/* *These* symbols are masked to allow static link. */
# define Perl_pregfree my_regfree
# define Perl_regnext my_regnext
-# define save_re_context my_save_re_context
+# define Perl_save_re_context my_save_re_context
+# define Perl_reginitcolors my_reginitcolors
#endif
/*SUPPRESS 112*/
while (scan && OP(scan) != END && scan < last) {
/* Peephole optimizer: */
- if (regkind[(U8)OP(scan)] == EXACT) {
+ if (PL_regkind[(U8)OP(scan)] == EXACT) {
regnode *n = regnext(scan);
U32 stringok = 1;
#ifdef DEBUGGING
next = scan + (*OPERAND(scan) + 2 - 1)/sizeof(regnode) + 2;
/* Skip NOTHING, merge EXACT*. */
while (n &&
- ( 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) {
if (OP(n) == TAIL || n > next)
stringok = 0;
- if (regkind[(U8)OP(n)] == NOTHING) {
+ if (PL_regkind[(U8)OP(n)] == NOTHING) {
NEXT_OFF(scan) += NEXT_OFF(n);
next = n + NODE_STEP_REGNODE;
#ifdef DEBUGGING
* don't initialize the OP() slot of a node when that node
* is occupied by just the trailing null of the string in
* an EXACT node */
- if (regkind[(U8)OP(n)] != NOTHING || OP(n) == NOTHING) {
+ if (PL_regkind[(U8)OP(n)] != NOTHING || OP(n) == NOTHING) {
OP(n) = OPTIMIZED;
NEXT_OFF(n) = 0;
}
/* Skip NOTHING and LONGJMP. */
while ((n = regnext(n))
- && ((regkind[(U8)OP(n)] == NOTHING && (noff = NEXT_OFF(n)))
+ && ((PL_regkind[(U8)OP(n)] == NOTHING && (noff = NEXT_OFF(n)))
|| ((OP(n) == LONGJMP) && (noff = ARG(n))))
&& off + noff < max)
off += noff;
data->flags &= ~SF_BEFORE_EOL;
}
}
- else if (regkind[(U8)OP(scan)] == EXACT) {
+ else if (PL_regkind[(U8)OP(scan)] == EXACT) {
I32 l = *OPERAND(scan);
if (flags & SCF_DO_SUBSTR)
scan_commit(data);
if (data && (flags & SCF_DO_SUBSTR))
data->pos_min += l;
}
- else if (strchr(varies,OP(scan))) {
+ else if (strchr(PL_varies,OP(scan))) {
I32 mincount, maxcount, minnext, deltanext, pos_before, fl;
regnode *oscan = scan;
- switch (regkind[(U8)OP(scan)]) {
+ switch (PL_regkind[(U8)OP(scan)]) {
case WHILEM:
scan = NEXTOPER(scan);
goto finish;
scan = next;
if (ckWARN(WARN_UNSAFE) && (minnext + deltanext == 0)
&& !(data->flags & (SF_HAS_PAR|SF_IN_PAR))
- && maxcount <= 10000) /* Complement check for big count */
+ && maxcount <= REG_INFTY/3) /* Complement check for big count */
warner(WARN_UNSAFE, "Strange *+?{} on zero-length expression");
min += minnext * mincount;
is_inf_internal |= (maxcount == REG_INFTY
/* Skip open. */
nxt = regnext(nxt);
- if (!strchr(simple,OP(nxt))
- && !(regkind[(U8)OP(nxt)] == EXACT
+ if (!strchr(PL_simple,OP(nxt))
+ && !(PL_regkind[(U8)OP(nxt)] == EXACT
&& *OPERAND(nxt) == 1))
goto nogo;
nxt2 = nxt;
data->flags |= SF_HAS_EVAL;
optimize_curly_tail:
if (OP(oscan) != CURLYX) {
- while (regkind[(U8)OP(next = regnext(oscan))] == NOTHING
+ while (PL_regkind[(U8)OP(next = regnext(oscan))] == NOTHING
&& NEXT_OFF(next))
NEXT_OFF(oscan) += NEXT_OFF(next);
}
break;
}
}
- else if (strchr(simple,OP(scan)) || regkind[(U8)OP(scan)] == ANYUTF8) {
+ else if (strchr(PL_simple,OP(scan)) || PL_regkind[(U8)OP(scan)] == ANYUTF8) {
if (flags & SCF_DO_SUBSTR) {
scan_commit(data);
data->pos_min++;
}
min++;
}
- else if (regkind[(U8)OP(scan)] == EOL && flags & SCF_DO_SUBSTR) {
+ else if (PL_regkind[(U8)OP(scan)] == EOL && flags & SCF_DO_SUBSTR) {
data->flags |= (OP(scan) == MEOL
? SF_BEFORE_MEOL
: SF_BEFORE_SEOL);
}
- else if (regkind[(U8)OP(scan)] == BRANCHJ
+ else if (PL_regkind[(U8)OP(scan)] == BRANCHJ
&& (scan->flags || data)
&& (OP(scan) == IFMATCH || OP(scan) == UNLESSM)) {
I32 deltanext, minnext;
return PL_regcomp_rx->data->count - n;
}
+void
+reginitcolors(void)
+{
+ dTHR;
+ int i = 0;
+ char *s = PerlEnv_getenv("PERL_RE_COLORS");
+
+ if (s) {
+ PL_colors[0] = s = savepv(s);
+ while (++i < 6) {
+ s = strchr(s, '\t');
+ if (s) {
+ *s = '\0';
+ PL_colors[i] = ++s;
+ }
+ else
+ PL_colors[i] = "";
+ }
+ } else {
+ while (i < 6)
+ PL_colors[i++] = "";
+ }
+ PL_colorset = 1;
+}
+
/*
- pregcomp - compile a regular expression into internal code
*
if (exp == NULL)
FAIL("NULL regexp argument");
- if (PL_curcop == &compiling ? (PL_hints & HINT_UTF8) : IN_UTF8)
+ if (PL_curcop == &PL_compiling ? (PL_hints & HINT_UTF8) : IN_UTF8)
PL_reg_flags |= RF_utf8;
else
PL_reg_flags = 0;
PL_regprecomp = savepvn(exp, xend - exp);
- DEBUG_r(
- if (!PL_colorset) {
- int i = 0;
- char *s = PerlEnv_getenv("PERL_RE_COLORS");
-
- if (s) {
- PL_colors[0] = s = savepv(s);
- while (++i < 6) {
- s = strchr(s, '\t');
- if (s) {
- *s = '\0';
- PL_colors[i] = ++s;
- }
- else
- PL_colors[i] = "";
- }
- } else {
- while (i < 6)
- PL_colors[i++] = "";
- }
- PL_colorset = 1;
- }
- );
+ DEBUG_r(if (!PL_colorset) reginitcolors());
DEBUG_r(PerlIO_printf(Perl_debug_log, "%sCompiling%s RE `%s%*s%s'\n",
- PL_colors[4],PL_colors[5],PL_colors[0],
- xend - exp, PL_regprecomp, PL_colors[1]));
+ PL_colors[4],PL_colors[5],PL_colors[0],
+ xend - exp, PL_regprecomp, PL_colors[1]));
PL_regflags = pm->op_pmflags;
PL_regsawback = 0;
PL_regnpar = 1;
PL_regsize = 0L;
PL_regcode = &PL_regdummy;
- regc((U8)MAGIC, (char*)PL_regcode);
+ regc((U8)REG_MAGIC, (char*)PL_regcode);
if (reg(0, &flags) == NULL) {
Safefree(PL_regprecomp);
PL_regprecomp = Nullch;
PL_regcode = r->program;
/* Store the count of eval-groups for security checks: */
PL_regcode->next_off = ((PL_seen_evals > U16_MAX) ? U16_MAX : PL_seen_evals);
- regc((U8)MAGIC, (char*) PL_regcode++);
+ regc((U8)REG_MAGIC, (char*) PL_regcode++);
r->data = 0;
if (reg(0, &flags) == NULL)
return(NULL);
(OP(first) == BRANCH && OP(regnext(first)) != BRANCH) ||
(OP(first) == PLUS) ||
(OP(first) == MINMOD) ||
- (regkind[(U8)OP(first)] == CURLY && ARG1(first) > 0) ) {
+ (PL_regkind[(U8)OP(first)] == CURLY && ARG1(first) > 0) ) {
if (OP(first) == PLUS)
sawplus = 1;
else
/* Starting-point info. */
again:
if (OP(first) == EXACT); /* Empty, get anchored substr later. */
- else if (strchr(simple+4,OP(first)))
+ else if (strchr(PL_simple+4,OP(first)))
r->regstclass = first;
- else if (regkind[(U8)OP(first)] == BOUND ||
- regkind[(U8)OP(first)] == NBOUND)
+ else if (PL_regkind[(U8)OP(first)] == BOUND ||
+ PL_regkind[(U8)OP(first)] == NBOUND)
r->regstclass = first;
- else if (regkind[(U8)OP(first)] == BOL) {
+ else if (PL_regkind[(U8)OP(first)] == BOL) {
r->reganch |= (OP(first) == MBOL ? ROPT_ANCH_MBOL: ROPT_ANCH_BOL);
first = NEXTOPER(first);
goto again;
goto again;
}
else if ((OP(first) == STAR &&
- regkind[(U8)OP(NEXTOPER(first))] == ANY) &&
+ PL_regkind[(U8)OP(NEXTOPER(first))] == REG_ANY) &&
!(r->reganch & ROPT_ANCH) )
{
/* turn .* into ^.* with an implied $*=1 */
PL_regcomp_rx->data->data[n+2] = (void*)sop;
SvREFCNT_dec(sv);
}
- else { /* First pass */
- if (PL_reginterp_cnt < ++PL_seen_evals && PL_curcop != &compiling)
+ else { /* First pass */
+ if (PL_reginterp_cnt < ++PL_seen_evals
+ && PL_curcop != &PL_compiling)
/* No compiled RE interpolated, has runtime
components ===> unsafe. */
FAIL("Eval-group not allowed at runtime, use re 'eval'");
else
regtail(br, reganode(LONGJMP, 0));
c = *nextchar();
+ if (flags&HASWIDTH)
+ *flagp |= HASWIDTH;
if (c == '|') {
lastbr = reganode(IFTHEN, 0); /* Fake one for optimizer. */
regbranch(&flags, 1);
regtail(ret, lastbr);
+ if (flags&HASWIDTH)
+ *flagp |= HASWIDTH;
c = *nextchar();
}
else
goto do_curly;
}
nest_check:
- if (ckWARN(WARN_UNSAFE) && !SIZE_ONLY && !(flags&HASWIDTH) && max > 10000) {
+ if (ckWARN(WARN_UNSAFE) && !SIZE_ONLY && !(flags&HASWIDTH) && max > REG_INFTY/3) {
warner(WARN_UNSAFE, "%.*s matches null string many times",
PL_regcomp_parse - origparse, origparse);
}
if (PL_regflags & PMf_SINGLELINE)
ret = reg_node(SANY);
else
- ret = reg_node(ANY);
+ ret = reg_node(REG_ANY);
*flagp |= HASWIDTH|SIMPLE;
}
PL_regnaughty++;
return p;
}
+/* parse POSIX character classes like [[:foo:]] */
+STATIC char*
+regpposixcc(I32 value)
+{
+ char *posixcc = 0;
+
+ if (value == '[' && PL_regcomp_parse + 1 < PL_regxend &&
+ /* I smell either [: or [= or [. -- POSIX has been here, right? */
+ (*PL_regcomp_parse == ':' ||
+ *PL_regcomp_parse == '=' ||
+ *PL_regcomp_parse == '.')) {
+ char c = *PL_regcomp_parse;
+ char* s = PL_regcomp_parse++;
+
+ while (PL_regcomp_parse < PL_regxend && *PL_regcomp_parse != c)
+ PL_regcomp_parse++;
+ if (PL_regcomp_parse == PL_regxend)
+ /* Grandfather lone [:, [=, [. */
+ PL_regcomp_parse = s;
+ else {
+ PL_regcomp_parse++; /* skip over the c */
+ if (*PL_regcomp_parse == ']') {
+ /* Not Implemented Yet.
+ * (POSIX Extended Character Classes, that is)
+ * The text between e.g. [: and :] would start
+ * at s + 1 and stop at regcomp_parse - 2. */
+ if (ckWARN(WARN_UNSAFE) && !SIZE_ONLY)
+ warner(WARN_UNSAFE,
+ "Character class syntax [%c %c] is reserved for future extensions", c, c);
+ PL_regcomp_parse++; /* skip over the ending ] */
+ posixcc = s + 1;
+ }
+ }
+ }
+
+ return posixcc;
+}
+
STATIC regnode *
regclass(void)
{
while (PL_regcomp_parse < PL_regxend && *PL_regcomp_parse != ']') {
skipcond:
value = UCHARAT(PL_regcomp_parse++);
- if (value == '[' && PL_regcomp_parse + 1 < PL_regxend &&
- /* I smell either [: or [= or [. -- POSIX has been here, right? */
- (*PL_regcomp_parse == ':' || *PL_regcomp_parse == '=' || *PL_regcomp_parse == '.')) {
- char posixccc = *PL_regcomp_parse;
- char* posixccs = PL_regcomp_parse++;
-
- while (PL_regcomp_parse < PL_regxend && *PL_regcomp_parse != posixccc)
- PL_regcomp_parse++;
- if (PL_regcomp_parse == PL_regxend)
- /* Grandfather lone [:, [=, [. */
- PL_regcomp_parse = posixccs;
- else {
- PL_regcomp_parse++; /* skip over the posixccc */
- if (*PL_regcomp_parse == ']') {
- /* Not Implemented Yet.
- * (POSIX Extended Character Classes, that is)
- * The text between e.g. [: and :] would start
- * at posixccs + 1 and stop at regcomp_parse - 2. */
- if (ckWARN(WARN_UNSAFE) && !SIZE_ONLY)
- warner(WARN_UNSAFE,
- "Character class syntax [%c %c] is reserved for future extensions", posixccc, posixccc);
- PL_regcomp_parse++; /* skip over the ending ] */
- }
- }
- }
- if (value == '\\') {
+ if (value == '[')
+ (void)regpposixcc(value); /* ignore the return value for now */
+ else if (value == '\\') {
value = UCHARAT(PL_regcomp_parse++);
switch (value) {
case 'w':
if (!SIZE_ONLY) {
#ifndef ASCIIish
if ((isLOWER(lastvalue) && isLOWER(value)) ||
- (isUPPER(lastvalue) && isUPPER(value))) {
+ (isUPPER(lastvalue) && isUPPER(value)))
+ {
+ I32 i;
if (isLOWER(lastvalue)) {
for (i = lastvalue; i <= value; i++)
if (isLOWER(i))
if (!SIZE_ONLY && (*opnd & (0xFF ^ ANYOF_INVERT)) == ANYOF_FOLD) {
for (value = 0; value < 256; ++value) {
if (ANYOF_TEST(opnd, value)) {
- I32 cf = fold[value];
+ I32 cf = PL_fold[value];
ANYOF_SET(opnd, cf);
}
}
value = utf8_to_uv((U8*)PL_regcomp_parse, &numlen);
PL_regcomp_parse += numlen;
- if (value == '[' && PL_regcomp_parse + 1 < PL_regxend &&
- /* I smell either [: or [= or [. -- POSIX has been here, right? */
- (*PL_regcomp_parse == ':' || *PL_regcomp_parse == '=' || *PL_regcomp_parse == '.')) {
- char posixccc = *PL_regcomp_parse;
- char* posixccs = PL_regcomp_parse++;
-
- while (PL_regcomp_parse < PL_regxend && *PL_regcomp_parse != posixccc)
- PL_regcomp_parse++;
- if (PL_regcomp_parse == PL_regxend)
- /* Grandfather lone [:, [=, [. */
- PL_regcomp_parse = posixccs;
- else {
- PL_regcomp_parse++; /* skip over the posixccc */
- if (*PL_regcomp_parse == ']') {
- /* Not Implemented Yet.
- * (POSIX Extended Character Classes, that is)
- * The text between e.g. [: and :] would start
- * at posixccs + 1 and stop at regcomp_parse - 2. */
- if (ckWARN(WARN_UNSAFE) && !SIZE_ONLY)
- warner(WARN_UNSAFE,
- "Character class syntax [%c %c] is reserved for future extensions", posixccc, posixccc);
- PL_regcomp_parse++; /* skip over the ending ] */
- }
- }
- }
-
- if (value == '\\') {
+ if (value == '[')
+ (void)regpposixcc(value); /* ignore the return value for now */
+ else if (value == '\\') {
value = utf8_to_uv((U8*)PL_regcomp_parse, &numlen);
PL_regcomp_parse += numlen;
switch (value) {
register regnode *place;
register int offset = regarglen[(U8)op];
-/* (regkind[(U8)op] == CURLY ? EXTRA_STEP_2ARGS : 0); */
+/* (PL_regkind[(U8)op] == CURLY ? EXTRA_STEP_2ARGS : 0); */
if (SIZE_ONLY) {
PL_regsize += NODE_STEP_REGNODE + offset;
/* "Operandless" and "op != BRANCH" are synonymous in practice. */
if (p == NULL || SIZE_ONLY)
return;
- if (regkind[(U8)OP(p)] == BRANCH) {
+ if (PL_regkind[(U8)OP(p)] == BRANCH) {
regtail(NEXTOPER(p), val);
}
- else if ( regkind[(U8)OP(p)] == BRANCHJ) {
+ else if ( PL_regkind[(U8)OP(p)] == BRANCHJ) {
regtail(NEXTOPER(NEXTOPER(p)), val);
}
else
PerlIO_printf(Perl_debug_log, "(%d)", next - start);
(void)PerlIO_putc(Perl_debug_log, '\n');
after_print:
- if (regkind[(U8)op] == BRANCHJ) {
+ if (PL_regkind[(U8)op] == BRANCHJ) {
register regnode *nnode = (OP(next) == LONGJMP
? regnext(next)
: next);
nnode = last;
node = dumpuntil(start, NEXTOPER(NEXTOPER(node)), nnode, sv, l + 1);
}
- else if (regkind[(U8)op] == BRANCH) {
+ else if (PL_regkind[(U8)op] == BRANCH) {
node = dumpuntil(start, NEXTOPER(node), next, sv, l + 1);
}
else if ( op == CURLY) { /* `next' might be very big: optimizer */
node = dumpuntil(start, NEXTOPER(node) + EXTRA_STEP_2ARGS,
NEXTOPER(node) + EXTRA_STEP_2ARGS + 1, sv, l + 1);
}
- else if (regkind[(U8)op] == CURLY && op != CURLYX) {
+ else if (PL_regkind[(U8)op] == CURLY && op != CURLYX) {
node = dumpuntil(start, NEXTOPER(node) + EXTRA_STEP_2ARGS,
next, sv, l + 1);
}
node = NEXTOPER(node);
node += ANY_SKIP;
}
- else if (regkind[(U8)op] == EXACT) {
+ else if (PL_regkind[(U8)op] == EXACT) {
/* Literal string, where present. */
node += ((*OPERAND(node)) + 2 + sizeof(regnode) - 1) / sizeof(regnode);
node = NEXTOPER(node);
case SEOL:
p = "SEOL";
break;
- case ANY:
+ case REG_ANY:
p = "ANY";
break;
case SANY:
SAVESPTR(PL_regcode); /* Code-emit pointer; ®dummy = don't */
SAVEPPTR(PL_regxend); /* End of input for compile */
SAVEPPTR(PL_regcomp_parse); /* Input-scan pointer. */
+ SAVESPTR(PL_reg_call_cc); /* from regexec.c */
+ SAVESPTR(PL_reg_re); /* from regexec.c */
+ SAVEPPTR(PL_reg_ganch); /* from regexec.c */
+ SAVESPTR(PL_reg_sv); /* from regexec.c */
+ SAVESPTR(PL_reg_magic); /* from regexec.c */
+ SAVEI32(PL_reg_oldpos); /* from regexec.c */
+ SAVESPTR(PL_reg_oldcurpm); /* from regexec.c */
+ SAVESPTR(PL_reg_curpm); /* from regexec.c */
+#ifdef DEBUGGING
+ SAVEPPTR(PL_reg_starttry); /* from regexec.c */
+#endif
}
https://perl5.git.perl.org / perl5.git / blobdiff