/* regexec.c */ /* * "One Ring to rule them all, One Ring to find them..." */ /* This file contains functions for executing a regular expression. See * also regcomp.c which funnily enough, contains functions for compiling * a regular expression. * * This file is also copied at build time to ext/re/re_exec.c, where * it's built with -DPERL_EXT_RE_BUILD -DPERL_EXT_RE_DEBUG -DPERL_EXT. * This causes the main functions to be compiled under new names and with * debugging support added, which makes "use re 'debug'" work. */ /* NOTE: this is derived from Henry Spencer's regexp code, and should not * confused with the original package (see point 3 below). Thanks, Henry! */ /* Additional note: this code is very heavily munged from Henry's version * in places. In some spots I've traded clarity for efficiency, so don't * blame Henry for some of the lack of readability. */ /* The names of the functions have been changed from regcomp and * regexec to pregcomp and pregexec in order to avoid conflicts * with the POSIX routines of the same names. */ #ifdef PERL_EXT_RE_BUILD /* need to replace pregcomp et al, so enable that */ # ifndef PERL_IN_XSUB_RE # define PERL_IN_XSUB_RE # endif /* need access to debugger hooks */ # if defined(PERL_EXT_RE_DEBUG) && !defined(DEBUGGING) # define DEBUGGING # endif #endif #ifdef PERL_IN_XSUB_RE /* We *really* need to overwrite these symbols: */ # define Perl_regexec_flags my_regexec # define Perl_regdump my_regdump # define Perl_regprop my_regprop # define Perl_re_intuit_start my_re_intuit_start /* *These* symbols are masked to allow static link. */ # define Perl_pregexec my_pregexec # define Perl_reginitcolors my_reginitcolors # define Perl_regclass_swash my_regclass_swash # define PERL_NO_GET_CONTEXT #endif /* * pregcomp and pregexec -- regsub and regerror are not used in perl * * Copyright (c) 1986 by University of Toronto. * Written by Henry Spencer. Not derived from licensed software. * * Permission is granted to anyone to use this software for any * purpose on any computer system, and to redistribute it freely, * subject to the following restrictions: * * 1. The author is not responsible for the consequences of use of * this software, no matter how awful, even if they arise * from defects in it. * * 2. The origin of this software must not be misrepresented, either * by explicit claim or by omission. * * 3. Altered versions must be plainly marked as such, and must not * be misrepresented as being the original software. * **** Alterations to Henry's code are... **** **** Copyright (C) 1991, 1992, 1993, 1994, 1995, 1996, 1997, 1998, 1999, **** 2000, 2001, 2002, 2003, 2004, 2005, by Larry Wall and others **** **** You may distribute under the terms of either the GNU General Public **** License or the Artistic License, as specified in the README file. * * Beware that some of this code is subtly aware of the way operator * precedence is structured in regular expressions. Serious changes in * regular-expression syntax might require a total rethink. */ #include "EXTERN.h" #define PERL_IN_REGEXEC_C #include "perl.h" #include "regcomp.h" #define RF_tainted 1 /* tainted information used? */ #define RF_warned 2 /* warned about big count? */ #define RF_evaled 4 /* Did an EVAL with setting? */ #define RF_utf8 8 /* String contains multibyte chars? */ #define UTF ((PL_reg_flags & RF_utf8) != 0) #define RS_init 1 /* eval environment created */ #define RS_set 2 /* replsv value is set */ #ifndef STATIC #define STATIC static #endif #define REGINCLASS(p,c) (ANYOF_FLAGS(p) ? reginclass(p,c,0,0) : ANYOF_BITMAP_TEST(p,*(c))) /* * Forwards. */ #define CHR_SVLEN(sv) (do_utf8 ? sv_len_utf8(sv) : SvCUR(sv)) #define CHR_DIST(a,b) (PL_reg_match_utf8 ? utf8_distance(a,b) : a - b) #define reghop_c(pos,off) ((char*)reghop((U8*)pos, off)) #define reghopmaybe_c(pos,off) ((char*)reghopmaybe((U8*)pos, off)) #define HOP(pos,off) (PL_reg_match_utf8 ? reghop((U8*)pos, off) : (U8*)(pos + off)) #define HOPMAYBE(pos,off) (PL_reg_match_utf8 ? reghopmaybe((U8*)pos, off) : (U8*)(pos + off)) #define HOPc(pos,off) ((char*)HOP(pos,off)) #define HOPMAYBEc(pos,off) ((char*)HOPMAYBE(pos,off)) #define HOPBACK(pos, off) ( \ (PL_reg_match_utf8) \ ? reghopmaybe((U8*)pos, -off) \ : (pos - off >= PL_bostr) \ ? (U8*)(pos - off) \ : (U8*)NULL \ ) #define HOPBACKc(pos, off) (char*)HOPBACK(pos, off) #define reghop3_c(pos,off,lim) ((char*)reghop3((U8*)pos, off, (U8*)lim)) #define reghopmaybe3_c(pos,off,lim) ((char*)reghopmaybe3((U8*)pos, off, (U8*)lim)) #define HOP3(pos,off,lim) (PL_reg_match_utf8 ? reghop3((U8*)pos, off, (U8*)lim) : (U8*)(pos + off)) #define HOPMAYBE3(pos,off,lim) (PL_reg_match_utf8 ? reghopmaybe3((U8*)pos, off, (U8*)lim) : (U8*)(pos + off)) #define HOP3c(pos,off,lim) ((char*)HOP3(pos,off,lim)) #define HOPMAYBE3c(pos,off,lim) ((char*)HOPMAYBE3(pos,off,lim)) #define LOAD_UTF8_CHARCLASS(class,str) STMT_START { \ if (!CAT2(PL_utf8_,class)) { bool ok; ENTER; save_re_context(); ok=CAT2(is_utf8_,class)((const U8*)str); assert(ok); LEAVE; } } STMT_END #define LOAD_UTF8_CHARCLASS_ALNUM() LOAD_UTF8_CHARCLASS(alnum,"a") #define LOAD_UTF8_CHARCLASS_DIGIT() LOAD_UTF8_CHARCLASS(digit,"0") #define LOAD_UTF8_CHARCLASS_SPACE() LOAD_UTF8_CHARCLASS(space," ") #define LOAD_UTF8_CHARCLASS_MARK() LOAD_UTF8_CHARCLASS(mark, "\xcd\x86") /* for use after a quantifier and before an EXACT-like node -- japhy */ #define JUMPABLE(rn) ( \ OP(rn) == OPEN || OP(rn) == CLOSE || OP(rn) == EVAL || \ OP(rn) == SUSPEND || OP(rn) == IFMATCH || \ OP(rn) == PLUS || OP(rn) == MINMOD || \ (PL_regkind[(U8)OP(rn)] == CURLY && ARG1(rn) > 0) \ ) #define HAS_TEXT(rn) ( \ PL_regkind[(U8)OP(rn)] == EXACT || PL_regkind[(U8)OP(rn)] == REF \ ) /* Search for mandatory following text node; for lookahead, the text must follow but for lookbehind (rn->flags != 0) we skip to the next step. */ #define FIND_NEXT_IMPT(rn) STMT_START { \ while (JUMPABLE(rn)) \ if (OP(rn) == SUSPEND || PL_regkind[(U8)OP(rn)] == CURLY) \ rn = NEXTOPER(NEXTOPER(rn)); \ else if (OP(rn) == PLUS) \ rn = NEXTOPER(rn); \ else if (OP(rn) == IFMATCH) \ rn = (rn->flags == 0) ? NEXTOPER(NEXTOPER(rn)) : rn + ARG(rn); \ else rn += NEXT_OFF(rn); \ } STMT_END static void restore_pos(pTHX_ void *arg); STATIC CHECKPOINT S_regcppush(pTHX_ I32 parenfloor) 1215116 { 1215116 const int retval = PL_savestack_ix; #define REGCP_PAREN_ELEMS 4 1215116 const int paren_elems_to_push = (PL_regsize - parenfloor) * REGCP_PAREN_ELEMS; 1215116 int p; 1215116 if (paren_elems_to_push < 0) ###### Perl_croak(aTHX_ "panic: paren_elems_to_push < 0"); #define REGCP_OTHER_ELEMS 6 1215116 SSGROW(paren_elems_to_push + REGCP_OTHER_ELEMS); 1393949 for (p = PL_regsize; p > parenfloor; p--) { /* REGCP_PARENS_ELEMS are pushed per pairs of parentheses. */ 178833 SSPUSHINT(PL_regendp[p]); 178833 SSPUSHINT(PL_regstartp[p]); 178833 SSPUSHPTR(PL_reg_start_tmp[p]); 178833 SSPUSHINT(p); } /* REGCP_OTHER_ELEMS are pushed in any case, parentheses or no. */ 1215116 SSPUSHINT(PL_regsize); 1215116 SSPUSHINT(*PL_reglastparen); 1215116 SSPUSHINT(*PL_reglastcloseparen); 1215116 SSPUSHPTR(PL_reginput); #define REGCP_FRAME_ELEMS 2 /* REGCP_FRAME_ELEMS are part of the REGCP_OTHER_ELEMS and * are needed for the regexp context stack bookkeeping. */ 1215116 SSPUSHINT(paren_elems_to_push + REGCP_OTHER_ELEMS - REGCP_FRAME_ELEMS); 1215116 SSPUSHINT(SAVEt_REGCONTEXT); /* Magic cookie. */ 1215116 return retval; } /* These are needed since we do not localize EVAL nodes: */ # define REGCP_SET(cp) DEBUG_EXECUTE_r(PerlIO_printf(Perl_debug_log, \ " Setting an EVAL scope, savestack=%"IVdf"\n", \ (IV)PL_savestack_ix)); cp = PL_savestack_ix # define REGCP_UNWIND(cp) DEBUG_EXECUTE_r(cp != PL_savestack_ix ? \ PerlIO_printf(Perl_debug_log, \ " Clearing an EVAL scope, savestack=%"IVdf"..%"IVdf"\n", \ (IV)(cp), (IV)PL_savestack_ix) : 0); regcpblow(cp) STATIC char * S_regcppop(pTHX) 679222 { 679222 I32 i; 679222 U32 paren = 0; 679222 char *input; 679222 GET_RE_DEBUG_FLAGS_DECL; /* Pop REGCP_OTHER_ELEMS before the parentheses loop starts. */ 679222 i = SSPOPINT; 679222 assert(i == SAVEt_REGCONTEXT); /* Check that the magic cookie is there. */ 679222 i = SSPOPINT; /* Parentheses elements to pop. */ 679222 input = (char *) SSPOPPTR; 679222 *PL_reglastcloseparen = SSPOPINT; 679222 *PL_reglastparen = SSPOPINT; 679222 PL_regsize = SSPOPINT; /* Now restore the parentheses context. */ 799840 for (i -= (REGCP_OTHER_ELEMS - REGCP_FRAME_ELEMS); i > 0; i -= REGCP_PAREN_ELEMS) { 120618 I32 tmps; 120618 paren = (U32)SSPOPINT; 120618 PL_reg_start_tmp[paren] = (char *) SSPOPPTR; 120618 PL_regstartp[paren] = SSPOPINT; 120618 tmps = SSPOPINT; 120618 if (paren <= *PL_reglastparen) 23876 PL_regendp[paren] = tmps; DEBUG_EXECUTE_r( PerlIO_printf(Perl_debug_log, " restoring \\%"UVuf" to %"IVdf"(%"IVdf")..%"IVdf"%s\n", (UV)paren, (IV)PL_regstartp[paren], (IV)(PL_reg_start_tmp[paren] - PL_bostr), (IV)PL_regendp[paren], (paren > *PL_reglastparen ? "(no)" : "")); 120618 ); } DEBUG_EXECUTE_r( if ((I32)(*PL_reglastparen + 1) <= PL_regnpar) { PerlIO_printf(Perl_debug_log, " restoring \\%"IVdf"..\\%"IVdf" to undef\n", (IV)(*PL_reglastparen + 1), (IV)PL_regnpar); } 679222 ); #if 1 /* It would seem that the similar code in regtry() * already takes care of this, and in fact it is in * a better location to since this code can #if 0-ed out * but the code in regtry() is needed or otherwise tests * requiring null fields (pat.t#187 and split.t#{13,14} * (as of patchlevel 7877) will fail. Then again, * this code seems to be necessary or otherwise * building DynaLoader will fail: * "Error: '*' not in typemap in DynaLoader.xs, line 164" * --jhi */ 1743941 for (paren = *PL_reglastparen + 1; (I32)paren <= PL_regnpar; paren++) { 1064719 if ((I32)paren > PL_regsize) 988140 PL_regstartp[paren] = -1; 1064719 PL_regendp[paren] = -1; } #endif 679222 return input; } STATIC char * S_regcp_set_to(pTHX_ I32 ss) 5544 { 5544 const I32 tmp = PL_savestack_ix; 5544 PL_savestack_ix = ss; 5544 regcppop(); 5544 PL_savestack_ix = tmp; 5544 return Nullch; } typedef struct re_cc_state { I32 ss; regnode *node; struct re_cc_state *prev; CURCUR *cc; regexp *re; } re_cc_state; #define regcpblow(cp) LEAVE_SCOPE(cp) /* Ignores regcppush()ed data. */ #define TRYPAREN(paren, n, input) { \ if (paren) { \ if (n) { \ PL_regstartp[paren] = HOPc(input, -1) - PL_bostr; \ PL_regendp[paren] = input - PL_bostr; \ } \ else \ PL_regendp[paren] = -1; \ } \ if (regmatch(next)) \ sayYES; \ if (paren && n) \ PL_regendp[paren] = -1; \ } /* * pregexec and friends */ /* - pregexec - match a regexp against a string */ I32 Perl_pregexec(pTHX_ register regexp *prog, char *stringarg, register char *strend, char *strbeg, I32 minend, SV *screamer, U32 nosave) /* strend: pointer to null at end of string */ /* strbeg: real beginning of string */ /* minend: end of match must be >=minend after stringarg. */ /* nosave: For optimizations. */ ###### { ###### return regexec_flags(prog, stringarg, strend, strbeg, minend, screamer, NULL, nosave ? 0 : REXEC_COPY_STR); } STATIC void S_cache_re(pTHX_ regexp *prog) 11716568 { 11716568 PL_regprecomp = prog->precomp; /* Needed for FAIL. */ #ifdef DEBUGGING 11716568 PL_regprogram = prog->program; #endif 11716568 PL_regnpar = prog->nparens; 11716568 PL_regdata = prog->data; 11716568 PL_reg_re = prog; } /* * Need to implement the following flags for reg_anch: * * USE_INTUIT_NOML - Useful to call re_intuit_start() first * USE_INTUIT_ML * INTUIT_AUTORITATIVE_NOML - Can trust a positive answer * INTUIT_AUTORITATIVE_ML * INTUIT_ONCE_NOML - Intuit can match in one location only. * INTUIT_ONCE_ML * * Another flag for this function: SECOND_TIME (so that float substrs * with giant delta may be not rechecked). */ /* Assumptions: if ANCH_GPOS, then strpos is anchored. XXXX Check GPOS logic */ /* If SCREAM, then SvPVX_const(sv) should be compatible with strpos and strend. Otherwise, only SvCUR(sv) is used to get strbeg. */ /* XXXX We assume that strpos is strbeg unless sv. */ /* XXXX Some places assume that there is a fixed substring. An update may be needed if optimizer marks as "INTUITable" RExen without fixed substrings. Similarly, it is assumed that lengths of all the strings are no more than minlen, thus they cannot come from lookahead. (Or minlen should take into account lookahead.) */ /* A failure to find a constant substring means that there is no need to make an expensive call to REx engine, thus we celebrate a failure. Similarly, finding a substring too deep into the string means that less calls to regtry() should be needed. REx compiler's optimizer found 4 possible hints: a) Anchored substring; b) Fixed substring; c) Whether we are anchored (beginning-of-line or \G); d) First node (of those at offset 0) which may distingush positions; We use a)b)d) and multiline-part of c), and try to find a position in the string which does not contradict any of them. */ /* Most of decisions we do here should have been done at compile time. The nodes of the REx which we used for the search should have been deleted from the finite automaton. */ char * Perl_re_intuit_start(pTHX_ regexp *prog, SV *sv, char *strpos, char *strend, U32 flags, re_scream_pos_data *data) 4555374 { 4555374 register I32 start_shift = 0; /* Should be nonnegative! */ 4555374 register I32 end_shift = 0; 4555374 register char *s; 4555374 register SV *check; 4555374 char *strbeg; 4555374 char *t; 4555374 const int do_utf8 = sv ? SvUTF8(sv) : 0; /* if no sv we have to assume bytes */ 4555374 I32 ml_anch; 4555374 register char *other_last = Nullch; /* other substr checked before this */ 4555374 char *check_at = Nullch; /* check substr found at this pos */ 4555374 const I32 multiline = prog->reganch & PMf_MULTILINE; #ifdef DEBUGGING 4555374 char *i_strpos = strpos; 4555374 SV *dsv = PERL_DEBUG_PAD_ZERO(0); #endif 4555374 GET_RE_DEBUG_FLAGS_DECL; 4555374 RX_MATCH_UTF8_set(prog,do_utf8); 4555374 if (prog->reganch & ROPT_UTF8) { DEBUG_EXECUTE_r(PerlIO_printf(Perl_debug_log, 2085 "UTF-8 regex...\n")); 2085 PL_reg_flags |= RF_utf8; } DEBUG_EXECUTE_r({ const char *s = PL_reg_match_utf8 ? sv_uni_display(dsv, sv, 60, UNI_DISPLAY_REGEX) : strpos; const int len = PL_reg_match_utf8 ? strlen(s) : strend - strpos; if (!PL_colorset) reginitcolors(); if (PL_reg_match_utf8) DEBUG_EXECUTE_r(PerlIO_printf(Perl_debug_log, "UTF-8 target...\n")); PerlIO_printf(Perl_debug_log, "%sGuessing start of match, REx%s \"%s%.60s%s%s\" against \"%s%.*s%s%s\"...\n", PL_colors[4], PL_colors[5], PL_colors[0], prog->precomp, PL_colors[1], (strlen(prog->precomp) > 60 ? "..." : ""), PL_colors[0], (int)(len > 60 ? 60 : len), s, PL_colors[1], (len > 60 ? "..." : "") ); 4555374 }); /* CHR_DIST() would be more correct here but it makes things slow. */ 4555374 if (prog->minlen > strend - strpos) { DEBUG_EXECUTE_r(PerlIO_printf(Perl_debug_log, 93641 "String too short... [re_intuit_start]\n")); ###### goto fail; } 4461733 strbeg = (sv && SvPOK(sv)) ? strend - SvCUR(sv) : strpos; 4461733 PL_regeol = strend; 4461733 if (do_utf8) { 4977 if (!prog->check_utf8 && prog->check_substr) 89 to_utf8_substr(prog); 4977 check = prog->check_utf8; } else { 4456756 if (!prog->check_substr && prog->check_utf8) 19 to_byte_substr(prog); 4456756 check = prog->check_substr; } 4461733 if (check == &PL_sv_undef) { DEBUG_EXECUTE_r(PerlIO_printf(Perl_debug_log, 2 "Non-utf string cannot match utf check string\n")); ###### goto fail; } 4461731 if (prog->reganch & ROPT_ANCH) { /* Match at beg-of-str or after \n */ 2624230 ml_anch = !( (prog->reganch & ROPT_ANCH_SINGLE) || ( (prog->reganch & ROPT_ANCH_BOL) && !multiline ) ); /* Check after \n? */ 2624230 if (!ml_anch) { 2551524 if ( !(prog->reganch & (ROPT_ANCH_GPOS /* Checked by the caller */ | ROPT_IMPLICIT)) /* not a real BOL */ /* SvCUR is not set on references: SvRV and SvPVX_const overlap */ && sv && !SvROK(sv) && (strpos != strbeg)) { 14 DEBUG_EXECUTE_r(PerlIO_printf(Perl_debug_log, "Not at start...\n")); ###### goto fail; } 2551510 if (prog->check_offset_min == prog->check_offset_max && !(prog->reganch & ROPT_CANY_SEEN)) { /* Substring at constant offset from beg-of-str... */ 1283680 I32 slen; 1283680 s = HOP3c(strpos, prog->check_offset_min, strend); 1283680 if (SvTAIL(check)) { 17998 slen = SvCUR(check); /* >= 1 */ 17998 if ( strend - s > slen || strend - s < slen - 1 || (strend - s == slen && strend[-1] != '\n')) { 3142 DEBUG_EXECUTE_r(PerlIO_printf(Perl_debug_log, "String too long...\n")); ###### goto fail_finish; } /* Now should match s[0..slen-2] */ 14856 slen--; 14856 if (slen && (*SvPVX_const(check) != *s || (slen > 1 && memNE(SvPVX_const(check), s, slen)))) { report_neq: 1078513 DEBUG_EXECUTE_r(PerlIO_printf(Perl_debug_log, "String not equal...\n")); ###### goto fail_finish; } } 1265682 else if (*SvPVX_const(check) != *s || ((slen = SvCUR(check)) > 1 && memNE(SvPVX_const(check), s, slen))) 1340536 goto report_neq; 1340536 goto success_at_start; } } /* Match is anchored, but substr is not anchored wrt beg-of-str. */ 1340536 s = strpos; 1340536 start_shift = prog->check_offset_min; /* okay to underestimate on CC */ 1340536 end_shift = prog->minlen - start_shift - CHR_SVLEN(check) + (SvTAIL(check) != 0); 1340536 if (!ml_anch) { 1267830 const I32 end = prog->check_offset_max + CHR_SVLEN(check) 1267830 - (SvTAIL(check) != 0); 1267830 const I32 eshift = CHR_DIST((U8*)strend, (U8*)s) - end; 1267830 if (end_shift < eshift) 120985 end_shift = eshift; } } else { /* Can match at random position */ 1837501 ml_anch = 0; 1837501 s = strpos; 1837501 start_shift = prog->check_offset_min; /* okay to underestimate on CC */ /* Should be nonnegative! */ 1837501 end_shift = prog->minlen - start_shift - CHR_SVLEN(check) + (SvTAIL(check) != 0); } #ifdef DEBUGGING /* 7/99: reports of failure (with the older version) */ 3178037 if (end_shift < 0) ###### Perl_croak(aTHX_ "panic: end_shift"); #endif restart: /* Find a possible match in the region s..strend by looking for the "check" substring in the region corrected by start/end_shift. */ 3193413 if (flags & REXEC_SCREAM) { 17 I32 p = -1; /* Internal iterator of scream. */ 17 I32 * const pp = data ? data->scream_pos : &p; 17 if (PL_screamfirst[BmRARE(check)] >= 0 || ( BmRARE(check) == '\n' && (BmPREVIOUS(check) == SvCUR(check) - 1) && SvTAIL(check) )) 17 s = screaminstr(sv, check, start_shift + (s - strbeg), end_shift, pp, 0); else 17 goto fail_finish; /* we may be pointing at the wrong string */ 17 if (s && RX_MATCH_COPIED(prog)) 4 s = strbeg + (s - SvPVX_const(sv)); 17 if (data) 2 *data->scream_olds = s; } 3193396 else if (prog->reganch & ROPT_CANY_SEEN) 3735 s = fbm_instr((U8*)(s + start_shift), (U8*)(strend - end_shift), check, multiline ? FBMrf_MULTILINE : 0); else 3189661 s = fbm_instr(HOP3(s, start_shift, strend), HOP3(strend, -end_shift, strbeg), check, multiline ? FBMrf_MULTILINE : 0); /* Update the count-of-usability, remove useless subpatterns, unshift s. */ DEBUG_EXECUTE_r(PerlIO_printf(Perl_debug_log, "%s %s substr \"%s%.*s%s\"%s%s", (s ? "Found" : "Did not find"), (check == (do_utf8 ? prog->anchored_utf8 : prog->anchored_substr) ? "anchored" : "floating"), PL_colors[0], (int)(SvCUR(check) - (SvTAIL(check)!=0)), SvPVX_const(check), PL_colors[1], (SvTAIL(check) ? "$" : ""), 3193413 (s ? " at offset " : "...\n") ) ); 3193413 if (!s) 1255756 goto fail_finish; 1937657 check_at = s; /* Finish the diagnostic message */ 1937657 DEBUG_EXECUTE_r(PerlIO_printf(Perl_debug_log, "%ld...\n", (long)(s - i_strpos)) ); /* Got a candidate. Check MBOL anchoring, and the *other* substr. Start with the other substr. XXXX no SCREAM optimization yet - and a very coarse implementation XXXX /ttx+/ results in anchored="ttx", floating="x". floating will *always* match. Probably should be marked during compile... Probably it is right to do no SCREAM here... */ 1937657 if (do_utf8 ? (prog->float_utf8 && prog->anchored_utf8) : (prog->float_substr && prog->anchored_substr)) { /* Take into account the "other" substring. */ /* XXXX May be hopelessly wrong for UTF... */ 435932 if (!other_last) 435406 other_last = strpos; 435932 if (check == (do_utf8 ? prog->float_utf8 : prog->float_substr)) { do_other_anchored: { 190642 char *last = HOP3c(s, -start_shift, strbeg), *last1, *last2; 190642 char *s1 = s; 190642 SV* must; 190642 t = s - prog->check_offset_max; 190642 if (s - strpos > prog->check_offset_max /* signed-corrected t > strpos */ && (!do_utf8 || ((t = reghopmaybe3_c(s, -(prog->check_offset_max), strpos)) && t > strpos))) /* EMPTY */; else 190571 t = strpos; 190642 t = HOP3c(t, prog->anchored_offset, strend); 190642 if (t < other_last) /* These positions already checked */ 452 t = other_last; 190642 last2 = last1 = HOP3c(strend, -prog->minlen, strbeg); 190642 if (last < last1) 188328 last1 = last; /* XXXX It is not documented what units *_offsets are in. Assume bytes. */ /* On end-of-str: see comment below. */ 190642 must = do_utf8 ? prog->anchored_utf8 : prog->anchored_substr; 190642 if (must == &PL_sv_undef) { 1 s = (char*)NULL; 1 DEBUG_EXECUTE_r(must = prog->anchored_utf8); /* for debug */ } else 190641 s = fbm_instr( (unsigned char*)t, HOP3(HOP3(last1, prog->anchored_offset, strend) + SvCUR(must), -(SvTAIL(must)!=0), strbeg), must, multiline ? FBMrf_MULTILINE : 0 ); DEBUG_EXECUTE_r(PerlIO_printf(Perl_debug_log, "%s anchored substr \"%s%.*s%s\"%s", (s ? "Found" : "Contradicts"), PL_colors[0], (int)(SvCUR(must) - (SvTAIL(must)!=0)), SvPVX_const(must), 190642 PL_colors[1], (SvTAIL(must) ? "$" : ""))); 190642 if (!s) { 1977 if (last1 >= last2) { DEBUG_EXECUTE_r(PerlIO_printf(Perl_debug_log, 130 ", giving up...\n")); ###### goto fail_finish; } DEBUG_EXECUTE_r(PerlIO_printf(Perl_debug_log, ", trying floating at offset %ld...\n", 1847 (long)(HOP3c(s1, 1, strend) - i_strpos))); 1847 other_last = HOP3c(last1, prog->anchored_offset+1, strend); 1847 s = HOP3c(last, 1, strend); 1847 goto restart; } else { DEBUG_EXECUTE_r(PerlIO_printf(Perl_debug_log, " at offset %ld...\n", 188665 (long)(s - i_strpos))); 188665 t = HOP3c(s, -prog->anchored_offset, strbeg); 188665 other_last = HOP3c(s, 1, strend); 188665 s = s1; 188665 if (t == strpos) 14695 goto try_at_start; 245349 goto try_at_offset; } } } else { /* Take into account the floating substring. */ 245349 char *last, *last1; 245349 char *s1 = s; 245349 SV* must; 245349 t = HOP3c(s, -start_shift, strbeg); 245349 last1 = last = HOP3c(strend, -prog->minlen + prog->float_min_offset, strbeg); 245349 if (CHR_DIST((U8*)last, (U8*)t) > prog->float_max_offset) 189 last = HOP3c(t, prog->float_max_offset, strend); 245349 s = HOP3c(t, prog->float_min_offset, strend); 245349 if (s < other_last) 10 s = other_last; /* XXXX It is not documented what units *_offsets are in. Assume bytes. */ 245349 must = do_utf8 ? prog->float_utf8 : prog->float_substr; /* fbm_instr() takes into account exact value of end-of-str if the check is SvTAIL(ed). Since false positives are OK, and end-of-str is not later than strend we are OK. */ 245349 if (must == &PL_sv_undef) { 1 s = (char*)NULL; 1 DEBUG_EXECUTE_r(must = prog->float_utf8); /* for debug message */ } else 245348 s = fbm_instr((unsigned char*)s, (unsigned char*)last + SvCUR(must) - (SvTAIL(must)!=0), must, multiline ? FBMrf_MULTILINE : 0); DEBUG_EXECUTE_r(PerlIO_printf(Perl_debug_log, "%s floating substr \"%s%.*s%s\"%s", (s ? "Found" : "Contradicts"), PL_colors[0], (int)(SvCUR(must) - (SvTAIL(must)!=0)), SvPVX_const(must), 245349 PL_colors[1], (SvTAIL(must) ? "$" : ""))); 245349 if (!s) { 1367 if (last1 == last) { DEBUG_EXECUTE_r(PerlIO_printf(Perl_debug_log, 1289 ", giving up...\n")); ###### goto fail_finish; } DEBUG_EXECUTE_r(PerlIO_printf(Perl_debug_log, ", trying anchored starting at offset %ld...\n", 78 (long)(s1 + 1 - i_strpos))); 78 other_last = last; 78 s = HOP3c(t, 1, strend); 78 goto restart; } else { DEBUG_EXECUTE_r(PerlIO_printf(Perl_debug_log, " at offset %ld...\n", 243982 (long)(s - i_strpos))); 243982 other_last = s; /* Fix this later. --Hugo */ 243982 s = s1; 243982 if (t == strpos) 5141 goto try_at_start; 1501725 goto try_at_offset; } } } 1501725 t = s - prog->check_offset_max; 1501725 if (s - strpos > prog->check_offset_max /* signed-corrected t > strpos */ && (!do_utf8 || ((t = reghopmaybe3_c(s, -prog->check_offset_max, strpos)) && t > strpos))) { /* Fixed substring is found far enough so that the match cannot start at strpos. */ try_at_offset: 762180 if (ml_anch && t[-1] != '\n') { /* Eventually fbm_*() should handle this, but often anchored_offset is not 0, so this check will not be wasted. */ /* XXXX In the code below we prefer to look for "^" even in presence of anchored substrings. And we search even beyond the found float position. These pessimizations are historical artefacts only. */ find_anchor: 996765 while (t < strend - prog->minlen) { 996601 if (*t == '\n') { 20039 if (t < check_at - prog->check_offset_min) { 12972 if (do_utf8 ? prog->anchored_utf8 : prog->anchored_substr) { /* Since we moved from the found position, we definitely contradict the found anchored substr. Due to the above check we do not contradict "check" substr. Thus we can arrive here only if check substr is float. Redo checking for "other"=="fixed". */ 1 strpos = t + 1; DEBUG_EXECUTE_r(PerlIO_printf(Perl_debug_log, "Found /%s^%s/m at offset %ld, rescanning for anchored from offset %ld...\n", 1 PL_colors[0], PL_colors[1], (long)(strpos - i_strpos), (long)(strpos - i_strpos + prog->anchored_offset))); ###### goto do_other_anchored; } /* We don't contradict the found floating substring. */ /* XXXX Why not check for STCLASS? */ 12971 s = t + 1; DEBUG_EXECUTE_r(PerlIO_printf(Perl_debug_log, "Found /%s^%s/m at offset %ld...\n", 12971 PL_colors[0], PL_colors[1], (long)(s - i_strpos))); ###### goto set_useful; } /* Position contradicts check-string */ /* XXXX probably better to look for check-string than for "\n", so one should lower the limit for t? */ DEBUG_EXECUTE_r(PerlIO_printf(Perl_debug_log, "Found /%s^%s/m, restarting lookup for check-string at offset %ld...\n", 7067 PL_colors[0], PL_colors[1], (long)(t + 1 - i_strpos))); 7067 other_last = strpos = s = t + 1; 7067 goto restart; } 976562 t++; } DEBUG_EXECUTE_r(PerlIO_printf(Perl_debug_log, "Did not find /%s^%s/m...\n", 164 PL_colors[0], PL_colors[1])); ###### goto fail_finish; } else { DEBUG_EXECUTE_r(PerlIO_printf(Perl_debug_log, "Starting position does not contradict /%s^%s/m...\n", 744622 PL_colors[0], PL_colors[1])); } 744622 s = t; set_useful: 757593 ++BmUSEFUL(do_utf8 ? prog->check_utf8 : prog->check_substr); /* hooray/5 */ } else { /* The found string does not prohibit matching at strpos, - no optimization of calling REx engine can be performed, unless it was an MBOL and we are not after MBOL, or a future STCLASS check will fail this. */ try_at_start: /* Even in this situation we may use MBOL flag if strpos is offset wrt the start of the string. */ 1183035 if (ml_anch && sv && !SvROK(sv) /* See prev comment on SvROK */ && (strpos != strbeg) && strpos[-1] != '\n' /* May be due to an implicit anchor of m{.*foo} */ && !(prog->reganch & ROPT_IMPLICIT)) { 2645 t = strpos; 2645 goto find_anchor; } DEBUG_EXECUTE_r( if (ml_anch) PerlIO_printf(Perl_debug_log, "Position at offset %ld does not contradict /%s^%s/m...\n", (long)(strpos - i_strpos), PL_colors[0], PL_colors[1]); 1180390 ); success_at_start: 1382415 if (!(prog->reganch & ROPT_NAUGHTY) /* XXXX If strpos moved? */ && (do_utf8 ? ( prog->check_utf8 /* Could be deleted already */ && --BmUSEFUL(prog->check_utf8) < 0 && (prog->check_utf8 == prog->float_utf8) ) : ( prog->check_substr /* Could be deleted already */ && --BmUSEFUL(prog->check_substr) < 0 && (prog->check_substr == prog->float_substr) ))) { /* If flags & SOMETHING - do not do it many times on the same match */ 600 DEBUG_EXECUTE_r(PerlIO_printf(Perl_debug_log, "... Disabling check substring...\n")); 600 SvREFCNT_dec(do_utf8 ? prog->check_utf8 : prog->check_substr); 600 if (do_utf8 ? prog->check_substr : prog->check_utf8) 5 SvREFCNT_dec(do_utf8 ? prog->check_substr : prog->check_utf8); 600 prog->check_substr = prog->check_utf8 = Nullsv; /* disable */ 600 prog->float_substr = prog->float_utf8 = Nullsv; /* clear */ 600 check = Nullsv; /* abort */ 600 s = strpos; /* XXXX This is a remnant of the old implementation. It looks wasteful, since now INTUIT can use many other heuristics. */ 600 prog->reganch &= ~RE_USE_INTUIT; } else 1381815 s = strpos; } /* Last resort... */ /* XXXX BmUSEFUL already changed, maybe multiple change is meaningful... */ 2140008 if (prog->regstclass) { /* minlen == 0 is possible if regstclass is \b or \B, and the fixed substr is ''$. Since minlen is already taken into account, s+1 is before strend; accidentally, minlen >= 1 guaranties no false positives at s + 1 even for \b or \B. But (minlen? 1 : 0) below assumes that regstclass does not come from lookahead... */ /* If regstclass takes bytelength more than 1: If charlength==1, OK. This leaves EXACTF only, which is dealt with in find_byclass(). */ 594787 const U8* str = (U8*)STRING(prog->regstclass); 594787 const int cl_l = (PL_regkind[(U8)OP(prog->regstclass)] == EXACT ? CHR_DIST(str+STR_LEN(prog->regstclass), str) 594787 : 1); 594787 const char * const endpos = (prog->anchored_substr || prog->anchored_utf8 || ml_anch) ? HOP3c(s, (prog->minlen ? cl_l : 0), strend) : (prog->float_substr || prog->float_utf8 ? HOP3c(HOP3c(check_at, -start_shift, strbeg), cl_l, strend) 594787 : strend); 594787 t = s; 594787 cache_re(prog); 594787 s = find_byclass(prog, prog->regstclass, s, endpos, 1); 594787 if (!s) { #ifdef DEBUGGING 33286 const char *what = 0; #endif 33286 if (endpos == strend) { DEBUG_EXECUTE_r( PerlIO_printf(Perl_debug_log, 9478 "Could not match STCLASS...\n") ); ###### goto fail; } DEBUG_EXECUTE_r( PerlIO_printf(Perl_debug_log, 23808 "This position contradicts STCLASS...\n") ); 23808 if ((prog->reganch & ROPT_ANCH) && !ml_anch) 5269 goto fail; /* Contradict one of substrings */ 18539 if (prog->anchored_substr || prog->anchored_utf8) { 862 if ((do_utf8 ? prog->anchored_utf8 : prog->anchored_substr) == check) { 797 DEBUG_EXECUTE_r( what = "anchored" ); hop_and_restart: 7636 s = HOP3c(t, 1, strend); 7636 if (s + start_shift + end_shift > strend) { /* XXXX Should be taken into account earlier? */ DEBUG_EXECUTE_r( PerlIO_printf(Perl_debug_log, 1252 "Could not match STCLASS...\n") ); ###### goto fail; } 6384 if (!check) ###### goto giveup; DEBUG_EXECUTE_r( PerlIO_printf(Perl_debug_log, "Looking for %s substr starting at offset %ld...\n", 6384 what, (long)(s + start_shift - i_strpos)) ); ###### goto restart; } /* Have both, check_string is floating */ 65 if (t + start_shift >= check_at) /* Contradicts floating=check */ 7 goto retry_floating_check; /* Recheck anchored substring, but not floating... */ 58 s = check_at; 58 if (!check) ###### goto giveup; DEBUG_EXECUTE_r( PerlIO_printf(Perl_debug_log, "Looking for anchored substr starting at offset %ld...\n", 58 (long)(other_last - i_strpos)) ); ###### goto do_other_anchored; } /* Another way we could have checked stclass at the current position only: */ 17677 if (ml_anch) { 10845 s = t = t + 1; 10845 if (!check) 2 goto giveup; DEBUG_EXECUTE_r( PerlIO_printf(Perl_debug_log, "Looking for /%s^%s/m starting at offset %ld...\n", 10843 PL_colors[0], PL_colors[1], (long)(t - i_strpos)) ); ###### goto try_at_offset; } 6832 if (!(do_utf8 ? prog->float_utf8 : prog->float_substr)) /* Could have been deleted */ ###### goto fail; /* Check is floating subtring. */ retry_floating_check: 6839 t = check_at - start_shift; 6839 DEBUG_EXECUTE_r( what = "floating" ); ###### goto hop_and_restart; } 561501 if (t != s) { DEBUG_EXECUTE_r(PerlIO_printf(Perl_debug_log, "By STCLASS: moving %ld --> %ld\n", (long)(t - i_strpos), (long)(s - i_strpos)) 151445 ); } else { DEBUG_EXECUTE_r(PerlIO_printf(Perl_debug_log, "Does not contradict STCLASS...\n"); 410056 ); } } giveup: DEBUG_EXECUTE_r(PerlIO_printf(Perl_debug_log, "%s%s:%s match at offset %ld\n", PL_colors[4], (check ? "Guessed" : "Giving up"), 2106724 PL_colors[5], (long)(s - i_strpos)) ); 2106724 return s; fail_finish: /* Substring not found */ 2338994 if (prog->check_substr || prog->check_utf8) /* could be removed already */ 2338994 BmUSEFUL(do_utf8 ? prog->check_utf8 : prog->check_substr) += 5; /* hooray */ fail: DEBUG_EXECUTE_r(PerlIO_printf(Perl_debug_log, "%sMatch rejected by optimizer%s\n", 2448650 PL_colors[4], PL_colors[5])); 2448650 return Nullch; } /* We know what class REx starts with. Try to find this position... */ STATIC char * S_find_byclass(pTHX_ regexp * prog, regnode *c, char *s, const char *strend, I32 norun) 5597636 { dVAR; 5597636 const I32 doevery = (prog->reganch & ROPT_SKIP) == 0; 5597636 char *m; 5597636 STRLEN ln; 5597636 STRLEN lnc; 5597636 register STRLEN uskip; 5597636 unsigned int c1; 5597636 unsigned int c2; 5597636 char *e; 5597636 register I32 tmp = 1; /* Scratch variable? */ 5597636 register const bool do_utf8 = PL_reg_match_utf8; /* We know what class it must start with. */ 5597636 switch (OP(c)) { case ANYOF: 3342502 if (do_utf8) { 1188325 while (s + (uskip = UTF8SKIP(s)) <= strend) { 1188136 if ((ANYOF_FLAGS(c) & ANYOF_UNICODE) || !UTF8_IS_INVARIANT((U8)s[0]) ? reginclass(c, (U8*)s, 0, do_utf8) : REGINCLASS(c, (U8*)s)) { 4142 if (tmp && (norun || regtry(prog, s))) 4 goto got_it; else 4 tmp = doevery; } else 1183994 tmp = 1; 1183998 s += uskip; } } else { 9410736 while (s < strend) { 7602014 STRLEN skip = 1; 7602014 if (REGINCLASS(c, (U8*)s) || (ANYOF_FOLD_SHARP_S(c, s, strend) && /* The assignment of 2 is intentional: * for the folded sharp s, the skip is 2. */ (skip = SHARP_S_SKIP))) { 1971618 if (tmp && (norun || regtry(prog, s))) 442165 goto got_it; else 442165 tmp = doevery; } else 5630396 tmp = 1; 6072561 s += skip; } } 35 break; case CANY: 35 while (s < strend) { 35 if (tmp && (norun || regtry(prog, s))) ###### goto got_it; else ###### tmp = doevery; ###### s++; } 25271 break; case EXACTF: 25271 m = STRING(c); 25271 ln = STR_LEN(c); /* length to match in octets/bytes */ 25271 lnc = (I32) ln; /* length to match in characters */ 25271 if (UTF) { 5925 STRLEN ulen1, ulen2; 5925 U8 *sm = (U8 *) m; 5925 U8 tmpbuf1[UTF8_MAXBYTES_CASE+1]; 5925 U8 tmpbuf2[UTF8_MAXBYTES_CASE+1]; 5925 to_utf8_lower((U8*)m, tmpbuf1, &ulen1); 5925 to_utf8_upper((U8*)m, tmpbuf2, &ulen2); 5925 c1 = utf8n_to_uvchr(tmpbuf1, UTF8_MAXBYTES_CASE, 0, ckWARN(WARN_UTF8) ? 0 : UTF8_ALLOW_ANY); 5925 c2 = utf8n_to_uvchr(tmpbuf2, UTF8_MAXBYTES_CASE, 0, ckWARN(WARN_UTF8) ? 0 : UTF8_ALLOW_ANY); 5925 lnc = 0; 18113 while (sm < ((U8 *) m + ln)) { 12188 lnc++; 12188 sm += UTF8SKIP(sm); } } else { 19346 c1 = *(U8*)m; 19346 c2 = PL_fold[c1]; } 19346 goto do_exactf; case EXACTFL: 1696824 m = STRING(c); 1696824 ln = STR_LEN(c); 1696824 lnc = (I32) ln; 1696824 c1 = *(U8*)m; 1696824 c2 = PL_fold_locale[c1]; do_exactf: 1722095 e = HOP3c(strend, -((I32)lnc), s); 1722095 if (norun && e < s) ###### e = s; /* Due to minlen logic of intuit() */ /* The idea in the EXACTF* cases is to first find the * first character of the EXACTF* node and then, if * necessary, case-insensitively compare the full * text of the node. The c1 and c2 are the first * characters (though in Unicode it gets a bit * more complicated because there are more cases * than just upper and lower: one needs to use * the so-called folding case for case-insensitive * matching (called "loose matching" in Unicode). * ibcmp_utf8() will do just that. */ 1722095 if (do_utf8) { 5931 UV c, f; 5931 U8 tmpbuf [UTF8_MAXBYTES+1]; 5931 STRLEN len, foldlen; 5931 if (c1 == c2) { /* Upper and lower of 1st char are equal - * probably not a "letter". */ 5894 while (s <= e) { 5893 c = utf8n_to_uvchr((U8*)s, UTF8_MAXBYTES, &len, ckWARN(WARN_UTF8) ? 0 : UTF8_ALLOW_ANY); 5893 if ( c == c1 && (ln == len || ibcmp_utf8(s, (char **)0, 0, do_utf8, m, (char **)0, ln, (bool)UTF)) && (norun || regtry(prog, s)) ) 1 goto got_it; else { 1 U8 foldbuf[UTF8_MAXBYTES_CASE+1]; 1 uvchr_to_utf8(tmpbuf, c); 1 f = to_utf8_fold(tmpbuf, foldbuf, &foldlen); 1 if ( f != c && (f == c1 || f == c2) && (ln == foldlen || !ibcmp_utf8((char *) foldbuf, (char **)0, foldlen, do_utf8, m, (char **)0, ln, (bool)UTF)) && (norun || regtry(prog, s)) ) 1 goto got_it; } 1 s += len; } } else { 47 while (s <= e) { 46 c = utf8n_to_uvchr((U8*)s, UTF8_MAXBYTES, &len, ckWARN(WARN_UTF8) ? 0 : UTF8_ALLOW_ANY); /* Handle some of the three Greek sigmas cases. * Note that not all the possible combinations * are handled here: some of them are handled * by the standard folding rules, and some of * them (the character class or ANYOF cases) * are handled during compiletime in * regexec.c:S_regclass(). */ 46 if (c == (UV)UNICODE_GREEK_CAPITAL_LETTER_SIGMA || c == (UV)UNICODE_GREEK_SMALL_LETTER_FINAL_SIGMA) 6 c = (UV)UNICODE_GREEK_SMALL_LETTER_SIGMA; 46 if ( (c == c1 || c == c2) && (ln == len || ibcmp_utf8(s, (char **)0, 0, do_utf8, m, (char **)0, ln, (bool)UTF)) && (norun || regtry(prog, s)) ) 12 goto got_it; else { 12 U8 foldbuf[UTF8_MAXBYTES_CASE+1]; 12 uvchr_to_utf8(tmpbuf, c); 12 f = to_utf8_fold(tmpbuf, foldbuf, &foldlen); 12 if ( f != c && (f == c1 || f == c2) && (ln == foldlen || !ibcmp_utf8((char *) foldbuf, (char **)0, foldlen, do_utf8, m, (char **)0, ln, (bool)UTF)) && (norun || regtry(prog, s)) ) 9 goto got_it; } 9 s += len; } } } else { 1716164 if (c1 == c2) 54779 while (s <= e) { 53393 if ( *(U8*)s == c1 && (ln == 1 || !(OP(c) == EXACTF ? ibcmp(s, m, ln) : ibcmp_locale(s, m, ln))) && (norun || regtry(prog, s)) ) 52919 goto got_it; 52919 s++; } else 1769337 while (s <= e) { 1753178 if ( (*(U8*)s == c1 || *(U8*)s == c2) && (ln == 1 || !(OP(c) == EXACTF ? ibcmp(s, m, ln) : ibcmp_locale(s, m, ln))) && (norun || regtry(prog, s)) ) 55033 goto got_it; 55033 s++; } } 8 break; case BOUNDL: 8 PL_reg_flags |= RF_tainted; /* FALL THROUGH */ case BOUND: 21426 if (do_utf8) { 11 if (s == PL_bostr) 10 tmp = '\n'; else { 1 U8 *r = reghop3((U8*)s, -1, (U8*)PL_bostr); 1 tmp = utf8n_to_uvchr(r, UTF8SKIP(r), 0, 0); } 11 tmp = ((OP(c) == BOUND ? isALNUM_uni(tmp) : isALNUM_LC_uvchr(UNI_TO_NATIVE(tmp))) != 0); 11 LOAD_UTF8_CHARCLASS_ALNUM(); 22 while (s + (uskip = UTF8SKIP(s)) <= strend) { 13 if (tmp == !(OP(c) == BOUND ? swash_fetch(PL_utf8_alnum, (U8*)s, do_utf8) : isALNUM_LC_utf8((U8*)s))) { 11 tmp = !tmp; 11 if ((norun || regtry(prog, s))) 11 goto got_it; } 11 s += uskip; } } else { 21415 tmp = (s != PL_bostr) ? UCHARAT(s - 1) : '\n'; 21415 tmp = ((OP(c) == BOUND ? isALNUM(tmp) : isALNUM_LC(tmp)) != 0); 127818 while (s < strend) { 115967 if (tmp == !(OP(c) == BOUND ? isALNUM(*s) : isALNUM_LC(*s))) { 38613 tmp = !tmp; 38613 if ((norun || regtry(prog, s))) 106403 goto got_it; } 106403 s++; } } 11860 if ((!prog->minlen && tmp) && (norun || regtry(prog, s))) ###### goto got_it; ###### break; case NBOUNDL: ###### PL_reg_flags |= RF_tainted; /* FALL THROUGH */ case NBOUND: 111 if (do_utf8) { ###### if (s == PL_bostr) ###### tmp = '\n'; else { ###### U8 *r = reghop3((U8*)s, -1, (U8*)PL_bostr); ###### tmp = utf8n_to_uvchr(r, UTF8SKIP(r), 0, 0); } ###### tmp = ((OP(c) == NBOUND ? isALNUM_uni(tmp) : isALNUM_LC_uvchr(UNI_TO_NATIVE(tmp))) != 0); ###### LOAD_UTF8_CHARCLASS_ALNUM(); ###### while (s + (uskip = UTF8SKIP(s)) <= strend) { ###### if (tmp == !(OP(c) == NBOUND ? swash_fetch(PL_utf8_alnum, (U8*)s, do_utf8) : isALNUM_LC_utf8((U8*)s))) ###### tmp = !tmp; ###### else if ((norun || regtry(prog, s))) ###### goto got_it; ###### s += uskip; } } else { 111 tmp = (s != PL_bostr) ? UCHARAT(s - 1) : '\n'; 111 tmp = ((OP(c) == NBOUND ? isALNUM(tmp) : isALNUM_LC(tmp)) != 0); 147 while (s < strend) { 117 if (tmp == !(OP(c) == NBOUND ? isALNUM(*s) : isALNUM_LC(*s))) 36 tmp = !tmp; 81 else if ((norun || regtry(prog, s))) 36 goto got_it; 36 s++; } } 30 if ((!prog->minlen && !tmp) && (norun || regtry(prog, s))) 1214 goto got_it; 1214 break; case ALNUM: 1214 if (do_utf8) { 139 LOAD_UTF8_CHARCLASS_ALNUM(); 239 while (s + (uskip = UTF8SKIP(s)) <= strend) { 143 if (swash_fetch(PL_utf8_alnum, (U8*)s, do_utf8)) { 43 if (tmp && (norun || regtry(prog, s))) ###### goto got_it; else ###### tmp = doevery; } else 100 tmp = 1; 100 s += uskip; } } else { 1821 while (s < strend) { 1351 if (isALNUM(*s)) { 663 if (tmp && (norun || regtry(prog, s))) 58 goto got_it; else 58 tmp = doevery; } else 688 tmp = 1; 746 s++; } } 176394 break; case ALNUML: 176394 PL_reg_flags |= RF_tainted; 176394 if (do_utf8) { ###### while (s + (uskip = UTF8SKIP(s)) <= strend) { ###### if (isALNUM_LC_utf8((U8*)s)) { ###### if (tmp && (norun || regtry(prog, s))) ###### goto got_it; else ###### tmp = doevery; } else ###### tmp = 1; ###### s += uskip; } } else { 275188 while (s < strend) { 176396 if (isALNUM_LC(*s)) { 77602 if (tmp && (norun || regtry(prog, s))) ###### goto got_it; else ###### tmp = doevery; } else 98794 tmp = 1; 98794 s++; } } 124421 break; case NALNUM: 124421 if (do_utf8) { 2 LOAD_UTF8_CHARCLASS_ALNUM(); 15 while (s + (uskip = UTF8SKIP(s)) <= strend) { 13 if (!swash_fetch(PL_utf8_alnum, (U8*)s, do_utf8)) { ###### if (tmp && (norun || regtry(prog, s))) ###### goto got_it; else ###### tmp = doevery; } else 13 tmp = 1; 13 s += uskip; } } else { 930036 while (s < strend) { 820609 if (!isALNUM(*s)) { 14992 if (tmp && (norun || regtry(prog, s))) ###### goto got_it; else ###### tmp = doevery; } else 805617 tmp = 1; 805617 s++; } } 47451 break; case NALNUML: 47451 PL_reg_flags |= RF_tainted; 47451 if (do_utf8) { ###### while (s + (uskip = UTF8SKIP(s)) <= strend) { ###### if (!isALNUM_LC_utf8((U8*)s)) { ###### if (tmp && (norun || regtry(prog, s))) ###### goto got_it; else ###### tmp = doevery; } else ###### tmp = 1; ###### s += uskip; } } else { 72399 while (s < strend) { 47912 if (!isALNUM_LC(*s)) { 22964 if (tmp && (norun || regtry(prog, s))) ###### goto got_it; else ###### tmp = doevery; } else 24948 tmp = 1; 24948 s++; } } 142455 break; case SPACE: 142455 if (do_utf8) { 1 LOAD_UTF8_CHARCLASS_SPACE(); 1 while (s + (uskip = UTF8SKIP(s)) <= strend) { 1 if (*s == ' ' || swash_fetch(PL_utf8_space,(U8*)s, do_utf8)) { 1 if (tmp && (norun || regtry(prog, s))) ###### goto got_it; else ###### tmp = doevery; } else ###### tmp = 1; ###### s += uskip; } } else { 2223340 while (s < strend) { 2162061 if (isSPACE(*s)) { 205662 if (tmp && (norun || regtry(prog, s))) 124487 goto got_it; else 124487 tmp = doevery; } else 1956399 tmp = 1; 2080886 s++; } } 573 break; case SPACEL: 573 PL_reg_flags |= RF_tainted; 573 if (do_utf8) { ###### while (s + (uskip = UTF8SKIP(s)) <= strend) { ###### if (*s == ' ' || isSPACE_LC_utf8((U8*)s)) { ###### if (tmp && (norun || regtry(prog, s))) ###### goto got_it; else ###### tmp = doevery; } else ###### tmp = 1; ###### s += uskip; } } else { 9216 while (s < strend) { 8983 if (isSPACE_LC(*s)) { 1617 if (tmp && (norun || regtry(prog, s))) 1277 goto got_it; else 1277 tmp = doevery; } else 7366 tmp = 1; 8643 s++; } } 14785 break; case NSPACE: 14785 if (do_utf8) { 2 LOAD_UTF8_CHARCLASS_SPACE(); 2 while (s + (uskip = UTF8SKIP(s)) <= strend) { 2 if (!(*s == ' ' || swash_fetch(PL_utf8_space,(U8*)s, do_utf8))) { 2 if (tmp && (norun || regtry(prog, s))) ###### goto got_it; else ###### tmp = doevery; } else ###### tmp = 1; ###### s += uskip; } } else { 28135 while (s < strend) { 27406 if (!isSPACE(*s)) { 14054 if (tmp && (norun || regtry(prog, s))) ###### goto got_it; else ###### tmp = doevery; } else 13352 tmp = 1; 13352 s++; } } 1 break; case NSPACEL: 1 PL_reg_flags |= RF_tainted; 1 if (do_utf8) { ###### while (s + (uskip = UTF8SKIP(s)) <= strend) { ###### if (!(*s == ' ' || isSPACE_LC_utf8((U8*)s))) { ###### if (tmp && (norun || regtry(prog, s))) ###### goto got_it; else ###### tmp = doevery; } else ###### tmp = 1; ###### s += uskip; } } else { 1 while (s < strend) { 1 if (!isSPACE_LC(*s)) { 1 if (tmp && (norun || regtry(prog, s))) ###### goto got_it; else ###### tmp = doevery; } else ###### tmp = 1; ###### s++; } } 2891 break; case DIGIT: 2891 if (do_utf8) { 1 LOAD_UTF8_CHARCLASS_DIGIT(); 1 while (s + (uskip = UTF8SKIP(s)) <= strend) { 1 if (swash_fetch(PL_utf8_digit,(U8*)s, do_utf8)) { 1 if (tmp && (norun || regtry(prog, s))) ###### goto got_it; else ###### tmp = doevery; } else ###### tmp = 1; ###### s += uskip; } } else { 8462 while (s < strend) { 7652 if (isDIGIT(*s)) { 2096 if (tmp && (norun || regtry(prog, s))) 16 goto got_it; else 16 tmp = doevery; } else 5556 tmp = 1; 5572 s++; } } ###### break; case DIGITL: ###### PL_reg_flags |= RF_tainted; ###### if (do_utf8) { ###### while (s + (uskip = UTF8SKIP(s)) <= strend) { ###### if (isDIGIT_LC_utf8((U8*)s)) { ###### if (tmp && (norun || regtry(prog, s))) ###### goto got_it; else ###### tmp = doevery; } else ###### tmp = 1; ###### s += uskip; } } else { ###### while (s < strend) { ###### if (isDIGIT_LC(*s)) { ###### if (tmp && (norun || regtry(prog, s))) ###### goto got_it; else ###### tmp = doevery; } else ###### tmp = 1; ###### s++; } } 1282 break; case NDIGIT: 1282 if (do_utf8) { ###### LOAD_UTF8_CHARCLASS_DIGIT(); ###### while (s + (uskip = UTF8SKIP(s)) <= strend) { ###### if (!swash_fetch(PL_utf8_digit,(U8*)s, do_utf8)) { ###### if (tmp && (norun || regtry(prog, s))) ###### goto got_it; else ###### tmp = doevery; } else ###### tmp = 1; ###### s += uskip; } } else { 3465 while (s < strend) { 2709 if (!isDIGIT(*s)) { 526 if (tmp && (norun || regtry(prog, s))) ###### goto got_it; else ###### tmp = doevery; } else 2183 tmp = 1; 2183 s++; } } ###### break; case NDIGITL: ###### PL_reg_flags |= RF_tainted; ###### if (do_utf8) { ###### while (s + (uskip = UTF8SKIP(s)) <= strend) { ###### if (!isDIGIT_LC_utf8((U8*)s)) { ###### if (tmp && (norun || regtry(prog, s))) ###### goto got_it; else ###### tmp = doevery; } else ###### tmp = 1; ###### s += uskip; } } else { ###### while (s < strend) { ###### if (!isDIGIT_LC(*s)) { ###### if (tmp && (norun || regtry(prog, s))) ###### goto got_it; else ###### tmp = doevery; } else ###### tmp = 1; ###### s++; } } ###### break; default: ###### Perl_croak(aTHX_ "panic: unknown regstclass %d", (int)OP(c)); 2135429 break; } 2135429 return 0; got_it: 3462207 return s; } /* - regexec_flags - match a regexp against a string */ I32 Perl_regexec_flags(pTHX_ register regexp *prog, char *stringarg, register char *strend, char *strbeg, I32 minend, SV *sv, void *data, U32 flags) /* strend: pointer to null at end of string */ /* strbeg: real beginning of string */ /* minend: end of match must be >=minend after stringarg. */ /* data: May be used for some additional optimizations. */ /* nosave: For optimizations. */ 11094673 { 11094673 register char *s; 11094673 register regnode *c; 11094673 register char *startpos = stringarg; 11094673 I32 minlen; /* must match at least this many chars */ 11094673 I32 dontbother = 0; /* how many characters not to try at end */ 11094673 I32 end_shift = 0; /* Same for the end. */ /* CC */ 11094673 I32 scream_pos = -1; /* Internal iterator of scream. */ 11094673 char *scream_olds; 11094673 SV* oreplsv = GvSV(PL_replgv); 11094673 const bool do_utf8 = DO_UTF8(sv); 11094673 const I32 multiline = prog->reganch & PMf_MULTILINE; #ifdef DEBUGGING 11094673 SV *dsv0 = PERL_DEBUG_PAD_ZERO(0); 11094673 SV *dsv1 = PERL_DEBUG_PAD_ZERO(1); #endif 11094673 GET_RE_DEBUG_FLAGS_DECL; 11094673 PERL_UNUSED_ARG(data); 11094673 RX_MATCH_UTF8_set(prog,do_utf8); 11094673 PL_regcc = 0; 11094673 cache_re(prog); #ifdef DEBUGGING 11094673 PL_regnarrate = DEBUG_r_TEST; #endif /* Be paranoid... */ 11094673 if (prog == NULL || startpos == NULL) { ###### Perl_croak(aTHX_ "NULL regexp parameter"); 11094673 return 0; } 11094673 minlen = prog->minlen; 11094673 if (strend - startpos < minlen) { DEBUG_EXECUTE_r(PerlIO_printf(Perl_debug_log, 139799 "String too short [regexec_flags]...\n")); ###### goto phooey; } /* Check validity of program. */ 10954874 if (UCHARAT(prog->program) != REG_MAGIC) { ###### Perl_croak(aTHX_ "corrupted regexp program"); } 10954874 PL_reg_flags = 0; 10954874 PL_reg_eval_set = 0; 10954874 PL_reg_maxiter = 0; 10954874 if (prog->reganch & ROPT_UTF8) 7724 PL_reg_flags |= RF_utf8; /* Mark beginning of line for ^ and lookbehind. */ 10954874 PL_regbol = startpos; 10954874 PL_bostr = strbeg; 10954874 PL_reg_sv = sv; /* Mark end of line for $ (and such) */ 10954874 PL_regeol = strend; /* see how far we have to get to not match where we matched before */ 10954874 PL_regtill = startpos+minend; /* We start without call_cc context. */ 10954874 PL_reg_call_cc = 0; /* If there is a "must appear" string, look for it. */ 10954874 s = startpos; 10954874 if (prog->reganch & ROPT_GPOS_SEEN) { /* Need to have PL_reg_ganch */ 974702 MAGIC *mg; 974702 if (flags & REXEC_IGNOREPOS) /* Means: check only at start */ 440733 PL_reg_ganch = startpos; 533969 else if (sv && SvTYPE(sv) >= SVt_PVMG && SvMAGIC(sv) && (mg = mg_find(sv, PERL_MAGIC_regex_global)) && mg->mg_len >= 0) { 47 PL_reg_ganch = strbeg + mg->mg_len; /* Defined pos() */ 47 if (prog->reganch & ROPT_ANCH_GPOS) { 22 if (s > PL_reg_ganch) ###### goto phooey; 22 s = PL_reg_ganch; } } else /* pos() not defined */ 533922 PL_reg_ganch = strbeg; } 10954874 if (!(flags & REXEC_CHECKED) && (prog->check_substr != Nullsv || prog->check_utf8 != Nullsv)) { 696064 re_scream_pos_data d; 696064 d.scream_olds = &scream_olds; 696064 d.scream_pos = &scream_pos; 696064 s = re_intuit_start(prog, sv, s, strend, flags, &d); 696064 if (!s) { 135194 DEBUG_EXECUTE_r(PerlIO_printf(Perl_debug_log, "Not present...\n")); ###### goto phooey; /* not present */ } } DEBUG_EXECUTE_r({ const char * const s0 = UTF ? pv_uni_display(dsv0, (U8*)prog->precomp, prog->prelen, 60, UNI_DISPLAY_REGEX) : prog->precomp; const int len0 = UTF ? SvCUR(dsv0) : prog->prelen; const char * const s1 = do_utf8 ? sv_uni_display(dsv1, sv, 60, UNI_DISPLAY_REGEX) : startpos; const int len1 = do_utf8 ? SvCUR(dsv1) : strend - startpos; if (!PL_colorset) reginitcolors(); PerlIO_printf(Perl_debug_log, "%sMatching REx%s \"%s%*.*s%s%s\" against \"%s%.*s%s%s\"\n", PL_colors[4], PL_colors[5], PL_colors[0], len0, len0, s0, PL_colors[1], len0 > 60 ? "..." : "", PL_colors[0], (int)(len1 > 60 ? 60 : len1), s1, PL_colors[1], (len1 > 60 ? "..." : "") ); 10819680 }); /* Simplest case: anchored match need be tried only once. */ /* [unless only anchor is BOL and multiline is set] */ 10819680 if (prog->reganch & (ROPT_ANCH & ~ROPT_ANCH_GPOS)) { 3521359 if (s == startpos && regtry(prog, startpos)) 1879531 goto got_it; 1641828 else if (multiline || (prog->reganch & ROPT_IMPLICIT) || (prog->reganch & ROPT_ANCH_MBOL)) /* XXXX SBOL? */ { 447852 char *end; 447852 if (minlen) 442132 dontbother = minlen - 1; 447852 end = HOP3c(strend, -dontbother, strbeg) - 1; /* for multiline we only have to try after newlines */ 447852 if (prog->check_substr || prog->check_utf8) { 16950 if (s == startpos) 2094 goto after_try; 19520 while (1) { 19520 if (regtry(prog, s)) 14919 goto got_it; after_try: 6695 if (s >= end) ###### goto phooey; 6695 if (prog->reganch & RE_USE_INTUIT) { 6695 s = re_intuit_start(prog, sv, s + 1, strend, flags, NULL); 6695 if (!s) 2031 goto phooey; } else ###### s++; } } else { 430902 if (s > startpos) ###### s--; 1567958 while (s < end) { 1514301 if (*s++ == '\n') { /* don't need PL_utf8skip here */ 381362 if (regtry(prog, s)) 377245 goto got_it; } } } } 7298321 goto phooey; 7298321 } else if (prog->reganch & ROPT_ANCH_GPOS) { 974521 if (regtry(prog, PL_reg_ganch)) 895459 goto got_it; 6323800 goto phooey; } /* Messy cases: unanchored match. */ 6323800 if ((prog->anchored_substr || prog->anchored_utf8) && prog->reganch & ROPT_SKIP) { /* we have /x+whatever/ */ /* it must be a one character string (XXXX Except UTF?) */ 57159 char ch; #ifdef DEBUGGING 57159 int did_match = 0; #endif 57159 if (!(do_utf8 ? prog->anchored_utf8 : prog->anchored_substr)) 2 do_utf8 ? to_utf8_substr(prog) : to_byte_substr(prog); 57159 ch = SvPVX_const(do_utf8 ? prog->anchored_utf8 : prog->anchored_substr)[0]; 57159 if (do_utf8) { 2 while (s < strend) { 2 if (*s == ch) { 2 DEBUG_EXECUTE_r( did_match = 1 ); 2 if (regtry(prog, s)) goto got_it; ###### s += UTF8SKIP(s); ###### while (s < strend && *s == ch) ###### s += UTF8SKIP(s); } ###### s += UTF8SKIP(s); } } else { 208614 while (s < strend) { 208059 if (*s == ch) { 60518 DEBUG_EXECUTE_r( did_match = 1 ); 60518 if (regtry(prog, s)) goto got_it; 3916 s++; 6128 while (s < strend && *s == ch) 2212 s++; } 151457 s++; } } DEBUG_EXECUTE_r(if (!did_match) PerlIO_printf(Perl_debug_log, "Did not find anchored character...\n") 555 ); } 6266641 else if (prog->anchored_substr != Nullsv || prog->anchored_utf8 != Nullsv || ((prog->float_substr != Nullsv || prog->float_utf8 != Nullsv) && prog->float_max_offset < strend - s)) { 765282 SV *must; 765282 I32 back_max; 765282 I32 back_min; 765282 char *last; 765282 char *last1; /* Last position checked before */ #ifdef DEBUGGING 765282 int did_match = 0; #endif 765282 if (prog->anchored_substr || prog->anchored_utf8) { 763003 if (!(do_utf8 ? prog->anchored_utf8 : prog->anchored_substr)) 63 do_utf8 ? to_utf8_substr(prog) : to_byte_substr(prog); 763003 must = do_utf8 ? prog->anchored_utf8 : prog->anchored_substr; 763003 back_max = back_min = prog->anchored_offset; } else { 2279 if (!(do_utf8 ? prog->float_utf8 : prog->float_substr)) ###### do_utf8 ? to_utf8_substr(prog) : to_byte_substr(prog); 2279 must = do_utf8 ? prog->float_utf8 : prog->float_substr; 2279 back_max = prog->float_max_offset; 2279 back_min = prog->float_min_offset; } 765282 if (must == &PL_sv_undef) /* could not downgrade utf8 check substring, so must fail */ 4 goto phooey; 765278 last = HOP3c(strend, /* Cannot start after this */ -(I32)(CHR_SVLEN(must) - (SvTAIL(must) != 0) + back_min), strbeg); 765278 if (s > PL_bostr) 681912 last1 = HOPc(s, -1); else 83366 last1 = s - 1; /* bogus */ /* XXXX check_substr already used to find "s", can optimize if check_substr==must. */ 765278 scream_pos = -1; 765278 dontbother = end_shift; 765278 strend = HOPc(strend, -dontbother); 1140301 while ( (s <= last) && ((flags & REXEC_SCREAM) ? (s = screaminstr(sv, must, HOP3c(s, back_min, strend) - strbeg, end_shift, &scream_pos, 0)) : (s = fbm_instr((unsigned char*)HOP3(s, back_min, strend), (unsigned char*)strend, must, multiline ? FBMrf_MULTILINE : 0))) ) { /* we may be pointing at the wrong string */ 1066268 if ((flags & REXEC_SCREAM) && RX_MATCH_COPIED(prog)) 3 s = strbeg + (s - SvPVX_const(sv)); 1066268 DEBUG_EXECUTE_r( did_match = 1 ); 1066268 if (HOPc(s, -back_max) > last1) { 1064471 last1 = HOPc(s, -back_min); 1064471 s = HOPc(s, -back_max); } else { 1797 char *t = (last1 >= PL_bostr) ? HOPc(last1, 1) : last1 + 1; 1797 last1 = HOPc(s, -back_min); 1797 s = t; } 1066268 if (do_utf8) { 1412 while (s <= last1) { 1391 if (regtry(prog, s)) 1370 goto got_it; 21 s += UTF8SKIP(s); } } else { 1440266 while (s <= last1) { 1065264 if (regtry(prog, s)) 689875 goto got_it; 375389 s++; } } } DEBUG_EXECUTE_r(if (!did_match) PerlIO_printf(Perl_debug_log, "Did not find %s substr \"%s%.*s%s\"%s...\n", ((must == prog->anchored_substr || must == prog->anchored_utf8) ? "anchored" : "floating"), PL_colors[0], (int)(SvCUR(must) - (SvTAIL(must)!=0)), SvPVX_const(must), PL_colors[1], (SvTAIL(must) ? "$" : "")) 74033 ); ###### goto phooey; } 5501359 else if ((c = prog->regstclass)) { 5002849 if (minlen) { 5002847 I32 op = (U8)OP(prog->regstclass); /* don't bother with what can't match */ 5002847 if (PL_regkind[op] != EXACT && op != CANY) 3286196 strend = HOPc(strend, -(minlen - 1)); } DEBUG_EXECUTE_r({ SV *prop = sv_newmortal(); const char *s0; const char *s1; int len0; int len1; regprop(prop, c); s0 = UTF ? pv_uni_display(dsv0, (U8*)SvPVX_const(prop), SvCUR(prop), 60, UNI_DISPLAY_REGEX) : SvPVX_const(prop); len0 = UTF ? SvCUR(dsv0) : SvCUR(prop); s1 = UTF ? sv_uni_display(dsv1, sv, 60, UNI_DISPLAY_REGEX) : s; len1 = UTF ? SvCUR(dsv1) : strend - s; PerlIO_printf(Perl_debug_log, "Matching stclass \"%*.*s\" against \"%*.*s\"\n", len0, len0, s0, len1, len1, s1); 5002849 }); 5002849 if (find_byclass(prog, c, s, strend, 0)) 2900706 goto got_it; 2102143 DEBUG_EXECUTE_r(PerlIO_printf(Perl_debug_log, "Contradicts stclass...\n")); } else { 498510 dontbother = 0; 498510 if (prog->float_substr != Nullsv || prog->float_utf8 != Nullsv) { /* Trim the end. */ 76982 char *last; 76982 SV* float_real; 76982 if (!(do_utf8 ? prog->float_utf8 : prog->float_substr)) ###### do_utf8 ? to_utf8_substr(prog) : to_byte_substr(prog); 76982 float_real = do_utf8 ? prog->float_utf8 : prog->float_substr; 76982 if (flags & REXEC_SCREAM) { ###### last = screaminstr(sv, float_real, s - strbeg, end_shift, &scream_pos, 1); /* last one */ ###### if (!last) ###### last = scream_olds; /* Only one occurrence. */ /* we may be pointing at the wrong string */ ###### else if (RX_MATCH_COPIED(prog)) ###### s = strbeg + (s - SvPVX_const(sv)); } else { 76982 STRLEN len; 76982 const char * const little = SvPV_const(float_real, len); 76982 if (SvTAIL(float_real)) { 12644 if (memEQ(strend - len + 1, little, len - 1)) 12644 last = strend - len + 1; ###### else if (!multiline) ###### last = memEQ(strend - len, little, len) ? strend - len : Nullch; else 64338 goto find_last; } else { find_last: 64338 if (len) 64338 last = rninstr(s, strend, little, little + len); else ###### last = strend; /* matching "$" */ } } 76982 if (last == NULL) { DEBUG_EXECUTE_r(PerlIO_printf(Perl_debug_log, "%sCan't trim the tail, match fails (should not happen)%s\n", ###### PL_colors[4], PL_colors[5])); ###### goto phooey; /* Should not happen! */ } 76982 dontbother = strend - last + prog->float_min_offset; } 498510 if (minlen && (dontbother < minlen)) 251067 dontbother = minlen - 1; 498510 strend -= dontbother; /* this one's always in bytes! */ /* We don't know much -- general case. */ 498510 if (do_utf8) { 34533 for (;;) { 32419 if (regtry(prog, s)) 30294 goto got_it; 2125 if (s >= strend) 11 break; 2114 s += UTF8SKIP(s); }; } else { 4180984 do { 4180984 if (regtry(prog, s)) 357468 goto got_it; 3823515 } while (s++ < strend); } } /* Failure. */ 7203471 goto phooey; got_it: 7203471 RX_MATCH_TAINTED_set(prog, PL_reg_flags & RF_tainted); 7203471 if (PL_reg_eval_set) { /* Preserve the current value of $^R */ 3206 if (oreplsv != GvSV(PL_replgv)) 306 sv_setsv(oreplsv, GvSV(PL_replgv));/* So that when GvSV(replgv) is restored, the value remains the same. */ 3206 restore_pos(aTHX_ 0); } /* make sure $`, $&, $', and $digit will work later */ 7203471 if ( !(flags & REXEC_NOT_FIRST) ) { 6811398 RX_MATCH_COPY_FREE(prog); 6811398 if (flags & REXEC_COPY_STR) { 3398779 I32 i = PL_regeol - startpos + (stringarg - strbeg); #ifdef PERL_OLD_COPY_ON_WRITE if ((SvIsCOW(sv) || (SvFLAGS(sv) & CAN_COW_MASK) == CAN_COW_FLAGS)) { if (DEBUG_C_TEST) { PerlIO_printf(Perl_debug_log, "Copy on write: regexp capture, type %d\n", (int) SvTYPE(sv)); } prog->saved_copy = sv_setsv_cow(prog->saved_copy, sv); prog->subbeg = (char *)SvPVX_const(prog->saved_copy); assert (SvPOKp(prog->saved_copy)); } else #endif { 3398779 RX_MATCH_COPIED_on(prog); 3398779 s = savepvn(strbeg, i); 3398779 prog->subbeg = s; } 3398779 prog->sublen = i; } else { 3412619 prog->subbeg = strbeg; 3412619 prog->sublen = PL_regeol - strbeg; /* strend may have been modified */ } } 7203471 return 1; phooey: DEBUG_EXECUTE_r(PerlIO_printf(Perl_debug_log, "%sMatch failed%s\n", 3891201 PL_colors[4], PL_colors[5])); 3891201 if (PL_reg_eval_set) 609 restore_pos(aTHX_ 0); 3891201 return 0; } /* - regtry - try match at specific point */ STATIC I32 /* 0 failure, 1 success */ S_regtry(pTHX_ regexp *prog, char *startpos) 13711296 { 13711296 register I32 i; 13711296 register I32 *sp; 13711296 register I32 *ep; 13711296 CHECKPOINT lastcp; 13711296 GET_RE_DEBUG_FLAGS_DECL; #ifdef DEBUGGING 13711296 PL_regindent = 0; /* XXXX Not good when matches are reenterable... */ #endif 13711296 if ((prog->reganch & ROPT_EVAL_SEEN) && !PL_reg_eval_set) { 3820 MAGIC *mg; 3820 PL_reg_eval_set = RS_init; DEBUG_EXECUTE_r(DEBUG_s( PerlIO_printf(Perl_debug_log, " setting stack tmpbase at %"IVdf"\n", (IV)(PL_stack_sp - PL_stack_base)); 3820 )); 3820 SAVEI32(cxstack[cxstack_ix].blk_oldsp); 3820 cxstack[cxstack_ix].blk_oldsp = PL_stack_sp - PL_stack_base; /* Otherwise OP_NEXTSTATE will free whatever on stack now. */ 3820 SAVETMPS; /* Apparently this is not needed, judging by wantarray. */ /* SAVEI8(cxstack[cxstack_ix].blk_gimme); cxstack[cxstack_ix].blk_gimme = G_SCALAR; */ 3820 if (PL_reg_sv) { /* Make $_ available to executed code. */ 3820 if (PL_reg_sv != DEFSV) { 3103 SAVE_DEFSV; 3103 DEFSV = PL_reg_sv; } 3820 if (!(SvTYPE(PL_reg_sv) >= SVt_PVMG && SvMAGIC(PL_reg_sv) && (mg = mg_find(PL_reg_sv, PERL_MAGIC_regex_global)))) { /* prepare for quick setting of pos */ 2403 sv_magic(PL_reg_sv, (SV*)0, PERL_MAGIC_regex_global, Nullch, 0); 2403 mg = mg_find(PL_reg_sv, PERL_MAGIC_regex_global); 2403 mg->mg_len = -1; } 3820 PL_reg_magic = mg; 3820 PL_reg_oldpos = mg->mg_len; 3820 SAVEDESTRUCTOR_X(restore_pos, 0); } 3820 if (!PL_reg_curpm) { 86 Newz(22, PL_reg_curpm, 1, PMOP); #ifdef USE_ITHREADS { SV* repointer = newSViv(0); /* so we know which PL_regex_padav element is PL_reg_curpm */ SvFLAGS(repointer) |= SVf_BREAK; av_push(PL_regex_padav,repointer); PL_reg_curpm->op_pmoffset = av_len(PL_regex_padav); PL_regex_pad = AvARRAY(PL_regex_padav); } #endif } 3820 PM_SETRE(PL_reg_curpm, prog); 3820 PL_reg_oldcurpm = PL_curpm; 3820 PL_curpm = PL_reg_curpm; 3820 if (RX_MATCH_COPIED(prog)) { /* Here is a serious problem: we cannot rewrite subbeg, since it may be needed if this match fails. Thus $` inside (?{}) could fail... */ 2460 PL_reg_oldsaved = prog->subbeg; 2460 PL_reg_oldsavedlen = prog->sublen; #ifdef PERL_OLD_COPY_ON_WRITE PL_nrs = prog->saved_copy; #endif 2460 RX_MATCH_COPIED_off(prog); } else 1360 PL_reg_oldsaved = Nullch; 3820 prog->subbeg = PL_bostr; 3820 prog->sublen = PL_regeol - PL_bostr; /* strend may have been modified */ } 13711296 prog->startp[0] = startpos - PL_bostr; 13711296 PL_reginput = startpos; 13711296 PL_regstartp = prog->startp; 13711296 PL_regendp = prog->endp; 13711296 PL_reglastparen = &prog->lastparen; 13711296 PL_reglastcloseparen = &prog->lastcloseparen; 13711296 prog->lastparen = 0; 13711296 prog->lastcloseparen = 0; 13711296 PL_regsize = 0; 13711296 DEBUG_EXECUTE_r(PL_reg_starttry = startpos); 13711296 if (PL_reg_start_tmpl <= prog->nparens) { 52353 PL_reg_start_tmpl = prog->nparens*3/2 + 3; 52353 if(PL_reg_start_tmp) 8495 Renew(PL_reg_start_tmp, PL_reg_start_tmpl, char*); else 43858 New(22, PL_reg_start_tmp, PL_reg_start_tmpl, char*); } /* XXXX What this code is doing here?!!! There should be no need to do this again and again, PL_reglastparen should take care of this! --ilya*/ /* Tests pat.t#187 and split.t#{13,14} seem to depend on this code. * Actually, the code in regcppop() (which Ilya may be meaning by * PL_reglastparen), is not needed at all by the test suite * (op/regexp, op/pat, op/split), but that code is needed, oddly * enough, for building DynaLoader, or otherwise this * "Error: '*' not in typemap in DynaLoader.xs, line 164" * will happen. Meanwhile, this code *is* needed for the * above-mentioned test suite tests to succeed. The common theme * on those tests seems to be returning null fields from matches. * --jhi */ #if 1 13711296 sp = prog->startp; 13711296 ep = prog->endp; 13711296 if (prog->nparens) { 13824751 for (i = prog->nparens; i > (I32)*PL_reglastparen; i--) { 8775065 *++sp = -1; 8775065 *++ep = -1; } } #endif 13711296 REGCP_SET(lastcp); 13711296 if (regmatch(prog->program + 1)) { 7203471 prog->endp[0] = PL_reginput - PL_bostr; 7203471 return 1; } 6507824 REGCP_UNWIND(lastcp); 6507824 return 0; } #define RE_UNWIND_BRANCH 1 #define RE_UNWIND_BRANCHJ 2 union re_unwind_t; typedef struct { /* XX: makes sense to enlarge it... */ I32 type; I32 prev; CHECKPOINT lastcp; } re_unwind_generic_t; typedef struct { I32 type; I32 prev; CHECKPOINT lastcp; I32 lastparen; regnode *next; char *locinput; I32 nextchr; #ifdef DEBUGGING int regindent; #endif } re_unwind_branch_t; typedef union re_unwind_t { I32 type; re_unwind_generic_t generic; re_unwind_branch_t branch; } re_unwind_t; #define sayYES goto yes #define sayNO goto no #define sayNO_ANYOF goto no_anyof #define sayYES_FINAL goto yes_final #define sayYES_LOUD goto yes_loud #define sayNO_FINAL goto no_final #define sayNO_SILENT goto do_no #define saySAME(x) if (x) goto yes; else goto no #define POSCACHE_SUCCESS 0 /* caching success rather than failure */ #define POSCACHE_SEEN 1 /* we know what we're caching */ #define POSCACHE_START 2 /* the real cache: this bit maps to pos 0 */ #define CACHEsayYES STMT_START { \ if (cache_offset | cache_bit) { \ if (!(PL_reg_poscache[0] & (1<states[ state ].wordnum ) { \ if ( !accepted ) { \ ENTER; \ SAVETMPS; \ bufflen = TRIE_INITAL_ACCEPT_BUFFLEN ; \ sv_accept_buff=NEWSV( 1234, \ bufflen * sizeof(reg_trie_accepted) - 1 ); \ SvCUR_set( sv_accept_buff, sizeof(reg_trie_accepted) ); \ SvPOK_on( sv_accept_buff ); \ sv_2mortal( sv_accept_buff ); \ accept_buff = (reg_trie_accepted*)SvPV_nolen( sv_accept_buff );\ } else { \ if ( accepted >= bufflen ) { \ bufflen *= 2; \ accept_buff =(reg_trie_accepted*)SvGROW( sv_accept_buff, \ bufflen * sizeof(reg_trie_accepted) ); \ } \ SvCUR_set( sv_accept_buff,SvCUR( sv_accept_buff ) \ + sizeof( reg_trie_accepted ) ); \ } \ accept_buff[ accepted ].wordnum = trie->states[ state ].wordnum; \ accept_buff[ accepted ].endpos = uc; \ ++accepted; \ } } STMT_END #define TRIE_HANDLE_CHAR STMT_START { \ if ( uvc < 256 ) { \ charid = trie->charmap[ uvc ]; \ } else { \ charid = 0; \ if( trie->widecharmap ) { \ SV** svpp = (SV**)NULL; \ svpp = hv_fetch( trie->widecharmap, (char*)&uvc, \ sizeof( UV ), 0 ); \ if ( svpp ) { \ charid = (U16)SvIV( *svpp ); \ } \ } \ } \ if ( charid && \ ( base + charid > trie->uniquecharcount ) && \ ( base + charid - 1 - trie->uniquecharcount < trie->lasttrans) && \ trie->trans[ base + charid - 1 - trie->uniquecharcount ].check == state ) \ { \ state = trie->trans[ base + charid - 1 - trie->uniquecharcount ].next; \ } else { \ state = 0; \ } \ uc += len; \ } STMT_END /* - regmatch - main matching routine * * Conceptually the strategy is simple: check to see whether the current * node matches, call self recursively to see whether the rest matches, * and then act accordingly. In practice we make some effort to avoid * recursion, in particular by going through "ordinary" nodes (that don't * need to know whether the rest of the match failed) by a loop instead of * by recursion. */ /* [lwall] I've hoisted the register declarations to the outer block in order to * maybe save a little bit of pushing and popping on the stack. It also takes * advantage of machines that use a register save mask on subroutine entry. */ STATIC I32 /* 0 failure, 1 success */ S_regmatch(pTHX_ regnode *prog) 28953002 { dVAR; 28953002 register regnode *scan; /* Current node. */ 28953002 regnode *next; /* Next node. */ 28953002 regnode *inner; /* Next node in internal branch. */ 28953002 register I32 nextchr; /* renamed nextchr - nextchar colides with function of same name */ 28953002 register I32 n; /* no or next */ 28953002 register I32 ln = 0; /* len or last */ 28953002 register char *s = Nullch; /* operand or save */ 28953002 register char *locinput = PL_reginput; 28953002 register I32 c1 = 0, c2 = 0, paren; /* case fold search, parenth */ 28953002 int minmod = 0, sw = 0, logical = 0; 28953002 I32 unwind = 0; /* used by the trie code */ 28953002 SV *sv_accept_buff = 0; /* accepting states we have traversed */ 28953002 reg_trie_accepted *accept_buff = 0; /* "" */ 28953002 reg_trie_data *trie; /* what trie are we using right now */ 28953002 U32 accepted = 0; /* how many accepting states we have seen*/ #if 0 I32 firstcp = PL_savestack_ix; #endif 28953002 const register bool do_utf8 = PL_reg_match_utf8; #ifdef DEBUGGING 28953002 SV *dsv0 = PERL_DEBUG_PAD_ZERO(0); 28953002 SV *dsv1 = PERL_DEBUG_PAD_ZERO(1); 28953002 SV *dsv2 = PERL_DEBUG_PAD_ZERO(2); 28953002 SV *re_debug_flags = NULL; #endif 28953002 GET_RE_DEBUG_FLAGS; #ifdef DEBUGGING 28953002 PL_regindent++; #endif /* Note that nextchr is a byte even in UTF */ 28953002 nextchr = UCHARAT(locinput); 28953002 scan = prog; 63107958 while (scan != NULL) { DEBUG_EXECUTE_r( { SV *prop = sv_newmortal(); const int docolor = *PL_colors[0]; const int taill = (docolor ? 10 : 7); /* 3 chars for "> <" */ int l = (PL_regeol - locinput) > taill ? taill : (PL_regeol - locinput); /* The part of the string before starttry has one color (pref0_len chars), between starttry and current position another one (pref_len - pref0_len chars), after the current position the third one. We assume that pref0_len <= pref_len, otherwise we decrease pref0_len. */ int pref_len = (locinput - PL_bostr) > (5 + taill) - l ? (5 + taill) - l : locinput - PL_bostr; int pref0_len; while (do_utf8 && UTF8_IS_CONTINUATION(*(U8*)(locinput - pref_len))) pref_len++; pref0_len = pref_len - (locinput - PL_reg_starttry); if (l + pref_len < (5 + taill) && l < PL_regeol - locinput) l = ( PL_regeol - locinput > (5 + taill) - pref_len ? (5 + taill) - pref_len : PL_regeol - locinput); while (do_utf8 && UTF8_IS_CONTINUATION(*(U8*)(locinput + l))) l--; if (pref0_len < 0) pref0_len = 0; if (pref0_len > pref_len) pref0_len = pref_len; regprop(prop, scan); { const char * const s0 = do_utf8 && OP(scan) != CANY ? pv_uni_display(dsv0, (U8*)(locinput - pref_len), pref0_len, 60, UNI_DISPLAY_REGEX) : locinput - pref_len; const int len0 = do_utf8 ? strlen(s0) : pref0_len; const char * const s1 = do_utf8 && OP(scan) != CANY ? pv_uni_display(dsv1, (U8*)(locinput - pref_len + pref0_len), pref_len - pref0_len, 60, UNI_DISPLAY_REGEX) : locinput - pref_len + pref0_len; const int len1 = do_utf8 ? strlen(s1) : pref_len - pref0_len; const char * const s2 = do_utf8 && OP(scan) != CANY ? pv_uni_display(dsv2, (U8*)locinput, PL_regeol - locinput, 60, UNI_DISPLAY_REGEX) : locinput; const int len2 = do_utf8 ? strlen(s2) : l; PerlIO_printf(Perl_debug_log, "%4"IVdf" <%s%.*s%s%s%.*s%s%s%s%.*s%s>%*s|%3"IVdf":%*s%s\n", (IV)(locinput - PL_bostr), PL_colors[4], len0, s0, PL_colors[5], PL_colors[2], len1, s1, PL_colors[3], (docolor ? "" : "> <"), PL_colors[0], len2, s2, PL_colors[1], 15 - l - pref_len + 1, "", (IV)(scan - PL_regprogram), PL_regindent*2, "", SvPVX_const(prop)); } 63107958 }); 63107958 next = scan + NEXT_OFF(scan); 63107958 if (next == scan) 10473303 next = NULL; 63107958 switch (OP(scan)) { case BOL: 4164077 if (locinput == PL_bostr) { /* regtill = regbol; */ 3026145 break; } 929551 sayNO; case MBOL: 929551 if (locinput == PL_bostr || ((nextchr || locinput < PL_regeol) && locinput[-1] == '\n')) { 43150 break; } 43150 sayNO; case SBOL: 43150 if (locinput == PL_bostr) 37509 break; 978427 sayNO; case GPOS: 978427 if (locinput == PL_reg_ganch) 974706 break; 911 sayNO; case EOL: 911 goto seol; case MEOL: 911 if ((nextchr || locinput < PL_regeol) && nextchr != '\n') 79 sayNO; 1345196 break; case SEOL: seol: 1345196 if ((nextchr || locinput < PL_regeol) && nextchr != '\n') 730887 sayNO; 614309 if (PL_regeol - locinput > 1) 610 sayNO; 129257 break; case EOS: 129257 if (PL_regeol != locinput) 89583 sayNO; 26157 break; case SANY: 26157 if (!nextchr && locinput >= PL_regeol) ###### sayNO; 26157 if (do_utf8) { 25 locinput += PL_utf8skip[nextchr]; 25 if (locinput > PL_regeol) ###### sayNO; 25 nextchr = UCHARAT(locinput); } else 26132 nextchr = UCHARAT(++locinput); 26132 break; case CANY: 67 if (!nextchr && locinput >= PL_regeol) ###### sayNO; 67 nextchr = UCHARAT(++locinput); 67 break; case REG_ANY: 92793 if ((!nextchr && locinput >= PL_regeol) || nextchr == '\n') 92640 sayNO; 92640 if (do_utf8) { 30015 locinput += PL_utf8skip[nextchr]; 30015 if (locinput > PL_regeol) ###### sayNO; 30015 nextchr = UCHARAT(locinput); } else 62625 nextchr = UCHARAT(++locinput); 62625 break; /* traverse the TRIE keeping track of all accepting states we transition through until we get to a failing node. we use two slightly different pieces of code to handle the traversal depending on whether its case sensitive or not. we reuse the accept code however. (this should probably be turned into a macro.) */ case TRIEF: case TRIEFL: { 11982 const U32 uniflags = ckWARN( WARN_UTF8 ) ? 0 : UTF8_ALLOW_ANY; 11982 U8 *uc = ( U8* )locinput; 11982 U32 state = 1; 11982 U16 charid = 0; 11982 U32 base = 0; 11982 UV uvc = 0; 11982 STRLEN len = 0; 11982 STRLEN foldlen = 0; 11982 U8 *uscan = (U8*)NULL; 11982 STRLEN bufflen=0; 11982 accepted = 0; 11982 trie = (reg_trie_data*)PL_regdata->data[ ARG( scan ) ]; 29707 while ( state && uc <= (U8*)PL_regeol ) { 17725 TRIE_CHECK_STATE_IS_ACCEPTING; 17725 base = trie->states[ state ].trans.base; DEBUG_TRIE_EXECUTE_r( PerlIO_printf( Perl_debug_log, "%*s %sState: %4"UVxf", Base: %4"UVxf", Accepted: %4"UVxf" ", REPORT_CODE_OFF + PL_regindent * 2, "", PL_colors[4], (UV)state, (UV)base, (UV)accepted ); 17725 ); 17725 if ( base ) { 16653 if ( do_utf8 || UTF ) { 128 if ( foldlen>0 ) { 5 uvc = utf8n_to_uvuni( uscan, UTF8_MAXLEN, &len, uniflags ); 5 foldlen -= len; 5 uscan += len; 5 len=0; } else { 123 U8 foldbuf[ UTF8_MAXBYTES_CASE + 1 ]; 123 uvc = utf8n_to_uvuni( (U8*)uc, UTF8_MAXLEN, &len, uniflags ); 123 uvc = to_uni_fold( uvc, foldbuf, &foldlen ); 123 foldlen -= UNISKIP( uvc ); 123 uscan = foldbuf + UNISKIP( uvc ); } } else { 16525 uvc = (UV)*uc; 16525 len = 1; } 16653 TRIE_HANDLE_CHAR; } else { 1072 state = 0; } DEBUG_TRIE_EXECUTE_r( PerlIO_printf( Perl_debug_log, "Charid:%3x CV:%4"UVxf" After State: %4"UVxf"%s\n", charid, uvc, (UV)state, PL_colors[5] ); 17725 ); } 11982 if ( !accepted ) { 10575 sayNO; } else { 315460 goto TrieAccept; } } /* unreached codepoint: we jump into the middle of the next case from previous if blocks */ case TRIE: { 315460 const U32 uniflags = ckWARN( WARN_UTF8 ) ? 0 : UTF8_ALLOW_ANY; 315460 U8 *uc = (U8*)locinput; 315460 U32 state = 1; 315460 U16 charid = 0; 315460 U32 base = 0; 315460 UV uvc = 0; 315460 STRLEN len = 0; 315460 STRLEN bufflen = 0; 315460 accepted = 0; 315460 trie = (reg_trie_data*)PL_regdata->data[ ARG( scan ) ]; 737722 while ( state && uc <= (U8*)PL_regeol ) { 422262 TRIE_CHECK_STATE_IS_ACCEPTING; 422262 base = trie->states[ state ].trans.base; DEBUG_TRIE_EXECUTE_r( PerlIO_printf( Perl_debug_log, "%*s %sState: %4"UVxf", Base: %4"UVxf", Accepted: %4"UVxf" ", REPORT_CODE_OFF + PL_regindent * 2, "", PL_colors[4], (UV)state, (UV)base, (UV)accepted ); 422262 ); 422262 if ( base ) { 401723 if ( do_utf8 || UTF ) { 1873 uvc = utf8n_to_uvuni( (U8*)uc, UTF8_MAXLEN, &len, uniflags ); } else { 399850 uvc = (U32)*uc; 399850 len = 1; } 401723 TRIE_HANDLE_CHAR; } else { 20539 state = 0; } DEBUG_TRIE_EXECUTE_r( PerlIO_printf( Perl_debug_log, "Charid:%3x CV:%4"UVxf" After State: %4"UVxf"%s\n", charid, uvc, (UV)state, PL_colors[5] ); 422262 ); } 315460 if ( !accepted ) { 290835 sayNO; } } /* There was at least one accepting state that we transitioned through. Presumably the number of accepting states is going to be low, typically one or two. So we simply scan through to find the one with lowest wordnum. Once we find it, we swap the last state into its place and decrement the size. We then try to match the rest of the pattern at the point where the word ends, if we succeed then we end the loop, otherwise the loop eventually terminates once all of the accepting states have been tried. */ TrieAccept: { 26032 int gotit = 0; 26032 if ( accepted == 1 ) { DEBUG_EXECUTE_r({ SV **tmp = av_fetch( trie->words, accept_buff[ 0 ].wordnum-1, 0 ); PerlIO_printf( Perl_debug_log, "%*s %sonly one match : #%d <%s>%s\n", REPORT_CODE_OFF+PL_regindent*2, "", PL_colors[4], accept_buff[ 0 ].wordnum, tmp ? SvPV_nolen_const( *tmp ) : "not compiled under -Dr", PL_colors[5] ); 24524 }); 24524 PL_reginput = (char *)accept_buff[ 0 ].endpos; /* in this case we free tmps/leave before we call regmatch as we wont be using accept_buff again. */ 24524 FREETMPS; 24524 LEAVE; 24524 gotit = regmatch( scan + NEXT_OFF( scan ) ); } else { DEBUG_EXECUTE_r( PerlIO_printf( Perl_debug_log,"%*s %sgot %"IVdf" possible matches%s\n", REPORT_CODE_OFF + PL_regindent * 2, "", PL_colors[4], (IV)accepted, PL_colors[5] ); 1508 ); 5128 while ( !gotit && accepted-- ) { 3620 U32 best = 0; 3620 U32 cur; 9506 for( cur = 1 ; cur <= accepted ; cur++ ) { DEBUG_TRIE_EXECUTE_r( PerlIO_printf( Perl_debug_log, "%*s %sgot %"IVdf" (%d) as best, looking at %"IVdf" (%d)%s\n", REPORT_CODE_OFF + PL_regindent * 2, "", PL_colors[4], (IV)best, accept_buff[ best ].wordnum, (IV)cur, accept_buff[ cur ].wordnum, PL_colors[5] ); 5886 ); 5886 if ( accept_buff[ cur ].wordnum < accept_buff[ best ].wordnum ) 2424 best = cur; } DEBUG_EXECUTE_r({ SV **tmp = av_fetch( trie->words, accept_buff[ best ].wordnum - 1, 0 ); PerlIO_printf( Perl_debug_log, "%*s %strying alternation #%d <%s> at 0x%p%s\n", REPORT_CODE_OFF+PL_regindent*2, "", PL_colors[4], accept_buff[best].wordnum, tmp ? SvPV_nolen_const( *tmp ) : "not compiled under -Dr",scan, PL_colors[5] ); 3620 }); 3620 if ( best= PL_regeol) 8 sayNO; 9270 if (NATIVE_TO_UNI(*(U8*)s) != utf8n_to_uvuni((U8*)l, UTF8_MAXBYTES, &ulen, ckWARN(WARN_UTF8) ? 0 : UTF8_ALLOW_ANY)) 2604 sayNO; 6666 l += ulen; 6666 s ++; } } else { /* The target is not utf8, the pattern is utf8. */ 292 while (s < e) { 243 STRLEN ulen; 243 if (l >= PL_regeol) 3 sayNO; 240 if (NATIVE_TO_UNI(*((U8*)l)) != utf8n_to_uvuni((U8*)s, UTF8_MAXBYTES, &ulen, ckWARN(WARN_UTF8) ? 0 : UTF8_ALLOW_ANY)) 31 sayNO; 209 s += ulen; 209 l ++; } } 3135 locinput = l; 3135 nextchr = UCHARAT(locinput); 3135 break; } /* The target and the pattern have the same utf8ness. */ /* Inline the first character, for speed. */ 6805445 if (UCHARAT(s) != nextchr) 2767374 sayNO; 4038071 if (PL_regeol - locinput < ln) 1566 sayNO; 4036505 if (ln > 1 && memNE(s, locinput, ln)) 403101 sayNO; 3633404 locinput += ln; 3633404 nextchr = UCHARAT(locinput); 3633404 break; case EXACTFL: 1696542 PL_reg_flags |= RF_tainted; /* FALL THROUGH */ case EXACTF: 1824337 s = STRING(scan); 1824337 ln = STR_LEN(scan); 1824337 if (do_utf8 || UTF) { /* Either target or the pattern are utf8. */ 8907 char *l = locinput; 8907 char *e = PL_regeol; 8907 if (ibcmp_utf8(s, 0, ln, (bool)UTF, l, &e, 0, do_utf8)) { /* One more case for the sharp s: * pack("U0U*", 0xDF) =~ /ss/i, * the 0xC3 0x9F are the UTF-8 * byte sequence for the U+00DF. */ 15 if (!(do_utf8 && toLOWER(s[0]) == 's' && ln >= 2 && toLOWER(s[1]) == 's' && (U8)l[0] == 0xC3 && e - l >= 2 && (U8)l[1] == 0x9F)) 8892 sayNO; } 8892 locinput = e; 8892 nextchr = UCHARAT(locinput); 8892 break; } /* Neither the target and the pattern are utf8. */ /* Inline the first character, for speed. */ 1815430 if (UCHARAT(s) != nextchr && UCHARAT(s) != ((OP(scan) == EXACTF) ? PL_fold : PL_fold_locale)[nextchr]) 107856 sayNO; 1707574 if (PL_regeol - locinput < ln) 355 sayNO; 1707219 if (ln > 1 && (OP(scan) == EXACTF ? ibcmp(s, locinput, ln) : ibcmp_locale(s, locinput, ln))) 1705025 sayNO; 1705025 locinput += ln; 1705025 nextchr = UCHARAT(locinput); 1705025 break; case ANYOF: 4437711 if (do_utf8) { 22504 STRLEN inclasslen = PL_regeol - locinput; 22504 if (!reginclass(scan, (U8*)locinput, &inclasslen, do_utf8)) 2370 sayNO_ANYOF; 20134 if (locinput >= PL_regeol) 3 sayNO; 20131 locinput += inclasslen ? inclasslen : UTF8SKIP(locinput); 20131 nextchr = UCHARAT(locinput); 20131 break; } else { 4415207 if (nextchr < 0) ###### nextchr = UCHARAT(locinput); 4415207 if (!REGINCLASS(scan, (U8*)locinput)) 1051159 sayNO_ANYOF; 2956545 if (!nextchr && locinput >= PL_regeol) 31328 sayNO; 2925217 nextchr = UCHARAT(++locinput); 2925217 break; } no_anyof: /* If we might have the case of the German sharp s * in a casefolding Unicode character class. */ 1461032 if (ANYOF_FOLD_SHARP_S(scan, locinput, PL_regeol)) { 2 locinput += SHARP_S_SKIP; 2 nextchr = UCHARAT(locinput); } else 77601 sayNO; 77601 break; case ALNUML: 77601 PL_reg_flags |= RF_tainted; /* FALL THROUGH */ case ALNUM: 84447 if (!nextchr) 2 sayNO; 84445 if (do_utf8) { 35 LOAD_UTF8_CHARCLASS_ALNUM(); 35 if (!(OP(scan) == ALNUM ? swash_fetch(PL_utf8_alnum, (U8*)locinput, do_utf8) : isALNUM_LC_utf8((U8*)locinput))) { ###### sayNO; } 35 locinput += PL_utf8skip[nextchr]; 35 nextchr = UCHARAT(locinput); 35 break; } 84410 if (!(OP(scan) == ALNUM ? isALNUM(nextchr) : isALNUM_LC(nextchr))) 84323 sayNO; 84323 nextchr = UCHARAT(++locinput); 84323 break; case NALNUML: 22964 PL_reg_flags |= RF_tainted; /* FALL THROUGH */ case NALNUM: 59276 if (!nextchr && locinput >= PL_regeol) 96 sayNO; 59180 if (do_utf8) { ###### LOAD_UTF8_CHARCLASS_ALNUM(); ###### if (OP(scan) == NALNUM ? swash_fetch(PL_utf8_alnum, (U8*)locinput, do_utf8) : isALNUM_LC_utf8((U8*)locinput)) { ###### sayNO; } ###### locinput += PL_utf8skip[nextchr]; ###### nextchr = UCHARAT(locinput); ###### break; } 59180 if (OP(scan) == NALNUM ? isALNUM(nextchr) : isALNUM_LC(nextchr)) 41624 sayNO; 41624 nextchr = UCHARAT(++locinput); 41624 break; case BOUNDL: case NBOUNDL: 128 PL_reg_flags |= RF_tainted; /* FALL THROUGH */ case BOUND: case NBOUND: /* was last char in word? */ 416361 if (do_utf8) { 14 if (locinput == PL_bostr) 10 ln = '\n'; else { 4 const U8 * const r = reghop3((U8*)locinput, -1, (U8*)PL_bostr); 4 ln = utf8n_to_uvchr(r, UTF8SKIP(r), 0, 0); } 14 if (OP(scan) == BOUND || OP(scan) == NBOUND) { 14 ln = isALNUM_uni(ln); 14 LOAD_UTF8_CHARCLASS_ALNUM(); 14 n = swash_fetch(PL_utf8_alnum, (U8*)locinput, do_utf8); } else { ###### ln = isALNUM_LC_uvchr(UNI_TO_NATIVE(ln)); ###### n = isALNUM_LC_utf8((U8*)locinput); } } else { 416347 ln = (locinput != PL_bostr) ? UCHARAT(locinput - 1) : '\n'; 416347 if (OP(scan) == BOUND || OP(scan) == NBOUND) { 416219 ln = isALNUM(ln); 416219 n = isALNUM(nextchr); } else { 128 ln = isALNUM_LC(ln); 128 n = isALNUM_LC(nextchr); } } 416361 if (((!ln) == (!n)) == (OP(scan) == BOUND || OP(scan) == BOUNDL)) 189276 sayNO; 7681 break; case SPACEL: 7681 PL_reg_flags |= RF_tainted; /* FALL THROUGH */ case SPACE: 224272 if (!nextchr) 350 sayNO; 223922 if (do_utf8) { 4 if (UTF8_IS_CONTINUED(nextchr)) { 4 LOAD_UTF8_CHARCLASS_SPACE(); 4 if (!(OP(scan) == SPACE ? swash_fetch(PL_utf8_space, (U8*)locinput, do_utf8) : isSPACE_LC_utf8((U8*)locinput))) { ###### sayNO; } 4 locinput += PL_utf8skip[nextchr]; 4 nextchr = UCHARAT(locinput); 4 break; } ###### if (!(OP(scan) == SPACE ? isSPACE(nextchr) : isSPACE_LC(nextchr))) ###### sayNO; ###### nextchr = UCHARAT(++locinput); } else { 223918 if (!(OP(scan) == SPACE ? isSPACE(nextchr) : isSPACE_LC(nextchr))) 5471 sayNO; 117258 nextchr = UCHARAT(++locinput); } 117258 break; case NSPACEL: 4947 PL_reg_flags |= RF_tainted; /* FALL THROUGH */ case NSPACE: 109241 if (!nextchr && locinput >= PL_regeol) 94 sayNO; 109147 if (do_utf8) { 3 LOAD_UTF8_CHARCLASS_SPACE(); 3 if (OP(scan) == NSPACE ? swash_fetch(PL_utf8_space, (U8*)locinput, do_utf8) : isSPACE_LC_utf8((U8*)locinput)) { 3 sayNO; } 3 locinput += PL_utf8skip[nextchr]; 3 nextchr = UCHARAT(locinput); 3 break; } 109144 if (OP(scan) == NSPACE ? isSPACE(nextchr) : isSPACE_LC(nextchr)) 95889 sayNO; 95889 nextchr = UCHARAT(++locinput); 95889 break; case DIGITL: ###### PL_reg_flags |= RF_tainted; /* FALL THROUGH */ case DIGIT: 63233 if (!nextchr) 6 sayNO; 63227 if (do_utf8) { 2 LOAD_UTF8_CHARCLASS_DIGIT(); 2 if (!(OP(scan) == DIGIT ? swash_fetch(PL_utf8_digit, (U8*)locinput, do_utf8) : isDIGIT_LC_utf8((U8*)locinput))) { ###### sayNO; } 1 locinput += PL_utf8skip[nextchr]; 1 nextchr = UCHARAT(locinput); 1 break; } 63225 if (!(OP(scan) == DIGIT ? isDIGIT(nextchr) : isDIGIT_LC(nextchr))) ###### sayNO; 3877 nextchr = UCHARAT(++locinput); 3877 break; case NDIGITL: ###### PL_reg_flags |= RF_tainted; /* FALL THROUGH */ case NDIGIT: 526 if (!nextchr && locinput >= PL_regeol) ###### sayNO; 526 if (do_utf8) { ###### LOAD_UTF8_CHARCLASS_DIGIT(); ###### if (OP(scan) == NDIGIT ? swash_fetch(PL_utf8_digit, (U8*)locinput, do_utf8) : isDIGIT_LC_utf8((U8*)locinput)) { ###### sayNO; } ###### locinput += PL_utf8skip[nextchr]; ###### nextchr = UCHARAT(locinput); ###### break; } 526 if (OP(scan) == NDIGIT ? isDIGIT(nextchr) : isDIGIT_LC(nextchr)) 526 sayNO; 526 nextchr = UCHARAT(++locinput); 526 break; case CLUMP: 11 if (locinput >= PL_regeol) ###### sayNO; 11 if (do_utf8) { 7 LOAD_UTF8_CHARCLASS_MARK(); 7 if (swash_fetch(PL_utf8_mark,(U8*)locinput, do_utf8)) ###### sayNO; 7 locinput += PL_utf8skip[nextchr]; 10 while (locinput < PL_regeol && swash_fetch(PL_utf8_mark,(U8*)locinput, do_utf8)) 3 locinput += UTF8SKIP(locinput); 7 if (locinput > PL_regeol) ###### sayNO; } else 4 locinput++; 11 nextchr = UCHARAT(locinput); 11 break; case REFFL: ###### PL_reg_flags |= RF_tainted; /* FALL THROUGH */ case REF: case REFF: 7136 n = ARG(scan); /* which paren pair */ 7136 ln = PL_regstartp[n]; 7136 PL_reg_leftiter = PL_reg_maxiter; /* Void cache */ 7136 if ((I32)*PL_reglastparen < n || ln == -1) 6686 sayNO; /* Do not match unless seen CLOSEn. */ 6686 if (ln == PL_regendp[n]) 10 break; 6676 s = PL_bostr + ln; 6676 if (do_utf8 && OP(scan) != REF) { /* REF can do byte comparison */ ###### char *l = locinput; ###### const char *e = PL_bostr + PL_regendp[n]; /* * Note that we can't do the "other character" lookup trick as * in the 8-bit case (no pun intended) because in Unicode we * have to map both upper and title case to lower case. */ ###### if (OP(scan) == REFF) { ###### while (s < e) { ###### STRLEN ulen1, ulen2; ###### U8 tmpbuf1[UTF8_MAXBYTES_CASE+1]; ###### U8 tmpbuf2[UTF8_MAXBYTES_CASE+1]; ###### if (l >= PL_regeol) ###### sayNO; ###### toLOWER_utf8((U8*)s, tmpbuf1, &ulen1); ###### toLOWER_utf8((U8*)l, tmpbuf2, &ulen2); ###### if (ulen1 != ulen2 || memNE((char *)tmpbuf1, (char *)tmpbuf2, ulen1)) ###### sayNO; ###### s += ulen1; ###### l += ulen2; } } ###### locinput = l; ###### nextchr = UCHARAT(locinput); ###### break; } /* Inline the first character, for speed. */ 6676 if (UCHARAT(s) != nextchr && (OP(scan) == REF || (UCHARAT(s) != ((OP(scan) == REFF ? PL_fold : PL_fold_locale)[nextchr])))) 2280 sayNO; 2280 ln = PL_regendp[n] - ln; 2280 if (locinput + ln > PL_regeol) 110 sayNO; 2170 if (ln > 1 && (OP(scan) == REF ? memNE(s, locinput, ln) : (OP(scan) == REFF ? ibcmp(s, locinput, ln) : ibcmp_locale(s, locinput, ln)))) 2170 sayNO; 2170 locinput += ln; 2170 nextchr = UCHARAT(locinput); 2170 break; case NOTHING: case TAIL: 10629 break; case BACK: 10629 break; case EVAL: { 10629 dSP; 10629 OP_4tree *oop = PL_op; 10629 COP *ocurcop = PL_curcop; 10629 PAD *old_comppad; 10629 SV *ret; 10629 struct regexp *oreg = PL_reg_re; 10629 n = ARG(scan); 10629 PL_op = (OP_4tree*)PL_regdata->data[n]; 10629 DEBUG_EXECUTE_r( PerlIO_printf(Perl_debug_log, " re_eval 0x%"UVxf"\n", PTR2UV(PL_op)) ); 10629 PAD_SAVE_LOCAL(old_comppad, (PAD*)PL_regdata->data[n + 2]); 10629 PL_regendp[0] = PL_reg_magic->mg_len = locinput - PL_bostr; { 10629 SV **before = SP; 10629 CALLRUNOPS(aTHX); /* Scalar context. */ 10628 SPAGAIN; 10628 if (SP == before) ###### ret = &PL_sv_undef; /* protect against empty (?{}) blocks. */ else { 10628 ret = POPs; 10628 PUTBACK; } } 10628 PL_op = oop; 10628 PAD_RESTORE_LOCAL(old_comppad); 10628 PL_curcop = ocurcop; 10628 if (logical) { 10099 if (logical == 2) { /* Postponed subexpression. */ 9905 regexp *re; 9905 MAGIC *mg = Null(MAGIC*); 9905 re_cc_state state; 9905 CHECKPOINT cp, lastcp; 9905 int toggleutf; 9905 register SV *sv; 9905 if(SvROK(ret) && SvSMAGICAL(sv = SvRV(ret))) 8492 mg = mg_find(sv, PERL_MAGIC_qr); 1413 else if (SvSMAGICAL(ret)) { 11 if (SvGMAGICAL(ret)) 4 sv_unmagic(ret, PERL_MAGIC_qr); else 7 mg = mg_find(ret, PERL_MAGIC_qr); } 9905 if (mg) { 8498 re = (regexp *)mg->mg_obj; 8498 (void)ReREFCNT_inc(re); } else { 1407 STRLEN len; 1407 const char *t = SvPV_const(ret, len); 1407 PMOP pm; 1407 char * const oprecomp = PL_regprecomp; 1407 const I32 osize = PL_regsize; 1407 const I32 onpar = PL_regnpar; 1407 Zero(&pm, 1, PMOP); 1407 if (DO_UTF8(ret)) pm.op_pmdynflags |= PMdf_DYN_UTF8; 1407 re = CALLREGCOMP(aTHX_ (char*)t, (char*)t + len, &pm); 1407 if (!(SvFLAGS(ret) & (SVs_TEMP | SVs_PADTMP | SVf_READONLY | SVs_GMG))) 6 sv_magic(ret,(SV*)ReREFCNT_inc(re), PERL_MAGIC_qr,0,0); 1407 PL_regprecomp = oprecomp; 1407 PL_regsize = osize; 1407 PL_regnpar = onpar; } DEBUG_EXECUTE_r( PerlIO_printf(Perl_debug_log, "Entering embedded \"%s%.60s%s%s\"\n", PL_colors[0], re->precomp, PL_colors[1], (strlen(re->precomp) > 60 ? "..." : "")) 9905 ); 9905 state.node = next; 9905 state.prev = PL_reg_call_cc; 9905 state.cc = PL_regcc; 9905 state.re = PL_reg_re; 9905 PL_regcc = 0; 9905 cp = regcppush(0); /* Save *all* the positions. */ 9905 REGCP_SET(lastcp); 9905 cache_re(re); 9905 state.ss = PL_savestack_ix; 9905 *PL_reglastparen = 0; 9905 *PL_reglastcloseparen = 0; 9905 PL_reg_call_cc = &state; 9905 PL_reginput = locinput; 9905 toggleutf = ((PL_reg_flags & RF_utf8) != 0) ^ ((re->reganch & ROPT_UTF8) != 0); 9905 if (toggleutf) PL_reg_flags ^= RF_utf8; /* XXXX This is too dramatic a measure... */ 9905 PL_reg_maxiter = 0; 9905 if (regmatch(re->program + 1)) { /* Even though we succeeded, we need to restore global variables, since we may be wrapped inside SUSPEND, thus the match may be not finished yet. */ /* XXXX Do this only if SUSPENDed? */ 4319 PL_reg_call_cc = state.prev; 4319 PL_regcc = state.cc; 4319 PL_reg_re = state.re; 4319 cache_re(PL_reg_re); 4319 if (toggleutf) PL_reg_flags ^= RF_utf8; /* XXXX This is too dramatic a measure... */ 4319 PL_reg_maxiter = 0; /* These are needed even if not SUSPEND. */ 4319 ReREFCNT_dec(re); 4319 regcpblow(cp); 4319 sayYES; } 5586 ReREFCNT_dec(re); 5586 REGCP_UNWIND(lastcp); 5586 regcppop(); 5586 PL_reg_call_cc = state.prev; 5586 PL_regcc = state.cc; 5586 PL_reg_re = state.re; 5586 cache_re(PL_reg_re); 5586 if (toggleutf) PL_reg_flags ^= RF_utf8; /* XXXX This is too dramatic a measure... */ 5586 PL_reg_maxiter = 0; 5586 logical = 0; 5586 sayNO; } 194 sw = SvTRUE(ret); 194 logical = 0; } else { 529 sv_setsv(save_scalar(PL_replgv), ret); 529 cache_re(oreg); } 529 break; } case OPEN: 7139107 n = ARG(scan); /* which paren pair */ 7139107 PL_reg_start_tmp[n] = locinput; 7139107 if (n > PL_regsize) 6305872 PL_regsize = n; 6305872 break; case CLOSE: 5823937 n = ARG(scan); /* which paren pair */ 5823937 PL_regstartp[n] = PL_reg_start_tmp[n] - PL_bostr; 5823937 PL_regendp[n] = locinput - PL_bostr; 5823937 if (n > (I32)*PL_reglastparen) 4773465 *PL_reglastparen = n; 5823937 *PL_reglastcloseparen = n; 5823937 break; case GROUPP: 380 n = ARG(scan); /* which paren pair */ 380 sw = ((I32)*PL_reglastparen >= n && PL_regendp[n] != -1); 380 break; case IFTHEN: 634 PL_reg_leftiter = PL_reg_maxiter; /* Void cache */ 634 if (sw) 306 next = NEXTOPER(NEXTOPER(scan)); else { 328 next = scan + ARG(scan); 328 if (OP(next) == IFTHEN) /* Fake one. */ 170 next = NEXTOPER(NEXTOPER(next)); } 170 break; case LOGICAL: 10159 logical = scan->flags; 10159 break; /******************************************************************* PL_regcc contains infoblock about the innermost (...)* loop, and a pointer to the next outer infoblock. Here is how Y(A)*Z is processed (if it is compiled into CURLYX/WHILEM): 1) After matching X, regnode for CURLYX is processed; 2) This regnode creates infoblock on the stack, and calls regmatch() recursively with the starting point at WHILEM node; 3) Each hit of WHILEM node tries to match A and Z (in the order depending on the current iteration, min/max of {min,max} and greediness). The information about where are nodes for "A" and "Z" is read from the infoblock, as is info on how many times "A" was already matched, and greediness. 4) After A matches, the same WHILEM node is hit again. 5) Each time WHILEM is hit, PL_regcc is the infoblock created by CURLYX of the same pair. Thus when WHILEM tries to match Z, it temporarily resets PL_regcc, since this Y(A)*Z can be a part of some other loop: as in (Y(A)*Z)*. If Z matches, the automaton will hit the WHILEM node of the external loop. Currently present infoblocks form a tree with a stem formed by PL_curcc and whatever it mentions via ->next, and additional attached trees corresponding to temporarily unset infoblocks as in "5" above. In the following picture infoblocks for outer loop of (Y(A)*?Z)*?T are denoted O, for inner I. NULL starting block is denoted by x. The matched string is YAAZYAZT. Temporarily postponed infoblocks are drawn below the "reset" infoblock. In fact in the picture below we do not show failed matches for Z and T by WHILEM blocks. [We illustrate minimal matches, since for them it is more obvious *why* one needs to *temporary* unset infoblocks.] Matched REx position InfoBlocks Comment (Y(A)*?Z)*?T x Y(A)*?Z)*?T x <- O Y (A)*?Z)*?T x <- O Y A)*?Z)*?T x <- O <- I YA )*?Z)*?T x <- O <- I YA A)*?Z)*?T x <- O <- I YAA )*?Z)*?T x <- O <- I YAA Z)*?T x <- O # Temporary unset I I YAAZ Y(A)*?Z)*?T x <- O I YAAZY (A)*?Z)*?T x <- O I YAAZY A)*?Z)*?T x <- O <- I I YAAZYA )*?Z)*?T x <- O <- I I YAAZYA Z)*?T x <- O # Temporary unset I I,I YAAZYAZ )*?T x <- O I,I YAAZYAZ T x # Temporary unset O O I,I YAAZYAZT x O I,I *******************************************************************/ case CURLYX: { 1170519 CURCUR cc; 1170519 CHECKPOINT cp = PL_savestack_ix; /* No need to save/restore up to this paren */ 1170519 I32 parenfloor = scan->flags; 1170519 if (OP(PREVOPER(next)) == NOTHING) /* LONGJMP */ 12 next += ARG(next); 1170519 cc.oldcc = PL_regcc; 1170519 PL_regcc = &cc; /* XXXX Probably it is better to teach regpush to support parenfloor > PL_regsize... */ 1170519 if (parenfloor > (I32)*PL_reglastparen) 236255 parenfloor = *PL_reglastparen; /* Pessimization... */ 1170519 cc.parenfloor = parenfloor; 1170519 cc.cur = -1; 1170519 cc.min = ARG1(scan); 1170519 cc.max = ARG2(scan); 1170519 cc.scan = NEXTOPER(scan) + EXTRA_STEP_2ARGS; 1170519 cc.next = next; 1170519 cc.minmod = minmod; 1170519 cc.lastloc = 0; 1170519 PL_reginput = locinput; 1170519 n = regmatch(PREVOPER(next)); /* start on the WHILEM */ 1170519 regcpblow(cp); 1170519 PL_regcc = cc.oldcc; 1170519 saySAME(n); } /* NOT REACHED */ case WHILEM: { /* * This is really hard to understand, because after we match * what we're trying to match, we must make sure th