOL
3837 else if ( PL_regkind[OP(scan)] == BRANCHJ
3838 /* Lookbehind, or need to calculate parens/evals/stclass: */
3839 && (scan->flags || data || (flags & SCF_DO_STCLASS))
3840 && (OP(scan) == IFMATCH || OP(scan) == UNLESSM)) {
3841 if ( !PERL_ENABLE_POSITIVE_ASSERTION_STUDY
3842 || OP(scan) == UNLESSM )
3844 /* Negative Lookahead/lookbehind
3845 In this case we can't do fixed string optimisation.
3848 I32 deltanext, minnext, fake = 0;
3850 struct regnode_charclass_class intrnl;
3853 data_fake.flags = 0;
3855 data_fake.whilem_c = data->whilem_c;
3856 data_fake.last_closep = data->last_closep;
3859 data_fake.last_closep = &fake;
3860 data_fake.pos_delta = delta;
3861 if ( flags & SCF_DO_STCLASS && !scan->flags
3862 && OP(scan) == IFMATCH ) { /* Lookahead */
3863 cl_init(pRExC_state, &intrnl);
3864 data_fake.start_class = &intrnl;
3865 f |= SCF_DO_STCLASS_AND;
3867 if (flags & SCF_WHILEM_VISITED_POS)
3868 f |= SCF_WHILEM_VISITED_POS;
3869 next = regnext(scan);
3870 nscan = NEXTOPER(NEXTOPER(scan));
3871 minnext = study_chunk(pRExC_state, &nscan, minlenp, &deltanext,
3872 last, &data_fake, stopparen, recursed, NULL, f, depth+1);
3875 FAIL("Variable length lookbehind not implemented");
3877 else if (minnext > (I32)U8_MAX) {
3878 FAIL2("Lookbehind longer than %"UVuf" not implemented", (UV)U8_MAX);
3880 scan->flags = (U8)minnext;
3883 if (data_fake.flags & (SF_HAS_PAR|SF_IN_PAR))
3885 if (data_fake.flags & SF_HAS_EVAL)
3886 data->flags |= SF_HAS_EVAL;
3887 data->whilem_c = data_fake.whilem_c;
3889 if (f & SCF_DO_STCLASS_AND) {
3890 if (flags & SCF_DO_STCLASS_OR) {
3891 /* OR before, AND after: ideally we would recurse with
3892 * data_fake to get the AND applied by study of the
3893 * remainder of the pattern, and then derecurse;
3894 * *** HACK *** for now just treat as "no information".
3895 * See [perl #56690].
3897 cl_init(pRExC_state, data->start_class);
3899 /* AND before and after: combine and continue */
3900 const int was = (data->start_class->flags & ANYOF_EOS);
3902 cl_and(data->start_class, &intrnl);
3904 data->start_class->flags |= ANYOF_EOS;
3908 #if PERL_ENABLE_POSITIVE_ASSERTION_STUDY
3910 /* Positive Lookahead/lookbehind
3911 In this case we can do fixed string optimisation,
3912 but we must be careful about it. Note in the case of
3913 lookbehind the positions will be offset by the minimum
3914 length of the pattern, something we won't know about
3915 until after the recurse.
3917 I32 deltanext, fake = 0;
3919 struct regnode_charclass_class intrnl;
3921 /* We use SAVEFREEPV so that when the full compile
3922 is finished perl will clean up the allocated
3923 minlens when it's all done. This way we don't
3924 have to worry about freeing them when we know
3925 they wont be used, which would be a pain.
3928 Newx( minnextp, 1, I32 );
3929 SAVEFREEPV(minnextp);
3932 StructCopy(data, &data_fake, scan_data_t);
3933 if ((flags & SCF_DO_SUBSTR) && data->last_found) {
3936 SCAN_COMMIT(pRExC_state, &data_fake,minlenp);
3937 data_fake.last_found=newSVsv(data->last_found);
3941 data_fake.last_closep = &fake;
3942 data_fake.flags = 0;
3943 data_fake.pos_delta = delta;
3945 data_fake.flags |= SF_IS_INF;
3946 if ( flags & SCF_DO_STCLASS && !scan->flags
3947 && OP(scan) == IFMATCH ) { /* Lookahead */
3948 cl_init(pRExC_state, &intrnl);
3949 data_fake.start_class = &intrnl;
3950 f |= SCF_DO_STCLASS_AND;
3952 if (flags & SCF_WHILEM_VISITED_POS)
3953 f |= SCF_WHILEM_VISITED_POS;
3954 next = regnext(scan);
3955 nscan = NEXTOPER(NEXTOPER(scan));
3957 *minnextp = study_chunk(pRExC_state, &nscan, minnextp, &deltanext,
3958 last, &data_fake, stopparen, recursed, NULL, f,depth+1);
3961 FAIL("Variable length lookbehind not implemented");
3963 else if (*minnextp > (I32)U8_MAX) {
3964 FAIL2("Lookbehind longer than %"UVuf" not implemented", (UV)U8_MAX);
3966 scan->flags = (U8)*minnextp;
3971 if (f & SCF_DO_STCLASS_AND) {
3972 const int was = (data->start_class->flags & ANYOF_EOS);
3974 cl_and(data->start_class, &intrnl);
3976 data->start_class->flags |= ANYOF_EOS;
3979 if (data_fake.flags & (SF_HAS_PAR|SF_IN_PAR))
3981 if (data_fake.flags & SF_HAS_EVAL)
3982 data->flags |= SF_HAS_EVAL;
3983 data->whilem_c = data_fake.whilem_c;
3984 if ((flags & SCF_DO_SUBSTR) && data_fake.last_found) {
3985 if (RExC_rx->minlen<*minnextp)
3986 RExC_rx->minlen=*minnextp;
3987 SCAN_COMMIT(pRExC_state, &data_fake, minnextp);
3988 SvREFCNT_dec(data_fake.last_found);
3990 if ( data_fake.minlen_fixed != minlenp )
3992 data->offset_fixed= data_fake.offset_fixed;
3993 data->minlen_fixed= data_fake.minlen_fixed;
3994 data->lookbehind_fixed+= scan->flags;
3996 if ( data_fake.minlen_float != minlenp )
3998 data->minlen_float= data_fake.minlen_float;
3999 data->offset_float_min=data_fake.offset_float_min;
4000 data->offset_float_max=data_fake.offset_float_max;
4001 data->lookbehind_float+= scan->flags;
4010 else if (OP(scan) == OPEN) {
4011 if (stopparen != (I32)ARG(scan))
4014 else if (OP(scan) == CLOSE) {
4015 if (stopparen == (I32)ARG(scan)) {
4018 if ((I32)ARG(scan) == is_par) {
4019 next = regnext(scan);
4021 if ( next && (OP(next) != WHILEM) && next < last)
4022 is_par = 0; /* Disable optimization */
4025 *(data->last_closep) = ARG(scan);
4027 else if (OP(scan) == EVAL) {
4029 data->flags |= SF_HAS_EVAL;
4031 else if ( PL_regkind[OP(scan)] == ENDLIKE ) {
4032 if (flags & SCF_DO_SUBSTR) {
4033 SCAN_COMMIT(pRExC_state,data,minlenp);
4034 flags &= ~SCF_DO_SUBSTR;
4036 if (data && OP(scan)==ACCEPT) {
4037 data->flags |= SCF_SEEN_ACCEPT;
4042 else if (OP(scan) == LOGICAL && scan->flags == 2) /* Embedded follows */
4044 if (flags & SCF_DO_SUBSTR) {
4045 SCAN_COMMIT(pRExC_state,data,minlenp);
4046 data->longest = &(data->longest_float);
4048 is_inf = is_inf_internal = 1;
4049 if (flags & SCF_DO_STCLASS_OR) /* Allow everything */
4050 cl_anything(pRExC_state, data->start_class);
4051 flags &= ~SCF_DO_STCLASS;
4053 else if (OP(scan) == GPOS) {
4054 if (!(RExC_rx->extflags & RXf_GPOS_FLOAT) &&
4055 !(delta || is_inf || (data && data->pos_delta)))
4057 if (!(RExC_rx->extflags & RXf_ANCH) && (flags & SCF_DO_SUBSTR))
4058 RExC_rx->extflags |= RXf_ANCH_GPOS;
4059 if (RExC_rx->gofs < (U32)min)
4060 RExC_rx->gofs = min;
4062 RExC_rx->extflags |= RXf_GPOS_FLOAT;
4066 #ifdef TRIE_STUDY_OPT
4067 #ifdef FULL_TRIE_STUDY
4068 else if (PL_regkind[OP(scan)] == TRIE) {
4069 /* NOTE - There is similar code to this block above for handling
4070 BRANCH nodes on the initial study. If you change stuff here
4072 regnode *trie_node= scan;
4073 regnode *tail= regnext(scan);
4074 reg_trie_data *trie = (reg_trie_data*)RExC_rxi->data->data[ ARG(scan) ];
4075 I32 max1 = 0, min1 = I32_MAX;
4076 struct regnode_charclass_class accum;
4078 if (flags & SCF_DO_SUBSTR) /* XXXX Add !SUSPEND? */
4079 SCAN_COMMIT(pRExC_state, data,minlenp); /* Cannot merge strings after this. */
4080 if (flags & SCF_DO_STCLASS)
4081 cl_init_zero(pRExC_state, &accum);
4087 const regnode *nextbranch= NULL;
4090 for ( word=1 ; word <= trie->wordcount ; word++)
4092 I32 deltanext=0, minnext=0, f = 0, fake;
4093 struct regnode_charclass_class this_class;
4095 data_fake.flags = 0;
4097 data_fake.whilem_c = data->whilem_c;
4098 data_fake.last_closep = data->last_closep;
4101 data_fake.last_closep = &fake;
4102 data_fake.pos_delta = delta;
4103 if (flags & SCF_DO_STCLASS) {
4104 cl_init(pRExC_state, &this_class);
4105 data_fake.start_class = &this_class;
4106 f = SCF_DO_STCLASS_AND;
4108 if (flags & SCF_WHILEM_VISITED_POS)
4109 f |= SCF_WHILEM_VISITED_POS;
4111 if (trie->jump[word]) {
4113 nextbranch = trie_node + trie->jump[0];
4114 scan= trie_node + trie->jump[word];
4115 /* We go from the jump point to the branch that follows
4116 it. Note this means we need the vestigal unused branches
4117 even though they arent otherwise used.
4119 minnext = study_chunk(pRExC_state, &scan, minlenp,
4120 &deltanext, (regnode *)nextbranch, &data_fake,
4121 stopparen, recursed, NULL, f,depth+1);
4123 if (nextbranch && PL_regkind[OP(nextbranch)]==BRANCH)
4124 nextbranch= regnext((regnode*)nextbranch);
4126 if (min1 > (I32)(minnext + trie->minlen))
4127 min1 = minnext + trie->minlen;
4128 if (max1 < (I32)(minnext + deltanext + trie->maxlen))
4129 max1 = minnext + deltanext + trie->maxlen;
4130 if (deltanext == I32_MAX)
4131 is_inf = is_inf_internal = 1;
4133 if (data_fake.flags & (SF_HAS_PAR|SF_IN_PAR))
4135 if (data_fake.flags & SCF_SEEN_ACCEPT) {
4136 if ( stopmin > min + min1)
4137 stopmin = min + min1;
4138 flags &= ~SCF_DO_SUBSTR;
4140 data->flags |= SCF_SEEN_ACCEPT;
4143 if (data_fake.flags & SF_HAS_EVAL)
4144 data->flags |= SF_HAS_EVAL;
4145 data->whilem_c = data_fake.whilem_c;
4147 if (flags & SCF_DO_STCLASS)
4148 cl_or(pRExC_state, &accum, &this_class);
4151 if (flags & SCF_DO_SUBSTR) {
4152 data->pos_min += min1;
4153 data->pos_delta += max1 - min1;
4154 if (max1 != min1 || is_inf)
4155 data->longest = &(data->longest_float);
4158 delta += max1 - min1;
4159 if (flags & SCF_DO_STCLASS_OR) {
4160 cl_or(pRExC_state, data->start_class, &accum);
4162 cl_and(data->start_class, and_withp);
4163 flags &= ~SCF_DO_STCLASS;
4166 else if (flags & SCF_DO_STCLASS_AND) {
4168 cl_and(data->start_class, &accum);
4169 flags &= ~SCF_DO_STCLASS;
4172 /* Switch to OR mode: cache the old value of
4173 * data->start_class */
4175 StructCopy(data->start_class, and_withp,
4176 struct regnode_charclass_class);
4177 flags &= ~SCF_DO_STCLASS_AND;
4178 StructCopy(&accum, data->start_class,
4179 struct regnode_charclass_class);
4180 flags |= SCF_DO_STCLASS_OR;
4181 data->start_class->flags |= ANYOF_EOS;
4188 else if (PL_regkind[OP(scan)] == TRIE) {
4189 reg_trie_data *trie = (reg_trie_data*)RExC_rxi->data->data[ ARG(scan) ];
4192 min += trie->minlen;
4193 delta += (trie->maxlen - trie->minlen);
4194 flags &= ~SCF_DO_STCLASS; /* xxx */
4195 if (flags & SCF_DO_SUBSTR) {
4196 SCAN_COMMIT(pRExC_state,data,minlenp); /* Cannot expect anything... */
4197 data->pos_min += trie->minlen;
4198 data->pos_delta += (trie->maxlen - trie->minlen);
4199 if (trie->maxlen != trie->minlen)
4200 data->longest = &(data->longest_float);
4202 if (trie->jump) /* no more substrings -- for now /grr*/
4203 flags &= ~SCF_DO_SUBSTR;
4205 #endif /* old or new */
4206 #endif /* TRIE_STUDY_OPT */
4208 /* Else: zero-length, ignore. */
4209 scan = regnext(scan);
4214 stopparen = frame->stop;
4215 frame = frame->prev;
4216 goto fake_study_recurse;
4221 DEBUG_STUDYDATA("pre-fin:",data,depth);
4224 *deltap = is_inf_internal ? I32_MAX : delta;
4225 if (flags & SCF_DO_SUBSTR && is_inf)
4226 data->pos_delta = I32_MAX - data->pos_min;
4227 if (is_par > (I32)U8_MAX)
4229 if (is_par && pars==1 && data) {
4230 data->flags |= SF_IN_PAR;
4231 data->flags &= ~SF_HAS_PAR;
4233 else if (pars && data) {
4234 data->flags |= SF_HAS_PAR;
4235 data->flags &= ~SF_IN_PAR;
4237 if (flags & SCF_DO_STCLASS_OR)
4238 cl_and(data->start_class, and_withp);
4239 if (flags & SCF_TRIE_RESTUDY)
4240 data->flags |= SCF_TRIE_RESTUDY;
4242 DEBUG_STUDYDATA("post-fin:",data,depth);
4244 return min < stopmin ? min : stopmin;
4248 S_add_data(RExC_state_t *pRExC_state, U32 n, const char *s)
4250 U32 count = RExC_rxi->data ? RExC_rxi->data->count : 0;
4252 PERL_ARGS_ASSERT_ADD_DATA;
4254 Renewc(RExC_rxi->data,
4255 sizeof(*RExC_rxi->data) + sizeof(void*) * (count + n - 1),
4256 char, struct reg_data);
4258 Renew(RExC_rxi->data->what, count + n, U8);
4260 Newx(RExC_rxi->data->what, n, U8);
4261 RExC_rxi->data->count = count + n;
4262 Copy(s, RExC_rxi->data->what + count, n, U8);
4266 /*XXX: todo make this not included in a non debugging perl */
4267 #ifndef PERL_IN_XSUB_RE
4269 Perl_reginitcolors(pTHX)
4272 const char * const s = PerlEnv_getenv("PERL_RE_COLORS");
4274 char *t = savepv(s);
4278 t = strchr(t, '\t');
4284 PL_colors[i] = t = (char *)"";
4289 PL_colors[i++] = (char *)"";
4296 #ifdef TRIE_STUDY_OPT
4297 #define CHECK_RESTUDY_GOTO \
4299 (data.flags & SCF_TRIE_RESTUDY) \
4303 #define CHECK_RESTUDY_GOTO
4307 - pregcomp - compile a regular expression into internal code
4309 * We can't allocate space until we know how big the compiled form will be,
4310 * but we can't compile it (and thus know how big it is) until we've got a
4311 * place to put the code. So we cheat: we compile it twice, once with code
4312 * generation turned off and size counting turned on, and once "for real".
4313 * This also means that we don't allocate space until we are sure that the
4314 * thing really will compile successfully, and we never have to move the
4315 * code and thus invalidate pointers into it. (Note that it has to be in
4316 * one piece because free() must be able to free it all.) [NB: not true in perl]
4318 * Beware that the optimization-preparation code in here knows about some
4319 * of the structure of the compiled regexp. [I'll say.]
4324 #ifndef PERL_IN_XSUB_RE
4325 #define RE_ENGINE_PTR &reh_regexp_engine
4327 extern const struct regexp_engine my_reg_engine;
4328 #define RE_ENGINE_PTR &my_reg_engine
4331 #ifndef PERL_IN_XSUB_RE
4333 Perl_pregcomp(pTHX_ SV * const pattern, const U32 flags)
4336 HV * const table = GvHV(PL_hintgv);
4338 PERL_ARGS_ASSERT_PREGCOMP;
4340 /* Dispatch a request to compile a regexp to correct
4343 SV **ptr= hv_fetchs(table, "regcomp", FALSE);
4344 GET_RE_DEBUG_FLAGS_DECL;
4345 if (ptr && SvIOK(*ptr) && SvIV(*ptr)) {
4346 const regexp_engine *eng=INT2PTR(regexp_engine*,SvIV(*ptr));
4348 PerlIO_printf(Perl_debug_log, "Using engine %"UVxf"\n",
4351 return CALLREGCOMP_ENG(eng, pattern, flags);
4354 return Perl_re_compile(aTHX_ pattern, flags);
4359 Perl_re_compile(pTHX_ SV * const pattern, U32 orig_pm_flags)
4364 register regexp_internal *ri;
4373 /* these are all flags - maybe they should be turned
4374 * into a single int with different bit masks */
4375 I32 sawlookahead = 0;
4378 bool used_setjump = FALSE;
4383 RExC_state_t RExC_state;
4384 RExC_state_t * const pRExC_state = &RExC_state;
4385 #ifdef TRIE_STUDY_OPT
4387 RExC_state_t copyRExC_state;
4389 GET_RE_DEBUG_FLAGS_DECL;
4391 PERL_ARGS_ASSERT_RE_COMPILE;
4393 DEBUG_r(if (!PL_colorset) reginitcolors());
4395 RExC_utf8 = RExC_orig_utf8 = SvUTF8(pattern);
4397 /****************** LONG JUMP TARGET HERE***********************/
4398 /* Longjmp back to here if have to switch in midstream to utf8 */
4399 if (! RExC_orig_utf8) {
4400 JMPENV_PUSH(jump_ret);
4401 used_setjump = TRUE;
4404 if (jump_ret == 0) { /* First time through */
4405 exp = SvPV(pattern, plen);
4407 /* ignore the utf8ness if the pattern is 0 length */
4409 RExC_utf8 = RExC_orig_utf8 = 0;
4413 SV *dsv= sv_newmortal();
4414 RE_PV_QUOTED_DECL(s, RExC_utf8,
4415 dsv, exp, plen, 60);
4416 PerlIO_printf(Perl_debug_log, "%sCompiling REx%s %s\n",
4417 PL_colors[4],PL_colors[5],s);
4420 else { /* longjumped back */
4423 /* If the cause for the longjmp was other than changing to utf8, pop
4424 * our own setjmp, and longjmp to the correct handler */
4425 if (jump_ret != UTF8_LONGJMP) {
4427 JMPENV_JUMP(jump_ret);
4432 /* It's possible to write a regexp in ascii that represents Unicode
4433 codepoints outside of the byte range, such as via \x{100}. If we
4434 detect such a sequence we have to convert the entire pattern to utf8
4435 and then recompile, as our sizing calculation will have been based
4436 on 1 byte == 1 character, but we will need to use utf8 to encode
4437 at least some part of the pattern, and therefore must convert the whole
4440 DEBUG_PARSE_r(PerlIO_printf(Perl_debug_log,
4441 "UTF8 mismatch! Converting to utf8 for resizing and compile\n"));
4442 exp = (char*)Perl_bytes_to_utf8(aTHX_ (U8*)SvPV(pattern, plen), &len);
4444 RExC_orig_utf8 = RExC_utf8 = 1;
4448 #ifdef TRIE_STUDY_OPT
4452 /* Set to use unicode semantics if the pattern is in utf8 and has the
4453 * 'depends' charset specified, as it means unicode when utf8 */
4454 pm_flags = orig_pm_flags;
4456 if (RExC_utf8 && get_regex_charset(pm_flags) == REGEX_DEPENDS_CHARSET) {
4457 set_regex_charset(&pm_flags, REGEX_UNICODE_CHARSET);
4461 RExC_flags = pm_flags;
4465 RExC_in_lookbehind = 0;
4466 RExC_seen_zerolen = *exp == '^' ? -1 : 0;
4467 RExC_seen_evals = 0;
4470 /* First pass: determine size, legality. */
4478 RExC_emit = &PL_regdummy;
4479 RExC_whilem_seen = 0;
4480 RExC_open_parens = NULL;
4481 RExC_close_parens = NULL;
4483 RExC_paren_names = NULL;
4485 RExC_paren_name_list = NULL;
4487 RExC_recurse = NULL;
4488 RExC_recurse_count = 0;
4490 #if 0 /* REGC() is (currently) a NOP at the first pass.
4491 * Clever compilers notice this and complain. --jhi */
4492 REGC((U8)REG_MAGIC, (char*)RExC_emit);
4494 DEBUG_PARSE_r(PerlIO_printf(Perl_debug_log, "Starting first pass (sizing)\n"));
4495 if (reg(pRExC_state, 0, &flags,1) == NULL) {
4496 RExC_precomp = NULL;
4500 /* Here, finished first pass. Get rid of any added setjmp */
4505 PerlIO_printf(Perl_debug_log,
4506 "Required size %"IVdf" nodes\n"
4507 "Starting second pass (creation)\n",
4510 RExC_lastparse=NULL;
4512 /* Small enough for pointer-storage convention?
4513 If extralen==0, this means that we will not need long jumps. */
4514 if (RExC_size >= 0x10000L && RExC_extralen)
4515 RExC_size += RExC_extralen;
4518 if (RExC_whilem_seen > 15)
4519 RExC_whilem_seen = 15;
4521 /* Allocate space and zero-initialize. Note, the two step process
4522 of zeroing when in debug mode, thus anything assigned has to
4523 happen after that */
4524 rx = (REGEXP*) newSV_type(SVt_REGEXP);
4525 r = (struct regexp*)SvANY(rx);
4526 Newxc(ri, sizeof(regexp_internal) + (unsigned)RExC_size * sizeof(regnode),
4527 char, regexp_internal);
4528 if ( r == NULL || ri == NULL )
4529 FAIL("Regexp out of space");
4531 /* avoid reading uninitialized memory in DEBUGGING code in study_chunk() */
4532 Zero(ri, sizeof(regexp_internal) + (unsigned)RExC_size * sizeof(regnode), char);
4534 /* bulk initialize base fields with 0. */
4535 Zero(ri, sizeof(regexp_internal), char);
4538 /* non-zero initialization begins here */
4540 r->engine= RE_ENGINE_PTR;
4541 r->extflags = pm_flags;
4543 bool has_p = ((r->extflags & RXf_PMf_KEEPCOPY) == RXf_PMf_KEEPCOPY);
4544 bool has_charset = (get_regex_charset(r->extflags) != REGEX_DEPENDS_CHARSET);
4546 /* The caret is output if there are any defaults: if not all the STD
4547 * flags are set, or if no character set specifier is needed */
4549 (((r->extflags & RXf_PMf_STD_PMMOD) != RXf_PMf_STD_PMMOD)
4551 bool has_runon = ((RExC_seen & REG_SEEN_RUN_ON_COMMENT)==REG_SEEN_RUN_ON_COMMENT);
4552 U16 reganch = (U16)((r->extflags & RXf_PMf_STD_PMMOD)
4553 >> RXf_PMf_STD_PMMOD_SHIFT);
4554 const char *fptr = STD_PAT_MODS; /*"msix"*/
4556 /* Allocate for the worst case, which is all the std flags are turned
4557 * on. If more precision is desired, we could do a population count of
4558 * the flags set. This could be done with a small lookup table, or by
4559 * shifting, masking and adding, or even, when available, assembly
4560 * language for a machine-language population count.
4561 * We never output a minus, as all those are defaults, so are
4562 * covered by the caret */
4563 const STRLEN wraplen = plen + has_p + has_runon
4564 + has_default /* If needs a caret */
4566 /* If needs a character set specifier */
4567 + ((has_charset) ? MAX_CHARSET_NAME_LENGTH : 0)
4568 + (sizeof(STD_PAT_MODS) - 1)
4569 + (sizeof("(?:)") - 1);
4571 p = sv_grow(MUTABLE_SV(rx), wraplen + 1); /* +1 for the ending NUL */
4573 SvFLAGS(rx) |= SvUTF8(pattern);
4576 /* If a default, cover it using the caret */
4578 *p++= DEFAULT_PAT_MOD;
4582 const char* const name = get_regex_charset_name(r->extflags, &len);
4583 Copy(name, p, len, char);
4587 *p++ = KEEPCOPY_PAT_MOD; /*'p'*/
4590 while((ch = *fptr++)) {
4598 Copy(RExC_precomp, p, plen, char);
4599 assert ((RX_WRAPPED(rx) - p) < 16);
4600 r->pre_prefix = p - RX_WRAPPED(rx);
4606 SvCUR_set(rx, p - SvPVX_const(rx));
4610 r->nparens = RExC_npar - 1; /* set early to validate backrefs */
4612 if (RExC_seen & REG_SEEN_RECURSE) {
4613 Newxz(RExC_open_parens, RExC_npar,regnode *);
4614 SAVEFREEPV(RExC_open_parens);
4615 Newxz(RExC_close_parens,RExC_npar,regnode *);
4616 SAVEFREEPV(RExC_close_parens);
4619 /* Useful during FAIL. */
4620 #ifdef RE_TRACK_PATTERN_OFFSETS
4621 Newxz(ri->u.offsets, 2*RExC_size+1, U32); /* MJD 20001228 */
4622 DEBUG_OFFSETS_r(PerlIO_printf(Perl_debug_log,
4623 "%s %"UVuf" bytes for offset annotations.\n",
4624 ri->u.offsets ? "Got" : "Couldn't get",
4625 (UV)((2*RExC_size+1) * sizeof(U32))));
4627 SetProgLen(ri,RExC_size);
4631 REH_CALL_COMP_BEGIN_HOOK(pRExC_state->rx);
4633 /* Second pass: emit code. */
4634 RExC_flags = pm_flags; /* don't let top level (?i) bleed */
4639 RExC_emit_start = ri->program;
4640 RExC_emit = ri->program;
4641 RExC_emit_bound = ri->program + RExC_size + 1;
4643 /* Store the count of eval-groups for security checks: */
4644 RExC_rx->seen_evals = RExC_seen_evals;
4645 REGC((U8)REG_MAGIC, (char*) RExC_emit++);
4646 if (reg(pRExC_state, 0, &flags,1) == NULL) {
4650 /* XXXX To minimize changes to RE engine we always allocate
4651 3-units-long substrs field. */
4652 Newx(r->substrs, 1, struct reg_substr_data);
4653 if (RExC_recurse_count) {
4654 Newxz(RExC_recurse,RExC_recurse_count,regnode *);
4655 SAVEFREEPV(RExC_recurse);
4659 r->minlen = minlen = sawlookahead = sawplus = sawopen = 0;
4660 Zero(r->substrs, 1, struct reg_substr_data);
4662 #ifdef TRIE_STUDY_OPT
4664 StructCopy(&zero_scan_data, &data, scan_data_t);
4665 copyRExC_state = RExC_state;
4668 DEBUG_OPTIMISE_r(PerlIO_printf(Perl_debug_log,"Restudying\n"));
4670 RExC_state = copyRExC_state;
4671 if (seen & REG_TOP_LEVEL_BRANCHES)
4672 RExC_seen |= REG_TOP_LEVEL_BRANCHES;
4674 RExC_seen &= ~REG_TOP_LEVEL_BRANCHES;
4675 if (data.last_found) {
4676 SvREFCNT_dec(data.longest_fixed);
4677 SvREFCNT_dec(data.longest_float);
4678 SvREFCNT_dec(data.last_found);
4680 StructCopy(&zero_scan_data, &data, scan_data_t);
4683 StructCopy(&zero_scan_data, &data, scan_data_t);
4686 /* Dig out information for optimizations. */
4687 r->extflags = RExC_flags; /* was pm_op */
4688 /*dmq: removed as part of de-PMOP: pm->op_pmflags = RExC_flags; */
4691 SvUTF8_on(rx); /* Unicode in it? */
4692 ri->regstclass = NULL;
4693 if (RExC_naughty >= 10) /* Probably an expensive pattern. */
4694 r->intflags |= PREGf_NAUGHTY;
4695 scan = ri->program + 1; /* First BRANCH. */
4697 /* testing for BRANCH here tells us whether there is "must appear"
4698 data in the pattern. If there is then we can use it for optimisations */
4699 if (!(RExC_seen & REG_TOP_LEVEL_BRANCHES)) { /* Only one top-level choice. */
4701 STRLEN longest_float_length, longest_fixed_length;
4702 struct regnode_charclass_class ch_class; /* pointed to by data */
4704 I32 last_close = 0; /* pointed to by data */
4705 regnode *first= scan;
4706 regnode *first_next= regnext(first);
4708 * Skip introductions and multiplicators >= 1
4709 * so that we can extract the 'meat' of the pattern that must
4710 * match in the large if() sequence following.
4711 * NOTE that EXACT is NOT covered here, as it is normally
4712 * picked up by the optimiser separately.
4714 * This is unfortunate as the optimiser isnt handling lookahead
4715 * properly currently.
4718 while ((OP(first) == OPEN && (sawopen = 1)) ||
4719 /* An OR of *one* alternative - should not happen now. */
4720 (OP(first) == BRANCH && OP(first_next) != BRANCH) ||
4721 /* for now we can't handle lookbehind IFMATCH*/
4722 (OP(first) == IFMATCH && !first->flags && (sawlookahead = 1)) ||
4723 (OP(first) == PLUS) ||
4724 (OP(first) == MINMOD) ||
4725 /* An {n,m} with n>0 */
4726 (PL_regkind[OP(first)] == CURLY && ARG1(first) > 0) ||
4727 (OP(first) == NOTHING && PL_regkind[OP(first_next)] != END ))
4730 * the only op that could be a regnode is PLUS, all the rest
4731 * will be regnode_1 or regnode_2.
4734 if (OP(first) == PLUS)
4737 first += regarglen[OP(first)];
4739 first = NEXTOPER(first);
4740 first_next= regnext(first);
4743 /* Starting-point info. */
4745 DEBUG_PEEP("first:",first,0);
4746 /* Ignore EXACT as we deal with it later. */
4747 if (PL_regkind[OP(first)] == EXACT) {
4748 if (OP(first) == EXACT)
4749 NOOP; /* Empty, get anchored substr later. */
4751 ri->regstclass = first;
4754 else if (PL_regkind[OP(first)] == TRIE &&
4755 ((reg_trie_data *)ri->data->data[ ARG(first) ])->minlen>0)
4758 /* this can happen only on restudy */
4759 if ( OP(first) == TRIE ) {
4760 struct regnode_1 *trieop = (struct regnode_1 *)
4761 PerlMemShared_calloc(1, sizeof(struct regnode_1));
4762 StructCopy(first,trieop,struct regnode_1);
4763 trie_op=(regnode *)trieop;
4765 struct regnode_charclass *trieop = (struct regnode_charclass *)
4766 PerlMemShared_calloc(1, sizeof(struct regnode_charclass));
4767 StructCopy(first,trieop,struct regnode_charclass);
4768 trie_op=(regnode *)trieop;
4771 make_trie_failtable(pRExC_state, (regnode *)first, trie_op, 0);
4772 ri->regstclass = trie_op;
4775 else if (REGNODE_SIMPLE(OP(first)))
4776 ri->regstclass = first;
4777 else if (PL_regkind[OP(first)] == BOUND ||
4778 PL_regkind[OP(first)] == NBOUND)
4779 ri->regstclass = first;
4780 else if (PL_regkind[OP(first)] == BOL) {
4781 r->extflags |= (OP(first) == MBOL
4783 : (OP(first) == SBOL
4786 first = NEXTOPER(first);
4789 else if (OP(first) == GPOS) {
4790 r->extflags |= RXf_ANCH_GPOS;
4791 first = NEXTOPER(first);
4794 else if ((!sawopen || !RExC_sawback) &&
4795 (OP(first) == STAR &&
4796 PL_regkind[OP(NEXTOPER(first))] == REG_ANY) &&
4797 !(r->extflags & RXf_ANCH) && !(RExC_seen & REG_SEEN_EVAL))
4799 /* turn .* into ^.* with an implied $*=1 */
4801 (OP(NEXTOPER(first)) == REG_ANY)
4804 r->extflags |= type;
4805 r->intflags |= PREGf_IMPLICIT;
4806 first = NEXTOPER(first);
4809 if (sawplus && !sawlookahead && (!sawopen || !RExC_sawback)
4810 && !(RExC_seen & REG_SEEN_EVAL)) /* May examine pos and $& */
4811 /* x+ must match at the 1st pos of run of x's */
4812 r->intflags |= PREGf_SKIP;
4814 /* Scan is after the zeroth branch, first is atomic matcher. */
4815 #ifdef TRIE_STUDY_OPT
4818 PerlIO_printf(Perl_debug_log, "first at %"IVdf"\n",
4819 (IV)(first - scan + 1))
4823 PerlIO_printf(Perl_debug_log, "first at %"IVdf"\n",
4824 (IV)(first - scan + 1))
4830 * If there's something expensive in the r.e., find the
4831 * longest literal string that must appear and make it the
4832 * regmust. Resolve ties in favor of later strings, since
4833 * the regstart check works with the beginning of the r.e.
4834 * and avoiding duplication strengthens checking. Not a
4835 * strong reason, but sufficient in the absence of others.
4836 * [Now we resolve ties in favor of the earlier string if
4837 * it happens that c_offset_min has been invalidated, since the
4838 * earlier string may buy us something the later one won't.]
4841 data.longest_fixed = newSVpvs("");
4842 data.longest_float = newSVpvs("");
4843 data.last_found = newSVpvs("");
4844 data.longest = &(data.longest_fixed);
4846 if (!ri->regstclass) {
4847 cl_init(pRExC_state, &ch_class);
4848 data.start_class = &ch_class;
4849 stclass_flag = SCF_DO_STCLASS_AND;
4850 } else /* XXXX Check for BOUND? */
4852 data.last_closep = &last_close;
4854 minlen = study_chunk(pRExC_state, &first, &minlen, &fake, scan + RExC_size, /* Up to end */
4855 &data, -1, NULL, NULL,
4856 SCF_DO_SUBSTR | SCF_WHILEM_VISITED_POS | stclass_flag,0);
4862 if ( RExC_npar == 1 && data.longest == &(data.longest_fixed)
4863 && data.last_start_min == 0 && data.last_end > 0
4864 && !RExC_seen_zerolen
4865 && !(RExC_seen & REG_SEEN_VERBARG)
4866 && (!(RExC_seen & REG_SEEN_GPOS) || (r->extflags & RXf_ANCH_GPOS)))
4867 r->extflags |= RXf_CHECK_ALL;
4868 scan_commit(pRExC_state, &data,&minlen,0);
4869 SvREFCNT_dec(data.last_found);
4871 /* Note that code very similar to this but for anchored string
4872 follows immediately below, changes may need to be made to both.
4875 longest_float_length = CHR_SVLEN(data.longest_float);
4876 if (longest_float_length
4877 || (data.flags & SF_FL_BEFORE_EOL
4878 && (!(data.flags & SF_FL_BEFORE_MEOL)
4879 || (RExC_flags & RXf_PMf_MULTILINE))))
4883 if (SvCUR(data.longest_fixed) /* ok to leave SvCUR */
4884 && data.offset_fixed == data.offset_float_min
4885 && SvCUR(data.longest_fixed) == SvCUR(data.longest_float))
4886 goto remove_float; /* As in (a)+. */
4888 /* copy the information about the longest float from the reg_scan_data
4889 over to the program. */
4890 if (SvUTF8(data.longest_float)) {
4891 r->float_utf8 = data.longest_float;
4892 r->float_substr = NULL;
4894 r->float_substr = data.longest_float;
4895 r->float_utf8 = NULL;
4897 /* float_end_shift is how many chars that must be matched that
4898 follow this item. We calculate it ahead of time as once the
4899 lookbehind offset is added in we lose the ability to correctly
4901 ml = data.minlen_float ? *(data.minlen_float)
4902 : (I32)longest_float_length;
4903 r->float_end_shift = ml - data.offset_float_min
4904 - longest_float_length + (SvTAIL(data.longest_float) != 0)
4905 + data.lookbehind_float;
4906 r->float_min_offset = data.offset_float_min - data.lookbehind_float;
4907 r->float_max_offset = data.offset_float_max;
4908 if (data.offset_float_max < I32_MAX) /* Don't offset infinity */
4909 r->float_max_offset -= data.lookbehind_float;
4911 t = (data.flags & SF_FL_BEFORE_EOL /* Can't have SEOL and MULTI */
4912 && (!(data.flags & SF_FL_BEFORE_MEOL)
4913 || (RExC_flags & RXf_PMf_MULTILINE)));
4914 fbm_compile(data.longest_float, t ? FBMcf_TAIL : 0);
4918 r->float_substr = r->float_utf8 = NULL;
4919 SvREFCNT_dec(data.longest_float);
4920 longest_float_length = 0;
4923 /* Note that code very similar to this but for floating string
4924 is immediately above, changes may need to be made to both.
4927 longest_fixed_length = CHR_SVLEN(data.longest_fixed);
4928 if (longest_fixed_length
4929 || (data.flags & SF_FIX_BEFORE_EOL /* Cannot have SEOL and MULTI */
4930 && (!(data.flags & SF_FIX_BEFORE_MEOL)
4931 || (RExC_flags & RXf_PMf_MULTILINE))))
4935 /* copy the information about the longest fixed
4936 from the reg_scan_data over to the program. */
4937 if (SvUTF8(data.longest_fixed)) {
4938 r->anchored_utf8 = data.longest_fixed;
4939 r->anchored_substr = NULL;
4941 r->anchored_substr = data.longest_fixed;
4942 r->anchored_utf8 = NULL;
4944 /* fixed_end_shift is how many chars that must be matched that
4945 follow this item. We calculate it ahead of time as once the
4946 lookbehind offset is added in we lose the ability to correctly
4948 ml = data.minlen_fixed ? *(data.minlen_fixed)
4949 : (I32)longest_fixed_length;
4950 r->anchored_end_shift = ml - data.offset_fixed
4951 - longest_fixed_length + (SvTAIL(data.longest_fixed) != 0)
4952 + data.lookbehind_fixed;
4953 r->anchored_offset = data.offset_fixed - data.lookbehind_fixed;
4955 t = (data.flags & SF_FIX_BEFORE_EOL /* Can't have SEOL and MULTI */
4956 && (!(data.flags & SF_FIX_BEFORE_MEOL)
4957 || (RExC_flags & RXf_PMf_MULTILINE)));
4958 fbm_compile(data.longest_fixed, t ? FBMcf_TAIL : 0);
4961 r->anchored_substr = r->anchored_utf8 = NULL;
4962 SvREFCNT_dec(data.longest_fixed);
4963 longest_fixed_length = 0;
4966 && (OP(ri->regstclass) == REG_ANY || OP(ri->regstclass) == SANY))
4967 ri->regstclass = NULL;
4969 /* If the synthetic start class were to ever be used when EOS is set,
4970 * that bit would have to be cleared, as it is shared with another */
4971 if ((!(r->anchored_substr || r->anchored_utf8) || r->anchored_offset)
4973 && !(data.start_class->flags & ANYOF_EOS)
4974 && !cl_is_anything(data.start_class))
4976 const U32 n = add_data(pRExC_state, 1, "f");
4978 Newx(RExC_rxi->data->data[n], 1,
4979 struct regnode_charclass_class);
4980 StructCopy(data.start_class,
4981 (struct regnode_charclass_class*)RExC_rxi->data->data[n],
4982 struct regnode_charclass_class);
4983 ri->regstclass = (regnode*)RExC_rxi->data->data[n];
4984 r->intflags &= ~PREGf_SKIP; /* Used in find_byclass(). */
4985 DEBUG_COMPILE_r({ SV *sv = sv_newmortal();
4986 regprop(r, sv, (regnode*)data.start_class);
4987 PerlIO_printf(Perl_debug_log,
4988 "synthetic stclass \"%s\".\n",
4989 SvPVX_const(sv));});
4992 /* A temporary algorithm prefers floated substr to fixed one to dig more info. */
4993 if (longest_fixed_length > longest_float_length) {
4994 r->check_end_shift = r->anchored_end_shift;
4995 r->check_substr = r->anchored_substr;
4996 r->check_utf8 = r->anchored_utf8;
4997 r->check_offset_min = r->check_offset_max = r->anchored_offset;
4998 if (r->extflags & RXf_ANCH_SINGLE)
4999 r->extflags |= RXf_NOSCAN;
5002 r->check_end_shift = r->float_end_shift;
5003 r->check_substr = r->float_substr;
5004 r->check_utf8 = r->float_utf8;
5005 r->check_offset_min = r->float_min_offset;
5006 r->check_offset_max = r->float_max_offset;
5008 /* XXXX Currently intuiting is not compatible with ANCH_GPOS.
5009 This should be changed ASAP! */
5010 if ((r->check_substr || r->check_utf8) && !(r->extflags & RXf_ANCH_GPOS)) {
5011 r->extflags |= RXf_USE_INTUIT;
5012 if (SvTAIL(r->check_substr ? r->check_substr : r->check_utf8))
5013 r->extflags |= RXf_INTUIT_TAIL;
5015 /* XXX Unneeded? dmq (shouldn't as this is handled elsewhere)
5016 if ( (STRLEN)minlen < longest_float_length )
5017 minlen= longest_float_length;
5018 if ( (STRLEN)minlen < longest_fixed_length )
5019 minlen= longest_fixed_length;
5023 /* Several toplevels. Best we can is to set minlen. */
5025 struct regnode_charclass_class ch_class;
5028 DEBUG_PARSE_r(PerlIO_printf(Perl_debug_log, "\nMulti Top Level\n"));
5030 scan = ri->program + 1;
5031 cl_init(pRExC_state, &ch_class);
5032 data.start_class = &ch_class;
5033 data.last_closep = &last_close;
5036 minlen = study_chunk(pRExC_state, &scan, &minlen, &fake, scan + RExC_size,
5037 &data, -1, NULL, NULL, SCF_DO_STCLASS_AND|SCF_WHILEM_VISITED_POS,0);
5041 r->check_substr = r->check_utf8 = r->anchored_substr = r->anchored_utf8
5042 = r->float_substr = r->float_utf8 = NULL;
5044 /* If the synthetic start class were to ever be used when EOS is set,
5045 * that bit would have to be cleared, as it is shared with another */
5046 if (!(data.start_class->flags & ANYOF_EOS)
5047 && !cl_is_anything(data.start_class))
5049 const U32 n = add_data(pRExC_state, 1, "f");
5051 Newx(RExC_rxi->data->data[n], 1,
5052 struct regnode_charclass_class);
5053 StructCopy(data.start_class,
5054 (struct regnode_charclass_class*)RExC_rxi->data->data[n],
5055 struct regnode_charclass_class);
5056 ri->regstclass = (regnode*)RExC_rxi->data->data[n];
5057 r->intflags &= ~PREGf_SKIP; /* Used in find_byclass(). */
5058 DEBUG_COMPILE_r({ SV* sv = sv_newmortal();
5059 regprop(r, sv, (regnode*)data.start_class);
5060 PerlIO_printf(Perl_debug_log,
5061 "synthetic stclass \"%s\".\n",
5062 SvPVX_const(sv));});
5066 /* Guard against an embedded (?=) or (?<=) with a longer minlen than
5067 the "real" pattern. */
5069 PerlIO_printf(Perl_debug_log,"minlen: %"IVdf" r->minlen:%"IVdf"\n",
5070 (IV)minlen, (IV)r->minlen);
5072 r->minlenret = minlen;
5073 if (r->minlen < minlen)
5076 if (RExC_seen & REG_SEEN_GPOS)
5077 r->extflags |= RXf_GPOS_SEEN;
5078 if (RExC_seen & REG_SEEN_LOOKBEHIND)
5079 r->extflags |= RXf_LOOKBEHIND_SEEN;
5080 if (RExC_seen & REG_SEEN_EVAL)
5081 r->extflags |= RXf_EVAL_SEEN;
5082 if (RExC_seen & REG_SEEN_CANY)
5083 r->extflags |= RXf_CANY_SEEN;
5084 if (RExC_seen & REG_SEEN_VERBARG)
5085 r->intflags |= PREGf_VERBARG_SEEN;
5086 if (RExC_seen & REG_SEEN_CUTGROUP)
5087 r->intflags |= PREGf_CUTGROUP_SEEN;
5088 if (RExC_paren_names)
5089 RXp_PAREN_NAMES(r) = MUTABLE_HV(SvREFCNT_inc(RExC_paren_names));
5091 RXp_PAREN_NAMES(r) = NULL;
5093 #ifdef STUPID_PATTERN_CHECKS
5094 if (RX_PRELEN(rx) == 0)
5095 r->extflags |= RXf_NULL;
5096 if (r->extflags & RXf_SPLIT && RX_PRELEN(rx) == 1 && RX_PRECOMP(rx)[0] == ' ')
5097 /* XXX: this should happen BEFORE we compile */
5098 r->extflags |= (RXf_SKIPWHITE|RXf_WHITE);
5099 else if (RX_PRELEN(rx) == 3 && memEQ("\\s+", RX_PRECOMP(rx), 3))
5100 r->extflags |= RXf_WHITE;
5101 else if (RX_PRELEN(rx) == 1 && RXp_PRECOMP(rx)[0] == '^')
5102 r->extflags |= RXf_START_ONLY;
5104 if (r->extflags & RXf_SPLIT && RX_PRELEN(rx) == 1 && RX_PRECOMP(rx)[0] == ' ')
5105 /* XXX: this should happen BEFORE we compile */
5106 r->extflags |= (RXf_SKIPWHITE|RXf_WHITE);
5108 regnode *first = ri->program + 1;
5110 U8 nop = OP(NEXTOPER(first));
5112 if (PL_regkind[fop] == NOTHING && nop == END)
5113 r->extflags |= RXf_NULL;
5114 else if (PL_regkind[fop] == BOL && nop == END)
5115 r->extflags |= RXf_START_ONLY;
5116 else if (fop == PLUS && nop ==SPACE && OP(regnext(first))==END)
5117 r->extflags |= RXf_WHITE;
5121 if (RExC_paren_names) {
5122 ri->name_list_idx = add_data( pRExC_state, 1, "a" );
5123 ri->data->data[ri->name_list_idx] = (void*)SvREFCNT_inc(RExC_paren_name_list);
5126 ri->name_list_idx = 0;
5128 if (RExC_recurse_count) {
5129 for ( ; RExC_recurse_count ; RExC_recurse_count-- ) {
5130 const regnode *scan = RExC_recurse[RExC_recurse_count-1];
5131 ARG2L_SET( scan, RExC_open_parens[ARG(scan)-1] - scan );
5134 Newxz(r->offs, RExC_npar, regexp_paren_pair);
5135 /* assume we don't need to swap parens around before we match */
5138 PerlIO_printf(Perl_debug_log,"Final program:\n");
5141 #ifdef RE_TRACK_PATTERN_OFFSETS
5142 DEBUG_OFFSETS_r(if (ri->u.offsets) {
5143 const U32 len = ri->u.offsets[0];
5145 GET_RE_DEBUG_FLAGS_DECL;
5146 PerlIO_printf(Perl_debug_log, "Offsets: [%"UVuf"]\n\t", (UV)ri->u.offsets[0]);
5147 for (i = 1; i <= len; i++) {
5148 if (ri->u.offsets[i*2-1] || ri->u.offsets[i*2])
5149 PerlIO_printf(Perl_debug_log, "%"UVuf":%"UVuf"[%"UVuf"] ",
5150 (UV)i, (UV)ri->u.offsets[i*2-1], (UV)ri->u.offsets[i*2]);
5152 PerlIO_printf(Perl_debug_log, "\n");
5158 #undef RE_ENGINE_PTR
5162 Perl_reg_named_buff(pTHX_ REGEXP * const rx, SV * const key, SV * const value,
5165 PERL_ARGS_ASSERT_REG_NAMED_BUFF;
5167 PERL_UNUSED_ARG(value);
5169 if (flags & RXapif_FETCH) {
5170 return reg_named_buff_fetch(rx, key, flags);
5171 } else if (flags & (RXapif_STORE | RXapif_DELETE | RXapif_CLEAR)) {
5172 Perl_croak_no_modify(aTHX);
5174 } else if (flags & RXapif_EXISTS) {
5175 return reg_named_buff_exists(rx, key, flags)
5178 } else if (flags & RXapif_REGNAMES) {
5179 return reg_named_buff_all(rx, flags);
5180 } else if (flags & (RXapif_SCALAR | RXapif_REGNAMES_COUNT)) {
5181 return reg_named_buff_scalar(rx, flags);
5183 Perl_croak(aTHX_ "panic: Unknown flags %d in named_buff", (int)flags);
5189 Perl_reg_named_buff_iter(pTHX_ REGEXP * const rx, const SV * const lastkey,
5192 PERL_ARGS_ASSERT_REG_NAMED_BUFF_ITER;
5193 PERL_UNUSED_ARG(lastkey);
5195 if (flags & RXapif_FIRSTKEY)
5196 return reg_named_buff_firstkey(rx, flags);
5197 else if (flags & RXapif_NEXTKEY)
5198 return reg_named_buff_nextkey(rx, flags);
5200 Perl_croak(aTHX_ "panic: Unknown flags %d in named_buff_iter", (int)flags);
5206 Perl_reg_named_buff_fetch(pTHX_ REGEXP * const r, SV * const namesv,
5209 AV *retarray = NULL;
5211 struct regexp *const rx = (struct regexp *)SvANY(r);
5213 PERL_ARGS_ASSERT_REG_NAMED_BUFF_FETCH;
5215 if (flags & RXapif_ALL)
5218 if (rx && RXp_PAREN_NAMES(rx)) {
5219 HE *he_str = hv_fetch_ent( RXp_PAREN_NAMES(rx), namesv, 0, 0 );
5222 SV* sv_dat=HeVAL(he_str);
5223 I32 *nums=(I32*)SvPVX(sv_dat);
5224 for ( i=0; i<SvIVX(sv_dat); i++ ) {
5225 if ((I32)(rx->nparens) >= nums[i]
5226 && rx->offs[nums[i]].start != -1
5227 && rx->offs[nums[i]].end != -1)
5230 CALLREG_NUMBUF_FETCH(r,nums[i],ret);
5234 ret = newSVsv(&PL_sv_undef);
5237 av_push(retarray, ret);
5240 return newRV_noinc(MUTABLE_SV(retarray));
5247 Perl_reg_named_buff_exists(pTHX_ REGEXP * const r, SV * const key,
5250 struct regexp *const rx = (struct regexp *)SvANY(r);
5252 PERL_ARGS_ASSERT_REG_NAMED_BUFF_EXISTS;
5254 if (rx && RXp_PAREN_NAMES(rx)) {
5255 if (flags & RXapif_ALL) {
5256 return hv_exists_ent(RXp_PAREN_NAMES(rx), key, 0);
5258 SV *sv = CALLREG_NAMED_BUFF_FETCH(r, key, flags);
5272 Perl_reg_named_buff_firstkey(pTHX_ REGEXP * const r, const U32 flags)
5274 struct regexp *const rx = (struct regexp *)SvANY(r);
5276 PERL_ARGS_ASSERT_REG_NAMED_BUFF_FIRSTKEY;
5278 if ( rx && RXp_PAREN_NAMES(rx) ) {
5279 (void)hv_iterinit(RXp_PAREN_NAMES(rx));
5281 return CALLREG_NAMED_BUFF_NEXTKEY(r, NULL, flags & ~RXapif_FIRSTKEY);
5288 Perl_reg_named_buff_nextkey(pTHX_ REGEXP * const r, const U32 flags)
5290 struct regexp *const rx = (struct regexp *)SvANY(r);
5291 GET_RE_DEBUG_FLAGS_DECL;
5293 PERL_ARGS_ASSERT_REG_NAMED_BUFF_NEXTKEY;
5295 if (rx && RXp_PAREN_NAMES(rx)) {
5296 HV *hv = RXp_PAREN_NAMES(rx);
5298 while ( (temphe = hv_iternext_flags(hv,0)) ) {
5301 SV* sv_dat = HeVAL(temphe);
5302 I32 *nums = (I32*)SvPVX(sv_dat);
5303 for ( i = 0; i < SvIVX(sv_dat); i++ ) {
5304 if ((I32)(rx->lastparen) >= nums[i] &&
5305 rx->offs[nums[i]].start != -1 &&
5306 rx->offs[nums[i]].end != -1)
5312 if (parno || flags & RXapif_ALL) {
5313 return newSVhek(HeKEY_hek(temphe));
5321 Perl_reg_named_buff_scalar(pTHX_ REGEXP * const r, const U32 flags)
5326 struct regexp *const rx = (struct regexp *)SvANY(r);
5328 PERL_ARGS_ASSERT_REG_NAMED_BUFF_SCALAR;
5330 if (rx && RXp_PAREN_NAMES(rx)) {
5331 if (flags & (RXapif_ALL | RXapif_REGNAMES_COUNT)) {
5332 return newSViv(HvTOTALKEYS(RXp_PAREN_NAMES(rx)));
5333 } else if (flags & RXapif_ONE) {
5334 ret = CALLREG_NAMED_BUFF_ALL(r, (flags | RXapif_REGNAMES));
5335 av = MUTABLE_AV(SvRV(ret));
5336 length = av_len(av);
5338 return newSViv(length + 1);
5340 Perl_croak(aTHX_ "panic: Unknown flags %d in named_buff_scalar", (int)flags);
5344 return &PL_sv_undef;
5348 Perl_reg_named_buff_all(pTHX_ REGEXP * const r, const U32 flags)
5350 struct regexp *const rx = (struct regexp *)SvANY(r);
5353 PERL_ARGS_ASSERT_REG_NAMED_BUFF_ALL;
5355 if (rx && RXp_PAREN_NAMES(rx)) {
5356 HV *hv= RXp_PAREN_NAMES(rx);
5358 (void)hv_iterinit(hv);
5359 while ( (temphe = hv_iternext_flags(hv,0)) ) {
5362 SV* sv_dat = HeVAL(temphe);
5363 I32 *nums = (I32*)SvPVX(sv_dat);
5364 for ( i = 0; i < SvIVX(sv_dat); i++ ) {
5365 if ((I32)(rx->lastparen) >= nums[i] &&
5366 rx->offs[nums[i]].start != -1 &&
5367 rx->offs[nums[i]].end != -1)
5373 if (parno || flags & RXapif_ALL) {
5374 av_push(av, newSVhek(HeKEY_hek(temphe)));
5379 return newRV_noinc(MUTABLE_SV(av));
5383 Perl_reg_numbered_buff_fetch(pTHX_ REGEXP * const r, const I32 paren,
5386 struct regexp *const rx = (struct regexp *)SvANY(r);
5391 PERL_ARGS_ASSERT_REG_NUMBERED_BUFF_FETCH;
5394 sv_setsv(sv,&PL_sv_undef);
5398 if (paren == RX_BUFF_IDX_PREMATCH && rx->offs[0].start != -1) {
5400 i = rx->offs[0].start;
5404 if (paren == RX_BUFF_IDX_POSTMATCH && rx->offs[0].end != -1) {
5406 s = rx->subbeg + rx->offs[0].end;
5407 i = rx->sublen - rx->offs[0].end;
5410 if ( 0 <= paren && paren <= (I32)rx->nparens &&
5411 (s1 = rx->offs[paren].start) != -1 &&
5412 (t1 = rx->offs[paren].end) != -1)
5416 s = rx->subbeg + s1;
5418 sv_setsv(sv,&PL_sv_undef);
5421 assert(rx->sublen >= (s - rx->subbeg) + i );
5423 const int oldtainted = PL_tainted;
5425 sv_setpvn(sv, s, i);
5426 PL_tainted = oldtainted;
5427 if ( (rx->extflags & RXf_CANY_SEEN)
5428 ? (RXp_MATCH_UTF8(rx)
5429 && (!i || is_utf8_string((U8*)s, i)))
5430 : (RXp_MATCH_UTF8(rx)) )
5437 if (RXp_MATCH_TAINTED(rx)) {
5438 if (SvTYPE(sv) >= SVt_PVMG) {
5439 MAGIC* const mg = SvMAGIC(sv);
5442 SvMAGIC_set(sv, mg->mg_moremagic);
5444 if ((mgt = SvMAGIC(sv))) {
5445 mg->mg_moremagic = mgt;
5446 SvMAGIC_set(sv, mg);
5456 sv_setsv(sv,&PL_sv_undef);
5462 Perl_reg_numbered_buff_store(pTHX_ REGEXP * const rx, const I32 paren,
5463 SV const * const value)
5465 PERL_ARGS_ASSERT_REG_NUMBERED_BUFF_STORE;
5467 PERL_UNUSED_ARG(rx);
5468 PERL_UNUSED_ARG(paren);
5469 PERL_UNUSED_ARG(value);
5472 Perl_croak_no_modify(aTHX);
5476 Perl_reg_numbered_buff_length(pTHX_ REGEXP * const r, const SV * const sv,
5479 struct regexp *const rx = (struct regexp *)SvANY(r);
5483 PERL_ARGS_ASSERT_REG_NUMBERED_BUFF_LENGTH;
5485 /* Some of this code was originally in C<Perl_magic_len> in F<mg.c> */
5487 /* $` / ${^PREMATCH} */
5488 case RX_BUFF_IDX_PREMATCH:
5489 if (rx->offs[0].start != -1) {
5490 i = rx->offs[0].start;
5498 /* $' / ${^POSTMATCH} */
5499 case RX_BUFF_IDX_POSTMATCH:
5500 if (rx->offs[0].end != -1) {
5501 i = rx->sublen - rx->offs[0].end;
5503 s1 = rx->offs[0].end;
5509 /* $& / ${^MATCH}, $1, $2, ... */
5511 if (paren <= (I32)rx->nparens &&
5512 (s1 = rx->offs[paren].start) != -1 &&
5513 (t1 = rx->offs[paren].end) != -1)
5518 if (ckWARN(WARN_UNINITIALIZED))
5519 report_uninit((const SV *)sv);
5524 if (i > 0 && RXp_MATCH_UTF8(rx)) {
5525 const char * const s = rx->subbeg + s1;
5530 if (is_utf8_string_loclen((U8*)s, i, &ep, &el))
5537 Perl_reg_qr_package(pTHX_ REGEXP * const rx)
5539 PERL_ARGS_ASSERT_REG_QR_PACKAGE;
5540 PERL_UNUSED_ARG(rx);
5544 return newSVpvs("Regexp");
5547 /* Scans the name of a named buffer from the pattern.
5548 * If flags is REG_RSN_RETURN_NULL returns null.
5549 * If flags is REG_RSN_RETURN_NAME returns an SV* containing the name
5550 * If flags is REG_RSN_RETURN_DATA returns the data SV* corresponding
5551 * to the parsed name as looked up in the RExC_paren_names hash.
5552 * If there is an error throws a vFAIL().. type exception.
5555 #define REG_RSN_RETURN_NULL 0
5556 #define REG_RSN_RETURN_NAME 1
5557 #define REG_RSN_RETURN_DATA 2
5560 S_reg_scan_name(pTHX_ RExC_state_t *pRExC_state, U32 flags)
5562 char *name_start = RExC_parse;
5564 PERL_ARGS_ASSERT_REG_SCAN_NAME;
5566 if (isIDFIRST_lazy_if(RExC_parse, UTF)) {
5567 /* skip IDFIRST by using do...while */
5570 RExC_parse += UTF8SKIP(RExC_parse);
5571 } while (isALNUM_utf8((U8*)RExC_parse));
5575 } while (isALNUM(*RExC_parse));
5580 = newSVpvn_flags(name_start, (int)(RExC_parse - name_start),
5581 SVs_TEMP | (UTF ? SVf_UTF8 : 0));
5582 if ( flags == REG_RSN_RETURN_NAME)
5584 else if (flags==REG_RSN_RETURN_DATA) {
5587 if ( ! sv_name ) /* should not happen*/
5588 Perl_croak(aTHX_ "panic: no svname in reg_scan_name");
5589 if (RExC_paren_names)
5590 he_str = hv_fetch_ent( RExC_paren_names, sv_name, 0, 0 );
5592 sv_dat = HeVAL(he_str);
5594 vFAIL("Reference to nonexistent named group");
5598 Perl_croak(aTHX_ "panic: bad flag in reg_scan_name");
5605 #define DEBUG_PARSE_MSG(funcname) DEBUG_PARSE_r({ \
5606 int rem=(int)(RExC_end - RExC_parse); \
5615 if (RExC_lastparse!=RExC_parse) \
5616 PerlIO_printf(Perl_debug_log," >%.*s%-*s", \
5619 iscut ? "..." : "<" \
5622 PerlIO_printf(Perl_debug_log,"%16s",""); \
5625 num = RExC_size + 1; \
5627 num=REG_NODE_NUM(RExC_emit); \
5628 if (RExC_lastnum!=num) \
5629 PerlIO_printf(Perl_debug_log,"|%4d",num); \
5631 PerlIO_printf(Perl_debug_log,"|%4s",""); \
5632 PerlIO_printf(Perl_debug_log,"|%*s%-4s", \
5633 (int)((depth*2)), "", \
5637 RExC_lastparse=RExC_parse; \
5642 #define DEBUG_PARSE(funcname) DEBUG_PARSE_r({ \
5643 DEBUG_PARSE_MSG((funcname)); \
5644 PerlIO_printf(Perl_debug_log,"%4s","\n"); \
5646 #define DEBUG_PARSE_FMT(funcname,fmt,args) DEBUG_PARSE_r({ \
5647 DEBUG_PARSE_MSG((funcname)); \
5648 PerlIO_printf(Perl_debug_log,fmt "\n",args); \
5651 - reg - regular expression, i.e. main body or parenthesized thing
5653 * Caller must absorb opening parenthesis.
5655 * Combining parenthesis handling with the base level of regular expression
5656 * is a trifle forced, but the need to tie the tails of the branches to what
5657 * follows makes it hard to avoid.
5659 #define REGTAIL(x,y,z) regtail((x),(y),(z),depth+1)
5661 #define REGTAIL_STUDY(x,y,z) regtail_study((x),(y),(z),depth+1)
5663 #define REGTAIL_STUDY(x,y,z) regtail((x),(y),(z),depth+1)
5667 S_reg(pTHX_ RExC_state_t *pRExC_state, I32 paren, I32 *flagp,U32 depth)
5668 /* paren: Parenthesized? 0=top, 1=(, inside: changed to letter. */
5671 register regnode *ret; /* Will be the head of the group. */
5672 register regnode *br;
5673 register regnode *lastbr;
5674 register regnode *ender = NULL;
5675 register I32 parno = 0;
5677 U32 oregflags = RExC_flags;
5678 bool have_branch = 0;
5680 I32 freeze_paren = 0;
5681 I32 after_freeze = 0;
5683 /* for (?g), (?gc), and (?o) warnings; warning
5684 about (?c) will warn about (?g) -- japhy */
5686 #define WASTED_O 0x01
5687 #define WASTED_G 0x02
5688 #define WASTED_C 0x04
5689 #define WASTED_GC (0x02|0x04)
5690 I32 wastedflags = 0x00;
5692 char * parse_start = RExC_parse; /* MJD */
5693 char * const oregcomp_parse = RExC_parse;
5695 GET_RE_DEBUG_FLAGS_DECL;
5697 PERL_ARGS_ASSERT_REG;
5698 DEBUG_PARSE("reg ");
5700 *flagp = 0; /* Tentatively. */
5703 /* Make an OPEN node, if parenthesized. */
5705 if ( *RExC_parse == '*') { /* (*VERB:ARG) */
5706 char *start_verb = RExC_parse;
5707 STRLEN verb_len = 0;
5708 char *start_arg = NULL;
5709 unsigned char op = 0;
5711 int internal_argval = 0; /* internal_argval is only useful if !argok */
5712 while ( *RExC_parse && *RExC_parse != ')' ) {
5713 if ( *RExC_parse == ':' ) {
5714 start_arg = RExC_parse + 1;
5720 verb_len = RExC_parse - start_verb;
5723 while ( *RExC_parse && *RExC_parse != ')' )
5725 if ( *RExC_parse != ')' )
5726 vFAIL("Unterminated verb pattern argument");
5727 if ( RExC_parse == start_arg )
5730 if ( *RExC_parse != ')' )
5731 vFAIL("Unterminated verb pattern");
5734 switch ( *start_verb ) {
5735 case 'A': /* (*ACCEPT) */
5736 if ( memEQs(start_verb,verb_len,"ACCEPT") ) {
5738 internal_argval = RExC_nestroot;
5741 case 'C': /* (*COMMIT) */
5742 if ( memEQs(start_verb,verb_len,"COMMIT") )
5745 case 'F': /* (*FAIL) */
5746 if ( verb_len==1 || memEQs(start_verb,verb_len,"FAIL") ) {
5751 case ':': /* (*:NAME) */
5752 case 'M': /* (*MARK:NAME) */
5753 if ( verb_len==0 || memEQs(start_verb,verb_len,"MARK") ) {
5758 case 'P': /* (*PRUNE) */
5759 if ( memEQs(start_verb,verb_len,"PRUNE") )
5762 case 'S': /* (*SKIP) */
5763 if ( memEQs(start_verb,verb_len,"SKIP") )
5766 case 'T': /* (*THEN) */
5767 /* [19:06] <TimToady> :: is then */
5768 if ( memEQs(start_verb,verb_len,"THEN") ) {
5770 RExC_seen |= REG_SEEN_CUTGROUP;
5776 vFAIL3("Unknown verb pattern '%.*s'",
5777 verb_len, start_verb);
5780 if ( start_arg && internal_argval ) {
5781 vFAIL3("Verb pattern '%.*s' may not have an argument",
5782 verb_len, start_verb);
5783 } else if ( argok < 0 && !start_arg ) {
5784 vFAIL3("Verb pattern '%.*s' has a mandatory argument",
5785 verb_len, start_verb);
5787 ret = reganode(pRExC_state, op, internal_argval);
5788 if ( ! internal_argval && ! SIZE_ONLY ) {
5790 SV *sv = newSVpvn( start_arg, RExC_parse - start_arg);
5791 ARG(ret) = add_data( pRExC_state, 1, "S" );
5792 RExC_rxi->data->data[ARG(ret)]=(void*)sv;
5799 if (!internal_argval)
5800 RExC_seen |= REG_SEEN_VERBARG;
5801 } else if ( start_arg ) {
5802 vFAIL3("Verb pattern '%.*s' may not have an argument",
5803 verb_len, start_verb);
5805 ret = reg_node(pRExC_state, op);
5807 nextchar(pRExC_state);
5810 if (*RExC_parse == '?') { /* (?...) */
5811 bool is_logical = 0;
5812 const char * const seqstart = RExC_parse;
5813 bool has_use_defaults = FALSE;
5816 paren = *RExC_parse++;
5817 ret = NULL; /* For look-ahead/behind. */
5820 case 'P': /* (?P...) variants for those used to PCRE/Python */
5821 paren = *RExC_parse++;
5822 if ( paren == '<') /* (?P<...>) named capture */
5824 else if (paren == '>') { /* (?P>name) named recursion */
5825 goto named_recursion;
5827 else if (paren == '=') { /* (?P=...) named backref */
5828 /* this pretty much dupes the code for \k<NAME> in regatom(), if
5829 you change this make sure you change that */
5830 char* name_start = RExC_parse;
5832 SV *sv_dat = reg_scan_name(pRExC_state,
5833 SIZE_ONLY ? REG_RSN_RETURN_NULL : REG_RSN_RETURN_DATA);
5834 if (RExC_parse == name_start || *RExC_parse != ')')
5835 vFAIL2("Sequence %.3s... not terminated",parse_start);
5838 num = add_data( pRExC_state, 1, "S" );
5839 RExC_rxi->data->data[num]=(void*)sv_dat;
5840 SvREFCNT_inc_simple_void(sv_dat);
5843 ret = reganode(pRExC_state,
5854 Set_Node_Offset(ret, parse_start+1);
5855 Set_Node_Cur_Length(ret); /* MJD */
5857 nextchar(pRExC_state);
5861 vFAIL3("Sequence (%.*s...) not recognized", RExC_parse-seqstart, seqstart);
5863 case '<': /* (?<...) */
5864 if (*RExC_parse == '!')
5866 else if (*RExC_parse != '=')
5872 case '\'': /* (?'...') */
5873 name_start= RExC_parse;
5874 svname = reg_scan_name(pRExC_state,
5875 SIZE_ONLY ? /* reverse test from the others */
5876 REG_RSN_RETURN_NAME :
5877 REG_RSN_RETURN_NULL);
5878 if (RExC_parse == name_start) {
5880 vFAIL3("Sequence (%.*s...) not recognized", RExC_parse-seqstart, seqstart);
5883 if (*RExC_parse != paren)
5884 vFAIL2("Sequence (?%c... not terminated",
5885 paren=='>' ? '<' : paren);
5889 if (!svname) /* shouldn't happen */
5891 "panic: reg_scan_name returned NULL");
5892 if (!RExC_paren_names) {
5893 RExC_paren_names= newHV();
5894 sv_2mortal(MUTABLE_SV(RExC_paren_names));
5896 RExC_paren_name_list= newAV();
5897 sv_2mortal(MUTABLE_SV(RExC_paren_name_list));
5900 he_str = hv_fetch_ent( RExC_paren_names, svname, 1, 0 );
5902 sv_dat = HeVAL(he_str);
5904 /* croak baby croak */
5906 "panic: paren_name hash element allocation failed");
5907 } else if ( SvPOK(sv_dat) ) {
5908 /* (?|...) can mean we have dupes so scan to check
5909 its already been stored. Maybe a flag indicating
5910 we are inside such a construct would be useful,
5911 but the arrays are likely to be quite small, so
5912 for now we punt -- dmq */
5913 IV count = SvIV(sv_dat);
5914 I32 *pv = (I32*)SvPVX(sv_dat);
5916 for ( i = 0 ; i < count ; i++ ) {
5917 if ( pv[i] == RExC_npar ) {
5923 pv = (I32*)SvGROW(sv_dat, SvCUR(sv_dat) + sizeof(I32)+1);
5924 SvCUR_set(sv_dat, SvCUR(sv_dat) + sizeof(I32));
5925 pv[count] = RExC_npar;
5926 SvIV_set(sv_dat, SvIVX(sv_dat) + 1);
5929 (void)SvUPGRADE(sv_dat,SVt_PVNV);
5930 sv_setpvn(sv_dat, (char *)&(RExC_npar), sizeof(I32));
5932 SvIV_set(sv_dat, 1);
5935 if (!av_store(RExC_paren_name_list, RExC_npar, SvREFCNT_inc(svname)))
5936 SvREFCNT_dec(svname);
5939 /*sv_dump(sv_dat);*/
5941 nextchar(pRExC_state);
5943 goto capturing_parens;
5945 RExC_seen |= REG_SEEN_LOOKBEHIND;
5946 RExC_in_lookbehind++;
5948 case '=': /* (?=...) */
5949 RExC_seen_zerolen++;
5951 case '!': /* (?!...) */
5952 RExC_seen_zerolen++;
5953 if (*RExC_parse == ')') {
5954 ret=reg_node(pRExC_state, OPFAIL);
5955 nextchar(pRExC_state);
5959 case '|': /* (?|...) */
5960 /* branch reset, behave like a (?:...) except that
5961 buffers in alternations share the same numbers */
5963 after_freeze = freeze_paren = RExC_npar;
5965 case ':': /* (?:...) */
5966 case '>': /* (?>...) */
5968 case '$': /* (?$...) */
5969 case '@': /* (?@...) */
5970 vFAIL2("Sequence (?%c...) not implemented", (int)paren);
5972 case '#': /* (?#...) */
5973 while (*RExC_parse && *RExC_parse != ')')
5975 if (*RExC_parse != ')')
5976 FAIL("Sequence (?#... not terminated");
5977 nextchar(pRExC_state);
5980 case '0' : /* (?0) */
5981 case 'R' : /* (?R) */
5982 if (*RExC_parse != ')')
5983 FAIL("Sequence (?R) not terminated");
5984 ret = reg_node(pRExC_state, GOSTART);
5985 *flagp |= POSTPONED;
5986 nextchar(pRExC_state);
5989 { /* named and numeric backreferences */
5991 case '&': /* (?&NAME) */
5992 parse_start = RExC_parse - 1;
5995 SV *sv_dat = reg_scan_name(pRExC_state,
5996 SIZE_ONLY ? REG_RSN_RETURN_NULL : REG_RSN_RETURN_DATA);
5997 num = sv_dat ? *((I32 *)SvPVX(sv_dat)) : 0;
5999 goto gen_recurse_regop;
6002 if (!(RExC_parse[0] >= '1' && RExC_parse[0] <= '9')) {
6004 vFAIL("Illegal pattern");
6006 goto parse_recursion;
6008 case '-': /* (?-1) */
6009 if (!(RExC_parse[0] >= '1' && RExC_parse[0] <= '9')) {
6010 RExC_parse--; /* rewind to let it be handled later */
6014 case '1': case '2': case '3': case '4': /* (?1) */
6015 case '5': case '6': case '7': case '8': case '9':
6018 num = atoi(RExC_parse);
6019 parse_start = RExC_parse - 1; /* MJD */
6020 if (*RExC_parse == '-')
6022 while (isDIGIT(*RExC_parse))
6024 if (*RExC_parse!=')')
6025 vFAIL("Expecting close bracket");
6028 if ( paren == '-' ) {
6030 Diagram of capture buffer numbering.
6031 Top line is the normal capture buffer numbers
6032 Bottom line is the negative indexing as from
6036 /(a(x)y)(a(b(c(?-2)d)e)f)(g(h))/
6040 num = RExC_npar + num;
6043 vFAIL("Reference to nonexistent group");
6045 } else if ( paren == '+' ) {
6046 num = RExC_npar + num - 1;
6049 ret = reganode(pRExC_state, GOSUB, num);
6051 if (num > (I32)RExC_rx->nparens) {
6053 vFAIL("Reference to nonexistent group");
6055 ARG2L_SET( ret, RExC_recurse_count++);
6057 DEBUG_OPTIMISE_MORE_r(PerlIO_printf(Perl_debug_log,
6058 "Recurse #%"UVuf" to %"IVdf"\n", (UV)ARG(ret), (IV)ARG2L(ret)));
6062 RExC_seen |= REG_SEEN_RECURSE;
6063 Set_Node_Length(ret, 1 + regarglen[OP(ret)]); /* MJD */
6064 Set_Node_Offset(ret, parse_start); /* MJD */
6066 *flagp |= POSTPONED;
6067 nextchar(pRExC_state);
6069 } /* named and numeric backreferences */
6072 case '?': /* (??...) */
6074 if (*RExC_parse != '{') {
6076 vFAIL3("Sequence (%.*s...) not recognized", RExC_parse-seqstart, seqstart);
6079 *flagp |= POSTPONED;
6080 paren = *RExC_parse++;
6082 case '{': /* (?{...}) */
6087 char *s = RExC_parse;
6089 RExC_seen_zerolen++;
6090 RExC_seen |= REG_SEEN_EVAL;
6091 while (count && (c = *RExC_parse)) {
6102 if (*RExC_parse != ')') {
6104 vFAIL("Sequence (?{...}) not terminated or not {}-balanced");
6108 OP_4tree *sop, *rop;
6109 SV * const sv = newSVpvn(s, RExC_parse - 1 - s);
6112 Perl_save_re_context(aTHX);
6113 rop = Perl_sv_compile_2op_is_broken(aTHX_ sv, &sop, "re", &pad);
6114 sop->op_private |= OPpREFCOUNTED;
6115 /* re_dup will OpREFCNT_inc */
6116 OpREFCNT_set(sop, 1);
6119 n = add_data(pRExC_state, 3, "nop");
6120 RExC_rxi->data->data[n] = (void*)rop;
6121 RExC_rxi->data->data[n+1] = (void*)sop;
6122 RExC_rxi->data->data[n+2] = (void*)pad;
6125 else { /* First pass */
6126 if (PL_reginterp_cnt < ++RExC_seen_evals
6128 /* No compiled RE interpolated, has runtime
6129 components ===> unsafe. */
6130 FAIL("Eval-group not allowed at runtime, use re 'eval'");
6131 if (PL_tainting && PL_tainted)
6132 FAIL("Eval-group in insecure regular expression");
6133 #if PERL_VERSION > 8
6134 if (IN_PERL_COMPILETIME)
6139 nextchar(pRExC_state);
6141 ret = reg_node(pRExC_state, LOGICAL);
6144 REGTAIL(pRExC_state, ret, reganode(pRExC_state, EVAL, n));
6145 /* deal with the length of this later - MJD */
6148 ret = reganode(pRExC_state, EVAL, n);
6149 Set_Node_Length(ret, RExC_parse - parse_start + 1);
6150 Set_Node_Offset(ret, parse_start);
6153 case '(': /* (?(?{...})...) and (?(?=...)...) */
6156 if (RExC_parse[0] == '?') { /* (?(?...)) */
6157 if (RExC_parse[1] == '=' || RExC_parse[1] == '!'
6158 || RExC_parse[1] == '<'
6159 || RExC_parse[1] == '{') { /* Lookahead or eval. */
6162 ret = reg_node(pRExC_state, LOGICAL);
6165 REGTAIL(pRExC_state, ret, reg(pRExC_state, 1, &flag,depth+1));
6169 else if ( RExC_parse[0] == '<' /* (?(<NAME>)...) */
6170 || RExC_parse[0] == '\'' ) /* (?('NAME')...) */
6172 char ch = RExC_parse[0] == '<' ? '>' : '\'';
6173 char *name_start= RExC_parse++;
6175 SV *sv_dat=reg_scan_name(pRExC_state,
6176 SIZE_ONLY ? REG_RSN_RETURN_NULL : REG_RSN_RETURN_DATA);
6177 if (RExC_parse == name_start || *RExC_parse != ch)
6178 vFAIL2("Sequence (?(%c... not terminated",
6179 (ch == '>' ? '<' : ch));
6182 num = add_data( pRExC_state, 1, "S" );
6183 RExC_rxi->data->data[num]=(void*)sv_dat;
6184 SvREFCNT_inc_simple_void(sv_dat);
6186 ret = reganode(pRExC_state,NGROUPP,num);
6187 goto insert_if_check_paren;
6189 else if (RExC_parse[0] == 'D' &&
6190 RExC_parse[1] == 'E' &&
6191 RExC_parse[2] == 'F' &&
6192 RExC_parse[3] == 'I' &&
6193 RExC_parse[4] == 'N' &&
6194 RExC_parse[5] == 'E')
6196 ret = reganode(pRExC_state,DEFINEP,0);
6199 goto insert_if_check_paren;
6201 else if (RExC_parse[0] == 'R') {
6204 if (RExC_parse[0] >= '1' && RExC_parse[0] <= '9' ) {
6205 parno = atoi(RExC_parse++);
6206 while (isDIGIT(*RExC_parse))
6208 } else if (RExC_parse[0] == '&') {
6211 sv_dat = reg_scan_name(pRExC_state,
6212 SIZE_ONLY ? REG_RSN_RETURN_NULL : REG_RSN_RETURN_DATA);
6213 parno = sv_dat ? *((I32 *)SvPVX(sv_dat)) : 0;
6215 ret = reganode(pRExC_state,INSUBP,parno);
6216 goto insert_if_check_paren;
6218 else if (RExC_parse[0] >= '1' && RExC_parse[0] <= '9' ) {
6221 parno = atoi(RExC_parse++);
6223 while (isDIGIT(*RExC_parse))
6225 ret = reganode(pRExC_state, GROUPP, parno);
6227 insert_if_check_paren:
6228 if ((c = *nextchar(pRExC_state)) != ')')
6229 vFAIL("Switch condition not recognized");
6231 REGTAIL(pRExC_state, ret, reganode(pRExC_state, IFTHEN, 0));
6232 br = regbranch(pRExC_state, &flags, 1,depth+1);
6234 br = reganode(pRExC_state, LONGJMP, 0);
6236 REGTAIL(pRExC_state, br, reganode(pRExC_state, LONGJMP, 0));
6237 c = *nextchar(pRExC_state);
6242 vFAIL("(?(DEFINE)....) does not allow branches");
6243 lastbr = reganode(pRExC_state, IFTHEN, 0); /* Fake one for optimizer. */
6244 regbranch(pRExC_state, &flags, 1,depth+1);
6245 REGTAIL(pRExC_state, ret, lastbr);
6248 c = *nextchar(pRExC_state);
6253 vFAIL("Switch (?(condition)... contains too many branches");
6254 ender = reg_node(pRExC_state, TAIL);
6255 REGTAIL(pRExC_state, br, ender);
6257 REGTAIL(pRExC_state, lastbr, ender);
6258 REGTAIL(pRExC_state, NEXTOPER(NEXTOPER(lastbr)), ender);
6261 REGTAIL(pRExC_state, ret, ender);
6262 RExC_size++; /* XXX WHY do we need this?!!
6263 For large programs it seems to be required
6264 but I can't figure out why. -- dmq*/
6268 vFAIL2("Unknown switch condition (?(%.2s", RExC_parse);
6272 RExC_parse--; /* for vFAIL to print correctly */
6273 vFAIL("Sequence (? incomplete");
6275 case DEFAULT_PAT_MOD: /* Use default flags with the exceptions
6277 has_use_defaults = TRUE;
6278 STD_PMMOD_FLAGS_CLEAR(&RExC_flags);
6279 if (RExC_utf8) { /* But the default for a utf8 pattern is
6280 unicode semantics */
6281 set_regex_charset(&RExC_flags, REGEX_UNICODE_CHARSET);
6286 parse_flags: /* (?i) */
6288 U32 posflags = 0, negflags = 0;
6289 U32 *flagsp = &posflags;
6290 bool has_charset_modifier = 0;
6291 regex_charset cs = REGEX_DEPENDS_CHARSET;
6293 while (*RExC_parse) {
6294 /* && strchr("iogcmsx", *RExC_parse) */
6295 /* (?g), (?gc) and (?o) are useless here
6296 and must be globally applied -- japhy */
6297 switch (*RExC_parse) {
6298 CASE_STD_PMMOD_FLAGS_PARSE_SET(flagsp);
6299 case LOCALE_PAT_MOD:
6300 if (has_charset_modifier || flagsp == &negflags) {
6301 goto fail_modifiers;
6303 cs = REGEX_LOCALE_CHARSET;
6304 has_charset_modifier = 1;
6306 case UNICODE_PAT_MOD:
6307 if (has_charset_modifier || flagsp == &negflags) {
6308 goto fail_modifiers;
6310 cs = REGEX_UNICODE_CHARSET;
6311 has_charset_modifier = 1;
6313 case ASCII_RESTRICT_PAT_MOD:
6314 if (has_charset_modifier || flagsp == &negflags) {
6315 goto fail_modifiers;
6317 cs = REGEX_ASCII_RESTRICTED_CHARSET;
6318 has_charset_modifier = 1;
6320 case DEPENDS_PAT_MOD:
6321 if (has_use_defaults
6322 || has_charset_modifier
6323 || flagsp == &negflags)
6325 goto fail_modifiers;
6328 /* The dual charset means unicode semantics if the
6329 * pattern (or target, not known until runtime) are
6332 ? REGEX_UNICODE_CHARSET
6333 : REGEX_DEPENDS_CHARSET;
6334 has_charset_modifier = 1;
6336 case ONCE_PAT_MOD: /* 'o' */
6337 case GLOBAL_PAT_MOD: /* 'g' */
6338 if (SIZE_ONLY && ckWARN(WARN_REGEXP)) {
6339 const I32 wflagbit = *RExC_parse == 'o' ? WASTED_O : WASTED_G;
6340 if (! (wastedflags & wflagbit) ) {
6341 wastedflags |= wflagbit;
6344 "Useless (%s%c) - %suse /%c modifier",
6345 flagsp == &negflags ? "?-" : "?",
6347 flagsp == &negflags ? "don't " : "",
6354 case CONTINUE_PAT_MOD: /* 'c' */
6355 if (SIZE_ONLY && ckWARN(WARN_REGEXP)) {
6356 if (! (wastedflags & WASTED_C) ) {
6357 wastedflags |= WASTED_GC;
6360 "Useless (%sc) - %suse /gc modifier",
6361 flagsp == &negflags ? "?-" : "?",
6362 flagsp == &negflags ? "don't " : ""
6367 case KEEPCOPY_PAT_MOD: /* 'p' */
6368 if (flagsp == &negflags) {
6370 ckWARNreg(RExC_parse + 1,"Useless use of (?-p)");
6372 *flagsp |= RXf_PMf_KEEPCOPY;
6376 /* A flag is a default iff it is following a minus, so
6377 * if there is a minus, it means will be trying to
6378 * re-specify a default which is an error */
6379 if (has_use_defaults || flagsp == &negflags) {
6382 vFAIL3("Sequence (%.*s...) not recognized", RExC_parse-seqstart, seqstart);
6386 wastedflags = 0; /* reset so (?g-c) warns twice */
6392 RExC_flags |= posflags;
6393 RExC_flags &= ~negflags;
6394 set_regex_charset(&RExC_flags, cs);
6396 oregflags |= posflags;
6397 oregflags &= ~negflags;
6398 set_regex_charset(&oregflags, cs);
6400 nextchar(pRExC_state);
6411 vFAIL3("Sequence (%.*s...) not recognized", RExC_parse-seqstart, seqstart);
6416 }} /* one for the default block, one for the switch */
6423 ret = reganode(pRExC_state, OPEN, parno);
6426 RExC_nestroot = parno;
6427 if (RExC_seen & REG_SEEN_RECURSE
6428 && !RExC_open_parens[parno-1])
6430 DEBUG_OPTIMISE_MORE_r(PerlIO_printf(Perl_debug_log,
6431 "Setting open paren #%"IVdf" to %d\n",
6432 (IV)parno, REG_NODE_NUM(ret)));
6433 RExC_open_parens[parno-1]= ret;
6436 Set_Node_Length(ret, 1); /* MJD */
6437 Set_Node_Offset(ret, RExC_parse); /* MJD */
6445 /* Pick up the branches, linking them together. */
6446 parse_start = RExC_parse; /* MJD */
6447 br = regbranch(pRExC_state, &flags, 1,depth+1);
6450 if (RExC_npar > after_freeze)
6451 after_freeze = RExC_npar;
6452 RExC_npar = freeze_paren;
6455 /* branch_len = (paren != 0); */
6459 if (*RExC_parse == '|') {
6460 if (!SIZE_ONLY && RExC_extralen) {
6461 reginsert(pRExC_state, BRANCHJ, br, depth+1);
6464 reginsert(pRExC_state, BRANCH, br, depth+1);
6465 Set_Node_Length(br, paren != 0);
6466 Set_Node_Offset_To_R(br-RExC_emit_start, parse_start-RExC_start);
6470 RExC_extralen += 1; /* For BRANCHJ-BRANCH. */
6472 else if (paren == ':') {
6473 *flagp |= flags&SIMPLE;
6475 if (is_open) { /* Starts with OPEN. */
6476 REGTAIL(pRExC_state, ret, br); /* OPEN -> first. */
6478 else if (paren != '?') /* Not Conditional */
6480 *flagp |= flags & (SPSTART | HASWIDTH | POSTPONED);
6482 while (*RExC_parse == '|') {
6483 if (!SIZE_ONLY && RExC_extralen) {
6484 ender = reganode(pRExC_state, LONGJMP,0);
6485 REGTAIL(pRExC_state, NEXTOPER(NEXTOPER(lastbr)), ender); /* Append to the previous. */
6488 RExC_extralen += 2; /* Account for LONGJMP. */
6489 nextchar(pRExC_state);
6491 if (RExC_npar > after_freeze)
6492 after_freeze = RExC_npar;
6493 RExC_npar = freeze_paren;
6495 br = regbranch(pRExC_state, &flags, 0, depth+1);
6499 REGTAIL(pRExC_state, lastbr, br); /* BRANCH -> BRANCH. */
6501 *flagp |= flags & (SPSTART | HASWIDTH | POSTPONED);
6504 if (have_branch || paren != ':') {
6505 /* Make a closing node, and hook it on the end. */
6508 ender = reg_node(pRExC_state, TAIL);
6511 ender = reganode(pRExC_state, CLOSE, parno);
6512 if (!SIZE_ONLY && RExC_seen & REG_SEEN_RECURSE) {
6513 DEBUG_OPTIMISE_MORE_r(PerlIO_printf(Perl_debug_log,
6514 "Setting close paren #%"IVdf" to %d\n",
6515 (IV)parno, REG_NODE_NUM(ender)));
6516 RExC_close_parens[parno-1]= ender;
6517 if (RExC_nestroot == parno)
6520 Set_Node_Offset(ender,RExC_parse+1); /* MJD */
6521 Set_Node_Length(ender,1); /* MJD */
6527 *flagp &= ~HASWIDTH;
6530 ender = reg_node(pRExC_state, SUCCEED);
6533 ender = reg_node(pRExC_state, END);
6535 assert(!RExC_opend); /* there can only be one! */
6540 REGTAIL(pRExC_state, lastbr, ender);
6542 if (have_branch && !SIZE_ONLY) {
6544 RExC_seen |= REG_TOP_LEVEL_BRANCHES;
6546 /* Hook the tails of the branches to the closing node. */
6547 for (br = ret; br; br = regnext(br)) {
6548 const U8 op = PL_regkind[OP(br)];
6550 REGTAIL_STUDY(pRExC_state, NEXTOPER(br), ender);
6552 else if (op == BRANCHJ) {
6553 REGTAIL_STUDY(pRExC_state, NEXTOPER(NEXTOPER(br)), ender);
6561 static const char parens[] = "=!<,>";
6563 if (paren && (p = strchr(parens, paren))) {
6564 U8 node = ((p - parens) % 2) ? UNLESSM : IFMATCH;
6565 int flag = (p - parens) > 1;
6568 node = SUSPEND, flag = 0;
6569 reginsert(pRExC_state, node,ret, depth+1);
6570 Set_Node_Cur_Length(ret);
6571 Set_Node_Offset(ret, parse_start + 1);
6573 REGTAIL_STUDY(pRExC_state, ret, reg_node(pRExC_state, TAIL));
6577 /* Check for proper termination. */
6579 RExC_flags = oregflags;
6580 if (RExC_parse >= RExC_end || *nextchar(pRExC_state) != ')') {
6581 RExC_parse = oregcomp_parse;
6582 vFAIL("Unmatched (");
6585 else if (!paren && RExC_parse < RExC_end) {
6586 if (*RExC_parse == ')') {
6588 vFAIL("Unmatched )");
6591 FAIL("Junk on end of regexp"); /* "Can't happen". */
6595 if (RExC_in_lookbehind) {
6596 RExC_in_lookbehind--;
6599 RExC_npar = after_freeze;
6604 - regbranch - one alternative of an | operator
6606 * Implements the concatenation operator.
6609 S_regbranch(pTHX_ RExC_state_t *pRExC_state, I32 *flagp, I32 first, U32 depth)
6612 register regnode *ret;
6613 register regnode *chain = NULL;
6614 register regnode *latest;
6615 I32 flags = 0, c = 0;
6616 GET_RE_DEBUG_FLAGS_DECL;
6618 PERL_ARGS_ASSERT_REGBRANCH;
6620 DEBUG_PARSE("brnc");
6625 if (!SIZE_ONLY && RExC_extralen)
6626 ret = reganode(pRExC_state, BRANCHJ,0);
6628 ret = reg_node(pRExC_state, BRANCH);
6629 Set_Node_Length(ret, 1);
6633 if (!first && SIZE_ONLY)
6634 RExC_extralen += 1; /* BRANCHJ */
6636 *flagp = WORST; /* Tentatively. */
6639 nextchar(pRExC_state);
6640 while (RExC_parse < RExC_end && *RExC_parse != '|' && *RExC_parse != ')') {
6642 latest = regpiece(pRExC_state, &flags,depth+1);
6643 if (latest == NULL) {
6644 if (flags & TRYAGAIN)
6648 else if (ret == NULL)
6650 *flagp |= flags&(HASWIDTH|POSTPONED);
6651 if (chain == NULL) /* First piece. */
6652 *flagp |= flags&SPSTART;
6655 REGTAIL(pRExC_state, chain, latest);
6660 if (chain == NULL) { /* Loop ran zero times. */
6661 chain = reg_node(pRExC_state, NOTHING);
6666 *flagp |= flags&SIMPLE;
6673 - regpiece - something followed by possible [*+?]
6675 * Note that the branching code sequences used for ? and the general cases
6676 * of * and + are somewhat optimized: they use the same NOTHING node as
6677 * both the endmarker for their branch list and the body of the last branch.
6678 * It might seem that this node could be dispensed with entirely, but the
6679 * endmarker role is not redundant.
6682 S_regpiece(pTHX_ RExC_state_t *pRExC_state, I32 *flagp, U32 depth)
6685 register regnode *ret;
6687 register char *next;
6689 const char * const origparse = RExC_parse;
6691 I32 max = REG_INFTY;
6693 const char *maxpos = NULL;
6694 GET_RE_DEBUG_FLAGS_DECL;
6696 PERL_ARGS_ASSERT_REGPIECE;
6698 DEBUG_PARSE("piec");
6700 ret = regatom(pRExC_state, &flags,depth+1);
6702 if (flags & TRYAGAIN)
6709 if (op == '{' && regcurly(RExC_parse)) {
6711 parse_start = RExC_parse; /* MJD */
6712 next = RExC_parse + 1;
6713 while (isDIGIT(*next) || *next == ',') {
6722 if (*next == '}') { /* got one */
6726 min = atoi(RExC_parse);
6730 maxpos = RExC_parse;
6732 if (!max && *maxpos != '0')
6733 max = REG_INFTY; /* meaning "infinity" */
6734 else if (max >= REG_INFTY)
6735 vFAIL2("Quantifier in {,} bigger than %d", REG_INFTY - 1);
6737 nextchar(pRExC_state);
6740 if ((flags&SIMPLE)) {
6741 RExC_naughty += 2 + RExC_naughty / 2;
6742 reginsert(pRExC_state, CURLY, ret, depth+1);
6743 Set_Node_Offset(ret, parse_start+1); /* MJD */
6744 Set_Node_Cur_Length(ret);
6747 regnode * const w = reg_node(pRExC_state, WHILEM);
6750 REGTAIL(pRExC_state, ret, w);
6751 if (!SIZE_ONLY && RExC_extralen) {
6752 reginsert(pRExC_state, LONGJMP,ret, depth+1);
6753 reginsert(pRExC_state, NOTHING,ret, depth+1);
6754 NEXT_OFF(ret) = 3; /* Go over LONGJMP. */
6756 reginsert(pRExC_state, CURLYX,ret, depth+1);
6758 Set_Node_Offset(ret, parse_start+1);
6759 Set_Node_Length(ret,
6760 op == '{' ? (RExC_parse - parse_start) : 1);
6762 if (!SIZE_ONLY && RExC_extralen)
6763 NEXT_OFF(ret) = 3; /* Go over NOTHING to LONGJMP. */
6764 REGTAIL(pRExC_state, ret, reg_node(pRExC_state, NOTHING));
6766 RExC_whilem_seen++, RExC_extralen += 3;
6767 RExC_naughty += 4 + RExC_naughty; /* compound interest */
6776 vFAIL("Can't do {n,m} with n > m");
6778 ARG1_SET(ret, (U16)min);
6779 ARG2_SET(ret, (U16)max);
6791 #if 0 /* Now runtime fix should be reliable. */
6793 /* if this is reinstated, don't forget to put this back into perldiag:
6795 =item Regexp *+ operand could be empty at {#} in regex m/%s/
6797 (F) The part of the regexp subject to either the * or + quantifier
6798 could match an empty string. The {#} shows in the regular
6799 expression about where the problem was discovered.
6803 if (!(flags&HASWIDTH) && op != '?')
6804 vFAIL("Regexp *+ operand could be empty");
6807 parse_start = RExC_parse;
6808 nextchar(pRExC_state);
6810 *flagp = (op != '+') ? (WORST|SPSTART|HASWIDTH) : (WORST|HASWIDTH);
6812 if (op == '*' && (flags&SIMPLE)) {
6813 reginsert(pRExC_state, STAR, ret, depth+1);
6817 else if (op == '*') {
6821 else if (op == '+' && (flags&SIMPLE)) {
6822 reginsert(pRExC_state, PLUS, ret, depth+1);
6826 else if (op == '+') {
6830 else if (op == '?') {
6835 if (!SIZE_ONLY && !(flags&(HASWIDTH|POSTPONED)) && max > REG_INFTY/3) {
6836 ckWARN3reg(RExC_parse,
6837 "%.*s matches null string many times",
6838 (int)(RExC_parse >= origparse ? RExC_parse - origparse : 0),
6842 if (RExC_parse < RExC_end && *RExC_parse == '?') {
6843 nextchar(pRExC_state);
6844 reginsert(pRExC_state, MINMOD, ret, depth+1);
6845 REGTAIL(pRExC_state, ret, ret + NODE_STEP_REGNODE);
6847 #ifndef REG_ALLOW_MINMOD_SUSPEND
6850 if (RExC_parse < RExC_end && *RExC_parse == '+') {
6852 nextchar(pRExC_state);
6853 ender = reg_node(pRExC_state, SUCCEED);
6854 REGTAIL(pRExC_state, ret, ender);
6855 reginsert(pRExC_state, SUSPEND, ret, depth+1);
6857 ender = reg_node(pRExC_state, TAIL);
6858 REGTAIL(pRExC_state, ret, ender);
6862 if (RExC_parse < RExC_end && ISMULT2(RExC_parse)) {
6864 vFAIL("Nested quantifiers");
6871 /* reg_namedseq(pRExC_state,UVp)
6873 This is expected to be called by a parser routine that has
6874 recognized '\N' and needs to handle the rest. RExC_parse is
6875 expected to point at the first char following the N at the time
6878 The \N may be inside (indicated by valuep not being NULL) or outside a
6881 \N may begin either a named sequence, or if outside a character class, mean
6882 to match a non-newline. For non single-quoted regexes, the tokenizer has
6883 attempted to decide which, and in the case of a named sequence converted it
6884 into one of the forms: \N{} (if the sequence is null), or \N{U+c1.c2...},
6885 where c1... are the characters in the sequence. For single-quoted regexes,
6886 the tokenizer passes the \N sequence through unchanged; this code will not
6887 attempt to determine this nor expand those. The net effect is that if the
6888 beginning of the passed-in pattern isn't '{U+' or there is no '}', it
6889 signals that this \N occurrence means to match a non-newline.
6891 Only the \N{U+...} form should occur in a character class, for the same
6892 reason that '.' inside a character class means to just match a period: it
6893 just doesn't make sense.
6895 If valuep is non-null then it is assumed that we are parsing inside
6896 of a charclass definition and the first codepoint in the resolved
6897 string is returned via *valuep and the routine will return NULL.
6898 In this mode if a multichar string is returned from the charnames
6899 handler, a warning will be issued, and only the first char in the
6900 sequence will be examined. If the string returned is zero length
6901 then the value of *valuep is undefined and NON-NULL will
6902 be returned to indicate failure. (This will NOT be a valid pointer
6905 If valuep is null then it is assumed that we are parsing normal text and a
6906 new EXACT node is inserted into the program containing the resolved string,
6907 and a pointer to the new node is returned. But if the string is zero length
6908 a NOTHING node is emitted instead.
6910 On success RExC_parse is set to the char following the endbrace.
6911 Parsing failures will generate a fatal error via vFAIL(...)
6914 S_reg_namedseq(pTHX_ RExC_state_t *pRExC_state, UV *valuep, I32 *flagp)
6916 char * endbrace; /* '}' following the name */
6917 regnode *ret = NULL;
6919 char* parse_start = RExC_parse - 2; /* points to the '\N' */
6923 GET_RE_DEBUG_FLAGS_DECL;
6925 PERL_ARGS_ASSERT_REG_NAMEDSEQ;
6929 /* The [^\n] meaning of \N ignores spaces and comments under the /x
6930 * modifier. The other meaning does not */
6931 p = (RExC_flags & RXf_PMf_EXTENDED)
6932 ? regwhite( pRExC_state, RExC_parse )
6935 /* Disambiguate between \N meaning a named character versus \N meaning
6936 * [^\n]. The former is assumed when it can't be the latter. */
6937 if (*p != '{' || regcurly(p)) {
6940 /* no bare \N in a charclass */
6941 vFAIL("\\N in a character class must be a named character: \\N{...}");
6943 nextchar(pRExC_state);
6944 ret = reg_node(pRExC_state, REG_ANY);
6945 *flagp |= HASWIDTH|SIMPLE;
6948 Set_Node_Length(ret, 1); /* MJD */
6952 /* Here, we have decided it should be a named sequence */
6954 /* The test above made sure that the next real character is a '{', but
6955 * under the /x modifier, it could be separated by space (or a comment and
6956 * \n) and this is not allowed (for consistency with \x{...} and the
6957 * tokenizer handling of \N{NAME}). */
6958 if (*RExC_parse != '{') {
6959 vFAIL("Missing braces on \\N{}");
6962 RExC_parse++; /* Skip past the '{' */
6964 if (! (endbrace = strchr(RExC_parse, '}')) /* no trailing brace */
6965 || ! (endbrace == RExC_parse /* nothing between the {} */
6966 || (endbrace - RExC_parse >= 2 /* U+ (bad hex is checked below */
6967 && strnEQ(RExC_parse, "U+", 2)))) /* for a better error msg) */
6969 if (endbrace) RExC_parse = endbrace; /* position msg's '<--HERE' */
6970 vFAIL("\\N{NAME} must be resolved by the lexer");
6973 if (endbrace == RExC_parse) { /* empty: \N{} */
6975 RExC_parse = endbrace + 1;
6976 return reg_node(pRExC_state,NOTHING);
6980 ckWARNreg(RExC_parse,
6981 "Ignoring zero length \\N{} in character class"
6983 RExC_parse = endbrace + 1;
6986 return (regnode *) &RExC_parse; /* Invalid regnode pointer */
6989 REQUIRE_UTF8; /* named sequences imply Unicode semantics */
6990 RExC_parse += 2; /* Skip past the 'U+' */
6992 if (valuep) { /* In a bracketed char class */
6993 /* We only pay attention to the first char of
6994 multichar strings being returned. I kinda wonder
6995 if this makes sense as it does change the behaviour
6996 from earlier versions, OTOH that behaviour was broken
6997 as well. XXX Solution is to recharacterize as
6998 [rest-of-class]|multi1|multi2... */
7000 STRLEN length_of_hex;
7001 I32 flags = PERL_SCAN_ALLOW_UNDERSCORES
7002 | PERL_SCAN_DISALLOW_PREFIX
7003 | (SIZE_ONLY ? PERL_SCAN_SILENT_ILLDIGIT : 0);
7005 char * endchar = RExC_parse + strcspn(RExC_parse, ".}");
7006 if (endchar < endbrace) {
7007 ckWARNreg(endchar, "Using just the first character returned by \\N{} in character class");
7010 length_of_hex = (STRLEN)(endchar - RExC_parse);
7011 *valuep = grok_hex(RExC_parse, &length_of_hex, &flags, NULL);
7013 /* The tokenizer should have guaranteed validity, but it's possible to
7014 * bypass it by using single quoting, so check */
7015 if (length_of_hex == 0
7016 || length_of_hex != (STRLEN)(endchar - RExC_parse) )
7018 RExC_parse += length_of_hex; /* Includes all the valid */
7019 RExC_parse += (RExC_orig_utf8) /* point to after 1st invalid */
7020 ? UTF8SKIP(RExC_parse)
7022 /* Guard against malformed utf8 */
7023 if (RExC_parse >= endchar) RExC_parse = endchar;
7024 vFAIL("Invalid hexadecimal number in \\N{U+...}");
7027 RExC_parse = endbrace + 1;
7028 if (endchar == endbrace) return NULL;
7030 ret = (regnode *) &RExC_parse; /* Invalid regnode pointer */
7032 else { /* Not a char class */
7033 char *s; /* String to put in generated EXACT node */
7034 STRLEN len = 0; /* Its current byte length */
7035 char *endchar; /* Points to '.' or '}' ending cur char in the input
7038 ret = reg_node(pRExC_state, (U8) ((! FOLD) ? EXACT
7046 /* Exact nodes can hold only a U8 length's of text = 255. Loop through
7047 * the input which is of the form now 'c1.c2.c3...}' until find the
7048 * ending brace or exceed length 255. The characters that exceed this
7049 * limit are dropped. The limit could be relaxed should it become
7050 * desirable by reparsing this as (?:\N{NAME}), so could generate
7051 * multiple EXACT nodes, as is done for just regular input. But this
7052 * is primarily a named character, and not intended to be a huge long
7053 * string, so 255 bytes should be good enough */
7055 STRLEN length_of_hex;
7056 I32 grok_flags = PERL_SCAN_ALLOW_UNDERSCORES
7057 | PERL_SCAN_DISALLOW_PREFIX
7058 | (SIZE_ONLY ? PERL_SCAN_SILENT_ILLDIGIT : 0);
7059 UV cp; /* Ord of current character */
7061 /* Code points are separated by dots. If none, there is only one
7062 * code point, and is terminated by the brace */
7063 endchar = RExC_parse + strcspn(RExC_parse, ".}");
7065 /* The values are Unicode even on EBCDIC machines */
7066 length_of_hex = (STRLEN)(endchar - RExC_parse);
7067 cp = grok_hex(RExC_parse, &length_of_hex, &grok_flags, NULL);
7068 if ( length_of_hex == 0
7069 || length_of_hex != (STRLEN)(endchar - RExC_parse) )
7071 RExC_parse += length_of_hex; /* Includes all the valid */
7072 RExC_parse += (RExC_orig_utf8) /* point to after 1st invalid */
7073 ? UTF8SKIP(RExC_parse)
7075 /* Guard against malformed utf8 */
7076 if (RExC_parse >= endchar) RExC_parse = endchar;
7077 vFAIL("Invalid hexadecimal number in \\N{U+...}");
7080 if (! FOLD) { /* Not folding, just append to the string */
7083 /* Quit before adding this character if would exceed limit */
7084 if (len + UNISKIP(cp) > U8_MAX) break;
7086 unilen = reguni(pRExC_state, cp, s);
7091 } else { /* Folding, output the folded equivalent */
7092 STRLEN foldlen,numlen;
7093 U8 tmpbuf[UTF8_MAXBYTES_CASE+1], *foldbuf;
7094 cp = toFOLD_uni(cp, tmpbuf, &foldlen);
7096 /* Quit before exceeding size limit */
7097 if (len + foldlen > U8_MAX) break;
7099 for (foldbuf = tmpbuf;
7103 cp = utf8_to_uvchr(foldbuf, &numlen);
7105 const STRLEN unilen = reguni(pRExC_state, cp, s);
7108 /* In EBCDIC the numlen and unilen can differ. */
7110 if (numlen >= foldlen)
7114 break; /* "Can't happen." */
7118 /* Point to the beginning of the next character in the sequence. */
7119 RExC_parse = endchar + 1;
7121 /* Quit if no more characters */
7122 if (RExC_parse >= endbrace) break;
7127 if (RExC_parse < endbrace) {
7128 ckWARNreg(RExC_parse - 1,
7129 "Using just the first characters returned by \\N{}");
7132 RExC_size += STR_SZ(len);
7135 RExC_emit += STR_SZ(len);
7138 RExC_parse = endbrace + 1;
7140 *flagp |= HASWIDTH; /* Not SIMPLE, as that causes the engine to fail
7141 with malformed in t/re/pat_advanced.t */
7143 Set_Node_Cur_Length(ret); /* MJD */
7144 nextchar(pRExC_state);
7154 * It returns the code point in utf8 for the value in *encp.
7155 * value: a code value in the source encoding
7156 * encp: a pointer to an Encode object
7158 * If the result from Encode is not a single character,
7159 * it returns U+FFFD (Replacement character) and sets *encp to NULL.
7162 S_reg_recode(pTHX_ const char value, SV **encp)
7165 SV * const sv = newSVpvn_flags(&value, numlen, SVs_TEMP);
7166 const char * const s = *encp ? sv_recode_to_utf8(sv, *encp) : SvPVX(sv);
7167 const STRLEN newlen = SvCUR(sv);
7168 UV uv = UNICODE_REPLACEMENT;
7170 PERL_ARGS_ASSERT_REG_RECODE;
7174 ? utf8n_to_uvchr((U8*)s, newlen, &numlen, UTF8_ALLOW_DEFAULT)
7177 if (!newlen || numlen != newlen) {
7178 uv = UNICODE_REPLACEMENT;
7186 - regatom - the lowest level
7188 Try to identify anything special at the start of the pattern. If there
7189 is, then handle it as required. This may involve generating a single regop,
7190 such as for an assertion; or it may involve recursing, such as to
7191 handle a () structure.
7193 If the string doesn't start with something special then we gobble up
7194 as much literal text as we can.
7196 Once we have been able to handle whatever type of thing started the
7197 sequence, we return.
7199 Note: we have to be careful with escapes, as they can be both literal
7200 and special, and in the case of \10 and friends can either, depending
7201 on context. Specifically there are two separate switches for handling
7202 escape sequences, with the one for handling literal escapes requiring
7203 a dummy entry for all of the special escapes that are actually handled
7208 S_regatom(pTHX_ RExC_state_t *pRExC_state, I32 *flagp, U32 depth)
7211 register regnode *ret = NULL;
7213 char *parse_start = RExC_parse;
7215 GET_RE_DEBUG_FLAGS_DECL;
7216 DEBUG_PARSE("atom");
7217 *flagp = WORST; /* Tentatively. */
7219 PERL_ARGS_ASSERT_REGATOM;
7222 switch ((U8)*RExC_parse) {
7224 RExC_seen_zerolen++;
7225 nextchar(pRExC_state);
7226 if (RExC_flags & RXf_PMf_MULTILINE)
7227 ret = reg_node(pRExC_state, MBOL);
7228 else if (RExC_flags & RXf_PMf_SINGLELINE)
7229 ret = reg_node(pRExC_state, SBOL);
7231 ret = reg_node(pRExC_state, BOL);
7232 Set_Node_Length(ret, 1); /* MJD */
7235 nextchar(pRExC_state);
7237 RExC_seen_zerolen++;
7238 if (RExC_flags & RXf_PMf_MULTILINE)
7239 ret = reg_node(pRExC_state, MEOL);
7240 else if (RExC_flags & RXf_PMf_SINGLELINE)
7241 ret = reg_node(pRExC_state, SEOL);
7243 ret = reg_node(pRExC_state, EOL);
7244 Set_Node_Length(ret, 1); /* MJD */
7247 nextchar(pRExC_state);
7248 if (RExC_flags & RXf_PMf_SINGLELINE)
7249 ret = reg_node(pRExC_state, SANY);
7251 ret = reg_node(pRExC_state, REG_ANY);
7252 *flagp |= HASWIDTH|SIMPLE;
7254 Set_Node_Length(ret, 1); /* MJD */
7258 char * const oregcomp_parse = ++RExC_parse;
7259 ret = regclass(pRExC_state,depth+1);
7260 if (*RExC_parse != ']') {
7261 RExC_parse = oregcomp_parse;
7262 vFAIL("Unmatched [");
7264 nextchar(pRExC_state);
7265 *flagp |= HASWIDTH|SIMPLE;
7266 Set_Node_Length(ret, RExC_parse - oregcomp_parse + 1); /* MJD */
7270 nextchar(pRExC_state);
7271 ret = reg(pRExC_state, 1, &flags,depth+1);
7273 if (flags & TRYAGAIN) {
7274 if (RExC_parse == RExC_end) {
7275 /* Make parent create an empty node if needed. */
7283 *flagp |= flags&(HASWIDTH|SPSTART|SIMPLE|POSTPONED);
7287 if (flags & TRYAGAIN) {
7291 vFAIL("Internal urp");
7292 /* Supposed to be caught earlier. */
7295 if (!regcurly(RExC_parse)) {
7304 vFAIL("Quantifier follows nothing");
7306 case LATIN_SMALL_LETTER_SHARP_S:
7307 case UTF8_TWO_BYTE_HI_nocast(LATIN_SMALL_LETTER_SHARP_S):
7308 case UTF8_TWO_BYTE_HI_nocast(IOTA_D_T):
7309 #if UTF8_TWO_BYTE_HI_nocast(UPSILON_D_T) != UTF8_TWO_BYTE_HI_nocast(IOTA_D_T)
7310 #error The beginning utf8 byte of IOTA_D_T and UPSILON_D_T unexpectedly differ. Other instances in this code should have the case statement below.
7311 case UTF8_TWO_BYTE_HI_nocast(UPSILON_D_T):
7316 len=0; /* silence a spurious compiler warning */
7317 if ((cp = what_len_TRICKYFOLD_safe(RExC_parse,RExC_end,UTF,len))) {
7318 *flagp |= HASWIDTH; /* could be SIMPLE too, but needs a handler in regexec.regrepeat */
7319 RExC_parse+=len-1; /* we get one from nextchar() as well. :-( */
7320 ret = reganode(pRExC_state, FOLDCHAR, cp);
7321 Set_Node_Length(ret, 1); /* MJD */
7322 nextchar(pRExC_state); /* kill whitespace under /x */
7330 This switch handles escape sequences that resolve to some kind
7331 of special regop and not to literal text. Escape sequnces that
7332 resolve to literal text are handled below in the switch marked
7335 Every entry in this switch *must* have a corresponding entry
7336 in the literal escape switch. However, the opposite is not
7337 required, as the default for this switch is to jump to the
7338 literal text handling code.
7340 switch ((U8)*++RExC_parse) {
7341 case LATIN_SMALL_LETTER_SHARP_S:
7342 case UTF8_TWO_BYTE_HI_nocast(LATIN_SMALL_LETTER_SHARP_S):
7343 case UTF8_TWO_BYTE_HI_nocast(IOTA_D_T):
7345 /* Special Escapes */
7347 RExC_seen_zerolen++;
7348 ret = reg_node(pRExC_state, SBOL);
7350 goto finish_meta_pat;
7352 ret = reg_node(pRExC_state, GPOS);
7353 RExC_seen |= REG_SEEN_GPOS;
7355 goto finish_meta_pat;
7357 RExC_seen_zerolen++;
7358 ret = reg_node(pRExC_state, KEEPS);
7360 /* XXX:dmq : disabling in-place substitution seems to
7361 * be necessary here to avoid cases of memory corruption, as
7362 * with: C<$_="x" x 80; s/x\K/y/> -- rgs
7364 RExC_seen |= REG_SEEN_LOOKBEHIND;
7365 goto finish_meta_pat;
7367 ret = reg_node(pRExC_state, SEOL);
7369 RExC_seen_zerolen++; /* Do not optimize RE away */
7370 goto finish_meta_pat;
7372 ret = reg_node(pRExC_state, EOS);
7374 RExC_seen_zerolen++; /* Do not optimize RE away */
7375 goto finish_meta_pat;
7377 ret = reg_node(pRExC_state, CANY);
7378 RExC_seen |= REG_SEEN_CANY;
7379 *flagp |= HASWIDTH|SIMPLE;
7380 goto finish_meta_pat;
7382 ret = reg_node(pRExC_state, CLUMP);
7384 goto finish_meta_pat;
7386 switch (get_regex_charset(RExC_flags)) {
7387 case REGEX_LOCALE_CHARSET:
7390 case REGEX_UNICODE_CHARSET:
7393 case REGEX_ASCII_RESTRICTED_CHARSET:
7396 case REGEX_DEPENDS_CHARSET:
7402 ret = reg_node(pRExC_state, op);
7403 *flagp |= HASWIDTH|SIMPLE;
7404 goto finish_meta_pat;
7406 switch (get_regex_charset(RExC_flags)) {
7407 case REGEX_LOCALE_CHARSET:
7410 case REGEX_UNICODE_CHARSET:
7413 case REGEX_ASCII_RESTRICTED_CHARSET:
7416 case REGEX_DEPENDS_CHARSET:
7422 ret = reg_node(pRExC_state, op);
7423 *flagp |= HASWIDTH|SIMPLE;
7424 goto finish_meta_pat;
7426 RExC_seen_zerolen++;
7427 RExC_seen |= REG_SEEN_LOOKBEHIND;
7428 switch (get_regex_charset(RExC_flags)) {
7429 case REGEX_LOCALE_CHARSET:
7432 case REGEX_UNICODE_CHARSET:
7435 case REGEX_ASCII_RESTRICTED_CHARSET:
7438 case REGEX_DEPENDS_CHARSET:
7444 ret = reg_node(pRExC_state, op);
7445 FLAGS(ret) = get_regex_charset(RExC_flags);
7447 goto finish_meta_pat;
7449 RExC_seen_zerolen++;
7450 RExC_seen |= REG_SEEN_LOOKBEHIND;
7451 switch (get_regex_charset(RExC_flags)) {
7452 case REGEX_LOCALE_CHARSET:
7455 case REGEX_UNICODE_CHARSET:
7458 case REGEX_ASCII_RESTRICTED_CHARSET:
7461 case REGEX_DEPENDS_CHARSET:
7467 ret = reg_node(pRExC_state, op);
7468 FLAGS(ret) = get_regex_charset(RExC_flags);
7470 goto finish_meta_pat;
7472 switch (get_regex_charset(RExC_flags)) {
7473 case REGEX_LOCALE_CHARSET:
7476 case REGEX_UNICODE_CHARSET:
7479 case REGEX_ASCII_RESTRICTED_CHARSET:
7482 case REGEX_DEPENDS_CHARSET:
7488 ret = reg_node(pRExC_state, op);
7489 *flagp |= HASWIDTH|SIMPLE;
7490 goto finish_meta_pat;
7492 switch (get_regex_charset(RExC_flags)) {
7493 case REGEX_LOCALE_CHARSET:
7496 case REGEX_UNICODE_CHARSET:
7499 case REGEX_ASCII_RESTRICTED_CHARSET:
7502 case REGEX_DEPENDS_CHARSET:
7508 ret = reg_node(pRExC_state, op);
7509 *flagp |= HASWIDTH|SIMPLE;
7510 goto finish_meta_pat;
7512 switch (get_regex_charset(RExC_flags)) {
7513 case REGEX_LOCALE_CHARSET:
7516 case REGEX_ASCII_RESTRICTED_CHARSET:
7519 case REGEX_DEPENDS_CHARSET: /* No difference between these */
7520 case REGEX_UNICODE_CHARSET:
7526 ret = reg_node(pRExC_state, op);
7527 *flagp |= HASWIDTH|SIMPLE;
7528 goto finish_meta_pat;
7530 switch (get_regex_charset(RExC_flags)) {
7531 case REGEX_LOCALE_CHARSET:
7534 case REGEX_ASCII_RESTRICTED_CHARSET:
7537 case REGEX_DEPENDS_CHARSET: /* No difference between these */
7538 case REGEX_UNICODE_CHARSET:
7544 ret = reg_node(pRExC_state, op);
7545 *flagp |= HASWIDTH|SIMPLE;
7546 goto finish_meta_pat;
7548 ret = reg_node(pRExC_state, LNBREAK);
7549 *flagp |= HASWIDTH|SIMPLE;
7550 goto finish_meta_pat;
7552 ret = reg_node(pRExC_state, HORIZWS);
7553 *flagp |= HASWIDTH|SIMPLE;
7554 goto finish_meta_pat;
7556 ret = reg_node(pRExC_state, NHORIZWS);
7557 *flagp |= HASWIDTH|SIMPLE;
7558 goto finish_meta_pat;
7560 ret = reg_node(pRExC_state, VERTWS);
7561 *flagp |= HASWIDTH|SIMPLE;
7562 goto finish_meta_pat;
7564 ret = reg_node(pRExC_state, NVERTWS);
7565 *flagp |= HASWIDTH|SIMPLE;
7567 nextchar(pRExC_state);
7568 Set_Node_Length(ret, 2); /* MJD */
7573 char* const oldregxend = RExC_end;
7575 char* parse_start = RExC_parse - 2;
7578 if (RExC_parse[1] == '{') {
7579 /* a lovely hack--pretend we saw [\pX] instead */
7580 RExC_end = strchr(RExC_parse, '}');
7582 const U8 c = (U8)*RExC_parse;
7584 RExC_end = oldregxend;
7585 vFAIL2("Missing right brace on \\%c{}", c);
7590 RExC_end = RExC_parse + 2;
7591 if (RExC_end > oldregxend)
7592 RExC_end = oldregxend;
7596 ret = regclass(pRExC_state,depth+1);
7598 RExC_end = oldregxend;
7601 Set_Node_Offset(ret, parse_start + 2);
7602 Set_Node_Cur_Length(ret);
7603 nextchar(pRExC_state);
7604 *flagp |= HASWIDTH|SIMPLE;
7608 /* Handle \N and \N{NAME} here and not below because it can be
7609 multicharacter. join_exact() will join them up later on.
7610 Also this makes sure that things like /\N{BLAH}+/ and
7611 \N{BLAH} being multi char Just Happen. dmq*/
7613 ret= reg_namedseq(pRExC_state, NULL, flagp);
7615 case 'k': /* Handle \k<NAME> and \k'NAME' */
7618 char ch= RExC_parse[1];
7619 if (ch != '<' && ch != '\'' && ch != '{') {
7621 vFAIL2("Sequence %.2s... not terminated",parse_start);
7623 /* this pretty much dupes the code for (?P=...) in reg(), if
7624 you change this make sure you change that */
7625 char* name_start = (RExC_parse += 2);
7627 SV *sv_dat = reg_scan_name(pRExC_state,
7628 SIZE_ONLY ? REG_RSN_RETURN_NULL : REG_RSN_RETURN_DATA);
7629 ch= (ch == '<') ? '>' : (ch == '{') ? '}' : '\'';
7630 if (RExC_parse == name_start || *RExC_parse != ch)
7631 vFAIL2("Sequence %.3s... not terminated",parse_start);
7634 num = add_data( pRExC_state, 1, "S" );
7635 RExC_rxi->data->data[num]=(void*)sv_dat;
7636 SvREFCNT_inc_simple_void(sv_dat);
7640 ret = reganode(pRExC_state,
7643 : (AT_LEAST_UNI_SEMANTICS)
7651 /* override incorrect value set in reganode MJD */
7652 Set_Node_Offset(ret, parse_start+1);
7653 Set_Node_Cur_Length(ret); /* MJD */
7654 nextchar(pRExC_state);
7660 case '1': case '2': case '3': case '4':
7661 case '5': case '6': case '7': case '8': case '9':
7664 bool isg = *RExC_parse == 'g';
7669 if (*RExC_parse == '{') {
7673 if (*RExC_parse == '-') {
7677 if (hasbrace && !isDIGIT(*RExC_parse)) {
7678 if (isrel) RExC_parse--;
7680 goto parse_named_seq;
7682 num = atoi(RExC_parse);
7683 if (isg && num == 0)
7684 vFAIL("Reference to invalid group 0");
7686 num = RExC_npar - num;
7688 vFAIL("Reference to nonexistent or unclosed group");
7690 if (!isg && num > 9 && num >= RExC_npar)
7693 char * const parse_start = RExC_parse - 1; /* MJD */
7694 while (isDIGIT(*RExC_parse))
7696 if (parse_start == RExC_parse - 1)
7697 vFAIL("Unterminated \\g... pattern");
7699 if (*RExC_parse != '}')
7700 vFAIL("Unterminated \\g{...} pattern");
7704 if (num > (I32)RExC_rx->nparens)
7705 vFAIL("Reference to nonexistent group");
7708 ret = reganode(pRExC_state,
7711 : (AT_LEAST_UNI_SEMANTICS)
7719 /* override incorrect value set in reganode MJD */
7720 Set_Node_Offset(ret, parse_start+1);
7721 Set_Node_Cur_Length(ret); /* MJD */
7723 nextchar(pRExC_state);
7728 if (RExC_parse >= RExC_end)
7729 FAIL("Trailing \\");
7732 /* Do not generate "unrecognized" warnings here, we fall
7733 back into the quick-grab loop below */
7740 if (RExC_flags & RXf_PMf_EXTENDED) {
7741 if ( reg_skipcomment( pRExC_state ) )
7748 register STRLEN len;
7753 U8 tmpbuf[UTF8_MAXBYTES_CASE+1], *foldbuf;
7755 parse_start = RExC_parse - 1;
7761 ret = reg_node(pRExC_state,
7762 (U8) ((! FOLD) ? EXACT
7765 : (AT_LEAST_UNI_SEMANTICS)
7770 for (len = 0, p = RExC_parse - 1;
7771 len < 127 && p < RExC_end;
7774 char * const oldp = p;
7776 if (RExC_flags & RXf_PMf_EXTENDED)
7777 p = regwhite( pRExC_state, p );
7779 case LATIN_SMALL_LETTER_SHARP_S:
7780 case UTF8_TWO_BYTE_HI_nocast(LATIN_SMALL_LETTER_SHARP_S):
7781 case UTF8_TWO_BYTE_HI_nocast(IOTA_D_T):
7782 if (LOC || !FOLD || !is_TRICKYFOLD_safe(p,RExC_end,UTF))
7783 goto normal_default;
7793 /* Literal Escapes Switch
7795 This switch is meant to handle escape sequences that
7796 resolve to a literal character.
7798 Every escape sequence that represents something
7799 else, like an assertion or a char class, is handled
7800 in the switch marked 'Special Escapes' above in this
7801 routine, but also has an entry here as anything that
7802 isn't explicitly mentioned here will be treated as
7803 an unescaped equivalent literal.
7807 /* These are all the special escapes. */
7808 case LATIN_SMALL_LETTER_SHARP_S:
7809 case UTF8_TWO_BYTE_HI_nocast(LATIN_SMALL_LETTER_SHARP_S):
7810 case UTF8_TWO_BYTE_HI_nocast(IOTA_D_T):
7811 if (LOC || !FOLD || !is_TRICKYFOLD_safe(p,RExC_end,UTF))
7812 goto normal_default;
7813 case 'A': /* Start assertion */
7814 case 'b': case 'B': /* Word-boundary assertion*/
7815 case 'C': /* Single char !DANGEROUS! */
7816 case 'd': case 'D': /* digit class */
7817 case 'g': case 'G': /* generic-backref, pos assertion */
7818 case 'h': case 'H': /* HORIZWS */
7819 case 'k': case 'K': /* named backref, keep marker */
7820 case 'N': /* named char sequence */
7821 case 'p': case 'P': /* Unicode property */
7822 case 'R': /* LNBREAK */
7823 case 's': case 'S': /* space class */
7824 case 'v': case 'V': /* VERTWS */
7825 case 'w': case 'W': /* word class */
7826 case 'X': /* eXtended Unicode "combining character sequence" */
7827 case 'z': case 'Z': /* End of line/string assertion */
7831 /* Anything after here is an escape that resolves to a
7832 literal. (Except digits, which may or may not)
7851 ender = ASCII_TO_NATIVE('\033');
7855 ender = ASCII_TO_NATIVE('\007');
7860 STRLEN brace_len = len;
7862 const char* error_msg;
7864 bool valid = grok_bslash_o(p,
7871 RExC_parse = p; /* going to die anyway; point
7872 to exact spot of failure */
7879 if (PL_encoding && ender < 0x100) {
7880 goto recode_encoding;
7889 char* const e = strchr(p, '}');
7893 vFAIL("Missing right brace on \\x{}");
7896 I32 flags = PERL_SCAN_ALLOW_UNDERSCORES
7897 | PERL_SCAN_DISALLOW_PREFIX;
7898 STRLEN numlen = e - p - 1;
7899 ender = grok_hex(p + 1, &numlen, &flags, NULL);
7906 I32 flags = PERL_SCAN_DISALLOW_PREFIX;
7908 ender = grok_hex(p, &numlen, &flags, NULL);
7911 if (PL_encoding && ender < 0x100)
7912 goto recode_encoding;
7916 ender = grok_bslash_c(*p++, SIZE_ONLY);
7918 case '0': case '1': case '2': case '3':case '4':
7919 case '5': case '6': case '7': case '8':case '9':
7921 (isDIGIT(p[1]) && atoi(p) >= RExC_npar))
7923 I32 flags = PERL_SCAN_SILENT_ILLDIGIT;
7925 ender = grok_oct(p, &numlen, &flags, NULL);
7935 if (PL_encoding && ender < 0x100)
7936 goto recode_encoding;
7940 SV* enc = PL_encoding;
7941 ender = reg_recode((const char)(U8)ender, &enc);
7942 if (!enc && SIZE_ONLY)
7943 ckWARNreg(p, "Invalid escape in the specified encoding");
7949 FAIL("Trailing \\");
7952 if (!SIZE_ONLY&& isALPHA(*p))
7953 ckWARN2reg(p + 1, "Unrecognized escape \\%c passed through", UCHARAT(p));
7954 goto normal_default;
7959 if (UTF8_IS_START(*p) && UTF) {
7961 ender = utf8n_to_uvchr((U8*)p, RExC_end - p,
7962 &numlen, UTF8_ALLOW_DEFAULT);
7969 if ( RExC_flags & RXf_PMf_EXTENDED)
7970 p = regwhite( pRExC_state, p );
7972 /* Prime the casefolded buffer. */
7973 ender = toFOLD_uni(ender, tmpbuf, &foldlen);
7975 if (p < RExC_end && ISMULT2(p)) { /* Back off on ?+*. */
7980 /* Emit all the Unicode characters. */
7982 for (foldbuf = tmpbuf;
7984 foldlen -= numlen) {
7985 ender = utf8_to_uvchr(foldbuf, &numlen);
7987 const STRLEN unilen = reguni(pRExC_state, ender, s);
7990 /* In EBCDIC the numlen
7991 * and unilen can differ. */
7993 if (numlen >= foldlen)
7997 break; /* "Can't happen." */
8001 const STRLEN unilen = reguni(pRExC_state, ender, s);
8010 REGC((char)ender, s++);
8016 /* Emit all the Unicode characters. */
8018 for (foldbuf = tmpbuf;
8020 foldlen -= numlen) {
8021 ender = utf8_to_uvchr(foldbuf, &numlen);
8023 const STRLEN unilen = reguni(pRExC_state, ender, s);
8026 /* In EBCDIC the numlen
8027 * and unilen can differ. */
8029 if (numlen >= foldlen)
8037 const STRLEN unilen = reguni(pRExC_state, ender, s);
8046 REGC((char)ender, s++);
8050 Set_Node_Cur_Length(ret); /* MJD */
8051 nextchar(pRExC_state);
8053 /* len is STRLEN which is unsigned, need to copy to signed */
8056 vFAIL("Internal disaster");
8060 if (len == 1 && UNI_IS_INVARIANT(ender))
8064 RExC_size += STR_SZ(len);
8067 RExC_emit += STR_SZ(len);
8075 /* Jumped to when an unrecognized character set is encountered */
8077 Perl_croak(aTHX_ "panic: Unknown regex character set encoding: %u", get_regex_charset(RExC_flags));
8082 S_regwhite( RExC_state_t *pRExC_state, char *p )
8084 const char *e = RExC_end;
8086 PERL_ARGS_ASSERT_REGWHITE;
8091 else if (*p == '#') {
8100 RExC_seen |= REG_SEEN_RUN_ON_COMMENT;
8108 /* Parse POSIX character classes: [[:foo:]], [[=foo=]], [[.foo.]].
8109 Character classes ([:foo:]) can also be negated ([:^foo:]).
8110 Returns a named class id (ANYOF_XXX) if successful, -1 otherwise.
8111 Equivalence classes ([=foo=]) and composites ([.foo.]) are parsed,
8112 but trigger failures because they are currently unimplemented. */
8114 #define POSIXCC_DONE(c) ((c) == ':')
8115 #define POSIXCC_NOTYET(c) ((c) == '=' || (c) == '.')
8116 #define POSIXCC(c) (POSIXCC_DONE(c) || POSIXCC_NOTYET(c))
8119 S_regpposixcc(pTHX_ RExC_state_t *pRExC_state, I32 value)
8122 I32 namedclass = OOB_NAMEDCLASS;
8124 PERL_ARGS_ASSERT_REGPPOSIXCC;
8126 if (value == '[' && RExC_parse + 1 < RExC_end &&
8127 /* I smell either [: or [= or [. -- POSIX has been here, right? */
8128 POSIXCC(UCHARAT(RExC_parse))) {
8129 const char c = UCHARAT(RExC_parse);
8130 char* const s = RExC_parse++;
8132 while (RExC_parse < RExC_end && UCHARAT(RExC_parse) != c)
8134 if (RExC_parse == RExC_end)
8135 /* Grandfather lone [:, [=, [. */
8138 const char* const t = RExC_parse++; /* skip over the c */
8141 if (UCHARAT(RExC_parse) == ']') {
8142 const char *posixcc = s + 1;
8143 RExC_parse++; /* skip over the ending ] */
8146 const I32 complement = *posixcc == '^' ? *posixcc++ : 0;
8147 const I32 skip = t - posixcc;
8149 /* Initially switch on the length of the name. */
8152 if (memEQ(posixcc, "word", 4)) /* this is not POSIX, this is the Perl \w */
8153 namedclass = complement ? ANYOF_NALNUM : ANYOF_ALNUM;
8156 /* Names all of length 5. */
8157 /* alnum alpha ascii blank cntrl digit graph lower
8158 print punct space upper */
8159 /* Offset 4 gives the best switch position. */
8160 switch (posixcc[4]) {
8162 if (memEQ(posixcc, "alph", 4)) /* alpha */
8163 namedclass = complement ? ANYOF_NALPHA : ANYOF_ALPHA;
8166 if (memEQ(posixcc, "spac", 4)) /* space */
8167 namedclass = complement ? ANYOF_NPSXSPC : ANYOF_PSXSPC;
8170 if (memEQ(posixcc, "grap", 4)) /* graph */
8171 namedclass = complement ? ANYOF_NGRAPH : ANYOF_GRAPH;
8174 if (memEQ(posixcc, "asci", 4)) /* ascii */
8175 namedclass = complement ? ANYOF_NASCII : ANYOF_ASCII;
8178 if (memEQ(posixcc, "blan", 4)) /* blank */
8179 namedclass = complement ? ANYOF_NBLANK : ANYOF_BLANK;
8182 if (memEQ(posixcc, "cntr", 4)) /* cntrl */
8183 namedclass = complement ? ANYOF_NCNTRL : ANYOF_CNTRL;
8186 if (memEQ(posixcc, "alnu", 4)) /* alnum */
8187 namedclass = complement ? ANYOF_NALNUMC : ANYOF_ALNUMC;
8190 if (memEQ(posixcc, "lowe", 4)) /* lower */
8191 namedclass = complement ? ANYOF_NLOWER : ANYOF_LOWER;
8192 else if (memEQ(posixcc, "uppe", 4)) /* upper */
8193 namedclass = complement ? ANYOF_NUPPER : ANYOF_UPPER;
8196 if (memEQ(posixcc, "digi", 4)) /* digit */
8197 namedclass = complement ? ANYOF_NDIGIT : ANYOF_DIGIT;
8198 else if (memEQ(posixcc, "prin", 4)) /* print */
8199 namedclass = complement ? ANYOF_NPRINT : ANYOF_PRINT;
8200 else if (memEQ(posixcc, "punc", 4)) /* punct */
8201 namedclass = complement ? ANYOF_NPUNCT : ANYOF_PUNCT;
8206 if (memEQ(posixcc, "xdigit", 6))
8207 namedclass = complement ? ANYOF_NXDIGIT : ANYOF_XDIGIT;
8211 if (namedclass == OOB_NAMEDCLASS)
8212 Simple_vFAIL3("POSIX class [:%.*s:] unknown",
8214 assert (posixcc[skip] == ':');
8215 assert (posixcc[skip+1] == ']');
8216 } else if (!SIZE_ONLY) {
8217 /* [[=foo=]] and [[.foo.]] are still future. */
8219 /* adjust RExC_parse so the warning shows after
8221 while (UCHARAT(RExC_parse) && UCHARAT(RExC_parse) != ']')
8223 Simple_vFAIL3("POSIX syntax [%c %c] is reserved for future extensions", c, c);
8226 /* Maternal grandfather:
8227 * "[:" ending in ":" but not in ":]" */
8237 S_checkposixcc(pTHX_ RExC_state_t *pRExC_state)
8241 PERL_ARGS_ASSERT_CHECKPOSIXCC;
8243 if (POSIXCC(UCHARAT(RExC_parse))) {
8244 const char *s = RExC_parse;
8245 const char c = *s++;
8249 if (*s && c == *s && s[1] == ']') {
8251 "POSIX syntax [%c %c] belongs inside character classes",
8254 /* [[=foo=]] and [[.foo.]] are still future. */
8255 if (POSIXCC_NOTYET(c)) {
8256 /* adjust RExC_parse so the error shows after
8258 while (UCHARAT(RExC_parse) && UCHARAT(RExC_parse++) != ']')
8260 Simple_vFAIL3("POSIX syntax [%c %c] is reserved for future extensions", c, c);
8266 /* No locale test, and always Unicode semantics */
8267 #define _C_C_T_NOLOC_(NAME,TEST,WORD) \
8269 for (value = 0; value < 256; value++) \
8271 stored += S_set_regclass_bit(aTHX_ pRExC_state, ret, (U8) value); \
8275 case ANYOF_N##NAME: \
8276 for (value = 0; value < 256; value++) \
8278 stored += S_set_regclass_bit(aTHX_ pRExC_state, ret, (U8) value); \
8283 /* Like the above, but there are differences if we are in uni-8-bit or not, so
8284 * there are two tests passed in, to use depending on that. There aren't any
8285 * cases where the label is different from the name, so no need for that
8287 #define _C_C_T_(NAME,TEST_8,TEST_7,WORD) \
8289 if (LOC) ANYOF_CLASS_SET(ret, ANYOF_##NAME); \
8290 else if (UNI_SEMANTICS) { \
8291 for (value = 0; value < 256; value++) { \
8292 if (TEST_8) stored += \
8293 S_set_regclass_bit(aTHX_ pRExC_state, ret, (U8) value); \
8297 for (value = 0; value < 128; value++) { \
8298 if (TEST_7) stored += \
8299 S_set_regclass_bit(aTHX_ pRExC_state, ret, \
8300 (U8) UNI_TO_NATIVE(value)); \
8306 case ANYOF_N##NAME: \
8307 if (LOC) ANYOF_CLASS_SET(ret, ANYOF_N##NAME); \
8308 else if (UNI_SEMANTICS) { \
8309 for (value = 0; value < 256; value++) { \
8310 if (! TEST_8) stored += \
8311 S_set_regclass_bit(aTHX_ pRExC_state, ret, (U8) value); \
8315 for (value = 0; value < 128; value++) { \
8316 if (! TEST_7) stored += \
8317 S_set_regclass_bit(aTHX_ pRExC_state, ret, (U8) value); \
8319 if (ASCII_RESTRICTED) { \
8320 for (value = 128; value < 256; value++) { \
8321 stored += S_set_regclass_bit(aTHX_ pRExC_state, ret, (U8) value); \
8323 ANYOF_FLAGS(ret) |= ANYOF_UNICODE_ALL|ANYOF_UTF8; \
8326 /* For a non-ut8 target string with DEPENDS semantics, all above \
8327 * ASCII Latin1 code points match the complement of any of the \
8328 * classes. But in utf8, they have their Unicode semantics, so \
8329 * can't just set them in the bitmap, or else regexec.c will think \
8330 * they matched when they shouldn't. */ \
8331 ANYOF_FLAGS(ret) |= ANYOF_NON_UTF8_LATIN1_ALL|ANYOF_UTF8; \
8339 We dont use PERL_LEGACY_UNICODE_CHARCLASS_MAPPINGS as the direct test
8340 so that it is possible to override the option here without having to
8341 rebuild the entire core. as we are required to do if we change regcomp.h
8342 which is where PERL_LEGACY_UNICODE_CHARCLASS_MAPPINGS is defined.
8344 #if PERL_LEGACY_UNICODE_CHARCLASS_MAPPINGS
8345 #define BROKEN_UNICODE_CHARCLASS_MAPPINGS
8348 #ifdef BROKEN_UNICODE_CHARCLASS_MAPPINGS
8349 #define POSIX_CC_UNI_NAME(CCNAME) CCNAME
8351 #define POSIX_CC_UNI_NAME(CCNAME) "Posix" CCNAME
8355 S_set_regclass_bit_fold(pTHX_ RExC_state_t *pRExC_state, regnode* node, const U8 value)
8358 /* Handle the setting of folds in the bitmap for non-locale ANYOF nodes.
8359 * Locale folding is done at run-time, so this function should not be
8360 * called for nodes that are for locales.
8362 * This function simply sets the bit corresponding to the fold of the input
8363 * 'value', if not already set. The fold of 'f' is 'F', and the fold of
8366 * It also sets any necessary flags, and returns the number of bits that
8367 * actually changed from 0 to 1 */
8372 fold = (AT_LEAST_UNI_SEMANTICS) ? PL_fold_latin1[value]
8375 /* It assumes the bit for 'value' has already been set */
8376 if (fold != value && ! ANYOF_BITMAP_TEST(node, fold)) {
8377 ANYOF_BITMAP_SET(node, fold);
8380 if (_HAS_NONLATIN1_FOLD_CLOSURE_ONLY_FOR_USE_BY_REGCOMP_DOT_C_AND_REGEXEC_DOT_C(value)
8383 && PL_fold_latin1[value] != value))
8384 { /* A character that has a fold outside of Latin1 matches outside the
8385 bitmap, but only when the target string is utf8. Similarly when we
8386 don't have unicode semantics for the above ASCII Latin-1 characters,
8387 and they have a fold, they should match if the target is utf8, and
8389 ANYOF_FLAGS(node) |= ANYOF_UTF8;
8396 PERL_STATIC_INLINE U8
8397 S_set_regclass_bit(pTHX_ RExC_state_t *pRExC_state, regnode* node, const U8 value)
8399 /* This inline function sets a bit in the bitmap if not already set, and if
8400 * appropriate, its fold, returning the number of bits that actually
8401 * changed from 0 to 1 */
8405 if (ANYOF_BITMAP_TEST(node, value)) { /* Already set */
8409 ANYOF_BITMAP_SET(node, value);
8412 if (FOLD && ! LOC) { /* Locale folds aren't known until runtime */
8413 stored += S_set_regclass_bit_fold(aTHX_ pRExC_state, node, value);
8420 parse a class specification and produce either an ANYOF node that
8421 matches the pattern or if the pattern matches a single char only and
8422 that char is < 256 and we are case insensitive then we produce an
8427 S_regclass(pTHX_ RExC_state_t *pRExC_state, U32 depth)
8430 register UV nextvalue;
8431 register IV prevvalue = OOB_UNICODE;
8432 register IV range = 0;
8433 UV value = 0; /* XXX:dmq: needs to be referenceable (unfortunately) */
8434 register regnode *ret;
8437 char *rangebegin = NULL;
8438 bool need_class = 0;
8441 AV* unicode_alternate = NULL;
8443 UV literal_endpoint = 0;
8445 UV stored = 0; /* how many chars stored in the bitmap */
8447 regnode * const orig_emit = RExC_emit; /* Save the original RExC_emit in
8448 case we need to change the emitted regop to an EXACT. */
8449 const char * orig_parse = RExC_parse;
8450 GET_RE_DEBUG_FLAGS_DECL;
8452 PERL_ARGS_ASSERT_REGCLASS;
8454 PERL_UNUSED_ARG(depth);
8457 DEBUG_PARSE("clas");
8459 /* Assume we are going to generate an ANYOF node. */
8460 ret = reganode(pRExC_state, ANYOF, 0);
8463 ANYOF_FLAGS(ret) = 0;
8465 if (UCHARAT(RExC_parse) == '^') { /* Complement of range. */
8469 ANYOF_FLAGS(ret) |= ANYOF_INVERT;
8473 RExC_size += ANYOF_SKIP;
8474 #ifdef ANYOF_ADD_LOC_SKIP
8476 RExC_size += ANYOF_ADD_LOC_SKIP;
8479 listsv = &PL_sv_undef; /* For code scanners: listsv always non-NULL. */
8482 RExC_emit += ANYOF_SKIP;
8484 ANYOF_FLAGS(ret) |= ANYOF_LOCALE;
8485 #ifdef ANYOF_ADD_LOC_SKIP
8486 RExC_emit += ANYOF_ADD_LOC_SKIP;
8489 ANYOF_BITMAP_ZERO(ret);
8490 listsv = newSVpvs("# comment\n");
8493 nextvalue = RExC_parse < RExC_end ? UCHARAT(RExC_parse) : 0;
8495 if (!SIZE_ONLY && POSIXCC(nextvalue))
8496 checkposixcc(pRExC_state);
8498 /* allow 1st char to be ] (allowing it to be - is dealt with later) */
8499 if (UCHARAT(RExC_parse) == ']')
8503 while (RExC_parse < RExC_end && UCHARAT(RExC_parse) != ']') {
8507 namedclass = OOB_NAMEDCLASS; /* initialize as illegal */
8510 rangebegin = RExC_parse;
8512 value = utf8n_to_uvchr((U8*)RExC_parse,
8513 RExC_end - RExC_parse,
8514 &numlen, UTF8_ALLOW_DEFAULT);
8515 RExC_parse += numlen;
8518 value = UCHARAT(RExC_parse++);
8520 nextvalue = RExC_parse < RExC_end ? UCHARAT(RExC_parse) : 0;
8521 if (value == '[' && POSIXCC(nextvalue))
8522 namedclass = regpposixcc(pRExC_state, value);
8523 else if (value == '\\') {
8525 value = utf8n_to_uvchr((U8*)RExC_parse,
8526 RExC_end - RExC_parse,
8527 &numlen, UTF8_ALLOW_DEFAULT);
8528 RExC_parse += numlen;
8531 value = UCHARAT(RExC_parse++);
8532 /* Some compilers cannot handle switching on 64-bit integer
8533 * values, therefore value cannot be an UV. Yes, this will
8534 * be a problem later if we want switch on Unicode.
8535 * A similar issue a little bit later when switching on
8536 * namedclass. --jhi */
8537 switch ((I32)value) {
8538 case 'w': namedclass = ANYOF_ALNUM; break;
8539 case 'W': namedclass = ANYOF_NALNUM; break;
8540 case 's': namedclass = ANYOF_SPACE; break;
8541 case 'S': namedclass = ANYOF_NSPACE; break;
8542 case 'd': namedclass = ANYOF_DIGIT; break;
8543 case 'D': namedclass = ANYOF_NDIGIT; break;
8544 case 'v': namedclass = ANYOF_VERTWS; break;
8545 case 'V': namedclass = ANYOF_NVERTWS; break;
8546 case 'h': namedclass = ANYOF_HORIZWS; break;
8547 case 'H': namedclass = ANYOF_NHORIZWS; break;
8548 case 'N': /* Handle \N{NAME} in class */
8550 /* We only pay attention to the first char of
8551 multichar strings being returned. I kinda wonder
8552 if this makes sense as it does change the behaviour
8553 from earlier versions, OTOH that behaviour was broken
8555 UV v; /* value is register so we cant & it /grrr */
8556 if (reg_namedseq(pRExC_state, &v, NULL)) {
8566 if (RExC_parse >= RExC_end)
8567 vFAIL2("Empty \\%c{}", (U8)value);
8568 if (*RExC_parse == '{') {
8569 const U8 c = (U8)value;
8570 e = strchr(RExC_parse++, '}');
8572 vFAIL2("Missing right brace on \\%c{}", c);
8573 while (isSPACE(UCHARAT(RExC_parse)))
8575 if (e == RExC_parse)
8576 vFAIL2("Empty \\%c{}", c);
8578 while (isSPACE(UCHARAT(RExC_parse + n - 1)))
8586 if (UCHARAT(RExC_parse) == '^') {
8589 value = value == 'p' ? 'P' : 'p'; /* toggle */
8590 while (isSPACE(UCHARAT(RExC_parse))) {
8595 Perl_sv_catpvf(aTHX_ listsv, "%cutf8::%.*s\n",
8596 (value=='p' ? '+' : '!'), (int)n, RExC_parse);
8600 /* The \p could match something in the Latin1 range, hence
8601 * something that isn't utf8 */
8602 ANYOF_FLAGS(ret) |= ANYOF_NONBITMAP;
8603 namedclass = ANYOF_MAX; /* no official name, but it's named */
8606 case 'n': value = '\n'; break;
8607 case 'r': value = '\r'; break;
8608 case 't': value = '\t'; break;
8609 case 'f': value = '\f'; break;
8610 case 'b': value = '\b'; break;
8611 case 'e': value = ASCII_TO_NATIVE('\033');break;
8612 case 'a': value = ASCII_TO_NATIVE('\007');break;
8614 RExC_parse--; /* function expects to be pointed at the 'o' */
8616 const char* error_msg;
8617 bool valid = grok_bslash_o(RExC_parse,
8622 RExC_parse += numlen;
8627 if (PL_encoding && value < 0x100) {
8628 goto recode_encoding;
8632 if (*RExC_parse == '{') {
8633 I32 flags = PERL_SCAN_ALLOW_UNDERSCORES
8634 | PERL_SCAN_DISALLOW_PREFIX;
8635 char * const e = strchr(RExC_parse++, '}');
8637 vFAIL("Missing right brace on \\x{}");
8639 numlen = e - RExC_parse;
8640 value = grok_hex(RExC_parse, &numlen, &flags, NULL);
8644 I32 flags = PERL_SCAN_DISALLOW_PREFIX;
8646 value = grok_hex(RExC_parse, &numlen, &flags, NULL);
8647 RExC_parse += numlen;
8649 if (PL_encoding && value < 0x100)
8650 goto recode_encoding;
8653 value = grok_bslash_c(*RExC_parse++, SIZE_ONLY);
8655 case '0': case '1': case '2': case '3': case '4':
8656 case '5': case '6': case '7':
8658 /* Take 1-3 octal digits */
8659 I32 flags = PERL_SCAN_SILENT_ILLDIGIT;
8661 value = grok_oct(--RExC_parse, &numlen, &flags, NULL);
8662 RExC_parse += numlen;
8663 if (PL_encoding && value < 0x100)
8664 goto recode_encoding;
8669 SV* enc = PL_encoding;
8670 value = reg_recode((const char)(U8)value, &enc);
8671 if (!enc && SIZE_ONLY)
8672 ckWARNreg(RExC_parse,
8673 "Invalid escape in the specified encoding");
8677 /* Allow \_ to not give an error */
8678 if (!SIZE_ONLY && isALNUM(value) && value != '_') {
8679 ckWARN2reg(RExC_parse,
8680 "Unrecognized escape \\%c in character class passed through",
8685 } /* end of \blah */
8691 if (namedclass > OOB_NAMEDCLASS) { /* this is a named class \blah */
8693 /* What matches in a locale is not known until runtime, so need to
8694 * (one time per class) allocate extra space to pass to regexec.
8695 * The space will contain a bit for each named class that is to be
8696 * matched against. This isn't needed for \p{} and pseudo-classes,
8697 * as they are not affected by locale, and hence are dealt with
8699 if (LOC && namedclass < ANYOF_MAX && ! need_class) {
8702 #ifdef ANYOF_CLASS_ADD_SKIP
8703 RExC_size += ANYOF_CLASS_ADD_SKIP;
8707 #ifdef ANYOF_CLASS_ADD_SKIP
8708 RExC_emit += ANYOF_CLASS_ADD_SKIP;
8710 ANYOF_CLASS_ZERO(ret);
8712 ANYOF_FLAGS(ret) |= ANYOF_CLASS;
8715 /* a bad range like a-\d, a-[:digit:]. The '-' is taken as a
8720 RExC_parse >= rangebegin ?
8721 RExC_parse - rangebegin : 0;
8722 ckWARN4reg(RExC_parse,
8723 "False [] range \"%*.*s\"",
8726 if (prevvalue < 256) {
8728 S_set_regclass_bit(aTHX_ pRExC_state, ret, (U8) prevvalue);
8730 S_set_regclass_bit(aTHX_ pRExC_state, ret, '-');
8733 ANYOF_FLAGS(ret) |= ANYOF_UTF8;
8734 Perl_sv_catpvf(aTHX_ listsv,
8735 "%04"UVxf"\n%04"UVxf"\n", (UV)prevvalue, (UV) '-');
8739 range = 0; /* this was not a true range */
8745 const char *what = NULL;
8748 /* Possible truncation here but in some 64-bit environments
8749 * the compiler gets heartburn about switch on 64-bit values.
8750 * A similar issue a little earlier when switching on value.
8752 switch ((I32)namedclass) {
8754 case _C_C_T_(ALNUMC, isALNUMC_L1(value), isALNUMC(value), "XPosixAlnum");
8755 case _C_C_T_(ALPHA, isALPHA_L1(value), isALPHA(value), "XPosixAlpha");
8756 case _C_C_T_(BLANK, isBLANK_L1(value), isBLANK(value), "XPosixBlank");
8757 case _C_C_T_(CNTRL, isCNTRL_L1(value), isCNTRL(value), "XPosixCntrl");
8758 case _C_C_T_(GRAPH, isGRAPH_L1(value), isGRAPH(value), "XPosixGraph");
8759 case _C_C_T_(LOWER, isLOWER_L1(value), isLOWER(value), "XPosixLower");
8760 case _C_C_T_(PRINT, isPRINT_L1(value), isPRINT(value), "XPosixPrint");
8761 case _C_C_T_(PSXSPC, isPSXSPC_L1(value), isPSXSPC(value), "XPosixSpace");
8762 case _C_C_T_(PUNCT, isPUNCT_L1(value), isPUNCT(value), "XPosixPunct");
8763 case _C_C_T_(UPPER, isUPPER_L1(value), isUPPER(value), "XPosixUpper");
8764 #ifdef BROKEN_UNICODE_CHARCLASS_MAPPINGS
8765 /* \s, \w match all unicode if utf8. */
8766 case _C_C_T_(SPACE, isSPACE_L1(value), isSPACE(value), "SpacePerl");
8767 case _C_C_T_(ALNUM, isWORDCHAR_L1(value), isALNUM(value), "Word");
8769 /* \s, \w match ascii and locale only */
8770 case _C_C_T_(SPACE, isSPACE_L1(value), isSPACE(value), "PerlSpace");
8771 case _C_C_T_(ALNUM, isWORDCHAR_L1(value), isALNUM(value), "PerlWord");
8773 case _C_C_T_(XDIGIT, isXDIGIT_L1(value), isXDIGIT(value), "XPosixXDigit");
8774 case _C_C_T_NOLOC_(VERTWS, is_VERTWS_latin1(&value), "VertSpace");
8775 case _C_C_T_NOLOC_(HORIZWS, is_HORIZWS_latin1(&value), "HorizSpace");
8778 ANYOF_CLASS_SET(ret, ANYOF_ASCII);
8780 for (value = 0; value < 128; value++)
8782 S_set_regclass_bit(aTHX_ pRExC_state, ret, (U8) ASCII_TO_NATIVE(value));
8785 what = NULL; /* Doesn't match outside ascii, so
8786 don't want to add +utf8:: */
8790 ANYOF_CLASS_SET(ret, ANYOF_NASCII);
8792 for (value = 128; value < 256; value++)
8794 S_set_regclass_bit(aTHX_ pRExC_state, ret, (U8) ASCII_TO_NATIVE(value));
8796 ANYOF_FLAGS(ret) |= ANYOF_UNICODE_ALL;
8802 ANYOF_CLASS_SET(ret, ANYOF_DIGIT);
8804 /* consecutive digits assumed */
8805 for (value = '0'; value <= '9'; value++)
8807 S_set_regclass_bit(aTHX_ pRExC_state, ret, (U8) value);
8810 what = POSIX_CC_UNI_NAME("Digit");
8814 ANYOF_CLASS_SET(ret, ANYOF_NDIGIT);
8816 /* consecutive digits assumed */
8817 for (value = 0; value < '0'; value++)
8819 S_set_regclass_bit(aTHX_ pRExC_state, ret, (U8) value);
8820 for (value = '9' + 1; value < 256; value++)
8822 S_set_regclass_bit(aTHX_ pRExC_state, ret, (U8) value);
8825 what = POSIX_CC_UNI_NAME("Digit");
8826 if (ASCII_RESTRICTED ) {
8827 ANYOF_FLAGS(ret) |= ANYOF_UNICODE_ALL;
8831 /* this is to handle \p and \P */
8834 vFAIL("Invalid [::] class");
8837 if (what && ! (ASCII_RESTRICTED)) {
8838 /* Strings such as "+utf8::isWord\n" */
8839 Perl_sv_catpvf(aTHX_ listsv, "%cutf8::Is%s\n", yesno, what);
8840 ANYOF_FLAGS(ret) |= ANYOF_UTF8;
8845 } /* end of namedclass \blah */
8848 if (prevvalue > (IV)value) /* b-a */ {
8849 const int w = RExC_parse - rangebegin;
8850 Simple_vFAIL4("Invalid [] range \"%*.*s\"", w, w, rangebegin);
8851 range = 0; /* not a valid range */
8855 prevvalue = value; /* save the beginning of the range */
8856 if (*RExC_parse == '-' && RExC_parse+1 < RExC_end &&
8857 RExC_parse[1] != ']') {
8860 /* a bad range like \w-, [:word:]- ? */
8861 if (namedclass > OOB_NAMEDCLASS) {
8862 if (ckWARN(WARN_REGEXP)) {
8864 RExC_parse >= rangebegin ?
8865 RExC_parse - rangebegin : 0;
8867 "False [] range \"%*.*s\"",
8872 S_set_regclass_bit(aTHX_ pRExC_state, ret, '-');
8874 range = 1; /* yeah, it's a range! */
8875 continue; /* but do it the next time */
8879 /* now is the next time */
8881 if (prevvalue < 256) {
8882 const IV ceilvalue = value < 256 ? value : 255;
8885 /* In EBCDIC [\x89-\x91] should include
8886 * the \x8e but [i-j] should not. */
8887 if (literal_endpoint == 2 &&
8888 ((isLOWER(prevvalue) && isLOWER(ceilvalue)) ||
8889 (isUPPER(prevvalue) && isUPPER(ceilvalue))))
8891 if (isLOWER(prevvalue)) {
8892 for (i = prevvalue; i <= ceilvalue; i++)
8893 if (isLOWER(i) && !ANYOF_BITMAP_TEST(ret,i)) {
8895 S_set_regclass_bit(aTHX_ pRExC_state, ret, (U8) i);
8898 for (i = prevvalue; i <= ceilvalue; i++)
8899 if (isUPPER(i) && !ANYOF_BITMAP_TEST(ret,i)) {
8901 S_set_regclass_bit(aTHX_ pRExC_state, ret, (U8) i);
8907 for (i = prevvalue; i <= ceilvalue; i++) {
8908 stored += S_set_regclass_bit(aTHX_ pRExC_state, ret, (U8) i);
8911 if (value > 255 || UTF) {
8912 const UV prevnatvalue = NATIVE_TO_UNI(prevvalue);
8913 const UV natvalue = NATIVE_TO_UNI(value);
8915 /* If the code point requires utf8 to represent, and we are not
8916 * folding, it can't match unless the target is in utf8. Only
8917 * a few code points above 255 fold to below it, so XXX an
8918 * optimization would be to know which ones and set the flag
8920 ANYOF_FLAGS(ret) |= (FOLD || value < 256)
8923 if (prevnatvalue < natvalue) { /* '>' case is fatal error above */
8925 /* The \t sets the whole range */
8926 Perl_sv_catpvf(aTHX_ listsv, "%04"UVxf"\t%04"UVxf"\n",
8927 prevnatvalue, natvalue);
8929 /* Currently, we don't look at every value in the range.
8930 * Therefore we have to assume the worst case: that if
8931 * folding, it will match more than one character. But in
8932 * lookbehind patterns, can only be single character
8933 * length, so disallow those folds */
8934 if (FOLD && ! RExC_in_lookbehind) {
8938 else if (prevnatvalue == natvalue) {
8939 Perl_sv_catpvf(aTHX_ listsv, "%04"UVxf"\n", natvalue);
8941 U8 foldbuf[UTF8_MAXBYTES_CASE+1];
8943 const UV f = to_uni_fold(natvalue, foldbuf, &foldlen);
8945 #ifdef EBCDIC /* RD t/uni/fold ff and 6b */
8946 if (RExC_precomp[0] == ':' &&
8947 RExC_precomp[1] == '[' &&
8948 (f == 0xDF || f == 0x92)) {
8949 f = NATIVE_TO_UNI(f);
8952 /* If folding and foldable and a single
8953 * character, insert also the folded version
8954 * to the charclass. */
8956 #ifdef EBCDIC /* RD tunifold ligatures s,t fb05, fb06 */
8957 if ((RExC_precomp[0] == ':' &&
8958 RExC_precomp[1] == '[' &&
8960 (value == 0xFB05 || value == 0xFB06))) ?
8961 foldlen == ((STRLEN)UNISKIP(f) - 1) :
8962 foldlen == (STRLEN)UNISKIP(f) )
8964 if (foldlen == (STRLEN)UNISKIP(f))
8966 Perl_sv_catpvf(aTHX_ listsv,
8968 else if (! RExC_in_lookbehind) {
8969 /* Any multicharacter foldings
8970 * (disallowed in lookbehind patterns)
8971 * require the following transform:
8972 * [ABCDEF] -> (?:[ABCabcDEFd]|pq|rst)
8973 * where E folds into "pq" and F folds
8974 * into "rst", all other characters
8975 * fold to single characters. We save
8976 * away these multicharacter foldings,
8977 * to be later saved as part of the
8978 * additional "s" data. */
8981 if (!unicode_alternate)
8982 unicode_alternate = newAV();
8983 sv = newSVpvn_utf8((char*)foldbuf, foldlen,
8985 av_push(unicode_alternate, sv);
8990 /* If folding and the value is one of the Greek
8991 * sigmas insert a few more sigmas to make the
8992 * folding rules of the sigmas to work right.
8993 * Note that not all the possible combinations
8994 * are handled here: some of them are handled
8995 * by the standard folding rules, and some of
8996 * them (literal or EXACTF cases) are handled
8997 * during runtime in regexec.c:S_find_byclass(). */
8998 if (value == UNICODE_GREEK_SMALL_LETTER_FINAL_SIGMA) {
8999 Perl_sv_catpvf(aTHX_ listsv, "%04"UVxf"\n",
9000 (UV)UNICODE_GREEK_CAPITAL_LETTER_SIGMA);
9001 Perl_sv_catpvf(aTHX_ listsv, "%04"UVxf"\n",
9002 (UV)UNICODE_GREEK_SMALL_LETTER_SIGMA);
9004 else if (value == UNICODE_GREEK_CAPITAL_LETTER_SIGMA)
9005 Perl_sv_catpvf(aTHX_ listsv, "%04"UVxf"\n",
9006 (UV)UNICODE_GREEK_SMALL_LETTER_SIGMA);
9011 literal_endpoint = 0;
9015 range = 0; /* this range (if it was one) is done now */
9022 /****** !SIZE_ONLY AFTER HERE *********/
9024 /* Optimize inverted simple patterns (e.g. [^a-z]). Note that we haven't
9025 * set the FOLD flag yet, so this this does optimize those. It doesn't
9026 * optimize locale. Doing so perhaps could be done as long as there is
9027 * nothing like \w in it; some thought also would have to be given to the
9028 * interaction with above 0x100 chars */
9029 if (! LOC && (ANYOF_FLAGS(ret) & ANYOF_FLAGS_ALL) == ANYOF_INVERT) {
9030 for (value = 0; value < ANYOF_BITMAP_SIZE; ++value)
9031 ANYOF_BITMAP(ret)[value] ^= 0xFF;
9032 stored = 256 - stored;
9034 /* The inversion means that everything above 255 is matched; and at the
9035 * same time we clear the invert flag */
9036 ANYOF_FLAGS(ret) = ANYOF_UTF8|ANYOF_UNICODE_ALL;
9042 /* This is the one character in the bitmap that needs special handling
9043 * under non-locale folding, as it folds to two characters 'ss'. This
9044 * happens if it is set and not inverting, or isn't set and are
9045 * inverting (disallowed in lookbehind patterns because they can't be
9046 * variable length) */
9048 && ! RExC_in_lookbehind
9049 && (cBOOL(ANYOF_BITMAP_TEST(ret, LATIN_SMALL_LETTER_SHARP_S))
9050 ^ cBOOL(ANYOF_FLAGS(ret) & ANYOF_INVERT)))
9052 OP(ret) = ANYOFV; /* Can match more than a single char */
9054 /* Under Unicode semantics), it can do this when the target string
9056 if (UNI_SEMANTICS) {
9057 ANYOF_FLAGS(ret) |= ANYOF_NONBITMAP_NON_UTF8;
9060 if (!unicode_alternate) {
9061 unicode_alternate = newAV();
9063 sv = newSVpvn_utf8("ss", 2, TRUE);
9064 av_push(unicode_alternate, sv);
9067 /* Folding in the bitmap is taken care of above, but not for locale
9068 * (for which we have to wait to see what folding is in effect at
9069 * runtime), and for things not in the bitmap. Set run-time fold flag
9071 if ((LOC || (ANYOF_FLAGS(ret) & ANYOF_NONBITMAP))) {
9072 ANYOF_FLAGS(ret) |= ANYOF_LOC_NONBITMAP_FOLD;
9076 /* A single character class can be "optimized" into an EXACTish node.
9077 * Note that since we don't currently count how many characters there are
9078 * outside the bitmap, we are XXX missing optimization possibilities for
9079 * them. This optimization can't happen unless this is a truly single
9080 * character class, which means that it can't be an inversion into a
9081 * many-character class, and there must be no possibility of there being
9082 * things outside the bitmap. 'stored' (only) for locales doesn't include
9083 * \w, etc, so have to make a special test that they aren't present
9085 * Similarly A 2-character class of the very special form like [bB] can be
9086 * optimized into an EXACTFish node, but only for non-locales, and for
9087 * characters which only have the two folds; so things like 'fF' and 'Ii'
9088 * wouldn't work because they are part of the fold of 'LATIN SMALL LIGATURE
9090 if (! (ANYOF_FLAGS(ret) & (ANYOF_NONBITMAP|ANYOF_INVERT|ANYOF_UNICODE_ALL))
9091 && (((stored == 1 && ((! (ANYOF_FLAGS(ret) & ANYOF_LOCALE))
9092 || (! ANYOF_CLASS_TEST_ANY_SET(ret)))))
9093 || (stored == 2 && ((! (ANYOF_FLAGS(ret) & ANYOF_LOCALE))
9094 && (! _HAS_NONLATIN1_FOLD_CLOSURE_ONLY_FOR_USE_BY_REGCOMP_DOT_C_AND_REGEXEC_DOT_C(value))
9095 /* If the latest code point has a fold whose
9096 * bit is set, it must be the only other one */
9097 && ((prevvalue = PL_fold_latin1[value]) != (IV)value)
9098 && ANYOF_BITMAP_TEST(ret, prevvalue)))))
9100 /* Note that the information needed to decide to do this optimization
9101 * is not currently available until the 2nd pass, and that the actually
9102 * used EXACTish node takes less space than the calculated ANYOF node,
9103 * and hence the amount of space calculated in the first pass is larger
9104 * than actually used, so this optimization doesn't gain us any space.
9105 * But an EXACT node is faster than an ANYOF node, and can be combined
9106 * with any adjacent EXACT nodes later by the optimizer for further
9107 * gains. The speed of executing an EXACTF is similar to an ANYOF
9108 * node, so the optimization advantage comes from the ability to join
9109 * it to adjacent EXACT nodes */
9111 const char * cur_parse= RExC_parse;
9113 RExC_emit = (regnode *)orig_emit;
9114 RExC_parse = (char *)orig_parse;
9118 /* A locale node with one point can be folded; all the other cases
9119 * with folding will have two points, since we calculate them above
9121 if (ANYOF_FLAGS(ret) & ANYOF_LOC_NONBITMAP_FOLD) {
9127 } /* else 2 chars in the bit map: the folds of each other */
9128 else if (AT_LEAST_UNI_SEMANTICS || !isASCII(value)) {
9130 /* To join adjacent nodes, they must be the exact EXACTish type.
9131 * Try to use the most likely type, by using EXACTFU if the regex
9132 * calls for them, or is required because the character is
9136 else { /* Otherwise, more likely to be EXACTF type */
9140 ret = reg_node(pRExC_state, op);
9141 RExC_parse = (char *)cur_parse;
9142 if (UTF && ! NATIVE_IS_INVARIANT(value)) {
9143 *STRING(ret)= UTF8_EIGHT_BIT_HI((U8) value);
9144 *(STRING(ret) + 1)= UTF8_EIGHT_BIT_LO((U8) value);
9146 RExC_emit += STR_SZ(2);
9149 *STRING(ret)= (char)value;
9151 RExC_emit += STR_SZ(1);
9153 SvREFCNT_dec(listsv);
9158 AV * const av = newAV();
9160 /* The 0th element stores the character class description
9161 * in its textual form: used later (regexec.c:Perl_regclass_swash())
9162 * to initialize the appropriate swash (which gets stored in
9163 * the 1st element), and also useful for dumping the regnode.
9164 * The 2nd element stores the multicharacter foldings,
9165 * used later (regexec.c:S_reginclass()). */
9166 av_store(av, 0, listsv);
9167 av_store(av, 1, NULL);
9168 av_store(av, 2, MUTABLE_SV(unicode_alternate));
9169 rv = newRV_noinc(MUTABLE_SV(av));
9170 n = add_data(pRExC_state, 1, "s");
9171 RExC_rxi->data->data[n] = (void*)rv;
9179 /* reg_skipcomment()
9181 Absorbs an /x style # comments from the input stream.
9182 Returns true if there is more text remaining in the stream.
9183 Will set the REG_SEEN_RUN_ON_COMMENT flag if the comment
9184 terminates the pattern without including a newline.
9186 Note its the callers responsibility to ensure that we are
9192 S_reg_skipcomment(pTHX_ RExC_state_t *pRExC_state)
9196 PERL_ARGS_ASSERT_REG_SKIPCOMMENT;
9198 while (RExC_parse < RExC_end)
9199 if (*RExC_parse++ == '\n') {
9204 /* we ran off the end of the pattern without ending
9205 the comment, so we have to add an \n when wrapping */
9206 RExC_seen |= REG_SEEN_RUN_ON_COMMENT;
9214 Advances the parse position, and optionally absorbs
9215 "whitespace" from the inputstream.
9217 Without /x "whitespace" means (?#...) style comments only,
9218 with /x this means (?#...) and # comments and whitespace proper.
9220 Returns the RExC_parse point from BEFORE the scan occurs.
9222 This is the /x friendly way of saying RExC_parse++.
9226 S_nextchar(pTHX_ RExC_state_t *pRExC_state)
9228 char* const retval = RExC_parse++;
9230 PERL_ARGS_ASSERT_NEXTCHAR;
9233 if (*RExC_parse == '(' && RExC_parse[1] == '?' &&
9234 RExC_parse[2] == '#') {
9235 while (*RExC_parse != ')') {
9236 if (RExC_parse == RExC_end)
9237 FAIL("Sequence (?#... not terminated");
9243 if (RExC_flags & RXf_PMf_EXTENDED) {
9244 if (isSPACE(*RExC_parse)) {
9248 else if (*RExC_parse == '#') {
9249 if ( reg_skipcomment( pRExC_state ) )
9258 - reg_node - emit a node
9260 STATIC regnode * /* Location. */
9261 S_reg_node(pTHX_ RExC_state_t *pRExC_state, U8 op)
9264 register regnode *ptr;
9265 regnode * const ret = RExC_emit;
9266 GET_RE_DEBUG_FLAGS_DECL;
9268 PERL_ARGS_ASSERT_REG_NODE;
9271 SIZE_ALIGN(RExC_size);
9275 if (RExC_emit >= RExC_emit_bound)
9276 Perl_croak(aTHX_ "panic: reg_node overrun trying to emit %d", op);
9278 NODE_ALIGN_FILL(ret);
9280 FILL_ADVANCE_NODE(ptr, op);
9281 REH_CALL_COMP_NODE_HOOK(pRExC_state->rx, (ptr) - 1);
9282 #ifdef RE_TRACK_PATTERN_OFFSETS
9283 if (RExC_offsets) { /* MJD */
9284 MJD_OFFSET_DEBUG(("%s:%d: (op %s) %s %"UVuf" (len %"UVuf") (max %"UVuf").\n",
9285 "reg_node", __LINE__,
9287 (UV)(RExC_emit - RExC_emit_start) > RExC_offsets[0]
9288 ? "Overwriting end of array!\n" : "OK",
9289 (UV)(RExC_emit - RExC_emit_start),
9290 (UV)(RExC_parse - RExC_start),
9291 (UV)RExC_offsets[0]));
9292 Set_Node_Offset(RExC_emit, RExC_parse + (op == END));
9300 - reganode - emit a node with an argument
9302 STATIC regnode * /* Location. */
9303 S_reganode(pTHX_ RExC_state_t *pRExC_state, U8 op, U32 arg)
9306 register regnode *ptr;
9307 regnode * const ret = RExC_emit;
9308 GET_RE_DEBUG_FLAGS_DECL;
9310 PERL_ARGS_ASSERT_REGANODE;
9313 SIZE_ALIGN(RExC_size);
9318 assert(2==regarglen[op]+1);
9320 Anything larger than this has to allocate the extra amount.
9321 If we changed this to be:
9323 RExC_size += (1 + regarglen[op]);
9325 then it wouldn't matter. Its not clear what side effect
9326 might come from that so its not done so far.
9331 if (RExC_emit >= RExC_emit_bound)
9332 Perl_croak(aTHX_ "panic: reg_node overrun trying to emit %d", op);
9334 NODE_ALIGN_FILL(ret);
9336 FILL_ADVANCE_NODE_ARG(ptr, op, arg);
9337 REH_CALL_COMP_NODE_HOOK(pRExC_state->rx, (ptr) - 2);
9338 #ifdef RE_TRACK_PATTERN_OFFSETS
9339 if (RExC_offsets) { /* MJD */
9340 MJD_OFFSET_DEBUG(("%s(%d): (op %s) %s %"UVuf" <- %"UVuf" (max %"UVuf").\n",
9344 (UV)(RExC_emit - RExC_emit_start) > RExC_offsets[0] ?
9345 "Overwriting end of array!\n" : "OK",
9346 (UV)(RExC_emit - RExC_emit_start),
9347 (UV)(RExC_parse - RExC_start),
9348 (UV)RExC_offsets[0]));
9349 Set_Cur_Node_Offset;
9357 - reguni - emit (if appropriate) a Unicode character
9360 S_reguni(pTHX_ const RExC_state_t *pRExC_state, UV uv, char* s)
9364 PERL_ARGS_ASSERT_REGUNI;
9366 return SIZE_ONLY ? UNISKIP(uv) : (uvchr_to_utf8((U8*)s, uv) - (U8*)s);
9370 - reginsert - insert an operator in front of already-emitted operand
9372 * Means relocating the operand.
9375 S_reginsert(pTHX_ RExC_state_t *pRExC_state, U8 op, regnode *opnd, U32 depth)
9378 register regnode *src;
9379 register regnode *dst;
9380 register regnode *place;
9381 const int offset = regarglen[(U8)op];
9382 const int size = NODE_STEP_REGNODE + offset;
9383 GET_RE_DEBUG_FLAGS_DECL;
9385 PERL_ARGS_ASSERT_REGINSERT;
9386 PERL_UNUSED_ARG(depth);
9387 /* (PL_regkind[(U8)op] == CURLY ? EXTRA_STEP_2ARGS : 0); */
9388 DEBUG_PARSE_FMT("inst"," - %s",PL_reg_name[op]);
9397 if (RExC_open_parens) {
9399 /*DEBUG_PARSE_FMT("inst"," - %"IVdf, (IV)RExC_npar);*/
9400 for ( paren=0 ; paren < RExC_npar ; paren++ ) {
9401 if ( RExC_open_parens[paren] >= opnd ) {
9402 /*DEBUG_PARSE_FMT("open"," - %d",size);*/
9403 RExC_open_parens[paren] += size;
9405 /*DEBUG_PARSE_FMT("open"," - %s","ok");*/
9407 if ( RExC_close_parens[paren] >= opnd ) {
9408 /*DEBUG_PARSE_FMT("close"," - %d",size);*/
9409 RExC_close_parens[paren] += size;
9411 /*DEBUG_PARSE_FMT("close"," - %s","ok");*/
9416 while (src > opnd) {
9417 StructCopy(--src, --dst, regnode);
9418 #ifdef RE_TRACK_PATTERN_OFFSETS
9419 if (RExC_offsets) { /* MJD 20010112 */
9420 MJD_OFFSET_DEBUG(("%s(%d): (op %s) %s copy %"UVuf" -> %"UVuf" (max %"UVuf").\n",
9424 (UV)(dst - RExC_emit_start) > RExC_offsets[0]
9425 ? "Overwriting end of array!\n" : "OK",
9426 (UV)(src - RExC_emit_start),
9427 (UV)(dst - RExC_emit_start),
9428 (UV)RExC_offsets[0]));
9429 Set_Node_Offset_To_R(dst-RExC_emit_start, Node_Offset(src));
9430 Set_Node_Length_To_R(dst-RExC_emit_start, Node_Length(src));
9436 place = opnd; /* Op node, where operand used to be. */
9437 #ifdef RE_TRACK_PATTERN_OFFSETS
9438 if (RExC_offsets) { /* MJD */
9439 MJD_OFFSET_DEBUG(("%s(%d): (op %s) %s %"UVuf" <- %"UVuf" (max %"UVuf").\n",
9443 (UV)(place - RExC_emit_start) > RExC_offsets[0]
9444 ? "Overwriting end of array!\n" : "OK",
9445 (UV)(place - RExC_emit_start),
9446 (UV)(RExC_parse - RExC_start),
9447 (UV)RExC_offsets[0]));
9448 Set_Node_Offset(place, RExC_parse);
9449 Set_Node_Length(place, 1);
9452 src = NEXTOPER(place);
9453 FILL_ADVANCE_NODE(place, op);
9454 REH_CALL_COMP_NODE_HOOK(pRExC_state->rx, (place) - 1);
9455 Zero(src, offset, regnode);
9459 - regtail - set the next-pointer at the end of a node chain of p to val.
9460 - SEE ALSO: regtail_study
9462 /* TODO: All three parms should be const */
9464 S_regtail(pTHX_ RExC_state_t *pRExC_state, regnode *p, const regnode *val,U32 depth)
9467 register regnode *scan;
9468 GET_RE_DEBUG_FLAGS_DECL;
9470 PERL_ARGS_ASSERT_REGTAIL;
9472 PERL_UNUSED_ARG(depth);
9478 /* Find last node. */
9481 regnode * const temp = regnext(scan);
9483 SV * const mysv=sv_newmortal();
9484 DEBUG_PARSE_MSG((scan==p ? "tail" : ""));
9485 regprop(RExC_rx, mysv, scan);
9486 PerlIO_printf(Perl_debug_log, "~ %s (%d) %s %s\n",
9487 SvPV_nolen_const(mysv), REG_NODE_NUM(scan),
9488 (temp == NULL ? "->" : ""),
9489 (temp == NULL ? PL_reg_name[OP(val)] : "")
9497 if (reg_off_by_arg[OP(scan)]) {
9498 ARG_SET(scan, val - scan);
9501 NEXT_OFF(scan) = val - scan;
9507 - regtail_study - set the next-pointer at the end of a node chain of p to val.
9508 - Look for optimizable sequences at the same time.
9509 - currently only looks for EXACT chains.
9511 This is experimental code. The idea is to use this routine to perform
9512 in place optimizations on branches and groups as they are constructed,
9513 with the long term intention of removing optimization from study_chunk so
9514 that it is purely analytical.
9516 Currently only used when in DEBUG mode. The macro REGTAIL_STUDY() is used
9517 to control which is which.
9520 /* TODO: All four parms should be const */
9523 S_regtail_study(pTHX_ RExC_state_t *pRExC_state, regnode *p, const regnode *val,U32 depth)
9526 register regnode *scan;
9528 #ifdef EXPERIMENTAL_INPLACESCAN
9531 GET_RE_DEBUG_FLAGS_DECL;
9533 PERL_ARGS_ASSERT_REGTAIL_STUDY;
9539 /* Find last node. */
9543 regnode * const temp = regnext(scan);
9544 #ifdef EXPERIMENTAL_INPLACESCAN
9545 if (PL_regkind[OP(scan)] == EXACT)
9546 if (join_exact(pRExC_state,scan,&min,1,val,depth+1))
9555 if( exact == PSEUDO )
9557 else if ( exact != OP(scan) )
9566 SV * const mysv=sv_newmortal();
9567 DEBUG_PARSE_MSG((scan==p ? "tsdy" : ""));
9568 regprop(RExC_rx, mysv, scan);
9569 PerlIO_printf(Perl_debug_log, "~ %s (%d) -> %s\n",
9570 SvPV_nolen_const(mysv),
9572 PL_reg_name[exact]);
9579 SV * const mysv_val=sv_newmortal();
9580 DEBUG_PARSE_MSG("");
9581 regprop(RExC_rx, mysv_val, val);
9582 PerlIO_printf(Perl_debug_log, "~ attach to %s (%"IVdf") offset to %"IVdf"\n",
9583 SvPV_nolen_const(mysv_val),
9584 (IV)REG_NODE_NUM(val),
9588 if (reg_off_by_arg[OP(scan)]) {
9589 ARG_SET(scan, val - scan);
9592 NEXT_OFF(scan) = val - scan;
9600 - regdump - dump a regexp onto Perl_debug_log in vaguely comprehensible form
9604 S_regdump_extflags(pTHX_ const char *lead, const U32 flags)
9610 for (bit=0; bit<32; bit++) {
9611 if (flags & (1<<bit)) {
9612 if ((1<<bit) & RXf_PMf_CHARSET) { /* Output separately, below */
9616 PerlIO_printf(Perl_debug_log, "%s",lead);
9617 PerlIO_printf(Perl_debug_log, "%s ",PL_reg_extflags_name[bit]);
9620 if ((cs = get_regex_charset(flags)) != REGEX_DEPENDS_CHARSET) {
9621 if (!set++ && lead) {
9622 PerlIO_printf(Perl_debug_log, "%s",lead);
9625 case REGEX_UNICODE_CHARSET:
9626 PerlIO_printf(Perl_debug_log, "UNICODE");
9628 case REGEX_LOCALE_CHARSET:
9629 PerlIO_printf(Perl_debug_log, "LOCALE");
9631 case REGEX_ASCII_RESTRICTED_CHARSET:
9632 PerlIO_printf(Perl_debug_log, "ASCII-RESTRICTED");
9635 PerlIO_printf(Perl_debug_log, "UNKNOWN CHARACTER SET");
9641 PerlIO_printf(Perl_debug_log, "\n");
9643 PerlIO_printf(Perl_debug_log, "%s[none-set]\n",lead);
9649 Perl_regdump(pTHX_ const regexp *r)
9653 SV * const sv = sv_newmortal();
9654 SV *dsv= sv_newmortal();
9656 GET_RE_DEBUG_FLAGS_DECL;
9658 PERL_ARGS_ASSERT_REGDUMP;
9660 (void)dumpuntil(r, ri->program, ri->program + 1, NULL, NULL, sv, 0, 0);
9662 /* Header fields of interest. */
9663 if (r->anchored_substr) {
9664 RE_PV_QUOTED_DECL(s, 0, dsv, SvPVX_const(r->anchored_substr),
9665 RE_SV_DUMPLEN(r->anchored_substr), 30);
9666 PerlIO_printf(Perl_debug_log,
9667 "anchored %s%s at %"IVdf" ",
9668 s, RE_SV_TAIL(r->anchored_substr),
9669 (IV)r->anchored_offset);
9670 } else if (r->anchored_utf8) {
9671 RE_PV_QUOTED_DECL(s, 1, dsv, SvPVX_const(r->anchored_utf8),
9672 RE_SV_DUMPLEN(r->anchored_utf8), 30);
9673 PerlIO_printf(Perl_debug_log,
9674 "anchored utf8 %s%s at %"IVdf" ",
9675 s, RE_SV_TAIL(r->anchored_utf8),
9676 (IV)r->anchored_offset);
9678 if (r->float_substr) {
9679 RE_PV_QUOTED_DECL(s, 0, dsv, SvPVX_const(r->float_substr),
9680 RE_SV_DUMPLEN(r->float_substr), 30);
9681 PerlIO_printf(Perl_debug_log,
9682 "floating %s%s at %"IVdf"..%"UVuf" ",
9683 s, RE_SV_TAIL(r->float_substr),
9684 (IV)r->float_min_offset, (UV)r->float_max_offset);
9685 } else if (r->float_utf8) {
9686 RE_PV_QUOTED_DECL(s, 1, dsv, SvPVX_const(r->float_utf8),
9687 RE_SV_DUMPLEN(r->float_utf8), 30);
9688 PerlIO_printf(Perl_debug_log,
9689 "floating utf8 %s%s at %"IVdf"..%"UVuf" ",
9690 s, RE_SV_TAIL(r->float_utf8),
9691 (IV)r->float_min_offset, (UV)r->float_max_offset);
9693 if (r->check_substr || r->check_utf8)
9694 PerlIO_printf(Perl_debug_log,
9696 (r->check_substr == r->float_substr
9697 && r->check_utf8 == r->float_utf8
9698 ? "(checking floating" : "(checking anchored"));
9699 if (r->extflags & RXf_NOSCAN)
9700 PerlIO_printf(Perl_debug_log, " noscan");
9701 if (r->extflags & RXf_CHECK_ALL)
9702 PerlIO_printf(Perl_debug_log, " isall");
9703 if (r->check_substr || r->check_utf8)
9704 PerlIO_printf(Perl_debug_log, ") ");
9706 if (ri->regstclass) {
9707 regprop(r, sv, ri->regstclass);
9708 PerlIO_printf(Perl_debug_log, "stclass %s ", SvPVX_const(sv));
9710 if (r->extflags & RXf_ANCH) {
9711 PerlIO_printf(Perl_debug_log, "anchored");
9712 if (r->extflags & RXf_ANCH_BOL)
9713 PerlIO_printf(Perl_debug_log, "(BOL)");
9714 if (r->extflags & RXf_ANCH_MBOL)
9715 PerlIO_printf(Perl_debug_log, "(MBOL)");
9716 if (r->extflags & RXf_ANCH_SBOL)
9717 PerlIO_printf(Perl_debug_log, "(SBOL)");
9718 if (r->extflags & RXf_ANCH_GPOS)
9719 PerlIO_printf(Perl_debug_log, "(GPOS)");
9720 PerlIO_putc(Perl_debug_log, ' ');
9722 if (r->extflags & RXf_GPOS_SEEN)
9723 PerlIO_printf(Perl_debug_log, "GPOS:%"UVuf" ", (UV)r->gofs);
9724 if (r->intflags & PREGf_SKIP)
9725 PerlIO_printf(Perl_debug_log, "plus ");
9726 if (r->intflags & PREGf_IMPLICIT)
9727 PerlIO_printf(Perl_debug_log, "implicit ");
9728 PerlIO_printf(Perl_debug_log, "minlen %"IVdf" ", (IV)r->minlen);
9729 if (r->extflags & RXf_EVAL_SEEN)
9730 PerlIO_printf(Perl_debug_log, "with eval ");
9731 PerlIO_printf(Perl_debug_log, "\n");
9732 DEBUG_FLAGS_r(regdump_extflags("r->extflags: ",r->extflags));
9734 PERL_ARGS_ASSERT_REGDUMP;
9735 PERL_UNUSED_CONTEXT;
9737 #endif /* DEBUGGING */
9741 - regprop - printable representation of opcode
9743 #define EMIT_ANYOF_TEST_SEPARATOR(do_sep,sv,flags) \
9746 Perl_sv_catpvf(aTHX_ sv,"%s][%s",PL_colors[1],PL_colors[0]); \
9747 if (flags & ANYOF_INVERT) \
9748 /*make sure the invert info is in each */ \
9749 sv_catpvs(sv, "^"); \
9755 Perl_regprop(pTHX_ const regexp *prog, SV *sv, const regnode *o)
9760 RXi_GET_DECL(prog,progi);
9761 GET_RE_DEBUG_FLAGS_DECL;
9763 PERL_ARGS_ASSERT_REGPROP;
9767 if (OP(o) > REGNODE_MAX) /* regnode.type is unsigned */
9768 /* It would be nice to FAIL() here, but this may be called from
9769 regexec.c, and it would be hard to supply pRExC_state. */
9770 Perl_croak(aTHX_ "Corrupted regexp opcode %d > %d", (int)OP(o), (int)REGNODE_MAX);
9771 sv_catpv(sv, PL_reg_name[OP(o)]); /* Take off const! */
9773 k = PL_regkind[OP(o)];
9777 /* Using is_utf8_string() (via PERL_PV_UNI_DETECT)
9778 * is a crude hack but it may be the best for now since
9779 * we have no flag "this EXACTish node was UTF-8"
9781 pv_pretty(sv, STRING(o), STR_LEN(o), 60, PL_colors[0], PL_colors[1],
9782 PERL_PV_ESCAPE_UNI_DETECT |
9783 PERL_PV_ESCAPE_NONASCII |
9784 PERL_PV_PRETTY_ELLIPSES |
9785 PERL_PV_PRETTY_LTGT |
9786 PERL_PV_PRETTY_NOCLEAR
9788 } else if (k == TRIE) {
9789 /* print the details of the trie in dumpuntil instead, as
9790 * progi->data isn't available here */
9791 const char op = OP(o);
9792 const U32 n = ARG(o);
9793 const reg_ac_data * const ac = IS_TRIE_AC(op) ?
9794 (reg_ac_data *)progi->data->data[n] :
9796 const reg_trie_data * const trie
9797 = (reg_trie_data*)progi->data->data[!IS_TRIE_AC(op) ? n : ac->trie];
9799 Perl_sv_catpvf(aTHX_ sv, "-%s",PL_reg_name[o->flags]);
9800 DEBUG_TRIE_COMPILE_r(
9801 Perl_sv_catpvf(aTHX_ sv,
9802 "<S:%"UVuf"/%"IVdf" W:%"UVuf" L:%"UVuf"/%"UVuf" C:%"UVuf"/%"UVuf">",
9803 (UV)trie->startstate,
9804 (IV)trie->statecount-1, /* -1 because of the unused 0 element */
9805 (UV)trie->wordcount,
9808 (UV)TRIE_CHARCOUNT(trie),
9809 (UV)trie->uniquecharcount
9812 if ( IS_ANYOF_TRIE(op) || trie->bitmap ) {
9814 int rangestart = -1;
9815 U8* bitmap = IS_ANYOF_TRIE(op) ? (U8*)ANYOF_BITMAP(o) : (U8*)TRIE_BITMAP(trie);
9817 for (i = 0; i <= 256; i++) {
9818 if (i < 256 && BITMAP_TEST(bitmap,i)) {
9819 if (rangestart == -1)
9821 } else if (rangestart != -1) {
9822 if (i <= rangestart + 3)
9823 for (; rangestart < i; rangestart++)
9824 put_byte(sv, rangestart);
9826 put_byte(sv, rangestart);
9828 put_byte(sv, i - 1);
9836 } else if (k == CURLY) {
9837 if (OP(o) == CURLYM || OP(o) == CURLYN || OP(o) == CURLYX)
9838 Perl_sv_catpvf(aTHX_ sv, "[%d]", o->flags); /* Parenth number */
9839 Perl_sv_catpvf(aTHX_ sv, " {%d,%d}", ARG1(o), ARG2(o));
9841 else if (k == WHILEM && o->flags) /* Ordinal/of */
9842 Perl_sv_catpvf(aTHX_ sv, "[%d/%d]", o->flags & 0xf, o->flags>>4);
9843 else if (k == REF || k == OPEN || k == CLOSE || k == GROUPP || OP(o)==ACCEPT) {
9844 Perl_sv_catpvf(aTHX_ sv, "%d", (int)ARG(o)); /* Parenth number */
9845 if ( RXp_PAREN_NAMES(prog) ) {
9846 if ( k != REF || (OP(o) < NREF)) {
9847 AV *list= MUTABLE_AV(progi->data->data[progi->name_list_idx]);
9848 SV **name= av_fetch(list, ARG(o), 0 );
9850 Perl_sv_catpvf(aTHX_ sv, " '%"SVf"'", SVfARG(*name));
9853 AV *list= MUTABLE_AV(progi->data->data[ progi->name_list_idx ]);
9854 SV *sv_dat= MUTABLE_SV(progi->data->data[ ARG( o ) ]);
9855 I32 *nums=(I32*)SvPVX(sv_dat);
9856 SV **name= av_fetch(list, nums[0], 0 );
9859 for ( n=0; n<SvIVX(sv_dat); n++ ) {
9860 Perl_sv_catpvf(aTHX_ sv, "%s%"IVdf,
9861 (n ? "," : ""), (IV)nums[n]);
9863 Perl_sv_catpvf(aTHX_ sv, " '%"SVf"'", SVfARG(*name));
9867 } else if (k == GOSUB)
9868 Perl_sv_catpvf(aTHX_ sv, "%d[%+d]", (int)ARG(o),(int)ARG2L(o)); /* Paren and offset */
9869 else if (k == VERB) {
9871 Perl_sv_catpvf(aTHX_ sv, ":%"SVf,
9872 SVfARG((MUTABLE_SV(progi->data->data[ ARG( o ) ]))));
9873 } else if (k == LOGICAL)
9874 Perl_sv_catpvf(aTHX_ sv, "[%d]", o->flags); /* 2: embedded, otherwise 1 */
9875 else if (k == FOLDCHAR)
9876 Perl_sv_catpvf(aTHX_ sv, "[0x%"UVXf"]", PTR2UV(ARG(o)) );
9877 else if (k == ANYOF) {
9878 int i, rangestart = -1;
9879 const U8 flags = ANYOF_FLAGS(o);
9882 /* Should be synchronized with * ANYOF_ #xdefines in regcomp.h */
9883 static const char * const anyofs[] = {
9916 if (flags & ANYOF_LOCALE)
9917 sv_catpvs(sv, "{loc}");
9918 if (flags & ANYOF_LOC_NONBITMAP_FOLD)
9919 sv_catpvs(sv, "{i}");
9920 Perl_sv_catpvf(aTHX_ sv, "[%s", PL_colors[0]);
9921 if (flags & ANYOF_INVERT)
9924 /* output what the standard cp 0-255 bitmap matches */
9925 for (i = 0; i <= 256; i++) {
9926 if (i < 256 && ANYOF_BITMAP_TEST(o,i)) {
9927 if (rangestart == -1)
9929 } else if (rangestart != -1) {
9930 if (i <= rangestart + 3)
9931 for (; rangestart < i; rangestart++)
9932 put_byte(sv, rangestart);
9934 put_byte(sv, rangestart);
9936 put_byte(sv, i - 1);
9943 EMIT_ANYOF_TEST_SEPARATOR(do_sep,sv,flags);
9944 /* output any special charclass tests (used entirely under use locale) */
9945 if (ANYOF_CLASS_TEST_ANY_SET(o))
9946 for (i = 0; i < (int)(sizeof(anyofs)/sizeof(char*)); i++)
9947 if (ANYOF_CLASS_TEST(o,i)) {
9948 sv_catpv(sv, anyofs[i]);
9952 EMIT_ANYOF_TEST_SEPARATOR(do_sep,sv,flags);
9954 if (flags & ANYOF_NON_UTF8_LATIN1_ALL) {
9955 sv_catpvs(sv, "{non-utf8-latin1-all}");
9958 /* output information about the unicode matching */
9959 if (flags & ANYOF_UNICODE_ALL)
9960 sv_catpvs(sv, "{unicode_all}");
9961 else if (flags & ANYOF_UTF8)
9962 sv_catpvs(sv, "{unicode}");
9963 if (flags & ANYOF_NONBITMAP_NON_UTF8)
9964 sv_catpvs(sv, "{outside bitmap}");
9968 SV * const sw = regclass_swash(prog, o, FALSE, &lv, 0);
9972 U8 s[UTF8_MAXBYTES_CASE+1];
9974 for (i = 0; i <= 256; i++) { /* just the first 256 */
9975 uvchr_to_utf8(s, i);
9977 if (i < 256 && swash_fetch(sw, s, TRUE)) {
9978 if (rangestart == -1)
9980 } else if (rangestart != -1) {
9981 if (i <= rangestart + 3)
9982 for (; rangestart < i; rangestart++) {
9983 const U8 * const e = uvchr_to_utf8(s,rangestart);
9985 for(p = s; p < e; p++)
9989 const U8 *e = uvchr_to_utf8(s,rangestart);
9991 for (p = s; p < e; p++)
9994 e = uvchr_to_utf8(s, i-1);
9995 for (p = s; p < e; p++)
10002 sv_catpvs(sv, "..."); /* et cetera */
10006 char *s = savesvpv(lv);
10007 char * const origs = s;
10009 while (*s && *s != '\n')
10013 const char * const t = ++s;
10031 Perl_sv_catpvf(aTHX_ sv, "%s]", PL_colors[1]);
10033 else if (k == BRANCHJ && (OP(o) == UNLESSM || OP(o) == IFMATCH))
10034 Perl_sv_catpvf(aTHX_ sv, "[%d]", -(o->flags));
10036 PERL_UNUSED_CONTEXT;
10037 PERL_UNUSED_ARG(sv);
10038 PERL_UNUSED_ARG(o);
10039 PERL_UNUSED_ARG(prog);
10040 #endif /* DEBUGGING */
10044 Perl_re_intuit_string(pTHX_ REGEXP * const r)
10045 { /* Assume that RE_INTUIT is set */
10047 struct regexp *const prog = (struct regexp *)SvANY(r);
10048 GET_RE_DEBUG_FLAGS_DECL;
10050 PERL_ARGS_ASSERT_RE_INTUIT_STRING;
10051 PERL_UNUSED_CONTEXT;
10055 const char * const s = SvPV_nolen_const(prog->check_substr
10056 ? prog->check_substr : prog->check_utf8);
10058 if (!PL_colorset) reginitcolors();
10059 PerlIO_printf(Perl_debug_log,
10060 "%sUsing REx %ssubstr:%s \"%s%.60s%s%s\"\n",
10062 prog->check_substr ? "" : "utf8 ",
10063 PL_colors[5],PL_colors[0],
10066 (strlen(s) > 60 ? "..." : ""));
10069 return prog->check_substr ? prog->check_substr : prog->check_utf8;
10075 handles refcounting and freeing the perl core regexp structure. When
10076 it is necessary to actually free the structure the first thing it
10077 does is call the 'free' method of the regexp_engine associated to
10078 the regexp, allowing the handling of the void *pprivate; member
10079 first. (This routine is not overridable by extensions, which is why
10080 the extensions free is called first.)
10082 See regdupe and regdupe_internal if you change anything here.
10084 #ifndef PERL_IN_XSUB_RE
10086 Perl_pregfree(pTHX_ REGEXP *r)
10092 Perl_pregfree2(pTHX_ REGEXP *rx)
10095 struct regexp *const r = (struct regexp *)SvANY(rx);
10096 GET_RE_DEBUG_FLAGS_DECL;
10098 PERL_ARGS_ASSERT_PREGFREE2;
10100 if (r->mother_re) {
10101 ReREFCNT_dec(r->mother_re);
10103 CALLREGFREE_PVT(rx); /* free the private data */
10104 SvREFCNT_dec(RXp_PAREN_NAMES(r));
10107 SvREFCNT_dec(r->anchored_substr);
10108 SvREFCNT_dec(r->anchored_utf8);
10109 SvREFCNT_dec(r->float_substr);
10110 SvREFCNT_dec(r->float_utf8);
10111 Safefree(r->substrs);
10113 RX_MATCH_COPY_FREE(rx);
10114 #ifdef PERL_OLD_COPY_ON_WRITE
10115 SvREFCNT_dec(r->saved_copy);
10122 This is a hacky workaround to the structural issue of match results
10123 being stored in the regexp structure which is in turn stored in
10124 PL_curpm/PL_reg_curpm. The problem is that due to qr// the pattern
10125 could be PL_curpm in multiple contexts, and could require multiple
10126 result sets being associated with the pattern simultaneously, such
10127 as when doing a recursive match with (??{$qr})
10129 The solution is to make a lightweight copy of the regexp structure
10130 when a qr// is returned from the code executed by (??{$qr}) this
10131 lightweight copy doesn't actually own any of its data except for
10132 the starp/end and the actual regexp structure itself.
10138 Perl_reg_temp_copy (pTHX_ REGEXP *ret_x, REGEXP *rx)
10140 struct regexp *ret;
10141 struct regexp *const r = (struct regexp *)SvANY(rx);
10142 register const I32 npar = r->nparens+1;
10144 PERL_ARGS_ASSERT_REG_TEMP_COPY;
10147 ret_x = (REGEXP*) newSV_type(SVt_REGEXP);
10148 ret = (struct regexp *)SvANY(ret_x);
10150 (void)ReREFCNT_inc(rx);
10151 /* We can take advantage of the existing "copied buffer" mechanism in SVs
10152 by pointing directly at the buffer, but flagging that the allocated
10153 space in the copy is zero. As we've just done a struct copy, it's now
10154 a case of zero-ing that, rather than copying the current length. */
10155 SvPV_set(ret_x, RX_WRAPPED(rx));
10156 SvFLAGS(ret_x) |= SvFLAGS(rx) & (SVf_POK|SVp_POK|SVf_UTF8);
10157 memcpy(&(ret->xpv_cur), &(r->xpv_cur),
10158 sizeof(regexp) - STRUCT_OFFSET(regexp, xpv_cur));
10159 SvLEN_set(ret_x, 0);
10160 SvSTASH_set(ret_x, NULL);
10161 SvMAGIC_set(ret_x, NULL);
10162 Newx(ret->offs, npar, regexp_paren_pair);
10163 Copy(r->offs, ret->offs, npar, regexp_paren_pair);
10165 Newx(ret->substrs, 1, struct reg_substr_data);
10166 StructCopy(r->substrs, ret->substrs, struct reg_substr_data);
10168 SvREFCNT_inc_void(ret->anchored_substr);
10169 SvREFCNT_inc_void(ret->anchored_utf8);
10170 SvREFCNT_inc_void(ret->float_substr);
10171 SvREFCNT_inc_void(ret->float_utf8);
10173 /* check_substr and check_utf8, if non-NULL, point to either their
10174 anchored or float namesakes, and don't hold a second reference. */
10176 RX_MATCH_COPIED_off(ret_x);
10177 #ifdef PERL_OLD_COPY_ON_WRITE
10178 ret->saved_copy = NULL;
10180 ret->mother_re = rx;
10186 /* regfree_internal()
10188 Free the private data in a regexp. This is overloadable by
10189 extensions. Perl takes care of the regexp structure in pregfree(),
10190 this covers the *pprivate pointer which technically perl doesn't
10191 know about, however of course we have to handle the
10192 regexp_internal structure when no extension is in use.
10194 Note this is called before freeing anything in the regexp
10199 Perl_regfree_internal(pTHX_ REGEXP * const rx)
10202 struct regexp *const r = (struct regexp *)SvANY(rx);
10203 RXi_GET_DECL(r,ri);
10204 GET_RE_DEBUG_FLAGS_DECL;
10206 PERL_ARGS_ASSERT_REGFREE_INTERNAL;
10212 SV *dsv= sv_newmortal();
10213 RE_PV_QUOTED_DECL(s, RX_UTF8(rx),
10214 dsv, RX_PRECOMP(rx), RX_PRELEN(rx), 60);
10215 PerlIO_printf(Perl_debug_log,"%sFreeing REx:%s %s\n",
10216 PL_colors[4],PL_colors[5],s);
10219 #ifdef RE_TRACK_PATTERN_OFFSETS
10221 Safefree(ri->u.offsets); /* 20010421 MJD */
10224 int n = ri->data->count;
10225 PAD* new_comppad = NULL;
10230 /* If you add a ->what type here, update the comment in regcomp.h */
10231 switch (ri->data->what[n]) {
10236 SvREFCNT_dec(MUTABLE_SV(ri->data->data[n]));
10239 Safefree(ri->data->data[n]);
10242 new_comppad = MUTABLE_AV(ri->data->data[n]);
10245 if (new_comppad == NULL)
10246 Perl_croak(aTHX_ "panic: pregfree comppad");
10247 PAD_SAVE_LOCAL(old_comppad,
10248 /* Watch out for global destruction's random ordering. */
10249 (SvTYPE(new_comppad) == SVt_PVAV) ? new_comppad : NULL
10252 refcnt = OpREFCNT_dec((OP_4tree*)ri->data->data[n]);
10255 op_free((OP_4tree*)ri->data->data[n]);
10257 PAD_RESTORE_LOCAL(old_comppad);
10258 SvREFCNT_dec(MUTABLE_SV(new_comppad));
10259 new_comppad = NULL;
10264 { /* Aho Corasick add-on structure for a trie node.
10265 Used in stclass optimization only */
10267 reg_ac_data *aho=(reg_ac_data*)ri->data->data[n];
10269 refcount = --aho->refcount;
10272 PerlMemShared_free(aho->states);
10273 PerlMemShared_free(aho->fail);
10274 /* do this last!!!! */
10275 PerlMemShared_free(ri->data->data[n]);
10276 PerlMemShared_free(ri->regstclass);
10282 /* trie structure. */
10284 reg_trie_data *trie=(reg_trie_data*)ri->data->data[n];
10286 refcount = --trie->refcount;
10289 PerlMemShared_free(trie->charmap);
10290 PerlMemShared_free(trie->states);
10291 PerlMemShared_free(trie->trans);
10293 PerlMemShared_free(trie->bitmap);
10295 PerlMemShared_free(trie->jump);
10296 PerlMemShared_free(trie->wordinfo);
10297 /* do this last!!!! */
10298 PerlMemShared_free(ri->data->data[n]);
10303 Perl_croak(aTHX_ "panic: regfree data code '%c'", ri->data->what[n]);
10306 Safefree(ri->data->what);
10307 Safefree(ri->data);
10313 #define av_dup_inc(s,t) MUTABLE_AV(sv_dup_inc((const SV *)s,t))
10314 #define hv_dup_inc(s,t) MUTABLE_HV(sv_dup_inc((const SV *)s,t))
10315 #define SAVEPVN(p,n) ((p) ? savepvn(p,n) : NULL)
10318 re_dup - duplicate a regexp.
10320 This routine is expected to clone a given regexp structure. It is only
10321 compiled under USE_ITHREADS.
10323 After all of the core data stored in struct regexp is duplicated
10324 the regexp_engine.dupe method is used to copy any private data
10325 stored in the *pprivate pointer. This allows extensions to handle
10326 any duplication it needs to do.
10328 See pregfree() and regfree_internal() if you change anything here.
10330 #if defined(USE_ITHREADS)
10331 #ifndef PERL_IN_XSUB_RE
10333 Perl_re_dup_guts(pTHX_ const REGEXP *sstr, REGEXP *dstr, CLONE_PARAMS *param)
10337 const struct regexp *r = (const struct regexp *)SvANY(sstr);
10338 struct regexp *ret = (struct regexp *)SvANY(dstr);
10340 PERL_ARGS_ASSERT_RE_DUP_GUTS;
10342 npar = r->nparens+1;
10343 Newx(ret->offs, npar, regexp_paren_pair);
10344 Copy(r->offs, ret->offs, npar, regexp_paren_pair);
10346 /* no need to copy these */
10347 Newx(ret->swap, npar, regexp_paren_pair);
10350 if (ret->substrs) {
10351 /* Do it this way to avoid reading from *r after the StructCopy().
10352 That way, if any of the sv_dup_inc()s dislodge *r from the L1
10353 cache, it doesn't matter. */
10354 const bool anchored = r->check_substr
10355 ? r->check_substr == r->anchored_substr
10356 : r->check_utf8 == r->anchored_utf8;
10357 Newx(ret->substrs, 1, struct reg_substr_data);
10358 StructCopy(r->substrs, ret->substrs, struct reg_substr_data);
10360 ret->anchored_substr = sv_dup_inc(ret->anchored_substr, param);
10361 ret->anchored_utf8 = sv_dup_inc(ret->anchored_utf8, param);
10362 ret->float_substr = sv_dup_inc(ret->float_substr, param);
10363 ret->float_utf8 = sv_dup_inc(ret->float_utf8, param);
10365 /* check_substr and check_utf8, if non-NULL, point to either their
10366 anchored or float namesakes, and don't hold a second reference. */
10368 if (ret->check_substr) {
10370 assert(r->check_utf8 == r->anchored_utf8);
10371 ret->check_substr = ret->anchored_substr;
10372 ret->check_utf8 = ret->anchored_utf8;
10374 assert(r->check_substr == r->float_substr);
10375 assert(r->check_utf8 == r->float_utf8);
10376 ret->check_substr = ret->float_substr;
10377 ret->check_utf8 = ret->float_utf8;
10379 } else if (ret->check_utf8) {
10381 ret->check_utf8 = ret->anchored_utf8;
10383 ret->check_utf8 = ret->float_utf8;
10388 RXp_PAREN_NAMES(ret) = hv_dup_inc(RXp_PAREN_NAMES(ret), param);
10391 RXi_SET(ret,CALLREGDUPE_PVT(dstr,param));
10393 if (RX_MATCH_COPIED(dstr))
10394 ret->subbeg = SAVEPVN(ret->subbeg, ret->sublen);
10396 ret->subbeg = NULL;
10397 #ifdef PERL_OLD_COPY_ON_WRITE
10398 ret->saved_copy = NULL;
10401 if (ret->mother_re) {
10402 if (SvPVX_const(dstr) == SvPVX_const(ret->mother_re)) {
10403 /* Our storage points directly to our mother regexp, but that's
10404 1: a buffer in a different thread
10405 2: something we no longer hold a reference on
10406 so we need to copy it locally. */
10407 /* Note we need to sue SvCUR() on our mother_re, because it, in
10408 turn, may well be pointing to its own mother_re. */
10409 SvPV_set(dstr, SAVEPVN(SvPVX_const(ret->mother_re),
10410 SvCUR(ret->mother_re)+1));
10411 SvLEN_set(dstr, SvCUR(ret->mother_re)+1);
10413 ret->mother_re = NULL;
10417 #endif /* PERL_IN_XSUB_RE */
10422 This is the internal complement to regdupe() which is used to copy
10423 the structure pointed to by the *pprivate pointer in the regexp.
10424 This is the core version of the extension overridable cloning hook.
10425 The regexp structure being duplicated will be copied by perl prior
10426 to this and will be provided as the regexp *r argument, however
10427 with the /old/ structures pprivate pointer value. Thus this routine
10428 may override any copying normally done by perl.
10430 It returns a pointer to the new regexp_internal structure.
10434 Perl_regdupe_internal(pTHX_ REGEXP * const rx, CLONE_PARAMS *param)
10437 struct regexp *const r = (struct regexp *)SvANY(rx);
10438 regexp_internal *reti;
10440 RXi_GET_DECL(r,ri);
10442 PERL_ARGS_ASSERT_REGDUPE_INTERNAL;
10444 npar = r->nparens+1;
10447 Newxc(reti, sizeof(regexp_internal) + len*sizeof(regnode), char, regexp_internal);
10448 Copy(ri->program, reti->program, len+1, regnode);
10451 reti->regstclass = NULL;
10454 struct reg_data *d;
10455 const int count = ri->data->count;
10458 Newxc(d, sizeof(struct reg_data) + count*sizeof(void *),
10459 char, struct reg_data);
10460 Newx(d->what, count, U8);
10463 for (i = 0; i < count; i++) {
10464 d->what[i] = ri->data->what[i];
10465 switch (d->what[i]) {
10466 /* legal options are one of: sSfpontTua
10467 see also regcomp.h and pregfree() */
10468 case 'a': /* actually an AV, but the dup function is identical. */
10471 case 'p': /* actually an AV, but the dup function is identical. */
10472 case 'u': /* actually an HV, but the dup function is identical. */
10473 d->data[i] = sv_dup_inc((const SV *)ri->data->data[i], param);
10476 /* This is cheating. */
10477 Newx(d->data[i], 1, struct regnode_charclass_class);
10478 StructCopy(ri->data->data[i], d->data[i],
10479 struct regnode_charclass_class);
10480 reti->regstclass = (regnode*)d->data[i];
10483 /* Compiled op trees are readonly and in shared memory,
10484 and can thus be shared without duplication. */
10486 d->data[i] = (void*)OpREFCNT_inc((OP*)ri->data->data[i]);
10490 /* Trie stclasses are readonly and can thus be shared
10491 * without duplication. We free the stclass in pregfree
10492 * when the corresponding reg_ac_data struct is freed.
10494 reti->regstclass= ri->regstclass;
10498 ((reg_trie_data*)ri->data->data[i])->refcount++;
10502 d->data[i] = ri->data->data[i];
10505 Perl_croak(aTHX_ "panic: re_dup unknown data code '%c'", ri->data->what[i]);
10514 reti->name_list_idx = ri->name_list_idx;
10516 #ifdef RE_TRACK_PATTERN_OFFSETS
10517 if (ri->u.offsets) {
10518 Newx(reti->u.offsets, 2*len+1, U32);
10519 Copy(ri->u.offsets, reti->u.offsets, 2*len+1, U32);
10522 SetProgLen(reti,len);
10525 return (void*)reti;
10528 #endif /* USE_ITHREADS */
10530 #ifndef PERL_IN_XSUB_RE
10533 - regnext - dig the "next" pointer out of a node
10536 Perl_regnext(pTHX_ register regnode *p)
10539 register I32 offset;
10544 if (OP(p) > REGNODE_MAX) { /* regnode.type is unsigned */
10545 Perl_croak(aTHX_ "Corrupted regexp opcode %d > %d", (int)OP(p), (int)REGNODE_MAX);
10548 offset = (reg_off_by_arg[OP(p)] ? ARG(p) : NEXT_OFF(p));
10557 S_re_croak2(pTHX_ const char* pat1,const char* pat2,...)
10560 STRLEN l1 = strlen(pat1);
10561 STRLEN l2 = strlen(pat2);
10564 const char *message;
10566 PERL_ARGS_ASSERT_RE_CROAK2;
10572 Copy(pat1, buf, l1 , char);
10573 Copy(pat2, buf + l1, l2 , char);
10574 buf[l1 + l2] = '\n';
10575 buf[l1 + l2 + 1] = '\0';
10577 /* ANSI variant takes additional second argument */
10578 va_start(args, pat2);
10582 msv = vmess(buf, &args);
10584 message = SvPV_const(msv,l1);
10587 Copy(message, buf, l1 , char);
10588 buf[l1-1] = '\0'; /* Overwrite \n */
10589 Perl_croak(aTHX_ "%s", buf);
10592 /* XXX Here's a total kludge. But we need to re-enter for swash routines. */
10594 #ifndef PERL_IN_XSUB_RE
10596 Perl_save_re_context(pTHX)
10600 struct re_save_state *state;
10602 SAVEVPTR(PL_curcop);
10603 SSGROW(SAVESTACK_ALLOC_FOR_RE_SAVE_STATE + 1);
10605 state = (struct re_save_state *)(PL_savestack + PL_savestack_ix);
10606 PL_savestack_ix += SAVESTACK_ALLOC_FOR_RE_SAVE_STATE;
10607 SSPUSHUV(SAVEt_RE_STATE);
10609 Copy(&PL_reg_state, state, 1, struct re_save_state);
10611 PL_reg_start_tmp = 0;
10612 PL_reg_start_tmpl = 0;
10613 PL_reg_oldsaved = NULL;
10614 PL_reg_oldsavedlen = 0;
10615 PL_reg_maxiter = 0;
10616 PL_reg_leftiter = 0;
10617 PL_reg_poscache = NULL;
10618 PL_reg_poscache_size = 0;
10619 #ifdef PERL_OLD_COPY_ON_WRITE
10623 /* Save $1..$n (#18107: UTF-8 s/(\w+)/uc($1)/e); AMS 20021106. */
10625 const REGEXP * const rx = PM_GETRE(PL_curpm);
10628 for (i = 1; i <= RX_NPARENS(rx); i++) {
10629 char digits[TYPE_CHARS(long)];
10630 const STRLEN len = my_snprintf(digits, sizeof(digits), "%lu", (long)i);
10631 GV *const *const gvp
10632 = (GV**)hv_fetch(PL_defstash, digits, len, 0);
10635 GV * const gv = *gvp;
10636 if (SvTYPE(gv) == SVt_PVGV && GvSV(gv))
10646 clear_re(pTHX_ void *r)
10649 ReREFCNT_dec((REGEXP *)r);
10655 S_put_byte(pTHX_ SV *sv, int c)
10657 PERL_ARGS_ASSERT_PUT_BYTE;
10659 /* Our definition of isPRINT() ignores locales, so only bytes that are
10660 not part of UTF-8 are considered printable. I assume that the same
10661 holds for UTF-EBCDIC.
10662 Also, code point 255 is not printable in either (it's E0 in EBCDIC,
10663 which Wikipedia says:
10665 EO, or Eight Ones, is an 8-bit EBCDIC character code represented as all
10666 ones (binary 1111 1111, hexadecimal FF). It is similar, but not
10667 identical, to the ASCII delete (DEL) or rubout control character.
10668 ) So the old condition can be simplified to !isPRINT(c) */
10671 Perl_sv_catpvf(aTHX_ sv, "\\x%02x", c);
10674 Perl_sv_catpvf(aTHX_ sv, "\\x{%x}", c);
10678 const char string = c;
10679 if (c == '-' || c == ']' || c == '\\' || c == '^')
10680 sv_catpvs(sv, "\\");
10681 sv_catpvn(sv, &string, 1);
10686 #define CLEAR_OPTSTART \
10687 if (optstart) STMT_START { \
10688 DEBUG_OPTIMISE_r(PerlIO_printf(Perl_debug_log, " (%"IVdf" nodes)\n", (IV)(node - optstart))); \
10692 #define DUMPUNTIL(b,e) CLEAR_OPTSTART; node=dumpuntil(r,start,(b),(e),last,sv,indent+1,depth+1);
10694 STATIC const regnode *
10695 S_dumpuntil(pTHX_ const regexp *r, const regnode *start, const regnode *node,
10696 const regnode *last, const regnode *plast,
10697 SV* sv, I32 indent, U32 depth)
10700 register U8 op = PSEUDO; /* Arbitrary non-END op. */
10701 register const regnode *next;
10702 const regnode *optstart= NULL;
10704 RXi_GET_DECL(r,ri);
10705 GET_RE_DEBUG_FLAGS_DECL;
10707 PERL_ARGS_ASSERT_DUMPUNTIL;
10709 #ifdef DEBUG_DUMPUNTIL
10710 PerlIO_printf(Perl_debug_log, "--- %d : %d - %d - %d\n",indent,node-start,
10711 last ? last-start : 0,plast ? plast-start : 0);
10714 if (plast && plast < last)
10717 while (PL_regkind[op] != END && (!last || node < last)) {
10718 /* While that wasn't END last time... */
10721 if (op == CLOSE || op == WHILEM)
10723 next = regnext((regnode *)node);
10726 if (OP(node) == OPTIMIZED) {
10727 if (!optstart && RE_DEBUG_FLAG(RE_DEBUG_COMPILE_OPTIMISE))
10734 regprop(r, sv, node);
10735 PerlIO_printf(Perl_debug_log, "%4"IVdf":%*s%s", (IV)(node - start),
10736 (int)(2*indent + 1), "", SvPVX_const(sv));
10738 if (OP(node) != OPTIMIZED) {
10739 if (next == NULL) /* Next ptr. */
10740 PerlIO_printf(Perl_debug_log, " (0)");
10741 else if (PL_regkind[(U8)op] == BRANCH && PL_regkind[OP(next)] != BRANCH )
10742 PerlIO_printf(Perl_debug_log, " (FAIL)");
10744 PerlIO_printf(Perl_debug_log, " (%"IVdf")", (IV)(next - start));
10745 (void)PerlIO_putc(Perl_debug_log, '\n');
10749 if (PL_regkind[(U8)op] == BRANCHJ) {
10752 register const regnode *nnode = (OP(next) == LONGJMP
10753 ? regnext((regnode *)next)
10755 if (last && nnode > last)
10757 DUMPUNTIL(NEXTOPER(NEXTOPER(node)), nnode);
10760 else if (PL_regkind[(U8)op] == BRANCH) {
10762 DUMPUNTIL(NEXTOPER(node), next);
10764 else if ( PL_regkind[(U8)op] == TRIE ) {
10765 const regnode *this_trie = node;
10766 const char op = OP(node);
10767 const U32 n = ARG(node);
10768 const reg_ac_data * const ac = op>=AHOCORASICK ?
10769 (reg_ac_data *)ri->data->data[n] :
10771 const reg_trie_data * const trie =
10772 (reg_trie_data*)ri->data->data[op<AHOCORASICK ? n : ac->trie];
10774 AV *const trie_words = MUTABLE_AV(ri->data->data[n + TRIE_WORDS_OFFSET]);
10776 const regnode *nextbranch= NULL;
10779 for (word_idx= 0; word_idx < (I32)trie->wordcount; word_idx++) {
10780 SV ** const elem_ptr = av_fetch(trie_words,word_idx,0);
10782 PerlIO_printf(Perl_debug_log, "%*s%s ",
10783 (int)(2*(indent+3)), "",
10784 elem_ptr ? pv_pretty(sv, SvPV_nolen_const(*elem_ptr), SvCUR(*elem_ptr), 60,
10785 PL_colors[0], PL_colors[1],
10786 (SvUTF8(*elem_ptr) ? PERL_PV_ESCAPE_UNI : 0) |
10787 PERL_PV_PRETTY_ELLIPSES |
10788 PERL_PV_PRETTY_LTGT
10793 U16 dist= trie->jump[word_idx+1];
10794 PerlIO_printf(Perl_debug_log, "(%"UVuf")\n",
10795 (UV)((dist ? this_trie + dist : next) - start));
10798 nextbranch= this_trie + trie->jump[0];
10799 DUMPUNTIL(this_trie + dist, nextbranch);
10801 if (nextbranch && PL_regkind[OP(nextbranch)]==BRANCH)
10802 nextbranch= regnext((regnode *)nextbranch);
10804 PerlIO_printf(Perl_debug_log, "\n");
10807 if (last && next > last)
10812 else if ( op == CURLY ) { /* "next" might be very big: optimizer */
10813 DUMPUNTIL(NEXTOPER(node) + EXTRA_STEP_2ARGS,
10814 NEXTOPER(node) + EXTRA_STEP_2ARGS + 1);
10816 else if (PL_regkind[(U8)op] == CURLY && op != CURLYX) {
10818 DUMPUNTIL(NEXTOPER(node) + EXTRA_STEP_2ARGS, next);
10820 else if ( op == PLUS || op == STAR) {
10821 DUMPUNTIL(NEXTOPER(node), NEXTOPER(node) + 1);
10823 else if (PL_regkind[(U8)op] == ANYOF) {
10824 /* arglen 1 + class block */
10825 node += 1 + ((ANYOF_FLAGS(node) & ANYOF_CLASS)
10826 ? ANYOF_CLASS_SKIP : ANYOF_SKIP);
10827 node = NEXTOPER(node);
10829 else if (PL_regkind[(U8)op] == EXACT) {
10830 /* Literal string, where present. */
10831 node += NODE_SZ_STR(node) - 1;
10832 node = NEXTOPER(node);
10835 node = NEXTOPER(node);
10836 node += regarglen[(U8)op];
10838 if (op == CURLYX || op == OPEN)
10842 #ifdef DEBUG_DUMPUNTIL
10843 PerlIO_printf(Perl_debug_log, "--- %d\n", (int)indent);
10848 #endif /* DEBUGGING */
10852 * c-indentation-style: bsd
10853 * c-basic-offset: 4
10854 * indent-tabs-mode: t
10857 * ex: set ts=8 sts=4 sw=4 noet:
projects / perl / modules / re-engine-Hooks.git / blob