* "One Ring to rule them all, One Ring to find them..."
*/
+/* This file contains functions for executing a regular expression. See
+ * also regcomp.c which funnily enough, contains functions for compiling
+ * a regular expression.
+ *
+ * This file is also copied at build time to ext/re/re_exec.c, where
+ * it's built with -DPERL_EXT_RE_BUILD -DPERL_EXT_RE_DEBUG -DPERL_EXT.
+ * This causes the main functions to be compiled under new names and with
+ * debugging support added, which makes "use re 'debug'" work.
+
+ */
+
/* NOTE: this is derived from Henry Spencer's regexp code, and should not
* confused with the original package (see point 3 below). Thanks, Henry!
*/
*
**** Alterations to Henry's code are...
****
- **** Copyright (c) 1991-2002, Larry Wall
+ **** Copyright (C) 1991, 1992, 1993, 1994, 1995, 1996, 1997, 1998, 1999,
+ **** 2000, 2001, 2002, 2003, 2004, 2005, by Larry Wall and others
****
**** You may distribute under the terms of either the GNU General Public
**** License or the Artistic License, as specified in the README file.
#define RF_warned 2 /* warned about big count? */
#define RF_evaled 4 /* Did an EVAL with setting? */
#define RF_utf8 8 /* String contains multibyte chars? */
+#define RF_false 16 /* odd number of nested negatives */
#define UTF ((PL_reg_flags & RF_utf8) != 0)
Perl_croak(aTHX_ "panic: paren_elems_to_push < 0");
#define REGCP_OTHER_ELEMS 6
- SSCHECK(paren_elems_to_push + REGCP_OTHER_ELEMS);
+ SSGROW(paren_elems_to_push + REGCP_OTHER_ELEMS);
for (p = PL_regsize; p > parenfloor; p--) {
/* REGCP_PARENS_ELEMS are pushed per pairs of parentheses. */
SSPUSHINT(PL_regendp[p]);
I32 ml_anch;
register char *other_last = Nullch; /* other substr checked before this */
char *check_at = Nullch; /* check substr found at this pos */
+ I32 multiline = prog->reganch & PMf_MULTILINE;
#ifdef DEBUGGING
char *i_strpos = strpos;
SV *dsv = PERL_DEBUG_PAD_ZERO(0);
#endif
+ RX_MATCH_UTF8_set(prog,do_utf8);
if (prog->reganch & ROPT_UTF8) {
DEBUG_r(PerlIO_printf(Perl_debug_log,
if (prog->reganch & ROPT_ANCH) { /* Match at beg-of-str or after \n */
ml_anch = !( (prog->reganch & ROPT_ANCH_SINGLE)
|| ( (prog->reganch & ROPT_ANCH_BOL)
- && !PL_multiline ) ); /* Check after \n? */
+ && !multiline ) ); /* Check after \n? */
if (!ml_anch) {
if ( !(prog->reganch & (ROPT_ANCH_GPOS /* Checked by the caller */
start_shift + (s - strbeg), end_shift, pp, 0);
else
goto fail_finish;
+ /* we may be pointing at the wrong string */
+ if (s && RX_MATCH_COPIED(prog))
+ s = strbeg + (s - SvPVX(sv));
if (data)
*data->scream_olds = s;
}
else if (prog->reganch & ROPT_CANY_SEEN)
s = fbm_instr((U8*)(s + start_shift),
(U8*)(strend - end_shift),
- check, PL_multiline ? FBMrf_MULTILINE : 0);
+ check, multiline ? FBMrf_MULTILINE : 0);
else
s = fbm_instr(HOP3(s, start_shift, strend),
HOP3(strend, -end_shift, strbeg),
- check, PL_multiline ? FBMrf_MULTILINE : 0);
+ check, multiline ? FBMrf_MULTILINE : 0);
/* Update the count-of-usability, remove useless subpatterns,
unshift s. */
t = s - prog->check_offset_max;
if (s - strpos > prog->check_offset_max /* signed-corrected t > strpos */
- && (!(prog->reganch & ROPT_UTF8)
+ && (!do_utf8
|| ((t = reghopmaybe3_c(s, -(prog->check_offset_max), strpos))
&& t > strpos)))
/* EMPTY */;
HOP3(HOP3(last1, prog->anchored_offset, strend)
+ SvCUR(must), -(SvTAIL(must)!=0), strbeg),
must,
- PL_multiline ? FBMrf_MULTILINE : 0
+ multiline ? FBMrf_MULTILINE : 0
);
DEBUG_r(PerlIO_printf(Perl_debug_log,
"%s anchored substr `%s%.*s%s'%s",
s = fbm_instr((unsigned char*)s,
(unsigned char*)last + SvCUR(must)
- (SvTAIL(must)!=0),
- must, PL_multiline ? FBMrf_MULTILINE : 0);
+ must, multiline ? FBMrf_MULTILINE : 0);
DEBUG_r(PerlIO_printf(Perl_debug_log, "%s floating substr `%s%.*s%s'%s",
(s ? "Found" : "Contradicts"),
PL_colors[0],
t = s - prog->check_offset_max;
if (s - strpos > prog->check_offset_max /* signed-corrected t > strpos */
- && (!(prog->reganch & ROPT_UTF8)
+ && (!do_utf8
|| ((t = reghopmaybe3_c(s, -prog->check_offset_max, strpos))
&& t > strpos))) {
/* Fixed substring is found far enough so that the match
char *startpos = strbeg;
t = s;
- if (prog->reganch & ROPT_UTF8) {
- PL_regdata = prog->data;
- PL_bostr = startpos;
- }
+ cache_re(prog);
s = find_byclass(prog, prog->regstclass, s, endpos, startpos, 1);
if (!s) {
#ifdef DEBUGGING
I32 doevery = (prog->reganch & ROPT_SKIP) == 0;
char *m;
STRLEN ln;
+ STRLEN lnc;
+ register STRLEN uskip;
unsigned int c1;
unsigned int c2;
char *e;
/* We know what class it must start with. */
switch (OP(c)) {
case ANYOF:
- while (s < strend) {
- STRLEN skip = do_utf8 ? UTF8SKIP(s) : 1;
-
- if (do_utf8 ?
- reginclass(c, (U8*)s, 0, do_utf8) :
- REGINCLASS(c, (U8*)s) ||
- (ANYOF_FOLD_SHARP_S(c, s, strend) &&
- /* The assignment of 2 is intentional:
- * for the sharp s, the skip is 2. */
- (skip = SHARP_S_SKIP)
- )) {
- if (tmp && (norun || regtry(prog, s)))
- goto got_it;
- else
- tmp = doevery;
- }
- else
- tmp = 1;
- s += skip;
+ if (do_utf8) {
+ while (s + (uskip = UTF8SKIP(s)) <= strend) {
+ if ((ANYOF_FLAGS(c) & ANYOF_UNICODE) ||
+ !UTF8_IS_INVARIANT((U8)s[0]) ?
+ reginclass(c, (U8*)s, 0, do_utf8) :
+ REGINCLASS(c, (U8*)s)) {
+ if (tmp && (norun || regtry(prog, s)))
+ goto got_it;
+ else
+ tmp = doevery;
+ }
+ else
+ tmp = 1;
+ s += uskip;
+ }
+ }
+ else {
+ while (s < strend) {
+ STRLEN skip = 1;
+
+ if (REGINCLASS(c, (U8*)s) ||
+ (ANYOF_FOLD_SHARP_S(c, s, strend) &&
+ /* The assignment of 2 is intentional:
+ * for the folded sharp s, the skip is 2. */
+ (skip = SHARP_S_SKIP))) {
+ if (tmp && (norun || regtry(prog, s)))
+ goto got_it;
+ else
+ tmp = doevery;
+ }
+ else
+ tmp = 1;
+ s += skip;
+ }
}
break;
case CANY:
}
break;
case EXACTF:
- m = STRING(c);
- ln = STR_LEN(c);
+ m = STRING(c);
+ ln = STR_LEN(c); /* length to match in octets/bytes */
+ lnc = (I32) ln; /* length to match in characters */
if (UTF) {
STRLEN ulen1, ulen2;
- U8 tmpbuf1[UTF8_MAXLEN_UCLC+1];
- U8 tmpbuf2[UTF8_MAXLEN_UCLC+1];
+ U8 *sm = (U8 *) m;
+ U8 tmpbuf1[UTF8_MAXBYTES_CASE+1];
+ U8 tmpbuf2[UTF8_MAXBYTES_CASE+1];
to_utf8_lower((U8*)m, tmpbuf1, &ulen1);
to_utf8_upper((U8*)m, tmpbuf2, &ulen2);
- c1 = utf8n_to_uvchr(tmpbuf1, UTF8_MAXLEN_UCLC,
+ c1 = utf8n_to_uvchr(tmpbuf1, UTF8_MAXBYTES_CASE,
0, ckWARN(WARN_UTF8) ? 0 : UTF8_ALLOW_ANY);
- c2 = utf8n_to_uvchr(tmpbuf2, UTF8_MAXLEN_UCLC,
+ c2 = utf8n_to_uvchr(tmpbuf2, UTF8_MAXBYTES_CASE,
0, ckWARN(WARN_UTF8) ? 0 : UTF8_ALLOW_ANY);
+ lnc = 0;
+ while (sm < ((U8 *) m + ln)) {
+ lnc++;
+ sm += UTF8SKIP(sm);
+ }
}
else {
c1 = *(U8*)m;
}
goto do_exactf;
case EXACTFL:
- m = STRING(c);
- ln = STR_LEN(c);
+ m = STRING(c);
+ ln = STR_LEN(c);
+ lnc = (I32) ln;
c1 = *(U8*)m;
c2 = PL_fold_locale[c1];
do_exactf:
- e = HOP3c(strend, -(I32)ln, s);
+ e = HOP3c(strend, -((I32)lnc), s);
if (norun && e < s)
e = s; /* Due to minlen logic of intuit() */
if (do_utf8) {
UV c, f;
- U8 tmpbuf [UTF8_MAXLEN+1];
- U8 foldbuf[UTF8_MAXLEN_FOLD+1];
+ U8 tmpbuf [UTF8_MAXBYTES+1];
+ U8 foldbuf[UTF8_MAXBYTES_CASE+1];
STRLEN len, foldlen;
if (c1 == c2) {
+ /* Upper and lower of 1st char are equal -
+ * probably not a "letter". */
while (s <= e) {
- c = utf8n_to_uvchr((U8*)s, UTF8_MAXLEN, &len,
+ c = utf8n_to_uvchr((U8*)s, UTF8_MAXBYTES, &len,
ckWARN(WARN_UTF8) ?
0 : UTF8_ALLOW_ANY);
if ( c == c1
}
else {
while (s <= e) {
- c = utf8n_to_uvchr((U8*)s, UTF8_MAXLEN, &len,
+ c = utf8n_to_uvchr((U8*)s, UTF8_MAXBYTES, &len,
ckWARN(WARN_UTF8) ?
0 : UTF8_ALLOW_ANY);
tmp = ((OP(c) == BOUND ?
isALNUM_uni(tmp) : isALNUM_LC_uvchr(UNI_TO_NATIVE(tmp))) != 0);
LOAD_UTF8_CHARCLASS(alnum,"a");
- while (s < strend) {
+ while (s + (uskip = UTF8SKIP(s)) <= strend) {
if (tmp == !(OP(c) == BOUND ?
swash_fetch(PL_utf8_alnum, (U8*)s, do_utf8) :
isALNUM_LC_utf8((U8*)s)))
if ((norun || regtry(prog, s)))
goto got_it;
}
- s += UTF8SKIP(s);
+ s += uskip;
}
}
else {
tmp = ((OP(c) == NBOUND ?
isALNUM_uni(tmp) : isALNUM_LC_uvchr(UNI_TO_NATIVE(tmp))) != 0);
LOAD_UTF8_CHARCLASS(alnum,"a");
- while (s < strend) {
+ while (s + (uskip = UTF8SKIP(s)) <= strend) {
if (tmp == !(OP(c) == NBOUND ?
swash_fetch(PL_utf8_alnum, (U8*)s, do_utf8) :
isALNUM_LC_utf8((U8*)s)))
tmp = !tmp;
else if ((norun || regtry(prog, s)))
goto got_it;
- s += UTF8SKIP(s);
+ s += uskip;
}
}
else {
case ALNUM:
if (do_utf8) {
LOAD_UTF8_CHARCLASS(alnum,"a");
- while (s < strend) {
+ while (s + (uskip = UTF8SKIP(s)) <= strend) {
if (swash_fetch(PL_utf8_alnum, (U8*)s, do_utf8)) {
if (tmp && (norun || regtry(prog, s)))
goto got_it;
}
else
tmp = 1;
- s += UTF8SKIP(s);
+ s += uskip;
}
}
else {
case ALNUML:
PL_reg_flags |= RF_tainted;
if (do_utf8) {
- while (s < strend) {
+ while (s + (uskip = UTF8SKIP(s)) <= strend) {
if (isALNUM_LC_utf8((U8*)s)) {
if (tmp && (norun || regtry(prog, s)))
goto got_it;
}
else
tmp = 1;
- s += UTF8SKIP(s);
+ s += uskip;
}
}
else {
case NALNUM:
if (do_utf8) {
LOAD_UTF8_CHARCLASS(alnum,"a");
- while (s < strend) {
+ while (s + (uskip = UTF8SKIP(s)) <= strend) {
if (!swash_fetch(PL_utf8_alnum, (U8*)s, do_utf8)) {
if (tmp && (norun || regtry(prog, s)))
goto got_it;
}
else
tmp = 1;
- s += UTF8SKIP(s);
+ s += uskip;
}
}
else {
case NALNUML:
PL_reg_flags |= RF_tainted;
if (do_utf8) {
- while (s < strend) {
+ while (s + (uskip = UTF8SKIP(s)) <= strend) {
if (!isALNUM_LC_utf8((U8*)s)) {
if (tmp && (norun || regtry(prog, s)))
goto got_it;
}
else
tmp = 1;
- s += UTF8SKIP(s);
+ s += uskip;
}
}
else {
case SPACE:
if (do_utf8) {
LOAD_UTF8_CHARCLASS(space," ");
- while (s < strend) {
+ while (s + (uskip = UTF8SKIP(s)) <= strend) {
if (*s == ' ' || swash_fetch(PL_utf8_space,(U8*)s, do_utf8)) {
if (tmp && (norun || regtry(prog, s)))
goto got_it;
}
else
tmp = 1;
- s += UTF8SKIP(s);
+ s += uskip;
}
}
else {
case SPACEL:
PL_reg_flags |= RF_tainted;
if (do_utf8) {
- while (s < strend) {
+ while (s + (uskip = UTF8SKIP(s)) <= strend) {
if (*s == ' ' || isSPACE_LC_utf8((U8*)s)) {
if (tmp && (norun || regtry(prog, s)))
goto got_it;
}
else
tmp = 1;
- s += UTF8SKIP(s);
+ s += uskip;
}
}
else {
case NSPACE:
if (do_utf8) {
LOAD_UTF8_CHARCLASS(space," ");
- while (s < strend) {
+ while (s + (uskip = UTF8SKIP(s)) <= strend) {
if (!(*s == ' ' || swash_fetch(PL_utf8_space,(U8*)s, do_utf8))) {
if (tmp && (norun || regtry(prog, s)))
goto got_it;
}
else
tmp = 1;
- s += UTF8SKIP(s);
+ s += uskip;
}
}
else {
case NSPACEL:
PL_reg_flags |= RF_tainted;
if (do_utf8) {
- while (s < strend) {
+ while (s + (uskip = UTF8SKIP(s)) <= strend) {
if (!(*s == ' ' || isSPACE_LC_utf8((U8*)s))) {
if (tmp && (norun || regtry(prog, s)))
goto got_it;
}
else
tmp = 1;
- s += UTF8SKIP(s);
+ s += uskip;
}
}
else {
case DIGIT:
if (do_utf8) {
LOAD_UTF8_CHARCLASS(digit,"0");
- while (s < strend) {
+ while (s + (uskip = UTF8SKIP(s)) <= strend) {
if (swash_fetch(PL_utf8_digit,(U8*)s, do_utf8)) {
if (tmp && (norun || regtry(prog, s)))
goto got_it;
}
else
tmp = 1;
- s += UTF8SKIP(s);
+ s += uskip;
}
}
else {
case DIGITL:
PL_reg_flags |= RF_tainted;
if (do_utf8) {
- while (s < strend) {
+ while (s + (uskip = UTF8SKIP(s)) <= strend) {
if (isDIGIT_LC_utf8((U8*)s)) {
if (tmp && (norun || regtry(prog, s)))
goto got_it;
}
else
tmp = 1;
- s += UTF8SKIP(s);
+ s += uskip;
}
}
else {
case NDIGIT:
if (do_utf8) {
LOAD_UTF8_CHARCLASS(digit,"0");
- while (s < strend) {
+ while (s + (uskip = UTF8SKIP(s)) <= strend) {
if (!swash_fetch(PL_utf8_digit,(U8*)s, do_utf8)) {
if (tmp && (norun || regtry(prog, s)))
goto got_it;
}
else
tmp = 1;
- s += UTF8SKIP(s);
+ s += uskip;
}
}
else {
case NDIGITL:
PL_reg_flags |= RF_tainted;
if (do_utf8) {
- while (s < strend) {
+ while (s + (uskip = UTF8SKIP(s)) <= strend) {
if (!isDIGIT_LC_utf8((U8*)s)) {
if (tmp && (norun || regtry(prog, s)))
goto got_it;
}
else
tmp = 1;
- s += UTF8SKIP(s);
+ s += uskip;
}
}
else {
char *scream_olds;
SV* oreplsv = GvSV(PL_replgv);
bool do_utf8 = DO_UTF8(sv);
+ I32 multiline = prog->reganch & PMf_MULTILINE;
#ifdef DEBUGGING
SV *dsv0 = PERL_DEBUG_PAD_ZERO(0);
SV *dsv1 = PERL_DEBUG_PAD_ZERO(1);
#endif
+ RX_MATCH_UTF8_set(prog,do_utf8);
PL_regcc = 0;
if (prog->reganch & (ROPT_ANCH & ~ROPT_ANCH_GPOS)) {
if (s == startpos && regtry(prog, startpos))
goto got_it;
- else if (PL_multiline || (prog->reganch & ROPT_IMPLICIT)
+ else if (multiline || (prog->reganch & ROPT_IMPLICIT)
|| (prog->reganch & ROPT_ANCH_MBOL)) /* XXXX SBOL? */
{
char *end;
end_shift, &scream_pos, 0))
: (s = fbm_instr((unsigned char*)HOP3(s, back_min, strend),
(unsigned char*)strend, must,
- PL_multiline ? FBMrf_MULTILINE : 0))) ) {
+ multiline ? FBMrf_MULTILINE : 0))) ) {
+ /* we may be pointing at the wrong string */
+ if ((flags & REXEC_SCREAM) && RX_MATCH_COPIED(prog))
+ s = strbeg + (s - SvPVX(sv));
DEBUG_r( did_match = 1 );
if (HOPc(s, -back_max) > last1) {
last1 = HOPc(s, -back_min);
goto phooey;
}
else if ((c = prog->regstclass)) {
- if (minlen && PL_regkind[(U8)OP(prog->regstclass)] != EXACT)
+ if (minlen) {
+ I32 op = (U8)OP(prog->regstclass);
/* don't bother with what can't match */
- strend = HOPc(strend, -(minlen - 1));
+ if (PL_regkind[op] != EXACT && op != CANY)
+ strend = HOPc(strend, -(minlen - 1));
+ }
DEBUG_r({
SV *prop = sv_newmortal();
char *s0;
end_shift, &scream_pos, 1); /* last one */
if (!last)
last = scream_olds; /* Only one occurrence. */
+ /* we may be pointing at the wrong string */
+ else if (RX_MATCH_COPIED(prog))
+ s = strbeg + (s - SvPVX(sv));
}
else {
STRLEN len;
if (SvTAIL(float_real)) {
if (memEQ(strend - len + 1, little, len - 1))
last = strend - len + 1;
- else if (!PL_multiline)
+ else if (!multiline)
last = memEQ(strend - len, little, len)
? strend - len : Nullch;
else
/* make sure $`, $&, $', and $digit will work later */
if ( !(flags & REXEC_NOT_FIRST) ) {
- if (RX_MATCH_COPIED(prog)) {
- Safefree(prog->subbeg);
- RX_MATCH_COPIED_off(prog);
- }
+ RX_MATCH_COPY_FREE(prog);
if (flags & REXEC_COPY_STR) {
I32 i = PL_regeol - startpos + (stringarg - strbeg);
-
- s = savepvn(strbeg, i);
- prog->subbeg = s;
+#ifdef PERL_COPY_ON_WRITE
+ if ((SvIsCOW(sv)
+ || (SvFLAGS(sv) & CAN_COW_MASK) == CAN_COW_FLAGS)) {
+ if (DEBUG_C_TEST) {
+ PerlIO_printf(Perl_debug_log,
+ "Copy on write: regexp capture, type %d\n",
+ (int) SvTYPE(sv));
+ }
+ prog->saved_copy = sv_setsv_cow(prog->saved_copy, sv);
+ prog->subbeg = SvPVX(prog->saved_copy);
+ assert (SvPOKp(prog->saved_copy));
+ } else
+#endif
+ {
+ RX_MATCH_COPIED_on(prog);
+ s = savepvn(strbeg, i);
+ prog->subbeg = s;
+ }
prog->sublen = i;
- RX_MATCH_COPIED_on(prog);
}
else {
prog->subbeg = strbeg;
if (PL_reg_sv) {
/* Make $_ available to executed code. */
if (PL_reg_sv != DEFSV) {
- /* SAVE_DEFSV does *not* suffice here for USE_5005THREADS */
- SAVESPTR(DEFSV);
+ SAVE_DEFSV;
DEFSV = PL_reg_sv;
}
$` inside (?{}) could fail... */
PL_reg_oldsaved = prog->subbeg;
PL_reg_oldsavedlen = prog->sublen;
+#ifdef PERL_COPY_ON_WRITE
+ PL_nrs = prog->saved_copy;
+#endif
RX_MATCH_COPIED_off(prog);
}
else
PL_reglastparen = &prog->lastparen;
PL_reglastcloseparen = &prog->lastcloseparen;
prog->lastparen = 0;
+ prog->lastcloseparen = 0;
PL_regsize = 0;
DEBUG_r(PL_reg_starttry = startpos);
if (PL_reg_start_tmpl <= prog->nparens) {
regprop(prop, scan);
{
char *s0 =
- do_utf8 ?
+ do_utf8 && OP(scan) != CANY ?
pv_uni_display(dsv0, (U8*)(locinput - pref_len),
pref0_len, 60, UNI_DISPLAY_REGEX) :
locinput - pref_len;
int len0 = do_utf8 ? strlen(s0) : pref0_len;
- char *s1 = do_utf8 ?
+ char *s1 = do_utf8 && OP(scan) != CANY ?
pv_uni_display(dsv1, (U8*)(locinput - pref_len + pref0_len),
pref_len - pref0_len, 60, UNI_DISPLAY_REGEX) :
locinput - pref_len + pref0_len;
int len1 = do_utf8 ? strlen(s1) : pref_len - pref0_len;
- char *s2 = do_utf8 ?
+ char *s2 = do_utf8 && OP(scan) != CANY ?
pv_uni_display(dsv2, (U8*)locinput,
PL_regeol - locinput, 60, UNI_DISPLAY_REGEX) :
locinput;
switch (OP(scan)) {
case BOL:
- if (locinput == PL_bostr || (PL_multiline &&
- (nextchr || locinput < PL_regeol) && locinput[-1] == '\n') )
+ if (locinput == PL_bostr)
{
/* regtill = regbol; */
break;
break;
sayNO;
case EOL:
- if (PL_multiline)
- goto meol;
- else
goto seol;
case MEOL:
- meol:
if ((nextchr || locinput < PL_regeol) && nextchr != '\n')
sayNO;
break;
if (l >= PL_regeol)
sayNO;
if (NATIVE_TO_UNI(*(U8*)s) !=
- utf8n_to_uvuni((U8*)l, UTF8_MAXLEN, &ulen,
+ utf8n_to_uvuni((U8*)l, UTF8_MAXBYTES, &ulen,
ckWARN(WARN_UTF8) ?
0 : UTF8_ALLOW_ANY))
sayNO;
if (l >= PL_regeol)
sayNO;
if (NATIVE_TO_UNI(*((U8*)l)) !=
- utf8n_to_uvuni((U8*)s, UTF8_MAXLEN, &ulen,
+ utf8n_to_uvuni((U8*)s, UTF8_MAXBYTES, &ulen,
ckWARN(WARN_UTF8) ?
0 : UTF8_ALLOW_ANY))
sayNO;
*/
if (OP(scan) == REFF) {
STRLEN ulen1, ulen2;
- U8 tmpbuf1[UTF8_MAXLEN_UCLC+1];
- U8 tmpbuf2[UTF8_MAXLEN_UCLC+1];
+ U8 tmpbuf1[UTF8_MAXBYTES_CASE+1];
+ U8 tmpbuf2[UTF8_MAXBYTES_CASE+1];
while (s < e) {
if (l >= PL_regeol)
sayNO;
dSP;
OP_4tree *oop = PL_op;
COP *ocurcop = PL_curcop;
- SV **ocurpad = PL_curpad;
+ PAD *old_comppad;
SV *ret;
+ struct regexp *oreg = PL_reg_re;
n = ARG(scan);
PL_op = (OP_4tree*)PL_regdata->data[n];
DEBUG_r( PerlIO_printf(Perl_debug_log, " re_eval 0x%"UVxf"\n", PTR2UV(PL_op)) );
- PL_curpad = AvARRAY((AV*)PL_regdata->data[n + 2]);
+ PAD_SAVE_LOCAL(old_comppad, (PAD*)PL_regdata->data[n + 2]);
PL_regendp[0] = PL_reg_magic->mg_len = locinput - PL_bostr;
{
CALLRUNOPS(aTHX); /* Scalar context. */
SPAGAIN;
if (SP == before)
- ret = Nullsv; /* protect against empty (?{}) blocks. */
+ ret = &PL_sv_undef; /* protect against empty (?{}) blocks. */
else {
ret = POPs;
PUTBACK;
}
PL_op = oop;
- PL_curpad = ocurpad;
+ PAD_RESTORE_LOCAL(old_comppad);
PL_curcop = ocurcop;
if (logical) {
if (logical == 2) { /* Postponed subexpression. */
MAGIC *mg = Null(MAGIC*);
re_cc_state state;
CHECKPOINT cp, lastcp;
-
- if(SvROK(ret) || SvRMAGICAL(ret)) {
- SV *sv = SvROK(ret) ? SvRV(ret) : ret;
-
- if(SvMAGICAL(sv))
- mg = mg_find(sv, PERL_MAGIC_qr);
+ int toggleutf;
+ register SV *sv;
+
+ if(SvROK(ret) && SvSMAGICAL(sv = SvRV(ret)))
+ mg = mg_find(sv, PERL_MAGIC_qr);
+ else if (SvSMAGICAL(ret)) {
+ if (SvGMAGICAL(ret))
+ sv_unmagic(ret, PERL_MAGIC_qr);
+ else
+ mg = mg_find(ret, PERL_MAGIC_qr);
}
+
if (mg) {
re = (regexp *)mg->mg_obj;
(void)ReREFCNT_inc(re);
I32 onpar = PL_regnpar;
Zero(&pm, 1, PMOP);
+ if (DO_UTF8(ret)) pm.op_pmdynflags |= PMdf_DYN_UTF8;
re = CALLREGCOMP(aTHX_ t, t + len, &pm);
if (!(SvFLAGS(ret)
- & (SVs_TEMP | SVs_PADTMP | SVf_READONLY)))
+ & (SVs_TEMP | SVs_PADTMP | SVf_READONLY
+ | SVs_GMG)))
sv_magic(ret,(SV*)ReREFCNT_inc(re),
PERL_MAGIC_qr,0,0);
PL_regprecomp = oprecomp;
*PL_reglastcloseparen = 0;
PL_reg_call_cc = &state;
PL_reginput = locinput;
+ toggleutf = ((PL_reg_flags & RF_utf8) != 0) ^
+ ((re->reganch & ROPT_UTF8) != 0);
+ if (toggleutf) PL_reg_flags ^= RF_utf8;
/* XXXX This is too dramatic a measure... */
PL_reg_maxiter = 0;
PL_regcc = state.cc;
PL_reg_re = state.re;
cache_re(PL_reg_re);
+ if (toggleutf) PL_reg_flags ^= RF_utf8;
/* XXXX This is too dramatic a measure... */
PL_reg_maxiter = 0;
PL_regcc = state.cc;
PL_reg_re = state.re;
cache_re(PL_reg_re);
+ if (toggleutf) PL_reg_flags ^= RF_utf8;
/* XXXX This is too dramatic a measure... */
PL_reg_maxiter = 0;
sw = SvTRUE(ret);
logical = 0;
}
- else
+ else {
sv_setsv(save_scalar(PL_replgv), ret);
+ cache_re(oreg);
+ }
break;
}
case OPEN:
"%*s already tried at this position...\n",
REPORT_CODE_OFF+PL_regindent*2, "")
);
- sayNO_SILENT;
+ if (PL_reg_flags & RF_false)
+ sayYES;
+ else
+ sayNO_SILENT;
}
PL_reg_poscache[o] |= (1<<b);
}
CHECKPOINT lastcp;
/* We suppose that the next guy does not need
- backtracking: in particular, it is of constant length,
+ backtracking: in particular, it is of constant non-zero length,
and has no parenths to influence future backrefs. */
ln = ARG1(scan); /* min to match */
n = ARG2(scan); /* max to match */
minmod = 0;
if (ln && regrepeat_hard(scan, ln, &l) < ln)
sayNO;
- /* 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 (HAS_TEXT(next) || JUMPABLE(next)) {
regnode *text_node = next;
if (! HAS_TEXT(text_node)) c1 = c2 = -1000;
else {
if (PL_regkind[(U8)OP(text_node)] == REF) {
- I32 n, ln;
- n = ARG(text_node); /* which paren pair */
- ln = PL_regstartp[n];
- /* assume yes if we haven't seen CLOSEn */
- if (
- (I32)*PL_reglastparen < n ||
- ln == -1 ||
- ln == PL_regendp[n]
- ) {
- c1 = c2 = -1000;
- goto assume_ok_MM;
- }
- c1 = *(PL_bostr + ln);
+ c1 = c2 = -1000;
+ goto assume_ok_MM;
}
else { c1 = (U8)*STRING(text_node); }
if (OP(text_node) == EXACTF || OP(text_node) == REFF)
c1 = c2 = -1000;
assume_ok_MM:
REGCP_SET(lastcp);
- /* This may be improved if l == 0. */
- while (n >= ln || (n == REG_INFTY && ln > 0 && l)) { /* ln overflow ? */
+ while (n >= ln || (n == REG_INFTY && ln > 0)) { /* ln overflow ? */
/* If it could work, try it. */
if (c1 == -1000 ||
UCHARAT(PL_reginput) == c1 ||
}
else {
n = regrepeat_hard(scan, n, &l);
- /* 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,
if (! HAS_TEXT(text_node)) c1 = c2 = -1000;
else {
if (PL_regkind[(U8)OP(text_node)] == REF) {
- I32 n, ln;
- n = ARG(text_node); /* which paren pair */
- ln = PL_regstartp[n];
- /* assume yes if we haven't seen CLOSEn */
- if (
- (I32)*PL_reglastparen < n ||
- ln == -1 ||
- ln == PL_regendp[n]
- ) {
- c1 = c2 = -1000;
- goto assume_ok_REG;
- }
- c1 = *(PL_bostr + ln);
+ c1 = c2 = -1000;
+ goto assume_ok_REG;
}
else { c1 = (U8)*STRING(text_node); }
if (! HAS_TEXT(text_node)) c1 = c2 = -1000;
else {
if (PL_regkind[(U8)OP(text_node)] == REF) {
- I32 n, ln;
- n = ARG(text_node); /* which paren pair */
- ln = PL_regstartp[n];
- /* assume yes if we haven't seen CLOSEn */
- if (
- (I32)*PL_reglastparen < n ||
- ln == -1 ||
- ln == PL_regendp[n]
- ) {
- c1 = c2 = -1000;
- goto assume_ok_easy;
- }
- s = (U8*)PL_bostr + ln;
+ c1 = c2 = -1000;
+ goto assume_ok_easy;
}
else { s = (U8*)STRING(text_node); }
else { /* UTF */
if (OP(text_node) == EXACTF || OP(text_node) == REFF) {
STRLEN ulen1, ulen2;
- U8 tmpbuf1[UTF8_MAXLEN_UCLC+1];
- U8 tmpbuf2[UTF8_MAXLEN_UCLC+1];
+ U8 tmpbuf1[UTF8_MAXBYTES_CASE+1];
+ U8 tmpbuf2[UTF8_MAXBYTES_CASE+1];
to_utf8_lower((U8*)s, tmpbuf1, &ulen1);
to_utf8_upper((U8*)s, tmpbuf2, &ulen2);
- c1 = utf8n_to_uvuni(tmpbuf1, UTF8_MAXLEN, 0,
+ c1 = utf8n_to_uvuni(tmpbuf1, UTF8_MAXBYTES, 0,
ckWARN(WARN_UTF8) ?
0 : UTF8_ALLOW_ANY);
- c2 = utf8n_to_uvuni(tmpbuf2, UTF8_MAXLEN, 0,
+ c2 = utf8n_to_uvuni(tmpbuf2, UTF8_MAXBYTES, 0,
ckWARN(WARN_UTF8) ?
0 : UTF8_ALLOW_ANY);
}
else {
- c2 = c1 = utf8n_to_uvchr(s, UTF8_MAXLEN, 0,
+ c2 = c1 = utf8n_to_uvchr(s, UTF8_MAXBYTES, 0,
ckWARN(WARN_UTF8) ?
0 : UTF8_ALLOW_ANY);
}
* utf8_distance(old, locinput) */
while (locinput <= e &&
utf8n_to_uvchr((U8*)locinput,
- UTF8_MAXLEN, &len,
+ UTF8_MAXBYTES, &len,
ckWARN(WARN_UTF8) ?
0 : UTF8_ALLOW_ANY) != (UV)c1) {
locinput += len;
* utf8_distance(old, locinput) */
while (locinput <= e) {
UV c = utf8n_to_uvchr((U8*)locinput,
- UTF8_MAXLEN, &len,
+ UTF8_MAXBYTES, &len,
ckWARN(WARN_UTF8) ?
0 : UTF8_ALLOW_ANY);
if (c == (UV)c1 || c == (UV)c2)
if (c1 != -1000) {
if (do_utf8)
c = utf8n_to_uvchr((U8*)PL_reginput,
- UTF8_MAXLEN, 0,
+ UTF8_MAXBYTES, 0,
ckWARN(WARN_UTF8) ?
0 : UTF8_ALLOW_ANY);
else
/* If it could work, try it. */
if (c == (UV)c1 || c == (UV)c2)
{
- TRYPAREN(paren, n, PL_reginput);
+ TRYPAREN(paren, ln, PL_reginput);
REGCP_UNWIND(lastcp);
}
}
/* If it could work, try it. */
else if (c1 == -1000)
{
- TRYPAREN(paren, n, PL_reginput);
+ TRYPAREN(paren, ln, PL_reginput);
REGCP_UNWIND(lastcp);
}
/* Couldn't or didn't -- move forward. */
n = regrepeat(scan, n);
locinput = PL_reginput;
if (ln < n && PL_regkind[(U8)OP(next)] == EOL &&
- ((!PL_multiline && OP(next) != MEOL) ||
+ (OP(next) != MEOL ||
OP(next) == SEOL || OP(next) == EOS))
{
ln = n; /* why back off? */
if (c1 != -1000) {
if (do_utf8)
c = utf8n_to_uvchr((U8*)PL_reginput,
- UTF8_MAXLEN, 0,
+ UTF8_MAXBYTES, 0,
ckWARN(WARN_UTF8) ?
0 : UTF8_ALLOW_ANY);
else
if (c1 != -1000) {
if (do_utf8)
c = utf8n_to_uvchr((U8*)PL_reginput,
- UTF8_MAXLEN, 0,
+ UTF8_MAXBYTES, 0,
ckWARN(WARN_UTF8) ?
0 : UTF8_ALLOW_ANY);
else
}
else
PL_reginput = locinput;
+ PL_reg_flags ^= RF_false;
goto do_ifmatch;
case IFMATCH:
n = 1;
do_ifmatch:
inner = NEXTOPER(NEXTOPER(scan));
if (regmatch(inner) != n) {
+ if (n == 0)
+ PL_reg_flags ^= RF_false;
say_no:
if (logical) {
logical = 0;
else
sayNO;
}
+ if (n == 0)
+ PL_reg_flags ^= RF_false;
say_yes:
if (logical) {
logical = 0;
register bool do_utf8 = PL_reg_match_utf8;
scan = PL_reginput;
- if (max != REG_INFTY && max < loceol - scan)
+ if (max == REG_INFTY)
+ max = I32_MAX;
+ else if (max < loceol - scan)
loceol = scan + max;
switch (OP(p)) {
case REG_ANY:
/*
- regrepeat_hard - repeatedly match something, report total lenth and length
*
- * The repeater is supposed to have constant length.
+ * The repeater is supposed to have constant non-zero length.
*/
STATIC I32
{
char flags = ANYOF_FLAGS(n);
bool match = FALSE;
- UV c;
+ UV c = *p;
STRLEN len = 0;
STRLEN plen;
- c = do_utf8 ? utf8n_to_uvchr(p, UTF8_MAXLEN, &len,
- ckWARN(WARN_UTF8) ? 0 : UTF8_ALLOW_ANY) : *p;
+ if (do_utf8 && !UTF8_IS_INVARIANT(c))
+ c = utf8n_to_uvchr(p, UTF8_MAXBYTES, &len,
+ ckWARN(WARN_UTF8) ? 0 : UTF8_ALLOW_ANY);
plen = lenp ? *lenp : UNISKIP(NATIVE_TO_UNI(c));
if (do_utf8 || (flags & ANYOF_UNICODE)) {
}
}
if (!match) {
- U8 tmpbuf[UTF8_MAXLEN_FOLD+1];
+ U8 tmpbuf[UTF8_MAXBYTES_CASE+1];
STRLEN tmplen;
to_utf8_fold(p, tmpbuf, &tmplen);
if (PL_reg_oldsaved) {
PL_reg_re->subbeg = PL_reg_oldsaved;
PL_reg_re->sublen = PL_reg_oldsavedlen;
+#ifdef PERL_COPY_ON_WRITE
+ PL_reg_re->saved_copy = PL_nrs;
+#endif
RX_MATCH_COPIED_on(PL_reg_re);
}
PL_reg_magic->mg_len = PL_reg_oldpos;
{
SV* sv;
if (prog->float_substr && !prog->float_utf8) {
- prog->float_utf8 = sv = NEWSV(117, 0);
- SvSetSV(sv, prog->float_substr);
+ prog->float_utf8 = sv = newSVsv(prog->float_substr);
sv_utf8_upgrade(sv);
if (SvTAIL(prog->float_substr))
SvTAIL_on(sv);
prog->check_utf8 = sv;
}
if (prog->anchored_substr && !prog->anchored_utf8) {
- prog->anchored_utf8 = sv = NEWSV(118, 0);
- SvSetSV(sv, prog->anchored_substr);
+ prog->anchored_utf8 = sv = newSVsv(prog->anchored_substr);
sv_utf8_upgrade(sv);
if (SvTAIL(prog->anchored_substr))
SvTAIL_on(sv);
{
SV* sv;
if (prog->float_utf8 && !prog->float_substr) {
- prog->float_substr = sv = NEWSV(117, 0);
- SvSetSV(sv, prog->float_utf8);
+ prog->float_substr = sv = newSVsv(prog->float_utf8);
if (sv_utf8_downgrade(sv, TRUE)) {
if (SvTAIL(prog->float_utf8))
SvTAIL_on(sv);
prog->check_substr = sv;
}
if (prog->anchored_utf8 && !prog->anchored_substr) {
- prog->anchored_substr = sv = NEWSV(118, 0);
- SvSetSV(sv, prog->anchored_utf8);
+ prog->anchored_substr = sv = newSVsv(prog->anchored_utf8);
if (sv_utf8_downgrade(sv, TRUE)) {
if (SvTAIL(prog->anchored_utf8))
SvTAIL_on(sv);