#define RF_evaled 4 /* Did an EVAL with setting? */
#define RF_utf8 8 /* String contains multibyte chars? */
-#define UTF (PL_reg_flags & RF_utf8)
+#define UTF ((PL_reg_flags & RF_utf8) != 0)
#define RS_init 1 /* eval environment created */
#define RS_set 2 /* replsv value is set */
);
}
DEBUG_r(
- if (*PL_reglastparen + 1 <= PL_regnpar) {
+ if ((I32)(*PL_reglastparen + 1) <= PL_regnpar) {
PerlIO_printf(Perl_debug_log,
" restoring \\%"IVdf"..\\%"IVdf" to undef\n",
(IV)(*PL_reglastparen + 1), (IV)PL_regnpar);
* building DynaLoader will fail:
* "Error: '*' not in typemap in DynaLoader.xs, line 164"
* --jhi */
- for (paren = *PL_reglastparen + 1; paren <= PL_regnpar; paren++) {
- if (paren > PL_regsize)
+ for (paren = *PL_reglastparen + 1; (I32)paren <= PL_regnpar; paren++) {
+ if ((I32)paren > PL_regsize)
PL_regstartp[paren] = -1;
PL_regendp[paren] = -1;
}
);
});
- if (prog->minlen > CHR_DIST((U8*)strend, (U8*)strpos)) {
+ /* CHR_DIST() would be more correct here but it makes things slow. */
+ if (prog->minlen > strend - strpos) {
DEBUG_r(PerlIO_printf(Perl_debug_log,
"String too short... [re_intuit_start]\n"));
goto fail;
t = s - prog->check_offset_max;
if (s - strpos > prog->check_offset_max /* signed-corrected t > strpos */
- && (!(prog->reganch & ROPT_UTF8)
+ && (!do_utf8
|| ((t = reghopmaybe3_c(s, -(prog->check_offset_max), strpos))
&& t > strpos)))
/* EMPTY */;
t = s - prog->check_offset_max;
if (s - strpos > prog->check_offset_max /* signed-corrected t > strpos */
- && (!(prog->reganch & ROPT_UTF8)
+ && (!do_utf8
|| ((t = reghopmaybe3_c(s, -prog->check_offset_max, strpos))
&& t > strpos))) {
/* Fixed substring is found far enough so that the match
c1 = *(U8*)m;
c2 = PL_fold_locale[c1];
do_exactf:
- e = HOP3c(strend, -ln, s);
+ e = HOP3c(strend, -(I32)ln, s);
if (norun && e < s)
e = s; /* Due to minlen logic of intuit() */
if ( c == c1
&& (ln == len ||
ibcmp_utf8(s, (char **)0, 0, do_utf8,
- m, (char **)0, ln, UTF))
+ m, (char **)0, ln, (bool)UTF))
&& (norun || regtry(prog, s)) )
goto got_it;
else {
!ibcmp_utf8((char *) foldbuf,
(char **)0, foldlen, do_utf8,
m,
- (char **)0, ln, UTF))
+ (char **)0, ln, (bool)UTF))
&& (norun || regtry(prog, s)) )
goto got_it;
}
if ( (c == c1 || c == c2)
&& (ln == len ||
ibcmp_utf8(s, (char **)0, 0, do_utf8,
- m, (char **)0, ln, UTF))
+ m, (char **)0, ln, (bool)UTF))
&& (norun || regtry(prog, s)) )
goto got_it;
else {
!ibcmp_utf8((char *) foldbuf,
(char **)0, foldlen, do_utf8,
m,
- (char **)0, ln, UTF))
+ (char **)0, ln, (bool)UTF))
&& (norun || regtry(prog, s)) )
goto got_it;
}
sp = prog->startp;
ep = prog->endp;
if (prog->nparens) {
- for (i = prog->nparens; i > *PL_reglastparen; i--) {
+ for (i = prog->nparens; i > (I32)*PL_reglastparen; i--) {
*++sp = -1;
*++ep = -1;
}
case EXACT:
s = STRING(scan);
ln = STR_LEN(scan);
- if (do_utf8 != (UTF!=0)) {
+ if (do_utf8 != UTF) {
/* The target and the pattern have differing utf8ness. */
char *l = locinput;
char *e = s + ln;
char *l = locinput;
char *e = PL_regeol;
- if (ibcmp_utf8(s, 0, ln, UTF,
+ if (ibcmp_utf8(s, 0, ln, (bool)UTF,
l, &e, 0, do_utf8)) {
/* One more case for the sharp s:
* pack("U0U*", 0xDF) =~ /ss/i,
n = ARG(scan); /* which paren pair */
ln = PL_regstartp[n];
PL_reg_leftiter = PL_reg_maxiter; /* Void cache */
- if (*PL_reglastparen < n || ln == -1)
+ if ((I32)*PL_reglastparen < n || ln == -1)
sayNO; /* Do not match unless seen CLOSEn. */
if (ln == PL_regendp[n])
break;
dSP;
OP_4tree *oop = PL_op;
COP *ocurcop = PL_curcop;
- SV **ocurpad = PL_curpad;
+ PAD *old_comppad;
SV *ret;
n = ARG(scan);
PL_op = (OP_4tree*)PL_regdata->data[n];
DEBUG_r( PerlIO_printf(Perl_debug_log, " re_eval 0x%"UVxf"\n", PTR2UV(PL_op)) );
- PL_curpad = AvARRAY((AV*)PL_regdata->data[n + 2]);
+ PAD_SAVE_LOCAL(old_comppad, (PAD*)PL_regdata->data[n + 2]);
PL_regendp[0] = PL_reg_magic->mg_len = locinput - PL_bostr;
{
CALLRUNOPS(aTHX); /* Scalar context. */
SPAGAIN;
if (SP == before)
- ret = Nullsv; /* protect against empty (?{}) blocks. */
+ ret = &PL_sv_undef; /* protect against empty (?{}) blocks. */
else {
ret = POPs;
PUTBACK;
}
PL_op = oop;
- PL_curpad = ocurpad;
+ PAD_RESTORE_LOCAL(old_comppad);
PL_curcop = ocurcop;
if (logical) {
if (logical == 2) { /* Postponed subexpression. */
MAGIC *mg = Null(MAGIC*);
re_cc_state state;
CHECKPOINT cp, lastcp;
+ int toggleutf;
if(SvROK(ret) || SvRMAGICAL(ret)) {
SV *sv = SvROK(ret) ? SvRV(ret) : ret;
I32 onpar = PL_regnpar;
Zero(&pm, 1, PMOP);
+ if (DO_UTF8(ret)) pm.op_pmdynflags |= PMdf_DYN_UTF8;
re = CALLREGCOMP(aTHX_ t, t + len, &pm);
if (!(SvFLAGS(ret)
& (SVs_TEMP | SVs_PADTMP | SVf_READONLY)))
*PL_reglastcloseparen = 0;
PL_reg_call_cc = &state;
PL_reginput = locinput;
+ toggleutf = ((PL_reg_flags & RF_utf8) != 0) ^
+ ((re->reganch & ROPT_UTF8) != 0);
+ if (toggleutf) PL_reg_flags ^= RF_utf8;
/* XXXX This is too dramatic a measure... */
PL_reg_maxiter = 0;
PL_regcc = state.cc;
PL_reg_re = state.re;
cache_re(PL_reg_re);
+ if (toggleutf) PL_reg_flags ^= RF_utf8;
/* XXXX This is too dramatic a measure... */
PL_reg_maxiter = 0;
PL_regcc = state.cc;
PL_reg_re = state.re;
cache_re(PL_reg_re);
+ if (toggleutf) PL_reg_flags ^= RF_utf8;
/* XXXX This is too dramatic a measure... */
PL_reg_maxiter = 0;
n = ARG(scan); /* which paren pair */
PL_regstartp[n] = PL_reg_start_tmp[n] - PL_bostr;
PL_regendp[n] = locinput - PL_bostr;
- if (n > *PL_reglastparen)
+ if (n > (I32)*PL_reglastparen)
*PL_reglastparen = n;
*PL_reglastcloseparen = n;
break;
case GROUPP:
n = ARG(scan); /* which paren pair */
- sw = (*PL_reglastparen >= n && PL_regendp[n] != -1);
+ sw = ((I32)*PL_reglastparen >= n && PL_regendp[n] != -1);
break;
case IFTHEN:
PL_reg_leftiter = PL_reg_maxiter; /* Void cache */
PL_regcc = &cc;
/* XXXX Probably it is better to teach regpush to support
parenfloor > PL_regsize... */
- if (parenfloor > *PL_reglastparen)
+ if (parenfloor > (I32)*PL_reglastparen)
parenfloor = *PL_reglastparen; /* Pessimization... */
cc.parenfloor = parenfloor;
cc.cur = -1;
DEBUG_r(
PerlIO_printf(Perl_debug_log,
- "%*s %ld out of %ld..%ld cc=%lx\n",
+ "%*s %ld out of %ld..%ld cc=%"UVxf"\n",
REPORT_CODE_OFF+PL_regindent*2, "",
(long)n, (long)cc->min,
- (long)cc->max, (long)cc)
+ (long)cc->max, PTR2UV(cc))
);
/* If degenerate scan matches "", assume scan done. */
if (PL_reg_leftiter-- == 0) {
I32 size = (PL_reg_maxiter + 7)/8;
if (PL_reg_poscache) {
- if (PL_reg_poscache_size < size) {
+ if ((I32)PL_reg_poscache_size < size) {
Renew(PL_reg_poscache, size, char);
PL_reg_poscache_size = size;
}
if (paren) {
if (paren > PL_regsize)
PL_regsize = paren;
- if (paren > *PL_reglastparen)
+ if (paren > (I32)*PL_reglastparen)
*PL_reglastparen = paren;
}
scan = NEXTOPER(scan) + NODE_STEP_REGNODE;
ln = PL_regstartp[n];
/* assume yes if we haven't seen CLOSEn */
if (
- *PL_reglastparen < n ||
+ (I32)*PL_reglastparen < n ||
ln == -1 ||
ln == PL_regendp[n]
) {
ln = PL_regstartp[n];
/* assume yes if we haven't seen CLOSEn */
if (
- *PL_reglastparen < n ||
+ (I32)*PL_reglastparen < n ||
ln == -1 ||
ln == PL_regendp[n]
) {
paren = scan->flags; /* Which paren to set */
if (paren > PL_regsize)
PL_regsize = paren;
- if (paren > *PL_reglastparen)
+ if (paren > (I32)*PL_reglastparen)
*PL_reglastparen = paren;
ln = ARG1(scan); /* min to match */
n = ARG2(scan); /* max to match */
ln = PL_regstartp[n];
/* assume yes if we haven't seen CLOSEn */
if (
- *PL_reglastparen < n ||
+ (I32)*PL_reglastparen < n ||
ln == -1 ||
ln == PL_regendp[n]
) {
utf8n_to_uvchr((U8*)locinput,
UTF8_MAXLEN, &len,
ckWARN(WARN_UTF8) ?
- 0 : UTF8_ALLOW_ANY) != c1) {
+ 0 : UTF8_ALLOW_ANY) != (UV)c1) {
locinput += len;
count++;
}
UTF8_MAXLEN, &len,
ckWARN(WARN_UTF8) ?
0 : UTF8_ALLOW_ANY);
- if (c == c1 || c == c2)
+ if (c == (UV)c1 || c == (UV)c2)
break;
locinput += len;
count++;
else
c = UCHARAT(PL_reginput);
/* If it could work, try it. */
- if (c == c1 || c == c2)
+ if (c == (UV)c1 || c == (UV)c2)
{
- TRYPAREN(paren, n, PL_reginput);
+ TRYPAREN(paren, ln, PL_reginput);
REGCP_UNWIND(lastcp);
}
}
/* If it could work, try it. */
else if (c1 == -1000)
{
- TRYPAREN(paren, n, PL_reginput);
+ TRYPAREN(paren, ln, PL_reginput);
REGCP_UNWIND(lastcp);
}
/* Couldn't or didn't -- move forward. */
c = UCHARAT(PL_reginput);
}
/* If it could work, try it. */
- if (c1 == -1000 || c == c1 || c == c2)
+ if (c1 == -1000 || c == (UV)c1 || c == (UV)c2)
{
TRYPAREN(paren, n, PL_reginput);
REGCP_UNWIND(lastcp);
c = UCHARAT(PL_reginput);
}
/* If it could work, try it. */
- if (c1 == -1000 || c == c1 || c == c2)
+ if (c1 == -1000 || c == (UV)c1 || c == (UV)c2)
{
TRYPAREN(paren, n, PL_reginput);
REGCP_UNWIND(lastcp);
register bool do_utf8 = PL_reg_match_utf8;
scan = PL_reginput;
- if (max != REG_INFTY && max < loceol - scan)
+ if (max == REG_INFTY)
+ max = I32_MAX;
+ else if (max < loceol - scan)
loceol = scan + max;
switch (OP(p)) {
case REG_ANY:
if (PL_regdata->what[n] == 's') {
SV *rv = (SV*)PL_regdata->data[n];
AV *av = (AV*)SvRV((SV*)rv);
+ SV **ary = AvARRAY(av);
SV **a, **b;
/* See the end of regcomp.c:S_reglass() for
* documentation of these array elements. */
- si = *av_fetch(av, 0, FALSE);
- a = av_fetch(av, 1, FALSE);
- b = av_fetch(av, 2, FALSE);
-
+ si = *ary;
+ a = SvTYPE(ary[1]) == SVt_RV ? &ary[1] : 0;
+ b = SvTYPE(ary[2]) == SVt_PVAV ? &ary[2] : 0;
+
if (a)
sw = *a;
else if (si && doinit) {
{
char flags = ANYOF_FLAGS(n);
bool match = FALSE;
- UV c;
+ UV c = *p;
STRLEN len = 0;
STRLEN plen;
- c = do_utf8 ? utf8n_to_uvchr(p, UTF8_MAXLEN, &len,
- ckWARN(WARN_UTF8) ? 0 : UTF8_ALLOW_ANY) : *p;
+ if (do_utf8 && !UTF8_IS_INVARIANT(c))
+ c = utf8n_to_uvchr(p, UTF8_MAXLEN, &len,
+ ckWARN(WARN_UTF8) ? 0 : UTF8_ALLOW_ANY);
plen = lenp ? *lenp : UNISKIP(NATIVE_TO_UNI(c));
if (do_utf8 || (flags & ANYOF_UNICODE)) {
if (ANYOF_BITMAP_TEST(n, c))
match = TRUE;
else if (flags & ANYOF_FOLD) {
- I32 f;
+ U8 f;
if (flags & ANYOF_LOCALE) {
PL_reg_flags |= RF_tainted;
SV* sv;
if (prog->float_substr && !prog->float_utf8) {
prog->float_utf8 = sv = NEWSV(117, 0);
- SvSetMagicSV(sv, prog->float_substr);
+ SvSetSV(sv, prog->float_substr);
sv_utf8_upgrade(sv);
if (SvTAIL(prog->float_substr))
SvTAIL_on(sv);
}
if (prog->anchored_substr && !prog->anchored_utf8) {
prog->anchored_utf8 = sv = NEWSV(118, 0);
- SvSetMagicSV(sv, prog->anchored_substr);
+ SvSetSV(sv, prog->anchored_substr);
sv_utf8_upgrade(sv);
if (SvTAIL(prog->anchored_substr))
SvTAIL_on(sv);
SV* sv;
if (prog->float_utf8 && !prog->float_substr) {
prog->float_substr = sv = NEWSV(117, 0);
- SvSetMagicSV(sv, prog->float_utf8);
+ SvSetSV(sv, prog->float_utf8);
if (sv_utf8_downgrade(sv, TRUE)) {
if (SvTAIL(prog->float_utf8))
SvTAIL_on(sv);
}
if (prog->anchored_utf8 && !prog->anchored_substr) {
prog->anchored_substr = sv = NEWSV(118, 0);
- SvSetMagicSV(sv, prog->anchored_utf8);
+ SvSetSV(sv, prog->anchored_utf8);
if (sv_utf8_downgrade(sv, TRUE)) {
if (SvTAIL(prog->anchored_utf8))
SvTAIL_on(sv);