X-Git-Url: http://git.shadowcat.co.uk/gitweb/gitweb.cgi?a=blobdiff_plain;f=regexec.c;h=bdbdb5918c4fdb8310e1fe2b0a79f6d3ce568730;hb=7df053ec69e901392ae6352566832be0a6917cfe;hp=1c599dee633aaf233ca08230306ff454e547ebec;hpb=dafc8851ab778a44213cb74896e93c0f90bed47d;p=p5sagit%2Fp5-mst-13.2.git diff --git a/regexec.c b/regexec.c index 1c599de..bdbdb59 100644 --- a/regexec.c +++ b/regexec.c @@ -39,6 +39,7 @@ /* *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 @@ -105,9 +106,6 @@ * Forwards. */ -#define REGINCLASS(p,c) (ANYOF_FLAGS(p) ? reginclass(p,c) : ANYOF_BITMAP_TEST(p,c)) -#define REGINCLASSUTF8(f,p) (ARG1(f) ? reginclassutf8(f,p) : swash_fetch((SV*)PL_regdata->data[ARG2(f)],p)) - #define CHR_SVLEN(sv) (UTF ? sv_len_utf8(sv) : SvCUR(sv)) #define CHR_DIST(a,b) (UTF ? utf8_distance(a,b) : a - b) @@ -124,7 +122,6 @@ static void restore_pos(pTHXo_ void *arg); STATIC CHECKPOINT S_regcppush(pTHX_ I32 parenfloor) { - dTHR; int retval = PL_savestack_ix; int i = (PL_regsize - parenfloor) * 4; int p; @@ -157,7 +154,6 @@ S_regcppush(pTHX_ I32 parenfloor) STATIC char * S_regcppop(pTHX) { - dTHR; I32 i = SSPOPINT; U32 paren = 0; char *input; @@ -190,13 +186,17 @@ S_regcppop(pTHX) (IV)(*PL_reglastparen + 1), (IV)PL_regnpar); } ); -#if 0 +#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 7877) will fail. --jhi */ + * (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 */ for (paren = *PL_reglastparen + 1; paren <= PL_regnpar; paren++) { if (paren > PL_regsize) PL_regstartp[paren] = -1; @@ -209,7 +209,6 @@ S_regcppop(pTHX) STATIC char * S_regcp_set_to(pTHX_ I32 ss) { - dTHR; I32 tmp = PL_savestack_ix; PL_savestack_ix = ss; @@ -268,7 +267,6 @@ Perl_pregexec(pTHX_ register regexp *prog, char *stringarg, register char *stren STATIC void S_cache_re(pTHX_ regexp *prog) { - dTHR; PL_regprecomp = prog->precomp; /* Needed for FAIL. */ #ifdef DEBUGGING PL_regprogram = prog->program; @@ -734,7 +732,7 @@ Perl_re_intuit_start(pTHX_ regexp *prog, SV *sv, char *strpos, t = s; if (prog->reganch & ROPT_UTF8) { - PL_regdata = prog->data; /* Used by REGINCLASS UTF logic */ + PL_regdata = prog->data; PL_bostr = startpos; } s = find_byclass(prog, prog->regstclass, s, endpos, startpos, 1); @@ -836,25 +834,13 @@ 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); /* We know what class it must start with. */ switch (OP(c)) { - case ANYOFUTF8: - while (s < strend) { - if (REGINCLASSUTF8(c, (U8*)s)) { - if (tmp && (norun || regtry(prog, s))) - goto got_it; - else - tmp = doevery; - } - else - tmp = 1; - s += UTF8SKIP(s); - } - break; case ANYOF: while (s < strend) { - if (REGINCLASS(c, *(U8*)s)) { + if (reginclass(c, (U8*)s, do_utf8)) { if (tmp && (norun || regtry(prog, s))) goto got_it; else @@ -862,7 +848,7 @@ S_find_byclass(pTHX_ regexp * prog, regnode *c, char *s, char *strend, char *sta } else tmp = 1; - s++; + s += do_utf8 ? UTF8SKIP(s) : 1; } break; case EXACTF: @@ -908,42 +894,40 @@ S_find_byclass(pTHX_ regexp * prog, regnode *c, char *s, char *strend, char *sta PL_reg_flags |= RF_tainted; /* FALL THROUGH */ case BOUND: - tmp = (s != startpos) ? UCHARAT(s - 1) : '\n'; - tmp = ((OP(c) == BOUND ? isALNUM(tmp) : isALNUM_LC(tmp)) != 0); - while (s < strend) { - if (tmp == !(OP(c) == BOUND ? isALNUM(*s) : isALNUM_LC(*s))) { - tmp = !tmp; - if ((norun || regtry(prog, s))) - goto got_it; + if (do_utf8) { + if (s == startpos) + tmp = '\n'; + else { + U8 *r = reghop((U8*)s, -1); + + tmp = (I32)utf8_to_uv(r, s - (char*)r, 0, 0); + } + tmp = ((OP(c) == BOUND ? + isALNUM_uni(tmp) : isALNUM_LC_uni(tmp)) != 0); + while (s < strend) { + if (tmp == !(OP(c) == BOUND ? + swash_fetch(PL_utf8_alnum, (U8*)s) : + isALNUM_LC_utf8((U8*)s))) + { + tmp = !tmp; + if ((norun || regtry(prog, s))) + goto got_it; + } + s += UTF8SKIP(s); } - s++; } - if ((!prog->minlen && tmp) && (norun || regtry(prog, s))) - goto got_it; - break; - case BOUNDLUTF8: - PL_reg_flags |= RF_tainted; - /* FALL THROUGH */ - case BOUNDUTF8: - if (s == startpos) - tmp = '\n'; else { - U8 *r = reghop((U8*)s, -1); - - tmp = (I32)utf8_to_uv(r, s - (char*)r, 0, 0); - } - tmp = ((OP(c) == BOUNDUTF8 ? - isALNUM_uni(tmp) : isALNUM_LC_uni(tmp)) != 0); - while (s < strend) { - if (tmp == !(OP(c) == BOUNDUTF8 ? - swash_fetch(PL_utf8_alnum, (U8*)s) : - isALNUM_LC_utf8((U8*)s))) - { - tmp = !tmp; - if ((norun || regtry(prog, s))) - goto got_it; + tmp = (s != startpos) ? UCHARAT(s - 1) : '\n'; + tmp = ((OP(c) == BOUND ? isALNUM(tmp) : isALNUM_LC(tmp)) != 0); + while (s < strend) { + if (tmp == + !(OP(c) == BOUND ? isALNUM(*s) : isALNUM_LC(*s))) { + tmp = !tmp; + if ((norun || regtry(prog, s))) + goto got_it; + } + s++; } - s += UTF8SKIP(s); } if ((!prog->minlen && tmp) && (norun || regtry(prog, s))) goto got_it; @@ -952,365 +936,382 @@ S_find_byclass(pTHX_ regexp * prog, regnode *c, char *s, char *strend, char *sta PL_reg_flags |= RF_tainted; /* FALL THROUGH */ case NBOUND: - tmp = (s != startpos) ? UCHARAT(s - 1) : '\n'; - tmp = ((OP(c) == NBOUND ? isALNUM(tmp) : isALNUM_LC(tmp)) != 0); - while (s < strend) { - if (tmp == !(OP(c) == NBOUND ? isALNUM(*s) : isALNUM_LC(*s))) - tmp = !tmp; - else if ((norun || regtry(prog, s))) - goto got_it; - s++; + if (do_utf8) { + if (s == startpos) + tmp = '\n'; + else { + U8 *r = reghop((U8*)s, -1); + + tmp = (I32)utf8_to_uv(r, s - (char*)r, 0, 0); + } + tmp = ((OP(c) == NBOUND ? + isALNUM_uni(tmp) : isALNUM_LC_uni(tmp)) != 0); + while (s < strend) { + if (tmp == !(OP(c) == NBOUND ? + swash_fetch(PL_utf8_alnum, (U8*)s) : + isALNUM_LC_utf8((U8*)s))) + tmp = !tmp; + else if ((norun || regtry(prog, s))) + goto got_it; + s += UTF8SKIP(s); + } } - if ((!prog->minlen && !tmp) && (norun || regtry(prog, s))) - goto got_it; - break; - case NBOUNDLUTF8: - PL_reg_flags |= RF_tainted; - /* FALL THROUGH */ - case NBOUNDUTF8: - if (s == startpos) - tmp = '\n'; else { - U8 *r = reghop((U8*)s, -1); - - tmp = (I32)utf8_to_uv(r, s - (char*)r, 0, 0); - } - tmp = ((OP(c) == NBOUNDUTF8 ? - isALNUM_uni(tmp) : isALNUM_LC_uni(tmp)) != 0); - while (s < strend) { - if (tmp == !(OP(c) == NBOUNDUTF8 ? - swash_fetch(PL_utf8_alnum, (U8*)s) : - isALNUM_LC_utf8((U8*)s))) - tmp = !tmp; - else if ((norun || regtry(prog, s))) - goto got_it; - s += UTF8SKIP(s); + tmp = (s != startpos) ? UCHARAT(s - 1) : '\n'; + tmp = ((OP(c) == NBOUND ? + isALNUM(tmp) : isALNUM_LC(tmp)) != 0); + while (s < strend) { + if (tmp == + !(OP(c) == NBOUND ? isALNUM(*s) : isALNUM_LC(*s))) + tmp = !tmp; + else if ((norun || regtry(prog, s))) + goto got_it; + s++; + } } if ((!prog->minlen && !tmp) && (norun || regtry(prog, s))) goto got_it; break; case ALNUM: - while (s < strend) { - if (isALNUM(*s)) { - if (tmp && (norun || regtry(prog, s))) - goto got_it; + if (do_utf8) { + while (s < strend) { + if (swash_fetch(PL_utf8_alnum, (U8*)s)) { + if (tmp && (norun || regtry(prog, s))) + goto got_it; + else + tmp = doevery; + } else - tmp = doevery; + tmp = 1; + s += UTF8SKIP(s); } - else - tmp = 1; - s++; } - break; - case ALNUMUTF8: - while (s < strend) { - if (swash_fetch(PL_utf8_alnum, (U8*)s)) { - if (tmp && (norun || regtry(prog, s))) - goto got_it; + else { + while (s < strend) { + if (isALNUM(*s)) { + if (tmp && (norun || regtry(prog, s))) + goto got_it; + else + tmp = doevery; + } else - tmp = doevery; + tmp = 1; + s++; } - else - tmp = 1; - s += UTF8SKIP(s); } break; case ALNUML: PL_reg_flags |= RF_tainted; - while (s < strend) { - if (isALNUM_LC(*s)) { - if (tmp && (norun || regtry(prog, s))) - goto got_it; + if (do_utf8) { + while (s < strend) { + if (isALNUM_LC_utf8((U8*)s)) { + if (tmp && (norun || regtry(prog, s))) + goto got_it; + else + tmp = doevery; + } else - tmp = doevery; + tmp = 1; + s += UTF8SKIP(s); } - else - tmp = 1; - s++; } - break; - case ALNUMLUTF8: - PL_reg_flags |= RF_tainted; - while (s < strend) { - if (isALNUM_LC_utf8((U8*)s)) { - if (tmp && (norun || regtry(prog, s))) - goto got_it; + else { + while (s < strend) { + if (isALNUM_LC(*s)) { + if (tmp && (norun || regtry(prog, s))) + goto got_it; + else + tmp = doevery; + } else - tmp = doevery; + tmp = 1; + s++; } - else - tmp = 1; - s += UTF8SKIP(s); } break; case NALNUM: - while (s < strend) { - if (!isALNUM(*s)) { - if (tmp && (norun || regtry(prog, s))) - goto got_it; + if (do_utf8) { + while (s < strend) { + if (!swash_fetch(PL_utf8_alnum, (U8*)s)) { + if (tmp && (norun || regtry(prog, s))) + goto got_it; + else + tmp = doevery; + } else - tmp = doevery; + tmp = 1; + s += UTF8SKIP(s); } - else - tmp = 1; - s++; } - break; - case NALNUMUTF8: - while (s < strend) { - if (!swash_fetch(PL_utf8_alnum, (U8*)s)) { - if (tmp && (norun || regtry(prog, s))) - goto got_it; + else { + while (s < strend) { + if (!isALNUM(*s)) { + if (tmp && (norun || regtry(prog, s))) + goto got_it; + else + tmp = doevery; + } else - tmp = doevery; + tmp = 1; + s++; } - else - tmp = 1; - s += UTF8SKIP(s); } break; case NALNUML: PL_reg_flags |= RF_tainted; - while (s < strend) { - if (!isALNUM_LC(*s)) { - if (tmp && (norun || regtry(prog, s))) - goto got_it; + if (do_utf8) { + while (s < strend) { + if (!isALNUM_LC_utf8((U8*)s)) { + if (tmp && (norun || regtry(prog, s))) + goto got_it; + else + tmp = doevery; + } else - tmp = doevery; + tmp = 1; + s += UTF8SKIP(s); } - else - tmp = 1; - s++; } - break; - case NALNUMLUTF8: - PL_reg_flags |= RF_tainted; - while (s < strend) { - if (!isALNUM_LC_utf8((U8*)s)) { - if (tmp && (norun || regtry(prog, s))) - goto got_it; + else { + while (s < strend) { + if (!isALNUM_LC(*s)) { + if (tmp && (norun || regtry(prog, s))) + goto got_it; + else + tmp = doevery; + } else - tmp = doevery; + tmp = 1; + s++; } - else - tmp = 1; - s += UTF8SKIP(s); } break; case SPACE: - while (s < strend) { - if (isSPACE(*s)) { - if (tmp && (norun || regtry(prog, s))) - goto got_it; + if (do_utf8) { + while (s < strend) { + if (*s == ' ' || swash_fetch(PL_utf8_space,(U8*)s)) { + if (tmp && (norun || regtry(prog, s))) + goto got_it; + else + tmp = doevery; + } else - tmp = doevery; + tmp = 1; + s += UTF8SKIP(s); } - else - tmp = 1; - s++; } - break; - case SPACEUTF8: - while (s < strend) { - if (*s == ' ' || swash_fetch(PL_utf8_space,(U8*)s)) { - if (tmp && (norun || regtry(prog, s))) - goto got_it; + else { + while (s < strend) { + if (isSPACE(*s)) { + if (tmp && (norun || regtry(prog, s))) + goto got_it; + else + tmp = doevery; + } else - tmp = doevery; + tmp = 1; + s++; } - else - tmp = 1; - s += UTF8SKIP(s); } break; case SPACEL: PL_reg_flags |= RF_tainted; - while (s < strend) { - if (isSPACE_LC(*s)) { - if (tmp && (norun || regtry(prog, s))) - goto got_it; + if (do_utf8) { + while (s < strend) { + if (*s == ' ' || isSPACE_LC_utf8((U8*)s)) { + if (tmp && (norun || regtry(prog, s))) + goto got_it; + else + tmp = doevery; + } else - tmp = doevery; + tmp = 1; + s += UTF8SKIP(s); } - else - tmp = 1; - s++; } - break; - case SPACELUTF8: - PL_reg_flags |= RF_tainted; - while (s < strend) { - if (*s == ' ' || isSPACE_LC_utf8((U8*)s)) { - if (tmp && (norun || regtry(prog, s))) - goto got_it; + else { + while (s < strend) { + if (isSPACE_LC(*s)) { + if (tmp && (norun || regtry(prog, s))) + goto got_it; + else + tmp = doevery; + } else - tmp = doevery; + tmp = 1; + s++; } - else - tmp = 1; - s += UTF8SKIP(s); } break; case NSPACE: - while (s < strend) { - if (!isSPACE(*s)) { - if (tmp && (norun || regtry(prog, s))) - goto got_it; + if (do_utf8) { + while (s < strend) { + if (!(*s == ' ' || swash_fetch(PL_utf8_space,(U8*)s))) { + if (tmp && (norun || regtry(prog, s))) + goto got_it; + else + tmp = doevery; + } else - tmp = doevery; + tmp = 1; + s += UTF8SKIP(s); } - else - tmp = 1; - s++; } - break; - case NSPACEUTF8: - while (s < strend) { - if (!(*s == ' ' || swash_fetch(PL_utf8_space,(U8*)s))) { - if (tmp && (norun || regtry(prog, s))) - goto got_it; + else { + while (s < strend) { + if (!isSPACE(*s)) { + if (tmp && (norun || regtry(prog, s))) + goto got_it; + else + tmp = doevery; + } else - tmp = doevery; + tmp = 1; + s++; } - else - tmp = 1; - s += UTF8SKIP(s); } break; case NSPACEL: PL_reg_flags |= RF_tainted; - while (s < strend) { - if (!isSPACE_LC(*s)) { - if (tmp && (norun || regtry(prog, s))) - goto got_it; + if (do_utf8) { + while (s < strend) { + if (!(*s == ' ' || isSPACE_LC_utf8((U8*)s))) { + if (tmp && (norun || regtry(prog, s))) + goto got_it; + else + tmp = doevery; + } else - tmp = doevery; + tmp = 1; + s += UTF8SKIP(s); } - else - tmp = 1; - s++; } - break; - case NSPACELUTF8: - PL_reg_flags |= RF_tainted; - while (s < strend) { - if (!(*s == ' ' || isSPACE_LC_utf8((U8*)s))) { - if (tmp && (norun || regtry(prog, s))) - goto got_it; + else { + while (s < strend) { + if (!isSPACE_LC(*s)) { + if (tmp && (norun || regtry(prog, s))) + goto got_it; + else + tmp = doevery; + } else - tmp = doevery; + tmp = 1; + s++; } - else - tmp = 1; - s += UTF8SKIP(s); } break; case DIGIT: - while (s < strend) { - if (isDIGIT(*s)) { - if (tmp && (norun || regtry(prog, s))) - goto got_it; + if (do_utf8) { + while (s < strend) { + if (swash_fetch(PL_utf8_digit,(U8*)s)) { + if (tmp && (norun || regtry(prog, s))) + goto got_it; + else + tmp = doevery; + } else - tmp = doevery; + tmp = 1; + s += UTF8SKIP(s); } - else - tmp = 1; - s++; } - break; - case DIGITUTF8: - while (s < strend) { - if (swash_fetch(PL_utf8_digit,(U8*)s)) { - if (tmp && (norun || regtry(prog, s))) - goto got_it; + else { + while (s < strend) { + if (isDIGIT(*s)) { + if (tmp && (norun || regtry(prog, s))) + goto got_it; + else + tmp = doevery; + } else - tmp = doevery; + tmp = 1; + s++; } - else - tmp = 1; - s += UTF8SKIP(s); } break; case DIGITL: PL_reg_flags |= RF_tainted; - while (s < strend) { - if (isDIGIT_LC(*s)) { - if (tmp && (norun || regtry(prog, s))) - goto got_it; + if (do_utf8) { + while (s < strend) { + if (isDIGIT_LC_utf8((U8*)s)) { + if (tmp && (norun || regtry(prog, s))) + goto got_it; + else + tmp = doevery; + } else - tmp = doevery; + tmp = 1; + s += UTF8SKIP(s); } - else - tmp = 1; - s++; } - break; - case DIGITLUTF8: - PL_reg_flags |= RF_tainted; - while (s < strend) { - if (isDIGIT_LC_utf8((U8*)s)) { - if (tmp && (norun || regtry(prog, s))) - goto got_it; + else { + while (s < strend) { + if (isDIGIT_LC(*s)) { + if (tmp && (norun || regtry(prog, s))) + goto got_it; + else + tmp = doevery; + } else - tmp = doevery; + tmp = 1; + s++; } - else - tmp = 1; - s += UTF8SKIP(s); } break; case NDIGIT: - while (s < strend) { - if (!isDIGIT(*s)) { - if (tmp && (norun || regtry(prog, s))) - goto got_it; + if (do_utf8) { + while (s < strend) { + if (!swash_fetch(PL_utf8_digit,(U8*)s)) { + if (tmp && (norun || regtry(prog, s))) + goto got_it; + else + tmp = doevery; + } else - tmp = doevery; + tmp = 1; + s += UTF8SKIP(s); } - else - tmp = 1; - s++; } - break; - case NDIGITUTF8: - while (s < strend) { - if (!swash_fetch(PL_utf8_digit,(U8*)s)) { - if (tmp && (norun || regtry(prog, s))) - goto got_it; + else { + while (s < strend) { + if (!isDIGIT(*s)) { + if (tmp && (norun || regtry(prog, s))) + goto got_it; + else + tmp = doevery; + } else - tmp = doevery; + tmp = 1; + s++; } - else - tmp = 1; - s += UTF8SKIP(s); } break; case NDIGITL: PL_reg_flags |= RF_tainted; - while (s < strend) { - if (!isDIGIT_LC(*s)) { - if (tmp && (norun || regtry(prog, s))) - goto got_it; + if (do_utf8) { + while (s < strend) { + if (!isDIGIT_LC_utf8((U8*)s)) { + if (tmp && (norun || regtry(prog, s))) + goto got_it; + else + tmp = doevery; + } else - tmp = doevery; + tmp = 1; + s += UTF8SKIP(s); } - else - tmp = 1; - s++; } - break; - case NDIGITLUTF8: - PL_reg_flags |= RF_tainted; - while (s < strend) { - if (!isDIGIT_LC_utf8((U8*)s)) { - if (tmp && (norun || regtry(prog, s))) - goto got_it; + else { + while (s < strend) { + if (!isDIGIT_LC(*s)) { + if (tmp && (norun || regtry(prog, s))) + goto got_it; + else + tmp = doevery; + } else - tmp = doevery; + tmp = 1; + s++; } - else - tmp = 1; - s += UTF8SKIP(s); } break; default: @@ -1334,7 +1335,6 @@ Perl_regexec_flags(pTHX_ register regexp *prog, char *stringarg, register char * /* data: May be used for some additional optimizations. */ /* nosave: For optimizations. */ { - dTHR; register char *s; register regnode *c; register char *startpos = stringarg; @@ -1603,6 +1603,11 @@ Perl_regexec_flags(pTHX_ register regexp *prog, char *stringarg, register char * if (minlen && PL_regkind[(U8)OP(prog->regstclass)] != EXACT) /* don't bother with what can't match */ strend = HOPc(strend, -(minlen - 1)); + DEBUG_r({ + SV *prop = sv_newmortal(); + regprop(prop, c); + PerlIO_printf(Perl_debug_log, "Matching stclass `%s' against `%s'\n", SvPVX(prop), s); + }); if (find_byclass(prog, c, s, strend, startpos, 0)) goto got_it; DEBUG_r(PerlIO_printf(Perl_debug_log, "Contradicts stclass...\n")); @@ -1616,7 +1621,7 @@ Perl_regexec_flags(pTHX_ register regexp *prog, char *stringarg, register char * last = screaminstr(sv, prog->float_substr, s - strbeg, end_shift, &scream_pos, 1); /* last one */ if (!last) - last = scream_olds; /* Only one occurence. */ + last = scream_olds; /* Only one occurrence. */ } else { STRLEN len; @@ -1718,7 +1723,6 @@ phooey: STATIC I32 /* 0 failure, 1 success */ S_regtry(pTHX_ regexp *prog, char *startpos) { - dTHR; register I32 i; register I32 *sp; register I32 *ep; @@ -1802,15 +1806,19 @@ S_regtry(pTHX_ regexp *prog, char *startpos) /* 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), does not seem to be needed at all (?!), whereas - * this code *is* needed for the above-mentioned tests to succeed. - * The common theme on those tests seems to be returning null fields - * from matches. --jhi */ + * 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 sp = prog->startp; ep = prog->endp; if (prog->nparens) { - for (i = prog->nparens; i >= 1; i--) { + for (i = prog->nparens; i > *PL_reglastparen; i--) { *++sp = -1; *++ep = -1; } @@ -1872,7 +1880,6 @@ typedef union re_unwind_t { STATIC I32 /* 0 failure, 1 success */ S_regmatch(pTHX_ regnode *prog) { - dTHR; register regnode *scan; /* Current node. */ regnode *next; /* Next node. */ regnode *inner; /* Next node in internal branch. */ @@ -1886,6 +1893,7 @@ S_regmatch(pTHX_ regnode *prog) int minmod = 0, sw = 0, logical = 0; I32 unwind = 0; I32 firstcp = PL_savestack_ix; + register bool do_utf8 = DO_UTF8(PL_reg_sv); #ifdef DEBUGGING PL_regindent++; @@ -2004,8 +2012,8 @@ S_regmatch(pTHX_ regnode *prog) if (PL_regeol != locinput) sayNO; break; - case SANYUTF8: - if (nextchr & 0x80) { + case SANY: + if (DO_UTF8(PL_reg_sv)) { locinput += PL_utf8skip[nextchr]; if (locinput > PL_regeol) sayNO; @@ -2016,13 +2024,8 @@ S_regmatch(pTHX_ regnode *prog) sayNO; nextchr = UCHARAT(++locinput); break; - case SANY: - if (!nextchr && locinput >= PL_regeol) - sayNO; - nextchr = UCHARAT(++locinput); - break; - case ANYUTF8: - if (nextchr & 0x80) { + case REG_ANY: + if (DO_UTF8(PL_reg_sv)) { locinput += PL_utf8skip[nextchr]; if (locinput > PL_regeol) sayNO; @@ -2033,11 +2036,6 @@ S_regmatch(pTHX_ regnode *prog) sayNO; nextchr = UCHARAT(++locinput); break; - case REG_ANY: - if ((!nextchr && locinput >= PL_regeol) || nextchr == '\n') - sayNO; - nextchr = UCHARAT(++locinput); - break; case EXACT: s = STRING(scan); ln = STR_LEN(scan); @@ -2094,22 +2092,24 @@ S_regmatch(pTHX_ regnode *prog) locinput += ln; nextchr = UCHARAT(locinput); break; - case ANYOFUTF8: - if (!REGINCLASSUTF8(scan, (U8*)locinput)) - sayNO; - if (locinput >= PL_regeol) - sayNO; - locinput += PL_utf8skip[nextchr]; - nextchr = UCHARAT(locinput); - break; case ANYOF: - if (nextchr < 0) + if (do_utf8) { + if (!reginclass(scan, (U8*)locinput, do_utf8)) + sayNO; + if (locinput >= PL_regeol) + sayNO; + locinput += PL_utf8skip[nextchr]; nextchr = UCHARAT(locinput); - if (!REGINCLASS(scan, nextchr)) - sayNO; - if (!nextchr && locinput >= PL_regeol) - sayNO; - nextchr = UCHARAT(++locinput); + } + else { + if (nextchr < 0) + nextchr = UCHARAT(locinput); + if (!reginclass(scan, (U8*)locinput, do_utf8)) + sayNO; + if (!nextchr && locinput >= PL_regeol) + sayNO; + nextchr = UCHARAT(++locinput); + } break; case ALNUML: PL_reg_flags |= RF_tainted; @@ -2117,19 +2117,8 @@ S_regmatch(pTHX_ regnode *prog) case ALNUM: if (!nextchr) sayNO; - if (!(OP(scan) == ALNUM - ? isALNUM(nextchr) : isALNUM_LC(nextchr))) - sayNO; - nextchr = UCHARAT(++locinput); - break; - case ALNUMLUTF8: - PL_reg_flags |= RF_tainted; - /* FALL THROUGH */ - case ALNUMUTF8: - if (!nextchr) - sayNO; - if (nextchr & 0x80) { - if (!(OP(scan) == ALNUMUTF8 + if (do_utf8) { + if (!(OP(scan) == ALNUM ? swash_fetch(PL_utf8_alnum, (U8*)locinput) : isALNUM_LC_utf8((U8*)locinput))) { @@ -2139,7 +2128,7 @@ S_regmatch(pTHX_ regnode *prog) nextchr = UCHARAT(locinput); break; } - if (!(OP(scan) == ALNUMUTF8 + if (!(OP(scan) == ALNUM ? isALNUM(nextchr) : isALNUM_LC(nextchr))) sayNO; nextchr = UCHARAT(++locinput); @@ -2150,19 +2139,8 @@ S_regmatch(pTHX_ regnode *prog) case NALNUM: if (!nextchr && locinput >= PL_regeol) sayNO; - if (OP(scan) == NALNUM - ? isALNUM(nextchr) : isALNUM_LC(nextchr)) - sayNO; - nextchr = UCHARAT(++locinput); - break; - case NALNUMLUTF8: - PL_reg_flags |= RF_tainted; - /* FALL THROUGH */ - case NALNUMUTF8: - if (!nextchr && locinput >= PL_regeol) - sayNO; - if (nextchr & 0x80) { - if (OP(scan) == NALNUMUTF8 + if (do_utf8) { + if (OP(scan) == NALNUM ? swash_fetch(PL_utf8_alnum, (U8*)locinput) : isALNUM_LC_utf8((U8*)locinput)) { @@ -2172,7 +2150,7 @@ S_regmatch(pTHX_ regnode *prog) nextchr = UCHARAT(locinput); break; } - if (OP(scan) == NALNUMUTF8 + if (OP(scan) == NALNUM ? isALNUM(nextchr) : isALNUM_LC(nextchr)) sayNO; nextchr = UCHARAT(++locinput); @@ -2184,42 +2162,38 @@ S_regmatch(pTHX_ regnode *prog) case BOUND: case NBOUND: /* was last char in word? */ - ln = (locinput != PL_regbol) ? UCHARAT(locinput - 1) : PL_regprev; - if (OP(scan) == BOUND || OP(scan) == NBOUND) { - ln = isALNUM(ln); - n = isALNUM(nextchr); - } - else { - ln = isALNUM_LC(ln); - n = isALNUM_LC(nextchr); - } - if (((!ln) == (!n)) == (OP(scan) == BOUND || OP(scan) == BOUNDL)) - sayNO; - break; - case BOUNDLUTF8: - case NBOUNDLUTF8: - PL_reg_flags |= RF_tainted; - /* FALL THROUGH */ - case BOUNDUTF8: - case NBOUNDUTF8: - /* was last char in word? */ - if (locinput == PL_regbol) - ln = PL_regprev; - else { - U8 *r = reghop((U8*)locinput, -1); - - ln = utf8_to_uv(r, s - (char*)r, 0, 0); - } - if (OP(scan) == BOUNDUTF8 || OP(scan) == NBOUNDUTF8) { - ln = isALNUM_uni(ln); - n = swash_fetch(PL_utf8_alnum, (U8*)locinput); + if (do_utf8) { + if (locinput == PL_regbol) + ln = PL_regprev; + else { + U8 *r = reghop((U8*)locinput, -1); + + ln = utf8_to_uv(r, s - (char*)r, 0, 0); + } + if (OP(scan) == BOUND || OP(scan) == NBOUND) { + ln = isALNUM_uni(ln); + n = swash_fetch(PL_utf8_alnum, (U8*)locinput); + } + else { + ln = isALNUM_LC_uni(ln); + n = isALNUM_LC_utf8((U8*)locinput); + } } else { - ln = isALNUM_LC_uni(ln); - n = isALNUM_LC_utf8((U8*)locinput); + ln = (locinput != PL_regbol) ? + UCHARAT(locinput - 1) : PL_regprev; + if (OP(scan) == BOUND || OP(scan) == NBOUND) { + ln = isALNUM(ln); + n = isALNUM(nextchr); + } + else { + ln = isALNUM_LC(ln); + n = isALNUM_LC(nextchr); + } } - if (((!ln) == (!n)) == (OP(scan) == BOUNDUTF8 || OP(scan) == BOUNDLUTF8)) - sayNO; + if (((!ln) == (!n)) == (OP(scan) == BOUND || + OP(scan) == BOUNDL)) + sayNO; break; case SPACEL: PL_reg_flags |= RF_tainted; @@ -2227,32 +2201,29 @@ S_regmatch(pTHX_ regnode *prog) case SPACE: if (!nextchr) sayNO; - if (!(OP(scan) == SPACE - ? isSPACE(nextchr) : isSPACE_LC(nextchr))) - sayNO; - nextchr = UCHARAT(++locinput); - break; - case SPACELUTF8: - PL_reg_flags |= RF_tainted; - /* FALL THROUGH */ - case SPACEUTF8: - if (!nextchr) - sayNO; - if (nextchr & 0x80) { - if (!(OP(scan) == SPACEUTF8 - ? swash_fetch(PL_utf8_space, (U8*)locinput) - : isSPACE_LC_utf8((U8*)locinput))) - { - sayNO; + if (DO_UTF8(PL_reg_sv)) { + if (nextchr & 0x80) { + if (!(OP(scan) == SPACE + ? swash_fetch(PL_utf8_space, (U8*)locinput) + : isSPACE_LC_utf8((U8*)locinput))) + { + sayNO; + } + locinput += PL_utf8skip[nextchr]; + nextchr = UCHARAT(locinput); + break; } - locinput += PL_utf8skip[nextchr]; - nextchr = UCHARAT(locinput); - break; + if (!(OP(scan) == SPACE + ? isSPACE(nextchr) : isSPACE_LC(nextchr))) + sayNO; + nextchr = UCHARAT(++locinput); + } + else { + if (!(OP(scan) == SPACE + ? isSPACE(nextchr) : isSPACE_LC(nextchr))) + sayNO; + nextchr = UCHARAT(++locinput); } - if (!(OP(scan) == SPACEUTF8 - ? isSPACE(nextchr) : isSPACE_LC(nextchr))) - sayNO; - nextchr = UCHARAT(++locinput); break; case NSPACEL: PL_reg_flags |= RF_tainted; @@ -2260,19 +2231,8 @@ S_regmatch(pTHX_ regnode *prog) case NSPACE: if (!nextchr && locinput >= PL_regeol) sayNO; - if (OP(scan) == NSPACE - ? isSPACE(nextchr) : isSPACE_LC(nextchr)) - sayNO; - nextchr = UCHARAT(++locinput); - break; - case NSPACELUTF8: - PL_reg_flags |= RF_tainted; - /* FALL THROUGH */ - case NSPACEUTF8: - if (!nextchr && locinput >= PL_regeol) - sayNO; - if (nextchr & 0x80) { - if (OP(scan) == NSPACEUTF8 + if (DO_UTF8(PL_reg_sv)) { + if (OP(scan) == NSPACE ? swash_fetch(PL_utf8_space, (U8*)locinput) : isSPACE_LC_utf8((U8*)locinput)) { @@ -2282,7 +2242,7 @@ S_regmatch(pTHX_ regnode *prog) nextchr = UCHARAT(locinput); break; } - if (OP(scan) == NSPACEUTF8 + if (OP(scan) == NSPACE ? isSPACE(nextchr) : isSPACE_LC(nextchr)) sayNO; nextchr = UCHARAT(++locinput); @@ -2293,19 +2253,8 @@ S_regmatch(pTHX_ regnode *prog) case DIGIT: if (!nextchr) sayNO; - if (!(OP(scan) == DIGIT - ? isDIGIT(nextchr) : isDIGIT_LC(nextchr))) - sayNO; - nextchr = UCHARAT(++locinput); - break; - case DIGITLUTF8: - PL_reg_flags |= RF_tainted; - /* FALL THROUGH */ - case DIGITUTF8: - if (!nextchr) - sayNO; - if (nextchr & 0x80) { - if (!(OP(scan) == DIGITUTF8 + if (DO_UTF8(PL_reg_sv)) { + if (!(OP(scan) == DIGIT ? swash_fetch(PL_utf8_digit, (U8*)locinput) : isDIGIT_LC_utf8((U8*)locinput))) { @@ -2315,7 +2264,7 @@ S_regmatch(pTHX_ regnode *prog) nextchr = UCHARAT(locinput); break; } - if (!(OP(scan) == DIGITUTF8 + if (!(OP(scan) == DIGIT ? isDIGIT(nextchr) : isDIGIT_LC(nextchr))) sayNO; nextchr = UCHARAT(++locinput); @@ -2326,19 +2275,8 @@ S_regmatch(pTHX_ regnode *prog) case NDIGIT: if (!nextchr && locinput >= PL_regeol) sayNO; - if (OP(scan) == NDIGIT - ? isDIGIT(nextchr) : isDIGIT_LC(nextchr)) - sayNO; - nextchr = UCHARAT(++locinput); - break; - case NDIGITLUTF8: - PL_reg_flags |= RF_tainted; - /* FALL THROUGH */ - case NDIGITUTF8: - if (!nextchr && locinput >= PL_regeol) - sayNO; - if (nextchr & 0x80) { - if (OP(scan) == NDIGITUTF8 + if (DO_UTF8(PL_reg_sv)) { + if (OP(scan) == NDIGIT ? swash_fetch(PL_utf8_digit, (U8*)locinput) : isDIGIT_LC_utf8((U8*)locinput)) { @@ -2348,7 +2286,7 @@ S_regmatch(pTHX_ regnode *prog) nextchr = UCHARAT(locinput); break; } - if (OP(scan) == NDIGITUTF8 + if (OP(scan) == NDIGIT ? isDIGIT(nextchr) : isDIGIT_LC(nextchr)) sayNO; nextchr = UCHARAT(++locinput); @@ -3452,35 +3390,37 @@ do_no: STATIC I32 S_regrepeat(pTHX_ regnode *p, I32 max) { - dTHR; register char *scan; register I32 c; register char *loceol = PL_regeol; register I32 hardcount = 0; + register bool do_utf8 = DO_UTF8(PL_reg_sv); scan = PL_reginput; if (max != REG_INFTY && max < loceol - scan) loceol = scan + max; switch (OP(p)) { case REG_ANY: - while (scan < loceol && *scan != '\n') - scan++; - break; - case SANY: - scan = loceol; - break; - case ANYUTF8: - loceol = PL_regeol; - while (scan < loceol && *scan != '\n') { - scan += UTF8SKIP(scan); - hardcount++; + if (DO_UTF8(PL_reg_sv)) { + loceol = PL_regeol; + while (scan < loceol && *scan != '\n') { + scan += UTF8SKIP(scan); + hardcount++; + } + } else { + while (scan < loceol && *scan != '\n') + scan++; } break; - case SANYUTF8: - loceol = PL_regeol; - while (scan < loceol) { - scan += UTF8SKIP(scan); - hardcount++; + case SANY: + if (DO_UTF8(PL_reg_sv)) { + loceol = PL_regeol; + while (scan < loceol) { + scan += UTF8SKIP(scan); + hardcount++; + } + } else { + scan = loceol; } break; case EXACT: /* length of string is 1 */ @@ -3501,135 +3441,144 @@ S_regrepeat(pTHX_ regnode *p, I32 max) (UCHARAT(scan) == c || UCHARAT(scan) == PL_fold_locale[c])) scan++; break; - case ANYOFUTF8: - loceol = PL_regeol; - while (scan < loceol && REGINCLASSUTF8(p, (U8*)scan)) { - scan += UTF8SKIP(scan); - hardcount++; - } - break; case ANYOF: - while (scan < loceol && REGINCLASS(p, *scan)) - scan++; + if (do_utf8) { + loceol = PL_regeol; + while (scan < loceol && reginclass(p, (U8*)scan, do_utf8)) { + scan += UTF8SKIP(scan); + hardcount++; + } + } else { + while (scan < loceol && reginclass(p, (U8*)scan, do_utf8)) + scan++; + } break; case ALNUM: - while (scan < loceol && isALNUM(*scan)) - scan++; - break; - case ALNUMUTF8: - loceol = PL_regeol; - while (scan < loceol && swash_fetch(PL_utf8_alnum, (U8*)scan)) { - scan += UTF8SKIP(scan); - hardcount++; + if (DO_UTF8(PL_reg_sv)) { + loceol = PL_regeol; + while (scan < loceol && swash_fetch(PL_utf8_alnum, (U8*)scan)) { + scan += UTF8SKIP(scan); + hardcount++; + } + } else { + while (scan < loceol && isALNUM(*scan)) + scan++; } break; case ALNUML: PL_reg_flags |= RF_tainted; - while (scan < loceol && isALNUM_LC(*scan)) - scan++; - break; - case ALNUMLUTF8: - PL_reg_flags |= RF_tainted; - loceol = PL_regeol; - while (scan < loceol && isALNUM_LC_utf8((U8*)scan)) { - scan += UTF8SKIP(scan); - hardcount++; + if (DO_UTF8(PL_reg_sv)) { + loceol = PL_regeol; + while (scan < loceol && isALNUM_LC_utf8((U8*)scan)) { + scan += UTF8SKIP(scan); + hardcount++; + } + } else { + while (scan < loceol && isALNUM_LC(*scan)) + scan++; } break; - break; case NALNUM: - while (scan < loceol && !isALNUM(*scan)) - scan++; - break; - case NALNUMUTF8: - loceol = PL_regeol; - while (scan < loceol && !swash_fetch(PL_utf8_alnum, (U8*)scan)) { - scan += UTF8SKIP(scan); - hardcount++; + if (DO_UTF8(PL_reg_sv)) { + loceol = PL_regeol; + while (scan < loceol && !swash_fetch(PL_utf8_alnum, (U8*)scan)) { + scan += UTF8SKIP(scan); + hardcount++; + } + } else { + while (scan < loceol && !isALNUM(*scan)) + scan++; } break; case NALNUML: PL_reg_flags |= RF_tainted; - while (scan < loceol && !isALNUM_LC(*scan)) - scan++; - break; - case NALNUMLUTF8: - PL_reg_flags |= RF_tainted; - loceol = PL_regeol; - while (scan < loceol && !isALNUM_LC_utf8((U8*)scan)) { - scan += UTF8SKIP(scan); - hardcount++; + if (DO_UTF8(PL_reg_sv)) { + loceol = PL_regeol; + while (scan < loceol && !isALNUM_LC_utf8((U8*)scan)) { + scan += UTF8SKIP(scan); + hardcount++; + } + } else { + while (scan < loceol && !isALNUM_LC(*scan)) + scan++; } break; case SPACE: - while (scan < loceol && isSPACE(*scan)) - scan++; - break; - case SPACEUTF8: - loceol = PL_regeol; - while (scan < loceol && (*scan == ' ' || swash_fetch(PL_utf8_space,(U8*)scan))) { - scan += UTF8SKIP(scan); - hardcount++; + if (DO_UTF8(PL_reg_sv)) { + loceol = PL_regeol; + while (scan < loceol && + (*scan == ' ' || swash_fetch(PL_utf8_space,(U8*)scan))) { + scan += UTF8SKIP(scan); + hardcount++; + } + } else { + while (scan < loceol && isSPACE(*scan)) + scan++; } break; case SPACEL: PL_reg_flags |= RF_tainted; - while (scan < loceol && isSPACE_LC(*scan)) - scan++; - break; - case SPACELUTF8: - PL_reg_flags |= RF_tainted; - loceol = PL_regeol; - while (scan < loceol && (*scan == ' ' || isSPACE_LC_utf8((U8*)scan))) { - scan += UTF8SKIP(scan); - hardcount++; + if (DO_UTF8(PL_reg_sv)) { + loceol = PL_regeol; + while (scan < loceol && + (*scan == ' ' || isSPACE_LC_utf8((U8*)scan))) { + scan += UTF8SKIP(scan); + hardcount++; + } + } else { + while (scan < loceol && isSPACE_LC(*scan)) + scan++; } break; case NSPACE: - while (scan < loceol && !isSPACE(*scan)) - scan++; - break; - case NSPACEUTF8: - loceol = PL_regeol; - while (scan < loceol && !(*scan == ' ' || swash_fetch(PL_utf8_space,(U8*)scan))) { - scan += UTF8SKIP(scan); - hardcount++; + if (DO_UTF8(PL_reg_sv)) { + loceol = PL_regeol; + while (scan < loceol && + !(*scan == ' ' || swash_fetch(PL_utf8_space,(U8*)scan))) { + scan += UTF8SKIP(scan); + hardcount++; + } + } else { + while (scan < loceol && !isSPACE(*scan)) + scan++; + break; } - break; case NSPACEL: PL_reg_flags |= RF_tainted; - while (scan < loceol && !isSPACE_LC(*scan)) - scan++; - break; - case NSPACELUTF8: - PL_reg_flags |= RF_tainted; - loceol = PL_regeol; - while (scan < loceol && !(*scan == ' ' || isSPACE_LC_utf8((U8*)scan))) { - scan += UTF8SKIP(scan); - hardcount++; + if (DO_UTF8(PL_reg_sv)) { + loceol = PL_regeol; + while (scan < loceol && + !(*scan == ' ' || isSPACE_LC_utf8((U8*)scan))) { + scan += UTF8SKIP(scan); + hardcount++; + } + } else { + while (scan < loceol && !isSPACE_LC(*scan)) + scan++; } break; case DIGIT: - while (scan < loceol && isDIGIT(*scan)) - scan++; - break; - case DIGITUTF8: - loceol = PL_regeol; - while (scan < loceol && swash_fetch(PL_utf8_digit,(U8*)scan)) { - scan += UTF8SKIP(scan); - hardcount++; + if (DO_UTF8(PL_reg_sv)) { + loceol = PL_regeol; + while (scan < loceol && swash_fetch(PL_utf8_digit,(U8*)scan)) { + scan += UTF8SKIP(scan); + hardcount++; + } + } else { + while (scan < loceol && isDIGIT(*scan)) + scan++; } break; - break; case NDIGIT: - while (scan < loceol && !isDIGIT(*scan)) - scan++; - break; - case NDIGITUTF8: - loceol = PL_regeol; - while (scan < loceol && !swash_fetch(PL_utf8_digit,(U8*)scan)) { - scan += UTF8SKIP(scan); - hardcount++; + if (DO_UTF8(PL_reg_sv)) { + loceol = PL_regeol; + while (scan < loceol && !swash_fetch(PL_utf8_digit,(U8*)scan)) { + scan += UTF8SKIP(scan); + hardcount++; + } + } else { + while (scan < loceol && !isDIGIT(*scan)) + scan++; } break; default: /* Called on something of 0 width. */ @@ -3664,7 +3613,6 @@ S_regrepeat(pTHX_ regnode *p, I32 max) STATIC I32 S_regrepeat_hard(pTHX_ regnode *p, I32 max, I32 *lp) { - dTHR; register char *scan; register char *start; register char *loceol = PL_regeol; @@ -3709,116 +3657,160 @@ S_regrepeat_hard(pTHX_ regnode *p, I32 max, I32 *lp) } /* +- regclass_swash - prepare the utf8 swash +*/ + +SV * +Perl_regclass_swash(pTHX_ register regnode* node, bool doinit, SV** initsvp) +{ + SV *sw = NULL; + SV *si = NULL; + + if (PL_regdata && PL_regdata->count) { + U32 n = ARG(node); + + if (PL_regdata->what[n] == 's') { + SV *rv = (SV*)PL_regdata->data[n]; + AV *av = (AV*)SvRV((SV*)rv); + SV **a; + + si = *av_fetch(av, 0, FALSE); + a = av_fetch(av, 1, FALSE); + + if (a) + sw = *a; + else if (si && doinit) { + sw = swash_init("utf8", "", si, 1, 0); + (void)av_store(av, 1, sw); + } + } + } + + if (initsvp) + *initsvp = si; + + return sw; +} + +/* - reginclass - determine if a character falls into a character class */ STATIC bool -S_reginclass(pTHX_ register regnode *p, register I32 c) +S_reginclass(pTHX_ register regnode *n, register U8* p, register bool do_utf8) { - dTHR; - char flags = ANYOF_FLAGS(p); + char flags = ANYOF_FLAGS(n); bool match = FALSE; - c &= 0xFF; - if (ANYOF_BITMAP_TEST(p, c)) - match = TRUE; - else if (flags & ANYOF_FOLD) { - I32 cf; - if (flags & ANYOF_LOCALE) { - PL_reg_flags |= RF_tainted; - cf = PL_fold_locale[c]; + if (do_utf8 || (flags & ANYOF_UNICODE)) { + if (do_utf8 && !ANYOF_RUNTIME(n)) { + STRLEN len; + UV c = utf8_to_uv_simple(p, &len); + + if (len != (STRLEN)-1 && c < 256 && ANYOF_BITMAP_TEST(n, c)) + match = TRUE; } - else - cf = PL_fold[c]; - if (ANYOF_BITMAP_TEST(p, cf)) - match = TRUE; - } - if (!match && (flags & ANYOF_CLASS)) { - PL_reg_flags |= RF_tainted; - if ( - (ANYOF_CLASS_TEST(p, ANYOF_ALNUM) && isALNUM_LC(c)) || - (ANYOF_CLASS_TEST(p, ANYOF_NALNUM) && !isALNUM_LC(c)) || - (ANYOF_CLASS_TEST(p, ANYOF_SPACE) && isSPACE_LC(c)) || - (ANYOF_CLASS_TEST(p, ANYOF_NSPACE) && !isSPACE_LC(c)) || - (ANYOF_CLASS_TEST(p, ANYOF_DIGIT) && isDIGIT_LC(c)) || - (ANYOF_CLASS_TEST(p, ANYOF_NDIGIT) && !isDIGIT_LC(c)) || - (ANYOF_CLASS_TEST(p, ANYOF_ALNUMC) && isALNUMC_LC(c)) || - (ANYOF_CLASS_TEST(p, ANYOF_NALNUMC) && !isALNUMC_LC(c)) || - (ANYOF_CLASS_TEST(p, ANYOF_ALPHA) && isALPHA_LC(c)) || - (ANYOF_CLASS_TEST(p, ANYOF_NALPHA) && !isALPHA_LC(c)) || - (ANYOF_CLASS_TEST(p, ANYOF_ASCII) && isASCII(c)) || - (ANYOF_CLASS_TEST(p, ANYOF_NASCII) && !isASCII(c)) || - (ANYOF_CLASS_TEST(p, ANYOF_CNTRL) && isCNTRL_LC(c)) || - (ANYOF_CLASS_TEST(p, ANYOF_NCNTRL) && !isCNTRL_LC(c)) || - (ANYOF_CLASS_TEST(p, ANYOF_GRAPH) && isGRAPH_LC(c)) || - (ANYOF_CLASS_TEST(p, ANYOF_NGRAPH) && !isGRAPH_LC(c)) || - (ANYOF_CLASS_TEST(p, ANYOF_LOWER) && isLOWER_LC(c)) || - (ANYOF_CLASS_TEST(p, ANYOF_NLOWER) && !isLOWER_LC(c)) || - (ANYOF_CLASS_TEST(p, ANYOF_PRINT) && isPRINT_LC(c)) || - (ANYOF_CLASS_TEST(p, ANYOF_NPRINT) && !isPRINT_LC(c)) || - (ANYOF_CLASS_TEST(p, ANYOF_PUNCT) && isPUNCT_LC(c)) || - (ANYOF_CLASS_TEST(p, ANYOF_NPUNCT) && !isPUNCT_LC(c)) || - (ANYOF_CLASS_TEST(p, ANYOF_UPPER) && isUPPER_LC(c)) || - (ANYOF_CLASS_TEST(p, ANYOF_NUPPER) && !isUPPER_LC(c)) || - (ANYOF_CLASS_TEST(p, ANYOF_XDIGIT) && isXDIGIT(c)) || - (ANYOF_CLASS_TEST(p, ANYOF_NXDIGIT) && !isXDIGIT(c)) || - (ANYOF_CLASS_TEST(p, ANYOF_PSXSPC) && isPSXSPC(c)) || - (ANYOF_CLASS_TEST(p, ANYOF_NPSXSPC) && !isPSXSPC(c)) || - (ANYOF_CLASS_TEST(p, ANYOF_BLANK) && isBLANK(c)) || - (ANYOF_CLASS_TEST(p, ANYOF_NBLANK) && !isBLANK(c)) - ) /* How's that for a conditional? */ - { - match = TRUE; + if (!match) { + SV *sw = regclass_swash(n, TRUE, 0); + + if (sw) { + if (swash_fetch(sw, p)) + match = TRUE; + else if (flags & ANYOF_FOLD) { + U8 tmpbuf[UTF8_MAXLEN+1]; + + if (flags & ANYOF_LOCALE) { + PL_reg_flags |= RF_tainted; + uv_to_utf8(tmpbuf, toLOWER_LC_utf8(p)); + } + else + uv_to_utf8(tmpbuf, toLOWER_utf8(p)); + if (swash_fetch(sw, tmpbuf)) + match = TRUE; + } + } } } + else { + U8 c = *p; - return (flags & ANYOF_INVERT) ? !match : match; -} - -STATIC bool -S_reginclassutf8(pTHX_ regnode *f, U8 *p) -{ - dTHR; - char flags = ARG1(f); - bool match = FALSE; - SV *sv = (SV*)PL_regdata->data[ARG2(f)]; + if (ANYOF_BITMAP_TEST(n, c)) + match = TRUE; + else if (flags & ANYOF_FOLD) { + I32 f; - if (swash_fetch(sv, p)) - match = TRUE; - else if (flags & ANYOF_FOLD) { - U8 tmpbuf[UTF8_MAXLEN]; - if (flags & ANYOF_LOCALE) { + if (flags & ANYOF_LOCALE) { + PL_reg_flags |= RF_tainted; + f = PL_fold_locale[c]; + } + else + f = PL_fold[c]; + if (f != c && ANYOF_BITMAP_TEST(n, f)) + match = TRUE; + } + + if (!match && (flags & ANYOF_CLASS)) { PL_reg_flags |= RF_tainted; - uv_to_utf8(tmpbuf, toLOWER_LC_utf8(p)); + if ( + (ANYOF_CLASS_TEST(n, ANYOF_ALNUM) && isALNUM_LC(c)) || + (ANYOF_CLASS_TEST(n, ANYOF_NALNUM) && !isALNUM_LC(c)) || + (ANYOF_CLASS_TEST(n, ANYOF_SPACE) && isSPACE_LC(c)) || + (ANYOF_CLASS_TEST(n, ANYOF_NSPACE) && !isSPACE_LC(c)) || + (ANYOF_CLASS_TEST(n, ANYOF_DIGIT) && isDIGIT_LC(c)) || + (ANYOF_CLASS_TEST(n, ANYOF_NDIGIT) && !isDIGIT_LC(c)) || + (ANYOF_CLASS_TEST(n, ANYOF_ALNUMC) && isALNUMC_LC(c)) || + (ANYOF_CLASS_TEST(n, ANYOF_NALNUMC) && !isALNUMC_LC(c)) || + (ANYOF_CLASS_TEST(n, ANYOF_ALPHA) && isALPHA_LC(c)) || + (ANYOF_CLASS_TEST(n, ANYOF_NALPHA) && !isALPHA_LC(c)) || + (ANYOF_CLASS_TEST(n, ANYOF_ASCII) && isASCII(c)) || + (ANYOF_CLASS_TEST(n, ANYOF_NASCII) && !isASCII(c)) || + (ANYOF_CLASS_TEST(n, ANYOF_CNTRL) && isCNTRL_LC(c)) || + (ANYOF_CLASS_TEST(n, ANYOF_NCNTRL) && !isCNTRL_LC(c)) || + (ANYOF_CLASS_TEST(n, ANYOF_GRAPH) && isGRAPH_LC(c)) || + (ANYOF_CLASS_TEST(n, ANYOF_NGRAPH) && !isGRAPH_LC(c)) || + (ANYOF_CLASS_TEST(n, ANYOF_LOWER) && isLOWER_LC(c)) || + (ANYOF_CLASS_TEST(n, ANYOF_NLOWER) && !isLOWER_LC(c)) || + (ANYOF_CLASS_TEST(n, ANYOF_PRINT) && isPRINT_LC(c)) || + (ANYOF_CLASS_TEST(n, ANYOF_NPRINT) && !isPRINT_LC(c)) || + (ANYOF_CLASS_TEST(n, ANYOF_PUNCT) && isPUNCT_LC(c)) || + (ANYOF_CLASS_TEST(n, ANYOF_NPUNCT) && !isPUNCT_LC(c)) || + (ANYOF_CLASS_TEST(n, ANYOF_UPPER) && isUPPER_LC(c)) || + (ANYOF_CLASS_TEST(n, ANYOF_NUPPER) && !isUPPER_LC(c)) || + (ANYOF_CLASS_TEST(n, ANYOF_XDIGIT) && isXDIGIT(c)) || + (ANYOF_CLASS_TEST(n, ANYOF_NXDIGIT) && !isXDIGIT(c)) || + (ANYOF_CLASS_TEST(n, ANYOF_PSXSPC) && isPSXSPC(c)) || + (ANYOF_CLASS_TEST(n, ANYOF_NPSXSPC) && !isPSXSPC(c)) || + (ANYOF_CLASS_TEST(n, ANYOF_BLANK) && isBLANK(c)) || + (ANYOF_CLASS_TEST(n, ANYOF_NBLANK) && !isBLANK(c)) + ) /* How's that for a conditional? */ + { + match = TRUE; + } } - else - uv_to_utf8(tmpbuf, toLOWER_utf8(p)); - if (swash_fetch(sv, tmpbuf)) - match = TRUE; } - /* UTF8 combined with ANYOF_CLASS is ill-defined. */ - return (flags & ANYOF_INVERT) ? !match : match; } STATIC U8 * S_reghop(pTHX_ U8 *s, I32 off) { - dTHR; if (off >= 0) { - while (off-- && s < (U8*)PL_regeol) + while (off-- && s < (U8*)PL_regeol) { + /* XXX could check well-formedness here */ s += UTF8SKIP(s); + } } else { while (off++) { if (s > (U8*)PL_bostr) { s--; - if (*s & 0x80) { - while (s > (U8*)PL_bostr && (*s & 0xc0) == 0x80) + if (UTF8_IS_CONTINUED(*s)) { + while (s > (U8*)PL_bostr && UTF8_IS_CONTINUATION(*s)) s--; - } /* XXX could check well-formedness here */ + } + /* XXX could check well-formedness here */ } } } @@ -3828,10 +3820,11 @@ S_reghop(pTHX_ U8 *s, I32 off) STATIC U8 * S_reghopmaybe(pTHX_ U8* s, I32 off) { - dTHR; if (off >= 0) { - while (off-- && s < (U8*)PL_regeol) + while (off-- && s < (U8*)PL_regeol) { + /* XXX could check well-formedness here */ s += UTF8SKIP(s); + } if (off >= 0) return 0; } @@ -3839,10 +3832,11 @@ S_reghopmaybe(pTHX_ U8* s, I32 off) while (off++) { if (s > (U8*)PL_bostr) { s--; - if (*s & 0x80) { - while (s > (U8*)PL_bostr && (*s & 0xc0) == 0x80) + if (UTF8_IS_CONTINUED(*s)) { + while (s > (U8*)PL_bostr && UTF8_IS_CONTINUATION(*s)) s--; - } /* XXX could check well-formedness here */ + } + /* XXX could check well-formedness here */ } else break; @@ -3860,7 +3854,6 @@ S_reghopmaybe(pTHX_ U8* s, I32 off) static void restore_pos(pTHXo_ void *arg) { - dTHR; if (PL_reg_eval_set) { if (PL_reg_oldsaved) { PL_reg_re->subbeg = PL_reg_oldsaved;