**** Alterations to Henry's code are...
****
**** Copyright (C) 1991, 1992, 1993, 1994, 1995, 1996, 1997, 1998, 1999,
- **** 2000, 2001, 2002, 2003, 2004, 2005, by Larry Wall and others
+ **** 2000, 2001, 2002, 2003, 2004, 2005, 2006, by Larry Wall and others
****
**** You may distribute under the terms of either the GNU General Public
**** License or the Artistic License, as specified in the README file.
STATIC CHECKPOINT
S_regcppush(pTHX_ I32 parenfloor)
{
+ dVAR;
const int retval = PL_savestack_ix;
#define REGCP_PAREN_ELEMS 4
const int paren_elems_to_push = (PL_regsize - parenfloor) * REGCP_PAREN_ELEMS;
STATIC char *
S_regcppop(pTHX)
{
+ dVAR;
I32 i;
U32 paren = 0;
char *input;
STATIC char *
S_regcp_set_to(pTHX_ I32 ss)
{
+ dVAR;
const I32 tmp = PL_savestack_ix;
PL_savestack_ix = ss;
STATIC void
S_cache_re(pTHX_ regexp *prog)
{
+ dVAR;
PL_regprecomp = prog->precomp; /* Needed for FAIL. */
#ifdef DEBUGGING
PL_regprogram = prog->program;
Perl_re_intuit_start(pTHX_ regexp *prog, SV *sv, char *strpos,
char *strend, U32 flags, re_scream_pos_data *data)
{
+ dVAR;
register I32 start_shift = 0;
/* Should be nonnegative! */
register I32 end_shift = 0;
|| ((slen = SvCUR(check)) > 1
&& memNE(SvPVX_const(check), s, slen)))
goto report_neq;
+ check_at = s;
goto success_at_start;
}
}
s = find_byclass(prog, prog->regstclass, s, endpos, 1);
if (!s) {
#ifdef DEBUGGING
- const char *what = 0;
+ const char *what = NULL;
#endif
if (endpos == strend) {
DEBUG_EXECUTE_r( PerlIO_printf(Perl_debug_log,
U8 *sm = (U8 *) m;
U8 tmpbuf1[UTF8_MAXBYTES_CASE+1];
U8 tmpbuf2[UTF8_MAXBYTES_CASE+1];
+ const U32 uniflags = ckWARN(WARN_UTF8) ? 0 : UTF8_ALLOW_ANY;
to_utf8_lower((U8*)m, tmpbuf1, &ulen1);
to_utf8_upper((U8*)m, tmpbuf2, &ulen2);
c1 = utf8n_to_uvchr(tmpbuf1, UTF8_MAXBYTES_CASE,
- 0, ckWARN(WARN_UTF8) ? 0 : UTF8_ALLOW_ANY);
+ 0, uniflags);
c2 = utf8n_to_uvchr(tmpbuf2, UTF8_MAXBYTES_CASE,
- 0, ckWARN(WARN_UTF8) ? 0 : UTF8_ALLOW_ANY);
+ 0, uniflags);
lnc = 0;
while (sm < ((U8 *) m + ln)) {
lnc++;
UV c, f;
U8 tmpbuf [UTF8_MAXBYTES+1];
STRLEN len, foldlen;
-
+ const U32 uniflags = ckWARN(WARN_UTF8) ? 0 : UTF8_ALLOW_ANY;
if (c1 == c2) {
/* Upper and lower of 1st char are equal -
* probably not a "letter". */
while (s <= e) {
c = utf8n_to_uvchr((U8*)s, UTF8_MAXBYTES, &len,
- ckWARN(WARN_UTF8) ?
- 0 : UTF8_ALLOW_ANY);
+ uniflags);
if ( c == c1
&& (ln == len ||
ibcmp_utf8(s, (char **)0, 0, do_utf8,
else {
while (s <= e) {
c = utf8n_to_uvchr((U8*)s, UTF8_MAXBYTES, &len,
- ckWARN(WARN_UTF8) ?
- 0 : UTF8_ALLOW_ANY);
+ uniflags);
/* Handle some of the three Greek sigmas cases.
* Note that not all the possible combinations
/* data: May be used for some additional optimizations. */
/* nosave: For optimizations. */
{
+ dVAR;
register char *s;
register regnode *c;
register char *startpos = stringarg;
STATIC I32 /* 0 failure, 1 success */
S_regtry(pTHX_ regexp *prog, char *startpos)
{
+ dVAR;
register I32 i;
register I32 *sp;
register I32 *ep;
I32 unwind = 0;
/* used by the trie code */
- SV *sv_accept_buff = 0; /* accepting states we have traversed */
- reg_trie_accepted *accept_buff = 0; /* "" */
+ SV *sv_accept_buff = NULL; /* accepting states we have traversed */
+ reg_trie_accepted *accept_buff = NULL; /* "" */
reg_trie_data *trie; /* what trie are we using right now */
U32 accepted = 0; /* how many accepting states we have seen*/
#if 0
I32 firstcp = PL_savestack_ix;
#endif
- const register bool do_utf8 = PL_reg_match_utf8;
+ register const bool do_utf8 = PL_reg_match_utf8;
#ifdef DEBUGGING
SV *dsv0 = PERL_DEBUG_PAD_ZERO(0);
SV *dsv1 = PERL_DEBUG_PAD_ZERO(1);
SV *re_debug_flags = NULL;
#endif
+ U32 uniflags = ckWARN(WARN_UTF8) ? 0 : UTF8_ALLOW_ANY;
GET_RE_DEBUG_FLAGS;
case TRIEF:
case TRIEFL:
{
-
- const U32 uniflags = ckWARN( WARN_UTF8 ) ? 0 : UTF8_ALLOW_ANY;
U8 *uc = ( U8* )locinput;
U32 state = 1;
U16 charid = 0;
from previous if blocks */
case TRIE:
{
- const U32 uniflags = ckWARN( WARN_UTF8 ) ? 0 : UTF8_ALLOW_ANY;
U8 *uc = (U8*)locinput;
U32 state = 1;
U16 charid = 0;
sayNO;
if (NATIVE_TO_UNI(*(U8*)s) !=
utf8n_to_uvuni((U8*)l, UTF8_MAXBYTES, &ulen,
- ckWARN(WARN_UTF8) ?
- 0 : UTF8_ALLOW_ANY))
+ uniflags))
sayNO;
l += ulen;
s ++;
sayNO;
if (NATIVE_TO_UNI(*((U8*)l)) !=
utf8n_to_uvuni((U8*)s, UTF8_MAXBYTES, &ulen,
- ckWARN(WARN_UTF8) ?
- 0 : UTF8_ALLOW_ANY))
+ uniflags))
sayNO;
s += ulen;
l ++;
to_utf8_upper((U8*)s, tmpbuf2, &ulen2);
c1 = utf8n_to_uvuni(tmpbuf1, UTF8_MAXBYTES, 0,
- ckWARN(WARN_UTF8) ?
- 0 : UTF8_ALLOW_ANY);
+ uniflags);
c2 = utf8n_to_uvuni(tmpbuf2, UTF8_MAXBYTES, 0,
- ckWARN(WARN_UTF8) ?
- 0 : UTF8_ALLOW_ANY);
+ uniflags);
}
else {
c2 = c1 = utf8n_to_uvchr(s, UTF8_MAXBYTES, 0,
- ckWARN(WARN_UTF8) ?
- 0 : UTF8_ALLOW_ANY);
+ uniflags);
}
}
}
while (locinput <= e &&
utf8n_to_uvchr((U8*)locinput,
UTF8_MAXBYTES, &len,
- ckWARN(WARN_UTF8) ?
- 0 : UTF8_ALLOW_ANY) != (UV)c1) {
+ uniflags) != (UV)c1) {
locinput += len;
count++;
}
while (locinput <= e) {
UV c = utf8n_to_uvchr((U8*)locinput,
UTF8_MAXBYTES, &len,
- ckWARN(WARN_UTF8) ?
- 0 : UTF8_ALLOW_ANY);
+ uniflags);
if (c == (UV)c1 || c == (UV)c2)
break;
locinput += len;
if (do_utf8)
c = utf8n_to_uvchr((U8*)PL_reginput,
UTF8_MAXBYTES, 0,
- ckWARN(WARN_UTF8) ?
- 0 : UTF8_ALLOW_ANY);
+ uniflags);
else
c = UCHARAT(PL_reginput);
/* If it could work, try it. */
if (do_utf8)
c = utf8n_to_uvchr((U8*)PL_reginput,
UTF8_MAXBYTES, 0,
- ckWARN(WARN_UTF8) ?
- 0 : UTF8_ALLOW_ANY);
+ uniflags);
else
c = UCHARAT(PL_reginput);
}
if (do_utf8)
c = utf8n_to_uvchr((U8*)PL_reginput,
UTF8_MAXBYTES, 0,
- ckWARN(WARN_UTF8) ?
- 0 : UTF8_ALLOW_ANY);
+ uniflags);
else
c = UCHARAT(PL_reginput);
}
STATIC I32
S_regrepeat_hard(pTHX_ regnode *p, I32 max, I32 *lp)
{
+ dVAR;
register char *scan = Nullch;
register char *start;
register char *loceol = PL_regeol;
SV *
Perl_regclass_swash(pTHX_ register const regnode* node, bool doinit, SV** listsvp, SV **altsvp)
{
+ dVAR;
SV *sw = NULL;
SV *si = NULL;
SV *alt = NULL;
if (PL_regdata->what[n] == 's') {
SV * const rv = (SV*)PL_regdata->data[n];
AV * const av = (AV*)SvRV((SV*)rv);
- SV **ary = AvARRAY(av);
+ SV **const ary = AvARRAY(av);
SV **a, **b;
- /* See the end of regcomp.c:S_reglass() for
+ /* See the end of regcomp.c:S_regclass() for
* documentation of these array elements. */
si = *ary;
- a = SvTYPE(ary[1]) == SVt_RV ? &ary[1] : 0;
+ a = SvROK(ary[1]) ? &ary[1] : 0;
b = SvTYPE(ary[2]) == SVt_PVAV ? &ary[2] : 0;
if (a)
STRLEN len = 0;
STRLEN plen;
- if (do_utf8 && !UTF8_IS_INVARIANT(c))
- c = utf8n_to_uvchr(p, UTF8_MAXBYTES, &len,
- ckWARN(WARN_UTF8) ? 0 : UTF8_ALLOW_ANY);
+ if (do_utf8 && !UTF8_IS_INVARIANT(c)) {
+ c = utf8n_to_uvchr(p, UTF8_MAXBYTES, &len,
+ ckWARN(WARN_UTF8) ? UTF8_CHECK_ONLY :
+ UTF8_ALLOW_ANYUV|UTF8_CHECK_ONLY);
+ if (len == (STRLEN)-1)
+ Perl_croak(aTHX_ "Malformed UTF-8 character (fatal)");
+ }
plen = lenp ? *lenp : UNISKIP(NATIVE_TO_UNI(c));
if (do_utf8 || (flags & ANYOF_UNICODE)) {
STATIC U8 *
S_reghop(pTHX_ U8 *s, I32 off)
{
+ dVAR;
return S_reghop3(aTHX_ s, off, (U8*)(off >= 0 ? PL_regeol : PL_bostr));
}
STATIC U8 *
S_reghop3(pTHX_ U8 *s, I32 off, U8* lim)
{
+ dVAR;
if (off >= 0) {
while (off-- && s < lim) {
/* XXX could check well-formedness here */
STATIC U8 *
S_reghopmaybe(pTHX_ U8 *s, I32 off)
{
+ dVAR;
return S_reghopmaybe3(aTHX_ s, off, (U8*)(off >= 0 ? PL_regeol : PL_bostr));
}
STATIC U8 *
S_reghopmaybe3(pTHX_ U8* s, I32 off, U8* lim)
{
+ dVAR;
if (off >= 0) {
while (off-- && s < lim) {
/* XXX could check well-formedness here */
static void
restore_pos(pTHX_ void *arg)
{
+ dVAR;
PERL_UNUSED_ARG(arg);
if (PL_reg_eval_set) {
if (PL_reg_oldsaved) {
STATIC void
S_to_byte_substr(pTHX_ register regexp *prog)
{
+ dVAR;
if (prog->float_utf8 && !prog->float_substr) {
SV* sv;
prog->float_substr = sv = newSVsv(prog->float_utf8);