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) 1997-2008 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 |
61 |
#define SP " " |
#define SP " " |
62 |
|
|
63 |
|
|
|
|
|
64 |
/************************************************* |
/************************************************* |
65 |
* Code parameters and static tables * |
* Code parameters and static tables * |
66 |
*************************************************/ |
*************************************************/ |
511 |
stateblock *current_state = active_states + i; |
stateblock *current_state = active_states + i; |
512 |
const uschar *code; |
const uschar *code; |
513 |
int state_offset = current_state->offset; |
int state_offset = current_state->offset; |
514 |
int count, codevalue; |
int count, codevalue, rrc; |
515 |
|
|
516 |
#ifdef DEBUG |
#ifdef DEBUG |
517 |
printf ("%.*sProcessing state %d c=", rlevel*2-2, SP, state_offset); |
printf ("%.*sProcessing state %d c=", rlevel*2-2, SP, state_offset); |
757 |
if ((md->moptions & PCRE_NOTEOL) == 0) |
if ((md->moptions & PCRE_NOTEOL) == 0) |
758 |
{ |
{ |
759 |
if (clen == 0 || |
if (clen == 0 || |
760 |
(IS_NEWLINE(ptr) && |
((md->poptions & PCRE_DOLLAR_ENDONLY) == 0 && IS_NEWLINE(ptr) && |
761 |
((ims & PCRE_MULTILINE) != 0 || ptr == end_subject - md->nllen) |
((ims & PCRE_MULTILINE) != 0 || ptr == end_subject - md->nllen) |
762 |
)) |
)) |
763 |
{ ADD_ACTIVE(state_offset + 1, 0); } |
{ ADD_ACTIVE(state_offset + 1, 0); } |
834 |
break; |
break; |
835 |
|
|
836 |
case PT_GC: |
case PT_GC: |
837 |
OK = ucp_gentype[prop->chartype] == code[2]; |
OK = _pcre_ucp_gentype[prop->chartype] == code[2]; |
838 |
break; |
break; |
839 |
|
|
840 |
case PT_PC: |
case PT_PC: |
1003 |
break; |
break; |
1004 |
|
|
1005 |
case PT_GC: |
case PT_GC: |
1006 |
OK = ucp_gentype[prop->chartype] == code[3]; |
OK = _pcre_ucp_gentype[prop->chartype] == code[3]; |
1007 |
break; |
break; |
1008 |
|
|
1009 |
case PT_PC: |
case PT_PC: |
1225 |
break; |
break; |
1226 |
|
|
1227 |
case PT_GC: |
case PT_GC: |
1228 |
OK = ucp_gentype[prop->chartype] == code[3]; |
OK = _pcre_ucp_gentype[prop->chartype] == code[3]; |
1229 |
break; |
break; |
1230 |
|
|
1231 |
case PT_PC: |
case PT_PC: |
1472 |
break; |
break; |
1473 |
|
|
1474 |
case PT_GC: |
case PT_GC: |
1475 |
OK = ucp_gentype[prop->chartype] == code[5]; |
OK = _pcre_ucp_gentype[prop->chartype] == code[5]; |
1476 |
break; |
break; |
1477 |
|
|
1478 |
case PT_PC: |
case PT_PC: |
2157 |
|
|
2158 |
/* ========================================================================== */ |
/* ========================================================================== */ |
2159 |
/* These are the opcodes for fancy brackets of various kinds. We have |
/* These are the opcodes for fancy brackets of various kinds. We have |
2160 |
to use recursion in order to handle them. The "always failing" assersion |
to use recursion in order to handle them. The "always failing" assertion |
2161 |
(?!) is optimised when compiling to OP_FAIL, so we have to support that, |
(?!) is optimised to OP_FAIL when compiling, so we have to support that, |
2162 |
though the other "backtracking verbs" are not supported. */ |
though the other "backtracking verbs" are not supported. */ |
2163 |
|
|
2164 |
case OP_FAIL: |
case OP_FAIL: |
2200 |
{ |
{ |
2201 |
int local_offsets[1000]; |
int local_offsets[1000]; |
2202 |
int local_workspace[1000]; |
int local_workspace[1000]; |
2203 |
int condcode = code[LINK_SIZE+1]; |
int codelink = GET(code, 1); |
2204 |
|
int condcode; |
2205 |
|
|
2206 |
|
/* Because of the way auto-callout works during compile, a callout item |
2207 |
|
is inserted between OP_COND and an assertion condition. This does not |
2208 |
|
happen for the other conditions. */ |
2209 |
|
|
2210 |
|
if (code[LINK_SIZE+1] == OP_CALLOUT) |
2211 |
|
{ |
2212 |
|
rrc = 0; |
2213 |
|
if (pcre_callout != NULL) |
2214 |
|
{ |
2215 |
|
pcre_callout_block cb; |
2216 |
|
cb.version = 1; /* Version 1 of the callout block */ |
2217 |
|
cb.callout_number = code[LINK_SIZE+2]; |
2218 |
|
cb.offset_vector = offsets; |
2219 |
|
cb.subject = (PCRE_SPTR)start_subject; |
2220 |
|
cb.subject_length = end_subject - start_subject; |
2221 |
|
cb.start_match = current_subject - start_subject; |
2222 |
|
cb.current_position = ptr - start_subject; |
2223 |
|
cb.pattern_position = GET(code, LINK_SIZE + 3); |
2224 |
|
cb.next_item_length = GET(code, 3 + 2*LINK_SIZE); |
2225 |
|
cb.capture_top = 1; |
2226 |
|
cb.capture_last = -1; |
2227 |
|
cb.callout_data = md->callout_data; |
2228 |
|
if ((rrc = (*pcre_callout)(&cb)) < 0) return rrc; /* Abandon */ |
2229 |
|
} |
2230 |
|
if (rrc > 0) break; /* Fail this thread */ |
2231 |
|
code += _pcre_OP_lengths[OP_CALLOUT]; /* Skip callout data */ |
2232 |
|
} |
2233 |
|
|
2234 |
|
condcode = code[LINK_SIZE+1]; |
2235 |
|
|
2236 |
/* Back reference conditions are not supported */ |
/* Back reference conditions are not supported */ |
2237 |
|
|
2240 |
/* The DEFINE condition is always false */ |
/* The DEFINE condition is always false */ |
2241 |
|
|
2242 |
if (condcode == OP_DEF) |
if (condcode == OP_DEF) |
2243 |
{ |
{ ADD_ACTIVE(state_offset + codelink + LINK_SIZE + 1, 0); } |
|
ADD_ACTIVE(state_offset + GET(code, 1) + LINK_SIZE + 1, 0); |
|
|
} |
|
2244 |
|
|
2245 |
/* 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, |
2246 |
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 |
2250 |
{ |
{ |
2251 |
int value = GET2(code, LINK_SIZE+2); |
int value = GET2(code, LINK_SIZE+2); |
2252 |
if (value != RREF_ANY) return PCRE_ERROR_DFA_UCOND; |
if (value != RREF_ANY) return PCRE_ERROR_DFA_UCOND; |
2253 |
if (recursing > 0) { ADD_ACTIVE(state_offset + LINK_SIZE + 4, 0); } |
if (recursing > 0) |
2254 |
else { ADD_ACTIVE(state_offset + GET(code, 1) + LINK_SIZE + 1, 0); } |
{ ADD_ACTIVE(state_offset + LINK_SIZE + 4, 0); } |
2255 |
|
else { ADD_ACTIVE(state_offset + codelink + LINK_SIZE + 1, 0); } |
2256 |
} |
} |
2257 |
|
|
2258 |
/* Otherwise, the condition is an assertion */ |
/* Otherwise, the condition is an assertion */ |
2282 |
(condcode == OP_ASSERT || condcode == OP_ASSERTBACK)) |
(condcode == OP_ASSERT || condcode == OP_ASSERTBACK)) |
2283 |
{ ADD_ACTIVE(endasscode + LINK_SIZE + 1 - start_code, 0); } |
{ ADD_ACTIVE(endasscode + LINK_SIZE + 1 - start_code, 0); } |
2284 |
else |
else |
2285 |
{ ADD_ACTIVE(state_offset + GET(code, 1) + LINK_SIZE + 1, 0); } |
{ ADD_ACTIVE(state_offset + codelink + LINK_SIZE + 1, 0); } |
2286 |
} |
} |
2287 |
} |
} |
2288 |
break; |
break; |
2434 |
/* Handle callouts */ |
/* Handle callouts */ |
2435 |
|
|
2436 |
case OP_CALLOUT: |
case OP_CALLOUT: |
2437 |
|
rrc = 0; |
2438 |
if (pcre_callout != NULL) |
if (pcre_callout != NULL) |
2439 |
{ |
{ |
|
int rrc; |
|
2440 |
pcre_callout_block cb; |
pcre_callout_block cb; |
2441 |
cb.version = 1; /* Version 1 of the callout block */ |
cb.version = 1; /* Version 1 of the callout block */ |
2442 |
cb.callout_number = code[1]; |
cb.callout_number = code[1]; |
2451 |
cb.capture_last = -1; |
cb.capture_last = -1; |
2452 |
cb.callout_data = md->callout_data; |
cb.callout_data = md->callout_data; |
2453 |
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); } |
|
2454 |
} |
} |
2455 |
|
if (rrc == 0) |
2456 |
|
{ ADD_ACTIVE(state_offset + _pcre_OP_lengths[OP_CALLOUT], 0); } |
2457 |
break; |
break; |
2458 |
|
|
2459 |
|
|
2536 |
< -1 => some kind of unexpected problem |
< -1 => some kind of unexpected problem |
2537 |
*/ |
*/ |
2538 |
|
|
2539 |
PCRE_EXP_DEFN int |
PCRE_EXP_DEFN int PCRE_CALL_CONVENTION |
2540 |
pcre_dfa_exec(const pcre *argument_re, const pcre_extra *extra_data, |
pcre_dfa_exec(const pcre *argument_re, const pcre_extra *extra_data, |
2541 |
const char *subject, int length, int start_offset, int options, int *offsets, |
const char *subject, int length, int start_offset, int options, int *offsets, |
2542 |
int offsetcount, int *workspace, int wscount) |
int offsetcount, int *workspace, int wscount) |
2645 |
PCRE_NEWLINE_BITS) |
PCRE_NEWLINE_BITS) |
2646 |
{ |
{ |
2647 |
case 0: newline = NEWLINE; break; /* Compile-time default */ |
case 0: newline = NEWLINE; break; /* Compile-time default */ |
2648 |
case PCRE_NEWLINE_CR: newline = '\r'; break; |
case PCRE_NEWLINE_CR: newline = CHAR_CR; break; |
2649 |
case PCRE_NEWLINE_LF: newline = '\n'; break; |
case PCRE_NEWLINE_LF: newline = CHAR_NL; break; |
2650 |
case PCRE_NEWLINE_CR+ |
case PCRE_NEWLINE_CR+ |
2651 |
PCRE_NEWLINE_LF: newline = ('\r' << 8) | '\n'; break; |
PCRE_NEWLINE_LF: newline = (CHAR_CR << 8) | CHAR_NL; break; |
2652 |
case PCRE_NEWLINE_ANY: newline = -1; break; |
case PCRE_NEWLINE_ANY: newline = -1; break; |
2653 |
case PCRE_NEWLINE_ANYCRLF: newline = -2; break; |
case PCRE_NEWLINE_ANYCRLF: newline = -2; break; |
2654 |
default: return PCRE_ERROR_BADNEWLINE; |
default: return PCRE_ERROR_BADNEWLINE; |
2744 |
} |
} |
2745 |
|
|
2746 |
/* 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 |
2747 |
failed match. Unless restarting, optimize by moving to the first match |
failed match. If not restarting, perform certain optimizations at the start of |
2748 |
character if possible, when not anchored. Then unless wanting a partial match, |
a match. */ |
|
check for a required later character. */ |
|
2749 |
|
|
2750 |
for (;;) |
for (;;) |
2751 |
{ |
{ |
2755 |
{ |
{ |
2756 |
const uschar *save_end_subject = end_subject; |
const uschar *save_end_subject = end_subject; |
2757 |
|
|
2758 |
/* 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 |
2759 |
start of the match is constrained to the first line of a multiline string. |
line of a multiline string. Implement this by temporarily adjusting |
2760 |
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 |
2761 |
scanning at a newline. If the match fails at the newline, later code breaks |
the newline, later code breaks this loop. */ |
|
this loop. */ |
|
2762 |
|
|
2763 |
if (firstline) |
if (firstline) |
2764 |
{ |
{ |
2765 |
const uschar *t = current_subject; |
USPTR t = current_subject; |
2766 |
|
#ifdef SUPPORT_UTF8 |
2767 |
|
if (utf8) |
2768 |
|
{ |
2769 |
|
while (t < md->end_subject && !IS_NEWLINE(t)) |
2770 |
|
{ |
2771 |
|
t++; |
2772 |
|
while (t < end_subject && (*t & 0xc0) == 0x80) t++; |
2773 |
|
} |
2774 |
|
} |
2775 |
|
else |
2776 |
|
#endif |
2777 |
while (t < md->end_subject && !IS_NEWLINE(t)) t++; |
while (t < md->end_subject && !IS_NEWLINE(t)) t++; |
2778 |
end_subject = t; |
end_subject = t; |
2779 |
} |
} |
2780 |
|
|
2781 |
if (first_byte >= 0) |
/* There are some optimizations that avoid running the match if a known |
2782 |
|
starting point is not found, or if a known later character is not present. |
2783 |
|
However, there is an option that disables these, for testing and for |
2784 |
|
ensuring that all callouts do actually occur. */ |
2785 |
|
|
2786 |
|
if ((options & PCRE_NO_START_OPTIMIZE) == 0) |
2787 |
{ |
{ |
|
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++; |
|
|
} |
|
2788 |
|
|
2789 |
/* Or to just after a linebreak for a multiline match if possible */ |
/* Advance to a known first byte. */ |
2790 |
|
|
2791 |
else if (startline) |
if (first_byte >= 0) |
|
{ |
|
|
if (current_subject > md->start_subject + start_offset) |
|
2792 |
{ |
{ |
2793 |
while (current_subject <= end_subject && !WAS_NEWLINE(current_subject)) |
if (first_byte_caseless) |
2794 |
current_subject++; |
while (current_subject < end_subject && |
2795 |
|
lcc[*current_subject] != first_byte) |
2796 |
|
current_subject++; |
2797 |
|
else |
2798 |
|
while (current_subject < end_subject && |
2799 |
|
*current_subject != first_byte) |
2800 |
|
current_subject++; |
2801 |
|
} |
2802 |
|
|
2803 |
|
/* Or to just after a linebreak for a multiline match if possible */ |
2804 |
|
|
2805 |
/* If we have just passed a CR and the newline option is ANY or |
else if (startline) |
2806 |
ANYCRLF, and we are now at a LF, advance the match position by one more |
{ |
2807 |
character. */ |
if (current_subject > md->start_subject + start_offset) |
2808 |
|
{ |
2809 |
if (current_subject[-1] == '\r' && |
#ifdef SUPPORT_UTF8 |
2810 |
(md->nltype == NLTYPE_ANY || md->nltype == NLTYPE_ANYCRLF) && |
if (utf8) |
2811 |
current_subject < end_subject && |
{ |
2812 |
*current_subject == '\n') |
while (current_subject < end_subject && |
2813 |
current_subject++; |
!WAS_NEWLINE(current_subject)) |
2814 |
|
{ |
2815 |
|
current_subject++; |
2816 |
|
while(current_subject < end_subject && |
2817 |
|
(*current_subject & 0xc0) == 0x80) |
2818 |
|
current_subject++; |
2819 |
|
} |
2820 |
|
} |
2821 |
|
else |
2822 |
|
#endif |
2823 |
|
while (current_subject < end_subject && !WAS_NEWLINE(current_subject)) |
2824 |
|
current_subject++; |
2825 |
|
|
2826 |
|
/* If we have just passed a CR and the newline option is ANY or |
2827 |
|
ANYCRLF, and we are now at a LF, advance the match position by one |
2828 |
|
more character. */ |
2829 |
|
|
2830 |
|
if (current_subject[-1] == CHAR_CR && |
2831 |
|
(md->nltype == NLTYPE_ANY || md->nltype == NLTYPE_ANYCRLF) && |
2832 |
|
current_subject < end_subject && |
2833 |
|
*current_subject == CHAR_NL) |
2834 |
|
current_subject++; |
2835 |
|
} |
2836 |
} |
} |
|
} |
|
2837 |
|
|
2838 |
/* Or to a non-unique first char after study */ |
/* Or to a non-unique first char after study */ |
2839 |
|
|
2840 |
else if (start_bits != NULL) |
else if (start_bits != NULL) |
|
{ |
|
|
while (current_subject < end_subject) |
|
2841 |
{ |
{ |
2842 |
register unsigned int c = *current_subject; |
while (current_subject < end_subject) |
2843 |
if ((start_bits[c/8] & (1 << (c&7))) == 0) current_subject++; |
{ |
2844 |
else break; |
register unsigned int c = *current_subject; |
2845 |
|
if ((start_bits[c/8] & (1 << (c&7))) == 0) current_subject++; |
2846 |
|
else break; |
2847 |
|
} |
2848 |
} |
} |
2849 |
} |
} |
2850 |
|
|
2866 |
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 |
2867 |
don't do this when the string is sufficiently long. |
don't do this when the string is sufficiently long. |
2868 |
|
|
2869 |
ALSO: this processing is disabled when partial matching is requested. |
ALSO: this processing is disabled when partial matching is requested, and can |
2870 |
*/ |
also be explicitly deactivated. */ |
2871 |
|
|
2872 |
if (req_byte >= 0 && |
if ((options & PCRE_NO_START_OPTIMIZE) == 0 && |
2873 |
|
req_byte >= 0 && |
2874 |
end_subject - current_subject < REQ_BYTE_MAX && |
end_subject - current_subject < REQ_BYTE_MAX && |
2875 |
(options & PCRE_PARTIAL) == 0) |
(options & PCRE_PARTIAL) == 0) |
2876 |
{ |
{ |
2946 |
not contain any explicit matches for \r or \n, and the newline option is CRLF |
not contain any explicit matches for \r or \n, and the newline option is CRLF |
2947 |
or ANY or ANYCRLF, advance the match position by one more character. */ |
or ANY or ANYCRLF, advance the match position by one more character. */ |
2948 |
|
|
2949 |
if (current_subject[-1] == '\r' && |
if (current_subject[-1] == CHAR_CR && |
2950 |
current_subject < end_subject && |
current_subject < end_subject && |
2951 |
*current_subject == '\n' && |
*current_subject == CHAR_NL && |
2952 |
(re->flags & PCRE_HASCRORLF) == 0 && |
(re->flags & PCRE_HASCRORLF) == 0 && |
2953 |
(md->nltype == NLTYPE_ANY || |
(md->nltype == NLTYPE_ANY || |
2954 |
md->nltype == NLTYPE_ANYCRLF || |
md->nltype == NLTYPE_ANYCRLF || |