*/
#ifdef PERL_EXT_RE_BUILD
-/* need to replace pregcomp et al, so enable that */
-# ifndef PERL_IN_XSUB_RE
-# define PERL_IN_XSUB_RE
-# endif
-/* need access to debugger hooks */
-# if defined(PERL_EXT_RE_DEBUG) && !defined(DEBUGGING)
-# define DEBUGGING
-# endif
-#endif
-
-#ifdef PERL_IN_XSUB_RE
-/* We *really* need to overwrite these symbols: */
-# 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
-/* *These* symbols are masked to allow static link. */
-# define Perl_pregexec my_pregexec
-# define Perl_reginitcolors my_reginitcolors
-# define Perl_regclass_swash my_regclass_swash
-
-# define PERL_NO_GET_CONTEXT
+#include "re_top.h"
#endif
/*
#define PERL_IN_REGEXEC_C
#include "perl.h"
-#include "regcomp.h"
+#ifdef PERL_IN_XSUB_RE
+# include "re_comp.h"
+#else
+# include "regcomp.h"
+#endif
#define RF_tainted 1 /* tainted information used? */
#define RF_warned 2 /* warned about big count? */
: (U8*)(pos + off)))
#define HOPBACKc(pos, off) ((char*) \
((PL_reg_match_utf8) \
- ? reghopmaybe3((U8*)pos, -off, ((U8*)(off < 0 ? PL_regeol : PL_bostr))) \
+ ? reghopmaybe3((U8*)pos, -off, (U8*)PL_bostr) \
: (pos - off >= PL_bostr) \
? (U8*)(pos - off) \
: (U8*)NULL) \
)
-#define reghopmaybe3_c(pos,off,lim) ((char*)reghopmaybe3((U8*)pos, off, (U8*)lim))
#define HOP3(pos,off,lim) (PL_reg_match_utf8 ? reghop3((U8*)pos, off, (U8*)lim) : (U8*)(pos + off))
#define HOP3c(pos,off,lim) ((char*)HOP3(pos,off,lim))
);
}
DEBUG_EXECUTE_r(
- if ((I32)(*PL_reglastparen + 1) <= rex->nparens) {
+ if (*PL_reglastparen + 1 <= rex->nparens) {
PerlIO_printf(Perl_debug_log,
" restoring \\%"IVdf"..\\%"IVdf" to undef\n",
(IV)(*PL_reglastparen + 1), (IV)rex->nparens);
* building DynaLoader will fail:
* "Error: '*' not in typemap in DynaLoader.xs, line 164"
* --jhi */
- for (i = *PL_reglastparen + 1; i <= rex->nparens; i++) {
+ for (i = *PL_reglastparen + 1; (U32)i <= rex->nparens; i++) {
if (i > PL_regsize)
PL_regstartp[i] = -1;
PL_regendp[i] = -1;
* pregexec and friends
*/
+#ifndef PERL_IN_XSUB_RE
/*
- pregexec - match a regexp against a string
*/
regexec_flags(prog, stringarg, strend, strbeg, minend, screamer, NULL,
nosave ? 0 : REXEC_COPY_STR);
}
-
+#endif
/*
* Need to implement the following flags for reg_anch:
sv_uni_display(dsv, sv, 60, UNI_DISPLAY_REGEX) :
strpos;
const int len = PL_reg_match_utf8 ?
- strlen(s) : strend - strpos;
+ (int)strlen(s) : strend - strpos;
if (!PL_colorset)
reginitcolors();
if (PL_reg_match_utf8)
t = s - prog->check_offset_max;
if (s - strpos > prog->check_offset_max /* signed-corrected t > strpos */
&& (!do_utf8
- || ((t = reghopmaybe3_c(s, -(prog->check_offset_max), strpos))
+ || ((t = (char*)reghopmaybe3((U8*)s, -(prog->check_offset_max), (U8*)strpos))
&& t > strpos)))
- /* EMPTY */;
+ NOOP;
else
t = strpos;
t = HOP3c(t, prog->anchored_offset, strend);
t = s - prog->check_offset_max;
if (s - strpos > prog->check_offset_max /* signed-corrected t > strpos */
&& (!do_utf8
- || ((t = reghopmaybe3_c(s, -prog->check_offset_max, strpos))
+ || ((t = (char*)reghopmaybe3((U8*)s, -prog->check_offset_max, (U8*)strpos))
&& t > strpos))) {
/* Fixed substring is found far enough so that the match
cannot start at strpos. */
LOAD_UTF8_CHARCLASS_ALNUM();
while (s + (uskip = UTF8SKIP(s)) <= strend) {
if (tmp == !(OP(c) == BOUND ?
- swash_fetch(PL_utf8_alnum, (U8*)s, do_utf8) :
+ (bool)swash_fetch(PL_utf8_alnum, (U8*)s, do_utf8) :
isALNUM_LC_utf8((U8*)s)))
{
tmp = !tmp;
LOAD_UTF8_CHARCLASS_ALNUM();
while (s + (uskip = UTF8SKIP(s)) <= strend) {
if (tmp == !(OP(c) == NBOUND ?
- swash_fetch(PL_utf8_alnum, (U8*)s, do_utf8) :
+ (bool)swash_fetch(PL_utf8_alnum, (U8*)s, do_utf8) :
isALNUM_LC_utf8((U8*)s)))
tmp = !tmp;
else if ((!reginfo || regtry(reginfo, s)))
? pv_uni_display(dsv0, (U8*)prog->precomp, prog->prelen, 60,
UNI_DISPLAY_REGEX)
: prog->precomp;
- const int len0 = UTF ? SvCUR(dsv0) : prog->prelen;
+ const int len0 = UTF ? (int)SvCUR(dsv0) : prog->prelen;
const char * const s1 = do_utf8 ? sv_uni_display(dsv1, sv, 60,
UNI_DISPLAY_REGEX) : startpos;
- const int len1 = do_utf8 ? SvCUR(dsv1) : strend - startpos;
+ const int len1 = do_utf8 ? (int)SvCUR(dsv1) : strend - startpos;
if (!PL_colorset)
reginitcolors();
PerlIO_printf(Perl_debug_log,
len0 = UTF ? SvCUR(dsv0) : SvCUR(prop);
s1 = UTF ?
sv_uni_display(dsv1, sv, 60, UNI_DISPLAY_REGEX) : s;
- len1 = UTF ? SvCUR(dsv1) : strend - s;
+ len1 = UTF ? (int)SvCUR(dsv1) : strend - s;
PerlIO_printf(Perl_debug_log,
"Matching stclass \"%*.*s\" against \"%*.*s\"\n",
len0, len0, s0,
#define sayNO goto no
#define sayNO_ANYOF goto no_anyof
#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 POSCACHE_SUCCESS 0 /* caching success rather than failure */
#define POSCACHE_SEEN 1 /* we know what we're caching */
#define POSCACHE_START 2 /* the real cache: this bit maps to pos 0 */
+
#define CACHEsayYES STMT_START { \
if (st->u.whilem.cache_offset | st->u.whilem.cache_bit) { \
- if (!(PL_reg_poscache[0] & (1<<POSCACHE_SEEN))) \
- PL_reg_poscache[0] |= (1<<POSCACHE_SUCCESS) || (1<<POSCACHE_SEEN); \
- else if (!(PL_reg_poscache[0] & (1<<POSCACHE_SUCCESS))) { \
+ if (!(PL_reg_poscache[0] & (1<<POSCACHE_SEEN))) { \
+ PL_reg_poscache[0] |= (1<<POSCACHE_SUCCESS) | (1<<POSCACHE_SEEN); \
+ PL_reg_poscache[st->u.whilem.cache_offset] |= (1<<st->u.whilem.cache_bit); \
+ } \
+ else if (PL_reg_poscache[0] & (1<<POSCACHE_SUCCESS)) { \
+ PL_reg_poscache[st->u.whilem.cache_offset] |= (1<<st->u.whilem.cache_bit); \
+ } \
+ else { \
/* cache records failure, but this is success */ \
DEBUG_r( \
PerlIO_printf(Perl_debug_log, \
} \
sayYES; \
} STMT_END
+
#define CACHEsayNO STMT_START { \
if (st->u.whilem.cache_offset | st->u.whilem.cache_bit) { \
- if (!(PL_reg_poscache[0] & (1<<POSCACHE_SEEN))) \
+ if (!(PL_reg_poscache[0] & (1<<POSCACHE_SEEN))) { \
PL_reg_poscache[0] |= (1<<POSCACHE_SEEN); \
- else if ((PL_reg_poscache[0] & (1<<POSCACHE_SUCCESS))) { \
+ PL_reg_poscache[st->u.whilem.cache_offset] |= (1<<st->u.whilem.cache_bit); \
+ } \
+ else if (!(PL_reg_poscache[0] & (1<<POSCACHE_SUCCESS))) { \
+ PL_reg_poscache[st->u.whilem.cache_offset] |= (1<<st->u.whilem.cache_bit); \
+ } \
+ else { \
/* cache records success, but this is failure */ \
DEBUG_r( \
PerlIO_printf(Perl_debug_log, \
/* Make sure there is a test for this +1 options in re_tests */
#define TRIE_INITAL_ACCEPT_BUFFLEN 4;
+/* this value indiciates that the c1/c2 "next char" test should be skipped */
+#define CHRTEST_VOID -1000
+
#define SLAB_FIRST(s) (&(s)->states[0])
#define SLAB_LAST(s) (&(s)->states[PERL_REGMATCH_SLAB_SLOTS-1])
STATIC regmatch_state *
S_push_slab(pTHX)
{
+#if PERL_VERSION < 9
+ dMY_CXT;
+#endif
regmatch_slab *s = PL_regmatch_slab->next;
if (!s) {
Newx(s, 1, regmatch_slab);
STATIC I32 /* 0 failure, 1 success */
S_regmatch(pTHX_ const regmatch_info *reginfo, regnode *prog)
{
+#if PERL_VERSION < 9
+ dMY_CXT;
+#endif
dVAR;
register const bool do_utf8 = PL_reg_match_utf8;
const U32 uniflags = UTF8_ALLOW_DEFAULT;
pv_uni_display(PERL_DEBUG_PAD(0), (U8*)(locinput - pref_len),
pref0_len, 60, UNI_DISPLAY_REGEX) :
locinput - pref_len;
- const int len0 = do_utf8 ? strlen(s0) : pref0_len;
+ const int len0 = do_utf8 ? (int)strlen(s0) : pref0_len;
const char * const s1 = do_utf8 && OP(scan) != CANY ?
pv_uni_display(PERL_DEBUG_PAD(1),
(U8*)(locinput - pref_len + pref0_len),
pref_len - pref0_len, 60, UNI_DISPLAY_REGEX) :
locinput - pref_len + pref0_len;
- const int len1 = do_utf8 ? strlen(s1) : pref_len - pref0_len;
+ const int len1 = do_utf8 ? (int)strlen(s1) : pref_len - pref0_len;
const char * const s2 = do_utf8 && OP(scan) != CANY ?
pv_uni_display(PERL_DEBUG_PAD(2), (U8*)locinput,
PL_regeol - locinput, 60, UNI_DISPLAY_REGEX) :
locinput;
- const int len2 = do_utf8 ? strlen(s2) : l;
+ const int len2 = do_utf8 ? (int)strlen(s2) : l;
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),
PerlIO_printf( Perl_debug_log, "%*s %strying alternation #%d <%s> at 0x%p%s\n",
REPORT_CODE_OFF+PL_regindent*2, "", PL_colors[4],
st->u.trie.accept_buff[best].wordnum,
- tmp ? SvPV_nolen_const( *tmp ) : "not compiled under -Dr",scan,
+ tmp ? SvPV_nolen_const( *tmp ) : "not compiled under -Dr", (void*)scan,
PL_colors[5] );
});
if ( best<st->u.trie.accepted ) {
if (do_utf8) {
LOAD_UTF8_CHARCLASS_ALNUM();
if (!(OP(scan) == ALNUM
- ? swash_fetch(PL_utf8_alnum, (U8*)locinput, do_utf8)
+ ? (bool)swash_fetch(PL_utf8_alnum, (U8*)locinput, do_utf8)
: isALNUM_LC_utf8((U8*)locinput)))
{
sayNO;
if (do_utf8) {
LOAD_UTF8_CHARCLASS_ALNUM();
if (OP(scan) == NALNUM
- ? swash_fetch(PL_utf8_alnum, (U8*)locinput, do_utf8)
+ ? (bool)swash_fetch(PL_utf8_alnum, (U8*)locinput, do_utf8)
: isALNUM_LC_utf8((U8*)locinput))
{
sayNO;
if (UTF8_IS_CONTINUED(nextchr)) {
LOAD_UTF8_CHARCLASS_SPACE();
if (!(OP(scan) == SPACE
- ? swash_fetch(PL_utf8_space, (U8*)locinput, do_utf8)
+ ? (bool)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, do_utf8)
+ ? (bool)swash_fetch(PL_utf8_space, (U8*)locinput, do_utf8)
: isSPACE_LC_utf8((U8*)locinput))
{
sayNO;
if (do_utf8) {
LOAD_UTF8_CHARCLASS_DIGIT();
if (!(OP(scan) == DIGIT
- ? swash_fetch(PL_utf8_digit, (U8*)locinput, do_utf8)
+ ? (bool)swash_fetch(PL_utf8_digit, (U8*)locinput, do_utf8)
: isDIGIT_LC_utf8((U8*)locinput)))
{
sayNO;
if (do_utf8) {
LOAD_UTF8_CHARCLASS_DIGIT();
if (OP(scan) == NDIGIT
- ? swash_fetch(PL_utf8_digit, (U8*)locinput, do_utf8)
+ ? (bool)swash_fetch(PL_utf8_digit, (U8*)locinput, do_utf8)
: isDIGIT_LC_utf8((U8*)locinput))
{
sayNO;
*that* much linear. */
if (!PL_reg_maxiter) {
PL_reg_maxiter = (PL_regeol - PL_bostr + 1) * (scan->flags>>4);
+ /* possible overflow for long strings and many CURLYX's */
+ if (PL_reg_maxiter < 0)
+ PL_reg_maxiter = I32_MAX;
PL_reg_leftiter = PL_reg_maxiter;
}
if (PL_reg_leftiter-- == 0) {
/* cache records failure */
sayNO_SILENT;
}
- PL_reg_poscache[st->u.whilem.cache_offset] |= (1<<st->u.whilem.cache_bit);
}
}
scan += NEXT_OFF(scan); /* Skip former OPEN. */
PL_reginput = locinput;
st->u.curlym.maxwanted = st->minmod ? st->ln : n;
- if (st->u.curlym.maxwanted) {
- while (PL_reginput < PL_regeol && st->u.curlym.matches < st->u.curlym.maxwanted) {
- REGMATCH(scan, CURLYM1);
- /*** all unsaved local vars undefined at this point */
- if (!result)
- break;
- /* on first match, determine length, u.curlym.l */
- if (!st->u.curlym.matches++) {
- if (PL_reg_match_utf8) {
- char *s = locinput;
- while (s < PL_reginput) {
- st->u.curlym.l++;
- s += UTF8SKIP(s);
- }
- }
- else {
- st->u.curlym.l = PL_reginput - locinput;
- }
- if (st->u.curlym.l == 0) {
- st->u.curlym.matches = st->u.curlym.maxwanted;
- break;
+ while (PL_reginput < PL_regeol && st->u.curlym.matches < st->u.curlym.maxwanted) {
+ /* resume to current state on success */
+ st->u.yes.prev_yes_state = yes_state;
+ yes_state = st;
+ REGMATCH(scan, CURLYM1);
+ yes_state = st->u.yes.prev_yes_state;
+ /*** all unsaved local vars undefined at this point */
+ if (!result)
+ break;
+ /* on first match, determine length, u.curlym.l */
+ if (!st->u.curlym.matches++) {
+ if (PL_reg_match_utf8) {
+ char *s = locinput;
+ while (s < PL_reginput) {
+ st->u.curlym.l++;
+ s += UTF8SKIP(s);
}
}
- locinput = PL_reginput;
+ else {
+ st->u.curlym.l = PL_reginput - locinput;
+ }
+ if (st->u.curlym.l == 0) {
+ st->u.curlym.matches = st->u.curlym.maxwanted;
+ break;
+ }
}
+ locinput = PL_reginput;
}
PL_reginput = locinput;
-
- if (st->minmod) {
+ if (st->u.curlym.matches < st->ln) {
st->minmod = 0;
- if (st->ln && st->u.curlym.matches < st->ln)
- sayNO;
- if (HAS_TEXT(next) || JUMPABLE(next)) {
- regnode *text_node = next;
+ sayNO;
+ }
+
+ DEBUG_EXECUTE_r(
+ PerlIO_printf(Perl_debug_log,
+ "%*s matched %"IVdf" times, len=%"IVdf"...\n",
+ (int)(REPORT_CODE_OFF+PL_regindent*2), "",
+ (IV) st->u.curlym.matches, (IV)st->u.curlym.l)
+ );
+
+ /* calculate c1 and c1 for possible match of 1st char
+ * following curly */
+ st->u.curlym.c1 = st->u.curlym.c2 = CHRTEST_VOID;
+ if (HAS_TEXT(next) || JUMPABLE(next)) {
+ regnode *text_node = next;
+ if (! HAS_TEXT(text_node)) FIND_NEXT_IMPT(text_node);
+ if (HAS_TEXT(text_node)
+ && PL_regkind[(U8)OP(text_node)] != REF)
+ {
+ st->u.curlym.c1 = (U8)*STRING(text_node);
+ st->u.curlym.c2 =
+ (OP(text_node) == EXACTF || OP(text_node) == REFF)
+ ? PL_fold[st->u.curlym.c1]
+ : (OP(text_node) == EXACTFL || OP(text_node) == REFFL)
+ ? PL_fold_locale[st->u.curlym.c1]
+ : st->u.curlym.c1;
+ }
+ }
- if (! HAS_TEXT(text_node)) FIND_NEXT_IMPT(text_node);
+ REGCP_SET(st->u.curlym.lastcp);
- if (! HAS_TEXT(text_node)) st->u.curlym.c1 = st->u.curlym.c2 = -1000;
- else {
- if (PL_regkind[(U8)OP(text_node)] == REF) {
- st->u.curlym.c1 = st->u.curlym.c2 = -1000;
- goto assume_ok_MM;
+ st->u.curlym.minmod = st->minmod;
+ st->minmod = 0;
+ while (st->u.curlym.matches >= st->ln
+ && (st->u.curlym.matches <= n
+ /* for REG_INFTY, ln could overflow to negative */
+ || (n == REG_INFTY && st->u.curlym.matches >= 0)))
+ {
+ /* If it could work, try it. */
+ if (st->u.curlym.c1 == CHRTEST_VOID ||
+ UCHARAT(PL_reginput) == st->u.curlym.c1 ||
+ UCHARAT(PL_reginput) == st->u.curlym.c2)
+ {
+ DEBUG_EXECUTE_r(
+ PerlIO_printf(Perl_debug_log,
+ "%*s trying tail with matches=%"IVdf"...\n",
+ (int)(REPORT_CODE_OFF+PL_regindent*2),
+ "", (IV)st->u.curlym.matches)
+ );
+ if (st->u.curlym.paren) {
+ if (st->u.curlym.matches) {
+ PL_regstartp[st->u.curlym.paren]
+ = HOPc(PL_reginput, -st->u.curlym.l) - PL_bostr;
+ PL_regendp[st->u.curlym.paren] = PL_reginput - PL_bostr;
}
- else { st->u.curlym.c1 = (U8)*STRING(text_node); }
- if (OP(text_node) == EXACTF || OP(text_node) == REFF)
- st->u.curlym.c2 = PL_fold[st->u.curlym.c1];
- else if (OP(text_node) == EXACTFL || OP(text_node) == REFFL)
- st->u.curlym.c2 = PL_fold_locale[st->u.curlym.c1];
else
- st->u.curlym.c2 = st->u.curlym.c1;
+ PL_regendp[st->u.curlym.paren] = -1;
}
+ /* resume to current state on success */
+ st->u.yes.prev_yes_state = yes_state;
+ yes_state = st;
+ REGMATCH(next, CURLYM2);
+ yes_state = st->u.yes.prev_yes_state;
+ /*** all unsaved local vars undefined at this point */
+ if (result)
+ /* XXX tmp sayYES; */
+ sayYES_FINAL;
+ REGCP_UNWIND(st->u.curlym.lastcp);
}
- else
- st->u.curlym.c1 = st->u.curlym.c2 = -1000;
- assume_ok_MM:
- REGCP_SET(st->u.curlym.lastcp);
- while (n >= st->ln || (n == REG_INFTY && st->ln > 0)) { /* ln overflow ? */
- /* If it could work, try it. */
- if (st->u.curlym.c1 == -1000 ||
- UCHARAT(PL_reginput) == st->u.curlym.c1 ||
- UCHARAT(PL_reginput) == st->u.curlym.c2)
- {
- if (st->u.curlym.paren) {
- if (st->ln) {
- PL_regstartp[st->u.curlym.paren] =
- HOPc(PL_reginput, -st->u.curlym.l) - PL_bostr;
- PL_regendp[st->u.curlym.paren] = PL_reginput - PL_bostr;
- }
- else
- PL_regendp[st->u.curlym.paren] = -1;
- }
- REGMATCH(next, CURLYM2);
- /*** all unsaved local vars undefined at this point */
- if (result)
- sayYES;
- REGCP_UNWIND(st->u.curlym.lastcp);
- }
- /* Couldn't or didn't -- move forward. */
+ /* Couldn't or didn't -- move forward/backward. */
+ if (st->u.curlym.minmod) {
PL_reginput = locinput;
+ /* resume to current state on success */
+ st->u.yes.prev_yes_state = yes_state;
+ yes_state = st;
REGMATCH(scan, CURLYM3);
+ yes_state = st->u.yes.prev_yes_state;
/*** all unsaved local vars undefined at this point */
if (result) {
- st->ln++;
+ st->u.curlym.matches++;
locinput = PL_reginput;
}
else
sayNO;
}
- }
- else {
- DEBUG_EXECUTE_r(
- PerlIO_printf(Perl_debug_log,
- "%*s matched %"IVdf" times, len=%"IVdf"...\n",
- (int)(REPORT_CODE_OFF+PL_regindent*2), "",
- (IV) st->u.curlym.matches, (IV)st->u.curlym.l)
- );
- if (st->u.curlym.matches >= st->ln) {
- if (HAS_TEXT(next) || JUMPABLE(next)) {
- regnode *text_node = next;
-
- if (! HAS_TEXT(text_node)) FIND_NEXT_IMPT(text_node);
-
- if (! HAS_TEXT(text_node)) st->u.curlym.c1 = st->u.curlym.c2 = -1000;
- else {
- if (PL_regkind[(U8)OP(text_node)] == REF) {
- st->u.curlym.c1 = st->u.curlym.c2 = -1000;
- goto assume_ok_REG;
- }
- else { st->u.curlym.c1 = (U8)*STRING(text_node); }
-
- if (OP(text_node) == EXACTF || OP(text_node) == REFF)
- st->u.curlym.c2 = PL_fold[st->u.curlym.c1];
- else if (OP(text_node) == EXACTFL || OP(text_node) == REFFL)
- st->u.curlym.c2 = PL_fold_locale[st->u.curlym.c1];
- else
- st->u.curlym.c2 = st->u.curlym.c1;
- }
- }
- else
- st->u.curlym.c1 = st->u.curlym.c2 = -1000;
- }
- assume_ok_REG:
- REGCP_SET(st->u.curlym.lastcp);
- while (st->u.curlym.matches >= st->ln) {
- /* If it could work, try it. */
- if (st->u.curlym.c1 == -1000 ||
- UCHARAT(PL_reginput) == st->u.curlym.c1 ||
- UCHARAT(PL_reginput) == st->u.curlym.c2)
- {
- DEBUG_EXECUTE_r(
- PerlIO_printf(Perl_debug_log,
- "%*s trying tail with matches=%"IVdf"...\n",
- (int)(REPORT_CODE_OFF+PL_regindent*2),
- "", (IV)st->u.curlym.matches)
- );
- if (st->u.curlym.paren) {
- if (st->u.curlym.matches) {
- PL_regstartp[st->u.curlym.paren]
- = HOPc(PL_reginput, -st->u.curlym.l) - PL_bostr;
- PL_regendp[st->u.curlym.paren] = PL_reginput - PL_bostr;
- }
- else
- PL_regendp[st->u.curlym.paren] = -1;
- }
- REGMATCH(next, CURLYM4);
- /*** all unsaved local vars undefined at this point */
- if (result)
- sayYES;
- REGCP_UNWIND(st->u.curlym.lastcp);
- }
- /* Couldn't or didn't -- back up. */
+ else {
st->u.curlym.matches--;
locinput = HOPc(locinput, -st->u.curlym.l);
PL_reginput = locinput;
if (! HAS_TEXT(text_node)) FIND_NEXT_IMPT(text_node);
- if (! HAS_TEXT(text_node)) st->u.plus.c1 = st->u.plus.c2 = -1000;
+ if (! HAS_TEXT(text_node))
+ st->u.plus.c1 = st->u.plus.c2 = CHRTEST_VOID;
else {
if (PL_regkind[(U8)OP(text_node)] == REF) {
- st->u.plus.c1 = st->u.plus.c2 = -1000;
+ st->u.plus.c1 = st->u.plus.c2 = CHRTEST_VOID;
goto assume_ok_easy;
}
else { s = (U8*)STRING(text_node); }
}
}
else
- st->u.plus.c1 = st->u.plus.c2 = -1000;
+ st->u.plus.c1 = st->u.plus.c2 = CHRTEST_VOID;
assume_ok_easy:
PL_reginput = locinput;
if (st->minmod) {
sayNO;
locinput = PL_reginput;
REGCP_SET(st->u.plus.lastcp);
- if (st->u.plus.c1 != -1000) {
+ if (st->u.plus.c1 != CHRTEST_VOID) {
st->u.plus.old = locinput;
st->u.plus.count = 0;
st->u.plus.count++;
}
} else {
- STRLEN len;
/* count initialised to
* utf8_distance(old, locinput) */
while (locinput <= st->u.plus.e) {
- UV c = utf8n_to_uvchr((U8*)locinput,
+ STRLEN len;
+ const UV c = utf8n_to_uvchr((U8*)locinput,
UTF8_MAXBYTES, &len,
uniflags);
if (c == (UV)st->u.plus.c1 || c == (UV)st->u.plus.c2)
else
while (n >= st->ln || (n == REG_INFTY && st->ln > 0)) { /* ln overflow ? */
UV c;
- if (st->u.plus.c1 != -1000) {
+ if (st->u.plus.c1 != CHRTEST_VOID) {
if (do_utf8)
c = utf8n_to_uvchr((U8*)PL_reginput,
UTF8_MAXBYTES, 0,
}
}
/* If it could work, try it. */
- else if (st->u.plus.c1 == -1000)
+ else if (st->u.plus.c1 == CHRTEST_VOID)
{
TRYPAREN(st->u.plus.paren, st->ln, PL_reginput, PLUS3);
/*** all unsaved local vars undefined at this point */
{
UV c = 0;
while (n >= st->ln) {
- if (st->u.plus.c1 != -1000) {
+ if (st->u.plus.c1 != CHRTEST_VOID) {
if (do_utf8)
c = utf8n_to_uvchr((U8*)PL_reginput,
UTF8_MAXBYTES, 0,
c = UCHARAT(PL_reginput);
}
/* If it could work, try it. */
- if (st->u.plus.c1 == -1000 || c == (UV)st->u.plus.c1 || c == (UV)st->u.plus.c2)
+ if (st->u.plus.c1 == CHRTEST_VOID || c == (UV)st->u.plus.c1 || c == (UV)st->u.plus.c2)
{
TRYPAREN(st->u.plus.paren, n, PL_reginput, PLUS4);
/*** all unsaved local vars undefined at this point */
}
PL_reginput = locinput; /* put where regtry can find it */
sayYES_FINAL; /* Success! */
- case SUCCEED:
+
+ case SUCCEED: /* successful SUSPEND/UNLESSM/IFMATCH/CURLYM */
+ DEBUG_EXECUTE_r(
+ PerlIO_printf(Perl_debug_log,
+ "%*s %ssubpattern success...%s\n",
+ REPORT_CODE_OFF+PL_regindent*2, "", PL_colors[4], PL_colors[5]));
PL_reginput = locinput; /* put where regtry can find it */
- sayYES_LOUD; /* Success! */
- case SUSPEND:
- n = 1;
+ sayYES_FINAL; /* Success! */
+
+ case SUSPEND: /* (?>FOO) */
+ st->u.ifmatch.wanted = 1;
PL_reginput = locinput;
goto do_ifmatch;
- case UNLESSM:
- n = 0;
- if (scan->flags) {
- char * const s = HOPBACKc(locinput, scan->flags);
- if (!s)
- goto say_yes;
- PL_reginput = s;
- }
- else
- PL_reginput = locinput;
- goto do_ifmatch;
- case IFMATCH:
- n = 1;
+
+ case UNLESSM: /* -ve lookaround: (?!FOO), or with flags, (?<!foo) */
+ st->u.ifmatch.wanted = 0;
+ goto ifmatch_trivial_fail_test;
+
+ case IFMATCH: /* +ve lookaround: (?=FOO), or with flags, (?<=foo) */
+ st->u.ifmatch.wanted = 1;
+ ifmatch_trivial_fail_test:
if (scan->flags) {
char * const s = HOPBACKc(locinput, scan->flags);
- if (!s)
- goto say_no;
+ if (!s) {
+ /* trivial fail */
+ if (st->logical) {
+ st->logical = 0;
+ st->sw = 1 - st->u.ifmatch.wanted;
+ }
+ else if (st->u.ifmatch.wanted)
+ sayNO;
+ next = scan + ARG(scan);
+ if (next == scan)
+ next = NULL;
+ break;
+ }
PL_reginput = s;
}
else
PL_reginput = locinput;
do_ifmatch:
- REGMATCH(NEXTOPER(NEXTOPER(scan)), IFMATCH);
- /*** all unsaved local vars undefined at this point */
- if (result != n) {
- say_no:
- if (st->logical) {
- st->logical = 0;
- st->sw = 0;
- goto do_longjump;
- }
- else
- sayNO;
- }
- say_yes:
- if (st->logical) {
- st->logical = 0;
- st->sw = 1;
- }
- if (OP(scan) == SUSPEND) {
- locinput = PL_reginput;
- nextchr = UCHARAT(locinput);
- }
- /* FALL THROUGH. */
+ /* resume to current state on success */
+ st->u.yes.prev_yes_state = yes_state;
+ yes_state = st;
+ PUSH_STATE(newst, resume_IFMATCH);
+ st = newst;
+ next = NEXTOPER(NEXTOPER(scan));
+ break;
+
case LONGJMP:
- do_longjump:
next = scan + ARG(scan);
if (next == scan)
next = NULL;
/*NOTREACHED*/
sayNO;
-yes_loud:
- DEBUG_EXECUTE_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:
if (yes_state) {
/* we have successfully completed a subexpression, but we must now
* pop to the state marked by yes_state and continue from there */
+ /*XXX tmp for CURLYM*/
+ regmatch_slab * const oslab = PL_regmatch_slab;
+ regmatch_state * const ost = st;
+ regmatch_state * const oys = yes_state;
+ int odepth = depth;
+
assert(st != yes_state);
while (yes_state < SLAB_FIRST(PL_regmatch_slab)
|| yes_state > SLAB_LAST(PL_regmatch_slab))
st = SLAB_LAST(PL_regmatch_slab);
}
depth -= (st - yes_state);
- DEBUG_EXECUTE_r(PerlIO_printf(Perl_debug_log, "POP STATE TO (%d)\n", depth)); \
+ DEBUG_EXECUTE_r(PerlIO_printf(Perl_debug_log, "POP STATE TO (%d)\n", depth));
st = yes_state;
yes_state = st->u.yes.prev_yes_state;
PL_regmatch_state = st;
/* Restore parens of the caller without popping the
* savestack */
{
- I32 tmp = PL_savestack_ix;
+ const I32 tmp = PL_savestack_ix;
PL_savestack_ix = st->u.eval.lastcp;
regcppop(rex);
PL_savestack_ix = tmp;
next = st->next;
goto reenter;
+ case resume_IFMATCH:
+ if (st->logical) {
+ st->logical = 0;
+ st->sw = st->u.ifmatch.wanted;
+ }
+ else if (!st->u.ifmatch.wanted)
+ sayNO;
+
+ if (OP(st->scan) == SUSPEND)
+ locinput = PL_reginput;
+ else {
+ locinput = PL_reginput = st->locinput;
+ nextchr = UCHARAT(locinput);
+ }
+ next = st->scan + ARG(st->scan);
+ if (next == st->scan)
+ next = NULL;
+ goto reenter;
+
+ /* XXX tmp don't handle yes_state yet */
+ case resume_CURLYM1:
+ case resume_CURLYM2:
+ case resume_CURLYM3:
+ PL_regmatch_slab =oslab;
+ st = ost;
+ PL_regmatch_state = st;
+ depth = odepth;
+ yes_state = oys;
+ DEBUG_EXECUTE_r(PerlIO_printf(Perl_debug_log, "XXX revering a CURLYM\n"));
+ goto yes;
+
default:
Perl_croak(aTHX_ "unexpected yes reume state");
}
goto resume_point_CURLYM2;
case resume_CURLYM3:
goto resume_point_CURLYM3;
- case resume_CURLYM4:
- goto resume_point_CURLYM4;
- case resume_IFMATCH:
- goto resume_point_IFMATCH;
case resume_PLUS1:
goto resume_point_PLUS1;
case resume_PLUS2:
case resume_PLUS4:
goto resume_point_PLUS4;
+ case resume_IFMATCH:
case resume_EVAL:
default:
Perl_croak(aTHX_ "regexp resume memory corruption");
goto resume_point_CURLYM2;
case resume_CURLYM3:
goto resume_point_CURLYM3;
- case resume_CURLYM4:
- goto resume_point_CURLYM4;
case resume_IFMATCH:
- goto resume_point_IFMATCH;
+ yes_state = st->u.yes.prev_yes_state;
+ if (st->logical) {
+ st->logical = 0;
+ st->sw = !st->u.ifmatch.wanted;
+ }
+ else if (st->u.ifmatch.wanted)
+ sayNO;
+
+ assert(OP(scan) != SUSPEND); /* XXX DAPM tmp */
+ locinput = PL_reginput = st->locinput;
+ nextchr = UCHARAT(locinput);
+ next = scan + ARG(scan);
+ if (next == scan)
+ next = NULL;
+ goto reenter;
+
case resume_PLUS1:
goto resume_point_PLUS1;
case resume_PLUS2:
/* free all slabs above current one */
if (orig_slab->next) {
- regmatch_slab *osl, *sl = orig_slab->next;
+ regmatch_slab *sl = orig_slab->next;
orig_slab->next = NULL;
while (sl) {
- osl = sl;
+ regmatch_slab * const osl = sl;
sl = sl->next;
Safefree(osl);
}
}
+#ifndef PERL_IN_XSUB_RE
/*
- regclass_swash - prepare the utf8 swash
*/
return sw;
}
+#endif
/*
- reginclass - determine if a character falls into a character class
}
STATIC U8 *
-S_reghop3(U8 *s, I32 off, U8* lim)
+S_reghop3(U8 *s, I32 off, const U8* lim)
{
dVAR;
if (off >= 0) {
}
STATIC U8 *
-S_reghopmaybe3(U8* s, I32 off, U8* lim)
+S_reghopmaybe3(U8* s, I32 off, const U8* lim)
{
dVAR;
if (off >= 0) {