5 * One Ring to rule them all, One Ring to find them
7 * [p.v of _The Lord of the Rings_, opening poem]
8 * [p.50 of _The Lord of the Rings_, I/iii: "The Shadow of the Past"]
9 * [p.254 of _The Lord of the Rings_, II/ii: "The Council of Elrond"]
12 /* This file contains functions for executing a regular expression. See
13 * also regcomp.c which funnily enough, contains functions for compiling
14 * a regular expression.
16 * This file is also copied at build time to ext/re/re_exec.c, where
17 * it's built with -DPERL_EXT_RE_BUILD -DPERL_EXT_RE_DEBUG -DPERL_EXT.
18 * This causes the main functions to be compiled under new names and with
19 * debugging support added, which makes "use re 'debug'" work.
22 /* NOTE: this is derived from Henry Spencer's regexp code, and should not
23 * confused with the original package (see point 3 below). Thanks, Henry!
26 /* Additional note: this code is very heavily munged from Henry's version
27 * in places. In some spots I've traded clarity for efficiency, so don't
28 * blame Henry for some of the lack of readability.
31 /* The names of the functions have been changed from regcomp and
32 * regexec to pregcomp and pregexec in order to avoid conflicts
33 * with the POSIX routines of the same names.
36 #ifdef PERL_EXT_RE_BUILD
41 * pregcomp and pregexec -- regsub and regerror are not used in perl
43 * Copyright (c) 1986 by University of Toronto.
44 * Written by Henry Spencer. Not derived from licensed software.
46 * Permission is granted to anyone to use this software for any
47 * purpose on any computer system, and to redistribute it freely,
48 * subject to the following restrictions:
50 * 1. The author is not responsible for the consequences of use of
51 * this software, no matter how awful, even if they arise
54 * 2. The origin of this software must not be misrepresented, either
55 * by explicit claim or by omission.
57 * 3. Altered versions must be plainly marked as such, and must not
58 * be misrepresented as being the original software.
60 **** Alterations to Henry's code are...
62 **** Copyright (C) 1991, 1992, 1993, 1994, 1995, 1996, 1997, 1998, 1999,
63 **** 2000, 2001, 2002, 2003, 2004, 2005, 2006, 2007, 2008
64 **** by Larry Wall and others
66 **** You may distribute under the terms of either the GNU General Public
67 **** 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_REGEXEC_C
77 #ifdef PERL_IN_XSUB_RE
83 #define RF_tainted 1 /* tainted information used? */
84 #define RF_warned 2 /* warned about big count? */
86 #define RF_utf8 8 /* Pattern contains multibyte chars? */
88 #define UTF_PATTERN ((PL_reg_flags & RF_utf8) != 0)
90 #define RS_init 1 /* eval environment created */
91 #define RS_set 2 /* replsv value is set */
97 /* Valid for non-utf8 strings only: avoids the reginclass call if there are no
98 * complications: i.e., if everything matchable is straight forward in the
100 #define REGINCLASS(prog,p,c) (ANYOF_FLAGS(p) ? reginclass(prog,p,c,0,0) \
101 : ANYOF_BITMAP_TEST(p,*(c)))
107 #define CHR_SVLEN(sv) (utf8_target ? sv_len_utf8(sv) : SvCUR(sv))
108 #define CHR_DIST(a,b) (PL_reg_match_utf8 ? utf8_distance(a,b) : a - b)
110 #define HOPc(pos,off) \
111 (char *)(PL_reg_match_utf8 \
112 ? reghop3((U8*)pos, off, (U8*)(off >= 0 ? PL_regeol : PL_bostr)) \
114 #define HOPBACKc(pos, off) \
115 (char*)(PL_reg_match_utf8\
116 ? reghopmaybe3((U8*)pos, -off, (U8*)PL_bostr) \
117 : (pos - off >= PL_bostr) \
121 #define HOP3(pos,off,lim) (PL_reg_match_utf8 ? reghop3((U8*)(pos), off, (U8*)(lim)) : (U8*)(pos + off))
122 #define HOP3c(pos,off,lim) ((char*)HOP3(pos,off,lim))
124 /* these are unrolled below in the CCC_TRY_XXX defined */
125 #define LOAD_UTF8_CHARCLASS(class,str) STMT_START { \
126 if (!CAT2(PL_utf8_,class)) { bool ok; ENTER; save_re_context(); ok=CAT2(is_utf8_,class)((const U8*)str); assert(ok); LEAVE; } } STMT_END
128 /* Doesn't do an assert to verify that is correct */
129 #define LOAD_UTF8_CHARCLASS_NO_CHECK(class) STMT_START { \
130 if (!CAT2(PL_utf8_,class)) { bool throw_away; ENTER; save_re_context(); throw_away = CAT2(is_utf8_,class)((const U8*)" "); LEAVE; } } STMT_END
132 #define LOAD_UTF8_CHARCLASS_ALNUM() LOAD_UTF8_CHARCLASS(alnum,"a")
133 #define LOAD_UTF8_CHARCLASS_DIGIT() LOAD_UTF8_CHARCLASS(digit,"0")
134 #define LOAD_UTF8_CHARCLASS_SPACE() LOAD_UTF8_CHARCLASS(space," ")
136 #define LOAD_UTF8_CHARCLASS_GCB() /* Grapheme cluster boundaries */ \
137 LOAD_UTF8_CHARCLASS(X_begin, " "); \
138 LOAD_UTF8_CHARCLASS(X_non_hangul, "A"); \
139 /* These are utf8 constants, and not utf-ebcdic constants, so the \
140 * assert should likely and hopefully fail on an EBCDIC machine */ \
141 LOAD_UTF8_CHARCLASS(X_extend, "\xcc\x80"); /* U+0300 */ \
143 /* No asserts are done for these, in case called on an early \
144 * Unicode version in which they map to nothing */ \
145 LOAD_UTF8_CHARCLASS_NO_CHECK(X_prepend);/* U+0E40 "\xe0\xb9\x80" */ \
146 LOAD_UTF8_CHARCLASS_NO_CHECK(X_L); /* U+1100 "\xe1\x84\x80" */ \
147 LOAD_UTF8_CHARCLASS_NO_CHECK(X_LV); /* U+AC00 "\xea\xb0\x80" */ \
148 LOAD_UTF8_CHARCLASS_NO_CHECK(X_LVT); /* U+AC01 "\xea\xb0\x81" */ \
149 LOAD_UTF8_CHARCLASS_NO_CHECK(X_LV_LVT_V);/* U+AC01 "\xea\xb0\x81" */\
150 LOAD_UTF8_CHARCLASS_NO_CHECK(X_T); /* U+11A8 "\xe1\x86\xa8" */ \
151 LOAD_UTF8_CHARCLASS_NO_CHECK(X_V) /* U+1160 "\xe1\x85\xa0" */
154 We dont use PERL_LEGACY_UNICODE_CHARCLASS_MAPPINGS as the direct test
155 so that it is possible to override the option here without having to
156 rebuild the entire core. as we are required to do if we change regcomp.h
157 which is where PERL_LEGACY_UNICODE_CHARCLASS_MAPPINGS is defined.
159 #if PERL_LEGACY_UNICODE_CHARCLASS_MAPPINGS
160 #define BROKEN_UNICODE_CHARCLASS_MAPPINGS
163 #ifdef BROKEN_UNICODE_CHARCLASS_MAPPINGS
164 #define LOAD_UTF8_CHARCLASS_PERL_WORD() LOAD_UTF8_CHARCLASS_ALNUM()
165 #define LOAD_UTF8_CHARCLASS_PERL_SPACE() LOAD_UTF8_CHARCLASS_SPACE()
166 #define LOAD_UTF8_CHARCLASS_POSIX_DIGIT() LOAD_UTF8_CHARCLASS_DIGIT()
167 #define RE_utf8_perl_word PL_utf8_alnum
168 #define RE_utf8_perl_space PL_utf8_space
169 #define RE_utf8_posix_digit PL_utf8_digit
170 #define perl_word alnum
171 #define perl_space space
172 #define posix_digit digit
174 #define LOAD_UTF8_CHARCLASS_PERL_WORD() LOAD_UTF8_CHARCLASS(perl_word,"a")
175 #define LOAD_UTF8_CHARCLASS_PERL_SPACE() LOAD_UTF8_CHARCLASS(perl_space," ")
176 #define LOAD_UTF8_CHARCLASS_POSIX_DIGIT() LOAD_UTF8_CHARCLASS(posix_digit,"0")
177 #define RE_utf8_perl_word PL_utf8_perl_word
178 #define RE_utf8_perl_space PL_utf8_perl_space
179 #define RE_utf8_posix_digit PL_utf8_posix_digit
182 #define PLACEHOLDER /* Something for the preprocessor to grab onto */
184 /* The actual code for CCC_TRY, which uses several variables from the routine
185 * it's callable from. It is designed to be the bulk of a case statement.
186 * FUNC is the macro or function to call on non-utf8 targets that indicate if
187 * nextchr matches the class.
188 * UTF8_TEST is the whole test string to use for utf8 targets
189 * LOAD is what to use to test, and if not present to load in the swash for the
191 * POS_OR_NEG is either empty or ! to complement the results of FUNC or
193 * The logic is: Fail if we're at the end-of-string; otherwise if the target is
194 * utf8 and a variant, load the swash if necessary and test using the utf8
195 * test. Advance to the next character if test is ok, otherwise fail; If not
196 * utf8 or an invariant under utf8, use the non-utf8 test, and fail if it
197 * fails, or advance to the next character */
199 #define _CCC_TRY_CODE(POS_OR_NEG, FUNC, UTF8_TEST, CLASS, STR) \
200 if (locinput >= PL_regeol) { \
203 if (utf8_target && UTF8_IS_CONTINUED(nextchr)) { \
204 LOAD_UTF8_CHARCLASS(CLASS, STR); \
205 if (POS_OR_NEG (UTF8_TEST)) { \
208 locinput += PL_utf8skip[nextchr]; \
209 nextchr = UCHARAT(locinput); \
212 if (POS_OR_NEG (FUNC(nextchr))) { \
215 nextchr = UCHARAT(++locinput); \
218 /* Handle the non-locale cases for a character class and its complement. It
219 * calls _CCC_TRY_CODE with a ! to complement the test for the character class.
220 * This is because that code fails when the test succeeds, so we want to have
221 * the test fail so that the code succeeds. The swash is stored in a
222 * predictable PL_ place */
223 #define _CCC_TRY_NONLOCALE(NAME, NNAME, FUNC, \
226 _CCC_TRY_CODE( !, FUNC, \
227 cBOOL(swash_fetch(CAT2(PL_utf8_,CLASS), \
228 (U8*)locinput, TRUE)), \
231 _CCC_TRY_CODE( PLACEHOLDER , FUNC, \
232 cBOOL(swash_fetch(CAT2(PL_utf8_,CLASS), \
233 (U8*)locinput, TRUE)), \
236 /* Generate the case statements for both locale and non-locale character
237 * classes in regmatch for classes that don't have special unicode semantics.
238 * Locales don't use an immediate swash, but an intermediary special locale
239 * function that is called on the pointer to the current place in the input
240 * string. That function will resolve to needing the same swash. One might
241 * think that because we don't know what the locale will match, we shouldn't
242 * check with the swash loading function that it loaded properly; ie, that we
243 * should use LOAD_UTF8_CHARCLASS_NO_CHECK for those, but what is passed to the
244 * regular LOAD_UTF8_CHARCLASS is in non-locale terms, and so locale is
246 #define CCC_TRY(NAME, NNAME, FUNC, \
247 NAMEL, NNAMEL, LCFUNC, LCFUNC_utf8, \
248 NAMEA, NNAMEA, FUNCA, \
251 PL_reg_flags |= RF_tainted; \
252 _CCC_TRY_CODE( !, LCFUNC, LCFUNC_utf8((U8*)locinput), CLASS, STR) \
254 PL_reg_flags |= RF_tainted; \
255 _CCC_TRY_CODE( PLACEHOLDER, LCFUNC, LCFUNC_utf8((U8*)locinput), \
258 if (locinput >= PL_regeol || ! FUNCA(nextchr)) { \
261 /* Matched a utf8-invariant, so don't have to worry about utf8 */ \
262 nextchr = UCHARAT(++locinput); \
265 if (locinput >= PL_regeol || FUNCA(nextchr)) { \
269 locinput += PL_utf8skip[nextchr]; \
270 nextchr = UCHARAT(locinput); \
273 nextchr = UCHARAT(++locinput); \
276 /* Generate the non-locale cases */ \
277 _CCC_TRY_NONLOCALE(NAME, NNAME, FUNC, CLASS, STR)
279 /* This is like CCC_TRY, but has an extra set of parameters for generating case
280 * statements to handle separate Unicode semantics nodes */
281 #define CCC_TRY_U(NAME, NNAME, FUNC, \
282 NAMEL, NNAMEL, LCFUNC, LCFUNC_utf8, \
283 NAMEU, NNAMEU, FUNCU, \
284 NAMEA, NNAMEA, FUNCA, \
286 CCC_TRY(NAME, NNAME, FUNC, \
287 NAMEL, NNAMEL, LCFUNC, LCFUNC_utf8, \
288 NAMEA, NNAMEA, FUNCA, \
290 _CCC_TRY_NONLOCALE(NAMEU, NNAMEU, FUNCU, CLASS, STR)
292 /* TODO: Combine JUMPABLE and HAS_TEXT to cache OP(rn) */
294 /* for use after a quantifier and before an EXACT-like node -- japhy */
295 /* it would be nice to rework regcomp.sym to generate this stuff. sigh
297 * NOTE that *nothing* that affects backtracking should be in here, specifically
298 * VERBS must NOT be included. JUMPABLE is used to determine if we can ignore a
299 * node that is in between two EXACT like nodes when ascertaining what the required
300 * "follow" character is. This should probably be moved to regex compile time
301 * although it may be done at run time beause of the REF possibility - more
302 * investigation required. -- demerphq
304 #define JUMPABLE(rn) ( \
306 (OP(rn) == CLOSE && (!cur_eval || cur_eval->u.eval.close_paren != ARG(rn))) || \
308 OP(rn) == SUSPEND || OP(rn) == IFMATCH || \
309 OP(rn) == PLUS || OP(rn) == MINMOD || \
311 (PL_regkind[OP(rn)] == CURLY && ARG1(rn) > 0) \
313 #define IS_EXACT(rn) (PL_regkind[OP(rn)] == EXACT)
315 #define HAS_TEXT(rn) ( IS_EXACT(rn) || PL_regkind[OP(rn)] == REF )
318 /* Currently these are only used when PL_regkind[OP(rn)] == EXACT so
319 we don't need this definition. */
320 #define IS_TEXT(rn) ( OP(rn)==EXACT || OP(rn)==REF || OP(rn)==NREF )
321 #define IS_TEXTF(rn) ( (OP(rn)==EXACTFU || OP(rn)==EXACTF) || OP(rn)==REFF || OP(rn)==NREFF )
322 #define IS_TEXTFL(rn) ( OP(rn)==EXACTFL || OP(rn)==REFFL || OP(rn)==NREFFL )
325 /* ... so we use this as its faster. */
326 #define IS_TEXT(rn) ( OP(rn)==EXACT )
327 #define IS_TEXTFU(rn) ( OP(rn)==EXACTFU )
328 #define IS_TEXTF(rn) ( OP(rn)==EXACTF )
329 #define IS_TEXTFL(rn) ( OP(rn)==EXACTFL )
334 Search for mandatory following text node; for lookahead, the text must
335 follow but for lookbehind (rn->flags != 0) we skip to the next step.
337 #define FIND_NEXT_IMPT(rn) STMT_START { \
338 while (JUMPABLE(rn)) { \
339 const OPCODE type = OP(rn); \
340 if (type == SUSPEND || PL_regkind[type] == CURLY) \
341 rn = NEXTOPER(NEXTOPER(rn)); \
342 else if (type == PLUS) \
344 else if (type == IFMATCH) \
345 rn = (rn->flags == 0) ? NEXTOPER(NEXTOPER(rn)) : rn + ARG(rn); \
346 else rn += NEXT_OFF(rn); \
351 static void restore_pos(pTHX_ void *arg);
353 #define REGCP_PAREN_ELEMS 4
354 #define REGCP_OTHER_ELEMS 5
355 #define REGCP_FRAME_ELEMS 1
356 /* REGCP_FRAME_ELEMS are not part of the REGCP_OTHER_ELEMS and
357 * are needed for the regexp context stack bookkeeping. */
360 S_regcppush(pTHX_ I32 parenfloor)
363 const int retval = PL_savestack_ix;
364 const int paren_elems_to_push = (PL_regsize - parenfloor) * REGCP_PAREN_ELEMS;
365 const UV total_elems = paren_elems_to_push + REGCP_OTHER_ELEMS;
366 const UV elems_shifted = total_elems << SAVE_TIGHT_SHIFT;
368 GET_RE_DEBUG_FLAGS_DECL;
370 if (paren_elems_to_push < 0)
371 Perl_croak(aTHX_ "panic: paren_elems_to_push < 0");
373 if ((elems_shifted >> SAVE_TIGHT_SHIFT) != total_elems)
374 Perl_croak(aTHX_ "panic: paren_elems_to_push offset %"UVuf
375 " out of range (%lu-%ld)",
376 total_elems, (unsigned long)PL_regsize, (long)parenfloor);
378 SSGROW(total_elems + REGCP_FRAME_ELEMS);
380 for (p = PL_regsize; p > parenfloor; p--) {
381 /* REGCP_PARENS_ELEMS are pushed per pairs of parentheses. */
382 SSPUSHINT(PL_regoffs[p].end);
383 SSPUSHINT(PL_regoffs[p].start);
384 SSPUSHPTR(PL_reg_start_tmp[p]);
386 DEBUG_BUFFERS_r(PerlIO_printf(Perl_debug_log,
387 " saving \\%"UVuf" %"IVdf"(%"IVdf")..%"IVdf"\n",
388 (UV)p, (IV)PL_regoffs[p].start,
389 (IV)(PL_reg_start_tmp[p] - PL_bostr),
390 (IV)PL_regoffs[p].end
393 /* REGCP_OTHER_ELEMS are pushed in any case, parentheses or no. */
394 SSPUSHPTR(PL_regoffs);
395 SSPUSHINT(PL_regsize);
396 SSPUSHINT(*PL_reglastparen);
397 SSPUSHINT(*PL_reglastcloseparen);
398 SSPUSHPTR(PL_reginput);
399 SSPUSHUV(SAVEt_REGCONTEXT | elems_shifted); /* Magic cookie. */
404 /* These are needed since we do not localize EVAL nodes: */
405 #define REGCP_SET(cp) \
407 PerlIO_printf(Perl_debug_log, \
408 " Setting an EVAL scope, savestack=%"IVdf"\n", \
409 (IV)PL_savestack_ix)); \
412 #define REGCP_UNWIND(cp) \
414 if (cp != PL_savestack_ix) \
415 PerlIO_printf(Perl_debug_log, \
416 " Clearing an EVAL scope, savestack=%"IVdf"..%"IVdf"\n", \
417 (IV)(cp), (IV)PL_savestack_ix)); \
421 S_regcppop(pTHX_ const regexp *rex)
426 GET_RE_DEBUG_FLAGS_DECL;
428 PERL_ARGS_ASSERT_REGCPPOP;
430 /* Pop REGCP_OTHER_ELEMS before the parentheses loop starts. */
432 assert((i & SAVE_MASK) == SAVEt_REGCONTEXT); /* Check that the magic cookie is there. */
433 i >>= SAVE_TIGHT_SHIFT; /* Parentheses elements to pop. */
434 input = (char *) SSPOPPTR;
435 *PL_reglastcloseparen = SSPOPINT;
436 *PL_reglastparen = SSPOPINT;
437 PL_regsize = SSPOPINT;
438 PL_regoffs=(regexp_paren_pair *) SSPOPPTR;
440 i -= REGCP_OTHER_ELEMS;
441 /* Now restore the parentheses context. */
442 for ( ; i > 0; i -= REGCP_PAREN_ELEMS) {
444 U32 paren = (U32)SSPOPINT;
445 PL_reg_start_tmp[paren] = (char *) SSPOPPTR;
446 PL_regoffs[paren].start = SSPOPINT;
448 if (paren <= *PL_reglastparen)
449 PL_regoffs[paren].end = tmps;
451 PerlIO_printf(Perl_debug_log,
452 " restoring \\%"UVuf" to %"IVdf"(%"IVdf")..%"IVdf"%s\n",
453 (UV)paren, (IV)PL_regoffs[paren].start,
454 (IV)(PL_reg_start_tmp[paren] - PL_bostr),
455 (IV)PL_regoffs[paren].end,
456 (paren > *PL_reglastparen ? "(no)" : ""));
460 if (*PL_reglastparen + 1 <= rex->nparens) {
461 PerlIO_printf(Perl_debug_log,
462 " restoring \\%"IVdf"..\\%"IVdf" to undef\n",
463 (IV)(*PL_reglastparen + 1), (IV)rex->nparens);
467 /* It would seem that the similar code in regtry()
468 * already takes care of this, and in fact it is in
469 * a better location to since this code can #if 0-ed out
470 * but the code in regtry() is needed or otherwise tests
471 * requiring null fields (pat.t#187 and split.t#{13,14}
472 * (as of patchlevel 7877) will fail. Then again,
473 * this code seems to be necessary or otherwise
474 * this erroneously leaves $1 defined: "1" =~ /^(?:(\d)x)?\d$/
475 * --jhi updated by dapm */
476 for (i = *PL_reglastparen + 1; i <= rex->nparens; i++) {
478 PL_regoffs[i].start = -1;
479 PL_regoffs[i].end = -1;
485 #define regcpblow(cp) LEAVE_SCOPE(cp) /* Ignores regcppush()ed data. */
488 * pregexec and friends
491 #ifndef PERL_IN_XSUB_RE
493 - pregexec - match a regexp against a string
496 Perl_pregexec(pTHX_ REGEXP * const prog, char* stringarg, register char *strend,
497 char *strbeg, I32 minend, SV *screamer, U32 nosave)
498 /* strend: pointer to null at end of string */
499 /* strbeg: real beginning of string */
500 /* minend: end of match must be >=minend after stringarg. */
501 /* nosave: For optimizations. */
503 PERL_ARGS_ASSERT_PREGEXEC;
506 regexec_flags(prog, stringarg, strend, strbeg, minend, screamer, NULL,
507 nosave ? 0 : REXEC_COPY_STR);
512 * Need to implement the following flags for reg_anch:
514 * USE_INTUIT_NOML - Useful to call re_intuit_start() first
516 * INTUIT_AUTORITATIVE_NOML - Can trust a positive answer
517 * INTUIT_AUTORITATIVE_ML
518 * INTUIT_ONCE_NOML - Intuit can match in one location only.
521 * Another flag for this function: SECOND_TIME (so that float substrs
522 * with giant delta may be not rechecked).
525 /* Assumptions: if ANCH_GPOS, then strpos is anchored. XXXX Check GPOS logic */
527 /* If SCREAM, then SvPVX_const(sv) should be compatible with strpos and strend.
528 Otherwise, only SvCUR(sv) is used to get strbeg. */
530 /* XXXX We assume that strpos is strbeg unless sv. */
532 /* XXXX Some places assume that there is a fixed substring.
533 An update may be needed if optimizer marks as "INTUITable"
534 RExen without fixed substrings. Similarly, it is assumed that
535 lengths of all the strings are no more than minlen, thus they
536 cannot come from lookahead.
537 (Or minlen should take into account lookahead.)
538 NOTE: Some of this comment is not correct. minlen does now take account
539 of lookahead/behind. Further research is required. -- demerphq
543 /* A failure to find a constant substring means that there is no need to make
544 an expensive call to REx engine, thus we celebrate a failure. Similarly,
545 finding a substring too deep into the string means that less calls to
546 regtry() should be needed.
548 REx compiler's optimizer found 4 possible hints:
549 a) Anchored substring;
551 c) Whether we are anchored (beginning-of-line or \G);
552 d) First node (of those at offset 0) which may distinguish positions;
553 We use a)b)d) and multiline-part of c), and try to find a position in the
554 string which does not contradict any of them.
557 /* Most of decisions we do here should have been done at compile time.
558 The nodes of the REx which we used for the search should have been
559 deleted from the finite automaton. */
562 Perl_re_intuit_start(pTHX_ REGEXP * const rx, SV *sv, char *strpos,
563 char *strend, const U32 flags, re_scream_pos_data *data)
566 struct regexp *const prog = (struct regexp *)SvANY(rx);
567 register I32 start_shift = 0;
568 /* Should be nonnegative! */
569 register I32 end_shift = 0;
574 const bool utf8_target = (sv && SvUTF8(sv)) ? 1 : 0; /* if no sv we have to assume bytes */
576 register char *other_last = NULL; /* other substr checked before this */
577 char *check_at = NULL; /* check substr found at this pos */
578 const I32 multiline = prog->extflags & RXf_PMf_MULTILINE;
579 RXi_GET_DECL(prog,progi);
581 const char * const i_strpos = strpos;
583 GET_RE_DEBUG_FLAGS_DECL;
585 PERL_ARGS_ASSERT_RE_INTUIT_START;
587 RX_MATCH_UTF8_set(rx,utf8_target);
590 PL_reg_flags |= RF_utf8;
593 debug_start_match(rx, utf8_target, strpos, strend,
594 sv ? "Guessing start of match in sv for"
595 : "Guessing start of match in string for");
598 /* CHR_DIST() would be more correct here but it makes things slow. */
599 if (prog->minlen > strend - strpos) {
600 DEBUG_EXECUTE_r(PerlIO_printf(Perl_debug_log,
601 "String too short... [re_intuit_start]\n"));
605 strbeg = (sv && SvPOK(sv)) ? strend - SvCUR(sv) : strpos;
608 if (!prog->check_utf8 && prog->check_substr)
609 to_utf8_substr(prog);
610 check = prog->check_utf8;
612 if (!prog->check_substr && prog->check_utf8)
613 to_byte_substr(prog);
614 check = prog->check_substr;
616 if (check == &PL_sv_undef) {
617 DEBUG_EXECUTE_r(PerlIO_printf(Perl_debug_log,
618 "Non-utf8 string cannot match utf8 check string\n"));
621 if (prog->extflags & RXf_ANCH) { /* Match at beg-of-str or after \n */
622 ml_anch = !( (prog->extflags & RXf_ANCH_SINGLE)
623 || ( (prog->extflags & RXf_ANCH_BOL)
624 && !multiline ) ); /* Check after \n? */
627 if ( !(prog->extflags & RXf_ANCH_GPOS) /* Checked by the caller */
628 && !(prog->intflags & PREGf_IMPLICIT) /* not a real BOL */
629 /* SvCUR is not set on references: SvRV and SvPVX_const overlap */
631 && (strpos != strbeg)) {
632 DEBUG_EXECUTE_r(PerlIO_printf(Perl_debug_log, "Not at start...\n"));
635 if (prog->check_offset_min == prog->check_offset_max &&
636 !(prog->extflags & RXf_CANY_SEEN)) {
637 /* Substring at constant offset from beg-of-str... */
640 s = HOP3c(strpos, prog->check_offset_min, strend);
643 slen = SvCUR(check); /* >= 1 */
645 if ( strend - s > slen || strend - s < slen - 1
646 || (strend - s == slen && strend[-1] != '\n')) {
647 DEBUG_EXECUTE_r(PerlIO_printf(Perl_debug_log, "String too long...\n"));
650 /* Now should match s[0..slen-2] */
652 if (slen && (*SvPVX_const(check) != *s
654 && memNE(SvPVX_const(check), s, slen)))) {
656 DEBUG_EXECUTE_r(PerlIO_printf(Perl_debug_log, "String not equal...\n"));
660 else if (*SvPVX_const(check) != *s
661 || ((slen = SvCUR(check)) > 1
662 && memNE(SvPVX_const(check), s, slen)))
665 goto success_at_start;
668 /* Match is anchored, but substr is not anchored wrt beg-of-str. */
670 start_shift = prog->check_offset_min; /* okay to underestimate on CC */
671 end_shift = prog->check_end_shift;
674 const I32 end = prog->check_offset_max + CHR_SVLEN(check)
675 - (SvTAIL(check) != 0);
676 const I32 eshift = CHR_DIST((U8*)strend, (U8*)s) - end;
678 if (end_shift < eshift)
682 else { /* Can match at random position */
685 start_shift = prog->check_offset_min; /* okay to underestimate on CC */
686 end_shift = prog->check_end_shift;
688 /* end shift should be non negative here */
691 #ifdef QDEBUGGING /* 7/99: reports of failure (with the older version) */
693 Perl_croak(aTHX_ "panic: end_shift: %"IVdf" pattern:\n%s\n ",
694 (IV)end_shift, RX_PRECOMP(prog));
698 /* Find a possible match in the region s..strend by looking for
699 the "check" substring in the region corrected by start/end_shift. */
702 I32 srch_start_shift = start_shift;
703 I32 srch_end_shift = end_shift;
704 if (srch_start_shift < 0 && strbeg - s > srch_start_shift) {
705 srch_end_shift -= ((strbeg - s) - srch_start_shift);
706 srch_start_shift = strbeg - s;
708 DEBUG_OPTIMISE_MORE_r({
709 PerlIO_printf(Perl_debug_log, "Check offset min: %"IVdf" Start shift: %"IVdf" End shift %"IVdf" Real End Shift: %"IVdf"\n",
710 (IV)prog->check_offset_min,
711 (IV)srch_start_shift,
713 (IV)prog->check_end_shift);
716 if (flags & REXEC_SCREAM) {
717 I32 p = -1; /* Internal iterator of scream. */
718 I32 * const pp = data ? data->scream_pos : &p;
720 if (PL_screamfirst[BmRARE(check)] >= 0
721 || ( BmRARE(check) == '\n'
722 && (BmPREVIOUS(check) == SvCUR(check) - 1)
724 s = screaminstr(sv, check,
725 srch_start_shift + (s - strbeg), srch_end_shift, pp, 0);
728 /* we may be pointing at the wrong string */
729 if (s && RXp_MATCH_COPIED(prog))
730 s = strbeg + (s - SvPVX_const(sv));
732 *data->scream_olds = s;
737 if (prog->extflags & RXf_CANY_SEEN) {
738 start_point= (U8*)(s + srch_start_shift);
739 end_point= (U8*)(strend - srch_end_shift);
741 start_point= HOP3(s, srch_start_shift, srch_start_shift < 0 ? strbeg : strend);
742 end_point= HOP3(strend, -srch_end_shift, strbeg);
744 DEBUG_OPTIMISE_MORE_r({
745 PerlIO_printf(Perl_debug_log, "fbm_instr len=%d str=<%.*s>\n",
746 (int)(end_point - start_point),
747 (int)(end_point - start_point) > 20 ? 20 : (int)(end_point - start_point),
751 s = fbm_instr( start_point, end_point,
752 check, multiline ? FBMrf_MULTILINE : 0);
755 /* Update the count-of-usability, remove useless subpatterns,
759 RE_PV_QUOTED_DECL(quoted, utf8_target, PERL_DEBUG_PAD_ZERO(0),
760 SvPVX_const(check), RE_SV_DUMPLEN(check), 30);
761 PerlIO_printf(Perl_debug_log, "%s %s substr %s%s%s",
762 (s ? "Found" : "Did not find"),
763 (check == (utf8_target ? prog->anchored_utf8 : prog->anchored_substr)
764 ? "anchored" : "floating"),
767 (s ? " at offset " : "...\n") );
772 /* Finish the diagnostic message */
773 DEBUG_EXECUTE_r(PerlIO_printf(Perl_debug_log, "%ld...\n", (long)(s - i_strpos)) );
775 /* XXX dmq: first branch is for positive lookbehind...
776 Our check string is offset from the beginning of the pattern.
777 So we need to do any stclass tests offset forward from that
786 /* Got a candidate. Check MBOL anchoring, and the *other* substr.
787 Start with the other substr.
788 XXXX no SCREAM optimization yet - and a very coarse implementation
789 XXXX /ttx+/ results in anchored="ttx", floating="x". floating will
790 *always* match. Probably should be marked during compile...
791 Probably it is right to do no SCREAM here...
794 if (utf8_target ? (prog->float_utf8 && prog->anchored_utf8)
795 : (prog->float_substr && prog->anchored_substr))
797 /* Take into account the "other" substring. */
798 /* XXXX May be hopelessly wrong for UTF... */
801 if (check == (utf8_target ? prog->float_utf8 : prog->float_substr)) {
804 char * const last = HOP3c(s, -start_shift, strbeg);
806 char * const saved_s = s;
809 t = s - prog->check_offset_max;
810 if (s - strpos > prog->check_offset_max /* signed-corrected t > strpos */
812 || ((t = (char*)reghopmaybe3((U8*)s, -(prog->check_offset_max), (U8*)strpos))
817 t = HOP3c(t, prog->anchored_offset, strend);
818 if (t < other_last) /* These positions already checked */
820 last2 = last1 = HOP3c(strend, -prog->minlen, strbeg);
823 /* XXXX It is not documented what units *_offsets are in.
824 We assume bytes, but this is clearly wrong.
825 Meaning this code needs to be carefully reviewed for errors.
829 /* On end-of-str: see comment below. */
830 must = utf8_target ? prog->anchored_utf8 : prog->anchored_substr;
831 if (must == &PL_sv_undef) {
833 DEBUG_r(must = prog->anchored_utf8); /* for debug */
838 HOP3(HOP3(last1, prog->anchored_offset, strend)
839 + SvCUR(must), -(SvTAIL(must)!=0), strbeg),
841 multiline ? FBMrf_MULTILINE : 0
844 RE_PV_QUOTED_DECL(quoted, utf8_target, PERL_DEBUG_PAD_ZERO(0),
845 SvPVX_const(must), RE_SV_DUMPLEN(must), 30);
846 PerlIO_printf(Perl_debug_log, "%s anchored substr %s%s",
847 (s ? "Found" : "Contradicts"),
848 quoted, RE_SV_TAIL(must));
853 if (last1 >= last2) {
854 DEBUG_EXECUTE_r(PerlIO_printf(Perl_debug_log,
855 ", giving up...\n"));
858 DEBUG_EXECUTE_r(PerlIO_printf(Perl_debug_log,
859 ", trying floating at offset %ld...\n",
860 (long)(HOP3c(saved_s, 1, strend) - i_strpos)));
861 other_last = HOP3c(last1, prog->anchored_offset+1, strend);
862 s = HOP3c(last, 1, strend);
866 DEBUG_EXECUTE_r(PerlIO_printf(Perl_debug_log, " at offset %ld...\n",
867 (long)(s - i_strpos)));
868 t = HOP3c(s, -prog->anchored_offset, strbeg);
869 other_last = HOP3c(s, 1, strend);
877 else { /* Take into account the floating substring. */
879 char * const saved_s = s;
882 t = HOP3c(s, -start_shift, strbeg);
884 HOP3c(strend, -prog->minlen + prog->float_min_offset, strbeg);
885 if (CHR_DIST((U8*)last, (U8*)t) > prog->float_max_offset)
886 last = HOP3c(t, prog->float_max_offset, strend);
887 s = HOP3c(t, prog->float_min_offset, strend);
890 /* XXXX It is not documented what units *_offsets are in. Assume bytes. */
891 must = utf8_target ? prog->float_utf8 : prog->float_substr;
892 /* fbm_instr() takes into account exact value of end-of-str
893 if the check is SvTAIL(ed). Since false positives are OK,
894 and end-of-str is not later than strend we are OK. */
895 if (must == &PL_sv_undef) {
897 DEBUG_r(must = prog->float_utf8); /* for debug message */
900 s = fbm_instr((unsigned char*)s,
901 (unsigned char*)last + SvCUR(must)
903 must, multiline ? FBMrf_MULTILINE : 0);
905 RE_PV_QUOTED_DECL(quoted, utf8_target, PERL_DEBUG_PAD_ZERO(0),
906 SvPVX_const(must), RE_SV_DUMPLEN(must), 30);
907 PerlIO_printf(Perl_debug_log, "%s floating substr %s%s",
908 (s ? "Found" : "Contradicts"),
909 quoted, RE_SV_TAIL(must));
913 DEBUG_EXECUTE_r(PerlIO_printf(Perl_debug_log,
914 ", giving up...\n"));
917 DEBUG_EXECUTE_r(PerlIO_printf(Perl_debug_log,
918 ", trying anchored starting at offset %ld...\n",
919 (long)(saved_s + 1 - i_strpos)));
921 s = HOP3c(t, 1, strend);
925 DEBUG_EXECUTE_r(PerlIO_printf(Perl_debug_log, " at offset %ld...\n",
926 (long)(s - i_strpos)));
927 other_last = s; /* Fix this later. --Hugo */
937 t= (char*)HOP3( s, -prog->check_offset_max, (prog->check_offset_max<0) ? strend : strpos);
939 DEBUG_OPTIMISE_MORE_r(
940 PerlIO_printf(Perl_debug_log,
941 "Check offset min:%"IVdf" max:%"IVdf" S:%"IVdf" t:%"IVdf" D:%"IVdf" end:%"IVdf"\n",
942 (IV)prog->check_offset_min,
943 (IV)prog->check_offset_max,
951 if (s - strpos > prog->check_offset_max /* signed-corrected t > strpos */
953 || ((t = (char*)reghopmaybe3((U8*)s, -prog->check_offset_max, (U8*) ((prog->check_offset_max<0) ? strend : strpos)))
956 /* Fixed substring is found far enough so that the match
957 cannot start at strpos. */
959 if (ml_anch && t[-1] != '\n') {
960 /* Eventually fbm_*() should handle this, but often
961 anchored_offset is not 0, so this check will not be wasted. */
962 /* XXXX In the code below we prefer to look for "^" even in
963 presence of anchored substrings. And we search even
964 beyond the found float position. These pessimizations
965 are historical artefacts only. */
967 while (t < strend - prog->minlen) {
969 if (t < check_at - prog->check_offset_min) {
970 if (utf8_target ? prog->anchored_utf8 : prog->anchored_substr) {
971 /* Since we moved from the found position,
972 we definitely contradict the found anchored
973 substr. Due to the above check we do not
974 contradict "check" substr.
975 Thus we can arrive here only if check substr
976 is float. Redo checking for "other"=="fixed".
979 DEBUG_EXECUTE_r(PerlIO_printf(Perl_debug_log, "Found /%s^%s/m at offset %ld, rescanning for anchored from offset %ld...\n",
980 PL_colors[0], PL_colors[1], (long)(strpos - i_strpos), (long)(strpos - i_strpos + prog->anchored_offset)));
981 goto do_other_anchored;
983 /* We don't contradict the found floating substring. */
984 /* XXXX Why not check for STCLASS? */
986 DEBUG_EXECUTE_r(PerlIO_printf(Perl_debug_log, "Found /%s^%s/m at offset %ld...\n",
987 PL_colors[0], PL_colors[1], (long)(s - i_strpos)));
990 /* Position contradicts check-string */
991 /* XXXX probably better to look for check-string
992 than for "\n", so one should lower the limit for t? */
993 DEBUG_EXECUTE_r(PerlIO_printf(Perl_debug_log, "Found /%s^%s/m, restarting lookup for check-string at offset %ld...\n",
994 PL_colors[0], PL_colors[1], (long)(t + 1 - i_strpos)));
995 other_last = strpos = s = t + 1;
1000 DEBUG_EXECUTE_r(PerlIO_printf(Perl_debug_log, "Did not find /%s^%s/m...\n",
1001 PL_colors[0], PL_colors[1]));
1005 DEBUG_EXECUTE_r(PerlIO_printf(Perl_debug_log, "Starting position does not contradict /%s^%s/m...\n",
1006 PL_colors[0], PL_colors[1]));
1010 ++BmUSEFUL(utf8_target ? prog->check_utf8 : prog->check_substr); /* hooray/5 */
1013 /* The found string does not prohibit matching at strpos,
1014 - no optimization of calling REx engine can be performed,
1015 unless it was an MBOL and we are not after MBOL,
1016 or a future STCLASS check will fail this. */
1018 /* Even in this situation we may use MBOL flag if strpos is offset
1019 wrt the start of the string. */
1020 if (ml_anch && sv && !SvROK(sv) /* See prev comment on SvROK */
1021 && (strpos != strbeg) && strpos[-1] != '\n'
1022 /* May be due to an implicit anchor of m{.*foo} */
1023 && !(prog->intflags & PREGf_IMPLICIT))
1028 DEBUG_EXECUTE_r( if (ml_anch)
1029 PerlIO_printf(Perl_debug_log, "Position at offset %ld does not contradict /%s^%s/m...\n",
1030 (long)(strpos - i_strpos), PL_colors[0], PL_colors[1]);
1033 if (!(prog->intflags & PREGf_NAUGHTY) /* XXXX If strpos moved? */
1035 prog->check_utf8 /* Could be deleted already */
1036 && --BmUSEFUL(prog->check_utf8) < 0
1037 && (prog->check_utf8 == prog->float_utf8)
1039 prog->check_substr /* Could be deleted already */
1040 && --BmUSEFUL(prog->check_substr) < 0
1041 && (prog->check_substr == prog->float_substr)
1044 /* If flags & SOMETHING - do not do it many times on the same match */
1045 DEBUG_EXECUTE_r(PerlIO_printf(Perl_debug_log, "... Disabling check substring...\n"));
1046 /* XXX Does the destruction order has to change with utf8_target? */
1047 SvREFCNT_dec(utf8_target ? prog->check_utf8 : prog->check_substr);
1048 SvREFCNT_dec(utf8_target ? prog->check_substr : prog->check_utf8);
1049 prog->check_substr = prog->check_utf8 = NULL; /* disable */
1050 prog->float_substr = prog->float_utf8 = NULL; /* clear */
1051 check = NULL; /* abort */
1053 /* XXXX If the check string was an implicit check MBOL, then we need to unset the relevant flag
1054 see http://bugs.activestate.com/show_bug.cgi?id=87173 */
1055 if (prog->intflags & PREGf_IMPLICIT)
1056 prog->extflags &= ~RXf_ANCH_MBOL;
1057 /* XXXX This is a remnant of the old implementation. It
1058 looks wasteful, since now INTUIT can use many
1059 other heuristics. */
1060 prog->extflags &= ~RXf_USE_INTUIT;
1061 /* XXXX What other flags might need to be cleared in this branch? */
1067 /* Last resort... */
1068 /* XXXX BmUSEFUL already changed, maybe multiple change is meaningful... */
1069 /* trie stclasses are too expensive to use here, we are better off to
1070 leave it to regmatch itself */
1071 if (progi->regstclass && PL_regkind[OP(progi->regstclass)]!=TRIE) {
1072 /* minlen == 0 is possible if regstclass is \b or \B,
1073 and the fixed substr is ''$.
1074 Since minlen is already taken into account, s+1 is before strend;
1075 accidentally, minlen >= 1 guaranties no false positives at s + 1
1076 even for \b or \B. But (minlen? 1 : 0) below assumes that
1077 regstclass does not come from lookahead... */
1078 /* If regstclass takes bytelength more than 1: If charlength==1, OK.
1079 This leaves EXACTF, EXACTFU only, which are dealt with in find_byclass(). */
1080 const U8* const str = (U8*)STRING(progi->regstclass);
1081 const int cl_l = (PL_regkind[OP(progi->regstclass)] == EXACT
1082 ? CHR_DIST(str+STR_LEN(progi->regstclass), str)
1085 if (prog->anchored_substr || prog->anchored_utf8 || ml_anch)
1086 endpos= HOP3c(s, (prog->minlen ? cl_l : 0), strend);
1087 else if (prog->float_substr || prog->float_utf8)
1088 endpos= HOP3c(HOP3c(check_at, -start_shift, strbeg), cl_l, strend);
1092 DEBUG_EXECUTE_r(PerlIO_printf(Perl_debug_log, "start_shift: %"IVdf" check_at: %"IVdf" s: %"IVdf" endpos: %"IVdf"\n",
1093 (IV)start_shift, (IV)(check_at - strbeg), (IV)(s - strbeg), (IV)(endpos - strbeg)));
1096 s = find_byclass(prog, progi->regstclass, s, endpos, NULL);
1099 const char *what = NULL;
1101 if (endpos == strend) {
1102 DEBUG_EXECUTE_r( PerlIO_printf(Perl_debug_log,
1103 "Could not match STCLASS...\n") );
1106 DEBUG_EXECUTE_r( PerlIO_printf(Perl_debug_log,
1107 "This position contradicts STCLASS...\n") );
1108 if ((prog->extflags & RXf_ANCH) && !ml_anch)
1110 /* Contradict one of substrings */
1111 if (prog->anchored_substr || prog->anchored_utf8) {
1112 if ((utf8_target ? prog->anchored_utf8 : prog->anchored_substr) == check) {
1113 DEBUG_EXECUTE_r( what = "anchored" );
1115 s = HOP3c(t, 1, strend);
1116 if (s + start_shift + end_shift > strend) {
1117 /* XXXX Should be taken into account earlier? */
1118 DEBUG_EXECUTE_r( PerlIO_printf(Perl_debug_log,
1119 "Could not match STCLASS...\n") );
1124 DEBUG_EXECUTE_r( PerlIO_printf(Perl_debug_log,
1125 "Looking for %s substr starting at offset %ld...\n",
1126 what, (long)(s + start_shift - i_strpos)) );
1129 /* Have both, check_string is floating */
1130 if (t + start_shift >= check_at) /* Contradicts floating=check */
1131 goto retry_floating_check;
1132 /* Recheck anchored substring, but not floating... */
1136 DEBUG_EXECUTE_r( PerlIO_printf(Perl_debug_log,
1137 "Looking for anchored substr starting at offset %ld...\n",
1138 (long)(other_last - i_strpos)) );
1139 goto do_other_anchored;
1141 /* Another way we could have checked stclass at the
1142 current position only: */
1147 DEBUG_EXECUTE_r( PerlIO_printf(Perl_debug_log,
1148 "Looking for /%s^%s/m starting at offset %ld...\n",
1149 PL_colors[0], PL_colors[1], (long)(t - i_strpos)) );
1152 if (!(utf8_target ? prog->float_utf8 : prog->float_substr)) /* Could have been deleted */
1154 /* Check is floating substring. */
1155 retry_floating_check:
1156 t = check_at - start_shift;
1157 DEBUG_EXECUTE_r( what = "floating" );
1158 goto hop_and_restart;
1161 DEBUG_EXECUTE_r(PerlIO_printf(Perl_debug_log,
1162 "By STCLASS: moving %ld --> %ld\n",
1163 (long)(t - i_strpos), (long)(s - i_strpos))
1167 DEBUG_EXECUTE_r(PerlIO_printf(Perl_debug_log,
1168 "Does not contradict STCLASS...\n");
1173 DEBUG_EXECUTE_r(PerlIO_printf(Perl_debug_log, "%s%s:%s match at offset %ld\n",
1174 PL_colors[4], (check ? "Guessed" : "Giving up"),
1175 PL_colors[5], (long)(s - i_strpos)) );
1178 fail_finish: /* Substring not found */
1179 if (prog->check_substr || prog->check_utf8) /* could be removed already */
1180 BmUSEFUL(utf8_target ? prog->check_utf8 : prog->check_substr) += 5; /* hooray */
1182 DEBUG_EXECUTE_r(PerlIO_printf(Perl_debug_log, "%sMatch rejected by optimizer%s\n",
1183 PL_colors[4], PL_colors[5]));
1187 #define DECL_TRIE_TYPE(scan) \
1188 const enum { trie_plain, trie_utf8, trie_utf8_fold, trie_latin_utf8_fold } \
1189 trie_type = (scan->flags != EXACT) \
1190 ? (utf8_target ? trie_utf8_fold : (UTF_PATTERN ? trie_latin_utf8_fold : trie_plain)) \
1191 : (utf8_target ? trie_utf8 : trie_plain)
1193 #define REXEC_TRIE_READ_CHAR(trie_type, trie, widecharmap, uc, uscan, len, \
1194 uvc, charid, foldlen, foldbuf, uniflags) STMT_START { \
1195 switch (trie_type) { \
1196 case trie_utf8_fold: \
1197 if ( foldlen>0 ) { \
1198 uvc = utf8n_to_uvuni( uscan, UTF8_MAXLEN, &len, uniflags ); \
1203 uvc = utf8n_to_uvuni( (U8*)uc, UTF8_MAXLEN, &len, uniflags ); \
1204 uvc = to_uni_fold( uvc, foldbuf, &foldlen ); \
1205 foldlen -= UNISKIP( uvc ); \
1206 uscan = foldbuf + UNISKIP( uvc ); \
1209 case trie_latin_utf8_fold: \
1210 if ( foldlen>0 ) { \
1211 uvc = utf8n_to_uvuni( uscan, UTF8_MAXLEN, &len, uniflags ); \
1217 uvc = to_uni_fold( *(U8*)uc, foldbuf, &foldlen ); \
1218 foldlen -= UNISKIP( uvc ); \
1219 uscan = foldbuf + UNISKIP( uvc ); \
1223 uvc = utf8n_to_uvuni( (U8*)uc, UTF8_MAXLEN, &len, uniflags ); \
1230 charid = trie->charmap[ uvc ]; \
1234 if (widecharmap) { \
1235 SV** const svpp = hv_fetch(widecharmap, \
1236 (char*)&uvc, sizeof(UV), 0); \
1238 charid = (U16)SvIV(*svpp); \
1243 #define REXEC_FBC_EXACTISH_CHECK(CoNd) \
1245 char *my_strend= (char *)strend; \
1248 foldEQ_utf8(s, &my_strend, 0, utf8_target, \
1249 m, NULL, ln, cBOOL(UTF_PATTERN))) \
1250 && (!reginfo || regtry(reginfo, &s)) ) \
1253 U8 foldbuf[UTF8_MAXBYTES_CASE+1]; \
1254 uvchr_to_utf8(tmpbuf, c); \
1255 f = to_utf8_fold(tmpbuf, foldbuf, &foldlen); \
1257 && (f == c1 || f == c2) \
1259 foldEQ_utf8(s, &my_strend, 0, utf8_target,\
1260 m, NULL, ln, cBOOL(UTF_PATTERN)))\
1261 && (!reginfo || regtry(reginfo, &s)) ) \
1267 #define REXEC_FBC_EXACTISH_SCAN(CoNd) \
1271 case EXACTFU: folder = foldEQ_latin1; break; \
1272 case EXACTFL: folder = foldEQ_locale; break; \
1273 case EXACTF: folder = foldEQ; break; \
1275 Perl_croak(aTHX_ "panic: Unexpected op %u", OP(c)); \
1279 && (ln == 1 || folder(s, m, ln)) \
1280 && (!reginfo || regtry(reginfo, &s)) ) \
1286 #define REXEC_FBC_UTF8_SCAN(CoDe) \
1288 while (s + (uskip = UTF8SKIP(s)) <= strend) { \
1294 #define REXEC_FBC_SCAN(CoDe) \
1296 while (s < strend) { \
1302 #define REXEC_FBC_UTF8_CLASS_SCAN(CoNd) \
1303 REXEC_FBC_UTF8_SCAN( \
1305 if (tmp && (!reginfo || regtry(reginfo, &s))) \
1314 #define REXEC_FBC_CLASS_SCAN(CoNd) \
1317 if (tmp && (!reginfo || regtry(reginfo, &s))) \
1326 #define REXEC_FBC_TRYIT \
1327 if ((!reginfo || regtry(reginfo, &s))) \
1330 #define REXEC_FBC_CSCAN(CoNdUtF8,CoNd) \
1331 if (utf8_target) { \
1332 REXEC_FBC_UTF8_CLASS_SCAN(CoNdUtF8); \
1335 REXEC_FBC_CLASS_SCAN(CoNd); \
1338 #define REXEC_FBC_CSCAN_PRELOAD(UtFpReLoAd,CoNdUtF8,CoNd) \
1339 if (utf8_target) { \
1341 REXEC_FBC_UTF8_CLASS_SCAN(CoNdUtF8); \
1344 REXEC_FBC_CLASS_SCAN(CoNd); \
1347 #define REXEC_FBC_CSCAN_TAINT(CoNdUtF8,CoNd) \
1348 PL_reg_flags |= RF_tainted; \
1349 if (utf8_target) { \
1350 REXEC_FBC_UTF8_CLASS_SCAN(CoNdUtF8); \
1353 REXEC_FBC_CLASS_SCAN(CoNd); \
1356 #define DUMP_EXEC_POS(li,s,doutf8) \
1357 dump_exec_pos(li,s,(PL_regeol),(PL_bostr),(PL_reg_starttry),doutf8)
1360 #define UTF8_NOLOAD(TEST_NON_UTF8, IF_SUCCESS, IF_FAIL) \
1361 tmp = (s != PL_bostr) ? UCHARAT(s - 1) : '\n'; \
1362 tmp = TEST_NON_UTF8(tmp); \
1363 REXEC_FBC_UTF8_SCAN( \
1364 if (tmp == ! TEST_NON_UTF8((U8) *s)) { \
1373 #define UTF8_LOAD(TeSt1_UtF8, TeSt2_UtF8, IF_SUCCESS, IF_FAIL) \
1374 if (s == PL_bostr) { \
1378 U8 * const r = reghop3((U8*)s, -1, (U8*)PL_bostr); \
1379 tmp = utf8n_to_uvchr(r, UTF8SKIP(r), 0, UTF8_ALLOW_DEFAULT); \
1382 LOAD_UTF8_CHARCLASS_ALNUM(); \
1383 REXEC_FBC_UTF8_SCAN( \
1384 if (tmp == ! (TeSt2_UtF8)) { \
1393 /* The only difference between the BOUND and NBOUND cases is that
1394 * REXEC_FBC_TRYIT is called when matched in BOUND, and when non-matched in
1395 * NBOUND. This is accomplished by passing it in either the if or else clause,
1396 * with the other one being empty */
1397 #define FBC_BOUND(TEST_NON_UTF8, TEST1_UTF8, TEST2_UTF8) \
1398 FBC_BOUND_COMMON(UTF8_LOAD(TEST1_UTF8, TEST2_UTF8, REXEC_FBC_TRYIT, PLACEHOLDER), TEST_NON_UTF8, REXEC_FBC_TRYIT, PLACEHOLDER)
1400 #define FBC_BOUND_NOLOAD(TEST_NON_UTF8, TEST1_UTF8, TEST2_UTF8) \
1401 FBC_BOUND_COMMON(UTF8_NOLOAD(TEST_NON_UTF8, REXEC_FBC_TRYIT, PLACEHOLDER), TEST_NON_UTF8, REXEC_FBC_TRYIT, PLACEHOLDER)
1403 #define FBC_NBOUND(TEST_NON_UTF8, TEST1_UTF8, TEST2_UTF8) \
1404 FBC_BOUND_COMMON(UTF8_LOAD(TEST1_UTF8, TEST2_UTF8, PLACEHOLDER, REXEC_FBC_TRYIT), TEST_NON_UTF8, PLACEHOLDER, REXEC_FBC_TRYIT)
1406 #define FBC_NBOUND_NOLOAD(TEST_NON_UTF8, TEST1_UTF8, TEST2_UTF8) \
1407 FBC_BOUND_COMMON(UTF8_NOLOAD(TEST_NON_UTF8, PLACEHOLDER, REXEC_FBC_TRYIT), TEST_NON_UTF8, PLACEHOLDER, REXEC_FBC_TRYIT)
1410 /* Common to the BOUND and NBOUND cases. Unfortunately the UTF8 tests need to
1411 * be passed in completely with the variable name being tested, which isn't
1412 * such a clean interface, but this is easier to read than it was before. We
1413 * are looking for the boundary (or non-boundary between a word and non-word
1414 * character. The utf8 and non-utf8 cases have the same logic, but the details
1415 * must be different. Find the "wordness" of the character just prior to this
1416 * one, and compare it with the wordness of this one. If they differ, we have
1417 * a boundary. At the beginning of the string, pretend that the previous
1418 * character was a new-line */
1419 #define FBC_BOUND_COMMON(UTF8_CODE, TEST_NON_UTF8, IF_SUCCESS, IF_FAIL) \
1420 if (utf8_target) { \
1423 else { /* Not utf8 */ \
1424 tmp = (s != PL_bostr) ? UCHARAT(s - 1) : '\n'; \
1425 tmp = TEST_NON_UTF8(tmp); \
1427 if (tmp == ! TEST_NON_UTF8((U8) *s)) { \
1436 if ((!prog->minlen && tmp) && (!reginfo || regtry(reginfo, &s))) \
1439 /* We know what class REx starts with. Try to find this position... */
1440 /* if reginfo is NULL, its a dryrun */
1441 /* annoyingly all the vars in this routine have different names from their counterparts
1442 in regmatch. /grrr */
1445 S_find_byclass(pTHX_ regexp * prog, const regnode *c, char *s,
1446 const char *strend, regmatch_info *reginfo)
1449 const I32 doevery = (prog->intflags & PREGf_SKIP) == 0;
1453 register STRLEN uskip;
1457 register I32 tmp = 1; /* Scratch variable? */
1458 register const bool utf8_target = PL_reg_match_utf8;
1459 RXi_GET_DECL(prog,progi);
1461 PERL_ARGS_ASSERT_FIND_BYCLASS;
1463 /* We know what class it must start with. */
1467 if (utf8_target || OP(c) == ANYOFV) {
1468 REXEC_FBC_UTF8_CLASS_SCAN((ANYOF_FLAGS(c) & ANYOF_NONBITMAP) ||
1469 !UTF8_IS_INVARIANT((U8)s[0]) ?
1470 reginclass(prog, c, (U8*)s, 0, utf8_target) :
1471 REGINCLASS(prog, c, (U8*)s));
1474 while (s < strend) {
1477 if (REGINCLASS(prog, c, (U8*)s) ||
1478 (ANYOF_FOLD_SHARP_S(c, s, strend) &&
1479 /* The assignment of 2 is intentional:
1480 * for the folded sharp s, the skip is 2. */
1481 (skip = SHARP_S_SKIP))) {
1482 if (tmp && (!reginfo || regtry(reginfo, &s)))
1495 if (tmp && (!reginfo || regtry(reginfo, &s)))
1504 ln = STR_LEN(c); /* length to match in octets/bytes */
1505 lnc = (I32) ln; /* length to match in characters */
1507 STRLEN ulen1, ulen2;
1509 U8 tmpbuf1[UTF8_MAXBYTES_CASE+1];
1510 U8 tmpbuf2[UTF8_MAXBYTES_CASE+1];
1511 /* used by commented-out code below */
1512 /*const U32 uniflags = UTF8_ALLOW_DEFAULT;*/
1514 /* XXX: Since the node will be case folded at compile
1515 time this logic is a little odd, although im not
1516 sure that its actually wrong. --dmq */
1518 c1 = to_utf8_lower((U8*)m, tmpbuf1, &ulen1);
1519 c2 = to_utf8_upper((U8*)m, tmpbuf2, &ulen2);
1521 /* XXX: This is kinda strange. to_utf8_XYZ returns the
1522 codepoint of the first character in the converted
1523 form, yet originally we did the extra step.
1524 No tests fail by commenting this code out however
1525 so Ive left it out. -- dmq.
1527 c1 = utf8n_to_uvchr(tmpbuf1, UTF8_MAXBYTES_CASE,
1529 c2 = utf8n_to_uvchr(tmpbuf2, UTF8_MAXBYTES_CASE,
1534 while (sm < ((U8 *) m + ln)) {
1541 if (utf8_target || OP(c) == EXACTFU) {
1543 /* Micro sign folds to GREEK SMALL LETTER MU;
1544 LATIN_SMALL_LETTER_SHARP_S folds to 'ss', and this sets
1545 c2 to the first 's' of the pair, and the code below will
1547 c2 = (c1 == MICRO_SIGN)
1548 ? GREEK_SMALL_LETTER_MU
1549 : (c1 == LATIN_SMALL_LETTER_SHARP_S)
1551 : PL_fold_latin1[c1];
1552 } else c2 = PL_fold[c1];
1560 c2 = PL_fold_locale[c1];
1562 e = HOP3c(strend, -((I32)lnc), s);
1564 if (!reginfo && e < s)
1565 e = s; /* Due to minlen logic of intuit() */
1567 /* The idea in the EXACTF* cases is to first find the
1568 * first character of the EXACTF* node and then, if
1569 * necessary, case-insensitively compare the full
1570 * text of the node. The c1 and c2 are the first
1571 * characters (though in Unicode it gets a bit
1572 * more complicated because there are more cases
1573 * than just upper and lower: one needs to use
1574 * the so-called folding case for case-insensitive
1575 * matching (called "loose matching" in Unicode).
1576 * foldEQ_utf8() will do just that. */
1578 if (utf8_target || UTF_PATTERN) {
1580 U8 tmpbuf [UTF8_MAXBYTES+1];
1583 const U32 uniflags = UTF8_ALLOW_DEFAULT;
1585 /* Upper and lower of 1st char are equal -
1586 * probably not a "letter". */
1589 c = utf8n_to_uvchr((U8*)s, UTF8_MAXBYTES, &len,
1594 REXEC_FBC_EXACTISH_CHECK(c == c1);
1600 c = utf8n_to_uvchr((U8*)s, UTF8_MAXBYTES, &len,
1606 /* Handle some of the three Greek sigmas cases.
1607 * Note that not all the possible combinations
1608 * are handled here: some of them are handled
1609 * by the standard folding rules, and some of
1610 * them (the character class or ANYOF cases)
1611 * are handled during compiletime in
1612 * regexec.c:S_regclass(). */
1613 if (c == (UV)UNICODE_GREEK_CAPITAL_LETTER_SIGMA ||
1614 c == (UV)UNICODE_GREEK_SMALL_LETTER_FINAL_SIGMA)
1615 c = (UV)UNICODE_GREEK_SMALL_LETTER_SIGMA;
1617 REXEC_FBC_EXACTISH_CHECK(c == c1 || c == c2);
1622 /* Neither pattern nor string are UTF8 */
1624 REXEC_FBC_EXACTISH_SCAN(*(U8*)s == c1);
1626 REXEC_FBC_EXACTISH_SCAN(*(U8*)s == c1 || *(U8*)s == c2);
1630 PL_reg_flags |= RF_tainted;
1631 FBC_BOUND(isALNUM_LC,
1632 isALNUM_LC_uvchr(UNI_TO_NATIVE(tmp)),
1633 isALNUM_LC_utf8((U8*)s));
1636 PL_reg_flags |= RF_tainted;
1637 FBC_NBOUND(isALNUM_LC,
1638 isALNUM_LC_uvchr(UNI_TO_NATIVE(tmp)),
1639 isALNUM_LC_utf8((U8*)s));
1642 FBC_BOUND(isWORDCHAR,
1644 cBOOL(swash_fetch(PL_utf8_alnum, (U8*)s, utf8_target)));
1647 FBC_BOUND_NOLOAD(isWORDCHAR_A,
1649 isWORDCHAR_A((U8*)s));
1652 FBC_NBOUND(isWORDCHAR,
1654 cBOOL(swash_fetch(PL_utf8_alnum, (U8*)s, utf8_target)));
1657 FBC_NBOUND_NOLOAD(isWORDCHAR_A,
1659 isWORDCHAR_A((U8*)s));
1662 FBC_BOUND(isWORDCHAR_L1,
1664 cBOOL(swash_fetch(PL_utf8_alnum, (U8*)s, utf8_target)));
1667 FBC_NBOUND(isWORDCHAR_L1,
1669 cBOOL(swash_fetch(PL_utf8_alnum, (U8*)s, utf8_target)));
1672 REXEC_FBC_CSCAN_TAINT(
1673 isALNUM_LC_utf8((U8*)s),
1678 REXEC_FBC_CSCAN_PRELOAD(
1679 LOAD_UTF8_CHARCLASS_PERL_WORD(),
1680 swash_fetch(RE_utf8_perl_word,(U8*)s, utf8_target),
1681 isWORDCHAR_L1((U8) *s)
1685 REXEC_FBC_CSCAN_PRELOAD(
1686 LOAD_UTF8_CHARCLASS_PERL_WORD(),
1687 swash_fetch(RE_utf8_perl_word,(U8*)s, utf8_target),
1692 /* Don't need to worry about utf8, as it can match only a single
1693 * byte invariant character */
1694 REXEC_FBC_CLASS_SCAN( isWORDCHAR_A(*s));
1697 REXEC_FBC_CSCAN_PRELOAD(
1698 LOAD_UTF8_CHARCLASS_PERL_WORD(),
1699 swash_fetch(RE_utf8_perl_word,(U8*)s, utf8_target),
1700 ! isWORDCHAR_L1((U8) *s)
1704 REXEC_FBC_CSCAN_PRELOAD(
1705 LOAD_UTF8_CHARCLASS_PERL_WORD(),
1706 !swash_fetch(RE_utf8_perl_word, (U8*)s, utf8_target),
1717 REXEC_FBC_CSCAN_TAINT(
1718 !isALNUM_LC_utf8((U8*)s),
1723 REXEC_FBC_CSCAN_PRELOAD(
1724 LOAD_UTF8_CHARCLASS_PERL_SPACE(),
1725 *s == ' ' || swash_fetch(RE_utf8_perl_space,(U8*)s, utf8_target),
1730 REXEC_FBC_CSCAN_PRELOAD(
1731 LOAD_UTF8_CHARCLASS_PERL_SPACE(),
1732 *s == ' ' || swash_fetch(RE_utf8_perl_space,(U8*)s, utf8_target),
1737 /* Don't need to worry about utf8, as it can match only a single
1738 * byte invariant character */
1739 REXEC_FBC_CLASS_SCAN( isSPACE_A(*s));
1742 REXEC_FBC_CSCAN_TAINT(
1743 isSPACE_LC_utf8((U8*)s),
1748 REXEC_FBC_CSCAN_PRELOAD(
1749 LOAD_UTF8_CHARCLASS_PERL_SPACE(),
1750 !( *s == ' ' || swash_fetch(RE_utf8_perl_space,(U8*)s, utf8_target)),
1751 ! isSPACE_L1((U8) *s)
1755 REXEC_FBC_CSCAN_PRELOAD(
1756 LOAD_UTF8_CHARCLASS_PERL_SPACE(),
1757 !(*s == ' ' || swash_fetch(RE_utf8_perl_space,(U8*)s, utf8_target)),
1768 REXEC_FBC_CSCAN_TAINT(
1769 !isSPACE_LC_utf8((U8*)s),
1774 REXEC_FBC_CSCAN_PRELOAD(
1775 LOAD_UTF8_CHARCLASS_POSIX_DIGIT(),
1776 swash_fetch(RE_utf8_posix_digit,(U8*)s, utf8_target),
1781 /* Don't need to worry about utf8, as it can match only a single
1782 * byte invariant character */
1783 REXEC_FBC_CLASS_SCAN( isDIGIT_A(*s));
1786 REXEC_FBC_CSCAN_TAINT(
1787 isDIGIT_LC_utf8((U8*)s),
1792 REXEC_FBC_CSCAN_PRELOAD(
1793 LOAD_UTF8_CHARCLASS_POSIX_DIGIT(),
1794 !swash_fetch(RE_utf8_posix_digit,(U8*)s, utf8_target),
1805 REXEC_FBC_CSCAN_TAINT(
1806 !isDIGIT_LC_utf8((U8*)s),
1813 is_LNBREAK_latin1(s)
1825 !is_VERTWS_latin1(s)
1831 is_HORIZWS_latin1(s)
1836 !is_HORIZWS_utf8(s),
1837 !is_HORIZWS_latin1(s)
1844 /* what trie are we using right now */
1846 = (reg_ac_data*)progi->data->data[ ARG( c ) ];
1848 = (reg_trie_data*)progi->data->data[ aho->trie ];
1849 HV *widecharmap = MUTABLE_HV(progi->data->data[ aho->trie + 1 ]);
1851 const char *last_start = strend - trie->minlen;
1853 const char *real_start = s;
1855 STRLEN maxlen = trie->maxlen;
1857 U8 **points; /* map of where we were in the input string
1858 when reading a given char. For ASCII this
1859 is unnecessary overhead as the relationship
1860 is always 1:1, but for Unicode, especially
1861 case folded Unicode this is not true. */
1862 U8 foldbuf[ UTF8_MAXBYTES_CASE + 1 ];
1866 GET_RE_DEBUG_FLAGS_DECL;
1868 /* We can't just allocate points here. We need to wrap it in
1869 * an SV so it gets freed properly if there is a croak while
1870 * running the match */
1873 sv_points=newSV(maxlen * sizeof(U8 *));
1874 SvCUR_set(sv_points,
1875 maxlen * sizeof(U8 *));
1876 SvPOK_on(sv_points);
1877 sv_2mortal(sv_points);
1878 points=(U8**)SvPV_nolen(sv_points );
1879 if ( trie_type != trie_utf8_fold
1880 && (trie->bitmap || OP(c)==AHOCORASICKC) )
1883 bitmap=(U8*)trie->bitmap;
1885 bitmap=(U8*)ANYOF_BITMAP(c);
1887 /* this is the Aho-Corasick algorithm modified a touch
1888 to include special handling for long "unknown char"
1889 sequences. The basic idea being that we use AC as long
1890 as we are dealing with a possible matching char, when
1891 we encounter an unknown char (and we have not encountered
1892 an accepting state) we scan forward until we find a legal
1894 AC matching is basically that of trie matching, except
1895 that when we encounter a failing transition, we fall back
1896 to the current states "fail state", and try the current char
1897 again, a process we repeat until we reach the root state,
1898 state 1, or a legal transition. If we fail on the root state
1899 then we can either terminate if we have reached an accepting
1900 state previously, or restart the entire process from the beginning
1904 while (s <= last_start) {
1905 const U32 uniflags = UTF8_ALLOW_DEFAULT;
1913 U8 *uscan = (U8*)NULL;
1914 U8 *leftmost = NULL;
1916 U32 accepted_word= 0;
1920 while ( state && uc <= (U8*)strend ) {
1922 U32 word = aho->states[ state ].wordnum;
1926 DEBUG_TRIE_EXECUTE_r(
1927 if ( uc <= (U8*)last_start && !BITMAP_TEST(bitmap,*uc) ) {
1928 dump_exec_pos( (char *)uc, c, strend, real_start,
1929 (char *)uc, utf8_target );
1930 PerlIO_printf( Perl_debug_log,
1931 " Scanning for legal start char...\n");
1935 while ( uc <= (U8*)last_start && !BITMAP_TEST(bitmap,*uc) ) {
1939 while ( uc <= (U8*)last_start && !BITMAP_TEST(bitmap,*uc) ) {
1945 if (uc >(U8*)last_start) break;
1949 U8 *lpos= points[ (pointpos - trie->wordinfo[word].len) % maxlen ];
1950 if (!leftmost || lpos < leftmost) {
1951 DEBUG_r(accepted_word=word);
1957 points[pointpos++ % maxlen]= uc;
1958 REXEC_TRIE_READ_CHAR(trie_type, trie, widecharmap, uc,
1959 uscan, len, uvc, charid, foldlen,
1961 DEBUG_TRIE_EXECUTE_r({
1962 dump_exec_pos( (char *)uc, c, strend, real_start,
1964 PerlIO_printf(Perl_debug_log,
1965 " Charid:%3u CP:%4"UVxf" ",
1971 word = aho->states[ state ].wordnum;
1973 base = aho->states[ state ].trans.base;
1975 DEBUG_TRIE_EXECUTE_r({
1977 dump_exec_pos( (char *)uc, c, strend, real_start,
1979 PerlIO_printf( Perl_debug_log,
1980 "%sState: %4"UVxf", word=%"UVxf,
1981 failed ? " Fail transition to " : "",
1982 (UV)state, (UV)word);
1988 ( ((offset = base + charid
1989 - 1 - trie->uniquecharcount)) >= 0)
1990 && ((U32)offset < trie->lasttrans)
1991 && trie->trans[offset].check == state
1992 && (tmp=trie->trans[offset].next))
1994 DEBUG_TRIE_EXECUTE_r(
1995 PerlIO_printf( Perl_debug_log," - legal\n"));
2000 DEBUG_TRIE_EXECUTE_r(
2001 PerlIO_printf( Perl_debug_log," - fail\n"));
2003 state = aho->fail[state];
2007 /* we must be accepting here */
2008 DEBUG_TRIE_EXECUTE_r(
2009 PerlIO_printf( Perl_debug_log," - accepting\n"));
2018 if (!state) state = 1;
2021 if ( aho->states[ state ].wordnum ) {
2022 U8 *lpos = points[ (pointpos - trie->wordinfo[aho->states[ state ].wordnum].len) % maxlen ];
2023 if (!leftmost || lpos < leftmost) {
2024 DEBUG_r(accepted_word=aho->states[ state ].wordnum);
2029 s = (char*)leftmost;
2030 DEBUG_TRIE_EXECUTE_r({
2032 Perl_debug_log,"Matches word #%"UVxf" at position %"IVdf". Trying full pattern...\n",
2033 (UV)accepted_word, (IV)(s - real_start)
2036 if (!reginfo || regtry(reginfo, &s)) {
2042 DEBUG_TRIE_EXECUTE_r({
2043 PerlIO_printf( Perl_debug_log,"Pattern failed. Looking for new start point...\n");
2046 DEBUG_TRIE_EXECUTE_r(
2047 PerlIO_printf( Perl_debug_log,"No match.\n"));
2056 Perl_croak(aTHX_ "panic: unknown regstclass %d", (int)OP(c));
2066 - regexec_flags - match a regexp against a string
2069 Perl_regexec_flags(pTHX_ REGEXP * const rx, char *stringarg, register char *strend,
2070 char *strbeg, I32 minend, SV *sv, void *data, U32 flags)
2071 /* strend: pointer to null at end of string */
2072 /* strbeg: real beginning of string */
2073 /* minend: end of match must be >=minend after stringarg. */
2074 /* data: May be used for some additional optimizations.
2075 Currently its only used, with a U32 cast, for transmitting
2076 the ganch offset when doing a /g match. This will change */
2077 /* nosave: For optimizations. */
2080 struct regexp *const prog = (struct regexp *)SvANY(rx);
2081 /*register*/ char *s;
2082 register regnode *c;
2083 /*register*/ char *startpos = stringarg;
2084 I32 minlen; /* must match at least this many chars */
2085 I32 dontbother = 0; /* how many characters not to try at end */
2086 I32 end_shift = 0; /* Same for the end. */ /* CC */
2087 I32 scream_pos = -1; /* Internal iterator of scream. */
2088 char *scream_olds = NULL;
2089 const bool utf8_target = cBOOL(DO_UTF8(sv));
2091 RXi_GET_DECL(prog,progi);
2092 regmatch_info reginfo; /* create some info to pass to regtry etc */
2093 regexp_paren_pair *swap = NULL;
2094 GET_RE_DEBUG_FLAGS_DECL;
2096 PERL_ARGS_ASSERT_REGEXEC_FLAGS;
2097 PERL_UNUSED_ARG(data);
2099 /* Be paranoid... */
2100 if (prog == NULL || startpos == NULL) {
2101 Perl_croak(aTHX_ "NULL regexp parameter");
2105 multiline = prog->extflags & RXf_PMf_MULTILINE;
2106 reginfo.prog = rx; /* Yes, sorry that this is confusing. */
2108 RX_MATCH_UTF8_set(rx, utf8_target);
2110 debug_start_match(rx, utf8_target, startpos, strend,
2114 minlen = prog->minlen;
2116 if (strend - startpos < (minlen+(prog->check_offset_min<0?prog->check_offset_min:0))) {
2117 DEBUG_EXECUTE_r(PerlIO_printf(Perl_debug_log,
2118 "String too short [regexec_flags]...\n"));
2123 /* Check validity of program. */
2124 if (UCHARAT(progi->program) != REG_MAGIC) {
2125 Perl_croak(aTHX_ "corrupted regexp program");
2129 PL_reg_eval_set = 0;
2133 PL_reg_flags |= RF_utf8;
2135 /* Mark beginning of line for ^ and lookbehind. */
2136 reginfo.bol = startpos; /* XXX not used ??? */
2140 /* Mark end of line for $ (and such) */
2143 /* see how far we have to get to not match where we matched before */
2144 reginfo.till = startpos+minend;
2146 /* If there is a "must appear" string, look for it. */
2149 if (prog->extflags & RXf_GPOS_SEEN) { /* Need to set reginfo->ganch */
2151 if (flags & REXEC_IGNOREPOS){ /* Means: check only at start */
2152 reginfo.ganch = startpos + prog->gofs;
2153 DEBUG_GPOS_r(PerlIO_printf(Perl_debug_log,
2154 "GPOS IGNOREPOS: reginfo.ganch = startpos + %"UVxf"\n",(UV)prog->gofs));
2155 } else if (sv && SvTYPE(sv) >= SVt_PVMG
2157 && (mg = mg_find(sv, PERL_MAGIC_regex_global))
2158 && mg->mg_len >= 0) {
2159 reginfo.ganch = strbeg + mg->mg_len; /* Defined pos() */
2160 DEBUG_GPOS_r(PerlIO_printf(Perl_debug_log,
2161 "GPOS MAGIC: reginfo.ganch = strbeg + %"IVdf"\n",(IV)mg->mg_len));
2163 if (prog->extflags & RXf_ANCH_GPOS) {
2164 if (s > reginfo.ganch)
2166 s = reginfo.ganch - prog->gofs;
2167 DEBUG_GPOS_r(PerlIO_printf(Perl_debug_log,
2168 "GPOS ANCH_GPOS: s = ganch - %"UVxf"\n",(UV)prog->gofs));
2174 reginfo.ganch = strbeg + PTR2UV(data);
2175 DEBUG_GPOS_r(PerlIO_printf(Perl_debug_log,
2176 "GPOS DATA: reginfo.ganch= strbeg + %"UVxf"\n",PTR2UV(data)));
2178 } else { /* pos() not defined */
2179 reginfo.ganch = strbeg;
2180 DEBUG_GPOS_r(PerlIO_printf(Perl_debug_log,
2181 "GPOS: reginfo.ganch = strbeg\n"));
2184 if (PL_curpm && (PM_GETRE(PL_curpm) == rx)) {
2185 /* We have to be careful. If the previous successful match
2186 was from this regex we don't want a subsequent partially
2187 successful match to clobber the old results.
2188 So when we detect this possibility we add a swap buffer
2189 to the re, and switch the buffer each match. If we fail
2190 we switch it back, otherwise we leave it swapped.
2193 /* do we need a save destructor here for eval dies? */
2194 Newxz(prog->offs, (prog->nparens + 1), regexp_paren_pair);
2196 if (!(flags & REXEC_CHECKED) && (prog->check_substr != NULL || prog->check_utf8 != NULL)) {
2197 re_scream_pos_data d;
2199 d.scream_olds = &scream_olds;
2200 d.scream_pos = &scream_pos;
2201 s = re_intuit_start(rx, sv, s, strend, flags, &d);
2203 DEBUG_EXECUTE_r(PerlIO_printf(Perl_debug_log, "Not present...\n"));
2204 goto phooey; /* not present */
2210 /* Simplest case: anchored match need be tried only once. */
2211 /* [unless only anchor is BOL and multiline is set] */
2212 if (prog->extflags & (RXf_ANCH & ~RXf_ANCH_GPOS)) {
2213 if (s == startpos && regtry(®info, &startpos))
2215 else if (multiline || (prog->intflags & PREGf_IMPLICIT)
2216 || (prog->extflags & RXf_ANCH_MBOL)) /* XXXX SBOL? */
2221 dontbother = minlen - 1;
2222 end = HOP3c(strend, -dontbother, strbeg) - 1;
2223 /* for multiline we only have to try after newlines */
2224 if (prog->check_substr || prog->check_utf8) {
2225 /* because of the goto we can not easily reuse the macros for bifurcating the
2226 unicode/non-unicode match modes here like we do elsewhere - demerphq */
2229 goto after_try_utf8;
2231 if (regtry(®info, &s)) {
2238 if (prog->extflags & RXf_USE_INTUIT) {
2239 s = re_intuit_start(rx, sv, s + UTF8SKIP(s), strend, flags, NULL);
2248 } /* end search for check string in unicode */
2250 if (s == startpos) {
2251 goto after_try_latin;
2254 if (regtry(®info, &s)) {
2261 if (prog->extflags & RXf_USE_INTUIT) {
2262 s = re_intuit_start(rx, sv, s + 1, strend, flags, NULL);
2271 } /* end search for check string in latin*/
2272 } /* end search for check string */
2273 else { /* search for newline */
2275 /*XXX: The s-- is almost definitely wrong here under unicode - demeprhq*/
2278 /* We can use a more efficient search as newlines are the same in unicode as they are in latin */
2280 if (*s++ == '\n') { /* don't need PL_utf8skip here */
2281 if (regtry(®info, &s))
2285 } /* end search for newline */
2286 } /* end anchored/multiline check string search */
2288 } else if (RXf_GPOS_CHECK == (prog->extflags & RXf_GPOS_CHECK))
2290 /* the warning about reginfo.ganch being used without initialization
2291 is bogus -- we set it above, when prog->extflags & RXf_GPOS_SEEN
2292 and we only enter this block when the same bit is set. */
2293 char *tmp_s = reginfo.ganch - prog->gofs;
2295 if (tmp_s >= strbeg && regtry(®info, &tmp_s))
2300 /* Messy cases: unanchored match. */
2301 if ((prog->anchored_substr || prog->anchored_utf8) && prog->intflags & PREGf_SKIP) {
2302 /* we have /x+whatever/ */
2303 /* it must be a one character string (XXXX Except UTF_PATTERN?) */
2308 if (!(utf8_target ? prog->anchored_utf8 : prog->anchored_substr))
2309 utf8_target ? to_utf8_substr(prog) : to_byte_substr(prog);
2310 ch = SvPVX_const(utf8_target ? prog->anchored_utf8 : prog->anchored_substr)[0];
2315 DEBUG_EXECUTE_r( did_match = 1 );
2316 if (regtry(®info, &s)) goto got_it;
2318 while (s < strend && *s == ch)
2326 DEBUG_EXECUTE_r( did_match = 1 );
2327 if (regtry(®info, &s)) goto got_it;
2329 while (s < strend && *s == ch)
2334 DEBUG_EXECUTE_r(if (!did_match)
2335 PerlIO_printf(Perl_debug_log,
2336 "Did not find anchored character...\n")
2339 else if (prog->anchored_substr != NULL
2340 || prog->anchored_utf8 != NULL
2341 || ((prog->float_substr != NULL || prog->float_utf8 != NULL)
2342 && prog->float_max_offset < strend - s)) {
2347 char *last1; /* Last position checked before */
2351 if (prog->anchored_substr || prog->anchored_utf8) {
2352 if (!(utf8_target ? prog->anchored_utf8 : prog->anchored_substr))
2353 utf8_target ? to_utf8_substr(prog) : to_byte_substr(prog);
2354 must = utf8_target ? prog->anchored_utf8 : prog->anchored_substr;
2355 back_max = back_min = prog->anchored_offset;
2357 if (!(utf8_target ? prog->float_utf8 : prog->float_substr))
2358 utf8_target ? to_utf8_substr(prog) : to_byte_substr(prog);
2359 must = utf8_target ? prog->float_utf8 : prog->float_substr;
2360 back_max = prog->float_max_offset;
2361 back_min = prog->float_min_offset;
2365 if (must == &PL_sv_undef)
2366 /* could not downgrade utf8 check substring, so must fail */
2372 last = HOP3c(strend, /* Cannot start after this */
2373 -(I32)(CHR_SVLEN(must)
2374 - (SvTAIL(must) != 0) + back_min), strbeg);
2377 last1 = HOPc(s, -1);
2379 last1 = s - 1; /* bogus */
2381 /* XXXX check_substr already used to find "s", can optimize if
2382 check_substr==must. */
2384 dontbother = end_shift;
2385 strend = HOPc(strend, -dontbother);
2386 while ( (s <= last) &&
2387 ((flags & REXEC_SCREAM)
2388 ? (s = screaminstr(sv, must, HOP3c(s, back_min, (back_min<0 ? strbeg : strend)) - strbeg,
2389 end_shift, &scream_pos, 0))
2390 : (s = fbm_instr((unsigned char*)HOP3(s, back_min, (back_min<0 ? strbeg : strend)),
2391 (unsigned char*)strend, must,
2392 multiline ? FBMrf_MULTILINE : 0))) ) {
2393 /* we may be pointing at the wrong string */
2394 if ((flags & REXEC_SCREAM) && RXp_MATCH_COPIED(prog))
2395 s = strbeg + (s - SvPVX_const(sv));
2396 DEBUG_EXECUTE_r( did_match = 1 );
2397 if (HOPc(s, -back_max) > last1) {
2398 last1 = HOPc(s, -back_min);
2399 s = HOPc(s, -back_max);
2402 char * const t = (last1 >= PL_bostr) ? HOPc(last1, 1) : last1 + 1;
2404 last1 = HOPc(s, -back_min);
2408 while (s <= last1) {
2409 if (regtry(®info, &s))
2415 while (s <= last1) {
2416 if (regtry(®info, &s))
2422 DEBUG_EXECUTE_r(if (!did_match) {
2423 RE_PV_QUOTED_DECL(quoted, utf8_target, PERL_DEBUG_PAD_ZERO(0),
2424 SvPVX_const(must), RE_SV_DUMPLEN(must), 30);
2425 PerlIO_printf(Perl_debug_log, "Did not find %s substr %s%s...\n",
2426 ((must == prog->anchored_substr || must == prog->anchored_utf8)
2427 ? "anchored" : "floating"),
2428 quoted, RE_SV_TAIL(must));
2432 else if ( (c = progi->regstclass) ) {
2434 const OPCODE op = OP(progi->regstclass);
2435 /* don't bother with what can't match */
2436 if (PL_regkind[op] != EXACT && op != CANY && PL_regkind[op] != TRIE)
2437 strend = HOPc(strend, -(minlen - 1));
2440 SV * const prop = sv_newmortal();
2441 regprop(prog, prop, c);
2443 RE_PV_QUOTED_DECL(quoted,utf8_target,PERL_DEBUG_PAD_ZERO(1),
2445 PerlIO_printf(Perl_debug_log,
2446 "Matching stclass %.*s against %s (%d bytes)\n",
2447 (int)SvCUR(prop), SvPVX_const(prop),
2448 quoted, (int)(strend - s));
2451 if (find_byclass(prog, c, s, strend, ®info))
2453 DEBUG_EXECUTE_r(PerlIO_printf(Perl_debug_log, "Contradicts stclass... [regexec_flags]\n"));
2457 if (prog->float_substr != NULL || prog->float_utf8 != NULL) {
2462 if (!(utf8_target ? prog->float_utf8 : prog->float_substr))
2463 utf8_target ? to_utf8_substr(prog) : to_byte_substr(prog);
2464 float_real = utf8_target ? prog->float_utf8 : prog->float_substr;
2466 if (flags & REXEC_SCREAM) {
2467 last = screaminstr(sv, float_real, s - strbeg,
2468 end_shift, &scream_pos, 1); /* last one */
2470 last = scream_olds; /* Only one occurrence. */
2471 /* we may be pointing at the wrong string */
2472 else if (RXp_MATCH_COPIED(prog))
2473 s = strbeg + (s - SvPVX_const(sv));
2477 const char * const little = SvPV_const(float_real, len);
2479 if (SvTAIL(float_real)) {
2480 if (memEQ(strend - len + 1, little, len - 1))
2481 last = strend - len + 1;
2482 else if (!multiline)
2483 last = memEQ(strend - len, little, len)
2484 ? strend - len : NULL;
2490 last = rninstr(s, strend, little, little + len);
2492 last = strend; /* matching "$" */
2497 PerlIO_printf(Perl_debug_log,
2498 "%sCan't trim the tail, match fails (should not happen)%s\n",
2499 PL_colors[4], PL_colors[5]));
2500 goto phooey; /* Should not happen! */
2502 dontbother = strend - last + prog->float_min_offset;
2504 if (minlen && (dontbother < minlen))
2505 dontbother = minlen - 1;
2506 strend -= dontbother; /* this one's always in bytes! */
2507 /* We don't know much -- general case. */
2510 if (regtry(®info, &s))
2519 if (regtry(®info, &s))
2521 } while (s++ < strend);
2530 RX_MATCH_TAINTED_set(rx, PL_reg_flags & RF_tainted);
2532 if (PL_reg_eval_set)
2533 restore_pos(aTHX_ prog);
2534 if (RXp_PAREN_NAMES(prog))
2535 (void)hv_iterinit(RXp_PAREN_NAMES(prog));
2537 /* make sure $`, $&, $', and $digit will work later */
2538 if ( !(flags & REXEC_NOT_FIRST) ) {
2539 RX_MATCH_COPY_FREE(rx);
2540 if (flags & REXEC_COPY_STR) {
2541 const I32 i = PL_regeol - startpos + (stringarg - strbeg);
2542 #ifdef PERL_OLD_COPY_ON_WRITE
2544 || (SvFLAGS(sv) & CAN_COW_MASK) == CAN_COW_FLAGS)) {
2546 PerlIO_printf(Perl_debug_log,
2547 "Copy on write: regexp capture, type %d\n",
2550 prog->saved_copy = sv_setsv_cow(prog->saved_copy, sv);
2551 prog->subbeg = (char *)SvPVX_const(prog->saved_copy);
2552 assert (SvPOKp(prog->saved_copy));
2556 RX_MATCH_COPIED_on(rx);
2557 s = savepvn(strbeg, i);
2563 prog->subbeg = strbeg;
2564 prog->sublen = PL_regeol - strbeg; /* strend may have been modified */
2571 DEBUG_EXECUTE_r(PerlIO_printf(Perl_debug_log, "%sMatch failed%s\n",
2572 PL_colors[4], PL_colors[5]));
2573 if (PL_reg_eval_set)
2574 restore_pos(aTHX_ prog);
2576 /* we failed :-( roll it back */
2577 Safefree(prog->offs);
2586 - regtry - try match at specific point
2588 STATIC I32 /* 0 failure, 1 success */
2589 S_regtry(pTHX_ regmatch_info *reginfo, char **startpos)
2593 REGEXP *const rx = reginfo->prog;
2594 regexp *const prog = (struct regexp *)SvANY(rx);
2595 RXi_GET_DECL(prog,progi);
2596 GET_RE_DEBUG_FLAGS_DECL;
2598 PERL_ARGS_ASSERT_REGTRY;
2600 reginfo->cutpoint=NULL;
2602 if ((prog->extflags & RXf_EVAL_SEEN) && !PL_reg_eval_set) {
2605 PL_reg_eval_set = RS_init;
2606 DEBUG_EXECUTE_r(DEBUG_s(
2607 PerlIO_printf(Perl_debug_log, " setting stack tmpbase at %"IVdf"\n",
2608 (IV)(PL_stack_sp - PL_stack_base));
2611 cxstack[cxstack_ix].blk_oldsp = PL_stack_sp - PL_stack_base;
2612 /* Otherwise OP_NEXTSTATE will free whatever on stack now. */
2614 /* Apparently this is not needed, judging by wantarray. */
2615 /* SAVEI8(cxstack[cxstack_ix].blk_gimme);
2616 cxstack[cxstack_ix].blk_gimme = G_SCALAR; */
2619 /* Make $_ available to executed code. */
2620 if (reginfo->sv != DEFSV) {
2622 DEFSV_set(reginfo->sv);
2625 if (!(SvTYPE(reginfo->sv) >= SVt_PVMG && SvMAGIC(reginfo->sv)
2626 && (mg = mg_find(reginfo->sv, PERL_MAGIC_regex_global)))) {
2627 /* prepare for quick setting of pos */
2628 #ifdef PERL_OLD_COPY_ON_WRITE
2629 if (SvIsCOW(reginfo->sv))
2630 sv_force_normal_flags(reginfo->sv, 0);
2632 mg = sv_magicext(reginfo->sv, NULL, PERL_MAGIC_regex_global,
2633 &PL_vtbl_mglob, NULL, 0);
2637 PL_reg_oldpos = mg->mg_len;
2638 SAVEDESTRUCTOR_X(restore_pos, prog);
2640 if (!PL_reg_curpm) {
2641 Newxz(PL_reg_curpm, 1, PMOP);
2644 SV* const repointer = &PL_sv_undef;
2645 /* this regexp is also owned by the new PL_reg_curpm, which
2646 will try to free it. */
2647 av_push(PL_regex_padav, repointer);
2648 PL_reg_curpm->op_pmoffset = av_len(PL_regex_padav);
2649 PL_regex_pad = AvARRAY(PL_regex_padav);
2654 /* It seems that non-ithreads works both with and without this code.
2655 So for efficiency reasons it seems best not to have the code
2656 compiled when it is not needed. */
2657 /* This is safe against NULLs: */
2658 ReREFCNT_dec(PM_GETRE(PL_reg_curpm));
2659 /* PM_reg_curpm owns a reference to this regexp. */
2662 PM_SETRE(PL_reg_curpm, rx);
2663 PL_reg_oldcurpm = PL_curpm;
2664 PL_curpm = PL_reg_curpm;
2665 if (RXp_MATCH_COPIED(prog)) {
2666 /* Here is a serious problem: we cannot rewrite subbeg,
2667 since it may be needed if this match fails. Thus
2668 $` inside (?{}) could fail... */
2669 PL_reg_oldsaved = prog->subbeg;
2670 PL_reg_oldsavedlen = prog->sublen;
2671 #ifdef PERL_OLD_COPY_ON_WRITE
2672 PL_nrs = prog->saved_copy;
2674 RXp_MATCH_COPIED_off(prog);
2677 PL_reg_oldsaved = NULL;
2678 prog->subbeg = PL_bostr;
2679 prog->sublen = PL_regeol - PL_bostr; /* strend may have been modified */
2681 DEBUG_EXECUTE_r(PL_reg_starttry = *startpos);
2682 prog->offs[0].start = *startpos - PL_bostr;
2683 PL_reginput = *startpos;
2684 PL_reglastparen = &prog->lastparen;
2685 PL_reglastcloseparen = &prog->lastcloseparen;
2686 prog->lastparen = 0;
2687 prog->lastcloseparen = 0;
2689 PL_regoffs = prog->offs;
2690 if (PL_reg_start_tmpl <= prog->nparens) {
2691 PL_reg_start_tmpl = prog->nparens*3/2 + 3;
2692 if(PL_reg_start_tmp)
2693 Renew(PL_reg_start_tmp, PL_reg_start_tmpl, char*);
2695 Newx(PL_reg_start_tmp, PL_reg_start_tmpl, char*);
2698 /* XXXX What this code is doing here?!!! There should be no need
2699 to do this again and again, PL_reglastparen should take care of
2702 /* Tests pat.t#187 and split.t#{13,14} seem to depend on this code.
2703 * Actually, the code in regcppop() (which Ilya may be meaning by
2704 * PL_reglastparen), is not needed at all by the test suite
2705 * (op/regexp, op/pat, op/split), but that code is needed otherwise
2706 * this erroneously leaves $1 defined: "1" =~ /^(?:(\d)x)?\d$/
2707 * Meanwhile, this code *is* needed for the
2708 * above-mentioned test suite tests to succeed. The common theme
2709 * on those tests seems to be returning null fields from matches.
2710 * --jhi updated by dapm */
2712 if (prog->nparens) {
2713 regexp_paren_pair *pp = PL_regoffs;
2715 for (i = prog->nparens; i > (I32)*PL_reglastparen; i--) {
2723 if (regmatch(reginfo, progi->program + 1)) {
2724 PL_regoffs[0].end = PL_reginput - PL_bostr;
2727 if (reginfo->cutpoint)
2728 *startpos= reginfo->cutpoint;
2729 REGCP_UNWIND(lastcp);
2734 #define sayYES goto yes
2735 #define sayNO goto no
2736 #define sayNO_SILENT goto no_silent
2738 /* we dont use STMT_START/END here because it leads to
2739 "unreachable code" warnings, which are bogus, but distracting. */
2740 #define CACHEsayNO \
2741 if (ST.cache_mask) \
2742 PL_reg_poscache[ST.cache_offset] |= ST.cache_mask; \
2745 /* this is used to determine how far from the left messages like
2746 'failed...' are printed. It should be set such that messages
2747 are inline with the regop output that created them.
2749 #define REPORT_CODE_OFF 32
2752 #define CHRTEST_UNINIT -1001 /* c1/c2 haven't been calculated yet */
2753 #define CHRTEST_VOID -1000 /* the c1/c2 "next char" test should be skipped */
2755 #define SLAB_FIRST(s) (&(s)->states[0])
2756 #define SLAB_LAST(s) (&(s)->states[PERL_REGMATCH_SLAB_SLOTS-1])
2758 /* grab a new slab and return the first slot in it */
2760 STATIC regmatch_state *
2763 #if PERL_VERSION < 9 && !defined(PERL_CORE)
2766 regmatch_slab *s = PL_regmatch_slab->next;
2768 Newx(s, 1, regmatch_slab);
2769 s->prev = PL_regmatch_slab;
2771 PL_regmatch_slab->next = s;
2773 PL_regmatch_slab = s;
2774 return SLAB_FIRST(s);
2778 /* push a new state then goto it */
2780 #define PUSH_STATE_GOTO(state, node) \
2782 st->resume_state = state; \
2785 /* push a new state with success backtracking, then goto it */
2787 #define PUSH_YES_STATE_GOTO(state, node) \
2789 st->resume_state = state; \
2790 goto push_yes_state;
2796 regmatch() - main matching routine
2798 This is basically one big switch statement in a loop. We execute an op,
2799 set 'next' to point the next op, and continue. If we come to a point which
2800 we may need to backtrack to on failure such as (A|B|C), we push a
2801 backtrack state onto the backtrack stack. On failure, we pop the top
2802 state, and re-enter the loop at the state indicated. If there are no more
2803 states to pop, we return failure.
2805 Sometimes we also need to backtrack on success; for example /A+/, where
2806 after successfully matching one A, we need to go back and try to
2807 match another one; similarly for lookahead assertions: if the assertion
2808 completes successfully, we backtrack to the state just before the assertion
2809 and then carry on. In these cases, the pushed state is marked as
2810 'backtrack on success too'. This marking is in fact done by a chain of
2811 pointers, each pointing to the previous 'yes' state. On success, we pop to
2812 the nearest yes state, discarding any intermediate failure-only states.
2813 Sometimes a yes state is pushed just to force some cleanup code to be
2814 called at the end of a successful match or submatch; e.g. (??{$re}) uses
2815 it to free the inner regex.
2817 Note that failure backtracking rewinds the cursor position, while
2818 success backtracking leaves it alone.
2820 A pattern is complete when the END op is executed, while a subpattern
2821 such as (?=foo) is complete when the SUCCESS op is executed. Both of these
2822 ops trigger the "pop to last yes state if any, otherwise return true"
2825 A common convention in this function is to use A and B to refer to the two
2826 subpatterns (or to the first nodes thereof) in patterns like /A*B/: so A is
2827 the subpattern to be matched possibly multiple times, while B is the entire
2828 rest of the pattern. Variable and state names reflect this convention.
2830 The states in the main switch are the union of ops and failure/success of
2831 substates associated with with that op. For example, IFMATCH is the op
2832 that does lookahead assertions /(?=A)B/ and so the IFMATCH state means
2833 'execute IFMATCH'; while IFMATCH_A is a state saying that we have just
2834 successfully matched A and IFMATCH_A_fail is a state saying that we have
2835 just failed to match A. Resume states always come in pairs. The backtrack
2836 state we push is marked as 'IFMATCH_A', but when that is popped, we resume
2837 at IFMATCH_A or IFMATCH_A_fail, depending on whether we are backtracking
2838 on success or failure.
2840 The struct that holds a backtracking state is actually a big union, with
2841 one variant for each major type of op. The variable st points to the
2842 top-most backtrack struct. To make the code clearer, within each
2843 block of code we #define ST to alias the relevant union.
2845 Here's a concrete example of a (vastly oversimplified) IFMATCH
2851 #define ST st->u.ifmatch
2853 case IFMATCH: // we are executing the IFMATCH op, (?=A)B
2854 ST.foo = ...; // some state we wish to save
2856 // push a yes backtrack state with a resume value of
2857 // IFMATCH_A/IFMATCH_A_fail, then continue execution at the
2859 PUSH_YES_STATE_GOTO(IFMATCH_A, A);
2862 case IFMATCH_A: // we have successfully executed A; now continue with B
2864 bar = ST.foo; // do something with the preserved value
2867 case IFMATCH_A_fail: // A failed, so the assertion failed
2868 ...; // do some housekeeping, then ...
2869 sayNO; // propagate the failure
2876 For any old-timers reading this who are familiar with the old recursive
2877 approach, the code above is equivalent to:
2879 case IFMATCH: // we are executing the IFMATCH op, (?=A)B
2888 ...; // do some housekeeping, then ...
2889 sayNO; // propagate the failure
2892 The topmost backtrack state, pointed to by st, is usually free. If you
2893 want to claim it, populate any ST.foo fields in it with values you wish to
2894 save, then do one of
2896 PUSH_STATE_GOTO(resume_state, node);
2897 PUSH_YES_STATE_GOTO(resume_state, node);
2899 which sets that backtrack state's resume value to 'resume_state', pushes a
2900 new free entry to the top of the backtrack stack, then goes to 'node'.
2901 On backtracking, the free slot is popped, and the saved state becomes the
2902 new free state. An ST.foo field in this new top state can be temporarily
2903 accessed to retrieve values, but once the main loop is re-entered, it
2904 becomes available for reuse.
2906 Note that the depth of the backtrack stack constantly increases during the
2907 left-to-right execution of the pattern, rather than going up and down with
2908 the pattern nesting. For example the stack is at its maximum at Z at the
2909 end of the pattern, rather than at X in the following:
2911 /(((X)+)+)+....(Y)+....Z/
2913 The only exceptions to this are lookahead/behind assertions and the cut,
2914 (?>A), which pop all the backtrack states associated with A before
2917 Backtrack state structs are allocated in slabs of about 4K in size.
2918 PL_regmatch_state and st always point to the currently active state,
2919 and PL_regmatch_slab points to the slab currently containing
2920 PL_regmatch_state. The first time regmatch() is called, the first slab is
2921 allocated, and is never freed until interpreter destruction. When the slab
2922 is full, a new one is allocated and chained to the end. At exit from
2923 regmatch(), slabs allocated since entry are freed.
2928 #define DEBUG_STATE_pp(pp) \
2930 DUMP_EXEC_POS(locinput, scan, utf8_target); \
2931 PerlIO_printf(Perl_debug_log, \
2932 " %*s"pp" %s%s%s%s%s\n", \
2934 PL_reg_name[st->resume_state], \
2935 ((st==yes_state||st==mark_state) ? "[" : ""), \
2936 ((st==yes_state) ? "Y" : ""), \
2937 ((st==mark_state) ? "M" : ""), \
2938 ((st==yes_state||st==mark_state) ? "]" : "") \
2943 #define REG_NODE_NUM(x) ((x) ? (int)((x)-prog) : -1)
2948 S_debug_start_match(pTHX_ const REGEXP *prog, const bool utf8_target,
2949 const char *start, const char *end, const char *blurb)
2951 const bool utf8_pat = RX_UTF8(prog) ? 1 : 0;
2953 PERL_ARGS_ASSERT_DEBUG_START_MATCH;
2958 RE_PV_QUOTED_DECL(s0, utf8_pat, PERL_DEBUG_PAD_ZERO(0),
2959 RX_PRECOMP_const(prog), RX_PRELEN(prog), 60);
2961 RE_PV_QUOTED_DECL(s1, utf8_target, PERL_DEBUG_PAD_ZERO(1),
2962 start, end - start, 60);
2964 PerlIO_printf(Perl_debug_log,
2965 "%s%s REx%s %s against %s\n",
2966 PL_colors[4], blurb, PL_colors[5], s0, s1);
2968 if (utf8_target||utf8_pat)
2969 PerlIO_printf(Perl_debug_log, "UTF-8 %s%s%s...\n",
2970 utf8_pat ? "pattern" : "",
2971 utf8_pat && utf8_target ? " and " : "",
2972 utf8_target ? "string" : ""
2978 S_dump_exec_pos(pTHX_ const char *locinput,
2979 const regnode *scan,
2980 const char *loc_regeol,
2981 const char *loc_bostr,
2982 const char *loc_reg_starttry,
2983 const bool utf8_target)
2985 const int docolor = *PL_colors[0] || *PL_colors[2] || *PL_colors[4];
2986 const int taill = (docolor ? 10 : 7); /* 3 chars for "> <" */
2987 int l = (loc_regeol - locinput) > taill ? taill : (loc_regeol - locinput);
2988 /* The part of the string before starttry has one color
2989 (pref0_len chars), between starttry and current
2990 position another one (pref_len - pref0_len chars),
2991 after the current position the third one.
2992 We assume that pref0_len <= pref_len, otherwise we
2993 decrease pref0_len. */
2994 int pref_len = (locinput - loc_bostr) > (5 + taill) - l
2995 ? (5 + taill) - l : locinput - loc_bostr;
2998 PERL_ARGS_ASSERT_DUMP_EXEC_POS;
3000 while (utf8_target && UTF8_IS_CONTINUATION(*(U8*)(locinput - pref_len)))
3002 pref0_len = pref_len - (locinput - loc_reg_starttry);
3003 if (l + pref_len < (5 + taill) && l < loc_regeol - locinput)
3004 l = ( loc_regeol - locinput > (5 + taill) - pref_len
3005 ? (5 + taill) - pref_len : loc_regeol - locinput);
3006 while (utf8_target && UTF8_IS_CONTINUATION(*(U8*)(locinput + l)))
3010 if (pref0_len > pref_len)
3011 pref0_len = pref_len;
3013 const int is_uni = (utf8_target && OP(scan) != CANY) ? 1 : 0;
3015 RE_PV_COLOR_DECL(s0,len0,is_uni,PERL_DEBUG_PAD(0),
3016 (locinput - pref_len),pref0_len, 60, 4, 5);
3018 RE_PV_COLOR_DECL(s1,len1,is_uni,PERL_DEBUG_PAD(1),
3019 (locinput - pref_len + pref0_len),
3020 pref_len - pref0_len, 60, 2, 3);
3022 RE_PV_COLOR_DECL(s2,len2,is_uni,PERL_DEBUG_PAD(2),
3023 locinput, loc_regeol - locinput, 10, 0, 1);
3025 const STRLEN tlen=len0+len1+len2;
3026 PerlIO_printf(Perl_debug_log,
3027 "%4"IVdf" <%.*s%.*s%s%.*s>%*s|",
3028 (IV)(locinput - loc_bostr),
3031 (docolor ? "" : "> <"),
3033 (int)(tlen > 19 ? 0 : 19 - tlen),
3040 /* reg_check_named_buff_matched()
3041 * Checks to see if a named buffer has matched. The data array of
3042 * buffer numbers corresponding to the buffer is expected to reside
3043 * in the regexp->data->data array in the slot stored in the ARG() of
3044 * node involved. Note that this routine doesn't actually care about the
3045 * name, that information is not preserved from compilation to execution.
3046 * Returns the index of the leftmost defined buffer with the given name
3047 * or 0 if non of the buffers matched.
3050 S_reg_check_named_buff_matched(pTHX_ const regexp *rex, const regnode *scan)
3053 RXi_GET_DECL(rex,rexi);
3054 SV *sv_dat= MUTABLE_SV(rexi->data->data[ ARG( scan ) ]);
3055 I32 *nums=(I32*)SvPVX(sv_dat);
3057 PERL_ARGS_ASSERT_REG_CHECK_NAMED_BUFF_MATCHED;
3059 for ( n=0; n<SvIVX(sv_dat); n++ ) {
3060 if ((I32)*PL_reglastparen >= nums[n] &&
3061 PL_regoffs[nums[n]].end != -1)
3070 /* free all slabs above current one - called during LEAVE_SCOPE */
3073 S_clear_backtrack_stack(pTHX_ void *p)
3075 regmatch_slab *s = PL_regmatch_slab->next;
3080 PL_regmatch_slab->next = NULL;
3082 regmatch_slab * const osl = s;
3089 #define SETREX(Re1,Re2) \
3090 if (PL_reg_eval_set) PM_SETRE((PL_reg_curpm), (Re2)); \
3093 STATIC I32 /* 0 failure, 1 success */
3094 S_regmatch(pTHX_ regmatch_info *reginfo, regnode *prog)
3096 #if PERL_VERSION < 9 && !defined(PERL_CORE)
3100 register const bool utf8_target = PL_reg_match_utf8;
3101 const U32 uniflags = UTF8_ALLOW_DEFAULT;
3102 REGEXP *rex_sv = reginfo->prog;
3103 regexp *rex = (struct regexp *)SvANY(rex_sv);
3104 RXi_GET_DECL(rex,rexi);
3106 /* the current state. This is a cached copy of PL_regmatch_state */
3107 register regmatch_state *st;
3108 /* cache heavy used fields of st in registers */
3109 register regnode *scan;
3110 register regnode *next;
3111 register U32 n = 0; /* general value; init to avoid compiler warning */
3112 register I32 ln = 0; /* len or last; init to avoid compiler warning */
3113 register char *locinput = PL_reginput;
3114 register I32 nextchr; /* is always set to UCHARAT(locinput) */
3116 bool result = 0; /* return value of S_regmatch */
3117 int depth = 0; /* depth of backtrack stack */
3118 U32 nochange_depth = 0; /* depth of GOSUB recursion with nochange */
3119 const U32 max_nochange_depth =
3120 (3 * rex->nparens > MAX_RECURSE_EVAL_NOCHANGE_DEPTH) ?
3121 3 * rex->nparens : MAX_RECURSE_EVAL_NOCHANGE_DEPTH;
3122 regmatch_state *yes_state = NULL; /* state to pop to on success of
3124 /* mark_state piggy backs on the yes_state logic so that when we unwind
3125 the stack on success we can update the mark_state as we go */
3126 regmatch_state *mark_state = NULL; /* last mark state we have seen */
3127 regmatch_state *cur_eval = NULL; /* most recent EVAL_AB state */
3128 struct regmatch_state *cur_curlyx = NULL; /* most recent curlyx */
3130 bool no_final = 0; /* prevent failure from backtracking? */
3131 bool do_cutgroup = 0; /* no_final only until next branch/trie entry */
3132 char *startpoint = PL_reginput;
3133 SV *popmark = NULL; /* are we looking for a mark? */
3134 SV *sv_commit = NULL; /* last mark name seen in failure */
3135 SV *sv_yes_mark = NULL; /* last mark name we have seen
3136 during a successful match */
3137 U32 lastopen = 0; /* last open we saw */
3138 bool has_cutgroup = RX_HAS_CUTGROUP(rex) ? 1 : 0;
3139 SV* const oreplsv = GvSV(PL_replgv);
3140 /* these three flags are set by various ops to signal information to
3141 * the very next op. They have a useful lifetime of exactly one loop
3142 * iteration, and are not preserved or restored by state pushes/pops
3144 bool sw = 0; /* the condition value in (?(cond)a|b) */
3145 bool minmod = 0; /* the next "{n,m}" is a "{n,m}?" */
3146 int logical = 0; /* the following EVAL is:
3150 or the following IFMATCH/UNLESSM is:
3151 false: plain (?=foo)
3152 true: used as a condition: (?(?=foo))
3155 GET_RE_DEBUG_FLAGS_DECL;
3158 PERL_ARGS_ASSERT_REGMATCH;
3160 DEBUG_OPTIMISE_r( DEBUG_EXECUTE_r({
3161 PerlIO_printf(Perl_debug_log,"regmatch start\n");
3163 /* on first ever call to regmatch, allocate first slab */
3164 if (!PL_regmatch_slab) {
3165 Newx(PL_regmatch_slab, 1, regmatch_slab);
3166 PL_regmatch_slab->prev = NULL;
3167 PL_regmatch_slab->next = NULL;
3168 PL_regmatch_state = SLAB_FIRST(PL_regmatch_slab);
3171 oldsave = PL_savestack_ix;
3172 SAVEDESTRUCTOR_X(S_clear_backtrack_stack, NULL);
3173 SAVEVPTR(PL_regmatch_slab);
3174 SAVEVPTR(PL_regmatch_state);
3176 /* grab next free state slot */
3177 st = ++PL_regmatch_state;
3178 if (st > SLAB_LAST(PL_regmatch_slab))
3179 st = PL_regmatch_state = S_push_slab(aTHX);
3181 /* Note that nextchr is a byte even in UTF */
3182 nextchr = UCHARAT(locinput);
3184 while (scan != NULL) {
3187 SV * const prop = sv_newmortal();
3188 regnode *rnext=regnext(scan);
3189 DUMP_EXEC_POS( locinput, scan, utf8_target );
3190 regprop(rex, prop, scan);
3192 PerlIO_printf(Perl_debug_log,
3193 "%3"IVdf":%*s%s(%"IVdf")\n",
3194 (IV)(scan - rexi->program), depth*2, "",
3196 (PL_regkind[OP(scan)] == END || !rnext) ?
3197 0 : (IV)(rnext - rexi->program));
3200 next = scan + NEXT_OFF(scan);
3203 state_num = OP(scan);
3207 assert(PL_reglastparen == &rex->lastparen);
3208 assert(PL_reglastcloseparen == &rex->lastcloseparen);
3209 assert(PL_regoffs == rex->offs);
3211 switch (state_num) {
3213 if (locinput == PL_bostr)
3215 /* reginfo->till = reginfo->bol; */
3220 if (locinput == PL_bostr ||
3221 ((nextchr || locinput < PL_regeol) && locinput[-1] == '\n'))
3227 if (locinput == PL_bostr)
3231 if (locinput == reginfo->ganch)
3236 /* update the startpoint */
3237 st->u.keeper.val = PL_regoffs[0].start;
3238 PL_reginput = locinput;
3239 PL_regoffs[0].start = locinput - PL_bostr;
3240 PUSH_STATE_GOTO(KEEPS_next, next);
3242 case KEEPS_next_fail:
3243 /* rollback the start point change */
3244 PL_regoffs[0].start = st->u.keeper.val;
3250 if ((nextchr || locinput < PL_regeol) && nextchr != '\n')
3255 if ((nextchr || locinput < PL_regeol) && nextchr != '\n')
3257 if (PL_regeol - locinput > 1)
3261 if (PL_regeol != locinput)
3265 if (!nextchr && locinput >= PL_regeol)
3268 locinput += PL_utf8skip[nextchr];
3269 if (locinput > PL_regeol)
3271 nextchr = UCHARAT(locinput);
3274 nextchr = UCHARAT(++locinput);
3277 if (!nextchr && locinput >= PL_regeol)
3279 nextchr = UCHARAT(++locinput);
3282 if ((!nextchr && locinput >= PL_regeol) || nextchr == '\n')
3285 locinput += PL_utf8skip[nextchr];
3286 if (locinput > PL_regeol)
3288 nextchr = UCHARAT(locinput);
3291 nextchr = UCHARAT(++locinput);
3295 #define ST st->u.trie
3297 /* In this case the charclass data is available inline so
3298 we can fail fast without a lot of extra overhead.
3300 if (scan->flags == EXACT || !utf8_target) {
3301 if(!ANYOF_BITMAP_TEST(scan, *locinput)) {
3303 PerlIO_printf(Perl_debug_log,
3304 "%*s %sfailed to match trie start class...%s\n",
3305 REPORT_CODE_OFF+depth*2, "", PL_colors[4], PL_colors[5])
3313 /* the basic plan of execution of the trie is:
3314 * At the beginning, run though all the states, and
3315 * find the longest-matching word. Also remember the position
3316 * of the shortest matching word. For example, this pattern:
3319 * when matched against the string "abcde", will generate
3320 * accept states for all words except 3, with the longest
3321 * matching word being 4, and the shortest being 1 (with
3322 * the position being after char 1 of the string).
3324 * Then for each matching word, in word order (i.e. 1,2,4,5),
3325 * we run the remainder of the pattern; on each try setting
3326 * the current position to the character following the word,
3327 * returning to try the next word on failure.
3329 * We avoid having to build a list of words at runtime by
3330 * using a compile-time structure, wordinfo[].prev, which
3331 * gives, for each word, the previous accepting word (if any).
3332 * In the case above it would contain the mappings 1->2, 2->0,
3333 * 3->0, 4->5, 5->1. We can use this table to generate, from
3334 * the longest word (4 above), a list of all words, by
3335 * following the list of prev pointers; this gives us the
3336 * unordered list 4,5,1,2. Then given the current word we have
3337 * just tried, we can go through the list and find the
3338 * next-biggest word to try (so if we just failed on word 2,
3339 * the next in the list is 4).
3341 * Since at runtime we don't record the matching position in
3342 * the string for each word, we have to work that out for
3343 * each word we're about to process. The wordinfo table holds
3344 * the character length of each word; given that we recorded
3345 * at the start: the position of the shortest word and its
3346 * length in chars, we just need to move the pointer the
3347 * difference between the two char lengths. Depending on
3348 * Unicode status and folding, that's cheap or expensive.
3350 * This algorithm is optimised for the case where are only a
3351 * small number of accept states, i.e. 0,1, or maybe 2.
3352 * With lots of accepts states, and having to try all of them,
3353 * it becomes quadratic on number of accept states to find all
3358 /* what type of TRIE am I? (utf8 makes this contextual) */
3359 DECL_TRIE_TYPE(scan);
3361 /* what trie are we using right now */
3362 reg_trie_data * const trie
3363 = (reg_trie_data*)rexi->data->data[ ARG( scan ) ];
3364 HV * widecharmap = MUTABLE_HV(rexi->data->data[ ARG( scan ) + 1 ]);
3365 U32 state = trie->startstate;
3367 if (trie->bitmap && trie_type != trie_utf8_fold &&
3368 !TRIE_BITMAP_TEST(trie,*locinput)
3370 if (trie->states[ state ].wordnum) {
3372 PerlIO_printf(Perl_debug_log,
3373 "%*s %smatched empty string...%s\n",
3374 REPORT_CODE_OFF+depth*2, "", PL_colors[4], PL_colors[5])
3380 PerlIO_printf(Perl_debug_log,
3381 "%*s %sfailed to match trie start class...%s\n",
3382 REPORT_CODE_OFF+depth*2, "", PL_colors[4], PL_colors[5])
3389 U8 *uc = ( U8* )locinput;
3393 U8 *uscan = (U8*)NULL;
3394 U8 foldbuf[ UTF8_MAXBYTES_CASE + 1 ];
3395 U32 charcount = 0; /* how many input chars we have matched */
3396 U32 accepted = 0; /* have we seen any accepting states? */
3399 ST.jump = trie->jump;
3402 ST.longfold = FALSE; /* char longer if folded => it's harder */
3405 /* fully traverse the TRIE; note the position of the
3406 shortest accept state and the wordnum of the longest
3409 while ( state && uc <= (U8*)PL_regeol ) {
3410 U32 base = trie->states[ state ].trans.base;
3414 wordnum = trie->states[ state ].wordnum;
3416 if (wordnum) { /* it's an accept state */
3419 /* record first match position */
3421 ST.firstpos = (U8*)locinput;
3426 ST.firstchars = charcount;
3429 if (!ST.nextword || wordnum < ST.nextword)
3430 ST.nextword = wordnum;
3431 ST.topword = wordnum;
3434 DEBUG_TRIE_EXECUTE_r({
3435 DUMP_EXEC_POS( (char *)uc, scan, utf8_target );
3436 PerlIO_printf( Perl_debug_log,
3437 "%*s %sState: %4"UVxf" Accepted: %c ",
3438 2+depth * 2, "", PL_colors[4],
3439 (UV)state, (accepted ? 'Y' : 'N'));
3442 /* read a char and goto next state */
3445 REXEC_TRIE_READ_CHAR(trie_type, trie, widecharmap, uc,
3446 uscan, len, uvc, charid, foldlen,
3453 base + charid - 1 - trie->uniquecharcount)) >= 0)
3455 && ((U32)offset < trie->lasttrans)
3456 && trie->trans[offset].check == state)
3458 state = trie->trans[offset].next;
3469 DEBUG_TRIE_EXECUTE_r(
3470 PerlIO_printf( Perl_debug_log,
3471 "Charid:%3x CP:%4"UVxf" After State: %4"UVxf"%s\n",
3472 charid, uvc, (UV)state, PL_colors[5] );
3478 /* calculate total number of accept states */
3483 w = trie->wordinfo[w].prev;
3486 ST.accepted = accepted;
3490 PerlIO_printf( Perl_debug_log,
3491 "%*s %sgot %"IVdf" possible matches%s\n",
3492 REPORT_CODE_OFF + depth * 2, "",
3493 PL_colors[4], (IV)ST.accepted, PL_colors[5] );
3495 goto trie_first_try; /* jump into the fail handler */
3499 case TRIE_next_fail: /* we failed - try next alternative */
3501 REGCP_UNWIND(ST.cp);
3502 for (n = *PL_reglastparen; n > ST.lastparen; n--)
3503 PL_regoffs[n].end = -1;
3504 *PL_reglastparen = n;
3506 if (!--ST.accepted) {
3508 PerlIO_printf( Perl_debug_log,
3509 "%*s %sTRIE failed...%s\n",
3510 REPORT_CODE_OFF+depth*2, "",
3517 /* Find next-highest word to process. Note that this code
3518 * is O(N^2) per trie run (O(N) per branch), so keep tight */
3519 register U16 min = 0;
3521 register U16 const nextword = ST.nextword;
3522 register reg_trie_wordinfo * const wordinfo
3523 = ((reg_trie_data*)rexi->data->data[ARG(ST.me)])->wordinfo;
3524 for (word=ST.topword; word; word=wordinfo[word].prev) {
3525 if (word > nextword && (!min || word < min))
3538 ST.lastparen = *PL_reglastparen;
3542 /* find start char of end of current word */
3544 U32 chars; /* how many chars to skip */
3545 U8 *uc = ST.firstpos;
3546 reg_trie_data * const trie
3547 = (reg_trie_data*)rexi->data->data[ARG(ST.me)];
3549 assert((trie->wordinfo[ST.nextword].len - trie->prefixlen)
3551 chars = (trie->wordinfo[ST.nextword].len - trie->prefixlen)
3555 /* the hard option - fold each char in turn and find
3556 * its folded length (which may be different */
3557 U8 foldbuf[UTF8_MAXBYTES_CASE + 1];
3565 uvc = utf8n_to_uvuni((U8*)uc, UTF8_MAXLEN, &len,
3573 uvc = to_uni_fold(uvc, foldbuf, &foldlen);
3578 uvc = utf8n_to_uvuni(uscan, UTF8_MAXLEN, &len,
3592 PL_reginput = (char *)uc;
3595 scan = (ST.jump && ST.jump[ST.nextword])
3596 ? ST.me + ST.jump[ST.nextword]
3600 PerlIO_printf( Perl_debug_log,
3601 "%*s %sTRIE matched word #%d, continuing%s\n",
3602 REPORT_CODE_OFF+depth*2, "",
3609 if (ST.accepted > 1 || has_cutgroup) {
3610 PUSH_STATE_GOTO(TRIE_next, scan);
3613 /* only one choice left - just continue */
3615 AV *const trie_words
3616 = MUTABLE_AV(rexi->data->data[ARG(ST.me)+TRIE_WORDS_OFFSET]);
3617 SV ** const tmp = av_fetch( trie_words,
3619 SV *sv= tmp ? sv_newmortal() : NULL;
3621 PerlIO_printf( Perl_debug_log,
3622 "%*s %sonly one match left, short-circuiting: #%d <%s>%s\n",
3623 REPORT_CODE_OFF+depth*2, "", PL_colors[4],
3625 tmp ? pv_pretty(sv, SvPV_nolen_const(*tmp), SvCUR(*tmp), 0,
3626 PL_colors[0], PL_colors[1],
3627 (SvUTF8(*tmp) ? PERL_PV_ESCAPE_UNI : 0)|PERL_PV_ESCAPE_NONASCII
3629 : "not compiled under -Dr",
3633 locinput = PL_reginput;
3634 nextchr = UCHARAT(locinput);
3635 continue; /* execute rest of RE */
3640 char *s = STRING(scan);
3642 if (utf8_target != UTF_PATTERN) {
3643 /* The target and the pattern have differing utf8ness. */
3645 const char * const e = s + ln;
3648 /* The target is utf8, the pattern is not utf8. */
3653 if (NATIVE_TO_UNI(*(U8*)s) !=
3654 utf8n_to_uvuni((U8*)l, UTF8_MAXBYTES, &ulen,
3662 /* The target is not utf8, the pattern is utf8. */
3667 if (NATIVE_TO_UNI(*((U8*)l)) !=
3668 utf8n_to_uvuni((U8*)s, UTF8_MAXBYTES, &ulen,
3676 nextchr = UCHARAT(locinput);
3679 /* The target and the pattern have the same utf8ness. */
3680 /* Inline the first character, for speed. */
3681 if (UCHARAT(s) != nextchr)
3683 if (PL_regeol - locinput < ln)
3685 if (ln > 1 && memNE(s, locinput, ln))
3688 nextchr = UCHARAT(locinput);
3693 const U8 * fold_array;
3696 PL_reg_flags |= RF_tainted;
3697 folder = foldEQ_locale;
3698 fold_array = PL_fold_locale;
3702 folder = foldEQ_latin1;
3703 fold_array = PL_fold_latin1;
3708 fold_array = PL_fold;
3714 if (utf8_target || UTF_PATTERN) {
3715 /* Either target or the pattern are utf8. */
3716 const char * const l = locinput;
3717 char *e = PL_regeol;
3719 if (! foldEQ_utf8(s, 0, ln, cBOOL(UTF_PATTERN),
3720 l, &e, 0, utf8_target)) {
3721 /* One more case for the sharp s:
3722 * pack("U0U*", 0xDF) =~ /ss/i,
3723 * the 0xC3 0x9F are the UTF-8
3724 * byte sequence for the U+00DF. */
3726 if (!(utf8_target &&
3727 toLOWER(s[0]) == 's' &&
3729 toLOWER(s[1]) == 's' &&
3736 nextchr = UCHARAT(locinput);
3740 /* Neither the target and the pattern are utf8. */
3742 /* Inline the first character, for speed. */
3743 if (UCHARAT(s) != nextchr &&
3744 UCHARAT(s) != fold_array[nextchr])
3748 if (PL_regeol - locinput < ln)
3750 if (ln > 1 && ! folder(s, locinput, ln))
3753 nextchr = UCHARAT(locinput);
3757 /* XXX Could improve efficiency by separating these all out using a
3758 * macro or in-line function. At that point regcomp.c would no longer
3759 * have to set the FLAGS fields of these */
3762 PL_reg_flags |= RF_tainted;
3770 /* was last char in word? */
3771 if (utf8_target && FLAGS(scan) != REGEX_ASCII_RESTRICTED_CHARSET) {
3772 if (locinput == PL_bostr)
3775 const U8 * const r = reghop3((U8*)locinput, -1, (U8*)PL_bostr);
3777 ln = utf8n_to_uvchr(r, UTF8SKIP(r), 0, uniflags);
3779 if (FLAGS(scan) != REGEX_LOCALE_CHARSET) {
3780 ln = isALNUM_uni(ln);
3781 LOAD_UTF8_CHARCLASS_ALNUM();
3782 n = swash_fetch(PL_utf8_alnum, (U8*)locinput, utf8_target);
3785 ln = isALNUM_LC_uvchr(UNI_TO_NATIVE(ln));
3786 n = isALNUM_LC_utf8((U8*)locinput);
3791 /* Here the string isn't utf8, or is utf8 and only ascii
3792 * characters are to match \w. In the latter case looking at
3793 * the byte just prior to the current one may be just the final
3794 * byte of a multi-byte character. This is ok. There are two
3796 * 1) it is a single byte character, and then the test is doing
3797 * just what it's supposed to.
3798 * 2) it is a multi-byte character, in which case the final
3799 * byte is never mistakable for ASCII, and so the test
3800 * will say it is not a word character, which is the
3801 * correct answer. */
3802 ln = (locinput != PL_bostr) ?
3803 UCHARAT(locinput - 1) : '\n';
3804 switch (FLAGS(scan)) {
3805 case REGEX_UNICODE_CHARSET:
3806 ln = isWORDCHAR_L1(ln);
3807 n = isWORDCHAR_L1(nextchr);
3809 case REGEX_LOCALE_CHARSET:
3810 ln = isALNUM_LC(ln);
3811 n = isALNUM_LC(nextchr);
3813 case REGEX_DEPENDS_CHARSET:
3815 n = isALNUM(nextchr);
3817 case REGEX_ASCII_RESTRICTED_CHARSET:
3818 ln = isWORDCHAR_A(ln);
3819 n = isWORDCHAR_A(nextchr);
3822 Perl_croak(aTHX_ "panic: Unexpected FLAGS %u in op %u", FLAGS(scan), OP(scan));
3826 /* Note requires that all BOUNDs be lower than all NBOUNDs in
3828 if (((!ln) == (!n)) == (OP(scan) < NBOUND))
3833 if (utf8_target || state_num == ANYOFV) {
3834 STRLEN inclasslen = PL_regeol - locinput;
3835 if (locinput >= PL_regeol)
3838 if (!reginclass(rex, scan, (U8*)locinput, &inclasslen, utf8_target))
3840 locinput += inclasslen;
3841 nextchr = UCHARAT(locinput);
3846 nextchr = UCHARAT(locinput);
3847 if (!nextchr && locinput >= PL_regeol)
3849 if (!REGINCLASS(rex, scan, (U8*)locinput))
3851 nextchr = UCHARAT(++locinput);
3855 /* Special char classes - The defines start on line 129 or so */
3856 CCC_TRY_U(ALNUM, NALNUM, isWORDCHAR,
3857 ALNUML, NALNUML, isALNUM_LC, isALNUM_LC_utf8,
3858 ALNUMU, NALNUMU, isWORDCHAR_L1,
3859 ALNUMA, NALNUMA, isWORDCHAR_A,
3862 CCC_TRY_U(SPACE, NSPACE, isSPACE,
3863 SPACEL, NSPACEL, isSPACE_LC, isSPACE_LC_utf8,
3864 SPACEU, NSPACEU, isSPACE_L1,
3865 SPACEA, NSPACEA, isSPACE_A,
3868 CCC_TRY(DIGIT, NDIGIT, isDIGIT,
3869 DIGITL, NDIGITL, isDIGIT_LC, isDIGIT_LC_utf8,
3870 DIGITA, NDIGITA, isDIGIT_A,
3873 case CLUMP: /* Match \X: logical Unicode character. This is defined as
3874 a Unicode extended Grapheme Cluster */
3875 /* From http://www.unicode.org/reports/tr29 (5.2 version). An
3876 extended Grapheme Cluster is:
3879 | Prepend* Begin Extend*
3882 Begin is (Hangul-syllable | ! Control)
3883 Extend is (Grapheme_Extend | Spacing_Mark)
3884 Control is [ GCB_Control CR LF ]
3886 The discussion below shows how the code for CLUMP is derived
3887 from this regex. Note that most of these concepts are from
3888 property values of the Grapheme Cluster Boundary (GCB) property.
3889 No code point can have multiple property values for a given
3890 property. Thus a code point in Prepend can't be in Control, but
3891 it must be in !Control. This is why Control above includes
3892 GCB_Control plus CR plus LF. The latter two are used in the GCB
3893 property separately, and so can't be in GCB_Control, even though
3894 they logically are controls. Control is not the same as gc=cc,
3895 but includes format and other characters as well.
3897 The Unicode definition of Hangul-syllable is:
3899 | (L* ( ( V | LV ) V* | LVT ) T*)
3902 Each of these is a value for the GCB property, and hence must be
3903 disjoint, so the order they are tested is immaterial, so the
3904 above can safely be changed to
3907 | (L* ( LVT | ( V | LV ) V*) T*)
3909 The last two terms can be combined like this:
3911 | (( LVT | ( V | LV ) V*) T*))
3913 And refactored into this:
3914 L* (L | LVT T* | V V* T* | LV V* T*)
3916 That means that if we have seen any L's at all we can quit
3917 there, but if the next character is a LVT, a V or and LV we
3920 There is a subtlety with Prepend* which showed up in testing.
3921 Note that the Begin, and only the Begin is required in:
3922 | Prepend* Begin Extend*
3923 Also, Begin contains '! Control'. A Prepend must be a '!
3924 Control', which means it must be a Begin. What it comes down to
3925 is that if we match Prepend* and then find no suitable Begin
3926 afterwards, that if we backtrack the last Prepend, that one will
3927 be a suitable Begin.
3930 if (locinput >= PL_regeol)
3932 if (! utf8_target) {
3934 /* Match either CR LF or '.', as all the other possibilities
3936 locinput++; /* Match the . or CR */
3938 && locinput < PL_regeol
3939 && UCHARAT(locinput) == '\n') locinput++;
3943 /* Utf8: See if is ( CR LF ); already know that locinput <
3944 * PL_regeol, so locinput+1 is in bounds */
3945 if (nextchr == '\r' && UCHARAT(locinput + 1) == '\n') {
3949 /* In case have to backtrack to beginning, then match '.' */
3950 char *starting = locinput;
3952 /* In case have to backtrack the last prepend */
3953 char *previous_prepend = 0;
3955 LOAD_UTF8_CHARCLASS_GCB();
3957 /* Match (prepend)* */
3958 while (locinput < PL_regeol
3959 && swash_fetch(PL_utf8_X_prepend,
3960 (U8*)locinput, utf8_target))
3962 previous_prepend = locinput;
3963 locinput += UTF8SKIP(locinput);
3966 /* As noted above, if we matched a prepend character, but
3967 * the next thing won't match, back off the last prepend we
3968 * matched, as it is guaranteed to match the begin */
3969 if (previous_prepend
3970 && (locinput >= PL_regeol
3971 || ! swash_fetch(PL_utf8_X_begin,
3972 (U8*)locinput, utf8_target)))
3974 locinput = previous_prepend;
3977 /* Note that here we know PL_regeol > locinput, as we
3978 * tested that upon input to this switch case, and if we
3979 * moved locinput forward, we tested the result just above
3980 * and it either passed, or we backed off so that it will
3982 if (! swash_fetch(PL_utf8_X_begin, (U8*)locinput, utf8_target)) {
3984 /* Here did not match the required 'Begin' in the
3985 * second term. So just match the very first
3986 * character, the '.' of the final term of the regex */
3987 locinput = starting + UTF8SKIP(starting);
3990 /* Here is the beginning of a character that can have
3991 * an extender. It is either a hangul syllable, or a
3993 if (swash_fetch(PL_utf8_X_non_hangul,
3994 (U8*)locinput, utf8_target))
3997 /* Here not a Hangul syllable, must be a
3998 * ('! * Control') */
3999 locinput += UTF8SKIP(locinput);
4002 /* Here is a Hangul syllable. It can be composed
4003 * of several individual characters. One
4004 * possibility is T+ */
4005 if (swash_fetch(PL_utf8_X_T,
4006 (U8*)locinput, utf8_target))
4008 while (locinput < PL_regeol
4009 && swash_fetch(PL_utf8_X_T,
4010 (U8*)locinput, utf8_target))
4012 locinput += UTF8SKIP(locinput);
4016 /* Here, not T+, but is a Hangul. That means
4017 * it is one of the others: L, LV, LVT or V,
4019 * L* (L | LVT T* | V V* T* | LV V* T*) */
4022 while (locinput < PL_regeol
4023 && swash_fetch(PL_utf8_X_L,
4024 (U8*)locinput, utf8_target))
4026 locinput += UTF8SKIP(locinput);
4029 /* Here, have exhausted L*. If the next
4030 * character is not an LV, LVT nor V, it means
4031 * we had to have at least one L, so matches L+
4032 * in the original equation, we have a complete
4033 * hangul syllable. Are done. */
4035 if (locinput < PL_regeol
4036 && swash_fetch(PL_utf8_X_LV_LVT_V,
4037 (U8*)locinput, utf8_target))
4040 /* Otherwise keep going. Must be LV, LVT
4041 * or V. See if LVT */
4042 if (swash_fetch(PL_utf8_X_LVT,
4043 (U8*)locinput, utf8_target))
4045 locinput += UTF8SKIP(locinput);
4048 /* Must be V or LV. Take it, then
4050 locinput += UTF8SKIP(locinput);
4051 while (locinput < PL_regeol
4052 && swash_fetch(PL_utf8_X_V,
4053 (U8*)locinput, utf8_target))
4055 locinput += UTF8SKIP(locinput);
4059 /* And any of LV, LVT, or V can be followed
4061 while (locinput < PL_regeol
4062 && swash_fetch(PL_utf8_X_T,
4066 locinput += UTF8SKIP(locinput);
4072 /* Match any extender */
4073 while (locinput < PL_regeol
4074 && swash_fetch(PL_utf8_X_extend,
4075 (U8*)locinput, utf8_target))
4077 locinput += UTF8SKIP(locinput);
4081 if (locinput > PL_regeol) sayNO;
4083 nextchr = UCHARAT(locinput);
4087 { /* The capture buffer cases. The ones beginning with N for the
4088 named buffers just convert to the equivalent numbered and
4089 pretend they were called as the corresponding numbered buffer
4091 /* don't initialize these, it makes C++ unhappy */
4095 const U8 *fold_array;
4097 PL_reg_flags |= RF_tainted;
4098 folder = foldEQ_locale;
4099 fold_array = PL_fold_locale;
4104 folder = foldEQ_latin1;
4105 fold_array = PL_fold_latin1;
4111 fold_array = PL_fold;
4121 /* For the named back references, find the corresponding buffer
4123 n = reg_check_named_buff_matched(rex,scan);
4128 goto do_nref_ref_common;
4131 PL_reg_flags |= RF_tainted;
4132 folder = foldEQ_locale;
4133 fold_array = PL_fold_locale;
4137 folder = foldEQ_latin1;
4138 fold_array = PL_fold_latin1;
4143 fold_array = PL_fold;
4152 n = ARG(scan); /* which paren pair */
4155 ln = PL_regoffs[n].start;
4156 PL_reg_leftiter = PL_reg_maxiter; /* Void cache */
4157 if (*PL_reglastparen < n || ln == -1)
4158 sayNO; /* Do not match unless seen CLOSEn. */
4159 if (ln == PL_regoffs[n].end)
4163 if (type != REF /* REF can do byte comparison */
4166 && (*s == (char) LATIN_SMALL_LETTER_SHARP_S
4167 || *locinput == (char) LATIN_SMALL_LETTER_SHARP_S))))
4168 { /* XXX handle REFFL better */
4169 char * limit = PL_regeol;
4171 /* This call case insensitively compares the entire buffer
4172 * at s, with the current input starting at locinput, but
4173 * not going off the end given by PL_regeol, and returns in
4174 * limit upon success, how much of the current input was
4176 if (! foldEQ_utf8(s, NULL, PL_regoffs[n].end - ln, utf8_target,
4177 locinput, &limit, 0, utf8_target))
4182 nextchr = UCHARAT(locinput);
4186 /* Not utf8: Inline the first character, for speed. */
4187 if (UCHARAT(s) != nextchr &&
4189 UCHARAT(s) != fold_array[nextchr]))
4191 ln = PL_regoffs[n].end - ln;
4192 if (locinput + ln > PL_regeol)
4194 if (ln > 1 && (type == REF
4195 ? memNE(s, locinput, ln)
4196 : ! folder(s, locinput, ln)))
4199 nextchr = UCHARAT(locinput);
4209 #define ST st->u.eval
4214 regexp_internal *rei;
4215 regnode *startpoint;
4218 case GOSUB: /* /(...(?1))/ /(...(?&foo))/ */
4219 if (cur_eval && cur_eval->locinput==locinput) {
4220 if (cur_eval->u.eval.close_paren == (U32)ARG(scan))
4221 Perl_croak(aTHX_ "Infinite recursion in regex");
4222 if ( ++nochange_depth > max_nochange_depth )
4224 "Pattern subroutine nesting without pos change"
4225 " exceeded limit in regex");
4232 (void)ReREFCNT_inc(rex_sv);
4233 if (OP(scan)==GOSUB) {
4234 startpoint = scan + ARG2L(scan);
4235 ST.close_paren = ARG(scan);
4237 startpoint = rei->program+1;
4240 goto eval_recurse_doit;
4242 case EVAL: /* /(?{A})B/ /(??{A})B/ and /(?(?{A})X|Y)B/ */
4243 if (cur_eval && cur_eval->locinput==locinput) {
4244 if ( ++nochange_depth > max_nochange_depth )
4245 Perl_croak(aTHX_ "EVAL without pos change exceeded limit in regex");
4250 /* execute the code in the {...} */
4252 SV ** const before = SP;
4253 OP_4tree * const oop = PL_op;
4254 COP * const ocurcop = PL_curcop;
4256 char *saved_regeol = PL_regeol;
4257 struct re_save_state saved_state;
4259 /* To not corrupt the existing regex state while executing the
4260 * eval we would normally put it on the save stack, like with
4261 * save_re_context. However, re-evals have a weird scoping so we
4262 * can't just add ENTER/LEAVE here. With that, things like
4264 * (?{$a=2})(a(?{local$a=$a+1}))*aak*c(?{$b=$a})
4266 * would break, as they expect the localisation to be unwound
4267 * only when the re-engine backtracks through the bit that
4270 * What we do instead is just saving the state in a local c
4273 Copy(&PL_reg_state, &saved_state, 1, struct re_save_state);
4276 PL_op = (OP_4tree*)rexi->data->data[n];
4277 DEBUG_STATE_r( PerlIO_printf(Perl_debug_log,
4278 " re_eval 0x%"UVxf"\n", PTR2UV(PL_op)) );
4279 PAD_SAVE_LOCAL(old_comppad, (PAD*)rexi->data->data[n + 2]);
4280 PL_regoffs[0].end = PL_reg_magic->mg_len = locinput - PL_bostr;
4283 SV *sv_mrk = get_sv("REGMARK", 1);
4284 sv_setsv(sv_mrk, sv_yes_mark);
4287 CALLRUNOPS(aTHX); /* Scalar context. */
4290 ret = &PL_sv_undef; /* protect against empty (?{}) blocks. */
4296 Copy(&saved_state, &PL_reg_state, 1, struct re_save_state);
4299 PAD_RESTORE_LOCAL(old_comppad);
4300 PL_curcop = ocurcop;
4301 PL_regeol = saved_regeol;
4304 sv_setsv(save_scalar(PL_replgv), ret);
4308 if (logical == 2) { /* Postponed subexpression: /(??{...})/ */
4311 /* extract RE object from returned value; compiling if
4317 SV *const sv = SvRV(ret);
4319 if (SvTYPE(sv) == SVt_REGEXP) {
4321 } else if (SvSMAGICAL(sv)) {
4322 mg = mg_find(sv, PERL_MAGIC_qr);
4325 } else if (SvTYPE(ret) == SVt_REGEXP) {
4327 } else if (SvSMAGICAL(ret)) {
4328 if (SvGMAGICAL(ret)) {
4329 /* I don't believe that there is ever qr magic
4331 assert(!mg_find(ret, PERL_MAGIC_qr));
4332 sv_unmagic(ret, PERL_MAGIC_qr);
4335 mg = mg_find(ret, PERL_MAGIC_qr);
4336 /* testing suggests mg only ends up non-NULL for
4337 scalars who were upgraded and compiled in the
4338 else block below. In turn, this is only
4339 triggered in the "postponed utf8 string" tests
4345 rx = (REGEXP *) mg->mg_obj; /*XXX:dmq*/
4349 rx = reg_temp_copy(NULL, rx);
4353 const I32 osize = PL_regsize;
4356 assert (SvUTF8(ret));
4357 } else if (SvUTF8(ret)) {
4358 /* Not doing UTF-8, despite what the SV says. Is
4359 this only if we're trapped in use 'bytes'? */
4360 /* Make a copy of the octet sequence, but without
4361 the flag on, as the compiler now honours the
4362 SvUTF8 flag on ret. */
4364 const char *const p = SvPV(ret, len);
4365 ret = newSVpvn_flags(p, len, SVs_TEMP);
4367 rx = CALLREGCOMP(ret, pm_flags);
4369 & (SVs_TEMP | SVs_PADTMP | SVf_READONLY
4371 /* This isn't a first class regexp. Instead, it's
4372 caching a regexp onto an existing, Perl visible
4374 sv_magic(ret, MUTABLE_SV(rx), PERL_MAGIC_qr, 0, 0);
4379 re = (struct regexp *)SvANY(rx);
4381 RXp_MATCH_COPIED_off(re);
4382 re->subbeg = rex->subbeg;
4383 re->sublen = rex->sublen;
4386 debug_start_match(re_sv, utf8_target, locinput, PL_regeol,
4387 "Matching embedded");
4389 startpoint = rei->program + 1;
4390 ST.close_paren = 0; /* only used for GOSUB */
4391 /* borrowed from regtry */
4392 if (PL_reg_start_tmpl <= re->nparens) {
4393 PL_reg_start_tmpl = re->nparens*3/2 + 3;
4394 if(PL_reg_start_tmp)
4395 Renew(PL_reg_start_tmp, PL_reg_start_tmpl, char*);
4397 Newx(PL_reg_start_tmp, PL_reg_start_tmpl, char*);
4400 eval_recurse_doit: /* Share code with GOSUB below this line */
4401 /* run the pattern returned from (??{...}) */
4402 ST.cp = regcppush(0); /* Save *all* the positions. */
4403 REGCP_SET(ST.lastcp);
4405 PL_regoffs = re->offs; /* essentially NOOP on GOSUB */
4407 /* see regtry, specifically PL_reglast(?:close)?paren is a pointer! (i dont know why) :dmq */
4408 PL_reglastparen = &re->lastparen;
4409 PL_reglastcloseparen = &re->lastcloseparen;
4411 re->lastcloseparen = 0;
4413 PL_reginput = locinput;
4416 /* XXXX This is too dramatic a measure... */
4419 ST.toggle_reg_flags = PL_reg_flags;
4421 PL_reg_flags |= RF_utf8;
4423 PL_reg_flags &= ~RF_utf8;
4424 ST.toggle_reg_flags ^= PL_reg_flags; /* diff of old and new */
4426 ST.prev_rex = rex_sv;
4427 ST.prev_curlyx = cur_curlyx;
4428 SETREX(rex_sv,re_sv);
4433 ST.prev_eval = cur_eval;
4435 /* now continue from first node in postoned RE */
4436 PUSH_YES_STATE_GOTO(EVAL_AB, startpoint);
4439 /* logical is 1, /(?(?{...})X|Y)/ */
4440 sw = cBOOL(SvTRUE(ret));
4445 case EVAL_AB: /* cleanup after a successful (??{A})B */
4446 /* note: this is called twice; first after popping B, then A */
4447 PL_reg_flags ^= ST.toggle_reg_flags;
4448 ReREFCNT_dec(rex_sv);
4449 SETREX(rex_sv,ST.prev_rex);
4450 rex = (struct regexp *)SvANY(rex_sv);
4451 rexi = RXi_GET(rex);
4453 cur_eval = ST.prev_eval;
4454 cur_curlyx = ST.prev_curlyx;
4456 /* rex was changed so update the pointer in PL_reglastparen and PL_reglastcloseparen */
4457 PL_reglastparen = &rex->lastparen;
4458 PL_reglastcloseparen = &rex->lastcloseparen;
4459 /* also update PL_regoffs */
4460 PL_regoffs = rex->offs;
4462 /* XXXX This is too dramatic a measure... */
4464 if ( nochange_depth )
4469 case EVAL_AB_fail: /* unsuccessfully ran A or B in (??{A})B */
4470 /* note: this is called twice; first after popping B, then A */
4471 PL_reg_flags ^= ST.toggle_reg_flags;
4472 ReREFCNT_dec(rex_sv);
4473 SETREX(rex_sv,ST.prev_rex);
4474 rex = (struct regexp *)SvANY(rex_sv);
4475 rexi = RXi_GET(rex);
4476 /* rex was changed so update the pointer in PL_reglastparen and PL_reglastcloseparen */
4477 PL_reglastparen = &rex->lastparen;
4478 PL_reglastcloseparen = &rex->lastcloseparen;
4480 PL_reginput = locinput;
4481 REGCP_UNWIND(ST.lastcp);
4483 cur_eval = ST.prev_eval;
4484 cur_curlyx = ST.prev_curlyx;
4485 /* XXXX This is too dramatic a measure... */
4487 if ( nochange_depth )
4493 n = ARG(scan); /* which paren pair */
4494 PL_reg_start_tmp[n] = locinput;
4500 n = ARG(scan); /* which paren pair */
4501 PL_regoffs[n].start = PL_reg_start_tmp[n] - PL_bostr;
4502 PL_regoffs[n].end = locinput - PL_bostr;
4503 /*if (n > PL_regsize)
4505 if (n > *PL_reglastparen)
4506 *PL_reglastparen = n;
4507 *PL_reglastcloseparen = n;
4508 if (cur_eval && cur_eval->u.eval.close_paren == n) {
4516 cursor && OP(cursor)!=END;
4517 cursor=regnext(cursor))
4519 if ( OP(cursor)==CLOSE ){
4521 if ( n <= lastopen ) {
4523 = PL_reg_start_tmp[n] - PL_bostr;
4524 PL_regoffs[n].end = locinput - PL_bostr;
4525 /*if (n > PL_regsize)
4527 if (n > *PL_reglastparen)
4528 *PL_reglastparen = n;
4529 *PL_reglastcloseparen = n;
4530 if ( n == ARG(scan) || (cur_eval &&
4531 cur_eval->u.eval.close_paren == n))
4540 n = ARG(scan); /* which paren pair */
4541 sw = cBOOL(*PL_reglastparen >= n && PL_regoffs[n].end != -1);
4544 /* reg_check_named_buff_matched returns 0 for no match */
4545 sw = cBOOL(0 < reg_check_named_buff_matched(rex,scan));
4549 sw = (cur_eval && (!n || cur_eval->u.eval.close_paren == n));
4555 PL_reg_leftiter = PL_reg_maxiter; /* Void cache */
4557 next = NEXTOPER(NEXTOPER(scan));
4559 next = scan + ARG(scan);
4560 if (OP(next) == IFTHEN) /* Fake one. */
4561 next = NEXTOPER(NEXTOPER(next));
4565 logical = scan->flags;
4568 /*******************************************************************
4570 The CURLYX/WHILEM pair of ops handle the most generic case of the /A*B/
4571 pattern, where A and B are subpatterns. (For simple A, CURLYM or
4572 STAR/PLUS/CURLY/CURLYN are used instead.)
4574 A*B is compiled as <CURLYX><A><WHILEM><B>
4576 On entry to the subpattern, CURLYX is called. This pushes a CURLYX
4577 state, which contains the current count, initialised to -1. It also sets
4578 cur_curlyx to point to this state, with any previous value saved in the
4581 CURLYX then jumps straight to the WHILEM op, rather than executing A,
4582 since the pattern may possibly match zero times (i.e. it's a while {} loop
4583 rather than a do {} while loop).
4585 Each entry to WHILEM represents a successful match of A. The count in the
4586 CURLYX block is incremented, another WHILEM state is pushed, and execution
4587 passes to A or B depending on greediness and the current count.
4589 For example, if matching against the string a1a2a3b (where the aN are
4590 substrings that match /A/), then the match progresses as follows: (the
4591 pushed states are interspersed with the bits of strings matched so far):
4594 <CURLYX cnt=0><WHILEM>
4595 <CURLYX cnt=1><WHILEM> a1 <WHILEM>
4596 <CURLYX cnt=2><WHILEM> a1 <WHILEM> a2 <WHILEM>
4597 <CURLYX cnt=3><WHILEM> a1 <WHILEM> a2 <WHILEM> a3 <WHILEM>
4598 <CURLYX cnt=3><WHILEM> a1 <WHILEM> a2 <WHILEM> a3 <WHILEM> b
4600 (Contrast this with something like CURLYM, which maintains only a single
4604 a1 <CURLYM cnt=1> a2
4605 a1 a2 <CURLYM cnt=2> a3
4606 a1 a2 a3 <CURLYM cnt=3> b
4609 Each WHILEM state block marks a point to backtrack to upon partial failure
4610 of A or B, and also contains some minor state data related to that
4611 iteration. The CURLYX block, pointed to by cur_curlyx, contains the
4612 overall state, such as the count, and pointers to the A and B ops.
4614 This is complicated slightly by nested CURLYX/WHILEM's. Since cur_curlyx
4615 must always point to the *current* CURLYX block, the rules are:
4617 When executing CURLYX, save the old cur_curlyx in the CURLYX state block,
4618 and set cur_curlyx to point the new block.
4620 When popping the CURLYX block after a successful or unsuccessful match,
4621 restore the previous cur_curlyx.
4623 When WHILEM is about to execute B, save the current cur_curlyx, and set it
4624 to the outer one saved in the CURLYX block.
4626 When popping the WHILEM block after a successful or unsuccessful B match,
4627 restore the previous cur_curlyx.
4629 Here's an example for the pattern (AI* BI)*BO
4630 I and O refer to inner and outer, C and W refer to CURLYX and WHILEM:
4633 curlyx backtrack stack
4634 ------ ---------------
4636 CO <CO prev=NULL> <WO>
4637 CI <CO prev=NULL> <WO> <CI prev=CO> <WI> ai
4638 CO <CO prev=NULL> <WO> <CI prev=CO> <WI> ai <WI prev=CI> bi
4639 NULL <CO prev=NULL> <WO> <CI prev=CO> <WI> ai <WI prev=CI> bi <WO prev=CO> bo
4641 At this point the pattern succeeds, and we work back down the stack to
4642 clean up, restoring as we go:
4644 CO <CO prev=NULL> <WO> <CI prev=CO> <WI> ai <WI prev=CI> bi
4645 CI <CO prev=NULL> <WO> <CI prev=CO> <WI> ai
4646 CO <CO prev=NULL> <WO>
4649 *******************************************************************/
4651 #define ST st->u.curlyx
4653 case CURLYX: /* start of /A*B/ (for complex A) */
4655 /* No need to save/restore up to this paren */
4656 I32 parenfloor = scan->flags;
4658 assert(next); /* keep Coverity happy */
4659 if (OP(PREVOPER(next)) == NOTHING) /* LONGJMP */
4662 /* XXXX Probably it is better to teach regpush to support
4663 parenfloor > PL_regsize... */
4664 if (parenfloor > (I32)*PL_reglastparen)
4665 parenfloor = *PL_reglastparen; /* Pessimization... */
4667 ST.prev_curlyx= cur_curlyx;
4669 ST.cp = PL_savestack_ix;
4671 /* these fields contain the state of the current curly.
4672 * they are accessed by subsequent WHILEMs */
4673 ST.parenfloor = parenfloor;
4678 ST.count = -1; /* this will be updated by WHILEM */
4679 ST.lastloc = NULL; /* this will be updated by WHILEM */
4681 PL_reginput = locinput;
4682 PUSH_YES_STATE_GOTO(CURLYX_end, PREVOPER(next));
4686 case CURLYX_end: /* just finished matching all of A*B */
4687 cur_curlyx = ST.prev_curlyx;
4691 case CURLYX_end_fail: /* just failed to match all of A*B */
4693 cur_curlyx = ST.prev_curlyx;
4699 #define ST st->u.whilem
4701 case WHILEM: /* just matched an A in /A*B/ (for complex A) */
4703 /* see the discussion above about CURLYX/WHILEM */
4705 int min = ARG1(cur_curlyx->u.curlyx.me);
4706 int max = ARG2(cur_curlyx->u.curlyx.me);
4707 regnode *A = NEXTOPER(cur_curlyx->u.curlyx.me) + EXTRA_STEP_2ARGS;
4709 assert(cur_curlyx); /* keep Coverity happy */
4710 n = ++cur_curlyx->u.curlyx.count; /* how many A's matched */
4711 ST.save_lastloc = cur_curlyx->u.curlyx.lastloc;
4712 ST.cache_offset = 0;
4715 PL_reginput = locinput;
4717 DEBUG_EXECUTE_r( PerlIO_printf(Perl_debug_log,
4718 "%*s whilem: matched %ld out of %d..%d\n",
4719 REPORT_CODE_OFF+depth*2, "", (long)n, min, max)
4722 /* First just match a string of min A's. */
4725 cur_curlyx->u.curlyx.lastloc = locinput;
4726 PUSH_STATE_GOTO(WHILEM_A_pre, A);
4730 /* If degenerate A matches "", assume A done. */
4732 if (locinput == cur_curlyx->u.curlyx.lastloc) {
4733 DEBUG_EXECUTE_r( PerlIO_printf(Perl_debug_log,
4734 "%*s whilem: empty match detected, trying continuation...\n",
4735 REPORT_CODE_OFF+depth*2, "")
4737 goto do_whilem_B_max;
4740 /* super-linear cache processing */
4744 if (!PL_reg_maxiter) {
4745 /* start the countdown: Postpone detection until we
4746 * know the match is not *that* much linear. */
4747 PL_reg_maxiter = (PL_regeol - PL_bostr + 1) * (scan->flags>>4);
4748 /* possible overflow for long strings and many CURLYX's */
4749 if (PL_reg_maxiter < 0)
4750 PL_reg_maxiter = I32_MAX;
4751 PL_reg_leftiter = PL_reg_maxiter;
4754 if (PL_reg_leftiter-- == 0) {
4755 /* initialise cache */
4756 const I32 size = (PL_reg_maxiter + 7)/8;
4757 if (PL_reg_poscache) {
4758 if ((I32)PL_reg_poscache_size < size) {
4759 Renew(PL_reg_poscache, size, char);
4760 PL_reg_poscache_size = size;
4762 Zero(PL_reg_poscache, size, char);
4765 PL_reg_poscache_size = size;
4766 Newxz(PL_reg_poscache, size, char);
4768 DEBUG_EXECUTE_r( PerlIO_printf(Perl_debug_log,
4769 "%swhilem: Detected a super-linear match, switching on caching%s...\n",
4770 PL_colors[4], PL_colors[5])
4774 if (PL_reg_leftiter < 0) {
4775 /* have we already failed at this position? */
4777 offset = (scan->flags & 0xf) - 1
4778 + (locinput - PL_bostr) * (scan->flags>>4);
4779 mask = 1 << (offset % 8);
4781 if (PL_reg_poscache[offset] & mask) {
4782 DEBUG_EXECUTE_r( PerlIO_printf(Perl_debug_log,
4783 "%*s whilem: (cache) already tried at this position...\n",
4784 REPORT_CODE_OFF+depth*2, "")
4786 sayNO; /* cache records failure */
4788 ST.cache_offset = offset;
4789 ST.cache_mask = mask;
4793 /* Prefer B over A for minimal matching. */
4795 if (cur_curlyx->u.curlyx.minmod) {
4796 ST.save_curlyx = cur_curlyx;
4797 cur_curlyx = cur_curlyx->u.curlyx.prev_curlyx;
4798 ST.cp = regcppush(ST.save_curlyx->u.curlyx.parenfloor);
4799 REGCP_SET(ST.lastcp);
4800 PUSH_YES_STATE_GOTO(WHILEM_B_min, ST.save_curlyx->u.curlyx.B);
4804 /* Prefer A over B for maximal matching. */
4806 if (n < max) { /* More greed allowed? */
4807 ST.cp = regcppush(cur_curlyx->u.curlyx.parenfloor);
4808 cur_curlyx->u.curlyx.lastloc = locinput;
4809 REGCP_SET(ST.lastcp);
4810 PUSH_STATE_GOTO(WHILEM_A_max, A);
4813 goto do_whilem_B_max;
4817 case WHILEM_B_min: /* just matched B in a minimal match */
4818 case WHILEM_B_max: /* just matched B in a maximal match */
4819 cur_curlyx = ST.save_curlyx;
4823 case WHILEM_B_max_fail: /* just failed to match B in a maximal match */
4824 cur_curlyx = ST.save_curlyx;
4825 cur_curlyx->u.curlyx.lastloc = ST.save_lastloc;
4826 cur_curlyx->u.curlyx.count--;
4830 case WHILEM_A_min_fail: /* just failed to match A in a minimal match */
4831 REGCP_UNWIND(ST.lastcp);
4834 case WHILEM_A_pre_fail: /* just failed to match even minimal A */
4835 cur_curlyx->u.curlyx.lastloc = ST.save_lastloc;
4836 cur_curlyx->u.curlyx.count--;
4840 case WHILEM_A_max_fail: /* just failed to match A in a maximal match */
4841 REGCP_UNWIND(ST.lastcp);
4842 regcppop(rex); /* Restore some previous $<digit>s? */
4843 PL_reginput = locinput;
4844 DEBUG_EXECUTE_r(PerlIO_printf(Perl_debug_log,
4845 "%*s whilem: failed, trying continuation...\n",
4846 REPORT_CODE_OFF+depth*2, "")
4849 if (cur_curlyx->u.curlyx.count >= REG_INFTY
4850 && ckWARN(WARN_REGEXP)
4851 && !(PL_reg_flags & RF_warned))
4853 PL_reg_flags |= RF_warned;
4854 Perl_warner(aTHX_ packWARN(WARN_REGEXP), "%s limit (%d) exceeded",
4855 "Complex regular subexpression recursion",
4860 ST.save_curlyx = cur_curlyx;
4861 cur_curlyx = cur_curlyx->u.curlyx.prev_curlyx;
4862 PUSH_YES_STATE_GOTO(WHILEM_B_max, ST.save_curlyx->u.curlyx.B);
4865 case WHILEM_B_min_fail: /* just failed to match B in a minimal match */
4866 cur_curlyx = ST.save_curlyx;
4867 REGCP_UNWIND(ST.lastcp);
4870 if (cur_curlyx->u.curlyx.count >= /*max*/ARG2(cur_curlyx->u.curlyx.me)) {
4871 /* Maximum greed exceeded */
4872 if (cur_curlyx->u.curlyx.count >= REG_INFTY
4873 && ckWARN(WARN_REGEXP)
4874 && !(PL_reg_flags & RF_warned))
4876 PL_reg_flags |= RF_warned;
4877 Perl_warner(aTHX_ packWARN(WARN_REGEXP),
4878 "%s limit (%d) exceeded",
4879 "Complex regular subexpression recursion",
4882 cur_curlyx->u.curlyx.count--;
4886 DEBUG_EXECUTE_r(PerlIO_printf(Perl_debug_log,
4887 "%*s trying longer...\n", REPORT_CODE_OFF+depth*2, "")
4889 /* Try grabbing another A and see if it helps. */
4890 PL_reginput = locinput;
4891 cur_curlyx->u.curlyx.lastloc = locinput;
4892 ST.cp = regcppush(cur_curlyx->u.curlyx.parenfloor);
4893 REGCP_SET(ST.lastcp);
4894 PUSH_STATE_GOTO(WHILEM_A_min,
4895 /*A*/ NEXTOPER(ST.save_curlyx->u.curlyx.me) + EXTRA_STEP_2ARGS);
4899 #define ST st->u.branch
4901 case BRANCHJ: /* /(...|A|...)/ with long next pointer */
4902 next = scan + ARG(scan);
4905 scan = NEXTOPER(scan);
4908 case BRANCH: /* /(...|A|...)/ */
4909 scan = NEXTOPER(scan); /* scan now points to inner node */
4910 ST.lastparen = *PL_reglastparen;
4911 ST.next_branch = next;
4913 PL_reginput = locinput;
4915 /* Now go into the branch */
4917 PUSH_YES_STATE_GOTO(BRANCH_next, scan);
4919 PUSH_STATE_GOTO(BRANCH_next, scan);
4923 PL_reginput = locinput;
4924 sv_yes_mark = st->u.mark.mark_name = scan->flags ? NULL :
4925 MUTABLE_SV(rexi->data->data[ ARG( scan ) ]);
4926 PUSH_STATE_GOTO(CUTGROUP_next,next);
4928 case CUTGROUP_next_fail:
4931 if (st->u.mark.mark_name)
4932 sv_commit = st->u.mark.mark_name;
4938 case BRANCH_next_fail: /* that branch failed; try the next, if any */
4943 REGCP_UNWIND(ST.cp);
4944 for (n = *PL_reglastparen; n > ST.lastparen; n--)
4945 PL_regoffs[n].end = -1;
4946 *PL_reglastparen = n;
4947 /*dmq: *PL_reglastcloseparen = n; */
4948 scan = ST.next_branch;
4949 /* no more branches? */
4950 if (!scan || (OP(scan) != BRANCH && OP(scan) != BRANCHJ)) {
4952 PerlIO_printf( Perl_debug_log,
4953 "%*s %sBRANCH failed...%s\n",
4954 REPORT_CODE_OFF+depth*2, "",
4960 continue; /* execute next BRANCH[J] op */
4968 #define ST st->u.curlym
4970 case CURLYM: /* /A{m,n}B/ where A is fixed-length */
4972 /* This is an optimisation of CURLYX that enables us to push
4973 * only a single backtracking state, no matter how many matches
4974 * there are in {m,n}. It relies on the pattern being constant
4975 * length, with no parens to influence future backrefs
4979 scan = NEXTOPER(scan) + NODE_STEP_REGNODE;
4981 /* if paren positive, emulate an OPEN/CLOSE around A */
4983 U32 paren = ST.me->flags;
4984 if (paren > PL_regsize)
4986 if (paren > *PL_reglastparen)
4987 *PL_reglastparen = paren;
4988 scan += NEXT_OFF(scan); /* Skip former OPEN. */
4996 ST.c1 = CHRTEST_UNINIT;
4999 if (!(ST.minmod ? ARG1(ST.me) : ARG2(ST.me))) /* min/max */
5002 curlym_do_A: /* execute the A in /A{m,n}B/ */
5003 PL_reginput = locinput;
5004 PUSH_YES_STATE_GOTO(CURLYM_A, ST.A); /* match A */
5007 case CURLYM_A: /* we've just matched an A */
5008 locinput = st->locinput;
5009 nextchr = UCHARAT(locinput);
5012 /* after first match, determine A's length: u.curlym.alen */
5013 if (ST.count == 1) {
5014 if (PL_reg_match_utf8) {
5016 while (s < PL_reginput) {
5022 ST.alen = PL_reginput - locinput;
5025 ST.count = ST.minmod ? ARG1(ST.me) : ARG2(ST.me);
5028 PerlIO_printf(Perl_debug_log,
5029 "%*s CURLYM now matched %"IVdf" times, len=%"IVdf"...\n",
5030 (int)(REPORT_CODE_OFF+(depth*2)), "",
5031 (IV) ST.count, (IV)ST.alen)
5034 locinput = PL_reginput;
5036 if (cur_eval && cur_eval->u.eval.close_paren &&
5037 cur_eval->u.eval.close_paren == (U32)ST.me->flags)
5041 I32 max = (ST.minmod ? ARG1(ST.me) : ARG2(ST.me));
5042 if ( max == REG_INFTY || ST.count < max )
5043 goto curlym_do_A; /* try to match another A */
5045 goto curlym_do_B; /* try to match B */
5047 case CURLYM_A_fail: /* just failed to match an A */
5048 REGCP_UNWIND(ST.cp);
5050 if (ST.minmod || ST.count < ARG1(ST.me) /* min*/
5051 || (cur_eval && cur_eval->u.eval.close_paren &&
5052 cur_eval->u.eval.close_paren == (U32)ST.me->flags))
5055 curlym_do_B: /* execute the B in /A{m,n}B/ */
5056 PL_reginput = locinput;
5057 if (ST.c1 == CHRTEST_UNINIT) {
5058 /* calculate c1 and c2 for possible match of 1st char
5059 * following curly */
5060 ST.c1 = ST.c2 = CHRTEST_VOID;
5061 if (HAS_TEXT(ST.B) || JUMPABLE(ST.B)) {
5062 regnode *text_node = ST.B;
5063 if (! HAS_TEXT(text_node))
5064 FIND_NEXT_IMPT(text_node);
5067 (HAS_TEXT(text_node) && PL_regkind[OP(text_node)] == EXACT)
5069 But the former is redundant in light of the latter.
5071 if this changes back then the macro for
5072 IS_TEXT and friends need to change.
5074 if (PL_regkind[OP(text_node)] == EXACT)
5077 ST.c1 = (U8)*STRING(text_node);
5078 switch (OP(text_node)) {
5079 case EXACTF: ST.c2 = PL_fold[ST.c1]; break;
5080 case EXACTFU: ST.c2 = PL_fold_latin1[ST.c1]; break;
5081 case EXACTFL: ST.c2 = PL_fold_locale[ST.c1]; break;
5082 default: ST.c2 = ST.c1;
5089 PerlIO_printf(Perl_debug_log,
5090 "%*s CURLYM trying tail with matches=%"IVdf"...\n",
5091 (int)(REPORT_CODE_OFF+(depth*2)),
5094 if (ST.c1 != CHRTEST_VOID
5095 && UCHARAT(PL_reginput) != ST.c1
5096 && UCHARAT(PL_reginput) != ST.c2)
5098 /* simulate B failing */
5100 PerlIO_printf(Perl_debug_log,
5101 "%*s CURLYM Fast bail c1=%"IVdf" c2=%"IVdf"\n",
5102 (int)(REPORT_CODE_OFF+(depth*2)),"",
5105 state_num = CURLYM_B_fail;
5106 goto reenter_switch;
5110 /* mark current A as captured */
5111 I32 paren = ST.me->flags;
5113 PL_regoffs[paren].start
5114 = HOPc(PL_reginput, -ST.alen) - PL_bostr;
5115 PL_regoffs[paren].end = PL_reginput - PL_bostr;
5116 /*dmq: *PL_reglastcloseparen = paren; */
5119 PL_regoffs[paren].end = -1;
5120 if (cur_eval && cur_eval->u.eval.close_paren &&
5121 cur_eval->u.eval.close_paren == (U32)ST.me->flags)
5130 PUSH_STATE_GOTO(CURLYM_B, ST.B); /* match B */
5133 case CURLYM_B_fail: /* just failed to match a B */
5134 REGCP_UNWIND(ST.cp);
5136 I32 max = ARG2(ST.me);
5137 if (max != REG_INFTY && ST.count == max)
5139 goto curlym_do_A; /* try to match a further A */
5141 /* backtrack one A */
5142 if (ST.count == ARG1(ST.me) /* min */)
5145 locinput = HOPc(locinput, -ST.alen);
5146 goto curlym_do_B; /* try to match B */
5149 #define ST st->u.curly
5151 #define CURLY_SETPAREN(paren, success) \
5154 PL_regoffs[paren].start = HOPc(locinput, -1) - PL_bostr; \
5155 PL_regoffs[paren].end = locinput - PL_bostr; \
5156 *PL_reglastcloseparen = paren; \
5159 PL_regoffs[paren].end = -1; \
5162 case STAR: /* /A*B/ where A is width 1 */
5166 scan = NEXTOPER(scan);
5168 case PLUS: /* /A+B/ where A is width 1 */
5172 scan = NEXTOPER(scan);
5174 case CURLYN: /* /(A){m,n}B/ where A is width 1 */
5175 ST.paren = scan->flags; /* Which paren to set */
5176 if (ST.paren > PL_regsize)
5177 PL_regsize = ST.paren;
5178 if (ST.paren > *PL_reglastparen)
5179 *PL_reglastparen = ST.paren;
5180 ST.min = ARG1(scan); /* min to match */
5181 ST.max = ARG2(scan); /* max to match */
5182 if (cur_eval && cur_eval->u.eval.close_paren &&
5183 cur_eval->u.eval.close_paren == (U32)ST.paren) {
5187 scan = regnext(NEXTOPER(scan) + NODE_STEP_REGNODE);
5189 case CURLY: /* /A{m,n}B/ where A is width 1 */
5191 ST.min = ARG1(scan); /* min to match */
5192 ST.max = ARG2(scan); /* max to match */
5193 scan = NEXTOPER(scan) + NODE_STEP_REGNODE;
5196 * Lookahead to avoid useless match attempts
5197 * when we know what character comes next.
5199 * Used to only do .*x and .*?x, but now it allows
5200 * for )'s, ('s and (?{ ... })'s to be in the way
5201 * of the quantifier and the EXACT-like node. -- japhy
5204 if (ST.min > ST.max) /* XXX make this a compile-time check? */
5206 if (HAS_TEXT(next) || JUMPABLE(next)) {
5208 regnode *text_node = next;
5210 if (! HAS_TEXT(text_node))
5211 FIND_NEXT_IMPT(text_node);
5213 if (! HAS_TEXT(text_node))
5214 ST.c1 = ST.c2 = CHRTEST_VOID;
5216 if ( PL_regkind[OP(text_node)] != EXACT ) {
5217 ST.c1 = ST.c2 = CHRTEST_VOID;
5218 goto assume_ok_easy;
5221 s = (U8*)STRING(text_node);
5223 /* Currently we only get here when
5225 PL_rekind[OP(text_node)] == EXACT
5227 if this changes back then the macro for IS_TEXT and
5228 friends need to change. */
5231 switch (OP(text_node)) {
5232 case EXACTF: ST.c2 = PL_fold[ST.c1]; break;
5233 case EXACTFU: ST.c2 = PL_fold_latin1[ST.c1]; break;
5234 case EXACTFL: ST.c2 = PL_fold_locale[ST.c1]; break;
5235 default: ST.c2 = ST.c1; break;
5238 else { /* UTF_PATTERN */
5239 if (IS_TEXTFU(text_node) || IS_TEXTF(text_node)) {
5240 STRLEN ulen1, ulen2;
5241 U8 tmpbuf1[UTF8_MAXBYTES_CASE+1];
5242 U8 tmpbuf2[UTF8_MAXBYTES_CASE+1];
5244 to_utf8_lower((U8*)s, tmpbuf1, &ulen1);
5245 to_utf8_upper((U8*)s, tmpbuf2, &ulen2);
5247 ST.c1 = utf8n_to_uvchr(tmpbuf1, UTF8_MAXLEN, 0,
5249 0 : UTF8_ALLOW_ANY);
5250 ST.c2 = utf8n_to_uvchr(tmpbuf2, UTF8_MAXLEN, 0,
5252 0 : UTF8_ALLOW_ANY);
5254 ST.c1 = utf8n_to_uvuni(tmpbuf1, UTF8_MAXBYTES, 0,
5256 ST.c2 = utf8n_to_uvuni(tmpbuf2, UTF8_MAXBYTES, 0,
5261 ST.c2 = ST.c1 = utf8n_to_uvchr(s, UTF8_MAXBYTES, 0,
5268 ST.c1 = ST.c2 = CHRTEST_VOID;
5273 PL_reginput = locinput;
5276 if (ST.min && regrepeat(rex, ST.A, ST.min, depth) < ST.min)
5279 locinput = PL_reginput;
5281 if (ST.c1 == CHRTEST_VOID)
5282 goto curly_try_B_min;
5284 ST.oldloc = locinput;
5286 /* set ST.maxpos to the furthest point along the
5287 * string that could possibly match */
5288 if (ST.max == REG_INFTY) {
5289 ST.maxpos = PL_regeol - 1;
5291 while (UTF8_IS_CONTINUATION(*(U8*)ST.maxpos))
5294 else if (utf8_target) {
5295 int m = ST.max - ST.min;
5296 for (ST.maxpos = locinput;
5297 m >0 && ST.maxpos + UTF8SKIP(ST.maxpos) <= PL_regeol; m--)
5298 ST.maxpos += UTF8SKIP(ST.maxpos);
5301 ST.maxpos = locinput + ST.max - ST.min;
5302 if (ST.maxpos >= PL_regeol)
5303 ST.maxpos = PL_regeol - 1;
5305 goto curly_try_B_min_known;
5309 ST.count = regrepeat(rex, ST.A, ST.max, depth);
5310 locinput = PL_reginput;
5311 if (ST.count < ST.min)
5313 if ((ST.count > ST.min)
5314 && (PL_regkind[OP(ST.B)] == EOL) && (OP(ST.B) != MEOL))
5316 /* A{m,n} must come at the end of the string, there's
5317 * no point in backing off ... */
5319 /* ...except that $ and \Z can match before *and* after
5320 newline at the end. Consider "\n\n" =~ /\n+\Z\n/.
5321 We may back off by one in this case. */
5322 if (UCHARAT(PL_reginput - 1) == '\n' && OP(ST.B) != EOS)
5326 goto curly_try_B_max;
5331 case CURLY_B_min_known_fail:
5332 /* failed to find B in a non-greedy match where c1,c2 valid */
5333 if (ST.paren && ST.count)
5334 PL_regoffs[ST.paren].end = -1;
5336 PL_reginput = locinput; /* Could be reset... */
5337 REGCP_UNWIND(ST.cp);
5338 /* Couldn't or didn't -- move forward. */
5339 ST.oldloc = locinput;
5341 locinput += UTF8SKIP(locinput);
5345 curly_try_B_min_known:
5346 /* find the next place where 'B' could work, then call B */
5350 n = (ST.oldloc == locinput) ? 0 : 1;
5351 if (ST.c1 == ST.c2) {
5353 /* set n to utf8_distance(oldloc, locinput) */
5354 while (locinput <= ST.maxpos &&
5355 utf8n_to_uvchr((U8*)locinput,
5356 UTF8_MAXBYTES, &len,
5357 uniflags) != (UV)ST.c1) {
5363 /* set n to utf8_distance(oldloc, locinput) */
5364 while (locinput <= ST.maxpos) {
5366 const UV c = utf8n_to_uvchr((U8*)locinput,
5367 UTF8_MAXBYTES, &len,
5369 if (c == (UV)ST.c1 || c == (UV)ST.c2)
5377 if (ST.c1 == ST.c2) {
5378 while (locinput <= ST.maxpos &&
5379 UCHARAT(locinput) != ST.c1)
5383 while (locinput <= ST.maxpos
5384 && UCHARAT(locinput) != ST.c1
5385 && UCHARAT(locinput) != ST.c2)
5388 n = locinput - ST.oldloc;
5390 if (locinput > ST.maxpos)
5392 /* PL_reginput == oldloc now */
5395 if (regrepeat(rex, ST.A, n, depth) < n)
5398 PL_reginput = locinput;
5399 CURLY_SETPAREN(ST.paren, ST.count);
5400 if (cur_eval && cur_eval->u.eval.close_paren &&
5401 cur_eval->u.eval.close_paren == (U32)ST.paren) {
5404 PUSH_STATE_GOTO(CURLY_B_min_known, ST.B);
5409 case CURLY_B_min_fail:
5410 /* failed to find B in a non-greedy match where c1,c2 invalid */
5411 if (ST.paren && ST.count)
5412 PL_regoffs[ST.paren].end = -1;
5414 REGCP_UNWIND(ST.cp);
5415 /* failed -- move forward one */
5416 PL_reginput = locinput;
5417 if (regrepeat(rex, ST.A, 1, depth)) {
5419 locinput = PL_reginput;
5420 if (ST.count <= ST.max || (ST.max == REG_INFTY &&
5421 ST.count > 0)) /* count overflow ? */
5424 CURLY_SETPAREN(ST.paren, ST.count);
5425 if (cur_eval && cur_eval->u.eval.close_paren &&
5426 cur_eval->u.eval.close_paren == (U32)ST.paren) {
5429 PUSH_STATE_GOTO(CURLY_B_min, ST.B);
5437 /* a successful greedy match: now try to match B */
5438 if (cur_eval && cur_eval->u.eval.close_paren &&
5439 cur_eval->u.eval.close_paren == (U32)ST.paren) {
5444 if (ST.c1 != CHRTEST_VOID)
5445 c = utf8_target ? utf8n_to_uvchr((U8*)PL_reginput,
5446 UTF8_MAXBYTES, 0, uniflags)
5447 : (UV) UCHARAT(PL_reginput);
5448 /* If it could work, try it. */
5449 if (ST.c1 == CHRTEST_VOID || c == (UV)ST.c1 || c == (UV)ST.c2) {
5450 CURLY_SETPAREN(ST.paren, ST.count);
5451 PUSH_STATE_GOTO(CURLY_B_max, ST.B);
5456 case CURLY_B_max_fail:
5457 /* failed to find B in a greedy match */
5458 if (ST.paren && ST.count)
5459 PL_regoffs[ST.paren].end = -1;
5461 REGCP_UNWIND(ST.cp);
5463 if (--ST.count < ST.min)
5465 PL_reginput = locinput = HOPc(locinput, -1);
5466 goto curly_try_B_max;
5473 /* we've just finished A in /(??{A})B/; now continue with B */
5475 st->u.eval.toggle_reg_flags
5476 = cur_eval->u.eval.toggle_reg_flags;
5477 PL_reg_flags ^= st->u.eval.toggle_reg_flags;
5479 st->u.eval.prev_rex = rex_sv; /* inner */
5480 SETREX(rex_sv,cur_eval->u.eval.prev_rex);
5481 rex = (struct regexp *)SvANY(rex_sv);
5482 rexi = RXi_GET(rex);
5483 cur_curlyx = cur_eval->u.eval.prev_curlyx;
5484 ReREFCNT_inc(rex_sv);
5485 st->u.eval.cp = regcppush(0); /* Save *all* the positions. */
5487 /* rex was changed so update the pointer in PL_reglastparen and PL_reglastcloseparen */
5488 PL_reglastparen = &rex->lastparen;
5489 PL_reglastcloseparen = &rex->lastcloseparen;
5491 REGCP_SET(st->u.eval.lastcp);
5492 PL_reginput = locinput;
5494 /* Restore parens of the outer rex without popping the
5496 tmpix = PL_savestack_ix;
5497 PL_savestack_ix = cur_eval->u.eval.lastcp;
5499 PL_savestack_ix = tmpix;
5501 st->u.eval.prev_eval = cur_eval;
5502 cur_eval = cur_eval->u.eval.prev_eval;
5504 PerlIO_printf(Perl_debug_log, "%*s EVAL trying tail ... %"UVxf"\n",
5505 REPORT_CODE_OFF+depth*2, "",PTR2UV(cur_eval)););
5506 if ( nochange_depth )
5509 PUSH_YES_STATE_GOTO(EVAL_AB,
5510 st->u.eval.prev_eval->u.eval.B); /* match B */
5513 if (locinput < reginfo->till) {
5514 DEBUG_EXECUTE_r(PerlIO_printf(Perl_debug_log,
5515 "%sMatch possible, but length=%ld is smaller than requested=%ld, failing!%s\n",
5517 (long)(locinput - PL_reg_starttry),
5518 (long)(reginfo->till - PL_reg_starttry),
5521 sayNO_SILENT; /* Cannot match: too short. */
5523 PL_reginput = locinput; /* put where regtry can find it */
5524 sayYES; /* Success! */
5526 case SUCCEED: /* successful SUSPEND/UNLESSM/IFMATCH/CURLYM */
5528 PerlIO_printf(Perl_debug_log,
5529 "%*s %ssubpattern success...%s\n",
5530 REPORT_CODE_OFF+depth*2, "", PL_colors[4], PL_colors[5]));
5531 PL_reginput = locinput; /* put where regtry can find it */
5532 sayYES; /* Success! */
5535 #define ST st->u.ifmatch
5537 case SUSPEND: /* (?>A) */
5539 PL_reginput = locinput;
5542 case UNLESSM: /* -ve lookaround: (?!A), or with flags, (?<!A) */
5544 goto ifmatch_trivial_fail_test;
5546 case IFMATCH: /* +ve lookaround: (?=A), or with flags, (?<=A) */
5548 ifmatch_trivial_fail_test:
5550 char * const s = HOPBACKc(locinput, scan->flags);
5555 sw = 1 - cBOOL(ST.wanted);
5559 next = scan + ARG(scan);
5567 PL_reginput = locinput;
5571 ST.logical = logical;
5572 logical = 0; /* XXX: reset state of logical once it has been saved into ST */
5574 /* execute body of (?...A) */
5575 PUSH_YES_STATE_GOTO(IFMATCH_A, NEXTOPER(NEXTOPER(scan)));
5578 case IFMATCH_A_fail: /* body of (?...A) failed */
5579 ST.wanted = !ST.wanted;
5582 case IFMATCH_A: /* body of (?...A) succeeded */
5584 sw = cBOOL(ST.wanted);
5586 else if (!ST.wanted)
5589 if (OP(ST.me) == SUSPEND)
5590 locinput = PL_reginput;
5592 locinput = PL_reginput = st->locinput;
5593 nextchr = UCHARAT(locinput);
5595 scan = ST.me + ARG(ST.me);
5598 continue; /* execute B */
5603 next = scan + ARG(scan);
5608 reginfo->cutpoint = PL_regeol;
5611 PL_reginput = locinput;
5613 sv_yes_mark = sv_commit = MUTABLE_SV(rexi->data->data[ ARG( scan ) ]);
5614 PUSH_STATE_GOTO(COMMIT_next,next);
5616 case COMMIT_next_fail:
5623 #define ST st->u.mark
5625 ST.prev_mark = mark_state;
5626 ST.mark_name = sv_commit = sv_yes_mark
5627 = MUTABLE_SV(rexi->data->data[ ARG( scan ) ]);
5629 ST.mark_loc = PL_reginput = locinput;
5630 PUSH_YES_STATE_GOTO(MARKPOINT_next,next);
5632 case MARKPOINT_next:
5633 mark_state = ST.prev_mark;
5636 case MARKPOINT_next_fail:
5637 if (popmark && sv_eq(ST.mark_name,popmark))
5639 if (ST.mark_loc > startpoint)
5640 reginfo->cutpoint = HOPBACKc(ST.mark_loc, 1);
5641 popmark = NULL; /* we found our mark */
5642 sv_commit = ST.mark_name;
5645 PerlIO_printf(Perl_debug_log,
5646 "%*s %ssetting cutpoint to mark:%"SVf"...%s\n",
5647 REPORT_CODE_OFF+depth*2, "",
5648 PL_colors[4], SVfARG(sv_commit), PL_colors[5]);
5651 mark_state = ST.prev_mark;
5652 sv_yes_mark = mark_state ?
5653 mark_state->u.mark.mark_name : NULL;
5657 PL_reginput = locinput;
5659 /* (*SKIP) : if we fail we cut here*/
5660 ST.mark_name = NULL;
5661 ST.mark_loc = locinput;
5662 PUSH_STATE_GOTO(SKIP_next,next);
5664 /* (*SKIP:NAME) : if there is a (*MARK:NAME) fail where it was,
5665 otherwise do nothing. Meaning we need to scan
5667 regmatch_state *cur = mark_state;
5668 SV *find = MUTABLE_SV(rexi->data->data[ ARG( scan ) ]);
5671 if ( sv_eq( cur->u.mark.mark_name,
5674 ST.mark_name = find;
5675 PUSH_STATE_GOTO( SKIP_next, next );
5677 cur = cur->u.mark.prev_mark;
5680 /* Didn't find our (*MARK:NAME) so ignore this (*SKIP:NAME) */
5682 case SKIP_next_fail:
5684 /* (*CUT:NAME) - Set up to search for the name as we
5685 collapse the stack*/
5686 popmark = ST.mark_name;
5688 /* (*CUT) - No name, we cut here.*/
5689 if (ST.mark_loc > startpoint)
5690 reginfo->cutpoint = HOPBACKc(ST.mark_loc, 1);
5691 /* but we set sv_commit to latest mark_name if there
5692 is one so they can test to see how things lead to this
5695 sv_commit=mark_state->u.mark.mark_name;
5703 if ( n == (U32)what_len_TRICKYFOLD(locinput,utf8_target,ln) ) {
5705 } else if ( LATIN_SMALL_LETTER_SHARP_S == n && !utf8_target && !UTF_PATTERN ) {
5708 U8 folded[UTF8_MAXBYTES_CASE+1];
5710 const char * const l = locinput;
5711 char *e = PL_regeol;
5712 to_uni_fold(n, folded, &foldlen);
5714 if (! foldEQ_utf8((const char*) folded, 0, foldlen, 1,
5715 l, &e, 0, utf8_target)) {
5720 nextchr = UCHARAT(locinput);
5723 if ((n=is_LNBREAK(locinput,utf8_target))) {
5725 nextchr = UCHARAT(locinput);
5730 #define CASE_CLASS(nAmE) \
5732 if ((n=is_##nAmE(locinput,utf8_target))) { \
5734 nextchr = UCHARAT(locinput); \
5739 if ((n=is_##nAmE(locinput,utf8_target))) { \
5742 locinput += UTF8SKIP(locinput); \
5743 nextchr = UCHARAT(locinput); \
5748 CASE_CLASS(HORIZWS);
5752 PerlIO_printf(Perl_error_log, "%"UVxf" %d\n",
5753 PTR2UV(scan), OP(scan));
5754 Perl_croak(aTHX_ "regexp memory corruption");
5758 /* switch break jumps here */
5759 scan = next; /* prepare to execute the next op and ... */
5760 continue; /* ... jump back to the top, reusing st */
5764 /* push a state that backtracks on success */
5765 st->u.yes.prev_yes_state = yes_state;
5769 /* push a new regex state, then continue at scan */
5771 regmatch_state *newst;
5774 regmatch_state *cur = st;
5775 regmatch_state *curyes = yes_state;
5777 regmatch_slab *slab = PL_regmatch_slab;
5778 for (;curd > -1;cur--,curd--) {
5779 if (cur < SLAB_FIRST(slab)) {
5781 cur = SLAB_LAST(slab);
5783 PerlIO_printf(Perl_error_log, "%*s#%-3d %-10s %s\n",
5784 REPORT_CODE_OFF + 2 + depth * 2,"",
5785 curd, PL_reg_name[cur->resume_state],
5786 (curyes == cur) ? "yes" : ""
5789 curyes = cur->u.yes.prev_yes_state;
5792 DEBUG_STATE_pp("push")
5795 st->locinput = locinput;
5797 if (newst > SLAB_LAST(PL_regmatch_slab))
5798 newst = S_push_slab(aTHX);
5799 PL_regmatch_state = newst;
5801 locinput = PL_reginput;
5802 nextchr = UCHARAT(locinput);
5810 * We get here only if there's trouble -- normally "case END" is
5811 * the terminating point.
5813 Perl_croak(aTHX_ "corrupted regexp pointers");
5819 /* we have successfully completed a subexpression, but we must now
5820 * pop to the state marked by yes_state and continue from there */
5821 assert(st != yes_state);
5823 while (st != yes_state) {
5825 if (st < SLAB_FIRST(PL_regmatch_slab)) {
5826 PL_regmatch_slab = PL_regmatch_slab->prev;
5827 st = SLAB_LAST(PL_regmatch_slab);
5831 DEBUG_STATE_pp("pop (no final)");
5833 DEBUG_STATE_pp("pop (yes)");
5839 while (yes_state < SLAB_FIRST(PL_regmatch_slab)
5840 || yes_state > SLAB_LAST(PL_regmatch_slab))
5842 /* not in this slab, pop slab */
5843 depth -= (st - SLAB_FIRST(PL_regmatch_slab) + 1);
5844 PL_regmatch_slab = PL_regmatch_slab->prev;
5845 st = SLAB_LAST(PL_regmatch_slab);
5847 depth -= (st - yes_state);
5850 yes_state = st->u.yes.prev_yes_state;
5851 PL_regmatch_state = st;
5854 locinput= st->locinput;
5855 nextchr = UCHARAT(locinput);
5857 state_num = st->resume_state + no_final;
5858 goto reenter_switch;
5861 DEBUG_EXECUTE_r(PerlIO_printf(Perl_debug_log, "%sMatch successful!%s\n",
5862 PL_colors[4], PL_colors[5]));
5864 if (PL_reg_eval_set) {
5865 /* each successfully executed (?{...}) block does the equivalent of
5866 * local $^R = do {...}
5867 * When popping the save stack, all these locals would be undone;
5868 * bypass this by setting the outermost saved $^R to the latest
5870 if (oreplsv != GvSV(PL_replgv))
5871 sv_setsv(oreplsv, GvSV(PL_replgv));
5878 PerlIO_printf(Perl_debug_log,
5879 "%*s %sfailed...%s\n",
5880 REPORT_CODE_OFF+depth*2, "",
5881 PL_colors[4], PL_colors[5])
5893 /* there's a previous state to backtrack to */
5895 if (st < SLAB_FIRST(PL_regmatch_slab)) {
5896 PL_regmatch_slab = PL_regmatch_slab->prev;
5897 st = SLAB_LAST(PL_regmatch_slab);
5899 PL_regmatch_state = st;
5900 locinput= st->locinput;
5901 nextchr = UCHARAT(locinput);
5903 DEBUG_STATE_pp("pop");
5905 if (yes_state == st)
5906 yes_state = st->u.yes.prev_yes_state;
5908 state_num = st->resume_state + 1; /* failure = success + 1 */
5909 goto reenter_switch;
5914 if (rex->intflags & PREGf_VERBARG_SEEN) {
5915 SV *sv_err = get_sv("REGERROR", 1);
5916 SV *sv_mrk = get_sv("REGMARK", 1);
5918 sv_commit = &PL_sv_no;
5920 sv_yes_mark = &PL_sv_yes;
5923 sv_commit = &PL_sv_yes;
5924 sv_yes_mark = &PL_sv_no;
5926 sv_setsv(sv_err, sv_commit);
5927 sv_setsv(sv_mrk, sv_yes_mark);
5930 /* clean up; in particular, free all slabs above current one */
5931 LEAVE_SCOPE(oldsave);
5937 - regrepeat - repeatedly match something simple, report how many
5940 * [This routine now assumes that it will only match on things of length 1.
5941 * That was true before, but now we assume scan - reginput is the count,
5942 * rather than incrementing count on every character. [Er, except utf8.]]
5945 S_regrepeat(pTHX_ const regexp *prog, const regnode *p, I32 max, int depth)
5948 register char *scan;
5950 register char *loceol = PL_regeol;
5951 register I32 hardcount = 0;
5952 register bool utf8_target = PL_reg_match_utf8;
5954 PERL_UNUSED_ARG(depth);
5957 PERL_ARGS_ASSERT_REGREPEAT;
5960 if (max == REG_INFTY)
5962 else if (max < loceol - scan)
5963 loceol = scan + max;
5968 while (scan < loceol && hardcount < max && *scan != '\n') {
5969 scan += UTF8SKIP(scan);
5973 while (scan < loceol && *scan != '\n')
5980 while (scan < loceol && hardcount < max) {
5981 scan += UTF8SKIP(scan);
5992 /* To get here, EXACTish nodes must have *byte* length == 1. That
5993 * means they match only characters in the string that can be expressed
5994 * as a single byte. For non-utf8 strings, that means a simple match.
5995 * For utf8 strings, the character matched must be an invariant, or
5996 * downgradable to a single byte. The pattern's utf8ness is
5997 * irrelevant, as since it's a single byte, it either isn't utf8, or if
5998 * it is, it's an invariant */
6001 assert(! UTF_PATTERN || UNI_IS_INVARIANT(c));
6003 if (! utf8_target || UNI_IS_INVARIANT(c)) {
6004 while (scan < loceol && UCHARAT(scan) == c) {
6010 /* Here, the string is utf8, and the pattern char is different
6011 * in utf8 than not, so can't compare them directly. Outside the
6012 * loop, find find the two utf8 bytes that represent c, and then
6013 * look for those in sequence in the utf8 string */
6014 U8 high = UTF8_TWO_BYTE_HI(c);
6015 U8 low = UTF8_TWO_BYTE_LO(c);
6018 while (hardcount < max
6019 && scan + 1 < loceol
6020 && UCHARAT(scan) == high
6021 && UCHARAT(scan + 1) == low)
6029 PL_reg_flags |= RF_tainted;
6034 /* The comments for the EXACT case above apply as well to these fold
6038 assert(! UTF_PATTERN || UNI_IS_INVARIANT(c));
6040 if (utf8_target) { /* Use full Unicode fold matching */
6042 /* For the EXACTFL case, It doesn't really make sense to compare
6043 * locale and utf8, but it is best we can do. The documents warn
6044 * against mixing them */
6046 char *tmpeol = loceol;
6047 while (hardcount < max
6048 && foldEQ_utf8(scan, &tmpeol, 0, utf8_target,
6049 STRING(p), NULL, 1, cBOOL(UTF_PATTERN)))
6056 /* XXX Note that the above handles properly the German sharp s in
6057 * the pattern matching ss in the string. But it doesn't handle
6058 * properly cases where the string contains say 'LIGATURE ff' and
6059 * the pattern is 'f+'. This would require, say, a new function or
6060 * revised interface to foldEQ_utf8(), in which the maximum number
6061 * of characters to match could be passed and it would return how
6062 * many actually did. This is just one of many cases where
6063 * multi-char folds don't work properly, and so the fix is being
6069 /* Here, the string isn't utf8 and c is a single byte; and either
6070 * the pattern isn't utf8 or c is an invariant, so its utf8ness
6071 * doesn't affect c. Can just do simple comparisons for exact or
6074 case EXACTF: folded = PL_fold[c]; break;
6075 case EXACTFU: folded = PL_fold_latin1[c]; break;
6076 case EXACTFL: folded = PL_fold_locale[c]; break;
6077 default: Perl_croak(aTHX_ "panic: Unexpected op %u", OP(p));
6079 while (scan < loceol &&
6080 (UCHARAT(scan) == c || UCHARAT(scan) == folded))
6090 while (hardcount < max && scan < loceol &&
6091 reginclass(prog, p, (U8*)scan, 0, utf8_target)) {
6092 scan += UTF8SKIP(scan);
6096 while (scan < loceol && REGINCLASS(prog, p, (U8*)scan))
6104 LOAD_UTF8_CHARCLASS_ALNUM();
6105 while (hardcount < max && scan < loceol &&
6106 swash_fetch(PL_utf8_alnum, (U8*)scan, utf8_target))
6108 scan += UTF8SKIP(scan);
6112 while (scan < loceol && isWORDCHAR_L1((U8) *scan)) {
6120 while (scan < loceol && isALNUM((U8) *scan)) {
6125 while (scan < loceol && isWORDCHAR_A((U8) *scan)) {
6130 PL_reg_flags |= RF_tainted;
6133 while (hardcount < max && scan < loceol &&
6134 isALNUM_LC_utf8((U8*)scan)) {
6135 scan += UTF8SKIP(scan);
6139 while (scan < loceol && isALNUM_LC(*scan))
6149 LOAD_UTF8_CHARCLASS_ALNUM();
6150 while (hardcount < max && scan < loceol &&
6151 ! swash_fetch(PL_utf8_alnum, (U8*)scan, utf8_target))
6153 scan += UTF8SKIP(scan);
6157 while (scan < loceol && ! isWORDCHAR_L1((U8) *scan)) {
6164 goto utf8_Nwordchar;
6165 while (scan < loceol && ! isALNUM((U8) *scan)) {
6171 while (scan < loceol && ! isWORDCHAR_A((U8) *scan)) {
6172 scan += UTF8SKIP(scan);
6176 while (scan < loceol && ! isWORDCHAR_A((U8) *scan)) {
6182 PL_reg_flags |= RF_tainted;
6185 while (hardcount < max && scan < loceol &&
6186 !isALNUM_LC_utf8((U8*)scan)) {
6187 scan += UTF8SKIP(scan);
6191 while (scan < loceol && !isALNUM_LC(*scan))
6201 LOAD_UTF8_CHARCLASS_SPACE();
6202 while (hardcount < max && scan < loceol &&
6204 swash_fetch(PL_utf8_space,(U8*)scan, utf8_target)))
6206 scan += UTF8SKIP(scan);
6212 while (scan < loceol && isSPACE_L1((U8) *scan)) {
6221 while (scan < loceol && isSPACE((U8) *scan)) {
6226 while (scan < loceol && isSPACE_A((U8) *scan)) {
6231 PL_reg_flags |= RF_tainted;
6234 while (hardcount < max && scan < loceol &&
6235 isSPACE_LC_utf8((U8*)scan)) {
6236 scan += UTF8SKIP(scan);
6240 while (scan < loceol && isSPACE_LC(*scan))
6250 LOAD_UTF8_CHARCLASS_SPACE();
6251 while (hardcount < max && scan < loceol &&
6253 swash_fetch(PL_utf8_space,(U8*)scan, utf8_target)))
6255 scan += UTF8SKIP(scan);
6261 while (scan < loceol && ! isSPACE_L1((U8) *scan)) {
6270 while (scan < loceol && ! isSPACE((U8) *scan)) {
6276 while (scan < loceol && ! isSPACE_A((U8) *scan)) {
6277 scan += UTF8SKIP(scan);
6281 while (scan < loceol && ! isSPACE_A((U8) *scan)) {
6287 PL_reg_flags |= RF_tainted;
6290 while (hardcount < max && scan < loceol &&
6291 !isSPACE_LC_utf8((U8*)scan)) {
6292 scan += UTF8SKIP(scan);
6296 while (scan < loceol && !isSPACE_LC(*scan))
6303 LOAD_UTF8_CHARCLASS_DIGIT();
6304 while (hardcount < max && scan < loceol &&
6305 swash_fetch(PL_utf8_digit, (U8*)scan, utf8_target)) {
6306 scan += UTF8SKIP(scan);
6310 while (scan < loceol && isDIGIT(*scan))
6315 while (scan < loceol && isDIGIT_A((U8) *scan)) {
6320 PL_reg_flags |= RF_tainted;
6323 while (hardcount < max && scan < loceol &&
6324 isDIGIT_LC_utf8((U8*)scan)) {
6325 scan += UTF8SKIP(scan);
6329 while (scan < loceol && isDIGIT_LC(*scan))
6336 LOAD_UTF8_CHARCLASS_DIGIT();
6337 while (hardcount < max && scan < loceol &&
6338 !swash_fetch(PL_utf8_digit, (U8*)scan, utf8_target)) {
6339 scan += UTF8SKIP(scan);
6343 while (scan < loceol && !isDIGIT(*scan))
6349 while (scan < loceol && ! isDIGIT_A((U8) *scan)) {
6350 scan += UTF8SKIP(scan);
6354 while (scan < loceol && ! isDIGIT_A((U8) *scan)) {
6360 PL_reg_flags |= RF_tainted;
6363 while (hardcount < max && scan < loceol &&
6364 !isDIGIT_LC_utf8((U8*)scan)) {
6365 scan += UTF8SKIP(scan);
6369 while (scan < loceol && !isDIGIT_LC(*scan))
6376 while (hardcount < max && scan < loceol && (c=is_LNBREAK_utf8(scan))) {
6382 LNBREAK can match two latin chars, which is ok,
6383 because we have a null terminated string, but we
6384 have to use hardcount in this situation
6386 while (scan < loceol && (c=is_LNBREAK_latin1(scan))) {
6395 while (hardcount < max && scan < loceol && (c=is_HORIZWS_utf8(scan))) {
6400 while (scan < loceol && is_HORIZWS_latin1(scan))
6407 while (hardcount < max && scan < loceol && !is_HORIZWS_utf8(scan)) {
6408 scan += UTF8SKIP(scan);
6412 while (scan < loceol && !is_HORIZWS_latin1(scan))
6420 while (hardcount < max && scan < loceol && (c=is_VERTWS_utf8(scan))) {
6425 while (scan < loceol && is_VERTWS_latin1(scan))
6433 while (hardcount < max && scan < loceol && !is_VERTWS_utf8(scan)) {
6434 scan += UTF8SKIP(scan);
6438 while (scan < loceol && !is_VERTWS_latin1(scan))
6444 default: /* Called on something of 0 width. */
6445 break; /* So match right here or not at all. */
6451 c = scan - PL_reginput;
6455 GET_RE_DEBUG_FLAGS_DECL;
6457 SV * const prop = sv_newmortal();
6458 regprop(prog, prop, p);
6459 PerlIO_printf(Perl_debug_log,
6460 "%*s %s can match %"IVdf" times out of %"IVdf"...\n",
6461 REPORT_CODE_OFF + depth*2, "", SvPVX_const(prop),(IV)c,(IV)max);
6469 #if !defined(PERL_IN_XSUB_RE) || defined(PLUGGABLE_RE_EXTENSION)
6471 - regclass_swash - prepare the utf8 swash
6475 Perl_regclass_swash(pTHX_ const regexp *prog, register const regnode* node, bool doinit, SV** listsvp, SV **altsvp)
6481 RXi_GET_DECL(prog,progi);
6482 const struct reg_data * const data = prog ? progi->data : NULL;
6484 PERL_ARGS_ASSERT_REGCLASS_SWASH;
6486 if (data && data->count) {
6487 const U32 n = ARG(node);
6489 if (data->what[n] == 's') {
6490 SV * const rv = MUTABLE_SV(data->data[n]);
6491 AV * const av = MUTABLE_AV(SvRV(rv));
6492 SV **const ary = AvARRAY(av);
6495 /* See the end of regcomp.c:S_regclass() for
6496 * documentation of these array elements. */
6499 a = SvROK(ary[1]) ? &ary[1] : NULL;
6500 b = SvTYPE(ary[2]) == SVt_PVAV ? &ary[2] : NULL;
6504 else if (si && doinit) {
6505 sw = swash_init("utf8", "", si, 1, 0);
6506 (void)av_store(av, 1, sw);
6523 - reginclass - determine if a character falls into a character class
6525 n is the ANYOF regnode
6526 p is the target string
6527 lenp is pointer to the maximum number of bytes of how far to go in p
6528 (This is assumed wthout checking to always be at least the current
6530 utf8_target tells whether p is in UTF-8.
6532 Returns true if matched; false otherwise. If lenp is not NULL, on return
6533 from a successful match, the value it points to will be updated to how many
6534 bytes in p were matched. If there was no match, the value is undefined,
6535 possibly changed from the input.
6537 Note that this can be a synthetic start class, a combination of various
6538 nodes, so things you think might be mutually exclusive, such as locale,
6539 aren't. It can match both locale and non-locale
6544 S_reginclass(pTHX_ const regexp * const prog, register const regnode * const n, register const U8* const p, STRLEN* lenp, register const bool utf8_target)
6547 const char flags = ANYOF_FLAGS(n);
6553 PERL_ARGS_ASSERT_REGINCLASS;
6555 /* If c is not already the code point, get it */
6556 if (utf8_target && !UTF8_IS_INVARIANT(c)) {
6557 c = utf8n_to_uvchr(p, UTF8_MAXBYTES, &c_len,
6558 (UTF8_ALLOW_DEFAULT & UTF8_ALLOW_ANYUV)
6559 | UTF8_ALLOW_FFFF | UTF8_CHECK_ONLY);
6560 /* see [perl #37836] for UTF8_ALLOW_ANYUV; [perl #38293] for
6561 * UTF8_ALLOW_FFFF */
6562 if (c_len == (STRLEN)-1)
6563 Perl_croak(aTHX_ "Malformed UTF-8 character (fatal)");
6569 /* Use passed in max length, or one character if none passed in or less
6570 * than one character. And assume will match just one character. This is
6571 * overwritten later if matched more. */
6573 maxlen = (*lenp > c_len) ? *lenp : c_len;
6581 /* If this character is potentially in the bitmap, check it */
6583 if (ANYOF_BITMAP_TEST(n, c))
6585 else if (flags & ANYOF_NON_UTF8_LATIN1_ALL
6592 else if (flags & ANYOF_LOCALE) {
6593 PL_reg_flags |= RF_tainted;
6595 if ((flags & ANYOF_LOC_NONBITMAP_FOLD)
6596 && ANYOF_BITMAP_TEST(n, PL_fold_locale[c]))
6600 else if (ANYOF_CLASS_TEST_ANY_SET(n) &&
6601 ((ANYOF_CLASS_TEST(n, ANYOF_ALNUM) && isALNUM_LC(c)) ||
6602 (ANYOF_CLASS_TEST(n, ANYOF_NALNUM) && !isALNUM_LC(c)) ||
6603 (ANYOF_CLASS_TEST(n, ANYOF_SPACE) && isSPACE_LC(c)) ||
6604 (ANYOF_CLASS_TEST(n, ANYOF_NSPACE) && !isSPACE_LC(c)) ||
6605 (ANYOF_CLASS_TEST(n, ANYOF_DIGIT) && isDIGIT_LC(c)) ||
6606 (ANYOF_CLASS_TEST(n, ANYOF_NDIGIT) && !isDIGIT_LC(c)) ||
6607 (ANYOF_CLASS_TEST(n, ANYOF_ALNUMC) && isALNUMC_LC(c)) ||
6608 (ANYOF_CLASS_TEST(n, ANYOF_NALNUMC) && !isALNUMC_LC(c)) ||
6609 (ANYOF_CLASS_TEST(n, ANYOF_ALPHA) && isALPHA_LC(c)) ||
6610 (ANYOF_CLASS_TEST(n, ANYOF_NALPHA) && !isALPHA_LC(c)) ||
6611 (ANYOF_CLASS_TEST(n, ANYOF_ASCII) && isASCII(c)) ||
6612 (ANYOF_CLASS_TEST(n, ANYOF_NASCII) && !isASCII(c)) ||
6613 (ANYOF_CLASS_TEST(n, ANYOF_CNTRL) && isCNTRL_LC(c)) ||
6614 (ANYOF_CLASS_TEST(n, ANYOF_NCNTRL) && !isCNTRL_LC(c)) ||
6615 (ANYOF_CLASS_TEST(n, ANYOF_GRAPH) && isGRAPH_LC(c)) ||
6616 (ANYOF_CLASS_TEST(n, ANYOF_NGRAPH) && !isGRAPH_LC(c)) ||
6617 (ANYOF_CLASS_TEST(n, ANYOF_LOWER) && isLOWER_LC(c)) ||
6618 (ANYOF_CLASS_TEST(n, ANYOF_NLOWER) && !isLOWER_LC(c)) ||
6619 (ANYOF_CLASS_TEST(n, ANYOF_PRINT) && isPRINT_LC(c)) ||
6620 (ANYOF_CLASS_TEST(n, ANYOF_NPRINT) && !isPRINT_LC(c)) ||
6621 (ANYOF_CLASS_TEST(n, ANYOF_PUNCT) && isPUNCT_LC(c)) ||
6622 (ANYOF_CLASS_TEST(n, ANYOF_NPUNCT) && !isPUNCT_LC(c)) ||
6623 (ANYOF_CLASS_TEST(n, ANYOF_UPPER) && isUPPER_LC(c)) ||
6624 (ANYOF_CLASS_TEST(n, ANYOF_NUPPER) && !isUPPER_LC(c)) ||
6625 (ANYOF_CLASS_TEST(n, ANYOF_XDIGIT) && isXDIGIT(c)) ||
6626 (ANYOF_CLASS_TEST(n, ANYOF_NXDIGIT) && !isXDIGIT(c)) ||
6627 (ANYOF_CLASS_TEST(n, ANYOF_PSXSPC) && isPSXSPC(c)) ||
6628 (ANYOF_CLASS_TEST(n, ANYOF_NPSXSPC) && !isPSXSPC(c)) ||
6629 (ANYOF_CLASS_TEST(n, ANYOF_BLANK) && isBLANK(c)) ||
6630 (ANYOF_CLASS_TEST(n, ANYOF_NBLANK) && !isBLANK(c))
6631 ) /* How's that for a conditional? */
6638 /* If the bitmap didn't (or couldn't) match, and something outside the
6639 * bitmap could match, try that */
6641 if (utf8_target && (flags & ANYOF_UNICODE_ALL)) {
6643 || ((flags & ANYOF_LOC_NONBITMAP_FOLD) /* Latin1 1 that has a
6646 && _HAS_NONLATIN1_FOLD_CLOSURE_ONLY_FOR_USE_BY_REGCOMP_DOT_C_AND_REGEXEC_DOT_C(c)))
6651 if (!match && ((flags & ANYOF_NONBITMAP_NON_UTF8)
6652 || (utf8_target && flags & ANYOF_UTF8)))
6655 SV * const sw = regclass_swash(prog, n, TRUE, 0, (SV**)&av);
6663 /* Not utf8. Convert as much of the string as available up
6664 * to the limit of how far the (single) character in the
6665 * pattern can possibly match (no need to go further). If
6666 * the node is a straight ANYOF or not folding, it can't
6667 * match more than one. Otherwise, It can match up to how
6668 * far a single char can fold to. Since not utf8, each
6669 * character is a single byte, so the max it can be in
6670 * bytes is the same as the max it can be in characters */
6671 STRLEN len = (OP(n) == ANYOF
6672 || ! (flags & ANYOF_LOC_NONBITMAP_FOLD))
6674 : (maxlen < UTF8_MAX_FOLD_CHAR_EXPAND)
6676 : UTF8_MAX_FOLD_CHAR_EXPAND;
6677 utf8_p = bytes_to_utf8(p, &len);
6680 if (swash_fetch(sw, utf8_p, TRUE))
6682 else if (flags & ANYOF_LOC_NONBITMAP_FOLD) {
6684 /* Here, we need to test if the fold of the target string
6685 * matches. In the case of a multi-char fold that is
6686 * caught by regcomp.c, it has stored all such folds into
6687 * 'av'; we linearly check to see if any match the target
6688 * string (folded). We know that the originals were each
6689 * one character, but we don't currently know how many
6690 * characters/bytes each folded to, except we do know that
6691 * there are small limits imposed by Unicode. XXX A
6692 * performance enhancement would be to have regcomp.c store
6693 * the max number of chars/bytes that are in an av entry,
6694 * as, say the 0th element. Even better would be to have a
6695 * hash of the few characters that can start a multi-char
6696 * fold to the max number of chars of those folds.
6698 * Further down, if there isn't a
6699 * match in the av, we will check if there is another
6700 * fold-type match. For that, we also need the fold, but
6701 * only the first character. No sense in folding it twice,
6702 * so we do it here, even if there isn't any multi-char
6703 * fold, so we always fold at least the first character.
6704 * If the node is a straight ANYOF node, or there is only
6705 * one character available in the string, or if there isn't
6706 * any av, that's all we have to fold. In the case of a
6707 * multi-char fold, we do have guarantees in Unicode that
6708 * it can only expand up to so many characters and so many
6709 * bytes. We keep track so don't exceed either.
6711 * If there is a match, we will need to advance (if lenp is
6712 * specified) the match pointer in the target string. But
6713 * what we are comparing here isn't that string directly,
6714 * but its fold, whose length may differ from the original.
6715 * As we go along in constructing the fold, therefore, we
6716 * create a map so that we know how many bytes in the
6717 * source to advance given that we have matched a certain
6718 * number of bytes in the fold. This map is stored in
6719 * 'map_fold_len_back'. The first character in the fold
6720 * has array element 1 contain the number of bytes in the
6721 * source that folded to it; the 2nd is the cumulative
6722 * number to match it; ... */
6723 U8 map_fold_len_back[UTF8_MAX_FOLD_CHAR_EXPAND] = { 0 };
6724 U8 folded[UTF8_MAXBYTES_CASE+1];
6725 STRLEN foldlen = 0; /* num bytes in fold of 1st char */
6726 STRLEN foldlen_for_av; /* num bytes in fold of all chars */
6728 if (OP(n) == ANYOF || maxlen == 1 || ! lenp || ! av) {
6730 /* Here, only need to fold the first char of the target
6732 to_utf8_fold(utf8_p, folded, &foldlen);
6733 foldlen_for_av = foldlen;
6734 map_fold_len_back[1] = UTF8SKIP(utf8_p);
6738 /* Here, need to fold more than the first char. Do so
6739 * up to the limits */
6741 U8* source_ptr = utf8_p; /* The source for the fold
6744 U8* folded_ptr = folded;
6745 U8* e = utf8_p + maxlen; /* Can't go beyond last
6746 available byte in the
6748 while (which_char < UTF8_MAX_FOLD_CHAR_EXPAND
6752 /* Fold the next character */
6753 U8 this_char_folded[UTF8_MAXBYTES_CASE+1];
6754 STRLEN this_char_foldlen;
6755 to_utf8_fold(source_ptr,
6757 &this_char_foldlen);
6759 /* Bail if it would exceed the byte limit for
6760 * folding a single char. */
6761 if (this_char_foldlen + folded_ptr - folded >
6767 /* Save the first character's folded length, in
6768 * case we have to use it later */
6770 foldlen = this_char_foldlen;
6773 /* Here, add the fold of this character */
6774 Copy(this_char_folded,
6779 map_fold_len_back[which_char] =
6780 map_fold_len_back[which_char - 1]
6781 + UTF8SKIP(source_ptr);
6782 folded_ptr += this_char_foldlen;
6783 source_ptr += UTF8SKIP(source_ptr);
6786 foldlen_for_av = folded_ptr - folded;
6790 /* Do the linear search to see if the fold is in the list
6791 * of multi-char folds. (Useless to look if won't be able
6792 * to store that it is a multi-char fold in *lenp) */
6795 for (i = 0; i <= av_len(av); i++) {
6796 SV* const sv = *av_fetch(av, i, FALSE);
6798 const char * const s = SvPV_const(sv, len);
6799 if (len <= foldlen_for_av && memEQ(s,
6804 /* Advance the target string ptr to account for
6805 * this fold, but have to translate from the
6806 * folded length to the corresponding source
6807 * length. The array is indexed by how many
6808 * characters in the match */
6809 *lenp = map_fold_len_back[
6810 utf8_length(folded, folded + len)];
6816 if (!match) { /* See if the folded version matches */
6819 /* Consider "k" =~ /[K]/i. The line above would have
6820 * just folded the 'k' to itself, and that isn't going
6821 * to match 'K'. So we look through the closure of
6822 * everything that folds to 'k'. That will find the
6823 * 'K'. Initialize the list, if necessary */
6824 if (! PL_utf8_foldclosures) {
6826 /* If the folds haven't been read in, call a fold
6827 * function to force that */
6828 if (! PL_utf8_tofold) {
6829 U8 dummy[UTF8_MAXBYTES+1];
6831 to_utf8_fold((U8*) "A", dummy, &dummy_len);
6833 PL_utf8_foldclosures =
6834 _swash_inversion_hash(PL_utf8_tofold);
6837 /* The data structure is a hash with the keys every
6838 * character that is folded to, like 'k', and the
6839 * values each an array of everything that folds to its
6840 * key. e.g. [ 'k', 'K', KELVIN_SIGN ] */
6841 if ((listp = hv_fetch(PL_utf8_foldclosures,
6842 (char *) folded, foldlen, FALSE)))
6844 AV* list = (AV*) *listp;
6846 for (i = 0; i <= av_len(list); i++) {
6847 SV** try_p = av_fetch(list, i, FALSE);
6849 if (try_p == NULL) {
6850 Perl_croak(aTHX_ "panic: invalid PL_utf8_foldclosures structure");
6852 /* Don't have to worry about embedded nulls
6853 * since NULL isn't folded or foldable */
6854 try_c = SvPVX(*try_p);
6856 /* The fold in a few cases of an above Latin1
6857 * char is in the Latin1 range, and hence may
6858 * be in the bitmap */
6859 if (UTF8_IS_INVARIANT(*try_c)
6860 && ANYOF_BITMAP_TEST(n,
6861 UNI_TO_NATIVE(*try_c)))
6867 (UTF8_IS_DOWNGRADEABLE_START(*try_c)
6868 && ANYOF_BITMAP_TEST(n, UNI_TO_NATIVE(
6869 TWO_BYTE_UTF8_TO_UNI(try_c[0],
6872 /* Since the fold comes from internally
6873 * generated data, we can safely assume it
6874 * is valid utf8 in the test above */
6877 } else if (swash_fetch(sw, (U8*) try_c, TRUE)) {
6886 /* If we allocated a string above, free it */
6887 if (! utf8_target) Safefree(utf8_p);
6892 return (flags & ANYOF_INVERT) ? !match : match;
6896 S_reghop3(U8 *s, I32 off, const U8* lim)
6900 PERL_ARGS_ASSERT_REGHOP3;
6903 while (off-- && s < lim) {
6904 /* XXX could check well-formedness here */
6909 while (off++ && s > lim) {
6911 if (UTF8_IS_CONTINUED(*s)) {
6912 while (s > lim && UTF8_IS_CONTINUATION(*s))
6915 /* XXX could check well-formedness here */
6922 /* there are a bunch of places where we use two reghop3's that should
6923 be replaced with this routine. but since thats not done yet
6924 we ifdef it out - dmq
6927 S_reghop4(U8 *s, I32 off, const U8* llim, const U8* rlim)
6931 PERL_ARGS_ASSERT_REGHOP4;
6934 while (off-- && s < rlim) {
6935 /* XXX could check well-formedness here */
6940 while (off++ && s > llim) {
6942 if (UTF8_IS_CONTINUED(*s)) {
6943 while (s > llim && UTF8_IS_CONTINUATION(*s))
6946 /* XXX could check well-formedness here */
6954 S_reghopmaybe3(U8* s, I32 off, const U8* lim)
6958 PERL_ARGS_ASSERT_REGHOPMAYBE3;
6961 while (off-- && s < lim) {
6962 /* XXX could check well-formedness here */
6969 while (off++ && s > lim) {
6971 if (UTF8_IS_CONTINUED(*s)) {
6972 while (s > lim && UTF8_IS_CONTINUATION(*s))
6975 /* XXX could check well-formedness here */
6984 restore_pos(pTHX_ void *arg)
6987 regexp * const rex = (regexp *)arg;
6988 if (PL_reg_eval_set) {
6989 if (PL_reg_oldsaved) {
6990 rex->subbeg = PL_reg_oldsaved;
6991 rex->sublen = PL_reg_oldsavedlen;
6992 #ifdef PERL_OLD_COPY_ON_WRITE
6993 rex->saved_copy = PL_nrs;
6995 RXp_MATCH_COPIED_on(rex);
6997 PL_reg_magic->mg_len = PL_reg_oldpos;
6998 PL_reg_eval_set = 0;
6999 PL_curpm = PL_reg_oldcurpm;
7004 S_to_utf8_substr(pTHX_ register regexp *prog)
7008 PERL_ARGS_ASSERT_TO_UTF8_SUBSTR;
7011 if (prog->substrs->data[i].substr
7012 && !prog->substrs->data[i].utf8_substr) {
7013 SV* const sv = newSVsv(prog->substrs->data[i].substr);
7014 prog->substrs->data[i].utf8_substr = sv;
7015 sv_utf8_upgrade(sv);
7016 if (SvVALID(prog->substrs->data[i].substr)) {
7017 const U8 flags = BmFLAGS(prog->substrs->data[i].substr);
7018 if (flags & FBMcf_TAIL) {
7019 /* Trim the trailing \n that fbm_compile added last
7021 SvCUR_set(sv, SvCUR(sv) - 1);
7022 /* Whilst this makes the SV technically "invalid" (as its
7023 buffer is no longer followed by "\0") when fbm_compile()
7024 adds the "\n" back, a "\0" is restored. */
7026 fbm_compile(sv, flags);
7028 if (prog->substrs->data[i].substr == prog->check_substr)
7029 prog->check_utf8 = sv;
7035 S_to_byte_substr(pTHX_ register regexp *prog)
7040 PERL_ARGS_ASSERT_TO_BYTE_SUBSTR;
7043 if (prog->substrs->data[i].utf8_substr
7044 && !prog->substrs->data[i].substr) {
7045 SV* sv = newSVsv(prog->substrs->data[i].utf8_substr);
7046 if (sv_utf8_downgrade(sv, TRUE)) {
7047 if (SvVALID(prog->substrs->data[i].utf8_substr)) {
7049 = BmFLAGS(prog->substrs->data[i].utf8_substr);
7050 if (flags & FBMcf_TAIL) {
7051 /* Trim the trailing \n that fbm_compile added last
7053 SvCUR_set(sv, SvCUR(sv) - 1);
7055 fbm_compile(sv, flags);
7061 prog->substrs->data[i].substr = sv;
7062 if (prog->substrs->data[i].utf8_substr == prog->check_utf8)
7063 prog->check_substr = sv;
7070 * c-indentation-style: bsd
7072 * indent-tabs-mode: t
7075 * ex: set ts=8 sts=4 sw=4 noet: