if (!SvIOK(re_trie_maxbuff)) {
sv_setiv(re_trie_maxbuff, RE_TRIE_MAXBUF_INIT);
}
- DEBUG_OPTIMISE_r({
+ DEBUG_TRIE_COMPILE_r({
PerlIO_printf( Perl_debug_log,
"%*smake_trie start==%d, first==%d, last==%d, tail==%d depth=%d\n",
(int)depth * 2 + 2, "",
*/
for ( cur = first ; cur < last ; cur = regnext( cur ) ) {
- regnode * const noper = NEXTOPER( cur );
+ regnode *noper = NEXTOPER( cur );
const U8 *uc = (U8*)STRING( noper );
- const U8 * const e = uc + STR_LEN( noper );
+ const U8 *e = uc + STR_LEN( noper );
STRLEN foldlen = 0;
U8 foldbuf[ UTF8_MAXBYTES_CASE + 1 ];
STRLEN skiplen = 0;
bool set_bit = trie->bitmap ? 1 : 0; /*store the first char in the bitmap?*/
if (OP(noper) == NOTHING) {
- trie->minlen= 0;
- continue;
+ regnode *noper_next= regnext(noper);
+ if (noper_next != tail && OP(noper_next) == flags) {
+ noper = noper_next;
+ uc= (U8*)STRING(noper);
+ e= uc + STR_LEN(noper);
+ trie->minlen= STR_LEN(noper);
+ } else {
+ trie->minlen= 0;
+ continue;
+ }
}
+
if ( set_bit ) { /* bitmap only alloced when !(UTF&&Folding) */
TRIE_BITMAP_SET(trie,*uc); /* store the raw first byte
regardless of encoding */
for ( cur = first ; cur < last ; cur = regnext( cur ) ) {
- regnode * const noper = NEXTOPER( cur );
+ regnode *noper = NEXTOPER( cur );
U8 *uc = (U8*)STRING( noper );
- const U8 * const e = uc + STR_LEN( noper );
+ const U8 *e = uc + STR_LEN( noper );
U32 state = 1; /* required init */
U16 charid = 0; /* sanity init */
U8 *scan = (U8*)NULL; /* sanity init */
U8 foldbuf[ UTF8_MAXBYTES_CASE + 1 ];
STRLEN skiplen = 0;
+ if (OP(noper) == NOTHING) {
+ regnode *noper_next= regnext(noper);
+ if (noper_next != tail && OP(noper_next) == flags) {
+ noper = noper_next;
+ uc= (U8*)STRING(noper);
+ e= uc + STR_LEN(noper);
+ }
+ }
+
if (OP(noper) != NOTHING) {
for ( ; uc < e ; uc += len ) {
for ( cur = first ; cur < last ; cur = regnext( cur ) ) {
- regnode * const noper = NEXTOPER( cur );
+ regnode *noper = NEXTOPER( cur );
const U8 *uc = (U8*)STRING( noper );
- const U8 * const e = uc + STR_LEN( noper );
+ const U8 *e = uc + STR_LEN( noper );
U32 state = 1; /* required init */
STRLEN skiplen = 0;
U8 foldbuf[ UTF8_MAXBYTES_CASE + 1 ];
+ if (OP(noper) == NOTHING) {
+ regnode *noper_next= regnext(noper);
+ if (noper_next != tail && OP(noper_next) == flags) {
+ noper = noper_next;
+ uc= (U8*)STRING(noper);
+ e= uc + STR_LEN(noper);
+ }
+ }
if ( OP(noper) != NOTHING ) {
for ( ; uc < e ; uc += len ) {
}
- DEBUG_OPTIMISE_r({
+ DEBUG_TRIE_COMPILE_r({
regprop(RExC_rx, mysv, tail );
PerlIO_printf( Perl_debug_log, "%*s%s%s\n",
(int)depth * 2 + 2, "",
U8 noper_trietype = TRIE_TYPE( noper_type );
#if defined(DEBUGGING) || defined(NOJUMPTRIE)
regnode * const noper_next = regnext( noper );
+ U8 noper_next_type = (noper_next && noper_next != tail) ? OP(noper_next) : 0;
+ U8 noper_next_trietype = (noper_next && noper_next != tail) ? TRIE_TYPE( noper_next_type ) :0;
#endif
- DEBUG_OPTIMISE_r({
+ DEBUG_TRIE_COMPILE_r({
regprop(RExC_rx, mysv, cur);
PerlIO_printf( Perl_debug_log, "%*s- %s (%d)",
(int)depth * 2 + 2,"", SvPV_nolen_const( mysv ), REG_NODE_NUM(cur) );
PerlIO_printf( Perl_debug_log,"\t=> %s\t",
SvPV_nolen_const(mysv));
}
- PerlIO_printf( Perl_debug_log, "(First==%d,Last==%d,Cur==%d)\n",
- REG_NODE_NUM(first), REG_NODE_NUM(last), REG_NODE_NUM(cur) );
+ PerlIO_printf( Perl_debug_log, "(First==%d,Last==%d,Cur==%d,tt==%s,nt==%s,nnt==%s)\n",
+ REG_NODE_NUM(first), REG_NODE_NUM(last), REG_NODE_NUM(cur),
+ PL_reg_name[trietype], PL_reg_name[noper_trietype], PL_reg_name[noper_next_trietype]
+ );
});
/* Is noper a trieable nodetype that can be merged with the
if ( noper_trietype
&&
(
- /* XXX: Currently we cannot allow a NOTHING node to be the first element
- * of a TRIEABLE sequence, Otherwise we will overwrite the regop following
- * the NOTHING with the TRIE regop later on. This is because a NOTHING node
- * is only one regnode wide, and a TRIE is two regnodes. An example of a
- * problematic pattern is: "x" =~ /\A(?>(?:(?:)A|B|C?x))\z/
- * At a later point of time we can somewhat workaround this by handling
- * NOTHING -> EXACT sequences as generated by /(?:)A|(?:)B/ type patterns,
- * as we can effectively ignore the NOTHING regop in that case.
- * This clause, which allows NOTHING to start a sequence is left commented
- * out as a reference.
- * - Yves
-
- ( noper_trietype == NOTHING)
- || ( trietype == NOTHING )
- */
- ( noper_trietype == NOTHING && trietype )
+ ( noper_trietype == NOTHING)
+ || ( trietype == NOTHING )
|| ( trietype == noper_trietype )
)
#ifdef NOJUMPTRIE
* Either we are the first node in a new trieable sequence,
* in which case we do some bookkeeping, otherwise we update
* the end pointer. */
- count++;
if ( !first ) {
first = cur;
trietype = noper_trietype;
+ if ( noper_trietype == NOTHING ) {
+#if !defined(DEBUGGING) && !defined(NOJUMPTRIE)
+ regnode * const noper_next = regnext( noper );
+ U8 noper_next_type = (noper_next && noper_next!=tail) ? OP(noper_next) : 0;
+ U8 noper_next_trietype = noper_next_type ? TRIE_TYPE( noper_next_type ) :0;
+#endif
+
+ if ( noper_next_trietype )
+ trietype = noper_next_trietype;
+ }
} else {
if ( trietype == NOTHING )
trietype = noper_trietype;
last = cur;
}
+ if (first)
+ count++;
} /* end handle mergable triable node */
else {
/* handle unmergable node -
}
} /* end handle unmergable node */
} /* loop over branches */
- DEBUG_OPTIMISE_r({
+ DEBUG_TRIE_COMPILE_r({
regprop(RExC_rx, mysv, cur);
PerlIO_printf( Perl_debug_log,
"%*s- %s (%d) <SCAN FINISHED>\n", (int)depth * 2 + 2,
"", SvPV_nolen_const( mysv ),REG_NODE_NUM(cur));
});
- if ( last && trietype != NOTHING ) {
- /* the last branch of the sequence was part of a trie,
- * so we have to construct it here outside of the loop
- */
- made= make_trie( pRExC_state, startbranch, first, scan, tail, count, trietype, depth+1 );
+ if ( last ) {
+ if ( trietype != NOTHING ) {
+ /* the last branch of the sequence was part of a trie,
+ * so we have to construct it here outside of the loop
+ */
+ made= make_trie( pRExC_state, startbranch, first, scan, tail, count, trietype, depth+1 );
#ifdef TRIE_STUDY_OPT
- if ( ((made == MADE_EXACT_TRIE &&
- startbranch == first)
- || ( first_non_open == first )) &&
- depth==0 ) {
- flags |= SCF_TRIE_RESTUDY;
- if ( startbranch == first
- && scan == tail )
- {
- RExC_seen &=~REG_TOP_LEVEL_BRANCHES;
+ if ( ((made == MADE_EXACT_TRIE &&
+ startbranch == first)
+ || ( first_non_open == first )) &&
+ depth==0 ) {
+ flags |= SCF_TRIE_RESTUDY;
+ if ( startbranch == first
+ && scan == tail )
+ {
+ RExC_seen &=~REG_TOP_LEVEL_BRANCHES;
+ }
}
- }
#endif
+ } else {
+ /* at this point we know whatever we have is a NOTHING sequence/branch
+ * AND if 'startbranch' is 'first' then we can turn the whole thing into a NOTHING
+ */
+ if ( startbranch == first ) {
+ regnode *opt;
+ /* the entire thing is a NOTHING sequence, something like this:
+ * (?:|) So we can turn it into a plain NOTHING op. */
+ DEBUG_TRIE_COMPILE_r({
+ regprop(RExC_rx, mysv, cur);
+ PerlIO_printf( Perl_debug_log,
+ "%*s- %s (%d) <NOTHING BRANCH SEQUENCE>\n", (int)depth * 2 + 2,
+ "", SvPV_nolen_const( mysv ),REG_NODE_NUM(cur));
+
+ });
+ OP(startbranch)= NOTHING;
+ NEXT_OFF(startbranch)= tail - startbranch;
+ for ( opt= startbranch + 1; opt < tail ; opt++ )
+ OP(opt)= OPTIMIZED;
+ }
+ }
} /* end if ( last) */
} /* TRIE_MAXBUF is non zero */
/* Lookbehind, or need to calculate parens/evals/stclass: */
&& (scan->flags || data || (flags & SCF_DO_STCLASS))
&& (OP(scan) == IFMATCH || OP(scan) == UNLESSM)) {
+ if ( OP(scan) == UNLESSM &&
+ scan->flags == 0 &&
+ OP(NEXTOPER(NEXTOPER(scan))) == NOTHING &&
+ OP(regnext(NEXTOPER(NEXTOPER(scan)))) == SUCCEED
+ ) {
+ regnode *opt;
+ regnode *upto= regnext(scan);
+ DEBUG_PARSE_r({
+ SV * const mysv_val=sv_newmortal();
+ DEBUG_STUDYDATA("OPFAIL",data,depth);
+
+ /*DEBUG_PARSE_MSG("opfail");*/
+ regprop(RExC_rx, mysv_val, upto);
+ PerlIO_printf(Perl_debug_log, "~ replace with OPFAIL pointed at %s (%"IVdf") offset %"IVdf"\n",
+ SvPV_nolen_const(mysv_val),
+ (IV)REG_NODE_NUM(upto),
+ (IV)(upto - scan)
+ );
+ });
+ OP(scan) = OPFAIL;
+ NEXT_OFF(scan) = upto - scan;
+ for (opt= scan + 1; opt < upto ; opt++)
+ OP(opt) = OPTIMIZED;
+ scan= upto;
+ continue;
+ }
if ( !PERL_ENABLE_POSITIVE_ASSERTION_STUDY
|| OP(scan) == UNLESSM )
{
}
}
}
-
-
}
#endif
}
* Clever compilers notice this and complain. --jhi */
REGC((U8)REG_MAGIC, (char*)RExC_emit);
#endif
- DEBUG_PARSE_r(
- PerlIO_printf(Perl_debug_log, "Starting first pass (sizing)\n");
- RExC_lastnum=0;
- RExC_lastparse=NULL;
- );
+ DEBUG_PARSE_r(PerlIO_printf(Perl_debug_log, "Starting first pass (sizing)\n"));
if (reg(pRExC_state, 0, &flags,1) == NULL) {
RExC_precomp = NULL;
return(NULL);
RExC_parse++;
}
if (*RExC_parse != ')') {
- RExC_parse = s;
+ RExC_parse = s;
vFAIL("Sequence (?{...}) not terminated or not {}-balanced");
}
if (!SIZE_ONLY) {
}
break;
}
+ DEBUG_PARSE_r(if (!SIZE_ONLY) {
+ SV * const mysv_val1=sv_newmortal();
+ SV * const mysv_val2=sv_newmortal();
+ DEBUG_PARSE_MSG("lsbr");
+ regprop(RExC_rx, mysv_val1, lastbr);
+ regprop(RExC_rx, mysv_val2, ender);
+ PerlIO_printf(Perl_debug_log, "~ tying lastbr %s (%"IVdf") to ender %s (%"IVdf") offset %"IVdf"\n",
+ SvPV_nolen_const(mysv_val1),
+ (IV)REG_NODE_NUM(lastbr),
+ SvPV_nolen_const(mysv_val2),
+ (IV)REG_NODE_NUM(ender),
+ (IV)(ender - lastbr)
+ );
+ });
REGTAIL(pRExC_state, lastbr, ender);
if (have_branch && !SIZE_ONLY) {
+ char is_nothing= 1;
if (depth==1)
RExC_seen |= REG_TOP_LEVEL_BRANCHES;
const U8 op = PL_regkind[OP(br)];
if (op == BRANCH) {
REGTAIL_STUDY(pRExC_state, NEXTOPER(br), ender);
+ if (OP(NEXTOPER(br)) != NOTHING || regnext(NEXTOPER(br)) != ender)
+ is_nothing= 0;
}
else if (op == BRANCHJ) {
REGTAIL_STUDY(pRExC_state, NEXTOPER(NEXTOPER(br)), ender);
+ /* for now we always disable this optimisation * /
+ if (OP(NEXTOPER(NEXTOPER(br))) != NOTHING || regnext(NEXTOPER(NEXTOPER(br))) != ender)
+ */
+ is_nothing= 0;
}
}
+ if (is_nothing) {
+ br= PL_regkind[OP(ret)] != BRANCH ? regnext(ret) : ret;
+ DEBUG_PARSE_r(if (!SIZE_ONLY) {
+ SV * const mysv_val1=sv_newmortal();
+ SV * const mysv_val2=sv_newmortal();
+ DEBUG_PARSE_MSG("NADA");
+ regprop(RExC_rx, mysv_val1, ret);
+ regprop(RExC_rx, mysv_val2, ender);
+ PerlIO_printf(Perl_debug_log, "~ converting ret %s (%"IVdf") to ender %s (%"IVdf") offset %"IVdf"\n",
+ SvPV_nolen_const(mysv_val1),
+ (IV)REG_NODE_NUM(ret),
+ SvPV_nolen_const(mysv_val2),
+ (IV)REG_NODE_NUM(ender),
+ (IV)(ender - ret)
+ );
+ });
+ OP(br)= NOTHING;
+ if (OP(ender) == TAIL) {
+ NEXT_OFF(br)= 0;
+ RExC_emit= br + 1;
+ } else {
+ regnode *opt;
+ for ( opt= br + 1; opt < ender ; opt++ )
+ OP(opt)= OPTIMIZED;
+ NEXT_OFF(br)= ender - br;
+ }
+ }
}
}
vFAIL("Internal urp");
/* Supposed to be caught earlier. */
break;
- case '{':
- if (!regcurly(RExC_parse)) {
- RExC_parse++;
- goto defchar;
- }
- /* FALL THROUGH */
case '?':
case '+':
case '*':
ret = reg_node(pRExC_state, op);
FLAGS(ret) = get_regex_charset(RExC_flags);
*flagp |= SIMPLE;
- if (! SIZE_ONLY && (U8) *(RExC_parse + 1) == '{') {
- ckWARNregdep(RExC_parse, "\"\\b{\" is deprecated; use \"\\b\\{\" instead");
- }
goto finish_meta_pat;
case 'B':
RExC_seen_zerolen++;
ret = reg_node(pRExC_state, op);
FLAGS(ret) = get_regex_charset(RExC_flags);
*flagp |= SIMPLE;
- if (! SIZE_ONLY && (U8) *(RExC_parse + 1) == '{') {
- ckWARNregdep(RExC_parse, "\"\\B{\" is deprecated; use \"\\B\\{\" instead");
- }
goto finish_meta_pat;
case 's':
switch (get_regex_charset(RExC_flags)) {
/* FALL THROUGH */
default:
if (!SIZE_ONLY&& isALPHA(*p)) {
- /* Include any { following the alpha to emphasize
- * that it could be part of an escape at some point
- * in the future */
- int len = (*(p + 1) == '{') ? 2 : 1;
- ckWARN3reg(p + len, "Unrecognized escape \\%.*s passed through", len, p);
+ ckWARN2reg(p + 1, "Unrecognized escape \\%.1s passed through", p);
}
goto normal_default;
}
break;
+ case '{':
+ /* Currently we don't warn when the lbrace is at the start
+ * of a construct. This catches it in the middle of a
+ * literal string, or when its the first thing after
+ * something like "\b" */
+ if (! SIZE_ONLY
+ && (len || (p > RExC_start && isALPHA_A(*(p -1)))))
+ {
+ ckWARNregdep(p + 1, "Unescaped left brace in regex is deprecated, passed through");
+ }
+ /*FALLTHROUGH*/
default:
normal_default:
if (UTF8_IS_START(*p) && UTF) {
* Local variables:
* c-indentation-style: bsd
* c-basic-offset: 4
- * indent-tabs-mode: t
+ * indent-tabs-mode: nil
* End:
*
- * ex: set ts=8 sts=4 sw=4 noet:
+ * ex: set ts=8 sts=4 sw=4 et:
*/