#define RF_utf8 8 /* Pattern contains multibyte chars? */
-#define UTF ((PL_reg_flags & RF_utf8) != 0)
+#define UTF_PATTERN ((PL_reg_flags & RF_utf8) != 0)
#define RS_init 1 /* eval environment created */
#define RS_set 2 /* replsv value is set */
#define STATIC static
#endif
-#define REGINCLASS(prog,p,c) (ANYOF_FLAGS(p) ? reginclass(prog,p,c,0,0) : ANYOF_BITMAP_TEST(p,*(c)))
+/* Valid for non-utf8 strings only: avoids the reginclass call if there are no
+ * complications: i.e., if everything matchable is straight forward in the
+ * bitmap */
+#define REGINCLASS(prog,p,c) (ANYOF_FLAGS(p) ? reginclass(prog,p,c,0,0) \
+ : ANYOF_BITMAP_TEST(p,*(c)))
/*
* Forwards.
*/
-#define CHR_SVLEN(sv) (do_utf8 ? sv_len_utf8(sv) : SvCUR(sv))
+#define CHR_SVLEN(sv) (utf8_target ? sv_len_utf8(sv) : SvCUR(sv))
#define CHR_DIST(a,b) (PL_reg_match_utf8 ? utf8_distance(a,b) : a - b)
#define HOPc(pos,off) \
#endif
-#define CCC_TRY_AFF(NAME,NAMEL,CLASS,STR,LCFUNC_utf8,FUNC,LCFUNC) \
- case NAMEL: \
- PL_reg_flags |= RF_tainted; \
- /* FALL THROUGH */ \
- case NAME: \
- if (!nextchr) \
- sayNO; \
- if (do_utf8 && UTF8_IS_CONTINUED(nextchr)) { \
- if (!CAT2(PL_utf8_,CLASS)) { \
- bool ok; \
- ENTER; \
- save_re_context(); \
- ok=CAT2(is_utf8_,CLASS)((const U8*)STR); \
- assert(ok); \
- LEAVE; \
- } \
- if (!(OP(scan) == NAME \
- ? cBOOL(swash_fetch(CAT2(PL_utf8_,CLASS), (U8*)locinput, do_utf8)) \
- : LCFUNC_utf8((U8*)locinput))) \
- { \
- sayNO; \
- } \
- locinput += PL_utf8skip[nextchr]; \
- nextchr = UCHARAT(locinput); \
- break; \
- } \
- if (!(OP(scan) == NAME ? FUNC(nextchr) : LCFUNC(nextchr))) \
- sayNO; \
- nextchr = UCHARAT(++locinput); \
+#define _CCC_TRY_AFF_COMMON(NAME,NAMEL,CLASS,STR,LCFUNC_utf8,FUNC) \
+ case NAMEL: \
+ PL_reg_flags |= RF_tainted; \
+ /* FALL THROUGH */ \
+ case NAME: \
+ if (!nextchr) \
+ sayNO; \
+ if (utf8_target && UTF8_IS_CONTINUED(nextchr)) { \
+ if (!CAT2(PL_utf8_,CLASS)) { \
+ bool ok; \
+ ENTER; \
+ save_re_context(); \
+ ok=CAT2(is_utf8_,CLASS)((const U8*)STR); \
+ assert(ok); \
+ LEAVE; \
+ } \
+ if (!(OP(scan) == NAME \
+ ? cBOOL(swash_fetch(CAT2(PL_utf8_,CLASS), (U8*)locinput, utf8_target)) \
+ : LCFUNC_utf8((U8*)locinput))) \
+ { \
+ sayNO; \
+ } \
+ locinput += PL_utf8skip[nextchr]; \
+ nextchr = UCHARAT(locinput); \
+ break; \
+ } \
+ /* Drops through to the macro that calls this one */
+
+#define CCC_TRY_AFF(NAME,NAMEL,CLASS,STR,LCFUNC_utf8,FUNC,LCFUNC) \
+ _CCC_TRY_AFF_COMMON(NAME,NAMEL,CLASS,STR,LCFUNC_utf8,FUNC) \
+ if (!(OP(scan) == NAME ? FUNC(nextchr) : LCFUNC(nextchr))) \
+ sayNO; \
+ nextchr = UCHARAT(++locinput); \
break
-#define CCC_TRY_NEG(NAME,NAMEL,CLASS,STR,LCFUNC_utf8,FUNC,LCFUNC) \
- case NAMEL: \
- PL_reg_flags |= RF_tainted; \
- /* FALL THROUGH */ \
- case NAME : \
- if (!nextchr && locinput >= PL_regeol) \
- sayNO; \
- if (do_utf8 && UTF8_IS_CONTINUED(nextchr)) { \
- if (!CAT2(PL_utf8_,CLASS)) { \
- bool ok; \
- ENTER; \
- save_re_context(); \
- ok=CAT2(is_utf8_,CLASS)((const U8*)STR); \
- assert(ok); \
- LEAVE; \
- } \
- if ((OP(scan) == NAME \
- ? cBOOL(swash_fetch(CAT2(PL_utf8_,CLASS), (U8*)locinput, do_utf8)) \
- : LCFUNC_utf8((U8*)locinput))) \
- { \
- sayNO; \
- } \
- locinput += PL_utf8skip[nextchr]; \
- nextchr = UCHARAT(locinput); \
- break; \
- } \
- if ((OP(scan) == NAME ? FUNC(nextchr) : LCFUNC(nextchr))) \
- sayNO; \
- nextchr = UCHARAT(++locinput); \
+/* Almost identical to the above, but has a case for a node that matches chars
+ * between 128 and 255 using Unicode (latin1) semantics. */
+#define CCC_TRY_AFF_U(NAME,NAMEL,CLASS,STR,LCFUNC_utf8,FUNCU,LCFUNC) \
+ _CCC_TRY_AFF_COMMON(NAME,NAMEL,CLASS,STR,LCFUNC_utf8,FUNC) \
+ if (!(OP(scan) == NAMEL ? LCFUNC(nextchr) : (FUNCU(nextchr) && (isASCII(nextchr) || (FLAGS(scan) & USE_UNI))))) \
+ sayNO; \
+ nextchr = UCHARAT(++locinput); \
break
+#define _CCC_TRY_NEG_COMMON(NAME,NAMEL,CLASS,STR,LCFUNC_utf8,FUNC) \
+ case NAMEL: \
+ PL_reg_flags |= RF_tainted; \
+ /* FALL THROUGH */ \
+ case NAME : \
+ if (!nextchr && locinput >= PL_regeol) \
+ sayNO; \
+ if (utf8_target && UTF8_IS_CONTINUED(nextchr)) { \
+ if (!CAT2(PL_utf8_,CLASS)) { \
+ bool ok; \
+ ENTER; \
+ save_re_context(); \
+ ok=CAT2(is_utf8_,CLASS)((const U8*)STR); \
+ assert(ok); \
+ LEAVE; \
+ } \
+ if ((OP(scan) == NAME \
+ ? cBOOL(swash_fetch(CAT2(PL_utf8_,CLASS), (U8*)locinput, utf8_target)) \
+ : LCFUNC_utf8((U8*)locinput))) \
+ { \
+ sayNO; \
+ } \
+ locinput += PL_utf8skip[nextchr]; \
+ nextchr = UCHARAT(locinput); \
+ break; \
+ }
+
+#define CCC_TRY_NEG(NAME,NAMEL,CLASS,STR,LCFUNC_utf8,FUNC,LCFUNC) \
+ _CCC_TRY_NEG_COMMON(NAME,NAMEL,CLASS,STR,LCFUNC_utf8,FUNC) \
+ if ((OP(scan) == NAME ? FUNC(nextchr) : LCFUNC(nextchr))) \
+ sayNO; \
+ nextchr = UCHARAT(++locinput); \
+ break
+#define CCC_TRY_NEG_U(NAME,NAMEL,CLASS,STR,LCFUNC_utf8,FUNCU,LCFUNC) \
+ _CCC_TRY_NEG_COMMON(NAME,NAMEL,CLASS,STR,LCFUNC_utf8,FUNCU) \
+ if ((OP(scan) == NAMEL ? LCFUNC(nextchr) : (FUNCU(nextchr) && (isASCII(nextchr) || (FLAGS(scan) & USE_UNI))))) \
+ sayNO; \
+ nextchr = UCHARAT(++locinput); \
+ break
+
/* TODO: Combine JUMPABLE and HAS_TEXT to cache OP(rn) */
/* for use after a quantifier and before an EXACT-like node -- japhy */
-/* it would be nice to rework regcomp.sym to generate this stuff. sigh */
+/* it would be nice to rework regcomp.sym to generate this stuff. sigh
+ *
+ * NOTE that *nothing* that affects backtracking should be in here, specifically
+ * VERBS must NOT be included. JUMPABLE is used to determine if we can ignore a
+ * node that is in between two EXACT like nodes when ascertaining what the required
+ * "follow" character is. This should probably be moved to regex compile time
+ * although it may be done at run time beause of the REF possibility - more
+ * investigation required. -- demerphq
+*/
#define JUMPABLE(rn) ( \
OP(rn) == OPEN || \
(OP(rn) == CLOSE && (!cur_eval || cur_eval->u.eval.close_paren != ARG(rn))) || \
OP(rn) == EVAL || \
OP(rn) == SUSPEND || OP(rn) == IFMATCH || \
OP(rn) == PLUS || OP(rn) == MINMOD || \
- OP(rn) == KEEPS || (PL_regkind[OP(rn)] == VERB) || \
+ OP(rn) == KEEPS || \
(PL_regkind[OP(rn)] == CURLY && ARG1(rn) > 0) \
)
#define IS_EXACT(rn) (PL_regkind[OP(rn)] == EXACT)
/* Currently these are only used when PL_regkind[OP(rn)] == EXACT so
we don't need this definition. */
#define IS_TEXT(rn) ( OP(rn)==EXACT || OP(rn)==REF || OP(rn)==NREF )
-#define IS_TEXTF(rn) ( OP(rn)==EXACTF || OP(rn)==REFF || OP(rn)==NREFF )
+#define IS_TEXTF(rn) ( (OP(rn)==EXACTFU || OP(rn)==EXACTF) || OP(rn)==REFF || OP(rn)==NREFF )
#define IS_TEXTFL(rn) ( OP(rn)==EXACTFL || OP(rn)==REFFL || OP(rn)==NREFFL )
#else
/* ... so we use this as its faster. */
#define IS_TEXT(rn) ( OP(rn)==EXACT )
+#define IS_TEXTFU(rn) ( OP(rn)==EXACTFU )
#define IS_TEXTF(rn) ( OP(rn)==EXACTF )
#define IS_TEXTFL(rn) ( OP(rn)==EXACTFL )
register SV *check;
char *strbeg;
char *t;
- const bool do_utf8 = (sv && SvUTF8(sv)) ? 1 : 0; /* if no sv we have to assume bytes */
+ const bool utf8_target = (sv && SvUTF8(sv)) ? 1 : 0; /* if no sv we have to assume bytes */
I32 ml_anch;
register char *other_last = NULL; /* other substr checked before this */
char *check_at = NULL; /* check substr found at this pos */
PERL_ARGS_ASSERT_RE_INTUIT_START;
- RX_MATCH_UTF8_set(rx,do_utf8);
+ RX_MATCH_UTF8_set(rx,utf8_target);
if (RX_UTF8(rx)) {
PL_reg_flags |= RF_utf8;
}
DEBUG_EXECUTE_r(
- debug_start_match(rx, do_utf8, strpos, strend,
+ debug_start_match(rx, utf8_target, strpos, strend,
sv ? "Guessing start of match in sv for"
: "Guessing start of match in string for");
);
strbeg = (sv && SvPOK(sv)) ? strend - SvCUR(sv) : strpos;
PL_regeol = strend;
- if (do_utf8) {
+ if (utf8_target) {
if (!prog->check_utf8 && prog->check_substr)
to_utf8_substr(prog);
check = prog->check_utf8;
unshift s. */
DEBUG_EXECUTE_r({
- RE_PV_QUOTED_DECL(quoted, do_utf8, PERL_DEBUG_PAD_ZERO(0),
+ RE_PV_QUOTED_DECL(quoted, utf8_target, PERL_DEBUG_PAD_ZERO(0),
SvPVX_const(check), RE_SV_DUMPLEN(check), 30);
PerlIO_printf(Perl_debug_log, "%s %s substr %s%s%s",
(s ? "Found" : "Did not find"),
- (check == (do_utf8 ? prog->anchored_utf8 : prog->anchored_substr)
+ (check == (utf8_target ? prog->anchored_utf8 : prog->anchored_substr)
? "anchored" : "floating"),
quoted,
RE_SV_TAIL(check),
Probably it is right to do no SCREAM here...
*/
- if (do_utf8 ? (prog->float_utf8 && prog->anchored_utf8)
+ if (utf8_target ? (prog->float_utf8 && prog->anchored_utf8)
: (prog->float_substr && prog->anchored_substr))
{
/* Take into account the "other" substring. */
/* XXXX May be hopelessly wrong for UTF... */
if (!other_last)
other_last = strpos;
- if (check == (do_utf8 ? prog->float_utf8 : prog->float_substr)) {
+ if (check == (utf8_target ? prog->float_utf8 : prog->float_substr)) {
do_other_anchored:
{
char * const last = HOP3c(s, -start_shift, strbeg);
t = s - prog->check_offset_max;
if (s - strpos > prog->check_offset_max /* signed-corrected t > strpos */
- && (!do_utf8
+ && (!utf8_target
|| ((t = (char*)reghopmaybe3((U8*)s, -(prog->check_offset_max), (U8*)strpos))
&& t > strpos)))
NOOP;
*/
/* On end-of-str: see comment below. */
- must = do_utf8 ? prog->anchored_utf8 : prog->anchored_substr;
+ must = utf8_target ? prog->anchored_utf8 : prog->anchored_substr;
if (must == &PL_sv_undef) {
s = (char*)NULL;
DEBUG_r(must = prog->anchored_utf8); /* for debug */
multiline ? FBMrf_MULTILINE : 0
);
DEBUG_EXECUTE_r({
- RE_PV_QUOTED_DECL(quoted, do_utf8, PERL_DEBUG_PAD_ZERO(0),
+ RE_PV_QUOTED_DECL(quoted, utf8_target, PERL_DEBUG_PAD_ZERO(0),
SvPVX_const(must), RE_SV_DUMPLEN(must), 30);
PerlIO_printf(Perl_debug_log, "%s anchored substr %s%s",
(s ? "Found" : "Contradicts"),
if (s < other_last)
s = other_last;
/* XXXX It is not documented what units *_offsets are in. Assume bytes. */
- must = do_utf8 ? prog->float_utf8 : prog->float_substr;
+ must = utf8_target ? prog->float_utf8 : prog->float_substr;
/* fbm_instr() takes into account exact value of end-of-str
if the check is SvTAIL(ed). Since false positives are OK,
and end-of-str is not later than strend we are OK. */
- (SvTAIL(must)!=0),
must, multiline ? FBMrf_MULTILINE : 0);
DEBUG_EXECUTE_r({
- RE_PV_QUOTED_DECL(quoted, do_utf8, PERL_DEBUG_PAD_ZERO(0),
+ RE_PV_QUOTED_DECL(quoted, utf8_target, PERL_DEBUG_PAD_ZERO(0),
SvPVX_const(must), RE_SV_DUMPLEN(must), 30);
PerlIO_printf(Perl_debug_log, "%s floating substr %s%s",
(s ? "Found" : "Contradicts"),
);
if (s - strpos > prog->check_offset_max /* signed-corrected t > strpos */
- && (!do_utf8
+ && (!utf8_target
|| ((t = (char*)reghopmaybe3((U8*)s, -prog->check_offset_max, (U8*) ((prog->check_offset_max<0) ? strend : strpos)))
&& t > strpos)))
{
while (t < strend - prog->minlen) {
if (*t == '\n') {
if (t < check_at - prog->check_offset_min) {
- if (do_utf8 ? prog->anchored_utf8 : prog->anchored_substr) {
+ if (utf8_target ? prog->anchored_utf8 : prog->anchored_substr) {
/* Since we moved from the found position,
we definitely contradict the found anchored
substr. Due to the above check we do not
}
s = t;
set_useful:
- ++BmUSEFUL(do_utf8 ? prog->check_utf8 : prog->check_substr); /* hooray/5 */
+ ++BmUSEFUL(utf8_target ? prog->check_utf8 : prog->check_substr); /* hooray/5 */
}
else {
/* The found string does not prohibit matching at strpos,
);
success_at_start:
if (!(prog->intflags & PREGf_NAUGHTY) /* XXXX If strpos moved? */
- && (do_utf8 ? (
+ && (utf8_target ? (
prog->check_utf8 /* Could be deleted already */
&& --BmUSEFUL(prog->check_utf8) < 0
&& (prog->check_utf8 == prog->float_utf8)
{
/* If flags & SOMETHING - do not do it many times on the same match */
DEBUG_EXECUTE_r(PerlIO_printf(Perl_debug_log, "... Disabling check substring...\n"));
- /* XXX Does the destruction order has to change with do_utf8? */
- SvREFCNT_dec(do_utf8 ? prog->check_utf8 : prog->check_substr);
- SvREFCNT_dec(do_utf8 ? prog->check_substr : prog->check_utf8);
+ /* XXX Does the destruction order has to change with utf8_target? */
+ SvREFCNT_dec(utf8_target ? prog->check_utf8 : prog->check_substr);
+ SvREFCNT_dec(utf8_target ? prog->check_substr : prog->check_utf8);
prog->check_substr = prog->check_utf8 = NULL; /* disable */
prog->float_substr = prog->float_utf8 = NULL; /* clear */
check = NULL; /* abort */
s = strpos;
+ /* XXXX If the check string was an implicit check MBOL, then we need to unset the relevent flag
+ see http://bugs.activestate.com/show_bug.cgi?id=87173 */
+ if (prog->intflags & PREGf_IMPLICIT)
+ prog->extflags &= ~RXf_ANCH_MBOL;
/* XXXX This is a remnant of the old implementation. It
looks wasteful, since now INTUIT can use many
other heuristics. */
prog->extflags &= ~RXf_USE_INTUIT;
+ /* XXXX What other flags might need to be cleared in this branch? */
}
else
s = strpos;
even for \b or \B. But (minlen? 1 : 0) below assumes that
regstclass does not come from lookahead... */
/* If regstclass takes bytelength more than 1: If charlength==1, OK.
- This leaves EXACTF only, which is dealt with in find_byclass(). */
+ This leaves EXACTF, EXACTFU only, which are dealt with in find_byclass(). */
const U8* const str = (U8*)STRING(progi->regstclass);
const int cl_l = (PL_regkind[OP(progi->regstclass)] == EXACT
? CHR_DIST(str+STR_LEN(progi->regstclass), str)
goto fail;
/* Contradict one of substrings */
if (prog->anchored_substr || prog->anchored_utf8) {
- if ((do_utf8 ? prog->anchored_utf8 : prog->anchored_substr) == check) {
+ if ((utf8_target ? prog->anchored_utf8 : prog->anchored_substr) == check) {
DEBUG_EXECUTE_r( what = "anchored" );
hop_and_restart:
s = HOP3c(t, 1, strend);
PL_colors[0], PL_colors[1], (long)(t - i_strpos)) );
goto try_at_offset;
}
- if (!(do_utf8 ? prog->float_utf8 : prog->float_substr)) /* Could have been deleted */
+ if (!(utf8_target ? prog->float_utf8 : prog->float_substr)) /* Could have been deleted */
goto fail;
/* Check is floating subtring. */
retry_floating_check:
fail_finish: /* Substring not found */
if (prog->check_substr || prog->check_utf8) /* could be removed already */
- BmUSEFUL(do_utf8 ? prog->check_utf8 : prog->check_substr) += 5; /* hooray */
+ BmUSEFUL(utf8_target ? prog->check_utf8 : prog->check_substr) += 5; /* hooray */
fail:
DEBUG_EXECUTE_r(PerlIO_printf(Perl_debug_log, "%sMatch rejected by optimizer%s\n",
PL_colors[4], PL_colors[5]));
#define DECL_TRIE_TYPE(scan) \
const enum { trie_plain, trie_utf8, trie_utf8_fold, trie_latin_utf8_fold } \
trie_type = (scan->flags != EXACT) \
- ? (do_utf8 ? trie_utf8_fold : (UTF ? trie_latin_utf8_fold : trie_plain)) \
- : (do_utf8 ? trie_utf8 : trie_plain)
+ ? (utf8_target ? trie_utf8_fold : (UTF_PATTERN ? trie_latin_utf8_fold : trie_plain)) \
+ : (utf8_target ? trie_utf8 : trie_plain)
#define REXEC_TRIE_READ_CHAR(trie_type, trie, widecharmap, uc, uscan, len, \
uvc, charid, foldlen, foldbuf, uniflags) STMT_START { \
char *my_strend= (char *)strend; \
if ( (CoNd) \
&& (ln == len || \
- foldEQ_utf8(s, &my_strend, 0, do_utf8, \
- m, NULL, ln, cBOOL(UTF))) \
+ foldEQ_utf8(s, &my_strend, 0, utf8_target, \
+ m, NULL, ln, cBOOL(UTF_PATTERN))) \
&& (!reginfo || regtry(reginfo, &s)) ) \
goto got_it; \
else { \
if ( f != c \
&& (f == c1 || f == c2) \
&& (ln == len || \
- foldEQ_utf8(s, &my_strend, 0, do_utf8,\
- m, NULL, ln, cBOOL(UTF)))\
+ foldEQ_utf8(s, &my_strend, 0, utf8_target,\
+ m, NULL, ln, cBOOL(UTF_PATTERN)))\
&& (!reginfo || regtry(reginfo, &s)) ) \
goto got_it; \
} \
#define REXEC_FBC_EXACTISH_SCAN(CoNd) \
STMT_START { \
+ re_fold_t folder; \
+ switch (OP(c)) { \
+ case EXACTFU: folder = foldEQ_latin1; break; \
+ case EXACTFL: folder = foldEQ_locale; break; \
+ case EXACTF: folder = foldEQ; break; \
+ default: \
+ Perl_croak(aTHX_ "panic: Unexpected op %u", OP(c)); \
+ } \
while (s <= e) { \
if ( (CoNd) \
- && (ln == 1 || (OP(c) == EXACTF \
- ? foldEQ(s, m, ln) \
- : foldEQ_locale(s, m, ln))) \
- && (!reginfo || regtry(reginfo, &s)) ) \
+ && (ln == 1 || folder(s, m, ln)) \
+ && (!reginfo || regtry(reginfo, &s)) ) \
goto got_it; \
s++; \
} \
goto got_it
#define REXEC_FBC_CSCAN(CoNdUtF8,CoNd) \
- if (do_utf8) { \
+ if (utf8_target) { \
REXEC_FBC_UTF8_CLASS_SCAN(CoNdUtF8); \
} \
else { \
break
#define REXEC_FBC_CSCAN_PRELOAD(UtFpReLoAd,CoNdUtF8,CoNd) \
- if (do_utf8) { \
+ if (utf8_target) { \
UtFpReLoAd; \
REXEC_FBC_UTF8_CLASS_SCAN(CoNdUtF8); \
} \
#define REXEC_FBC_CSCAN_TAINT(CoNdUtF8,CoNd) \
PL_reg_flags |= RF_tainted; \
- if (do_utf8) { \
+ if (utf8_target) { \
REXEC_FBC_UTF8_CLASS_SCAN(CoNdUtF8); \
} \
else { \
unsigned int c2;
char *e;
register I32 tmp = 1; /* Scratch variable? */
- register const bool do_utf8 = PL_reg_match_utf8;
+ register const bool utf8_target = PL_reg_match_utf8;
RXi_GET_DECL(prog,progi);
PERL_ARGS_ASSERT_FIND_BYCLASS;
/* We know what class it must start with. */
switch (OP(c)) {
case ANYOF:
- if (do_utf8) {
- REXEC_FBC_UTF8_CLASS_SCAN((ANYOF_FLAGS(c) & ANYOF_UNICODE) ||
+ if (utf8_target) {
+ REXEC_FBC_UTF8_CLASS_SCAN((ANYOF_FLAGS(c) & ANYOF_NONBITMAP) ||
!UTF8_IS_INVARIANT((U8)s[0]) ?
- reginclass(prog, c, (U8*)s, 0, do_utf8) :
+ reginclass(prog, c, (U8*)s, 0, utf8_target) :
REGINCLASS(prog, c, (U8*)s));
}
else {
tmp = doevery;
);
break;
+ case EXACTFU:
case EXACTF:
m = STRING(c);
ln = STR_LEN(c); /* length to match in octets/bytes */
lnc = (I32) ln; /* length to match in characters */
- if (UTF) {
+ if (UTF_PATTERN) {
STRLEN ulen1, ulen2;
U8 *sm = (U8 *) m;
U8 tmpbuf1[UTF8_MAXBYTES_CASE+1];
}
else {
c1 = *(U8*)m;
- c2 = PL_fold[c1];
+ if (utf8_target || OP(c) == EXACTFU) {
+
+ /* Micro sign folds to GREEK SMALL LETTER MU;
+ LATIN_SMALL_LETTER_SHARP_S folds to 'ss', and this sets
+ c2 to the first 's' of the pair, and the code below will
+ look for others */
+ c2 = (c1 == MICRO_SIGN)
+ ? GREEK_SMALL_LETTER_MU
+ : (c1 == LATIN_SMALL_LETTER_SHARP_S)
+ ? 's'
+ : PL_fold_latin1[c1];
+ } else c2 = PL_fold[c1];
}
goto do_exactf;
case EXACTFL:
* matching (called "loose matching" in Unicode).
* foldEQ_utf8() will do just that. */
- if (do_utf8 || UTF) {
+ if (utf8_target || UTF_PATTERN) {
UV c, f;
U8 tmpbuf [UTF8_MAXBYTES+1];
STRLEN len = 1;
/* Upper and lower of 1st char are equal -
* probably not a "letter". */
while (s <= e) {
- if (do_utf8) {
+ if (utf8_target) {
c = utf8n_to_uvchr((U8*)s, UTF8_MAXBYTES, &len,
uniflags);
} else {
}
else {
while (s <= e) {
- if (do_utf8) {
+ if (utf8_target) {
c = utf8n_to_uvchr((U8*)s, UTF8_MAXBYTES, &len,
uniflags);
} else {
PL_reg_flags |= RF_tainted;
/* FALL THROUGH */
case BOUND:
- if (do_utf8) {
+ if (utf8_target) {
if (s == PL_bostr)
tmp = '\n';
else {
LOAD_UTF8_CHARCLASS_ALNUM();
REXEC_FBC_UTF8_SCAN(
if (tmp == !(OP(c) == BOUND ?
- cBOOL(swash_fetch(PL_utf8_alnum, (U8*)s, do_utf8)) :
+ cBOOL(swash_fetch(PL_utf8_alnum, (U8*)s, utf8_target)) :
isALNUM_LC_utf8((U8*)s)))
{
tmp = !tmp;
}
);
}
- else {
+ else { /* Not utf8 */
tmp = (s != PL_bostr) ? UCHARAT(s - 1) : '\n';
- tmp = ((OP(c) == BOUND ? isALNUM(tmp) : isALNUM_LC(tmp)) != 0);
+ tmp = cBOOL((OP(c) == BOUNDL)
+ ? isALNUM_LC(tmp)
+ : (isWORDCHAR_L1(tmp)
+ && (isASCII(tmp) || (FLAGS(c) & USE_UNI))));
REXEC_FBC_SCAN(
if (tmp ==
- !(OP(c) == BOUND ? isALNUM(*s) : isALNUM_LC(*s))) {
+ !((OP(c) == BOUNDL)
+ ? isALNUM_LC(*s)
+ : (isWORDCHAR_L1((U8) *s)
+ && (isASCII((U8) *s) || (FLAGS(c) & USE_UNI)))))
+ {
tmp = !tmp;
REXEC_FBC_TRYIT;
}
PL_reg_flags |= RF_tainted;
/* FALL THROUGH */
case NBOUND:
- if (do_utf8) {
+ if (utf8_target) {
if (s == PL_bostr)
tmp = '\n';
else {
LOAD_UTF8_CHARCLASS_ALNUM();
REXEC_FBC_UTF8_SCAN(
if (tmp == !(OP(c) == NBOUND ?
- cBOOL(swash_fetch(PL_utf8_alnum, (U8*)s, do_utf8)) :
+ cBOOL(swash_fetch(PL_utf8_alnum, (U8*)s, utf8_target)) :
isALNUM_LC_utf8((U8*)s)))
tmp = !tmp;
else REXEC_FBC_TRYIT;
}
else {
tmp = (s != PL_bostr) ? UCHARAT(s - 1) : '\n';
- tmp = ((OP(c) == NBOUND ?
- isALNUM(tmp) : isALNUM_LC(tmp)) != 0);
+ tmp = cBOOL((OP(c) == NBOUNDL)
+ ? isALNUM_LC(tmp)
+ : (isWORDCHAR_L1(tmp)
+ && (isASCII(tmp) || (FLAGS(c) & USE_UNI))));
REXEC_FBC_SCAN(
- if (tmp ==
- !(OP(c) == NBOUND ? isALNUM(*s) : isALNUM_LC(*s)))
+ if (tmp == ! cBOOL(
+ (OP(c) == NBOUNDL)
+ ? isALNUM_LC(*s)
+ : (isWORDCHAR_L1((U8) *s)
+ && (isASCII((U8) *s) || (FLAGS(c) & USE_UNI)))))
+ {
tmp = !tmp;
+ }
else REXEC_FBC_TRYIT;
);
}
case ALNUM:
REXEC_FBC_CSCAN_PRELOAD(
LOAD_UTF8_CHARCLASS_PERL_WORD(),
- swash_fetch(RE_utf8_perl_word, (U8*)s, do_utf8),
- isALNUM(*s)
+ swash_fetch(RE_utf8_perl_word, (U8*)s, utf8_target),
+ (FLAGS(c) & USE_UNI) ? isWORDCHAR_L1((U8) *s) : isALNUM(*s)
);
case ALNUML:
REXEC_FBC_CSCAN_TAINT(
case NALNUM:
REXEC_FBC_CSCAN_PRELOAD(
LOAD_UTF8_CHARCLASS_PERL_WORD(),
- !swash_fetch(RE_utf8_perl_word, (U8*)s, do_utf8),
- !isALNUM(*s)
+ !swash_fetch(RE_utf8_perl_word, (U8*)s, utf8_target),
+ ! ((FLAGS(c) & USE_UNI) ? isWORDCHAR_L1((U8) *s) : isALNUM(*s))
);
case NALNUML:
REXEC_FBC_CSCAN_TAINT(
case SPACE:
REXEC_FBC_CSCAN_PRELOAD(
LOAD_UTF8_CHARCLASS_PERL_SPACE(),
- *s == ' ' || swash_fetch(RE_utf8_perl_space,(U8*)s, do_utf8),
- isSPACE(*s)
+ *s == ' ' || swash_fetch(RE_utf8_perl_space,(U8*)s, utf8_target),
+ isSPACE_L1((U8) *s) && (isASCII((U8) *s) || (FLAGS(c) & USE_UNI))
);
case SPACEL:
REXEC_FBC_CSCAN_TAINT(
case NSPACE:
REXEC_FBC_CSCAN_PRELOAD(
LOAD_UTF8_CHARCLASS_PERL_SPACE(),
- !(*s == ' ' || swash_fetch(RE_utf8_perl_space,(U8*)s, do_utf8)),
- !isSPACE(*s)
+ !(*s == ' ' || swash_fetch(RE_utf8_perl_space,(U8*)s, utf8_target)),
+ !(isSPACE_L1((U8) *s) && (isASCII((U8) *s) || (FLAGS(c) & USE_UNI)))
);
case NSPACEL:
REXEC_FBC_CSCAN_TAINT(
case DIGIT:
REXEC_FBC_CSCAN_PRELOAD(
LOAD_UTF8_CHARCLASS_POSIX_DIGIT(),
- swash_fetch(RE_utf8_posix_digit,(U8*)s, do_utf8),
+ swash_fetch(RE_utf8_posix_digit,(U8*)s, utf8_target),
isDIGIT(*s)
);
case DIGITL:
case NDIGIT:
REXEC_FBC_CSCAN_PRELOAD(
LOAD_UTF8_CHARCLASS_POSIX_DIGIT(),
- !swash_fetch(RE_utf8_posix_digit,(U8*)s, do_utf8),
+ !swash_fetch(RE_utf8_posix_digit,(U8*)s, utf8_target),
!isDIGIT(*s)
);
case NDIGITL:
DEBUG_TRIE_EXECUTE_r(
if ( uc <= (U8*)last_start && !BITMAP_TEST(bitmap,*uc) ) {
dump_exec_pos( (char *)uc, c, strend, real_start,
- (char *)uc, do_utf8 );
+ (char *)uc, utf8_target );
PerlIO_printf( Perl_debug_log,
" Scanning for legal start char...\n");
}
- );
- while ( uc <= (U8*)last_start && !BITMAP_TEST(bitmap,*uc) ) {
- uc++;
- }
+ );
+ if (utf8_target) {
+ while ( uc <= (U8*)last_start && !BITMAP_TEST(bitmap,*uc) ) {
+ uc += UTF8SKIP(uc);
+ }
+ } else {
+ while ( uc <= (U8*)last_start && !BITMAP_TEST(bitmap,*uc) ) {
+ uc++;
+ }
+ }
s= (char *)uc;
}
if (uc >(U8*)last_start) break;
foldbuf, uniflags);
DEBUG_TRIE_EXECUTE_r({
dump_exec_pos( (char *)uc, c, strend, real_start,
- s, do_utf8 );
+ s, utf8_target );
PerlIO_printf(Perl_debug_log,
" Charid:%3u CP:%4"UVxf" ",
charid, uvc);
DEBUG_TRIE_EXECUTE_r({
if (failed)
dump_exec_pos( (char *)uc, c, strend, real_start,
- s, do_utf8 );
+ s, utf8_target );
PerlIO_printf( Perl_debug_log,
"%sState: %4"UVxf", word=%"UVxf,
failed ? " Fail transition to " : "",
I32 end_shift = 0; /* Same for the end. */ /* CC */
I32 scream_pos = -1; /* Internal iterator of scream. */
char *scream_olds = NULL;
- const bool do_utf8 = cBOOL(DO_UTF8(sv));
+ const bool utf8_target = cBOOL(DO_UTF8(sv));
I32 multiline;
RXi_GET_DECL(prog,progi);
regmatch_info reginfo; /* create some info to pass to regtry etc */
multiline = prog->extflags & RXf_PMf_MULTILINE;
reginfo.prog = rx; /* Yes, sorry that this is confusing. */
- RX_MATCH_UTF8_set(rx, do_utf8);
+ RX_MATCH_UTF8_set(rx, utf8_target);
DEBUG_EXECUTE_r(
- debug_start_match(rx, do_utf8, startpos, strend,
+ debug_start_match(rx, utf8_target, startpos, strend,
"Matching");
);
end = HOP3c(strend, -dontbother, strbeg) - 1;
/* for multiline we only have to try after newlines */
if (prog->check_substr || prog->check_utf8) {
- if (s == startpos)
- goto after_try;
- while (1) {
- if (regtry(®info, &s))
- goto got_it;
- after_try:
- if (s > end)
- goto phooey;
- if (prog->extflags & RXf_USE_INTUIT) {
- s = re_intuit_start(rx, sv, s + 1, strend, flags, NULL);
- if (!s)
- goto phooey;
- }
- else
- s++;
- }
- } else {
- if (s > startpos)
+ /* because of the goto we can not easily reuse the macros for bifurcating the
+ unicode/non-unicode match modes here like we do elsewhere - demerphq */
+ if (utf8_target) {
+ if (s == startpos)
+ goto after_try_utf8;
+ while (1) {
+ if (regtry(®info, &s)) {
+ goto got_it;
+ }
+ after_try_utf8:
+ if (s > end) {
+ goto phooey;
+ }
+ if (prog->extflags & RXf_USE_INTUIT) {
+ s = re_intuit_start(rx, sv, s + UTF8SKIP(s), strend, flags, NULL);
+ if (!s) {
+ goto phooey;
+ }
+ }
+ else {
+ s += UTF8SKIP(s);
+ }
+ }
+ } /* end search for check string in unicode */
+ else {
+ if (s == startpos) {
+ goto after_try_latin;
+ }
+ while (1) {
+ if (regtry(®info, &s)) {
+ goto got_it;
+ }
+ after_try_latin:
+ if (s > end) {
+ goto phooey;
+ }
+ if (prog->extflags & RXf_USE_INTUIT) {
+ s = re_intuit_start(rx, sv, s + 1, strend, flags, NULL);
+ if (!s) {
+ goto phooey;
+ }
+ }
+ else {
+ s++;
+ }
+ }
+ } /* end search for check string in latin*/
+ } /* end search for check string */
+ else { /* search for newline */
+ if (s > startpos) {
+ /*XXX: The s-- is almost definitely wrong here under unicode - demeprhq*/
s--;
+ }
+ /* We can use a more efficient search as newlines are the same in unicode as they are in latin */
while (s < end) {
if (*s++ == '\n') { /* don't need PL_utf8skip here */
if (regtry(®info, &s))
goto got_it;
}
- }
- }
- }
+ }
+ } /* end search for newline */
+ } /* end anchored/multiline check string search */
goto phooey;
} else if (RXf_GPOS_CHECK == (prog->extflags & RXf_GPOS_CHECK))
{
/* Messy cases: unanchored match. */
if ((prog->anchored_substr || prog->anchored_utf8) && prog->intflags & PREGf_SKIP) {
/* we have /x+whatever/ */
- /* it must be a one character string (XXXX Except UTF?) */
+ /* it must be a one character string (XXXX Except UTF_PATTERN?) */
char ch;
#ifdef DEBUGGING
int did_match = 0;
#endif
- if (!(do_utf8 ? prog->anchored_utf8 : prog->anchored_substr))
- do_utf8 ? to_utf8_substr(prog) : to_byte_substr(prog);
- ch = SvPVX_const(do_utf8 ? prog->anchored_utf8 : prog->anchored_substr)[0];
+ if (!(utf8_target ? prog->anchored_utf8 : prog->anchored_substr))
+ utf8_target ? to_utf8_substr(prog) : to_byte_substr(prog);
+ ch = SvPVX_const(utf8_target ? prog->anchored_utf8 : prog->anchored_substr)[0];
- if (do_utf8) {
+ if (utf8_target) {
REXEC_FBC_SCAN(
if (*s == ch) {
DEBUG_EXECUTE_r( did_match = 1 );
int did_match = 0;
#endif
if (prog->anchored_substr || prog->anchored_utf8) {
- if (!(do_utf8 ? prog->anchored_utf8 : prog->anchored_substr))
- do_utf8 ? to_utf8_substr(prog) : to_byte_substr(prog);
- must = do_utf8 ? prog->anchored_utf8 : prog->anchored_substr;
+ if (!(utf8_target ? prog->anchored_utf8 : prog->anchored_substr))
+ utf8_target ? to_utf8_substr(prog) : to_byte_substr(prog);
+ must = utf8_target ? prog->anchored_utf8 : prog->anchored_substr;
back_max = back_min = prog->anchored_offset;
} else {
- if (!(do_utf8 ? prog->float_utf8 : prog->float_substr))
- do_utf8 ? to_utf8_substr(prog) : to_byte_substr(prog);
- must = do_utf8 ? prog->float_utf8 : prog->float_substr;
+ if (!(utf8_target ? prog->float_utf8 : prog->float_substr))
+ utf8_target ? to_utf8_substr(prog) : to_byte_substr(prog);
+ must = utf8_target ? prog->float_utf8 : prog->float_substr;
back_max = prog->float_max_offset;
back_min = prog->float_min_offset;
}
last1 = HOPc(s, -back_min);
s = t;
}
- if (do_utf8) {
+ if (utf8_target) {
while (s <= last1) {
if (regtry(®info, &s))
goto got_it;
}
}
DEBUG_EXECUTE_r(if (!did_match) {
- RE_PV_QUOTED_DECL(quoted, do_utf8, PERL_DEBUG_PAD_ZERO(0),
+ RE_PV_QUOTED_DECL(quoted, utf8_target, PERL_DEBUG_PAD_ZERO(0),
SvPVX_const(must), RE_SV_DUMPLEN(must), 30);
PerlIO_printf(Perl_debug_log, "Did not find %s substr %s%s...\n",
((must == prog->anchored_substr || must == prog->anchored_utf8)
SV * const prop = sv_newmortal();
regprop(prog, prop, c);
{
- RE_PV_QUOTED_DECL(quoted,do_utf8,PERL_DEBUG_PAD_ZERO(1),
+ RE_PV_QUOTED_DECL(quoted,utf8_target,PERL_DEBUG_PAD_ZERO(1),
s,strend-s,60);
PerlIO_printf(Perl_debug_log,
"Matching stclass %.*s against %s (%d bytes)\n",
char *last;
SV* float_real;
- if (!(do_utf8 ? prog->float_utf8 : prog->float_substr))
- do_utf8 ? to_utf8_substr(prog) : to_byte_substr(prog);
- float_real = do_utf8 ? prog->float_utf8 : prog->float_substr;
+ if (!(utf8_target ? prog->float_utf8 : prog->float_substr))
+ utf8_target ? to_utf8_substr(prog) : to_byte_substr(prog);
+ float_real = utf8_target ? prog->float_utf8 : prog->float_substr;
if (flags & REXEC_SCREAM) {
last = screaminstr(sv, float_real, s - strbeg,
dontbother = minlen - 1;
strend -= dontbother; /* this one's always in bytes! */
/* We don't know much -- general case. */
- if (do_utf8) {
+ if (utf8_target) {
for (;;) {
if (regtry(®info, &s))
goto got_it;
#define DEBUG_STATE_pp(pp) \
DEBUG_STATE_r({ \
- DUMP_EXEC_POS(locinput, scan, do_utf8); \
+ DUMP_EXEC_POS(locinput, scan, utf8_target); \
PerlIO_printf(Perl_debug_log, \
" %*s"pp" %s%s%s%s%s\n", \
depth*2, "", \
#ifdef DEBUGGING
STATIC void
-S_debug_start_match(pTHX_ const REGEXP *prog, const bool do_utf8,
+S_debug_start_match(pTHX_ const REGEXP *prog, const bool utf8_target,
const char *start, const char *end, const char *blurb)
{
const bool utf8_pat = RX_UTF8(prog) ? 1 : 0;
RE_PV_QUOTED_DECL(s0, utf8_pat, PERL_DEBUG_PAD_ZERO(0),
RX_PRECOMP_const(prog), RX_PRELEN(prog), 60);
- RE_PV_QUOTED_DECL(s1, do_utf8, PERL_DEBUG_PAD_ZERO(1),
+ RE_PV_QUOTED_DECL(s1, utf8_target, PERL_DEBUG_PAD_ZERO(1),
start, end - start, 60);
PerlIO_printf(Perl_debug_log,
"%s%s REx%s %s against %s\n",
PL_colors[4], blurb, PL_colors[5], s0, s1);
- if (do_utf8||utf8_pat)
+ if (utf8_target||utf8_pat)
PerlIO_printf(Perl_debug_log, "UTF-8 %s%s%s...\n",
utf8_pat ? "pattern" : "",
- utf8_pat && do_utf8 ? " and " : "",
- do_utf8 ? "string" : ""
+ utf8_pat && utf8_target ? " and " : "",
+ utf8_target ? "string" : ""
);
}
}
const char *loc_regeol,
const char *loc_bostr,
const char *loc_reg_starttry,
- const bool do_utf8)
+ const bool utf8_target)
{
const int docolor = *PL_colors[0] || *PL_colors[2] || *PL_colors[4];
const int taill = (docolor ? 10 : 7); /* 3 chars for "> <" */
PERL_ARGS_ASSERT_DUMP_EXEC_POS;
- while (do_utf8 && UTF8_IS_CONTINUATION(*(U8*)(locinput - pref_len)))
+ while (utf8_target && UTF8_IS_CONTINUATION(*(U8*)(locinput - pref_len)))
pref_len++;
pref0_len = pref_len - (locinput - loc_reg_starttry);
if (l + pref_len < (5 + taill) && l < loc_regeol - locinput)
l = ( loc_regeol - locinput > (5 + taill) - pref_len
? (5 + taill) - pref_len : loc_regeol - locinput);
- while (do_utf8 && UTF8_IS_CONTINUATION(*(U8*)(locinput + l)))
+ while (utf8_target && UTF8_IS_CONTINUATION(*(U8*)(locinput + l)))
l--;
if (pref0_len < 0)
pref0_len = 0;
if (pref0_len > pref_len)
pref0_len = pref_len;
{
- const int is_uni = (do_utf8 && OP(scan) != CANY) ? 1 : 0;
+ const int is_uni = (utf8_target && OP(scan) != CANY) ? 1 : 0;
RE_PV_COLOR_DECL(s0,len0,is_uni,PERL_DEBUG_PAD(0),
(locinput - pref_len),pref0_len, 60, 4, 5);
dMY_CXT;
#endif
dVAR;
- register const bool do_utf8 = PL_reg_match_utf8;
+ register const bool utf8_target = PL_reg_match_utf8;
const U32 uniflags = UTF8_ALLOW_DEFAULT;
REGEXP *rex_sv = reginfo->prog;
regexp *rex = (struct regexp *)SvANY(rex_sv);
DEBUG_EXECUTE_r( {
SV * const prop = sv_newmortal();
regnode *rnext=regnext(scan);
- DUMP_EXEC_POS( locinput, scan, do_utf8 );
+ DUMP_EXEC_POS( locinput, scan, utf8_target );
regprop(rex, prop, scan);
PerlIO_printf(Perl_debug_log,
case SANY:
if (!nextchr && locinput >= PL_regeol)
sayNO;
- if (do_utf8) {
+ if (utf8_target) {
locinput += PL_utf8skip[nextchr];
if (locinput > PL_regeol)
sayNO;
case REG_ANY:
if ((!nextchr && locinput >= PL_regeol) || nextchr == '\n')
sayNO;
- if (do_utf8) {
+ if (utf8_target) {
locinput += PL_utf8skip[nextchr];
if (locinput > PL_regeol)
sayNO;
/* In this case the charclass data is available inline so
we can fail fast without a lot of extra overhead.
*/
- if (scan->flags == EXACT || !do_utf8) {
+ if (scan->flags == EXACT || !utf8_target) {
if(!ANYOF_BITMAP_TEST(scan, *locinput)) {
DEBUG_EXECUTE_r(
PerlIO_printf(Perl_debug_log,
"%*s %smatched empty string...%s\n",
REPORT_CODE_OFF+depth*2, "", PL_colors[4], PL_colors[5])
);
- break;
+ if (!trie->jump)
+ break;
} else {
DEBUG_EXECUTE_r(
PerlIO_printf(Perl_debug_log,
while ( state && uc <= (U8*)PL_regeol ) {
U32 base = trie->states[ state ].trans.base;
UV uvc = 0;
- U16 charid;
+ U16 charid = 0;
U16 wordnum;
wordnum = trie->states[ state ].wordnum;
}
DEBUG_TRIE_EXECUTE_r({
- DUMP_EXEC_POS( (char *)uc, scan, do_utf8 );
+ DUMP_EXEC_POS( (char *)uc, scan, utf8_target );
PerlIO_printf( Perl_debug_log,
"%*s %sState: %4"UVxf" Accepted: %c ",
2+depth * 2, "", PL_colors[4],
U8 *uscan;
while (chars) {
- if (do_utf8) {
+ if (utf8_target) {
uvc = utf8n_to_uvuni((U8*)uc, UTF8_MAXLEN, &len,
uniflags);
uc += len;
}
}
else {
- if (do_utf8)
+ if (utf8_target)
while (chars--)
uc += UTF8SKIP(uc);
else
case EXACT: {
char *s = STRING(scan);
ln = STR_LEN(scan);
- if (do_utf8 != UTF) {
+ if (utf8_target != UTF_PATTERN) {
/* The target and the pattern have differing utf8ness. */
char *l = locinput;
const char * const e = s + ln;
- if (do_utf8) {
+ if (utf8_target) {
/* The target is utf8, the pattern is not utf8. */
while (s < e) {
STRLEN ulen;
nextchr = UCHARAT(locinput);
break;
}
- case EXACTFL:
+ case EXACTFL: {
+ re_fold_t folder;
+ const U8 * fold_array;
+ const char * s;
+
PL_reg_flags |= RF_tainted;
- /* FALL THROUGH */
- case EXACTF: {
- char * const s = STRING(scan);
+ folder = foldEQ_locale;
+ fold_array = PL_fold_locale;
+ goto do_exactf;
+
+ case EXACTFU:
+ folder = foldEQ_latin1;
+ fold_array = PL_fold_latin1;
+ goto do_exactf;
+
+ case EXACTF:
+ folder = foldEQ;
+ fold_array = PL_fold;
+
+ do_exactf:
+ s = STRING(scan);
ln = STR_LEN(scan);
- if (do_utf8 || UTF) {
+ if (utf8_target || UTF_PATTERN) {
/* Either target or the pattern are utf8. */
const char * const l = locinput;
char *e = PL_regeol;
- if (! foldEQ_utf8(s, 0, ln, cBOOL(UTF),
- l, &e, 0, do_utf8)) {
+ if (! foldEQ_utf8(s, 0, ln, cBOOL(UTF_PATTERN),
+ l, &e, 0, utf8_target)) {
/* One more case for the sharp s:
* pack("U0U*", 0xDF) =~ /ss/i,
* the 0xC3 0x9F are the UTF-8
* byte sequence for the U+00DF. */
- if (!(do_utf8 &&
+ if (!(utf8_target &&
toLOWER(s[0]) == 's' &&
ln >= 2 &&
toLOWER(s[1]) == 's' &&
/* Inline the first character, for speed. */
if (UCHARAT(s) != nextchr &&
- UCHARAT(s) != ((OP(scan) == EXACTF)
- ? PL_fold : PL_fold_locale)[nextchr])
+ UCHARAT(s) != fold_array[nextchr])
+ {
sayNO;
+ }
if (PL_regeol - locinput < ln)
sayNO;
- if (ln > 1 && (OP(scan) == EXACTF
- ? ! foldEQ(s, locinput, ln)
- : ! foldEQ_locale(s, locinput, ln)))
+ if (ln > 1 && ! folder(s, locinput, ln))
sayNO;
locinput += ln;
nextchr = UCHARAT(locinput);
break;
- }
+ }
case BOUNDL:
case NBOUNDL:
PL_reg_flags |= RF_tainted;
case BOUND:
case NBOUND:
/* was last char in word? */
- if (do_utf8) {
+ if (utf8_target) {
if (locinput == PL_bostr)
ln = '\n';
else {
if (OP(scan) == BOUND || OP(scan) == NBOUND) {
ln = isALNUM_uni(ln);
LOAD_UTF8_CHARCLASS_ALNUM();
- n = swash_fetch(PL_utf8_alnum, (U8*)locinput, do_utf8);
+ n = swash_fetch(PL_utf8_alnum, (U8*)locinput, utf8_target);
}
else {
ln = isALNUM_LC_uvchr(UNI_TO_NATIVE(ln));
else {
ln = (locinput != PL_bostr) ?
UCHARAT(locinput - 1) : '\n';
- if (OP(scan) == BOUND || OP(scan) == NBOUND) {
+ if (FLAGS(scan) & USE_UNI) {
+
+ /* Here, can't be BOUNDL or NBOUNDL because they never set
+ * the flags to USE_UNI */
+ ln = isWORDCHAR_L1(ln);
+ n = isWORDCHAR_L1(nextchr);
+ }
+ else if (OP(scan) == BOUND || OP(scan) == NBOUND) {
ln = isALNUM(ln);
n = isALNUM(nextchr);
}
sayNO;
break;
case ANYOF:
- if (do_utf8) {
+ if (utf8_target) {
STRLEN inclasslen = PL_regeol - locinput;
-
- if (!reginclass(rex, scan, (U8*)locinput, &inclasslen, do_utf8))
- goto anyof_fail;
if (locinput >= PL_regeol)
sayNO;
- locinput += inclasslen ? inclasslen : UTF8SKIP(locinput);
+
+ if (!reginclass(rex, scan, (U8*)locinput, &inclasslen, utf8_target))
+ goto anyof_fail;
+ locinput += inclasslen;
nextchr = UCHARAT(locinput);
break;
}
else {
if (nextchr < 0)
nextchr = UCHARAT(locinput);
- if (!REGINCLASS(rex, scan, (U8*)locinput))
- goto anyof_fail;
if (!nextchr && locinput >= PL_regeol)
sayNO;
+ if (!REGINCLASS(rex, scan, (U8*)locinput))
+ goto anyof_fail;
nextchr = UCHARAT(++locinput);
break;
}
sayNO;
break;
/* Special char classes - The defines start on line 129 or so */
- CCC_TRY_AFF( ALNUM, ALNUML, perl_word, "a", isALNUM_LC_utf8, isALNUM, isALNUM_LC);
- CCC_TRY_NEG(NALNUM, NALNUML, perl_word, "a", isALNUM_LC_utf8, isALNUM, isALNUM_LC);
+ CCC_TRY_AFF_U( ALNUM, ALNUML, perl_word, "a", isALNUM_LC_utf8, isWORDCHAR_L1, isALNUM_LC);
+ CCC_TRY_NEG_U(NALNUM, NALNUML, perl_word, "a", isALNUM_LC_utf8, isWORDCHAR_L1, isALNUM_LC);
- CCC_TRY_AFF( SPACE, SPACEL, perl_space, " ", isSPACE_LC_utf8, isSPACE, isSPACE_LC);
- CCC_TRY_NEG(NSPACE, NSPACEL, perl_space, " ", isSPACE_LC_utf8, isSPACE, isSPACE_LC);
+ CCC_TRY_AFF_U( SPACE, SPACEL, perl_space, " ", isSPACE_LC_utf8, isSPACE_L1, isSPACE_LC);
+ CCC_TRY_NEG_U(NSPACE, NSPACEL, perl_space, " ", isSPACE_LC_utf8, isSPACE_L1, isSPACE_LC);
CCC_TRY_AFF( DIGIT, DIGITL, posix_digit, "0", isDIGIT_LC_utf8, isDIGIT, isDIGIT_LC);
CCC_TRY_NEG(NDIGIT, NDIGITL, posix_digit, "0", isDIGIT_LC_utf8, isDIGIT, isDIGIT_LC);
if (locinput >= PL_regeol)
sayNO;
- if (! do_utf8) {
+ if (! utf8_target) {
/* Match either CR LF or '.', as all the other possibilities
* require utf8 */
/* Match (prepend)* */
while (locinput < PL_regeol
&& swash_fetch(PL_utf8_X_prepend,
- (U8*)locinput, do_utf8))
+ (U8*)locinput, utf8_target))
{
previous_prepend = locinput;
locinput += UTF8SKIP(locinput);
if (previous_prepend
&& (locinput >= PL_regeol
|| ! swash_fetch(PL_utf8_X_begin,
- (U8*)locinput, do_utf8)))
+ (U8*)locinput, utf8_target)))
{
locinput = previous_prepend;
}
* moved locinput forward, we tested the result just above
* and it either passed, or we backed off so that it will
* now pass */
- if (! swash_fetch(PL_utf8_X_begin, (U8*)locinput, do_utf8)) {
+ if (! swash_fetch(PL_utf8_X_begin, (U8*)locinput, utf8_target)) {
/* Here did not match the required 'Begin' in the
* second term. So just match the very first
* an extender. It is either a hangul syllable, or a
* non-control */
if (swash_fetch(PL_utf8_X_non_hangul,
- (U8*)locinput, do_utf8))
+ (U8*)locinput, utf8_target))
{
/* Here not a Hangul syllable, must be a
* of several individual characters. One
* possibility is T+ */
if (swash_fetch(PL_utf8_X_T,
- (U8*)locinput, do_utf8))
+ (U8*)locinput, utf8_target))
{
while (locinput < PL_regeol
&& swash_fetch(PL_utf8_X_T,
- (U8*)locinput, do_utf8))
+ (U8*)locinput, utf8_target))
{
locinput += UTF8SKIP(locinput);
}
/* Match L* */
while (locinput < PL_regeol
&& swash_fetch(PL_utf8_X_L,
- (U8*)locinput, do_utf8))
+ (U8*)locinput, utf8_target))
{
locinput += UTF8SKIP(locinput);
}
if (locinput < PL_regeol
&& swash_fetch(PL_utf8_X_LV_LVT_V,
- (U8*)locinput, do_utf8))
+ (U8*)locinput, utf8_target))
{
/* Otherwise keep going. Must be LV, LVT
* or V. See if LVT */
if (swash_fetch(PL_utf8_X_LVT,
- (U8*)locinput, do_utf8))
+ (U8*)locinput, utf8_target))
{
locinput += UTF8SKIP(locinput);
} else {
locinput += UTF8SKIP(locinput);
while (locinput < PL_regeol
&& swash_fetch(PL_utf8_X_V,
- (U8*)locinput, do_utf8))
+ (U8*)locinput, utf8_target))
{
locinput += UTF8SKIP(locinput);
}
while (locinput < PL_regeol
&& swash_fetch(PL_utf8_X_T,
(U8*)locinput,
- do_utf8))
+ utf8_target))
{
locinput += UTF8SKIP(locinput);
}
/* Match any extender */
while (locinput < PL_regeol
&& swash_fetch(PL_utf8_X_extend,
- (U8*)locinput, do_utf8))
+ (U8*)locinput, utf8_target))
{
locinput += UTF8SKIP(locinput);
}
break;
case NREFFL:
- {
+ { /* The capture buffer cases. The ones beginning with N for the
+ named buffers just convert to the equivalent numbered and
+ pretend they were called as the corresponding numbered buffer
+ op. */
+ /* don't initialize these, it makes C++ unhappy */
char *s;
char type;
+ re_fold_t folder;
+ const U8 *fold_array;
+
+ folder = NULL; /* NULL assumes will be NREF, REF: no
+ folding */
+ fold_array = NULL;
+
PL_reg_flags |= RF_tainted;
- /* FALL THROUGH */
- case NREF:
+ folder = foldEQ_locale;
+ fold_array = PL_fold_locale;
+ type = REFFL;
+ goto do_nref;
+
+ case NREFFU:
+ folder = foldEQ_latin1;
+ fold_array = PL_fold_latin1;
+ type = REFFU;
+ goto do_nref;
+
case NREFF:
- type = OP(scan);
+ folder = foldEQ;
+ fold_array = PL_fold;
+ type = REFF;
+ goto do_nref;
+
+ case NREF:
+ type = REF;
+ do_nref:
+
+ /* For the named back references, find the corresponding buffer
+ * number */
n = reg_check_named_buff_matched(rex,scan);
- if ( n ) {
- type = REF + ( type - NREF );
- goto do_ref;
- } else {
+ if ( ! n ) {
sayNO;
- }
- /* unreached */
+ }
+ goto do_nref_ref_common;
+
case REFFL:
PL_reg_flags |= RF_tainted;
+ folder = foldEQ_locale;
+ fold_array = PL_fold_locale;
+ goto do_ref;
+
+ case REFFU:
+ folder = foldEQ_latin1;
+ fold_array = PL_fold_latin1;
+ goto do_ref;
+
+ case REFF:
+ folder = foldEQ;
+ fold_array = PL_fold;
/* FALL THROUGH */
+
case REF:
- case REFF:
- n = ARG(scan); /* which paren pair */
+ do_ref:
type = OP(scan);
- do_ref:
+ n = ARG(scan); /* which paren pair */
+
+ do_nref_ref_common:
ln = PL_regoffs[n].start;
PL_reg_leftiter = PL_reg_maxiter; /* Void cache */
if (*PL_reglastparen < n || ln == -1)
break;
s = PL_bostr + ln;
- if (do_utf8 && type != REF) { /* REF can do byte comparison */
- char *l = locinput;
- const char *e = PL_bostr + PL_regoffs[n].end;
- /*
- * Note that we can't do the "other character" lookup trick as
- * in the 8-bit case (no pun intended) because in Unicode we
- * have to map both upper and title case to lower case.
- */
- if (type == REFF) {
- while (s < e) {
- STRLEN ulen1, ulen2;
- U8 tmpbuf1[UTF8_MAXBYTES_CASE+1];
- U8 tmpbuf2[UTF8_MAXBYTES_CASE+1];
-
- if (l >= PL_regeol)
- sayNO;
- toLOWER_utf8((U8*)s, tmpbuf1, &ulen1);
- toLOWER_utf8((U8*)l, tmpbuf2, &ulen2);
- if (ulen1 != ulen2 || memNE((char *)tmpbuf1, (char *)tmpbuf2, ulen1))
- sayNO;
- s += ulen1;
- l += ulen2;
- }
+ if (type != REF /* REF can do byte comparison */
+ && (utf8_target
+ || (type == REFFU
+ && (*s == (char) LATIN_SMALL_LETTER_SHARP_S
+ || *locinput == (char) LATIN_SMALL_LETTER_SHARP_S))))
+ { /* XXX handle REFFL better */
+ char * limit = PL_regeol;
+
+ /* This call case insensitively compares the entire buffer
+ * at s, with the current input starting at locinput, but
+ * not going off the end given by PL_regeol, and returns in
+ * limit upon success, how much of the current input was
+ * matched */
+ if (! foldEQ_utf8(s, NULL, PL_regoffs[n].end - ln, utf8_target,
+ locinput, &limit, 0, utf8_target))
+ {
+ sayNO;
}
- locinput = l;
+ locinput = limit;
nextchr = UCHARAT(locinput);
break;
}
- /* Inline the first character, for speed. */
+ /* Not utf8: Inline the first character, for speed. */
if (UCHARAT(s) != nextchr &&
(type == REF ||
- (UCHARAT(s) != (type == REFF
- ? PL_fold : PL_fold_locale)[nextchr])))
+ UCHARAT(s) != fold_array[nextchr]))
sayNO;
ln = PL_regoffs[n].end - ln;
if (locinput + ln > PL_regeol)
sayNO;
if (ln > 1 && (type == REF
? memNE(s, locinput, ln)
- : (type == REFF
- ? ! foldEQ(s, locinput, ln)
- : ! foldEQ_locale(s, locinput, ln))))
+ : ! folder(s, locinput, ln)))
sayNO;
locinput += ln;
nextchr = UCHARAT(locinput);
COP * const ocurcop = PL_curcop;
PAD *old_comppad;
char *saved_regeol = PL_regeol;
-
+ struct re_save_state saved_state;
+
+ /* To not corrupt the existing regex state while executing the
+ * eval we would normally put it on the save stack, like with
+ * save_re_context. However, re-evals have a weird scoping so we
+ * can't just add ENTER/LEAVE here. With that, things like
+ *
+ * (?{$a=2})(a(?{local$a=$a+1}))*aak*c(?{$b=$a})
+ *
+ * would break, as they expect the localisation to be unwound
+ * only when the re-engine backtracks through the bit that
+ * localised it.
+ *
+ * What we do instead is just saving the state in a local c
+ * variable.
+ */
+ Copy(&PL_reg_state, &saved_state, 1, struct re_save_state);
+
n = ARG(scan);
PL_op = (OP_4tree*)rexi->data->data[n];
DEBUG_STATE_r( PerlIO_printf(Perl_debug_log,
PUTBACK;
}
+ Copy(&saved_state, &PL_reg_state, 1, struct re_save_state);
+
PL_op = oop;
PAD_RESTORE_LOCAL(old_comppad);
PL_curcop = ocurcop;
re->sublen = rex->sublen;
rei = RXi_GET(re);
DEBUG_EXECUTE_r(
- debug_start_match(re_sv, do_utf8, locinput, PL_regeol,
+ debug_start_match(re_sv, utf8_target, locinput, PL_regeol,
"Matching embedded");
);
startpoint = rei->program + 1;
{
ST.c1 = (U8)*STRING(text_node);
- ST.c2 =
- (IS_TEXTF(text_node))
- ? PL_fold[ST.c1]
- : (IS_TEXTFL(text_node))
- ? PL_fold_locale[ST.c1]
- : ST.c1;
+ switch (OP(text_node)) {
+ case EXACTF: ST.c2 = PL_fold[ST.c1]; break;
+ case EXACTFU: ST.c2 = PL_fold_latin1[ST.c1]; break;
+ case EXACTFL: ST.c2 = PL_fold_locale[ST.c1]; break;
+ default: ST.c2 = ST.c1;
+ }
}
}
}
if this changes back then the macro for IS_TEXT and
friends need to change. */
- if (!UTF) {
- ST.c2 = ST.c1 = *s;
- if (IS_TEXTF(text_node))
- ST.c2 = PL_fold[ST.c1];
- else if (IS_TEXTFL(text_node))
- ST.c2 = PL_fold_locale[ST.c1];
+ if (!UTF_PATTERN) {
+ ST.c1 = *s;
+ switch (OP(text_node)) {
+ case EXACTF: ST.c2 = PL_fold[ST.c1]; break;
+ case EXACTFU: ST.c2 = PL_fold_latin1[ST.c1]; break;
+ case EXACTFL: ST.c2 = PL_fold_locale[ST.c1]; break;
+ default: ST.c2 = ST.c1; break;
+ }
}
- else { /* UTF */
- if (IS_TEXTF(text_node)) {
+ else { /* UTF_PATTERN */
+ if (IS_TEXTFU(text_node) || IS_TEXTF(text_node)) {
STRLEN ulen1, ulen2;
U8 tmpbuf1[UTF8_MAXBYTES_CASE+1];
U8 tmpbuf2[UTF8_MAXBYTES_CASE+1];
* string that could possibly match */
if (ST.max == REG_INFTY) {
ST.maxpos = PL_regeol - 1;
- if (do_utf8)
+ if (utf8_target)
while (UTF8_IS_CONTINUATION(*(U8*)ST.maxpos))
ST.maxpos--;
}
- else if (do_utf8) {
+ else if (utf8_target) {
int m = ST.max - ST.min;
for (ST.maxpos = locinput;
m >0 && ST.maxpos + UTF8SKIP(ST.maxpos) <= PL_regeol; m--)
REGCP_UNWIND(ST.cp);
/* Couldn't or didn't -- move forward. */
ST.oldloc = locinput;
- if (do_utf8)
+ if (utf8_target)
locinput += UTF8SKIP(locinput);
else
locinput++;
/* find the next place where 'B' could work, then call B */
{
int n;
- if (do_utf8) {
+ if (utf8_target) {
n = (ST.oldloc == locinput) ? 0 : 1;
if (ST.c1 == ST.c2) {
STRLEN len;
{
UV c = 0;
if (ST.c1 != CHRTEST_VOID)
- c = do_utf8 ? utf8n_to_uvchr((U8*)PL_reginput,
+ c = utf8_target ? utf8n_to_uvchr((U8*)PL_reginput,
UTF8_MAXBYTES, 0, uniflags)
: (UV) UCHARAT(PL_reginput);
/* If it could work, try it. */
#undef ST
case FOLDCHAR:
n = ARG(scan);
- if ( n == (U32)what_len_TRICKYFOLD(locinput,do_utf8,ln) ) {
+ if ( n == (U32)what_len_TRICKYFOLD(locinput,utf8_target,ln) ) {
locinput += ln;
- } else if ( 0xDF == n && !do_utf8 && !UTF ) {
+ } else if ( LATIN_SMALL_LETTER_SHARP_S == n && !utf8_target && !UTF_PATTERN ) {
sayNO;
} else {
U8 folded[UTF8_MAXBYTES_CASE+1];
to_uni_fold(n, folded, &foldlen);
if (! foldEQ_utf8((const char*) folded, 0, foldlen, 1,
- l, &e, 0, do_utf8)) {
+ l, &e, 0, utf8_target)) {
sayNO;
}
locinput = e;
nextchr = UCHARAT(locinput);
break;
case LNBREAK:
- if ((n=is_LNBREAK(locinput,do_utf8))) {
+ if ((n=is_LNBREAK(locinput,utf8_target))) {
locinput += n;
nextchr = UCHARAT(locinput);
} else
#define CASE_CLASS(nAmE) \
case nAmE: \
- if ((n=is_##nAmE(locinput,do_utf8))) { \
+ if ((n=is_##nAmE(locinput,utf8_target))) { \
locinput += n; \
nextchr = UCHARAT(locinput); \
} else \
sayNO; \
break; \
case N##nAmE: \
- if ((n=is_##nAmE(locinput,do_utf8))) { \
+ if ((n=is_##nAmE(locinput,utf8_target))) { \
sayNO; \
} else { \
locinput += UTF8SKIP(locinput); \
register I32 c;
register char *loceol = PL_regeol;
register I32 hardcount = 0;
- register bool do_utf8 = PL_reg_match_utf8;
+ register bool utf8_target = PL_reg_match_utf8;
#ifndef DEBUGGING
PERL_UNUSED_ARG(depth);
#endif
loceol = scan + max;
switch (OP(p)) {
case REG_ANY:
- if (do_utf8) {
+ if (utf8_target) {
loceol = PL_regeol;
while (scan < loceol && hardcount < max && *scan != '\n') {
scan += UTF8SKIP(scan);
}
break;
case SANY:
- if (do_utf8) {
+ if (utf8_target) {
loceol = PL_regeol;
while (scan < loceol && hardcount < max) {
scan += UTF8SKIP(scan);
case CANY:
scan = loceol;
break;
- case EXACT: /* length of string is 1 */
- c = (U8)*STRING(p);
- while (scan < loceol && UCHARAT(scan) == c)
- scan++;
- break;
- case EXACTF: /* length of string is 1 */
+ case EXACT:
+ /* To get here, EXACTish nodes must have *byte* length == 1. That
+ * means they match only characters in the string that can be expressed
+ * as a single byte. For non-utf8 strings, that means a simple match.
+ * For utf8 strings, the character matched must be an invariant, or
+ * downgradable to a single byte. The pattern's utf8ness is
+ * irrelevant, as since it's a single byte, it either isn't utf8, or if
+ * it is, it's an invariant */
+
c = (U8)*STRING(p);
- while (scan < loceol &&
- (UCHARAT(scan) == c || UCHARAT(scan) == PL_fold[c]))
- scan++;
+ assert(! UTF_PATTERN || UNI_IS_INVARIANT(c));
+
+ if (! utf8_target || UNI_IS_INVARIANT(c)) {
+ while (scan < loceol && UCHARAT(scan) == c) {
+ scan++;
+ }
+ }
+ else {
+
+ /* Here, the string is utf8, and the pattern char is different
+ * in utf8 than not, so can't compare them directly. Outside the
+ * loop, find find the two utf8 bytes that represent c, and then
+ * look for those in sequence in the utf8 string */
+ U8 high = UTF8_TWO_BYTE_HI(c);
+ U8 low = UTF8_TWO_BYTE_LO(c);
+ loceol = PL_regeol;
+
+ while (hardcount < max
+ && scan + 1 < loceol
+ && UCHARAT(scan) == high
+ && UCHARAT(scan + 1) == low)
+ {
+ scan += 2;
+ hardcount++;
+ }
+ }
break;
- case EXACTFL: /* length of string is 1 */
+ case EXACTFL:
PL_reg_flags |= RF_tainted;
+ /* FALL THROUGH */
+ case EXACTF:
+ case EXACTFU:
+
+ /* The comments for the EXACT case above apply as well to these fold
+ * ones */
+
c = (U8)*STRING(p);
- while (scan < loceol &&
- (UCHARAT(scan) == c || UCHARAT(scan) == PL_fold_locale[c]))
- scan++;
+ assert(! UTF_PATTERN || UNI_IS_INVARIANT(c));
+
+ if (utf8_target) { /* Use full Unicode fold matching */
+
+ /* For the EXACTFL case, It doesn't really make sense to compare
+ * locale and utf8, but it is best we can do. The documents warn
+ * against mixing them */
+
+ char *tmpeol = loceol;
+ while (hardcount < max
+ && foldEQ_utf8(scan, &tmpeol, 0, utf8_target,
+ STRING(p), NULL, 1, UTF_PATTERN))
+ {
+ scan = tmpeol;
+ tmpeol = loceol;
+ hardcount++;
+ }
+
+ /* XXX Note that the above handles properly the German sharp s in
+ * the pattern matching ss in the string. But it doesn't handle
+ * properly cases where the string contains say 'LIGATURE ff' and
+ * the pattern is 'f+'. This would require, say, a new function or
+ * revised interface to foldEQ_utf8(), in which the maximum number
+ * of characters to match could be passed and it would return how
+ * many actually did. This is just one of many cases where
+ * multi-char folds don't work properly, and so the fix is being
+ * deferred */
+ }
+ else {
+ U8 folded;
+
+ /* Here, the string isn't utf8 and c is a single byte; and either
+ * the pattern isn't utf8 or c is an invariant, so its utf8ness
+ * doesn't affect c. Can just do simple comparisons for exact or
+ * fold matching. */
+ switch (OP(p)) {
+ case EXACTF: folded = PL_fold[c]; break;
+ case EXACTFU: folded = PL_fold_latin1[c]; break;
+ case EXACTFL: folded = PL_fold_locale[c]; break;
+ default: Perl_croak(aTHX_ "panic: Unexpected op %u", OP(p));
+ }
+ while (scan < loceol &&
+ (UCHARAT(scan) == c || UCHARAT(scan) == folded))
+ {
+ scan++;
+ }
+ }
break;
case ANYOF:
- if (do_utf8) {
+ if (utf8_target) {
loceol = PL_regeol;
while (hardcount < max && scan < loceol &&
- reginclass(prog, p, (U8*)scan, 0, do_utf8)) {
+ reginclass(prog, p, (U8*)scan, 0, utf8_target)) {
scan += UTF8SKIP(scan);
hardcount++;
}
}
break;
case ALNUM:
- if (do_utf8) {
+ if (utf8_target) {
loceol = PL_regeol;
LOAD_UTF8_CHARCLASS_ALNUM();
while (hardcount < max && scan < loceol &&
- swash_fetch(PL_utf8_alnum, (U8*)scan, do_utf8)) {
+ swash_fetch(PL_utf8_alnum, (U8*)scan, utf8_target))
+ {
scan += UTF8SKIP(scan);
hardcount++;
}
+ } else if (FLAGS(p) & USE_UNI) {
+ while (scan < loceol && isWORDCHAR_L1((U8) *scan)) {
+ scan++;
+ }
} else {
- while (scan < loceol && isALNUM(*scan))
- scan++;
+ while (scan < loceol && isALNUM((U8) *scan)) {
+ scan++;
+ }
}
break;
case ALNUML:
PL_reg_flags |= RF_tainted;
- if (do_utf8) {
+ if (utf8_target) {
loceol = PL_regeol;
while (hardcount < max && scan < loceol &&
isALNUM_LC_utf8((U8*)scan)) {
}
break;
case NALNUM:
- if (do_utf8) {
+ if (utf8_target) {
loceol = PL_regeol;
LOAD_UTF8_CHARCLASS_ALNUM();
while (hardcount < max && scan < loceol &&
- !swash_fetch(PL_utf8_alnum, (U8*)scan, do_utf8)) {
+ !swash_fetch(PL_utf8_alnum, (U8*)scan, utf8_target))
+ {
scan += UTF8SKIP(scan);
hardcount++;
}
+ } else if (FLAGS(p) & USE_UNI) {
+ while (scan < loceol && ! isWORDCHAR_L1((U8) *scan)) {
+ scan++;
+ }
} else {
- while (scan < loceol && !isALNUM(*scan))
- scan++;
+ while (scan < loceol && ! isALNUM((U8) *scan)) {
+ scan++;
+ }
}
break;
case NALNUML:
PL_reg_flags |= RF_tainted;
- if (do_utf8) {
+ if (utf8_target) {
loceol = PL_regeol;
while (hardcount < max && scan < loceol &&
!isALNUM_LC_utf8((U8*)scan)) {
}
break;
case SPACE:
- if (do_utf8) {
+ if (utf8_target) {
loceol = PL_regeol;
LOAD_UTF8_CHARCLASS_SPACE();
while (hardcount < max && scan < loceol &&
(*scan == ' ' ||
- swash_fetch(PL_utf8_space,(U8*)scan, do_utf8))) {
+ swash_fetch(PL_utf8_space,(U8*)scan, utf8_target)))
+ {
scan += UTF8SKIP(scan);
hardcount++;
}
+ } else if (FLAGS(p) & USE_UNI) {
+ while (scan < loceol && isSPACE_L1((U8) *scan)) {
+ scan++;
+ }
} else {
- while (scan < loceol && isSPACE(*scan))
- scan++;
+ while (scan < loceol && isSPACE((U8) *scan))
+ scan++;
}
break;
case SPACEL:
PL_reg_flags |= RF_tainted;
- if (do_utf8) {
+ if (utf8_target) {
loceol = PL_regeol;
while (hardcount < max && scan < loceol &&
(*scan == ' ' || isSPACE_LC_utf8((U8*)scan))) {
}
break;
case NSPACE:
- if (do_utf8) {
+ if (utf8_target) {
loceol = PL_regeol;
LOAD_UTF8_CHARCLASS_SPACE();
while (hardcount < max && scan < loceol &&
!(*scan == ' ' ||
- swash_fetch(PL_utf8_space,(U8*)scan, do_utf8))) {
+ swash_fetch(PL_utf8_space,(U8*)scan, utf8_target)))
+ {
scan += UTF8SKIP(scan);
hardcount++;
}
+ } else if (FLAGS(p) & USE_UNI) {
+ while (scan < loceol && ! isSPACE_L1((U8) *scan)) {
+ scan++;
+ }
} else {
- while (scan < loceol && !isSPACE(*scan))
- scan++;
+ while (scan < loceol && ! isSPACE((U8) *scan)) {
+ scan++;
+ }
}
break;
case NSPACEL:
PL_reg_flags |= RF_tainted;
- if (do_utf8) {
+ if (utf8_target) {
loceol = PL_regeol;
while (hardcount < max && scan < loceol &&
!(*scan == ' ' || isSPACE_LC_utf8((U8*)scan))) {
}
break;
case DIGIT:
- if (do_utf8) {
+ if (utf8_target) {
loceol = PL_regeol;
LOAD_UTF8_CHARCLASS_DIGIT();
while (hardcount < max && scan < loceol &&
- swash_fetch(PL_utf8_digit, (U8*)scan, do_utf8)) {
+ swash_fetch(PL_utf8_digit, (U8*)scan, utf8_target)) {
scan += UTF8SKIP(scan);
hardcount++;
}
}
break;
case NDIGIT:
- if (do_utf8) {
+ if (utf8_target) {
loceol = PL_regeol;
LOAD_UTF8_CHARCLASS_DIGIT();
while (hardcount < max && scan < loceol &&
- !swash_fetch(PL_utf8_digit, (U8*)scan, do_utf8)) {
+ !swash_fetch(PL_utf8_digit, (U8*)scan, utf8_target)) {
scan += UTF8SKIP(scan);
hardcount++;
}
scan++;
}
case LNBREAK:
- if (do_utf8) {
+ if (utf8_target) {
loceol = PL_regeol;
while (hardcount < max && scan < loceol && (c=is_LNBREAK_utf8(scan))) {
scan += c;
}
break;
case HORIZWS:
- if (do_utf8) {
+ if (utf8_target) {
loceol = PL_regeol;
while (hardcount < max && scan < loceol && (c=is_HORIZWS_utf8(scan))) {
scan += c;
}
break;
case NHORIZWS:
- if (do_utf8) {
+ if (utf8_target) {
loceol = PL_regeol;
while (hardcount < max && scan < loceol && !is_HORIZWS_utf8(scan)) {
scan += UTF8SKIP(scan);
}
break;
case VERTWS:
- if (do_utf8) {
+ if (utf8_target) {
loceol = PL_regeol;
while (hardcount < max && scan < loceol && (c=is_VERTWS_utf8(scan))) {
scan += c;
}
break;
case NVERTWS:
- if (do_utf8) {
+ if (utf8_target) {
loceol = PL_regeol;
while (hardcount < max && scan < loceol && !is_VERTWS_utf8(scan)) {
scan += UTF8SKIP(scan);
/*
- reginclass - determine if a character falls into a character class
- The n is the ANYOF regnode, the p is the target string, lenp
- is pointer to the maximum length of how far to go in the p
- (if the lenp is zero, UTF8SKIP(p) is used),
- do_utf8 tells whether the target string is in UTF-8.
+ n is the ANYOF regnode
+ p is the target string
+ lenp is pointer to the maximum number of bytes of how far to go in p
+ (This is assumed wthout checking to always be at least the current
+ character's size)
+ utf8_target tells whether p is in UTF-8.
+
+ Returns true if matched; false otherwise. If lenp is not NULL, on return
+ from a successful match, the value it points to will be updated to how many
+ bytes in p were matched. If there was no match, the value is undefined,
+ possibly changed from the input.
*/
STATIC bool
-S_reginclass(pTHX_ const regexp *prog, register const regnode *n, register const U8* p, STRLEN* lenp, register bool do_utf8)
+S_reginclass(pTHX_ const regexp * const prog, register const regnode * const n, register const U8* const p, STRLEN* lenp, register const bool utf8_target)
{
dVAR;
const char flags = ANYOF_FLAGS(n);
bool match = FALSE;
UV c = *p;
- STRLEN len = 0;
- STRLEN plen;
+ STRLEN c_len = 0;
+ STRLEN maxlen;
PERL_ARGS_ASSERT_REGINCLASS;
- if (do_utf8 && !UTF8_IS_INVARIANT(c)) {
- c = utf8n_to_uvchr(p, UTF8_MAXBYTES, &len,
+ /* If c is not already the code point, get it */
+ if (utf8_target && !UTF8_IS_INVARIANT(c)) {
+ c = utf8n_to_uvchr(p, UTF8_MAXBYTES, &c_len,
(UTF8_ALLOW_DEFAULT & UTF8_ALLOW_ANYUV)
| UTF8_ALLOW_FFFF | UTF8_CHECK_ONLY);
/* see [perl #37836] for UTF8_ALLOW_ANYUV; [perl #38293] for
* UTF8_ALLOW_FFFF */
- if (len == (STRLEN)-1)
+ if (c_len == (STRLEN)-1)
Perl_croak(aTHX_ "Malformed UTF-8 character (fatal)");
}
+ else {
+ c_len = 1;
+ }
- plen = lenp ? *lenp : UNISKIP(NATIVE_TO_UNI(c));
- if (do_utf8 || (flags & ANYOF_UNICODE)) {
- if (lenp)
- *lenp = 0;
- if (do_utf8 && !ANYOF_RUNTIME(n)) {
- if (len != (STRLEN)-1 && c < 256 && ANYOF_BITMAP_TEST(n, c))
- match = TRUE;
- }
- if (!match && do_utf8 && (flags & ANYOF_UNICODE_ALL) && c >= 256)
- match = TRUE;
- if (!match) {
- AV *av;
- SV * const sw = regclass_swash(prog, n, TRUE, 0, (SV**)&av);
-
- if (sw) {
- U8 * utf8_p;
- if (do_utf8) {
- utf8_p = (U8 *) p;
- } else {
- STRLEN len = 1;
- utf8_p = bytes_to_utf8(p, &len);
- }
- if (swash_fetch(sw, utf8_p, 1))
- match = TRUE;
- else if (flags & ANYOF_FOLD) {
- if (!match && lenp && av) {
- I32 i;
- for (i = 0; i <= av_len(av); i++) {
- SV* const sv = *av_fetch(av, i, FALSE);
- STRLEN len;
- const char * const s = SvPV_const(sv, len);
- if (len <= plen && memEQ(s, (char*)utf8_p, len)) {
- *lenp = len;
- match = TRUE;
- break;
- }
- }
- }
- if (!match) {
- U8 tmpbuf[UTF8_MAXBYTES_CASE+1];
-
- STRLEN tmplen;
- to_utf8_fold(utf8_p, tmpbuf, &tmplen);
- if (swash_fetch(sw, tmpbuf, 1))
- match = TRUE;
- }
- }
+ /* Use passed in max length, or one character if none passed in or less
+ * than one character. And assume will match just one character. This is
+ * overwritten later if matched more. */
+ if (lenp) {
+ maxlen = (*lenp > c_len) ? *lenp : c_len;
+ *lenp = c_len;
- /* If we allocated a string above, free it */
- if (! do_utf8) Safefree(utf8_p);
- }
- }
- if (match && lenp && *lenp == 0)
- *lenp = UNISKIP(NATIVE_TO_UNI(c));
}
- if (!match && c < 256) {
+ else {
+ maxlen = c_len;
+ }
+
+ /* If this character is potentially in the bitmap, check it */
+ if (c < 256) {
if (ANYOF_BITMAP_TEST(n, c))
match = TRUE;
else if (flags & ANYOF_FOLD) {
match = TRUE;
}
- if (!match && (flags & ANYOF_CLASS)) {
+ if (!match && (flags & ANYOF_CLASS) && ANYOF_CLASS_TEST_ANY_SET(n)) {
PL_reg_flags |= RF_tainted;
if (
(ANYOF_CLASS_TEST(n, ANYOF_ALNUM) && isALNUM_LC(c)) ||
}
}
+ /* If the bitmap didn't (or couldn't) match, and something outside the
+ * bitmap could match, try that */
+ if (!match) {
+ if (utf8_target && (flags & ANYOF_UNICODE_ALL)) {
+ if (c >= 256
+ || ((flags & ANYOF_FOLD) /* Latin1 1 that has a non-Latin1 fold
+ should match */
+ && _HAS_NONLATIN1_FOLD_CLOSURE_ONLY_FOR_USE_BY_REGCOMP_DOT_C_AND_REGEXEC_DOT_C(c)))
+ {
+ match = TRUE;
+ }
+ }
+ if (!match && ((flags & ANYOF_NONBITMAP_NON_UTF8)
+ || (utf8_target && flags & ANYOF_UTF8)))
+ {
+ AV *av;
+ SV * const sw = regclass_swash(prog, n, TRUE, 0, (SV**)&av);
+
+ if (sw) {
+ U8 * utf8_p;
+ if (utf8_target) {
+ utf8_p = (U8 *) p;
+ } else {
+ STRLEN len = 1;
+ utf8_p = bytes_to_utf8(p, &len);
+ }
+ if (swash_fetch(sw, utf8_p, 1))
+ match = TRUE;
+ else if (flags & ANYOF_FOLD) {
+ if (!match && lenp && av) {
+ I32 i;
+ for (i = 0; i <= av_len(av); i++) {
+ SV* const sv = *av_fetch(av, i, FALSE);
+ STRLEN len;
+ const char * const s = SvPV_const(sv, len);
+ if (len <= maxlen && memEQ(s, (char*)utf8_p, len)) {
+ *lenp = len;
+ match = TRUE;
+ break;
+ }
+ }
+ }
+ if (!match) {
+ U8 folded[UTF8_MAXBYTES_CASE+1];
+
+ /* See if the folded version matches */
+ STRLEN foldlen;
+ to_utf8_fold(utf8_p, folded, &foldlen);
+ if (swash_fetch(sw, folded, 1)) { /* 1 => is utf8 */
+ match = TRUE;
+ }
+ else {
+ /* The fold in a few cases of an above Latin1 char
+ * is in the Latin1 range, and hence may be in the
+ * bitmap */
+ if (UTF8_IS_INVARIANT(*folded)
+ && ANYOF_BITMAP_TEST(n, UNI_TO_NATIVE(*folded)))
+ {
+ match = TRUE;
+ }
+ else if (UTF8_IS_DOWNGRADEABLE_START(*folded)
+ && ANYOF_BITMAP_TEST(n,
+ UNI_TO_NATIVE(
+ TWO_BYTE_UTF8_TO_UNI(folded[0],
+ folded[1]))))
+ { /* Since the fold comes from internally
+ * generated data, we can safely assume it is
+ * valid utf8 in the test above */
+
+ match = TRUE;
+ }
+ if (! match) {
+ SV** listp;
+
+ /* Consider "k" =~ /[K]/i. The line above
+ * would have just folded the 'k' to itself,
+ * and that isn't going to match 'K'. So we
+ * look through the closure of everything that
+ * folds to 'k'. That will find the 'K'.
+ * Initialize the list, if necessary */
+ if (! PL_utf8_foldclosures) {
+
+ /* If the folds haven't been read in, call a
+ * fold function to force that */
+ if (! PL_utf8_tofold) {
+ U8 dummy[UTF8_MAXBYTES+1];
+ STRLEN dummy_len;
+ to_utf8_fold((U8*) "A",
+ dummy, &dummy_len);
+ }
+ PL_utf8_foldclosures =
+ _swash_inversion_hash(PL_utf8_tofold);
+ }
+
+ /* The data structure is a hash with the keys
+ * every character that is folded to, like 'k',
+ * and the values each an array of everything
+ * that folds to its key. e.g. [ 'k', 'K',
+ * KELVIN_SIGN ] */
+ if ((listp = hv_fetch(PL_utf8_foldclosures,
+ (char *) folded, foldlen, FALSE)))
+ {
+ AV* list = (AV*) *listp;
+ IV i;
+ for (i = 0; i <= av_len(list); i++) {
+ SV** try_p = av_fetch(list, i, FALSE);
+ char* try_c;
+ if (try_p == NULL) {
+ Perl_croak(aTHX_ "panic: invalid PL_utf8_foldclosures structure");
+ }
+ /* Don't have to worry about embeded
+ * nulls since NULL isn't folded or
+ * foldable */
+ try_c = SvPVX(*try_p);
+ if (UTF8_IS_INVARIANT(*try_c)
+ && ANYOF_BITMAP_TEST(n,
+ UNI_TO_NATIVE(*try_c)))
+ {
+ match = TRUE;
+ break;
+ }
+ else if
+ (UTF8_IS_DOWNGRADEABLE_START(*try_c)
+ && ANYOF_BITMAP_TEST(n,
+ UNI_TO_NATIVE(
+ TWO_BYTE_UTF8_TO_UNI(try_c[0],
+ try_c[1]))))
+ {
+ match = TRUE;
+ break;
+ } else if (swash_fetch(sw,
+ (U8*) try_c, 1))
+ {
+ match = TRUE;
+ break;
+ }
+ }
+ }
+ }
+ }
+ }
+ }
+
+ /* If we allocated a string above, free it */
+ if (! utf8_target) Safefree(utf8_p);
+ }
+ }
+ }
+
return (flags & ANYOF_INVERT) ? !match : match;
}