/* these are unrolled below in the CCC_TRY_XXX defined */
#define LOAD_UTF8_CHARCLASS(class,str) STMT_START { \
- if (!CAT2(PL_utf8_,class)) { bool ok; ENTER; save_re_context(); ok=CAT2(is_utf8_,class)((const U8*)str); assert(ok); LEAVE; } } STMT_END
+ if (!CAT2(PL_utf8_,class)) { \
+ bool ok; \
+ ENTER; save_re_context(); \
+ ok=CAT2(is_utf8_,class)((const U8*)str); \
+ assert(ok); LEAVE; } } STMT_END
/* Doesn't do an assert to verify that is correct */
#define LOAD_UTF8_CHARCLASS_NO_CHECK(class) STMT_START { \
- if (!CAT2(PL_utf8_,class)) { bool throw_away; ENTER; save_re_context(); throw_away = CAT2(is_utf8_,class)((const U8*)" "); LEAVE; } } STMT_END
+ if (!CAT2(PL_utf8_,class)) { \
+ bool throw_away __attribute__unused__; \
+ ENTER; save_re_context(); \
+ throw_away = CAT2(is_utf8_,class)((const U8*)" "); \
+ LEAVE; } } STMT_END
#define LOAD_UTF8_CHARCLASS_ALNUM() LOAD_UTF8_CHARCLASS(alnum,"a")
#define LOAD_UTF8_CHARCLASS_DIGIT() LOAD_UTF8_CHARCLASS(digit,"0")
reginclass(prog, c, (U8*)s, &inclasslen, utf8_target));
}
else {
- while (s < strend) {
- STRLEN skip = 1;
-
- if (REGINCLASS(prog, c, (U8*)s)) {
- if (tmp && (!reginfo || regtry(reginfo, &s)))
- goto got_it;
- else
- tmp = doevery;
- }
- else
- tmp = 1;
- s += skip;
- }
+ REXEC_FBC_CLASS_SCAN(REGINCLASS(prog, c, (U8*)s));
}
break;
case CANY:
utf8_fold_flags = 0;
goto do_exactf_utf8;
}
+
+ /* Any 'ss' in the pattern should have been replaced by regcomp,
+ * so we don't have to worry here about this single special case
+ * in the Latin1 range */
fold_array = PL_fold_latin1;
folder = foldEQ_latin1;
- /* XXX This uses the full utf8 fold because if the pattern contains
- * 'ss' it could match LATIN_SMALL_LETTER SHARP_S in the string.
- * There could be a new node type, say EXACTFU_SS, which is
- * generated by regcomp only if there is an 'ss', and then every
- * other case could goto do_exactf_non_utf8;*/
- goto do_exactf_utf8;
+ goto do_exactf_non_utf8;
case EXACTF:
if (UTF_PATTERN || utf8_target) {
case NALNUMU:
REXEC_FBC_CSCAN_PRELOAD(
LOAD_UTF8_CHARCLASS_ALNUM(),
- swash_fetch(PL_utf8_alnum,(U8*)s, utf8_target),
+ !swash_fetch(PL_utf8_alnum,(U8*)s, utf8_target),
! isWORDCHAR_L1((U8) *s)
);
break;
/* This is safe against NULLs: */
ReREFCNT_dec(PM_GETRE(PL_reg_curpm));
/* PM_reg_curpm owns a reference to this regexp. */
- ReREFCNT_inc(rx);
+ (void)ReREFCNT_inc(rx);
#endif
PM_SETRE(PL_reg_curpm, rx);
PL_reg_oldcurpm = PL_curpm;
char *e = PL_regeol;
if (! foldEQ_utf8_flags(s, 0, ln, cBOOL(UTF_PATTERN),
- l, &e, 0, utf8_target, fold_utf8_flags)) {
- /* 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 (!(utf8_target &&
- toLOWER(s[0]) == 's' &&
- ln >= 2 &&
- toLOWER(s[1]) == 's' &&
- (U8)l[0] == 0xC3 &&
- e - l >= 2 &&
- (U8)l[1] == 0x9F))
- sayNO;
+ l, &e, 0, utf8_target, fold_utf8_flags))
+ {
+ sayNO;
}
locinput = e;
nextchr = UCHARAT(locinput);
/* First just match a string of min A's. */
if (n < min) {
+ ST.cp = regcppush(cur_curlyx->u.curlyx.parenfloor);
cur_curlyx->u.curlyx.lastloc = locinput;
+ REGCP_SET(ST.lastcp);
+
PUSH_STATE_GOTO(WHILEM_A_pre, A);
/* NOTREACHED */
}
/* NOTREACHED */
case WHILEM_A_min_fail: /* just failed to match A in a minimal match */
- REGCP_UNWIND(ST.lastcp);
- regcppop(rex);
/* FALL THROUGH */
case WHILEM_A_pre_fail: /* just failed to match even minimal A */
+ REGCP_UNWIND(ST.lastcp);
+ regcppop(rex);
cur_curlyx->u.curlyx.lastloc = ST.save_lastloc;
cur_curlyx->u.curlyx.count--;
CACHEsayNO;
rex = (struct regexp *)SvANY(rex_sv);
rexi = RXi_GET(rex);
cur_curlyx = cur_eval->u.eval.prev_curlyx;
- ReREFCNT_inc(rex_sv);
+ (void)ReREFCNT_inc(rex_sv);
st->u.eval.cp = regcppush(0); /* Save *all* the positions. */
/* rex was changed so update the pointer in PL_reglastparen and PL_reglastcloseparen */
PERL_ARGS_ASSERT_REGCLASS_SWASH;
+ assert(ANYOF_NONBITMAP(node));
+
if (data && data->count) {
const U32 n = ARG(node);
/* If the bitmap didn't (or couldn't) match, and something outside the
* bitmap could match, try that. Locale nodes specifiy completely the
* behavior of code points in the bit map (otherwise, a utf8 target would
- * cause them to be treated as Unicode and not locale), except XXX in
+ * cause them to be treated as Unicode and not locale), except in
* the very unlikely event when this node is a synthetic start class, which
- * could be a combination of locale and non-locale nodes */
+ * could be a combination of locale and non-locale nodes. So allow locale
+ * to match for the synthetic start class, which will give a false
+ * positive that will be resolved when the match is done again as not part
+ * of the synthetic start class */
if (!match) {
if (utf8_target && (flags & ANYOF_UNICODE_ALL) && c >= 256) {
match = TRUE; /* Everything above 255 matches */
}
- else if ((flags & ANYOF_NONBITMAP_NON_UTF8
- || (utf8_target && ANYOF_NONBITMAP(n)
- && (c >=256 || ! (flags & ANYOF_LOCALE)))))
+ else if (ANYOF_NONBITMAP(n)
+ && ((flags & ANYOF_NONBITMAP_NON_UTF8)
+ || (utf8_target
+ && (c >=256
+ || (! (flags & ANYOF_LOCALE))
+ || (flags & ANYOF_IS_SYNTHETIC)))))
{
AV *av;
SV * const sw = regclass_swash(prog, n, TRUE, 0, (SV**)&av);
else if (flags & ANYOF_LOC_NONBITMAP_FOLD) {
/* Here, we need to test if the fold of the target string
- * matches. In the case of a multi-char fold that is
- * caught by regcomp.c, it has stored all such folds into
- * 'av'; we linearly check to see if any match the target
- * string (folded). We know that the originals were each
- * one character, but we don't currently know how many
- * characters/bytes each folded to, except we do know that
- * there are small limits imposed by Unicode. XXX A
- * performance enhancement would be to have regcomp.c store
- * the max number of chars/bytes that are in an av entry,
- * as, say the 0th element. Even better would be to have a
- * hash of the few characters that can start a multi-char
- * fold to the max number of chars of those folds.
- *
- * Further down, if there isn't a
- * match in the av, we will check if there is another
- * fold-type match. For that, we also need the fold, but
- * only the first character. No sense in folding it twice,
- * so we do it here, even if there isn't any multi-char
- * fold, so we always fold at least the first character.
- * If the node is a straight ANYOF node, or there is only
- * one character available in the string, or if there isn't
- * any av, that's all we have to fold. In the case of a
- * multi-char fold, we do have guarantees in Unicode that
- * it can only expand up to so many characters and so many
- * bytes. We keep track so don't exceed either.
+ * matches. The non-multi char folds have all been moved to
+ * the compilation phase, and the multi-char folds have
+ * been stored by regcomp into 'av'; we linearly check to
+ * see if any match the target string (folded). We know
+ * that the originals were each one character, but we don't
+ * currently know how many characters/bytes each folded to,
+ * except we do know that there are small limits imposed by
+ * Unicode. XXX A performance enhancement would be to have
+ * regcomp.c store the max number of chars/bytes that are
+ * in an av entry, as, say the 0th element. Even better
+ * would be to have a hash of the few characters that can
+ * start a multi-char fold to the max number of chars of
+ * those folds.
*
* If there is a match, we will need to advance (if lenp is
* specified) the match pointer in the target string. But
* create a map so that we know how many bytes in the
* source to advance given that we have matched a certain
* number of bytes in the fold. This map is stored in
- * 'map_fold_len_back'. The first character in the fold
- * has array element 1 contain the number of bytes in the
- * source that folded to it; the 2nd is the cumulative
- * number to match it; ... */
- U8 map_fold_len_back[UTF8_MAX_FOLD_CHAR_EXPAND+1] = { 0 };
+ * 'map_fold_len_back'. Let n mean the number of bytes in
+ * the fold of the first character that we are folding.
+ * Then map_fold_len_back[n] is set to the number of bytes
+ * in that first character. Similarly let m be the
+ * corresponding number for the second character to be
+ * folded. Then map_fold_len_back[n+m] is set to the
+ * number of bytes occupied by the first two source
+ * characters. ... */
+ U8 map_fold_len_back[UTF8_MAXBYTES_CASE+1] = { 0 };
U8 folded[UTF8_MAXBYTES_CASE+1];
STRLEN foldlen = 0; /* num bytes in fold of 1st char */
- STRLEN foldlen_for_av; /* num bytes in fold of all chars */
+ STRLEN total_foldlen = 0; /* num bytes in fold of all
+ chars */
if (OP(n) == ANYOF || maxlen == 1 || ! lenp || ! av) {
/* Here, only need to fold the first char of the target
- * string */
+ * string. It the source wasn't utf8, is 1 byte long */
to_utf8_fold(utf8_p, folded, &foldlen);
- foldlen_for_av = foldlen;
- map_fold_len_back[1] = UTF8SKIP(utf8_p);
+ total_foldlen = foldlen;
+ map_fold_len_back[foldlen] = (utf8_target)
+ ? UTF8SKIP(utf8_p)
+ : 1;
}
else {
/* Here, need to fold more than the first char. Do so
* up to the limits */
- UV which_char = 0;
U8* source_ptr = utf8_p; /* The source for the fold
is the regex target
string */
U8* e = utf8_p + maxlen; /* Can't go beyond last
available byte in the
target string */
- while (which_char < UTF8_MAX_FOLD_CHAR_EXPAND
- && source_ptr < e)
+ U8 i;
+ for (i = 0;
+ i < UTF8_MAX_FOLD_CHAR_EXPAND && source_ptr < e;
+ i++)
{
/* Fold the next character */
break;
}
- /* Save the first character's folded length, in
- * case we have to use it later */
- if (! foldlen) {
- foldlen = this_char_foldlen;
- }
-
- /* Here, add the fold of this character */
+ /* Add the fold of this character */
Copy(this_char_folded,
folded_ptr,
this_char_foldlen,
U8);
- which_char++;
- map_fold_len_back[which_char] =
- map_fold_len_back[which_char - 1]
- + UTF8SKIP(source_ptr);
- folded_ptr += this_char_foldlen;
source_ptr += UTF8SKIP(source_ptr);
+ folded_ptr += this_char_foldlen;
+ total_foldlen = folded_ptr - folded;
+
+ /* Create map from the number of bytes in the fold
+ * back to the number of bytes in the source. If
+ * the source isn't utf8, the byte count is just
+ * the number of characters so far */
+ map_fold_len_back[total_foldlen]
+ = (utf8_target)
+ ? source_ptr - utf8_p
+ : i + 1;
}
*folded_ptr = '\0';
- foldlen_for_av = folded_ptr - folded;
}
/* Do the linear search to see if the fold is in the list
- * of multi-char folds. (Useless to look if won't be able
- * to store that it is a multi-char fold in *lenp) */
- if (lenp && av) {
+ * of multi-char folds. */
+ if (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 <= foldlen_for_av && memEQ(s,
- (char*)folded,
- len))
+
+ if (len <= total_foldlen
+ && memEQ(s, (char*)folded, len)
+
+ /* If 0, means matched a partial char. See
+ * [perl #90536] */
+ && map_fold_len_back[len])
{
/* Advance the target string ptr to account for
* this fold, but have to translate from the
* folded length to the corresponding source
- * length. The array is indexed by how many
- * characters in the match */
- *lenp = map_fold_len_back[
- utf8_length(folded, folded + len)];
+ * length. */
+ if (lenp) {
+ *lenp = map_fold_len_back[len];
+ }
match = TRUE;
break;
}
prog->substrs->data[i].utf8_substr = sv;
sv_utf8_upgrade(sv);
if (SvVALID(prog->substrs->data[i].substr)) {
- const U8 flags = BmFLAGS(prog->substrs->data[i].substr);
- if (flags & FBMcf_TAIL) {
+ if (SvTAIL(prog->substrs->data[i].substr)) {
/* Trim the trailing \n that fbm_compile added last
time. */
SvCUR_set(sv, SvCUR(sv) - 1);
/* Whilst this makes the SV technically "invalid" (as its
buffer is no longer followed by "\0") when fbm_compile()
adds the "\n" back, a "\0" is restored. */
- }
- fbm_compile(sv, flags);
+ fbm_compile(sv, FBMcf_TAIL);
+ } else
+ fbm_compile(sv, 0);
}
if (prog->substrs->data[i].substr == prog->check_substr)
prog->check_utf8 = sv;
SV* sv = newSVsv(prog->substrs->data[i].utf8_substr);
if (sv_utf8_downgrade(sv, TRUE)) {
if (SvVALID(prog->substrs->data[i].utf8_substr)) {
- const U8 flags
- = BmFLAGS(prog->substrs->data[i].utf8_substr);
- if (flags & FBMcf_TAIL) {
+ if (SvTAIL(prog->substrs->data[i].utf8_substr)) {
/* Trim the trailing \n that fbm_compile added last
time. */
SvCUR_set(sv, SvCUR(sv) - 1);
- }
- fbm_compile(sv, flags);
- }
+ fbm_compile(sv, FBMcf_TAIL);
+ } else
+ fbm_compile(sv, 0);
+ }
} else {
SvREFCNT_dec(sv);
sv = &PL_sv_undef;