X-Git-Url: http://git.shadowcat.co.uk/gitweb/gitweb.cgi?a=blobdiff_plain;f=regexec.c;h=d65d70c916d75068f2c775f15c75e00cc96ba824;hb=79af22e4562a90bdaa3aa9d22a11a3fa41f9c6bc;hp=c795a5b3d5bd4f9b7e3f6aec220542d6e11a116d;hpb=0064a8a9866779dceb087452b9bfaa733c51adce;p=p5sagit%2Fp5-mst-13.2.git diff --git a/regexec.c b/regexec.c index c795a5b..d65d70c 100644 --- a/regexec.c +++ b/regexec.c @@ -80,12 +80,6 @@ #define PERL_IN_REGEXEC_C #include "perl.h" -#ifdef PERL_IN_XSUB_RE -# if defined(PERL_CAPI) || defined(PERL_OBJECT) -# include "XSUB.h" -# endif -#endif - #include "regcomp.h" #define RF_tainted 1 /* tainted information used? */ @@ -107,17 +101,17 @@ */ #define CHR_SVLEN(sv) (UTF ? sv_len_utf8(sv) : SvCUR(sv)) -#define CHR_DIST(a,b) (DO_UTF8(PL_reg_sv) ? utf8_distance(a,b) : a - b) +#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) (DO_UTF8(PL_reg_sv) ? reghop((U8*)pos, off) : (U8*)(pos + off)) -#define HOPMAYBE(pos,off) (DO_UTF8(PL_reg_sv) ? reghopmaybe((U8*)pos, off) : (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) ( \ - (UTF && DO_UTF8(PL_reg_sv)) \ + (UTF && PL_reg_match_utf8) \ ? reghopmaybe((U8*)pos, -off) \ : (pos - off >= PL_bostr) \ ? (U8*)(pos - off) \ @@ -127,14 +121,29 @@ #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) (DO_UTF8(PL_reg_sv) ? reghop3((U8*)pos, off, (U8*)lim) : (U8*)(pos + off)) -#define HOPMAYBE3(pos,off,lim) (DO_UTF8(PL_reg_sv) ? reghopmaybe3((U8*)pos, off, (U8*)lim) : (U8*)(pos + off)) +#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(a,b) STMT_START { if (!CAT2(PL_utf8_,a)) (void)CAT2(is_utf8_, a)((U8*)b); } STMT_END -static void restore_pos(pTHXo_ void *arg); +/* 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 \ +) + +#define NEAR_EXACT(rn) (PL_regkind[(U8)OP(rn)] == EXACT || JUMPABLE(rn)) + +#define NEXT_IMPT(rn) STMT_START { \ + while (JUMPABLE(rn)) \ + if (OP(rn) == SUSPEND || OP(rn) == IFMATCH) \ + rn = NEXTOPER(NEXTOPER(rn)); \ + else rn += NEXT_OFF(rn); \ +} STMT_END + +static void restore_pos(pTHX_ void *arg); STATIC CHECKPOINT S_regcppush(pTHX_ I32 parenfloor) @@ -147,7 +156,7 @@ S_regcppush(pTHX_ I32 parenfloor) if (paren_elems_to_push < 0) Perl_croak(aTHX_ "panic: paren_elems_to_push < 0"); -#define REGCP_OTHER_ELEMS 5 +#define REGCP_OTHER_ELEMS 6 SSCHECK(paren_elems_to_push + REGCP_OTHER_ELEMS); for (p = PL_regsize; p > parenfloor; p--) { /* REGCP_PARENS_ELEMS are pushed per pairs of parentheses. */ @@ -159,6 +168,7 @@ S_regcppush(pTHX_ I32 parenfloor) /* REGCP_OTHER_ELEMS are pushed in any case, parentheses or no. */ SSPUSHINT(PL_regsize); SSPUSHINT(*PL_reglastparen); + SSPUSHINT(*PL_reglastcloseparen); SSPUSHPTR(PL_reginput); #define REGCP_FRAME_ELEMS 2 /* REGCP_FRAME_ELEMS are part of the REGCP_OTHER_ELEMS and @@ -192,6 +202,7 @@ S_regcppop(pTHX) assert(i == SAVEt_REGCONTEXT); /* Check that the magic cookie is there. */ i = SSPOPINT; /* Parentheses elements to pop. */ input = (char *) SSPOPPTR; + *PL_reglastcloseparen = SSPOPINT; *PL_reglastparen = SSPOPINT; PL_regsize = SSPOPINT; @@ -412,7 +423,7 @@ Perl_re_intuit_start(pTHX_ regexp *prog, SV *sv, char *strpos, goto fail; } if (prog->check_offset_min == prog->check_offset_max && - !(prog->reganch & ROPT_SANY_SEEN)) { + !(prog->reganch & ROPT_CANY_SEEN)) { /* Substring at constant offset from beg-of-str... */ I32 slen; @@ -488,7 +499,7 @@ Perl_re_intuit_start(pTHX_ regexp *prog, SV *sv, char *strpos, if (data) *data->scream_olds = s; } - else if (prog->reganch & ROPT_SANY_SEEN) + else if (prog->reganch & ROPT_CANY_SEEN) s = fbm_instr((U8*)(s + start_shift), (U8*)(strend - end_shift), check, PL_multiline ? FBMrf_MULTILINE : 0); @@ -774,7 +785,7 @@ Perl_re_intuit_start(pTHX_ regexp *prog, SV *sv, char *strpos, PL_regdata = prog->data; PL_bostr = startpos; } - s = find_byclass(prog, prog->regstclass, s, endpos, startpos, 1); + s = find_byclass(prog, prog->regstclass, s, endpos, startpos, 1); if (!s) { #ifdef DEBUGGING char *what = 0; @@ -876,7 +887,7 @@ S_find_byclass(pTHX_ regexp * prog, regnode *c, char *s, char *strend, char *sta unsigned int c2; char *e; register I32 tmp = 1; /* Scratch variable? */ - register bool do_utf8 = DO_UTF8(PL_reg_sv); + register bool do_utf8 = PL_reg_match_utf8; /* We know what class it must start with. */ switch (OP(c)) { @@ -893,12 +904,28 @@ S_find_byclass(pTHX_ regexp * prog, regnode *c, char *s, char *strend, char *sta s += do_utf8 ? UTF8SKIP(s) : 1; } break; + case CANY: + while (s < strend) { + if (tmp && (norun || regtry(prog, s))) + goto got_it; + else + tmp = doevery; + s++; + } + break; case EXACTF: m = STRING(c); ln = STR_LEN(c); if (UTF) { - c1 = to_utf8_lower((U8*)m); - c2 = to_utf8_upper((U8*)m); + STRLEN ulen1, ulen2; + U8 tmpbuf1[UTF8_MAXLEN*2+1]; + U8 tmpbuf2[UTF8_MAXLEN*2+1]; + + to_utf8_lower((U8*)m, tmpbuf1, &ulen1); + to_utf8_upper((U8*)m, tmpbuf2, &ulen2); + + c1 = utf8_to_uvuni(tmpbuf1, 0); + c2 = utf8_to_uvuni(tmpbuf2, 0); } else { c1 = *(U8*)m; @@ -1438,12 +1465,11 @@ Perl_regexec_flags(pTHX_ register regexp *prog, char *stringarg, register char * } minlen = prog->minlen; - if (do_utf8) { - if (!(prog->reganch & ROPT_SANY_SEEN)) + if (do_utf8 && !(prog->reganch & ROPT_CANY_SEEN)) { if (utf8_distance((U8*)strend, (U8*)startpos) < minlen) goto phooey; } else { - if (strend - startpos < minlen) goto phooey; + if (strend - startpos < minlen) goto phooey; } /* Check validity of program. */ @@ -1486,7 +1512,7 @@ Perl_regexec_flags(pTHX_ register regexp *prog, char *stringarg, register char * && mg->mg_len >= 0) { PL_reg_ganch = strbeg + mg->mg_len; /* Defined pos() */ if (prog->reganch & ROPT_ANCH_GPOS) { - if (s > PL_reg_ganch) + if (s > PL_reg_ganch) goto phooey; s = PL_reg_ganch; } @@ -1765,7 +1791,7 @@ got_it: sv_setsv(oreplsv, GvSV(PL_replgv));/* So that when GvSV(replgv) is restored, the value remains the same. */ - restore_pos(aTHXo_ 0); + restore_pos(aTHX_ 0); } /* make sure $`, $&, $', and $digit will work later */ @@ -1794,7 +1820,7 @@ phooey: DEBUG_r(PerlIO_printf(Perl_debug_log, "%sMatch failed%s\n", PL_colors[4],PL_colors[5])); if (PL_reg_eval_set) - restore_pos(aTHXo_ 0); + restore_pos(aTHX_ 0); return 0; } @@ -1831,7 +1857,7 @@ S_regtry(pTHX_ regexp *prog, char *startpos) if (PL_reg_sv) { /* Make $_ available to executed code. */ if (PL_reg_sv != DEFSV) { - /* SAVE_DEFSV does *not* suffice here for USE_THREADS */ + /* SAVE_DEFSV does *not* suffice here for USE_5005THREADS */ SAVESPTR(DEFSV); DEFSV = PL_reg_sv; } @@ -1848,9 +1874,20 @@ S_regtry(pTHX_ regexp *prog, char *startpos) PL_reg_oldpos = mg->mg_len; SAVEDESTRUCTOR_X(restore_pos, 0); } - if (!PL_reg_curpm) + if (!PL_reg_curpm) { Newz(22,PL_reg_curpm, 1, PMOP); - PL_reg_curpm->op_pmregexp = prog; +#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 + } + PM_SETRE(PL_reg_curpm, prog); PL_reg_oldcurpm = PL_curpm; PL_curpm = PL_reg_curpm; if (RX_MATCH_COPIED(prog)) { @@ -1871,6 +1908,7 @@ S_regtry(pTHX_ regexp *prog, char *startpos) PL_regstartp = prog->startp; PL_regendp = prog->endp; PL_reglastparen = &prog->lastparen; + PL_reglastcloseparen = &prog->lastcloseparen; prog->lastparen = 0; PL_regsize = 0; DEBUG_r(PL_reg_starttry = startpos); @@ -1945,6 +1983,16 @@ typedef union re_unwind_t { re_unwind_branch_t branch; } re_unwind_t; +#define sayYES goto yes +#define sayNO goto no +#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 REPORT_CODE_OFF 24 + /* - regmatch - main matching routine * @@ -1977,7 +2025,7 @@ S_regmatch(pTHX_ regnode *prog) #if 0 I32 firstcp = PL_savestack_ix; #endif - register bool do_utf8 = DO_UTF8(PL_reg_sv); + register bool do_utf8 = PL_reg_match_utf8; #ifdef DEBUGGING PL_regindent++; @@ -1987,25 +2035,7 @@ S_regmatch(pTHX_ regnode *prog) nextchr = UCHARAT(locinput); scan = prog; while (scan != NULL) { -#define sayNO_L (logical ? (logical = 0, sw = 0, goto cont) : sayNO) -#if 1 -# define sayYES goto yes -# define sayNO goto no -# 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 REPORT_CODE_OFF 24 -#else -# define sayYES return 1 -# define sayNO return 0 -# define sayYES_FINAL return 1 -# define sayYES_LOUD return 1 -# define sayNO_FINAL return 0 -# define sayNO_SILENT return 0 -# define saySAME(x) return x -#endif + DEBUG_r( { SV *prop = sv_newmortal(); int docolor = *PL_colors[0]; @@ -2101,6 +2131,18 @@ S_regmatch(pTHX_ regnode *prog) case SANY: if (!nextchr && locinput >= PL_regeol) sayNO; + if (do_utf8) { + locinput += PL_utf8skip[nextchr]; + if (locinput > PL_regeol) + sayNO; + nextchr = UCHARAT(locinput); + } + else + nextchr = UCHARAT(++locinput); + break; + case CANY: + if (!nextchr && locinput >= PL_regeol) + sayNO; nextchr = UCHARAT(++locinput); break; case REG_ANY: @@ -2164,17 +2206,17 @@ S_regmatch(pTHX_ regnode *prog) if (do_utf8) { char *l = locinput; char *e; + STRLEN ulen; + U8 tmpbuf[UTF8_MAXLEN*2+1]; e = s + ln; - c1 = OP(scan) == EXACTF; while (s < e) { - if (l >= PL_regeol) { + if (l >= PL_regeol) sayNO; - } - if ((UTF ? utf8n_to_uvchr((U8*)s, e - s, 0, 0) : *((U8*)s)) != - (c1 ? toLOWER_utf8((U8*)l) : toLOWER_LC_utf8((U8*)l))) - sayNO; - s += UTF ? UTF8SKIP(s) : 1; - l += UTF8SKIP(l); + toLOWER_utf8((U8*)l, tmpbuf, &ulen); + if (memNE(s, tmpbuf, ulen)) + sayNO; + s += UTF8SKIP(s); + l += ulen; } locinput = l; nextchr = UCHARAT(locinput); @@ -2437,23 +2479,18 @@ S_regmatch(pTHX_ regnode *prog) * have to map both upper and title case to lower case. */ if (OP(scan) == REFF) { + STRLEN ulen1, ulen2; + U8 tmpbuf1[UTF8_MAXLEN*2+1]; + U8 tmpbuf2[UTF8_MAXLEN*2+1]; while (s < e) { if (l >= PL_regeol) sayNO; - if (toLOWER_utf8((U8*)s) != toLOWER_utf8((U8*)l)) - sayNO; - s += UTF8SKIP(s); - l += UTF8SKIP(l); - } - } - else { - while (s < e) { - if (l >= PL_regeol) - sayNO; - if (toLOWER_LC_utf8((U8*)s) != toLOWER_LC_utf8((U8*)l)) + toLOWER_utf8((U8*)s, tmpbuf1, &ulen1); + toLOWER_utf8((U8*)l, tmpbuf2, &ulen2); + if (ulen1 != ulen2 || memNE(tmpbuf1, tmpbuf2, ulen1)) sayNO; - s += UTF8SKIP(s); - l += UTF8SKIP(l); + s += ulen1; + l += ulen2; } } locinput = l; @@ -2499,11 +2536,18 @@ S_regmatch(pTHX_ regnode *prog) PL_curpad = AvARRAY((AV*)PL_regdata->data[n + 2]); PL_regendp[0] = PL_reg_magic->mg_len = locinput - PL_bostr; - CALLRUNOPS(aTHX); /* Scalar context. */ - SPAGAIN; - ret = POPs; - PUTBACK; - + { + SV **before = SP; + CALLRUNOPS(aTHX); /* Scalar context. */ + SPAGAIN; + if (SP == before) + ret = Nullsv; /* protect against empty (?{}) blocks. */ + else { + ret = POPs; + PUTBACK; + } + } + PL_op = oop; PL_curpad = ocurpad; PL_curcop = ocurcop; @@ -2562,6 +2606,7 @@ S_regmatch(pTHX_ regnode *prog) cache_re(re); state.ss = PL_savestack_ix; *PL_reglastparen = 0; + *PL_reglastcloseparen = 0; PL_reg_call_cc = &state; PL_reginput = locinput; @@ -2598,6 +2643,7 @@ S_regmatch(pTHX_ regnode *prog) /* XXXX This is too dramatic a measure... */ PL_reg_maxiter = 0; + logical = 0; sayNO; } sw = SvTRUE(ret); @@ -2619,6 +2665,7 @@ S_regmatch(pTHX_ regnode *prog) PL_regendp[n] = locinput - PL_bostr; if (n > *PL_reglastparen) *PL_reglastparen = n; + *PL_reglastcloseparen = n; break; case GROUPP: n = ARG(scan); /* which paren pair */ @@ -3004,21 +3051,34 @@ S_regmatch(pTHX_ regnode *prog) minmod = 0; if (ln && regrepeat_hard(scan, ln, &l) < ln) sayNO; - if (ln && l == 0 && n >= ln - /* In fact, this is tricky. If paren, then the - fact that we did/didnot match may influence - future execution. */ - && !(paren && ln == 0)) - ln = n; + /* if we matched something zero-length we don't need to + backtrack - capturing parens are already defined, so + the caveat in the maximal case doesn't apply + + XXXX if ln == 0, we can redo this check first time + through the following loop + */ + if (ln && l == 0) + n = ln; /* don't backtrack */ locinput = PL_reginput; - if (PL_regkind[(U8)OP(next)] == EXACT) { - c1 = (U8)*STRING(next); - if (OP(next) == EXACTF) - c2 = PL_fold[c1]; - else if (OP(next) == EXACTFL) - c2 = PL_fold_locale[c1]; - else - c2 = c1; + if (NEAR_EXACT(next)) { + regnode *text_node = next; + + if (PL_regkind[(U8)OP(next)] != EXACT) + NEXT_IMPT(text_node); + + if (PL_regkind[(U8)OP(text_node)] != EXACT) { + c1 = c2 = -1000; + } + else { + c1 = (U8)*STRING(text_node); + if (OP(next) == EXACTF) + c2 = PL_fold[c1]; + else if (OP(text_node) == EXACTFL) + c2 = PL_fold_locale[c1]; + else + c2 = c1; + } } else c1 = c2 = -1000; @@ -3031,7 +3091,7 @@ S_regmatch(pTHX_ regnode *prog) UCHARAT(PL_reginput) == c2) { if (paren) { - if (n) { + if (ln) { PL_regstartp[paren] = HOPc(PL_reginput, -l) - PL_bostr; PL_regendp[paren] = PL_reginput - PL_bostr; @@ -3055,12 +3115,13 @@ S_regmatch(pTHX_ regnode *prog) } else { n = regrepeat_hard(scan, n, &l); - if (n != 0 && l == 0 - /* In fact, this is tricky. If paren, then the - fact that we did/didnot match may influence - future execution. */ - && !(paren && ln == 0)) - ln = n; + /* if we matched something zero-length we don't need to + backtrack, unless the minimum count is zero and we + are capturing the result - in that case the capture + being defined or not may affect later execution + */ + if (n != 0 && l == 0 && !(paren && ln == 0)) + ln = n; /* don't backtrack */ locinput = PL_reginput; DEBUG_r( PerlIO_printf(Perl_debug_log, @@ -3069,14 +3130,24 @@ S_regmatch(pTHX_ regnode *prog) (IV) n, (IV)l) ); if (n >= ln) { - if (PL_regkind[(U8)OP(next)] == EXACT) { - c1 = (U8)*STRING(next); - if (OP(next) == EXACTF) - c2 = PL_fold[c1]; - else if (OP(next) == EXACTFL) - c2 = PL_fold_locale[c1]; - else - c2 = c1; + if (NEAR_EXACT(next)) { + regnode *text_node = next; + + if (PL_regkind[(U8)OP(next)] != EXACT) + NEXT_IMPT(text_node); + + if (PL_regkind[(U8)OP(text_node)] != EXACT) { + c1 = c2 = -1000; + } + else { + c1 = (U8)*STRING(text_node); + if (OP(text_node) == EXACTF) + c2 = PL_fold[c1]; + else if (OP(text_node) == EXACTFL) + c2 = PL_fold_locale[c1]; + else + c2 = c1; + } } else c1 = c2 = -1000; @@ -3146,22 +3217,48 @@ S_regmatch(pTHX_ regnode *prog) * Lookahead to avoid useless match attempts * when we know what character comes next. */ - if (PL_regkind[(U8)OP(next)] == EXACT) { - U8 *s = (U8*)STRING(next); - if (!UTF) { - c2 = c1 = *s; - if (OP(next) == EXACTF) - c2 = PL_fold[c1]; - else if (OP(next) == EXACTFL) - c2 = PL_fold_locale[c1]; - } - else { /* UTF */ - if (OP(next) == EXACTF) { - c1 = to_utf8_lower(s); - c2 = to_utf8_upper(s); + + /* + * Used to only do .*x and .*?x, but now it allows + * for )'s, ('s and (?{ ... })'s to be in the way + * of the quantifier and the EXACT-like node. -- japhy + */ + + if (NEAR_EXACT(next)) { + U8 *s; + regnode *text_node = next; + + if (PL_regkind[(U8)OP(next)] != EXACT) + NEXT_IMPT(text_node); + + if (PL_regkind[(U8)OP(text_node)] != EXACT) { + c1 = c2 = -1000; + } + else { + s = (U8*)STRING(text_node); + + if (!UTF) { + c2 = c1 = *s; + if (OP(text_node) == EXACTF) + c2 = PL_fold[c1]; + else if (OP(text_node) == EXACTFL) + c2 = PL_fold_locale[c1]; } - else { - c2 = c1 = utf8_to_uvchr(s, NULL); + else { /* UTF */ + if (OP(text_node) == EXACTF) { + STRLEN ulen1, ulen2; + U8 tmpbuf1[UTF8_MAXLEN*2+1]; + U8 tmpbuf2[UTF8_MAXLEN*2+1]; + + to_utf8_lower((U8*)s, tmpbuf1, &ulen1); + to_utf8_upper((U8*)s, tmpbuf2, &ulen2); + + c1 = utf8_to_uvuni(tmpbuf1, 0); + c2 = utf8_to_uvuni(tmpbuf2, 0); + } + else { + c2 = c1 = utf8_to_uvchr(s, NULL); + } } } } @@ -3201,12 +3298,13 @@ S_regmatch(pTHX_ regnode *prog) /* Find place 'next' could work */ if (!do_utf8) { if (c1 == c2) { - while (locinput <= e && *locinput != c1) + while (locinput <= e && + UCHARAT(locinput) != c1) locinput++; } else { while (locinput <= e - && *locinput != c1 - && *locinput != c2) + && UCHARAT(locinput) != c1 + && UCHARAT(locinput) != c2) locinput++; } count = locinput - old; @@ -3295,7 +3393,7 @@ S_regmatch(pTHX_ regnode *prog) } REGCP_SET(lastcp); if (paren) { - UV c; + UV c = 0; while (n >= ln) { if (c1 != -1000) { if (do_utf8) @@ -3315,7 +3413,7 @@ S_regmatch(pTHX_ regnode *prog) } } else { - UV c; + UV c = 0; while (n >= ln) { if (c1 != -1000) { if (do_utf8) @@ -3557,7 +3655,7 @@ S_regrepeat(pTHX_ regnode *p, I32 max) register I32 c; register char *loceol = PL_regeol; register I32 hardcount = 0; - register bool do_utf8 = DO_UTF8(PL_reg_sv); + register bool do_utf8 = PL_reg_match_utf8; scan = PL_reginput; if (max != REG_INFTY && max < loceol - scan) @@ -3578,6 +3676,9 @@ S_regrepeat(pTHX_ regnode *p, I32 max) case SANY: scan = loceol; break; + case CANY: + scan = loceol; + break; case EXACT: /* length of string is 1 */ c = (U8)*STRING(p); while (scan < loceol && UCHARAT(scan) == c) @@ -3793,7 +3894,7 @@ S_regrepeat_hard(pTHX_ regnode *p, I32 max, I32 *lp) return 0; start = PL_reginput; - if (DO_UTF8(PL_reg_sv)) { + if (PL_reg_match_utf8) { while (PL_reginput < loceol && (scan = PL_reginput, res = regmatch(p))) { if (!count++) { l = 0; @@ -3890,14 +3991,10 @@ S_reginclass(pTHX_ register regnode *n, register U8* p, register bool do_utf8) if (swash_fetch(sw, p, do_utf8)) match = TRUE; else if (flags & ANYOF_FOLD) { - U8 tmpbuf[UTF8_MAXLEN+1]; - - if (flags & ANYOF_LOCALE) { - PL_reg_flags |= RF_tainted; - uvchr_to_utf8(tmpbuf, toLOWER_LC_utf8(p)); - } - else - uvchr_to_utf8(tmpbuf, toLOWER_utf8(p)); + STRLEN ulen; + U8 tmpbuf[UTF8_MAXLEN*2+1]; + + toLOWER_utf8(p, tmpbuf, &ulen); if (swash_fetch(sw, tmpbuf, do_utf8)) match = TRUE; } @@ -4029,12 +4126,8 @@ S_reghopmaybe3(pTHX_ U8* s, I32 off, U8* lim) return s; } -#ifdef PERL_OBJECT -#include "XSUB.h" -#endif - static void -restore_pos(pTHXo_ void *arg) +restore_pos(pTHX_ void *arg) { if (PL_reg_eval_set) { if (PL_reg_oldsaved) {