#ifdef PERL_EXT_RE_BUILD
# ifndef PERL_IN_XSUB_RE
# define PERL_IN_XSUB_RE
# endif
# if defined(PERL_EXT_RE_DEBUG) && !defined(DEBUGGING)
# define DEBUGGING
# endif
#endif
#ifdef PERL_IN_XSUB_RE
# define Perl_regexec_flags my_regexec
# define Perl_regdump my_regdump
# define Perl_regprop my_regprop
# define Perl_re_intuit_start my_re_intuit_start
# define Perl_pregexec my_pregexec
# define Perl_reginitcolors my_reginitcolors
# define PERL_NO_GET_CONTEXT
#endif
#include "EXTERN.h"
#define PERL_IN_REGEXEC_C
#include "perl.h"
#ifdef PERL_IN_XSUB_RE
# if defined(PERL_CAPI) || defined(PERL_OBJECT)
# include "XSUB.h"
# endif
#endif
#include "regcomp.h"
#define RF_tainted 1
#define RF_warned 2
#define RF_evaled 4
#define RF_utf8 8
#define UTF (PL_reg_flags & RF_utf8)
#define RS_init 1
#define RS_set 2
#ifndef STATIC
#define STATIC static
#endif
#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)
#define reghop_c(pos,off) ((char*)reghop((U8*)pos, off))
#define reghopmaybe_c(pos,off) ((char*)reghopmaybe((U8*)pos, off))
#define HOP(pos,off) (UTF ? reghop((U8*)pos, off) : (U8*)(pos + off))
#define HOPMAYBE(pos,off) (UTF ? reghopmaybe((U8*)pos, off) : (U8*)(pos + off))
#define HOPc(pos,off) ((char*)HOP(pos,off))
#define HOPMAYBEc(pos,off) ((char*)HOPMAYBE(pos,off))
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;
SSCHECK(i + 5);
for (p = PL_regsize; p > parenfloor; p--) {
SSPUSHINT(PL_regendp[p]);
SSPUSHINT(PL_regstartp[p]);
SSPUSHPTR(PL_reg_start_tmp[p]);
SSPUSHINT(p);
}
SSPUSHINT(PL_regsize);
SSPUSHINT(*PL_reglastparen);
SSPUSHPTR(PL_reginput);
SSPUSHINT(i + 3);
SSPUSHINT(SAVEt_REGCONTEXT);
return retval;
}
# define REGCP_SET DEBUG_r(PerlIO_printf(Perl_debug_log, \
" Setting an EVAL scope, savestack=%"IVdf"\n", \
(IV)PL_savestack_ix)); lastcp = PL_savestack_ix
# define REGCP_UNWIND DEBUG_r(lastcp != PL_savestack_ix ? \
PerlIO_printf(Perl_debug_log, \
" Clearing an EVAL scope, savestack=%"IVdf"..%"IVdf"\n", \
(IV)lastcp, (IV)PL_savestack_ix) : 0); regcpblow(lastcp)
STATIC char *
S_regcppop(pTHX)
{
dTHR;
I32 i = SSPOPINT;
U32 paren = 0;
char *input;
I32 tmps;
assert(i == SAVEt_REGCONTEXT);
i = SSPOPINT;
input = (char *) SSPOPPTR;
*PL_reglastparen = SSPOPINT;
PL_regsize = SSPOPINT;
for (i -= 3; i > 0; i -= 4) {
paren = (U32)SSPOPINT;
PL_reg_start_tmp[paren] = (char *) SSPOPPTR;
PL_regstartp[paren] = SSPOPINT;
tmps = SSPOPINT;
if (paren <= *PL_reglastparen)
PL_regendp[paren] = tmps;
DEBUG_r(
PerlIO_printf(Perl_debug_log,
" restoring \\%"UVuf" to %"IVdf"(%"IVdf")..%"IVdf"%s\n",
(UV)paren, (IV)PL_regstartp[paren],
(IV)(PL_reg_start_tmp[paren] - PL_bostr),
(IV)PL_regendp[paren],
(paren > *PL_reglastparen ? "(no)" : ""));
);
}
DEBUG_r(
if (*PL_reglastparen + 1 <= PL_regnpar) {
PerlIO_printf(Perl_debug_log,
" restoring \\%"IVdf"..\\%"IVdf" to undef\n",
(IV)(*PL_reglastparen + 1), (IV)PL_regnpar);
}
);
for (paren = *PL_reglastparen + 1; paren <= PL_regnpar; paren++) {
if (paren > PL_regsize)
PL_regstartp[paren] = -1;
PL_regendp[paren] = -1;
}
return input;
}
STATIC char *
S_regcp_set_to(pTHX_ I32 ss)
{
dTHR;
I32 tmp = PL_savestack_ix;
PL_savestack_ix = ss;
regcppop();
PL_savestack_ix = tmp;
return Nullch;
}
typedef struct re_cc_state
{
I32 ss;
regnode *node;
struct re_cc_state *prev;
CURCUR *cc;
regexp *re;
} re_cc_state;
#define regcpblow(cp) LEAVE_SCOPE(cp)
I32
Perl_pregexec(pTHX_ register regexp *prog, char *stringarg, register char *strend,
char *strbeg, I32 minend, SV *screamer, U32 nosave)
{
return
regexec_flags(prog, stringarg, strend, strbeg, minend, screamer, NULL,
nosave ? 0 : REXEC_COPY_STR);
}
STATIC void
S_cache_re(pTHX_ regexp *prog)
{
dTHR;
PL_regprecomp = prog->precomp;
#ifdef DEBUGGING
PL_regprogram = prog->program;
#endif
PL_regnpar = prog->nparens;
PL_regdata = prog->data;
PL_reg_re = prog;
}
char *
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 end_shift;
register char *s;
register SV *check;
char *t;
I32 ml_anch;
char *tmp;
register char *other_last = Nullch;
char *check_at;
#ifdef DEBUGGING
char *i_strpos = strpos;
#endif
DEBUG_r( if (!PL_colorset) reginitcolors() );
DEBUG_r(PerlIO_printf(Perl_debug_log,
"%sGuessing start of match, REx%s `%s%.60s%s%s' against `%s%.*s%s%s'...\n",
PL_colors[4],PL_colors[5],PL_colors[0],
prog->precomp,
PL_colors[1],
(strlen(prog->precomp) > 60 ? "..." : ""),
PL_colors[0],
(int)(strend - strpos > 60 ? 60 : strend - strpos),
strpos, PL_colors[1],
(strend - strpos > 60 ? "..." : ""))
);
if (prog->minlen > strend - strpos) {
DEBUG_r(PerlIO_printf(Perl_debug_log, "String too short...\n"));
goto fail;
}
check = prog->check_substr;
if (prog->reganch & ROPT_ANCH) {
ml_anch = !( (prog->reganch & ROPT_ANCH_SINGLE)
|| ( (prog->reganch & ROPT_ANCH_BOL)
&& !PL_multiline ) );
if ((prog->check_offset_min == prog->check_offset_max) && !ml_anch) {
I32 slen;
if ( !(prog->reganch & ROPT_ANCH_GPOS)
&& (sv && (strpos + SvCUR(sv) != strend)) ) {
DEBUG_r(PerlIO_printf(Perl_debug_log, "Not at start...\n"));
goto fail;
}
PL_regeol = strend;
s = HOPc(strpos, prog->check_offset_min);
if (SvTAIL(check)) {
slen = SvCUR(check);
if ( strend - s > slen || strend - s < slen - 1
|| (strend - s == slen && strend[-1] != '\n')) {
DEBUG_r(PerlIO_printf(Perl_debug_log, "String too long...\n"));
goto fail_finish;
}
slen--;
if (slen && (*SvPVX(check) != *s
|| (slen > 1
&& memNE(SvPVX(check), s, slen)))) {
report_neq:
DEBUG_r(PerlIO_printf(Perl_debug_log, "String not equal...\n"));
goto fail_finish;
}
}
else if (*SvPVX(check) != *s
|| ((slen = SvCUR(check)) > 1
&& memNE(SvPVX(check), s, slen)))
goto report_neq;
goto success_at_start;
}
s = strpos;
start_shift = prog->check_offset_min;
end_shift = prog->minlen - start_shift -
CHR_SVLEN(check) + (SvTAIL(check) != 0);
if (!ml_anch) {
I32 end = prog->check_offset_max + CHR_SVLEN(check)
- (SvTAIL(check) != 0);
I32 eshift = strend - s - end;
if (end_shift < eshift)
end_shift = eshift;
}
}
else {
ml_anch = 0;
s = strpos;
start_shift = prog->check_offset_min;
end_shift = prog->minlen - start_shift -
CHR_SVLEN(check) + (SvTAIL(check) != 0);
}
#ifdef DEBUGGING
if (end_shift < 0)
Perl_croak(aTHX_ "panic: end_shift");
#endif
restart:
if (flags & REXEC_SCREAM) {
char *strbeg = SvPVX(sv);
I32 p = -1;
I32 *pp = data ? data->scream_pos : &p;
if (PL_screamfirst[BmRARE(check)] >= 0
|| ( BmRARE(check) == '\n'
&& (BmPREVIOUS(check) == SvCUR(check) - 1)
&& SvTAIL(check) ))
s = screaminstr(sv, check,
start_shift + (s - strbeg), end_shift, pp, 0);
else
goto fail_finish;
if (data)
*data->scream_olds = s;
}
else
s = fbm_instr((unsigned char*)s + start_shift,
(unsigned char*)strend - end_shift,
check, PL_multiline ? FBMrf_MULTILINE : 0);
DEBUG_r(PerlIO_printf(Perl_debug_log, "%s %s substr `%s%.*s%s'%s%s",
(s ? "Found" : "Did not find"),
((check == prog->anchored_substr) ? "anchored" : "floating"),
PL_colors[0],
(int)(SvCUR(check) - (SvTAIL(check)!=0)),
SvPVX(check),
PL_colors[1], (SvTAIL(check) ? "$" : ""),
(s ? " at offset " : "...\n") ) );
if (!s)
goto fail_finish;
check_at = s;
DEBUG_r(PerlIO_printf(Perl_debug_log, "%ld...\n", (long)(s - i_strpos)) );
if (prog->float_substr && prog->anchored_substr) {
if (!other_last)
other_last = strpos;
if (check == prog->float_substr) {
do_other_anchored:
{
char *last = s - start_shift, *last1, *last2;
char *s1 = s;
tmp = PL_bostr;
t = s - prog->check_offset_max;
if (s - strpos > prog->check_offset_max
&& (!(prog->reganch & ROPT_UTF8)
|| (PL_bostr = strpos,
(t = reghopmaybe_c(s, -(prog->check_offset_max)))
&& t > strpos)))
;
else
t = strpos;
t += prog->anchored_offset;
if (t < other_last)
t = other_last;
PL_bostr = tmp;
last2 = last1 = strend - prog->minlen;
if (last < last1)
last1 = last;
s = fbm_instr((unsigned char*)t,
(unsigned char*)last1 + prog->anchored_offset
+ SvCUR(prog->anchored_substr)
- (SvTAIL(prog->anchored_substr)!=0),
prog->anchored_substr, PL_multiline ? FBMrf_MULTILINE : 0);
DEBUG_r(PerlIO_printf(Perl_debug_log, "%s anchored substr `%s%.*s%s'%s",
(s ? "Found" : "Contradicts"),
PL_colors[0],
(int)(SvCUR(prog->anchored_substr)
- (SvTAIL(prog->anchored_substr)!=0)),
SvPVX(prog->anchored_substr),
PL_colors[1], (SvTAIL(prog->anchored_substr) ? "$" : "")));
if (!s) {
if (last1 >= last2) {
DEBUG_r(PerlIO_printf(Perl_debug_log,
", giving up...\n"));
goto fail_finish;
}
DEBUG_r(PerlIO_printf(Perl_debug_log,
", trying floating at offset %ld...\n",
(long)(s1 + 1 - i_strpos)));
PL_regeol = strend;
other_last = last1 + prog->anchored_offset + 1;
s = HOPc(last, 1);
goto restart;
}
else {
DEBUG_r(PerlIO_printf(Perl_debug_log, " at offset %ld...\n",
(long)(s - i_strpos)));
t = s - prog->anchored_offset;
other_last = s + 1;
s = s1;
if (t == strpos)
goto try_at_start;
goto try_at_offset;
}
}
}
else {
char *last, *last1;
char *s1 = s;
t = s - start_shift;
last1 = last = strend - prog->minlen + prog->float_min_offset;
if (last - t > prog->float_max_offset)
last = t + prog->float_max_offset;
s = t + prog->float_min_offset;
if (s < other_last)
s = other_last;
s = fbm_instr((unsigned char*)s,
(unsigned char*)last + SvCUR(prog->float_substr)
- (SvTAIL(prog->float_substr)!=0),
prog->float_substr, PL_multiline ? FBMrf_MULTILINE : 0);
DEBUG_r(PerlIO_printf(Perl_debug_log, "%s floating substr `%s%.*s%s'%s",
(s ? "Found" : "Contradicts"),
PL_colors[0],
(int)(SvCUR(prog->float_substr)
- (SvTAIL(prog->float_substr)!=0)),
SvPVX(prog->float_substr),
PL_colors[1], (SvTAIL(prog->float_substr) ? "$" : "")));
if (!s) {
if (last1 == last) {
DEBUG_r(PerlIO_printf(Perl_debug_log,
", giving up...\n"));
goto fail_finish;
}
DEBUG_r(PerlIO_printf(Perl_debug_log,
", trying anchored starting at offset %ld...\n",
(long)(s1 + 1 - i_strpos)));
other_last = last + 1;
PL_regeol = strend;
s = HOPc(t, 1);
goto restart;
}
else {
DEBUG_r(PerlIO_printf(Perl_debug_log, " at offset %ld...\n",
(long)(s - i_strpos)));
other_last = s + 1;
s = s1;
if (t == strpos)
goto try_at_start;
goto try_at_offset;
}
}
}
t = s - prog->check_offset_max;
tmp = PL_bostr;
if (s - strpos > prog->check_offset_max
&& (!(prog->reganch & ROPT_UTF8)
|| (PL_bostr = strpos,
((t = reghopmaybe_c(s, -(prog->check_offset_max)))
&& t > strpos)))) {
PL_bostr = tmp;
try_at_offset:
if (ml_anch && t[-1] != '\n') {
find_anchor:
while (t < strend - prog->minlen) {
if (*t == '\n') {
if (t < s - prog->check_offset_min) {
if (prog->anchored_substr) {
strpos = t + 1;
DEBUG_r(PerlIO_printf(Perl_debug_log, "Found /%s^%s/m at offset %ld, rescanning for anchored from offset %ld...\n",
PL_colors[0],PL_colors[1], (long)(strpos - i_strpos), (long)(strpos - i_strpos + prog->anchored_offset)));
goto do_other_anchored;
}
s = t + 1;
DEBUG_r(PerlIO_printf(Perl_debug_log, "Found /%s^%s/m at offset %ld...\n",
PL_colors[0],PL_colors[1], (long)(s - i_strpos)));
goto set_useful;
}
DEBUG_r(PerlIO_printf(Perl_debug_log, "Found /%s^%s/m, restarting at offset %ld...\n",
PL_colors[0],PL_colors[1], (long)(t + 1 - i_strpos)));
strpos = s = t + 1;
goto restart;
}
t++;
}
DEBUG_r(PerlIO_printf(Perl_debug_log, "Did not find /%s^%s/m...\n",
PL_colors[0],PL_colors[1]));
goto fail_finish;
}
s = t;
set_useful:
++BmUSEFUL(prog->check_substr);
}
else {
PL_bostr = tmp;
try_at_start:
if (ml_anch && sv
&& (strpos + SvCUR(sv) != strend) && strpos[-1] != '\n'
&& !(prog->reganch & ROPT_IMPLICIT))
{
t = strpos;
goto find_anchor;
}
DEBUG_r( if (ml_anch)
PerlIO_printf(Perl_debug_log, "Does not contradict /%s^%s/m...\n",
PL_colors[0],PL_colors[1]);
);
success_at_start:
if (!(prog->reganch & ROPT_NAUGHTY)
&& prog->check_substr
&& --BmUSEFUL(prog->check_substr) < 0
&& prog->check_substr == prog->float_substr)
{
SvREFCNT_dec(prog->check_substr);
prog->check_substr = Nullsv;
prog->float_substr = Nullsv;
s = strpos;
prog->reganch &= ~RE_USE_INTUIT;
}
else
s = strpos;
}
if (prog->regstclass) {
int cl_l = (PL_regkind[(U8)OP(prog->regstclass)] == EXACT
? STR_LEN(prog->regstclass)
: 1);
char *endpos = (prog->anchored_substr || ml_anch)
? s + (prog->minlen? cl_l : 0)
: (prog->float_substr ? check_at - start_shift + cl_l
: strend) ;
char *startpos = sv && SvPOK(sv) ? strend - SvCUR(sv) : s;
t = s;
if (prog->reganch & ROPT_UTF8) {
PL_regdata = prog->data;
PL_bostr = startpos;
}
s = find_byclass(prog, prog->regstclass, s, endpos, startpos, 1);
if (!s) {
#ifdef DEBUGGING
char *what;
#endif
if (endpos == strend) {
DEBUG_r( PerlIO_printf(Perl_debug_log,
"Could not match STCLASS...\n") );
goto fail;
}
DEBUG_r( PerlIO_printf(Perl_debug_log,
"This position contradicts STCLASS...\n") );
if ((prog->reganch & ROPT_ANCH) && !ml_anch)
goto fail;
if (prog->anchored_substr) {
if (prog->anchored_substr == check) {
DEBUG_r( what = "anchored" );
hop_and_restart:
PL_regeol = strend;
s = HOPc(t, 1);
if (s + start_shift + end_shift > strend) {
DEBUG_r( PerlIO_printf(Perl_debug_log,
"Could not match STCLASS...\n") );
goto fail;
}
DEBUG_r( PerlIO_printf(Perl_debug_log,
"Trying %s substr starting at offset %ld...\n",
what, (long)(s + start_shift - i_strpos)) );
goto restart;
}
if (t + start_shift >= check_at)
goto retry_floating_check;
s = check_at;
DEBUG_r( PerlIO_printf(Perl_debug_log,
"Trying anchored substr starting at offset %ld...\n",
(long)(other_last - i_strpos)) );
goto do_other_anchored;
}
if (ml_anch) {
s = t = t + 1;
DEBUG_r( PerlIO_printf(Perl_debug_log,
"Trying /^/m starting at offset %ld...\n",
(long)(t - i_strpos)) );
goto try_at_offset;
}
if (!prog->float_substr)
goto fail;
retry_floating_check:
t = check_at - start_shift;
DEBUG_r( what = "floating" );
goto hop_and_restart;
}
DEBUG_r( if (t != s)
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") );
}
DEBUG_r(PerlIO_printf(Perl_debug_log, "%sGuessed:%s match at offset %ld\n",
PL_colors[4], PL_colors[5], (long)(s - i_strpos)) );
return s;
fail_finish:
if (prog->check_substr)
BmUSEFUL(prog->check_substr) += 5;
fail:
DEBUG_r(PerlIO_printf(Perl_debug_log, "%sMatch rejected by optimizer%s\n",
PL_colors[4],PL_colors[5]));
return Nullch;
}
STATIC char *
S_find_byclass(pTHX_ regexp * prog, regnode *c, char *s, char *strend, char *startpos, I32 norun)
{
I32 doevery = (prog->reganch & ROPT_SKIP) == 0;
char *m;
STRLEN ln;
unsigned int c1;
unsigned int c2;
char *e;
register I32 tmp = 1;
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 (tmp && (norun || regtry(prog, s)))
goto got_it;
else
tmp = doevery;
}
else
tmp = 1;
s++;
}
break;
case EXACTF:
m = STRING(c);
ln = STR_LEN(c);
c1 = *(U8*)m;
c2 = PL_fold[c1];
goto do_exactf;
case EXACTFL:
m = STRING(c);
ln = STR_LEN(c);
c1 = *(U8*)m;
c2 = PL_fold_locale[c1];
do_exactf:
e = strend - ln;
if (norun && e < s)
e = s;
if (c1 == c2) {
while (s <= e) {
if ( *(U8*)s == c1
&& (ln == 1 || !(OP(c) == EXACTF
? ibcmp(s, m, ln)
: ibcmp_locale(s, m, ln)))
&& (norun || regtry(prog, s)) )
goto got_it;
s++;
}
} else {
while (s <= e) {
if ( (*(U8*)s == c1 || *(U8*)s == c2)
&& (ln == 1 || !(OP(c) == EXACTF
? ibcmp(s, m, ln)
: ibcmp_locale(s, m, ln)))
&& (norun || regtry(prog, s)) )
goto got_it;
s++;
}
}
break;
case BOUNDL:
PL_reg_flags |= RF_tainted;
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;
}
s++;
}
if ((!prog->minlen && tmp) && (norun || regtry(prog, s)))
goto got_it;
break;
case BOUNDLUTF8:
PL_reg_flags |= RF_tainted;
case BOUNDUTF8:
tmp = (I32)(s != startpos) ? utf8_to_uv(reghop((U8*)s, -1), 0) : '\n';
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;
}
s += UTF8SKIP(s);
}
if ((!prog->minlen && tmp) && (norun || regtry(prog, s)))
goto got_it;
break;
case NBOUNDL:
PL_reg_flags |= RF_tainted;
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 ((!prog->minlen && !tmp) && (norun || regtry(prog, s)))
goto got_it;
break;
case NBOUNDLUTF8:
PL_reg_flags |= RF_tainted;
case NBOUNDUTF8:
tmp = (I32)(s != startpos) ? utf8_to_uv(reghop((U8*)s, -1), 0) : '\n';
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);
}
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;
else
tmp = doevery;
}
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
tmp = doevery;
}
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;
else
tmp = doevery;
}
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
tmp = doevery;
}
else
tmp = 1;
s += UTF8SKIP(s);
}
break;
case NALNUM:
while (s < strend) {
if (!isALNUM(*s)) {
if (tmp && (norun || regtry(prog, s)))
goto got_it;
else
tmp = doevery;
}
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
tmp = doevery;
}
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;
else
tmp = doevery;
}
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
tmp = doevery;
}
else
tmp = 1;
s += UTF8SKIP(s);
}
break;
case SPACE:
while (s < strend) {
if (isSPACE(*s)) {
if (tmp && (norun || regtry(prog, s)))
goto got_it;
else
tmp = doevery;
}
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
tmp = doevery;
}
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;
else
tmp = doevery;
}
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
tmp = doevery;
}
else
tmp = 1;
s += UTF8SKIP(s);
}
break;
case NSPACE:
while (s < strend) {
if (!isSPACE(*s)) {
if (tmp && (norun || regtry(prog, s)))
goto got_it;
else
tmp = doevery;
}
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
tmp = doevery;
}
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;
else
tmp = doevery;
}
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
tmp = doevery;
}
else
tmp = 1;
s += UTF8SKIP(s);
}
break;
case DIGIT:
while (s < strend) {
if (isDIGIT(*s)) {
if (tmp && (norun || regtry(prog, s)))
goto got_it;
else
tmp = doevery;
}
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
tmp = doevery;
}
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;
else
tmp = doevery;
}
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
tmp = doevery;
}
else
tmp = 1;
s += UTF8SKIP(s);
}
break;
case NDIGIT:
while (s < strend) {
if (!isDIGIT(*s)) {
if (tmp && (norun || regtry(prog, s)))
goto got_it;
else
tmp = doevery;
}
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
tmp = doevery;
}
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;
else
tmp = doevery;
}
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
tmp = doevery;
}
else
tmp = 1;
s += UTF8SKIP(s);
}
break;
default:
Perl_croak(aTHX_ "panic: unknown regstclass %d", (int)OP(c));
break;
}
return 0;
got_it:
return s;
}
I32
Perl_regexec_flags(pTHX_ register regexp *prog, char *stringarg, register char *strend,
char *strbeg, I32 minend, SV *sv, void *data, U32 flags)
{
dTHR;
register char *s;
register regnode *c;
register char *startpos = stringarg;
I32 minlen;
I32 dontbother = 0;
I32 end_shift = 0;
I32 scream_pos = -1;
char *scream_olds;
SV* oreplsv = GvSV(PL_replgv);
PL_regcc = 0;
cache_re(prog);
#ifdef DEBUGGING
PL_regnarrate = PL_debug & 512;
#endif
if (prog == NULL || startpos == NULL) {
Perl_croak(aTHX_ "NULL regexp parameter");
return 0;
}
minlen = prog->minlen;
if (strend - startpos < minlen) goto phooey;
if (startpos == strbeg)
PL_regprev = '\n';
else {
PL_regprev = (U32)stringarg[-1];
if (!PL_multiline && PL_regprev == '\n')
PL_regprev = '\0';
}
if (UCHARAT(prog->program) != REG_MAGIC) {
Perl_croak(aTHX_ "corrupted regexp program");
}
PL_reg_flags = 0;
PL_reg_eval_set = 0;
PL_reg_maxiter = 0;
if (prog->reganch & ROPT_UTF8)
PL_reg_flags |= RF_utf8;
PL_regbol = startpos;
PL_bostr = strbeg;
PL_reg_sv = sv;
PL_regeol = strend;
PL_regtill = startpos+minend;
PL_reg_call_cc = 0;
s = startpos;
if (prog->reganch & ROPT_GPOS_SEEN) {
MAGIC *mg;
if (flags & REXEC_IGNOREPOS)
PL_reg_ganch = startpos;
else if (sv && SvTYPE(sv) >= SVt_PVMG
&& SvMAGIC(sv)
&& (mg = mg_find(sv, 'g')) && mg->mg_len >= 0) {
PL_reg_ganch = strbeg + mg->mg_len;
if (prog->reganch & ROPT_ANCH_GPOS) {
if (s > PL_reg_ganch)
goto phooey;
s = PL_reg_ganch;
}
}
else
PL_reg_ganch = strbeg;
}
if (!(flags & REXEC_CHECKED) && prog->check_substr != Nullsv) {
re_scream_pos_data d;
d.scream_olds = &scream_olds;
d.scream_pos = &scream_pos;
s = re_intuit_start(prog, sv, s, strend, flags, &d);
if (!s)
goto phooey;
}
DEBUG_r( if (!PL_colorset) reginitcolors() );
DEBUG_r(PerlIO_printf(Perl_debug_log,
"%sMatching REx%s `%s%.60s%s%s' against `%s%.*s%s%s'\n",
PL_colors[4],PL_colors[5],PL_colors[0],
prog->precomp,
PL_colors[1],
(strlen(prog->precomp) > 60 ? "..." : ""),
PL_colors[0],
(int)(strend - startpos > 60 ? 60 : strend - startpos),
startpos, PL_colors[1],
(strend - startpos > 60 ? "..." : ""))
);
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)
|| (prog->reganch & ROPT_ANCH_MBOL))
{
char *end;
if (minlen)
dontbother = minlen - 1;
end = HOPc(strend, -dontbother) - 1;
if (prog->check_substr) {
if (s == startpos)
goto after_try;
while (1) {
if (regtry(prog, s))
goto got_it;
after_try:
if (s >= end)
goto phooey;
if (prog->reganch & RE_USE_INTUIT) {
s = re_intuit_start(prog, sv, s + 1, strend, flags, NULL);
if (!s)
goto phooey;
}
else
s++;
}
} else {
if (s > startpos)
s--;
while (s < end) {
if (*s++ == '\n') {
if (regtry(prog, s))
goto got_it;
}
}
}
}
goto phooey;
} else if (prog->reganch & ROPT_ANCH_GPOS) {
if (regtry(prog, PL_reg_ganch))
goto got_it;
goto phooey;
}
if (prog->anchored_substr && prog->reganch & ROPT_SKIP) {
char ch = SvPVX(prog->anchored_substr)[0];
if (UTF) {
while (s < strend) {
if (*s == ch) {
if (regtry(prog, s)) goto got_it;
s += UTF8SKIP(s);
while (s < strend && *s == ch)
s += UTF8SKIP(s);
}
s += UTF8SKIP(s);
}
}
else {
while (s < strend) {
if (*s == ch) {
if (regtry(prog, s)) goto got_it;
s++;
while (s < strend && *s == ch)
s++;
}
s++;
}
}
}
else if (prog->anchored_substr != Nullsv
|| (prog->float_substr != Nullsv
&& prog->float_max_offset < strend - s)) {
SV *must = prog->anchored_substr
? prog->anchored_substr : prog->float_substr;
I32 back_max =
prog->anchored_substr ? prog->anchored_offset : prog->float_max_offset;
I32 back_min =
prog->anchored_substr ? prog->anchored_offset : prog->float_min_offset;
char *last = HOPc(strend,
-(I32)(CHR_SVLEN(must)
- (SvTAIL(must) != 0) + back_min));
char *last1;
if (s > PL_bostr)
last1 = HOPc(s, -1);
else
last1 = s - 1;
scream_pos = -1;
dontbother = end_shift;
strend = HOPc(strend, -dontbother);
while ( (s <= last) &&
((flags & REXEC_SCREAM)
? (s = screaminstr(sv, must, HOPc(s, back_min) - strbeg,
end_shift, &scream_pos, 0))
: (s = fbm_instr((unsigned char*)HOP(s, back_min),
(unsigned char*)strend, must,
PL_multiline ? FBMrf_MULTILINE : 0))) ) {
if (HOPc(s, -back_max) > last1) {
last1 = HOPc(s, -back_min);
s = HOPc(s, -back_max);
}
else {
char *t = (last1 >= PL_bostr) ? HOPc(last1, 1) : last1 + 1;
last1 = HOPc(s, -back_min);
s = t;
}
if (UTF) {
while (s <= last1) {
if (regtry(prog, s))
goto got_it;
s += UTF8SKIP(s);
}
}
else {
while (s <= last1) {
if (regtry(prog, s))
goto got_it;
s++;
}
}
}
goto phooey;
}
else if ((c = prog->regstclass)) {
if (minlen && PL_regkind[(U8)OP(prog->regstclass)] != EXACT)
strend = HOPc(strend, -(minlen - 1));
if (find_byclass(prog, c, s, strend, startpos, 0))
goto got_it;
}
else {
dontbother = 0;
if (prog->float_substr != Nullsv) {
char *last;
if (flags & REXEC_SCREAM) {
last = screaminstr(sv, prog->float_substr, s - strbeg,
end_shift, &scream_pos, 1);
if (!last)
last = scream_olds;
}
else {
STRLEN len;
char *little = SvPV(prog->float_substr, len);
if (SvTAIL(prog->float_substr)) {
if (memEQ(strend - len + 1, little, len - 1))
last = strend - len + 1;
else if (!PL_multiline)
last = memEQ(strend - len, little, len)
? strend - len : Nullch;
else
goto find_last;
} else {
find_last:
if (len)
last = rninstr(s, strend, little, little + len);
else
last = strend;
}
}
if (last == NULL) goto phooey;
dontbother = strend - last + prog->float_min_offset;
}
if (minlen && (dontbother < minlen))
dontbother = minlen - 1;
strend -= dontbother;
if (UTF) {
for (;;) {
if (regtry(prog, s))
goto got_it;
if (s >= strend)
break;
s += UTF8SKIP(s);
};
}
else {
do {
if (regtry(prog, s))
goto got_it;
} while (s++ < strend);
}
}
goto phooey;
got_it:
RX_MATCH_TAINTED_set(prog, PL_reg_flags & RF_tainted);
if (PL_reg_eval_set) {
if (oreplsv != GvSV(PL_replgv))
sv_setsv(oreplsv, GvSV(PL_replgv));
restore_pos(aTHXo_ 0);
}
if ( !(flags & REXEC_NOT_FIRST) ) {
if (RX_MATCH_COPIED(prog)) {
Safefree(prog->subbeg);
RX_MATCH_COPIED_off(prog);
}
if (flags & REXEC_COPY_STR) {
I32 i = PL_regeol - startpos + (stringarg - strbeg);
s = savepvn(strbeg, i);
prog->subbeg = s;
prog->sublen = i;
RX_MATCH_COPIED_on(prog);
}
else {
prog->subbeg = strbeg;
prog->sublen = PL_regeol - strbeg;
}
}
return 1;
phooey:
if (PL_reg_eval_set)
restore_pos(aTHXo_ 0);
return 0;
}
STATIC I32
S_regtry(pTHX_ regexp *prog, char *startpos)
{
dTHR;
register I32 i;
register I32 *sp;
register I32 *ep;
CHECKPOINT lastcp;
if ((prog->reganch & ROPT_EVAL_SEEN) && !PL_reg_eval_set) {
MAGIC *mg;
PL_reg_eval_set = RS_init;
DEBUG_r(DEBUG_s(
PerlIO_printf(Perl_debug_log, " setting stack tmpbase at %"IVdf"\n",
(IV)(PL_stack_sp - PL_stack_base));
));
SAVEI32(cxstack[cxstack_ix].blk_oldsp);
cxstack[cxstack_ix].blk_oldsp = PL_stack_sp - PL_stack_base;
SAVETMPS;
if (PL_reg_sv) {
if (PL_reg_sv != DEFSV) {
SAVESPTR(DEFSV);
DEFSV = PL_reg_sv;
}
if (!(SvTYPE(PL_reg_sv) >= SVt_PVMG && SvMAGIC(PL_reg_sv)
&& (mg = mg_find(PL_reg_sv, 'g')))) {
sv_magic(PL_reg_sv, (SV*)0, 'g', Nullch, 0);
mg = mg_find(PL_reg_sv, 'g');
mg->mg_len = -1;
}
PL_reg_magic = mg;
PL_reg_oldpos = mg->mg_len;
SAVEDESTRUCTOR_X(restore_pos, 0);
}
if (!PL_reg_curpm)
Newz(22,PL_reg_curpm, 1, PMOP);
PL_reg_curpm->op_pmregexp = prog;
PL_reg_oldcurpm = PL_curpm;
PL_curpm = PL_reg_curpm;
if (RX_MATCH_COPIED(prog)) {
PL_reg_oldsaved = prog->subbeg;
PL_reg_oldsavedlen = prog->sublen;
RX_MATCH_COPIED_off(prog);
}
else
PL_reg_oldsaved = Nullch;
prog->subbeg = PL_bostr;
prog->sublen = PL_regeol - PL_bostr;
}
prog->startp[0] = startpos - PL_bostr;
PL_reginput = startpos;
PL_regstartp = prog->startp;
PL_regendp = prog->endp;
PL_reglastparen = &prog->lastparen;
prog->lastparen = 0;
PL_regsize = 0;
DEBUG_r(PL_reg_starttry = startpos);
if (PL_reg_start_tmpl <= prog->nparens) {
PL_reg_start_tmpl = prog->nparens*3/2 + 3;
if(PL_reg_start_tmp)
Renew(PL_reg_start_tmp, PL_reg_start_tmpl, char*);
else
New(22,PL_reg_start_tmp, PL_reg_start_tmpl, char*);
}
sp = prog->startp;
ep = prog->endp;
if (prog->nparens) {
for (i = prog->nparens; i >= 1; i--) {
*++sp = -1;
*++ep = -1;
}
}
REGCP_SET;
if (regmatch(prog->program + 1)) {
prog->endp[0] = PL_reginput - PL_bostr;
return 1;
}
REGCP_UNWIND;
return 0;
}
STATIC I32
S_regmatch(pTHX_ regnode *prog)
{
dTHR;
register regnode *scan;
regnode *next;
regnode *inner;
register I32 nextchr;
register I32 n;
register I32 ln;
register char *s;
register char *locinput = PL_reginput;
register I32 c1, c2, paren;
int minmod = 0, sw = 0, logical = 0;
#ifdef DEBUGGING
PL_regindent++;
#endif
nextchr = UCHARAT(locinput);
scan = prog;
while (scan != NULL) {
#define sayNO_L (logical ? (logical = 0, sw = 0, goto cont) : sayNO)
#ifdef DEBUGGING
# 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];
int taill = (docolor ? 10 : 7);
int l = (PL_regeol - locinput > taill ? taill : PL_regeol - locinput);
int pref_len = (locinput - PL_bostr > (5 + taill) - l
? (5 + taill) - l : locinput - PL_bostr);
int pref0_len = pref_len - (locinput - PL_reg_starttry);
if (l + pref_len < (5 + taill) && l < PL_regeol - locinput)
l = ( PL_regeol - locinput > (5 + taill) - pref_len
? (5 + taill) - pref_len : PL_regeol - locinput);
if (pref0_len < 0)
pref0_len = 0;
if (pref0_len > pref_len)
pref0_len = pref_len;
regprop(prop, scan);
PerlIO_printf(Perl_debug_log,
"%4"IVdf" <%s%.*s%s%s%.*s%s%s%s%.*s%s>%*s|%3"IVdf":%*s%s\n",
(IV)(locinput - PL_bostr),
PL_colors[4], pref0_len,
locinput - pref_len, PL_colors[5],
PL_colors[2], pref_len - pref0_len,
locinput - pref_len + pref0_len, PL_colors[3],
(docolor ? "" : "> <"),
PL_colors[0], l, locinput, PL_colors[1],
15 - l - pref_len + 1,
"",
(IV)(scan - PL_regprogram), PL_regindent*2, "",
SvPVX(prop));
} );
next = scan + NEXT_OFF(scan);
if (next == scan)
next = NULL;
switch (OP(scan)) {
case BOL:
if (locinput == PL_bostr
? PL_regprev == '\n'
: (PL_multiline &&
(nextchr || locinput < PL_regeol) && locinput[-1] == '\n') )
{
break;
}
sayNO;
case MBOL:
if (locinput == PL_bostr
? PL_regprev == '\n'
: ((nextchr || locinput < PL_regeol) && locinput[-1] == '\n') )
{
break;
}
sayNO;
case SBOL:
if (locinput == PL_regbol && PL_regprev == '\n')
break;
sayNO;
case GPOS:
if (locinput == PL_reg_ganch)
break;
sayNO;
case EOL:
if (PL_multiline)
goto meol;
else
goto seol;
case MEOL:
meol:
if ((nextchr || locinput < PL_regeol) && nextchr != '\n')
sayNO;
break;
case SEOL:
seol:
if ((nextchr || locinput < PL_regeol) && nextchr != '\n')
sayNO;
if (PL_regeol - locinput > 1)
sayNO;
break;
case EOS:
if (PL_regeol != locinput)
sayNO;
break;
case SANYUTF8:
if (nextchr & 0x80) {
locinput += PL_utf8skip[nextchr];
if (locinput > PL_regeol)
sayNO;
nextchr = UCHARAT(locinput);
break;
}
if (!nextchr && locinput >= PL_regeol)
sayNO;
nextchr = UCHARAT(++locinput);
break;
case SANY:
if (!nextchr && locinput >= PL_regeol)
sayNO;
nextchr = UCHARAT(++locinput);
break;
case ANYUTF8:
if (nextchr & 0x80) {
locinput += PL_utf8skip[nextchr];
if (locinput > PL_regeol)
sayNO;
nextchr = UCHARAT(locinput);
break;
}
if ((!nextchr && locinput >= PL_regeol) || nextchr == '\n')
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);
if (UCHARAT(s) != nextchr)
sayNO;
if (PL_regeol - locinput < ln)
sayNO;
if (ln > 1 && memNE(s, locinput, ln))
sayNO;
locinput += ln;
nextchr = UCHARAT(locinput);
break;
case EXACTFL:
PL_reg_flags |= RF_tainted;
case EXACTF:
s = STRING(scan);
ln = STR_LEN(scan);
if (UTF) {
char *l = locinput;
char *e = s + ln;
c1 = OP(scan) == EXACTF;
while (s < e) {
if (l >= PL_regeol)
sayNO;
if (utf8_to_uv((U8*)s, 0) != (c1 ?
toLOWER_utf8((U8*)l) :
toLOWER_LC_utf8((U8*)l)))
{
sayNO;
}
s += UTF8SKIP(s);
l += UTF8SKIP(l);
}
locinput = l;
nextchr = UCHARAT(locinput);
break;
}
if (UCHARAT(s) != nextchr &&
UCHARAT(s) != ((OP(scan) == EXACTF)
? PL_fold : PL_fold_locale)[nextchr])
sayNO;
if (PL_regeol - locinput < ln)
sayNO;
if (ln > 1 && (OP(scan) == EXACTF
? ibcmp(s, locinput, ln)
: ibcmp_locale(s, locinput, ln)))
sayNO;
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)
nextchr = UCHARAT(locinput);
if (!REGINCLASS(scan, nextchr))
sayNO;
if (!nextchr && locinput >= PL_regeol)
sayNO;
nextchr = UCHARAT(++locinput);
break;
case ALNUML:
PL_reg_flags |= RF_tainted;
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;
case ALNUMUTF8:
if (!nextchr)
sayNO;
if (nextchr & 0x80) {
if (!(OP(scan) == ALNUMUTF8
? swash_fetch(PL_utf8_alnum, (U8*)locinput)
: isALNUM_LC_utf8((U8*)locinput)))
{
sayNO;
}
locinput += PL_utf8skip[nextchr];
nextchr = UCHARAT(locinput);
break;
}
if (!(OP(scan) == ALNUMUTF8
? isALNUM(nextchr) : isALNUM_LC(nextchr)))
sayNO;
nextchr = UCHARAT(++locinput);
break;
case NALNUML:
PL_reg_flags |= RF_tainted;
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;
case NALNUMUTF8:
if (!nextchr && locinput >= PL_regeol)
sayNO;
if (nextchr & 0x80) {
if (OP(scan) == NALNUMUTF8
? swash_fetch(PL_utf8_alnum, (U8*)locinput)
: isALNUM_LC_utf8((U8*)locinput))
{
sayNO;
}
locinput += PL_utf8skip[nextchr];
nextchr = UCHARAT(locinput);
break;
}
if (OP(scan) == NALNUMUTF8
? isALNUM(nextchr) : isALNUM_LC(nextchr))
sayNO;
nextchr = UCHARAT(++locinput);
break;
case BOUNDL:
case NBOUNDL:
PL_reg_flags |= RF_tainted;
case BOUND:
case NBOUND:
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;
case BOUNDUTF8:
case NBOUNDUTF8:
ln = (locinput != PL_regbol)
? utf8_to_uv(reghop((U8*)locinput, -1), 0) : PL_regprev;
if (OP(scan) == BOUNDUTF8 || OP(scan) == NBOUNDUTF8) {
ln = isALNUM_uni(ln);
n = swash_fetch(PL_utf8_alnum, (U8*)locinput);
}
else {
ln = isALNUM_LC_uni(ln);
n = isALNUM_LC_utf8((U8*)locinput);
}
if (((!ln) == (!n)) == (OP(scan) == BOUNDUTF8 || OP(scan) == BOUNDLUTF8))
sayNO;
break;
case SPACEL:
PL_reg_flags |= RF_tainted;
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;
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;
}
locinput += PL_utf8skip[nextchr];
nextchr = UCHARAT(locinput);
break;
}
if (!(OP(scan) == SPACEUTF8
? isSPACE(nextchr) : isSPACE_LC(nextchr)))
sayNO;
nextchr = UCHARAT(++locinput);
break;
case NSPACEL:
PL_reg_flags |= RF_tainted;
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;
case NSPACEUTF8:
if (!nextchr && locinput >= PL_regeol)
sayNO;
if (nextchr & 0x80) {
if (OP(scan) == NSPACEUTF8
? swash_fetch(PL_utf8_space, (U8*)locinput)
: isSPACE_LC_utf8((U8*)locinput))
{
sayNO;
}
locinput += PL_utf8skip[nextchr];
nextchr = UCHARAT(locinput);
break;
}
if (OP(scan) == NSPACEUTF8
? isSPACE(nextchr) : isSPACE_LC(nextchr))
sayNO;
nextchr = UCHARAT(++locinput);
break;
case DIGITL:
PL_reg_flags |= RF_tainted;
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;
case DIGITUTF8:
if (!nextchr)
sayNO;
if (nextchr & 0x80) {
if (!(OP(scan) == DIGITUTF8
? swash_fetch(PL_utf8_digit, (U8*)locinput)
: isDIGIT_LC_utf8((U8*)locinput)))
{
sayNO;
}
locinput += PL_utf8skip[nextchr];
nextchr = UCHARAT(locinput);
break;
}
if (!(OP(scan) == DIGITUTF8
? isDIGIT(nextchr) : isDIGIT_LC(nextchr)))
sayNO;
nextchr = UCHARAT(++locinput);
break;
case NDIGITL:
PL_reg_flags |= RF_tainted;
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;
case NDIGITUTF8:
if (!nextchr && locinput >= PL_regeol)
sayNO;
if (nextchr & 0x80) {
if (OP(scan) == NDIGITUTF8
? swash_fetch(PL_utf8_digit, (U8*)locinput)
: isDIGIT_LC_utf8((U8*)locinput))
{
sayNO;
}
locinput += PL_utf8skip[nextchr];
nextchr = UCHARAT(locinput);
break;
}
if (OP(scan) == NDIGITUTF8
? isDIGIT(nextchr) : isDIGIT_LC(nextchr))
sayNO;
nextchr = UCHARAT(++locinput);
break;
case CLUMP:
if (locinput >= PL_regeol || swash_fetch(PL_utf8_mark,(U8*)locinput))
sayNO;
locinput += PL_utf8skip[nextchr];
while (locinput < PL_regeol && swash_fetch(PL_utf8_mark,(U8*)locinput))
locinput += UTF8SKIP(locinput);
if (locinput > PL_regeol)
sayNO;
nextchr = UCHARAT(locinput);
break;
case REFFL:
PL_reg_flags |= RF_tainted;
case REF:
case REFF:
n = ARG(scan);
ln = PL_regstartp[n];
PL_reg_leftiter = PL_reg_maxiter;
if (*PL_reglastparen < n || ln == -1)
sayNO;
if (ln == PL_regendp[n])
break;
s = PL_bostr + ln;
if (UTF && OP(scan) != REF) {
char *l = locinput;
char *e = PL_bostr + PL_regendp[n];
if (OP(scan) == REFF) {
while (s < e) {
if (l >= PL_regeol)
sayNO;
if (toLOWER_utf8((U8*)s) != toLOWER_utf8((U8*)l))
sayNO;
s += UTF8SKIP(s);
l += UTF8SKIP(l);
}
}
else {
while (s < e) {
if (l >= PL_regeol)
sayNO;
if (toLOWER_LC_utf8((U8*)s) != toLOWER_LC_utf8((U8*)l))
sayNO;
s += UTF8SKIP(s);
l += UTF8SKIP(l);
}
}
locinput = l;
nextchr = UCHARAT(locinput);
break;
}
if (UCHARAT(s) != nextchr &&
(OP(scan) == REF ||
(UCHARAT(s) != ((OP(scan) == REFF
? PL_fold : PL_fold_locale)[nextchr]))))
sayNO;
ln = PL_regendp[n] - ln;
if (locinput + ln > PL_regeol)
sayNO;
if (ln > 1 && (OP(scan) == REF
? memNE(s, locinput, ln)
: (OP(scan) == REFF
? ibcmp(s, locinput, ln)
: ibcmp_locale(s, locinput, ln))))
sayNO;
locinput += ln;
nextchr = UCHARAT(locinput);
break;
case NOTHING:
case TAIL:
break;
case BACK:
break;
case EVAL:
{
dSP;
OP_4tree *oop = PL_op;
COP *ocurcop = PL_curcop;
SV **ocurpad = PL_curpad;
SV *ret;
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]);
PL_regendp[0] = PL_reg_magic->mg_len = locinput - PL_bostr;
CALLRUNOPS(aTHX);
SPAGAIN;
ret = POPs;
PUTBACK;
PL_op = oop;
PL_curpad = ocurpad;
PL_curcop = ocurcop;
if (logical) {
if (logical == 2) {
regexp *re;
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, 'r');
}
if (mg) {
re = (regexp *)mg->mg_obj;
(void)ReREFCNT_inc(re);
}
else {
STRLEN len;
char *t = SvPV(ret, len);
PMOP pm;
char *oprecomp = PL_regprecomp;
I32 osize = PL_regsize;
I32 onpar = PL_regnpar;
pm.op_pmflags = 0;
pm.op_pmdynflags = (UTF||DO_UTF8(ret) ? PMdf_UTF8 : 0);
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);
PL_regprecomp = oprecomp;
PL_regsize = osize;
PL_regnpar = onpar;
}
DEBUG_r(
PerlIO_printf(Perl_debug_log,
"Entering embedded `%s%.60s%s%s'\n",
PL_colors[0],
re->precomp,
PL_colors[1],
(strlen(re->precomp) > 60 ? "..." : ""))
);
state.node = next;
state.prev = PL_reg_call_cc;
state.cc = PL_regcc;
state.re = PL_reg_re;
PL_regcc = 0;
cp = regcppush(0);
REGCP_SET;
cache_re(re);
state.ss = PL_savestack_ix;
*PL_reglastparen = 0;
PL_reg_call_cc = &state;
PL_reginput = locinput;
PL_reg_maxiter = 0;
if (regmatch(re->program + 1)) {
PL_reg_call_cc = state.prev;
PL_regcc = state.cc;
PL_reg_re = state.re;
cache_re(PL_reg_re);
PL_reg_maxiter = 0;
ReREFCNT_dec(re);
regcpblow(cp);
sayYES;
}
ReREFCNT_dec(re);
REGCP_UNWIND;
regcppop();
PL_reg_call_cc = state.prev;
PL_regcc = state.cc;
PL_reg_re = state.re;
cache_re(PL_reg_re);
PL_reg_maxiter = 0;
sayNO;
}
sw = SvTRUE(ret);
logical = 0;
}
else
sv_setsv(save_scalar(PL_replgv), ret);
break;
}
case OPEN:
n = ARG(scan);
PL_reg_start_tmp[n] = locinput;
if (n > PL_regsize)
PL_regsize = n;
break;
case CLOSE:
n = ARG(scan);
PL_regstartp[n] = PL_reg_start_tmp[n] - PL_bostr;
PL_regendp[n] = locinput - PL_bostr;
if (n > *PL_reglastparen)
*PL_reglastparen = n;
break;
case GROUPP:
n = ARG(scan);
sw = (*PL_reglastparen >= n && PL_regendp[n] != -1);
break;
case IFTHEN:
PL_reg_leftiter = PL_reg_maxiter;
if (sw)
next = NEXTOPER(NEXTOPER(scan));
else {
next = scan + ARG(scan);
if (OP(next) == IFTHEN)
next = NEXTOPER(NEXTOPER(next));
}
break;
case LOGICAL:
logical = scan->flags;
break;
case CURLYX: {
CURCUR cc;
CHECKPOINT cp = PL_savestack_ix;
if (OP(PREVOPER(next)) == NOTHING)
next += ARG(next);
cc.oldcc = PL_regcc;
PL_regcc = &cc;
cc.parenfloor = *PL_reglastparen;
cc.cur = -1;
cc.min = ARG1(scan);
cc.max = ARG2(scan);
cc.scan = NEXTOPER(scan) + EXTRA_STEP_2ARGS;
cc.next = next;
cc.minmod = minmod;
cc.lastloc = 0;
PL_reginput = locinput;
n = regmatch(PREVOPER(next));
regcpblow(cp);
PL_regcc = cc.oldcc;
saySAME(n);
}
case WHILEM: {
CHECKPOINT cp, lastcp;
CURCUR* cc = PL_regcc;
char *lastloc = cc->lastloc;
n = cc->cur + 1;
PL_reginput = locinput;
DEBUG_r(
PerlIO_printf(Perl_debug_log,
"%*s %ld out of %ld..%ld cc=%lx\n",
REPORT_CODE_OFF+PL_regindent*2, "",
(long)n, (long)cc->min,
(long)cc->max, (long)cc)
);
if (locinput == cc->lastloc && n >= cc->min) {
PL_regcc = cc->oldcc;
if (PL_regcc)
ln = PL_regcc->cur;
DEBUG_r(
PerlIO_printf(Perl_debug_log,
"%*s empty match detected, try continuation...\n",
REPORT_CODE_OFF+PL_regindent*2, "")
);
if (regmatch(cc->next))
sayYES;
if (PL_regcc)
PL_regcc->cur = ln;
PL_regcc = cc;
sayNO;
}
if (n < cc->min) {
cc->cur = n;
cc->lastloc = locinput;
if (regmatch(cc->scan))
sayYES;
cc->cur = n - 1;
cc->lastloc = lastloc;
sayNO;
}
if (scan->flags) {
if (!PL_reg_maxiter) {
PL_reg_maxiter = (PL_regeol - PL_bostr + 1) * (scan->flags>>4);
PL_reg_leftiter = PL_reg_maxiter;
}
if (PL_reg_leftiter-- == 0) {
I32 size = (PL_reg_maxiter + 7)/8;
if (PL_reg_poscache) {
if (PL_reg_poscache_size < size) {
Renew(PL_reg_poscache, size, char);
PL_reg_poscache_size = size;
}
Zero(PL_reg_poscache, size, char);
}
else {
PL_reg_poscache_size = size;
Newz(29, PL_reg_poscache, size, char);
}
DEBUG_r(
PerlIO_printf(Perl_debug_log,
"%sDetected a super-linear match, switching on caching%s...\n",
PL_colors[4], PL_colors[5])
);
}
if (PL_reg_leftiter < 0) {
I32 o = locinput - PL_bostr, b;
o = (scan->flags & 0xf) - 1 + o * (scan->flags>>4);
b = o % 8;
o /= 8;
if (PL_reg_poscache[o] & (1<<b)) {
DEBUG_r(
PerlIO_printf(Perl_debug_log,
"%*s already tried at this position...\n",
REPORT_CODE_OFF+PL_regindent*2, "")
);
sayNO_SILENT;
}
PL_reg_poscache[o] |= (1<<b);
}
}
if (cc->minmod) {
PL_regcc = cc->oldcc;
if (PL_regcc)
ln = PL_regcc->cur;
cp = regcppush(cc->parenfloor);
REGCP_SET;
if (regmatch(cc->next)) {
regcpblow(cp);
sayYES;
}
REGCP_UNWIND;
regcppop();
if (PL_regcc)
PL_regcc->cur = ln;
PL_regcc = cc;
if (n >= cc->max) {
if (ckWARN(WARN_REGEXP) && n >= REG_INFTY
&& !(PL_reg_flags & RF_warned)) {
PL_reg_flags |= RF_warned;
Perl_warner(aTHX_ WARN_REGEXP, "%s limit (%d) exceeded",
"Complex regular subexpression recursion",
REG_INFTY - 1);
}
sayNO;
}
DEBUG_r(
PerlIO_printf(Perl_debug_log,
"%*s trying longer...\n",
REPORT_CODE_OFF+PL_regindent*2, "")
);
PL_reginput = locinput;
cc->cur = n;
cc->lastloc = locinput;
cp = regcppush(cc->parenfloor);
REGCP_SET;
if (regmatch(cc->scan)) {
regcpblow(cp);
sayYES;
}
REGCP_UNWIND;
regcppop();
cc->cur = n - 1;
cc->lastloc = lastloc;
sayNO;
}
if (n < cc->max) {
cp = regcppush(cc->parenfloor);
cc->cur = n;
cc->lastloc = locinput;
REGCP_SET;
if (regmatch(cc->scan)) {
regcpblow(cp);
sayYES;
}
REGCP_UNWIND;
regcppop();
PL_reginput = locinput;
DEBUG_r(
PerlIO_printf(Perl_debug_log,
"%*s failed, try continuation...\n",
REPORT_CODE_OFF+PL_regindent*2, "")
);
}
if (ckWARN(WARN_REGEXP) && n >= REG_INFTY
&& !(PL_reg_flags & RF_warned)) {
PL_reg_flags |= RF_warned;
Perl_warner(aTHX_ WARN_REGEXP, "%s limit (%d) exceeded",
"Complex regular subexpression recursion",
REG_INFTY - 1);
}
PL_regcc = cc->oldcc;
if (PL_regcc)
ln = PL_regcc->cur;
if (regmatch(cc->next))
sayYES;
if (PL_regcc)
PL_regcc->cur = ln;
PL_regcc = cc;
cc->cur = n - 1;
cc->lastloc = lastloc;
sayNO;
}
case BRANCHJ:
next = scan + ARG(scan);
if (next == scan)
next = NULL;
inner = NEXTOPER(NEXTOPER(scan));
goto do_branch;
case BRANCH:
inner = NEXTOPER(scan);
do_branch:
{
CHECKPOINT lastcp;
c1 = OP(scan);
if (OP(next) != c1)
next = inner;
else {
int lastparen = *PL_reglastparen;
REGCP_SET;
do {
PL_reginput = locinput;
if (regmatch(inner))
sayYES;
REGCP_UNWIND;
for (n = *PL_reglastparen; n > lastparen; n--)
PL_regendp[n] = -1;
*PL_reglastparen = n;
scan = next;
if ((n = (c1 == BRANCH ? NEXT_OFF(next) : ARG(next))))
next += n;
else
next = NULL;
inner = NEXTOPER(scan);
if (c1 == BRANCHJ) {
inner = NEXTOPER(inner);
}
} while (scan != NULL && OP(scan) == c1);
sayNO;
}
}
break;
case MINMOD:
minmod = 1;
break;
case CURLYM:
{
I32 l = 0;
CHECKPOINT lastcp;
ln = ARG1(scan);
n = ARG2(scan);
paren = scan->flags;
if (paren) {
if (paren > PL_regsize)
PL_regsize = paren;
if (paren > *PL_reglastparen)
*PL_reglastparen = paren;
}
scan = NEXTOPER(scan) + NODE_STEP_REGNODE;
if (paren)
scan += NEXT_OFF(scan);
PL_reginput = locinput;
if (minmod) {
minmod = 0;
if (ln && regrepeat_hard(scan, ln, &l) < ln)
sayNO;
if (ln && l == 0 && n >= ln
&& !(paren && ln == 0))
ln = n;
locinput = PL_reginput;
if (PL_regkind[(U8)OP(next)] == EXACT) {
c1 = (U8)*STRING(next);
if (OP(next) == EXACTF)
c2 = PL_fold[c1];
else if (OP(next) == EXACTFL)
c2 = PL_fold_locale[c1];
else
c2 = c1;
}
else
c1 = c2 = -1000;
REGCP_SET;
while (n >= ln || (n == REG_INFTY && ln > 0 && l)) {
if (c1 == -1000 ||
UCHARAT(PL_reginput) == c1 ||
UCHARAT(PL_reginput) == c2)
{
if (paren) {
if (n) {
PL_regstartp[paren] =
HOPc(PL_reginput, -l) - PL_bostr;
PL_regendp[paren] = PL_reginput - PL_bostr;
}
else
PL_regendp[paren] = -1;
}
if (regmatch(next))
sayYES;
REGCP_UNWIND;
}
PL_reginput = locinput;
if (regrepeat_hard(scan, 1, &l)) {
ln++;
locinput = PL_reginput;
}
else
sayNO;
}
}
else {
n = regrepeat_hard(scan, n, &l);
if (n != 0 && l == 0
&& !(paren && ln == 0))
ln = n;
locinput = PL_reginput;
DEBUG_r(
PerlIO_printf(Perl_debug_log,
"%*s matched %"IVdf" times, len=%"IVdf"...\n",
(int)(REPORT_CODE_OFF+PL_regindent*2), "",
(IV) n, (IV)l)
);
if (n >= ln) {
if (PL_regkind[(U8)OP(next)] == EXACT) {
c1 = (U8)*STRING(next);
if (OP(next) == EXACTF)
c2 = PL_fold[c1];
else if (OP(next) == EXACTFL)
c2 = PL_fold_locale[c1];
else
c2 = c1;
}
else
c1 = c2 = -1000;
}
REGCP_SET;
while (n >= ln) {
if (c1 == -1000 ||
UCHARAT(PL_reginput) == c1 ||
UCHARAT(PL_reginput) == c2)
{
DEBUG_r(
PerlIO_printf(Perl_debug_log,
"%*s trying tail with n=%"IVdf"...\n",
(int)(REPORT_CODE_OFF+PL_regindent*2), "", (IV)n)
);
if (paren) {
if (n) {
PL_regstartp[paren] = HOPc(PL_reginput, -l) - PL_bostr;
PL_regendp[paren] = PL_reginput - PL_bostr;
}
else
PL_regendp[paren] = -1;
}
if (regmatch(next))
sayYES;
REGCP_UNWIND;
}
n--;
locinput = HOPc(locinput, -l);
PL_reginput = locinput;
}
}
sayNO;
break;
}
case CURLYN:
paren = scan->flags;
if (paren > PL_regsize)
PL_regsize = paren;
if (paren > *PL_reglastparen)
*PL_reglastparen = paren;
ln = ARG1(scan);
n = ARG2(scan);
scan = regnext(NEXTOPER(scan) + NODE_STEP_REGNODE);
goto repeat;
case CURLY:
paren = 0;
ln = ARG1(scan);
n = ARG2(scan);
scan = NEXTOPER(scan) + NODE_STEP_REGNODE;
goto repeat;
case STAR:
ln = 0;
n = REG_INFTY;
scan = NEXTOPER(scan);
paren = 0;
goto repeat;
case PLUS:
ln = 1;
n = REG_INFTY;
scan = NEXTOPER(scan);
paren = 0;
repeat:
if (PL_regkind[(U8)OP(next)] == EXACT) {
c1 = (U8)*STRING(next);
if (OP(next) == EXACTF)
c2 = PL_fold[c1];
else if (OP(next) == EXACTFL)
c2 = PL_fold_locale[c1];
else
c2 = c1;
}
else
c1 = c2 = -1000;
PL_reginput = locinput;
if (minmod) {
CHECKPOINT lastcp;
minmod = 0;
if (ln && regrepeat(scan, ln) < ln)
sayNO;
locinput = PL_reginput;
REGCP_SET;
if (c1 != -1000) {
char *e = locinput + n - ln;
char *old = locinput;
if (e >= PL_regeol || (n == REG_INFTY))
e = PL_regeol - 1;
while (1) {
if (c1 == c2) {
while (locinput <= e && *locinput != c1)
locinput++;
} else {
while (locinput <= e
&& *locinput != c1
&& *locinput != c2)
locinput++;
}
if (locinput > e)
sayNO;
if (locinput != old) {
ln = 1;
if (regrepeat(scan, locinput - old) <
locinput - old)
sayNO;
}
if (paren) {
if (ln) {
PL_regstartp[paren] = HOPc(locinput, -1) - PL_bostr;
PL_regendp[paren] = locinput - PL_bostr;
}
else
PL_regendp[paren] = -1;
}
if (regmatch(next))
sayYES;
PL_reginput = locinput;
REGCP_UNWIND;
old = locinput++;
}
}
else
while (n >= ln || (n == REG_INFTY && ln > 0)) {
if (c1 == -1000 ||
UCHARAT(PL_reginput) == c1 ||
UCHARAT(PL_reginput) == c2)
{
if (paren) {
if (n) {
PL_regstartp[paren] = HOPc(PL_reginput, -1) - PL_bostr;
PL_regendp[paren] = PL_reginput - PL_bostr;
}
else
PL_regendp[paren] = -1;
}
if (regmatch(next))
sayYES;
REGCP_UNWIND;
}
PL_reginput = locinput;
if (regrepeat(scan, 1)) {
ln++;
locinput = PL_reginput;
}
else
sayNO;
}
}
else {
CHECKPOINT lastcp;
n = regrepeat(scan, n);
locinput = PL_reginput;
if (ln < n && PL_regkind[(U8)OP(next)] == EOL &&
(!PL_multiline || OP(next) == SEOL || OP(next) == EOS)) {
ln = n;
if (UCHARAT(PL_reginput - 1) == '\n' && OP(next) != EOS)
ln--;
}
REGCP_SET;
if (paren) {
while (n >= ln) {
if (c1 == -1000 ||
UCHARAT(PL_reginput) == c1 ||
UCHARAT(PL_reginput) == c2)
{
if (paren && n) {
if (n) {
PL_regstartp[paren] = HOPc(PL_reginput, -1) - PL_bostr;
PL_regendp[paren] = PL_reginput - PL_bostr;
}
else
PL_regendp[paren] = -1;
}
if (regmatch(next))
sayYES;
REGCP_UNWIND;
}
n--;
PL_reginput = locinput = HOPc(locinput, -1);
}
}
else {
while (n >= ln) {
if (c1 == -1000 ||
UCHARAT(PL_reginput) == c1 ||
UCHARAT(PL_reginput) == c2)
{
if (regmatch(next))
sayYES;
REGCP_UNWIND;
}
n--;
PL_reginput = locinput = HOPc(locinput, -1);
}
}
}
sayNO;
break;
case END:
if (PL_reg_call_cc) {
re_cc_state *cur_call_cc = PL_reg_call_cc;
CURCUR *cctmp = PL_regcc;
regexp *re = PL_reg_re;
CHECKPOINT cp, lastcp;
cp = regcppush(0);
REGCP_SET;
regcp_set_to(PL_reg_call_cc->ss);
PL_reginput = locinput;
cache_re(PL_reg_call_cc->re);
PL_regcc = PL_reg_call_cc->cc;
PL_reg_call_cc = PL_reg_call_cc->prev;
if (regmatch(cur_call_cc->node)) {
PL_reg_call_cc = cur_call_cc;
regcpblow(cp);
sayYES;
}
REGCP_UNWIND;
regcppop();
PL_reg_call_cc = cur_call_cc;
PL_regcc = cctmp;
PL_reg_re = re;
cache_re(re);
DEBUG_r(
PerlIO_printf(Perl_debug_log,
"%*s continuation failed...\n",
REPORT_CODE_OFF+PL_regindent*2, "")
);
sayNO_SILENT;
}
if (locinput < PL_regtill) {
DEBUG_r(PerlIO_printf(Perl_debug_log,
"%sMatch possible, but length=%ld is smaller than requested=%ld, failing!%s\n",
PL_colors[4],
(long)(locinput - PL_reg_starttry),
(long)(PL_regtill - PL_reg_starttry),
PL_colors[5]));
sayNO_FINAL;
}
PL_reginput = locinput;
sayYES_FINAL;
case SUCCEED:
PL_reginput = locinput;
sayYES_LOUD;
case SUSPEND:
n = 1;
PL_reginput = locinput;
goto do_ifmatch;
case UNLESSM:
n = 0;
if (scan->flags) {
if (UTF) {
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;
}
}
else
PL_reginput = locinput;
goto do_ifmatch;
case IFMATCH:
n = 1;
if (scan->flags) {
if (UTF) {
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;
}
}
else
PL_reginput = locinput;
do_ifmatch:
inner = NEXTOPER(NEXTOPER(scan));
if (regmatch(inner) != n) {
say_no:
if (logical) {
logical = 0;
sw = 0;
goto do_longjump;
}
else
sayNO;
}
say_yes:
if (logical) {
logical = 0;
sw = 1;
}
if (OP(scan) == SUSPEND) {
locinput = PL_reginput;
nextchr = UCHARAT(locinput);
}
case LONGJMP:
do_longjump:
next = scan + ARG(scan);
if (next == scan)
next = NULL;
break;
default:
PerlIO_printf(Perl_error_log, "%"UVxf" %d\n",
PTR2UV(scan), OP(scan));
Perl_croak(aTHX_ "regexp memory corruption");
}
scan = next;
}
Perl_croak(aTHX_ "corrupted regexp pointers");
sayNO;
yes_loud:
DEBUG_r(
PerlIO_printf(Perl_debug_log,
"%*s %scould match...%s\n",
REPORT_CODE_OFF+PL_regindent*2, "", PL_colors[4],PL_colors[5])
);
goto yes;
yes_final:
DEBUG_r(PerlIO_printf(Perl_debug_log, "%sMatch successful!%s\n",
PL_colors[4],PL_colors[5]));
yes:
#ifdef DEBUGGING
PL_regindent--;
#endif
return 1;
no:
DEBUG_r(
PerlIO_printf(Perl_debug_log,
"%*s %sfailed...%s\n",
REPORT_CODE_OFF+PL_regindent*2, "",PL_colors[4],PL_colors[5])
);
goto do_no;
no_final:
do_no:
#ifdef DEBUGGING
PL_regindent--;
#endif
return 0;
}
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;
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++;
}
break;
case SANYUTF8:
loceol = PL_regeol;
while (scan < loceol) {
scan += UTF8SKIP(scan);
hardcount++;
}
break;
case EXACT:
c = (U8)*STRING(p);
while (scan < loceol && UCHARAT(scan) == c)
scan++;
break;
case EXACTF:
c = (U8)*STRING(p);
while (scan < loceol &&
(UCHARAT(scan) == c || UCHARAT(scan) == PL_fold[c]))
scan++;
break;
case EXACTFL:
PL_reg_flags |= RF_tainted;
c = (U8)*STRING(p);
while (scan < loceol &&
(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++;
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++;
}
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++;
}
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++;
}
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++;
}
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++;
}
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++;
}
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++;
}
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++;
}
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++;
}
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++;
}
break;
default:
break;
}
if (hardcount)
c = hardcount;
else
c = scan - PL_reginput;
PL_reginput = scan;
DEBUG_r(
{
SV *prop = sv_newmortal();
regprop(prop, p);
PerlIO_printf(Perl_debug_log,
"%*s %s can match %"IVdf" times out of %"IVdf"...\n",
REPORT_CODE_OFF+1, "", SvPVX(prop),(IV)c,(IV)max);
});
return(c);
}
STATIC I32
S_regrepeat_hard(pTHX_ regnode *p, I32 max, I32 *lp)
{
dTHR;
register char *scan;
register char *start;
register char *loceol = PL_regeol;
I32 l = 0;
I32 count = 0, res = 1;
if (!max)
return 0;
start = PL_reginput;
if (UTF) {
while (PL_reginput < loceol && (scan = PL_reginput, res = regmatch(p))) {
if (!count++) {
l = 0;
while (start < PL_reginput) {
l++;
start += UTF8SKIP(start);
}
*lp = l;
if (l == 0)
return max;
}
if (count == max)
return count;
}
}
else {
while (PL_reginput < loceol && (scan = PL_reginput, res = regmatch(p))) {
if (!count++) {
*lp = l = PL_reginput - start;
if (max != REG_INFTY && l*max < loceol - scan)
loceol = scan + l*max;
if (l == 0)
return max;
}
}
}
if (!res)
PL_reginput = scan;
return count;
}
STATIC bool
S_reginclass(pTHX_ register regnode *p, register I32 c)
{
dTHR;
char flags = ANYOF_FLAGS(p);
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];
}
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))
)
{
match = TRUE;
}
}
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 (swash_fetch(sv, p))
match = TRUE;
else if (flags & ANYOF_FOLD) {
U8 tmpbuf[UTF8_MAXLEN];
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(sv, tmpbuf))
match = TRUE;
}
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)
s += UTF8SKIP(s);
}
else {
while (off++) {
if (s > (U8*)PL_bostr) {
s--;
if (*s & 0x80) {
while (s > (U8*)PL_bostr && (*s & 0xc0) == 0x80)
s--;
}
}
}
}
return s;
}
STATIC U8 *
S_reghopmaybe(pTHX_ U8* s, I32 off)
{
dTHR;
if (off >= 0) {
while (off-- && s < (U8*)PL_regeol)
s += UTF8SKIP(s);
if (off >= 0)
return 0;
}
else {
while (off++) {
if (s > (U8*)PL_bostr) {
s--;
if (*s & 0x80) {
while (s > (U8*)PL_bostr && (*s & 0xc0) == 0x80)
s--;
}
}
else
break;
}
if (off <= 0)
return 0;
}
return s;
}
#ifdef PERL_OBJECT
#include "XSUB.h"
#endif
static void
restore_pos(pTHXo_ void *arg)
{
dTHR;
if (PL_reg_eval_set) {
if (PL_reg_oldsaved) {
PL_reg_re->subbeg = PL_reg_oldsaved;
PL_reg_re->sublen = PL_reg_oldsavedlen;
RX_MATCH_COPIED_on(PL_reg_re);
}
PL_reg_magic->mg_len = PL_reg_oldpos;
PL_reg_eval_set = 0;
PL_curpm = PL_reg_oldcurpm;
}
}