/* *These* symbols are masked to allow static link. */
# define Perl_regnext my_regnext
# define Perl_save_re_context my_save_re_context
-# define Perl_reginitcolors my_reginitcolors
+# define Perl_reginitcolors my_reginitcolors
# define PERL_NO_GET_CONTEXT
-#endif
+#endif
/*SUPPRESS 112*/
/*
*
**** Alterations to Henry's code are...
****
- **** Copyright (c) 1991-2000, Larry Wall
+ **** Copyright (c) 1991-2001, Larry Wall
****
**** You may distribute under the terms of either the GNU General Public
**** License or the Artistic License, as specified in the README file.
#define ISMULT1(c) ((c) == '*' || (c) == '+' || (c) == '?')
#define ISMULT2(s) ((*s) == '*' || (*s) == '+' || (*s) == '?' || \
((*s) == '{' && regcurly(s)))
-#ifdef atarist
-#define PERL_META "^$.[()|?+*\\"
-#else
-#define META "^$.[()|?+*\\"
-#endif
#ifdef SPSTART
#undef SPSTART /* dratted cpp namespace... */
* Forward declarations for pregcomp()'s friends.
*/
-static scan_data_t zero_scan_data = { 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
+static scan_data_t zero_scan_data = { 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
0, 0, 0, 0, 0, 0};
#define SF_BEFORE_EOL (SF_BEFORE_SEOL|SF_BEFORE_MEOL)
*/
#define MARKER1 "HERE" /* marker as it appears in the description */
#define MARKER2 " << HERE " /* marker as it appears within the regex */
-
+
#define REPORT_LOCATION " before " MARKER1 " mark in regex m/%.*s" MARKER2 "%s/"
/*
{
STRLEN l = CHR_SVLEN(data->last_found);
STRLEN old_l = CHR_SVLEN(*data->longest);
-
+
if ((l >= old_l) && ((l > old_l) || (data->flags & SF_BEFORE_EOL))) {
sv_setsv(*data->longest, data->last_found);
if (*data->longest == data->longest_fixed) {
data->offset_fixed = l ? data->last_start_min : data->pos_min;
if (data->flags & SF_BEFORE_EOL)
- data->flags
+ data->flags
|= ((data->flags & SF_BEFORE_EOL) << SF_FIX_SHIFT_EOL);
else
data->flags &= ~SF_FIX_BEFORE_EOL;
}
else {
data->offset_float_min = l ? data->last_start_min : data->pos_min;
- data->offset_float_max = (l
- ? data->last_start_max
+ data->offset_float_max = (l
+ ? data->last_start_max
: data->pos_min + data->pos_delta);
if (data->flags & SF_BEFORE_EOL)
- data->flags
+ data->flags
|= ((data->flags & SF_BEFORE_EOL) << SF_FL_SHIFT_EOL);
else
data->flags &= ~SF_FL_BEFORE_EOL;
} else {
/* (B1 | CL1) | (B2 | CL2) = (B1 | B2) | (CL1 | CL2)) */
if ( (or_with->flags & ANYOF_LOCALE) == (cl->flags & ANYOF_LOCALE)
- && (!(or_with->flags & ANYOF_FOLD)
+ && (!(or_with->flags & ANYOF_FOLD)
|| (cl->flags & ANYOF_FOLD)) ) {
int i;
I32 is_par = OP(scan) == OPEN ? ARG(scan) : 0;
scan_data_t data_fake;
struct regnode_charclass_class and_with; /* Valid if flags & SCF_DO_STCLASS_OR */
-
+
while (scan && OP(scan) != END && scan < last) {
/* Peephole optimizer: */
U32 stringok = 1;
#ifdef DEBUGGING
regnode *stop = scan;
-#endif
+#endif
next = scan + NODE_SZ_STR(scan);
/* Skip NOTHING, merge EXACT*. */
while (n &&
- ( PL_regkind[(U8)OP(n)] == NOTHING ||
+ ( PL_regkind[(U8)OP(n)] == NOTHING ||
(stringok && (OP(n) == OP(scan))))
&& NEXT_OFF(n)
&& NEXT_OFF(scan) + NEXT_OFF(n) < I16_MAX) {
#ifdef DEBUGGING
if (stringok)
stop = n;
-#endif
+#endif
n = regnext(n);
}
- else {
+ else if (stringok) {
int oldl = STR_LEN(scan);
regnode *nnext = regnext(n);
-
- if (oldl + STR_LEN(n) > U8_MAX)
+
+ if (oldl + STR_LEN(n) > U8_MAX)
break;
NEXT_OFF(scan) += NEXT_OFF(n);
STR_LEN(scan) += STR_LEN(n);
next = n + NODE_SZ_STR(n);
/* Now we can overwrite *n : */
- Move(STRING(n), STRING(scan) + oldl,
- STR_LEN(n), char);
+ Move(STRING(n), STRING(scan) + oldl, STR_LEN(n), char);
#ifdef DEBUGGING
- if (stringok)
- stop = next - 1;
-#endif
+ stop = next - 1;
+#endif
n = nnext;
}
}
int off = (reg_off_by_arg[OP(scan)] ? ARG(scan) : NEXT_OFF(scan));
int noff;
regnode *n = scan;
-
+
/* Skip NOTHING and LONGJMP. */
while ((n = regnext(n))
&& ((PL_regkind[(U8)OP(n)] == NOTHING && (noff = NEXT_OFF(n)))
off += noff;
if (reg_off_by_arg[OP(scan)])
ARG(scan) = off;
- else
+ else
NEXT_OFF(scan) = off;
}
/* The principal pseudo-switch. Cannot be a switch, since we
look into several different things. */
- if (OP(scan) == BRANCH || OP(scan) == BRANCHJ
+ if (OP(scan) == BRANCH || OP(scan) == BRANCHJ
|| OP(scan) == IFTHEN || OP(scan) == SUSPEND) {
next = regnext(scan);
code = OP(scan);
-
- if (OP(next) == code || code == IFTHEN || code == SUSPEND) {
+
+ if (OP(next) == code || code == IFTHEN || code == SUSPEND) {
I32 max1 = 0, min1 = I32_MAX, num = 0;
struct regnode_charclass_class accum;
num++;
data_fake.flags = 0;
- if (data) {
+ if (data) {
data_fake.whilem_c = data->whilem_c;
data_fake.last_closep = data->last_closep;
}
cl_init(pRExC_state, &this_class);
data_fake.start_class = &this_class;
f = SCF_DO_STCLASS_AND;
- }
+ }
if (flags & SCF_WHILEM_VISITED_POS)
f |= SCF_WHILEM_VISITED_POS;
/* we suppose the run is continuous, last=next...*/
minnext = study_chunk(pRExC_state, &scan, &deltanext,
next, &data_fake, f);
- if (min1 > minnext)
+ if (min1 > minnext)
min1 = minnext;
if (max1 < minnext + deltanext)
max1 = minnext + deltanext;
data->whilem_c = data_fake.whilem_c;
if (flags & SCF_DO_STCLASS)
cl_or(pRExC_state, &accum, &this_class);
- if (code == SUSPEND)
+ if (code == SUSPEND)
break;
}
if (code == IFTHEN && num < 2) /* Empty ELSE branch */
flags &= ~SCF_DO_STCLASS;
}
else {
- /* Switch to OR mode: cache the old value of
+ /* Switch to OR mode: cache the old value of
* data->start_class */
StructCopy(data->start_class, &and_with,
struct regnode_charclass_class);
if (UTF) {
U8 *s = (U8*)STRING(scan);
l = utf8_length(s, s + l);
- uc = utf8_to_uv_simple(s, NULL);
+ uc = utf8_to_uvchr(s, NULL);
}
min += l;
if (flags & SCF_DO_SUBSTR) { /* Update longest substr. */
if (data->last_end == -1) { /* Update the start info. */
data->last_start_min = data->pos_min;
data->last_start_max = is_inf
- ? I32_MAX : data->pos_min + data->pos_delta;
+ ? I32_MAX : data->pos_min + data->pos_delta;
}
sv_catpvn(data->last_found, STRING(scan), STR_LEN(scan));
data->last_end = data->pos_min + l;
int compat = 1;
if (uc >= 0x100 ||
- !(data->start_class->flags & (ANYOF_CLASS | ANYOF_LOCALE))
+ !(data->start_class->flags & (ANYOF_CLASS | ANYOF_LOCALE))
&& !ANYOF_BITMAP_TEST(data->start_class, uc)
&& (!(data->start_class->flags & ANYOF_FOLD)
|| !ANYOF_BITMAP_TEST(data->start_class, PL_fold[uc])))
if (compat)
ANYOF_BITMAP_SET(data->start_class, uc);
data->start_class->flags &= ~ANYOF_EOS;
+ if (uc < 0x100)
+ data->start_class->flags &= ~ANYOF_UNICODE_ALL;
}
else if (flags & SCF_DO_STCLASS_OR) {
/* false positive possible if the class is case-folded */
if (uc < 0x100)
- ANYOF_BITMAP_SET(data->start_class, uc);
+ ANYOF_BITMAP_SET(data->start_class, uc);
+ else
+ data->start_class->flags |= ANYOF_UNICODE_ALL;
data->start_class->flags &= ~ANYOF_EOS;
cl_and(data->start_class, &and_with);
}
UV uc = *((U8*)STRING(scan));
/* Search for fixed substrings supports EXACT only. */
- if (flags & SCF_DO_SUBSTR)
+ if (flags & SCF_DO_SUBSTR)
scan_commit(pRExC_state, data);
if (UTF) {
U8 *s = (U8 *)STRING(scan);
l = utf8_length(s, s + l);
- uc = utf8_to_uv_simple(s, NULL);
+ uc = utf8_to_uvchr(s, NULL);
}
min += l;
if (data && (flags & SCF_DO_SUBSTR))
int compat = 1;
if (uc >= 0x100 ||
- !(data->start_class->flags & (ANYOF_CLASS | ANYOF_LOCALE))
+ !(data->start_class->flags & (ANYOF_CLASS | ANYOF_LOCALE))
&& !ANYOF_BITMAP_TEST(data->start_class, uc)
&& !ANYOF_BITMAP_TEST(data->start_class, PL_fold[uc]))
compat = 0;
if (flags & (SCF_DO_SUBSTR | SCF_DO_STCLASS)) {
next = NEXTOPER(scan);
if (OP(next) == EXACT || (flags & SCF_DO_STCLASS)) {
- mincount = 1;
- maxcount = REG_INFTY;
+ mincount = 1;
+ maxcount = REG_INFTY;
next = regnext(scan);
scan = NEXTOPER(scan);
goto do_curly;
case STAR:
if (flags & SCF_DO_STCLASS) {
mincount = 0;
- maxcount = REG_INFTY;
+ maxcount = REG_INFTY;
next = regnext(scan);
scan = NEXTOPER(scan);
goto do_curly;
}
- is_inf = is_inf_internal = 1;
+ is_inf = is_inf_internal = 1;
scan = regnext(scan);
if (flags & SCF_DO_SUBSTR) {
scan_commit(pRExC_state, data); /* Cannot extend fixed substrings */
}
goto optimize_curly_tail;
case CURLY:
- mincount = ARG1(scan);
+ mincount = ARG1(scan);
maxcount = ARG2(scan);
next = regnext(scan);
if (OP(scan) == CURLYX) {
/* These are the cases when once a subexpression
fails at a particular position, it cannot succeed
even after backtracking at the enclosing scope.
-
+
XXXX what if minimal match and we are at the
initial run of {n,m}? */
if ((mincount != maxcount - 1) && (maxcount != REG_INFTY))
f &= ~SCF_WHILEM_VISITED_POS;
/* This will finish on WHILEM, setting scan, or on NULL: */
- minnext = study_chunk(pRExC_state, &scan, &deltanext, last, data,
- mincount == 0
+ minnext = study_chunk(pRExC_state, &scan, &deltanext, last, data,
+ mincount == 0
? (f & ~SCF_DO_SUBSTR) : f);
if (flags & SCF_DO_STCLASS)
cl_or(pRExC_state, data->start_class, &this_class);
}
else if (flags & SCF_DO_STCLASS_AND) {
- /* Switch to OR mode: cache the old value of
+ /* Switch to OR mode: cache the old value of
* data->start_class */
StructCopy(data->start_class, &and_with,
struct regnode_charclass_class);
}
if (!scan) /* It was not CURLYX, but CURLY. */
scan = next;
- if (ckWARN(WARN_REGEXP) && (minnext + deltanext == 0)
+ if (ckWARN(WARN_REGEXP) && (minnext + deltanext == 0)
&& !(data->flags & (SF_HAS_PAR|SF_IN_PAR))
&& maxcount <= REG_INFTY/3) /* Complement check for big count */
{
}
min += minnext * mincount;
- is_inf_internal |= ((maxcount == REG_INFTY
+ is_inf_internal |= ((maxcount == REG_INFTY
&& (minnext + deltanext) > 0)
|| deltanext == I32_MAX);
is_inf |= is_inf_internal;
delta += (minnext + deltanext) * maxcount - minnext * mincount;
/* Try powerful optimization CURLYX => CURLYN. */
- if ( OP(oscan) == CURLYX && data
+ if ( OP(oscan) == CURLYX && data
&& data->flags & SF_IN_PAR
&& !(data->flags & SF_HAS_EVAL)
&& !deltanext && minnext == 1 ) {
nxt = regnext(nxt);
if (!strchr((char*)PL_simple,OP(nxt))
&& !(PL_regkind[(U8)OP(nxt)] == EXACT
- && STR_LEN(nxt) == 1))
+ && STR_LEN(nxt) == 1))
goto nogo;
nxt2 = nxt;
nxt = regnext(nxt);
- if (OP(nxt) != CLOSE)
+ if (OP(nxt) != CLOSE)
goto nogo;
/* Now we know that nxt2 is the only contents: */
oscan->flags = ARG(nxt);
OP(nxt) = OPTIMIZED; /* was CLOSE. */
OP(nxt + 1) = OPTIMIZED; /* was count. */
NEXT_OFF(nxt+ 1) = 0; /* just for consistancy. */
-#endif
+#endif
}
nogo:
/* Try optimization CURLYX => CURLYM. */
- if ( OP(oscan) == CURLYX && data
+ if ( OP(oscan) == CURLYX && data
&& !(data->flags & SF_HAS_PAR)
&& !(data->flags & SF_HAS_EVAL)
&& !deltanext ) {
OP(oscan) = CURLYM;
while ( (nxt2 = regnext(nxt)) /* skip over embedded stuff*/
- && (OP(nxt2) != WHILEM))
+ && (OP(nxt2) != WHILEM))
nxt = nxt2;
OP(nxt2) = SUCCEED; /* Whas WHILEM */
/* Need to optimize away parenths. */
/* Set the parenth number. */
regnode *nxt1 = NEXTOPER(oscan) + EXTRA_STEP_2ARGS; /* OPEN*/
- if (OP(nxt) != CLOSE)
+ if (OP(nxt) != CLOSE)
FAIL("Panic opt close");
oscan->flags = ARG(nxt);
OP(nxt1) = OPTIMIZED; /* was OPEN. */
OP(nxt + 1) = OPTIMIZED; /* was count. */
NEXT_OFF(nxt1 + 1) = 0; /* just for consistancy. */
NEXT_OFF(nxt + 1) = 0; /* just for consistancy. */
-#endif
+#endif
#if 0
while ( nxt1 && (OP(nxt1) != WHILEM)) {
regnode *nnxt = regnext(nxt1);
-
+
if (nnxt == nxt) {
if (reg_off_by_arg[OP(nxt1)])
ARG_SET(nxt1, nxt2 - nxt1);
}
#endif
/* Optimize again: */
- study_chunk(pRExC_state, &nxt1, &deltanext, nxt,
+ study_chunk(pRExC_state, &nxt1, &deltanext, nxt,
NULL, 0);
}
else
PREVOPER(nxt)->flags = data->whilem_c
| (RExC_whilem_seen << 4); /* On WHILEM */
}
- if (data && fl & (SF_HAS_PAR|SF_IN_PAR))
+ if (data && fl & (SF_HAS_PAR|SF_IN_PAR))
pars++;
if (flags & SCF_DO_SUBSTR) {
SV *last_str = Nullsv;
int counted = mincount != 0;
if (data->last_end > 0 && mincount != 0) { /* Ends with a string. */
- I32 b = pos_before >= data->last_start_min
+ I32 b = pos_before >= data->last_start_min
? pos_before : data->last_start_min;
STRLEN l;
char *s = SvPV(data->last_found, l);
/* What was added is a constant string */
if (mincount > 1) {
SvGROW(last_str, (mincount * l) + 1);
- repeatcpy(SvPVX(last_str) + l,
+ repeatcpy(SvPVX(last_str) + l,
SvPVX(last_str), l, mincount - 1);
SvCUR(last_str) *= mincount;
/* Add additional parts. */
- SvCUR_set(data->last_found,
+ SvCUR_set(data->last_found,
SvCUR(data->last_found) - l);
sv_catsv(data->last_found, last_str);
data->last_end += l * (mincount - 1);
if (mincount && last_str) {
sv_setsv(data->last_found, last_str);
data->last_end = data->pos_min;
- data->last_start_min =
+ data->last_start_min =
data->pos_min - CHR_SVLEN(last_str);
- data->last_start_max = is_inf
- ? I32_MAX
+ data->last_start_max = is_inf
+ ? I32_MAX
: data->pos_min + data->pos_delta
- CHR_SVLEN(last_str);
}
else {
for (value = 0; value < 256; value++)
if (isALNUM(value))
- ANYOF_BITMAP_SET(data->start_class, value);
+ ANYOF_BITMAP_SET(data->start_class, value);
}
}
break;
else {
for (value = 0; value < 256; value++)
if (!isALNUM(value))
- ANYOF_BITMAP_SET(data->start_class, value);
+ ANYOF_BITMAP_SET(data->start_class, value);
}
}
break;
else {
for (value = 0; value < 256; value++)
if (isSPACE(value))
- ANYOF_BITMAP_SET(data->start_class, value);
+ ANYOF_BITMAP_SET(data->start_class, value);
}
}
break;
else {
for (value = 0; value < 256; value++)
if (!isSPACE(value))
- ANYOF_BITMAP_SET(data->start_class, value);
+ ANYOF_BITMAP_SET(data->start_class, value);
}
}
break;
else {
for (value = 0; value < 256; value++)
if (isDIGIT(value))
- ANYOF_BITMAP_SET(data->start_class, value);
+ ANYOF_BITMAP_SET(data->start_class, value);
}
}
break;
else {
for (value = 0; value < 256; value++)
if (!isDIGIT(value))
- ANYOF_BITMAP_SET(data->start_class, value);
+ ANYOF_BITMAP_SET(data->start_class, value);
}
}
break;
int f = 0;
data_fake.flags = 0;
- if (data) {
+ if (data) {
data_fake.whilem_c = data->whilem_c;
data_fake.last_closep = data->last_closep;
}
finish:
*scanp = scan;
*deltap = is_inf_internal ? I32_MAX : delta;
- if (flags & SCF_DO_SUBSTR && is_inf)
+ if (flags & SCF_DO_SUBSTR && is_inf)
data->pos_delta = I32_MAX - data->pos_min;
if (is_par > U8_MAX)
is_par = 0;
S_add_data(pTHX_ RExC_state_t *pRExC_state, I32 n, char *s)
{
if (RExC_rx->data) {
- Renewc(RExC_rx->data,
- sizeof(*RExC_rx->data) + sizeof(void*) * (RExC_rx->data->count + n - 1),
+ Renewc(RExC_rx->data,
+ sizeof(*RExC_rx->data) + sizeof(void*) * (RExC_rx->data->count + n - 1),
char, struct reg_data);
Renew(RExC_rx->data->what, RExC_rx->data->count + n, U8);
RExC_rx->data->count += n;
{
int i = 0;
char *s = PerlEnv_getenv("PERL_RE_COLORS");
-
+
if (s) {
PL_colors[0] = s = savepv(s);
while (++i < 6) {
PL_colors[i] = s = "";
}
} else {
- while (i < 6)
+ while (i < 6)
PL_colors[i++] = "";
}
PL_colorset = 1;
first = NEXTOPER(first);
goto again;
}
- if (sawplus && (!sawopen || !RExC_sawback)
+ if (sawplus && (!sawopen || !RExC_sawback)
&& !(RExC_seen & REG_SEEN_EVAL)) /* May examine pos and $& */
/* x+ must match at the 1st pos of run of x's */
r->reganch |= ROPT_SKIP;
/* Scan is after the zeroth branch, first is atomic matcher. */
- DEBUG_r(PerlIO_printf(Perl_debug_log, "first at %"IVdf"\n",
+ DEBUG_r(PerlIO_printf(Perl_debug_log, "first at %"IVdf"\n",
(IV)(first - scan + 1)));
/*
* If there's something expensive in the r.e., find the
minlen = study_chunk(pRExC_state, &first, &fake, scan + RExC_size, /* Up to end */
&data, SCF_DO_SUBSTR | SCF_WHILEM_VISITED_POS | stclass_flag);
if ( RExC_npar == 1 && data.longest == &(data.longest_fixed)
- && data.last_start_min == 0 && data.last_end > 0
+ && data.last_start_min == 0 && data.last_end > 0
&& !RExC_seen_zerolen
&& (!(RExC_seen & REG_SEEN_GPOS) || (r->reganch & ROPT_ANCH_GPOS)))
r->reganch |= ROPT_CHECK_ALL;
SvREFCNT_dec(data.longest_fixed);
longest_fixed_length = 0;
}
- if (r->regstclass
+ if (r->regstclass
&& (OP(r->regstclass) == REG_ANY || OP(r->regstclass) == SANY))
r->regstclass = NULL;
if ((!r->anchored_substr || r->anchored_offset) && stclass_flag
SV *sv;
I32 n = add_data(pRExC_state, 1, "f");
- New(1006, RExC_rx->data->data[n], 1,
+ New(1006, RExC_rx->data->data[n], 1,
struct regnode_charclass_class);
StructCopy(data.start_class,
(struct regnode_charclass_class*)RExC_rx->data->data[n],
SV *sv;
I32 n = add_data(pRExC_state, 1, "f");
- New(1006, RExC_rx->data->data[n], 1,
+ New(1006, RExC_rx->data->data[n], 1,
struct regnode_charclass_class);
StructCopy(data.start_class,
(struct regnode_charclass_class*)RExC_rx->data->data[n],
}
r->minlen = minlen;
- if (RExC_seen & REG_SEEN_GPOS)
+ if (RExC_seen & REG_SEEN_GPOS)
r->reganch |= ROPT_GPOS_SEEN;
if (RExC_seen & REG_SEEN_LOOKBEHIND)
r->reganch |= ROPT_LOOKBEHIND_SEEN;
if (RExC_seen & REG_SEEN_EVAL)
r->reganch |= ROPT_EVAL_SEEN;
+ if (RExC_seen & REG_SEEN_SANY)
+ r->reganch |= ROPT_SANY_SEEN;
Newz(1002, r->startp, RExC_npar, I32);
Newz(1002, r->endp, RExC_npar, I32);
PL_regdata = r->data; /* for regprop() */
switch (paren) {
case '<':
RExC_seen |= REG_SEEN_LOOKBEHIND;
- if (*RExC_parse == '!')
+ if (*RExC_parse == '!')
paren = ',';
- if (*RExC_parse != '=' && *RExC_parse != '!')
+ if (*RExC_parse != '=' && *RExC_parse != '!')
goto unknown;
RExC_parse++;
case '=':
while (count && (c = *RExC_parse)) {
if (c == '\\' && RExC_parse[1])
RExC_parse++;
- else if (c == '{')
+ else if (c == '{')
count++;
- else if (c == '}')
+ else if (c == '}')
count--;
RExC_parse++;
}
if (*RExC_parse != ')')
{
- RExC_parse = s;
+ RExC_parse = s;
vFAIL("Sequence (?{...}) not terminated or not {}-balanced");
}
if (!SIZE_ONLY) {
AV *av;
-
- if (RExC_parse - 1 - s)
+
+ if (RExC_parse - 1 - s)
sv = newSVpvn(s, RExC_parse - 1 - s);
else
sv = newSVpvn("", 0);
case '(':
{
if (RExC_parse[0] == '?') {
- if (RExC_parse[1] == '=' || RExC_parse[1] == '!'
- || RExC_parse[1] == '<'
+ if (RExC_parse[1] == '=' || RExC_parse[1] == '!'
+ || RExC_parse[1] == '<'
|| RExC_parse[1] == '{') { /* Lookahead or eval. */
I32 flag;
ret->flags = 1;
regtail(pRExC_state, ret, reg(pRExC_state, 1, &flag));
goto insert_if;
- }
+ }
}
else if (RExC_parse[0] >= '1' && RExC_parse[0] <= '9' ) {
parno = atoi(RExC_parse++);
register regnode *latest;
I32 flags = 0, c = 0;
- if (first)
+ if (first)
ret = NULL;
else {
- if (!SIZE_ONLY && RExC_extralen)
+ if (!SIZE_ONLY && RExC_extralen)
ret = reganode(pRExC_state, BRANCHJ,0);
else
ret = reg_node(pRExC_state, BRANCH);
}
- if (!first && SIZE_ONLY)
+ if (!first && SIZE_ONLY)
RExC_extralen += 1; /* BRANCHJ */
-
+
*flagp = WORST; /* Tentatively. */
RExC_parse--;
if (!(flags&HASWIDTH) && op != '?')
vFAIL("Regexp *+ operand could be empty");
-#endif
+#endif
nextchar(pRExC_state);
break;
case '$':
nextchar(pRExC_state);
- if (*RExC_parse)
+ if (*RExC_parse)
RExC_seen_zerolen++;
if (RExC_flags16 & PMf_MULTILINE)
ret = reg_node(pRExC_state, MEOL);
break;
case 'C':
ret = reg_node(pRExC_state, SANY);
+ RExC_seen |= REG_SEEN_SANY;
*flagp |= HASWIDTH|SIMPLE;
nextchar(pRExC_state);
break;
ret = reg_node(pRExC_state, CLUMP);
*flagp |= HASWIDTH;
nextchar(pRExC_state);
- if (UTF && !PL_utf8_mark)
- is_utf8_mark((U8*)"~"); /* preload table */
break;
case 'w':
ret = reg_node(pRExC_state, LOC ? ALNUML : ALNUM);
*flagp |= HASWIDTH|SIMPLE;
nextchar(pRExC_state);
- if (UTF && !PL_utf8_alnum)
- is_utf8_alnum((U8*)"a"); /* preload table */
break;
case 'W':
ret = reg_node(pRExC_state, LOC ? NALNUML : NALNUM);
*flagp |= HASWIDTH|SIMPLE;
nextchar(pRExC_state);
- if (UTF && !PL_utf8_alnum)
- is_utf8_alnum((U8*)"a"); /* preload table */
break;
case 'b':
RExC_seen_zerolen++;
ret = reg_node(pRExC_state, LOC ? BOUNDL : BOUND);
*flagp |= SIMPLE;
nextchar(pRExC_state);
- if (UTF && !PL_utf8_alnum)
- is_utf8_alnum((U8*)"a"); /* preload table */
break;
case 'B':
RExC_seen_zerolen++;
ret = reg_node(pRExC_state, LOC ? NBOUNDL : NBOUND);
*flagp |= SIMPLE;
nextchar(pRExC_state);
- if (UTF && !PL_utf8_alnum)
- is_utf8_alnum((U8*)"a"); /* preload table */
break;
case 's':
ret = reg_node(pRExC_state, LOC ? SPACEL : SPACE);
*flagp |= HASWIDTH|SIMPLE;
nextchar(pRExC_state);
- if (UTF && !PL_utf8_space)
- is_utf8_space((U8*)" "); /* preload table */
break;
case 'S':
ret = reg_node(pRExC_state, LOC ? NSPACEL : NSPACE);
*flagp |= HASWIDTH|SIMPLE;
nextchar(pRExC_state);
- if (UTF && !PL_utf8_space)
- is_utf8_space((U8*)" "); /* preload table */
break;
case 'd':
ret = reg_node(pRExC_state, DIGIT);
*flagp |= HASWIDTH|SIMPLE;
nextchar(pRExC_state);
- if (UTF && !PL_utf8_digit)
- is_utf8_digit((U8*)"1"); /* preload table */
break;
case 'D':
ret = reg_node(pRExC_state, NDIGIT);
*flagp |= HASWIDTH|SIMPLE;
nextchar(pRExC_state);
- if (UTF && !PL_utf8_digit)
- is_utf8_digit((U8*)"1"); /* preload table */
break;
case 'p':
case 'P':
p++;
break;
case 'e':
-#ifdef ASCIIish
- ender = '\033';
-#else
- ender = '\047';
-#endif
+ ender = ASCII_TO_NATIVE('\033');
p++;
break;
case 'a':
-#ifdef ASCIIish
- ender = '\007';
-#else
- ender = '\057';
-#endif
+ ender = ASCII_TO_NATIVE('\007');
p++;
break;
case 'x':
if (*++p == '{') {
char* e = strchr(p, '}');
-
+
if (!e) {
RExC_parse = p + 1;
vFAIL("Missing right brace on \\x{}");
else {
numlen = 1; /* allow underscores */
ender = (UV)scan_hex(p + 1, e - p - 1, &numlen);
+ if (ender > 0xff)
+ RExC_utf8 = 1;
/* numlen is generous */
if (numlen + len >= 127) {
p--;
break;
default:
normal_default:
- if ((*p & 0xc0) == 0xc0 && UTF) {
- ender = utf8_to_uv((U8*)p, RExC_end - p,
+ if (UTF8_IS_START(*p) && UTF) {
+ ender = utf8n_to_uvchr((U8*)p, RExC_end - p,
&numlen, 0);
p += numlen;
}
p = regwhite(p, RExC_end);
if (UTF && FOLD) {
if (LOC)
- ender = toLOWER_LC_uni(ender);
+ ender = toLOWER_LC_uvchr(ender);
else
ender = toLOWER_uni(ender);
}
if (ISMULT2(p)) { /* Back off on ?+*. */
if (len)
p = oldp;
- else if (ender >= 0x80 && UTF) {
+ /* ender is a Unicode value so it can be > 0xff --
+ * in other words, do not use UTF8_IS_CONTINUED(). */
+ else if (NATIVE_TO_ASCII(ender) >= 0x80 && UTF) {
reguni(pRExC_state, ender, s, &numlen);
s += numlen;
len += numlen;
}
break;
}
- if (ender >= 0x80 && UTF) {
+ /* ender is a Unicode value so it can be > 0xff --
+ * in other words, do not use UTF8_IS_CONTINUED(). */
+ if (NATIVE_TO_ASCII(ender) >= 0x80 && UTF) {
reguni(pRExC_state, ender, s, &numlen);
s += numlen;
len += numlen - 1;
*RExC_parse == '.')) {
char c = *RExC_parse;
char* s = RExC_parse++;
-
+
while (RExC_parse < RExC_end && *RExC_parse != c)
RExC_parse++;
if (RExC_parse == RExC_end)
if (!range)
rangebegin = RExC_parse;
if (UTF) {
- value = utf8_to_uv((U8*)RExC_parse,
+ value = utf8n_to_uvchr((U8*)RExC_parse,
RExC_end - RExC_parse,
&numlen, 0);
RExC_parse += numlen;
namedclass = regpposixcc(pRExC_state, value);
else if (value == '\\') {
if (UTF) {
- value = utf8_to_uv((U8*)RExC_parse,
+ value = utf8n_to_uvchr((U8*)RExC_parse,
RExC_end - RExC_parse,
&numlen, 0);
RExC_parse += numlen;
case 't': value = '\t'; break;
case 'f': value = '\f'; break;
case 'b': value = '\b'; break;
-#ifdef ASCIIish
- case 'e': value = '\033'; break;
- case 'a': value = '\007'; break;
-#else
- case 'e': value = '\047'; break;
- case 'a': value = '\057'; break;
-#endif
+ case 'e': value = ASCII_TO_NATIVE('\033');break;
+ case 'a': value = ASCII_TO_NATIVE('\007');break;
case 'x':
if (*RExC_parse == '{') {
e = strchr(RExC_parse++, '}');
- if (!e)
+ if (!e)
vFAIL("Missing right brace on \\x{}");
numlen = 1; /* allow underscores */
value = (UV)scan_hex(RExC_parse,
else {
ANYOF_FLAGS(ret) |= ANYOF_UNICODE;
Perl_sv_catpvf(aTHX_ listsv,
- /* 0x002D is Unicode for '-' */
- "%04"UVxf"\n002D\n", (UV)lastvalue);
+ "%04"UVxf"\n%04"UVxf"\n", (UV)lastvalue, (UV) '-');
}
}
if (LOC)
ANYOF_CLASS_SET(ret, ANYOF_ASCII);
else {
-#ifdef ASCIIish
+#ifndef EBCDIC
for (value = 0; value < 128; value++)
ANYOF_BITMAP_SET(ret, value);
#else /* EBCDIC */
if (LOC)
ANYOF_CLASS_SET(ret, ANYOF_NASCII);
else {
-#ifdef ASCIIish
+#ifndef EBCDIC
for (value = 128; value < 256; value++)
ANYOF_BITMAP_SET(ret, value);
#else /* EBCDIC */
/* now is the next time */
if (!SIZE_ONLY) {
if (lastvalue < 256 && value < 256) {
-#ifndef ASCIIish /* EBCDIC, for example. */
+#ifdef EBCDIC /* EBCDIC, for example. */
if ((isLOWER(lastvalue) && isLOWER(value)) ||
(isUPPER(lastvalue) && isUPPER(value)))
{
ANYOF_FLAGS(ret) = ANYOF_UNICODE_ALL;
}
- if (!SIZE_ONLY) {
+ if (!SIZE_ONLY) {
AV *av = newAV();
SV *rv;
STATIC void
S_reguni(pTHX_ RExC_state_t *pRExC_state, UV uv, char* s, STRLEN* lenp)
{
- *lenp = SIZE_ONLY ? UNISKIP(uv) : (uv_to_utf8((U8*)s, uv) - (U8*)s);
+ *lenp = SIZE_ONLY ? UNISKIP(uv) : (uvchr_to_utf8((U8*)s, uv) - (U8*)s);
}
/*
register regnode *dst;
register regnode *place;
register int offset = regarglen[(U8)op];
-
+
/* (PL_regkind[(U8)op] == CURLY ? EXTRA_STEP_2ARGS : 0); */
if (SIZE_ONLY) {
(int)(2*l + 1), "", SvPVX(sv));
if (next == NULL) /* Next ptr. */
PerlIO_printf(Perl_debug_log, "(0)");
- else
+ else
PerlIO_printf(Perl_debug_log, "(%"IVdf")", (IV)(next - start));
(void)PerlIO_putc(Perl_debug_log, '\n');
after_print:
if (PL_regkind[(U8)op] == BRANCHJ) {
- register regnode *nnode = (OP(next) == LONGJMP
- ? regnext(next)
+ register regnode *nnode = (OP(next) == LONGJMP
+ ? regnext(next)
: next);
if (last && nnode > last)
nnode = last;
/* Header fields of interest. */
if (r->anchored_substr)
PerlIO_printf(Perl_debug_log,
- "anchored `%s%.*s%s'%s at %"IVdf" ",
+ "anchored `%s%.*s%s'%s at %"IVdf" ",
PL_colors[0],
(int)(SvCUR(r->anchored_substr) - (SvTAIL(r->anchored_substr)!=0)),
- SvPVX(r->anchored_substr),
+ SvPVX(r->anchored_substr),
PL_colors[1],
SvTAIL(r->anchored_substr) ? "$" : "",
(IV)r->anchored_offset);
if (r->float_substr)
PerlIO_printf(Perl_debug_log,
- "floating `%s%.*s%s'%s at %"IVdf"..%"UVuf" ",
+ "floating `%s%.*s%s'%s at %"IVdf"..%"UVuf" ",
PL_colors[0],
- (int)(SvCUR(r->float_substr) - (SvTAIL(r->float_substr)!=0)),
+ (int)(SvCUR(r->float_substr) - (SvTAIL(r->float_substr)!=0)),
SvPVX(r->float_substr),
PL_colors[1],
SvTAIL(r->float_substr) ? "$" : "",
(IV)r->float_min_offset, (UV)r->float_max_offset);
if (r->check_substr)
- PerlIO_printf(Perl_debug_log,
- r->check_substr == r->float_substr
+ PerlIO_printf(Perl_debug_log,
+ r->check_substr == r->float_substr
? "(checking floating" : "(checking anchored");
if (r->reganch & ROPT_NOSCAN)
PerlIO_printf(Perl_debug_log, " noscan");
{
SV *lv;
SV *sw = regclass_swash(o, FALSE, &lv);
-
+
if (lv) {
if (sw) {
UV i;
U8 s[UTF8_MAXLEN+1];
-
+
for (i = 0; i <= 256; i++) { /* just the first 256 */
- U8 *e = uv_to_utf8(s, i);
+ U8 *e = uvuni_to_utf8(s, i);
if (i < 256 && swash_fetch(sw, s)) {
if (rangestart == -1)
rangestart = i;
} else if (rangestart != -1) {
U8 *p;
-
+
if (i <= rangestart + 3)
for (; rangestart < i; rangestart++) {
- for(e = uv_to_utf8(s, rangestart), p = s; p < e; p++)
+ for(e = uvuni_to_utf8(s, rangestart), p = s; p < e; p++)
put_byte(sv, *p);
}
else {
- for (e = uv_to_utf8(s, rangestart), p = s; p < e; p++)
+ for (e = uvuni_to_utf8(s, rangestart), p = s; p < e; p++)
put_byte(sv, *p);
sv_catpv(sv, "-");
- for (e = uv_to_utf8(s, i - 1), p = s; p < e; p++)
+ for (e = uvuni_to_utf8(s, i - 1), p = s; p < e; p++)
put_byte(sv, *p);
}
rangestart = -1;
{
char *s = savepv(SvPVX(lv));
char *origs = s;
-
+
while(*s && *s != '\n') s++;
-
+
if (*s == '\n') {
char *t = ++s;
sv_catpv(sv, t);
}
-
+
Safefree(origs);
}
}
void
Perl_save_re_context(pTHX)
-{
+{
#if 0
SAVEPPTR(RExC_precomp); /* uncompiled string. */
SAVEI32(RExC_npar); /* () count. */
SAVEVPTR(PL_reglastparen); /* Similarly for lastparen. */
SAVEPPTR(PL_regtill); /* How far we are required to go. */
SAVEI8(PL_regprev); /* char before regbol, \n if none */
- SAVEVPTR(PL_reg_start_tmp); /* from regexec.c */
+ SAVEGENERICPV(PL_reg_start_tmp); /* from regexec.c */
PL_reg_start_tmp = 0;
- SAVEFREEPV(PL_reg_start_tmp);
SAVEI32(PL_reg_start_tmpl); /* from regexec.c */
PL_reg_start_tmpl = 0;
SAVEVPTR(PL_regdata);
SAVEVPTR(PL_reg_curpm); /* from regexec.c */
SAVEI32(PL_regnpar); /* () count. */
#ifdef DEBUGGING
- SAVEPPTR(PL_reg_starttry); /* from regexec.c */
+ SAVEPPTR(PL_reg_starttry); /* from regexec.c */
#endif
}
projects / p5sagit/p5-mst-13.2.git / blobdiff