#define LOAD_UTF8_CHARCLASS(swash_ptr, property_name) STMT_START { \
if (!swash_ptr) { \
U8 flags = _CORE_SWASH_INIT_ACCEPT_INVLIST; \
- ENTER; save_re_context(); \
swash_ptr = _core_swash_init("utf8", property_name, &PL_sv_undef, \
1, 0, NULL, &flags); \
assert(swash_ptr); \
switch ((_char_class_number) classnum) {
case _CC_ENUM_ALPHANUMERIC: return isALPHANUMERIC_LC(character);
case _CC_ENUM_ALPHA: return isALPHA_LC(character);
+ case _CC_ENUM_ASCII: return isASCII_LC(character);
+ case _CC_ENUM_BLANK: return isBLANK_LC(character);
+ case _CC_ENUM_CASED: return isLOWER_LC(character)
+ || isUPPER_LC(character);
+ case _CC_ENUM_CNTRL: return isCNTRL_LC(character);
case _CC_ENUM_DIGIT: return isDIGIT_LC(character);
case _CC_ENUM_GRAPH: return isGRAPH_LC(character);
case _CC_ENUM_LOWER: return isLOWER_LC(character);
case _CC_ENUM_PRINT: return isPRINT_LC(character);
+ case _CC_ENUM_PSXSPC: return isPSXSPC_LC(character);
case _CC_ENUM_PUNCT: return isPUNCT_LC(character);
+ case _CC_ENUM_SPACE: return isSPACE_LC(character);
case _CC_ENUM_UPPER: return isUPPER_LC(character);
case _CC_ENUM_WORDCHAR: return isWORDCHAR_LC(character);
- case _CC_ENUM_SPACE: return isSPACE_LC(character);
- case _CC_ENUM_BLANK: return isBLANK_LC(character);
case _CC_ENUM_XDIGIT: return isXDIGIT_LC(character);
- case _CC_ENUM_CNTRL: return isCNTRL_LC(character);
- case _CC_ENUM_PSXSPC: return isPSXSPC_LC(character);
- case _CC_ENUM_ASCII: return isASCII_LC(character);
default: /* VERTSPACE should never occur in locales */
Perl_croak(aTHX_ "panic: isFOO_lc() has an unexpected character class '%d'", classnum);
}
swash_property_names[classnum], &PL_sv_undef, 1, 0, NULL, &flags);
}
- return swash_fetch(PL_utf8_swash_ptrs[classnum], (U8 *) character, TRUE);
+ return cBOOL(swash_fetch(PL_utf8_swash_ptrs[classnum], (U8 *)
+ character,
+ TRUE /* is UTF */ ));
}
switch ((_char_class_number) classnum) {
? (utf8_target ? trie_utf8 : trie_plain) \
: (utf8_target ? trie_utf8_fold : trie_latin_utf8_fold))
-#define REXEC_TRIE_READ_CHAR(trie_type, trie, widecharmap, uc, uscan, len, \
-uvc, charid, foldlen, foldbuf, uniflags) STMT_START { \
+#define REXEC_TRIE_READ_CHAR(trie_type, trie, widecharmap, uc, uscan, len, uvc, charid, foldlen, foldbuf, uniflags) \
+STMT_START { \
STRLEN skiplen; \
switch (trie_type) { \
case trie_utf8_fold: \
/* We know what class it must start with. */
switch (OP(c)) {
case ANYOF:
+ case ANYOF_SYNTHETIC:
case ANYOF_WARN_SUPER:
if (utf8_target) {
REXEC_FBC_UTF8_CLASS_SCAN(
}
case BOUNDL:
RXp_MATCH_TAINTED_on(prog);
- FBC_BOUND(isALNUM_LC,
- isALNUM_LC_uvchr(UNI_TO_NATIVE(tmp)),
- isALNUM_LC_utf8((U8*)s));
+ FBC_BOUND(isWORDCHAR_LC,
+ isWORDCHAR_LC_uvchr(UNI_TO_NATIVE(tmp)),
+ isWORDCHAR_LC_utf8((U8*)s));
break;
case NBOUNDL:
RXp_MATCH_TAINTED_on(prog);
- FBC_NBOUND(isALNUM_LC,
- isALNUM_LC_uvchr(UNI_TO_NATIVE(tmp)),
- isALNUM_LC_utf8((U8*)s));
+ FBC_NBOUND(isWORDCHAR_LC,
+ isWORDCHAR_LC_uvchr(UNI_TO_NATIVE(tmp)),
+ isWORDCHAR_LC_utf8((U8*)s));
break;
case BOUND:
FBC_BOUND(isWORDCHAR,
- isALNUM_uni(tmp),
+ isWORDCHAR_uni(tmp),
cBOOL(swash_fetch(PL_utf8_swash_ptrs[_CC_WORDCHAR], (U8*)s, utf8_target)));
break;
case BOUNDA:
break;
case NBOUND:
FBC_NBOUND(isWORDCHAR,
- isALNUM_uni(tmp),
+ isWORDCHAR_uni(tmp),
cBOOL(swash_fetch(PL_utf8_swash_ptrs[_CC_WORDCHAR], (U8*)s, utf8_target)));
break;
case NBOUNDA:
break;
case BOUNDU:
FBC_BOUND(isWORDCHAR_L1,
- isALNUM_uni(tmp),
+ isWORDCHAR_uni(tmp),
cBOOL(swash_fetch(PL_utf8_swash_ptrs[_CC_WORDCHAR], (U8*)s, utf8_target)));
break;
case NBOUNDU:
FBC_NBOUND(isWORDCHAR_L1,
- isALNUM_uni(tmp),
+ isWORDCHAR_uni(tmp),
cBOOL(swash_fetch(PL_utf8_swash_ptrs[_CC_WORDCHAR], (U8*)s, utf8_target)));
break;
case LNBREAK:
ln = utf8n_to_uvchr(r, UTF8SKIP(r), 0, uniflags);
}
if (FLAGS(scan) != REGEX_LOCALE_CHARSET) {
- ln = isALNUM_uni(ln);
+ ln = isWORDCHAR_uni(ln);
if (NEXTCHR_IS_EOS)
n = 0;
else {
}
}
else {
- ln = isALNUM_LC_uvchr(UNI_TO_NATIVE(ln));
- n = NEXTCHR_IS_EOS ? 0 : isALNUM_LC_utf8((U8*)locinput);
+ ln = isWORDCHAR_LC_uvchr(UNI_TO_NATIVE(ln));
+ n = NEXTCHR_IS_EOS ? 0 : isWORDCHAR_LC_utf8((U8*)locinput);
}
}
else {
n = NEXTCHR_IS_EOS ? 0 : isWORDCHAR_L1(nextchr);
break;
case REGEX_LOCALE_CHARSET:
- ln = isALNUM_LC(ln);
- n = NEXTCHR_IS_EOS ? 0 : isALNUM_LC(nextchr);
+ ln = isWORDCHAR_LC(ln);
+ n = NEXTCHR_IS_EOS ? 0 : isWORDCHAR_LC(nextchr);
break;
case REGEX_DEPENDS_CHARSET:
- ln = isALNUM(ln);
- n = NEXTCHR_IS_EOS ? 0 : isALNUM(nextchr);
+ ln = isWORDCHAR(ln);
+ n = NEXTCHR_IS_EOS ? 0 : isWORDCHAR(nextchr);
break;
case REGEX_ASCII_RESTRICTED_CHARSET:
case REGEX_ASCII_MORE_RESTRICTED_CHARSET:
* UTF8_IS_INVARIANT works even on non-UTF-8 strings, or else
* wouldn't be invariant) */
if (UTF8_IS_INVARIANT(nextchr) || ! utf8_target) {
- if (! (to_complement ^ cBOOL(isFOO_lc(FLAGS(scan), nextchr)))) {
+ if (! (to_complement ^ cBOOL(isFOO_lc(FLAGS(scan), (U8) nextchr)))) {
sayNO;
}
}
else if (UTF8_IS_DOWNGRADEABLE_START(nextchr)) {
if (! (to_complement ^ cBOOL(isFOO_lc(FLAGS(scan),
- TWO_BYTE_UTF8_TO_UNI(nextchr,
+ (U8) TWO_BYTE_UTF8_TO_UNI(nextchr,
*(locinput + 1))))))
{
sayNO;
*/
Copy(&PL_reg_state, &saved_state, 1, struct re_save_state);
- PL_reg_state.re_reparsing = FALSE;
-
if (!caller_cv)
caller_cv = find_runcv(NULL);
* points to newcv's pad. */
if (newcv != last_pushed_cv || PL_comppad != last_pad)
{
- I32 depth = (newcv == caller_cv) ? 0 : 1;
+ U8 flags = (CXp_SUB_RE |
+ ((newcv == caller_cv) ? CXp_SUB_RE_FAKE : 0));
if (last_pushed_cv) {
- CHANGE_MULTICALL_WITHDEPTH(newcv, depth);
+ CHANGE_MULTICALL_FLAGS(newcv, flags);
}
else {
- PUSH_MULTICALL_WITHDEPTH(newcv, depth);
+ PUSH_MULTICALL_FLAGS(newcv, flags);
}
last_pushed_cv = newcv;
}
* will be 1, so the exclusive or will reverse things, so we
* are testing for \W. On the third iteration, 'to_complement'
* will be 0, and we would be testing for \s; the fourth
- * iteration would test for \S, etc. */
+ * iteration would test for \S, etc.
+ *
+ * Note that this code assumes that all the classes are closed
+ * under folding. For example, if a character matches \w, then
+ * its fold does too; and vice versa. This should be true for
+ * any well-behaved locale for all the currently defined Posix
+ * classes, except for :lower: and :upper:, which are handled
+ * by the pseudo-class :cased: which matches if either of the
+ * other two does. To get rid of this assumption, an outer
+ * loop could be used below to iterate over both the source
+ * character, and its fold (if different) */
int count = 0;
int to_complement = 0;
|| (utf8_target
&& (c >=256
|| (! (flags & ANYOF_LOCALE))
- || (flags & ANYOF_IS_SYNTHETIC)))))
+ || OP(n) == ANYOF_SYNTHETIC))))
{
SV * const sw = core_regclass_swash(prog, n, TRUE, 0);
if (sw) {
https://perl5.git.perl.org / perl5.git / blobdiff