#define FAIL(msg) \
STMT_START { \
char *ellipses = ""; \
- unsigned len = strlen(RExC_precomp); \
+ IV len = RExC_end - RExC_precomp; \
\
if (!SIZE_ONLY) \
SAVEDESTRUCTOR_X(clear_re,(void*)RExC_rx); \
#define FAIL2(pat,msg) \
STMT_START { \
char *ellipses = ""; \
- unsigned len = strlen(RExC_precomp); \
+ IV len = RExC_end - RExC_precomp; \
\
if (!SIZE_ONLY) \
SAVEDESTRUCTOR_X(clear_re,(void*)RExC_rx); \
*/
#define Simple_vFAIL(m) \
STMT_START { \
- unsigned offset = strlen(RExC_precomp)-(RExC_end-RExC_parse); \
+ IV offset = RExC_parse - RExC_precomp; \
\
Perl_croak(aTHX_ "%s" REPORT_LOCATION, \
m, (int)offset, RExC_precomp, RExC_precomp + offset); \
*/
#define Simple_vFAIL2(m,a1) \
STMT_START { \
- unsigned offset = strlen(RExC_precomp)-(RExC_end-RExC_parse); \
+ IV offset = RExC_parse - RExC_precomp; \
\
S_re_croak2(aTHX_ m, REPORT_LOCATION, a1, \
(int)offset, RExC_precomp, RExC_precomp + offset); \
*/
#define Simple_vFAIL3(m, a1, a2) \
STMT_START { \
- unsigned offset = strlen(RExC_precomp)-(RExC_end-RExC_parse); \
+ IV offset = RExC_parse - RExC_precomp; \
\
S_re_croak2(aTHX_ m, REPORT_LOCATION, a1, a2, \
(int)offset, RExC_precomp, RExC_precomp + offset); \
*/
#define Simple_vFAIL4(m, a1, a2, a3) \
STMT_START { \
- unsigned offset = strlen(RExC_precomp)-(RExC_end-RExC_parse); \
+ IV offset = RExC_parse - RExC_precomp; \
\
S_re_croak2(aTHX_ m, REPORT_LOCATION, a1, a2, a3,\
(int)offset, RExC_precomp, RExC_precomp + offset); \
*/
#define Simple_vFAIL5(m, a1, a2, a3, a4) \
STMT_START { \
- unsigned offset = strlen(RExC_precomp)-(RExC_end-RExC_parse); \
+ IV offset = RExC_parse - RExC_precomp; \
S_re_croak2(aTHX_ m, REPORT_LOCATION, a1, a2, a3, a4,\
(int)offset, RExC_precomp, RExC_precomp + offset); \
} STMT_END
#define vWARN(loc,m) \
STMT_START { \
- unsigned offset = strlen(RExC_precomp)-(RExC_end-(loc)); \
+ IV offset = loc - RExC_precomp; \
Perl_warner(aTHX_ WARN_REGEXP, "%s" REPORT_LOCATION,\
m, (int)offset, RExC_precomp, RExC_precomp + offset); \
} STMT_END \
#define vWARNdep(loc,m) \
STMT_START { \
- unsigned offset = strlen(RExC_precomp)-(RExC_end-(loc)); \
+ IV offset = loc - RExC_precomp; \
int warn_cat = ckWARN(WARN_REGEXP) ? WARN_REGEXP : WARN_DEPRECATED; \
Perl_warner(aTHX_ warn_cat, "%s" REPORT_LOCATION,\
m, (int)offset, RExC_precomp, RExC_precomp + offset); \
#define vWARN2(loc, m, a1) \
STMT_START { \
- unsigned offset = strlen(RExC_precomp)-(RExC_end-(loc)); \
+ IV offset = loc - RExC_precomp; \
Perl_warner(aTHX_ WARN_REGEXP, m REPORT_LOCATION,\
a1, \
(int)offset, RExC_precomp, RExC_precomp + offset); \
#define vWARN3(loc, m, a1, a2) \
STMT_START { \
- unsigned offset = strlen(RExC_precomp) - (RExC_end - (loc)); \
+ IV offset = loc - RExC_precomp; \
Perl_warner(aTHX_ WARN_REGEXP, m REPORT_LOCATION, \
a1, a2, \
(int)offset, RExC_precomp, RExC_precomp + offset); \
#define vWARN4(loc, m, a1, a2, a3) \
STMT_START { \
- unsigned offset = strlen(RExC_precomp)-(RExC_end-(loc)); \
+ IV offset = loc - RExC_precomp; \
Perl_warner(aTHX_ WARN_REGEXP, m REPORT_LOCATION,\
a1, a2, a3, \
(int)offset, RExC_precomp, RExC_precomp + offset); \
/* used for the parse_flags section for (?c) -- japhy */
#define vWARN5(loc, m, a1, a2, a3, a4) \
STMT_START { \
- unsigned offset = strlen(RExC_precomp)-(RExC_end-(loc)); \
+ IV offset = loc - RExC_precomp; \
Perl_warner(aTHX_ WARN_REGEXP, m REPORT_LOCATION, \
a1, a2, a3, a4, \
(int)offset, RExC_precomp, RExC_precomp + offset); \
STATIC void
S_cl_anything(pTHX_ RExC_state_t *pRExC_state, struct regnode_charclass_class *cl)
{
- int value;
-
ANYOF_CLASS_ZERO(cl);
- for (value = 0; value < 256; ++value)
- ANYOF_BITMAP_SET(cl, value);
+ ANYOF_BITMAP_SETALL(cl);
cl->flags = ANYOF_EOS|ANYOF_UNICODE_ALL;
if (LOC)
cl->flags |= ANYOF_LOCALE;
return 1;
if (!(cl->flags & ANYOF_UNICODE_ALL))
return 0;
- for (value = 0; value < 256; ++value)
- if (!ANYOF_BITMAP_TEST(cl, value))
- return 0;
+ if (!ANYOF_BITMAP_TESTALLSET(cl))
+ return 0;
return 1;
}
}
}
+/*
+ * There are strange code-generation bugs caused on sparc64 by gcc-2.95.2.
+ * These need to be revisited when a newer toolchain becomes available.
+ */
+#if defined(__sparc64__) && defined(__GNUC__)
+# if __GNUC__ < 2 || (__GNUC__ == 2 && __GNUC_MINOR__ < 96)
+# undef SPARC64_GCC_WORKAROUND
+# define SPARC64_GCC_WORKAROUND 1
+# endif
+#endif
+
/* REx optimizer. Converts nodes into quickier variants "in place".
Finds fixed substrings. */
regnode *oscan = scan;
struct regnode_charclass_class this_class;
struct regnode_charclass_class *oclass = NULL;
+ I32 next_is_eval = 0;
switch (PL_regkind[(U8)OP(scan)]) {
case WHILEM: /* End of (?:...)* . */
scan->flags = ((lp <= U8_MAX) ? lp : U8_MAX);
}
scan = NEXTOPER(scan) + EXTRA_STEP_2ARGS;
+ next_is_eval = (OP(scan) == EVAL);
do_curly:
if (flags & SCF_DO_SUBSTR) {
if (mincount == 0) scan_commit(pRExC_state,data); /* Cannot extend fixed substrings */
if (!scan) /* It was not CURLYX, but CURLY. */
scan = next;
if (ckWARN(WARN_REGEXP)
+ /* ? quantifier ok, except for (?{ ... }) */
+ && (next_is_eval || !(mincount == 0 && maxcount == 1))
&& (minnext == 0) && (deltanext == 0)
&& data && !(data->flags & (SF_HAS_PAR|SF_IN_PAR))
&& maxcount <= REG_INFTY/3) /* Complement check for big count */
int counted = mincount != 0;
if (data->last_end > 0 && mincount != 0) { /* Ends with a string. */
+#if defined(SPARC64_GCC_WORKAROUND)
+ I32 b = 0;
+ STRLEN l = 0;
+ char *s = NULL;
+ I32 old = 0;
+
+ if (pos_before >= data->last_start_min)
+ b = pos_before;
+ else
+ b = data->last_start_min;
+
+ l = 0;
+ s = SvPV(data->last_found, l);
+ old = b - data->last_start_min;
+
+#else
I32 b = pos_before >= data->last_start_min
? pos_before : data->last_start_min;
STRLEN l;
char *s = SvPV(data->last_found, l);
I32 old = b - data->last_start_min;
+#endif
if (UTF)
old = utf8_hop((U8*)s, old) - (U8*)s;
if (exp == NULL)
FAIL("NULL regexp argument");
- /* XXXX This looks very suspicious... */
- if (pm->op_pmdynflags & PMdf_CMP_UTF8)
- RExC_utf8 = 1;
- else
- RExC_utf8 = 0;
+ RExC_utf8 = pm->op_pmdynflags & PMdf_CMP_UTF8;
RExC_precomp = exp;
- DEBUG_r(if (!PL_colorset) reginitcolors());
- DEBUG_r(PerlIO_printf(Perl_debug_log, "%sCompiling REx%s `%s%*s%s'\n",
- PL_colors[4],PL_colors[5],PL_colors[0],
- (int)(xend - exp), RExC_precomp, PL_colors[1]));
+ DEBUG_r({
+ if (!PL_colorset) reginitcolors();
+ PerlIO_printf(Perl_debug_log, "%sCompiling REx%s `%s%*s%s'\n",
+ PL_colors[4],PL_colors[5],PL_colors[0],
+ (int)(xend - exp), RExC_precomp, PL_colors[1]);
+ });
RExC_flags16 = pm->op_pmflags;
RExC_sawback = 0;
r->reganch = pm->op_pmflags & PMf_COMPILETIME; /* Again? */
pm->op_pmflags = RExC_flags16;
if (UTF)
- r->reganch |= ROPT_UTF8;
+ r->reganch |= ROPT_UTF8; /* Unicode in it? */
r->regstclass = NULL;
if (RExC_naughty >= 10) /* Probably an expensive pattern. */
r->reganch |= ROPT_NAUGHTY;
case 'Z':
ret = reg_node(pRExC_state, SEOL);
*flagp |= SIMPLE;
+ RExC_seen_zerolen++; /* Do not optimize RE away */
nextchar(pRExC_state);
break;
case 'z':
/* a lovely hack--pretend we saw [\pX] instead */
RExC_end = strchr(RExC_parse, '}');
if (!RExC_end) {
+ U8 c = (U8)*RExC_parse;
RExC_parse += 2;
RExC_end = oldregxend;
- vFAIL("Missing right brace on \\p{}");
+ vFAIL2("Missing right brace on \\%c{}", c);
}
RExC_end++;
}
register char *p;
char *oldp, *s;
STRLEN numlen;
+ STRLEN foldlen;
+ U8 tmpbuf[UTF8_MAXLEN_FOLD+1], *foldbuf;
parse_start = RExC_parse - 1;
case '\\':
switch (*++p) {
case 'A':
+ case 'C':
+ case 'X':
case 'G':
case 'Z':
case 'z':
vFAIL("Missing right brace on \\x{}");
}
else {
- I32 flags = PERL_SCAN_ALLOW_UNDERSCORES;
+ I32 flags = PERL_SCAN_ALLOW_UNDERSCORES
+ | PERL_SCAN_DISALLOW_PREFIX;
numlen = e - p - 1;
ender = grok_hex(p + 1, &numlen, &flags, NULL);
if (ender > 0xff)
}
}
else {
- I32 flags = 0;
+ I32 flags = PERL_SCAN_DISALLOW_PREFIX;
numlen = 2;
ender = grok_hex(p, &numlen, &flags, NULL);
p += numlen;
/* FALL THROUGH */
default:
if (!SIZE_ONLY && ckWARN(WARN_REGEXP) && isALPHA(*p))
- vWARN2(p +1, "Unrecognized escape \\%c passed through", *p);
+ vWARN2(p + 1, "Unrecognized escape \\%c passed through", UCHARAT(p));
goto normal_default;
}
break;
if (RExC_flags16 & PMf_EXTENDED)
p = regwhite(p, RExC_end);
if (UTF && FOLD) {
- if (LOC)
- ender = toLOWER_LC_uvchr(ender);
- else
- ender = toLOWER_uni(ender);
+ /* Prime the casefolded buffer. */
+ ender = toFOLD_uni(ender, tmpbuf, &foldlen);
}
if (ISMULT2(p)) { /* Back off on ?+*. */
if (len)
p = oldp;
- else if (!UNI_IS_INVARIANT(NATIVE_TO_UNI(ender)) && UTF) {
- reguni(pRExC_state, ender, s, &numlen);
- s += numlen;
- len += numlen;
+ else if (UTF) {
+ if (FOLD) {
+ /* Emit all the Unicode characters. */
+ for (foldbuf = tmpbuf;
+ foldlen;
+ foldlen -= numlen) {
+ ender = utf8_to_uvchr(foldbuf, &numlen);
+ reguni(pRExC_state, ender, s, &numlen);
+ s += numlen;
+ len += numlen;
+ foldbuf += numlen;
+ }
+ }
+ else {
+ reguni(pRExC_state, ender, s, &numlen);
+ s += numlen;
+ len += numlen;
+ }
}
else {
len++;
}
break;
}
- if (!UNI_IS_INVARIANT(NATIVE_TO_UNI(ender)) && UTF) {
- reguni(pRExC_state, ender, s, &numlen);
- s += numlen;
- len += numlen - 1;
+ if (UTF) {
+ if (FOLD) {
+ /* Emit all the Unicode characters. */
+ for (foldbuf = tmpbuf;
+ foldlen;
+ foldlen -= numlen) {
+ ender = utf8_to_uvchr(foldbuf, &numlen);
+ reguni(pRExC_state, ender, s, &numlen);
+ s += numlen;
+ len += numlen;
+ foldbuf += numlen;
+ }
+ }
+ else {
+ reguni(pRExC_state, ender, s, &numlen);
+ s += numlen;
+ len += numlen;
+ }
+ len--;
}
else
REGC(ender, s++);
break;
}
+ /* If the encoding pragma is in effect recode the text of
+ * any EXACT-kind nodes. */
+ if (PL_encoding && PL_regkind[(U8)OP(ret)] == EXACT) {
+ STRLEN oldlen = STR_LEN(ret);
+ SV *sv = sv_2mortal(newSVpvn(STRING(ret), oldlen));
+
+ if (RExC_utf8)
+ SvUTF8_on(sv);
+ if (sv_utf8_downgrade(sv, TRUE)) {
+ char *s = Perl_sv_recode_to_utf8(aTHX_ sv, PL_encoding);
+ STRLEN newlen = SvCUR(sv);
+
+ if (!SIZE_ONLY) {
+ DEBUG_r(PerlIO_printf(Perl_debug_log, "recode %*s to %*s\n",
+ (int)oldlen, STRING(ret),
+ (int)newlen, s));
+ Copy(s, STRING(ret), newlen, char);
+ STR_LEN(ret) += newlen - oldlen;
+ RExC_emit += STR_SZ(newlen) - STR_SZ(oldlen);
+ } else
+ RExC_size += STR_SZ(newlen) - STR_SZ(oldlen);
+ }
+ }
+
return(ret);
}
Character classes ([:foo:]) can also be negated ([:^foo:]).
Returns a named class id (ANYOF_XXX) if successful, -1 otherwise.
Equivalence classes ([=foo=]) and composites ([.foo.]) are parsed,
- but trigger warnings because they are currently unimplemented. */
+ but trigger failures because they are currently unimplemented. */
+
+#define POSIXCC_DONE(c) ((c) == ':')
+#define POSIXCC_NOTYET(c) ((c) == '=' || (c) == '.')
+#define POSIXCC(c) (POSIXCC_DONE(c) || POSIXCC_NOTYET(c))
+
STATIC I32
S_regpposixcc(pTHX_ RExC_state_t *pRExC_state, I32 value)
{
if (value == '[' && RExC_parse + 1 < RExC_end &&
/* I smell either [: or [= or [. -- POSIX has been here, right? */
- (*RExC_parse == ':' ||
- *RExC_parse == '=' ||
- *RExC_parse == '.')) {
- char c = *RExC_parse;
+ POSIXCC(UCHARAT(RExC_parse))) {
+ char c = UCHARAT(RExC_parse);
char* s = RExC_parse++;
- while (RExC_parse < RExC_end && *RExC_parse != c)
+ while (RExC_parse < RExC_end && UCHARAT(RExC_parse) != c)
RExC_parse++;
if (RExC_parse == RExC_end)
/* Grandfather lone [:, [=, [. */
else {
char* t = RExC_parse++; /* skip over the c */
- if (*RExC_parse == ']') {
+ if (UCHARAT(RExC_parse) == ']') {
RExC_parse++; /* skip over the ending ] */
posixcc = s + 1;
if (*s == ':') {
/* adjust RExC_parse so the warning shows after
the class closes */
- while (*RExC_parse && *RExC_parse != ']')
+ while (UCHARAT(RExC_parse) && UCHARAT(RExC_parse) != ']')
RExC_parse++;
Simple_vFAIL3("POSIX syntax [%c %c] is reserved for future extensions", c, c);
}
S_checkposixcc(pTHX_ RExC_state_t *pRExC_state)
{
if (!SIZE_ONLY && ckWARN(WARN_REGEXP) &&
- (*RExC_parse == ':' ||
- *RExC_parse == '=' ||
- *RExC_parse == '.')) {
+ POSIXCC(UCHARAT(RExC_parse))) {
char *s = RExC_parse;
char c = *s++;
vWARN3(s+2, "POSIX syntax [%c %c] belongs inside character classes", c, c);
/* [[=foo=]] and [[.foo.]] are still future. */
- if (c == '=' || c == '.')
- {
+ if (POSIXCC_NOTYET(c)) {
/* adjust RExC_parse so the error shows after
the class closes */
- while (*RExC_parse && *RExC_parse++ != ']')
+ while (UCHARAT(RExC_parse) && UCHARAT(RExC_parse++) != ']')
;
Simple_vFAIL3("POSIX syntax [%c %c] is reserved for future extensions", c, c);
}
S_regclass(pTHX_ RExC_state_t *pRExC_state)
{
register UV value;
+ register UV nextvalue;
register IV prevvalue = OOB_UNICODE;
register IV range = 0;
register regnode *ret;
SV *listsv = Nullsv;
register char *e;
UV n;
- bool optimize_invert = TRUE;
+ bool optimize_invert = TRUE;
+ AV* unicode_alternate = 0;
ret = reganode(pRExC_state, ANYOF, 0);
if (!SIZE_ONLY)
ANYOF_FLAGS(ret) = 0;
- if (*RExC_parse == '^') { /* Complement of range. */
+ if (UCHARAT(RExC_parse) == '^') { /* Complement of range. */
RExC_naughty++;
RExC_parse++;
if (!SIZE_ONLY)
listsv = newSVpvn("# comment\n", 10);
}
- if (!SIZE_ONLY && ckWARN(WARN_REGEXP))
+ nextvalue = RExC_parse < RExC_end ? UCHARAT(RExC_parse) : 0;
+
+ if (!SIZE_ONLY && ckWARN(WARN_REGEXP) && POSIXCC(nextvalue))
checkposixcc(pRExC_state);
- if (*RExC_parse == ']' || *RExC_parse == '-')
+ if (UCHARAT(RExC_parse) == ']' || UCHARAT(RExC_parse) == '-')
goto charclassloop; /* allow 1st char to be ] or - */
- while (RExC_parse < RExC_end && *RExC_parse != ']') {
+ while (RExC_parse < RExC_end && UCHARAT(RExC_parse) != ']') {
charclassloop:
}
else
value = UCHARAT(RExC_parse++);
- if (value == '[')
+ nextvalue = RExC_parse < RExC_end ? UCHARAT(RExC_parse) : 0;
+ if (value == '[' && POSIXCC(nextvalue))
namedclass = regpposixcc(pRExC_state, value);
else if (value == '\\') {
if (UTF) {
case 'p':
case 'P':
if (*RExC_parse == '{') {
+ U8 c = (U8)value;
e = strchr(RExC_parse++, '}');
if (!e)
- vFAIL("Missing right brace on \\p{}");
+ vFAIL2("Missing right brace on \\%c{}", c);
+ while (isSPACE(UCHARAT(RExC_parse)))
+ RExC_parse++;
+ if (e == RExC_parse)
+ vFAIL2("Empty \\%c{}", c);
n = e - RExC_parse;
+ while (isSPACE(UCHARAT(RExC_parse + n - 1)))
+ n--;
}
else {
e = RExC_parse;
n = 1;
}
if (!SIZE_ONLY) {
+ if (UCHARAT(RExC_parse) == '^') {
+ RExC_parse++;
+ n--;
+ value = value == 'p' ? 'P' : 'p'; /* toggle */
+ while (isSPACE(UCHARAT(RExC_parse))) {
+ RExC_parse++;
+ n--;
+ }
+ }
if (value == 'p')
- Perl_sv_catpvf(aTHX_ listsv,
- "+utf8::%.*s\n", (int)n, RExC_parse);
+ Perl_sv_catpvf(aTHX_ listsv,
+ "+utf8::%.*s\n", (int)n, RExC_parse);
else
- Perl_sv_catpvf(aTHX_ listsv,
- "!utf8::%.*s\n", (int)n, RExC_parse);
+ Perl_sv_catpvf(aTHX_ listsv,
+ "!utf8::%.*s\n", (int)n, RExC_parse);
}
RExC_parse = e + 1;
ANYOF_FLAGS(ret) |= ANYOF_UNICODE;
case 'a': value = ASCII_TO_NATIVE('\007');break;
case 'x':
if (*RExC_parse == '{') {
- I32 flags = PERL_SCAN_ALLOW_UNDERSCORES;
+ I32 flags = PERL_SCAN_ALLOW_UNDERSCORES
+ | PERL_SCAN_DISALLOW_PREFIX;
e = strchr(RExC_parse++, '}');
if (!e)
vFAIL("Missing right brace on \\x{}");
RExC_parse = e + 1;
}
else {
- I32 flags = 0;
+ I32 flags = PERL_SCAN_DISALLOW_PREFIX;
numlen = 2;
value = grok_hex(RExC_parse, &numlen, &flags, NULL);
RExC_parse += numlen;
}
else
#endif
- for (i = prevvalue; i <= ceilvalue; i++)
- ANYOF_BITMAP_SET(ret, i);
+ for (i = prevvalue; i <= ceilvalue; i++)
+ ANYOF_BITMAP_SET(ret, i);
}
- if (value > 255) {
+ if (value > 255 || UTF) {
ANYOF_FLAGS(ret) |= ANYOF_UNICODE;
if (prevvalue < value)
Perl_sv_catpvf(aTHX_ listsv, "%04"UVxf"\t%04"UVxf"\n",
(UV)prevvalue, (UV)value);
- else if (prevvalue == value)
+ else if (prevvalue == value) {
Perl_sv_catpvf(aTHX_ listsv, "%04"UVxf"\n",
(UV)value);
+ if (FOLD) {
+ U8 tmpbuf [UTF8_MAXLEN+1];
+ U8 foldbuf[UTF8_MAXLEN_FOLD+1];
+ STRLEN foldlen;
+ UV f;
+
+ uvchr_to_utf8(tmpbuf, value);
+ to_utf8_fold(tmpbuf, foldbuf, &foldlen);
+ f = utf8_to_uvchr(foldbuf, 0);
+
+ /* If folding and foldable and a single
+ * character, insert also the folded version
+ * to the charclass. */
+ if (f != value) {
+ if (foldlen == UNISKIP(f))
+ Perl_sv_catpvf(aTHX_ listsv,
+ "%04"UVxf"\n", f);
+ else {
+ /* Any multicharacter foldings
+ * require the following transform:
+ * [ABCDEF] -> (?:[ABCabcDEFd]|pq|rst)
+ * where E folds into "pq" and F folds
+ * into "rst", all other characters
+ * fold to single characters. We save
+ * away these multicharacter foldings,
+ * to be later saved as part of the
+ * additional "s" data. */
+ SV *sv;
+
+ if (!unicode_alternate)
+ unicode_alternate = newAV();
+ sv = newSVpvn((char*)foldbuf, foldlen);
+ SvUTF8_on(sv);
+ av_push(unicode_alternate, sv);
+ }
+ }
+
+ /* If folding and the value is one of the Greek
+ * sigmas insert a few more sigmas to make the
+ * folding rules of the sigmas to work right.
+ * Note that not all the possible combinations
+ * are handled here: some of them are handled
+ * by the standard folding rules, and some of
+ * them (literal or EXACTF cases) are handled
+ * during runtime in regexec.c:S_find_byclass(). */
+ if (value == UNICODE_GREEK_SMALL_LETTER_FINAL_SIGMA) {
+ Perl_sv_catpvf(aTHX_ listsv, "%04"UVxf"\n",
+ (UV)UNICODE_GREEK_CAPITAL_LETTER_SIGMA);
+ Perl_sv_catpvf(aTHX_ listsv, "%04"UVxf"\n",
+ (UV)UNICODE_GREEK_SMALL_LETTER_SIGMA);
+ }
+ else if (value == UNICODE_GREEK_CAPITAL_LETTER_SIGMA)
+ Perl_sv_catpvf(aTHX_ listsv, "%04"UVxf"\n",
+ (UV)UNICODE_GREEK_SMALL_LETTER_SIGMA);
+ }
+ }
}
}
AV *av = newAV();
SV *rv;
+ /* The 0th element stores the character class description
+ * in its textual form: used later (regexec.c:Perl_regclass_swatch())
+ * to initialize the appropriate swash (which gets stored in
+ * the 1st element), and also useful for dumping the regnode.
+ * The 2nd element stores the multicharacter foldings,
+ * used later (regexec.c:s_reginclasslen()). */
av_store(av, 0, listsv);
av_store(av, 1, NULL);
+ av_store(av, 2, (SV*)unicode_alternate);
rv = newRV_noinc((SV*)av);
n = add_data(pRExC_state, 1, "s");
RExC_rx->data->data[n] = (void*)rv;
k = PL_regkind[(U8)OP(o)];
- if (k == EXACT)
- Perl_sv_catpvf(aTHX_ sv, " <%s%.*s%s>", PL_colors[0],
- STR_LEN(o), STRING(o), PL_colors[1]);
+ if (k == EXACT) {
+ SV *dsv = sv_2mortal(newSVpvn("", 0));
+ bool do_utf8 = DO_UTF8(sv);
+ char *s = do_utf8 ?
+ pv_uni_display(dsv, (U8*)STRING(o), STR_LEN(o), 60, 0) :
+ STRING(o);
+ int len = do_utf8 ?
+ strlen(s) :
+ STR_LEN(o);
+ Perl_sv_catpvf(aTHX_ sv, " <%s%.*s%s>",
+ PL_colors[0],
+ len, s,
+ PL_colors[1]);
+ }
else if (k == CURLY) {
if (OP(o) == CURLYM || OP(o) == CURLYN || OP(o) == CURLYX)
Perl_sv_catpvf(aTHX_ sv, "[%d]", o->flags); /* Parenth number */
if (flags & ANYOF_UNICODE)
sv_catpv(sv, "{unicode}");
else if (flags & ANYOF_UNICODE_ALL)
- sv_catpv(sv, "{all-unicode}");
+ sv_catpv(sv, "{unicode_all}");
{
SV *lv;
- SV *sw = regclass_swash(o, FALSE, &lv);
+ SV *sw = regclass_swash(o, FALSE, &lv, 0);
if (lv) {
if (sw) {
void
Perl_pregfree(pTHX_ struct regexp *r)
{
- DEBUG_r(if (!PL_colorset) reginitcolors());
+#ifdef DEBUGGING
+ SV *dsv = PERL_DEBUG_PAD_ZERO(0);
+#endif
if (!r || (--r->refcnt > 0))
return;
- DEBUG_r(PerlIO_printf(Perl_debug_log,
- "%sFreeing REx:%s `%s%.60s%s%s'\n",
- PL_colors[4],PL_colors[5],PL_colors[0],
- r->precomp,
- PL_colors[1],
- (strlen(r->precomp) > 60 ? "..." : "")));
+ DEBUG_r({
+ bool utf8 = r->reganch & ROPT_UTF8;
+ char *s = pv_uni_display(dsv, (U8*)r->precomp, r->prelen, 60,
+ UNI_DISPLAY_ISPRINT);
+ int len = SvCUR(dsv);
+ if (!PL_colorset)
+ reginitcolors();
+ PerlIO_printf(Perl_debug_log,
+ "%sFreeing REx:%s `%s%*.*s%s%s'\n",
+ PL_colors[4],PL_colors[5],PL_colors[0],
+ len, len, s,
+ PL_colors[1],
+ len > 60 ? "..." : "");
+ });
if (r->precomp)
Safefree(r->precomp);
new_comppad = NULL;
break;
case 'n':
- break;
+ break;
default:
Perl_croak(aTHX_ "panic: regfree data code '%c'", r->data->what[n]);
}