3 |
*************************************************/ |
*************************************************/ |
4 |
|
|
5 |
/* PCRE is a library of functions to support regular expressions whose syntax |
/* PCRE is a library of functions to support regular expressions whose syntax |
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 (but see |
7 |
|
below for why this module is different). |
8 |
|
|
9 |
Written by Philip Hazel |
Written by Philip Hazel |
10 |
Copyright (c) 1997-2007 University of Cambridge |
Copyright (c) 1997-2009 University of Cambridge |
11 |
|
|
12 |
----------------------------------------------------------------------------- |
----------------------------------------------------------------------------- |
13 |
Redistribution and use in source and binary forms, with or without |
Redistribution and use in source and binary forms, with or without |
45 |
applications. */ |
applications. */ |
46 |
|
|
47 |
|
|
48 |
|
/* NOTE ABOUT PERFORMANCE: A user of this function sent some code that improved |
49 |
|
the performance of his patterns greatly. I could not use it as it stood, as it |
50 |
|
was not thread safe, and made assumptions about pattern sizes. Also, it caused |
51 |
|
test 7 to loop, and test 9 to crash with a segfault. |
52 |
|
|
53 |
|
The issue is the check for duplicate states, which is done by a simple linear |
54 |
|
search up the state list. (Grep for "duplicate" below to find the code.) For |
55 |
|
many patterns, there will never be many states active at one time, so a simple |
56 |
|
linear search is fine. In patterns that have many active states, it might be a |
57 |
|
bottleneck. The suggested code used an indexing scheme to remember which states |
58 |
|
had previously been used for each character, and avoided the linear search when |
59 |
|
it knew there was no chance of a duplicate. This was implemented when adding |
60 |
|
states to the state lists. |
61 |
|
|
62 |
|
I wrote some thread-safe, not-limited code to try something similar at the time |
63 |
|
of checking for duplicates (instead of when adding states), using index vectors |
64 |
|
on the stack. It did give a 13% improvement with one specially constructed |
65 |
|
pattern for certain subject strings, but on other strings and on many of the |
66 |
|
simpler patterns in the test suite it did worse. The major problem, I think, |
67 |
|
was the extra time to initialize the index. This had to be done for each call |
68 |
|
of internal_dfa_exec(). (The supplied patch used a static vector, initialized |
69 |
|
only once - I suspect this was the cause of the problems with the tests.) |
70 |
|
|
71 |
|
Overall, I concluded that the gains in some cases did not outweigh the losses |
72 |
|
in others, so I abandoned this code. */ |
73 |
|
|
74 |
|
|
75 |
|
|
76 |
|
#ifdef HAVE_CONFIG_H |
77 |
|
#include "config.h" |
78 |
|
#endif |
79 |
|
|
80 |
#define NLBLOCK md /* Block containing newline information */ |
#define NLBLOCK md /* Block containing newline information */ |
81 |
#define PSSTART start_subject /* Field containing processed string start */ |
#define PSSTART start_subject /* Field containing processed string start */ |
82 |
#define PSEND end_subject /* Field containing processed string end */ |
#define PSEND end_subject /* Field containing processed string end */ |
89 |
#define SP " " |
#define SP " " |
90 |
|
|
91 |
|
|
|
|
|
92 |
/************************************************* |
/************************************************* |
93 |
* Code parameters and static tables * |
* Code parameters and static tables * |
94 |
*************************************************/ |
*************************************************/ |
95 |
|
|
96 |
/* 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 |
97 |
into others, under special conditions. A gap of 20 between the blocks should be |
into others, under special conditions. A gap of 20 between the blocks should be |
98 |
enough. */ |
enough. The resulting opcodes don't have to be less than 256 because they are |
99 |
|
never stored, so we push them well clear of the normal opcodes. */ |
100 |
|
|
101 |
#define OP_PROP_EXTRA 100 |
#define OP_PROP_EXTRA 300 |
102 |
#define OP_EXTUNI_EXTRA 120 |
#define OP_EXTUNI_EXTRA 320 |
103 |
#define OP_ANYNL_EXTRA 140 |
#define OP_ANYNL_EXTRA 340 |
104 |
|
#define OP_HSPACE_EXTRA 360 |
105 |
|
#define OP_VSPACE_EXTRA 380 |
106 |
|
|
107 |
|
|
108 |
/* This table identifies those opcodes that are followed immediately by a |
/* This table identifies those opcodes that are followed immediately by a |
109 |
character that is to be tested in some way. This makes is possible to |
character that is to be tested in some way. This makes is possible to |
110 |
centralize the loading of these characters. In the case of Type * etc, the |
centralize the loading of these characters. In the case of Type * etc, the |
111 |
"character" is the opcode for \D, \d, \S, \s, \W, or \w, which will always be a |
"character" is the opcode for \D, \d, \S, \s, \W, or \w, which will always be a |
112 |
small value. ***NOTE*** If the start of this table is modified, the two tables |
small value. ***NOTE*** If the start of this table is modified, the two tables |
113 |
that follow must also be modified. */ |
that follow must also be modified. */ |
114 |
|
|
115 |
static uschar coptable[] = { |
static const uschar coptable[] = { |
116 |
0, /* End */ |
0, /* End */ |
117 |
0, 0, 0, 0, 0, /* \A, \G, \K, \B, \b */ |
0, 0, 0, 0, 0, /* \A, \G, \K, \B, \b */ |
118 |
0, 0, 0, 0, 0, 0, /* \D, \d, \S, \s, \W, \w */ |
0, 0, 0, 0, 0, 0, /* \D, \d, \S, \s, \W, \w */ |
119 |
0, 0, /* Any, Anybyte */ |
0, 0, 0, /* Any, AllAny, Anybyte */ |
120 |
0, 0, 0, 0, /* NOTPROP, PROP, EXTUNI, ANYNL */ |
0, 0, 0, /* NOTPROP, PROP, EXTUNI */ |
121 |
|
0, 0, 0, 0, 0, /* \R, \H, \h, \V, \v */ |
122 |
0, 0, 0, 0, 0, /* \Z, \z, Opt, ^, $ */ |
0, 0, 0, 0, 0, /* \Z, \z, Opt, ^, $ */ |
123 |
1, /* Char */ |
1, /* Char */ |
124 |
1, /* Charnc */ |
1, /* Charnc */ |
158 |
0, /* CREF */ |
0, /* CREF */ |
159 |
0, /* RREF */ |
0, /* RREF */ |
160 |
0, /* DEF */ |
0, /* DEF */ |
161 |
0, 0 /* BRAZERO, BRAMINZERO */ |
0, 0, /* BRAZERO, BRAMINZERO */ |
162 |
|
0, 0, 0, 0, /* PRUNE, SKIP, THEN, COMMIT */ |
163 |
|
0, 0, 0 /* FAIL, ACCEPT, SKIPZERO */ |
164 |
}; |
}; |
165 |
|
|
166 |
/* 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, |
167 |
and \w */ |
and \w */ |
168 |
|
|
169 |
static uschar toptable1[] = { |
static const uschar toptable1[] = { |
170 |
0, 0, 0, 0, 0, 0, |
0, 0, 0, 0, 0, 0, |
171 |
ctype_digit, ctype_digit, |
ctype_digit, ctype_digit, |
172 |
ctype_space, ctype_space, |
ctype_space, ctype_space, |
173 |
ctype_word, ctype_word, |
ctype_word, ctype_word, |
174 |
0 /* OP_ANY */ |
0, 0 /* OP_ANY, OP_ALLANY */ |
175 |
}; |
}; |
176 |
|
|
177 |
static uschar toptable2[] = { |
static const uschar toptable2[] = { |
178 |
0, 0, 0, 0, 0, 0, |
0, 0, 0, 0, 0, 0, |
179 |
ctype_digit, 0, |
ctype_digit, 0, |
180 |
ctype_space, 0, |
ctype_space, 0, |
181 |
ctype_word, 0, |
ctype_word, 0, |
182 |
1 /* OP_ANY */ |
1, 1 /* OP_ANY, OP_ALLANY */ |
183 |
}; |
}; |
184 |
|
|
185 |
|
|
251 |
rlevel function call recursion level |
rlevel function call recursion level |
252 |
recursing regex recursive call level |
recursing regex recursive call level |
253 |
|
|
254 |
Returns: > 0 => |
Returns: > 0 => number of match offset pairs placed in offsets |
255 |
= 0 => |
= 0 => offsets overflowed; longest matches are present |
256 |
-1 => failed to match |
-1 => failed to match |
257 |
< -1 => some kind of unexpected problem |
< -1 => some kind of unexpected problem |
258 |
|
|
417 |
current_subject - start_subject : max_back; |
current_subject - start_subject : max_back; |
418 |
current_subject -= gone_back; |
current_subject -= gone_back; |
419 |
} |
} |
420 |
|
|
421 |
|
/* Save the earliest consulted character */ |
422 |
|
|
423 |
|
if (current_subject < md->start_used_ptr) |
424 |
|
md->start_used_ptr = current_subject; |
425 |
|
|
426 |
/* Now we can process the individual branches. */ |
/* Now we can process the individual branches. */ |
427 |
|
|
487 |
int i, j; |
int i, j; |
488 |
int clen, dlen; |
int clen, dlen; |
489 |
unsigned int c, d; |
unsigned int c, d; |
490 |
|
int forced_fail = 0; |
491 |
|
int reached_end = 0; |
492 |
|
|
493 |
/* 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 |
494 |
new state list. */ |
new state list. */ |
546 |
stateblock *current_state = active_states + i; |
stateblock *current_state = active_states + i; |
547 |
const uschar *code; |
const uschar *code; |
548 |
int state_offset = current_state->offset; |
int state_offset = current_state->offset; |
549 |
int count, codevalue; |
int count, codevalue, rrc; |
|
#ifdef SUPPORT_UCP |
|
|
int chartype, script; |
|
|
#endif |
|
550 |
|
|
551 |
#ifdef DEBUG |
#ifdef DEBUG |
552 |
printf ("%.*sProcessing state %d c=", rlevel*2-2, SP, state_offset); |
printf ("%.*sProcessing state %d c=", rlevel*2-2, SP, state_offset); |
578 |
} |
} |
579 |
} |
} |
580 |
|
|
581 |
/* Check for a duplicate state with the same count, and skip if found. */ |
/* Check for a duplicate state with the same count, and skip if found. |
582 |
|
See the note at the head of this module about the possibility of improving |
583 |
|
performance here. */ |
584 |
|
|
585 |
for (j = 0; j < i; j++) |
for (j = 0; j < i; j++) |
586 |
{ |
{ |
603 |
permitted. |
permitted. |
604 |
|
|
605 |
We also use this mechanism for opcodes such as OP_TYPEPLUS that take an |
We also use this mechanism for opcodes such as OP_TYPEPLUS that take an |
606 |
argument that is not a data character - but is always one byte long. |
argument that is not a data character - but is always one byte long. We |
607 |
Unfortunately, we have to take special action to deal with \P, \p, and |
have to take special action to deal with \P, \p, \H, \h, \V, \v and \X in |
608 |
\X in this case. To keep the other cases fast, convert these ones to new |
this case. To keep the other cases fast, convert these ones to new opcodes. |
609 |
opcodes. */ |
*/ |
610 |
|
|
611 |
if (coptable[codevalue] > 0) |
if (coptable[codevalue] > 0) |
612 |
{ |
{ |
624 |
case OP_PROP: codevalue += OP_PROP_EXTRA; break; |
case OP_PROP: codevalue += OP_PROP_EXTRA; break; |
625 |
case OP_ANYNL: codevalue += OP_ANYNL_EXTRA; break; |
case OP_ANYNL: codevalue += OP_ANYNL_EXTRA; break; |
626 |
case OP_EXTUNI: codevalue += OP_EXTUNI_EXTRA; break; |
case OP_EXTUNI: codevalue += OP_EXTUNI_EXTRA; break; |
627 |
|
case OP_NOT_HSPACE: |
628 |
|
case OP_HSPACE: codevalue += OP_HSPACE_EXTRA; break; |
629 |
|
case OP_NOT_VSPACE: |
630 |
|
case OP_VSPACE: codevalue += OP_VSPACE_EXTRA; break; |
631 |
default: break; |
default: break; |
632 |
} |
} |
633 |
} |
} |
647 |
/* ========================================================================== */ |
/* ========================================================================== */ |
648 |
/* Reached a closing bracket. If not at the end of the pattern, carry |
/* Reached a closing bracket. If not at the end of the pattern, carry |
649 |
on with the next opcode. Otherwise, unless we have an empty string and |
on with the next opcode. Otherwise, unless we have an empty string and |
650 |
PCRE_NOTEMPTY is set, save the match data, shifting up all previous |
PCRE_NOTEMPTY is set, or PCRE_NOTEMPTY_ATSTART is set and we are at the |
651 |
|
start of the subject, save the match data, shifting up all previous |
652 |
matches so we always have the longest first. */ |
matches so we always have the longest first. */ |
653 |
|
|
654 |
case OP_KET: |
case OP_KET: |
662 |
ADD_ACTIVE(state_offset - GET(code, 1), 0); |
ADD_ACTIVE(state_offset - GET(code, 1), 0); |
663 |
} |
} |
664 |
} |
} |
665 |
else if (ptr > current_subject || (md->moptions & PCRE_NOTEMPTY) == 0) |
else |
666 |
{ |
{ |
667 |
if (match_count < 0) match_count = (offsetcount >= 2)? 1 : 0; |
reached_end++; /* Count branches that reach the end */ |
668 |
else if (match_count > 0 && ++match_count * 2 >= offsetcount) |
if (ptr > current_subject || |
669 |
match_count = 0; |
((md->moptions & PCRE_NOTEMPTY) == 0 && |
670 |
count = ((match_count == 0)? offsetcount : match_count * 2) - 2; |
((md->moptions & PCRE_NOTEMPTY_ATSTART) == 0 || |
671 |
if (count > 0) memmove(offsets + 2, offsets, count * sizeof(int)); |
current_subject > start_subject + md->start_offset))) |
672 |
if (offsetcount >= 2) |
{ |
673 |
{ |
if (match_count < 0) match_count = (offsetcount >= 2)? 1 : 0; |
674 |
offsets[0] = current_subject - start_subject; |
else if (match_count > 0 && ++match_count * 2 >= offsetcount) |
675 |
offsets[1] = ptr - start_subject; |
match_count = 0; |
676 |
DPRINTF(("%.*sSet matched string = \"%.*s\"\n", rlevel*2-2, SP, |
count = ((match_count == 0)? offsetcount : match_count * 2) - 2; |
677 |
offsets[1] - offsets[0], current_subject)); |
if (count > 0) memmove(offsets + 2, offsets, count * sizeof(int)); |
678 |
} |
if (offsetcount >= 2) |
679 |
if ((md->moptions & PCRE_DFA_SHORTEST) != 0) |
{ |
680 |
{ |
offsets[0] = current_subject - start_subject; |
681 |
DPRINTF(("%.*sEnd of internal_dfa_exec %d: returning %d\n" |
offsets[1] = ptr - start_subject; |
682 |
"%.*s---------------------\n\n", rlevel*2-2, SP, rlevel, |
DPRINTF(("%.*sSet matched string = \"%.*s\"\n", rlevel*2-2, SP, |
683 |
match_count, rlevel*2-2, SP)); |
offsets[1] - offsets[0], current_subject)); |
684 |
return match_count; |
} |
685 |
} |
if ((md->moptions & PCRE_DFA_SHORTEST) != 0) |
686 |
|
{ |
687 |
|
DPRINTF(("%.*sEnd of internal_dfa_exec %d: returning %d\n" |
688 |
|
"%.*s---------------------\n\n", rlevel*2-2, SP, rlevel, |
689 |
|
match_count, rlevel*2-2, SP)); |
690 |
|
return match_count; |
691 |
|
} |
692 |
|
} |
693 |
} |
} |
694 |
break; |
break; |
695 |
|
|
736 |
break; |
break; |
737 |
|
|
738 |
/*-----------------------------------------------------------------*/ |
/*-----------------------------------------------------------------*/ |
739 |
|
case OP_SKIPZERO: |
740 |
|
code += 1 + GET(code, 2); |
741 |
|
while (*code == OP_ALT) code += GET(code, 1); |
742 |
|
ADD_ACTIVE(code - start_code + 1 + LINK_SIZE, 0); |
743 |
|
break; |
744 |
|
|
745 |
|
/*-----------------------------------------------------------------*/ |
746 |
case OP_CIRC: |
case OP_CIRC: |
747 |
if ((ptr == start_subject && (md->moptions & PCRE_NOTBOL) == 0) || |
if ((ptr == start_subject && (md->moptions & PCRE_NOTBOL) == 0) || |
748 |
((ims & PCRE_MULTILINE) != 0 && |
((ims & PCRE_MULTILINE) != 0 && |
781 |
|
|
782 |
/*-----------------------------------------------------------------*/ |
/*-----------------------------------------------------------------*/ |
783 |
case OP_ANY: |
case OP_ANY: |
784 |
if (clen > 0 && ((ims & PCRE_DOTALL) != 0 || !IS_NEWLINE(ptr))) |
if (clen > 0 && !IS_NEWLINE(ptr)) |
785 |
|
{ ADD_NEW(state_offset + 1, 0); } |
786 |
|
break; |
787 |
|
|
788 |
|
/*-----------------------------------------------------------------*/ |
789 |
|
case OP_ALLANY: |
790 |
|
if (clen > 0) |
791 |
{ ADD_NEW(state_offset + 1, 0); } |
{ ADD_NEW(state_offset + 1, 0); } |
792 |
break; |
break; |
793 |
|
|
802 |
if ((md->moptions & PCRE_NOTEOL) == 0) |
if ((md->moptions & PCRE_NOTEOL) == 0) |
803 |
{ |
{ |
804 |
if (clen == 0 || |
if (clen == 0 || |
805 |
(IS_NEWLINE(ptr) && |
((md->poptions & PCRE_DOLLAR_ENDONLY) == 0 && IS_NEWLINE(ptr) && |
806 |
((ims & PCRE_MULTILINE) != 0 || ptr == end_subject - md->nllen) |
((ims & PCRE_MULTILINE) != 0 || ptr == end_subject - md->nllen) |
807 |
)) |
)) |
808 |
{ ADD_ACTIVE(state_offset + 1, 0); } |
{ ADD_ACTIVE(state_offset + 1, 0); } |
839 |
if (ptr > start_subject) |
if (ptr > start_subject) |
840 |
{ |
{ |
841 |
const uschar *temp = ptr - 1; |
const uschar *temp = ptr - 1; |
842 |
|
if (temp < md->start_used_ptr) md->start_used_ptr = temp; |
843 |
#ifdef SUPPORT_UTF8 |
#ifdef SUPPORT_UTF8 |
844 |
if (utf8) BACKCHAR(temp); |
if (utf8) BACKCHAR(temp); |
845 |
#endif |
#endif |
848 |
} |
} |
849 |
else left_word = 0; |
else left_word = 0; |
850 |
|
|
851 |
if (clen > 0) right_word = c < 256 && (ctypes[c] & ctype_word) != 0; |
if (clen > 0) |
852 |
else right_word = 0; |
right_word = c < 256 && (ctypes[c] & ctype_word) != 0; |
853 |
|
else /* This is a fudge to ensure that if this is the */ |
854 |
|
{ /* last item in the pattern, we don't count it as */ |
855 |
|
reached_end--; /* reached, thus disabling a partial match. */ |
856 |
|
right_word = 0; |
857 |
|
} |
858 |
|
|
859 |
if ((left_word == right_word) == (codevalue == OP_NOT_WORD_BOUNDARY)) |
if ((left_word == right_word) == (codevalue == OP_NOT_WORD_BOUNDARY)) |
860 |
{ ADD_ACTIVE(state_offset + 1, 0); } |
{ ADD_ACTIVE(state_offset + 1, 0); } |
873 |
if (clen > 0) |
if (clen > 0) |
874 |
{ |
{ |
875 |
BOOL OK; |
BOOL OK; |
876 |
int category = _pcre_ucp_findprop(c, &chartype, &script); |
const ucd_record * prop = GET_UCD(c); |
877 |
switch(code[1]) |
switch(code[1]) |
878 |
{ |
{ |
879 |
case PT_ANY: |
case PT_ANY: |
881 |
break; |
break; |
882 |
|
|
883 |
case PT_LAMP: |
case PT_LAMP: |
884 |
OK = chartype == ucp_Lu || chartype == ucp_Ll || chartype == ucp_Lt; |
OK = prop->chartype == ucp_Lu || prop->chartype == ucp_Ll || prop->chartype == ucp_Lt; |
885 |
break; |
break; |
886 |
|
|
887 |
case PT_GC: |
case PT_GC: |
888 |
OK = category == code[2]; |
OK = _pcre_ucp_gentype[prop->chartype] == code[2]; |
889 |
break; |
break; |
890 |
|
|
891 |
case PT_PC: |
case PT_PC: |
892 |
OK = chartype == code[2]; |
OK = prop->chartype == code[2]; |
893 |
break; |
break; |
894 |
|
|
895 |
case PT_SC: |
case PT_SC: |
896 |
OK = script == code[2]; |
OK = prop->script == code[2]; |
897 |
break; |
break; |
898 |
|
|
899 |
/* Should never occur, but keep compilers from grumbling. */ |
/* Should never occur, but keep compilers from grumbling. */ |
913 |
/* ========================================================================== */ |
/* ========================================================================== */ |
914 |
/* These opcodes likewise inspect the subject character, but have an |
/* These opcodes likewise inspect the subject character, but have an |
915 |
argument that is not a data character. It is one of these opcodes: |
argument that is not a data character. It is one of these opcodes: |
916 |
OP_ANY, OP_DIGIT, OP_NOT_DIGIT, OP_WHITESPACE, OP_NOT_SPACE, OP_WORDCHAR, |
OP_ANY, OP_ALLANY, OP_DIGIT, OP_NOT_DIGIT, OP_WHITESPACE, OP_NOT_SPACE, |
917 |
OP_NOT_WORDCHAR. The value is loaded into d. */ |
OP_WORDCHAR, OP_NOT_WORDCHAR. The value is loaded into d. */ |
918 |
|
|
919 |
case OP_TYPEPLUS: |
case OP_TYPEPLUS: |
920 |
case OP_TYPEMINPLUS: |
case OP_TYPEMINPLUS: |
925 |
{ |
{ |
926 |
if ((c >= 256 && d != OP_DIGIT && d != OP_WHITESPACE && d != OP_WORDCHAR) || |
if ((c >= 256 && d != OP_DIGIT && d != OP_WHITESPACE && d != OP_WORDCHAR) || |
927 |
(c < 256 && |
(c < 256 && |
928 |
(d != OP_ANY || |
(d != OP_ANY || !IS_NEWLINE(ptr)) && |
|
(ims & PCRE_DOTALL) != 0 || |
|
|
!IS_NEWLINE(ptr) |
|
|
) && |
|
929 |
((ctypes[c] & toptable1[d]) ^ toptable2[d]) != 0)) |
((ctypes[c] & toptable1[d]) ^ toptable2[d]) != 0)) |
930 |
{ |
{ |
931 |
if (count > 0 && codevalue == OP_TYPEPOSPLUS) |
if (count > 0 && codevalue == OP_TYPEPOSPLUS) |
948 |
{ |
{ |
949 |
if ((c >= 256 && d != OP_DIGIT && d != OP_WHITESPACE && d != OP_WORDCHAR) || |
if ((c >= 256 && d != OP_DIGIT && d != OP_WHITESPACE && d != OP_WORDCHAR) || |
950 |
(c < 256 && |
(c < 256 && |
951 |
(d != OP_ANY || |
(d != OP_ANY || !IS_NEWLINE(ptr)) && |
|
(ims & PCRE_DOTALL) != 0 || |
|
|
!IS_NEWLINE(ptr) |
|
|
) && |
|
952 |
((ctypes[c] & toptable1[d]) ^ toptable2[d]) != 0)) |
((ctypes[c] & toptable1[d]) ^ toptable2[d]) != 0)) |
953 |
{ |
{ |
954 |
if (codevalue == OP_TYPEPOSQUERY) |
if (codevalue == OP_TYPEPOSQUERY) |
970 |
{ |
{ |
971 |
if ((c >= 256 && d != OP_DIGIT && d != OP_WHITESPACE && d != OP_WORDCHAR) || |
if ((c >= 256 && d != OP_DIGIT && d != OP_WHITESPACE && d != OP_WORDCHAR) || |
972 |
(c < 256 && |
(c < 256 && |
973 |
(d != OP_ANY || |
(d != OP_ANY || !IS_NEWLINE(ptr)) && |
|
(ims & PCRE_DOTALL) != 0 || |
|
|
!IS_NEWLINE(ptr) |
|
|
) && |
|
974 |
((ctypes[c] & toptable1[d]) ^ toptable2[d]) != 0)) |
((ctypes[c] & toptable1[d]) ^ toptable2[d]) != 0)) |
975 |
{ |
{ |
976 |
if (codevalue == OP_TYPEPOSSTAR) |
if (codevalue == OP_TYPEPOSSTAR) |
990 |
{ |
{ |
991 |
if ((c >= 256 && d != OP_DIGIT && d != OP_WHITESPACE && d != OP_WORDCHAR) || |
if ((c >= 256 && d != OP_DIGIT && d != OP_WHITESPACE && d != OP_WORDCHAR) || |
992 |
(c < 256 && |
(c < 256 && |
993 |
(d != OP_ANY || |
(d != OP_ANY || !IS_NEWLINE(ptr)) && |
|
(ims & PCRE_DOTALL) != 0 || |
|
|
!IS_NEWLINE(ptr) |
|
|
) && |
|
994 |
((ctypes[c] & toptable1[d]) ^ toptable2[d]) != 0)) |
((ctypes[c] & toptable1[d]) ^ toptable2[d]) != 0)) |
995 |
{ |
{ |
996 |
if (++count >= GET2(code, 1)) |
if (++count >= GET2(code, 1)) |
1011 |
{ |
{ |
1012 |
if ((c >= 256 && d != OP_DIGIT && d != OP_WHITESPACE && d != OP_WORDCHAR) || |
if ((c >= 256 && d != OP_DIGIT && d != OP_WHITESPACE && d != OP_WORDCHAR) || |
1013 |
(c < 256 && |
(c < 256 && |
1014 |
(d != OP_ANY || |
(d != OP_ANY || !IS_NEWLINE(ptr)) && |
|
(ims & PCRE_DOTALL) != 0 || |
|
|
!IS_NEWLINE(ptr) |
|
|
) && |
|
1015 |
((ctypes[c] & toptable1[d]) ^ toptable2[d]) != 0)) |
((ctypes[c] & toptable1[d]) ^ toptable2[d]) != 0)) |
1016 |
{ |
{ |
1017 |
if (codevalue == OP_TYPEPOSUPTO) |
if (codevalue == OP_TYPEPOSUPTO) |
1042 |
if (clen > 0) |
if (clen > 0) |
1043 |
{ |
{ |
1044 |
BOOL OK; |
BOOL OK; |
1045 |
int category = _pcre_ucp_findprop(c, &chartype, &script); |
const ucd_record * prop = GET_UCD(c); |
1046 |
switch(code[2]) |
switch(code[2]) |
1047 |
{ |
{ |
1048 |
case PT_ANY: |
case PT_ANY: |
1050 |
break; |
break; |
1051 |
|
|
1052 |
case PT_LAMP: |
case PT_LAMP: |
1053 |
OK = chartype == ucp_Lu || chartype == ucp_Ll || chartype == ucp_Lt; |
OK = prop->chartype == ucp_Lu || prop->chartype == ucp_Ll || prop->chartype == ucp_Lt; |
1054 |
break; |
break; |
1055 |
|
|
1056 |
case PT_GC: |
case PT_GC: |
1057 |
OK = category == code[3]; |
OK = _pcre_ucp_gentype[prop->chartype] == code[3]; |
1058 |
break; |
break; |
1059 |
|
|
1060 |
case PT_PC: |
case PT_PC: |
1061 |
OK = chartype == code[3]; |
OK = prop->chartype == code[3]; |
1062 |
break; |
break; |
1063 |
|
|
1064 |
case PT_SC: |
case PT_SC: |
1065 |
OK = script == code[3]; |
OK = prop->script == code[3]; |
1066 |
break; |
break; |
1067 |
|
|
1068 |
/* Should never occur, but keep compilers from grumbling. */ |
/* Should never occur, but keep compilers from grumbling. */ |
1091 |
case OP_EXTUNI_EXTRA + OP_TYPEPOSPLUS: |
case OP_EXTUNI_EXTRA + OP_TYPEPOSPLUS: |
1092 |
count = current_state->count; /* Already matched */ |
count = current_state->count; /* Already matched */ |
1093 |
if (count > 0) { ADD_ACTIVE(state_offset + 2, 0); } |
if (count > 0) { ADD_ACTIVE(state_offset + 2, 0); } |
1094 |
if (clen > 0 && _pcre_ucp_findprop(c, &chartype, &script) != ucp_M) |
if (clen > 0 && UCD_CATEGORY(c) != ucp_M) |
1095 |
{ |
{ |
1096 |
const uschar *nptr = ptr + clen; |
const uschar *nptr = ptr + clen; |
1097 |
int ncount = 0; |
int ncount = 0; |
1105 |
int nd; |
int nd; |
1106 |
int ndlen = 1; |
int ndlen = 1; |
1107 |
GETCHARLEN(nd, nptr, ndlen); |
GETCHARLEN(nd, nptr, ndlen); |
1108 |
if (_pcre_ucp_findprop(nd, &chartype, &script) != ucp_M) break; |
if (UCD_CATEGORY(nd) != ucp_M) break; |
1109 |
ncount++; |
ncount++; |
1110 |
nptr += ndlen; |
nptr += ndlen; |
1111 |
} |
} |
1126 |
int ncount = 0; |
int ncount = 0; |
1127 |
switch (c) |
switch (c) |
1128 |
{ |
{ |
1129 |
|
case 0x000b: |
1130 |
|
case 0x000c: |
1131 |
|
case 0x0085: |
1132 |
|
case 0x2028: |
1133 |
|
case 0x2029: |
1134 |
|
if ((md->moptions & PCRE_BSR_ANYCRLF) != 0) break; |
1135 |
|
goto ANYNL01; |
1136 |
|
|
1137 |
case 0x000d: |
case 0x000d: |
1138 |
if (ptr + 1 < end_subject && ptr[1] == 0x0a) ncount = 1; |
if (ptr + 1 < end_subject && ptr[1] == 0x0a) ncount = 1; |
1139 |
/* Fall through */ |
/* Fall through */ |
1140 |
|
|
1141 |
|
ANYNL01: |
1142 |
|
case 0x000a: |
1143 |
|
if (count > 0 && codevalue == OP_ANYNL_EXTRA + OP_TYPEPOSPLUS) |
1144 |
|
{ |
1145 |
|
active_count--; /* Remove non-match possibility */ |
1146 |
|
next_active_state--; |
1147 |
|
} |
1148 |
|
count++; |
1149 |
|
ADD_NEW_DATA(-state_offset, count, ncount); |
1150 |
|
break; |
1151 |
|
|
1152 |
|
default: |
1153 |
|
break; |
1154 |
|
} |
1155 |
|
} |
1156 |
|
break; |
1157 |
|
|
1158 |
|
/*-----------------------------------------------------------------*/ |
1159 |
|
case OP_VSPACE_EXTRA + OP_TYPEPLUS: |
1160 |
|
case OP_VSPACE_EXTRA + OP_TYPEMINPLUS: |
1161 |
|
case OP_VSPACE_EXTRA + OP_TYPEPOSPLUS: |
1162 |
|
count = current_state->count; /* Already matched */ |
1163 |
|
if (count > 0) { ADD_ACTIVE(state_offset + 2, 0); } |
1164 |
|
if (clen > 0) |
1165 |
|
{ |
1166 |
|
BOOL OK; |
1167 |
|
switch (c) |
1168 |
|
{ |
1169 |
case 0x000a: |
case 0x000a: |
1170 |
case 0x000b: |
case 0x000b: |
1171 |
case 0x000c: |
case 0x000c: |
1172 |
|
case 0x000d: |
1173 |
case 0x0085: |
case 0x0085: |
1174 |
case 0x2028: |
case 0x2028: |
1175 |
case 0x2029: |
case 0x2029: |
1176 |
if (count > 0 && codevalue == OP_ANYNL_EXTRA + OP_TYPEPOSPLUS) |
OK = TRUE; |
1177 |
|
break; |
1178 |
|
|
1179 |
|
default: |
1180 |
|
OK = FALSE; |
1181 |
|
break; |
1182 |
|
} |
1183 |
|
|
1184 |
|
if (OK == (d == OP_VSPACE)) |
1185 |
|
{ |
1186 |
|
if (count > 0 && codevalue == OP_VSPACE_EXTRA + OP_TYPEPOSPLUS) |
1187 |
{ |
{ |
1188 |
active_count--; /* Remove non-match possibility */ |
active_count--; /* Remove non-match possibility */ |
1189 |
next_active_state--; |
next_active_state--; |
1190 |
} |
} |
1191 |
count++; |
count++; |
1192 |
ADD_NEW_DATA(-state_offset, count, ncount); |
ADD_NEW_DATA(-state_offset, count, 0); |
1193 |
|
} |
1194 |
|
} |
1195 |
|
break; |
1196 |
|
|
1197 |
|
/*-----------------------------------------------------------------*/ |
1198 |
|
case OP_HSPACE_EXTRA + OP_TYPEPLUS: |
1199 |
|
case OP_HSPACE_EXTRA + OP_TYPEMINPLUS: |
1200 |
|
case OP_HSPACE_EXTRA + OP_TYPEPOSPLUS: |
1201 |
|
count = current_state->count; /* Already matched */ |
1202 |
|
if (count > 0) { ADD_ACTIVE(state_offset + 2, 0); } |
1203 |
|
if (clen > 0) |
1204 |
|
{ |
1205 |
|
BOOL OK; |
1206 |
|
switch (c) |
1207 |
|
{ |
1208 |
|
case 0x09: /* HT */ |
1209 |
|
case 0x20: /* SPACE */ |
1210 |
|
case 0xa0: /* NBSP */ |
1211 |
|
case 0x1680: /* OGHAM SPACE MARK */ |
1212 |
|
case 0x180e: /* MONGOLIAN VOWEL SEPARATOR */ |
1213 |
|
case 0x2000: /* EN QUAD */ |
1214 |
|
case 0x2001: /* EM QUAD */ |
1215 |
|
case 0x2002: /* EN SPACE */ |
1216 |
|
case 0x2003: /* EM SPACE */ |
1217 |
|
case 0x2004: /* THREE-PER-EM SPACE */ |
1218 |
|
case 0x2005: /* FOUR-PER-EM SPACE */ |
1219 |
|
case 0x2006: /* SIX-PER-EM SPACE */ |
1220 |
|
case 0x2007: /* FIGURE SPACE */ |
1221 |
|
case 0x2008: /* PUNCTUATION SPACE */ |
1222 |
|
case 0x2009: /* THIN SPACE */ |
1223 |
|
case 0x200A: /* HAIR SPACE */ |
1224 |
|
case 0x202f: /* NARROW NO-BREAK SPACE */ |
1225 |
|
case 0x205f: /* MEDIUM MATHEMATICAL SPACE */ |
1226 |
|
case 0x3000: /* IDEOGRAPHIC SPACE */ |
1227 |
|
OK = TRUE; |
1228 |
break; |
break; |
1229 |
|
|
1230 |
default: |
default: |
1231 |
|
OK = FALSE; |
1232 |
break; |
break; |
1233 |
} |
} |
1234 |
|
|
1235 |
|
if (OK == (d == OP_HSPACE)) |
1236 |
|
{ |
1237 |
|
if (count > 0 && codevalue == OP_HSPACE_EXTRA + OP_TYPEPOSPLUS) |
1238 |
|
{ |
1239 |
|
active_count--; /* Remove non-match possibility */ |
1240 |
|
next_active_state--; |
1241 |
|
} |
1242 |
|
count++; |
1243 |
|
ADD_NEW_DATA(-state_offset, count, 0); |
1244 |
|
} |
1245 |
} |
} |
1246 |
break; |
break; |
1247 |
|
|
1264 |
if (clen > 0) |
if (clen > 0) |
1265 |
{ |
{ |
1266 |
BOOL OK; |
BOOL OK; |
1267 |
int category = _pcre_ucp_findprop(c, &chartype, &script); |
const ucd_record * prop = GET_UCD(c); |
1268 |
switch(code[2]) |
switch(code[2]) |
1269 |
{ |
{ |
1270 |
case PT_ANY: |
case PT_ANY: |
1272 |
break; |
break; |
1273 |
|
|
1274 |
case PT_LAMP: |
case PT_LAMP: |
1275 |
OK = chartype == ucp_Lu || chartype == ucp_Ll || chartype == ucp_Lt; |
OK = prop->chartype == ucp_Lu || prop->chartype == ucp_Ll || prop->chartype == ucp_Lt; |
1276 |
break; |
break; |
1277 |
|
|
1278 |
case PT_GC: |
case PT_GC: |
1279 |
OK = category == code[3]; |
OK = _pcre_ucp_gentype[prop->chartype] == code[3]; |
1280 |
break; |
break; |
1281 |
|
|
1282 |
case PT_PC: |
case PT_PC: |
1283 |
OK = chartype == code[3]; |
OK = prop->chartype == code[3]; |
1284 |
break; |
break; |
1285 |
|
|
1286 |
case PT_SC: |
case PT_SC: |
1287 |
OK = script == code[3]; |
OK = prop->script == code[3]; |
1288 |
break; |
break; |
1289 |
|
|
1290 |
/* Should never occur, but keep compilers from grumbling. */ |
/* Should never occur, but keep compilers from grumbling. */ |
1322 |
QS2: |
QS2: |
1323 |
|
|
1324 |
ADD_ACTIVE(state_offset + 2, 0); |
ADD_ACTIVE(state_offset + 2, 0); |
1325 |
if (clen > 0 && _pcre_ucp_findprop(c, &chartype, &script) != ucp_M) |
if (clen > 0 && UCD_CATEGORY(c) != ucp_M) |
1326 |
{ |
{ |
1327 |
const uschar *nptr = ptr + clen; |
const uschar *nptr = ptr + clen; |
1328 |
int ncount = 0; |
int ncount = 0; |
1337 |
int nd; |
int nd; |
1338 |
int ndlen = 1; |
int ndlen = 1; |
1339 |
GETCHARLEN(nd, nptr, ndlen); |
GETCHARLEN(nd, nptr, ndlen); |
1340 |
if (_pcre_ucp_findprop(nd, &chartype, &script) != ucp_M) break; |
if (UCD_CATEGORY(nd) != ucp_M) break; |
1341 |
ncount++; |
ncount++; |
1342 |
nptr += ndlen; |
nptr += ndlen; |
1343 |
} |
} |
1365 |
int ncount = 0; |
int ncount = 0; |
1366 |
switch (c) |
switch (c) |
1367 |
{ |
{ |
1368 |
|
case 0x000b: |
1369 |
|
case 0x000c: |
1370 |
|
case 0x0085: |
1371 |
|
case 0x2028: |
1372 |
|
case 0x2029: |
1373 |
|
if ((md->moptions & PCRE_BSR_ANYCRLF) != 0) break; |
1374 |
|
goto ANYNL02; |
1375 |
|
|
1376 |
case 0x000d: |
case 0x000d: |
1377 |
if (ptr + 1 < end_subject && ptr[1] == 0x0a) ncount = 1; |
if (ptr + 1 < end_subject && ptr[1] == 0x0a) ncount = 1; |
1378 |
/* Fall through */ |
/* Fall through */ |
1379 |
|
|
1380 |
|
ANYNL02: |
1381 |
|
case 0x000a: |
1382 |
|
if (codevalue == OP_ANYNL_EXTRA + OP_TYPEPOSSTAR || |
1383 |
|
codevalue == OP_ANYNL_EXTRA + OP_TYPEPOSQUERY) |
1384 |
|
{ |
1385 |
|
active_count--; /* Remove non-match possibility */ |
1386 |
|
next_active_state--; |
1387 |
|
} |
1388 |
|
ADD_NEW_DATA(-(state_offset + count), 0, ncount); |
1389 |
|
break; |
1390 |
|
|
1391 |
|
default: |
1392 |
|
break; |
1393 |
|
} |
1394 |
|
} |
1395 |
|
break; |
1396 |
|
|
1397 |
|
/*-----------------------------------------------------------------*/ |
1398 |
|
case OP_VSPACE_EXTRA + OP_TYPEQUERY: |
1399 |
|
case OP_VSPACE_EXTRA + OP_TYPEMINQUERY: |
1400 |
|
case OP_VSPACE_EXTRA + OP_TYPEPOSQUERY: |
1401 |
|
count = 2; |
1402 |
|
goto QS4; |
1403 |
|
|
1404 |
|
case OP_VSPACE_EXTRA + OP_TYPESTAR: |
1405 |
|
case OP_VSPACE_EXTRA + OP_TYPEMINSTAR: |
1406 |
|
case OP_VSPACE_EXTRA + OP_TYPEPOSSTAR: |
1407 |
|
count = 0; |
1408 |
|
|
1409 |
|
QS4: |
1410 |
|
ADD_ACTIVE(state_offset + 2, 0); |
1411 |
|
if (clen > 0) |
1412 |
|
{ |
1413 |
|
BOOL OK; |
1414 |
|
switch (c) |
1415 |
|
{ |
1416 |
case 0x000a: |
case 0x000a: |
1417 |
case 0x000b: |
case 0x000b: |
1418 |
case 0x000c: |
case 0x000c: |
1419 |
|
case 0x000d: |
1420 |
case 0x0085: |
case 0x0085: |
1421 |
case 0x2028: |
case 0x2028: |
1422 |
case 0x2029: |
case 0x2029: |
1423 |
if (codevalue == OP_ANYNL_EXTRA + OP_TYPEPOSSTAR || |
OK = TRUE; |
1424 |
codevalue == OP_ANYNL_EXTRA + OP_TYPEPOSQUERY) |
break; |
1425 |
|
|
1426 |
|
default: |
1427 |
|
OK = FALSE; |
1428 |
|
break; |
1429 |
|
} |
1430 |
|
if (OK == (d == OP_VSPACE)) |
1431 |
|
{ |
1432 |
|
if (codevalue == OP_VSPACE_EXTRA + OP_TYPEPOSSTAR || |
1433 |
|
codevalue == OP_VSPACE_EXTRA + OP_TYPEPOSQUERY) |
1434 |
{ |
{ |
1435 |
active_count--; /* Remove non-match possibility */ |
active_count--; /* Remove non-match possibility */ |
1436 |
next_active_state--; |
next_active_state--; |
1437 |
} |
} |
1438 |
ADD_NEW_DATA(-(state_offset + count), 0, ncount); |
ADD_NEW_DATA(-(state_offset + count), 0, 0); |
1439 |
|
} |
1440 |
|
} |
1441 |
|
break; |
1442 |
|
|
1443 |
|
/*-----------------------------------------------------------------*/ |
1444 |
|
case OP_HSPACE_EXTRA + OP_TYPEQUERY: |
1445 |
|
case OP_HSPACE_EXTRA + OP_TYPEMINQUERY: |
1446 |
|
case OP_HSPACE_EXTRA + OP_TYPEPOSQUERY: |
1447 |
|
count = 2; |
1448 |
|
goto QS5; |
1449 |
|
|
1450 |
|
case OP_HSPACE_EXTRA + OP_TYPESTAR: |
1451 |
|
case OP_HSPACE_EXTRA + OP_TYPEMINSTAR: |
1452 |
|
case OP_HSPACE_EXTRA + OP_TYPEPOSSTAR: |
1453 |
|
count = 0; |
1454 |
|
|
1455 |
|
QS5: |
1456 |
|
ADD_ACTIVE(state_offset + 2, 0); |
1457 |
|
if (clen > 0) |
1458 |
|
{ |
1459 |
|
BOOL OK; |
1460 |
|
switch (c) |
1461 |
|
{ |
1462 |
|
case 0x09: /* HT */ |
1463 |
|
case 0x20: /* SPACE */ |
1464 |
|
case 0xa0: /* NBSP */ |
1465 |
|
case 0x1680: /* OGHAM SPACE MARK */ |
1466 |
|
case 0x180e: /* MONGOLIAN VOWEL SEPARATOR */ |
1467 |
|
case 0x2000: /* EN QUAD */ |
1468 |
|
case 0x2001: /* EM QUAD */ |
1469 |
|
case 0x2002: /* EN SPACE */ |
1470 |
|
case 0x2003: /* EM SPACE */ |
1471 |
|
case 0x2004: /* THREE-PER-EM SPACE */ |
1472 |
|
case 0x2005: /* FOUR-PER-EM SPACE */ |
1473 |
|
case 0x2006: /* SIX-PER-EM SPACE */ |
1474 |
|
case 0x2007: /* FIGURE SPACE */ |
1475 |
|
case 0x2008: /* PUNCTUATION SPACE */ |
1476 |
|
case 0x2009: /* THIN SPACE */ |
1477 |
|
case 0x200A: /* HAIR SPACE */ |
1478 |
|
case 0x202f: /* NARROW NO-BREAK SPACE */ |
1479 |
|
case 0x205f: /* MEDIUM MATHEMATICAL SPACE */ |
1480 |
|
case 0x3000: /* IDEOGRAPHIC SPACE */ |
1481 |
|
OK = TRUE; |
1482 |
break; |
break; |
1483 |
|
|
1484 |
default: |
default: |
1485 |
|
OK = FALSE; |
1486 |
break; |
break; |
1487 |
} |
} |
1488 |
|
|
1489 |
|
if (OK == (d == OP_HSPACE)) |
1490 |
|
{ |
1491 |
|
if (codevalue == OP_HSPACE_EXTRA + OP_TYPEPOSSTAR || |
1492 |
|
codevalue == OP_HSPACE_EXTRA + OP_TYPEPOSQUERY) |
1493 |
|
{ |
1494 |
|
active_count--; /* Remove non-match possibility */ |
1495 |
|
next_active_state--; |
1496 |
|
} |
1497 |
|
ADD_NEW_DATA(-(state_offset + count), 0, 0); |
1498 |
|
} |
1499 |
} |
} |
1500 |
break; |
break; |
1501 |
|
|
1511 |
if (clen > 0) |
if (clen > 0) |
1512 |
{ |
{ |
1513 |
BOOL OK; |
BOOL OK; |
1514 |
int category = _pcre_ucp_findprop(c, &chartype, &script); |
const ucd_record * prop = GET_UCD(c); |
1515 |
switch(code[4]) |
switch(code[4]) |
1516 |
{ |
{ |
1517 |
case PT_ANY: |
case PT_ANY: |
1519 |
break; |
break; |
1520 |
|
|
1521 |
case PT_LAMP: |
case PT_LAMP: |
1522 |
OK = chartype == ucp_Lu || chartype == ucp_Ll || chartype == ucp_Lt; |
OK = prop->chartype == ucp_Lu || prop->chartype == ucp_Ll || prop->chartype == ucp_Lt; |
1523 |
break; |
break; |
1524 |
|
|
1525 |
case PT_GC: |
case PT_GC: |
1526 |
OK = category == code[5]; |
OK = _pcre_ucp_gentype[prop->chartype] == code[5]; |
1527 |
break; |
break; |
1528 |
|
|
1529 |
case PT_PC: |
case PT_PC: |
1530 |
OK = chartype == code[5]; |
OK = prop->chartype == code[5]; |
1531 |
break; |
break; |
1532 |
|
|
1533 |
case PT_SC: |
case PT_SC: |
1534 |
OK = script == code[5]; |
OK = prop->script == code[5]; |
1535 |
break; |
break; |
1536 |
|
|
1537 |
/* Should never occur, but keep compilers from grumbling. */ |
/* Should never occur, but keep compilers from grumbling. */ |
1564 |
if (codevalue != OP_EXTUNI_EXTRA + OP_TYPEEXACT) |
if (codevalue != OP_EXTUNI_EXTRA + OP_TYPEEXACT) |
1565 |
{ ADD_ACTIVE(state_offset + 4, 0); } |
{ ADD_ACTIVE(state_offset + 4, 0); } |
1566 |
count = current_state->count; /* Number already matched */ |
count = current_state->count; /* Number already matched */ |
1567 |
if (clen > 0 && _pcre_ucp_findprop(c, &chartype, &script) != ucp_M) |
if (clen > 0 && UCD_CATEGORY(c) != ucp_M) |
1568 |
{ |
{ |
1569 |
const uschar *nptr = ptr + clen; |
const uschar *nptr = ptr + clen; |
1570 |
int ncount = 0; |
int ncount = 0; |
1578 |
int nd; |
int nd; |
1579 |
int ndlen = 1; |
int ndlen = 1; |
1580 |
GETCHARLEN(nd, nptr, ndlen); |
GETCHARLEN(nd, nptr, ndlen); |
1581 |
if (_pcre_ucp_findprop(nd, &chartype, &script) != ucp_M) break; |
if (UCD_CATEGORY(nd) != ucp_M) break; |
1582 |
ncount++; |
ncount++; |
1583 |
nptr += ndlen; |
nptr += ndlen; |
1584 |
} |
} |
1603 |
int ncount = 0; |
int ncount = 0; |
1604 |
switch (c) |
switch (c) |
1605 |
{ |
{ |
1606 |
|
case 0x000b: |
1607 |
|
case 0x000c: |
1608 |
|
case 0x0085: |
1609 |
|
case 0x2028: |
1610 |
|
case 0x2029: |
1611 |
|
if ((md->moptions & PCRE_BSR_ANYCRLF) != 0) break; |
1612 |
|
goto ANYNL03; |
1613 |
|
|
1614 |
case 0x000d: |
case 0x000d: |
1615 |
if (ptr + 1 < end_subject && ptr[1] == 0x0a) ncount = 1; |
if (ptr + 1 < end_subject && ptr[1] == 0x0a) ncount = 1; |
1616 |
/* Fall through */ |
/* Fall through */ |
1617 |
|
|
1618 |
|
ANYNL03: |
1619 |
|
case 0x000a: |
1620 |
|
if (codevalue == OP_ANYNL_EXTRA + OP_TYPEPOSUPTO) |
1621 |
|
{ |
1622 |
|
active_count--; /* Remove non-match possibility */ |
1623 |
|
next_active_state--; |
1624 |
|
} |
1625 |
|
if (++count >= GET2(code, 1)) |
1626 |
|
{ ADD_NEW_DATA(-(state_offset + 4), 0, ncount); } |
1627 |
|
else |
1628 |
|
{ ADD_NEW_DATA(-state_offset, count, ncount); } |
1629 |
|
break; |
1630 |
|
|
1631 |
|
default: |
1632 |
|
break; |
1633 |
|
} |
1634 |
|
} |
1635 |
|
break; |
1636 |
|
|
1637 |
|
/*-----------------------------------------------------------------*/ |
1638 |
|
case OP_VSPACE_EXTRA + OP_TYPEEXACT: |
1639 |
|
case OP_VSPACE_EXTRA + OP_TYPEUPTO: |
1640 |
|
case OP_VSPACE_EXTRA + OP_TYPEMINUPTO: |
1641 |
|
case OP_VSPACE_EXTRA + OP_TYPEPOSUPTO: |
1642 |
|
if (codevalue != OP_VSPACE_EXTRA + OP_TYPEEXACT) |
1643 |
|
{ ADD_ACTIVE(state_offset + 4, 0); } |
1644 |
|
count = current_state->count; /* Number already matched */ |
1645 |
|
if (clen > 0) |
1646 |
|
{ |
1647 |
|
BOOL OK; |
1648 |
|
switch (c) |
1649 |
|
{ |
1650 |
case 0x000a: |
case 0x000a: |
1651 |
case 0x000b: |
case 0x000b: |
1652 |
case 0x000c: |
case 0x000c: |
1653 |
|
case 0x000d: |
1654 |
case 0x0085: |
case 0x0085: |
1655 |
case 0x2028: |
case 0x2028: |
1656 |
case 0x2029: |
case 0x2029: |
1657 |
if (codevalue == OP_ANYNL_EXTRA + OP_TYPEPOSUPTO) |
OK = TRUE; |
1658 |
|
break; |
1659 |
|
|
1660 |
|
default: |
1661 |
|
OK = FALSE; |
1662 |
|
} |
1663 |
|
|
1664 |
|
if (OK == (d == OP_VSPACE)) |
1665 |
|
{ |
1666 |
|
if (codevalue == OP_VSPACE_EXTRA + OP_TYPEPOSUPTO) |
1667 |
{ |
{ |
1668 |
active_count--; /* Remove non-match possibility */ |
active_count--; /* Remove non-match possibility */ |
1669 |
next_active_state--; |
next_active_state--; |
1670 |
} |
} |
1671 |
if (++count >= GET2(code, 1)) |
if (++count >= GET2(code, 1)) |
1672 |
{ ADD_NEW_DATA(-(state_offset + 4), 0, ncount); } |
{ ADD_NEW_DATA(-(state_offset + 4), 0, 0); } |
1673 |
else |
else |
1674 |
{ ADD_NEW_DATA(-state_offset, count, ncount); } |
{ ADD_NEW_DATA(-state_offset, count, 0); } |
1675 |
|
} |
1676 |
|
} |
1677 |
|
break; |
1678 |
|
|
1679 |
|
/*-----------------------------------------------------------------*/ |
1680 |
|
case OP_HSPACE_EXTRA + OP_TYPEEXACT: |
1681 |
|
case OP_HSPACE_EXTRA + OP_TYPEUPTO: |
1682 |
|
case OP_HSPACE_EXTRA + OP_TYPEMINUPTO: |
1683 |
|
case OP_HSPACE_EXTRA + OP_TYPEPOSUPTO: |
1684 |
|
if (codevalue != OP_HSPACE_EXTRA + OP_TYPEEXACT) |
1685 |
|
{ ADD_ACTIVE(state_offset + 4, 0); } |
1686 |
|
count = current_state->count; /* Number already matched */ |
1687 |
|
if (clen > 0) |
1688 |
|
{ |
1689 |
|
BOOL OK; |
1690 |
|
switch (c) |
1691 |
|
{ |
1692 |
|
case 0x09: /* HT */ |
1693 |
|
case 0x20: /* SPACE */ |
1694 |
|
case 0xa0: /* NBSP */ |
1695 |
|
case 0x1680: /* OGHAM SPACE MARK */ |
1696 |
|
case 0x180e: /* MONGOLIAN VOWEL SEPARATOR */ |
1697 |
|
case 0x2000: /* EN QUAD */ |
1698 |
|
case 0x2001: /* EM QUAD */ |
1699 |
|
case 0x2002: /* EN SPACE */ |
1700 |
|
case 0x2003: /* EM SPACE */ |
1701 |
|
case 0x2004: /* THREE-PER-EM SPACE */ |
1702 |
|
case 0x2005: /* FOUR-PER-EM SPACE */ |
1703 |
|
case 0x2006: /* SIX-PER-EM SPACE */ |
1704 |
|
case 0x2007: /* FIGURE SPACE */ |
1705 |
|
case 0x2008: /* PUNCTUATION SPACE */ |
1706 |
|
case 0x2009: /* THIN SPACE */ |
1707 |
|
case 0x200A: /* HAIR SPACE */ |
1708 |
|
case 0x202f: /* NARROW NO-BREAK SPACE */ |
1709 |
|
case 0x205f: /* MEDIUM MATHEMATICAL SPACE */ |
1710 |
|
case 0x3000: /* IDEOGRAPHIC SPACE */ |
1711 |
|
OK = TRUE; |
1712 |
break; |
break; |
1713 |
|
|
1714 |
default: |
default: |
1715 |
|
OK = FALSE; |
1716 |
break; |
break; |
1717 |
} |
} |
1718 |
|
|
1719 |
|
if (OK == (d == OP_HSPACE)) |
1720 |
|
{ |
1721 |
|
if (codevalue == OP_HSPACE_EXTRA + OP_TYPEPOSUPTO) |
1722 |
|
{ |
1723 |
|
active_count--; /* Remove non-match possibility */ |
1724 |
|
next_active_state--; |
1725 |
|
} |
1726 |
|
if (++count >= GET2(code, 1)) |
1727 |
|
{ ADD_NEW_DATA(-(state_offset + 4), 0, 0); } |
1728 |
|
else |
1729 |
|
{ ADD_NEW_DATA(-state_offset, count, 0); } |
1730 |
|
} |
1731 |
} |
} |
1732 |
break; |
break; |
1733 |
|
|
1758 |
other case of the character. */ |
other case of the character. */ |
1759 |
|
|
1760 |
#ifdef SUPPORT_UCP |
#ifdef SUPPORT_UCP |
1761 |
othercase = _pcre_ucp_othercase(c); |
othercase = UCD_OTHERCASE(c); |
1762 |
#else |
#else |
1763 |
othercase = NOTACHAR; |
othercase = NOTACHAR; |
1764 |
#endif |
#endif |
1783 |
to wait for them to pass before continuing. */ |
to wait for them to pass before continuing. */ |
1784 |
|
|
1785 |
case OP_EXTUNI: |
case OP_EXTUNI: |
1786 |
if (clen > 0 && _pcre_ucp_findprop(c, &chartype, &script) != ucp_M) |
if (clen > 0 && UCD_CATEGORY(c) != ucp_M) |
1787 |
{ |
{ |
1788 |
const uschar *nptr = ptr + clen; |
const uschar *nptr = ptr + clen; |
1789 |
int ncount = 0; |
int ncount = 0; |
1791 |
{ |
{ |
1792 |
int nclen = 1; |
int nclen = 1; |
1793 |
GETCHARLEN(c, nptr, nclen); |
GETCHARLEN(c, nptr, nclen); |
1794 |
if (_pcre_ucp_findprop(c, &chartype, &script) != ucp_M) break; |
if (UCD_CATEGORY(c) != ucp_M) break; |
1795 |
ncount++; |
ncount++; |
1796 |
nptr += nclen; |
nptr += nclen; |
1797 |
} |
} |
1808 |
case OP_ANYNL: |
case OP_ANYNL: |
1809 |
if (clen > 0) switch(c) |
if (clen > 0) switch(c) |
1810 |
{ |
{ |
|
case 0x000a: |
|
1811 |
case 0x000b: |
case 0x000b: |
1812 |
case 0x000c: |
case 0x000c: |
1813 |
case 0x0085: |
case 0x0085: |
1814 |
case 0x2028: |
case 0x2028: |
1815 |
case 0x2029: |
case 0x2029: |
1816 |
|
if ((md->moptions & PCRE_BSR_ANYCRLF) != 0) break; |
1817 |
|
|
1818 |
|
case 0x000a: |
1819 |
ADD_NEW(state_offset + 1, 0); |
ADD_NEW(state_offset + 1, 0); |
1820 |
break; |
break; |
1821 |
|
|
1822 |
case 0x000d: |
case 0x000d: |
1823 |
if (ptr + 1 < end_subject && ptr[1] == 0x0a) |
if (ptr + 1 < end_subject && ptr[1] == 0x0a) |
1824 |
{ |
{ |
1833 |
break; |
break; |
1834 |
|
|
1835 |
/*-----------------------------------------------------------------*/ |
/*-----------------------------------------------------------------*/ |
1836 |
|
case OP_NOT_VSPACE: |
1837 |
|
if (clen > 0) switch(c) |
1838 |
|
{ |
1839 |
|
case 0x000a: |
1840 |
|
case 0x000b: |
1841 |
|
case 0x000c: |
1842 |
|
case 0x000d: |
1843 |
|
case 0x0085: |
1844 |
|
case 0x2028: |
1845 |
|
case 0x2029: |
1846 |
|
break; |
1847 |
|
|
1848 |
|
default: |
1849 |
|
ADD_NEW(state_offset + 1, 0); |
1850 |
|
break; |
1851 |
|
} |
1852 |
|
break; |
1853 |
|
|
1854 |
|
/*-----------------------------------------------------------------*/ |
1855 |
|
case OP_VSPACE: |
1856 |
|
if (clen > 0) switch(c) |
1857 |
|
{ |
1858 |
|
case 0x000a: |
1859 |
|
case 0x000b: |
1860 |
|
case 0x000c: |
1861 |
|
case 0x000d: |
1862 |
|
case 0x0085: |
1863 |
|
case 0x2028: |
1864 |
|
case 0x2029: |
1865 |
|
ADD_NEW(state_offset + 1, 0); |
1866 |
|
break; |
1867 |
|
|
1868 |
|
default: break; |
1869 |
|
} |
1870 |
|
break; |
1871 |
|
|
1872 |
|
/*-----------------------------------------------------------------*/ |
1873 |
|
case OP_NOT_HSPACE: |
1874 |
|
if (clen > 0) switch(c) |
1875 |
|
{ |
1876 |
|
case 0x09: /* HT */ |
1877 |
|
case 0x20: /* SPACE */ |
1878 |
|
case 0xa0: /* NBSP */ |
1879 |
|
case 0x1680: /* OGHAM SPACE MARK */ |
1880 |
|
case 0x180e: /* MONGOLIAN VOWEL SEPARATOR */ |
1881 |
|
case 0x2000: /* EN QUAD */ |
1882 |
|
case 0x2001: /* EM QUAD */ |
1883 |
|
case 0x2002: /* EN SPACE */ |
1884 |
|
case 0x2003: /* EM SPACE */ |
1885 |
|
case 0x2004: /* THREE-PER-EM SPACE */ |
1886 |
|
case 0x2005: /* FOUR-PER-EM SPACE */ |
1887 |
|
case 0x2006: /* SIX-PER-EM SPACE */ |
1888 |
|
case 0x2007: /* FIGURE SPACE */ |
1889 |
|
case 0x2008: /* PUNCTUATION SPACE */ |
1890 |
|
case 0x2009: /* THIN SPACE */ |
1891 |
|
case 0x200A: /* HAIR SPACE */ |
1892 |
|
case 0x202f: /* NARROW NO-BREAK SPACE */ |
1893 |
|
case 0x205f: /* MEDIUM MATHEMATICAL SPACE */ |
1894 |
|
case 0x3000: /* IDEOGRAPHIC SPACE */ |
1895 |
|
break; |
1896 |
|
|
1897 |
|
default: |
1898 |
|
ADD_NEW(state_offset + 1, 0); |
1899 |
|
break; |
1900 |
|
} |
1901 |
|
break; |
1902 |
|
|
1903 |
|
/*-----------------------------------------------------------------*/ |
1904 |
|
case OP_HSPACE: |
1905 |
|
if (clen > 0) switch(c) |
1906 |
|
{ |
1907 |
|
case 0x09: /* HT */ |
1908 |
|
case 0x20: /* SPACE */ |
1909 |
|
case 0xa0: /* NBSP */ |
1910 |
|
case 0x1680: /* OGHAM SPACE MARK */ |
1911 |
|
case 0x180e: /* MONGOLIAN VOWEL SEPARATOR */ |
1912 |
|
case 0x2000: /* EN QUAD */ |
1913 |
|
case 0x2001: /* EM QUAD */ |
1914 |
|
case 0x2002: /* EN SPACE */ |
1915 |
|
case 0x2003: /* EM SPACE */ |
1916 |
|
case 0x2004: /* THREE-PER-EM SPACE */ |
1917 |
|
case 0x2005: /* FOUR-PER-EM SPACE */ |
1918 |
|
case 0x2006: /* SIX-PER-EM SPACE */ |
1919 |
|
case 0x2007: /* FIGURE SPACE */ |
1920 |
|
case 0x2008: /* PUNCTUATION SPACE */ |
1921 |
|
case 0x2009: /* THIN SPACE */ |
1922 |
|
case 0x200A: /* HAIR SPACE */ |
1923 |
|
case 0x202f: /* NARROW NO-BREAK SPACE */ |
1924 |
|
case 0x205f: /* MEDIUM MATHEMATICAL SPACE */ |
1925 |
|
case 0x3000: /* IDEOGRAPHIC SPACE */ |
1926 |
|
ADD_NEW(state_offset + 1, 0); |
1927 |
|
break; |
1928 |
|
} |
1929 |
|
break; |
1930 |
|
|
1931 |
|
/*-----------------------------------------------------------------*/ |
1932 |
/* Match a negated single character. This is only used for one-byte |
/* Match a negated single character. This is only used for one-byte |
1933 |
characters, that is, we know that d < 256. The character we are |
characters, that is, we know that d < 256. The character we are |
1934 |
checking (c) can be multibyte. */ |
checking (c) can be multibyte. */ |
1959 |
if (utf8 && d >= 128) |
if (utf8 && d >= 128) |
1960 |
{ |
{ |
1961 |
#ifdef SUPPORT_UCP |
#ifdef SUPPORT_UCP |
1962 |
otherd = _pcre_ucp_othercase(d); |
otherd = UCD_OTHERCASE(d); |
1963 |
#endif /* SUPPORT_UCP */ |
#endif /* SUPPORT_UCP */ |
1964 |
} |
} |
1965 |
else |
else |
1997 |
if (utf8 && d >= 128) |
if (utf8 && d >= 128) |
1998 |
{ |
{ |
1999 |
#ifdef SUPPORT_UCP |
#ifdef SUPPORT_UCP |
2000 |
otherd = _pcre_ucp_othercase(d); |
otherd = UCD_OTHERCASE(d); |
2001 |
#endif /* SUPPORT_UCP */ |
#endif /* SUPPORT_UCP */ |
2002 |
} |
} |
2003 |
else |
else |
2033 |
if (utf8 && d >= 128) |
if (utf8 && d >= 128) |
2034 |
{ |
{ |
2035 |
#ifdef SUPPORT_UCP |
#ifdef SUPPORT_UCP |
2036 |
otherd = _pcre_ucp_othercase(d); |
otherd = UCD_OTHERCASE(d); |
2037 |
#endif /* SUPPORT_UCP */ |
#endif /* SUPPORT_UCP */ |
2038 |
} |
} |
2039 |
else |
else |
2065 |
if (utf8 && d >= 128) |
if (utf8 && d >= 128) |
2066 |
{ |
{ |
2067 |
#ifdef SUPPORT_UCP |
#ifdef SUPPORT_UCP |
2068 |
otherd = _pcre_ucp_othercase(d); |
otherd = UCD_OTHERCASE(d); |
2069 |
#endif /* SUPPORT_UCP */ |
#endif /* SUPPORT_UCP */ |
2070 |
} |
} |
2071 |
else |
else |
2100 |
if (utf8 && d >= 128) |
if (utf8 && d >= 128) |
2101 |
{ |
{ |
2102 |
#ifdef SUPPORT_UCP |
#ifdef SUPPORT_UCP |
2103 |
otherd = _pcre_ucp_othercase(d); |
otherd = UCD_OTHERCASE(d); |
2104 |
#endif /* SUPPORT_UCP */ |
#endif /* SUPPORT_UCP */ |
2105 |
} |
} |
2106 |
else |
else |
2208 |
|
|
2209 |
/* ========================================================================== */ |
/* ========================================================================== */ |
2210 |
/* These are the opcodes for fancy brackets of various kinds. We have |
/* These are the opcodes for fancy brackets of various kinds. We have |
2211 |
to use recursion in order to handle them. */ |
to use recursion in order to handle them. The "always failing" assertion |
2212 |
|
(?!) is optimised to OP_FAIL when compiling, so we have to support that, |
2213 |
|
though the other "backtracking verbs" are not supported. */ |
2214 |
|
|
2215 |
|
case OP_FAIL: |
2216 |
|
forced_fail++; /* Count FAILs for multiple states */ |
2217 |
|
break; |
2218 |
|
|
2219 |
case OP_ASSERT: |
case OP_ASSERT: |
2220 |
case OP_ASSERT_NOT: |
case OP_ASSERT_NOT: |
2252 |
{ |
{ |
2253 |
int local_offsets[1000]; |
int local_offsets[1000]; |
2254 |
int local_workspace[1000]; |
int local_workspace[1000]; |
2255 |
int condcode = code[LINK_SIZE+1]; |
int codelink = GET(code, 1); |
2256 |
|
int condcode; |
2257 |
|
|
2258 |
|
/* Because of the way auto-callout works during compile, a callout item |
2259 |
|
is inserted between OP_COND and an assertion condition. This does not |
2260 |
|
happen for the other conditions. */ |
2261 |
|
|
2262 |
|
if (code[LINK_SIZE+1] == OP_CALLOUT) |
2263 |
|
{ |
2264 |
|
rrc = 0; |
2265 |
|
if (pcre_callout != NULL) |
2266 |
|
{ |
2267 |
|
pcre_callout_block cb; |
2268 |
|
cb.version = 1; /* Version 1 of the callout block */ |
2269 |
|
cb.callout_number = code[LINK_SIZE+2]; |
2270 |
|
cb.offset_vector = offsets; |
2271 |
|
cb.subject = (PCRE_SPTR)start_subject; |
2272 |
|
cb.subject_length = end_subject - start_subject; |
2273 |
|
cb.start_match = current_subject - start_subject; |
2274 |
|
cb.current_position = ptr - start_subject; |
2275 |
|
cb.pattern_position = GET(code, LINK_SIZE + 3); |
2276 |
|
cb.next_item_length = GET(code, 3 + 2*LINK_SIZE); |
2277 |
|
cb.capture_top = 1; |
2278 |
|
cb.capture_last = -1; |
2279 |
|
cb.callout_data = md->callout_data; |
2280 |
|
if ((rrc = (*pcre_callout)(&cb)) < 0) return rrc; /* Abandon */ |
2281 |
|
} |
2282 |
|
if (rrc > 0) break; /* Fail this thread */ |
2283 |
|
code += _pcre_OP_lengths[OP_CALLOUT]; /* Skip callout data */ |
2284 |
|
} |
2285 |
|
|
2286 |
|
condcode = code[LINK_SIZE+1]; |
2287 |
|
|
2288 |
/* Back reference conditions are not supported */ |
/* Back reference conditions are not supported */ |
2289 |
|
|
2292 |
/* The DEFINE condition is always false */ |
/* The DEFINE condition is always false */ |
2293 |
|
|
2294 |
if (condcode == OP_DEF) |
if (condcode == OP_DEF) |
2295 |
{ |
{ ADD_ACTIVE(state_offset + codelink + LINK_SIZE + 1, 0); } |
|
ADD_ACTIVE(state_offset + GET(code, 1) + LINK_SIZE + 1, 0); |
|
|
} |
|
2296 |
|
|
2297 |
/* The only supported version of OP_RREF is for the value RREF_ANY, |
/* The only supported version of OP_RREF is for the value RREF_ANY, |
2298 |
which means "test if in any recursion". We can't test for specifically |
which means "test if in any recursion". We can't test for specifically |
2302 |
{ |
{ |
2303 |
int value = GET2(code, LINK_SIZE+2); |
int value = GET2(code, LINK_SIZE+2); |
2304 |
if (value != RREF_ANY) return PCRE_ERROR_DFA_UCOND; |
if (value != RREF_ANY) return PCRE_ERROR_DFA_UCOND; |
2305 |
if (recursing > 0) { ADD_ACTIVE(state_offset + LINK_SIZE + 4, 0); } |
if (recursing > 0) |
2306 |
else { ADD_ACTIVE(state_offset + GET(code, 1) + LINK_SIZE + 1, 0); } |
{ ADD_ACTIVE(state_offset + LINK_SIZE + 4, 0); } |
2307 |
|
else { ADD_ACTIVE(state_offset + codelink + LINK_SIZE + 1, 0); } |
2308 |
} |
} |
2309 |
|
|
2310 |
/* Otherwise, the condition is an assertion */ |
/* Otherwise, the condition is an assertion */ |
2334 |
(condcode == OP_ASSERT || condcode == OP_ASSERTBACK)) |
(condcode == OP_ASSERT || condcode == OP_ASSERTBACK)) |
2335 |
{ ADD_ACTIVE(endasscode + LINK_SIZE + 1 - start_code, 0); } |
{ ADD_ACTIVE(endasscode + LINK_SIZE + 1 - start_code, 0); } |
2336 |
else |
else |
2337 |
{ ADD_ACTIVE(state_offset + GET(code, 1) + LINK_SIZE + 1, 0); } |
{ ADD_ACTIVE(state_offset + codelink + LINK_SIZE + 1, 0); } |
2338 |
} |
} |
2339 |
} |
} |
2340 |
break; |
break; |
2486 |
/* Handle callouts */ |
/* Handle callouts */ |
2487 |
|
|
2488 |
case OP_CALLOUT: |
case OP_CALLOUT: |
2489 |
|
rrc = 0; |
2490 |
if (pcre_callout != NULL) |
if (pcre_callout != NULL) |
2491 |
{ |
{ |
|
int rrc; |
|
2492 |
pcre_callout_block cb; |
pcre_callout_block cb; |
2493 |
cb.version = 1; /* Version 1 of the callout block */ |
cb.version = 1; /* Version 1 of the callout block */ |
2494 |
cb.callout_number = code[1]; |
cb.callout_number = code[1]; |
2503 |
cb.capture_last = -1; |
cb.capture_last = -1; |
2504 |
cb.callout_data = md->callout_data; |
cb.callout_data = md->callout_data; |
2505 |
if ((rrc = (*pcre_callout)(&cb)) < 0) return rrc; /* Abandon */ |
if ((rrc = (*pcre_callout)(&cb)) < 0) return rrc; /* Abandon */ |
|
if (rrc == 0) { ADD_ACTIVE(state_offset + 2 + 2*LINK_SIZE, 0); } |
|
2506 |
} |
} |
2507 |
|
if (rrc == 0) |
2508 |
|
{ ADD_ACTIVE(state_offset + _pcre_OP_lengths[OP_CALLOUT], 0); } |
2509 |
break; |
break; |
2510 |
|
|
2511 |
|
|
2521 |
/* We have finished the processing at the current subject character. If no |
/* We have finished the processing at the current subject character. If no |
2522 |
new states have been set for the next character, we have found all the |
new states have been set for the next character, we have found all the |
2523 |
matches that we are going to find. If we are at the top level and partial |
matches that we are going to find. If we are at the top level and partial |
2524 |
matching has been requested, check for appropriate conditions. */ |
matching has been requested, check for appropriate conditions. The "forced_ |
2525 |
|
fail" variable counts the number of (*F) encountered for the character. If it |
2526 |
|
is equal to the original active_count (saved in workspace[1]) it means that |
2527 |
|
(*F) was found on every active state. In this case we don't want to give a |
2528 |
|
partial match. */ |
2529 |
|
|
2530 |
if (new_count <= 0) |
if (new_count <= 0) |
2531 |
{ |
{ |
2532 |
if (match_count < 0 && /* No matches found */ |
if (rlevel == 1 && /* Top level, and */ |
2533 |
rlevel == 1 && /* Top level match function */ |
reached_end != workspace[1] && /* Not all reached end */ |
2534 |
(md->moptions & PCRE_PARTIAL) != 0 && /* Want partial matching */ |
forced_fail != workspace[1] && /* Not all forced fail & */ |
2535 |
ptr >= end_subject && /* Reached end of subject */ |
( /* either... */ |
2536 |
ptr > current_subject) /* Matched non-empty string */ |
(md->moptions & PCRE_PARTIAL_HARD) != 0 /* Hard partial */ |
2537 |
|
|| /* or... */ |
2538 |
|
((md->moptions & PCRE_PARTIAL_SOFT) != 0 && /* Soft partial and */ |
2539 |
|
match_count < 0) /* no matches */ |
2540 |
|
) && /* And... */ |
2541 |
|
ptr >= end_subject && /* Reached end of subject */ |
2542 |
|
ptr > current_subject) /* Matched non-empty string */ |
2543 |
{ |
{ |
2544 |
if (offsetcount >= 2) |
if (offsetcount >= 2) |
2545 |
{ |
{ |
2546 |
offsets[0] = current_subject - start_subject; |
offsets[0] = md->start_used_ptr - start_subject; |
2547 |
offsets[1] = end_subject - start_subject; |
offsets[1] = end_subject - start_subject; |
2548 |
} |
} |
2549 |
match_count = PCRE_ERROR_PARTIAL; |
match_count = PCRE_ERROR_PARTIAL; |
2598 |
< -1 => some kind of unexpected problem |
< -1 => some kind of unexpected problem |
2599 |
*/ |
*/ |
2600 |
|
|
2601 |
PCRE_EXP_DEFN int |
PCRE_EXP_DEFN int PCRE_CALL_CONVENTION |
2602 |
pcre_dfa_exec(const pcre *argument_re, const pcre_extra *extra_data, |
pcre_dfa_exec(const pcre *argument_re, const pcre_extra *extra_data, |
2603 |
const char *subject, int length, int start_offset, int options, int *offsets, |
const char *subject, int length, int start_offset, int options, int *offsets, |
2604 |
int offsetcount, int *workspace, int wscount) |
int offsetcount, int *workspace, int wscount) |
2685 |
re->name_table_offset + re->name_count * re->name_entry_size; |
re->name_table_offset + re->name_count * re->name_entry_size; |
2686 |
md->start_subject = (const unsigned char *)subject; |
md->start_subject = (const unsigned char *)subject; |
2687 |
md->end_subject = end_subject; |
md->end_subject = end_subject; |
2688 |
|
md->start_offset = start_offset; |
2689 |
md->moptions = options; |
md->moptions = options; |
2690 |
md->poptions = re->options; |
md->poptions = re->options; |
2691 |
|
|
2692 |
|
/* If the BSR option is not set at match time, copy what was set |
2693 |
|
at compile time. */ |
2694 |
|
|
2695 |
|
if ((md->moptions & (PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE)) == 0) |
2696 |
|
{ |
2697 |
|
if ((re->options & (PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE)) != 0) |
2698 |
|
md->moptions |= re->options & (PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE); |
2699 |
|
#ifdef BSR_ANYCRLF |
2700 |
|
else md->moptions |= PCRE_BSR_ANYCRLF; |
2701 |
|
#endif |
2702 |
|
} |
2703 |
|
|
2704 |
/* Handle different types of newline. The three bits give eight cases. If |
/* Handle different types of newline. The three bits give eight cases. If |
2705 |
nothing is set at run time, whatever was used at compile time applies. */ |
nothing is set at run time, whatever was used at compile time applies. */ |
2706 |
|
|
2708 |
PCRE_NEWLINE_BITS) |
PCRE_NEWLINE_BITS) |
2709 |
{ |
{ |
2710 |
case 0: newline = NEWLINE; break; /* Compile-time default */ |
case 0: newline = NEWLINE; break; /* Compile-time default */ |
2711 |
case PCRE_NEWLINE_CR: newline = '\r'; break; |
case PCRE_NEWLINE_CR: newline = CHAR_CR; break; |
2712 |
case PCRE_NEWLINE_LF: newline = '\n'; break; |
case PCRE_NEWLINE_LF: newline = CHAR_NL; break; |
2713 |
case PCRE_NEWLINE_CR+ |
case PCRE_NEWLINE_CR+ |
2714 |
PCRE_NEWLINE_LF: newline = ('\r' << 8) | '\n'; break; |
PCRE_NEWLINE_LF: newline = (CHAR_CR << 8) | CHAR_NL; break; |
2715 |
case PCRE_NEWLINE_ANY: newline = -1; break; |
case PCRE_NEWLINE_ANY: newline = -1; break; |
2716 |
case PCRE_NEWLINE_ANYCRLF: newline = -2; break; |
case PCRE_NEWLINE_ANYCRLF: newline = -2; break; |
2717 |
default: return PCRE_ERROR_BADNEWLINE; |
default: return PCRE_ERROR_BADNEWLINE; |
2771 |
used in a loop when finding where to start. */ |
used in a loop when finding where to start. */ |
2772 |
|
|
2773 |
lcc = md->tables + lcc_offset; |
lcc = md->tables + lcc_offset; |
2774 |
startline = (re->options & PCRE_STARTLINE) != 0; |
startline = (re->flags & PCRE_STARTLINE) != 0; |
2775 |
firstline = (re->options & PCRE_FIRSTLINE) != 0; |
firstline = (re->options & PCRE_FIRSTLINE) != 0; |
2776 |
|
|
2777 |
/* Set up the first character to match, if available. The first_byte value is |
/* Set up the first character to match, if available. The first_byte value is |
2782 |
|
|
2783 |
if (!anchored) |
if (!anchored) |
2784 |
{ |
{ |
2785 |
if ((re->options & PCRE_FIRSTSET) != 0) |
if ((re->flags & PCRE_FIRSTSET) != 0) |
2786 |
{ |
{ |
2787 |
first_byte = re->first_byte & 255; |
first_byte = re->first_byte & 255; |
2788 |
if ((first_byte_caseless = ((re->first_byte & REQ_CASELESS) != 0)) == TRUE) |
if ((first_byte_caseless = ((re->first_byte & REQ_CASELESS) != 0)) == TRUE) |
2799 |
/* For anchored or unanchored matches, there may be a "last known required |
/* For anchored or unanchored matches, there may be a "last known required |
2800 |
character" set. */ |
character" set. */ |
2801 |
|
|
2802 |
if ((re->options & PCRE_REQCHSET) != 0) |
if ((re->flags & PCRE_REQCHSET) != 0) |
2803 |
{ |
{ |
2804 |
req_byte = re->req_byte & 255; |
req_byte = re->req_byte & 255; |
2805 |
req_byte_caseless = (re->req_byte & REQ_CASELESS) != 0; |
req_byte_caseless = (re->req_byte & REQ_CASELESS) != 0; |
2807 |
} |
} |
2808 |
|
|
2809 |
/* 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 |
2810 |
failed match. Unless restarting, optimize by moving to the first match |
failed match. If not restarting, perform certain optimizations at the start of |
2811 |
character if possible, when not anchored. Then unless wanting a partial match, |
a match. */ |
|
check for a required later character. */ |
|
2812 |
|
|
2813 |
for (;;) |
for (;;) |
2814 |
{ |
{ |
2818 |
{ |
{ |
2819 |
const uschar *save_end_subject = end_subject; |
const uschar *save_end_subject = end_subject; |
2820 |
|
|
2821 |
/* Advance to a unique first char if possible. If firstline is TRUE, the |
/* If firstline is TRUE, the start of the match is constrained to the first |
2822 |
start of the match is constrained to the first line of a multiline string. |
line of a multiline string. Implement this by temporarily adjusting |
2823 |
Implement this by temporarily adjusting end_subject so that we stop |
end_subject so that we stop scanning at a newline. If the match fails at |
2824 |
scanning at a newline. If the match fails at the newline, later code breaks |
the newline, later code breaks this loop. */ |
|
this loop. */ |
|
2825 |
|
|
2826 |
if (firstline) |
if (firstline) |
2827 |
{ |
{ |
2828 |
const uschar *t = current_subject; |
USPTR t = current_subject; |
2829 |
|
#ifdef SUPPORT_UTF8 |
2830 |
|
if (utf8) |
2831 |
|
{ |
2832 |
|
while (t < md->end_subject && !IS_NEWLINE(t)) |
2833 |
|
{ |
2834 |
|
t++; |
2835 |
|
while (t < end_subject && (*t & 0xc0) == 0x80) t++; |
2836 |
|
} |
2837 |
|
} |
2838 |
|
else |
2839 |
|
#endif |
2840 |
while (t < md->end_subject && !IS_NEWLINE(t)) t++; |
while (t < md->end_subject && !IS_NEWLINE(t)) t++; |
2841 |
end_subject = t; |
end_subject = t; |
2842 |
} |
} |
2843 |
|
|
2844 |
if (first_byte >= 0) |
/* There are some optimizations that avoid running the match if a known |
2845 |
|
starting point is not found, or if a known later character is not present. |
2846 |
|
However, there is an option that disables these, for testing and for |
2847 |
|
ensuring that all callouts do actually occur. */ |
2848 |
|
|
2849 |
|
if ((options & PCRE_NO_START_OPTIMIZE) == 0) |
2850 |
{ |
{ |
|
if (first_byte_caseless) |
|
|
while (current_subject < end_subject && |
|
|
lcc[*current_subject] != first_byte) |
|
|
current_subject++; |
|
|
else |
|
|
while (current_subject < end_subject && *current_subject != first_byte) |
|
|
current_subject++; |
|
|
} |
|
2851 |
|
|
2852 |
/* Or to just after a linebreak for a multiline match if possible */ |
/* Advance to a known first byte. */ |
2853 |
|
|
2854 |
else if (startline) |
if (first_byte >= 0) |
2855 |
{ |
{ |
2856 |
if (current_subject > md->start_subject + start_offset) |
if (first_byte_caseless) |
2857 |
|
while (current_subject < end_subject && |
2858 |
|
lcc[*current_subject] != first_byte) |
2859 |
|
current_subject++; |
2860 |
|
else |
2861 |
|
while (current_subject < end_subject && |
2862 |
|
*current_subject != first_byte) |
2863 |
|
current_subject++; |
2864 |
|
} |
2865 |
|
|
2866 |
|
/* Or to just after a linebreak for a multiline match if possible */ |
2867 |
|
|
2868 |
|
else if (startline) |
2869 |
{ |
{ |
2870 |
while (current_subject <= end_subject && !WAS_NEWLINE(current_subject)) |
if (current_subject > md->start_subject + start_offset) |
2871 |
current_subject++; |
{ |
2872 |
|
#ifdef SUPPORT_UTF8 |
2873 |
|
if (utf8) |
2874 |
|
{ |
2875 |
|
while (current_subject < end_subject && |
2876 |
|
!WAS_NEWLINE(current_subject)) |
2877 |
|
{ |
2878 |
|
current_subject++; |
2879 |
|
while(current_subject < end_subject && |
2880 |
|
(*current_subject & 0xc0) == 0x80) |
2881 |
|
current_subject++; |
2882 |
|
} |
2883 |
|
} |
2884 |
|
else |
2885 |
|
#endif |
2886 |
|
while (current_subject < end_subject && !WAS_NEWLINE(current_subject)) |
2887 |
|
current_subject++; |
2888 |
|
|
2889 |
|
/* If we have just passed a CR and the newline option is ANY or |
2890 |
|
ANYCRLF, and we are now at a LF, advance the match position by one |
2891 |
|
more character. */ |
2892 |
|
|
2893 |
/* If we have just passed a CR and the newline option is ANY or |
if (current_subject[-1] == CHAR_CR && |
2894 |
ANYCRLF, and we are now at a LF, advance the match position by one more |
(md->nltype == NLTYPE_ANY || md->nltype == NLTYPE_ANYCRLF) && |
2895 |
character. */ |
current_subject < end_subject && |
2896 |
|
*current_subject == CHAR_NL) |
2897 |
if (current_subject[-1] == '\r' && |
current_subject++; |
2898 |
(md->nltype == NLTYPE_ANY || md->nltype == NLTYPE_ANYCRLF) && |
} |
|
current_subject < end_subject && |
|
|
*current_subject == '\n') |
|
|
current_subject++; |
|
2899 |
} |
} |
|
} |
|
2900 |
|
|
2901 |
/* Or to a non-unique first char after study */ |
/* Or to a non-unique first char after study */ |
2902 |
|
|
2903 |
else if (start_bits != NULL) |
else if (start_bits != NULL) |
|
{ |
|
|
while (current_subject < end_subject) |
|
2904 |
{ |
{ |
2905 |
register unsigned int c = *current_subject; |
while (current_subject < end_subject) |
2906 |
if ((start_bits[c/8] & (1 << (c&7))) == 0) current_subject++; |
{ |
2907 |
else break; |
register unsigned int c = *current_subject; |
2908 |
|
if ((start_bits[c/8] & (1 << (c&7))) == 0) current_subject++; |
2909 |
|
else break; |
2910 |
|
} |
2911 |
} |
} |
2912 |
} |
} |
2913 |
|
|
2929 |
showed up when somebody was matching /^C/ on a 32-megabyte string... so we |
showed up when somebody was matching /^C/ on a 32-megabyte string... so we |
2930 |
don't do this when the string is sufficiently long. |
don't do this when the string is sufficiently long. |
2931 |
|
|
2932 |
ALSO: this processing is disabled when partial matching is requested. |
ALSO: this processing is disabled when partial matching is requested, and can |
2933 |
*/ |
also be explicitly deactivated. Furthermore, we have to disable when |
2934 |
|
restarting after a partial match, because the required character may have |
2935 |
|
already been matched. */ |
2936 |
|
|
2937 |
if (req_byte >= 0 && |
if ((options & PCRE_NO_START_OPTIMIZE) == 0 && |
2938 |
|
req_byte >= 0 && |
2939 |
end_subject - current_subject < REQ_BYTE_MAX && |
end_subject - current_subject < REQ_BYTE_MAX && |
2940 |
(options & PCRE_PARTIAL) == 0) |
(options & (PCRE_PARTIAL_HARD|PCRE_PARTIAL_SOFT|PCRE_DFA_RESTART)) == 0) |
2941 |
{ |
{ |
2942 |
register const uschar *p = current_subject + ((first_byte >= 0)? 1 : 0); |
register const uschar *p = current_subject + ((first_byte >= 0)? 1 : 0); |
2943 |
|
|
2977 |
|
|
2978 |
/* OK, now we can do the business */ |
/* OK, now we can do the business */ |
2979 |
|
|
2980 |
|
md->start_used_ptr = current_subject; |
2981 |
|
|
2982 |
rc = internal_dfa_exec( |
rc = internal_dfa_exec( |
2983 |
md, /* fixed match data */ |
md, /* fixed match data */ |
2984 |
md->start_code, /* this subexpression's code */ |
md->start_code, /* this subexpression's code */ |
3009 |
} |
} |
3010 |
if (current_subject > end_subject) break; |
if (current_subject > end_subject) break; |
3011 |
|
|
3012 |
/* If we have just passed a CR and the newline option is CRLF or ANY or |
/* If we have just passed a CR and we are now at a LF, and the pattern does |
3013 |
ANYCRLF, and we are now at a LF, advance the match position by one more |
not contain any explicit matches for \r or \n, and the newline option is CRLF |
3014 |
character. */ |
or ANY or ANYCRLF, advance the match position by one more character. */ |
3015 |
|
|
3016 |
if (current_subject[-1] == '\r' && |
if (current_subject[-1] == CHAR_CR && |
3017 |
(md->nltype == NLTYPE_ANY || |
current_subject < end_subject && |
3018 |
md->nltype == NLTYPE_ANYCRLF || |
*current_subject == CHAR_NL && |
3019 |
md->nllen == 2) && |
(re->flags & PCRE_HASCRORLF) == 0 && |
3020 |
current_subject < end_subject && |
(md->nltype == NLTYPE_ANY || |
3021 |
*current_subject == '\n') |
md->nltype == NLTYPE_ANYCRLF || |
3022 |
|
md->nllen == 2)) |
3023 |
current_subject++; |
current_subject++; |
3024 |
|
|
3025 |
} /* "Bumpalong" loop */ |
} /* "Bumpalong" loop */ |