X-Git-Url: http://git.shadowcat.co.uk/gitweb/gitweb.cgi?a=blobdiff_plain;f=regcomp.c;h=fe54a6c45a38d4b2066c20361199983524509930;hb=4a2d328fcdc90a0fcbd435b479571aad60159771;hp=59fe5a7d9ff9493e6b9adabf6263a217c2223274;hpb=f722798beaa437498e1f879b01996be44f2fdc1a;p=p5sagit%2Fp5-mst-13.2.git diff --git a/regcomp.c b/regcomp.c index 59fe5a7..fe54a6c 100644 --- a/regcomp.c +++ b/regcomp.c @@ -41,6 +41,8 @@ # define Perl_regnext my_regnext # define Perl_save_re_context my_save_re_context # define Perl_reginitcolors my_reginitcolors + +# define PERL_NO_GET_CONTEXT #endif /*SUPPRESS 112*/ @@ -81,7 +83,11 @@ #define PERL_IN_REGCOMP_C #include "perl.h" -#ifndef PERL_IN_XSUB_RE +#ifdef PERL_IN_XSUB_RE +# if defined(PERL_CAPI) || defined(PERL_OBJECT) +# include "XSUB.h" +# endif +#else # include "INTERN.h" #endif @@ -126,12 +132,33 @@ #define SPSTART 0x4 /* Starts with * or +. */ #define TRYAGAIN 0x8 /* Weeded out a declaration. */ +/* Length of a variant. */ + +typedef struct scan_data_t { + I32 len_min; + I32 len_delta; + I32 pos_min; + I32 pos_delta; + SV *last_found; + I32 last_end; /* min value, <0 unless valid. */ + I32 last_start_min; + I32 last_start_max; + SV **longest; /* Either &l_fixed, or &l_float. */ + SV *longest_fixed; + I32 offset_fixed; + SV *longest_float; + I32 offset_float_min; + I32 offset_float_max; + I32 flags; + I32 whilem_c; +} scan_data_t; + /* * Forward declarations for pregcomp()'s friends. */ static scan_data_t zero_scan_data = { 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, - 0, 0, 0 }; + 0, 0, 0, 0 }; #define SF_BEFORE_EOL (SF_BEFORE_SEOL|SF_BEFORE_MEOL) #define SF_BEFORE_SEOL 0x1 @@ -163,14 +190,17 @@ static scan_data_t zero_scan_data = { 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, #define LOC (PL_regflags & PMf_LOCALE) #define FOLD (PL_regflags & PMf_FOLD) +#define OOB_CHAR8 1234 +#define OOB_UTF8 123456 +#define OOB_NAMEDCLASS -1 + #define CHR_SVLEN(sv) (UTF ? sv_len_utf8(sv) : SvCUR(sv)) #define CHR_DIST(a,b) (UTF ? utf8_distance(a,b) : a - b) -STATIC void -S_clear_re(pTHX_ void *r) -{ - ReREFCNT_dec((regexp *)r); -} +/* Allow for side effects in s */ +#define REGC(c,s) STMT_START { if (!SIZE_ONLY) *(s) = (c); else (s);} STMT_END + +static void clear_re(pTHXo_ void *r); STATIC void S_scan_commit(pTHX_ scan_data_t *data) @@ -234,7 +264,7 @@ S_study_chunk(pTHX_ regnode **scanp, I32 *deltap, regnode *last, scan_data_t *da regnode *stop = scan; #endif - next = scan + (*OPERAND(scan) + 2 - 1)/sizeof(regnode) + 2; + next = scan + NODE_SZ_STR(scan); /* Skip NOTHING, merge EXACT*. */ while (n && ( PL_regkind[(U8)OP(n)] == NOTHING || @@ -253,17 +283,17 @@ S_study_chunk(pTHX_ regnode **scanp, I32 *deltap, regnode *last, scan_data_t *da n = regnext(n); } else { - int oldl = *OPERAND(scan); + int oldl = STR_LEN(scan); regnode *nnext = regnext(n); - if (oldl + *OPERAND(n) > U8_MAX) + if (oldl + STR_LEN(n) > U8_MAX) break; NEXT_OFF(scan) += NEXT_OFF(n); - *OPERAND(scan) += *OPERAND(n); - next = n + (*OPERAND(n) + 2 - 1)/sizeof(regnode) + 2; + STR_LEN(scan) += STR_LEN(n); + next = n + NODE_SZ_STR(n); /* Now we can overwrite *n : */ - Move(OPERAND(n) + 1, OPERAND(scan) + oldl + 1, - *OPERAND(n) + 1, char); + Move(STRING(n), STRING(scan) + oldl, + STR_LEN(n), char); #ifdef DEBUGGING if (stringok) stop = next - 1; @@ -273,7 +303,7 @@ S_study_chunk(pTHX_ regnode **scanp, I32 *deltap, regnode *last, scan_data_t *da } #ifdef DEBUGGING /* Allow dumping */ - n = scan + (*OPERAND(scan) + 2 - 1)/sizeof(regnode) + 2; + n = scan + NODE_SZ_STR(scan); while (n <= stop) { /* Purify reports a benign UMR here sometimes, because we * don't initialize the OP() slot of a node when that node @@ -323,6 +353,8 @@ S_study_chunk(pTHX_ regnode **scanp, I32 *deltap, regnode *last, scan_data_t *da num++; data_fake.flags = 0; + if (data) + data_fake.whilem_c = data->whilem_c; next = regnext(scan); scan = NEXTOPER(scan); if (code != BRANCH) @@ -341,6 +373,8 @@ S_study_chunk(pTHX_ regnode **scanp, I32 *deltap, regnode *last, scan_data_t *da pars++; if (data && (data_fake.flags & SF_HAS_EVAL)) data->flags |= SF_HAS_EVAL; + if (data) + data->whilem_c = data_fake.whilem_c; if (code == SUSPEND) break; } @@ -362,9 +396,9 @@ S_study_chunk(pTHX_ regnode **scanp, I32 *deltap, regnode *last, scan_data_t *da continue; } else if (OP(scan) == EXACT) { - I32 l = *OPERAND(scan); + I32 l = STR_LEN(scan); if (UTF) { - unsigned char *s = (unsigned char *)(OPERAND(scan)+1); + unsigned char *s = (unsigned char *)STRING(scan); unsigned char *e = s + l; I32 newl = 0; while (s < e) { @@ -382,18 +416,18 @@ S_study_chunk(pTHX_ regnode **scanp, I32 *deltap, regnode *last, scan_data_t *da data->last_start_max = is_inf ? I32_MAX : data->pos_min + data->pos_delta; } - sv_catpvn(data->last_found, (char *)(OPERAND(scan)+1), *OPERAND(scan)); + sv_catpvn(data->last_found, STRING(scan), STR_LEN(scan)); data->last_end = data->pos_min + l; data->pos_min += l; /* As in the first entry. */ data->flags &= ~SF_BEFORE_EOL; } } else if (PL_regkind[(U8)OP(scan)] == EXACT) { - I32 l = *OPERAND(scan); + I32 l = STR_LEN(scan); if (flags & SCF_DO_SUBSTR) scan_commit(data); if (UTF) { - unsigned char *s = (unsigned char *)(OPERAND(scan)+1); + unsigned char *s = (unsigned char *)STRING(scan); unsigned char *e = s + l; I32 newl = 0; while (s < e) { @@ -406,7 +440,7 @@ S_study_chunk(pTHX_ regnode **scanp, I32 *deltap, regnode *last, scan_data_t *da if (data && (flags & SCF_DO_SUBSTR)) data->pos_min += l; } - else if (strchr(PL_varies,OP(scan))) { + else if (strchr((char*)PL_varies,OP(scan))) { I32 mincount, maxcount, minnext, deltanext, pos_before, fl; regnode *oscan = scan; @@ -462,7 +496,8 @@ S_study_chunk(pTHX_ regnode **scanp, I32 *deltap, regnode *last, scan_data_t *da if (ckWARN(WARN_UNSAFE) && (minnext + deltanext == 0) && !(data->flags & (SF_HAS_PAR|SF_IN_PAR)) && maxcount <= REG_INFTY/3) /* Complement check for big count */ - Perl_warner(aTHX_ WARN_UNSAFE, "Strange *+?{} on zero-length expression"); + Perl_warner(aTHX_ WARN_UNSAFE, + "Strange *+?{} on zero-length expression"); min += minnext * mincount; is_inf_internal |= (maxcount == REG_INFTY && (minnext + deltanext) > 0 @@ -481,9 +516,9 @@ S_study_chunk(pTHX_ regnode **scanp, I32 *deltap, regnode *last, scan_data_t *da /* Skip open. */ nxt = regnext(nxt); - if (!strchr(PL_simple,OP(nxt)) + if (!strchr((char*)PL_simple,OP(nxt)) && !(PL_regkind[(U8)OP(nxt)] == EXACT - && *OPERAND(nxt) == 1)) + && STR_LEN(nxt) == 1)) goto nogo; nxt2 = nxt; nxt = regnext(nxt); @@ -556,6 +591,16 @@ S_study_chunk(pTHX_ regnode **scanp, I32 *deltap, regnode *last, scan_data_t *da else oscan->flags = 0; } + else if (OP(oscan) == CURLYX && data && ++data->whilem_c < 16) { + /* This stays as CURLYX, and can put the count/of pair. */ + /* Find WHILEM (as in regexec.c) */ + regnode *nxt = oscan + NEXT_OFF(oscan); + + if (OP(PREVOPER(nxt)) == NOTHING) /* LONGJMP */ + nxt += ARG(nxt); + PREVOPER(nxt)->flags = data->whilem_c + | (PL_reg_whilem_seen << 4); /* On WHILEM */ + } if (data && fl & (SF_HAS_PAR|SF_IN_PAR)) pars++; if (flags & SCF_DO_SUBSTR) { @@ -628,7 +673,7 @@ S_study_chunk(pTHX_ regnode **scanp, I32 *deltap, regnode *last, scan_data_t *da break; } } - else if (strchr(PL_simple,OP(scan)) || PL_regkind[(U8)OP(scan)] == ANYUTF8) { + else if (strchr((char*)PL_simple,OP(scan)) || PL_regkind[(U8)OP(scan)] == ANYUTF8) { if (flags & SCF_DO_SUBSTR) { scan_commit(data); data->pos_min++; @@ -647,6 +692,8 @@ S_study_chunk(pTHX_ regnode **scanp, I32 *deltap, regnode *last, scan_data_t *da regnode *nscan; data_fake.flags = 0; + if (data) + data_fake.whilem_c = data->whilem_c; next = regnext(scan); nscan = NEXTOPER(NEXTOPER(scan)); minnext = study_chunk(&nscan, &deltanext, last, &data_fake, 0); @@ -655,7 +702,11 @@ S_study_chunk(pTHX_ regnode **scanp, I32 *deltap, regnode *last, scan_data_t *da FAIL("variable length lookbehind not implemented"); } else if (minnext > U8_MAX) { +#ifdef UV_IS_QUAD + FAIL2("lookbehind longer than %" PERL_PRIu64 " not implemented", (UV)U8_MAX); +#else FAIL2("lookbehind longer than %d not implemented", U8_MAX); +#endif } scan->flags = minnext; } @@ -663,6 +714,8 @@ S_study_chunk(pTHX_ regnode **scanp, I32 *deltap, regnode *last, scan_data_t *da pars++; if (data && (data_fake.flags & SF_HAS_EVAL)) data->flags |= SF_HAS_EVAL; + if (data) + data->whilem_c = data_fake.whilem_c; } else if (OP(scan) == OPEN) { pars++; @@ -781,6 +834,7 @@ Perl_pregcomp(pTHX_ char *exp, char *xend, PMOP *pm) I32 minlen = 0; I32 sawplus = 0; I32 sawopen = 0; + scan_data_t data; if (exp == NULL) FAIL("NULL regexp argument"); @@ -792,7 +846,7 @@ Perl_pregcomp(pTHX_ char *exp, char *xend, PMOP *pm) PL_regprecomp = savepvn(exp, xend - exp); DEBUG_r(if (!PL_colorset) reginitcolors()); - DEBUG_r(PerlIO_printf(Perl_debug_log, "%sCompiling%s RE `%s%*s%s'\n", + DEBUG_r(PerlIO_printf(Perl_debug_log, "%sCompiling REx%s `%s%*s%s'\n", PL_colors[4],PL_colors[5],PL_colors[0], xend - exp, PL_regprecomp, PL_colors[1])); PL_regflags = pm->op_pmflags; @@ -810,7 +864,8 @@ Perl_pregcomp(pTHX_ char *exp, char *xend, PMOP *pm) PL_regnpar = 1; PL_regsize = 0L; PL_regcode = &PL_regdummy; - regc((U8)REG_MAGIC, (char*)PL_regcode); + PL_reg_whilem_seen = 0; + REGC((U8)REG_MAGIC, (char*)PL_regcode); if (reg(0, &flags) == NULL) { Safefree(PL_regprecomp); PL_regprecomp = Nullch; @@ -824,6 +879,8 @@ Perl_pregcomp(pTHX_ char *exp, char *xend, PMOP *pm) PL_regsize += PL_extralen; else PL_extralen = 0; + if (PL_reg_whilem_seen > 15) + PL_reg_whilem_seen = 15; /* Allocate space and initialize. */ Newc(1001, r, sizeof(regexp) + (unsigned)PL_regsize * sizeof(regnode), @@ -851,7 +908,7 @@ Perl_pregcomp(pTHX_ char *exp, char *xend, PMOP *pm) PL_regcode = r->program; /* Store the count of eval-groups for security checks: */ PL_regcode->next_off = ((PL_seen_evals > U16_MAX) ? U16_MAX : PL_seen_evals); - regc((U8)REG_MAGIC, (char*) PL_regcode++); + REGC((U8)REG_MAGIC, (char*) PL_regcode++); r->data = 0; if (reg(0, &flags) == NULL) return(NULL); @@ -870,12 +927,11 @@ Perl_pregcomp(pTHX_ char *exp, char *xend, PMOP *pm) 3-units-long substrs field. */ Newz(1004, r->substrs, 1, struct reg_substr_data); + StructCopy(&zero_scan_data, &data, scan_data_t); if (OP(scan) != BRANCH) { /* Only one top-level choice. */ - scan_data_t data; I32 fake; STRLEN longest_float_length, longest_fixed_length; - StructCopy(&zero_scan_data, &data, scan_data_t); first = scan; /* Skip introductions and multiplicators >= 1. */ while ((OP(first) == OPEN && (sawopen = 1)) || @@ -893,7 +949,7 @@ Perl_pregcomp(pTHX_ char *exp, char *xend, PMOP *pm) /* Starting-point info. */ again: if (OP(first) == EXACT); /* Empty, get anchored substr later. */ - else if (strchr(PL_simple+4,OP(first))) + else if (strchr((char*)PL_simple+4,OP(first))) r->regstclass = first; else if (PL_regkind[(U8)OP(first)] == BOUND || PL_regkind[(U8)OP(first)] == NBOUND) @@ -1036,7 +1092,7 @@ Perl_pregcomp(pTHX_ char *exp, char *xend, PMOP *pm) DEBUG_r(PerlIO_printf(Perl_debug_log, "\n")); scan = r->program + 1; - minlen = study_chunk(&scan, &fake, scan + PL_regsize, NULL, 0); + minlen = study_chunk(&scan, &fake, scan + PL_regsize, &data, 0); r->check_substr = r->anchored_substr = r->float_substr = Nullsv; } @@ -1514,8 +1570,10 @@ S_regpiece(pTHX_ I32 *flagp) reginsert(CURLY, ret); } else { - PL_regnaughty += 4 + PL_regnaughty; /* compound interest */ - regtail(ret, reg_node(WHILEM)); + regnode *w = reg_node(WHILEM); + + w->flags = 0; + regtail(ret, w); if (!SIZE_ONLY && PL_extralen) { reginsert(LONGJMP,ret); reginsert(NOTHING,ret); @@ -1526,7 +1584,8 @@ S_regpiece(pTHX_ I32 *flagp) NEXT_OFF(ret) = 3; /* Go over NOTHING to LONGJMP. */ regtail(ret, reg_node(NOTHING)); if (SIZE_ONLY) - PL_extralen += 3; + PL_reg_whilem_seen++, PL_extralen += 3; + PL_regnaughty += 4 + PL_regnaughty; /* compound interest */ } ret->flags = 0; @@ -1898,8 +1957,7 @@ tryagain: ret = reg_node(FOLD ? (LOC ? EXACTFL : EXACTF) : EXACT); - s = (char *) OPERAND(ret); - regc(0, s++); /* save spot for len */ + s = STRING(ret); for (len = 0, p = PL_regcomp_parse - 1; len < 127 && p < PL_regxend; len++) @@ -2039,7 +2097,7 @@ tryagain: } else { len++; - regc(ender, s++); + REGC(ender, s++); } break; } @@ -2049,7 +2107,7 @@ tryagain: len += numlen - 1; } else - regc(ender, s++); + REGC(ender, s++); } loopdone: PL_regcomp_parse = p - 1; @@ -2061,14 +2119,11 @@ tryagain: if (len == 1) *flagp |= SIMPLE; if (!SIZE_ONLY) - *OPERAND(ret) = len; - regc('\0', s++); - if (SIZE_ONLY) { - PL_regsize += (len + 2 + sizeof(regnode) - 1) / sizeof(regnode); - } - else { - PL_regcode += (len + 2 + sizeof(regnode) - 1) / sizeof(regnode); - } + STR_LEN(ret) = len; + if (SIZE_ONLY) + PL_regsize += STR_SZ(len); + else + PL_regcode += STR_SZ(len); } break; } @@ -2093,12 +2148,17 @@ S_regwhite(pTHX_ char *p, char *e) return p; } -/* parse POSIX character classes like [[:foo:]] */ -STATIC char* +/* Parse POSIX character classes: [[:foo:]], [[=foo=]], [[.foo.]]. + Character classes ([:foo:]) can also be negated ([:^foo:]). + Returns a named class id (ANYOF_XXX) if successful, -1 otherwise. + Equivalence classes ([=foo=]) and composites ([.foo.]) are parsed, + but trigger warnings because they are currently unimplemented. */ +STATIC I32 S_regpposixcc(pTHX_ I32 value) { dTHR; char *posixcc = 0; + I32 namedclass = -1; if (value == '[' && PL_regcomp_parse + 1 < PL_regxend && /* I smell either [: or [= or [. -- POSIX has been here, right? */ @@ -2114,26 +2174,120 @@ S_regpposixcc(pTHX_ I32 value) /* Grandfather lone [:, [=, [. */ PL_regcomp_parse = s; else { - PL_regcomp_parse++; /* skip over the c */ - if (*PL_regcomp_parse == ']') { - /* Not Implemented Yet. - * (POSIX Extended Character Classes, that is) - * The text between e.g. [: and :] would start - * at s + 1 and stop at regcomp_parse - 2. */ - if (ckWARN(WARN_UNSAFE) && !SIZE_ONLY) + char* t = PL_regcomp_parse++; /* skip over the c */ + + if (*PL_regcomp_parse == ']') { + PL_regcomp_parse++; /* skip over the ending ] */ + posixcc = s + 1; + if (*s == ':') { + I32 complement = *posixcc == '^' ? *posixcc++ : 0; + I32 skip = 5; /* the most common skip */ + + switch (*posixcc) { + case 'a': + if (strnEQ(posixcc, "alnum", 5)) + namedclass = + complement ? ANYOF_NALNUMC : ANYOF_ALNUMC; + else if (strnEQ(posixcc, "alpha", 5)) + namedclass = + complement ? ANYOF_NALPHA : ANYOF_ALPHA; + else if (strnEQ(posixcc, "ascii", 5)) + namedclass = + complement ? ANYOF_NASCII : ANYOF_ASCII; + break; + case 'c': + if (strnEQ(posixcc, "cntrl", 5)) + namedclass = + complement ? ANYOF_NCNTRL : ANYOF_CNTRL; + break; + case 'd': + if (strnEQ(posixcc, "digit", 5)) + namedclass = + complement ? ANYOF_NDIGIT : ANYOF_DIGIT; + break; + case 'g': + if (strnEQ(posixcc, "graph", 5)) + namedclass = + complement ? ANYOF_NGRAPH : ANYOF_GRAPH; + break; + case 'l': + if (strnEQ(posixcc, "lower", 5)) + namedclass = + complement ? ANYOF_NLOWER : ANYOF_LOWER; + break; + case 'p': + if (strnEQ(posixcc, "print", 5)) + namedclass = + complement ? ANYOF_NPRINT : ANYOF_PRINT; + else if (strnEQ(posixcc, "punct", 5)) + namedclass = + complement ? ANYOF_NPUNCT : ANYOF_PUNCT; + break; + case 's': + if (strnEQ(posixcc, "space", 5)) + namedclass = + complement ? ANYOF_NSPACE : ANYOF_SPACE; + case 'u': + if (strnEQ(posixcc, "upper", 5)) + namedclass = + complement ? ANYOF_NUPPER : ANYOF_UPPER; + break; + case 'w': /* this is not POSIX, this is the Perl \w */ + if (strnEQ(posixcc, "word", 4)) { + namedclass = + complement ? ANYOF_NALNUM : ANYOF_ALNUM; + skip = 4; + } + break; + case 'x': + if (strnEQ(posixcc, "xdigit", 6)) { + namedclass = + complement ? ANYOF_NXDIGIT : ANYOF_XDIGIT; + skip = 6; + } + break; + } + if ((namedclass == OOB_NAMEDCLASS || + !(posixcc + skip + 2 < PL_regxend && + (posixcc[skip] == ':' && + posixcc[skip + 1] == ']')))) + Perl_croak(aTHX_ "Character class [:%.*s:] unknown", + t - s - 1, s + 1); + } else if (ckWARN(WARN_UNSAFE) && !SIZE_ONLY) + /* [[=foo=]] and [[.foo.]] are still future. */ Perl_warner(aTHX_ WARN_UNSAFE, - "Character class syntax [%c %c] is reserved for future extensions", c, c); - PL_regcomp_parse++; /* skip over the ending ] */ - posixcc = s + 1; - } - else { - /* maternal grandfather */ + "Character class syntax [%c %c] is reserved for future extensions", c, c); + } else { + /* Maternal grandfather: + * "[:" ending in ":" but not in ":]" */ PL_regcomp_parse = s; } } } - return posixcc; + return namedclass; +} + +STATIC void +S_checkposixcc(pTHX) +{ + if (ckWARN(WARN_UNSAFE) && !SIZE_ONLY && + (*PL_regcomp_parse == ':' || + *PL_regcomp_parse == '=' || + *PL_regcomp_parse == '.')) { + char *s = PL_regcomp_parse; + char c = *s++; + + while(*s && isALNUM(*s)) + s++; + if (*s && c == *s && s[1] == ']') { + Perl_warner(aTHX_ WARN_UNSAFE, + "Character class syntax [%c %c] belongs inside character classes", c, c); + if (c == '=' || c == '.') + Perl_warner(aTHX_ WARN_UNSAFE, + "Character class syntax [%c %c] is reserved for future extensions", c, c); + } + } } STATIC regnode * @@ -2142,158 +2296,339 @@ S_regclass(pTHX) dTHR; register char *opnd, *s; register I32 value; - register I32 lastvalue = 1234; + register I32 lastvalue = OOB_CHAR8; register I32 range = 0; register regnode *ret; register I32 def; I32 numlen; + I32 namedclass; - s = opnd = (char *) OPERAND(PL_regcode); + s = opnd = MASK(PL_regcode); ret = reg_node(ANYOF); - for (value = 0; value < 33; value++) - regc(0, s++); + for (value = 0; value < ANYOF_SIZE; value++) + REGC(0, s++); if (*PL_regcomp_parse == '^') { /* Complement of range. */ PL_regnaughty++; PL_regcomp_parse++; if (!SIZE_ONLY) - *opnd |= ANYOF_INVERT; + ANYOF_FLAGS(opnd) |= ANYOF_INVERT; } if (!SIZE_ONLY) { PL_regcode += ANY_SKIP; if (FOLD) - *opnd |= ANYOF_FOLD; + ANYOF_FLAGS(opnd) |= ANYOF_FOLD; if (LOC) - *opnd |= ANYOF_LOCALE; + ANYOF_FLAGS(opnd) |= ANYOF_LOCALE; } else { PL_regsize += ANY_SKIP; } + + checkposixcc(); + if (*PL_regcomp_parse == ']' || *PL_regcomp_parse == '-') goto skipcond; /* allow 1st char to be ] or - */ while (PL_regcomp_parse < PL_regxend && *PL_regcomp_parse != ']') { skipcond: + namedclass = OOB_NAMEDCLASS; value = UCHARAT(PL_regcomp_parse++); if (value == '[') - (void)regpposixcc(value); /* ignore the return value for now */ + namedclass = regpposixcc(value); else if (value == '\\') { value = UCHARAT(PL_regcomp_parse++); switch (value) { - case 'w': - if (!SIZE_ONLY) { - if (LOC) - *opnd |= ANYOF_ALNUML; - else { - for (value = 0; value < 256; value++) - if (isALNUM(value)) - ANYOF_SET(opnd, value); - } + case 'w': namedclass = ANYOF_ALNUM; break; + case 'W': namedclass = ANYOF_NALNUM; break; + case 's': namedclass = ANYOF_SPACE; break; + case 'S': namedclass = ANYOF_NSPACE; break; + case 'd': namedclass = ANYOF_DIGIT; break; + case 'D': namedclass = ANYOF_NDIGIT; break; + case 'n': value = '\n'; break; + case 'r': value = '\r'; break; + case 't': value = '\t'; break; + case 'f': value = '\f'; break; + case 'b': value = '\b'; break; + case 'e': value = '\033'; break; + case 'a': value = '\007'; break; + case 'x': + value = scan_hex(PL_regcomp_parse, 2, &numlen); + PL_regcomp_parse += numlen; + break; + case 'c': + value = UCHARAT(PL_regcomp_parse++); + value = toCTRL(value); + break; + case '0': case '1': case '2': case '3': case '4': + case '5': case '6': case '7': case '8': case '9': + value = scan_oct(--PL_regcomp_parse, 3, &numlen); + PL_regcomp_parse += numlen; + break; + } + } + if (!SIZE_ONLY && namedclass > OOB_NAMEDCLASS) { + if (range) + FAIL("invalid [] range in regexp"); /* [a-\w], [a-[:word:]] */ + switch (namedclass) { + case ANYOF_ALNUM: + if (LOC) + ANYOF_CLASS_SET(opnd, ANYOF_ALNUM); + else { + for (value = 0; value < 256; value++) + if (isALNUM(value)) + ANYOF_BITMAP_SET(opnd, value); } - lastvalue = 1234; - continue; - case 'W': - if (!SIZE_ONLY) { - if (LOC) - *opnd |= ANYOF_NALNUML; - else { - for (value = 0; value < 256; value++) - if (!isALNUM(value)) - ANYOF_SET(opnd, value); - } + break; + case ANYOF_NALNUM: + if (LOC) + ANYOF_CLASS_SET(opnd, ANYOF_NALNUM); + else { + for (value = 0; value < 256; value++) + if (!isALNUM(value)) + ANYOF_BITMAP_SET(opnd, value); } - lastvalue = 1234; - continue; - case 's': - if (!SIZE_ONLY) { - if (LOC) - *opnd |= ANYOF_SPACEL; - else { - for (value = 0; value < 256; value++) - if (isSPACE(value)) - ANYOF_SET(opnd, value); - } + break; + case ANYOF_SPACE: + if (LOC) + ANYOF_CLASS_SET(opnd, ANYOF_SPACE); + else { + for (value = 0; value < 256; value++) + if (isSPACE(value)) + ANYOF_BITMAP_SET(opnd, value); } - lastvalue = 1234; - continue; - case 'S': - if (!SIZE_ONLY) { - if (LOC) - *opnd |= ANYOF_NSPACEL; - else { - for (value = 0; value < 256; value++) - if (!isSPACE(value)) - ANYOF_SET(opnd, value); - } + break; + case ANYOF_NSPACE: + if (LOC) + ANYOF_CLASS_SET(opnd, ANYOF_NSPACE); + else { + for (value = 0; value < 256; value++) + if (!isSPACE(value)) + ANYOF_BITMAP_SET(opnd, value); } - lastvalue = 1234; - continue; - case 'd': - if (!SIZE_ONLY) { + break; + case ANYOF_DIGIT: + if (LOC) + ANYOF_CLASS_SET(opnd, ANYOF_DIGIT); + else { for (value = '0'; value <= '9'; value++) - ANYOF_SET(opnd, value); + ANYOF_BITMAP_SET(opnd, value); } - lastvalue = 1234; - continue; - case 'D': - if (!SIZE_ONLY) { + break; + case ANYOF_NDIGIT: + if (LOC) + ANYOF_CLASS_SET(opnd, ANYOF_NDIGIT); + else { for (value = 0; value < '0'; value++) - ANYOF_SET(opnd, value); + ANYOF_BITMAP_SET(opnd, value); for (value = '9' + 1; value < 256; value++) - ANYOF_SET(opnd, value); + ANYOF_BITMAP_SET(opnd, value); + } + break; + case ANYOF_NALNUMC: + if (LOC) + ANYOF_CLASS_SET(opnd, ANYOF_NALNUMC); + else { + for (value = 0; value < 256; value++) + if (!isALNUMC(value)) + ANYOF_BITMAP_SET(opnd, value); + } + break; + case ANYOF_ALNUMC: + if (LOC) + ANYOF_CLASS_SET(opnd, ANYOF_ALNUMC); + else { + for (value = 0; value < 256; value++) + if (isALNUMC(value)) + ANYOF_BITMAP_SET(opnd, value); + } + break; + case ANYOF_ALPHA: + if (LOC) + ANYOF_CLASS_SET(opnd, ANYOF_ALPHA); + else { + for (value = 0; value < 256; value++) + if (isALPHA(value)) + ANYOF_BITMAP_SET(opnd, value); } - lastvalue = 1234; - continue; - case 'n': - value = '\n'; break; - case 'r': - value = '\r'; + case ANYOF_NALPHA: + if (LOC) + ANYOF_CLASS_SET(opnd, ANYOF_NALPHA); + else { + for (value = 0; value < 256; value++) + if (!isALPHA(value)) + ANYOF_BITMAP_SET(opnd, value); + } break; - case 't': - value = '\t'; + case ANYOF_ASCII: + if (LOC) + ANYOF_CLASS_SET(opnd, ANYOF_ASCII); + else { + for (value = 0; value < 128; value++) + ANYOF_BITMAP_SET(opnd, value); + } break; - case 'f': - value = '\f'; + case ANYOF_NASCII: + if (LOC) + ANYOF_CLASS_SET(opnd, ANYOF_NASCII); + else { + for (value = 128; value < 256; value++) + ANYOF_BITMAP_SET(opnd, value); + } break; - case 'b': - value = '\b'; + case ANYOF_CNTRL: + if (LOC) + ANYOF_CLASS_SET(opnd, ANYOF_CNTRL); + else { + for (value = 0; value < 256; value++) + if (isCNTRL(value)) + ANYOF_BITMAP_SET(opnd, value); + } + lastvalue = OOB_CHAR8; break; - case 'e': - value = '\033'; + case ANYOF_NCNTRL: + if (LOC) + ANYOF_CLASS_SET(opnd, ANYOF_NCNTRL); + else { + for (value = 0; value < 256; value++) + if (!isCNTRL(value)) + ANYOF_BITMAP_SET(opnd, value); + } break; - case 'a': - value = '\007'; + case ANYOF_GRAPH: + if (LOC) + ANYOF_CLASS_SET(opnd, ANYOF_GRAPH); + else { + for (value = 0; value < 256; value++) + if (isGRAPH(value)) + ANYOF_BITMAP_SET(opnd, value); + } break; - case 'x': - value = scan_hex(PL_regcomp_parse, 2, &numlen); - PL_regcomp_parse += numlen; + case ANYOF_NGRAPH: + if (LOC) + ANYOF_CLASS_SET(opnd, ANYOF_NGRAPH); + else { + for (value = 0; value < 256; value++) + if (!isGRAPH(value)) + ANYOF_BITMAP_SET(opnd, value); + } break; - case 'c': - value = UCHARAT(PL_regcomp_parse++); - value = toCTRL(value); + case ANYOF_LOWER: + if (LOC) + ANYOF_CLASS_SET(opnd, ANYOF_LOWER); + else { + for (value = 0; value < 256; value++) + if (isLOWER(value)) + ANYOF_BITMAP_SET(opnd, value); + } break; - case '0': case '1': case '2': case '3': case '4': - case '5': case '6': case '7': case '8': case '9': - value = scan_oct(--PL_regcomp_parse, 3, &numlen); - PL_regcomp_parse += numlen; + case ANYOF_NLOWER: + if (LOC) + ANYOF_CLASS_SET(opnd, ANYOF_NLOWER); + else { + for (value = 0; value < 256; value++) + if (!isLOWER(value)) + ANYOF_BITMAP_SET(opnd, value); + } + break; + case ANYOF_PRINT: + if (LOC) + ANYOF_CLASS_SET(opnd, ANYOF_PRINT); + else { + for (value = 0; value < 256; value++) + if (isPRINT(value)) + ANYOF_BITMAP_SET(opnd, value); + } + break; + case ANYOF_NPRINT: + if (LOC) + ANYOF_CLASS_SET(opnd, ANYOF_NPRINT); + else { + for (value = 0; value < 256; value++) + if (!isPRINT(value)) + ANYOF_BITMAP_SET(opnd, value); + } + break; + case ANYOF_PUNCT: + if (LOC) + ANYOF_CLASS_SET(opnd, ANYOF_PUNCT); + else { + for (value = 0; value < 256; value++) + if (isPUNCT(value)) + ANYOF_BITMAP_SET(opnd, value); + } + break; + case ANYOF_NPUNCT: + if (LOC) + ANYOF_CLASS_SET(opnd, ANYOF_NPUNCT); + else { + for (value = 0; value < 256; value++) + if (!isPUNCT(value)) + ANYOF_BITMAP_SET(opnd, value); + } + break; + case ANYOF_UPPER: + if (LOC) + ANYOF_CLASS_SET(opnd, ANYOF_UPPER); + else { + for (value = 0; value < 256; value++) + if (isUPPER(value)) + ANYOF_BITMAP_SET(opnd, value); + } + break; + case ANYOF_NUPPER: + if (LOC) + ANYOF_CLASS_SET(opnd, ANYOF_NUPPER); + else { + for (value = 0; value < 256; value++) + if (!isUPPER(value)) + ANYOF_BITMAP_SET(opnd, value); + } + break; + case ANYOF_XDIGIT: + if (LOC) + ANYOF_CLASS_SET(opnd, ANYOF_XDIGIT); + else { + for (value = 0; value < 256; value++) + if (isXDIGIT(value)) + ANYOF_BITMAP_SET(opnd, value); + } + break; + case ANYOF_NXDIGIT: + if (LOC) + ANYOF_CLASS_SET(opnd, ANYOF_NXDIGIT); + else { + for (value = 0; value < 256; value++) + if (!isXDIGIT(value)) + ANYOF_BITMAP_SET(opnd, value); + } + break; + default: + FAIL("invalid [::] class in regexp"); break; } + if (LOC) + ANYOF_FLAGS(opnd) |= ANYOF_CLASS; + continue; } if (range) { if (lastvalue > value) - FAIL("invalid [] range in regexp"); + FAIL("invalid [] range in regexp"); /* [b-a] */ range = 0; } else { lastvalue = value; if (*PL_regcomp_parse == '-' && PL_regcomp_parse+1 < PL_regxend && - PL_regcomp_parse[1] != ']') { + PL_regcomp_parse[1] != ']') { + if (namedclass > OOB_NAMEDCLASS) + FAIL("invalid [] range in regexp"); /* [\w-a] */ PL_regcomp_parse++; range = 1; continue; /* do it next time */ } } + /* now is the next time */ if (!SIZE_ONLY) { -#ifndef ASCIIish +#ifndef ASCIIish /* EBCDIC, for example. */ if ((isLOWER(lastvalue) && isLOWER(value)) || (isUPPER(lastvalue) && isUPPER(value))) { @@ -2301,35 +2636,36 @@ S_regclass(pTHX) if (isLOWER(lastvalue)) { for (i = lastvalue; i <= value; i++) if (isLOWER(i)) - ANYOF_SET(opnd, i); + ANYOF_BITMAP_SET(opnd, i); } else { for (i = lastvalue; i <= value; i++) if (isUPPER(i)) - ANYOF_SET(opnd, i); + ANYOF_BITMAP_SET(opnd, i); } } else #endif for ( ; lastvalue <= value; lastvalue++) - ANYOF_SET(opnd, lastvalue); + ANYOF_BITMAP_SET(opnd, lastvalue); } - lastvalue = value; + range = 0; } /* optimize case-insensitive simple patterns (e.g. /[a-z]/i) */ - if (!SIZE_ONLY && (*opnd & (0xFF ^ ANYOF_INVERT)) == ANYOF_FOLD) { + if (!SIZE_ONLY && + (ANYOF_FLAGS(opnd) & (ANYOF_FLAGS_ALL ^ ANYOF_INVERT)) == ANYOF_FOLD) { for (value = 0; value < 256; ++value) { - if (ANYOF_TEST(opnd, value)) { + if (ANYOF_BITMAP_TEST(opnd, value)) { I32 cf = PL_fold[value]; - ANYOF_SET(opnd, cf); + ANYOF_BITMAP_SET(opnd, cf); } } - *opnd &= ~ANYOF_FOLD; + ANYOF_FLAGS(opnd) &= ~ANYOF_FOLD; } /* optimize inverted simple patterns (e.g. [^a-z]) */ - if (!SIZE_ONLY && (*opnd & 0xFF) == ANYOF_INVERT) { - for (value = 0; value < 32; ++value) - opnd[1 + value] ^= 0xFF; - *opnd = 0; + if (!SIZE_ONLY && (ANYOF_FLAGS(opnd) & ANYOF_FLAGS_ALL) == ANYOF_INVERT) { + for (value = 0; value < ANYOF_BITMAP_SIZE; ++value) + opnd[ANYOF_BITMAP_OFFSET + value] ^= ANYOF_FLAGS_ALL; + ANYOF_FLAGS(opnd) = 0; } return ret; } @@ -2337,16 +2673,17 @@ S_regclass(pTHX) STATIC regnode * S_regclassutf8(pTHX) { + dTHR; register char *opnd, *e; register U32 value; - register U32 lastvalue = 123456; + register U32 lastvalue = OOB_UTF8; register I32 range = 0; register regnode *ret; I32 numlen; I32 n; SV *listsv; U8 flags = 0; - dTHR; + I32 namedclass; if (*PL_regcomp_parse == '^') { /* Complement of range. */ PL_regnaughty++; @@ -2362,75 +2699,29 @@ S_regclassutf8(pTHX) listsv = newSVpvn("# comment\n",10); } + checkposixcc(); + if (*PL_regcomp_parse == ']' || *PL_regcomp_parse == '-') goto skipcond; /* allow 1st char to be ] or - */ while (PL_regcomp_parse < PL_regxend && *PL_regcomp_parse != ']') { skipcond: + namedclass = OOB_NAMEDCLASS; value = utf8_to_uv((U8*)PL_regcomp_parse, &numlen); PL_regcomp_parse += numlen; if (value == '[') - (void)regpposixcc(value); /* ignore the return value for now */ + namedclass = regpposixcc(value); else if (value == '\\') { value = utf8_to_uv((U8*)PL_regcomp_parse, &numlen); PL_regcomp_parse += numlen; switch (value) { - case 'w': - if (!SIZE_ONLY) { - if (LOC) - flags |= ANYOF_ALNUML; - - Perl_sv_catpvf(aTHX_ listsv, "+utf8::IsAlnum\n"); - } - lastvalue = 123456; - continue; - case 'W': - if (!SIZE_ONLY) { - if (LOC) - flags |= ANYOF_NALNUML; - - Perl_sv_catpvf(aTHX_ listsv, - "-utf8::IsAlpha\n-utf8::IsDigit\n0000\t%04x\n%04x\tffff\n", - '_' - 1, - '_' + 1); - } - lastvalue = 123456; - continue; - case 's': - if (!SIZE_ONLY) { - if (LOC) - flags |= ANYOF_SPACEL; - Perl_sv_catpvf(aTHX_ listsv, "+utf8::IsSpace\n"); - if (!PL_utf8_space) - is_utf8_space((U8*)" "); - } - lastvalue = 123456; - continue; - case 'S': - if (!SIZE_ONLY) { - if (LOC) - flags |= ANYOF_NSPACEL; - Perl_sv_catpvf(aTHX_ listsv, - "!utf8::IsSpace\n"); - if (!PL_utf8_space) - is_utf8_space((U8*)" "); - } - lastvalue = 123456; - continue; - case 'd': - if (!SIZE_ONLY) { - Perl_sv_catpvf(aTHX_ listsv, "+utf8::IsDigit\n"); - } - lastvalue = 123456; - continue; - case 'D': - if (!SIZE_ONLY) { - Perl_sv_catpvf(aTHX_ listsv, - "!utf8::IsDigit\n"); - } - lastvalue = 123456; - continue; + case 'w': namedclass = ANYOF_ALNUM; break; + case 'W': namedclass = ANYOF_NALNUM; break; + case 's': namedclass = ANYOF_SPACE; break; + case 'S': namedclass = ANYOF_NSPACE; break; + case 'd': namedclass = ANYOF_DIGIT; break; + case 'D': namedclass = ANYOF_NDIGIT; break; case 'p': case 'P': if (*PL_regcomp_parse == '{') { @@ -2445,41 +2736,30 @@ S_regclassutf8(pTHX) } if (!SIZE_ONLY) { if (value == 'p') - Perl_sv_catpvf(aTHX_ listsv, "+utf8::%.*s\n", n, PL_regcomp_parse); + Perl_sv_catpvf(aTHX_ listsv, + "+utf8::%.*s\n", n, PL_regcomp_parse); else Perl_sv_catpvf(aTHX_ listsv, - "!utf8::%.*s\n", n, PL_regcomp_parse); + "!utf8::%.*s\n", n, PL_regcomp_parse); } PL_regcomp_parse = e + 1; - lastvalue = 123456; + lastvalue = OOB_UTF8; continue; - case 'n': - value = '\n'; - break; - case 'r': - value = '\r'; - break; - case 't': - value = '\t'; - break; - case 'f': - value = '\f'; - break; - case 'b': - value = '\b'; - break; - case 'e': - value = '\033'; - break; - case 'a': - value = '\007'; - break; + case 'n': value = '\n'; break; + case 'r': value = '\r'; break; + case 't': value = '\t'; break; + case 'f': value = '\f'; break; + case 'b': value = '\b'; break; + case 'e': value = '\033'; break; + case 'a': value = '\007'; break; case 'x': if (*PL_regcomp_parse == '{') { e = strchr(PL_regcomp_parse++, '}'); if (!e) FAIL("Missing right brace on \\x{}"); - value = scan_hex(PL_regcomp_parse, e - PL_regcomp_parse, &numlen); + value = scan_hex(PL_regcomp_parse, + e - PL_regcomp_parse, + &numlen); PL_regcomp_parse = e + 1; } else { @@ -2498,25 +2778,97 @@ S_regclassutf8(pTHX) break; } } - if (range) { + if (!SIZE_ONLY && namedclass > OOB_NAMEDCLASS) { + if (range) + FAIL("invalid [] range in regexp"); /* [a-\w], [a-[:word:]] */ + switch (namedclass) { + case ANYOF_ALNUM: + Perl_sv_catpvf(aTHX_ listsv, "+utf8::IsWord\n"); break; + case ANYOF_NALNUM: + Perl_sv_catpvf(aTHX_ listsv, "!utf8::IsWord\n"); break; + case ANYOF_ALNUMC: + Perl_sv_catpvf(aTHX_ listsv, "+utf8::IsAlnum\n"); break; + case ANYOF_NALNUMC: + Perl_sv_catpvf(aTHX_ listsv, "!utf8::IsAlnum\n"); break; + case ANYOF_ALPHA: + Perl_sv_catpvf(aTHX_ listsv, "+utf8::IsAlpha\n"); break; + case ANYOF_NALPHA: + Perl_sv_catpvf(aTHX_ listsv, "!utf8::IsAlpha\n"); break; + case ANYOF_ASCII: + Perl_sv_catpvf(aTHX_ listsv, "+utf8::IsASCII\n"); break; + case ANYOF_NASCII: + Perl_sv_catpvf(aTHX_ listsv, "!utf8::IsASCII\n"); break; + case ANYOF_CNTRL: + Perl_sv_catpvf(aTHX_ listsv, "+utf8::IsCntrl\n"); break; + case ANYOF_NCNTRL: + Perl_sv_catpvf(aTHX_ listsv, "!utf8::IsCntrl\n"); break; + case ANYOF_GRAPH: + Perl_sv_catpvf(aTHX_ listsv, "+utf8::IsGraph\n"); break; + case ANYOF_NGRAPH: + Perl_sv_catpvf(aTHX_ listsv, "!utf8::IsGraph\n"); break; + case ANYOF_DIGIT: + Perl_sv_catpvf(aTHX_ listsv, "+utf8::IsDigit\n"); break; + case ANYOF_NDIGIT: + Perl_sv_catpvf(aTHX_ listsv, "!utf8::IsDigit\n"); break; + case ANYOF_LOWER: + Perl_sv_catpvf(aTHX_ listsv, "+utf8::IsLower\n"); break; + case ANYOF_NLOWER: + Perl_sv_catpvf(aTHX_ listsv, "!utf8::IsLower\n"); break; + case ANYOF_PRINT: + Perl_sv_catpvf(aTHX_ listsv, "+utf8::IsPrint\n"); break; + case ANYOF_NPRINT: + Perl_sv_catpvf(aTHX_ listsv, "!utf8::IsPrint\n"); break; + case ANYOF_PUNCT: + Perl_sv_catpvf(aTHX_ listsv, "+utf8::IsPunct\n"); break; + case ANYOF_NPUNCT: + Perl_sv_catpvf(aTHX_ listsv, "!utf8::IsPunct\n"); break; + case ANYOF_SPACE: + Perl_sv_catpvf(aTHX_ listsv, "+utf8::IsSpace\n"); break; + case ANYOF_NSPACE: + Perl_sv_catpvf(aTHX_ listsv, "!utf8::IsSpace\n"); break; + case ANYOF_UPPER: + Perl_sv_catpvf(aTHX_ listsv, "+utf8::IsUpper\n"); break; + case ANYOF_NUPPER: + Perl_sv_catpvf(aTHX_ listsv, "!utf8::IsUpper\n"); break; + case ANYOF_XDIGIT: + Perl_sv_catpvf(aTHX_ listsv, "+utf8::IsXDigit\n"); break; + case ANYOF_NXDIGIT: + Perl_sv_catpvf(aTHX_ listsv, "!utf8::IsXDigit\n"); break; + } + continue; + } + if (range) { if (lastvalue > value) - FAIL("invalid [] range in regexp"); + FAIL("invalid [] range in regexp"); /* [b-a] */ +#ifdef UV_IS_QUAD if (!SIZE_ONLY) - Perl_sv_catpvf(aTHX_ listsv, "%04x\t%04x\n", lastvalue, value); - lastvalue = value; + Perl_sv_catpvf(aTHX_ listsv, "%04" PERL_PRIx64 "\t%04" PERL_PRIx64 "\n", (UV)lastvalue, (UV)value); +#else + if (!SIZE_ONLY) + Perl_sv_catpvf(aTHX_ listsv, "%04x\t%04x\n", lastvalue, value); +#endif range = 0; } else { lastvalue = value; if (*PL_regcomp_parse == '-' && PL_regcomp_parse+1 < PL_regxend && - PL_regcomp_parse[1] != ']') { + PL_regcomp_parse[1] != ']') { + if (namedclass > OOB_NAMEDCLASS) + FAIL("invalid [] range in regexp"); /* [\w-a] */ PL_regcomp_parse++; range = 1; continue; /* do it next time */ } - if (!SIZE_ONLY) - Perl_sv_catpvf(aTHX_ listsv, "%04x\n", value); } + /* now is the next time */ +#ifdef UV_IS_QUAD + if (!SIZE_ONLY) + Perl_sv_catpvf(aTHX_ listsv, "%04" PERL_PRIx64 "\n", (UV)value); +#else + if (!SIZE_ONLY) + Perl_sv_catpvf(aTHX_ listsv, "%04x\n", value); +#endif + range = 0; } ret = reganode(ANYOFUTF8, 0); @@ -2614,7 +2966,7 @@ S_reganode(pTHX_ U8 op, U32 arg) } /* -- regc - emit (if appropriate) a Unicode character +- reguni - emit (if appropriate) a Unicode character */ STATIC void S_reguni(pTHX_ UV uv, char* s, I32* lenp) @@ -2630,17 +2982,6 @@ S_reguni(pTHX_ UV uv, char* s, I32* lenp) } /* -- regc - emit (if appropriate) a byte of code -*/ -STATIC void -S_regc(pTHX_ U8 b, char* s) -{ - dTHR; - if (!SIZE_ONLY) - *s = b; -} - -/* - reginsert - insert an operator in front of already-emitted operand * * Means relocating the operand. @@ -2750,7 +3091,7 @@ STATIC regnode * S_dumpuntil(pTHX_ regnode *start, regnode *node, regnode *last, SV* sv, I32 l) { #ifdef DEBUGGING - register char op = EXACT; /* Arbitrary non-END op. */ + register U8 op = EXACT; /* Arbitrary non-END op. */ register regnode *next, *onode; while (op != END && (!last || node < last)) { @@ -2801,7 +3142,7 @@ S_dumpuntil(pTHX_ regnode *start, regnode *node, regnode *last, SV* sv, I32 l) } else if (PL_regkind[(U8)op] == EXACT) { /* Literal string, where present. */ - node += ((*OPERAND(node)) + 2 + sizeof(regnode) - 1) / sizeof(regnode); + node += NODE_SZ_STR(node) - 1; node = NEXTOPER(node); } else { @@ -2831,16 +3172,18 @@ Perl_regdump(pTHX_ regexp *r) /* Header fields of interest. */ if (r->anchored_substr) - PerlIO_printf(Perl_debug_log, "anchored `%s%s%s'%s at %d ", + PerlIO_printf(Perl_debug_log, "anchored `%s%.*s%s'%s at %d ", PL_colors[0], + SvCUR(r->anchored_substr) - (SvTAIL(r->anchored_substr)!=0), SvPVX(r->anchored_substr), PL_colors[1], SvTAIL(r->anchored_substr) ? "$" : "", r->anchored_offset); if (r->float_substr) - PerlIO_printf(Perl_debug_log, "floating `%s%s%s'%s at %d..%u ", + PerlIO_printf(Perl_debug_log, "floating `%s%.*s%s'%s at %d..%u ", PL_colors[0], - SvPVX(r->float_substr), + SvCUR(r->float_substr) - (SvTAIL(r->float_substr)!=0), + SvPVX(r->float_substr), PL_colors[1], SvTAIL(r->float_substr) ? "$" : "", r->float_min_offset, r->float_max_offset); @@ -2902,12 +3245,15 @@ Perl_regprop(pTHX_ SV *sv, regnode *o) k = PL_regkind[(U8)OP(o)]; if (k == EXACT) - Perl_sv_catpvf(aTHX_ sv, " <%s%s%s>", PL_colors[0], OPERAND(o) + 1, PL_colors[1]); + Perl_sv_catpvf(aTHX_ sv, " <%s%*s%s>", PL_colors[0], + STR_LEN(o), STRING(o), PL_colors[1]); else if (k == CURLY) { if (OP(o) == CURLYM || OP(o) == CURLYN) Perl_sv_catpvf(aTHX_ sv, "[%d]", o->flags); /* Parenth number */ Perl_sv_catpvf(aTHX_ sv, " {%d,%d}", ARG1(o), ARG2(o)); } + else if (k == WHILEM && o->flags) /* Ordinal/of */ + Perl_sv_catpvf(aTHX_ sv, "[%d/%d]", o->flags & 0xf, o->flags>>4); else if (k == REF || k == OPEN || k == CLOSE || k == GROUPP ) Perl_sv_catpvf(aTHX_ sv, "%d", ARG(o)); /* Parenth number */ else if (k == LOGICAL) @@ -2940,6 +3286,7 @@ void Perl_pregfree(pTHX_ struct regexp *r) { dTHR; + DEBUG_r(if (!PL_colorset) reginitcolors()); DEBUG_r(PerlIO_printf(Perl_debug_log, "%sFreeing REx:%s `%s%.60s%s%s'\n", PL_colors[4],PL_colors[5],PL_colors[0], @@ -3109,3 +3456,17 @@ Perl_save_re_context(pTHX) SAVEPPTR(PL_reg_starttry); /* from regexec.c */ #endif } + +#ifdef PERL_OBJECT +#define NO_XSLOCKS +#include "XSUB.h" +#undef this +#define this pPerl +#endif + +static void +clear_re(pTHXo_ void *r) +{ + ReREFCNT_dec((regexp *)r); +} +