have unique chars.
We use an array of integers to represent the character codes 0..255
- (trie->charmap) and we use a an HV* to store unicode characters. We use the
+ (trie->charmap) and we use a an HV* to store Unicode characters. We use the
native representation of the character value as the key and IV's for the
coded index.
return(NULL);
}
if (RExC_utf8 && !RExC_orig_utf8) {
- /* It's possible to write a regexp in ascii that represents unicode
+ /* It's possible to write a regexp in ascii that represents Unicode
codepoints outside of the byte range, such as via \x{100}. If we
detect such a sequence we have to convert the entire pattern to utf8
and then recompile, as our sizing calculation will have been based
r->prelen = plen;
r->extflags = pm_flags;
{
- bool has_k = ((r->extflags & RXf_PMf_KEEPCOPY) == RXf_PMf_KEEPCOPY);
+ bool has_p = ((r->extflags & RXf_PMf_KEEPCOPY) == RXf_PMf_KEEPCOPY);
bool has_minus = ((r->extflags & RXf_PMf_STD_PMMOD) != RXf_PMf_STD_PMMOD);
bool has_runon = ((RExC_seen & REG_SEEN_RUN_ON_COMMENT)==REG_SEEN_RUN_ON_COMMENT);
U16 reganch = (U16)((r->extflags & RXf_PMf_STD_PMMOD) >> 12);
const char *fptr = STD_PAT_MODS; /*"msix"*/
char *p;
- r->wraplen = r->prelen + has_minus + has_k + has_runon
+ r->wraplen = r->prelen + has_minus + has_p + has_runon
+ (sizeof(STD_PAT_MODS) - 1)
+ (sizeof("(?:)") - 1);
Newx(r->wrapped, r->wraplen + 1, char );
p = r->wrapped;
*p++='('; *p++='?';
- if (has_k)
- *p++ = KEEPCOPY_PAT_MOD; /*'k'*/
+ if (has_p)
+ *p++ = KEEPCOPY_PAT_MOD; /*'p'*/
{
char *r = p + (sizeof(STD_PAT_MODS) - 1) + has_minus - 1;
char *colon = r + 1;
/* Store the count of eval-groups for security checks: */
RExC_rx->seen_evals = RExC_seen_evals;
REGC((U8)REG_MAGIC, (char*) RExC_emit++);
- if (reg(pRExC_state, 0, &flags,1) == NULL)
+ if (reg(pRExC_state, 0, &flags,1) == NULL) {
+ ReREFCNT_dec(r);
return(NULL);
-
+ }
/* XXXX To minimize changes to RE engine we always allocate
3-units-long substrs field. */
Newx(r->substrs, 1, struct reg_substr_data);
#endif
/* Dig out information for optimizations. */
- r->extflags = pm_flags; /* Again? */
+ r->extflags = RExC_flags; /* was pm_op */
/*dmq: removed as part of de-PMOP: pm->op_pmflags = RExC_flags; */
if (UTF)
r->paren_names = (HV*)SvREFCNT_inc(RExC_paren_names);
else
r->paren_names = NULL;
- if (r->prelen == 3 && strnEQ("\\s+", r->precomp, 3)) /* precomp = "\\s+)" */
- r->extflags |= RXf_WHITE;
+
+#ifdef STUPID_PATTERN_CHECKS
+ if (r->extflags & RXf_SPLIT && r->prelen == 1 && r->precomp[0] == ' ')
+ /* XXX: this should happen BEFORE we compile */
+ r->extflags |= (RXf_SKIPWHITE|RXf_WHITE);
+ else if (r->prelen == 3 && memEQ("\\s+", r->precomp, 3))
+ r->extflags |= RXf_WHITE;
else if (r->prelen == 1 && r->precomp[0] == '^')
r->extflags |= RXf_START_ONLY;
-
+#else
+ if (r->extflags & RXf_SPLIT && r->prelen == 1 && r->precomp[0] == ' ')
+ /* XXX: this should happen BEFORE we compile */
+ r->extflags |= (RXf_SKIPWHITE|RXf_WHITE);
+ else {
+ regnode *first = ri->program + 1;
+ U8 fop = OP(first);
+ U8 nop = OP(NEXTOPER(first));
+
+ if (PL_regkind[fop] == BOL && nop == END)
+ r->extflags |= RXf_START_ONLY;
+ else if (fop == PLUS && nop ==SPACE && OP(regnext(first))==END)
+ r->extflags |= RXf_WHITE;
+ }
+#endif
#ifdef DEBUGGING
if (RExC_paren_names) {
ri->name_list_idx = add_data( pRExC_state, 1, "p" );
{
PERL_UNUSED_ARG(value);
- if (flags & RXf_HASH_FETCH) {
+ if (flags & RXapif_FETCH) {
return reg_named_buff_fetch(rx, key, flags);
- } else if (flags & (RXf_HASH_STORE | RXf_HASH_DELETE | RXf_HASH_CLEAR)) {
+ } else if (flags & (RXapif_STORE | RXapif_DELETE | RXapif_CLEAR)) {
Perl_croak(aTHX_ PL_no_modify);
return NULL;
- } else if (flags & RXf_HASH_EXISTS) {
+ } else if (flags & RXapif_EXISTS) {
return reg_named_buff_exists(rx, key, flags)
? &PL_sv_yes
: &PL_sv_no;
- } else if (flags & RXf_HASH_REGNAMES) {
+ } else if (flags & RXapif_REGNAMES) {
return reg_named_buff_all(rx, flags);
- } else if (flags & (RXf_HASH_SCALAR | RXf_HASH_REGNAMES_COUNT)) {
+ } else if (flags & (RXapif_SCALAR | RXapif_REGNAMES_COUNT)) {
return reg_named_buff_scalar(rx, flags);
} else {
Perl_croak(aTHX_ "panic: Unknown flags %d in named_buff", (int)flags);
{
PERL_UNUSED_ARG(lastkey);
- if (flags & RXf_HASH_FIRSTKEY)
+ if (flags & RXapif_FIRSTKEY)
return reg_named_buff_firstkey(rx, flags);
- else if (flags & RXf_HASH_NEXTKEY)
+ else if (flags & RXapif_NEXTKEY)
return reg_named_buff_nextkey(rx, flags);
else {
Perl_croak(aTHX_ "panic: Unknown flags %d in named_buff_iter", (int)flags);
{
AV *retarray = NULL;
SV *ret;
- if (flags & RXf_HASH_ALL)
+ if (flags & RXapif_ALL)
retarray=newAV();
if (rx && rx->paren_names) {
const U32 flags)
{
if (rx && rx->paren_names) {
- if (flags & RXf_HASH_ALL) {
+ if (flags & RXapif_ALL) {
return hv_exists_ent(rx->paren_names, key, 0);
} else {
- if (CALLREG_NAMED_BUFF_FETCH(rx, key, flags)) {
+ SV *sv = CALLREG_NAMED_BUFF_FETCH(rx, key, flags);
+ if (sv) {
+ SvREFCNT_dec(sv);
return TRUE;
} else {
return FALSE;
{
(void)hv_iterinit(rx->paren_names);
- return CALLREG_NAMED_BUFF_NEXTKEY(rx, NULL, flags & ~RXf_HASH_FIRSTKEY);
+ return CALLREG_NAMED_BUFF_NEXTKEY(rx, NULL, flags & ~RXapif_FIRSTKEY);
}
SV*
break;
}
}
- if (parno || flags & RXf_HASH_ALL) {
+ if (parno || flags & RXapif_ALL) {
STRLEN len;
char *pv = HePV(temphe, len);
return newSVpvn(pv,len);
I32 length;
if (rx && rx->paren_names) {
- if (flags & (RXf_HASH_ALL | RXf_HASH_REGNAMES_COUNT)) {
+ if (flags & (RXapif_ALL | RXapif_REGNAMES_COUNT)) {
return newSViv(HvTOTALKEYS(rx->paren_names));
- } else if (flags & RXf_HASH_ONE) {
- ret = CALLREG_NAMED_BUFF_ALL(rx, (flags | RXf_HASH_REGNAMES));
+ } else if (flags & RXapif_ONE) {
+ ret = CALLREG_NAMED_BUFF_ALL(rx, (flags | RXapif_REGNAMES));
av = (AV*)SvRV(ret);
length = av_len(av);
return newSViv(length + 1);
break;
}
}
- if (parno || flags & RXf_HASH_ALL) {
+ if (parno || flags & RXapif_ALL) {
STRLEN len;
char *pv = HePV(temphe, len);
av_push(av, newSVpvn(pv,len));
return;
}
else
- if (paren == RXf_PREMATCH && rx->offs[0].start != -1) {
+ if (paren == RX_BUFF_IDX_PREMATCH && rx->offs[0].start != -1) {
/* $` */
i = rx->offs[0].start;
s = rx->subbeg;
}
else
- if (paren == RXf_POSTMATCH && rx->offs[0].end != -1) {
+ if (paren == RX_BUFF_IDX_POSTMATCH && rx->offs[0].end != -1) {
/* $' */
s = rx->subbeg + rx->offs[0].end;
i = rx->sublen - rx->offs[0].end;
/* Some of this code was originally in C<Perl_magic_len> in F<mg.c> */
switch (paren) {
/* $` / ${^PREMATCH} */
- case RXf_PREMATCH:
+ case RX_BUFF_IDX_PREMATCH:
if (rx->offs[0].start != -1) {
i = rx->offs[0].start;
if (i > 0) {
}
return 0;
/* $' / ${^POSTMATCH} */
- case RXf_POSTMATCH:
+ case RX_BUFF_IDX_POSTMATCH:
if (rx->offs[0].end != -1) {
i = rx->sublen - rx->offs[0].end;
if (i > 0) {
register regnode *ender = NULL;
register I32 parno = 0;
I32 flags;
- const I32 oregflags = RExC_flags;
+ U32 oregflags = RExC_flags;
bool have_branch = 0;
bool is_open = 0;
I32 freeze_paren = 0;
and must be globally applied -- japhy */
switch (*RExC_parse) {
CASE_STD_PMMOD_FLAGS_PARSE_SET(flagsp);
- case 'o':
- case 'g':
+ case ONCE_PAT_MOD: /* 'o' */
+ case GLOBAL_PAT_MOD: /* 'g' */
if (SIZE_ONLY && ckWARN(WARN_REGEXP)) {
const I32 wflagbit = *RExC_parse == 'o' ? WASTED_O : WASTED_G;
if (! (wastedflags & wflagbit) ) {
}
break;
- case 'c':
+ case CONTINUE_PAT_MOD: /* 'c' */
if (SIZE_ONLY && ckWARN(WARN_REGEXP)) {
if (! (wastedflags & WASTED_C) ) {
wastedflags |= WASTED_GC;
}
}
break;
- case 'k':
+ case KEEPCOPY_PAT_MOD: /* 'p' */
if (flagsp == &negflags) {
if (SIZE_ONLY && ckWARN(WARN_REGEXP))
- vWARN(RExC_parse + 1,"Useless use of (?-k)");
+ vWARN(RExC_parse + 1,"Useless use of (?-p)");
} else {
*flagsp |= RXf_PMf_KEEPCOPY;
}
case ')':
RExC_flags |= posflags;
RExC_flags &= ~negflags;
+ if (paren != ':') {
+ oregflags |= posflags;
+ oregflags &= ~negflags;
+ }
nextchar(pRExC_state);
if (paren != ':') {
*flagp = TRYAGAIN;
/* RExC_parse points at the beginning brace,
endbrace points at the last */
if ( name[0]=='U' && name[1]=='+' ) {
- /* its a "unicode hex" notation {U+89AB} */
+ /* its a "Unicode hex" notation {U+89AB} */
I32 fl = PERL_SCAN_ALLOW_UNDERSCORES
| PERL_SCAN_DISALLOW_PREFIX
| (SIZE_ONLY ? PERL_SCAN_SILENT_ILLDIGIT : 0);
goto parse_named_seq;
} }
num = atoi(RExC_parse);
+ if (isg && num == 0)
+ vFAIL("Reference to invalid group 0");
if (isrel) {
num = RExC_npar - num;
if (num < 1)
case 'h': case 'H': /* HORIZWS */
case 'k': case 'K': /* named backref, keep marker */
case 'N': /* named char sequence */
- case 'p': case 'P': /* unicode property */
+ case 'p': case 'P': /* Unicode property */
case 'R': /* LNBREAK */
case 's': case 'S': /* space class */
case 'v': case 'V': /* VERTWS */
{
if (isLOWER(prevvalue)) {
for (i = prevvalue; i <= ceilvalue; i++)
- if (isLOWER(i))
+ if (isLOWER(i) && !ANYOF_BITMAP_TEST(ret,i)) {
+ stored++;
ANYOF_BITMAP_SET(ret, i);
+ }
} else {
for (i = prevvalue; i <= ceilvalue; i++)
- if (isUPPER(i))
+ if (isUPPER(i) && !ANYOF_BITMAP_TEST(ret,i)) {
+ stored++;
ANYOF_BITMAP_SET(ret, i);
+ }
}
}
else
/*
- regdump - dump a regexp onto Perl_debug_log in vaguely comprehensible form
*/
+#ifdef DEBUGGING
+void
+S_regdump_extflags(pTHX_ const char *lead, const U32 flags) {
+ int bit;
+ int set=0;
+ for (bit=0; bit<32; bit++) {
+ if (flags & (1<<bit)) {
+ if (!set++ && lead)
+ PerlIO_printf(Perl_debug_log, "%s",lead);
+ PerlIO_printf(Perl_debug_log, "%s ",PL_reg_extflags_name[bit]);
+ }
+ }
+ if (lead) {
+ if (set)
+ PerlIO_printf(Perl_debug_log, "\n");
+ else
+ PerlIO_printf(Perl_debug_log, "%s[none-set]\n",lead);
+ }
+}
+#endif
+
void
Perl_regdump(pTHX_ const regexp *r)
{
SV * const sv = sv_newmortal();
SV *dsv= sv_newmortal();
RXi_GET_DECL(r,ri);
+ GET_RE_DEBUG_FLAGS_DECL;
(void)dumpuntil(r, ri->program, ri->program + 1, NULL, NULL, sv, 0, 0);
if (r->extflags & RXf_EVAL_SEEN)
PerlIO_printf(Perl_debug_log, "with eval ");
PerlIO_printf(Perl_debug_log, "\n");
+ DEBUG_FLAGS_r(regdump_extflags("r->extflags: ",r->extflags));
#else
PERL_UNUSED_CONTEXT;
PERL_UNUSED_ARG(r);
projects / p5sagit/p5-mst-13.2.git / blobdiff