5 * "One Ring to rule them all, One Ring to find them..."
8 /* This file contains functions for executing a regular expression. See
9 * also regcomp.c which funnily enough, contains functions for compiling
10 * a regular expression.
12 * This file is also copied at build time to ext/re/re_exec.c, where
13 * it's built with -DPERL_EXT_RE_BUILD -DPERL_EXT_RE_DEBUG -DPERL_EXT.
14 * This causes the main functions to be compiled under new names and with
15 * debugging support added, which makes "use re 'debug'" work.
19 /* NOTE: this is derived from Henry Spencer's regexp code, and should not
20 * confused with the original package (see point 3 below). Thanks, Henry!
23 /* Additional note: this code is very heavily munged from Henry's version
24 * in places. In some spots I've traded clarity for efficiency, so don't
25 * blame Henry for some of the lack of readability.
28 /* The names of the functions have been changed from regcomp and
29 * regexec to pregcomp and pregexec in order to avoid conflicts
30 * with the POSIX routines of the same names.
33 #ifdef PERL_EXT_RE_BUILD
34 /* need to replace pregcomp et al, so enable that */
35 # ifndef PERL_IN_XSUB_RE
36 # define PERL_IN_XSUB_RE
38 /* need access to debugger hooks */
39 # if defined(PERL_EXT_RE_DEBUG) && !defined(DEBUGGING)
44 #ifdef PERL_IN_XSUB_RE
45 /* We *really* need to overwrite these symbols: */
46 # define Perl_regexec_flags my_regexec
47 # define Perl_regdump my_regdump
48 # define Perl_regprop my_regprop
49 # define Perl_re_intuit_start my_re_intuit_start
50 /* *These* symbols are masked to allow static link. */
51 # define Perl_pregexec my_pregexec
52 # define Perl_reginitcolors my_reginitcolors
53 # define Perl_regclass_swash my_regclass_swash
55 # define PERL_NO_GET_CONTEXT
59 * pregcomp and pregexec -- regsub and regerror are not used in perl
61 * Copyright (c) 1986 by University of Toronto.
62 * Written by Henry Spencer. Not derived from licensed software.
64 * Permission is granted to anyone to use this software for any
65 * purpose on any computer system, and to redistribute it freely,
66 * subject to the following restrictions:
68 * 1. The author is not responsible for the consequences of use of
69 * this software, no matter how awful, even if they arise
72 * 2. The origin of this software must not be misrepresented, either
73 * by explicit claim or by omission.
75 * 3. Altered versions must be plainly marked as such, and must not
76 * be misrepresented as being the original software.
78 **** Alterations to Henry's code are...
80 **** Copyright (C) 1991, 1992, 1993, 1994, 1995, 1996, 1997, 1998, 1999,
81 **** 2000, 2001, 2002, 2003, 2004, 2005, 2006, by Larry Wall and others
83 **** You may distribute under the terms of either the GNU General Public
84 **** License or the Artistic License, as specified in the README file.
86 * Beware that some of this code is subtly aware of the way operator
87 * precedence is structured in regular expressions. Serious changes in
88 * regular-expression syntax might require a total rethink.
91 #define PERL_IN_REGEXEC_C
96 #define RF_tainted 1 /* tainted information used? */
97 #define RF_warned 2 /* warned about big count? */
98 #define RF_evaled 4 /* Did an EVAL with setting? */
99 #define RF_utf8 8 /* String contains multibyte chars? */
101 #define UTF ((PL_reg_flags & RF_utf8) != 0)
103 #define RS_init 1 /* eval environment created */
104 #define RS_set 2 /* replsv value is set */
107 #define STATIC static
110 #define REGINCLASS(p,c) (ANYOF_FLAGS(p) ? reginclass(p,c,0,0) : ANYOF_BITMAP_TEST(p,*(c)))
116 #define CHR_SVLEN(sv) (do_utf8 ? sv_len_utf8(sv) : SvCUR(sv))
117 #define CHR_DIST(a,b) (PL_reg_match_utf8 ? utf8_distance(a,b) : a - b)
119 #define reghop_c(pos,off) ((char*)reghop((U8*)pos, off))
120 #define reghopmaybe_c(pos,off) ((char*)reghopmaybe((U8*)pos, off))
121 #define HOP(pos,off) (PL_reg_match_utf8 ? reghop((U8*)pos, off) : (U8*)(pos + off))
122 #define HOPMAYBE(pos,off) (PL_reg_match_utf8 ? reghopmaybe((U8*)pos, off) : (U8*)(pos + off))
123 #define HOPc(pos,off) ((char*)HOP(pos,off))
124 #define HOPMAYBEc(pos,off) ((char*)HOPMAYBE(pos,off))
126 #define HOPBACK(pos, off) ( \
127 (PL_reg_match_utf8) \
128 ? reghopmaybe((U8*)pos, -off) \
129 : (pos - off >= PL_bostr) \
133 #define HOPBACKc(pos, off) (char*)HOPBACK(pos, off)
135 #define reghop3_c(pos,off,lim) ((char*)reghop3((U8*)pos, off, (U8*)lim))
136 #define reghopmaybe3_c(pos,off,lim) ((char*)reghopmaybe3((U8*)pos, off, (U8*)lim))
137 #define HOP3(pos,off,lim) (PL_reg_match_utf8 ? reghop3((U8*)pos, off, (U8*)lim) : (U8*)(pos + off))
138 #define HOPMAYBE3(pos,off,lim) (PL_reg_match_utf8 ? reghopmaybe3((U8*)pos, off, (U8*)lim) : (U8*)(pos + off))
139 #define HOP3c(pos,off,lim) ((char*)HOP3(pos,off,lim))
140 #define HOPMAYBE3c(pos,off,lim) ((char*)HOPMAYBE3(pos,off,lim))
142 #define LOAD_UTF8_CHARCLASS(class,str) STMT_START { \
143 if (!CAT2(PL_utf8_,class)) { bool ok; ENTER; save_re_context(); ok=CAT2(is_utf8_,class)((const U8*)str); assert(ok); LEAVE; } } STMT_END
144 #define LOAD_UTF8_CHARCLASS_ALNUM() LOAD_UTF8_CHARCLASS(alnum,"a")
145 #define LOAD_UTF8_CHARCLASS_DIGIT() LOAD_UTF8_CHARCLASS(digit,"0")
146 #define LOAD_UTF8_CHARCLASS_SPACE() LOAD_UTF8_CHARCLASS(space," ")
147 #define LOAD_UTF8_CHARCLASS_MARK() LOAD_UTF8_CHARCLASS(mark, "\xcd\x86")
149 /* for use after a quantifier and before an EXACT-like node -- japhy */
150 #define JUMPABLE(rn) ( \
151 OP(rn) == OPEN || OP(rn) == CLOSE || OP(rn) == EVAL || \
152 OP(rn) == SUSPEND || OP(rn) == IFMATCH || \
153 OP(rn) == PLUS || OP(rn) == MINMOD || \
154 (PL_regkind[(U8)OP(rn)] == CURLY && ARG1(rn) > 0) \
157 #define HAS_TEXT(rn) ( \
158 PL_regkind[(U8)OP(rn)] == EXACT || PL_regkind[(U8)OP(rn)] == REF \
162 Search for mandatory following text node; for lookahead, the text must
163 follow but for lookbehind (rn->flags != 0) we skip to the next step.
165 #define FIND_NEXT_IMPT(rn) STMT_START { \
166 while (JUMPABLE(rn)) \
167 if (OP(rn) == SUSPEND || PL_regkind[(U8)OP(rn)] == CURLY) \
168 rn = NEXTOPER(NEXTOPER(rn)); \
169 else if (OP(rn) == PLUS) \
171 else if (OP(rn) == IFMATCH) \
172 rn = (rn->flags == 0) ? NEXTOPER(NEXTOPER(rn)) : rn + ARG(rn); \
173 else rn += NEXT_OFF(rn); \
176 static void restore_pos(pTHX_ void *arg);
179 S_regcppush(pTHX_ I32 parenfloor)
182 const int retval = PL_savestack_ix;
183 #define REGCP_PAREN_ELEMS 4
184 const int paren_elems_to_push = (PL_regsize - parenfloor) * REGCP_PAREN_ELEMS;
187 if (paren_elems_to_push < 0)
188 Perl_croak(aTHX_ "panic: paren_elems_to_push < 0");
190 #define REGCP_OTHER_ELEMS 6
191 SSGROW(paren_elems_to_push + REGCP_OTHER_ELEMS);
192 for (p = PL_regsize; p > parenfloor; p--) {
193 /* REGCP_PARENS_ELEMS are pushed per pairs of parentheses. */
194 SSPUSHINT(PL_regendp[p]);
195 SSPUSHINT(PL_regstartp[p]);
196 SSPUSHPTR(PL_reg_start_tmp[p]);
199 /* REGCP_OTHER_ELEMS are pushed in any case, parentheses or no. */
200 SSPUSHINT(PL_regsize);
201 SSPUSHINT(*PL_reglastparen);
202 SSPUSHINT(*PL_reglastcloseparen);
203 SSPUSHPTR(PL_reginput);
204 #define REGCP_FRAME_ELEMS 2
205 /* REGCP_FRAME_ELEMS are part of the REGCP_OTHER_ELEMS and
206 * are needed for the regexp context stack bookkeeping. */
207 SSPUSHINT(paren_elems_to_push + REGCP_OTHER_ELEMS - REGCP_FRAME_ELEMS);
208 SSPUSHINT(SAVEt_REGCONTEXT); /* Magic cookie. */
213 /* These are needed since we do not localize EVAL nodes: */
214 # define REGCP_SET(cp) DEBUG_EXECUTE_r(PerlIO_printf(Perl_debug_log, \
215 " Setting an EVAL scope, savestack=%"IVdf"\n", \
216 (IV)PL_savestack_ix)); cp = PL_savestack_ix
218 # define REGCP_UNWIND(cp) DEBUG_EXECUTE_r(cp != PL_savestack_ix ? \
219 PerlIO_printf(Perl_debug_log, \
220 " Clearing an EVAL scope, savestack=%"IVdf"..%"IVdf"\n", \
221 (IV)(cp), (IV)PL_savestack_ix) : 0); regcpblow(cp)
231 GET_RE_DEBUG_FLAGS_DECL;
233 /* Pop REGCP_OTHER_ELEMS before the parentheses loop starts. */
235 assert(i == SAVEt_REGCONTEXT); /* Check that the magic cookie is there. */
236 i = SSPOPINT; /* Parentheses elements to pop. */
237 input = (char *) SSPOPPTR;
238 *PL_reglastcloseparen = SSPOPINT;
239 *PL_reglastparen = SSPOPINT;
240 PL_regsize = SSPOPINT;
242 /* Now restore the parentheses context. */
243 for (i -= (REGCP_OTHER_ELEMS - REGCP_FRAME_ELEMS);
244 i > 0; i -= REGCP_PAREN_ELEMS) {
246 paren = (U32)SSPOPINT;
247 PL_reg_start_tmp[paren] = (char *) SSPOPPTR;
248 PL_regstartp[paren] = SSPOPINT;
250 if (paren <= *PL_reglastparen)
251 PL_regendp[paren] = tmps;
253 PerlIO_printf(Perl_debug_log,
254 " restoring \\%"UVuf" to %"IVdf"(%"IVdf")..%"IVdf"%s\n",
255 (UV)paren, (IV)PL_regstartp[paren],
256 (IV)(PL_reg_start_tmp[paren] - PL_bostr),
257 (IV)PL_regendp[paren],
258 (paren > *PL_reglastparen ? "(no)" : ""));
262 if ((I32)(*PL_reglastparen + 1) <= PL_regnpar) {
263 PerlIO_printf(Perl_debug_log,
264 " restoring \\%"IVdf"..\\%"IVdf" to undef\n",
265 (IV)(*PL_reglastparen + 1), (IV)PL_regnpar);
269 /* It would seem that the similar code in regtry()
270 * already takes care of this, and in fact it is in
271 * a better location to since this code can #if 0-ed out
272 * but the code in regtry() is needed or otherwise tests
273 * requiring null fields (pat.t#187 and split.t#{13,14}
274 * (as of patchlevel 7877) will fail. Then again,
275 * this code seems to be necessary or otherwise
276 * building DynaLoader will fail:
277 * "Error: '*' not in typemap in DynaLoader.xs, line 164"
279 for (paren = *PL_reglastparen + 1; (I32)paren <= PL_regnpar; paren++) {
280 if ((I32)paren > PL_regsize)
281 PL_regstartp[paren] = -1;
282 PL_regendp[paren] = -1;
288 typedef struct re_cc_state
292 struct re_cc_state *prev;
297 #define regcpblow(cp) LEAVE_SCOPE(cp) /* Ignores regcppush()ed data. */
299 #define TRYPAREN(paren, n, input) { \
302 PL_regstartp[paren] = HOPc(input, -1) - PL_bostr; \
303 PL_regendp[paren] = input - PL_bostr; \
306 PL_regendp[paren] = -1; \
308 if (regmatch(next)) \
311 PL_regendp[paren] = -1; \
316 * pregexec and friends
320 - pregexec - match a regexp against a string
323 Perl_pregexec(pTHX_ register regexp *prog, char *stringarg, register char *strend,
324 char *strbeg, I32 minend, SV *screamer, U32 nosave)
325 /* strend: pointer to null at end of string */
326 /* strbeg: real beginning of string */
327 /* minend: end of match must be >=minend after stringarg. */
328 /* nosave: For optimizations. */
331 regexec_flags(prog, stringarg, strend, strbeg, minend, screamer, NULL,
332 nosave ? 0 : REXEC_COPY_STR);
336 S_cache_re(pTHX_ regexp *prog)
339 PL_regprecomp = prog->precomp; /* Needed for FAIL. */
341 PL_regprogram = prog->program;
343 PL_regnpar = prog->nparens;
344 PL_regdata = prog->data;
349 * Need to implement the following flags for reg_anch:
351 * USE_INTUIT_NOML - Useful to call re_intuit_start() first
353 * INTUIT_AUTORITATIVE_NOML - Can trust a positive answer
354 * INTUIT_AUTORITATIVE_ML
355 * INTUIT_ONCE_NOML - Intuit can match in one location only.
358 * Another flag for this function: SECOND_TIME (so that float substrs
359 * with giant delta may be not rechecked).
362 /* Assumptions: if ANCH_GPOS, then strpos is anchored. XXXX Check GPOS logic */
364 /* If SCREAM, then SvPVX_const(sv) should be compatible with strpos and strend.
365 Otherwise, only SvCUR(sv) is used to get strbeg. */
367 /* XXXX We assume that strpos is strbeg unless sv. */
369 /* XXXX Some places assume that there is a fixed substring.
370 An update may be needed if optimizer marks as "INTUITable"
371 RExen without fixed substrings. Similarly, it is assumed that
372 lengths of all the strings are no more than minlen, thus they
373 cannot come from lookahead.
374 (Or minlen should take into account lookahead.) */
376 /* A failure to find a constant substring means that there is no need to make
377 an expensive call to REx engine, thus we celebrate a failure. Similarly,
378 finding a substring too deep into the string means that less calls to
379 regtry() should be needed.
381 REx compiler's optimizer found 4 possible hints:
382 a) Anchored substring;
384 c) Whether we are anchored (beginning-of-line or \G);
385 d) First node (of those at offset 0) which may distingush positions;
386 We use a)b)d) and multiline-part of c), and try to find a position in the
387 string which does not contradict any of them.
390 /* Most of decisions we do here should have been done at compile time.
391 The nodes of the REx which we used for the search should have been
392 deleted from the finite automaton. */
395 Perl_re_intuit_start(pTHX_ regexp *prog, SV *sv, char *strpos,
396 char *strend, U32 flags, re_scream_pos_data *data)
399 register I32 start_shift = 0;
400 /* Should be nonnegative! */
401 register I32 end_shift = 0;
406 const int do_utf8 = sv ? SvUTF8(sv) : 0; /* if no sv we have to assume bytes */
408 register char *other_last = NULL; /* other substr checked before this */
409 char *check_at = NULL; /* check substr found at this pos */
410 const I32 multiline = prog->reganch & PMf_MULTILINE;
412 const char * const i_strpos = strpos;
413 SV * const dsv = PERL_DEBUG_PAD_ZERO(0);
416 GET_RE_DEBUG_FLAGS_DECL;
418 RX_MATCH_UTF8_set(prog,do_utf8);
420 if (prog->reganch & ROPT_UTF8) {
421 DEBUG_EXECUTE_r(PerlIO_printf(Perl_debug_log,
422 "UTF-8 regex...\n"));
423 PL_reg_flags |= RF_utf8;
427 const char *s = PL_reg_match_utf8 ?
428 sv_uni_display(dsv, sv, 60, UNI_DISPLAY_REGEX) :
430 const int len = PL_reg_match_utf8 ?
431 strlen(s) : strend - strpos;
434 if (PL_reg_match_utf8)
435 DEBUG_EXECUTE_r(PerlIO_printf(Perl_debug_log,
436 "UTF-8 target...\n"));
437 PerlIO_printf(Perl_debug_log,
438 "%sGuessing start of match, REx%s \"%s%.60s%s%s\" against \"%s%.*s%s%s\"...\n",
439 PL_colors[4], PL_colors[5], PL_colors[0],
442 (strlen(prog->precomp) > 60 ? "..." : ""),
444 (int)(len > 60 ? 60 : len),
446 (len > 60 ? "..." : "")
450 /* CHR_DIST() would be more correct here but it makes things slow. */
451 if (prog->minlen > strend - strpos) {
452 DEBUG_EXECUTE_r(PerlIO_printf(Perl_debug_log,
453 "String too short... [re_intuit_start]\n"));
456 strbeg = (sv && SvPOK(sv)) ? strend - SvCUR(sv) : strpos;
459 if (!prog->check_utf8 && prog->check_substr)
460 to_utf8_substr(prog);
461 check = prog->check_utf8;
463 if (!prog->check_substr && prog->check_utf8)
464 to_byte_substr(prog);
465 check = prog->check_substr;
467 if (check == &PL_sv_undef) {
468 DEBUG_EXECUTE_r(PerlIO_printf(Perl_debug_log,
469 "Non-utf string cannot match utf check string\n"));
472 if (prog->reganch & ROPT_ANCH) { /* Match at beg-of-str or after \n */
473 ml_anch = !( (prog->reganch & ROPT_ANCH_SINGLE)
474 || ( (prog->reganch & ROPT_ANCH_BOL)
475 && !multiline ) ); /* Check after \n? */
478 if ( !(prog->reganch & (ROPT_ANCH_GPOS /* Checked by the caller */
479 | ROPT_IMPLICIT)) /* not a real BOL */
480 /* SvCUR is not set on references: SvRV and SvPVX_const overlap */
482 && (strpos != strbeg)) {
483 DEBUG_EXECUTE_r(PerlIO_printf(Perl_debug_log, "Not at start...\n"));
486 if (prog->check_offset_min == prog->check_offset_max &&
487 !(prog->reganch & ROPT_CANY_SEEN)) {
488 /* Substring at constant offset from beg-of-str... */
491 s = HOP3c(strpos, prog->check_offset_min, strend);
493 slen = SvCUR(check); /* >= 1 */
495 if ( strend - s > slen || strend - s < slen - 1
496 || (strend - s == slen && strend[-1] != '\n')) {
497 DEBUG_EXECUTE_r(PerlIO_printf(Perl_debug_log, "String too long...\n"));
500 /* Now should match s[0..slen-2] */
502 if (slen && (*SvPVX_const(check) != *s
504 && memNE(SvPVX_const(check), s, slen)))) {
506 DEBUG_EXECUTE_r(PerlIO_printf(Perl_debug_log, "String not equal...\n"));
510 else if (*SvPVX_const(check) != *s
511 || ((slen = SvCUR(check)) > 1
512 && memNE(SvPVX_const(check), s, slen)))
515 goto success_at_start;
518 /* Match is anchored, but substr is not anchored wrt beg-of-str. */
520 start_shift = prog->check_offset_min; /* okay to underestimate on CC */
521 end_shift = prog->minlen - start_shift -
522 CHR_SVLEN(check) + (SvTAIL(check) != 0);
524 const I32 end = prog->check_offset_max + CHR_SVLEN(check)
525 - (SvTAIL(check) != 0);
526 const I32 eshift = CHR_DIST((U8*)strend, (U8*)s) - end;
528 if (end_shift < eshift)
532 else { /* Can match at random position */
535 start_shift = prog->check_offset_min; /* okay to underestimate on CC */
536 /* Should be nonnegative! */
537 end_shift = prog->minlen - start_shift -
538 CHR_SVLEN(check) + (SvTAIL(check) != 0);
541 #ifdef DEBUGGING /* 7/99: reports of failure (with the older version) */
543 Perl_croak(aTHX_ "panic: end_shift");
547 /* Find a possible match in the region s..strend by looking for
548 the "check" substring in the region corrected by start/end_shift. */
549 if (flags & REXEC_SCREAM) {
550 I32 p = -1; /* Internal iterator of scream. */
551 I32 * const pp = data ? data->scream_pos : &p;
553 if (PL_screamfirst[BmRARE(check)] >= 0
554 || ( BmRARE(check) == '\n'
555 && (BmPREVIOUS(check) == SvCUR(check) - 1)
557 s = screaminstr(sv, check,
558 start_shift + (s - strbeg), end_shift, pp, 0);
561 /* we may be pointing at the wrong string */
562 if (s && RX_MATCH_COPIED(prog))
563 s = strbeg + (s - SvPVX_const(sv));
565 *data->scream_olds = s;
567 else if (prog->reganch & ROPT_CANY_SEEN)
568 s = fbm_instr((U8*)(s + start_shift),
569 (U8*)(strend - end_shift),
570 check, multiline ? FBMrf_MULTILINE : 0);
572 s = fbm_instr(HOP3(s, start_shift, strend),
573 HOP3(strend, -end_shift, strbeg),
574 check, multiline ? FBMrf_MULTILINE : 0);
576 /* Update the count-of-usability, remove useless subpatterns,
579 DEBUG_EXECUTE_r(PerlIO_printf(Perl_debug_log, "%s %s substr \"%s%.*s%s\"%s%s",
580 (s ? "Found" : "Did not find"),
581 (check == (do_utf8 ? prog->anchored_utf8 : prog->anchored_substr) ? "anchored" : "floating"),
583 (int)(SvCUR(check) - (SvTAIL(check)!=0)),
585 PL_colors[1], (SvTAIL(check) ? "$" : ""),
586 (s ? " at offset " : "...\n") ) );
593 /* Finish the diagnostic message */
594 DEBUG_EXECUTE_r(PerlIO_printf(Perl_debug_log, "%ld...\n", (long)(s - i_strpos)) );
596 /* Got a candidate. Check MBOL anchoring, and the *other* substr.
597 Start with the other substr.
598 XXXX no SCREAM optimization yet - and a very coarse implementation
599 XXXX /ttx+/ results in anchored="ttx", floating="x". floating will
600 *always* match. Probably should be marked during compile...
601 Probably it is right to do no SCREAM here...
604 if (do_utf8 ? (prog->float_utf8 && prog->anchored_utf8) : (prog->float_substr && prog->anchored_substr)) {
605 /* Take into account the "other" substring. */
606 /* XXXX May be hopelessly wrong for UTF... */
609 if (check == (do_utf8 ? prog->float_utf8 : prog->float_substr)) {
612 char * const last = HOP3c(s, -start_shift, strbeg);
617 t = s - prog->check_offset_max;
618 if (s - strpos > prog->check_offset_max /* signed-corrected t > strpos */
620 || ((t = reghopmaybe3_c(s, -(prog->check_offset_max), strpos))
625 t = HOP3c(t, prog->anchored_offset, strend);
626 if (t < other_last) /* These positions already checked */
628 last2 = last1 = HOP3c(strend, -prog->minlen, strbeg);
631 /* XXXX It is not documented what units *_offsets are in. Assume bytes. */
632 /* On end-of-str: see comment below. */
633 must = do_utf8 ? prog->anchored_utf8 : prog->anchored_substr;
634 if (must == &PL_sv_undef) {
636 DEBUG_EXECUTE_r(must = prog->anchored_utf8); /* for debug */
641 HOP3(HOP3(last1, prog->anchored_offset, strend)
642 + SvCUR(must), -(SvTAIL(must)!=0), strbeg),
644 multiline ? FBMrf_MULTILINE : 0
646 DEBUG_EXECUTE_r(PerlIO_printf(Perl_debug_log,
647 "%s anchored substr \"%s%.*s%s\"%s",
648 (s ? "Found" : "Contradicts"),
651 - (SvTAIL(must)!=0)),
653 PL_colors[1], (SvTAIL(must) ? "$" : "")));
655 if (last1 >= last2) {
656 DEBUG_EXECUTE_r(PerlIO_printf(Perl_debug_log,
657 ", giving up...\n"));
660 DEBUG_EXECUTE_r(PerlIO_printf(Perl_debug_log,
661 ", trying floating at offset %ld...\n",
662 (long)(HOP3c(s1, 1, strend) - i_strpos)));
663 other_last = HOP3c(last1, prog->anchored_offset+1, strend);
664 s = HOP3c(last, 1, strend);
668 DEBUG_EXECUTE_r(PerlIO_printf(Perl_debug_log, " at offset %ld...\n",
669 (long)(s - i_strpos)));
670 t = HOP3c(s, -prog->anchored_offset, strbeg);
671 other_last = HOP3c(s, 1, strend);
679 else { /* Take into account the floating substring. */
684 t = HOP3c(s, -start_shift, strbeg);
686 HOP3c(strend, -prog->minlen + prog->float_min_offset, strbeg);
687 if (CHR_DIST((U8*)last, (U8*)t) > prog->float_max_offset)
688 last = HOP3c(t, prog->float_max_offset, strend);
689 s = HOP3c(t, prog->float_min_offset, strend);
692 /* XXXX It is not documented what units *_offsets are in. Assume bytes. */
693 must = do_utf8 ? prog->float_utf8 : prog->float_substr;
694 /* fbm_instr() takes into account exact value of end-of-str
695 if the check is SvTAIL(ed). Since false positives are OK,
696 and end-of-str is not later than strend we are OK. */
697 if (must == &PL_sv_undef) {
699 DEBUG_EXECUTE_r(must = prog->float_utf8); /* for debug message */
702 s = fbm_instr((unsigned char*)s,
703 (unsigned char*)last + SvCUR(must)
705 must, multiline ? FBMrf_MULTILINE : 0);
706 DEBUG_EXECUTE_r(PerlIO_printf(Perl_debug_log, "%s floating substr \"%s%.*s%s\"%s",
707 (s ? "Found" : "Contradicts"),
709 (int)(SvCUR(must) - (SvTAIL(must)!=0)),
711 PL_colors[1], (SvTAIL(must) ? "$" : "")));
714 DEBUG_EXECUTE_r(PerlIO_printf(Perl_debug_log,
715 ", giving up...\n"));
718 DEBUG_EXECUTE_r(PerlIO_printf(Perl_debug_log,
719 ", trying anchored starting at offset %ld...\n",
720 (long)(s1 + 1 - i_strpos)));
722 s = HOP3c(t, 1, strend);
726 DEBUG_EXECUTE_r(PerlIO_printf(Perl_debug_log, " at offset %ld...\n",
727 (long)(s - i_strpos)));
728 other_last = s; /* Fix this later. --Hugo */
737 t = s - prog->check_offset_max;
738 if (s - strpos > prog->check_offset_max /* signed-corrected t > strpos */
740 || ((t = reghopmaybe3_c(s, -prog->check_offset_max, strpos))
742 /* Fixed substring is found far enough so that the match
743 cannot start at strpos. */
745 if (ml_anch && t[-1] != '\n') {
746 /* Eventually fbm_*() should handle this, but often
747 anchored_offset is not 0, so this check will not be wasted. */
748 /* XXXX In the code below we prefer to look for "^" even in
749 presence of anchored substrings. And we search even
750 beyond the found float position. These pessimizations
751 are historical artefacts only. */
753 while (t < strend - prog->minlen) {
755 if (t < check_at - prog->check_offset_min) {
756 if (do_utf8 ? prog->anchored_utf8 : prog->anchored_substr) {
757 /* Since we moved from the found position,
758 we definitely contradict the found anchored
759 substr. Due to the above check we do not
760 contradict "check" substr.
761 Thus we can arrive here only if check substr
762 is float. Redo checking for "other"=="fixed".
765 DEBUG_EXECUTE_r(PerlIO_printf(Perl_debug_log, "Found /%s^%s/m at offset %ld, rescanning for anchored from offset %ld...\n",
766 PL_colors[0], PL_colors[1], (long)(strpos - i_strpos), (long)(strpos - i_strpos + prog->anchored_offset)));
767 goto do_other_anchored;
769 /* We don't contradict the found floating substring. */
770 /* XXXX Why not check for STCLASS? */
772 DEBUG_EXECUTE_r(PerlIO_printf(Perl_debug_log, "Found /%s^%s/m at offset %ld...\n",
773 PL_colors[0], PL_colors[1], (long)(s - i_strpos)));
776 /* Position contradicts check-string */
777 /* XXXX probably better to look for check-string
778 than for "\n", so one should lower the limit for t? */
779 DEBUG_EXECUTE_r(PerlIO_printf(Perl_debug_log, "Found /%s^%s/m, restarting lookup for check-string at offset %ld...\n",
780 PL_colors[0], PL_colors[1], (long)(t + 1 - i_strpos)));
781 other_last = strpos = s = t + 1;
786 DEBUG_EXECUTE_r(PerlIO_printf(Perl_debug_log, "Did not find /%s^%s/m...\n",
787 PL_colors[0], PL_colors[1]));
791 DEBUG_EXECUTE_r(PerlIO_printf(Perl_debug_log, "Starting position does not contradict /%s^%s/m...\n",
792 PL_colors[0], PL_colors[1]));
796 ++BmUSEFUL(do_utf8 ? prog->check_utf8 : prog->check_substr); /* hooray/5 */
799 /* The found string does not prohibit matching at strpos,
800 - no optimization of calling REx engine can be performed,
801 unless it was an MBOL and we are not after MBOL,
802 or a future STCLASS check will fail this. */
804 /* Even in this situation we may use MBOL flag if strpos is offset
805 wrt the start of the string. */
806 if (ml_anch && sv && !SvROK(sv) /* See prev comment on SvROK */
807 && (strpos != strbeg) && strpos[-1] != '\n'
808 /* May be due to an implicit anchor of m{.*foo} */
809 && !(prog->reganch & ROPT_IMPLICIT))
814 DEBUG_EXECUTE_r( if (ml_anch)
815 PerlIO_printf(Perl_debug_log, "Position at offset %ld does not contradict /%s^%s/m...\n",
816 (long)(strpos - i_strpos), PL_colors[0], PL_colors[1]);
819 if (!(prog->reganch & ROPT_NAUGHTY) /* XXXX If strpos moved? */
821 prog->check_utf8 /* Could be deleted already */
822 && --BmUSEFUL(prog->check_utf8) < 0
823 && (prog->check_utf8 == prog->float_utf8)
825 prog->check_substr /* Could be deleted already */
826 && --BmUSEFUL(prog->check_substr) < 0
827 && (prog->check_substr == prog->float_substr)
830 /* If flags & SOMETHING - do not do it many times on the same match */
831 DEBUG_EXECUTE_r(PerlIO_printf(Perl_debug_log, "... Disabling check substring...\n"));
832 SvREFCNT_dec(do_utf8 ? prog->check_utf8 : prog->check_substr);
833 if (do_utf8 ? prog->check_substr : prog->check_utf8)
834 SvREFCNT_dec(do_utf8 ? prog->check_substr : prog->check_utf8);
835 prog->check_substr = prog->check_utf8 = NULL; /* disable */
836 prog->float_substr = prog->float_utf8 = NULL; /* clear */
837 check = NULL; /* abort */
839 /* XXXX This is a remnant of the old implementation. It
840 looks wasteful, since now INTUIT can use many
842 prog->reganch &= ~RE_USE_INTUIT;
849 /* XXXX BmUSEFUL already changed, maybe multiple change is meaningful... */
850 if (prog->regstclass) {
851 /* minlen == 0 is possible if regstclass is \b or \B,
852 and the fixed substr is ''$.
853 Since minlen is already taken into account, s+1 is before strend;
854 accidentally, minlen >= 1 guaranties no false positives at s + 1
855 even for \b or \B. But (minlen? 1 : 0) below assumes that
856 regstclass does not come from lookahead... */
857 /* If regstclass takes bytelength more than 1: If charlength==1, OK.
858 This leaves EXACTF only, which is dealt with in find_byclass(). */
859 const U8* const str = (U8*)STRING(prog->regstclass);
860 const int cl_l = (PL_regkind[(U8)OP(prog->regstclass)] == EXACT
861 ? CHR_DIST(str+STR_LEN(prog->regstclass), str)
863 const char * const endpos = (prog->anchored_substr || prog->anchored_utf8 || ml_anch)
864 ? HOP3c(s, (prog->minlen ? cl_l : 0), strend)
865 : (prog->float_substr || prog->float_utf8
866 ? HOP3c(HOP3c(check_at, -start_shift, strbeg),
872 s = find_byclass(prog, prog->regstclass, s, endpos, 1);
875 const char *what = NULL;
877 if (endpos == strend) {
878 DEBUG_EXECUTE_r( PerlIO_printf(Perl_debug_log,
879 "Could not match STCLASS...\n") );
882 DEBUG_EXECUTE_r( PerlIO_printf(Perl_debug_log,
883 "This position contradicts STCLASS...\n") );
884 if ((prog->reganch & ROPT_ANCH) && !ml_anch)
886 /* Contradict one of substrings */
887 if (prog->anchored_substr || prog->anchored_utf8) {
888 if ((do_utf8 ? prog->anchored_utf8 : prog->anchored_substr) == check) {
889 DEBUG_EXECUTE_r( what = "anchored" );
891 s = HOP3c(t, 1, strend);
892 if (s + start_shift + end_shift > strend) {
893 /* XXXX Should be taken into account earlier? */
894 DEBUG_EXECUTE_r( PerlIO_printf(Perl_debug_log,
895 "Could not match STCLASS...\n") );
900 DEBUG_EXECUTE_r( PerlIO_printf(Perl_debug_log,
901 "Looking for %s substr starting at offset %ld...\n",
902 what, (long)(s + start_shift - i_strpos)) );
905 /* Have both, check_string is floating */
906 if (t + start_shift >= check_at) /* Contradicts floating=check */
907 goto retry_floating_check;
908 /* Recheck anchored substring, but not floating... */
912 DEBUG_EXECUTE_r( PerlIO_printf(Perl_debug_log,
913 "Looking for anchored substr starting at offset %ld...\n",
914 (long)(other_last - i_strpos)) );
915 goto do_other_anchored;
917 /* Another way we could have checked stclass at the
918 current position only: */
923 DEBUG_EXECUTE_r( PerlIO_printf(Perl_debug_log,
924 "Looking for /%s^%s/m starting at offset %ld...\n",
925 PL_colors[0], PL_colors[1], (long)(t - i_strpos)) );
928 if (!(do_utf8 ? prog->float_utf8 : prog->float_substr)) /* Could have been deleted */
930 /* Check is floating subtring. */
931 retry_floating_check:
932 t = check_at - start_shift;
933 DEBUG_EXECUTE_r( what = "floating" );
934 goto hop_and_restart;
937 DEBUG_EXECUTE_r(PerlIO_printf(Perl_debug_log,
938 "By STCLASS: moving %ld --> %ld\n",
939 (long)(t - i_strpos), (long)(s - i_strpos))
943 DEBUG_EXECUTE_r(PerlIO_printf(Perl_debug_log,
944 "Does not contradict STCLASS...\n");
949 DEBUG_EXECUTE_r(PerlIO_printf(Perl_debug_log, "%s%s:%s match at offset %ld\n",
950 PL_colors[4], (check ? "Guessed" : "Giving up"),
951 PL_colors[5], (long)(s - i_strpos)) );
954 fail_finish: /* Substring not found */
955 if (prog->check_substr || prog->check_utf8) /* could be removed already */
956 BmUSEFUL(do_utf8 ? prog->check_utf8 : prog->check_substr) += 5; /* hooray */
958 DEBUG_EXECUTE_r(PerlIO_printf(Perl_debug_log, "%sMatch rejected by optimizer%s\n",
959 PL_colors[4], PL_colors[5]));
963 /* We know what class REx starts with. Try to find this position... */
965 S_find_byclass(pTHX_ regexp * prog, regnode *c, char *s, const char *strend, I32 norun)
968 const I32 doevery = (prog->reganch & ROPT_SKIP) == 0;
972 register STRLEN uskip;
976 register I32 tmp = 1; /* Scratch variable? */
977 register const bool do_utf8 = PL_reg_match_utf8;
979 /* We know what class it must start with. */
983 while (s + (uskip = UTF8SKIP(s)) <= strend) {
984 if ((ANYOF_FLAGS(c) & ANYOF_UNICODE) ||
985 !UTF8_IS_INVARIANT((U8)s[0]) ?
986 reginclass(c, (U8*)s, 0, do_utf8) :
987 REGINCLASS(c, (U8*)s)) {
988 if (tmp && (norun || regtry(prog, s)))
1002 if (REGINCLASS(c, (U8*)s) ||
1003 (ANYOF_FOLD_SHARP_S(c, s, strend) &&
1004 /* The assignment of 2 is intentional:
1005 * for the folded sharp s, the skip is 2. */
1006 (skip = SHARP_S_SKIP))) {
1007 if (tmp && (norun || regtry(prog, s)))
1019 while (s < strend) {
1020 if (tmp && (norun || regtry(prog, s)))
1029 ln = STR_LEN(c); /* length to match in octets/bytes */
1030 lnc = (I32) ln; /* length to match in characters */
1032 STRLEN ulen1, ulen2;
1034 U8 tmpbuf1[UTF8_MAXBYTES_CASE+1];
1035 U8 tmpbuf2[UTF8_MAXBYTES_CASE+1];
1036 const U32 uniflags = ckWARN(WARN_UTF8) ? 0 : UTF8_ALLOW_ANY;
1038 to_utf8_lower((U8*)m, tmpbuf1, &ulen1);
1039 to_utf8_upper((U8*)m, tmpbuf2, &ulen2);
1041 c1 = utf8n_to_uvchr(tmpbuf1, UTF8_MAXBYTES_CASE,
1043 c2 = utf8n_to_uvchr(tmpbuf2, UTF8_MAXBYTES_CASE,
1046 while (sm < ((U8 *) m + ln)) {
1061 c2 = PL_fold_locale[c1];
1063 e = HOP3c(strend, -((I32)lnc), s);
1066 e = s; /* Due to minlen logic of intuit() */
1068 /* The idea in the EXACTF* cases is to first find the
1069 * first character of the EXACTF* node and then, if
1070 * necessary, case-insensitively compare the full
1071 * text of the node. The c1 and c2 are the first
1072 * characters (though in Unicode it gets a bit
1073 * more complicated because there are more cases
1074 * than just upper and lower: one needs to use
1075 * the so-called folding case for case-insensitive
1076 * matching (called "loose matching" in Unicode).
1077 * ibcmp_utf8() will do just that. */
1081 U8 tmpbuf [UTF8_MAXBYTES+1];
1082 STRLEN len, foldlen;
1083 const U32 uniflags = ckWARN(WARN_UTF8) ? 0 : UTF8_ALLOW_ANY;
1085 /* Upper and lower of 1st char are equal -
1086 * probably not a "letter". */
1088 c = utf8n_to_uvchr((U8*)s, UTF8_MAXBYTES, &len,
1092 ibcmp_utf8(s, (char **)0, 0, do_utf8,
1093 m, (char **)0, ln, (bool)UTF))
1094 && (norun || regtry(prog, s)) )
1097 U8 foldbuf[UTF8_MAXBYTES_CASE+1];
1098 uvchr_to_utf8(tmpbuf, c);
1099 f = to_utf8_fold(tmpbuf, foldbuf, &foldlen);
1101 && (f == c1 || f == c2)
1102 && (ln == foldlen ||
1103 !ibcmp_utf8((char *) foldbuf,
1104 (char **)0, foldlen, do_utf8,
1106 (char **)0, ln, (bool)UTF))
1107 && (norun || regtry(prog, s)) )
1115 c = utf8n_to_uvchr((U8*)s, UTF8_MAXBYTES, &len,
1118 /* Handle some of the three Greek sigmas cases.
1119 * Note that not all the possible combinations
1120 * are handled here: some of them are handled
1121 * by the standard folding rules, and some of
1122 * them (the character class or ANYOF cases)
1123 * are handled during compiletime in
1124 * regexec.c:S_regclass(). */
1125 if (c == (UV)UNICODE_GREEK_CAPITAL_LETTER_SIGMA ||
1126 c == (UV)UNICODE_GREEK_SMALL_LETTER_FINAL_SIGMA)
1127 c = (UV)UNICODE_GREEK_SMALL_LETTER_SIGMA;
1129 if ( (c == c1 || c == c2)
1131 ibcmp_utf8(s, (char **)0, 0, do_utf8,
1132 m, (char **)0, ln, (bool)UTF))
1133 && (norun || regtry(prog, s)) )
1136 U8 foldbuf[UTF8_MAXBYTES_CASE+1];
1137 uvchr_to_utf8(tmpbuf, c);
1138 f = to_utf8_fold(tmpbuf, foldbuf, &foldlen);
1140 && (f == c1 || f == c2)
1141 && (ln == foldlen ||
1142 !ibcmp_utf8((char *) foldbuf,
1143 (char **)0, foldlen, do_utf8,
1145 (char **)0, ln, (bool)UTF))
1146 && (norun || regtry(prog, s)) )
1157 && (ln == 1 || !(OP(c) == EXACTF
1159 : ibcmp_locale(s, m, ln)))
1160 && (norun || regtry(prog, s)) )
1166 if ( (*(U8*)s == c1 || *(U8*)s == c2)
1167 && (ln == 1 || !(OP(c) == EXACTF
1169 : ibcmp_locale(s, m, ln)))
1170 && (norun || regtry(prog, s)) )
1177 PL_reg_flags |= RF_tainted;
1184 U8 * const r = reghop3((U8*)s, -1, (U8*)PL_bostr);
1185 tmp = utf8n_to_uvchr(r, UTF8SKIP(r), 0, 0);
1187 tmp = ((OP(c) == BOUND ?
1188 isALNUM_uni(tmp) : isALNUM_LC_uvchr(UNI_TO_NATIVE(tmp))) != 0);
1189 LOAD_UTF8_CHARCLASS_ALNUM();
1190 while (s + (uskip = UTF8SKIP(s)) <= strend) {
1191 if (tmp == !(OP(c) == BOUND ?
1192 swash_fetch(PL_utf8_alnum, (U8*)s, do_utf8) :
1193 isALNUM_LC_utf8((U8*)s)))
1196 if ((norun || regtry(prog, s)))
1203 tmp = (s != PL_bostr) ? UCHARAT(s - 1) : '\n';
1204 tmp = ((OP(c) == BOUND ? isALNUM(tmp) : isALNUM_LC(tmp)) != 0);
1205 while (s < strend) {
1207 !(OP(c) == BOUND ? isALNUM(*s) : isALNUM_LC(*s))) {
1209 if ((norun || regtry(prog, s)))
1215 if ((!prog->minlen && tmp) && (norun || regtry(prog, s)))
1219 PL_reg_flags |= RF_tainted;
1226 U8 * const r = reghop3((U8*)s, -1, (U8*)PL_bostr);
1227 tmp = utf8n_to_uvchr(r, UTF8SKIP(r), 0, 0);
1229 tmp = ((OP(c) == NBOUND ?
1230 isALNUM_uni(tmp) : isALNUM_LC_uvchr(UNI_TO_NATIVE(tmp))) != 0);
1231 LOAD_UTF8_CHARCLASS_ALNUM();
1232 while (s + (uskip = UTF8SKIP(s)) <= strend) {
1233 if (tmp == !(OP(c) == NBOUND ?
1234 swash_fetch(PL_utf8_alnum, (U8*)s, do_utf8) :
1235 isALNUM_LC_utf8((U8*)s)))
1237 else if ((norun || regtry(prog, s)))
1243 tmp = (s != PL_bostr) ? UCHARAT(s - 1) : '\n';
1244 tmp = ((OP(c) == NBOUND ?
1245 isALNUM(tmp) : isALNUM_LC(tmp)) != 0);
1246 while (s < strend) {
1248 !(OP(c) == NBOUND ? isALNUM(*s) : isALNUM_LC(*s)))
1250 else if ((norun || regtry(prog, s)))
1255 if ((!prog->minlen && !tmp) && (norun || regtry(prog, s)))
1260 LOAD_UTF8_CHARCLASS_ALNUM();
1261 while (s + (uskip = UTF8SKIP(s)) <= strend) {
1262 if (swash_fetch(PL_utf8_alnum, (U8*)s, do_utf8)) {
1263 if (tmp && (norun || regtry(prog, s)))
1274 while (s < strend) {
1276 if (tmp && (norun || regtry(prog, s)))
1288 PL_reg_flags |= RF_tainted;
1290 while (s + (uskip = UTF8SKIP(s)) <= strend) {
1291 if (isALNUM_LC_utf8((U8*)s)) {
1292 if (tmp && (norun || regtry(prog, s)))
1303 while (s < strend) {
1304 if (isALNUM_LC(*s)) {
1305 if (tmp && (norun || regtry(prog, s)))
1318 LOAD_UTF8_CHARCLASS_ALNUM();
1319 while (s + (uskip = UTF8SKIP(s)) <= strend) {
1320 if (!swash_fetch(PL_utf8_alnum, (U8*)s, do_utf8)) {
1321 if (tmp && (norun || regtry(prog, s)))
1332 while (s < strend) {
1334 if (tmp && (norun || regtry(prog, s)))
1346 PL_reg_flags |= RF_tainted;
1348 while (s + (uskip = UTF8SKIP(s)) <= strend) {
1349 if (!isALNUM_LC_utf8((U8*)s)) {
1350 if (tmp && (norun || regtry(prog, s)))
1361 while (s < strend) {
1362 if (!isALNUM_LC(*s)) {
1363 if (tmp && (norun || regtry(prog, s)))
1376 LOAD_UTF8_CHARCLASS_SPACE();
1377 while (s + (uskip = UTF8SKIP(s)) <= strend) {
1378 if (*s == ' ' || swash_fetch(PL_utf8_space,(U8*)s, do_utf8)) {
1379 if (tmp && (norun || regtry(prog, s)))
1390 while (s < strend) {
1392 if (tmp && (norun || regtry(prog, s)))
1404 PL_reg_flags |= RF_tainted;
1406 while (s + (uskip = UTF8SKIP(s)) <= strend) {
1407 if (*s == ' ' || isSPACE_LC_utf8((U8*)s)) {
1408 if (tmp && (norun || regtry(prog, s)))
1419 while (s < strend) {
1420 if (isSPACE_LC(*s)) {
1421 if (tmp && (norun || regtry(prog, s)))
1434 LOAD_UTF8_CHARCLASS_SPACE();
1435 while (s + (uskip = UTF8SKIP(s)) <= strend) {
1436 if (!(*s == ' ' || swash_fetch(PL_utf8_space,(U8*)s, do_utf8))) {
1437 if (tmp && (norun || regtry(prog, s)))
1448 while (s < strend) {
1450 if (tmp && (norun || regtry(prog, s)))
1462 PL_reg_flags |= RF_tainted;
1464 while (s + (uskip = UTF8SKIP(s)) <= strend) {
1465 if (!(*s == ' ' || isSPACE_LC_utf8((U8*)s))) {
1466 if (tmp && (norun || regtry(prog, s)))
1477 while (s < strend) {
1478 if (!isSPACE_LC(*s)) {
1479 if (tmp && (norun || regtry(prog, s)))
1492 LOAD_UTF8_CHARCLASS_DIGIT();
1493 while (s + (uskip = UTF8SKIP(s)) <= strend) {
1494 if (swash_fetch(PL_utf8_digit,(U8*)s, do_utf8)) {
1495 if (tmp && (norun || regtry(prog, s)))
1506 while (s < strend) {
1508 if (tmp && (norun || regtry(prog, s)))
1520 PL_reg_flags |= RF_tainted;
1522 while (s + (uskip = UTF8SKIP(s)) <= strend) {
1523 if (isDIGIT_LC_utf8((U8*)s)) {
1524 if (tmp && (norun || regtry(prog, s)))
1535 while (s < strend) {
1536 if (isDIGIT_LC(*s)) {
1537 if (tmp && (norun || regtry(prog, s)))
1550 LOAD_UTF8_CHARCLASS_DIGIT();
1551 while (s + (uskip = UTF8SKIP(s)) <= strend) {
1552 if (!swash_fetch(PL_utf8_digit,(U8*)s, do_utf8)) {
1553 if (tmp && (norun || regtry(prog, s)))
1564 while (s < strend) {
1566 if (tmp && (norun || regtry(prog, s)))
1578 PL_reg_flags |= RF_tainted;
1580 while (s + (uskip = UTF8SKIP(s)) <= strend) {
1581 if (!isDIGIT_LC_utf8((U8*)s)) {
1582 if (tmp && (norun || regtry(prog, s)))
1593 while (s < strend) {
1594 if (!isDIGIT_LC(*s)) {
1595 if (tmp && (norun || regtry(prog, s)))
1607 Perl_croak(aTHX_ "panic: unknown regstclass %d", (int)OP(c));
1616 - regexec_flags - match a regexp against a string
1619 Perl_regexec_flags(pTHX_ register regexp *prog, char *stringarg, register char *strend,
1620 char *strbeg, I32 minend, SV *sv, void *data, U32 flags)
1621 /* strend: pointer to null at end of string */
1622 /* strbeg: real beginning of string */
1623 /* minend: end of match must be >=minend after stringarg. */
1624 /* data: May be used for some additional optimizations. */
1625 /* nosave: For optimizations. */
1629 register regnode *c;
1630 register char *startpos = stringarg;
1631 I32 minlen; /* must match at least this many chars */
1632 I32 dontbother = 0; /* how many characters not to try at end */
1633 I32 end_shift = 0; /* Same for the end. */ /* CC */
1634 I32 scream_pos = -1; /* Internal iterator of scream. */
1635 char *scream_olds = NULL;
1636 SV* oreplsv = GvSV(PL_replgv);
1637 const bool do_utf8 = DO_UTF8(sv);
1638 const I32 multiline = prog->reganch & PMf_MULTILINE;
1640 SV * const dsv0 = PERL_DEBUG_PAD_ZERO(0);
1641 SV * const dsv1 = PERL_DEBUG_PAD_ZERO(1);
1644 GET_RE_DEBUG_FLAGS_DECL;
1646 PERL_UNUSED_ARG(data);
1647 RX_MATCH_UTF8_set(prog,do_utf8);
1653 PL_regnarrate = DEBUG_r_TEST;
1656 /* Be paranoid... */
1657 if (prog == NULL || startpos == NULL) {
1658 Perl_croak(aTHX_ "NULL regexp parameter");
1662 minlen = prog->minlen;
1663 if (strend - startpos < minlen) {
1664 DEBUG_EXECUTE_r(PerlIO_printf(Perl_debug_log,
1665 "String too short [regexec_flags]...\n"));
1669 /* Check validity of program. */
1670 if (UCHARAT(prog->program) != REG_MAGIC) {
1671 Perl_croak(aTHX_ "corrupted regexp program");
1675 PL_reg_eval_set = 0;
1678 if (prog->reganch & ROPT_UTF8)
1679 PL_reg_flags |= RF_utf8;
1681 /* Mark beginning of line for ^ and lookbehind. */
1682 PL_regbol = startpos;
1686 /* Mark end of line for $ (and such) */
1689 /* see how far we have to get to not match where we matched before */
1690 PL_regtill = startpos+minend;
1692 /* We start without call_cc context. */
1695 /* If there is a "must appear" string, look for it. */
1698 if (prog->reganch & ROPT_GPOS_SEEN) { /* Need to have PL_reg_ganch */
1701 if (flags & REXEC_IGNOREPOS) /* Means: check only at start */
1702 PL_reg_ganch = startpos;
1703 else if (sv && SvTYPE(sv) >= SVt_PVMG
1705 && (mg = mg_find(sv, PERL_MAGIC_regex_global))
1706 && mg->mg_len >= 0) {
1707 PL_reg_ganch = strbeg + mg->mg_len; /* Defined pos() */
1708 if (prog->reganch & ROPT_ANCH_GPOS) {
1709 if (s > PL_reg_ganch)
1714 else /* pos() not defined */
1715 PL_reg_ganch = strbeg;
1718 if (!(flags & REXEC_CHECKED) && (prog->check_substr != NULL || prog->check_utf8 != NULL)) {
1719 re_scream_pos_data d;
1721 d.scream_olds = &scream_olds;
1722 d.scream_pos = &scream_pos;
1723 s = re_intuit_start(prog, sv, s, strend, flags, &d);
1725 DEBUG_EXECUTE_r(PerlIO_printf(Perl_debug_log, "Not present...\n"));
1726 goto phooey; /* not present */
1731 const char * const s0 = UTF
1732 ? pv_uni_display(dsv0, (U8*)prog->precomp, prog->prelen, 60,
1735 const int len0 = UTF ? SvCUR(dsv0) : prog->prelen;
1736 const char * const s1 = do_utf8 ? sv_uni_display(dsv1, sv, 60,
1737 UNI_DISPLAY_REGEX) : startpos;
1738 const int len1 = do_utf8 ? SvCUR(dsv1) : strend - startpos;
1741 PerlIO_printf(Perl_debug_log,
1742 "%sMatching REx%s \"%s%*.*s%s%s\" against \"%s%.*s%s%s\"\n",
1743 PL_colors[4], PL_colors[5], PL_colors[0],
1746 len0 > 60 ? "..." : "",
1748 (int)(len1 > 60 ? 60 : len1),
1750 (len1 > 60 ? "..." : "")
1754 /* Simplest case: anchored match need be tried only once. */
1755 /* [unless only anchor is BOL and multiline is set] */
1756 if (prog->reganch & (ROPT_ANCH & ~ROPT_ANCH_GPOS)) {
1757 if (s == startpos && regtry(prog, startpos))
1759 else if (multiline || (prog->reganch & ROPT_IMPLICIT)
1760 || (prog->reganch & ROPT_ANCH_MBOL)) /* XXXX SBOL? */
1765 dontbother = minlen - 1;
1766 end = HOP3c(strend, -dontbother, strbeg) - 1;
1767 /* for multiline we only have to try after newlines */
1768 if (prog->check_substr || prog->check_utf8) {
1772 if (regtry(prog, s))
1777 if (prog->reganch & RE_USE_INTUIT) {
1778 s = re_intuit_start(prog, sv, s + 1, strend, flags, NULL);
1789 if (*s++ == '\n') { /* don't need PL_utf8skip here */
1790 if (regtry(prog, s))
1797 } else if (prog->reganch & ROPT_ANCH_GPOS) {
1798 if (regtry(prog, PL_reg_ganch))
1803 /* Messy cases: unanchored match. */
1804 if ((prog->anchored_substr || prog->anchored_utf8) && prog->reganch & ROPT_SKIP) {
1805 /* we have /x+whatever/ */
1806 /* it must be a one character string (XXXX Except UTF?) */
1811 if (!(do_utf8 ? prog->anchored_utf8 : prog->anchored_substr))
1812 do_utf8 ? to_utf8_substr(prog) : to_byte_substr(prog);
1813 ch = SvPVX_const(do_utf8 ? prog->anchored_utf8 : prog->anchored_substr)[0];
1816 while (s < strend) {
1818 DEBUG_EXECUTE_r( did_match = 1 );
1819 if (regtry(prog, s)) goto got_it;
1821 while (s < strend && *s == ch)
1828 while (s < strend) {
1830 DEBUG_EXECUTE_r( did_match = 1 );
1831 if (regtry(prog, s)) goto got_it;
1833 while (s < strend && *s == ch)
1839 DEBUG_EXECUTE_r(if (!did_match)
1840 PerlIO_printf(Perl_debug_log,
1841 "Did not find anchored character...\n")
1844 else if (prog->anchored_substr != NULL
1845 || prog->anchored_utf8 != NULL
1846 || ((prog->float_substr != NULL || prog->float_utf8 != NULL)
1847 && prog->float_max_offset < strend - s)) {
1852 char *last1; /* Last position checked before */
1856 if (prog->anchored_substr || prog->anchored_utf8) {
1857 if (!(do_utf8 ? prog->anchored_utf8 : prog->anchored_substr))
1858 do_utf8 ? to_utf8_substr(prog) : to_byte_substr(prog);
1859 must = do_utf8 ? prog->anchored_utf8 : prog->anchored_substr;
1860 back_max = back_min = prog->anchored_offset;
1862 if (!(do_utf8 ? prog->float_utf8 : prog->float_substr))
1863 do_utf8 ? to_utf8_substr(prog) : to_byte_substr(prog);
1864 must = do_utf8 ? prog->float_utf8 : prog->float_substr;
1865 back_max = prog->float_max_offset;
1866 back_min = prog->float_min_offset;
1868 if (must == &PL_sv_undef)
1869 /* could not downgrade utf8 check substring, so must fail */
1872 last = HOP3c(strend, /* Cannot start after this */
1873 -(I32)(CHR_SVLEN(must)
1874 - (SvTAIL(must) != 0) + back_min), strbeg);
1877 last1 = HOPc(s, -1);
1879 last1 = s - 1; /* bogus */
1881 /* XXXX check_substr already used to find "s", can optimize if
1882 check_substr==must. */
1884 dontbother = end_shift;
1885 strend = HOPc(strend, -dontbother);
1886 while ( (s <= last) &&
1887 ((flags & REXEC_SCREAM)
1888 ? (s = screaminstr(sv, must, HOP3c(s, back_min, strend) - strbeg,
1889 end_shift, &scream_pos, 0))
1890 : (s = fbm_instr((unsigned char*)HOP3(s, back_min, strend),
1891 (unsigned char*)strend, must,
1892 multiline ? FBMrf_MULTILINE : 0))) ) {
1893 /* we may be pointing at the wrong string */
1894 if ((flags & REXEC_SCREAM) && RX_MATCH_COPIED(prog))
1895 s = strbeg + (s - SvPVX_const(sv));
1896 DEBUG_EXECUTE_r( did_match = 1 );
1897 if (HOPc(s, -back_max) > last1) {
1898 last1 = HOPc(s, -back_min);
1899 s = HOPc(s, -back_max);
1902 char *t = (last1 >= PL_bostr) ? HOPc(last1, 1) : last1 + 1;
1904 last1 = HOPc(s, -back_min);
1908 while (s <= last1) {
1909 if (regtry(prog, s))
1915 while (s <= last1) {
1916 if (regtry(prog, s))
1922 DEBUG_EXECUTE_r(if (!did_match)
1923 PerlIO_printf(Perl_debug_log,
1924 "Did not find %s substr \"%s%.*s%s\"%s...\n",
1925 ((must == prog->anchored_substr || must == prog->anchored_utf8)
1926 ? "anchored" : "floating"),
1928 (int)(SvCUR(must) - (SvTAIL(must)!=0)),
1930 PL_colors[1], (SvTAIL(must) ? "$" : ""))
1934 else if ((c = prog->regstclass)) {
1936 I32 op = (U8)OP(prog->regstclass);
1937 /* don't bother with what can't match */
1938 if (PL_regkind[op] != EXACT && op != CANY)
1939 strend = HOPc(strend, -(minlen - 1));
1942 SV *prop = sv_newmortal();
1950 pv_uni_display(dsv0, (U8*)SvPVX_const(prop), SvCUR(prop), 60,
1951 UNI_DISPLAY_REGEX) :
1953 len0 = UTF ? SvCUR(dsv0) : SvCUR(prop);
1955 sv_uni_display(dsv1, sv, 60, UNI_DISPLAY_REGEX) : s;
1956 len1 = UTF ? SvCUR(dsv1) : strend - s;
1957 PerlIO_printf(Perl_debug_log,
1958 "Matching stclass \"%*.*s\" against \"%*.*s\"\n",
1962 if (find_byclass(prog, c, s, strend, 0))
1964 DEBUG_EXECUTE_r(PerlIO_printf(Perl_debug_log, "Contradicts stclass...\n"));
1968 if (prog->float_substr != NULL || prog->float_utf8 != NULL) {
1973 if (!(do_utf8 ? prog->float_utf8 : prog->float_substr))
1974 do_utf8 ? to_utf8_substr(prog) : to_byte_substr(prog);
1975 float_real = do_utf8 ? prog->float_utf8 : prog->float_substr;
1977 if (flags & REXEC_SCREAM) {
1978 last = screaminstr(sv, float_real, s - strbeg,
1979 end_shift, &scream_pos, 1); /* last one */
1981 last = scream_olds; /* Only one occurrence. */
1982 /* we may be pointing at the wrong string */
1983 else if (RX_MATCH_COPIED(prog))
1984 s = strbeg + (s - SvPVX_const(sv));
1988 const char * const little = SvPV_const(float_real, len);
1990 if (SvTAIL(float_real)) {
1991 if (memEQ(strend - len + 1, little, len - 1))
1992 last = strend - len + 1;
1993 else if (!multiline)
1994 last = memEQ(strend - len, little, len)
1995 ? strend - len : NULL;
2001 last = rninstr(s, strend, little, little + len);
2003 last = strend; /* matching "$" */
2007 DEBUG_EXECUTE_r(PerlIO_printf(Perl_debug_log,
2008 "%sCan't trim the tail, match fails (should not happen)%s\n",
2009 PL_colors[4], PL_colors[5]));
2010 goto phooey; /* Should not happen! */
2012 dontbother = strend - last + prog->float_min_offset;
2014 if (minlen && (dontbother < minlen))
2015 dontbother = minlen - 1;
2016 strend -= dontbother; /* this one's always in bytes! */
2017 /* We don't know much -- general case. */
2020 if (regtry(prog, s))
2029 if (regtry(prog, s))
2031 } while (s++ < strend);
2039 RX_MATCH_TAINTED_set(prog, PL_reg_flags & RF_tainted);
2041 if (PL_reg_eval_set) {
2042 /* Preserve the current value of $^R */
2043 if (oreplsv != GvSV(PL_replgv))
2044 sv_setsv(oreplsv, GvSV(PL_replgv));/* So that when GvSV(replgv) is
2045 restored, the value remains
2047 restore_pos(aTHX_ 0);
2050 /* make sure $`, $&, $', and $digit will work later */
2051 if ( !(flags & REXEC_NOT_FIRST) ) {
2052 RX_MATCH_COPY_FREE(prog);
2053 if (flags & REXEC_COPY_STR) {
2054 I32 i = PL_regeol - startpos + (stringarg - strbeg);
2055 #ifdef PERL_OLD_COPY_ON_WRITE
2057 || (SvFLAGS(sv) & CAN_COW_MASK) == CAN_COW_FLAGS)) {
2059 PerlIO_printf(Perl_debug_log,
2060 "Copy on write: regexp capture, type %d\n",
2063 prog->saved_copy = sv_setsv_cow(prog->saved_copy, sv);
2064 prog->subbeg = (char *)SvPVX_const(prog->saved_copy);
2065 assert (SvPOKp(prog->saved_copy));
2069 RX_MATCH_COPIED_on(prog);
2070 s = savepvn(strbeg, i);
2076 prog->subbeg = strbeg;
2077 prog->sublen = PL_regeol - strbeg; /* strend may have been modified */
2084 DEBUG_EXECUTE_r(PerlIO_printf(Perl_debug_log, "%sMatch failed%s\n",
2085 PL_colors[4], PL_colors[5]));
2086 if (PL_reg_eval_set)
2087 restore_pos(aTHX_ 0);
2092 - regtry - try match at specific point
2094 STATIC I32 /* 0 failure, 1 success */
2095 S_regtry(pTHX_ regexp *prog, char *startpos)
2102 GET_RE_DEBUG_FLAGS_DECL;
2105 PL_regindent = 0; /* XXXX Not good when matches are reenterable... */
2107 if ((prog->reganch & ROPT_EVAL_SEEN) && !PL_reg_eval_set) {
2110 PL_reg_eval_set = RS_init;
2111 DEBUG_EXECUTE_r(DEBUG_s(
2112 PerlIO_printf(Perl_debug_log, " setting stack tmpbase at %"IVdf"\n",
2113 (IV)(PL_stack_sp - PL_stack_base));
2115 SAVEI32(cxstack[cxstack_ix].blk_oldsp);
2116 cxstack[cxstack_ix].blk_oldsp = PL_stack_sp - PL_stack_base;
2117 /* Otherwise OP_NEXTSTATE will free whatever on stack now. */
2119 /* Apparently this is not needed, judging by wantarray. */
2120 /* SAVEI8(cxstack[cxstack_ix].blk_gimme);
2121 cxstack[cxstack_ix].blk_gimme = G_SCALAR; */
2124 /* Make $_ available to executed code. */
2125 if (PL_reg_sv != DEFSV) {
2130 if (!(SvTYPE(PL_reg_sv) >= SVt_PVMG && SvMAGIC(PL_reg_sv)
2131 && (mg = mg_find(PL_reg_sv, PERL_MAGIC_regex_global)))) {
2132 /* prepare for quick setting of pos */
2133 sv_magic(PL_reg_sv, (SV*)0,
2134 PERL_MAGIC_regex_global, NULL, 0);
2135 mg = mg_find(PL_reg_sv, PERL_MAGIC_regex_global);
2139 PL_reg_oldpos = mg->mg_len;
2140 SAVEDESTRUCTOR_X(restore_pos, 0);
2142 if (!PL_reg_curpm) {
2143 Newxz(PL_reg_curpm, 1, PMOP);
2146 SV* repointer = newSViv(0);
2147 /* so we know which PL_regex_padav element is PL_reg_curpm */
2148 SvFLAGS(repointer) |= SVf_BREAK;
2149 av_push(PL_regex_padav,repointer);
2150 PL_reg_curpm->op_pmoffset = av_len(PL_regex_padav);
2151 PL_regex_pad = AvARRAY(PL_regex_padav);
2155 PM_SETRE(PL_reg_curpm, prog);
2156 PL_reg_oldcurpm = PL_curpm;
2157 PL_curpm = PL_reg_curpm;
2158 if (RX_MATCH_COPIED(prog)) {
2159 /* Here is a serious problem: we cannot rewrite subbeg,
2160 since it may be needed if this match fails. Thus
2161 $` inside (?{}) could fail... */
2162 PL_reg_oldsaved = prog->subbeg;
2163 PL_reg_oldsavedlen = prog->sublen;
2164 #ifdef PERL_OLD_COPY_ON_WRITE
2165 PL_nrs = prog->saved_copy;
2167 RX_MATCH_COPIED_off(prog);
2170 PL_reg_oldsaved = NULL;
2171 prog->subbeg = PL_bostr;
2172 prog->sublen = PL_regeol - PL_bostr; /* strend may have been modified */
2174 prog->startp[0] = startpos - PL_bostr;
2175 PL_reginput = startpos;
2176 PL_regstartp = prog->startp;
2177 PL_regendp = prog->endp;
2178 PL_reglastparen = &prog->lastparen;
2179 PL_reglastcloseparen = &prog->lastcloseparen;
2180 prog->lastparen = 0;
2181 prog->lastcloseparen = 0;
2183 DEBUG_EXECUTE_r(PL_reg_starttry = startpos);
2184 if (PL_reg_start_tmpl <= prog->nparens) {
2185 PL_reg_start_tmpl = prog->nparens*3/2 + 3;
2186 if(PL_reg_start_tmp)
2187 Renew(PL_reg_start_tmp, PL_reg_start_tmpl, char*);
2189 Newx(PL_reg_start_tmp, PL_reg_start_tmpl, char*);
2192 /* XXXX What this code is doing here?!!! There should be no need
2193 to do this again and again, PL_reglastparen should take care of
2196 /* Tests pat.t#187 and split.t#{13,14} seem to depend on this code.
2197 * Actually, the code in regcppop() (which Ilya may be meaning by
2198 * PL_reglastparen), is not needed at all by the test suite
2199 * (op/regexp, op/pat, op/split), but that code is needed, oddly
2200 * enough, for building DynaLoader, or otherwise this
2201 * "Error: '*' not in typemap in DynaLoader.xs, line 164"
2202 * will happen. Meanwhile, this code *is* needed for the
2203 * above-mentioned test suite tests to succeed. The common theme
2204 * on those tests seems to be returning null fields from matches.
2209 if (prog->nparens) {
2210 for (i = prog->nparens; i > (I32)*PL_reglastparen; i--) {
2217 if (regmatch(prog->program + 1)) {
2218 prog->endp[0] = PL_reginput - PL_bostr;
2221 REGCP_UNWIND(lastcp);
2225 #define RE_UNWIND_BRANCH 1
2226 #define RE_UNWIND_BRANCHJ 2
2230 typedef struct { /* XX: makes sense to enlarge it... */
2234 } re_unwind_generic_t;
2247 } re_unwind_branch_t;
2249 typedef union re_unwind_t {
2251 re_unwind_generic_t generic;
2252 re_unwind_branch_t branch;
2255 #define sayYES goto yes
2256 #define sayNO goto no
2257 #define sayNO_ANYOF goto no_anyof
2258 #define sayYES_FINAL goto yes_final
2259 #define sayYES_LOUD goto yes_loud
2260 #define sayNO_FINAL goto no_final
2261 #define sayNO_SILENT goto do_no
2262 #define saySAME(x) if (x) goto yes; else goto no
2264 #define POSCACHE_SUCCESS 0 /* caching success rather than failure */
2265 #define POSCACHE_SEEN 1 /* we know what we're caching */
2266 #define POSCACHE_START 2 /* the real cache: this bit maps to pos 0 */
2267 #define CACHEsayYES STMT_START { \
2268 if (cache_offset | cache_bit) { \
2269 if (!(PL_reg_poscache[0] & (1<<POSCACHE_SEEN))) \
2270 PL_reg_poscache[0] |= (1<<POSCACHE_SUCCESS) || (1<<POSCACHE_SEEN); \
2271 else if (!(PL_reg_poscache[0] & (1<<POSCACHE_SUCCESS))) { \
2272 /* cache records failure, but this is success */ \
2274 PerlIO_printf(Perl_debug_log, \
2275 "%*s (remove success from failure cache)\n", \
2276 REPORT_CODE_OFF+PL_regindent*2, "") \
2278 PL_reg_poscache[cache_offset] &= ~(1<<cache_bit); \
2283 #define CACHEsayNO STMT_START { \
2284 if (cache_offset | cache_bit) { \
2285 if (!(PL_reg_poscache[0] & (1<<POSCACHE_SEEN))) \
2286 PL_reg_poscache[0] |= (1<<POSCACHE_SEEN); \
2287 else if ((PL_reg_poscache[0] & (1<<POSCACHE_SUCCESS))) { \
2288 /* cache records success, but this is failure */ \
2290 PerlIO_printf(Perl_debug_log, \
2291 "%*s (remove failure from success cache)\n", \
2292 REPORT_CODE_OFF+PL_regindent*2, "") \
2294 PL_reg_poscache[cache_offset] &= ~(1<<cache_bit); \
2300 /* this is used to determine how far from the left messages like
2301 'failed...' are printed. Currently 29 makes these messages line
2302 up with the opcode they refer to. Earlier perls used 25 which
2303 left these messages outdented making reviewing a debug output
2306 #define REPORT_CODE_OFF 29
2309 /* Make sure there is a test for this +1 options in re_tests */
2310 #define TRIE_INITAL_ACCEPT_BUFFLEN 4;
2312 #define TRIE_CHECK_STATE_IS_ACCEPTING STMT_START { \
2313 if ( trie->states[ state ].wordnum ) { \
2314 if ( !accepted ) { \
2317 bufflen = TRIE_INITAL_ACCEPT_BUFFLEN ; \
2318 sv_accept_buff=newSV(bufflen * sizeof(reg_trie_accepted) - 1 );\
2319 SvCUR_set( sv_accept_buff, sizeof(reg_trie_accepted) ); \
2320 SvPOK_on( sv_accept_buff ); \
2321 sv_2mortal( sv_accept_buff ); \
2322 accept_buff = (reg_trie_accepted*)SvPV_nolen( sv_accept_buff );\
2324 if ( accepted >= bufflen ) { \
2326 accept_buff =(reg_trie_accepted*)SvGROW( sv_accept_buff, \
2327 bufflen * sizeof(reg_trie_accepted) ); \
2329 SvCUR_set( sv_accept_buff,SvCUR( sv_accept_buff ) \
2330 + sizeof( reg_trie_accepted ) ); \
2332 accept_buff[ accepted ].wordnum = trie->states[ state ].wordnum; \
2333 accept_buff[ accepted ].endpos = uc; \
2337 #define TRIE_HANDLE_CHAR STMT_START { \
2338 if ( uvc < 256 ) { \
2339 charid = trie->charmap[ uvc ]; \
2342 if( trie->widecharmap ) { \
2343 SV** svpp = (SV**)NULL; \
2344 svpp = hv_fetch( trie->widecharmap, (char*)&uvc, \
2345 sizeof( UV ), 0 ); \
2347 charid = (U16)SvIV( *svpp ); \
2352 ( base + charid > trie->uniquecharcount ) && \
2353 ( base + charid - 1 - trie->uniquecharcount < trie->lasttrans) && \
2354 trie->trans[ base + charid - 1 - trie->uniquecharcount ].check == state ) \
2356 state = trie->trans[ base + charid - 1 - trie->uniquecharcount ].next; \
2364 - regmatch - main matching routine
2366 * Conceptually the strategy is simple: check to see whether the current
2367 * node matches, call self recursively to see whether the rest matches,
2368 * and then act accordingly. In practice we make some effort to avoid
2369 * recursion, in particular by going through "ordinary" nodes (that don't
2370 * need to know whether the rest of the match failed) by a loop instead of
2373 /* [lwall] I've hoisted the register declarations to the outer block in order to
2374 * maybe save a little bit of pushing and popping on the stack. It also takes
2375 * advantage of machines that use a register save mask on subroutine entry.
2377 STATIC I32 /* 0 failure, 1 success */
2378 S_regmatch(pTHX_ regnode *prog)
2381 register regnode *scan; /* Current node. */
2382 regnode *next; /* Next node. */
2383 regnode *inner; /* Next node in internal branch. */
2384 register I32 nextchr; /* renamed nextchr - nextchar colides with
2385 function of same name */
2386 register I32 n; /* no or next */
2387 register I32 ln = 0; /* len or last */
2388 register char *s = NULL; /* operand or save */
2389 register char *locinput = PL_reginput;
2390 register I32 c1 = 0, c2 = 0, paren; /* case fold search, parenth */
2391 int minmod = 0, sw = 0, logical = 0;
2394 /* used by the trie code */
2395 SV *sv_accept_buff = NULL; /* accepting states we have traversed */
2396 reg_trie_accepted *accept_buff = NULL; /* "" */
2397 reg_trie_data *trie; /* what trie are we using right now */
2398 U32 accepted = 0; /* how many accepting states we have seen*/
2401 I32 firstcp = PL_savestack_ix;
2403 register const bool do_utf8 = PL_reg_match_utf8;
2405 SV * const dsv0 = PERL_DEBUG_PAD_ZERO(0);
2406 SV * const dsv1 = PERL_DEBUG_PAD_ZERO(1);
2407 SV * const dsv2 = PERL_DEBUG_PAD_ZERO(2);
2409 SV *re_debug_flags = NULL;
2411 U32 uniflags = ckWARN(WARN_UTF8) ? 0 : UTF8_ALLOW_ANY;
2420 /* Note that nextchr is a byte even in UTF */
2421 nextchr = UCHARAT(locinput);
2423 while (scan != NULL) {
2426 SV * const prop = sv_newmortal();
2427 const int docolor = *PL_colors[0];
2428 const int taill = (docolor ? 10 : 7); /* 3 chars for "> <" */
2429 int l = (PL_regeol - locinput) > taill ? taill : (PL_regeol - locinput);
2430 /* The part of the string before starttry has one color
2431 (pref0_len chars), between starttry and current
2432 position another one (pref_len - pref0_len chars),
2433 after the current position the third one.
2434 We assume that pref0_len <= pref_len, otherwise we
2435 decrease pref0_len. */
2436 int pref_len = (locinput - PL_bostr) > (5 + taill) - l
2437 ? (5 + taill) - l : locinput - PL_bostr;
2440 while (do_utf8 && UTF8_IS_CONTINUATION(*(U8*)(locinput - pref_len)))
2442 pref0_len = pref_len - (locinput - PL_reg_starttry);
2443 if (l + pref_len < (5 + taill) && l < PL_regeol - locinput)
2444 l = ( PL_regeol - locinput > (5 + taill) - pref_len
2445 ? (5 + taill) - pref_len : PL_regeol - locinput);
2446 while (do_utf8 && UTF8_IS_CONTINUATION(*(U8*)(locinput + l)))
2450 if (pref0_len > pref_len)
2451 pref0_len = pref_len;
2452 regprop(prop, scan);
2454 const char * const s0 =
2455 do_utf8 && OP(scan) != CANY ?
2456 pv_uni_display(dsv0, (U8*)(locinput - pref_len),
2457 pref0_len, 60, UNI_DISPLAY_REGEX) :
2458 locinput - pref_len;
2459 const int len0 = do_utf8 ? strlen(s0) : pref0_len;
2460 const char * const s1 = do_utf8 && OP(scan) != CANY ?
2461 pv_uni_display(dsv1, (U8*)(locinput - pref_len + pref0_len),
2462 pref_len - pref0_len, 60, UNI_DISPLAY_REGEX) :
2463 locinput - pref_len + pref0_len;
2464 const int len1 = do_utf8 ? strlen(s1) : pref_len - pref0_len;
2465 const char * const s2 = do_utf8 && OP(scan) != CANY ?
2466 pv_uni_display(dsv2, (U8*)locinput,
2467 PL_regeol - locinput, 60, UNI_DISPLAY_REGEX) :
2469 const int len2 = do_utf8 ? strlen(s2) : l;
2470 PerlIO_printf(Perl_debug_log,
2471 "%4"IVdf" <%s%.*s%s%s%.*s%s%s%s%.*s%s>%*s|%3"IVdf":%*s%s\n",
2472 (IV)(locinput - PL_bostr),
2479 (docolor ? "" : "> <"),
2483 15 - l - pref_len + 1,
2485 (IV)(scan - PL_regprogram), PL_regindent*2, "",
2490 next = scan + NEXT_OFF(scan);
2496 if (locinput == PL_bostr)
2498 /* regtill = regbol; */
2503 if (locinput == PL_bostr ||
2504 ((nextchr || locinput < PL_regeol) && locinput[-1] == '\n'))
2510 if (locinput == PL_bostr)
2514 if (locinput == PL_reg_ganch)
2520 if ((nextchr || locinput < PL_regeol) && nextchr != '\n')
2525 if ((nextchr || locinput < PL_regeol) && nextchr != '\n')
2527 if (PL_regeol - locinput > 1)
2531 if (PL_regeol != locinput)
2535 if (!nextchr && locinput >= PL_regeol)
2538 locinput += PL_utf8skip[nextchr];
2539 if (locinput > PL_regeol)
2541 nextchr = UCHARAT(locinput);
2544 nextchr = UCHARAT(++locinput);
2547 if (!nextchr && locinput >= PL_regeol)
2549 nextchr = UCHARAT(++locinput);
2552 if ((!nextchr && locinput >= PL_regeol) || nextchr == '\n')
2555 locinput += PL_utf8skip[nextchr];
2556 if (locinput > PL_regeol)
2558 nextchr = UCHARAT(locinput);
2561 nextchr = UCHARAT(++locinput);
2567 traverse the TRIE keeping track of all accepting states
2568 we transition through until we get to a failing node.
2570 we use two slightly different pieces of code to handle
2571 the traversal depending on whether its case sensitive or
2572 not. we reuse the accept code however. (this should probably
2573 be turned into a macro.)
2579 U8 *uc = ( U8* )locinput;
2586 U8 *uscan = (U8*)NULL;
2590 trie = (reg_trie_data*)PL_regdata->data[ ARG( scan ) ];
2592 while ( state && uc <= (U8*)PL_regeol ) {
2594 TRIE_CHECK_STATE_IS_ACCEPTING;
2596 base = trie->states[ state ].trans.base;
2598 DEBUG_TRIE_EXECUTE_r(
2599 PerlIO_printf( Perl_debug_log,
2600 "%*s %sState: %4"UVxf", Base: %4"UVxf", Accepted: %4"UVxf" ",
2601 REPORT_CODE_OFF + PL_regindent * 2, "", PL_colors[4],
2602 (UV)state, (UV)base, (UV)accepted );
2609 uvc = utf8n_to_uvuni( uscan, UTF8_MAXLEN, &len, uniflags );
2614 U8 foldbuf[ UTF8_MAXBYTES_CASE + 1 ];
2615 uvc = utf8n_to_uvuni( (U8*)uc, UTF8_MAXLEN, &len, uniflags );
2616 uvc = to_uni_fold( uvc, foldbuf, &foldlen );
2617 foldlen -= UNISKIP( uvc );
2618 uscan = foldbuf + UNISKIP( uvc );
2630 DEBUG_TRIE_EXECUTE_r(
2631 PerlIO_printf( Perl_debug_log,
2632 "Charid:%3x CV:%4"UVxf" After State: %4"UVxf"%s\n",
2633 charid, uvc, (UV)state, PL_colors[5] );
2642 /* unreached codepoint: we jump into the middle of the next case
2643 from previous if blocks */
2646 U8 *uc = (U8*)locinput;
2655 trie = (reg_trie_data*)PL_regdata->data[ ARG( scan ) ];
2657 while ( state && uc <= (U8*)PL_regeol ) {
2659 TRIE_CHECK_STATE_IS_ACCEPTING;
2661 base = trie->states[ state ].trans.base;
2663 DEBUG_TRIE_EXECUTE_r(
2664 PerlIO_printf( Perl_debug_log,
2665 "%*s %sState: %4"UVxf", Base: %4"UVxf", Accepted: %4"UVxf" ",
2666 REPORT_CODE_OFF + PL_regindent * 2, "", PL_colors[4],
2667 (UV)state, (UV)base, (UV)accepted );
2673 uvc = utf8n_to_uvuni( (U8*)uc, UTF8_MAXLEN, &len, uniflags );
2684 DEBUG_TRIE_EXECUTE_r(
2685 PerlIO_printf( Perl_debug_log,
2686 "Charid:%3x CV:%4"UVxf" After State: %4"UVxf"%s\n",
2687 charid, uvc, (UV)state, PL_colors[5] );
2697 There was at least one accepting state that we
2698 transitioned through. Presumably the number of accepting
2699 states is going to be low, typically one or two. So we
2700 simply scan through to find the one with lowest wordnum.
2701 Once we find it, we swap the last state into its place
2702 and decrement the size. We then try to match the rest of
2703 the pattern at the point where the word ends, if we
2704 succeed then we end the loop, otherwise the loop
2705 eventually terminates once all of the accepting states
2712 if ( accepted == 1 ) {
2714 SV **tmp = av_fetch( trie->words, accept_buff[ 0 ].wordnum-1, 0 );
2715 PerlIO_printf( Perl_debug_log,
2716 "%*s %sonly one match : #%d <%s>%s\n",
2717 REPORT_CODE_OFF+PL_regindent*2, "", PL_colors[4],
2718 accept_buff[ 0 ].wordnum,
2719 tmp ? SvPV_nolen_const( *tmp ) : "not compiled under -Dr",
2722 PL_reginput = (char *)accept_buff[ 0 ].endpos;
2723 /* in this case we free tmps/leave before we call regmatch
2724 as we wont be using accept_buff again. */
2727 gotit = regmatch( scan + NEXT_OFF( scan ) );
2730 PerlIO_printf( Perl_debug_log,"%*s %sgot %"IVdf" possible matches%s\n",
2731 REPORT_CODE_OFF + PL_regindent * 2, "", PL_colors[4], (IV)accepted,
2734 while ( !gotit && accepted-- ) {
2737 for( cur = 1 ; cur <= accepted ; cur++ ) {
2738 DEBUG_TRIE_EXECUTE_r(
2739 PerlIO_printf( Perl_debug_log,
2740 "%*s %sgot %"IVdf" (%d) as best, looking at %"IVdf" (%d)%s\n",
2741 REPORT_CODE_OFF + PL_regindent * 2, "", PL_colors[4],
2742 (IV)best, accept_buff[ best ].wordnum, (IV)cur,
2743 accept_buff[ cur ].wordnum, PL_colors[5] );
2746 if ( accept_buff[ cur ].wordnum < accept_buff[ best ].wordnum )
2750 SV ** const tmp = av_fetch( trie->words, accept_buff[ best ].wordnum - 1, 0 );
2751 PerlIO_printf( Perl_debug_log, "%*s %strying alternation #%d <%s> at 0x%p%s\n",
2752 REPORT_CODE_OFF+PL_regindent*2, "", PL_colors[4],
2753 accept_buff[best].wordnum,
2754 tmp ? SvPV_nolen_const( *tmp ) : "not compiled under -Dr",scan,
2757 if ( best<accepted ) {
2758 reg_trie_accepted tmp = accept_buff[ best ];
2759 accept_buff[ best ] = accept_buff[ accepted ];
2760 accept_buff[ accepted ] = tmp;
2763 PL_reginput = (char *)accept_buff[ best ].endpos;
2766 as far as I can tell we only need the SAVETMPS/FREETMPS
2767 for re's with EVAL in them but I'm leaving them in for
2768 all until I can be sure.
2771 gotit = regmatch( scan + NEXT_OFF( scan ) ) ;
2784 /* unreached codepoint */
2788 if (do_utf8 != UTF) {
2789 /* The target and the pattern have differing utf8ness. */
2791 const char *e = s + ln;
2794 /* The target is utf8, the pattern is not utf8. */
2799 if (NATIVE_TO_UNI(*(U8*)s) !=
2800 utf8n_to_uvuni((U8*)l, UTF8_MAXBYTES, &ulen,
2808 /* The target is not utf8, the pattern is utf8. */
2813 if (NATIVE_TO_UNI(*((U8*)l)) !=
2814 utf8n_to_uvuni((U8*)s, UTF8_MAXBYTES, &ulen,
2822 nextchr = UCHARAT(locinput);
2825 /* The target and the pattern have the same utf8ness. */
2826 /* Inline the first character, for speed. */
2827 if (UCHARAT(s) != nextchr)
2829 if (PL_regeol - locinput < ln)
2831 if (ln > 1 && memNE(s, locinput, ln))
2834 nextchr = UCHARAT(locinput);
2837 PL_reg_flags |= RF_tainted;
2843 if (do_utf8 || UTF) {
2844 /* Either target or the pattern are utf8. */
2846 char *e = PL_regeol;
2848 if (ibcmp_utf8(s, 0, ln, (bool)UTF,
2849 l, &e, 0, do_utf8)) {
2850 /* One more case for the sharp s:
2851 * pack("U0U*", 0xDF) =~ /ss/i,
2852 * the 0xC3 0x9F are the UTF-8
2853 * byte sequence for the U+00DF. */
2855 toLOWER(s[0]) == 's' &&
2857 toLOWER(s[1]) == 's' &&
2864 nextchr = UCHARAT(locinput);
2868 /* Neither the target and the pattern are utf8. */
2870 /* Inline the first character, for speed. */
2871 if (UCHARAT(s) != nextchr &&
2872 UCHARAT(s) != ((OP(scan) == EXACTF)
2873 ? PL_fold : PL_fold_locale)[nextchr])
2875 if (PL_regeol - locinput < ln)
2877 if (ln > 1 && (OP(scan) == EXACTF
2878 ? ibcmp(s, locinput, ln)
2879 : ibcmp_locale(s, locinput, ln)))
2882 nextchr = UCHARAT(locinput);
2886 STRLEN inclasslen = PL_regeol - locinput;
2888 if (!reginclass(scan, (U8*)locinput, &inclasslen, do_utf8))
2890 if (locinput >= PL_regeol)
2892 locinput += inclasslen ? inclasslen : UTF8SKIP(locinput);
2893 nextchr = UCHARAT(locinput);
2898 nextchr = UCHARAT(locinput);
2899 if (!REGINCLASS(scan, (U8*)locinput))
2901 if (!nextchr && locinput >= PL_regeol)
2903 nextchr = UCHARAT(++locinput);
2907 /* If we might have the case of the German sharp s
2908 * in a casefolding Unicode character class. */
2910 if (ANYOF_FOLD_SHARP_S(scan, locinput, PL_regeol)) {
2911 locinput += SHARP_S_SKIP;
2912 nextchr = UCHARAT(locinput);
2918 PL_reg_flags |= RF_tainted;
2924 LOAD_UTF8_CHARCLASS_ALNUM();
2925 if (!(OP(scan) == ALNUM
2926 ? swash_fetch(PL_utf8_alnum, (U8*)locinput, do_utf8)
2927 : isALNUM_LC_utf8((U8*)locinput)))
2931 locinput += PL_utf8skip[nextchr];
2932 nextchr = UCHARAT(locinput);
2935 if (!(OP(scan) == ALNUM
2936 ? isALNUM(nextchr) : isALNUM_LC(nextchr)))
2938 nextchr = UCHARAT(++locinput);
2941 PL_reg_flags |= RF_tainted;
2944 if (!nextchr && locinput >= PL_regeol)
2947 LOAD_UTF8_CHARCLASS_ALNUM();
2948 if (OP(scan) == NALNUM
2949 ? swash_fetch(PL_utf8_alnum, (U8*)locinput, do_utf8)
2950 : isALNUM_LC_utf8((U8*)locinput))
2954 locinput += PL_utf8skip[nextchr];
2955 nextchr = UCHARAT(locinput);
2958 if (OP(scan) == NALNUM
2959 ? isALNUM(nextchr) : isALNUM_LC(nextchr))
2961 nextchr = UCHARAT(++locinput);
2965 PL_reg_flags |= RF_tainted;
2969 /* was last char in word? */
2971 if (locinput == PL_bostr)
2974 const U8 * const r = reghop3((U8*)locinput, -1, (U8*)PL_bostr);
2976 ln = utf8n_to_uvchr(r, UTF8SKIP(r), 0, 0);
2978 if (OP(scan) == BOUND || OP(scan) == NBOUND) {
2979 ln = isALNUM_uni(ln);
2980 LOAD_UTF8_CHARCLASS_ALNUM();
2981 n = swash_fetch(PL_utf8_alnum, (U8*)locinput, do_utf8);
2984 ln = isALNUM_LC_uvchr(UNI_TO_NATIVE(ln));
2985 n = isALNUM_LC_utf8((U8*)locinput);
2989 ln = (locinput != PL_bostr) ?
2990 UCHARAT(locinput - 1) : '\n';
2991 if (OP(scan) == BOUND || OP(scan) == NBOUND) {
2993 n = isALNUM(nextchr);
2996 ln = isALNUM_LC(ln);
2997 n = isALNUM_LC(nextchr);
3000 if (((!ln) == (!n)) == (OP(scan) == BOUND ||
3001 OP(scan) == BOUNDL))
3005 PL_reg_flags |= RF_tainted;
3011 if (UTF8_IS_CONTINUED(nextchr)) {
3012 LOAD_UTF8_CHARCLASS_SPACE();
3013 if (!(OP(scan) == SPACE
3014 ? swash_fetch(PL_utf8_space, (U8*)locinput, do_utf8)
3015 : isSPACE_LC_utf8((U8*)locinput)))
3019 locinput += PL_utf8skip[nextchr];
3020 nextchr = UCHARAT(locinput);
3023 if (!(OP(scan) == SPACE
3024 ? isSPACE(nextchr) : isSPACE_LC(nextchr)))
3026 nextchr = UCHARAT(++locinput);
3029 if (!(OP(scan) == SPACE
3030 ? isSPACE(nextchr) : isSPACE_LC(nextchr)))
3032 nextchr = UCHARAT(++locinput);
3036 PL_reg_flags |= RF_tainted;
3039 if (!nextchr && locinput >= PL_regeol)
3042 LOAD_UTF8_CHARCLASS_SPACE();
3043 if (OP(scan) == NSPACE
3044 ? swash_fetch(PL_utf8_space, (U8*)locinput, do_utf8)
3045 : isSPACE_LC_utf8((U8*)locinput))
3049 locinput += PL_utf8skip[nextchr];
3050 nextchr = UCHARAT(locinput);
3053 if (OP(scan) == NSPACE
3054 ? isSPACE(nextchr) : isSPACE_LC(nextchr))
3056 nextchr = UCHARAT(++locinput);
3059 PL_reg_flags |= RF_tainted;
3065 LOAD_UTF8_CHARCLASS_DIGIT();
3066 if (!(OP(scan) == DIGIT
3067 ? swash_fetch(PL_utf8_digit, (U8*)locinput, do_utf8)
3068 : isDIGIT_LC_utf8((U8*)locinput)))
3072 locinput += PL_utf8skip[nextchr];
3073 nextchr = UCHARAT(locinput);
3076 if (!(OP(scan) == DIGIT
3077 ? isDIGIT(nextchr) : isDIGIT_LC(nextchr)))
3079 nextchr = UCHARAT(++locinput);
3082 PL_reg_flags |= RF_tainted;
3085 if (!nextchr && locinput >= PL_regeol)
3088 LOAD_UTF8_CHARCLASS_DIGIT();
3089 if (OP(scan) == NDIGIT
3090 ? swash_fetch(PL_utf8_digit, (U8*)locinput, do_utf8)
3091 : isDIGIT_LC_utf8((U8*)locinput))
3095 locinput += PL_utf8skip[nextchr];
3096 nextchr = UCHARAT(locinput);
3099 if (OP(scan) == NDIGIT
3100 ? isDIGIT(nextchr) : isDIGIT_LC(nextchr))
3102 nextchr = UCHARAT(++locinput);
3105 if (locinput >= PL_regeol)
3108 LOAD_UTF8_CHARCLASS_MARK();
3109 if (swash_fetch(PL_utf8_mark,(U8*)locinput, do_utf8))
3111 locinput += PL_utf8skip[nextchr];
3112 while (locinput < PL_regeol &&
3113 swash_fetch(PL_utf8_mark,(U8*)locinput, do_utf8))
3114 locinput += UTF8SKIP(locinput);
3115 if (locinput > PL_regeol)
3120 nextchr = UCHARAT(locinput);
3123 PL_reg_flags |= RF_tainted;
3127 n = ARG(scan); /* which paren pair */
3128 ln = PL_regstartp[n];
3129 PL_reg_leftiter = PL_reg_maxiter; /* Void cache */
3130 if ((I32)*PL_reglastparen < n || ln == -1)
3131 sayNO; /* Do not match unless seen CLOSEn. */
3132 if (ln == PL_regendp[n])
3136 if (do_utf8 && OP(scan) != REF) { /* REF can do byte comparison */
3138 const char *e = PL_bostr + PL_regendp[n];
3140 * Note that we can't do the "other character" lookup trick as
3141 * in the 8-bit case (no pun intended) because in Unicode we
3142 * have to map both upper and title case to lower case.
3144 if (OP(scan) == REFF) {
3146 STRLEN ulen1, ulen2;
3147 U8 tmpbuf1[UTF8_MAXBYTES_CASE+1];
3148 U8 tmpbuf2[UTF8_MAXBYTES_CASE+1];
3152 toLOWER_utf8((U8*)s, tmpbuf1, &ulen1);
3153 toLOWER_utf8((U8*)l, tmpbuf2, &ulen2);
3154 if (ulen1 != ulen2 || memNE((char *)tmpbuf1, (char *)tmpbuf2, ulen1))
3161 nextchr = UCHARAT(locinput);
3165 /* Inline the first character, for speed. */
3166 if (UCHARAT(s) != nextchr &&
3168 (UCHARAT(s) != ((OP(scan) == REFF
3169 ? PL_fold : PL_fold_locale)[nextchr]))))
3171 ln = PL_regendp[n] - ln;
3172 if (locinput + ln > PL_regeol)
3174 if (ln > 1 && (OP(scan) == REF
3175 ? memNE(s, locinput, ln)
3177 ? ibcmp(s, locinput, ln)
3178 : ibcmp_locale(s, locinput, ln))))
3181 nextchr = UCHARAT(locinput);
3192 OP_4tree * const oop = PL_op;
3193 COP * const ocurcop = PL_curcop;
3196 struct regexp * const oreg = PL_reg_re;
3199 PL_op = (OP_4tree*)PL_regdata->data[n];
3200 DEBUG_EXECUTE_r( PerlIO_printf(Perl_debug_log, " re_eval 0x%"UVxf"\n", PTR2UV(PL_op)) );
3201 PAD_SAVE_LOCAL(old_comppad, (PAD*)PL_regdata->data[n + 2]);
3202 PL_regendp[0] = PL_reg_magic->mg_len = locinput - PL_bostr;
3205 SV ** const before = SP;
3206 CALLRUNOPS(aTHX); /* Scalar context. */
3209 ret = &PL_sv_undef; /* protect against empty (?{}) blocks. */
3217 PAD_RESTORE_LOCAL(old_comppad);
3218 PL_curcop = ocurcop;
3220 if (logical == 2) { /* Postponed subexpression. */
3224 CHECKPOINT cp, lastcp;
3228 if(SvROK(ret) && SvSMAGICAL(sv = SvRV(ret)))
3229 mg = mg_find(sv, PERL_MAGIC_qr);
3230 else if (SvSMAGICAL(ret)) {
3231 if (SvGMAGICAL(ret))
3232 sv_unmagic(ret, PERL_MAGIC_qr);
3234 mg = mg_find(ret, PERL_MAGIC_qr);
3238 re = (regexp *)mg->mg_obj;
3239 (void)ReREFCNT_inc(re);
3243 const char * const t = SvPV_const(ret, len);
3245 char * const oprecomp = PL_regprecomp;
3246 const I32 osize = PL_regsize;
3247 const I32 onpar = PL_regnpar;
3250 if (DO_UTF8(ret)) pm.op_pmdynflags |= PMdf_DYN_UTF8;
3251 re = CALLREGCOMP(aTHX_ (char*)t, (char*)t + len, &pm);
3253 & (SVs_TEMP | SVs_PADTMP | SVf_READONLY
3255 sv_magic(ret,(SV*)ReREFCNT_inc(re),
3257 PL_regprecomp = oprecomp;
3262 PerlIO_printf(Perl_debug_log,
3263 "Entering embedded \"%s%.60s%s%s\"\n",
3267 (strlen(re->precomp) > 60 ? "..." : ""))
3270 state.prev = PL_reg_call_cc;
3271 state.cc = PL_regcc;
3272 state.re = PL_reg_re;
3276 cp = regcppush(0); /* Save *all* the positions. */
3279 state.ss = PL_savestack_ix;
3280 *PL_reglastparen = 0;
3281 *PL_reglastcloseparen = 0;
3282 PL_reg_call_cc = &state;
3283 PL_reginput = locinput;
3284 toggleutf = ((PL_reg_flags & RF_utf8) != 0) ^
3285 ((re->reganch & ROPT_UTF8) != 0);
3286 if (toggleutf) PL_reg_flags ^= RF_utf8;
3288 /* XXXX This is too dramatic a measure... */
3291 if (regmatch(re->program + 1)) {
3292 /* Even though we succeeded, we need to restore
3293 global variables, since we may be wrapped inside
3294 SUSPEND, thus the match may be not finished yet. */
3296 /* XXXX Do this only if SUSPENDed? */
3297 PL_reg_call_cc = state.prev;
3298 PL_regcc = state.cc;
3299 PL_reg_re = state.re;
3300 cache_re(PL_reg_re);
3301 if (toggleutf) PL_reg_flags ^= RF_utf8;
3303 /* XXXX This is too dramatic a measure... */
3306 /* These are needed even if not SUSPEND. */
3312 REGCP_UNWIND(lastcp);
3314 PL_reg_call_cc = state.prev;
3315 PL_regcc = state.cc;
3316 PL_reg_re = state.re;
3317 cache_re(PL_reg_re);
3318 if (toggleutf) PL_reg_flags ^= RF_utf8;
3320 /* XXXX This is too dramatic a measure... */
3330 sv_setsv(save_scalar(PL_replgv), ret);
3336 n = ARG(scan); /* which paren pair */
3337 PL_reg_start_tmp[n] = locinput;
3342 n = ARG(scan); /* which paren pair */
3343 PL_regstartp[n] = PL_reg_start_tmp[n] - PL_bostr;
3344 PL_regendp[n] = locinput - PL_bostr;
3345 if (n > (I32)*PL_reglastparen)
3346 *PL_reglastparen = n;
3347 *PL_reglastcloseparen = n;
3350 n = ARG(scan); /* which paren pair */
3351 sw = ((I32)*PL_reglastparen >= n && PL_regendp[n] != -1);
3354 PL_reg_leftiter = PL_reg_maxiter; /* Void cache */
3356 next = NEXTOPER(NEXTOPER(scan));
3358 next = scan + ARG(scan);
3359 if (OP(next) == IFTHEN) /* Fake one. */
3360 next = NEXTOPER(NEXTOPER(next));
3364 logical = scan->flags;
3366 /*******************************************************************
3367 PL_regcc contains infoblock about the innermost (...)* loop, and
3368 a pointer to the next outer infoblock.
3370 Here is how Y(A)*Z is processed (if it is compiled into CURLYX/WHILEM):
3372 1) After matching X, regnode for CURLYX is processed;
3374 2) This regnode creates infoblock on the stack, and calls
3375 regmatch() recursively with the starting point at WHILEM node;
3377 3) Each hit of WHILEM node tries to match A and Z (in the order
3378 depending on the current iteration, min/max of {min,max} and
3379 greediness). The information about where are nodes for "A"
3380 and "Z" is read from the infoblock, as is info on how many times "A"
3381 was already matched, and greediness.
3383 4) After A matches, the same WHILEM node is hit again.
3385 5) Each time WHILEM is hit, PL_regcc is the infoblock created by CURLYX
3386 of the same pair. Thus when WHILEM tries to match Z, it temporarily
3387 resets PL_regcc, since this Y(A)*Z can be a part of some other loop:
3388 as in (Y(A)*Z)*. If Z matches, the automaton will hit the WHILEM node
3389 of the external loop.
3391 Currently present infoblocks form a tree with a stem formed by PL_curcc
3392 and whatever it mentions via ->next, and additional attached trees
3393 corresponding to temporarily unset infoblocks as in "5" above.
3395 In the following picture infoblocks for outer loop of
3396 (Y(A)*?Z)*?T are denoted O, for inner I. NULL starting block
3397 is denoted by x. The matched string is YAAZYAZT. Temporarily postponed
3398 infoblocks are drawn below the "reset" infoblock.
3400 In fact in the picture below we do not show failed matches for Z and T
3401 by WHILEM blocks. [We illustrate minimal matches, since for them it is
3402 more obvious *why* one needs to *temporary* unset infoblocks.]
3404 Matched REx position InfoBlocks Comment
3408 Y A)*?Z)*?T x <- O <- I
3409 YA )*?Z)*?T x <- O <- I
3410 YA A)*?Z)*?T x <- O <- I
3411 YAA )*?Z)*?T x <- O <- I
3412 YAA Z)*?T x <- O # Temporary unset I
3415 YAAZ Y(A)*?Z)*?T x <- O
3418 YAAZY (A)*?Z)*?T x <- O
3421 YAAZY A)*?Z)*?T x <- O <- I
3424 YAAZYA )*?Z)*?T x <- O <- I
3427 YAAZYA Z)*?T x <- O # Temporary unset I
3433 YAAZYAZ T x # Temporary unset O
3440 *******************************************************************/
3443 CHECKPOINT cp = PL_savestack_ix;
3444 /* No need to save/restore up to this paren */
3445 I32 parenfloor = scan->flags;
3447 if (OP(PREVOPER(next)) == NOTHING) /* LONGJMP */
3449 cc.oldcc = PL_regcc;
3451 /* XXXX Probably it is better to teach regpush to support
3452 parenfloor > PL_regsize... */
3453 if (parenfloor > (I32)*PL_reglastparen)
3454 parenfloor = *PL_reglastparen; /* Pessimization... */
3455 cc.parenfloor = parenfloor;
3457 cc.min = ARG1(scan);
3458 cc.max = ARG2(scan);
3459 cc.scan = NEXTOPER(scan) + EXTRA_STEP_2ARGS;
3463 PL_reginput = locinput;
3464 n = regmatch(PREVOPER(next)); /* start on the WHILEM */
3466 PL_regcc = cc.oldcc;
3472 * This is really hard to understand, because after we match
3473 * what we're trying to match, we must make sure the rest of
3474 * the REx is going to match for sure, and to do that we have
3475 * to go back UP the parse tree by recursing ever deeper. And
3476 * if it fails, we have to reset our parent's current state
3477 * that we can try again after backing off.
3480 CHECKPOINT cp, lastcp;
3481 CURCUR* cc = PL_regcc;
3482 char * const lastloc = cc->lastloc; /* Detection of 0-len. */
3483 I32 cache_offset = 0, cache_bit = 0;
3485 n = cc->cur + 1; /* how many we know we matched */
3486 PL_reginput = locinput;
3489 PerlIO_printf(Perl_debug_log,
3490 "%*s %ld out of %ld..%ld cc=%"UVxf"\n",
3491 REPORT_CODE_OFF+PL_regindent*2, "",
3492 (long)n, (long)cc->min,
3493 (long)cc->max, PTR2UV(cc))
3496 /* If degenerate scan matches "", assume scan done. */
3498 if (locinput == cc->lastloc && n >= cc->min) {
3499 PL_regcc = cc->oldcc;
3503 PerlIO_printf(Perl_debug_log,
3504 "%*s empty match detected, try continuation...\n",
3505 REPORT_CODE_OFF+PL_regindent*2, "")
3507 if (regmatch(cc->next))
3515 /* First just match a string of min scans. */
3519 cc->lastloc = locinput;
3520 if (regmatch(cc->scan))
3523 cc->lastloc = lastloc;
3528 /* Check whether we already were at this position.
3529 Postpone detection until we know the match is not
3530 *that* much linear. */
3531 if (!PL_reg_maxiter) {
3532 PL_reg_maxiter = (PL_regeol - PL_bostr + 1) * (scan->flags>>4);
3533 PL_reg_leftiter = PL_reg_maxiter;
3535 if (PL_reg_leftiter-- == 0) {
3536 const I32 size = (PL_reg_maxiter + 7 + POSCACHE_START)/8;
3537 if (PL_reg_poscache) {
3538 if ((I32)PL_reg_poscache_size < size) {
3539 Renew(PL_reg_poscache, size, char);
3540 PL_reg_poscache_size = size;
3542 Zero(PL_reg_poscache, size, char);
3545 PL_reg_poscache_size = size;
3546 Newxz(PL_reg_poscache, size, char);
3549 PerlIO_printf(Perl_debug_log,
3550 "%sDetected a super-linear match, switching on caching%s...\n",
3551 PL_colors[4], PL_colors[5])
3554 if (PL_reg_leftiter < 0) {
3555 cache_offset = locinput - PL_bostr;
3557 cache_offset = (scan->flags & 0xf) - 1 + POSCACHE_START
3558 + cache_offset * (scan->flags>>4);
3559 cache_bit = cache_offset % 8;
3561 if (PL_reg_poscache[cache_offset] & (1<<cache_bit)) {
3563 PerlIO_printf(Perl_debug_log,
3564 "%*s already tried at this position...\n",
3565 REPORT_CODE_OFF+PL_regindent*2, "")
3567 if (PL_reg_poscache[0] & (1<<POSCACHE_SUCCESS))
3568 /* cache records success */
3571 /* cache records failure */
3574 PL_reg_poscache[cache_offset] |= (1<<cache_bit);
3578 /* Prefer next over scan for minimal matching. */
3581 PL_regcc = cc->oldcc;
3584 cp = regcppush(cc->parenfloor);
3586 if (regmatch(cc->next)) {
3588 CACHEsayYES; /* All done. */
3590 REGCP_UNWIND(lastcp);
3596 if (n >= cc->max) { /* Maximum greed exceeded? */
3597 if (ckWARN(WARN_REGEXP) && n >= REG_INFTY
3598 && !(PL_reg_flags & RF_warned)) {
3599 PL_reg_flags |= RF_warned;
3600 Perl_warner(aTHX_ packWARN(WARN_REGEXP), "%s limit (%d) exceeded",
3601 "Complex regular subexpression recursion",
3608 PerlIO_printf(Perl_debug_log,
3609 "%*s trying longer...\n",
3610 REPORT_CODE_OFF+PL_regindent*2, "")
3612 /* Try scanning more and see if it helps. */
3613 PL_reginput = locinput;
3615 cc->lastloc = locinput;
3616 cp = regcppush(cc->parenfloor);
3618 if (regmatch(cc->scan)) {
3622 REGCP_UNWIND(lastcp);
3625 cc->lastloc = lastloc;
3629 /* Prefer scan over next for maximal matching. */
3631 if (n < cc->max) { /* More greed allowed? */
3632 cp = regcppush(cc->parenfloor);
3634 cc->lastloc = locinput;
3636 if (regmatch(cc->scan)) {
3640 REGCP_UNWIND(lastcp);
3641 regcppop(); /* Restore some previous $<digit>s? */
3642 PL_reginput = locinput;
3644 PerlIO_printf(Perl_debug_log,
3645 "%*s failed, try continuation...\n",
3646 REPORT_CODE_OFF+PL_regindent*2, "")
3649 if (ckWARN(WARN_REGEXP) && n >= REG_INFTY
3650 && !(PL_reg_flags & RF_warned)) {
3651 PL_reg_flags |= RF_warned;
3652 Perl_warner(aTHX_ packWARN(WARN_REGEXP), "%s limit (%d) exceeded",
3653 "Complex regular subexpression recursion",
3657 /* Failed deeper matches of scan, so see if this one works. */
3658 PL_regcc = cc->oldcc;
3661 if (regmatch(cc->next))
3667 cc->lastloc = lastloc;
3672 next = scan + ARG(scan);
3675 inner = NEXTOPER(NEXTOPER(scan));
3678 inner = NEXTOPER(scan);
3682 if (OP(next) != c1) /* No choice. */
3683 next = inner; /* Avoid recursion. */
3685 const I32 lastparen = *PL_reglastparen;
3686 /* Put unwinding data on stack */
3687 const I32 unwind1 = SSNEWt(1,re_unwind_branch_t);
3688 re_unwind_branch_t * const uw = SSPTRt(unwind1,re_unwind_branch_t);
3692 uw->type = ((c1 == BRANCH)
3694 : RE_UNWIND_BRANCHJ);
3695 uw->lastparen = lastparen;
3697 uw->locinput = locinput;
3698 uw->nextchr = nextchr;
3700 uw->regindent = ++PL_regindent;
3703 REGCP_SET(uw->lastcp);
3705 /* Now go into the first branch */
3720 /* We suppose that the next guy does not need
3721 backtracking: in particular, it is of constant non-zero length,
3722 and has no parenths to influence future backrefs. */
3723 ln = ARG1(scan); /* min to match */
3724 n = ARG2(scan); /* max to match */
3725 paren = scan->flags;
3727 if (paren > PL_regsize)
3729 if (paren > (I32)*PL_reglastparen)
3730 *PL_reglastparen = paren;
3732 scan = NEXTOPER(scan) + NODE_STEP_REGNODE;
3734 scan += NEXT_OFF(scan); /* Skip former OPEN. */
3735 PL_reginput = locinput;
3736 maxwanted = minmod ? ln : n;
3738 while (PL_reginput < PL_regeol && matches < maxwanted) {
3739 if (!regmatch(scan))
3741 /* on first match, determine length, l */
3743 if (PL_reg_match_utf8) {
3745 while (s < PL_reginput) {
3751 l = PL_reginput - locinput;
3754 matches = maxwanted;
3758 locinput = PL_reginput;
3762 PL_reginput = locinput;
3766 if (ln && matches < ln)
3768 if (HAS_TEXT(next) || JUMPABLE(next)) {
3769 regnode *text_node = next;
3771 if (! HAS_TEXT(text_node)) FIND_NEXT_IMPT(text_node);
3773 if (! HAS_TEXT(text_node)) c1 = c2 = -1000;
3775 if (PL_regkind[(U8)OP(text_node)] == REF) {
3779 else { c1 = (U8)*STRING(text_node); }
3780 if (OP(text_node) == EXACTF || OP(text_node) == REFF)
3782 else if (OP(text_node) == EXACTFL || OP(text_node) == REFFL)
3783 c2 = PL_fold_locale[c1];
3792 while (n >= ln || (n == REG_INFTY && ln > 0)) { /* ln overflow ? */
3793 /* If it could work, try it. */
3795 UCHARAT(PL_reginput) == c1 ||
3796 UCHARAT(PL_reginput) == c2)
3800 PL_regstartp[paren] =
3801 HOPc(PL_reginput, -l) - PL_bostr;
3802 PL_regendp[paren] = PL_reginput - PL_bostr;
3805 PL_regendp[paren] = -1;
3809 REGCP_UNWIND(lastcp);
3811 /* Couldn't or didn't -- move forward. */
3812 PL_reginput = locinput;
3813 if (regmatch(scan)) {
3815 locinput = PL_reginput;
3823 PerlIO_printf(Perl_debug_log,
3824 "%*s matched %"IVdf" times, len=%"IVdf"...\n",
3825 (int)(REPORT_CODE_OFF+PL_regindent*2), "",
3826 (IV) matches, (IV)l)
3828 if (matches >= ln) {
3829 if (HAS_TEXT(next) || JUMPABLE(next)) {
3830 regnode *text_node = next;
3832 if (! HAS_TEXT(text_node)) FIND_NEXT_IMPT(text_node);
3834 if (! HAS_TEXT(text_node)) c1 = c2 = -1000;
3836 if (PL_regkind[(U8)OP(text_node)] == REF) {
3840 else { c1 = (U8)*STRING(text_node); }
3842 if (OP(text_node) == EXACTF || OP(text_node) == REFF)
3844 else if (OP(text_node) == EXACTFL || OP(text_node) == REFFL)
3845 c2 = PL_fold_locale[c1];
3855 while (matches >= ln) {
3856 /* If it could work, try it. */
3858 UCHARAT(PL_reginput) == c1 ||
3859 UCHARAT(PL_reginput) == c2)