/[pcre]/code/trunk/pcre_exec.c
ViewVC logotype

Diff of /code/trunk/pcre_exec.c

Parent Directory Parent Directory | Revision Log Revision Log | View Patch Patch

revision 91 by nigel, Sat Feb 24 21:41:34 2007 UTC revision 227 by ph10, Tue Aug 21 15:00:15 2007 UTC
# Line 6  Line 6 
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) 1997-2006 University of Cambridge             Copyright (c) 1997-2007 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
# Line 42  POSSIBILITY OF SUCH DAMAGE. Line 42  POSSIBILITY OF SUCH DAMAGE.
42  pattern matching using an NFA algorithm, trying to mimic Perl as closely as  pattern matching using an NFA algorithm, trying to mimic Perl as closely as
43  possible. There are also some static supporting functions. */  possible. There are also some static supporting functions. */
44    
45  #define NLBLOCK md           /* The block containing newline information */  #ifdef HAVE_CONFIG_H
46    #include <config.h>
47    #endif
48    
49    #define NLBLOCK md             /* Block containing newline information */
50    #define PSSTART start_subject  /* Field containing processed string start */
51    #define PSEND   end_subject    /* Field containing processed string end */
52    
53  #include "pcre_internal.h"  #include "pcre_internal.h"
54    
55    /* Undefine some potentially clashing cpp symbols */
56    
57  /* Structure for building a chain of data that actually lives on the  #undef min
58  stack, for holding the values of the subject pointer at the start of each  #undef max
 subpattern, so as to detect when an empty string has been matched by a  
 subpattern - to break infinite loops. When NO_RECURSE is set, these blocks  
 are on the heap, not on the stack. */  
   
 typedef struct eptrblock {  
   struct eptrblock *epb_prev;  
   USPTR epb_saved_eptr;  
 } eptrblock;  
59    
60  /* Flag bits for the match() function */  /* Flag bits for the match() function */
61    
62  #define match_condassert   0x01    /* Called to check a condition assertion */  #define match_condassert     0x01  /* Called to check a condition assertion */
63  #define match_isgroup      0x02    /* Set if start of bracketed group */  #define match_cbegroup       0x02  /* Could-be-empty unlimited repeat group */
64    
65  /* Non-error returns from the match() function. Error returns are externally  /* Non-error returns from the match() function. Error returns are externally
66  defined PCRE_ERROR_xxx codes, which are all negative. */  defined PCRE_ERROR_xxx codes, which are all negative. */
# Line 68  defined PCRE_ERROR_xxx codes, which are Line 68  defined PCRE_ERROR_xxx codes, which are
68  #define MATCH_MATCH        1  #define MATCH_MATCH        1
69  #define MATCH_NOMATCH      0  #define MATCH_NOMATCH      0
70    
71    /* Special internal returns from the match() function. Make them sufficiently
72    negative to avoid the external error codes. */
73    
74    #define MATCH_COMMIT       (-999)
75    #define MATCH_PRUNE        (-998)
76    #define MATCH_SKIP         (-997)
77    #define MATCH_THEN         (-996)
78    
79  /* Maximum number of ints of offset to save on the stack for recursive calls.  /* Maximum number of ints of offset to save on the stack for recursive calls.
80  If the offset vector is bigger, malloc is used. This should be a multiple of 3,  If the offset vector is bigger, malloc is used. This should be a multiple of 3,
81  because the offset vector is always a multiple of 3 long. */  because the offset vector is always a multiple of 3 long. */
# Line 101  Returns:     nothing Line 109  Returns:     nothing
109  static void  static void
110  pchars(const uschar *p, int length, BOOL is_subject, match_data *md)  pchars(const uschar *p, int length, BOOL is_subject, match_data *md)
111  {  {
112  int c;  unsigned int c;
113  if (is_subject && length > md->end_subject - p) length = md->end_subject - p;  if (is_subject && length > md->end_subject - p) length = md->end_subject - p;
114  while (length-- > 0)  while (length-- > 0)
115    if (isprint(c = *(p++))) printf("%c", c); else printf("\\x%02x", c);    if (isprint(c = *(p++))) printf("%c", c); else printf("\\x%02x", c);
# Line 186  calls by keeping local variables that ne Line 194  calls by keeping local variables that ne
194  obtained from malloc() instead instead of on the stack. Macros are used to  obtained from malloc() instead instead of on the stack. Macros are used to
195  achieve this so that the actual code doesn't look very different to what it  achieve this so that the actual code doesn't look very different to what it
196  always used to.  always used to.
197    
198    The original heap-recursive code used longjmp(). However, it seems that this
199    can be very slow on some operating systems. Following a suggestion from Stan
200    Switzer, the use of longjmp() has been abolished, at the cost of having to
201    provide a unique number for each call to RMATCH. There is no way of generating
202    a sequence of numbers at compile time in C. I have given them names, to make
203    them stand out more clearly.
204    
205    Crude tests on x86 Linux show a small speedup of around 5-8%. However, on
206    FreeBSD, avoiding longjmp() more than halves the time taken to run the standard
207    tests. Furthermore, not using longjmp() means that local dynamic variables
208    don't have indeterminate values; this has meant that the frame size can be
209    reduced because the result can be "passed back" by straight setting of the
210    variable instead of being passed in the frame.
211  ****************************************************************************  ****************************************************************************
212  ***************************************************************************/  ***************************************************************************/
213    
214    /* Numbers for RMATCH calls. When this list is changed, the code at HEAP_RETURN
215    below must be updated in sync.  */
216    
217    enum { RM1=1, RM2,  RM3,  RM4,  RM5,  RM6,  RM7,  RM8,  RM9,  RM10,
218           RM11,  RM12, RM13, RM14, RM15, RM16, RM17, RM18, RM19, RM20,
219           RM21,  RM22, RM23, RM24, RM25, RM26, RM27, RM28, RM29, RM30,
220           RM31,  RM32, RM33, RM34, RM35, RM36, RM37, RM38, RM39, RM40,
221           RM41,  RM42, RM43, RM44, RM45, RM46, RM47, RM48, RM49, RM50,
222           RM51,  RM52, RM53, RM54 };
223    
224  /* 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
225  versions and production versions. */  versions and production versions. Note that the "rw" argument of RMATCH isn't
226    actuall used in this definition. */
227    
228  #ifndef NO_RECURSE  #ifndef NO_RECURSE
229  #define REGISTER register  #define REGISTER register
230    
231  #ifdef DEBUG  #ifdef DEBUG
232  #define RMATCH(rx,ra,rb,rc,rd,re,rf,rg) \  #define RMATCH(ra,rb,rc,rd,re,rf,rg,rw) \
233    { \    { \
234    printf("match() called in line %d\n", __LINE__); \    printf("match() called in line %d\n", __LINE__); \
235    rx = match(ra,rb,rc,rd,re,rf,rg,rdepth+1); \    rrc = match(ra,rb,mstart,rc,rd,re,rf,rg,rdepth+1); \
236    printf("to line %d\n", __LINE__); \    printf("to line %d\n", __LINE__); \
237    }    }
238  #define RRETURN(ra) \  #define RRETURN(ra) \
# Line 208  versions and production versions. */ Line 241  versions and production versions. */
241    return ra; \    return ra; \
242    }    }
243  #else  #else
244  #define RMATCH(rx,ra,rb,rc,rd,re,rf,rg) \  #define RMATCH(ra,rb,rc,rd,re,rf,rg,rw) \
245    rx = match(ra,rb,rc,rd,re,rf,rg,rdepth+1)    rrc = match(ra,rb,mstart,rc,rd,re,rf,rg,rdepth+1)
246  #define RRETURN(ra) return ra  #define RRETURN(ra) return ra
247  #endif  #endif
248    
249  #else  #else
250    
251    
252  /* These versions of the macros manage a private stack on the heap. Note  /* These versions of the macros manage a private stack on the heap. Note that
253  that the rd argument of RMATCH isn't actually used. It's the md argument of  the "rd" argument of RMATCH isn't actually used in this definition. It's the md
254  match(), which never changes. */  argument of match(), which never changes. */
255    
256  #define REGISTER  #define REGISTER
257    
258  #define RMATCH(rx,ra,rb,rc,rd,re,rf,rg)\  #define RMATCH(ra,rb,rc,rd,re,rf,rg,rw)\
259    {\    {\
260    heapframe *newframe = (pcre_stack_malloc)(sizeof(heapframe));\    heapframe *newframe = (pcre_stack_malloc)(sizeof(heapframe));\
261    if (setjmp(frame->Xwhere) == 0)\    frame->Xwhere = rw; \
262      {\    newframe->Xeptr = ra;\
263      newframe->Xeptr = ra;\    newframe->Xecode = rb;\
264      newframe->Xecode = rb;\    newframe->Xmstart = mstart;\
265      newframe->Xoffset_top = rc;\    newframe->Xoffset_top = rc;\
266      newframe->Xims = re;\    newframe->Xims = re;\
267      newframe->Xeptrb = rf;\    newframe->Xeptrb = rf;\
268      newframe->Xflags = rg;\    newframe->Xflags = rg;\
269      newframe->Xrdepth = frame->Xrdepth + 1;\    newframe->Xrdepth = frame->Xrdepth + 1;\
270      newframe->Xprevframe = frame;\    newframe->Xprevframe = frame;\
271      frame = newframe;\    frame = newframe;\
272      DPRINTF(("restarting from line %d\n", __LINE__));\    DPRINTF(("restarting from line %d\n", __LINE__));\
273      goto HEAP_RECURSE;\    goto HEAP_RECURSE;\
274      }\    L_##rw:\
275    else\    DPRINTF(("jumped back to line %d\n", __LINE__));\
     {\  
     DPRINTF(("longjumped back to line %d\n", __LINE__));\  
     frame = md->thisframe;\  
     rx = frame->Xresult;\  
     }\  
276    }    }
277    
278  #define RRETURN(ra)\  #define RRETURN(ra)\
# Line 254  match(), which never changes. */ Line 282  match(), which never changes. */
282    (pcre_stack_free)(newframe);\    (pcre_stack_free)(newframe);\
283    if (frame != NULL)\    if (frame != NULL)\
284      {\      {\
285      frame->Xresult = ra;\      rrc = ra;\
286      md->thisframe = frame;\      goto HEAP_RETURN;\
     longjmp(frame->Xwhere, 1);\  
287      }\      }\
288    return ra;\    return ra;\
289    }    }
# Line 271  typedef struct heapframe { Line 298  typedef struct heapframe {
298    
299    const uschar *Xeptr;    const uschar *Xeptr;
300    const uschar *Xecode;    const uschar *Xecode;
301      const uschar *Xmstart;
302    int Xoffset_top;    int Xoffset_top;
303    long int Xims;    long int Xims;
304    eptrblock *Xeptrb;    eptrblock *Xeptrb;
# Line 291  typedef struct heapframe { Line 319  typedef struct heapframe {
319    
320    BOOL Xcur_is_word;    BOOL Xcur_is_word;
321    BOOL Xcondition;    BOOL Xcondition;
   BOOL Xminimize;  
322    BOOL Xprev_is_word;    BOOL Xprev_is_word;
323    
324    unsigned long int Xoriginal_ims;    unsigned long int Xoriginal_ims;
# Line 303  typedef struct heapframe { Line 330  typedef struct heapframe {
330    int Xprop_category;    int Xprop_category;
331    int Xprop_chartype;    int Xprop_chartype;
332    int Xprop_script;    int Xprop_script;
333    int *Xprop_test_variable;    int Xoclength;
334      uschar Xocchars[8];
335  #endif  #endif
336    
337    int Xctype;    int Xctype;
338    int Xfc;    unsigned int Xfc;
339    int Xfi;    int Xfi;
340    int Xlength;    int Xlength;
341    int Xmax;    int Xmax;
# Line 321  typedef struct heapframe { Line 349  typedef struct heapframe {
349    
350    eptrblock Xnewptrb;    eptrblock Xnewptrb;
351    
352    /* Place to pass back result, and where to jump back to */    /* Where to jump back to */
353    
354    int  Xresult;    int Xwhere;
   jmp_buf Xwhere;  
355    
356  } heapframe;  } heapframe;
357    
# Line 340  typedef struct heapframe { Line 367  typedef struct heapframe {
367  *         Match from current position            *  *         Match from current position            *
368  *************************************************/  *************************************************/
369    
370  /* On entry ecode points to the first opcode, and eptr to the first character  /* This function is called recursively in many circumstances. Whenever it
 in the subject string, while eptrb holds the value of eptr at the start of the  
 last bracketed group - used for breaking infinite loops matching zero-length  
 strings. This function is called recursively in many circumstances. Whenever it  
371  returns a negative (error) response, the outer incarnation must also return the  returns a negative (error) response, the outer incarnation must also return the
372  same response.  same response.
373    
# Line 353  performance. Tests using gcc on a SPARC Line 377  performance. Tests using gcc on a SPARC
377  made performance worse.  made performance worse.
378    
379  Arguments:  Arguments:
380     eptr        pointer in subject     eptr        pointer to current character in subject
381     ecode       position in code     ecode       pointer to current position in compiled code
382       mstart      pointer to the current match start position (can be modified
383                     by encountering \K)
384     offset_top  current top pointer     offset_top  current top pointer
385     md          pointer to "static" info for the match     md          pointer to "static" info for the match
386     ims         current /i, /m, and /s options     ims         current /i, /m, and /s options
# Line 362  Arguments: Line 388  Arguments:
388                   brackets - for testing for empty matches                   brackets - for testing for empty matches
389     flags       can contain     flags       can contain
390                   match_condassert - this is an assertion condition                   match_condassert - this is an assertion condition
391                   match_isgroup - this is the start of a bracketed group                   match_cbegroup - this is the start of an unlimited repeat
392                       group that can match an empty string
393     rdepth      the recursion depth     rdepth      the recursion depth
394    
395  Returns:       MATCH_MATCH if matched            )  these values are >= 0  Returns:       MATCH_MATCH if matched            )  these values are >= 0
# Line 372  Returns:       MATCH_MATCH if matched Line 399  Returns:       MATCH_MATCH if matched
399  */  */
400    
401  static int  static int
402  match(REGISTER USPTR eptr, REGISTER const uschar *ecode,  match(REGISTER USPTR eptr, REGISTER const uschar *ecode, const uschar *mstart,
403    int offset_top, match_data *md, unsigned long int ims, eptrblock *eptrb,    int offset_top, match_data *md, unsigned long int ims, eptrblock *eptrb,
404    int flags, unsigned int rdepth)    int flags, unsigned int rdepth)
405  {  {
406  /* 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,
407  so they can be ordinary variables in all cases. Mark them with "register"  so they can be ordinary variables in all cases. Mark some of them with
408  because they are used a lot in loops. */  "register" because they are used a lot in loops. */
409    
410  register int  rrc;         /* Returns from recursive calls */  register int  rrc;         /* Returns from recursive calls */
411  register int  i;           /* Used for loops not involving calls to RMATCH() */  register int  i;           /* Used for loops not involving calls to RMATCH() */
412  register unsigned int  c;  /* Character values not kept over RMATCH() calls */  register unsigned int c;   /* Character values not kept over RMATCH() calls */
413  register BOOL utf8;        /* Local copy of UTF-8 flag for speed */  register BOOL utf8;        /* Local copy of UTF-8 flag for speed */
414    
415    BOOL minimize, possessive; /* Quantifier options */
416    
417  /* 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
418  preserved over calls to RMATCH() are part of a "frame" which is obtained from  preserved over calls to RMATCH() are part of a "frame" which is obtained from
419  heap storage. Set up the top-level frame here; others are obtained from the  heap storage. Set up the top-level frame here; others are obtained from the
# Line 398  frame->Xprevframe = NULL;            /* Line 427  frame->Xprevframe = NULL;            /*
427    
428  frame->Xeptr = eptr;  frame->Xeptr = eptr;
429  frame->Xecode = ecode;  frame->Xecode = ecode;
430    frame->Xmstart = mstart;
431  frame->Xoffset_top = offset_top;  frame->Xoffset_top = offset_top;
432  frame->Xims = ims;  frame->Xims = ims;
433  frame->Xeptrb = eptrb;  frame->Xeptrb = eptrb;
# Line 412  HEAP_RECURSE: Line 442  HEAP_RECURSE:
442    
443  #define eptr               frame->Xeptr  #define eptr               frame->Xeptr
444  #define ecode              frame->Xecode  #define ecode              frame->Xecode
445    #define mstart             frame->Xmstart
446  #define offset_top         frame->Xoffset_top  #define offset_top         frame->Xoffset_top
447  #define ims                frame->Xims  #define ims                frame->Xims
448  #define eptrb              frame->Xeptrb  #define eptrb              frame->Xeptrb
# Line 434  HEAP_RECURSE: Line 465  HEAP_RECURSE:
465    
466  #define cur_is_word        frame->Xcur_is_word  #define cur_is_word        frame->Xcur_is_word
467  #define condition          frame->Xcondition  #define condition          frame->Xcondition
 #define minimize           frame->Xminimize  
468  #define prev_is_word       frame->Xprev_is_word  #define prev_is_word       frame->Xprev_is_word
469    
470  #define original_ims       frame->Xoriginal_ims  #define original_ims       frame->Xoriginal_ims
# Line 446  HEAP_RECURSE: Line 476  HEAP_RECURSE:
476  #define prop_category      frame->Xprop_category  #define prop_category      frame->Xprop_category
477  #define prop_chartype      frame->Xprop_chartype  #define prop_chartype      frame->Xprop_chartype
478  #define prop_script        frame->Xprop_script  #define prop_script        frame->Xprop_script
479  #define prop_test_variable frame->Xprop_test_variable  #define oclength           frame->Xoclength
480    #define occhars            frame->Xocchars
481  #endif  #endif
482    
483  #define ctype              frame->Xctype  #define ctype              frame->Xctype
# Line 470  HEAP_RECURSE: Line 501  HEAP_RECURSE:
501  get preserved during recursion in the normal way. In this environment, fi and  get preserved during recursion in the normal way. In this environment, fi and
502  i, and fc and c, can be the same variables. */  i, and fc and c, can be the same variables. */
503    
504  #else  #else         /* NO_RECURSE not defined */
505  #define fi i  #define fi i
506  #define fc c  #define fc c
507    
# Line 489  recursion_info new_recursive;      /* wi Line 520  recursion_info new_recursive;      /* wi
520                                     /* that do not have to be preserved over  */                                     /* that do not have to be preserved over  */
521  BOOL cur_is_word;                  /* a recursive call to RMATCH().          */  BOOL cur_is_word;                  /* a recursive call to RMATCH().          */
522  BOOL condition;  BOOL condition;
 BOOL minimize;  
523  BOOL prev_is_word;  BOOL prev_is_word;
524    
525  unsigned long int original_ims;  unsigned long int original_ims;
# Line 501  int prop_fail_result; Line 531  int prop_fail_result;
531  int prop_category;  int prop_category;
532  int prop_chartype;  int prop_chartype;
533  int prop_script;  int prop_script;
534  int *prop_test_variable;  int oclength;
535    uschar occhars[8];
536  #endif  #endif
537    
538  int ctype;  int ctype;
# Line 516  int save_offset1, save_offset2, save_off Line 547  int save_offset1, save_offset2, save_off
547  int stacksave[REC_STACK_SAVE_MAX];  int stacksave[REC_STACK_SAVE_MAX];
548    
549  eptrblock newptrb;  eptrblock newptrb;
550  #endif  #endif     /* NO_RECURSE */
551    
552  /* These statements are here to stop the compiler complaining about unitialized  /* These statements are here to stop the compiler complaining about unitialized
553  variables. */  variables. */
# Line 524  variables. */ Line 555  variables. */
555  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
556  prop_value = 0;  prop_value = 0;
557  prop_fail_result = 0;  prop_fail_result = 0;
 prop_test_variable = NULL;  
558  #endif  #endif
559    
560    
561  /* This label is used for tail recursion, which is used in a few cases even  /* This label is used for tail recursion, which is used in a few cases even
562  when NO_RECURSE is not defined, in order to reduce the amount of stack that is  when NO_RECURSE is not defined, in order to reduce the amount of stack that is
563  used. Thanks to Ian Taylor for noticing this possibility and sending the  used. Thanks to Ian Taylor for noticing this possibility and sending the
# Line 542  defined). However, RMATCH isn't like a f Line 573  defined). However, RMATCH isn't like a f
573  complicated macro. It has to be used in one particular way. This shouldn't,  complicated macro. It has to be used in one particular way. This shouldn't,
574  however, impact performance when true recursion is being used. */  however, impact performance when true recursion is being used. */
575    
576    #ifdef SUPPORT_UTF8
577    utf8 = md->utf8;       /* Local copy of the flag */
578    #else
579    utf8 = FALSE;
580    #endif
581    
582  /* First check that we haven't called match() too many times, or that we  /* First check that we haven't called match() too many times, or that we
583  haven't exceeded the recursive call limit. */  haven't exceeded the recursive call limit. */
584    
# Line 550  if (rdepth >= md->match_limit_recursion) Line 587  if (rdepth >= md->match_limit_recursion)
587    
588  original_ims = ims;    /* Save for resetting on ')' */  original_ims = ims;    /* Save for resetting on ')' */
589    
590  #ifdef SUPPORT_UTF8  /* At the start of a group with an unlimited repeat that may match an empty
591  utf8 = md->utf8;       /* Local copy of the flag */  string, the match_cbegroup flag is set. When this is the case, add the current
592  #else  subject pointer to the chain of such remembered pointers, to be checked when we
593  utf8 = FALSE;  hit the closing ket, in order to break infinite loops that match no characters.
594  #endif  When match() is called in other circumstances, don't add to the chain. The
595    match_cbegroup flag must NOT be used with tail recursion, because the memory
596  /* At the start of a bracketed group, add the current subject pointer to the  block that is used is on the stack, so a new one may be required for each
597  stack of such pointers, to be re-instated at the end of the group when we hit  match(). */
 the closing ket. When match() is called in other circumstances, we don't add to  
 this stack. */  
598    
599  if ((flags & match_isgroup) != 0)  if ((flags & match_cbegroup) != 0)
600    {    {
   newptrb.epb_prev = eptrb;  
601    newptrb.epb_saved_eptr = eptr;    newptrb.epb_saved_eptr = eptr;
602      newptrb.epb_prev = eptrb;
603    eptrb = &newptrb;    eptrb = &newptrb;
604    }    }
605    
606  /* Now start processing the operations. */  /* Now start processing the opcodes. */
607    
608  for (;;)  for (;;)
609    {    {
610      minimize = possessive = FALSE;
611    op = *ecode;    op = *ecode;
   minimize = FALSE;  
612    
613    /* For partial matching, remember if we ever hit the end of the subject after    /* For partial matching, remember if we ever hit the end of the subject after
614    matching at least one subject character. */    matching at least one subject character. */
615    
616    if (md->partial &&    if (md->partial &&
617        eptr >= md->end_subject &&        eptr >= md->end_subject &&
618        eptr > md->start_match)        eptr > mstart)
619      md->hitend = TRUE;      md->hitend = TRUE;
620    
621    /* Opening capturing bracket. If there is space in the offset vector, save    switch(op)
   the current subject position in the working slot at the top of the vector. We  
   mustn't change the current values of the data slot, because they may be set  
   from a previous iteration of this group, and be referred to by a reference  
   inside the group.  
   
   If the bracket fails to match, we need to restore this value and also the  
   values of the final offsets, in case they were set by a previous iteration of  
   the same bracket.  
   
   If there isn't enough space in the offset vector, treat this as if it were a  
   non-capturing bracket. Don't worry about setting the flag for the error case  
   here; that is handled in the code for KET. */  
   
   if (op > OP_BRA)  
622      {      {
623      number = op - OP_BRA;      case OP_FAIL:
624        RRETURN(MATCH_NOMATCH);
625    
626      /* For extended extraction brackets (large number), we have to fish out the      case OP_PRUNE:
627      number from a dummy opcode at the start. */      RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,
628          ims, eptrb, flags, RM51);
629      if (number > EXTRACT_BASIC_MAX)      if (rrc != MATCH_NOMATCH) RRETURN(rrc);
630        number = GET2(ecode, 2+LINK_SIZE);      RRETURN(MATCH_PRUNE);
631    
632        case OP_COMMIT:
633        RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,
634          ims, eptrb, flags, RM52);
635        if (rrc != MATCH_NOMATCH) RRETURN(rrc);
636        RRETURN(MATCH_COMMIT);
637    
638        case OP_SKIP:
639        RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,
640          ims, eptrb, flags, RM53);
641        if (rrc != MATCH_NOMATCH) RRETURN(rrc);
642        md->start_match_ptr = eptr;   /* Pass back current position */
643        RRETURN(MATCH_SKIP);
644    
645        case OP_THEN:
646        RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,
647          ims, eptrb, flags, RM54);
648        if (rrc != MATCH_NOMATCH) RRETURN(rrc);
649        RRETURN(MATCH_THEN);
650    
651        /* Handle a capturing bracket. If there is space in the offset vector, save
652        the current subject position in the working slot at the top of the vector.
653        We mustn't change the current values of the data slot, because they may be
654        set from a previous iteration of this group, and be referred to by a
655        reference inside the group.
656    
657        If the bracket fails to match, we need to restore this value and also the
658        values of the final offsets, in case they were set by a previous iteration
659        of the same bracket.
660    
661        If there isn't enough space in the offset vector, treat this as if it were
662        a non-capturing bracket. Don't worry about setting the flag for the error
663        case here; that is handled in the code for KET. */
664    
665        case OP_CBRA:
666        case OP_SCBRA:
667        number = GET2(ecode, 1+LINK_SIZE);
668      offset = number << 1;      offset = number << 1;
669    
670  #ifdef DEBUG  #ifdef DEBUG
671      printf("start bracket %d subject=", number);      printf("start bracket %d\n", number);
672        printf("subject=");
673      pchars(eptr, 16, TRUE, md);      pchars(eptr, 16, TRUE, md);
674      printf("\n");      printf("\n");
675  #endif  #endif
# Line 624  for (;;) Line 684  for (;;)
684        DPRINTF(("saving %d %d %d\n", save_offset1, save_offset2, save_offset3));        DPRINTF(("saving %d %d %d\n", save_offset1, save_offset2, save_offset3));
685        md->offset_vector[md->offset_end - number] = eptr - md->start_subject;        md->offset_vector[md->offset_end - number] = eptr - md->start_subject;
686    
687          flags = (op == OP_SCBRA)? match_cbegroup : 0;
688        do        do
689          {          {
690          RMATCH(rrc, eptr, ecode + 1 + LINK_SIZE, offset_top, md, ims, eptrb,          RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,
691            match_isgroup);            ims, eptrb, flags, RM1);
692          if (rrc != MATCH_NOMATCH) RRETURN(rrc);          if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN) RRETURN(rrc);
693          md->capture_last = save_capture_last;          md->capture_last = save_capture_last;
694          ecode += GET(ecode, 1);          ecode += GET(ecode, 1);
695          }          }
# Line 643  for (;;) Line 704  for (;;)
704        RRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
705        }        }
706    
707      /* Insufficient room for saving captured contents */      /* FALL THROUGH ... Insufficient room for saving captured contents. Treat
708        as a non-capturing bracket. */
     else op = OP_BRA;  
     }  
   
   /* Other types of node can be handled by a switch */  
709    
710    switch(op)      /* VVVVVVVVVVVVVVVVVVVVVVVVV */
711      {      /* VVVVVVVVVVVVVVVVVVVVVVVVV */
     case OP_BRA:     /* Non-capturing bracket: optimized */  
     DPRINTF(("start bracket 0\n"));  
712    
713      /* Loop for all the alternatives */      DPRINTF(("insufficient capture room: treat as non-capturing\n"));
714    
715        /* VVVVVVVVVVVVVVVVVVVVVVVVV */
716        /* VVVVVVVVVVVVVVVVVVVVVVVVV */
717    
718        /* Non-capturing bracket. Loop for all the alternatives. When we get to the
719        final alternative within the brackets, we would return the result of a
720        recursive call to match() whatever happened. We can reduce stack usage by
721        turning this into a tail recursion, except in the case when match_cbegroup
722        is set.*/
723    
724        case OP_BRA:
725        case OP_SBRA:
726        DPRINTF(("start non-capturing bracket\n"));
727        flags = (op >= OP_SBRA)? match_cbegroup : 0;
728      for (;;)      for (;;)
729        {        {
730        /* When we get to the final alternative within the brackets, we would        if (ecode[GET(ecode, 1)] != OP_ALT)   /* Final alternative */
731        return the result of a recursive call to match() whatever happened. We          {
732        can reduce stack usage by turning this into a tail recursion. */          if (flags == 0)    /* Not a possibly empty group */
733              {
734        if (ecode[GET(ecode, 1)] != OP_ALT)            ecode += _pcre_OP_lengths[*ecode];
735         {            DPRINTF(("bracket 0 tail recursion\n"));
736         ecode += 1 + LINK_SIZE;            goto TAIL_RECURSE;
737         flags = match_isgroup;            }
738         DPRINTF(("bracket 0 tail recursion\n"));  
739         goto TAIL_RECURSE;          /* Possibly empty group; can't use tail recursion. */
740         }  
741            RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md, ims,
742              eptrb, flags, RM48);
743            RRETURN(rrc);
744            }
745    
746        /* For non-final alternatives, continue the loop for a NOMATCH result;        /* For non-final alternatives, continue the loop for a NOMATCH result;
747        otherwise return. */        otherwise return. */
748    
749        RMATCH(rrc, eptr, ecode + 1 + LINK_SIZE, offset_top, md, ims, eptrb,        RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md, ims,
750          match_isgroup);          eptrb, flags, RM2);
751        if (rrc != MATCH_NOMATCH) RRETURN(rrc);        if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN) RRETURN(rrc);
752        ecode += GET(ecode, 1);        ecode += GET(ecode, 1);
753        }        }
754      /* Control never reaches here. */      /* Control never reaches here. */
# Line 688  for (;;) Line 760  for (;;)
760      obeyed, we can use tail recursion to avoid using another stack frame. */      obeyed, we can use tail recursion to avoid using another stack frame. */
761    
762      case OP_COND:      case OP_COND:
763      if (ecode[LINK_SIZE+1] == OP_CREF) /* Condition extract or recurse test */      case OP_SCOND:
764        if (ecode[LINK_SIZE+1] == OP_RREF)         /* Recursion test */
765          {
766          offset = GET2(ecode, LINK_SIZE + 2);     /* Recursion group number*/
767          condition = md->recursive != NULL &&
768            (offset == RREF_ANY || offset == md->recursive->group_num);
769          ecode += condition? 3 : GET(ecode, 1);
770          }
771    
772        else if (ecode[LINK_SIZE+1] == OP_CREF)    /* Group used test */
773        {        {
774        offset = GET2(ecode, LINK_SIZE+2) << 1;  /* Doubled ref number */        offset = GET2(ecode, LINK_SIZE+2) << 1;  /* Doubled ref number */
775        condition = (offset == CREF_RECURSE * 2)?        condition = offset < offset_top && md->offset_vector[offset] >= 0;
776          (md->recursive != NULL) :        ecode += condition? 3 : GET(ecode, 1);
777          (offset < offset_top && md->offset_vector[offset] >= 0);        }
778        ecode += condition? (LINK_SIZE + 4) : (LINK_SIZE + 1 + GET(ecode, 1));  
779        flags = match_isgroup;      else if (ecode[LINK_SIZE+1] == OP_DEF)     /* DEFINE - always false */
780        goto TAIL_RECURSE;        {
781          condition = FALSE;
782          ecode += GET(ecode, 1);
783        }        }
784    
785      /* The condition is an assertion. Call match() to evaluate it - setting      /* The condition is an assertion. Call match() to evaluate it - setting
786      the final argument TRUE causes it to stop at the end of an assertion. */      the final argument match_condassert causes it to stop at the end of an
787        assertion. */
788    
789      else      else
790        {        {
791        RMATCH(rrc, eptr, ecode + 1 + LINK_SIZE, offset_top, md, ims, NULL,        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, ims, NULL,
792            match_condassert | match_isgroup);            match_condassert, RM3);
793        if (rrc == MATCH_MATCH)        if (rrc == MATCH_MATCH)
794          {          {
795          ecode += 1 + LINK_SIZE + GET(ecode, LINK_SIZE+2);          condition = TRUE;
796            ecode += 1 + LINK_SIZE + GET(ecode, LINK_SIZE + 2);
797          while (*ecode == OP_ALT) ecode += GET(ecode, 1);          while (*ecode == OP_ALT) ecode += GET(ecode, 1);
798          }          }
799        else if (rrc != MATCH_NOMATCH)        else if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN)
800          {          {
801          RRETURN(rrc);         /* Need braces because of following else */          RRETURN(rrc);         /* Need braces because of following else */
802          }          }
803        else ecode += GET(ecode, 1);        else
804            {
805            condition = FALSE;
806            ecode += GET(ecode, 1);
807            }
808          }
809    
810        /* 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,
811        we can use tail recursion to avoid using another stack frame. */      we can use tail recursion to avoid using another stack frame, except when
812        match_cbegroup is required for an unlimited repeat of a possibly empty
813        group. If the second alternative doesn't exist, we can just plough on. */
814    
815        if (condition || *ecode == OP_ALT)
816          {
817          ecode += 1 + LINK_SIZE;
818          if (op == OP_SCOND)        /* Possibly empty group */
819            {
820            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, match_cbegroup, RM49);
821            RRETURN(rrc);
822            }
823          else                       /* Group must match something */
824            {
825            flags = 0;
826            goto TAIL_RECURSE;
827            }
828          }
829        else                         /* Condition false & no 2nd alternative */
830          {
831        ecode += 1 + LINK_SIZE;        ecode += 1 + LINK_SIZE;
       flags = match_isgroup;  
       goto TAIL_RECURSE;  
832        }        }
     /* Control never reaches here */  
   
     /* Skip over conditional reference or large extraction number data if  
     encountered. */  
   
     case OP_CREF:  
     case OP_BRANUMBER:  
     ecode += 3;  
833      break;      break;
834    
     /* End of the pattern. If we are in a recursion, we should restore the  
     offsets appropriately and continue from after the call. */  
835    
836        /* End of the pattern, either real or forced. If we are in a top-level
837        recursion, we should restore the offsets appropriately and continue from
838        after the call. */
839    
840        case OP_ACCEPT:
841      case OP_END:      case OP_END:
842      if (md->recursive != NULL && md->recursive->group_num == 0)      if (md->recursive != NULL && md->recursive->group_num == 0)
843        {        {
# Line 745  for (;;) Line 846  for (;;)
846        md->recursive = rec->prevrec;        md->recursive = rec->prevrec;
847        memmove(md->offset_vector, rec->offset_save,        memmove(md->offset_vector, rec->offset_save,
848          rec->saved_max * sizeof(int));          rec->saved_max * sizeof(int));
849        md->start_match = rec->save_start;        mstart = rec->save_start;
850        ims = original_ims;        ims = original_ims;
851        ecode = rec->after_call;        ecode = rec->after_call;
852        break;        break;
# Line 754  for (;;) Line 855  for (;;)
855      /* Otherwise, if PCRE_NOTEMPTY is set, fail if we have matched an empty      /* Otherwise, if PCRE_NOTEMPTY is set, fail if we have matched an empty
856      string - backtracking will then try other alternatives, if any. */      string - backtracking will then try other alternatives, if any. */
857    
858      if (md->notempty && eptr == md->start_match) RRETURN(MATCH_NOMATCH);      if (md->notempty && eptr == mstart) RRETURN(MATCH_NOMATCH);
859      md->end_match_ptr = eptr;          /* Record where we ended */      md->end_match_ptr = eptr;           /* Record where we ended */
860      md->end_offset_top = offset_top;   /* and how many extracts were taken */      md->end_offset_top = offset_top;    /* and how many extracts were taken */
861        md->start_match_ptr = mstart;       /* and the start (\K can modify) */
862      RRETURN(MATCH_MATCH);      RRETURN(MATCH_MATCH);
863    
864      /* Change option settings */      /* Change option settings */
# Line 777  for (;;) Line 879  for (;;)
879      case OP_ASSERTBACK:      case OP_ASSERTBACK:
880      do      do
881        {        {
882        RMATCH(rrc, eptr, ecode + 1 + LINK_SIZE, offset_top, md, ims, NULL,        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, ims, NULL, 0,
883          match_isgroup);          RM4);
884        if (rrc == MATCH_MATCH) break;        if (rrc == MATCH_MATCH) break;
885        if (rrc != MATCH_NOMATCH) RRETURN(rrc);        if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN) RRETURN(rrc);
886        ecode += GET(ecode, 1);        ecode += GET(ecode, 1);
887        }        }
888      while (*ecode == OP_ALT);      while (*ecode == OP_ALT);
# Line 804  for (;;) Line 906  for (;;)
906      case OP_ASSERTBACK_NOT:      case OP_ASSERTBACK_NOT:
907      do      do
908        {        {
909        RMATCH(rrc, eptr, ecode + 1 + LINK_SIZE, offset_top, md, ims, NULL,        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, ims, NULL, 0,
910          match_isgroup);          RM5);
911        if (rrc == MATCH_MATCH) RRETURN(MATCH_NOMATCH);        if (rrc == MATCH_MATCH) RRETURN(MATCH_NOMATCH);
912        if (rrc != MATCH_NOMATCH) RRETURN(rrc);        if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN) RRETURN(rrc);
913        ecode += GET(ecode,1);        ecode += GET(ecode,1);
914        }        }
915      while (*ecode == OP_ALT);      while (*ecode == OP_ALT);
# Line 826  for (;;) Line 928  for (;;)
928  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF8
929      if (utf8)      if (utf8)
930        {        {
931        c = GET(ecode,1);        i = GET(ecode, 1);
932        for (i = 0; i < c; i++)        while (i-- > 0)
933          {          {
934          eptr--;          eptr--;
935          if (eptr < md->start_subject) RRETURN(MATCH_NOMATCH);          if (eptr < md->start_subject) RRETURN(MATCH_NOMATCH);
936          BACKCHAR(eptr)          BACKCHAR(eptr);
937          }          }
938        }        }
939      else      else
# Line 840  for (;;) Line 942  for (;;)
942      /* No UTF-8 support, or not in UTF-8 mode: count is byte count */      /* No UTF-8 support, or not in UTF-8 mode: count is byte count */
943    
944        {        {
945        eptr -= GET(ecode,1);        eptr -= GET(ecode, 1);
946        if (eptr < md->start_subject) RRETURN(MATCH_NOMATCH);        if (eptr < md->start_subject) RRETURN(MATCH_NOMATCH);
947        }        }
948    
# Line 862  for (;;) Line 964  for (;;)
964        cb.offset_vector    = md->offset_vector;        cb.offset_vector    = md->offset_vector;
965        cb.subject          = (PCRE_SPTR)md->start_subject;        cb.subject          = (PCRE_SPTR)md->start_subject;
966        cb.subject_length   = md->end_subject - md->start_subject;        cb.subject_length   = md->end_subject - md->start_subject;
967        cb.start_match      = md->start_match - md->start_subject;        cb.start_match      = mstart - md->start_subject;
968        cb.current_position = eptr - md->start_subject;        cb.current_position = eptr - md->start_subject;
969        cb.pattern_position = GET(ecode, 2);        cb.pattern_position = GET(ecode, 2);
970        cb.next_item_length = GET(ecode, 2 + LINK_SIZE);        cb.next_item_length = GET(ecode, 2 + LINK_SIZE);
# Line 897  for (;;) Line 999  for (;;)
999      case OP_RECURSE:      case OP_RECURSE:
1000        {        {
1001        callpat = md->start_code + GET(ecode, 1);        callpat = md->start_code + GET(ecode, 1);
1002        new_recursive.group_num = *callpat - OP_BRA;        new_recursive.group_num = (callpat == md->start_code)? 0 :
1003            GET2(callpat, 1 + LINK_SIZE);
       /* For extended extraction brackets (large number), we have to fish out  
       the number from a dummy opcode at the start. */  
   
       if (new_recursive.group_num > EXTRACT_BASIC_MAX)  
         new_recursive.group_num = GET2(callpat, 2+LINK_SIZE);  
1004    
1005        /* Add to "recursing stack" */        /* Add to "recursing stack" */
1006    
# Line 929  for (;;) Line 1026  for (;;)
1026    
1027        memcpy(new_recursive.offset_save, md->offset_vector,        memcpy(new_recursive.offset_save, md->offset_vector,
1028              new_recursive.saved_max * sizeof(int));              new_recursive.saved_max * sizeof(int));
1029        new_recursive.save_start = md->start_match;        new_recursive.save_start = mstart;
1030        md->start_match = eptr;        mstart = eptr;
1031    
1032        /* OK, now we can do the recursion. For each top-level alternative we        /* OK, now we can do the recursion. For each top-level alternative we
1033        restore the offset and recursion data. */        restore the offset and recursion data. */
1034    
1035        DPRINTF(("Recursing into group %d\n", new_recursive.group_num));        DPRINTF(("Recursing into group %d\n", new_recursive.group_num));
1036          flags = (*callpat >= OP_SBRA)? match_cbegroup : 0;
1037        do        do
1038          {          {
1039          RMATCH(rrc, eptr, callpat + 1 + LINK_SIZE, offset_top, md, ims,          RMATCH(eptr, callpat + _pcre_OP_lengths[*callpat], offset_top,
1040              eptrb, match_isgroup);            md, ims, eptrb, flags, RM6);
1041          if (rrc == MATCH_MATCH)          if (rrc == MATCH_MATCH)
1042            {            {
1043            DPRINTF(("Recursion matched\n"));            DPRINTF(("Recursion matched\n"));
# Line 948  for (;;) Line 1046  for (;;)
1046              (pcre_free)(new_recursive.offset_save);              (pcre_free)(new_recursive.offset_save);
1047            RRETURN(MATCH_MATCH);            RRETURN(MATCH_MATCH);
1048            }            }
1049          else if (rrc != MATCH_NOMATCH)          else if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN)
1050            {            {
1051            DPRINTF(("Recursion gave error %d\n", rrc));            DPRINTF(("Recursion gave error %d\n", rrc));
1052            RRETURN(rrc);            RRETURN(rrc);
# Line 982  for (;;) Line 1080  for (;;)
1080    
1081      do      do
1082        {        {
1083        RMATCH(rrc, eptr, ecode + 1 + LINK_SIZE, offset_top, md, ims,        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, ims, eptrb, 0, RM7);
         eptrb, match_isgroup);  
1084        if (rrc == MATCH_MATCH) break;        if (rrc == MATCH_MATCH) break;
1085        if (rrc != MATCH_NOMATCH) RRETURN(rrc);        if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN) RRETURN(rrc);
1086        ecode += GET(ecode,1);        ecode += GET(ecode,1);
1087        }        }
1088      while (*ecode == OP_ALT);      while (*ecode == OP_ALT);
# Line 997  for (;;) Line 1094  for (;;)
1094      /* Continue as from after the assertion, updating the offsets high water      /* Continue as from after the assertion, updating the offsets high water
1095      mark, since extracts may have been taken. */      mark, since extracts may have been taken. */
1096    
1097      do ecode += GET(ecode,1); while (*ecode == OP_ALT);      do ecode += GET(ecode, 1); while (*ecode == OP_ALT);
1098    
1099      offset_top = md->end_offset_top;      offset_top = md->end_offset_top;
1100      eptr = md->end_match_ptr;      eptr = md->end_match_ptr;
# Line 1028  for (;;) Line 1125  for (;;)
1125    
1126      if (*ecode == OP_KETRMIN)      if (*ecode == OP_KETRMIN)
1127        {        {
1128        RMATCH(rrc, eptr, ecode + 1 + LINK_SIZE, offset_top, md, ims, eptrb, 0);        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, ims, eptrb, 0, RM8);
1129        if (rrc != MATCH_NOMATCH) RRETURN(rrc);        if (rrc != MATCH_NOMATCH) RRETURN(rrc);
1130        ecode = prev;        ecode = prev;
1131        flags = match_isgroup;        flags = 0;
1132        goto TAIL_RECURSE;        goto TAIL_RECURSE;
1133        }        }
1134      else  /* OP_KETRMAX */      else  /* OP_KETRMAX */
1135        {        {
1136        RMATCH(rrc, eptr, prev, offset_top, md, ims, eptrb, match_isgroup);        RMATCH(eptr, prev, offset_top, md, ims, eptrb, match_cbegroup, RM9);
1137        if (rrc != MATCH_NOMATCH) RRETURN(rrc);        if (rrc != MATCH_NOMATCH) RRETURN(rrc);
1138        ecode += 1 + LINK_SIZE;        ecode += 1 + LINK_SIZE;
1139        flags = 0;        flags = 0;
# Line 1060  for (;;) Line 1157  for (;;)
1157      case OP_BRAZERO:      case OP_BRAZERO:
1158        {        {
1159        next = ecode+1;        next = ecode+1;
1160        RMATCH(rrc, eptr, next, offset_top, md, ims, eptrb, match_isgroup);        RMATCH(eptr, next, offset_top, md, ims, eptrb, 0, RM10);
1161        if (rrc != MATCH_NOMATCH) RRETURN(rrc);        if (rrc != MATCH_NOMATCH) RRETURN(rrc);
1162        do next += GET(next,1); while (*next == OP_ALT);        do next += GET(next,1); while (*next == OP_ALT);
1163        ecode = next + 1+LINK_SIZE;        ecode = next + 1 + LINK_SIZE;
1164        }        }
1165      break;      break;
1166    
1167      case OP_BRAMINZERO:      case OP_BRAMINZERO:
1168        {        {
1169        next = ecode+1;        next = ecode+1;
1170        do next += GET(next,1); while (*next == OP_ALT);        do next += GET(next, 1); while (*next == OP_ALT);
1171        RMATCH(rrc, eptr, next + 1+LINK_SIZE, offset_top, md, ims, eptrb,        RMATCH(eptr, next + 1+LINK_SIZE, offset_top, md, ims, eptrb, 0, RM11);
         match_isgroup);  
1172        if (rrc != MATCH_NOMATCH) RRETURN(rrc);        if (rrc != MATCH_NOMATCH) RRETURN(rrc);
1173        ecode++;        ecode++;
1174        }        }
1175      break;      break;
1176    
1177      /* End of a group, repeated or non-repeating. If we are at the end of      /* End of a group, repeated or non-repeating. */
     an assertion "group", stop matching and return MATCH_MATCH, but record the  
     current high water mark for use by positive assertions. Do this also  
     for the "once" (not-backup up) groups. */  
1178    
1179      case OP_KET:      case OP_KET:
1180      case OP_KETRMIN:      case OP_KETRMIN:
1181      case OP_KETRMAX:      case OP_KETRMAX:
1182      prev = ecode - GET(ecode, 1);      prev = ecode - GET(ecode, 1);
     saved_eptr = eptrb->epb_saved_eptr;  
1183    
1184      /* Back up the stack of bracket start pointers. */      /* If this was a group that remembered the subject start, in order to break
1185        infinite repeats of empty string matches, retrieve the subject start from
1186        the chain. Otherwise, set it NULL. */
1187    
1188        if (*prev >= OP_SBRA)
1189          {
1190          saved_eptr = eptrb->epb_saved_eptr;   /* Value at start of group */
1191          eptrb = eptrb->epb_prev;              /* Backup to previous group */
1192          }
1193        else saved_eptr = NULL;
1194    
1195      eptrb = eptrb->epb_prev;      /* If we are at the end of an assertion group, stop matching and return
1196        MATCH_MATCH, but record the current high water mark for use by positive
1197        assertions. Do this also for the "once" (atomic) groups. */
1198    
1199      if (*prev == OP_ASSERT || *prev == OP_ASSERT_NOT ||      if (*prev == OP_ASSERT || *prev == OP_ASSERT_NOT ||
1200          *prev == OP_ASSERTBACK || *prev == OP_ASSERTBACK_NOT ||          *prev == OP_ASSERTBACK || *prev == OP_ASSERTBACK_NOT ||
# Line 1102  for (;;) Line 1205  for (;;)
1205        RRETURN(MATCH_MATCH);        RRETURN(MATCH_MATCH);
1206        }        }
1207    
1208      /* In all other cases except a conditional group we have to check the      /* For capturing groups we have to check the group number back at the start
1209      group number back at the start and if necessary complete handling an      and if necessary complete handling an extraction by setting the offsets and
1210      extraction by setting the offsets and bumping the high water mark. */      bumping the high water mark. Note that whole-pattern recursion is coded as
1211        a recurse into group 0, so it won't be picked up here. Instead, we catch it
1212        when the OP_END is reached. Other recursion is handled here. */
1213    
1214      if (*prev != OP_COND)      if (*prev == OP_CBRA || *prev == OP_SCBRA)
1215        {        {
1216        number = *prev - OP_BRA;        number = GET2(prev, 1+LINK_SIZE);
   
       /* For extended extraction brackets (large number), we have to fish out  
       the number from a dummy opcode at the start. */  
   
       if (number > EXTRACT_BASIC_MAX) number = GET2(prev, 2+LINK_SIZE);  
1217        offset = number << 1;        offset = number << 1;
1218    
1219  #ifdef DEBUG  #ifdef DEBUG
# Line 1121  for (;;) Line 1221  for (;;)
1221        printf("\n");        printf("\n");
1222  #endif  #endif
1223    
1224        /* Test for a numbered group. This includes groups called as a result        md->capture_last = number;
1225        of recursion. Note that whole-pattern recursion is coded as a recurse        if (offset >= md->offset_max) md->offset_overflow = TRUE; else
       into group 0, so it won't be picked up here. Instead, we catch it when  
       the OP_END is reached. */  
   
       if (number > 0)  
1226          {          {
1227          md->capture_last = number;          md->offset_vector[offset] =
1228          if (offset >= md->offset_max) md->offset_overflow = TRUE; else            md->offset_vector[md->offset_end - number];
1229            {          md->offset_vector[offset+1] = eptr - md->start_subject;
1230            md->offset_vector[offset] =          if (offset_top <= offset) offset_top = offset + 2;
1231              md->offset_vector[md->offset_end - number];          }
1232            md->offset_vector[offset+1] = eptr - md->start_subject;  
1233            if (offset_top <= offset) offset_top = offset + 2;        /* Handle a recursively called group. Restore the offsets
1234            }        appropriately and continue from after the call. */
1235    
1236          /* Handle a recursively called group. Restore the offsets        if (md->recursive != NULL && md->recursive->group_num == number)
1237          appropriately and continue from after the call. */          {
1238            recursion_info *rec = md->recursive;
1239          if (md->recursive != NULL && md->recursive->group_num == number)          DPRINTF(("Recursion (%d) succeeded - continuing\n", number));
1240            {          md->recursive = rec->prevrec;
1241            recursion_info *rec = md->recursive;          mstart = rec->save_start;
1242            DPRINTF(("Recursion (%d) succeeded - continuing\n", number));          memcpy(md->offset_vector, rec->offset_save,
1243            md->recursive = rec->prevrec;            rec->saved_max * sizeof(int));
1244            md->start_match = rec->save_start;          ecode = rec->after_call;
1245            memcpy(md->offset_vector, rec->offset_save,          ims = original_ims;
1246              rec->saved_max * sizeof(int));          break;
           ecode = rec->after_call;  
           ims = original_ims;  
           break;  
           }  
1247          }          }
1248        }        }
1249    
1250      /* Reset the value of the ims flags, in case they got changed during      /* For both capturing and non-capturing groups, reset the value of the ims
1251      the group. */      flags, in case they got changed during the group. */
1252    
1253      ims = original_ims;      ims = original_ims;
1254      DPRINTF(("ims reset to %02lx\n", ims));      DPRINTF(("ims reset to %02lx\n", ims));
# Line 1175  for (;;) Line 1267  for (;;)
1267    
1268      /* The repeating kets try the rest of the pattern or restart from the      /* The repeating kets try the rest of the pattern or restart from the
1269      preceding bracket, in the appropriate order. In the second case, we can use      preceding bracket, in the appropriate order. In the second case, we can use
1270      tail recursion to avoid using another stack frame. */      tail recursion to avoid using another stack frame, unless we have an
1271        unlimited repeat of a group that can match an empty string. */
1272    
1273        flags = (*prev >= OP_SBRA)? match_cbegroup : 0;
1274    
1275      if (*ecode == OP_KETRMIN)      if (*ecode == OP_KETRMIN)
1276        {        {
1277        RMATCH(rrc, eptr, ecode + 1+LINK_SIZE, offset_top, md, ims, eptrb, 0);        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, ims, eptrb, 0, RM12);
1278        if (rrc != MATCH_NOMATCH) RRETURN(rrc);        if (rrc != MATCH_NOMATCH) RRETURN(rrc);
1279          if (flags != 0)    /* Could match an empty string */
1280            {
1281            RMATCH(eptr, prev, offset_top, md, ims, eptrb, flags, RM50);
1282            RRETURN(rrc);
1283            }
1284        ecode = prev;        ecode = prev;
       flags = match_isgroup;  
1285        goto TAIL_RECURSE;        goto TAIL_RECURSE;
1286        }        }
1287      else  /* OP_KETRMAX */      else  /* OP_KETRMAX */
1288        {        {
1289        RMATCH(rrc, eptr, prev, offset_top, md, ims, eptrb, match_isgroup);        RMATCH(eptr, prev, offset_top, md, ims, eptrb, flags, RM13);
1290        if (rrc != MATCH_NOMATCH) RRETURN(rrc);        if (rrc != MATCH_NOMATCH) RRETURN(rrc);
1291        ecode += 1 + LINK_SIZE;        ecode += 1 + LINK_SIZE;
1292        flags = 0;        flags = 0;
# Line 1202  for (;;) Line 1301  for (;;)
1301      if ((ims & PCRE_MULTILINE) != 0)      if ((ims & PCRE_MULTILINE) != 0)
1302        {        {
1303        if (eptr != md->start_subject &&        if (eptr != md->start_subject &&
1304            (eptr == md->end_subject ||            (eptr == md->end_subject || !WAS_NEWLINE(eptr)))
            eptr < md->start_subject + md->nllen ||  
            !IS_NEWLINE(eptr - md->nllen)))  
1305          RRETURN(MATCH_NOMATCH);          RRETURN(MATCH_NOMATCH);
1306        ecode++;        ecode++;
1307        break;        break;
# Line 1225  for (;;) Line 1322  for (;;)
1322      ecode++;      ecode++;
1323      break;      break;
1324    
1325        /* Reset the start of match point */
1326    
1327        case OP_SET_SOM:
1328        mstart = eptr;
1329        ecode++;
1330        break;
1331    
1332      /* Assert before internal newline if multiline, or before a terminating      /* Assert before internal newline if multiline, or before a terminating
1333      newline unless endonly is set, else end of subject unless noteol is set. */      newline unless endonly is set, else end of subject unless noteol is set. */
1334    
# Line 1244  for (;;) Line 1348  for (;;)
1348        if (!md->endonly)        if (!md->endonly)
1349          {          {
1350          if (eptr != md->end_subject &&          if (eptr != md->end_subject &&
1351              (eptr != md->end_subject - md->nllen || !IS_NEWLINE(eptr)))              (!IS_NEWLINE(eptr) || eptr != md->end_subject - md->nllen))
1352            RRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
1353          ecode++;          ecode++;
1354          break;          break;
# Line 1263  for (;;) Line 1367  for (;;)
1367    
1368      case OP_EODN:      case OP_EODN:
1369      if (eptr != md->end_subject &&      if (eptr != md->end_subject &&
1370          (eptr != md->end_subject - md->nllen || !IS_NEWLINE(eptr)))          (!IS_NEWLINE(eptr) || eptr != md->end_subject - md->nllen))
1371        RRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
1372      ecode++;      ecode++;
1373      break;      break;
# Line 1319  for (;;) Line 1423  for (;;)
1423      case OP_ANY:      case OP_ANY:
1424      if ((ims & PCRE_DOTALL) == 0)      if ((ims & PCRE_DOTALL) == 0)
1425        {        {
1426        if (eptr <= md->end_subject - md->nllen && IS_NEWLINE(eptr))        if (IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH);
         RRETURN(MATCH_NOMATCH);  
1427        }        }
1428      if (eptr++ >= md->end_subject) RRETURN(MATCH_NOMATCH);      if (eptr++ >= md->end_subject) RRETURN(MATCH_NOMATCH);
1429      if (utf8)      if (utf8)
# Line 1414  for (;;) Line 1517  for (;;)
1517      ecode++;      ecode++;
1518      break;      break;
1519    
1520        case OP_ANYNL:
1521        if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
1522        GETCHARINCTEST(c, eptr);
1523        switch(c)
1524          {
1525          default: RRETURN(MATCH_NOMATCH);
1526          case 0x000d:
1527          if (eptr < md->end_subject && *eptr == 0x0a) eptr++;
1528          break;
1529          case 0x000a:
1530          case 0x000b:
1531          case 0x000c:
1532          case 0x0085:
1533          case 0x2028:
1534          case 0x2029:
1535          break;
1536          }
1537        ecode++;
1538        break;
1539    
1540        case OP_NOT_HSPACE:
1541        if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
1542        GETCHARINCTEST(c, eptr);
1543        switch(c)
1544          {
1545          default: break;
1546          case 0x09:      /* HT */
1547          case 0x20:      /* SPACE */
1548          case 0xa0:      /* NBSP */
1549          case 0x1680:    /* OGHAM SPACE MARK */
1550          case 0x180e:    /* MONGOLIAN VOWEL SEPARATOR */
1551          case 0x2000:    /* EN QUAD */
1552          case 0x2001:    /* EM QUAD */
1553          case 0x2002:    /* EN SPACE */
1554          case 0x2003:    /* EM SPACE */
1555          case 0x2004:    /* THREE-PER-EM SPACE */
1556          case 0x2005:    /* FOUR-PER-EM SPACE */
1557          case 0x2006:    /* SIX-PER-EM SPACE */
1558          case 0x2007:    /* FIGURE SPACE */
1559          case 0x2008:    /* PUNCTUATION SPACE */
1560          case 0x2009:    /* THIN SPACE */
1561          case 0x200A:    /* HAIR SPACE */
1562          case 0x202f:    /* NARROW NO-BREAK SPACE */
1563          case 0x205f:    /* MEDIUM MATHEMATICAL SPACE */
1564          case 0x3000:    /* IDEOGRAPHIC SPACE */
1565          RRETURN(MATCH_NOMATCH);
1566          }
1567        ecode++;
1568        break;
1569    
1570        case OP_HSPACE:
1571        if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
1572        GETCHARINCTEST(c, eptr);
1573        switch(c)
1574          {
1575          default: RRETURN(MATCH_NOMATCH);
1576          case 0x09:      /* HT */
1577          case 0x20:      /* SPACE */
1578          case 0xa0:      /* NBSP */
1579          case 0x1680:    /* OGHAM SPACE MARK */
1580          case 0x180e:    /* MONGOLIAN VOWEL SEPARATOR */
1581          case 0x2000:    /* EN QUAD */
1582          case 0x2001:    /* EM QUAD */
1583          case 0x2002:    /* EN SPACE */
1584          case 0x2003:    /* EM SPACE */
1585          case 0x2004:    /* THREE-PER-EM SPACE */
1586          case 0x2005:    /* FOUR-PER-EM SPACE */
1587          case 0x2006:    /* SIX-PER-EM SPACE */
1588          case 0x2007:    /* FIGURE SPACE */
1589          case 0x2008:    /* PUNCTUATION SPACE */
1590          case 0x2009:    /* THIN SPACE */
1591          case 0x200A:    /* HAIR SPACE */
1592          case 0x202f:    /* NARROW NO-BREAK SPACE */
1593          case 0x205f:    /* MEDIUM MATHEMATICAL SPACE */
1594          case 0x3000:    /* IDEOGRAPHIC SPACE */
1595          break;
1596          }
1597        ecode++;
1598        break;
1599    
1600        case OP_NOT_VSPACE:
1601        if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
1602        GETCHARINCTEST(c, eptr);
1603        switch(c)
1604          {
1605          default: break;
1606          case 0x0a:      /* LF */
1607          case 0x0b:      /* VT */
1608          case 0x0c:      /* FF */
1609          case 0x0d:      /* CR */
1610          case 0x85:      /* NEL */
1611          case 0x2028:    /* LINE SEPARATOR */
1612          case 0x2029:    /* PARAGRAPH SEPARATOR */
1613          RRETURN(MATCH_NOMATCH);
1614          }
1615        ecode++;
1616        break;
1617    
1618        case OP_VSPACE:
1619        if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
1620        GETCHARINCTEST(c, eptr);
1621        switch(c)
1622          {
1623          default: RRETURN(MATCH_NOMATCH);
1624          case 0x0a:      /* LF */
1625          case 0x0b:      /* VT */
1626          case 0x0c:      /* FF */
1627          case 0x0d:      /* CR */
1628          case 0x85:      /* NEL */
1629          case 0x2028:    /* LINE SEPARATOR */
1630          case 0x2029:    /* PARAGRAPH SEPARATOR */
1631          break;
1632          }
1633        ecode++;
1634        break;
1635    
1636  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
1637      /* Check the next character by Unicode property. We will get here only      /* Check the next character by Unicode property. We will get here only
1638      if the support is in the binary; otherwise a compile-time error occurs. */      if the support is in the binary; otherwise a compile-time error occurs. */
# Line 1456  for (;;) Line 1675  for (;;)
1675    
1676          default:          default:
1677          RRETURN(PCRE_ERROR_INTERNAL);          RRETURN(PCRE_ERROR_INTERNAL);
         break;  
1678          }          }
1679    
1680        ecode += 3;        ecode += 3;
# Line 1570  for (;;) Line 1788  for (;;)
1788          {          {
1789          for (fi = min;; fi++)          for (fi = min;; fi++)
1790            {            {
1791            RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM14);
1792            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
1793            if (fi >= max || !match_ref(offset, eptr, length, md, ims))            if (fi >= max || !match_ref(offset, eptr, length, md, ims))
1794              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
# Line 1591  for (;;) Line 1809  for (;;)
1809            }            }
1810          while (eptr >= pp)          while (eptr >= pp)
1811            {            {
1812            RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM15);
1813            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
1814            eptr -= length;            eptr -= length;
1815            }            }
# Line 1696  for (;;) Line 1914  for (;;)
1914            {            {
1915            for (fi = min;; fi++)            for (fi = min;; fi++)
1916              {              {
1917              RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM16);
1918              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
1919              if (fi >= max || eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);              if (fi >= max || eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
1920              GETCHARINC(c, eptr);              GETCHARINC(c, eptr);
# Line 1716  for (;;) Line 1934  for (;;)
1934            {            {
1935            for (fi = min;; fi++)            for (fi = min;; fi++)
1936              {              {
1937              RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM17);
1938              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
1939              if (fi >= max || eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);              if (fi >= max || eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
1940              c = *eptr++;              c = *eptr++;
# Line 1753  for (;;) Line 1971  for (;;)
1971              }              }
1972            for (;;)            for (;;)
1973              {              {
1974              RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM18);
1975              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
1976              if (eptr-- == pp) break;        /* Stop if tried at original pos */              if (eptr-- == pp) break;        /* Stop if tried at original pos */
1977              BACKCHAR(eptr);              BACKCHAR(eptr);
# Line 1772  for (;;) Line 1990  for (;;)
1990              }              }
1991            while (eptr >= pp)            while (eptr >= pp)
1992              {              {
1993              RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM19);
1994              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
1995              eptr--;              eptr--;
1996              }              }
# Line 1843  for (;;) Line 2061  for (;;)
2061          {          {
2062          for (fi = min;; fi++)          for (fi = min;; fi++)
2063            {            {
2064            RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM20);
2065            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2066            if (fi >= max || eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);            if (fi >= max || eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
2067            GETCHARINC(c, eptr);            GETCHARINC(c, eptr);
# Line 1867  for (;;) Line 2085  for (;;)
2085            }            }
2086          for(;;)          for(;;)
2087            {            {
2088            RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM21);
2089            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2090            if (eptr-- == pp) break;        /* Stop if tried at original pos */            if (eptr-- == pp) break;        /* Stop if tried at original pos */
2091            BACKCHAR(eptr)            if (utf8) BACKCHAR(eptr);
2092            }            }
2093          RRETURN(MATCH_NOMATCH);          RRETURN(MATCH_NOMATCH);
2094          }          }
# Line 1926  for (;;) Line 2144  for (;;)
2144    
2145        else        else
2146          {          {
2147          int dc;          unsigned int dc;
2148          GETCHARINC(dc, eptr);          GETCHARINC(dc, eptr);
2149          ecode += length;          ecode += length;
2150    
# Line 1953  for (;;) Line 2171  for (;;)
2171        }        }
2172      break;      break;
2173    
2174      /* Match a single character repeatedly; different opcodes share code. */      /* Match a single character repeatedly. */
2175    
2176      case OP_EXACT:      case OP_EXACT:
2177      min = max = GET2(ecode, 1);      min = max = GET2(ecode, 1);
2178      ecode += 3;      ecode += 3;
2179      goto REPEATCHAR;      goto REPEATCHAR;
2180    
2181        case OP_POSUPTO:
2182        possessive = TRUE;
2183        /* Fall through */
2184    
2185      case OP_UPTO:      case OP_UPTO:
2186      case OP_MINUPTO:      case OP_MINUPTO:
2187      min = 0;      min = 0;
# Line 1968  for (;;) Line 2190  for (;;)
2190      ecode += 3;      ecode += 3;
2191      goto REPEATCHAR;      goto REPEATCHAR;
2192    
2193        case OP_POSSTAR:
2194        possessive = TRUE;
2195        min = 0;
2196        max = INT_MAX;
2197        ecode++;
2198        goto REPEATCHAR;
2199    
2200        case OP_POSPLUS:
2201        possessive = TRUE;
2202        min = 1;
2203        max = INT_MAX;
2204        ecode++;
2205        goto REPEATCHAR;
2206    
2207        case OP_POSQUERY:
2208        possessive = TRUE;
2209        min = 0;
2210        max = 1;
2211        ecode++;
2212        goto REPEATCHAR;
2213    
2214      case OP_STAR:      case OP_STAR:
2215      case OP_MINSTAR:      case OP_MINSTAR:
2216      case OP_PLUS:      case OP_PLUS:
# Line 1999  for (;;) Line 2242  for (;;)
2242    
2243        if (length > 1)        if (length > 1)
2244          {          {
         int oclength = 0;  
         uschar occhars[8];  
   
2245  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
2246          int othercase;          unsigned int othercase;
2247          if ((ims & PCRE_CASELESS) != 0 &&          if ((ims & PCRE_CASELESS) != 0 &&
2248              (othercase = _pcre_ucp_othercase(fc)) >= 0 &&              (othercase = _pcre_ucp_othercase(fc)) != NOTACHAR)
              othercase >= 0)  
2249            oclength = _pcre_ord2utf8(othercase, occhars);            oclength = _pcre_ord2utf8(othercase, occhars);
2250            else oclength = 0;
2251  #endif  /* SUPPORT_UCP */  #endif  /* SUPPORT_UCP */
2252    
2253          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
2254            {            {
2255            if (memcmp(eptr, charptr, length) == 0) eptr += length;            if (memcmp(eptr, charptr, length) == 0) eptr += length;
2256    #ifdef SUPPORT_UCP
2257            /* Need braces because of following else */            /* Need braces because of following else */
2258            else if (oclength == 0) { RRETURN(MATCH_NOMATCH); }            else if (oclength == 0) { RRETURN(MATCH_NOMATCH); }
2259            else            else
# Line 2020  for (;;) Line 2261  for (;;)
2261              if (memcmp(eptr, occhars, oclength) != 0) RRETURN(MATCH_NOMATCH);              if (memcmp(eptr, occhars, oclength) != 0) RRETURN(MATCH_NOMATCH);
2262              eptr += oclength;              eptr += oclength;
2263              }              }
2264    #else   /* without SUPPORT_UCP */
2265              else { RRETURN(MATCH_NOMATCH); }
2266    #endif  /* SUPPORT_UCP */
2267            }            }
2268    
2269          if (min == max) continue;          if (min == max) continue;
# Line 2028  for (;;) Line 2272  for (;;)
2272            {            {
2273            for (fi = min;; fi++)            for (fi = min;; fi++)
2274              {              {
2275              RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM22);
2276              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2277              if (fi >= max || eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);              if (fi >= max || eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
2278              if (memcmp(eptr, charptr, length) == 0) eptr += length;              if (memcmp(eptr, charptr, length) == 0) eptr += length;
2279    #ifdef SUPPORT_UCP
2280              /* Need braces because of following else */              /* Need braces because of following else */
2281              else if (oclength == 0) { RRETURN(MATCH_NOMATCH); }              else if (oclength == 0) { RRETURN(MATCH_NOMATCH); }
2282              else              else
# Line 2039  for (;;) Line 2284  for (;;)
2284                if (memcmp(eptr, occhars, oclength) != 0) RRETURN(MATCH_NOMATCH);                if (memcmp(eptr, occhars, oclength) != 0) RRETURN(MATCH_NOMATCH);
2285                eptr += oclength;                eptr += oclength;
2286                }                }
2287    #else   /* without SUPPORT_UCP */
2288                else { RRETURN (MATCH_NOMATCH); }
2289    #endif  /* SUPPORT_UCP */
2290              }              }
2291            /* Control never gets here */            /* Control never gets here */
2292            }            }
2293          else  
2294            else  /* Maximize */
2295            {            {
2296            pp = eptr;            pp = eptr;
2297            for (i = min; i < max; i++)            for (i = min; i < max; i++)
2298              {              {
2299              if (eptr > md->end_subject - length) break;              if (eptr > md->end_subject - length) break;
2300              if (memcmp(eptr, charptr, length) == 0) eptr += length;              if (memcmp(eptr, charptr, length) == 0) eptr += length;
2301    #ifdef SUPPORT_UCP
2302              else if (oclength == 0) break;              else if (oclength == 0) break;
2303              else              else
2304                {                {
2305                if (memcmp(eptr, occhars, oclength) != 0) break;                if (memcmp(eptr, occhars, oclength) != 0) break;
2306                eptr += oclength;                eptr += oclength;
2307                }                }
2308    #else   /* without SUPPORT_UCP */
2309                else break;
2310    #endif  /* SUPPORT_UCP */
2311              }              }
2312            while (eptr >= pp)  
2313              if (possessive) continue;
2314              for(;;)
2315             {             {
2316             RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);             RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM23);
2317             if (rrc != MATCH_NOMATCH) RRETURN(rrc);             if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2318               if (eptr == pp) RRETURN(MATCH_NOMATCH);
2319    #ifdef SUPPORT_UCP
2320               eptr--;
2321               BACKCHAR(eptr);
2322    #else   /* without SUPPORT_UCP */
2323             eptr -= length;             eptr -= length;
2324    #endif  /* SUPPORT_UCP */
2325             }             }
           RRETURN(MATCH_NOMATCH);  
2326            }            }
2327          /* Control never gets here */          /* Control never gets here */
2328          }          }
# Line 2102  for (;;) Line 2362  for (;;)
2362          {          {
2363          for (fi = min;; fi++)          for (fi = min;; fi++)
2364            {            {
2365            RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM24);
2366            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2367            if (fi >= max || eptr >= md->end_subject ||            if (fi >= max || eptr >= md->end_subject ||
2368                fc != md->lcc[*eptr++])                fc != md->lcc[*eptr++])
# Line 2110  for (;;) Line 2370  for (;;)
2370            }            }
2371          /* Control never gets here */          /* Control never gets here */
2372          }          }
2373        else        else  /* Maximize */
2374          {          {
2375          pp = eptr;          pp = eptr;
2376          for (i = min; i < max; i++)          for (i = min; i < max; i++)
# Line 2118  for (;;) Line 2378  for (;;)
2378            if (eptr >= md->end_subject || fc != md->lcc[*eptr]) break;            if (eptr >= md->end_subject || fc != md->lcc[*eptr]) break;
2379            eptr++;            eptr++;
2380            }            }
2381            if (possessive) continue;
2382          while (eptr >= pp)          while (eptr >= pp)
2383            {            {
2384            RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM25);
2385            eptr--;            eptr--;
2386            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2387            }            }
# Line 2139  for (;;) Line 2400  for (;;)
2400          {          {
2401          for (fi = min;; fi++)          for (fi = min;; fi++)
2402            {            {
2403            RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM26);
2404            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2405            if (fi >= max || eptr >= md->end_subject || fc != *eptr++)            if (fi >= max || eptr >= md->end_subject || fc != *eptr++)
2406              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
2407            }            }
2408          /* Control never gets here */          /* Control never gets here */
2409          }          }
2410        else        else  /* Maximize */
2411          {          {
2412          pp = eptr;          pp = eptr;
2413          for (i = min; i < max; i++)          for (i = min; i < max; i++)
# Line 2154  for (;;) Line 2415  for (;;)
2415            if (eptr >= md->end_subject || fc != *eptr) break;            if (eptr >= md->end_subject || fc != *eptr) break;
2416            eptr++;            eptr++;
2417            }            }
2418            if (possessive) continue;
2419          while (eptr >= pp)          while (eptr >= pp)
2420            {            {
2421            RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM27);
2422            eptr--;            eptr--;
2423            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2424            }            }
# Line 2206  for (;;) Line 2468  for (;;)
2468      ecode += 3;      ecode += 3;
2469      goto REPEATNOTCHAR;      goto REPEATNOTCHAR;
2470    
2471        case OP_NOTPOSSTAR:
2472        possessive = TRUE;
2473        min = 0;
2474        max = INT_MAX;
2475        ecode++;
2476        goto REPEATNOTCHAR;
2477    
2478        case OP_NOTPOSPLUS:
2479        possessive = TRUE;
2480        min = 1;
2481        max = INT_MAX;
2482        ecode++;
2483        goto REPEATNOTCHAR;
2484    
2485        case OP_NOTPOSQUERY:
2486        possessive = TRUE;
2487        min = 0;
2488        max = 1;
2489        ecode++;
2490        goto REPEATNOTCHAR;
2491    
2492        case OP_NOTPOSUPTO:
2493        possessive = TRUE;
2494        min = 0;
2495        max = GET2(ecode, 1);
2496        ecode += 3;
2497        goto REPEATNOTCHAR;
2498    
2499      case OP_NOTSTAR:      case OP_NOTSTAR:
2500      case OP_NOTMINSTAR:      case OP_NOTMINSTAR:
2501      case OP_NOTPLUS:      case OP_NOTPLUS:
# Line 2245  for (;;) Line 2535  for (;;)
2535        /* UTF-8 mode */        /* UTF-8 mode */
2536        if (utf8)        if (utf8)
2537          {          {
2538          register int d;          register unsigned int d;
2539          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
2540            {            {
2541            GETCHARINC(d, eptr);            GETCHARINC(d, eptr);
# Line 2270  for (;;) Line 2560  for (;;)
2560          /* UTF-8 mode */          /* UTF-8 mode */
2561          if (utf8)          if (utf8)
2562            {            {
2563            register int d;            register unsigned int d;
2564            for (fi = min;; fi++)            for (fi = min;; fi++)
2565              {              {
2566              RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM28);
2567              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2568              GETCHARINC(d, eptr);              GETCHARINC(d, eptr);
2569              if (d < 256) d = md->lcc[d];              if (d < 256) d = md->lcc[d];
# Line 2287  for (;;) Line 2577  for (;;)
2577            {            {
2578            for (fi = min;; fi++)            for (fi = min;; fi++)
2579              {              {
2580              RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM29);
2581              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2582              if (fi >= max || eptr >= md->end_subject || fc == md->lcc[*eptr++])              if (fi >= max || eptr >= md->end_subject || fc == md->lcc[*eptr++])
2583                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
# Line 2306  for (;;) Line 2596  for (;;)
2596          /* UTF-8 mode */          /* UTF-8 mode */
2597          if (utf8)          if (utf8)
2598            {            {
2599            register int d;            register unsigned int d;
2600            for (i = min; i < max; i++)            for (i = min; i < max; i++)
2601              {              {
2602              int len = 1;              int len = 1;
# Line 2316  for (;;) Line 2606  for (;;)
2606              if (fc == d) break;              if (fc == d) break;
2607              eptr += len;              eptr += len;
2608              }              }
2609            for(;;)          if (possessive) continue;
2610            for(;;)
2611              {              {
2612              RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM30);
2613              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2614              if (eptr-- == pp) break;        /* Stop if tried at original pos */              if (eptr-- == pp) break;        /* Stop if tried at original pos */
2615              BACKCHAR(eptr);              BACKCHAR(eptr);
# Line 2333  for (;;) Line 2624  for (;;)
2624              if (eptr >= md->end_subject || fc == md->lcc[*eptr]) break;              if (eptr >= md->end_subject || fc == md->lcc[*eptr]) break;
2625              eptr++;              eptr++;
2626              }              }
2627              if (possessive) continue;
2628            while (eptr >= pp)            while (eptr >= pp)
2629              {              {
2630              RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM31);
2631              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2632              eptr--;              eptr--;
2633              }              }
# Line 2354  for (;;) Line 2646  for (;;)
2646        /* UTF-8 mode */        /* UTF-8 mode */
2647        if (utf8)        if (utf8)
2648          {          {
2649          register int d;          register unsigned int d;
2650          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
2651            {            {
2652            GETCHARINC(d, eptr);            GETCHARINC(d, eptr);
# Line 2377  for (;;) Line 2669  for (;;)
2669          /* UTF-8 mode */          /* UTF-8 mode */
2670          if (utf8)          if (utf8)
2671            {            {
2672            register int d;            register unsigned int d;
2673            for (fi = min;; fi++)            for (fi = min;; fi++)
2674              {              {
2675              RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM32);
2676              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2677              GETCHARINC(d, eptr);              GETCHARINC(d, eptr);
2678              if (fi >= max || eptr >= md->end_subject || fc == d)              if (fi >= max || eptr >= md->end_subject || fc == d)
# Line 2393  for (;;) Line 2685  for (;;)
2685            {            {
2686            for (fi = min;; fi++)            for (fi = min;; fi++)
2687              {              {
2688              RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM33);
2689              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2690              if (fi >= max || eptr >= md->end_subject || fc == *eptr++)              if (fi >= max || eptr >= md->end_subject || fc == *eptr++)
2691                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
# Line 2412  for (;;) Line 2704  for (;;)
2704          /* UTF-8 mode */          /* UTF-8 mode */
2705          if (utf8)          if (utf8)
2706            {            {
2707            register int d;            register unsigned int d;
2708            for (i = min; i < max; i++)            for (i = min; i < max; i++)
2709              {              {
2710              int len = 1;              int len = 1;
# Line 2421  for (;;) Line 2713  for (;;)
2713              if (fc == d) break;              if (fc == d) break;
2714              eptr += len;              eptr += len;
2715              }              }
2716              if (possessive) continue;
2717            for(;;)            for(;;)
2718              {              {
2719              RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM34);
2720              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2721              if (eptr-- == pp) break;        /* Stop if tried at original pos */              if (eptr-- == pp) break;        /* Stop if tried at original pos */
2722              BACKCHAR(eptr);              BACKCHAR(eptr);
# Line 2438  for (;;) Line 2731  for (;;)
2731              if (eptr >= md->end_subject || fc == *eptr) break;              if (eptr >= md->end_subject || fc == *eptr) break;
2732              eptr++;              eptr++;
2733              }              }
2734              if (possessive) continue;
2735            while (eptr >= pp)            while (eptr >= pp)
2736              {              {
2737              RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM35);
2738              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2739              eptr--;              eptr--;
2740              }              }
# Line 2469  for (;;) Line 2763  for (;;)
2763      ecode += 3;      ecode += 3;
2764      goto REPEATTYPE;      goto REPEATTYPE;
2765    
2766        case OP_TYPEPOSSTAR:
2767        possessive = TRUE;
2768        min = 0;
2769        max = INT_MAX;
2770        ecode++;
2771        goto REPEATTYPE;
2772    
2773        case OP_TYPEPOSPLUS:
2774        possessive = TRUE;
2775        min = 1;
2776        max = INT_MAX;
2777        ecode++;
2778        goto REPEATTYPE;
2779    
2780        case OP_TYPEPOSQUERY:
2781        possessive = TRUE;
2782        min = 0;
2783        max = 1;
2784        ecode++;
2785        goto REPEATTYPE;
2786    
2787        case OP_TYPEPOSUPTO:
2788        possessive = TRUE;
2789        min = 0;
2790        max = GET2(ecode, 1);
2791        ecode += 3;
2792        goto REPEATTYPE;
2793    
2794      case OP_TYPESTAR:      case OP_TYPESTAR:
2795      case OP_TYPEMINSTAR:      case OP_TYPEMINSTAR:
2796      case OP_TYPEPLUS:      case OP_TYPEPLUS:
# Line 2519  for (;;) Line 2841  for (;;)
2841            for (i = 1; i <= min; i++)            for (i = 1; i <= min; i++)
2842              {              {
2843              if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);              if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
2844              GETCHARINC(c, eptr);              GETCHARINCTEST(c, eptr);
2845              }              }
2846            break;            break;
2847    
# Line 2527  for (;;) Line 2849  for (;;)
2849            for (i = 1; i <= min; i++)            for (i = 1; i <= min; i++)
2850              {              {
2851              if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);              if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
2852              GETCHARINC(c, eptr);              GETCHARINCTEST(c, eptr);
2853              prop_category = _pcre_ucp_findprop(c, &prop_chartype, &prop_script);              prop_category = _pcre_ucp_findprop(c, &prop_chartype, &prop_script);
2854              if ((prop_chartype == ucp_Lu ||              if ((prop_chartype == ucp_Lu ||
2855                   prop_chartype == ucp_Ll ||                   prop_chartype == ucp_Ll ||
# Line 2540  for (;;) Line 2862  for (;;)
2862            for (i = 1; i <= min; i++)            for (i = 1; i <= min; i++)
2863              {              {
2864              if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);              if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
2865              GETCHARINC(c, eptr);              GETCHARINCTEST(c, eptr);
2866              prop_category = _pcre_ucp_findprop(c, &prop_chartype, &prop_script);              prop_category = _pcre_ucp_findprop(c, &prop_chartype, &prop_script);
2867              if ((prop_category == prop_value) == prop_fail_result)              if ((prop_category == prop_value) == prop_fail_result)
2868                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
# Line 2551  for (;;) Line 2873  for (;;)
2873            for (i = 1; i <= min; i++)            for (i = 1; i <= min; i++)
2874              {              {
2875              if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);              if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
2876              GETCHARINC(c, eptr);              GETCHARINCTEST(c, eptr);
2877              prop_category = _pcre_ucp_findprop(c, &prop_chartype, &prop_script);              prop_category = _pcre_ucp_findprop(c, &prop_chartype, &prop_script);
2878              if ((prop_chartype == prop_value) == prop_fail_result)              if ((prop_chartype == prop_value) == prop_fail_result)
2879                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
# Line 2562  for (;;) Line 2884  for (;;)
2884            for (i = 1; i <= min; i++)            for (i = 1; i <= min; i++)
2885              {              {
2886              if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);              if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
2887              GETCHARINC(c, eptr);              GETCHARINCTEST(c, eptr);
2888              prop_category = _pcre_ucp_findprop(c, &prop_chartype, &prop_script);              prop_category = _pcre_ucp_findprop(c, &prop_chartype, &prop_script);
2889              if ((prop_script == prop_value) == prop_fail_result)              if ((prop_script == prop_value) == prop_fail_result)
2890                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
# Line 2571  for (;;) Line 2893  for (;;)
2893    
2894            default:            default:
2895            RRETURN(PCRE_ERROR_INTERNAL);            RRETURN(PCRE_ERROR_INTERNAL);
           break;  
2896            }            }
2897          }          }
2898    
# Line 2611  for (;;) Line 2932  for (;;)
2932          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
2933            {            {
2934            if (eptr >= md->end_subject ||            if (eptr >= md->end_subject ||
2935                 ((ims & PCRE_DOTALL) == 0 &&                 ((ims & PCRE_DOTALL) == 0 && IS_NEWLINE(eptr)))
                  eptr <= md->end_subject - md->nllen &&  
                  IS_NEWLINE(eptr)))  
2936              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
2937            eptr++;            eptr++;
2938            while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;            while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;
# Line 2624  for (;;) Line 2943  for (;;)
2943          eptr += min;          eptr += min;
2944          break;          break;
2945    
2946          case OP_NOT_DIGIT:          case OP_ANYNL:
2947          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
2948            {            {
2949            if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);            if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
2950            GETCHARINC(c, eptr);            GETCHARINC(c, eptr);
2951            if (c < 128 && (md->ctypes[c] & ctype_digit) != 0)            switch(c)
2952              RRETURN(MATCH_NOMATCH);              {
2953                default: RRETURN(MATCH_NOMATCH);
2954                case 0x000d:
2955                if (eptr < md->end_subject && *eptr == 0x0a) eptr++;
2956                break;
2957                case 0x000a:
2958                case 0x000b:
2959                case 0x000c:
2960                case 0x0085:
2961                case 0x2028:
2962                case 0x2029:
2963                break;
2964                }
2965            }            }
2966          break;          break;
2967    
2968          case OP_DIGIT:          case OP_NOT_HSPACE:
2969          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
2970            {            {
2971            if (eptr >= md->end_subject ||            if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
2972               *eptr >= 128 || (md->ctypes[*eptr++] & ctype_digit) == 0)            GETCHARINC(c, eptr);
2973              switch(c)
2974                {
2975                default: break;
2976                case 0x09:      /* HT */
2977                case 0x20:      /* SPACE */
2978                case 0xa0:      /* NBSP */
2979                case 0x1680:    /* OGHAM SPACE MARK */
2980                case 0x180e:    /* MONGOLIAN VOWEL SEPARATOR */
2981                case 0x2000:    /* EN QUAD */
2982                case 0x2001:    /* EM QUAD */
2983                case 0x2002:    /* EN SPACE */
2984                case 0x2003:    /* EM SPACE */
2985                case 0x2004:    /* THREE-PER-EM SPACE */
2986                case 0x2005:    /* FOUR-PER-EM SPACE */
2987                case 0x2006:    /* SIX-PER-EM SPACE */
2988                case 0x2007:    /* FIGURE SPACE */
2989                case 0x2008:    /* PUNCTUATION SPACE */
2990                case 0x2009:    /* THIN SPACE */
2991                case 0x200A:    /* HAIR SPACE */
2992                case 0x202f:    /* NARROW NO-BREAK SPACE */
2993                case 0x205f:    /* MEDIUM MATHEMATICAL SPACE */
2994                case 0x3000:    /* IDEOGRAPHIC SPACE */
2995              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
2996            /* No need to skip more bytes - we know it's a 1-byte character */              }
2997            }            }
2998          break;          break;
2999    
3000          case OP_NOT_WHITESPACE:          case OP_HSPACE:
3001          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
3002            {            {
3003            if (eptr >= md->end_subject ||            if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
3004               (*eptr < 128 && (md->ctypes[*eptr++] & ctype_space) != 0))            GETCHARINC(c, eptr);
3005              RRETURN(MATCH_NOMATCH);            switch(c)
3006            while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;              {
3007                default: RRETURN(MATCH_NOMATCH);
3008                case 0x09:      /* HT */
3009                case 0x20:      /* SPACE */
3010                case 0xa0:      /* NBSP */
3011                case 0x1680:    /* OGHAM SPACE MARK */
3012                case 0x180e:    /* MONGOLIAN VOWEL SEPARATOR */
3013                case 0x2000:    /* EN QUAD */
3014                case 0x2001:    /* EM QUAD */
3015                case 0x2002:    /* EN SPACE */
3016                case 0x2003:    /* EM SPACE */
3017                case 0x2004:    /* THREE-PER-EM SPACE */
3018                case 0x2005:    /* FOUR-PER-EM SPACE */
3019                case 0x2006:    /* SIX-PER-EM SPACE */
3020                case 0x2007:    /* FIGURE SPACE */
3021                case 0x2008:    /* PUNCTUATION SPACE */
3022                case 0x2009:    /* THIN SPACE */
3023                case 0x200A:    /* HAIR SPACE */
3024                case 0x202f:    /* NARROW NO-BREAK SPACE */
3025                case 0x205f:    /* MEDIUM MATHEMATICAL SPACE */
3026                case 0x3000:    /* IDEOGRAPHIC SPACE */
3027                break;
3028                }
3029            }            }
3030          break;          break;
3031    
3032          case OP_WHITESPACE:          case OP_NOT_VSPACE:
3033          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
3034            {            {
3035            if (eptr >= md->end_subject ||            if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
3036               *eptr >= 128 || (md->ctypes[*eptr++] & ctype_space) == 0)            GETCHARINC(c, eptr);
3037              switch(c)
3038                {
3039                default: break;
3040                case 0x0a:      /* LF */
3041                case 0x0b:      /* VT */
3042                case 0x0c:      /* FF */
3043                case 0x0d:      /* CR */
3044                case 0x85:      /* NEL */
3045                case 0x2028:    /* LINE SEPARATOR */
3046                case 0x2029:    /* PARAGRAPH SEPARATOR */
3047              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
3048            /* No need to skip more bytes - we know it's a 1-byte character */              }
3049            }            }
3050          break;          break;
3051    
3052          case OP_NOT_WORDCHAR:          case OP_VSPACE:
3053          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
3054            {            {
3055            if (eptr >= md->end_subject ||            if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
3056               (*eptr < 128 && (md->ctypes[*eptr++] & ctype_word) != 0))            GETCHARINC(c, eptr);
3057              RRETURN(MATCH_NOMATCH);            switch(c)
3058            while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;              {
3059                default: RRETURN(MATCH_NOMATCH);
3060                case 0x0a:      /* LF */
3061                case 0x0b:      /* VT */
3062                case 0x0c:      /* FF */
3063                case 0x0d:      /* CR */
3064                case 0x85:      /* NEL */
3065                case 0x2028:    /* LINE SEPARATOR */
3066                case 0x2029:    /* PARAGRAPH SEPARATOR */
3067                break;
3068                }
3069              }
3070            break;
3071    
3072            case OP_NOT_DIGIT:
3073            for (i = 1; i <= min; i++)
3074              {
3075              if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
3076              GETCHARINC(c, eptr);
3077              if (c < 128 && (md->ctypes[c] & ctype_digit) != 0)
3078                RRETURN(MATCH_NOMATCH);
3079              }
3080            break;
3081    
3082            case OP_DIGIT:
3083            for (i = 1; i <= min; i++)
3084              {
3085              if (eptr >= md->end_subject ||
3086                 *eptr >= 128 || (md->ctypes[*eptr++] & ctype_digit) == 0)
3087                RRETURN(MATCH_NOMATCH);
3088              /* No need to skip more bytes - we know it's a 1-byte character */
3089              }
3090            break;
3091    
3092            case OP_NOT_WHITESPACE:
3093            for (i = 1; i <= min; i++)
3094              {
3095              if (eptr >= md->end_subject ||
3096                 (*eptr < 128 && (md->ctypes[*eptr] & ctype_space) != 0))
3097                RRETURN(MATCH_NOMATCH);
3098              while (++eptr < md->end_subject && (*eptr & 0xc0) == 0x80);
3099              }
3100            break;
3101    
3102            case OP_WHITESPACE:
3103            for (i = 1; i <= min; i++)
3104              {
3105              if (eptr >= md->end_subject ||
3106                 *eptr >= 128 || (md->ctypes[*eptr++] & ctype_space) == 0)
3107                RRETURN(MATCH_NOMATCH);
3108              /* No need to skip more bytes - we know it's a 1-byte character */
3109              }
3110            break;
3111    
3112            case OP_NOT_WORDCHAR:
3113            for (i = 1; i <= min; i++)
3114              {
3115              if (eptr >= md->end_subject ||
3116                 (*eptr < 128 && (md->ctypes[*eptr] & ctype_word) != 0))
3117                RRETURN(MATCH_NOMATCH);
3118              while (++eptr < md->end_subject && (*eptr & 0xc0) == 0x80);
3119            }            }
3120          break;          break;
3121    
# Line 2692  for (;;) Line 3137  for (;;)
3137  #endif     /* SUPPORT_UTF8 */  #endif     /* SUPPORT_UTF8 */
3138    
3139        /* Code for the non-UTF-8 case for minimum matching of operators other        /* Code for the non-UTF-8 case for minimum matching of operators other
3140        than OP_PROP and OP_NOTPROP. */        than OP_PROP and OP_NOTPROP. We can assume that there are the minimum
3141          number of bytes present, as this was tested above. */
3142    
3143        switch(ctype)        switch(ctype)
3144          {          {
# Line 2701  for (;;) Line 3147  for (;;)
3147            {            {
3148            for (i = 1; i <= min; i++)            for (i = 1; i <= min; i++)
3149              {              {
3150              if (eptr <= md->end_subject - md->nllen && IS_NEWLINE(eptr))              if (IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH);
               RRETURN(MATCH_NOMATCH);  
3151              eptr++;              eptr++;
3152              }              }
3153            }            }
# Line 2713  for (;;) Line 3158  for (;;)
3158          eptr += min;          eptr += min;
3159          break;          break;
3160    
3161            /* Because of the CRLF case, we can't assume the minimum number of
3162            bytes are present in this case. */
3163    
3164            case OP_ANYNL:
3165            for (i = 1; i <= min; i++)
3166              {
3167              if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
3168              switch(*eptr++)
3169                {
3170                default: RRETURN(MATCH_NOMATCH);
3171                case 0x000d:
3172                if (eptr < md->end_subject && *eptr == 0x0a) eptr++;
3173                break;
3174                case 0x000a:
3175                case 0x000b:
3176                case 0x000c:
3177                case 0x0085:
3178                break;
3179                }
3180              }
3181            break;
3182    
3183            case OP_NOT_HSPACE:
3184            for (i = 1; i <= min; i++)
3185              {
3186              if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
3187              switch(*eptr++)
3188                {
3189                default: break;
3190                case 0x09:      /* HT */
3191                case 0x20:      /* SPACE */
3192                case 0xa0:      /* NBSP */
3193                RRETURN(MATCH_NOMATCH);
3194                }
3195              }
3196            break;
3197    
3198            case OP_HSPACE:
3199            for (i = 1; i <= min; i++)
3200              {
3201              if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
3202              switch(*eptr++)
3203                {
3204                default: RRETURN(MATCH_NOMATCH);
3205                case 0x09:      /* HT */
3206                case 0x20:      /* SPACE */
3207                case 0xa0:      /* NBSP */
3208                break;
3209                }
3210              }
3211            break;
3212    
3213            case OP_NOT_VSPACE:
3214            for (i = 1; i <= min; i++)
3215              {
3216              if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
3217              switch(*eptr++)
3218                {
3219                default: break;
3220                case 0x0a:      /* LF */
3221                case 0x0b:      /* VT */
3222                case 0x0c:      /* FF */
3223                case 0x0d:      /* CR */
3224                case 0x85:      /* NEL */
3225                RRETURN(MATCH_NOMATCH);
3226                }
3227              }
3228            break;
3229    
3230            case OP_VSPACE:
3231            for (i = 1; i <= min; i++)
3232              {
3233              if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
3234              switch(*eptr++)
3235                {
3236                default: RRETURN(MATCH_NOMATCH);
3237                case 0x0a:      /* LF */
3238                case 0x0b:      /* VT */
3239                case 0x0c:      /* FF */
3240                case 0x0d:      /* CR */
3241                case 0x85:      /* NEL */
3242                break;
3243                }
3244              }
3245            break;
3246    
3247          case OP_NOT_DIGIT:          case OP_NOT_DIGIT:
3248          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
3249            if ((md->ctypes[*eptr++] & ctype_digit) != 0) RRETURN(MATCH_NOMATCH);            if ((md->ctypes[*eptr++] & ctype_digit) != 0) RRETURN(MATCH_NOMATCH);
# Line 2768  for (;;) Line 3299  for (;;)
3299            case PT_ANY:            case PT_ANY:
3300            for (fi = min;; fi++)            for (fi = min;; fi++)
3301              {              {
3302              RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM36);
3303              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3304              if (fi >= max || eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);              if (fi >= max || eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
3305              GETCHARINC(c, eptr);              GETCHARINC(c, eptr);
3306              if (prop_fail_result) RRETURN(MATCH_NOMATCH);              if (prop_fail_result) RRETURN(MATCH_NOMATCH);
3307              }              }
3308            break;            /* Control never gets here */
3309    
3310            case PT_LAMP:            case PT_LAMP:
3311            for (fi = min;; fi++)            for (fi = min;; fi++)
3312              {              {
3313              RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM37);
3314              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3315              if (fi >= max || eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);              if (fi >= max || eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
3316              GETCHARINC(c, eptr);              GETCHARINC(c, eptr);
# Line 2789  for (;;) Line 3320  for (;;)
3320                   prop_chartype == ucp_Lt) == prop_fail_result)                   prop_chartype == ucp_Lt) == prop_fail_result)
3321                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
3322              }              }
3323            break;            /* Control never gets here */
3324    
3325            case PT_GC:            case PT_GC:
3326            for (fi = min;; fi++)            for (fi = min;; fi++)
3327              {              {
3328              RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM38);
3329              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3330              if (fi >= max || eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);              if (fi >= max || eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
3331              GETCHARINC(c, eptr);              GETCHARINC(c, eptr);
# Line 2802  for (;;) Line 3333  for (;;)
3333              if ((prop_category == prop_value) == prop_fail_result)              if ((prop_category == prop_value) == prop_fail_result)
3334                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
3335              }              }
3336            break;            /* Control never gets here */
3337    
3338            case PT_PC:            case PT_PC:
3339            for (fi = min;; fi++)            for (fi = min;; fi++)
3340              {              {
3341              RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM39);
3342              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3343              if (fi >= max || eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);              if (fi >= max || eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
3344              GETCHARINC(c, eptr);              GETCHARINC(c, eptr);
# Line 2815  for (;;) Line 3346  for (;;)
3346              if ((prop_chartype == prop_value) == prop_fail_result)              if ((prop_chartype == prop_value) == prop_fail_result)
3347                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
3348              }              }
3349            break;            /* Control never gets here */
3350    
3351            case PT_SC:            case PT_SC:
3352            for (fi = min;; fi++)            for (fi = min;; fi++)
3353              {              {
3354              RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM40);
3355              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3356              if (fi >= max || eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);              if (fi >= max || eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
3357              GETCHARINC(c, eptr);              GETCHARINC(c, eptr);
# Line 2828  for (;;) Line 3359  for (;;)
3359              if ((prop_script == prop_value) == prop_fail_result)              if ((prop_script == prop_value) == prop_fail_result)
3360                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
3361              }              }
3362            break;            /* Control never gets here */
3363    
3364            default:            default:
3365            RRETURN(PCRE_ERROR_INTERNAL);            RRETURN(PCRE_ERROR_INTERNAL);
           break;  
3366            }            }
3367          }          }
3368    
# Line 2843  for (;;) Line 3373  for (;;)
3373          {          {
3374          for (fi = min;; fi++)          for (fi = min;; fi++)
3375            {            {
3376            RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM41);
3377            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3378            if (fi >= max || eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);            if (fi >= max || eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
3379            GETCHARINCTEST(c, eptr);            GETCHARINCTEST(c, eptr);
# Line 2872  for (;;) Line 3402  for (;;)
3402          {          {
3403          for (fi = min;; fi++)          for (fi = min;; fi++)
3404            {            {
3405            RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM42);
3406            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3407            if (fi >= max || eptr >= md->end_subject ||            if (fi >= max || eptr >= md->end_subject ||
3408                 (ctype == OP_ANY && (ims & PCRE_DOTALL) == 0 &&                 (ctype == OP_ANY && (ims & PCRE_DOTALL) == 0 &&
3409                  eptr <= md->end_subject - md->nllen && IS_NEWLINE(eptr)))                  IS_NEWLINE(eptr)))
3410              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
3411    
3412            GETCHARINC(c, eptr);            GETCHARINC(c, eptr);
# Line 2888  for (;;) Line 3418  for (;;)
3418              case OP_ANYBYTE:              case OP_ANYBYTE:
3419              break;              break;
3420    
3421                case OP_ANYNL:
3422                switch(c)
3423                  {
3424                  default: RRETURN(MATCH_NOMATCH);
3425                  case 0x000d:
3426                  if (eptr < md->end_subject && *eptr == 0x0a) eptr++;
3427                  break;
3428                  case 0x000a:
3429                  case 0x000b:
3430                  case 0x000c:
3431                  case 0x0085:
3432                  case 0x2028:
3433                  case 0x2029:
3434                  break;
3435                  }
3436                break;
3437    
3438                case OP_NOT_HSPACE:
3439                switch(c)
3440                  {
3441                  default: break;
3442                  case 0x09:      /* HT */
3443                  case 0x20:      /* SPACE */
3444                  case 0xa0:      /* NBSP */
3445                  case 0x1680:    /* OGHAM SPACE MARK */
3446                  case 0x180e:    /* MONGOLIAN VOWEL SEPARATOR */
3447                  case 0x2000:    /* EN QUAD */
3448                  case 0x2001:    /* EM QUAD */
3449                  case 0x2002:    /* EN SPACE */
3450                  case 0x2003:    /* EM SPACE */
3451                  case 0x2004:    /* THREE-PER-EM SPACE */
3452                  case 0x2005:    /* FOUR-PER-EM SPACE */
3453                  case 0x2006:    /* SIX-PER-EM SPACE */
3454                  case 0x2007:    /* FIGURE SPACE */
3455                  case 0x2008:    /* PUNCTUATION SPACE */
3456                  case 0x2009:    /* THIN SPACE */
3457                  case 0x200A:    /* HAIR SPACE */
3458                  case 0x202f:    /* NARROW NO-BREAK SPACE */
3459                  case 0x205f:    /* MEDIUM MATHEMATICAL SPACE */
3460                  case 0x3000:    /* IDEOGRAPHIC SPACE */
3461                  RRETURN(MATCH_NOMATCH);
3462                  }
3463                break;
3464    
3465                case OP_HSPACE:
3466                switch(c)
3467                  {
3468                  default: RRETURN(MATCH_NOMATCH);
3469                  case 0x09:      /* HT */
3470                  case 0x20:      /* SPACE */
3471                  case 0xa0:      /* NBSP */
3472                  case 0x1680:    /* OGHAM SPACE MARK */
3473                  case 0x180e:    /* MONGOLIAN VOWEL SEPARATOR */
3474                  case 0x2000:    /* EN QUAD */
3475                  case 0x2001:    /* EM QUAD */
3476                  case 0x2002:    /* EN SPACE */
3477                  case 0x2003:    /* EM SPACE */
3478                  case 0x2004:    /* THREE-PER-EM SPACE */
3479                  case 0x2005:    /* FOUR-PER-EM SPACE */
3480                  case 0x2006:    /* SIX-PER-EM SPACE */
3481                  case 0x2007:    /* FIGURE SPACE */
3482                  case 0x2008:    /* PUNCTUATION SPACE */
3483                  case 0x2009:    /* THIN SPACE */
3484                  case 0x200A:    /* HAIR SPACE */
3485                  case 0x202f:    /* NARROW NO-BREAK SPACE */
3486                  case 0x205f:    /* MEDIUM MATHEMATICAL SPACE */
3487                  case 0x3000:    /* IDEOGRAPHIC SPACE */
3488                  break;
3489                  }
3490                break;
3491    
3492                case OP_NOT_VSPACE:
3493                switch(c)
3494                  {
3495                  default: break;
3496                  case 0x0a:      /* LF */
3497                  case 0x0b:      /* VT */
3498                  case 0x0c:      /* FF */
3499                  case 0x0d:      /* CR */
3500                  case 0x85:      /* NEL */
3501                  case 0x2028:    /* LINE SEPARATOR */
3502                  case 0x2029:    /* PARAGRAPH SEPARATOR */
3503                  RRETURN(MATCH_NOMATCH);
3504                  }
3505                break;
3506    
3507                case OP_VSPACE:
3508                switch(c)
3509                  {
3510                  default: RRETURN(MATCH_NOMATCH);
3511                  case 0x0a:      /* LF */
3512                  case 0x0b:      /* VT */
3513                  case 0x0c:      /* FF */
3514                  case 0x0d:      /* CR */
3515                  case 0x85:      /* NEL */
3516                  case 0x2028:    /* LINE SEPARATOR */
3517                  case 0x2029:    /* PARAGRAPH SEPARATOR */
3518                  break;
3519                  }
3520                break;
3521    
3522              case OP_NOT_DIGIT:              case OP_NOT_DIGIT:
3523              if (c < 256 && (md->ctypes[c] & ctype_digit) != 0)              if (c < 256 && (md->ctypes[c] & ctype_digit) != 0)
3524                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
# Line 2929  for (;;) Line 3560  for (;;)
3560          {          {
3561          for (fi = min;; fi++)          for (fi = min;; fi++)
3562            {            {
3563            RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM43);
3564            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3565            if (fi >= max || eptr >= md->end_subject ||            if (fi >= max || eptr >= md->end_subject ||
3566                 ((ims & PCRE_DOTALL) == 0 &&                 ((ims & PCRE_DOTALL) == 0 && IS_NEWLINE(eptr)))
                  eptr <= md->end_subject - md->nllen && IS_NEWLINE(eptr)))  
3567              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
3568    
3569            c = *eptr++;            c = *eptr++;
# Line 2945  for (;;) Line 3575  for (;;)
3575              case OP_ANYBYTE:              case OP_ANYBYTE:
3576              break;              break;
3577    
3578                case OP_ANYNL:
3579                switch(c)
3580                  {
3581                  default: RRETURN(MATCH_NOMATCH);
3582                  case 0x000d:
3583                  if (eptr < md->end_subject && *eptr == 0x0a) eptr++;
3584                  break;
3585                  case 0x000a:
3586                  case 0x000b:
3587                  case 0x000c:
3588                  case 0x0085:
3589                  break;
3590                  }
3591                break;
3592    
3593                case OP_NOT_HSPACE:
3594                switch(c)
3595                  {
3596                  default: break;
3597                  case 0x09:      /* HT */
3598                  case 0x20:      /* SPACE */
3599                  case 0xa0:      /* NBSP */
3600                  RRETURN(MATCH_NOMATCH);
3601                  }
3602                break;
3603    
3604                case OP_HSPACE:
3605                switch(c)
3606                  {
3607                  default: RRETURN(MATCH_NOMATCH);
3608                  case 0x09:      /* HT */
3609                  case 0x20:      /* SPACE */
3610                  case 0xa0:      /* NBSP */
3611                  break;
3612                  }
3613                break;
3614    
3615                case OP_NOT_VSPACE:
3616                switch(c)
3617                  {
3618                  default: break;
3619                  case 0x0a:      /* LF */
3620                  case 0x0b:      /* VT */
3621                  case 0x0c:      /* FF */
3622                  case 0x0d:      /* CR */
3623                  case 0x85:      /* NEL */
3624                  RRETURN(MATCH_NOMATCH);
3625                  }
3626                break;
3627    
3628                case OP_VSPACE:
3629                switch(c)
3630                  {
3631                  default: RRETURN(MATCH_NOMATCH);
3632                  case 0x0a:      /* LF */
3633                  case 0x0b:      /* VT */
3634                  case 0x0c:      /* FF */
3635                  case 0x0d:      /* CR */
3636                  case 0x85:      /* NEL */
3637                  break;
3638                  }
3639                break;
3640    
3641              case OP_NOT_DIGIT:              case OP_NOT_DIGIT:
3642              if ((md->ctypes[c] & ctype_digit) != 0) RRETURN(MATCH_NOMATCH);              if ((md->ctypes[c] & ctype_digit) != 0) RRETURN(MATCH_NOMATCH);
3643              break;              break;
# Line 2977  for (;;) Line 3670  for (;;)
3670        /* Control never gets here */        /* Control never gets here */
3671        }        }
3672    
3673      /* If maximizing it is worth using inline code for speed, doing the type      /* If maximizing, it is worth using inline code for speed, doing the type
3674      test once at the start (i.e. keep it out of the loop). Again, keep the      test once at the start (i.e. keep it out of the loop). Again, keep the
3675      UTF-8 and UCP stuff separate. */      UTF-8 and UCP stuff separate. */
3676    
# Line 3058  for (;;) Line 3751  for (;;)
3751    
3752          /* eptr is now past the end of the maximum run */          /* eptr is now past the end of the maximum run */
3753    
3754            if (possessive) continue;
3755          for(;;)          for(;;)
3756            {            {
3757            RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM44);
3758            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3759            if (eptr-- == pp) break;        /* Stop if tried at original pos */            if (eptr-- == pp) break;        /* Stop if tried at original pos */
3760            BACKCHAR(eptr);            if (utf8) BACKCHAR(eptr);
3761            }            }
3762          }          }
3763    
# Line 3093  for (;;) Line 3787  for (;;)
3787    
3788          /* eptr is now past the end of the maximum run */          /* eptr is now past the end of the maximum run */
3789    
3790            if (possessive) continue;
3791          for(;;)          for(;;)
3792            {            {
3793            RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM45);
3794            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3795            if (eptr-- == pp) break;        /* Stop if tried at original pos */            if (eptr-- == pp) break;        /* Stop if tried at original pos */
3796            for (;;)                        /* Move back over one extended */            for (;;)                        /* Move back over one extended */
3797              {              {
3798              int len = 1;              int len = 1;
             BACKCHAR(eptr);  
3799              if (!utf8) c = *eptr; else              if (!utf8) c = *eptr; else
3800                {                {
3801                  BACKCHAR(eptr);
3802                GETCHARLEN(c, eptr, len);                GETCHARLEN(c, eptr, len);
3803                }                }
3804              prop_category = _pcre_ucp_findprop(c, &prop_chartype, &prop_script);              prop_category = _pcre_ucp_findprop(c, &prop_chartype, &prop_script);
# Line 3124  for (;;) Line 3819  for (;;)
3819          switch(ctype)          switch(ctype)
3820            {            {
3821            case OP_ANY:            case OP_ANY:
   
           /* Special code is required for UTF8, but when the maximum is  
           unlimited we don't need it, so we repeat the non-UTF8 code. This is  
           probably worth it, because .* is quite a common idiom. */  
   
3822            if (max < INT_MAX)            if (max < INT_MAX)
3823              {              {
3824              if ((ims & PCRE_DOTALL) == 0)              if ((ims & PCRE_DOTALL) == 0)
3825                {                {
3826                for (i = min; i < max; i++)                for (i = min; i < max; i++)
3827                  {                  {
3828                  if (eptr >= md->end_subject ||                  if (eptr >= md->end_subject || IS_NEWLINE(eptr)) break;
                     (eptr <= md->end_subject - md->nllen && IS_NEWLINE(eptr)))  
                   break;  
3829                  eptr++;                  eptr++;
3830                  while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;                  while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;
3831                  }                  }
# Line 3161  for (;;) Line 3849  for (;;)
3849                {                {
3850                for (i = min; i < max; i++)                for (i = min; i < max; i++)
3851                  {                  {
3852                  if (eptr >= md->end_subject ||                  if (eptr >= md->end_subject || IS_NEWLINE(eptr)) break;
                     (eptr <= md->end_subject - md->nllen && IS_NEWLINE(eptr)))  
                   break;  
3853                  eptr++;                  eptr++;
3854                    while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;
3855                  }                  }
               break;  
3856                }                }
3857              else              else
3858                {                {
3859                c = max - min;                eptr = md->end_subject;
               if (c > md->end_subject - eptr) c = md->end_subject - eptr;  
               eptr += c;  
3860                }                }
3861              }              }
3862            break;            break;
# Line 3181  for (;;) Line 3865  for (;;)
3865    
3866            case OP_ANYBYTE:            case OP_ANYBYTE:
3867            c = max - min;            c = max - min;
3868            if (c > md->end_subject - eptr) c = md->end_subject - eptr;            if (c > (unsigned int)(md->end_subject - eptr))
3869                c = md->end_subject - eptr;
3870            eptr += c;            eptr += c;
3871            break;            break;
3872    
3873              case OP_ANYNL:
3874              for (i = min; i < max; i++)
3875                {
3876                int len = 1;
3877                if (eptr >= md->end_subject) break;
3878                GETCHARLEN(c, eptr, len);
3879                if (c == 0x000d)
3880                  {
3881                  if (++eptr >= md->end_subject) break;
3882                  if (*eptr == 0x000a) eptr++;
3883                  }
3884                else
3885                  {
3886                  if (c != 0x000a && c != 0x000b && c != 0x000c &&
3887                      c != 0x0085 && c != 0x2028 && c != 0x2029)
3888                    break;
3889                  eptr += len;
3890                  }
3891                }
3892              break;
3893    
3894              case OP_NOT_HSPACE:
3895              case OP_HSPACE:
3896              for (i = min; i < max; i++)
3897                {
3898                BOOL gotspace;
3899                int len = 1;
3900                if (eptr >= md->end_subject) break;
3901                GETCHARLEN(c, eptr, len);
3902                switch(c)
3903                  {
3904                  default: gotspace = FALSE; break;
3905                  case 0x09:      /* HT */
3906                  case 0x20:      /* SPACE */
3907                  case 0xa0:      /* NBSP */
3908                  case 0x1680:    /* OGHAM SPACE MARK */
3909                  case 0x180e:    /* MONGOLIAN VOWEL SEPARATOR */
3910                  case 0x2000:    /* EN QUAD */
3911                  case 0x2001:    /* EM QUAD */
3912                  case 0x2002:    /* EN SPACE */
3913                  case 0x2003:    /* EM SPACE */
3914                  case 0x2004:    /* THREE-PER-EM SPACE */
3915                  case 0x2005:    /* FOUR-PER-EM SPACE */
3916                  case 0x2006:    /* SIX-PER-EM SPACE */
3917                  case 0x2007:    /* FIGURE SPACE */
3918                  case 0x2008:    /* PUNCTUATION SPACE */
3919                  case 0x2009:    /* THIN SPACE */
3920                  case 0x200A:    /* HAIR SPACE */
3921                  case 0x202f:    /* NARROW NO-BREAK SPACE */
3922                  case 0x205f:    /* MEDIUM MATHEMATICAL SPACE */
3923                  case 0x3000:    /* IDEOGRAPHIC SPACE */
3924                  gotspace = TRUE;
3925                  break;
3926                  }
3927                if (gotspace == (ctype == OP_NOT_HSPACE)) break;
3928                eptr += len;
3929                }
3930              break;
3931    
3932              case OP_NOT_VSPACE:
3933              case OP_VSPACE:
3934              for (i = min; i < max; i++)
3935                {
3936                BOOL gotspace;
3937                int len = 1;
3938                if (eptr >= md->end_subject) break;
3939                GETCHARLEN(c, eptr, len);
3940                switch(c)
3941                  {
3942                  default: gotspace = FALSE; break;
3943                  case 0x0a:      /* LF */
3944                  case 0x0b:      /* VT */
3945                  case 0x0c:      /* FF */
3946                  case 0x0d:      /* CR */
3947                  case 0x85:      /* NEL */
3948                  case 0x2028:    /* LINE SEPARATOR */
3949                  case 0x2029:    /* PARAGRAPH SEPARATOR */
3950                  gotspace = TRUE;
3951                  break;
3952                  }
3953                if (gotspace == (ctype == OP_NOT_VSPACE)) break;
3954                eptr += len;
3955                }
3956              break;
3957    
3958            case OP_NOT_DIGIT:            case OP_NOT_DIGIT:
3959            for (i = min; i < max; i++)            for (i = min; i < max; i++)
3960              {              {
# Line 3257  for (;;) Line 4027  for (;;)
4027    
4028          /* eptr is now past the end of the maximum run */          /* eptr is now past the end of the maximum run */
4029    
4030            if (possessive) continue;
4031          for(;;)          for(;;)
4032            {            {
4033            RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM46);
4034            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
4035            if (eptr-- == pp) break;        /* Stop if tried at original pos */            if (eptr-- == pp) break;        /* Stop if tried at original pos */
4036            BACKCHAR(eptr);            BACKCHAR(eptr);
4037            }            }
4038          }          }
4039        else        else
4040  #endif  #endif  /* SUPPORT_UTF8 */
4041    
4042        /* Not UTF-8 mode */        /* Not UTF-8 mode */
4043          {          {
# Line 3277  for (;;) Line 4048  for (;;)
4048              {              {
4049              for (i = min; i < max; i++)              for (i = min; i < max; i++)
4050                {                {
4051                if (eptr >= md->end_subject ||                if (eptr >= md->end_subject || IS_NEWLINE(eptr)) break;
                   (eptr <= md->end_subject - md->nllen && IS_NEWLINE(eptr)))  
                 break;  
4052                eptr++;                eptr++;
4053                }                }
4054              break;              break;
# Line 3288  for (;;) Line 4057  for (;;)
4057    
4058            case OP_ANYBYTE:            case OP_ANYBYTE:
4059            c = max - min;            c = max - min;
4060            if (c > md->end_subject - eptr) c = md->end_subject - eptr;            if (c > (unsigned int)(md->end_subject - eptr))
4061                c = md->end_subject - eptr;
4062            eptr += c;            eptr += c;
4063            break;            break;
4064    
4065              case OP_ANYNL:
4066              for (i = min; i < max; i++)
4067                {
4068                if (eptr >= md->end_subject) break;
4069                c = *eptr;
4070                if (c == 0x000d)
4071                  {
4072                  if (++eptr >= md->end_subject) break;
4073                  if (*eptr == 0x000a) eptr++;
4074                  }
4075                else
4076                  {
4077                  if (c != 0x000a && c != 0x000b && c != 0x000c && c != 0x0085)
4078                    break;
4079                  eptr++;
4080                  }
4081                }
4082              break;
4083    
4084              case OP_NOT_HSPACE:
4085              for (i = min; i < max; i++)
4086                {
4087                if (eptr >= md->end_subject) break;
4088                c = *eptr;
4089                if (c == 0x09 || c == 0x20 || c == 0xa0) break;
4090                eptr++;
4091                }
4092              break;
4093    
4094              case OP_HSPACE:
4095              for (i = min; i < max; i++)
4096                {
4097                if (eptr >= md->end_subject) break;
4098                c = *eptr;
4099                if (c != 0x09 && c != 0x20 && c != 0xa0) break;
4100                eptr++;
4101                }
4102              break;
4103    
4104              case OP_NOT_VSPACE:
4105              for (i = min; i < max; i++)
4106                {
4107                if (eptr >= md->end_subject) break;
4108                c = *eptr;
4109                if (c == 0x0a || c == 0x0b || c == 0x0c || c == 0x0d || c == 0x85)
4110                  break;
4111                eptr++;
4112                }
4113              break;
4114    
4115              case OP_VSPACE:
4116              for (i = min; i < max; i++)
4117                {
4118                if (eptr >= md->end_subject) break;
4119                c = *eptr;
4120                if (c != 0x0a && c != 0x0b && c != 0x0c && c != 0x0d && c != 0x85)
4121                  break;
4122                eptr++;
4123                }
4124              break;
4125    
4126            case OP_NOT_DIGIT:            case OP_NOT_DIGIT:
4127            for (i = min; i < max; i++)            for (i = min; i < max; i++)
4128              {              {
# Line 3352  for (;;) Line 4183  for (;;)
4183    
4184          /* eptr is now past the end of the maximum run */          /* eptr is now past the end of the maximum run */
4185    
4186            if (possessive) continue;
4187          while (eptr >= pp)          while (eptr >= pp)
4188            {            {
4189            RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM47);
4190            eptr--;            eptr--;
4191            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
4192            }            }
# Line 3366  for (;;) Line 4198  for (;;)
4198        }        }
4199      /* Control never gets here */      /* Control never gets here */
4200    
4201      /* There's been some horrible disaster. Since all codes > OP_BRA are      /* There's been some horrible disaster. Arrival here can only mean there is
4202      for capturing brackets, and there shouldn't be any gaps between 0 and      something seriously wrong in the code above or the OP_xxx definitions. */
     OP_BRA, arrival here can only mean there is something seriously wrong  
     in the code above or the OP_xxx definitions. */  
4203    
4204      default:      default:
4205      DPRINTF(("Unknown opcode %d\n", *ecode));      DPRINTF(("Unknown opcode %d\n", *ecode));
4206      RRETURN(PCRE_ERROR_UNKNOWN_NODE);      RRETURN(PCRE_ERROR_UNKNOWN_OPCODE);
4207      }      }
4208    
4209    /* Do not stick any code in here without much thought; it is assumed    /* Do not stick any code in here without much thought; it is assumed
# Line 3382  for (;;) Line 4212  for (;;)
4212    
4213    }             /* End of main loop */    }             /* End of main loop */
4214  /* Control never reaches here */  /* Control never reaches here */
4215    
4216    
4217    /* When compiling to use the heap rather than the stack for recursive calls to
4218    match(), the RRETURN() macro jumps here. The number that is saved in
4219    frame->Xwhere indicates which label we actually want to return to. */
4220    
4221    #ifdef NO_RECURSE
4222    #define LBL(val) case val: goto L_RM##val;
4223    HEAP_RETURN:
4224    switch (frame->Xwhere)
4225      {
4226      LBL( 1) LBL( 2) LBL( 3) LBL( 4) LBL( 5) LBL( 6) LBL( 7) LBL( 8)
4227      LBL( 9) LBL(10) LBL(11) LBL(12) LBL(13) LBL(14) LBL(15) LBL(16)
4228      LBL(17) LBL(18) LBL(19) LBL(20) LBL(21) LBL(22) LBL(23) LBL(24)
4229      LBL(25) LBL(26) LBL(27) LBL(28) LBL(29) LBL(30) LBL(31) LBL(32)
4230      LBL(33) LBL(34) LBL(35) LBL(36) LBL(37) LBL(38) LBL(39) LBL(40)
4231      LBL(41) LBL(42) LBL(43) LBL(44) LBL(45) LBL(46) LBL(47) LBL(48)
4232      LBL(49) LBL(50) LBL(51) LBL(52) LBL(53) LBL(54)
4233      default:
4234      DPRINTF(("jump error in pcre match: label %d non-existent\n", frame->Xwhere));
4235      return PCRE_ERROR_INTERNAL;
4236      }
4237    #undef LBL
4238    #endif  /* NO_RECURSE */
4239  }  }
4240    
4241    
# Line 3394  Undefine all the macros that were define Line 4248  Undefine all the macros that were define
4248  #ifdef NO_RECURSE  #ifdef NO_RECURSE
4249  #undef eptr  #undef eptr
4250  #undef ecode  #undef ecode
4251    #undef mstart
4252  #undef offset_top  #undef offset_top
4253  #undef ims  #undef ims
4254  #undef eptrb  #undef eptrb
# Line 3411  Undefine all the macros that were define Line 4266  Undefine all the macros that were define
4266    
4267  #undef cur_is_word  #undef cur_is_word
4268  #undef condition  #undef condition
 #undef minimize  
4269  #undef prev_is_word  #undef prev_is_word
4270    
4271  #undef original_ims  #undef original_ims
# Line 3467  Returns:          > 0 => success; value Line 4321  Returns:          > 0 => success; value
4321                   < -1 => some kind of unexpected problem                   < -1 => some kind of unexpected problem
4322  */  */
4323    
4324  PCRE_DATA_SCOPE int  PCRE_EXP_DEFN int
4325  pcre_exec(const pcre *argument_re, const pcre_extra *extra_data,  pcre_exec(const pcre *argument_re, const pcre_extra *extra_data,
4326    PCRE_SPTR subject, int length, int start_offset, int options, int *offsets,    PCRE_SPTR subject, int length, int start_offset, int options, int *offsets,
4327    int offsetcount)    int offsetcount)
# Line 3484  BOOL startline; Line 4338  BOOL startline;
4338  BOOL firstline;  BOOL firstline;
4339  BOOL first_byte_caseless = FALSE;  BOOL first_byte_caseless = FALSE;
4340  BOOL req_byte_caseless = FALSE;  BOOL req_byte_caseless = FALSE;
4341    BOOL utf8;
4342  match_data match_block;  match_data match_block;
4343  match_data *md = &match_block;  match_data *md = &match_block;
4344  const uschar *tables;  const uschar *tables;
# Line 3567  md->end_subject = md->start_subject + le Line 4422  md->end_subject = md->start_subject + le
4422  end_subject = md->end_subject;  end_subject = md->end_subject;
4423    
4424  md->endonly = (re->options & PCRE_DOLLAR_ENDONLY) != 0;  md->endonly = (re->options & PCRE_DOLLAR_ENDONLY) != 0;
4425  md->utf8 = (re->options & PCRE_UTF8) != 0;  utf8 = md->utf8 = (re->options & PCRE_UTF8) != 0;
4426    
4427  md->notbol = (options & PCRE_NOTBOL) != 0;  md->notbol = (options & PCRE_NOTBOL) != 0;
4428  md->noteol = (options & PCRE_NOTEOL) != 0;  md->noteol = (options & PCRE_NOTEOL) != 0;
# Line 3580  md->recursive = NULL; Line 4435  md->recursive = NULL;
4435  md->lcc = tables + lcc_offset;  md->lcc = tables + lcc_offset;
4436  md->ctypes = tables + ctypes_offset;  md->ctypes = tables + ctypes_offset;
4437    
4438  /* Handle different types of newline. The two bits give four cases. If nothing  /* Handle different types of newline. The three bits give eight cases. If
4439  is set at run time, whatever was used at compile time applies. */  nothing is set at run time, whatever was used at compile time applies. */
4440    
4441  switch ((((options & PCRE_NEWLINE_CRLF) == 0)? re->options : options) &  switch ((((options & PCRE_NEWLINE_BITS) == 0)? re->options : (pcre_uint32)options) &
4442           PCRE_NEWLINE_CRLF)         PCRE_NEWLINE_BITS)
4443    {    {
4444    default:              newline = NEWLINE; break;   /* Compile-time default */    case 0: newline = NEWLINE; break;   /* Compile-time default */
4445    case PCRE_NEWLINE_CR: newline = '\r'; break;    case PCRE_NEWLINE_CR: newline = '\r'; break;
4446    case PCRE_NEWLINE_LF: newline = '\n'; break;    case PCRE_NEWLINE_LF: newline = '\n'; break;
4447    case PCRE_NEWLINE_CR+    case PCRE_NEWLINE_CR+
4448         PCRE_NEWLINE_LF: newline = ('\r' << 8) | '\n'; break;         PCRE_NEWLINE_LF: newline = ('\r' << 8) | '\n'; break;
4449      case PCRE_NEWLINE_ANY: newline = -1; break;
4450      case PCRE_NEWLINE_ANYCRLF: newline = -2; break;
4451      default: return PCRE_ERROR_BADNEWLINE;
4452    }    }
4453    
4454  if (newline > 255)  if (newline == -2)
4455    {    {
4456    md->nllen = 2;    md->nltype = NLTYPE_ANYCRLF;
4457    md->nl[0] = (newline >> 8) & 255;    }
4458    md->nl[1] = newline & 255;  else if (newline < 0)
4459      {
4460      md->nltype = NLTYPE_ANY;
4461    }    }
4462  else  else
4463    {    {
4464    md->nllen = 1;    md->nltype = NLTYPE_FIXED;
4465    md->nl[0] = newline;    if (newline > 255)
4466        {
4467        md->nllen = 2;
4468        md->nl[0] = (newline >> 8) & 255;
4469        md->nl[1] = newline & 255;
4470        }
4471      else
4472        {
4473        md->nllen = 1;
4474        md->nl[0] = newline;
4475        }
4476    }    }
4477    
4478  /* Partial matching is supported only for a restricted set of regexes at the  /* Partial matching is supported only for a restricted set of regexes at the
# Line 3615  if (md->partial && (re->options & PCRE_N Line 4485  if (md->partial && (re->options & PCRE_N
4485  back the character offset. */  back the character offset. */
4486    
4487  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF8
4488  if (md->utf8 && (options & PCRE_NO_UTF8_CHECK) == 0)  if (utf8 && (options & PCRE_NO_UTF8_CHECK) == 0)
4489    {    {
4490    if (_pcre_valid_utf8((uschar *)subject, length) >= 0)    if (_pcre_valid_utf8((uschar *)subject, length) >= 0)
4491      return PCRE_ERROR_BADUTF8;      return PCRE_ERROR_BADUTF8;
# Line 3707  if ((re->options & PCRE_REQCHSET) != 0) Line 4577  if ((re->options & PCRE_REQCHSET) != 0)
4577    req_byte2 = (tables + fcc_offset)[req_byte];  /* case flipped */    req_byte2 = (tables + fcc_offset)[req_byte];  /* case flipped */
4578    }    }
4579    
4580    
4581    /* ==========================================================================*/
4582    
4583  /* Loop for handling unanchored repeated matching attempts; for anchored regexs  /* Loop for handling unanchored repeated matching attempts; for anchored regexs
4584  the loop runs just once. */  the loop runs just once. */
4585    
4586  do  for(;;)
4587    {    {
4588    USPTR save_end_subject = end_subject;    USPTR save_end_subject = end_subject;
4589      USPTR new_start_match;
4590    
4591    /* Reset the maximum number of extractions we might see. */    /* Reset the maximum number of extractions we might see. */
4592    
# Line 3725  do Line 4599  do
4599    
4600    /* Advance to a unique first char if possible. If firstline is TRUE, the    /* Advance to a unique first char if possible. If firstline is TRUE, the
4601    start of the match is constrained to the first line of a multiline string.    start of the match is constrained to the first line of a multiline string.
4602    Implement this by temporarily adjusting end_subject so that we stop scanning    That is, the match must be before or at the first newline. Implement this by
4603    at a newline. If the match fails at the newline, later code breaks this loop.    temporarily adjusting end_subject so that we stop scanning at a newline. If
4604    */    the match fails at the newline, later code breaks this loop. */
4605    
4606    if (firstline)    if (firstline)
4607      {      {
4608      USPTR t = start_match;      USPTR t = start_match;
4609      while (t <= save_end_subject - md->nllen && !IS_NEWLINE(t)) t++;      while (t < md->end_subject && !IS_NEWLINE(t)) t++;
4610      end_subject = t;      end_subject = t;
4611      }      }
4612    
# Line 3753  do Line 4627  do
4627    
4628    else if (startline)    else if (startline)
4629      {      {
4630      if (start_match >= md->start_subject + md->nllen +      if (start_match > md->start_subject + start_offset)
           start_offset)  
4631        {        {
4632        while (start_match <= end_subject &&        while (start_match <= end_subject && !WAS_NEWLINE(start_match))
4633               !IS_NEWLINE(start_match - md->nllen))          start_match++;
4634    
4635          /* If we have just passed a CR and the newline option is ANY or ANYCRLF,
4636          and we are now at a LF, advance the match position by one more character.
4637          */
4638    
4639          if (start_match[-1] == '\r' &&
4640               (md->nltype == NLTYPE_ANY || md->nltype == NLTYPE_ANYCRLF) &&
4641               start_match < end_subject &&
4642               *start_match == '\n')
4643          start_match++;          start_match++;
4644        }        }
4645      }      }
# Line 3793  do Line 4675  do
4675    
4676    HOWEVER: when the subject string is very, very long, searching to its end can    HOWEVER: when the subject string is very, very long, searching to its end can
4677    take a long time, and give bad performance on quite ordinary patterns. This    take a long time, and give bad performance on quite ordinary patterns. This
4678    showed up when somebody was matching /^C/ on a 32-megabyte string... so we    showed up when somebody was matching something like /^\d+C/ on a 32-megabyte
4679    don't do this when the string is sufficiently long.    string... so we don't do this when the string is sufficiently long.
4680    
4681    ALSO: this processing is disabled when partial matching is requested.    ALSO: this processing is disabled when partial matching is requested.
4682    */    */
# Line 3826  do Line 4708  do
4708            }            }
4709          }          }
4710    
4711        /* If we can't find the required character, break the matching loop */        /* If we can't find the required character, break the matching loop,
4712          forcing a match failure. */
4713    
4714        if (p >= end_subject) break;        if (p >= end_subject)
4715            {
4716            rc = MATCH_NOMATCH;
4717            break;
4718            }
4719    
4720        /* If we have found the required character, save the point where we        /* If we have found the required character, save the point where we
4721        found it, so that we don't search again next time round the loop if        found it, so that we don't search again next time round the loop if
# Line 3838  do Line 4725  do
4725        }        }
4726      }      }
4727    
4728    /* When a match occurs, substrings will be set for all internal extractions;    /* OK, we can now run the match. */
   we just need to set up the whole thing as substring 0 before returning. If  
   there were too many extractions, set the return code to zero. In the case  
   where we had to get some local store to hold offsets for backreferences, copy  
   those back references that we can. In this case there need not be overflow  
   if certain parts of the pattern were not used. */  
4729    
4730    md->start_match = start_match;    md->start_match_ptr = start_match;
4731    md->match_call_count = 0;    md->match_call_count = 0;
4732      rc = match(start_match, md->start_code, start_match, 2, md, ims, NULL, 0, 0);
4733    
4734    rc = match(start_match, md->start_code, 2, md, ims, NULL, match_isgroup, 0);    switch(rc)
   
   /* When the result is no match, if the subject's first character was a  
   newline and the PCRE_FIRSTLINE option is set, break (which will return  
   PCRE_ERROR_NOMATCH). The option requests that a match occur before the first  
   newline in the subject. Otherwise, advance the pointer to the next character  
   and continue - but the continuation will actually happen only when the  
   pattern is not anchored. */  
   
   if (rc == MATCH_NOMATCH)  
4735      {      {
4736      if (firstline &&      /* NOMATCH and PRUNE advance by one character. THEN at this level acts
4737          start_match <= md->end_subject - md->nllen &&      exactly like PRUNE. */
4738          IS_NEWLINE(start_match))  
4739        break;      case MATCH_NOMATCH:
4740      start_match++;      case MATCH_PRUNE:
4741        case MATCH_THEN:
4742        new_start_match = start_match + 1;
4743  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF8
4744      if (md->utf8)      if (utf8)
4745        while(start_match < end_subject && (*start_match & 0xc0) == 0x80)        while(new_start_match < end_subject && (*new_start_match & 0xc0) == 0x80)
4746          start_match++;          new_start_match++;
4747  #endif  #endif
4748      continue;      break;
     }  
4749    
4750    if (rc != MATCH_MATCH)      /* SKIP passes back the next starting point explicitly. */
4751      {  
4752      DPRINTF((">>>> error: returning %d\n", rc));      case MATCH_SKIP:
4753      return rc;      new_start_match = md->start_match_ptr;
4754        break;
4755    
4756        /* COMMIT disables the bumpalong, but otherwise behaves as NOMATCH. */
4757    
4758        case MATCH_COMMIT:
4759        rc = MATCH_NOMATCH;
4760        goto ENDLOOP;
4761    
4762        /* Any other return is some kind of error. */
4763    
4764        default:
4765        goto ENDLOOP;
4766      }      }
4767    
4768    /* We have a match! Copy the offset information from temporary store if    /* Control reaches here for the various types of "no match at this point"
4769    necessary */    result. Reset the code to MATCH_NOMATCH for subsequent checking. */
4770    
4771      rc = MATCH_NOMATCH;
4772    
4773      /* If PCRE_FIRSTLINE is set, the match must happen before or at the first
4774      newline in the subject (though it may continue over the newline). Therefore,
4775      if we have just failed to match, starting at a newline, do not continue. */
4776    
4777      if (firstline && IS_NEWLINE(start_match)) break;
4778    
4779      /* Advance to new matching position */
4780    
4781      start_match = new_start_match;
4782    
4783      /* Break the loop if the pattern is anchored or if we have passed the end of
4784      the subject. */
4785    
4786      if (anchored || start_match > end_subject) break;
4787    
4788      /* If we have just passed a CR and we are now at a LF, and the pattern does
4789      not contain any explicit matches for \r or \n, and the newline option is CRLF
4790      or ANY or ANYCRLF, advance the match position by one more character. */
4791    
4792      if (start_match[-1] == '\r' &&
4793          start_match < end_subject &&
4794          *start_match == '\n' &&
4795          (re->options & PCRE_HASCRORLF) == 0 &&
4796            (md->nltype == NLTYPE_ANY ||
4797             md->nltype == NLTYPE_ANYCRLF ||
4798             md->nllen == 2))
4799        start_match++;
4800    
4801      }   /* End of for(;;) "bumpalong" loop */
4802    
4803    /* ==========================================================================*/
4804    
4805    /* We reach here when rc is not MATCH_NOMATCH, or if one of the stopping
4806    conditions is true:
4807    
4808    (1) The pattern is anchored or the match was failed by (*COMMIT);
4809    
4810    (2) We are past the end of the subject;
4811    
4812    (3) PCRE_FIRSTLINE is set and we have failed to match at a newline, because
4813        this option requests that a match occur at or before the first newline in
4814        the subject.
4815    
4816    When we have a match and the offset vector is big enough to deal with any
4817    backreferences, captured substring offsets will already be set up. In the case
4818    where we had to get some local store to hold offsets for backreference
4819    processing, copy those that we can. In this case there need not be overflow if
4820    certain parts of the pattern were not used, even though there are more
4821    capturing parentheses than vector slots. */
4822    
4823    ENDLOOP:
4824    
4825    if (rc == MATCH_MATCH)
4826      {
4827    if (using_temporary_offsets)    if (using_temporary_offsets)
4828      {      {
4829      if (offsetcount >= 4)      if (offsetcount >= 4)
# Line 3889  do Line 4832  do
4832          (offsetcount - 2) * sizeof(int));          (offsetcount - 2) * sizeof(int));
4833        DPRINTF(("Copied offsets from temporary memory\n"));        DPRINTF(("Copied offsets from temporary memory\n"));
4834        }        }
4835      if (md->end_offset_top > offsetcount)      if (md->end_offset_top > offsetcount) md->offset_overflow = TRUE;
       md->offset_overflow = TRUE;  
   
4836      DPRINTF(("Freeing temporary memory\n"));      DPRINTF(("Freeing temporary memory\n"));
4837      (pcre_free)(md->offset_vector);      (pcre_free)(md->offset_vector);
4838      }      }
4839    
4840      /* Set the return code to the number of captured strings, or 0 if there are
4841      too many to fit into the vector. */
4842    
4843    rc = md->offset_overflow? 0 : md->end_offset_top/2;    rc = md->offset_overflow? 0 : md->end_offset_top/2;
4844    
4845      /* If there is space, set up the whole thing as substring 0. The value of
4846      md->start_match_ptr might be modified if \K was encountered on the success
4847      matching path. */
4848    
4849    if (offsetcount < 2) rc = 0; else    if (offsetcount < 2) rc = 0; else
4850      {      {
4851      offsets[0] = start_match - md->start_subject;      offsets[0] = md->start_match_ptr - md->start_subject;
4852      offsets[1] = md->end_match_ptr - md->start_subject;      offsets[1] = md->end_match_ptr - md->start_subject;
4853      }      }
4854    
# Line 3908  do Line 4856  do
4856    return rc;    return rc;
4857    }    }
4858    
4859  /* This "while" is the end of the "do" above */  /* Control gets here if there has been an error, or if the overall match
4860    attempt has failed at all permitted starting positions. */
 while (!anchored && start_match <= end_subject);  
4861    
4862  if (using_temporary_offsets)  if (using_temporary_offsets)
4863    {    {
# Line 3918  if (using_temporary_offsets) Line 4865  if (using_temporary_offsets)
4865    (pcre_free)(md->offset_vector);    (pcre_free)(md->offset_vector);
4866    }    }
4867    
4868  if (md->partial && md->hitend)  if (rc != MATCH_NOMATCH)
4869      {
4870      DPRINTF((">>>> error: returning %d\n", rc));
4871      return rc;
4872      }
4873    else if (md->partial && md->hitend)
4874    {    {
4875    DPRINTF((">>>> returning PCRE_ERROR_PARTIAL\n"));    DPRINTF((">>>> returning PCRE_ERROR_PARTIAL\n"));
4876    return PCRE_ERROR_PARTIAL;    return PCRE_ERROR_PARTIAL;

Legend:
Removed from v.91  
changed lines
  Added in v.227

  ViewVC Help
Powered by ViewVC 1.1.5