#endif
bool seen_unfolded_sharp_s;
bool strict;
+ bool study_started;
};
#define RExC_flags (pRExC_state->flags)
#define RExC_frame_last (pRExC_state->frame_last)
#define RExC_frame_count (pRExC_state->frame_count)
#define RExC_strict (pRExC_state->strict)
+#define RExC_study_started (pRExC_state->study_started)
/* Heuristic check on the complexity of the pattern: if TOO_NAUGHTY, we set
* a flag to disable back-off on the fixed/floating substrings - if it's
#define SF_HAS_PAR 0x0080
#define SF_IN_PAR 0x0100
#define SF_HAS_EVAL 0x0200
+
+
+/* SCF_DO_SUBSTR is the flag that tells the regexp analyzer to track the
+ * longest substring in the pattern. When it is not set the optimiser keeps
+ * track of position, but does not keep track of the actual strings seen,
+ *
+ * So for instance /foo/ will be parsed with SCF_DO_SUBSTR being true, but
+ * /foo/i will not.
+ *
+ * Similarly, /foo.*(blah|erm|huh).*fnorble/ will have "foo" and "fnorble"
+ * parsed with SCF_DO_SUBSTR on, but while processing the (...) it will be
+ * turned off because of the alternation (BRANCH). */
#define SCF_DO_SUBSTR 0x0400
+
#define SCF_DO_STCLASS_AND 0x0800
#define SCF_DO_STCLASS_OR 0x1000
#define SCF_DO_STCLASS (SCF_DO_STCLASS_AND|SCF_DO_STCLASS_OR)
#define EXPERIMENTAL_INPLACESCAN
#endif /*PERL_ENABLE_EXPERIMENTAL_REGEX_OPTIMISATIONS*/
-#define DEBUG_RExC_seen() \
+#ifdef DEBUGGING
+int
+Perl_re_printf(pTHX_ const char *fmt, ...)
+{
+ va_list ap;
+ int result;
+ PerlIO *f= Perl_debug_log;
+ PERL_ARGS_ASSERT_RE_PRINTF;
+ va_start(ap, fmt);
+ result = PerlIO_vprintf(f, fmt, ap);
+ va_end(ap);
+ return result;
+}
+
+int
+Perl_re_indentf(pTHX_ const char *fmt, U32 depth, ...)
+{
+ va_list ap;
+ int result;
+ PerlIO *f= Perl_debug_log;
+ PERL_ARGS_ASSERT_RE_INDENTF;
+ va_start(ap, depth);
+ PerlIO_printf(f, "%*s", ( (int)depth % 20 ) * 2, "");
+ result = PerlIO_vprintf(f, fmt, ap);
+ va_end(ap);
+ return result;
+}
+#endif /* DEBUGGING */
+
+#define DEBUG_RExC_seen() \
DEBUG_OPTIMISE_MORE_r({ \
- PerlIO_printf(Perl_debug_log,"RExC_seen: "); \
+ Perl_re_printf( aTHX_ "RExC_seen: "); \
\
if (RExC_seen & REG_ZERO_LEN_SEEN) \
- PerlIO_printf(Perl_debug_log,"REG_ZERO_LEN_SEEN "); \
+ Perl_re_printf( aTHX_ "REG_ZERO_LEN_SEEN "); \
\
if (RExC_seen & REG_LOOKBEHIND_SEEN) \
- PerlIO_printf(Perl_debug_log,"REG_LOOKBEHIND_SEEN "); \
+ Perl_re_printf( aTHX_ "REG_LOOKBEHIND_SEEN "); \
\
if (RExC_seen & REG_GPOS_SEEN) \
- PerlIO_printf(Perl_debug_log,"REG_GPOS_SEEN "); \
+ Perl_re_printf( aTHX_ "REG_GPOS_SEEN "); \
\
if (RExC_seen & REG_RECURSE_SEEN) \
- PerlIO_printf(Perl_debug_log,"REG_RECURSE_SEEN "); \
+ Perl_re_printf( aTHX_ "REG_RECURSE_SEEN "); \
\
- if (RExC_seen & REG_TOP_LEVEL_BRANCHES_SEEN) \
- PerlIO_printf(Perl_debug_log,"REG_TOP_LEVEL_BRANCHES_SEEN "); \
+ if (RExC_seen & REG_TOP_LEVEL_BRANCHES_SEEN) \
+ Perl_re_printf( aTHX_ "REG_TOP_LEVEL_BRANCHES_SEEN "); \
\
if (RExC_seen & REG_VERBARG_SEEN) \
- PerlIO_printf(Perl_debug_log,"REG_VERBARG_SEEN "); \
+ Perl_re_printf( aTHX_ "REG_VERBARG_SEEN "); \
\
if (RExC_seen & REG_CUTGROUP_SEEN) \
- PerlIO_printf(Perl_debug_log,"REG_CUTGROUP_SEEN "); \
+ Perl_re_printf( aTHX_ "REG_CUTGROUP_SEEN "); \
\
if (RExC_seen & REG_RUN_ON_COMMENT_SEEN) \
- PerlIO_printf(Perl_debug_log,"REG_RUN_ON_COMMENT_SEEN "); \
+ Perl_re_printf( aTHX_ "REG_RUN_ON_COMMENT_SEEN "); \
\
if (RExC_seen & REG_UNFOLDED_MULTI_SEEN) \
- PerlIO_printf(Perl_debug_log,"REG_UNFOLDED_MULTI_SEEN "); \
+ Perl_re_printf( aTHX_ "REG_UNFOLDED_MULTI_SEEN "); \
\
- if (RExC_seen & REG_UNBOUNDED_QUANTIFIER_SEEN) \
- PerlIO_printf(Perl_debug_log,"REG_UNBOUNDED_QUANTIFIER_SEEN "); \
+ if (RExC_seen & REG_UNBOUNDED_QUANTIFIER_SEEN) \
+ Perl_re_printf( aTHX_ "REG_UNBOUNDED_QUANTIFIER_SEEN "); \
\
- PerlIO_printf(Perl_debug_log,"\n"); \
+ Perl_re_printf( aTHX_ "\n"); \
});
#define DEBUG_SHOW_STUDY_FLAG(flags,flag) \
- if ((flags) & flag) PerlIO_printf(Perl_debug_log, "%s ", #flag)
+ if ((flags) & flag) Perl_re_printf( aTHX_ "%s ", #flag)
#define DEBUG_SHOW_STUDY_FLAGS(flags,open_str,close_str) \
if ( ( flags ) ) { \
- PerlIO_printf(Perl_debug_log, "%s", open_str); \
+ Perl_re_printf( aTHX_ "%s", open_str); \
DEBUG_SHOW_STUDY_FLAG(flags,SF_FL_BEFORE_SEOL); \
DEBUG_SHOW_STUDY_FLAG(flags,SF_FL_BEFORE_MEOL); \
DEBUG_SHOW_STUDY_FLAG(flags,SF_IS_INF); \
DEBUG_SHOW_STUDY_FLAG(flags,SCF_SEEN_ACCEPT); \
DEBUG_SHOW_STUDY_FLAG(flags,SCF_TRIE_DOING_RESTUDY); \
DEBUG_SHOW_STUDY_FLAG(flags,SCF_IN_DEFINE); \
- PerlIO_printf(Perl_debug_log, "%s", close_str); \
+ Perl_re_printf( aTHX_ "%s", close_str); \
}
#define DEBUG_STUDYDATA(str,data,depth) \
DEBUG_OPTIMISE_MORE_r(if(data){ \
- PerlIO_printf(Perl_debug_log, \
- "%*s" str "Pos:%"IVdf"/%"IVdf \
+ Perl_re_indentf( aTHX_ "" str "Pos:%"IVdf"/%"IVdf \
" Flags: 0x%"UVXf, \
- (int)(depth)*2, "", \
+ depth, \
(IV)((data)->pos_min), \
(IV)((data)->pos_delta), \
(UV)((data)->flags) \
); \
DEBUG_SHOW_STUDY_FLAGS((data)->flags," [ ","]"); \
- PerlIO_printf(Perl_debug_log, \
+ Perl_re_printf( aTHX_ \
" Whilem_c: %"IVdf" Lcp: %"IVdf" %s", \
(IV)((data)->whilem_c), \
(IV)((data)->last_closep ? *((data)->last_closep) : -1), \
is_inf ? "INF " : "" \
); \
if ((data)->last_found) \
- PerlIO_printf(Perl_debug_log, \
+ Perl_re_printf( aTHX_ \
"Last:'%s' %"IVdf":%"IVdf"/%"IVdf" %sFixed:'%s' @ %"IVdf \
" %sFloat: '%s' @ %"IVdf"/%"IVdf"", \
SvPVX_const((data)->last_found), \
(IV)((data)->offset_float_min), \
(IV)((data)->offset_float_max) \
); \
- PerlIO_printf(Perl_debug_log,"\n"); \
+ Perl_re_printf( aTHX_ "\n"); \
});
+
/* =========================================================
* BEGIN edit_distance stuff.
*
PERL_ARGS_ASSERT_DUMP_TRIE;
- PerlIO_printf( Perl_debug_log, "%*sChar : %-6s%-6s%-4s ",
- (int)depth * 2 + 2,"",
- "Match","Base","Ofs" );
+ Perl_re_indentf( aTHX_ "Char : %-6s%-6s%-4s ",
+ depth+1, "Match","Base","Ofs" );
for( state = 0 ; state < trie->uniquecharcount ; state++ ) {
SV ** const tmp = av_fetch( revcharmap, state, 0);
if ( tmp ) {
- PerlIO_printf( Perl_debug_log, "%*s",
+ Perl_re_printf( aTHX_ "%*s",
colwidth,
pv_pretty(sv, SvPV_nolen_const(*tmp), SvCUR(*tmp), colwidth,
PL_colors[0], PL_colors[1],
);
}
}
- PerlIO_printf( Perl_debug_log, "\n%*sState|-----------------------",
- (int)depth * 2 + 2,"");
+ Perl_re_printf( aTHX_ "\n");
+ Perl_re_indentf( aTHX_ "State|-----------------------", depth+1);
for( state = 0 ; state < trie->uniquecharcount ; state++ )
- PerlIO_printf( Perl_debug_log, "%.*s", colwidth, "--------");
- PerlIO_printf( Perl_debug_log, "\n");
+ Perl_re_printf( aTHX_ "%.*s", colwidth, "--------");
+ Perl_re_printf( aTHX_ "\n");
for( state = 1 ; state < trie->statecount ; state++ ) {
const U32 base = trie->states[ state ].trans.base;
- PerlIO_printf( Perl_debug_log, "%*s#%4"UVXf"|",
- (int)depth * 2 + 2,"", (UV)state);
+ Perl_re_indentf( aTHX_ "#%4"UVXf"|", depth+1, (UV)state);
if ( trie->states[ state ].wordnum ) {
- PerlIO_printf( Perl_debug_log, " W%4X",
- trie->states[ state ].wordnum );
+ Perl_re_printf( aTHX_ " W%4X", trie->states[ state ].wordnum );
} else {
- PerlIO_printf( Perl_debug_log, "%6s", "" );
+ Perl_re_printf( aTHX_ "%6s", "" );
}
- PerlIO_printf( Perl_debug_log, " @%4"UVXf" ", (UV)base );
+ Perl_re_printf( aTHX_ " @%4"UVXf" ", (UV)base );
if ( base ) {
U32 ofs = 0;
!= state))
ofs++;
- PerlIO_printf( Perl_debug_log, "+%2"UVXf"[ ", (UV)ofs);
+ Perl_re_printf( aTHX_ "+%2"UVXf"[ ", (UV)ofs);
for ( ofs = 0 ; ofs < trie->uniquecharcount ; ofs++ ) {
if ( ( base + ofs >= trie->uniquecharcount )
&& trie->trans[ base + ofs
- trie->uniquecharcount ].check == state )
{
- PerlIO_printf( Perl_debug_log, "%*"UVXf,
- colwidth,
- (UV)trie->trans[ base + ofs
- - trie->uniquecharcount ].next );
+ Perl_re_printf( aTHX_ "%*"UVXf, colwidth,
+ (UV)trie->trans[ base + ofs - trie->uniquecharcount ].next
+ );
} else {
- PerlIO_printf( Perl_debug_log, "%*s",colwidth," ." );
+ Perl_re_printf( aTHX_ "%*s",colwidth," ." );
}
}
- PerlIO_printf( Perl_debug_log, "]");
+ Perl_re_printf( aTHX_ "]");
}
- PerlIO_printf( Perl_debug_log, "\n" );
+ Perl_re_printf( aTHX_ "\n" );
}
- PerlIO_printf(Perl_debug_log, "%*sword_info N:(prev,len)=",
- (int)depth*2, "");
+ Perl_re_indentf( aTHX_ "word_info N:(prev,len)=",
+ depth);
for (word=1; word <= trie->wordcount; word++) {
- PerlIO_printf(Perl_debug_log, " %d:(%d,%d)",
+ Perl_re_printf( aTHX_ " %d:(%d,%d)",
(int)word, (int)(trie->wordinfo[word].prev),
(int)(trie->wordinfo[word].len));
}
- PerlIO_printf(Perl_debug_log, "\n" );
+ Perl_re_printf( aTHX_ "\n" );
}
/*
Dumps a fully constructed but uncompressed trie in list form.
PERL_ARGS_ASSERT_DUMP_TRIE_INTERIM_LIST;
/* print out the table precompression. */
- PerlIO_printf( Perl_debug_log, "%*sState :Word | Transition Data\n%*s%s",
- (int)depth * 2 + 2,"", (int)depth * 2 + 2,"",
- "------:-----+-----------------\n" );
+ Perl_re_indentf( aTHX_ "State :Word | Transition Data\n",
+ depth+1 );
+ Perl_re_indentf( aTHX_ "%s",
+ depth+1, "------:-----+-----------------\n" );
for( state=1 ; state < next_alloc ; state ++ ) {
U16 charid;
- PerlIO_printf( Perl_debug_log, "%*s %4"UVXf" :",
- (int)depth * 2 + 2,"", (UV)state );
+ Perl_re_indentf( aTHX_ " %4"UVXf" :",
+ depth+1, (UV)state );
if ( ! trie->states[ state ].wordnum ) {
- PerlIO_printf( Perl_debug_log, "%5s| ","");
+ Perl_re_printf( aTHX_ "%5s| ","");
} else {
- PerlIO_printf( Perl_debug_log, "W%4x| ",
+ Perl_re_printf( aTHX_ "W%4x| ",
trie->states[ state ].wordnum
);
}
SV ** const tmp = av_fetch( revcharmap,
TRIE_LIST_ITEM(state,charid).forid, 0);
if ( tmp ) {
- PerlIO_printf( Perl_debug_log, "%*s:%3X=%4"UVXf" | ",
+ Perl_re_printf( aTHX_ "%*s:%3X=%4"UVXf" | ",
colwidth,
pv_pretty(sv, SvPV_nolen_const(*tmp), SvCUR(*tmp),
colwidth,
(UV)TRIE_LIST_ITEM(state,charid).newstate
);
if (!(charid % 10))
- PerlIO_printf(Perl_debug_log, "\n%*s| ",
+ Perl_re_printf( aTHX_ "\n%*s| ",
(int)((depth * 2) + 14), "");
}
}
- PerlIO_printf( Perl_debug_log, "\n");
+ Perl_re_printf( aTHX_ "\n");
}
}
that they are identical.
*/
- PerlIO_printf( Perl_debug_log, "%*sChar : ",(int)depth * 2 + 2,"" );
+ Perl_re_indentf( aTHX_ "Char : ", depth+1 );
for( charid = 0 ; charid < trie->uniquecharcount ; charid++ ) {
SV ** const tmp = av_fetch( revcharmap, charid, 0);
if ( tmp ) {
- PerlIO_printf( Perl_debug_log, "%*s",
+ Perl_re_printf( aTHX_ "%*s",
colwidth,
pv_pretty(sv, SvPV_nolen_const(*tmp), SvCUR(*tmp), colwidth,
PL_colors[0], PL_colors[1],
}
}
- PerlIO_printf( Perl_debug_log, "\n%*sState+-",(int)depth * 2 + 2,"" );
+ Perl_re_printf( aTHX_ "\n");
+ Perl_re_indentf( aTHX_ "State+-", depth+1 );
for( charid=0 ; charid < trie->uniquecharcount ; charid++ ) {
- PerlIO_printf( Perl_debug_log, "%.*s", colwidth,"--------");
+ Perl_re_printf( aTHX_ "%.*s", colwidth,"--------");
}
- PerlIO_printf( Perl_debug_log, "\n" );
+ Perl_re_printf( aTHX_ "\n" );
for( state=1 ; state < next_alloc ; state += trie->uniquecharcount ) {
- PerlIO_printf( Perl_debug_log, "%*s%4"UVXf" : ",
- (int)depth * 2 + 2,"",
+ Perl_re_indentf( aTHX_ "%4"UVXf" : ",
+ depth+1,
(UV)TRIE_NODENUM( state ) );
for( charid = 0 ; charid < trie->uniquecharcount ; charid++ ) {
UV v=(UV)SAFE_TRIE_NODENUM( trie->trans[ state + charid ].next );
if (v)
- PerlIO_printf( Perl_debug_log, "%*"UVXf, colwidth, v );
+ Perl_re_printf( aTHX_ "%*"UVXf, colwidth, v );
else
- PerlIO_printf( Perl_debug_log, "%*s", colwidth, "." );
+ Perl_re_printf( aTHX_ "%*s", colwidth, "." );
}
if ( ! trie->states[ TRIE_NODENUM( state ) ].wordnum ) {
- PerlIO_printf( Perl_debug_log, " (%4"UVXf")\n",
+ Perl_re_printf( aTHX_ " (%4"UVXf")\n",
(UV)trie->trans[ state ].check );
} else {
- PerlIO_printf( Perl_debug_log, " (%4"UVXf") W%4X\n",
+ Perl_re_printf( aTHX_ " (%4"UVXf") W%4X\n",
(UV)trie->trans[ state ].check,
trie->states[ TRIE_NODENUM( state ) ].wordnum );
}
sv_setiv(re_trie_maxbuff, RE_TRIE_MAXBUF_INIT);
}
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, "",
+ Perl_re_indentf( aTHX_
+ "make_trie start==%d, first==%d, last==%d, tail==%d depth=%d\n",
+ depth+1,
REG_NODE_NUM(startbranch),REG_NODE_NUM(first),
REG_NODE_NUM(last), REG_NODE_NUM(tail), (int)depth);
});
for ( cur = first ; cur < last ; cur = regnext( cur ) ) {
regnode *noper = NEXTOPER( cur );
- const U8 *uc = (U8*)STRING( noper );
- const U8 *e = uc + STR_LEN( noper );
+ const U8 *uc;
+ const U8 *e;
int foldlen = 0;
U32 wordlen = 0; /* required init */
STRLEN minchars = 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);
- trie->minlen= STR_LEN(noper);
- } else {
- trie->minlen= 0;
- continue;
- }
+ if (noper_next < tail)
+ noper= noper_next;
+ }
+
+ if ( noper < tail && ( OP(noper) == flags || ( flags == EXACTFU && OP(noper) == EXACTFU_SS ) ) ) {
+ uc= (U8*)STRING(noper);
+ e= uc + 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 */
}
} /* end first pass */
DEBUG_TRIE_COMPILE_r(
- PerlIO_printf( Perl_debug_log,
- "%*sTRIE(%s): W:%d C:%d Uq:%d Min:%d Max:%d\n",
- (int)depth * 2 + 2,"",
+ Perl_re_indentf( aTHX_
+ "TRIE(%s): W:%d C:%d Uq:%d Min:%d Max:%d\n",
+ depth+1,
( widecharmap ? "UTF8" : "NATIVE" ), (int)word_count,
(int)TRIE_CHARCOUNT(trie), trie->uniquecharcount,
(int)trie->minlen, (int)trie->maxlen )
STRLEN transcount = 1;
- DEBUG_TRIE_COMPILE_MORE_r( PerlIO_printf( Perl_debug_log,
- "%*sCompiling trie using list compiler\n",
- (int)depth * 2 + 2, ""));
+ DEBUG_TRIE_COMPILE_MORE_r( Perl_re_indentf( aTHX_ "Compiling trie using list compiler\n",
+ depth+1));
trie->states = (reg_trie_state *)
PerlMemShared_calloc( TRIE_CHARCOUNT(trie) + 2,
for ( cur = first ; cur < last ; cur = regnext( cur ) ) {
regnode *noper = NEXTOPER( cur );
- U8 *uc = (U8*)STRING( noper );
- const U8 *e = uc + STR_LEN( noper );
U32 state = 1; /* required init */
U16 charid = 0; /* sanity init */
U32 wordlen = 0; /* required init */
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 (noper_next < tail)
+ noper= noper_next;
}
- if (OP(noper) != NOTHING) {
+ if ( noper < tail && ( OP(noper) == flags || ( flags == EXACTFU && OP(noper) == EXACTFU_SS ) ) ) {
+ const U8 *uc= (U8*)STRING(noper);
+ const U8 *e= uc + STR_LEN(noper);
+
for ( ; uc < e ; uc += len ) {
TRIE_READ_CHAR;
/*
DEBUG_TRIE_COMPILE_MORE_r(
- PerlIO_printf( Perl_debug_log, "tp: %d zp: %d ",tp,zp)
+ Perl_re_printf( aTHX_ "tp: %d zp: %d ",tp,zp)
);
*/
}
/*
DEBUG_TRIE_COMPILE_MORE_r(
- PerlIO_printf( Perl_debug_log, " base: %d\n",base);
+ Perl_re_printf( aTHX_ " base: %d\n",base);
);
*/
trie->states[ state ].trans.base=base;
we have to use TRIE_NODENUM() to convert.
*/
- DEBUG_TRIE_COMPILE_MORE_r( PerlIO_printf( Perl_debug_log,
- "%*sCompiling trie using table compiler\n",
- (int)depth * 2 + 2, ""));
+ DEBUG_TRIE_COMPILE_MORE_r( Perl_re_indentf( aTHX_ "Compiling trie using table compiler\n",
+ depth+1));
trie->trans = (reg_trie_trans *)
PerlMemShared_calloc( ( TRIE_CHARCOUNT(trie) + 1 )
for ( cur = first ; cur < last ; cur = regnext( cur ) ) {
regnode *noper = NEXTOPER( cur );
- const U8 *uc = (U8*)STRING( noper );
- const U8 *e = uc + STR_LEN( noper );
U32 state = 1; /* required init */
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 (noper_next < tail)
+ noper= noper_next;
}
- if ( OP(noper) != NOTHING ) {
+ if ( noper < tail && ( OP(noper) == flags || ( flags == EXACTFU && OP(noper) == EXACTFU_SS ) ) ) {
+ const U8 *uc= (U8*)STRING(noper);
+ const U8 *e= uc + STR_LEN(noper);
+
for ( ; uc < e ; uc += len ) {
TRIE_READ_CHAR;
PerlMemShared_realloc( trie->states, laststate
* sizeof(reg_trie_state) );
DEBUG_TRIE_COMPILE_MORE_r(
- PerlIO_printf( Perl_debug_log,
- "%*sAlloc: %d Orig: %"IVdf" elements, Final:%"IVdf". Savings of %%%5.2f\n",
- (int)depth * 2 + 2,"",
+ Perl_re_indentf( aTHX_ "Alloc: %d Orig: %"IVdf" elements, Final:%"IVdf". Savings of %%%5.2f\n",
+ depth+1,
(int)( ( TRIE_CHARCOUNT(trie) + 1 ) * trie->uniquecharcount
+ 1 ),
(IV)next_alloc,
} /* end table compress */
}
DEBUG_TRIE_COMPILE_MORE_r(
- PerlIO_printf(Perl_debug_log,
- "%*sStatecount:%"UVxf" Lasttrans:%"UVxf"\n",
- (int)depth * 2 + 2, "",
+ Perl_re_indentf( aTHX_ "Statecount:%"UVxf" Lasttrans:%"UVxf"\n",
+ depth+1,
(UV)trie->statecount,
(UV)trie->lasttrans)
);
});
}
DEBUG_OPTIMISE_r(
- PerlIO_printf(Perl_debug_log,
- "%*sMJD offset:%"UVuf" MJD length:%"UVuf"\n",
- (int)depth * 2 + 2, "",
+ Perl_re_indentf( aTHX_ "MJD offset:%"UVuf" MJD length:%"UVuf"\n",
+ depth+1,
(UV)mjd_offset, (UV)mjd_nodelen)
);
#endif
if ( count == 2 ) {
Zero(trie->bitmap, ANYOF_BITMAP_SIZE, char);
DEBUG_OPTIMISE_r(
- PerlIO_printf(Perl_debug_log,
- "%*sNew Start State=%"UVuf" Class: [",
- (int)depth * 2 + 2, "",
+ Perl_re_indentf( aTHX_ "New Start State=%"UVuf" Class: [",
+ depth+1,
(UV)state));
if (idx >= 0) {
SV ** const tmp = av_fetch( revcharmap, idx, 0);
if ( folder )
TRIE_BITMAP_SET(trie, folder[ *ch ]);
DEBUG_OPTIMISE_r(
- PerlIO_printf(Perl_debug_log, "%s", (char*)ch)
+ Perl_re_printf( aTHX_ "%s", (char*)ch)
);
}
}
TRIE_BITMAP_SET(trie,*ch);
if ( folder )
TRIE_BITMAP_SET(trie,folder[ *ch ]);
- DEBUG_OPTIMISE_r(PerlIO_printf( Perl_debug_log,"%s", ch));
+ DEBUG_OPTIMISE_r(Perl_re_printf( aTHX_ "%s", ch));
}
idx = ofs;
}
char *ch = SvPV( *tmp, len );
DEBUG_OPTIMISE_r({
SV *sv=sv_newmortal();
- PerlIO_printf( Perl_debug_log,
- "%*sPrefix State: %"UVuf" Idx:%"UVuf" Char='%s'\n",
- (int)depth * 2 + 2, "",
+ Perl_re_indentf( aTHX_ "Prefix State: %"UVuf" Idx:%"UVuf" Char='%s'\n",
+ depth+1,
(UV)state, (UV)idx,
pv_pretty(sv, SvPV_nolen_const(*tmp), SvCUR(*tmp), 6,
PL_colors[0], PL_colors[1],
} else {
#ifdef DEBUGGING
if (state>1)
- DEBUG_OPTIMISE_r(PerlIO_printf( Perl_debug_log,"]\n"));
+ DEBUG_OPTIMISE_r(Perl_re_printf( aTHX_ "]\n"));
#endif
break;
}
*/
fail[ 0 ] = fail[ 1 ] = 0;
DEBUG_TRIE_COMPILE_r({
- PerlIO_printf(Perl_debug_log,
- "%*sStclass Failtable (%"UVuf" states): 0",
- (int)(depth * 2), "", (UV)numstates
+ Perl_re_indentf( aTHX_ "Stclass Failtable (%"UVuf" states): 0",
+ depth, (UV)numstates
);
for( q_read=1; q_read<numstates; q_read++ ) {
- PerlIO_printf(Perl_debug_log, ", %"UVuf, (UV)fail[q_read]);
+ Perl_re_printf( aTHX_ ", %"UVuf, (UV)fail[q_read]);
}
- PerlIO_printf(Perl_debug_log, "\n");
+ Perl_re_printf( aTHX_ "\n");
});
Safefree(q);
/*RExC_seen |= REG_TRIEDFA_SEEN;*/
}
-#define DEBUG_PEEP(str,scan,depth) \
- DEBUG_OPTIMISE_r({if (scan){ \
- regnode *Next = regnext(scan); \
- regprop(RExC_rx, RExC_mysv, scan, NULL, pRExC_state); \
- PerlIO_printf(Perl_debug_log, "%*s" str ">%3d: %s (%d)", \
- (int)depth*2, "", REG_NODE_NUM(scan), SvPV_nolen_const(RExC_mysv),\
- Next ? (REG_NODE_NUM(Next)) : 0 ); \
+#define DEBUG_PEEP(str,scan,depth) \
+ DEBUG_OPTIMISE_r({if (scan){ \
+ regnode *Next = regnext(scan); \
+ regprop(RExC_rx, RExC_mysv, scan, NULL, pRExC_state);\
+ Perl_re_indentf( aTHX_ "" str ">%3d: %s (%d)", \
+ depth, REG_NODE_NUM(scan), SvPV_nolen_const(RExC_mysv),\
+ Next ? (REG_NODE_NUM(Next)) : 0 );\
DEBUG_SHOW_STUDY_FLAGS(flags," [ ","]");\
- PerlIO_printf(Perl_debug_log, "\n"); \
+ Perl_re_printf( aTHX_ "\n"); \
}});
/* The below joins as many adjacent EXACTish nodes as possible into a single
GET_RE_DEBUG_FLAGS_DECL;
PERL_ARGS_ASSERT_STUDY_CHUNK;
+ RExC_study_started= 1;
if ( depth == 0 ) {
);
DEBUG_OPTIMISE_MORE_r(
{
- PerlIO_printf(Perl_debug_log,
- "%*sstudy_chunk stopparen=%ld recursed_count=%lu depth=%lu recursed_depth=%lu scan=%p last=%p",
- (int)(depth*2), "", (long)stopparen,
+ Perl_re_indentf( aTHX_ "study_chunk stopparen=%ld recursed_count=%lu depth=%lu recursed_depth=%lu scan=%p last=%p",
+ depth, (long)stopparen,
(unsigned long)RExC_study_chunk_recursed_count,
(unsigned long)depth, (unsigned long)recursed_depth,
scan,
(( j - 1 ) * RExC_study_chunk_recursed_bytes), i)
)
) {
- PerlIO_printf(Perl_debug_log," %d",(int)i);
+ Perl_re_printf( aTHX_ " %d",(int)i);
break;
}
}
if ( j + 1 < recursed_depth ) {
- PerlIO_printf(Perl_debug_log, ",");
+ Perl_re_printf( aTHX_ ",");
}
}
}
- PerlIO_printf(Perl_debug_log,"\n");
+ Perl_re_printf( aTHX_ "\n");
}
);
while ( scan && OP(scan) != END && scan < last ){
DEBUG_PEEP("Peep", scan, depth);
- /* The reason we do this here we need to deal with things like /(?:f)(?:o)(?:o)/
- * which cant be dealt with by the normal EXACT parsing code, as each (?:..) is handled
- * by a different invocation of reg() -- Yves
+ /* The reason we do this here is that we need to deal with things like
+ * /(?:f)(?:o)(?:o)/ which cant be dealt with by the normal EXACT
+ * parsing code, as each (?:..) is handled by a different invocation of
+ * reg() -- Yves
*/
JOIN_EXACT(scan,&min_subtract, &unfolded_multi_char, 0);
DEBUG_TRIE_COMPILE_r({
regprop(RExC_rx, RExC_mysv, tail, NULL, pRExC_state);
- PerlIO_printf( Perl_debug_log, "%*s%s %"UVuf":%s\n",
- (int)depth * 2 + 2, "",
+ Perl_re_indentf( aTHX_ "%s %"UVuf":%s\n",
+ depth+1,
"Looking for TRIE'able sequences. Tail node is ",
(UV)(tail - RExC_emit_start),
SvPV_nolen_const( RExC_mysv )
DEBUG_TRIE_COMPILE_r({
regprop(RExC_rx, RExC_mysv, cur, NULL, pRExC_state);
- PerlIO_printf( Perl_debug_log, "%*s- %d:%s (%d)",
- (int)depth * 2 + 2,"",
+ Perl_re_indentf( aTHX_ "- %d:%s (%d)",
+ depth+1,
REG_NODE_NUM(cur), SvPV_nolen_const( RExC_mysv ), REG_NODE_NUM(cur) );
regprop(RExC_rx, RExC_mysv, noper, NULL, pRExC_state);
- PerlIO_printf( Perl_debug_log, " -> %d:%s",
+ Perl_re_printf( aTHX_ " -> %d:%s",
REG_NODE_NUM(noper), SvPV_nolen_const(RExC_mysv));
if ( noper_next ) {
regprop(RExC_rx, RExC_mysv, noper_next, NULL, pRExC_state);
- PerlIO_printf( Perl_debug_log,"\t=> %d:%s\t",
+ Perl_re_printf( aTHX_ "\t=> %d:%s\t",
REG_NODE_NUM(noper_next), SvPV_nolen_const(RExC_mysv));
}
- PerlIO_printf( Perl_debug_log, "(First==%d,Last==%d,Cur==%d,tt==%s,ntt==%s,nntt==%s)\n",
+ Perl_re_printf( aTHX_ "(First==%d,Last==%d,Cur==%d,tt==%s,ntt==%s,nntt==%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]
);
|| ( trietype == noper_trietype )
)
#ifdef NOJUMPTRIE
- && noper_next == tail
+ && noper_next >= tail
#endif
&& count < U16_MAX)
{
}
if ( noper_trietype
#ifdef NOJUMPTRIE
- && noper_next == tail
+ && noper_next >= tail
#endif
){
/* noper is triable, so we can start a new
} /* loop over branches */
DEBUG_TRIE_COMPILE_r({
regprop(RExC_rx, RExC_mysv, cur, NULL, pRExC_state);
- PerlIO_printf( Perl_debug_log,
- "%*s- %s (%d) <SCAN FINISHED> ",
- (int)depth * 2 + 2,
- "", SvPV_nolen_const( RExC_mysv ),REG_NODE_NUM(cur));
- PerlIO_printf( Perl_debug_log, "(First==%d, Last==%d, Cur==%d, tt==%s)\n",
+ Perl_re_indentf( aTHX_ "- %s (%d) <SCAN FINISHED> ",
+ depth+1, SvPV_nolen_const( RExC_mysv ),REG_NODE_NUM(cur));
+ Perl_re_printf( aTHX_ "(First==%d, Last==%d, Cur==%d, tt==%s)\n",
REG_NODE_NUM(first), REG_NODE_NUM(last), REG_NODE_NUM(cur),
PL_reg_name[trietype]
);
depth==0 ) {
flags |= SCF_TRIE_RESTUDY;
if ( startbranch == first
- && scan == tail )
+ && scan >= tail )
{
RExC_seen &=~REG_TOP_LEVEL_BRANCHES_SEEN;
}
* turn it into a plain NOTHING op. */
DEBUG_TRIE_COMPILE_r({
regprop(RExC_rx, RExC_mysv, cur, NULL, pRExC_state);
- PerlIO_printf( Perl_debug_log,
- "%*s- %s (%d) <NOTHING BRANCH SEQUENCE>\n", (int)depth * 2 + 2,
- "", SvPV_nolen_const( RExC_mysv ),REG_NODE_NUM(cur));
+ Perl_re_indentf( aTHX_ "- %s (%d) <NOTHING BRANCH SEQUENCE>\n",
+ depth+1,
+ SvPV_nolen_const( RExC_mysv ),REG_NODE_NUM(cur));
});
OP(startbranch)= NOTHING;
/* It is counted once already... */
data->pos_min += minnext * (mincount - counted);
#if 0
-PerlIO_printf(Perl_debug_log, "counted=%"UVuf" deltanext=%"UVuf
+Perl_re_printf( aTHX_ "counted=%"UVuf" deltanext=%"UVuf
" SSize_t_MAX=%"UVuf" minnext=%"UVuf
" maxcount=%"UVuf" mincount=%"UVuf"\n",
(UV)counted, (UV)deltanext, (UV)SSize_t_MAX, (UV)minnext, (UV)maxcount,
(UV)mincount);
if (deltanext != SSize_t_MAX)
-PerlIO_printf(Perl_debug_log, "LHS=%"UVuf" RHS=%"UVuf"\n",
+Perl_re_printf( aTHX_ "LHS=%"UVuf" RHS=%"UVuf"\n",
(UV)(-counted * deltanext + (minnext + deltanext) * maxcount
- minnext * mincount), (UV)(SSize_t_MAX - data->pos_delta));
#endif
/* Else: zero-length, ignore. */
scan = regnext(scan);
}
- /* If we are exiting a recursion we can unset its recursed bit
- * and allow ourselves to enter it again - no danger of an
- * infinite loop there.
- if (stopparen > -1 && recursed) {
- DEBUG_STUDYDATA("unset:", data,depth);
- PAREN_UNSET( recursed, stopparen);
- }
- */
+
+ finish:
if (frame) {
+ /* we need to unwind recursion. */
depth = depth - 1;
DEBUG_STUDYDATA("frame-end:",data,depth);
goto fake_study_recurse;
}
- finish:
assert(!frame);
DEBUG_STUDYDATA("pre-fin:",data,depth);
/* Dispatch a request to compile a regexp to correct regexp engine. */
DEBUG_COMPILE_r({
- PerlIO_printf(Perl_debug_log, "Using engine %"UVxf"\n",
+ Perl_re_printf( aTHX_ "Using engine %"UVxf"\n",
PTR2UV(eng));
});
return CALLREGCOMP_ENG(eng, pattern, flags);
bool do_end = 0;
GET_RE_DEBUG_FLAGS_DECL;
- DEBUG_PARSE_r(PerlIO_printf(Perl_debug_log,
+ DEBUG_PARSE_r(Perl_re_printf( aTHX_
"UTF8 mismatch! Converting to utf8 for resizing and compile\n"));
Newx(dst, *plen_p * 2 + 1, U8);
*p++ = 'x';
*p++ = '\0';
DEBUG_COMPILE_r({
- PerlIO_printf(Perl_debug_log,
+ Perl_re_printf( aTHX_
"%sre-parsing pattern for runtime code:%s %s\n",
PL_colors[4],PL_colors[5],newpat);
});
}
- DEBUG_PARSE_r(PerlIO_printf(Perl_debug_log,
+ DEBUG_PARSE_r(Perl_re_printf( aTHX_
"Assembling pattern from %d elements%s\n", pat_count,
orig_rx_flags & RXf_SPLIT ? " for split" : ""));
*is_bare_re = TRUE;
SvREFCNT_inc(re);
Safefree(pRExC_state->code_blocks);
- DEBUG_PARSE_r(PerlIO_printf(Perl_debug_log,
+ DEBUG_PARSE_r(Perl_re_printf( aTHX_
"Precompiled pattern%s\n",
orig_rx_flags & RXf_SPLIT ? " for split" : ""));
RExC_contains_locale = 0;
RExC_contains_i = 0;
RExC_strict = cBOOL(pm_flags & RXf_PMf_STRICT);
+ RExC_study_started = 0;
pRExC_state->runtime_code_qr = NULL;
RExC_frame_head= NULL;
RExC_frame_last= NULL;
DEBUG_COMPILE_r({
SV *dsv= sv_newmortal();
RE_PV_QUOTED_DECL(s, RExC_utf8, dsv, exp, plen, 60);
- PerlIO_printf(Perl_debug_log, "%sCompiling REx%s %s\n",
+ Perl_re_printf( aTHX_ "%sCompiling REx%s %s\n",
PL_colors[4],PL_colors[5],s);
});
assert(*RExC_end == '\0');
DEBUG_PARSE_r(
- PerlIO_printf(Perl_debug_log, "Starting first pass (sizing)\n");
+ Perl_re_printf( aTHX_ "Starting first pass (sizing)\n");
RExC_lastnum=0;
RExC_lastparse=NULL;
);
pRExC_state->num_code_blocks);
}
else {
- DEBUG_PARSE_r(PerlIO_printf(Perl_debug_log,
+ DEBUG_PARSE_r(Perl_re_printf( aTHX_
"Need to redo pass 1\n"));
}
SvLEN_set(code_blocksv,0); /* no you can't have it, sv_clear */
DEBUG_PARSE_r({
- PerlIO_printf(Perl_debug_log,
+ Perl_re_printf( aTHX_
"Required size %"IVdf" nodes\n"
"Starting second pass (creation)\n",
(IV)RExC_size);
/* Useful during FAIL. */
#ifdef RE_TRACK_PATTERN_OFFSETS
Newxz(ri->u.offsets, 2*RExC_size+1, U32); /* MJD 20001228 */
- DEBUG_OFFSETS_r(PerlIO_printf(Perl_debug_log,
+ DEBUG_OFFSETS_r(Perl_re_printf( aTHX_
"%s %"UVuf" bytes for offset annotations.\n",
ri->u.offsets ? "Got" : "Couldn't get",
(UV)((2*RExC_size+1) * sizeof(U32))));
/* setup various meta data about recursion, this all requires
* RExC_npar to be correctly set, and a bit later on we clear it */
if (RExC_seen & REG_RECURSE_SEEN) {
- DEBUG_OPTIMISE_MORE_r(PerlIO_printf(Perl_debug_log,
+ DEBUG_OPTIMISE_MORE_r(Perl_re_printf( aTHX_
"%*s%*s Setting up open/close parens\n",
22, "| |", (int)(0 * 2 + 1), ""));
Perl_croak(aTHX_ "panic: reg returned NULL to re_op_compile for generation pass, flags=%#"UVxf"", (UV) flags);
}
DEBUG_OPTIMISE_r(
- PerlIO_printf(Perl_debug_log, "Starting post parse optimization\n");
+ Perl_re_printf( aTHX_ "Starting post parse optimization\n");
);
/* XXXX To minimize changes to RE engine we always allocate
copyRExC_state = RExC_state;
} else {
U32 seen=RExC_seen;
- DEBUG_OPTIMISE_r(PerlIO_printf(Perl_debug_log,"Restudying\n"));
+ DEBUG_OPTIMISE_r(Perl_re_printf( aTHX_ "Restudying\n"));
RExC_state = copyRExC_state;
if (seen & REG_TOP_LEVEL_BRANCHES_SEEN)
#ifdef TRIE_STUDY_OPT
DEBUG_PARSE_r(
if (!restudied)
- PerlIO_printf(Perl_debug_log, "first at %"IVdf"\n",
+ Perl_re_printf( aTHX_ "first at %"IVdf"\n",
(IV)(first - scan + 1))
);
#else
DEBUG_PARSE_r(
- PerlIO_printf(Perl_debug_log, "first at %"IVdf"\n",
+ Perl_re_printf( aTHX_ "first at %"IVdf"\n",
(IV)(first - scan + 1))
);
#endif
r->intflags &= ~PREGf_SKIP; /* Used in find_byclass(). */
DEBUG_COMPILE_r({ SV *sv = sv_newmortal();
regprop(r, sv, (regnode*)data.start_class, NULL, pRExC_state);
- PerlIO_printf(Perl_debug_log,
+ Perl_re_printf( aTHX_
"synthetic stclass \"%s\".\n",
SvPVX_const(sv));});
data.start_class = NULL;
regnode_ssc ch_class;
SSize_t last_close = 0;
- DEBUG_PARSE_r(PerlIO_printf(Perl_debug_log, "\nMulti Top Level\n"));
+ DEBUG_PARSE_r(Perl_re_printf( aTHX_ "\nMulti Top Level\n"));
scan = ri->program + 1;
ssc_init(pRExC_state, &ch_class);
r->intflags &= ~PREGf_SKIP; /* Used in find_byclass(). */
DEBUG_COMPILE_r({ SV* sv = sv_newmortal();
regprop(r, sv, (regnode*)data.start_class, NULL, pRExC_state);
- PerlIO_printf(Perl_debug_log,
+ Perl_re_printf( aTHX_
"synthetic stclass \"%s\".\n",
SvPVX_const(sv));});
data.start_class = NULL;
/* Guard against an embedded (?=) or (?<=) with a longer minlen than
the "real" pattern. */
DEBUG_OPTIMISE_r({
- PerlIO_printf(Perl_debug_log,"minlen: %"IVdf" r->minlen:%"IVdf" maxlen:%"IVdf"\n",
+ Perl_re_printf( aTHX_ "minlen: %"IVdf" r->minlen:%"IVdf" maxlen:%"IVdf"\n",
(IV)minlen, (IV)r->minlen, (IV)RExC_maxlen);
});
r->minlenret = minlen;
Newxz(r->offs, RExC_npar, regexp_paren_pair);
/* assume we don't need to swap parens around before we match */
DEBUG_TEST_r({
- PerlIO_printf(Perl_debug_log,"study_chunk_recursed_count: %lu\n",
+ Perl_re_printf( aTHX_ "study_chunk_recursed_count: %lu\n",
(unsigned long)RExC_study_chunk_recursed_count);
});
DEBUG_DUMP_r({
DEBUG_RExC_seen();
- PerlIO_printf(Perl_debug_log,"Final program:\n");
+ Perl_re_printf( aTHX_ "Final program:\n");
regdump(r);
});
#ifdef RE_TRACK_PATTERN_OFFSETS
const STRLEN len = ri->u.offsets[0];
STRLEN i;
GET_RE_DEBUG_FLAGS_DECL;
- PerlIO_printf(Perl_debug_log,
+ Perl_re_printf( aTHX_
"Offsets: [%"UVuf"]\n\t", (UV)ri->u.offsets[0]);
for (i = 1; i <= len; i++) {
if (ri->u.offsets[i*2-1] || ri->u.offsets[i*2])
- PerlIO_printf(Perl_debug_log, "%"UVuf":%"UVuf"[%"UVuf"] ",
+ Perl_re_printf( aTHX_ "%"UVuf":%"UVuf"[%"UVuf"] ",
(UV)i, (UV)ri->u.offsets[i*2-1], (UV)ri->u.offsets[i*2]);
}
- PerlIO_printf(Perl_debug_log, "\n");
+ Perl_re_printf( aTHX_ "\n");
});
#endif
#define DEBUG_PARSE_MSG(funcname) DEBUG_PARSE_r({ \
int num; \
if (RExC_lastparse!=RExC_parse) { \
- PerlIO_printf(Perl_debug_log, "%s", \
+ Perl_re_printf( aTHX_ "%s", \
Perl_pv_pretty(aTHX_ RExC_mysv1, RExC_parse, \
RExC_end - RExC_parse, 16, \
"", "", \
) \
); \
} else \
- PerlIO_printf(Perl_debug_log,"%16s",""); \
+ Perl_re_printf( aTHX_ "%16s",""); \
\
if (SIZE_ONLY) \
num = RExC_size + 1; \
else \
num=REG_NODE_NUM(RExC_emit); \
if (RExC_lastnum!=num) \
- PerlIO_printf(Perl_debug_log,"|%4d",num); \
+ Perl_re_printf( aTHX_ "|%4d",num); \
else \
- PerlIO_printf(Perl_debug_log,"|%4s",""); \
- PerlIO_printf(Perl_debug_log,"|%*s%-4s", \
+ Perl_re_printf( aTHX_ "|%4s",""); \
+ Perl_re_printf( aTHX_ "|%*s%-4s", \
(int)((depth*2)), "", \
(funcname) \
); \
#define DEBUG_PARSE(funcname) DEBUG_PARSE_r({ \
DEBUG_PARSE_MSG((funcname)); \
- PerlIO_printf(Perl_debug_log,"%4s","\n"); \
+ Perl_re_printf( aTHX_ "%4s","\n"); \
})
-#define DEBUG_PARSE_FMT(funcname,fmt,args) DEBUG_PARSE_r({ \
+#define DEBUG_PARSE_FMT(funcname,fmt,args) DEBUG_PARSE_r({\
DEBUG_PARSE_MSG((funcname)); \
- PerlIO_printf(Perl_debug_log,fmt "\n",args); \
+ Perl_re_printf( aTHX_ fmt "\n",args); \
})
/* This section of code defines the inversion list object and its methods. The
if (RExC_flags & RXf_PMf_FOLD) {
RExC_contains_i = 1;
}
- if (PASS2) {
- STD_PMMOD_FLAGS_PARSE_X_WARN(x_mod_count);
+
+ if (UNLIKELY((x_mod_count) > 1)) {
+ vFAIL("Only one /x regex modifier is allowed");
}
return;
/*NOTREACHED*/
vFAIL("Reference to nonexistent group");
}
RExC_recurse_count++;
- DEBUG_OPTIMISE_MORE_r(PerlIO_printf(Perl_debug_log,
+ DEBUG_OPTIMISE_MORE_r(Perl_re_printf( aTHX_
"%*s%*s Recurse #%"UVuf" to %"IVdf"\n",
22, "| |", (int)(depth * 2 + 1), "",
(UV)ARG(ret), (IV)ARG2L(ret)));
RExC_nestroot = parno;
if (RExC_open_parens && !RExC_open_parens[parno])
{
- DEBUG_OPTIMISE_MORE_r(PerlIO_printf(Perl_debug_log,
+ DEBUG_OPTIMISE_MORE_r(Perl_re_printf( aTHX_
"%*s%*s Setting open paren #%"IVdf" to %d\n",
22, "| |", (int)(depth * 2 + 1), "",
(IV)parno, REG_NODE_NUM(ret)));
case 1: case 2:
ender = reganode(pRExC_state, CLOSE, parno);
if ( RExC_close_parens ) {
- DEBUG_OPTIMISE_MORE_r(PerlIO_printf(Perl_debug_log,
+ DEBUG_OPTIMISE_MORE_r(Perl_re_printf( aTHX_
"%*s%*s Setting close paren #%"IVdf" to %d\n",
22, "| |", (int)(depth * 2 + 1), "", (IV)parno, REG_NODE_NUM(ender)));
RExC_close_parens[parno]= ender;
assert(!RExC_end_op); /* there can only be one! */
RExC_end_op = ender;
if (RExC_close_parens) {
- DEBUG_OPTIMISE_MORE_r(PerlIO_printf(Perl_debug_log,
+ DEBUG_OPTIMISE_MORE_r(Perl_re_printf( aTHX_
"%*s%*s Setting close paren #0 (END) to %d\n",
22, "| |", (int)(depth * 2 + 1), "", REG_NODE_NUM(ender)));
DEBUG_PARSE_MSG("lsbr");
regprop(RExC_rx, RExC_mysv1, lastbr, NULL, pRExC_state);
regprop(RExC_rx, RExC_mysv2, ender, NULL, pRExC_state);
- PerlIO_printf(Perl_debug_log, "~ tying lastbr %s (%"IVdf") to ender %s (%"IVdf") offset %"IVdf"\n",
+ Perl_re_printf( aTHX_ "~ tying lastbr %s (%"IVdf") to ender %s (%"IVdf") offset %"IVdf"\n",
SvPV_nolen_const(RExC_mysv1),
(IV)REG_NODE_NUM(lastbr),
SvPV_nolen_const(RExC_mysv2),
DEBUG_PARSE_MSG("NADA");
regprop(RExC_rx, RExC_mysv1, ret, NULL, pRExC_state);
regprop(RExC_rx, RExC_mysv2, ender, NULL, pRExC_state);
- PerlIO_printf(Perl_debug_log, "~ converting ret %s (%"IVdf") to ender %s (%"IVdf") offset %"IVdf"\n",
+ Perl_re_printf( aTHX_ "~ converting ret %s (%"IVdf") to ender %s (%"IVdf") offset %"IVdf"\n",
SvPV_nolen_const(RExC_mysv1),
(IV)REG_NODE_NUM(ret),
SvPV_nolen_const(RExC_mysv2),
} /* End of switch on '\' */
break;
case '{':
- /* Currently we don't warn when the lbrace is at the start
+ /* Currently we don't care if the lbrace is at the start
* of a construct. This catches it in the middle of a
* literal string, or when it's 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");
+ if (len || (p > RExC_start && isALPHA_A(*(p -1)))) {
+ RExC_parse = p + 1;
+ vFAIL("Unescaped left brace in regex is illegal");
}
/*FALLTHROUGH*/
default: /* A literal character */
'stack' of where the undealt-with left
parens would be if they were actually
put there */
- IV fence = 0; /* Position of where most recent undealt-
+ /* The 'VOL' (expanding to 'volatile') is a workaround for an optimiser bug
+ * in Solaris Studio 12.3. See RT #127455 */
+ VOL IV fence = 0; /* Position of where most recent undealt-
with left paren in stack is; -1 if none.
*/
STRLEN len; /* Temporary */
* stack, we have to check if it is a !. But first, the code above
* may have altered the stack in the time since we earlier set
* 'top_index'. */
+
top_index = av_tindex_nomg(stack);
if (top_index - fence >= 0) {
/* If the top entry on the stack is an operator, it had better
SAVEFREEPV(name);
if (FOLD) {
lookup_name = savepv(Perl_form(aTHX_ "__%s_i", name));
+
+ /* The function call just below that uses this can fail
+ * to return, leaking memory if we don't do this */
+ SAVEFREEPV(lookup_name);
}
/* Look up the property name, and get its swash and
NULL, /* No inversion list */
&swash_init_flags
);
- if (lookup_name) {
- Safefree(lookup_name);
- }
if (! swash || ! (invlist = _get_swash_invlist(swash))) {
HV* curpkg = (IN_PERL_COMPILETIME)
? PL_curstash
RExC_size += size;
return;
}
-
+ assert(!RExC_study_started); /* I believe we should never use reginsert once we have started
+ studying. If this is wrong then we need to adjust RExC_recurse
+ below like we do with RExC_open_parens/RExC_close_parens. */
src = RExC_emit;
RExC_emit += size;
dst = RExC_emit;
* iow it is 1 more than the number of parens seen in
* the pattern so far. */
for ( paren=0 ; paren < RExC_npar ; paren++ ) {
- if ( RExC_open_parens[paren] >= opnd ) {
+ /* note, RExC_open_parens[0] is the start of the
+ * regex, it can't move. RExC_close_parens[0] is the end
+ * of the regex, it *can* move. */
+ if ( paren && RExC_open_parens[paren] >= opnd ) {
/*DEBUG_PARSE_FMT("open"," - %d",size);*/
RExC_open_parens[paren] += size;
} else {
DEBUG_PARSE_r({
DEBUG_PARSE_MSG((scan==p ? "tail" : ""));
regprop(RExC_rx, RExC_mysv, scan, NULL, pRExC_state);
- PerlIO_printf(Perl_debug_log, "~ %s (%d) %s %s\n",
+ Perl_re_printf( aTHX_ "~ %s (%d) %s %s\n",
SvPV_nolen_const(RExC_mysv), REG_NODE_NUM(scan),
(temp == NULL ? "->" : ""),
(temp == NULL ? PL_reg_name[OP(val)] : "")
DEBUG_PARSE_r({
DEBUG_PARSE_MSG((scan==p ? "tsdy" : ""));
regprop(RExC_rx, RExC_mysv, scan, NULL, pRExC_state);
- PerlIO_printf(Perl_debug_log, "~ %s (%d) -> %s\n",
+ Perl_re_printf( aTHX_ "~ %s (%d) -> %s\n",
SvPV_nolen_const(RExC_mysv),
REG_NODE_NUM(scan),
PL_reg_name[exact]);
DEBUG_PARSE_r({
DEBUG_PARSE_MSG("");
regprop(RExC_rx, RExC_mysv, val, NULL, pRExC_state);
- PerlIO_printf(Perl_debug_log,
+ Perl_re_printf( aTHX_
"~ attach to %s (%"IVdf") offset to %"IVdf"\n",
SvPV_nolen_const(RExC_mysv),
(IV)REG_NODE_NUM(val),
for (bit=0; bit<REG_INTFLAGS_NAME_SIZE; bit++) {
if (flags & (1<<bit)) {
if (!set++ && lead)
- PerlIO_printf(Perl_debug_log, "%s",lead);
- PerlIO_printf(Perl_debug_log, "%s ",PL_reg_intflags_name[bit]);
+ Perl_re_printf( aTHX_ "%s",lead);
+ Perl_re_printf( aTHX_ "%s ",PL_reg_intflags_name[bit]);
}
}
if (lead) {
if (set)
- PerlIO_printf(Perl_debug_log, "\n");
+ Perl_re_printf( aTHX_ "\n");
else
- PerlIO_printf(Perl_debug_log, "%s[none-set]\n",lead);
+ Perl_re_printf( aTHX_ "%s[none-set]\n",lead);
}
}
continue;
}
if (!set++ && lead)
- PerlIO_printf(Perl_debug_log, "%s",lead);
- PerlIO_printf(Perl_debug_log, "%s ",PL_reg_extflags_name[bit]);
+ Perl_re_printf( aTHX_ "%s",lead);
+ Perl_re_printf( aTHX_ "%s ",PL_reg_extflags_name[bit]);
}
}
if ((cs = get_regex_charset(flags)) != REGEX_DEPENDS_CHARSET) {
if (!set++ && lead) {
- PerlIO_printf(Perl_debug_log, "%s",lead);
+ Perl_re_printf( aTHX_ "%s",lead);
}
switch (cs) {
case REGEX_UNICODE_CHARSET:
- PerlIO_printf(Perl_debug_log, "UNICODE");
+ Perl_re_printf( aTHX_ "UNICODE");
break;
case REGEX_LOCALE_CHARSET:
- PerlIO_printf(Perl_debug_log, "LOCALE");
+ Perl_re_printf( aTHX_ "LOCALE");
break;
case REGEX_ASCII_RESTRICTED_CHARSET:
- PerlIO_printf(Perl_debug_log, "ASCII-RESTRICTED");
+ Perl_re_printf( aTHX_ "ASCII-RESTRICTED");
break;
case REGEX_ASCII_MORE_RESTRICTED_CHARSET:
- PerlIO_printf(Perl_debug_log, "ASCII-MORE_RESTRICTED");
+ Perl_re_printf( aTHX_ "ASCII-MORE_RESTRICTED");
break;
default:
- PerlIO_printf(Perl_debug_log, "UNKNOWN CHARACTER SET");
+ Perl_re_printf( aTHX_ "UNKNOWN CHARACTER SET");
break;
}
}
if (lead) {
if (set)
- PerlIO_printf(Perl_debug_log, "\n");
+ Perl_re_printf( aTHX_ "\n");
else
- PerlIO_printf(Perl_debug_log, "%s[none-set]\n",lead);
+ Perl_re_printf( aTHX_ "%s[none-set]\n",lead);
}
}
#endif
if (r->anchored_substr) {
RE_PV_QUOTED_DECL(s, 0, dsv, SvPVX_const(r->anchored_substr),
RE_SV_DUMPLEN(r->anchored_substr), 30);
- PerlIO_printf(Perl_debug_log,
+ Perl_re_printf( aTHX_
"anchored %s%s at %"IVdf" ",
s, RE_SV_TAIL(r->anchored_substr),
(IV)r->anchored_offset);
} else if (r->anchored_utf8) {
RE_PV_QUOTED_DECL(s, 1, dsv, SvPVX_const(r->anchored_utf8),
RE_SV_DUMPLEN(r->anchored_utf8), 30);
- PerlIO_printf(Perl_debug_log,
+ Perl_re_printf( aTHX_
"anchored utf8 %s%s at %"IVdf" ",
s, RE_SV_TAIL(r->anchored_utf8),
(IV)r->anchored_offset);
if (r->float_substr) {
RE_PV_QUOTED_DECL(s, 0, dsv, SvPVX_const(r->float_substr),
RE_SV_DUMPLEN(r->float_substr), 30);
- PerlIO_printf(Perl_debug_log,
+ Perl_re_printf( aTHX_
"floating %s%s at %"IVdf"..%"UVuf" ",
s, RE_SV_TAIL(r->float_substr),
(IV)r->float_min_offset, (UV)r->float_max_offset);
} else if (r->float_utf8) {
RE_PV_QUOTED_DECL(s, 1, dsv, SvPVX_const(r->float_utf8),
RE_SV_DUMPLEN(r->float_utf8), 30);
- PerlIO_printf(Perl_debug_log,
+ Perl_re_printf( aTHX_
"floating utf8 %s%s at %"IVdf"..%"UVuf" ",
s, RE_SV_TAIL(r->float_utf8),
(IV)r->float_min_offset, (UV)r->float_max_offset);
}
if (r->check_substr || r->check_utf8)
- PerlIO_printf(Perl_debug_log,
+ Perl_re_printf( aTHX_
(const char *)
(r->check_substr == r->float_substr
&& r->check_utf8 == r->float_utf8
? "(checking floating" : "(checking anchored"));
if (r->intflags & PREGf_NOSCAN)
- PerlIO_printf(Perl_debug_log, " noscan");
+ Perl_re_printf( aTHX_ " noscan");
if (r->extflags & RXf_CHECK_ALL)
- PerlIO_printf(Perl_debug_log, " isall");
+ Perl_re_printf( aTHX_ " isall");
if (r->check_substr || r->check_utf8)
- PerlIO_printf(Perl_debug_log, ") ");
+ Perl_re_printf( aTHX_ ") ");
if (ri->regstclass) {
regprop(r, sv, ri->regstclass, NULL, NULL);
- PerlIO_printf(Perl_debug_log, "stclass %s ", SvPVX_const(sv));
+ Perl_re_printf( aTHX_ "stclass %s ", SvPVX_const(sv));
}
if (r->intflags & PREGf_ANCH) {
- PerlIO_printf(Perl_debug_log, "anchored");
+ Perl_re_printf( aTHX_ "anchored");
if (r->intflags & PREGf_ANCH_MBOL)
- PerlIO_printf(Perl_debug_log, "(MBOL)");
+ Perl_re_printf( aTHX_ "(MBOL)");
if (r->intflags & PREGf_ANCH_SBOL)
- PerlIO_printf(Perl_debug_log, "(SBOL)");
+ Perl_re_printf( aTHX_ "(SBOL)");
if (r->intflags & PREGf_ANCH_GPOS)
- PerlIO_printf(Perl_debug_log, "(GPOS)");
- (void)PerlIO_putc(Perl_debug_log, ' ');
+ Perl_re_printf( aTHX_ "(GPOS)");
+ Perl_re_printf( aTHX_ " ");
}
if (r->intflags & PREGf_GPOS_SEEN)
- PerlIO_printf(Perl_debug_log, "GPOS:%"UVuf" ", (UV)r->gofs);
+ Perl_re_printf( aTHX_ "GPOS:%"UVuf" ", (UV)r->gofs);
if (r->intflags & PREGf_SKIP)
- PerlIO_printf(Perl_debug_log, "plus ");
+ Perl_re_printf( aTHX_ "plus ");
if (r->intflags & PREGf_IMPLICIT)
- PerlIO_printf(Perl_debug_log, "implicit ");
- PerlIO_printf(Perl_debug_log, "minlen %"IVdf" ", (IV)r->minlen);
+ Perl_re_printf( aTHX_ "implicit ");
+ Perl_re_printf( aTHX_ "minlen %"IVdf" ", (IV)r->minlen);
if (r->extflags & RXf_EVAL_SEEN)
- PerlIO_printf(Perl_debug_log, "with eval ");
- PerlIO_printf(Perl_debug_log, "\n");
+ Perl_re_printf( aTHX_ "with eval ");
+ Perl_re_printf( aTHX_ "\n");
DEBUG_FLAGS_r({
regdump_extflags("r->extflags: ",r->extflags);
regdump_intflags("r->intflags: ",r->intflags);
? prog->check_utf8 : prog->check_substr);
if (!PL_colorset) reginitcolors();
- PerlIO_printf(Perl_debug_log,
+ Perl_re_printf( aTHX_
"%sUsing REx %ssubstr:%s \"%s%.60s%s%s\"\n",
PL_colors[4],
RX_UTF8(r) ? "utf8 " : "",
SV *dsv= sv_newmortal();
RE_PV_QUOTED_DECL(s, RX_UTF8(rx),
dsv, RX_PRECOMP(rx), RX_PRELEN(rx), 60);
- PerlIO_printf(Perl_debug_log,"%sFreeing REx:%s %s\n",
+ Perl_re_printf( aTHX_ "%sFreeing REx:%s %s\n",
PL_colors[4],PL_colors[5],s);
}
});
/* And this flag for matching all non-ASCII 0xFF and below */
if (flags & ANYOF_SHARED_d_MATCHES_ALL_NON_UTF8_NON_ASCII_non_d_WARN_SUPER)
{
- if (invert) {
- not_utf8 = _new_invlist(0);
- }
- else {
- not_utf8 = invlist_clone(PL_UpperLatin1);
- }
- inverting_allowed = FALSE; /* XXX needs more work to be able
- to allow this */
+ not_utf8 = invlist_clone(PL_UpperLatin1);
}
}
else if (OP(node) == ANYOFL) {
* conditional code points, so that when inverted, they will be gone
* from it */
_invlist_union(only_utf8, invlist, &invlist);
+ _invlist_union(not_utf8, invlist, &invlist);
_invlist_union(only_utf8_locale, invlist, &invlist);
_invlist_invert(invlist);
_invlist_intersection(invlist, PL_InBitmap, &invlist);
return SvCUR(sv) > orig_sv_cur;
}
-#define CLEAR_OPTSTART \
+#define CLEAR_OPTSTART \
if (optstart) STMT_START { \
- DEBUG_OPTIMISE_r(PerlIO_printf(Perl_debug_log, \
+ DEBUG_OPTIMISE_r(Perl_re_printf( aTHX_ \
" (%"IVdf" nodes)\n", (IV)(node - optstart))); \
- optstart=NULL; \
+ optstart=NULL; \
} STMT_END
#define DUMPUNTIL(b,e) \
PERL_ARGS_ASSERT_DUMPUNTIL;
#ifdef DEBUG_DUMPUNTIL
- PerlIO_printf(Perl_debug_log, "--- %d : %d - %d - %d\n",indent,node-start,
+ Perl_re_printf( aTHX_ "--- %d : %d - %d - %d\n",indent,node-start,
last ? last-start : 0,plast ? plast-start : 0);
#endif
CLEAR_OPTSTART;
regprop(r, sv, node, NULL, NULL);
- PerlIO_printf(Perl_debug_log, "%4"IVdf":%*s%s", (IV)(node - start),
+ Perl_re_printf( aTHX_ "%4"IVdf":%*s%s", (IV)(node - start),
(int)(2*indent + 1), "", SvPVX_const(sv));
if (OP(node) != OPTIMIZED) {
if (next == NULL) /* Next ptr. */
- PerlIO_printf(Perl_debug_log, " (0)");
+ Perl_re_printf( aTHX_ " (0)");
else if (PL_regkind[(U8)op] == BRANCH
&& PL_regkind[OP(next)] != BRANCH )
- PerlIO_printf(Perl_debug_log, " (FAIL)");
+ Perl_re_printf( aTHX_ " (FAIL)");
else
- PerlIO_printf(Perl_debug_log, " (%"IVdf")", (IV)(next - start));
- (void)PerlIO_putc(Perl_debug_log, '\n');
+ Perl_re_printf( aTHX_ " (%"IVdf")", (IV)(next - start));
+ Perl_re_printf( aTHX_ "\n");
}
after_print:
for (word_idx= 0; word_idx < (I32)trie->wordcount; word_idx++) {
SV ** const elem_ptr = av_fetch(trie_words,word_idx,0);
- PerlIO_printf(Perl_debug_log, "%*s%s ",
- (int)(2*(indent+3)), "",
+ Perl_re_indentf( aTHX_ "%s ",
+ indent+3,
elem_ptr
? pv_pretty(sv, SvPV_nolen_const(*elem_ptr),
SvCUR(*elem_ptr), 60,
);
if (trie->jump) {
U16 dist= trie->jump[word_idx+1];
- PerlIO_printf(Perl_debug_log, "(%"UVuf")\n",
+ Perl_re_printf( aTHX_ "(%"UVuf")\n",
(UV)((dist ? this_trie + dist : next) - start));
if (dist) {
if (!nextbranch)
if (nextbranch && PL_regkind[OP(nextbranch)]==BRANCH)
nextbranch= regnext((regnode *)nextbranch);
} else {
- PerlIO_printf(Perl_debug_log, "\n");
+ Perl_re_printf( aTHX_ "\n");
}
}
if (last && next > last)
}
CLEAR_OPTSTART;
#ifdef DEBUG_DUMPUNTIL
- PerlIO_printf(Perl_debug_log, "--- %d\n", (int)indent);
+ Perl_re_printf( aTHX_ "--- %d\n", (int)indent);
#endif
return node;
}
https://perl5.git.perl.org / perl5.git / blobdiff