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) 1997-2010 University of Cambridge |
Copyright (c) 1997-2011 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 |
57 |
#undef min |
#undef min |
58 |
#undef max |
#undef max |
59 |
|
|
60 |
/* Flag bits for the match() function */ |
/* Values for setting in md->match_function_type to indicate two special types |
61 |
|
of call to match(). We do it this way to save on using another stack variable, |
62 |
|
as stack usage is to be discouraged. */ |
63 |
|
|
64 |
#define match_condassert 0x01 /* Called to check a condition assertion */ |
#define MATCH_CONDASSERT 1 /* Called to check a condition assertion */ |
65 |
#define match_cbegroup 0x02 /* Could-be-empty unlimited repeat group */ |
#define MATCH_CBEGROUP 2 /* Could-be-empty unlimited repeat group */ |
66 |
|
|
67 |
/* Non-error returns from the match() function. Error returns are externally |
/* Non-error returns from the match() function. Error returns are externally |
68 |
defined PCRE_ERROR_xxx codes, which are all negative. */ |
defined PCRE_ERROR_xxx codes, which are all negative. */ |
75 |
|
|
76 |
#define MATCH_ACCEPT (-999) |
#define MATCH_ACCEPT (-999) |
77 |
#define MATCH_COMMIT (-998) |
#define MATCH_COMMIT (-998) |
78 |
#define MATCH_PRUNE (-997) |
#define MATCH_KETRPOS (-997) |
79 |
#define MATCH_SKIP (-996) |
#define MATCH_ONCE (-996) |
80 |
#define MATCH_SKIP_ARG (-995) |
#define MATCH_PRUNE (-995) |
81 |
#define MATCH_THEN (-994) |
#define MATCH_SKIP (-994) |
82 |
|
#define MATCH_SKIP_ARG (-993) |
83 |
|
#define MATCH_THEN (-992) |
84 |
|
|
85 |
/* This is a convenience macro for code that occurs many times. */ |
/* This is a convenience macro for code that occurs many times. */ |
86 |
|
|
136 |
* Match a back-reference * |
* Match a back-reference * |
137 |
*************************************************/ |
*************************************************/ |
138 |
|
|
139 |
/* If a back reference hasn't been set, the length that is passed is greater |
/* Normally, if a back reference hasn't been set, the length that is passed is |
140 |
than the number of characters left in the string, so the match fails. |
negative, so the match always fails. However, in JavaScript compatibility mode, |
141 |
|
the length passed is zero. Note that in caseless UTF-8 mode, the number of |
142 |
|
subject bytes matched may be different to the number of reference bytes. |
143 |
|
|
144 |
Arguments: |
Arguments: |
145 |
offset index into the offset vector |
offset index into the offset vector |
146 |
eptr points into the subject |
eptr pointer into the subject |
147 |
length length to be matched |
length length of reference to be matched (number of bytes) |
148 |
md points to match data block |
md points to match data block |
149 |
ims the ims flags |
caseless TRUE if caseless |
150 |
|
|
151 |
Returns: TRUE if matched |
Returns: < 0 if not matched, otherwise the number of subject bytes matched |
152 |
*/ |
*/ |
153 |
|
|
154 |
static BOOL |
static int |
155 |
match_ref(int offset, register USPTR eptr, int length, match_data *md, |
match_ref(int offset, register USPTR eptr, int length, match_data *md, |
156 |
unsigned long int ims) |
BOOL caseless) |
157 |
{ |
{ |
158 |
USPTR p = md->start_subject + md->offset_vector[offset]; |
USPTR eptr_start = eptr; |
159 |
|
register USPTR p = md->start_subject + md->offset_vector[offset]; |
160 |
|
|
161 |
#ifdef PCRE_DEBUG |
#ifdef PCRE_DEBUG |
162 |
if (eptr >= md->end_subject) |
if (eptr >= md->end_subject) |
171 |
printf("\n"); |
printf("\n"); |
172 |
#endif |
#endif |
173 |
|
|
174 |
/* Always fail if not enough characters left */ |
/* Always fail if reference not set (and not JavaScript compatible). */ |
175 |
|
|
176 |
if (length > md->end_subject - eptr) return FALSE; |
if (length < 0) return -1; |
177 |
|
|
178 |
/* Separate the caseless case for speed. In UTF-8 mode we can only do this |
/* Separate the caseless case for speed. In UTF-8 mode we can only do this |
179 |
properly if Unicode properties are supported. Otherwise, we can check only |
properly if Unicode properties are supported. Otherwise, we can check only |
180 |
ASCII characters. */ |
ASCII characters. */ |
181 |
|
|
182 |
if ((ims & PCRE_CASELESS) != 0) |
if (caseless) |
183 |
{ |
{ |
184 |
#ifdef SUPPORT_UTF8 |
#ifdef SUPPORT_UTF8 |
185 |
#ifdef SUPPORT_UCP |
#ifdef SUPPORT_UCP |
186 |
if (md->utf8) |
if (md->utf8) |
187 |
{ |
{ |
188 |
USPTR endptr = eptr + length; |
/* Match characters up to the end of the reference. NOTE: the number of |
189 |
while (eptr < endptr) |
bytes matched may differ, because there are some characters whose upper and |
190 |
|
lower case versions code as different numbers of bytes. For example, U+023A |
191 |
|
(2 bytes in UTF-8) is the upper case version of U+2C65 (3 bytes in UTF-8); |
192 |
|
a sequence of 3 of the former uses 6 bytes, as does a sequence of two of |
193 |
|
the latter. It is important, therefore, to check the length along the |
194 |
|
reference, not along the subject (earlier code did this wrong). */ |
195 |
|
|
196 |
|
USPTR endptr = p + length; |
197 |
|
while (p < endptr) |
198 |
{ |
{ |
199 |
int c, d; |
int c, d; |
200 |
|
if (eptr >= md->end_subject) return -1; |
201 |
GETCHARINC(c, eptr); |
GETCHARINC(c, eptr); |
202 |
GETCHARINC(d, p); |
GETCHARINC(d, p); |
203 |
if (c != d && c != UCD_OTHERCASE(d)) return FALSE; |
if (c != d && c != UCD_OTHERCASE(d)) return -1; |
204 |
} |
} |
205 |
} |
} |
206 |
else |
else |
209 |
|
|
210 |
/* The same code works when not in UTF-8 mode and in UTF-8 mode when there |
/* The same code works when not in UTF-8 mode and in UTF-8 mode when there |
211 |
is no UCP support. */ |
is no UCP support. */ |
212 |
|
{ |
213 |
while (length-- > 0) |
if (eptr + length > md->end_subject) return -1; |
214 |
{ if (md->lcc[*p++] != md->lcc[*eptr++]) return FALSE; } |
while (length-- > 0) |
215 |
|
{ if (md->lcc[*p++] != md->lcc[*eptr++]) return -1; } |
216 |
|
} |
217 |
} |
} |
218 |
|
|
219 |
/* In the caseful case, we can just compare the bytes, whether or not we |
/* In the caseful case, we can just compare the bytes, whether or not we |
220 |
are in UTF-8 mode. */ |
are in UTF-8 mode. */ |
221 |
|
|
222 |
else |
else |
223 |
{ while (length-- > 0) if (*p++ != *eptr++) return FALSE; } |
{ |
224 |
|
if (eptr + length > md->end_subject) return -1; |
225 |
|
while (length-- > 0) if (*p++ != *eptr++) return -1; |
226 |
|
} |
227 |
|
|
228 |
return TRUE; |
return eptr - eptr_start; |
229 |
} |
} |
230 |
|
|
231 |
|
|
277 |
RM31, RM32, RM33, RM34, RM35, RM36, RM37, RM38, RM39, RM40, |
RM31, RM32, RM33, RM34, RM35, RM36, RM37, RM38, RM39, RM40, |
278 |
RM41, RM42, RM43, RM44, RM45, RM46, RM47, RM48, RM49, RM50, |
RM41, RM42, RM43, RM44, RM45, RM46, RM47, RM48, RM49, RM50, |
279 |
RM51, RM52, RM53, RM54, RM55, RM56, RM57, RM58, RM59, RM60, |
RM51, RM52, RM53, RM54, RM55, RM56, RM57, RM58, RM59, RM60, |
280 |
RM61, RM62 }; |
RM61, RM62, RM63 }; |
281 |
|
|
282 |
/* These versions of the macros use the stack, as normal. There are debugging |
/* These versions of the macros use the stack, as normal. There are debugging |
283 |
versions and production versions. Note that the "rw" argument of RMATCH isn't |
versions and production versions. Note that the "rw" argument of RMATCH isn't |
287 |
#define REGISTER register |
#define REGISTER register |
288 |
|
|
289 |
#ifdef PCRE_DEBUG |
#ifdef PCRE_DEBUG |
290 |
#define RMATCH(ra,rb,rc,rd,re,rf,rg,rw) \ |
#define RMATCH(ra,rb,rc,rd,re,rw) \ |
291 |
{ \ |
{ \ |
292 |
printf("match() called in line %d\n", __LINE__); \ |
printf("match() called in line %d\n", __LINE__); \ |
293 |
rrc = match(ra,rb,mstart,markptr,rc,rd,re,rf,rg,rdepth+1); \ |
rrc = match(ra,rb,mstart,markptr,rc,rd,re,rdepth+1); \ |
294 |
printf("to line %d\n", __LINE__); \ |
printf("to line %d\n", __LINE__); \ |
295 |
} |
} |
296 |
#define RRETURN(ra) \ |
#define RRETURN(ra) \ |
299 |
return ra; \ |
return ra; \ |
300 |
} |
} |
301 |
#else |
#else |
302 |
#define RMATCH(ra,rb,rc,rd,re,rf,rg,rw) \ |
#define RMATCH(ra,rb,rc,rd,re,rw) \ |
303 |
rrc = match(ra,rb,mstart,markptr,rc,rd,re,rf,rg,rdepth+1) |
rrc = match(ra,rb,mstart,markptr,rc,rd,re,rdepth+1) |
304 |
#define RRETURN(ra) return ra |
#define RRETURN(ra) return ra |
305 |
#endif |
#endif |
306 |
|
|
313 |
|
|
314 |
#define REGISTER |
#define REGISTER |
315 |
|
|
316 |
#define RMATCH(ra,rb,rc,rd,re,rf,rg,rw)\ |
#define RMATCH(ra,rb,rc,rd,re,rw)\ |
317 |
{\ |
{\ |
318 |
heapframe *newframe = (heapframe *)(pcre_stack_malloc)(sizeof(heapframe));\ |
heapframe *newframe = (heapframe *)(pcre_stack_malloc)(sizeof(heapframe));\ |
319 |
if (newframe == NULL) RRETURN(PCRE_ERROR_NOMEMORY);\ |
if (newframe == NULL) RRETURN(PCRE_ERROR_NOMEMORY);\ |
323 |
newframe->Xmstart = mstart;\ |
newframe->Xmstart = mstart;\ |
324 |
newframe->Xmarkptr = markptr;\ |
newframe->Xmarkptr = markptr;\ |
325 |
newframe->Xoffset_top = rc;\ |
newframe->Xoffset_top = rc;\ |
326 |
newframe->Xims = re;\ |
newframe->Xeptrb = re;\ |
|
newframe->Xeptrb = rf;\ |
|
|
newframe->Xflags = rg;\ |
|
327 |
newframe->Xrdepth = frame->Xrdepth + 1;\ |
newframe->Xrdepth = frame->Xrdepth + 1;\ |
328 |
newframe->Xprevframe = frame;\ |
newframe->Xprevframe = frame;\ |
329 |
frame = newframe;\ |
frame = newframe;\ |
359 |
USPTR Xmstart; |
USPTR Xmstart; |
360 |
USPTR Xmarkptr; |
USPTR Xmarkptr; |
361 |
int Xoffset_top; |
int Xoffset_top; |
|
long int Xims; |
|
362 |
eptrblock *Xeptrb; |
eptrblock *Xeptrb; |
|
int Xflags; |
|
363 |
unsigned int Xrdepth; |
unsigned int Xrdepth; |
364 |
|
|
365 |
/* Function local variables */ |
/* Function local variables */ |
380 |
BOOL Xcondition; |
BOOL Xcondition; |
381 |
BOOL Xprev_is_word; |
BOOL Xprev_is_word; |
382 |
|
|
|
unsigned long int Xoriginal_ims; |
|
|
|
|
383 |
#ifdef SUPPORT_UCP |
#ifdef SUPPORT_UCP |
384 |
int Xprop_type; |
int Xprop_type; |
385 |
int Xprop_value; |
int Xprop_value; |
386 |
int Xprop_fail_result; |
int Xprop_fail_result; |
|
int Xprop_category; |
|
|
int Xprop_chartype; |
|
|
int Xprop_script; |
|
387 |
int Xoclength; |
int Xoclength; |
388 |
uschar Xocchars[8]; |
uschar Xocchars[8]; |
389 |
#endif |
#endif |
462 |
markptr pointer to the most recent MARK name, or NULL |
markptr pointer to the most recent MARK name, or NULL |
463 |
offset_top current top pointer |
offset_top current top pointer |
464 |
md pointer to "static" info for the match |
md pointer to "static" info for the match |
|
ims current /i, /m, and /s options |
|
465 |
eptrb pointer to chain of blocks containing eptr at start of |
eptrb pointer to chain of blocks containing eptr at start of |
466 |
brackets - for testing for empty matches |
brackets - for testing for empty matches |
|
flags can contain |
|
|
match_condassert - this is an assertion condition |
|
|
match_cbegroup - this is the start of an unlimited repeat |
|
|
group that can match an empty string |
|
467 |
rdepth the recursion depth |
rdepth the recursion depth |
468 |
|
|
469 |
Returns: MATCH_MATCH if matched ) these values are >= 0 |
Returns: MATCH_MATCH if matched ) these values are >= 0 |
475 |
|
|
476 |
static int |
static int |
477 |
match(REGISTER USPTR eptr, REGISTER const uschar *ecode, USPTR mstart, |
match(REGISTER USPTR eptr, REGISTER const uschar *ecode, USPTR mstart, |
478 |
const uschar *markptr, int offset_top, match_data *md, unsigned long int ims, |
const uschar *markptr, int offset_top, match_data *md, eptrblock *eptrb, |
479 |
eptrblock *eptrb, int flags, unsigned int rdepth) |
unsigned int rdepth) |
480 |
{ |
{ |
481 |
/* These variables do not need to be preserved over recursion in this function, |
/* These variables do not need to be preserved over recursion in this function, |
482 |
so they can be ordinary variables in all cases. Mark some of them with |
so they can be ordinary variables in all cases. Mark some of them with |
488 |
register BOOL utf8; /* Local copy of UTF-8 flag for speed */ |
register BOOL utf8; /* Local copy of UTF-8 flag for speed */ |
489 |
|
|
490 |
BOOL minimize, possessive; /* Quantifier options */ |
BOOL minimize, possessive; /* Quantifier options */ |
491 |
|
BOOL caseless; |
492 |
int condcode; |
int condcode; |
493 |
|
|
494 |
/* When recursion is not being used, all "local" variables that have to be |
/* When recursion is not being used, all "local" variables that have to be |
508 |
frame->Xmstart = mstart; |
frame->Xmstart = mstart; |
509 |
frame->Xmarkptr = markptr; |
frame->Xmarkptr = markptr; |
510 |
frame->Xoffset_top = offset_top; |
frame->Xoffset_top = offset_top; |
|
frame->Xims = ims; |
|
511 |
frame->Xeptrb = eptrb; |
frame->Xeptrb = eptrb; |
|
frame->Xflags = flags; |
|
512 |
frame->Xrdepth = rdepth; |
frame->Xrdepth = rdepth; |
513 |
|
|
514 |
/* This is where control jumps back to to effect "recursion" */ |
/* This is where control jumps back to to effect "recursion" */ |
522 |
#define mstart frame->Xmstart |
#define mstart frame->Xmstart |
523 |
#define markptr frame->Xmarkptr |
#define markptr frame->Xmarkptr |
524 |
#define offset_top frame->Xoffset_top |
#define offset_top frame->Xoffset_top |
|
#define ims frame->Xims |
|
525 |
#define eptrb frame->Xeptrb |
#define eptrb frame->Xeptrb |
|
#define flags frame->Xflags |
|
526 |
#define rdepth frame->Xrdepth |
#define rdepth frame->Xrdepth |
527 |
|
|
528 |
/* Ditto for the local variables */ |
/* Ditto for the local variables */ |
544 |
#define condition frame->Xcondition |
#define condition frame->Xcondition |
545 |
#define prev_is_word frame->Xprev_is_word |
#define prev_is_word frame->Xprev_is_word |
546 |
|
|
|
#define original_ims frame->Xoriginal_ims |
|
|
|
|
547 |
#ifdef SUPPORT_UCP |
#ifdef SUPPORT_UCP |
548 |
#define prop_type frame->Xprop_type |
#define prop_type frame->Xprop_type |
549 |
#define prop_value frame->Xprop_value |
#define prop_value frame->Xprop_value |
550 |
#define prop_fail_result frame->Xprop_fail_result |
#define prop_fail_result frame->Xprop_fail_result |
|
#define prop_category frame->Xprop_category |
|
|
#define prop_chartype frame->Xprop_chartype |
|
|
#define prop_script frame->Xprop_script |
|
551 |
#define oclength frame->Xoclength |
#define oclength frame->Xoclength |
552 |
#define occhars frame->Xocchars |
#define occhars frame->Xocchars |
553 |
#endif |
#endif |
577 |
#define fi i |
#define fi i |
578 |
#define fc c |
#define fc c |
579 |
|
|
580 |
|
/* Many of the following variables are used only in small blocks of the code. |
581 |
|
My normal style of coding would have declared them within each of those blocks. |
582 |
|
However, in order to accommodate the version of this code that uses an external |
583 |
|
"stack" implemented on the heap, it is easier to declare them all here, so the |
584 |
|
declarations can be cut out in a block. The only declarations within blocks |
585 |
|
below are for variables that do not have to be preserved over a recursive call |
586 |
|
to RMATCH(). */ |
587 |
|
|
588 |
|
#ifdef SUPPORT_UTF8 |
589 |
|
const uschar *charptr; |
590 |
|
#endif |
591 |
|
const uschar *callpat; |
592 |
|
const uschar *data; |
593 |
|
const uschar *next; |
594 |
|
USPTR pp; |
595 |
|
const uschar *prev; |
596 |
|
USPTR saved_eptr; |
597 |
|
|
598 |
#ifdef SUPPORT_UTF8 /* Many of these variables are used only */ |
recursion_info new_recursive; |
599 |
const uschar *charptr; /* in small blocks of the code. My normal */ |
|
600 |
#endif /* style of coding would have declared */ |
BOOL cur_is_word; |
|
const uschar *callpat; /* them within each of those blocks. */ |
|
|
const uschar *data; /* However, in order to accommodate the */ |
|
|
const uschar *next; /* version of this code that uses an */ |
|
|
USPTR pp; /* external "stack" implemented on the */ |
|
|
const uschar *prev; /* heap, it is easier to declare them all */ |
|
|
USPTR saved_eptr; /* here, so the declarations can be cut */ |
|
|
/* out in a block. The only declarations */ |
|
|
recursion_info new_recursive; /* within blocks below are for variables */ |
|
|
/* that do not have to be preserved over */ |
|
|
BOOL cur_is_word; /* a recursive call to RMATCH(). */ |
|
601 |
BOOL condition; |
BOOL condition; |
602 |
BOOL prev_is_word; |
BOOL prev_is_word; |
603 |
|
|
|
unsigned long int original_ims; |
|
|
|
|
604 |
#ifdef SUPPORT_UCP |
#ifdef SUPPORT_UCP |
605 |
int prop_type; |
int prop_type; |
606 |
int prop_value; |
int prop_value; |
607 |
int prop_fail_result; |
int prop_fail_result; |
|
int prop_category; |
|
|
int prop_chartype; |
|
|
int prop_script; |
|
608 |
int oclength; |
int oclength; |
609 |
uschar occhars[8]; |
uschar occhars[8]; |
610 |
#endif |
#endif |
624 |
eptrblock newptrb; |
eptrblock newptrb; |
625 |
#endif /* NO_RECURSE */ |
#endif /* NO_RECURSE */ |
626 |
|
|
627 |
|
/* To save space on the stack and in the heap frame, I have doubled up on some |
628 |
|
of the local variables that are used only in localised parts of the code, but |
629 |
|
still need to be preserved over recursive calls of match(). These macros define |
630 |
|
the alternative names that are used. */ |
631 |
|
|
632 |
|
#define allow_zero cur_is_word |
633 |
|
#define cbegroup condition |
634 |
|
#define code_offset codelink |
635 |
|
#define condassert condition |
636 |
|
#define matched_once prev_is_word |
637 |
|
|
638 |
/* These statements are here to stop the compiler complaining about unitialized |
/* These statements are here to stop the compiler complaining about unitialized |
639 |
variables. */ |
variables. */ |
640 |
|
|
671 |
if (md->match_call_count++ >= md->match_limit) RRETURN(PCRE_ERROR_MATCHLIMIT); |
if (md->match_call_count++ >= md->match_limit) RRETURN(PCRE_ERROR_MATCHLIMIT); |
672 |
if (rdepth >= md->match_limit_recursion) RRETURN(PCRE_ERROR_RECURSIONLIMIT); |
if (rdepth >= md->match_limit_recursion) RRETURN(PCRE_ERROR_RECURSIONLIMIT); |
673 |
|
|
|
original_ims = ims; /* Save for resetting on ')' */ |
|
|
|
|
674 |
/* At the start of a group with an unlimited repeat that may match an empty |
/* At the start of a group with an unlimited repeat that may match an empty |
675 |
string, the match_cbegroup flag is set. When this is the case, add the current |
string, the variable md->match_function_type is set to MATCH_CBEGROUP. It is |
676 |
subject pointer to the chain of such remembered pointers, to be checked when we |
done this way to save having to use another function argument, which would take |
677 |
hit the closing ket, in order to break infinite loops that match no characters. |
up space on the stack. See also MATCH_CONDASSERT below. |
678 |
When match() is called in other circumstances, don't add to the chain. The |
|
679 |
match_cbegroup flag must NOT be used with tail recursion, because the memory |
When MATCH_CBEGROUP is set, add the current subject pointer to the chain of |
680 |
block that is used is on the stack, so a new one may be required for each |
such remembered pointers, to be checked when we hit the closing ket, in order |
681 |
match(). */ |
to break infinite loops that match no characters. When match() is called in |
682 |
|
other circumstances, don't add to the chain. The MATCH_CBEGROUP feature must |
683 |
|
NOT be used with tail recursion, because the memory block that is used is on |
684 |
|
the stack, so a new one may be required for each match(). */ |
685 |
|
|
686 |
if ((flags & match_cbegroup) != 0) |
if (md->match_function_type == MATCH_CBEGROUP) |
687 |
{ |
{ |
688 |
newptrb.epb_saved_eptr = eptr; |
newptrb.epb_saved_eptr = eptr; |
689 |
newptrb.epb_prev = eptrb; |
newptrb.epb_prev = eptrb; |
690 |
eptrb = &newptrb; |
eptrb = &newptrb; |
691 |
|
md->match_function_type = 0; |
692 |
} |
} |
693 |
|
|
694 |
/* Now start processing the opcodes. */ |
/* Now start processing the opcodes. */ |
703 |
case OP_MARK: |
case OP_MARK: |
704 |
markptr = ecode + 2; |
markptr = ecode + 2; |
705 |
RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode] + ecode[1], offset_top, md, |
RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode] + ecode[1], offset_top, md, |
706 |
ims, eptrb, flags, RM55); |
eptrb, RM55); |
707 |
|
|
708 |
/* A return of MATCH_SKIP_ARG means that matching failed at SKIP with an |
/* A return of MATCH_SKIP_ARG means that matching failed at SKIP with an |
709 |
argument, and we must check whether that argument matches this MARK's |
argument, and we must check whether that argument matches this MARK's |
729 |
|
|
730 |
case OP_COMMIT: |
case OP_COMMIT: |
731 |
RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md, |
RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md, |
732 |
ims, eptrb, flags, RM52); |
eptrb, RM52); |
733 |
if (rrc != MATCH_NOMATCH && rrc != MATCH_PRUNE && |
if (rrc != MATCH_NOMATCH && rrc != MATCH_PRUNE && |
734 |
rrc != MATCH_SKIP && rrc != MATCH_SKIP_ARG && |
rrc != MATCH_SKIP && rrc != MATCH_SKIP_ARG && |
735 |
rrc != MATCH_THEN) |
rrc != MATCH_THEN) |
740 |
|
|
741 |
case OP_PRUNE: |
case OP_PRUNE: |
742 |
RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md, |
RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md, |
743 |
ims, eptrb, flags, RM51); |
eptrb, RM51); |
744 |
if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN) RRETURN(rrc); |
if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN) RRETURN(rrc); |
745 |
MRRETURN(MATCH_PRUNE); |
MRRETURN(MATCH_PRUNE); |
746 |
|
|
747 |
case OP_PRUNE_ARG: |
case OP_PRUNE_ARG: |
748 |
RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode] + ecode[1], offset_top, md, |
RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode] + ecode[1], offset_top, md, |
749 |
ims, eptrb, flags, RM56); |
eptrb, RM56); |
750 |
if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN) RRETURN(rrc); |
if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN) RRETURN(rrc); |
751 |
md->mark = ecode + 2; |
md->mark = ecode + 2; |
752 |
RRETURN(MATCH_PRUNE); |
RRETURN(MATCH_PRUNE); |
755 |
|
|
756 |
case OP_SKIP: |
case OP_SKIP: |
757 |
RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md, |
RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md, |
758 |
ims, eptrb, flags, RM53); |
eptrb, RM53); |
759 |
if (rrc != MATCH_NOMATCH && rrc != MATCH_PRUNE && rrc != MATCH_THEN) |
if (rrc != MATCH_NOMATCH && rrc != MATCH_PRUNE && rrc != MATCH_THEN) |
760 |
RRETURN(rrc); |
RRETURN(rrc); |
761 |
md->start_match_ptr = eptr; /* Pass back current position */ |
md->start_match_ptr = eptr; /* Pass back current position */ |
763 |
|
|
764 |
case OP_SKIP_ARG: |
case OP_SKIP_ARG: |
765 |
RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode] + ecode[1], offset_top, md, |
RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode] + ecode[1], offset_top, md, |
766 |
ims, eptrb, flags, RM57); |
eptrb, RM57); |
767 |
if (rrc != MATCH_NOMATCH && rrc != MATCH_PRUNE && rrc != MATCH_THEN) |
if (rrc != MATCH_NOMATCH && rrc != MATCH_PRUNE && rrc != MATCH_THEN) |
768 |
RRETURN(rrc); |
RRETURN(rrc); |
769 |
|
|
782 |
|
|
783 |
case OP_THEN: |
case OP_THEN: |
784 |
RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md, |
RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md, |
785 |
ims, eptrb, flags, RM54); |
eptrb, RM54); |
786 |
if (rrc != MATCH_NOMATCH) RRETURN(rrc); |
if (rrc != MATCH_NOMATCH) RRETURN(rrc); |
787 |
md->start_match_ptr = ecode - GET(ecode, 1); |
md->start_match_ptr = ecode - GET(ecode, 1); |
788 |
MRRETURN(MATCH_THEN); |
MRRETURN(MATCH_THEN); |
789 |
|
|
790 |
case OP_THEN_ARG: |
case OP_THEN_ARG: |
791 |
RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode] + ecode[1+LINK_SIZE], |
RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode] + ecode[1+LINK_SIZE], |
792 |
offset_top, md, ims, eptrb, flags, RM58); |
offset_top, md, eptrb, RM58); |
793 |
if (rrc != MATCH_NOMATCH) RRETURN(rrc); |
if (rrc != MATCH_NOMATCH) RRETURN(rrc); |
794 |
md->start_match_ptr = ecode - GET(ecode, 1); |
md->start_match_ptr = ecode - GET(ecode, 1); |
795 |
md->mark = ecode + LINK_SIZE + 2; |
md->mark = ecode + LINK_SIZE + 2; |
796 |
RRETURN(MATCH_THEN); |
RRETURN(MATCH_THEN); |
797 |
|
|
798 |
/* Handle a capturing bracket. If there is space in the offset vector, save |
/* Handle a capturing bracket, other than those that are possessive with an |
799 |
the current subject position in the working slot at the top of the vector. |
unlimited repeat. If there is space in the offset vector, save the current |
800 |
We mustn't change the current values of the data slot, because they may be |
subject position in the working slot at the top of the vector. We mustn't |
801 |
set from a previous iteration of this group, and be referred to by a |
change the current values of the data slot, because they may be set from a |
802 |
reference inside the group. |
previous iteration of this group, and be referred to by a reference inside |
803 |
|
the group. A failure to match might occur after the group has succeeded, |
804 |
If the bracket fails to match, we need to restore this value and also the |
if something later on doesn't match. For this reason, we need to restore |
805 |
values of the final offsets, in case they were set by a previous iteration |
the working value and also the values of the final offsets, in case they |
806 |
of the same bracket. |
were set by a previous iteration of the same bracket. |
807 |
|
|
808 |
If there isn't enough space in the offset vector, treat this as if it were |
If there isn't enough space in the offset vector, treat this as if it were |
809 |
a non-capturing bracket. Don't worry about setting the flag for the error |
a non-capturing bracket. Don't worry about setting the flag for the error |
832 |
md->offset_vector[md->offset_end - number] = |
md->offset_vector[md->offset_end - number] = |
833 |
(int)(eptr - md->start_subject); |
(int)(eptr - md->start_subject); |
834 |
|
|
835 |
flags = (op == OP_SCBRA)? match_cbegroup : 0; |
for (;;) |
|
do |
|
836 |
{ |
{ |
837 |
|
if (op >= OP_SBRA) md->match_function_type = MATCH_CBEGROUP; |
838 |
RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md, |
RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md, |
839 |
ims, eptrb, flags, RM1); |
eptrb, RM1); |
840 |
|
if (rrc == MATCH_ONCE) break; /* Backing up through an atomic group */ |
841 |
if (rrc != MATCH_NOMATCH && |
if (rrc != MATCH_NOMATCH && |
842 |
(rrc != MATCH_THEN || md->start_match_ptr != ecode)) |
(rrc != MATCH_THEN || md->start_match_ptr != ecode)) |
843 |
RRETURN(rrc); |
RRETURN(rrc); |
844 |
md->capture_last = save_capture_last; |
md->capture_last = save_capture_last; |
845 |
ecode += GET(ecode, 1); |
ecode += GET(ecode, 1); |
846 |
|
if (*ecode != OP_ALT) break; |
847 |
} |
} |
|
while (*ecode == OP_ALT); |
|
848 |
|
|
849 |
DPRINTF(("bracket %d failed\n", number)); |
DPRINTF(("bracket %d failed\n", number)); |
|
|
|
850 |
md->offset_vector[offset] = save_offset1; |
md->offset_vector[offset] = save_offset1; |
851 |
md->offset_vector[offset+1] = save_offset2; |
md->offset_vector[offset+1] = save_offset2; |
852 |
md->offset_vector[md->offset_end - number] = save_offset3; |
md->offset_vector[md->offset_end - number] = save_offset3; |
853 |
|
|
854 |
if (rrc != MATCH_THEN) md->mark = markptr; |
/* At this point, rrc will be one of MATCH_ONCE, MATCH_NOMATCH, or |
855 |
RRETURN(MATCH_NOMATCH); |
MATCH_THEN. */ |
856 |
|
|
857 |
|
if (rrc != MATCH_THEN && md->mark == NULL) md->mark = markptr; |
858 |
|
RRETURN(((rrc == MATCH_ONCE)? MATCH_ONCE:MATCH_NOMATCH)); |
859 |
} |
} |
860 |
|
|
861 |
/* FALL THROUGH ... Insufficient room for saving captured contents. Treat |
/* FALL THROUGH ... Insufficient room for saving captured contents. Treat |
869 |
/* VVVVVVVVVVVVVVVVVVVVVVVVV */ |
/* VVVVVVVVVVVVVVVVVVVVVVVVV */ |
870 |
/* VVVVVVVVVVVVVVVVVVVVVVVVV */ |
/* VVVVVVVVVVVVVVVVVVVVVVVVV */ |
871 |
|
|
872 |
/* Non-capturing bracket. Loop for all the alternatives. When we get to the |
/* Non-capturing or atomic group, except for possessive with unlimited |
873 |
final alternative within the brackets, we would return the result of a |
repeat. Loop for all the alternatives. When we get to the final alternative |
874 |
recursive call to match() whatever happened. We can reduce stack usage by |
within the brackets, we used to return the result of a recursive call to |
875 |
turning this into a tail recursion, except in the case when match_cbegroup |
match() whatever happened so it was possible to reduce stack usage by |
876 |
is set.*/ |
turning this into a tail recursion, except in the case of a possibly empty |
877 |
|
group. However, now that there is the possiblity of (*THEN) occurring in |
878 |
|
the final alternative, this optimization is no longer possible. |
879 |
|
|
880 |
|
MATCH_ONCE is returned when the end of an atomic group is successfully |
881 |
|
reached, but subsequent matching fails. It passes back up the tree (causing |
882 |
|
captured values to be reset) until the original atomic group level is |
883 |
|
reached. This is tested by comparing md->once_target with the start of the |
884 |
|
group. At this point, the return is converted into MATCH_NOMATCH so that |
885 |
|
previous backup points can be taken. */ |
886 |
|
|
887 |
|
case OP_ONCE: |
888 |
case OP_BRA: |
case OP_BRA: |
889 |
case OP_SBRA: |
case OP_SBRA: |
890 |
DPRINTF(("start non-capturing bracket\n")); |
DPRINTF(("start non-capturing bracket\n")); |
891 |
flags = (op >= OP_SBRA)? match_cbegroup : 0; |
|
892 |
for (;;) |
for (;;) |
893 |
{ |
{ |
894 |
if (ecode[GET(ecode, 1)] != OP_ALT) /* Final alternative */ |
if (op >= OP_SBRA || op == OP_ONCE) md->match_function_type = MATCH_CBEGROUP; |
895 |
|
RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md, eptrb, |
896 |
|
RM2); |
897 |
|
if (rrc != MATCH_NOMATCH && |
898 |
|
(rrc != MATCH_THEN || md->start_match_ptr != ecode)) |
899 |
{ |
{ |
900 |
if (flags == 0) /* Not a possibly empty group */ |
if (rrc == MATCH_ONCE) |
901 |
{ |
{ |
902 |
ecode += _pcre_OP_lengths[*ecode]; |
const uschar *scode = ecode; |
903 |
DPRINTF(("bracket 0 tail recursion\n")); |
if (*scode != OP_ONCE) /* If not at start, find it */ |
904 |
goto TAIL_RECURSE; |
{ |
905 |
|
while (*scode == OP_ALT) scode += GET(scode, 1); |
906 |
|
scode -= GET(scode, 1); |
907 |
|
} |
908 |
|
if (md->once_target == scode) rrc = MATCH_NOMATCH; |
909 |
} |
} |
910 |
|
RRETURN(rrc); |
911 |
|
} |
912 |
|
ecode += GET(ecode, 1); |
913 |
|
if (*ecode != OP_ALT) break; |
914 |
|
} |
915 |
|
if (rrc != MATCH_THEN && md->mark == NULL) md->mark = markptr; |
916 |
|
RRETURN(MATCH_NOMATCH); |
917 |
|
|
918 |
/* Possibly empty group; can't use tail recursion. */ |
/* Handle possessive capturing brackets with an unlimited repeat. We come |
919 |
|
here from BRAZERO with allow_zero set TRUE. The offset_vector values are |
920 |
|
handled similarly to the normal case above. However, the matching is |
921 |
|
different. The end of these brackets will always be OP_KETRPOS, which |
922 |
|
returns MATCH_KETRPOS without going further in the pattern. By this means |
923 |
|
we can handle the group by iteration rather than recursion, thereby |
924 |
|
reducing the amount of stack needed. */ |
925 |
|
|
926 |
|
case OP_CBRAPOS: |
927 |
|
case OP_SCBRAPOS: |
928 |
|
allow_zero = FALSE; |
929 |
|
|
930 |
RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md, ims, |
POSSESSIVE_CAPTURE: |
931 |
eptrb, flags, RM48); |
number = GET2(ecode, 1+LINK_SIZE); |
932 |
if (rrc == MATCH_NOMATCH) md->mark = markptr; |
offset = number << 1; |
933 |
RRETURN(rrc); |
|
934 |
|
#ifdef PCRE_DEBUG |
935 |
|
printf("start possessive bracket %d\n", number); |
936 |
|
printf("subject="); |
937 |
|
pchars(eptr, 16, TRUE, md); |
938 |
|
printf("\n"); |
939 |
|
#endif |
940 |
|
|
941 |
|
if (offset < md->offset_max) |
942 |
|
{ |
943 |
|
matched_once = FALSE; |
944 |
|
code_offset = ecode - md->start_code; |
945 |
|
|
946 |
|
save_offset1 = md->offset_vector[offset]; |
947 |
|
save_offset2 = md->offset_vector[offset+1]; |
948 |
|
save_offset3 = md->offset_vector[md->offset_end - number]; |
949 |
|
save_capture_last = md->capture_last; |
950 |
|
|
951 |
|
DPRINTF(("saving %d %d %d\n", save_offset1, save_offset2, save_offset3)); |
952 |
|
|
953 |
|
/* Each time round the loop, save the current subject position for use |
954 |
|
when the group matches. For MATCH_MATCH, the group has matched, so we |
955 |
|
restart it with a new subject starting position, remembering that we had |
956 |
|
at least one match. For MATCH_NOMATCH, carry on with the alternatives, as |
957 |
|
usual. If we haven't matched any alternatives in any iteration, check to |
958 |
|
see if a previous iteration matched. If so, the group has matched; |
959 |
|
continue from afterwards. Otherwise it has failed; restore the previous |
960 |
|
capture values before returning NOMATCH. */ |
961 |
|
|
962 |
|
for (;;) |
963 |
|
{ |
964 |
|
md->offset_vector[md->offset_end - number] = |
965 |
|
(int)(eptr - md->start_subject); |
966 |
|
if (op >= OP_SBRA) md->match_function_type = MATCH_CBEGROUP; |
967 |
|
RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md, |
968 |
|
eptrb, RM63); |
969 |
|
if (rrc == MATCH_KETRPOS) |
970 |
|
{ |
971 |
|
offset_top = md->end_offset_top; |
972 |
|
eptr = md->end_match_ptr; |
973 |
|
ecode = md->start_code + code_offset; |
974 |
|
save_capture_last = md->capture_last; |
975 |
|
matched_once = TRUE; |
976 |
|
continue; |
977 |
|
} |
978 |
|
if (rrc != MATCH_NOMATCH && |
979 |
|
(rrc != MATCH_THEN || md->start_match_ptr != ecode)) |
980 |
|
RRETURN(rrc); |
981 |
|
md->capture_last = save_capture_last; |
982 |
|
ecode += GET(ecode, 1); |
983 |
|
if (*ecode != OP_ALT) break; |
984 |
} |
} |
985 |
|
|
986 |
/* For non-final alternatives, continue the loop for a NOMATCH result; |
if (!matched_once) |
987 |
otherwise return. */ |
{ |
988 |
|
md->offset_vector[offset] = save_offset1; |
989 |
|
md->offset_vector[offset+1] = save_offset2; |
990 |
|
md->offset_vector[md->offset_end - number] = save_offset3; |
991 |
|
} |
992 |
|
|
993 |
|
if (rrc != MATCH_THEN && md->mark == NULL) md->mark = markptr; |
994 |
|
if (allow_zero || matched_once) |
995 |
|
{ |
996 |
|
ecode += 1 + LINK_SIZE; |
997 |
|
break; |
998 |
|
} |
999 |
|
|
1000 |
RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md, ims, |
RRETURN(MATCH_NOMATCH); |
1001 |
eptrb, flags, RM2); |
} |
1002 |
|
|
1003 |
|
/* FALL THROUGH ... Insufficient room for saving captured contents. Treat |
1004 |
|
as a non-capturing bracket. */ |
1005 |
|
|
1006 |
|
/* VVVVVVVVVVVVVVVVVVVVVVVVV */ |
1007 |
|
/* VVVVVVVVVVVVVVVVVVVVVVVVV */ |
1008 |
|
|
1009 |
|
DPRINTF(("insufficient capture room: treat as non-capturing\n")); |
1010 |
|
|
1011 |
|
/* VVVVVVVVVVVVVVVVVVVVVVVVV */ |
1012 |
|
/* VVVVVVVVVVVVVVVVVVVVVVVVV */ |
1013 |
|
|
1014 |
|
/* Non-capturing possessive bracket with unlimited repeat. We come here |
1015 |
|
from BRAZERO with allow_zero = TRUE. The code is similar to the above, |
1016 |
|
without the capturing complication. It is written out separately for speed |
1017 |
|
and cleanliness. */ |
1018 |
|
|
1019 |
|
case OP_BRAPOS: |
1020 |
|
case OP_SBRAPOS: |
1021 |
|
allow_zero = FALSE; |
1022 |
|
|
1023 |
|
POSSESSIVE_NON_CAPTURE: |
1024 |
|
matched_once = FALSE; |
1025 |
|
code_offset = ecode - md->start_code; |
1026 |
|
|
1027 |
|
for (;;) |
1028 |
|
{ |
1029 |
|
if (op >= OP_SBRA) md->match_function_type = MATCH_CBEGROUP; |
1030 |
|
RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md, |
1031 |
|
eptrb, RM48); |
1032 |
|
if (rrc == MATCH_KETRPOS) |
1033 |
|
{ |
1034 |
|
offset_top = md->end_offset_top; |
1035 |
|
eptr = md->end_match_ptr; |
1036 |
|
ecode = md->start_code + code_offset; |
1037 |
|
matched_once = TRUE; |
1038 |
|
continue; |
1039 |
|
} |
1040 |
if (rrc != MATCH_NOMATCH && |
if (rrc != MATCH_NOMATCH && |
1041 |
(rrc != MATCH_THEN || md->start_match_ptr != ecode)) |
(rrc != MATCH_THEN || md->start_match_ptr != ecode)) |
1042 |
RRETURN(rrc); |
RRETURN(rrc); |
1043 |
ecode += GET(ecode, 1); |
ecode += GET(ecode, 1); |
1044 |
|
if (*ecode != OP_ALT) break; |
1045 |
} |
} |
1046 |
|
|
1047 |
|
if (matched_once || allow_zero) |
1048 |
|
{ |
1049 |
|
ecode += 1 + LINK_SIZE; |
1050 |
|
break; |
1051 |
|
} |
1052 |
|
RRETURN(MATCH_NOMATCH); |
1053 |
|
|
1054 |
/* Control never reaches here. */ |
/* Control never reaches here. */ |
1055 |
|
|
1056 |
/* Conditional group: compilation checked that there are no more than |
/* Conditional group: compilation checked that there are no more than |
1057 |
two branches. If the condition is false, skipping the first branch takes us |
two branches. If the condition is false, skipping the first branch takes us |
1058 |
past the end if there is only one branch, but that's OK because that is |
past the end if there is only one branch, but that's OK because that is |
1059 |
exactly what going to the ket would do. As there is only one branch to be |
exactly what going to the ket would do. */ |
|
obeyed, we can use tail recursion to avoid using another stack frame. */ |
|
1060 |
|
|
1061 |
case OP_COND: |
case OP_COND: |
1062 |
case OP_SCOND: |
case OP_SCOND: |
1063 |
codelink= GET(ecode, 1); |
codelink = GET(ecode, 1); |
1064 |
|
|
1065 |
/* Because of the way auto-callout works during compile, a callout item is |
/* Because of the way auto-callout works during compile, a callout item is |
1066 |
inserted between OP_COND and an assertion condition. */ |
inserted between OP_COND and an assertion condition. */ |
1235 |
} |
} |
1236 |
|
|
1237 |
/* The condition is an assertion. Call match() to evaluate it - setting |
/* The condition is an assertion. Call match() to evaluate it - setting |
1238 |
the final argument match_condassert causes it to stop at the end of an |
md->match_function_type to MATCH_CONDASSERT causes it to stop at the end of |
1239 |
assertion. */ |
an assertion. */ |
1240 |
|
|
1241 |
else |
else |
1242 |
{ |
{ |
1243 |
RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, ims, NULL, |
md->match_function_type = MATCH_CONDASSERT; |
1244 |
match_condassert, RM3); |
RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, NULL, RM3); |
1245 |
if (rrc == MATCH_MATCH) |
if (rrc == MATCH_MATCH) |
1246 |
{ |
{ |
1247 |
|
if (md->end_offset_top > offset_top) |
1248 |
|
offset_top = md->end_offset_top; /* Captures may have happened */ |
1249 |
condition = TRUE; |
condition = TRUE; |
1250 |
ecode += 1 + LINK_SIZE + GET(ecode, LINK_SIZE + 2); |
ecode += 1 + LINK_SIZE + GET(ecode, LINK_SIZE + 2); |
1251 |
while (*ecode == OP_ALT) ecode += GET(ecode, 1); |
while (*ecode == OP_ALT) ecode += GET(ecode, 1); |
1263 |
} |
} |
1264 |
|
|
1265 |
/* We are now at the branch that is to be obeyed. As there is only one, |
/* We are now at the branch that is to be obeyed. As there is only one, |
1266 |
we can use tail recursion to avoid using another stack frame, except when |
we used to use tail recursion to avoid using another stack frame, except |
1267 |
match_cbegroup is required for an unlimited repeat of a possibly empty |
when there was unlimited repeat of a possibly empty group. However, that |
1268 |
group. If the second alternative doesn't exist, we can just plough on. */ |
strategy no longer works because of the possibilty of (*THEN) being |
1269 |
|
encountered in the branch. A recursive call to match() is always required, |
1270 |
|
unless the second alternative doesn't exist, in which case we can just |
1271 |
|
plough on. */ |
1272 |
|
|
1273 |
if (condition || *ecode == OP_ALT) |
if (condition || *ecode == OP_ALT) |
1274 |
{ |
{ |
1275 |
ecode += 1 + LINK_SIZE; |
if (op == OP_SCOND) md->match_function_type = MATCH_CBEGROUP; |
1276 |
if (op == OP_SCOND) /* Possibly empty group */ |
RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, eptrb, RM49); |
1277 |
{ |
if (rrc == MATCH_THEN && md->start_match_ptr == ecode) |
1278 |
RMATCH(eptr, ecode, offset_top, md, ims, eptrb, match_cbegroup, RM49); |
rrc = MATCH_NOMATCH; |
1279 |
RRETURN(rrc); |
RRETURN(rrc); |
|
} |
|
|
else /* Group must match something */ |
|
|
{ |
|
|
flags = 0; |
|
|
goto TAIL_RECURSE; |
|
|
} |
|
1280 |
} |
} |
1281 |
else /* Condition false & no alternative */ |
else /* Condition false & no alternative */ |
1282 |
{ |
{ |
1309 |
break; |
break; |
1310 |
|
|
1311 |
|
|
1312 |
/* End of the pattern, either real or forced. If we are in a top-level |
/* End of the pattern, either real or forced. */ |
|
recursion, we should restore the offsets appropriately and continue from |
|
|
after the call. */ |
|
1313 |
|
|
|
case OP_ACCEPT: |
|
1314 |
case OP_END: |
case OP_END: |
1315 |
if (md->recursive != NULL && md->recursive->group_num == 0) |
case OP_ACCEPT: |
1316 |
{ |
case OP_ASSERT_ACCEPT: |
|
recursion_info *rec = md->recursive; |
|
|
DPRINTF(("End of pattern in a (?0) recursion\n")); |
|
|
md->recursive = rec->prevrec; |
|
|
memmove(md->offset_vector, rec->offset_save, |
|
|
rec->saved_max * sizeof(int)); |
|
|
offset_top = rec->save_offset_top; |
|
|
ims = original_ims; |
|
|
ecode = rec->after_call; |
|
|
break; |
|
|
} |
|
1317 |
|
|
1318 |
/* Otherwise, if we have matched an empty string, fail if PCRE_NOTEMPTY is |
/* If we have matched an empty string, fail if not in an assertion and not |
1319 |
set, or if PCRE_NOTEMPTY_ATSTART is set and we have matched at the start of |
in a recursion if either PCRE_NOTEMPTY is set, or if PCRE_NOTEMPTY_ATSTART |
1320 |
the subject. In both cases, backtracking will then try other alternatives, |
is set and we have matched at the start of the subject. In both cases, |
1321 |
if any. */ |
backtracking will then try other alternatives, if any. */ |
1322 |
|
|
1323 |
if (eptr == mstart && |
if (eptr == mstart && op != OP_ASSERT_ACCEPT && |
1324 |
(md->notempty || |
md->recursive == NULL && |
1325 |
(md->notempty_atstart && |
(md->notempty || |
1326 |
mstart == md->start_subject + md->start_offset))) |
(md->notempty_atstart && |
1327 |
|
mstart == md->start_subject + md->start_offset))) |
1328 |
MRRETURN(MATCH_NOMATCH); |
MRRETURN(MATCH_NOMATCH); |
1329 |
|
|
1330 |
/* Otherwise, we have a match. */ |
/* Otherwise, we have a match. */ |
1339 |
rrc = (op == OP_END)? MATCH_MATCH : MATCH_ACCEPT; |
rrc = (op == OP_END)? MATCH_MATCH : MATCH_ACCEPT; |
1340 |
MRRETURN(rrc); |
MRRETURN(rrc); |
1341 |
|
|
|
/* Change option settings */ |
|
|
|
|
|
case OP_OPT: |
|
|
ims = ecode[1]; |
|
|
ecode += 2; |
|
|
DPRINTF(("ims set to %02lx\n", ims)); |
|
|
break; |
|
|
|
|
1342 |
/* Assertion brackets. Check the alternative branches in turn - the |
/* Assertion brackets. Check the alternative branches in turn - the |
1343 |
matching won't pass the KET for an assertion. If any one branch matches, |
matching won't pass the KET for an assertion. If any one branch matches, |
1344 |
the assertion is true. Lookbehind assertions have an OP_REVERSE item at the |
the assertion is true. Lookbehind assertions have an OP_REVERSE item at the |
1345 |
start of each branch to move the current point backwards, so the code at |
start of each branch to move the current point backwards, so the code at |
1346 |
this level is identical to the lookahead case. */ |
this level is identical to the lookahead case. When the assertion is part |
1347 |
|
of a condition, we want to return immediately afterwards. The caller of |
1348 |
|
this incarnation of the match() function will have set MATCH_CONDASSERT in |
1349 |
|
md->match_function type, and one of these opcodes will be the first opcode |
1350 |
|
that is processed. We use a local variable that is preserved over calls to |
1351 |
|
match() to remember this case. */ |
1352 |
|
|
1353 |
case OP_ASSERT: |
case OP_ASSERT: |
1354 |
case OP_ASSERTBACK: |
case OP_ASSERTBACK: |
1355 |
|
if (md->match_function_type == MATCH_CONDASSERT) |
1356 |
|
{ |
1357 |
|
condassert = TRUE; |
1358 |
|
md->match_function_type = 0; |
1359 |
|
} |
1360 |
|
else condassert = FALSE; |
1361 |
|
|
1362 |
do |
do |
1363 |
{ |
{ |
1364 |
RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, ims, NULL, 0, |
RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, NULL, RM4); |
|
RM4); |
|
1365 |
if (rrc == MATCH_MATCH || rrc == MATCH_ACCEPT) |
if (rrc == MATCH_MATCH || rrc == MATCH_ACCEPT) |
1366 |
{ |
{ |
1367 |
mstart = md->start_match_ptr; /* In case \K reset it */ |
mstart = md->start_match_ptr; /* In case \K reset it */ |
1368 |
|
markptr = md->mark; |
1369 |
break; |
break; |
1370 |
} |
} |
1371 |
if (rrc != MATCH_NOMATCH && |
if (rrc != MATCH_NOMATCH && |
1374 |
ecode += GET(ecode, 1); |
ecode += GET(ecode, 1); |
1375 |
} |
} |
1376 |
while (*ecode == OP_ALT); |
while (*ecode == OP_ALT); |
1377 |
|
|
1378 |
if (*ecode == OP_KET) MRRETURN(MATCH_NOMATCH); |
if (*ecode == OP_KET) MRRETURN(MATCH_NOMATCH); |
1379 |
|
|
1380 |
/* If checking an assertion for a condition, return MATCH_MATCH. */ |
/* If checking an assertion for a condition, return MATCH_MATCH. */ |
1381 |
|
|
1382 |
if ((flags & match_condassert) != 0) RRETURN(MATCH_MATCH); |
if (condassert) RRETURN(MATCH_MATCH); |
1383 |
|
|
1384 |
/* Continue from after the assertion, updating the offsets high water |
/* Continue from after the assertion, updating the offsets high water |
1385 |
mark, since extracts may have been taken during the assertion. */ |
mark, since extracts may have been taken during the assertion. */ |
1395 |
|
|
1396 |
case OP_ASSERT_NOT: |
case OP_ASSERT_NOT: |
1397 |
case OP_ASSERTBACK_NOT: |
case OP_ASSERTBACK_NOT: |
1398 |
|
if (md->match_function_type == MATCH_CONDASSERT) |
1399 |
|
{ |
1400 |
|
condassert = TRUE; |
1401 |
|
md->match_function_type = 0; |
1402 |
|
} |
1403 |
|
else condassert = FALSE; |
1404 |
|
|
1405 |
do |
do |
1406 |
{ |
{ |
1407 |
RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, ims, NULL, 0, |
RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, NULL, RM5); |
|
RM5); |
|
1408 |
if (rrc == MATCH_MATCH || rrc == MATCH_ACCEPT) MRRETURN(MATCH_NOMATCH); |
if (rrc == MATCH_MATCH || rrc == MATCH_ACCEPT) MRRETURN(MATCH_NOMATCH); |
1409 |
if (rrc == MATCH_SKIP || rrc == MATCH_PRUNE || rrc == MATCH_COMMIT) |
if (rrc == MATCH_SKIP || rrc == MATCH_PRUNE || rrc == MATCH_COMMIT) |
1410 |
{ |
{ |
1418 |
} |
} |
1419 |
while (*ecode == OP_ALT); |
while (*ecode == OP_ALT); |
1420 |
|
|
1421 |
if ((flags & match_condassert) != 0) RRETURN(MATCH_MATCH); |
if (condassert) RRETURN(MATCH_MATCH); /* Condition assertion */ |
1422 |
|
|
1423 |
ecode += 1 + LINK_SIZE; |
ecode += 1 + LINK_SIZE; |
1424 |
continue; |
continue; |
1486 |
offset data is the offset to the starting bracket from the start of the |
offset data is the offset to the starting bracket from the start of the |
1487 |
whole pattern. (This is so that it works from duplicated subpatterns.) |
whole pattern. (This is so that it works from duplicated subpatterns.) |
1488 |
|
|
1489 |
If there are any capturing brackets started but not finished, we have to |
The state of the capturing groups is preserved over recursion, and |
1490 |
save their starting points and reinstate them after the recursion. However, |
re-instated afterwards. We don't know how many are started and not yet |
1491 |
we don't know how many such there are (offset_top records the completed |
finished (offset_top records the completed total) so we just have to save |
1492 |
total) so we just have to save all the potential data. There may be up to |
all the potential data. There may be up to 65535 such values, which is too |
1493 |
65535 such values, which is too large to put on the stack, but using malloc |
large to put on the stack, but using malloc for small numbers seems |
1494 |
for small numbers seems expensive. As a compromise, the stack is used when |
expensive. As a compromise, the stack is used when there are no more than |
1495 |
there are no more than REC_STACK_SAVE_MAX values to store; otherwise malloc |
REC_STACK_SAVE_MAX values to store; otherwise malloc is used. |
|
is used. A problem is what to do if the malloc fails ... there is no way of |
|
|
returning to the top level with an error. Save the top REC_STACK_SAVE_MAX |
|
|
values on the stack, and accept that the rest may be wrong. |
|
1496 |
|
|
1497 |
There are also other values that have to be saved. We use a chained |
There are also other values that have to be saved. We use a chained |
1498 |
sequence of blocks that actually live on the stack. Thanks to Robin Houston |
sequence of blocks that actually live on the stack. Thanks to Robin Houston |
1499 |
for the original version of this logic. */ |
for the original version of this logic. It has, however, been hacked around |
1500 |
|
a lot, so he is not to blame for the current way it works. */ |
1501 |
|
|
1502 |
case OP_RECURSE: |
case OP_RECURSE: |
1503 |
{ |
{ |
1504 |
|
recursion_info *ri; |
1505 |
|
int recno; |
1506 |
|
|
1507 |
callpat = md->start_code + GET(ecode, 1); |
callpat = md->start_code + GET(ecode, 1); |
1508 |
new_recursive.group_num = (callpat == md->start_code)? 0 : |
recno = (callpat == md->start_code)? 0 : |
1509 |
GET2(callpat, 1 + LINK_SIZE); |
GET2(callpat, 1 + LINK_SIZE); |
1510 |
|
|
1511 |
|
/* Check for repeating a recursion without advancing the subject pointer. |
1512 |
|
This should catch convoluted mutual recursions. (Some simple cases are |
1513 |
|
caught at compile time.) */ |
1514 |
|
|
1515 |
|
for (ri = md->recursive; ri != NULL; ri = ri->prevrec) |
1516 |
|
if (recno == ri->group_num && eptr == ri->subject_position) |
1517 |
|
RRETURN(PCRE_ERROR_RECURSELOOP); |
1518 |
|
|
1519 |
/* Add to "recursing stack" */ |
/* Add to "recursing stack" */ |
1520 |
|
|
1521 |
|
new_recursive.group_num = recno; |
1522 |
|
new_recursive.subject_position = eptr; |
1523 |
new_recursive.prevrec = md->recursive; |
new_recursive.prevrec = md->recursive; |
1524 |
md->recursive = &new_recursive; |
md->recursive = &new_recursive; |
1525 |
|
|
1526 |
/* Find where to continue from afterwards */ |
/* Where to continue from afterwards */ |
1527 |
|
|
1528 |
ecode += 1 + LINK_SIZE; |
ecode += 1 + LINK_SIZE; |
|
new_recursive.after_call = ecode; |
|
1529 |
|
|
1530 |
/* Now save the offset data. */ |
/* Now save the offset data */ |
1531 |
|
|
1532 |
new_recursive.saved_max = md->offset_end; |
new_recursive.saved_max = md->offset_end; |
1533 |
if (new_recursive.saved_max <= REC_STACK_SAVE_MAX) |
if (new_recursive.saved_max <= REC_STACK_SAVE_MAX) |
1538 |
(int *)(pcre_malloc)(new_recursive.saved_max * sizeof(int)); |
(int *)(pcre_malloc)(new_recursive.saved_max * sizeof(int)); |
1539 |
if (new_recursive.offset_save == NULL) RRETURN(PCRE_ERROR_NOMEMORY); |
if (new_recursive.offset_save == NULL) RRETURN(PCRE_ERROR_NOMEMORY); |
1540 |
} |
} |
|
|
|
1541 |
memcpy(new_recursive.offset_save, md->offset_vector, |
memcpy(new_recursive.offset_save, md->offset_vector, |
1542 |
new_recursive.saved_max * sizeof(int)); |
new_recursive.saved_max * sizeof(int)); |
|
new_recursive.save_offset_top = offset_top; |
|
1543 |
|
|
1544 |
/* OK, now we can do the recursion. For each top-level alternative we |
/* OK, now we can do the recursion. After processing each alternative, |
1545 |
restore the offset and recursion data. */ |
restore the offset data. If there were nested recursions, md->recursive |
1546 |
|
might be changed, so reset it before looping. */ |
1547 |
|
|
1548 |
DPRINTF(("Recursing into group %d\n", new_recursive.group_num)); |
DPRINTF(("Recursing into group %d\n", new_recursive.group_num)); |
1549 |
flags = (*callpat >= OP_SBRA)? match_cbegroup : 0; |
cbegroup = (*callpat >= OP_SBRA); |
1550 |
do |
do |
1551 |
{ |
{ |
1552 |
|
if (cbegroup) md->match_function_type = MATCH_CBEGROUP; |
1553 |
RMATCH(eptr, callpat + _pcre_OP_lengths[*callpat], offset_top, |
RMATCH(eptr, callpat + _pcre_OP_lengths[*callpat], offset_top, |
1554 |
md, ims, eptrb, flags, RM6); |
md, eptrb, RM6); |
1555 |
|
memcpy(md->offset_vector, new_recursive.offset_save, |
1556 |
|
new_recursive.saved_max * sizeof(int)); |
1557 |
if (rrc == MATCH_MATCH || rrc == MATCH_ACCEPT) |
if (rrc == MATCH_MATCH || rrc == MATCH_ACCEPT) |
1558 |
{ |
{ |
1559 |
DPRINTF(("Recursion matched\n")); |
DPRINTF(("Recursion matched\n")); |
1560 |
md->recursive = new_recursive.prevrec; |
md->recursive = new_recursive.prevrec; |
1561 |
if (new_recursive.offset_save != stacksave) |
if (new_recursive.offset_save != stacksave) |
1562 |
(pcre_free)(new_recursive.offset_save); |
(pcre_free)(new_recursive.offset_save); |
1563 |
MRRETURN(MATCH_MATCH); |
|
1564 |
|
/* Set where we got to in the subject, and reset the start in case |
1565 |
|
it was changed by \K. This *is* propagated back out of a recursion, |
1566 |
|
for Perl compatibility. */ |
1567 |
|
|
1568 |
|
eptr = md->end_match_ptr; |
1569 |
|
mstart = md->start_match_ptr; |
1570 |
|
goto RECURSION_MATCHED; /* Exit loop; end processing */ |
1571 |
} |
} |
1572 |
else if (rrc != MATCH_NOMATCH && |
else if (rrc != MATCH_NOMATCH && |
1573 |
(rrc != MATCH_THEN || md->start_match_ptr != ecode)) |
(rrc != MATCH_THEN || md->start_match_ptr != ecode)) |
1579 |
} |
} |
1580 |
|
|
1581 |
md->recursive = &new_recursive; |
md->recursive = &new_recursive; |
|
memcpy(md->offset_vector, new_recursive.offset_save, |
|
|
new_recursive.saved_max * sizeof(int)); |
|
1582 |
callpat += GET(callpat, 1); |
callpat += GET(callpat, 1); |
1583 |
} |
} |
1584 |
while (*callpat == OP_ALT); |
while (*callpat == OP_ALT); |
1589 |
(pcre_free)(new_recursive.offset_save); |
(pcre_free)(new_recursive.offset_save); |
1590 |
MRRETURN(MATCH_NOMATCH); |
MRRETURN(MATCH_NOMATCH); |
1591 |
} |
} |
|
/* Control never reaches here */ |
|
|
|
|
|
/* "Once" brackets are like assertion brackets except that after a match, |
|
|
the point in the subject string is not moved back. Thus there can never be |
|
|
a move back into the brackets. Friedl calls these "atomic" subpatterns. |
|
|
Check the alternative branches in turn - the matching won't pass the KET |
|
|
for this kind of subpattern. If any one branch matches, we carry on as at |
|
|
the end of a normal bracket, leaving the subject pointer, but resetting |
|
|
the start-of-match value in case it was changed by \K. */ |
|
|
|
|
|
case OP_ONCE: |
|
|
prev = ecode; |
|
|
saved_eptr = eptr; |
|
|
|
|
|
do |
|
|
{ |
|
|
RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, ims, eptrb, 0, RM7); |
|
|
if (rrc == MATCH_MATCH) /* Note: _not_ MATCH_ACCEPT */ |
|
|
{ |
|
|
mstart = md->start_match_ptr; |
|
|
break; |
|
|
} |
|
|
if (rrc != MATCH_NOMATCH && |
|
|
(rrc != MATCH_THEN || md->start_match_ptr != ecode)) |
|
|
RRETURN(rrc); |
|
|
ecode += GET(ecode,1); |
|
|
} |
|
|
while (*ecode == OP_ALT); |
|
|
|
|
|
/* If hit the end of the group (which could be repeated), fail */ |
|
|
|
|
|
if (*ecode != OP_ONCE && *ecode != OP_ALT) RRETURN(MATCH_NOMATCH); |
|
|
|
|
|
/* Continue as from after the assertion, updating the offsets high water |
|
|
mark, since extracts may have been taken. */ |
|
|
|
|
|
do ecode += GET(ecode, 1); while (*ecode == OP_ALT); |
|
|
|
|
|
offset_top = md->end_offset_top; |
|
|
eptr = md->end_match_ptr; |
|
|
|
|
|
/* For a non-repeating ket, just continue at this level. This also |
|
|
happens for a repeating ket if no characters were matched in the group. |
|
|
This is the forcible breaking of infinite loops as implemented in Perl |
|
|
5.005. If there is an options reset, it will get obeyed in the normal |
|
|
course of events. */ |
|
|
|
|
|
if (*ecode == OP_KET || eptr == saved_eptr) |
|
|
{ |
|
|
ecode += 1+LINK_SIZE; |
|
|
break; |
|
|
} |
|
|
|
|
|
/* The repeating kets try the rest of the pattern or restart from the |
|
|
preceding bracket, in the appropriate order. The second "call" of match() |
|
|
uses tail recursion, to avoid using another stack frame. We need to reset |
|
|
any options that changed within the bracket before re-running it, so |
|
|
check the next opcode. */ |
|
|
|
|
|
if (ecode[1+LINK_SIZE] == OP_OPT) |
|
|
{ |
|
|
ims = (ims & ~PCRE_IMS) | ecode[4]; |
|
|
DPRINTF(("ims set to %02lx at group repeat\n", ims)); |
|
|
} |
|
1592 |
|
|
1593 |
if (*ecode == OP_KETRMIN) |
RECURSION_MATCHED: |
1594 |
{ |
break; |
|
RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, ims, eptrb, 0, RM8); |
|
|
if (rrc != MATCH_NOMATCH) RRETURN(rrc); |
|
|
ecode = prev; |
|
|
flags = 0; |
|
|
goto TAIL_RECURSE; |
|
|
} |
|
|
else /* OP_KETRMAX */ |
|
|
{ |
|
|
RMATCH(eptr, prev, offset_top, md, ims, eptrb, match_cbegroup, RM9); |
|
|
if (rrc != MATCH_NOMATCH) RRETURN(rrc); |
|
|
ecode += 1 + LINK_SIZE; |
|
|
flags = 0; |
|
|
goto TAIL_RECURSE; |
|
|
} |
|
|
/* Control never gets here */ |
|
1595 |
|
|
1596 |
/* An alternation is the end of a branch; scan along to find the end of the |
/* An alternation is the end of a branch; scan along to find the end of the |
1597 |
bracketed group and go to there. */ |
bracketed group and go to there. */ |
1607 |
optional ones preceded by BRAZERO or BRAMINZERO. */ |
optional ones preceded by BRAZERO or BRAMINZERO. */ |
1608 |
|
|
1609 |
case OP_BRAZERO: |
case OP_BRAZERO: |
1610 |
{ |
next = ecode + 1; |
1611 |
next = ecode+1; |
RMATCH(eptr, next, offset_top, md, eptrb, RM10); |
1612 |
RMATCH(eptr, next, offset_top, md, ims, eptrb, 0, RM10); |
if (rrc != MATCH_NOMATCH) RRETURN(rrc); |
1613 |
if (rrc != MATCH_NOMATCH) RRETURN(rrc); |
do next += GET(next, 1); while (*next == OP_ALT); |
1614 |
do next += GET(next,1); while (*next == OP_ALT); |
ecode = next + 1 + LINK_SIZE; |
|
ecode = next + 1 + LINK_SIZE; |
|
|
} |
|
1615 |
break; |
break; |
1616 |
|
|
1617 |
case OP_BRAMINZERO: |
case OP_BRAMINZERO: |
1618 |
{ |
next = ecode + 1; |
1619 |
next = ecode+1; |
do next += GET(next, 1); while (*next == OP_ALT); |
1620 |
do next += GET(next, 1); while (*next == OP_ALT); |
RMATCH(eptr, next + 1+LINK_SIZE, offset_top, md, eptrb, RM11); |
1621 |
RMATCH(eptr, next + 1+LINK_SIZE, offset_top, md, ims, eptrb, 0, RM11); |
if (rrc != MATCH_NOMATCH) RRETURN(rrc); |
1622 |
if (rrc != MATCH_NOMATCH) RRETURN(rrc); |
ecode++; |
|
ecode++; |
|
|
} |
|
1623 |
break; |
break; |
1624 |
|
|
1625 |
case OP_SKIPZERO: |
case OP_SKIPZERO: |
1626 |
{ |
next = ecode+1; |
1627 |
next = ecode+1; |
do next += GET(next,1); while (*next == OP_ALT); |
1628 |
do next += GET(next,1); while (*next == OP_ALT); |
ecode = next + 1 + LINK_SIZE; |
|
ecode = next + 1 + LINK_SIZE; |
|
|
} |
|
1629 |
break; |
break; |
1630 |
|
|
1631 |
|
/* BRAPOSZERO occurs before a possessive bracket group. Don't do anything |
1632 |
|
here; just jump to the group, with allow_zero set TRUE. */ |
1633 |
|
|
1634 |
|
case OP_BRAPOSZERO: |
1635 |
|
op = *(++ecode); |
1636 |
|
allow_zero = TRUE; |
1637 |
|
if (op == OP_CBRAPOS || op == OP_SCBRAPOS) goto POSSESSIVE_CAPTURE; |
1638 |
|
goto POSSESSIVE_NON_CAPTURE; |
1639 |
|
|
1640 |
/* End of a group, repeated or non-repeating. */ |
/* End of a group, repeated or non-repeating. */ |
1641 |
|
|
1642 |
case OP_KET: |
case OP_KET: |
1643 |
case OP_KETRMIN: |
case OP_KETRMIN: |
1644 |
case OP_KETRMAX: |
case OP_KETRMAX: |
1645 |
|
case OP_KETRPOS: |
1646 |
prev = ecode - GET(ecode, 1); |
prev = ecode - GET(ecode, 1); |
1647 |
|
|
1648 |
/* If this was a group that remembered the subject start, in order to break |
/* If this was a group that remembered the subject start, in order to break |
1649 |
infinite repeats of empty string matches, retrieve the subject start from |
infinite repeats of empty string matches, retrieve the subject start from |
1650 |
the chain. Otherwise, set it NULL. */ |
the chain. Otherwise, set it NULL. */ |
1651 |
|
|
1652 |
if (*prev >= OP_SBRA) |
if (*prev >= OP_SBRA || *prev == OP_ONCE) |
1653 |
{ |
{ |
1654 |
saved_eptr = eptrb->epb_saved_eptr; /* Value at start of group */ |
saved_eptr = eptrb->epb_saved_eptr; /* Value at start of group */ |
1655 |
eptrb = eptrb->epb_prev; /* Backup to previous group */ |
eptrb = eptrb->epb_prev; /* Backup to previous group */ |
1656 |
} |
} |
1657 |
else saved_eptr = NULL; |
else saved_eptr = NULL; |
1658 |
|
|
1659 |
/* If we are at the end of an assertion group or an atomic group, stop |
/* If we are at the end of an assertion group, stop matching and return |
1660 |
matching and return MATCH_MATCH, but record the current high water mark for |
MATCH_MATCH, but record the current high water mark for use by positive |
1661 |
use by positive assertions. We also need to record the match start in case |
assertions. We also need to record the match start in case it was changed |
1662 |
it was changed by \K. */ |
by \K. */ |
1663 |
|
|
1664 |
if (*prev == OP_ASSERT || *prev == OP_ASSERT_NOT || |
if (*prev == OP_ASSERT || *prev == OP_ASSERT_NOT || |
1665 |
*prev == OP_ASSERTBACK || *prev == OP_ASSERTBACK_NOT || |
*prev == OP_ASSERTBACK || *prev == OP_ASSERTBACK_NOT) |
|
*prev == OP_ONCE) |
|
1666 |
{ |
{ |
1667 |
md->end_match_ptr = eptr; /* For ONCE */ |
md->end_match_ptr = eptr; /* For ONCE */ |
1668 |
md->end_offset_top = offset_top; |
md->end_offset_top = offset_top; |
1669 |
md->start_match_ptr = mstart; |
md->start_match_ptr = mstart; |
1670 |
MRRETURN(MATCH_MATCH); |
MRRETURN(MATCH_MATCH); /* Sets md->mark */ |
1671 |
} |
} |
1672 |
|
|
1673 |
/* For capturing groups we have to check the group number back at the start |
/* For capturing groups we have to check the group number back at the start |
1674 |
and if necessary complete handling an extraction by setting the offsets and |
and if necessary complete handling an extraction by setting the offsets and |
1675 |
bumping the high water mark. Note that whole-pattern recursion is coded as |
bumping the high water mark. Whole-pattern recursion is coded as a recurse |
1676 |
a recurse into group 0, so it won't be picked up here. Instead, we catch it |
into group 0, so it won't be picked up here. Instead, we catch it when the |
1677 |
when the OP_END is reached. Other recursion is handled here. */ |
OP_END is reached. Other recursion is handled here. We just have to record |
1678 |
|
the current subject position and start match pointer and give a MATCH |
1679 |
|
return. */ |
1680 |
|
|
1681 |
if (*prev == OP_CBRA || *prev == OP_SCBRA) |
if (*prev == OP_CBRA || *prev == OP_SCBRA || |
1682 |
|
*prev == OP_CBRAPOS || *prev == OP_SCBRAPOS) |
1683 |
{ |
{ |
1684 |
number = GET2(prev, 1+LINK_SIZE); |
number = GET2(prev, 1+LINK_SIZE); |
1685 |
offset = number << 1; |
offset = number << 1; |
1689 |
printf("\n"); |
printf("\n"); |
1690 |
#endif |
#endif |
1691 |
|
|
1692 |
|
/* Handle a recursively called group. */ |
1693 |
|
|
1694 |
|
if (md->recursive != NULL && md->recursive->group_num == number) |
1695 |
|
{ |
1696 |
|
md->end_match_ptr = eptr; |
1697 |
|
md->start_match_ptr = mstart; |
1698 |
|
RRETURN(MATCH_MATCH); |
1699 |
|
} |
1700 |
|
|
1701 |
|
/* Deal with capturing */ |
1702 |
|
|
1703 |
md->capture_last = number; |
md->capture_last = number; |
1704 |
if (offset >= md->offset_max) md->offset_overflow = TRUE; else |
if (offset >= md->offset_max) md->offset_overflow = TRUE; else |
1705 |
{ |
{ |
1706 |
|
/* If offset is greater than offset_top, it means that we are |
1707 |
|
"skipping" a capturing group, and that group's offsets must be marked |
1708 |
|
unset. In earlier versions of PCRE, all the offsets were unset at the |
1709 |
|
start of matching, but this doesn't work because atomic groups and |
1710 |
|
assertions can cause a value to be set that should later be unset. |
1711 |
|
Example: matching /(?>(a))b|(a)c/ against "ac". This sets group 1 as |
1712 |
|
part of the atomic group, but this is not on the final matching path, |
1713 |
|
so must be unset when 2 is set. (If there is no group 2, there is no |
1714 |
|
problem, because offset_top will then be 2, indicating no capture.) */ |
1715 |
|
|
1716 |
|
if (offset > offset_top) |
1717 |
|
{ |
1718 |
|
register int *iptr = md->offset_vector + offset_top; |
1719 |
|
register int *iend = md->offset_vector + offset; |
1720 |
|
while (iptr < iend) *iptr++ = -1; |
1721 |
|
} |
1722 |
|
|
1723 |
|
/* Now make the extraction */ |
1724 |
|
|
1725 |
md->offset_vector[offset] = |
md->offset_vector[offset] = |
1726 |
md->offset_vector[md->offset_end - number]; |
md->offset_vector[md->offset_end - number]; |
1727 |
md->offset_vector[offset+1] = (int)(eptr - md->start_subject); |
md->offset_vector[offset+1] = (int)(eptr - md->start_subject); |
1728 |
if (offset_top <= offset) offset_top = offset + 2; |
if (offset_top <= offset) offset_top = offset + 2; |
1729 |
} |
} |
1730 |
|
} |
1731 |
|
|
1732 |
/* Handle a recursively called group. Restore the offsets |
/* For an ordinary non-repeating ket, just continue at this level. This |
1733 |
appropriately and continue from after the call. */ |
also happens for a repeating ket if no characters were matched in the |
1734 |
|
group. This is the forcible breaking of infinite loops as implemented in |
1735 |
|
Perl 5.005. For a non-repeating atomic group, establish a backup point by |
1736 |
|
processing the rest of the pattern at a lower level. If this results in a |
1737 |
|
NOMATCH return, pass MATCH_ONCE back to the original OP_ONCE level, thereby |
1738 |
|
bypassing intermediate backup points, but resetting any captures that |
1739 |
|
happened along the way. */ |
1740 |
|
|
1741 |
if (md->recursive != NULL && md->recursive->group_num == number) |
if (*ecode == OP_KET || eptr == saved_eptr) |
1742 |
|
{ |
1743 |
|
if (*prev == OP_ONCE) |
1744 |
{ |
{ |
1745 |
recursion_info *rec = md->recursive; |
RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, eptrb, RM12); |
1746 |
DPRINTF(("Recursion (%d) succeeded - continuing\n", number)); |
if (rrc != MATCH_NOMATCH) RRETURN(rrc); |
1747 |
md->recursive = rec->prevrec; |
md->once_target = prev; /* Level at which to change to MATCH_NOMATCH */ |
1748 |
memcpy(md->offset_vector, rec->offset_save, |
RRETURN(MATCH_ONCE); |
|
rec->saved_max * sizeof(int)); |
|
|
offset_top = rec->save_offset_top; |
|
|
ecode = rec->after_call; |
|
|
ims = original_ims; |
|
|
break; |
|
1749 |
} |
} |
1750 |
|
ecode += 1 + LINK_SIZE; /* Carry on at this level */ |
1751 |
|
break; |
1752 |
} |
} |
1753 |
|
|
1754 |
/* For both capturing and non-capturing groups, reset the value of the ims |
/* OP_KETRPOS is a possessive repeating ket. Remember the current position, |
1755 |
flags, in case they got changed during the group. */ |
and return the MATCH_KETRPOS. This makes it possible to do the repeats one |
1756 |
|
at a time from the outer level, thus saving stack. */ |
|
ims = original_ims; |
|
|
DPRINTF(("ims reset to %02lx\n", ims)); |
|
|
|
|
|
/* For a non-repeating ket, just continue at this level. This also |
|
|
happens for a repeating ket if no characters were matched in the group. |
|
|
This is the forcible breaking of infinite loops as implemented in Perl |
|
|
5.005. If there is an options reset, it will get obeyed in the normal |
|
|
course of events. */ |
|
1757 |
|
|
1758 |
if (*ecode == OP_KET || eptr == saved_eptr) |
if (*ecode == OP_KETRPOS) |
1759 |
{ |
{ |
1760 |
ecode += 1 + LINK_SIZE; |
md->end_match_ptr = eptr; |
1761 |
break; |
md->end_offset_top = offset_top; |
1762 |
|
RRETURN(MATCH_KETRPOS); |
1763 |
} |
} |
1764 |
|
|
1765 |
/* The repeating kets try the rest of the pattern or restart from the |
/* The normal repeating kets try the rest of the pattern or restart from |
1766 |
preceding bracket, in the appropriate order. In the second case, we can use |
the preceding bracket, in the appropriate order. In the second case, we can |
1767 |
tail recursion to avoid using another stack frame, unless we have an |
use tail recursion to avoid using another stack frame, unless we have an |
1768 |
unlimited repeat of a group that can match an empty string. */ |
an atomic group or an unlimited repeat of a group that can match an empty |
1769 |
|
string. */ |
|
flags = (*prev >= OP_SBRA)? match_cbegroup : 0; |
|
1770 |
|
|
1771 |
if (*ecode == OP_KETRMIN) |
if (*ecode == OP_KETRMIN) |
1772 |
{ |
{ |
1773 |
RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, ims, eptrb, 0, RM12); |
RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, eptrb, RM7); |
1774 |
if (rrc != MATCH_NOMATCH) RRETURN(rrc); |
if (rrc != MATCH_NOMATCH) RRETURN(rrc); |
1775 |
if (flags != 0) /* Could match an empty string */ |
if (*prev == OP_ONCE) |
1776 |
{ |
{ |
1777 |
RMATCH(eptr, prev, offset_top, md, ims, eptrb, flags, RM50); |
RMATCH(eptr, prev, offset_top, md, eptrb, RM8); |
1778 |
|
if (rrc != MATCH_NOMATCH) RRETURN(rrc); |
1779 |
|
md->once_target = prev; /* Level at which to change to MATCH_NOMATCH */ |
1780 |
|
RRETURN(MATCH_ONCE); |
1781 |
|
} |
1782 |
|
if (*prev >= OP_SBRA) /* Could match an empty string */ |
1783 |
|
{ |
1784 |
|
md->match_function_type = MATCH_CBEGROUP; |
1785 |
|
RMATCH(eptr, prev, offset_top, md, eptrb, RM50); |
1786 |
RRETURN(rrc); |
RRETURN(rrc); |
1787 |
} |
} |
1788 |
ecode = prev; |
ecode = prev; |
1790 |
} |
} |
1791 |
else /* OP_KETRMAX */ |
else /* OP_KETRMAX */ |
1792 |
{ |
{ |
1793 |
RMATCH(eptr, prev, offset_top, md, ims, eptrb, flags, RM13); |
if (*prev >= OP_SBRA) md->match_function_type = MATCH_CBEGROUP; |
1794 |
|
RMATCH(eptr, prev, offset_top, md, eptrb, RM13); |
1795 |
|
if (rrc == MATCH_ONCE && md->once_target == prev) rrc = MATCH_NOMATCH; |
1796 |
if (rrc != MATCH_NOMATCH) RRETURN(rrc); |
if (rrc != MATCH_NOMATCH) RRETURN(rrc); |
1797 |
|
if (*prev == OP_ONCE) |
1798 |
|
{ |
1799 |
|
RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, eptrb, RM9); |
1800 |
|
if (rrc != MATCH_NOMATCH) RRETURN(rrc); |
1801 |
|
md->once_target = prev; |
1802 |
|
RRETURN(MATCH_ONCE); |
1803 |
|
} |
1804 |
ecode += 1 + LINK_SIZE; |
ecode += 1 + LINK_SIZE; |
|
flags = 0; |
|
1805 |
goto TAIL_RECURSE; |
goto TAIL_RECURSE; |
1806 |
} |
} |
1807 |
/* Control never gets here */ |
/* Control never gets here */ |
1808 |
|
|
1809 |
/* Start of subject unless notbol, or after internal newline if multiline */ |
/* Not multiline mode: start of subject assertion, unless notbol. */ |
1810 |
|
|
1811 |
case OP_CIRC: |
case OP_CIRC: |
1812 |
if (md->notbol && eptr == md->start_subject) MRRETURN(MATCH_NOMATCH); |
if (md->notbol && eptr == md->start_subject) MRRETURN(MATCH_NOMATCH); |
|
if ((ims & PCRE_MULTILINE) != 0) |
|
|
{ |
|
|
if (eptr != md->start_subject && |
|
|
(eptr == md->end_subject || !WAS_NEWLINE(eptr))) |
|
|
MRRETURN(MATCH_NOMATCH); |
|
|
ecode++; |
|
|
break; |
|
|
} |
|
|
/* ... else fall through */ |
|
1813 |
|
|
1814 |
/* Start of subject assertion */ |
/* Start of subject assertion */ |
1815 |
|
|
1818 |
ecode++; |
ecode++; |
1819 |
break; |
break; |
1820 |
|
|
1821 |
|
/* Multiline mode: start of subject unless notbol, or after any newline. */ |
1822 |
|
|
1823 |
|
case OP_CIRCM: |
1824 |
|
if (md->notbol && eptr == md->start_subject) MRRETURN(MATCH_NOMATCH); |
1825 |
|
if (eptr != md->start_subject && |
1826 |
|
(eptr == md->end_subject || !WAS_NEWLINE(eptr))) |
1827 |
|
MRRETURN(MATCH_NOMATCH); |
1828 |
|
ecode++; |
1829 |
|
break; |
1830 |
|
|
1831 |
/* Start of match assertion */ |
/* Start of match assertion */ |
1832 |
|
|
1833 |
case OP_SOM: |
case OP_SOM: |
1842 |
ecode++; |
ecode++; |
1843 |
break; |
break; |
1844 |
|
|
1845 |
/* Assert before internal newline if multiline, or before a terminating |
/* Multiline mode: assert before any newline, or before end of subject |
1846 |
newline unless endonly is set, else end of subject unless noteol is set. */ |
unless noteol is set. */ |
1847 |
|
|
1848 |
case OP_DOLL: |
case OP_DOLLM: |
1849 |
if ((ims & PCRE_MULTILINE) != 0) |
if (eptr < md->end_subject) |
1850 |
{ |
{ if (!IS_NEWLINE(eptr)) MRRETURN(MATCH_NOMATCH); } |
1851 |
if (eptr < md->end_subject) |
else |
|
{ if (!IS_NEWLINE(eptr)) MRRETURN(MATCH_NOMATCH); } |
|
|
else |
|
|
{ |
|
|
if (md->noteol) MRRETURN(MATCH_NOMATCH); |
|
|
SCHECK_PARTIAL(); |
|
|
} |
|
|
ecode++; |
|
|
break; |
|
|
} |
|
|
else /* Not multiline */ |
|
1852 |
{ |
{ |
1853 |
if (md->noteol) MRRETURN(MATCH_NOMATCH); |
if (md->noteol) MRRETURN(MATCH_NOMATCH); |
1854 |
if (!md->endonly) goto ASSERT_NL_OR_EOS; |
SCHECK_PARTIAL(); |
1855 |
} |
} |
1856 |
|
ecode++; |
1857 |
|
break; |
1858 |
|
|
1859 |
|
/* Not multiline mode: assert before a terminating newline or before end of |
1860 |
|
subject unless noteol is set. */ |
1861 |
|
|
1862 |
|
case OP_DOLL: |
1863 |
|
if (md->noteol) MRRETURN(MATCH_NOMATCH); |
1864 |
|
if (!md->endonly) goto ASSERT_NL_OR_EOS; |
1865 |
|
|
1866 |
/* ... else fall through for endonly */ |
/* ... else fall through for endonly */ |
1867 |
|
|
1868 |
/* End of subject assertion (\z) */ |
/* End of subject assertion (\z) */ |
1880 |
if (eptr < md->end_subject && |
if (eptr < md->end_subject && |
1881 |
(!IS_NEWLINE(eptr) || eptr != md->end_subject - md->nllen)) |
(!IS_NEWLINE(eptr) || eptr != md->end_subject - md->nllen)) |
1882 |
MRRETURN(MATCH_NOMATCH); |
MRRETURN(MATCH_NOMATCH); |
1883 |
|
|
1884 |
/* Either at end of string or \n before end. */ |
/* Either at end of string or \n before end. */ |
1885 |
|
|
1886 |
SCHECK_PARTIAL(); |
SCHECK_PARTIAL(); |
1887 |
ecode++; |
ecode++; |
1888 |
break; |
break; |
2145 |
switch(c) |
switch(c) |
2146 |
{ |
{ |
2147 |
default: MRRETURN(MATCH_NOMATCH); |
default: MRRETURN(MATCH_NOMATCH); |
2148 |
|
|
2149 |
case 0x000d: |
case 0x000d: |
2150 |
if (eptr < md->end_subject && *eptr == 0x0a) eptr++; |
if (eptr < md->end_subject && *eptr == 0x0a) eptr++; |
2151 |
break; |
break; |
2369 |
MRRETURN(MATCH_NOMATCH); |
MRRETURN(MATCH_NOMATCH); |
2370 |
} |
} |
2371 |
GETCHARINCTEST(c, eptr); |
GETCHARINCTEST(c, eptr); |
2372 |
|
if (UCD_CATEGORY(c) == ucp_M) MRRETURN(MATCH_NOMATCH); |
2373 |
|
while (eptr < md->end_subject) |
2374 |
{ |
{ |
2375 |
int category = UCD_CATEGORY(c); |
int len = 1; |
2376 |
if (category == ucp_M) MRRETURN(MATCH_NOMATCH); |
if (!utf8) c = *eptr; else { GETCHARLEN(c, eptr, len); } |
2377 |
while (eptr < md->end_subject) |
if (UCD_CATEGORY(c) != ucp_M) break; |
2378 |
{ |
eptr += len; |
|
int len = 1; |
|
|
if (!utf8) c = *eptr; else |
|
|
{ |
|
|
GETCHARLEN(c, eptr, len); |
|
|
} |
|
|
category = UCD_CATEGORY(c); |
|
|
if (category != ucp_M) break; |
|
|
eptr += len; |
|
|
} |
|
2379 |
} |
} |
2380 |
ecode++; |
ecode++; |
2381 |
break; |
break; |
2391 |
loops). */ |
loops). */ |
2392 |
|
|
2393 |
case OP_REF: |
case OP_REF: |
2394 |
{ |
case OP_REFI: |
2395 |
offset = GET2(ecode, 1) << 1; /* Doubled ref number */ |
caseless = op == OP_REFI; |
2396 |
ecode += 3; |
offset = GET2(ecode, 1) << 1; /* Doubled ref number */ |
2397 |
|
ecode += 3; |
2398 |
|
|
2399 |
/* If the reference is unset, there are two possibilities: |
/* If the reference is unset, there are two possibilities: |
2400 |
|
|
2401 |
(a) In the default, Perl-compatible state, set the length to be longer |
(a) In the default, Perl-compatible state, set the length negative; |
2402 |
than the amount of subject left; this ensures that every attempt at a |
this ensures that every attempt at a match fails. We can't just fail |
2403 |
match fails. We can't just fail here, because of the possibility of |
here, because of the possibility of quantifiers with zero minima. |
|
quantifiers with zero minima. |
|
2404 |
|
|
2405 |
(b) If the JavaScript compatibility flag is set, set the length to zero |
(b) If the JavaScript compatibility flag is set, set the length to zero |
2406 |
so that the back reference matches an empty string. |
so that the back reference matches an empty string. |
2407 |
|
|
2408 |
Otherwise, set the length to the length of what was matched by the |
Otherwise, set the length to the length of what was matched by the |
2409 |
referenced subpattern. */ |
referenced subpattern. */ |
2410 |
|
|
2411 |
if (offset >= offset_top || md->offset_vector[offset] < 0) |
if (offset >= offset_top || md->offset_vector[offset] < 0) |
2412 |
length = (md->jscript_compat)? 0 : (int)(md->end_subject - eptr + 1); |
length = (md->jscript_compat)? 0 : -1; |
2413 |
else |
else |
2414 |
length = md->offset_vector[offset+1] - md->offset_vector[offset]; |
length = md->offset_vector[offset+1] - md->offset_vector[offset]; |
2415 |
|
|
2416 |
/* Set up for repetition, or handle the non-repeated case */ |
/* Set up for repetition, or handle the non-repeated case */ |
2417 |
|
|
2418 |
switch (*ecode) |
switch (*ecode) |
2419 |
{ |
{ |
2420 |
case OP_CRSTAR: |
case OP_CRSTAR: |
2421 |
case OP_CRMINSTAR: |
case OP_CRMINSTAR: |
2422 |
case OP_CRPLUS: |
case OP_CRPLUS: |
2423 |
case OP_CRMINPLUS: |
case OP_CRMINPLUS: |
2424 |
case OP_CRQUERY: |
case OP_CRQUERY: |
2425 |
case OP_CRMINQUERY: |
case OP_CRMINQUERY: |
2426 |
c = *ecode++ - OP_CRSTAR; |
c = *ecode++ - OP_CRSTAR; |
2427 |
minimize = (c & 1) != 0; |
minimize = (c & 1) != 0; |
2428 |
min = rep_min[c]; /* Pick up values from tables; */ |
min = rep_min[c]; /* Pick up values from tables; */ |
2429 |
max = rep_max[c]; /* zero for max => infinity */ |
max = rep_max[c]; /* zero for max => infinity */ |
2430 |
if (max == 0) max = INT_MAX; |
if (max == 0) max = INT_MAX; |
2431 |
break; |
break; |
2432 |
|
|
2433 |
case OP_CRRANGE: |
case OP_CRRANGE: |
2434 |
case OP_CRMINRANGE: |
case OP_CRMINRANGE: |
2435 |
minimize = (*ecode == OP_CRMINRANGE); |
minimize = (*ecode == OP_CRMINRANGE); |
2436 |
min = GET2(ecode, 1); |
min = GET2(ecode, 1); |
2437 |
max = GET2(ecode, 3); |
max = GET2(ecode, 3); |
2438 |
if (max == 0) max = INT_MAX; |
if (max == 0) max = INT_MAX; |
2439 |
ecode += 5; |
ecode += 5; |
2440 |
break; |
break; |
2441 |
|
|
2442 |
default: /* No repeat follows */ |
default: /* No repeat follows */ |
2443 |
if (!match_ref(offset, eptr, length, md, ims)) |
if ((length = match_ref(offset, eptr, length, md, caseless)) < 0) |
2444 |
{ |
{ |
2445 |
CHECK_PARTIAL(); |
CHECK_PARTIAL(); |
2446 |
MRRETURN(MATCH_NOMATCH); |
MRRETURN(MATCH_NOMATCH); |
|
} |
|
|
eptr += length; |
|
|
continue; /* With the main loop */ |
|
2447 |
} |
} |
2448 |
|
eptr += length; |
2449 |
|
continue; /* With the main loop */ |
2450 |
|
} |
2451 |
|
|
2452 |
/* If the length of the reference is zero, just continue with the |
/* Handle repeated back references. If the length of the reference is |
2453 |
main loop. */ |
zero, just continue with the main loop. */ |
2454 |
|
|
2455 |
if (length == 0) continue; |
if (length == 0) continue; |
2456 |
|
|
2457 |
/* First, ensure the minimum number of matches are present. We get back |
/* First, ensure the minimum number of matches are present. We get back |
2458 |
the length of the reference string explicitly rather than passing the |
the length of the reference string explicitly rather than passing the |
2459 |
address of eptr, so that eptr can be a register variable. */ |
address of eptr, so that eptr can be a register variable. */ |
2460 |
|
|
2461 |
for (i = 1; i <= min; i++) |
for (i = 1; i <= min; i++) |
2462 |
|
{ |
2463 |
|
int slength; |
2464 |
|
if ((slength = match_ref(offset, eptr, length, md, caseless)) < 0) |
2465 |
{ |
{ |
2466 |
if (!match_ref(offset, eptr, length, md, ims)) |
CHECK_PARTIAL(); |
2467 |
{ |
MRRETURN(MATCH_NOMATCH); |
|
CHECK_PARTIAL(); |
|
|
MRRETURN(MATCH_NOMATCH); |
|
|
} |
|
|
eptr += length; |
|
2468 |
} |
} |
2469 |
|
eptr += slength; |
2470 |
|
} |
2471 |
|
|
2472 |
/* If min = max, continue at the same level without recursion. |
/* If min = max, continue at the same level without recursion. |
2473 |
They are not both allowed to be zero. */ |
They are not both allowed to be zero. */ |
2474 |
|
|
2475 |
if (min == max) continue; |
if (min == max) continue; |
2476 |
|
|
2477 |
/* If minimizing, keep trying and advancing the pointer */ |
/* If minimizing, keep trying and advancing the pointer */ |
2478 |
|
|
2479 |
if (minimize) |
if (minimize) |
2480 |
|
{ |
2481 |
|
for (fi = min;; fi++) |
2482 |
{ |
{ |
2483 |
for (fi = min;; fi++) |
int slength; |
2484 |
|
RMATCH(eptr, ecode, offset_top, md, eptrb, RM14); |
2485 |
|
if (rrc != MATCH_NOMATCH) RRETURN(rrc); |
2486 |
|
if (fi >= max) MRRETURN(MATCH_NOMATCH); |
2487 |
|
if ((slength = match_ref(offset, eptr, length, md, caseless)) < 0) |
2488 |
{ |
{ |
2489 |
RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM14); |
CHECK_PARTIAL(); |
2490 |
if (rrc != MATCH_NOMATCH) RRETURN(rrc); |
MRRETURN(MATCH_NOMATCH); |
|
if (fi >= max) MRRETURN(MATCH_NOMATCH); |
|
|
if (!match_ref(offset, eptr, length, md, ims)) |
|
|
{ |
|
|
CHECK_PARTIAL(); |
|
|
MRRETURN(MATCH_NOMATCH); |
|
|
} |
|
|
eptr += length; |
|
2491 |
} |
} |
2492 |
/* Control never gets here */ |
eptr += slength; |
2493 |
} |
} |
2494 |
|
/* Control never gets here */ |
2495 |
|
} |
2496 |
|
|
2497 |
/* If maximizing, find the longest string and work backwards */ |
/* If maximizing, find the longest string and work backwards */ |
2498 |
|
|
2499 |
else |
else |
2500 |
|
{ |
2501 |
|
pp = eptr; |
2502 |
|
for (i = min; i < max; i++) |
2503 |
{ |
{ |
2504 |
pp = eptr; |
int slength; |
2505 |
for (i = min; i < max; i++) |
if ((slength = match_ref(offset, eptr, length, md, caseless)) < 0) |
|
{ |
|
|
if (!match_ref(offset, eptr, length, md, ims)) |
|
|
{ |
|
|
CHECK_PARTIAL(); |
|
|
break; |
|
|
} |
|
|
eptr += length; |
|
|
} |
|
|
while (eptr >= pp) |
|
2506 |
{ |
{ |
2507 |
RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM15); |
CHECK_PARTIAL(); |
2508 |
if (rrc != MATCH_NOMATCH) RRETURN(rrc); |
break; |
|
eptr -= length; |
|
2509 |
} |
} |
2510 |
MRRETURN(MATCH_NOMATCH); |
eptr += slength; |
2511 |
} |
} |
2512 |
|
while (eptr >= pp) |
2513 |
|
{ |
2514 |
|
RMATCH(eptr, ecode, offset_top, md, eptrb, RM15); |
2515 |
|
if (rrc != MATCH_NOMATCH) RRETURN(rrc); |
2516 |
|
eptr -= length; |
2517 |
|
} |
2518 |
|
MRRETURN(MATCH_NOMATCH); |
2519 |
} |
} |
2520 |
/* Control never gets here */ |
/* Control never gets here */ |
2521 |
|
|
2621 |
{ |
{ |
2622 |
for (fi = min;; fi++) |
for (fi = min;; fi++) |
2623 |
{ |
{ |
2624 |
RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM16); |
RMATCH(eptr, ecode, offset_top, md, eptrb, RM16); |
2625 |
if (rrc != MATCH_NOMATCH) RRETURN(rrc); |
if (rrc != MATCH_NOMATCH) RRETURN(rrc); |
2626 |
if (fi >= max) MRRETURN(MATCH_NOMATCH); |
if (fi >= max) MRRETURN(MATCH_NOMATCH); |
2627 |
if (eptr >= md->end_subject) |
if (eptr >= md->end_subject) |
2646 |
{ |
{ |
2647 |
for (fi = min;; fi++) |
for (fi = min;; fi++) |
2648 |
{ |
{ |
2649 |
RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM17); |
RMATCH(eptr, ecode, offset_top, md, eptrb, RM17); |
2650 |
if (rrc != MATCH_NOMATCH) RRETURN(rrc); |
if (rrc != MATCH_NOMATCH) RRETURN(rrc); |
2651 |
if (fi >= max) MRRETURN(MATCH_NOMATCH); |
if (fi >= max) MRRETURN(MATCH_NOMATCH); |
2652 |
if (eptr >= md->end_subject) |
if (eptr >= md->end_subject) |
2692 |
} |
} |
2693 |
for (;;) |
for (;;) |
2694 |
{ |
{ |
2695 |
RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM18); |
RMATCH(eptr, ecode, offset_top, md, eptrb, RM18); |
2696 |
if (rrc != MATCH_NOMATCH) RRETURN(rrc); |
if (rrc != MATCH_NOMATCH) RRETURN(rrc); |
2697 |
if (eptr-- == pp) break; /* Stop if tried at original pos */ |
if (eptr-- == pp) break; /* Stop if tried at original pos */ |
2698 |
BACKCHAR(eptr); |
BACKCHAR(eptr); |
2715 |
} |
} |
2716 |
while (eptr >= pp) |
while (eptr >= pp) |
2717 |
{ |
{ |
2718 |
RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM19); |
RMATCH(eptr, ecode, offset_top, md, eptrb, RM19); |
2719 |
if (rrc != MATCH_NOMATCH) RRETURN(rrc); |
if (rrc != MATCH_NOMATCH) RRETURN(rrc); |
2720 |
eptr--; |
eptr--; |
2721 |
} |
} |
2791 |
{ |
{ |
2792 |
for (fi = min;; fi++) |
for (fi = min;; fi++) |
2793 |
{ |
{ |
2794 |
RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM20); |
RMATCH(eptr, ecode, offset_top, md, eptrb, RM20); |
2795 |
if (rrc != MATCH_NOMATCH) RRETURN(rrc); |
if (rrc != MATCH_NOMATCH) RRETURN(rrc); |
2796 |
if (fi >= max) MRRETURN(MATCH_NOMATCH); |
if (fi >= max) MRRETURN(MATCH_NOMATCH); |
2797 |
if (eptr >= md->end_subject) |
if (eptr >= md->end_subject) |
2824 |
} |
} |
2825 |
for(;;) |
for(;;) |
2826 |
{ |
{ |
2827 |
RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM21); |
RMATCH(eptr, ecode, offset_top, md, eptrb, RM21); |
2828 |
if (rrc != MATCH_NOMATCH) RRETURN(rrc); |
if (rrc != MATCH_NOMATCH) RRETURN(rrc); |
2829 |
if (eptr-- == pp) break; /* Stop if tried at original pos */ |
if (eptr-- == pp) break; /* Stop if tried at original pos */ |
2830 |
if (utf8) BACKCHAR(eptr); |
if (utf8) BACKCHAR(eptr); |
2869 |
|
|
2870 |
/* Match a single character, caselessly */ |
/* Match a single character, caselessly */ |
2871 |
|
|
2872 |
case OP_CHARNC: |
case OP_CHARI: |
2873 |
#ifdef SUPPORT_UTF8 |
#ifdef SUPPORT_UTF8 |
2874 |
if (utf8) |
if (utf8) |
2875 |
{ |
{ |
2929 |
/* Match a single character repeatedly. */ |
/* Match a single character repeatedly. */ |
2930 |
|
|
2931 |
case OP_EXACT: |
case OP_EXACT: |
2932 |
|
case OP_EXACTI: |
2933 |
min = max = GET2(ecode, 1); |
min = max = GET2(ecode, 1); |
2934 |
ecode += 3; |
ecode += 3; |
2935 |
goto REPEATCHAR; |
goto REPEATCHAR; |
2936 |
|
|
2937 |
case OP_POSUPTO: |
case OP_POSUPTO: |
2938 |
|
case OP_POSUPTOI: |
2939 |
possessive = TRUE; |
possessive = TRUE; |
2940 |
/* Fall through */ |
/* Fall through */ |
2941 |
|
|
2942 |
case OP_UPTO: |
case OP_UPTO: |
2943 |
|
case OP_UPTOI: |
2944 |
case OP_MINUPTO: |
case OP_MINUPTO: |
2945 |
|
case OP_MINUPTOI: |
2946 |
min = 0; |
min = 0; |
2947 |
max = GET2(ecode, 1); |
max = GET2(ecode, 1); |
2948 |
minimize = *ecode == OP_MINUPTO; |
minimize = *ecode == OP_MINUPTO || *ecode == OP_MINUPTOI; |
2949 |
ecode += 3; |
ecode += 3; |
2950 |
goto REPEATCHAR; |
goto REPEATCHAR; |
2951 |
|
|
2952 |
case OP_POSSTAR: |
case OP_POSSTAR: |
2953 |
|
case OP_POSSTARI: |
2954 |
possessive = TRUE; |
possessive = TRUE; |
2955 |
min = 0; |
min = 0; |
2956 |
max = INT_MAX; |
max = INT_MAX; |
2958 |
goto REPEATCHAR; |
goto REPEATCHAR; |
2959 |
|
|
2960 |
case OP_POSPLUS: |
case OP_POSPLUS: |
2961 |
|
case OP_POSPLUSI: |
2962 |
possessive = TRUE; |
possessive = TRUE; |
2963 |
min = 1; |
min = 1; |
2964 |
max = INT_MAX; |
max = INT_MAX; |
2966 |
goto REPEATCHAR; |
goto REPEATCHAR; |
2967 |
|
|
2968 |
case OP_POSQUERY: |
case OP_POSQUERY: |
2969 |
|
case OP_POSQUERYI: |
2970 |
possessive = TRUE; |
possessive = TRUE; |
2971 |
min = 0; |
min = 0; |
2972 |
max = 1; |
max = 1; |
2974 |
goto REPEATCHAR; |
goto REPEATCHAR; |
2975 |
|
|
2976 |
case OP_STAR: |
case OP_STAR: |
2977 |
|
case OP_STARI: |
2978 |
case OP_MINSTAR: |
case OP_MINSTAR: |
2979 |
|
case OP_MINSTARI: |
2980 |
case OP_PLUS: |
case OP_PLUS: |
2981 |
|
case OP_PLUSI: |
2982 |
case OP_MINPLUS: |
case OP_MINPLUS: |
2983 |
|
case OP_MINPLUSI: |
2984 |
case OP_QUERY: |
case OP_QUERY: |
2985 |
|
case OP_QUERYI: |
2986 |
case OP_MINQUERY: |
case OP_MINQUERY: |
2987 |
c = *ecode++ - OP_STAR; |
case OP_MINQUERYI: |
2988 |
|
c = *ecode++ - ((op < OP_STARI)? OP_STAR : OP_STARI); |
2989 |
minimize = (c & 1) != 0; |
minimize = (c & 1) != 0; |
|
|
|
2990 |
min = rep_min[c]; /* Pick up values from tables; */ |
min = rep_min[c]; /* Pick up values from tables; */ |
2991 |
max = rep_max[c]; /* zero for max => infinity */ |
max = rep_max[c]; /* zero for max => infinity */ |
2992 |
if (max == 0) max = INT_MAX; |
if (max == 0) max = INT_MAX; |
3009 |
{ |
{ |
3010 |
#ifdef SUPPORT_UCP |
#ifdef SUPPORT_UCP |
3011 |
unsigned int othercase; |
unsigned int othercase; |
3012 |
if ((ims & PCRE_CASELESS) != 0 && |
if (op >= OP_STARI && /* Caseless */ |
3013 |
(othercase = UCD_OTHERCASE(fc)) != fc) |
(othercase = UCD_OTHERCASE(fc)) != fc) |
3014 |
oclength = _pcre_ord2utf8(othercase, occhars); |
oclength = _pcre_ord2utf8(othercase, occhars); |
3015 |
else oclength = 0; |
else oclength = 0; |
3037 |
{ |
{ |
3038 |
for (fi = min;; fi++) |
for (fi = min;; fi++) |
3039 |
{ |
{ |
3040 |
RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM22); |
RMATCH(eptr, ecode, offset_top, md, eptrb, RM22); |
3041 |
if (rrc != MATCH_NOMATCH) RRETURN(rrc); |
if (rrc != MATCH_NOMATCH) RRETURN(rrc); |
3042 |
if (fi >= max) MRRETURN(MATCH_NOMATCH); |
if (fi >= max) MRRETURN(MATCH_NOMATCH); |
3043 |
if (eptr <= md->end_subject - length && |
if (eptr <= md->end_subject - length && |
3079 |
|
|
3080 |
for(;;) |
for(;;) |
3081 |
{ |
{ |
3082 |
RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM23); |
RMATCH(eptr, ecode, offset_top, md, eptrb, RM23); |
3083 |
if (rrc != MATCH_NOMATCH) RRETURN(rrc); |
if (rrc != MATCH_NOMATCH) RRETURN(rrc); |
3084 |
if (eptr == pp) { MRRETURN(MATCH_NOMATCH); } |
if (eptr == pp) { MRRETURN(MATCH_NOMATCH); } |
3085 |
#ifdef SUPPORT_UCP |
#ifdef SUPPORT_UCP |
3116 |
DPRINTF(("matching %c{%d,%d} against subject %.*s\n", fc, min, max, |
DPRINTF(("matching %c{%d,%d} against subject %.*s\n", fc, min, max, |
3117 |
max, eptr)); |
max, eptr)); |
3118 |
|
|
3119 |
if ((ims & PCRE_CASELESS) != 0) |
if (op >= OP_STARI) /* Caseless */ |
3120 |
{ |
{ |
3121 |
fc = md->lcc[fc]; |
fc = md->lcc[fc]; |
3122 |
for (i = 1; i <= min; i++) |
for (i = 1; i <= min; i++) |
3133 |
{ |
{ |
3134 |
for (fi = min;; fi++) |
for (fi = min;; fi++) |
3135 |
{ |
{ |
3136 |
RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM24); |
RMATCH(eptr, ecode, offset_top, md, eptrb, RM24); |
3137 |
if (rrc != MATCH_NOMATCH) RRETURN(rrc); |
if (rrc != MATCH_NOMATCH) RRETURN(rrc); |
3138 |
if (fi >= max) MRRETURN(MATCH_NOMATCH); |
if (fi >= max) MRRETURN(MATCH_NOMATCH); |
3139 |
if (eptr >= md->end_subject) |
if (eptr >= md->end_subject) |
3163 |
|
|
3164 |
while (eptr >= pp) |
while (eptr >= pp) |
3165 |
{ |
{ |
3166 |
RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM25); |
RMATCH(eptr, ecode, offset_top, md, eptrb, RM25); |
3167 |
eptr--; |
eptr--; |
3168 |
if (rrc != MATCH_NOMATCH) RRETURN(rrc); |
if (rrc != MATCH_NOMATCH) RRETURN(rrc); |
3169 |
} |
} |
3192 |
{ |
{ |
3193 |
for (fi = min;; fi++) |
for (fi = min;; fi++) |
3194 |
{ |
{ |
3195 |
RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM26); |
RMATCH(eptr, ecode, offset_top, md, eptrb, RM26); |
3196 |
if (rrc != MATCH_NOMATCH) RRETURN(rrc); |
if (rrc != MATCH_NOMATCH) RRETURN(rrc); |
3197 |
if (fi >= max) MRRETURN(MATCH_NOMATCH); |
if (fi >= max) MRRETURN(MATCH_NOMATCH); |
3198 |
if (eptr >= md->end_subject) |
if (eptr >= md->end_subject) |
3221 |
|
|
3222 |
while (eptr >= pp) |
while (eptr >= pp) |
3223 |
{ |
{ |
3224 |
RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM27); |
RMATCH(eptr, ecode, offset_top, md, eptrb, RM27); |
3225 |
eptr--; |
eptr--; |
3226 |
if (rrc != MATCH_NOMATCH) RRETURN(rrc); |
if (rrc != MATCH_NOMATCH) RRETURN(rrc); |
3227 |
} |
} |
3234 |
checking can be multibyte. */ |
checking can be multibyte. */ |
3235 |
|
|
3236 |
case OP_NOT: |
case OP_NOT: |
3237 |
|
case OP_NOTI: |
3238 |
if (eptr >= md->end_subject) |
if (eptr >= md->end_subject) |
3239 |
{ |
{ |
3240 |
SCHECK_PARTIAL(); |
SCHECK_PARTIAL(); |
3242 |
} |
} |
3243 |
ecode++; |
ecode++; |
3244 |
GETCHARINCTEST(c, eptr); |
GETCHARINCTEST(c, eptr); |
3245 |
if ((ims & PCRE_CASELESS) != 0) |
if (op == OP_NOTI) /* The caseless case */ |
3246 |
{ |
{ |
3247 |
#ifdef SUPPORT_UTF8 |
#ifdef SUPPORT_UTF8 |
3248 |
if (c < 256) |
if (c < 256) |
3250 |
c = md->lcc[c]; |
c = md->lcc[c]; |
3251 |
if (md->lcc[*ecode++] == c) MRRETURN(MATCH_NOMATCH); |
if (md->lcc[*ecode++] == c) MRRETURN(MATCH_NOMATCH); |
3252 |
} |
} |
3253 |
else |
else /* Caseful */ |
3254 |
{ |
{ |
3255 |
if (*ecode++ == c) MRRETURN(MATCH_NOMATCH); |
if (*ecode++ == c) MRRETURN(MATCH_NOMATCH); |
3256 |
} |
} |
3264 |
about... */ |
about... */ |
3265 |
|
|
3266 |
case OP_NOTEXACT: |
case OP_NOTEXACT: |
3267 |
|
case OP_NOTEXACTI: |
3268 |
min = max = GET2(ecode, 1); |
min = max = GET2(ecode, 1); |
3269 |
ecode += 3; |
ecode += 3; |
3270 |
goto REPEATNOTCHAR; |
goto REPEATNOTCHAR; |
3271 |
|
|
3272 |
case OP_NOTUPTO: |
case OP_NOTUPTO: |
3273 |
|
case OP_NOTUPTOI: |
3274 |
case OP_NOTMINUPTO: |
case OP_NOTMINUPTO: |
3275 |
|
case OP_NOTMINUPTOI: |
3276 |
min = 0; |
min = 0; |
3277 |
max = GET2(ecode, 1); |
max = GET2(ecode, 1); |
3278 |
minimize = *ecode == OP_NOTMINUPTO; |
minimize = *ecode == OP_NOTMINUPTO || *ecode == OP_NOTMINUPTOI; |
3279 |
ecode += 3; |
ecode += 3; |
3280 |
goto REPEATNOTCHAR; |
goto REPEATNOTCHAR; |
3281 |
|
|
3282 |
case OP_NOTPOSSTAR: |
case OP_NOTPOSSTAR: |
3283 |
|
case OP_NOTPOSSTARI: |
3284 |
possessive = TRUE; |
possessive = TRUE; |
3285 |
min = 0; |
min = 0; |
3286 |
max = INT_MAX; |
max = INT_MAX; |
3288 |
goto REPEATNOTCHAR; |
goto REPEATNOTCHAR; |
3289 |
|
|
3290 |
case OP_NOTPOSPLUS: |
case OP_NOTPOSPLUS: |
3291 |
|
case OP_NOTPOSPLUSI: |
3292 |
possessive = TRUE; |
possessive = TRUE; |
3293 |
min = 1; |
min = 1; |
3294 |
max = INT_MAX; |
max = INT_MAX; |
3296 |
goto REPEATNOTCHAR; |
goto REPEATNOTCHAR; |
3297 |
|
|
3298 |
case OP_NOTPOSQUERY: |
case OP_NOTPOSQUERY: |
3299 |
|
case OP_NOTPOSQUERYI: |
3300 |
possessive = TRUE; |
possessive = TRUE; |
3301 |
min = 0; |
min = 0; |
3302 |
max = 1; |
max = 1; |
3304 |
goto REPEATNOTCHAR; |
goto REPEATNOTCHAR; |
3305 |
|
|
3306 |
case OP_NOTPOSUPTO: |
case OP_NOTPOSUPTO: |
3307 |
|
case OP_NOTPOSUPTOI: |
3308 |
possessive = TRUE; |
possessive = TRUE; |
3309 |
min = 0; |
min = 0; |
3310 |
max = GET2(ecode, 1); |
max = GET2(ecode, 1); |
3312 |
goto REPEATNOTCHAR; |
goto REPEATNOTCHAR; |
3313 |
|
|
3314 |
case OP_NOTSTAR: |
case OP_NOTSTAR: |
3315 |
|
case OP_NOTSTARI: |
3316 |
case OP_NOTMINSTAR: |
case OP_NOTMINSTAR: |
3317 |
|
case OP_NOTMINSTARI: |
3318 |
case OP_NOTPLUS: |
case OP_NOTPLUS: |
3319 |
|
case OP_NOTPLUSI: |
3320 |
case OP_NOTMINPLUS: |
case OP_NOTMINPLUS: |
3321 |
|
case OP_NOTMINPLUSI: |
3322 |
case OP_NOTQUERY: |
case OP_NOTQUERY: |
3323 |
|
case OP_NOTQUERYI: |
3324 |
case OP_NOTMINQUERY: |
case OP_NOTMINQUERY: |
3325 |
c = *ecode++ - OP_NOTSTAR; |
case OP_NOTMINQUERYI: |
3326 |
|
c = *ecode++ - ((op >= OP_NOTSTARI)? OP_NOTSTARI: OP_NOTSTAR); |
3327 |
minimize = (c & 1) != 0; |
minimize = (c & 1) != 0; |
3328 |
min = rep_min[c]; /* Pick up values from tables; */ |
min = rep_min[c]; /* Pick up values from tables; */ |
3329 |
max = rep_max[c]; /* zero for max => infinity */ |
max = rep_max[c]; /* zero for max => infinity */ |
3345 |
DPRINTF(("negative matching %c{%d,%d} against subject %.*s\n", fc, min, max, |
DPRINTF(("negative matching %c{%d,%d} against subject %.*s\n", fc, min, max, |
3346 |
max, eptr)); |
max, eptr)); |
3347 |
|
|
3348 |
if ((ims & PCRE_CASELESS) != 0) |
if (op >= OP_NOTSTARI) /* Caseless */ |
3349 |
{ |
{ |
3350 |
fc = md->lcc[fc]; |
fc = md->lcc[fc]; |
3351 |
|
|
3393 |
register unsigned int d; |
register unsigned int d; |
3394 |
for (fi = min;; fi++) |
for (fi = min;; fi++) |
3395 |
{ |
{ |
3396 |
RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM28); |
RMATCH(eptr, ecode, offset_top, md, eptrb, RM28); |
3397 |
if (rrc != MATCH_NOMATCH) RRETURN(rrc); |
if (rrc != MATCH_NOMATCH) RRETURN(rrc); |
3398 |
if (fi >= max) MRRETURN(MATCH_NOMATCH); |
if (fi >= max) MRRETURN(MATCH_NOMATCH); |
3399 |
if (eptr >= md->end_subject) |
if (eptr >= md->end_subject) |
3412 |
{ |
{ |
3413 |
for (fi = min;; fi++) |
for (fi = min;; fi++) |
3414 |
{ |
{ |
3415 |
RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM29); |
RMATCH(eptr, ecode, offset_top, md, eptrb, RM29); |
3416 |
if (rrc != MATCH_NOMATCH) RRETURN(rrc); |
if (rrc != MATCH_NOMATCH) RRETURN(rrc); |
3417 |
if (fi >= max) MRRETURN(MATCH_NOMATCH); |
if (fi >= max) MRRETURN(MATCH_NOMATCH); |
3418 |
if (eptr >= md->end_subject) |
if (eptr >= md->end_subject) |
3453 |
if (possessive) continue; |
if (possessive) continue; |
3454 |
for(;;) |
for(;;) |
3455 |
{ |
{ |
3456 |
RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM30); |
RMATCH(eptr, ecode, offset_top, md, eptrb, RM30); |
3457 |
if (rrc != MATCH_NOMATCH) RRETURN(rrc); |
if (rrc != MATCH_NOMATCH) RRETURN(rrc); |
3458 |
if (eptr-- == pp) break; /* Stop if tried at original pos */ |
if (eptr-- == pp) break; /* Stop if tried at original pos */ |
3459 |
BACKCHAR(eptr); |
BACKCHAR(eptr); |
3476 |
if (possessive) continue; |
if (possessive) continue; |
3477 |
while (eptr >= pp) |
while (eptr >= pp) |
3478 |
{ |
{ |
3479 |
RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM31); |
RMATCH(eptr, ecode, offset_top, md, eptrb, RM31); |
3480 |
if (rrc != MATCH_NOMATCH) RRETURN(rrc); |
if (rrc != MATCH_NOMATCH) RRETURN(rrc); |
3481 |
eptr--; |
eptr--; |
3482 |
} |
} |
3533 |
register unsigned int d; |
register unsigned int d; |
3534 |
for (fi = min;; fi++) |
for (fi = min;; fi++) |
3535 |
{ |
{ |
3536 |
RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM32); |
RMATCH(eptr, ecode, offset_top, md, eptrb, RM32); |
3537 |
if (rrc != MATCH_NOMATCH) RRETURN(rrc); |
if (rrc != MATCH_NOMATCH) RRETURN(rrc); |
3538 |
if (fi >= max) MRRETURN(MATCH_NOMATCH); |
if (fi >= max) MRRETURN(MATCH_NOMATCH); |
3539 |
if (eptr >= md->end_subject) |
if (eptr >= md->end_subject) |
3551 |
{ |
{ |
3552 |
for (fi = min;; fi++) |
for (fi = min;; fi++) |
3553 |
{ |
{ |
3554 |
RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM33); |
RMATCH(eptr, ecode, offset_top, md, eptrb, RM33); |
3555 |
if (rrc != MATCH_NOMATCH) RRETURN(rrc); |
if (rrc != MATCH_NOMATCH) RRETURN(rrc); |
3556 |
if (fi >= max) MRRETURN(MATCH_NOMATCH); |
if (fi >= max) MRRETURN(MATCH_NOMATCH); |
3557 |
if (eptr >= md->end_subject) |
if (eptr >= md->end_subject) |
3591 |
if (possessive) continue; |
if (possessive) continue; |
3592 |
for(;;) |
for(;;) |
3593 |
{ |
{ |
3594 |
RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM34); |
RMATCH(eptr, ecode, offset_top, md, eptrb, RM34); |
3595 |
if (rrc != MATCH_NOMATCH) RRETURN(rrc); |
if (rrc != MATCH_NOMATCH) RRETURN(rrc); |
3596 |
if (eptr-- == pp) break; /* Stop if tried at original pos */ |
if (eptr-- == pp) break; /* Stop if tried at original pos */ |
3597 |
BACKCHAR(eptr); |
BACKCHAR(eptr); |
3614 |
if (possessive) continue; |
if (possessive) continue; |
3615 |
while (eptr >= pp) |
while (eptr >= pp) |
3616 |
{ |
{ |
3617 |
RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM35); |
RMATCH(eptr, ecode, offset_top, md, eptrb, RM35); |
3618 |
if (rrc != MATCH_NOMATCH) RRETURN(rrc); |
if (rrc != MATCH_NOMATCH) RRETURN(rrc); |
3619 |
eptr--; |
eptr--; |
3620 |
} |
} |
3729 |
case PT_LAMP: |
case PT_LAMP: |
3730 |
for (i = 1; i <= min; i++) |
for (i = 1; i <= min; i++) |
3731 |
{ |
{ |
3732 |
|
int chartype; |
3733 |
if (eptr >= md->end_subject) |
if (eptr >= md->end_subject) |
3734 |
{ |
{ |
3735 |
SCHECK_PARTIAL(); |
SCHECK_PARTIAL(); |
3736 |
MRRETURN(MATCH_NOMATCH); |
MRRETURN(MATCH_NOMATCH); |
3737 |
} |
} |
3738 |
GETCHARINCTEST(c, eptr); |
GETCHARINCTEST(c, eptr); |
3739 |
prop_chartype = UCD_CHARTYPE(c); |
chartype = UCD_CHARTYPE(c); |
3740 |
if ((prop_chartype == ucp_Lu || |
if ((chartype == ucp_Lu || |
3741 |
prop_chartype == ucp_Ll || |
chartype == ucp_Ll || |
3742 |
prop_chartype == ucp_Lt) == prop_fail_result) |
chartype == ucp_Lt) == prop_fail_result) |
3743 |
MRRETURN(MATCH_NOMATCH); |
MRRETURN(MATCH_NOMATCH); |
3744 |
} |
} |
3745 |
break; |
break; |
3753 |
MRRETURN(MATCH_NOMATCH); |
MRRETURN(MATCH_NOMATCH); |
3754 |
} |
} |
3755 |
GETCHARINCTEST(c, eptr); |
GETCHARINCTEST(c, eptr); |
3756 |
prop_category = UCD_CATEGORY(c); |
if ((UCD_CATEGORY(c) == prop_value) == prop_fail_result) |
|
if ((prop_category == prop_value) == prop_fail_result) |
|
3757 |
MRRETURN(MATCH_NOMATCH); |
MRRETURN(MATCH_NOMATCH); |
3758 |
} |
} |
3759 |
break; |
break; |
3767 |
MRRETURN(MATCH_NOMATCH); |
MRRETURN(MATCH_NOMATCH); |
3768 |
} |
} |
3769 |
GETCHARINCTEST(c, eptr); |
GETCHARINCTEST(c, eptr); |
3770 |
prop_chartype = UCD_CHARTYPE(c); |
if ((UCD_CHARTYPE(c) == prop_value) == prop_fail_result) |
|
if ((prop_chartype == prop_value) == prop_fail_result) |
|
3771 |
MRRETURN(MATCH_NOMATCH); |
MRRETURN(MATCH_NOMATCH); |
3772 |
} |
} |
3773 |
break; |
break; |
3781 |
MRRETURN(MATCH_NOMATCH); |
MRRETURN(MATCH_NOMATCH); |
3782 |
} |
} |
3783 |
GETCHARINCTEST(c, eptr); |
GETCHARINCTEST(c, eptr); |
3784 |
prop_script = UCD_SCRIPT(c); |
if ((UCD_SCRIPT(c) == prop_value) == prop_fail_result) |
|
if ((prop_script == prop_value) == prop_fail_result) |
|
3785 |
MRRETURN(MATCH_NOMATCH); |
MRRETURN(MATCH_NOMATCH); |
3786 |
} |
} |
3787 |
break; |
break; |
3789 |
case PT_ALNUM: |
case PT_ALNUM: |
3790 |
for (i = 1; i <= min; i++) |
for (i = 1; i <= min; i++) |
3791 |
{ |
{ |
3792 |
|
int category; |
3793 |
if (eptr >= md->end_subject) |
if (eptr >= md->end_subject) |
3794 |
{ |
{ |
3795 |
SCHECK_PARTIAL(); |
SCHECK_PARTIAL(); |
3796 |
MRRETURN(MATCH_NOMATCH); |
MRRETURN(MATCH_NOMATCH); |
3797 |
} |
} |
3798 |
GETCHARINCTEST(c, eptr); |
GETCHARINCTEST(c, eptr); |
3799 |
prop_category = UCD_CATEGORY(c); |
category = UCD_CATEGORY(c); |
3800 |
if ((prop_category == ucp_L || prop_category == ucp_N) |
if ((category == ucp_L || category == ucp_N) == prop_fail_result) |
|
== prop_fail_result) |
|
3801 |
MRRETURN(MATCH_NOMATCH); |
MRRETURN(MATCH_NOMATCH); |
3802 |
} |
} |
3803 |
break; |
break; |
3811 |
MRRETURN(MATCH_NOMATCH); |
MRRETURN(MATCH_NOMATCH); |
3812 |
} |
} |
3813 |
GETCHARINCTEST(c, eptr); |
GETCHARINCTEST(c, eptr); |
3814 |
prop_category = UCD_CATEGORY(c); |
if ((UCD_CATEGORY(c) == ucp_Z || c == CHAR_HT || c == CHAR_NL || |
|
if ((prop_category == ucp_Z || c == CHAR_HT || c == CHAR_NL || |
|
3815 |
c == CHAR_FF || c == CHAR_CR) |
c == CHAR_FF || c == CHAR_CR) |
3816 |
== prop_fail_result) |
== prop_fail_result) |
3817 |
MRRETURN(MATCH_NOMATCH); |
MRRETURN(MATCH_NOMATCH); |
3827 |
MRRETURN(MATCH_NOMATCH); |
MRRETURN(MATCH_NOMATCH); |
3828 |
} |
} |
3829 |
GETCHARINCTEST(c, eptr); |
GETCHARINCTEST(c, eptr); |
3830 |
prop_category = UCD_CATEGORY(c); |
if ((UCD_CATEGORY(c) == ucp_Z || c == CHAR_HT || c == CHAR_NL || |
|
if ((prop_category == ucp_Z || c == CHAR_HT || c == CHAR_NL || |
|
3831 |
c == CHAR_VT || c == CHAR_FF || c == CHAR_CR) |
c == CHAR_VT || c == CHAR_FF || c == CHAR_CR) |
3832 |
== prop_fail_result) |
== prop_fail_result) |
3833 |
MRRETURN(MATCH_NOMATCH); |
MRRETURN(MATCH_NOMATCH); |
3837 |
case PT_WORD: |
case PT_WORD: |
3838 |
for (i = 1; i <= min; i++) |
for (i = 1; i <= min; i++) |
3839 |
{ |
{ |
3840 |
|
int category; |
3841 |
if (eptr >= md->end_subject) |
if (eptr >= md->end_subject) |
3842 |
{ |
{ |
3843 |
SCHECK_PARTIAL(); |
SCHECK_PARTIAL(); |
3844 |
MRRETURN(MATCH_NOMATCH); |
MRRETURN(MATCH_NOMATCH); |
3845 |
} |
} |
3846 |
GETCHARINCTEST(c, eptr); |
GETCHARINCTEST(c, eptr); |
3847 |
prop_category = UCD_CATEGORY(c); |
category = UCD_CATEGORY(c); |
3848 |
if ((prop_category == ucp_L || prop_category == ucp_N || |
if ((category == ucp_L || category == ucp_N || c == CHAR_UNDERSCORE) |
|
c == CHAR_UNDERSCORE) |
|
3849 |
== prop_fail_result) |
== prop_fail_result) |
3850 |
MRRETURN(MATCH_NOMATCH); |
MRRETURN(MATCH_NOMATCH); |
3851 |
} |
} |
3871 |
MRRETURN(MATCH_NOMATCH); |
MRRETURN(MATCH_NOMATCH); |
3872 |
} |
} |
3873 |
GETCHARINCTEST(c, eptr); |
GETCHARINCTEST(c, eptr); |
3874 |
prop_category = UCD_CATEGORY(c); |
if (UCD_CATEGORY(c) == ucp_M) MRRETURN(MATCH_NOMATCH); |
|
if (prop_category == ucp_M) MRRETURN(MATCH_NOMATCH); |
|
3875 |
while (eptr < md->end_subject) |
while (eptr < md->end_subject) |
3876 |
{ |
{ |
3877 |
int len = 1; |
int len = 1; |
3878 |
if (!utf8) c = *eptr; |
if (!utf8) c = *eptr; else { GETCHARLEN(c, eptr, len); } |
3879 |
else { GETCHARLEN(c, eptr, len); } |
if (UCD_CATEGORY(c) != ucp_M) break; |
|
prop_category = UCD_CATEGORY(c); |
|
|
if (prop_category != ucp_M) break; |
|
3880 |
eptr += len; |
eptr += len; |
3881 |
} |
} |
3882 |
} |
} |
3934 |
switch(c) |
switch(c) |
3935 |
{ |
{ |
3936 |
default: MRRETURN(MATCH_NOMATCH); |
default: MRRETURN(MATCH_NOMATCH); |
3937 |
|
|
3938 |
case 0x000d: |
case 0x000d: |
3939 |
if (eptr < md->end_subject && *eptr == 0x0a) eptr++; |
if (eptr < md->end_subject && *eptr == 0x0a) eptr++; |
3940 |
break; |
break; |
4211 |
switch(*eptr++) |
switch(*eptr++) |
4212 |
{ |
{ |
4213 |
default: MRRETURN(MATCH_NOMATCH); |
default: MRRETURN(MATCH_NOMATCH); |
4214 |
|
|
4215 |
case 0x000d: |
case 0x000d: |
4216 |
if (eptr < md->end_subject && *eptr == 0x0a) eptr++; |
if (eptr < md->end_subject && *eptr == 0x0a) eptr++; |
4217 |
break; |
break; |
4218 |
|
|
4219 |
case 0x000a: |
case 0x000a: |
4220 |
break; |
break; |
4221 |
|
|
4405 |
case PT_ANY: |
case PT_ANY: |
4406 |
for (fi = min;; fi++) |
for (fi = min;; fi++) |
4407 |
{ |
{ |
4408 |
RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM36); |
RMATCH(eptr, ecode, offset_top, md, eptrb, RM36); |
4409 |
if (rrc != MATCH_NOMATCH) RRETURN(rrc); |
if (rrc != MATCH_NOMATCH) RRETURN(rrc); |
4410 |
if (fi >= max) MRRETURN(MATCH_NOMATCH); |
if (fi >= max) MRRETURN(MATCH_NOMATCH); |
4411 |
if (eptr >= md->end_subject) |
if (eptr >= md->end_subject) |
4421 |
case PT_LAMP: |
case PT_LAMP: |
4422 |
for (fi = min;; fi++) |
for (fi = min;; fi++) |
4423 |
{ |
{ |
4424 |
RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM37); |
int chartype; |
4425 |
|
RMATCH(eptr, ecode, offset_top, md, eptrb, RM37); |
4426 |
if (rrc != MATCH_NOMATCH) RRETURN(rrc); |
if (rrc != MATCH_NOMATCH) RRETURN(rrc); |
4427 |
if (fi >= max) MRRETURN(MATCH_NOMATCH); |
if (fi >= max) MRRETURN(MATCH_NOMATCH); |
4428 |
if (eptr >= md->end_subject) |
if (eptr >= md->end_subject) |
4431 |
MRRETURN(MATCH_NOMATCH); |
MRRETURN(MATCH_NOMATCH); |
4432 |
} |
} |
4433 |
GETCHARINCTEST(c, eptr); |
GETCHARINCTEST(c, eptr); |
4434 |
prop_chartype = UCD_CHARTYPE(c); |
chartype = UCD_CHARTYPE(c); |
4435 |
if ((prop_chartype == ucp_Lu || |
if ((chartype == ucp_Lu || |
4436 |
prop_chartype == ucp_Ll || |
chartype == ucp_Ll || |
4437 |
prop_chartype == ucp_Lt) == prop_fail_result) |
chartype == ucp_Lt) == prop_fail_result) |
4438 |
MRRETURN(MATCH_NOMATCH); |
MRRETURN(MATCH_NOMATCH); |
4439 |
} |
} |
4440 |
/* Control never gets here */ |
/* Control never gets here */ |
4442 |
case PT_GC: |
case PT_GC: |
4443 |
for (fi = min;; fi++) |
for (fi = min;; fi++) |
4444 |
{ |
{ |
4445 |
RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM38); |
RMATCH(eptr, ecode, offset_top, md, eptrb, RM38); |
4446 |
if (rrc != MATCH_NOMATCH) RRETURN(rrc); |
if (rrc != MATCH_NOMATCH) RRETURN(rrc); |
4447 |
if (fi >= max) MRRETURN(MATCH_NOMATCH); |
if (fi >= max) MRRETURN(MATCH_NOMATCH); |
4448 |
if (eptr >= md->end_subject) |
if (eptr >= md->end_subject) |
4451 |
MRRETURN(MATCH_NOMATCH); |
MRRETURN(MATCH_NOMATCH); |
4452 |
} |
} |
4453 |
GETCHARINCTEST(c, eptr); |
GETCHARINCTEST(c, eptr); |
4454 |
prop_category = UCD_CATEGORY(c); |
if ((UCD_CATEGORY(c) == prop_value) == prop_fail_result) |
|
if ((prop_category == prop_value) == prop_fail_result) |
|
4455 |
MRRETURN(MATCH_NOMATCH); |
MRRETURN(MATCH_NOMATCH); |
4456 |
} |
} |
4457 |
/* Control never gets here */ |
/* Control never gets here */ |
4459 |
case PT_PC: |
case PT_PC: |
4460 |
for (fi = min;; fi++) |
for (fi = min;; fi++) |
4461 |
{ |
{ |
4462 |
RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM39); |
RMATCH(eptr, ecode, offset_top, md, eptrb, RM39); |
4463 |
if (rrc != MATCH_NOMATCH) RRETURN(rrc); |
if (rrc != MATCH_NOMATCH) RRETURN(rrc); |
4464 |
if (fi >= max) MRRETURN(MATCH_NOMATCH); |
if (fi >= max) MRRETURN(MATCH_NOMATCH); |
4465 |
if (eptr >= md->end_subject) |
if (eptr >= md->end_subject) |
4468 |
MRRETURN(MATCH_NOMATCH); |
MRRETURN(MATCH_NOMATCH); |
4469 |
} |
} |
4470 |
GETCHARINCTEST(c, eptr); |
GETCHARINCTEST(c, eptr); |
4471 |
prop_chartype = UCD_CHARTYPE(c); |
if ((UCD_CHARTYPE(c) == prop_value) == prop_fail_result) |
|
if ((prop_chartype == prop_value) == prop_fail_result) |
|
4472 |
MRRETURN(MATCH_NOMATCH); |
MRRETURN(MATCH_NOMATCH); |
4473 |
} |
} |
4474 |
/* Control never gets here */ |
/* Control never gets here */ |
4476 |
case PT_SC: |
case PT_SC: |
4477 |
for (fi = min;; fi++) |
for (fi = min;; fi++) |
4478 |
{ |
{ |
4479 |
RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM40); |
RMATCH(eptr, ecode, offset_top, md, eptrb, RM40); |
4480 |
if (rrc != MATCH_NOMATCH) RRETURN(rrc); |
if (rrc != MATCH_NOMATCH) RRETURN(rrc); |
4481 |
if (fi >= max) MRRETURN(MATCH_NOMATCH); |
if (fi >= max) MRRETURN(MATCH_NOMATCH); |
4482 |
if (eptr >= md->end_subject) |
if (eptr >= md->end_subject) |
4485 |
MRRETURN(MATCH_NOMATCH); |
MRRETURN(MATCH_NOMATCH); |
4486 |
} |
} |
4487 |
GETCHARINCTEST(c, eptr); |
GETCHARINCTEST(c, eptr); |
4488 |
prop_script = UCD_SCRIPT(c); |
if ((UCD_SCRIPT(c) == prop_value) == prop_fail_result) |
|
if ((prop_script == prop_value) == prop_fail_result) |
|
4489 |
MRRETURN(MATCH_NOMATCH); |
MRRETURN(MATCH_NOMATCH); |
4490 |
} |
} |
4491 |
/* Control never gets here */ |
/* Control never gets here */ |
4493 |
case PT_ALNUM: |
case PT_ALNUM: |
4494 |
for (fi = min;; fi++) |
for (fi = min;; fi++) |
4495 |
{ |
{ |
4496 |
RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM59); |
int category; |
4497 |
|
RMATCH(eptr, ecode, offset_top, md, eptrb, RM59); |
4498 |
if (rrc != MATCH_NOMATCH) RRETURN(rrc); |
if (rrc != MATCH_NOMATCH) RRETURN(rrc); |
4499 |
if (fi >= max) MRRETURN(MATCH_NOMATCH); |
if (fi >= max) MRRETURN(MATCH_NOMATCH); |
4500 |
if (eptr >= md->end_subject) |
if (eptr >= md->end_subject) |
4503 |
MRRETURN(MATCH_NOMATCH); |
MRRETURN(MATCH_NOMATCH); |
4504 |
} |
} |
4505 |
GETCHARINCTEST(c, eptr); |
GETCHARINCTEST(c, eptr); |
4506 |
prop_category = UCD_CATEGORY(c); |
category = UCD_CATEGORY(c); |
4507 |
if ((prop_category == ucp_L || prop_category == ucp_N) |
if ((category == ucp_L || category == ucp_N) == prop_fail_result) |
|
== prop_fail_result) |
|
4508 |
MRRETURN(MATCH_NOMATCH); |
MRRETURN(MATCH_NOMATCH); |
4509 |
} |
} |
4510 |
/* Control never gets here */ |
/* Control never gets here */ |
4512 |
case PT_SPACE: /* Perl space */ |
case PT_SPACE: /* Perl space */ |
4513 |
for (fi = min;; fi++) |
for (fi = min;; fi++) |
4514 |
{ |
{ |
4515 |
RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM60); |
RMATCH(eptr, ecode, offset_top, md, eptrb, RM60); |
4516 |
if (rrc != MATCH_NOMATCH) RRETURN(rrc); |
if (rrc != MATCH_NOMATCH) RRETURN(rrc); |
4517 |
if (fi >= max) MRRETURN(MATCH_NOMATCH); |
if (fi >= max) MRRETURN(MATCH_NOMATCH); |
4518 |
if (eptr >= md->end_subject) |
if (eptr >= md->end_subject) |
4521 |
MRRETURN(MATCH_NOMATCH); |
MRRETURN(MATCH_NOMATCH); |
4522 |
} |
} |
4523 |
GETCHARINCTEST(c, eptr); |
GETCHARINCTEST(c, eptr); |
4524 |
prop_category = UCD_CATEGORY(c); |
if ((UCD_CATEGORY(c) == ucp_Z || c == CHAR_HT || c == CHAR_NL || |
|
if ((prop_category == ucp_Z || c == CHAR_HT || c == CHAR_NL || |
|
4525 |
c == CHAR_FF || c == CHAR_CR) |
c == CHAR_FF || c == CHAR_CR) |
4526 |
== prop_fail_result) |
== prop_fail_result) |
4527 |
MRRETURN(MATCH_NOMATCH); |
MRRETURN(MATCH_NOMATCH); |
4531 |
case PT_PXSPACE: /* POSIX space */ |
case PT_PXSPACE: /* POSIX space */ |
4532 |
for (fi = min;; fi++) |
for (fi = min;; fi++) |
4533 |
{ |
{ |
4534 |
RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM61); |
RMATCH(eptr, ecode, offset_top, md, eptrb, RM61); |
4535 |
if (rrc != MATCH_NOMATCH) RRETURN(rrc); |
if (rrc != MATCH_NOMATCH) RRETURN(rrc); |
4536 |
if (fi >= max) MRRETURN(MATCH_NOMATCH); |
if (fi >= max) MRRETURN(MATCH_NOMATCH); |
4537 |
if (eptr >= md->end_subject) |
if (eptr >= md->end_subject) |
4540 |
MRRETURN(MATCH_NOMATCH); |
MRRETURN(MATCH_NOMATCH); |
4541 |
} |
} |
4542 |
GETCHARINCTEST(c, eptr); |
GETCHARINCTEST(c, eptr); |
4543 |
prop_category = UCD_CATEGORY(c); |
if ((UCD_CATEGORY(c) == ucp_Z || c == CHAR_HT || c == CHAR_NL || |
|
if ((prop_category == ucp_Z || c == CHAR_HT || c == CHAR_NL || |
|
4544 |
c == CHAR_VT || c == CHAR_FF || c == CHAR_CR) |
c == CHAR_VT || c == CHAR_FF || c == CHAR_CR) |
4545 |
== prop_fail_result) |
== prop_fail_result) |
4546 |
MRRETURN(MATCH_NOMATCH); |
MRRETURN(MATCH_NOMATCH); |
4550 |
case PT_WORD: |
case PT_WORD: |
4551 |
for (fi = min;; fi++) |
for (fi = min;; fi++) |
4552 |
{ |
{ |
4553 |
RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM62); |
int category; |
4554 |
|
RMATCH(eptr, ecode, offset_top, md, eptrb, RM62); |
4555 |
if (rrc != MATCH_NOMATCH) RRETURN(rrc); |
if (rrc != MATCH_NOMATCH) RRETURN(rrc); |
4556 |
if (fi >= max) MRRETURN(MATCH_NOMATCH); |
if (fi >= max) MRRETURN(MATCH_NOMATCH); |
4557 |
if (eptr >= md->end_subject) |
if (eptr >= md->end_subject) |
4560 |
MRRETURN(MATCH_NOMATCH); |
MRRETURN(MATCH_NOMATCH); |
4561 |
} |
} |
4562 |
GETCHARINCTEST(c, eptr); |
GETCHARINCTEST(c, eptr); |
4563 |
prop_category = UCD_CATEGORY(c); |
category = UCD_CATEGORY(c); |
4564 |
if ((prop_category == ucp_L || |
if ((category == ucp_L || |
4565 |
prop_category == ucp_N || |
category == ucp_N || |
4566 |
c == CHAR_UNDERSCORE) |
c == CHAR_UNDERSCORE) |
4567 |
== prop_fail_result) |
== prop_fail_result) |
4568 |
MRRETURN(MATCH_NOMATCH); |
MRRETURN(MATCH_NOMATCH); |
4583 |
{ |
{ |
4584 |
for (fi = min;; fi++) |
for (fi = min;; fi++) |
4585 |
{ |
{ |
4586 |
RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM41); |
RMATCH(eptr, ecode, offset_top, md, eptrb, RM41); |
4587 |
if (rrc != MATCH_NOMATCH) RRETURN(rrc); |
if (rrc != MATCH_NOMATCH) RRETURN(rrc); |
4588 |
if (fi >= max) MRRETURN(MATCH_NOMATCH); |
if (fi >= max) MRRETURN(MATCH_NOMATCH); |
4589 |
if (eptr >= md->end_subject) |
if (eptr >= md->end_subject) |
4592 |
MRRETURN(MATCH_NOMATCH); |
MRRETURN(MATCH_NOMATCH); |
4593 |
} |
} |
4594 |
GETCHARINCTEST(c, eptr); |
GETCHARINCTEST(c, eptr); |
4595 |
prop_category = UCD_CATEGORY(c); |
if (UCD_CATEGORY(c) == ucp_M) MRRETURN(MATCH_NOMATCH); |
|
if (prop_category == ucp_M) MRRETURN(MATCH_NOMATCH); |
|
4596 |
while (eptr < md->end_subject) |
while (eptr < md->end_subject) |
4597 |
{ |
{ |
4598 |
int len = 1; |
int len = 1; |
4599 |
if (!utf8) c = *eptr; |
if (!utf8) c = *eptr; else { GETCHARLEN(c, eptr, len); } |
4600 |
else { GETCHARLEN(c, eptr, len); } |
if (UCD_CATEGORY(c) != ucp_M) break; |
|
prop_category = UCD_CATEGORY(c); |
|
|
if (prop_category != ucp_M) break; |
|
4601 |
eptr += len; |
eptr += len; |
4602 |
} |
} |
4603 |
} |
} |
4604 |
} |
} |
|
|
|
4605 |
else |
else |
4606 |
#endif /* SUPPORT_UCP */ |
#endif /* SUPPORT_UCP */ |
4607 |
|
|
4611 |
{ |
{ |
4612 |
for (fi = min;; fi++) |
for (fi = min;; fi++) |
4613 |
{ |
{ |
4614 |
RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM42); |
RMATCH(eptr, ecode, offset_top, md, eptrb, RM42); |
4615 |
if (rrc != MATCH_NOMATCH) RRETURN(rrc); |
if (rrc != MATCH_NOMATCH) RRETURN(rrc); |
4616 |
if (fi >= max) MRRETURN(MATCH_NOMATCH); |
if (fi >= max) MRRETURN(MATCH_NOMATCH); |
4617 |
if (eptr >= md->end_subject) |
if (eptr >= md->end_subject) |
4774 |
{ |
{ |
4775 |
for (fi = min;; fi++) |
for (fi = min;; fi++) |
4776 |
{ |
{ |
4777 |
RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM43); |
RMATCH(eptr, ecode, offset_top, md, eptrb, RM43); |
4778 |
if (rrc != MATCH_NOMATCH) RRETURN(rrc); |
if (rrc != MATCH_NOMATCH) RRETURN(rrc); |
4779 |
if (fi >= max) MRRETURN(MATCH_NOMATCH); |
if (fi >= max) MRRETURN(MATCH_NOMATCH); |
4780 |
if (eptr >= md->end_subject) |
if (eptr >= md->end_subject) |
4922 |
case PT_LAMP: |
case PT_LAMP: |
4923 |
for (i = min; i < max; i++) |
for (i = min; i < max; i++) |
4924 |
{ |
{ |
4925 |
|
int chartype; |
4926 |
int len = 1; |
int len = 1; |
4927 |
if (eptr >= md->end_subject) |
if (eptr >= md->end_subject) |
4928 |
{ |
{ |
4930 |
break; |
break; |
4931 |
} |
} |
4932 |
GETCHARLENTEST(c, eptr, len); |
GETCHARLENTEST(c, eptr, len); |
4933 |
prop_chartype = UCD_CHARTYPE(c); |
chartype = UCD_CHARTYPE(c); |
4934 |
if ((prop_chartype == ucp_Lu || |
if ((chartype == ucp_Lu || |
4935 |
prop_chartype == ucp_Ll || |
chartype == ucp_Ll || |
4936 |
prop_chartype == ucp_Lt) == prop_fail_result) |
chartype == ucp_Lt) == prop_fail_result) |
4937 |
break; |
break; |
4938 |
eptr+= len; |
eptr+= len; |
4939 |
} |
} |
4949 |
break; |
break; |
4950 |
} |
} |
4951 |
GETCHARLENTEST(c, eptr, len); |
GETCHARLENTEST(c, eptr, len); |
4952 |
prop_category = UCD_CATEGORY(c); |
if ((UCD_CATEGORY(c) == prop_value) == prop_fail_result) break; |
|
if ((prop_category == prop_value) == prop_fail_result) |
|
|
break; |
|
4953 |
eptr+= len; |
eptr+= len; |
4954 |
} |
} |
4955 |
break; |
break; |
4964 |
break; |
break; |
4965 |
} |
} |
4966 |
GETCHARLENTEST(c, eptr, len); |
GETCHARLENTEST(c, eptr, len); |
4967 |
prop_chartype = UCD_CHARTYPE(c); |
if ((UCD_CHARTYPE(c) == prop_value) == prop_fail_result) break; |
|
if ((prop_chartype == prop_value) == prop_fail_result) |
|
|
break; |
|
4968 |
eptr+= len; |
eptr+= len; |
4969 |
} |
} |
4970 |
break; |
break; |
4979 |
break; |
break; |
4980 |
} |
} |
4981 |
GETCHARLENTEST(c, eptr, len); |
GETCHARLENTEST(c, eptr, len); |
4982 |
prop_script = UCD_SCRIPT(c); |
if ((UCD_SCRIPT(c) == prop_value) == prop_fail_result) break; |
|
if ((prop_script == prop_value) == prop_fail_result) |
|
|
break; |
|
4983 |
eptr+= len; |
eptr+= len; |
4984 |
} |
} |
4985 |
break; |
break; |
4987 |
case PT_ALNUM: |
case PT_ALNUM: |
4988 |
for (i = min; i < max; i++) |
for (i = min; i < max; i++) |
4989 |
{ |
{ |
4990 |
|
int category; |
4991 |
int len = 1; |
int len = 1; |
4992 |
if (eptr >= md->end_subject) |
if (eptr >= md->end_subject) |
4993 |
{ |
{ |
4995 |
break; |
break; |
4996 |
} |
} |
4997 |
GETCHARLENTEST(c, eptr, len); |
GETCHARLENTEST(c, eptr, len); |
4998 |
prop_category = UCD_CATEGORY(c); |
category = UCD_CATEGORY(c); |
4999 |
if ((prop_category == ucp_L || prop_category == ucp_N) |
if ((category == ucp_L || category == ucp_N) == prop_fail_result) |
|
== prop_fail_result) |
|
5000 |
break; |
break; |
5001 |
eptr+= len; |
eptr+= len; |
5002 |
} |
} |
5012 |
break; |
break; |
5013 |
} |
} |
5014 |
GETCHARLENTEST(c, eptr, len); |
GETCHARLENTEST(c, eptr, len); |
5015 |
prop_category = UCD_CATEGORY(c); |
if ((UCD_CATEGORY(c) == ucp_Z || c == CHAR_HT || c == CHAR_NL || |
|
if ((prop_category == ucp_Z || c == CHAR_HT || c == CHAR_NL || |
|
5016 |
c == CHAR_FF || c == CHAR_CR) |
c == CHAR_FF || c == CHAR_CR) |
5017 |
== prop_fail_result) |
== prop_fail_result) |
5018 |
break; |
break; |
5030 |
break; |
break; |
5031 |
} |
} |
5032 |
GETCHARLENTEST(c, eptr, len); |
GETCHARLENTEST(c, eptr, len); |
5033 |
prop_category = UCD_CATEGORY(c); |
if ((UCD_CATEGORY(c) == ucp_Z || c == CHAR_HT || c == CHAR_NL || |
|
if ((prop_category == ucp_Z || c == CHAR_HT || c == CHAR_NL || |
|
5034 |
c == CHAR_VT || c == CHAR_FF || c == CHAR_CR) |
c == CHAR_VT || c == CHAR_FF || c == CHAR_CR) |
5035 |
== prop_fail_result) |
== prop_fail_result) |
5036 |
break; |
break; |
5041 |
case PT_WORD: |
case PT_WORD: |
5042 |
for (i = min; i < max; i++) |
for (i = min; i < max; i++) |
5043 |
{ |
{ |
5044 |
|
int category; |
5045 |
int len = 1; |
int len = 1; |
5046 |
if (eptr >= md->end_subject) |
if (eptr >= md->end_subject) |
5047 |
{ |
{ |
5049 |
break; |
break; |
5050 |
} |
} |
5051 |
GETCHARLENTEST(c, eptr, len); |
GETCHARLENTEST(c, eptr, len); |
5052 |
prop_category = UCD_CATEGORY(c); |
category = UCD_CATEGORY(c); |
5053 |
if ((prop_category == ucp_L || prop_category == ucp_N || |
if ((category == ucp_L || category == ucp_N || |
5054 |
c == CHAR_UNDERSCORE) == prop_fail_result) |
c == CHAR_UNDERSCORE) == prop_fail_result) |
5055 |
break; |
break; |
5056 |
eptr+= len; |
eptr+= len; |
5066 |
if (possessive) continue; |
if (possessive) continue; |
5067 |
for(;;) |
for(;;) |
5068 |
{ |
{ |
5069 |
RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM44); |
RMATCH(eptr, ecode, offset_top, md, eptrb, RM44); |
5070 |
if (rrc != MATCH_NOMATCH) RRETURN(rrc); |
if (rrc != MATCH_NOMATCH) RRETURN(rrc); |
5071 |
if (eptr-- == pp) break; /* Stop if tried at original pos */ |
if (eptr-- == pp) break; /* Stop if tried at original pos */ |
5072 |
if (utf8) BACKCHAR(eptr); |
if (utf8) BACKCHAR(eptr); |
5080 |
{ |
{ |
5081 |
for (i = min; i < max; i++) |
for (i = min; i < max; i++) |
5082 |
{ |
{ |
5083 |
|
int len = 1; |
5084 |
if (eptr >= md->end_subject) |
if (eptr >= md->end_subject) |
5085 |
{ |
{ |
5086 |
SCHECK_PARTIAL(); |
SCHECK_PARTIAL(); |
5087 |
break; |
break; |
5088 |
} |
} |
5089 |
GETCHARINCTEST(c, eptr); |
if (!utf8) c = *eptr; else { GETCHARLEN(c, eptr, len); } |
5090 |
prop_category = UCD_CATEGORY(c); |
if (UCD_CATEGORY(c) == ucp_M) break; |
5091 |
if (prop_category == ucp_M) break; |
eptr += len; |
5092 |
while (eptr < md->end_subject) |
while (eptr < md->end_subject) |
5093 |
{ |
{ |
5094 |
int len = 1; |
len = 1; |
5095 |
if (!utf8) c = *eptr; else |
if (!utf8) c = *eptr; else { GETCHARLEN(c, eptr, len); } |
5096 |
{ |
if (UCD_CATEGORY(c) != ucp_M) break; |
|
GETCHARLEN(c, eptr, len); |
|
|
} |
|
|
prop_category = UCD_CATEGORY(c); |
|
|
if (prop_category != ucp_M) break; |
|
5097 |
eptr += len; |
eptr += len; |
5098 |
} |
} |
5099 |
} |
} |
5104 |
|
|
5105 |
for(;;) |
for(;;) |
5106 |
{ |
{ |
5107 |
RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM45); |
RMATCH(eptr, ecode, offset_top, md, eptrb, RM45); |
5108 |
if (rrc != MATCH_NOMATCH) RRETURN(rrc); |
if (rrc != MATCH_NOMATCH) RRETURN(rrc); |
5109 |
if (eptr-- == pp) break; /* Stop if tried at original pos */ |
if (eptr-- == pp) break; /* Stop if tried at original pos */ |
5110 |
for (;;) /* Move back over one extended */ |
for (;;) /* Move back over one extended */ |
5111 |
{ |
{ |
|
int len = 1; |
|
5112 |
if (!utf8) c = *eptr; else |
if (!utf8) c = *eptr; else |
5113 |
{ |
{ |
5114 |
BACKCHAR(eptr); |
BACKCHAR(eptr); |
5115 |
GETCHARLEN(c, eptr, len); |
GETCHAR(c, eptr); |
5116 |
} |
} |
5117 |
prop_category = UCD_CATEGORY(c); |
if (UCD_CATEGORY(c) != ucp_M) break; |
|
if (prop_category != ucp_M) break; |
|
5118 |
eptr--; |
eptr--; |
5119 |
} |
} |
5120 |
} |
} |
5386 |
RRETURN(PCRE_ERROR_INTERNAL); |
RRETURN(PCRE_ERROR_INTERNAL); |
5387 |
} |
} |
5388 |
|
|
5389 |
/* eptr is now past the end of the maximum run */ |
/* eptr is now past the end of the maximum run. If possessive, we are |
5390 |
|
done (no backing up). Otherwise, match at this position; anything other |
5391 |
|
than no match is immediately returned. For nomatch, back up one |
5392 |
|
character, unless we are matching \R and the last thing matched was |
5393 |
|
\r\n, in which case, back up two bytes. */ |
5394 |
|
|
5395 |
if (possessive) continue; |
if (possessive) continue; |
5396 |
for(;;) |
for(;;) |
5397 |
{ |
{ |
5398 |
RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM46); |
RMATCH(eptr, ecode, offset_top, md, eptrb, RM46); |
5399 |
if (rrc != MATCH_NOMATCH) RRETURN(rrc); |
if (rrc != MATCH_NOMATCH) RRETURN(rrc); |
5400 |
if (eptr-- == pp) break; /* Stop if tried at original pos */ |
if (eptr-- == pp) break; /* Stop if tried at original pos */ |
5401 |
BACKCHAR(eptr); |
BACKCHAR(eptr); |
5402 |
|
if (ctype == OP_ANYNL && eptr > pp && *eptr == '\n' && |
5403 |
|
eptr[-1] == '\r') eptr--; |
5404 |
} |
} |
5405 |
} |
} |
5406 |
else |
else |
5599 |
RRETURN(PCRE_ERROR_INTERNAL); |
RRETURN(PCRE_ERROR_INTERNAL); |
5600 |
} |
} |
5601 |
|
|
5602 |
/* eptr is now past the end of the maximum run */ |
/* eptr is now past the end of the maximum run. If possessive, we are |
5603 |
|
done (no backing up). Otherwise, match at this position; anything other |
5604 |
|
than no match is immediately returned. For nomatch, back up one |
5605 |
|
character (byte), unless we are matching \R and the last thing matched |
5606 |
|
was \r\n, in which case, back up two bytes. */ |
5607 |
|
|
5608 |
if (possessive) continue; |
if (possessive) continue; |
5609 |
while (eptr >= pp) |
while (eptr >= pp) |
5610 |
{ |
{ |
5611 |
RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM47); |
RMATCH(eptr, ecode, offset_top, md, eptrb, RM47); |
|
eptr--; |
|
5612 |
if (rrc != MATCH_NOMATCH) RRETURN(rrc); |
if (rrc != MATCH_NOMATCH) RRETURN(rrc); |
5613 |
|
eptr--; |
5614 |
|
if (ctype == OP_ANYNL && eptr > pp && *eptr == '\n' && |
5615 |
|
eptr[-1] == '\r') eptr--; |
5616 |
} |
} |
5617 |
} |
} |
5618 |
|
|
5651 |
LBL( 9) LBL(10) LBL(11) LBL(12) LBL(13) LBL(14) LBL(15) LBL(17) |
LBL( 9) LBL(10) LBL(11) LBL(12) LBL(13) LBL(14) LBL(15) LBL(17) |
5652 |
LBL(19) LBL(24) LBL(25) LBL(26) LBL(27) LBL(29) LBL(31) LBL(33) |
LBL(19) LBL(24) LBL(25) LBL(26) LBL(27) LBL(29) LBL(31) LBL(33) |
5653 |
LBL(35) LBL(43) LBL(47) LBL(48) LBL(49) LBL(50) LBL(51) LBL(52) |
LBL(35) LBL(43) LBL(47) LBL(48) LBL(49) LBL(50) LBL(51) LBL(52) |
5654 |
LBL(53) LBL(54) LBL(55) LBL(56) LBL(57) LBL(58) |
LBL(53) LBL(54) LBL(55) LBL(56) LBL(57) LBL(58) LBL(63) |
5655 |
#ifdef SUPPORT_UTF8 |
#ifdef SUPPORT_UTF8 |
5656 |
LBL(16) LBL(18) LBL(20) LBL(21) LBL(22) LBL(23) LBL(28) LBL(30) |
LBL(16) LBL(18) LBL(20) LBL(21) LBL(22) LBL(23) LBL(28) LBL(30) |
5657 |
LBL(32) LBL(34) LBL(42) LBL(46) |
LBL(32) LBL(34) LBL(42) LBL(46) |
5680 |
#undef ecode |
#undef ecode |
5681 |
#undef mstart |
#undef mstart |
5682 |
#undef offset_top |
#undef offset_top |
|
#undef ims |
|
5683 |
#undef eptrb |
#undef eptrb |
5684 |
#undef flags |
#undef flags |
5685 |
|
|
5697 |
#undef condition |
#undef condition |
5698 |
#undef prev_is_word |
#undef prev_is_word |
5699 |
|
|
|
#undef original_ims |
|
|
|
|
5700 |
#undef ctype |
#undef ctype |
5701 |
#undef length |
#undef length |
5702 |
#undef max |
#undef max |
5753 |
PCRE_SPTR subject, int length, int start_offset, int options, int *offsets, |
PCRE_SPTR subject, int length, int start_offset, int options, int *offsets, |
5754 |
int offsetcount) |
int offsetcount) |
5755 |
{ |
{ |
5756 |
int rc, resetcount, ocount; |
int rc, ocount; |
5757 |
int first_byte = -1; |
int first_byte = -1; |
5758 |
int req_byte = -1; |
int req_byte = -1; |
5759 |
int req_byte2 = -1; |
int req_byte2 = -1; |
5760 |
int newline; |
int newline; |
|
unsigned long int ims; |
|
5761 |
BOOL using_temporary_offsets = FALSE; |
BOOL using_temporary_offsets = FALSE; |
5762 |
BOOL anchored; |
BOOL anchored; |
5763 |
BOOL startline; |
BOOL startline; |
5787 |
if (re == NULL || subject == NULL || |
if (re == NULL || subject == NULL || |
5788 |
(offsets == NULL && offsetcount > 0)) return PCRE_ERROR_NULL; |
(offsets == NULL && offsetcount > 0)) return PCRE_ERROR_NULL; |
5789 |
if (offsetcount < 0) return PCRE_ERROR_BADCOUNT; |
if (offsetcount < 0) return PCRE_ERROR_BADCOUNT; |
5790 |
|
if (start_offset < 0 || start_offset > length) return PCRE_ERROR_BADOFFSET; |
5791 |
|
|
5792 |
/* This information is for finding all the numbers associated with a given |
/* This information is for finding all the numbers associated with a given |
5793 |
name, for condition testing. */ |
name, for condition testing. */ |
5861 |
md->use_ucp = (re->options & PCRE_UCP) != 0; |
md->use_ucp = (re->options & PCRE_UCP) != 0; |
5862 |
md->jscript_compat = (re->options & PCRE_JAVASCRIPT_COMPAT) != 0; |
md->jscript_compat = (re->options & PCRE_JAVASCRIPT_COMPAT) != 0; |
5863 |
|
|
5864 |
|
/* Some options are unpacked into BOOL variables in the hope that testing |
5865 |
|
them will be faster than individual option bits. */ |
5866 |
|
|
5867 |
md->notbol = (options & PCRE_NOTBOL) != 0; |
md->notbol = (options & PCRE_NOTBOL) != 0; |
5868 |
md->noteol = (options & PCRE_NOTEOL) != 0; |
md->noteol = (options & PCRE_NOTEOL) != 0; |
5869 |
md->notempty = (options & PCRE_NOTEMPTY) != 0; |
md->notempty = (options & PCRE_NOTEMPTY) != 0; |
5870 |
md->notempty_atstart = (options & PCRE_NOTEMPTY_ATSTART) != 0; |
md->notempty_atstart = (options & PCRE_NOTEMPTY_ATSTART) != 0; |
5871 |
md->partial = ((options & PCRE_PARTIAL_HARD) != 0)? 2 : |
md->partial = ((options & PCRE_PARTIAL_HARD) != 0)? 2 : |
5872 |
((options & PCRE_PARTIAL_SOFT) != 0)? 1 : 0; |
((options & PCRE_PARTIAL_SOFT) != 0)? 1 : 0; |
5873 |
|
|
5874 |
|
|
5875 |
md->hitend = FALSE; |
md->hitend = FALSE; |
5876 |
md->mark = NULL; /* In case never set */ |
md->mark = NULL; /* In case never set */ |
5877 |
|
|
5953 |
if (md->partial && (re->flags & PCRE_NOPARTIAL) != 0) |
if (md->partial && (re->flags & PCRE_NOPARTIAL) != 0) |
5954 |
return PCRE_ERROR_BADPARTIAL; |
return PCRE_ERROR_BADPARTIAL; |
5955 |
|
|
5956 |
/* Check a UTF-8 string if required. Unfortunately there's no way of passing |
/* Check a UTF-8 string if required. Pass back the character offset and error |
5957 |
back the character offset. */ |
code for an invalid string if a results vector is available. */ |
5958 |
|
|
5959 |
#ifdef SUPPORT_UTF8 |
#ifdef SUPPORT_UTF8 |
5960 |
if (utf8 && (options & PCRE_NO_UTF8_CHECK) == 0) |
if (utf8 && (options & PCRE_NO_UTF8_CHECK) == 0) |
5961 |
{ |
{ |
5962 |
if (_pcre_valid_utf8((USPTR)subject, length) >= 0) |
int erroroffset; |
5963 |
return PCRE_ERROR_BADUTF8; |
int errorcode = _pcre_valid_utf8((USPTR)subject, length, &erroroffset); |
5964 |
if (start_offset > 0 && start_offset < length) |
if (errorcode != 0) |
5965 |
{ |
{ |
5966 |
int tb = ((USPTR)subject)[start_offset]; |
if (offsetcount >= 2) |
|
if (tb > 127) |
|
5967 |
{ |
{ |
5968 |
tb &= 0xc0; |
offsets[0] = erroroffset; |
5969 |
if (tb != 0 && tb != 0xc0) return PCRE_ERROR_BADUTF8_OFFSET; |
offsets[1] = errorcode; |
5970 |
} |
} |
5971 |
|
return (errorcode <= PCRE_UTF8_ERR5 && md->partial > 1)? |
5972 |
|
PCRE_ERROR_SHORTUTF8 : PCRE_ERROR_BADUTF8; |
5973 |
} |
} |
|
} |
|
|
#endif |
|
5974 |
|
|
5975 |
/* The ims options can vary during the matching as a result of the presence |
/* Check that a start_offset points to the start of a UTF-8 character. */ |
|
of (?ims) items in the pattern. They are kept in a local variable so that |
|
|
restoring at the exit of a group is easy. */ |
|
5976 |
|
|
5977 |
ims = re->options & (PCRE_CASELESS|PCRE_MULTILINE|PCRE_DOTALL); |
if (start_offset > 0 && start_offset < length && |
5978 |
|
(((USPTR)subject)[start_offset] & 0xc0) == 0x80) |
5979 |
|
return PCRE_ERROR_BADUTF8_OFFSET; |
5980 |
|
} |
5981 |
|
#endif |
5982 |
|
|
5983 |
/* If the expression has got more back references than the offsets supplied can |
/* If the expression has got more back references than the offsets supplied can |
5984 |
hold, we get a temporary chunk of working store to use during the matching. |
hold, we get a temporary chunk of working store to use during the matching. |
6002 |
md->offset_overflow = FALSE; |
md->offset_overflow = FALSE; |
6003 |
md->capture_last = -1; |
md->capture_last = -1; |
6004 |
|
|
|
/* Compute the minimum number of offsets that we need to reset each time. Doing |
|
|
this makes a huge difference to execution time when there aren't many brackets |
|
|
in the pattern. */ |
|
|
|
|
|
resetcount = 2 + re->top_bracket * 2; |
|
|
if (resetcount > offsetcount) resetcount = ocount; |
|
|
|
|
6005 |
/* Reset the working variable associated with each extraction. These should |
/* Reset the working variable associated with each extraction. These should |
6006 |
never be used unless previously set, but they get saved and restored, and so we |
never be used unless previously set, but they get saved and restored, and so we |
6007 |
initialize them to avoid reading uninitialized locations. */ |
initialize them to avoid reading uninitialized locations. Also, unset the |
6008 |
|
offsets for the matched string. This is really just for tidiness with callouts, |
6009 |
|
in case they inspect these fields. */ |
6010 |
|
|
6011 |
if (md->offset_vector != NULL) |
if (md->offset_vector != NULL) |
6012 |
{ |
{ |
6013 |
register int *iptr = md->offset_vector + ocount; |
register int *iptr = md->offset_vector + ocount; |
6014 |
register int *iend = iptr - resetcount/2 + 1; |
register int *iend = iptr - re->top_bracket; |
6015 |
|
if (iend < md->offset_vector + 2) iend = md->offset_vector + 2; |
6016 |
while (--iptr >= iend) *iptr = -1; |
while (--iptr >= iend) *iptr = -1; |
6017 |
|
md->offset_vector[0] = md->offset_vector[1] = -1; |
6018 |
} |
} |
6019 |
|
|
6020 |
/* 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 |
6048 |
} |
} |
6049 |
|
|
6050 |
|
|
6051 |
|
|
6052 |
|
|
6053 |
/* ==========================================================================*/ |
/* ==========================================================================*/ |
6054 |
|
|
6055 |
/* Loop for handling unanchored repeated matching attempts; for anchored regexs |
/* Loop for handling unanchored repeated matching attempts; for anchored regexs |
6060 |
USPTR save_end_subject = end_subject; |
USPTR save_end_subject = end_subject; |
6061 |
USPTR new_start_match; |
USPTR new_start_match; |
6062 |
|
|
|
/* Reset the maximum number of extractions we might see. */ |
|
|
|
|
|
if (md->offset_vector != NULL) |
|
|
{ |
|
|
register int *iptr = md->offset_vector; |
|
|
register int *iend = iptr + resetcount; |
|
|
while (iptr < iend) *iptr++ = -1; |
|
|
} |
|
|
|
|
6063 |
/* If firstline is TRUE, the start of the match is constrained to the first |
/* If firstline is TRUE, the start of the match is constrained to the first |
6064 |
line of a multiline string. That is, the match must be before or at the first |
line of a multiline string. That is, the match must be before or at the first |
6065 |
newline. Implement this by temporarily adjusting end_subject so that we stop |
newline. Implement this by temporarily adjusting end_subject so that we stop |
6087 |
/* There are some optimizations that avoid running the match if a known |
/* There are some optimizations that avoid running the match if a known |
6088 |
starting point is not found, or if a known later character is not present. |
starting point is not found, or if a known later character is not present. |
6089 |
However, there is an option that disables these, for testing and for ensuring |
However, there is an option that disables these, for testing and for ensuring |
6090 |
that all callouts do actually occur. */ |
that all callouts do actually occur. The option can be set in the regex by |
6091 |
|
(*NO_START_OPT) or passed in match-time options. */ |
6092 |
|
|
6093 |
if ((options & PCRE_NO_START_OPTIMIZE) == 0) |
if (((options | re->options) & PCRE_NO_START_OPTIMIZE) == 0) |
6094 |
{ |
{ |
6095 |
/* Advance to a unique first byte if there is one. */ |
/* Advance to a unique first byte if there is one. */ |
6096 |
|
|
6248 |
md->start_match_ptr = start_match; |
md->start_match_ptr = start_match; |
6249 |
md->start_used_ptr = start_match; |
md->start_used_ptr = start_match; |
6250 |
md->match_call_count = 0; |
md->match_call_count = 0; |
6251 |
rc = match(start_match, md->start_code, start_match, NULL, 2, md, ims, NULL, |
md->match_function_type = 0; |
6252 |
0, 0); |
md->end_offset_top = 0; |
6253 |
|
rc = match(start_match, md->start_code, start_match, NULL, 2, md, NULL, 0); |
6254 |
if (md->hitend && start_partial == NULL) start_partial = md->start_used_ptr; |
if (md->hitend && start_partial == NULL) start_partial = md->start_used_ptr; |
6255 |
|
|
6256 |
switch(rc) |
switch(rc) |
6376 |
|
|
6377 |
rc = md->offset_overflow? 0 : md->end_offset_top/2; |
rc = md->offset_overflow? 0 : md->end_offset_top/2; |
6378 |
|
|
6379 |
|
/* If there is space in the offset vector, set any unused pairs at the end of |
6380 |
|
the pattern to -1 for backwards compatibility. It is documented that this |
6381 |
|
happens. In earlier versions, the whole set of potential capturing offsets |
6382 |
|
was set to -1 each time round the loop, but this is handled differently now. |
6383 |
|
"Gaps" are set to -1 dynamically instead (this fixes a bug). Thus, it is only |
6384 |
|
those at the end that need unsetting here. We can't just unset them all at |
6385 |
|
the start of the whole thing because they may get set in one branch that is |
6386 |
|
not the final matching branch. */ |
6387 |
|
|
6388 |
|
if (md->end_offset_top/2 <= re->top_bracket && offsets != NULL) |
6389 |
|
{ |
6390 |
|
register int *iptr, *iend; |
6391 |
|
int resetcount = 2 + re->top_bracket * 2; |
6392 |
|
if (resetcount > offsetcount) resetcount = ocount; |
6393 |
|
iptr = offsets + md->end_offset_top; |
6394 |
|
iend = offsets + resetcount; |
6395 |
|
while (iptr < iend) *iptr++ = -1; |
6396 |
|
} |
6397 |
|
|
6398 |
/* If there is space, set up the whole thing as substring 0. The value of |
/* If there is space, set up the whole thing as substring 0. The value of |
6399 |
md->start_match_ptr might be modified if \K was encountered on the success |
md->start_match_ptr might be modified if \K was encountered on the success |
6400 |
matching path. */ |
matching path. */ |