/[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 165 by ph10, Wed May 9 10:50:57 2007 UTC revision 345 by ph10, Mon Apr 28 15:10:02 2008 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) 1997-2007 University of Cambridge             Copyright (c) 1997-2008 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    #ifdef HAVE_CONFIG_H
46    #include "config.h"
47    #endif
48    
49  #define NLBLOCK md             /* Block containing newline information */  #define NLBLOCK md             /* Block containing newline information */
50  #define PSSTART start_subject  /* Field containing processed string start */  #define PSSTART start_subject  /* Field containing processed string start */
51  #define PSEND   end_subject    /* Field containing processed string end */  #define PSEND   end_subject    /* Field containing processed string end */
# Line 53  possible. There are also some static sup Line 57  possible. There are also some static sup
57  #undef min  #undef min
58  #undef max  #undef max
59    
 /* The chain of eptrblocks for tail recursions uses memory in stack workspace,  
 obtained at top level, the size of which is defined by EPTR_WORK_SIZE. */  
   
 #define EPTR_WORK_SIZE (1000)  
   
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_cbegroup       0x02  /* Could-be-empty unlimited repeat group */  #define match_cbegroup       0x02  /* Could-be-empty unlimited repeat group */
 #define match_tail_recursed  0x04  /* Tail recursive call */  
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 70  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 205  variable instead of being passed in the Line 211  variable instead of being passed in the
211  ****************************************************************************  ****************************************************************************
212  ***************************************************************************/  ***************************************************************************/
213    
214    /* Numbers for RMATCH calls. When this list is changed, the code at HEAP_RETURN
215  /* Numbers for RMATCH calls */  below must be updated in sync.  */
216    
217  enum { RM1=1, RM2,  RM3,  RM4,  RM5,  RM6,  RM7,  RM8,  RM9,  RM10,  enum { RM1=1, RM2,  RM3,  RM4,  RM5,  RM6,  RM7,  RM8,  RM9,  RM10,
218         RM11,  RM12, RM13, RM14, RM15, RM16, RM17, RM18, RM19, RM20,         RM11,  RM12, RM13, RM14, RM15, RM16, RM17, RM18, RM19, RM20,
219         RM21,  RM22, RM23, RM24, RM25, RM26, RM27, RM28, RM29, RM30,         RM21,  RM22, RM23, RM24, RM25, RM26, RM27, RM28, RM29, RM30,
220         RM31,  RM32, RM33, RM34, RM35, RM36, RM37, RM38, RM39, RM40,         RM31,  RM32, RM33, RM34, RM35, RM36, RM37, RM38, RM39, RM40,
221         RM41,  RM42, RM43, RM44, RM45, RM46, RM47 };         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. Note that the "rw" argument of RMATCH isn't  versions and production versions. Note that the "rw" argument of RMATCH isn't
# Line 226  actuall used in this definition. */ Line 232  actuall used in this definition. */
232  #define RMATCH(ra,rb,rc,rd,re,rf,rg,rw) \  #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    rrc = 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 236  actuall used in this definition. */ Line 242  actuall used in this definition. */
242    }    }
243  #else  #else
244  #define RMATCH(ra,rb,rc,rd,re,rf,rg,rw) \  #define RMATCH(ra,rb,rc,rd,re,rf,rg,rw) \
245    rrc = 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    
# Line 255  argument of match(), which never changes Line 261  argument of match(), which never changes
261    frame->Xwhere = rw; \    frame->Xwhere = rw; \
262    newframe->Xeptr = ra;\    newframe->Xeptr = ra;\
263    newframe->Xecode = rb;\    newframe->Xecode = rb;\
264      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;\
# Line 291  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 371  made performance worse. Line 379  made performance worse.
379  Arguments:  Arguments:
380     eptr        pointer to current character in subject     eptr        pointer to current character in subject
381     ecode       pointer to current position in compiled 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 380  Arguments: Line 390  Arguments:
390                   match_condassert - this is an assertion condition                   match_condassert - this is an assertion condition
391                   match_cbegroup - this is the start of an unlimited repeat                   match_cbegroup - this is the start of an unlimited repeat
392                     group that can match an empty string                     group that can match an empty string
                  match_tail_recursed - this is a tail_recursed group  
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 390  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  {  {
# Line 418  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 432  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 580  original_ims = ims;    /* Save for reset Line 591  original_ims = ims;    /* Save for reset
591  string, the match_cbegroup flag is set. When this is the case, add the current  string, the match_cbegroup flag is set. When this is the case, add the current
592  subject pointer to the chain of such remembered pointers, to be checked when we  subject pointer to the chain of such remembered pointers, to be checked when we
593  hit the closing ket, in order to break infinite loops that match no characters.  hit the closing ket, in order to break infinite loops that match no characters.
594  When match() is called in other circumstances, don't add to the chain. If this  When match() is called in other circumstances, don't add to the chain. The
595  is a tail recursion, use a block from the workspace, as the one on the stack is  match_cbegroup flag must NOT be used with tail recursion, because the memory
596  already used. */  block that is used is on the stack, so a new one may be required for each
597    match(). */
598    
599  if ((flags & match_cbegroup) != 0)  if ((flags & match_cbegroup) != 0)
600    {    {
601    eptrblock *p;    newptrb.epb_saved_eptr = eptr;
602    if ((flags & match_tail_recursed) != 0)    newptrb.epb_prev = eptrb;
603      {    eptrb = &newptrb;
     if (md->eptrn >= EPTR_WORK_SIZE) RRETURN(PCRE_ERROR_NULLWSLIMIT);  
     p = md->eptrchain + md->eptrn++;  
     }  
   else p = &newptrb;  
   p->epb_saved_eptr = eptr;  
   p->epb_prev = eptrb;  
   eptrb = p;  
604    }    }
605    
606  /* Now start processing the opcodes. */  /* Now start processing the opcodes. */
# Line 610  for (;;) Line 615  for (;;)
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    switch(op)    switch(op)
622      {      {
623        case OP_FAIL:
624        RRETURN(MATCH_NOMATCH);
625    
626        case OP_PRUNE:
627        RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,
628          ims, eptrb, flags, RM51);
629        if (rrc != MATCH_NOMATCH) RRETURN(rrc);
630        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      /* 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.      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      We mustn't change the current values of the data slot, because they may be
# Line 656  for (;;) Line 689  for (;;)
689          {          {
690          RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,          RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,
691            ims, eptrb, flags, RM1);            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 671  for (;;) Line 704  for (;;)
704        RRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
705        }        }
706    
707      /* Insufficient room for saving captured contents. Treat as a non-capturing      /* FALL THROUGH ... Insufficient room for saving captured contents. Treat
708      bracket. */      as a non-capturing bracket. */
709    
710        /* VVVVVVVVVVVVVVVVVVVVVVVVV */
711        /* VVVVVVVVVVVVVVVVVVVVVVVVV */
712    
713      DPRINTF(("insufficient capture room: treat as non-capturing\n"));      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      /* 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      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      recursive call to match() whatever happened. We can reduce stack usage by
721      turning this into a tail recursion. */      turning this into a tail recursion, except in the case when match_cbegroup
722        is set.*/
723    
724      case OP_BRA:      case OP_BRA:
725      case OP_SBRA:      case OP_SBRA:
# Line 687  for (;;) Line 727  for (;;)
727      flags = (op >= OP_SBRA)? match_cbegroup : 0;      flags = (op >= OP_SBRA)? match_cbegroup : 0;
728      for (;;)      for (;;)
729        {        {
730        if (ecode[GET(ecode, 1)] != OP_ALT)        if (ecode[GET(ecode, 1)] != OP_ALT)   /* Final alternative */
731          {          {
732          ecode += _pcre_OP_lengths[*ecode];          if (flags == 0)    /* Not a possibly empty group */
733          flags |= match_tail_recursed;            {
734          DPRINTF(("bracket 0 tail recursion\n"));            ecode += _pcre_OP_lengths[*ecode];
735          goto TAIL_RECURSE;            DPRINTF(("bracket 0 tail recursion\n"));
736              goto TAIL_RECURSE;
737              }
738    
739            /* 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;
# Line 700  for (;;) Line 748  for (;;)
748    
749        RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md, ims,        RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md, ims,
750          eptrb, flags, RM2);          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 748  for (;;) Line 796  for (;;)
796          ecode += 1 + LINK_SIZE + GET(ecode, LINK_SIZE + 2);          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          }          }
# Line 760  for (;;) Line 808  for (;;)
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. If the second      we can use tail recursion to avoid using another stack frame, except when
812      alternative doesn't exist, we can just plough on. */      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)      if (condition || *ecode == OP_ALT)
816        {        {
817        ecode += 1 + LINK_SIZE;        ecode += 1 + LINK_SIZE;
818        flags = match_tail_recursed | ((op == OP_SCOND)? match_cbegroup : 0);        if (op == OP_SCOND)        /* Possibly empty group */
819        goto TAIL_RECURSE;          {
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      else                         /* Condition false & no 2nd alternative */
830        {        {
831        ecode += 1 + LINK_SIZE;        ecode += 1 + LINK_SIZE;
832        }        }
833      break;      break;
834    
835    
836      /* End of the pattern. If we are in a top-level recursion, we should      /* End of the pattern, either real or forced. If we are in a top-level
837      restore the offsets appropriately and continue from after the call. */      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 787  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 796  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 822  for (;;) Line 882  for (;;)
882        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, ims, NULL, 0,        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, ims, NULL, 0,
883          RM4);          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 849  for (;;) Line 909  for (;;)
909        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, ims, NULL, 0,        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, ims, NULL, 0,
910          RM5);          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 873  for (;;) Line 933  for (;;)
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 904  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 966  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. */
# Line 986  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 1020  for (;;) Line 1080  for (;;)
1080    
1081      do      do
1082        {        {
1083        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, ims,        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, ims, eptrb, 0, RM7);
         eptrb, 0, RM7);  
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 1066  for (;;) Line 1125  for (;;)
1125    
1126      if (*ecode == OP_KETRMIN)      if (*ecode == OP_KETRMIN)
1127        {        {
1128        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, ims, eptrb, 0,        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, ims, eptrb, 0, RM8);
         RM8);  
1129        if (rrc != MATCH_NOMATCH) RRETURN(rrc);        if (rrc != MATCH_NOMATCH) RRETURN(rrc);
1130        ecode = prev;        ecode = prev;
1131        flags = match_tail_recursed;        flags = 0;
1132        goto TAIL_RECURSE;        goto TAIL_RECURSE;
1133        }        }
1134      else  /* OP_KETRMAX */      else  /* OP_KETRMAX */
# Line 1078  for (;;) Line 1136  for (;;)
1136        RMATCH(eptr, prev, offset_top, md, ims, eptrb, match_cbegroup, RM9);        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 = match_tail_recursed;        flags = 0;
1140        goto TAIL_RECURSE;        goto TAIL_RECURSE;
1141        }        }
1142      /* Control never gets here */      /* Control never gets here */
# Line 1090  for (;;) Line 1148  for (;;)
1148      do ecode += GET(ecode,1); while (*ecode == OP_ALT);      do ecode += GET(ecode,1); while (*ecode == OP_ALT);
1149      break;      break;
1150    
1151      /* BRAZERO and BRAMINZERO occur just before a bracket group, indicating      /* BRAZERO, BRAMINZERO and SKIPZERO occur just before a bracket group,
1152      that it may occur zero times. It may repeat infinitely, or not at all -      indicating that it may occur zero times. It may repeat infinitely, or not
1153      i.e. it could be ()* or ()? in the pattern. Brackets with fixed upper      at all - i.e. it could be ()* or ()? or even (){0} in the pattern. Brackets
1154      repeat limits are compiled as a number of copies, with the optional ones      with fixed upper repeat limits are compiled as a number of copies, with the
1155      preceded by BRAZERO or BRAMINZERO. */      optional ones preceded by BRAZERO or BRAMINZERO. */
1156    
1157      case OP_BRAZERO:      case OP_BRAZERO:
1158        {        {
# Line 1116  for (;;) Line 1174  for (;;)
1174        }        }
1175      break;      break;
1176    
1177        case OP_SKIPZERO:
1178          {
1179          next = ecode+1;
1180          do next += GET(next,1); while (*next == OP_ALT);
1181          ecode = next + 1 + LINK_SIZE;
1182          }
1183        break;
1184    
1185      /* End of a group, repeated or non-repeating. */      /* End of a group, repeated or non-repeating. */
1186    
1187      case OP_KET:      case OP_KET:
# Line 1180  for (;;) Line 1246  for (;;)
1246          recursion_info *rec = md->recursive;          recursion_info *rec = md->recursive;
1247          DPRINTF(("Recursion (%d) succeeded - continuing\n", number));          DPRINTF(("Recursion (%d) succeeded - continuing\n", number));
1248          md->recursive = rec->prevrec;          md->recursive = rec->prevrec;
1249          md->start_match = rec->save_start;          mstart = rec->save_start;
1250          memcpy(md->offset_vector, rec->offset_save,          memcpy(md->offset_vector, rec->offset_save,
1251            rec->saved_max * sizeof(int));            rec->saved_max * sizeof(int));
1252          ecode = rec->after_call;          ecode = rec->after_call;
# Line 1209  for (;;) Line 1275  for (;;)
1275    
1276      /* 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
1277      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
1278      tail recursion to avoid using another stack frame. */      tail recursion to avoid using another stack frame, unless we have an
1279        unlimited repeat of a group that can match an empty string. */
1280    
1281      flags = (*prev >= OP_SBRA)? match_cbegroup : 0;      flags = (*prev >= OP_SBRA)? match_cbegroup : 0;
1282    
1283      if (*ecode == OP_KETRMIN)      if (*ecode == OP_KETRMIN)
1284        {        {
1285        RMATCH(eptr, ecode + 1+LINK_SIZE, offset_top, md, ims, eptrb, 0,        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, ims, eptrb, 0, RM12);
         RM12);  
1286        if (rrc != MATCH_NOMATCH) RRETURN(rrc);        if (rrc != MATCH_NOMATCH) RRETURN(rrc);
1287          if (flags != 0)    /* Could match an empty string */
1288            {
1289            RMATCH(eptr, prev, offset_top, md, ims, eptrb, flags, RM50);
1290            RRETURN(rrc);
1291            }
1292        ecode = prev;        ecode = prev;
       flags |= match_tail_recursed;  
1293        goto TAIL_RECURSE;        goto TAIL_RECURSE;
1294        }        }
1295      else  /* OP_KETRMAX */      else  /* OP_KETRMAX */
# Line 1227  for (;;) Line 1297  for (;;)
1297        RMATCH(eptr, prev, offset_top, md, ims, eptrb, flags, RM13);        RMATCH(eptr, prev, offset_top, md, ims, eptrb, flags, RM13);
1298        if (rrc != MATCH_NOMATCH) RRETURN(rrc);        if (rrc != MATCH_NOMATCH) RRETURN(rrc);
1299        ecode += 1 + LINK_SIZE;        ecode += 1 + LINK_SIZE;
1300        flags = match_tail_recursed;        flags = 0;
1301        goto TAIL_RECURSE;        goto TAIL_RECURSE;
1302        }        }
1303      /* Control never gets here */      /* Control never gets here */
# Line 1260  for (;;) Line 1330  for (;;)
1330      ecode++;      ecode++;
1331      break;      break;
1332    
1333        /* Reset the start of match point */
1334    
1335        case OP_SET_SOM:
1336        mstart = eptr;
1337        ecode++;
1338        break;
1339    
1340      /* Assert before internal newline if multiline, or before a terminating      /* Assert before internal newline if multiline, or before a terminating
1341      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. */
1342    
# Line 1352  for (;;) Line 1429  for (;;)
1429      /* Match a single character type; inline for speed */      /* Match a single character type; inline for speed */
1430    
1431      case OP_ANY:      case OP_ANY:
1432      if ((ims & PCRE_DOTALL) == 0)      if (IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH);
1433        {      /* Fall through */
1434        if (IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH);  
1435        }      case OP_ALLANY:
1436      if (eptr++ >= md->end_subject) RRETURN(MATCH_NOMATCH);      if (eptr++ >= md->end_subject) RRETURN(MATCH_NOMATCH);
1437      if (utf8)      if (utf8) while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;
       while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;  
1438      ecode++;      ecode++;
1439      break;      break;
1440    
# Line 1457  for (;;) Line 1533  for (;;)
1533        case 0x000d:        case 0x000d:
1534        if (eptr < md->end_subject && *eptr == 0x0a) eptr++;        if (eptr < md->end_subject && *eptr == 0x0a) eptr++;
1535        break;        break;
1536    
1537        case 0x000a:        case 0x000a:
1538          break;
1539    
1540        case 0x000b:        case 0x000b:
1541        case 0x000c:        case 0x000c:
1542        case 0x0085:        case 0x0085:
1543        case 0x2028:        case 0x2028:
1544        case 0x2029:        case 0x2029:
1545          if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);
1546          break;
1547          }
1548        ecode++;
1549        break;
1550    
1551        case OP_NOT_HSPACE:
1552        if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
1553        GETCHARINCTEST(c, eptr);
1554        switch(c)
1555          {
1556          default: break;
1557          case 0x09:      /* HT */
1558          case 0x20:      /* SPACE */
1559          case 0xa0:      /* NBSP */
1560          case 0x1680:    /* OGHAM SPACE MARK */
1561          case 0x180e:    /* MONGOLIAN VOWEL SEPARATOR */
1562          case 0x2000:    /* EN QUAD */
1563          case 0x2001:    /* EM QUAD */
1564          case 0x2002:    /* EN SPACE */
1565          case 0x2003:    /* EM SPACE */
1566          case 0x2004:    /* THREE-PER-EM SPACE */
1567          case 0x2005:    /* FOUR-PER-EM SPACE */
1568          case 0x2006:    /* SIX-PER-EM SPACE */
1569          case 0x2007:    /* FIGURE SPACE */
1570          case 0x2008:    /* PUNCTUATION SPACE */
1571          case 0x2009:    /* THIN SPACE */
1572          case 0x200A:    /* HAIR SPACE */
1573          case 0x202f:    /* NARROW NO-BREAK SPACE */
1574          case 0x205f:    /* MEDIUM MATHEMATICAL SPACE */
1575          case 0x3000:    /* IDEOGRAPHIC SPACE */
1576          RRETURN(MATCH_NOMATCH);
1577          }
1578        ecode++;
1579        break;
1580    
1581        case OP_HSPACE:
1582        if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
1583        GETCHARINCTEST(c, eptr);
1584        switch(c)
1585          {
1586          default: RRETURN(MATCH_NOMATCH);
1587          case 0x09:      /* HT */
1588          case 0x20:      /* SPACE */
1589          case 0xa0:      /* NBSP */
1590          case 0x1680:    /* OGHAM SPACE MARK */
1591          case 0x180e:    /* MONGOLIAN VOWEL SEPARATOR */
1592          case 0x2000:    /* EN QUAD */
1593          case 0x2001:    /* EM QUAD */
1594          case 0x2002:    /* EN SPACE */
1595          case 0x2003:    /* EM SPACE */
1596          case 0x2004:    /* THREE-PER-EM SPACE */
1597          case 0x2005:    /* FOUR-PER-EM SPACE */
1598          case 0x2006:    /* SIX-PER-EM SPACE */
1599          case 0x2007:    /* FIGURE SPACE */
1600          case 0x2008:    /* PUNCTUATION SPACE */
1601          case 0x2009:    /* THIN SPACE */
1602          case 0x200A:    /* HAIR SPACE */
1603          case 0x202f:    /* NARROW NO-BREAK SPACE */
1604          case 0x205f:    /* MEDIUM MATHEMATICAL SPACE */
1605          case 0x3000:    /* IDEOGRAPHIC SPACE */
1606          break;
1607          }
1608        ecode++;
1609        break;
1610    
1611        case OP_NOT_VSPACE:
1612        if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
1613        GETCHARINCTEST(c, eptr);
1614        switch(c)
1615          {
1616          default: break;
1617          case 0x0a:      /* LF */
1618          case 0x0b:      /* VT */
1619          case 0x0c:      /* FF */
1620          case 0x0d:      /* CR */
1621          case 0x85:      /* NEL */
1622          case 0x2028:    /* LINE SEPARATOR */
1623          case 0x2029:    /* PARAGRAPH SEPARATOR */
1624          RRETURN(MATCH_NOMATCH);
1625          }
1626        ecode++;
1627        break;
1628    
1629        case OP_VSPACE:
1630        if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
1631        GETCHARINCTEST(c, eptr);
1632        switch(c)
1633          {
1634          default: RRETURN(MATCH_NOMATCH);
1635          case 0x0a:      /* LF */
1636          case 0x0b:      /* VT */
1637          case 0x0c:      /* FF */
1638          case 0x0d:      /* CR */
1639          case 0x85:      /* NEL */
1640          case 0x2028:    /* LINE SEPARATOR */
1641          case 0x2029:    /* PARAGRAPH SEPARATOR */
1642        break;        break;
1643        }        }
1644      ecode++;      ecode++;
# Line 1554  for (;;) Line 1730  for (;;)
1730      case OP_REF:      case OP_REF:
1731        {        {
1732        offset = GET2(ecode, 1) << 1;               /* Doubled ref number */        offset = GET2(ecode, 1) << 1;               /* Doubled ref number */
1733        ecode += 3;                                 /* Advance past item */        ecode += 3;
1734    
1735          /* If the reference is unset, there are two possibilities:
1736    
1737          (a) In the default, Perl-compatible state, set the length to be longer
1738          than the amount of subject left; this ensures that every attempt at a
1739          match fails. We can't just fail here, because of the possibility of
1740          quantifiers with zero minima.
1741    
1742        /* If the reference is unset, set the length to be longer than the amount        (b) If the JavaScript compatibility flag is set, set the length to zero
1743        of subject left; this ensures that every attempt at a match fails. We        so that the back reference matches an empty string.
1744        can't just fail here, because of the possibility of quantifiers with zero  
1745        minima. */        Otherwise, set the length to the length of what was matched by the
1746          referenced subpattern. */
1747        length = (offset >= offset_top || md->offset_vector[offset] < 0)?  
1748          md->end_subject - eptr + 1 :        if (offset >= offset_top || md->offset_vector[offset] < 0)
1749          md->offset_vector[offset+1] - md->offset_vector[offset];          length = (md->jscript_compat)? 0 : md->end_subject - eptr + 1;
1750          else
1751            length = md->offset_vector[offset+1] - md->offset_vector[offset];
1752    
1753        /* Set up for repetition, or handle the non-repeated case */        /* Set up for repetition, or handle the non-repeated case */
1754    
# Line 1923  for (;;) Line 2108  for (;;)
2108            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM21);            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM21);
2109            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2110            if (eptr-- == pp) break;        /* Stop if tried at original pos */            if (eptr-- == pp) break;        /* Stop if tried at original pos */
2111            BACKCHAR(eptr)            if (utf8) BACKCHAR(eptr);
2112            }            }
2113          RRETURN(MATCH_NOMATCH);          RRETURN(MATCH_NOMATCH);
2114          }          }
# Line 2676  for (;;) Line 2861  for (;;)
2861            for (i = 1; i <= min; i++)            for (i = 1; i <= min; i++)
2862              {              {
2863              if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);              if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
2864              GETCHARINC(c, eptr);              GETCHARINCTEST(c, eptr);
2865              }              }
2866            break;            break;
2867    
# Line 2684  for (;;) Line 2869  for (;;)
2869            for (i = 1; i <= min; i++)            for (i = 1; i <= min; i++)
2870              {              {
2871              if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);              if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
2872              GETCHARINC(c, eptr);              GETCHARINCTEST(c, eptr);
2873              prop_category = _pcre_ucp_findprop(c, &prop_chartype, &prop_script);              prop_category = _pcre_ucp_findprop(c, &prop_chartype, &prop_script);
2874              if ((prop_chartype == ucp_Lu ||              if ((prop_chartype == ucp_Lu ||
2875                   prop_chartype == ucp_Ll ||                   prop_chartype == ucp_Ll ||
# Line 2697  for (;;) Line 2882  for (;;)
2882            for (i = 1; i <= min; i++)            for (i = 1; i <= min; i++)
2883              {              {
2884              if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);              if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
2885              GETCHARINC(c, eptr);              GETCHARINCTEST(c, eptr);
2886              prop_category = _pcre_ucp_findprop(c, &prop_chartype, &prop_script);              prop_category = _pcre_ucp_findprop(c, &prop_chartype, &prop_script);
2887              if ((prop_category == prop_value) == prop_fail_result)              if ((prop_category == prop_value) == prop_fail_result)
2888                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
# Line 2708  for (;;) Line 2893  for (;;)
2893            for (i = 1; i <= min; i++)            for (i = 1; i <= min; i++)
2894              {              {
2895              if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);              if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
2896              GETCHARINC(c, eptr);              GETCHARINCTEST(c, eptr);
2897              prop_category = _pcre_ucp_findprop(c, &prop_chartype, &prop_script);              prop_category = _pcre_ucp_findprop(c, &prop_chartype, &prop_script);
2898              if ((prop_chartype == prop_value) == prop_fail_result)              if ((prop_chartype == prop_value) == prop_fail_result)
2899                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
# Line 2719  for (;;) Line 2904  for (;;)
2904            for (i = 1; i <= min; i++)            for (i = 1; i <= min; i++)
2905              {              {
2906              if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);              if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
2907              GETCHARINC(c, eptr);              GETCHARINCTEST(c, eptr);
2908              prop_category = _pcre_ucp_findprop(c, &prop_chartype, &prop_script);              prop_category = _pcre_ucp_findprop(c, &prop_chartype, &prop_script);
2909              if ((prop_script == prop_value) == prop_fail_result)              if ((prop_script == prop_value) == prop_fail_result)
2910                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
# Line 2766  for (;;) Line 2951  for (;;)
2951          case OP_ANY:          case OP_ANY:
2952          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
2953            {            {
2954            if (eptr >= md->end_subject ||            if (eptr >= md->end_subject || IS_NEWLINE(eptr))
                ((ims & PCRE_DOTALL) == 0 && IS_NEWLINE(eptr)))  
2955              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
2956            eptr++;            eptr++;
2957            while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;            while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;
2958            }            }
2959          break;          break;
2960    
2961            case OP_ALLANY:
2962            for (i = 1; i <= min; i++)
2963              {
2964              if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
2965              eptr++;
2966              while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;
2967              }
2968            break;
2969    
2970          case OP_ANYBYTE:          case OP_ANYBYTE:
2971          eptr += min;          eptr += min;
2972          break;          break;
# Line 2789  for (;;) Line 2982  for (;;)
2982              case 0x000d:              case 0x000d:
2983              if (eptr < md->end_subject && *eptr == 0x0a) eptr++;              if (eptr < md->end_subject && *eptr == 0x0a) eptr++;
2984              break;              break;
2985    
2986              case 0x000a:              case 0x000a:
2987                break;
2988    
2989              case 0x000b:              case 0x000b:
2990              case 0x000c:              case 0x000c:
2991              case 0x0085:              case 0x0085:
2992              case 0x2028:              case 0x2028:
2993              case 0x2029:              case 0x2029:
2994                if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);
2995                break;
2996                }
2997              }
2998            break;
2999    
3000            case OP_NOT_HSPACE:
3001            for (i = 1; i <= min; i++)
3002              {
3003              if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
3004              GETCHARINC(c, eptr);
3005              switch(c)
3006                {
3007                default: break;
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                RRETURN(MATCH_NOMATCH);
3028                }
3029              }
3030            break;
3031    
3032            case OP_HSPACE:
3033            for (i = 1; i <= min; i++)
3034              {
3035              if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
3036              GETCHARINC(c, eptr);
3037              switch(c)
3038                {
3039                default: RRETURN(MATCH_NOMATCH);
3040                case 0x09:      /* HT */
3041                case 0x20:      /* SPACE */
3042                case 0xa0:      /* NBSP */
3043                case 0x1680:    /* OGHAM SPACE MARK */
3044                case 0x180e:    /* MONGOLIAN VOWEL SEPARATOR */
3045                case 0x2000:    /* EN QUAD */
3046                case 0x2001:    /* EM QUAD */
3047                case 0x2002:    /* EN SPACE */
3048                case 0x2003:    /* EM SPACE */
3049                case 0x2004:    /* THREE-PER-EM SPACE */
3050                case 0x2005:    /* FOUR-PER-EM SPACE */
3051                case 0x2006:    /* SIX-PER-EM SPACE */
3052                case 0x2007:    /* FIGURE SPACE */
3053                case 0x2008:    /* PUNCTUATION SPACE */
3054                case 0x2009:    /* THIN SPACE */
3055                case 0x200A:    /* HAIR SPACE */
3056                case 0x202f:    /* NARROW NO-BREAK SPACE */
3057                case 0x205f:    /* MEDIUM MATHEMATICAL SPACE */
3058                case 0x3000:    /* IDEOGRAPHIC SPACE */
3059                break;
3060                }
3061              }
3062            break;
3063    
3064            case OP_NOT_VSPACE:
3065            for (i = 1; i <= min; i++)
3066              {
3067              if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
3068              GETCHARINC(c, eptr);
3069              switch(c)
3070                {
3071                default: break;
3072                case 0x0a:      /* LF */
3073                case 0x0b:      /* VT */
3074                case 0x0c:      /* FF */
3075                case 0x0d:      /* CR */
3076                case 0x85:      /* NEL */
3077                case 0x2028:    /* LINE SEPARATOR */
3078                case 0x2029:    /* PARAGRAPH SEPARATOR */
3079                RRETURN(MATCH_NOMATCH);
3080                }
3081              }
3082            break;
3083    
3084            case OP_VSPACE:
3085            for (i = 1; i <= min; i++)
3086              {
3087              if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
3088              GETCHARINC(c, eptr);
3089              switch(c)
3090                {
3091                default: RRETURN(MATCH_NOMATCH);
3092                case 0x0a:      /* LF */
3093                case 0x0b:      /* VT */
3094                case 0x0c:      /* FF */
3095                case 0x0d:      /* CR */
3096                case 0x85:      /* NEL */
3097                case 0x2028:    /* LINE SEPARATOR */
3098                case 0x2029:    /* PARAGRAPH SEPARATOR */
3099              break;              break;
3100              }              }
3101            }            }
# Line 2824  for (;;) Line 3125  for (;;)
3125          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
3126            {            {
3127            if (eptr >= md->end_subject ||            if (eptr >= md->end_subject ||
3128               (*eptr < 128 && (md->ctypes[*eptr++] & ctype_space) != 0))               (*eptr < 128 && (md->ctypes[*eptr] & ctype_space) != 0))
3129              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
3130            while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;            while (++eptr < md->end_subject && (*eptr & 0xc0) == 0x80);
3131            }            }
3132          break;          break;
3133    
# Line 2844  for (;;) Line 3145  for (;;)
3145          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
3146            {            {
3147            if (eptr >= md->end_subject ||            if (eptr >= md->end_subject ||
3148               (*eptr < 128 && (md->ctypes[*eptr++] & ctype_word) != 0))               (*eptr < 128 && (md->ctypes[*eptr] & ctype_word) != 0))
3149              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
3150            while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;            while (++eptr < md->end_subject && (*eptr & 0xc0) == 0x80);
3151            }            }
3152          break;          break;
3153    
# Line 2874  for (;;) Line 3175  for (;;)
3175        switch(ctype)        switch(ctype)
3176          {          {
3177          case OP_ANY:          case OP_ANY:
3178          if ((ims & PCRE_DOTALL) == 0)          for (i = 1; i <= min; i++)
3179            {            {
3180            for (i = 1; i <= min; i++)            if (IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH);
3181              {            eptr++;
             if (IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH);  
             eptr++;  
             }  
3182            }            }
3183          else eptr += min;          break;
3184    
3185            case OP_ALLANY:
3186            eptr += min;
3187          break;          break;
3188    
3189          case OP_ANYBYTE:          case OP_ANYBYTE:
# Line 2903  for (;;) Line 3204  for (;;)
3204              if (eptr < md->end_subject && *eptr == 0x0a) eptr++;              if (eptr < md->end_subject && *eptr == 0x0a) eptr++;
3205              break;              break;
3206              case 0x000a:              case 0x000a:
3207                break;
3208    
3209              case 0x000b:              case 0x000b:
3210              case 0x000c:              case 0x000c:
3211              case 0x0085:              case 0x0085:
3212                if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);
3213                break;
3214                }
3215              }
3216            break;
3217    
3218            case OP_NOT_HSPACE:
3219            for (i = 1; i <= min; i++)
3220              {
3221              if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
3222              switch(*eptr++)
3223                {
3224                default: break;
3225                case 0x09:      /* HT */
3226                case 0x20:      /* SPACE */
3227                case 0xa0:      /* NBSP */
3228                RRETURN(MATCH_NOMATCH);
3229                }
3230              }
3231            break;
3232    
3233            case OP_HSPACE:
3234            for (i = 1; i <= min; i++)
3235              {
3236              if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
3237              switch(*eptr++)
3238                {
3239                default: RRETURN(MATCH_NOMATCH);
3240                case 0x09:      /* HT */
3241                case 0x20:      /* SPACE */
3242                case 0xa0:      /* NBSP */
3243                break;
3244                }
3245              }
3246            break;
3247    
3248            case OP_NOT_VSPACE:
3249            for (i = 1; i <= min; i++)
3250              {
3251              if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
3252              switch(*eptr++)
3253                {
3254                default: break;
3255                case 0x0a:      /* LF */
3256                case 0x0b:      /* VT */
3257                case 0x0c:      /* FF */
3258                case 0x0d:      /* CR */
3259                case 0x85:      /* NEL */
3260                RRETURN(MATCH_NOMATCH);
3261                }
3262              }
3263            break;
3264    
3265            case OP_VSPACE:
3266            for (i = 1; i <= min; i++)
3267              {
3268              if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
3269              switch(*eptr++)
3270                {
3271                default: RRETURN(MATCH_NOMATCH);
3272                case 0x0a:      /* LF */
3273                case 0x0b:      /* VT */
3274                case 0x0c:      /* FF */
3275                case 0x0d:      /* CR */
3276                case 0x85:      /* NEL */
3277              break;              break;
3278              }              }
3279            }            }
# Line 3072  for (;;) Line 3440  for (;;)
3440            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM42);            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM42);
3441            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3442            if (fi >= max || eptr >= md->end_subject ||            if (fi >= max || eptr >= md->end_subject ||
3443                 (ctype == OP_ANY && (ims & PCRE_DOTALL) == 0 &&                 (ctype == OP_ANY && IS_NEWLINE(eptr)))
                 IS_NEWLINE(eptr)))  
3444              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
3445    
3446            GETCHARINC(c, eptr);            GETCHARINC(c, eptr);
3447            switch(ctype)            switch(ctype)
3448              {              {
3449              case OP_ANY:        /* This is the DOTALL case */              case OP_ANY:        /* This is the non-NL case */
3450              break;              case OP_ALLANY:
   
3451              case OP_ANYBYTE:              case OP_ANYBYTE:
3452              break;              break;
3453    
# Line 3093  for (;;) Line 3459  for (;;)
3459                if (eptr < md->end_subject && *eptr == 0x0a) eptr++;                if (eptr < md->end_subject && *eptr == 0x0a) eptr++;
3460                break;                break;
3461                case 0x000a:                case 0x000a:
3462                  break;
3463    
3464                case 0x000b:                case 0x000b:
3465                case 0x000c:                case 0x000c:
3466                case 0x0085:                case 0x0085:
3467                case 0x2028:                case 0x2028:
3468                case 0x2029:                case 0x2029:
3469                  if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);
3470                  break;
3471                  }
3472                break;
3473    
3474                case OP_NOT_HSPACE:
3475                switch(c)
3476                  {
3477                  default: break;
3478                  case 0x09:      /* HT */
3479                  case 0x20:      /* SPACE */
3480                  case 0xa0:      /* NBSP */
3481                  case 0x1680:    /* OGHAM SPACE MARK */
3482                  case 0x180e:    /* MONGOLIAN VOWEL SEPARATOR */
3483                  case 0x2000:    /* EN QUAD */
3484                  case 0x2001:    /* EM QUAD */
3485                  case 0x2002:    /* EN SPACE */
3486                  case 0x2003:    /* EM SPACE */
3487                  case 0x2004:    /* THREE-PER-EM SPACE */
3488                  case 0x2005:    /* FOUR-PER-EM SPACE */
3489                  case 0x2006:    /* SIX-PER-EM SPACE */
3490                  case 0x2007:    /* FIGURE SPACE */
3491                  case 0x2008:    /* PUNCTUATION SPACE */
3492                  case 0x2009:    /* THIN SPACE */
3493                  case 0x200A:    /* HAIR SPACE */
3494                  case 0x202f:    /* NARROW NO-BREAK SPACE */
3495                  case 0x205f:    /* MEDIUM MATHEMATICAL SPACE */
3496                  case 0x3000:    /* IDEOGRAPHIC SPACE */
3497                  RRETURN(MATCH_NOMATCH);
3498                  }
3499                break;
3500    
3501                case OP_HSPACE:
3502                switch(c)
3503                  {
3504                  default: RRETURN(MATCH_NOMATCH);
3505                  case 0x09:      /* HT */
3506                  case 0x20:      /* SPACE */
3507                  case 0xa0:      /* NBSP */
3508                  case 0x1680:    /* OGHAM SPACE MARK */
3509                  case 0x180e:    /* MONGOLIAN VOWEL SEPARATOR */
3510                  case 0x2000:    /* EN QUAD */
3511                  case 0x2001:    /* EM QUAD */
3512                  case 0x2002:    /* EN SPACE */
3513                  case 0x2003:    /* EM SPACE */
3514                  case 0x2004:    /* THREE-PER-EM SPACE */
3515                  case 0x2005:    /* FOUR-PER-EM SPACE */
3516                  case 0x2006:    /* SIX-PER-EM SPACE */
3517                  case 0x2007:    /* FIGURE SPACE */
3518                  case 0x2008:    /* PUNCTUATION SPACE */
3519                  case 0x2009:    /* THIN SPACE */
3520                  case 0x200A:    /* HAIR SPACE */
3521                  case 0x202f:    /* NARROW NO-BREAK SPACE */
3522                  case 0x205f:    /* MEDIUM MATHEMATICAL SPACE */
3523                  case 0x3000:    /* IDEOGRAPHIC SPACE */
3524                  break;
3525                  }
3526                break;
3527    
3528                case OP_NOT_VSPACE:
3529                switch(c)
3530                  {
3531                  default: break;
3532                  case 0x0a:      /* LF */
3533                  case 0x0b:      /* VT */
3534                  case 0x0c:      /* FF */
3535                  case 0x0d:      /* CR */
3536                  case 0x85:      /* NEL */
3537                  case 0x2028:    /* LINE SEPARATOR */
3538                  case 0x2029:    /* PARAGRAPH SEPARATOR */
3539                  RRETURN(MATCH_NOMATCH);
3540                  }
3541                break;
3542    
3543                case OP_VSPACE:
3544                switch(c)
3545                  {
3546                  default: RRETURN(MATCH_NOMATCH);
3547                  case 0x0a:      /* LF */
3548                  case 0x0b:      /* VT */
3549                  case 0x0c:      /* FF */
3550                  case 0x0d:      /* CR */
3551                  case 0x85:      /* NEL */
3552                  case 0x2028:    /* LINE SEPARATOR */
3553                  case 0x2029:    /* PARAGRAPH SEPARATOR */
3554                break;                break;
3555                }                }
3556              break;              break;
# Line 3146  for (;;) Line 3599  for (;;)
3599            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM43);            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM43);
3600            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3601            if (fi >= max || eptr >= md->end_subject ||            if (fi >= max || eptr >= md->end_subject ||
3602                 ((ims & PCRE_DOTALL) == 0 && IS_NEWLINE(eptr)))                 (ctype == OP_ANY && IS_NEWLINE(eptr)))
3603              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
3604    
3605            c = *eptr++;            c = *eptr++;
3606            switch(ctype)            switch(ctype)
3607              {              {
3608              case OP_ANY:   /* This is the DOTALL case */              case OP_ANY:     /* This is the non-NL case */
3609              break;              case OP_ALLANY:
   
3610              case OP_ANYBYTE:              case OP_ANYBYTE:
3611              break;              break;
3612    
# Line 3165  for (;;) Line 3617  for (;;)
3617                case 0x000d:                case 0x000d:
3618                if (eptr < md->end_subject && *eptr == 0x0a) eptr++;                if (eptr < md->end_subject && *eptr == 0x0a) eptr++;
3619                break;                break;
3620    
3621                case 0x000a:                case 0x000a:
3622                  break;
3623    
3624                case 0x000b:                case 0x000b:
3625                case 0x000c:                case 0x000c:
3626                case 0x0085:                case 0x0085:
3627                  if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);
3628                  break;
3629                  }
3630                break;
3631    
3632                case OP_NOT_HSPACE:
3633                switch(c)
3634                  {
3635                  default: break;
3636                  case 0x09:      /* HT */
3637                  case 0x20:      /* SPACE */
3638                  case 0xa0:      /* NBSP */
3639                  RRETURN(MATCH_NOMATCH);
3640                  }
3641                break;
3642    
3643                case OP_HSPACE:
3644                switch(c)
3645                  {
3646                  default: RRETURN(MATCH_NOMATCH);
3647                  case 0x09:      /* HT */
3648                  case 0x20:      /* SPACE */
3649                  case 0xa0:      /* NBSP */
3650                  break;
3651                  }
3652                break;
3653    
3654                case OP_NOT_VSPACE:
3655                switch(c)
3656                  {
3657                  default: break;
3658                  case 0x0a:      /* LF */
3659                  case 0x0b:      /* VT */
3660                  case 0x0c:      /* FF */
3661                  case 0x0d:      /* CR */
3662                  case 0x85:      /* NEL */
3663                  RRETURN(MATCH_NOMATCH);
3664                  }
3665                break;
3666    
3667                case OP_VSPACE:
3668                switch(c)
3669                  {
3670                  default: RRETURN(MATCH_NOMATCH);
3671                  case 0x0a:      /* LF */
3672                  case 0x0b:      /* VT */
3673                  case 0x0c:      /* FF */
3674                  case 0x0d:      /* CR */
3675                  case 0x85:      /* NEL */
3676                break;                break;
3677                }                }
3678              break;              break;
# Line 3292  for (;;) Line 3796  for (;;)
3796            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM44);            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM44);
3797            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3798            if (eptr-- == pp) break;        /* Stop if tried at original pos */            if (eptr-- == pp) break;        /* Stop if tried at original pos */
3799            BACKCHAR(eptr);            if (utf8) BACKCHAR(eptr);
3800            }            }
3801          }          }
3802    
# Line 3331  for (;;) Line 3835  for (;;)
3835            for (;;)                        /* Move back over one extended */            for (;;)                        /* Move back over one extended */
3836              {              {
3837              int len = 1;              int len = 1;
             BACKCHAR(eptr);  
3838              if (!utf8) c = *eptr; else              if (!utf8) c = *eptr; else
3839                {                {
3840                  BACKCHAR(eptr);
3841                GETCHARLEN(c, eptr, len);                GETCHARLEN(c, eptr, len);
3842                }                }
3843              prop_category = _pcre_ucp_findprop(c, &prop_chartype, &prop_script);              prop_category = _pcre_ucp_findprop(c, &prop_chartype, &prop_script);
# Line 3354  for (;;) Line 3858  for (;;)
3858          switch(ctype)          switch(ctype)
3859            {            {
3860            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. */  
   
3861            if (max < INT_MAX)            if (max < INT_MAX)
3862              {              {
3863              if ((ims & PCRE_DOTALL) == 0)              for (i = min; i < max; i++)
               {  
               for (i = min; i < max; i++)  
                 {  
                 if (eptr >= md->end_subject || IS_NEWLINE(eptr)) break;  
                 eptr++;  
                 while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;  
                 }  
               }  
             else  
3864                {                {
3865                for (i = min; i < max; i++)                if (eptr >= md->end_subject || IS_NEWLINE(eptr)) break;
3866                  {                eptr++;
3867                  if (eptr >= md->end_subject) break;                while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;
                 eptr++;  
                 while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;  
                 }  
3868                }                }
3869              }              }
3870    
# Line 3385  for (;;) Line 3872  for (;;)
3872    
3873            else            else
3874              {              {
3875              if ((ims & PCRE_DOTALL) == 0)              for (i = min; i < max; i++)
3876                {                {
3877                for (i = min; i < max; i++)                if (eptr >= md->end_subject || IS_NEWLINE(eptr)) break;
3878                  {                eptr++;
3879                  if (eptr >= md->end_subject || IS_NEWLINE(eptr)) break;                while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;
                 eptr++;  
                 }  
               break;  
3880                }                }
3881              else              }
3882              break;
3883    
3884              case OP_ALLANY:
3885              if (max < INT_MAX)
3886                {
3887                for (i = min; i < max; i++)
3888                {                {
3889                c = max - min;                if (eptr >= md->end_subject) break;
3890                if (c > (unsigned int)(md->end_subject - eptr))                eptr++;
3891                  c = md->end_subject - eptr;                while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;
               eptr += c;  
3892                }                }
3893              }              }
3894              else eptr = md->end_subject;   /* Unlimited UTF-8 repeat */
3895            break;            break;
3896    
3897            /* The byte case is the same as non-UTF8 */            /* The byte case is the same as non-UTF8 */
# Line 3426  for (;;) Line 3916  for (;;)
3916                }                }
3917              else              else
3918                {                {
3919                if (c != 0x000a && c != 0x000b && c != 0x000c &&                if (c != 0x000a &&
3920                    c != 0x0085 && c != 0x2028 && c != 0x2029)                    (md->bsr_anycrlf ||
3921                       (c != 0x000b && c != 0x000c &&
3922                        c != 0x0085 && c != 0x2028 && c != 0x2029)))
3923                  break;                  break;
3924                eptr += len;                eptr += len;
3925                }                }
3926              }              }
3927            break;            break;
3928    
3929              case OP_NOT_HSPACE:
3930              case OP_HSPACE:
3931              for (i = min; i < max; i++)
3932                {
3933                BOOL gotspace;
3934                int len = 1;
3935                if (eptr >= md->end_subject) break;
3936                GETCHARLEN(c, eptr, len);
3937                switch(c)
3938                  {
3939                  default: gotspace = FALSE; break;
3940                  case 0x09:      /* HT */
3941                  case 0x20:      /* SPACE */
3942                  case 0xa0:      /* NBSP */
3943                  case 0x1680:    /* OGHAM SPACE MARK */
3944                  case 0x180e:    /* MONGOLIAN VOWEL SEPARATOR */
3945                  case 0x2000:    /* EN QUAD */
3946                  case 0x2001:    /* EM QUAD */
3947                  case 0x2002:    /* EN SPACE */
3948                  case 0x2003:    /* EM SPACE */
3949                  case 0x2004:    /* THREE-PER-EM SPACE */
3950                  case 0x2005:    /* FOUR-PER-EM SPACE */
3951                  case 0x2006:    /* SIX-PER-EM SPACE */
3952                  case 0x2007:    /* FIGURE SPACE */
3953                  case 0x2008:    /* PUNCTUATION SPACE */
3954                  case 0x2009:    /* THIN SPACE */
3955                  case 0x200A:    /* HAIR SPACE */
3956                  case 0x202f:    /* NARROW NO-BREAK SPACE */
3957                  case 0x205f:    /* MEDIUM MATHEMATICAL SPACE */
3958                  case 0x3000:    /* IDEOGRAPHIC SPACE */
3959                  gotspace = TRUE;
3960                  break;
3961                  }
3962                if (gotspace == (ctype == OP_NOT_HSPACE)) break;
3963                eptr += len;
3964                }
3965              break;
3966    
3967              case OP_NOT_VSPACE:
3968              case OP_VSPACE:
3969              for (i = min; i < max; i++)
3970                {
3971                BOOL gotspace;
3972                int len = 1;
3973                if (eptr >= md->end_subject) break;
3974                GETCHARLEN(c, eptr, len);
3975                switch(c)
3976                  {
3977                  default: gotspace = FALSE; break;
3978                  case 0x0a:      /* LF */
3979                  case 0x0b:      /* VT */
3980                  case 0x0c:      /* FF */
3981                  case 0x0d:      /* CR */
3982                  case 0x85:      /* NEL */
3983                  case 0x2028:    /* LINE SEPARATOR */
3984                  case 0x2029:    /* PARAGRAPH SEPARATOR */
3985                  gotspace = TRUE;
3986                  break;
3987                  }
3988                if (gotspace == (ctype == OP_NOT_VSPACE)) break;
3989                eptr += len;
3990                }
3991              break;
3992    
3993            case OP_NOT_DIGIT:            case OP_NOT_DIGIT:
3994            for (i = min; i < max; i++)            for (i = min; i < max; i++)
3995              {              {
# Line 3516  for (;;) Line 4072  for (;;)
4072            }            }
4073          }          }
4074        else        else
4075  #endif  #endif  /* SUPPORT_UTF8 */
4076    
4077        /* Not UTF-8 mode */        /* Not UTF-8 mode */
4078          {          {
4079          switch(ctype)          switch(ctype)
4080            {            {
4081            case OP_ANY:            case OP_ANY:
4082            if ((ims & PCRE_DOTALL) == 0)            for (i = min; i < max; i++)
4083              {              {
4084              for (i = min; i < max; i++)              if (eptr >= md->end_subject || IS_NEWLINE(eptr)) break;
4085                {              eptr++;
               if (eptr >= md->end_subject || IS_NEWLINE(eptr)) break;  
               eptr++;  
               }  
             break;  
4086              }              }
4087            /* For DOTALL case, fall through and treat as \C */            break;
4088    
4089              case OP_ALLANY:
4090            case OP_ANYBYTE:            case OP_ANYBYTE:
4091            c = max - min;            c = max - min;
4092            if (c > (unsigned int)(md->end_subject - eptr))            if (c > (unsigned int)(md->end_subject - eptr))
# Line 3553  for (;;) Line 4106  for (;;)
4106                }                }
4107              else              else
4108                {                {
4109                if (c != 0x000a && c != 0x000b && c != 0x000c && c != 0x0085)                if (c != 0x000a &&
4110                      (md->bsr_anycrlf ||
4111                        (c != 0x000b && c != 0x000c && c != 0x0085)))
4112                  break;                  break;
4113                eptr++;                eptr++;
4114                }                }
4115              }              }
4116            break;            break;
4117    
4118              case OP_NOT_HSPACE:
4119              for (i = min; i < max; i++)
4120                {
4121                if (eptr >= md->end_subject) break;
4122                c = *eptr;
4123                if (c == 0x09 || c == 0x20 || c == 0xa0) break;
4124                eptr++;
4125                }
4126              break;
4127    
4128              case OP_HSPACE:
4129              for (i = min; i < max; i++)
4130                {
4131                if (eptr >= md->end_subject) break;
4132                c = *eptr;
4133                if (c != 0x09 && c != 0x20 && c != 0xa0) break;
4134                eptr++;
4135                }
4136              break;
4137    
4138              case OP_NOT_VSPACE:
4139              for (i = min; i < max; i++)
4140                {
4141                if (eptr >= md->end_subject) break;
4142                c = *eptr;
4143                if (c == 0x0a || c == 0x0b || c == 0x0c || c == 0x0d || c == 0x85)
4144                  break;
4145                eptr++;
4146                }
4147              break;
4148    
4149              case OP_VSPACE:
4150              for (i = min; i < max; i++)
4151                {
4152                if (eptr >= md->end_subject) break;
4153                c = *eptr;
4154                if (c != 0x0a && c != 0x0b && c != 0x0c && c != 0x0d && c != 0x85)
4155                  break;
4156                eptr++;
4157                }
4158              break;
4159    
4160            case OP_NOT_DIGIT:            case OP_NOT_DIGIT:
4161            for (i = min; i < max; i++)            for (i = min; i < max; i++)
4162              {              {
# Line 3661  HEAP_RETURN: Line 4258  HEAP_RETURN:
4258  switch (frame->Xwhere)  switch (frame->Xwhere)
4259    {    {
4260    LBL( 1) LBL( 2) LBL( 3) LBL( 4) LBL( 5) LBL( 6) LBL( 7) LBL( 8)    LBL( 1) LBL( 2) LBL( 3) LBL( 4) LBL( 5) LBL( 6) LBL( 7) LBL( 8)
4261    LBL( 9) LBL(10) LBL(11) LBL(12) LBL(13) LBL(14) LBL(15) LBL(16)    LBL( 9) LBL(10) LBL(11) LBL(12) LBL(13) LBL(14) LBL(15) LBL(17)
4262    LBL(17) LBL(18) LBL(19) LBL(20) LBL(21) LBL(22) LBL(23) LBL(24)    LBL(19) LBL(24) LBL(25) LBL(26) LBL(27) LBL(29) LBL(31) LBL(33)
4263    LBL(25) LBL(26) LBL(27) LBL(28) LBL(29) LBL(30) LBL(31) LBL(32)    LBL(35) LBL(43) LBL(47) LBL(48) LBL(49) LBL(50) LBL(51) LBL(52)
4264    LBL(33) LBL(34) LBL(35) LBL(36) LBL(37) LBL(38) LBL(39) LBL(40)    LBL(53) LBL(54)
4265    LBL(41) LBL(42) LBL(43) LBL(44) LBL(45) LBL(46) LBL(47)  #ifdef SUPPORT_UTF8
4266      LBL(16) LBL(18) LBL(20) LBL(21) LBL(22) LBL(23) LBL(28) LBL(30)
4267      LBL(32) LBL(34) LBL(42) LBL(46)
4268    #ifdef SUPPORT_UCP
4269      LBL(36) LBL(37) LBL(38) LBL(39) LBL(40) LBL(41) LBL(44) LBL(45)
4270    #endif  /* SUPPORT_UCP */
4271    #endif  /* SUPPORT_UTF8 */
4272    default:    default:
4273    DPRINTF(("jump error in pcre match: label %d non-existent\n", frame->Xwhere));    DPRINTF(("jump error in pcre match: label %d non-existent\n", frame->Xwhere));
4274    return PCRE_ERROR_INTERNAL;    return PCRE_ERROR_INTERNAL;
# Line 3684  Undefine all the macros that were define Line 4287  Undefine all the macros that were define
4287  #ifdef NO_RECURSE  #ifdef NO_RECURSE
4288  #undef eptr  #undef eptr
4289  #undef ecode  #undef ecode
4290    #undef mstart
4291  #undef offset_top  #undef offset_top
4292  #undef ims  #undef ims
4293  #undef eptrb  #undef eptrb
# Line 3781  const uschar *start_bits = NULL; Line 4385  const uschar *start_bits = NULL;
4385  USPTR start_match = (USPTR)subject + start_offset;  USPTR start_match = (USPTR)subject + start_offset;
4386  USPTR end_subject;  USPTR end_subject;
4387  USPTR req_byte_ptr = start_match - 1;  USPTR req_byte_ptr = start_match - 1;
 eptrblock eptrchain[EPTR_WORK_SIZE];  
4388    
4389  pcre_study_data internal_study;  pcre_study_data internal_study;
4390  const pcre_study_data *study;  const pcre_study_data *study;
# Line 3844  if (re->magic_number != MAGIC_NUMBER) Line 4447  if (re->magic_number != MAGIC_NUMBER)
4447  /* Set up other data */  /* Set up other data */
4448    
4449  anchored = ((re->options | options) & PCRE_ANCHORED) != 0;  anchored = ((re->options | options) & PCRE_ANCHORED) != 0;
4450  startline = (re->options & PCRE_STARTLINE) != 0;  startline = (re->flags & PCRE_STARTLINE) != 0;
4451  firstline = (re->options & PCRE_FIRSTLINE) != 0;  firstline = (re->options & PCRE_FIRSTLINE) != 0;
4452    
4453  /* The code starts after the real_pcre block and the capture name table. */  /* The code starts after the real_pcre block and the capture name table. */
# Line 3859  end_subject = md->end_subject; Line 4462  end_subject = md->end_subject;
4462    
4463  md->endonly = (re->options & PCRE_DOLLAR_ENDONLY) != 0;  md->endonly = (re->options & PCRE_DOLLAR_ENDONLY) != 0;
4464  utf8 = md->utf8 = (re->options & PCRE_UTF8) != 0;  utf8 = md->utf8 = (re->options & PCRE_UTF8) != 0;
4465    md->jscript_compat = (re->options & PCRE_JAVASCRIPT_COMPAT) != 0;
4466    
4467  md->notbol = (options & PCRE_NOTBOL) != 0;  md->notbol = (options & PCRE_NOTBOL) != 0;
4468  md->noteol = (options & PCRE_NOTEOL) != 0;  md->noteol = (options & PCRE_NOTEOL) != 0;
# Line 3867  md->partial = (options & PCRE_PARTIAL) ! Line 4471  md->partial = (options & PCRE_PARTIAL) !
4471  md->hitend = FALSE;  md->hitend = FALSE;
4472    
4473  md->recursive = NULL;                   /* No recursion at top level */  md->recursive = NULL;                   /* No recursion at top level */
 md->eptrchain = eptrchain;              /* Make workspace generally available */  
4474    
4475  md->lcc = tables + lcc_offset;  md->lcc = tables + lcc_offset;
4476  md->ctypes = tables + ctypes_offset;  md->ctypes = tables + ctypes_offset;
4477    
4478    /* Handle different \R options. */
4479    
4480    switch (options & (PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE))
4481      {
4482      case 0:
4483      if ((re->options & (PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE)) != 0)
4484        md->bsr_anycrlf = (re->options & PCRE_BSR_ANYCRLF) != 0;
4485      else
4486    #ifdef BSR_ANYCRLF
4487      md->bsr_anycrlf = TRUE;
4488    #else
4489      md->bsr_anycrlf = FALSE;
4490    #endif
4491      break;
4492    
4493      case PCRE_BSR_ANYCRLF:
4494      md->bsr_anycrlf = TRUE;
4495      break;
4496    
4497      case PCRE_BSR_UNICODE:
4498      md->bsr_anycrlf = FALSE;
4499      break;
4500    
4501      default: return PCRE_ERROR_BADNEWLINE;
4502      }
4503    
4504  /* Handle different types of newline. The three bits give eight cases. If  /* Handle different types of newline. The three bits give eight cases. If
4505  nothing is set at run time, whatever was used at compile time applies. */  nothing is set at run time, whatever was used at compile time applies. */
4506    
4507  switch ((((options & PCRE_NEWLINE_BITS) == 0)? re->options : (pcre_uint32)options) &  switch ((((options & PCRE_NEWLINE_BITS) == 0)? re->options :
4508         PCRE_NEWLINE_BITS)          (pcre_uint32)options) & PCRE_NEWLINE_BITS)
4509    {    {
4510    case 0: newline = NEWLINE; break;   /* Compile-time default */    case 0: newline = NEWLINE; break;   /* Compile-time default */
4511    case PCRE_NEWLINE_CR: newline = '\r'; break;    case PCRE_NEWLINE_CR: newline = '\r'; break;
# Line 3915  else Line 4544  else
4544  /* 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
4545  moment. */  moment. */
4546    
4547  if (md->partial && (re->options & PCRE_NOPARTIAL) != 0)  if (md->partial && (re->flags & PCRE_NOPARTIAL) != 0)
4548    return PCRE_ERROR_BADPARTIAL;    return PCRE_ERROR_BADPARTIAL;
4549    
4550  /* Check a UTF-8 string if required. Unfortunately there's no way of passing  /* Check a UTF-8 string if required. Unfortunately there's no way of passing
# Line 3992  studied, there may be a bitmap of possib Line 4621  studied, there may be a bitmap of possib
4621    
4622  if (!anchored)  if (!anchored)
4623    {    {
4624    if ((re->options & PCRE_FIRSTSET) != 0)    if ((re->flags & PCRE_FIRSTSET) != 0)
4625      {      {
4626      first_byte = re->first_byte & 255;      first_byte = re->first_byte & 255;
4627      if ((first_byte_caseless = ((re->first_byte & REQ_CASELESS) != 0)) == TRUE)      if ((first_byte_caseless = ((re->first_byte & REQ_CASELESS) != 0)) == TRUE)
# Line 4007  if (!anchored) Line 4636  if (!anchored)
4636  /* For anchored or unanchored matches, there may be a "last known required  /* For anchored or unanchored matches, there may be a "last known required
4637  character" set. */  character" set. */
4638    
4639  if ((re->options & PCRE_REQCHSET) != 0)  if ((re->flags & PCRE_REQCHSET) != 0)
4640    {    {
4641    req_byte = re->req_byte & 255;    req_byte = re->req_byte & 255;
4642    req_byte_caseless = (re->req_byte & REQ_CASELESS) != 0;    req_byte_caseless = (re->req_byte & REQ_CASELESS) != 0;
# Line 4023  the loop runs just once. */ Line 4652  the loop runs just once. */
4652  for(;;)  for(;;)
4653    {    {
4654    USPTR save_end_subject = end_subject;    USPTR save_end_subject = end_subject;
4655      USPTR new_start_match;
4656    
4657    /* Reset the maximum number of extractions we might see. */    /* Reset the maximum number of extractions we might see. */
4658    
# Line 4053  for(;;) Line 4683  for(;;)
4683      if (first_byte_caseless)      if (first_byte_caseless)
4684        while (start_match < end_subject &&        while (start_match < end_subject &&
4685               md->lcc[*start_match] != first_byte)               md->lcc[*start_match] != first_byte)
4686          start_match++;          { NEXTCHAR(start_match); }
4687      else      else
4688        while (start_match < end_subject && *start_match != first_byte)        while (start_match < end_subject && *start_match != first_byte)
4689          start_match++;          { NEXTCHAR(start_match); }
4690      }      }
4691    
4692    /* Or to just after a linebreak for a multiline match if possible */    /* Or to just after a linebreak for a multiline match if possible */
# Line 4066  for(;;) Line 4696  for(;;)
4696      if (start_match > md->start_subject + start_offset)      if (start_match > md->start_subject + start_offset)
4697        {        {
4698        while (start_match <= end_subject && !WAS_NEWLINE(start_match))        while (start_match <= end_subject && !WAS_NEWLINE(start_match))
4699          start_match++;          { NEXTCHAR(start_match); }
4700    
4701        /* If we have just passed a CR and the newline option is ANY or ANYCRLF,        /* If we have just passed a CR and the newline option is ANY or ANYCRLF,
4702        and we are now at a LF, advance the match position by one more character.        and we are now at a LF, advance the match position by one more character.
# Line 4087  for(;;) Line 4717  for(;;)
4717      while (start_match < end_subject)      while (start_match < end_subject)
4718        {        {
4719        register unsigned int c = *start_match;        register unsigned int c = *start_match;
4720        if ((start_bits[c/8] & (1 << (c&7))) == 0) start_match++; else break;        if ((start_bits[c/8] & (1 << (c&7))) == 0)
4721            { NEXTCHAR(start_match); }
4722          else break;
4723        }        }
4724      }      }
4725    
# Line 4163  for(;;) Line 4795  for(;;)
4795    
4796    /* OK, we can now run the match. */    /* OK, we can now run the match. */
4797    
4798    md->start_match = start_match;    md->start_match_ptr = start_match;
4799    md->match_call_count = 0;    md->match_call_count = 0;
4800    md->eptrn = 0;                          /* Next free eptrchain slot */    rc = match(start_match, md->start_code, start_match, 2, md, ims, NULL, 0, 0);
   rc = match(start_match, md->start_code, 2, md, ims, NULL, 0, 0);  
4801    
4802    /* Any return other than MATCH_NOMATCH breaks the loop. */    switch(rc)
4803        {
4804        /* NOMATCH and PRUNE advance by one character. THEN at this level acts
4805        exactly like PRUNE. */
4806    
4807    if (rc != MATCH_NOMATCH) break;      case MATCH_NOMATCH:
4808        case MATCH_PRUNE:
4809        case MATCH_THEN:
4810        new_start_match = start_match + 1;
4811    #ifdef SUPPORT_UTF8
4812        if (utf8)
4813          while(new_start_match < end_subject && (*new_start_match & 0xc0) == 0x80)
4814            new_start_match++;
4815    #endif
4816        break;
4817    
4818        /* SKIP passes back the next starting point explicitly. */
4819    
4820        case MATCH_SKIP:
4821        new_start_match = md->start_match_ptr;
4822        break;
4823    
4824        /* COMMIT disables the bumpalong, but otherwise behaves as NOMATCH. */
4825    
4826        case MATCH_COMMIT:
4827        rc = MATCH_NOMATCH;
4828        goto ENDLOOP;
4829    
4830        /* Any other return is some kind of error. */
4831    
4832        default:
4833        goto ENDLOOP;
4834        }
4835    
4836      /* Control reaches here for the various types of "no match at this point"
4837      result. Reset the code to MATCH_NOMATCH for subsequent checking. */
4838    
4839      rc = MATCH_NOMATCH;
4840    
4841    /* If PCRE_FIRSTLINE is set, the match must happen before or at the first    /* If PCRE_FIRSTLINE is set, the match must happen before or at the first
4842    newline in the subject (though it may continue over the newline). Therefore,    newline in the subject (though it may continue over the newline). Therefore,
# Line 4178  for(;;) Line 4844  for(;;)
4844    
4845    if (firstline && IS_NEWLINE(start_match)) break;    if (firstline && IS_NEWLINE(start_match)) break;
4846    
4847    /* Advance the match position by one character. */    /* Advance to new matching position */
4848    
4849    start_match++;    start_match = new_start_match;
 #ifdef SUPPORT_UTF8  
   if (utf8)  
     while(start_match < end_subject && (*start_match & 0xc0) == 0x80)  
       start_match++;  
 #endif  
4850    
4851    /* Break the loop if the pattern is anchored or if we have passed the end of    /* Break the loop if the pattern is anchored or if we have passed the end of
4852    the subject. */    the subject. */
4853    
4854    if (anchored || start_match > end_subject) break;    if (anchored || start_match > end_subject) break;
4855    
4856    /* If we have just passed a CR and the newline option is CRLF or ANY or    /* If we have just passed a CR and we are now at a LF, and the pattern does
4857    ANYCRLF, and we are now at a LF, advance the match position by one more    not contain any explicit matches for \r or \n, and the newline option is CRLF
4858    character. */    or ANY or ANYCRLF, advance the match position by one more character. */
4859    
4860    if (start_match[-1] == '\r' &&    if (start_match[-1] == '\r' &&
4861         (md->nltype == NLTYPE_ANY ||        start_match < end_subject &&
4862          md->nltype == NLTYPE_ANYCRLF ||        *start_match == '\n' &&
4863          md->nllen == 2) &&        (re->flags & PCRE_HASCRORLF) == 0 &&
4864         start_match < end_subject &&          (md->nltype == NLTYPE_ANY ||
4865         *start_match == '\n')           md->nltype == NLTYPE_ANYCRLF ||
4866             md->nllen == 2))
4867      start_match++;      start_match++;
4868    
4869    }   /* End of for(;;) "bumpalong" loop */    }   /* End of for(;;) "bumpalong" loop */
# Line 4211  for(;;) Line 4873  for(;;)
4873  /* We reach here when rc is not MATCH_NOMATCH, or if one of the stopping  /* We reach here when rc is not MATCH_NOMATCH, or if one of the stopping
4874  conditions is true:  conditions is true:
4875    
4876  (1) The pattern is anchored;  (1) The pattern is anchored or the match was failed by (*COMMIT);
4877    
4878  (2) We are past the end of the subject;  (2) We are past the end of the subject;
4879    
# Line 4226  processing, copy those that we can. In t Line 4888  processing, copy those that we can. In t
4888  certain parts of the pattern were not used, even though there are more  certain parts of the pattern were not used, even though there are more
4889  capturing parentheses than vector slots. */  capturing parentheses than vector slots. */
4890    
4891    ENDLOOP:
4892    
4893  if (rc == MATCH_MATCH)  if (rc == MATCH_MATCH)
4894    {    {
4895    if (using_temporary_offsets)    if (using_temporary_offsets)
# Line 4246  if (rc == MATCH_MATCH) Line 4910  if (rc == MATCH_MATCH)
4910    
4911    rc = md->offset_overflow? 0 : md->end_offset_top/2;    rc = md->offset_overflow? 0 : md->end_offset_top/2;
4912    
4913    /* If there is space, set up the whole thing as substring 0. */    /* If there is space, set up the whole thing as substring 0. The value of
4914      md->start_match_ptr might be modified if \K was encountered on the success
4915      matching path. */
4916    
4917    if (offsetcount < 2) rc = 0; else    if (offsetcount < 2) rc = 0; else
4918      {      {
4919      offsets[0] = start_match - md->start_subject;      offsets[0] = md->start_match_ptr - md->start_subject;
4920      offsets[1] = md->end_match_ptr - md->start_subject;      offsets[1] = md->end_match_ptr - md->start_subject;
4921      }      }
4922    

Legend:
Removed from v.165  
changed lines
  Added in v.345

  ViewVC Help
Powered by ViewVC 1.1.5