/[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 123 by ph10, Mon Mar 12 15:19:06 2007 UTC revision 349 by ph10, Wed Jul 2 18:42:11 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 */
52    
53  #include "pcre_internal.h"  #include "pcre_internal.h"
54    
55  /* The chain of eptrblocks for tail recursions uses memory in stack workspace,  /* Undefine some potentially clashing cpp symbols */
 obtained at top level, the size of which is defined by EPTR_WORK_SIZE. */  
56    
57  #define EPTR_WORK_SIZE (1000)  #undef min
58    #undef max
59    
60  /* Flag bits for the match() function */  /* Flag bits for the match() function */
61    
62  #define match_condassert     0x01  /* Called to check a condition assertion */  #define match_condassert     0x01  /* Called to check a condition assertion */
63  #define match_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 65  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 183  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 205  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 251  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 268  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 318  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 349  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 358  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 368  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 396  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 410  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 540  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 548  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 588  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 632  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 649  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 665  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 718  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 738  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 765  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 774  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 797  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 823  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 849  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 880  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 942  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 952  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 962  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 996  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 1042  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 1065  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 1085  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 1155  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 1184  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 1234  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 1326  for (;;) Line 1429  for (;;)
1429      /* Match a single character type; inline for speed */      /* Match a single character type; inline for speed */
1430    
1431      case OP_ANY:      case OP_ANY:
1432      if ((ims & PCRE_DOTALL) == 0)      if (IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH);
1433        {      /* Fall through */
1434        if (IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH);  
1435        }      case OP_ALLANY:
1436      if (eptr++ >= md->end_subject) RRETURN(MATCH_NOMATCH);      if (eptr++ >= md->end_subject) RRETURN(MATCH_NOMATCH);
1437      if (utf8)      if (utf8) while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;
       while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;  
1438      ecode++;      ecode++;
1439      break;      break;
1440    
# Line 1431  for (;;) Line 1533  for (;;)
1533        case 0x000d:        case 0x000d:
1534        if (eptr < md->end_subject && *eptr == 0x0a) eptr++;        if (eptr < md->end_subject && *eptr == 0x0a) eptr++;
1535        break;        break;
1536    
1537        case 0x000a:        case 0x000a:
1538          break;
1539    
1540        case 0x000b:        case 0x000b:
1541        case 0x000c:        case 0x000c:
1542        case 0x0085:        case 0x0085:
1543        case 0x2028:        case 0x2028:
1544        case 0x2029:        case 0x2029:
1545          if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);
1546          break;
1547          }
1548        ecode++;
1549        break;
1550    
1551        case OP_NOT_HSPACE:
1552        if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
1553        GETCHARINCTEST(c, eptr);
1554        switch(c)
1555          {
1556          default: break;
1557          case 0x09:      /* HT */
1558          case 0x20:      /* SPACE */
1559          case 0xa0:      /* NBSP */
1560          case 0x1680:    /* OGHAM SPACE MARK */
1561          case 0x180e:    /* MONGOLIAN VOWEL SEPARATOR */
1562          case 0x2000:    /* EN QUAD */
1563          case 0x2001:    /* EM QUAD */
1564          case 0x2002:    /* EN SPACE */
1565          case 0x2003:    /* EM SPACE */
1566          case 0x2004:    /* THREE-PER-EM SPACE */
1567          case 0x2005:    /* FOUR-PER-EM SPACE */
1568          case 0x2006:    /* SIX-PER-EM SPACE */
1569          case 0x2007:    /* FIGURE SPACE */
1570          case 0x2008:    /* PUNCTUATION SPACE */
1571          case 0x2009:    /* THIN SPACE */
1572          case 0x200A:    /* HAIR SPACE */
1573          case 0x202f:    /* NARROW NO-BREAK SPACE */
1574          case 0x205f:    /* MEDIUM MATHEMATICAL SPACE */
1575          case 0x3000:    /* IDEOGRAPHIC SPACE */
1576          RRETURN(MATCH_NOMATCH);
1577          }
1578        ecode++;
1579        break;
1580    
1581        case OP_HSPACE:
1582        if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
1583        GETCHARINCTEST(c, eptr);
1584        switch(c)
1585          {
1586          default: RRETURN(MATCH_NOMATCH);
1587          case 0x09:      /* HT */
1588          case 0x20:      /* SPACE */
1589          case 0xa0:      /* NBSP */
1590          case 0x1680:    /* OGHAM SPACE MARK */
1591          case 0x180e:    /* MONGOLIAN VOWEL SEPARATOR */
1592          case 0x2000:    /* EN QUAD */
1593          case 0x2001:    /* EM QUAD */
1594          case 0x2002:    /* EN SPACE */
1595          case 0x2003:    /* EM SPACE */
1596          case 0x2004:    /* THREE-PER-EM SPACE */
1597          case 0x2005:    /* FOUR-PER-EM SPACE */
1598          case 0x2006:    /* SIX-PER-EM SPACE */
1599          case 0x2007:    /* FIGURE SPACE */
1600          case 0x2008:    /* PUNCTUATION SPACE */
1601          case 0x2009:    /* THIN SPACE */
1602          case 0x200A:    /* HAIR SPACE */
1603          case 0x202f:    /* NARROW NO-BREAK SPACE */
1604          case 0x205f:    /* MEDIUM MATHEMATICAL SPACE */
1605          case 0x3000:    /* IDEOGRAPHIC SPACE */
1606          break;
1607          }
1608        ecode++;
1609        break;
1610    
1611        case OP_NOT_VSPACE:
1612        if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
1613        GETCHARINCTEST(c, eptr);
1614        switch(c)
1615          {
1616          default: break;
1617          case 0x0a:      /* LF */
1618          case 0x0b:      /* VT */
1619          case 0x0c:      /* FF */
1620          case 0x0d:      /* CR */
1621          case 0x85:      /* NEL */
1622          case 0x2028:    /* LINE SEPARATOR */
1623          case 0x2029:    /* PARAGRAPH SEPARATOR */
1624          RRETURN(MATCH_NOMATCH);
1625          }
1626        ecode++;
1627        break;
1628    
1629        case OP_VSPACE:
1630        if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
1631        GETCHARINCTEST(c, eptr);
1632        switch(c)
1633          {
1634          default: RRETURN(MATCH_NOMATCH);
1635          case 0x0a:      /* LF */
1636          case 0x0b:      /* VT */
1637          case 0x0c:      /* FF */
1638          case 0x0d:      /* CR */
1639          case 0x85:      /* NEL */
1640          case 0x2028:    /* LINE SEPARATOR */
1641          case 0x2029:    /* PARAGRAPH SEPARATOR */
1642        break;        break;
1643        }        }
1644      ecode++;      ecode++;
# Line 1451  for (;;) Line 1653  for (;;)
1653      if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);      if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
1654      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
1655        {        {
1656        int chartype, script;        const ucd_record * prop = GET_UCD(c);
       int category = _pcre_ucp_findprop(c, &chartype, &script);  
1657    
1658        switch(ecode[1])        switch(ecode[1])
1659          {          {
# Line 1461  for (;;) Line 1662  for (;;)
1662          break;          break;
1663    
1664          case PT_LAMP:          case PT_LAMP:
1665          if ((chartype == ucp_Lu ||          if ((prop->chartype == ucp_Lu ||
1666               chartype == ucp_Ll ||               prop->chartype == ucp_Ll ||
1667               chartype == ucp_Lt) == (op == OP_NOTPROP))               prop->chartype == ucp_Lt) == (op == OP_NOTPROP))
1668            RRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
1669           break;           break;
1670    
1671          case PT_GC:          case PT_GC:
1672          if ((ecode[2] != category) == (op == OP_PROP))          if ((ecode[2] != ucp_gentype[prop->chartype]) == (op == OP_PROP))
1673            RRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
1674          break;          break;
1675    
1676          case PT_PC:          case PT_PC:
1677          if ((ecode[2] != chartype) == (op == OP_PROP))          if ((ecode[2] != prop->chartype) == (op == OP_PROP))
1678            RRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
1679          break;          break;
1680    
1681          case PT_SC:          case PT_SC:
1682          if ((ecode[2] != script) == (op == OP_PROP))          if ((ecode[2] != prop->script) == (op == OP_PROP))
1683            RRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
1684          break;          break;
1685    
# Line 1497  for (;;) Line 1698  for (;;)
1698      if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);      if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
1699      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
1700        {        {
1701        int chartype, script;        int category = UCD_CATEGORY(c);
       int category = _pcre_ucp_findprop(c, &chartype, &script);  
1702        if (category == ucp_M) RRETURN(MATCH_NOMATCH);        if (category == ucp_M) RRETURN(MATCH_NOMATCH);
1703        while (eptr < md->end_subject)        while (eptr < md->end_subject)
1704          {          {
# Line 1507  for (;;) Line 1707  for (;;)
1707            {            {
1708            GETCHARLEN(c, eptr, len);            GETCHARLEN(c, eptr, len);
1709            }            }
1710          category = _pcre_ucp_findprop(c, &chartype, &script);          category = UCD_CATEGORY(c);
1711          if (category != ucp_M) break;          if (category != ucp_M) break;
1712          eptr += len;          eptr += len;
1713          }          }
# Line 1528  for (;;) Line 1728  for (;;)
1728      case OP_REF:      case OP_REF:
1729        {        {
1730        offset = GET2(ecode, 1) << 1;               /* Doubled ref number */        offset = GET2(ecode, 1) << 1;               /* Doubled ref number */
1731        ecode += 3;                                 /* Advance past item */        ecode += 3;
1732    
1733        /* If the reference is unset, set the length to be longer than the amount        /* If the reference is unset, there are two possibilities:
1734        of subject left; this ensures that every attempt at a match fails. We  
1735        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
1736        minima. */        than the amount of subject left; this ensures that every attempt at a
1737          match fails. We can't just fail here, because of the possibility of
1738        length = (offset >= offset_top || md->offset_vector[offset] < 0)?        quantifiers with zero minima.
1739          md->end_subject - eptr + 1 :  
1740          md->offset_vector[offset+1] - md->offset_vector[offset];        (b) If the JavaScript compatibility flag is set, set the length to zero
1741          so that the back reference matches an empty string.
1742    
1743          Otherwise, set the length to the length of what was matched by the
1744          referenced subpattern. */
1745    
1746          if (offset >= offset_top || md->offset_vector[offset] < 0)
1747            length = (md->jscript_compat)? 0 : md->end_subject - eptr + 1;
1748          else
1749            length = md->offset_vector[offset+1] - md->offset_vector[offset];
1750    
1751        /* Set up for repetition, or handle the non-repeated case */        /* Set up for repetition, or handle the non-repeated case */
1752    
# Line 1597  for (;;) Line 1806  for (;;)
1806          {          {
1807          for (fi = min;; fi++)          for (fi = min;; fi++)
1808            {            {
1809            RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM14);
1810            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
1811            if (fi >= max || !match_ref(offset, eptr, length, md, ims))            if (fi >= max || !match_ref(offset, eptr, length, md, ims))
1812              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
# Line 1618  for (;;) Line 1827  for (;;)
1827            }            }
1828          while (eptr >= pp)          while (eptr >= pp)
1829            {            {
1830            RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM15);
1831            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
1832            eptr -= length;            eptr -= length;
1833            }            }
# Line 1723  for (;;) Line 1932  for (;;)
1932            {            {
1933            for (fi = min;; fi++)            for (fi = min;; fi++)
1934              {              {
1935              RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM16);
1936              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
1937              if (fi >= max || eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);              if (fi >= max || eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
1938              GETCHARINC(c, eptr);              GETCHARINC(c, eptr);
# Line 1743  for (;;) Line 1952  for (;;)
1952            {            {
1953            for (fi = min;; fi++)            for (fi = min;; fi++)
1954              {              {
1955              RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM17);
1956              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
1957              if (fi >= max || eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);              if (fi >= max || eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
1958              c = *eptr++;              c = *eptr++;
# Line 1780  for (;;) Line 1989  for (;;)
1989              }              }
1990            for (;;)            for (;;)
1991              {              {
1992              RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM18);
1993              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
1994              if (eptr-- == pp) break;        /* Stop if tried at original pos */              if (eptr-- == pp) break;        /* Stop if tried at original pos */
1995              BACKCHAR(eptr);              BACKCHAR(eptr);
# Line 1799  for (;;) Line 2008  for (;;)
2008              }              }
2009            while (eptr >= pp)            while (eptr >= pp)
2010              {              {
2011              RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM19);
2012              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2013              eptr--;              eptr--;
2014              }              }
# Line 1870  for (;;) Line 2079  for (;;)
2079          {          {
2080          for (fi = min;; fi++)          for (fi = min;; fi++)
2081            {            {
2082            RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM20);
2083            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2084            if (fi >= max || eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);            if (fi >= max || eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
2085            GETCHARINC(c, eptr);            GETCHARINC(c, eptr);
# Line 1894  for (;;) Line 2103  for (;;)
2103            }            }
2104          for(;;)          for(;;)
2105            {            {
2106            RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM21);
2107            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2108            if (eptr-- == pp) break;        /* Stop if tried at original pos */            if (eptr-- == pp) break;        /* Stop if tried at original pos */
2109            BACKCHAR(eptr)            if (utf8) BACKCHAR(eptr);
2110            }            }
2111          RRETURN(MATCH_NOMATCH);          RRETURN(MATCH_NOMATCH);
2112          }          }
# Line 1963  for (;;) Line 2172  for (;;)
2172          if (fc != dc)          if (fc != dc)
2173            {            {
2174  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
2175            if (dc != _pcre_ucp_othercase(fc))            if (dc != UCD_OTHERCASE(fc))
2176  #endif  #endif
2177              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
2178            }            }
# Line 2054  for (;;) Line 2263  for (;;)
2263  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
2264          unsigned int othercase;          unsigned int othercase;
2265          if ((ims & PCRE_CASELESS) != 0 &&          if ((ims & PCRE_CASELESS) != 0 &&
2266              (othercase = _pcre_ucp_othercase(fc)) != NOTACHAR)              (othercase = UCD_OTHERCASE(fc)) != fc)
2267            oclength = _pcre_ord2utf8(othercase, occhars);            oclength = _pcre_ord2utf8(othercase, occhars);
2268          else oclength = 0;          else oclength = 0;
2269  #endif  /* SUPPORT_UCP */  #endif  /* SUPPORT_UCP */
# Line 2081  for (;;) Line 2290  for (;;)
2290            {            {
2291            for (fi = min;; fi++)            for (fi = min;; fi++)
2292              {              {
2293              RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM22);
2294              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2295              if (fi >= max || eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);              if (fi >= max || eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
2296              if (memcmp(eptr, charptr, length) == 0) eptr += length;              if (memcmp(eptr, charptr, length) == 0) eptr += length;
# Line 2122  for (;;) Line 2331  for (;;)
2331            if (possessive) continue;            if (possessive) continue;
2332            for(;;)            for(;;)
2333             {             {
2334             RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);             RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM23);
2335             if (rrc != MATCH_NOMATCH) RRETURN(rrc);             if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2336             if (eptr == pp) RRETURN(MATCH_NOMATCH);             if (eptr == pp) RRETURN(MATCH_NOMATCH);
2337  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
# Line 2171  for (;;) Line 2380  for (;;)
2380          {          {
2381          for (fi = min;; fi++)          for (fi = min;; fi++)
2382            {            {
2383            RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM24);
2384            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2385            if (fi >= max || eptr >= md->end_subject ||            if (fi >= max || eptr >= md->end_subject ||
2386                fc != md->lcc[*eptr++])                fc != md->lcc[*eptr++])
# Line 2190  for (;;) Line 2399  for (;;)
2399          if (possessive) continue;          if (possessive) continue;
2400          while (eptr >= pp)          while (eptr >= pp)
2401            {            {
2402            RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM25);
2403            eptr--;            eptr--;
2404            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2405            }            }
# Line 2209  for (;;) Line 2418  for (;;)
2418          {          {
2419          for (fi = min;; fi++)          for (fi = min;; fi++)
2420            {            {
2421            RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM26);
2422            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2423            if (fi >= max || eptr >= md->end_subject || fc != *eptr++)            if (fi >= max || eptr >= md->end_subject || fc != *eptr++)
2424              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
# Line 2227  for (;;) Line 2436  for (;;)
2436          if (possessive) continue;          if (possessive) continue;
2437          while (eptr >= pp)          while (eptr >= pp)
2438            {            {
2439            RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM27);
2440            eptr--;            eptr--;
2441            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2442            }            }
# Line 2372  for (;;) Line 2581  for (;;)
2581            register unsigned int d;            register unsigned int d;
2582            for (fi = min;; fi++)            for (fi = min;; fi++)
2583              {              {
2584              RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM28);
2585              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2586              GETCHARINC(d, eptr);              GETCHARINC(d, eptr);
2587              if (d < 256) d = md->lcc[d];              if (d < 256) d = md->lcc[d];
# Line 2386  for (;;) Line 2595  for (;;)
2595            {            {
2596            for (fi = min;; fi++)            for (fi = min;; fi++)
2597              {              {
2598              RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM29);
2599              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2600              if (fi >= max || eptr >= md->end_subject || fc == md->lcc[*eptr++])              if (fi >= max || eptr >= md->end_subject || fc == md->lcc[*eptr++])
2601                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
# Line 2418  for (;;) Line 2627  for (;;)
2627          if (possessive) continue;          if (possessive) continue;
2628          for(;;)          for(;;)
2629              {              {
2630              RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM30);
2631              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2632              if (eptr-- == pp) break;        /* Stop if tried at original pos */              if (eptr-- == pp) break;        /* Stop if tried at original pos */
2633              BACKCHAR(eptr);              BACKCHAR(eptr);
# Line 2436  for (;;) Line 2645  for (;;)
2645            if (possessive) continue;            if (possessive) continue;
2646            while (eptr >= pp)            while (eptr >= pp)
2647              {              {
2648              RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM31);
2649              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2650              eptr--;              eptr--;
2651              }              }
# Line 2481  for (;;) Line 2690  for (;;)
2690            register unsigned int d;            register unsigned int d;
2691            for (fi = min;; fi++)            for (fi = min;; fi++)
2692              {              {
2693              RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM32);
2694              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2695              GETCHARINC(d, eptr);              GETCHARINC(d, eptr);
2696              if (fi >= max || eptr >= md->end_subject || fc == d)              if (fi >= max || eptr >= md->end_subject || fc == d)
# Line 2494  for (;;) Line 2703  for (;;)
2703            {            {
2704            for (fi = min;; fi++)            for (fi = min;; fi++)
2705              {              {
2706              RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM33);
2707              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2708              if (fi >= max || eptr >= md->end_subject || fc == *eptr++)              if (fi >= max || eptr >= md->end_subject || fc == *eptr++)
2709                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
# Line 2525  for (;;) Line 2734  for (;;)
2734            if (possessive) continue;            if (possessive) continue;
2735            for(;;)            for(;;)
2736              {              {
2737              RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM34);
2738              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2739              if (eptr-- == pp) break;        /* Stop if tried at original pos */              if (eptr-- == pp) break;        /* Stop if tried at original pos */
2740              BACKCHAR(eptr);              BACKCHAR(eptr);
# Line 2543  for (;;) Line 2752  for (;;)
2752            if (possessive) continue;            if (possessive) continue;
2753            while (eptr >= pp)            while (eptr >= pp)
2754              {              {
2755              RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM35);
2756              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2757              eptr--;              eptr--;
2758              }              }
# Line 2650  for (;;) Line 2859  for (;;)
2859            for (i = 1; i <= min; i++)            for (i = 1; i <= min; i++)
2860              {              {
2861              if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);              if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
2862              GETCHARINC(c, eptr);              GETCHARINCTEST(c, eptr);
2863              }              }
2864            break;            break;
2865    
# Line 2658  for (;;) Line 2867  for (;;)
2867            for (i = 1; i <= min; i++)            for (i = 1; i <= min; i++)
2868              {              {
2869              if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);              if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
2870              GETCHARINC(c, eptr);              GETCHARINCTEST(c, eptr);
2871              prop_category = _pcre_ucp_findprop(c, &prop_chartype, &prop_script);              prop_chartype = UCD_CHARTYPE(c);
2872              if ((prop_chartype == ucp_Lu ||              if ((prop_chartype == ucp_Lu ||
2873                   prop_chartype == ucp_Ll ||                   prop_chartype == ucp_Ll ||
2874                   prop_chartype == ucp_Lt) == prop_fail_result)                   prop_chartype == ucp_Lt) == prop_fail_result)
# Line 2671  for (;;) Line 2880  for (;;)
2880            for (i = 1; i <= min; i++)            for (i = 1; i <= min; i++)
2881              {              {
2882              if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);              if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
2883              GETCHARINC(c, eptr);              GETCHARINCTEST(c, eptr);
2884              prop_category = _pcre_ucp_findprop(c, &prop_chartype, &prop_script);              prop_category = UCD_CATEGORY(c);
2885              if ((prop_category == prop_value) == prop_fail_result)              if ((prop_category == prop_value) == prop_fail_result)
2886                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
2887              }              }
# Line 2682  for (;;) Line 2891  for (;;)
2891            for (i = 1; i <= min; i++)            for (i = 1; i <= min; i++)
2892              {              {
2893              if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);              if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
2894              GETCHARINC(c, eptr);              GETCHARINCTEST(c, eptr);
2895              prop_category = _pcre_ucp_findprop(c, &prop_chartype, &prop_script);              prop_chartype = UCD_CHARTYPE(c);
2896              if ((prop_chartype == prop_value) == prop_fail_result)              if ((prop_chartype == prop_value) == prop_fail_result)
2897                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
2898              }              }
# Line 2693  for (;;) Line 2902  for (;;)
2902            for (i = 1; i <= min; i++)            for (i = 1; i <= min; i++)
2903              {              {
2904              if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);              if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
2905              GETCHARINC(c, eptr);              GETCHARINCTEST(c, eptr);
2906              prop_category = _pcre_ucp_findprop(c, &prop_chartype, &prop_script);              prop_script = UCD_SCRIPT(c);
2907              if ((prop_script == prop_value) == prop_fail_result)              if ((prop_script == prop_value) == prop_fail_result)
2908                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
2909              }              }
# Line 2713  for (;;) Line 2922  for (;;)
2922          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
2923            {            {
2924            GETCHARINCTEST(c, eptr);            GETCHARINCTEST(c, eptr);
2925            prop_category = _pcre_ucp_findprop(c, &prop_chartype, &prop_script);            prop_category = UCD_CATEGORY(c);
2926            if (prop_category == ucp_M) RRETURN(MATCH_NOMATCH);            if (prop_category == ucp_M) RRETURN(MATCH_NOMATCH);
2927            while (eptr < md->end_subject)            while (eptr < md->end_subject)
2928              {              {
# Line 2722  for (;;) Line 2931  for (;;)
2931                {                {
2932                GETCHARLEN(c, eptr, len);                GETCHARLEN(c, eptr, len);
2933                }                }
2934              prop_category = _pcre_ucp_findprop(c, &prop_chartype, &prop_script);              prop_category = UCD_CATEGORY(c);
2935              if (prop_category != ucp_M) break;              if (prop_category != ucp_M) break;
2936              eptr += len;              eptr += len;
2937              }              }
# Line 2740  for (;;) Line 2949  for (;;)
2949          case OP_ANY:          case OP_ANY:
2950          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
2951            {            {
2952            if (eptr >= md->end_subject ||            if (eptr >= md->end_subject || IS_NEWLINE(eptr))
                ((ims & PCRE_DOTALL) == 0 && IS_NEWLINE(eptr)))  
2953              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
2954            eptr++;            eptr++;
2955            while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;            while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;
2956            }            }
2957          break;          break;
2958    
2959            case OP_ALLANY:
2960            for (i = 1; i <= min; i++)
2961              {
2962              if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
2963              eptr++;
2964              while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;
2965              }
2966            break;
2967    
2968          case OP_ANYBYTE:          case OP_ANYBYTE:
2969          eptr += min;          eptr += min;
2970          break;          break;
# Line 2763  for (;;) Line 2980  for (;;)
2980              case 0x000d:              case 0x000d:
2981              if (eptr < md->end_subject && *eptr == 0x0a) eptr++;              if (eptr < md->end_subject && *eptr == 0x0a) eptr++;
2982              break;              break;
2983    
2984              case 0x000a:              case 0x000a:
2985                break;
2986    
2987              case 0x000b:              case 0x000b:
2988              case 0x000c:              case 0x000c:
2989              case 0x0085:              case 0x0085:
2990              case 0x2028:              case 0x2028:
2991              case 0x2029:              case 0x2029:
2992                if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);
2993                break;
2994                }
2995              }
2996            break;
2997    
2998            case OP_NOT_HSPACE:
2999            for (i = 1; i <= min; i++)
3000              {
3001              if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
3002              GETCHARINC(c, eptr);
3003              switch(c)
3004                {
3005                default: break;
3006                case 0x09:      /* HT */
3007                case 0x20:      /* SPACE */
3008                case 0xa0:      /* NBSP */
3009                case 0x1680:    /* OGHAM SPACE MARK */
3010                case 0x180e:    /* MONGOLIAN VOWEL SEPARATOR */
3011                case 0x2000:    /* EN QUAD */
3012                case 0x2001:    /* EM QUAD */
3013                case 0x2002:    /* EN SPACE */
3014                case 0x2003:    /* EM SPACE */
3015                case 0x2004:    /* THREE-PER-EM SPACE */
3016                case 0x2005:    /* FOUR-PER-EM SPACE */
3017                case 0x2006:    /* SIX-PER-EM SPACE */
3018                case 0x2007:    /* FIGURE SPACE */
3019                case 0x2008:    /* PUNCTUATION SPACE */
3020                case 0x2009:    /* THIN SPACE */
3021                case 0x200A:    /* HAIR SPACE */
3022                case 0x202f:    /* NARROW NO-BREAK SPACE */
3023                case 0x205f:    /* MEDIUM MATHEMATICAL SPACE */
3024                case 0x3000:    /* IDEOGRAPHIC SPACE */
3025                RRETURN(MATCH_NOMATCH);
3026                }
3027              }
3028            break;
3029    
3030            case OP_HSPACE:
3031            for (i = 1; i <= min; i++)
3032              {
3033              if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
3034              GETCHARINC(c, eptr);
3035              switch(c)
3036                {
3037                default: RRETURN(MATCH_NOMATCH);
3038                case 0x09:      /* HT */
3039                case 0x20:      /* SPACE */
3040                case 0xa0:      /* NBSP */
3041                case 0x1680:    /* OGHAM SPACE MARK */
3042                case 0x180e:    /* MONGOLIAN VOWEL SEPARATOR */
3043                case 0x2000:    /* EN QUAD */
3044                case 0x2001:    /* EM QUAD */
3045                case 0x2002:    /* EN SPACE */
3046                case 0x2003:    /* EM SPACE */
3047                case 0x2004:    /* THREE-PER-EM SPACE */
3048                case 0x2005:    /* FOUR-PER-EM SPACE */
3049                case 0x2006:    /* SIX-PER-EM SPACE */
3050                case 0x2007:    /* FIGURE SPACE */
3051                case 0x2008:    /* PUNCTUATION SPACE */
3052                case 0x2009:    /* THIN SPACE */
3053                case 0x200A:    /* HAIR SPACE */
3054                case 0x202f:    /* NARROW NO-BREAK SPACE */
3055                case 0x205f:    /* MEDIUM MATHEMATICAL SPACE */
3056                case 0x3000:    /* IDEOGRAPHIC SPACE */
3057                break;
3058                }
3059              }
3060            break;
3061    
3062            case OP_NOT_VSPACE:
3063            for (i = 1; i <= min; i++)
3064              {
3065              if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
3066              GETCHARINC(c, eptr);
3067              switch(c)
3068                {
3069                default: break;
3070                case 0x0a:      /* LF */
3071                case 0x0b:      /* VT */
3072                case 0x0c:      /* FF */
3073                case 0x0d:      /* CR */
3074                case 0x85:      /* NEL */
3075                case 0x2028:    /* LINE SEPARATOR */
3076                case 0x2029:    /* PARAGRAPH SEPARATOR */
3077                RRETURN(MATCH_NOMATCH);
3078                }
3079              }
3080            break;
3081    
3082            case OP_VSPACE:
3083            for (i = 1; i <= min; i++)
3084              {
3085              if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
3086              GETCHARINC(c, eptr);
3087              switch(c)
3088                {
3089                default: RRETURN(MATCH_NOMATCH);
3090                case 0x0a:      /* LF */
3091                case 0x0b:      /* VT */
3092                case 0x0c:      /* FF */
3093                case 0x0d:      /* CR */
3094                case 0x85:      /* NEL */
3095                case 0x2028:    /* LINE SEPARATOR */
3096                case 0x2029:    /* PARAGRAPH SEPARATOR */
3097              break;              break;
3098              }              }
3099            }            }
# Line 2798  for (;;) Line 3123  for (;;)
3123          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
3124            {            {
3125            if (eptr >= md->end_subject ||            if (eptr >= md->end_subject ||
3126               (*eptr < 128 && (md->ctypes[*eptr++] & ctype_space) != 0))               (*eptr < 128 && (md->ctypes[*eptr] & ctype_space) != 0))
3127              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
3128            while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;            while (++eptr < md->end_subject && (*eptr & 0xc0) == 0x80);
3129            }            }
3130          break;          break;
3131    
# Line 2818  for (;;) Line 3143  for (;;)
3143          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
3144            {            {
3145            if (eptr >= md->end_subject ||            if (eptr >= md->end_subject ||
3146               (*eptr < 128 && (md->ctypes[*eptr++] & ctype_word) != 0))               (*eptr < 128 && (md->ctypes[*eptr] & ctype_word) != 0))
3147              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
3148            while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;            while (++eptr < md->end_subject && (*eptr & 0xc0) == 0x80);
3149            }            }
3150          break;          break;
3151    
# Line 2848  for (;;) Line 3173  for (;;)
3173        switch(ctype)        switch(ctype)
3174          {          {
3175          case OP_ANY:          case OP_ANY:
3176          if ((ims & PCRE_DOTALL) == 0)          for (i = 1; i <= min; i++)
3177            {            {
3178            for (i = 1; i <= min; i++)            if (IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH);
3179              {            eptr++;
             if (IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH);  
             eptr++;  
             }  
3180            }            }
3181          else eptr += min;          break;
3182    
3183            case OP_ALLANY:
3184            eptr += min;
3185          break;          break;
3186    
3187          case OP_ANYBYTE:          case OP_ANYBYTE:
# Line 2877  for (;;) Line 3202  for (;;)
3202              if (eptr < md->end_subject && *eptr == 0x0a) eptr++;              if (eptr < md->end_subject && *eptr == 0x0a) eptr++;
3203              break;              break;
3204              case 0x000a:              case 0x000a:
3205                break;
3206    
3207              case 0x000b:              case 0x000b:
3208              case 0x000c:              case 0x000c:
3209              case 0x0085:              case 0x0085:
3210                if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);
3211                break;
3212                }
3213              }
3214            break;
3215    
3216            case OP_NOT_HSPACE:
3217            for (i = 1; i <= min; i++)
3218              {
3219              if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
3220              switch(*eptr++)
3221                {
3222                default: break;
3223                case 0x09:      /* HT */
3224                case 0x20:      /* SPACE */
3225                case 0xa0:      /* NBSP */
3226                RRETURN(MATCH_NOMATCH);
3227                }
3228              }
3229            break;
3230    
3231            case OP_HSPACE:
3232            for (i = 1; i <= min; i++)
3233              {
3234              if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
3235              switch(*eptr++)
3236                {
3237                default: RRETURN(MATCH_NOMATCH);
3238                case 0x09:      /* HT */
3239                case 0x20:      /* SPACE */
3240                case 0xa0:      /* NBSP */
3241                break;
3242                }
3243              }
3244            break;
3245    
3246            case OP_NOT_VSPACE:
3247            for (i = 1; i <= min; i++)
3248              {
3249              if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
3250              switch(*eptr++)
3251                {
3252                default: break;
3253                case 0x0a:      /* LF */
3254                case 0x0b:      /* VT */
3255                case 0x0c:      /* FF */
3256                case 0x0d:      /* CR */
3257                case 0x85:      /* NEL */
3258                RRETURN(MATCH_NOMATCH);
3259                }
3260              }
3261            break;
3262    
3263            case OP_VSPACE:
3264            for (i = 1; i <= min; i++)
3265              {
3266              if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
3267              switch(*eptr++)
3268                {
3269                default: RRETURN(MATCH_NOMATCH);
3270                case 0x0a:      /* LF */
3271                case 0x0b:      /* VT */
3272                case 0x0c:      /* FF */
3273                case 0x0d:      /* CR */
3274                case 0x85:      /* NEL */
3275              break;              break;
3276              }              }
3277            }            }
# Line 2940  for (;;) Line 3332  for (;;)
3332            case PT_ANY:            case PT_ANY:
3333            for (fi = min;; fi++)            for (fi = min;; fi++)
3334              {              {
3335              RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM36);
3336              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3337              if (fi >= max || eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);              if (fi >= max || eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
3338              GETCHARINC(c, eptr);              GETCHARINC(c, eptr);
# Line 2951  for (;;) Line 3343  for (;;)
3343            case PT_LAMP:            case PT_LAMP:
3344            for (fi = min;; fi++)            for (fi = min;; fi++)
3345              {              {
3346              RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM37);
3347              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3348              if (fi >= max || eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);              if (fi >= max || eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
3349              GETCHARINC(c, eptr);              GETCHARINC(c, eptr);
3350              prop_category = _pcre_ucp_findprop(c, &prop_chartype, &prop_script);              prop_chartype = UCD_CHARTYPE(c);
3351              if ((prop_chartype == ucp_Lu ||              if ((prop_chartype == ucp_Lu ||
3352                   prop_chartype == ucp_Ll ||                   prop_chartype == ucp_Ll ||
3353                   prop_chartype == ucp_Lt) == prop_fail_result)                   prop_chartype == ucp_Lt) == prop_fail_result)
# Line 2966  for (;;) Line 3358  for (;;)
3358            case PT_GC:            case PT_GC:
3359            for (fi = min;; fi++)            for (fi = min;; fi++)
3360              {              {
3361              RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM38);
3362              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3363              if (fi >= max || eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);              if (fi >= max || eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
3364              GETCHARINC(c, eptr);              GETCHARINC(c, eptr);
3365              prop_category = _pcre_ucp_findprop(c, &prop_chartype, &prop_script);              prop_category = UCD_CATEGORY(c);
3366              if ((prop_category == prop_value) == prop_fail_result)              if ((prop_category == prop_value) == prop_fail_result)
3367                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
3368              }              }
# Line 2979  for (;;) Line 3371  for (;;)
3371            case PT_PC:            case PT_PC:
3372            for (fi = min;; fi++)            for (fi = min;; fi++)
3373              {              {
3374              RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM39);
3375              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3376              if (fi >= max || eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);              if (fi >= max || eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
3377              GETCHARINC(c, eptr);              GETCHARINC(c, eptr);
3378              prop_category = _pcre_ucp_findprop(c, &prop_chartype, &prop_script);              prop_chartype = UCD_CHARTYPE(c);
3379              if ((prop_chartype == prop_value) == prop_fail_result)              if ((prop_chartype == prop_value) == prop_fail_result)
3380                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
3381              }              }
# Line 2992  for (;;) Line 3384  for (;;)
3384            case PT_SC:            case PT_SC:
3385            for (fi = min;; fi++)            for (fi = min;; fi++)
3386              {              {
3387              RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM40);
3388              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3389              if (fi >= max || eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);              if (fi >= max || eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
3390              GETCHARINC(c, eptr);              GETCHARINC(c, eptr);
3391              prop_category = _pcre_ucp_findprop(c, &prop_chartype, &prop_script);              prop_script = UCD_SCRIPT(c);
3392              if ((prop_script == prop_value) == prop_fail_result)              if ((prop_script == prop_value) == prop_fail_result)
3393                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
3394              }              }
# Line 3014  for (;;) Line 3406  for (;;)
3406          {          {
3407          for (fi = min;; fi++)          for (fi = min;; fi++)
3408            {            {
3409            RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM41);
3410            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3411            if (fi >= max || eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);            if (fi >= max || eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);
3412            GETCHARINCTEST(c, eptr);            GETCHARINCTEST(c, eptr);
3413            prop_category = _pcre_ucp_findprop(c, &prop_chartype, &prop_script);            prop_category = UCD_CATEGORY(c);
3414            if (prop_category == ucp_M) RRETURN(MATCH_NOMATCH);            if (prop_category == ucp_M) RRETURN(MATCH_NOMATCH);
3415            while (eptr < md->end_subject)            while (eptr < md->end_subject)
3416              {              {
# Line 3027  for (;;) Line 3419  for (;;)
3419                {                {
3420                GETCHARLEN(c, eptr, len);                GETCHARLEN(c, eptr, len);
3421                }                }
3422              prop_category = _pcre_ucp_findprop(c, &prop_chartype, &prop_script);              prop_category = UCD_CATEGORY(c);
3423              if (prop_category != ucp_M) break;              if (prop_category != ucp_M) break;
3424              eptr += len;              eptr += len;
3425              }              }
# Line 3043  for (;;) Line 3435  for (;;)
3435          {          {
3436          for (fi = min;; fi++)          for (fi = min;; fi++)
3437            {            {
3438            RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM42);
3439            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3440            if (fi >= max || eptr >= md->end_subject ||            if (fi >= max || eptr >= md->end_subject ||
3441                 (ctype == OP_ANY && (ims & PCRE_DOTALL) == 0 &&                 (ctype == OP_ANY && IS_NEWLINE(eptr)))
                 IS_NEWLINE(eptr)))  
3442              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
3443    
3444            GETCHARINC(c, eptr);            GETCHARINC(c, eptr);
3445            switch(ctype)            switch(ctype)
3446              {              {
3447              case OP_ANY:        /* This is the DOTALL case */              case OP_ANY:        /* This is the non-NL case */
3448              break;              case OP_ALLANY:
   
3449              case OP_ANYBYTE:              case OP_ANYBYTE:
3450              break;              break;
3451    
# Line 3067  for (;;) Line 3457  for (;;)
3457                if (eptr < md->end_subject && *eptr == 0x0a) eptr++;                if (eptr < md->end_subject && *eptr == 0x0a) eptr++;
3458                break;                break;
3459                case 0x000a:                case 0x000a:
3460                  break;
3461    
3462                case 0x000b:                case 0x000b:
3463                case 0x000c:                case 0x000c:
3464                case 0x0085:                case 0x0085:
3465                case 0x2028:                case 0x2028:
3466                case 0x2029:                case 0x2029:
3467                  if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);
3468                  break;
3469                  }
3470                break;
3471    
3472                case OP_NOT_HSPACE:
3473                switch(c)
3474                  {
3475                  default: break;
3476                  case 0x09:      /* HT */
3477                  case 0x20:      /* SPACE */
3478                  case 0xa0:      /* NBSP */
3479                  case 0x1680:    /* OGHAM SPACE MARK */
3480                  case 0x180e:    /* MONGOLIAN VOWEL SEPARATOR */
3481                  case 0x2000:    /* EN QUAD */
3482                  case 0x2001:    /* EM QUAD */
3483                  case 0x2002:    /* EN SPACE */
3484                  case 0x2003:    /* EM SPACE */
3485                  case 0x2004:    /* THREE-PER-EM SPACE */
3486                  case 0x2005:    /* FOUR-PER-EM SPACE */
3487                  case 0x2006:    /* SIX-PER-EM SPACE */
3488                  case 0x2007:    /* FIGURE SPACE */
3489                  case 0x2008:    /* PUNCTUATION SPACE */
3490                  case 0x2009:    /* THIN SPACE */
3491                  case 0x200A:    /* HAIR SPACE */
3492                  case 0x202f:    /* NARROW NO-BREAK SPACE */
3493                  case 0x205f:    /* MEDIUM MATHEMATICAL SPACE */
3494                  case 0x3000:    /* IDEOGRAPHIC SPACE */
3495                  RRETURN(MATCH_NOMATCH);
3496                  }
3497                break;
3498    
3499                case OP_HSPACE:
3500                switch(c)
3501                  {
3502                  default: RRETURN(MATCH_NOMATCH);
3503                  case 0x09:      /* HT */
3504                  case 0x20:      /* SPACE */
3505                  case 0xa0:      /* NBSP */
3506                  case 0x1680:    /* OGHAM SPACE MARK */
3507                  case 0x180e:    /* MONGOLIAN VOWEL SEPARATOR */
3508                  case 0x2000:    /* EN QUAD */
3509                  case 0x2001:    /* EM QUAD */
3510                  case 0x2002:    /* EN SPACE */
3511                  case 0x2003:    /* EM SPACE */
3512                  case 0x2004:    /* THREE-PER-EM SPACE */
3513                  case 0x2005:    /* FOUR-PER-EM SPACE */
3514                  case 0x2006:    /* SIX-PER-EM SPACE */
3515                  case 0x2007:    /* FIGURE SPACE */
3516                  case 0x2008:    /* PUNCTUATION SPACE */
3517                  case 0x2009:    /* THIN SPACE */
3518                  case 0x200A:    /* HAIR SPACE */
3519                  case 0x202f:    /* NARROW NO-BREAK SPACE */
3520                  case 0x205f:    /* MEDIUM MATHEMATICAL SPACE */
3521                  case 0x3000:    /* IDEOGRAPHIC SPACE */
3522                  break;
3523                  }
3524                break;
3525    
3526                case OP_NOT_VSPACE:
3527                switch(c)
3528                  {
3529                  default: break;
3530                  case 0x0a:      /* LF */
3531                  case 0x0b:      /* VT */
3532                  case 0x0c:      /* FF */
3533                  case 0x0d:      /* CR */
3534                  case 0x85:      /* NEL */
3535                  case 0x2028:    /* LINE SEPARATOR */
3536                  case 0x2029:    /* PARAGRAPH SEPARATOR */
3537                  RRETURN(MATCH_NOMATCH);
3538                  }
3539                break;
3540    
3541                case OP_VSPACE:
3542                switch(c)
3543                  {
3544                  default: RRETURN(MATCH_NOMATCH);
3545                  case 0x0a:      /* LF */
3546                  case 0x0b:      /* VT */
3547                  case 0x0c:      /* FF */
3548                  case 0x0d:      /* CR */
3549                  case 0x85:      /* NEL */
3550                  case 0x2028:    /* LINE SEPARATOR */
3551                  case 0x2029:    /* PARAGRAPH SEPARATOR */
3552                break;                break;
3553                }                }
3554              break;              break;
# Line 3117  for (;;) Line 3594  for (;;)
3594          {          {
3595          for (fi = min;; fi++)          for (fi = min;; fi++)
3596            {            {
3597            RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM43);
3598            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3599            if (fi >= max || eptr >= md->end_subject ||            if (fi >= max || eptr >= md->end_subject ||
3600                 ((ims & PCRE_DOTALL) == 0 && IS_NEWLINE(eptr)))                 (ctype == OP_ANY && IS_NEWLINE(eptr)))
3601              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
3602    
3603            c = *eptr++;            c = *eptr++;
3604            switch(ctype)            switch(ctype)
3605              {              {
3606              case OP_ANY:   /* This is the DOTALL case */              case OP_ANY:     /* This is the non-NL case */
3607              break;              case OP_ALLANY:
   
3608              case OP_ANYBYTE:              case OP_ANYBYTE:
3609              break;              break;
3610    
# Line 3139  for (;;) Line 3615  for (;;)
3615                case 0x000d:                case 0x000d:
3616                if (eptr < md->end_subject && *eptr == 0x0a) eptr++;                if (eptr < md->end_subject && *eptr == 0x0a) eptr++;
3617                break;                break;
3618    
3619                case 0x000a:                case 0x000a:
3620                  break;
3621    
3622                case 0x000b:                case 0x000b:
3623                case 0x000c:                case 0x000c:
3624                case 0x0085:                case 0x0085:
3625                  if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);
3626                  break;
3627                  }
3628                break;
3629    
3630                case OP_NOT_HSPACE:
3631                switch(c)
3632                  {
3633                  default: break;
3634                  case 0x09:      /* HT */
3635                  case 0x20:      /* SPACE */
3636                  case 0xa0:      /* NBSP */
3637                  RRETURN(MATCH_NOMATCH);
3638                  }
3639                break;
3640    
3641                case OP_HSPACE:
3642                switch(c)
3643                  {
3644                  default: RRETURN(MATCH_NOMATCH);
3645                  case 0x09:      /* HT */
3646                  case 0x20:      /* SPACE */
3647                  case 0xa0:      /* NBSP */
3648                  break;
3649                  }
3650                break;
3651    
3652                case OP_NOT_VSPACE:
3653                switch(c)
3654                  {
3655                  default: break;
3656                  case 0x0a:      /* LF */
3657                  case 0x0b:      /* VT */
3658                  case 0x0c:      /* FF */
3659                  case 0x0d:      /* CR */
3660                  case 0x85:      /* NEL */
3661                  RRETURN(MATCH_NOMATCH);
3662                  }
3663                break;
3664    
3665                case OP_VSPACE:
3666                switch(c)
3667                  {
3668                  default: RRETURN(MATCH_NOMATCH);
3669                  case 0x0a:      /* LF */
3670                  case 0x0b:      /* VT */
3671                  case 0x0c:      /* FF */
3672                  case 0x0d:      /* CR */
3673                  case 0x85:      /* NEL */
3674                break;                break;
3675                }                }
3676              break;              break;
# Line 3209  for (;;) Line 3737  for (;;)
3737              int len = 1;              int len = 1;
3738              if (eptr >= md->end_subject) break;              if (eptr >= md->end_subject) break;
3739              GETCHARLEN(c, eptr, len);              GETCHARLEN(c, eptr, len);
3740              prop_category = _pcre_ucp_findprop(c, &prop_chartype, &prop_script);              prop_chartype = UCD_CHARTYPE(c);
3741              if ((prop_chartype == ucp_Lu ||              if ((prop_chartype == ucp_Lu ||
3742                   prop_chartype == ucp_Ll ||                   prop_chartype == ucp_Ll ||
3743                   prop_chartype == ucp_Lt) == prop_fail_result)                   prop_chartype == ucp_Lt) == prop_fail_result)
# Line 3224  for (;;) Line 3752  for (;;)
3752              int len = 1;              int len = 1;
3753              if (eptr >= md->end_subject) break;              if (eptr >= md->end_subject) break;
3754              GETCHARLEN(c, eptr, len);              GETCHARLEN(c, eptr, len);
3755              prop_category = _pcre_ucp_findprop(c, &prop_chartype, &prop_script);              prop_category = UCD_CATEGORY(c);
3756              if ((prop_category == prop_value) == prop_fail_result)              if ((prop_category == prop_value) == prop_fail_result)
3757                break;                break;
3758              eptr+= len;              eptr+= len;
# Line 3237  for (;;) Line 3765  for (;;)
3765              int len = 1;              int len = 1;
3766              if (eptr >= md->end_subject) break;              if (eptr >= md->end_subject) break;
3767              GETCHARLEN(c, eptr, len);              GETCHARLEN(c, eptr, len);
3768              prop_category = _pcre_ucp_findprop(c, &prop_chartype, &prop_script);              prop_chartype = UCD_CHARTYPE(c);
3769              if ((prop_chartype == prop_value) == prop_fail_result)              if ((prop_chartype == prop_value) == prop_fail_result)
3770                break;                break;
3771              eptr+= len;              eptr+= len;
# Line 3250  for (;;) Line 3778  for (;;)
3778              int len = 1;              int len = 1;
3779              if (eptr >= md->end_subject) break;              if (eptr >= md->end_subject) break;
3780              GETCHARLEN(c, eptr, len);              GETCHARLEN(c, eptr, len);
3781              prop_category = _pcre_ucp_findprop(c, &prop_chartype, &prop_script);              prop_script = UCD_SCRIPT(c);
3782              if ((prop_script == prop_value) == prop_fail_result)              if ((prop_script == prop_value) == prop_fail_result)
3783                break;                break;
3784              eptr+= len;              eptr+= len;
# Line 3263  for (;;) Line 3791  for (;;)
3791          if (possessive) continue;          if (possessive) continue;
3792          for(;;)          for(;;)
3793            {            {
3794            RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM44);
3795            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3796            if (eptr-- == pp) break;        /* Stop if tried at original pos */            if (eptr-- == pp) break;        /* Stop if tried at original pos */
3797            BACKCHAR(eptr);            if (utf8) BACKCHAR(eptr);
3798            }            }
3799          }          }
3800    
# Line 3279  for (;;) Line 3807  for (;;)
3807            {            {
3808            if (eptr >= md->end_subject) break;            if (eptr >= md->end_subject) break;
3809            GETCHARINCTEST(c, eptr);            GETCHARINCTEST(c, eptr);
3810            prop_category = _pcre_ucp_findprop(c, &prop_chartype, &prop_script);            prop_category = UCD_CATEGORY(c);
3811            if (prop_category == ucp_M) break;            if (prop_category == ucp_M) break;
3812            while (eptr < md->end_subject)            while (eptr < md->end_subject)
3813              {              {
# Line 3288  for (;;) Line 3816  for (;;)
3816                {                {
3817                GETCHARLEN(c, eptr, len);                GETCHARLEN(c, eptr, len);
3818                }                }
3819              prop_category = _pcre_ucp_findprop(c, &prop_chartype, &prop_script);              prop_category = UCD_CATEGORY(c);
3820              if (prop_category != ucp_M) break;              if (prop_category != ucp_M) break;
3821              eptr += len;              eptr += len;
3822              }              }
# Line 3299  for (;;) Line 3827  for (;;)
3827          if (possessive) continue;          if (possessive) continue;
3828          for(;;)          for(;;)
3829            {            {
3830            RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM45);
3831            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3832            if (eptr-- == pp) break;        /* Stop if tried at original pos */            if (eptr-- == pp) break;        /* Stop if tried at original pos */
3833            for (;;)                        /* Move back over one extended */            for (;;)                        /* Move back over one extended */
3834              {              {
3835              int len = 1;              int len = 1;
             BACKCHAR(eptr);  
3836              if (!utf8) c = *eptr; else              if (!utf8) c = *eptr; else
3837                {                {
3838                  BACKCHAR(eptr);
3839                GETCHARLEN(c, eptr, len);                GETCHARLEN(c, eptr, len);
3840                }                }
3841              prop_category = _pcre_ucp_findprop(c, &prop_chartype, &prop_script);              prop_category = UCD_CATEGORY(c);
3842              if (prop_category != ucp_M) break;              if (prop_category != ucp_M) break;
3843              eptr--;              eptr--;
3844              }              }
# Line 3328  for (;;) Line 3856  for (;;)
3856          switch(ctype)          switch(ctype)
3857            {            {
3858            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. */  
   
3859            if (max < INT_MAX)            if (max < INT_MAX)
3860              {              {
3861              if ((ims & PCRE_DOTALL) == 0)              for (i = min; i < max; i++)
               {  
               for (i = min; i < max; i++)  
                 {  
                 if (eptr >= md->end_subject || IS_NEWLINE(eptr)) break;  
                 eptr++;  
                 while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;  
                 }  
               }  
             else  
3862                {                {
3863                for (i = min; i < max; i++)                if (eptr >= md->end_subject || IS_NEWLINE(eptr)) break;
3864                  {                eptr++;
3865                  if (eptr >= md->end_subject) break;                while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;
                 eptr++;  
                 while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;  
                 }  
3866                }                }
3867              }              }
3868    
# Line 3359  for (;;) Line 3870  for (;;)
3870    
3871            else            else
3872              {              {
3873              if ((ims & PCRE_DOTALL) == 0)              for (i = min; i < max; i++)
3874                {                {
3875                for (i = min; i < max; i++)                if (eptr >= md->end_subject || IS_NEWLINE(eptr)) break;
3876                  {                eptr++;
3877                  if (eptr >= md->end_subject || IS_NEWLINE(eptr)) break;                while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;
                 eptr++;  
                 }  
               break;  
3878                }                }
3879              else              }
3880              break;
3881    
3882              case OP_ALLANY:
3883              if (max < INT_MAX)
3884                {
3885                for (i = min; i < max; i++)
3886                {                {
3887                c = max - min;                if (eptr >= md->end_subject) break;
3888                if (c > (unsigned int)(md->end_subject - eptr))                eptr++;
3889                  c = md->end_subject - eptr;                while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;
               eptr += c;  
3890                }                }
3891              }              }
3892              else eptr = md->end_subject;   /* Unlimited UTF-8 repeat */
3893            break;            break;
3894    
3895            /* The byte case is the same as non-UTF8 */            /* The byte case is the same as non-UTF8 */
# Line 3400  for (;;) Line 3914  for (;;)
3914                }                }
3915              else              else
3916                {                {
3917                if (c != 0x000a && c != 0x000b && c != 0x000c &&                if (c != 0x000a &&
3918                    c != 0x0085 && c != 0x2028 && c != 0x2029)                    (md->bsr_anycrlf ||
3919                       (c != 0x000b && c != 0x000c &&
3920                        c != 0x0085 && c != 0x2028 && c != 0x2029)))
3921                  break;                  break;
3922                eptr += len;                eptr += len;
3923                }                }
3924              }              }
3925            break;            break;
3926    
3927              case OP_NOT_HSPACE:
3928              case OP_HSPACE:
3929              for (i = min; i < max; i++)
3930                {
3931                BOOL gotspace;
3932                int len = 1;
3933                if (eptr >= md->end_subject) break;
3934                GETCHARLEN(c, eptr, len);
3935                switch(c)
3936                  {
3937                  default: gotspace = FALSE; break;
3938                  case 0x09:      /* HT */
3939                  case 0x20:      /* SPACE */
3940                  case 0xa0:      /* NBSP */
3941                  case 0x1680:    /* OGHAM SPACE MARK */
3942                  case 0x180e:    /* MONGOLIAN VOWEL SEPARATOR */
3943                  case 0x2000:    /* EN QUAD */
3944                  case 0x2001:    /* EM QUAD */
3945                  case 0x2002:    /* EN SPACE */
3946                  case 0x2003:    /* EM SPACE */
3947                  case 0x2004:    /* THREE-PER-EM SPACE */
3948                  case 0x2005:    /* FOUR-PER-EM SPACE */
3949                  case 0x2006:    /* SIX-PER-EM SPACE */
3950                  case 0x2007:    /* FIGURE SPACE */
3951                  case 0x2008:    /* PUNCTUATION SPACE */
3952                  case 0x2009:    /* THIN SPACE */
3953                  case 0x200A:    /* HAIR SPACE */
3954                  case 0x202f:    /* NARROW NO-BREAK SPACE */
3955                  case 0x205f:    /* MEDIUM MATHEMATICAL SPACE */
3956                  case 0x3000:    /* IDEOGRAPHIC SPACE */
3957                  gotspace = TRUE;
3958                  break;
3959                  }
3960                if (gotspace == (ctype == OP_NOT_HSPACE)) break;
3961                eptr += len;
3962                }
3963              break;
3964    
3965              case OP_NOT_VSPACE:
3966              case OP_VSPACE:
3967              for (i = min; i < max; i++)
3968                {
3969                BOOL gotspace;
3970                int len = 1;
3971                if (eptr >= md->end_subject) break;
3972                GETCHARLEN(c, eptr, len);
3973                switch(c)
3974                  {
3975                  default: gotspace = FALSE; break;
3976                  case 0x0a:      /* LF */
3977                  case 0x0b:      /* VT */
3978                  case 0x0c:      /* FF */
3979                  case 0x0d:      /* CR */
3980                  case 0x85:      /* NEL */
3981                  case 0x2028:    /* LINE SEPARATOR */
3982                  case 0x2029:    /* PARAGRAPH SEPARATOR */
3983                  gotspace = TRUE;
3984                  break;
3985                  }
3986                if (gotspace == (ctype == OP_NOT_VSPACE)) break;
3987                eptr += len;
3988                }
3989              break;
3990    
3991            case OP_NOT_DIGIT:            case OP_NOT_DIGIT:
3992            for (i = min; i < max; i++)            for (i = min; i < max; i++)
3993              {              {
# Line 3483  for (;;) Line 4063  for (;;)
4063          if (possessive) continue;          if (possessive) continue;
4064          for(;;)          for(;;)
4065            {            {
4066            RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM46);
4067            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
4068            if (eptr-- == pp) break;        /* Stop if tried at original pos */            if (eptr-- == pp) break;        /* Stop if tried at original pos */
4069            BACKCHAR(eptr);            BACKCHAR(eptr);
4070            }            }
4071          }          }
4072        else        else
4073  #endif  #endif  /* SUPPORT_UTF8 */
4074    
4075        /* Not UTF-8 mode */        /* Not UTF-8 mode */
4076          {          {
4077          switch(ctype)          switch(ctype)
4078            {            {
4079            case OP_ANY:            case OP_ANY:
4080            if ((ims & PCRE_DOTALL) == 0)            for (i = min; i < max; i++)
4081              {              {
4082              for (i = min; i < max; i++)              if (eptr >= md->end_subject || IS_NEWLINE(eptr)) break;
4083                {              eptr++;
               if (eptr >= md->end_subject || IS_NEWLINE(eptr)) break;  
               eptr++;  
               }  
             break;  
4084              }              }
4085            /* For DOTALL case, fall through and treat as \C */            break;
4086    
4087              case OP_ALLANY:
4088            case OP_ANYBYTE:            case OP_ANYBYTE:
4089            c = max - min;            c = max - min;
4090            if (c > (unsigned int)(md->end_subject - eptr))            if (c > (unsigned int)(md->end_subject - eptr))
# Line 3527  for (;;) Line 4104  for (;;)
4104                }                }
4105              else              else
4106                {                {
4107                if (c != 0x000a && c != 0x000b && c != 0x000c && c != 0x0085)                if (c != 0x000a &&
4108                      (md->bsr_anycrlf ||
4109                        (c != 0x000b && c != 0x000c && c != 0x0085)))
4110                  break;                  break;
4111                eptr++;                eptr++;
4112                }                }
4113              }              }
4114            break;            break;
4115    
4116              case OP_NOT_HSPACE:
4117              for (i = min; i < max; i++)
4118                {
4119                if (eptr >= md->end_subject) break;
4120                c = *eptr;
4121                if (c == 0x09 || c == 0x20 || c == 0xa0) break;
4122                eptr++;
4123                }
4124              break;
4125    
4126              case OP_HSPACE:
4127              for (i = min; i < max; i++)
4128                {
4129                if (eptr >= md->end_subject) break;
4130                c = *eptr;
4131                if (c != 0x09 && c != 0x20 && c != 0xa0) break;
4132                eptr++;
4133                }
4134              break;
4135    
4136              case OP_NOT_VSPACE:
4137              for (i = min; i < max; i++)
4138                {
4139                if (eptr >= md->end_subject) break;
4140                c = *eptr;
4141                if (c == 0x0a || c == 0x0b || c == 0x0c || c == 0x0d || c == 0x85)
4142                  break;
4143                eptr++;
4144                }
4145              break;
4146    
4147              case OP_VSPACE:
4148              for (i = min; i < max; i++)
4149                {
4150                if (eptr >= md->end_subject) break;
4151                c = *eptr;
4152                if (c != 0x0a && c != 0x0b && c != 0x0c && c != 0x0d && c != 0x85)
4153                  break;
4154                eptr++;
4155                }
4156              break;
4157    
4158            case OP_NOT_DIGIT:            case OP_NOT_DIGIT:
4159            for (i = min; i < max; i++)            for (i = min; i < max; i++)
4160              {              {
# Line 3597  for (;;) Line 4218  for (;;)
4218          if (possessive) continue;          if (possessive) continue;
4219          while (eptr >= pp)          while (eptr >= pp)
4220            {            {
4221            RMATCH(rrc, eptr, ecode, offset_top, md, ims, eptrb, 0);            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM47);
4222            eptr--;            eptr--;
4223            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
4224            }            }
# Line 3623  for (;;) Line 4244  for (;;)
4244    
4245    }             /* End of main loop */    }             /* End of main loop */
4246  /* Control never reaches here */  /* Control never reaches here */
4247    
4248    
4249    /* When compiling to use the heap rather than the stack for recursive calls to
4250    match(), the RRETURN() macro jumps here. The number that is saved in
4251    frame->Xwhere indicates which label we actually want to return to. */
4252    
4253    #ifdef NO_RECURSE
4254    #define LBL(val) case val: goto L_RM##val;
4255    HEAP_RETURN:
4256    switch (frame->Xwhere)
4257      {
4258      LBL( 1) LBL( 2) LBL( 3) LBL( 4) LBL( 5) LBL( 6) LBL( 7) LBL( 8)
4259      LBL( 9) LBL(10) LBL(11) LBL(12) LBL(13) LBL(14) LBL(15) LBL(17)
4260      LBL(19) LBL(24) LBL(25) LBL(26) LBL(27) LBL(29) LBL(31) LBL(33)
4261      LBL(35) LBL(43) LBL(47) LBL(48) LBL(49) LBL(50) LBL(51) LBL(52)
4262      LBL(53) LBL(54)
4263    #ifdef SUPPORT_UTF8
4264      LBL(16) LBL(18) LBL(20) LBL(21) LBL(22) LBL(23) LBL(28) LBL(30)
4265      LBL(32) LBL(34) LBL(42) LBL(46)
4266    #ifdef SUPPORT_UCP
4267      LBL(36) LBL(37) LBL(38) LBL(39) LBL(40) LBL(41) LBL(44) LBL(45)
4268    #endif  /* SUPPORT_UCP */
4269    #endif  /* SUPPORT_UTF8 */
4270      default:
4271      DPRINTF(("jump error in pcre match: label %d non-existent\n", frame->Xwhere));
4272      return PCRE_ERROR_INTERNAL;
4273      }
4274    #undef LBL
4275    #endif  /* NO_RECURSE */
4276  }  }
4277    
4278    
# Line 3635  Undefine all the macros that were define Line 4285  Undefine all the macros that were define
4285  #ifdef NO_RECURSE  #ifdef NO_RECURSE
4286  #undef eptr  #undef eptr
4287  #undef ecode  #undef ecode
4288    #undef mstart
4289  #undef offset_top  #undef offset_top
4290  #undef ims  #undef ims
4291  #undef eptrb  #undef eptrb
# Line 3707  Returns:          > 0 => success; value Line 4358  Returns:          > 0 => success; value
4358                   < -1 => some kind of unexpected problem                   < -1 => some kind of unexpected problem
4359  */  */
4360    
4361  PCRE_DATA_SCOPE int  PCRE_EXP_DEFN int
4362  pcre_exec(const pcre *argument_re, const pcre_extra *extra_data,  pcre_exec(const pcre *argument_re, const pcre_extra *extra_data,
4363    PCRE_SPTR subject, int length, int start_offset, int options, int *offsets,    PCRE_SPTR subject, int length, int start_offset, int options, int *offsets,
4364    int offsetcount)    int offsetcount)
# Line 3732  const uschar *start_bits = NULL; Line 4383  const uschar *start_bits = NULL;
4383  USPTR start_match = (USPTR)subject + start_offset;  USPTR start_match = (USPTR)subject + start_offset;
4384  USPTR end_subject;  USPTR end_subject;
4385  USPTR req_byte_ptr = start_match - 1;  USPTR req_byte_ptr = start_match - 1;
 eptrblock eptrchain[EPTR_WORK_SIZE];  
4386    
4387  pcre_study_data internal_study;  pcre_study_data internal_study;
4388  const pcre_study_data *study;  const pcre_study_data *study;
# Line 3795  if (re->magic_number != MAGIC_NUMBER) Line 4445  if (re->magic_number != MAGIC_NUMBER)
4445  /* Set up other data */  /* Set up other data */
4446    
4447  anchored = ((re->options | options) & PCRE_ANCHORED) != 0;  anchored = ((re->options | options) & PCRE_ANCHORED) != 0;
4448  startline = (re->options & PCRE_STARTLINE) != 0;  startline = (re->flags & PCRE_STARTLINE) != 0;
4449  firstline = (re->options & PCRE_FIRSTLINE) != 0;  firstline = (re->options & PCRE_FIRSTLINE) != 0;
4450    
4451  /* 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 3810  end_subject = md->end_subject; Line 4460  end_subject = md->end_subject;
4460    
4461  md->endonly = (re->options & PCRE_DOLLAR_ENDONLY) != 0;  md->endonly = (re->options & PCRE_DOLLAR_ENDONLY) != 0;
4462  utf8 = md->utf8 = (re->options & PCRE_UTF8) != 0;  utf8 = md->utf8 = (re->options & PCRE_UTF8) != 0;
4463    md->jscript_compat = (re->options & PCRE_JAVASCRIPT_COMPAT) != 0;
4464    
4465  md->notbol = (options & PCRE_NOTBOL) != 0;  md->notbol = (options & PCRE_NOTBOL) != 0;
4466  md->noteol = (options & PCRE_NOTEOL) != 0;  md->noteol = (options & PCRE_NOTEOL) != 0;
# Line 3818  md->partial = (options & PCRE_PARTIAL) ! Line 4469  md->partial = (options & PCRE_PARTIAL) !
4469  md->hitend = FALSE;  md->hitend = FALSE;
4470    
4471  md->recursive = NULL;                   /* No recursion at top level */  md->recursive = NULL;                   /* No recursion at top level */
 md->eptrchain = eptrchain;              /* Make workspace generally available */  
4472    
4473  md->lcc = tables + lcc_offset;  md->lcc = tables + lcc_offset;
4474  md->ctypes = tables + ctypes_offset;  md->ctypes = tables + ctypes_offset;
4475    
4476    /* Handle different \R options. */
4477    
4478    switch (options & (PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE))
4479      {
4480      case 0:
4481      if ((re->options & (PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE)) != 0)
4482        md->bsr_anycrlf = (re->options & PCRE_BSR_ANYCRLF) != 0;
4483      else
4484    #ifdef BSR_ANYCRLF
4485      md->bsr_anycrlf = TRUE;
4486    #else
4487      md->bsr_anycrlf = FALSE;
4488    #endif
4489      break;
4490    
4491      case PCRE_BSR_ANYCRLF:
4492      md->bsr_anycrlf = TRUE;
4493      break;
4494    
4495      case PCRE_BSR_UNICODE:
4496      md->bsr_anycrlf = FALSE;
4497      break;
4498    
4499      default: return PCRE_ERROR_BADNEWLINE;
4500      }
4501    
4502  /* Handle different types of newline. The three bits give eight cases. If  /* Handle different types of newline. The three bits give eight cases. If
4503  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. */
4504    
4505  switch ((((options & PCRE_NEWLINE_BITS) == 0)? re->options : options) &  switch ((((options & PCRE_NEWLINE_BITS) == 0)? re->options :
4506         PCRE_NEWLINE_BITS)          (pcre_uint32)options) & PCRE_NEWLINE_BITS)
4507    {    {
4508    case 0: newline = NEWLINE; break;   /* Compile-time default */    case 0: newline = NEWLINE; break;   /* Compile-time default */
4509    case PCRE_NEWLINE_CR: newline = '\r'; break;    case PCRE_NEWLINE_CR: newline = '\r'; break;
# Line 3835  switch ((((options & PCRE_NEWLINE_BITS) Line 4511  switch ((((options & PCRE_NEWLINE_BITS)
4511    case PCRE_NEWLINE_CR+    case PCRE_NEWLINE_CR+
4512         PCRE_NEWLINE_LF: newline = ('\r' << 8) | '\n'; break;         PCRE_NEWLINE_LF: newline = ('\r' << 8) | '\n'; break;
4513    case PCRE_NEWLINE_ANY: newline = -1; break;    case PCRE_NEWLINE_ANY: newline = -1; break;
4514      case PCRE_NEWLINE_ANYCRLF: newline = -2; break;
4515    default: return PCRE_ERROR_BADNEWLINE;    default: return PCRE_ERROR_BADNEWLINE;
4516    }    }
4517    
4518  if (newline < 0)  if (newline == -2)
4519      {
4520      md->nltype = NLTYPE_ANYCRLF;
4521      }
4522    else if (newline < 0)
4523    {    {
4524    md->nltype = NLTYPE_ANY;    md->nltype = NLTYPE_ANY;
4525    }    }
# Line 3861  else Line 4542  else
4542  /* 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
4543  moment. */  moment. */
4544    
4545  if (md->partial && (re->options & PCRE_NOPARTIAL) != 0)  if (md->partial && (re->flags & PCRE_NOPARTIAL) != 0)
4546    return PCRE_ERROR_BADPARTIAL;    return PCRE_ERROR_BADPARTIAL;
4547    
4548  /* 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 3938  studied, there may be a bitmap of possib Line 4619  studied, there may be a bitmap of possib
4619    
4620  if (!anchored)  if (!anchored)
4621    {    {
4622    if ((re->options & PCRE_FIRSTSET) != 0)    if ((re->flags & PCRE_FIRSTSET) != 0)
4623      {      {
4624      first_byte = re->first_byte & 255;      first_byte = re->first_byte & 255;
4625      if ((first_byte_caseless = ((re->first_byte & REQ_CASELESS) != 0)) == TRUE)      if ((first_byte_caseless = ((re->first_byte & REQ_CASELESS) != 0)) == TRUE)
# Line 3953  if (!anchored) Line 4634  if (!anchored)
4634  /* For anchored or unanchored matches, there may be a "last known required  /* For anchored or unanchored matches, there may be a "last known required
4635  character" set. */  character" set. */
4636    
4637  if ((re->options & PCRE_REQCHSET) != 0)  if ((re->flags & PCRE_REQCHSET) != 0)
4638    {    {
4639    req_byte = re->req_byte & 255;    req_byte = re->req_byte & 255;
4640    req_byte_caseless = (re->req_byte & REQ_CASELESS) != 0;    req_byte_caseless = (re->req_byte & REQ_CASELESS) != 0;
# Line 3969  the loop runs just once. */ Line 4650  the loop runs just once. */
4650  for(;;)  for(;;)
4651    {    {
4652    USPTR save_end_subject = end_subject;    USPTR save_end_subject = end_subject;
4653      USPTR new_start_match;
4654    
4655    /* Reset the maximum number of extractions we might see. */    /* Reset the maximum number of extractions we might see. */
4656    
# Line 3999  for(;;) Line 4681  for(;;)
4681      if (first_byte_caseless)      if (first_byte_caseless)
4682        while (start_match < end_subject &&        while (start_match < end_subject &&
4683               md->lcc[*start_match] != first_byte)               md->lcc[*start_match] != first_byte)
4684          start_match++;          { NEXTCHAR(start_match); }
4685      else      else
4686        while (start_match < end_subject && *start_match != first_byte)        while (start_match < end_subject && *start_match != first_byte)
4687          start_match++;          { NEXTCHAR(start_match); }
4688      }      }
4689    
4690    /* 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 4012  for(;;) Line 4694  for(;;)
4694      if (start_match > md->start_subject + start_offset)      if (start_match > md->start_subject + start_offset)
4695        {        {
4696        while (start_match <= end_subject && !WAS_NEWLINE(start_match))        while (start_match <= end_subject && !WAS_NEWLINE(start_match))
4697            { NEXTCHAR(start_match); }
4698    
4699          /* If we have just passed a CR and the newline option is ANY or ANYCRLF,
4700          and we are now at a LF, advance the match position by one more character.
4701          */
4702    
4703          if (start_match[-1] == '\r' &&
4704               (md->nltype == NLTYPE_ANY || md->nltype == NLTYPE_ANYCRLF) &&
4705               start_match < end_subject &&
4706               *start_match == '\n')
4707          start_match++;          start_match++;
4708        }        }
4709      }      }
# Line 4023  for(;;) Line 4715  for(;;)
4715      while (start_match < end_subject)      while (start_match < end_subject)
4716        {        {
4717        register unsigned int c = *start_match;        register unsigned int c = *start_match;
4718        if ((start_bits[c/8] & (1 << (c&7))) == 0) start_match++; else break;        if ((start_bits[c/8] & (1 << (c&7))) == 0)
4719            { NEXTCHAR(start_match); }
4720          else break;
4721        }        }
4722      }      }
4723    
# Line 4099  for(;;) Line 4793  for(;;)
4793    
4794    /* OK, we can now run the match. */    /* OK, we can now run the match. */
4795    
4796    md->start_match = start_match;    md->start_match_ptr = start_match;
4797    md->match_call_count = 0;    md->match_call_count = 0;
4798    md->eptrn = 0;                          /* Next free eptrchain slot */    rc = match(start_match, md->start_code, start_match, 2, md, ims, NULL, 0, 0);
4799    rc = match(start_match, md->start_code, 2, md, ims, NULL, 0, 0);  
4800      switch(rc)
4801        {
4802        /* NOMATCH and PRUNE advance by one character. THEN at this level acts
4803        exactly like PRUNE. */
4804    
4805        case MATCH_NOMATCH:
4806        case MATCH_PRUNE:
4807        case MATCH_THEN:
4808        new_start_match = start_match + 1;
4809    #ifdef SUPPORT_UTF8
4810        if (utf8)
4811          while(new_start_match < end_subject && (*new_start_match & 0xc0) == 0x80)
4812            new_start_match++;
4813    #endif
4814        break;
4815    
4816        /* SKIP passes back the next starting point explicitly. */
4817    
4818        case MATCH_SKIP:
4819        new_start_match = md->start_match_ptr;
4820        break;
4821    
4822    /* Any return other than MATCH_NOMATCH breaks the loop. */      /* COMMIT disables the bumpalong, but otherwise behaves as NOMATCH. */
4823    
4824    if (rc != MATCH_NOMATCH) break;      case MATCH_COMMIT:
4825        rc = MATCH_NOMATCH;
4826        goto ENDLOOP;
4827    
4828        /* Any other return is some kind of error. */
4829    
4830        default:
4831        goto ENDLOOP;
4832        }
4833    
4834      /* Control reaches here for the various types of "no match at this point"
4835      result. Reset the code to MATCH_NOMATCH for subsequent checking. */
4836    
4837      rc = MATCH_NOMATCH;
4838    
4839    /* 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
4840    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 4114  for(;;) Line 4842  for(;;)
4842    
4843    if (firstline && IS_NEWLINE(start_match)) break;    if (firstline && IS_NEWLINE(start_match)) break;
4844    
4845    /* Advance the match position by one character. */    /* Advance to new matching position */
4846    
4847    start_match++;    start_match = new_start_match;
 #ifdef SUPPORT_UTF8  
   if (utf8)  
     while(start_match < end_subject && (*start_match & 0xc0) == 0x80)  
       start_match++;  
 #endif  
4848    
4849    /* 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
4850    the subject. */    the subject. */
4851    
4852    if (anchored || start_match > end_subject) break;    if (anchored || start_match > end_subject) break;
4853    
4854    /* If we have just passed a CR and the newline option is CRLF or ANY, and we    /* If we have just passed a CR and we are now at a LF, and the pattern does
4855    are now at a LF, advance the match position by one more character. */    not contain any explicit matches for \r or \n, and the newline option is CRLF
4856      or ANY or ANYCRLF, advance the match position by one more character. */
4857    
4858    if (start_match[-1] == '\r' &&    if (start_match[-1] == '\r' &&
4859         (md->nltype == NLTYPE_ANY || md->nllen == 2) &&        start_match < end_subject &&
4860         start_match < end_subject &&        *start_match == '\n' &&
4861         *start_match == '\n')        (re->flags & PCRE_HASCRORLF) == 0 &&
4862            (md->nltype == NLTYPE_ANY ||
4863             md->nltype == NLTYPE_ANYCRLF ||
4864             md->nllen == 2))
4865      start_match++;      start_match++;
4866    
4867    }   /* End of for(;;) "bumpalong" loop */    }   /* End of for(;;) "bumpalong" loop */
# Line 4144  for(;;) Line 4871  for(;;)
4871  /* 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
4872  conditions is true:  conditions is true:
4873    
4874  (1) The pattern is anchored;  (1) The pattern is anchored or the match was failed by (*COMMIT);
4875    
4876  (2) We are past the end of the subject;  (2) We are past the end of the subject;
4877    
# Line 4159  processing, copy those that we can. In t Line 4886  processing, copy those that we can. In t
4886  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
4887  capturing parentheses than vector slots. */  capturing parentheses than vector slots. */
4888    
4889    ENDLOOP:
4890    
4891  if (rc == MATCH_MATCH)  if (rc == MATCH_MATCH)
4892    {    {
4893    if (using_temporary_offsets)    if (using_temporary_offsets)
# Line 4179  if (rc == MATCH_MATCH) Line 4908  if (rc == MATCH_MATCH)
4908    
4909    rc = md->offset_overflow? 0 : md->end_offset_top/2;    rc = md->offset_overflow? 0 : md->end_offset_top/2;
4910    
4911    /* 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
4912      md->start_match_ptr might be modified if \K was encountered on the success
4913      matching path. */
4914    
4915    if (offsetcount < 2) rc = 0; else    if (offsetcount < 2) rc = 0; else
4916      {      {
4917      offsets[0] = start_match - md->start_subject;      offsets[0] = md->start_match_ptr - md->start_subject;
4918      offsets[1] = md->end_match_ptr - md->start_subject;      offsets[1] = md->end_match_ptr - md->start_subject;
4919      }      }
4920    
Legend:
Removed from v.123  
changed lines
  Added in v.349
  ViewVC Help
Powered by ViewVC 1.1.5