/[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 1316 by zherczeg, Sun Apr 28 08:54:42 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 preceed 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);
831        RRETURN(rrc);  
   
832      /* Pass back the current skip name by overloading md->start_match_ptr and      /* Pass back the current skip name by overloading md->start_match_ptr and
833      returning the special MATCH_SKIP_ARG return code. This will either be      returning the special MATCH_SKIP_ARG return code. This will either be
834      caught by a matching MARK, or get to the top, where it causes a rematch      caught by a matching MARK, or get to the top, where it causes a rematch
835      with the md->ignore_skip_arg flag set. */      with md->ignore_skip_arg set to the value of md->skip_arg_count. */
836    
837      md->start_match_ptr = ecode + 2;      md->start_match_ptr = ecode + 2;
838      RRETURN(MATCH_SKIP_ARG);      RRETURN(MATCH_SKIP_ARG);
# Line 838  for (;;) Line 875  for (;;)
875      case OP_ONCE_NC:      case OP_ONCE_NC:
876      prev = ecode;      prev = ecode;
877      saved_eptr = eptr;      saved_eptr = eptr;
878      save_mark = md->mark;      save_mark = md->mark;
879      do      do
880        {        {
881        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, eptrb, RM64);        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, eptrb, RM64);
# Line 857  for (;;) Line 894  for (;;)
894    
895        if (rrc != MATCH_NOMATCH) RRETURN(rrc);        if (rrc != MATCH_NOMATCH) RRETURN(rrc);
896        ecode += GET(ecode,1);        ecode += GET(ecode,1);
897        md->mark = save_mark;        md->mark = save_mark;
898        }        }
899      while (*ecode == OP_ALT);      while (*ecode == OP_ALT);
900    
# Line 897  for (;;) Line 934  for (;;)
934        }        }
935      else  /* OP_KETRMAX */      else  /* OP_KETRMAX */
936        {        {
       md->match_function_type = MATCH_CBEGROUP;  
937        RMATCH(eptr, prev, offset_top, md, eptrb, RM66);        RMATCH(eptr, prev, offset_top, md, eptrb, RM66);
938        if (rrc != MATCH_NOMATCH) RRETURN(rrc);        if (rrc != MATCH_NOMATCH) RRETURN(rrc);
939        ecode += 1 + LINK_SIZE;        ecode += 1 + LINK_SIZE;
# Line 937  for (;;) Line 973  for (;;)
973        save_offset2 = md->offset_vector[offset+1];        save_offset2 = md->offset_vector[offset+1];
974        save_offset3 = md->offset_vector[md->offset_end - number];        save_offset3 = md->offset_vector[md->offset_end - number];
975        save_capture_last = md->capture_last;        save_capture_last = md->capture_last;
976        save_mark = md->mark;        save_mark = md->mark;
977    
978        DPRINTF(("saving %d %d %d\n", save_offset1, save_offset2, save_offset3));        DPRINTF(("saving %d %d %d\n", save_offset1, save_offset2, save_offset3));
979        md->offset_vector[md->offset_end - number] =        md->offset_vector[md->offset_end - number] =
# Line 1026  for (;;) Line 1062  for (;;)
1062    
1063      for (;;)      for (;;)
1064        {        {
1065        if (op >= OP_SBRA || op == OP_ONCE) md->match_function_type = MATCH_CBEGROUP;        if (op >= OP_SBRA || op == OP_ONCE)
1066            md->match_function_type = MATCH_CBEGROUP;
1067    
1068        /* If this is not a possibly empty group, and there are no (*THEN)s in        /* If this is not a possibly empty group, and there are no (*THEN)s in
1069        the pattern, and this is the final alternative, optimize as described        the pattern, and this is the final alternative, optimize as described
# Line 1041  for (;;) Line 1078  for (;;)
1078        /* In all other cases, we have to make another call to match(). */        /* In all other cases, we have to make another call to match(). */
1079    
1080        save_mark = md->mark;        save_mark = md->mark;
1081          save_capture_last = md->capture_last;
1082        RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md, eptrb,        RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md, eptrb,
1083          RM2);          RM2);
1084    
1085        /* See comment in the code for capturing groups above about handling        /* See comment in the code for capturing groups above about handling
1086        THEN. */        THEN. */
1087    
# Line 1070  for (;;) Line 1108  for (;;)
1108          RRETURN(rrc);          RRETURN(rrc);
1109          }          }
1110        ecode += GET(ecode, 1);        ecode += GET(ecode, 1);
1111        md->mark = save_mark;        md->mark = save_mark;
1112        if (*ecode != OP_ALT) break;        if (*ecode != OP_ALT) break;
1113          md->capture_last = save_capture_last;
1114        }        }
1115    
1116      RRETURN(MATCH_NOMATCH);      RRETURN(MATCH_NOMATCH);
# Line 1193  for (;;) Line 1232  for (;;)
1232      POSSESSIVE_NON_CAPTURE:      POSSESSIVE_NON_CAPTURE:
1233      matched_once = FALSE;      matched_once = FALSE;
1234      code_offset = (int)(ecode - md->start_code);      code_offset = (int)(ecode - md->start_code);
1235        save_capture_last = md->capture_last;
1236    
1237      for (;;)      for (;;)
1238        {        {
# Line 1222  for (;;) Line 1262  for (;;)
1262        if (rrc != MATCH_NOMATCH) RRETURN(rrc);        if (rrc != MATCH_NOMATCH) RRETURN(rrc);
1263        ecode += GET(ecode, 1);        ecode += GET(ecode, 1);
1264        if (*ecode != OP_ALT) break;        if (*ecode != OP_ALT) break;
1265          md->capture_last = save_capture_last;
1266        }        }
1267    
1268      if (matched_once || allow_zero)      if (matched_once || allow_zero)
# Line 1253  for (;;) Line 1294  for (;;)
1294          cb.version          = 2;   /* Version 1 of the callout block */          cb.version          = 2;   /* Version 1 of the callout block */
1295          cb.callout_number   = ecode[LINK_SIZE+2];          cb.callout_number   = ecode[LINK_SIZE+2];
1296          cb.offset_vector    = md->offset_vector;          cb.offset_vector    = md->offset_vector;
1297  #ifdef COMPILE_PCRE8  #if defined COMPILE_PCRE8
1298          cb.subject          = (PCRE_SPTR)md->start_subject;          cb.subject          = (PCRE_SPTR)md->start_subject;
1299  #else  #elif defined COMPILE_PCRE16
1300          cb.subject          = (PCRE_SPTR16)md->start_subject;          cb.subject          = (PCRE_SPTR16)md->start_subject;
1301    #elif defined COMPILE_PCRE32
1302            cb.subject          = (PCRE_SPTR32)md->start_subject;
1303  #endif  #endif
1304          cb.subject_length   = (int)(md->end_subject - md->start_subject);          cb.subject_length   = (int)(md->end_subject - md->start_subject);
1305          cb.start_match      = (int)(mstart - md->start_subject);          cb.start_match      = (int)(mstart - md->start_subject);
# Line 1264  for (;;) Line 1307  for (;;)
1307          cb.pattern_position = GET(ecode, LINK_SIZE + 3);          cb.pattern_position = GET(ecode, LINK_SIZE + 3);
1308          cb.next_item_length = GET(ecode, 3 + 2*LINK_SIZE);          cb.next_item_length = GET(ecode, 3 + 2*LINK_SIZE);
1309          cb.capture_top      = offset_top/2;          cb.capture_top      = offset_top/2;
1310          cb.capture_last     = md->capture_last;          cb.capture_last     = md->capture_last & CAPLMASK;
1311            /* Internal change requires this for API compatibility. */
1312            if (cb.capture_last == 0) cb.capture_last = -1;
1313          cb.callout_data     = md->callout_data;          cb.callout_data     = md->callout_data;
1314          cb.mark             = md->nomatch_mark;          cb.mark             = md->nomatch_mark;
1315          if ((rrc = (*PUBL(callout))(&cb)) > 0) RRETURN(MATCH_NOMATCH);          if ((rrc = (*PUBL(callout))(&cb)) > 0) RRETURN(MATCH_NOMATCH);
1316          if (rrc < 0) RRETURN(rrc);          if (rrc < 0) RRETURN(rrc);
1317          }          }
1318        ecode += PRIV(OP_lengths)[OP_CALLOUT];        ecode += PRIV(OP_lengths)[OP_CALLOUT];
1319          codelink -= PRIV(OP_lengths)[OP_CALLOUT];
1320        }        }
1321    
1322      condcode = ecode[LINK_SIZE+1];      condcode = ecode[LINK_SIZE+1];
# Line 1286  for (;;) Line 1332  for (;;)
1332          }          }
1333        else        else
1334          {          {
1335          int recno = GET2(ecode, LINK_SIZE + 2);   /* Recursion group number*/          unsigned int recno = GET2(ecode, LINK_SIZE + 2);   /* Recursion group number*/
1336          condition = (recno == RREF_ANY || recno == md->recursive->group_num);          condition = (recno == RREF_ANY || recno == md->recursive->group_num);
1337    
1338          /* If the test is for recursion into a specific subpattern, and it is          /* If the test is for recursion into a specific subpattern, and it is
# Line 1358  for (;;) Line 1404  for (;;)
1404    
1405        if (!condition && condcode == OP_NCREF)        if (!condition && condcode == OP_NCREF)
1406          {          {
1407          int refno = offset >> 1;          unsigned int refno = offset >> 1;
1408          pcre_uchar *slotA = md->name_table;          pcre_uchar *slotA = md->name_table;
1409    
1410          for (i = 0; i < md->name_count; i++)          for (i = 0; i < md->name_count; i++)
# Line 1486  for (;;) Line 1532  for (;;)
1532      to close any currently open capturing brackets. */      to close any currently open capturing brackets. */
1533    
1534      case OP_CLOSE:      case OP_CLOSE:
1535      number = GET2(ecode, 1);      number = GET2(ecode, 1);   /* Must be less than 65536 */
1536      offset = number << 1;      offset = number << 1;
1537    
1538  #ifdef PCRE_DEBUG  #ifdef PCRE_DEBUG
# Line 1494  for (;;) Line 1540  for (;;)
1540        printf("\n");        printf("\n");
1541  #endif  #endif
1542    
1543      md->capture_last = number;      md->capture_last = (md->capture_last & OVFLMASK) | number;
1544      if (offset >= md->offset_max) md->offset_overflow = TRUE; else      if (offset >= md->offset_max) md->capture_last |= OVFLBIT; else
1545        {        {
1546        md->offset_vector[offset] =        md->offset_vector[offset] =
1547          md->offset_vector[md->offset_end - number];          md->offset_vector[md->offset_end - number];
# Line 1549  for (;;) Line 1595  for (;;)
1595    
1596      case OP_ASSERT:      case OP_ASSERT:
1597      case OP_ASSERTBACK:      case OP_ASSERTBACK:
1598      save_mark = md->mark;      save_mark = md->mark;
1599      if (md->match_function_type == MATCH_CONDASSERT)      if (md->match_function_type == MATCH_CONDASSERT)
1600        {        {
1601        condassert = TRUE;        condassert = TRUE;
# Line 1557  for (;;) Line 1603  for (;;)
1603        }        }
1604      else condassert = FALSE;      else condassert = FALSE;
1605    
1606        /* Loop for each branch */
1607    
1608      do      do
1609        {        {
1610        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, NULL, RM4);        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, NULL, RM4);
1611    
1612          /* A match means that the assertion is true; break out of the loop
1613          that matches its alternatives. */
1614    
1615        if (rrc == MATCH_MATCH || rrc == MATCH_ACCEPT)        if (rrc == MATCH_MATCH || rrc == MATCH_ACCEPT)
1616          {          {
1617          mstart = md->start_match_ptr;   /* In case \K reset it */          mstart = md->start_match_ptr;   /* In case \K reset it */
1618          break;          break;
1619          }          }
1620    
1621          /* If not matched, restore the previous mark setting. */
1622    
1623          md->mark = save_mark;
1624    
1625        /* PCRE does not allow THEN to escape beyond an assertion; it is treated        /* See comment in the code for capturing groups above about handling
1626        as NOMATCH. */        THEN. */
1627    
1628        if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN) RRETURN(rrc);        if (rrc == MATCH_THEN)
1629            {
1630            next = ecode + GET(ecode,1);
1631            if (md->start_match_ptr < next &&
1632                (*ecode == OP_ALT || *next == OP_ALT))
1633              rrc = MATCH_NOMATCH;
1634            }
1635    
1636          /* Anything other than NOMATCH causes the entire assertion to fail,
1637          passing back the return code. This includes COMMIT, SKIP, PRUNE and an
1638          uncaptured THEN, which means they take their normal effect. This
1639          consistent approach does not always have exactly the same effect as in
1640          Perl. */
1641    
1642          if (rrc != MATCH_NOMATCH) RRETURN(rrc);
1643        ecode += GET(ecode, 1);        ecode += GET(ecode, 1);
       md->mark = save_mark;  
1644        }        }
1645      while (*ecode == OP_ALT);      while (*ecode == OP_ALT);   /* Continue for next alternative */
1646    
1647        /* If we have tried all the alternative branches, the assertion has
1648        failed. If not, we broke out after a match. */
1649    
1650      if (*ecode == OP_KET) RRETURN(MATCH_NOMATCH);      if (*ecode == OP_KET) RRETURN(MATCH_NOMATCH);
1651    
# Line 1581  for (;;) Line 1653  for (;;)
1653    
1654      if (condassert) RRETURN(MATCH_MATCH);      if (condassert) RRETURN(MATCH_MATCH);
1655    
1656      /* Continue from after the assertion, updating the offsets high water      /* Continue from after a successful assertion, updating the offsets high
1657      mark, since extracts may have been taken during the assertion. */      water mark, since extracts may have been taken during the assertion. */
1658    
1659      do ecode += GET(ecode,1); while (*ecode == OP_ALT);      do ecode += GET(ecode,1); while (*ecode == OP_ALT);
1660      ecode += 1 + LINK_SIZE;      ecode += 1 + LINK_SIZE;
1661      offset_top = md->end_offset_top;      offset_top = md->end_offset_top;
1662      continue;      continue;
1663    
1664      /* Negative assertion: all branches must fail to match. Encountering SKIP,      /* Negative assertion: all branches must fail to match for the assertion to
1665      PRUNE, or COMMIT means we must assume failure without checking subsequent      succeed. */
     branches. */  
1666    
1667      case OP_ASSERT_NOT:      case OP_ASSERT_NOT:
1668      case OP_ASSERTBACK_NOT:      case OP_ASSERTBACK_NOT:
1669      save_mark = md->mark;      save_mark = md->mark;
1670      if (md->match_function_type == MATCH_CONDASSERT)      if (md->match_function_type == MATCH_CONDASSERT)
1671        {        {
1672        condassert = TRUE;        condassert = TRUE;
# Line 1603  for (;;) Line 1674  for (;;)
1674        }        }
1675      else condassert = FALSE;      else condassert = FALSE;
1676    
1677        /* Loop for each alternative branch. */
1678    
1679      do      do
1680        {        {
1681        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, NULL, RM5);        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, NULL, RM5);
1682        md->mark = save_mark;        md->mark = save_mark;   /* Always restore the mark setting */
1683        if (rrc == MATCH_MATCH || rrc == MATCH_ACCEPT) RRETURN(MATCH_NOMATCH);  
1684        if (rrc == MATCH_SKIP || rrc == MATCH_PRUNE || rrc == MATCH_COMMIT)        switch(rrc)
1685          {          {
1686          do ecode += GET(ecode,1); while (*ecode == OP_ALT);          case MATCH_MATCH:            /* A successful match means */
1687          break;          case MATCH_ACCEPT:           /* the assertion has failed. */
1688          }          RRETURN(MATCH_NOMATCH);
1689    
1690            case MATCH_NOMATCH:          /* Carry on with next branch */
1691            break;
1692    
1693        /* PCRE does not allow THEN to escape beyond an assertion; it is treated          /* See comment in the code for capturing groups above about handling
1694        as NOMATCH. */          THEN. */
1695    
1696        if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN) RRETURN(rrc);          case MATCH_THEN:
1697            next = ecode + GET(ecode,1);
1698            if (md->start_match_ptr < next &&
1699                (*ecode == OP_ALT || *next == OP_ALT))
1700              {
1701              rrc = MATCH_NOMATCH;
1702              break;
1703              }
1704            /* Otherwise fall through. */
1705    
1706            /* COMMIT, SKIP, PRUNE, and an uncaptured THEN cause the whole
1707            assertion to fail to match, without considering any more alternatives.
1708            Failing to match means the assertion is true. This is a consistent
1709            approach, but does not always have the same effect as in Perl. */
1710    
1711            case MATCH_COMMIT:
1712            case MATCH_SKIP:
1713            case MATCH_SKIP_ARG:
1714            case MATCH_PRUNE:
1715            do ecode += GET(ecode,1); while (*ecode == OP_ALT);
1716            goto NEG_ASSERT_TRUE;   /* Break out of alternation loop */
1717    
1718            /* Anything else is an error */
1719    
1720            default:
1721            RRETURN(rrc);
1722            }
1723    
1724          /* Continue with next branch */
1725    
1726        ecode += GET(ecode,1);        ecode += GET(ecode,1);
1727        }        }
1728      while (*ecode == OP_ALT);      while (*ecode == OP_ALT);
1729    
1730        /* All branches in the assertion failed to match. */
1731    
1732        NEG_ASSERT_TRUE:
1733      if (condassert) RRETURN(MATCH_MATCH);  /* Condition assertion */      if (condassert) RRETURN(MATCH_MATCH);  /* Condition assertion */
1734        ecode += 1 + LINK_SIZE;                /* Continue with current branch */
     ecode += 1 + LINK_SIZE;  
1735      continue;      continue;
1736    
1737      /* Move the subject pointer back. This occurs only at the start of      /* Move the subject pointer back. This occurs only at the start of
# Line 1671  for (;;) Line 1778  for (;;)
1778        cb.version          = 2;   /* Version 1 of the callout block */        cb.version          = 2;   /* Version 1 of the callout block */
1779        cb.callout_number   = ecode[1];        cb.callout_number   = ecode[1];
1780        cb.offset_vector    = md->offset_vector;        cb.offset_vector    = md->offset_vector;
1781  #ifdef COMPILE_PCRE8  #if defined COMPILE_PCRE8
1782        cb.subject          = (PCRE_SPTR)md->start_subject;        cb.subject          = (PCRE_SPTR)md->start_subject;
1783  #else  #elif defined COMPILE_PCRE16
1784        cb.subject          = (PCRE_SPTR16)md->start_subject;        cb.subject          = (PCRE_SPTR16)md->start_subject;
1785    #elif defined COMPILE_PCRE32
1786          cb.subject          = (PCRE_SPTR32)md->start_subject;
1787  #endif  #endif
1788        cb.subject_length   = (int)(md->end_subject - md->start_subject);        cb.subject_length   = (int)(md->end_subject - md->start_subject);
1789        cb.start_match      = (int)(mstart - md->start_subject);        cb.start_match      = (int)(mstart - md->start_subject);
# Line 1682  for (;;) Line 1791  for (;;)
1791        cb.pattern_position = GET(ecode, 2);        cb.pattern_position = GET(ecode, 2);
1792        cb.next_item_length = GET(ecode, 2 + LINK_SIZE);        cb.next_item_length = GET(ecode, 2 + LINK_SIZE);
1793        cb.capture_top      = offset_top/2;        cb.capture_top      = offset_top/2;
1794        cb.capture_last     = md->capture_last;        cb.capture_last     = md->capture_last & CAPLMASK;
1795          /* Internal change requires this for API compatibility. */
1796          if (cb.capture_last == 0) cb.capture_last = -1;
1797        cb.callout_data     = md->callout_data;        cb.callout_data     = md->callout_data;
1798        cb.mark             = md->nomatch_mark;        cb.mark             = md->nomatch_mark;
1799        if ((rrc = (*PUBL(callout))(&cb)) > 0) RRETURN(MATCH_NOMATCH);        if ((rrc = (*PUBL(callout))(&cb)) > 0) RRETURN(MATCH_NOMATCH);
# Line 1711  for (;;) Line 1822  for (;;)
1822      case OP_RECURSE:      case OP_RECURSE:
1823        {        {
1824        recursion_info *ri;        recursion_info *ri;
1825        int recno;        unsigned int recno;
1826    
1827        callpat = md->start_code + GET(ecode, 1);        callpat = md->start_code + GET(ecode, 1);
1828        recno = (callpat == md->start_code)? 0 :        recno = (callpat == md->start_code)? 0 :
# Line 1728  for (;;) Line 1839  for (;;)
1839        /* Add to "recursing stack" */        /* Add to "recursing stack" */
1840    
1841        new_recursive.group_num = recno;        new_recursive.group_num = recno;
1842          new_recursive.saved_capture_last = md->capture_last;
1843        new_recursive.subject_position = eptr;        new_recursive.subject_position = eptr;
1844        new_recursive.prevrec = md->recursive;        new_recursive.prevrec = md->recursive;
1845        md->recursive = &new_recursive;        md->recursive = &new_recursive;
# Line 1751  for (;;) Line 1863  for (;;)
1863              new_recursive.saved_max * sizeof(int));              new_recursive.saved_max * sizeof(int));
1864    
1865        /* OK, now we can do the recursion. After processing each alternative,        /* OK, now we can do the recursion. After processing each alternative,
1866        restore the offset data. If there were nested recursions, md->recursive        restore the offset data and the last captured value. If there were nested
1867        might be changed, so reset it before looping. */        recursions, md->recursive might be changed, so reset it before looping.
1868          */
1869    
1870        DPRINTF(("Recursing into group %d\n", new_recursive.group_num));        DPRINTF(("Recursing into group %d\n", new_recursive.group_num));
1871        cbegroup = (*callpat >= OP_SBRA);        cbegroup = (*callpat >= OP_SBRA);
# Line 1763  for (;;) Line 1876  for (;;)
1876            md, eptrb, RM6);            md, eptrb, RM6);
1877          memcpy(md->offset_vector, new_recursive.offset_save,          memcpy(md->offset_vector, new_recursive.offset_save,
1878              new_recursive.saved_max * sizeof(int));              new_recursive.saved_max * sizeof(int));
1879            md->capture_last = new_recursive.saved_capture_last;
1880          md->recursive = new_recursive.prevrec;          md->recursive = new_recursive.prevrec;
1881          if (rrc == MATCH_MATCH || rrc == MATCH_ACCEPT)          if (rrc == MATCH_MATCH || rrc == MATCH_ACCEPT)
1882            {            {
# Line 1779  for (;;) Line 1893  for (;;)
1893            goto RECURSION_MATCHED;        /* Exit loop; end processing */            goto RECURSION_MATCHED;        /* Exit loop; end processing */
1894            }            }
1895    
1896          /* PCRE does not allow THEN to escape beyond a recursion; it is treated          /* PCRE does not allow THEN, SKIP, PRUNE or COMMIT to escape beyond a
1897          as NOMATCH. */          recursion; they cause a NOMATCH for the entire recursion. These codes
1898            are defined in a range that can be tested for. */
1899    
1900            if (rrc >= MATCH_BACKTRACK_MIN && rrc <= MATCH_BACKTRACK_MAX)
1901              RRETURN(MATCH_NOMATCH);
1902    
1903            /* Any return code other than NOMATCH is an error. */
1904    
1905          else if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN)          if (rrc != MATCH_NOMATCH)
1906            {            {
1907            DPRINTF(("Recursion gave error %d\n", rrc));            DPRINTF(("Recursion gave error %d\n", rrc));
1908            if (new_recursive.offset_save != stacksave)            if (new_recursive.offset_save != stacksave)
# Line 1912  for (;;) Line 2032  for (;;)
2032    
2033        /* Deal with capturing */        /* Deal with capturing */
2034    
2035        md->capture_last = number;        md->capture_last = (md->capture_last & OVFLMASK) | number;
2036        if (offset >= md->offset_max) md->offset_overflow = TRUE; else        if (offset >= md->offset_max) md->capture_last |= OVFLBIT; else
2037          {          {
2038          /* If offset is greater than offset_top, it means that we are          /* If offset is greater than offset_top, it means that we are
2039          "skipping" a capturing group, and that group's offsets must be marked          "skipping" a capturing group, and that group's offsets must be marked
# Line 1993  for (;;) Line 2113  for (;;)
2113          }          }
2114        if (*prev >= OP_SBRA)    /* Could match an empty string */        if (*prev >= OP_SBRA)    /* Could match an empty string */
2115          {          {
         md->match_function_type = MATCH_CBEGROUP;  
2116          RMATCH(eptr, prev, offset_top, md, eptrb, RM50);          RMATCH(eptr, prev, offset_top, md, eptrb, RM50);
2117          RRETURN(rrc);          RRETURN(rrc);
2118          }          }
# Line 2002  for (;;) Line 2121  for (;;)
2121        }        }
2122      else  /* OP_KETRMAX */      else  /* OP_KETRMAX */
2123        {        {
       if (*prev >= OP_SBRA) md->match_function_type = MATCH_CBEGROUP;  
2124        RMATCH(eptr, prev, offset_top, md, eptrb, RM13);        RMATCH(eptr, prev, offset_top, md, eptrb, RM13);
2125        if (rrc == MATCH_ONCE && md->once_target == prev) rrc = MATCH_NOMATCH;        if (rrc == MATCH_ONCE && md->once_target == prev) rrc = MATCH_NOMATCH;
2126        if (rrc != MATCH_NOMATCH) RRETURN(rrc);        if (rrc != MATCH_NOMATCH) RRETURN(rrc);
# Line 2059  for (;;) Line 2177  for (;;)
2177    
2178      case OP_DOLLM:      case OP_DOLLM:
2179      if (eptr < md->end_subject)      if (eptr < md->end_subject)
2180        { if (!IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH); }        {
2181          if (!IS_NEWLINE(eptr))
2182            {
2183            if (md->partial != 0 &&
2184                eptr + 1 >= md->end_subject &&
2185                NLBLOCK->nltype == NLTYPE_FIXED &&
2186                NLBLOCK->nllen == 2 &&
2187                RAWUCHARTEST(eptr) == NLBLOCK->nl[0])
2188              {
2189              md->hitend = TRUE;
2190              if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
2191              }
2192            RRETURN(MATCH_NOMATCH);
2193            }
2194          }
2195      else      else
2196        {        {
2197        if (md->noteol) RRETURN(MATCH_NOMATCH);        if (md->noteol) RRETURN(MATCH_NOMATCH);
# Line 2091  for (;;) Line 2223  for (;;)
2223      ASSERT_NL_OR_EOS:      ASSERT_NL_OR_EOS:
2224      if (eptr < md->end_subject &&      if (eptr < md->end_subject &&
2225          (!IS_NEWLINE(eptr) || eptr != md->end_subject - md->nllen))          (!IS_NEWLINE(eptr) || eptr != md->end_subject - md->nllen))
2226          {
2227          if (md->partial != 0 &&
2228              eptr + 1 >= md->end_subject &&
2229              NLBLOCK->nltype == NLTYPE_FIXED &&
2230              NLBLOCK->nllen == 2 &&
2231              RAWUCHARTEST(eptr) == NLBLOCK->nl[0])
2232            {
2233            md->hitend = TRUE;
2234            if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
2235            }
2236        RRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2237          }
2238    
2239      /* Either at end of string or \n before end. */      /* Either at end of string or \n before end. */
2240    
# Line 2219  for (;;) Line 2362  for (;;)
2362        }        }
2363      break;      break;
2364    
2365      /* Match a single character type; inline for speed */      /* Match any single character type except newline; have to take care with
2366        CRLF newlines and partial matching. */
2367    
2368      case OP_ANY:      case OP_ANY:
2369      if (IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH);      if (IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH);
2370        if (md->partial != 0 &&
2371            eptr + 1 >= md->end_subject &&
2372            NLBLOCK->nltype == NLTYPE_FIXED &&
2373            NLBLOCK->nllen == 2 &&
2374            RAWUCHARTEST(eptr) == NLBLOCK->nl[0])
2375          {
2376          md->hitend = TRUE;
2377          if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
2378          }
2379    
2380      /* Fall through */      /* Fall through */
2381    
2382        /* Match any single character whatsoever. */
2383    
2384      case OP_ALLANY:      case OP_ALLANY:
2385      if (eptr >= md->end_subject)   /* DO NOT merge the eptr++ here; it must */      if (eptr >= md->end_subject)   /* DO NOT merge the eptr++ here; it must */
2386        {                            /* not be updated before SCHECK_PARTIAL. */        {                            /* not be updated before SCHECK_PARTIAL. */
# Line 2364  for (;;) Line 2520  for (;;)
2520        {        {
2521        default: RRETURN(MATCH_NOMATCH);        default: RRETURN(MATCH_NOMATCH);
2522    
2523        case 0x000d:        case CHAR_CR:
2524        if (eptr < md->end_subject && *eptr == 0x0a) eptr++;        if (eptr >= md->end_subject)
2525            {
2526            SCHECK_PARTIAL();
2527            }
2528          else if (RAWUCHARTEST(eptr) == CHAR_LF) eptr++;
2529        break;        break;
2530    
2531        case 0x000a:        case CHAR_LF:
2532        break;        break;
2533    
2534        case 0x000b:        case CHAR_VT:
2535        case 0x000c:        case CHAR_FF:
2536        case 0x0085:        case CHAR_NEL:
2537    #ifndef EBCDIC
2538        case 0x2028:        case 0x2028:
2539        case 0x2029:        case 0x2029:
2540    #endif  /* Not EBCDIC */
2541        if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);        if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);
2542        break;        break;
2543        }        }
# Line 2391  for (;;) Line 2553  for (;;)
2553      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2554      switch(c)      switch(c)
2555        {        {
2556          HSPACE_CASES: RRETURN(MATCH_NOMATCH);  /* Byte and multibyte cases */
2557        default: break;        default: break;
       case 0x09:      /* HT */  
       case 0x20:      /* SPACE */  
       case 0xa0:      /* NBSP */  
       case 0x1680:    /* OGHAM SPACE MARK */  
       case 0x180e:    /* MONGOLIAN VOWEL SEPARATOR */  
       case 0x2000:    /* EN QUAD */  
       case 0x2001:    /* EM QUAD */  
       case 0x2002:    /* EN SPACE */  
       case 0x2003:    /* EM SPACE */  
       case 0x2004:    /* THREE-PER-EM SPACE */  
       case 0x2005:    /* FOUR-PER-EM SPACE */  
       case 0x2006:    /* SIX-PER-EM SPACE */  
       case 0x2007:    /* FIGURE SPACE */  
       case 0x2008:    /* PUNCTUATION SPACE */  
       case 0x2009:    /* THIN SPACE */  
       case 0x200A:    /* HAIR SPACE */  
       case 0x202f:    /* NARROW NO-BREAK SPACE */  
       case 0x205f:    /* MEDIUM MATHEMATICAL SPACE */  
       case 0x3000:    /* IDEOGRAPHIC SPACE */  
       RRETURN(MATCH_NOMATCH);  
2558        }        }
2559      ecode++;      ecode++;
2560      break;      break;
# Line 2425  for (;;) Line 2568  for (;;)
2568      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2569      switch(c)      switch(c)
2570        {        {
2571          HSPACE_CASES: break;  /* Byte and multibyte cases */
2572        default: RRETURN(MATCH_NOMATCH);        default: RRETURN(MATCH_NOMATCH);
       case 0x09:      /* HT */  
       case 0x20:      /* SPACE */  
       case 0xa0:      /* NBSP */  
       case 0x1680:    /* OGHAM SPACE MARK */  
       case 0x180e:    /* MONGOLIAN VOWEL SEPARATOR */  
       case 0x2000:    /* EN QUAD */  
       case 0x2001:    /* EM QUAD */  
       case 0x2002:    /* EN SPACE */  
       case 0x2003:    /* EM SPACE */  
       case 0x2004:    /* THREE-PER-EM SPACE */  
       case 0x2005:    /* FOUR-PER-EM SPACE */  
       case 0x2006:    /* SIX-PER-EM SPACE */  
       case 0x2007:    /* FIGURE SPACE */  
       case 0x2008:    /* PUNCTUATION SPACE */  
       case 0x2009:    /* THIN SPACE */  
       case 0x200A:    /* HAIR SPACE */  
       case 0x202f:    /* NARROW NO-BREAK SPACE */  
       case 0x205f:    /* MEDIUM MATHEMATICAL SPACE */  
       case 0x3000:    /* IDEOGRAPHIC SPACE */  
       break;  
2573        }        }
2574      ecode++;      ecode++;
2575      break;      break;
# Line 2459  for (;;) Line 2583  for (;;)
2583      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2584      switch(c)      switch(c)
2585        {        {
2586          VSPACE_CASES: RRETURN(MATCH_NOMATCH);
2587        default: break;        default: break;
       case 0x0a:      /* LF */  
       case 0x0b:      /* VT */  
       case 0x0c:      /* FF */  
       case 0x0d:      /* CR */  
       case 0x85:      /* NEL */  
       case 0x2028:    /* LINE SEPARATOR */  
       case 0x2029:    /* PARAGRAPH SEPARATOR */  
       RRETURN(MATCH_NOMATCH);  
2588        }        }
2589      ecode++;      ecode++;
2590      break;      break;
# Line 2481  for (;;) Line 2598  for (;;)
2598      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2599      switch(c)      switch(c)
2600        {        {
2601          VSPACE_CASES: break;
2602        default: RRETURN(MATCH_NOMATCH);        default: RRETURN(MATCH_NOMATCH);
       case 0x0a:      /* LF */  
       case 0x0b:      /* VT */  
       case 0x0c:      /* FF */  
       case 0x0d:      /* CR */  
       case 0x85:      /* NEL */  
       case 0x2028:    /* LINE SEPARATOR */  
       case 0x2029:    /* PARAGRAPH SEPARATOR */  
       break;  
2603        }        }
2604      ecode++;      ecode++;
2605      break;      break;
# Line 2507  for (;;) Line 2617  for (;;)
2617        }        }
2618      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2619        {        {
2620          const pcre_uint32 *cp;
2621        const ucd_record *prop = GET_UCD(c);        const ucd_record *prop = GET_UCD(c);
2622    
2623        switch(ecode[1])        switch(ecode[1])
# Line 2567  for (;;) Line 2678  for (;;)
2678            RRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
2679          break;          break;
2680    
2681            case PT_CLIST:
2682            cp = PRIV(ucd_caseless_sets) + ecode[2];
2683            for (;;)
2684              {
2685              if (c < *cp)
2686                { if (op == OP_PROP) { RRETURN(MATCH_NOMATCH); } else break; }
2687              if (c == *cp++)
2688                { if (op == OP_PROP) break; else { RRETURN(MATCH_NOMATCH); } }
2689              }
2690            break;
2691    
2692            case PT_UCNC:
2693            if ((c == CHAR_DOLLAR_SIGN || c == CHAR_COMMERCIAL_AT ||
2694                 c == CHAR_GRAVE_ACCENT || (c >= 0xa0 && c <= 0xd7ff) ||
2695                 c >= 0xe000) == (op == OP_NOTPROP))
2696              RRETURN(MATCH_NOMATCH);
2697            break;
2698    
2699          /* This should never occur */          /* This should never occur */
2700    
2701          default:          default:
# Line 2586  for (;;) Line 2715  for (;;)
2715        SCHECK_PARTIAL();        SCHECK_PARTIAL();
2716        RRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2717        }        }
2718      GETCHARINCTEST(c, eptr);      else
     if (UCD_CATEGORY(c) == ucp_M) RRETURN(MATCH_NOMATCH);  
     while (eptr < md->end_subject)  
2719        {        {
2720        int len = 1;        int lgb, rgb;
2721        if (!utf) c = *eptr; else { GETCHARLEN(c, eptr, len); }        GETCHARINCTEST(c, eptr);
2722        if (UCD_CATEGORY(c) != ucp_M) break;        lgb = UCD_GRAPHBREAK(c);
2723        eptr += len;        while (eptr < md->end_subject)
2724            {
2725            int len = 1;
2726            if (!utf) c = *eptr; else { GETCHARLEN(c, eptr, len); }
2727            rgb = UCD_GRAPHBREAK(c);
2728            if ((PRIV(ucp_gbtable)[lgb] & (1 << rgb)) == 0) break;
2729            lgb = rgb;
2730            eptr += len;
2731            }
2732        }        }
2733        CHECK_PARTIAL();
2734      ecode++;      ecode++;
2735      break;      break;
2736  #endif  #endif  /* SUPPORT_UCP */
2737    
2738    
2739      /* Match a back reference, possibly repeatedly. Look past the end of the      /* Match a back reference, possibly repeatedly. Look past the end of the
# Line 2660  for (;;) Line 2796  for (;;)
2796        default:               /* No repeat follows */        default:               /* No repeat follows */
2797        if ((length = match_ref(offset, eptr, length, md, caseless)) < 0)        if ((length = match_ref(offset, eptr, length, md, caseless)) < 0)
2798          {          {
2799            if (length == -2) eptr = md->end_subject;   /* Partial match */
2800          CHECK_PARTIAL();          CHECK_PARTIAL();
2801          RRETURN(MATCH_NOMATCH);          RRETURN(MATCH_NOMATCH);
2802          }          }
# Line 2685  for (;;) Line 2822  for (;;)
2822        int slength;        int slength;
2823        if ((slength = match_ref(offset, eptr, length, md, caseless)) < 0)        if ((slength = match_ref(offset, eptr, length, md, caseless)) < 0)
2824          {          {
2825            if (slength == -2) eptr = md->end_subject;   /* Partial match */
2826          CHECK_PARTIAL();          CHECK_PARTIAL();
2827          RRETURN(MATCH_NOMATCH);          RRETURN(MATCH_NOMATCH);
2828          }          }
# Line 2708  for (;;) Line 2846  for (;;)
2846          if (fi >= max) RRETURN(MATCH_NOMATCH);          if (fi >= max) RRETURN(MATCH_NOMATCH);
2847          if ((slength = match_ref(offset, eptr, length, md, caseless)) < 0)          if ((slength = match_ref(offset, eptr, length, md, caseless)) < 0)
2848            {            {
2849              if (slength == -2) eptr = md->end_subject;   /* Partial match */
2850            CHECK_PARTIAL();            CHECK_PARTIAL();
2851            RRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
2852            }            }
# Line 2726  for (;;) Line 2865  for (;;)
2865          int slength;          int slength;
2866          if ((slength = match_ref(offset, eptr, length, md, caseless)) < 0)          if ((slength = match_ref(offset, eptr, length, md, caseless)) < 0)
2867            {            {
2868            CHECK_PARTIAL();            /* Can't use CHECK_PARTIAL because we don't want to update eptr in
2869              the soft partial matching case. */
2870    
2871              if (slength == -2 && md->partial != 0 &&
2872                  md->end_subject > md->start_used_ptr)
2873                {
2874                md->hitend = TRUE;
2875                if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
2876                }
2877            break;            break;
2878            }            }
2879          eptr += slength;          eptr += slength;
2880          }          }
2881    
2882        while (eptr >= pp)        while (eptr >= pp)
2883          {          {
2884          RMATCH(eptr, ecode, offset_top, md, eptrb, RM15);          RMATCH(eptr, ecode, offset_top, md, eptrb, RM15);
# Line 3094  for (;;) Line 3242  for (;;)
3242          CHECK_PARTIAL();             /* Not SCHECK_PARTIAL() */          CHECK_PARTIAL();             /* Not SCHECK_PARTIAL() */
3243          RRETURN(MATCH_NOMATCH);          RRETURN(MATCH_NOMATCH);
3244          }          }
3245        while (length-- > 0) if (*ecode++ != *eptr++) RRETURN(MATCH_NOMATCH);        while (length-- > 0) if (*ecode++ != RAWUCHARINC(eptr)) RRETURN(MATCH_NOMATCH);
3246        }        }
3247      else      else
3248  #endif  #endif
# Line 3134  for (;;) Line 3282  for (;;)
3282    
3283        if (fc < 128)        if (fc < 128)
3284          {          {
3285          if (md->lcc[fc]          pcre_uint32 cc = RAWUCHAR(eptr);
3286              != TABLE_GET(*eptr, md->lcc, *eptr)) RRETURN(MATCH_NOMATCH);          if (md->lcc[fc] != TABLE_GET(cc, md->lcc, cc)) RRETURN(MATCH_NOMATCH);
3287          ecode++;          ecode++;
3288          eptr++;          eptr++;
3289          }          }
# Line 3146  for (;;) Line 3294  for (;;)
3294    
3295        else        else
3296          {          {
3297          unsigned int dc;          pcre_uint32 dc;
3298          GETCHARINC(dc, eptr);          GETCHARINC(dc, eptr);
3299          ecode += length;          ecode += length;
3300    
# Line 3239  for (;;) Line 3387  for (;;)
3387      max = rep_max[c];                 /* zero for max => infinity */      max = rep_max[c];                 /* zero for max => infinity */
3388      if (max == 0) max = INT_MAX;      if (max == 0) max = INT_MAX;
3389    
3390      /* Common code for all repeated single-character matches. */      /* Common code for all repeated single-character matches. We first check
3391        for the minimum number of characters. If the minimum equals the maximum, we
3392        are done. Otherwise, if minimizing, check the rest of the pattern for a
3393        match; if there isn't one, advance up to the maximum, one character at a
3394        time.
3395    
3396        If maximizing, advance up to the maximum number of matching characters,
3397        until eptr is past the end of the maximum run. If possessive, we are
3398        then done (no backing up). Otherwise, match at this position; anything
3399        other than no match is immediately returned. For nomatch, back up one
3400        character, unless we are matching \R and the last thing matched was
3401        \r\n, in which case, back up two bytes. When we reach the first optional
3402        character position, we can save stack by doing a tail recurse.
3403    
3404        The various UTF/non-UTF and caseful/caseless cases are handled separately,
3405        for speed. */
3406    
3407      REPEATCHAR:      REPEATCHAR:
3408  #ifdef SUPPORT_UTF  #ifdef SUPPORT_UTF
# Line 3256  for (;;) Line 3419  for (;;)
3419        if (length > 1)        if (length > 1)
3420          {          {
3421  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
3422          unsigned int othercase;          pcre_uint32 othercase;
3423          if (op >= OP_STARI &&     /* Caseless */          if (op >= OP_STARI &&     /* Caseless */
3424              (othercase = UCD_OTHERCASE(fc)) != fc)              (othercase = UCD_OTHERCASE(fc)) != fc)
3425            oclength = PRIV(ord2utf)(othercase, occhars);            oclength = PRIV(ord2utf)(othercase, occhars);
# Line 3323  for (;;) Line 3486  for (;;)
3486                }                }
3487              }              }
3488    
3489            if (possessive) continue;            if (possessive) continue;    /* No backtracking */
   
3490            for(;;)            for(;;)
3491              {              {
3492                if (eptr == pp) goto TAIL_RECURSE;
3493              RMATCH(eptr, ecode, offset_top, md, eptrb, RM23);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM23);
3494              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
             if (eptr == pp) { RRETURN(MATCH_NOMATCH); }  
3495  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
3496              eptr--;              eptr--;
3497              BACKCHAR(eptr);              BACKCHAR(eptr);
# Line 3360  for (;;) Line 3522  for (;;)
3522      maximizing, find the maximum number of characters and work backwards. */      maximizing, find the maximum number of characters and work backwards. */
3523    
3524      DPRINTF(("matching %c{%d,%d} against subject %.*s\n", fc, min, max,      DPRINTF(("matching %c{%d,%d} against subject %.*s\n", fc, min, max,
3525        max, eptr));        max, (char *)eptr));
3526    
3527      if (op >= OP_STARI)  /* Caseless */      if (op >= OP_STARI)  /* Caseless */
3528        {        {
# Line 3383  for (;;) Line 3545  for (;;)
3545    
3546        for (i = 1; i <= min; i++)        for (i = 1; i <= min; i++)
3547          {          {
3548            pcre_uint32 cc;                 /* Faster than pcre_uchar */
3549          if (eptr >= md->end_subject)          if (eptr >= md->end_subject)
3550            {            {
3551            SCHECK_PARTIAL();            SCHECK_PARTIAL();
3552            RRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
3553            }            }
3554          if (fc != *eptr && foc != *eptr) RRETURN(MATCH_NOMATCH);          cc = RAWUCHARTEST(eptr);
3555            if (fc != cc && foc != cc) RRETURN(MATCH_NOMATCH);
3556          eptr++;          eptr++;
3557          }          }
3558        if (min == max) continue;        if (min == max) continue;
# Line 3396  for (;;) Line 3560  for (;;)
3560          {          {
3561          for (fi = min;; fi++)          for (fi = min;; fi++)
3562            {            {
3563              pcre_uint32 cc;               /* Faster than pcre_uchar */
3564            RMATCH(eptr, ecode, offset_top, md, eptrb, RM24);            RMATCH(eptr, ecode, offset_top, md, eptrb, RM24);
3565            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3566            if (fi >= max) RRETURN(MATCH_NOMATCH);            if (fi >= max) RRETURN(MATCH_NOMATCH);
# Line 3404  for (;;) Line 3569  for (;;)
3569              SCHECK_PARTIAL();              SCHECK_PARTIAL();
3570              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
3571              }              }
3572            if (fc != *eptr && foc != *eptr) RRETURN(MATCH_NOMATCH);            cc = RAWUCHARTEST(eptr);
3573              if (fc != cc && foc != cc) RRETURN(MATCH_NOMATCH);
3574            eptr++;            eptr++;
3575            }            }
3576          /* Control never gets here */          /* Control never gets here */
# Line 3414  for (;;) Line 3580  for (;;)
3580          pp = eptr;          pp = eptr;
3581          for (i = min; i < max; i++)          for (i = min; i < max; i++)
3582            {            {
3583              pcre_uint32 cc;               /* Faster than pcre_uchar */
3584            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
3585              {              {
3586              SCHECK_PARTIAL();              SCHECK_PARTIAL();
3587              break;              break;
3588              }              }
3589            if (fc != *eptr && foc != *eptr) break;            cc = RAWUCHARTEST(eptr);
3590              if (fc != cc && foc != cc) break;
3591            eptr++;            eptr++;
3592            }            }
3593    
3594          if (possessive) continue;          if (possessive) continue;       /* No backtracking */
3595            for (;;)
         while (eptr >= pp)  
3596            {            {
3597              if (eptr == pp) goto TAIL_RECURSE;
3598            RMATCH(eptr, ecode, offset_top, md, eptrb, RM25);            RMATCH(eptr, ecode, offset_top, md, eptrb, RM25);
3599            eptr--;            eptr--;
3600            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
# Line 3447  for (;;) Line 3615  for (;;)
3615            SCHECK_PARTIAL();            SCHECK_PARTIAL();
3616            RRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
3617            }            }
3618          if (fc != *eptr++) RRETURN(MATCH_NOMATCH);          if (fc != RAWUCHARINCTEST(eptr)) RRETURN(MATCH_NOMATCH);
3619          }          }
3620    
3621        if (min == max) continue;        if (min == max) continue;
# Line 3464  for (;;) Line 3632  for (;;)
3632              SCHECK_PARTIAL();              SCHECK_PARTIAL();
3633              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
3634              }              }
3635            if (fc != *eptr++) RRETURN(MATCH_NOMATCH);            if (fc != RAWUCHARINCTEST(eptr)) RRETURN(MATCH_NOMATCH);
3636            }            }
3637          /* Control never gets here */          /* Control never gets here */
3638          }          }
# Line 3478  for (;;) Line 3646  for (;;)
3646              SCHECK_PARTIAL();              SCHECK_PARTIAL();
3647              break;              break;
3648              }              }
3649            if (fc != *eptr) break;            if (fc != RAWUCHARTEST(eptr)) break;
3650            eptr++;            eptr++;
3651            }            }
3652          if (possessive) continue;          if (possessive) continue;    /* No backtracking */
3653            for (;;)
         while (eptr >= pp)  
3654            {            {
3655              if (eptr == pp) goto TAIL_RECURSE;
3656            RMATCH(eptr, ecode, offset_top, md, eptrb, RM27);            RMATCH(eptr, ecode, offset_top, md, eptrb, RM27);
3657            eptr--;            eptr--;
3658            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
# Line 3504  for (;;) Line 3672  for (;;)
3672        SCHECK_PARTIAL();        SCHECK_PARTIAL();
3673        RRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
3674        }        }
     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  
3675  #ifdef SUPPORT_UTF  #ifdef SUPPORT_UTF
3676        if (utf)
3677          {
3678          register pcre_uint32 ch, och;
3679    
3680          ecode++;
3681          GETCHARINC(ch, ecode);
3682          GETCHARINC(c, eptr);
3683    
3684          if (op == OP_NOT)
3685            {
3686            if (ch == c) RRETURN(MATCH_NOMATCH);
3687            }
3688          else
3689            {
3690  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
3691        if (utf && ch > 127)          if (ch > 127)
3692          och = UCD_OTHERCASE(ch);            och = UCD_OTHERCASE(ch);
3693  #else  #else
3694        if (utf && ch > 127)          if (ch > 127)
3695          och = ch;            och = ch;
3696  #endif /* SUPPORT_UCP */  #endif /* SUPPORT_UCP */
3697        else          else
3698  #endif /* SUPPORT_UTF */            och = TABLE_GET(ch, md->fcc, ch);
3699          och = TABLE_GET(ch, md->fcc, ch);          if (ch == c || och == c) RRETURN(MATCH_NOMATCH);
3700  #endif /* COMPILE_PCRE8 */          }
       if (ch == c || och == c) RRETURN(MATCH_NOMATCH);  
3701        }        }
3702      else    /* Caseful */      else
3703    #endif
3704        {        {
3705        if (*ecode++ == c) RRETURN(MATCH_NOMATCH);        register pcre_uint32 ch = ecode[1];
3706          c = *eptr++;
3707          if (ch == c || (op == OP_NOTI && TABLE_GET(ch, md->fcc, ch) == c))
3708            RRETURN(MATCH_NOMATCH);
3709          ecode += 2;
3710        }        }
3711      break;      break;
3712    
# Line 3610  for (;;) Line 3786  for (;;)
3786      /* Common code for all repeated single-byte matches. */      /* Common code for all repeated single-byte matches. */
3787    
3788      REPEATNOTCHAR:      REPEATNOTCHAR:
3789      fc = *ecode++;      GETCHARINCTEST(fc, ecode);
3790    
3791      /* The code is duplicated for the caseless and caseful cases, for speed,      /* The code is duplicated for the caseless and caseful cases, for speed,
3792      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 3797  for (;;)
3797      characters and work backwards. */      characters and work backwards. */
3798    
3799      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,
3800        max, eptr));        max, (char *)eptr));
3801    
3802      if (op >= OP_NOTSTARI)     /* Caseless */      if (op >= OP_NOTSTARI)     /* Caseless */
3803        {        {
 #ifdef COMPILE_PCRE8  
       /* fc must be < 128 if UTF is enabled. */  
       foc = md->fcc[fc];  
 #else  
3804  #ifdef SUPPORT_UTF  #ifdef SUPPORT_UTF
3805  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
3806        if (utf && fc > 127)        if (utf && fc > 127)
# Line 3640  for (;;) Line 3812  for (;;)
3812        else        else
3813  #endif /* SUPPORT_UTF */  #endif /* SUPPORT_UTF */
3814          foc = TABLE_GET(fc, md->fcc, fc);          foc = TABLE_GET(fc, md->fcc, fc);
 #endif /* COMPILE_PCRE8 */  
3815    
3816  #ifdef SUPPORT_UTF  #ifdef SUPPORT_UTF
3817        if (utf)        if (utf)
3818          {          {
3819          register unsigned int d;          register pcre_uint32 d;
3820          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
3821            {            {
3822            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
# Line 3654  for (;;) Line 3825  for (;;)
3825              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
3826              }              }
3827            GETCHARINC(d, eptr);            GETCHARINC(d, eptr);
3828            if (fc == d || foc == d) RRETURN(MATCH_NOMATCH);            if (fc == d || (unsigned int)foc == d) RRETURN(MATCH_NOMATCH);
3829            }            }
3830          }          }
3831        else        else
3832  #endif  #endif  /* SUPPORT_UTF */
3833        /* Not UTF mode */        /* Not UTF mode */
3834          {          {
3835          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
# Line 3680  for (;;) Line 3851  for (;;)
3851  #ifdef SUPPORT_UTF  #ifdef SUPPORT_UTF
3852          if (utf)          if (utf)
3853            {            {
3854            register unsigned int d;            register pcre_uint32 d;
3855            for (fi = min;; fi++)            for (fi = min;; fi++)
3856              {              {
3857              RMATCH(eptr, ecode, offset_top, md, eptrb, RM28);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM28);
# Line 3692  for (;;) Line 3863  for (;;)
3863                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
3864                }                }
3865              GETCHARINC(d, eptr);              GETCHARINC(d, eptr);
3866              if (fc == d || foc == d) RRETURN(MATCH_NOMATCH);              if (fc == d || (unsigned int)foc == d) RRETURN(MATCH_NOMATCH);
3867              }              }
3868            }            }
3869          else          else
3870  #endif  #endif  /*SUPPORT_UTF */
3871          /* Not UTF mode */          /* Not UTF mode */
3872            {            {
3873            for (fi = min;; fi++)            for (fi = min;; fi++)
# Line 3725  for (;;) Line 3896  for (;;)
3896  #ifdef SUPPORT_UTF  #ifdef SUPPORT_UTF
3897          if (utf)          if (utf)
3898            {            {
3899            register unsigned int d;            register pcre_uint32 d;
3900            for (i = min; i < max; i++)            for (i = min; i < max; i++)
3901              {              {
3902              int len = 1;              int len = 1;
# Line 3735  for (;;) Line 3906  for (;;)
3906                break;                break;
3907                }                }
3908              GETCHARLEN(d, eptr, len);              GETCHARLEN(d, eptr, len);
3909              if (fc == d || foc == d) break;              if (fc == d || (unsigned int)foc == d) break;
3910              eptr += len;              eptr += len;
3911              }              }
3912            if (possessive) continue;            if (possessive) continue;    /* No backtracking */
3913            for(;;)            for(;;)
3914              {              {
3915                if (eptr == pp) goto TAIL_RECURSE;
3916              RMATCH(eptr, ecode, offset_top, md, eptrb, RM30);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM30);
3917              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3918              if (eptr-- == pp) break;        /* Stop if tried at original pos */              eptr--;
3919              BACKCHAR(eptr);              BACKCHAR(eptr);
3920              }              }
3921            }            }
3922          else          else
3923  #endif  #endif  /* SUPPORT_UTF */
3924          /* Not UTF mode */          /* Not UTF mode */
3925            {            {
3926            for (i = min; i < max; i++)            for (i = min; i < max; i++)
# Line 3761  for (;;) Line 3933  for (;;)
3933              if (fc == *eptr || foc == *eptr) break;              if (fc == *eptr || foc == *eptr) break;
3934              eptr++;              eptr++;
3935              }              }
3936            if (possessive) continue;            if (possessive) continue;    /* No backtracking */
3937            while (eptr >= pp)            for (;;)
3938              {              {
3939                if (eptr == pp) goto TAIL_RECURSE;
3940              RMATCH(eptr, ecode, offset_top, md, eptrb, RM31);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM31);
3941              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3942              eptr--;              eptr--;
# Line 3782  for (;;) Line 3955  for (;;)
3955  #ifdef SUPPORT_UTF  #ifdef SUPPORT_UTF
3956        if (utf)        if (utf)
3957          {          {
3958          register unsigned int d;          register pcre_uint32 d;
3959          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
3960            {            {
3961            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
# Line 3816  for (;;) Line 3989  for (;;)
3989  #ifdef SUPPORT_UTF  #ifdef SUPPORT_UTF
3990          if (utf)          if (utf)
3991            {            {
3992            register unsigned int d;            register pcre_uint32 d;
3993            for (fi = min;; fi++)            for (fi = min;; fi++)
3994              {              {
3995              RMATCH(eptr, ecode, offset_top, md, eptrb, RM32);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM32);
# Line 3860  for (;;) Line 4033  for (;;)
4033  #ifdef SUPPORT_UTF  #ifdef SUPPORT_UTF
4034          if (utf)          if (utf)
4035            {            {
4036            register unsigned int d;            register pcre_uint32 d;
4037            for (i = min; i < max; i++)            for (i = min; i < max; i++)
4038              {              {
4039              int len = 1;              int len = 1;
# Line 3873  for (;;) Line 4046  for (;;)
4046              if (fc == d) break;              if (fc == d) break;
4047              eptr += len;              eptr += len;
4048              }              }
4049            if (possessive) continue;            if (possessive) continue;    /* No backtracking */
4050            for(;;)            for(;;)
4051              {              {
4052                if (eptr == pp) goto TAIL_RECURSE;
4053              RMATCH(eptr, ecode, offset_top, md, eptrb, RM34);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM34);
4054              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
4055              if (eptr-- == pp) break;        /* Stop if tried at original pos */              eptr--;
4056              BACKCHAR(eptr);              BACKCHAR(eptr);
4057              }              }
4058            }            }
# Line 3896  for (;;) Line 4070  for (;;)
4070              if (fc == *eptr) break;              if (fc == *eptr) break;
4071              eptr++;              eptr++;
4072              }              }
4073            if (possessive) continue;            if (possessive) continue;    /* No backtracking */
4074            while (eptr >= pp)            for (;;)
4075              {              {
4076                if (eptr == pp) goto TAIL_RECURSE;
4077              RMATCH(eptr, ecode, offset_top, md, eptrb, RM35);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM35);
4078              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
4079              eptr--;              eptr--;
# Line 4136  for (;;) Line 4311  for (;;)
4311              }              }
4312            break;            break;
4313    
4314              case PT_CLIST:
4315              for (i = 1; i <= min; i++)
4316                {
4317                const pcre_uint32 *cp;
4318                if (eptr >= md->end_subject)
4319                  {
4320                  SCHECK_PARTIAL();
4321                  RRETURN(MATCH_NOMATCH);
4322                  }
4323                GETCHARINCTEST(c, eptr);
4324                cp = PRIV(ucd_caseless_sets) + prop_value;
4325                for (;;)
4326                  {
4327                  if (c < *cp)
4328                    { if (prop_fail_result) break; else { RRETURN(MATCH_NOMATCH); } }
4329                  if (c == *cp++)
4330                    { if (prop_fail_result) { RRETURN(MATCH_NOMATCH); } else break; }
4331                  }
4332                }
4333              break;
4334    
4335              case PT_UCNC:
4336              for (i = 1; i <= min; i++)
4337                {
4338                if (eptr >= md->end_subject)
4339                  {
4340                  SCHECK_PARTIAL();
4341                  RRETURN(MATCH_NOMATCH);
4342                  }
4343                GETCHARINCTEST(c, eptr);
4344                if ((c == CHAR_DOLLAR_SIGN || c == CHAR_COMMERCIAL_AT ||
4345                     c == CHAR_GRAVE_ACCENT || (c >= 0xa0 && c <= 0xd7ff) ||
4346                     c >= 0xe000) == prop_fail_result)
4347                  RRETURN(MATCH_NOMATCH);
4348                }
4349              break;
4350    
4351            /* This should not occur */            /* This should not occur */
4352    
4353            default:            default:
# Line 4155  for (;;) Line 4367  for (;;)
4367              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4368              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4369              }              }
4370            GETCHARINCTEST(c, eptr);            else
           if (UCD_CATEGORY(c) == ucp_M) RRETURN(MATCH_NOMATCH);  
           while (eptr < md->end_subject)  
4371              {              {
4372              int len = 1;              int lgb, rgb;
4373              if (!utf) c = *eptr; else { GETCHARLEN(c, eptr, len); }              GETCHARINCTEST(c, eptr);
4374              if (UCD_CATEGORY(c) != ucp_M) break;              lgb = UCD_GRAPHBREAK(c);
4375              eptr += len;             while (eptr < md->end_subject)
4376                  {
4377                  int len = 1;
4378                  if (!utf) c = *eptr; else { GETCHARLEN(c, eptr, len); }
4379                  rgb = UCD_GRAPHBREAK(c);
4380                  if ((PRIV(ucp_gbtable)[lgb] & (1 << rgb)) == 0) break;
4381                  lgb = rgb;
4382                  eptr += len;
4383                  }
4384              }              }
4385              CHECK_PARTIAL();
4386            }            }
4387          }          }
4388    
# Line 4184  for (;;) Line 4403  for (;;)
4403              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4404              }              }
4405            if (IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH);            if (IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH);
4406              if (md->partial != 0 &&
4407                  eptr + 1 >= md->end_subject &&
4408                  NLBLOCK->nltype == NLTYPE_FIXED &&
4409                  NLBLOCK->nllen == 2 &&
4410                  RAWUCHAR(eptr) == NLBLOCK->nl[0])
4411                {
4412                md->hitend = TRUE;
4413                if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
4414                }
4415            eptr++;            eptr++;
4416            ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);            ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);
4417            }            }
# Line 4220  for (;;) Line 4448  for (;;)
4448              {              {
4449              default: RRETURN(MATCH_NOMATCH);              default: RRETURN(MATCH_NOMATCH);
4450    
4451              case 0x000d:              case CHAR_CR:
4452              if (eptr < md->end_subject && *eptr == 0x0a) eptr++;              if (eptr < md->end_subject && RAWUCHAR(eptr) == CHAR_LF) eptr++;
4453              break;              break;
4454    
4455              case 0x000a:              case CHAR_LF:
4456              break;              break;
4457    
4458              case 0x000b:              case CHAR_VT:
4459              case 0x000c:              case CHAR_FF:
4460              case 0x0085:              case CHAR_NEL:
4461    #ifndef EBCDIC
4462              case 0x2028:              case 0x2028:
4463              case 0x2029:              case 0x2029:
4464    #endif  /* Not EBCDIC */
4465              if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);              if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);
4466              break;              break;
4467              }              }
# Line 4249  for (;;) Line 4479  for (;;)
4479            GETCHARINC(c, eptr);            GETCHARINC(c, eptr);
4480            switch(c)            switch(c)
4481              {              {
4482                HSPACE_CASES: RRETURN(MATCH_NOMATCH);  /* Byte and multibyte cases */
4483              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);  
4484              }              }
4485            }            }
4486          break;          break;
# Line 4285  for (;;) Line 4496  for (;;)
4496            GETCHARINC(c, eptr);            GETCHARINC(c, eptr);
4497            switch(c)            switch(c)
4498              {              {
4499                HSPACE_CASES: break;  /* Byte and multibyte cases */
4500              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;  
4501              }              }
4502            }            }
4503          break;          break;
# Line 4321  for (;;) Line 4513  for (;;)
4513            GETCHARINC(c, eptr);            GETCHARINC(c, eptr);
4514            switch(c)            switch(c)
4515              {              {
4516                VSPACE_CASES: RRETURN(MATCH_NOMATCH);
4517              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);  
4518              }              }
4519            }            }
4520          break;          break;
# Line 4345  for (;;) Line 4530  for (;;)
4530            GETCHARINC(c, eptr);            GETCHARINC(c, eptr);
4531            switch(c)            switch(c)
4532              {              {
4533                VSPACE_CASES: break;
4534              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;  
4535              }              }
4536            }            }
4537          break;          break;
# Line 4375  for (;;) Line 4553  for (;;)
4553          case OP_DIGIT:          case OP_DIGIT:
4554          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
4555            {            {
4556              pcre_uint32 cc;
4557            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
4558              {              {
4559              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4560              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4561              }              }
4562            if (*eptr >= 128 || (md->ctypes[*eptr] & ctype_digit) == 0)            cc = RAWUCHAR(eptr);
4563              if (cc >= 128 || (md->ctypes[cc] & ctype_digit) == 0)
4564              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4565            eptr++;            eptr++;
4566            /* 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 4570  for (;;)
4570          case OP_NOT_WHITESPACE:          case OP_NOT_WHITESPACE:
4571          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
4572            {            {
4573              pcre_uint32 cc;
4574            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
4575              {              {
4576              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4577              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4578              }              }
4579            if (*eptr < 128 && (md->ctypes[*eptr] & ctype_space) != 0)            cc = RAWUCHAR(eptr);
4580              if (cc < 128 && (md->ctypes[cc] & ctype_space) != 0)
4581              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4582            eptr++;            eptr++;
4583            ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);            ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);
# Line 4405  for (;;) Line 4587  for (;;)
4587          case OP_WHITESPACE:          case OP_WHITESPACE:
4588          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
4589            {            {
4590              pcre_uint32 cc;
4591            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
4592              {              {
4593              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4594              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4595              }              }
4596            if (*eptr >= 128 || (md->ctypes[*eptr] & ctype_space) == 0)            cc = RAWUCHAR(eptr);
4597              if (cc >= 128 || (md->ctypes[cc] & ctype_space) == 0)
4598              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4599            eptr++;            eptr++;
4600            /* 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 4604  for (;;)
4604          case OP_NOT_WORDCHAR:          case OP_NOT_WORDCHAR:
4605          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
4606            {            {
4607              pcre_uint32 cc;
4608            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
4609              {              {
4610              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4611              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4612              }              }
4613            if (*eptr < 128 && (md->ctypes[*eptr] & ctype_word) != 0)            cc = RAWUCHAR(eptr);
4614              if (cc < 128 && (md->ctypes[cc] & ctype_word) != 0)
4615              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4616            eptr++;            eptr++;
4617            ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);            ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);
# Line 4435  for (;;) Line 4621  for (;;)
4621          case OP_WORDCHAR:          case OP_WORDCHAR:
4622          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
4623            {            {
4624              pcre_uint32 cc;
4625            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
4626              {              {
4627              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4628              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4629              }              }
4630            if (*eptr >= 128 || (md->ctypes[*eptr] & ctype_word) == 0)            cc = RAWUCHAR(eptr);
4631              if (cc >= 128 || (md->ctypes[cc] & ctype_word) == 0)
4632              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4633            eptr++;            eptr++;
4634            /* 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 4656  for (;;)
4656              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4657              }              }
4658            if (IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH);            if (IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH);
4659              if (md->partial != 0 &&
4660                  eptr + 1 >= md->end_subject &&
4661                  NLBLOCK->nltype == NLTYPE_FIXED &&
4662                  NLBLOCK->nllen == 2 &&
4663                  *eptr == NLBLOCK->nl[0])
4664                {
4665                md->hitend = TRUE;
4666                if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
4667                }
4668            eptr++;            eptr++;
4669            }            }
4670          break;          break;
# Line 4502  for (;;) Line 4699  for (;;)
4699              {              {
4700              default: RRETURN(MATCH_NOMATCH);              default: RRETURN(MATCH_NOMATCH);
4701    
4702              case 0x000d:              case CHAR_CR:
4703              if (eptr < md->end_subject && *eptr == 0x0a) eptr++;              if (eptr < md->end_subject && *eptr == CHAR_LF) eptr++;
4704              break;              break;
4705    
4706              case 0x000a:              case CHAR_LF:
4707              break;              break;
4708    
4709              case 0x000b:              case CHAR_VT:
4710              case 0x000c:              case CHAR_FF:
4711              case 0x0085:              case CHAR_NEL:
4712  #ifdef COMPILE_PCRE16  #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
4713              case 0x2028:              case 0x2028:
4714              case 0x2029:              case 0x2029:
4715  #endif  #endif
# Line 4533  for (;;) Line 4730  for (;;)
4730            switch(*eptr++)            switch(*eptr++)
4731              {              {
4732              default: break;              default: break;
4733              case 0x09:      /* HT */              HSPACE_BYTE_CASES:
4734              case 0x20:      /* SPACE */  #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
4735              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 */  
4736  #endif  #endif
4737              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4738              }              }
# Line 4570  for (;;) Line 4750  for (;;)
4750            switch(*eptr++)            switch(*eptr++)
4751              {              {
4752              default: RRETURN(MATCH_NOMATCH);              default: RRETURN(MATCH_NOMATCH);
4753              case 0x09:      /* HT */              HSPACE_BYTE_CASES:
4754              case 0x20:      /* SPACE */  #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
4755              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 */  
4756  #endif  #endif
4757              break;              break;
4758              }              }
# Line 4606  for (;;) Line 4769  for (;;)
4769              }              }
4770            switch(*eptr++)            switch(*eptr++)
4771              {              {
4772              default: break;              VSPACE_BYTE_CASES:
4773              case 0x0a:      /* LF */  #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
4774              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 */  
4775  #endif  #endif
4776              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4777                default: break;
4778              }              }
4779            }            }
4780          break;          break;
# Line 4632  for (;;) Line 4790  for (;;)
4790            switch(*eptr++)            switch(*eptr++)
4791              {              {
4792              default: RRETURN(MATCH_NOMATCH);              default: RRETURN(MATCH_NOMATCH);
4793              case 0x0a:      /* LF */              VSPACE_BYTE_CASES:
4794              case 0x0b:      /* VT */  #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
4795              case 0x0c:      /* FF */              VSPACE_MULTIBYTE_CASES:
             case 0x0d:      /* CR */  
             case 0x85:      /* NEL */  
 #ifdef COMPILE_PCRE16  
             case 0x2028:    /* LINE SEPARATOR */  
             case 0x2029:    /* PARAGRAPH SEPARATOR */  
4796  #endif  #endif
4797              break;              break;
4798              }              }
# Line 4917  for (;;) Line 5070  for (;;)
5070              }              }
5071            /* Control never gets here */            /* Control never gets here */
5072    
5073            /* This should never occur */            case PT_CLIST:
5074              for (fi = min;; fi++)
5075                {
5076                const pcre_uint32 *cp;
5077                RMATCH(eptr, ecode, offset_top, md, eptrb, RM67);
5078                if (rrc != MATCH_NOMATCH) RRETURN(rrc);
5079                if (fi >= max) RRETURN(MATCH_NOMATCH);
5080                if (eptr >= md->end_subject)
5081                  {
5082                  SCHECK_PARTIAL();
5083                  RRETURN(MATCH_NOMATCH);
5084                  }
5085                GETCHARINCTEST(c, eptr);
5086                cp = PRIV(ucd_caseless_sets) + prop_value;
5087                for (;;)
5088                  {
5089                  if (c < *cp)
5090                    { if (prop_fail_result) break; else { RRETURN(MATCH_NOMATCH); } }
5091                  if (c == *cp++)
5092                    { if (prop_fail_result) { RRETURN(MATCH_NOMATCH); } else break; }
5093                  }
5094                }
5095              /* Control never gets here */
5096    
5097              case PT_UCNC:
5098              for (fi = min;; fi++)
5099                {
5100                RMATCH(eptr, ecode, offset_top, md, eptrb, RM68);
5101                if (rrc != MATCH_NOMATCH) RRETURN(rrc);
5102                if (fi >= max) RRETURN(MATCH_NOMATCH);
5103                if (eptr >= md->end_subject)
5104                  {
5105                  SCHECK_PARTIAL();
5106                  RRETURN(MATCH_NOMATCH);
5107                  }
5108                GETCHARINCTEST(c, eptr);
5109                if ((c == CHAR_DOLLAR_SIGN || c == CHAR_COMMERCIAL_AT ||
5110                     c == CHAR_GRAVE_ACCENT || (c >= 0xa0 && c <= 0xd7ff) ||
5111                     c >= 0xe000) == prop_fail_result)
5112                  RRETURN(MATCH_NOMATCH);
5113                }
5114              /* Control never gets here */
5115    
5116              /* This should never occur */
5117            default:            default:
5118            RRETURN(PCRE_ERROR_INTERNAL);            RRETURN(PCRE_ERROR_INTERNAL);
5119            }            }
# Line 4939  for (;;) Line 5134  for (;;)
5134              SCHECK_PARTIAL();              SCHECK_PARTIAL();
5135              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
5136              }              }
5137            GETCHARINCTEST(c, eptr);            else
           if (UCD_CATEGORY(c) == ucp_M) RRETURN(MATCH_NOMATCH);  
           while (eptr < md->end_subject)  
5138              {              {
5139              int len = 1;              int lgb, rgb;
5140              if (!utf) c = *eptr; else { GETCHARLEN(c, eptr, len); }              GETCHARINCTEST(c, eptr);
5141              if (UCD_CATEGORY(c) != ucp_M) break;              lgb = UCD_GRAPHBREAK(c);
5142              eptr += len;              while (eptr < md->end_subject)
5143                  {
5144                  int len = 1;
5145                  if (!utf) c = *eptr; else { GETCHARLEN(c, eptr, len); }
5146                  rgb = UCD_GRAPHBREAK(c);
5147                  if ((PRIV(ucp_gbtable)[lgb] & (1 << rgb)) == 0) break;
5148                  lgb = rgb;
5149                  eptr += len;
5150                  }
5151              }              }
5152              CHECK_PARTIAL();
5153            }            }
5154          }          }
5155        else        else
# Line 4971  for (;;) Line 5173  for (;;)
5173            GETCHARINC(c, eptr);            GETCHARINC(c, eptr);
5174            switch(ctype)            switch(ctype)
5175              {              {
5176              case OP_ANY:        /* This is the non-NL case */              case OP_ANY:               /* This is the non-NL case */
5177                if (md->partial != 0 &&    /* Take care with CRLF partial */
5178                    eptr >= md->end_subject &&
5179                    NLBLOCK->nltype == NLTYPE_FIXED &&
5180                    NLBLOCK->nllen == 2 &&
5181                    c == NLBLOCK->nl[0])
5182                  {
5183                  md->hitend = TRUE;
5184                  if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
5185                  }
5186                break;
5187    
5188              case OP_ALLANY:              case OP_ALLANY:
5189              case OP_ANYBYTE:              case OP_ANYBYTE:
5190              break;              break;
# Line 4980  for (;;) Line 5193  for (;;)
5193              switch(c)              switch(c)
5194                {                {
5195                default: RRETURN(MATCH_NOMATCH);                default: RRETURN(MATCH_NOMATCH);
5196                case 0x000d:                case CHAR_CR:
5197                if (eptr < md->end_subject && *eptr == 0x0a) eptr++;                if (eptr < md->end_subject && RAWUCHAR(eptr) == CHAR_LF) eptr++;
5198                break;                break;
5199                case 0x000a:  
5200                  case CHAR_LF:
5201                break;                break;
5202    
5203                case 0x000b:                case CHAR_VT:
5204                case 0x000c:                case CHAR_FF:
5205                case 0x0085:                case CHAR_NEL:
5206    #ifndef EBCDIC
5207                case 0x2028:                case 0x2028:
5208                case 0x2029:                case 0x2029:
5209    #endif  /* Not EBCDIC */
5210                if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);                if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);
5211                break;                break;
5212                }                }
# Line 4999  for (;;) Line 5215  for (;;)
5215              case OP_NOT_HSPACE:              case OP_NOT_HSPACE:
5216              switch(c)              switch(c)
5217                {                {
5218                  HSPACE_CASES: RRETURN(MATCH_NOMATCH);
5219                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);  
5220                }                }
5221              break;              break;
5222    
5223              case OP_HSPACE:              case OP_HSPACE:
5224              switch(c)              switch(c)
5225                {                {
5226                  HSPACE_CASES: break;
5227                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;  
5228                }                }
5229              break;              break;
5230    
5231              case OP_NOT_VSPACE:              case OP_NOT_VSPACE:
5232              switch(c)              switch(c)
5233                {                {
5234                  VSPACE_CASES: RRETURN(MATCH_NOMATCH);
5235                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);  
5236                }                }
5237              break;              break;
5238    
5239              case OP_VSPACE:              case OP_VSPACE:
5240              switch(c)              switch(c)
5241                {                {
5242                  VSPACE_CASES: break;
5243                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;  
5244                }                }
5245              break;              break;
5246    
# Line 5134  for (;;) Line 5298  for (;;)
5298            c = *eptr++;            c = *eptr++;
5299            switch(ctype)            switch(ctype)
5300              {              {
5301              case OP_ANY:     /* This is the non-NL case */              case OP_ANY:               /* This is the non-NL case */
5302                if (md->partial != 0 &&    /* Take care with CRLF partial */
5303                    eptr >= md->end_subject &&
5304                    NLBLOCK->nltype == NLTYPE_FIXED &&
5305                    NLBLOCK->nllen == 2 &&
5306                    c == NLBLOCK->nl[0])
5307                  {
5308                  md->hitend = TRUE;
5309                  if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
5310                  }
5311                break;
5312    
5313              case OP_ALLANY:              case OP_ALLANY:
5314              case OP_ANYBYTE:              case OP_ANYBYTE:
5315              break;              break;
# Line 5143  for (;;) Line 5318  for (;;)
5318              switch(c)              switch(c)
5319                {                {
5320                default: RRETURN(MATCH_NOMATCH);                default: RRETURN(MATCH_NOMATCH);
5321                case 0x000d:                case CHAR_CR:
5322                if (eptr < md->end_subject && *eptr == 0x0a) eptr++;                if (eptr < md->end_subject && *eptr == CHAR_LF) eptr++;
5323                break;                break;
5324    
5325                case 0x000a:                case CHAR_LF:
5326                break;                break;
5327    
5328                case 0x000b:                case CHAR_VT:
5329                case 0x000c:                case CHAR_FF:
5330                case 0x0085:                case CHAR_NEL:
5331  #ifdef COMPILE_PCRE16  #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
5332                case 0x2028:                case 0x2028:
5333                case 0x2029:                case 0x2029:
5334  #endif  #endif
# Line 5166  for (;;) Line 5341  for (;;)
5341              switch(c)              switch(c)
5342                {                {
5343                default: break;                default: break;
5344                case 0x09:      /* HT */                HSPACE_BYTE_CASES:
5345                case 0x20:      /* SPACE */  #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
5346                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 */  
5347  #endif  #endif
5348                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
5349                }                }
# Line 5195  for (;;) Line 5353  for (;;)
5353              switch(c)              switch(c)
5354                {                {
5355                default: RRETURN(MATCH_NOMATCH);                default: RRETURN(MATCH_NOMATCH);
5356                case 0x09:      /* HT */                HSPACE_BYTE_CASES:
5357                case 0x20:      /* SPACE */  #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
5358                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 */  
5359  #endif  #endif
5360                break;                break;
5361                }                }
# Line 5224  for (;;) Line 5365  for (;;)
5365              switch(c)              switch(c)
5366                {                {
5367                default: break;                default: break;
5368                case 0x0a:      /* LF */                VSPACE_BYTE_CASES:
5369                case 0x0b:      /* VT */  #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
5370                case 0x0c:      /* FF */                VSPACE_MULTIBYTE_CASES:
               case 0x0d:      /* CR */  
               case 0x85:      /* NEL */  
 #ifdef COMPILE_PCRE16  
               case 0x2028:    /* LINE SEPARATOR */  
               case 0x2029:    /* PARAGRAPH SEPARATOR */  
5371  #endif  #endif
5372                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
5373                }                }
# Line 5241  for (;;) Line 5377  for (;;)
5377              switch(c)              switch(c)
5378                {                {
5379                default: RRETURN(MATCH_NOMATCH);                default: RRETURN(MATCH_NOMATCH);
5380                case 0x0a:      /* LF */                VSPACE_BYTE_CASES:
5381                case 0x0b:      /* VT */  #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
5382                case 0x0c:      /* FF */                VSPACE_MULTIBYTE_CASES:
               case 0x0d:      /* CR */  
               case 0x85:      /* NEL */  
 #ifdef COMPILE_PCRE16  
               case 0x2028:    /* LINE SEPARATOR */  
               case 0x2029:    /* PARAGRAPH SEPARATOR */  
5383  #endif  #endif
5384                break;                break;
5385                }                }
# Line 5452  for (;;) Line 5583  for (;;)
5583              }              }
5584            break;            break;
5585    
5586              case PT_CLIST:
5587              for (i = min; i < max; i++)
5588                {
5589                const pcre_uint32 *cp;
5590                int len = 1;
5591                if (eptr >= md->end_subject)
5592                  {
5593                  SCHECK_PARTIAL();
5594                  break;
5595                  }
5596                GETCHARLENTEST(c, eptr, len);
5597                cp = PRIV(ucd_caseless_sets) + prop_value;
5598                for (;;)
5599                  {
5600                  if (c < *cp)
5601                    { if (prop_fail_result) break; else goto GOT_MAX; }
5602                  if (c == *cp++)
5603                    { if (prop_fail_result) goto GOT_MAX; else break; }
5604                  }
5605                eptr += len;
5606                }
5607              GOT_MAX:
5608              break;
5609    
5610              case PT_UCNC:
5611              for (i = min; i < max; i++)
5612                {
5613                int len = 1;
5614                if (eptr >= md->end_subject)
5615                  {
5616                  SCHECK_PARTIAL();
5617                  break;
5618                  }
5619                GETCHARLENTEST(c, eptr, len);
5620                if ((c == CHAR_DOLLAR_SIGN || c == CHAR_COMMERCIAL_AT ||
5621                     c == CHAR_GRAVE_ACCENT || (c >= 0xa0 && c <= 0xd7ff) ||
5622                     c >= 0xe000) == prop_fail_result)
5623                  break;
5624                eptr += len;
5625                }
5626              break;
5627    
5628            default:            default:
5629            RRETURN(PCRE_ERROR_INTERNAL);            RRETURN(PCRE_ERROR_INTERNAL);
5630            }            }
5631    
5632          /* eptr is now past the end of the maximum run */          /* eptr is now past the end of the maximum run */
5633    
5634          if (possessive) continue;          if (possessive) continue;    /* No backtracking */
5635          for(;;)          for(;;)
5636            {            {
5637              if (eptr == pp) goto TAIL_RECURSE;
5638            RMATCH(eptr, ecode, offset_top, md, eptrb, RM44);            RMATCH(eptr, ecode, offset_top, md, eptrb, RM44);
5639            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
5640            if (eptr-- == pp) break;        /* Stop if tried at original pos */            eptr--;
5641            if (utf) BACKCHAR(eptr);            if (utf) BACKCHAR(eptr);
5642            }            }
5643          }          }
# Line 5475  for (;;) Line 5649  for (;;)
5649          {          {
5650          for (i = min; i < max; i++)          for (i = min; i < max; i++)
5651            {            {
           int len = 1;  
5652            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
5653              {              {
5654              SCHECK_PARTIAL();              SCHECK_PARTIAL();
5655              break;              break;
5656              }              }
5657            if (!utf) c = *eptr; else { GETCHARLEN(c, eptr, len); }            else
           if (UCD_CATEGORY(c) == ucp_M) break;  
           eptr += len;  
           while (eptr < md->end_subject)  
5658              {              {
5659              len = 1;              int lgb, rgb;
5660              if (!utf) c = *eptr; else { GETCHARLEN(c, eptr, len); }              GETCHARINCTEST(c, eptr);
5661              if (UCD_CATEGORY(c) != ucp_M) break;              lgb = UCD_GRAPHBREAK(c);
5662              eptr += len;              while (eptr < md->end_subject)
5663                  {
5664                  int len = 1;
5665                  if (!utf) c = *eptr; else { GETCHARLEN(c, eptr, len); }
5666                  rgb = UCD_GRAPHBREAK(c);
5667                  if ((PRIV(ucp_gbtable)[lgb] & (1 << rgb)) == 0) break;
5668                  lgb = rgb;
5669                  eptr += len;
5670                  }
5671              }              }
5672              CHECK_PARTIAL();
5673            }            }
5674    
5675          /* eptr is now past the end of the maximum run */          /* eptr is now past the end of the maximum run */
5676    
5677          if (possessive) continue;          if (possessive) continue;    /* No backtracking */
   
5678          for(;;)          for(;;)
5679            {            {
5680              if (eptr == pp) goto TAIL_RECURSE;
5681            RMATCH(eptr, ecode, offset_top, md, eptrb, RM45);            RMATCH(eptr, ecode, offset_top, md, eptrb, RM45);
5682            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
5683            if (eptr-- == pp) break;        /* Stop if tried at original pos */            eptr--;
5684            for (;;)                        /* Move back over one extended */            for (;;)                        /* Move back over one extended */
5685              {              {
5686              if (!utf) c = *eptr; else              if (!utf) c = *eptr; else
# Line 5534  for (;;) Line 5713  for (;;)
5713                  break;                  break;
5714                  }                  }
5715                if (IS_NEWLINE(eptr)) break;                if (IS_NEWLINE(eptr)) break;
5716                  if (md->partial != 0 &&    /* Take care with CRLF partial */
5717                      eptr + 1 >= md->end_subject &&
5718                      NLBLOCK->nltype == NLTYPE_FIXED &&
5719                      NLBLOCK->nllen == 2 &&
5720                      RAWUCHAR(eptr) == NLBLOCK->nl[0])
5721                    {
5722                    md->hitend = TRUE;
5723                    if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
5724                    }
5725                eptr++;                eptr++;
5726                ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);                ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);
5727                }                }
# Line 5551  for (;;) Line 5739  for (;;)
5739                  break;                  break;
5740                  }                  }
5741                if (IS_NEWLINE(eptr)) break;                if (IS_NEWLINE(eptr)) break;
5742                  if (md->partial != 0 &&    /* Take care with CRLF partial */
5743                      eptr + 1 >= md->end_subject &&
5744                      NLBLOCK->nltype == NLTYPE_FIXED &&
5745                      NLBLOCK->nllen == 2 &&
5746                      RAWUCHAR(eptr) == NLBLOCK->nl[0])
5747                    {
5748                    md->hitend = TRUE;
5749                    if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
5750                    }
5751                eptr++;                eptr++;
5752                ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);                ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);
5753                }                }
# Line 5600  for (;;) Line 5797  for (;;)
5797                break;                break;
5798                }                }
5799              GETCHARLEN(c, eptr, len);              GETCHARLEN(c, eptr, len);
5800              if (c == 0x000d)              if (c == CHAR_CR)
5801                {                {
5802                if (++eptr >= md->end_subject) break;                if (++eptr >= md->end_subject) break;
5803                if (*eptr == 0x000a) eptr++;                if (RAWUCHAR(eptr) == CHAR_LF) eptr++;
5804                }                }
5805              else              else
5806                {                {
5807                if (c != 0x000a &&                if (c != CHAR_LF &&
5808                    (md->bsr_anycrlf ||                    (md->bsr_anycrlf ||
5809                     (c != 0x000b && c != 0x000c &&                     (c != CHAR_VT && c != CHAR_FF && c != CHAR_NEL
5810                      c != 0x0085 && c != 0x2028 && c != 0x2029)))  #ifndef EBCDIC
5811                        && c != 0x2028 && c != 0x2029
5812    #endif  /* Not EBCDIC */
5813                        )))
5814                  break;                  break;
5815                eptr += len;                eptr += len;
5816                }                }
# Line 5631  for (;;) Line 5831  for (;;)
5831              GETCHARLEN(c, eptr, len);              GETCHARLEN(c, eptr, len);
5832              switch(c)              switch(c)
5833                {                {
5834                  HSPACE_CASES: gotspace = TRUE; break;
5835                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;  
5836                }                }
5837              if (gotspace == (ctype == OP_NOT_HSPACE)) break;              if (gotspace == (ctype == OP_NOT_HSPACE)) break;
5838              eptr += len;              eptr += len;
# Line 5673  for (;;) Line 5853  for (;;)
5853              GETCHARLEN(c, eptr, len);              GETCHARLEN(c, eptr, len);
5854              switch(c)              switch(c)
5855                {                {
5856                  VSPACE_CASES: gotspace = TRUE; break;
5857                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;  
5858                }                }
5859              if (gotspace == (ctype == OP_NOT_VSPACE)) break;              if (gotspace == (ctype == OP_NOT_VSPACE)) break;
5860              eptr += len;              eptr += len;
# Line 5783  for (;;) Line 5955  for (;;)
5955            RRETURN(PCRE_ERROR_INTERNAL);            RRETURN(PCRE_ERROR_INTERNAL);
5956            }            }
5957    
5958          /* 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;  
5959          for(;;)          for(;;)
5960            {            {
5961              if (eptr == pp) goto TAIL_RECURSE;
5962            RMATCH(eptr, ecode, offset_top, md, eptrb, RM46);            RMATCH(eptr, ecode, offset_top, md, eptrb, RM46);
5963            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
5964            if (eptr-- == pp) break;        /* Stop if tried at original pos */            eptr--;
5965            BACKCHAR(eptr);            BACKCHAR(eptr);
5966            if (ctype == OP_ANYNL && eptr > pp  && *eptr == '\n' &&            if (ctype == OP_ANYNL && eptr > pp  && RAWUCHAR(eptr) == CHAR_NL &&
5967                eptr[-1] == '\r') eptr--;                RAWUCHAR(eptr - 1) == CHAR_CR) eptr--;
5968            }            }
5969          }          }
5970        else        else
# Line 5815  for (;;) Line 5982  for (;;)
5982                break;                break;
5983                }                }
5984              if (IS_NEWLINE(eptr)) break;              if (IS_NEWLINE(eptr)) break;
5985                if (md->partial != 0 &&    /* Take care with CRLF partial */
5986                    eptr + 1 >= md->end_subject &&
5987                    NLBLOCK->nltype == NLTYPE_FIXED &&
5988                    NLBLOCK->nllen == 2 &&
5989                    *eptr == NLBLOCK->nl[0])
5990                  {
5991                  md->hitend = TRUE;
5992                  if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
5993                  }
5994              eptr++;              eptr++;
5995              }              }
5996            break;            break;
# Line 5839  for (;;) Line 6015  for (;;)
6015                break;                break;
6016                }                }
6017              c = *eptr;              c = *eptr;
6018              if (c == 0x000d)              if (c == CHAR_CR)
6019                {                {
6020                if (++eptr >= md->end_subject) break;                if (++eptr >= md->end_subject) break;
6021                if (*eptr == 0x000a) eptr++;                if (*eptr == CHAR_LF) eptr++;
6022                }                }
6023              else              else
6024                {                {
6025                if (c != 0x000a && (md->bsr_anycrlf ||                if (c != CHAR_LF && (md->bsr_anycrlf ||
6026                  (c != 0x000b && c != 0x000c && c != 0x0085                   (c != CHAR_VT && c != CHAR_FF && c != CHAR_NEL
6027  #ifdef COMPILE_PCRE16  #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
6028                  && c != 0x2028 && c != 0x2029                   && c != 0x2028 && c != 0x2029
6029  #endif  #endif
6030                  ))) break;                   ))) break;
6031                eptr++;                eptr++;
6032                }                }
6033              }              }
# Line 5865  for (;;) Line 6041  for (;;)
6041                SCHECK_PARTIAL();                SCHECK_PARTIAL();
6042                break;                break;
6043                }                }
6044              c = *eptr;              switch(*eptr)
6045              if (c == 0x09 || c == 0x20 || c == 0xa0                {
6046  #ifdef COMPILE_PCRE16                default: eptr++; break;
6047                || c == 0x1680 || c == 0x180e || (c >= 0x2000 && c <= 0x200A)                HSPACE_BYTE_CASES:
6048                || c == 0x202f || c == 0x205f || c == 0x3000  #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
6049                  HSPACE_MULTIBYTE_CASES:
6050  #endif  #endif
6051                ) break;                goto ENDLOOP00;
6052              eptr++;                }
6053              }              }
6054              ENDLOOP00:
6055            break;            break;
6056    
6057            case OP_HSPACE:            case OP_HSPACE:
# Line 5884  for (;;) Line 6062  for (;;)
6062                SCHECK_PARTIAL();                SCHECK_PARTIAL();
6063                break;                break;
6064                }                }
6065              c = *eptr;              switch(*eptr)
6066              if (c != 0x09 && c != 0x20 && c != 0xa0                {
6067  #ifdef COMPILE_PCRE16                default: goto ENDLOOP01;
6068                && c != 0x1680 && c != 0x180e && (c < 0x2000 || c > 0x200A)                HSPACE_BYTE_CASES:
6069                && c != 0x202f && c != 0x205f && c != 0x3000  #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
6070                  HSPACE_MULTIBYTE_CASES:
6071  #endif  #endif
6072                ) break;                eptr++; break;
6073              eptr++;                }
6074              }              }
6075              ENDLOOP01:
6076            break;            break;
6077    
6078            case OP_NOT_VSPACE:            case OP_NOT_VSPACE:
# Line 5903  for (;;) Line 6083  for (;;)
6083                SCHECK_PARTIAL();                SCHECK_PARTIAL();
6084                break;                break;
6085                }                }
6086              c = *eptr;              switch(*eptr)
6087              if (c == 0x0a || c == 0x0b || c == 0x0c || c == 0x0d || c == 0x85                {
6088  #ifdef COMPILE_PCRE16                default: eptr++; break;
6089                || c == 0x2028 || c == 0x2029                VSPACE_BYTE_CASES:
6090    #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
6091                  VSPACE_MULTIBYTE_CASES:
6092  #endif  #endif
6093                ) break;                goto ENDLOOP02;
6094              eptr++;                }
6095              }              }
6096              ENDLOOP02:
6097            break;            break;
6098    
6099            case OP_VSPACE:            case OP_VSPACE:
# Line 5921  for (;;) Line 6104  for (;;)
6104                SCHECK_PARTIAL();                SCHECK_PARTIAL();
6105                break;                break;
6106                }                }
6107              c = *eptr;              switch(*eptr)
6108              if (c != 0x0a && c != 0x0b && c != 0x0c && c != 0x0d && c != 0x85                {
6109  #ifdef COMPILE_PCRE16                default: goto ENDLOOP03;
6110                && c != 0x2028 && c != 0x2029                VSPACE_BYTE_CASES:
6111    #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
6112                  VSPACE_MULTIBYTE_CASES:
6113  #endif  #endif
6114                ) break;                eptr++; break;
6115              eptr++;                }
6116              }              }
6117              ENDLOOP03:
6118            break;            break;
6119    
6120            case OP_NOT_DIGIT:            case OP_NOT_DIGIT:
# Line 6013  for (;;) Line 6199  for (;;)
6199            RRETURN(PCRE_ERROR_INTERNAL);            RRETURN(PCRE_ERROR_INTERNAL);
6200            }            }
6201    
6202          /* eptr is now past the end of the maximum run. If possessive, we are          if (possessive) continue;    /* No backtracking */
6203          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)  
6204            {            {
6205              if (eptr == pp) goto TAIL_RECURSE;
6206            RMATCH(eptr, ecode, offset_top, md, eptrb, RM47);            RMATCH(eptr, ecode, offset_top, md, eptrb, RM47);
6207            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
6208            eptr--;            eptr--;
6209            if (ctype == OP_ANYNL && eptr > pp  && *eptr == '\n' &&            if (ctype == OP_ANYNL && eptr > pp  && *eptr == CHAR_LF &&
6210                eptr[-1] == '\r') eptr--;                eptr[-1] == CHAR_CR) eptr--;
6211            }            }
6212          }          }
6213    
# Line 6076  switch (frame->Xwhere) Line 6257  switch (frame->Xwhere)
6257    LBL(32) LBL(34) LBL(42) LBL(46)    LBL(32) LBL(34) LBL(42) LBL(46)
6258  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
6259    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)
6260    LBL(59) LBL(60) LBL(61) LBL(62)    LBL(59) LBL(60) LBL(61) LBL(62) LBL(67) LBL(68)
6261  #endif  /* SUPPORT_UCP */  #endif  /* SUPPORT_UCP */
6262  #endif  /* SUPPORT_UTF */  #endif  /* SUPPORT_UTF */
6263    default:    default:
6264    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);  
   
6265    return PCRE_ERROR_INTERNAL;    return PCRE_ERROR_INTERNAL;
6266    }    }
6267  #undef LBL  #undef LBL
# Line 6145  Undefine all the macros that were define Line 6323  Undefine all the macros that were define
6323  ***************************************************************************/  ***************************************************************************/
6324    
6325    
6326    #ifdef NO_RECURSE
6327    /*************************************************
6328    *          Release allocated heap frames         *
6329    *************************************************/
6330    
6331    /* This function releases all the allocated frames. The base frame is on the
6332    machine stack, and so must not be freed.
6333    
6334    Argument: the address of the base frame
6335    Returns:  nothing
6336    */
6337    
6338    static void
6339    release_match_heapframes (heapframe *frame_base)
6340    {
6341    heapframe *nextframe = frame_base->Xnextframe;
6342    while (nextframe != NULL)
6343      {
6344      heapframe *oldframe = nextframe;
6345      nextframe = nextframe->Xnextframe;
6346      (PUBL(stack_free))(oldframe);
6347      }
6348    }
6349    #endif
6350    
6351    
6352  /*************************************************  /*************************************************
6353  *         Execute a Regular Expression           *  *         Execute a Regular Expression           *
# Line 6170  Returns:          > 0 => success; value Line 6373  Returns:          > 0 => success; value
6373                   < -1 => some kind of unexpected problem                   < -1 => some kind of unexpected problem
6374  */  */
6375    
6376  #ifdef COMPILE_PCRE8  #if defined COMPILE_PCRE8
6377  PCRE_EXP_DEFN int PCRE_CALL_CONVENTION  PCRE_EXP_DEFN int PCRE_CALL_CONVENTION
6378  pcre_exec(const pcre *argument_re, const pcre_extra *extra_data,  pcre_exec(const pcre *argument_re, const pcre_extra *extra_data,
6379    PCRE_SPTR subject, int length, int start_offset, int options, int *offsets,    PCRE_SPTR subject, int length, int start_offset, int options, int *offsets,
6380    int offsetcount)    int offsetcount)
6381  #else  #elif defined COMPILE_PCRE16
6382  PCRE_EXP_DEFN int PCRE_CALL_CONVENTION  PCRE_EXP_DEFN int PCRE_CALL_CONVENTION
6383  pcre16_exec(const pcre16 *argument_re, const pcre16_extra *extra_data,  pcre16_exec(const pcre16 *argument_re, const pcre16_extra *extra_data,
6384    PCRE_SPTR16 subject, int length, int start_offset, int options, int *offsets,    PCRE_SPTR16 subject, int length, int start_offset, int options, int *offsets,
6385    int offsetcount)    int offsetcount)
6386    #elif defined COMPILE_PCRE32
6387    PCRE_EXP_DEFN int PCRE_CALL_CONVENTION
6388    pcre32_exec(const pcre32 *argument_re, const pcre32_extra *extra_data,
6389      PCRE_SPTR32 subject, int length, int start_offset, int options, int *offsets,
6390      int offsetcount)
6391  #endif  #endif
6392  {  {
6393  int rc, ocount, arg_offset_max;  int rc, ocount, arg_offset_max;
# Line 6202  const pcre_uint8 *start_bits = NULL; Line 6410  const pcre_uint8 *start_bits = NULL;
6410  PCRE_PUCHAR start_match = (PCRE_PUCHAR)subject + start_offset;  PCRE_PUCHAR start_match = (PCRE_PUCHAR)subject + start_offset;
6411  PCRE_PUCHAR end_subject;  PCRE_PUCHAR end_subject;
6412  PCRE_PUCHAR start_partial = NULL;  PCRE_PUCHAR start_partial = NULL;
6413    PCRE_PUCHAR match_partial;
6414  PCRE_PUCHAR req_char_ptr = start_match - 1;  PCRE_PUCHAR req_char_ptr = start_match - 1;
6415    
6416  const pcre_study_data *study;  const pcre_study_data *study;
6417  const REAL_PCRE *re = (const REAL_PCRE *)argument_re;  const REAL_PCRE *re = (const REAL_PCRE *)argument_re;
6418    
6419  /* Check for the special magic call that measures the size of the stack used  #ifdef NO_RECURSE
6420  per recursive call of match(). */  heapframe frame_zero;
6421    frame_zero.Xprevframe = NULL;            /* Marks the top level */
6422    frame_zero.Xnextframe = NULL;            /* None are allocated yet */
6423    md->match_frames_base = &frame_zero;
6424    #endif
6425    
6426    /* Check for the special magic call that measures the size of the stack used
6427    per recursive call of match(). Without the funny casting for sizeof, a Windows
6428    compiler gave this error: "unary minus operator applied to unsigned type,
6429    result still unsigned". Hopefully the cast fixes that. */
6430    
6431  if (re == NULL && extra_data == NULL && subject == NULL && length == -1)  if (re == NULL && extra_data == NULL && subject == NULL && length == -999 &&
6432        start_offset == -999)
6433  #ifdef NO_RECURSE  #ifdef NO_RECURSE
6434    return -sizeof(heapframe);    return -((int)sizeof(heapframe));
6435  #else  #else
6436    return match(NULL, NULL, NULL, 0, NULL, NULL, 0);    return match(NULL, NULL, NULL, 0, NULL, NULL, 0);
6437  #endif  #endif
6438    
6439  /* Plausibility checks */  /* Plausibility checks */
6440    
6441  if ((options & ~PUBLIC_EXEC_OPTIONS) != 0) return PCRE_ERROR_BADOPTION;  if ((options & ~PUBLIC_EXEC_OPTIONS) != 0) return PCRE_ERROR_BADOPTION;
6442  if (re == NULL || subject == NULL || (offsets == NULL && offsetcount > 0))  if (re == NULL || subject == NULL || (offsets == NULL && offsetcount > 0))
6443    return PCRE_ERROR_NULL;    return PCRE_ERROR_NULL;
6444  if (offsetcount < 0) return PCRE_ERROR_BADCOUNT;  if (offsetcount < 0) return PCRE_ERROR_BADCOUNT;
6445    if (length < 0) return PCRE_ERROR_BADLENGTH;
6446  if (start_offset < 0 || start_offset > length) return PCRE_ERROR_BADOFFSET;  if (start_offset < 0 || start_offset > length) return PCRE_ERROR_BADOFFSET;
6447    
6448  /* 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 6480  if (utf && (options & PCRE_NO_UTF8_CHECK
6480        offsets[0] = erroroffset;        offsets[0] = erroroffset;
6481        offsets[1] = errorcode;        offsets[1] = errorcode;
6482        }        }
6483  #ifdef COMPILE_PCRE16  #if defined COMPILE_PCRE8
     return (errorcode <= PCRE_UTF16_ERR1 && md->partial > 1)?  
       PCRE_ERROR_SHORTUTF16 : PCRE_ERROR_BADUTF16;  
 #else  
6484      return (errorcode <= PCRE_UTF8_ERR5 && md->partial > 1)?      return (errorcode <= PCRE_UTF8_ERR5 && md->partial > 1)?
6485        PCRE_ERROR_SHORTUTF8 : PCRE_ERROR_BADUTF8;        PCRE_ERROR_SHORTUTF8 : PCRE_ERROR_BADUTF8;
6486    #elif defined COMPILE_PCRE16
6487        return (errorcode <= PCRE_UTF16_ERR1 && md->partial > 1)?
6488          PCRE_ERROR_SHORTUTF16 : PCRE_ERROR_BADUTF16;
6489    #elif defined COMPILE_PCRE32
6490        return PCRE_ERROR_BADUTF32;
6491  #endif  #endif
6492      }      }
6493    #if defined COMPILE_PCRE8 || defined COMPILE_PCRE16
6494    /* 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. */
6495    if (start_offset > 0 && start_offset < length &&    if (start_offset > 0 && start_offset < length &&
6496        NOT_FIRSTCHAR(((PCRE_PUCHAR)subject)[start_offset]))        NOT_FIRSTCHAR(((PCRE_PUCHAR)subject)[start_offset]))
6497      return PCRE_ERROR_BADUTF8_OFFSET;      return PCRE_ERROR_BADUTF8_OFFSET;
6498    #endif
6499    }    }
6500  #endif  #endif
6501    
6502  /* If the pattern was successfully studied with JIT support, run the JIT  /* If the pattern was successfully studied with JIT support, run the JIT
6503  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
6504  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
6505  an unsupported flag is set. In particular, JIT does not support partial  an unsupported flag is set. */
 matching. */  
6506    
6507  #ifdef SUPPORT_JIT  #ifdef SUPPORT_JIT
6508  if (extra_data != NULL  if (extra_data != NULL
6509      && (extra_data->flags & PCRE_EXTRA_EXECUTABLE_JIT) != 0      && (extra_data->flags & (PCRE_EXTRA_EXECUTABLE_JIT |
6510                                 PCRE_EXTRA_TABLES)) == PCRE_EXTRA_EXECUTABLE_JIT
6511      && extra_data->executable_jit != NULL      && extra_data->executable_jit != NULL
6512      && (extra_data->flags & PCRE_EXTRA_TABLES) == 0      && (options & ~PUBLIC_JIT_EXEC_OPTIONS) == 0)
6513      && (options & ~(PCRE_NO_UTF8_CHECK | PCRE_NOTBOL | PCRE_NOTEOL |    {
6514                      PCRE_NOTEMPTY | PCRE_NOTEMPTY_ATSTART)) == 0)    rc = PRIV(jit_exec)(extra_data, (const pcre_uchar *)subject, length,
6515    return PRIV(jit_exec)(re, extra_data->executable_jit,         start_offset, options, offsets, offsetcount);
6516      (const pcre_uchar *)subject, length, start_offset, options,  
6517      ((extra_data->flags & PCRE_EXTRA_MATCH_LIMIT) == 0)    /* PCRE_ERROR_NULL means that the selected normal or partial matching
6518      ? MATCH_LIMIT : extra_data->match_limit, offsets, offsetcount);    mode is not compiled. In this case we simply fallback to interpreter. */
6519    
6520      if (rc != PCRE_ERROR_JIT_BADOPTION) return rc;
6521      }
6522  #endif  #endif
6523    
6524  /* 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 6540  md->callout_data = NULL;
6540    
6541  tables = re->tables;  tables = re->tables;
6542    
6543    /* The two limit values override the defaults, whatever their value. */
6544    
6545  if (extra_data != NULL)  if (extra_data != NULL)
6546    {    {
6547    register unsigned int flags = extra_data->flags;    register unsigned int flags = extra_data->flags;
# Line 6328  if (extra_data != NULL) Line 6556  if (extra_data != NULL)
6556    if ((flags & PCRE_EXTRA_TABLES) != 0) tables = extra_data->tables;    if ((flags & PCRE_EXTRA_TABLES) != 0) tables = extra_data->tables;
6557    }    }
6558    
6559    /* Limits in the regex override only if they are smaller. */
6560    
6561    if ((re->flags & PCRE_MLSET) != 0 && re->limit_match < md->match_limit)
6562      md->match_limit = re->limit_match;
6563    
6564    if ((re->flags & PCRE_RLSET) != 0 &&
6565        re->limit_recursion < md->match_limit_recursion)
6566      md->match_limit_recursion = re->limit_recursion;
6567    
6568  /* 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
6569  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
6570  in other programs later. */  in other programs later. */
# Line 6353  end_subject = md->end_subject; Line 6590  end_subject = md->end_subject;
6590  md->endonly = (re->options & PCRE_DOLLAR_ENDONLY) != 0;  md->endonly = (re->options & PCRE_DOLLAR_ENDONLY) != 0;
6591  md->use_ucp = (re->options & PCRE_UCP) != 0;  md->use_ucp = (re->options & PCRE_UCP) != 0;
6592  md->jscript_compat = (re->options & PCRE_JAVASCRIPT_COMPAT) != 0;  md->jscript_compat = (re->options & PCRE_JAVASCRIPT_COMPAT) != 0;
6593  md->ignore_skip_arg = FALSE;  md->ignore_skip_arg = 0;
6594    
6595  /* Some options are unpacked into BOOL variables in the hope that testing  /* Some options are unpacked into BOOL variables in the hope that testing
6596  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 6700  if (re->top_backref > 0 && re->top_backr
6700    DPRINTF(("Got memory to hold back references\n"));    DPRINTF(("Got memory to hold back references\n"));
6701    }    }
6702  else md->offset_vector = offsets;  else md->offset_vector = offsets;
   
6703  md->offset_end = ocount;  md->offset_end = ocount;
6704  md->offset_max = (2*ocount)/3;  md->offset_max = (2*ocount)/3;
6705  md->offset_overflow = FALSE;  md->capture_last = 0;
 md->capture_last = -1;  
6706    
6707  /* Reset the working variable associated with each extraction. These should  /* Reset the working variable associated with each extraction. These should
6708  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 6730  if (!anchored)
6730    if ((re->flags & PCRE_FIRSTSET) != 0)    if ((re->flags & PCRE_FIRSTSET) != 0)
6731      {      {
6732      has_first_char = TRUE;      has_first_char = TRUE;
6733      first_char = first_char2 = re->first_char;      first_char = first_char2 = (pcre_uchar)(re->first_char);
6734      if ((re->flags & PCRE_FCH_CASELESS) != 0)      if ((re->flags & PCRE_FCH_CASELESS) != 0)
6735        {        {
6736        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 6752  character" set. */
6752  if ((re->flags & PCRE_REQCHSET) != 0)  if ((re->flags & PCRE_REQCHSET) != 0)
6753    {    {
6754    has_req_char = TRUE;    has_req_char = TRUE;
6755    req_char = req_char2 = re->req_char;    req_char = req_char2 = (pcre_uchar)(re->req_char);
6756    if ((re->flags & PCRE_RCH_CASELESS) != 0)    if ((re->flags & PCRE_RCH_CASELESS) != 0)
6757      {      {
6758      req_char2 = TABLE_GET(req_char, md->fcc, req_char);      req_char2 = TABLE_GET(req_char, md->fcc, req_char);
# Line 6575  for(;;) Line 6810  for(;;)
6810    
6811      if (has_first_char)      if (has_first_char)
6812        {        {
6813          pcre_uchar smc;
6814    
6815        if (first_char != first_char2)        if (first_char != first_char2)
6816          while (start_match < end_subject &&          while (start_match < end_subject &&
6817              *start_match != first_char && *start_match != first_char2)            (smc = RAWUCHARTEST(start_match)) != first_char && smc != first_char2)
6818            start_match++;            start_match++;
6819        else        else
6820          while (start_match < end_subject && *start_match != first_char)          while (start_match < end_subject && RAWUCHARTEST(start_match) != first_char)
6821            start_match++;            start_match++;
6822        }        }
6823    
# Line 6612  for(;;) Line 6849  for(;;)
6849          if (start_match[-1] == CHAR_CR &&          if (start_match[-1] == CHAR_CR &&
6850               (md->nltype == NLTYPE_ANY || md->nltype == NLTYPE_ANYCRLF) &&               (md->nltype == NLTYPE_ANY || md->nltype == NLTYPE_ANYCRLF) &&
6851               start_match < end_subject &&               start_match < end_subject &&
6852               *start_match == CHAR_NL)               RAWUCHARTEST(start_match) == CHAR_NL)
6853            start_match++;            start_match++;
6854          }          }
6855        }        }
# Line 6623  for(;;) Line 6860  for(;;)
6860        {        {
6861        while (start_match < end_subject)        while (start_match < end_subject)
6862          {          {
6863          register unsigned int c = *start_match;          register pcre_uint32 c = RAWUCHARTEST(start_match);
6864  #ifndef COMPILE_PCRE8  #ifndef COMPILE_PCRE8
6865          if (c > 255) c = 255;          if (c > 255) c = 255;
6866  #endif  #endif
# Line 6691  for(;;) Line 6928  for(;;)
6928            {            {
6929            while (p < end_subject)            while (p < end_subject)
6930              {              {
6931              register int pp = *p++;              register pcre_uint32 pp = RAWUCHARINCTEST(p);
6932              if (pp == req_char || pp == req_char2) { p--; break; }              if (pp == req_char || pp == req_char2) { p--; break; }
6933              }              }
6934            }            }
# Line 6699  for(;;) Line 6936  for(;;)
6936            {            {
6937            while (p < end_subject)            while (p < end_subject)
6938              {              {
6939              if (*p++ == req_char) { p--; break; }              if (RAWUCHARINCTEST(p) == req_char) { p--; break; }
6940              }              }
6941            }            }
6942    
# Line 6735  for(;;) Line 6972  for(;;)
6972    md->match_call_count = 0;    md->match_call_count = 0;
6973    md->match_function_type = 0;    md->match_function_type = 0;
6974    md->end_offset_top = 0;    md->end_offset_top = 0;
6975      md->skip_arg_count = 0;
6976    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);
6977    if (md->hitend && start_partial == NULL) start_partial = md->start_used_ptr;    if (md->hitend && start_partial == NULL)
6978        {
6979        start_partial = md->start_used_ptr;
6980        match_partial = start_match;
6981        }
6982    
6983    switch(rc)    switch(rc)
6984      {      {
# Line 6749  for(;;) Line 6991  for(;;)
6991    
6992      case MATCH_SKIP_ARG:      case MATCH_SKIP_ARG:
6993      new_start_match = start_match;      new_start_match = start_match;
6994      md->ignore_skip_arg = TRUE;      md->ignore_skip_arg = md->skip_arg_count;
6995      break;      break;
6996    
6997      /* 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
6998      same as the match we have just done, treat it as NOMATCH. */      greater than the match we have just done, treat it as NOMATCH. */
6999    
7000      case MATCH_SKIP:      case MATCH_SKIP:
7001      if (md->start_match_ptr != start_match)      if (md->start_match_ptr > start_match)
7002        {        {
7003        new_start_match = md->start_match_ptr;        new_start_match = md->start_match_ptr;
7004        break;        break;
# Line 6764  for(;;) Line 7006  for(;;)
7006      /* Fall through */      /* Fall through */
7007    
7008      /* NOMATCH and PRUNE advance by one character. THEN at this level acts      /* NOMATCH and PRUNE advance by one character. THEN at this level acts
7009      exactly like PRUNE. Unset the ignore SKIP-with-argument flag. */      exactly like PRUNE. Unset ignore SKIP-with-argument. */
7010    
7011      case MATCH_NOMATCH:      case MATCH_NOMATCH:
7012      case MATCH_PRUNE:      case MATCH_PRUNE:
7013      case MATCH_THEN:      case MATCH_THEN:
7014      md->ignore_skip_arg = FALSE;      md->ignore_skip_arg = 0;
7015      new_start_match = start_match + 1;      new_start_match = start_match + 1;
7016  #ifdef SUPPORT_UTF  #ifdef SUPPORT_UTF
7017      if (utf)      if (utf)
# Line 6862  if (rc == MATCH_MATCH || rc == MATCH_ACC Line 7104  if (rc == MATCH_MATCH || rc == MATCH_ACC
7104          (arg_offset_max - 2) * sizeof(int));          (arg_offset_max - 2) * sizeof(int));
7105        DPRINTF(("Copied offsets from temporary memory\n"));        DPRINTF(("Copied offsets from temporary memory\n"));
7106        }        }
7107      if (md->end_offset_top > arg_offset_max) md->offset_overflow = TRUE;      if (md->end_offset_top > arg_offset_max) md->capture_last |= OVFLBIT;
7108      DPRINTF(("Freeing temporary memory\n"));      DPRINTF(("Freeing temporary memory\n"));
7109      (PUBL(free))(md->offset_vector);      (PUBL(free))(md->offset_vector);
7110      }      }
# Line 6870  if (rc == MATCH_MATCH || rc == MATCH_ACC Line 7112  if (rc == MATCH_MATCH || rc == MATCH_ACC
7112    /* Set the return code to the number of captured strings, or 0 if there were    /* Set the return code to the number of captured strings, or 0 if there were
7113    too many to fit into the vector. */    too many to fit into the vector. */
7114    
7115    rc = (md->offset_overflow && md->end_offset_top >= arg_offset_max)?    rc = ((md->capture_last & OVFLBIT) != 0 &&
7116             md->end_offset_top >= arg_offset_max)?
7117      0 : md->end_offset_top/2;      0 : md->end_offset_top/2;
7118    
7119    /* If there is space in the offset vector, set any unused pairs at the end of    /* If there is space in the offset vector, set any unused pairs at the end of
# Line 6886  if (rc == MATCH_MATCH || rc == MATCH_ACC Line 7129  if (rc == MATCH_MATCH || rc == MATCH_ACC
7129      {      {
7130      register int *iptr, *iend;      register int *iptr, *iend;
7131      int resetcount = 2 + re->top_bracket * 2;      int resetcount = 2 + re->top_bracket * 2;
7132      if (resetcount > offsetcount) resetcount = ocount;      if (resetcount > offsetcount) resetcount = offsetcount;
7133      iptr = offsets + md->end_offset_top;      iptr = offsets + md->end_offset_top;
7134      iend = offsets + resetcount;      iend = offsets + resetcount;
7135      while (iptr < iend) *iptr++ = -1;      while (iptr < iend) *iptr++ = -1;
# Line 6907  if (rc == MATCH_MATCH || rc == MATCH_ACC Line 7150  if (rc == MATCH_MATCH || rc == MATCH_ACC
7150    if (extra_data != NULL && (extra_data->flags & PCRE_EXTRA_MARK) != 0)    if (extra_data != NULL && (extra_data->flags & PCRE_EXTRA_MARK) != 0)
7151      *(extra_data->mark) = (pcre_uchar *)md->mark;      *(extra_data->mark) = (pcre_uchar *)md->mark;
7152    DPRINTF((">>>> returning %d\n", rc));    DPRINTF((">>>> returning %d\n", rc));
7153    #ifdef NO_RECURSE
7154      release_match_heapframes(&frame_zero);
7155    #endif
7156    return rc;    return rc;
7157    }    }
7158    
# Line 6924  if (using_temporary_offsets) Line 7170  if (using_temporary_offsets)
7170  if (rc != MATCH_NOMATCH && rc != PCRE_ERROR_PARTIAL)  if (rc != MATCH_NOMATCH && rc != PCRE_ERROR_PARTIAL)
7171    {    {
7172    DPRINTF((">>>> error: returning %d\n", rc));    DPRINTF((">>>> error: returning %d\n", rc));
7173    #ifdef NO_RECURSE
7174      release_match_heapframes(&frame_zero);
7175    #endif
7176    return rc;    return rc;
7177    }    }
7178    
# Line 6937  if (start_partial != NULL) Line 7186  if (start_partial != NULL)
7186      {      {
7187      offsets[0] = (int)(start_partial - (PCRE_PUCHAR)subject);      offsets[0] = (int)(start_partial - (PCRE_PUCHAR)subject);
7188      offsets[1] = (int)(end_subject - (PCRE_PUCHAR)subject);      offsets[1] = (int)(end_subject - (PCRE_PUCHAR)subject);
7189        if (offsetcount > 2)
7190          offsets[2] = (int)(match_partial - (PCRE_PUCHAR)subject);
7191      }      }
7192    rc = PCRE_ERROR_PARTIAL;    rc = PCRE_ERROR_PARTIAL;
7193    }    }
# Line 6953  else Line 7204  else
7204    
7205  if (extra_data != NULL && (extra_data->flags & PCRE_EXTRA_MARK) != 0)  if (extra_data != NULL && (extra_data->flags & PCRE_EXTRA_MARK) != 0)
7206    *(extra_data->mark) = (pcre_uchar *)md->nomatch_mark;    *(extra_data->mark) = (pcre_uchar *)md->nomatch_mark;
7207    #ifdef NO_RECURSE
7208      release_match_heapframes(&frame_zero);
7209    #endif
7210  return rc;  return rc;
7211  }  }
7212    

Legend:
Removed from v.895  
changed lines
  Added in v.1316

  ViewVC Help
Powered by ViewVC 1.1.5