/[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 895 by ph10, Fri Jan 20 12:12:03 2012 UTC revision 1425 by ph10, Tue Dec 31 17:44:40 2013 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) 1997-2012 University of Cambridge             Copyright (c) 1997-2013 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 37  POSSIBILITY OF SUCH DAMAGE. Line 37  POSSIBILITY OF SUCH DAMAGE.
37  -----------------------------------------------------------------------------  -----------------------------------------------------------------------------
38  */  */
39    
   
40  /* This module contains pcre_exec(), the externally visible function that does  /* This module contains pcre_exec(), the externally visible function that does
41  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
42  possible. There are also some static supporting functions. */  possible. There are also some static supporting functions. */
# Line 57  possible. There are also some static sup Line 56  possible. There are also some static sup
56  #undef min  #undef min
57  #undef max  #undef max
58    
59    /* The md->capture_last field uses the lower 16 bits for the last captured
60    substring (which can never be greater than 65535) and a bit in the top half
61    to mean "capture vector overflowed". This odd way of doing things was
62    implemented when it was realized that preserving and restoring the overflow bit
63    whenever the last capture number was saved/restored made for a neater
64    interface, and doing it this way saved on (a) another variable, which would
65    have increased the stack frame size (a big NO-NO in PCRE) and (b) another
66    separate set of save/restore instructions. The following defines are used in
67    implementing this. */
68    
69    #define CAPLMASK    0x0000ffff    /* The bits used for last_capture */
70    #define OVFLMASK    0xffff0000    /* The bits used for the overflow flag */
71    #define OVFLBIT     0x00010000    /* The bit that is set for overflow */
72    
73  /* Values for setting in md->match_function_type to indicate two special types  /* Values for setting in md->match_function_type to indicate two special types
74  of call to match(). We do it this way to save on using another stack variable,  of call to match(). We do it this way to save on using another stack variable,
75  as stack usage is to be discouraged. */  as stack usage is to be discouraged. */
# Line 74  defined PCRE_ERROR_xxx codes, which are Line 87  defined PCRE_ERROR_xxx codes, which are
87  negative to avoid the external error codes. */  negative to avoid the external error codes. */
88    
89  #define MATCH_ACCEPT       (-999)  #define MATCH_ACCEPT       (-999)
90  #define MATCH_COMMIT       (-998)  #define MATCH_KETRPOS      (-998)
91  #define MATCH_KETRPOS      (-997)  #define MATCH_ONCE         (-997)
92  #define MATCH_ONCE         (-996)  /* The next 5 must be kept together and in sequence so that a test that checks
93    for any one of them can use a range. */
94    #define MATCH_COMMIT       (-996)
95  #define MATCH_PRUNE        (-995)  #define MATCH_PRUNE        (-995)
96  #define MATCH_SKIP         (-994)  #define MATCH_SKIP         (-994)
97  #define MATCH_SKIP_ARG     (-993)  #define MATCH_SKIP_ARG     (-993)
98  #define MATCH_THEN         (-992)  #define MATCH_THEN         (-992)
99    #define MATCH_BACKTRACK_MAX MATCH_THEN
100    #define MATCH_BACKTRACK_MIN MATCH_COMMIT
101    
102  /* 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.
103  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,
# Line 90  because the offset vector is always a mu Line 107  because the offset vector is always a mu
107    
108  /* Min and max values for the common repeats; for the maxima, 0 => infinity */  /* Min and max values for the common repeats; for the maxima, 0 => infinity */
109    
110  static const char rep_min[] = { 0, 0, 1, 1, 0, 0 };  static const char rep_min[] = { 0, 0, 1, 1, 0, 0, 0, 0, 0, 1, 0, };
111  static const char rep_max[] = { 0, 0, 0, 0, 1, 1 };  static const char rep_max[] = { 0, 0, 0, 0, 1, 1, 0, 0, 0, 0, 1, };
   
   
112    
113  #ifdef PCRE_DEBUG  #ifdef PCRE_DEBUG
114  /*************************************************  /*************************************************
# Line 115  Returns:     nothing Line 130  Returns:     nothing
130  static void  static void
131  pchars(const pcre_uchar *p, int length, BOOL is_subject, match_data *md)  pchars(const pcre_uchar *p, int length, BOOL is_subject, match_data *md)
132  {  {
133  unsigned int c;  pcre_uint32 c;
134    BOOL utf = md->utf;
135  if (is_subject && length > md->end_subject - p) length = md->end_subject - p;  if (is_subject && length > md->end_subject - p) length = md->end_subject - p;
136  while (length-- > 0)  while (length-- > 0)
137    if (isprint(c = *(p++))) printf("%c", c); else printf("\\x%02x", c);    if (isprint(c = *p++)) printf("%c", (char)c); else printf("\\x{%02x}", c);
138  }  }
139  #endif  #endif
140    
# Line 140  Arguments: Line 156  Arguments:
156    md          points to match data block    md          points to match data block
157    caseless    TRUE if caseless    caseless    TRUE if caseless
158    
159  Returns:      < 0 if not matched, otherwise the number of subject bytes matched  Returns:      >= 0 the number of subject bytes matched
160                  -1 no match
161                  -2 partial match; always given if at end subject
162  */  */
163    
164  static int  static int
# Line 149  match_ref(int offset, register PCRE_PUCH Line 167  match_ref(int offset, register PCRE_PUCH
167  {  {
168  PCRE_PUCHAR eptr_start = eptr;  PCRE_PUCHAR eptr_start = eptr;
169  register PCRE_PUCHAR p = md->start_subject + md->offset_vector[offset];  register PCRE_PUCHAR p = md->start_subject + md->offset_vector[offset];
170    #if defined SUPPORT_UTF && defined SUPPORT_UCP
171    BOOL utf = md->utf;
172    #endif
173    
174  #ifdef PCRE_DEBUG  #ifdef PCRE_DEBUG
175  if (eptr >= md->end_subject)  if (eptr >= md->end_subject)
# Line 163  pchars(p, length, FALSE, md); Line 184  pchars(p, length, FALSE, md);
184  printf("\n");  printf("\n");
185  #endif  #endif
186    
187  /* Always fail if reference not set (and not JavaScript compatible). */  /* Always fail if reference not set (and not JavaScript compatible - in that
188    case the length is passed as zero). */
189    
190  if (length < 0) return -1;  if (length < 0) return -1;
191    
# Line 173  ASCII characters. */ Line 195  ASCII characters. */
195    
196  if (caseless)  if (caseless)
197    {    {
198  #ifdef SUPPORT_UTF  #if defined SUPPORT_UTF && defined SUPPORT_UCP
199  #ifdef SUPPORT_UCP    if (utf)
   if (md->utf)  
200      {      {
201      /* Match characters up to the end of the reference. NOTE: the number of      /* Match characters up to the end of the reference. NOTE: the number of
202      bytes matched may differ, because there are some characters whose upper and      data units matched may differ, because in UTF-8 there are some characters
203      lower case versions code as different numbers of bytes. For example, U+023A      whose upper and lower case versions code have different numbers of bytes.
204      (2 bytes in UTF-8) is the upper case version of U+2C65 (3 bytes in UTF-8);      For example, U+023A (2 bytes in UTF-8) is the upper case version of U+2C65
205      a sequence of 3 of the former uses 6 bytes, as does a sequence of two of      (3 bytes in UTF-8); a sequence of 3 of the former uses 6 bytes, as does a
206      the latter. It is important, therefore, to check the length along the      sequence of two of the latter. It is important, therefore, to check the
207      reference, not along the subject (earlier code did this wrong). */      length along the reference, not along the subject (earlier code did this
208        wrong). */
209    
210      PCRE_PUCHAR endptr = p + length;      PCRE_PUCHAR endptr = p + length;
211      while (p < endptr)      while (p < endptr)
212        {        {
213        int c, d;        pcre_uint32 c, d;
214        if (eptr >= md->end_subject) return -1;        const ucd_record *ur;
215          if (eptr >= md->end_subject) return -2;   /* Partial match */
216        GETCHARINC(c, eptr);        GETCHARINC(c, eptr);
217        GETCHARINC(d, p);        GETCHARINC(d, p);
218        if (c != d && c != UCD_OTHERCASE(d)) return -1;        ur = GET_UCD(d);
219          if (c != d && c != d + ur->other_case)
220            {
221            const pcre_uint32 *pp = PRIV(ucd_caseless_sets) + ur->caseset;
222            for (;;)
223              {
224              if (c < *pp) return -1;
225              if (c == *pp++) break;
226              }
227            }
228        }        }
229      }      }
230    else    else
231  #endif  #endif
 #endif  
232    
233    /* The same code works when not in UTF-8 mode and in UTF-8 mode when there    /* The same code works when not in UTF-8 mode and in UTF-8 mode when there
234    is no UCP support. */    is no UCP support. */
235      {      {
     if (eptr + length > md->end_subject) return -1;  
236      while (length-- > 0)      while (length-- > 0)
237        {        {
238        if (TABLE_GET(*p, md->lcc, *p) != TABLE_GET(*eptr, md->lcc, *eptr)) return -1;        pcre_uint32 cc, cp;
239          if (eptr >= md->end_subject) return -2;   /* Partial match */
240          cc = *eptr;
241          cp = *p;
242          if (TABLE_GET(cp, md->lcc, cp) != TABLE_GET(cc, md->lcc, cc)) return -1;
243        p++;        p++;
244        eptr++;        eptr++;
245        }        }
# Line 217  are in UTF-8 mode. */ Line 251  are in UTF-8 mode. */
251    
252  else  else
253    {    {
254    if (eptr + length > md->end_subject) return -1;    while (length-- > 0)
255    while (length-- > 0) if (*p++ != *eptr++) return -1;      {
256        if (eptr >= md->end_subject) return -2;   /* Partial match */
257        if (*p++ != *eptr++) return -1;
258        }
259    }    }
260    
261  return (int)(eptr - eptr_start);  return (int)(eptr - eptr_start);
# Line 273  enum { RM1=1, RM2,  RM3,  RM4,  RM5,  RM Line 310  enum { RM1=1, RM2,  RM3,  RM4,  RM5,  RM
310         RM31,  RM32, RM33, RM34, RM35, RM36, RM37, RM38, RM39, RM40,         RM31,  RM32, RM33, RM34, RM35, RM36, RM37, RM38, RM39, RM40,
311         RM41,  RM42, RM43, RM44, RM45, RM46, RM47, RM48, RM49, RM50,         RM41,  RM42, RM43, RM44, RM45, RM46, RM47, RM48, RM49, RM50,
312         RM51,  RM52, RM53, RM54, RM55, RM56, RM57, RM58, RM59, RM60,         RM51,  RM52, RM53, RM54, RM55, RM56, RM57, RM58, RM59, RM60,
313         RM61,  RM62, RM63, RM64, RM65, RM66 };         RM61,  RM62, RM63, RM64, RM65, RM66, RM67 };
314    
315  /* 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
316  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 291  actually used in this definition. */ Line 328  actually used in this definition. */
328    }    }
329  #define RRETURN(ra) \  #define RRETURN(ra) \
330    { \    { \
331    printf("match() returned %d from line %d ", ra, __LINE__); \    printf("match() returned %d from line %d\n", ra, __LINE__); \
332    return ra; \    return ra; \
333    }    }
334  #else  #else
# Line 311  argument of match(), which never changes Line 348  argument of match(), which never changes
348    
349  #define RMATCH(ra,rb,rc,rd,re,rw)\  #define RMATCH(ra,rb,rc,rd,re,rw)\
350    {\    {\
351    heapframe *newframe = (heapframe *)(PUBL(stack_malloc))(sizeof(heapframe));\    heapframe *newframe = frame->Xnextframe;\
352    if (newframe == NULL) RRETURN(PCRE_ERROR_NOMEMORY);\    if (newframe == NULL)\
353    frame->Xwhere = rw; \      {\
354        newframe = (heapframe *)(PUBL(stack_malloc))(sizeof(heapframe));\
355        if (newframe == NULL) RRETURN(PCRE_ERROR_NOMEMORY);\
356        newframe->Xnextframe = NULL;\
357        frame->Xnextframe = newframe;\
358        }\
359      frame->Xwhere = rw;\
360    newframe->Xeptr = ra;\    newframe->Xeptr = ra;\
361    newframe->Xecode = rb;\    newframe->Xecode = rb;\
362    newframe->Xmstart = mstart;\    newframe->Xmstart = mstart;\
# Line 332  argument of match(), which never changes Line 375  argument of match(), which never changes
375    {\    {\
376    heapframe *oldframe = frame;\    heapframe *oldframe = frame;\
377    frame = oldframe->Xprevframe;\    frame = oldframe->Xprevframe;\
   if (oldframe != &frame_zero) (PUBL(stack_free))(oldframe);\  
378    if (frame != NULL)\    if (frame != NULL)\
379      {\      {\
380      rrc = ra;\      rrc = ra;\
# Line 346  argument of match(), which never changes Line 388  argument of match(), which never changes
388    
389  typedef struct heapframe {  typedef struct heapframe {
390    struct heapframe *Xprevframe;    struct heapframe *Xprevframe;
391      struct heapframe *Xnextframe;
392    
393    /* Function arguments that may change */    /* Function arguments that may change */
394    
# Line 376  typedef struct heapframe { Line 419  typedef struct heapframe {
419    
420  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
421    int Xprop_type;    int Xprop_type;
422    int Xprop_value;    unsigned int Xprop_value;
423    int Xprop_fail_result;    int Xprop_fail_result;
424    int Xoclength;    int Xoclength;
425    pcre_uchar Xocchars[6];    pcre_uchar Xocchars[6];
# Line 389  typedef struct heapframe { Line 432  typedef struct heapframe {
432    int Xlength;    int Xlength;
433    int Xmax;    int Xmax;
434    int Xmin;    int Xmin;
435    int Xnumber;    unsigned int Xnumber;
436    int Xoffset;    int Xoffset;
437    int Xop;    unsigned int Xop;
438    int Xsave_capture_last;    pcre_int32 Xsave_capture_last;
439    int Xsave_offset1, Xsave_offset2, Xsave_offset3;    int Xsave_offset1, Xsave_offset2, Xsave_offset3;
440    int Xstacksave[REC_STACK_SAVE_MAX];    int Xstacksave[REC_STACK_SAVE_MAX];
441    
# Line 477  so they can be ordinary variables in all Line 520  so they can be ordinary variables in all
520    
521  register int  rrc;         /* Returns from recursive calls */  register int  rrc;         /* Returns from recursive calls */
522  register int  i;           /* Used for loops not involving calls to RMATCH() */  register int  i;           /* Used for loops not involving calls to RMATCH() */
523  register unsigned int c;   /* Character values not kept over RMATCH() calls */  register pcre_uint32 c;    /* Character values not kept over RMATCH() calls */
524  register BOOL utf;         /* Local copy of UTF flag for speed */  register BOOL utf;         /* Local copy of UTF flag for speed */
525    
526  BOOL minimize, possessive; /* Quantifier options */  BOOL minimize, possessive; /* Quantifier options */
# Line 487  int condcode; Line 530  int condcode;
530  /* 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
531  preserved over calls to RMATCH() are part of a "frame". We set up the top-level  preserved over calls to RMATCH() are part of a "frame". We set up the top-level
532  frame on the stack here; subsequent instantiations are obtained from the heap  frame on the stack here; subsequent instantiations are obtained from the heap
533  whenever RMATCH() does a "recursion". See the macro definitions above. Putting  whenever RMATCH() does a "recursion". See the macro definitions above. Putting
534  the top-level on the stack rather than malloc-ing them all gives a performance  the top-level on the stack rather than malloc-ing them all gives a performance
535  boost in many cases where there is not much "recursion". */  boost in many cases where there is not much "recursion". */
536    
537  #ifdef NO_RECURSE  #ifdef NO_RECURSE
538  heapframe frame_zero;  heapframe *frame = (heapframe *)md->match_frames_base;
 heapframe *frame = &frame_zero;  
 frame->Xprevframe = NULL;            /* Marks the top level */  
539    
540  /* Copy in the original argument variables */  /* Copy in the original argument variables */
541    
# Line 596  BOOL prev_is_word; Line 637  BOOL prev_is_word;
637    
638  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
639  int prop_type;  int prop_type;
640  int prop_value;  unsigned int prop_value;
641  int prop_fail_result;  int prop_fail_result;
642  int oclength;  int oclength;
643  pcre_uchar occhars[6];  pcre_uchar occhars[6];
# Line 607  int ctype; Line 648  int ctype;
648  int length;  int length;
649  int max;  int max;
650  int min;  int min;
651  int number;  unsigned int number;
652  int offset;  int offset;
653  int op;  unsigned int op;
654  int save_capture_last;  pcre_int32 save_capture_last;
655  int save_offset1, save_offset2, save_offset3;  int save_offset1, save_offset2, save_offset3;
656  int stacksave[REC_STACK_SAVE_MAX];  int stacksave[REC_STACK_SAVE_MAX];
657    
658  eptrblock newptrb;  eptrblock newptrb;
659    
660  /* There is a special fudge for calling match() in a way that causes it to  /* There is a special fudge for calling match() in a way that causes it to
661  measure the size of its basic stack frame when the stack is being used for  measure the size of its basic stack frame when the stack is being used for
662  recursion. The second argument (ecode) being NULL triggers this behaviour. It  recursion. The second argument (ecode) being NULL triggers this behaviour. It
663  cannot normally every be NULL. The return is the negated value of the frame  cannot normally ever be NULL. The return is the negated value of the frame
664  size. */  size. */
665    
666  if (ecode == NULL)  if (ecode == NULL)
# Line 631  if (ecode == NULL) Line 672  if (ecode == NULL)
672      int len = (char *)&rdepth - (char *)eptr;      int len = (char *)&rdepth - (char *)eptr;
673      return (len > 0)? -len : len;      return (len > 0)? -len : len;
674      }      }
675    }    }
676  #endif     /* NO_RECURSE */  #endif     /* NO_RECURSE */
677    
678  /* To save space on the stack and in the heap frame, I have doubled up on some  /* To save space on the stack and in the heap frame, I have doubled up on some
# Line 728  for (;;) Line 769  for (;;)
769      unaltered. */      unaltered. */
770    
771      else if (rrc == MATCH_SKIP_ARG &&      else if (rrc == MATCH_SKIP_ARG &&
772          STRCMP_UC_UC(ecode + 2, md->start_match_ptr) == 0)          STRCMP_UC_UC_TEST(ecode + 2, md->start_match_ptr) == 0)
773        {        {
774        md->start_match_ptr = eptr;        md->start_match_ptr = eptr;
775        RRETURN(MATCH_SKIP);        RRETURN(MATCH_SKIP);
# Line 738  for (;;) Line 779  for (;;)
779      case OP_FAIL:      case OP_FAIL:
780      RRETURN(MATCH_NOMATCH);      RRETURN(MATCH_NOMATCH);
781    
     /* COMMIT overrides PRUNE, SKIP, and THEN */  
   
782      case OP_COMMIT:      case OP_COMMIT:
783      RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md,      RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md,
784        eptrb, RM52);        eptrb, RM52);
785      if (rrc != MATCH_NOMATCH && rrc != MATCH_PRUNE &&      if (rrc != MATCH_NOMATCH) RRETURN(rrc);
         rrc != MATCH_SKIP && rrc != MATCH_SKIP_ARG &&  
         rrc != MATCH_THEN)  
       RRETURN(rrc);  
786      RRETURN(MATCH_COMMIT);      RRETURN(MATCH_COMMIT);
787    
     /* PRUNE overrides THEN */  
   
788      case OP_PRUNE:      case OP_PRUNE:
789      RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md,      RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md,
790        eptrb, RM51);        eptrb, RM51);
791      if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN) RRETURN(rrc);      if (rrc != MATCH_NOMATCH) RRETURN(rrc);
792      RRETURN(MATCH_PRUNE);      RRETURN(MATCH_PRUNE);
793    
794      case OP_PRUNE_ARG:      case OP_PRUNE_ARG:
# Line 764  for (;;) Line 798  for (;;)
798        eptrb, RM56);        eptrb, RM56);
799      if ((rrc == MATCH_MATCH || rrc == MATCH_ACCEPT) &&      if ((rrc == MATCH_MATCH || rrc == MATCH_ACCEPT) &&
800           md->mark == NULL) md->mark = ecode + 2;           md->mark == NULL) md->mark = ecode + 2;
801      if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN) RRETURN(rrc);      if (rrc != MATCH_NOMATCH) RRETURN(rrc);
802      RRETURN(MATCH_PRUNE);      RRETURN(MATCH_PRUNE);
803    
     /* SKIP overrides PRUNE and THEN */  
   
804      case OP_SKIP:      case OP_SKIP:
805      RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md,      RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md,
806        eptrb, RM53);        eptrb, RM53);
807      if (rrc != MATCH_NOMATCH && rrc != MATCH_PRUNE && rrc != MATCH_THEN)      if (rrc != MATCH_NOMATCH) RRETURN(rrc);
       RRETURN(rrc);  
808      md->start_match_ptr = eptr;   /* Pass back current position */      md->start_match_ptr = eptr;   /* Pass back current position */
809      RRETURN(MATCH_SKIP);      RRETURN(MATCH_SKIP);
810    
811      /* Note that, for Perl compatibility, SKIP with an argument does NOT set      /* Note that, for Perl compatibility, SKIP with an argument does NOT set
812      nomatch_mark. There is a flag that disables this opcode when re-matching a      nomatch_mark. When a pattern match ends with a SKIP_ARG for which there was
813      pattern that ended with a SKIP for which there was not a matching MARK. */      not a matching mark, we have to re-run the match, ignoring the SKIP_ARG
814        that failed and any that precede it (either they also failed, or were not
815        triggered). To do this, we maintain a count of executed SKIP_ARGs. If a
816        SKIP_ARG gets to top level, the match is re-run with md->ignore_skip_arg
817        set to the count of the one that failed. */
818    
819      case OP_SKIP_ARG:      case OP_SKIP_ARG:
820      if (md->ignore_skip_arg)      md->skip_arg_count++;
821        if (md->skip_arg_count <= md->ignore_skip_arg)
822        {        {
823        ecode += PRIV(OP_lengths)[*ecode] + ecode[1];        ecode += PRIV(OP_lengths)[*ecode] + ecode[1];
824        break;        break;
825        }        }
826      RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode] + ecode[1], offset_top, md,      RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode] + ecode[1], offset_top, md,
827        eptrb, RM57);        eptrb, RM57);
828      if (rrc != MATCH_NOMATCH && rrc != MATCH_PRUNE && rrc != MATCH_THEN)      if (rrc != MATCH_NOMATCH) RRETURN(rrc);
       RRETURN(rrc);  
829    
830      /* Pass back the current skip name by overloading md->start_match_ptr and      /* Pass back the current skip name by overloading md->start_match_ptr and
831      returning the special MATCH_SKIP_ARG return code. This will either be      returning the special MATCH_SKIP_ARG return code. This will either be
832      caught by a matching MARK, or get to the top, where it causes a rematch      caught by a matching MARK, or get to the top, where it causes a rematch
833      with the md->ignore_skip_arg flag set. */      with md->ignore_skip_arg set to the value of md->skip_arg_count. */
834    
835      md->start_match_ptr = ecode + 2;      md->start_match_ptr = ecode + 2;
836      RRETURN(MATCH_SKIP_ARG);      RRETURN(MATCH_SKIP_ARG);
# Line 838  for (;;) Line 873  for (;;)
873      case OP_ONCE_NC:      case OP_ONCE_NC:
874      prev = ecode;      prev = ecode;
875      saved_eptr = eptr;      saved_eptr = eptr;
876      save_mark = md->mark;      save_mark = md->mark;
877      do      do
878        {        {
879        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, eptrb, RM64);        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, eptrb, RM64);
# Line 857  for (;;) Line 892  for (;;)
892    
893        if (rrc != MATCH_NOMATCH) RRETURN(rrc);        if (rrc != MATCH_NOMATCH) RRETURN(rrc);
894        ecode += GET(ecode,1);        ecode += GET(ecode,1);
895        md->mark = save_mark;        md->mark = save_mark;
896        }        }
897      while (*ecode == OP_ALT);      while (*ecode == OP_ALT);
898    
# Line 897  for (;;) Line 932  for (;;)
932        }        }
933      else  /* OP_KETRMAX */      else  /* OP_KETRMAX */
934        {        {
       md->match_function_type = MATCH_CBEGROUP;  
935        RMATCH(eptr, prev, offset_top, md, eptrb, RM66);        RMATCH(eptr, prev, offset_top, md, eptrb, RM66);
936        if (rrc != MATCH_NOMATCH) RRETURN(rrc);        if (rrc != MATCH_NOMATCH) RRETURN(rrc);
937        ecode += 1 + LINK_SIZE;        ecode += 1 + LINK_SIZE;
# Line 937  for (;;) Line 971  for (;;)
971        save_offset2 = md->offset_vector[offset+1];        save_offset2 = md->offset_vector[offset+1];
972        save_offset3 = md->offset_vector[md->offset_end - number];        save_offset3 = md->offset_vector[md->offset_end - number];
973        save_capture_last = md->capture_last;        save_capture_last = md->capture_last;
974        save_mark = md->mark;        save_mark = md->mark;
975    
976        DPRINTF(("saving %d %d %d\n", save_offset1, save_offset2, save_offset3));        DPRINTF(("saving %d %d %d\n", save_offset1, save_offset2, save_offset3));
977        md->offset_vector[md->offset_end - number] =        md->offset_vector[md->offset_end - number] =
# Line 1026  for (;;) Line 1060  for (;;)
1060    
1061      for (;;)      for (;;)
1062        {        {
1063        if (op >= OP_SBRA || op == OP_ONCE) md->match_function_type = MATCH_CBEGROUP;        if (op >= OP_SBRA || op == OP_ONCE)
1064            md->match_function_type = MATCH_CBEGROUP;
1065    
1066        /* If this is not a possibly empty group, and there are no (*THEN)s in        /* If this is not a possibly empty group, and there are no (*THEN)s in
1067        the pattern, and this is the final alternative, optimize as described        the pattern, and this is the final alternative, optimize as described
# Line 1041  for (;;) Line 1076  for (;;)
1076        /* In all other cases, we have to make another call to match(). */        /* In all other cases, we have to make another call to match(). */
1077    
1078        save_mark = md->mark;        save_mark = md->mark;
1079          save_capture_last = md->capture_last;
1080        RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md, eptrb,        RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md, eptrb,
1081          RM2);          RM2);
1082    
1083        /* See comment in the code for capturing groups above about handling        /* See comment in the code for capturing groups above about handling
1084        THEN. */        THEN. */
1085    
# Line 1070  for (;;) Line 1106  for (;;)
1106          RRETURN(rrc);          RRETURN(rrc);
1107          }          }
1108        ecode += GET(ecode, 1);        ecode += GET(ecode, 1);
1109        md->mark = save_mark;        md->mark = save_mark;
1110        if (*ecode != OP_ALT) break;        if (*ecode != OP_ALT) break;
1111          md->capture_last = save_capture_last;
1112        }        }
1113    
1114      RRETURN(MATCH_NOMATCH);      RRETURN(MATCH_NOMATCH);
# Line 1134  for (;;) Line 1171  for (;;)
1171            ecode = md->start_code + code_offset;            ecode = md->start_code + code_offset;
1172            save_capture_last = md->capture_last;            save_capture_last = md->capture_last;
1173            matched_once = TRUE;            matched_once = TRUE;
1174              mstart = md->start_match_ptr;    /* In case \K changed it */
1175            continue;            continue;
1176            }            }
1177    
# Line 1193  for (;;) Line 1231  for (;;)
1231      POSSESSIVE_NON_CAPTURE:      POSSESSIVE_NON_CAPTURE:
1232      matched_once = FALSE;      matched_once = FALSE;
1233      code_offset = (int)(ecode - md->start_code);      code_offset = (int)(ecode - md->start_code);
1234        save_capture_last = md->capture_last;
1235    
1236      for (;;)      for (;;)
1237        {        {
# Line 1205  for (;;) Line 1244  for (;;)
1244          eptr = md->end_match_ptr;          eptr = md->end_match_ptr;
1245          ecode = md->start_code + code_offset;          ecode = md->start_code + code_offset;
1246          matched_once = TRUE;          matched_once = TRUE;
1247            mstart = md->start_match_ptr;   /* In case \K reset it */
1248          continue;          continue;
1249          }          }
1250    
# Line 1222  for (;;) Line 1262  for (;;)
1262        if (rrc != MATCH_NOMATCH) RRETURN(rrc);        if (rrc != MATCH_NOMATCH) RRETURN(rrc);
1263        ecode += GET(ecode, 1);        ecode += GET(ecode, 1);
1264        if (*ecode != OP_ALT) break;        if (*ecode != OP_ALT) break;
1265          md->capture_last = save_capture_last;
1266        }        }
1267    
1268      if (matched_once || allow_zero)      if (matched_once || allow_zero)
# Line 1233  for (;;) Line 1274  for (;;)
1274    
1275      /* Control never reaches here. */      /* Control never reaches here. */
1276    
1277      /* Conditional group: compilation checked that there are no more than      /* Conditional group: compilation checked that there are no more than two
1278      two branches. If the condition is false, skipping the first branch takes us      branches. If the condition is false, skipping the first branch takes us
1279      past the end if there is only one branch, but that's OK because that is      past the end of the item if there is only one branch, but that's exactly
1280      exactly what going to the ket would do. */      what we want. */
1281    
1282      case OP_COND:      case OP_COND:
1283      case OP_SCOND:      case OP_SCOND:
1284      codelink = GET(ecode, 1);  
1285        /* The variable codelink will be added to ecode when the condition is
1286        false, to get to the second branch. Setting it to the offset to the ALT
1287        or KET, then incrementing ecode achieves this effect. We now have ecode
1288        pointing to the condition or callout. */
1289    
1290        codelink = GET(ecode, 1);   /* Offset to the second branch */
1291        ecode += 1 + LINK_SIZE;     /* From this opcode */
1292    
1293      /* Because of the way auto-callout works during compile, a callout item is      /* Because of the way auto-callout works during compile, a callout item is
1294      inserted between OP_COND and an assertion condition. */      inserted between OP_COND and an assertion condition. */
1295    
1296      if (ecode[LINK_SIZE+1] == OP_CALLOUT)      if (*ecode == OP_CALLOUT)
1297        {        {
1298        if (PUBL(callout) != NULL)        if (PUBL(callout) != NULL)
1299          {          {
1300          PUBL(callout_block) cb;          PUBL(callout_block) cb;
1301          cb.version          = 2;   /* Version 1 of the callout block */          cb.version          = 2;   /* Version 1 of the callout block */
1302          cb.callout_number   = ecode[LINK_SIZE+2];          cb.callout_number   = ecode[1];
1303          cb.offset_vector    = md->offset_vector;          cb.offset_vector    = md->offset_vector;
1304  #ifdef COMPILE_PCRE8  #if defined COMPILE_PCRE8
1305          cb.subject          = (PCRE_SPTR)md->start_subject;          cb.subject          = (PCRE_SPTR)md->start_subject;
1306  #else  #elif defined COMPILE_PCRE16
1307          cb.subject          = (PCRE_SPTR16)md->start_subject;          cb.subject          = (PCRE_SPTR16)md->start_subject;
1308    #elif defined COMPILE_PCRE32
1309            cb.subject          = (PCRE_SPTR32)md->start_subject;
1310  #endif  #endif
1311          cb.subject_length   = (int)(md->end_subject - md->start_subject);          cb.subject_length   = (int)(md->end_subject - md->start_subject);
1312          cb.start_match      = (int)(mstart - md->start_subject);          cb.start_match      = (int)(mstart - md->start_subject);
1313          cb.current_position = (int)(eptr - md->start_subject);          cb.current_position = (int)(eptr - md->start_subject);
1314          cb.pattern_position = GET(ecode, LINK_SIZE + 3);          cb.pattern_position = GET(ecode, 2);
1315          cb.next_item_length = GET(ecode, 3 + 2*LINK_SIZE);          cb.next_item_length = GET(ecode, 2 + LINK_SIZE);
1316          cb.capture_top      = offset_top/2;          cb.capture_top      = offset_top/2;
1317          cb.capture_last     = md->capture_last;          cb.capture_last     = md->capture_last & CAPLMASK;
1318            /* Internal change requires this for API compatibility. */
1319            if (cb.capture_last == 0) cb.capture_last = -1;
1320          cb.callout_data     = md->callout_data;          cb.callout_data     = md->callout_data;
1321          cb.mark             = md->nomatch_mark;          cb.mark             = md->nomatch_mark;
1322          if ((rrc = (*PUBL(callout))(&cb)) > 0) RRETURN(MATCH_NOMATCH);          if ((rrc = (*PUBL(callout))(&cb)) > 0) RRETURN(MATCH_NOMATCH);
1323          if (rrc < 0) RRETURN(rrc);          if (rrc < 0) RRETURN(rrc);
1324          }          }
1325    
1326          /* Advance ecode past the callout, so it now points to the condition. We
1327          must adjust codelink so that the value of ecode+codelink is unchanged. */
1328    
1329        ecode += PRIV(OP_lengths)[OP_CALLOUT];        ecode += PRIV(OP_lengths)[OP_CALLOUT];
1330          codelink -= PRIV(OP_lengths)[OP_CALLOUT];
1331        }        }
1332    
1333      condcode = ecode[LINK_SIZE+1];      /* Test the various possible conditions */
1334    
1335      /* Now see what the actual condition is */      condition = FALSE;
1336        switch(condcode = *ecode)
     if (condcode == OP_RREF || condcode == OP_NRREF)    /* Recursion test */  
1337        {        {
1338        if (md->recursive == NULL)                /* Not recursing => FALSE */        case OP_RREF:         /* Numbered group recursion test */
1339          {        if (md->recursive != NULL)     /* Not recursing => FALSE */
         condition = FALSE;  
         ecode += GET(ecode, 1);  
         }  
       else  
1340          {          {
1341          int recno = GET2(ecode, LINK_SIZE + 2);   /* Recursion group number*/          unsigned int recno = GET2(ecode, 1);   /* Recursion group number*/
1342          condition = (recno == RREF_ANY || recno == md->recursive->group_num);          condition = (recno == RREF_ANY || recno == md->recursive->group_num);
1343            }
1344          break;
1345    
1346          /* If the test is for recursion into a specific subpattern, and it is        case OP_DNRREF:       /* Duplicate named group recursion test */
1347          false, but the test was set up by name, scan the table to see if the        if (md->recursive != NULL)
1348          name refers to any other numbers, and test them. The condition is true          {
1349          if any one is set. */          int count = GET2(ecode, 1 + IMM2_SIZE);
1350            pcre_uchar *slot = md->name_table + GET2(ecode, 1) * md->name_entry_size;
1351          if (!condition && condcode == OP_NRREF)          while (count-- > 0)
1352            {            {
1353            pcre_uchar *slotA = md->name_table;            unsigned int recno = GET2(slot, 0);
1354            for (i = 0; i < md->name_count; i++)            condition = recno == md->recursive->group_num;
1355              {            if (condition) break;
1356              if (GET2(slotA, 0) == recno) break;            slot += md->name_entry_size;
             slotA += md->name_entry_size;  
             }  
   
           /* Found a name for the number - there can be only one; duplicate  
           names for different numbers are allowed, but not vice versa. First  
           scan down for duplicates. */  
   
           if (i < md->name_count)  
             {  
             pcre_uchar *slotB = slotA;  
             while (slotB > md->name_table)  
               {  
               slotB -= md->name_entry_size;  
               if (STRCMP_UC_UC(slotA + IMM2_SIZE, slotB + IMM2_SIZE) == 0)  
                 {  
                 condition = GET2(slotB, 0) == md->recursive->group_num;  
                 if (condition) break;  
                 }  
               else break;  
               }  
   
             /* Scan up for duplicates */  
   
             if (!condition)  
               {  
               slotB = slotA;  
               for (i++; i < md->name_count; i++)  
                 {  
                 slotB += md->name_entry_size;  
                 if (STRCMP_UC_UC(slotA + IMM2_SIZE, slotB + IMM2_SIZE) == 0)  
                   {  
                   condition = GET2(slotB, 0) == md->recursive->group_num;  
                   if (condition) break;  
                   }  
                 else break;  
                 }  
               }  
             }  
1357            }            }
   
         /* Chose branch according to the condition */  
   
         ecode += condition? 1 + IMM2_SIZE : GET(ecode, 1);  
1358          }          }
1359        }        break;
1360    
1361      else if (condcode == OP_CREF || condcode == OP_NCREF)  /* Group used test */        case OP_CREF:         /* Numbered group used test */
1362        {        offset = GET2(ecode, 1) << 1;  /* Doubled ref number */
       offset = GET2(ecode, LINK_SIZE+2) << 1;  /* Doubled ref number */  
1363        condition = offset < offset_top && md->offset_vector[offset] >= 0;        condition = offset < offset_top && md->offset_vector[offset] >= 0;
1364          break;
1365    
1366        /* If the numbered capture is unset, but the reference was by name,        case OP_DNCREF:      /* Duplicate named group used test */
       scan the table to see if the name refers to any other numbers, and test  
       them. The condition is true if any one is set. This is tediously similar  
       to the code above, but not close enough to try to amalgamate. */  
   
       if (!condition && condcode == OP_NCREF)  
1367          {          {
1368          int refno = offset >> 1;          int count = GET2(ecode, 1 + IMM2_SIZE);
1369          pcre_uchar *slotA = md->name_table;          pcre_uchar *slot = md->name_table + GET2(ecode, 1) * md->name_entry_size;
1370            while (count-- > 0)
1371          for (i = 0; i < md->name_count; i++)            {
1372            {            offset = GET2(slot, 0) << 1;
1373            if (GET2(slotA, 0) == refno) break;            condition = offset < offset_top && md->offset_vector[offset] >= 0;
1374            slotA += md->name_entry_size;            if (condition) break;
1375            }            slot += md->name_entry_size;
   
         /* Found a name for the number - there can be only one; duplicate names  
         for different numbers are allowed, but not vice versa. First scan down  
         for duplicates. */  
   
         if (i < md->name_count)  
           {  
           pcre_uchar *slotB = slotA;  
           while (slotB > md->name_table)  
             {  
             slotB -= md->name_entry_size;  
             if (STRCMP_UC_UC(slotA + IMM2_SIZE, slotB + IMM2_SIZE) == 0)  
               {  
               offset = GET2(slotB, 0) << 1;  
               condition = offset < offset_top &&  
                 md->offset_vector[offset] >= 0;  
               if (condition) break;  
               }  
             else break;  
             }  
   
           /* Scan up for duplicates */  
   
           if (!condition)  
             {  
             slotB = slotA;  
             for (i++; i < md->name_count; i++)  
               {  
               slotB += md->name_entry_size;  
               if (STRCMP_UC_UC(slotA + IMM2_SIZE, slotB + IMM2_SIZE) == 0)  
                 {  
                 offset = GET2(slotB, 0) << 1;  
                 condition = offset < offset_top &&  
                   md->offset_vector[offset] >= 0;  
                 if (condition) break;  
                 }  
               else break;  
               }  
             }  
1376            }            }
1377          }          }
1378          break;
1379    
1380        /* Chose branch according to the condition */        case OP_DEF:     /* DEFINE - always false */
1381          break;
       ecode += condition? 1 + IMM2_SIZE : GET(ecode, 1);  
       }  
   
     else if (condcode == OP_DEF)     /* DEFINE - always false */  
       {  
       condition = FALSE;  
       ecode += GET(ecode, 1);  
       }  
1382    
1383      /* The condition is an assertion. Call match() to evaluate it - setting        /* The condition is an assertion. Call match() to evaluate it - setting
1384      md->match_function_type to MATCH_CONDASSERT causes it to stop at the end of        md->match_function_type to MATCH_CONDASSERT causes it to stop at the end
1385      an assertion. */        of an assertion. */
1386    
1387      else        default:
       {  
1388        md->match_function_type = MATCH_CONDASSERT;        md->match_function_type = MATCH_CONDASSERT;
1389        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, NULL, RM3);        RMATCH(eptr, ecode, offset_top, md, NULL, RM3);
1390        if (rrc == MATCH_MATCH)        if (rrc == MATCH_MATCH)
1391          {          {
1392          if (md->end_offset_top > offset_top)          if (md->end_offset_top > offset_top)
1393            offset_top = md->end_offset_top;  /* Captures may have happened */            offset_top = md->end_offset_top;  /* Captures may have happened */
1394          condition = TRUE;          condition = TRUE;
1395          ecode += 1 + LINK_SIZE + GET(ecode, LINK_SIZE + 2);  
1396            /* Advance ecode past the assertion to the start of the first branch,
1397            but adjust it so that the general choosing code below works. */
1398    
1399            ecode += GET(ecode, 1);
1400          while (*ecode == OP_ALT) ecode += GET(ecode, 1);          while (*ecode == OP_ALT) ecode += GET(ecode, 1);
1401            ecode += 1 + LINK_SIZE - PRIV(OP_lengths)[condcode];
1402          }          }
1403    
1404        /* PCRE doesn't allow the effect of (*THEN) to escape beyond an        /* PCRE doesn't allow the effect of (*THEN) to escape beyond an
1405        assertion; it is therefore treated as NOMATCH. */        assertion; it is therefore treated as NOMATCH. Any other return is an
1406          error. */
1407    
1408        else if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN)        else if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN)
1409          {          {
1410          RRETURN(rrc);         /* Need braces because of following else */          RRETURN(rrc);         /* Need braces because of following else */
1411          }          }
1412        else        break;
         {  
         condition = FALSE;  
         ecode += codelink;  
         }  
1413        }        }
1414    
1415      /* We are now at the branch that is to be obeyed. As there is only one, can      /* Choose branch according to the condition */
     use tail recursion to avoid using another stack frame, except when there is  
     unlimited repeat of a possibly empty group. In the latter case, a recursive  
     call to match() is always required, unless the second alternative doesn't  
     exist, in which case we can just plough on. Note that, for compatibility  
     with Perl, the | in a conditional group is NOT treated as creating two  
     alternatives. If a THEN is encountered in the branch, it propagates out to  
     the enclosing alternative (unless nested in a deeper set of alternatives,  
     of course). */  
1416    
1417      if (condition || *ecode == OP_ALT)      ecode += condition? PRIV(OP_lengths)[condcode] : codelink;
1418    
1419        /* We are now at the branch that is to be obeyed. As there is only one, we
1420        can use tail recursion to avoid using another stack frame, except when
1421        there is unlimited repeat of a possibly empty group. In the latter case, a
1422        recursive call to match() is always required, unless the second alternative
1423        doesn't exist, in which case we can just plough on. Note that, for
1424        compatibility with Perl, the | in a conditional group is NOT treated as
1425        creating two alternatives. If a THEN is encountered in the branch, it
1426        propagates out to the enclosing alternative (unless nested in a deeper set
1427        of alternatives, of course). */
1428    
1429        if (condition || ecode[-(1+LINK_SIZE)] == OP_ALT)
1430        {        {
1431        if (op != OP_SCOND)        if (op != OP_SCOND)
1432          {          {
         ecode += 1 + LINK_SIZE;  
1433          goto TAIL_RECURSE;          goto TAIL_RECURSE;
1434          }          }
1435    
1436        md->match_function_type = MATCH_CBEGROUP;        md->match_function_type = MATCH_CBEGROUP;
1437        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, eptrb, RM49);        RMATCH(eptr, ecode, offset_top, md, eptrb, RM49);
1438        RRETURN(rrc);        RRETURN(rrc);
1439        }        }
1440    
# Line 1477  for (;;) Line 1442  for (;;)
1442    
1443      else      else
1444        {        {
       ecode += 1 + LINK_SIZE;  
1445        }        }
1446      break;      break;
1447    
# Line 1486  for (;;) Line 1450  for (;;)
1450      to close any currently open capturing brackets. */      to close any currently open capturing brackets. */
1451    
1452      case OP_CLOSE:      case OP_CLOSE:
1453      number = GET2(ecode, 1);      number = GET2(ecode, 1);   /* Must be less than 65536 */
1454      offset = number << 1;      offset = number << 1;
1455    
1456  #ifdef PCRE_DEBUG  #ifdef PCRE_DEBUG
# Line 1494  for (;;) Line 1458  for (;;)
1458        printf("\n");        printf("\n");
1459  #endif  #endif
1460    
1461      md->capture_last = number;      md->capture_last = (md->capture_last & OVFLMASK) | number;
1462      if (offset >= md->offset_max) md->offset_overflow = TRUE; else      if (offset >= md->offset_max) md->capture_last |= OVFLBIT; else
1463        {        {
1464        md->offset_vector[offset] =        md->offset_vector[offset] =
1465          md->offset_vector[md->offset_end - number];          md->offset_vector[md->offset_end - number];
# Line 1549  for (;;) Line 1513  for (;;)
1513    
1514      case OP_ASSERT:      case OP_ASSERT:
1515      case OP_ASSERTBACK:      case OP_ASSERTBACK:
1516      save_mark = md->mark;      save_mark = md->mark;
1517      if (md->match_function_type == MATCH_CONDASSERT)      if (md->match_function_type == MATCH_CONDASSERT)
1518        {        {
1519        condassert = TRUE;        condassert = TRUE;
# Line 1557  for (;;) Line 1521  for (;;)
1521        }        }
1522      else condassert = FALSE;      else condassert = FALSE;
1523    
1524        /* Loop for each branch */
1525    
1526      do      do
1527        {        {
1528        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, NULL, RM4);        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, NULL, RM4);
1529    
1530          /* A match means that the assertion is true; break out of the loop
1531          that matches its alternatives. */
1532    
1533        if (rrc == MATCH_MATCH || rrc == MATCH_ACCEPT)        if (rrc == MATCH_MATCH || rrc == MATCH_ACCEPT)
1534          {          {
1535          mstart = md->start_match_ptr;   /* In case \K reset it */          mstart = md->start_match_ptr;   /* In case \K reset it */
1536          break;          break;
1537          }          }
1538    
1539        /* PCRE does not allow THEN to escape beyond an assertion; it is treated        /* If not matched, restore the previous mark setting. */
1540        as NOMATCH. */  
1541          md->mark = save_mark;
1542    
1543          /* See comment in the code for capturing groups above about handling
1544          THEN. */
1545    
1546          if (rrc == MATCH_THEN)
1547            {
1548            next = ecode + GET(ecode,1);
1549            if (md->start_match_ptr < next &&
1550                (*ecode == OP_ALT || *next == OP_ALT))
1551              rrc = MATCH_NOMATCH;
1552            }
1553    
1554          /* Anything other than NOMATCH causes the entire assertion to fail,
1555          passing back the return code. This includes COMMIT, SKIP, PRUNE and an
1556          uncaptured THEN, which means they take their normal effect. This
1557          consistent approach does not always have exactly the same effect as in
1558          Perl. */
1559    
1560        if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN) RRETURN(rrc);        if (rrc != MATCH_NOMATCH) RRETURN(rrc);
1561        ecode += GET(ecode, 1);        ecode += GET(ecode, 1);
       md->mark = save_mark;  
1562        }        }
1563      while (*ecode == OP_ALT);      while (*ecode == OP_ALT);   /* Continue for next alternative */
1564    
1565        /* If we have tried all the alternative branches, the assertion has
1566        failed. If not, we broke out after a match. */
1567    
1568      if (*ecode == OP_KET) RRETURN(MATCH_NOMATCH);      if (*ecode == OP_KET) RRETURN(MATCH_NOMATCH);
1569    
# Line 1581  for (;;) Line 1571  for (;;)
1571    
1572      if (condassert) RRETURN(MATCH_MATCH);      if (condassert) RRETURN(MATCH_MATCH);
1573    
1574      /* Continue from after the assertion, updating the offsets high water      /* Continue from after a successful assertion, updating the offsets high
1575      mark, since extracts may have been taken during the assertion. */      water mark, since extracts may have been taken during the assertion. */
1576    
1577      do ecode += GET(ecode,1); while (*ecode == OP_ALT);      do ecode += GET(ecode,1); while (*ecode == OP_ALT);
1578      ecode += 1 + LINK_SIZE;      ecode += 1 + LINK_SIZE;
1579      offset_top = md->end_offset_top;      offset_top = md->end_offset_top;
1580      continue;      continue;
1581    
1582      /* Negative assertion: all branches must fail to match. Encountering SKIP,      /* Negative assertion: all branches must fail to match for the assertion to
1583      PRUNE, or COMMIT means we must assume failure without checking subsequent      succeed. */
     branches. */  
1584    
1585      case OP_ASSERT_NOT:      case OP_ASSERT_NOT:
1586      case OP_ASSERTBACK_NOT:      case OP_ASSERTBACK_NOT:
1587      save_mark = md->mark;      save_mark = md->mark;
1588      if (md->match_function_type == MATCH_CONDASSERT)      if (md->match_function_type == MATCH_CONDASSERT)
1589        {        {
1590        condassert = TRUE;        condassert = TRUE;
# Line 1603  for (;;) Line 1592  for (;;)
1592        }        }
1593      else condassert = FALSE;      else condassert = FALSE;
1594    
1595        /* Loop for each alternative branch. */
1596    
1597      do      do
1598        {        {
1599        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, NULL, RM5);        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, NULL, RM5);
1600        md->mark = save_mark;        md->mark = save_mark;   /* Always restore the mark setting */
1601        if (rrc == MATCH_MATCH || rrc == MATCH_ACCEPT) RRETURN(MATCH_NOMATCH);  
1602        if (rrc == MATCH_SKIP || rrc == MATCH_PRUNE || rrc == MATCH_COMMIT)        switch(rrc)
1603          {          {
1604          do ecode += GET(ecode,1); while (*ecode == OP_ALT);          case MATCH_MATCH:            /* A successful match means */
1605            case MATCH_ACCEPT:           /* the assertion has failed. */
1606            RRETURN(MATCH_NOMATCH);
1607    
1608            case MATCH_NOMATCH:          /* Carry on with next branch */
1609          break;          break;
1610    
1611            /* See comment in the code for capturing groups above about handling
1612            THEN. */
1613    
1614            case MATCH_THEN:
1615            next = ecode + GET(ecode,1);
1616            if (md->start_match_ptr < next &&
1617                (*ecode == OP_ALT || *next == OP_ALT))
1618              {
1619              rrc = MATCH_NOMATCH;
1620              break;
1621              }
1622            /* Otherwise fall through. */
1623    
1624            /* COMMIT, SKIP, PRUNE, and an uncaptured THEN cause the whole
1625            assertion to fail to match, without considering any more alternatives.
1626            Failing to match means the assertion is true. This is a consistent
1627            approach, but does not always have the same effect as in Perl. */
1628    
1629            case MATCH_COMMIT:
1630            case MATCH_SKIP:
1631            case MATCH_SKIP_ARG:
1632            case MATCH_PRUNE:
1633            do ecode += GET(ecode,1); while (*ecode == OP_ALT);
1634            goto NEG_ASSERT_TRUE;   /* Break out of alternation loop */
1635    
1636            /* Anything else is an error */
1637    
1638            default:
1639            RRETURN(rrc);
1640          }          }
1641    
1642        /* PCRE does not allow THEN to escape beyond an assertion; it is treated        /* Continue with next branch */
       as NOMATCH. */  
1643    
       if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN) RRETURN(rrc);  
1644        ecode += GET(ecode,1);        ecode += GET(ecode,1);
1645        }        }
1646      while (*ecode == OP_ALT);      while (*ecode == OP_ALT);
1647    
1648      if (condassert) RRETURN(MATCH_MATCH);  /* Condition assertion */      /* All branches in the assertion failed to match. */
1649    
1650      ecode += 1 + LINK_SIZE;      NEG_ASSERT_TRUE:
1651        if (condassert) RRETURN(MATCH_MATCH);  /* Condition assertion */
1652        ecode += 1 + LINK_SIZE;                /* Continue with current branch */
1653      continue;      continue;
1654    
1655      /* Move the subject pointer back. This occurs only at the start of      /* Move the subject pointer back. This occurs only at the start of
# Line 1671  for (;;) Line 1696  for (;;)
1696        cb.version          = 2;   /* Version 1 of the callout block */        cb.version          = 2;   /* Version 1 of the callout block */
1697        cb.callout_number   = ecode[1];        cb.callout_number   = ecode[1];
1698        cb.offset_vector    = md->offset_vector;        cb.offset_vector    = md->offset_vector;
1699  #ifdef COMPILE_PCRE8  #if defined COMPILE_PCRE8
1700        cb.subject          = (PCRE_SPTR)md->start_subject;        cb.subject          = (PCRE_SPTR)md->start_subject;
1701  #else  #elif defined COMPILE_PCRE16
1702        cb.subject          = (PCRE_SPTR16)md->start_subject;        cb.subject          = (PCRE_SPTR16)md->start_subject;
1703    #elif defined COMPILE_PCRE32
1704          cb.subject          = (PCRE_SPTR32)md->start_subject;
1705  #endif  #endif
1706        cb.subject_length   = (int)(md->end_subject - md->start_subject);        cb.subject_length   = (int)(md->end_subject - md->start_subject);
1707        cb.start_match      = (int)(mstart - md->start_subject);        cb.start_match      = (int)(mstart - md->start_subject);
# Line 1682  for (;;) Line 1709  for (;;)
1709        cb.pattern_position = GET(ecode, 2);        cb.pattern_position = GET(ecode, 2);
1710        cb.next_item_length = GET(ecode, 2 + LINK_SIZE);        cb.next_item_length = GET(ecode, 2 + LINK_SIZE);
1711        cb.capture_top      = offset_top/2;        cb.capture_top      = offset_top/2;
1712        cb.capture_last     = md->capture_last;        cb.capture_last     = md->capture_last & CAPLMASK;
1713          /* Internal change requires this for API compatibility. */
1714          if (cb.capture_last == 0) cb.capture_last = -1;
1715        cb.callout_data     = md->callout_data;        cb.callout_data     = md->callout_data;
1716        cb.mark             = md->nomatch_mark;        cb.mark             = md->nomatch_mark;
1717        if ((rrc = (*PUBL(callout))(&cb)) > 0) RRETURN(MATCH_NOMATCH);        if ((rrc = (*PUBL(callout))(&cb)) > 0) RRETURN(MATCH_NOMATCH);
# Line 1711  for (;;) Line 1740  for (;;)
1740      case OP_RECURSE:      case OP_RECURSE:
1741        {        {
1742        recursion_info *ri;        recursion_info *ri;
1743        int recno;        unsigned int recno;
1744    
1745        callpat = md->start_code + GET(ecode, 1);        callpat = md->start_code + GET(ecode, 1);
1746        recno = (callpat == md->start_code)? 0 :        recno = (callpat == md->start_code)? 0 :
# Line 1728  for (;;) Line 1757  for (;;)
1757        /* Add to "recursing stack" */        /* Add to "recursing stack" */
1758    
1759        new_recursive.group_num = recno;        new_recursive.group_num = recno;
1760          new_recursive.saved_capture_last = md->capture_last;
1761        new_recursive.subject_position = eptr;        new_recursive.subject_position = eptr;
1762        new_recursive.prevrec = md->recursive;        new_recursive.prevrec = md->recursive;
1763        md->recursive = &new_recursive;        md->recursive = &new_recursive;
# Line 1751  for (;;) Line 1781  for (;;)
1781              new_recursive.saved_max * sizeof(int));              new_recursive.saved_max * sizeof(int));
1782    
1783        /* OK, now we can do the recursion. After processing each alternative,        /* OK, now we can do the recursion. After processing each alternative,
1784        restore the offset data. If there were nested recursions, md->recursive        restore the offset data and the last captured value. If there were nested
1785        might be changed, so reset it before looping. */        recursions, md->recursive might be changed, so reset it before looping.
1786          */
1787    
1788        DPRINTF(("Recursing into group %d\n", new_recursive.group_num));        DPRINTF(("Recursing into group %d\n", new_recursive.group_num));
1789        cbegroup = (*callpat >= OP_SBRA);        cbegroup = (*callpat >= OP_SBRA);
# Line 1763  for (;;) Line 1794  for (;;)
1794            md, eptrb, RM6);            md, eptrb, RM6);
1795          memcpy(md->offset_vector, new_recursive.offset_save,          memcpy(md->offset_vector, new_recursive.offset_save,
1796              new_recursive.saved_max * sizeof(int));              new_recursive.saved_max * sizeof(int));
1797            md->capture_last = new_recursive.saved_capture_last;
1798          md->recursive = new_recursive.prevrec;          md->recursive = new_recursive.prevrec;
1799          if (rrc == MATCH_MATCH || rrc == MATCH_ACCEPT)          if (rrc == MATCH_MATCH || rrc == MATCH_ACCEPT)
1800            {            {
# Line 1779  for (;;) Line 1811  for (;;)
1811            goto RECURSION_MATCHED;        /* Exit loop; end processing */            goto RECURSION_MATCHED;        /* Exit loop; end processing */
1812            }            }
1813    
1814          /* PCRE does not allow THEN to escape beyond a recursion; it is treated          /* PCRE does not allow THEN, SKIP, PRUNE or COMMIT to escape beyond a
1815          as NOMATCH. */          recursion; they cause a NOMATCH for the entire recursion. These codes
1816            are defined in a range that can be tested for. */
1817    
1818            if (rrc >= MATCH_BACKTRACK_MIN && rrc <= MATCH_BACKTRACK_MAX)
1819              RRETURN(MATCH_NOMATCH);
1820    
1821            /* Any return code other than NOMATCH is an error. */
1822    
1823          else if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN)          if (rrc != MATCH_NOMATCH)
1824            {            {
1825            DPRINTF(("Recursion gave error %d\n", rrc));            DPRINTF(("Recursion gave error %d\n", rrc));
1826            if (new_recursive.offset_save != stacksave)            if (new_recursive.offset_save != stacksave)
# Line 1912  for (;;) Line 1950  for (;;)
1950    
1951        /* Deal with capturing */        /* Deal with capturing */
1952    
1953        md->capture_last = number;        md->capture_last = (md->capture_last & OVFLMASK) | number;
1954        if (offset >= md->offset_max) md->offset_overflow = TRUE; else        if (offset >= md->offset_max) md->capture_last |= OVFLBIT; else
1955          {          {
1956          /* If offset is greater than offset_top, it means that we are          /* If offset is greater than offset_top, it means that we are
1957          "skipping" a capturing group, and that group's offsets must be marked          "skipping" a capturing group, and that group's offsets must be marked
# Line 1969  for (;;) Line 2007  for (;;)
2007    
2008      if (*ecode == OP_KETRPOS)      if (*ecode == OP_KETRPOS)
2009        {        {
2010          md->start_match_ptr = mstart;    /* In case \K reset it */
2011        md->end_match_ptr = eptr;        md->end_match_ptr = eptr;
2012        md->end_offset_top = offset_top;        md->end_offset_top = offset_top;
2013        RRETURN(MATCH_KETRPOS);        RRETURN(MATCH_KETRPOS);
# Line 1993  for (;;) Line 2032  for (;;)
2032          }          }
2033        if (*prev >= OP_SBRA)    /* Could match an empty string */        if (*prev >= OP_SBRA)    /* Could match an empty string */
2034          {          {
         md->match_function_type = MATCH_CBEGROUP;  
2035          RMATCH(eptr, prev, offset_top, md, eptrb, RM50);          RMATCH(eptr, prev, offset_top, md, eptrb, RM50);
2036          RRETURN(rrc);          RRETURN(rrc);
2037          }          }
# Line 2002  for (;;) Line 2040  for (;;)
2040        }        }
2041      else  /* OP_KETRMAX */      else  /* OP_KETRMAX */
2042        {        {
       if (*prev >= OP_SBRA) md->match_function_type = MATCH_CBEGROUP;  
2043        RMATCH(eptr, prev, offset_top, md, eptrb, RM13);        RMATCH(eptr, prev, offset_top, md, eptrb, RM13);
2044        if (rrc == MATCH_ONCE && md->once_target == prev) rrc = MATCH_NOMATCH;        if (rrc == MATCH_ONCE && md->once_target == prev) rrc = MATCH_NOMATCH;
2045        if (rrc != MATCH_NOMATCH) RRETURN(rrc);        if (rrc != MATCH_NOMATCH) RRETURN(rrc);
# Line 2059  for (;;) Line 2096  for (;;)
2096    
2097      case OP_DOLLM:      case OP_DOLLM:
2098      if (eptr < md->end_subject)      if (eptr < md->end_subject)
2099        { if (!IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH); }        {
2100          if (!IS_NEWLINE(eptr))
2101            {
2102            if (md->partial != 0 &&
2103                eptr + 1 >= md->end_subject &&
2104                NLBLOCK->nltype == NLTYPE_FIXED &&
2105                NLBLOCK->nllen == 2 &&
2106                *eptr == NLBLOCK->nl[0])
2107              {
2108              md->hitend = TRUE;
2109              if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
2110              }
2111            RRETURN(MATCH_NOMATCH);
2112            }
2113          }
2114      else      else
2115        {        {
2116        if (md->noteol) RRETURN(MATCH_NOMATCH);        if (md->noteol) RRETURN(MATCH_NOMATCH);
# Line 2091  for (;;) Line 2142  for (;;)
2142      ASSERT_NL_OR_EOS:      ASSERT_NL_OR_EOS:
2143      if (eptr < md->end_subject &&      if (eptr < md->end_subject &&
2144          (!IS_NEWLINE(eptr) || eptr != md->end_subject - md->nllen))          (!IS_NEWLINE(eptr) || eptr != md->end_subject - md->nllen))
2145          {
2146          if (md->partial != 0 &&
2147              eptr + 1 >= md->end_subject &&
2148              NLBLOCK->nltype == NLTYPE_FIXED &&
2149              NLBLOCK->nllen == 2 &&
2150              *eptr == NLBLOCK->nl[0])
2151            {
2152            md->hitend = TRUE;
2153            if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
2154            }
2155        RRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2156          }
2157    
2158      /* Either at end of string or \n before end. */      /* Either at end of string or \n before end. */
2159    
# Line 2219  for (;;) Line 2281  for (;;)
2281        }        }
2282      break;      break;
2283    
2284      /* Match a single character type; inline for speed */      /* Match any single character type except newline; have to take care with
2285        CRLF newlines and partial matching. */
2286    
2287      case OP_ANY:      case OP_ANY:
2288      if (IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH);      if (IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH);
2289        if (md->partial != 0 &&
2290            eptr + 1 >= md->end_subject &&
2291            NLBLOCK->nltype == NLTYPE_FIXED &&
2292            NLBLOCK->nllen == 2 &&
2293            *eptr == NLBLOCK->nl[0])
2294          {
2295          md->hitend = TRUE;
2296          if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
2297          }
2298    
2299      /* Fall through */      /* Fall through */
2300    
2301        /* Match any single character whatsoever. */
2302    
2303      case OP_ALLANY:      case OP_ALLANY:
2304      if (eptr >= md->end_subject)   /* DO NOT merge the eptr++ here; it must */      if (eptr >= md->end_subject)   /* DO NOT merge the eptr++ here; it must */
2305        {                            /* not be updated before SCHECK_PARTIAL. */        {                            /* not be updated before SCHECK_PARTIAL. */
# Line 2364  for (;;) Line 2439  for (;;)
2439        {        {
2440        default: RRETURN(MATCH_NOMATCH);        default: RRETURN(MATCH_NOMATCH);
2441    
2442        case 0x000d:        case CHAR_CR:
2443        if (eptr < md->end_subject && *eptr == 0x0a) eptr++;        if (eptr >= md->end_subject)
2444            {
2445            SCHECK_PARTIAL();
2446            }
2447          else if (*eptr == CHAR_LF) eptr++;
2448        break;        break;
2449    
2450        case 0x000a:        case CHAR_LF:
2451        break;        break;
2452    
2453        case 0x000b:        case CHAR_VT:
2454        case 0x000c:        case CHAR_FF:
2455        case 0x0085:        case CHAR_NEL:
2456    #ifndef EBCDIC
2457        case 0x2028:        case 0x2028:
2458        case 0x2029:        case 0x2029:
2459    #endif  /* Not EBCDIC */
2460        if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);        if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);
2461        break;        break;
2462        }        }
# Line 2391  for (;;) Line 2472  for (;;)
2472      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2473      switch(c)      switch(c)
2474        {        {
2475          HSPACE_CASES: RRETURN(MATCH_NOMATCH);  /* Byte and multibyte cases */
2476        default: break;        default: break;
       case 0x09:      /* HT */  
       case 0x20:      /* SPACE */  
       case 0xa0:      /* NBSP */  
       case 0x1680:    /* OGHAM SPACE MARK */  
       case 0x180e:    /* MONGOLIAN VOWEL SEPARATOR */  
       case 0x2000:    /* EN QUAD */  
       case 0x2001:    /* EM QUAD */  
       case 0x2002:    /* EN SPACE */  
       case 0x2003:    /* EM SPACE */  
       case 0x2004:    /* THREE-PER-EM SPACE */  
       case 0x2005:    /* FOUR-PER-EM SPACE */  
       case 0x2006:    /* SIX-PER-EM SPACE */  
       case 0x2007:    /* FIGURE SPACE */  
       case 0x2008:    /* PUNCTUATION SPACE */  
       case 0x2009:    /* THIN SPACE */  
       case 0x200A:    /* HAIR SPACE */  
       case 0x202f:    /* NARROW NO-BREAK SPACE */  
       case 0x205f:    /* MEDIUM MATHEMATICAL SPACE */  
       case 0x3000:    /* IDEOGRAPHIC SPACE */  
       RRETURN(MATCH_NOMATCH);  
2477        }        }
2478      ecode++;      ecode++;
2479      break;      break;
# Line 2425  for (;;) Line 2487  for (;;)
2487      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2488      switch(c)      switch(c)
2489        {        {
2490          HSPACE_CASES: break;  /* Byte and multibyte cases */
2491        default: RRETURN(MATCH_NOMATCH);        default: RRETURN(MATCH_NOMATCH);
       case 0x09:      /* HT */  
       case 0x20:      /* SPACE */  
       case 0xa0:      /* NBSP */  
       case 0x1680:    /* OGHAM SPACE MARK */  
       case 0x180e:    /* MONGOLIAN VOWEL SEPARATOR */  
       case 0x2000:    /* EN QUAD */  
       case 0x2001:    /* EM QUAD */  
       case 0x2002:    /* EN SPACE */  
       case 0x2003:    /* EM SPACE */  
       case 0x2004:    /* THREE-PER-EM SPACE */  
       case 0x2005:    /* FOUR-PER-EM SPACE */  
       case 0x2006:    /* SIX-PER-EM SPACE */  
       case 0x2007:    /* FIGURE SPACE */  
       case 0x2008:    /* PUNCTUATION SPACE */  
       case 0x2009:    /* THIN SPACE */  
       case 0x200A:    /* HAIR SPACE */  
       case 0x202f:    /* NARROW NO-BREAK SPACE */  
       case 0x205f:    /* MEDIUM MATHEMATICAL SPACE */  
       case 0x3000:    /* IDEOGRAPHIC SPACE */  
       break;  
2492        }        }
2493      ecode++;      ecode++;
2494      break;      break;
# Line 2459  for (;;) Line 2502  for (;;)
2502      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2503      switch(c)      switch(c)
2504        {        {
2505          VSPACE_CASES: RRETURN(MATCH_NOMATCH);
2506        default: break;        default: break;
       case 0x0a:      /* LF */  
       case 0x0b:      /* VT */  
       case 0x0c:      /* FF */  
       case 0x0d:      /* CR */  
       case 0x85:      /* NEL */  
       case 0x2028:    /* LINE SEPARATOR */  
       case 0x2029:    /* PARAGRAPH SEPARATOR */  
       RRETURN(MATCH_NOMATCH);  
2507        }        }
2508      ecode++;      ecode++;
2509      break;      break;
# Line 2481  for (;;) Line 2517  for (;;)
2517      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2518      switch(c)      switch(c)
2519        {        {
2520          VSPACE_CASES: break;
2521        default: RRETURN(MATCH_NOMATCH);        default: RRETURN(MATCH_NOMATCH);
       case 0x0a:      /* LF */  
       case 0x0b:      /* VT */  
       case 0x0c:      /* FF */  
       case 0x0d:      /* CR */  
       case 0x85:      /* NEL */  
       case 0x2028:    /* LINE SEPARATOR */  
       case 0x2029:    /* PARAGRAPH SEPARATOR */  
       break;  
2522        }        }
2523      ecode++;      ecode++;
2524      break;      break;
# Line 2507  for (;;) Line 2536  for (;;)
2536        }        }
2537      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2538        {        {
2539          const pcre_uint32 *cp;
2540        const ucd_record *prop = GET_UCD(c);        const ucd_record *prop = GET_UCD(c);
2541    
2542        switch(ecode[1])        switch(ecode[1])
# Line 2545  for (;;) Line 2575  for (;;)
2575            RRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
2576          break;          break;
2577    
2578          case PT_SPACE:    /* Perl space */          /* Perl space used to exclude VT, but from Perl 5.18 it is included,
2579          if ((PRIV(ucp_gentype)[prop->chartype] == ucp_Z ||          which means that Perl space and POSIX space are now identical. PCRE
2580               c == CHAR_HT || c == CHAR_NL || c == CHAR_FF || c == CHAR_CR)          was changed at release 8.34. */
                == (op == OP_NOTPROP))  
           RRETURN(MATCH_NOMATCH);  
         break;  
2581    
2582            case PT_SPACE:    /* Perl space */
2583          case PT_PXSPACE:  /* POSIX space */          case PT_PXSPACE:  /* POSIX space */
2584          if ((PRIV(ucp_gentype)[prop->chartype] == ucp_Z ||          switch(c)
2585               c == CHAR_HT || c == CHAR_NL || c == CHAR_VT ||            {
2586               c == CHAR_FF || c == CHAR_CR)            HSPACE_CASES:
2587                 == (op == OP_NOTPROP))            VSPACE_CASES:
2588            RRETURN(MATCH_NOMATCH);            if (op == OP_NOTPROP) RRETURN(MATCH_NOMATCH);
2589              break;
2590    
2591              default:
2592              if ((PRIV(ucp_gentype)[prop->chartype] == ucp_Z) ==
2593                (op == OP_NOTPROP)) RRETURN(MATCH_NOMATCH);
2594              break;
2595              }
2596          break;          break;
2597    
2598          case PT_WORD:          case PT_WORD:
# Line 2567  for (;;) Line 2602  for (;;)
2602            RRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
2603          break;          break;
2604    
2605            case PT_CLIST:
2606            cp = PRIV(ucd_caseless_sets) + ecode[2];
2607            for (;;)
2608              {
2609              if (c < *cp)
2610                { if (op == OP_PROP) { RRETURN(MATCH_NOMATCH); } else break; }
2611              if (c == *cp++)
2612                { if (op == OP_PROP) break; else { RRETURN(MATCH_NOMATCH); } }
2613              }
2614            break;
2615    
2616            case PT_UCNC:
2617            if ((c == CHAR_DOLLAR_SIGN || c == CHAR_COMMERCIAL_AT ||
2618                 c == CHAR_GRAVE_ACCENT || (c >= 0xa0 && c <= 0xd7ff) ||
2619                 c >= 0xe000) == (op == OP_NOTPROP))
2620              RRETURN(MATCH_NOMATCH);
2621            break;
2622    
2623          /* This should never occur */          /* This should never occur */
2624    
2625          default:          default:
# Line 2586  for (;;) Line 2639  for (;;)
2639        SCHECK_PARTIAL();        SCHECK_PARTIAL();
2640        RRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2641        }        }
2642      GETCHARINCTEST(c, eptr);      else
     if (UCD_CATEGORY(c) == ucp_M) RRETURN(MATCH_NOMATCH);  
     while (eptr < md->end_subject)  
2643        {        {
2644        int len = 1;        int lgb, rgb;
2645        if (!utf) c = *eptr; else { GETCHARLEN(c, eptr, len); }        GETCHARINCTEST(c, eptr);
2646        if (UCD_CATEGORY(c) != ucp_M) break;        lgb = UCD_GRAPHBREAK(c);
2647        eptr += len;        while (eptr < md->end_subject)
2648            {
2649            int len = 1;
2650            if (!utf) c = *eptr; else { GETCHARLEN(c, eptr, len); }
2651            rgb = UCD_GRAPHBREAK(c);
2652            if ((PRIV(ucp_gbtable)[lgb] & (1 << rgb)) == 0) break;
2653            lgb = rgb;
2654            eptr += len;
2655            }
2656        }        }
2657        CHECK_PARTIAL();
2658      ecode++;      ecode++;
2659      break;      break;
2660  #endif  #endif  /* SUPPORT_UCP */
2661    
2662    
2663      /* Match a back reference, possibly repeatedly. Look past the end of the      /* Match a back reference, possibly repeatedly. Look past the end of the
# Line 2606  for (;;) Line 2666  for (;;)
2666      similar code to character type repeats - written out again for speed.      similar code to character type repeats - written out again for speed.
2667      However, if the referenced string is the empty string, always treat      However, if the referenced string is the empty string, always treat
2668      it as matched, any number of times (otherwise there could be infinite      it as matched, any number of times (otherwise there could be infinite
2669      loops). */      loops). If the reference is unset, there are two possibilities:
   
     case OP_REF:  
     case OP_REFI:  
     caseless = op == OP_REFI;  
     offset = GET2(ecode, 1) << 1;               /* Doubled ref number */  
     ecode += 1 + IMM2_SIZE;  
   
     /* If the reference is unset, there are two possibilities:  
2670    
2671      (a) In the default, Perl-compatible state, set the length negative;      (a) In the default, Perl-compatible state, set the length negative;
2672      this ensures that every attempt at a match fails. We can't just fail      this ensures that every attempt at a match fails. We can't just fail
# Line 2624  for (;;) Line 2676  for (;;)
2676      so that the back reference matches an empty string.      so that the back reference matches an empty string.
2677    
2678      Otherwise, set the length to the length of what was matched by the      Otherwise, set the length to the length of what was matched by the
2679      referenced subpattern. */      referenced subpattern.
2680    
2681        The OP_REF and OP_REFI opcodes are used for a reference to a numbered group
2682        or to a non-duplicated named group. For a duplicated named group, OP_DNREF
2683        and OP_DNREFI are used. In this case we must scan the list of groups to
2684        which the name refers, and use the first one that is set. */
2685    
2686        case OP_DNREF:
2687        case OP_DNREFI:
2688        caseless = op == OP_DNREFI;
2689          {
2690          int count = GET2(ecode, 1+IMM2_SIZE);
2691          pcre_uchar *slot = md->name_table + GET2(ecode, 1) * md->name_entry_size;
2692          ecode += 1 + 2*IMM2_SIZE;
2693    
2694          /* Setting the default length first and initializing 'offset' avoids
2695          compiler warnings in the REF_REPEAT code. */
2696    
2697          length = (md->jscript_compat)? 0 : -1;
2698          offset = 0;
2699    
2700          while (count-- > 0)
2701            {
2702            offset = GET2(slot, 0) << 1;
2703            if (offset < offset_top && md->offset_vector[offset] >= 0)
2704              {
2705              length = md->offset_vector[offset+1] - md->offset_vector[offset];
2706              break;
2707              }
2708            slot += md->name_entry_size;
2709            }
2710          }
2711        goto REF_REPEAT;
2712    
2713        case OP_REF:
2714        case OP_REFI:
2715        caseless = op == OP_REFI;
2716        offset = GET2(ecode, 1) << 1;               /* Doubled ref number */
2717        ecode += 1 + IMM2_SIZE;
2718      if (offset >= offset_top || md->offset_vector[offset] < 0)      if (offset >= offset_top || md->offset_vector[offset] < 0)
2719        length = (md->jscript_compat)? 0 : -1;        length = (md->jscript_compat)? 0 : -1;
2720      else      else
# Line 2633  for (;;) Line 2722  for (;;)
2722    
2723      /* Set up for repetition, or handle the non-repeated case */      /* Set up for repetition, or handle the non-repeated case */
2724    
2725        REF_REPEAT:
2726      switch (*ecode)      switch (*ecode)
2727        {        {
2728        case OP_CRSTAR:        case OP_CRSTAR:
# Line 2660  for (;;) Line 2750  for (;;)
2750        default:               /* No repeat follows */        default:               /* No repeat follows */
2751        if ((length = match_ref(offset, eptr, length, md, caseless)) < 0)        if ((length = match_ref(offset, eptr, length, md, caseless)) < 0)
2752          {          {
2753            if (length == -2) eptr = md->end_subject;   /* Partial match */
2754          CHECK_PARTIAL();          CHECK_PARTIAL();
2755          RRETURN(MATCH_NOMATCH);          RRETURN(MATCH_NOMATCH);
2756          }          }
# Line 2685  for (;;) Line 2776  for (;;)
2776        int slength;        int slength;
2777        if ((slength = match_ref(offset, eptr, length, md, caseless)) < 0)        if ((slength = match_ref(offset, eptr, length, md, caseless)) < 0)
2778          {          {
2779            if (slength == -2) eptr = md->end_subject;   /* Partial match */
2780          CHECK_PARTIAL();          CHECK_PARTIAL();
2781          RRETURN(MATCH_NOMATCH);          RRETURN(MATCH_NOMATCH);
2782          }          }
# Line 2708  for (;;) Line 2800  for (;;)
2800          if (fi >= max) RRETURN(MATCH_NOMATCH);          if (fi >= max) RRETURN(MATCH_NOMATCH);
2801          if ((slength = match_ref(offset, eptr, length, md, caseless)) < 0)          if ((slength = match_ref(offset, eptr, length, md, caseless)) < 0)
2802            {            {
2803              if (slength == -2) eptr = md->end_subject;   /* Partial match */
2804            CHECK_PARTIAL();            CHECK_PARTIAL();
2805            RRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
2806            }            }
# Line 2726  for (;;) Line 2819  for (;;)
2819          int slength;          int slength;
2820          if ((slength = match_ref(offset, eptr, length, md, caseless)) < 0)          if ((slength = match_ref(offset, eptr, length, md, caseless)) < 0)
2821            {            {
2822            CHECK_PARTIAL();            /* Can't use CHECK_PARTIAL because we don't want to update eptr in
2823              the soft partial matching case. */
2824    
2825              if (slength == -2 && md->partial != 0 &&
2826                  md->end_subject > md->start_used_ptr)
2827                {
2828                md->hitend = TRUE;
2829                if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
2830                }
2831            break;            break;
2832            }            }
2833          eptr += slength;          eptr += slength;
2834          }          }
2835    
2836        while (eptr >= pp)        while (eptr >= pp)
2837          {          {
2838          RMATCH(eptr, ecode, offset_top, md, eptrb, RM15);          RMATCH(eptr, ecode, offset_top, md, eptrb, RM15);
# Line 2769  for (;;) Line 2871  for (;;)
2871          case OP_CRMINPLUS:          case OP_CRMINPLUS:
2872          case OP_CRQUERY:          case OP_CRQUERY:
2873          case OP_CRMINQUERY:          case OP_CRMINQUERY:
2874            case OP_CRPOSSTAR:
2875            case OP_CRPOSPLUS:
2876            case OP_CRPOSQUERY:
2877          c = *ecode++ - OP_CRSTAR;          c = *ecode++ - OP_CRSTAR;
2878          minimize = (c & 1) != 0;          if (c < OP_CRPOSSTAR - OP_CRSTAR) minimize = (c & 1) != 0;
2879            else possessive = TRUE;
2880          min = rep_min[c];                 /* Pick up values from tables; */          min = rep_min[c];                 /* Pick up values from tables; */
2881          max = rep_max[c];                 /* zero for max => infinity */          max = rep_max[c];                 /* zero for max => infinity */
2882          if (max == 0) max = INT_MAX;          if (max == 0) max = INT_MAX;
# Line 2778  for (;;) Line 2884  for (;;)
2884    
2885          case OP_CRRANGE:          case OP_CRRANGE:
2886          case OP_CRMINRANGE:          case OP_CRMINRANGE:
2887            case OP_CRPOSRANGE:
2888          minimize = (*ecode == OP_CRMINRANGE);          minimize = (*ecode == OP_CRMINRANGE);
2889            possessive = (*ecode == OP_CRPOSRANGE);
2890          min = GET2(ecode, 1);          min = GET2(ecode, 1);
2891          max = GET2(ecode, 1 + IMM2_SIZE);          max = GET2(ecode, 1 + IMM2_SIZE);
2892          if (max == 0) max = INT_MAX;          if (max == 0) max = INT_MAX;
# Line 2920  for (;;) Line 3028  for (;;)
3028                if ((BYTE_MAP[c/8] & (1 << (c&7))) == 0) break;                if ((BYTE_MAP[c/8] & (1 << (c&7))) == 0) break;
3029              eptr += len;              eptr += len;
3030              }              }
3031    
3032              if (possessive) continue;    /* No backtracking */
3033    
3034            for (;;)            for (;;)
3035              {              {
3036              RMATCH(eptr, ecode, offset_top, md, eptrb, RM18);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM18);
# Line 2950  for (;;) Line 3061  for (;;)
3061                if ((BYTE_MAP[c/8] & (1 << (c&7))) == 0) break;                if ((BYTE_MAP[c/8] & (1 << (c&7))) == 0) break;
3062              eptr++;              eptr++;
3063              }              }
3064    
3065              if (possessive) continue;    /* No backtracking */
3066    
3067            while (eptr >= pp)            while (eptr >= pp)
3068              {              {
3069              RMATCH(eptr, ecode, offset_top, md, eptrb, RM19);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM19);
# Line 2965  for (;;) Line 3079  for (;;)
3079      /* Control never gets here */      /* Control never gets here */
3080    
3081    
3082      /* Match an extended character class. This opcode is encountered only      /* Match an extended character class. In the 8-bit library, this opcode is
3083      when UTF-8 mode mode is supported. Nevertheless, we may not be in UTF-8      encountered only when UTF-8 mode mode is supported. In the 16-bit and
3084      mode, because Unicode properties are supported in non-UTF-8 mode. */      32-bit libraries, codepoints greater than 255 may be encountered even when
3085        UTF is not supported. */
3086    
3087  #if defined SUPPORT_UTF || !defined COMPILE_PCRE8  #if defined SUPPORT_UTF || !defined COMPILE_PCRE8
3088      case OP_XCLASS:      case OP_XCLASS:
# Line 2983  for (;;) Line 3098  for (;;)
3098          case OP_CRMINPLUS:          case OP_CRMINPLUS:
3099          case OP_CRQUERY:          case OP_CRQUERY:
3100          case OP_CRMINQUERY:          case OP_CRMINQUERY:
3101            case OP_CRPOSSTAR:
3102            case OP_CRPOSPLUS:
3103            case OP_CRPOSQUERY:
3104          c = *ecode++ - OP_CRSTAR;          c = *ecode++ - OP_CRSTAR;
3105          minimize = (c & 1) != 0;          if (c < OP_CRPOSSTAR - OP_CRSTAR) minimize = (c & 1) != 0;
3106            else possessive = TRUE;
3107          min = rep_min[c];                 /* Pick up values from tables; */          min = rep_min[c];                 /* Pick up values from tables; */
3108          max = rep_max[c];                 /* zero for max => infinity */          max = rep_max[c];                 /* zero for max => infinity */
3109          if (max == 0) max = INT_MAX;          if (max == 0) max = INT_MAX;
# Line 2992  for (;;) Line 3111  for (;;)
3111    
3112          case OP_CRRANGE:          case OP_CRRANGE:
3113          case OP_CRMINRANGE:          case OP_CRMINRANGE:
3114            case OP_CRPOSRANGE:
3115          minimize = (*ecode == OP_CRMINRANGE);          minimize = (*ecode == OP_CRMINRANGE);
3116            possessive = (*ecode == OP_CRPOSRANGE);
3117          min = GET2(ecode, 1);          min = GET2(ecode, 1);
3118          max = GET2(ecode, 1 + IMM2_SIZE);          max = GET2(ecode, 1 + IMM2_SIZE);
3119          if (max == 0) max = INT_MAX;          if (max == 0) max = INT_MAX;
# Line 3064  for (;;) Line 3185  for (;;)
3185            if (!PRIV(xclass)(c, data, utf)) break;            if (!PRIV(xclass)(c, data, utf)) break;
3186            eptr += len;            eptr += len;
3187            }            }
3188    
3189            if (possessive) continue;    /* No backtracking */
3190    
3191          for(;;)          for(;;)
3192            {            {
3193            RMATCH(eptr, ecode, offset_top, md, eptrb, RM21);            RMATCH(eptr, ecode, offset_top, md, eptrb, RM21);
# Line 3134  for (;;) Line 3258  for (;;)
3258    
3259        if (fc < 128)        if (fc < 128)
3260          {          {
3261          if (md->lcc[fc]          pcre_uint32 cc = *eptr;
3262              != TABLE_GET(*eptr, md->lcc, *eptr)) RRETURN(MATCH_NOMATCH);          if (md->lcc[fc] != TABLE_GET(cc, md->lcc, cc)) RRETURN(MATCH_NOMATCH);
3263          ecode++;          ecode++;
3264          eptr++;          eptr++;
3265          }          }
# Line 3146  for (;;) Line 3270  for (;;)
3270    
3271        else        else
3272          {          {
3273          unsigned int dc;          pcre_uint32 dc;
3274          GETCHARINC(dc, eptr);          GETCHARINC(dc, eptr);
3275          ecode += length;          ecode += length;
3276    
# Line 3239  for (;;) Line 3363  for (;;)
3363      max = rep_max[c];                 /* zero for max => infinity */      max = rep_max[c];                 /* zero for max => infinity */
3364      if (max == 0) max = INT_MAX;      if (max == 0) max = INT_MAX;
3365    
3366      /* Common code for all repeated single-character matches. */      /* Common code for all repeated single-character matches. We first check
3367        for the minimum number of characters. If the minimum equals the maximum, we
3368        are done. Otherwise, if minimizing, check the rest of the pattern for a
3369        match; if there isn't one, advance up to the maximum, one character at a
3370        time.
3371    
3372        If maximizing, advance up to the maximum number of matching characters,
3373        until eptr is past the end of the maximum run. If possessive, we are
3374        then done (no backing up). Otherwise, match at this position; anything
3375        other than no match is immediately returned. For nomatch, back up one
3376        character, unless we are matching \R and the last thing matched was
3377        \r\n, in which case, back up two bytes. When we reach the first optional
3378        character position, we can save stack by doing a tail recurse.
3379    
3380        The various UTF/non-UTF and caseful/caseless cases are handled separately,
3381        for speed. */
3382    
3383      REPEATCHAR:      REPEATCHAR:
3384  #ifdef SUPPORT_UTF  #ifdef SUPPORT_UTF
# Line 3256  for (;;) Line 3395  for (;;)
3395        if (length > 1)        if (length > 1)
3396          {          {
3397  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
3398          unsigned int othercase;          pcre_uint32 othercase;
3399          if (op >= OP_STARI &&     /* Caseless */          if (op >= OP_STARI &&     /* Caseless */
3400              (othercase = UCD_OTHERCASE(fc)) != fc)              (othercase = UCD_OTHERCASE(fc)) != fc)
3401            oclength = PRIV(ord2utf)(othercase, occhars);            oclength = PRIV(ord2utf)(othercase, occhars);
# Line 3323  for (;;) Line 3462  for (;;)
3462                }                }
3463              }              }
3464    
3465            if (possessive) continue;            if (possessive) continue;    /* No backtracking */
   
3466            for(;;)            for(;;)
3467              {              {
3468                if (eptr == pp) goto TAIL_RECURSE;
3469              RMATCH(eptr, ecode, offset_top, md, eptrb, RM23);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM23);
3470              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
             if (eptr == pp) { RRETURN(MATCH_NOMATCH); }  
3471  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
3472              eptr--;              eptr--;
3473              BACKCHAR(eptr);              BACKCHAR(eptr);
# Line 3360  for (;;) Line 3498  for (;;)
3498      maximizing, find the maximum number of characters and work backwards. */      maximizing, find the maximum number of characters and work backwards. */
3499    
3500      DPRINTF(("matching %c{%d,%d} against subject %.*s\n", fc, min, max,      DPRINTF(("matching %c{%d,%d} against subject %.*s\n", fc, min, max,
3501        max, eptr));        max, (char *)eptr));
3502    
3503      if (op >= OP_STARI)  /* Caseless */      if (op >= OP_STARI)  /* Caseless */
3504        {        {
# Line 3383  for (;;) Line 3521  for (;;)
3521    
3522        for (i = 1; i <= min; i++)        for (i = 1; i <= min; i++)
3523          {          {
3524            pcre_uint32 cc;                 /* Faster than pcre_uchar */
3525          if (eptr >= md->end_subject)          if (eptr >= md->end_subject)
3526            {            {
3527            SCHECK_PARTIAL();            SCHECK_PARTIAL();
3528            RRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
3529            }            }
3530          if (fc != *eptr && foc != *eptr) RRETURN(MATCH_NOMATCH);          cc = *eptr;
3531            if (fc != cc && foc != cc) RRETURN(MATCH_NOMATCH);
3532          eptr++;          eptr++;
3533          }          }
3534        if (min == max) continue;        if (min == max) continue;
# Line 3396  for (;;) Line 3536  for (;;)
3536          {          {
3537          for (fi = min;; fi++)          for (fi = min;; fi++)
3538            {            {
3539              pcre_uint32 cc;               /* Faster than pcre_uchar */
3540            RMATCH(eptr, ecode, offset_top, md, eptrb, RM24);            RMATCH(eptr, ecode, offset_top, md, eptrb, RM24);
3541            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3542            if (fi >= max) RRETURN(MATCH_NOMATCH);            if (fi >= max) RRETURN(MATCH_NOMATCH);
# Line 3404  for (;;) Line 3545  for (;;)
3545              SCHECK_PARTIAL();              SCHECK_PARTIAL();
3546              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
3547              }              }
3548            if (fc != *eptr && foc != *eptr) RRETURN(MATCH_NOMATCH);            cc = *eptr;
3549              if (fc != cc && foc != cc) RRETURN(MATCH_NOMATCH);
3550            eptr++;            eptr++;
3551            }            }
3552          /* Control never gets here */          /* Control never gets here */
# Line 3414  for (;;) Line 3556  for (;;)
3556          pp = eptr;          pp = eptr;
3557          for (i = min; i < max; i++)          for (i = min; i < max; i++)
3558            {            {
3559              pcre_uint32 cc;               /* Faster than pcre_uchar */
3560            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
3561              {              {
3562              SCHECK_PARTIAL();              SCHECK_PARTIAL();
3563              break;              break;
3564              }              }
3565            if (fc != *eptr && foc != *eptr) break;            cc = *eptr;
3566              if (fc != cc && foc != cc) break;
3567            eptr++;            eptr++;
3568            }            }
3569            if (possessive) continue;       /* No backtracking */
3570          if (possessive) continue;          for (;;)
   
         while (eptr >= pp)  
3571            {            {
3572              if (eptr == pp) goto TAIL_RECURSE;
3573            RMATCH(eptr, ecode, offset_top, md, eptrb, RM25);            RMATCH(eptr, ecode, offset_top, md, eptrb, RM25);
3574            eptr--;            eptr--;
3575            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3576            }            }
3577          RRETURN(MATCH_NOMATCH);          /* Control never gets here */
3578          }          }
       /* Control never gets here */  
3579        }        }
3580    
3581      /* Caseful comparisons (includes all multi-byte characters) */      /* Caseful comparisons (includes all multi-byte characters) */
# Line 3481  for (;;) Line 3623  for (;;)
3623            if (fc != *eptr) break;            if (fc != *eptr) break;
3624            eptr++;            eptr++;
3625            }            }
3626          if (possessive) continue;          if (possessive) continue;    /* No backtracking */
3627            for (;;)
         while (eptr >= pp)  
3628            {            {
3629              if (eptr == pp) goto TAIL_RECURSE;
3630            RMATCH(eptr, ecode, offset_top, md, eptrb, RM27);            RMATCH(eptr, ecode, offset_top, md, eptrb, RM27);
3631            eptr--;            eptr--;
3632            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3633            }            }
3634          RRETURN(MATCH_NOMATCH);          /* Control never gets here */
3635          }          }
3636        }        }
3637      /* Control never gets here */      /* Control never gets here */
# Line 3504  for (;;) Line 3646  for (;;)
3646        SCHECK_PARTIAL();        SCHECK_PARTIAL();
3647        RRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
3648        }        }
     ecode++;  
     GETCHARINCTEST(c, eptr);  
     if (op == OP_NOTI)         /* The caseless case */  
       {  
       register int ch, och;  
       ch = *ecode++;  
 #ifdef COMPILE_PCRE8  
       /* ch must be < 128 if UTF is enabled. */  
       och = md->fcc[ch];  
 #else  
3649  #ifdef SUPPORT_UTF  #ifdef SUPPORT_UTF
3650  #ifdef SUPPORT_UCP      if (utf)
3651        if (utf && ch > 127)        {
3652          och = UCD_OTHERCASE(ch);        register pcre_uint32 ch, och;
3653  #else  
3654        if (utf && ch > 127)        ecode++;
3655          och = ch;        GETCHARINC(ch, ecode);
3656  #endif /* SUPPORT_UCP */        GETCHARINC(c, eptr);
3657    
3658          if (op == OP_NOT)
3659            {
3660            if (ch == c) RRETURN(MATCH_NOMATCH);
3661            }
3662        else        else
3663  #endif /* SUPPORT_UTF */          {
3664          och = TABLE_GET(ch, md->fcc, ch);  #ifdef SUPPORT_UCP
3665  #endif /* COMPILE_PCRE8 */          if (ch > 127)
3666        if (ch == c || och == c) RRETURN(MATCH_NOMATCH);            och = UCD_OTHERCASE(ch);
3667    #else
3668            if (ch > 127)
3669              och = ch;
3670    #endif /* SUPPORT_UCP */
3671            else
3672              och = TABLE_GET(ch, md->fcc, ch);
3673            if (ch == c || och == c) RRETURN(MATCH_NOMATCH);
3674            }
3675        }        }
3676      else    /* Caseful */      else
3677    #endif
3678        {        {
3679        if (*ecode++ == c) RRETURN(MATCH_NOMATCH);        register pcre_uint32 ch = ecode[1];
3680          c = *eptr++;
3681          if (ch == c || (op == OP_NOTI && TABLE_GET(ch, md->fcc, ch) == c))
3682            RRETURN(MATCH_NOMATCH);
3683          ecode += 2;
3684        }        }
3685      break;      break;
3686    
# Line 3610  for (;;) Line 3760  for (;;)
3760      /* Common code for all repeated single-byte matches. */      /* Common code for all repeated single-byte matches. */
3761    
3762      REPEATNOTCHAR:      REPEATNOTCHAR:
3763      fc = *ecode++;      GETCHARINCTEST(fc, ecode);
3764    
3765      /* The code is duplicated for the caseless and caseful cases, for speed,      /* The code is duplicated for the caseless and caseful cases, for speed,
3766      since matching characters is likely to be quite common. First, ensure the      since matching characters is likely to be quite common. First, ensure the
# Line 3621  for (;;) Line 3771  for (;;)
3771      characters and work backwards. */      characters and work backwards. */
3772    
3773      DPRINTF(("negative matching %c{%d,%d} against subject %.*s\n", fc, min, max,      DPRINTF(("negative matching %c{%d,%d} against subject %.*s\n", fc, min, max,
3774        max, eptr));        max, (char *)eptr));
3775    
3776      if (op >= OP_NOTSTARI)     /* Caseless */      if (op >= OP_NOTSTARI)     /* Caseless */
3777        {        {
 #ifdef COMPILE_PCRE8  
       /* fc must be < 128 if UTF is enabled. */  
       foc = md->fcc[fc];  
 #else  
3778  #ifdef SUPPORT_UTF  #ifdef SUPPORT_UTF
3779  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
3780        if (utf && fc > 127)        if (utf && fc > 127)
# Line 3640  for (;;) Line 3786  for (;;)
3786        else        else
3787  #endif /* SUPPORT_UTF */  #endif /* SUPPORT_UTF */
3788          foc = TABLE_GET(fc, md->fcc, fc);          foc = TABLE_GET(fc, md->fcc, fc);
 #endif /* COMPILE_PCRE8 */  
3789    
3790  #ifdef SUPPORT_UTF  #ifdef SUPPORT_UTF
3791        if (utf)        if (utf)
3792          {          {
3793          register unsigned int d;          register pcre_uint32 d;
3794          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
3795            {            {
3796            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
# Line 3654  for (;;) Line 3799  for (;;)
3799              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
3800              }              }
3801            GETCHARINC(d, eptr);            GETCHARINC(d, eptr);
3802            if (fc == d || foc == d) RRETURN(MATCH_NOMATCH);            if (fc == d || (unsigned int)foc == d) RRETURN(MATCH_NOMATCH);
3803            }            }
3804          }          }
3805        else        else
3806  #endif  #endif  /* SUPPORT_UTF */
3807        /* Not UTF mode */        /* Not UTF mode */
3808          {          {
3809          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
# Line 3680  for (;;) Line 3825  for (;;)
3825  #ifdef SUPPORT_UTF  #ifdef SUPPORT_UTF
3826          if (utf)          if (utf)
3827            {            {
3828            register unsigned int d;            register pcre_uint32 d;
3829            for (fi = min;; fi++)            for (fi = min;; fi++)
3830              {              {
3831              RMATCH(eptr, ecode, offset_top, md, eptrb, RM28);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM28);
# Line 3692  for (;;) Line 3837  for (;;)
3837                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
3838                }                }
3839              GETCHARINC(d, eptr);              GETCHARINC(d, eptr);
3840              if (fc == d || foc == d) RRETURN(MATCH_NOMATCH);              if (fc == d || (unsigned int)foc == d) RRETURN(MATCH_NOMATCH);
3841              }              }
3842            }            }
3843          else          else
3844  #endif  #endif  /*SUPPORT_UTF */
3845          /* Not UTF mode */          /* Not UTF mode */
3846            {            {
3847            for (fi = min;; fi++)            for (fi = min;; fi++)
# Line 3725  for (;;) Line 3870  for (;;)
3870  #ifdef SUPPORT_UTF  #ifdef SUPPORT_UTF
3871          if (utf)          if (utf)
3872            {            {
3873            register unsigned int d;            register pcre_uint32 d;
3874            for (i = min; i < max; i++)            for (i = min; i < max; i++)
3875              {              {
3876              int len = 1;              int len = 1;
# Line 3735  for (;;) Line 3880  for (;;)
3880                break;                break;
3881                }                }
3882              GETCHARLEN(d, eptr, len);              GETCHARLEN(d, eptr, len);
3883              if (fc == d || foc == d) break;              if (fc == d || (unsigned int)foc == d) break;
3884              eptr += len;              eptr += len;
3885              }              }
3886            if (possessive) continue;            if (possessive) continue;    /* No backtracking */
3887            for(;;)            for(;;)
3888              {              {
3889                if (eptr == pp) goto TAIL_RECURSE;
3890              RMATCH(eptr, ecode, offset_top, md, eptrb, RM30);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM30);
3891              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3892              if (eptr-- == pp) break;        /* Stop if tried at original pos */              eptr--;
3893              BACKCHAR(eptr);              BACKCHAR(eptr);
3894              }              }
3895            }            }
3896          else          else
3897  #endif  #endif  /* SUPPORT_UTF */
3898          /* Not UTF mode */          /* Not UTF mode */
3899            {            {
3900            for (i = min; i < max; i++)            for (i = min; i < max; i++)
# Line 3761  for (;;) Line 3907  for (;;)
3907              if (fc == *eptr || foc == *eptr) break;              if (fc == *eptr || foc == *eptr) break;
3908              eptr++;              eptr++;
3909              }              }
3910            if (possessive) continue;            if (possessive) continue;    /* No backtracking */
3911            while (eptr >= pp)            for (;;)
3912              {              {
3913                if (eptr == pp) goto TAIL_RECURSE;
3914              RMATCH(eptr, ecode, offset_top, md, eptrb, RM31);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM31);
3915              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3916              eptr--;              eptr--;
3917              }              }
3918            }            }
3919            /* Control never gets here */
         RRETURN(MATCH_NOMATCH);  
3920          }          }
       /* Control never gets here */  
3921        }        }
3922    
3923      /* Caseful comparisons */      /* Caseful comparisons */
# Line 3782  for (;;) Line 3927  for (;;)
3927  #ifdef SUPPORT_UTF  #ifdef SUPPORT_UTF
3928        if (utf)        if (utf)
3929          {          {
3930          register unsigned int d;          register pcre_uint32 d;
3931          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
3932            {            {
3933            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
# Line 3816  for (;;) Line 3961  for (;;)
3961  #ifdef SUPPORT_UTF  #ifdef SUPPORT_UTF
3962          if (utf)          if (utf)
3963            {            {
3964            register unsigned int d;            register pcre_uint32 d;
3965            for (fi = min;; fi++)            for (fi = min;; fi++)
3966              {              {
3967              RMATCH(eptr, ecode, offset_top, md, eptrb, RM32);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM32);
# Line 3860  for (;;) Line 4005  for (;;)
4005  #ifdef SUPPORT_UTF  #ifdef SUPPORT_UTF
4006          if (utf)          if (utf)
4007            {            {
4008            register unsigned int d;            register pcre_uint32 d;
4009            for (i = min; i < max; i++)            for (i = min; i < max; i++)
4010              {              {
4011              int len = 1;              int len = 1;
# Line 3873  for (;;) Line 4018  for (;;)
4018              if (fc == d) break;              if (fc == d) break;
4019              eptr += len;              eptr += len;
4020              }              }
4021            if (possessive) continue;            if (possessive) continue;    /* No backtracking */
4022            for(;;)            for(;;)
4023              {              {
4024                if (eptr == pp) goto TAIL_RECURSE;
4025              RMATCH(eptr, ecode, offset_top, md, eptrb, RM34);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM34);
4026              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
4027              if (eptr-- == pp) break;        /* Stop if tried at original pos */              eptr--;
4028              BACKCHAR(eptr);              BACKCHAR(eptr);
4029              }              }
4030            }            }
# Line 3896  for (;;) Line 4042  for (;;)
4042              if (fc == *eptr) break;              if (fc == *eptr) break;
4043              eptr++;              eptr++;
4044              }              }
4045            if (possessive) continue;            if (possessive) continue;    /* No backtracking */
4046            while (eptr >= pp)            for (;;)
4047              {              {
4048                if (eptr == pp) goto TAIL_RECURSE;
4049              RMATCH(eptr, ecode, offset_top, md, eptrb, RM35);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM35);
4050              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
4051              eptr--;              eptr--;
4052              }              }
4053            }            }
4054            /* Control never gets here */
         RRETURN(MATCH_NOMATCH);  
4055          }          }
4056        }        }
4057      /* Control never gets here */      /* Control never gets here */
# Line 4087  for (;;) Line 4233  for (;;)
4233              }              }
4234            break;            break;
4235    
4236              /* Perl space used to exclude VT, but from Perl 5.18 it is included,
4237              which means that Perl space and POSIX space are now identical. PCRE
4238              was changed at release 8.34. */
4239    
4240            case PT_SPACE:    /* Perl space */            case PT_SPACE:    /* Perl space */
4241              case PT_PXSPACE:  /* POSIX space */
4242            for (i = 1; i <= min; i++)            for (i = 1; i <= min; i++)
4243              {              {
4244              if (eptr >= md->end_subject)              if (eptr >= md->end_subject)
# Line 4096  for (;;) Line 4247  for (;;)
4247                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
4248                }                }
4249              GETCHARINCTEST(c, eptr);              GETCHARINCTEST(c, eptr);
4250              if ((UCD_CATEGORY(c) == ucp_Z || c == CHAR_HT || c == CHAR_NL ||              switch(c)
4251                   c == CHAR_FF || c == CHAR_CR)                {
4252                     == prop_fail_result)                HSPACE_CASES:
4253                RRETURN(MATCH_NOMATCH);                VSPACE_CASES:
4254                  if (prop_fail_result) RRETURN(MATCH_NOMATCH);
4255                  break;
4256    
4257                  default:
4258                  if ((UCD_CATEGORY(c) == ucp_Z) == prop_fail_result)
4259                    RRETURN(MATCH_NOMATCH);
4260                  break;
4261                  }
4262              }              }
4263            break;            break;
4264    
4265            case PT_PXSPACE:  /* POSIX space */            case PT_WORD:
4266            for (i = 1; i <= min; i++)            for (i = 1; i <= min; i++)
4267              {              {
4268                int category;
4269              if (eptr >= md->end_subject)              if (eptr >= md->end_subject)
4270                {                {
4271                SCHECK_PARTIAL();                SCHECK_PARTIAL();
4272                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
4273                }                }
4274              GETCHARINCTEST(c, eptr);              GETCHARINCTEST(c, eptr);
4275              if ((UCD_CATEGORY(c) == ucp_Z || c == CHAR_HT || c == CHAR_NL ||              category = UCD_CATEGORY(c);
4276                   c == CHAR_VT || c == CHAR_FF || c == CHAR_CR)              if ((category == ucp_L || category == ucp_N || c == CHAR_UNDERSCORE)
4277                     == prop_fail_result)                     == prop_fail_result)
4278                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
4279              }              }
4280            break;            break;
4281    
4282            case PT_WORD:            case PT_CLIST:
4283            for (i = 1; i <= min; i++)            for (i = 1; i <= min; i++)
4284              {              {
4285              int category;              const pcre_uint32 *cp;
4286              if (eptr >= md->end_subject)              if (eptr >= md->end_subject)
4287                {                {
4288                SCHECK_PARTIAL();                SCHECK_PARTIAL();
4289                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
4290                }                }
4291              GETCHARINCTEST(c, eptr);              GETCHARINCTEST(c, eptr);
4292              category = UCD_CATEGORY(c);              cp = PRIV(ucd_caseless_sets) + prop_value;
4293              if ((category == ucp_L || category == ucp_N || c == CHAR_UNDERSCORE)              for (;;)
4294                     == prop_fail_result)                {
4295                  if (c < *cp)
4296                    { if (prop_fail_result) break; else { RRETURN(MATCH_NOMATCH); } }
4297                  if (c == *cp++)
4298                    { if (prop_fail_result) { RRETURN(MATCH_NOMATCH); } else break; }
4299                  }
4300                }
4301              break;
4302    
4303              case PT_UCNC:
4304              for (i = 1; i <= min; i++)
4305                {
4306                if (eptr >= md->end_subject)
4307                  {
4308                  SCHECK_PARTIAL();
4309                  RRETURN(MATCH_NOMATCH);
4310                  }
4311                GETCHARINCTEST(c, eptr);
4312                if ((c == CHAR_DOLLAR_SIGN || c == CHAR_COMMERCIAL_AT ||
4313                     c == CHAR_GRAVE_ACCENT || (c >= 0xa0 && c <= 0xd7ff) ||
4314                     c >= 0xe000) == prop_fail_result)
4315                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
4316              }              }
4317            break;            break;
# Line 4155  for (;;) Line 4335  for (;;)
4335              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4336              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4337              }              }
4338            GETCHARINCTEST(c, eptr);            else
           if (UCD_CATEGORY(c) == ucp_M) RRETURN(MATCH_NOMATCH);  
           while (eptr < md->end_subject)  
4339              {              {
4340              int len = 1;              int lgb, rgb;
4341              if (!utf) c = *eptr; else { GETCHARLEN(c, eptr, len); }              GETCHARINCTEST(c, eptr);
4342              if (UCD_CATEGORY(c) != ucp_M) break;              lgb = UCD_GRAPHBREAK(c);
4343              eptr += len;             while (eptr < md->end_subject)
4344                  {
4345                  int len = 1;
4346                  if (!utf) c = *eptr; else { GETCHARLEN(c, eptr, len); }
4347                  rgb = UCD_GRAPHBREAK(c);
4348                  if ((PRIV(ucp_gbtable)[lgb] & (1 << rgb)) == 0) break;
4349                  lgb = rgb;
4350                  eptr += len;
4351                  }
4352              }              }
4353              CHECK_PARTIAL();
4354            }            }
4355          }          }
4356    
# Line 4184  for (;;) Line 4371  for (;;)
4371              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4372              }              }
4373            if (IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH);            if (IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH);
4374              if (md->partial != 0 &&
4375                  eptr + 1 >= md->end_subject &&
4376                  NLBLOCK->nltype == NLTYPE_FIXED &&
4377                  NLBLOCK->nllen == 2 &&
4378                  *eptr == NLBLOCK->nl[0])
4379                {
4380                md->hitend = TRUE;
4381                if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
4382                }
4383            eptr++;            eptr++;
4384            ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);            ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);
4385            }            }
# Line 4220  for (;;) Line 4416  for (;;)
4416              {              {
4417              default: RRETURN(MATCH_NOMATCH);              default: RRETURN(MATCH_NOMATCH);
4418    
4419              case 0x000d:              case CHAR_CR:
4420              if (eptr < md->end_subject && *eptr == 0x0a) eptr++;              if (eptr < md->end_subject && *eptr == CHAR_LF) eptr++;
4421              break;              break;
4422    
4423              case 0x000a:              case CHAR_LF:
4424              break;              break;
4425    
4426              case 0x000b:              case CHAR_VT:
4427              case 0x000c:              case CHAR_FF:
4428              case 0x0085:              case CHAR_NEL:
4429    #ifndef EBCDIC
4430              case 0x2028:              case 0x2028:
4431              case 0x2029:              case 0x2029:
4432    #endif  /* Not EBCDIC */
4433              if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);              if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);
4434              break;              break;
4435              }              }
# Line 4249  for (;;) Line 4447  for (;;)
4447            GETCHARINC(c, eptr);            GETCHARINC(c, eptr);
4448            switch(c)            switch(c)
4449              {              {
4450                HSPACE_CASES: RRETURN(MATCH_NOMATCH);  /* Byte and multibyte cases */
4451              default: break;              default: break;
             case 0x09:      /* HT */  
             case 0x20:      /* SPACE */  
             case 0xa0:      /* NBSP */  
             case 0x1680:    /* OGHAM SPACE MARK */  
             case 0x180e:    /* MONGOLIAN VOWEL SEPARATOR */  
             case 0x2000:    /* EN QUAD */  
             case 0x2001:    /* EM QUAD */  
             case 0x2002:    /* EN SPACE */  
             case 0x2003:    /* EM SPACE */  
             case 0x2004:    /* THREE-PER-EM SPACE */  
             case 0x2005:    /* FOUR-PER-EM SPACE */  
             case 0x2006:    /* SIX-PER-EM SPACE */  
             case 0x2007:    /* FIGURE SPACE */  
             case 0x2008:    /* PUNCTUATION SPACE */  
             case 0x2009:    /* THIN SPACE */  
             case 0x200A:    /* HAIR SPACE */  
             case 0x202f:    /* NARROW NO-BREAK SPACE */  
             case 0x205f:    /* MEDIUM MATHEMATICAL SPACE */  
             case 0x3000:    /* IDEOGRAPHIC SPACE */  
             RRETURN(MATCH_NOMATCH);  
4452              }              }
4453            }            }
4454          break;          break;
# Line 4285  for (;;) Line 4464  for (;;)
4464            GETCHARINC(c, eptr);            GETCHARINC(c, eptr);
4465            switch(c)            switch(c)
4466              {              {
4467                HSPACE_CASES: break;  /* Byte and multibyte cases */
4468              default: RRETURN(MATCH_NOMATCH);              default: RRETURN(MATCH_NOMATCH);
             case 0x09:      /* HT */  
             case 0x20:      /* SPACE */  
             case 0xa0:      /* NBSP */  
             case 0x1680:    /* OGHAM SPACE MARK */  
             case 0x180e:    /* MONGOLIAN VOWEL SEPARATOR */  
             case 0x2000:    /* EN QUAD */  
             case 0x2001:    /* EM QUAD */  
             case 0x2002:    /* EN SPACE */  
             case 0x2003:    /* EM SPACE */  
             case 0x2004:    /* THREE-PER-EM SPACE */  
             case 0x2005:    /* FOUR-PER-EM SPACE */  
             case 0x2006:    /* SIX-PER-EM SPACE */  
             case 0x2007:    /* FIGURE SPACE */  
             case 0x2008:    /* PUNCTUATION SPACE */  
             case 0x2009:    /* THIN SPACE */  
             case 0x200A:    /* HAIR SPACE */  
             case 0x202f:    /* NARROW NO-BREAK SPACE */  
             case 0x205f:    /* MEDIUM MATHEMATICAL SPACE */  
             case 0x3000:    /* IDEOGRAPHIC SPACE */  
             break;  
4469              }              }
4470            }            }
4471          break;          break;
# Line 4321  for (;;) Line 4481  for (;;)
4481            GETCHARINC(c, eptr);            GETCHARINC(c, eptr);
4482            switch(c)            switch(c)
4483              {              {
4484                VSPACE_CASES: RRETURN(MATCH_NOMATCH);
4485              default: break;              default: break;
             case 0x0a:      /* LF */  
             case 0x0b:      /* VT */  
             case 0x0c:      /* FF */  
             case 0x0d:      /* CR */  
             case 0x85:      /* NEL */  
             case 0x2028:    /* LINE SEPARATOR */  
             case 0x2029:    /* PARAGRAPH SEPARATOR */  
             RRETURN(MATCH_NOMATCH);  
4486              }              }
4487            }            }
4488          break;          break;
# Line 4345  for (;;) Line 4498  for (;;)
4498            GETCHARINC(c, eptr);            GETCHARINC(c, eptr);
4499            switch(c)            switch(c)
4500              {              {
4501                VSPACE_CASES: break;
4502              default: RRETURN(MATCH_NOMATCH);              default: RRETURN(MATCH_NOMATCH);
             case 0x0a:      /* LF */  
             case 0x0b:      /* VT */  
             case 0x0c:      /* FF */  
             case 0x0d:      /* CR */  
             case 0x85:      /* NEL */  
             case 0x2028:    /* LINE SEPARATOR */  
             case 0x2029:    /* PARAGRAPH SEPARATOR */  
             break;  
4503              }              }
4504            }            }
4505          break;          break;
# Line 4375  for (;;) Line 4521  for (;;)
4521          case OP_DIGIT:          case OP_DIGIT:
4522          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
4523            {            {
4524              pcre_uint32 cc;
4525            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
4526              {              {
4527              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4528              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4529              }              }
4530            if (*eptr >= 128 || (md->ctypes[*eptr] & ctype_digit) == 0)            cc = *eptr;
4531              if (cc >= 128 || (md->ctypes[cc] & ctype_digit) == 0)
4532              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4533            eptr++;            eptr++;
4534            /* 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 */
# Line 4390  for (;;) Line 4538  for (;;)
4538          case OP_NOT_WHITESPACE:          case OP_NOT_WHITESPACE:
4539          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
4540            {            {
4541              pcre_uint32 cc;
4542            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
4543              {              {
4544              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4545              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4546              }              }
4547            if (*eptr < 128 && (md->ctypes[*eptr] & ctype_space) != 0)            cc = *eptr;
4548              if (cc < 128 && (md->ctypes[cc] & ctype_space) != 0)
4549              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4550            eptr++;            eptr++;
4551            ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);            ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);
# Line 4405  for (;;) Line 4555  for (;;)
4555          case OP_WHITESPACE:          case OP_WHITESPACE:
4556          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
4557            {            {
4558              pcre_uint32 cc;
4559            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
4560              {              {
4561              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4562              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4563              }              }
4564            if (*eptr >= 128 || (md->ctypes[*eptr] & ctype_space) == 0)            cc = *eptr;
4565              if (cc >= 128 || (md->ctypes[cc] & ctype_space) == 0)
4566              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4567            eptr++;            eptr++;
4568            /* 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 */
# Line 4420  for (;;) Line 4572  for (;;)
4572          case OP_NOT_WORDCHAR:          case OP_NOT_WORDCHAR:
4573          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
4574            {            {
4575              pcre_uint32 cc;
4576            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
4577              {              {
4578              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4579              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4580              }              }
4581            if (*eptr < 128 && (md->ctypes[*eptr] & ctype_word) != 0)            cc = *eptr;
4582              if (cc < 128 && (md->ctypes[cc] & ctype_word) != 0)
4583              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4584            eptr++;            eptr++;
4585            ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);            ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);
# Line 4435  for (;;) Line 4589  for (;;)
4589          case OP_WORDCHAR:          case OP_WORDCHAR:
4590          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
4591            {            {
4592              pcre_uint32 cc;
4593            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
4594              {              {
4595              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4596              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4597              }              }
4598            if (*eptr >= 128 || (md->ctypes[*eptr] & ctype_word) == 0)            cc = *eptr;
4599              if (cc >= 128 || (md->ctypes[cc] & ctype_word) == 0)
4600              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4601            eptr++;            eptr++;
4602            /* 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 */
# Line 4468  for (;;) Line 4624  for (;;)
4624              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4625              }              }
4626            if (IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH);            if (IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH);
4627              if (md->partial != 0 &&
4628                  eptr + 1 >= md->end_subject &&
4629                  NLBLOCK->nltype == NLTYPE_FIXED &&
4630                  NLBLOCK->nllen == 2 &&
4631                  *eptr == NLBLOCK->nl[0])
4632                {
4633                md->hitend = TRUE;
4634                if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
4635                }
4636            eptr++;            eptr++;
4637            }            }
4638          break;          break;
# Line 4502  for (;;) Line 4667  for (;;)
4667              {              {
4668              default: RRETURN(MATCH_NOMATCH);              default: RRETURN(MATCH_NOMATCH);
4669    
4670              case 0x000d:              case CHAR_CR:
4671              if (eptr < md->end_subject && *eptr == 0x0a) eptr++;              if (eptr < md->end_subject && *eptr == CHAR_LF) eptr++;
4672              break;              break;
4673    
4674              case 0x000a:              case CHAR_LF:
4675              break;              break;
4676    
4677              case 0x000b:              case CHAR_VT:
4678              case 0x000c:              case CHAR_FF:
4679              case 0x0085:              case CHAR_NEL:
4680  #ifdef COMPILE_PCRE16  #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
4681              case 0x2028:              case 0x2028:
4682              case 0x2029:              case 0x2029:
4683  #endif  #endif
# Line 4533  for (;;) Line 4698  for (;;)
4698            switch(*eptr++)            switch(*eptr++)
4699              {              {
4700              default: break;              default: break;
4701              case 0x09:      /* HT */              HSPACE_BYTE_CASES:
4702              case 0x20:      /* SPACE */  #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
4703              case 0xa0:      /* NBSP */              HSPACE_MULTIBYTE_CASES:
 #ifdef COMPILE_PCRE16  
             case 0x1680:    /* OGHAM SPACE MARK */  
             case 0x180e:    /* MONGOLIAN VOWEL SEPARATOR */  
             case 0x2000:    /* EN QUAD */  
             case 0x2001:    /* EM QUAD */  
             case 0x2002:    /* EN SPACE */  
             case 0x2003:    /* EM SPACE */  
             case 0x2004:    /* THREE-PER-EM SPACE */  
             case 0x2005:    /* FOUR-PER-EM SPACE */  
             case 0x2006:    /* SIX-PER-EM SPACE */  
             case 0x2007:    /* FIGURE SPACE */  
             case 0x2008:    /* PUNCTUATION SPACE */  
             case 0x2009:    /* THIN SPACE */  
             case 0x200A:    /* HAIR SPACE */  
             case 0x202f:    /* NARROW NO-BREAK SPACE */  
             case 0x205f:    /* MEDIUM MATHEMATICAL SPACE */  
             case 0x3000:    /* IDEOGRAPHIC SPACE */  
4704  #endif  #endif
4705              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4706              }              }
# Line 4570  for (;;) Line 4718  for (;;)
4718            switch(*eptr++)            switch(*eptr++)
4719              {              {
4720              default: RRETURN(MATCH_NOMATCH);              default: RRETURN(MATCH_NOMATCH);
4721              case 0x09:      /* HT */              HSPACE_BYTE_CASES:
4722              case 0x20:      /* SPACE */  #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
4723              case 0xa0:      /* NBSP */              HSPACE_MULTIBYTE_CASES:
 #ifdef COMPILE_PCRE16  
             case 0x1680:    /* OGHAM SPACE MARK */  
             case 0x180e:    /* MONGOLIAN VOWEL SEPARATOR */  
             case 0x2000:    /* EN QUAD */  
             case 0x2001:    /* EM QUAD */  
             case 0x2002:    /* EN SPACE */  
             case 0x2003:    /* EM SPACE */  
             case 0x2004:    /* THREE-PER-EM SPACE */  
             case 0x2005:    /* FOUR-PER-EM SPACE */  
             case 0x2006:    /* SIX-PER-EM SPACE */  
             case 0x2007:    /* FIGURE SPACE */  
             case 0x2008:    /* PUNCTUATION SPACE */  
             case 0x2009:    /* THIN SPACE */  
             case 0x200A:    /* HAIR SPACE */  
             case 0x202f:    /* NARROW NO-BREAK SPACE */  
             case 0x205f:    /* MEDIUM MATHEMATICAL SPACE */  
             case 0x3000:    /* IDEOGRAPHIC SPACE */  
4724  #endif  #endif
4725              break;              break;
4726              }              }
# Line 4606  for (;;) Line 4737  for (;;)
4737              }              }
4738            switch(*eptr++)            switch(*eptr++)
4739              {              {
4740              default: break;              VSPACE_BYTE_CASES:
4741              case 0x0a:      /* LF */  #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
4742              case 0x0b:      /* VT */              VSPACE_MULTIBYTE_CASES:
             case 0x0c:      /* FF */  
             case 0x0d:      /* CR */  
             case 0x85:      /* NEL */  
 #ifdef COMPILE_PCRE16  
             case 0x2028:    /* LINE SEPARATOR */  
             case 0x2029:    /* PARAGRAPH SEPARATOR */  
4743  #endif  #endif
4744              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4745                default: break;
4746              }              }
4747            }            }
4748          break;          break;
# Line 4632  for (;;) Line 4758  for (;;)
4758            switch(*eptr++)            switch(*eptr++)
4759              {              {
4760              default: RRETURN(MATCH_NOMATCH);              default: RRETURN(MATCH_NOMATCH);
4761              case 0x0a:      /* LF */              VSPACE_BYTE_CASES:
4762              case 0x0b:      /* VT */  #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
4763              case 0x0c:      /* FF */              VSPACE_MULTIBYTE_CASES:
             case 0x0d:      /* CR */  
             case 0x85:      /* NEL */  
 #ifdef COMPILE_PCRE16  
             case 0x2028:    /* LINE SEPARATOR */  
             case 0x2029:    /* PARAGRAPH SEPARATOR */  
4764  #endif  #endif
4765              break;              break;
4766              }              }
# Line 4857  for (;;) Line 4978  for (;;)
4978              }              }
4979            /* Control never gets here */            /* Control never gets here */
4980    
4981              /* Perl space used to exclude VT, but from Perl 5.18 it is included,
4982              which means that Perl space and POSIX space are now identical. PCRE
4983              was changed at release 8.34. */
4984    
4985            case PT_SPACE:    /* Perl space */            case PT_SPACE:    /* Perl space */
4986              case PT_PXSPACE:  /* POSIX space */
4987            for (fi = min;; fi++)            for (fi = min;; fi++)
4988              {              {
4989              RMATCH(eptr, ecode, offset_top, md, eptrb, RM60);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM61);
4990              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
4991              if (fi >= max) RRETURN(MATCH_NOMATCH);              if (fi >= max) RRETURN(MATCH_NOMATCH);
4992              if (eptr >= md->end_subject)              if (eptr >= md->end_subject)
# Line 4869  for (;;) Line 4995  for (;;)
4995                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
4996                }                }
4997              GETCHARINCTEST(c, eptr);              GETCHARINCTEST(c, eptr);
4998              if ((UCD_CATEGORY(c) == ucp_Z || c == CHAR_HT || c == CHAR_NL ||              switch(c)
4999                   c == CHAR_FF || c == CHAR_CR)                {
5000                     == prop_fail_result)                HSPACE_CASES:
5001                RRETURN(MATCH_NOMATCH);                VSPACE_CASES:
5002                  if (prop_fail_result) RRETURN(MATCH_NOMATCH);
5003                  break;
5004    
5005                  default:
5006                  if ((UCD_CATEGORY(c) == ucp_Z) == prop_fail_result)
5007                    RRETURN(MATCH_NOMATCH);
5008                  break;
5009                  }
5010              }              }
5011            /* Control never gets here */            /* Control never gets here */
5012    
5013            case PT_PXSPACE:  /* POSIX space */            case PT_WORD:
5014            for (fi = min;; fi++)            for (fi = min;; fi++)
5015              {              {
5016              RMATCH(eptr, ecode, offset_top, md, eptrb, RM61);              int category;
5017                RMATCH(eptr, ecode, offset_top, md, eptrb, RM62);
5018              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
5019              if (fi >= max) RRETURN(MATCH_NOMATCH);              if (fi >= max) RRETURN(MATCH_NOMATCH);
5020              if (eptr >= md->end_subject)              if (eptr >= md->end_subject)
# Line 4888  for (;;) Line 5023  for (;;)
5023                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
5024                }                }
5025              GETCHARINCTEST(c, eptr);              GETCHARINCTEST(c, eptr);
5026              if ((UCD_CATEGORY(c) == ucp_Z || c == CHAR_HT || c == CHAR_NL ||              category = UCD_CATEGORY(c);
5027                   c == CHAR_VT || c == CHAR_FF || c == CHAR_CR)              if ((category == ucp_L ||
5028                     category == ucp_N ||
5029                     c == CHAR_UNDERSCORE)
5030                     == prop_fail_result)                     == prop_fail_result)
5031                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
5032              }              }
5033            /* Control never gets here */            /* Control never gets here */
5034    
5035            case PT_WORD:            case PT_CLIST:
5036            for (fi = min;; fi++)            for (fi = min;; fi++)
5037              {              {
5038              int category;              const pcre_uint32 *cp;
5039              RMATCH(eptr, ecode, offset_top, md, eptrb, RM62);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM67);
5040              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
5041              if (fi >= max) RRETURN(MATCH_NOMATCH);              if (fi >= max) RRETURN(MATCH_NOMATCH);
5042              if (eptr >= md->end_subject)              if (eptr >= md->end_subject)
# Line 4908  for (;;) Line 5045  for (;;)
5045                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
5046                }                }
5047              GETCHARINCTEST(c, eptr);              GETCHARINCTEST(c, eptr);
5048              category = UCD_CATEGORY(c);              cp = PRIV(ucd_caseless_sets) + prop_value;
5049              if ((category == ucp_L ||              for (;;)
5050                   category == ucp_N ||                {
5051                   c == CHAR_UNDERSCORE)                if (c < *cp)
5052                     == prop_fail_result)                  { if (prop_fail_result) break; else { RRETURN(MATCH_NOMATCH); } }
5053                  if (c == *cp++)
5054                    { if (prop_fail_result) { RRETURN(MATCH_NOMATCH); } else break; }
5055                  }
5056                }
5057              /* Control never gets here */
5058    
5059              case PT_UCNC:
5060              for (fi = min;; fi++)
5061                {
5062                RMATCH(eptr, ecode, offset_top, md, eptrb, RM60);
5063                if (rrc != MATCH_NOMATCH) RRETURN(rrc);
5064                if (fi >= max) RRETURN(MATCH_NOMATCH);
5065                if (eptr >= md->end_subject)
5066                  {
5067                  SCHECK_PARTIAL();
5068                  RRETURN(MATCH_NOMATCH);
5069                  }
5070                GETCHARINCTEST(c, eptr);
5071                if ((c == CHAR_DOLLAR_SIGN || c == CHAR_COMMERCIAL_AT ||
5072                     c == CHAR_GRAVE_ACCENT || (c >= 0xa0 && c <= 0xd7ff) ||
5073                     c >= 0xe000) == prop_fail_result)
5074                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
5075              }              }
5076            /* Control never gets here */            /* Control never gets here */
5077    
5078            /* This should never occur */            /* This should never occur */
   
5079            default:            default:
5080            RRETURN(PCRE_ERROR_INTERNAL);            RRETURN(PCRE_ERROR_INTERNAL);
5081            }            }
# Line 4939  for (;;) Line 5096  for (;;)
5096              SCHECK_PARTIAL();              SCHECK_PARTIAL();
5097              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
5098              }              }
5099            GETCHARINCTEST(c, eptr);            else
           if (UCD_CATEGORY(c) == ucp_M) RRETURN(MATCH_NOMATCH);  
           while (eptr < md->end_subject)  
5100              {              {
5101              int len = 1;              int lgb, rgb;
5102              if (!utf) c = *eptr; else { GETCHARLEN(c, eptr, len); }              GETCHARINCTEST(c, eptr);
5103              if (UCD_CATEGORY(c) != ucp_M) break;              lgb = UCD_GRAPHBREAK(c);
5104              eptr += len;              while (eptr < md->end_subject)
5105                  {
5106                  int len = 1;
5107                  if (!utf) c = *eptr; else { GETCHARLEN(c, eptr, len); }
5108                  rgb = UCD_GRAPHBREAK(c);
5109                  if ((PRIV(ucp_gbtable)[lgb] & (1 << rgb)) == 0) break;
5110                  lgb = rgb;
5111                  eptr += len;
5112                  }
5113              }              }
5114              CHECK_PARTIAL();
5115            }            }
5116          }          }
5117        else        else
# Line 4971  for (;;) Line 5135  for (;;)
5135            GETCHARINC(c, eptr);            GETCHARINC(c, eptr);
5136            switch(ctype)            switch(ctype)
5137              {              {
5138              case OP_ANY:        /* This is the non-NL case */              case OP_ANY:               /* This is the non-NL case */
5139                if (md->partial != 0 &&    /* Take care with CRLF partial */
5140                    eptr >= md->end_subject &&
5141                    NLBLOCK->nltype == NLTYPE_FIXED &&
5142                    NLBLOCK->nllen == 2 &&
5143                    c == NLBLOCK->nl[0])
5144                  {
5145                  md->hitend = TRUE;
5146                  if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
5147                  }
5148                break;
5149    
5150              case OP_ALLANY:              case OP_ALLANY:
5151              case OP_ANYBYTE:              case OP_ANYBYTE:
5152              break;              break;
# Line 4980  for (;;) Line 5155  for (;;)
5155              switch(c)              switch(c)
5156                {                {
5157                default: RRETURN(MATCH_NOMATCH);                default: RRETURN(MATCH_NOMATCH);
5158                case 0x000d:                case CHAR_CR:
5159                if (eptr < md->end_subject && *eptr == 0x0a) eptr++;                if (eptr < md->end_subject && *eptr == CHAR_LF) eptr++;
5160                break;                break;
5161                case 0x000a:  
5162                  case CHAR_LF:
5163                break;                break;
5164    
5165                case 0x000b:                case CHAR_VT:
5166                case 0x000c:                case CHAR_FF:
5167                case 0x0085:                case CHAR_NEL:
5168    #ifndef EBCDIC
5169                case 0x2028:                case 0x2028:
5170                case 0x2029:                case 0x2029:
5171    #endif  /* Not EBCDIC */
5172                if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);                if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);
5173                break;                break;
5174                }                }
# Line 4999  for (;;) Line 5177  for (;;)
5177              case OP_NOT_HSPACE:              case OP_NOT_HSPACE:
5178              switch(c)              switch(c)
5179                {                {
5180                  HSPACE_CASES: RRETURN(MATCH_NOMATCH);
5181                default: break;                default: break;
               case 0x09:      /* HT */  
               case 0x20:      /* SPACE */  
               case 0xa0:      /* NBSP */  
               case 0x1680:    /* OGHAM SPACE MARK */  
               case 0x180e:    /* MONGOLIAN VOWEL SEPARATOR */  
               case 0x2000:    /* EN QUAD */  
               case 0x2001:    /* EM QUAD */  
               case 0x2002:    /* EN SPACE */  
               case 0x2003:    /* EM SPACE */  
               case 0x2004:    /* THREE-PER-EM SPACE */  
               case 0x2005:    /* FOUR-PER-EM SPACE */  
               case 0x2006:    /* SIX-PER-EM SPACE */  
               case 0x2007:    /* FIGURE SPACE */  
               case 0x2008:    /* PUNCTUATION SPACE */  
               case 0x2009:    /* THIN SPACE */  
               case 0x200A:    /* HAIR SPACE */  
               case 0x202f:    /* NARROW NO-BREAK SPACE */  
               case 0x205f:    /* MEDIUM MATHEMATICAL SPACE */  
               case 0x3000:    /* IDEOGRAPHIC SPACE */  
               RRETURN(MATCH_NOMATCH);  
5182                }                }
5183              break;              break;
5184    
5185              case OP_HSPACE:              case OP_HSPACE:
5186              switch(c)              switch(c)
5187                {                {
5188                  HSPACE_CASES: break;
5189                default: RRETURN(MATCH_NOMATCH);                default: RRETURN(MATCH_NOMATCH);
               case 0x09:      /* HT */  
               case 0x20:      /* SPACE */  
               case 0xa0:      /* NBSP */  
               case 0x1680:    /* OGHAM SPACE MARK */  
               case 0x180e:    /* MONGOLIAN VOWEL SEPARATOR */  
               case 0x2000:    /* EN QUAD */  
               case 0x2001:    /* EM QUAD */  
               case 0x2002:    /* EN SPACE */  
               case 0x2003:    /* EM SPACE */  
               case 0x2004:    /* THREE-PER-EM SPACE */  
               case 0x2005:    /* FOUR-PER-EM SPACE */  
               case 0x2006:    /* SIX-PER-EM SPACE */  
               case 0x2007:    /* FIGURE SPACE */  
               case 0x2008:    /* PUNCTUATION SPACE */  
               case 0x2009:    /* THIN SPACE */  
               case 0x200A:    /* HAIR SPACE */  
               case 0x202f:    /* NARROW NO-BREAK SPACE */  
               case 0x205f:    /* MEDIUM MATHEMATICAL SPACE */  
               case 0x3000:    /* IDEOGRAPHIC SPACE */  
               break;  
5190                }                }
5191              break;              break;
5192    
5193              case OP_NOT_VSPACE:              case OP_NOT_VSPACE:
5194              switch(c)              switch(c)
5195                {                {
5196                  VSPACE_CASES: RRETURN(MATCH_NOMATCH);
5197                default: break;                default: break;
               case 0x0a:      /* LF */  
               case 0x0b:      /* VT */  
               case 0x0c:      /* FF */  
               case 0x0d:      /* CR */  
               case 0x85:      /* NEL */  
               case 0x2028:    /* LINE SEPARATOR */  
               case 0x2029:    /* PARAGRAPH SEPARATOR */  
               RRETURN(MATCH_NOMATCH);  
5198                }                }
5199              break;              break;
5200    
5201              case OP_VSPACE:              case OP_VSPACE:
5202              switch(c)              switch(c)
5203                {                {
5204                  VSPACE_CASES: break;
5205                default: RRETURN(MATCH_NOMATCH);                default: RRETURN(MATCH_NOMATCH);
               case 0x0a:      /* LF */  
               case 0x0b:      /* VT */  
               case 0x0c:      /* FF */  
               case 0x0d:      /* CR */  
               case 0x85:      /* NEL */  
               case 0x2028:    /* LINE SEPARATOR */  
               case 0x2029:    /* PARAGRAPH SEPARATOR */  
               break;  
5206                }                }
5207              break;              break;
5208    
# Line 5134  for (;;) Line 5260  for (;;)
5260            c = *eptr++;            c = *eptr++;
5261            switch(ctype)            switch(ctype)
5262              {              {
5263              case OP_ANY:     /* This is the non-NL case */              case OP_ANY:               /* This is the non-NL case */
5264                if (md->partial != 0 &&    /* Take care with CRLF partial */
5265                    eptr >= md->end_subject &&
5266                    NLBLOCK->nltype == NLTYPE_FIXED &&
5267                    NLBLOCK->nllen == 2 &&
5268                    c == NLBLOCK->nl[0])
5269                  {
5270                  md->hitend = TRUE;
5271                  if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
5272                  }
5273                break;
5274    
5275              case OP_ALLANY:              case OP_ALLANY:
5276              case OP_ANYBYTE:              case OP_ANYBYTE:
5277              break;              break;
# Line 5143  for (;;) Line 5280  for (;;)
5280              switch(c)              switch(c)
5281                {                {
5282                default: RRETURN(MATCH_NOMATCH);                default: RRETURN(MATCH_NOMATCH);
5283                case 0x000d:                case CHAR_CR:
5284                if (eptr < md->end_subject && *eptr == 0x0a) eptr++;                if (eptr < md->end_subject && *eptr == CHAR_LF) eptr++;
5285                break;                break;
5286    
5287                case 0x000a:                case CHAR_LF:
5288                break;                break;
5289    
5290                case 0x000b:                case CHAR_VT:
5291                case 0x000c:                case CHAR_FF:
5292                case 0x0085:                case CHAR_NEL:
5293  #ifdef COMPILE_PCRE16  #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
5294                case 0x2028:                case 0x2028:
5295                case 0x2029:                case 0x2029:
5296  #endif  #endif
# Line 5166  for (;;) Line 5303  for (;;)
5303              switch(c)              switch(c)
5304                {                {
5305                default: break;                default: break;
5306                case 0x09:      /* HT */                HSPACE_BYTE_CASES:
5307                case 0x20:      /* SPACE */  #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
5308                case 0xa0:      /* NBSP */                HSPACE_MULTIBYTE_CASES:
 #ifdef COMPILE_PCRE16  
               case 0x1680:    /* OGHAM SPACE MARK */  
               case 0x180e:    /* MONGOLIAN VOWEL SEPARATOR */  
               case 0x2000:    /* EN QUAD */  
               case 0x2001:    /* EM QUAD */  
               case 0x2002:    /* EN SPACE */  
               case 0x2003:    /* EM SPACE */  
               case 0x2004:    /* THREE-PER-EM SPACE */  
               case 0x2005:    /* FOUR-PER-EM SPACE */  
               case 0x2006:    /* SIX-PER-EM SPACE */  
               case 0x2007:    /* FIGURE SPACE */  
               case 0x2008:    /* PUNCTUATION SPACE */  
               case 0x2009:    /* THIN SPACE */  
               case 0x200A:    /* HAIR SPACE */  
               case 0x202f:    /* NARROW NO-BREAK SPACE */  
               case 0x205f:    /* MEDIUM MATHEMATICAL SPACE */  
               case 0x3000:    /* IDEOGRAPHIC SPACE */  
5309  #endif  #endif
5310                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
5311                }                }
# Line 5195  for (;;) Line 5315  for (;;)
5315              switch(c)              switch(c)
5316                {                {
5317                default: RRETURN(MATCH_NOMATCH);                default: RRETURN(MATCH_NOMATCH);
5318                case 0x09:      /* HT */                HSPACE_BYTE_CASES:
5319                case 0x20:      /* SPACE */  #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
5320                case 0xa0:      /* NBSP */                HSPACE_MULTIBYTE_CASES:
 #ifdef COMPILE_PCRE16  
               case 0x1680:    /* OGHAM SPACE MARK */  
               case 0x180e:    /* MONGOLIAN VOWEL SEPARATOR */  
               case 0x2000:    /* EN QUAD */  
               case 0x2001:    /* EM QUAD */  
               case 0x2002:    /* EN SPACE */  
               case 0x2003:    /* EM SPACE */  
               case 0x2004:    /* THREE-PER-EM SPACE */  
               case 0x2005:    /* FOUR-PER-EM SPACE */  
               case 0x2006:    /* SIX-PER-EM SPACE */  
               case 0x2007:    /* FIGURE SPACE */  
               case 0x2008:    /* PUNCTUATION SPACE */  
               case 0x2009:    /* THIN SPACE */  
               case 0x200A:    /* HAIR SPACE */  
               case 0x202f:    /* NARROW NO-BREAK SPACE */  
               case 0x205f:    /* MEDIUM MATHEMATICAL SPACE */  
               case 0x3000:    /* IDEOGRAPHIC SPACE */  
5321  #endif  #endif
5322                break;                break;
5323                }                }
# Line 5224  for (;;) Line 5327  for (;;)
5327              switch(c)              switch(c)
5328                {                {
5329                default: break;                default: break;
5330                case 0x0a:      /* LF */                VSPACE_BYTE_CASES:
5331                case 0x0b:      /* VT */  #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
5332                case 0x0c:      /* FF */                VSPACE_MULTIBYTE_CASES:
               case 0x0d:      /* CR */  
               case 0x85:      /* NEL */  
 #ifdef COMPILE_PCRE16  
               case 0x2028:    /* LINE SEPARATOR */  
               case 0x2029:    /* PARAGRAPH SEPARATOR */  
5333  #endif  #endif
5334                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
5335                }                }
# Line 5241  for (;;) Line 5339  for (;;)
5339              switch(c)              switch(c)
5340                {                {
5341                default: RRETURN(MATCH_NOMATCH);                default: RRETURN(MATCH_NOMATCH);
5342                case 0x0a:      /* LF */                VSPACE_BYTE_CASES:
5343                case 0x0b:      /* VT */  #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
5344                case 0x0c:      /* FF */                VSPACE_MULTIBYTE_CASES:
               case 0x0d:      /* CR */  
               case 0x85:      /* NEL */  
 #ifdef COMPILE_PCRE16  
               case 0x2028:    /* LINE SEPARATOR */  
               case 0x2029:    /* PARAGRAPH SEPARATOR */  
5345  #endif  #endif
5346                break;                break;
5347                }                }
# Line 5397  for (;;) Line 5490  for (;;)
5490              }              }
5491            break;            break;
5492    
5493              /* Perl space used to exclude VT, but from Perl 5.18 it is included,
5494              which means that Perl space and POSIX space are now identical. PCRE
5495              was changed at release 8.34. */
5496    
5497            case PT_SPACE:    /* Perl space */            case PT_SPACE:    /* Perl space */
5498              case PT_PXSPACE:  /* POSIX space */
5499            for (i = min; i < max; i++)            for (i = min; i < max; i++)
5500              {              {
5501              int len = 1;              int len = 1;
# Line 5407  for (;;) Line 5505  for (;;)
5505                break;                break;
5506                }                }
5507              GETCHARLENTEST(c, eptr, len);              GETCHARLENTEST(c, eptr, len);
5508              if ((UCD_CATEGORY(c) == ucp_Z || c == CHAR_HT || c == CHAR_NL ||              switch(c)
5509                   c == CHAR_FF || c == CHAR_CR)                {
5510                   == prop_fail_result)                HSPACE_CASES:
5511                  VSPACE_CASES:
5512                  if (prop_fail_result) goto ENDLOOP99;  /* Break the loop */
5513                  break;
5514    
5515                  default:
5516                  if ((UCD_CATEGORY(c) == ucp_Z) == prop_fail_result)
5517                    goto ENDLOOP99;   /* Break the loop */
5518                break;                break;
5519                  }
5520              eptr+= len;              eptr+= len;
5521              }              }
5522              ENDLOOP99:
5523            break;            break;
5524    
5525            case PT_PXSPACE:  /* POSIX space */            case PT_WORD:
5526            for (i = min; i < max; i++)            for (i = min; i < max; i++)
5527              {              {
5528                int category;
5529              int len = 1;              int len = 1;
5530              if (eptr >= md->end_subject)              if (eptr >= md->end_subject)
5531                {                {
# Line 5425  for (;;) Line 5533  for (;;)
5533                break;                break;
5534                }                }
5535              GETCHARLENTEST(c, eptr, len);              GETCHARLENTEST(c, eptr, len);
5536              if ((UCD_CATEGORY(c) == ucp_Z || c == CHAR_HT || c == CHAR_NL ||              category = UCD_CATEGORY(c);
5537                   c == CHAR_VT || c == CHAR_FF || c == CHAR_CR)              if ((category == ucp_L || category == ucp_N ||
5538                   == prop_fail_result)                   c == CHAR_UNDERSCORE) == prop_fail_result)
5539                break;                break;
5540              eptr+= len;              eptr+= len;
5541              }              }
5542            break;            break;
5543    
5544            case PT_WORD:            case PT_CLIST:
5545            for (i = min; i < max; i++)            for (i = min; i < max; i++)
5546              {              {
5547              int category;              const pcre_uint32 *cp;
5548              int len = 1;              int len = 1;
5549              if (eptr >= md->end_subject)              if (eptr >= md->end_subject)
5550                {                {
# Line 5444  for (;;) Line 5552  for (;;)
5552                break;                break;
5553                }                }
5554              GETCHARLENTEST(c, eptr, len);              GETCHARLENTEST(c, eptr, len);
5555              category = UCD_CATEGORY(c);              cp = PRIV(ucd_caseless_sets) + prop_value;
5556              if ((category == ucp_L || category == ucp_N ||              for (;;)
5557                   c == CHAR_UNDERSCORE) == prop_fail_result)                {
5558                  if (c < *cp)
5559                    { if (prop_fail_result) break; else goto GOT_MAX; }
5560                  if (c == *cp++)
5561                    { if (prop_fail_result) goto GOT_MAX; else break; }
5562                  }
5563                eptr += len;
5564                }
5565              GOT_MAX:
5566              break;
5567    
5568              case PT_UCNC:
5569              for (i = min; i < max; i++)
5570                {
5571                int len = 1;
5572                if (eptr >= md->end_subject)
5573                  {
5574                  SCHECK_PARTIAL();
5575                break;                break;
5576              eptr+= len;                }
5577                GETCHARLENTEST(c, eptr, len);
5578                if ((c == CHAR_DOLLAR_SIGN || c == CHAR_COMMERCIAL_AT ||
5579                     c == CHAR_GRAVE_ACCENT || (c >= 0xa0 && c <= 0xd7ff) ||
5580                     c >= 0xe000) == prop_fail_result)
5581                  break;
5582                eptr += len;
5583              }              }
5584            break;            break;
5585    
# Line 5458  for (;;) Line 5589  for (;;)
5589    
5590          /* eptr is now past the end of the maximum run */          /* eptr is now past the end of the maximum run */
5591    
5592          if (possessive) continue;          if (possessive) continue;    /* No backtracking */
5593          for(;;)          for(;;)
5594            {            {
5595              if (eptr == pp) goto TAIL_RECURSE;
5596            RMATCH(eptr, ecode, offset_top, md, eptrb, RM44);            RMATCH(eptr, ecode, offset_top, md, eptrb, RM44);
5597            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
5598            if (eptr-- == pp) break;        /* Stop if tried at original pos */            eptr--;
5599            if (utf) BACKCHAR(eptr);            if (utf) BACKCHAR(eptr);
5600            }            }
5601          }          }
5602    
5603        /* Match extended Unicode sequences. We will get here only if the        /* Match extended Unicode grapheme clusters. We will get here only if the
5604        support is in the binary; otherwise a compile-time error occurs. */        support is in the binary; otherwise a compile-time error occurs. */
5605    
5606        else if (ctype == OP_EXTUNI)        else if (ctype == OP_EXTUNI)
5607          {          {
5608          for (i = min; i < max; i++)          for (i = min; i < max; i++)
5609            {            {
           int len = 1;  
5610            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
5611              {              {
5612              SCHECK_PARTIAL();              SCHECK_PARTIAL();
5613              break;              break;
5614              }              }
5615            if (!utf) c = *eptr; else { GETCHARLEN(c, eptr, len); }            else
           if (UCD_CATEGORY(c) == ucp_M) break;  
           eptr += len;  
           while (eptr < md->end_subject)  
5616              {              {
5617              len = 1;              int lgb, rgb;
5618              if (!utf) c = *eptr; else { GETCHARLEN(c, eptr, len); }              GETCHARINCTEST(c, eptr);
5619              if (UCD_CATEGORY(c) != ucp_M) break;              lgb = UCD_GRAPHBREAK(c);
5620              eptr += len;              while (eptr < md->end_subject)
5621                  {
5622                  int len = 1;
5623                  if (!utf) c = *eptr; else { GETCHARLEN(c, eptr, len); }
5624                  rgb = UCD_GRAPHBREAK(c);
5625                  if ((PRIV(ucp_gbtable)[lgb] & (1 << rgb)) == 0) break;
5626                  lgb = rgb;
5627                  eptr += len;
5628                  }
5629              }              }
5630              CHECK_PARTIAL();
5631            }            }
5632    
5633          /* eptr is now past the end of the maximum run */          /* eptr is now past the end of the maximum run */
5634    
5635          if (possessive) continue;          if (possessive) continue;    /* No backtracking */
5636    
5637          for(;;)          for(;;)
5638            {            {
5639              int lgb, rgb;
5640              PCRE_PUCHAR fptr;
5641    
5642              if (eptr == pp) goto TAIL_RECURSE;   /* At start of char run */
5643            RMATCH(eptr, ecode, offset_top, md, eptrb, RM45);            RMATCH(eptr, ecode, offset_top, md, eptrb, RM45);
5644            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
5645            if (eptr-- == pp) break;        /* Stop if tried at original pos */  
5646            for (;;)                        /* Move back over one extended */            /* Backtracking over an extended grapheme cluster involves inspecting
5647              the previous two characters (if present) to see if a break is
5648              permitted between them. */
5649    
5650              eptr--;
5651              if (!utf) c = *eptr; else
5652              {              {
5653              if (!utf) c = *eptr; else              BACKCHAR(eptr);
5654                {              GETCHAR(c, eptr);
5655                BACKCHAR(eptr);              }
5656                GETCHAR(c, eptr);            rgb = UCD_GRAPHBREAK(c);
5657                }  
5658              if (UCD_CATEGORY(c) != ucp_M) break;            for (;;)
5659              eptr--;              {
5660                if (eptr == pp) goto TAIL_RECURSE;   /* At start of char run */
5661                fptr = eptr - 1;
5662                if (!utf) c = *fptr; else
5663                  {
5664                  BACKCHAR(fptr);
5665                  GETCHAR(c, fptr);
5666                  }
5667                lgb = UCD_GRAPHBREAK(c);
5668                if ((PRIV(ucp_gbtable)[lgb] & (1 << rgb)) == 0) break;
5669                eptr = fptr;
5670                rgb = lgb;
5671              }              }
5672            }            }
5673          }          }
# Line 5534  for (;;) Line 5691  for (;;)
5691                  break;                  break;
5692                  }                  }
5693                if (IS_NEWLINE(eptr)) break;                if (IS_NEWLINE(eptr)) break;
5694                  if (md->partial != 0 &&    /* Take care with CRLF partial */
5695                      eptr + 1 >= md->end_subject &&
5696                      NLBLOCK->nltype == NLTYPE_FIXED &&
5697                      NLBLOCK->nllen == 2 &&
5698                      *eptr == NLBLOCK->nl[0])
5699                    {
5700                    md->hitend = TRUE;
5701                    if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
5702                    }
5703                eptr++;                eptr++;
5704                ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);                ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);
5705                }                }
# Line 5551  for (;;) Line 5717  for (;;)
5717                  break;                  break;
5718                  }                  }
5719                if (IS_NEWLINE(eptr)) break;                if (IS_NEWLINE(eptr)) break;
5720                  if (md->partial != 0 &&    /* Take care with CRLF partial */
5721                      eptr + 1 >= md->end_subject &&
5722                      NLBLOCK->nltype == NLTYPE_FIXED &&
5723                      NLBLOCK->nllen == 2 &&
5724                      *eptr == NLBLOCK->nl[0])
5725                    {
5726                    md->hitend = TRUE;
5727                    if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
5728                    }
5729                eptr++;                eptr++;
5730                ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);                ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);
5731                }                }
# Line 5600  for (;;) Line 5775  for (;;)
5775                break;                break;
5776                }                }
5777              GETCHARLEN(c, eptr, len);              GETCHARLEN(c, eptr, len);
5778              if (c == 0x000d)              if (c == CHAR_CR)
5779                {                {
5780                if (++eptr >= md->end_subject) break;                if (++eptr >= md->end_subject) break;
5781                if (*eptr == 0x000a) eptr++;                if (*eptr == CHAR_LF) eptr++;
5782                }                }
5783              else              else
5784                {                {
5785                if (c != 0x000a &&                if (c != CHAR_LF &&
5786                    (md->bsr_anycrlf ||                    (md->bsr_anycrlf ||
5787                     (c != 0x000b && c != 0x000c &&                     (c != CHAR_VT && c != CHAR_FF && c != CHAR_NEL
5788                      c != 0x0085 && c != 0x2028 && c != 0x2029)))  #ifndef EBCDIC
5789                        && c != 0x2028 && c != 0x2029
5790    #endif  /* Not EBCDIC */
5791                        )))
5792                  break;                  break;
5793                eptr += len;                eptr += len;
5794                }                }
# Line 5631  for (;;) Line 5809  for (;;)
5809              GETCHARLEN(c, eptr, len);              GETCHARLEN(c, eptr, len);
5810              switch(c)              switch(c)
5811                {                {
5812                  HSPACE_CASES: gotspace = TRUE; break;
5813                default: gotspace = FALSE; break;                default: gotspace = FALSE; break;
               case 0x09:      /* HT */  
               case 0x20:      /* SPACE */  
               case 0xa0:      /* NBSP */  
               case 0x1680:    /* OGHAM SPACE MARK */  
               case 0x180e:    /* MONGOLIAN VOWEL SEPARATOR */  
               case 0x2000:    /* EN QUAD */  
               case 0x2001:    /* EM QUAD */  
               case 0x2002:    /* EN SPACE */  
               case 0x2003:    /* EM SPACE */  
               case 0x2004:    /* THREE-PER-EM SPACE */  
               case 0x2005:    /* FOUR-PER-EM SPACE */  
               case 0x2006:    /* SIX-PER-EM SPACE */  
               case 0x2007:    /* FIGURE SPACE */  
               case 0x2008:    /* PUNCTUATION SPACE */  
               case 0x2009:    /* THIN SPACE */  
               case 0x200A:    /* HAIR SPACE */  
               case 0x202f:    /* NARROW NO-BREAK SPACE */  
               case 0x205f:    /* MEDIUM MATHEMATICAL SPACE */  
               case 0x3000:    /* IDEOGRAPHIC SPACE */  
               gotspace = TRUE;  
               break;  
5814                }                }
5815              if (gotspace == (ctype == OP_NOT_HSPACE)) break;              if (gotspace == (ctype == OP_NOT_HSPACE)) break;
5816              eptr += len;              eptr += len;
# Line 5673  for (;;) Line 5831  for (;;)
5831              GETCHARLEN(c, eptr, len);              GETCHARLEN(c, eptr, len);
5832              switch(c)              switch(c)
5833                {                {
5834                  VSPACE_CASES: gotspace = TRUE; break;
5835                default: gotspace = FALSE; break;                default: gotspace = FALSE; break;
               case 0x0a:      /* LF */  
               case 0x0b:      /* VT */  
               case 0x0c:      /* FF */  
               case 0x0d:      /* CR */  
               case 0x85:      /* NEL */  
               case 0x2028:    /* LINE SEPARATOR */  
               case 0x2029:    /* PARAGRAPH SEPARATOR */  
               gotspace = TRUE;  
               break;  
5836                }                }
5837              if (gotspace == (ctype == OP_NOT_VSPACE)) break;              if (gotspace == (ctype == OP_NOT_VSPACE)) break;
5838              eptr += len;              eptr += len;
# Line 5783  for (;;) Line 5933  for (;;)
5933            RRETURN(PCRE_ERROR_INTERNAL);            RRETURN(PCRE_ERROR_INTERNAL);
5934            }            }
5935    
5936          /* eptr is now past the end of the maximum run. If possessive, we are          if (possessive) continue;    /* No backtracking */
         done (no backing up). Otherwise, match at this position; anything other  
         than no match is immediately returned. For nomatch, back up one  
         character, unless we are matching \R and the last thing matched was  
         \r\n, in which case, back up two bytes. */  
   
         if (possessive) continue;  
5937          for(;;)          for(;;)
5938            {            {
5939              if (eptr == pp) goto TAIL_RECURSE;
5940            RMATCH(eptr, ecode, offset_top, md, eptrb, RM46);            RMATCH(eptr, ecode, offset_top, md, eptrb, RM46);
5941            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
5942            if (eptr-- == pp) break;        /* Stop if tried at original pos */            eptr--;
5943            BACKCHAR(eptr);            BACKCHAR(eptr);
5944            if (ctype == OP_ANYNL && eptr > pp  && *eptr == '\n' &&            if (ctype == OP_ANYNL && eptr > pp  && *eptr == CHAR_NL &&
5945                eptr[-1] == '\r') eptr--;                eptr[-1] == CHAR_CR) eptr--;
5946            }            }
5947          }          }
5948        else        else
# Line 5815  for (;;) Line 5960  for (;;)
5960                break;                break;
5961                }                }
5962              if (IS_NEWLINE(eptr)) break;              if (IS_NEWLINE(eptr)) break;
5963                if (md->partial != 0 &&    /* Take care with CRLF partial */
5964                    eptr + 1 >= md->end_subject &&
5965                    NLBLOCK->nltype == NLTYPE_FIXED &&
5966                    NLBLOCK->nllen == 2 &&
5967                    *eptr == NLBLOCK->nl[0])
5968                  {
5969                  md->hitend = TRUE;
5970                  if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
5971                  }
5972              eptr++;              eptr++;
5973              }              }
5974            break;            break;
# Line 5839  for (;;) Line 5993  for (;;)
5993                break;                break;
5994                }                }
5995              c = *eptr;              c = *eptr;
5996              if (c == 0x000d)              if (c == CHAR_CR)
5997                {                {
5998                if (++eptr >= md->end_subject) break;                if (++eptr >= md->end_subject) break;
5999                if (*eptr == 0x000a) eptr++;                if (*eptr == CHAR_LF) eptr++;
6000                }                }
6001              else              else
6002                {                {
6003                if (c != 0x000a && (md->bsr_anycrlf ||                if (c != CHAR_LF && (md->bsr_anycrlf ||
6004                  (c != 0x000b && c != 0x000c && c != 0x0085                   (c != CHAR_VT && c != CHAR_FF && c != CHAR_NEL
6005  #ifdef COMPILE_PCRE16  #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
6006                  && c != 0x2028 && c != 0x2029                   && c != 0x2028 && c != 0x2029
6007  #endif  #endif
6008                  ))) break;                   ))) break;
6009                eptr++;                eptr++;
6010                }                }
6011              }              }
# Line 5865  for (;;) Line 6019  for (;;)
6019                SCHECK_PARTIAL();                SCHECK_PARTIAL();
6020                break;                break;
6021                }                }
6022              c = *eptr;              switch(*eptr)
6023              if (c == 0x09 || c == 0x20 || c == 0xa0                {
6024  #ifdef COMPILE_PCRE16                default: eptr++; break;
6025                || c == 0x1680 || c == 0x180e || (c >= 0x2000 && c <= 0x200A)                HSPACE_BYTE_CASES:
6026                || c == 0x202f || c == 0x205f || c == 0x3000  #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
6027                  HSPACE_MULTIBYTE_CASES:
6028  #endif  #endif
6029                ) break;                goto ENDLOOP00;
6030              eptr++;                }
6031              }              }
6032              ENDLOOP00:
6033            break;            break;
6034    
6035            case OP_HSPACE:            case OP_HSPACE:
# Line 5884  for (;;) Line 6040  for (;;)
6040                SCHECK_PARTIAL();                SCHECK_PARTIAL();
6041                break;                break;
6042                }                }
6043              c = *eptr;              switch(*eptr)
6044              if (c != 0x09 && c != 0x20 && c != 0xa0                {
6045  #ifdef COMPILE_PCRE16                default: goto ENDLOOP01;
6046                && c != 0x1680 && c != 0x180e && (c < 0x2000 || c > 0x200A)                HSPACE_BYTE_CASES:
6047                && c != 0x202f && c != 0x205f && c != 0x3000  #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
6048                  HSPACE_MULTIBYTE_CASES:
6049  #endif  #endif
6050                ) break;                eptr++; break;
6051              eptr++;                }
6052              }              }
6053              ENDLOOP01:
6054            break;            break;
6055    
6056            case OP_NOT_VSPACE:            case OP_NOT_VSPACE:
# Line 5903  for (;;) Line 6061  for (;;)
6061                SCHECK_PARTIAL();                SCHECK_PARTIAL();
6062                break;                break;
6063                }                }
6064              c = *eptr;              switch(*eptr)
6065              if (c == 0x0a || c == 0x0b || c == 0x0c || c == 0x0d || c == 0x85                {
6066  #ifdef COMPILE_PCRE16                default: eptr++; break;
6067                || c == 0x2028 || c == 0x2029                VSPACE_BYTE_CASES:
6068    #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
6069                  VSPACE_MULTIBYTE_CASES:
6070  #endif  #endif
6071                ) break;                goto ENDLOOP02;
6072              eptr++;                }
6073              }              }
6074              ENDLOOP02:
6075            break;            break;
6076    
6077            case OP_VSPACE:            case OP_VSPACE:
# Line 5921  for (;;) Line 6082  for (;;)
6082                SCHECK_PARTIAL();                SCHECK_PARTIAL();
6083                break;                break;
6084                }                }
6085              c = *eptr;              switch(*eptr)
6086              if (c != 0x0a && c != 0x0b && c != 0x0c && c != 0x0d && c != 0x85                {
6087  #ifdef COMPILE_PCRE16                default: goto ENDLOOP03;
6088                && c != 0x2028 && c != 0x2029                VSPACE_BYTE_CASES:
6089    #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
6090                  VSPACE_MULTIBYTE_CASES:
6091  #endif  #endif
6092                ) break;                eptr++; break;
6093              eptr++;                }
6094              }              }
6095              ENDLOOP03:
6096            break;            break;
6097    
6098            case OP_NOT_DIGIT:            case OP_NOT_DIGIT:
# Line 6013  for (;;) Line 6177  for (;;)
6177            RRETURN(PCRE_ERROR_INTERNAL);            RRETURN(PCRE_ERROR_INTERNAL);
6178            }            }
6179    
6180          /* eptr is now past the end of the maximum run. If possessive, we are          if (possessive) continue;    /* No backtracking */
6181          done (no backing up). Otherwise, match at this position; anything other          for (;;)
         than no match is immediately returned. For nomatch, back up one  
         character (byte), unless we are matching \R and the last thing matched  
         was \r\n, in which case, back up two bytes. */  
   
         if (possessive) continue;  
         while (eptr >= pp)  
6182            {            {
6183              if (eptr == pp) goto TAIL_RECURSE;
6184            RMATCH(eptr, ecode, offset_top, md, eptrb, RM47);            RMATCH(eptr, ecode, offset_top, md, eptrb, RM47);
6185            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
6186            eptr--;            eptr--;
6187            if (ctype == OP_ANYNL && eptr > pp  && *eptr == '\n' &&            if (ctype == OP_ANYNL && eptr > pp  && *eptr == CHAR_LF &&
6188                eptr[-1] == '\r') eptr--;                eptr[-1] == CHAR_CR) eptr--;
6189            }            }
6190          }          }
6191    
6192        /* Get here if we can't make it match with any permitted repetitions */        /* Control never gets here */
   
       RRETURN(MATCH_NOMATCH);  
6193        }        }
     /* Control never gets here */  
6194    
6195      /* There's been some horrible disaster. Arrival here can only mean there is      /* There's been some horrible disaster. Arrival here can only mean there is
6196      something seriously wrong in the code above or the OP_xxx definitions. */      something seriously wrong in the code above or the OP_xxx definitions. */
# Line 6068  switch (frame->Xwhere) Line 6224  switch (frame->Xwhere)
6224    LBL(53) LBL(54) LBL(55) LBL(56) LBL(57) LBL(58) LBL(63) LBL(64)    LBL(53) LBL(54) LBL(55) LBL(56) LBL(57) LBL(58) LBL(63) LBL(64)
6225    LBL(65) LBL(66)    LBL(65) LBL(66)
6226  #if defined SUPPORT_UTF || !defined COMPILE_PCRE8  #if defined SUPPORT_UTF || !defined COMPILE_PCRE8
6227    LBL(21)    LBL(20) LBL(21)
6228  #endif  #endif
6229  #ifdef SUPPORT_UTF  #ifdef SUPPORT_UTF
6230    LBL(16) LBL(18) LBL(20)    LBL(16) LBL(18)
6231    LBL(22) LBL(23) LBL(28) LBL(30)    LBL(22) LBL(23) LBL(28) LBL(30)
6232    LBL(32) LBL(34) LBL(42) LBL(46)    LBL(32) LBL(34) LBL(42) LBL(46)
6233  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
6234    LBL(36) LBL(37) LBL(38) LBL(39) LBL(40) LBL(41) LBL(44) LBL(45)    LBL(36) LBL(37) LBL(38) LBL(39) LBL(40) LBL(41) LBL(44) LBL(45)
6235    LBL(59) LBL(60) LBL(61) LBL(62)    LBL(59) LBL(60) LBL(61) LBL(62) LBL(67)
6236  #endif  /* SUPPORT_UCP */  #endif  /* SUPPORT_UCP */
6237  #endif  /* SUPPORT_UTF */  #endif  /* SUPPORT_UTF */
6238    default:    default:
6239    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));
   
 printf("+++jump error in pcre match: label %d non-existent\n", frame->Xwhere);  
   
6240    return PCRE_ERROR_INTERNAL;    return PCRE_ERROR_INTERNAL;
6241    }    }
6242  #undef LBL  #undef LBL
# Line 6145  Undefine all the macros that were define Line 6298  Undefine all the macros that were define
6298  ***************************************************************************/  ***************************************************************************/
6299    
6300    
6301    #ifdef NO_RECURSE
6302    /*************************************************
6303    *          Release allocated heap frames         *
6304    *************************************************/
6305    
6306    /* This function releases all the allocated frames. The base frame is on the
6307    machine stack, and so must not be freed.
6308    
6309    Argument: the address of the base frame
6310    Returns:  nothing
6311    */
6312    
6313    static void
6314    release_match_heapframes (heapframe *frame_base)
6315    {
6316    heapframe *nextframe = frame_base->Xnextframe;
6317    while (nextframe != NULL)
6318      {
6319      heapframe *oldframe = nextframe;
6320      nextframe = nextframe->Xnextframe;
6321      (PUBL(stack_free))(oldframe);
6322      }
6323    }
6324    #endif
6325    
6326    
6327  /*************************************************  /*************************************************
6328  *         Execute a Regular Expression           *  *         Execute a Regular Expression           *
# Line 6170  Returns:          > 0 => success; value Line 6348  Returns:          > 0 => success; value
6348                   < -1 => some kind of unexpected problem                   < -1 => some kind of unexpected problem
6349  */  */
6350    
6351  #ifdef COMPILE_PCRE8  #if defined COMPILE_PCRE8
6352  PCRE_EXP_DEFN int PCRE_CALL_CONVENTION  PCRE_EXP_DEFN int PCRE_CALL_CONVENTION
6353  pcre_exec(const pcre *argument_re, const pcre_extra *extra_data,  pcre_exec(const pcre *argument_re, const pcre_extra *extra_data,
6354    PCRE_SPTR subject, int length, int start_offset, int options, int *offsets,    PCRE_SPTR subject, int length, int start_offset, int options, int *offsets,
6355    int offsetcount)    int offsetcount)
6356  #else  #elif defined COMPILE_PCRE16
6357  PCRE_EXP_DEFN int PCRE_CALL_CONVENTION  PCRE_EXP_DEFN int PCRE_CALL_CONVENTION
6358  pcre16_exec(const pcre16 *argument_re, const pcre16_extra *extra_data,  pcre16_exec(const pcre16 *argument_re, const pcre16_extra *extra_data,
6359    PCRE_SPTR16 subject, int length, int start_offset, int options, int *offsets,    PCRE_SPTR16 subject, int length, int start_offset, int options, int *offsets,
6360    int offsetcount)    int offsetcount)
6361    #elif defined COMPILE_PCRE32
6362    PCRE_EXP_DEFN int PCRE_CALL_CONVENTION
6363    pcre32_exec(const pcre32 *argument_re, const pcre32_extra *extra_data,
6364      PCRE_SPTR32 subject, int length, int start_offset, int options, int *offsets,
6365      int offsetcount)
6366  #endif  #endif
6367  {  {
6368  int rc, ocount, arg_offset_max;  int rc, ocount, arg_offset_max;
# Line 6202  const pcre_uint8 *start_bits = NULL; Line 6385  const pcre_uint8 *start_bits = NULL;
6385  PCRE_PUCHAR start_match = (PCRE_PUCHAR)subject + start_offset;  PCRE_PUCHAR start_match = (PCRE_PUCHAR)subject + start_offset;
6386  PCRE_PUCHAR end_subject;  PCRE_PUCHAR end_subject;
6387  PCRE_PUCHAR start_partial = NULL;  PCRE_PUCHAR start_partial = NULL;
6388    PCRE_PUCHAR match_partial = NULL;
6389  PCRE_PUCHAR req_char_ptr = start_match - 1;  PCRE_PUCHAR req_char_ptr = start_match - 1;
6390    
6391  const pcre_study_data *study;  const pcre_study_data *study;
6392  const REAL_PCRE *re = (const REAL_PCRE *)argument_re;  const REAL_PCRE *re = (const REAL_PCRE *)argument_re;
6393    
6394  /* Check for the special magic call that measures the size of the stack used  #ifdef NO_RECURSE
6395  per recursive call of match(). */  heapframe frame_zero;
6396    frame_zero.Xprevframe = NULL;            /* Marks the top level */
6397    frame_zero.Xnextframe = NULL;            /* None are allocated yet */
6398    md->match_frames_base = &frame_zero;
6399    #endif
6400    
6401    /* Check for the special magic call that measures the size of the stack used
6402    per recursive call of match(). Without the funny casting for sizeof, a Windows
6403    compiler gave this error: "unary minus operator applied to unsigned type,
6404    result still unsigned". Hopefully the cast fixes that. */
6405    
6406  if (re == NULL && extra_data == NULL && subject == NULL && length == -1)  if (re == NULL && extra_data == NULL && subject == NULL && length == -999 &&
6407        start_offset == -999)
6408  #ifdef NO_RECURSE  #ifdef NO_RECURSE
6409    return -sizeof(heapframe);    return -((int)sizeof(heapframe));
6410  #else  #else
6411    return match(NULL, NULL, NULL, 0, NULL, NULL, 0);    return match(NULL, NULL, NULL, 0, NULL, NULL, 0);
6412  #endif  #endif
6413    
6414  /* Plausibility checks */  /* Plausibility checks */
6415    
6416  if ((options & ~PUBLIC_EXEC_OPTIONS) != 0) return PCRE_ERROR_BADOPTION;  if ((options & ~PUBLIC_EXEC_OPTIONS) != 0) return PCRE_ERROR_BADOPTION;
6417  if (re == NULL || subject == NULL || (offsets == NULL && offsetcount > 0))  if (re == NULL || subject == NULL || (offsets == NULL && offsetcount > 0))
6418    return PCRE_ERROR_NULL;    return PCRE_ERROR_NULL;
6419  if (offsetcount < 0) return PCRE_ERROR_BADCOUNT;  if (offsetcount < 0) return PCRE_ERROR_BADCOUNT;
6420    if (length < 0) return PCRE_ERROR_BADLENGTH;
6421  if (start_offset < 0 || start_offset > length) return PCRE_ERROR_BADOFFSET;  if (start_offset < 0 || start_offset > length) return PCRE_ERROR_BADOFFSET;
6422    
6423  /* Check that the first field in the block is the magic number. If it is not,  /* Check that the first field in the block is the magic number. If it is not,
# Line 6260  if (utf && (options & PCRE_NO_UTF8_CHECK Line 6455  if (utf && (options & PCRE_NO_UTF8_CHECK
6455        offsets[0] = erroroffset;        offsets[0] = erroroffset;
6456        offsets[1] = errorcode;        offsets[1] = errorcode;
6457        }        }
6458  #ifdef COMPILE_PCRE16  #if defined COMPILE_PCRE8
     return (errorcode <= PCRE_UTF16_ERR1 && md->partial > 1)?  
       PCRE_ERROR_SHORTUTF16 : PCRE_ERROR_BADUTF16;  
 #else  
6459      return (errorcode <= PCRE_UTF8_ERR5 && md->partial > 1)?      return (errorcode <= PCRE_UTF8_ERR5 && md->partial > 1)?
6460        PCRE_ERROR_SHORTUTF8 : PCRE_ERROR_BADUTF8;        PCRE_ERROR_SHORTUTF8 : PCRE_ERROR_BADUTF8;
6461    #elif defined COMPILE_PCRE16
6462        return (errorcode <= PCRE_UTF16_ERR1 && md->partial > 1)?
6463          PCRE_ERROR_SHORTUTF16 : PCRE_ERROR_BADUTF16;
6464    #elif defined COMPILE_PCRE32
6465        return PCRE_ERROR_BADUTF32;
6466  #endif  #endif
6467      }      }
6468    #if defined COMPILE_PCRE8 || defined COMPILE_PCRE16
6469    /* Check that a start_offset points to the start of a UTF character. */    /* Check that a start_offset points to the start of a UTF character. */
6470    if (start_offset > 0 && start_offset < length &&    if (start_offset > 0 && start_offset < length &&
6471        NOT_FIRSTCHAR(((PCRE_PUCHAR)subject)[start_offset]))        NOT_FIRSTCHAR(((PCRE_PUCHAR)subject)[start_offset]))
6472      return PCRE_ERROR_BADUTF8_OFFSET;      return PCRE_ERROR_BADUTF8_OFFSET;
6473    #endif
6474    }    }
6475  #endif  #endif
6476    
6477  /* If the pattern was successfully studied with JIT support, run the JIT  /* If the pattern was successfully studied with JIT support, run the JIT
6478  executable instead of the rest of this function. Most options must be set at  executable instead of the rest of this function. Most options must be set at
6479  compile time for the JIT code to be usable. Fallback to the normal code path if  compile time for the JIT code to be usable. Fallback to the normal code path if
6480  an unsupported flag is set. In particular, JIT does not support partial  an unsupported flag is set. */
 matching. */  
6481    
6482  #ifdef SUPPORT_JIT  #ifdef SUPPORT_JIT
6483  if (extra_data != NULL  if (extra_data != NULL
6484      && (extra_data->flags & PCRE_EXTRA_EXECUTABLE_JIT) != 0      && (extra_data->flags & (PCRE_EXTRA_EXECUTABLE_JIT |
6485                                 PCRE_EXTRA_TABLES)) == PCRE_EXTRA_EXECUTABLE_JIT
6486      && extra_data->executable_jit != NULL      && extra_data->executable_jit != NULL
6487      && (extra_data->flags & PCRE_EXTRA_TABLES) == 0      && (options & ~PUBLIC_JIT_EXEC_OPTIONS) == 0)
6488      && (options & ~(PCRE_NO_UTF8_CHECK | PCRE_NOTBOL | PCRE_NOTEOL |    {
6489                      PCRE_NOTEMPTY | PCRE_NOTEMPTY_ATSTART)) == 0)    rc = PRIV(jit_exec)(extra_data, (const pcre_uchar *)subject, length,
6490    return PRIV(jit_exec)(re, extra_data->executable_jit,         start_offset, options, offsets, offsetcount);
6491      (const pcre_uchar *)subject, length, start_offset, options,  
6492      ((extra_data->flags & PCRE_EXTRA_MATCH_LIMIT) == 0)    /* PCRE_ERROR_NULL means that the selected normal or partial matching
6493      ? MATCH_LIMIT : extra_data->match_limit, offsets, offsetcount);    mode is not compiled. In this case we simply fallback to interpreter. */
6494    
6495      if (rc != PCRE_ERROR_JIT_BADOPTION) return rc;
6496      }
6497  #endif  #endif
6498    
6499  /* Carry on with non-JIT matching. This information is for finding all the  /* Carry on with non-JIT matching. This information is for finding all the
# Line 6314  md->callout_data = NULL; Line 6515  md->callout_data = NULL;
6515    
6516  tables = re->tables;  tables = re->tables;
6517    
6518    /* The two limit values override the defaults, whatever their value. */
6519    
6520  if (extra_data != NULL)  if (extra_data != NULL)
6521    {    {
6522    register unsigned int flags = extra_data->flags;    register unsigned int flags = extra_data->flags;
# Line 6328  if (extra_data != NULL) Line 6531  if (extra_data != NULL)
6531    if ((flags & PCRE_EXTRA_TABLES) != 0) tables = extra_data->tables;    if ((flags & PCRE_EXTRA_TABLES) != 0) tables = extra_data->tables;
6532    }    }
6533    
6534    /* Limits in the regex override only if they are smaller. */
6535    
6536    if ((re->flags & PCRE_MLSET) != 0 && re->limit_match < md->match_limit)
6537      md->match_limit = re->limit_match;
6538    
6539    if ((re->flags & PCRE_RLSET) != 0 &&
6540        re->limit_recursion < md->match_limit_recursion)
6541      md->match_limit_recursion = re->limit_recursion;
6542    
6543  /* If the exec call supplied NULL for tables, use the inbuilt ones. This  /* If the exec call supplied NULL for tables, use the inbuilt ones. This
6544  is a feature that makes it possible to save compiled regex and re-use them  is a feature that makes it possible to save compiled regex and re-use them
6545  in other programs later. */  in other programs later. */
# Line 6353  end_subject = md->end_subject; Line 6565  end_subject = md->end_subject;
6565  md->endonly = (re->options & PCRE_DOLLAR_ENDONLY) != 0;  md->endonly = (re->options & PCRE_DOLLAR_ENDONLY) != 0;
6566  md->use_ucp = (re->options & PCRE_UCP) != 0;  md->use_ucp = (re->options & PCRE_UCP) != 0;
6567  md->jscript_compat = (re->options & PCRE_JAVASCRIPT_COMPAT) != 0;  md->jscript_compat = (re->options & PCRE_JAVASCRIPT_COMPAT) != 0;
6568  md->ignore_skip_arg = FALSE;  md->ignore_skip_arg = 0;
6569    
6570  /* Some options are unpacked into BOOL variables in the hope that testing  /* Some options are unpacked into BOOL variables in the hope that testing
6571  them will be faster than individual option bits. */  them will be faster than individual option bits. */
# Line 6463  if (re->top_backref > 0 && re->top_backr Line 6675  if (re->top_backref > 0 && re->top_backr
6675    DPRINTF(("Got memory to hold back references\n"));    DPRINTF(("Got memory to hold back references\n"));
6676    }    }
6677  else md->offset_vector = offsets;  else md->offset_vector = offsets;
   
6678  md->offset_end = ocount;  md->offset_end = ocount;
6679  md->offset_max = (2*ocount)/3;  md->offset_max = (2*ocount)/3;
6680  md->offset_overflow = FALSE;  md->capture_last = 0;
 md->capture_last = -1;  
6681    
6682  /* Reset the working variable associated with each extraction. These should  /* Reset the working variable associated with each extraction. These should
6683  never be used unless previously set, but they get saved and restored, and so we  never be used unless previously set, but they get saved and restored, and so we
# Line 6495  if (!anchored) Line 6705  if (!anchored)
6705    if ((re->flags & PCRE_FIRSTSET) != 0)    if ((re->flags & PCRE_FIRSTSET) != 0)
6706      {      {
6707      has_first_char = TRUE;      has_first_char = TRUE;
6708      first_char = first_char2 = re->first_char;      first_char = first_char2 = (pcre_uchar)(re->first_char);
6709      if ((re->flags & PCRE_FCH_CASELESS) != 0)      if ((re->flags & PCRE_FCH_CASELESS) != 0)
6710        {        {
6711        first_char2 = TABLE_GET(first_char, md->fcc, first_char);        first_char2 = TABLE_GET(first_char, md->fcc, first_char);
# Line 6517  character" set. */ Line 6727  character" set. */
6727  if ((re->flags & PCRE_REQCHSET) != 0)  if ((re->flags & PCRE_REQCHSET) != 0)
6728    {    {
6729    has_req_char = TRUE;    has_req_char = TRUE;
6730    req_char = req_char2 = re->req_char;    req_char = req_char2 = (pcre_uchar)(re->req_char);
6731    if ((re->flags & PCRE_RCH_CASELESS) != 0)    if ((re->flags & PCRE_RCH_CASELESS) != 0)
6732      {      {
6733      req_char2 = TABLE_GET(req_char, md->fcc, req_char);      req_char2 = TABLE_GET(req_char, md->fcc, req_char);
# Line 6575  for(;;) Line 6785  for(;;)
6785    
6786      if (has_first_char)      if (has_first_char)
6787        {        {
6788          pcre_uchar smc;
6789    
6790        if (first_char != first_char2)        if (first_char != first_char2)
6791          while (start_match < end_subject &&          while (start_match < end_subject &&
6792              *start_match != first_char && *start_match != first_char2)            (smc = *start_match) != first_char && smc != first_char2)
6793            start_match++;            start_match++;
6794        else        else
6795          while (start_match < end_subject && *start_match != first_char)          while (start_match < end_subject && *start_match != first_char)
# Line 6623  for(;;) Line 6835  for(;;)
6835        {        {
6836        while (start_match < end_subject)        while (start_match < end_subject)
6837          {          {
6838          register unsigned int c = *start_match;          register pcre_uint32 c = *start_match;
6839  #ifndef COMPILE_PCRE8  #ifndef COMPILE_PCRE8
6840          if (c > 255) c = 255;          if (c > 255) c = 255;
6841  #endif  #endif
6842          if ((start_bits[c/8] & (1 << (c&7))) == 0)          if ((start_bits[c/8] & (1 << (c&7))) != 0) break;
6843            {          start_match++;
           start_match++;  
 #if defined SUPPORT_UTF && defined COMPILE_PCRE8  
           /* In non 8-bit mode, the iteration will stop for  
           characters > 255 at the beginning or not stop at all. */  
           if (utf)  
             ACROSSCHAR(start_match < end_subject, *start_match,  
               start_match++);  
 #endif  
           }  
         else break;  
6844          }          }
6845        }        }
6846      }   /* Starting optimizations */      }   /* Starting optimizations */
# Line 6691  for(;;) Line 6893  for(;;)
6893            {            {
6894            while (p < end_subject)            while (p < end_subject)
6895              {              {
6896              register int pp = *p++;              register pcre_uint32 pp = *p++;
6897              if (pp == req_char || pp == req_char2) { p--; break; }              if (pp == req_char || pp == req_char2) { p--; break; }
6898              }              }
6899            }            }
# Line 6735  for(;;) Line 6937  for(;;)
6937    md->match_call_count = 0;    md->match_call_count = 0;
6938    md->match_function_type = 0;    md->match_function_type = 0;
6939    md->end_offset_top = 0;    md->end_offset_top = 0;
6940      md->skip_arg_count = 0;
6941    rc = match(start_match, md->start_code, start_match, 2, md, NULL, 0);    rc = match(start_match, md->start_code, start_match, 2, md, NULL, 0);
6942    if (md->hitend && start_partial == NULL) start_partial = md->start_used_ptr;    if (md->hitend && start_partial == NULL)
6943        {
6944        start_partial = md->start_used_ptr;
6945        match_partial = start_match;
6946        }
6947    
6948    switch(rc)    switch(rc)
6949      {      {
# Line 6749  for(;;) Line 6956  for(;;)
6956    
6957      case MATCH_SKIP_ARG:      case MATCH_SKIP_ARG:
6958      new_start_match = start_match;      new_start_match = start_match;
6959      md->ignore_skip_arg = TRUE;      md->ignore_skip_arg = md->skip_arg_count;
6960      break;      break;
6961    
6962      /* SKIP passes back the next starting point explicitly, but if it is the      /* SKIP passes back the next starting point explicitly, but if it is no
6963      same as the match we have just done, treat it as NOMATCH. */      greater than the match we have just done, treat it as NOMATCH. */
6964    
6965      case MATCH_SKIP:      case MATCH_SKIP:
6966      if (md->start_match_ptr != start_match)      if (md->start_match_ptr > start_match)
6967        {        {
6968        new_start_match = md->start_match_ptr;        new_start_match = md->start_match_ptr;
6969        break;        break;
# Line 6764  for(;;) Line 6971  for(;;)
6971      /* Fall through */      /* Fall through */
6972    
6973      /* NOMATCH and PRUNE advance by one character. THEN at this level acts      /* NOMATCH and PRUNE advance by one character. THEN at this level acts
6974      exactly like PRUNE. Unset the ignore SKIP-with-argument flag. */      exactly like PRUNE. Unset ignore SKIP-with-argument. */
6975    
6976      case MATCH_NOMATCH:      case MATCH_NOMATCH:
6977      case MATCH_PRUNE:      case MATCH_PRUNE:
6978      case MATCH_THEN:      case MATCH_THEN:
6979      md->ignore_skip_arg = FALSE;      md->ignore_skip_arg = 0;
6980      new_start_match = start_match + 1;      new_start_match = start_match + 1;
6981  #ifdef SUPPORT_UTF  #ifdef SUPPORT_UTF
6982      if (utf)      if (utf)