/* *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.
U16 flags16; /* are we folding, multilining? */
char *precomp; /* uncompiled string. */
regexp *rx;
+ char *start; /* Start of input for compile */
char *end; /* End of input for compile */
char *parse; /* Input-scan pointer. */
I32 whilem_seen; /* number of WHILEM in this expr */
+ regnode *emit_start; /* Start of emitted-code area */
regnode *emit; /* Code-emit pointer; ®dummy = don't = compiling */
I32 naughty; /* How bad is this pattern? */
I32 sawback; /* Did we see \1, ...? */
I32 extralen;
I32 seen_zerolen;
I32 seen_evals;
+ I32 utf8;
#if ADD_TO_REGEXEC
char *starttry; /* -Dr: where regtry was called. */
#define RExC_starttry (pRExC_state->starttry)
#define RExC_flags16 (pRExC_state->flags16)
#define RExC_precomp (pRExC_state->precomp)
#define RExC_rx (pRExC_state->rx)
+#define RExC_start (pRExC_state->start)
#define RExC_end (pRExC_state->end)
#define RExC_parse (pRExC_state->parse)
#define RExC_whilem_seen (pRExC_state->whilem_seen)
+#define RExC_offsets (pRExC_state->rx->offsets) /* I am not like the others */
#define RExC_emit (pRExC_state->emit)
+#define RExC_emit_start (pRExC_state->emit_start)
#define RExC_naughty (pRExC_state->naughty)
#define RExC_sawback (pRExC_state->sawback)
#define RExC_seen (pRExC_state->seen)
#define RExC_extralen (pRExC_state->extralen)
#define RExC_seen_zerolen (pRExC_state->seen_zerolen)
#define RExC_seen_evals (pRExC_state->seen_evals)
+#define RExC_utf8 (pRExC_state->utf8)
#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 SCF_DO_STCLASS (SCF_DO_STCLASS_AND|SCF_DO_STCLASS_OR)
#define SCF_WHILEM_VISITED_POS 0x2000
-#define RF_utf8 8
-#define UTF (PL_reg_flags & RF_utf8)
+#define UTF RExC_utf8
#define LOC (RExC_flags16 & PMf_LOCALE)
#define FOLD (RExC_flags16 & PMf_FOLD)
* of t/op/regmesg.t, the tests in t/op/re_tests, and those in
* op/pragma/warn/regcomp.
*/
-#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/"
+#define MARKER1 "<-- HERE" /* marker as it appears in the description */
+#define MARKER2 " <-- HERE " /* marker as it appears within the regex */
+
+#define REPORT_LOCATION " in regex; marked by " MARKER1 " in m/%.*s" MARKER2 "%s/"
/*
* Calls SAVEDESTRUCTOR_X if needed, then calls Perl_croak with the given
m, (int)offset, RExC_precomp, RExC_precomp + offset); \
} STMT_END \
+#define vWARNdep(loc,m) \
+ STMT_START { \
+ unsigned offset = strlen(RExC_precomp)-(RExC_end-(loc)); \
+ 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); \
+ } STMT_END \
+
#define vWARN2(loc, m, a1) \
STMT_START { \
(int)offset, RExC_precomp, RExC_precomp + offset); \
} STMT_END
+/* 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)); \
+ Perl_warner(aTHX_ WARN_REGEXP, m REPORT_LOCATION, \
+ a1, a2, a3, a4, \
+ (int)offset, RExC_precomp, RExC_precomp + offset); \
+ } STMT_END
+
/* Allow for side effects in s */
-#define REGC(c,s) STMT_START { if (!SIZE_ONLY) *(s) = (c); else (s);} STMT_END
+#define REGC(c,s) STMT_START { if (!SIZE_ONLY) *(s) = (c); else (void)(s);} STMT_END
+
+/* Macros for recording node offsets. 20001227 mjd@plover.com
+ * Nodes are numbered 1, 2, 3, 4. Node #n's position is recorded in
+ * element 2*n-1 of the array. Element #2n holds the byte length node #n.
+ * Element 0 holds the number n.
+ */
+
+#define MJD_OFFSET_DEBUG(x)
+/* #define MJD_OFFSET_DEBUG(x) fprintf x */
+
+
+# define Set_Node_Offset_To_R(node,byte) \
+ STMT_START { \
+ if (! SIZE_ONLY) { \
+ if((node) < 0) { \
+ Perl_croak(aTHX_ "value of node is %d in Offset macro", node); \
+ } else { \
+ RExC_offsets[2*(node)-1] = (byte); \
+ } \
+ } \
+ } STMT_END
+
+# define Set_Node_Offset(node,byte) Set_Node_Offset_To_R((node)-RExC_emit_start, (byte)-RExC_start)
+# define Set_Cur_Node_Offset Set_Node_Offset(RExC_emit, RExC_parse)
+
+# define Set_Node_Length_To_R(node,len) \
+ STMT_START { \
+ if (! SIZE_ONLY) { \
+ MJD_OFFSET_DEBUG((stderr, "** (%d) size of node %d is %d.\n", __LINE__, (node), (len))); \
+ if((node) < 0) { \
+ Perl_croak(aTHX_ "value of node is %d in Length macro", node); \
+ } else { \
+ RExC_offsets[2*(node)] = (len); \
+ } \
+ } \
+ } STMT_END
+
+# define Set_Node_Length(node,len) Set_Node_Length_To_R((node)-RExC_emit_start, len)
+# define Set_Cur_Node_Length(len) Set_Node_Length(RExC_emit, len)
+# define Set_Node_Cur_Length(node) Set_Node_Length(node, RExC_parse - parse_start)
+
+/* Get offsets and lengths */
+#define Node_Offset(n) (RExC_offsets[2*((n)-RExC_emit_start)-1])
+#define Node_Length(n) (RExC_offsets[2*((n)-RExC_emit_start)])
static void clear_re(pTHXo_ void *r);
{
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;
ANYOF_CLASS_ZERO(cl);
for (value = 0; value < 256; ++value)
ANYOF_BITMAP_SET(cl, value);
- cl->flags = ANYOF_EOS;
+ cl->flags = ANYOF_EOS|ANYOF_UNICODE_ALL;
if (LOC)
cl->flags |= ANYOF_LOCALE;
}
for (value = 0; value <= ANYOF_MAX; value += 2)
if (ANYOF_CLASS_TEST(cl, value) && ANYOF_CLASS_TEST(cl, value + 1))
return 1;
+ if (!(cl->flags & ANYOF_UNICODE_ALL))
+ return 0;
for (value = 0; value < 256; ++value)
if (!ANYOF_BITMAP_TEST(cl, value))
return 0;
} /* XXXX: logic is complicated otherwise, leave it along for a moment. */
if (!(and_with->flags & ANYOF_EOS))
cl->flags &= ~ANYOF_EOS;
+
+ if (cl->flags & ANYOF_UNICODE_ALL && and_with->flags & ANYOF_UNICODE) {
+ cl->flags &= ~ANYOF_UNICODE_ALL;
+ cl->flags |= ANYOF_UNICODE;
+ ARG_SET(cl, ARG(and_with));
+ }
+ if (!(and_with->flags & ANYOF_UNICODE_ALL))
+ cl->flags &= ~ANYOF_UNICODE_ALL;
+ if (!(and_with->flags & (ANYOF_UNICODE|ANYOF_UNICODE_ALL)))
+ cl->flags &= ~ANYOF_UNICODE;
}
/* 'OR' a given class with another one. Can create false positives */
} 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;
}
if (or_with->flags & ANYOF_EOS)
cl->flags |= ANYOF_EOS;
+
+ if (cl->flags & ANYOF_UNICODE && or_with->flags & ANYOF_UNICODE &&
+ ARG(cl) != ARG(or_with)) {
+ cl->flags |= ANYOF_UNICODE_ALL;
+ cl->flags &= ~ANYOF_UNICODE;
+ }
+ if (or_with->flags & ANYOF_UNICODE_ALL) {
+ cl->flags |= ANYOF_UNICODE_ALL;
+ cl->flags &= ~ANYOF_UNICODE;
+ }
}
/* REx optimizer. Converts nodes into quickier variants "in place".
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);
}
else if (OP(scan) == EXACT) {
I32 l = STR_LEN(scan);
+ UV uc = *((U8*)STRING(scan));
if (UTF) {
- unsigned char *s = (unsigned char *)STRING(scan);
- unsigned char *e = s + l;
- I32 newl = 0;
- while (s < e) {
- newl++;
- s += UTF8SKIP(s);
- }
- l = newl;
+ U8 *s = (U8*)STRING(scan);
+ l = utf8_length(s, s + l);
+ 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;
/* Check whether it is compatible with what we know already! */
int compat = 1;
- if (!(data->start_class->flags & (ANYOF_CLASS | ANYOF_LOCALE))
- && !ANYOF_BITMAP_TEST(data->start_class, *STRING(scan))
+ if (uc >= 0x100 ||
+ (!(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[*(U8*)STRING(scan)])))
+ || !ANYOF_BITMAP_TEST(data->start_class, PL_fold[uc])))
+ )
compat = 0;
ANYOF_CLASS_ZERO(data->start_class);
ANYOF_BITMAP_ZERO(data->start_class);
if (compat)
- ANYOF_BITMAP_SET(data->start_class, *STRING(scan));
+ 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 */
- ANYOF_BITMAP_SET(data->start_class, *STRING(scan));
+ if (uc < 0x100)
+ 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);
}
}
else if (PL_regkind[(U8)OP(scan)] == EXACT) { /* But OP != EXACT! */
I32 l = STR_LEN(scan);
+ 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) {
- unsigned char *s = (unsigned char *)STRING(scan);
- unsigned char *e = s + l;
- I32 newl = 0;
- while (s < e) {
- newl++;
- s += UTF8SKIP(s);
- }
- l = newl;
+ U8 *s = (U8 *)STRING(scan);
+ l = utf8_length(s, s + l);
+ uc = utf8_to_uvchr(s, NULL);
}
min += l;
if (data && (flags & SCF_DO_SUBSTR))
/* Check whether it is compatible with what we know already! */
int compat = 1;
- if (!(data->start_class->flags & (ANYOF_CLASS | ANYOF_LOCALE))
- && !ANYOF_BITMAP_TEST(data->start_class, *STRING(scan))
- && !ANYOF_BITMAP_TEST(data->start_class,
- PL_fold[*(U8*)STRING(scan)]))
+ if (uc >= 0x100 ||
+ (!(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;
ANYOF_CLASS_ZERO(data->start_class);
ANYOF_BITMAP_ZERO(data->start_class);
if (compat) {
- ANYOF_BITMAP_SET(data->start_class, *STRING(scan));
+ ANYOF_BITMAP_SET(data->start_class, uc);
data->start_class->flags &= ~ANYOF_EOS;
data->start_class->flags |= ANYOF_FOLD;
if (OP(scan) == EXACTFL)
if (data->start_class->flags & ANYOF_FOLD) {
/* false positive possible if the class is case-folded.
Assume that the locale settings are the same... */
- ANYOF_BITMAP_SET(data->start_class, *STRING(scan));
+ if (uc < 0x100)
+ ANYOF_BITMAP_SET(data->start_class, uc);
data->start_class->flags &= ~ANYOF_EOS;
}
cl_and(data->start_class, &and_with);
flags &= ~SCF_DO_STCLASS;
}
else if (strchr((char*)PL_varies,OP(scan))) {
- I32 mincount, maxcount, minnext, deltanext, fl;
+ I32 mincount, maxcount, minnext, deltanext, fl = 0;
I32 f = flags, pos_before = 0;
regnode *oscan = scan;
struct regnode_charclass_class this_class;
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)
- && !(data->flags & (SF_HAS_PAR|SF_IN_PAR))
+ if (ckWARN(WARN_REGEXP)
+ && (minnext == 0) && (deltanext == 0)
+ && data && !(data->flags & (SF_HAS_PAR|SF_IN_PAR))
&& maxcount <= REG_INFTY/3) /* Complement check for big count */
{
vWARN(RExC_parse,
}
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 if (strchr((char*)PL_simple,OP(scan))) {
- int value;
+ int value = 0;
if (flags & SCF_DO_SUBSTR) {
scan_commit(pRExC_state,data);
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;
FAIL("NULL regexp argument");
/* XXXX This looks very suspicious... */
- if (pm->op_pmdynflags & PMdf_UTF8) {
- PL_reg_flags |= RF_utf8;
- }
+ if (pm->op_pmdynflags & PMdf_CMP_UTF8)
+ RExC_utf8 = 1;
else
- PL_reg_flags = 0;
+ RExC_utf8 = 0;
- RExC_precomp = savepvn(exp, xend - exp);
+ 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],
/* First pass: determine size, legality. */
RExC_parse = exp;
+ RExC_start = exp;
RExC_end = xend;
RExC_naughty = 0;
RExC_npar = 1;
REGC((U8)REG_MAGIC, (char*)RExC_emit);
#endif
if (reg(pRExC_state, 0, &flags) == NULL) {
- Safefree(RExC_precomp);
RExC_precomp = Nullch;
return(NULL);
}
#endif
r->refcnt = 1;
r->prelen = xend - exp;
- r->precomp = RExC_precomp;
+ r->precomp = savepvn(RExC_precomp, r->prelen);
r->subbeg = NULL;
r->reganch = pm->op_pmflags & PMf_COMPILETIME;
r->nparens = RExC_npar - 1; /* set early to validate backrefs */
r->startp = 0; /* Useful during FAIL. */
r->endp = 0; /* Useful during FAIL. */
+ Newz(1304, r->offsets, 2*RExC_size+1, U32); /* MJD 20001228 */
+ if (r->offsets) {
+ r->offsets[0] = RExC_size;
+ }
+ DEBUG_r(PerlIO_printf(Perl_debug_log,
+ "%s %"UVuf" bytes for offset annotations.\n",
+ r->offsets ? "Got" : "Couldn't get",
+ (UV)((2*RExC_size+1) * sizeof(U32))));
+
RExC_rx = r;
/* Second pass: emit code. */
RExC_end = xend;
RExC_naughty = 0;
RExC_npar = 1;
+ RExC_emit_start = r->program;
RExC_emit = r->program;
/* Store the count of eval-groups for security checks: */
RExC_emit->next_off = ((RExC_seen_evals > U16_MAX) ? U16_MAX : RExC_seen_evals);
/* Starting-point info. */
again:
if (PL_regkind[(U8)OP(first)] == EXACT) {
- if (OP(first) == EXACT); /* Empty, get anchored substr later. */
- else if ((OP(first) == EXACTF || OP(first) == EXACTFL)
- && !UTF)
+ if (OP(first) == EXACT)
+ ; /* Empty, get anchored substr later. */
+ else if ((OP(first) == EXACTF || OP(first) == EXACTFL))
r->regstclass = first;
}
else if (strchr((char*)PL_simple,OP(first)))
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
&& !(data.start_class->flags & ANYOF_EOS)
&& !cl_is_anything(data.start_class)) {
- 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->regstclass = (regnode*)RExC_rx->data->data[n];
r->reganch &= ~ROPT_SKIP; /* Used in find_byclass(). */
PL_regdata = r->data; /* for regprop() */
- DEBUG_r((sv = sv_newmortal(),
- regprop(sv, (regnode*)data.start_class),
- PerlIO_printf(Perl_debug_log, "synthetic stclass `%s'.\n",
- SvPVX(sv))));
+ DEBUG_r({ SV *sv = sv_newmortal();
+ regprop(sv, (regnode*)data.start_class);
+ PerlIO_printf(Perl_debug_log,
+ "synthetic stclass `%s'.\n",
+ SvPVX(sv));});
}
/* A temporary algorithm prefers floated substr to fixed one to dig more info. */
r->check_substr = r->anchored_substr = r->float_substr = Nullsv;
if (!(data.start_class->flags & ANYOF_EOS)
&& !cl_is_anything(data.start_class)) {
- 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],
struct regnode_charclass_class);
r->regstclass = (regnode*)RExC_rx->data->data[n];
r->reganch &= ~ROPT_SKIP; /* Used in find_byclass(). */
- DEBUG_r((sv = sv_newmortal(),
- regprop(sv, (regnode*)data.start_class),
- PerlIO_printf(Perl_debug_log, "synthetic stclass `%s'.\n",
- SvPVX(sv))));
+ DEBUG_r({ SV* sv = sv_newmortal();
+ regprop(sv, (regnode*)data.start_class);
+ PerlIO_printf(Perl_debug_log,
+ "synthetic stclass `%s'.\n",
+ SvPVX(sv));});
}
}
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() */
register regnode *ender = 0;
register I32 parno = 0;
I32 flags, oregflags = RExC_flags16, have_branch = 0, open = 0;
+
+ /* for (?g), (?gc), and (?o) warnings; warning
+ about (?c) will warn about (?g) -- japhy */
+
+ I32 wastedflags = 0x00,
+ wasted_o = 0x01,
+ wasted_g = 0x02,
+ wasted_gc = 0x02 | 0x04,
+ wasted_c = 0x04;
+
+ char * parse_start = RExC_parse; /* MJD */
char *oregcomp_parse = RExC_parse;
char c;
*flagp = 0; /* Tentatively. */
+
/* Make an OPEN node, if parenthesized. */
if (paren) {
- if (*RExC_parse == '?') {
+ if (*RExC_parse == '?') { /* (?...) */
U16 posflags = 0, negflags = 0;
U16 *flagsp = &posflags;
int logical = 0;
paren = *RExC_parse++;
ret = NULL; /* For look-ahead/behind. */
switch (paren) {
- case '<':
+ 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 '=':
- case '!':
+ case '=': /* (?=...) */
+ case '!': /* (?!...) */
RExC_seen_zerolen++;
- case ':':
- case '>':
+ case ':': /* (?:...) */
+ case '>': /* (?>...) */
break;
- case '$':
- case '@':
+ case '$': /* (?$...) */
+ case '@': /* (?@...) */
vFAIL2("Sequence (?%c...) not implemented", (int)paren);
break;
- case '#':
+ case '#': /* (?#...) */
while (*RExC_parse && *RExC_parse != ')')
RExC_parse++;
if (*RExC_parse != ')')
nextchar(pRExC_state);
*flagp = TRYAGAIN;
return NULL;
- case 'p':
- if (SIZE_ONLY)
- vWARN(RExC_parse, "(?p{}) is deprecated - use (??{})");
+ case 'p': /* (?p...) */
+ if (SIZE_ONLY && ckWARN2(WARN_DEPRECATED, WARN_REGEXP))
+ vWARNdep(RExC_parse, "(?p{}) is deprecated - use (??{})");
/* FALL THROUGH*/
- case '?':
+ case '?': /* (??...) */
logical = 1;
paren = *RExC_parse++;
/* FALL THROUGH */
- case '{':
+ case '{': /* (?{...}) */
{
I32 count = 1, n = 0;
char c;
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);
/* No compiled RE interpolated, has runtime
components ===> unsafe. */
FAIL("Eval-group not allowed at runtime, use re 'eval'");
- if (PL_tainted)
+ if (PL_tainting && PL_tainted)
FAIL("Eval-group in insecure regular expression");
}
if (!SIZE_ONLY)
ret->flags = 2;
regtail(pRExC_state, ret, reganode(pRExC_state, EVAL, n));
+ /* deal with the length of this later - MJD */
return ret;
}
return reganode(pRExC_state, EVAL, n);
}
- case '(':
+ case '(': /* (?(?{...})...) and (?(?=...)...) */
{
- if (RExC_parse[0] == '?') {
- if (RExC_parse[1] == '=' || RExC_parse[1] == '!'
- || RExC_parse[1] == '<'
+ if (RExC_parse[0] == '?') { /* (?(?...)) */
+ 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' ) {
+ /* (?(1)...) */
parno = atoi(RExC_parse++);
while (isDIGIT(*RExC_parse))
RExC_parse++;
- ret = reganode(pRExC_state, GROUPP, parno);
+ ret = reganode(pRExC_state, GROUPP, parno);
+
if ((c = *nextchar(pRExC_state)) != ')')
vFAIL("Switch condition not recognized");
insert_if:
break;
default:
--RExC_parse;
- parse_flags:
+ parse_flags: /* (?i) */
while (*RExC_parse && strchr("iogcmsx", *RExC_parse)) {
- if (*RExC_parse != 'o')
- pmflag(flagsp, *RExC_parse);
+ /* (?g), (?gc) and (?o) are useless here
+ and must be globally applied -- japhy */
+
+ if (*RExC_parse == 'o' || *RExC_parse == 'g') {
+ if (SIZE_ONLY && ckWARN(WARN_REGEXP)) {
+ I32 wflagbit = *RExC_parse == 'o' ? wasted_o : wasted_g;
+ if (! (wastedflags & wflagbit) ) {
+ wastedflags |= wflagbit;
+ vWARN5(
+ RExC_parse + 1,
+ "Useless (%s%c) - %suse /%c modifier",
+ flagsp == &negflags ? "?-" : "?",
+ *RExC_parse,
+ flagsp == &negflags ? "don't " : "",
+ *RExC_parse
+ );
+ }
+ }
+ }
+ else if (*RExC_parse == 'c') {
+ if (SIZE_ONLY && ckWARN(WARN_REGEXP)) {
+ if (! (wastedflags & wasted_c) ) {
+ wastedflags |= wasted_gc;
+ vWARN3(
+ RExC_parse + 1,
+ "Useless (%sc) - %suse /gc modifier",
+ flagsp == &negflags ? "?-" : "?",
+ flagsp == &negflags ? "don't " : ""
+ );
+ }
+ }
+ }
+ else { pmflag(flagsp, *RExC_parse); }
+
++RExC_parse;
}
if (*RExC_parse == '-') {
flagsp = &negflags;
+ wastedflags = 0; /* reset so (?g-c) warns twice */
++RExC_parse;
goto parse_flags;
}
return NULL;
}
}
- else {
+ else { /* (...) */
parno = RExC_npar;
RExC_npar++;
ret = reganode(pRExC_state, OPEN, parno);
+ Set_Node_Length(ret, 1); /* MJD */
+ Set_Node_Offset(ret, RExC_parse); /* MJD */
open = 1;
}
}
- else
+ else /* ! paren */
ret = NULL;
/* Pick up the branches, linking them together. */
+ parse_start = RExC_parse; /* MJD */
br = regbranch(pRExC_state, &flags, 1);
+ /* branch_len = (paren != 0); */
+
if (br == NULL)
return(NULL);
if (*RExC_parse == '|') {
if (!SIZE_ONLY && RExC_extralen) {
reginsert(pRExC_state, BRANCHJ, br);
}
- else
+ else { /* MJD */
reginsert(pRExC_state, BRANCH, br);
+ Set_Node_Length(br, paren != 0);
+ Set_Node_Offset_To_R(br-RExC_emit_start, parse_start-RExC_start);
+ }
have_branch = 1;
if (SIZE_ONLY)
RExC_extralen += 1; /* For BRANCHJ-BRANCH. */
RExC_extralen += 2; /* Account for LONGJMP. */
nextchar(pRExC_state);
br = regbranch(pRExC_state, &flags, 0);
+
if (br == NULL)
return(NULL);
regtail(pRExC_state, lastbr, br); /* BRANCH -> BRANCH. */
break;
case 1:
ender = reganode(pRExC_state, CLOSE, parno);
+ Set_Node_Offset(ender,RExC_parse+1); /* MJD */
+ Set_Node_Length(ender,1); /* MJD */
break;
case '<':
case ',':
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
+ else {
ret = reg_node(pRExC_state, BRANCH);
+ Set_Node_Length(ret, 1);
+ }
}
- if (!first && SIZE_ONLY)
+ if (!first && SIZE_ONLY)
RExC_extralen += 1; /* BRANCHJ */
-
+
*flagp = WORST; /* Tentatively. */
RExC_parse--;
char *maxpos;
I32 min;
I32 max = REG_INFTY;
+ char *parse_start;
ret = regatom(pRExC_state, &flags);
if (ret == NULL) {
op = *RExC_parse;
if (op == '{' && regcurly(RExC_parse)) {
+ parse_start = RExC_parse; /* MJD */
next = RExC_parse + 1;
maxpos = Nullch;
while (isDIGIT(*next) || *next == ',') {
if ((flags&SIMPLE)) {
RExC_naughty += 2 + RExC_naughty / 2;
reginsert(pRExC_state, CURLY, ret);
+ Set_Node_Offset(ret, parse_start+1); /* MJD */
+ Set_Node_Cur_Length(ret);
}
else {
regnode *w = reg_node(pRExC_state, WHILEM);
NEXT_OFF(ret) = 3; /* Go over LONGJMP. */
}
reginsert(pRExC_state, CURLYX,ret);
+ /* MJD hk */
+ Set_Node_Offset(ret, parse_start+1);
+ Set_Node_Length(ret,
+ op == '{' ? (RExC_parse - parse_start) : 1);
+
if (!SIZE_ONLY && RExC_extralen)
NEXT_OFF(ret) = 3; /* Go over NOTHING to LONGJMP. */
regtail(pRExC_state, ret, reg_node(pRExC_state, NOTHING));
if (!(flags&HASWIDTH) && op != '?')
vFAIL("Regexp *+ operand could be empty");
-#endif
+#endif
+ parse_start = RExC_parse;
nextchar(pRExC_state);
*flagp = (op != '+') ? (WORST|SPSTART|HASWIDTH) : (WORST|HASWIDTH);
ret = reg_node(pRExC_state, SBOL);
else
ret = reg_node(pRExC_state, BOL);
+ Set_Node_Length(ret, 1); /* MJD */
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);
ret = reg_node(pRExC_state, SEOL);
else
ret = reg_node(pRExC_state, EOL);
+ Set_Node_Length(ret, 1); /* MJD */
break;
case '.':
nextchar(pRExC_state);
ret = reg_node(pRExC_state, REG_ANY);
*flagp |= HASWIDTH|SIMPLE;
RExC_naughty++;
+ Set_Node_Length(ret, 1); /* MJD */
break;
case '[':
{
}
nextchar(pRExC_state);
*flagp |= HASWIDTH|SIMPLE;
+ Set_Node_Length(ret, RExC_parse - oregcomp_parse + 1); /* MJD */
break;
}
case '(':
ret = reg_node(pRExC_state, SBOL);
*flagp |= SIMPLE;
nextchar(pRExC_state);
+ Set_Node_Length(ret, 2); /* MJD */
break;
case 'G':
ret = reg_node(pRExC_state, GPOS);
RExC_seen |= REG_SEEN_GPOS;
*flagp |= SIMPLE;
nextchar(pRExC_state);
+ Set_Node_Length(ret, 2); /* MJD */
break;
case 'Z':
ret = reg_node(pRExC_state, SEOL);
*flagp |= SIMPLE;
RExC_seen_zerolen++; /* Do not optimize RE away */
nextchar(pRExC_state);
+ Set_Node_Length(ret, 2); /* MJD */
break;
case 'C':
ret = reg_node(pRExC_state, SANY);
+ RExC_seen |= REG_SEEN_SANY;
*flagp |= HASWIDTH|SIMPLE;
nextchar(pRExC_state);
+ Set_Node_Length(ret, 2); /* MJD */
break;
case 'X':
ret = reg_node(pRExC_state, CLUMP);
*flagp |= HASWIDTH;
nextchar(pRExC_state);
- if (UTF && !PL_utf8_mark)
- is_utf8_mark((U8*)"~"); /* preload table */
+ Set_Node_Length(ret, 2); /* MJD */
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 */
+ Set_Node_Length(ret, 2); /* MJD */
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 */
+ Set_Node_Length(ret, 2); /* MJD */
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 */
+ Set_Node_Length(ret, 2); /* MJD */
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 */
+ Set_Node_Length(ret, 2); /* MJD */
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 */
+ Set_Node_Length(ret, 2); /* MJD */
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 */
+ Set_Node_Length(ret, 2); /* MJD */
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 */
+ Set_Node_Length(ret, 2); /* MJD */
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 */
+ Set_Node_Length(ret, 2); /* MJD */
break;
case 'p':
case 'P':
- { /* a lovely hack--pretend we saw [\pX] instead */
+ {
char* oldregxend = RExC_end;
+ char* parse_start = RExC_parse;
if (RExC_parse[1] == '{') {
+ /* a lovely hack--pretend we saw [\pX] instead */
RExC_end = strchr(RExC_parse, '}');
if (!RExC_end) {
RExC_parse += 2;
RExC_end = oldregxend;
RExC_parse--;
+ Set_Node_Cur_Length(ret); /* MJD */
nextchar(pRExC_state);
*flagp |= HASWIDTH|SIMPLE;
}
if (num > 9 && num >= RExC_npar)
goto defchar;
else {
+ char * parse_start = RExC_parse - 1; /* MJD */
while (isDIGIT(*RExC_parse))
RExC_parse++;
? (LOC ? REFFL : REFF)
: REF, num);
*flagp |= HASWIDTH;
+
+ /* override incorrect value set in reganode MJD */
+ Set_Node_Offset(ret, parse_start+1);
+ Set_Node_Cur_Length(ret); /* MJD */
RExC_parse--;
nextchar(pRExC_state);
}
register char *p;
char *oldp, *s;
STRLEN numlen;
+ char *parse_start = RExC_parse - 1;
RExC_parse++;
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) {
+ else if (!UNI_IS_INVARIANT(NATIVE_TO_UNI(ender)) && UTF) {
reguni(pRExC_state, ender, s, &numlen);
s += numlen;
len += numlen;
}
break;
}
- if (ender >= 0x80 && UTF) {
+ if (!UNI_IS_INVARIANT(NATIVE_TO_UNI(ender)) && UTF) {
reguni(pRExC_state, ender, s, &numlen);
s += numlen;
len += numlen - 1;
}
loopdone:
RExC_parse = p - 1;
+ Set_Node_Cur_Length(ret); /* MJD */
nextchar(pRExC_state);
{
/* len is STRLEN which is unsigned, need to copy to signed */
*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)
S_regclass(pTHX_ RExC_state_t *pRExC_state)
{
register UV value;
- register IV lastvalue = OOB_UNICODE;
+ register IV prevvalue = OOB_UNICODE;
register IV range = 0;
register regnode *ret;
STRLEN numlen;
IV namedclass;
- char *rangebegin;
+ char *rangebegin = 0;
bool need_class = 0;
- SV *listsv;
+ SV *listsv = Nullsv;
register char *e;
UV n;
+ bool optimize_invert = TRUE;
ret = reganode(pRExC_state, ANYOF, 0);
if (!range)
rangebegin = RExC_parse;
if (UTF) {
- value = utf8_to_uv((U8*)RExC_parse,
- RExC_end - RExC_parse,
- &numlen, 0);
+ value = utf8n_to_uvchr((U8*)RExC_parse,
+ RExC_end - RExC_parse,
+ &numlen, 0);
RExC_parse += numlen;
}
else
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,
RExC_parse - rangebegin,
RExC_parse - rangebegin,
rangebegin);
- if (lastvalue < 256) {
- ANYOF_BITMAP_SET(ret, lastvalue);
+ if (prevvalue < 256) {
+ ANYOF_BITMAP_SET(ret, prevvalue);
ANYOF_BITMAP_SET(ret, '-');
}
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)prevvalue, (UV) '-');
}
}
}
if (!SIZE_ONLY) {
+ if (namedclass > OOB_NAMEDCLASS)
+ optimize_invert = FALSE;
/* Possible truncation here but in some 64-bit environments
* the compiler gets heartburn about switch on 64-bit values.
* A similar issue a little earlier when switching on value.
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 */
- for (value = 0; value < 256; value++)
+ for (value = 0; value < 256; value++) {
if (isASCII(value))
- ANYOF_BITMAP_SET(ret, value);
+ ANYOF_BITMAP_SET(ret, value);
+ }
#endif /* EBCDIC */
}
Perl_sv_catpvf(aTHX_ listsv, "+utf8::IsASCII\n");
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 */
- for (value = 0; value < 256; value++)
+ for (value = 0; value < 256; value++) {
if (!isASCII(value))
- ANYOF_BITMAP_SET(ret, value);
+ ANYOF_BITMAP_SET(ret, value);
+ }
#endif /* EBCDIC */
}
Perl_sv_catpvf(aTHX_ listsv, "!utf8::IsASCII\n");
} /* end of namedclass \blah */
if (range) {
- if (lastvalue > value) /* b-a */ {
+ if (prevvalue > value) /* b-a */ {
Simple_vFAIL4("Invalid [] range \"%*.*s\"",
RExC_parse - rangebegin,
RExC_parse - rangebegin,
rangebegin);
+ range = 0; /* not a valid range */
}
- range = 0; /* not a true range */
}
else {
- lastvalue = value; /* save the beginning of the range */
+ prevvalue = value; /* save the beginning of the range */
if (*RExC_parse == '-' && RExC_parse+1 < RExC_end &&
RExC_parse[1] != ']') {
RExC_parse++;
/* now is the next time */
if (!SIZE_ONLY) {
- if (lastvalue < 256 && value < 256) {
-#ifndef ASCIIish /* EBCDIC, for example. */
- if ((isLOWER(lastvalue) && isLOWER(value)) ||
- (isUPPER(lastvalue) && isUPPER(value)))
+ IV i;
+
+ if (prevvalue < 256) {
+ IV ceilvalue = value < 256 ? value : 255;
+
+#ifdef EBCDIC
+ if ((isLOWER(prevvalue) && isLOWER(ceilvalue)) ||
+ (isUPPER(prevvalue) && isUPPER(ceilvalue)))
{
- IV i;
- if (isLOWER(lastvalue)) {
- for (i = lastvalue; i <= value; i++)
+ if (isLOWER(prevvalue)) {
+ for (i = prevvalue; i <= ceilvalue; i++)
if (isLOWER(i))
ANYOF_BITMAP_SET(ret, i);
} else {
- for (i = lastvalue; i <= value; i++)
+ for (i = prevvalue; i <= ceilvalue; i++)
if (isUPPER(i))
ANYOF_BITMAP_SET(ret, i);
}
}
else
#endif
- for ( ; lastvalue <= value; lastvalue++)
- ANYOF_BITMAP_SET(ret, lastvalue);
- } else {
+ for (i = prevvalue; i <= ceilvalue; i++)
+ ANYOF_BITMAP_SET(ret, i);
+ }
+ if (value > 255) {
ANYOF_FLAGS(ret) |= ANYOF_UNICODE;
- if (lastvalue < value)
+ if (prevvalue < value)
Perl_sv_catpvf(aTHX_ listsv, "%04"UVxf"\t%04"UVxf"\n",
- (UV)lastvalue, (UV)value);
- else
+ (UV)prevvalue, (UV)value);
+ else if (prevvalue == value)
Perl_sv_catpvf(aTHX_ listsv, "%04"UVxf"\n",
(UV)value);
}
}
if (need_class) {
+ ANYOF_FLAGS(ret) |= ANYOF_LARGE;
if (SIZE_ONLY)
RExC_size += ANYOF_CLASS_ADD_SKIP;
else
/* optimize case-insensitive simple patterns (e.g. /[a-z]/i) */
if (!SIZE_ONLY &&
- (ANYOF_FLAGS(ret) &
/* If the only flag is folding (plus possibly inversion). */
- (ANYOF_FLAGS_ALL ^ ANYOF_INVERT) == ANYOF_FOLD)) {
+ ((ANYOF_FLAGS(ret) & (ANYOF_FLAGS_ALL ^ ANYOF_INVERT)) == ANYOF_FOLD)
+ ) {
for (value = 0; value < 256; ++value) {
if (ANYOF_BITMAP_TEST(ret, value)) {
IV fold = PL_fold[value];
}
/* optimize inverted simple patterns (e.g. [^a-z]) */
- if (!SIZE_ONLY &&
+ if (!SIZE_ONLY && optimize_invert &&
/* If the only flag is inversion. */
(ANYOF_FLAGS(ret) & ANYOF_FLAGS_ALL) == ANYOF_INVERT) {
for (value = 0; value < ANYOF_BITMAP_SIZE; ++value)
ANYOF_BITMAP(ret)[value] ^= ANYOF_FLAGS_ALL;
- ANYOF_FLAGS(ret) = 0;
+ ANYOF_FLAGS(ret) = ANYOF_UNICODE_ALL;
}
- if (!SIZE_ONLY) {
+ if (!SIZE_ONLY) {
AV *av = newAV();
SV *rv;
NODE_ALIGN_FILL(ret);
ptr = ret;
FILL_ADVANCE_NODE(ptr, op);
+ if (RExC_offsets) { /* MJD */
+ MJD_OFFSET_DEBUG((stderr, "%s:%u: (op %s) %s %u <- %u (len %u) (max %u).\n",
+ "reg_node", __LINE__,
+ reg_name[op],
+ RExC_emit - RExC_emit_start > RExC_offsets[0]
+ ? "Overwriting end of array!\n" : "OK",
+ RExC_emit - RExC_emit_start,
+ RExC_parse - RExC_start,
+ RExC_offsets[0]));
+ Set_Node_Offset(RExC_emit, RExC_parse + (op == END));
+ }
+
RExC_emit = ptr;
return(ret);
NODE_ALIGN_FILL(ret);
ptr = ret;
FILL_ADVANCE_NODE_ARG(ptr, op, arg);
+ if (RExC_offsets) { /* MJD */
+ MJD_OFFSET_DEBUG((stderr, "%s: %s %u <- %u (max %u).\n",
+ "reganode",
+ RExC_emit - RExC_emit_start > RExC_offsets[0] ?
+ "Overwriting end of array!\n" : "OK",
+ RExC_emit - RExC_emit_start,
+ RExC_parse - RExC_start,
+ RExC_offsets[0]));
+ Set_Cur_Node_Offset;
+ }
+
RExC_emit = ptr;
return(ret);
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) {
src = RExC_emit;
RExC_emit += NODE_STEP_REGNODE + offset;
dst = RExC_emit;
- while (src > opnd)
+ while (src > opnd) {
StructCopy(--src, --dst, regnode);
+ if (RExC_offsets) { /* MJD 20010112 */
+ MJD_OFFSET_DEBUG((stderr, "%s: %s copy %u -> %u (max %u).\n",
+ "reg_insert",
+ dst - RExC_emit_start > RExC_offsets[0]
+ ? "Overwriting end of array!\n" : "OK",
+ src - RExC_emit_start,
+ dst - RExC_emit_start,
+ RExC_offsets[0]));
+ Set_Node_Offset_To_R(dst-RExC_emit_start, Node_Offset(src));
+ Set_Node_Length_To_R(dst-RExC_emit_start, Node_Length(src));
+ }
+ }
+
place = opnd; /* Op node, where operand used to be. */
+ if (RExC_offsets) { /* MJD */
+ MJD_OFFSET_DEBUG((stderr, "%s: %s %u <- %u (max %u).\n",
+ "reginsert",
+ place - RExC_emit_start > RExC_offsets[0]
+ ? "Overwriting end of array!\n" : "OK",
+ place - RExC_emit_start,
+ RExC_parse - RExC_start,
+ RExC_offsets[0]));
+ Set_Node_Offset(place, RExC_parse);
+ }
src = NEXTOPER(place);
FILL_ADVANCE_NODE(place, op);
Zero(src, offset, regnode);
}
+#ifdef DEBUGGING
+
STATIC regnode *
S_dumpuntil(pTHX_ regnode *start, regnode *node, regnode *last, SV* sv, I32 l)
{
-#ifdef DEBUGGING
register U8 op = EXACT; /* Arbitrary non-END op. */
register regnode *next;
(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;
node = dumpuntil(start, NEXTOPER(node), NEXTOPER(node) + 1, sv, l + 1);
}
else if (op == ANYOF) {
+ /* arglen 1 + class block */
+ node += 1 + ((ANYOF_FLAGS(node) & ANYOF_LARGE)
+ ? ANYOF_CLASS_SKIP : ANYOF_SKIP);
node = NEXTOPER(node);
- node += ANYOF_SKIP;
}
else if (PL_regkind[(U8)op] == EXACT) {
/* Literal string, where present. */
else if (op == WHILEM)
l--;
}
-#endif /* DEBUGGING */
return node;
}
+#endif /* DEBUGGING */
+
/*
- regdump - dump a regexp onto Perl_debug_log in vaguely comprehensible form
*/
/* 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");
if (r->reganch & ROPT_EVAL_SEEN)
PerlIO_printf(Perl_debug_log, "with eval ");
PerlIO_printf(Perl_debug_log, "\n");
+ if (r->offsets) {
+ U32 i;
+ U32 len = r->offsets[0];
+ PerlIO_printf(Perl_debug_log, "Offsets: [%"UVuf"]\n\t", (UV)r->offsets[0]);
+ for (i = 1; i <= len; i++)
+ PerlIO_printf(Perl_debug_log, "%"UVuf"[%"UVuf"] ",
+ (UV)r->offsets[i*2-1],
+ (UV)r->offsets[i*2]);
+ PerlIO_printf(Perl_debug_log, "\n");
+ }
#endif /* DEBUGGING */
}
+#ifdef DEBUGGING
+
STATIC void
S_put_byte(pTHX_ SV *sv, int c)
{
Perl_sv_catpvf(aTHX_ sv, "%c", c);
}
+#endif /* DEBUGGING */
+
/*
- regprop - printable representation of opcode
*/
if (flags & ANYOF_UNICODE)
sv_catpv(sv, "{unicode}");
+ else if (flags & ANYOF_UNICODE_ALL)
+ sv_catpv(sv, "{all-unicode}");
{
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 = uvchr_to_utf8(s, i);
- if (i < 256 && swash_fetch(sw, s)) {
+ if (i < 256 && swash_fetch(sw, s, TRUE)) {
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 = uvchr_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 = uvchr_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 = uvchr_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);
}
}
if (r->precomp)
Safefree(r->precomp);
+ if (r->offsets) /* 20010421 MJD */
+ Safefree(r->offsets);
if (RX_MATCH_COPIED(r))
Safefree(r->subbeg);
if (r->substrs) {
void
Perl_save_re_context(pTHX)
-{
+{
#if 0
SAVEPPTR(RExC_precomp); /* uncompiled string. */
SAVEI32(RExC_npar); /* () count. */
SAVEVPTR(PL_regendp); /* Ditto for endp. */
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_oldcurpm); /* from regexec.c */
SAVEVPTR(PL_reg_curpm); /* from regexec.c */
SAVEI32(PL_regnpar); /* () count. */
+ SAVEI32(PL_regsize); /* from regexec.c */
#ifdef DEBUGGING
- SAVEPPTR(PL_reg_starttry); /* from regexec.c */
+ SAVEPPTR(PL_reg_starttry); /* from regexec.c */
#endif
}
{
ReREFCNT_dec((regexp *)r);
}
+
https://perl5.git.perl.org / perl5.git / blobdiff