/[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 893 by ph10, Thu Jan 19 17:15:11 2012 UTC revision 1284 by ph10, Fri Mar 15 11:54:58 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 first argument (eptr) points to a pointer that is used  recursion. The second argument (ecode) being NULL triggers this behaviour. It
665  "statically" for doing the calculation. The second argument (ecode) being NULL  cannot normally ever be NULL. The return is the negated value of the frame
666  triggers this behaviour. It cannot normally every be NULL. The return is the  size. */
 negated value of the frame size. */  
667    
668  if (ecode == NULL)  if (ecode == NULL)
669    {    {
   char **aptr = (char **)eptr;  
670    if (rdepth == 0)    if (rdepth == 0)
671      {      return match((PCRE_PUCHAR)&rdepth, NULL, NULL, 0, NULL, NULL, 1);
     *aptr = (char *)&rdepth;  
     return match(eptr, NULL, NULL, 0, NULL, NULL, 1);  
     }  
672    else    else
673      {      {
674      int len = (char *)&rdepth - *aptr;      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 733  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 743  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 769  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 843  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 862  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 902  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 942  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 1031  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 1046  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 1075  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 1198  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 1227  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 1258  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 1269  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 1291  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 1363  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 1491  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 1499  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 1554  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 1570  for (;;) Line 1611  for (;;)
1611          mstart = md->start_match_ptr;   /* In case \K reset it */          mstart = md->start_match_ptr;   /* In case \K reset it */
1612          break;          break;
1613          }          }
1614          md->mark = save_mark;
1615    
1616        /* PCRE does not allow THEN to escape beyond an assertion; it is treated        /* A COMMIT failure must fail the entire assertion, without trying any
1617        as NOMATCH. */        subsequent branches. */
1618    
1619          if (rrc == MATCH_COMMIT) RRETURN(MATCH_NOMATCH);
1620    
1621          /* PCRE does not allow THEN to escape beyond an assertion; it
1622          is treated as NOMATCH. */
1623    
1624        if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN) RRETURN(rrc);        if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN) RRETURN(rrc);
1625        ecode += GET(ecode, 1);        ecode += GET(ecode, 1);
       md->mark = save_mark;  
1626        }        }
1627      while (*ecode == OP_ALT);      while (*ecode == OP_ALT);
1628    
# Line 1600  for (;;) Line 1646  for (;;)
1646    
1647      case OP_ASSERT_NOT:      case OP_ASSERT_NOT:
1648      case OP_ASSERTBACK_NOT:      case OP_ASSERTBACK_NOT:
1649      save_mark = md->mark;      save_mark = md->mark;
1650      if (md->match_function_type == MATCH_CONDASSERT)      if (md->match_function_type == MATCH_CONDASSERT)
1651        {        {
1652        condassert = TRUE;        condassert = TRUE;
# Line 1611  for (;;) Line 1657  for (;;)
1657      do      do
1658        {        {
1659        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, NULL, RM5);        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, NULL, RM5);
1660        md->mark = save_mark;        md->mark = save_mark;
1661        if (rrc == MATCH_MATCH || rrc == MATCH_ACCEPT) RRETURN(MATCH_NOMATCH);        if (rrc == MATCH_MATCH || rrc == MATCH_ACCEPT) RRETURN(MATCH_NOMATCH);
1662        if (rrc == MATCH_SKIP || rrc == MATCH_PRUNE || rrc == MATCH_COMMIT)        if (rrc == MATCH_SKIP || rrc == MATCH_PRUNE || rrc == MATCH_COMMIT)
1663          {          {
# Line 1676  for (;;) Line 1722  for (;;)
1722        cb.version          = 2;   /* Version 1 of the callout block */        cb.version          = 2;   /* Version 1 of the callout block */
1723        cb.callout_number   = ecode[1];        cb.callout_number   = ecode[1];
1724        cb.offset_vector    = md->offset_vector;        cb.offset_vector    = md->offset_vector;
1725  #ifdef COMPILE_PCRE8  #if defined COMPILE_PCRE8
1726        cb.subject          = (PCRE_SPTR)md->start_subject;        cb.subject          = (PCRE_SPTR)md->start_subject;
1727  #else  #elif defined COMPILE_PCRE16
1728        cb.subject          = (PCRE_SPTR16)md->start_subject;        cb.subject          = (PCRE_SPTR16)md->start_subject;
1729    #elif defined COMPILE_PCRE32
1730          cb.subject          = (PCRE_SPTR32)md->start_subject;
1731  #endif  #endif
1732        cb.subject_length   = (int)(md->end_subject - md->start_subject);        cb.subject_length   = (int)(md->end_subject - md->start_subject);
1733        cb.start_match      = (int)(mstart - md->start_subject);        cb.start_match      = (int)(mstart - md->start_subject);
# Line 1687  for (;;) Line 1735  for (;;)
1735        cb.pattern_position = GET(ecode, 2);        cb.pattern_position = GET(ecode, 2);
1736        cb.next_item_length = GET(ecode, 2 + LINK_SIZE);        cb.next_item_length = GET(ecode, 2 + LINK_SIZE);
1737        cb.capture_top      = offset_top/2;        cb.capture_top      = offset_top/2;
1738        cb.capture_last     = md->capture_last;        cb.capture_last     = md->capture_last & CAPLMASK;
1739          /* Internal change requires this for API compatibility. */
1740          if (cb.capture_last == 0) cb.capture_last = -1;
1741        cb.callout_data     = md->callout_data;        cb.callout_data     = md->callout_data;
1742        cb.mark             = md->nomatch_mark;        cb.mark             = md->nomatch_mark;
1743        if ((rrc = (*PUBL(callout))(&cb)) > 0) RRETURN(MATCH_NOMATCH);        if ((rrc = (*PUBL(callout))(&cb)) > 0) RRETURN(MATCH_NOMATCH);
# Line 1716  for (;;) Line 1766  for (;;)
1766      case OP_RECURSE:      case OP_RECURSE:
1767        {        {
1768        recursion_info *ri;        recursion_info *ri;
1769        int recno;        unsigned int recno;
1770    
1771        callpat = md->start_code + GET(ecode, 1);        callpat = md->start_code + GET(ecode, 1);
1772        recno = (callpat == md->start_code)? 0 :        recno = (callpat == md->start_code)? 0 :
# Line 1733  for (;;) Line 1783  for (;;)
1783        /* Add to "recursing stack" */        /* Add to "recursing stack" */
1784    
1785        new_recursive.group_num = recno;        new_recursive.group_num = recno;
1786          new_recursive.saved_capture_last = md->capture_last;
1787        new_recursive.subject_position = eptr;        new_recursive.subject_position = eptr;
1788        new_recursive.prevrec = md->recursive;        new_recursive.prevrec = md->recursive;
1789        md->recursive = &new_recursive;        md->recursive = &new_recursive;
# Line 1756  for (;;) Line 1807  for (;;)
1807              new_recursive.saved_max * sizeof(int));              new_recursive.saved_max * sizeof(int));
1808    
1809        /* OK, now we can do the recursion. After processing each alternative,        /* OK, now we can do the recursion. After processing each alternative,
1810        restore the offset data. If there were nested recursions, md->recursive        restore the offset data and the last captured value. If there were nested
1811        might be changed, so reset it before looping. */        recursions, md->recursive might be changed, so reset it before looping.
1812          */
1813    
1814        DPRINTF(("Recursing into group %d\n", new_recursive.group_num));        DPRINTF(("Recursing into group %d\n", new_recursive.group_num));
1815        cbegroup = (*callpat >= OP_SBRA);        cbegroup = (*callpat >= OP_SBRA);
# Line 1768  for (;;) Line 1820  for (;;)
1820            md, eptrb, RM6);            md, eptrb, RM6);
1821          memcpy(md->offset_vector, new_recursive.offset_save,          memcpy(md->offset_vector, new_recursive.offset_save,
1822              new_recursive.saved_max * sizeof(int));              new_recursive.saved_max * sizeof(int));
1823            md->capture_last = new_recursive.saved_capture_last;
1824          md->recursive = new_recursive.prevrec;          md->recursive = new_recursive.prevrec;
1825          if (rrc == MATCH_MATCH || rrc == MATCH_ACCEPT)          if (rrc == MATCH_MATCH || rrc == MATCH_ACCEPT)
1826            {            {
# Line 1784  for (;;) Line 1837  for (;;)
1837            goto RECURSION_MATCHED;        /* Exit loop; end processing */            goto RECURSION_MATCHED;        /* Exit loop; end processing */
1838            }            }
1839    
1840          /* 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
1841          as NOMATCH. */          recursion; they are treated as NOMATCH. These codes are defined in a
1842            range that can be tested for. Any other return code is an error. */
1843    
1844          else if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN)          else if (rrc != MATCH_NOMATCH &&
1845                     (rrc < MATCH_BACKTRACK_MIN || rrc > MATCH_BACKTRACK_MAX))
1846            {            {
1847            DPRINTF(("Recursion gave error %d\n", rrc));            DPRINTF(("Recursion gave error %d\n", rrc));
1848            if (new_recursive.offset_save != stacksave)            if (new_recursive.offset_save != stacksave)
# Line 1917  for (;;) Line 1972  for (;;)
1972    
1973        /* Deal with capturing */        /* Deal with capturing */
1974    
1975        md->capture_last = number;        md->capture_last = (md->capture_last & OVFLMASK) | number;
1976        if (offset >= md->offset_max) md->offset_overflow = TRUE; else        if (offset >= md->offset_max) md->capture_last |= OVFLBIT; else
1977          {          {
1978          /* If offset is greater than offset_top, it means that we are          /* If offset is greater than offset_top, it means that we are
1979          "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 1998  for (;;) Line 2053  for (;;)
2053          }          }
2054        if (*prev >= OP_SBRA)    /* Could match an empty string */        if (*prev >= OP_SBRA)    /* Could match an empty string */
2055          {          {
         md->match_function_type = MATCH_CBEGROUP;  
2056          RMATCH(eptr, prev, offset_top, md, eptrb, RM50);          RMATCH(eptr, prev, offset_top, md, eptrb, RM50);
2057          RRETURN(rrc);          RRETURN(rrc);
2058          }          }
# Line 2007  for (;;) Line 2061  for (;;)
2061        }        }
2062      else  /* OP_KETRMAX */      else  /* OP_KETRMAX */
2063        {        {
       if (*prev >= OP_SBRA) md->match_function_type = MATCH_CBEGROUP;  
2064        RMATCH(eptr, prev, offset_top, md, eptrb, RM13);        RMATCH(eptr, prev, offset_top, md, eptrb, RM13);
2065        if (rrc == MATCH_ONCE && md->once_target == prev) rrc = MATCH_NOMATCH;        if (rrc == MATCH_ONCE && md->once_target == prev) rrc = MATCH_NOMATCH;
2066        if (rrc != MATCH_NOMATCH) RRETURN(rrc);        if (rrc != MATCH_NOMATCH) RRETURN(rrc);
# Line 2064  for (;;) Line 2117  for (;;)
2117    
2118      case OP_DOLLM:      case OP_DOLLM:
2119      if (eptr < md->end_subject)      if (eptr < md->end_subject)
2120        { if (!IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH); }        {
2121          if (!IS_NEWLINE(eptr))
2122            {
2123            if (md->partial != 0 &&
2124                eptr + 1 >= md->end_subject &&
2125                NLBLOCK->nltype == NLTYPE_FIXED &&
2126                NLBLOCK->nllen == 2 &&
2127                RAWUCHARTEST(eptr) == NLBLOCK->nl[0])
2128              {
2129              md->hitend = TRUE;
2130              if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
2131              }
2132            RRETURN(MATCH_NOMATCH);
2133            }
2134          }
2135      else      else
2136        {        {
2137        if (md->noteol) RRETURN(MATCH_NOMATCH);        if (md->noteol) RRETURN(MATCH_NOMATCH);
# Line 2096  for (;;) Line 2163  for (;;)
2163      ASSERT_NL_OR_EOS:      ASSERT_NL_OR_EOS:
2164      if (eptr < md->end_subject &&      if (eptr < md->end_subject &&
2165          (!IS_NEWLINE(eptr) || eptr != md->end_subject - md->nllen))          (!IS_NEWLINE(eptr) || eptr != md->end_subject - md->nllen))
2166          {
2167          if (md->partial != 0 &&
2168              eptr + 1 >= md->end_subject &&
2169              NLBLOCK->nltype == NLTYPE_FIXED &&
2170              NLBLOCK->nllen == 2 &&
2171              RAWUCHARTEST(eptr) == NLBLOCK->nl[0])
2172            {
2173            md->hitend = TRUE;
2174            if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
2175            }
2176        RRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2177          }
2178    
2179      /* Either at end of string or \n before end. */      /* Either at end of string or \n before end. */
2180    
# Line 2224  for (;;) Line 2302  for (;;)
2302        }        }
2303      break;      break;
2304    
2305      /* Match a single character type; inline for speed */      /* Match any single character type except newline; have to take care with
2306        CRLF newlines and partial matching. */
2307    
2308      case OP_ANY:      case OP_ANY:
2309      if (IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH);      if (IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH);
2310        if (md->partial != 0 &&
2311            eptr + 1 >= md->end_subject &&
2312            NLBLOCK->nltype == NLTYPE_FIXED &&
2313            NLBLOCK->nllen == 2 &&
2314            RAWUCHARTEST(eptr) == NLBLOCK->nl[0])
2315          {
2316          md->hitend = TRUE;
2317          if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
2318          }
2319    
2320      /* Fall through */      /* Fall through */
2321    
2322        /* Match any single character whatsoever. */
2323    
2324      case OP_ALLANY:      case OP_ALLANY:
2325      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 */
2326        {                            /* not be updated before SCHECK_PARTIAL. */        {                            /* not be updated before SCHECK_PARTIAL. */
# Line 2369  for (;;) Line 2460  for (;;)
2460        {        {
2461        default: RRETURN(MATCH_NOMATCH);        default: RRETURN(MATCH_NOMATCH);
2462    
2463        case 0x000d:        case CHAR_CR:
2464        if (eptr < md->end_subject && *eptr == 0x0a) eptr++;        if (eptr >= md->end_subject)
2465            {
2466            SCHECK_PARTIAL();
2467            }
2468          else if (RAWUCHARTEST(eptr) == CHAR_LF) eptr++;
2469        break;        break;
2470    
2471        case 0x000a:        case CHAR_LF:
2472        break;        break;
2473    
2474        case 0x000b:        case CHAR_VT:
2475        case 0x000c:        case CHAR_FF:
2476        case 0x0085:        case CHAR_NEL:
2477    #ifndef EBCDIC
2478        case 0x2028:        case 0x2028:
2479        case 0x2029:        case 0x2029:
2480    #endif  /* Not EBCDIC */
2481        if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);        if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);
2482        break;        break;
2483        }        }
# Line 2396  for (;;) Line 2493  for (;;)
2493      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2494      switch(c)      switch(c)
2495        {        {
2496          HSPACE_CASES: RRETURN(MATCH_NOMATCH);  /* Byte and multibyte cases */
2497        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);  
2498        }        }
2499      ecode++;      ecode++;
2500      break;      break;
# Line 2430  for (;;) Line 2508  for (;;)
2508      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2509      switch(c)      switch(c)
2510        {        {
2511          HSPACE_CASES: break;  /* Byte and multibyte cases */
2512        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;  
2513        }        }
2514      ecode++;      ecode++;
2515      break;      break;
# Line 2464  for (;;) Line 2523  for (;;)
2523      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2524      switch(c)      switch(c)
2525        {        {
2526          VSPACE_CASES: RRETURN(MATCH_NOMATCH);
2527        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);  
2528        }        }
2529      ecode++;      ecode++;
2530      break;      break;
# Line 2486  for (;;) Line 2538  for (;;)
2538      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2539      switch(c)      switch(c)
2540        {        {
2541          VSPACE_CASES: break;
2542        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;  
2543        }        }
2544      ecode++;      ecode++;
2545      break;      break;
# Line 2512  for (;;) Line 2557  for (;;)
2557        }        }
2558      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2559        {        {
2560          const pcre_uint32 *cp;
2561        const ucd_record *prop = GET_UCD(c);        const ucd_record *prop = GET_UCD(c);
2562    
2563        switch(ecode[1])        switch(ecode[1])
# Line 2572  for (;;) Line 2618  for (;;)
2618            RRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
2619          break;          break;
2620    
2621            case PT_CLIST:
2622            cp = PRIV(ucd_caseless_sets) + ecode[2];
2623            for (;;)
2624              {
2625              if (c < *cp)
2626                { if (op == OP_PROP) { RRETURN(MATCH_NOMATCH); } else break; }
2627              if (c == *cp++)
2628                { if (op == OP_PROP) break; else { RRETURN(MATCH_NOMATCH); } }
2629              }
2630            break;
2631    
2632            case PT_UCNC:
2633            if ((c == CHAR_DOLLAR_SIGN || c == CHAR_COMMERCIAL_AT ||
2634                 c == CHAR_GRAVE_ACCENT || (c >= 0xa0 && c <= 0xd7ff) ||
2635                 c >= 0xe000) == (op == OP_NOTPROP))
2636              RRETURN(MATCH_NOMATCH);
2637            break;
2638    
2639          /* This should never occur */          /* This should never occur */
2640    
2641          default:          default:
# Line 2591  for (;;) Line 2655  for (;;)
2655        SCHECK_PARTIAL();        SCHECK_PARTIAL();
2656        RRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2657        }        }
2658      GETCHARINCTEST(c, eptr);      else
     if (UCD_CATEGORY(c) == ucp_M) RRETURN(MATCH_NOMATCH);  
     while (eptr < md->end_subject)  
2659        {        {
2660        int len = 1;        int lgb, rgb;
2661        if (!utf) c = *eptr; else { GETCHARLEN(c, eptr, len); }        GETCHARINCTEST(c, eptr);
2662        if (UCD_CATEGORY(c) != ucp_M) break;        lgb = UCD_GRAPHBREAK(c);
2663        eptr += len;        while (eptr < md->end_subject)
2664            {
2665            int len = 1;
2666            if (!utf) c = *eptr; else { GETCHARLEN(c, eptr, len); }
2667            rgb = UCD_GRAPHBREAK(c);
2668            if ((PRIV(ucp_gbtable)[lgb] & (1 << rgb)) == 0) break;
2669            lgb = rgb;
2670            eptr += len;
2671            }
2672        }        }
2673        CHECK_PARTIAL();
2674      ecode++;      ecode++;
2675      break;      break;
2676  #endif  #endif  /* SUPPORT_UCP */
2677    
2678    
2679      /* 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 2665  for (;;) Line 2736  for (;;)
2736        default:               /* No repeat follows */        default:               /* No repeat follows */
2737        if ((length = match_ref(offset, eptr, length, md, caseless)) < 0)        if ((length = match_ref(offset, eptr, length, md, caseless)) < 0)
2738          {          {
2739            if (length == -2) eptr = md->end_subject;   /* Partial match */
2740          CHECK_PARTIAL();          CHECK_PARTIAL();
2741          RRETURN(MATCH_NOMATCH);          RRETURN(MATCH_NOMATCH);
2742          }          }
# Line 2690  for (;;) Line 2762  for (;;)
2762        int slength;        int slength;
2763        if ((slength = match_ref(offset, eptr, length, md, caseless)) < 0)        if ((slength = match_ref(offset, eptr, length, md, caseless)) < 0)
2764          {          {
2765            if (slength == -2) eptr = md->end_subject;   /* Partial match */
2766          CHECK_PARTIAL();          CHECK_PARTIAL();
2767          RRETURN(MATCH_NOMATCH);          RRETURN(MATCH_NOMATCH);
2768          }          }
# Line 2713  for (;;) Line 2786  for (;;)
2786          if (fi >= max) RRETURN(MATCH_NOMATCH);          if (fi >= max) RRETURN(MATCH_NOMATCH);
2787          if ((slength = match_ref(offset, eptr, length, md, caseless)) < 0)          if ((slength = match_ref(offset, eptr, length, md, caseless)) < 0)
2788            {            {
2789              if (slength == -2) eptr = md->end_subject;   /* Partial match */
2790            CHECK_PARTIAL();            CHECK_PARTIAL();
2791            RRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
2792            }            }
# Line 2731  for (;;) Line 2805  for (;;)
2805          int slength;          int slength;
2806          if ((slength = match_ref(offset, eptr, length, md, caseless)) < 0)          if ((slength = match_ref(offset, eptr, length, md, caseless)) < 0)
2807            {            {
2808            CHECK_PARTIAL();            /* Can't use CHECK_PARTIAL because we don't want to update eptr in
2809              the soft partial matching case. */
2810    
2811              if (slength == -2 && md->partial != 0 &&
2812                  md->end_subject > md->start_used_ptr)
2813                {
2814                md->hitend = TRUE;
2815                if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
2816                }
2817            break;            break;
2818            }            }
2819          eptr += slength;          eptr += slength;
2820          }          }
2821    
2822        while (eptr >= pp)        while (eptr >= pp)
2823          {          {
2824          RMATCH(eptr, ecode, offset_top, md, eptrb, RM15);          RMATCH(eptr, ecode, offset_top, md, eptrb, RM15);
# Line 3099  for (;;) Line 3182  for (;;)
3182          CHECK_PARTIAL();             /* Not SCHECK_PARTIAL() */          CHECK_PARTIAL();             /* Not SCHECK_PARTIAL() */
3183          RRETURN(MATCH_NOMATCH);          RRETURN(MATCH_NOMATCH);
3184          }          }
3185        while (length-- > 0) if (*ecode++ != *eptr++) RRETURN(MATCH_NOMATCH);        while (length-- > 0) if (*ecode++ != RAWUCHARINC(eptr)) RRETURN(MATCH_NOMATCH);
3186        }        }
3187      else      else
3188  #endif  #endif
# Line 3139  for (;;) Line 3222  for (;;)
3222    
3223        if (fc < 128)        if (fc < 128)
3224          {          {
3225          if (md->lcc[fc]          pcre_uint32 cc = RAWUCHAR(eptr);
3226              != TABLE_GET(*eptr, md->lcc, *eptr)) RRETURN(MATCH_NOMATCH);          if (md->lcc[fc] != TABLE_GET(cc, md->lcc, cc)) RRETURN(MATCH_NOMATCH);
3227          ecode++;          ecode++;
3228          eptr++;          eptr++;
3229          }          }
# Line 3151  for (;;) Line 3234  for (;;)
3234    
3235        else        else
3236          {          {
3237          unsigned int dc;          pcre_uint32 dc;
3238          GETCHARINC(dc, eptr);          GETCHARINC(dc, eptr);
3239          ecode += length;          ecode += length;
3240    
# Line 3261  for (;;) Line 3344  for (;;)
3344        if (length > 1)        if (length > 1)
3345          {          {
3346  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
3347          unsigned int othercase;          pcre_uint32 othercase;
3348          if (op >= OP_STARI &&     /* Caseless */          if (op >= OP_STARI &&     /* Caseless */
3349              (othercase = UCD_OTHERCASE(fc)) != fc)              (othercase = UCD_OTHERCASE(fc)) != fc)
3350            oclength = PRIV(ord2utf)(othercase, occhars);            oclength = PRIV(ord2utf)(othercase, occhars);
# Line 3365  for (;;) Line 3448  for (;;)
3448      maximizing, find the maximum number of characters and work backwards. */      maximizing, find the maximum number of characters and work backwards. */
3449    
3450      DPRINTF(("matching %c{%d,%d} against subject %.*s\n", fc, min, max,      DPRINTF(("matching %c{%d,%d} against subject %.*s\n", fc, min, max,
3451        max, eptr));        max, (char *)eptr));
3452    
3453      if (op >= OP_STARI)  /* Caseless */      if (op >= OP_STARI)  /* Caseless */
3454        {        {
# Line 3388  for (;;) Line 3471  for (;;)
3471    
3472        for (i = 1; i <= min; i++)        for (i = 1; i <= min; i++)
3473          {          {
3474            pcre_uint32 cc;                 /* Faster than pcre_uchar */
3475          if (eptr >= md->end_subject)          if (eptr >= md->end_subject)
3476            {            {
3477            SCHECK_PARTIAL();            SCHECK_PARTIAL();
3478            RRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
3479            }            }
3480          if (fc != *eptr && foc != *eptr) RRETURN(MATCH_NOMATCH);          cc = RAWUCHARTEST(eptr);
3481            if (fc != cc && foc != cc) RRETURN(MATCH_NOMATCH);
3482          eptr++;          eptr++;
3483          }          }
3484        if (min == max) continue;        if (min == max) continue;
# Line 3401  for (;;) Line 3486  for (;;)
3486          {          {
3487          for (fi = min;; fi++)          for (fi = min;; fi++)
3488            {            {
3489              pcre_uint32 cc;               /* Faster than pcre_uchar */
3490            RMATCH(eptr, ecode, offset_top, md, eptrb, RM24);            RMATCH(eptr, ecode, offset_top, md, eptrb, RM24);
3491            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3492            if (fi >= max) RRETURN(MATCH_NOMATCH);            if (fi >= max) RRETURN(MATCH_NOMATCH);
# Line 3409  for (;;) Line 3495  for (;;)
3495              SCHECK_PARTIAL();              SCHECK_PARTIAL();
3496              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
3497              }              }
3498            if (fc != *eptr && foc != *eptr) RRETURN(MATCH_NOMATCH);            cc = RAWUCHARTEST(eptr);
3499              if (fc != cc && foc != cc) RRETURN(MATCH_NOMATCH);
3500            eptr++;            eptr++;
3501            }            }
3502          /* Control never gets here */          /* Control never gets here */
# Line 3419  for (;;) Line 3506  for (;;)
3506          pp = eptr;          pp = eptr;
3507          for (i = min; i < max; i++)          for (i = min; i < max; i++)
3508            {            {
3509              pcre_uint32 cc;               /* Faster than pcre_uchar */
3510            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
3511              {              {
3512              SCHECK_PARTIAL();              SCHECK_PARTIAL();
3513              break;              break;
3514              }              }
3515            if (fc != *eptr && foc != *eptr) break;            cc = RAWUCHARTEST(eptr);
3516              if (fc != cc && foc != cc) break;
3517            eptr++;            eptr++;
3518            }            }
3519    
# Line 3452  for (;;) Line 3541  for (;;)
3541            SCHECK_PARTIAL();            SCHECK_PARTIAL();
3542            RRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
3543            }            }
3544          if (fc != *eptr++) RRETURN(MATCH_NOMATCH);          if (fc != RAWUCHARINCTEST(eptr)) RRETURN(MATCH_NOMATCH);
3545          }          }
3546    
3547        if (min == max) continue;        if (min == max) continue;
# Line 3469  for (;;) Line 3558  for (;;)
3558              SCHECK_PARTIAL();              SCHECK_PARTIAL();
3559              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
3560              }              }
3561            if (fc != *eptr++) RRETURN(MATCH_NOMATCH);            if (fc != RAWUCHARINCTEST(eptr)) RRETURN(MATCH_NOMATCH);
3562            }            }
3563          /* Control never gets here */          /* Control never gets here */
3564          }          }
# Line 3483  for (;;) Line 3572  for (;;)
3572              SCHECK_PARTIAL();              SCHECK_PARTIAL();
3573              break;              break;
3574              }              }
3575            if (fc != *eptr) break;            if (fc != RAWUCHARTEST(eptr)) break;
3576            eptr++;            eptr++;
3577            }            }
3578          if (possessive) continue;          if (possessive) continue;
# Line 3509  for (;;) Line 3598  for (;;)
3598        SCHECK_PARTIAL();        SCHECK_PARTIAL();
3599        RRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
3600        }        }
     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  
3601  #ifdef SUPPORT_UTF  #ifdef SUPPORT_UTF
3602        if (utf)
3603          {
3604          register pcre_uint32 ch, och;
3605    
3606          ecode++;
3607          GETCHARINC(ch, ecode);
3608          GETCHARINC(c, eptr);
3609    
3610          if (op == OP_NOT)
3611            {
3612            if (ch == c) RRETURN(MATCH_NOMATCH);
3613            }
3614          else
3615            {
3616  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
3617        if (utf && ch > 127)          if (ch > 127)
3618          och = UCD_OTHERCASE(ch);            och = UCD_OTHERCASE(ch);
3619  #else  #else
3620        if (utf && ch > 127)          if (ch > 127)
3621          och = ch;            och = ch;
3622  #endif /* SUPPORT_UCP */  #endif /* SUPPORT_UCP */
3623        else          else
3624  #endif /* SUPPORT_UTF */            och = TABLE_GET(ch, md->fcc, ch);
3625          och = TABLE_GET(ch, md->fcc, ch);          if (ch == c || och == c) RRETURN(MATCH_NOMATCH);
3626  #endif /* COMPILE_PCRE8 */          }
       if (ch == c || och == c) RRETURN(MATCH_NOMATCH);  
3627        }        }
3628      else    /* Caseful */      else
3629    #endif
3630        {        {
3631        if (*ecode++ == c) RRETURN(MATCH_NOMATCH);        register pcre_uint32 ch = ecode[1];
3632          c = *eptr++;
3633          if (ch == c || (op == OP_NOTI && TABLE_GET(ch, md->fcc, ch) == c))
3634            RRETURN(MATCH_NOMATCH);
3635          ecode += 2;
3636        }        }
3637      break;      break;
3638    
# Line 3615  for (;;) Line 3712  for (;;)
3712      /* Common code for all repeated single-byte matches. */      /* Common code for all repeated single-byte matches. */
3713    
3714      REPEATNOTCHAR:      REPEATNOTCHAR:
3715      fc = *ecode++;      GETCHARINCTEST(fc, ecode);
3716    
3717      /* The code is duplicated for the caseless and caseful cases, for speed,      /* The code is duplicated for the caseless and caseful cases, for speed,
3718      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 3626  for (;;) Line 3723  for (;;)
3723      characters and work backwards. */      characters and work backwards. */
3724    
3725      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,
3726        max, eptr));        max, (char *)eptr));
3727    
3728      if (op >= OP_NOTSTARI)     /* Caseless */      if (op >= OP_NOTSTARI)     /* Caseless */
3729        {        {
 #ifdef COMPILE_PCRE8  
       /* fc must be < 128 if UTF is enabled. */  
       foc = md->fcc[fc];  
 #else  
3730  #ifdef SUPPORT_UTF  #ifdef SUPPORT_UTF
3731  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
3732        if (utf && fc > 127)        if (utf && fc > 127)
# Line 3645  for (;;) Line 3738  for (;;)
3738        else        else
3739  #endif /* SUPPORT_UTF */  #endif /* SUPPORT_UTF */
3740          foc = TABLE_GET(fc, md->fcc, fc);          foc = TABLE_GET(fc, md->fcc, fc);
 #endif /* COMPILE_PCRE8 */  
3741    
3742  #ifdef SUPPORT_UTF  #ifdef SUPPORT_UTF
3743        if (utf)        if (utf)
3744          {          {
3745          register unsigned int d;          register pcre_uint32 d;
3746          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
3747            {            {
3748            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
# Line 3659  for (;;) Line 3751  for (;;)
3751              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
3752              }              }
3753            GETCHARINC(d, eptr);            GETCHARINC(d, eptr);
3754            if (fc == d || foc == d) RRETURN(MATCH_NOMATCH);            if (fc == d || (unsigned int)foc == d) RRETURN(MATCH_NOMATCH);
3755            }            }
3756          }          }
3757        else        else
# Line 3685  for (;;) Line 3777  for (;;)
3777  #ifdef SUPPORT_UTF  #ifdef SUPPORT_UTF
3778          if (utf)          if (utf)
3779            {            {
3780            register unsigned int d;            register pcre_uint32 d;
3781            for (fi = min;; fi++)            for (fi = min;; fi++)
3782              {              {
3783              RMATCH(eptr, ecode, offset_top, md, eptrb, RM28);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM28);
# Line 3697  for (;;) Line 3789  for (;;)
3789                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
3790                }                }
3791              GETCHARINC(d, eptr);              GETCHARINC(d, eptr);
3792              if (fc == d || foc == d) RRETURN(MATCH_NOMATCH);              if (fc == d || (unsigned int)foc == d) RRETURN(MATCH_NOMATCH);
3793              }              }
3794            }            }
3795          else          else
# Line 3730  for (;;) Line 3822  for (;;)
3822  #ifdef SUPPORT_UTF  #ifdef SUPPORT_UTF
3823          if (utf)          if (utf)
3824            {            {
3825            register unsigned int d;            register pcre_uint32 d;
3826            for (i = min; i < max; i++)            for (i = min; i < max; i++)
3827              {              {
3828              int len = 1;              int len = 1;
# Line 3740  for (;;) Line 3832  for (;;)
3832                break;                break;
3833                }                }
3834              GETCHARLEN(d, eptr, len);              GETCHARLEN(d, eptr, len);
3835              if (fc == d || foc == d) break;              if (fc == d || (unsigned int)foc == d) break;
3836              eptr += len;              eptr += len;
3837              }              }
3838            if (possessive) continue;            if (possessive) continue;
# Line 3787  for (;;) Line 3879  for (;;)
3879  #ifdef SUPPORT_UTF  #ifdef SUPPORT_UTF
3880        if (utf)        if (utf)
3881          {          {
3882          register unsigned int d;          register pcre_uint32 d;
3883          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
3884            {            {
3885            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
# Line 3821  for (;;) Line 3913  for (;;)
3913  #ifdef SUPPORT_UTF  #ifdef SUPPORT_UTF
3914          if (utf)          if (utf)
3915            {            {
3916            register unsigned int d;            register pcre_uint32 d;
3917            for (fi = min;; fi++)            for (fi = min;; fi++)
3918              {              {
3919              RMATCH(eptr, ecode, offset_top, md, eptrb, RM32);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM32);
# Line 3865  for (;;) Line 3957  for (;;)
3957  #ifdef SUPPORT_UTF  #ifdef SUPPORT_UTF
3958          if (utf)          if (utf)
3959            {            {
3960            register unsigned int d;            register pcre_uint32 d;
3961            for (i = min; i < max; i++)            for (i = min; i < max; i++)
3962              {              {
3963              int len = 1;              int len = 1;
# Line 4141  for (;;) Line 4233  for (;;)
4233              }              }
4234            break;            break;
4235    
4236              case PT_CLIST:
4237              for (i = 1; i <= min; i++)
4238                {
4239                const pcre_uint32 *cp;
4240                if (eptr >= md->end_subject)
4241                  {
4242                  SCHECK_PARTIAL();
4243                  RRETURN(MATCH_NOMATCH);
4244                  }
4245                GETCHARINCTEST(c, eptr);
4246                cp = PRIV(ucd_caseless_sets) + prop_value;
4247                for (;;)
4248                  {
4249                  if (c < *cp)
4250                    { if (prop_fail_result) break; else { RRETURN(MATCH_NOMATCH); } }
4251                  if (c == *cp++)
4252                    { if (prop_fail_result) { RRETURN(MATCH_NOMATCH); } else break; }
4253                  }
4254                }
4255              break;
4256    
4257              case PT_UCNC:
4258              for (i = 1; i <= min; i++)
4259                {
4260                if (eptr >= md->end_subject)
4261                  {
4262                  SCHECK_PARTIAL();
4263                  RRETURN(MATCH_NOMATCH);
4264                  }
4265                GETCHARINCTEST(c, eptr);
4266                if ((c == CHAR_DOLLAR_SIGN || c == CHAR_COMMERCIAL_AT ||
4267                     c == CHAR_GRAVE_ACCENT || (c >= 0xa0 && c <= 0xd7ff) ||
4268                     c >= 0xe000) == prop_fail_result)
4269                  RRETURN(MATCH_NOMATCH);
4270                }
4271              break;
4272    
4273            /* This should not occur */            /* This should not occur */
4274    
4275            default:            default:
# Line 4160  for (;;) Line 4289  for (;;)
4289              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4290              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4291              }              }
4292            GETCHARINCTEST(c, eptr);            else
           if (UCD_CATEGORY(c) == ucp_M) RRETURN(MATCH_NOMATCH);  
           while (eptr < md->end_subject)  
4293              {              {
4294              int len = 1;              int lgb, rgb;
4295              if (!utf) c = *eptr; else { GETCHARLEN(c, eptr, len); }              GETCHARINCTEST(c, eptr);
4296              if (UCD_CATEGORY(c) != ucp_M) break;              lgb = UCD_GRAPHBREAK(c);
4297              eptr += len;             while (eptr < md->end_subject)
4298                  {
4299                  int len = 1;
4300                  if (!utf) c = *eptr; else { GETCHARLEN(c, eptr, len); }
4301                  rgb = UCD_GRAPHBREAK(c);
4302                  if ((PRIV(ucp_gbtable)[lgb] & (1 << rgb)) == 0) break;
4303                  lgb = rgb;
4304                  eptr += len;
4305                  }
4306              }              }
4307              CHECK_PARTIAL();
4308            }            }
4309          }          }
4310    
# Line 4189  for (;;) Line 4325  for (;;)
4325              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4326              }              }
4327            if (IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH);            if (IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH);
4328              if (md->partial != 0 &&
4329                  eptr + 1 >= md->end_subject &&
4330                  NLBLOCK->nltype == NLTYPE_FIXED &&
4331                  NLBLOCK->nllen == 2 &&
4332                  RAWUCHAR(eptr) == NLBLOCK->nl[0])
4333                {
4334                md->hitend = TRUE;
4335                if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
4336                }
4337            eptr++;            eptr++;
4338            ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);            ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);
4339            }            }
# Line 4225  for (;;) Line 4370  for (;;)
4370              {              {
4371              default: RRETURN(MATCH_NOMATCH);              default: RRETURN(MATCH_NOMATCH);
4372    
4373              case 0x000d:              case CHAR_CR:
4374              if (eptr < md->end_subject && *eptr == 0x0a) eptr++;              if (eptr < md->end_subject && RAWUCHAR(eptr) == CHAR_LF) eptr++;
4375              break;              break;
4376    
4377              case 0x000a:              case CHAR_LF:
4378              break;              break;
4379    
4380              case 0x000b:              case CHAR_VT:
4381              case 0x000c:              case CHAR_FF:
4382              case 0x0085:              case CHAR_NEL:
4383    #ifndef EBCDIC
4384              case 0x2028:              case 0x2028:
4385              case 0x2029:              case 0x2029:
4386    #endif  /* Not EBCDIC */
4387              if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);              if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);
4388              break;              break;
4389              }              }
# Line 4254  for (;;) Line 4401  for (;;)
4401            GETCHARINC(c, eptr);            GETCHARINC(c, eptr);
4402            switch(c)            switch(c)
4403              {              {
4404                HSPACE_CASES: RRETURN(MATCH_NOMATCH);  /* Byte and multibyte cases */
4405              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);  
4406              }              }
4407            }            }
4408          break;          break;
# Line 4290  for (;;) Line 4418  for (;;)
4418            GETCHARINC(c, eptr);            GETCHARINC(c, eptr);
4419            switch(c)            switch(c)
4420              {              {
4421                HSPACE_CASES: break;  /* Byte and multibyte cases */
4422              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;  
4423              }              }
4424            }            }
4425          break;          break;
# Line 4326  for (;;) Line 4435  for (;;)
4435            GETCHARINC(c, eptr);            GETCHARINC(c, eptr);
4436            switch(c)            switch(c)
4437              {              {
4438                VSPACE_CASES: RRETURN(MATCH_NOMATCH);
4439              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);  
4440              }              }
4441            }            }
4442          break;          break;
# Line 4350  for (;;) Line 4452  for (;;)
4452            GETCHARINC(c, eptr);            GETCHARINC(c, eptr);
4453            switch(c)            switch(c)
4454              {              {
4455                VSPACE_CASES: break;
4456              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;  
4457              }              }
4458            }            }
4459          break;          break;
# Line 4380  for (;;) Line 4475  for (;;)
4475          case OP_DIGIT:          case OP_DIGIT:
4476          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
4477            {            {
4478              pcre_uint32 cc;
4479            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
4480              {              {
4481              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4482              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4483              }              }
4484            if (*eptr >= 128 || (md->ctypes[*eptr] & ctype_digit) == 0)            cc = RAWUCHAR(eptr);
4485              if (cc >= 128 || (md->ctypes[cc] & ctype_digit) == 0)
4486              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4487            eptr++;            eptr++;
4488            /* 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 4395  for (;;) Line 4492  for (;;)
4492          case OP_NOT_WHITESPACE:          case OP_NOT_WHITESPACE:
4493          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
4494            {            {
4495              pcre_uint32 cc;
4496            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
4497              {              {
4498              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4499              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4500              }              }
4501            if (*eptr < 128 && (md->ctypes[*eptr] & ctype_space) != 0)            cc = RAWUCHAR(eptr);
4502              if (cc < 128 && (md->ctypes[cc] & ctype_space) != 0)
4503              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4504            eptr++;            eptr++;
4505            ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);            ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);
# Line 4410  for (;;) Line 4509  for (;;)
4509          case OP_WHITESPACE:          case OP_WHITESPACE:
4510          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
4511            {            {
4512              pcre_uint32 cc;
4513            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
4514              {              {
4515              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4516              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4517              }              }
4518            if (*eptr >= 128 || (md->ctypes[*eptr] & ctype_space) == 0)            cc = RAWUCHAR(eptr);
4519              if (cc >= 128 || (md->ctypes[cc] & ctype_space) == 0)
4520              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4521            eptr++;            eptr++;
4522            /* 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 4425  for (;;) Line 4526  for (;;)
4526          case OP_NOT_WORDCHAR:          case OP_NOT_WORDCHAR:
4527          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
4528            {            {
4529              pcre_uint32 cc;
4530            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
4531              {              {
4532              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4533              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4534              }              }
4535            if (*eptr < 128 && (md->ctypes[*eptr] & ctype_word) != 0)            cc = RAWUCHAR(eptr);
4536              if (cc < 128 && (md->ctypes[cc] & ctype_word) != 0)
4537              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4538            eptr++;            eptr++;
4539            ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);            ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);
# Line 4440  for (;;) Line 4543  for (;;)
4543          case OP_WORDCHAR:          case OP_WORDCHAR:
4544          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
4545            {            {
4546              pcre_uint32 cc;
4547            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
4548              {              {
4549              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4550              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4551              }              }
4552            if (*eptr >= 128 || (md->ctypes[*eptr] & ctype_word) == 0)            cc = RAWUCHAR(eptr);
4553              if (cc >= 128 || (md->ctypes[cc] & ctype_word) == 0)
4554              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4555            eptr++;            eptr++;
4556            /* 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 4473  for (;;) Line 4578  for (;;)
4578              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4579              }              }
4580            if (IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH);            if (IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH);
4581              if (md->partial != 0 &&
4582                  eptr + 1 >= md->end_subject &&
4583                  NLBLOCK->nltype == NLTYPE_FIXED &&
4584                  NLBLOCK->nllen == 2 &&
4585                  *eptr == NLBLOCK->nl[0])
4586                {
4587                md->hitend = TRUE;
4588                if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
4589                }
4590            eptr++;            eptr++;
4591            }            }
4592          break;          break;
# Line 4507  for (;;) Line 4621  for (;;)
4621              {              {
4622              default: RRETURN(MATCH_NOMATCH);              default: RRETURN(MATCH_NOMATCH);
4623    
4624              case 0x000d:              case CHAR_CR:
4625              if (eptr < md->end_subject && *eptr == 0x0a) eptr++;              if (eptr < md->end_subject && *eptr == CHAR_LF) eptr++;
4626              break;              break;
4627    
4628              case 0x000a:              case CHAR_LF:
4629              break;              break;
4630    
4631              case 0x000b:              case CHAR_VT:
4632              case 0x000c:              case CHAR_FF:
4633              case 0x0085:              case CHAR_NEL:
4634  #ifdef COMPILE_PCRE16  #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
4635              case 0x2028:              case 0x2028:
4636              case 0x2029:              case 0x2029:
4637  #endif  #endif
# Line 4538  for (;;) Line 4652  for (;;)
4652            switch(*eptr++)            switch(*eptr++)
4653              {              {
4654              default: break;              default: break;
4655              case 0x09:      /* HT */              HSPACE_BYTE_CASES:
4656              case 0x20:      /* SPACE */  #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
4657              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 */  
4658  #endif  #endif
4659              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4660              }              }
# Line 4575  for (;;) Line 4672  for (;;)
4672            switch(*eptr++)            switch(*eptr++)
4673              {              {
4674              default: RRETURN(MATCH_NOMATCH);              default: RRETURN(MATCH_NOMATCH);
4675              case 0x09:      /* HT */              HSPACE_BYTE_CASES:
4676              case 0x20:      /* SPACE */  #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
4677              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 */  
4678  #endif  #endif
4679              break;              break;
4680              }              }
# Line 4611  for (;;) Line 4691  for (;;)
4691              }              }
4692            switch(*eptr++)            switch(*eptr++)
4693              {              {
4694              default: break;              VSPACE_BYTE_CASES:
4695              case 0x0a:      /* LF */  #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
4696              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 */  
4697  #endif  #endif
4698              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4699                default: break;
4700              }              }
4701            }            }
4702          break;          break;
# Line 4637  for (;;) Line 4712  for (;;)
4712            switch(*eptr++)            switch(*eptr++)
4713              {              {
4714              default: RRETURN(MATCH_NOMATCH);              default: RRETURN(MATCH_NOMATCH);
4715              case 0x0a:      /* LF */              VSPACE_BYTE_CASES:
4716              case 0x0b:      /* VT */  #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
4717              case 0x0c:      /* FF */              VSPACE_MULTIBYTE_CASES:
             case 0x0d:      /* CR */  
             case 0x85:      /* NEL */  
 #ifdef COMPILE_PCRE16  
             case 0x2028:    /* LINE SEPARATOR */  
             case 0x2029:    /* PARAGRAPH SEPARATOR */  
4718  #endif  #endif
4719              break;              break;
4720              }              }
# Line 4922  for (;;) Line 4992  for (;;)
4992              }              }
4993            /* Control never gets here */            /* Control never gets here */
4994    
4995            /* This should never occur */            case PT_CLIST:
4996              for (fi = min;; fi++)
4997                {
4998                const pcre_uint32 *cp;
4999                RMATCH(eptr, ecode, offset_top, md, eptrb, RM67);
5000                if (rrc != MATCH_NOMATCH) RRETURN(rrc);
5001                if (fi >= max) RRETURN(MATCH_NOMATCH);
5002                if (eptr >= md->end_subject)
5003                  {
5004                  SCHECK_PARTIAL();
5005                  RRETURN(MATCH_NOMATCH);
5006                  }
5007                GETCHARINCTEST(c, eptr);
5008                cp = PRIV(ucd_caseless_sets) + prop_value;
5009                for (;;)
5010                  {
5011                  if (c < *cp)
5012                    { if (prop_fail_result) break; else { RRETURN(MATCH_NOMATCH); } }
5013                  if (c == *cp++)
5014                    { if (prop_fail_result) { RRETURN(MATCH_NOMATCH); } else break; }
5015                  }
5016                }
5017              /* Control never gets here */
5018    
5019              case PT_UCNC:
5020              for (fi = min;; fi++)
5021                {
5022                RMATCH(eptr, ecode, offset_top, md, eptrb, RM68);
5023                if (rrc != MATCH_NOMATCH) RRETURN(rrc);
5024                if (fi >= max) RRETURN(MATCH_NOMATCH);
5025                if (eptr >= md->end_subject)
5026                  {
5027                  SCHECK_PARTIAL();
5028                  RRETURN(MATCH_NOMATCH);
5029                  }
5030                GETCHARINCTEST(c, eptr);
5031                if ((c == CHAR_DOLLAR_SIGN || c == CHAR_COMMERCIAL_AT ||
5032                     c == CHAR_GRAVE_ACCENT || (c >= 0xa0 && c <= 0xd7ff) ||
5033                     c >= 0xe000) == prop_fail_result)
5034                  RRETURN(MATCH_NOMATCH);
5035                }
5036              /* Control never gets here */
5037    
5038              /* This should never occur */
5039            default:            default:
5040            RRETURN(PCRE_ERROR_INTERNAL);            RRETURN(PCRE_ERROR_INTERNAL);
5041            }            }
# Line 4944  for (;;) Line 5056  for (;;)
5056              SCHECK_PARTIAL();              SCHECK_PARTIAL();
5057              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
5058              }              }
5059            GETCHARINCTEST(c, eptr);            else
           if (UCD_CATEGORY(c) == ucp_M) RRETURN(MATCH_NOMATCH);  
           while (eptr < md->end_subject)  
5060              {              {
5061              int len = 1;              int lgb, rgb;
5062              if (!utf) c = *eptr; else { GETCHARLEN(c, eptr, len); }              GETCHARINCTEST(c, eptr);
5063              if (UCD_CATEGORY(c) != ucp_M) break;              lgb = UCD_GRAPHBREAK(c);
5064              eptr += len;              while (eptr < md->end_subject)
5065                  {
5066                  int len = 1;
5067                  if (!utf) c = *eptr; else { GETCHARLEN(c, eptr, len); }
5068                  rgb = UCD_GRAPHBREAK(c);
5069                  if ((PRIV(ucp_gbtable)[lgb] & (1 << rgb)) == 0) break;
5070                  lgb = rgb;
5071                  eptr += len;
5072                  }
5073              }              }
5074              CHECK_PARTIAL();
5075            }            }
5076          }          }
5077        else        else
# Line 4976  for (;;) Line 5095  for (;;)
5095            GETCHARINC(c, eptr);            GETCHARINC(c, eptr);
5096            switch(ctype)            switch(ctype)
5097              {              {
5098              case OP_ANY:        /* This is the non-NL case */              case OP_ANY:               /* This is the non-NL case */
5099                if (md->partial != 0 &&    /* Take care with CRLF partial */
5100                    eptr >= md->end_subject &&
5101                    NLBLOCK->nltype == NLTYPE_FIXED &&
5102                    NLBLOCK->nllen == 2 &&
5103                    c == NLBLOCK->nl[0])
5104                  {
5105                  md->hitend = TRUE;
5106                  if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
5107                  }
5108                break;
5109    
5110              case OP_ALLANY:              case OP_ALLANY:
5111              case OP_ANYBYTE:              case OP_ANYBYTE:
5112              break;              break;
# Line 4985  for (;;) Line 5115  for (;;)
5115              switch(c)              switch(c)
5116                {                {
5117                default: RRETURN(MATCH_NOMATCH);                default: RRETURN(MATCH_NOMATCH);
5118                case 0x000d:                case CHAR_CR:
5119                if (eptr < md->end_subject && *eptr == 0x0a) eptr++;                if (eptr < md->end_subject && RAWUCHAR(eptr) == CHAR_LF) eptr++;
5120                break;                break;
5121                case 0x000a:  
5122                  case CHAR_LF:
5123                break;                break;
5124    
5125                case 0x000b:                case CHAR_VT:
5126                case 0x000c:                case CHAR_FF:
5127                case 0x0085:                case CHAR_NEL:
5128    #ifndef EBCDIC
5129                case 0x2028:                case 0x2028:
5130                case 0x2029:                case 0x2029:
5131    #endif  /* Not EBCDIC */
5132                if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);                if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);
5133                break;                break;
5134                }                }
# Line 5004  for (;;) Line 5137  for (;;)
5137              case OP_NOT_HSPACE:              case OP_NOT_HSPACE:
5138              switch(c)              switch(c)
5139                {                {
5140                  HSPACE_CASES: RRETURN(MATCH_NOMATCH);
5141                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);  
5142                }                }
5143              break;              break;
5144    
5145              case OP_HSPACE:              case OP_HSPACE:
5146              switch(c)              switch(c)
5147                {                {
5148                  HSPACE_CASES: break;
5149                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;  
5150                }                }
5151              break;              break;
5152    
5153              case OP_NOT_VSPACE:              case OP_NOT_VSPACE:
5154              switch(c)              switch(c)
5155                {                {
5156                  VSPACE_CASES: RRETURN(MATCH_NOMATCH);
5157                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);  
5158                }                }
5159              break;              break;
5160    
5161              case OP_VSPACE:              case OP_VSPACE:
5162              switch(c)              switch(c)
5163                {                {
5164                  VSPACE_CASES: break;
5165                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;  
5166                }                }
5167              break;              break;
5168    
# Line 5139  for (;;) Line 5220  for (;;)
5220            c = *eptr++;            c = *eptr++;
5221            switch(ctype)            switch(ctype)
5222              {              {
5223              case OP_ANY:     /* This is the non-NL case */              case OP_ANY:               /* This is the non-NL case */
5224                if (md->partial != 0 &&    /* Take care with CRLF partial */
5225                    eptr >= md->end_subject &&
5226                    NLBLOCK->nltype == NLTYPE_FIXED &&
5227                    NLBLOCK->nllen == 2 &&
5228                    c == NLBLOCK->nl[0])
5229                  {
5230                  md->hitend = TRUE;
5231                  if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
5232                  }
5233                break;
5234    
5235              case OP_ALLANY:              case OP_ALLANY:
5236              case OP_ANYBYTE:              case OP_ANYBYTE:
5237              break;              break;
# Line 5148  for (;;) Line 5240  for (;;)
5240              switch(c)              switch(c)
5241                {                {
5242                default: RRETURN(MATCH_NOMATCH);                default: RRETURN(MATCH_NOMATCH);
5243                case 0x000d:                case CHAR_CR:
5244                if (eptr < md->end_subject && *eptr == 0x0a) eptr++;                if (eptr < md->end_subject && *eptr == CHAR_LF) eptr++;
5245                break;                break;
5246    
5247                case 0x000a:                case CHAR_LF:
5248                break;                break;
5249    
5250                case 0x000b:                case CHAR_VT:
5251                case 0x000c:                case CHAR_FF:
5252                case 0x0085:                case CHAR_NEL:
5253  #ifdef COMPILE_PCRE16  #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
5254                case 0x2028:                case 0x2028:
5255                case 0x2029:                case 0x2029:
5256  #endif  #endif
# Line 5171  for (;;) Line 5263  for (;;)
5263              switch(c)              switch(c)
5264                {                {
5265                default: break;                default: break;
5266                case 0x09:      /* HT */                HSPACE_BYTE_CASES:
5267                case 0x20:      /* SPACE */  #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
5268                case 0xa0:      /* NBSP */                HSPACE_MULTIBYTE_CASES:
 #ifdef COMPILE_PCRE16  
               case 0x1680:    /* OGHAM SPACE MARK */  
               case 0x180e:    /* MONGOLIAN VOWEL SEPARATOR */  
               case 0x2000:    /* EN QUAD */  
               case 0x2001:    /* EM QUAD */  
               case 0x2002:    /* EN SPACE */  
               case 0x2003:    /* EM SPACE */  
               case 0x2004:    /* THREE-PER-EM SPACE */  
               case 0x2005:    /* FOUR-PER-EM SPACE */  
               case 0x2006:    /* SIX-PER-EM SPACE */  
               case 0x2007:    /* FIGURE SPACE */  
               case 0x2008:    /* PUNCTUATION SPACE */  
               case 0x2009:    /* THIN SPACE */  
               case 0x200A:    /* HAIR SPACE */  
               case 0x202f:    /* NARROW NO-BREAK SPACE */  
               case 0x205f:    /* MEDIUM MATHEMATICAL SPACE */  
               case 0x3000:    /* IDEOGRAPHIC SPACE */  
5269  #endif  #endif
5270                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
5271                }                }
# Line 5200  for (;;) Line 5275  for (;;)
5275              switch(c)              switch(c)
5276                {                {
5277                default: RRETURN(MATCH_NOMATCH);                default: RRETURN(MATCH_NOMATCH);
5278                case 0x09:      /* HT */                HSPACE_BYTE_CASES:
5279                case 0x20:      /* SPACE */  #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
5280                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 */  
5281  #endif  #endif
5282                break;                break;
5283                }                }
# Line 5229  for (;;) Line 5287  for (;;)
5287              switch(c)              switch(c)
5288                {                {
5289                default: break;                default: break;
5290                case 0x0a:      /* LF */                VSPACE_BYTE_CASES:
5291                case 0x0b:      /* VT */  #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
5292                case 0x0c:      /* FF */                VSPACE_MULTIBYTE_CASES:
               case 0x0d:      /* CR */  
               case 0x85:      /* NEL */  
 #ifdef COMPILE_PCRE16  
               case 0x2028:    /* LINE SEPARATOR */  
               case 0x2029:    /* PARAGRAPH SEPARATOR */  
5293  #endif  #endif
5294                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
5295                }                }
# Line 5246  for (;;) Line 5299  for (;;)
5299              switch(c)              switch(c)
5300                {                {
5301                default: RRETURN(MATCH_NOMATCH);                default: RRETURN(MATCH_NOMATCH);
5302                case 0x0a:      /* LF */                VSPACE_BYTE_CASES:
5303                case 0x0b:      /* VT */  #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
5304                case 0x0c:      /* FF */                VSPACE_MULTIBYTE_CASES:
               case 0x0d:      /* CR */  
               case 0x85:      /* NEL */  
 #ifdef COMPILE_PCRE16  
               case 0x2028:    /* LINE SEPARATOR */  
               case 0x2029:    /* PARAGRAPH SEPARATOR */  
5305  #endif  #endif
5306                break;                break;
5307                }                }
# Line 5457  for (;;) Line 5505  for (;;)
5505              }              }
5506            break;            break;
5507    
5508              case PT_CLIST:
5509              for (i = min; i < max; i++)
5510                {
5511                const pcre_uint32 *cp;
5512                int len = 1;
5513                if (eptr >= md->end_subject)
5514                  {
5515                  SCHECK_PARTIAL();
5516                  break;
5517                  }
5518                GETCHARLENTEST(c, eptr, len);
5519                cp = PRIV(ucd_caseless_sets) + prop_value;
5520                for (;;)
5521                  {
5522                  if (c < *cp)
5523                    { if (prop_fail_result) break; else goto GOT_MAX; }
5524                  if (c == *cp++)
5525                    { if (prop_fail_result) goto GOT_MAX; else break; }
5526                  }
5527                eptr += len;
5528                }
5529              GOT_MAX:
5530              break;
5531    
5532              case PT_UCNC:
5533              for (i = min; i < max; i++)
5534                {
5535                int len = 1;
5536                if (eptr >= md->end_subject)
5537                  {
5538                  SCHECK_PARTIAL();
5539                  break;
5540                  }
5541                GETCHARLENTEST(c, eptr, len);
5542                if ((c == CHAR_DOLLAR_SIGN || c == CHAR_COMMERCIAL_AT ||
5543                     c == CHAR_GRAVE_ACCENT || (c >= 0xa0 && c <= 0xd7ff) ||
5544                     c >= 0xe000) == prop_fail_result)
5545                  break;
5546                eptr += len;
5547                }
5548              break;
5549    
5550            default:            default:
5551            RRETURN(PCRE_ERROR_INTERNAL);            RRETURN(PCRE_ERROR_INTERNAL);
5552            }            }
# Line 5480  for (;;) Line 5570  for (;;)
5570          {          {
5571          for (i = min; i < max; i++)          for (i = min; i < max; i++)
5572            {            {
           int len = 1;  
5573            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
5574              {              {
5575              SCHECK_PARTIAL();              SCHECK_PARTIAL();
5576              break;              break;
5577              }              }
5578            if (!utf) c = *eptr; else { GETCHARLEN(c, eptr, len); }            else
           if (UCD_CATEGORY(c) == ucp_M) break;  
           eptr += len;  
           while (eptr < md->end_subject)  
5579              {              {
5580              len = 1;              int lgb, rgb;
5581              if (!utf) c = *eptr; else { GETCHARLEN(c, eptr, len); }              GETCHARINCTEST(c, eptr);
5582              if (UCD_CATEGORY(c) != ucp_M) break;              lgb = UCD_GRAPHBREAK(c);
5583              eptr += len;              while (eptr < md->end_subject)
5584                  {
5585                  int len = 1;
5586                  if (!utf) c = *eptr; else { GETCHARLEN(c, eptr, len); }
5587                  rgb = UCD_GRAPHBREAK(c);
5588                  if ((PRIV(ucp_gbtable)[lgb] & (1 << rgb)) == 0) break;
5589                  lgb = rgb;
5590                  eptr += len;
5591                  }
5592              }              }
5593              CHECK_PARTIAL();
5594            }            }
5595    
5596          /* eptr is now past the end of the maximum run */          /* eptr is now past the end of the maximum run */
# Line 5539  for (;;) Line 5634  for (;;)
5634                  break;                  break;
5635                  }                  }
5636                if (IS_NEWLINE(eptr)) break;                if (IS_NEWLINE(eptr)) break;
5637                  if (md->partial != 0 &&    /* Take care with CRLF partial */
5638                      eptr + 1 >= md->end_subject &&
5639                      NLBLOCK->nltype == NLTYPE_FIXED &&
5640                      NLBLOCK->nllen == 2 &&
5641                      RAWUCHAR(eptr) == NLBLOCK->nl[0])
5642                    {
5643                    md->hitend = TRUE;
5644                    if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
5645                    }
5646                eptr++;                eptr++;
5647                ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);                ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);
5648                }                }
# Line 5556  for (;;) Line 5660  for (;;)
5660                  break;                  break;
5661                  }                  }
5662                if (IS_NEWLINE(eptr)) break;                if (IS_NEWLINE(eptr)) break;
5663                  if (md->partial != 0 &&    /* Take care with CRLF partial */
5664                      eptr + 1 >= md->end_subject &&
5665                      NLBLOCK->nltype == NLTYPE_FIXED &&
5666                      NLBLOCK->nllen == 2 &&
5667                      RAWUCHAR(eptr) == NLBLOCK->nl[0])
5668                    {
5669                    md->hitend = TRUE;
5670                    if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
5671                    }
5672                eptr++;                eptr++;
5673                ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);                ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);
5674                }                }
# Line 5605  for (;;) Line 5718  for (;;)
5718                break;                break;
5719                }                }
5720              GETCHARLEN(c, eptr, len);              GETCHARLEN(c, eptr, len);
5721              if (c == 0x000d)              if (c == CHAR_CR)
5722                {                {
5723                if (++eptr >= md->end_subject) break;                if (++eptr >= md->end_subject) break;
5724                if (*eptr == 0x000a) eptr++;                if (RAWUCHAR(eptr) == CHAR_LF) eptr++;
5725                }                }
5726              else              else
5727                {                {
5728                if (c != 0x000a &&                if (c != CHAR_LF &&
5729                    (md->bsr_anycrlf ||                    (md->bsr_anycrlf ||
5730                     (c != 0x000b && c != 0x000c &&                     (c != CHAR_VT && c != CHAR_FF && c != CHAR_NEL
5731                      c != 0x0085 && c != 0x2028 && c != 0x2029)))  #ifndef EBCDIC
5732                        && c != 0x2028 && c != 0x2029
5733    #endif  /* Not EBCDIC */
5734                        )))
5735                  break;                  break;
5736                eptr += len;                eptr += len;
5737                }                }
# Line 5636  for (;;) Line 5752  for (;;)
5752              GETCHARLEN(c, eptr, len);              GETCHARLEN(c, eptr, len);
5753              switch(c)              switch(c)
5754                {                {
5755                  HSPACE_CASES: gotspace = TRUE; break;
5756                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;  
5757                }                }
5758              if (gotspace == (ctype == OP_NOT_HSPACE)) break;              if (gotspace == (ctype == OP_NOT_HSPACE)) break;
5759              eptr += len;              eptr += len;
# Line 5678  for (;;) Line 5774  for (;;)
5774              GETCHARLEN(c, eptr, len);              GETCHARLEN(c, eptr, len);
5775              switch(c)              switch(c)
5776                {                {
5777                  VSPACE_CASES: gotspace = TRUE; break;
5778                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;  
5779                }                }
5780              if (gotspace == (ctype == OP_NOT_VSPACE)) break;              if (gotspace == (ctype == OP_NOT_VSPACE)) break;
5781              eptr += len;              eptr += len;
# Line 5801  for (;;) Line 5889  for (;;)
5889            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
5890            if (eptr-- == pp) break;        /* Stop if tried at original pos */            if (eptr-- == pp) break;        /* Stop if tried at original pos */
5891            BACKCHAR(eptr);            BACKCHAR(eptr);
5892            if (ctype == OP_ANYNL && eptr > pp  && *eptr == '\n' &&            if (ctype == OP_ANYNL && eptr > pp  && RAWUCHAR(eptr) == CHAR_NL &&
5893                eptr[-1] == '\r') eptr--;                RAWUCHAR(eptr - 1) == CHAR_CR) eptr--;
5894            }            }
5895          }          }
5896        else        else
# Line 5820  for (;;) Line 5908  for (;;)
5908                break;                break;
5909                }                }
5910              if (IS_NEWLINE(eptr)) break;              if (IS_NEWLINE(eptr)) break;
5911                if (md->partial != 0 &&    /* Take care with CRLF partial */
5912                    eptr + 1 >= md->end_subject &&
5913                    NLBLOCK->nltype == NLTYPE_FIXED &&
5914                    NLBLOCK->nllen == 2 &&
5915                    *eptr == NLBLOCK->nl[0])
5916                  {
5917                  md->hitend = TRUE;
5918                  if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
5919                  }
5920              eptr++;              eptr++;
5921              }              }
5922            break;            break;
# Line 5844  for (;;) Line 5941  for (;;)
5941                break;                break;
5942                }                }
5943              c = *eptr;              c = *eptr;
5944              if (c == 0x000d)              if (c == CHAR_CR)
5945                {                {
5946                if (++eptr >= md->end_subject) break;                if (++eptr >= md->end_subject) break;
5947                if (*eptr == 0x000a) eptr++;                if (*eptr == CHAR_LF) eptr++;
5948                }                }
5949              else              else
5950                {                {
5951                if (c != 0x000a && (md->bsr_anycrlf ||                if (c != CHAR_LF && (md->bsr_anycrlf ||
5952                  (c != 0x000b && c != 0x000c && c != 0x0085                   (c != CHAR_VT && c != CHAR_FF && c != CHAR_NEL
5953  #ifdef COMPILE_PCRE16  #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
5954                  && c != 0x2028 && c != 0x2029                   && c != 0x2028 && c != 0x2029
5955  #endif  #endif
5956                  ))) break;                   ))) break;
5957                eptr++;                eptr++;
5958                }                }
5959              }              }
# Line 5870  for (;;) Line 5967  for (;;)
5967                SCHECK_PARTIAL();                SCHECK_PARTIAL();
5968                break;                break;
5969                }                }
5970              c = *eptr;              switch(*eptr)
5971              if (c == 0x09 || c == 0x20 || c == 0xa0                {
5972  #ifdef COMPILE_PCRE16                default: eptr++; break;
5973                || c == 0x1680 || c == 0x180e || (c >= 0x2000 && c <= 0x200A)                HSPACE_BYTE_CASES:
5974                || c == 0x202f || c == 0x205f || c == 0x3000  #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
5975                  HSPACE_MULTIBYTE_CASES:
5976  #endif  #endif
5977                ) break;                goto ENDLOOP00;
5978              eptr++;                }
5979              }              }
5980              ENDLOOP00:
5981            break;            break;
5982    
5983            case OP_HSPACE:            case OP_HSPACE:
# Line 5889  for (;;) Line 5988  for (;;)
5988                SCHECK_PARTIAL();                SCHECK_PARTIAL();
5989                break;                break;
5990                }                }
5991              c = *eptr;              switch(*eptr)
5992              if (c != 0x09 && c != 0x20 && c != 0xa0                {
5993  #ifdef COMPILE_PCRE16                default: goto ENDLOOP01;
5994                && c != 0x1680 && c != 0x180e && (c < 0x2000 || c > 0x200A)                HSPACE_BYTE_CASES:
5995                && c != 0x202f && c != 0x205f && c != 0x3000  #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
5996                  HSPACE_MULTIBYTE_CASES:
5997  #endif  #endif
5998                ) break;                eptr++; break;
5999              eptr++;                }
6000              }              }
6001              ENDLOOP01:
6002            break;            break;
6003    
6004            case OP_NOT_VSPACE:            case OP_NOT_VSPACE:
# Line 5908  for (;;) Line 6009  for (;;)
6009                SCHECK_PARTIAL();                SCHECK_PARTIAL();
6010                break;                break;
6011                }                }
6012              c = *eptr;              switch(*eptr)
6013              if (c == 0x0a || c == 0x0b || c == 0x0c || c == 0x0d || c == 0x85                {
6014  #ifdef COMPILE_PCRE16                default: eptr++; break;
6015                || c == 0x2028 || c == 0x2029                VSPACE_BYTE_CASES:
6016    #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
6017                  VSPACE_MULTIBYTE_CASES:
6018  #endif  #endif
6019                ) break;                goto ENDLOOP02;
6020              eptr++;                }
6021              }              }
6022              ENDLOOP02:
6023            break;            break;
6024    
6025            case OP_VSPACE:            case OP_VSPACE:
# Line 5926  for (;;) Line 6030  for (;;)
6030                SCHECK_PARTIAL();                SCHECK_PARTIAL();
6031                break;                break;
6032                }                }
6033              c = *eptr;              switch(*eptr)
6034              if (c != 0x0a && c != 0x0b && c != 0x0c && c != 0x0d && c != 0x85                {
6035  #ifdef COMPILE_PCRE16                default: goto ENDLOOP03;
6036                && c != 0x2028 && c != 0x2029                VSPACE_BYTE_CASES:
6037    #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
6038                  VSPACE_MULTIBYTE_CASES:
6039  #endif  #endif
6040                ) break;                eptr++; break;
6041              eptr++;                }
6042              }              }
6043              ENDLOOP03:
6044            break;            break;
6045    
6046            case OP_NOT_DIGIT:            case OP_NOT_DIGIT:
# Line 6030  for (;;) Line 6137  for (;;)
6137            RMATCH(eptr, ecode, offset_top, md, eptrb, RM47);            RMATCH(eptr, ecode, offset_top, md, eptrb, RM47);
6138            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
6139            eptr--;            eptr--;
6140            if (ctype == OP_ANYNL && eptr > pp  && *eptr == '\n' &&            if (ctype == OP_ANYNL && eptr > pp  && *eptr == CHAR_LF &&
6141                eptr[-1] == '\r') eptr--;                eptr[-1] == CHAR_CR) eptr--;
6142            }            }
6143          }          }
6144    
# Line 6081  switch (frame->Xwhere) Line 6188  switch (frame->Xwhere)
6188    LBL(32) LBL(34) LBL(42) LBL(46)    LBL(32) LBL(34) LBL(42) LBL(46)
6189  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
6190    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)
6191    LBL(59) LBL(60) LBL(61) LBL(62)    LBL(59) LBL(60) LBL(61) LBL(62) LBL(67) LBL(68)
6192  #endif  /* SUPPORT_UCP */  #endif  /* SUPPORT_UCP */
6193  #endif  /* SUPPORT_UTF */  #endif  /* SUPPORT_UTF */
6194    default:    default:
6195    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);  
   
6196    return PCRE_ERROR_INTERNAL;    return PCRE_ERROR_INTERNAL;
6197    }    }
6198  #undef LBL  #undef LBL
# Line 6150  Undefine all the macros that were define Line 6254  Undefine all the macros that were define
6254  ***************************************************************************/  ***************************************************************************/
6255    
6256    
6257    #ifdef NO_RECURSE
6258    /*************************************************
6259    *          Release allocated heap frames         *
6260    *************************************************/
6261    
6262    /* This function releases all the allocated frames. The base frame is on the
6263    machine stack, and so must not be freed.
6264    
6265    Argument: the address of the base frame
6266    Returns:  nothing
6267    */
6268    
6269    static void
6270    release_match_heapframes (heapframe *frame_base)
6271    {
6272    heapframe *nextframe = frame_base->Xnextframe;
6273    while (nextframe != NULL)
6274      {
6275      heapframe *oldframe = nextframe;
6276      nextframe = nextframe->Xnextframe;
6277      (PUBL(stack_free))(oldframe);
6278      }
6279    }
6280    #endif
6281    
6282    
6283  /*************************************************  /*************************************************
6284  *         Execute a Regular Expression           *  *         Execute a Regular Expression           *
# Line 6175  Returns:          > 0 => success; value Line 6304  Returns:          > 0 => success; value
6304                   < -1 => some kind of unexpected problem                   < -1 => some kind of unexpected problem
6305  */  */
6306    
6307  #ifdef COMPILE_PCRE8  #if defined COMPILE_PCRE8
6308  PCRE_EXP_DEFN int PCRE_CALL_CONVENTION  PCRE_EXP_DEFN int PCRE_CALL_CONVENTION
6309  pcre_exec(const pcre *argument_re, const pcre_extra *extra_data,  pcre_exec(const pcre *argument_re, const pcre_extra *extra_data,
6310    PCRE_SPTR subject, int length, int start_offset, int options, int *offsets,    PCRE_SPTR subject, int length, int start_offset, int options, int *offsets,
6311    int offsetcount)    int offsetcount)
6312  #else  #elif defined COMPILE_PCRE16
6313  PCRE_EXP_DEFN int PCRE_CALL_CONVENTION  PCRE_EXP_DEFN int PCRE_CALL_CONVENTION
6314  pcre16_exec(const pcre16 *argument_re, const pcre16_extra *extra_data,  pcre16_exec(const pcre16 *argument_re, const pcre16_extra *extra_data,
6315    PCRE_SPTR16 subject, int length, int start_offset, int options, int *offsets,    PCRE_SPTR16 subject, int length, int start_offset, int options, int *offsets,
6316    int offsetcount)    int offsetcount)
6317    #elif defined COMPILE_PCRE32
6318    PCRE_EXP_DEFN int PCRE_CALL_CONVENTION
6319    pcre32_exec(const pcre32 *argument_re, const pcre32_extra *extra_data,
6320      PCRE_SPTR32 subject, int length, int start_offset, int options, int *offsets,
6321      int offsetcount)
6322  #endif  #endif
6323  {  {
6324  int rc, ocount, arg_offset_max;  int rc, ocount, arg_offset_max;
# Line 6207  const pcre_uint8 *start_bits = NULL; Line 6341  const pcre_uint8 *start_bits = NULL;
6341  PCRE_PUCHAR start_match = (PCRE_PUCHAR)subject + start_offset;  PCRE_PUCHAR start_match = (PCRE_PUCHAR)subject + start_offset;
6342  PCRE_PUCHAR end_subject;  PCRE_PUCHAR end_subject;
6343  PCRE_PUCHAR start_partial = NULL;  PCRE_PUCHAR start_partial = NULL;
6344    PCRE_PUCHAR match_partial;
6345  PCRE_PUCHAR req_char_ptr = start_match - 1;  PCRE_PUCHAR req_char_ptr = start_match - 1;
6346    
6347  const pcre_study_data *study;  const pcre_study_data *study;
6348  const REAL_PCRE *re = (const REAL_PCRE *)argument_re;  const REAL_PCRE *re = (const REAL_PCRE *)argument_re;
6349    
6350  /* Check for the special magic call that measures the size of the stack used  #ifdef NO_RECURSE
6351  per recursive call of match(). */  heapframe frame_zero;
6352    frame_zero.Xprevframe = NULL;            /* Marks the top level */
6353    frame_zero.Xnextframe = NULL;            /* None are allocated yet */
6354    md->match_frames_base = &frame_zero;
6355    #endif
6356    
6357    /* Check for the special magic call that measures the size of the stack used
6358    per recursive call of match(). Without the funny casting for sizeof, a Windows
6359    compiler gave this error: "unary minus operator applied to unsigned type,
6360    result still unsigned". Hopefully the cast fixes that. */
6361    
6362  if (re == NULL && extra_data == NULL && subject == NULL && length == -1)  if (re == NULL && extra_data == NULL && subject == NULL && length == -999 &&
6363        start_offset == -999)
6364  #ifdef NO_RECURSE  #ifdef NO_RECURSE
6365    return -sizeof(heapframe);    return -((int)sizeof(heapframe));
6366  #else  #else
6367    return match((PCRE_PUCHAR)&start_partial, NULL, NULL, 0, NULL, NULL, 0);    return match(NULL, NULL, NULL, 0, NULL, NULL, 0);
6368  #endif  #endif
6369    
6370  /* Plausibility checks */  /* Plausibility checks */
6371    
6372  if ((options & ~PUBLIC_EXEC_OPTIONS) != 0) return PCRE_ERROR_BADOPTION;  if ((options & ~PUBLIC_EXEC_OPTIONS) != 0) return PCRE_ERROR_BADOPTION;
6373  if (re == NULL || subject == NULL || (offsets == NULL && offsetcount > 0))  if (re == NULL || subject == NULL || (offsets == NULL && offsetcount > 0))
6374    return PCRE_ERROR_NULL;    return PCRE_ERROR_NULL;
6375  if (offsetcount < 0) return PCRE_ERROR_BADCOUNT;  if (offsetcount < 0) return PCRE_ERROR_BADCOUNT;
6376    if (length < 0) return PCRE_ERROR_BADLENGTH;
6377  if (start_offset < 0 || start_offset > length) return PCRE_ERROR_BADOFFSET;  if (start_offset < 0 || start_offset > length) return PCRE_ERROR_BADOFFSET;
6378    
6379  /* 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 6265  if (utf && (options & PCRE_NO_UTF8_CHECK Line 6411  if (utf && (options & PCRE_NO_UTF8_CHECK
6411        offsets[0] = erroroffset;        offsets[0] = erroroffset;
6412        offsets[1] = errorcode;        offsets[1] = errorcode;
6413        }        }
6414  #ifdef COMPILE_PCRE16  #if defined COMPILE_PCRE8
     return (errorcode <= PCRE_UTF16_ERR1 && md->partial > 1)?  
       PCRE_ERROR_SHORTUTF16 : PCRE_ERROR_BADUTF16;  
 #else  
6415      return (errorcode <= PCRE_UTF8_ERR5 && md->partial > 1)?      return (errorcode <= PCRE_UTF8_ERR5 && md->partial > 1)?
6416        PCRE_ERROR_SHORTUTF8 : PCRE_ERROR_BADUTF8;        PCRE_ERROR_SHORTUTF8 : PCRE_ERROR_BADUTF8;
6417    #elif defined COMPILE_PCRE16
6418        return (errorcode <= PCRE_UTF16_ERR1 && md->partial > 1)?
6419          PCRE_ERROR_SHORTUTF16 : PCRE_ERROR_BADUTF16;
6420    #elif defined COMPILE_PCRE32
6421        return PCRE_ERROR_BADUTF32;
6422  #endif  #endif
6423      }      }
6424    #if defined COMPILE_PCRE8 || defined COMPILE_PCRE16
6425    /* 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. */
6426    if (start_offset > 0 && start_offset < length &&    if (start_offset > 0 && start_offset < length &&
6427        NOT_FIRSTCHAR(((PCRE_PUCHAR)subject)[start_offset]))        NOT_FIRSTCHAR(((PCRE_PUCHAR)subject)[start_offset]))
6428      return PCRE_ERROR_BADUTF8_OFFSET;      return PCRE_ERROR_BADUTF8_OFFSET;
6429    #endif
6430    }    }
6431  #endif  #endif
6432    
6433  /* If the pattern was successfully studied with JIT support, run the JIT  /* If the pattern was successfully studied with JIT support, run the JIT
6434  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
6435  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
6436  an unsupported flag is set. In particular, JIT does not support partial  an unsupported flag is set. */
 matching. */  
6437    
6438  #ifdef SUPPORT_JIT  #ifdef SUPPORT_JIT
6439  if (extra_data != NULL  if (extra_data != NULL
6440      && (extra_data->flags & PCRE_EXTRA_EXECUTABLE_JIT) != 0      && (extra_data->flags & (PCRE_EXTRA_EXECUTABLE_JIT |
6441                                 PCRE_EXTRA_TABLES)) == PCRE_EXTRA_EXECUTABLE_JIT
6442      && extra_data->executable_jit != NULL      && extra_data->executable_jit != NULL
6443      && (extra_data->flags & PCRE_EXTRA_TABLES) == 0      && (options & ~PUBLIC_JIT_EXEC_OPTIONS) == 0)
6444      && (options & ~(PCRE_NO_UTF8_CHECK | PCRE_NOTBOL | PCRE_NOTEOL |    {
6445                      PCRE_NOTEMPTY | PCRE_NOTEMPTY_ATSTART)) == 0)    rc = PRIV(jit_exec)(extra_data, (const pcre_uchar *)subject, length,
6446    return PRIV(jit_exec)(re, extra_data->executable_jit,         start_offset, options, offsets, offsetcount);
6447      (const pcre_uchar *)subject, length, start_offset, options,  
6448      ((extra_data->flags & PCRE_EXTRA_MATCH_LIMIT) == 0)    /* PCRE_ERROR_NULL means that the selected normal or partial matching
6449      ? MATCH_LIMIT : extra_data->match_limit, offsets, offsetcount);    mode is not compiled. In this case we simply fallback to interpreter. */
6450    
6451      if (rc != PCRE_ERROR_JIT_BADOPTION) return rc;
6452      }
6453  #endif  #endif
6454    
6455  /* 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 6358  end_subject = md->end_subject; Line 6510  end_subject = md->end_subject;
6510  md->endonly = (re->options & PCRE_DOLLAR_ENDONLY) != 0;  md->endonly = (re->options & PCRE_DOLLAR_ENDONLY) != 0;
6511  md->use_ucp = (re->options & PCRE_UCP) != 0;  md->use_ucp = (re->options & PCRE_UCP) != 0;
6512  md->jscript_compat = (re->options & PCRE_JAVASCRIPT_COMPAT) != 0;  md->jscript_compat = (re->options & PCRE_JAVASCRIPT_COMPAT) != 0;
6513  md->ignore_skip_arg = FALSE;  md->ignore_skip_arg = 0;
6514    
6515  /* Some options are unpacked into BOOL variables in the hope that testing  /* Some options are unpacked into BOOL variables in the hope that testing
6516  them will be faster than individual option bits. */  them will be faster than individual option bits. */
# Line 6468  if (re->top_backref > 0 && re->top_backr Line 6620  if (re->top_backref > 0 && re->top_backr
6620    DPRINTF(("Got memory to hold back references\n"));    DPRINTF(("Got memory to hold back references\n"));
6621    }    }
6622  else md->offset_vector = offsets;  else md->offset_vector = offsets;
   
6623  md->offset_end = ocount;  md->offset_end = ocount;
6624  md->offset_max = (2*ocount)/3;  md->offset_max = (2*ocount)/3;
6625  md->offset_overflow = FALSE;  md->capture_last = 0;
 md->capture_last = -1;  
6626    
6627  /* Reset the working variable associated with each extraction. These should  /* Reset the working variable associated with each extraction. These should
6628  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 6500  if (!anchored) Line 6650  if (!anchored)
6650    if ((re->flags & PCRE_FIRSTSET) != 0)    if ((re->flags & PCRE_FIRSTSET) != 0)
6651      {      {
6652      has_first_char = TRUE;      has_first_char = TRUE;
6653      first_char = first_char2 = re->first_char;      first_char = first_char2 = (pcre_uchar)(re->first_char);
6654      if ((re->flags & PCRE_FCH_CASELESS) != 0)      if ((re->flags & PCRE_FCH_CASELESS) != 0)
6655        {        {
6656        first_char2 = TABLE_GET(first_char, md->fcc, first_char);        first_char2 = TABLE_GET(first_char, md->fcc, first_char);
# Line 6522  character" set. */ Line 6672  character" set. */
6672  if ((re->flags & PCRE_REQCHSET) != 0)  if ((re->flags & PCRE_REQCHSET) != 0)
6673    {    {
6674    has_req_char = TRUE;    has_req_char = TRUE;
6675    req_char = req_char2 = re->req_char;    req_char = req_char2 = (pcre_uchar)(re->req_char);
6676    if ((re->flags & PCRE_RCH_CASELESS) != 0)    if ((re->flags & PCRE_RCH_CASELESS) != 0)
6677      {      {
6678      req_char2 = TABLE_GET(req_char, md->fcc, req_char);      req_char2 = TABLE_GET(req_char, md->fcc, req_char);
# Line 6580  for(;;) Line 6730  for(;;)
6730    
6731      if (has_first_char)      if (has_first_char)
6732        {        {
6733          pcre_uchar smc;
6734    
6735        if (first_char != first_char2)        if (first_char != first_char2)
6736          while (start_match < end_subject &&          while (start_match < end_subject &&
6737              *start_match != first_char && *start_match != first_char2)            (smc = RAWUCHARTEST(start_match)) != first_char && smc != first_char2)
6738            start_match++;            start_match++;
6739        else        else
6740          while (start_match < end_subject && *start_match != first_char)          while (start_match < end_subject && RAWUCHARTEST(start_match) != first_char)
6741            start_match++;            start_match++;
6742        }        }
6743    
# Line 6617  for(;;) Line 6769  for(;;)
6769          if (start_match[-1] == CHAR_CR &&          if (start_match[-1] == CHAR_CR &&
6770               (md->nltype == NLTYPE_ANY || md->nltype == NLTYPE_ANYCRLF) &&               (md->nltype == NLTYPE_ANY || md->nltype == NLTYPE_ANYCRLF) &&
6771               start_match < end_subject &&               start_match < end_subject &&
6772               *start_match == CHAR_NL)               RAWUCHARTEST(start_match) == CHAR_NL)
6773            start_match++;            start_match++;
6774          }          }
6775        }        }
# Line 6628  for(;;) Line 6780  for(;;)
6780        {        {
6781        while (start_match < end_subject)        while (start_match < end_subject)
6782          {          {
6783          register unsigned int c = *start_match;          register pcre_uint32 c = RAWUCHARTEST(start_match);
6784  #ifndef COMPILE_PCRE8  #ifndef COMPILE_PCRE8
6785          if (c > 255) c = 255;          if (c > 255) c = 255;
6786  #endif  #endif
# Line 6696  for(;;) Line 6848  for(;;)
6848            {            {
6849            while (p < end_subject)            while (p < end_subject)
6850              {              {
6851              register int pp = *p++;              register pcre_uint32 pp = RAWUCHARINCTEST(p);
6852              if (pp == req_char || pp == req_char2) { p--; break; }              if (pp == req_char || pp == req_char2) { p--; break; }
6853              }              }
6854            }            }
# Line 6704  for(;;) Line 6856  for(;;)
6856            {            {
6857            while (p < end_subject)            while (p < end_subject)
6858              {              {
6859              if (*p++ == req_char) { p--; break; }              if (RAWUCHARINCTEST(p) == req_char) { p--; break; }
6860              }              }
6861            }            }
6862    
# Line 6740  for(;;) Line 6892  for(;;)
6892    md->match_call_count = 0;    md->match_call_count = 0;
6893    md->match_function_type = 0;    md->match_function_type = 0;
6894    md->end_offset_top = 0;    md->end_offset_top = 0;
6895      md->skip_arg_count = 0;
6896    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);
6897    if (md->hitend && start_partial == NULL) start_partial = md->start_used_ptr;    if (md->hitend && start_partial == NULL)
6898        {
6899        start_partial = md->start_used_ptr;
6900        match_partial = start_match;
6901        }
6902    
6903    switch(rc)    switch(rc)
6904      {      {
# Line 6754  for(;;) Line 6911  for(;;)
6911    
6912      case MATCH_SKIP_ARG:      case MATCH_SKIP_ARG:
6913      new_start_match = start_match;      new_start_match = start_match;
6914      md->ignore_skip_arg = TRUE;      md->ignore_skip_arg = md->skip_arg_count;
6915      break;      break;
6916    
6917      /* 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
6918      same as the match we have just done, treat it as NOMATCH. */      greater than the match we have just done, treat it as NOMATCH. */
6919    
6920      case MATCH_SKIP:      case MATCH_SKIP:
6921      if (md->start_match_ptr != start_match)      if (md->start_match_ptr > start_match)
6922        {        {
6923        new_start_match = md->start_match_ptr;        new_start_match = md->start_match_ptr;
6924        break;        break;
# Line 6769  for(;;) Line 6926  for(;;)
6926      /* Fall through */      /* Fall through */
6927    
6928      /* NOMATCH and PRUNE advance by one character. THEN at this level acts      /* NOMATCH and PRUNE advance by one character. THEN at this level acts
6929      exactly like PRUNE. Unset the ignore SKIP-with-argument flag. */      exactly like PRUNE. Unset ignore SKIP-with-argument. */
6930    
6931      case MATCH_NOMATCH:      case MATCH_NOMATCH:
6932      case MATCH_PRUNE:      case MATCH_PRUNE:
6933      case MATCH_THEN:      case MATCH_THEN:
6934      md->ignore_skip_arg = FALSE;      md->ignore_skip_arg = 0;
6935      new_start_match = start_match + 1;      new_start_match = start_match + 1;
6936  #ifdef SUPPORT_UTF  #ifdef SUPPORT_UTF
6937      if (utf)      if (utf)
# Line 6867  if (rc == MATCH_MATCH || rc == MATCH_ACC Line 7024  if (rc == MATCH_MATCH || rc == MATCH_ACC
7024          (arg_offset_max - 2) * sizeof(int));          (arg_offset_max - 2) * sizeof(int));
7025        DPRINTF(("Copied offsets from temporary memory\n"));        DPRINTF(("Copied offsets from temporary memory\n"));
7026        }        }
7027      if (md->end_offset_top > arg_offset_max) md->offset_overflow = TRUE;      if (md->end_offset_top > arg_offset_max) md->capture_last |= OVFLBIT;
7028      DPRINTF(("Freeing temporary memory\n"));      DPRINTF(("Freeing temporary memory\n"));
7029      (PUBL(free))(md->offset_vector);      (PUBL(free))(md->offset_vector);
7030      }      }
# Line 6875  if (rc == MATCH_MATCH || rc == MATCH_ACC Line 7032  if (rc == MATCH_MATCH || rc == MATCH_ACC
7032    /* 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
7033    too many to fit into the vector. */    too many to fit into the vector. */
7034    
7035    rc = (md->offset_overflow && md->end_offset_top >= arg_offset_max)?    rc = ((md->capture_last & OVFLBIT) != 0 &&
7036             md->end_offset_top >= arg_offset_max)?
7037      0 : md->end_offset_top/2;      0 : md->end_offset_top/2;
7038    
7039    /* 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 6891  if (rc == MATCH_MATCH || rc == MATCH_ACC Line 7049  if (rc == MATCH_MATCH || rc == MATCH_ACC
7049      {      {
7050      register int *iptr, *iend;      register int *iptr, *iend;
7051      int resetcount = 2 + re->top_bracket * 2;      int resetcount = 2 + re->top_bracket * 2;
7052      if (resetcount > offsetcount) resetcount = ocount;      if (resetcount > offsetcount) resetcount = offsetcount;
7053      iptr = offsets + md->end_offset_top;      iptr = offsets + md->end_offset_top;
7054      iend = offsets + resetcount;      iend = offsets + resetcount;
7055      while (iptr < iend) *iptr++ = -1;      while (iptr < iend) *iptr++ = -1;
# Line 6912  if (rc == MATCH_MATCH || rc == MATCH_ACC Line 7070  if (rc == MATCH_MATCH || rc == MATCH_ACC
7070    if (extra_data != NULL && (extra_data->flags & PCRE_EXTRA_MARK) != 0)    if (extra_data != NULL && (extra_data->flags & PCRE_EXTRA_MARK) != 0)
7071      *(extra_data->mark) = (pcre_uchar *)md->mark;      *(extra_data->mark) = (pcre_uchar *)md->mark;
7072    DPRINTF((">>>> returning %d\n", rc));    DPRINTF((">>>> returning %d\n", rc));
7073    #ifdef NO_RECURSE
7074      release_match_heapframes(&frame_zero);
7075    #endif
7076    return rc;    return rc;
7077    }    }
7078    
# Line 6929  if (using_temporary_offsets) Line 7090  if (using_temporary_offsets)
7090  if (rc != MATCH_NOMATCH && rc != PCRE_ERROR_PARTIAL)  if (rc != MATCH_NOMATCH && rc != PCRE_ERROR_PARTIAL)
7091    {    {
7092    DPRINTF((">>>> error: returning %d\n", rc));    DPRINTF((">>>> error: returning %d\n", rc));
7093    #ifdef NO_RECURSE
7094      release_match_heapframes(&frame_zero);
7095    #endif
7096    return rc;    return rc;
7097    }    }
7098    
# Line 6942  if (start_partial != NULL) Line 7106  if (start_partial != NULL)
7106      {      {
7107      offsets[0] = (int)(start_partial - (PCRE_PUCHAR)subject);      offsets[0] = (int)(start_partial - (PCRE_PUCHAR)subject);
7108      offsets[1] = (int)(end_subject - (PCRE_PUCHAR)subject);      offsets[1] = (int)(end_subject - (PCRE_PUCHAR)subject);
7109        if (offsetcount > 2)
7110          offsets[2] = (int)(match_partial - (PCRE_PUCHAR)subject);
7111      }      }
7112    rc = PCRE_ERROR_PARTIAL;    rc = PCRE_ERROR_PARTIAL;
7113    }    }
# Line 6958  else Line 7124  else
7124    
7125  if (extra_data != NULL && (extra_data->flags & PCRE_EXTRA_MARK) != 0)  if (extra_data != NULL && (extra_data->flags & PCRE_EXTRA_MARK) != 0)
7126    *(extra_data->mark) = (pcre_uchar *)md->nomatch_mark;    *(extra_data->mark) = (pcre_uchar *)md->nomatch_mark;
7127    #ifdef NO_RECURSE
7128      release_match_heapframes(&frame_zero);
7129    #endif
7130  return rc;  return rc;
7131  }  }
7132    

Legend:
Removed from v.893  
changed lines
  Added in v.1284

  ViewVC Help
Powered by ViewVC 1.1.5