/[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 149 by ph10, Mon Apr 16 15:28:08 2007 UTC revision 336 by ph10, Sat Apr 12 15:59:03 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 188  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 210  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 256  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 273  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 323  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 354  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 363  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 373  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 401  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 415  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 545  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 553  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    
 #ifdef SUPPORT_UTF8  
 utf8 = md->utf8;       /* Local copy of the flag */  
 #else  
 utf8 = FALSE;  
 #endif  
   
590  /* At the start of a group with an unlimited repeat that may match an empty  /* At the start of a group with an unlimited repeat that may match an empty
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 593  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 637  for (;;) Line 687  for (;;)
687        flags = (op == OP_SCBRA)? match_cbegroup : 0;        flags = (op == OP_SCBRA)? match_cbegroup : 0;
688        do        do
689          {          {
690          RMATCH(rrc, eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,          RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,
691            ims, eptrb, flags);            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 654  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 670  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;
747        otherwise return. */        otherwise return. */
748    
749        RMATCH(rrc, eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md, ims,        RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md, ims,
750          eptrb, flags);          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 723  for (;;) Line 788  for (;;)
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_condassert, RM3);
793        if (rrc == MATCH_MATCH)        if (rrc == MATCH_MATCH)
794          {          {
795          condition = TRUE;          condition = TRUE;
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 743  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 770  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 779  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 802  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, 0);        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, ims, NULL, 0,
883            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 828  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, 0);        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, ims, NULL, 0,
910            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 854  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 885  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 947  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 957  for (;;) Line 1036  for (;;)
1036        flags = (*callpat >= OP_SBRA)? match_cbegroup : 0;        flags = (*callpat >= OP_SBRA)? match_cbegroup : 0;
1037        do        do
1038          {          {
1039          RMATCH(rrc, eptr, callpat + _pcre_OP_lengths[*callpat], offset_top,          RMATCH(eptr, callpat + _pcre_OP_lengths[*callpat], offset_top,
1040            md, ims, eptrb, flags);            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 967  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 1001  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, 0);  
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 1047  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_tail_recursed;        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_cbegroup);        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 1070  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        {        {
1159        next = ecode+1;        next = ecode+1;
1160        RMATCH(rrc, eptr, next, offset_top, md, ims, eptrb, 0);        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;
# Line 1090  for (;;) Line 1168  for (;;)
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, 0);        RMATCH(eptr, next + 1+LINK_SIZE, offset_top, md, ims, eptrb, 0, RM11);
1172        if (rrc != MATCH_NOMATCH) RRETURN(rrc);        if (rrc != MATCH_NOMATCH) RRETURN(rrc);
1173        ecode++;        ecode++;
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 1160  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 1189  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(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);
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 */
1296        {        {
1297        RMATCH(rrc, eptr, prev, offset_top, md, ims, eptrb, flags);        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 1239  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 1436  for (;;) Line 1534  for (;;)
1534        case 0x000d:        case 0x000d:
1535        if (eptr < md->end_subject && *eptr == 0x0a) eptr++;        if (eptr < md->end_subject && *eptr == 0x0a) eptr++;
1536        break;        break;
1537    
1538        case 0x000a:        case 0x000a:
1539          break;
1540    
1541        case 0x000b:        case 0x000b:
1542        case 0x000c:        case 0x000c:
1543        case 0x0085:        case 0x0085:
1544        case 0x2028:        case 0x2028:
1545        case 0x2029:        case 0x2029:
1546          if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);
1547          break;
1548          }
1549        ecode++;
1550        break;
1551    
1552        case OP_NOT_HSPACE:
1553        if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
1554        GETCHARINCTEST(c, eptr);
1555        switch(c)
1556          {
1557          default: break;
1558          case 0x09:      /* HT */
1559          case 0x20:      /* SPACE */
1560          case 0xa0:      /* NBSP */
1561          case 0x1680:    /* OGHAM SPACE MARK */
1562          case 0x180e:    /* MONGOLIAN VOWEL SEPARATOR */
1563          case 0x2000:    /* EN QUAD */
1564          case 0x2001:    /* EM QUAD */
1565          case 0x2002:    /* EN SPACE */
1566          case 0x2003:    /* EM SPACE */
1567          case 0x2004:    /* THREE-PER-EM SPACE */
1568          case 0x2005:    /* FOUR-PER-EM SPACE */
1569          case 0x2006:    /* SIX-PER-EM SPACE */
1570          case 0x2007:    /* FIGURE SPACE */
1571          case 0x2008:    /* PUNCTUATION SPACE */
1572          case 0x2009:    /* THIN SPACE */
1573          case 0x200A:    /* HAIR SPACE */
1574          case 0x202f:    /* NARROW NO-BREAK SPACE */
1575          case 0x205f:    /* MEDIUM MATHEMATICAL SPACE */
1576          case 0x3000:    /* IDEOGRAPHIC SPACE */
1577          RRETURN(MATCH_NOMATCH);
1578          }
1579        ecode++;
1580        break;
1581    
1582        case OP_HSPACE:
1583        if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
1584        GETCHARINCTEST(c, eptr);
1585        switch(c)
1586          {
1587          default: RRETURN(MATCH_NOMATCH);
1588          case 0x09:      /* HT */
1589          case 0x20:      /* SPACE */
1590          case 0xa0:      /* NBSP */
1591          case 0x1680:    /* OGHAM SPACE MARK */
1592          case 0x180e:    /* MONGOLIAN VOWEL SEPARATOR */
1593          case 0x2000:    /* EN QUAD */
1594          case 0x2001:    /* EM QUAD */
1595          case 0x2002:    /* EN SPACE */
1596          case 0x2003:    /* EM SPACE */
1597          case 0x2004:    /* THREE-PER-EM SPACE */
1598          case 0x2005:    /* FOUR-PER-EM SPACE */
1599          case 0x2006:    /* SIX-PER-EM SPACE */
1600          case 0x2007:    /* FIGURE SPACE */
1601          case 0x2008:    /* PUNCTUATION SPACE */
1602          case 0x2009:    /* THIN SPACE */
1603          case 0x200A:    /* HAIR SPACE */
1604          case 0x202f:    /* NARROW NO-BREAK SPACE */
1605          case 0x205f:    /* MEDIUM MATHEMATICAL SPACE */
1606          case 0x3000:    /* IDEOGRAPHIC SPACE */
1607          break;
1608          }
1609        ecode++;
1610        break;
1611    
1612        case OP_NOT_VSPACE:
1613        if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
1614        GETCHARINCTEST(c, eptr);
1615        switch(c)
1616          {
1617          default: break;
1618          case 0x0a:      /* LF */
1619          case 0x0b:      /* VT */
1620          case 0x0c:      /* FF */
1621          case 0x0d:      /* CR */
1622          case 0x85:      /* NEL */
1623          case 0x2028:    /* LINE SEPARATOR */
1624          case 0x2029:    /* PARAGRAPH SEPARATOR */
1625          RRETURN(MATCH_NOMATCH);
1626          }
1627        ecode++;
1628        break;
1629    
1630        case OP_VSPACE:
1631        if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
1632        GETCHARINCTEST(c, eptr);
1633        switch(c)
1634          {
1635          default: RRETURN(MATCH_NOMATCH);
1636          case 0x0a:      /* LF */
1637          case 0x0b:      /* VT */
1638          case 0x0c:      /* FF */
1639          case 0x0d:      /* CR */
1640          case 0x85:      /* NEL */
1641          case 0x2028:    /* LINE SEPARATOR */
1642          case 0x2029:    /* PARAGRAPH SEPARATOR */
1643        break;        break;
1644        }        }
1645      ecode++;      ecode++;
# Line 1533  for (;;) Line 1731  for (;;)
1731      case OP_REF:      case OP_REF:
1732        {        {
1733        offset = GET2(ecode, 1) << 1;               /* Doubled ref number */        offset = GET2(ecode, 1) << 1;               /* Doubled ref number */
1734        ecode += 3;                                 /* Advance past item */        ecode += 3;
1735    
1736        /* If the reference is unset, set the length to be longer than the amount        /* If the reference is unset, there are two possibilities:
1737        of subject left; this ensures that every attempt at a match fails. We  
1738        can't just fail here, because of the possibility of quantifiers with zero        (a) In the default, Perl-compatible state, set the length to be longer
1739        minima. */        than the amount of subject left; this ensures that every attempt at a
1740          match fails. We can't just fail here, because of the possibility of
1741        length = (offset >= offset_top || md->offset_vector[offset] < 0)?        quantifiers with zero minima.
1742          md->end_subject - eptr + 1 :  
1743          md->offset_vector[offset+1] - md->offset_vector[offset];        (b) If the JavaScript compatibility flag is set, set the length to zero
1744          so that the back reference matches an empty string.
1745    
1746          Otherwise, set the length to the length of what was matched by the
1747          referenced subpattern. */
1748    
1749          if (offset >= offset_top || md->offset_vector[offset] < 0)
1750            length = (md->jscript_compat)? 0 : md->end_subject - eptr + 1;
1751          else
1752            length = md->offset_vector[offset+1] - md->offset_vector[offset];
1753    
1754        /* Set up for repetition, or handle the non-repeated case */        /* Set up for repetition, or handle the non-repeated case */
1755    
# Line 1602  for (;;) Line 1809  for (;;)
1809          {          {
1810          for (fi = min;; fi++)          for (fi = min;; fi++)
1811            {            {
1812            RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM14);
1813            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
1814            if (fi >= max || !match_ref(offset, eptr, length, md, ims))            if (fi >= max || !match_ref(offset, eptr, length, md, ims))
1815              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
# Line 1623  for (;;) Line 1830  for (;;)
1830            }            }
1831          while (eptr >= pp)          while (eptr >= pp)
1832            {            {
1833            RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM15);
1834            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
1835            eptr -= length;            eptr -= length;
1836            }            }
# Line 1728  for (;;) Line 1935  for (;;)
1935            {            {
1936            for (fi = min;; fi++)            for (fi = min;; fi++)
1937              {              {
1938              RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM16);
1939              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
1940              if (fi >= max || eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);              if (fi >= max || eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
1941              GETCHARINC(c, eptr);              GETCHARINC(c, eptr);
# Line 1748  for (;;) Line 1955  for (;;)
1955            {            {
1956            for (fi = min;; fi++)            for (fi = min;; fi++)
1957              {              {
1958              RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM17);
1959              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
1960              if (fi >= max || eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);              if (fi >= max || eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
1961              c = *eptr++;              c = *eptr++;
# Line 1785  for (;;) Line 1992  for (;;)
1992              }              }
1993            for (;;)            for (;;)
1994              {              {
1995              RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM18);
1996              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
1997              if (eptr-- == pp) break;        /* Stop if tried at original pos */              if (eptr-- == pp) break;        /* Stop if tried at original pos */
1998              BACKCHAR(eptr);              BACKCHAR(eptr);
# Line 1804  for (;;) Line 2011  for (;;)
2011              }              }
2012            while (eptr >= pp)            while (eptr >= pp)
2013              {              {
2014              RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM19);
2015              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2016              eptr--;              eptr--;
2017              }              }
# Line 1875  for (;;) Line 2082  for (;;)
2082          {          {
2083          for (fi = min;; fi++)          for (fi = min;; fi++)
2084            {            {
2085            RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM20);
2086            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2087            if (fi >= max || eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);            if (fi >= max || eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
2088            GETCHARINC(c, eptr);            GETCHARINC(c, eptr);
# Line 1899  for (;;) Line 2106  for (;;)
2106            }            }
2107          for(;;)          for(;;)
2108            {            {
2109            RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM21);
2110            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2111            if (eptr-- == pp) break;        /* Stop if tried at original pos */            if (eptr-- == pp) break;        /* Stop if tried at original pos */
2112            BACKCHAR(eptr)            if (utf8) BACKCHAR(eptr);
2113            }            }
2114          RRETURN(MATCH_NOMATCH);          RRETURN(MATCH_NOMATCH);
2115          }          }
# Line 2086  for (;;) Line 2293  for (;;)
2293            {            {
2294            for (fi = min;; fi++)            for (fi = min;; fi++)
2295              {              {
2296              RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM22);
2297              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2298              if (fi >= max || eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);              if (fi >= max || eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
2299              if (memcmp(eptr, charptr, length) == 0) eptr += length;              if (memcmp(eptr, charptr, length) == 0) eptr += length;
# Line 2127  for (;;) Line 2334  for (;;)
2334            if (possessive) continue;            if (possessive) continue;
2335            for(;;)            for(;;)
2336             {             {
2337             RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);             RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM23);
2338             if (rrc != MATCH_NOMATCH) RRETURN(rrc);             if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2339             if (eptr == pp) RRETURN(MATCH_NOMATCH);             if (eptr == pp) RRETURN(MATCH_NOMATCH);
2340  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
# Line 2176  for (;;) Line 2383  for (;;)
2383          {          {
2384          for (fi = min;; fi++)          for (fi = min;; fi++)
2385            {            {
2386            RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM24);
2387            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2388            if (fi >= max || eptr >= md->end_subject ||            if (fi >= max || eptr >= md->end_subject ||
2389                fc != md->lcc[*eptr++])                fc != md->lcc[*eptr++])
# Line 2195  for (;;) Line 2402  for (;;)
2402          if (possessive) continue;          if (possessive) continue;
2403          while (eptr >= pp)          while (eptr >= pp)
2404            {            {
2405            RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM25);
2406            eptr--;            eptr--;
2407            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2408            }            }
# Line 2214  for (;;) Line 2421  for (;;)
2421          {          {
2422          for (fi = min;; fi++)          for (fi = min;; fi++)
2423            {            {
2424            RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM26);
2425            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2426            if (fi >= max || eptr >= md->end_subject || fc != *eptr++)            if (fi >= max || eptr >= md->end_subject || fc != *eptr++)
2427              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
# Line 2232  for (;;) Line 2439  for (;;)
2439          if (possessive) continue;          if (possessive) continue;
2440          while (eptr >= pp)          while (eptr >= pp)
2441            {            {
2442            RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM27);
2443            eptr--;            eptr--;
2444            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2445            }            }
# Line 2377  for (;;) Line 2584  for (;;)
2584            register unsigned int d;            register unsigned int d;
2585            for (fi = min;; fi++)            for (fi = min;; fi++)
2586              {              {
2587              RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM28);
2588              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2589              GETCHARINC(d, eptr);              GETCHARINC(d, eptr);
2590              if (d < 256) d = md->lcc[d];              if (d < 256) d = md->lcc[d];
# Line 2391  for (;;) Line 2598  for (;;)
2598            {            {
2599            for (fi = min;; fi++)            for (fi = min;; fi++)
2600              {              {
2601              RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM29);
2602              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2603              if (fi >= max || eptr >= md->end_subject || fc == md->lcc[*eptr++])              if (fi >= max || eptr >= md->end_subject || fc == md->lcc[*eptr++])
2604                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
# Line 2423  for (;;) Line 2630  for (;;)
2630          if (possessive) continue;          if (possessive) continue;
2631          for(;;)          for(;;)
2632              {              {
2633              RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM30);
2634              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2635              if (eptr-- == pp) break;        /* Stop if tried at original pos */              if (eptr-- == pp) break;        /* Stop if tried at original pos */
2636              BACKCHAR(eptr);              BACKCHAR(eptr);
# Line 2441  for (;;) Line 2648  for (;;)
2648            if (possessive) continue;            if (possessive) continue;
2649            while (eptr >= pp)            while (eptr >= pp)
2650              {              {
2651              RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM31);
2652              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2653              eptr--;              eptr--;
2654              }              }
# Line 2486  for (;;) Line 2693  for (;;)
2693            register unsigned int d;            register unsigned int d;
2694            for (fi = min;; fi++)            for (fi = min;; fi++)
2695              {              {
2696              RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM32);
2697              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2698              GETCHARINC(d, eptr);              GETCHARINC(d, eptr);
2699              if (fi >= max || eptr >= md->end_subject || fc == d)              if (fi >= max || eptr >= md->end_subject || fc == d)
# Line 2499  for (;;) Line 2706  for (;;)
2706            {            {
2707            for (fi = min;; fi++)            for (fi = min;; fi++)
2708              {              {
2709              RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM33);
2710              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2711              if (fi >= max || eptr >= md->end_subject || fc == *eptr++)              if (fi >= max || eptr >= md->end_subject || fc == *eptr++)
2712                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
# Line 2530  for (;;) Line 2737  for (;;)
2737            if (possessive) continue;            if (possessive) continue;
2738            for(;;)            for(;;)
2739              {              {
2740              RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM34);
2741              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2742              if (eptr-- == pp) break;        /* Stop if tried at original pos */              if (eptr-- == pp) break;        /* Stop if tried at original pos */
2743              BACKCHAR(eptr);              BACKCHAR(eptr);
# Line 2548  for (;;) Line 2755  for (;;)
2755            if (possessive) continue;            if (possessive) continue;
2756            while (eptr >= pp)            while (eptr >= pp)
2757              {              {
2758              RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM35);
2759              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2760              eptr--;              eptr--;
2761              }              }
# Line 2655  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              }              }
2867            break;            break;
2868    
# Line 2663  for (;;) Line 2870  for (;;)
2870            for (i = 1; i <= min; i++)            for (i = 1; i <= min; i++)
2871              {              {
2872              if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);              if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
2873              GETCHARINC(c, eptr);              GETCHARINCTEST(c, eptr);
2874              prop_category = _pcre_ucp_findprop(c, &prop_chartype, &prop_script);              prop_category = _pcre_ucp_findprop(c, &prop_chartype, &prop_script);
2875              if ((prop_chartype == ucp_Lu ||              if ((prop_chartype == ucp_Lu ||
2876                   prop_chartype == ucp_Ll ||                   prop_chartype == ucp_Ll ||
# Line 2676  for (;;) Line 2883  for (;;)
2883            for (i = 1; i <= min; i++)            for (i = 1; i <= min; i++)
2884              {              {
2885              if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);              if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
2886              GETCHARINC(c, eptr);              GETCHARINCTEST(c, eptr);
2887              prop_category = _pcre_ucp_findprop(c, &prop_chartype, &prop_script);              prop_category = _pcre_ucp_findprop(c, &prop_chartype, &prop_script);
2888              if ((prop_category == prop_value) == prop_fail_result)              if ((prop_category == prop_value) == prop_fail_result)
2889                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
# Line 2687  for (;;) Line 2894  for (;;)
2894            for (i = 1; i <= min; i++)            for (i = 1; i <= min; i++)
2895              {              {
2896              if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);              if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
2897              GETCHARINC(c, eptr);              GETCHARINCTEST(c, eptr);
2898              prop_category = _pcre_ucp_findprop(c, &prop_chartype, &prop_script);              prop_category = _pcre_ucp_findprop(c, &prop_chartype, &prop_script);
2899              if ((prop_chartype == prop_value) == prop_fail_result)              if ((prop_chartype == prop_value) == prop_fail_result)
2900                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
# Line 2698  for (;;) Line 2905  for (;;)
2905            for (i = 1; i <= min; i++)            for (i = 1; i <= min; i++)
2906              {              {
2907              if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);              if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
2908              GETCHARINC(c, eptr);              GETCHARINCTEST(c, eptr);
2909              prop_category = _pcre_ucp_findprop(c, &prop_chartype, &prop_script);              prop_category = _pcre_ucp_findprop(c, &prop_chartype, &prop_script);
2910              if ((prop_script == prop_value) == prop_fail_result)              if ((prop_script == prop_value) == prop_fail_result)
2911                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
# Line 2768  for (;;) Line 2975  for (;;)
2975              case 0x000d:              case 0x000d:
2976              if (eptr < md->end_subject && *eptr == 0x0a) eptr++;              if (eptr < md->end_subject && *eptr == 0x0a) eptr++;
2977              break;              break;
2978    
2979              case 0x000a:              case 0x000a:
2980                break;
2981    
2982              case 0x000b:              case 0x000b:
2983              case 0x000c:              case 0x000c:
2984              case 0x0085:              case 0x0085:
2985              case 0x2028:              case 0x2028:
2986              case 0x2029:              case 0x2029:
2987                if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);
2988                break;
2989                }
2990              }
2991            break;
2992    
2993            case OP_NOT_HSPACE:
2994            for (i = 1; i <= min; i++)
2995              {
2996              if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
2997              GETCHARINC(c, eptr);
2998              switch(c)
2999                {
3000                default: break;
3001                case 0x09:      /* HT */
3002                case 0x20:      /* SPACE */
3003                case 0xa0:      /* NBSP */
3004                case 0x1680:    /* OGHAM SPACE MARK */
3005                case 0x180e:    /* MONGOLIAN VOWEL SEPARATOR */
3006                case 0x2000:    /* EN QUAD */
3007                case 0x2001:    /* EM QUAD */
3008                case 0x2002:    /* EN SPACE */
3009                case 0x2003:    /* EM SPACE */
3010                case 0x2004:    /* THREE-PER-EM SPACE */
3011                case 0x2005:    /* FOUR-PER-EM SPACE */
3012                case 0x2006:    /* SIX-PER-EM SPACE */
3013                case 0x2007:    /* FIGURE SPACE */
3014                case 0x2008:    /* PUNCTUATION SPACE */
3015                case 0x2009:    /* THIN SPACE */
3016                case 0x200A:    /* HAIR SPACE */
3017                case 0x202f:    /* NARROW NO-BREAK SPACE */
3018                case 0x205f:    /* MEDIUM MATHEMATICAL SPACE */
3019                case 0x3000:    /* IDEOGRAPHIC SPACE */
3020                RRETURN(MATCH_NOMATCH);
3021                }
3022              }
3023            break;
3024    
3025            case OP_HSPACE:
3026            for (i = 1; i <= min; i++)
3027              {
3028              if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
3029              GETCHARINC(c, eptr);
3030              switch(c)
3031                {
3032                default: RRETURN(MATCH_NOMATCH);
3033                case 0x09:      /* HT */
3034                case 0x20:      /* SPACE */
3035                case 0xa0:      /* NBSP */
3036                case 0x1680:    /* OGHAM SPACE MARK */
3037                case 0x180e:    /* MONGOLIAN VOWEL SEPARATOR */
3038                case 0x2000:    /* EN QUAD */
3039                case 0x2001:    /* EM QUAD */
3040                case 0x2002:    /* EN SPACE */
3041                case 0x2003:    /* EM SPACE */
3042                case 0x2004:    /* THREE-PER-EM SPACE */
3043                case 0x2005:    /* FOUR-PER-EM SPACE */
3044                case 0x2006:    /* SIX-PER-EM SPACE */
3045                case 0x2007:    /* FIGURE SPACE */
3046                case 0x2008:    /* PUNCTUATION SPACE */
3047                case 0x2009:    /* THIN SPACE */
3048                case 0x200A:    /* HAIR SPACE */
3049                case 0x202f:    /* NARROW NO-BREAK SPACE */
3050                case 0x205f:    /* MEDIUM MATHEMATICAL SPACE */
3051                case 0x3000:    /* IDEOGRAPHIC SPACE */
3052                break;
3053                }
3054              }
3055            break;
3056    
3057            case OP_NOT_VSPACE:
3058            for (i = 1; i <= min; i++)
3059              {
3060              if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
3061              GETCHARINC(c, eptr);
3062              switch(c)
3063                {
3064                default: break;
3065                case 0x0a:      /* LF */
3066                case 0x0b:      /* VT */
3067                case 0x0c:      /* FF */
3068                case 0x0d:      /* CR */
3069                case 0x85:      /* NEL */
3070                case 0x2028:    /* LINE SEPARATOR */
3071                case 0x2029:    /* PARAGRAPH SEPARATOR */
3072                RRETURN(MATCH_NOMATCH);
3073                }
3074              }
3075            break;
3076    
3077            case OP_VSPACE:
3078            for (i = 1; i <= min; i++)
3079              {
3080              if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
3081              GETCHARINC(c, eptr);
3082              switch(c)
3083                {
3084                default: RRETURN(MATCH_NOMATCH);
3085                case 0x0a:      /* LF */
3086                case 0x0b:      /* VT */
3087                case 0x0c:      /* FF */
3088                case 0x0d:      /* CR */
3089                case 0x85:      /* NEL */
3090                case 0x2028:    /* LINE SEPARATOR */
3091                case 0x2029:    /* PARAGRAPH SEPARATOR */
3092              break;              break;
3093              }              }
3094            }            }
# Line 2803  for (;;) Line 3118  for (;;)
3118          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
3119            {            {
3120            if (eptr >= md->end_subject ||            if (eptr >= md->end_subject ||
3121               (*eptr < 128 && (md->ctypes[*eptr++] & ctype_space) != 0))               (*eptr < 128 && (md->ctypes[*eptr] & ctype_space) != 0))
3122              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
3123            while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;            while (++eptr < md->end_subject && (*eptr & 0xc0) == 0x80);
3124            }            }
3125          break;          break;
3126    
# Line 2823  for (;;) Line 3138  for (;;)
3138          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
3139            {            {
3140            if (eptr >= md->end_subject ||            if (eptr >= md->end_subject ||
3141               (*eptr < 128 && (md->ctypes[*eptr++] & ctype_word) != 0))               (*eptr < 128 && (md->ctypes[*eptr] & ctype_word) != 0))
3142              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
3143            while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;            while (++eptr < md->end_subject && (*eptr & 0xc0) == 0x80);
3144            }            }
3145          break;          break;
3146    
# Line 2882  for (;;) Line 3197  for (;;)
3197              if (eptr < md->end_subject && *eptr == 0x0a) eptr++;              if (eptr < md->end_subject && *eptr == 0x0a) eptr++;
3198              break;              break;
3199              case 0x000a:              case 0x000a:
3200                break;
3201    
3202              case 0x000b:              case 0x000b:
3203              case 0x000c:              case 0x000c:
3204              case 0x0085:              case 0x0085:
3205                if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);
3206                break;
3207                }
3208              }
3209            break;
3210    
3211            case OP_NOT_HSPACE:
3212            for (i = 1; i <= min; i++)
3213              {
3214              if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
3215              switch(*eptr++)
3216                {
3217                default: break;
3218                case 0x09:      /* HT */
3219                case 0x20:      /* SPACE */
3220                case 0xa0:      /* NBSP */
3221                RRETURN(MATCH_NOMATCH);
3222                }
3223              }
3224            break;
3225    
3226            case OP_HSPACE:
3227            for (i = 1; i <= min; i++)
3228              {
3229              if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
3230              switch(*eptr++)
3231                {
3232                default: RRETURN(MATCH_NOMATCH);
3233                case 0x09:      /* HT */
3234                case 0x20:      /* SPACE */
3235                case 0xa0:      /* NBSP */
3236                break;
3237                }
3238              }
3239            break;
3240    
3241            case OP_NOT_VSPACE:
3242            for (i = 1; i <= min; i++)
3243              {
3244              if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
3245              switch(*eptr++)
3246                {
3247                default: break;
3248                case 0x0a:      /* LF */
3249                case 0x0b:      /* VT */
3250                case 0x0c:      /* FF */
3251                case 0x0d:      /* CR */
3252                case 0x85:      /* NEL */
3253                RRETURN(MATCH_NOMATCH);
3254                }
3255              }
3256            break;
3257    
3258            case OP_VSPACE:
3259            for (i = 1; i <= min; i++)
3260              {
3261              if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
3262              switch(*eptr++)
3263                {
3264                default: RRETURN(MATCH_NOMATCH);
3265                case 0x0a:      /* LF */
3266                case 0x0b:      /* VT */
3267                case 0x0c:      /* FF */
3268                case 0x0d:      /* CR */
3269                case 0x85:      /* NEL */
3270              break;              break;
3271              }              }
3272            }            }
# Line 2945  for (;;) Line 3327  for (;;)
3327            case PT_ANY:            case PT_ANY:
3328            for (fi = min;; fi++)            for (fi = min;; fi++)
3329              {              {
3330              RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM36);
3331              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3332              if (fi >= max || eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);              if (fi >= max || eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
3333              GETCHARINC(c, eptr);              GETCHARINC(c, eptr);
# Line 2956  for (;;) Line 3338  for (;;)
3338            case PT_LAMP:            case PT_LAMP:
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, RM37);
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 2971  for (;;) Line 3353  for (;;)
3353            case PT_GC:            case PT_GC:
3354            for (fi = min;; fi++)            for (fi = min;; fi++)
3355              {              {
3356              RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM38);
3357              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3358              if (fi >= max || eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);              if (fi >= max || eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
3359              GETCHARINC(c, eptr);              GETCHARINC(c, eptr);
# Line 2984  for (;;) Line 3366  for (;;)
3366            case PT_PC:            case PT_PC:
3367            for (fi = min;; fi++)            for (fi = min;; fi++)
3368              {              {
3369              RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM39);
3370              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3371              if (fi >= max || eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);              if (fi >= max || eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
3372              GETCHARINC(c, eptr);              GETCHARINC(c, eptr);
# Line 2997  for (;;) Line 3379  for (;;)
3379            case PT_SC:            case PT_SC:
3380            for (fi = min;; fi++)            for (fi = min;; fi++)
3381              {              {
3382              RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM40);
3383              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3384              if (fi >= max || eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);              if (fi >= max || eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
3385              GETCHARINC(c, eptr);              GETCHARINC(c, eptr);
# Line 3019  for (;;) Line 3401  for (;;)
3401          {          {
3402          for (fi = min;; fi++)          for (fi = min;; fi++)
3403            {            {
3404            RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM41);
3405            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3406            if (fi >= max || eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);            if (fi >= max || eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
3407            GETCHARINCTEST(c, eptr);            GETCHARINCTEST(c, eptr);
# Line 3048  for (;;) Line 3430  for (;;)
3430          {          {
3431          for (fi = min;; fi++)          for (fi = min;; fi++)
3432            {            {
3433            RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM42);
3434            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3435            if (fi >= max || eptr >= md->end_subject ||            if (fi >= max || eptr >= md->end_subject ||
3436                 (ctype == OP_ANY && (ims & PCRE_DOTALL) == 0 &&                 (ctype == OP_ANY && (ims & PCRE_DOTALL) == 0 &&
# Line 3072  for (;;) Line 3454  for (;;)
3454                if (eptr < md->end_subject && *eptr == 0x0a) eptr++;                if (eptr < md->end_subject && *eptr == 0x0a) eptr++;
3455                break;                break;
3456                case 0x000a:                case 0x000a:
3457                  break;
3458    
3459                case 0x000b:                case 0x000b:
3460                case 0x000c:                case 0x000c:
3461                case 0x0085:                case 0x0085:
3462                case 0x2028:                case 0x2028:
3463                case 0x2029:                case 0x2029:
3464                  if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);
3465                  break;
3466                  }
3467                break;
3468    
3469                case OP_NOT_HSPACE:
3470                switch(c)
3471                  {
3472                  default: break;
3473                  case 0x09:      /* HT */
3474                  case 0x20:      /* SPACE */
3475                  case 0xa0:      /* NBSP */
3476                  case 0x1680:    /* OGHAM SPACE MARK */
3477                  case 0x180e:    /* MONGOLIAN VOWEL SEPARATOR */
3478                  case 0x2000:    /* EN QUAD */
3479                  case 0x2001:    /* EM QUAD */
3480                  case 0x2002:    /* EN SPACE */
3481                  case 0x2003:    /* EM SPACE */
3482                  case 0x2004:    /* THREE-PER-EM SPACE */
3483                  case 0x2005:    /* FOUR-PER-EM SPACE */
3484                  case 0x2006:    /* SIX-PER-EM SPACE */
3485                  case 0x2007:    /* FIGURE SPACE */
3486                  case 0x2008:    /* PUNCTUATION SPACE */
3487                  case 0x2009:    /* THIN SPACE */
3488                  case 0x200A:    /* HAIR SPACE */
3489                  case 0x202f:    /* NARROW NO-BREAK SPACE */
3490                  case 0x205f:    /* MEDIUM MATHEMATICAL SPACE */
3491                  case 0x3000:    /* IDEOGRAPHIC SPACE */
3492                  RRETURN(MATCH_NOMATCH);
3493                  }
3494                break;
3495    
3496                case OP_HSPACE:
3497                switch(c)
3498                  {
3499                  default: RRETURN(MATCH_NOMATCH);
3500                  case 0x09:      /* HT */
3501                  case 0x20:      /* SPACE */
3502                  case 0xa0:      /* NBSP */
3503                  case 0x1680:    /* OGHAM SPACE MARK */
3504                  case 0x180e:    /* MONGOLIAN VOWEL SEPARATOR */
3505                  case 0x2000:    /* EN QUAD */
3506                  case 0x2001:    /* EM QUAD */
3507                  case 0x2002:    /* EN SPACE */
3508                  case 0x2003:    /* EM SPACE */
3509                  case 0x2004:    /* THREE-PER-EM SPACE */
3510                  case 0x2005:    /* FOUR-PER-EM SPACE */
3511                  case 0x2006:    /* SIX-PER-EM SPACE */
3512                  case 0x2007:    /* FIGURE SPACE */
3513                  case 0x2008:    /* PUNCTUATION SPACE */
3514                  case 0x2009:    /* THIN SPACE */
3515                  case 0x200A:    /* HAIR SPACE */
3516                  case 0x202f:    /* NARROW NO-BREAK SPACE */
3517                  case 0x205f:    /* MEDIUM MATHEMATICAL SPACE */
3518                  case 0x3000:    /* IDEOGRAPHIC SPACE */
3519                  break;
3520                  }
3521                break;
3522    
3523                case OP_NOT_VSPACE:
3524                switch(c)
3525                  {
3526                  default: break;
3527                  case 0x0a:      /* LF */
3528                  case 0x0b:      /* VT */
3529                  case 0x0c:      /* FF */
3530                  case 0x0d:      /* CR */
3531                  case 0x85:      /* NEL */
3532                  case 0x2028:    /* LINE SEPARATOR */
3533                  case 0x2029:    /* PARAGRAPH SEPARATOR */
3534                  RRETURN(MATCH_NOMATCH);
3535                  }
3536                break;
3537    
3538                case OP_VSPACE:
3539                switch(c)
3540                  {
3541                  default: RRETURN(MATCH_NOMATCH);
3542                  case 0x0a:      /* LF */
3543                  case 0x0b:      /* VT */
3544                  case 0x0c:      /* FF */
3545                  case 0x0d:      /* CR */
3546                  case 0x85:      /* NEL */
3547                  case 0x2028:    /* LINE SEPARATOR */
3548                  case 0x2029:    /* PARAGRAPH SEPARATOR */
3549                break;                break;
3550                }                }
3551              break;              break;
# Line 3122  for (;;) Line 3591  for (;;)
3591          {          {
3592          for (fi = min;; fi++)          for (fi = min;; fi++)
3593            {            {
3594            RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM43);
3595            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3596            if (fi >= max || eptr >= md->end_subject ||            if (fi >= max || eptr >= md->end_subject ||
3597                 ((ims & PCRE_DOTALL) == 0 && IS_NEWLINE(eptr)))                 ((ims & PCRE_DOTALL) == 0 && IS_NEWLINE(eptr)))
# Line 3144  for (;;) Line 3613  for (;;)
3613                case 0x000d:                case 0x000d:
3614                if (eptr < md->end_subject && *eptr == 0x0a) eptr++;                if (eptr < md->end_subject && *eptr == 0x0a) eptr++;
3615                break;                break;
3616    
3617                case 0x000a:                case 0x000a:
3618                  break;
3619    
3620                case 0x000b:                case 0x000b:
3621                case 0x000c:                case 0x000c:
3622                case 0x0085:                case 0x0085:
3623                  if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);
3624                  break;
3625                  }
3626                break;
3627    
3628                case OP_NOT_HSPACE:
3629                switch(c)
3630                  {
3631                  default: break;
3632                  case 0x09:      /* HT */
3633                  case 0x20:      /* SPACE */
3634                  case 0xa0:      /* NBSP */
3635                  RRETURN(MATCH_NOMATCH);
3636                  }
3637                break;
3638    
3639                case OP_HSPACE:
3640                switch(c)
3641                  {
3642                  default: RRETURN(MATCH_NOMATCH);
3643                  case 0x09:      /* HT */
3644                  case 0x20:      /* SPACE */
3645                  case 0xa0:      /* NBSP */
3646                  break;
3647                  }
3648                break;
3649    
3650                case OP_NOT_VSPACE:
3651                switch(c)
3652                  {
3653                  default: break;
3654                  case 0x0a:      /* LF */
3655                  case 0x0b:      /* VT */
3656                  case 0x0c:      /* FF */
3657                  case 0x0d:      /* CR */
3658                  case 0x85:      /* NEL */
3659                  RRETURN(MATCH_NOMATCH);
3660                  }
3661                break;
3662    
3663                case OP_VSPACE:
3664                switch(c)
3665                  {
3666                  default: RRETURN(MATCH_NOMATCH);
3667                  case 0x0a:      /* LF */
3668                  case 0x0b:      /* VT */
3669                  case 0x0c:      /* FF */
3670                  case 0x0d:      /* CR */
3671                  case 0x85:      /* NEL */
3672                break;                break;
3673                }                }
3674              break;              break;
# Line 3268  for (;;) Line 3789  for (;;)
3789          if (possessive) continue;          if (possessive) continue;
3790          for(;;)          for(;;)
3791            {            {
3792            RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM44);
3793            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3794            if (eptr-- == pp) break;        /* Stop if tried at original pos */            if (eptr-- == pp) break;        /* Stop if tried at original pos */
3795            BACKCHAR(eptr);            if (utf8) BACKCHAR(eptr);
3796            }            }
3797          }          }
3798    
# Line 3304  for (;;) Line 3825  for (;;)
3825          if (possessive) continue;          if (possessive) continue;
3826          for(;;)          for(;;)
3827            {            {
3828            RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM45);
3829            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3830            if (eptr-- == pp) break;        /* Stop if tried at original pos */            if (eptr-- == pp) break;        /* Stop if tried at original pos */
3831            for (;;)                        /* Move back over one extended */            for (;;)                        /* Move back over one extended */
3832              {              {
3833              int len = 1;              int len = 1;
             BACKCHAR(eptr);  
3834              if (!utf8) c = *eptr; else              if (!utf8) c = *eptr; else
3835                {                {
3836                  BACKCHAR(eptr);
3837                GETCHARLEN(c, eptr, len);                GETCHARLEN(c, eptr, len);
3838                }                }
3839              prop_category = _pcre_ucp_findprop(c, &prop_chartype, &prop_script);              prop_category = _pcre_ucp_findprop(c, &prop_chartype, &prop_script);
# Line 3333  for (;;) Line 3854  for (;;)
3854          switch(ctype)          switch(ctype)
3855            {            {
3856            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. */  
   
3857            if (max < INT_MAX)            if (max < INT_MAX)
3858              {              {
3859              if ((ims & PCRE_DOTALL) == 0)              if ((ims & PCRE_DOTALL) == 0)
# Line 3370  for (;;) Line 3886  for (;;)
3886                  {                  {
3887                  if (eptr >= md->end_subject || IS_NEWLINE(eptr)) break;                  if (eptr >= md->end_subject || IS_NEWLINE(eptr)) break;
3888                  eptr++;                  eptr++;
3889                    while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;
3890                  }                  }
               break;  
3891                }                }
3892              else              else
3893                {                {
3894                c = max - min;                eptr = md->end_subject;
               if (c > (unsigned int)(md->end_subject - eptr))  
                 c = md->end_subject - eptr;  
               eptr += c;  
3895                }                }
3896              }              }
3897            break;            break;
# Line 3405  for (;;) Line 3918  for (;;)
3918                }                }
3919              else              else
3920                {                {
3921                if (c != 0x000a && c != 0x000b && c != 0x000c &&                if (c != 0x000a &&
3922                    c != 0x0085 && c != 0x2028 && c != 0x2029)                    (md->bsr_anycrlf ||
3923                       (c != 0x000b && c != 0x000c &&
3924                        c != 0x0085 && c != 0x2028 && c != 0x2029)))
3925                  break;                  break;
3926                eptr += len;                eptr += len;
3927                }                }
3928              }              }
3929            break;            break;
3930    
3931              case OP_NOT_HSPACE:
3932              case OP_HSPACE:
3933              for (i = min; i < max; i++)
3934                {
3935                BOOL gotspace;
3936                int len = 1;
3937                if (eptr >= md->end_subject) break;
3938                GETCHARLEN(c, eptr, len);
3939                switch(c)
3940                  {
3941                  default: gotspace = FALSE; break;
3942                  case 0x09:      /* HT */
3943                  case 0x20:      /* SPACE */
3944                  case 0xa0:      /* NBSP */
3945                  case 0x1680:    /* OGHAM SPACE MARK */
3946                  case 0x180e:    /* MONGOLIAN VOWEL SEPARATOR */
3947                  case 0x2000:    /* EN QUAD */
3948                  case 0x2001:    /* EM QUAD */
3949                  case 0x2002:    /* EN SPACE */
3950                  case 0x2003:    /* EM SPACE */
3951                  case 0x2004:    /* THREE-PER-EM SPACE */
3952                  case 0x2005:    /* FOUR-PER-EM SPACE */
3953                  case 0x2006:    /* SIX-PER-EM SPACE */
3954                  case 0x2007:    /* FIGURE SPACE */
3955                  case 0x2008:    /* PUNCTUATION SPACE */
3956                  case 0x2009:    /* THIN SPACE */
3957                  case 0x200A:    /* HAIR SPACE */
3958                  case 0x202f:    /* NARROW NO-BREAK SPACE */
3959                  case 0x205f:    /* MEDIUM MATHEMATICAL SPACE */
3960                  case 0x3000:    /* IDEOGRAPHIC SPACE */
3961                  gotspace = TRUE;
3962                  break;
3963                  }
3964                if (gotspace == (ctype == OP_NOT_HSPACE)) break;
3965                eptr += len;
3966                }
3967              break;
3968    
3969              case OP_NOT_VSPACE:
3970              case OP_VSPACE:
3971              for (i = min; i < max; i++)
3972                {
3973                BOOL gotspace;
3974                int len = 1;
3975                if (eptr >= md->end_subject) break;
3976                GETCHARLEN(c, eptr, len);
3977                switch(c)
3978                  {
3979                  default: gotspace = FALSE; break;
3980                  case 0x0a:      /* LF */
3981                  case 0x0b:      /* VT */
3982                  case 0x0c:      /* FF */
3983                  case 0x0d:      /* CR */
3984                  case 0x85:      /* NEL */
3985                  case 0x2028:    /* LINE SEPARATOR */
3986                  case 0x2029:    /* PARAGRAPH SEPARATOR */
3987                  gotspace = TRUE;
3988                  break;
3989                  }
3990                if (gotspace == (ctype == OP_NOT_VSPACE)) break;
3991                eptr += len;
3992                }
3993              break;
3994    
3995            case OP_NOT_DIGIT:            case OP_NOT_DIGIT:
3996            for (i = min; i < max; i++)            for (i = min; i < max; i++)
3997              {              {
# Line 3488  for (;;) Line 4067  for (;;)
4067          if (possessive) continue;          if (possessive) continue;
4068          for(;;)          for(;;)
4069            {            {
4070            RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM46);
4071            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
4072            if (eptr-- == pp) break;        /* Stop if tried at original pos */            if (eptr-- == pp) break;        /* Stop if tried at original pos */
4073            BACKCHAR(eptr);            BACKCHAR(eptr);
4074            }            }
4075          }          }
4076        else        else
4077  #endif  #endif  /* SUPPORT_UTF8 */
4078    
4079        /* Not UTF-8 mode */        /* Not UTF-8 mode */
4080          {          {
# Line 3532  for (;;) Line 4111  for (;;)
4111                }                }
4112              else              else
4113                {                {
4114                if (c != 0x000a && c != 0x000b && c != 0x000c && c != 0x0085)                if (c != 0x000a &&
4115                      (md->bsr_anycrlf ||
4116                        (c != 0x000b && c != 0x000c && c != 0x0085)))
4117                  break;                  break;
4118                eptr++;                eptr++;
4119                }                }
4120              }              }
4121            break;            break;
4122    
4123              case OP_NOT_HSPACE:
4124              for (i = min; i < max; i++)
4125                {
4126                if (eptr >= md->end_subject) break;
4127                c = *eptr;
4128                if (c == 0x09 || c == 0x20 || c == 0xa0) break;
4129                eptr++;
4130                }
4131              break;
4132    
4133              case OP_HSPACE:
4134              for (i = min; i < max; i++)
4135                {
4136                if (eptr >= md->end_subject) break;
4137                c = *eptr;
4138                if (c != 0x09 && c != 0x20 && c != 0xa0) break;
4139                eptr++;
4140                }
4141              break;
4142    
4143              case OP_NOT_VSPACE:
4144              for (i = min; i < max; i++)
4145                {
4146                if (eptr >= md->end_subject) break;
4147                c = *eptr;
4148                if (c == 0x0a || c == 0x0b || c == 0x0c || c == 0x0d || c == 0x85)
4149                  break;
4150                eptr++;
4151                }
4152              break;
4153    
4154              case OP_VSPACE:
4155              for (i = min; i < max; i++)
4156                {
4157                if (eptr >= md->end_subject) break;
4158                c = *eptr;
4159                if (c != 0x0a && c != 0x0b && c != 0x0c && c != 0x0d && c != 0x85)
4160                  break;
4161                eptr++;
4162                }
4163              break;
4164    
4165            case OP_NOT_DIGIT:            case OP_NOT_DIGIT:
4166            for (i = min; i < max; i++)            for (i = min; i < max; i++)
4167              {              {
# Line 3602  for (;;) Line 4225  for (;;)
4225          if (possessive) continue;          if (possessive) continue;
4226          while (eptr >= pp)          while (eptr >= pp)
4227            {            {
4228            RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM47);
4229            eptr--;            eptr--;
4230            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
4231            }            }
# Line 3628  for (;;) Line 4251  for (;;)
4251    
4252    }             /* End of main loop */    }             /* End of main loop */
4253  /* Control never reaches here */  /* Control never reaches here */
4254    
4255    
4256    /* When compiling to use the heap rather than the stack for recursive calls to
4257    match(), the RRETURN() macro jumps here. The number that is saved in
4258    frame->Xwhere indicates which label we actually want to return to. */
4259    
4260    #ifdef NO_RECURSE
4261    #define LBL(val) case val: goto L_RM##val;
4262    HEAP_RETURN:
4263    switch (frame->Xwhere)
4264      {
4265      LBL( 1) LBL( 2) LBL( 3) LBL( 4) LBL( 5) LBL( 6) LBL( 7) LBL( 8)
4266      LBL( 9) LBL(10) LBL(11) LBL(12) LBL(13) LBL(14) LBL(15) LBL(17)
4267      LBL(19) LBL(24) LBL(25) LBL(26) LBL(27) LBL(29) LBL(31) LBL(33)
4268      LBL(35) LBL(43) LBL(47) LBL(48) LBL(49) LBL(50) LBL(51) LBL(52)
4269      LBL(53) LBL(54)
4270    #ifdef SUPPORT_UTF8
4271      LBL(16) LBL(18) LBL(20) LBL(21) LBL(22) LBL(23) LBL(28) LBL(30)
4272      LBL(32) LBL(34) LBL(42) LBL(46)
4273    #ifdef SUPPORT_UCP
4274      LBL(36) LBL(37) LBL(38) LBL(39) LBL(40) LBL(41) LBL(44) LBL(45)
4275    #endif  /* SUPPORT_UCP */
4276    #endif  /* SUPPORT_UTF8 */
4277      default:
4278      DPRINTF(("jump error in pcre match: label %d non-existent\n", frame->Xwhere));
4279      return PCRE_ERROR_INTERNAL;
4280      }
4281    #undef LBL
4282    #endif  /* NO_RECURSE */
4283  }  }
4284    
4285    
# Line 3640  Undefine all the macros that were define Line 4292  Undefine all the macros that were define
4292  #ifdef NO_RECURSE  #ifdef NO_RECURSE
4293  #undef eptr  #undef eptr
4294  #undef ecode  #undef ecode
4295    #undef mstart
4296  #undef offset_top  #undef offset_top
4297  #undef ims  #undef ims
4298  #undef eptrb  #undef eptrb
# Line 3737  const uschar *start_bits = NULL; Line 4390  const uschar *start_bits = NULL;
4390  USPTR start_match = (USPTR)subject + start_offset;  USPTR start_match = (USPTR)subject + start_offset;
4391  USPTR end_subject;  USPTR end_subject;
4392  USPTR req_byte_ptr = start_match - 1;  USPTR req_byte_ptr = start_match - 1;
 eptrblock eptrchain[EPTR_WORK_SIZE];  
4393    
4394  pcre_study_data internal_study;  pcre_study_data internal_study;
4395  const pcre_study_data *study;  const pcre_study_data *study;
# Line 3800  if (re->magic_number != MAGIC_NUMBER) Line 4452  if (re->magic_number != MAGIC_NUMBER)
4452  /* Set up other data */  /* Set up other data */
4453    
4454  anchored = ((re->options | options) & PCRE_ANCHORED) != 0;  anchored = ((re->options | options) & PCRE_ANCHORED) != 0;
4455  startline = (re->options & PCRE_STARTLINE) != 0;  startline = (re->flags & PCRE_STARTLINE) != 0;
4456  firstline = (re->options & PCRE_FIRSTLINE) != 0;  firstline = (re->options & PCRE_FIRSTLINE) != 0;
4457    
4458  /* 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 3815  end_subject = md->end_subject; Line 4467  end_subject = md->end_subject;
4467    
4468  md->endonly = (re->options & PCRE_DOLLAR_ENDONLY) != 0;  md->endonly = (re->options & PCRE_DOLLAR_ENDONLY) != 0;
4469  utf8 = md->utf8 = (re->options & PCRE_UTF8) != 0;  utf8 = md->utf8 = (re->options & PCRE_UTF8) != 0;
4470    md->jscript_compat = (re->options & PCRE_JAVASCRIPT_COMPAT) != 0;
4471    
4472  md->notbol = (options & PCRE_NOTBOL) != 0;  md->notbol = (options & PCRE_NOTBOL) != 0;
4473  md->noteol = (options & PCRE_NOTEOL) != 0;  md->noteol = (options & PCRE_NOTEOL) != 0;
# Line 3823  md->partial = (options & PCRE_PARTIAL) ! Line 4476  md->partial = (options & PCRE_PARTIAL) !
4476  md->hitend = FALSE;  md->hitend = FALSE;
4477    
4478  md->recursive = NULL;                   /* No recursion at top level */  md->recursive = NULL;                   /* No recursion at top level */
 md->eptrchain = eptrchain;              /* Make workspace generally available */  
4479    
4480  md->lcc = tables + lcc_offset;  md->lcc = tables + lcc_offset;
4481  md->ctypes = tables + ctypes_offset;  md->ctypes = tables + ctypes_offset;
4482    
4483    /* Handle different \R options. */
4484    
4485    switch (options & (PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE))
4486      {
4487      case 0:
4488      if ((re->options & (PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE)) != 0)
4489        md->bsr_anycrlf = (re->options & PCRE_BSR_ANYCRLF) != 0;
4490      else
4491    #ifdef BSR_ANYCRLF
4492      md->bsr_anycrlf = TRUE;
4493    #else
4494      md->bsr_anycrlf = FALSE;
4495    #endif
4496      break;
4497    
4498      case PCRE_BSR_ANYCRLF:
4499      md->bsr_anycrlf = TRUE;
4500      break;
4501    
4502      case PCRE_BSR_UNICODE:
4503      md->bsr_anycrlf = FALSE;
4504      break;
4505    
4506      default: return PCRE_ERROR_BADNEWLINE;
4507      }
4508    
4509  /* Handle different types of newline. The three bits give eight cases. If  /* Handle different types of newline. The three bits give eight cases. If
4510  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. */
4511    
4512  switch ((((options & PCRE_NEWLINE_BITS) == 0)? re->options : (pcre_uint32)options) &  switch ((((options & PCRE_NEWLINE_BITS) == 0)? re->options :
4513         PCRE_NEWLINE_BITS)          (pcre_uint32)options) & PCRE_NEWLINE_BITS)
4514    {    {
4515    case 0: newline = NEWLINE; break;   /* Compile-time default */    case 0: newline = NEWLINE; break;   /* Compile-time default */
4516    case PCRE_NEWLINE_CR: newline = '\r'; break;    case PCRE_NEWLINE_CR: newline = '\r'; break;
# Line 3840  switch ((((options & PCRE_NEWLINE_BITS) Line 4518  switch ((((options & PCRE_NEWLINE_BITS)
4518    case PCRE_NEWLINE_CR+    case PCRE_NEWLINE_CR+
4519         PCRE_NEWLINE_LF: newline = ('\r' << 8) | '\n'; break;         PCRE_NEWLINE_LF: newline = ('\r' << 8) | '\n'; break;
4520    case PCRE_NEWLINE_ANY: newline = -1; break;    case PCRE_NEWLINE_ANY: newline = -1; break;
4521    case PCRE_NEWLINE_ANYCRLF: newline = -2; break;    case PCRE_NEWLINE_ANYCRLF: newline = -2; break;
4522    default: return PCRE_ERROR_BADNEWLINE;    default: return PCRE_ERROR_BADNEWLINE;
4523    }    }
4524    
# Line 3871  else Line 4549  else
4549  /* 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
4550  moment. */  moment. */
4551    
4552  if (md->partial && (re->options & PCRE_NOPARTIAL) != 0)  if (md->partial && (re->flags & PCRE_NOPARTIAL) != 0)
4553    return PCRE_ERROR_BADPARTIAL;    return PCRE_ERROR_BADPARTIAL;
4554    
4555  /* 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 3948  studied, there may be a bitmap of possib Line 4626  studied, there may be a bitmap of possib
4626    
4627  if (!anchored)  if (!anchored)
4628    {    {
4629    if ((re->options & PCRE_FIRSTSET) != 0)    if ((re->flags & PCRE_FIRSTSET) != 0)
4630      {      {
4631      first_byte = re->first_byte & 255;      first_byte = re->first_byte & 255;
4632      if ((first_byte_caseless = ((re->first_byte & REQ_CASELESS) != 0)) == TRUE)      if ((first_byte_caseless = ((re->first_byte & REQ_CASELESS) != 0)) == TRUE)
# Line 3963  if (!anchored) Line 4641  if (!anchored)
4641  /* For anchored or unanchored matches, there may be a "last known required  /* For anchored or unanchored matches, there may be a "last known required
4642  character" set. */  character" set. */
4643    
4644  if ((re->options & PCRE_REQCHSET) != 0)  if ((re->flags & PCRE_REQCHSET) != 0)
4645    {    {
4646    req_byte = re->req_byte & 255;    req_byte = re->req_byte & 255;
4647    req_byte_caseless = (re->req_byte & REQ_CASELESS) != 0;    req_byte_caseless = (re->req_byte & REQ_CASELESS) != 0;
# Line 3979  the loop runs just once. */ Line 4657  the loop runs just once. */
4657  for(;;)  for(;;)
4658    {    {
4659    USPTR save_end_subject = end_subject;    USPTR save_end_subject = end_subject;
4660      USPTR new_start_match;
4661    
4662    /* Reset the maximum number of extractions we might see. */    /* Reset the maximum number of extractions we might see. */
4663    
# Line 4009  for(;;) Line 4688  for(;;)
4688      if (first_byte_caseless)      if (first_byte_caseless)
4689        while (start_match < end_subject &&        while (start_match < end_subject &&
4690               md->lcc[*start_match] != first_byte)               md->lcc[*start_match] != first_byte)
4691          start_match++;          { NEXTCHAR(start_match); }
4692      else      else
4693        while (start_match < end_subject && *start_match != first_byte)        while (start_match < end_subject && *start_match != first_byte)
4694          start_match++;          { NEXTCHAR(start_match); }
4695      }      }
4696    
4697    /* 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 4022  for(;;) Line 4701  for(;;)
4701      if (start_match > md->start_subject + start_offset)      if (start_match > md->start_subject + start_offset)
4702        {        {
4703        while (start_match <= end_subject && !WAS_NEWLINE(start_match))        while (start_match <= end_subject && !WAS_NEWLINE(start_match))
4704          start_match++;          { NEXTCHAR(start_match); }
4705    
4706        /* 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,
4707        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 4043  for(;;) Line 4722  for(;;)
4722      while (start_match < end_subject)      while (start_match < end_subject)
4723        {        {
4724        register unsigned int c = *start_match;        register unsigned int c = *start_match;
4725        if ((start_bits[c/8] & (1 << (c&7))) == 0) start_match++; else break;        if ((start_bits[c/8] & (1 << (c&7))) == 0)
4726            { NEXTCHAR(start_match); }
4727          else break;
4728        }        }
4729      }      }
4730    
# Line 4119  for(;;) Line 4800  for(;;)
4800    
4801    /* OK, we can now run the match. */    /* OK, we can now run the match. */
4802    
4803    md->start_match = start_match;    md->start_match_ptr = start_match;
4804    md->match_call_count = 0;    md->match_call_count = 0;
4805    md->eptrn = 0;                          /* Next free eptrchain slot */    rc = match(start_match, md->start_code, start_match, 2, md, ims, NULL, 0, 0);
4806    rc = match(start_match, md->start_code, 2, md, ims, NULL, 0, 0);  
4807      switch(rc)
4808        {
4809        /* NOMATCH and PRUNE advance by one character. THEN at this level acts
4810        exactly like PRUNE. */
4811    
4812    /* Any return other than MATCH_NOMATCH breaks the loop. */      case MATCH_NOMATCH:
4813        case MATCH_PRUNE:
4814        case MATCH_THEN:
4815        new_start_match = start_match + 1;
4816    #ifdef SUPPORT_UTF8
4817        if (utf8)
4818          while(new_start_match < end_subject && (*new_start_match & 0xc0) == 0x80)
4819            new_start_match++;
4820    #endif
4821        break;
4822    
4823        /* SKIP passes back the next starting point explicitly. */
4824    
4825    if (rc != MATCH_NOMATCH) break;      case MATCH_SKIP:
4826        new_start_match = md->start_match_ptr;
4827        break;
4828    
4829        /* COMMIT disables the bumpalong, but otherwise behaves as NOMATCH. */
4830    
4831        case MATCH_COMMIT:
4832        rc = MATCH_NOMATCH;
4833        goto ENDLOOP;
4834    
4835        /* Any other return is some kind of error. */
4836    
4837        default:
4838        goto ENDLOOP;
4839        }
4840    
4841      /* Control reaches here for the various types of "no match at this point"
4842      result. Reset the code to MATCH_NOMATCH for subsequent checking. */
4843    
4844      rc = MATCH_NOMATCH;
4845    
4846    /* 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
4847    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 4134  for(;;) Line 4849  for(;;)
4849    
4850    if (firstline && IS_NEWLINE(start_match)) break;    if (firstline && IS_NEWLINE(start_match)) break;
4851    
4852    /* Advance the match position by one character. */    /* Advance to new matching position */
4853    
4854    start_match++;    start_match = new_start_match;
 #ifdef SUPPORT_UTF8  
   if (utf8)  
     while(start_match < end_subject && (*start_match & 0xc0) == 0x80)  
       start_match++;  
 #endif  
4855    
4856    /* 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
4857    the subject. */    the subject. */
4858    
4859    if (anchored || start_match > end_subject) break;    if (anchored || start_match > end_subject) break;
4860    
4861    /* 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
4862    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
4863    character. */    or ANY or ANYCRLF, advance the match position by one more character. */
4864    
4865    if (start_match[-1] == '\r' &&    if (start_match[-1] == '\r' &&
4866         (md->nltype == NLTYPE_ANY ||        start_match < end_subject &&
4867          md->nltype == NLTYPE_ANYCRLF ||        *start_match == '\n' &&
4868          md->nllen == 2) &&        (re->flags & PCRE_HASCRORLF) == 0 &&
4869         start_match < end_subject &&          (md->nltype == NLTYPE_ANY ||
4870         *start_match == '\n')           md->nltype == NLTYPE_ANYCRLF ||
4871             md->nllen == 2))
4872      start_match++;      start_match++;
4873    
4874    }   /* End of for(;;) "bumpalong" loop */    }   /* End of for(;;) "bumpalong" loop */
# Line 4167  for(;;) Line 4878  for(;;)
4878  /* 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
4879  conditions is true:  conditions is true:
4880    
4881  (1) The pattern is anchored;  (1) The pattern is anchored or the match was failed by (*COMMIT);
4882    
4883  (2) We are past the end of the subject;  (2) We are past the end of the subject;
4884    
# Line 4182  processing, copy those that we can. In t Line 4893  processing, copy those that we can. In t
4893  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
4894  capturing parentheses than vector slots. */  capturing parentheses than vector slots. */
4895    
4896    ENDLOOP:
4897    
4898  if (rc == MATCH_MATCH)  if (rc == MATCH_MATCH)
4899    {    {
4900    if (using_temporary_offsets)    if (using_temporary_offsets)
# Line 4202  if (rc == MATCH_MATCH) Line 4915  if (rc == MATCH_MATCH)
4915    
4916    rc = md->offset_overflow? 0 : md->end_offset_top/2;    rc = md->offset_overflow? 0 : md->end_offset_top/2;
4917    
4918    /* 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
4919      md->start_match_ptr might be modified if \K was encountered on the success
4920      matching path. */
4921    
4922    if (offsetcount < 2) rc = 0; else    if (offsetcount < 2) rc = 0; else
4923      {      {
4924      offsets[0] = start_match - md->start_subject;      offsets[0] = md->start_match_ptr - md->start_subject;
4925      offsets[1] = md->end_match_ptr - md->start_subject;      offsets[1] = md->end_match_ptr - md->start_subject;
4926      }      }
4927    

Legend:
Removed from v.149  
changed lines
  Added in v.336

  ViewVC Help
Powered by ViewVC 1.1.5