6 |
and semantics are as close as possible to those of the Perl 5 language. |
and semantics are as close as possible to those of the Perl 5 language. |
7 |
|
|
8 |
Written by Philip Hazel |
Written by Philip Hazel |
9 |
Copyright (c) 1997-2005 University of Cambridge |
Copyright (c) 1997-2006 University of Cambridge |
10 |
|
|
11 |
----------------------------------------------------------------------------- |
----------------------------------------------------------------------------- |
12 |
Redistribution and use in source and binary forms, with or without |
Redistribution and use in source and binary forms, with or without |
39 |
|
|
40 |
|
|
41 |
/* This module contains the external function pcre_dfa_exec(), which is an |
/* This module contains the external function pcre_dfa_exec(), which is an |
42 |
alternative matching function that uses a DFA algorithm. This is NOT Perl- |
alternative matching function that uses a sort of DFA algorithm (not a true |
43 |
compatible, but it has advantages in certain applications. */ |
FSM). This is NOT Perl- compatible, but it has advantages in certain |
44 |
|
applications. */ |
45 |
|
|
46 |
|
|
47 |
|
#define NLBLOCK md /* Block containing newline information */ |
48 |
|
#define PSSTART start_subject /* Field containing processed string start */ |
49 |
|
#define PSEND end_subject /* Field containing processed string end */ |
50 |
|
|
51 |
#include "pcre_internal.h" |
#include "pcre_internal.h" |
52 |
|
|
53 |
|
|
62 |
*************************************************/ |
*************************************************/ |
63 |
|
|
64 |
/* These are offsets that are used to turn the OP_TYPESTAR and friends opcodes |
/* These are offsets that are used to turn the OP_TYPESTAR and friends opcodes |
65 |
into others, under special conditions. A gap of 10 between the blocks should be |
into others, under special conditions. A gap of 20 between the blocks should be |
66 |
enough. */ |
enough. */ |
67 |
|
|
68 |
#define OP_PROP_EXTRA (EXTRACT_BASIC_MAX+1) |
#define OP_PROP_EXTRA 100 |
69 |
#define OP_EXTUNI_EXTRA (EXTRACT_BASIC_MAX+11) |
#define OP_EXTUNI_EXTRA 120 |
70 |
|
#define OP_ANYNL_EXTRA 140 |
71 |
|
|
72 |
|
|
73 |
/* This table identifies those opcodes that are followed immediately by a |
/* This table identifies those opcodes that are followed immediately by a |
80 |
0, /* End */ |
0, /* End */ |
81 |
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, /* \A, \G, \B, \b, \D, \d, \S, \s, \W, \w */ |
0, 0, 0, 0, 0, 0, 0, 0, 0, 0, /* \A, \G, \B, \b, \D, \d, \S, \s, \W, \w */ |
82 |
0, 0, /* Any, Anybyte */ |
0, 0, /* Any, Anybyte */ |
83 |
0, 0, 0, /* NOTPROP, PROP, EXTUNI */ |
0, 0, 0, 0, /* NOTPROP, PROP, EXTUNI, ANYNL */ |
84 |
0, 0, 0, 0, 0, /* \Z, \z, Opt, ^, $ */ |
0, 0, 0, 0, 0, /* \Z, \z, Opt, ^, $ */ |
85 |
1, /* Char */ |
1, /* Char */ |
86 |
1, /* Charnc */ |
1, /* Charnc */ |
88 |
/* Positive single-char repeats */ |
/* Positive single-char repeats */ |
89 |
1, 1, 1, 1, 1, 1, /* *, *?, +, +?, ?, ?? */ |
1, 1, 1, 1, 1, 1, /* *, *?, +, +?, ?, ?? */ |
90 |
3, 3, 3, /* upto, minupto, exact */ |
3, 3, 3, /* upto, minupto, exact */ |
91 |
|
1, 1, 1, 3, /* *+, ++, ?+, upto+ */ |
92 |
/* Negative single-char repeats - only for chars < 256 */ |
/* Negative single-char repeats - only for chars < 256 */ |
93 |
1, 1, 1, 1, 1, 1, /* NOT *, *?, +, +?, ?, ?? */ |
1, 1, 1, 1, 1, 1, /* NOT *, *?, +, +?, ?, ?? */ |
94 |
3, 3, 3, /* NOT upto, minupto, exact */ |
3, 3, 3, /* NOT upto, minupto, exact */ |
95 |
|
1, 1, 1, 3, /* NOT *+, ++, ?+, updo+ */ |
96 |
/* Positive type repeats */ |
/* Positive type repeats */ |
97 |
1, 1, 1, 1, 1, 1, /* Type *, *?, +, +?, ?, ?? */ |
1, 1, 1, 1, 1, 1, /* Type *, *?, +, +?, ?, ?? */ |
98 |
3, 3, 3, /* Type upto, minupto, exact */ |
3, 3, 3, /* Type upto, minupto, exact */ |
99 |
|
1, 1, 1, 3, /* Type *+, ++, ?+, upto+ */ |
100 |
/* Character class & ref repeats */ |
/* Character class & ref repeats */ |
101 |
0, 0, 0, 0, 0, 0, /* *, *?, +, +?, ?, ?? */ |
0, 0, 0, 0, 0, 0, /* *, *?, +, +?, ?, ?? */ |
102 |
0, 0, /* CRRANGE, CRMINRANGE */ |
0, 0, /* CRRANGE, CRMINRANGE */ |
115 |
0, /* Assert behind */ |
0, /* Assert behind */ |
116 |
0, /* Assert behind not */ |
0, /* Assert behind not */ |
117 |
0, /* Reverse */ |
0, /* Reverse */ |
118 |
0, /* Once */ |
0, 0, 0, 0, /* ONCE, BRA, CBRA, COND */ |
119 |
0, /* COND */ |
0, 0, 0, /* SBRA, SCBRA, SCOND */ |
120 |
0, /* CREF */ |
0, /* CREF */ |
121 |
0, 0, /* BRAZERO, BRAMINZERO */ |
0, /* RREF */ |
122 |
0, /* BRANUMBER */ |
0, /* DEF */ |
123 |
0 /* BRA */ |
0, 0 /* BRAZERO, BRAMINZERO */ |
124 |
}; |
}; |
125 |
|
|
126 |
/* These 2 tables allow for compact code for testing for \D, \d, \S, \s, \W, |
/* These 2 tables allow for compact code for testing for \D, \d, \S, \s, \W, |
286 |
|
|
287 |
const uschar *ctypes, *lcc, *fcc; |
const uschar *ctypes, *lcc, *fcc; |
288 |
const uschar *ptr; |
const uschar *ptr; |
289 |
const uschar *end_code; |
const uschar *end_code, *first_op; |
290 |
|
|
291 |
int active_count, new_count, match_count; |
int active_count, new_count, match_count; |
292 |
|
|
297 |
const uschar *end_subject = md->end_subject; |
const uschar *end_subject = md->end_subject; |
298 |
const uschar *start_code = md->start_code; |
const uschar *start_code = md->start_code; |
299 |
|
|
300 |
|
#ifdef SUPPORT_UTF8 |
301 |
BOOL utf8 = (md->poptions & PCRE_UTF8) != 0; |
BOOL utf8 = (md->poptions & PCRE_UTF8) != 0; |
302 |
|
#else |
303 |
|
BOOL utf8 = FALSE; |
304 |
|
#endif |
305 |
|
|
306 |
rlevel++; |
rlevel++; |
307 |
offsetcount &= (-2); |
offsetcount &= (-2); |
324 |
next_new_state = new_states = active_states + wscount; |
next_new_state = new_states = active_states + wscount; |
325 |
new_count = 0; |
new_count = 0; |
326 |
|
|
327 |
|
first_op = this_start_code + 1 + LINK_SIZE + |
328 |
|
((*this_start_code == OP_CBRA || *this_start_code == OP_SCBRA)? 2:0); |
329 |
|
|
330 |
/* The first thing in any (sub) pattern is a bracket of some sort. Push all |
/* The first thing in any (sub) pattern is a bracket of some sort. Push all |
331 |
the alternative states onto the list, and find out where the end is. This |
the alternative states onto the list, and find out where the end is. This |
332 |
makes is possible to use this function recursively, when we want to stop at a |
makes is possible to use this function recursively, when we want to stop at a |
336 |
a backward assertion. In that case, we have to find out the maximum amount to |
a backward assertion. In that case, we have to find out the maximum amount to |
337 |
move back, and set up each alternative appropriately. */ |
move back, and set up each alternative appropriately. */ |
338 |
|
|
339 |
if (this_start_code[1+LINK_SIZE] == OP_REVERSE) |
if (*first_op == OP_REVERSE) |
340 |
{ |
{ |
341 |
int max_back = 0; |
int max_back = 0; |
342 |
int gone_back; |
int gone_back; |
418 |
|
|
419 |
else |
else |
420 |
{ |
{ |
421 |
|
int length = 1 + LINK_SIZE + |
422 |
|
((*this_start_code == OP_CBRA || *this_start_code == OP_SCBRA)? 2:0); |
423 |
do |
do |
424 |
{ |
{ |
425 |
ADD_NEW(end_code - start_code + 1 + LINK_SIZE, 0); |
ADD_NEW(end_code - start_code + length, 0); |
426 |
end_code += GET(end_code, 1); |
end_code += GET(end_code, 1); |
427 |
|
length = 1 + LINK_SIZE; |
428 |
} |
} |
429 |
while (*end_code == OP_ALT); |
while (*end_code == OP_ALT); |
430 |
} |
} |
440 |
for (;;) |
for (;;) |
441 |
{ |
{ |
442 |
int i, j; |
int i, j; |
443 |
int c, d, clen, dlen; |
int clen, dlen; |
444 |
|
unsigned int c, d; |
445 |
|
|
446 |
/* Make the new state list into the active state list and empty the |
/* Make the new state list into the active state list and empty the |
447 |
new state list. */ |
new state list. */ |
477 |
|
|
478 |
if (ptr < end_subject) |
if (ptr < end_subject) |
479 |
{ |
{ |
480 |
clen = 1; |
clen = 1; /* Number of bytes in the character */ |
481 |
#ifdef SUPPORT_UTF8 |
#ifdef SUPPORT_UTF8 |
482 |
if (utf8) { GETCHARLEN(c, ptr, clen); } else |
if (utf8) { GETCHARLEN(c, ptr, clen); } else |
483 |
#endif /* SUPPORT_UTF8 */ |
#endif /* SUPPORT_UTF8 */ |
485 |
} |
} |
486 |
else |
else |
487 |
{ |
{ |
488 |
clen = 0; /* At end subject */ |
clen = 0; /* This indicates the end of the subject */ |
489 |
c = -1; |
c = NOTACHAR; /* This value should never actually be used */ |
490 |
} |
} |
491 |
|
|
492 |
/* Scan up the active states and act on each one. The result of an action |
/* Scan up the active states and act on each one. The result of an action |
500 |
const uschar *code; |
const uschar *code; |
501 |
int state_offset = current_state->offset; |
int state_offset = current_state->offset; |
502 |
int count, codevalue; |
int count, codevalue; |
503 |
int chartype, othercase; |
int chartype, script; |
504 |
|
|
505 |
#ifdef DEBUG |
#ifdef DEBUG |
506 |
printf ("%.*sProcessing state %d c=", rlevel*2-2, SP, state_offset); |
printf ("%.*sProcessing state %d c=", rlevel*2-2, SP, state_offset); |
507 |
if (c < 0) printf("-1\n"); |
if (clen == 0) printf("EOL\n"); |
508 |
else if (c > 32 && c < 127) printf("'%c'\n", c); |
else if (c > 32 && c < 127) printf("'%c'\n", c); |
509 |
else printf("0x%02x\n", c); |
else printf("0x%02x\n", c); |
510 |
#endif |
#endif |
548 |
|
|
549 |
code = start_code + state_offset; |
code = start_code + state_offset; |
550 |
codevalue = *code; |
codevalue = *code; |
|
if (codevalue >= OP_BRA) codevalue = OP_BRA; /* All brackets are equal */ |
|
551 |
|
|
552 |
/* If this opcode is followed by an inline character, load it. It is |
/* If this opcode is followed by an inline character, load it. It is |
553 |
tempting to test for the presence of a subject character here, but that |
tempting to test for the presence of a subject character here, but that |
569 |
d = code[coptable[codevalue]]; |
d = code[coptable[codevalue]]; |
570 |
if (codevalue >= OP_TYPESTAR) |
if (codevalue >= OP_TYPESTAR) |
571 |
{ |
{ |
572 |
if (d == OP_ANYBYTE) return PCRE_ERROR_DFA_UITEM; |
switch(d) |
573 |
if (d >= OP_NOTPROP) |
{ |
574 |
codevalue += (d == OP_EXTUNI)? OP_EXTUNI_EXTRA : OP_PROP_EXTRA; |
case OP_ANYBYTE: return PCRE_ERROR_DFA_UITEM; |
575 |
|
case OP_NOTPROP: |
576 |
|
case OP_PROP: codevalue += OP_PROP_EXTRA; break; |
577 |
|
case OP_ANYNL: codevalue += OP_ANYNL_EXTRA; break; |
578 |
|
case OP_EXTUNI: codevalue += OP_EXTUNI_EXTRA; break; |
579 |
|
default: break; |
580 |
|
} |
581 |
} |
} |
582 |
} |
} |
583 |
else |
else |
584 |
{ |
{ |
585 |
dlen = 0; /* Not strictly necessary, but compilers moan */ |
dlen = 0; /* Not strictly necessary, but compilers moan */ |
586 |
d = -1; /* if these variables are not set. */ |
d = NOTACHAR; /* if these variables are not set. */ |
587 |
} |
} |
588 |
|
|
589 |
|
|
645 |
|
|
646 |
/*-----------------------------------------------------------------*/ |
/*-----------------------------------------------------------------*/ |
647 |
case OP_BRA: |
case OP_BRA: |
648 |
|
case OP_SBRA: |
649 |
do |
do |
650 |
{ |
{ |
651 |
ADD_ACTIVE(code - start_code + 1 + LINK_SIZE, 0); |
ADD_ACTIVE(code - start_code + 1 + LINK_SIZE, 0); |
655 |
break; |
break; |
656 |
|
|
657 |
/*-----------------------------------------------------------------*/ |
/*-----------------------------------------------------------------*/ |
658 |
|
case OP_CBRA: |
659 |
|
case OP_SCBRA: |
660 |
|
ADD_ACTIVE(code - start_code + 3 + LINK_SIZE, 0); |
661 |
|
code += GET(code, 1); |
662 |
|
while (*code == OP_ALT) |
663 |
|
{ |
664 |
|
ADD_ACTIVE(code - start_code + 1 + LINK_SIZE, 0); |
665 |
|
code += GET(code, 1); |
666 |
|
} |
667 |
|
break; |
668 |
|
|
669 |
|
/*-----------------------------------------------------------------*/ |
670 |
case OP_BRAZERO: |
case OP_BRAZERO: |
671 |
case OP_BRAMINZERO: |
case OP_BRAMINZERO: |
672 |
ADD_ACTIVE(state_offset + 1, 0); |
ADD_ACTIVE(state_offset + 1, 0); |
676 |
break; |
break; |
677 |
|
|
678 |
/*-----------------------------------------------------------------*/ |
/*-----------------------------------------------------------------*/ |
|
case OP_BRANUMBER: |
|
|
ADD_ACTIVE(state_offset + 1 + LINK_SIZE, 0); |
|
|
break; |
|
|
|
|
|
/*-----------------------------------------------------------------*/ |
|
679 |
case OP_CIRC: |
case OP_CIRC: |
680 |
if ((ptr == start_subject && (md->moptions & PCRE_NOTBOL) == 0) || |
if ((ptr == start_subject && (md->moptions & PCRE_NOTBOL) == 0) || |
681 |
((ims & PCRE_MULTILINE) != 0 && ptr[-1] == NEWLINE)) |
((ims & PCRE_MULTILINE) != 0 && |
682 |
|
ptr != end_subject && |
683 |
|
WAS_NEWLINE(ptr))) |
684 |
{ ADD_ACTIVE(state_offset + 1, 0); } |
{ ADD_ACTIVE(state_offset + 1, 0); } |
685 |
break; |
break; |
686 |
|
|
714 |
|
|
715 |
/*-----------------------------------------------------------------*/ |
/*-----------------------------------------------------------------*/ |
716 |
case OP_ANY: |
case OP_ANY: |
717 |
if (clen > 0 && (c != NEWLINE || (ims & PCRE_DOTALL) != 0)) |
if (clen > 0 && ((ims & PCRE_DOTALL) != 0 || !IS_NEWLINE(ptr))) |
718 |
{ ADD_NEW(state_offset + 1, 0); } |
{ ADD_NEW(state_offset + 1, 0); } |
719 |
break; |
break; |
720 |
|
|
721 |
/*-----------------------------------------------------------------*/ |
/*-----------------------------------------------------------------*/ |
722 |
case OP_EODN: |
case OP_EODN: |
723 |
if (clen == 0 || (c == NEWLINE && ptr + 1 == end_subject)) |
if (clen == 0 || (IS_NEWLINE(ptr) && ptr == end_subject - md->nllen)) |
724 |
{ ADD_ACTIVE(state_offset + 1, 0); } |
{ ADD_ACTIVE(state_offset + 1, 0); } |
725 |
break; |
break; |
726 |
|
|
728 |
case OP_DOLL: |
case OP_DOLL: |
729 |
if ((md->moptions & PCRE_NOTEOL) == 0) |
if ((md->moptions & PCRE_NOTEOL) == 0) |
730 |
{ |
{ |
731 |
if (clen == 0 || (c == NEWLINE && (ptr + 1 == end_subject || |
if (clen == 0 || |
732 |
(ims & PCRE_MULTILINE) != 0))) |
(IS_NEWLINE(ptr) && |
733 |
|
((ims & PCRE_MULTILINE) != 0 || ptr == end_subject - md->nllen) |
734 |
|
)) |
735 |
{ ADD_ACTIVE(state_offset + 1, 0); } |
{ ADD_ACTIVE(state_offset + 1, 0); } |
736 |
} |
} |
737 |
else if (c == NEWLINE && (ims & PCRE_MULTILINE) != 0) |
else if ((ims & PCRE_MULTILINE) != 0 && IS_NEWLINE(ptr)) |
738 |
{ ADD_ACTIVE(state_offset + 1, 0); } |
{ ADD_ACTIVE(state_offset + 1, 0); } |
739 |
break; |
break; |
740 |
|
|
794 |
case OP_NOTPROP: |
case OP_NOTPROP: |
795 |
if (clen > 0) |
if (clen > 0) |
796 |
{ |
{ |
797 |
int rqdtype, category; |
BOOL OK; |
798 |
category = ucp_findchar(c, &chartype, &othercase); |
int category = _pcre_ucp_findprop(c, &chartype, &script); |
799 |
rqdtype = code[1]; |
switch(code[1]) |
|
if (rqdtype >= 128) |
|
|
{ |
|
|
if ((rqdtype - 128 == category) == (codevalue == OP_PROP)) |
|
|
{ ADD_NEW(state_offset + 2, 0); } |
|
|
} |
|
|
else |
|
800 |
{ |
{ |
801 |
if ((rqdtype == chartype) == (codevalue == OP_PROP)) |
case PT_ANY: |
802 |
{ ADD_NEW(state_offset + 2, 0); } |
OK = TRUE; |
803 |
|
break; |
804 |
|
|
805 |
|
case PT_LAMP: |
806 |
|
OK = chartype == ucp_Lu || chartype == ucp_Ll || chartype == ucp_Lt; |
807 |
|
break; |
808 |
|
|
809 |
|
case PT_GC: |
810 |
|
OK = category == code[2]; |
811 |
|
break; |
812 |
|
|
813 |
|
case PT_PC: |
814 |
|
OK = chartype == code[2]; |
815 |
|
break; |
816 |
|
|
817 |
|
case PT_SC: |
818 |
|
OK = script == code[2]; |
819 |
|
break; |
820 |
|
|
821 |
|
/* Should never occur, but keep compilers from grumbling. */ |
822 |
|
|
823 |
|
default: |
824 |
|
OK = codevalue != OP_PROP; |
825 |
|
break; |
826 |
} |
} |
827 |
|
|
828 |
|
if (OK == (codevalue == OP_PROP)) { ADD_NEW(state_offset + 3, 0); } |
829 |
} |
} |
830 |
break; |
break; |
831 |
#endif |
#endif |
840 |
|
|
841 |
case OP_TYPEPLUS: |
case OP_TYPEPLUS: |
842 |
case OP_TYPEMINPLUS: |
case OP_TYPEMINPLUS: |
843 |
|
case OP_TYPEPOSPLUS: |
844 |
count = current_state->count; /* Already matched */ |
count = current_state->count; /* Already matched */ |
845 |
if (count > 0) { ADD_ACTIVE(state_offset + 2, 0); } |
if (count > 0) { ADD_ACTIVE(state_offset + 2, 0); } |
846 |
if (clen > 0) |
if (clen > 0) |
847 |
{ |
{ |
848 |
if ((c >= 256 && d != OP_DIGIT && d != OP_WHITESPACE && d != OP_WORDCHAR) || |
if ((c >= 256 && d != OP_DIGIT && d != OP_WHITESPACE && d != OP_WORDCHAR) || |
849 |
(c < 256 && |
(c < 256 && |
850 |
(d != OP_ANY || c != '\n' || (ims & PCRE_DOTALL) != 0) && |
(d != OP_ANY || |
851 |
|
(ims & PCRE_DOTALL) != 0 || |
852 |
|
!IS_NEWLINE(ptr) |
853 |
|
) && |
854 |
((ctypes[c] & toptable1[d]) ^ toptable2[d]) != 0)) |
((ctypes[c] & toptable1[d]) ^ toptable2[d]) != 0)) |
855 |
{ |
{ |
856 |
|
if (count > 0 && codevalue == OP_TYPEPOSPLUS) |
857 |
|
{ |
858 |
|
active_count--; /* Remove non-match possibility */ |
859 |
|
next_active_state--; |
860 |
|
} |
861 |
count++; |
count++; |
862 |
ADD_NEW(state_offset, count); |
ADD_NEW(state_offset, count); |
863 |
} |
} |
867 |
/*-----------------------------------------------------------------*/ |
/*-----------------------------------------------------------------*/ |
868 |
case OP_TYPEQUERY: |
case OP_TYPEQUERY: |
869 |
case OP_TYPEMINQUERY: |
case OP_TYPEMINQUERY: |
870 |
|
case OP_TYPEPOSQUERY: |
871 |
ADD_ACTIVE(state_offset + 2, 0); |
ADD_ACTIVE(state_offset + 2, 0); |
872 |
if (clen > 0) |
if (clen > 0) |
873 |
{ |
{ |
874 |
if ((c >= 256 && d != OP_DIGIT && d != OP_WHITESPACE && d != OP_WORDCHAR) || |
if ((c >= 256 && d != OP_DIGIT && d != OP_WHITESPACE && d != OP_WORDCHAR) || |
875 |
(c < 256 && |
(c < 256 && |
876 |
(d != OP_ANY || c != '\n' || (ims & PCRE_DOTALL) != 0) && |
(d != OP_ANY || |
877 |
|
(ims & PCRE_DOTALL) != 0 || |
878 |
|
!IS_NEWLINE(ptr) |
879 |
|
) && |
880 |
((ctypes[c] & toptable1[d]) ^ toptable2[d]) != 0)) |
((ctypes[c] & toptable1[d]) ^ toptable2[d]) != 0)) |
881 |
{ |
{ |
882 |
|
if (codevalue == OP_TYPEPOSQUERY) |
883 |
|
{ |
884 |
|
active_count--; /* Remove non-match possibility */ |
885 |
|
next_active_state--; |
886 |
|
} |
887 |
ADD_NEW(state_offset + 2, 0); |
ADD_NEW(state_offset + 2, 0); |
888 |
} |
} |
889 |
} |
} |
892 |
/*-----------------------------------------------------------------*/ |
/*-----------------------------------------------------------------*/ |
893 |
case OP_TYPESTAR: |
case OP_TYPESTAR: |
894 |
case OP_TYPEMINSTAR: |
case OP_TYPEMINSTAR: |
895 |
|
case OP_TYPEPOSSTAR: |
896 |
ADD_ACTIVE(state_offset + 2, 0); |
ADD_ACTIVE(state_offset + 2, 0); |
897 |
if (clen > 0) |
if (clen > 0) |
898 |
{ |
{ |
899 |
if ((c >= 256 && d != OP_DIGIT && d != OP_WHITESPACE && d != OP_WORDCHAR) || |
if ((c >= 256 && d != OP_DIGIT && d != OP_WHITESPACE && d != OP_WORDCHAR) || |
900 |
(c < 256 && |
(c < 256 && |
901 |
(d != OP_ANY || c != '\n' || (ims & PCRE_DOTALL) != 0) && |
(d != OP_ANY || |
902 |
|
(ims & PCRE_DOTALL) != 0 || |
903 |
|
!IS_NEWLINE(ptr) |
904 |
|
) && |
905 |
((ctypes[c] & toptable1[d]) ^ toptable2[d]) != 0)) |
((ctypes[c] & toptable1[d]) ^ toptable2[d]) != 0)) |
906 |
{ |
{ |
907 |
|
if (codevalue == OP_TYPEPOSSTAR) |
908 |
|
{ |
909 |
|
active_count--; /* Remove non-match possibility */ |
910 |
|
next_active_state--; |
911 |
|
} |
912 |
ADD_NEW(state_offset, 0); |
ADD_NEW(state_offset, 0); |
913 |
} |
} |
914 |
} |
} |
916 |
|
|
917 |
/*-----------------------------------------------------------------*/ |
/*-----------------------------------------------------------------*/ |
918 |
case OP_TYPEEXACT: |
case OP_TYPEEXACT: |
919 |
|
count = current_state->count; /* Number already matched */ |
920 |
|
if (clen > 0) |
921 |
|
{ |
922 |
|
if ((c >= 256 && d != OP_DIGIT && d != OP_WHITESPACE && d != OP_WORDCHAR) || |
923 |
|
(c < 256 && |
924 |
|
(d != OP_ANY || |
925 |
|
(ims & PCRE_DOTALL) != 0 || |
926 |
|
!IS_NEWLINE(ptr) |
927 |
|
) && |
928 |
|
((ctypes[c] & toptable1[d]) ^ toptable2[d]) != 0)) |
929 |
|
{ |
930 |
|
if (++count >= GET2(code, 1)) |
931 |
|
{ ADD_NEW(state_offset + 4, 0); } |
932 |
|
else |
933 |
|
{ ADD_NEW(state_offset, count); } |
934 |
|
} |
935 |
|
} |
936 |
|
break; |
937 |
|
|
938 |
|
/*-----------------------------------------------------------------*/ |
939 |
case OP_TYPEUPTO: |
case OP_TYPEUPTO: |
940 |
case OP_TYPEMINUPTO: |
case OP_TYPEMINUPTO: |
941 |
if (codevalue != OP_TYPEEXACT) |
case OP_TYPEPOSUPTO: |
942 |
{ ADD_ACTIVE(state_offset + 4, 0); } |
ADD_ACTIVE(state_offset + 4, 0); |
943 |
count = current_state->count; /* Number already matched */ |
count = current_state->count; /* Number already matched */ |
944 |
if (clen > 0) |
if (clen > 0) |
945 |
{ |
{ |
946 |
if ((c >= 256 && d != OP_DIGIT && d != OP_WHITESPACE && d != OP_WORDCHAR) || |
if ((c >= 256 && d != OP_DIGIT && d != OP_WHITESPACE && d != OP_WORDCHAR) || |
947 |
(c < 256 && |
(c < 256 && |
948 |
(d != OP_ANY || c != '\n' || (ims & PCRE_DOTALL) != 0) && |
(d != OP_ANY || |
949 |
|
(ims & PCRE_DOTALL) != 0 || |
950 |
|
!IS_NEWLINE(ptr) |
951 |
|
) && |
952 |
((ctypes[c] & toptable1[d]) ^ toptable2[d]) != 0)) |
((ctypes[c] & toptable1[d]) ^ toptable2[d]) != 0)) |
953 |
{ |
{ |
954 |
|
if (codevalue == OP_TYPEPOSUPTO) |
955 |
|
{ |
956 |
|
active_count--; /* Remove non-match possibility */ |
957 |
|
next_active_state--; |
958 |
|
} |
959 |
if (++count >= GET2(code, 1)) |
if (++count >= GET2(code, 1)) |
960 |
{ ADD_NEW(state_offset + 4, 0); } |
{ ADD_NEW(state_offset + 4, 0); } |
961 |
else |
else |
966 |
|
|
967 |
/* ========================================================================== */ |
/* ========================================================================== */ |
968 |
/* These are virtual opcodes that are used when something like |
/* These are virtual opcodes that are used when something like |
969 |
OP_TYPEPLUS has OP_PROP, OP_NOTPROP, or OP_EXTUNI as its argument. It |
OP_TYPEPLUS has OP_PROP, OP_NOTPROP, OP_ANYNL, or OP_EXTUNI as its |
970 |
keeps the code above fast for the other cases. The argument is in the |
argument. It keeps the code above fast for the other cases. The argument |
971 |
d variable. */ |
is in the d variable. */ |
972 |
|
|
973 |
case OP_PROP_EXTRA + OP_TYPEPLUS: |
case OP_PROP_EXTRA + OP_TYPEPLUS: |
974 |
case OP_PROP_EXTRA + OP_TYPEMINPLUS: |
case OP_PROP_EXTRA + OP_TYPEMINPLUS: |
975 |
|
case OP_PROP_EXTRA + OP_TYPEPOSPLUS: |
976 |
count = current_state->count; /* Already matched */ |
count = current_state->count; /* Already matched */ |
977 |
if (count > 0) { ADD_ACTIVE(state_offset + 3, 0); } |
if (count > 0) { ADD_ACTIVE(state_offset + 4, 0); } |
978 |
if (clen > 0) |
if (clen > 0) |
979 |
{ |
{ |
980 |
int category = ucp_findchar(c, &chartype, &othercase); |
BOOL OK; |
981 |
int rqdtype = code[2]; |
int category = _pcre_ucp_findprop(c, &chartype, &script); |
982 |
if ((d == OP_PROP) == |
switch(code[2]) |
983 |
(rqdtype == ((rqdtype >= 128)? (category + 128) : chartype))) |
{ |
984 |
{ count++; ADD_NEW(state_offset, count); } |
case PT_ANY: |
985 |
|
OK = TRUE; |
986 |
|
break; |
987 |
|
|
988 |
|
case PT_LAMP: |
989 |
|
OK = chartype == ucp_Lu || chartype == ucp_Ll || chartype == ucp_Lt; |
990 |
|
break; |
991 |
|
|
992 |
|
case PT_GC: |
993 |
|
OK = category == code[3]; |
994 |
|
break; |
995 |
|
|
996 |
|
case PT_PC: |
997 |
|
OK = chartype == code[3]; |
998 |
|
break; |
999 |
|
|
1000 |
|
case PT_SC: |
1001 |
|
OK = script == code[3]; |
1002 |
|
break; |
1003 |
|
|
1004 |
|
/* Should never occur, but keep compilers from grumbling. */ |
1005 |
|
|
1006 |
|
default: |
1007 |
|
OK = codevalue != OP_PROP; |
1008 |
|
break; |
1009 |
|
} |
1010 |
|
|
1011 |
|
if (OK == (d == OP_PROP)) |
1012 |
|
{ |
1013 |
|
if (count > 0 && codevalue == OP_PROP_EXTRA + OP_TYPEPOSPLUS) |
1014 |
|
{ |
1015 |
|
active_count--; /* Remove non-match possibility */ |
1016 |
|
next_active_state--; |
1017 |
|
} |
1018 |
|
count++; |
1019 |
|
ADD_NEW(state_offset, count); |
1020 |
|
} |
1021 |
} |
} |
1022 |
break; |
break; |
1023 |
|
|
1024 |
/*-----------------------------------------------------------------*/ |
/*-----------------------------------------------------------------*/ |
1025 |
case OP_EXTUNI_EXTRA + OP_TYPEPLUS: |
case OP_EXTUNI_EXTRA + OP_TYPEPLUS: |
1026 |
case OP_EXTUNI_EXTRA + OP_TYPEMINPLUS: |
case OP_EXTUNI_EXTRA + OP_TYPEMINPLUS: |
1027 |
|
case OP_EXTUNI_EXTRA + OP_TYPEPOSPLUS: |
1028 |
count = current_state->count; /* Already matched */ |
count = current_state->count; /* Already matched */ |
1029 |
if (count > 0) { ADD_ACTIVE(state_offset + 2, 0); } |
if (count > 0) { ADD_ACTIVE(state_offset + 2, 0); } |
1030 |
if (clen > 0 && ucp_findchar(c, &chartype, &othercase) != ucp_M) |
if (clen > 0 && _pcre_ucp_findprop(c, &chartype, &script) != ucp_M) |
1031 |
{ |
{ |
1032 |
const uschar *nptr = ptr + clen; |
const uschar *nptr = ptr + clen; |
1033 |
int ncount = 0; |
int ncount = 0; |
1034 |
|
if (count > 0 && codevalue == OP_EXTUNI_EXTRA + OP_TYPEPOSPLUS) |
1035 |
|
{ |
1036 |
|
active_count--; /* Remove non-match possibility */ |
1037 |
|
next_active_state--; |
1038 |
|
} |
1039 |
while (nptr < end_subject) |
while (nptr < end_subject) |
1040 |
{ |
{ |
1041 |
int nd; |
int nd; |
1042 |
int ndlen = 1; |
int ndlen = 1; |
1043 |
GETCHARLEN(nd, nptr, ndlen); |
GETCHARLEN(nd, nptr, ndlen); |
1044 |
if (ucp_findchar(nd, &chartype, &othercase) != ucp_M) break; |
if (_pcre_ucp_findprop(nd, &chartype, &script) != ucp_M) break; |
1045 |
ncount++; |
ncount++; |
1046 |
nptr += ndlen; |
nptr += ndlen; |
1047 |
} |
} |
1051 |
break; |
break; |
1052 |
|
|
1053 |
/*-----------------------------------------------------------------*/ |
/*-----------------------------------------------------------------*/ |
1054 |
|
case OP_ANYNL_EXTRA + OP_TYPEPLUS: |
1055 |
|
case OP_ANYNL_EXTRA + OP_TYPEMINPLUS: |
1056 |
|
case OP_ANYNL_EXTRA + OP_TYPEPOSPLUS: |
1057 |
|
count = current_state->count; /* Already matched */ |
1058 |
|
if (count > 0) { ADD_ACTIVE(state_offset + 2, 0); } |
1059 |
|
if (clen > 0) |
1060 |
|
{ |
1061 |
|
int ncount = 0; |
1062 |
|
switch (c) |
1063 |
|
{ |
1064 |
|
case 0x000d: |
1065 |
|
if (ptr + 1 < end_subject && ptr[1] == 0x0a) ncount = 1; |
1066 |
|
/* Fall through */ |
1067 |
|
case 0x000a: |
1068 |
|
case 0x000b: |
1069 |
|
case 0x000c: |
1070 |
|
case 0x0085: |
1071 |
|
case 0x2028: |
1072 |
|
case 0x2029: |
1073 |
|
if (count > 0 && codevalue == OP_ANYNL_EXTRA + OP_TYPEPOSPLUS) |
1074 |
|
{ |
1075 |
|
active_count--; /* Remove non-match possibility */ |
1076 |
|
next_active_state--; |
1077 |
|
} |
1078 |
|
count++; |
1079 |
|
ADD_NEW_DATA(-state_offset, count, ncount); |
1080 |
|
break; |
1081 |
|
default: |
1082 |
|
break; |
1083 |
|
} |
1084 |
|
} |
1085 |
|
break; |
1086 |
|
|
1087 |
|
/*-----------------------------------------------------------------*/ |
1088 |
case OP_PROP_EXTRA + OP_TYPEQUERY: |
case OP_PROP_EXTRA + OP_TYPEQUERY: |
1089 |
case OP_PROP_EXTRA + OP_TYPEMINQUERY: |
case OP_PROP_EXTRA + OP_TYPEMINQUERY: |
1090 |
count = 3; |
case OP_PROP_EXTRA + OP_TYPEPOSQUERY: |
1091 |
|
count = 4; |
1092 |
goto QS1; |
goto QS1; |
1093 |
|
|
1094 |
case OP_PROP_EXTRA + OP_TYPESTAR: |
case OP_PROP_EXTRA + OP_TYPESTAR: |
1095 |
case OP_PROP_EXTRA + OP_TYPEMINSTAR: |
case OP_PROP_EXTRA + OP_TYPEMINSTAR: |
1096 |
|
case OP_PROP_EXTRA + OP_TYPEPOSSTAR: |
1097 |
count = 0; |
count = 0; |
1098 |
|
|
1099 |
QS1: |
QS1: |
1100 |
|
|
1101 |
ADD_ACTIVE(state_offset + 3, 0); |
ADD_ACTIVE(state_offset + 4, 0); |
1102 |
if (clen > 0) |
if (clen > 0) |
1103 |
{ |
{ |
1104 |
int category = ucp_findchar(c, &chartype, &othercase); |
BOOL OK; |
1105 |
int rqdtype = code[2]; |
int category = _pcre_ucp_findprop(c, &chartype, &script); |
1106 |
if ((d == OP_PROP) == |
switch(code[2]) |
1107 |
(rqdtype == ((rqdtype >= 128)? (category + 128) : chartype))) |
{ |
1108 |
{ ADD_NEW(state_offset + count, 0); } |
case PT_ANY: |
1109 |
|
OK = TRUE; |
1110 |
|
break; |
1111 |
|
|
1112 |
|
case PT_LAMP: |
1113 |
|
OK = chartype == ucp_Lu || chartype == ucp_Ll || chartype == ucp_Lt; |
1114 |
|
break; |
1115 |
|
|
1116 |
|
case PT_GC: |
1117 |
|
OK = category == code[3]; |
1118 |
|
break; |
1119 |
|
|
1120 |
|
case PT_PC: |
1121 |
|
OK = chartype == code[3]; |
1122 |
|
break; |
1123 |
|
|
1124 |
|
case PT_SC: |
1125 |
|
OK = script == code[3]; |
1126 |
|
break; |
1127 |
|
|
1128 |
|
/* Should never occur, but keep compilers from grumbling. */ |
1129 |
|
|
1130 |
|
default: |
1131 |
|
OK = codevalue != OP_PROP; |
1132 |
|
break; |
1133 |
|
} |
1134 |
|
|
1135 |
|
if (OK == (d == OP_PROP)) |
1136 |
|
{ |
1137 |
|
if (codevalue == OP_PROP_EXTRA + OP_TYPEPOSSTAR || |
1138 |
|
codevalue == OP_PROP_EXTRA + OP_TYPEPOSQUERY) |
1139 |
|
{ |
1140 |
|
active_count--; /* Remove non-match possibility */ |
1141 |
|
next_active_state--; |
1142 |
|
} |
1143 |
|
ADD_NEW(state_offset + count, 0); |
1144 |
|
} |
1145 |
} |
} |
1146 |
break; |
break; |
1147 |
|
|
1148 |
/*-----------------------------------------------------------------*/ |
/*-----------------------------------------------------------------*/ |
1149 |
case OP_EXTUNI_EXTRA + OP_TYPEQUERY: |
case OP_EXTUNI_EXTRA + OP_TYPEQUERY: |
1150 |
case OP_EXTUNI_EXTRA + OP_TYPEMINQUERY: |
case OP_EXTUNI_EXTRA + OP_TYPEMINQUERY: |
1151 |
|
case OP_EXTUNI_EXTRA + OP_TYPEPOSQUERY: |
1152 |
count = 2; |
count = 2; |
1153 |
goto QS2; |
goto QS2; |
1154 |
|
|
1155 |
case OP_EXTUNI_EXTRA + OP_TYPESTAR: |
case OP_EXTUNI_EXTRA + OP_TYPESTAR: |
1156 |
case OP_EXTUNI_EXTRA + OP_TYPEMINSTAR: |
case OP_EXTUNI_EXTRA + OP_TYPEMINSTAR: |
1157 |
|
case OP_EXTUNI_EXTRA + OP_TYPEPOSSTAR: |
1158 |
count = 0; |
count = 0; |
1159 |
|
|
1160 |
QS2: |
QS2: |
1161 |
|
|
1162 |
ADD_ACTIVE(state_offset + 2, 0); |
ADD_ACTIVE(state_offset + 2, 0); |
1163 |
if (clen > 0 && ucp_findchar(c, &chartype, &othercase) != ucp_M) |
if (clen > 0 && _pcre_ucp_findprop(c, &chartype, &script) != ucp_M) |
1164 |
{ |
{ |
1165 |
const uschar *nptr = ptr + clen; |
const uschar *nptr = ptr + clen; |
1166 |
int ncount = 0; |
int ncount = 0; |
1167 |
|
if (codevalue == OP_EXTUNI_EXTRA + OP_TYPEPOSSTAR || |
1168 |
|
codevalue == OP_EXTUNI_EXTRA + OP_TYPEPOSQUERY) |
1169 |
|
{ |
1170 |
|
active_count--; /* Remove non-match possibility */ |
1171 |
|
next_active_state--; |
1172 |
|
} |
1173 |
while (nptr < end_subject) |
while (nptr < end_subject) |
1174 |
{ |
{ |
1175 |
int nd; |
int nd; |
1176 |
int ndlen = 1; |
int ndlen = 1; |
1177 |
GETCHARLEN(nd, nptr, ndlen); |
GETCHARLEN(nd, nptr, ndlen); |
1178 |
if (ucp_findchar(nd, &chartype, &othercase) != ucp_M) break; |
if (_pcre_ucp_findprop(nd, &chartype, &script) != ucp_M) break; |
1179 |
ncount++; |
ncount++; |
1180 |
nptr += ndlen; |
nptr += ndlen; |
1181 |
} |
} |
1184 |
break; |
break; |
1185 |
|
|
1186 |
/*-----------------------------------------------------------------*/ |
/*-----------------------------------------------------------------*/ |
1187 |
|
case OP_ANYNL_EXTRA + OP_TYPEQUERY: |
1188 |
|
case OP_ANYNL_EXTRA + OP_TYPEMINQUERY: |
1189 |
|
case OP_ANYNL_EXTRA + OP_TYPEPOSQUERY: |
1190 |
|
count = 2; |
1191 |
|
goto QS3; |
1192 |
|
|
1193 |
|
case OP_ANYNL_EXTRA + OP_TYPESTAR: |
1194 |
|
case OP_ANYNL_EXTRA + OP_TYPEMINSTAR: |
1195 |
|
case OP_ANYNL_EXTRA + OP_TYPEPOSSTAR: |
1196 |
|
count = 0; |
1197 |
|
|
1198 |
|
QS3: |
1199 |
|
ADD_ACTIVE(state_offset + 2, 0); |
1200 |
|
if (clen > 0) |
1201 |
|
{ |
1202 |
|
int ncount = 0; |
1203 |
|
switch (c) |
1204 |
|
{ |
1205 |
|
case 0x000d: |
1206 |
|
if (ptr + 1 < end_subject && ptr[1] == 0x0a) ncount = 1; |
1207 |
|
/* Fall through */ |
1208 |
|
case 0x000a: |
1209 |
|
case 0x000b: |
1210 |
|
case 0x000c: |
1211 |
|
case 0x0085: |
1212 |
|
case 0x2028: |
1213 |
|
case 0x2029: |
1214 |
|
if (codevalue == OP_ANYNL_EXTRA + OP_TYPEPOSSTAR || |
1215 |
|
codevalue == OP_ANYNL_EXTRA + OP_TYPEPOSQUERY) |
1216 |
|
{ |
1217 |
|
active_count--; /* Remove non-match possibility */ |
1218 |
|
next_active_state--; |
1219 |
|
} |
1220 |
|
ADD_NEW_DATA(-(state_offset + count), 0, ncount); |
1221 |
|
break; |
1222 |
|
default: |
1223 |
|
break; |
1224 |
|
} |
1225 |
|
} |
1226 |
|
break; |
1227 |
|
|
1228 |
|
/*-----------------------------------------------------------------*/ |
1229 |
case OP_PROP_EXTRA + OP_TYPEEXACT: |
case OP_PROP_EXTRA + OP_TYPEEXACT: |
1230 |
case OP_PROP_EXTRA + OP_TYPEUPTO: |
case OP_PROP_EXTRA + OP_TYPEUPTO: |
1231 |
case OP_PROP_EXTRA + OP_TYPEMINUPTO: |
case OP_PROP_EXTRA + OP_TYPEMINUPTO: |
1232 |
|
case OP_PROP_EXTRA + OP_TYPEPOSUPTO: |
1233 |
if (codevalue != OP_PROP_EXTRA + OP_TYPEEXACT) |
if (codevalue != OP_PROP_EXTRA + OP_TYPEEXACT) |
1234 |
{ ADD_ACTIVE(state_offset + 5, 0); } |
{ ADD_ACTIVE(state_offset + 6, 0); } |
1235 |
count = current_state->count; /* Number already matched */ |
count = current_state->count; /* Number already matched */ |
1236 |
if (clen > 0) |
if (clen > 0) |
1237 |
{ |
{ |
1238 |
int category = ucp_findchar(c, &chartype, &othercase); |
BOOL OK; |
1239 |
int rqdtype = code[4]; |
int category = _pcre_ucp_findprop(c, &chartype, &script); |
1240 |
if ((d == OP_PROP) == |
switch(code[4]) |
1241 |
(rqdtype == ((rqdtype >= 128)? (category + 128) : chartype))) |
{ |
1242 |
|
case PT_ANY: |
1243 |
|
OK = TRUE; |
1244 |
|
break; |
1245 |
|
|
1246 |
|
case PT_LAMP: |
1247 |
|
OK = chartype == ucp_Lu || chartype == ucp_Ll || chartype == ucp_Lt; |
1248 |
|
break; |
1249 |
|
|
1250 |
|
case PT_GC: |
1251 |
|
OK = category == code[5]; |
1252 |
|
break; |
1253 |
|
|
1254 |
|
case PT_PC: |
1255 |
|
OK = chartype == code[5]; |
1256 |
|
break; |
1257 |
|
|
1258 |
|
case PT_SC: |
1259 |
|
OK = script == code[5]; |
1260 |
|
break; |
1261 |
|
|
1262 |
|
/* Should never occur, but keep compilers from grumbling. */ |
1263 |
|
|
1264 |
|
default: |
1265 |
|
OK = codevalue != OP_PROP; |
1266 |
|
break; |
1267 |
|
} |
1268 |
|
|
1269 |
|
if (OK == (d == OP_PROP)) |
1270 |
{ |
{ |
1271 |
|
if (codevalue == OP_PROP_EXTRA + OP_TYPEPOSUPTO) |
1272 |
|
{ |
1273 |
|
active_count--; /* Remove non-match possibility */ |
1274 |
|
next_active_state--; |
1275 |
|
} |
1276 |
if (++count >= GET2(code, 1)) |
if (++count >= GET2(code, 1)) |
1277 |
{ ADD_NEW(state_offset + 5, 0); } |
{ ADD_NEW(state_offset + 6, 0); } |
1278 |
else |
else |
1279 |
{ ADD_NEW(state_offset, count); } |
{ ADD_NEW(state_offset, count); } |
1280 |
} |
} |
1285 |
case OP_EXTUNI_EXTRA + OP_TYPEEXACT: |
case OP_EXTUNI_EXTRA + OP_TYPEEXACT: |
1286 |
case OP_EXTUNI_EXTRA + OP_TYPEUPTO: |
case OP_EXTUNI_EXTRA + OP_TYPEUPTO: |
1287 |
case OP_EXTUNI_EXTRA + OP_TYPEMINUPTO: |
case OP_EXTUNI_EXTRA + OP_TYPEMINUPTO: |
1288 |
|
case OP_EXTUNI_EXTRA + OP_TYPEPOSUPTO: |
1289 |
if (codevalue != OP_EXTUNI_EXTRA + OP_TYPEEXACT) |
if (codevalue != OP_EXTUNI_EXTRA + OP_TYPEEXACT) |
1290 |
{ ADD_ACTIVE(state_offset + 4, 0); } |
{ ADD_ACTIVE(state_offset + 4, 0); } |
1291 |
count = current_state->count; /* Number already matched */ |
count = current_state->count; /* Number already matched */ |
1292 |
if (clen > 0 && ucp_findchar(c, &chartype, &othercase) != ucp_M) |
if (clen > 0 && _pcre_ucp_findprop(c, &chartype, &script) != ucp_M) |
1293 |
{ |
{ |
1294 |
const uschar *nptr = ptr + clen; |
const uschar *nptr = ptr + clen; |
1295 |
int ncount = 0; |
int ncount = 0; |
1296 |
|
if (codevalue == OP_EXTUNI_EXTRA + OP_TYPEPOSUPTO) |
1297 |
|
{ |
1298 |
|
active_count--; /* Remove non-match possibility */ |
1299 |
|
next_active_state--; |
1300 |
|
} |
1301 |
while (nptr < end_subject) |
while (nptr < end_subject) |
1302 |
{ |
{ |
1303 |
int nd; |
int nd; |
1304 |
int ndlen = 1; |
int ndlen = 1; |
1305 |
GETCHARLEN(nd, nptr, ndlen); |
GETCHARLEN(nd, nptr, ndlen); |
1306 |
if (ucp_findchar(nd, &chartype, &othercase) != ucp_M) break; |
if (_pcre_ucp_findprop(nd, &chartype, &script) != ucp_M) break; |
1307 |
ncount++; |
ncount++; |
1308 |
nptr += ndlen; |
nptr += ndlen; |
1309 |
} |
} |
1314 |
} |
} |
1315 |
break; |
break; |
1316 |
|
|
1317 |
|
/*-----------------------------------------------------------------*/ |
1318 |
|
case OP_ANYNL_EXTRA + OP_TYPEEXACT: |
1319 |
|
case OP_ANYNL_EXTRA + OP_TYPEUPTO: |
1320 |
|
case OP_ANYNL_EXTRA + OP_TYPEMINUPTO: |
1321 |
|
case OP_ANYNL_EXTRA + OP_TYPEPOSUPTO: |
1322 |
|
if (codevalue != OP_ANYNL_EXTRA + OP_TYPEEXACT) |
1323 |
|
{ ADD_ACTIVE(state_offset + 4, 0); } |
1324 |
|
count = current_state->count; /* Number already matched */ |
1325 |
|
if (clen > 0) |
1326 |
|
{ |
1327 |
|
int ncount = 0; |
1328 |
|
switch (c) |
1329 |
|
{ |
1330 |
|
case 0x000d: |
1331 |
|
if (ptr + 1 < end_subject && ptr[1] == 0x0a) ncount = 1; |
1332 |
|
/* Fall through */ |
1333 |
|
case 0x000a: |
1334 |
|
case 0x000b: |
1335 |
|
case 0x000c: |
1336 |
|
case 0x0085: |
1337 |
|
case 0x2028: |
1338 |
|
case 0x2029: |
1339 |
|
if (codevalue == OP_ANYNL_EXTRA + OP_TYPEPOSUPTO) |
1340 |
|
{ |
1341 |
|
active_count--; /* Remove non-match possibility */ |
1342 |
|
next_active_state--; |
1343 |
|
} |
1344 |
|
if (++count >= GET2(code, 1)) |
1345 |
|
{ ADD_NEW_DATA(-(state_offset + 4), 0, ncount); } |
1346 |
|
else |
1347 |
|
{ ADD_NEW_DATA(-state_offset, count, ncount); } |
1348 |
|
break; |
1349 |
|
default: |
1350 |
|
break; |
1351 |
|
} |
1352 |
|
} |
1353 |
|
break; |
1354 |
|
|
1355 |
/* ========================================================================== */ |
/* ========================================================================== */ |
1356 |
/* These opcodes are followed by a character that is usually compared |
/* These opcodes are followed by a character that is usually compared |
1357 |
to the current subject character; it is loaded into d. We still get |
to the current subject character; it is loaded into d. We still get |
1372 |
{ |
{ |
1373 |
if (c == d) { ADD_NEW(state_offset + dlen + 1, 0); } else |
if (c == d) { ADD_NEW(state_offset + dlen + 1, 0); } else |
1374 |
{ |
{ |
1375 |
|
unsigned int othercase; |
1376 |
if (c < 128) othercase = fcc[c]; else |
if (c < 128) othercase = fcc[c]; else |
1377 |
|
|
1378 |
/* If we have Unicode property support, we can use it to test the |
/* If we have Unicode property support, we can use it to test the |
1379 |
other case of the character, if there is one. The result of |
other case of the character. */ |
|
ucp_findchar() is < 0 if the char isn't found, and othercase is |
|
|
returned as zero if there isn't another case. */ |
|
1380 |
|
|
1381 |
#ifdef SUPPORT_UCP |
#ifdef SUPPORT_UCP |
1382 |
if (ucp_findchar(c, &chartype, &othercase) < 0) |
othercase = _pcre_ucp_othercase(c); |
1383 |
|
#else |
1384 |
|
othercase = NOTACHAR; |
1385 |
#endif |
#endif |
|
othercase = -1; |
|
1386 |
|
|
1387 |
if (d == othercase) { ADD_NEW(state_offset + dlen + 1, 0); } |
if (d == othercase) { ADD_NEW(state_offset + dlen + 1, 0); } |
1388 |
} |
} |
1404 |
to wait for them to pass before continuing. */ |
to wait for them to pass before continuing. */ |
1405 |
|
|
1406 |
case OP_EXTUNI: |
case OP_EXTUNI: |
1407 |
if (clen > 0 && ucp_findchar(c, &chartype, &othercase) != ucp_M) |
if (clen > 0 && _pcre_ucp_findprop(c, &chartype, &script) != ucp_M) |
1408 |
{ |
{ |
1409 |
const uschar *nptr = ptr + clen; |
const uschar *nptr = ptr + clen; |
1410 |
int ncount = 0; |
int ncount = 0; |
1412 |
{ |
{ |
1413 |
int nclen = 1; |
int nclen = 1; |
1414 |
GETCHARLEN(c, nptr, nclen); |
GETCHARLEN(c, nptr, nclen); |
1415 |
if (ucp_findchar(c, &chartype, &othercase) != ucp_M) break; |
if (_pcre_ucp_findprop(c, &chartype, &script) != ucp_M) break; |
1416 |
ncount++; |
ncount++; |
1417 |
nptr += nclen; |
nptr += nclen; |
1418 |
} |
} |
1422 |
#endif |
#endif |
1423 |
|
|
1424 |
/*-----------------------------------------------------------------*/ |
/*-----------------------------------------------------------------*/ |
1425 |
|
/* This is a tricky like EXTUNI because it too can match more than one |
1426 |
|
character (when CR is followed by LF). In this case, set up a negative |
1427 |
|
state to wait for one character to pass before continuing. */ |
1428 |
|
|
1429 |
|
case OP_ANYNL: |
1430 |
|
if (clen > 0) switch(c) |
1431 |
|
{ |
1432 |
|
case 0x000a: |
1433 |
|
case 0x000b: |
1434 |
|
case 0x000c: |
1435 |
|
case 0x0085: |
1436 |
|
case 0x2028: |
1437 |
|
case 0x2029: |
1438 |
|
ADD_NEW(state_offset + 1, 0); |
1439 |
|
break; |
1440 |
|
case 0x000d: |
1441 |
|
if (ptr + 1 < end_subject && ptr[1] == 0x0a) |
1442 |
|
{ |
1443 |
|
ADD_NEW_DATA(-(state_offset + 1), 0, 1); |
1444 |
|
} |
1445 |
|
else |
1446 |
|
{ |
1447 |
|
ADD_NEW(state_offset + 1, 0); |
1448 |
|
} |
1449 |
|
break; |
1450 |
|
} |
1451 |
|
break; |
1452 |
|
|
1453 |
|
/*-----------------------------------------------------------------*/ |
1454 |
/* Match a negated single character. This is only used for one-byte |
/* Match a negated single character. This is only used for one-byte |
1455 |
characters, that is, we know that d < 256. The character we are |
characters, that is, we know that d < 256. The character we are |
1456 |
checking (c) can be multibyte. */ |
checking (c) can be multibyte. */ |
1458 |
case OP_NOT: |
case OP_NOT: |
1459 |
if (clen > 0) |
if (clen > 0) |
1460 |
{ |
{ |
1461 |
int otherd = ((ims & PCRE_CASELESS) != 0)? fcc[d] : d; |
unsigned int otherd = ((ims & PCRE_CASELESS) != 0)? fcc[d] : d; |
1462 |
if (c != d && c != otherd) { ADD_NEW(state_offset + dlen + 1, 0); } |
if (c != d && c != otherd) { ADD_NEW(state_offset + dlen + 1, 0); } |
1463 |
} |
} |
1464 |
break; |
break; |
1466 |
/*-----------------------------------------------------------------*/ |
/*-----------------------------------------------------------------*/ |
1467 |
case OP_PLUS: |
case OP_PLUS: |
1468 |
case OP_MINPLUS: |
case OP_MINPLUS: |
1469 |
|
case OP_POSPLUS: |
1470 |
case OP_NOTPLUS: |
case OP_NOTPLUS: |
1471 |
case OP_NOTMINPLUS: |
case OP_NOTMINPLUS: |
1472 |
|
case OP_NOTPOSPLUS: |
1473 |
count = current_state->count; /* Already matched */ |
count = current_state->count; /* Already matched */ |
1474 |
if (count > 0) { ADD_ACTIVE(state_offset + dlen + 1, 0); } |
if (count > 0) { ADD_ACTIVE(state_offset + dlen + 1, 0); } |
1475 |
if (clen > 0) |
if (clen > 0) |
1476 |
{ |
{ |
1477 |
int otherd = -1; |
unsigned int otherd = NOTACHAR; |
1478 |
if ((ims & PCRE_CASELESS) != 0) |
if ((ims & PCRE_CASELESS) != 0) |
1479 |
{ |
{ |
1480 |
#ifdef SUPPORT_UTF8 |
#ifdef SUPPORT_UTF8 |
1481 |
if (utf8 && c >= 128) |
if (utf8 && d >= 128) |
1482 |
{ |
{ |
1483 |
#ifdef SUPPORT_UCP |
#ifdef SUPPORT_UCP |
1484 |
if (ucp_findchar(d, &chartype, &otherd) < 0) otherd = -1; |
otherd = _pcre_ucp_othercase(d); |
1485 |
#endif /* SUPPORT_UCP */ |
#endif /* SUPPORT_UCP */ |
1486 |
} |
} |
1487 |
else |
else |
1489 |
otherd = fcc[d]; |
otherd = fcc[d]; |
1490 |
} |
} |
1491 |
if ((c == d || c == otherd) == (codevalue < OP_NOTSTAR)) |
if ((c == d || c == otherd) == (codevalue < OP_NOTSTAR)) |
1492 |
{ count++; ADD_NEW(state_offset, count); } |
{ |
1493 |
|
if (count > 0 && |
1494 |
|
(codevalue == OP_POSPLUS || codevalue == OP_NOTPOSPLUS)) |
1495 |
|
{ |
1496 |
|
active_count--; /* Remove non-match possibility */ |
1497 |
|
next_active_state--; |
1498 |
|
} |
1499 |
|
count++; |
1500 |
|
ADD_NEW(state_offset, count); |
1501 |
|
} |
1502 |
} |
} |
1503 |
break; |
break; |
1504 |
|
|
1505 |
/*-----------------------------------------------------------------*/ |
/*-----------------------------------------------------------------*/ |
1506 |
case OP_QUERY: |
case OP_QUERY: |
1507 |
case OP_MINQUERY: |
case OP_MINQUERY: |
1508 |
|
case OP_POSQUERY: |
1509 |
case OP_NOTQUERY: |
case OP_NOTQUERY: |
1510 |
case OP_NOTMINQUERY: |
case OP_NOTMINQUERY: |
1511 |
|
case OP_NOTPOSQUERY: |
1512 |
ADD_ACTIVE(state_offset + dlen + 1, 0); |
ADD_ACTIVE(state_offset + dlen + 1, 0); |
1513 |
if (clen > 0) |
if (clen > 0) |
1514 |
{ |
{ |
1515 |
int otherd = -1; |
unsigned int otherd = NOTACHAR; |
1516 |
if ((ims && PCRE_CASELESS) != 0) |
if ((ims & PCRE_CASELESS) != 0) |
1517 |
{ |
{ |
1518 |
#ifdef SUPPORT_UTF8 |
#ifdef SUPPORT_UTF8 |
1519 |
if (utf8 && c >= 128) |
if (utf8 && d >= 128) |
1520 |
{ |
{ |
1521 |
#ifdef SUPPORT_UCP |
#ifdef SUPPORT_UCP |
1522 |
if (ucp_findchar(c, &chartype, &otherd) < 0) otherd = -1; |
otherd = _pcre_ucp_othercase(d); |
1523 |
#endif /* SUPPORT_UCP */ |
#endif /* SUPPORT_UCP */ |
1524 |
} |
} |
1525 |
else |
else |
1527 |
otherd = fcc[d]; |
otherd = fcc[d]; |
1528 |
} |
} |
1529 |
if ((c == d || c == otherd) == (codevalue < OP_NOTSTAR)) |
if ((c == d || c == otherd) == (codevalue < OP_NOTSTAR)) |
1530 |
{ ADD_NEW(state_offset + dlen + 1, 0); } |
{ |
1531 |
|
if (codevalue == OP_POSQUERY || codevalue == OP_NOTPOSQUERY) |
1532 |
|
{ |
1533 |
|
active_count--; /* Remove non-match possibility */ |
1534 |
|
next_active_state--; |
1535 |
|
} |
1536 |
|
ADD_NEW(state_offset + dlen + 1, 0); |
1537 |
|
} |
1538 |
} |
} |
1539 |
break; |
break; |
1540 |
|
|
1541 |
/*-----------------------------------------------------------------*/ |
/*-----------------------------------------------------------------*/ |
1542 |
case OP_STAR: |
case OP_STAR: |
1543 |
case OP_MINSTAR: |
case OP_MINSTAR: |
1544 |
|
case OP_POSSTAR: |
1545 |
case OP_NOTSTAR: |
case OP_NOTSTAR: |
1546 |
case OP_NOTMINSTAR: |
case OP_NOTMINSTAR: |
1547 |
|
case OP_NOTPOSSTAR: |
1548 |
ADD_ACTIVE(state_offset + dlen + 1, 0); |
ADD_ACTIVE(state_offset + dlen + 1, 0); |
1549 |
if (clen > 0) |
if (clen > 0) |
1550 |
{ |
{ |
1551 |
int otherd = -1; |
unsigned int otherd = NOTACHAR; |
1552 |
if ((ims && PCRE_CASELESS) != 0) |
if ((ims & PCRE_CASELESS) != 0) |
1553 |
{ |
{ |
1554 |
#ifdef SUPPORT_UTF8 |
#ifdef SUPPORT_UTF8 |
1555 |
if (utf8 && c >= 128) |
if (utf8 && d >= 128) |
1556 |
{ |
{ |
1557 |
#ifdef SUPPORT_UCP |
#ifdef SUPPORT_UCP |
1558 |
if (ucp_findchar(c, &chartype, &otherd) < 0) otherd = -1; |
otherd = _pcre_ucp_othercase(d); |
1559 |
#endif /* SUPPORT_UCP */ |
#endif /* SUPPORT_UCP */ |
1560 |
} |
} |
1561 |
else |
else |
1563 |
otherd = fcc[d]; |
otherd = fcc[d]; |
1564 |
} |
} |
1565 |
if ((c == d || c == otherd) == (codevalue < OP_NOTSTAR)) |
if ((c == d || c == otherd) == (codevalue < OP_NOTSTAR)) |
1566 |
{ ADD_NEW(state_offset, 0); } |
{ |
1567 |
|
if (codevalue == OP_POSSTAR || codevalue == OP_NOTPOSSTAR) |
1568 |
|
{ |
1569 |
|
active_count--; /* Remove non-match possibility */ |
1570 |
|
next_active_state--; |
1571 |
|
} |
1572 |
|
ADD_NEW(state_offset, 0); |
1573 |
|
} |
1574 |
} |
} |
1575 |
break; |
break; |
1576 |
|
|
1577 |
/*-----------------------------------------------------------------*/ |
/*-----------------------------------------------------------------*/ |
1578 |
case OP_EXACT: |
case OP_EXACT: |
1579 |
|
case OP_NOTEXACT: |
1580 |
|
count = current_state->count; /* Number already matched */ |
1581 |
|
if (clen > 0) |
1582 |
|
{ |
1583 |
|
unsigned int otherd = NOTACHAR; |
1584 |
|
if ((ims & PCRE_CASELESS) != 0) |
1585 |
|
{ |
1586 |
|
#ifdef SUPPORT_UTF8 |
1587 |
|
if (utf8 && d >= 128) |
1588 |
|
{ |
1589 |
|
#ifdef SUPPORT_UCP |
1590 |
|
otherd = _pcre_ucp_othercase(d); |
1591 |
|
#endif /* SUPPORT_UCP */ |
1592 |
|
} |
1593 |
|
else |
1594 |
|
#endif /* SUPPORT_UTF8 */ |
1595 |
|
otherd = fcc[d]; |
1596 |
|
} |
1597 |
|
if ((c == d || c == otherd) == (codevalue < OP_NOTSTAR)) |
1598 |
|
{ |
1599 |
|
if (++count >= GET2(code, 1)) |
1600 |
|
{ ADD_NEW(state_offset + dlen + 3, 0); } |
1601 |
|
else |
1602 |
|
{ ADD_NEW(state_offset, count); } |
1603 |
|
} |
1604 |
|
} |
1605 |
|
break; |
1606 |
|
|
1607 |
|
/*-----------------------------------------------------------------*/ |
1608 |
case OP_UPTO: |
case OP_UPTO: |
1609 |
case OP_MINUPTO: |
case OP_MINUPTO: |
1610 |
case OP_NOTEXACT: |
case OP_POSUPTO: |
1611 |
case OP_NOTUPTO: |
case OP_NOTUPTO: |
1612 |
case OP_NOTMINUPTO: |
case OP_NOTMINUPTO: |
1613 |
if (codevalue != OP_EXACT && codevalue != OP_NOTEXACT) |
case OP_NOTPOSUPTO: |
1614 |
{ ADD_ACTIVE(state_offset + dlen + 3, 0); } |
ADD_ACTIVE(state_offset + dlen + 3, 0); |
1615 |
count = current_state->count; /* Number already matched */ |
count = current_state->count; /* Number already matched */ |
1616 |
if (clen > 0) |
if (clen > 0) |
1617 |
{ |
{ |
1618 |
int otherd = -1; |
unsigned int otherd = NOTACHAR; |
1619 |
if ((ims & PCRE_CASELESS) != 0) |
if ((ims & PCRE_CASELESS) != 0) |
1620 |
{ |
{ |
1621 |
#ifdef SUPPORT_UTF8 |
#ifdef SUPPORT_UTF8 |
1622 |
if (utf8 && c >= 128) |
if (utf8 && d >= 128) |
1623 |
{ |
{ |
1624 |
#ifdef SUPPORT_UCP |
#ifdef SUPPORT_UCP |
1625 |
if (ucp_findchar(d, &chartype, &otherd) < 0) otherd = -1; |
otherd = _pcre_ucp_othercase(d); |
1626 |
#endif /* SUPPORT_UCP */ |
#endif /* SUPPORT_UCP */ |
1627 |
} |
} |
1628 |
else |
else |
1631 |
} |
} |
1632 |
if ((c == d || c == otherd) == (codevalue < OP_NOTSTAR)) |
if ((c == d || c == otherd) == (codevalue < OP_NOTSTAR)) |
1633 |
{ |
{ |
1634 |
|
if (codevalue == OP_POSUPTO || codevalue == OP_NOTPOSUPTO) |
1635 |
|
{ |
1636 |
|
active_count--; /* Remove non-match possibility */ |
1637 |
|
next_active_state--; |
1638 |
|
} |
1639 |
if (++count >= GET2(code, 1)) |
if (++count >= GET2(code, 1)) |
1640 |
{ ADD_NEW(state_offset + dlen + 3, 0); } |
{ ADD_NEW(state_offset + dlen + 3, 0); } |
1641 |
else |
else |
1713 |
{ ADD_ACTIVE(next_state_offset + 5, 0); } |
{ ADD_ACTIVE(next_state_offset + 5, 0); } |
1714 |
if (isinclass) |
if (isinclass) |
1715 |
{ |
{ |
1716 |
if (++count >= GET2(ecode, 3)) |
int max = GET2(ecode, 3); |
1717 |
|
if (++count >= max && max != 0) /* Max 0 => no limit */ |
1718 |
{ ADD_NEW(next_state_offset + 5, 0); } |
{ ADD_NEW(next_state_offset + 5, 0); } |
1719 |
else |
else |
1720 |
{ ADD_NEW(state_offset, count); } |
{ ADD_NEW(state_offset, count); } |
1764 |
|
|
1765 |
/*-----------------------------------------------------------------*/ |
/*-----------------------------------------------------------------*/ |
1766 |
case OP_COND: |
case OP_COND: |
1767 |
|
case OP_SCOND: |
1768 |
{ |
{ |
1769 |
int local_offsets[1000]; |
int local_offsets[1000]; |
1770 |
int local_workspace[1000]; |
int local_workspace[1000]; |
1771 |
int condcode = code[LINK_SIZE+1]; |
int condcode = code[LINK_SIZE+1]; |
1772 |
|
|
1773 |
/* The only supported version of OP_CREF is for the value 0xffff, which |
/* Back reference conditions are not supported */ |
1774 |
means "test if in a recursion". */ |
|
1775 |
|
if (condcode == OP_CREF) return PCRE_ERROR_DFA_UCOND; |
1776 |
|
|
1777 |
if (condcode == OP_CREF) |
/* The DEFINE condition is always false */ |
1778 |
|
|
1779 |
|
if (condcode == OP_DEF) |
1780 |
|
{ |
1781 |
|
ADD_ACTIVE(state_offset + GET(code, 1) + LINK_SIZE + 1, 0); |
1782 |
|
} |
1783 |
|
|
1784 |
|
/* The only supported version of OP_RREF is for the value RREF_ANY, |
1785 |
|
which means "test if in any recursion". We can't test for specifically |
1786 |
|
recursed groups. */ |
1787 |
|
|
1788 |
|
else if (condcode == OP_RREF) |
1789 |
{ |
{ |
1790 |
int value = GET2(code, LINK_SIZE+2); |
int value = GET2(code, LINK_SIZE+2); |
1791 |
if (value != 0xffff) return PCRE_ERROR_DFA_UCOND; |
if (value != RREF_ANY) return PCRE_ERROR_DFA_UCOND; |
1792 |
if (recursing > 0) { ADD_ACTIVE(state_offset + LINK_SIZE + 4, 0); } |
if (recursing > 0) { ADD_ACTIVE(state_offset + LINK_SIZE + 4, 0); } |
1793 |
else { ADD_ACTIVE(state_offset + GET(code, 1) + LINK_SIZE + 1, 0); } |
else { ADD_ACTIVE(state_offset + GET(code, 1) + LINK_SIZE + 1, 0); } |
1794 |
} |
} |
1884 |
/*-----------------------------------------------------------------*/ |
/*-----------------------------------------------------------------*/ |
1885 |
case OP_ONCE: |
case OP_ONCE: |
1886 |
{ |
{ |
|
const uschar *endcode; |
|
1887 |
int local_offsets[2]; |
int local_offsets[2]; |
1888 |
int local_workspace[1000]; |
int local_workspace[1000]; |
1889 |
|
|
1905 |
const uschar *end_subpattern = code; |
const uschar *end_subpattern = code; |
1906 |
int charcount = local_offsets[1] - local_offsets[0]; |
int charcount = local_offsets[1] - local_offsets[0]; |
1907 |
int next_state_offset, repeat_state_offset; |
int next_state_offset, repeat_state_offset; |
|
BOOL is_repeated; |
|
1908 |
|
|
1909 |
do { end_subpattern += GET(end_subpattern, 1); } |
do { end_subpattern += GET(end_subpattern, 1); } |
1910 |
while (*end_subpattern == OP_ALT); |
while (*end_subpattern == OP_ALT); |
1979 |
cb.version = 1; /* Version 1 of the callout block */ |
cb.version = 1; /* Version 1 of the callout block */ |
1980 |
cb.callout_number = code[1]; |
cb.callout_number = code[1]; |
1981 |
cb.offset_vector = offsets; |
cb.offset_vector = offsets; |
1982 |
cb.subject = (char *)start_subject; |
cb.subject = (PCRE_SPTR)start_subject; |
1983 |
cb.subject_length = end_subject - start_subject; |
cb.subject_length = end_subject - start_subject; |
1984 |
cb.start_match = current_subject - start_subject; |
cb.start_match = current_subject - start_subject; |
1985 |
cb.current_position = ptr - start_subject; |
cb.current_position = ptr - start_subject; |
2027 |
DPRINTF(("%.*sEnd of internal_dfa_exec %d: returning %d\n" |
DPRINTF(("%.*sEnd of internal_dfa_exec %d: returning %d\n" |
2028 |
"%.*s---------------------\n\n", rlevel*2-2, SP, rlevel, match_count, |
"%.*s---------------------\n\n", rlevel*2-2, SP, rlevel, match_count, |
2029 |
rlevel*2-2, SP)); |
rlevel*2-2, SP)); |
2030 |
return match_count; |
break; /* In effect, "return", but see the comment below */ |
2031 |
} |
} |
2032 |
|
|
2033 |
/* One or more states are active for the next character. */ |
/* One or more states are active for the next character. */ |
2035 |
ptr += clen; /* Advance to next subject character */ |
ptr += clen; /* Advance to next subject character */ |
2036 |
} /* Loop to move along the subject string */ |
} /* Loop to move along the subject string */ |
2037 |
|
|
2038 |
/* Control never gets here, but we must keep the compiler happy. */ |
/* Control gets here from "break" a few lines above. We do it this way because |
2039 |
|
if we use "return" above, we have compiler trouble. Some compilers warn if |
2040 |
|
there's nothing here because they think the function doesn't return a value. On |
2041 |
|
the other hand, if we put a dummy statement here, some more clever compilers |
2042 |
|
complain that it can't be reached. Sigh. */ |
2043 |
|
|
2044 |
DPRINTF(("%.*s+++ Unexpected end of internal_dfa_exec %d +++\n" |
return match_count; |
|
"%.*s---------------------\n\n", rlevel*2-2, SP, rlevel, rlevel*2-2, SP)); |
|
|
return PCRE_ERROR_NOMATCH; |
|
2045 |
} |
} |
2046 |
|
|
2047 |
|
|
2057 |
|
|
2058 |
Arguments: |
Arguments: |
2059 |
argument_re points to the compiled expression |
argument_re points to the compiled expression |
2060 |
extra_data points to extra data or is NULL (not currently used) |
extra_data points to extra data or is NULL |
2061 |
subject points to the subject string |
subject points to the subject string |
2062 |
length length of subject string (may contain binary zeros) |
length length of subject string (may contain binary zeros) |
2063 |
start_offset where to start in the subject string |
start_offset where to start in the subject string |
2073 |
< -1 => some kind of unexpected problem |
< -1 => some kind of unexpected problem |
2074 |
*/ |
*/ |
2075 |
|
|
2076 |
EXPORT int |
PCRE_DATA_SCOPE int |
2077 |
pcre_dfa_exec(const pcre *argument_re, const pcre_extra *extra_data, |
pcre_dfa_exec(const pcre *argument_re, const pcre_extra *extra_data, |
2078 |
const char *subject, int length, int start_offset, int options, int *offsets, |
const char *subject, int length, int start_offset, int options, int *offsets, |
2079 |
int offsetcount, int *workspace, int wscount) |
int offsetcount, int *workspace, int wscount) |
2080 |
{ |
{ |
2081 |
real_pcre *re = (real_pcre *)argument_re; |
real_pcre *re = (real_pcre *)argument_re; |
2082 |
dfa_match_data match_block; |
dfa_match_data match_block; |
2083 |
|
dfa_match_data *md = &match_block; |
2084 |
BOOL utf8, anchored, startline, firstline; |
BOOL utf8, anchored, startline, firstline; |
2085 |
const uschar *current_subject, *end_subject, *lcc; |
const uschar *current_subject, *end_subject, *lcc; |
2086 |
|
|
2095 |
int first_byte = -1; |
int first_byte = -1; |
2096 |
int req_byte = -1; |
int req_byte = -1; |
2097 |
int req_byte2 = -1; |
int req_byte2 = -1; |
2098 |
|
int newline; |
2099 |
|
|
2100 |
/* Plausibility checks */ |
/* Plausibility checks */ |
2101 |
|
|
2110 |
match block, so we must initialize them beforehand. However, the other fields |
match block, so we must initialize them beforehand. However, the other fields |
2111 |
in the match block must not be set until after the byte flipping. */ |
in the match block must not be set until after the byte flipping. */ |
2112 |
|
|
2113 |
match_block.tables = re->tables; |
md->tables = re->tables; |
2114 |
match_block.callout_data = NULL; |
md->callout_data = NULL; |
2115 |
|
|
2116 |
if (extra_data != NULL) |
if (extra_data != NULL) |
2117 |
{ |
{ |
2119 |
if ((flags & PCRE_EXTRA_STUDY_DATA) != 0) |
if ((flags & PCRE_EXTRA_STUDY_DATA) != 0) |
2120 |
study = (const pcre_study_data *)extra_data->study_data; |
study = (const pcre_study_data *)extra_data->study_data; |
2121 |
if ((flags & PCRE_EXTRA_MATCH_LIMIT) != 0) return PCRE_ERROR_DFA_UMLIMIT; |
if ((flags & PCRE_EXTRA_MATCH_LIMIT) != 0) return PCRE_ERROR_DFA_UMLIMIT; |
2122 |
|
if ((flags & PCRE_EXTRA_MATCH_LIMIT_RECURSION) != 0) |
2123 |
|
return PCRE_ERROR_DFA_UMLIMIT; |
2124 |
if ((flags & PCRE_EXTRA_CALLOUT_DATA) != 0) |
if ((flags & PCRE_EXTRA_CALLOUT_DATA) != 0) |
2125 |
match_block.callout_data = extra_data->callout_data; |
md->callout_data = extra_data->callout_data; |
2126 |
if ((flags & PCRE_EXTRA_TABLES) != 0) |
if ((flags & PCRE_EXTRA_TABLES) != 0) |
2127 |
match_block.tables = extra_data->tables; |
md->tables = extra_data->tables; |
2128 |
} |
} |
2129 |
|
|
2130 |
/* Check that the first field in the block is the magic number. If it is not, |
/* Check that the first field in the block is the magic number. If it is not, |
2145 |
end_subject = (const unsigned char *)subject + length; |
end_subject = (const unsigned char *)subject + length; |
2146 |
req_byte_ptr = current_subject - 1; |
req_byte_ptr = current_subject - 1; |
2147 |
|
|
2148 |
|
#ifdef SUPPORT_UTF8 |
2149 |
utf8 = (re->options & PCRE_UTF8) != 0; |
utf8 = (re->options & PCRE_UTF8) != 0; |
2150 |
anchored = (options & PCRE_ANCHORED) != 0 || (re->options & PCRE_ANCHORED) != 0; |
#else |
2151 |
|
utf8 = FALSE; |
2152 |
|
#endif |
2153 |
|
|
2154 |
|
anchored = (options & (PCRE_ANCHORED|PCRE_DFA_RESTART)) != 0 || |
2155 |
|
(re->options & PCRE_ANCHORED) != 0; |
2156 |
|
|
2157 |
/* The remaining fixed data for passing around. */ |
/* The remaining fixed data for passing around. */ |
2158 |
|
|
2159 |
match_block.start_code = (const uschar *)argument_re + |
md->start_code = (const uschar *)argument_re + |
2160 |
re->name_table_offset + re->name_count * re->name_entry_size; |
re->name_table_offset + re->name_count * re->name_entry_size; |
2161 |
match_block.start_subject = (const unsigned char *)subject; |
md->start_subject = (const unsigned char *)subject; |
2162 |
match_block.end_subject = end_subject; |
md->end_subject = end_subject; |
2163 |
match_block.moptions = options; |
md->moptions = options; |
2164 |
match_block.poptions = re->options; |
md->poptions = re->options; |
2165 |
|
|
2166 |
|
/* Handle different types of newline. The three bits give eight cases. If |
2167 |
|
nothing is set at run time, whatever was used at compile time applies. */ |
2168 |
|
|
2169 |
|
switch ((((options & PCRE_NEWLINE_BITS) == 0)? re->options : options) & |
2170 |
|
PCRE_NEWLINE_BITS) |
2171 |
|
{ |
2172 |
|
case 0: newline = NEWLINE; break; /* Compile-time default */ |
2173 |
|
case PCRE_NEWLINE_CR: newline = '\r'; break; |
2174 |
|
case PCRE_NEWLINE_LF: newline = '\n'; break; |
2175 |
|
case PCRE_NEWLINE_CR+ |
2176 |
|
PCRE_NEWLINE_LF: newline = ('\r' << 8) | '\n'; break; |
2177 |
|
case PCRE_NEWLINE_ANY: newline = -1; break; |
2178 |
|
default: return PCRE_ERROR_BADNEWLINE; |
2179 |
|
} |
2180 |
|
|
2181 |
|
if (newline < 0) |
2182 |
|
{ |
2183 |
|
md->nltype = NLTYPE_ANY; |
2184 |
|
} |
2185 |
|
else |
2186 |
|
{ |
2187 |
|
md->nltype = NLTYPE_FIXED; |
2188 |
|
if (newline > 255) |
2189 |
|
{ |
2190 |
|
md->nllen = 2; |
2191 |
|
md->nl[0] = (newline >> 8) & 255; |
2192 |
|
md->nl[1] = newline & 255; |
2193 |
|
} |
2194 |
|
else |
2195 |
|
{ |
2196 |
|
md->nllen = 1; |
2197 |
|
md->nl[0] = newline; |
2198 |
|
} |
2199 |
|
} |
2200 |
|
|
2201 |
/* Check a UTF-8 string if required. Unfortunately there's no way of passing |
/* Check a UTF-8 string if required. Unfortunately there's no way of passing |
2202 |
back the character offset. */ |
back the character offset. */ |
2222 |
is a feature that makes it possible to save compiled regex and re-use them |
is a feature that makes it possible to save compiled regex and re-use them |
2223 |
in other programs later. */ |
in other programs later. */ |
2224 |
|
|
2225 |
if (match_block.tables == NULL) match_block.tables = _pcre_default_tables; |
if (md->tables == NULL) md->tables = _pcre_default_tables; |
2226 |
|
|
2227 |
/* The lower casing table and the "must be at the start of a line" flag are |
/* The lower casing table and the "must be at the start of a line" flag are |
2228 |
used in a loop when finding where to start. */ |
used in a loop when finding where to start. */ |
2229 |
|
|
2230 |
lcc = match_block.tables + lcc_offset; |
lcc = md->tables + lcc_offset; |
2231 |
startline = (re->options & PCRE_STARTLINE) != 0; |
startline = (re->options & PCRE_STARTLINE) != 0; |
2232 |
firstline = (re->options & PCRE_FIRSTLINE) != 0; |
firstline = (re->options & PCRE_FIRSTLINE) != 0; |
2233 |
|
|
2260 |
{ |
{ |
2261 |
req_byte = re->req_byte & 255; |
req_byte = re->req_byte & 255; |
2262 |
req_byte_caseless = (re->req_byte & REQ_CASELESS) != 0; |
req_byte_caseless = (re->req_byte & REQ_CASELESS) != 0; |
2263 |
req_byte2 = (match_block.tables + fcc_offset)[req_byte]; /* case flipped */ |
req_byte2 = (md->tables + fcc_offset)[req_byte]; /* case flipped */ |
2264 |
} |
} |
2265 |
|
|
2266 |
/* Call the main matching function, looping for a non-anchored regex after a |
/* Call the main matching function, looping for a non-anchored regex after a |
2278 |
|
|
2279 |
/* Advance to a unique first char if possible. If firstline is TRUE, the |
/* Advance to a unique first char if possible. If firstline is TRUE, the |
2280 |
start of the match is constrained to the first line of a multiline string. |
start of the match is constrained to the first line of a multiline string. |
2281 |
Implement this by temporarily adjusting end_subject so that we stop scanning |
Implement this by temporarily adjusting end_subject so that we stop |
2282 |
at a newline. If the match fails at the newline, later code breaks this loop. |
scanning at a newline. If the match fails at the newline, later code breaks |
2283 |
*/ |
this loop. */ |
2284 |
|
|
2285 |
if (firstline) |
if (firstline) |
2286 |
{ |
{ |
2287 |
const uschar *t = current_subject; |
const uschar *t = current_subject; |
2288 |
while (t < save_end_subject && *t != '\n') t++; |
while (t < md->end_subject && !IS_NEWLINE(t)) t++; |
2289 |
end_subject = t; |
end_subject = t; |
2290 |
} |
} |
2291 |
|
|
2300 |
current_subject++; |
current_subject++; |
2301 |
} |
} |
2302 |
|
|
2303 |
/* Or to just after \n for a multiline match if possible */ |
/* Or to just after a linebreak for a multiline match if possible */ |
2304 |
|
|
2305 |
else if (startline) |
else if (startline) |
2306 |
{ |
{ |
2307 |
if (current_subject > match_block.start_subject + start_offset) |
if (current_subject > md->start_subject + start_offset) |
2308 |
{ |
{ |
2309 |
while (current_subject < end_subject && current_subject[-1] != NEWLINE) |
while (current_subject <= end_subject && !WAS_NEWLINE(current_subject)) |
2310 |
current_subject++; |
current_subject++; |
2311 |
} |
} |
2312 |
} |
} |
2387 |
/* OK, now we can do the business */ |
/* OK, now we can do the business */ |
2388 |
|
|
2389 |
rc = internal_dfa_exec( |
rc = internal_dfa_exec( |
2390 |
&match_block, /* fixed match data */ |
md, /* fixed match data */ |
2391 |
match_block.start_code, /* this subexpression's code */ |
md->start_code, /* this subexpression's code */ |
2392 |
current_subject, /* where we currently are */ |
current_subject, /* where we currently are */ |
2393 |
start_offset, /* start offset in subject */ |
start_offset, /* start offset in subject */ |
2394 |
offsets, /* offset vector */ |
offsets, /* offset vector */ |
2395 |
offsetcount, /* size of same */ |
offsetcount, /* size of same */ |
2396 |
workspace, /* workspace vector */ |
workspace, /* workspace vector */ |
2397 |
wscount, /* size of same */ |
wscount, /* size of same */ |
2398 |
re->options & (PCRE_CASELESS|PCRE_MULTILINE|PCRE_DOTALL), /* ims flags */ |
re->options & (PCRE_CASELESS|PCRE_MULTILINE|PCRE_DOTALL), /* ims flags */ |
2399 |
0, /* function recurse level */ |
0, /* function recurse level */ |
2400 |
0); /* regex recurse level */ |
0); /* regex recurse level */ |
2401 |
|
|
2402 |
/* Anything other than "no match" means we are done, always; otherwise, carry |
/* Anything other than "no match" means we are done, always; otherwise, carry |
2403 |
on only if not anchored. */ |
on only if not anchored. */ |
2407 |
/* Advance to the next subject character unless we are at the end of a line |
/* Advance to the next subject character unless we are at the end of a line |
2408 |
and firstline is set. */ |
and firstline is set. */ |
2409 |
|
|
2410 |
if (firstline && *current_subject == NEWLINE) break; |
if (firstline && IS_NEWLINE(current_subject)) break; |
2411 |
current_subject++; |
current_subject++; |
|
|
|
|
#ifdef SUPPORT_UTF8 |
|
2412 |
if (utf8) |
if (utf8) |
2413 |
{ |
{ |
2414 |
while (current_subject < end_subject && (*current_subject & 0xc0) == 0x80) |
while (current_subject < end_subject && (*current_subject & 0xc0) == 0x80) |
2415 |
current_subject++; |
current_subject++; |
2416 |
} |
} |
|
#endif |
|
|
|
|
2417 |
if (current_subject > end_subject) break; |
if (current_subject > end_subject) break; |
2418 |
} |
|
2419 |
|
/* If we have just passed a CR and the newline option is CRLF or ANY, and we |
2420 |
|
are now at a LF, advance the match position by one more character. */ |
2421 |
|
|
2422 |
|
if (current_subject[-1] == '\r' && |
2423 |
|
(md->nltype == NLTYPE_ANY || md->nllen == 2) && |
2424 |
|
current_subject < end_subject && |
2425 |
|
*current_subject == '\n') |
2426 |
|
current_subject++; |
2427 |
|
|
2428 |
|
} /* "Bumpalong" loop */ |
2429 |
|
|
2430 |
return PCRE_ERROR_NOMATCH; |
return PCRE_ERROR_NOMATCH; |
2431 |
} |
} |