/[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 1365 by ph10, Sun Oct 6 18:33:56 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 1233  for (;;) Line 1274  for (;;)
1274    
1275      /* Control never reaches here. */      /* Control never reaches here. */
1276    
1277      /* Conditional group: compilation checked that there are no more than      /* Conditional group: compilation checked that there are no more than two
1278      two branches. If the condition is false, skipping the first branch takes us      branches. If the condition is false, skipping the first branch takes us
1279      past the end if there is only one branch, but that's OK because that is      past the end of the item if there is only one branch, but that's exactly
1280      exactly what going to the ket would do. */      what we want. */
1281    
1282      case OP_COND:      case OP_COND:
1283      case OP_SCOND:      case OP_SCOND:
1284      codelink = GET(ecode, 1);  
1285        /* The variable codelink will be added to ecode when the condition is
1286        false, to get to the second branch. Setting it to the offset to the ALT
1287        or KET, then incrementing ecode achieves this effect. We now have ecode
1288        pointing to the condition or callout. */
1289    
1290        codelink = GET(ecode, 1);   /* Offset to the second branch */
1291        ecode += 1 + LINK_SIZE;     /* From this opcode */
1292    
1293      /* Because of the way auto-callout works during compile, a callout item is      /* Because of the way auto-callout works during compile, a callout item is
1294      inserted between OP_COND and an assertion condition. */      inserted between OP_COND and an assertion condition. */
1295    
1296      if (ecode[LINK_SIZE+1] == OP_CALLOUT)      if (*ecode == OP_CALLOUT)
1297        {        {
1298        if (PUBL(callout) != NULL)        if (PUBL(callout) != NULL)
1299          {          {
1300          PUBL(callout_block) cb;          PUBL(callout_block) cb;
1301          cb.version          = 2;   /* Version 1 of the callout block */          cb.version          = 2;   /* Version 1 of the callout block */
1302          cb.callout_number   = ecode[LINK_SIZE+2];          cb.callout_number   = ecode[1];
1303          cb.offset_vector    = md->offset_vector;          cb.offset_vector    = md->offset_vector;
1304  #ifdef COMPILE_PCRE8  #if defined COMPILE_PCRE8
1305          cb.subject          = (PCRE_SPTR)md->start_subject;          cb.subject          = (PCRE_SPTR)md->start_subject;
1306  #else  #elif defined COMPILE_PCRE16
1307          cb.subject          = (PCRE_SPTR16)md->start_subject;          cb.subject          = (PCRE_SPTR16)md->start_subject;
1308    #elif defined COMPILE_PCRE32
1309            cb.subject          = (PCRE_SPTR32)md->start_subject;
1310  #endif  #endif
1311          cb.subject_length   = (int)(md->end_subject - md->start_subject);          cb.subject_length   = (int)(md->end_subject - md->start_subject);
1312          cb.start_match      = (int)(mstart - md->start_subject);          cb.start_match      = (int)(mstart - md->start_subject);
1313          cb.current_position = (int)(eptr - md->start_subject);          cb.current_position = (int)(eptr - md->start_subject);
1314          cb.pattern_position = GET(ecode, LINK_SIZE + 3);          cb.pattern_position = GET(ecode, 2);
1315          cb.next_item_length = GET(ecode, 3 + 2*LINK_SIZE);          cb.next_item_length = GET(ecode, 2 + LINK_SIZE);
1316          cb.capture_top      = offset_top/2;          cb.capture_top      = offset_top/2;
1317          cb.capture_last     = md->capture_last;          cb.capture_last     = md->capture_last & CAPLMASK;
1318            /* Internal change requires this for API compatibility. */
1319            if (cb.capture_last == 0) cb.capture_last = -1;
1320          cb.callout_data     = md->callout_data;          cb.callout_data     = md->callout_data;
1321          cb.mark             = md->nomatch_mark;          cb.mark             = md->nomatch_mark;
1322          if ((rrc = (*PUBL(callout))(&cb)) > 0) RRETURN(MATCH_NOMATCH);          if ((rrc = (*PUBL(callout))(&cb)) > 0) RRETURN(MATCH_NOMATCH);
1323          if (rrc < 0) RRETURN(rrc);          if (rrc < 0) RRETURN(rrc);
1324          }          }
1325    
1326          /* Advance ecode past the callout, so it now points to the condition. We
1327          must adjust codelink so that the value of ecode+codelink is unchanged. */
1328    
1329        ecode += PRIV(OP_lengths)[OP_CALLOUT];        ecode += PRIV(OP_lengths)[OP_CALLOUT];
1330          codelink -= PRIV(OP_lengths)[OP_CALLOUT];
1331        }        }
1332    
1333      condcode = ecode[LINK_SIZE+1];      /* Test the various possible conditions */
1334    
1335      /* Now see what the actual condition is */      condition = FALSE;
1336        switch(condcode = *ecode)
1337      if (condcode == OP_RREF || condcode == OP_NRREF)    /* Recursion test */        {
1338        {        case OP_RREF:         /* Numbered group recursion test */
1339        if (md->recursive == NULL)                /* Not recursing => FALSE */        if (md->recursive != NULL)     /* Not recursing => FALSE */
         {  
         condition = FALSE;  
         ecode += GET(ecode, 1);  
         }  
       else  
1340          {          {
1341          int recno = GET2(ecode, LINK_SIZE + 2);   /* Recursion group number*/          unsigned int recno = GET2(ecode, 1);   /* Recursion group number*/
1342          condition = (recno == RREF_ANY || recno == md->recursive->group_num);          condition = (recno == RREF_ANY || recno == md->recursive->group_num);
1343            }
1344          break;
1345    
1346          /* If the test is for recursion into a specific subpattern, and it is        case OP_DNRREF:       /* Duplicate named group recursion test */
1347          false, but the test was set up by name, scan the table to see if the        if (md->recursive != NULL)
1348          name refers to any other numbers, and test them. The condition is true          {
1349          if any one is set. */          int count = GET2(ecode, 1 + IMM2_SIZE);
1350            pcre_uchar *slot = md->name_table + GET2(ecode, 1) * md->name_entry_size;
1351          if (!condition && condcode == OP_NRREF)          while (count-- > 0)
1352            {            {
1353            pcre_uchar *slotA = md->name_table;            unsigned int recno = GET2(slot, 0);
1354            for (i = 0; i < md->name_count; i++)            condition = recno == md->recursive->group_num;
1355              {            if (condition) break;
1356              if (GET2(slotA, 0) == recno) break;            slot += md->name_entry_size;
             slotA += md->name_entry_size;  
             }  
   
           /* Found a name for the number - there can be only one; duplicate  
           names for different numbers are allowed, but not vice versa. First  
           scan down for duplicates. */  
   
           if (i < md->name_count)  
             {  
             pcre_uchar *slotB = slotA;  
             while (slotB > md->name_table)  
               {  
               slotB -= md->name_entry_size;  
               if (STRCMP_UC_UC(slotA + IMM2_SIZE, slotB + IMM2_SIZE) == 0)  
                 {  
                 condition = GET2(slotB, 0) == md->recursive->group_num;  
                 if (condition) break;  
                 }  
               else break;  
               }  
   
             /* Scan up for duplicates */  
   
             if (!condition)  
               {  
               slotB = slotA;  
               for (i++; i < md->name_count; i++)  
                 {  
                 slotB += md->name_entry_size;  
                 if (STRCMP_UC_UC(slotA + IMM2_SIZE, slotB + IMM2_SIZE) == 0)  
                   {  
                   condition = GET2(slotB, 0) == md->recursive->group_num;  
                   if (condition) break;  
                   }  
                 else break;  
                 }  
               }  
             }  
1357            }            }
1358            }
1359          break;
1360    
1361          /* Chose branch according to the condition */        case OP_CREF:         /* Numbered group used test */
1362          offset = GET2(ecode, 1) << 1;  /* Doubled ref number */
         ecode += condition? 1 + IMM2_SIZE : GET(ecode, 1);  
         }  
       }  
   
     else if (condcode == OP_CREF || condcode == OP_NCREF)  /* Group used test */  
       {  
       offset = GET2(ecode, LINK_SIZE+2) << 1;  /* Doubled ref number */  
1363        condition = offset < offset_top && md->offset_vector[offset] >= 0;        condition = offset < offset_top && md->offset_vector[offset] >= 0;
1364          break;
1365    
1366        /* If the numbered capture is unset, but the reference was by name,        case OP_DNCREF:      /* Duplicate named group used test */
       scan the table to see if the name refers to any other numbers, and test  
       them. The condition is true if any one is set. This is tediously similar  
       to the code above, but not close enough to try to amalgamate. */  
   
       if (!condition && condcode == OP_NCREF)  
1367          {          {
1368          int refno = offset >> 1;          int count = GET2(ecode, 1 + IMM2_SIZE);
1369          pcre_uchar *slotA = md->name_table;          pcre_uchar *slot = md->name_table + GET2(ecode, 1) * md->name_entry_size;
1370            while (count-- > 0)
         for (i = 0; i < md->name_count; i++)  
           {  
           if (GET2(slotA, 0) == refno) break;  
           slotA += md->name_entry_size;  
           }  
   
         /* Found a name for the number - there can be only one; duplicate names  
         for different numbers are allowed, but not vice versa. First scan down  
         for duplicates. */  
   
         if (i < md->name_count)  
1371            {            {
1372            pcre_uchar *slotB = slotA;            offset = GET2(slot, 0) << 1;
1373            while (slotB > md->name_table)            condition = offset < offset_top && md->offset_vector[offset] >= 0;
1374              {            if (condition) break;
1375              slotB -= md->name_entry_size;            slot += md->name_entry_size;
             if (STRCMP_UC_UC(slotA + IMM2_SIZE, slotB + IMM2_SIZE) == 0)  
               {  
               offset = GET2(slotB, 0) << 1;  
               condition = offset < offset_top &&  
                 md->offset_vector[offset] >= 0;  
               if (condition) break;  
               }  
             else break;  
             }  
   
           /* Scan up for duplicates */  
   
           if (!condition)  
             {  
             slotB = slotA;  
             for (i++; i < md->name_count; i++)  
               {  
               slotB += md->name_entry_size;  
               if (STRCMP_UC_UC(slotA + IMM2_SIZE, slotB + IMM2_SIZE) == 0)  
                 {  
                 offset = GET2(slotB, 0) << 1;  
                 condition = offset < offset_top &&  
                   md->offset_vector[offset] >= 0;  
                 if (condition) break;  
                 }  
               else break;  
               }  
             }  
1376            }            }
1377          }          }
1378          break;
1379    
1380        /* Chose branch according to the condition */        case OP_DEF:     /* DEFINE - always false */
1381          break;
       ecode += condition? 1 + IMM2_SIZE : GET(ecode, 1);  
       }  
   
     else if (condcode == OP_DEF)     /* DEFINE - always false */  
       {  
       condition = FALSE;  
       ecode += GET(ecode, 1);  
       }  
   
     /* The condition is an assertion. Call match() to evaluate it - setting  
     md->match_function_type to MATCH_CONDASSERT causes it to stop at the end of  
     an assertion. */  
1382    
1383      else        /* The condition is an assertion. Call match() to evaluate it - setting
1384        {        md->match_function_type to MATCH_CONDASSERT causes it to stop at the end
1385          of an assertion. */
1386    
1387          default:
1388        md->match_function_type = MATCH_CONDASSERT;        md->match_function_type = MATCH_CONDASSERT;
1389        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, NULL, RM3);        RMATCH(eptr, ecode, offset_top, md, NULL, RM3);
1390        if (rrc == MATCH_MATCH)        if (rrc == MATCH_MATCH)
1391          {          {
1392          if (md->end_offset_top > offset_top)          if (md->end_offset_top > offset_top)
1393            offset_top = md->end_offset_top;  /* Captures may have happened */            offset_top = md->end_offset_top;  /* Captures may have happened */
1394          condition = TRUE;          condition = TRUE;
1395          ecode += 1 + LINK_SIZE + GET(ecode, LINK_SIZE + 2);  
1396            /* Advance ecode past the assertion to the start of the first branch,
1397            but adjust it so that the general choosing code below works. */
1398    
1399            ecode += GET(ecode, 1);
1400          while (*ecode == OP_ALT) ecode += GET(ecode, 1);          while (*ecode == OP_ALT) ecode += GET(ecode, 1);
1401            ecode += 1 + LINK_SIZE - PRIV(OP_lengths)[condcode];
1402          }          }
1403    
1404        /* PCRE doesn't allow the effect of (*THEN) to escape beyond an        /* PCRE doesn't allow the effect of (*THEN) to escape beyond an
1405        assertion; it is therefore treated as NOMATCH. */        assertion; it is therefore treated as NOMATCH. Any other return is an
1406          error. */
1407    
1408        else if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN)        else if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN)
1409          {          {
1410          RRETURN(rrc);         /* Need braces because of following else */          RRETURN(rrc);         /* Need braces because of following else */
1411          }          }
1412        else        break;
         {  
         condition = FALSE;  
         ecode += codelink;  
         }  
1413        }        }
1414    
1415      /* We are now at the branch that is to be obeyed. As there is only one, can      /* Choose branch according to the condition */
1416      use tail recursion to avoid using another stack frame, except when there is  
1417      unlimited repeat of a possibly empty group. In the latter case, a recursive      ecode += condition? PRIV(OP_lengths)[condcode] : codelink;
1418      call to match() is always required, unless the second alternative doesn't  
1419      exist, in which case we can just plough on. Note that, for compatibility      /* We are now at the branch that is to be obeyed. As there is only one, we
1420      with Perl, the | in a conditional group is NOT treated as creating two      can use tail recursion to avoid using another stack frame, except when
1421      alternatives. If a THEN is encountered in the branch, it propagates out to      there is unlimited repeat of a possibly empty group. In the latter case, a
1422      the enclosing alternative (unless nested in a deeper set of alternatives,      recursive call to match() is always required, unless the second alternative
1423      of course). */      doesn't exist, in which case we can just plough on. Note that, for
1424        compatibility with Perl, the | in a conditional group is NOT treated as
1425      if (condition || *ecode == OP_ALT)      creating two alternatives. If a THEN is encountered in the branch, it
1426        propagates out to the enclosing alternative (unless nested in a deeper set
1427        of alternatives, of course). */
1428    
1429        if (condition || ecode[-(1+LINK_SIZE)] == OP_ALT)
1430        {        {
1431        if (op != OP_SCOND)        if (op != OP_SCOND)
1432          {          {
         ecode += 1 + LINK_SIZE;  
1433          goto TAIL_RECURSE;          goto TAIL_RECURSE;
1434          }          }
1435    
1436        md->match_function_type = MATCH_CBEGROUP;        md->match_function_type = MATCH_CBEGROUP;
1437        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, eptrb, RM49);        RMATCH(eptr, ecode, offset_top, md, eptrb, RM49);
1438        RRETURN(rrc);        RRETURN(rrc);
1439        }        }
1440    
# Line 1477  for (;;) Line 1442  for (;;)
1442    
1443      else      else
1444        {        {
       ecode += 1 + LINK_SIZE;  
1445        }        }
1446      break;      break;
1447    
# Line 1486  for (;;) Line 1450  for (;;)
1450      to close any currently open capturing brackets. */      to close any currently open capturing brackets. */
1451    
1452      case OP_CLOSE:      case OP_CLOSE:
1453      number = GET2(ecode, 1);      number = GET2(ecode, 1);   /* Must be less than 65536 */
1454      offset = number << 1;      offset = number << 1;
1455    
1456  #ifdef PCRE_DEBUG  #ifdef PCRE_DEBUG
# Line 1494  for (;;) Line 1458  for (;;)
1458        printf("\n");        printf("\n");
1459  #endif  #endif
1460    
1461      md->capture_last = number;      md->capture_last = (md->capture_last & OVFLMASK) | number;
1462      if (offset >= md->offset_max) md->offset_overflow = TRUE; else      if (offset >= md->offset_max) md->capture_last |= OVFLBIT; else
1463        {        {
1464        md->offset_vector[offset] =        md->offset_vector[offset] =
1465          md->offset_vector[md->offset_end - number];          md->offset_vector[md->offset_end - number];
# Line 1549  for (;;) Line 1513  for (;;)
1513    
1514      case OP_ASSERT:      case OP_ASSERT:
1515      case OP_ASSERTBACK:      case OP_ASSERTBACK:
1516      save_mark = md->mark;      save_mark = md->mark;
1517      if (md->match_function_type == MATCH_CONDASSERT)      if (md->match_function_type == MATCH_CONDASSERT)
1518        {        {
1519        condassert = TRUE;        condassert = TRUE;
# Line 1557  for (;;) Line 1521  for (;;)
1521        }        }
1522      else condassert = FALSE;      else condassert = FALSE;
1523    
1524        /* Loop for each branch */
1525    
1526      do      do
1527        {        {
1528        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, NULL, RM4);        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, NULL, RM4);
1529    
1530          /* A match means that the assertion is true; break out of the loop
1531          that matches its alternatives. */
1532    
1533        if (rrc == MATCH_MATCH || rrc == MATCH_ACCEPT)        if (rrc == MATCH_MATCH || rrc == MATCH_ACCEPT)
1534          {          {
1535          mstart = md->start_match_ptr;   /* In case \K reset it */          mstart = md->start_match_ptr;   /* In case \K reset it */
1536          break;          break;
1537          }          }
1538    
1539        /* PCRE does not allow THEN to escape beyond an assertion; it is treated        /* If not matched, restore the previous mark setting. */
1540        as NOMATCH. */  
1541          md->mark = save_mark;
1542    
1543          /* See comment in the code for capturing groups above about handling
1544          THEN. */
1545    
1546          if (rrc == MATCH_THEN)
1547            {
1548            next = ecode + GET(ecode,1);
1549            if (md->start_match_ptr < next &&
1550                (*ecode == OP_ALT || *next == OP_ALT))
1551              rrc = MATCH_NOMATCH;
1552            }
1553    
1554          /* Anything other than NOMATCH causes the entire assertion to fail,
1555          passing back the return code. This includes COMMIT, SKIP, PRUNE and an
1556          uncaptured THEN, which means they take their normal effect. This
1557          consistent approach does not always have exactly the same effect as in
1558          Perl. */
1559    
1560        if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN) RRETURN(rrc);        if (rrc != MATCH_NOMATCH) RRETURN(rrc);
1561        ecode += GET(ecode, 1);        ecode += GET(ecode, 1);
       md->mark = save_mark;  
1562        }        }
1563      while (*ecode == OP_ALT);      while (*ecode == OP_ALT);   /* Continue for next alternative */
1564    
1565        /* If we have tried all the alternative branches, the assertion has
1566        failed. If not, we broke out after a match. */
1567    
1568      if (*ecode == OP_KET) RRETURN(MATCH_NOMATCH);      if (*ecode == OP_KET) RRETURN(MATCH_NOMATCH);
1569    
# Line 1581  for (;;) Line 1571  for (;;)
1571    
1572      if (condassert) RRETURN(MATCH_MATCH);      if (condassert) RRETURN(MATCH_MATCH);
1573    
1574      /* Continue from after the assertion, updating the offsets high water      /* Continue from after a successful assertion, updating the offsets high
1575      mark, since extracts may have been taken during the assertion. */      water mark, since extracts may have been taken during the assertion. */
1576    
1577      do ecode += GET(ecode,1); while (*ecode == OP_ALT);      do ecode += GET(ecode,1); while (*ecode == OP_ALT);
1578      ecode += 1 + LINK_SIZE;      ecode += 1 + LINK_SIZE;
1579      offset_top = md->end_offset_top;      offset_top = md->end_offset_top;
1580      continue;      continue;
1581    
1582      /* Negative assertion: all branches must fail to match. Encountering SKIP,      /* Negative assertion: all branches must fail to match for the assertion to
1583      PRUNE, or COMMIT means we must assume failure without checking subsequent      succeed. */
     branches. */  
1584    
1585      case OP_ASSERT_NOT:      case OP_ASSERT_NOT:
1586      case OP_ASSERTBACK_NOT:      case OP_ASSERTBACK_NOT:
1587      save_mark = md->mark;      save_mark = md->mark;
1588      if (md->match_function_type == MATCH_CONDASSERT)      if (md->match_function_type == MATCH_CONDASSERT)
1589        {        {
1590        condassert = TRUE;        condassert = TRUE;
# Line 1603  for (;;) Line 1592  for (;;)
1592        }        }
1593      else condassert = FALSE;      else condassert = FALSE;
1594    
1595        /* Loop for each alternative branch. */
1596    
1597      do      do
1598        {        {
1599        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, NULL, RM5);        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, NULL, RM5);
1600        md->mark = save_mark;        md->mark = save_mark;   /* Always restore the mark setting */
1601        if (rrc == MATCH_MATCH || rrc == MATCH_ACCEPT) RRETURN(MATCH_NOMATCH);  
1602        if (rrc == MATCH_SKIP || rrc == MATCH_PRUNE || rrc == MATCH_COMMIT)        switch(rrc)
1603          {          {
1604          do ecode += GET(ecode,1); while (*ecode == OP_ALT);          case MATCH_MATCH:            /* A successful match means */
1605            case MATCH_ACCEPT:           /* the assertion has failed. */
1606            RRETURN(MATCH_NOMATCH);
1607    
1608            case MATCH_NOMATCH:          /* Carry on with next branch */
1609          break;          break;
1610    
1611            /* See comment in the code for capturing groups above about handling
1612            THEN. */
1613    
1614            case MATCH_THEN:
1615            next = ecode + GET(ecode,1);
1616            if (md->start_match_ptr < next &&
1617                (*ecode == OP_ALT || *next == OP_ALT))
1618              {
1619              rrc = MATCH_NOMATCH;
1620              break;
1621              }
1622            /* Otherwise fall through. */
1623    
1624            /* COMMIT, SKIP, PRUNE, and an uncaptured THEN cause the whole
1625            assertion to fail to match, without considering any more alternatives.
1626            Failing to match means the assertion is true. This is a consistent
1627            approach, but does not always have the same effect as in Perl. */
1628    
1629            case MATCH_COMMIT:
1630            case MATCH_SKIP:
1631            case MATCH_SKIP_ARG:
1632            case MATCH_PRUNE:
1633            do ecode += GET(ecode,1); while (*ecode == OP_ALT);
1634            goto NEG_ASSERT_TRUE;   /* Break out of alternation loop */
1635    
1636            /* Anything else is an error */
1637    
1638            default:
1639            RRETURN(rrc);
1640          }          }
1641    
1642        /* PCRE does not allow THEN to escape beyond an assertion; it is treated        /* Continue with next branch */
       as NOMATCH. */  
1643    
       if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN) RRETURN(rrc);  
1644        ecode += GET(ecode,1);        ecode += GET(ecode,1);
1645        }        }
1646      while (*ecode == OP_ALT);      while (*ecode == OP_ALT);
1647    
1648      if (condassert) RRETURN(MATCH_MATCH);  /* Condition assertion */      /* All branches in the assertion failed to match. */
1649    
1650      ecode += 1 + LINK_SIZE;      NEG_ASSERT_TRUE:
1651        if (condassert) RRETURN(MATCH_MATCH);  /* Condition assertion */
1652        ecode += 1 + LINK_SIZE;                /* Continue with current branch */
1653      continue;      continue;
1654    
1655      /* Move the subject pointer back. This occurs only at the start of      /* Move the subject pointer back. This occurs only at the start of
# Line 1671  for (;;) Line 1696  for (;;)
1696        cb.version          = 2;   /* Version 1 of the callout block */        cb.version          = 2;   /* Version 1 of the callout block */
1697        cb.callout_number   = ecode[1];        cb.callout_number   = ecode[1];
1698        cb.offset_vector    = md->offset_vector;        cb.offset_vector    = md->offset_vector;
1699  #ifdef COMPILE_PCRE8  #if defined COMPILE_PCRE8
1700        cb.subject          = (PCRE_SPTR)md->start_subject;        cb.subject          = (PCRE_SPTR)md->start_subject;
1701  #else  #elif defined COMPILE_PCRE16
1702        cb.subject          = (PCRE_SPTR16)md->start_subject;        cb.subject          = (PCRE_SPTR16)md->start_subject;
1703    #elif defined COMPILE_PCRE32
1704          cb.subject          = (PCRE_SPTR32)md->start_subject;
1705  #endif  #endif
1706        cb.subject_length   = (int)(md->end_subject - md->start_subject);        cb.subject_length   = (int)(md->end_subject - md->start_subject);
1707        cb.start_match      = (int)(mstart - md->start_subject);        cb.start_match      = (int)(mstart - md->start_subject);
# Line 1682  for (;;) Line 1709  for (;;)
1709        cb.pattern_position = GET(ecode, 2);        cb.pattern_position = GET(ecode, 2);
1710        cb.next_item_length = GET(ecode, 2 + LINK_SIZE);        cb.next_item_length = GET(ecode, 2 + LINK_SIZE);
1711        cb.capture_top      = offset_top/2;        cb.capture_top      = offset_top/2;
1712        cb.capture_last     = md->capture_last;        cb.capture_last     = md->capture_last & CAPLMASK;
1713          /* Internal change requires this for API compatibility. */
1714          if (cb.capture_last == 0) cb.capture_last = -1;
1715        cb.callout_data     = md->callout_data;        cb.callout_data     = md->callout_data;
1716        cb.mark             = md->nomatch_mark;        cb.mark             = md->nomatch_mark;
1717        if ((rrc = (*PUBL(callout))(&cb)) > 0) RRETURN(MATCH_NOMATCH);        if ((rrc = (*PUBL(callout))(&cb)) > 0) RRETURN(MATCH_NOMATCH);
# Line 1711  for (;;) Line 1740  for (;;)
1740      case OP_RECURSE:      case OP_RECURSE:
1741        {        {
1742        recursion_info *ri;        recursion_info *ri;
1743        int recno;        unsigned int recno;
1744    
1745        callpat = md->start_code + GET(ecode, 1);        callpat = md->start_code + GET(ecode, 1);
1746        recno = (callpat == md->start_code)? 0 :        recno = (callpat == md->start_code)? 0 :
# Line 1728  for (;;) Line 1757  for (;;)
1757        /* Add to "recursing stack" */        /* Add to "recursing stack" */
1758    
1759        new_recursive.group_num = recno;        new_recursive.group_num = recno;
1760          new_recursive.saved_capture_last = md->capture_last;
1761        new_recursive.subject_position = eptr;        new_recursive.subject_position = eptr;
1762        new_recursive.prevrec = md->recursive;        new_recursive.prevrec = md->recursive;
1763        md->recursive = &new_recursive;        md->recursive = &new_recursive;
# Line 1751  for (;;) Line 1781  for (;;)
1781              new_recursive.saved_max * sizeof(int));              new_recursive.saved_max * sizeof(int));
1782    
1783        /* OK, now we can do the recursion. After processing each alternative,        /* OK, now we can do the recursion. After processing each alternative,
1784        restore the offset data. If there were nested recursions, md->recursive        restore the offset data and the last captured value. If there were nested
1785        might be changed, so reset it before looping. */        recursions, md->recursive might be changed, so reset it before looping.
1786          */
1787    
1788        DPRINTF(("Recursing into group %d\n", new_recursive.group_num));        DPRINTF(("Recursing into group %d\n", new_recursive.group_num));
1789        cbegroup = (*callpat >= OP_SBRA);        cbegroup = (*callpat >= OP_SBRA);
# Line 1763  for (;;) Line 1794  for (;;)
1794            md, eptrb, RM6);            md, eptrb, RM6);
1795          memcpy(md->offset_vector, new_recursive.offset_save,          memcpy(md->offset_vector, new_recursive.offset_save,
1796              new_recursive.saved_max * sizeof(int));              new_recursive.saved_max * sizeof(int));
1797            md->capture_last = new_recursive.saved_capture_last;
1798          md->recursive = new_recursive.prevrec;          md->recursive = new_recursive.prevrec;
1799          if (rrc == MATCH_MATCH || rrc == MATCH_ACCEPT)          if (rrc == MATCH_MATCH || rrc == MATCH_ACCEPT)
1800            {            {
# Line 1779  for (;;) Line 1811  for (;;)
1811            goto RECURSION_MATCHED;        /* Exit loop; end processing */            goto RECURSION_MATCHED;        /* Exit loop; end processing */
1812            }            }
1813    
1814          /* PCRE does not allow THEN to escape beyond a recursion; it is treated          /* PCRE does not allow THEN, SKIP, PRUNE or COMMIT to escape beyond a
1815          as NOMATCH. */          recursion; they cause a NOMATCH for the entire recursion. These codes
1816            are defined in a range that can be tested for. */
1817    
1818            if (rrc >= MATCH_BACKTRACK_MIN && rrc <= MATCH_BACKTRACK_MAX)
1819              RRETURN(MATCH_NOMATCH);
1820    
1821            /* Any return code other than NOMATCH is an error. */
1822    
1823          else if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN)          if (rrc != MATCH_NOMATCH)
1824            {            {
1825            DPRINTF(("Recursion gave error %d\n", rrc));            DPRINTF(("Recursion gave error %d\n", rrc));
1826            if (new_recursive.offset_save != stacksave)            if (new_recursive.offset_save != stacksave)
# Line 1912  for (;;) Line 1950  for (;;)
1950    
1951        /* Deal with capturing */        /* Deal with capturing */
1952    
1953        md->capture_last = number;        md->capture_last = (md->capture_last & OVFLMASK) | number;
1954        if (offset >= md->offset_max) md->offset_overflow = TRUE; else        if (offset >= md->offset_max) md->capture_last |= OVFLBIT; else
1955          {          {
1956          /* If offset is greater than offset_top, it means that we are          /* If offset is greater than offset_top, it means that we are
1957          "skipping" a capturing group, and that group's offsets must be marked          "skipping" a capturing group, and that group's offsets must be marked
# Line 1993  for (;;) Line 2031  for (;;)
2031          }          }
2032        if (*prev >= OP_SBRA)    /* Could match an empty string */        if (*prev >= OP_SBRA)    /* Could match an empty string */
2033          {          {
         md->match_function_type = MATCH_CBEGROUP;  
2034          RMATCH(eptr, prev, offset_top, md, eptrb, RM50);          RMATCH(eptr, prev, offset_top, md, eptrb, RM50);
2035          RRETURN(rrc);          RRETURN(rrc);
2036          }          }
# Line 2002  for (;;) Line 2039  for (;;)
2039        }        }
2040      else  /* OP_KETRMAX */      else  /* OP_KETRMAX */
2041        {        {
       if (*prev >= OP_SBRA) md->match_function_type = MATCH_CBEGROUP;  
2042        RMATCH(eptr, prev, offset_top, md, eptrb, RM13);        RMATCH(eptr, prev, offset_top, md, eptrb, RM13);
2043        if (rrc == MATCH_ONCE && md->once_target == prev) rrc = MATCH_NOMATCH;        if (rrc == MATCH_ONCE && md->once_target == prev) rrc = MATCH_NOMATCH;
2044        if (rrc != MATCH_NOMATCH) RRETURN(rrc);        if (rrc != MATCH_NOMATCH) RRETURN(rrc);
# Line 2059  for (;;) Line 2095  for (;;)
2095    
2096      case OP_DOLLM:      case OP_DOLLM:
2097      if (eptr < md->end_subject)      if (eptr < md->end_subject)
2098        { if (!IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH); }        {
2099          if (!IS_NEWLINE(eptr))
2100            {
2101            if (md->partial != 0 &&
2102                eptr + 1 >= md->end_subject &&
2103                NLBLOCK->nltype == NLTYPE_FIXED &&
2104                NLBLOCK->nllen == 2 &&
2105                RAWUCHARTEST(eptr) == NLBLOCK->nl[0])
2106              {
2107              md->hitend = TRUE;
2108              if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
2109              }
2110            RRETURN(MATCH_NOMATCH);
2111            }
2112          }
2113      else      else
2114        {        {
2115        if (md->noteol) RRETURN(MATCH_NOMATCH);        if (md->noteol) RRETURN(MATCH_NOMATCH);
# Line 2091  for (;;) Line 2141  for (;;)
2141      ASSERT_NL_OR_EOS:      ASSERT_NL_OR_EOS:
2142      if (eptr < md->end_subject &&      if (eptr < md->end_subject &&
2143          (!IS_NEWLINE(eptr) || eptr != md->end_subject - md->nllen))          (!IS_NEWLINE(eptr) || eptr != md->end_subject - md->nllen))
2144          {
2145          if (md->partial != 0 &&
2146              eptr + 1 >= md->end_subject &&
2147              NLBLOCK->nltype == NLTYPE_FIXED &&
2148              NLBLOCK->nllen == 2 &&
2149              RAWUCHARTEST(eptr) == NLBLOCK->nl[0])
2150            {
2151            md->hitend = TRUE;
2152            if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
2153            }
2154        RRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2155          }
2156    
2157      /* Either at end of string or \n before end. */      /* Either at end of string or \n before end. */
2158    
# Line 2219  for (;;) Line 2280  for (;;)
2280        }        }
2281      break;      break;
2282    
2283      /* Match a single character type; inline for speed */      /* Match any single character type except newline; have to take care with
2284        CRLF newlines and partial matching. */
2285    
2286      case OP_ANY:      case OP_ANY:
2287      if (IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH);      if (IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH);
2288        if (md->partial != 0 &&
2289            eptr + 1 >= md->end_subject &&
2290            NLBLOCK->nltype == NLTYPE_FIXED &&
2291            NLBLOCK->nllen == 2 &&
2292            RAWUCHARTEST(eptr) == NLBLOCK->nl[0])
2293          {
2294          md->hitend = TRUE;
2295          if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
2296          }
2297    
2298      /* Fall through */      /* Fall through */
2299    
2300        /* Match any single character whatsoever. */
2301    
2302      case OP_ALLANY:      case OP_ALLANY:
2303      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 */
2304        {                            /* not be updated before SCHECK_PARTIAL. */        {                            /* not be updated before SCHECK_PARTIAL. */
# Line 2364  for (;;) Line 2438  for (;;)
2438        {        {
2439        default: RRETURN(MATCH_NOMATCH);        default: RRETURN(MATCH_NOMATCH);
2440    
2441        case 0x000d:        case CHAR_CR:
2442        if (eptr < md->end_subject && *eptr == 0x0a) eptr++;        if (eptr >= md->end_subject)
2443            {
2444            SCHECK_PARTIAL();
2445            }
2446          else if (RAWUCHARTEST(eptr) == CHAR_LF) eptr++;
2447        break;        break;
2448    
2449        case 0x000a:        case CHAR_LF:
2450        break;        break;
2451    
2452        case 0x000b:        case CHAR_VT:
2453        case 0x000c:        case CHAR_FF:
2454        case 0x0085:        case CHAR_NEL:
2455    #ifndef EBCDIC
2456        case 0x2028:        case 0x2028:
2457        case 0x2029:        case 0x2029:
2458    #endif  /* Not EBCDIC */
2459        if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);        if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);
2460        break;        break;
2461        }        }
# Line 2391  for (;;) Line 2471  for (;;)
2471      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2472      switch(c)      switch(c)
2473        {        {
2474          HSPACE_CASES: RRETURN(MATCH_NOMATCH);  /* Byte and multibyte cases */
2475        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);  
2476        }        }
2477      ecode++;      ecode++;
2478      break;      break;
# Line 2425  for (;;) Line 2486  for (;;)
2486      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2487      switch(c)      switch(c)
2488        {        {
2489          HSPACE_CASES: break;  /* Byte and multibyte cases */
2490        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;  
2491        }        }
2492      ecode++;      ecode++;
2493      break;      break;
# Line 2459  for (;;) Line 2501  for (;;)
2501      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2502      switch(c)      switch(c)
2503        {        {
2504          VSPACE_CASES: RRETURN(MATCH_NOMATCH);
2505        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);  
2506        }        }
2507      ecode++;      ecode++;
2508      break;      break;
# Line 2481  for (;;) Line 2516  for (;;)
2516      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2517      switch(c)      switch(c)
2518        {        {
2519          VSPACE_CASES: break;
2520        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;  
2521        }        }
2522      ecode++;      ecode++;
2523      break;      break;
# Line 2507  for (;;) Line 2535  for (;;)
2535        }        }
2536      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2537        {        {
2538          const pcre_uint32 *cp;
2539        const ucd_record *prop = GET_UCD(c);        const ucd_record *prop = GET_UCD(c);
2540    
2541        switch(ecode[1])        switch(ecode[1])
# Line 2545  for (;;) Line 2574  for (;;)
2574            RRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
2575          break;          break;
2576    
2577            /* Perl space used to exclude VT, but from Perl 5.18 it is included,
2578            which means that Perl space and POSIX space are now identical. PCRE
2579            was changed at release 8.34. */
2580    
2581          case PT_SPACE:    /* Perl space */          case PT_SPACE:    /* Perl space */
         if ((PRIV(ucp_gentype)[prop->chartype] == ucp_Z ||  
              c == CHAR_HT || c == CHAR_NL || c == CHAR_FF || c == CHAR_CR)  
                == (op == OP_NOTPROP))  
           RRETURN(MATCH_NOMATCH);  
         break;  
   
2582          case PT_PXSPACE:  /* POSIX space */          case PT_PXSPACE:  /* POSIX space */
2583          if ((PRIV(ucp_gentype)[prop->chartype] == ucp_Z ||          if ((PRIV(ucp_gentype)[prop->chartype] == ucp_Z ||
2584               c == CHAR_HT || c == CHAR_NL || c == CHAR_VT ||               c == CHAR_HT || c == CHAR_NL || c == CHAR_VT ||
# Line 2567  for (;;) Line 2594  for (;;)
2594            RRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
2595          break;          break;
2596    
2597            case PT_CLIST:
2598            cp = PRIV(ucd_caseless_sets) + ecode[2];
2599            for (;;)
2600              {
2601              if (c < *cp)
2602                { if (op == OP_PROP) { RRETURN(MATCH_NOMATCH); } else break; }
2603              if (c == *cp++)
2604                { if (op == OP_PROP) break; else { RRETURN(MATCH_NOMATCH); } }
2605              }
2606            break;
2607    
2608            case PT_UCNC:
2609            if ((c == CHAR_DOLLAR_SIGN || c == CHAR_COMMERCIAL_AT ||
2610                 c == CHAR_GRAVE_ACCENT || (c >= 0xa0 && c <= 0xd7ff) ||
2611                 c >= 0xe000) == (op == OP_NOTPROP))
2612              RRETURN(MATCH_NOMATCH);
2613            break;
2614    
2615          /* This should never occur */          /* This should never occur */
2616    
2617          default:          default:
# Line 2586  for (;;) Line 2631  for (;;)
2631        SCHECK_PARTIAL();        SCHECK_PARTIAL();
2632        RRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2633        }        }
2634      GETCHARINCTEST(c, eptr);      else
     if (UCD_CATEGORY(c) == ucp_M) RRETURN(MATCH_NOMATCH);  
     while (eptr < md->end_subject)  
2635        {        {
2636        int len = 1;        int lgb, rgb;
2637        if (!utf) c = *eptr; else { GETCHARLEN(c, eptr, len); }        GETCHARINCTEST(c, eptr);
2638        if (UCD_CATEGORY(c) != ucp_M) break;        lgb = UCD_GRAPHBREAK(c);
2639        eptr += len;        while (eptr < md->end_subject)
2640            {
2641            int len = 1;
2642            if (!utf) c = *eptr; else { GETCHARLEN(c, eptr, len); }
2643            rgb = UCD_GRAPHBREAK(c);
2644            if ((PRIV(ucp_gbtable)[lgb] & (1 << rgb)) == 0) break;
2645            lgb = rgb;
2646            eptr += len;
2647            }
2648        }        }
2649        CHECK_PARTIAL();
2650      ecode++;      ecode++;
2651      break;      break;
2652  #endif  #endif  /* SUPPORT_UCP */
2653    
2654    
2655      /* 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 2658  for (;;)
2658      similar code to character type repeats - written out again for speed.      similar code to character type repeats - written out again for speed.
2659      However, if the referenced string is the empty string, always treat      However, if the referenced string is the empty string, always treat
2660      it as matched, any number of times (otherwise there could be infinite      it as matched, any number of times (otherwise there could be infinite
2661      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:  
2662    
2663      (a) In the default, Perl-compatible state, set the length negative;      (a) In the default, Perl-compatible state, set the length negative;
2664      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 2668  for (;;)
2668      so that the back reference matches an empty string.      so that the back reference matches an empty string.
2669    
2670      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
2671      referenced subpattern. */      referenced subpattern.
2672    
2673        The OP_REF and OP_REFI opcodes are used for a reference to a numbered group
2674        or to a non-duplicated named group. For a duplicated named group, OP_DNREF
2675        and OP_DNREFI are used. In this case we must scan the list of groups to
2676        which the name refers, and use the first one that is set. */
2677    
2678        case OP_DNREF:
2679        case OP_DNREFI:
2680        caseless = op == OP_DNREFI;
2681          {
2682          int count = GET2(ecode, 1+IMM2_SIZE);
2683          pcre_uchar *slot = md->name_table + GET2(ecode, 1) * md->name_entry_size;
2684          ecode += 1 + 2*IMM2_SIZE;
2685    
2686          while (count-- > 0)
2687            {
2688            offset = GET2(slot, 0) << 1;
2689            if (offset < offset_top && md->offset_vector[offset] >= 0) break;
2690            slot += md->name_entry_size;
2691            }
2692          if (count < 0)
2693            length = (md->jscript_compat)? 0 : -1;
2694          else
2695            length = md->offset_vector[offset+1] - md->offset_vector[offset];
2696          }
2697        goto REF_REPEAT;
2698    
2699        case OP_REF:
2700        case OP_REFI:
2701        caseless = op == OP_REFI;
2702        offset = GET2(ecode, 1) << 1;               /* Doubled ref number */
2703        ecode += 1 + IMM2_SIZE;
2704      if (offset >= offset_top || md->offset_vector[offset] < 0)      if (offset >= offset_top || md->offset_vector[offset] < 0)
2705        length = (md->jscript_compat)? 0 : -1;        length = (md->jscript_compat)? 0 : -1;
2706      else      else
# Line 2633  for (;;) Line 2708  for (;;)
2708    
2709      /* Set up for repetition, or handle the non-repeated case */      /* Set up for repetition, or handle the non-repeated case */
2710    
2711        REF_REPEAT:
2712      switch (*ecode)      switch (*ecode)
2713        {        {
2714        case OP_CRSTAR:        case OP_CRSTAR:
# Line 2660  for (;;) Line 2736  for (;;)
2736        default:               /* No repeat follows */        default:               /* No repeat follows */
2737        if ((length = match_ref(offset, eptr, length, md, caseless)) < 0)        if ((length = match_ref(offset, eptr, length, md, caseless)) < 0)
2738          {          {
2739            if (length == -2) eptr = md->end_subject;   /* Partial match */
2740          CHECK_PARTIAL();          CHECK_PARTIAL();
2741          RRETURN(MATCH_NOMATCH);          RRETURN(MATCH_NOMATCH);
2742          }          }
# Line 2685  for (;;) Line 2762  for (;;)
2762        int slength;        int slength;
2763        if ((slength = match_ref(offset, eptr, length, md, caseless)) < 0)        if ((slength = match_ref(offset, eptr, length, md, caseless)) < 0)
2764          {          {
2765            if (slength == -2) eptr = md->end_subject;   /* Partial match */
2766          CHECK_PARTIAL();          CHECK_PARTIAL();
2767          RRETURN(MATCH_NOMATCH);          RRETURN(MATCH_NOMATCH);
2768          }          }
# Line 2708  for (;;) Line 2786  for (;;)
2786          if (fi >= max) RRETURN(MATCH_NOMATCH);          if (fi >= max) RRETURN(MATCH_NOMATCH);
2787          if ((slength = match_ref(offset, eptr, length, md, caseless)) < 0)          if ((slength = match_ref(offset, eptr, length, md, caseless)) < 0)
2788            {            {
2789              if (slength == -2) eptr = md->end_subject;   /* Partial match */
2790            CHECK_PARTIAL();            CHECK_PARTIAL();
2791            RRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
2792            }            }
# Line 2726  for (;;) Line 2805  for (;;)
2805          int slength;          int slength;
2806          if ((slength = match_ref(offset, eptr, length, md, caseless)) < 0)          if ((slength = match_ref(offset, eptr, length, md, caseless)) < 0)
2807            {            {
2808            CHECK_PARTIAL();            /* Can't use CHECK_PARTIAL because we don't want to update eptr in
2809              the soft partial matching case. */
2810    
2811              if (slength == -2 && md->partial != 0 &&
2812                  md->end_subject > md->start_used_ptr)
2813                {
2814                md->hitend = TRUE;
2815                if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
2816                }
2817            break;            break;
2818            }            }
2819          eptr += slength;          eptr += slength;
2820          }          }
2821    
2822        while (eptr >= pp)        while (eptr >= pp)
2823          {          {
2824          RMATCH(eptr, ecode, offset_top, md, eptrb, RM15);          RMATCH(eptr, ecode, offset_top, md, eptrb, RM15);
# Line 3094  for (;;) Line 3182  for (;;)
3182          CHECK_PARTIAL();             /* Not SCHECK_PARTIAL() */          CHECK_PARTIAL();             /* Not SCHECK_PARTIAL() */
3183          RRETURN(MATCH_NOMATCH);          RRETURN(MATCH_NOMATCH);
3184          }          }
3185        while (length-- > 0) if (*ecode++ != *eptr++) RRETURN(MATCH_NOMATCH);        while (length-- > 0) if (*ecode++ != RAWUCHARINC(eptr)) RRETURN(MATCH_NOMATCH);
3186        }        }
3187      else      else
3188  #endif  #endif
# Line 3134  for (;;) Line 3222  for (;;)
3222    
3223        if (fc < 128)        if (fc < 128)
3224          {          {
3225          if (md->lcc[fc]          pcre_uint32 cc = RAWUCHAR(eptr);
3226              != TABLE_GET(*eptr, md->lcc, *eptr)) RRETURN(MATCH_NOMATCH);          if (md->lcc[fc] != TABLE_GET(cc, md->lcc, cc)) RRETURN(MATCH_NOMATCH);
3227          ecode++;          ecode++;
3228          eptr++;          eptr++;
3229          }          }
# Line 3146  for (;;) Line 3234  for (;;)
3234    
3235        else        else
3236          {          {
3237          unsigned int dc;          pcre_uint32 dc;
3238          GETCHARINC(dc, eptr);          GETCHARINC(dc, eptr);
3239          ecode += length;          ecode += length;
3240    
# Line 3239  for (;;) Line 3327  for (;;)
3327      max = rep_max[c];                 /* zero for max => infinity */      max = rep_max[c];                 /* zero for max => infinity */
3328      if (max == 0) max = INT_MAX;      if (max == 0) max = INT_MAX;
3329    
3330      /* Common code for all repeated single-character matches. */      /* Common code for all repeated single-character matches. We first check
3331        for the minimum number of characters. If the minimum equals the maximum, we
3332        are done. Otherwise, if minimizing, check the rest of the pattern for a
3333        match; if there isn't one, advance up to the maximum, one character at a
3334        time.
3335    
3336        If maximizing, advance up to the maximum number of matching characters,
3337        until eptr is past the end of the maximum run. If possessive, we are
3338        then done (no backing up). Otherwise, match at this position; anything
3339        other than no match is immediately returned. For nomatch, back up one
3340        character, unless we are matching \R and the last thing matched was
3341        \r\n, in which case, back up two bytes. When we reach the first optional
3342        character position, we can save stack by doing a tail recurse.
3343    
3344        The various UTF/non-UTF and caseful/caseless cases are handled separately,
3345        for speed. */
3346    
3347      REPEATCHAR:      REPEATCHAR:
3348  #ifdef SUPPORT_UTF  #ifdef SUPPORT_UTF
# Line 3256  for (;;) Line 3359  for (;;)
3359        if (length > 1)        if (length > 1)
3360          {          {
3361  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
3362          unsigned int othercase;          pcre_uint32 othercase;
3363          if (op >= OP_STARI &&     /* Caseless */          if (op >= OP_STARI &&     /* Caseless */
3364              (othercase = UCD_OTHERCASE(fc)) != fc)              (othercase = UCD_OTHERCASE(fc)) != fc)
3365            oclength = PRIV(ord2utf)(othercase, occhars);            oclength = PRIV(ord2utf)(othercase, occhars);
# Line 3323  for (;;) Line 3426  for (;;)
3426                }                }
3427              }              }
3428    
3429            if (possessive) continue;            if (possessive) continue;    /* No backtracking */
   
3430            for(;;)            for(;;)
3431              {              {
3432                if (eptr == pp) goto TAIL_RECURSE;
3433              RMATCH(eptr, ecode, offset_top, md, eptrb, RM23);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM23);
3434              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
             if (eptr == pp) { RRETURN(MATCH_NOMATCH); }  
3435  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
3436              eptr--;              eptr--;
3437              BACKCHAR(eptr);              BACKCHAR(eptr);
# Line 3360  for (;;) Line 3462  for (;;)
3462      maximizing, find the maximum number of characters and work backwards. */      maximizing, find the maximum number of characters and work backwards. */
3463    
3464      DPRINTF(("matching %c{%d,%d} against subject %.*s\n", fc, min, max,      DPRINTF(("matching %c{%d,%d} against subject %.*s\n", fc, min, max,
3465        max, eptr));        max, (char *)eptr));
3466    
3467      if (op >= OP_STARI)  /* Caseless */      if (op >= OP_STARI)  /* Caseless */
3468        {        {
# Line 3383  for (;;) Line 3485  for (;;)
3485    
3486        for (i = 1; i <= min; i++)        for (i = 1; i <= min; i++)
3487          {          {
3488            pcre_uint32 cc;                 /* Faster than pcre_uchar */
3489          if (eptr >= md->end_subject)          if (eptr >= md->end_subject)
3490            {            {
3491            SCHECK_PARTIAL();            SCHECK_PARTIAL();
3492            RRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
3493            }            }
3494          if (fc != *eptr && foc != *eptr) RRETURN(MATCH_NOMATCH);          cc = RAWUCHARTEST(eptr);
3495            if (fc != cc && foc != cc) RRETURN(MATCH_NOMATCH);
3496          eptr++;          eptr++;
3497          }          }
3498        if (min == max) continue;        if (min == max) continue;
# Line 3396  for (;;) Line 3500  for (;;)
3500          {          {
3501          for (fi = min;; fi++)          for (fi = min;; fi++)
3502            {            {
3503              pcre_uint32 cc;               /* Faster than pcre_uchar */
3504            RMATCH(eptr, ecode, offset_top, md, eptrb, RM24);            RMATCH(eptr, ecode, offset_top, md, eptrb, RM24);
3505            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3506            if (fi >= max) RRETURN(MATCH_NOMATCH);            if (fi >= max) RRETURN(MATCH_NOMATCH);
# Line 3404  for (;;) Line 3509  for (;;)
3509              SCHECK_PARTIAL();              SCHECK_PARTIAL();
3510              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
3511              }              }
3512            if (fc != *eptr && foc != *eptr) RRETURN(MATCH_NOMATCH);            cc = RAWUCHARTEST(eptr);
3513              if (fc != cc && foc != cc) RRETURN(MATCH_NOMATCH);
3514            eptr++;            eptr++;
3515            }            }
3516          /* Control never gets here */          /* Control never gets here */
# Line 3414  for (;;) Line 3520  for (;;)
3520          pp = eptr;          pp = eptr;
3521          for (i = min; i < max; i++)          for (i = min; i < max; i++)
3522            {            {
3523              pcre_uint32 cc;               /* Faster than pcre_uchar */
3524            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
3525              {              {
3526              SCHECK_PARTIAL();              SCHECK_PARTIAL();
3527              break;              break;
3528              }              }
3529            if (fc != *eptr && foc != *eptr) break;            cc = RAWUCHARTEST(eptr);
3530              if (fc != cc && foc != cc) break;
3531            eptr++;            eptr++;
3532            }            }
3533            if (possessive) continue;       /* No backtracking */
3534          if (possessive) continue;          for (;;)
   
         while (eptr >= pp)  
3535            {            {
3536              if (eptr == pp) goto TAIL_RECURSE;
3537            RMATCH(eptr, ecode, offset_top, md, eptrb, RM25);            RMATCH(eptr, ecode, offset_top, md, eptrb, RM25);
3538            eptr--;            eptr--;
3539            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3540            }            }
3541          RRETURN(MATCH_NOMATCH);          /* Control never gets here */
3542          }          }
       /* Control never gets here */  
3543        }        }
3544    
3545      /* Caseful comparisons (includes all multi-byte characters) */      /* Caseful comparisons (includes all multi-byte characters) */
# Line 3447  for (;;) Line 3553  for (;;)
3553            SCHECK_PARTIAL();            SCHECK_PARTIAL();
3554            RRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
3555            }            }
3556          if (fc != *eptr++) RRETURN(MATCH_NOMATCH);          if (fc != RAWUCHARINCTEST(eptr)) RRETURN(MATCH_NOMATCH);
3557          }          }
3558    
3559        if (min == max) continue;        if (min == max) continue;
# Line 3464  for (;;) Line 3570  for (;;)
3570              SCHECK_PARTIAL();              SCHECK_PARTIAL();
3571              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
3572              }              }
3573            if (fc != *eptr++) RRETURN(MATCH_NOMATCH);            if (fc != RAWUCHARINCTEST(eptr)) RRETURN(MATCH_NOMATCH);
3574            }            }
3575          /* Control never gets here */          /* Control never gets here */
3576          }          }
# Line 3478  for (;;) Line 3584  for (;;)
3584              SCHECK_PARTIAL();              SCHECK_PARTIAL();
3585              break;              break;
3586              }              }
3587            if (fc != *eptr) break;            if (fc != RAWUCHARTEST(eptr)) break;
3588            eptr++;            eptr++;
3589            }            }
3590          if (possessive) continue;          if (possessive) continue;    /* No backtracking */
3591            for (;;)
         while (eptr >= pp)  
3592            {            {
3593              if (eptr == pp) goto TAIL_RECURSE;
3594            RMATCH(eptr, ecode, offset_top, md, eptrb, RM27);            RMATCH(eptr, ecode, offset_top, md, eptrb, RM27);
3595            eptr--;            eptr--;
3596            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3597            }            }
3598          RRETURN(MATCH_NOMATCH);          /* Control never gets here */
3599          }          }
3600        }        }
3601      /* Control never gets here */      /* Control never gets here */
# Line 3504  for (;;) Line 3610  for (;;)
3610        SCHECK_PARTIAL();        SCHECK_PARTIAL();
3611        RRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
3612        }        }
     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  
3613  #ifdef SUPPORT_UTF  #ifdef SUPPORT_UTF
3614        if (utf)
3615          {
3616          register pcre_uint32 ch, och;
3617    
3618          ecode++;
3619          GETCHARINC(ch, ecode);
3620          GETCHARINC(c, eptr);
3621    
3622          if (op == OP_NOT)
3623            {
3624            if (ch == c) RRETURN(MATCH_NOMATCH);
3625            }
3626          else
3627            {
3628  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
3629        if (utf && ch > 127)          if (ch > 127)
3630          och = UCD_OTHERCASE(ch);            och = UCD_OTHERCASE(ch);
3631  #else  #else
3632        if (utf && ch > 127)          if (ch > 127)
3633          och = ch;            och = ch;
3634  #endif /* SUPPORT_UCP */  #endif /* SUPPORT_UCP */
3635        else          else
3636  #endif /* SUPPORT_UTF */            och = TABLE_GET(ch, md->fcc, ch);
3637          och = TABLE_GET(ch, md->fcc, ch);          if (ch == c || och == c) RRETURN(MATCH_NOMATCH);
3638  #endif /* COMPILE_PCRE8 */          }
       if (ch == c || och == c) RRETURN(MATCH_NOMATCH);  
3639        }        }
3640      else    /* Caseful */      else
3641    #endif
3642        {        {
3643        if (*ecode++ == c) RRETURN(MATCH_NOMATCH);        register pcre_uint32 ch = ecode[1];
3644          c = *eptr++;
3645          if (ch == c || (op == OP_NOTI && TABLE_GET(ch, md->fcc, ch) == c))
3646            RRETURN(MATCH_NOMATCH);
3647          ecode += 2;
3648        }        }
3649      break;      break;
3650    
# Line 3610  for (;;) Line 3724  for (;;)
3724      /* Common code for all repeated single-byte matches. */      /* Common code for all repeated single-byte matches. */
3725    
3726      REPEATNOTCHAR:      REPEATNOTCHAR:
3727      fc = *ecode++;      GETCHARINCTEST(fc, ecode);
3728    
3729      /* The code is duplicated for the caseless and caseful cases, for speed,      /* The code is duplicated for the caseless and caseful cases, for speed,
3730      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 3735  for (;;)
3735      characters and work backwards. */      characters and work backwards. */
3736    
3737      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,
3738        max, eptr));        max, (char *)eptr));
3739    
3740      if (op >= OP_NOTSTARI)     /* Caseless */      if (op >= OP_NOTSTARI)     /* Caseless */
3741        {        {
 #ifdef COMPILE_PCRE8  
       /* fc must be < 128 if UTF is enabled. */  
       foc = md->fcc[fc];  
 #else  
3742  #ifdef SUPPORT_UTF  #ifdef SUPPORT_UTF
3743  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
3744        if (utf && fc > 127)        if (utf && fc > 127)
# Line 3640  for (;;) Line 3750  for (;;)
3750        else        else
3751  #endif /* SUPPORT_UTF */  #endif /* SUPPORT_UTF */
3752          foc = TABLE_GET(fc, md->fcc, fc);          foc = TABLE_GET(fc, md->fcc, fc);
 #endif /* COMPILE_PCRE8 */  
3753    
3754  #ifdef SUPPORT_UTF  #ifdef SUPPORT_UTF
3755        if (utf)        if (utf)
3756          {          {
3757          register unsigned int d;          register pcre_uint32 d;
3758          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
3759            {            {
3760            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
# Line 3654  for (;;) Line 3763  for (;;)
3763              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
3764              }              }
3765            GETCHARINC(d, eptr);            GETCHARINC(d, eptr);
3766            if (fc == d || foc == d) RRETURN(MATCH_NOMATCH);            if (fc == d || (unsigned int)foc == d) RRETURN(MATCH_NOMATCH);
3767            }            }
3768          }          }
3769        else        else
3770  #endif  #endif  /* SUPPORT_UTF */
3771        /* Not UTF mode */        /* Not UTF mode */
3772          {          {
3773          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
# Line 3680  for (;;) Line 3789  for (;;)
3789  #ifdef SUPPORT_UTF  #ifdef SUPPORT_UTF
3790          if (utf)          if (utf)
3791            {            {
3792            register unsigned int d;            register pcre_uint32 d;
3793            for (fi = min;; fi++)            for (fi = min;; fi++)
3794              {              {
3795              RMATCH(eptr, ecode, offset_top, md, eptrb, RM28);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM28);
# Line 3692  for (;;) Line 3801  for (;;)
3801                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
3802                }                }
3803              GETCHARINC(d, eptr);              GETCHARINC(d, eptr);
3804              if (fc == d || foc == d) RRETURN(MATCH_NOMATCH);              if (fc == d || (unsigned int)foc == d) RRETURN(MATCH_NOMATCH);
3805              }              }
3806            }            }
3807          else          else
3808  #endif  #endif  /*SUPPORT_UTF */
3809          /* Not UTF mode */          /* Not UTF mode */
3810            {            {
3811            for (fi = min;; fi++)            for (fi = min;; fi++)
# Line 3725  for (;;) Line 3834  for (;;)
3834  #ifdef SUPPORT_UTF  #ifdef SUPPORT_UTF
3835          if (utf)          if (utf)
3836            {            {
3837            register unsigned int d;            register pcre_uint32 d;
3838            for (i = min; i < max; i++)            for (i = min; i < max; i++)
3839              {              {
3840              int len = 1;              int len = 1;
# Line 3735  for (;;) Line 3844  for (;;)
3844                break;                break;
3845                }                }
3846              GETCHARLEN(d, eptr, len);              GETCHARLEN(d, eptr, len);
3847              if (fc == d || foc == d) break;              if (fc == d || (unsigned int)foc == d) break;
3848              eptr += len;              eptr += len;
3849              }              }
3850            if (possessive) continue;            if (possessive) continue;    /* No backtracking */
3851            for(;;)            for(;;)
3852              {              {
3853                if (eptr == pp) goto TAIL_RECURSE;
3854              RMATCH(eptr, ecode, offset_top, md, eptrb, RM30);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM30);
3855              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3856              if (eptr-- == pp) break;        /* Stop if tried at original pos */              eptr--;
3857              BACKCHAR(eptr);              BACKCHAR(eptr);
3858              }              }
3859            }            }
3860          else          else
3861  #endif  #endif  /* SUPPORT_UTF */
3862          /* Not UTF mode */          /* Not UTF mode */
3863            {            {
3864            for (i = min; i < max; i++)            for (i = min; i < max; i++)
# Line 3761  for (;;) Line 3871  for (;;)
3871              if (fc == *eptr || foc == *eptr) break;              if (fc == *eptr || foc == *eptr) break;
3872              eptr++;              eptr++;
3873              }              }
3874            if (possessive) continue;            if (possessive) continue;    /* No backtracking */
3875            while (eptr >= pp)            for (;;)
3876              {              {
3877                if (eptr == pp) goto TAIL_RECURSE;
3878              RMATCH(eptr, ecode, offset_top, md, eptrb, RM31);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM31);
3879              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3880              eptr--;              eptr--;
3881              }              }
3882            }            }
3883            /* Control never gets here */
         RRETURN(MATCH_NOMATCH);  
3884          }          }
       /* Control never gets here */  
3885        }        }
3886    
3887      /* Caseful comparisons */      /* Caseful comparisons */
# Line 3782  for (;;) Line 3891  for (;;)
3891  #ifdef SUPPORT_UTF  #ifdef SUPPORT_UTF
3892        if (utf)        if (utf)
3893          {          {
3894          register unsigned int d;          register pcre_uint32 d;
3895          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
3896            {            {
3897            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
# Line 3816  for (;;) Line 3925  for (;;)
3925  #ifdef SUPPORT_UTF  #ifdef SUPPORT_UTF
3926          if (utf)          if (utf)
3927            {            {
3928            register unsigned int d;            register pcre_uint32 d;
3929            for (fi = min;; fi++)            for (fi = min;; fi++)
3930              {              {
3931              RMATCH(eptr, ecode, offset_top, md, eptrb, RM32);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM32);
# Line 3860  for (;;) Line 3969  for (;;)
3969  #ifdef SUPPORT_UTF  #ifdef SUPPORT_UTF
3970          if (utf)          if (utf)
3971            {            {
3972            register unsigned int d;            register pcre_uint32 d;
3973            for (i = min; i < max; i++)            for (i = min; i < max; i++)
3974              {              {
3975              int len = 1;              int len = 1;
# Line 3873  for (;;) Line 3982  for (;;)
3982              if (fc == d) break;              if (fc == d) break;
3983              eptr += len;              eptr += len;
3984              }              }
3985            if (possessive) continue;            if (possessive) continue;    /* No backtracking */
3986            for(;;)            for(;;)
3987              {              {
3988                if (eptr == pp) goto TAIL_RECURSE;
3989              RMATCH(eptr, ecode, offset_top, md, eptrb, RM34);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM34);
3990              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3991              if (eptr-- == pp) break;        /* Stop if tried at original pos */              eptr--;
3992              BACKCHAR(eptr);              BACKCHAR(eptr);
3993              }              }
3994            }            }
# Line 3896  for (;;) Line 4006  for (;;)
4006              if (fc == *eptr) break;              if (fc == *eptr) break;
4007              eptr++;              eptr++;
4008              }              }
4009            if (possessive) continue;            if (possessive) continue;    /* No backtracking */
4010            while (eptr >= pp)            for (;;)
4011              {              {
4012                if (eptr == pp) goto TAIL_RECURSE;
4013              RMATCH(eptr, ecode, offset_top, md, eptrb, RM35);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM35);
4014              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
4015              eptr--;              eptr--;
4016              }              }
4017            }            }
4018            /* Control never gets here */
         RRETURN(MATCH_NOMATCH);  
4019          }          }
4020        }        }
4021      /* Control never gets here */      /* Control never gets here */
# Line 4087  for (;;) Line 4197  for (;;)
4197              }              }
4198            break;            break;
4199    
4200              /* Perl space used to exclude VT, but from Perl 5.18 it is included,
4201              which means that Perl space and POSIX space are now identical. PCRE
4202              was changed at release 8.34. */
4203    
4204            case PT_SPACE:    /* Perl space */            case PT_SPACE:    /* Perl space */
4205              case PT_PXSPACE:  /* POSIX space */
4206            for (i = 1; i <= min; i++)            for (i = 1; i <= min; i++)
4207              {              {
4208              if (eptr >= md->end_subject)              if (eptr >= md->end_subject)
# Line 4097  for (;;) Line 4212  for (;;)
4212                }                }
4213              GETCHARINCTEST(c, eptr);              GETCHARINCTEST(c, eptr);
4214              if ((UCD_CATEGORY(c) == ucp_Z || c == CHAR_HT || c == CHAR_NL ||              if ((UCD_CATEGORY(c) == ucp_Z || c == CHAR_HT || c == CHAR_NL ||
4215                   c == CHAR_FF || c == CHAR_CR)                   c == CHAR_VT || c == CHAR_FF || c == CHAR_CR)
4216                     == prop_fail_result)                     == prop_fail_result)
4217                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
4218              }              }
4219            break;            break;
4220    
4221            case PT_PXSPACE:  /* POSIX space */            case PT_WORD:
4222            for (i = 1; i <= min; i++)            for (i = 1; i <= min; i++)
4223              {              {
4224                int category;
4225              if (eptr >= md->end_subject)              if (eptr >= md->end_subject)
4226                {                {
4227                SCHECK_PARTIAL();                SCHECK_PARTIAL();
4228                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
4229                }                }
4230              GETCHARINCTEST(c, eptr);              GETCHARINCTEST(c, eptr);
4231              if ((UCD_CATEGORY(c) == ucp_Z || c == CHAR_HT || c == CHAR_NL ||              category = UCD_CATEGORY(c);
4232                   c == CHAR_VT || c == CHAR_FF || c == CHAR_CR)              if ((category == ucp_L || category == ucp_N || c == CHAR_UNDERSCORE)
4233                     == prop_fail_result)                     == prop_fail_result)
4234                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
4235              }              }
4236            break;            break;
4237    
4238            case PT_WORD:            case PT_CLIST:
4239            for (i = 1; i <= min; i++)            for (i = 1; i <= min; i++)
4240              {              {
4241              int category;              const pcre_uint32 *cp;
4242              if (eptr >= md->end_subject)              if (eptr >= md->end_subject)
4243                {                {
4244                SCHECK_PARTIAL();                SCHECK_PARTIAL();
4245                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
4246                }                }
4247              GETCHARINCTEST(c, eptr);              GETCHARINCTEST(c, eptr);
4248              category = UCD_CATEGORY(c);              cp = PRIV(ucd_caseless_sets) + prop_value;
4249              if ((category == ucp_L || category == ucp_N || c == CHAR_UNDERSCORE)              for (;;)
4250                     == prop_fail_result)                {
4251                  if (c < *cp)
4252                    { if (prop_fail_result) break; else { RRETURN(MATCH_NOMATCH); } }
4253                  if (c == *cp++)
4254                    { if (prop_fail_result) { RRETURN(MATCH_NOMATCH); } else break; }
4255                  }
4256                }
4257              break;
4258    
4259              case PT_UCNC:
4260              for (i = 1; i <= min; i++)
4261                {
4262                if (eptr >= md->end_subject)
4263                  {
4264                  SCHECK_PARTIAL();
4265                  RRETURN(MATCH_NOMATCH);
4266                  }
4267                GETCHARINCTEST(c, eptr);
4268                if ((c == CHAR_DOLLAR_SIGN || c == CHAR_COMMERCIAL_AT ||
4269                     c == CHAR_GRAVE_ACCENT || (c >= 0xa0 && c <= 0xd7ff) ||
4270                     c >= 0xe000) == prop_fail_result)
4271                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
4272              }              }
4273            break;            break;
# Line 4155  for (;;) Line 4291  for (;;)
4291              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4292              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4293              }              }
4294            GETCHARINCTEST(c, eptr);            else
           if (UCD_CATEGORY(c) == ucp_M) RRETURN(MATCH_NOMATCH);  
           while (eptr < md->end_subject)  
4295              {              {
4296              int len = 1;              int lgb, rgb;
4297              if (!utf) c = *eptr; else { GETCHARLEN(c, eptr, len); }              GETCHARINCTEST(c, eptr);
4298              if (UCD_CATEGORY(c) != ucp_M) break;              lgb = UCD_GRAPHBREAK(c);
4299              eptr += len;             while (eptr < md->end_subject)
4300                  {
4301                  int len = 1;
4302                  if (!utf) c = *eptr; else { GETCHARLEN(c, eptr, len); }
4303                  rgb = UCD_GRAPHBREAK(c);
4304                  if ((PRIV(ucp_gbtable)[lgb] & (1 << rgb)) == 0) break;
4305                  lgb = rgb;
4306                  eptr += len;
4307                  }
4308              }              }
4309              CHECK_PARTIAL();
4310            }            }
4311          }          }
4312    
# Line 4184  for (;;) Line 4327  for (;;)
4327              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4328              }              }
4329            if (IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH);            if (IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH);
4330              if (md->partial != 0 &&
4331                  eptr + 1 >= md->end_subject &&
4332                  NLBLOCK->nltype == NLTYPE_FIXED &&
4333                  NLBLOCK->nllen == 2 &&
4334                  RAWUCHAR(eptr) == NLBLOCK->nl[0])
4335                {
4336                md->hitend = TRUE;
4337                if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
4338                }
4339            eptr++;            eptr++;
4340            ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);            ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);
4341            }            }
# Line 4220  for (;;) Line 4372  for (;;)
4372              {              {
4373              default: RRETURN(MATCH_NOMATCH);              default: RRETURN(MATCH_NOMATCH);
4374    
4375              case 0x000d:              case CHAR_CR:
4376              if (eptr < md->end_subject && *eptr == 0x0a) eptr++;              if (eptr < md->end_subject && RAWUCHAR(eptr) == CHAR_LF) eptr++;
4377              break;              break;
4378    
4379              case 0x000a:              case CHAR_LF:
4380              break;              break;
4381    
4382              case 0x000b:              case CHAR_VT:
4383              case 0x000c:              case CHAR_FF:
4384              case 0x0085:              case CHAR_NEL:
4385    #ifndef EBCDIC
4386              case 0x2028:              case 0x2028:
4387              case 0x2029:              case 0x2029:
4388    #endif  /* Not EBCDIC */
4389              if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);              if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);
4390              break;              break;
4391              }              }
# Line 4249  for (;;) Line 4403  for (;;)
4403            GETCHARINC(c, eptr);            GETCHARINC(c, eptr);
4404            switch(c)            switch(c)
4405              {              {
4406                HSPACE_CASES: RRETURN(MATCH_NOMATCH);  /* Byte and multibyte cases */
4407              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);  
4408              }              }
4409            }            }
4410          break;          break;
# Line 4285  for (;;) Line 4420  for (;;)
4420            GETCHARINC(c, eptr);            GETCHARINC(c, eptr);
4421            switch(c)            switch(c)
4422              {              {
4423                HSPACE_CASES: break;  /* Byte and multibyte cases */
4424              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;  
4425              }              }
4426            }            }
4427          break;          break;
# Line 4321  for (;;) Line 4437  for (;;)
4437            GETCHARINC(c, eptr);            GETCHARINC(c, eptr);
4438            switch(c)            switch(c)
4439              {              {
4440                VSPACE_CASES: RRETURN(MATCH_NOMATCH);
4441              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);  
4442              }              }
4443            }            }
4444          break;          break;
# Line 4345  for (;;) Line 4454  for (;;)
4454            GETCHARINC(c, eptr);            GETCHARINC(c, eptr);
4455            switch(c)            switch(c)
4456              {              {
4457                VSPACE_CASES: break;
4458              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;  
4459              }              }
4460            }            }
4461          break;          break;
# Line 4375  for (;;) Line 4477  for (;;)
4477          case OP_DIGIT:          case OP_DIGIT:
4478          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
4479            {            {
4480              pcre_uint32 cc;
4481            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
4482              {              {
4483              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4484              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4485              }              }
4486            if (*eptr >= 128 || (md->ctypes[*eptr] & ctype_digit) == 0)            cc = RAWUCHAR(eptr);
4487              if (cc >= 128 || (md->ctypes[cc] & ctype_digit) == 0)
4488              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4489            eptr++;            eptr++;
4490            /* 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 4494  for (;;)
4494          case OP_NOT_WHITESPACE:          case OP_NOT_WHITESPACE:
4495          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
4496            {            {
4497              pcre_uint32 cc;
4498            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
4499              {              {
4500              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4501              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4502              }              }
4503            if (*eptr < 128 && (md->ctypes[*eptr] & ctype_space) != 0)            cc = RAWUCHAR(eptr);
4504              if (cc < 128 && (md->ctypes[cc] & ctype_space) != 0)
4505              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4506            eptr++;            eptr++;
4507            ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);            ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);
# Line 4405  for (;;) Line 4511  for (;;)
4511          case OP_WHITESPACE:          case OP_WHITESPACE:
4512          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
4513            {            {
4514              pcre_uint32 cc;
4515            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
4516              {              {
4517              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4518              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4519              }              }
4520            if (*eptr >= 128 || (md->ctypes[*eptr] & ctype_space) == 0)            cc = RAWUCHAR(eptr);
4521              if (cc >= 128 || (md->ctypes[cc] & ctype_space) == 0)
4522              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4523            eptr++;            eptr++;
4524            /* 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 4528  for (;;)
4528          case OP_NOT_WORDCHAR:          case OP_NOT_WORDCHAR:
4529          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
4530            {            {
4531              pcre_uint32 cc;
4532            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
4533              {              {
4534              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4535              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4536              }              }
4537            if (*eptr < 128 && (md->ctypes[*eptr] & ctype_word) != 0)            cc = RAWUCHAR(eptr);
4538              if (cc < 128 && (md->ctypes[cc] & ctype_word) != 0)
4539              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4540            eptr++;            eptr++;
4541            ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);            ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);
# Line 4435  for (;;) Line 4545  for (;;)
4545          case OP_WORDCHAR:          case OP_WORDCHAR:
4546          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
4547            {            {
4548              pcre_uint32 cc;
4549            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
4550              {              {
4551              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4552              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4553              }              }
4554            if (*eptr >= 128 || (md->ctypes[*eptr] & ctype_word) == 0)            cc = RAWUCHAR(eptr);
4555              if (cc >= 128 || (md->ctypes[cc] & ctype_word) == 0)
4556              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4557            eptr++;            eptr++;
4558            /* 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 4580  for (;;)
4580              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4581              }              }
4582            if (IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH);            if (IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH);
4583              if (md->partial != 0 &&
4584                  eptr + 1 >= md->end_subject &&
4585                  NLBLOCK->nltype == NLTYPE_FIXED &&
4586                  NLBLOCK->nllen == 2 &&
4587                  *eptr == NLBLOCK->nl[0])
4588                {
4589                md->hitend = TRUE;
4590                if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
4591                }
4592            eptr++;            eptr++;
4593            }            }
4594          break;          break;
# Line 4502  for (;;) Line 4623  for (;;)
4623              {              {
4624              default: RRETURN(MATCH_NOMATCH);              default: RRETURN(MATCH_NOMATCH);
4625    
4626              case 0x000d:              case CHAR_CR:
4627              if (eptr < md->end_subject && *eptr == 0x0a) eptr++;              if (eptr < md->end_subject && *eptr == CHAR_LF) eptr++;
4628              break;              break;
4629    
4630              case 0x000a:              case CHAR_LF:
4631              break;              break;
4632    
4633              case 0x000b:              case CHAR_VT:
4634              case 0x000c:              case CHAR_FF:
4635              case 0x0085:              case CHAR_NEL:
4636  #ifdef COMPILE_PCRE16  #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
4637              case 0x2028:              case 0x2028:
4638              case 0x2029:              case 0x2029:
4639  #endif  #endif
# Line 4533  for (;;) Line 4654  for (;;)
4654            switch(*eptr++)            switch(*eptr++)
4655              {              {
4656              default: break;              default: break;
4657              case 0x09:      /* HT */              HSPACE_BYTE_CASES:
4658              case 0x20:      /* SPACE */  #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
4659              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 */  
4660  #endif  #endif
4661              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4662              }              }
# Line 4570  for (;;) Line 4674  for (;;)
4674            switch(*eptr++)            switch(*eptr++)
4675              {              {
4676              default: RRETURN(MATCH_NOMATCH);              default: RRETURN(MATCH_NOMATCH);
4677              case 0x09:      /* HT */              HSPACE_BYTE_CASES:
4678              case 0x20:      /* SPACE */  #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
4679              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 */  
4680  #endif  #endif
4681              break;              break;
4682              }              }
# Line 4606  for (;;) Line 4693  for (;;)
4693              }              }
4694            switch(*eptr++)            switch(*eptr++)
4695              {              {
4696              default: break;              VSPACE_BYTE_CASES:
4697              case 0x0a:      /* LF */  #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
4698              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 */  
4699  #endif  #endif
4700              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4701                default: break;
4702              }              }
4703            }            }
4704          break;          break;
# Line 4632  for (;;) Line 4714  for (;;)
4714            switch(*eptr++)            switch(*eptr++)
4715              {              {
4716              default: RRETURN(MATCH_NOMATCH);              default: RRETURN(MATCH_NOMATCH);
4717              case 0x0a:      /* LF */              VSPACE_BYTE_CASES:
4718              case 0x0b:      /* VT */  #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
4719              case 0x0c:      /* FF */              VSPACE_MULTIBYTE_CASES:
             case 0x0d:      /* CR */  
             case 0x85:      /* NEL */  
 #ifdef COMPILE_PCRE16  
             case 0x2028:    /* LINE SEPARATOR */  
             case 0x2029:    /* PARAGRAPH SEPARATOR */  
4720  #endif  #endif
4721              break;              break;
4722              }              }
# Line 4857  for (;;) Line 4934  for (;;)
4934              }              }
4935            /* Control never gets here */            /* Control never gets here */
4936    
4937              /* Perl space used to exclude VT, but from Perl 5.18 it is included,
4938              which means that Perl space and POSIX space are now identical. PCRE
4939              was changed at release 8.34. */
4940    
4941            case PT_SPACE:    /* Perl space */            case PT_SPACE:    /* Perl space */
4942              case PT_PXSPACE:  /* POSIX space */
4943            for (fi = min;; fi++)            for (fi = min;; fi++)
4944              {              {
4945              RMATCH(eptr, ecode, offset_top, md, eptrb, RM60);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM61);
4946              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
4947              if (fi >= max) RRETURN(MATCH_NOMATCH);              if (fi >= max) RRETURN(MATCH_NOMATCH);
4948              if (eptr >= md->end_subject)              if (eptr >= md->end_subject)
# Line 4870  for (;;) Line 4952  for (;;)
4952                }                }
4953              GETCHARINCTEST(c, eptr);              GETCHARINCTEST(c, eptr);
4954              if ((UCD_CATEGORY(c) == ucp_Z || c == CHAR_HT || c == CHAR_NL ||              if ((UCD_CATEGORY(c) == ucp_Z || c == CHAR_HT || c == CHAR_NL ||
4955                   c == CHAR_FF || c == CHAR_CR)                   c == CHAR_VT || c == CHAR_FF || c == CHAR_CR)
4956                     == prop_fail_result)                     == prop_fail_result)
4957                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
4958              }              }
4959            /* Control never gets here */            /* Control never gets here */
4960    
4961            case PT_PXSPACE:  /* POSIX space */            case PT_WORD:
4962            for (fi = min;; fi++)            for (fi = min;; fi++)
4963              {              {
4964              RMATCH(eptr, ecode, offset_top, md, eptrb, RM61);              int category;
4965                RMATCH(eptr, ecode, offset_top, md, eptrb, RM62);
4966              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
4967              if (fi >= max) RRETURN(MATCH_NOMATCH);              if (fi >= max) RRETURN(MATCH_NOMATCH);
4968              if (eptr >= md->end_subject)              if (eptr >= md->end_subject)
# Line 4888  for (;;) Line 4971  for (;;)
4971                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
4972                }                }
4973              GETCHARINCTEST(c, eptr);              GETCHARINCTEST(c, eptr);
4974              if ((UCD_CATEGORY(c) == ucp_Z || c == CHAR_HT || c == CHAR_NL ||              category = UCD_CATEGORY(c);
4975                   c == CHAR_VT || c == CHAR_FF || c == CHAR_CR)              if ((category == ucp_L ||
4976                     category == ucp_N ||
4977                     c == CHAR_UNDERSCORE)
4978                     == prop_fail_result)                     == prop_fail_result)
4979                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
4980              }              }
4981            /* Control never gets here */            /* Control never gets here */
4982    
4983            case PT_WORD:            case PT_CLIST:
4984            for (fi = min;; fi++)            for (fi = min;; fi++)
4985              {              {
4986              int category;              const pcre_uint32 *cp;
4987              RMATCH(eptr, ecode, offset_top, md, eptrb, RM62);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM67);
4988              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
4989              if (fi >= max) RRETURN(MATCH_NOMATCH);              if (fi >= max) RRETURN(MATCH_NOMATCH);
4990              if (eptr >= md->end_subject)              if (eptr >= md->end_subject)
# Line 4908  for (;;) Line 4993  for (;;)
4993                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
4994                }                }
4995              GETCHARINCTEST(c, eptr);              GETCHARINCTEST(c, eptr);
4996              category = UCD_CATEGORY(c);              cp = PRIV(ucd_caseless_sets) + prop_value;
4997              if ((category == ucp_L ||              for (;;)
4998                   category == ucp_N ||                {
4999                   c == CHAR_UNDERSCORE)                if (c < *cp)
5000                     == prop_fail_result)                  { if (prop_fail_result) break; else { RRETURN(MATCH_NOMATCH); } }
5001                  if (c == *cp++)
5002                    { if (prop_fail_result) { RRETURN(MATCH_NOMATCH); } else break; }
5003                  }
5004                }
5005              /* Control never gets here */
5006    
5007              case PT_UCNC:
5008              for (fi = min;; fi++)
5009                {
5010                RMATCH(eptr, ecode, offset_top, md, eptrb, RM68);
5011                if (rrc != MATCH_NOMATCH) RRETURN(rrc);
5012                if (fi >= max) RRETURN(MATCH_NOMATCH);
5013                if (eptr >= md->end_subject)
5014                  {
5015                  SCHECK_PARTIAL();
5016                  RRETURN(MATCH_NOMATCH);
5017                  }
5018                GETCHARINCTEST(c, eptr);
5019                if ((c == CHAR_DOLLAR_SIGN || c == CHAR_COMMERCIAL_AT ||
5020                     c == CHAR_GRAVE_ACCENT || (c >= 0xa0 && c <= 0xd7ff) ||
5021                     c >= 0xe000) == prop_fail_result)
5022                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
5023              }              }
5024            /* Control never gets here */            /* Control never gets here */
5025    
5026            /* This should never occur */            /* This should never occur */
   
5027            default:            default:
5028            RRETURN(PCRE_ERROR_INTERNAL);            RRETURN(PCRE_ERROR_INTERNAL);
5029            }            }
# Line 4939  for (;;) Line 5044  for (;;)
5044              SCHECK_PARTIAL();              SCHECK_PARTIAL();
5045              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
5046              }              }
5047            GETCHARINCTEST(c, eptr);            else
           if (UCD_CATEGORY(c) == ucp_M) RRETURN(MATCH_NOMATCH);  
           while (eptr < md->end_subject)  
5048              {              {
5049              int len = 1;              int lgb, rgb;
5050              if (!utf) c = *eptr; else { GETCHARLEN(c, eptr, len); }              GETCHARINCTEST(c, eptr);
5051              if (UCD_CATEGORY(c) != ucp_M) break;              lgb = UCD_GRAPHBREAK(c);
5052              eptr += len;              while (eptr < md->end_subject)
5053                  {
5054                  int len = 1;
5055                  if (!utf) c = *eptr; else { GETCHARLEN(c, eptr, len); }
5056                  rgb = UCD_GRAPHBREAK(c);
5057                  if ((PRIV(ucp_gbtable)[lgb] & (1 << rgb)) == 0) break;
5058                  lgb = rgb;
5059                  eptr += len;
5060                  }
5061              }              }
5062              CHECK_PARTIAL();
5063            }            }
5064          }          }
5065        else        else
# Line 4971  for (;;) Line 5083  for (;;)
5083            GETCHARINC(c, eptr);            GETCHARINC(c, eptr);
5084            switch(ctype)            switch(ctype)
5085              {              {
5086              case OP_ANY:        /* This is the non-NL case */              case OP_ANY:               /* This is the non-NL case */
5087                if (md->partial != 0 &&    /* Take care with CRLF partial */
5088                    eptr >= md->end_subject &&
5089                    NLBLOCK->nltype == NLTYPE_FIXED &&
5090                    NLBLOCK->nllen == 2 &&
5091                    c == NLBLOCK->nl[0])
5092                  {
5093                  md->hitend = TRUE;
5094                  if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
5095                  }
5096                break;
5097    
5098              case OP_ALLANY:              case OP_ALLANY:
5099              case OP_ANYBYTE:              case OP_ANYBYTE:
5100              break;              break;
# Line 4980  for (;;) Line 5103  for (;;)
5103              switch(c)              switch(c)
5104                {                {
5105                default: RRETURN(MATCH_NOMATCH);                default: RRETURN(MATCH_NOMATCH);
5106                case 0x000d:                case CHAR_CR:
5107                if (eptr < md->end_subject && *eptr == 0x0a) eptr++;                if (eptr < md->end_subject && RAWUCHAR(eptr) == CHAR_LF) eptr++;
5108                break;                break;
5109                case 0x000a:  
5110                  case CHAR_LF:
5111                break;                break;
5112    
5113                case 0x000b:                case CHAR_VT:
5114                case 0x000c:                case CHAR_FF:
5115                case 0x0085:                case CHAR_NEL:
5116    #ifndef EBCDIC
5117                case 0x2028:                case 0x2028:
5118                case 0x2029:                case 0x2029:
5119    #endif  /* Not EBCDIC */
5120                if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);                if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);
5121                break;                break;
5122                }                }
# Line 4999  for (;;) Line 5125  for (;;)
5125              case OP_NOT_HSPACE:              case OP_NOT_HSPACE:
5126              switch(c)              switch(c)
5127                {                {
5128                  HSPACE_CASES: RRETURN(MATCH_NOMATCH);
5129                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);  
5130                }                }
5131              break;              break;
5132    
5133              case OP_HSPACE:              case OP_HSPACE:
5134              switch(c)              switch(c)
5135                {                {
5136                  HSPACE_CASES: break;
5137                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;  
5138                }                }
5139              break;              break;
5140    
5141              case OP_NOT_VSPACE:              case OP_NOT_VSPACE:
5142              switch(c)              switch(c)
5143                {                {
5144                  VSPACE_CASES: RRETURN(MATCH_NOMATCH);
5145                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);  
5146                }                }
5147              break;              break;
5148    
5149              case OP_VSPACE:              case OP_VSPACE:
5150              switch(c)              switch(c)
5151                {                {
5152                  VSPACE_CASES: break;
5153                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;  
5154                }                }
5155              break;              break;
5156    
# Line 5134  for (;;) Line 5208  for (;;)
5208            c = *eptr++;            c = *eptr++;
5209            switch(ctype)            switch(ctype)
5210              {              {
5211              case OP_ANY:     /* This is the non-NL case */              case OP_ANY:               /* This is the non-NL case */
5212                if (md->partial != 0 &&    /* Take care with CRLF partial */
5213                    eptr >= md->end_subject &&
5214                    NLBLOCK->nltype == NLTYPE_FIXED &&
5215                    NLBLOCK->nllen == 2 &&
5216                    c == NLBLOCK->nl[0])
5217                  {
5218                  md->hitend = TRUE;
5219                  if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
5220                  }
5221                break;
5222    
5223              case OP_ALLANY:              case OP_ALLANY:
5224              case OP_ANYBYTE:              case OP_ANYBYTE:
5225              break;              break;
# Line 5143  for (;;) Line 5228  for (;;)
5228              switch(c)              switch(c)
5229                {                {
5230                default: RRETURN(MATCH_NOMATCH);                default: RRETURN(MATCH_NOMATCH);
5231                case 0x000d:                case CHAR_CR:
5232                if (eptr < md->end_subject && *eptr == 0x0a) eptr++;                if (eptr < md->end_subject && *eptr == CHAR_LF) eptr++;
5233                break;                break;
5234    
5235                case 0x000a:                case CHAR_LF:
5236                break;                break;
5237    
5238                case 0x000b:                case CHAR_VT:
5239                case 0x000c:                case CHAR_FF:
5240                case 0x0085:                case CHAR_NEL:
5241  #ifdef COMPILE_PCRE16  #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
5242                case 0x2028:                case 0x2028:
5243                case 0x2029:                case 0x2029:
5244  #endif  #endif
# Line 5166  for (;;) Line 5251  for (;;)
5251              switch(c)              switch(c)
5252                {                {
5253                default: break;                default: break;
5254                case 0x09:      /* HT */                HSPACE_BYTE_CASES:
5255                case 0x20:      /* SPACE */  #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
5256                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 */  
5257  #endif  #endif
5258                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
5259                }                }
# Line 5195  for (;;) Line 5263  for (;;)
5263              switch(c)              switch(c)
5264                {                {
5265                default: RRETURN(MATCH_NOMATCH);                default: RRETURN(MATCH_NOMATCH);
5266                case 0x09:      /* HT */                HSPACE_BYTE_CASES:
5267                case 0x20:      /* SPACE */  #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
5268                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 */  
5269  #endif  #endif
5270                break;                break;
5271                }                }
# Line 5224  for (;;) Line 5275  for (;;)
5275              switch(c)              switch(c)
5276                {                {
5277                default: break;                default: break;
5278                case 0x0a:      /* LF */                VSPACE_BYTE_CASES:
5279                case 0x0b:      /* VT */  #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
5280                case 0x0c:      /* FF */                VSPACE_MULTIBYTE_CASES:
               case 0x0d:      /* CR */  
               case 0x85:      /* NEL */  
 #ifdef COMPILE_PCRE16  
               case 0x2028:    /* LINE SEPARATOR */  
               case 0x2029:    /* PARAGRAPH SEPARATOR */  
5281  #endif  #endif
5282                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
5283                }                }
# Line 5241  for (;;) Line 5287  for (;;)
5287              switch(c)              switch(c)
5288                {                {
5289                default: RRETURN(MATCH_NOMATCH);                default: RRETURN(MATCH_NOMATCH);
5290                case 0x0a:      /* LF */                VSPACE_BYTE_CASES:
5291                case 0x0b:      /* VT */  #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
5292                case 0x0c:      /* FF */                VSPACE_MULTIBYTE_CASES:
               case 0x0d:      /* CR */  
               case 0x85:      /* NEL */  
 #ifdef COMPILE_PCRE16  
               case 0x2028:    /* LINE SEPARATOR */  
               case 0x2029:    /* PARAGRAPH SEPARATOR */  
5293  #endif  #endif
5294                break;                break;
5295                }                }
# Line 5397  for (;;) Line 5438  for (;;)
5438              }              }
5439            break;            break;
5440    
5441              /* Perl space used to exclude VT, but from Perl 5.18 it is included,
5442              which means that Perl space and POSIX space are now identical. PCRE
5443              was changed at release 8.34. */
5444    
5445            case PT_SPACE:    /* Perl space */            case PT_SPACE:    /* Perl space */
5446              case PT_PXSPACE:  /* POSIX space */
5447            for (i = min; i < max; i++)            for (i = min; i < max; i++)
5448              {              {
5449              int len = 1;              int len = 1;
# Line 5408  for (;;) Line 5454  for (;;)
5454                }                }
5455              GETCHARLENTEST(c, eptr, len);              GETCHARLENTEST(c, eptr, len);
5456              if ((UCD_CATEGORY(c) == ucp_Z || c == CHAR_HT || c == CHAR_NL ||              if ((UCD_CATEGORY(c) == ucp_Z || c == CHAR_HT || c == CHAR_NL ||
5457                   c == CHAR_FF || c == CHAR_CR)                   c == CHAR_VT || c == CHAR_FF || c == CHAR_CR)
5458                   == prop_fail_result)                   == prop_fail_result)
5459                break;                break;
5460              eptr+= len;              eptr+= len;
5461              }              }
5462            break;            break;
5463    
5464            case PT_PXSPACE:  /* POSIX space */            case PT_WORD:
5465            for (i = min; i < max; i++)            for (i = min; i < max; i++)
5466              {              {
5467                int category;
5468              int len = 1;              int len = 1;
5469              if (eptr >= md->end_subject)              if (eptr >= md->end_subject)
5470                {                {
# Line 5425  for (;;) Line 5472  for (;;)
5472                break;                break;
5473                }                }
5474              GETCHARLENTEST(c, eptr, len);              GETCHARLENTEST(c, eptr, len);
5475              if ((UCD_CATEGORY(c) == ucp_Z || c == CHAR_HT || c == CHAR_NL ||              category = UCD_CATEGORY(c);
5476                   c == CHAR_VT || c == CHAR_FF || c == CHAR_CR)              if ((category == ucp_L || category == ucp_N ||
5477                   == prop_fail_result)                   c == CHAR_UNDERSCORE) == prop_fail_result)
5478                break;                break;
5479              eptr+= len;              eptr+= len;
5480              }              }
5481            break;            break;
5482    
5483            case PT_WORD:            case PT_CLIST:
5484            for (i = min; i < max; i++)            for (i = min; i < max; i++)
5485              {              {
5486              int category;              const pcre_uint32 *cp;
5487              int len = 1;              int len = 1;
5488              if (eptr >= md->end_subject)              if (eptr >= md->end_subject)
5489                {                {
# Line 5444  for (;;) Line 5491  for (;;)
5491                break;                break;
5492                }                }
5493              GETCHARLENTEST(c, eptr, len);              GETCHARLENTEST(c, eptr, len);
5494              category = UCD_CATEGORY(c);              cp = PRIV(ucd_caseless_sets) + prop_value;
5495              if ((category == ucp_L || category == ucp_N ||              for (;;)
5496                   c == CHAR_UNDERSCORE) == prop_fail_result)                {
5497                  if (c < *cp)
5498                    { if (prop_fail_result) break; else goto GOT_MAX; }
5499                  if (c == *cp++)
5500                    { if (prop_fail_result) goto GOT_MAX; else break; }
5501                  }
5502                eptr += len;
5503                }
5504              GOT_MAX:
5505              break;
5506    
5507              case PT_UCNC:
5508              for (i = min; i < max; i++)
5509                {
5510                int len = 1;
5511                if (eptr >= md->end_subject)
5512                  {
5513                  SCHECK_PARTIAL();
5514                break;                break;
5515              eptr+= len;                }
5516                GETCHARLENTEST(c, eptr, len);
5517                if ((c == CHAR_DOLLAR_SIGN || c == CHAR_COMMERCIAL_AT ||
5518                     c == CHAR_GRAVE_ACCENT || (c >= 0xa0 && c <= 0xd7ff) ||
5519                     c >= 0xe000) == prop_fail_result)
5520                  break;
5521                eptr += len;
5522              }              }
5523            break;            break;
5524    
# Line 5458  for (;;) Line 5528  for (;;)
5528    
5529          /* eptr is now past the end of the maximum run */          /* eptr is now past the end of the maximum run */
5530    
5531          if (possessive) continue;          if (possessive) continue;    /* No backtracking */
5532          for(;;)          for(;;)
5533            {            {
5534              if (eptr == pp) goto TAIL_RECURSE;
5535            RMATCH(eptr, ecode, offset_top, md, eptrb, RM44);            RMATCH(eptr, ecode, offset_top, md, eptrb, RM44);
5536            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
5537            if (eptr-- == pp) break;        /* Stop if tried at original pos */            eptr--;
5538            if (utf) BACKCHAR(eptr);            if (utf) BACKCHAR(eptr);
5539            }            }
5540          }          }
5541    
5542        /* Match extended Unicode sequences. We will get here only if the        /* Match extended Unicode grapheme clusters. We will get here only if the
5543        support is in the binary; otherwise a compile-time error occurs. */        support is in the binary; otherwise a compile-time error occurs. */
5544    
5545        else if (ctype == OP_EXTUNI)        else if (ctype == OP_EXTUNI)
5546          {          {
5547          for (i = min; i < max; i++)          for (i = min; i < max; i++)
5548            {            {
           int len = 1;  
5549            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
5550              {              {
5551              SCHECK_PARTIAL();              SCHECK_PARTIAL();
5552              break;              break;
5553              }              }
5554            if (!utf) c = *eptr; else { GETCHARLEN(c, eptr, len); }            else
           if (UCD_CATEGORY(c) == ucp_M) break;  
           eptr += len;  
           while (eptr < md->end_subject)  
5555              {              {
5556              len = 1;              int lgb, rgb;
5557              if (!utf) c = *eptr; else { GETCHARLEN(c, eptr, len); }              GETCHARINCTEST(c, eptr);
5558              if (UCD_CATEGORY(c) != ucp_M) break;              lgb = UCD_GRAPHBREAK(c);
5559              eptr += len;              while (eptr < md->end_subject)
5560                  {
5561                  int len = 1;
5562                  if (!utf) c = *eptr; else { GETCHARLEN(c, eptr, len); }
5563                  rgb = UCD_GRAPHBREAK(c);
5564                  if ((PRIV(ucp_gbtable)[lgb] & (1 << rgb)) == 0) break;
5565                  lgb = rgb;
5566                  eptr += len;
5567                  }
5568              }              }
5569              CHECK_PARTIAL();
5570            }            }
5571    
5572          /* eptr is now past the end of the maximum run */          /* eptr is now past the end of the maximum run */
5573    
5574          if (possessive) continue;          if (possessive) continue;    /* No backtracking */
5575    
5576          for(;;)          for(;;)
5577            {            {
5578              int lgb, rgb;
5579              PCRE_PUCHAR fptr;
5580    
5581              if (eptr == pp) goto TAIL_RECURSE;   /* At start of char run */
5582            RMATCH(eptr, ecode, offset_top, md, eptrb, RM45);            RMATCH(eptr, ecode, offset_top, md, eptrb, RM45);
5583            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
5584            if (eptr-- == pp) break;        /* Stop if tried at original pos */  
5585            for (;;)                        /* Move back over one extended */            /* Backtracking over an extended grapheme cluster involves inspecting
5586              the previous two characters (if present) to see if a break is
5587              permitted between them. */
5588    
5589              eptr--;
5590              if (!utf) c = *eptr; else
5591              {              {
5592              if (!utf) c = *eptr; else              BACKCHAR(eptr);
5593                {              GETCHAR(c, eptr);
5594                BACKCHAR(eptr);              }
5595                GETCHAR(c, eptr);            rgb = UCD_GRAPHBREAK(c);
5596                }  
5597              if (UCD_CATEGORY(c) != ucp_M) break;            for (;;)
5598              eptr--;              {
5599                if (eptr == pp) goto TAIL_RECURSE;   /* At start of char run */
5600                fptr = eptr - 1;
5601                if (!utf) c = *fptr; else
5602                  {
5603                  BACKCHAR(fptr);
5604                  GETCHAR(c, fptr);
5605                  }
5606                lgb = UCD_GRAPHBREAK(c);
5607                if ((PRIV(ucp_gbtable)[lgb] & (1 << rgb)) == 0) break;
5608                eptr = fptr;
5609                rgb = lgb;
5610              }              }
5611            }            }
5612          }          }
# Line 5534  for (;;) Line 5630  for (;;)
5630                  break;                  break;
5631                  }                  }
5632                if (IS_NEWLINE(eptr)) break;                if (IS_NEWLINE(eptr)) break;
5633                  if (md->partial != 0 &&    /* Take care with CRLF partial */
5634                      eptr + 1 >= md->end_subject &&
5635                      NLBLOCK->nltype == NLTYPE_FIXED &&
5636                      NLBLOCK->nllen == 2 &&
5637                      RAWUCHAR(eptr) == NLBLOCK->nl[0])
5638                    {
5639                    md->hitend = TRUE;
5640                    if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
5641                    }
5642                eptr++;                eptr++;
5643                ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);                ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);
5644                }                }
# Line 5551  for (;;) Line 5656  for (;;)
5656                  break;                  break;
5657                  }                  }
5658                if (IS_NEWLINE(eptr)) break;                if (IS_NEWLINE(eptr)) break;
5659                  if (md->partial != 0 &&    /* Take care with CRLF partial */
5660                      eptr + 1 >= md->end_subject &&
5661                      NLBLOCK->nltype == NLTYPE_FIXED &&
5662                      NLBLOCK->nllen == 2 &&
5663                      RAWUCHAR(eptr) == NLBLOCK->nl[0])
5664                    {
5665                    md->hitend = TRUE;
5666                    if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
5667                    }
5668                eptr++;                eptr++;
5669                ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);                ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);
5670                }                }
# Line 5600  for (;;) Line 5714  for (;;)
5714                break;                break;
5715                }                }
5716              GETCHARLEN(c, eptr, len);              GETCHARLEN(c, eptr, len);
5717              if (c == 0x000d)              if (c == CHAR_CR)
5718                {                {
5719                if (++eptr >= md->end_subject) break;                if (++eptr >= md->end_subject) break;
5720                if (*eptr == 0x000a) eptr++;                if (RAWUCHAR(eptr) == CHAR_LF) eptr++;
5721                }                }
5722              else              else
5723                {                {
5724                if (c != 0x000a &&                if (c != CHAR_LF &&
5725                    (md->bsr_anycrlf ||                    (md->bsr_anycrlf ||
5726                     (c != 0x000b && c != 0x000c &&                     (c != CHAR_VT && c != CHAR_FF && c != CHAR_NEL
5727                      c != 0x0085 && c != 0x2028 && c != 0x2029)))  #ifndef EBCDIC
5728                        && c != 0x2028 && c != 0x2029
5729    #endif  /* Not EBCDIC */
5730                        )))
5731                  break;                  break;
5732                eptr += len;                eptr += len;
5733                }                }
# Line 5631  for (;;) Line 5748  for (;;)
5748              GETCHARLEN(c, eptr, len);              GETCHARLEN(c, eptr, len);
5749              switch(c)              switch(c)
5750                {                {
5751                  HSPACE_CASES: gotspace = TRUE; break;
5752                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;  
5753                }                }
5754              if (gotspace == (ctype == OP_NOT_HSPACE)) break;              if (gotspace == (ctype == OP_NOT_HSPACE)) break;
5755              eptr += len;              eptr += len;
# Line 5673  for (;;) Line 5770  for (;;)
5770              GETCHARLEN(c, eptr, len);              GETCHARLEN(c, eptr, len);
5771              switch(c)              switch(c)
5772                {                {
5773                  VSPACE_CASES: gotspace = TRUE; break;
5774                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;  
5775                }                }
5776              if (gotspace == (ctype == OP_NOT_VSPACE)) break;              if (gotspace == (ctype == OP_NOT_VSPACE)) break;
5777              eptr += len;              eptr += len;
# Line 5783  for (;;) Line 5872  for (;;)
5872            RRETURN(PCRE_ERROR_INTERNAL);            RRETURN(PCRE_ERROR_INTERNAL);
5873            }            }
5874    
5875          /* 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;  
5876          for(;;)          for(;;)
5877            {            {
5878              if (eptr == pp) goto TAIL_RECURSE;
5879            RMATCH(eptr, ecode, offset_top, md, eptrb, RM46);            RMATCH(eptr, ecode, offset_top, md, eptrb, RM46);
5880            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
5881            if (eptr-- == pp) break;        /* Stop if tried at original pos */            eptr--;
5882            BACKCHAR(eptr);            BACKCHAR(eptr);
5883            if (ctype == OP_ANYNL && eptr > pp  && *eptr == '\n' &&            if (ctype == OP_ANYNL && eptr > pp  && RAWUCHAR(eptr) == CHAR_NL &&
5884                eptr[-1] == '\r') eptr--;                RAWUCHAR(eptr - 1) == CHAR_CR) eptr--;
5885            }            }
5886          }          }
5887        else        else
# Line 5815  for (;;) Line 5899  for (;;)
5899                break;                break;
5900                }                }
5901              if (IS_NEWLINE(eptr)) break;              if (IS_NEWLINE(eptr)) break;
5902                if (md->partial != 0 &&    /* Take care with CRLF partial */
5903                    eptr + 1 >= md->end_subject &&
5904                    NLBLOCK->nltype == NLTYPE_FIXED &&
5905                    NLBLOCK->nllen == 2 &&
5906                    *eptr == NLBLOCK->nl[0])
5907                  {
5908                  md->hitend = TRUE;
5909                  if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
5910                  }
5911              eptr++;              eptr++;
5912              }              }
5913            break;            break;
# Line 5839  for (;;) Line 5932  for (;;)
5932                break;                break;
5933                }                }
5934              c = *eptr;              c = *eptr;
5935              if (c == 0x000d)              if (c == CHAR_CR)
5936                {                {
5937                if (++eptr >= md->end_subject) break;                if (++eptr >= md->end_subject) break;
5938                if (*eptr == 0x000a) eptr++;                if (*eptr == CHAR_LF) eptr++;
5939                }                }
5940              else              else
5941                {                {
5942                if (c != 0x000a && (md->bsr_anycrlf ||                if (c != CHAR_LF && (md->bsr_anycrlf ||
5943                  (c != 0x000b && c != 0x000c && c != 0x0085                   (c != CHAR_VT && c != CHAR_FF && c != CHAR_NEL
5944  #ifdef COMPILE_PCRE16  #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
5945                  && c != 0x2028 && c != 0x2029                   && c != 0x2028 && c != 0x2029
5946  #endif  #endif
5947                  ))) break;                   ))) break;
5948                eptr++;                eptr++;
5949                }                }
5950              }              }
# Line 5865  for (;;) Line 5958  for (;;)
5958                SCHECK_PARTIAL();                SCHECK_PARTIAL();
5959                break;                break;
5960                }                }
5961              c = *eptr;              switch(*eptr)
5962              if (c == 0x09 || c == 0x20 || c == 0xa0                {
5963  #ifdef COMPILE_PCRE16                default: eptr++; break;
5964                || c == 0x1680 || c == 0x180e || (c >= 0x2000 && c <= 0x200A)                HSPACE_BYTE_CASES:
5965                || c == 0x202f || c == 0x205f || c == 0x3000  #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
5966                  HSPACE_MULTIBYTE_CASES:
5967  #endif  #endif
5968                ) break;                goto ENDLOOP00;
5969              eptr++;                }
5970              }              }
5971              ENDLOOP00:
5972            break;            break;
5973    
5974            case OP_HSPACE:            case OP_HSPACE:
# Line 5884  for (;;) Line 5979  for (;;)
5979                SCHECK_PARTIAL();                SCHECK_PARTIAL();
5980                break;                break;
5981                }                }
5982              c = *eptr;              switch(*eptr)
5983              if (c != 0x09 && c != 0x20 && c != 0xa0                {
5984  #ifdef COMPILE_PCRE16                default: goto ENDLOOP01;
5985                && c != 0x1680 && c != 0x180e && (c < 0x2000 || c > 0x200A)                HSPACE_BYTE_CASES:
5986                && c != 0x202f && c != 0x205f && c != 0x3000  #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
5987                  HSPACE_MULTIBYTE_CASES:
5988  #endif  #endif
5989                ) break;                eptr++; break;
5990              eptr++;                }
5991              }              }
5992              ENDLOOP01:
5993            break;            break;
5994    
5995            case OP_NOT_VSPACE:            case OP_NOT_VSPACE:
# Line 5903  for (;;) Line 6000  for (;;)
6000                SCHECK_PARTIAL();                SCHECK_PARTIAL();
6001                break;                break;
6002                }                }
6003              c = *eptr;              switch(*eptr)
6004              if (c == 0x0a || c == 0x0b || c == 0x0c || c == 0x0d || c == 0x85                {
6005  #ifdef COMPILE_PCRE16                default: eptr++; break;
6006                || c == 0x2028 || c == 0x2029                VSPACE_BYTE_CASES:
6007    #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
6008                  VSPACE_MULTIBYTE_CASES:
6009  #endif  #endif
6010                ) break;                goto ENDLOOP02;
6011              eptr++;                }
6012              }              }
6013              ENDLOOP02:
6014            break;            break;
6015    
6016            case OP_VSPACE:            case OP_VSPACE:
# Line 5921  for (;;) Line 6021  for (;;)
6021                SCHECK_PARTIAL();                SCHECK_PARTIAL();
6022                break;                break;
6023                }                }
6024              c = *eptr;              switch(*eptr)
6025              if (c != 0x0a && c != 0x0b && c != 0x0c && c != 0x0d && c != 0x85                {
6026  #ifdef COMPILE_PCRE16                default: goto ENDLOOP03;
6027                && c != 0x2028 && c != 0x2029                VSPACE_BYTE_CASES:
6028    #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
6029                  VSPACE_MULTIBYTE_CASES:
6030  #endif  #endif
6031                ) break;                eptr++; break;
6032              eptr++;                }
6033              }              }
6034              ENDLOOP03:
6035            break;            break;
6036    
6037            case OP_NOT_DIGIT:            case OP_NOT_DIGIT:
# Line 6013  for (;;) Line 6116  for (;;)
6116            RRETURN(PCRE_ERROR_INTERNAL);            RRETURN(PCRE_ERROR_INTERNAL);
6117            }            }
6118    
6119          /* eptr is now past the end of the maximum run. If possessive, we are          if (possessive) continue;    /* No backtracking */
6120          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)  
6121            {            {
6122              if (eptr == pp) goto TAIL_RECURSE;
6123            RMATCH(eptr, ecode, offset_top, md, eptrb, RM47);            RMATCH(eptr, ecode, offset_top, md, eptrb, RM47);
6124            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
6125            eptr--;            eptr--;
6126            if (ctype == OP_ANYNL && eptr > pp  && *eptr == '\n' &&            if (ctype == OP_ANYNL && eptr > pp  && *eptr == CHAR_LF &&
6127                eptr[-1] == '\r') eptr--;                eptr[-1] == CHAR_CR) eptr--;
6128            }            }
6129          }          }
6130    
6131        /* Get here if we can't make it match with any permitted repetitions */        /* Control never gets here */
   
       RRETURN(MATCH_NOMATCH);  
6132        }        }
     /* Control never gets here */  
6133    
6134      /* 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
6135      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 6171  switch (frame->Xwhere)
6171    LBL(32) LBL(34) LBL(42) LBL(46)    LBL(32) LBL(34) LBL(42) LBL(46)
6172  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
6173    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)
6174    LBL(59) LBL(60) LBL(61) LBL(62)    LBL(59) LBL(60) LBL(61) LBL(62) LBL(67) LBL(68)
6175  #endif  /* SUPPORT_UCP */  #endif  /* SUPPORT_UCP */
6176  #endif  /* SUPPORT_UTF */  #endif  /* SUPPORT_UTF */
6177    default:    default:
6178    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);  
   
6179    return PCRE_ERROR_INTERNAL;    return PCRE_ERROR_INTERNAL;
6180    }    }
6181  #undef LBL  #undef LBL
# Line 6145  Undefine all the macros that were define Line 6237  Undefine all the macros that were define
6237  ***************************************************************************/  ***************************************************************************/
6238    
6239    
6240    #ifdef NO_RECURSE
6241    /*************************************************
6242    *          Release allocated heap frames         *
6243    *************************************************/
6244    
6245    /* This function releases all the allocated frames. The base frame is on the
6246    machine stack, and so must not be freed.
6247    
6248    Argument: the address of the base frame
6249    Returns:  nothing
6250    */
6251    
6252    static void
6253    release_match_heapframes (heapframe *frame_base)
6254    {
6255    heapframe *nextframe = frame_base->Xnextframe;
6256    while (nextframe != NULL)
6257      {
6258      heapframe *oldframe = nextframe;
6259      nextframe = nextframe->Xnextframe;
6260      (PUBL(stack_free))(oldframe);
6261      }
6262    }
6263    #endif
6264    
6265    
6266  /*************************************************  /*************************************************
6267  *         Execute a Regular Expression           *  *         Execute a Regular Expression           *
# Line 6170  Returns:          > 0 => success; value Line 6287  Returns:          > 0 => success; value
6287                   < -1 => some kind of unexpected problem                   < -1 => some kind of unexpected problem
6288  */  */
6289    
6290  #ifdef COMPILE_PCRE8  #if defined COMPILE_PCRE8
6291  PCRE_EXP_DEFN int PCRE_CALL_CONVENTION  PCRE_EXP_DEFN int PCRE_CALL_CONVENTION
6292  pcre_exec(const pcre *argument_re, const pcre_extra *extra_data,  pcre_exec(const pcre *argument_re, const pcre_extra *extra_data,
6293    PCRE_SPTR subject, int length, int start_offset, int options, int *offsets,    PCRE_SPTR subject, int length, int start_offset, int options, int *offsets,
6294    int offsetcount)    int offsetcount)
6295  #else  #elif defined COMPILE_PCRE16
6296  PCRE_EXP_DEFN int PCRE_CALL_CONVENTION  PCRE_EXP_DEFN int PCRE_CALL_CONVENTION
6297  pcre16_exec(const pcre16 *argument_re, const pcre16_extra *extra_data,  pcre16_exec(const pcre16 *argument_re, const pcre16_extra *extra_data,
6298    PCRE_SPTR16 subject, int length, int start_offset, int options, int *offsets,    PCRE_SPTR16 subject, int length, int start_offset, int options, int *offsets,
6299    int offsetcount)    int offsetcount)
6300    #elif defined COMPILE_PCRE32
6301    PCRE_EXP_DEFN int PCRE_CALL_CONVENTION
6302    pcre32_exec(const pcre32 *argument_re, const pcre32_extra *extra_data,
6303      PCRE_SPTR32 subject, int length, int start_offset, int options, int *offsets,
6304      int offsetcount)
6305  #endif  #endif
6306  {  {
6307  int rc, ocount, arg_offset_max;  int rc, ocount, arg_offset_max;
# Line 6202  const pcre_uint8 *start_bits = NULL; Line 6324  const pcre_uint8 *start_bits = NULL;
6324  PCRE_PUCHAR start_match = (PCRE_PUCHAR)subject + start_offset;  PCRE_PUCHAR start_match = (PCRE_PUCHAR)subject + start_offset;
6325  PCRE_PUCHAR end_subject;  PCRE_PUCHAR end_subject;
6326  PCRE_PUCHAR start_partial = NULL;  PCRE_PUCHAR start_partial = NULL;
6327    PCRE_PUCHAR match_partial = NULL;
6328  PCRE_PUCHAR req_char_ptr = start_match - 1;  PCRE_PUCHAR req_char_ptr = start_match - 1;
6329    
6330  const pcre_study_data *study;  const pcre_study_data *study;
6331  const REAL_PCRE *re = (const REAL_PCRE *)argument_re;  const REAL_PCRE *re = (const REAL_PCRE *)argument_re;
6332    
6333  /* Check for the special magic call that measures the size of the stack used  #ifdef NO_RECURSE
6334  per recursive call of match(). */  heapframe frame_zero;
6335    frame_zero.Xprevframe = NULL;            /* Marks the top level */
6336    frame_zero.Xnextframe = NULL;            /* None are allocated yet */
6337    md->match_frames_base = &frame_zero;
6338    #endif
6339    
6340    /* Check for the special magic call that measures the size of the stack used
6341    per recursive call of match(). Without the funny casting for sizeof, a Windows
6342    compiler gave this error: "unary minus operator applied to unsigned type,
6343    result still unsigned". Hopefully the cast fixes that. */
6344    
6345  if (re == NULL && extra_data == NULL && subject == NULL && length == -1)  if (re == NULL && extra_data == NULL && subject == NULL && length == -999 &&
6346        start_offset == -999)
6347  #ifdef NO_RECURSE  #ifdef NO_RECURSE
6348    return -sizeof(heapframe);    return -((int)sizeof(heapframe));
6349  #else  #else
6350    return match(NULL, NULL, NULL, 0, NULL, NULL, 0);    return match(NULL, NULL, NULL, 0, NULL, NULL, 0);
6351  #endif  #endif
6352    
6353  /* Plausibility checks */  /* Plausibility checks */
6354    
6355  if ((options & ~PUBLIC_EXEC_OPTIONS) != 0) return PCRE_ERROR_BADOPTION;  if ((options & ~PUBLIC_EXEC_OPTIONS) != 0) return PCRE_ERROR_BADOPTION;
6356  if (re == NULL || subject == NULL || (offsets == NULL && offsetcount > 0))  if (re == NULL || subject == NULL || (offsets == NULL && offsetcount > 0))
6357    return PCRE_ERROR_NULL;    return PCRE_ERROR_NULL;
6358  if (offsetcount < 0) return PCRE_ERROR_BADCOUNT;  if (offsetcount < 0) return PCRE_ERROR_BADCOUNT;
6359    if (length < 0) return PCRE_ERROR_BADLENGTH;
6360  if (start_offset < 0 || start_offset > length) return PCRE_ERROR_BADOFFSET;  if (start_offset < 0 || start_offset > length) return PCRE_ERROR_BADOFFSET;
6361    
6362  /* 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 6394  if (utf && (options & PCRE_NO_UTF8_CHECK
6394        offsets[0] = erroroffset;        offsets[0] = erroroffset;
6395        offsets[1] = errorcode;        offsets[1] = errorcode;
6396        }        }
6397  #ifdef COMPILE_PCRE16  #if defined COMPILE_PCRE8
     return (errorcode <= PCRE_UTF16_ERR1 && md->partial > 1)?  
       PCRE_ERROR_SHORTUTF16 : PCRE_ERROR_BADUTF16;  
 #else  
6398      return (errorcode <= PCRE_UTF8_ERR5 && md->partial > 1)?      return (errorcode <= PCRE_UTF8_ERR5 && md->partial > 1)?
6399        PCRE_ERROR_SHORTUTF8 : PCRE_ERROR_BADUTF8;        PCRE_ERROR_SHORTUTF8 : PCRE_ERROR_BADUTF8;
6400    #elif defined COMPILE_PCRE16
6401        return (errorcode <= PCRE_UTF16_ERR1 && md->partial > 1)?
6402          PCRE_ERROR_SHORTUTF16 : PCRE_ERROR_BADUTF16;
6403    #elif defined COMPILE_PCRE32
6404        return PCRE_ERROR_BADUTF32;
6405  #endif  #endif
6406      }      }
6407    #if defined COMPILE_PCRE8 || defined COMPILE_PCRE16
6408    /* 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. */
6409    if (start_offset > 0 && start_offset < length &&    if (start_offset > 0 && start_offset < length &&
6410        NOT_FIRSTCHAR(((PCRE_PUCHAR)subject)[start_offset]))        NOT_FIRSTCHAR(((PCRE_PUCHAR)subject)[start_offset]))
6411      return PCRE_ERROR_BADUTF8_OFFSET;      return PCRE_ERROR_BADUTF8_OFFSET;
6412    #endif
6413    }    }
6414  #endif  #endif
6415    
6416  /* If the pattern was successfully studied with JIT support, run the JIT  /* If the pattern was successfully studied with JIT support, run the JIT
6417  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
6418  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
6419  an unsupported flag is set. In particular, JIT does not support partial  an unsupported flag is set. */
 matching. */  
6420    
6421  #ifdef SUPPORT_JIT  #ifdef SUPPORT_JIT
6422  if (extra_data != NULL  if (extra_data != NULL
6423      && (extra_data->flags & PCRE_EXTRA_EXECUTABLE_JIT) != 0      && (extra_data->flags & (PCRE_EXTRA_EXECUTABLE_JIT |
6424                                 PCRE_EXTRA_TABLES)) == PCRE_EXTRA_EXECUTABLE_JIT
6425      && extra_data->executable_jit != NULL      && extra_data->executable_jit != NULL
6426      && (extra_data->flags & PCRE_EXTRA_TABLES) == 0      && (options & ~PUBLIC_JIT_EXEC_OPTIONS) == 0)
6427      && (options & ~(PCRE_NO_UTF8_CHECK | PCRE_NOTBOL | PCRE_NOTEOL |    {
6428                      PCRE_NOTEMPTY | PCRE_NOTEMPTY_ATSTART)) == 0)    rc = PRIV(jit_exec)(extra_data, (const pcre_uchar *)subject, length,
6429    return PRIV(jit_exec)(re, extra_data->executable_jit,         start_offset, options, offsets, offsetcount);
6430      (const pcre_uchar *)subject, length, start_offset, options,  
6431      ((extra_data->flags & PCRE_EXTRA_MATCH_LIMIT) == 0)    /* PCRE_ERROR_NULL means that the selected normal or partial matching
6432      ? MATCH_LIMIT : extra_data->match_limit, offsets, offsetcount);    mode is not compiled. In this case we simply fallback to interpreter. */
6433    
6434      if (rc != PCRE_ERROR_JIT_BADOPTION) return rc;
6435      }
6436  #endif  #endif
6437    
6438  /* 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 6454  md->callout_data = NULL;
6454    
6455  tables = re->tables;  tables = re->tables;
6456    
6457    /* The two limit values override the defaults, whatever their value. */
6458    
6459  if (extra_data != NULL)  if (extra_data != NULL)
6460    {    {
6461    register unsigned int flags = extra_data->flags;    register unsigned int flags = extra_data->flags;
# Line 6328  if (extra_data != NULL) Line 6470  if (extra_data != NULL)
6470    if ((flags & PCRE_EXTRA_TABLES) != 0) tables = extra_data->tables;    if ((flags & PCRE_EXTRA_TABLES) != 0) tables = extra_data->tables;
6471    }    }
6472    
6473    /* Limits in the regex override only if they are smaller. */
6474    
6475    if ((re->flags & PCRE_MLSET) != 0 && re->limit_match < md->match_limit)
6476      md->match_limit = re->limit_match;
6477    
6478    if ((re->flags & PCRE_RLSET) != 0 &&
6479        re->limit_recursion < md->match_limit_recursion)
6480      md->match_limit_recursion = re->limit_recursion;
6481    
6482  /* 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
6483  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
6484  in other programs later. */  in other programs later. */
# Line 6353  end_subject = md->end_subject; Line 6504  end_subject = md->end_subject;
6504  md->endonly = (re->options & PCRE_DOLLAR_ENDONLY) != 0;  md->endonly = (re->options & PCRE_DOLLAR_ENDONLY) != 0;
6505  md->use_ucp = (re->options & PCRE_UCP) != 0;  md->use_ucp = (re->options & PCRE_UCP) != 0;
6506  md->jscript_compat = (re->options & PCRE_JAVASCRIPT_COMPAT) != 0;  md->jscript_compat = (re->options & PCRE_JAVASCRIPT_COMPAT) != 0;
6507  md->ignore_skip_arg = FALSE;  md->ignore_skip_arg = 0;
6508    
6509  /* Some options are unpacked into BOOL variables in the hope that testing  /* Some options are unpacked into BOOL variables in the hope that testing
6510  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 6614  if (re->top_backref > 0 && re->top_backr
6614    DPRINTF(("Got memory to hold back references\n"));    DPRINTF(("Got memory to hold back references\n"));
6615    }    }
6616  else md->offset_vector = offsets;  else md->offset_vector = offsets;
   
6617  md->offset_end = ocount;  md->offset_end = ocount;
6618  md->offset_max = (2*ocount)/3;  md->offset_max = (2*ocount)/3;
6619  md->offset_overflow = FALSE;  md->capture_last = 0;
 md->capture_last = -1;  
6620    
6621  /* Reset the working variable associated with each extraction. These should  /* Reset the working variable associated with each extraction. These should
6622  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 6644  if (!anchored)
6644    if ((re->flags & PCRE_FIRSTSET) != 0)    if ((re->flags & PCRE_FIRSTSET) != 0)
6645      {      {
6646      has_first_char = TRUE;      has_first_char = TRUE;
6647      first_char = first_char2 = re->first_char;      first_char = first_char2 = (pcre_uchar)(re->first_char);
6648      if ((re->flags & PCRE_FCH_CASELESS) != 0)      if ((re->flags & PCRE_FCH_CASELESS) != 0)
6649        {        {
6650        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 6666  character" set. */
6666  if ((re->flags & PCRE_REQCHSET) != 0)  if ((re->flags & PCRE_REQCHSET) != 0)
6667    {    {
6668    has_req_char = TRUE;    has_req_char = TRUE;
6669    req_char = req_char2 = re->req_char;    req_char = req_char2 = (pcre_uchar)(re->req_char);
6670    if ((re->flags & PCRE_RCH_CASELESS) != 0)    if ((re->flags & PCRE_RCH_CASELESS) != 0)
6671      {      {
6672      req_char2 = TABLE_GET(req_char, md->fcc, req_char);      req_char2 = TABLE_GET(req_char, md->fcc, req_char);
# Line 6575  for(;;) Line 6724  for(;;)
6724    
6725      if (has_first_char)      if (has_first_char)
6726        {        {
6727          pcre_uchar smc;
6728    
6729        if (first_char != first_char2)        if (first_char != first_char2)
6730          while (start_match < end_subject &&          while (start_match < end_subject &&
6731              *start_match != first_char && *start_match != first_char2)            (smc = RAWUCHARTEST(start_match)) != first_char && smc != first_char2)
6732            start_match++;            start_match++;
6733        else        else
6734          while (start_match < end_subject && *start_match != first_char)          while (start_match < end_subject && RAWUCHARTEST(start_match) != first_char)
6735            start_match++;            start_match++;
6736        }        }
6737    
# Line 6612  for(;;) Line 6763  for(;;)
6763          if (start_match[-1] == CHAR_CR &&          if (start_match[-1] == CHAR_CR &&
6764               (md->nltype == NLTYPE_ANY || md->nltype == NLTYPE_ANYCRLF) &&               (md->nltype == NLTYPE_ANY || md->nltype == NLTYPE_ANYCRLF) &&
6765               start_match < end_subject &&               start_match < end_subject &&
6766               *start_match == CHAR_NL)               RAWUCHARTEST(start_match) == CHAR_NL)
6767            start_match++;            start_match++;
6768          }          }
6769        }        }
# Line 6623  for(;;) Line 6774  for(;;)
6774        {        {
6775        while (start_match < end_subject)        while (start_match < end_subject)
6776          {          {
6777          register unsigned int c = *start_match;          register pcre_uint32 c = RAWUCHARTEST(start_match);
6778  #ifndef COMPILE_PCRE8  #ifndef COMPILE_PCRE8
6779          if (c > 255) c = 255;          if (c > 255) c = 255;
6780  #endif  #endif
# Line 6691  for(;;) Line 6842  for(;;)
6842            {            {
6843            while (p < end_subject)            while (p < end_subject)
6844              {              {
6845              register int pp = *p++;              register pcre_uint32 pp = RAWUCHARINCTEST(p);
6846              if (pp == req_char || pp == req_char2) { p--; break; }              if (pp == req_char || pp == req_char2) { p--; break; }
6847              }              }
6848            }            }
# Line 6699  for(;;) Line 6850  for(;;)
6850            {            {
6851            while (p < end_subject)            while (p < end_subject)
6852              {              {
6853              if (*p++ == req_char) { p--; break; }              if (RAWUCHARINCTEST(p) == req_char) { p--; break; }
6854              }              }
6855            }            }
6856    
# Line 6735  for(;;) Line 6886  for(;;)
6886    md->match_call_count = 0;    md->match_call_count = 0;
6887    md->match_function_type = 0;    md->match_function_type = 0;
6888    md->end_offset_top = 0;    md->end_offset_top = 0;
6889      md->skip_arg_count = 0;
6890    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);
6891    if (md->hitend && start_partial == NULL) start_partial = md->start_used_ptr;    if (md->hitend && start_partial == NULL)
6892        {
6893        start_partial = md->start_used_ptr;
6894        match_partial = start_match;
6895        }
6896    
6897    switch(rc)    switch(rc)
6898      {      {
# Line 6749  for(;;) Line 6905  for(;;)
6905    
6906      case MATCH_SKIP_ARG:      case MATCH_SKIP_ARG:
6907      new_start_match = start_match;      new_start_match = start_match;
6908      md->ignore_skip_arg = TRUE;      md->ignore_skip_arg = md->skip_arg_count;
6909      break;      break;
6910    
6911      /* 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
6912      same as the match we have just done, treat it as NOMATCH. */      greater than the match we have just done, treat it as NOMATCH. */
6913