*/
#define CHR_SVLEN(sv) (UTF ? sv_len_utf8(sv) : SvCUR(sv))
-#define CHR_DIST(a,b) (DO_UTF8(PL_reg_sv) ? utf8_distance(a,b) : a - b)
+#define CHR_DIST(a,b) (PL_reg_sv_utf8 ? utf8_distance(a,b) : a - b)
#define reghop_c(pos,off) ((char*)reghop((U8*)pos, off))
#define reghopmaybe_c(pos,off) ((char*)reghopmaybe((U8*)pos, off))
-#define HOP(pos,off) (DO_UTF8(PL_reg_sv) ? reghop((U8*)pos, off) : (U8*)(pos + off))
-#define HOPMAYBE(pos,off) (DO_UTF8(PL_reg_sv) ? reghopmaybe((U8*)pos, off) : (U8*)(pos + off))
+#define HOP(pos,off) (PL_reg_sv_utf8 ? reghop((U8*)pos, off) : (U8*)(pos + off))
+#define HOPMAYBE(pos,off) (PL_reg_sv_utf8 ? reghopmaybe((U8*)pos, off) : (U8*)(pos + off))
#define HOPc(pos,off) ((char*)HOP(pos,off))
#define HOPMAYBEc(pos,off) ((char*)HOPMAYBE(pos,off))
+#define HOPBACK(pos, off) ( \
+ (UTF && PL_reg_sv_utf8) \
+ ? reghopmaybe((U8*)pos, -off) \
+ : (pos - off >= PL_bostr) \
+ ? (U8*)(pos - off) \
+ : (U8*)NULL \
+)
+#define HOPBACKc(pos, off) (char*)HOPBACK(pos, off)
+
#define reghop3_c(pos,off,lim) ((char*)reghop3((U8*)pos, off, (U8*)lim))
#define reghopmaybe3_c(pos,off,lim) ((char*)reghopmaybe3((U8*)pos, off, (U8*)lim))
-#define HOP3(pos,off,lim) (DO_UTF8(PL_reg_sv) ? reghop3((U8*)pos, off, (U8*)lim) : (U8*)(pos + off))
-#define HOPMAYBE3(pos,off,lim) (DO_UTF8(PL_reg_sv) ? reghopmaybe3((U8*)pos, off, (U8*)lim) : (U8*)(pos + off))
+#define HOP3(pos,off,lim) (PL_reg_sv_utf8 ? reghop3((U8*)pos, off, (U8*)lim) : (U8*)(pos + off))
+#define HOPMAYBE3(pos,off,lim) (PL_reg_sv_utf8 ? reghopmaybe3((U8*)pos, off, (U8*)lim) : (U8*)(pos + off))
#define HOP3c(pos,off,lim) ((char*)HOP3(pos,off,lim))
#define HOPMAYBE3c(pos,off,lim) ((char*)HOPMAYBE3(pos,off,lim))
int paren_elems_to_push = (PL_regsize - parenfloor) * REGCP_PAREN_ELEMS;
int p;
-#define REGCP_OTHER_ELEMS 5
+ if (paren_elems_to_push < 0)
+ Perl_croak(aTHX_ "panic: paren_elems_to_push < 0");
+
+#define REGCP_OTHER_ELEMS 6
SSCHECK(paren_elems_to_push + REGCP_OTHER_ELEMS);
for (p = PL_regsize; p > parenfloor; p--) {
/* REGCP_PARENS_ELEMS are pushed per pairs of parentheses. */
/* REGCP_OTHER_ELEMS are pushed in any case, parentheses or no. */
SSPUSHINT(PL_regsize);
SSPUSHINT(*PL_reglastparen);
+ SSPUSHINT(*PL_reglastcloseparen);
SSPUSHPTR(PL_reginput);
#define REGCP_FRAME_ELEMS 2
/* REGCP_FRAME_ELEMS are part of the REGCP_OTHER_ELEMS and
assert(i == SAVEt_REGCONTEXT); /* Check that the magic cookie is there. */
i = SSPOPINT; /* Parentheses elements to pop. */
input = (char *) SSPOPPTR;
+ *PL_reglastcloseparen = SSPOPINT;
*PL_reglastparen = SSPOPINT;
PL_regsize = SSPOPINT;
Perl_re_intuit_start(pTHX_ regexp *prog, SV *sv, char *strpos,
char *strend, U32 flags, re_scream_pos_data *data)
{
- register I32 start_shift;
+ register I32 start_shift = 0;
/* Should be nonnegative! */
- register I32 end_shift;
+ register I32 end_shift = 0;
register char *s;
register SV *check;
char *strbeg;
char *t;
I32 ml_anch;
- char *tmp;
register char *other_last = Nullch; /* other substr checked before this */
- char *check_at; /* check substr found at this pos */
+ char *check_at = Nullch; /* check substr found at this pos */
#ifdef DEBUGGING
char *i_strpos = strpos;
#endif
&& !PL_multiline ) ); /* Check after \n? */
if (!ml_anch) {
- if ( !(prog->reganch & ROPT_ANCH_GPOS) /* Checked by the caller */
+ if ( !(prog->reganch & (ROPT_ANCH_GPOS /* Checked by the caller */
+ | ROPT_IMPLICIT)) /* not a real BOL */
/* SvCUR is not set on references: SvRV and SvPVX overlap */
&& sv && !SvROK(sv)
&& (strpos != strbeg)) {
goto fail;
}
if (prog->check_offset_min == prog->check_offset_max &&
- !(prog->reganch & ROPT_SANY_SEEN)) {
+ !(prog->reganch & ROPT_CANY_SEEN)) {
/* Substring at constant offset from beg-of-str... */
I32 slen;
if (data)
*data->scream_olds = s;
}
- else if (prog->reganch & ROPT_SANY_SEEN)
+ else if (prog->reganch & ROPT_CANY_SEEN)
s = fbm_instr((U8*)(s + start_shift),
(U8*)(strend - end_shift),
check, PL_multiline ? FBMrf_MULTILINE : 0);
PL_regdata = prog->data;
PL_bostr = startpos;
}
- s = find_byclass(prog, prog->regstclass, s, endpos, startpos, 1);
+ s = find_byclass(prog, prog->regstclass, s, endpos, startpos, 1);
if (!s) {
#ifdef DEBUGGING
- char *what;
+ char *what = 0;
#endif
if (endpos == strend) {
DEBUG_r( PerlIO_printf(Perl_debug_log,
DEBUG_r( what = "floating" );
goto hop_and_restart;
}
- DEBUG_r( if (t != s)
- PerlIO_printf(Perl_debug_log,
+ if (t != s) {
+ DEBUG_r(PerlIO_printf(Perl_debug_log,
"By STCLASS: moving %ld --> %ld\n",
- (long)(t - i_strpos), (long)(s - i_strpos));
- else
- PerlIO_printf(Perl_debug_log,
- "Does not contradict STCLASS...\n") );
+ (long)(t - i_strpos), (long)(s - i_strpos))
+ );
+ }
+ else {
+ DEBUG_r(PerlIO_printf(Perl_debug_log,
+ "Does not contradict STCLASS...\n");
+ );
+ }
}
giveup:
DEBUG_r(PerlIO_printf(Perl_debug_log, "%s%s:%s match at offset %ld\n",
unsigned int c2;
char *e;
register I32 tmp = 1; /* Scratch variable? */
- register bool do_utf8 = DO_UTF8(PL_reg_sv);
+ register bool do_utf8 = PL_reg_sv_utf8;
/* We know what class it must start with. */
switch (OP(c)) {
s += do_utf8 ? UTF8SKIP(s) : 1;
}
break;
+ case CANY:
+ while (s < strend) {
+ if (tmp && (norun || regtry(prog, s)))
+ goto got_it;
+ else
+ tmp = doevery;
+ s++;
+ }
+ break;
case EXACTF:
m = STRING(c);
ln = STR_LEN(c);
/* FALL THROUGH */
case BOUND:
if (do_utf8) {
- if (s == startpos)
+ if (s == PL_bostr)
tmp = '\n';
else {
U8 *r = reghop3((U8*)s, -1, (U8*)startpos);
- tmp = (I32)utf8n_to_uvuni(r, s - (char*)r, 0, 0);
+ if (s > (char*)r)
+ tmp = (I32)utf8n_to_uvchr(r, s - (char*)r, 0, 0);
}
tmp = ((OP(c) == BOUND ?
isALNUM_uni(tmp) : isALNUM_LC_uvchr(UNI_TO_NATIVE(tmp))) != 0);
LOAD_UTF8_CHARCLASS(alnum,"a");
while (s < strend) {
if (tmp == !(OP(c) == BOUND ?
- swash_fetch(PL_utf8_alnum, (U8*)s) :
+ swash_fetch(PL_utf8_alnum, (U8*)s, do_utf8) :
isALNUM_LC_utf8((U8*)s)))
{
tmp = !tmp;
}
}
else {
- tmp = (s != startpos) ? UCHARAT(s - 1) : '\n';
+ tmp = (s != PL_bostr) ? UCHARAT(s - 1) : '\n';
tmp = ((OP(c) == BOUND ? isALNUM(tmp) : isALNUM_LC(tmp)) != 0);
while (s < strend) {
if (tmp ==
/* FALL THROUGH */
case NBOUND:
if (do_utf8) {
- if (s == startpos)
+ if (s == PL_bostr)
tmp = '\n';
else {
U8 *r = reghop3((U8*)s, -1, (U8*)startpos);
- tmp = (I32)utf8n_to_uvuni(r, s - (char*)r, 0, 0);
+ if (s > (char*)r)
+ tmp = (I32)utf8n_to_uvchr(r, s - (char*)r, 0, 0);
}
tmp = ((OP(c) == NBOUND ?
isALNUM_uni(tmp) : isALNUM_LC_uvchr(UNI_TO_NATIVE(tmp))) != 0);
LOAD_UTF8_CHARCLASS(alnum,"a");
while (s < strend) {
if (tmp == !(OP(c) == NBOUND ?
- swash_fetch(PL_utf8_alnum, (U8*)s) :
+ swash_fetch(PL_utf8_alnum, (U8*)s, do_utf8) :
isALNUM_LC_utf8((U8*)s)))
tmp = !tmp;
else if ((norun || regtry(prog, s)))
}
}
else {
- tmp = (s != startpos) ? UCHARAT(s - 1) : '\n';
+ tmp = (s != PL_bostr) ? UCHARAT(s - 1) : '\n';
tmp = ((OP(c) == NBOUND ?
isALNUM(tmp) : isALNUM_LC(tmp)) != 0);
while (s < strend) {
if (do_utf8) {
LOAD_UTF8_CHARCLASS(alnum,"a");
while (s < strend) {
- if (swash_fetch(PL_utf8_alnum, (U8*)s)) {
+ if (swash_fetch(PL_utf8_alnum, (U8*)s, do_utf8)) {
if (tmp && (norun || regtry(prog, s)))
goto got_it;
else
if (do_utf8) {
LOAD_UTF8_CHARCLASS(alnum,"a");
while (s < strend) {
- if (!swash_fetch(PL_utf8_alnum, (U8*)s)) {
+ if (!swash_fetch(PL_utf8_alnum, (U8*)s, do_utf8)) {
if (tmp && (norun || regtry(prog, s)))
goto got_it;
else
if (do_utf8) {
LOAD_UTF8_CHARCLASS(space," ");
while (s < strend) {
- if (*s == ' ' || swash_fetch(PL_utf8_space,(U8*)s)) {
+ if (*s == ' ' || swash_fetch(PL_utf8_space,(U8*)s, do_utf8)) {
if (tmp && (norun || regtry(prog, s)))
goto got_it;
else
if (do_utf8) {
LOAD_UTF8_CHARCLASS(space," ");
while (s < strend) {
- if (!(*s == ' ' || swash_fetch(PL_utf8_space,(U8*)s))) {
+ if (!(*s == ' ' || swash_fetch(PL_utf8_space,(U8*)s, do_utf8))) {
if (tmp && (norun || regtry(prog, s)))
goto got_it;
else
if (do_utf8) {
LOAD_UTF8_CHARCLASS(digit,"0");
while (s < strend) {
- if (swash_fetch(PL_utf8_digit,(U8*)s)) {
+ if (swash_fetch(PL_utf8_digit,(U8*)s, do_utf8)) {
if (tmp && (norun || regtry(prog, s)))
goto got_it;
else
if (do_utf8) {
LOAD_UTF8_CHARCLASS(digit,"0");
while (s < strend) {
- if (!swash_fetch(PL_utf8_digit,(U8*)s)) {
+ if (!swash_fetch(PL_utf8_digit,(U8*)s, do_utf8)) {
if (tmp && (norun || regtry(prog, s)))
goto got_it;
else
}
minlen = prog->minlen;
- if (do_utf8) {
- if (!(prog->reganch & ROPT_SANY_SEEN))
+ if (do_utf8 && !(prog->reganch & ROPT_CANY_SEEN)) {
if (utf8_distance((U8*)strend, (U8*)startpos) < minlen) goto phooey;
}
else {
- if (strend - startpos < minlen) goto phooey;
- }
-
- if (startpos == strbeg) /* is ^ valid at stringarg? */
- PL_regprev = '\n';
- else {
- if (prog->reganch & ROPT_UTF8 && do_utf8) {
- U8 *s = reghop3((U8*)stringarg, -1, (U8*)strbeg);
- PL_regprev = utf8n_to_uvuni(s, (U8*)stringarg - s, NULL, 0);
- }
- else
- PL_regprev = (U32)stringarg[-1];
- if (!PL_multiline && PL_regprev == '\n')
- PL_regprev = '\0'; /* force ^ to NOT match */
+ if (strend - startpos < minlen) goto phooey;
}
/* Check validity of program. */
PL_reg_ganch = startpos;
else if (sv && SvTYPE(sv) >= SVt_PVMG
&& SvMAGIC(sv)
- && (mg = mg_find(sv, 'g')) && mg->mg_len >= 0) {
+ && (mg = mg_find(sv, PERL_MAGIC_regex_global))
+ && mg->mg_len >= 0) {
PL_reg_ganch = strbeg + mg->mg_len; /* Defined pos() */
if (prog->reganch & ROPT_ANCH_GPOS) {
- if (s > PL_reg_ganch)
+ if (s > PL_reg_ganch)
goto phooey;
s = PL_reg_ganch;
}
s++;
}
}
- DEBUG_r(did_match ||
+ DEBUG_r(if (!did_match)
PerlIO_printf(Perl_debug_log,
- "Did not find anchored character...\n"));
+ "Did not find anchored character...\n")
+ );
}
/*SUPPRESS 560*/
else if (do_utf8 == (UTF!=0) &&
}
}
}
- DEBUG_r(did_match ||
- PerlIO_printf(Perl_debug_log, "Did not find %s substr `%s%.*s%s'%s...\n",
+ DEBUG_r(if (!did_match)
+ PerlIO_printf(Perl_debug_log,
+ "Did not find %s substr `%s%.*s%s'%s...\n",
((must == prog->anchored_substr)
? "anchored" : "floating"),
PL_colors[0],
(int)(SvCUR(must) - (SvTAIL(must)!=0)),
SvPVX(must),
- PL_colors[1], (SvTAIL(must) ? "$" : "")));
+ PL_colors[1], (SvTAIL(must) ? "$" : ""))
+ );
goto phooey;
}
else if ((c = prog->regstclass)) {
}
if (!(SvTYPE(PL_reg_sv) >= SVt_PVMG && SvMAGIC(PL_reg_sv)
- && (mg = mg_find(PL_reg_sv, 'g')))) {
+ && (mg = mg_find(PL_reg_sv, PERL_MAGIC_regex_global)))) {
/* prepare for quick setting of pos */
- sv_magic(PL_reg_sv, (SV*)0, 'g', Nullch, 0);
- mg = mg_find(PL_reg_sv, 'g');
+ sv_magic(PL_reg_sv, (SV*)0,
+ PERL_MAGIC_regex_global, Nullch, 0);
+ mg = mg_find(PL_reg_sv, PERL_MAGIC_regex_global);
mg->mg_len = -1;
}
PL_reg_magic = mg;
PL_reg_oldpos = mg->mg_len;
SAVEDESTRUCTOR_X(restore_pos, 0);
}
- if (!PL_reg_curpm)
+ if (!PL_reg_curpm) {
Newz(22,PL_reg_curpm, 1, PMOP);
- PL_reg_curpm->op_pmregexp = prog;
+#ifdef USE_ITHREADS
+ {
+ SV* repointer = newSViv(0);
+ /* so we know which PL_regex_padav element is PL_reg_curpm */
+ SvFLAGS(repointer) |= SVf_BREAK;
+ av_push(PL_regex_padav,repointer);
+ PL_reg_curpm->op_pmoffset = av_len(PL_regex_padav);
+ PL_regex_pad = AvARRAY(PL_regex_padav);
+ }
+#endif
+ }
+ PM_SETRE(PL_reg_curpm, prog);
PL_reg_oldcurpm = PL_curpm;
PL_curpm = PL_reg_curpm;
if (RX_MATCH_COPIED(prog)) {
PL_regstartp = prog->startp;
PL_regendp = prog->endp;
PL_reglastparen = &prog->lastparen;
+ PL_reglastcloseparen = &prog->lastcloseparen;
prog->lastparen = 0;
PL_regsize = 0;
DEBUG_r(PL_reg_starttry = startpos);
re_unwind_branch_t branch;
} re_unwind_t;
+#define sayYES goto yes
+#define sayNO goto no
+#define sayYES_FINAL goto yes_final
+#define sayYES_LOUD goto yes_loud
+#define sayNO_FINAL goto no_final
+#define sayNO_SILENT goto do_no
+#define saySAME(x) if (x) goto yes; else goto no
+
+#define REPORT_CODE_OFF 24
+
/*
- regmatch - main matching routine
*
register I32 nextchr; /* renamed nextchr - nextchar colides with
function of same name */
register I32 n; /* no or next */
- register I32 ln; /* len or last */
- register char *s; /* operand or save */
+ register I32 ln = 0; /* len or last */
+ register char *s = Nullch; /* operand or save */
register char *locinput = PL_reginput;
- register I32 c1, c2, paren; /* case fold search, parenth */
+ register I32 c1 = 0, c2 = 0, paren; /* case fold search, parenth */
int minmod = 0, sw = 0, logical = 0;
I32 unwind = 0;
+#if 0
I32 firstcp = PL_savestack_ix;
- register bool do_utf8 = DO_UTF8(PL_reg_sv);
+#endif
+ register bool do_utf8 = PL_reg_sv_utf8;
#ifdef DEBUGGING
PL_regindent++;
nextchr = UCHARAT(locinput);
scan = prog;
while (scan != NULL) {
-#define sayNO_L (logical ? (logical = 0, sw = 0, goto cont) : sayNO)
-#if 1
-# define sayYES goto yes
-# define sayNO goto no
-# define sayYES_FINAL goto yes_final
-# define sayYES_LOUD goto yes_loud
-# define sayNO_FINAL goto no_final
-# define sayNO_SILENT goto do_no
-# define saySAME(x) if (x) goto yes; else goto no
-# define REPORT_CODE_OFF 24
-#else
-# define sayYES return 1
-# define sayNO return 0
-# define sayYES_FINAL return 1
-# define sayYES_LOUD return 1
-# define sayNO_FINAL return 0
-# define sayNO_SILENT return 0
-# define saySAME(x) return x
-#endif
+
DEBUG_r( {
SV *prop = sv_newmortal();
int docolor = *PL_colors[0];
switch (OP(scan)) {
case BOL:
- if (locinput == PL_bostr
- ? PL_regprev == '\n'
- : (PL_multiline &&
- (nextchr || locinput < PL_regeol) && locinput[-1] == '\n') )
+ if (locinput == PL_bostr || (PL_multiline &&
+ (nextchr || locinput < PL_regeol) && locinput[-1] == '\n') )
{
/* regtill = regbol; */
break;
}
sayNO;
case MBOL:
- if (locinput == PL_bostr
- ? PL_regprev == '\n'
- : ((nextchr || locinput < PL_regeol) && locinput[-1] == '\n') )
+ if (locinput == PL_bostr ||
+ ((nextchr || locinput < PL_regeol) && locinput[-1] == '\n'))
{
break;
}
case SANY:
if (!nextchr && locinput >= PL_regeol)
sayNO;
+ if (do_utf8) {
+ locinput += PL_utf8skip[nextchr];
+ if (locinput > PL_regeol)
+ sayNO;
+ nextchr = UCHARAT(locinput);
+ }
+ else
+ nextchr = UCHARAT(++locinput);
+ break;
+ case CANY:
+ if (!nextchr && locinput >= PL_regeol)
+ sayNO;
nextchr = UCHARAT(++locinput);
break;
case REG_ANY:
if (l >= PL_regeol) {
sayNO;
}
- if ((UTF ? utf8n_to_uvuni((U8*)s, e - s, 0, 0) : *((U8*)s)) !=
+ if ((UTF ? utf8n_to_uvchr((U8*)s, e - s, 0, 0) : *((U8*)s)) !=
(c1 ? toLOWER_utf8((U8*)l) : toLOWER_LC_utf8((U8*)l)))
sayNO;
s += UTF ? UTF8SKIP(s) : 1;
if (!nextchr)
sayNO;
if (do_utf8) {
+ LOAD_UTF8_CHARCLASS(alnum,"a");
if (!(OP(scan) == ALNUM
- ? swash_fetch(PL_utf8_alnum, (U8*)locinput)
+ ? swash_fetch(PL_utf8_alnum, (U8*)locinput, do_utf8)
: isALNUM_LC_utf8((U8*)locinput)))
{
sayNO;
if (do_utf8) {
LOAD_UTF8_CHARCLASS(alnum,"a");
if (OP(scan) == NALNUM
- ? swash_fetch(PL_utf8_alnum, (U8*)locinput)
+ ? swash_fetch(PL_utf8_alnum, (U8*)locinput, do_utf8)
: isALNUM_LC_utf8((U8*)locinput))
{
sayNO;
case NBOUND:
/* was last char in word? */
if (do_utf8) {
- if (locinput == PL_regbol)
- ln = PL_regprev;
+ if (locinput == PL_bostr)
+ ln = '\n';
else {
U8 *r = reghop((U8*)locinput, -1);
- ln = utf8n_to_uvuni(r, s - (char*)r, 0, 0);
+ ln = utf8n_to_uvchr(r, s - (char*)r, 0, 0);
}
if (OP(scan) == BOUND || OP(scan) == NBOUND) {
ln = isALNUM_uni(ln);
LOAD_UTF8_CHARCLASS(alnum,"a");
- n = swash_fetch(PL_utf8_alnum, (U8*)locinput);
+ n = swash_fetch(PL_utf8_alnum, (U8*)locinput, do_utf8);
}
else {
ln = isALNUM_LC_uvchr(UNI_TO_NATIVE(ln));
}
}
else {
- ln = (locinput != PL_regbol) ?
- UCHARAT(locinput - 1) : PL_regprev;
+ ln = (locinput != PL_bostr) ?
+ UCHARAT(locinput - 1) : '\n';
if (OP(scan) == BOUND || OP(scan) == NBOUND) {
ln = isALNUM(ln);
n = isALNUM(nextchr);
if (UTF8_IS_CONTINUED(nextchr)) {
LOAD_UTF8_CHARCLASS(space," ");
if (!(OP(scan) == SPACE
- ? swash_fetch(PL_utf8_space, (U8*)locinput)
+ ? swash_fetch(PL_utf8_space, (U8*)locinput, do_utf8)
: isSPACE_LC_utf8((U8*)locinput)))
{
sayNO;
if (do_utf8) {
LOAD_UTF8_CHARCLASS(space," ");
if (OP(scan) == NSPACE
- ? swash_fetch(PL_utf8_space, (U8*)locinput)
+ ? swash_fetch(PL_utf8_space, (U8*)locinput, do_utf8)
: isSPACE_LC_utf8((U8*)locinput))
{
sayNO;
if (do_utf8) {
LOAD_UTF8_CHARCLASS(digit,"0");
if (!(OP(scan) == DIGIT
- ? swash_fetch(PL_utf8_digit, (U8*)locinput)
+ ? swash_fetch(PL_utf8_digit, (U8*)locinput, do_utf8)
: isDIGIT_LC_utf8((U8*)locinput)))
{
sayNO;
if (do_utf8) {
LOAD_UTF8_CHARCLASS(digit,"0");
if (OP(scan) == NDIGIT
- ? swash_fetch(PL_utf8_digit, (U8*)locinput)
+ ? swash_fetch(PL_utf8_digit, (U8*)locinput, do_utf8)
: isDIGIT_LC_utf8((U8*)locinput))
{
sayNO;
break;
case CLUMP:
LOAD_UTF8_CHARCLASS(mark,"~");
- if (locinput >= PL_regeol || swash_fetch(PL_utf8_mark,(U8*)locinput))
+ if (locinput >= PL_regeol ||
+ swash_fetch(PL_utf8_mark,(U8*)locinput, do_utf8))
sayNO;
locinput += PL_utf8skip[nextchr];
- while (locinput < PL_regeol && swash_fetch(PL_utf8_mark,(U8*)locinput))
+ while (locinput < PL_regeol &&
+ swash_fetch(PL_utf8_mark,(U8*)locinput, do_utf8))
locinput += UTF8SKIP(locinput);
if (locinput > PL_regeol)
sayNO;
SV *sv = SvROK(ret) ? SvRV(ret) : ret;
if(SvMAGICAL(sv))
- mg = mg_find(sv, 'r');
+ mg = mg_find(sv, PERL_MAGIC_qr);
}
if (mg) {
re = (regexp *)mg->mg_obj;
I32 osize = PL_regsize;
I32 onpar = PL_regnpar;
- pm.op_pmflags = 0;
+ Zero(&pm, 1, PMOP);
re = CALLREGCOMP(aTHX_ t, t + len, &pm);
if (!(SvFLAGS(ret)
& (SVs_TEMP | SVs_PADTMP | SVf_READONLY)))
- sv_magic(ret,(SV*)ReREFCNT_inc(re),'r',0,0);
+ sv_magic(ret,(SV*)ReREFCNT_inc(re),
+ PERL_MAGIC_qr,0,0);
PL_regprecomp = oprecomp;
PL_regsize = osize;
PL_regnpar = onpar;
cache_re(re);
state.ss = PL_savestack_ix;
*PL_reglastparen = 0;
+ *PL_reglastcloseparen = 0;
PL_reg_call_cc = &state;
PL_reginput = locinput;
/* XXXX This is too dramatic a measure... */
PL_reg_maxiter = 0;
+ logical = 0;
sayNO;
}
sw = SvTRUE(ret);
PL_regendp[n] = locinput - PL_bostr;
if (n > *PL_reglastparen)
*PL_reglastparen = n;
+ *PL_reglastcloseparen = n;
break;
case GROUPP:
n = ARG(scan); /* which paren pair */
inner = NEXTOPER(scan);
do_branch:
{
- CHECKPOINT lastcp;
c1 = OP(scan);
if (OP(next) != c1) /* No choice. */
next = inner; /* Avoid recursion. */
minmod = 0;
if (ln && regrepeat_hard(scan, ln, &l) < ln)
sayNO;
- if (ln && l == 0 && n >= ln
- /* In fact, this is tricky. If paren, then the
- fact that we did/didnot match may influence
- future execution. */
- && !(paren && ln == 0))
- ln = n;
+ /* if we matched something zero-length we don't need to
+ backtrack - capturing parens are already defined, so
+ the caveat in the maximal case doesn't apply
+
+ XXXX if ln == 0, we can redo this check first time
+ through the following loop
+ */
+ if (ln && l == 0)
+ n = ln; /* don't backtrack */
locinput = PL_reginput;
if (PL_regkind[(U8)OP(next)] == EXACT) {
c1 = (U8)*STRING(next);
UCHARAT(PL_reginput) == c2)
{
if (paren) {
- if (n) {
+ if (ln) {
PL_regstartp[paren] =
HOPc(PL_reginput, -l) - PL_bostr;
PL_regendp[paren] = PL_reginput - PL_bostr;
}
else {
n = regrepeat_hard(scan, n, &l);
- if (n != 0 && l == 0
- /* In fact, this is tricky. If paren, then the
- fact that we did/didnot match may influence
- future execution. */
- && !(paren && ln == 0))
- ln = n;
+ /* if we matched something zero-length we don't need to
+ backtrack, unless the minimum count is zero and we
+ are capturing the result - in that case the capture
+ being defined or not may affect later execution
+ */
+ if (n != 0 && l == 0 && !(paren && ln == 0))
+ ln = n; /* don't backtrack */
locinput = PL_reginput;
DEBUG_r(
PerlIO_printf(Perl_debug_log,
c = utf8_to_uvchr((U8*)PL_reginput, NULL);
else
c = UCHARAT(PL_reginput);
+ /* If it could work, try it. */
+ if (c == c1 || c == c2)
+ {
+ TRYPAREN(paren, n, PL_reginput);
+ REGCP_UNWIND(lastcp);
+ }
}
/* If it could work, try it. */
- if (c1 == -1000 || c == c1 || c == c2)
+ else if (c1 == -1000)
{
TRYPAREN(paren, n, PL_reginput);
REGCP_UNWIND(lastcp);
}
REGCP_SET(lastcp);
if (paren) {
- UV c;
+ UV c = 0;
while (n >= ln) {
if (c1 != -1000) {
if (do_utf8)
}
}
else {
- UV c;
+ UV c = 0;
while (n >= ln) {
if (c1 != -1000) {
if (do_utf8)
case UNLESSM:
n = 0;
if (scan->flags) {
- if (UTF) { /* XXXX This is absolutely
- broken, we read before
- start of string. */
- s = HOPMAYBEc(locinput, -scan->flags);
- if (!s)
- goto say_yes;
- PL_reginput = s;
- }
- else {
- if (locinput < PL_bostr + scan->flags)
- goto say_yes;
- PL_reginput = locinput - scan->flags;
- goto do_ifmatch;
- }
+ s = HOPBACKc(locinput, scan->flags);
+ if (!s)
+ goto say_yes;
+ PL_reginput = s;
}
else
PL_reginput = locinput;
case IFMATCH:
n = 1;
if (scan->flags) {
- if (UTF) { /* XXXX This is absolutely
- broken, we read before
- start of string. */
- s = HOPMAYBEc(locinput, -scan->flags);
- if (!s || s < PL_bostr)
- goto say_no;
- PL_reginput = s;
- }
- else {
- if (locinput < PL_bostr + scan->flags)
- goto say_no;
- PL_reginput = locinput - scan->flags;
- goto do_ifmatch;
- }
+ s = HOPBACKc(locinput, scan->flags);
+ if (!s)
+ goto say_no;
+ PL_reginput = s;
}
else
PL_reginput = locinput;
register I32 c;
register char *loceol = PL_regeol;
register I32 hardcount = 0;
- register bool do_utf8 = DO_UTF8(PL_reg_sv);
+ register bool do_utf8 = PL_reg_sv_utf8;
scan = PL_reginput;
if (max != REG_INFTY && max < loceol - scan)
case SANY:
scan = loceol;
break;
+ case CANY:
+ scan = loceol;
+ break;
case EXACT: /* length of string is 1 */
c = (U8)*STRING(p);
while (scan < loceol && UCHARAT(scan) == c)
loceol = PL_regeol;
LOAD_UTF8_CHARCLASS(alnum,"a");
while (hardcount < max && scan < loceol &&
- swash_fetch(PL_utf8_alnum, (U8*)scan)) {
+ swash_fetch(PL_utf8_alnum, (U8*)scan, do_utf8)) {
scan += UTF8SKIP(scan);
hardcount++;
}
loceol = PL_regeol;
LOAD_UTF8_CHARCLASS(alnum,"a");
while (hardcount < max && scan < loceol &&
- !swash_fetch(PL_utf8_alnum, (U8*)scan)) {
+ !swash_fetch(PL_utf8_alnum, (U8*)scan, do_utf8)) {
scan += UTF8SKIP(scan);
hardcount++;
}
loceol = PL_regeol;
LOAD_UTF8_CHARCLASS(space," ");
while (hardcount < max && scan < loceol &&
- (*scan == ' ' || swash_fetch(PL_utf8_space,(U8*)scan))) {
+ (*scan == ' ' ||
+ swash_fetch(PL_utf8_space,(U8*)scan, do_utf8))) {
scan += UTF8SKIP(scan);
hardcount++;
}
loceol = PL_regeol;
LOAD_UTF8_CHARCLASS(space," ");
while (hardcount < max && scan < loceol &&
- !(*scan == ' ' || swash_fetch(PL_utf8_space,(U8*)scan))) {
+ !(*scan == ' ' ||
+ swash_fetch(PL_utf8_space,(U8*)scan, do_utf8))) {
scan += UTF8SKIP(scan);
hardcount++;
}
loceol = PL_regeol;
LOAD_UTF8_CHARCLASS(digit,"0");
while (hardcount < max && scan < loceol &&
- swash_fetch(PL_utf8_digit,(U8*)scan)) {
+ swash_fetch(PL_utf8_digit, (U8*)scan, do_utf8)) {
scan += UTF8SKIP(scan);
hardcount++;
}
loceol = PL_regeol;
LOAD_UTF8_CHARCLASS(digit,"0");
while (hardcount < max && scan < loceol &&
- !swash_fetch(PL_utf8_digit,(U8*)scan)) {
+ !swash_fetch(PL_utf8_digit, (U8*)scan, do_utf8)) {
scan += UTF8SKIP(scan);
hardcount++;
}
STATIC I32
S_regrepeat_hard(pTHX_ regnode *p, I32 max, I32 *lp)
{
- register char *scan;
+ register char *scan = Nullch;
register char *start;
register char *loceol = PL_regeol;
I32 l = 0;
return 0;
start = PL_reginput;
- if (DO_UTF8(PL_reg_sv)) {
+ if (PL_reg_sv_utf8) {
while (PL_reginput < loceol && (scan = PL_reginput, res = regmatch(p))) {
if (!count++) {
l = 0;
char flags = ANYOF_FLAGS(n);
bool match = FALSE;
UV c;
- STRLEN len;
+ STRLEN len = 0;
- if (do_utf8)
- c = utf8_to_uvchr(p, &len);
- else
- c = *p;
+ c = do_utf8 ? utf8_to_uvchr(p, &len) : *p;
if (do_utf8 || (flags & ANYOF_UNICODE)) {
if (do_utf8 && !ANYOF_RUNTIME(n)) {
if (len != (STRLEN)-1 && c < 256 && ANYOF_BITMAP_TEST(n, c))
match = TRUE;
}
- if (do_utf8 && flags & ANYOF_UNICODE_ALL && c >= 256)
+ if (!match && do_utf8 && (flags & ANYOF_UNICODE_ALL) && c >= 256)
match = TRUE;
if (!match) {
SV *sw = regclass_swash(n, TRUE, 0);
if (sw) {
- if (swash_fetch(sw, p))
+ if (swash_fetch(sw, p, do_utf8))
match = TRUE;
else if (flags & ANYOF_FOLD) {
U8 tmpbuf[UTF8_MAXLEN+1];
}
else
uvchr_to_utf8(tmpbuf, toLOWER_utf8(p));
- if (swash_fetch(sw, tmpbuf))
+ if (swash_fetch(sw, tmpbuf, do_utf8))
match = TRUE;
}
}
if (ANYOF_BITMAP_TEST(n, c))
match = TRUE;
else if (flags & ANYOF_FOLD) {
- I32 f;
+ I32 f;
if (flags & ANYOF_LOCALE) {
PL_reg_flags |= RF_tainted;
https://perl5.git.perl.org / perl5.git / blobdiff