/[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 1361 by ph10, Fri Sep 6 17:47:32 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 93  because the offset vector is always a mu Line 110  because the offset vector is always a mu
110  static const char rep_min[] = { 0, 0, 1, 1, 0, 0 };  static const char rep_min[] = { 0, 0, 1, 1, 0, 0 };
111  static const char rep_max[] = { 0, 0, 0, 0, 1, 1 };  static const char rep_max[] = { 0, 0, 0, 0, 1, 1 };
112    
   
   
113  #ifdef PCRE_DEBUG  #ifdef PCRE_DEBUG
114  /*************************************************  /*************************************************
115  *        Debugging function to print chars       *  *        Debugging function to print chars       *
# 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 = RAWUCHARINCTEST(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    #ifdef SUPPORT_UTF
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 175  if (caseless) Line 197  if (caseless)
197    {    {
198  #ifdef SUPPORT_UTF  #ifdef SUPPORT_UTF
199  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
200    if (md->utf)    if (utf)
201      {      {
202      /* 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
203      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
204      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.
205      (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
206      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
207      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
208      reference, not along the subject (earlier code did this wrong). */      length along the reference, not along the subject (earlier code did this
209        wrong). */
210    
211      PCRE_PUCHAR endptr = p + length;      PCRE_PUCHAR endptr = p + length;
212      while (p < endptr)      while (p < endptr)
213        {        {
214        int c, d;        pcre_uint32 c, d;
215        if (eptr >= md->end_subject) return -1;        const ucd_record *ur;
216          if (eptr >= md->end_subject) return -2;   /* Partial match */
217        GETCHARINC(c, eptr);        GETCHARINC(c, eptr);
218        GETCHARINC(d, p);        GETCHARINC(d, p);
219        if (c != d && c != UCD_OTHERCASE(d)) return -1;        ur = GET_UCD(d);
220          if (c != d && c != d + ur->other_case)
221            {
222            const pcre_uint32 *pp = PRIV(ucd_caseless_sets) + ur->caseset;
223            for (;;)
224              {
225              if (c < *pp) return -1;
226              if (c == *pp++) break;
227              }
228            }
229        }        }
230      }      }
231    else    else
# Line 202  if (caseless) Line 235  if (caseless)
235    /* 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
236    is no UCP support. */    is no UCP support. */
237      {      {
     if (eptr + length > md->end_subject) return -1;  
238      while (length-- > 0)      while (length-- > 0)
239        {        {
240        if (TABLE_GET(*p, md->lcc, *p) != TABLE_GET(*eptr, md->lcc, *eptr)) return -1;        pcre_uint32 cc, cp;
241          if (eptr >= md->end_subject) return -2;   /* Partial match */
242          cc = RAWUCHARTEST(eptr);
243          cp = RAWUCHARTEST(p);
244          if (TABLE_GET(cp, md->lcc, cp) != TABLE_GET(cc, md->lcc, cc)) return -1;
245        p++;        p++;
246        eptr++;        eptr++;
247        }        }
# Line 217  are in UTF-8 mode. */ Line 253  are in UTF-8 mode. */
253    
254  else  else
255    {    {
256    if (eptr + length > md->end_subject) return -1;    while (length-- > 0)
257    while (length-- > 0) if (*p++ != *eptr++) return -1;      {
258        if (eptr >= md->end_subject) return -2;   /* Partial match */
259        if (RAWUCHARINCTEST(p) != RAWUCHARINCTEST(eptr)) return -1;
260        }
261    }    }
262    
263  return (int)(eptr - eptr_start);  return (int)(eptr - eptr_start);
# Line 273  enum { RM1=1, RM2,  RM3,  RM4,  RM5,  RM Line 312  enum { RM1=1, RM2,  RM3,  RM4,  RM5,  RM
312         RM31,  RM32, RM33, RM34, RM35, RM36, RM37, RM38, RM39, RM40,         RM31,  RM32, RM33, RM34, RM35, RM36, RM37, RM38, RM39, RM40,
313         RM41,  RM42, RM43, RM44, RM45, RM46, RM47, RM48, RM49, RM50,         RM41,  RM42, RM43, RM44, RM45, RM46, RM47, RM48, RM49, RM50,
314         RM51,  RM52, RM53, RM54, RM55, RM56, RM57, RM58, RM59, RM60,         RM51,  RM52, RM53, RM54, RM55, RM56, RM57, RM58, RM59, RM60,
315         RM61,  RM62, RM63, RM64, RM65, RM66 };         RM61,  RM62, RM63, RM64, RM65, RM66, RM67, RM68 };
316    
317  /* 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
318  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 330  actually used in this definition. */
330    }    }
331  #define RRETURN(ra) \  #define RRETURN(ra) \
332    { \    { \
333    printf("match() returned %d from line %d ", ra, __LINE__); \    printf("match() returned %d from line %d\n", ra, __LINE__); \
334    return ra; \    return ra; \
335    }    }
336  #else  #else
# Line 311  argument of match(), which never changes Line 350  argument of match(), which never changes
350    
351  #define RMATCH(ra,rb,rc,rd,re,rw)\  #define RMATCH(ra,rb,rc,rd,re,rw)\
352    {\    {\
353    heapframe *newframe = (heapframe *)(PUBL(stack_malloc))(sizeof(heapframe));\    heapframe *newframe = frame->Xnextframe;\
354    if (newframe == NULL) RRETURN(PCRE_ERROR_NOMEMORY);\    if (newframe == NULL)\
355    frame->Xwhere = rw; \      {\
356        newframe = (heapframe *)(PUBL(stack_malloc))(sizeof(heapframe));\
357        if (newframe == NULL) RRETURN(PCRE_ERROR_NOMEMORY);\
358        newframe->Xnextframe = NULL;\
359        frame->Xnextframe = newframe;\
360        }\
361      frame->Xwhere = rw;\
362    newframe->Xeptr = ra;\    newframe->Xeptr = ra;\
363    newframe->Xecode = rb;\    newframe->Xecode = rb;\
364    newframe->Xmstart = mstart;\    newframe->Xmstart = mstart;\
# Line 332  argument of match(), which never changes Line 377  argument of match(), which never changes
377    {\    {\
378    heapframe *oldframe = frame;\    heapframe *oldframe = frame;\
379    frame = oldframe->Xprevframe;\    frame = oldframe->Xprevframe;\
   if (oldframe != &frame_zero) (PUBL(stack_free))(oldframe);\  
380    if (frame != NULL)\    if (frame != NULL)\
381      {\      {\
382      rrc = ra;\      rrc = ra;\
# Line 346  argument of match(), which never changes Line 390  argument of match(), which never changes
390    
391  typedef struct heapframe {  typedef struct heapframe {
392    struct heapframe *Xprevframe;    struct heapframe *Xprevframe;
393      struct heapframe *Xnextframe;
394    
395    /* Function arguments that may change */    /* Function arguments that may change */
396    
# Line 376  typedef struct heapframe { Line 421  typedef struct heapframe {
421    
422  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
423    int Xprop_type;    int Xprop_type;
424    int Xprop_value;    unsigned int Xprop_value;
425    int Xprop_fail_result;    int Xprop_fail_result;
426    int Xoclength;    int Xoclength;
427    pcre_uchar Xocchars[6];    pcre_uchar Xocchars[6];
# Line 389  typedef struct heapframe { Line 434  typedef struct heapframe {
434    int Xlength;    int Xlength;
435    int Xmax;    int Xmax;
436    int Xmin;    int Xmin;
437    int Xnumber;    unsigned int Xnumber;
438    int Xoffset;    int Xoffset;
439    int Xop;    unsigned int Xop;
440    int Xsave_capture_last;    pcre_int32 Xsave_capture_last;
441    int Xsave_offset1, Xsave_offset2, Xsave_offset3;    int Xsave_offset1, Xsave_offset2, Xsave_offset3;
442    int Xstacksave[REC_STACK_SAVE_MAX];    int Xstacksave[REC_STACK_SAVE_MAX];
443    
# Line 477  so they can be ordinary variables in all Line 522  so they can be ordinary variables in all
522    
523  register int  rrc;         /* Returns from recursive calls */  register int  rrc;         /* Returns from recursive calls */
524  register int  i;           /* Used for loops not involving calls to RMATCH() */  register int  i;           /* Used for loops not involving calls to RMATCH() */
525  register unsigned int c;   /* Character values not kept over RMATCH() calls */  register pcre_uint32 c;    /* Character values not kept over RMATCH() calls */
526  register BOOL utf;         /* Local copy of UTF flag for speed */  register BOOL utf;         /* Local copy of UTF flag for speed */
527    
528  BOOL minimize, possessive; /* Quantifier options */  BOOL minimize, possessive; /* Quantifier options */
# Line 487  int condcode; Line 532  int condcode;
532  /* 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
533  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
534  frame on the stack here; subsequent instantiations are obtained from the heap  frame on the stack here; subsequent instantiations are obtained from the heap
535  whenever RMATCH() does a "recursion". See the macro definitions above. Putting  whenever RMATCH() does a "recursion". See the macro definitions above. Putting
536  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
537  boost in many cases where there is not much "recursion". */  boost in many cases where there is not much "recursion". */
538    
539  #ifdef NO_RECURSE  #ifdef NO_RECURSE
540  heapframe frame_zero;  heapframe *frame = (heapframe *)md->match_frames_base;
 heapframe *frame = &frame_zero;  
 frame->Xprevframe = NULL;            /* Marks the top level */  
541    
542  /* Copy in the original argument variables */  /* Copy in the original argument variables */
543    
# Line 596  BOOL prev_is_word; Line 639  BOOL prev_is_word;
639    
640  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
641  int prop_type;  int prop_type;
642  int prop_value;  unsigned int prop_value;
643  int prop_fail_result;  int prop_fail_result;
644  int oclength;  int oclength;
645  pcre_uchar occhars[6];  pcre_uchar occhars[6];
# Line 607  int ctype; Line 650  int ctype;
650  int length;  int length;
651  int max;  int max;
652  int min;  int min;
653  int number;  unsigned int number;
654  int offset;  int offset;
655  int op;  unsigned int op;
656  int save_capture_last;  pcre_int32 save_capture_last;
657  int save_offset1, save_offset2, save_offset3;  int save_offset1, save_offset2, save_offset3;
658  int stacksave[REC_STACK_SAVE_MAX];  int stacksave[REC_STACK_SAVE_MAX];
659    
660  eptrblock newptrb;  eptrblock newptrb;
661    
662  /* 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
663  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
664  recursion. The second argument (ecode) being NULL triggers this behaviour. It  recursion. The second argument (ecode) being NULL triggers this behaviour. It
665  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
666  size. */  size. */
667    
668  if (ecode == NULL)  if (ecode == NULL)
# Line 631  if (ecode == NULL) Line 674  if (ecode == NULL)
674      int len = (char *)&rdepth - (char *)eptr;      int len = (char *)&rdepth - (char *)eptr;
675      return (len > 0)? -len : len;      return (len > 0)? -len : len;
676      }      }
677    }    }
678  #endif     /* NO_RECURSE */  #endif     /* NO_RECURSE */
679    
680  /* 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 771  for (;;)
771      unaltered. */      unaltered. */
772    
773      else if (rrc == MATCH_SKIP_ARG &&      else if (rrc == MATCH_SKIP_ARG &&
774          STRCMP_UC_UC(ecode + 2, md->start_match_ptr) == 0)          STRCMP_UC_UC_TEST(ecode + 2, md->start_match_ptr) == 0)
775        {        {
776        md->start_match_ptr = eptr;        md->start_match_ptr = eptr;
777        RRETURN(MATCH_SKIP);        RRETURN(MATCH_SKIP);
# Line 738  for (;;) Line 781  for (;;)
781      case OP_FAIL:      case OP_FAIL:
782      RRETURN(MATCH_NOMATCH);      RRETURN(MATCH_NOMATCH);
783    
     /* COMMIT overrides PRUNE, SKIP, and THEN */  
   
784      case OP_COMMIT:      case OP_COMMIT:
785      RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md,      RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md,
786        eptrb, RM52);        eptrb, RM52);
787      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);  
788      RRETURN(MATCH_COMMIT);      RRETURN(MATCH_COMMIT);
789    
     /* PRUNE overrides THEN */  
   
790      case OP_PRUNE:      case OP_PRUNE:
791      RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md,      RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md,
792        eptrb, RM51);        eptrb, RM51);
793      if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN) RRETURN(rrc);      if (rrc != MATCH_NOMATCH) RRETURN(rrc);
794      RRETURN(MATCH_PRUNE);      RRETURN(MATCH_PRUNE);
795    
796      case OP_PRUNE_ARG:      case OP_PRUNE_ARG:
# Line 764  for (;;) Line 800  for (;;)
800        eptrb, RM56);        eptrb, RM56);
801      if ((rrc == MATCH_MATCH || rrc == MATCH_ACCEPT) &&      if ((rrc == MATCH_MATCH || rrc == MATCH_ACCEPT) &&
802           md->mark == NULL) md->mark = ecode + 2;           md->mark == NULL) md->mark = ecode + 2;
803      if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN) RRETURN(rrc);      if (rrc != MATCH_NOMATCH) RRETURN(rrc);
804      RRETURN(MATCH_PRUNE);      RRETURN(MATCH_PRUNE);
805    
     /* SKIP overrides PRUNE and THEN */  
   
806      case OP_SKIP:      case OP_SKIP:
807      RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md,      RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md,
808        eptrb, RM53);        eptrb, RM53);
809      if (rrc != MATCH_NOMATCH && rrc != MATCH_PRUNE && rrc != MATCH_THEN)      if (rrc != MATCH_NOMATCH) RRETURN(rrc);
       RRETURN(rrc);  
810      md->start_match_ptr = eptr;   /* Pass back current position */      md->start_match_ptr = eptr;   /* Pass back current position */
811      RRETURN(MATCH_SKIP);      RRETURN(MATCH_SKIP);
812    
813      /* Note that, for Perl compatibility, SKIP with an argument does NOT set      /* Note that, for Perl compatibility, SKIP with an argument does NOT set
814      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
815      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
816        that failed and any that precede it (either they also failed, or were not
817        triggered). To do this, we maintain a count of executed SKIP_ARGs. If a
818        SKIP_ARG gets to top level, the match is re-run with md->ignore_skip_arg
819        set to the count of the one that failed. */
820    
821      case OP_SKIP_ARG:      case OP_SKIP_ARG:
822      if (md->ignore_skip_arg)      md->skip_arg_count++;
823        if (md->skip_arg_count <= md->ignore_skip_arg)
824        {        {
825        ecode += PRIV(OP_lengths)[*ecode] + ecode[1];        ecode += PRIV(OP_lengths)[*ecode] + ecode[1];
826        break;        break;
827        }        }
828      RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode] + ecode[1], offset_top, md,      RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode] + ecode[1], offset_top, md,
829        eptrb, RM57);        eptrb, RM57);
830      if (rrc != MATCH_NOMATCH && rrc != MATCH_PRUNE && rrc != MATCH_THEN)      if (rrc != MATCH_NOMATCH) RRETURN(rrc);
       RRETURN(rrc);  
831    
832      /* 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
833      returning the special MATCH_SKIP_ARG return code. This will either be      returning the special MATCH_SKIP_ARG return code. This will either be
834      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
835      with the md->ignore_skip_arg flag set. */      with md->ignore_skip_arg set to the value of md->skip_arg_count. */
836    
837      md->start_match_ptr = ecode + 2;      md->start_match_ptr = ecode + 2;
838      RRETURN(MATCH_SKIP_ARG);      RRETURN(MATCH_SKIP_ARG);
# Line 838  for (;;) Line 875  for (;;)
875      case OP_ONCE_NC:      case OP_ONCE_NC:
876      prev = ecode;      prev = ecode;
877      saved_eptr = eptr;      saved_eptr = eptr;
878      save_mark = md->mark;      save_mark = md->mark;
879      do      do
880        {        {
881        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 894  for (;;)
894    
895        if (rrc != MATCH_NOMATCH) RRETURN(rrc);        if (rrc != MATCH_NOMATCH) RRETURN(rrc);
896        ecode += GET(ecode,1);        ecode += GET(ecode,1);
897        md->mark = save_mark;        md->mark = save_mark;
898        }        }
899      while (*ecode == OP_ALT);      while (*ecode == OP_ALT);
900    
# Line 897  for (;;) Line 934  for (;;)
934        }        }
935      else  /* OP_KETRMAX */      else  /* OP_KETRMAX */
936        {        {
       md->match_function_type = MATCH_CBEGROUP;  
937        RMATCH(eptr, prev, offset_top, md, eptrb, RM66);        RMATCH(eptr, prev, offset_top, md, eptrb, RM66);
938        if (rrc != MATCH_NOMATCH) RRETURN(rrc);        if (rrc != MATCH_NOMATCH) RRETURN(rrc);
939        ecode += 1 + LINK_SIZE;        ecode += 1 + LINK_SIZE;
# Line 937  for (;;) Line 973  for (;;)
973        save_offset2 = md->offset_vector[offset+1];        save_offset2 = md->offset_vector[offset+1];
974        save_offset3 = md->offset_vector[md->offset_end - number];        save_offset3 = md->offset_vector[md->offset_end - number];
975        save_capture_last = md->capture_last;        save_capture_last = md->capture_last;
976        save_mark = md->mark;        save_mark = md->mark;
977    
978        DPRINTF(("saving %d %d %d\n", save_offset1, save_offset2, save_offset3));        DPRINTF(("saving %d %d %d\n", save_offset1, save_offset2, save_offset3));
979        md->offset_vector[md->offset_end - number] =        md->offset_vector[md->offset_end - number] =
# Line 1026  for (;;) Line 1062  for (;;)
1062    
1063      for (;;)      for (;;)
1064        {        {
1065        if (op >= OP_SBRA || op == OP_ONCE) md->match_function_type = MATCH_CBEGROUP;        if (op >= OP_SBRA || op == OP_ONCE)
1066            md->match_function_type = MATCH_CBEGROUP;
1067    
1068        /* 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
1069        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 1078  for (;;)
1078        /* In all other cases, we have to make another call to match(). */        /* In all other cases, we have to make another call to match(). */
1079    
1080        save_mark = md->mark;        save_mark = md->mark;
1081          save_capture_last = md->capture_last;
1082        RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md, eptrb,        RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md, eptrb,
1083          RM2);          RM2);
1084    
1085        /* See comment in the code for capturing groups above about handling        /* See comment in the code for capturing groups above about handling
1086        THEN. */        THEN. */
1087    
# Line 1070  for (;;) Line 1108  for (;;)
1108          RRETURN(rrc);          RRETURN(rrc);
1109          }          }
1110        ecode += GET(ecode, 1);        ecode += GET(ecode, 1);
1111        md->mark = save_mark;        md->mark = save_mark;
1112        if (*ecode != OP_ALT) break;        if (*ecode != OP_ALT) break;
1113          md->capture_last = save_capture_last;
1114        }        }
1115    
1116      RRETURN(MATCH_NOMATCH);      RRETURN(MATCH_NOMATCH);
# Line 1193  for (;;) Line 1232  for (;;)
1232      POSSESSIVE_NON_CAPTURE:      POSSESSIVE_NON_CAPTURE:
1233      matched_once = FALSE;      matched_once = FALSE;
1234      code_offset = (int)(ecode - md->start_code);      code_offset = (int)(ecode - md->start_code);
1235        save_capture_last = md->capture_last;
1236    
1237      for (;;)      for (;;)
1238        {        {
# 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 1253  for (;;) Line 1294  for (;;)
1294          cb.version          = 2;   /* Version 1 of the callout block */          cb.version          = 2;   /* Version 1 of the callout block */
1295          cb.callout_number   = ecode[LINK_SIZE+2];          cb.callout_number   = ecode[LINK_SIZE+2];
1296          cb.offset_vector    = md->offset_vector;          cb.offset_vector    = md->offset_vector;
1297  #ifdef COMPILE_PCRE8  #if defined COMPILE_PCRE8
1298          cb.subject          = (PCRE_SPTR)md->start_subject;          cb.subject          = (PCRE_SPTR)md->start_subject;
1299  #else  #elif defined COMPILE_PCRE16
1300          cb.subject          = (PCRE_SPTR16)md->start_subject;          cb.subject          = (PCRE_SPTR16)md->start_subject;
1301    #elif defined COMPILE_PCRE32
1302            cb.subject          = (PCRE_SPTR32)md->start_subject;
1303  #endif  #endif
1304          cb.subject_length   = (int)(md->end_subject - md->start_subject);          cb.subject_length   = (int)(md->end_subject - md->start_subject);
1305          cb.start_match      = (int)(mstart - md->start_subject);          cb.start_match      = (int)(mstart - md->start_subject);
# Line 1264  for (;;) Line 1307  for (;;)
1307          cb.pattern_position = GET(ecode, LINK_SIZE + 3);          cb.pattern_position = GET(ecode, LINK_SIZE + 3);
1308          cb.next_item_length = GET(ecode, 3 + 2*LINK_SIZE);          cb.next_item_length = GET(ecode, 3 + 2*LINK_SIZE);
1309          cb.capture_top      = offset_top/2;          cb.capture_top      = offset_top/2;
1310          cb.capture_last     = md->capture_last;          cb.capture_last     = md->capture_last & CAPLMASK;
1311            /* Internal change requires this for API compatibility. */
1312            if (cb.capture_last == 0) cb.capture_last = -1;
1313          cb.callout_data     = md->callout_data;          cb.callout_data     = md->callout_data;
1314          cb.mark             = md->nomatch_mark;          cb.mark             = md->nomatch_mark;
1315          if ((rrc = (*PUBL(callout))(&cb)) > 0) RRETURN(MATCH_NOMATCH);          if ((rrc = (*PUBL(callout))(&cb)) > 0) RRETURN(MATCH_NOMATCH);
1316          if (rrc < 0) RRETURN(rrc);          if (rrc < 0) RRETURN(rrc);
1317          }          }
1318        ecode += PRIV(OP_lengths)[OP_CALLOUT];        ecode += PRIV(OP_lengths)[OP_CALLOUT];
1319          codelink -= PRIV(OP_lengths)[OP_CALLOUT];
1320        }        }
1321    
1322      condcode = ecode[LINK_SIZE+1];      condcode = ecode[LINK_SIZE+1];
# Line 1286  for (;;) Line 1332  for (;;)
1332          }          }
1333        else        else
1334          {          {
1335          int recno = GET2(ecode, LINK_SIZE + 2);   /* Recursion group number*/          unsigned int recno = GET2(ecode, LINK_SIZE + 2);   /* Recursion group number*/
1336          condition = (recno == RREF_ANY || recno == md->recursive->group_num);          condition = (recno == RREF_ANY || recno == md->recursive->group_num);
1337    
1338          /* If the test is for recursion into a specific subpattern, and it is          /* If the test is for recursion into a specific subpattern, and it is
# Line 1358  for (;;) Line 1404  for (;;)
1404    
1405        if (!condition && condcode == OP_NCREF)        if (!condition && condcode == OP_NCREF)
1406          {          {
1407          int refno = offset >> 1;          unsigned int refno = offset >> 1;
1408          pcre_uchar *slotA = md->name_table;          pcre_uchar *slotA = md->name_table;
1409    
1410          for (i = 0; i < md->name_count; i++)          for (i = 0; i < md->name_count; i++)
# Line 1486  for (;;) Line 1532  for (;;)
1532      to close any currently open capturing brackets. */      to close any currently open capturing brackets. */
1533    
1534      case OP_CLOSE:      case OP_CLOSE:
1535      number = GET2(ecode, 1);      number = GET2(ecode, 1);   /* Must be less than 65536 */
1536      offset = number << 1;      offset = number << 1;
1537    
1538  #ifdef PCRE_DEBUG  #ifdef PCRE_DEBUG
# Line 1494  for (;;) Line 1540  for (;;)
1540        printf("\n");        printf("\n");
1541  #endif  #endif
1542    
1543      md->capture_last = number;      md->capture_last = (md->capture_last & OVFLMASK) | number;
1544      if (offset >= md->offset_max) md->offset_overflow = TRUE; else      if (offset >= md->offset_max) md->capture_last |= OVFLBIT; else
1545        {        {
1546        md->offset_vector[offset] =        md->offset_vector[offset] =
1547          md->offset_vector[md->offset_end - number];          md->offset_vector[md->offset_end - number];
# Line 1549  for (;;) Line 1595  for (;;)
1595    
1596      case OP_ASSERT:      case OP_ASSERT:
1597      case OP_ASSERTBACK:      case OP_ASSERTBACK:
1598      save_mark = md->mark;      save_mark = md->mark;
1599      if (md->match_function_type == MATCH_CONDASSERT)      if (md->match_function_type == MATCH_CONDASSERT)
1600        {        {
1601        condassert = TRUE;        condassert = TRUE;
# Line 1557  for (;;) Line 1603  for (;;)
1603        }        }
1604      else condassert = FALSE;      else condassert = FALSE;
1605    
1606        /* Loop for each branch */
1607    
1608      do      do
1609        {        {
1610        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, NULL, RM4);        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, NULL, RM4);
1611    
1612          /* A match means that the assertion is true; break out of the loop
1613          that matches its alternatives. */
1614    
1615        if (rrc == MATCH_MATCH || rrc == MATCH_ACCEPT)        if (rrc == MATCH_MATCH || rrc == MATCH_ACCEPT)
1616          {          {
1617          mstart = md->start_match_ptr;   /* In case \K reset it */          mstart = md->start_match_ptr;   /* In case \K reset it */
1618          break;          break;
1619          }          }
1620    
1621        /* PCRE does not allow THEN to escape beyond an assertion; it is treated        /* If not matched, restore the previous mark setting. */
1622        as NOMATCH. */  
1623          md->mark = save_mark;
1624    
1625          /* See comment in the code for capturing groups above about handling
1626          THEN. */
1627    
1628        if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN) RRETURN(rrc);        if (rrc == MATCH_THEN)
1629            {
1630            next = ecode + GET(ecode,1);
1631            if (md->start_match_ptr < next &&
1632                (*ecode == OP_ALT || *next == OP_ALT))
1633              rrc = MATCH_NOMATCH;
1634            }
1635    
1636          /* Anything other than NOMATCH causes the entire assertion to fail,
1637          passing back the return code. This includes COMMIT, SKIP, PRUNE and an
1638          uncaptured THEN, which means they take their normal effect. This
1639          consistent approach does not always have exactly the same effect as in
1640          Perl. */
1641    
1642          if (rrc != MATCH_NOMATCH) RRETURN(rrc);
1643        ecode += GET(ecode, 1);        ecode += GET(ecode, 1);
       md->mark = save_mark;  
1644        }        }
1645      while (*ecode == OP_ALT);      while (*ecode == OP_ALT);   /* Continue for next alternative */
1646    
1647        /* If we have tried all the alternative branches, the assertion has
1648        failed. If not, we broke out after a match. */
1649    
1650      if (*ecode == OP_KET) RRETURN(MATCH_NOMATCH);      if (*ecode == OP_KET) RRETURN(MATCH_NOMATCH);
1651    
# Line 1581  for (;;) Line 1653  for (;;)
1653    
1654      if (condassert) RRETURN(MATCH_MATCH);      if (condassert) RRETURN(MATCH_MATCH);
1655    
1656      /* Continue from after the assertion, updating the offsets high water      /* Continue from after a successful assertion, updating the offsets high
1657      mark, since extracts may have been taken during the assertion. */      water mark, since extracts may have been taken during the assertion. */
1658    
1659      do ecode += GET(ecode,1); while (*ecode == OP_ALT);      do ecode += GET(ecode,1); while (*ecode == OP_ALT);
1660      ecode += 1 + LINK_SIZE;      ecode += 1 + LINK_SIZE;
1661      offset_top = md->end_offset_top;      offset_top = md->end_offset_top;
1662      continue;      continue;
1663    
1664      /* Negative assertion: all branches must fail to match. Encountering SKIP,      /* Negative assertion: all branches must fail to match for the assertion to
1665      PRUNE, or COMMIT means we must assume failure without checking subsequent      succeed. */
     branches. */  
1666    
1667      case OP_ASSERT_NOT:      case OP_ASSERT_NOT:
1668      case OP_ASSERTBACK_NOT:      case OP_ASSERTBACK_NOT:
1669      save_mark = md->mark;      save_mark = md->mark;
1670      if (md->match_function_type == MATCH_CONDASSERT)      if (md->match_function_type == MATCH_CONDASSERT)
1671        {        {
1672        condassert = TRUE;        condassert = TRUE;
# Line 1603  for (;;) Line 1674  for (;;)
1674        }        }
1675      else condassert = FALSE;      else condassert = FALSE;
1676    
1677        /* Loop for each alternative branch. */
1678    
1679      do      do
1680        {        {
1681        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, NULL, RM5);        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, NULL, RM5);
1682        md->mark = save_mark;        md->mark = save_mark;   /* Always restore the mark setting */
1683        if (rrc == MATCH_MATCH || rrc == MATCH_ACCEPT) RRETURN(MATCH_NOMATCH);  
1684        if (rrc == MATCH_SKIP || rrc == MATCH_PRUNE || rrc == MATCH_COMMIT)        switch(rrc)
1685          {          {
1686          do ecode += GET(ecode,1); while (*ecode == OP_ALT);          case MATCH_MATCH:            /* A successful match means */
1687            case MATCH_ACCEPT:           /* the assertion has failed. */
1688            RRETURN(MATCH_NOMATCH);
1689    
1690            case MATCH_NOMATCH:          /* Carry on with next branch */
1691          break;          break;
1692    
1693            /* See comment in the code for capturing groups above about handling
1694            THEN. */
1695    
1696            case MATCH_THEN:
1697            next = ecode + GET(ecode,1);
1698            if (md->start_match_ptr < next &&
1699                (*ecode == OP_ALT || *next == OP_ALT))
1700              {
1701              rrc = MATCH_NOMATCH;
1702              break;
1703              }
1704            /* Otherwise fall through. */
1705    
1706            /* COMMIT, SKIP, PRUNE, and an uncaptured THEN cause the whole
1707            assertion to fail to match, without considering any more alternatives.
1708            Failing to match means the assertion is true. This is a consistent
1709            approach, but does not always have the same effect as in Perl. */
1710    
1711            case MATCH_COMMIT:
1712            case MATCH_SKIP:
1713            case MATCH_SKIP_ARG:
1714            case MATCH_PRUNE:
1715            do ecode += GET(ecode,1); while (*ecode == OP_ALT);
1716            goto NEG_ASSERT_TRUE;   /* Break out of alternation loop */
1717    
1718            /* Anything else is an error */
1719    
1720            default:
1721            RRETURN(rrc);
1722          }          }
1723    
1724        /* PCRE does not allow THEN to escape beyond an assertion; it is treated        /* Continue with next branch */
       as NOMATCH. */  
1725    
       if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN) RRETURN(rrc);  
1726        ecode += GET(ecode,1);        ecode += GET(ecode,1);
1727        }        }
1728      while (*ecode == OP_ALT);      while (*ecode == OP_ALT);
1729    
1730      if (condassert) RRETURN(MATCH_MATCH);  /* Condition assertion */      /* All branches in the assertion failed to match. */
1731    
1732      ecode += 1 + LINK_SIZE;      NEG_ASSERT_TRUE:
1733        if (condassert) RRETURN(MATCH_MATCH);  /* Condition assertion */
1734        ecode += 1 + LINK_SIZE;                /* Continue with current branch */
1735      continue;      continue;
1736    
1737      /* 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 1778  for (;;)
1778        cb.version          = 2;   /* Version 1 of the callout block */        cb.version          = 2;   /* Version 1 of the callout block */
1779        cb.callout_number   = ecode[1];        cb.callout_number   = ecode[1];
1780        cb.offset_vector    = md->offset_vector;        cb.offset_vector    = md->offset_vector;
1781  #ifdef COMPILE_PCRE8  #if defined COMPILE_PCRE8
1782        cb.subject          = (PCRE_SPTR)md->start_subject;        cb.subject          = (PCRE_SPTR)md->start_subject;
1783  #else  #elif defined COMPILE_PCRE16
1784        cb.subject          = (PCRE_SPTR16)md->start_subject;        cb.subject          = (PCRE_SPTR16)md->start_subject;
1785    #elif defined COMPILE_PCRE32
1786          cb.subject          = (PCRE_SPTR32)md->start_subject;
1787  #endif  #endif
1788        cb.subject_length   = (int)(md->end_subject - md->start_subject);        cb.subject_length   = (int)(md->end_subject - md->start_subject);
1789        cb.start_match      = (int)(mstart - md->start_subject);        cb.start_match      = (int)(mstart - md->start_subject);
# Line 1682  for (;;) Line 1791  for (;;)
1791        cb.pattern_position = GET(ecode, 2);        cb.pattern_position = GET(ecode, 2);
1792        cb.next_item_length = GET(ecode, 2 + LINK_SIZE);        cb.next_item_length = GET(ecode, 2 + LINK_SIZE);
1793        cb.capture_top      = offset_top/2;        cb.capture_top      = offset_top/2;
1794        cb.capture_last     = md->capture_last;        cb.capture_last     = md->capture_last & CAPLMASK;
1795          /* Internal change requires this for API compatibility. */
1796          if (cb.capture_last == 0) cb.capture_last = -1;
1797        cb.callout_data     = md->callout_data;        cb.callout_data     = md->callout_data;
1798        cb.mark             = md->nomatch_mark;        cb.mark             = md->nomatch_mark;
1799        if ((rrc = (*PUBL(callout))(&cb)) > 0) RRETURN(MATCH_NOMATCH);        if ((rrc = (*PUBL(callout))(&cb)) > 0) RRETURN(MATCH_NOMATCH);
# Line 1711  for (;;) Line 1822  for (;;)
1822      case OP_RECURSE:      case OP_RECURSE:
1823        {        {
1824        recursion_info *ri;        recursion_info *ri;
1825        int recno;        unsigned int recno;
1826    
1827        callpat = md->start_code + GET(ecode, 1);        callpat = md->start_code + GET(ecode, 1);
1828        recno = (callpat == md->start_code)? 0 :        recno = (callpat == md->start_code)? 0 :
# Line 1728  for (;;) Line 1839  for (;;)
1839        /* Add to "recursing stack" */        /* Add to "recursing stack" */
1840    
1841        new_recursive.group_num = recno;        new_recursive.group_num = recno;
1842          new_recursive.saved_capture_last = md->capture_last;
1843        new_recursive.subject_position = eptr;        new_recursive.subject_position = eptr;
1844        new_recursive.prevrec = md->recursive;        new_recursive.prevrec = md->recursive;
1845        md->recursive = &new_recursive;        md->recursive = &new_recursive;
# Line 1751  for (;;) Line 1863  for (;;)
1863              new_recursive.saved_max * sizeof(int));              new_recursive.saved_max * sizeof(int));
1864    
1865        /* OK, now we can do the recursion. After processing each alternative,        /* OK, now we can do the recursion. After processing each alternative,
1866        restore the offset data. If there were nested recursions, md->recursive        restore the offset data and the last captured value. If there were nested
1867        might be changed, so reset it before looping. */        recursions, md->recursive might be changed, so reset it before looping.
1868          */
1869    
1870        DPRINTF(("Recursing into group %d\n", new_recursive.group_num));        DPRINTF(("Recursing into group %d\n", new_recursive.group_num));
1871        cbegroup = (*callpat >= OP_SBRA);        cbegroup = (*callpat >= OP_SBRA);
# Line 1763  for (;;) Line 1876  for (;;)
1876            md, eptrb, RM6);            md, eptrb, RM6);
1877          memcpy(md->offset_vector, new_recursive.offset_save,          memcpy(md->offset_vector, new_recursive.offset_save,
1878              new_recursive.saved_max * sizeof(int));              new_recursive.saved_max * sizeof(int));
1879            md->capture_last = new_recursive.saved_capture_last;
1880          md->recursive = new_recursive.prevrec;          md->recursive = new_recursive.prevrec;
1881          if (rrc == MATCH_MATCH || rrc == MATCH_ACCEPT)          if (rrc == MATCH_MATCH || rrc == MATCH_ACCEPT)
1882            {            {
# Line 1779  for (;;) Line 1893  for (;;)
1893            goto RECURSION_MATCHED;        /* Exit loop; end processing */            goto RECURSION_MATCHED;        /* Exit loop; end processing */
1894            }            }
1895    
1896          /* 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
1897          as NOMATCH. */          recursion; they cause a NOMATCH for the entire recursion. These codes
1898            are defined in a range that can be tested for. */
1899    
1900            if (rrc >= MATCH_BACKTRACK_MIN && rrc <= MATCH_BACKTRACK_MAX)
1901              RRETURN(MATCH_NOMATCH);
1902    
1903            /* Any return code other than NOMATCH is an error. */
1904    
1905          else if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN)          if (rrc != MATCH_NOMATCH)
1906            {            {
1907            DPRINTF(("Recursion gave error %d\n", rrc));            DPRINTF(("Recursion gave error %d\n", rrc));
1908            if (new_recursive.offset_save != stacksave)            if (new_recursive.offset_save != stacksave)
# Line 1912  for (;;) Line 2032  for (;;)
2032    
2033        /* Deal with capturing */        /* Deal with capturing */
2034    
2035        md->capture_last = number;        md->capture_last = (md->capture_last & OVFLMASK) | number;
2036        if (offset >= md->offset_max) md->offset_overflow = TRUE; else        if (offset >= md->offset_max) md->capture_last |= OVFLBIT; else
2037          {          {
2038          /* If offset is greater than offset_top, it means that we are          /* If offset is greater than offset_top, it means that we are
2039          "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 1993  for (;;) Line 2113  for (;;)
2113          }          }
2114        if (*prev >= OP_SBRA)    /* Could match an empty string */        if (*prev >= OP_SBRA)    /* Could match an empty string */
2115          {          {
         md->match_function_type = MATCH_CBEGROUP;  
2116          RMATCH(eptr, prev, offset_top, md, eptrb, RM50);          RMATCH(eptr, prev, offset_top, md, eptrb, RM50);
2117          RRETURN(rrc);          RRETURN(rrc);
2118          }          }
# Line 2002  for (;;) Line 2121  for (;;)
2121        }        }
2122      else  /* OP_KETRMAX */      else  /* OP_KETRMAX */
2123        {        {
       if (*prev >= OP_SBRA) md->match_function_type = MATCH_CBEGROUP;  
2124        RMATCH(eptr, prev, offset_top, md, eptrb, RM13);        RMATCH(eptr, prev, offset_top, md, eptrb, RM13);
2125        if (rrc == MATCH_ONCE && md->once_target == prev) rrc = MATCH_NOMATCH;        if (rrc == MATCH_ONCE && md->once_target == prev) rrc = MATCH_NOMATCH;
2126        if (rrc != MATCH_NOMATCH) RRETURN(rrc);        if (rrc != MATCH_NOMATCH) RRETURN(rrc);
# Line 2059  for (;;) Line 2177  for (;;)
2177    
2178      case OP_DOLLM:      case OP_DOLLM:
2179      if (eptr < md->end_subject)      if (eptr < md->end_subject)
2180        { if (!IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH); }        {
2181          if (!IS_NEWLINE(eptr))
2182            {
2183            if (md->partial != 0 &&
2184                eptr + 1 >= md->end_subject &&
2185                NLBLOCK->nltype == NLTYPE_FIXED &&
2186                NLBLOCK->nllen == 2 &&
2187                RAWUCHARTEST(eptr) == NLBLOCK->nl[0])
2188              {
2189              md->hitend = TRUE;
2190              if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
2191              }
2192            RRETURN(MATCH_NOMATCH);
2193            }
2194          }
2195      else      else
2196        {        {
2197        if (md->noteol) RRETURN(MATCH_NOMATCH);        if (md->noteol) RRETURN(MATCH_NOMATCH);
# Line 2091  for (;;) Line 2223  for (;;)
2223      ASSERT_NL_OR_EOS:      ASSERT_NL_OR_EOS:
2224      if (eptr < md->end_subject &&      if (eptr < md->end_subject &&
2225          (!IS_NEWLINE(eptr) || eptr != md->end_subject - md->nllen))          (!IS_NEWLINE(eptr) || eptr != md->end_subject - md->nllen))
2226          {
2227          if (md->partial != 0 &&
2228              eptr + 1 >= md->end_subject &&
2229              NLBLOCK->nltype == NLTYPE_FIXED &&
2230              NLBLOCK->nllen == 2 &&
2231              RAWUCHARTEST(eptr) == NLBLOCK->nl[0])
2232            {
2233            md->hitend = TRUE;
2234            if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
2235            }
2236        RRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2237          }
2238    
2239      /* Either at end of string or \n before end. */      /* Either at end of string or \n before end. */
2240    
# Line 2219  for (;;) Line 2362  for (;;)
2362        }        }
2363      break;      break;
2364    
2365      /* Match a single character type; inline for speed */      /* Match any single character type except newline; have to take care with
2366        CRLF newlines and partial matching. */
2367    
2368      case OP_ANY:      case OP_ANY:
2369      if (IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH);      if (IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH);
2370        if (md->partial != 0 &&
2371            eptr + 1 >= md->end_subject &&
2372            NLBLOCK->nltype == NLTYPE_FIXED &&
2373            NLBLOCK->nllen == 2 &&
2374            RAWUCHARTEST(eptr) == NLBLOCK->nl[0])
2375          {
2376          md->hitend = TRUE;
2377          if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
2378          }
2379    
2380      /* Fall through */      /* Fall through */
2381    
2382        /* Match any single character whatsoever. */
2383    
2384      case OP_ALLANY:      case OP_ALLANY:
2385      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 */
2386        {                            /* not be updated before SCHECK_PARTIAL. */        {                            /* not be updated before SCHECK_PARTIAL. */
# Line 2364  for (;;) Line 2520  for (;;)
2520        {        {
2521        default: RRETURN(MATCH_NOMATCH);        default: RRETURN(MATCH_NOMATCH);
2522    
2523        case 0x000d:        case CHAR_CR:
2524        if (eptr < md->end_subject && *eptr == 0x0a) eptr++;        if (eptr >= md->end_subject)
2525            {
2526            SCHECK_PARTIAL();
2527            }
2528          else if (RAWUCHARTEST(eptr) == CHAR_LF) eptr++;
2529        break;        break;
2530    
2531        case 0x000a:        case CHAR_LF:
2532        break;        break;
2533    
2534        case 0x000b:        case CHAR_VT:
2535        case 0x000c:        case CHAR_FF:
2536        case 0x0085:        case CHAR_NEL:
2537    #ifndef EBCDIC
2538        case 0x2028:        case 0x2028:
2539        case 0x2029:        case 0x2029:
2540    #endif  /* Not EBCDIC */
2541        if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);        if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);
2542        break;        break;
2543        }        }
# Line 2391  for (;;) Line 2553  for (;;)
2553      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2554      switch(c)      switch(c)
2555        {        {
2556          HSPACE_CASES: RRETURN(MATCH_NOMATCH);  /* Byte and multibyte cases */
2557        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);  
2558        }        }
2559      ecode++;      ecode++;
2560      break;      break;
# Line 2425  for (;;) Line 2568  for (;;)
2568      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2569      switch(c)      switch(c)
2570        {        {
2571          HSPACE_CASES: break;  /* Byte and multibyte cases */
2572        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;  
2573        }        }
2574      ecode++;      ecode++;
2575      break;      break;
# Line 2459  for (;;) Line 2583  for (;;)
2583      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2584      switch(c)      switch(c)
2585        {        {
2586          VSPACE_CASES: RRETURN(MATCH_NOMATCH);
2587        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);  
2588        }        }
2589      ecode++;      ecode++;
2590      break;      break;
# Line 2481  for (;;) Line 2598  for (;;)
2598      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2599      switch(c)      switch(c)
2600        {        {
2601          VSPACE_CASES: break;
2602        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;  
2603        }        }
2604      ecode++;      ecode++;
2605      break;      break;
# Line 2507  for (;;) Line 2617  for (;;)
2617        }        }
2618      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2619        {        {
2620          const pcre_uint32 *cp;
2621        const ucd_record *prop = GET_UCD(c);        const ucd_record *prop = GET_UCD(c);
2622    
2623        switch(ecode[1])        switch(ecode[1])
# Line 2567  for (;;) Line 2678  for (;;)
2678            RRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
2679          break;          break;
2680    
2681            case PT_CLIST:
2682            cp = PRIV(ucd_caseless_sets) + ecode[2];
2683            for (;;)
2684              {
2685              if (c < *cp)
2686                { if (op == OP_PROP) { RRETURN(MATCH_NOMATCH); } else break; }
2687              if (c == *cp++)
2688                { if (op == OP_PROP) break; else { RRETURN(MATCH_NOMATCH); } }
2689              }
2690            break;
2691    
2692            case PT_UCNC:
2693            if ((c == CHAR_DOLLAR_SIGN || c == CHAR_COMMERCIAL_AT ||
2694                 c == CHAR_GRAVE_ACCENT || (c >= 0xa0 && c <= 0xd7ff) ||
2695                 c >= 0xe000) == (op == OP_NOTPROP))
2696              RRETURN(MATCH_NOMATCH);
2697            break;
2698    
2699          /* This should never occur */          /* This should never occur */
2700    
2701          default:          default:
# Line 2586  for (;;) Line 2715  for (;;)
2715        SCHECK_PARTIAL();        SCHECK_PARTIAL();
2716        RRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2717        }        }
2718      GETCHARINCTEST(c, eptr);      else
     if (UCD_CATEGORY(c) == ucp_M) RRETURN(MATCH_NOMATCH);  
     while (eptr < md->end_subject)  
2719        {        {
2720        int len = 1;        int lgb, rgb;
2721        if (!utf) c = *eptr; else { GETCHARLEN(c, eptr, len); }        GETCHARINCTEST(c, eptr);
2722        if (UCD_CATEGORY(c) != ucp_M) break;        lgb = UCD_GRAPHBREAK(c);
2723        eptr += len;        while (eptr < md->end_subject)
2724            {
2725            int len = 1;
2726            if (!utf) c = *eptr; else { GETCHARLEN(c, eptr, len); }
2727            rgb = UCD_GRAPHBREAK(c);
2728            if ((PRIV(ucp_gbtable)[lgb] & (1 << rgb)) == 0) break;
2729            lgb = rgb;
2730            eptr += len;
2731            }
2732        }        }
2733        CHECK_PARTIAL();
2734      ecode++;      ecode++;
2735      break;      break;
2736  #endif  #endif  /* SUPPORT_UCP */
2737    
2738    
2739      /* 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 2742  for (;;)
2742      similar code to character type repeats - written out again for speed.      similar code to character type repeats - written out again for speed.
2743      However, if the referenced string is the empty string, always treat      However, if the referenced string is the empty string, always treat
2744      it as matched, any number of times (otherwise there could be infinite      it as matched, any number of times (otherwise there could be infinite
2745      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:  
2746    
2747      (a) In the default, Perl-compatible state, set the length negative;      (a) In the default, Perl-compatible state, set the length negative;
2748      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 2752  for (;;)
2752      so that the back reference matches an empty string.      so that the back reference matches an empty string.
2753    
2754      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
2755      referenced subpattern. */      referenced subpattern.
2756    
2757        The OP_REF and OP_REFI opcodes are used for a reference to a numbered group
2758        or to a non-duplicated named group. For a duplicated named group, OP_DNREF
2759        and OP_DNREFI are used. In this case we must scan the list of groups to
2760        which the name refers, and use the first one that is set. */
2761    
2762        case OP_DNREF:
2763        case OP_DNREFI:
2764        caseless = op == OP_DNREFI;
2765          {
2766          int count = GET2(ecode, 1+IMM2_SIZE);
2767          pcre_uchar *slot = md->name_table + GET2(ecode, 1) * md->name_entry_size;
2768          ecode += 1 + 2*IMM2_SIZE;
2769    
2770          while (count-- > 0)
2771            {
2772            offset = GET2(slot, 0) << 1;
2773            if (offset < offset_top && md->offset_vector[offset] >= 0) break;
2774            slot += md->name_entry_size;
2775            }
2776          if (count < 0)
2777            length = (md->jscript_compat)? 0 : -1;
2778          else
2779            length = md->offset_vector[offset+1] - md->offset_vector[offset];
2780          }
2781        goto REF_REPEAT;
2782    
2783        case OP_REF:
2784        case OP_REFI:
2785        caseless = op == OP_REFI;
2786        offset = GET2(ecode, 1) << 1;               /* Doubled ref number */
2787        ecode += 1 + IMM2_SIZE;
2788    
2789    
2790      if (offset >= offset_top || md->offset_vector[offset] < 0)      if (offset >= offset_top || md->offset_vector[offset] < 0)
2791        length = (md->jscript_compat)? 0 : -1;        length = (md->jscript_compat)? 0 : -1;
# Line 2633  for (;;) Line 2794  for (;;)
2794    
2795      /* Set up for repetition, or handle the non-repeated case */      /* Set up for repetition, or handle the non-repeated case */
2796    
2797        REF_REPEAT:
2798      switch (*ecode)      switch (*ecode)
2799        {        {
2800        case OP_CRSTAR:        case OP_CRSTAR:
# Line 2660  for (;;) Line 2822  for (;;)
2822        default:               /* No repeat follows */        default:               /* No repeat follows */
2823        if ((length = match_ref(offset, eptr, length, md, caseless)) < 0)        if ((length = match_ref(offset, eptr, length, md, caseless)) < 0)
2824          {          {
2825            if (length == -2) eptr = md->end_subject;   /* Partial match */
2826          CHECK_PARTIAL();          CHECK_PARTIAL();
2827          RRETURN(MATCH_NOMATCH);          RRETURN(MATCH_NOMATCH);
2828          }          }
# Line 2685  for (;;) Line 2848  for (;;)
2848        int slength;        int slength;
2849        if ((slength = match_ref(offset, eptr, length, md, caseless)) < 0)        if ((slength = match_ref(offset, eptr, length, md, caseless)) < 0)
2850          {          {
2851            if (slength == -2) eptr = md->end_subject;   /* Partial match */
2852          CHECK_PARTIAL();          CHECK_PARTIAL();
2853          RRETURN(MATCH_NOMATCH);          RRETURN(MATCH_NOMATCH);
2854          }          }
# Line 2708  for (;;) Line 2872  for (;;)
2872          if (fi >= max) RRETURN(MATCH_NOMATCH);          if (fi >= max) RRETURN(MATCH_NOMATCH);
2873          if ((slength = match_ref(offset, eptr, length, md, caseless)) < 0)          if ((slength = match_ref(offset, eptr, length, md, caseless)) < 0)
2874            {            {
2875              if (slength == -2) eptr = md->end_subject;   /* Partial match */
2876            CHECK_PARTIAL();            CHECK_PARTIAL();
2877            RRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
2878            }            }
# Line 2726  for (;;) Line 2891  for (;;)
2891          int slength;          int slength;
2892          if ((slength = match_ref(offset, eptr, length, md, caseless)) < 0)          if ((slength = match_ref(offset, eptr, length, md, caseless)) < 0)
2893            {            {
2894            CHECK_PARTIAL();            /* Can't use CHECK_PARTIAL because we don't want to update eptr in
2895              the soft partial matching case. */
2896    
2897              if (slength == -2 && md->partial != 0 &&
2898                  md->end_subject > md->start_used_ptr)
2899                {
2900                md->hitend = TRUE;
2901                if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
2902                }
2903            break;            break;
2904            }            }
2905          eptr += slength;          eptr += slength;
2906          }          }
2907    
2908        while (eptr >= pp)        while (eptr >= pp)
2909          {          {
2910          RMATCH(eptr, ecode, offset_top, md, eptrb, RM15);          RMATCH(eptr, ecode, offset_top, md, eptrb, RM15);
# Line 3094  for (;;) Line 3268  for (;;)
3268          CHECK_PARTIAL();             /* Not SCHECK_PARTIAL() */          CHECK_PARTIAL();             /* Not SCHECK_PARTIAL() */
3269          RRETURN(MATCH_NOMATCH);          RRETURN(MATCH_NOMATCH);
3270          }          }
3271        while (length-- > 0) if (*ecode++ != *eptr++) RRETURN(MATCH_NOMATCH);        while (length-- > 0) if (*ecode++ != RAWUCHARINC(eptr)) RRETURN(MATCH_NOMATCH);
3272        }        }
3273      else      else
3274  #endif  #endif
# Line 3134  for (;;) Line 3308  for (;;)
3308    
3309        if (fc < 128)        if (fc < 128)
3310          {          {
3311          if (md->lcc[fc]          pcre_uint32 cc = RAWUCHAR(eptr);
3312              != TABLE_GET(*eptr, md->lcc, *eptr)) RRETURN(MATCH_NOMATCH);          if (md->lcc[fc] != TABLE_GET(cc, md->lcc, cc)) RRETURN(MATCH_NOMATCH);
3313          ecode++;          ecode++;
3314          eptr++;          eptr++;
3315          }          }
# Line 3146  for (;;) Line 3320  for (;;)
3320    
3321        else        else
3322          {          {
3323          unsigned int dc;          pcre_uint32 dc;
3324          GETCHARINC(dc, eptr);          GETCHARINC(dc, eptr);
3325          ecode += length;          ecode += length;
3326    
# Line 3239  for (;;) Line 3413  for (;;)
3413      max = rep_max[c];                 /* zero for max => infinity */      max = rep_max[c];                 /* zero for max => infinity */
3414      if (max == 0) max = INT_MAX;      if (max == 0) max = INT_MAX;
3415    
3416      /* Common code for all repeated single-character matches. */      /* Common code for all repeated single-character matches. We first check
3417        for the minimum number of characters. If the minimum equals the maximum, we
3418        are done. Otherwise, if minimizing, check the rest of the pattern for a
3419        match; if there isn't one, advance up to the maximum, one character at a
3420        time.
3421    
3422        If maximizing, advance up to the maximum number of matching characters,
3423        until eptr is past the end of the maximum run. If possessive, we are
3424        then done (no backing up). Otherwise, match at this position; anything
3425        other than no match is immediately returned. For nomatch, back up one
3426        character, unless we are matching \R and the last thing matched was
3427        \r\n, in which case, back up two bytes. When we reach the first optional
3428        character position, we can save stack by doing a tail recurse.
3429    
3430        The various UTF/non-UTF and caseful/caseless cases are handled separately,
3431        for speed. */
3432    
3433      REPEATCHAR:      REPEATCHAR:
3434  #ifdef SUPPORT_UTF  #ifdef SUPPORT_UTF
# Line 3256  for (;;) Line 3445  for (;;)
3445        if (length > 1)        if (length > 1)
3446          {          {
3447  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
3448          unsigned int othercase;          pcre_uint32 othercase;
3449          if (op >= OP_STARI &&     /* Caseless */          if (op >= OP_STARI &&     /* Caseless */
3450              (othercase = UCD_OTHERCASE(fc)) != fc)              (othercase = UCD_OTHERCASE(fc)) != fc)
3451            oclength = PRIV(ord2utf)(othercase, occhars);            oclength = PRIV(ord2utf)(othercase, occhars);
# Line 3323  for (;;) Line 3512  for (;;)
3512                }                }
3513              }              }
3514    
3515            if (possessive) continue;            if (possessive) continue;    /* No backtracking */
   
3516            for(;;)            for(;;)
3517              {              {
3518                if (eptr == pp) goto TAIL_RECURSE;
3519              RMATCH(eptr, ecode, offset_top, md, eptrb, RM23);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM23);
3520              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
             if (eptr == pp) { RRETURN(MATCH_NOMATCH); }  
3521  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
3522              eptr--;              eptr--;
3523              BACKCHAR(eptr);              BACKCHAR(eptr);
# Line 3360  for (;;) Line 3548  for (;;)
3548      maximizing, find the maximum number of characters and work backwards. */      maximizing, find the maximum number of characters and work backwards. */
3549    
3550      DPRINTF(("matching %c{%d,%d} against subject %.*s\n", fc, min, max,      DPRINTF(("matching %c{%d,%d} against subject %.*s\n", fc, min, max,
3551        max, eptr));        max, (char *)eptr));
3552    
3553      if (op >= OP_STARI)  /* Caseless */      if (op >= OP_STARI)  /* Caseless */
3554        {        {
# Line 3383  for (;;) Line 3571  for (;;)
3571    
3572        for (i = 1; i <= min; i++)        for (i = 1; i <= min; i++)
3573          {          {
3574            pcre_uint32 cc;                 /* Faster than pcre_uchar */
3575          if (eptr >= md->end_subject)          if (eptr >= md->end_subject)
3576            {            {
3577            SCHECK_PARTIAL();            SCHECK_PARTIAL();
3578            RRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
3579            }            }
3580          if (fc != *eptr && foc != *eptr) RRETURN(MATCH_NOMATCH);          cc = RAWUCHARTEST(eptr);
3581            if (fc != cc && foc != cc) RRETURN(MATCH_NOMATCH);
3582          eptr++;          eptr++;
3583          }          }
3584        if (min == max) continue;        if (min == max) continue;
# Line 3396  for (;;) Line 3586  for (;;)
3586          {          {
3587          for (fi = min;; fi++)          for (fi = min;; fi++)
3588            {            {
3589              pcre_uint32 cc;               /* Faster than pcre_uchar */
3590            RMATCH(eptr, ecode, offset_top, md, eptrb, RM24);            RMATCH(eptr, ecode, offset_top, md, eptrb, RM24);
3591            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3592            if (fi >= max) RRETURN(MATCH_NOMATCH);            if (fi >= max) RRETURN(MATCH_NOMATCH);
# Line 3404  for (;;) Line 3595  for (;;)
3595              SCHECK_PARTIAL();              SCHECK_PARTIAL();
3596              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
3597              }              }
3598            if (fc != *eptr && foc != *eptr) RRETURN(MATCH_NOMATCH);            cc = RAWUCHARTEST(eptr);
3599              if (fc != cc && foc != cc) RRETURN(MATCH_NOMATCH);
3600            eptr++;            eptr++;
3601            }            }
3602          /* Control never gets here */          /* Control never gets here */
# Line 3414  for (;;) Line 3606  for (;;)
3606          pp = eptr;          pp = eptr;
3607          for (i = min; i < max; i++)          for (i = min; i < max; i++)
3608            {            {
3609              pcre_uint32 cc;               /* Faster than pcre_uchar */
3610            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
3611              {              {
3612              SCHECK_PARTIAL();              SCHECK_PARTIAL();
3613              break;              break;
3614              }              }
3615            if (fc != *eptr && foc != *eptr) break;            cc = RAWUCHARTEST(eptr);
3616              if (fc != cc && foc != cc) break;
3617            eptr++;            eptr++;
3618            }            }
3619            if (possessive) continue;       /* No backtracking */
3620          if (possessive) continue;          for (;;)
   
         while (eptr >= pp)  
3621            {            {
3622              if (eptr == pp) goto TAIL_RECURSE;
3623            RMATCH(eptr, ecode, offset_top, md, eptrb, RM25);            RMATCH(eptr, ecode, offset_top, md, eptrb, RM25);
3624            eptr--;            eptr--;
3625            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3626            }            }
3627          RRETURN(MATCH_NOMATCH);          /* Control never gets here */
3628          }          }
       /* Control never gets here */  
3629        }        }
3630    
3631      /* Caseful comparisons (includes all multi-byte characters) */      /* Caseful comparisons (includes all multi-byte characters) */
# Line 3447  for (;;) Line 3639  for (;;)
3639            SCHECK_PARTIAL();            SCHECK_PARTIAL();
3640            RRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
3641            }            }
3642          if (fc != *eptr++) RRETURN(MATCH_NOMATCH);          if (fc != RAWUCHARINCTEST(eptr)) RRETURN(MATCH_NOMATCH);
3643          }          }
3644    
3645        if (min == max) continue;        if (min == max) continue;
# Line 3464  for (;;) Line 3656  for (;;)
3656              SCHECK_PARTIAL();              SCHECK_PARTIAL();
3657              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
3658              }              }
3659            if (fc != *eptr++) RRETURN(MATCH_NOMATCH);            if (fc != RAWUCHARINCTEST(eptr)) RRETURN(MATCH_NOMATCH);
3660            }            }
3661          /* Control never gets here */          /* Control never gets here */
3662          }          }
# Line 3478  for (;;) Line 3670  for (;;)
3670              SCHECK_PARTIAL();              SCHECK_PARTIAL();
3671              break;              break;
3672              }              }
3673            if (fc != *eptr) break;            if (fc != RAWUCHARTEST(eptr)) break;
3674            eptr++;            eptr++;
3675            }            }
3676          if (possessive) continue;          if (possessive) continue;    /* No backtracking */
3677            for (;;)
         while (eptr >= pp)  
3678            {            {
3679              if (eptr == pp) goto TAIL_RECURSE;
3680            RMATCH(eptr, ecode, offset_top, md, eptrb, RM27);            RMATCH(eptr, ecode, offset_top, md, eptrb, RM27);
3681            eptr--;            eptr--;
3682            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3683            }            }
3684          RRETURN(MATCH_NOMATCH);          /* Control never gets here */
3685          }          }
3686        }        }
3687      /* Control never gets here */      /* Control never gets here */
# Line 3504  for (;;) Line 3696  for (;;)
3696        SCHECK_PARTIAL();        SCHECK_PARTIAL();
3697        RRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
3698        }        }
     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  
3699  #ifdef SUPPORT_UTF  #ifdef SUPPORT_UTF
3700        if (utf)
3701          {
3702          register pcre_uint32 ch, och;
3703    
3704          ecode++;
3705          GETCHARINC(ch, ecode);
3706          GETCHARINC(c, eptr);
3707    
3708          if (op == OP_NOT)
3709            {
3710            if (ch == c) RRETURN(MATCH_NOMATCH);
3711            }
3712          else
3713            {
3714  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
3715        if (utf && ch > 127)          if (ch > 127)
3716          och = UCD_OTHERCASE(ch);            och = UCD_OTHERCASE(ch);
3717  #else  #else
3718        if (utf && ch > 127)          if (ch > 127)
3719          och = ch;            och = ch;
3720  #endif /* SUPPORT_UCP */  #endif /* SUPPORT_UCP */
3721        else          else
3722  #endif /* SUPPORT_UTF */            och = TABLE_GET(ch, md->fcc, ch);
3723          och = TABLE_GET(ch, md->fcc, ch);          if (ch == c || och == c) RRETURN(MATCH_NOMATCH);
3724  #endif /* COMPILE_PCRE8 */          }
       if (ch == c || och == c) RRETURN(MATCH_NOMATCH);  
3725        }        }
3726      else    /* Caseful */      else
3727    #endif
3728        {        {
3729        if (*ecode++ == c) RRETURN(MATCH_NOMATCH);        register pcre_uint32 ch = ecode[1];
3730          c = *eptr++;
3731          if (ch == c || (op == OP_NOTI && TABLE_GET(ch, md->fcc, ch) == c))
3732            RRETURN(MATCH_NOMATCH);
3733          ecode += 2;
3734        }        }
3735      break;      break;
3736    
# Line 3610  for (;;) Line 3810  for (;;)
3810      /* Common code for all repeated single-byte matches. */      /* Common code for all repeated single-byte matches. */
3811    
3812      REPEATNOTCHAR:      REPEATNOTCHAR:
3813      fc = *ecode++;      GETCHARINCTEST(fc, ecode);
3814    
3815      /* The code is duplicated for the caseless and caseful cases, for speed,      /* The code is duplicated for the caseless and caseful cases, for speed,
3816      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 3821  for (;;)
3821      characters and work backwards. */      characters and work backwards. */
3822    
3823      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,
3824        max, eptr));        max, (char *)eptr));
3825    
3826      if (op >= OP_NOTSTARI)     /* Caseless */      if (op >= OP_NOTSTARI)     /* Caseless */
3827        {        {
 #ifdef COMPILE_PCRE8  
       /* fc must be < 128 if UTF is enabled. */  
       foc = md->fcc[fc];  
 #else  
3828  #ifdef SUPPORT_UTF  #ifdef SUPPORT_UTF
3829  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
3830        if (utf && fc > 127)        if (utf && fc > 127)
# Line 3640  for (;;) Line 3836  for (;;)
3836        else        else
3837  #endif /* SUPPORT_UTF */  #endif /* SUPPORT_UTF */
3838          foc = TABLE_GET(fc, md->fcc, fc);          foc = TABLE_GET(fc, md->fcc, fc);
 #endif /* COMPILE_PCRE8 */  
3839    
3840  #ifdef SUPPORT_UTF  #ifdef SUPPORT_UTF
3841        if (utf)        if (utf)
3842          {          {
3843          register unsigned int d;          register pcre_uint32 d;
3844          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
3845            {            {
3846            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
# Line 3654  for (;;) Line 3849  for (;;)
3849              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
3850              }              }
3851            GETCHARINC(d, eptr);            GETCHARINC(d, eptr);
3852            if (fc == d || foc == d) RRETURN(MATCH_NOMATCH);            if (fc == d || (unsigned int)foc == d) RRETURN(MATCH_NOMATCH);
3853            }            }
3854          }          }
3855        else        else
3856  #endif  #endif  /* SUPPORT_UTF */
3857        /* Not UTF mode */        /* Not UTF mode */
3858          {          {
3859          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
# Line 3680  for (;;) Line 3875  for (;;)
3875  #ifdef SUPPORT_UTF  #ifdef SUPPORT_UTF
3876          if (utf)          if (utf)
3877            {            {
3878            register unsigned int d;            register pcre_uint32 d;
3879            for (fi = min;; fi++)            for (fi = min;; fi++)
3880              {              {
3881              RMATCH(eptr, ecode, offset_top, md, eptrb, RM28);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM28);
# Line 3692  for (;;) Line 3887  for (;;)
3887                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
3888                }                }
3889              GETCHARINC(d, eptr);              GETCHARINC(d, eptr);
3890              if (fc == d || foc == d) RRETURN(MATCH_NOMATCH);              if (fc == d || (unsigned int)foc == d) RRETURN(MATCH_NOMATCH);
3891              }              }
3892            }            }
3893          else          else
3894  #endif  #endif  /*SUPPORT_UTF */
3895          /* Not UTF mode */          /* Not UTF mode */
3896            {            {
3897            for (fi = min;; fi++)            for (fi = min;; fi++)
# Line 3725  for (;;) Line 3920  for (;;)
3920  #ifdef SUPPORT_UTF  #ifdef SUPPORT_UTF
3921          if (utf)          if (utf)
3922            {            {
3923            register unsigned int d;            register pcre_uint32 d;
3924            for (i = min; i < max; i++)            for (i = min; i < max; i++)
3925              {              {
3926              int len = 1;              int len = 1;
# Line 3735  for (;;) Line 3930  for (;;)
3930                break;                break;
3931                }                }
3932              GETCHARLEN(d, eptr, len);              GETCHARLEN(d, eptr, len);
3933              if (fc == d || foc == d) break;              if (fc == d || (unsigned int)foc == d) break;
3934              eptr += len;              eptr += len;
3935              }              }
3936            if (possessive) continue;            if (possessive) continue;    /* No backtracking */
3937            for(;;)            for(;;)
3938              {              {
3939                if (eptr == pp) goto TAIL_RECURSE;
3940              RMATCH(eptr, ecode, offset_top, md, eptrb, RM30);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM30);
3941              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3942              if (eptr-- == pp) break;        /* Stop if tried at original pos */              eptr--;
3943              BACKCHAR(eptr);              BACKCHAR(eptr);
3944              }              }
3945            }            }
3946          else          else
3947  #endif  #endif  /* SUPPORT_UTF */
3948          /* Not UTF mode */          /* Not UTF mode */
3949            {            {
3950            for (i = min; i < max; i++)            for (i = min; i < max; i++)
# Line 3761  for (;;) Line 3957  for (;;)
3957              if (fc == *eptr || foc == *eptr) break;              if (fc == *eptr || foc == *eptr) break;
3958              eptr++;              eptr++;
3959              }              }
3960            if (possessive) continue;            if (possessive) continue;    /* No backtracking */
3961            while (eptr >= pp)            for (;;)
3962              {              {
3963                if (eptr == pp) goto TAIL_RECURSE;
3964              RMATCH(eptr, ecode, offset_top, md, eptrb, RM31);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM31);
3965              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3966              eptr--;              eptr--;
3967              }              }
3968            }            }
3969            /* Control never gets here */
         RRETURN(MATCH_NOMATCH);  
3970          }          }
       /* Control never gets here */  
3971        }        }
3972    
3973      /* Caseful comparisons */      /* Caseful comparisons */
# Line 3782  for (;;) Line 3977  for (;;)
3977  #ifdef SUPPORT_UTF  #ifdef SUPPORT_UTF
3978        if (utf)        if (utf)
3979          {          {
3980          register unsigned int d;          register pcre_uint32 d;
3981          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
3982            {            {
3983            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
# Line 3816  for (;;) Line 4011  for (;;)
4011  #ifdef SUPPORT_UTF  #ifdef SUPPORT_UTF
4012          if (utf)          if (utf)
4013            {            {
4014            register unsigned int d;            register pcre_uint32 d;
4015            for (fi = min;; fi++)            for (fi = min;; fi++)
4016              {              {
4017              RMATCH(eptr, ecode, offset_top, md, eptrb, RM32);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM32);
# Line 3860  for (;;) Line 4055  for (;;)
4055  #ifdef SUPPORT_UTF  #ifdef SUPPORT_UTF
4056          if (utf)          if (utf)
4057            {            {
4058            register unsigned int d;            register pcre_uint32 d;
4059            for (i = min; i < max; i++)            for (i = min; i < max; i++)
4060              {              {
4061              int len = 1;              int len = 1;
# Line 3873  for (;;) Line 4068  for (;;)
4068              if (fc == d) break;              if (fc == d) break;
4069              eptr += len;              eptr += len;
4070              }              }
4071            if (possessive) continue;            if (possessive) continue;    /* No backtracking */
4072            for(;;)            for(;;)
4073              {              {
4074                if (eptr == pp) goto TAIL_RECURSE;
4075              RMATCH(eptr, ecode, offset_top, md, eptrb, RM34);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM34);
4076              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
4077              if (eptr-- == pp) break;        /* Stop if tried at original pos */              eptr--;
4078              BACKCHAR(eptr);              BACKCHAR(eptr);
4079              }              }
4080            }            }
# Line 3896  for (;;) Line 4092  for (;;)
4092              if (fc == *eptr) break;              if (fc == *eptr) break;
4093              eptr++;              eptr++;
4094              }              }
4095            if (possessive) continue;            if (possessive) continue;    /* No backtracking */
4096            while (eptr >= pp)            for (;;)
4097              {              {
4098                if (eptr == pp) goto TAIL_RECURSE;
4099              RMATCH(eptr, ecode, offset_top, md, eptrb, RM35);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM35);
4100              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
4101              eptr--;              eptr--;
4102              }              }
4103            }            }
4104            /* Control never gets here */
         RRETURN(MATCH_NOMATCH);  
4105          }          }
4106        }        }
4107      /* Control never gets here */      /* Control never gets here */
# Line 4136  for (;;) Line 4332  for (;;)
4332              }              }
4333            break;            break;
4334    
4335              case PT_CLIST:
4336              for (i = 1; i <= min; i++)
4337                {
4338                const pcre_uint32 *cp;
4339                if (eptr >= md->end_subject)
4340                  {
4341                  SCHECK_PARTIAL();
4342                  RRETURN(MATCH_NOMATCH);
4343                  }
4344                GETCHARINCTEST(c, eptr);
4345                cp = PRIV(ucd_caseless_sets) + prop_value;
4346                for (;;)
4347                  {
4348                  if (c < *cp)
4349                    { if (prop_fail_result) break; else { RRETURN(MATCH_NOMATCH); } }
4350                  if (c == *cp++)
4351                    { if (prop_fail_result) { RRETURN(MATCH_NOMATCH); } else break; }
4352                  }
4353                }
4354              break;
4355    
4356              case PT_UCNC:
4357              for (i = 1; i <= min; i++)
4358                {
4359                if (eptr >= md->end_subject)
4360                  {
4361                  SCHECK_PARTIAL();
4362                  RRETURN(MATCH_NOMATCH);
4363                  }
4364                GETCHARINCTEST(c, eptr);
4365                if ((c == CHAR_DOLLAR_SIGN || c == CHAR_COMMERCIAL_AT ||
4366                     c == CHAR_GRAVE_ACCENT || (c >= 0xa0 && c <= 0xd7ff) ||
4367                     c >= 0xe000) == prop_fail_result)
4368                  RRETURN(MATCH_NOMATCH);
4369                }
4370              break;
4371    
4372            /* This should not occur */            /* This should not occur */
4373    
4374            default:            default:
# Line 4155  for (;;) Line 4388  for (;;)
4388              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4389              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4390              }              }
4391            GETCHARINCTEST(c, eptr);            else
           if (UCD_CATEGORY(c) == ucp_M) RRETURN(MATCH_NOMATCH);  
           while (eptr < md->end_subject)  
4392              {              {
4393              int len = 1;              int lgb, rgb;
4394              if (!utf) c = *eptr; else { GETCHARLEN(c, eptr, len); }              GETCHARINCTEST(c, eptr);
4395              if (UCD_CATEGORY(c) != ucp_M) break;              lgb = UCD_GRAPHBREAK(c);
4396              eptr += len;             while (eptr < md->end_subject)
4397                  {
4398                  int len = 1;
4399                  if (!utf) c = *eptr; else { GETCHARLEN(c, eptr, len); }
4400                  rgb = UCD_GRAPHBREAK(c);
4401                  if ((PRIV(ucp_gbtable)[lgb] & (1 << rgb)) == 0) break;
4402                  lgb = rgb;
4403                  eptr += len;
4404                  }
4405              }              }
4406              CHECK_PARTIAL();
4407            }            }
4408          }          }
4409    
# Line 4184  for (;;) Line 4424  for (;;)
4424              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4425              }              }
4426            if (IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH);            if (IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH);
4427              if (md->partial != 0 &&
4428                  eptr + 1 >= md->end_subject &&
4429                  NLBLOCK->nltype == NLTYPE_FIXED &&
4430                  NLBLOCK->nllen == 2 &&
4431                  RAWUCHAR(eptr) == NLBLOCK->nl[0])
4432                {
4433                md->hitend = TRUE;
4434                if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
4435                }
4436            eptr++;            eptr++;
4437            ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);            ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);
4438            }            }
# Line 4220  for (;;) Line 4469  for (;;)
4469              {              {
4470              default: RRETURN(MATCH_NOMATCH);              default: RRETURN(MATCH_NOMATCH);
4471    
4472              case 0x000d:              case CHAR_CR:
4473              if (eptr < md->end_subject && *eptr == 0x0a) eptr++;              if (eptr < md->end_subject && RAWUCHAR(eptr) == CHAR_LF) eptr++;
4474              break;              break;
4475    
4476              case 0x000a:              case CHAR_LF:
4477              break;              break;
4478    
4479              case 0x000b:              case CHAR_VT:
4480              case 0x000c:              case CHAR_FF:
4481              case 0x0085:              case CHAR_NEL:
4482    #ifndef EBCDIC
4483              case 0x2028:              case 0x2028:
4484              case 0x2029:              case 0x2029:
4485    #endif  /* Not EBCDIC */
4486              if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);              if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);
4487              break;              break;
4488              }              }
# Line 4249  for (;;) Line 4500  for (;;)
4500            GETCHARINC(c, eptr);            GETCHARINC(c, eptr);
4501            switch(c)            switch(c)
4502              {              {
4503                HSPACE_CASES: RRETURN(MATCH_NOMATCH);  /* Byte and multibyte cases */
4504              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);  
4505              }              }
4506            }            }
4507          break;          break;
# Line 4285  for (;;) Line 4517  for (;;)
4517            GETCHARINC(c, eptr);            GETCHARINC(c, eptr);
4518            switch(c)            switch(c)
4519              {              {
4520                HSPACE_CASES: break;  /* Byte and multibyte cases */
4521              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;  
4522              }              }
4523            }            }
4524          break;          break;
# Line 4321  for (;;) Line 4534  for (;;)
4534            GETCHARINC(c, eptr);            GETCHARINC(c, eptr);
4535            switch(c)            switch(c)
4536              {              {
4537                VSPACE_CASES: RRETURN(MATCH_NOMATCH);
4538              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);  
4539              }              }
4540            }            }
4541          break;          break;
# Line 4345  for (;;) Line 4551  for (;;)
4551            GETCHARINC(c, eptr);            GETCHARINC(c, eptr);
4552            switch(c)            switch(c)
4553              {              {
4554                VSPACE_CASES: break;
4555              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;  
4556              }              }
4557            }            }
4558          break;          break;
# Line 4375  for (;;) Line 4574  for (;;)
4574          case OP_DIGIT:          case OP_DIGIT:
4575          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
4576            {            {
4577              pcre_uint32 cc;
4578            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
4579              {              {
4580              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4581              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4582              }              }
4583            if (*eptr >= 128 || (md->ctypes[*eptr] & ctype_digit) == 0)            cc = RAWUCHAR(eptr);
4584              if (cc >= 128 || (md->ctypes[cc] & ctype_digit) == 0)
4585              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4586            eptr++;            eptr++;
4587            /* 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 4591  for (;;)
4591          case OP_NOT_WHITESPACE:          case OP_NOT_WHITESPACE:
4592          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
4593            {            {
4594              pcre_uint32 cc;
4595            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
4596              {              {
4597              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4598              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4599              }              }
4600            if (*eptr < 128 && (md->ctypes[*eptr] & ctype_space) != 0)            cc = RAWUCHAR(eptr);
4601              if (cc < 128 && (md->ctypes[cc] & ctype_space) != 0)
4602              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4603            eptr++;            eptr++;
4604            ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);            ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);
# Line 4405  for (;;) Line 4608  for (;;)
4608          case OP_WHITESPACE:          case OP_WHITESPACE:
4609          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
4610            {            {
4611              pcre_uint32 cc;
4612            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
4613              {              {
4614              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4615              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4616              }              }
4617            if (*eptr >= 128 || (md->ctypes[*eptr] & ctype_space) == 0)            cc = RAWUCHAR(eptr);
4618              if (cc >= 128 || (md->ctypes[cc] & ctype_space) == 0)
4619              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4620            eptr++;            eptr++;
4621            /* 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 4625  for (;;)
4625          case OP_NOT_WORDCHAR:          case OP_NOT_WORDCHAR:
4626          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
4627            {            {
4628              pcre_uint32 cc;
4629            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
4630              {              {
4631              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4632              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4633              }              }
4634            if (*eptr < 128 && (md->ctypes[*eptr] & ctype_word) != 0)            cc = RAWUCHAR(eptr);
4635              if (cc < 128 && (md->ctypes[cc] & ctype_word) != 0)
4636              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4637            eptr++;            eptr++;
4638            ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);            ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);
# Line 4435  for (;;) Line 4642  for (;;)
4642          case OP_WORDCHAR:          case OP_WORDCHAR:
4643          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
4644            {            {
4645              pcre_uint32 cc;
4646            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
4647              {              {
4648              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4649              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4650              }              }
4651            if (*eptr >= 128 || (md->ctypes[*eptr] & ctype_word) == 0)            cc = RAWUCHAR(eptr);
4652              if (cc >= 128 || (md->ctypes[cc] & ctype_word) == 0)
4653              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4654            eptr++;            eptr++;
4655            /* 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 4677  for (;;)
4677              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4678              }              }
4679            if (IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH);            if (IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH);
4680              if (md->partial != 0 &&
4681                  eptr + 1 >= md->end_subject &&
4682                  NLBLOCK->nltype == NLTYPE_FIXED &&
4683                  NLBLOCK->nllen == 2 &&
4684                  *eptr == NLBLOCK->nl[0])
4685                {
4686                md->hitend = TRUE;
4687                if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
4688                }
4689            eptr++;            eptr++;
4690            }            }
4691          break;          break;
# Line 4502  for (;;) Line 4720  for (;;)
4720              {              {
4721              default: RRETURN(MATCH_NOMATCH);              default: RRETURN(MATCH_NOMATCH);
4722    
4723              case 0x000d:              case CHAR_CR:
4724              if (eptr < md->end_subject && *eptr == 0x0a) eptr++;              if (eptr < md->end_subject && *eptr == CHAR_LF) eptr++;
4725              break;              break;
4726    
4727              case 0x000a:              case CHAR_LF:
4728              break;              break;
4729    
4730              case 0x000b:              case CHAR_VT:
4731              case 0x000c:              case CHAR_FF:
4732              case 0x0085:              case CHAR_NEL:
4733  #ifdef COMPILE_PCRE16  #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
4734              case 0x2028:              case 0x2028:
4735              case 0x2029:              case 0x2029:
4736  #endif  #endif
# Line 4533  for (;;) Line 4751  for (;;)
4751            switch(*eptr++)            switch(*eptr++)
4752              {              {
4753              default: break;              default: break;
4754              case 0x09:      /* HT */              HSPACE_BYTE_CASES:
4755              case 0x20:      /* SPACE */  #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
4756              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 */  
4757  #endif  #endif
4758              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4759              }              }
# Line 4570  for (;;) Line 4771  for (;;)
4771            switch(*eptr++)            switch(*eptr++)
4772              {              {
4773              default: RRETURN(MATCH_NOMATCH);              default: RRETURN(MATCH_NOMATCH);
4774              case 0x09:      /* HT */              HSPACE_BYTE_CASES:
4775              case 0x20:      /* SPACE */  #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
4776              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 */  
4777  #endif  #endif
4778              break;              break;
4779              }              }
# Line 4606  for (;;) Line 4790  for (;;)
4790              }              }
4791            switch(*eptr++)            switch(*eptr++)
4792              {              {
4793              default: break;              VSPACE_BYTE_CASES:
4794              case 0x0a:      /* LF */  #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
4795              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 */  
4796  #endif  #endif
4797              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4798                default: break;
4799              }              }
4800            }            }
4801          break;          break;
# Line 4632  for (;;) Line 4811  for (;;)
4811            switch(*eptr++)            switch(*eptr++)
4812              {              {
4813              default: RRETURN(MATCH_NOMATCH);              default: RRETURN(MATCH_NOMATCH);
4814              case 0x0a:      /* LF */              VSPACE_BYTE_CASES:
4815              case 0x0b:      /* VT */  #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
4816              case 0x0c:      /* FF */              VSPACE_MULTIBYTE_CASES:
             case 0x0d:      /* CR */  
             case 0x85:      /* NEL */  
 #ifdef COMPILE_PCRE16  
             case 0x2028:    /* LINE SEPARATOR */  
             case 0x2029:    /* PARAGRAPH SEPARATOR */  
4817  #endif  #endif
4818              break;              break;
4819              }              }
# Line 4917  for (;;) Line 5091  for (;;)
5091              }              }
5092            /* Control never gets here */            /* Control never gets here */
5093    
5094            /* This should never occur */            case PT_CLIST:
5095              for (fi = min;; fi++)
5096                {
5097                const pcre_uint32 *cp;
5098                RMATCH(eptr, ecode, offset_top, md, eptrb, RM67);
5099                if (rrc != MATCH_NOMATCH) RRETURN(rrc);
5100                if (fi >= max) RRETURN(MATCH_NOMATCH);
5101                if (eptr >= md->end_subject)
5102                  {
5103                  SCHECK_PARTIAL();
5104                  RRETURN(MATCH_NOMATCH);
5105                  }
5106                GETCHARINCTEST(c, eptr);
5107                cp = PRIV(ucd_caseless_sets) + prop_value;
5108                for (;;)
5109                  {
5110                  if (c < *cp)
5111                    { if (prop_fail_result) break; else { RRETURN(MATCH_NOMATCH); } }
5112                  if (c == *cp++)
5113                    { if (prop_fail_result) { RRETURN(MATCH_NOMATCH); } else break; }
5114                  }
5115                }
5116              /* Control never gets here */
5117    
5118              case PT_UCNC:
5119              for (fi = min;; fi++)
5120                {
5121                RMATCH(eptr, ecode, offset_top, md, eptrb, RM68);
5122                if (rrc != MATCH_NOMATCH) RRETURN(rrc);
5123                if (fi >= max) RRETURN(MATCH_NOMATCH);
5124                if (eptr >= md->end_subject)
5125                  {
5126                  SCHECK_PARTIAL();
5127                  RRETURN(MATCH_NOMATCH);
5128                  }
5129                GETCHARINCTEST(c, eptr);
5130                if ((c == CHAR_DOLLAR_SIGN || c == CHAR_COMMERCIAL_AT ||
5131                     c == CHAR_GRAVE_ACCENT || (c >= 0xa0 && c <= 0xd7ff) ||
5132                     c >= 0xe000) == prop_fail_result)
5133                  RRETURN(MATCH_NOMATCH);
5134                }
5135              /* Control never gets here */
5136    
5137              /* This should never occur */
5138            default:            default:
5139            RRETURN(PCRE_ERROR_INTERNAL);            RRETURN(PCRE_ERROR_INTERNAL);
5140            }            }
# Line 4939  for (;;) Line 5155  for (;;)
5155              SCHECK_PARTIAL();              SCHECK_PARTIAL();
5156              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
5157              }              }
5158            GETCHARINCTEST(c, eptr);            else
           if (UCD_CATEGORY(c) == ucp_M) RRETURN(MATCH_NOMATCH);  
           while (eptr < md->end_subject)  
5159              {              {
5160              int len = 1;              int lgb, rgb;
5161              if (!utf) c = *eptr; else { GETCHARLEN(c, eptr, len); }              GETCHARINCTEST(c, eptr);
5162              if (UCD_CATEGORY(c) != ucp_M) break;              lgb = UCD_GRAPHBREAK(c);
5163              eptr += len;              while (eptr < md->end_subject)
5164                  {
5165                  int len = 1;
5166                  if (!utf) c = *eptr; else { GETCHARLEN(c, eptr, len); }
5167                  rgb = UCD_GRAPHBREAK(c);
5168                  if ((PRIV(ucp_gbtable)[lgb] & (1 << rgb)) == 0) break;
5169                  lgb = rgb;
5170                  eptr += len;
5171                  }
5172              }              }
5173              CHECK_PARTIAL();
5174            }            }
5175          }          }
5176        else        else
# Line 4971  for (;;) Line 5194  for (;;)
5194            GETCHARINC(c, eptr);            GETCHARINC(c, eptr);
5195            switch(ctype)            switch(ctype)
5196              {              {
5197              case OP_ANY:        /* This is the non-NL case */              case OP_ANY:               /* This is the non-NL case */
5198                if (md->partial != 0 &&    /* Take care with CRLF partial */
5199                    eptr >= md->end_subject &&
5200                    NLBLOCK->nltype == NLTYPE_FIXED &&
5201                    NLBLOCK->nllen == 2 &&
5202                    c == NLBLOCK->nl[0])
5203                  {
5204                  md->hitend = TRUE;
5205                  if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
5206                  }
5207                break;
5208    
5209              case OP_ALLANY:              case OP_ALLANY:
5210              case OP_ANYBYTE:              case OP_ANYBYTE:
5211              break;              break;
# Line 4980  for (;;) Line 5214  for (;;)
5214              switch(c)              switch(c)
5215                {                {
5216                default: RRETURN(MATCH_NOMATCH);                default: RRETURN(MATCH_NOMATCH);
5217                case 0x000d:                case CHAR_CR:
5218                if (eptr < md->end_subject && *eptr == 0x0a) eptr++;                if (eptr < md->end_subject && RAWUCHAR(eptr) == CHAR_LF) eptr++;
5219                break;                break;
5220                case 0x000a:  
5221                  case CHAR_LF:
5222                break;                break;
5223    
5224                case 0x000b:                case CHAR_VT:
5225                case 0x000c:                case CHAR_FF:
5226                case 0x0085:                case CHAR_NEL:
5227    #ifndef EBCDIC
5228                case 0x2028:                case 0x2028:
5229                case 0x2029:                case 0x2029:
5230    #endif  /* Not EBCDIC */
5231                if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);                if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);
5232                break;                break;
5233                }                }
# Line 4999  for (;;) Line 5236  for (;;)
5236              case OP_NOT_HSPACE:              case OP_NOT_HSPACE:
5237              switch(c)              switch(c)
5238                {                {
5239                  HSPACE_CASES: RRETURN(MATCH_NOMATCH);
5240                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);  
5241                }                }
5242              break;              break;
5243    
5244              case OP_HSPACE:              case OP_HSPACE:
5245              switch(c)              switch(c)
5246                {                {
5247                  HSPACE_CASES: break;
5248                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;  
5249                }                }
5250              break;              break;
5251    
5252              case OP_NOT_VSPACE:              case OP_NOT_VSPACE:
5253              switch(c)              switch(c)
5254                {                {
5255                  VSPACE_CASES: RRETURN(MATCH_NOMATCH);
5256                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);  
5257                }                }
5258              break;              break;
5259    
5260              case OP_VSPACE:              case OP_VSPACE:
5261              switch(c)              switch(c)
5262                {                {
5263                  VSPACE_CASES: break;
5264                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;  
5265                }                }
5266              break;              break;
5267    
# Line 5134  for (;;) Line 5319  for (;;)
5319            c = *eptr++;            c = *eptr++;
5320            switch(ctype)            switch(ctype)
5321              {              {
5322              case OP_ANY:     /* This is the non-NL case */              case OP_ANY:               /* This is the non-NL case */
5323                if (md->partial != 0 &&    /* Take care with CRLF partial */
5324                    eptr >= md->end_subject &&
5325                    NLBLOCK->nltype == NLTYPE_FIXED &&
5326                    NLBLOCK->nllen == 2 &&
5327                    c == NLBLOCK->nl[0])
5328                  {
5329                  md->hitend = TRUE;
5330                  if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
5331                  }
5332                break;
5333    
5334              case OP_ALLANY:              case OP_ALLANY:
5335              case OP_ANYBYTE:              case OP_ANYBYTE:
5336              break;              break;
# Line 5143  for (;;) Line 5339  for (;;)
5339              switch(c)              switch(c)
5340                {                {
5341                default: RRETURN(MATCH_NOMATCH);                default: RRETURN(MATCH_NOMATCH);
5342                case 0x000d:                case CHAR_CR:
5343                if (eptr < md->end_subject && *eptr == 0x0a) eptr++;                if (eptr < md->end_subject && *eptr == CHAR_LF) eptr++;
5344                break;                break;
5345    
5346                case 0x000a:                case CHAR_LF:
5347                break;                break;
5348    
5349                case 0x000b:                case CHAR_VT:
5350                case 0x000c:                case CHAR_FF:
5351                case 0x0085:                case CHAR_NEL:
5352  #ifdef COMPILE_PCRE16  #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
5353                case 0x2028:                case 0x2028:
5354                case 0x2029:                case 0x2029:
5355  #endif  #endif
# Line 5166  for (;;) Line 5362  for (;;)
5362              switch(c)              switch(c)
5363                {                {
5364                default: break;                default: break;
5365                case 0x09:      /* HT */                HSPACE_BYTE_CASES:
5366                case 0x20:      /* SPACE */  #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
5367                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 */  
5368  #endif  #endif
5369                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
5370                }                }
# Line 5195  for (;;) Line 5374  for (;;)
5374              switch(c)              switch(c)
5375                {                {
5376                default: RRETURN(MATCH_NOMATCH);                default: RRETURN(MATCH_NOMATCH);
5377                case 0x09:      /* HT */                HSPACE_BYTE_CASES:
5378                case 0x20:      /* SPACE */  #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
5379                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 */  
5380  #endif  #endif
5381                break;                break;
5382                }                }
# Line 5224  for (;;) Line 5386  for (;;)
5386              switch(c)              switch(c)
5387                {                {
5388                default: break;                default: break;
5389                case 0x0a:      /* LF */                VSPACE_BYTE_CASES:
5390                case 0x0b:      /* VT */  #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
5391                case 0x0c:      /* FF */                VSPACE_MULTIBYTE_CASES:
               case 0x0d:      /* CR */  
               case 0x85:      /* NEL */  
 #ifdef COMPILE_PCRE16  
               case 0x2028:    /* LINE SEPARATOR */  
               case 0x2029:    /* PARAGRAPH SEPARATOR */  
5392  #endif  #endif
5393                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
5394                }                }
# Line 5241  for (;;) Line 5398  for (;;)
5398              switch(c)              switch(c)
5399                {                {
5400                default: RRETURN(MATCH_NOMATCH);                default: RRETURN(MATCH_NOMATCH);
5401                case 0x0a:      /* LF */                VSPACE_BYTE_CASES:
5402                case 0x0b:      /* VT */  #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
5403                case 0x0c:      /* FF */                VSPACE_MULTIBYTE_CASES:
               case 0x0d:      /* CR */  
               case 0x85:      /* NEL */  
 #ifdef COMPILE_PCRE16  
               case 0x2028:    /* LINE SEPARATOR */  
               case 0x2029:    /* PARAGRAPH SEPARATOR */  
5404  #endif  #endif
5405                break;                break;
5406                }                }
# Line 5452  for (;;) Line 5604  for (;;)
5604              }              }
5605            break;            break;
5606    
5607              case PT_CLIST:
5608              for (i = min; i < max; i++)
5609                {
5610                const pcre_uint32 *cp;
5611                int len = 1;
5612                if (eptr >= md->end_subject)
5613                  {
5614                  SCHECK_PARTIAL();
5615                  break;
5616                  }
5617                GETCHARLENTEST(c, eptr, len);
5618                cp = PRIV(ucd_caseless_sets) + prop_value;
5619                for (;;)
5620                  {
5621                  if (c < *cp)
5622                    { if (prop_fail_result) break; else goto GOT_MAX; }
5623                  if (c == *cp++)
5624                    { if (prop_fail_result) goto GOT_MAX; else break; }
5625                  }
5626                eptr += len;
5627                }
5628              GOT_MAX:
5629              break;
5630    
5631              case PT_UCNC:
5632              for (i = min; i < max; i++)
5633                {
5634                int len = 1;
5635                if (eptr >= md->end_subject)
5636                  {
5637                  SCHECK_PARTIAL();
5638                  break;
5639                  }
5640                GETCHARLENTEST(c, eptr, len);
5641                if ((c == CHAR_DOLLAR_SIGN || c == CHAR_COMMERCIAL_AT ||
5642                     c == CHAR_GRAVE_ACCENT || (c >= 0xa0 && c <= 0xd7ff) ||
5643                     c >= 0xe000) == prop_fail_result)
5644                  break;
5645                eptr += len;
5646                }
5647              break;
5648    
5649            default:            default:
5650            RRETURN(PCRE_ERROR_INTERNAL);            RRETURN(PCRE_ERROR_INTERNAL);
5651            }            }
5652    
5653          /* eptr is now past the end of the maximum run */          /* eptr is now past the end of the maximum run */
5654    
5655          if (possessive) continue;          if (possessive) continue;    /* No backtracking */
5656          for(;;)          for(;;)
5657            {            {
5658              if (eptr == pp) goto TAIL_RECURSE;
5659            RMATCH(eptr, ecode, offset_top, md, eptrb, RM44);            RMATCH(eptr, ecode, offset_top, md, eptrb, RM44);
5660            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
5661            if (eptr-- == pp) break;        /* Stop if tried at original pos */            eptr--;
5662            if (utf) BACKCHAR(eptr);            if (utf) BACKCHAR(eptr);
5663            }            }
5664          }          }
5665    
5666        /* Match extended Unicode sequences. We will get here only if the        /* Match extended Unicode grapheme clusters. We will get here only if the
5667        support is in the binary; otherwise a compile-time error occurs. */        support is in the binary; otherwise a compile-time error occurs. */
5668    
5669        else if (ctype == OP_EXTUNI)        else if (ctype == OP_EXTUNI)
5670          {          {
5671          for (i = min; i < max; i++)          for (i = min; i < max; i++)
5672            {            {
           int len = 1;  
5673            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
5674              {              {
5675              SCHECK_PARTIAL();              SCHECK_PARTIAL();
5676              break;              break;
5677              }              }
5678            if (!utf) c = *eptr; else { GETCHARLEN(c, eptr, len); }            else
           if (UCD_CATEGORY(c) == ucp_M) break;  
           eptr += len;  
           while (eptr < md->end_subject)  
5679              {              {
5680              len = 1;              int lgb, rgb;
5681              if (!utf) c = *eptr; else { GETCHARLEN(c, eptr, len); }              GETCHARINCTEST(c, eptr);
5682              if (UCD_CATEGORY(c) != ucp_M) break;              lgb = UCD_GRAPHBREAK(c);
5683              eptr += len;              while (eptr < md->end_subject)
5684                  {
5685                  int len = 1;
5686                  if (!utf) c = *eptr; else { GETCHARLEN(c, eptr, len); }
5687                  rgb = UCD_GRAPHBREAK(c);
5688                  if ((PRIV(ucp_gbtable)[lgb] & (1 << rgb)) == 0) break;
5689                  lgb = rgb;
5690                  eptr += len;
5691                  }
5692              }              }
5693              CHECK_PARTIAL();
5694            }            }
5695    
5696          /* eptr is now past the end of the maximum run */          /* eptr is now past the end of the maximum run */
5697    
5698          if (possessive) continue;          if (possessive) continue;    /* No backtracking */
5699    
5700          for(;;)          for(;;)
5701            {            {
5702              int lgb, rgb;
5703              PCRE_PUCHAR fptr;
5704    
5705              if (eptr == pp) goto TAIL_RECURSE;   /* At start of char run */
5706            RMATCH(eptr, ecode, offset_top, md, eptrb, RM45);            RMATCH(eptr, ecode, offset_top, md, eptrb, RM45);
5707            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
5708            if (eptr-- == pp) break;        /* Stop if tried at original pos */  
5709            for (;;)                        /* Move back over one extended */            /* Backtracking over an extended grapheme cluster involves inspecting
5710              the previous two characters (if present) to see if a break is
5711              permitted between them. */
5712    
5713              eptr--;
5714              if (!utf) c = *eptr; else
5715              {              {
5716              if (!utf) c = *eptr; else              BACKCHAR(eptr);
5717                {              GETCHAR(c, eptr);
5718                BACKCHAR(eptr);              }
5719                GETCHAR(c, eptr);            rgb = UCD_GRAPHBREAK(c);
5720                }  
5721              if (UCD_CATEGORY(c) != ucp_M) break;            for (;;)
5722              eptr--;              {
5723                if (eptr == pp) goto TAIL_RECURSE;   /* At start of char run */
5724                fptr = eptr - 1;
5725                if (!utf) c = *fptr; else
5726                  {
5727                  BACKCHAR(fptr);
5728                  GETCHAR(c, fptr);
5729                  }
5730                lgb = UCD_GRAPHBREAK(c);
5731                if ((PRIV(ucp_gbtable)[lgb] & (1 << rgb)) == 0) break;
5732                eptr = fptr;
5733                rgb = lgb;
5734              }              }
5735            }            }
5736          }          }
# Line 5534  for (;;) Line 5754  for (;;)
5754                  break;                  break;
5755                  }                  }
5756                if (IS_NEWLINE(eptr)) break;                if (IS_NEWLINE(eptr)) break;
5757                  if (md->partial != 0 &&    /* Take care with CRLF partial */
5758                      eptr + 1 >= md->end_subject &&
5759                      NLBLOCK->nltype == NLTYPE_FIXED &&
5760                      NLBLOCK->nllen == 2 &&
5761                      RAWUCHAR(eptr) == NLBLOCK->nl[0])
5762                    {
5763                    md->hitend = TRUE;
5764                    if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
5765                    }
5766                eptr++;                eptr++;
5767                ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);                ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);
5768                }                }
# Line 5551  for (;;) Line 5780  for (;;)
5780                  break;                  break;
5781                  }                  }
5782                if (IS_NEWLINE(eptr)) break;                if (IS_NEWLINE(eptr)) break;
5783                  if (md->partial != 0 &&    /* Take care with CRLF partial */
5784                      eptr + 1 >= md->end_subject &&
5785                      NLBLOCK->nltype == NLTYPE_FIXED &&
5786                      NLBLOCK->nllen == 2 &&
5787                      RAWUCHAR(eptr) == NLBLOCK->nl[0])
5788                    {
5789                    md->hitend = TRUE;
5790                    if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
5791                    }
5792                eptr++;                eptr++;
5793                ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);                ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);
5794                }                }
# Line 5600  for (;;) Line 5838  for (;;)
5838                break;                break;
5839                }                }
5840              GETCHARLEN(c, eptr, len);              GETCHARLEN(c, eptr, len);
5841              if (c == 0x000d)              if (c == CHAR_CR)
5842                {                {
5843                if (++eptr >= md->end_subject) break;                if (++eptr >= md->end_subject) break;
5844                if (*eptr == 0x000a) eptr++;                if (RAWUCHAR(eptr) == CHAR_LF) eptr++;
5845                }                }
5846              else              else
5847                {                {
5848                if (c != 0x000a &&                if (c != CHAR_LF &&
5849                    (md->bsr_anycrlf ||                    (md->bsr_anycrlf ||
5850                     (c != 0x000b && c != 0x000c &&                     (c != CHAR_VT && c != CHAR_FF && c != CHAR_NEL
5851                      c != 0x0085 && c != 0x2028 && c != 0x2029)))  #ifndef EBCDIC
5852                        && c != 0x2028 && c != 0x2029
5853    #endif  /* Not EBCDIC */
5854                        )))
5855                  break;                  break;
5856                eptr += len;                eptr += len;
5857                }                }
# Line 5631  for (;;) Line 5872  for (;;)
5872              GETCHARLEN(c, eptr, len);              GETCHARLEN(c, eptr, len);
5873              switch(c)              switch(c)
5874                {                {
5875                  HSPACE_CASES: gotspace = TRUE; break;
5876                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;  
5877                }                }
5878              if (gotspace == (ctype == OP_NOT_HSPACE)) break;              if (gotspace == (ctype == OP_NOT_HSPACE)) break;
5879              eptr += len;              eptr += len;
# Line 5673  for (;;) Line 5894  for (;;)
5894              GETCHARLEN(c, eptr, len);              GETCHARLEN(c, eptr, len);
5895              switch(c)              switch(c)
5896                {                {
5897                  VSPACE_CASES: gotspace = TRUE; break;
5898                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;  
5899                }                }
5900              if (gotspace == (ctype == OP_NOT_VSPACE)) break;              if (gotspace == (ctype == OP_NOT_VSPACE)) break;
5901              eptr += len;              eptr += len;
# Line 5783  for (;;) Line 5996  for (;;)
5996            RRETURN(PCRE_ERROR_INTERNAL);            RRETURN(PCRE_ERROR_INTERNAL);
5997            }            }
5998    
5999          /* 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;  
6000          for(;;)          for(;;)
6001            {            {
6002              if (eptr == pp) goto TAIL_RECURSE;
6003            RMATCH(eptr, ecode, offset_top, md, eptrb, RM46);            RMATCH(eptr, ecode, offset_top, md, eptrb, RM46);
6004            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
6005            if (eptr-- == pp) break;        /* Stop if tried at original pos */            eptr--;
6006            BACKCHAR(eptr);            BACKCHAR(eptr);
6007            if (ctype == OP_ANYNL && eptr > pp  && *eptr == '\n' &&            if (ctype == OP_ANYNL && eptr > pp  && RAWUCHAR(eptr) == CHAR_NL &&
6008                eptr[-1] == '\r') eptr--;                RAWUCHAR(eptr - 1) == CHAR_CR) eptr--;
6009            }            }
6010          }          }
6011        else        else
# Line 5815  for (;;) Line 6023  for (;;)
6023                break;                break;
6024                }                }
6025              if (IS_NEWLINE(eptr)) break;              if (IS_NEWLINE(eptr)) break;
6026                if (md->partial != 0 &&    /* Take care with CRLF partial */
6027                    eptr + 1 >= md->end_subject &&
6028                    NLBLOCK->nltype == NLTYPE_FIXED &&
6029                    NLBLOCK->nllen == 2 &&
6030                    *eptr == NLBLOCK->nl[0])
6031                  {
6032                  md->hitend = TRUE;
6033                  if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
6034                  }
6035              eptr++;              eptr++;
6036              }              }
6037            break;            break;
# Line 5839  for (;;) Line 6056  for (;;)
6056                break;                break;
6057                }                }
6058              c = *eptr;              c = *eptr;
6059              if (c == 0x000d)              if (c == CHAR_CR)
6060                {                {
6061                if (++eptr >= md->end_subject) break;                if (++eptr >= md->end_subject) break;
6062                if (*eptr == 0x000a) eptr++;                if (*eptr == CHAR_LF) eptr++;
6063                }                }
6064              else              else
6065                {                {
6066                if (c != 0x000a && (md->bsr_anycrlf ||                if (c != CHAR_LF && (md->bsr_anycrlf ||
6067                  (c != 0x000b && c != 0x000c && c != 0x0085                   (c != CHAR_VT && c != CHAR_FF && c != CHAR_NEL
6068  #ifdef COMPILE_PCRE16  #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
6069                  && c != 0x2028 && c != 0x2029                   && c != 0x2028 && c != 0x2029
6070  #endif  #endif
6071                  ))) break;                   ))) break;
6072                eptr++;                eptr++;
6073                }                }
6074              }              }
# Line 5865  for (;;) Line 6082  for (;;)
6082                SCHECK_PARTIAL();                SCHECK_PARTIAL();
6083                break;                break;
6084                }                }
6085              c = *eptr;              switch(*eptr)
6086              if (c == 0x09 || c == 0x20 || c == 0xa0                {
6087  #ifdef COMPILE_PCRE16                default: eptr++; break;
6088                || c == 0x1680 || c == 0x180e || (c >= 0x2000 && c <= 0x200A)                HSPACE_BYTE_CASES:
6089                || c == 0x202f || c == 0x205f || c == 0x3000  #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
6090                  HSPACE_MULTIBYTE_CASES:
6091  #endif  #endif
6092                ) break;                goto ENDLOOP00;
6093              eptr++;                }
6094              }              }
6095              ENDLOOP00:
6096            break;            break;
6097    
6098            case OP_HSPACE:            case OP_HSPACE:
# Line 5884  for (;;) Line 6103  for (;;)
6103                SCHECK_PARTIAL();                SCHECK_PARTIAL();
6104                break;                break;
6105                }                }
6106              c = *eptr;              switch(*eptr)
6107              if (c != 0x09 && c != 0x20 && c != 0xa0                {
6108  #ifdef COMPILE_PCRE16                default: goto ENDLOOP01;
6109                && c != 0x1680 && c != 0x180e && (c < 0x2000 || c > 0x200A)                HSPACE_BYTE_CASES:
6110                && c != 0x202f && c != 0x205f && c != 0x3000  #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
6111                  HSPACE_MULTIBYTE_CASES:
6112  #endif  #endif
6113                ) break;                eptr++; break;
6114              eptr++;                }
6115              }              }
6116              ENDLOOP01:
6117            break;            break;
6118    
6119            case OP_NOT_VSPACE:            case OP_NOT_VSPACE:
# Line 5903  for (;;) Line 6124  for (;;)
6124                SCHECK_PARTIAL();                SCHECK_PARTIAL();
6125                break;                break;
6126                }                }
6127              c = *eptr;              switch(*eptr)
6128              if (c == 0x0a || c == 0x0b || c == 0x0c || c == 0x0d || c == 0x85                {
6129  #ifdef COMPILE_PCRE16                default: eptr++; break;
6130                || c == 0x2028 || c == 0x2029                VSPACE_BYTE_CASES:
6131    #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
6132                  VSPACE_MULTIBYTE_CASES:
6133  #endif  #endif
6134                ) break;                goto ENDLOOP02;
6135              eptr++;                }
6136              }              }
6137              ENDLOOP02:
6138            break;            break;
6139    
6140            case OP_VSPACE:            case OP_VSPACE:
# Line 5921  for (;;) Line 6145  for (;;)
6145                SCHECK_PARTIAL();                SCHECK_PARTIAL();
6146                break;                break;
6147                }                }
6148              c = *eptr;              switch(*eptr)
6149              if (c != 0x0a && c != 0x0b && c != 0x0c && c != 0x0d && c != 0x85                {
6150  #ifdef COMPILE_PCRE16                default: goto ENDLOOP03;
6151                && c != 0x2028 && c != 0x2029                VSPACE_BYTE_CASES:
6152    #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
6153                  VSPACE_MULTIBYTE_CASES:
6154  #endif  #endif
6155                ) break;                eptr++; break;
6156              eptr++;                }
6157              }              }
6158              ENDLOOP03:
6159            break;            break;
6160    
6161            case OP_NOT_DIGIT:            case OP_NOT_DIGIT:
# Line 6013  for (;;) Line 6240  for (;;)
6240            RRETURN(PCRE_ERROR_INTERNAL);            RRETURN(PCRE_ERROR_INTERNAL);
6241            }            }
6242    
6243          /* eptr is now past the end of the maximum run. If possessive, we are          if (possessive) continue;    /* No backtracking */
6244          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)  
6245            {            {
6246              if (eptr == pp) goto TAIL_RECURSE;
6247            RMATCH(eptr, ecode, offset_top, md, eptrb, RM47);            RMATCH(eptr, ecode, offset_top, md, eptrb, RM47);
6248            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
6249            eptr--;            eptr--;
6250            if (ctype == OP_ANYNL && eptr > pp  && *eptr == '\n' &&            if (ctype == OP_ANYNL && eptr > pp  && *eptr == CHAR_LF &&
6251                eptr[-1] == '\r') eptr--;                eptr[-1] == CHAR_CR) eptr--;
6252            }            }
6253          }          }
6254    
6255        /* Get here if we can't make it match with any permitted repetitions */        /* Control never gets here */
   
       RRETURN(MATCH_NOMATCH);  
6256        }        }
     /* Control never gets here */  
6257    
6258      /* 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
6259      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 6076  switch (frame->Xwhere) Line 6295  switch (frame->Xwhere)
6295    LBL(32) LBL(34) LBL(42) LBL(46)    LBL(32) LBL(34) LBL(42) LBL(46)
6296  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
6297    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)
6298    LBL(59) LBL(60) LBL(61) LBL(62)    LBL(59) LBL(60) LBL(61) LBL(62) LBL(67) LBL(68)
6299  #endif  /* SUPPORT_UCP */  #endif  /* SUPPORT_UCP */
6300  #endif  /* SUPPORT_UTF */  #endif  /* SUPPORT_UTF */
6301    default:    default:
6302    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);  
   
6303    return PCRE_ERROR_INTERNAL;    return PCRE_ERROR_INTERNAL;
6304    }    }
6305  #undef LBL  #undef LBL
# Line 6145  Undefine all the macros that were define Line 6361  Undefine all the macros that were define
6361  ***************************************************************************/  ***************************************************************************/
6362    
6363    
6364    #ifdef NO_RECURSE
6365    /*************************************************
6366    *          Release allocated heap frames         *
6367    *************************************************/
6368    
6369    /* This function releases all the allocated frames. The base frame is on the
6370    machine stack, and so must not be freed.
6371    
6372    Argument: the address of the base frame
6373    Returns:  nothing
6374    */
6375    
6376    static void
6377    release_match_heapframes (heapframe *frame_base)
6378    {
6379    heapframe *nextframe = frame_base->Xnextframe;
6380    while (nextframe != NULL)
6381      {
6382      heapframe *oldframe = nextframe;
6383      nextframe = nextframe->Xnextframe;
6384      (PUBL(stack_free))(oldframe);
6385      }
6386    }
6387    #endif
6388    
6389    
6390  /*************************************************  /*************************************************
6391  *         Execute a Regular Expression           *  *         Execute a Regular Expression           *
# Line 6170  Returns:          > 0 => success; value Line 6411  Returns:          > 0 => success; value
6411                   < -1 => some kind of unexpected problem                   < -1 => some kind of unexpected problem
6412  */  */
6413    
6414  #ifdef COMPILE_PCRE8  #if defined COMPILE_PCRE8
6415  PCRE_EXP_DEFN int PCRE_CALL_CONVENTION  PCRE_EXP_DEFN int PCRE_CALL_CONVENTION
6416  pcre_exec(const pcre *argument_re, const pcre_extra *extra_data,  pcre_exec(const pcre *argument_re, const pcre_extra *extra_data,
6417    PCRE_SPTR subject, int length, int start_offset, int options, int *offsets,    PCRE_SPTR subject, int length, int start_offset, int options, int *offsets,
6418    int offsetcount)    int offsetcount)
6419  #else  #elif defined COMPILE_PCRE16
6420  PCRE_EXP_DEFN int PCRE_CALL_CONVENTION  PCRE_EXP_DEFN int PCRE_CALL_CONVENTION
6421  pcre16_exec(const pcre16 *argument_re, const pcre16_extra *extra_data,  pcre16_exec(const pcre16 *argument_re, const pcre16_extra *extra_data,
6422    PCRE_SPTR16 subject, int length, int start_offset, int options, int *offsets,    PCRE_SPTR16 subject, int length, int start_offset, int options, int *offsets,
6423    int offsetcount)    int offsetcount)
6424    #elif defined COMPILE_PCRE32
6425    PCRE_EXP_DEFN int PCRE_CALL_CONVENTION
6426    pcre32_exec(const pcre32 *argument_re, const pcre32_extra *extra_data,
6427      PCRE_SPTR32 subject, int length, int start_offset, int options, int *offsets,
6428      int offsetcount)
6429  #endif  #endif
6430  {  {
6431  int rc, ocount, arg_offset_max;  int rc, ocount, arg_offset_max;
# Line 6202  const pcre_uint8 *start_bits = NULL; Line 6448  const pcre_uint8 *start_bits = NULL;
6448  PCRE_PUCHAR start_match = (PCRE_PUCHAR)subject + start_offset;  PCRE_PUCHAR start_match = (PCRE_PUCHAR)subject + start_offset;
6449  PCRE_PUCHAR end_subject;  PCRE_PUCHAR end_subject;
6450  PCRE_PUCHAR start_partial = NULL;  PCRE_PUCHAR start_partial = NULL;
6451    PCRE_PUCHAR match_partial = NULL;
6452  PCRE_PUCHAR req_char_ptr = start_match - 1;  PCRE_PUCHAR req_char_ptr = start_match - 1;
6453    
6454  const pcre_study_data *study;  const pcre_study_data *study;
6455  const REAL_PCRE *re = (const REAL_PCRE *)argument_re;  const REAL_PCRE *re = (const REAL_PCRE *)argument_re;
6456    
6457  /* Check for the special magic call that measures the size of the stack used  #ifdef NO_RECURSE
6458  per recursive call of match(). */  heapframe frame_zero;
6459    frame_zero.Xprevframe = NULL;            /* Marks the top level */
6460    frame_zero.Xnextframe = NULL;            /* None are allocated yet */
6461    md->match_frames_base = &frame_zero;
6462    #endif
6463    
6464    /* Check for the special magic call that measures the size of the stack used
6465    per recursive call of match(). Without the funny casting for sizeof, a Windows
6466    compiler gave this error: "unary minus operator applied to unsigned type,
6467    result still unsigned". Hopefully the cast fixes that. */
6468    
6469  if (re == NULL && extra_data == NULL && subject == NULL && length == -1)  if (re == NULL && extra_data == NULL && subject == NULL && length == -999 &&
6470        start_offset == -999)
6471  #ifdef NO_RECURSE  #ifdef NO_RECURSE
6472    return -sizeof(heapframe);    return -((int)sizeof(heapframe));
6473  #else  #else
6474    return match(NULL, NULL, NULL, 0, NULL, NULL, 0);    return match(NULL, NULL, NULL, 0, NULL, NULL, 0);
6475  #endif  #endif
6476    
6477  /* Plausibility checks */  /* Plausibility checks */
6478    
6479  if ((options & ~PUBLIC_EXEC_OPTIONS) != 0) return PCRE_ERROR_BADOPTION;  if ((options & ~PUBLIC_EXEC_OPTIONS) != 0) return PCRE_ERROR_BADOPTION;
6480  if (re == NULL || subject == NULL || (offsets == NULL && offsetcount > 0))  if (re == NULL || subject == NULL || (offsets == NULL && offsetcount > 0))
6481    return PCRE_ERROR_NULL;    return PCRE_ERROR_NULL;
6482  if (offsetcount < 0) return PCRE_ERROR_BADCOUNT;  if (offsetcount < 0) return PCRE_ERROR_BADCOUNT;
6483    if (length < 0) return PCRE_ERROR_BADLENGTH;
6484  if (start_offset < 0 || start_offset > length) return PCRE_ERROR_BADOFFSET;  if (start_offset < 0 || start_offset > length) return PCRE_ERROR_BADOFFSET;
6485    
6486  /* 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 6518  if (utf && (options & PCRE_NO_UTF8_CHECK
6518        offsets[0] = erroroffset;        offsets[0] = erroroffset;
6519        offsets[1] = errorcode;        offsets[1] = errorcode;
6520        }        }
6521  #ifdef COMPILE_PCRE16  #if defined COMPILE_PCRE8
     return (errorcode <= PCRE_UTF16_ERR1 && md->partial > 1)?  
       PCRE_ERROR_SHORTUTF16 : PCRE_ERROR_BADUTF16;  
 #else  
6522      return (errorcode <= PCRE_UTF8_ERR5 && md->partial > 1)?      return (errorcode <= PCRE_UTF8_ERR5 && md->partial > 1)?
6523        PCRE_ERROR_SHORTUTF8 : PCRE_ERROR_BADUTF8;        PCRE_ERROR_SHORTUTF8 : PCRE_ERROR_BADUTF8;
6524    #elif defined COMPILE_PCRE16
6525        return (errorcode <= PCRE_UTF16_ERR1 && md->partial > 1)?
6526          PCRE_ERROR_SHORTUTF16 : PCRE_ERROR_BADUTF16;
6527    #elif defined COMPILE_PCRE32
6528        return PCRE_ERROR_BADUTF32;
6529  #endif  #endif
6530      }      }
6531    #if defined COMPILE_PCRE8 || defined COMPILE_PCRE16
6532    /* 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. */
6533    if (start_offset > 0 && start_offset < length &&    if (start_offset > 0 && start_offset < length &&
6534        NOT_FIRSTCHAR(((PCRE_PUCHAR)subject)[start_offset]))        NOT_FIRSTCHAR(((PCRE_PUCHAR)subject)[start_offset]))
6535      return PCRE_ERROR_BADUTF8_OFFSET;      return PCRE_ERROR_BADUTF8_OFFSET;
6536    #endif
6537    }    }
6538  #endif  #endif
6539    
6540  /* If the pattern was successfully studied with JIT support, run the JIT  /* If the pattern was successfully studied with JIT support, run the JIT
6541  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
6542  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
6543  an unsupported flag is set. In particular, JIT does not support partial  an unsupported flag is set. */
 matching. */  
6544    
6545  #ifdef SUPPORT_JIT  #ifdef SUPPORT_JIT
6546  if (extra_data != NULL  if (extra_data != NULL
6547      && (extra_data->flags & PCRE_EXTRA_EXECUTABLE_JIT) != 0      && (extra_data->flags & (PCRE_EXTRA_EXECUTABLE_JIT |
6548                                 PCRE_EXTRA_TABLES)) == PCRE_EXTRA_EXECUTABLE_JIT
6549      && extra_data->executable_jit != NULL      && extra_data->executable_jit != NULL
6550      && (extra_data->flags & PCRE_EXTRA_TABLES) == 0      && (options & ~PUBLIC_JIT_EXEC_OPTIONS) == 0)
6551      && (options & ~(PCRE_NO_UTF8_CHECK | PCRE_NOTBOL | PCRE_NOTEOL |    {
6552                      PCRE_NOTEMPTY | PCRE_NOTEMPTY_ATSTART)) == 0)    rc = PRIV(jit_exec)(extra_data, (const pcre_uchar *)subject, length,
6553    return PRIV(jit_exec)(re, extra_data->executable_jit,         start_offset, options, offsets, offsetcount);
6554      (const pcre_uchar *)subject, length, start_offset, options,  
6555      ((extra_data->flags & PCRE_EXTRA_MATCH_LIMIT) == 0)    /* PCRE_ERROR_NULL means that the selected normal or partial matching
6556      ? MATCH_LIMIT : extra_data->match_limit, offsets, offsetcount);    mode is not compiled. In this case we simply fallback to interpreter. */
6557    
6558      if (rc != PCRE_ERROR_JIT_BADOPTION) return rc;
6559      }
6560  #endif  #endif
6561    
6562  /* 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 6578  md->callout_data = NULL;
6578    
6579  tables = re->tables;  tables = re->tables;
6580    
6581    /* The two limit values override the defaults, whatever their value. */
6582    
6583  if (extra_data != NULL)  if (extra_data != NULL)
6584    {    {
6585    register unsigned int flags = extra_data->flags;    register unsigned int flags = extra_data->flags;
# Line 6328  if (extra_data != NULL) Line 6594  if (extra_data != NULL)
6594    if ((flags & PCRE_EXTRA_TABLES) != 0) tables = extra_data->tables;    if ((flags & PCRE_EXTRA_TABLES) != 0) tables = extra_data->tables;
6595    }    }
6596    
6597    /* Limits in the regex override only if they are smaller. */
6598    
6599    if ((re->flags & PCRE_MLSET) != 0 && re->limit_match < md->match_limit)
6600      md->match_limit = re->limit_match;
6601    
6602    if ((re->flags & PCRE_RLSET) != 0 &&
6603        re->limit_recursion < md->match_limit_recursion)
6604      md->match_limit_recursion = re->limit_recursion;
6605    
6606  /* 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
6607  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
6608  in other programs later. */  in other programs later. */
# Line 6353  end_subject = md->end_subject; Line 6628  end_subject = md->end_subject;
6628  md->endonly = (re->options & PCRE_DOLLAR_ENDONLY) != 0;  md->endonly = (re->options & PCRE_DOLLAR_ENDONLY) != 0;
6629  md->use_ucp = (re->options & PCRE_UCP) != 0;  md->use_ucp = (re->options & PCRE_UCP) != 0;
6630  md->jscript_compat = (re->options & PCRE_JAVASCRIPT_COMPAT) != 0;  md->jscript_compat = (re->options & PCRE_JAVASCRIPT_COMPAT) != 0;
6631  md->ignore_skip_arg = FALSE;  md->ignore_skip_arg = 0;
6632    
6633  /* Some options are unpacked into BOOL variables in the hope that testing  /* Some options are unpacked into BOOL variables in the hope that testing
6634  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 6738  if (re->top_backref > 0 && re->top_backr
6738    DPRINTF(("Got memory to hold back references\n"));    DPRINTF(("Got memory to hold back references\n"));
6739    }    }
6740  else md->offset_vector = offsets;  else md->offset_vector = offsets;
   
6741  md->offset_end = ocount;  md->offset_end = ocount;
6742  md->offset_max = (2*ocount)/3;  md->offset_max = (2*ocount)/3;
6743  md->offset_overflow = FALSE;  md->capture_last = 0;
 md->capture_last = -1;  
6744    
6745  /* Reset the working variable associated with each extraction. These should  /* Reset the working variable associated with each extraction. These should
6746  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 6768  if (!anchored)
6768    if ((re->flags & PCRE_FIRSTSET) != 0)    if ((re->flags & PCRE_FIRSTSET) != 0)
6769      {      {
6770      has_first_char = TRUE;      has_first_char = TRUE;
6771      first_char = first_char2 = re->first_char;      first_char = first_char2 = (pcre_uchar)(re->first_char);
6772      if ((re->flags & PCRE_FCH_CASELESS) != 0)      if ((re->flags & PCRE_FCH_CASELESS) != 0)
6773        {        {
6774        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 6790  character" set. */
6790  if ((re->flags & PCRE_REQCHSET) != 0)  if ((re->flags & PCRE_REQCHSET) != 0)
6791    {    {
6792    has_req_char = TRUE;    has_req_char = TRUE;
6793    req_char = req_char2 = re->req_char;    req_char = req_char2 = (pcre_uchar)(re->req_char);
6794    if ((re->flags & PCRE_RCH_CASELESS) != 0)    if ((re->flags & PCRE_RCH_CASELESS) != 0)
6795      {      {
6796      req_char2 = TABLE_GET(req_char, md->fcc, req_char);      req_char2 = TABLE_GET(req_char, md->fcc, req_char);
# Line 6575  for(;;) Line 6848  for(;;)
6848    
6849      if (has_first_char)      if (has_first_char)
6850        {        {
6851          pcre_uchar smc;
6852    
6853        if (first_char != first_char2)        if (first_char != first_char2)
6854          while (start_match < end_subject &&          while (start_match < end_subject &&
6855              *start_match != first_char && *start_match != first_char2)            (smc = RAWUCHARTEST(start_match)) != first_char && smc != first_char2)
6856            start_match++;            start_match++;
6857        else        else
6858          while (start_match < end_subject && *start_match != first_char)          while (start_match < end_subject && RAWUCHARTEST(start_match) != first_char)
6859            start_match++;            start_match++;
6860        }        }
6861    
# Line 6612  for(;;) Line 6887  for(;;)
6887          if (start_match[-1] == CHAR_CR &&          if (start_match[-1] == CHAR_CR &&
6888               (md->nltype == NLTYPE_ANY || md->nltype == NLTYPE_ANYCRLF) &&               (md->nltype == NLTYPE_ANY || md->nltype == NLTYPE_ANYCRLF) &&
6889               start_match < end_subject &&               start_match < end_subject &&
6890               *start_match == CHAR_NL)               RAWUCHARTEST(start_match) == CHAR_NL)
6891            start_match++;            start_match++;
6892          }          }
6893        }        }
# Line 6623  for(;;) Line 6898  for(;;)
6898        {        {
6899        while (start_match < end_subject)        while (start_match < end_subject)
6900          {          {
6901          register unsigned int c = *start_match;          register pcre_uint32 c = RAWUCHARTEST(start_match);
6902  #ifndef COMPILE_PCRE8  #ifndef COMPILE_PCRE8
6903          if (c > 255) c = 255;          if (c > 255) c = 255;
6904  #endif  #endif
# Line 6691  for(;;) Line 6966  for(;;)
6966            {            {
6967            while (p < end_subject)            while (p < end_subject)
6968              {              {
6969              register int pp = *p++;              register pcre_uint32 pp = RAWUCHARINCTEST(p);
6970              if (pp == req_char || pp == req_char2) { p--; break; }              if (pp == req_char || pp == req_char2) { p--; break; }
6971              }              }
6972            }            }
# Line 6699  for(;;) Line 6974  for(;;)
6974            {            {
6975            while (p < end_subject)            while (p < end_subject)
6976              {              {
6977              if (*p++ == req_char) { p--; break; }              if (RAWUCHARINCTEST(p) == req_char) { p--; break; }
6978              }              }
6979            }            }
6980    
# Line 6735  for(;;) Line 7010  for(;;)
7010    md->match_call_count = 0;    md->match_call_count = 0;
7011    md->match_function_type = 0;    md->match_function_type = 0;
7012    md->end_offset_top = 0;    md->end_offset_top = 0;
7013      md->skip_arg_count = 0;
7014    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);
7015    if (md->hitend && start_partial == NULL) start_partial = md->start_used_ptr;    if (md->hitend && start_partial == NULL)
7016        {
7017        start_partial = md->start_used_ptr;
7018        match_partial = start_match;
7019        }
7020    
7021    switch(rc)    switch(rc)
7022      {      {
# Line 6749  for(;;) Line 7029  for(;;)
7029    
7030      case MATCH_SKIP_ARG:      case MATCH_SKIP_ARG:
7031      new_start_match = start_match;      new_start_match = start_match;
7032      md->ignore_skip_arg = TRUE;      md->ignore_skip_arg = md->skip_arg_count;
7033      break;      break;
7034    
7035      /* 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
7036      same as the match we have just done, treat it as NOMATCH. */      greater than the match we have just done, treat it as NOMATCH. */
7037    
7038      case MATCH_SKIP:      case MATCH_SKIP:
7039      if (md->start_match_ptr != start_match)      if (md->start_match_ptr > start_match)
7040        {        {
7041        new_start_match = md->start_match_ptr;        new_start_match = md->start_match_ptr;
7042        break;        break;
# Line 6764  for(;;) Line 7044  for(;;)
7044      /* Fall through */      /* Fall through */
7045    
7046      /* NOMATCH and PRUNE advance by one character. THEN at this level acts      /* NOMATCH and PRUNE advance by one character. THEN at this level acts
7047      exactly like PRUNE. Unset the ignore SKIP-with-argument flag. */      exactly like PRUNE. Unset ignore SKIP-with-argument. */
7048    
7049      case MATCH_NOMATCH:      case MATCH_NOMATCH:
7050      case MATCH_PRUNE:      case MATCH_PRUNE:
7051      case MATCH_THEN:      case MATCH_THEN:
7052      md->ignore_skip_arg = FALSE;      md->ignore_skip_arg = 0;
7053      new_start_match = start_match + 1;      new_start_match = start_match + 1;
7054  #ifdef SUPPORT_UTF  #ifdef SUPPORT_UTF
7055      if (utf)      if (utf)
# Line 6862  if (rc == MATCH_MATCH || rc == MATCH_ACC Line 7142  if (rc == MATCH_MATCH || rc == MATCH_ACC
7142          (arg_offset_max - 2) * sizeof(int));          (arg_offset_max - 2) * sizeof(int));
7143        DPRINTF(("Copied offsets from temporary memory\n"));        DPRINTF(("Copied offsets from temporary memory\n"));
7144        }        }
7145      if (md->end_offset_top > arg_offset_max) md->offset_overflow = TRUE;      if (md->end_offset_top > arg_offset_max) md->capture_last |= OVFLBIT;
7146      DPRINTF(("Freeing temporary memory\n"));      DPRINTF(("Freeing temporary memory\n"));
7147      (PUBL(free))(md->offset_vector);      (PUBL(free))(md->offset_vector);
7148      }      }
# Line 6870  if (rc == MATCH_MATCH || rc == MATCH_ACC Line 7150  if (rc == MATCH_MATCH || rc == MATCH_ACC
7150    /* Set the return code to the number of captured strings, or 0 if there were    /* Set the return code to the number of captured strings, or 0 if there were
7151    too many to fit into the vector. */    too many to fit into the vector. */
7152    
7153    rc = (md->offset_overflow && md->end_offset_top >= arg_offset_max)?    rc = ((md->capture_last & OVFLBIT) != 0 &&
7154             md->end_offset_top >= arg_offset_max)?
7155      0 : md->end_offset_top/2;      0 : md->end_offset_top/2;
7156    
7157    /* If there is space in the offset vector, set any unused pairs at the end of    /* If there is space in the offset vector, set any unused pairs at the end of
# Line 6886  if (rc == MATCH_MATCH || rc == MATCH_ACC Line 7167  if (rc == MATCH_MATCH || rc == MATCH_ACC
7167      {      {
7168      register int *iptr, *iend;      register int *iptr, *iend;
7169      int resetcount = 2 + re->top_bracket * 2;      int resetcount = 2 + re->top_bracket * 2;
7170      if (resetcount > offsetcount) resetcount = ocount;      if (resetcount > offsetcount) resetcount = offsetcount;
7171      iptr = offsets + md->end_offset_top;      iptr = offsets + md->end_offset_top;
7172      iend = offsets + resetcount;      iend = offsets + resetcount;
7173      while (iptr < iend) *iptr++ = -1;      while (iptr < iend) *iptr++ = -1;
# Line 6907  if (rc == MATCH_MATCH || rc == MATCH_ACC Line 7188  if (rc == MATCH_MATCH || rc == MATCH_ACC
7188    if (extra_data != NULL && (extra_data->flags & PCRE_EXTRA_MARK) != 0)    if (extra_data != NULL && (extra_data->flags & PCRE_EXTRA_MARK) != 0)
7189      *(extra_data->mark) = (pcre_uchar *)md->mark;      *(extra_data->mark) = (pcre_uchar *)md->mark;
7190    DPRINTF((">>>> returning %d\n", rc));    DPRINTF((">>>> returning %d\n", rc));
7191    #ifdef NO_RECURSE
7192      release_match_heapframes(&frame_zero);
7193    #endif
7194    return rc;    return rc;
7195    }    }
7196    
# Line 6924  if (using_temporary_offsets) Line 7208  if (using_temporary_offsets)
7208  if (rc != MATCH_NOMATCH && rc != PCRE_ERROR_PARTIAL)  if (rc != MATCH_NOMATCH && rc != PCRE_ERROR_PARTIAL)
7209    {    {
7210    DPRINTF((">>>> error: returning %d\n", rc));    DPRINTF((">>>> error: returning %d\n", rc));
7211    #ifdef NO_RECURSE
7212      release_match_heapframes(&frame_zero);
7213    #endif
7214    return rc;    return rc;
7215    }    }
7216    
7217  /* Handle partial matches - disable any mark data */  /* Handle partial matches - disable any mark data */
7218    
7219  if (start_partial != NULL)  if (match_partial != NULL)
7220    {    {
7221    DPRINTF((">>>> returning PCRE_ERROR_PARTIAL\n"));    DPRINTF((">>>> returning PCRE_ERROR_PARTIAL\n"));
7222    md->mark = NULL;    md->mark = NULL;
# Line 6937  if (start_partial != NULL) Line 7224  if (start_partial != NULL)
7224      {      {
7225      offsets[0] = (int)(start_partial - (PCRE_PUCHAR)subject);      offsets[0] = (int)(start_partial - (PCRE_PUCHAR)subject);
7226      offsets[1] = (int)(end_subject - (PCRE_PUCHAR)subject);      offsets[1] = (int)(end_subject - (PCRE_PUCHAR)subject);
7227        if (offsetcount > 2)
7228          offsets[2] = (int)(match_partial - (PCRE_PUCHAR)subject);
7229      }      }
7230    rc = PCRE_ERROR_PARTIAL;    rc = PCRE_ERROR_PARTIAL;
7231    }    }
# Line 6953  else Line 7242  else
7242    
7243  if (extra_data != NULL && (extra_data->flags & PCRE_EXTRA_MARK) != 0)  if (extra_data != NULL && (extra_data->flags & PCRE_EXTRA_MARK) != 0)
7244    *(extra_data->mark) = (pcre_uchar *)md->nomatch_mark;    *(extra_data->mark) = (pcre_uchar *)md->nomatch_mark;
7245    #ifdef NO_RECURSE
7246      release_match_heapframes(&frame_zero);
7247    #endif
7248  return rc;  return rc;
7249  }  }
7250    

Legend:
Removed from v.895  
changed lines
  Added in v.1361

  ViewVC Help
Powered by ViewVC 1.1.5