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

Diff of /code/trunk/pcre_exec.c

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

revision 165 by ph10, Wed May 9 10:50:57 2007 UTC revision 459 by ph10, Sun Oct 4 09:21:39 2009 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) 1997-2009 University of Cambridge
10    
11  -----------------------------------------------------------------------------  -----------------------------------------------------------------------------
12  Redistribution and use in source and binary forms, with or without  Redistribution and use in source and binary forms, with or without
# Line 42  POSSIBILITY OF SUCH DAMAGE. Line 42  POSSIBILITY OF SUCH DAMAGE.
42  pattern matching using an NFA algorithm, trying to mimic Perl as closely as  pattern matching using an NFA algorithm, trying to mimic Perl as closely as
43  possible. There are also some static supporting functions. */  possible. There are also some static supporting functions. */
44    
45    #ifdef HAVE_CONFIG_H
46    #include "config.h"
47    #endif
48    
49  #define NLBLOCK md             /* Block containing newline information */  #define NLBLOCK md             /* Block containing newline information */
50  #define PSSTART start_subject  /* Field containing processed string start */  #define PSSTART start_subject  /* Field containing processed string start */
51  #define PSEND   end_subject    /* Field containing processed string end */  #define PSEND   end_subject    /* Field containing processed string end */
# Line 53  possible. There are also some static sup Line 57  possible. There are also some static sup
57  #undef min  #undef min
58  #undef max  #undef max
59    
 /* The chain of eptrblocks for tail recursions uses memory in stack workspace,  
 obtained at top level, the size of which is defined by EPTR_WORK_SIZE. */  
   
 #define EPTR_WORK_SIZE (1000)  
   
60  /* Flag bits for the match() function */  /* Flag bits for the match() function */
61    
62  #define match_condassert     0x01  /* Called to check a condition assertion */  #define match_condassert     0x01  /* Called to check a condition assertion */
63  #define match_cbegroup       0x02  /* Could-be-empty unlimited repeat group */  #define match_cbegroup       0x02  /* Could-be-empty unlimited repeat group */
 #define match_tail_recursed  0x04  /* Tail recursive call */  
64    
65  /* Non-error returns from the match() function. Error returns are externally  /* Non-error returns from the match() function. Error returns are externally
66  defined PCRE_ERROR_xxx codes, which are all negative. */  defined PCRE_ERROR_xxx codes, which are all negative. */
# Line 70  defined PCRE_ERROR_xxx codes, which are Line 68  defined PCRE_ERROR_xxx codes, which are
68  #define MATCH_MATCH        1  #define MATCH_MATCH        1
69  #define MATCH_NOMATCH      0  #define MATCH_NOMATCH      0
70    
71    /* Special internal returns from the match() function. Make them sufficiently
72    negative to avoid the external error codes. */
73    
74    #define MATCH_COMMIT       (-999)
75    #define MATCH_PRUNE        (-998)
76    #define MATCH_SKIP         (-997)
77    #define MATCH_THEN         (-996)
78    
79  /* Maximum number of ints of offset to save on the stack for recursive calls.  /* Maximum number of ints of offset to save on the stack for recursive calls.
80  If the offset vector is bigger, malloc is used. This should be a multiple of 3,  If the offset vector is bigger, malloc is used. This should be a multiple of 3,
81  because the offset vector is always a multiple of 3 long. */  because the offset vector is always a multiple of 3 long. */
# Line 152  printf("\n"); Line 158  printf("\n");
158    
159  if (length > md->end_subject - eptr) return FALSE;  if (length > md->end_subject - eptr) return FALSE;
160    
161  /* Separate the caselesss case for speed */  /* Separate the caseless case for speed. In UTF-8 mode we can only do this
162    properly if Unicode properties are supported. Otherwise, we can check only
163    ASCII characters. */
164    
165  if ((ims & PCRE_CASELESS) != 0)  if ((ims & PCRE_CASELESS) != 0)
166    {    {
167    #ifdef SUPPORT_UTF8
168    #ifdef SUPPORT_UCP
169      if (md->utf8)
170        {
171        USPTR endptr = eptr + length;
172        while (eptr < endptr)
173          {
174          int c, d;
175          GETCHARINC(c, eptr);
176          GETCHARINC(d, p);
177          if (c != d && c != UCD_OTHERCASE(d)) return FALSE;
178          }
179        }
180      else
181    #endif
182    #endif
183    
184      /* The same code works when not in UTF-8 mode and in UTF-8 mode when there
185      is no UCP support. */
186    
187    while (length-- > 0)    while (length-- > 0)
188      if (md->lcc[*p++] != md->lcc[*eptr++]) return FALSE;      { if (md->lcc[*p++] != md->lcc[*eptr++]) return FALSE; }
189    }    }
190    
191    /* In the caseful case, we can just compare the bytes, whether or not we
192    are in UTF-8 mode. */
193    
194  else  else
195    { while (length-- > 0) if (*p++ != *eptr++) return FALSE; }    { while (length-- > 0) if (*p++ != *eptr++) return FALSE; }
196    
# Line 205  variable instead of being passed in the Line 237  variable instead of being passed in the
237  ****************************************************************************  ****************************************************************************
238  ***************************************************************************/  ***************************************************************************/
239    
240    /* Numbers for RMATCH calls. When this list is changed, the code at HEAP_RETURN
241  /* Numbers for RMATCH calls */  below must be updated in sync.  */
242    
243  enum { RM1=1, RM2,  RM3,  RM4,  RM5,  RM6,  RM7,  RM8,  RM9,  RM10,  enum { RM1=1, RM2,  RM3,  RM4,  RM5,  RM6,  RM7,  RM8,  RM9,  RM10,
244         RM11,  RM12, RM13, RM14, RM15, RM16, RM17, RM18, RM19, RM20,         RM11,  RM12, RM13, RM14, RM15, RM16, RM17, RM18, RM19, RM20,
245         RM21,  RM22, RM23, RM24, RM25, RM26, RM27, RM28, RM29, RM30,         RM21,  RM22, RM23, RM24, RM25, RM26, RM27, RM28, RM29, RM30,
246         RM31,  RM32, RM33, RM34, RM35, RM36, RM37, RM38, RM39, RM40,         RM31,  RM32, RM33, RM34, RM35, RM36, RM37, RM38, RM39, RM40,
247         RM41,  RM42, RM43, RM44, RM45, RM46, RM47 };         RM41,  RM42, RM43, RM44, RM45, RM46, RM47, RM48, RM49, RM50,
248           RM51,  RM52, RM53, RM54 };
249    
250  /* 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
251  versions and production versions. Note that the "rw" argument of RMATCH isn't  versions and production versions. Note that the "rw" argument of RMATCH isn't
# Line 226  actuall used in this definition. */ Line 258  actuall used in this definition. */
258  #define RMATCH(ra,rb,rc,rd,re,rf,rg,rw) \  #define RMATCH(ra,rb,rc,rd,re,rf,rg,rw) \
259    { \    { \
260    printf("match() called in line %d\n", __LINE__); \    printf("match() called in line %d\n", __LINE__); \
261    rrc = match(ra,rb,rc,rd,re,rf,rg,rdepth+1); \    rrc = match(ra,rb,mstart,rc,rd,re,rf,rg,rdepth+1); \
262    printf("to line %d\n", __LINE__); \    printf("to line %d\n", __LINE__); \
263    }    }
264  #define RRETURN(ra) \  #define RRETURN(ra) \
# Line 236  actuall used in this definition. */ Line 268  actuall used in this definition. */
268    }    }
269  #else  #else
270  #define RMATCH(ra,rb,rc,rd,re,rf,rg,rw) \  #define RMATCH(ra,rb,rc,rd,re,rf,rg,rw) \
271    rrc = match(ra,rb,rc,rd,re,rf,rg,rdepth+1)    rrc = match(ra,rb,mstart,rc,rd,re,rf,rg,rdepth+1)
272  #define RRETURN(ra) return ra  #define RRETURN(ra) return ra
273  #endif  #endif
274    
# Line 255  argument of match(), which never changes Line 287  argument of match(), which never changes
287    frame->Xwhere = rw; \    frame->Xwhere = rw; \
288    newframe->Xeptr = ra;\    newframe->Xeptr = ra;\
289    newframe->Xecode = rb;\    newframe->Xecode = rb;\
290      newframe->Xmstart = mstart;\
291    newframe->Xoffset_top = rc;\    newframe->Xoffset_top = rc;\
292    newframe->Xims = re;\    newframe->Xims = re;\
293    newframe->Xeptrb = rf;\    newframe->Xeptrb = rf;\
# Line 289  typedef struct heapframe { Line 322  typedef struct heapframe {
322    
323    /* Function arguments that may change */    /* Function arguments that may change */
324    
325    const uschar *Xeptr;    USPTR Xeptr;
326    const uschar *Xecode;    const uschar *Xecode;
327      USPTR Xmstart;
328    int Xoffset_top;    int Xoffset_top;
329    long int Xims;    long int Xims;
330    eptrblock *Xeptrb;    eptrblock *Xeptrb;
# Line 299  typedef struct heapframe { Line 333  typedef struct heapframe {
333    
334    /* Function local variables */    /* Function local variables */
335    
336    const uschar *Xcallpat;    USPTR Xcallpat;
337    const uschar *Xcharptr;  #ifdef SUPPORT_UTF8
338    const uschar *Xdata;    USPTR Xcharptr;
339    const uschar *Xnext;  #endif
340    const uschar *Xpp;    USPTR Xdata;
341    const uschar *Xprev;    USPTR Xnext;
342    const uschar *Xsaved_eptr;    USPTR Xpp;
343      USPTR Xprev;
344      USPTR Xsaved_eptr;
345    
346    recursion_info Xnew_recursive;    recursion_info Xnew_recursive;
347    
# Line 326  typedef struct heapframe { Line 362  typedef struct heapframe {
362    uschar Xocchars[8];    uschar Xocchars[8];
363  #endif  #endif
364    
365      int Xcodelink;
366    int Xctype;    int Xctype;
367    unsigned int Xfc;    unsigned int Xfc;
368    int Xfi;    int Xfi;
# Line 361  typedef struct heapframe { Line 398  typedef struct heapframe {
398    
399  /* This function is called recursively in many circumstances. Whenever it  /* This function is called recursively in many circumstances. Whenever it
400  returns a negative (error) response, the outer incarnation must also return the  returns a negative (error) response, the outer incarnation must also return the
401  same response.  same response. */
402    
403    /* These macros pack up tests that are used for partial matching, and which
404    appears several times in the code. We set the "hit end" flag if the pointer is
405    at the end of the subject and also past the start of the subject (i.e.
406    something has been matched). For hard partial matching, we then return
407    immediately. The second one is used when we already know we are past the end of
408    the subject. */
409    
410  Performance note: It might be tempting to extract commonly used fields from the  #define CHECK_PARTIAL()\
411  md structure (e.g. utf8, end_subject) into individual variables to improve    if (md->partial != 0 && eptr >= md->end_subject && eptr > mstart)\
412        {\
413        md->hitend = TRUE;\
414        if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);\
415        }
416    
417    #define SCHECK_PARTIAL()\
418      if (md->partial && eptr > mstart)\
419        {\
420        md->hitend = TRUE;\
421        if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);\
422        }
423    
424    
425    /* Performance note: It might be tempting to extract commonly used fields from
426    the md structure (e.g. utf8, end_subject) into individual variables to improve
427  performance. Tests using gcc on a SPARC disproved this; in the first case, it  performance. Tests using gcc on a SPARC disproved this; in the first case, it
428  made performance worse.  made performance worse.
429    
430  Arguments:  Arguments:
431     eptr        pointer to current character in subject     eptr        pointer to current character in subject
432     ecode       pointer to current position in compiled code     ecode       pointer to current position in compiled code
433       mstart      pointer to the current match start position (can be modified
434                     by encountering \K)
435     offset_top  current top pointer     offset_top  current top pointer
436     md          pointer to "static" info for the match     md          pointer to "static" info for the match
437     ims         current /i, /m, and /s options     ims         current /i, /m, and /s options
# Line 380  Arguments: Line 441  Arguments:
441                   match_condassert - this is an assertion condition                   match_condassert - this is an assertion condition
442                   match_cbegroup - this is the start of an unlimited repeat                   match_cbegroup - this is the start of an unlimited repeat
443                     group that can match an empty string                     group that can match an empty string
                  match_tail_recursed - this is a tail_recursed group  
444     rdepth      the recursion depth     rdepth      the recursion depth
445    
446  Returns:       MATCH_MATCH if matched            )  these values are >= 0  Returns:       MATCH_MATCH if matched            )  these values are >= 0
# Line 390  Returns:       MATCH_MATCH if matched Line 450  Returns:       MATCH_MATCH if matched
450  */  */
451    
452  static int  static int
453  match(REGISTER USPTR eptr, REGISTER const uschar *ecode,  match(REGISTER USPTR eptr, REGISTER const uschar *ecode, USPTR mstart,
454    int offset_top, match_data *md, unsigned long int ims, eptrblock *eptrb,    int offset_top, match_data *md, unsigned long int ims, eptrblock *eptrb,
455    int flags, unsigned int rdepth)    int flags, unsigned int rdepth)
456  {  {
# Line 404  register unsigned int c;   /* Character Line 464  register unsigned int c;   /* Character
464  register BOOL utf8;        /* Local copy of UTF-8 flag for speed */  register BOOL utf8;        /* Local copy of UTF-8 flag for speed */
465    
466  BOOL minimize, possessive; /* Quantifier options */  BOOL minimize, possessive; /* Quantifier options */
467    int condcode;
468    
469  /* When recursion is not being used, all "local" variables that have to be  /* When recursion is not being used, all "local" variables that have to be
470  preserved over calls to RMATCH() are part of a "frame" which is obtained from  preserved over calls to RMATCH() are part of a "frame" which is obtained from
# Line 418  frame->Xprevframe = NULL;            /* Line 479  frame->Xprevframe = NULL;            /*
479    
480  frame->Xeptr = eptr;  frame->Xeptr = eptr;
481  frame->Xecode = ecode;  frame->Xecode = ecode;
482    frame->Xmstart = mstart;
483  frame->Xoffset_top = offset_top;  frame->Xoffset_top = offset_top;
484  frame->Xims = ims;  frame->Xims = ims;
485  frame->Xeptrb = eptrb;  frame->Xeptrb = eptrb;
# Line 432  HEAP_RECURSE: Line 494  HEAP_RECURSE:
494    
495  #define eptr               frame->Xeptr  #define eptr               frame->Xeptr
496  #define ecode              frame->Xecode  #define ecode              frame->Xecode
497    #define mstart             frame->Xmstart
498  #define offset_top         frame->Xoffset_top  #define offset_top         frame->Xoffset_top
499  #define ims                frame->Xims  #define ims                frame->Xims
500  #define eptrb              frame->Xeptrb  #define eptrb              frame->Xeptrb
# Line 444  HEAP_RECURSE: Line 507  HEAP_RECURSE:
507  #define charptr            frame->Xcharptr  #define charptr            frame->Xcharptr
508  #endif  #endif
509  #define callpat            frame->Xcallpat  #define callpat            frame->Xcallpat
510    #define codelink           frame->Xcodelink
511  #define data               frame->Xdata  #define data               frame->Xdata
512  #define next               frame->Xnext  #define next               frame->Xnext
513  #define pp                 frame->Xpp  #define pp                 frame->Xpp
# Line 524  int oclength; Line 588  int oclength;
588  uschar occhars[8];  uschar occhars[8];
589  #endif  #endif
590    
591    int codelink;
592  int ctype;  int ctype;
593  int length;  int length;
594  int max;  int max;
# Line 580  original_ims = ims;    /* Save for reset Line 645  original_ims = ims;    /* Save for reset
645  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
646  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
647  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.
648  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
649  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
650  already used. */  block that is used is on the stack, so a new one may be required for each
651    match(). */
652    
653  if ((flags & match_cbegroup) != 0)  if ((flags & match_cbegroup) != 0)
654    {    {
655    eptrblock *p;    newptrb.epb_saved_eptr = eptr;
656    if ((flags & match_tail_recursed) != 0)    newptrb.epb_prev = eptrb;
657      {    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;  
658    }    }
659    
660  /* Now start processing the opcodes. */  /* Now start processing the opcodes. */
# Line 605  for (;;) Line 664  for (;;)
664    minimize = possessive = FALSE;    minimize = possessive = FALSE;
665    op = *ecode;    op = *ecode;
666    
   /* For partial matching, remember if we ever hit the end of the subject after  
   matching at least one subject character. */  
   
   if (md->partial &&  
       eptr >= md->end_subject &&  
       eptr > md->start_match)  
     md->hitend = TRUE;  
   
667    switch(op)    switch(op)
668      {      {
669        case OP_FAIL:
670        RRETURN(MATCH_NOMATCH);
671    
672        case OP_PRUNE:
673        RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,
674          ims, eptrb, flags, RM51);
675        if (rrc != MATCH_NOMATCH) RRETURN(rrc);
676        RRETURN(MATCH_PRUNE);
677    
678        case OP_COMMIT:
679        RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,
680          ims, eptrb, flags, RM52);
681        if (rrc != MATCH_NOMATCH) RRETURN(rrc);
682        RRETURN(MATCH_COMMIT);
683    
684        case OP_SKIP:
685        RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,
686          ims, eptrb, flags, RM53);
687        if (rrc != MATCH_NOMATCH) RRETURN(rrc);
688        md->start_match_ptr = eptr;   /* Pass back current position */
689        RRETURN(MATCH_SKIP);
690    
691        case OP_THEN:
692        RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,
693          ims, eptrb, flags, RM54);
694        if (rrc != MATCH_NOMATCH) RRETURN(rrc);
695        RRETURN(MATCH_THEN);
696    
697      /* 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
698      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.
699      We mustn't change the current values of the data slot, because they may be      We mustn't change the current values of the data slot, because they may be
# Line 656  for (;;) Line 735  for (;;)
735          {          {
736          RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,          RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,
737            ims, eptrb, flags, RM1);            ims, eptrb, flags, RM1);
738          if (rrc != MATCH_NOMATCH) RRETURN(rrc);          if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN) RRETURN(rrc);
739          md->capture_last = save_capture_last;          md->capture_last = save_capture_last;
740          ecode += GET(ecode, 1);          ecode += GET(ecode, 1);
741          }          }
# Line 671  for (;;) Line 750  for (;;)
750        RRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
751        }        }
752    
753      /* Insufficient room for saving captured contents. Treat as a non-capturing      /* FALL THROUGH ... Insufficient room for saving captured contents. Treat
754      bracket. */      as a non-capturing bracket. */
755    
756        /* VVVVVVVVVVVVVVVVVVVVVVVVV */
757        /* VVVVVVVVVVVVVVVVVVVVVVVVV */
758    
759      DPRINTF(("insufficient capture room: treat as non-capturing\n"));      DPRINTF(("insufficient capture room: treat as non-capturing\n"));
760    
761        /* VVVVVVVVVVVVVVVVVVVVVVVVV */
762        /* VVVVVVVVVVVVVVVVVVVVVVVVV */
763    
764      /* 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
765      final alternative within the brackets, we would return the result of a      final alternative within the brackets, we would return the result of a
766      recursive call to match() whatever happened. We can reduce stack usage by      recursive call to match() whatever happened. We can reduce stack usage by
767      turning this into a tail recursion. */      turning this into a tail recursion, except in the case when match_cbegroup
768        is set.*/
769    
770      case OP_BRA:      case OP_BRA:
771      case OP_SBRA:      case OP_SBRA:
# Line 687  for (;;) Line 773  for (;;)
773      flags = (op >= OP_SBRA)? match_cbegroup : 0;      flags = (op >= OP_SBRA)? match_cbegroup : 0;
774      for (;;)      for (;;)
775        {        {
776        if (ecode[GET(ecode, 1)] != OP_ALT)        if (ecode[GET(ecode, 1)] != OP_ALT)   /* Final alternative */
777          {          {
778          ecode += _pcre_OP_lengths[*ecode];          if (flags == 0)    /* Not a possibly empty group */
779          flags |= match_tail_recursed;            {
780          DPRINTF(("bracket 0 tail recursion\n"));            ecode += _pcre_OP_lengths[*ecode];
781          goto TAIL_RECURSE;            DPRINTF(("bracket 0 tail recursion\n"));
782              goto TAIL_RECURSE;
783              }
784    
785            /* Possibly empty group; can't use tail recursion. */
786    
787            RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md, ims,
788              eptrb, flags, RM48);
789            RRETURN(rrc);
790          }          }
791    
792        /* For non-final alternatives, continue the loop for a NOMATCH result;        /* For non-final alternatives, continue the loop for a NOMATCH result;
# Line 700  for (;;) Line 794  for (;;)
794    
795        RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md, ims,        RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md, ims,
796          eptrb, flags, RM2);          eptrb, flags, RM2);
797        if (rrc != MATCH_NOMATCH) RRETURN(rrc);        if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN) RRETURN(rrc);
798        ecode += GET(ecode, 1);        ecode += GET(ecode, 1);
799        }        }
800      /* Control never reaches here. */      /* Control never reaches here. */
# Line 713  for (;;) Line 807  for (;;)
807    
808      case OP_COND:      case OP_COND:
809      case OP_SCOND:      case OP_SCOND:
810      if (ecode[LINK_SIZE+1] == OP_RREF)         /* Recursion test */      codelink= GET(ecode, 1);
811    
812        /* Because of the way auto-callout works during compile, a callout item is
813        inserted between OP_COND and an assertion condition. */
814    
815        if (ecode[LINK_SIZE+1] == OP_CALLOUT)
816        {        {
817        offset = GET2(ecode, LINK_SIZE + 2);     /* Recursion group number*/        if (pcre_callout != NULL)
818        condition = md->recursive != NULL &&          {
819          (offset == RREF_ANY || offset == md->recursive->group_num);          pcre_callout_block cb;
820        ecode += condition? 3 : GET(ecode, 1);          cb.version          = 1;   /* Version 1 of the callout block */
821            cb.callout_number   = ecode[LINK_SIZE+2];
822            cb.offset_vector    = md->offset_vector;
823            cb.subject          = (PCRE_SPTR)md->start_subject;
824            cb.subject_length   = md->end_subject - md->start_subject;
825            cb.start_match      = mstart - md->start_subject;
826            cb.current_position = eptr - md->start_subject;
827            cb.pattern_position = GET(ecode, LINK_SIZE + 3);
828            cb.next_item_length = GET(ecode, 3 + 2*LINK_SIZE);
829            cb.capture_top      = offset_top/2;
830            cb.capture_last     = md->capture_last;
831            cb.callout_data     = md->callout_data;
832            if ((rrc = (*pcre_callout)(&cb)) > 0) RRETURN(MATCH_NOMATCH);
833            if (rrc < 0) RRETURN(rrc);
834            }
835          ecode += _pcre_OP_lengths[OP_CALLOUT];
836        }        }
837    
838      else if (ecode[LINK_SIZE+1] == OP_CREF)    /* Group used test */      condcode = ecode[LINK_SIZE+1];
839    
840        /* Now see what the actual condition is */
841    
842        if (condcode == OP_RREF || condcode == OP_NRREF)    /* Recursion test */
843          {
844          if (md->recursive == NULL)                /* Not recursing => FALSE */
845            {
846            condition = FALSE;
847            ecode += GET(ecode, 1);
848            }
849          else
850            {
851            int recno = GET2(ecode, LINK_SIZE + 2);   /* Recursion group number*/
852            condition =  (recno == RREF_ANY || recno == md->recursive->group_num);
853    
854            /* If the test is for recursion into a specific subpattern, and it is
855            false, but the test was set up by name, scan the table to see if the
856            name refers to any other numbers, and test them. The condition is true
857            if any one is set. */
858    
859            if (!condition && condcode == OP_NRREF && recno != RREF_ANY)
860              {
861              uschar *slotA = md->name_table;
862              for (i = 0; i < md->name_count; i++)
863                {
864                if (GET2(slotA, 0) == recno) break;
865                slotA += md->name_entry_size;
866                }
867    
868              /* Found a name for the number - there can be only one; duplicate
869              names for different numbers are allowed, but not vice versa. First
870              scan down for duplicates. */
871    
872              if (i < md->name_count)
873                {
874                uschar *slotB = slotA;
875                while (slotB > md->name_table)
876                  {
877                  slotB -= md->name_entry_size;
878                  if (strcmp((char *)slotA + 2, (char *)slotB + 2) == 0)
879                    {
880                    condition = GET2(slotB, 0) == md->recursive->group_num;
881                    if (condition) break;
882                    }
883                  else break;
884                  }
885    
886                /* Scan up for duplicates */
887    
888                if (!condition)
889                  {
890                  slotB = slotA;
891                  for (i++; i < md->name_count; i++)
892                    {
893                    slotB += md->name_entry_size;
894                    if (strcmp((char *)slotA + 2, (char *)slotB + 2) == 0)
895                      {
896                      condition = GET2(slotB, 0) == md->recursive->group_num;
897                      if (condition) break;
898                      }
899                    else break;
900                    }
901                  }
902                }
903              }
904    
905            /* Chose branch according to the condition */
906    
907            ecode += condition? 3 : GET(ecode, 1);
908            }
909          }
910    
911        else if (condcode == OP_CREF || condcode == OP_NCREF)  /* Group used test */
912        {        {
913        offset = GET2(ecode, LINK_SIZE+2) << 1;  /* Doubled ref number */        offset = GET2(ecode, LINK_SIZE+2) << 1;  /* Doubled ref number */
914        condition = offset < offset_top && md->offset_vector[offset] >= 0;        condition = offset < offset_top && md->offset_vector[offset] >= 0;
915    
916          /* If the numbered capture is unset, but the reference was by name,
917          scan the table to see if the name refers to any other numbers, and test
918          them. The condition is true if any one is set. This is tediously similar
919          to the code above, but not close enough to try to amalgamate. */
920    
921          if (!condition && condcode == OP_NCREF)
922            {
923            int refno = offset >> 1;
924            uschar *slotA = md->name_table;
925    
926            for (i = 0; i < md->name_count; i++)
927              {
928              if (GET2(slotA, 0) == refno) break;
929              slotA += md->name_entry_size;
930              }
931    
932            /* Found a name for the number - there can be only one; duplicate names
933            for different numbers are allowed, but not vice versa. First scan down
934            for duplicates. */
935    
936            if (i < md->name_count)
937              {
938              uschar *slotB = slotA;
939              while (slotB > md->name_table)
940                {
941                slotB -= md->name_entry_size;
942                if (strcmp((char *)slotA + 2, (char *)slotB + 2) == 0)
943                  {
944                  offset = GET2(slotB, 0) << 1;
945                  condition = offset < offset_top &&
946                    md->offset_vector[offset] >= 0;
947                  if (condition) break;
948                  }
949                else break;
950                }
951    
952              /* Scan up for duplicates */
953    
954              if (!condition)
955                {
956                slotB = slotA;
957                for (i++; i < md->name_count; i++)
958                  {
959                  slotB += md->name_entry_size;
960                  if (strcmp((char *)slotA + 2, (char *)slotB + 2) == 0)
961                    {
962                    offset = GET2(slotB, 0) << 1;
963                    condition = offset < offset_top &&
964                      md->offset_vector[offset] >= 0;
965                    if (condition) break;
966                    }
967                  else break;
968                  }
969                }
970              }
971            }
972    
973          /* Chose branch according to the condition */
974    
975        ecode += condition? 3 : GET(ecode, 1);        ecode += condition? 3 : GET(ecode, 1);
976        }        }
977    
978      else if (ecode[LINK_SIZE+1] == OP_DEF)     /* DEFINE - always false */      else if (condcode == OP_DEF)     /* DEFINE - always false */
979        {        {
980        condition = FALSE;        condition = FALSE;
981        ecode += GET(ecode, 1);        ecode += GET(ecode, 1);
# Line 748  for (;;) Line 995  for (;;)
995          ecode += 1 + LINK_SIZE + GET(ecode, LINK_SIZE + 2);          ecode += 1 + LINK_SIZE + GET(ecode, LINK_SIZE + 2);
996          while (*ecode == OP_ALT) ecode += GET(ecode, 1);          while (*ecode == OP_ALT) ecode += GET(ecode, 1);
997          }          }
998        else if (rrc != MATCH_NOMATCH)        else if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN)
999          {          {
1000          RRETURN(rrc);         /* Need braces because of following else */          RRETURN(rrc);         /* Need braces because of following else */
1001          }          }
1002        else        else
1003          {          {
1004          condition = FALSE;          condition = FALSE;
1005          ecode += GET(ecode, 1);          ecode += codelink;
1006          }          }
1007        }        }
1008    
1009      /* 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,
1010      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
1011      alternative doesn't exist, we can just plough on. */      match_cbegroup is required for an unlimited repeat of a possibly empty
1012        group. If the second alternative doesn't exist, we can just plough on. */
1013    
1014      if (condition || *ecode == OP_ALT)      if (condition || *ecode == OP_ALT)
1015        {        {
1016        ecode += 1 + LINK_SIZE;        ecode += 1 + LINK_SIZE;
1017        flags = match_tail_recursed | ((op == OP_SCOND)? match_cbegroup : 0);        if (op == OP_SCOND)        /* Possibly empty group */
1018        goto TAIL_RECURSE;          {
1019            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, match_cbegroup, RM49);
1020            RRETURN(rrc);
1021            }
1022          else                       /* Group must match something */
1023            {
1024            flags = 0;
1025            goto TAIL_RECURSE;
1026            }
1027        }        }
1028      else      else                         /* Condition false & no alternative */
1029        {        {
1030        ecode += 1 + LINK_SIZE;        ecode += 1 + LINK_SIZE;
1031        }        }
1032      break;      break;
1033    
1034    
1035        /* Before OP_ACCEPT there may be any number of OP_CLOSE opcodes,
1036        to close any currently open capturing brackets. */
1037    
1038        case OP_CLOSE:
1039        number = GET2(ecode, 1);
1040        offset = number << 1;
1041    
1042    #ifdef DEBUG
1043          printf("end bracket %d at *ACCEPT", number);
1044          printf("\n");
1045    #endif
1046    
1047        md->capture_last = number;
1048        if (offset >= md->offset_max) md->offset_overflow = TRUE; else
1049          {
1050          md->offset_vector[offset] =
1051            md->offset_vector[md->offset_end - number];
1052          md->offset_vector[offset+1] = eptr - md->start_subject;
1053          if (offset_top <= offset) offset_top = offset + 2;
1054          }
1055        ecode += 3;
1056        break;
1057    
1058    
1059      /* 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
1060      restore the offsets appropriately and continue from after the call. */      recursion, we should restore the offsets appropriately and continue from
1061        after the call. */
1062    
1063        case OP_ACCEPT:
1064      case OP_END:      case OP_END:
1065      if (md->recursive != NULL && md->recursive->group_num == 0)      if (md->recursive != NULL && md->recursive->group_num == 0)
1066        {        {
# Line 787  for (;;) Line 1069  for (;;)
1069        md->recursive = rec->prevrec;        md->recursive = rec->prevrec;
1070        memmove(md->offset_vector, rec->offset_save,        memmove(md->offset_vector, rec->offset_save,
1071          rec->saved_max * sizeof(int));          rec->saved_max * sizeof(int));
1072        md->start_match = rec->save_start;        offset_top = rec->offset_top;
1073          mstart = rec->save_start;
1074        ims = original_ims;        ims = original_ims;
1075        ecode = rec->after_call;        ecode = rec->after_call;
1076        break;        break;
1077        }        }
1078    
1079      /* Otherwise, if PCRE_NOTEMPTY is set, fail if we have matched an empty      /* Otherwise, if we have matched an empty string, fail if PCRE_NOTEMPTY is
1080      string - backtracking will then try other alternatives, if any. */      set, or if PCRE_NOTEMPTY_ATSTART is set and we have matched at the start of
1081        the subject. In both cases, backtracking will then try other alternatives,
1082        if any. */
1083    
1084        if (eptr == mstart &&
1085            (md->notempty ||
1086              (md->notempty_atstart &&
1087                mstart == md->start_subject + md->start_offset)))
1088          RRETURN(MATCH_NOMATCH);
1089    
1090        /* Otherwise, we have a match. */
1091    
1092      if (md->notempty && eptr == md->start_match) RRETURN(MATCH_NOMATCH);      md->end_match_ptr = eptr;           /* Record where we ended */
1093      md->end_match_ptr = eptr;          /* Record where we ended */      md->end_offset_top = offset_top;    /* and how many extracts were taken */
1094      md->end_offset_top = offset_top;   /* and how many extracts were taken */      md->start_match_ptr = mstart;       /* and the start (\K can modify) */
1095      RRETURN(MATCH_MATCH);      RRETURN(MATCH_MATCH);
1096    
1097      /* Change option settings */      /* Change option settings */
# Line 822  for (;;) Line 1115  for (;;)
1115        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, ims, NULL, 0,        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, ims, NULL, 0,
1116          RM4);          RM4);
1117        if (rrc == MATCH_MATCH) break;        if (rrc == MATCH_MATCH) break;
1118        if (rrc != MATCH_NOMATCH) RRETURN(rrc);        if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN) RRETURN(rrc);
1119        ecode += GET(ecode, 1);        ecode += GET(ecode, 1);
1120        }        }
1121      while (*ecode == OP_ALT);      while (*ecode == OP_ALT);
# Line 849  for (;;) Line 1142  for (;;)
1142        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, ims, NULL, 0,        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, ims, NULL, 0,
1143          RM5);          RM5);
1144        if (rrc == MATCH_MATCH) RRETURN(MATCH_NOMATCH);        if (rrc == MATCH_MATCH) RRETURN(MATCH_NOMATCH);
1145        if (rrc != MATCH_NOMATCH) RRETURN(rrc);        if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN) RRETURN(rrc);
1146        ecode += GET(ecode,1);        ecode += GET(ecode,1);
1147        }        }
1148      while (*ecode == OP_ALT);      while (*ecode == OP_ALT);
# Line 873  for (;;) Line 1166  for (;;)
1166          {          {
1167          eptr--;          eptr--;
1168          if (eptr < md->start_subject) RRETURN(MATCH_NOMATCH);          if (eptr < md->start_subject) RRETURN(MATCH_NOMATCH);
1169          BACKCHAR(eptr)          BACKCHAR(eptr);
1170          }          }
1171        }        }
1172      else      else
# Line 886  for (;;) Line 1179  for (;;)
1179        if (eptr < md->start_subject) RRETURN(MATCH_NOMATCH);        if (eptr < md->start_subject) RRETURN(MATCH_NOMATCH);
1180        }        }
1181    
1182      /* Skip to next op code */      /* Save the earliest consulted character, then skip to next op code */
1183    
1184        if (eptr < md->start_used_ptr) md->start_used_ptr = eptr;
1185      ecode += 1 + LINK_SIZE;      ecode += 1 + LINK_SIZE;
1186      break;      break;
1187    
# Line 904  for (;;) Line 1198  for (;;)
1198        cb.offset_vector    = md->offset_vector;        cb.offset_vector    = md->offset_vector;
1199        cb.subject          = (PCRE_SPTR)md->start_subject;        cb.subject          = (PCRE_SPTR)md->start_subject;
1200        cb.subject_length   = md->end_subject - md->start_subject;        cb.subject_length   = md->end_subject - md->start_subject;
1201        cb.start_match      = md->start_match - md->start_subject;        cb.start_match      = mstart - md->start_subject;
1202        cb.current_position = eptr - md->start_subject;        cb.current_position = eptr - md->start_subject;
1203        cb.pattern_position = GET(ecode, 2);        cb.pattern_position = GET(ecode, 2);
1204        cb.next_item_length = GET(ecode, 2 + LINK_SIZE);        cb.next_item_length = GET(ecode, 2 + LINK_SIZE);
# Line 966  for (;;) Line 1260  for (;;)
1260    
1261        memcpy(new_recursive.offset_save, md->offset_vector,        memcpy(new_recursive.offset_save, md->offset_vector,
1262              new_recursive.saved_max * sizeof(int));              new_recursive.saved_max * sizeof(int));
1263        new_recursive.save_start = md->start_match;        new_recursive.save_start = mstart;
1264        md->start_match = eptr;        new_recursive.offset_top = offset_top;
1265          mstart = eptr;
1266    
1267        /* 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
1268        restore the offset and recursion data. */        restore the offset and recursion data. */
# Line 986  for (;;) Line 1281  for (;;)
1281              (pcre_free)(new_recursive.offset_save);              (pcre_free)(new_recursive.offset_save);
1282            RRETURN(MATCH_MATCH);            RRETURN(MATCH_MATCH);
1283            }            }
1284          else if (rrc != MATCH_NOMATCH)          else if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN)
1285            {            {
1286            DPRINTF(("Recursion gave error %d\n", rrc));            DPRINTF(("Recursion gave error %d\n", rrc));
1287              if (new_recursive.offset_save != stacksave)
1288                (pcre_free)(new_recursive.offset_save);
1289            RRETURN(rrc);            RRETURN(rrc);
1290            }            }
1291    
# Line 1020  for (;;) Line 1317  for (;;)
1317    
1318      do      do
1319        {        {
1320        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, ims,        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, ims, eptrb, 0, RM7);
         eptrb, 0, RM7);  
1321        if (rrc == MATCH_MATCH) break;        if (rrc == MATCH_MATCH) break;
1322        if (rrc != MATCH_NOMATCH) RRETURN(rrc);        if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN) RRETURN(rrc);
1323        ecode += GET(ecode,1);        ecode += GET(ecode,1);
1324        }        }
1325      while (*ecode == OP_ALT);      while (*ecode == OP_ALT);
# Line 1066  for (;;) Line 1362  for (;;)
1362    
1363      if (*ecode == OP_KETRMIN)      if (*ecode == OP_KETRMIN)
1364        {        {
1365        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, ims, eptrb, 0,        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, ims, eptrb, 0, RM8);
         RM8);  
1366        if (rrc != MATCH_NOMATCH) RRETURN(rrc);        if (rrc != MATCH_NOMATCH) RRETURN(rrc);
1367        ecode = prev;        ecode = prev;
1368        flags = match_tail_recursed;        flags = 0;
1369        goto TAIL_RECURSE;        goto TAIL_RECURSE;
1370        }        }
1371      else  /* OP_KETRMAX */      else  /* OP_KETRMAX */
# Line 1078  for (;;) Line 1373  for (;;)
1373        RMATCH(eptr, prev, offset_top, md, ims, eptrb, match_cbegroup, RM9);        RMATCH(eptr, prev, offset_top, md, ims, eptrb, match_cbegroup, RM9);
1374        if (rrc != MATCH_NOMATCH) RRETURN(rrc);        if (rrc != MATCH_NOMATCH) RRETURN(rrc);
1375        ecode += 1 + LINK_SIZE;        ecode += 1 + LINK_SIZE;
1376        flags = match_tail_recursed;        flags = 0;
1377        goto TAIL_RECURSE;        goto TAIL_RECURSE;
1378        }        }
1379      /* Control never gets here */      /* Control never gets here */
# Line 1090  for (;;) Line 1385  for (;;)
1385      do ecode += GET(ecode,1); while (*ecode == OP_ALT);      do ecode += GET(ecode,1); while (*ecode == OP_ALT);
1386      break;      break;
1387    
1388      /* BRAZERO and BRAMINZERO occur just before a bracket group, indicating      /* BRAZERO, BRAMINZERO and SKIPZERO occur just before a bracket group,
1389      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
1390      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
1391      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
1392      preceded by BRAZERO or BRAMINZERO. */      optional ones preceded by BRAZERO or BRAMINZERO. */
1393    
1394      case OP_BRAZERO:      case OP_BRAZERO:
1395        {        {
# Line 1116  for (;;) Line 1411  for (;;)
1411        }        }
1412      break;      break;
1413    
1414        case OP_SKIPZERO:
1415          {
1416          next = ecode+1;
1417          do next += GET(next,1); while (*next == OP_ALT);
1418          ecode = next + 1 + LINK_SIZE;
1419          }
1420        break;
1421    
1422      /* End of a group, repeated or non-repeating. */      /* End of a group, repeated or non-repeating. */
1423    
1424      case OP_KET:      case OP_KET:
# Line 1157  for (;;) Line 1460  for (;;)
1460        {        {
1461        number = GET2(prev, 1+LINK_SIZE);        number = GET2(prev, 1+LINK_SIZE);
1462        offset = number << 1;        offset = number << 1;
1463    
1464  #ifdef DEBUG  #ifdef DEBUG
1465        printf("end bracket %d", number);        printf("end bracket %d", number);
1466        printf("\n");        printf("\n");
# Line 1180  for (;;) Line 1483  for (;;)
1483          recursion_info *rec = md->recursive;          recursion_info *rec = md->recursive;
1484          DPRINTF(("Recursion (%d) succeeded - continuing\n", number));          DPRINTF(("Recursion (%d) succeeded - continuing\n", number));
1485          md->recursive = rec->prevrec;          md->recursive = rec->prevrec;
1486          md->start_match = rec->save_start;          mstart = rec->save_start;
1487          memcpy(md->offset_vector, rec->offset_save,          memcpy(md->offset_vector, rec->offset_save,
1488            rec->saved_max * sizeof(int));            rec->saved_max * sizeof(int));
1489            offset_top = rec->offset_top;
1490          ecode = rec->after_call;          ecode = rec->after_call;
1491          ims = original_ims;          ims = original_ims;
1492          break;          break;
# Line 1209  for (;;) Line 1513  for (;;)
1513    
1514      /* 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
1515      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
1516      tail recursion to avoid using another stack frame. */      tail recursion to avoid using another stack frame, unless we have an
1517        unlimited repeat of a group that can match an empty string. */
1518    
1519      flags = (*prev >= OP_SBRA)? match_cbegroup : 0;      flags = (*prev >= OP_SBRA)? match_cbegroup : 0;
1520    
1521      if (*ecode == OP_KETRMIN)      if (*ecode == OP_KETRMIN)
1522        {        {
1523        RMATCH(eptr, ecode + 1+LINK_SIZE, offset_top, md, ims, eptrb, 0,        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, ims, eptrb, 0, RM12);
         RM12);  
1524        if (rrc != MATCH_NOMATCH) RRETURN(rrc);        if (rrc != MATCH_NOMATCH) RRETURN(rrc);
1525          if (flags != 0)    /* Could match an empty string */
1526            {
1527            RMATCH(eptr, prev, offset_top, md, ims, eptrb, flags, RM50);
1528            RRETURN(rrc);
1529            }
1530        ecode = prev;        ecode = prev;
       flags |= match_tail_recursed;  
1531        goto TAIL_RECURSE;        goto TAIL_RECURSE;
1532        }        }
1533      else  /* OP_KETRMAX */      else  /* OP_KETRMAX */
# Line 1227  for (;;) Line 1535  for (;;)
1535        RMATCH(eptr, prev, offset_top, md, ims, eptrb, flags, RM13);        RMATCH(eptr, prev, offset_top, md, ims, eptrb, flags, RM13);
1536        if (rrc != MATCH_NOMATCH) RRETURN(rrc);        if (rrc != MATCH_NOMATCH) RRETURN(rrc);
1537        ecode += 1 + LINK_SIZE;        ecode += 1 + LINK_SIZE;
1538        flags = match_tail_recursed;        flags = 0;
1539        goto TAIL_RECURSE;        goto TAIL_RECURSE;
1540        }        }
1541      /* Control never gets here */      /* Control never gets here */
# Line 1260  for (;;) Line 1568  for (;;)
1568      ecode++;      ecode++;
1569      break;      break;
1570    
1571        /* Reset the start of match point */
1572    
1573        case OP_SET_SOM:
1574        mstart = eptr;
1575        ecode++;
1576        break;
1577    
1578      /* Assert before internal newline if multiline, or before a terminating      /* Assert before internal newline if multiline, or before a terminating
1579      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. */
1580    
# Line 1311  for (;;) Line 1626  for (;;)
1626    
1627        /* Find out if the previous and current characters are "word" characters.        /* Find out if the previous and current characters are "word" characters.
1628        It takes a bit more work in UTF-8 mode. Characters > 255 are assumed to        It takes a bit more work in UTF-8 mode. Characters > 255 are assumed to
1629        be "non-word" characters. */        be "non-word" characters. Remember the earliest consulted character for
1630          partial matching. */
1631    
1632  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF8
1633        if (utf8)        if (utf8)
1634          {          {
1635          if (eptr == md->start_subject) prev_is_word = FALSE; else          if (eptr == md->start_subject) prev_is_word = FALSE; else
1636            {            {
1637            const uschar *lastptr = eptr - 1;            USPTR lastptr = eptr - 1;
1638            while((*lastptr & 0xc0) == 0x80) lastptr--;            while((*lastptr & 0xc0) == 0x80) lastptr--;
1639              if (lastptr < md->start_used_ptr) md->start_used_ptr = lastptr;
1640            GETCHAR(c, lastptr);            GETCHAR(c, lastptr);
1641            prev_is_word = c < 256 && (md->ctypes[c] & ctype_word) != 0;            prev_is_word = c < 256 && (md->ctypes[c] & ctype_word) != 0;
1642            }            }
1643          if (eptr >= md->end_subject) cur_is_word = FALSE; else          if (eptr >= md->end_subject)
1644              {
1645              SCHECK_PARTIAL();
1646              cur_is_word = FALSE;
1647              }
1648            else
1649            {            {
1650            GETCHAR(c, eptr);            GETCHAR(c, eptr);
1651            cur_is_word = c < 256 && (md->ctypes[c] & ctype_word) != 0;            cur_is_word = c < 256 && (md->ctypes[c] & ctype_word) != 0;
# Line 1332  for (;;) Line 1654  for (;;)
1654        else        else
1655  #endif  #endif
1656    
1657        /* More streamlined when not in UTF-8 mode */        /* Not in UTF-8 mode */
1658    
1659          {          {
1660          prev_is_word = (eptr != md->start_subject) &&          if (eptr == md->start_subject) prev_is_word = FALSE; else
1661            ((md->ctypes[eptr[-1]] & ctype_word) != 0);            {
1662          cur_is_word = (eptr < md->end_subject) &&            if (eptr <= md->start_used_ptr) md->start_used_ptr = eptr - 1;
1663            ((md->ctypes[*eptr] & ctype_word) != 0);            prev_is_word = ((md->ctypes[eptr[-1]] & ctype_word) != 0);
1664              }
1665            if (eptr >= md->end_subject)
1666              {
1667              SCHECK_PARTIAL();
1668              cur_is_word = FALSE;
1669              }
1670            else cur_is_word = ((md->ctypes[*eptr] & ctype_word) != 0);
1671          }          }
1672    
1673        /* Now see if the situation is what we want */        /* Now see if the situation is what we want */
# Line 1352  for (;;) Line 1681  for (;;)
1681      /* Match a single character type; inline for speed */      /* Match a single character type; inline for speed */
1682    
1683      case OP_ANY:      case OP_ANY:
1684      if ((ims & PCRE_DOTALL) == 0)      if (IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH);
1685        /* Fall through */
1686    
1687        case OP_ALLANY:
1688        if (eptr++ >= md->end_subject)
1689        {        {
1690        if (IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH);        SCHECK_PARTIAL();
1691          RRETURN(MATCH_NOMATCH);
1692        }        }
1693      if (eptr++ >= md->end_subject) RRETURN(MATCH_NOMATCH);      if (utf8) while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;
     if (utf8)  
       while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;  
1694      ecode++;      ecode++;
1695      break;      break;
1696    
# Line 1366  for (;;) Line 1698  for (;;)
1698      any byte, even newline, independent of the setting of PCRE_DOTALL. */      any byte, even newline, independent of the setting of PCRE_DOTALL. */
1699    
1700      case OP_ANYBYTE:      case OP_ANYBYTE:
1701      if (eptr++ >= md->end_subject) RRETURN(MATCH_NOMATCH);      if (eptr++ >= md->end_subject)
1702          {
1703          SCHECK_PARTIAL();
1704          RRETURN(MATCH_NOMATCH);
1705          }
1706      ecode++;      ecode++;
1707      break;      break;
1708    
1709      case OP_NOT_DIGIT:      case OP_NOT_DIGIT:
1710      if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);      if (eptr >= md->end_subject)
1711          {
1712          SCHECK_PARTIAL();
1713          RRETURN(MATCH_NOMATCH);
1714          }
1715      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
1716      if (      if (
1717  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF8
# Line 1384  for (;;) Line 1724  for (;;)
1724      break;      break;
1725    
1726      case OP_DIGIT:      case OP_DIGIT:
1727      if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);      if (eptr >= md->end_subject)
1728          {
1729          SCHECK_PARTIAL();
1730          RRETURN(MATCH_NOMATCH);
1731          }
1732      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
1733      if (      if (
1734  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF8
# Line 1397  for (;;) Line 1741  for (;;)
1741      break;      break;
1742    
1743      case OP_NOT_WHITESPACE:      case OP_NOT_WHITESPACE:
1744      if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);      if (eptr >= md->end_subject)
1745          {
1746          SCHECK_PARTIAL();
1747          RRETURN(MATCH_NOMATCH);
1748          }
1749      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
1750      if (      if (
1751  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF8
# Line 1410  for (;;) Line 1758  for (;;)
1758      break;      break;
1759    
1760      case OP_WHITESPACE:      case OP_WHITESPACE:
1761      if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);      if (eptr >= md->end_subject)
1762          {
1763          SCHECK_PARTIAL();
1764          RRETURN(MATCH_NOMATCH);
1765          }
1766      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
1767      if (      if (
1768  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF8
# Line 1423  for (;;) Line 1775  for (;;)
1775      break;      break;
1776    
1777      case OP_NOT_WORDCHAR:      case OP_NOT_WORDCHAR:
1778      if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);      if (eptr >= md->end_subject)
1779          {
1780          SCHECK_PARTIAL();
1781          RRETURN(MATCH_NOMATCH);
1782          }
1783      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
1784      if (      if (
1785  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF8
# Line 1436  for (;;) Line 1792  for (;;)
1792      break;      break;
1793    
1794      case OP_WORDCHAR:      case OP_WORDCHAR:
1795      if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);      if (eptr >= md->end_subject)
1796          {
1797          SCHECK_PARTIAL();
1798          RRETURN(MATCH_NOMATCH);
1799          }
1800      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
1801      if (      if (
1802  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF8
# Line 1449  for (;;) Line 1809  for (;;)
1809      break;      break;
1810    
1811      case OP_ANYNL:      case OP_ANYNL:
1812      if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);      if (eptr >= md->end_subject)
1813          {
1814          SCHECK_PARTIAL();
1815          RRETURN(MATCH_NOMATCH);
1816          }
1817      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
1818      switch(c)      switch(c)
1819        {        {
# Line 1457  for (;;) Line 1821  for (;;)
1821        case 0x000d:        case 0x000d:
1822        if (eptr < md->end_subject && *eptr == 0x0a) eptr++;        if (eptr < md->end_subject && *eptr == 0x0a) eptr++;
1823        break;        break;
1824    
1825        case 0x000a:        case 0x000a:
1826          break;
1827    
1828        case 0x000b:        case 0x000b:
1829        case 0x000c:        case 0x000c:
1830        case 0x0085:        case 0x0085:
1831        case 0x2028:        case 0x2028:
1832        case 0x2029:        case 0x2029:
1833          if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);
1834          break;
1835          }
1836        ecode++;
1837        break;
1838    
1839        case OP_NOT_HSPACE:
1840        if (eptr >= md->end_subject)
1841          {
1842          SCHECK_PARTIAL();
1843          RRETURN(MATCH_NOMATCH);
1844          }
1845        GETCHARINCTEST(c, eptr);
1846        switch(c)
1847          {
1848          default: break;
1849          case 0x09:      /* HT */
1850          case 0x20:      /* SPACE */
1851          case 0xa0:      /* NBSP */
1852          case 0x1680:    /* OGHAM SPACE MARK */
1853          case 0x180e:    /* MONGOLIAN VOWEL SEPARATOR */
1854          case 0x2000:    /* EN QUAD */
1855          case 0x2001:    /* EM QUAD */
1856          case 0x2002:    /* EN SPACE */
1857          case 0x2003:    /* EM SPACE */
1858          case 0x2004:    /* THREE-PER-EM SPACE */
1859          case 0x2005:    /* FOUR-PER-EM SPACE */
1860          case 0x2006:    /* SIX-PER-EM SPACE */
1861          case 0x2007:    /* FIGURE SPACE */
1862          case 0x2008:    /* PUNCTUATION SPACE */
1863          case 0x2009:    /* THIN SPACE */
1864          case 0x200A:    /* HAIR SPACE */
1865          case 0x202f:    /* NARROW NO-BREAK SPACE */
1866          case 0x205f:    /* MEDIUM MATHEMATICAL SPACE */
1867          case 0x3000:    /* IDEOGRAPHIC SPACE */
1868          RRETURN(MATCH_NOMATCH);
1869          }
1870        ecode++;
1871        break;
1872    
1873        case OP_HSPACE:
1874        if (eptr >= md->end_subject)
1875          {
1876          SCHECK_PARTIAL();
1877          RRETURN(MATCH_NOMATCH);
1878          }
1879        GETCHARINCTEST(c, eptr);
1880        switch(c)
1881          {
1882          default: RRETURN(MATCH_NOMATCH);
1883          case 0x09:      /* HT */
1884          case 0x20:      /* SPACE */
1885          case 0xa0:      /* NBSP */
1886          case 0x1680:    /* OGHAM SPACE MARK */
1887          case 0x180e:    /* MONGOLIAN VOWEL SEPARATOR */
1888          case 0x2000:    /* EN QUAD */
1889          case 0x2001:    /* EM QUAD */
1890          case 0x2002:    /* EN SPACE */
1891          case 0x2003:    /* EM SPACE */
1892          case 0x2004:    /* THREE-PER-EM SPACE */
1893          case 0x2005:    /* FOUR-PER-EM SPACE */
1894          case 0x2006:    /* SIX-PER-EM SPACE */
1895          case 0x2007:    /* FIGURE SPACE */
1896          case 0x2008:    /* PUNCTUATION SPACE */
1897          case 0x2009:    /* THIN SPACE */
1898          case 0x200A:    /* HAIR SPACE */
1899          case 0x202f:    /* NARROW NO-BREAK SPACE */
1900          case 0x205f:    /* MEDIUM MATHEMATICAL SPACE */
1901          case 0x3000:    /* IDEOGRAPHIC SPACE */
1902          break;
1903          }
1904        ecode++;
1905        break;
1906    
1907        case OP_NOT_VSPACE:
1908        if (eptr >= md->end_subject)
1909          {
1910          SCHECK_PARTIAL();
1911          RRETURN(MATCH_NOMATCH);
1912          }
1913        GETCHARINCTEST(c, eptr);
1914        switch(c)
1915          {
1916          default: break;
1917          case 0x0a:      /* LF */
1918          case 0x0b:      /* VT */
1919          case 0x0c:      /* FF */
1920          case 0x0d:      /* CR */
1921          case 0x85:      /* NEL */
1922          case 0x2028:    /* LINE SEPARATOR */
1923          case 0x2029:    /* PARAGRAPH SEPARATOR */
1924          RRETURN(MATCH_NOMATCH);
1925          }
1926        ecode++;
1927        break;
1928    
1929        case OP_VSPACE:
1930        if (eptr >= md->end_subject)
1931          {
1932          SCHECK_PARTIAL();
1933          RRETURN(MATCH_NOMATCH);
1934          }
1935        GETCHARINCTEST(c, eptr);
1936        switch(c)
1937          {
1938          default: RRETURN(MATCH_NOMATCH);
1939          case 0x0a:      /* LF */
1940          case 0x0b:      /* VT */
1941          case 0x0c:      /* FF */
1942          case 0x0d:      /* CR */
1943          case 0x85:      /* NEL */
1944          case 0x2028:    /* LINE SEPARATOR */
1945          case 0x2029:    /* PARAGRAPH SEPARATOR */
1946        break;        break;
1947        }        }
1948      ecode++;      ecode++;
# Line 1474  for (;;) Line 1954  for (;;)
1954    
1955      case OP_PROP:      case OP_PROP:
1956      case OP_NOTPROP:      case OP_NOTPROP:
1957      if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);      if (eptr >= md->end_subject)
1958          {
1959          SCHECK_PARTIAL();
1960          RRETURN(MATCH_NOMATCH);
1961          }
1962      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
1963        {        {
1964        int chartype, script;        const ucd_record *prop = GET_UCD(c);
       int category = _pcre_ucp_findprop(c, &chartype, &script);  
1965    
1966        switch(ecode[1])        switch(ecode[1])
1967          {          {
# Line 1487  for (;;) Line 1970  for (;;)
1970          break;          break;
1971    
1972          case PT_LAMP:          case PT_LAMP:
1973          if ((chartype == ucp_Lu ||          if ((prop->chartype == ucp_Lu ||
1974               chartype == ucp_Ll ||               prop->chartype == ucp_Ll ||
1975               chartype == ucp_Lt) == (op == OP_NOTPROP))               prop->chartype == ucp_Lt) == (op == OP_NOTPROP))
1976            RRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
1977           break;           break;
1978    
1979          case PT_GC:          case PT_GC:
1980          if ((ecode[2] != category) == (op == OP_PROP))          if ((ecode[2] != _pcre_ucp_gentype[prop->chartype]) == (op == OP_PROP))
1981            RRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
1982          break;          break;
1983    
1984          case PT_PC:          case PT_PC:
1985          if ((ecode[2] != chartype) == (op == OP_PROP))          if ((ecode[2] != prop->chartype) == (op == OP_PROP))
1986            RRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
1987          break;          break;
1988    
1989          case PT_SC:          case PT_SC:
1990          if ((ecode[2] != script) == (op == OP_PROP))          if ((ecode[2] != prop->script) == (op == OP_PROP))
1991            RRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
1992          break;          break;
1993    
# Line 1520  for (;;) Line 2003  for (;;)
2003      is in the binary; otherwise a compile-time error occurs. */      is in the binary; otherwise a compile-time error occurs. */
2004    
2005      case OP_EXTUNI:      case OP_EXTUNI:
2006      if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);      if (eptr >= md->end_subject)
2007          {
2008          SCHECK_PARTIAL();
2009          RRETURN(MATCH_NOMATCH);
2010          }
2011      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2012        {        {
2013        int chartype, script;        int category = UCD_CATEGORY(c);
       int category = _pcre_ucp_findprop(c, &chartype, &script);  
2014        if (category == ucp_M) RRETURN(MATCH_NOMATCH);        if (category == ucp_M) RRETURN(MATCH_NOMATCH);
2015        while (eptr < md->end_subject)        while (eptr < md->end_subject)
2016          {          {
# Line 1533  for (;;) Line 2019  for (;;)
2019            {            {
2020            GETCHARLEN(c, eptr, len);            GETCHARLEN(c, eptr, len);
2021            }            }
2022          category = _pcre_ucp_findprop(c, &chartype, &script);          category = UCD_CATEGORY(c);
2023          if (category != ucp_M) break;          if (category != ucp_M) break;
2024          eptr += len;          eptr += len;
2025          }          }
# Line 1554  for (;;) Line 2040  for (;;)
2040      case OP_REF:      case OP_REF:
2041        {        {
2042        offset = GET2(ecode, 1) << 1;               /* Doubled ref number */        offset = GET2(ecode, 1) << 1;               /* Doubled ref number */
2043        ecode += 3;                                 /* Advance past item */        ecode += 3;
2044    
2045          /* If the reference is unset, there are two possibilities:
2046    
2047          (a) In the default, Perl-compatible state, set the length to be longer
2048          than the amount of subject left; this ensures that every attempt at a
2049          match fails. We can't just fail here, because of the possibility of
2050          quantifiers with zero minima.
2051    
2052          (b) If the JavaScript compatibility flag is set, set the length to zero
2053          so that the back reference matches an empty string.
2054    
2055          Otherwise, set the length to the length of what was matched by the
2056          referenced subpattern. */
2057    
2058        /* If the reference is unset, set the length to be longer than the amount        if (offset >= offset_top || md->offset_vector[offset] < 0)
2059        of subject left; this ensures that every attempt at a match fails. We          length = (md->jscript_compat)? 0 : md->end_subject - eptr + 1;
2060        can't just fail here, because of the possibility of quantifiers with zero        else
2061        minima. */          length = md->offset_vector[offset+1] - md->offset_vector[offset];
   
       length = (offset >= offset_top || md->offset_vector[offset] < 0)?  
         md->end_subject - eptr + 1 :  
         md->offset_vector[offset+1] - md->offset_vector[offset];  
2062    
2063        /* Set up for repetition, or handle the non-repeated case */        /* Set up for repetition, or handle the non-repeated case */
2064    
# Line 1592  for (;;) Line 2087  for (;;)
2087          break;          break;
2088    
2089          default:               /* No repeat follows */          default:               /* No repeat follows */
2090          if (!match_ref(offset, eptr, length, md, ims)) RRETURN(MATCH_NOMATCH);          if (!match_ref(offset, eptr, length, md, ims))
2091              {
2092              CHECK_PARTIAL();
2093              RRETURN(MATCH_NOMATCH);
2094              }
2095          eptr += length;          eptr += length;
2096          continue;              /* With the main loop */          continue;              /* With the main loop */
2097          }          }
# Line 1608  for (;;) Line 2107  for (;;)
2107    
2108        for (i = 1; i <= min; i++)        for (i = 1; i <= min; i++)
2109          {          {
2110          if (!match_ref(offset, eptr, length, md, ims)) RRETURN(MATCH_NOMATCH);          if (!match_ref(offset, eptr, length, md, ims))
2111              {
2112              CHECK_PARTIAL();
2113              RRETURN(MATCH_NOMATCH);
2114              }
2115          eptr += length;          eptr += length;
2116          }          }
2117    
# Line 1625  for (;;) Line 2128  for (;;)
2128            {            {
2129            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM14);            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM14);
2130            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2131            if (fi >= max || !match_ref(offset, eptr, length, md, ims))            if (fi >= max) RRETURN(MATCH_NOMATCH);
2132              if (!match_ref(offset, eptr, length, md, ims))
2133                {
2134                CHECK_PARTIAL();
2135              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
2136                }
2137            eptr += length;            eptr += length;
2138            }            }
2139          /* Control never gets here */          /* Control never gets here */
# Line 1653  for (;;) Line 2160  for (;;)
2160        }        }
2161      /* Control never gets here */      /* Control never gets here */
2162    
   
   
2163      /* Match a bit-mapped character class, possibly repeatedly. This op code is      /* Match a bit-mapped character class, possibly repeatedly. This op code is
2164      used when all the characters in the class have values in the range 0-255,      used when all the characters in the class have values in the range 0-255,
2165      and either the matching is caseful, or the characters are in the range      and either the matching is caseful, or the characters are in the range
# Line 1709  for (;;) Line 2214  for (;;)
2214          {          {
2215          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
2216            {            {
2217            if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);            if (eptr >= md->end_subject)
2218                {
2219                SCHECK_PARTIAL();
2220                RRETURN(MATCH_NOMATCH);
2221                }
2222            GETCHARINC(c, eptr);            GETCHARINC(c, eptr);
2223            if (c > 255)            if (c > 255)
2224              {              {
# Line 1727  for (;;) Line 2236  for (;;)
2236          {          {
2237          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
2238            {            {
2239            if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);            if (eptr >= md->end_subject)
2240                {
2241                SCHECK_PARTIAL();
2242                RRETURN(MATCH_NOMATCH);
2243                }
2244            c = *eptr++;            c = *eptr++;
2245            if ((data[c/8] & (1 << (c&7))) == 0) RRETURN(MATCH_NOMATCH);            if ((data[c/8] & (1 << (c&7))) == 0) RRETURN(MATCH_NOMATCH);
2246            }            }
# Line 1751  for (;;) Line 2264  for (;;)
2264              {              {
2265              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM16);              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM16);
2266              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2267              if (fi >= max || eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);              if (fi >= max) RRETURN(MATCH_NOMATCH);
2268                if (eptr >= md->end_subject)
2269                  {
2270                  SCHECK_PARTIAL();
2271                  RRETURN(MATCH_NOMATCH);
2272                  }
2273              GETCHARINC(c, eptr);              GETCHARINC(c, eptr);
2274              if (c > 255)              if (c > 255)
2275                {                {
# Line 1771  for (;;) Line 2289  for (;;)
2289              {              {
2290              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM17);              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM17);
2291              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2292              if (fi >= max || eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);              if (fi >= max) RRETURN(MATCH_NOMATCH);
2293                if (eptr >= md->end_subject)
2294                  {
2295                  SCHECK_PARTIAL();
2296                  RRETURN(MATCH_NOMATCH);
2297                  }
2298              c = *eptr++;              c = *eptr++;
2299              if ((data[c/8] & (1 << (c&7))) == 0) RRETURN(MATCH_NOMATCH);              if ((data[c/8] & (1 << (c&7))) == 0) RRETURN(MATCH_NOMATCH);
2300              }              }
# Line 1838  for (;;) Line 2361  for (;;)
2361    
2362    
2363      /* Match an extended character class. This opcode is encountered only      /* Match an extended character class. This opcode is encountered only
2364      in UTF-8 mode, because that's the only time it is compiled. */      when UTF-8 mode mode is supported. Nevertheless, we may not be in UTF-8
2365        mode, because Unicode properties are supported in non-UTF-8 mode. */
2366    
2367  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF8
2368      case OP_XCLASS:      case OP_XCLASS:
# Line 1879  for (;;) Line 2403  for (;;)
2403    
2404        for (i = 1; i <= min; i++)        for (i = 1; i <= min; i++)
2405          {          {
2406          if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);          if (eptr >= md->end_subject)
2407          GETCHARINC(c, eptr);            {
2408              SCHECK_PARTIAL();
2409              RRETURN(MATCH_NOMATCH);
2410              }
2411            GETCHARINCTEST(c, eptr);
2412          if (!_pcre_xclass(c, data)) RRETURN(MATCH_NOMATCH);          if (!_pcre_xclass(c, data)) RRETURN(MATCH_NOMATCH);
2413          }          }
2414    
# Line 1898  for (;;) Line 2426  for (;;)
2426            {            {
2427            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM20);            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM20);
2428            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2429            if (fi >= max || eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);            if (fi >= max) RRETURN(MATCH_NOMATCH);
2430            GETCHARINC(c, eptr);            if (eptr >= md->end_subject)
2431                {
2432                SCHECK_PARTIAL();
2433                RRETURN(MATCH_NOMATCH);
2434                }
2435              GETCHARINCTEST(c, eptr);
2436            if (!_pcre_xclass(c, data)) RRETURN(MATCH_NOMATCH);            if (!_pcre_xclass(c, data)) RRETURN(MATCH_NOMATCH);
2437            }            }
2438          /* Control never gets here */          /* Control never gets here */
# Line 1914  for (;;) Line 2447  for (;;)
2447            {            {
2448            int len = 1;            int len = 1;
2449            if (eptr >= md->end_subject) break;            if (eptr >= md->end_subject) break;
2450            GETCHARLEN(c, eptr, len);            GETCHARLENTEST(c, eptr, len);
2451            if (!_pcre_xclass(c, data)) break;            if (!_pcre_xclass(c, data)) break;
2452            eptr += len;            eptr += len;
2453            }            }
# Line 1923  for (;;) Line 2456  for (;;)
2456            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM21);            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM21);
2457            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2458            if (eptr-- == pp) break;        /* Stop if tried at original pos */            if (eptr-- == pp) break;        /* Stop if tried at original pos */
2459            BACKCHAR(eptr)            if (utf8) BACKCHAR(eptr);
2460            }            }
2461          RRETURN(MATCH_NOMATCH);          RRETURN(MATCH_NOMATCH);
2462          }          }
# Line 1941  for (;;) Line 2474  for (;;)
2474        length = 1;        length = 1;
2475        ecode++;        ecode++;
2476        GETCHARLEN(fc, ecode, length);        GETCHARLEN(fc, ecode, length);
2477        if (length > md->end_subject - eptr) RRETURN(MATCH_NOMATCH);        if (length > md->end_subject - eptr)
2478            {
2479            CHECK_PARTIAL();             /* Not SCHECK_PARTIAL() */
2480            RRETURN(MATCH_NOMATCH);
2481            }
2482        while (length-- > 0) if (*ecode++ != *eptr++) RRETURN(MATCH_NOMATCH);        while (length-- > 0) if (*ecode++ != *eptr++) RRETURN(MATCH_NOMATCH);
2483        }        }
2484      else      else
# Line 1949  for (;;) Line 2486  for (;;)
2486    
2487      /* Non-UTF-8 mode */      /* Non-UTF-8 mode */
2488        {        {
2489        if (md->end_subject - eptr < 1) RRETURN(MATCH_NOMATCH);        if (md->end_subject - eptr < 1)
2490            {
2491            SCHECK_PARTIAL();            /* This one can use SCHECK_PARTIAL() */
2492            RRETURN(MATCH_NOMATCH);
2493            }
2494        if (ecode[1] != *eptr++) RRETURN(MATCH_NOMATCH);        if (ecode[1] != *eptr++) RRETURN(MATCH_NOMATCH);
2495        ecode += 2;        ecode += 2;
2496        }        }
# Line 1965  for (;;) Line 2506  for (;;)
2506        ecode++;        ecode++;
2507        GETCHARLEN(fc, ecode, length);        GETCHARLEN(fc, ecode, length);
2508    
2509        if (length > md->end_subject - eptr) RRETURN(MATCH_NOMATCH);        if (length > md->end_subject - eptr)
2510            {
2511            CHECK_PARTIAL();             /* Not SCHECK_PARTIAL() */
2512            RRETURN(MATCH_NOMATCH);
2513            }
2514    
2515        /* If the pattern character's value is < 128, we have only one byte, and        /* If the pattern character's value is < 128, we have only one byte, and
2516        can use the fast lookup table. */        can use the fast lookup table. */
# Line 1989  for (;;) Line 2534  for (;;)
2534          if (fc != dc)          if (fc != dc)
2535            {            {
2536  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
2537            if (dc != _pcre_ucp_othercase(fc))            if (dc != UCD_OTHERCASE(fc))
2538  #endif  #endif
2539              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
2540            }            }
# Line 2000  for (;;) Line 2545  for (;;)
2545    
2546      /* Non-UTF-8 mode */      /* Non-UTF-8 mode */
2547        {        {
2548        if (md->end_subject - eptr < 1) RRETURN(MATCH_NOMATCH);        if (md->end_subject - eptr < 1)
2549            {
2550            SCHECK_PARTIAL();            /* This one can use SCHECK_PARTIAL() */
2551            RRETURN(MATCH_NOMATCH);
2552            }
2553        if (md->lcc[ecode[1]] != md->lcc[*eptr++]) RRETURN(MATCH_NOMATCH);        if (md->lcc[ecode[1]] != md->lcc[*eptr++]) RRETURN(MATCH_NOMATCH);
2554        ecode += 2;        ecode += 2;
2555        }        }
# Line 2054  for (;;) Line 2603  for (;;)
2603      case OP_MINQUERY:      case OP_MINQUERY:
2604      c = *ecode++ - OP_STAR;      c = *ecode++ - OP_STAR;
2605      minimize = (c & 1) != 0;      minimize = (c & 1) != 0;
2606    
2607      min = rep_min[c];                 /* Pick up values from tables; */      min = rep_min[c];                 /* Pick up values from tables; */
2608      max = rep_max[c];                 /* zero for max => infinity */      max = rep_max[c];                 /* zero for max => infinity */
2609      if (max == 0) max = INT_MAX;      if (max == 0) max = INT_MAX;
2610    
2611      /* Common code for all repeated single-character matches. We can give      /* Common code for all repeated single-character matches. */
     up quickly if there are fewer than the minimum number of characters left in  
     the subject. */  
2612    
2613      REPEATCHAR:      REPEATCHAR:
2614  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF8
# Line 2069  for (;;) Line 2617  for (;;)
2617        length = 1;        length = 1;
2618        charptr = ecode;        charptr = ecode;
2619        GETCHARLEN(fc, ecode, length);        GETCHARLEN(fc, ecode, length);
       if (min * length > md->end_subject - eptr) RRETURN(MATCH_NOMATCH);  
2620        ecode += length;        ecode += length;
2621    
2622        /* Handle multibyte character matching specially here. There is        /* Handle multibyte character matching specially here. There is
# Line 2080  for (;;) Line 2627  for (;;)
2627  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
2628          unsigned int othercase;          unsigned int othercase;
2629          if ((ims & PCRE_CASELESS) != 0 &&          if ((ims & PCRE_CASELESS) != 0 &&
2630              (othercase = _pcre_ucp_othercase(fc)) != NOTACHAR)              (othercase = UCD_OTHERCASE(fc)) != fc)
2631            oclength = _pcre_ord2utf8(othercase, occhars);            oclength = _pcre_ord2utf8(othercase, occhars);
2632          else oclength = 0;          else oclength = 0;
2633  #endif  /* SUPPORT_UCP */  #endif  /* SUPPORT_UCP */
2634    
2635          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
2636            {            {
2637            if (memcmp(eptr, charptr, length) == 0) eptr += length;            if (eptr <= md->end_subject - length &&
2638                memcmp(eptr, charptr, length) == 0) eptr += length;
2639  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
2640            /* Need braces because of following else */            else if (oclength > 0 &&
2641            else if (oclength == 0) { RRETURN(MATCH_NOMATCH); }                     eptr <= md->end_subject - oclength &&
2642                       memcmp(eptr, occhars, oclength) == 0) eptr += oclength;
2643    #endif  /* SUPPORT_UCP */
2644            else            else
2645              {              {
2646              if (memcmp(eptr, occhars, oclength) != 0) RRETURN(MATCH_NOMATCH);              CHECK_PARTIAL();
2647              eptr += oclength;              RRETURN(MATCH_NOMATCH);
2648              }              }
 #else   /* without SUPPORT_UCP */  
           else { RRETURN(MATCH_NOMATCH); }  
 #endif  /* SUPPORT_UCP */  
2649            }            }
2650    
2651          if (min == max) continue;          if (min == max) continue;
# Line 2109  for (;;) Line 2656  for (;;)
2656              {              {
2657              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM22);              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM22);
2658              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2659              if (fi >= max || eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);              if (fi >= max) RRETURN(MATCH_NOMATCH);
2660              if (memcmp(eptr, charptr, length) == 0) eptr += length;              if (eptr <= md->end_subject - length &&
2661                  memcmp(eptr, charptr, length) == 0) eptr += length;
2662  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
2663              /* Need braces because of following else */              else if (oclength > 0 &&
2664              else if (oclength == 0) { RRETURN(MATCH_NOMATCH); }                       eptr <= md->end_subject - oclength &&
2665                         memcmp(eptr, occhars, oclength) == 0) eptr += oclength;
2666    #endif  /* SUPPORT_UCP */
2667              else              else
2668                {                {
2669                if (memcmp(eptr, occhars, oclength) != 0) RRETURN(MATCH_NOMATCH);                CHECK_PARTIAL();
2670                eptr += oclength;                RRETURN(MATCH_NOMATCH);
2671                }                }
 #else   /* without SUPPORT_UCP */  
             else { RRETURN (MATCH_NOMATCH); }  
 #endif  /* SUPPORT_UCP */  
2672              }              }
2673            /* Control never gets here */            /* Control never gets here */
2674            }            }
# Line 2131  for (;;) Line 2678  for (;;)
2678            pp = eptr;            pp = eptr;
2679            for (i = min; i < max; i++)            for (i = min; i < max; i++)
2680              {              {
2681              if (eptr > md->end_subject - length) break;              if (eptr <= md->end_subject - length &&
2682              if (memcmp(eptr, charptr, length) == 0) eptr += length;                  memcmp(eptr, charptr, length) == 0) eptr += length;
2683  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
2684              else if (oclength == 0) break;              else if (oclength > 0 &&
2685              else                       eptr <= md->end_subject - oclength &&
2686                {                       memcmp(eptr, occhars, oclength) == 0) eptr += oclength;
               if (memcmp(eptr, occhars, oclength) != 0) break;  
               eptr += oclength;  
               }  
 #else   /* without SUPPORT_UCP */  
             else break;  
2687  #endif  /* SUPPORT_UCP */  #endif  /* SUPPORT_UCP */
2688                else break;
2689              }              }
2690    
2691            if (possessive) continue;            if (possessive) continue;
2692    
2693            for(;;)            for(;;)
2694             {              {
2695             RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM23);              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM23);
2696             if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2697             if (eptr == pp) RRETURN(MATCH_NOMATCH);              if (eptr == pp) { RRETURN(MATCH_NOMATCH); }
2698  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
2699             eptr--;              eptr--;
2700             BACKCHAR(eptr);              BACKCHAR(eptr);
2701  #else   /* without SUPPORT_UCP */  #else   /* without SUPPORT_UCP */
2702             eptr -= length;              eptr -= length;
2703  #endif  /* SUPPORT_UCP */  #endif  /* SUPPORT_UCP */
2704             }              }
2705            }            }
2706          /* Control never gets here */          /* Control never gets here */
2707          }          }
# Line 2170  for (;;) Line 2714  for (;;)
2714  #endif  /* SUPPORT_UTF8 */  #endif  /* SUPPORT_UTF8 */
2715    
2716      /* When not in UTF-8 mode, load a single-byte character. */      /* When not in UTF-8 mode, load a single-byte character. */
2717        {  
2718        if (min > md->end_subject - eptr) RRETURN(MATCH_NOMATCH);      fc = *ecode++;
       fc = *ecode++;  
       }  
2719    
2720      /* The value of fc at this point is always less than 256, though we may or      /* The value of fc at this point is always less than 256, though we may or
2721      may not be in UTF-8 mode. The code is duplicated for the caseless and      may not be in UTF-8 mode. The code is duplicated for the caseless and
# Line 2191  for (;;) Line 2733  for (;;)
2733        {        {
2734        fc = md->lcc[fc];        fc = md->lcc[fc];
2735        for (i = 1; i <= min; i++)        for (i = 1; i <= min; i++)
2736            {
2737            if (eptr >= md->end_subject)
2738              {
2739              SCHECK_PARTIAL();
2740              RRETURN(MATCH_NOMATCH);
2741              }
2742          if (fc != md->lcc[*eptr++]) RRETURN(MATCH_NOMATCH);          if (fc != md->lcc[*eptr++]) RRETURN(MATCH_NOMATCH);
2743            }
2744        if (min == max) continue;        if (min == max) continue;
2745        if (minimize)        if (minimize)
2746          {          {
# Line 2199  for (;;) Line 2748  for (;;)
2748            {            {
2749            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM24);            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM24);
2750            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2751            if (fi >= max || eptr >= md->end_subject ||            if (fi >= max) RRETURN(MATCH_NOMATCH);
2752                fc != md->lcc[*eptr++])            if (eptr >= md->end_subject)
2753                {
2754                SCHECK_PARTIAL();
2755              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
2756                }
2757              if (fc != md->lcc[*eptr++]) RRETURN(MATCH_NOMATCH);
2758            }            }
2759          /* Control never gets here */          /* Control never gets here */
2760          }          }
# Line 2213  for (;;) Line 2766  for (;;)
2766            if (eptr >= md->end_subject || fc != md->lcc[*eptr]) break;            if (eptr >= md->end_subject || fc != md->lcc[*eptr]) break;
2767            eptr++;            eptr++;
2768            }            }
2769    
2770          if (possessive) continue;          if (possessive) continue;
2771    
2772          while (eptr >= pp)          while (eptr >= pp)
2773            {            {
2774            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM25);            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM25);
# Line 2229  for (;;) Line 2784  for (;;)
2784    
2785      else      else
2786        {        {
2787        for (i = 1; i <= min; i++) if (fc != *eptr++) RRETURN(MATCH_NOMATCH);        for (i = 1; i <= min; i++)
2788            {
2789            if (eptr >= md->end_subject)
2790              {
2791              SCHECK_PARTIAL();
2792              RRETURN(MATCH_NOMATCH);
2793              }
2794            if (fc != *eptr++) RRETURN(MATCH_NOMATCH);
2795            }
2796    
2797        if (min == max) continue;        if (min == max) continue;
2798    
2799        if (minimize)        if (minimize)
2800          {          {
2801          for (fi = min;; fi++)          for (fi = min;; fi++)
2802            {            {
2803            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM26);            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM26);
2804            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2805            if (fi >= max || eptr >= md->end_subject || fc != *eptr++)            if (fi >= max) RRETURN(MATCH_NOMATCH);
2806              if (eptr >= md->end_subject)
2807                {
2808                SCHECK_PARTIAL();
2809              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
2810                }
2811              if (fc != *eptr++) RRETURN(MATCH_NOMATCH);
2812            }            }
2813          /* Control never gets here */          /* Control never gets here */
2814          }          }
# Line 2251  for (;;) Line 2821  for (;;)
2821            eptr++;            eptr++;
2822            }            }
2823          if (possessive) continue;          if (possessive) continue;
2824    
2825          while (eptr >= pp)          while (eptr >= pp)
2826            {            {
2827            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM27);            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM27);
# Line 2266  for (;;) Line 2837  for (;;)
2837      checking can be multibyte. */      checking can be multibyte. */
2838    
2839      case OP_NOT:      case OP_NOT:
2840      if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);      if (eptr >= md->end_subject)
2841          {
2842          SCHECK_PARTIAL();
2843          RRETURN(MATCH_NOMATCH);
2844          }
2845      ecode++;      ecode++;
2846      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2847      if ((ims & PCRE_CASELESS) != 0)      if ((ims & PCRE_CASELESS) != 0)
# Line 2343  for (;;) Line 2918  for (;;)
2918      max = rep_max[c];                 /* zero for max => infinity */      max = rep_max[c];                 /* zero for max => infinity */
2919      if (max == 0) max = INT_MAX;      if (max == 0) max = INT_MAX;
2920    
2921      /* Common code for all repeated single-byte matches. We can give up quickly      /* Common code for all repeated single-byte matches. */
     if there are fewer than the minimum number of bytes left in the  
     subject. */  
2922    
2923      REPEATNOTCHAR:      REPEATNOTCHAR:
     if (min > md->end_subject - eptr) RRETURN(MATCH_NOMATCH);  
2924      fc = *ecode++;      fc = *ecode++;
2925    
2926      /* The code is duplicated for the caseless and caseful cases, for speed,      /* The code is duplicated for the caseless and caseful cases, for speed,
# Line 2373  for (;;) Line 2945  for (;;)
2945          register unsigned int d;          register unsigned int d;
2946          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
2947            {            {
2948              if (eptr >= md->end_subject)
2949                {
2950                SCHECK_PARTIAL();
2951                RRETURN(MATCH_NOMATCH);
2952                }
2953            GETCHARINC(d, eptr);            GETCHARINC(d, eptr);
2954            if (d < 256) d = md->lcc[d];            if (d < 256) d = md->lcc[d];
2955            if (fc == d) RRETURN(MATCH_NOMATCH);            if (fc == d) RRETURN(MATCH_NOMATCH);
# Line 2384  for (;;) Line 2961  for (;;)
2961        /* Not UTF-8 mode */        /* Not UTF-8 mode */
2962          {          {
2963          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
2964              {
2965              if (eptr >= md->end_subject)
2966                {
2967                SCHECK_PARTIAL();
2968                RRETURN(MATCH_NOMATCH);
2969                }
2970            if (fc == md->lcc[*eptr++]) RRETURN(MATCH_NOMATCH);            if (fc == md->lcc[*eptr++]) RRETURN(MATCH_NOMATCH);
2971              }
2972          }          }
2973    
2974        if (min == max) continue;        if (min == max) continue;
# Line 2400  for (;;) Line 2984  for (;;)
2984              {              {
2985              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM28);              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM28);
2986              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2987                if (fi >= max) RRETURN(MATCH_NOMATCH);
2988                if (eptr >= md->end_subject)
2989                  {
2990                  SCHECK_PARTIAL();
2991                  RRETURN(MATCH_NOMATCH);
2992                  }
2993              GETCHARINC(d, eptr);              GETCHARINC(d, eptr);
2994              if (d < 256) d = md->lcc[d];              if (d < 256) d = md->lcc[d];
2995              if (fi >= max || eptr >= md->end_subject || fc == d)              if (fc == d) RRETURN(MATCH_NOMATCH);
               RRETURN(MATCH_NOMATCH);  
2996              }              }
2997            }            }
2998          else          else
# Line 2414  for (;;) Line 3003  for (;;)
3003              {              {
3004              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM29);              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM29);
3005              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3006              if (fi >= max || eptr >= md->end_subject || fc == md->lcc[*eptr++])              if (fi >= max) RRETURN(MATCH_NOMATCH);
3007                if (eptr >= md->end_subject)
3008                  {
3009                  SCHECK_PARTIAL();
3010                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
3011                  }
3012                if (fc == md->lcc[*eptr++]) RRETURN(MATCH_NOMATCH);
3013              }              }
3014            }            }
3015          /* Control never gets here */          /* Control never gets here */
# Line 2484  for (;;) Line 3078  for (;;)
3078          register unsigned int d;          register unsigned int d;
3079          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
3080            {            {
3081              if (eptr >= md->end_subject)
3082                {
3083                SCHECK_PARTIAL();
3084                RRETURN(MATCH_NOMATCH);
3085                }
3086            GETCHARINC(d, eptr);            GETCHARINC(d, eptr);
3087            if (fc == d) RRETURN(MATCH_NOMATCH);            if (fc == d) RRETURN(MATCH_NOMATCH);
3088            }            }
# Line 2493  for (;;) Line 3092  for (;;)
3092        /* Not UTF-8 mode */        /* Not UTF-8 mode */
3093          {          {
3094          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
3095              {
3096              if (eptr >= md->end_subject)
3097                {
3098                SCHECK_PARTIAL();
3099                RRETURN(MATCH_NOMATCH);
3100                }
3101            if (fc == *eptr++) RRETURN(MATCH_NOMATCH);            if (fc == *eptr++) RRETURN(MATCH_NOMATCH);
3102              }
3103          }          }
3104    
3105        if (min == max) continue;        if (min == max) continue;
# Line 2509  for (;;) Line 3115  for (;;)
3115              {              {
3116              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM32);              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM32);
3117              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3118              GETCHARINC(d, eptr);              if (fi >= max) RRETURN(MATCH_NOMATCH);
3119              if (fi >= max || eptr >= md->end_subject || fc == d)              if (eptr >= md->end_subject)
3120                  {
3121                  SCHECK_PARTIAL();
3122                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
3123                  }
3124                GETCHARINC(d, eptr);
3125                if (fc == d) RRETURN(MATCH_NOMATCH);
3126              }              }
3127            }            }
3128          else          else
# Line 2522  for (;;) Line 3133  for (;;)
3133              {              {
3134              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM33);              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM33);
3135              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3136              if (fi >= max || eptr >= md->end_subject || fc == *eptr++)              if (fi >= max) RRETURN(MATCH_NOMATCH);
3137                if (eptr >= md->end_subject)
3138                  {
3139                  SCHECK_PARTIAL();
3140                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
3141                  }
3142                if (fc == *eptr++) RRETURN(MATCH_NOMATCH);
3143              }              }
3144            }            }
3145          /* Control never gets here */          /* Control never gets here */
# Line 2657  for (;;) Line 3273  for (;;)
3273    
3274      /* First, ensure the minimum number of matches are present. Use inline      /* First, ensure the minimum number of matches are present. Use inline
3275      code for maximizing the speed, and do the type test once at the start      code for maximizing the speed, and do the type test once at the start
3276      (i.e. keep it out of the loop). Also we can test that there are at least      (i.e. keep it out of the loop). Separate the UTF-8 code completely as that
     the minimum number of bytes before we start. This isn't as effective in  
     UTF-8 mode, but it does no harm. Separate the UTF-8 code completely as that  
3277      is tidier. Also separate the UCP code, which can be the same for both UTF-8      is tidier. Also separate the UCP code, which can be the same for both UTF-8
3278      and single-bytes. */      and single-bytes. */
3279    
     if (min > md->end_subject - eptr) RRETURN(MATCH_NOMATCH);  
3280      if (min > 0)      if (min > 0)
3281        {        {
3282  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
# Line 2675  for (;;) Line 3288  for (;;)
3288            if (prop_fail_result) RRETURN(MATCH_NOMATCH);            if (prop_fail_result) RRETURN(MATCH_NOMATCH);
3289            for (i = 1; i <= min; i++)            for (i = 1; i <= min; i++)
3290              {              {
3291              if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);              if (eptr >= md->end_subject)
3292              GETCHARINC(c, eptr);                {
3293                  SCHECK_PARTIAL();
3294                  RRETURN(MATCH_NOMATCH);
3295                  }
3296                GETCHARINCTEST(c, eptr);
3297              }              }
3298            break;            break;
3299    
3300            case PT_LAMP:            case PT_LAMP:
3301            for (i = 1; i <= min; i++)            for (i = 1; i <= min; i++)
3302              {              {
3303              if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);              if (eptr >= md->end_subject)
3304              GETCHARINC(c, eptr);                {
3305              prop_category = _pcre_ucp_findprop(c, &prop_chartype, &prop_script);                SCHECK_PARTIAL();
3306                  RRETURN(MATCH_NOMATCH);
3307                  }
3308                GETCHARINCTEST(c, eptr);
3309                prop_chartype = UCD_CHARTYPE(c);
3310              if ((prop_chartype == ucp_Lu ||              if ((prop_chartype == ucp_Lu ||
3311                   prop_chartype == ucp_Ll ||                   prop_chartype == ucp_Ll ||
3312                   prop_chartype == ucp_Lt) == prop_fail_result)                   prop_chartype == ucp_Lt) == prop_fail_result)
# Line 2696  for (;;) Line 3317  for (;;)
3317            case PT_GC:            case PT_GC:
3318            for (i = 1; i <= min; i++)            for (i = 1; i <= min; i++)
3319              {              {
3320              if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);              if (eptr >= md->end_subject)
3321              GETCHARINC(c, eptr);                {
3322              prop_category = _pcre_ucp_findprop(c, &prop_chartype, &prop_script);                SCHECK_PARTIAL();
3323                  RRETURN(MATCH_NOMATCH);
3324                  }
3325                GETCHARINCTEST(c, eptr);
3326                prop_category = UCD_CATEGORY(c);
3327              if ((prop_category == prop_value) == prop_fail_result)              if ((prop_category == prop_value) == prop_fail_result)
3328                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
3329              }              }
# Line 2707  for (;;) Line 3332  for (;;)
3332            case PT_PC:            case PT_PC:
3333            for (i = 1; i <= min; i++)            for (i = 1; i <= min; i++)
3334              {              {
3335              if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);              if (eptr >= md->end_subject)
3336              GETCHARINC(c, eptr);                {
3337              prop_category = _pcre_ucp_findprop(c, &prop_chartype, &prop_script);                SCHECK_PARTIAL();
3338                  RRETURN(MATCH_NOMATCH);
3339                  }
3340                GETCHARINCTEST(c, eptr);
3341                prop_chartype = UCD_CHARTYPE(c);
3342              if ((prop_chartype == prop_value) == prop_fail_result)              if ((prop_chartype == prop_value) == prop_fail_result)
3343                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
3344              }              }
# Line 2718  for (;;) Line 3347  for (;;)
3347            case PT_SC:            case PT_SC:
3348            for (i = 1; i <= min; i++)            for (i = 1; i <= min; i++)
3349              {              {
3350              if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);              if (eptr >= md->end_subject)
3351              GETCHARINC(c, eptr);                {
3352              prop_category = _pcre_ucp_findprop(c, &prop_chartype, &prop_script);                SCHECK_PARTIAL();
3353                  RRETURN(MATCH_NOMATCH);
3354                  }
3355                GETCHARINCTEST(c, eptr);
3356                prop_script = UCD_SCRIPT(c);
3357              if ((prop_script == prop_value) == prop_fail_result)              if ((prop_script == prop_value) == prop_fail_result)
3358                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
3359              }              }
# Line 2738  for (;;) Line 3371  for (;;)
3371          {          {
3372          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
3373            {            {
3374              if (eptr >= md->end_subject)
3375                {
3376                SCHECK_PARTIAL();
3377                RRETURN(MATCH_NOMATCH);
3378                }
3379            GETCHARINCTEST(c, eptr);            GETCHARINCTEST(c, eptr);
3380            prop_category = _pcre_ucp_findprop(c, &prop_chartype, &prop_script);            prop_category = UCD_CATEGORY(c);
3381            if (prop_category == ucp_M) RRETURN(MATCH_NOMATCH);            if (prop_category == ucp_M) RRETURN(MATCH_NOMATCH);
3382            while (eptr < md->end_subject)            while (eptr < md->end_subject)
3383              {              {
3384              int len = 1;              int len = 1;
3385              if (!utf8) c = *eptr; else              if (!utf8) c = *eptr;
3386                {                else { GETCHARLEN(c, eptr, len); }
3387                GETCHARLEN(c, eptr, len);              prop_category = UCD_CATEGORY(c);
               }  
             prop_category = _pcre_ucp_findprop(c, &prop_chartype, &prop_script);  
3388              if (prop_category != ucp_M) break;              if (prop_category != ucp_M) break;
3389              eptr += len;              eptr += len;
3390              }              }
# Line 2766  for (;;) Line 3402  for (;;)
3402          case OP_ANY:          case OP_ANY:
3403          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
3404            {            {
3405            if (eptr >= md->end_subject ||            if (eptr >= md->end_subject)
3406                 ((ims & PCRE_DOTALL) == 0 && IS_NEWLINE(eptr)))              {
3407                SCHECK_PARTIAL();
3408              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
3409                }
3410              if (IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH);
3411              eptr++;
3412              while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;
3413              }
3414            break;
3415    
3416            case OP_ALLANY:
3417            for (i = 1; i <= min; i++)
3418              {
3419              if (eptr >= md->end_subject)
3420                {
3421                SCHECK_PARTIAL();
3422                RRETURN(MATCH_NOMATCH);
3423                }
3424            eptr++;            eptr++;
3425            while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;            while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;
3426            }            }
3427          break;          break;
3428    
3429          case OP_ANYBYTE:          case OP_ANYBYTE:
3430            if (eptr > md->end_subject - min) RRETURN(MATCH_NOMATCH);
3431          eptr += min;          eptr += min;
3432          break;          break;
3433    
3434          case OP_ANYNL:          case OP_ANYNL:
3435          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
3436            {            {
3437            if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);            if (eptr >= md->end_subject)
3438                {
3439                SCHECK_PARTIAL();
3440                RRETURN(MATCH_NOMATCH);
3441                }
3442            GETCHARINC(c, eptr);            GETCHARINC(c, eptr);
3443            switch(c)            switch(c)
3444              {              {
# Line 2789  for (;;) Line 3446  for (;;)
3446              case 0x000d:              case 0x000d:
3447              if (eptr < md->end_subject && *eptr == 0x0a) eptr++;              if (eptr < md->end_subject && *eptr == 0x0a) eptr++;
3448              break;              break;
3449    
3450              case 0x000a:              case 0x000a:
3451                break;
3452    
3453              case 0x000b:              case 0x000b:
3454              case 0x000c:              case 0x000c:
3455              case 0x0085:              case 0x0085:
3456              case 0x2028:              case 0x2028:
3457              case 0x2029:              case 0x2029:
3458                if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);
3459                break;
3460                }
3461              }
3462            break;
3463    
3464            case OP_NOT_HSPACE:
3465            for (i = 1; i <= min; i++)
3466              {
3467              if (eptr >= md->end_subject)
3468                {
3469                SCHECK_PARTIAL();
3470                RRETURN(MATCH_NOMATCH);
3471                }
3472              GETCHARINC(c, eptr);
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              }
3498            break;
3499    
3500            case OP_HSPACE:
3501            for (i = 1; i <= min; i++)
3502              {
3503              if (eptr >= md->end_subject)
3504                {
3505                SCHECK_PARTIAL();
3506                RRETURN(MATCH_NOMATCH);
3507                }
3508              GETCHARINC(c, eptr);
3509              switch(c)
3510                {
3511                default: RRETURN(MATCH_NOMATCH);
3512                case 0x09:      /* HT */
3513                case 0x20:      /* SPACE */
3514                case 0xa0:      /* NBSP */
3515                case 0x1680:    /* OGHAM SPACE MARK */
3516                case 0x180e:    /* MONGOLIAN VOWEL SEPARATOR */
3517                case 0x2000:    /* EN QUAD */
3518                case 0x2001:    /* EM QUAD */
3519                case 0x2002:    /* EN SPACE */
3520                case 0x2003:    /* EM SPACE */
3521                case 0x2004:    /* THREE-PER-EM SPACE */
3522                case 0x2005:    /* FOUR-PER-EM SPACE */
3523                case 0x2006:    /* SIX-PER-EM SPACE */
3524                case 0x2007:    /* FIGURE SPACE */
3525                case 0x2008:    /* PUNCTUATION SPACE */
3526                case 0x2009:    /* THIN SPACE */
3527                case 0x200A:    /* HAIR SPACE */
3528                case 0x202f:    /* NARROW NO-BREAK SPACE */
3529                case 0x205f:    /* MEDIUM MATHEMATICAL SPACE */
3530                case 0x3000:    /* IDEOGRAPHIC SPACE */
3531                break;
3532                }
3533              }
3534            break;
3535    
3536            case OP_NOT_VSPACE:
3537            for (i = 1; i <= min; i++)
3538              {
3539              if (eptr >= md->end_subject)
3540                {
3541                SCHECK_PARTIAL();
3542                RRETURN(MATCH_NOMATCH);
3543                }
3544              GETCHARINC(c, eptr);
3545              switch(c)
3546                {
3547                default: break;
3548                case 0x0a:      /* LF */
3549                case 0x0b:      /* VT */
3550                case 0x0c:      /* FF */
3551                case 0x0d:      /* CR */
3552                case 0x85:      /* NEL */
3553                case 0x2028:    /* LINE SEPARATOR */
3554                case 0x2029:    /* PARAGRAPH SEPARATOR */
3555                RRETURN(MATCH_NOMATCH);
3556                }
3557              }
3558            break;
3559    
3560            case OP_VSPACE:
3561            for (i = 1; i <= min; i++)
3562              {
3563              if (eptr >= md->end_subject)
3564                {
3565                SCHECK_PARTIAL();
3566                RRETURN(MATCH_NOMATCH);
3567                }
3568              GETCHARINC(c, eptr);
3569              switch(c)
3570                {
3571                default: RRETURN(MATCH_NOMATCH);
3572                case 0x0a:      /* LF */
3573                case 0x0b:      /* VT */
3574                case 0x0c:      /* FF */
3575                case 0x0d:      /* CR */
3576                case 0x85:      /* NEL */
3577                case 0x2028:    /* LINE SEPARATOR */
3578                case 0x2029:    /* PARAGRAPH SEPARATOR */
3579              break;              break;
3580              }              }
3581            }            }
# Line 2803  for (;;) Line 3584  for (;;)
3584          case OP_NOT_DIGIT:          case OP_NOT_DIGIT:
3585          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
3586            {            {
3587            if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);            if (eptr >= md->end_subject)
3588                {
3589                SCHECK_PARTIAL();
3590                RRETURN(MATCH_NOMATCH);
3591                }
3592            GETCHARINC(c, eptr);            GETCHARINC(c, eptr);
3593            if (c < 128 && (md->ctypes[c] & ctype_digit) != 0)            if (c < 128 && (md->ctypes[c] & ctype_digit) != 0)
3594              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
# Line 2813  for (;;) Line 3598  for (;;)
3598          case OP_DIGIT:          case OP_DIGIT:
3599          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
3600            {            {
3601            if (eptr >= md->end_subject ||            if (eptr >= md->end_subject)
3602               *eptr >= 128 || (md->ctypes[*eptr++] & ctype_digit) == 0)              {
3603                SCHECK_PARTIAL();
3604                RRETURN(MATCH_NOMATCH);
3605                }
3606              if (*eptr >= 128 || (md->ctypes[*eptr++] & ctype_digit) == 0)
3607              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
3608            /* No need to skip more bytes - we know it's a 1-byte character */            /* No need to skip more bytes - we know it's a 1-byte character */
3609            }            }
# Line 2823  for (;;) Line 3612  for (;;)
3612          case OP_NOT_WHITESPACE:          case OP_NOT_WHITESPACE:
3613          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
3614            {            {
3615            if (eptr >= md->end_subject ||            if (eptr >= md->end_subject)
3616               (*eptr < 128 && (md->ctypes[*eptr++] & ctype_space) != 0))              {
3617                SCHECK_PARTIAL();
3618              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
3619            while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;              }
3620              if (*eptr < 128 && (md->ctypes[*eptr] & ctype_space) != 0)
3621                RRETURN(MATCH_NOMATCH);
3622              while (++eptr < md->end_subject && (*eptr & 0xc0) == 0x80);
3623            }            }
3624          break;          break;
3625    
3626          case OP_WHITESPACE:          case OP_WHITESPACE:
3627          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
3628            {            {
3629            if (eptr >= md->end_subject ||            if (eptr >= md->end_subject)
3630               *eptr >= 128 || (md->ctypes[*eptr++] & ctype_space) == 0)              {
3631                SCHECK_PARTIAL();
3632                RRETURN(MATCH_NOMATCH);
3633                }
3634              if (*eptr >= 128 || (md->ctypes[*eptr++] & ctype_space) == 0)
3635              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
3636            /* No need to skip more bytes - we know it's a 1-byte character */            /* No need to skip more bytes - we know it's a 1-byte character */
3637            }            }
# Line 2844  for (;;) Line 3641  for (;;)
3641          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
3642            {            {
3643            if (eptr >= md->end_subject ||            if (eptr >= md->end_subject ||
3644               (*eptr < 128 && (md->ctypes[*eptr++] & ctype_word) != 0))               (*eptr < 128 && (md->ctypes[*eptr] & ctype_word) != 0))
3645              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
3646            while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;            while (++eptr < md->end_subject && (*eptr & 0xc0) == 0x80);
3647            }            }
3648          break;          break;
3649    
3650          case OP_WORDCHAR:          case OP_WORDCHAR:
3651          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
3652            {            {
3653            if (eptr >= md->end_subject ||            if (eptr >= md->end_subject)
3654               *eptr >= 128 || (md->ctypes[*eptr++] & ctype_word) == 0)              {
3655                SCHECK_PARTIAL();
3656                RRETURN(MATCH_NOMATCH);
3657                }
3658              if (*eptr >= 128 || (md->ctypes[*eptr++] & ctype_word) == 0)
3659              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
3660            /* No need to skip more bytes - we know it's a 1-byte character */            /* No need to skip more bytes - we know it's a 1-byte character */
3661            }            }
# Line 2868  for (;;) Line 3669  for (;;)
3669  #endif     /* SUPPORT_UTF8 */  #endif     /* SUPPORT_UTF8 */
3670    
3671        /* Code for the non-UTF-8 case for minimum matching of operators other        /* Code for the non-UTF-8 case for minimum matching of operators other
3672        than OP_PROP and OP_NOTPROP. We can assume that there are the minimum        than OP_PROP and OP_NOTPROP. */
       number of bytes present, as this was tested above. */  
3673    
3674        switch(ctype)        switch(ctype)
3675          {          {
3676          case OP_ANY:          case OP_ANY:
3677          if ((ims & PCRE_DOTALL) == 0)          for (i = 1; i <= min; i++)
3678            {            {
3679            for (i = 1; i <= min; i++)            if (eptr >= md->end_subject)
3680              {              {
3681              if (IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH);              SCHECK_PARTIAL();
3682              eptr++;              RRETURN(MATCH_NOMATCH);
3683              }              }
3684              if (IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH);
3685              eptr++;
3686            }            }
         else eptr += min;  
3687          break;          break;
3688    
3689          case OP_ANYBYTE:          case OP_ALLANY:
3690            if (eptr > md->end_subject - min)
3691              {
3692              SCHECK_PARTIAL();
3693              RRETURN(MATCH_NOMATCH);
3694              }
3695          eptr += min;          eptr += min;
3696          break;          break;
3697    
3698          /* Because of the CRLF case, we can't assume the minimum number of          case OP_ANYBYTE:
3699          bytes are present in this case. */          if (eptr > md->end_subject - min)
3700              {
3701              SCHECK_PARTIAL();
3702              RRETURN(MATCH_NOMATCH);
3703              }
3704            eptr += min;
3705            break;
3706    
3707          case OP_ANYNL:          case OP_ANYNL:
3708          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
3709            {            {
3710            if (eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);            if (eptr >= md->end_subject)
3711                {
3712                SCHECK_PARTIAL();
3713                RRETURN(MATCH_NOMATCH);
3714                }
3715            switch(*eptr++)            switch(*eptr++)
3716              {              {
3717              default: RRETURN(MATCH_NOMATCH);              default: RRETURN(MATCH_NOMATCH);
# Line 2903  for (;;) Line 3719  for (;;)
3719              if (eptr < md->end_subject && *eptr == 0x0a) eptr++;              if (eptr < md->end_subject && *eptr == 0x0a) eptr++;
3720              break;              break;
3721              case 0x000a:              case 0x000a:
3722                break;
3723    
3724              case 0x000b:              case 0x000b:
3725              case 0x000c:              case 0x000c:
3726              case 0x0085:              case 0x0085:
3727                if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);
3728                break;
3729                }
3730              }
3731            break;
3732    
3733            case OP_NOT_HSPACE:
3734            for (i = 1; i <= min; i++)
3735              {
3736              if (eptr >= md->end_subject)
3737                {
3738                SCHECK_PARTIAL();
3739                RRETURN(MATCH_NOMATCH);
3740                }
3741              switch(*eptr++)
3742                {
3743                default: break;
3744                case 0x09:      /* HT */
3745                case 0x20:      /* SPACE */
3746                case 0xa0:      /* NBSP */
3747                RRETURN(MATCH_NOMATCH);
3748                }
3749              }
3750            break;
3751    
3752            case OP_HSPACE:
3753            for (i = 1; i <= min; i++)
3754              {
3755              if (eptr >= md->end_subject)
3756                {
3757                SCHECK_PARTIAL();
3758                RRETURN(MATCH_NOMATCH);
3759                }
3760              switch(*eptr++)
3761                {
3762                default: RRETURN(MATCH_NOMATCH);
3763                case 0x09:      /* HT */
3764                case 0x20:      /* SPACE */
3765                case 0xa0:      /* NBSP */
3766                break;
3767                }
3768              }
3769            break;
3770    
3771            case OP_NOT_VSPACE:
3772            for (i = 1; i <= min; i++)
3773              {
3774              if (eptr >= md->end_subject)
3775                {
3776                SCHECK_PARTIAL();
3777                RRETURN(MATCH_NOMATCH);
3778                }
3779              switch(*eptr++)
3780                {
3781                default: break;
3782                case 0x0a:      /* LF */
3783                case 0x0b:      /* VT */
3784                case 0x0c:      /* FF */
3785                case 0x0d:      /* CR */
3786                case 0x85:      /* NEL */
3787                RRETURN(MATCH_NOMATCH);
3788                }
3789              }
3790            break;
3791    
3792            case OP_VSPACE:
3793            for (i = 1; i <= min; i++)
3794              {
3795              if (eptr >= md->end_subject)
3796                {
3797                SCHECK_PARTIAL();
3798                RRETURN(MATCH_NOMATCH);
3799                }
3800              switch(*eptr++)
3801                {
3802                default: RRETURN(MATCH_NOMATCH);
3803                case 0x0a:      /* LF */
3804                case 0x0b:      /* VT */
3805                case 0x0c:      /* FF */
3806                case 0x0d:      /* CR */
3807                case 0x85:      /* NEL */
3808              break;              break;
3809              }              }
3810            }            }
# Line 2913  for (;;) Line 3812  for (;;)
3812    
3813          case OP_NOT_DIGIT:          case OP_NOT_DIGIT:
3814          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
3815              {
3816              if (eptr >= md->end_subject)
3817                {
3818                SCHECK_PARTIAL();
3819                RRETURN(MATCH_NOMATCH);
3820                }
3821            if ((md->ctypes[*eptr++] & ctype_digit) != 0) RRETURN(MATCH_NOMATCH);            if ((md->ctypes[*eptr++] & ctype_digit) != 0) RRETURN(MATCH_NOMATCH);
3822              }
3823          break;          break;
3824    
3825          case OP_DIGIT:          case OP_DIGIT:
3826          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
3827              {
3828              if (eptr >= md->end_subject)
3829                {
3830                SCHECK_PARTIAL();
3831                RRETURN(MATCH_NOMATCH);
3832                }
3833            if ((md->ctypes[*eptr++] & ctype_digit) == 0) RRETURN(MATCH_NOMATCH);            if ((md->ctypes[*eptr++] & ctype_digit) == 0) RRETURN(MATCH_NOMATCH);
3834              }
3835          break;          break;
3836    
3837          case OP_NOT_WHITESPACE:          case OP_NOT_WHITESPACE:
3838          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
3839              {
3840              if (eptr >= md->end_subject)
3841                {
3842                SCHECK_PARTIAL();
3843                RRETURN(MATCH_NOMATCH);
3844                }
3845            if ((md->ctypes[*eptr++] & ctype_space) != 0) RRETURN(MATCH_NOMATCH);            if ((md->ctypes[*eptr++] & ctype_space) != 0) RRETURN(MATCH_NOMATCH);
3846              }
3847          break;          break;
3848    
3849          case OP_WHITESPACE:          case OP_WHITESPACE:
3850          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
3851              {
3852              if (eptr >= md->end_subject)
3853                {
3854                SCHECK_PARTIAL();
3855                RRETURN(MATCH_NOMATCH);
3856                }
3857            if ((md->ctypes[*eptr++] & ctype_space) == 0) RRETURN(MATCH_NOMATCH);            if ((md->ctypes[*eptr++] & ctype_space) == 0) RRETURN(MATCH_NOMATCH);
3858              }
3859          break;          break;
3860    
3861          case OP_NOT_WORDCHAR:          case OP_NOT_WORDCHAR:
3862          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
3863              {
3864              if (eptr >= md->end_subject)
3865                {
3866                SCHECK_PARTIAL();
3867                RRETURN(MATCH_NOMATCH);
3868                }
3869            if ((md->ctypes[*eptr++] & ctype_word) != 0)            if ((md->ctypes[*eptr++] & ctype_word) != 0)
3870              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
3871              }
3872          break;          break;
3873    
3874          case OP_WORDCHAR:          case OP_WORDCHAR:
3875          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
3876              {
3877              if (eptr >= md->end_subject)
3878                {
3879                SCHECK_PARTIAL();
3880                RRETURN(MATCH_NOMATCH);
3881                }
3882            if ((md->ctypes[*eptr++] & ctype_word) == 0)            if ((md->ctypes[*eptr++] & ctype_word) == 0)
3883              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
3884              }
3885          break;          break;
3886    
3887          default:          default:
# Line 2968  for (;;) Line 3909  for (;;)
3909              {              {
3910              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM36);              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM36);
3911              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3912              if (fi >= max || eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);              if (fi >= max) RRETURN(MATCH_NOMATCH);
3913                if (eptr >= md->end_subject)
3914                  {
3915                  SCHECK_PARTIAL();
3916                  RRETURN(MATCH_NOMATCH);
3917                  }
3918              GETCHARINC(c, eptr);              GETCHARINC(c, eptr);
3919              if (prop_fail_result) RRETURN(MATCH_NOMATCH);              if (prop_fail_result) RRETURN(MATCH_NOMATCH);
3920              }              }
# Line 2979  for (;;) Line 3925  for (;;)
3925              {              {
3926              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM37);              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM37);
3927              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3928              if (fi >= max || eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);              if (fi >= max) RRETURN(MATCH_NOMATCH);
3929                if (eptr >= md->end_subject)
3930                  {
3931                  SCHECK_PARTIAL();
3932                  RRETURN(MATCH_NOMATCH);
3933                  }
3934              GETCHARINC(c, eptr);              GETCHARINC(c, eptr);
3935              prop_category = _pcre_ucp_findprop(c, &prop_chartype, &prop_script);              prop_chartype = UCD_CHARTYPE(c);
3936              if ((prop_chartype == ucp_Lu ||              if ((prop_chartype == ucp_Lu ||
3937                   prop_chartype == ucp_Ll ||                   prop_chartype == ucp_Ll ||
3938                   prop_chartype == ucp_Lt) == prop_fail_result)                   prop_chartype == ucp_Lt) == prop_fail_result)
# Line 2994  for (;;) Line 3945  for (;;)
3945              {              {
3946              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM38);              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM38);
3947              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3948              if (fi >= max || eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);              if (fi >= max) RRETURN(MATCH_NOMATCH);
3949                if (eptr >= md->end_subject)
3950                  {
3951                  SCHECK_PARTIAL();
3952                  RRETURN(MATCH_NOMATCH);
3953                  }
3954              GETCHARINC(c, eptr);              GETCHARINC(c, eptr);
3955              prop_category = _pcre_ucp_findprop(c, &prop_chartype, &prop_script);              prop_category = UCD_CATEGORY(c);
3956              if ((prop_category == prop_value) == prop_fail_result)              if ((prop_category == prop_value) == prop_fail_result)
3957                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
3958              }              }
# Line 3007  for (;;) Line 3963  for (;;)
3963              {              {
3964              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM39);              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM39);
3965              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3966              if (fi >= max || eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);              if (fi >= max) RRETURN(MATCH_NOMATCH);
3967                if (eptr >= md->end_subject)
3968                  {
3969                  SCHECK_PARTIAL();
3970                  RRETURN(MATCH_NOMATCH);
3971                  }
3972              GETCHARINC(c, eptr);              GETCHARINC(c, eptr);
3973              prop_category = _pcre_ucp_findprop(c, &prop_chartype, &prop_script);              prop_chartype = UCD_CHARTYPE(c);
3974              if ((prop_chartype == prop_value) == prop_fail_result)              if ((prop_chartype == prop_value) == prop_fail_result)
3975                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
3976              }              }
# Line 3020  for (;;) Line 3981  for (;;)
3981              {              {
3982              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM40);              RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM40);
3983              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3984              if (fi >= max || eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);              if (fi >= max) RRETURN(MATCH_NOMATCH);
3985                if (eptr >= md->end_subject)
3986                  {
3987                  SCHECK_PARTIAL();
3988                  RRETURN(MATCH_NOMATCH);
3989                  }
3990              GETCHARINC(c, eptr);              GETCHARINC(c, eptr);
3991              prop_category = _pcre_ucp_findprop(c, &prop_chartype, &prop_script);              prop_script = UCD_SCRIPT(c);
3992              if ((prop_script == prop_value) == prop_fail_result)              if ((prop_script == prop_value) == prop_fail_result)
3993                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
3994              }              }
# Line 3042  for (;;) Line 4008  for (;;)
4008            {            {
4009            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM41);            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM41);
4010            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
4011            if (fi >= max || eptr >= md->end_subject) RRETURN(MATCH_NOMATCH);            if (fi >= max) RRETURN(MATCH_NOMATCH);
4012              if (eptr >= md->end_subject)
4013                {
4014                SCHECK_PARTIAL();
4015                RRETURN(MATCH_NOMATCH);
4016                }
4017            GETCHARINCTEST(c, eptr);            GETCHARINCTEST(c, eptr);
4018            prop_category = _pcre_ucp_findprop(c, &prop_chartype, &prop_script);            prop_category = UCD_CATEGORY(c);
4019            if (prop_category == ucp_M) RRETURN(MATCH_NOMATCH);            if (prop_category == ucp_M) RRETURN(MATCH_NOMATCH);
4020            while (eptr < md->end_subject)            while (eptr < md->end_subject)
4021              {              {
4022              int len = 1;              int len = 1;
4023              if (!utf8) c = *eptr; else              if (!utf8) c = *eptr;
4024                {                else { GETCHARLEN(c, eptr, len); }
4025                GETCHARLEN(c, eptr, len);              prop_category = UCD_CATEGORY(c);
               }  
             prop_category = _pcre_ucp_findprop(c, &prop_chartype, &prop_script);  
4026              if (prop_category != ucp_M) break;              if (prop_category != ucp_M) break;
4027              eptr += len;              eptr += len;
4028              }              }
# Line 3071  for (;;) Line 4040  for (;;)
4040            {            {
4041            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM42);            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM42);
4042            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
4043            if (fi >= max || eptr >= md->end_subject ||            if (fi >= max) RRETURN(MATCH_NOMATCH);
4044                 (ctype == OP_ANY && (ims & PCRE_DOTALL) == 0 &&            if (eptr >= md->end_subject)
4045                  IS_NEWLINE(eptr)))              {
4046                SCHECK_PARTIAL();
4047                RRETURN(MATCH_NOMATCH);
4048                }
4049              if (ctype == OP_ANY && IS_NEWLINE(eptr))
4050              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
   
4051            GETCHARINC(c, eptr);            GETCHARINC(c, eptr);
4052            switch(ctype)            switch(ctype)
4053              {              {
4054              case OP_ANY:        /* This is the DOTALL case */              case OP_ANY:        /* This is the non-NL case */
4055              break;              case OP_ALLANY:
   
4056              case OP_ANYBYTE:              case OP_ANYBYTE:
4057              break;              break;
4058    
# Line 3093  for (;;) Line 4064  for (;;)
4064                if (eptr < md->end_subject && *eptr == 0x0a) eptr++;                if (eptr < md->end_subject && *eptr == 0x0a) eptr++;
4065                break;                break;
4066                case 0x000a:                case 0x000a:
4067                  break;
4068    
4069                case 0x000b:                case 0x000b:
4070                case 0x000c:                case 0x000c:
4071                case 0x0085:                case 0x0085:
4072                case 0x2028:                case 0x2028:
4073                case 0x2029:                case 0x2029:
4074                  if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);
4075                  break;
4076                  }
4077                break;
4078    
4079                case OP_NOT_HSPACE:
4080                switch(c)
4081                  {
4082                  default: break;
4083                  case 0x09:      /* HT */
4084                  case 0x20:      /* SPACE */
4085                  case 0xa0:      /* NBSP */
4086                  case 0x1680:    /* OGHAM SPACE MARK */
4087                  case 0x180e:    /* MONGOLIAN VOWEL SEPARATOR */
4088                  case 0x2000:    /* EN QUAD */
4089                  case 0x2001:    /* EM QUAD */
4090                  case 0x2002:    /* EN SPACE */
4091                  case 0x2003:    /* EM SPACE */
4092                  case 0x2004:    /* THREE-PER-EM SPACE */
4093                  case 0x2005:    /* FOUR-PER-EM SPACE */
4094                  case 0x2006:    /* SIX-PER-EM SPACE */
4095                  case 0x2007:    /* FIGURE SPACE */
4096                  case 0x2008:    /* PUNCTUATION SPACE */
4097                  case 0x2009:    /* THIN SPACE */
4098                  case 0x200A:    /* HAIR SPACE */
4099                  case 0x202f:    /* NARROW NO-BREAK SPACE */
4100                  case 0x205f:    /* MEDIUM MATHEMATICAL SPACE */
4101                  case 0x3000:    /* IDEOGRAPHIC SPACE */
4102                  RRETURN(MATCH_NOMATCH);
4103                  }
4104                break;
4105    
4106                case OP_HSPACE:
4107                switch(c)
4108                  {
4109                  default: RRETURN(MATCH_NOMATCH);
4110                  case 0x09:      /* HT */
4111                  case 0x20:      /* SPACE */
4112                  case 0xa0:      /* NBSP */
4113                  case 0x1680:    /* OGHAM SPACE MARK */
4114                  case 0x180e:    /* MONGOLIAN VOWEL SEPARATOR */
4115                  case 0x2000:    /* EN QUAD */
4116                  case 0x2001:    /* EM QUAD */
4117                  case 0x2002:    /* EN SPACE */
4118                  case 0x2003:    /* EM SPACE */
4119                  case 0x2004:    /* THREE-PER-EM SPACE */
4120                  case 0x2005:    /* FOUR-PER-EM SPACE */
4121                  case 0x2006:    /* SIX-PER-EM SPACE */
4122                  case 0x2007:    /* FIGURE SPACE */
4123                  case 0x2008:    /* PUNCTUATION SPACE */
4124                  case 0x2009:    /* THIN SPACE */
4125                  case 0x200A:    /* HAIR SPACE */
4126                  case 0x202f:    /* NARROW NO-BREAK SPACE */
4127                  case 0x205f:    /* MEDIUM MATHEMATICAL SPACE */
4128                  case 0x3000:    /* IDEOGRAPHIC SPACE */
4129                  break;
4130                  }
4131                break;
4132    
4133                case OP_NOT_VSPACE:
4134                switch(c)
4135                  {
4136                  default: break;
4137                  case 0x0a:      /* LF */
4138                  case 0x0b:      /* VT */
4139                  case 0x0c:      /* FF */
4140                  case 0x0d:      /* CR */
4141                  case 0x85:      /* NEL */
4142                  case 0x2028:    /* LINE SEPARATOR */
4143                  case 0x2029:    /* PARAGRAPH SEPARATOR */
4144                  RRETURN(MATCH_NOMATCH);
4145                  }
4146                break;
4147    
4148                case OP_VSPACE:
4149                switch(c)
4150                  {
4151                  default: RRETURN(MATCH_NOMATCH);
4152                  case 0x0a:      /* LF */
4153                  case 0x0b:      /* VT */
4154                  case 0x0c:      /* FF */
4155                  case 0x0d:      /* CR */
4156                  case 0x85:      /* NEL */
4157                  case 0x2028:    /* LINE SEPARATOR */
4158                  case 0x2029:    /* PARAGRAPH SEPARATOR */
4159                break;                break;
4160                }                }
4161              break;              break;
# Line 3145  for (;;) Line 4203  for (;;)
4203            {            {
4204            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM43);            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM43);
4205            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
4206            if (fi >= max || eptr >= md->end_subject ||            if (fi >= max) RRETURN(MATCH_NOMATCH);
4207                 ((ims & PCRE_DOTALL) == 0 && IS_NEWLINE(eptr)))            if (eptr >= md->end_subject)
4208                {
4209                SCHECK_PARTIAL();
4210                RRETURN(MATCH_NOMATCH);
4211                }
4212              if (ctype == OP_ANY && IS_NEWLINE(eptr))
4213              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
   
4214            c = *eptr++;            c = *eptr++;
4215            switch(ctype)            switch(ctype)
4216              {              {
4217              case OP_ANY:   /* This is the DOTALL case */              case OP_ANY:     /* This is the non-NL case */
4218              break;              case OP_ALLANY:
   
4219              case OP_ANYBYTE:              case OP_ANYBYTE:
4220              break;              break;
4221    
# Line 3165  for (;;) Line 4226  for (;;)
4226                case 0x000d:                case 0x000d:
4227                if (eptr < md->end_subject && *eptr == 0x0a) eptr++;                if (eptr < md->end_subject && *eptr == 0x0a) eptr++;
4228                break;                break;
4229    
4230                case 0x000a:                case 0x000a:
4231                  break;
4232    
4233                case 0x000b:                case 0x000b:
4234                case 0x000c:                case 0x000c:
4235                case 0x0085:                case 0x0085:
4236                  if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);
4237                  break;
4238                  }
4239                break;
4240    
4241                case OP_NOT_HSPACE:
4242                switch(c)
4243                  {
4244                  default: break;
4245                  case 0x09:      /* HT */
4246                  case 0x20:      /* SPACE */
4247                  case 0xa0:      /* NBSP */
4248                  RRETURN(MATCH_NOMATCH);
4249                  }
4250                break;
4251    
4252                case OP_HSPACE:
4253                switch(c)
4254                  {
4255                  default: RRETURN(MATCH_NOMATCH);
4256                  case 0x09:      /* HT */
4257                  case 0x20:      /* SPACE */
4258                  case 0xa0:      /* NBSP */
4259                  break;
4260                  }
4261                break;
4262    
4263                case OP_NOT_VSPACE:
4264                switch(c)
4265                  {
4266                  default: break;
4267                  case 0x0a:      /* LF */
4268                  case 0x0b:      /* VT */
4269                  case 0x0c:      /* FF */
4270                  case 0x0d:      /* CR */
4271                  case 0x85:      /* NEL */
4272                  RRETURN(MATCH_NOMATCH);
4273                  }
4274                break;
4275    
4276                case OP_VSPACE:
4277                switch(c)
4278                  {
4279                  default: RRETURN(MATCH_NOMATCH);
4280                  case 0x0a:      /* LF */
4281                  case 0x0b:      /* VT */
4282                  case 0x0c:      /* FF */
4283                  case 0x0d:      /* CR */
4284                  case 0x85:      /* NEL */
4285                break;                break;
4286                }                }
4287              break;              break;
# Line 3235  for (;;) Line 4348  for (;;)
4348              int len = 1;              int len = 1;
4349              if (eptr >= md->end_subject) break;              if (eptr >= md->end_subject) break;
4350              GETCHARLEN(c, eptr, len);              GETCHARLEN(c, eptr, len);
4351              prop_category = _pcre_ucp_findprop(c, &prop_chartype, &prop_script);              prop_chartype = UCD_CHARTYPE(c);
4352              if ((prop_chartype == ucp_Lu ||              if ((prop_chartype == ucp_Lu ||
4353                   prop_chartype == ucp_Ll ||                   prop_chartype == ucp_Ll ||
4354                   prop_chartype == ucp_Lt) == prop_fail_result)                   prop_chartype == ucp_Lt) == prop_fail_result)
# Line 3250  for (;;) Line 4363  for (;;)
4363              int len = 1;              int len = 1;
4364              if (eptr >= md->end_subject) break;              if (eptr >= md->end_subject) break;
4365              GETCHARLEN(c, eptr, len);              GETCHARLEN(c, eptr, len);
4366              prop_category = _pcre_ucp_findprop(c, &prop_chartype, &prop_script);              prop_category = UCD_CATEGORY(c);
4367              if ((prop_category == prop_value) == prop_fail_result)              if ((prop_category == prop_value) == prop_fail_result)
4368                break;                break;
4369              eptr+= len;              eptr+= len;
# Line 3263  for (;;) Line 4376  for (;;)
4376              int len = 1;              int len = 1;
4377              if (eptr >= md->end_subject) break;              if (eptr >= md->end_subject) break;
4378              GETCHARLEN(c, eptr, len);              GETCHARLEN(c, eptr, len);
4379              prop_category = _pcre_ucp_findprop(c, &prop_chartype, &prop_script);              prop_chartype = UCD_CHARTYPE(c);
4380              if ((prop_chartype == prop_value) == prop_fail_result)              if ((prop_chartype == prop_value) == prop_fail_result)
4381                break;                break;
4382              eptr+= len;              eptr+= len;
# Line 3276  for (;;) Line 4389  for (;;)
4389              int len = 1;              int len = 1;
4390              if (eptr >= md->end_subject) break;              if (eptr >= md->end_subject) break;
4391              GETCHARLEN(c, eptr, len);              GETCHARLEN(c, eptr, len);
4392              prop_category = _pcre_ucp_findprop(c, &prop_chartype, &prop_script);              prop_script = UCD_SCRIPT(c);
4393              if ((prop_script == prop_value) == prop_fail_result)              if ((prop_script == prop_value) == prop_fail_result)
4394                break;                break;
4395              eptr+= len;              eptr+= len;
# Line 3292  for (;;) Line 4405  for (;;)
4405            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM44);            RMATCH(eptr, ecode, offset_top, md, ims, eptrb, 0, RM44);
4406            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
4407            if (eptr-- == pp) break;        /* Stop if tried at original pos */            if (eptr-- == pp) break;        /* Stop if tried at original pos */
4408            BACKCHAR(eptr);            if (utf8) BACKCHAR(eptr);
4409            }            }
4410          }          }
4411    
# Line 3305  for (;;) Line 4418  for (;;)
4418            {            {
4419            if (eptr >= md->end_subject) break;            if (eptr >= md->end_subject) break;
4420            GETCHARINCTEST(c, eptr);            GETCHARINCTEST(c, eptr);
4421            prop_category = _pcre_ucp_findprop(c, &prop_chartype, &prop_script);            prop_category = UCD_CATEGORY(c);
4422            if (prop_category == ucp_M) break;            if (prop_category == ucp_M) break;
4423            while (eptr < md->end_subject)            while (eptr < md->end_subject)
4424              {              {
# Line 3314  for (;;) Line 4427  for (;;)
4427                {                {
4428                GETCHARLEN(c, eptr, len);                GETCHARLEN(c, eptr, len);
4429                }                }
4430              prop_category = _pcre_ucp_findprop(c, &prop_chartype, &prop_script);              prop_category = UCD_CATEGORY(c);
4431              if (prop_category != ucp_M) break;              if (prop_category != ucp_M) break;
4432              eptr += len;              eptr += len;
4433              }              }
# Line 3331  for (;;) Line 4444  for (;;)
4444            for (;;)                        /* Move back over one extended */            for (;;)                        /* Move back over one extended */
4445              {              {
4446              int len = 1;              int len = 1;
             BACKCHAR(eptr);  
4447              if (!utf8) c = *eptr; else              if (!utf8) c = *eptr; else
4448                {                {
4449                  BACKCHAR(eptr);
4450                GETCHARLEN(c, eptr, len);                GETCHARLEN(c, eptr, len);
4451                }                }
4452              prop_category = _pcre_ucp_findprop(c, &prop_chartype, &prop_script);              prop_category = UCD_CATEGORY(c);
4453              if (prop_category != ucp_M) break;              if (prop_category != ucp_M) break;
4454              eptr--;              eptr--;
4455              }              }
# Line 3354  for (;;) Line 4467  for (;;)
4467          switch(ctype)          switch(ctype)
4468            {            {
4469            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. */  
   
4470            if (max < INT_MAX)            if (max < INT_MAX)
4471              {              {
4472              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  
4473                {                {
4474                for (i = min; i < max; i++)                if (eptr >= md->end_subject || IS_NEWLINE(eptr)) break;
4475                  {                eptr++;
4476                  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++;  
                 }  
4477                }                }
4478              }              }
4479    
# Line 3385  for (;;) Line 4481  for (;;)
4481    
4482            else            else
4483              {              {
4484              if ((ims & PCRE_DOTALL) == 0)              for (i = min; i < max; i++)
4485                {                {
4486                for (i = min; i < max; i++)                if (eptr >= md->end_subject || IS_NEWLINE(eptr)) break;
4487                  {                eptr++;
4488                  if (eptr >= md->end_subject || IS_NEWLINE(eptr)) break;                while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;
                 eptr++;  
                 }  
               break;  
4489                }                }
4490              else              }
4491              break;
4492    
4493              case OP_ALLANY:
4494              if (max < INT_MAX)
4495                {
4496                for (i = min; i < max; i++)
4497                {                {
4498                c = max - min;                if (eptr >= md->end_subject) break;
4499                if (c > (unsigned int)(md->end_subject - eptr))                eptr++;
4500                  c = md->end_subject - eptr;                while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;
               eptr += c;  
4501                }                }
4502              }              }
4503              else eptr = md->end_subject;   /* Unlimited UTF-8 repeat */
4504            break;            break;
4505    
4506            /* The byte case is the same as non-UTF8 */            /* The byte case is the same as non-UTF8 */
# Line 3426  for (;;) Line 4525  for (;;)
4525                }                }
4526              else              else
4527                {                {
4528                if (c != 0x000a && c != 0x000b && c != 0x000c &&                if (c != 0x000a &&
4529                    c != 0x0085 && c != 0x2028 && c != 0x2029)                    (md->bsr_anycrlf ||
4530                       (c != 0x000b && c != 0x000c &&
4531                        c != 0x0085 && c != 0x2028 && c != 0x2029)))
4532                  break;                  break;
4533                eptr += len;                eptr += len;
4534                }                }
4535              }              }
4536            break;            break;
4537    
4538              case OP_NOT_HSPACE:
4539              case OP_HSPACE:
4540              for (i = min; i < max; i++)
4541                {
4542                BOOL gotspace;
4543                int len = 1;
4544                if (eptr >= md->end_subject) break;
4545                GETCHARLEN(c, eptr, len);
4546                switch(c)
4547                  {
4548                  default: gotspace = FALSE; break;
4549                  case 0x09:      /* HT */
4550                  case 0x20:      /* SPACE */
4551                  case 0xa0:      /* NBSP */
4552                  case 0x1680:    /* OGHAM SPACE MARK */
4553                  case 0x180e:    /* MONGOLIAN VOWEL SEPARATOR */
4554                  case 0x2000:    /* EN QUAD */
4555                  case 0x2001:    /* EM QUAD */
4556                  case 0x2002:    /* EN SPACE */
4557                  case 0x2003:    /* EM SPACE */
4558                  case 0x2004:    /* THREE-PER-EM SPACE */
4559                  case 0x2005:    /* FOUR-PER-EM SPACE */
4560                  case 0x2006:    /* SIX-PER-EM SPACE */
4561                  case 0x2007:    /* FIGURE SPACE */
4562                  case 0x2008:    /* PUNCTUATION SPACE */
4563                  case 0x2009:    /* THIN SPACE */
4564                  case 0x200A:    /* HAIR SPACE */
4565                  case 0x202f:    /* NARROW NO-BREAK SPACE */
4566                  case 0x205f:    /* MEDIUM MATHEMATICAL SPACE */
4567                  case 0x3000:    /* IDEOGRAPHIC SPACE */
4568                  gotspace = TRUE;
4569                  break;
4570                  }
4571                if (gotspace == (ctype == OP_NOT_HSPACE)) break;
4572                eptr += len;
4573                }
4574              break;
4575    
4576              case OP_NOT_VSPACE:
4577              case OP_VSPACE:
4578              for (i = min; i < max; i++)
4579                {
4580                BOOL gotspace;
4581                int len = 1;
4582                if (eptr >= md->end_subject) break;
4583                GETCHARLEN(c, eptr, len);
4584                switch(c)
4585                  {
4586                  default: gotspace = FALSE; break;
4587                  case 0x0a:      /* LF */
4588                  case 0x0b:      /* VT */
4589                  case 0x0c:      /* FF */
4590                  case 0x0d:      /* CR */
4591                  case 0x85:      /* NEL */
4592                  case 0x2028:    /* LINE SEPARATOR */
4593                  case 0x2029:    /* PARAGRAPH SEPARATOR */
4594                  gotspace = TRUE;
4595                  break;
4596                  }
4597                if (gotspace == (ctype == OP_NOT_VSPACE)) break;
4598                eptr += len;
4599                }
4600              break;
4601    
4602            case OP_NOT_DIGIT:            case OP_NOT_DIGIT:
4603            for (i = min; i < max; i++)            for (i = min; i < max; i++)
4604              {              {
# Line 3516  for (;;) Line 4681  for (;;)
4681            }            }
4682          }          }
4683        else        else
4684  #endif  #endif  /* SUPPORT_UTF8 */
4685    
4686        /* Not UTF-8 mode */        /* Not UTF-8 mode */
4687          {          {
4688          switch(ctype)          switch(ctype)
4689            {            {
4690            case OP_ANY:            case OP_ANY:
4691            if ((ims & PCRE_DOTALL) == 0)            for (i = min; i < max; i++)
4692              {              {
4693              for (i = min; i < max; i++)              if (eptr >= md->end_subject || IS_NEWLINE(eptr)) break;
4694                {              eptr++;
               if (eptr >= md->end_subject || IS_NEWLINE(eptr)) break;  
               eptr++;  
               }  
             break;  
4695              }              }
4696            /* For DOTALL case, fall through and treat as \C */            break;
4697    
4698              case OP_ALLANY:
4699            case OP_ANYBYTE:            case OP_ANYBYTE:
4700            c = max - min;            c = max - min;
4701            if (c > (unsigned int)(md->end_subject - eptr))            if (c > (unsigned int)(md->end_subject - eptr))
# Line 3553  for (;;) Line 4715  for (;;)
4715                }                }
4716              else              else
4717                {                {
4718                if (c != 0x000a && c != 0x000b && c != 0x000c && c != 0x0085)                if (c != 0x000a &&
4719                      (md->bsr_anycrlf ||
4720                        (c != 0x000b && c != 0x000c && c != 0x0085)))
4721                  break;                  break;
4722                eptr++;                eptr++;
4723                }                }
4724              }              }
4725            break;            break;
4726    
4727              case OP_NOT_HSPACE:
4728              for (i = min; i < max; i++)
4729                {
4730                if (eptr >= md->end_subject) break;
4731                c = *eptr;
4732                if (c == 0x09 || c == 0x20 || c == 0xa0) break;
4733                eptr++;
4734                }
4735              break;
4736    
4737              case OP_HSPACE:
4738              for (i = min; i < max; i++)
4739                {
4740                if (eptr >= md->end_subject) break;
4741                c = *eptr;
4742                if (c != 0x09 && c != 0x20 && c != 0xa0) break;
4743                eptr++;
4744                }
4745              break;
4746    
4747              case OP_NOT_VSPACE:
4748              for (i = min; i < max; i++)
4749                {
4750                if (eptr >= md->end_subject) break;
4751                c = *eptr;
4752                if (c == 0x0a || c == 0x0b || c == 0x0c || c == 0x0d || c == 0x85)
4753                  break;
4754                eptr++;
4755                }
4756              break;
4757    
4758              case OP_VSPACE:
4759              for (i = min; i < max; i++)
4760                {
4761                if (eptr >= md->end_subject) break;
4762                c = *eptr;
4763                if (c != 0x0a && c != 0x0b && c != 0x0c && c != 0x0d && c != 0x85)
4764                  break;
4765                eptr++;
4766                }
4767              break;
4768    
4769            case OP_NOT_DIGIT:            case OP_NOT_DIGIT:
4770            for (i = min; i < max; i++)            for (i = min; i < max; i++)
4771              {              {
# Line 3661  HEAP_RETURN: Line 4867  HEAP_RETURN:
4867  switch (frame->Xwhere)  switch (frame->Xwhere)
4868    {    {
4869    LBL( 1) LBL( 2) LBL( 3) LBL( 4) LBL( 5) LBL( 6) LBL( 7) LBL( 8)    LBL( 1) LBL( 2) LBL( 3) LBL( 4) LBL( 5) LBL( 6) LBL( 7) LBL( 8)
4870    LBL( 9) LBL(10) LBL(11) LBL(12) LBL(13) LBL(14) LBL(15) LBL(16)    LBL( 9) LBL(10) LBL(11) LBL(12) LBL(13) LBL(14) LBL(15) LBL(17)
4871    LBL(17) LBL(18) LBL(19) LBL(20) LBL(21) LBL(22) LBL(23) LBL(24)    LBL(19) LBL(24) LBL(25) LBL(26) LBL(27) LBL(29) LBL(31) LBL(33)
4872    LBL(25) LBL(26) LBL(27) LBL(28) LBL(29) LBL(30) LBL(31) LBL(32)    LBL(35) LBL(43) LBL(47) LBL(48) LBL(49) LBL(50) LBL(51) LBL(52)
4873    LBL(33) LBL(34) LBL(35) LBL(36) LBL(37) LBL(38) LBL(39) LBL(40)    LBL(53) LBL(54)
4874    LBL(41) LBL(42) LBL(43) LBL(44) LBL(45) LBL(46) LBL(47)  #ifdef SUPPORT_UTF8
4875      LBL(16) LBL(18) LBL(20) LBL(21) LBL(22) LBL(23) LBL(28) LBL(30)
4876      LBL(32) LBL(34) LBL(42) LBL(46)
4877    #ifdef SUPPORT_UCP
4878      LBL(36) LBL(37) LBL(38) LBL(39) LBL(40) LBL(41) LBL(44) LBL(45)
4879    #endif  /* SUPPORT_UCP */
4880    #endif  /* SUPPORT_UTF8 */
4881    default:    default:
4882    DPRINTF(("jump error in pcre match: label %d non-existent\n", frame->Xwhere));    DPRINTF(("jump error in pcre match: label %d non-existent\n", frame->Xwhere));
4883    return PCRE_ERROR_INTERNAL;    return PCRE_ERROR_INTERNAL;
# Line 3684  Undefine all the macros that were define Line 4896  Undefine all the macros that were define
4896  #ifdef NO_RECURSE  #ifdef NO_RECURSE
4897  #undef eptr  #undef eptr
4898  #undef ecode  #undef ecode
4899    #undef mstart
4900  #undef offset_top  #undef offset_top
4901  #undef ims  #undef ims
4902  #undef eptrb  #undef eptrb
# Line 3756  Returns:          > 0 => success; value Line 4969  Returns:          > 0 => success; value
4969                   < -1 => some kind of unexpected problem                   < -1 => some kind of unexpected problem
4970  */  */
4971    
4972  PCRE_EXP_DEFN int  PCRE_EXP_DEFN int PCRE_CALL_CONVENTION
4973  pcre_exec(const pcre *argument_re, const pcre_extra *extra_data,  pcre_exec(const pcre *argument_re, const pcre_extra *extra_data,
4974    PCRE_SPTR subject, int length, int start_offset, int options, int *offsets,    PCRE_SPTR subject, int length, int start_offset, int options, int *offsets,
4975    int offsetcount)    int offsetcount)
# Line 3780  const uschar *tables; Line 4993  const uschar *tables;
4993  const uschar *start_bits = NULL;  const uschar *start_bits = NULL;
4994  USPTR start_match = (USPTR)subject + start_offset;  USPTR start_match = (USPTR)subject + start_offset;
4995  USPTR end_subject;  USPTR end_subject;
4996    USPTR start_partial = NULL;
4997  USPTR req_byte_ptr = start_match - 1;  USPTR req_byte_ptr = start_match - 1;
 eptrblock eptrchain[EPTR_WORK_SIZE];  
4998    
4999  pcre_study_data internal_study;  pcre_study_data internal_study;
5000  const pcre_study_data *study;  const pcre_study_data *study;
# Line 3797  if (re == NULL || subject == NULL || Line 5010  if (re == NULL || subject == NULL ||
5010     (offsets == NULL && offsetcount > 0)) return PCRE_ERROR_NULL;     (offsets == NULL && offsetcount > 0)) return PCRE_ERROR_NULL;
5011  if (offsetcount < 0) return PCRE_ERROR_BADCOUNT;  if (offsetcount < 0) return PCRE_ERROR_BADCOUNT;
5012    
5013    /* This information is for finding all the numbers associated with a given
5014    name, for condition testing. */
5015    
5016    md->name_table = (uschar *)re + re->name_table_offset;
5017    md->name_count = re->name_count;
5018    md->name_entry_size = re->name_entry_size;
5019    
5020  /* Fish out the optional data from the extra_data structure, first setting  /* Fish out the optional data from the extra_data structure, first setting
5021  the default values. */  the default values. */
5022    
# Line 3844  if (re->magic_number != MAGIC_NUMBER) Line 5064  if (re->magic_number != MAGIC_NUMBER)
5064  /* Set up other data */  /* Set up other data */
5065    
5066  anchored = ((re->options | options) & PCRE_ANCHORED) != 0;  anchored = ((re->options | options) & PCRE_ANCHORED) != 0;
5067  startline = (re->options & PCRE_STARTLINE) != 0;  startline = (re->flags & PCRE_STARTLINE) != 0;
5068  firstline = (re->options & PCRE_FIRSTLINE) != 0;  firstline = (re->options & PCRE_FIRSTLINE) != 0;
5069    
5070  /* The code starts after the real_pcre block and the capture name table. */  /* The code starts after the real_pcre block and the capture name table. */
# Line 3859  end_subject = md->end_subject; Line 5079  end_subject = md->end_subject;
5079    
5080  md->endonly = (re->options & PCRE_DOLLAR_ENDONLY) != 0;  md->endonly = (re->options & PCRE_DOLLAR_ENDONLY) != 0;
5081  utf8 = md->utf8 = (re->options & PCRE_UTF8) != 0;  utf8 = md->utf8 = (re->options & PCRE_UTF8) != 0;
5082    md->jscript_compat = (re->options & PCRE_JAVASCRIPT_COMPAT) != 0;
5083    
5084  md->notbol = (options & PCRE_NOTBOL) != 0;  md->notbol = (options & PCRE_NOTBOL) != 0;
5085  md->noteol = (options & PCRE_NOTEOL) != 0;  md->noteol = (options & PCRE_NOTEOL) != 0;
5086  md->notempty = (options & PCRE_NOTEMPTY) != 0;  md->notempty = (options & PCRE_NOTEMPTY) != 0;
5087  md->partial = (options & PCRE_PARTIAL) != 0;  md->notempty_atstart = (options & PCRE_NOTEMPTY_ATSTART) != 0;
5088    md->partial = ((options & PCRE_PARTIAL_HARD) != 0)? 2 :
5089                  ((options & PCRE_PARTIAL_SOFT) != 0)? 1 : 0;
5090  md->hitend = FALSE;  md->hitend = FALSE;
5091    
5092  md->recursive = NULL;                   /* No recursion at top level */  md->recursive = NULL;                   /* No recursion at top level */
 md->eptrchain = eptrchain;              /* Make workspace generally available */  
5093    
5094  md->lcc = tables + lcc_offset;  md->lcc = tables + lcc_offset;
5095  md->ctypes = tables + ctypes_offset;  md->ctypes = tables + ctypes_offset;
5096    
5097    /* Handle different \R options. */
5098    
5099    switch (options & (PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE))
5100      {
5101      case 0:
5102      if ((re->options & (PCRE_BSR_ANYCRLF|PCRE_BSR_UNICODE)) != 0)
5103        md->bsr_anycrlf = (re->options & PCRE_BSR_ANYCRLF) != 0;
5104      else
5105    #ifdef BSR_ANYCRLF
5106      md->bsr_anycrlf = TRUE;
5107    #else
5108      md->bsr_anycrlf = FALSE;
5109    #endif
5110      break;
5111    
5112      case PCRE_BSR_ANYCRLF:
5113      md->bsr_anycrlf = TRUE;
5114      break;
5115    
5116      case PCRE_BSR_UNICODE:
5117      md->bsr_anycrlf = FALSE;
5118      break;
5119    
5120      default: return PCRE_ERROR_BADNEWLINE;
5121      }
5122    
5123  /* Handle different types of newline. The three bits give eight cases. If  /* Handle different types of newline. The three bits give eight cases. If
5124  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. */
5125    
5126  switch ((((options & PCRE_NEWLINE_BITS) == 0)? re->options : (pcre_uint32)options) &  switch ((((options & PCRE_NEWLINE_BITS) == 0)? re->options :
5127         PCRE_NEWLINE_BITS)          (pcre_uint32)options) & PCRE_NEWLINE_BITS)
5128    {    {
5129    case 0: newline = NEWLINE; break;   /* Compile-time default */    case 0: newline = NEWLINE; break;   /* Compile-time default */
5130    case PCRE_NEWLINE_CR: newline = '\r'; break;    case PCRE_NEWLINE_CR: newline = CHAR_CR; break;
5131    case PCRE_NEWLINE_LF: newline = '\n'; break;    case PCRE_NEWLINE_LF: newline = CHAR_NL; break;
5132    case PCRE_NEWLINE_CR+    case PCRE_NEWLINE_CR+
5133         PCRE_NEWLINE_LF: newline = ('\r' << 8) | '\n'; break;         PCRE_NEWLINE_LF: newline = (CHAR_CR << 8) | CHAR_NL; break;
5134    case PCRE_NEWLINE_ANY: newline = -1; break;    case PCRE_NEWLINE_ANY: newline = -1; break;
5135    case PCRE_NEWLINE_ANYCRLF: newline = -2; break;    case PCRE_NEWLINE_ANYCRLF: newline = -2; break;
5136    default: return PCRE_ERROR_BADNEWLINE;    default: return PCRE_ERROR_BADNEWLINE;
# Line 3912  else Line 5160  else
5160      }      }
5161    }    }
5162    
5163  /* Partial matching is supported only for a restricted set of regexes at the  /* Partial matching was originally supported only for a restricted set of
5164  moment. */  regexes; from release 8.00 there are no restrictions, but the bits are still
5165    defined (though never set). So there's no harm in leaving this code. */
5166    
5167  if (md->partial && (re->options & PCRE_NOPARTIAL) != 0)  if (md->partial && (re->flags & PCRE_NOPARTIAL) != 0)
5168    return PCRE_ERROR_BADPARTIAL;    return PCRE_ERROR_BADPARTIAL;
5169    
5170  /* 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 3924  back the character offset. */ Line 5173  back the character offset. */
5173  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF8
5174  if (utf8 && (options & PCRE_NO_UTF8_CHECK) == 0)  if (utf8 && (options & PCRE_NO_UTF8_CHECK) == 0)
5175    {    {
5176    if (_pcre_valid_utf8((uschar *)subject, length) >= 0)    if (_pcre_valid_utf8((USPTR)subject, length) >= 0)
5177      return PCRE_ERROR_BADUTF8;      return PCRE_ERROR_BADUTF8;
5178    if (start_offset > 0 && start_offset < length)    if (start_offset > 0 && start_offset < length)
5179      {      {
5180      int tb = ((uschar *)subject)[start_offset];      int tb = ((USPTR)subject)[start_offset];
5181      if (tb > 127)      if (tb > 127)
5182        {        {
5183        tb &= 0xc0;        tb &= 0xc0;
# Line 3992  studied, there may be a bitmap of possib Line 5241  studied, there may be a bitmap of possib
5241    
5242  if (!anchored)  if (!anchored)
5243    {    {
5244    if ((re->options & PCRE_FIRSTSET) != 0)    if ((re->flags & PCRE_FIRSTSET) != 0)
5245      {      {
5246      first_byte = re->first_byte & 255;      first_byte = re->first_byte & 255;
5247      if ((first_byte_caseless = ((re->first_byte & REQ_CASELESS) != 0)) == TRUE)      if ((first_byte_caseless = ((re->first_byte & REQ_CASELESS) != 0)) == TRUE)
# Line 4000  if (!anchored) Line 5249  if (!anchored)
5249      }      }
5250    else    else
5251      if (!startline && study != NULL &&      if (!startline && study != NULL &&
5252        (study->options & PCRE_STUDY_MAPPED) != 0)        (study->flags & PCRE_STUDY_MAPPED) != 0)
5253          start_bits = study->start_bits;          start_bits = study->start_bits;
5254    }    }
5255    
5256  /* For anchored or unanchored matches, there may be a "last known required  /* For anchored or unanchored matches, there may be a "last known required
5257  character" set. */  character" set. */
5258    
5259  if ((re->options & PCRE_REQCHSET) != 0)  if ((re->flags & PCRE_REQCHSET) != 0)
5260    {    {
5261    req_byte = re->req_byte & 255;    req_byte = re->req_byte & 255;
5262    req_byte_caseless = (re->req_byte & REQ_CASELESS) != 0;    req_byte_caseless = (re->req_byte & REQ_CASELESS) != 0;
# Line 4023  the loop runs just once. */ Line 5272  the loop runs just once. */
5272  for(;;)  for(;;)
5273    {    {
5274    USPTR save_end_subject = end_subject;    USPTR save_end_subject = end_subject;
5275      USPTR new_start_match;
5276    
5277    /* Reset the maximum number of extractions we might see. */    /* Reset the maximum number of extractions we might see. */
5278    
# Line 4033  for(;;) Line 5283  for(;;)
5283      while (iptr < iend) *iptr++ = -1;      while (iptr < iend) *iptr++ = -1;
5284      }      }
5285    
5286    /* Advance to a unique first char if possible. If firstline is TRUE, the    /* If firstline is TRUE, the start of the match is constrained to the first
5287    start of the match is constrained to the first line of a multiline string.    line of a multiline string. That is, the match must be before or at the first
5288    That is, the match must be before or at the first newline. Implement this by    newline. Implement this by temporarily adjusting end_subject so that we stop
5289    temporarily adjusting end_subject so that we stop scanning at a newline. If    scanning at a newline. If the match fails at the newline, later code breaks
5290    the match fails at the newline, later code breaks this loop. */    this loop. */
5291    
5292    if (firstline)    if (firstline)
5293      {      {
5294      USPTR t = start_match;      USPTR t = start_match;
5295    #ifdef SUPPORT_UTF8
5296        if (utf8)
5297          {
5298          while (t < md->end_subject && !IS_NEWLINE(t))
5299            {
5300            t++;
5301            while (t < end_subject && (*t & 0xc0) == 0x80) t++;
5302            }
5303          }
5304        else
5305    #endif
5306      while (t < md->end_subject && !IS_NEWLINE(t)) t++;      while (t < md->end_subject && !IS_NEWLINE(t)) t++;
5307      end_subject = t;      end_subject = t;
5308      }      }
5309    
5310    /* Now test for a unique first byte */    /* There are some optimizations that avoid running the match if a known
5311      starting point is not found, or if a known later character is not present.
5312      However, there is an option that disables these, for testing and for ensuring
5313      that all callouts do actually occur. */
5314    
5315    if (first_byte >= 0)    if ((options & PCRE_NO_START_OPTIMIZE) == 0)
5316      {      {
5317      if (first_byte_caseless)      /* Advance to a unique first byte if there is one. */
5318        while (start_match < end_subject &&  
5319               md->lcc[*start_match] != first_byte)      if (first_byte >= 0)
5320          start_match++;        {
5321      else        if (first_byte_caseless)
5322        while (start_match < end_subject && *start_match != first_byte)          while (start_match < end_subject && md->lcc[*start_match] != first_byte)
5323          start_match++;            start_match++;
5324      }        else
5325            while (start_match < end_subject && *start_match != first_byte)
5326              start_match++;
5327          }
5328    
5329    /* Or to just after a linebreak for a multiline match if possible */      /* Or to just after a linebreak for a multiline match */
5330    
5331    else if (startline)      else if (startline)
     {  
     if (start_match > md->start_subject + start_offset)  
5332        {        {
5333        while (start_match <= end_subject && !WAS_NEWLINE(start_match))        if (start_match > md->start_subject + start_offset)
5334          start_match++;          {
5335    #ifdef SUPPORT_UTF8
5336            if (utf8)
5337              {
5338              while (start_match < end_subject && !WAS_NEWLINE(start_match))
5339                {
5340                start_match++;
5341                while(start_match < end_subject && (*start_match & 0xc0) == 0x80)
5342                  start_match++;
5343                }
5344              }
5345            else
5346    #endif
5347            while (start_match < end_subject && !WAS_NEWLINE(start_match))
5348              start_match++;
5349    
5350            /* If we have just passed a CR and the newline option is ANY or ANYCRLF,
5351            and we are now at a LF, advance the match position by one more character.
5352            */
5353    
5354        /* If we have just passed a CR and the newline option is ANY or ANYCRLF,          if (start_match[-1] == CHAR_CR &&
5355        and we are now at a LF, advance the match position by one more character.               (md->nltype == NLTYPE_ANY || md->nltype == NLTYPE_ANYCRLF) &&
5356        */               start_match < end_subject &&
5357                 *start_match == CHAR_NL)
5358        if (start_match[-1] == '\r' &&            start_match++;
5359             (md->nltype == NLTYPE_ANY || md->nltype == NLTYPE_ANYCRLF) &&          }
            start_match < end_subject &&  
            *start_match == '\n')  
         start_match++;  
5360        }        }
     }  
5361    
5362    /* Or to a non-unique first char after study */      /* Or to a non-unique first byte after study */
5363    
5364    else if (start_bits != NULL)      else if (start_bits != NULL)
     {  
     while (start_match < end_subject)  
5365        {        {
5366        register unsigned int c = *start_match;        while (start_match < end_subject)
5367        if ((start_bits[c/8] & (1 << (c&7))) == 0) start_match++; else break;          {
5368            register unsigned int c = *start_match;
5369            if ((start_bits[c/8] & (1 << (c&7))) == 0) start_match++;
5370              else break;
5371            }
5372        }        }
5373      }      }   /* Starting optimizations */
5374    
5375    /* Restore fudged end_subject */    /* Restore fudged end_subject */
5376    
5377    end_subject = save_end_subject;    end_subject = save_end_subject;
5378    
5379      /* The following two optimizations are disabled for partial matching or if
5380      disabling is explicitly requested. */
5381    
5382      if ((options & PCRE_NO_START_OPTIMIZE) == 0 && !md->partial)
5383        {
5384        /* If the pattern was studied, a minimum subject length may be set. This is
5385        a lower bound; no actual string of that length may actually match the
5386        pattern. Although the value is, strictly, in characters, we treat it as
5387        bytes to avoid spending too much time in this optimization. */
5388    
5389        if (study != NULL && (study->flags & PCRE_STUDY_MINLEN) != 0 &&
5390            end_subject - start_match < study->minlength)
5391          {
5392          rc = MATCH_NOMATCH;
5393          break;
5394          }
5395    
5396        /* If req_byte is set, we know that that character must appear in the
5397        subject for the match to succeed. If the first character is set, req_byte
5398        must be later in the subject; otherwise the test starts at the match point.
5399        This optimization can save a huge amount of backtracking in patterns with
5400        nested unlimited repeats that aren't going to match. Writing separate code
5401        for cased/caseless versions makes it go faster, as does using an
5402        autoincrement and backing off on a match.
5403    
5404        HOWEVER: when the subject string is very, very long, searching to its end
5405        can take a long time, and give bad performance on quite ordinary patterns.
5406        This showed up when somebody was matching something like /^\d+C/ on a
5407        32-megabyte string... so we don't do this when the string is sufficiently
5408        long. */
5409    
5410        if (req_byte >= 0 && end_subject - start_match < REQ_BYTE_MAX)
5411          {
5412          register USPTR p = start_match + ((first_byte >= 0)? 1 : 0);
5413    
5414          /* We don't need to repeat the search if we haven't yet reached the
5415          place we found it at last time. */
5416    
5417          if (p > req_byte_ptr)
5418            {
5419            if (req_byte_caseless)
5420              {
5421              while (p < end_subject)
5422                {
5423                register int pp = *p++;
5424                if (pp == req_byte || pp == req_byte2) { p--; break; }
5425                }
5426              }
5427            else
5428              {
5429              while (p < end_subject)
5430                {
5431                if (*p++ == req_byte) { p--; break; }
5432                }
5433              }
5434    
5435            /* If we can't find the required character, break the matching loop,
5436            forcing a match failure. */
5437    
5438            if (p >= end_subject)
5439              {
5440              rc = MATCH_NOMATCH;
5441              break;
5442              }
5443    
5444            /* If we have found the required character, save the point where we
5445            found it, so that we don't search again next time round the loop if
5446            the start hasn't passed this character yet. */
5447    
5448            req_byte_ptr = p;
5449            }
5450          }
5451        }
5452    
5453  #ifdef DEBUG  /* Sigh. Some compilers never learn. */  #ifdef DEBUG  /* Sigh. Some compilers never learn. */
5454    printf(">>>> Match against: ");    printf(">>>> Match against: ");
# Line 4101  for(;;) Line 5456  for(;;)
5456    printf("\n");    printf("\n");
5457  #endif  #endif
5458    
5459    /* If req_byte is set, we know that that character must appear in the subject    /* OK, we can now run the match. If "hitend" is set afterwards, remember the
5460    for the match to succeed. If the first character is set, req_byte must be    first starting point for which a partial match was found. */
   later in the subject; otherwise the test starts at the match point. This  
   optimization can save a huge amount of backtracking in patterns with nested  
   unlimited repeats that aren't going to match. Writing separate code for  
   cased/caseless versions makes it go faster, as does using an autoincrement  
   and backing off on a match.  
   
   HOWEVER: when the subject string is very, very long, searching to its end can  
   take a long time, and give bad performance on quite ordinary patterns. This  
   showed up when somebody was matching something like /^\d+C/ on a 32-megabyte  
   string... so we don't do this when the string is sufficiently long.  
   
   ALSO: this processing is disabled when partial matching is requested.  
   */  
   
   if (req_byte >= 0 &&  
       end_subject - start_match < REQ_BYTE_MAX &&  
       !md->partial)  
     {  
     register USPTR p = start_match + ((first_byte >= 0)? 1 : 0);  
5461    
5462      /* We don't need to repeat the search if we haven't yet reached the    md->start_match_ptr = start_match;
5463      place we found it at last time. */    md->start_used_ptr = start_match;
5464      md->match_call_count = 0;
5465      rc = match(start_match, md->start_code, start_match, 2, md, ims, NULL, 0, 0);
5466      if (md->hitend && start_partial == NULL) start_partial = md->start_used_ptr;
5467    
5468      if (p > req_byte_ptr)    switch(rc)
5469        {      {
5470        if (req_byte_caseless)      /* NOMATCH and PRUNE advance by one character. THEN at this level acts
5471          {      exactly like PRUNE. */
5472          while (p < end_subject)  
5473            {      case MATCH_NOMATCH:
5474            register int pp = *p++;      case MATCH_PRUNE:
5475            if (pp == req_byte || pp == req_byte2) { p--; break; }      case MATCH_THEN:
5476            }      new_start_match = start_match + 1;
5477          }  #ifdef SUPPORT_UTF8
5478        else      if (utf8)
5479          {        while(new_start_match < end_subject && (*new_start_match & 0xc0) == 0x80)
5480          while (p < end_subject)          new_start_match++;
5481            {  #endif
5482            if (*p++ == req_byte) { p--; break; }      break;
           }  
         }  
5483    
5484        /* If we can't find the required character, break the matching loop,      /* SKIP passes back the next starting point explicitly. */
       forcing a match failure. */  
5485    
5486        if (p >= end_subject)      case MATCH_SKIP:
5487          {      new_start_match = md->start_match_ptr;
5488          rc = MATCH_NOMATCH;      break;
         break;  
         }  
5489    
5490        /* If we have found the required character, save the point where we      /* COMMIT disables the bumpalong, but otherwise behaves as NOMATCH. */
       found it, so that we don't search again next time round the loop if  
       the start hasn't passed this character yet. */  
5491    
5492        req_byte_ptr = p;      case MATCH_COMMIT:
5493        }      rc = MATCH_NOMATCH;
5494      }      goto ENDLOOP;
5495    
5496    /* OK, we can now run the match. */      /* Any other return is either a match, or some kind of error. */
5497    
5498    md->start_match = start_match;      default:
5499    md->match_call_count = 0;      goto ENDLOOP;
5500    md->eptrn = 0;                          /* Next free eptrchain slot */      }
   rc = match(start_match, md->start_code, 2, md, ims, NULL, 0, 0);  
5501    
5502    /* Any return other than MATCH_NOMATCH breaks the loop. */    /* Control reaches here for the various types of "no match at this point"
5503      result. Reset the code to MATCH_NOMATCH for subsequent checking. */
5504    
5505    if (rc != MATCH_NOMATCH) break;    rc = MATCH_NOMATCH;
5506    
5507    /* 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
5508    newline in the subject (though it may continue over the newline). Therefore,    newline in the subject (though it may continue over the newline). Therefore,
# Line 4178  for(;;) Line 5510  for(;;)
5510    
5511    if (firstline && IS_NEWLINE(start_match)) break;    if (firstline && IS_NEWLINE(start_match)) break;
5512    
5513    /* Advance the match position by one character. */    /* Advance to new matching position */
5514    
5515    start_match++;    start_match = new_start_match;
 #ifdef SUPPORT_UTF8  
   if (utf8)  
     while(start_match < end_subject && (*start_match & 0xc0) == 0x80)  
       start_match++;  
 #endif  
5516    
5517    /* 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
5518    the subject. */    the subject. */
5519    
5520    if (anchored || start_match > end_subject) break;    if (anchored || start_match > end_subject) break;
5521    
5522    /* If we have just passed a CR and the newline option is CRLF or ANY or    /* If we have just passed a CR and we are now at a LF, and the pattern does
5523    ANYCRLF, and we are now at a LF, advance the match position by one more    not contain any explicit matches for \r or \n, and the newline option is CRLF
5524    character. */    or ANY or ANYCRLF, advance the match position by one more character. */
5525    
5526    if (start_match[-1] == '\r' &&    if (start_match[-1] == CHAR_CR &&
5527         (md->nltype == NLTYPE_ANY ||        start_match