/[pcre]/code/trunk/pcre_exec.c
ViewVC logotype

Diff of /code/trunk/pcre_exec.c

Parent Directory Parent Directory | Revision Log Revision Log | View Patch Patch

revision 895 by ph10, Fri Jan 20 12:12:03 2012 UTC revision 1296 by ph10, Tue Mar 19 16:29:12 2013 UTC
# Line 6  Line 6 
6  and semantics are as close as possible to those of the Perl 5 language.  and semantics are as close as possible to those of the Perl 5 language.
7    
8                         Written by Philip Hazel                         Written by Philip Hazel
9             Copyright (c) 1997-2012 University of Cambridge             Copyright (c) 1997-2013 University of Cambridge
10    
11  -----------------------------------------------------------------------------  -----------------------------------------------------------------------------
12  Redistribution and use in source and binary forms, with or without  Redistribution and use in source and binary forms, with or without
# Line 37  POSSIBILITY OF SUCH DAMAGE. Line 37  POSSIBILITY OF SUCH DAMAGE.
37  -----------------------------------------------------------------------------  -----------------------------------------------------------------------------
38  */  */
39    
   
40  /* This module contains pcre_exec(), the externally visible function that does  /* This module contains pcre_exec(), the externally visible function that does
41  pattern matching using an NFA algorithm, trying to mimic Perl as closely as  pattern matching using an NFA algorithm, trying to mimic Perl as closely as
42  possible. There are also some static supporting functions. */  possible. There are also some static supporting functions. */
# Line 57  possible. There are also some static sup Line 56  possible. There are also some static sup
56  #undef min  #undef min
57  #undef max  #undef max
58    
59    /* The md->capture_last field uses the lower 16 bits for the last captured
60    substring (which can never be greater than 65535) and a bit in the top half
61    to mean "capture vector overflowed". This odd way of doing things was
62    implemented when it was realized that preserving and restoring the overflow bit
63    whenever the last capture number was saved/restored made for a neater
64    interface, and doing it this way saved on (a) another variable, which would
65    have increased the stack frame size (a big NO-NO in PCRE) and (b) another
66    separate set of save/restore instructions. The following defines are used in
67    implementing this. */
68    
69    #define CAPLMASK    0x0000ffff    /* The bits used for last_capture */
70    #define OVFLMASK    0xffff0000    /* The bits used for the overflow flag */
71    #define OVFLBIT     0x00010000    /* The bit that is set for overflow */
72    
73  /* Values for setting in md->match_function_type to indicate two special types  /* Values for setting in md->match_function_type to indicate two special types
74  of call to match(). We do it this way to save on using another stack variable,  of call to match(). We do it this way to save on using another stack variable,
75  as stack usage is to be discouraged. */  as stack usage is to be discouraged. */
# Line 74  defined PCRE_ERROR_xxx codes, which are Line 87  defined PCRE_ERROR_xxx codes, which are
87  negative to avoid the external error codes. */  negative to avoid the external error codes. */
88    
89  #define MATCH_ACCEPT       (-999)  #define MATCH_ACCEPT       (-999)
90  #define MATCH_COMMIT       (-998)  #define MATCH_KETRPOS      (-998)
91  #define MATCH_KETRPOS      (-997)  #define MATCH_ONCE         (-997)
92  #define MATCH_ONCE         (-996)  /* The next 5 must be kept together and in sequence so that a test that checks
93    for any one of them can use a range. */
94    #define MATCH_COMMIT       (-996)
95  #define MATCH_PRUNE        (-995)  #define MATCH_PRUNE        (-995)
96  #define MATCH_SKIP         (-994)  #define MATCH_SKIP         (-994)
97  #define MATCH_SKIP_ARG     (-993)  #define MATCH_SKIP_ARG     (-993)
98  #define MATCH_THEN         (-992)  #define MATCH_THEN         (-992)
99    #define MATCH_BACKTRACK_MAX MATCH_THEN
100    #define MATCH_BACKTRACK_MIN MATCH_COMMIT
101    
102  /* Maximum number of ints of offset to save on the stack for recursive calls.  /* Maximum number of ints of offset to save on the stack for recursive calls.
103  If the offset vector is bigger, malloc is used. This should be a multiple of 3,  If the offset vector is bigger, malloc is used. This should be a multiple of 3,
# Line 93  because the offset vector is always a mu Line 110  because the offset vector is always a mu
110  static const char rep_min[] = { 0, 0, 1, 1, 0, 0 };  static const char rep_min[] = { 0, 0, 1, 1, 0, 0 };
111  static const char rep_max[] = { 0, 0, 0, 0, 1, 1 };  static const char rep_max[] = { 0, 0, 0, 0, 1, 1 };
112    
   
   
113  #ifdef PCRE_DEBUG  #ifdef PCRE_DEBUG
114  /*************************************************  /*************************************************
115  *        Debugging function to print chars       *  *        Debugging function to print chars       *
# Line 115  Returns:     nothing Line 130  Returns:     nothing
130  static void  static void
131  pchars(const pcre_uchar *p, int length, BOOL is_subject, match_data *md)  pchars(const pcre_uchar *p, int length, BOOL is_subject, match_data *md)
132  {  {
133  unsigned int c;  pcre_uint32 c;
134    BOOL utf = md->utf;
135  if (is_subject && length > md->end_subject - p) length = md->end_subject - p;  if (is_subject && length > md->end_subject - p) length = md->end_subject - p;
136  while (length-- > 0)  while (length-- > 0)
137    if (isprint(c = *(p++))) printf("%c", c); else printf("\\x%02x", c);    if (isprint(c = RAWUCHARINCTEST(p))) printf("%c", (char)c); else printf("\\x{%02x}", c);
138  }  }
139  #endif  #endif
140    
# Line 140  Arguments: Line 156  Arguments:
156    md          points to match data block    md          points to match data block
157    caseless    TRUE if caseless    caseless    TRUE if caseless
158    
159  Returns:      < 0 if not matched, otherwise the number of subject bytes matched  Returns:      >= 0 the number of subject bytes matched
160                  -1 no match
161                  -2 partial match; always given if at end subject
162  */  */
163    
164  static int  static int
# Line 149  match_ref(int offset, register PCRE_PUCH Line 167  match_ref(int offset, register PCRE_PUCH
167  {  {
168  PCRE_PUCHAR eptr_start = eptr;  PCRE_PUCHAR eptr_start = eptr;
169  register PCRE_PUCHAR p = md->start_subject + md->offset_vector[offset];  register PCRE_PUCHAR p = md->start_subject + md->offset_vector[offset];
170    #ifdef SUPPORT_UTF
171    BOOL utf = md->utf;
172    #endif
173    
174  #ifdef PCRE_DEBUG  #ifdef PCRE_DEBUG
175  if (eptr >= md->end_subject)  if (eptr >= md->end_subject)
# Line 163  pchars(p, length, FALSE, md); Line 184  pchars(p, length, FALSE, md);
184  printf("\n");  printf("\n");
185  #endif  #endif
186    
187  /* Always fail if reference not set (and not JavaScript compatible). */  /* Always fail if reference not set (and not JavaScript compatible - in that
188    case the length is passed as zero). */
189    
190  if (length < 0) return -1;  if (length < 0) return -1;
191    
# Line 175  if (caseless) Line 197  if (caseless)
197    {    {
198  #ifdef SUPPORT_UTF  #ifdef SUPPORT_UTF
199  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
200    if (md->utf)    if (utf)
201      {      {
202      /* Match characters up to the end of the reference. NOTE: the number of      /* Match characters up to the end of the reference. NOTE: the number of
203      bytes matched may differ, because there are some characters whose upper and      data units matched may differ, because in UTF-8 there are some characters
204      lower case versions code as different numbers of bytes. For example, U+023A      whose upper and lower case versions code have different numbers of bytes.
205      (2 bytes in UTF-8) is the upper case version of U+2C65 (3 bytes in UTF-8);      For example, U+023A (2 bytes in UTF-8) is the upper case version of U+2C65
206      a sequence of 3 of the former uses 6 bytes, as does a sequence of two of      (3 bytes in UTF-8); a sequence of 3 of the former uses 6 bytes, as does a
207      the latter. It is important, therefore, to check the length along the      sequence of two of the latter. It is important, therefore, to check the
208      reference, not along the subject (earlier code did this wrong). */      length along the reference, not along the subject (earlier code did this
209        wrong). */
210    
211      PCRE_PUCHAR endptr = p + length;      PCRE_PUCHAR endptr = p + length;
212      while (p < endptr)      while (p < endptr)
213        {        {
214        int c, d;        pcre_uint32 c, d;
215        if (eptr >= md->end_subject) return -1;        const ucd_record *ur;
216          if (eptr >= md->end_subject) return -2;   /* Partial match */
217        GETCHARINC(c, eptr);        GETCHARINC(c, eptr);
218        GETCHARINC(d, p);        GETCHARINC(d, p);
219        if (c != d && c != UCD_OTHERCASE(d)) return -1;        ur = GET_UCD(d);
220          if (c != d && c != d + ur->other_case)
221            {
222            const pcre_uint32 *pp = PRIV(ucd_caseless_sets) + ur->caseset;
223            for (;;)
224              {
225              if (c < *pp) return -1;
226              if (c == *pp++) break;
227              }
228            }
229        }        }
230      }      }
231    else    else
# Line 202  if (caseless) Line 235  if (caseless)
235    /* The same code works when not in UTF-8 mode and in UTF-8 mode when there    /* The same code works when not in UTF-8 mode and in UTF-8 mode when there
236    is no UCP support. */    is no UCP support. */
237      {      {
     if (eptr + length > md->end_subject) return -1;  
238      while (length-- > 0)      while (length-- > 0)
239        {        {
240        if (TABLE_GET(*p, md->lcc, *p) != TABLE_GET(*eptr, md->lcc, *eptr)) return -1;        pcre_uint32 cc, cp;
241          if (eptr >= md->end_subject) return -2;   /* Partial match */
242          cc = RAWUCHARTEST(eptr);
243          cp = RAWUCHARTEST(p);
244          if (TABLE_GET(cp, md->lcc, cp) != TABLE_GET(cc, md->lcc, cc)) return -1;
245        p++;        p++;
246        eptr++;        eptr++;
247        }        }
# Line 217  are in UTF-8 mode. */ Line 253  are in UTF-8 mode. */
253    
254  else  else
255    {    {
256    if (eptr + length > md->end_subject) return -1;    while (length-- > 0)
257    while (length-- > 0) if (*p++ != *eptr++) return -1;      {
258        if (eptr >= md->end_subject) return -2;   /* Partial match */
259        if (RAWUCHARINCTEST(p) != RAWUCHARINCTEST(eptr)) return -1;
260        }
261    }    }
262    
263  return (int)(eptr - eptr_start);  return (int)(eptr - eptr_start);
# Line 273  enum { RM1=1, RM2,  RM3,  RM4,  RM5,  RM Line 312  enum { RM1=1, RM2,  RM3,  RM4,  RM5,  RM
312         RM31,  RM32, RM33, RM34, RM35, RM36, RM37, RM38, RM39, RM40,         RM31,  RM32, RM33, RM34, RM35, RM36, RM37, RM38, RM39, RM40,
313         RM41,  RM42, RM43, RM44, RM45, RM46, RM47, RM48, RM49, RM50,         RM41,  RM42, RM43, RM44, RM45, RM46, RM47, RM48, RM49, RM50,
314         RM51,  RM52, RM53, RM54, RM55, RM56, RM57, RM58, RM59, RM60,         RM51,  RM52, RM53, RM54, RM55, RM56, RM57, RM58, RM59, RM60,
315         RM61,  RM62, RM63, RM64, RM65, RM66 };         RM61,  RM62, RM63, RM64, RM65, RM66, RM67, RM68 };
316    
317  /* These versions of the macros use the stack, as normal. There are debugging  /* These versions of the macros use the stack, as normal. There are debugging
318  versions and production versions. Note that the "rw" argument of RMATCH isn't  versions and production versions. Note that the "rw" argument of RMATCH isn't
# Line 291  actually used in this definition. */ Line 330  actually used in this definition. */
330    }    }
331  #define RRETURN(ra) \  #define RRETURN(ra) \
332    { \    { \
333    printf("match() returned %d from line %d ", ra, __LINE__); \    printf("match() returned %d from line %d\n", ra, __LINE__); \
334    return ra; \    return ra; \
335    }    }
336  #else  #else
# Line 311  argument of match(), which never changes Line 350  argument of match(), which never changes
350    
351  #define RMATCH(ra,rb,rc,rd,re,rw)\  #define RMATCH(ra,rb,rc,rd,re,rw)\
352    {\    {\
353    heapframe *newframe = (heapframe *)(PUBL(stack_malloc))(sizeof(heapframe));\    heapframe *newframe = frame->Xnextframe;\
354    if (newframe == NULL) RRETURN(PCRE_ERROR_NOMEMORY);\    if (newframe == NULL)\
355    frame->Xwhere = rw; \      {\
356        newframe = (heapframe *)(PUBL(stack_malloc))(sizeof(heapframe));\
357        if (newframe == NULL) RRETURN(PCRE_ERROR_NOMEMORY);\
358        newframe->Xnextframe = NULL;\
359        frame->Xnextframe = newframe;\
360        }\
361      frame->Xwhere = rw;\
362    newframe->Xeptr = ra;\    newframe->Xeptr = ra;\
363    newframe->Xecode = rb;\    newframe->Xecode = rb;\
364    newframe->Xmstart = mstart;\    newframe->Xmstart = mstart;\
# Line 332  argument of match(), which never changes Line 377  argument of match(), which never changes
377    {\    {\
378    heapframe *oldframe = frame;\    heapframe *oldframe = frame;\
379    frame = oldframe->Xprevframe;\    frame = oldframe->Xprevframe;\
   if (oldframe != &frame_zero) (PUBL(stack_free))(oldframe);\  
380    if (frame != NULL)\    if (frame != NULL)\
381      {\      {\
382      rrc = ra;\      rrc = ra;\
# Line 346  argument of match(), which never changes Line 390  argument of match(), which never changes
390    
391  typedef struct heapframe {  typedef struct heapframe {
392    struct heapframe *Xprevframe;    struct heapframe *Xprevframe;
393      struct heapframe *Xnextframe;
394    
395    /* Function arguments that may change */    /* Function arguments that may change */
396    
# Line 376  typedef struct heapframe { Line 421  typedef struct heapframe {
421    
422  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
423    int Xprop_type;    int Xprop_type;
424    int Xprop_value;    unsigned int Xprop_value;
425    int Xprop_fail_result;    int Xprop_fail_result;
426    int Xoclength;    int Xoclength;
427    pcre_uchar Xocchars[6];    pcre_uchar Xocchars[6];
# Line 389  typedef struct heapframe { Line 434  typedef struct heapframe {
434    int Xlength;    int Xlength;
435    int Xmax;    int Xmax;
436    int Xmin;    int Xmin;
437    int Xnumber;    unsigned int Xnumber;
438    int Xoffset;    int Xoffset;
439    int Xop;    unsigned int Xop;
440    int Xsave_capture_last;    pcre_int32 Xsave_capture_last;
441    int Xsave_offset1, Xsave_offset2, Xsave_offset3;    int Xsave_offset1, Xsave_offset2, Xsave_offset3;
442    int Xstacksave[REC_STACK_SAVE_MAX];    int Xstacksave[REC_STACK_SAVE_MAX];
443    
# Line 477  so they can be ordinary variables in all Line 522  so they can be ordinary variables in all
522    
523  register int  rrc;         /* Returns from recursive calls */  register int  rrc;         /* Returns from recursive calls */
524  register int  i;           /* Used for loops not involving calls to RMATCH() */  register int  i;           /* Used for loops not involving calls to RMATCH() */
525  register unsigned int c;   /* Character values not kept over RMATCH() calls */  register pcre_uint32 c;    /* Character values not kept over RMATCH() calls */
526  register BOOL utf;         /* Local copy of UTF flag for speed */  register BOOL utf;         /* Local copy of UTF flag for speed */
527    
528  BOOL minimize, possessive; /* Quantifier options */  BOOL minimize, possessive; /* Quantifier options */
# Line 487  int condcode; Line 532  int condcode;
532  /* When recursion is not being used, all "local" variables that have to be  /* When recursion is not being used, all "local" variables that have to be
533  preserved over calls to RMATCH() are part of a "frame". We set up the top-level  preserved over calls to RMATCH() are part of a "frame". We set up the top-level
534  frame on the stack here; subsequent instantiations are obtained from the heap  frame on the stack here; subsequent instantiations are obtained from the heap
535  whenever RMATCH() does a "recursion". See the macro definitions above. Putting  whenever RMATCH() does a "recursion". See the macro definitions above. Putting
536  the top-level on the stack rather than malloc-ing them all gives a performance  the top-level on the stack rather than malloc-ing them all gives a performance
537  boost in many cases where there is not much "recursion". */  boost in many cases where there is not much "recursion". */
538    
539  #ifdef NO_RECURSE  #ifdef NO_RECURSE
540  heapframe frame_zero;  heapframe *frame = (heapframe *)md->match_frames_base;
 heapframe *frame = &frame_zero;  
 frame->Xprevframe = NULL;            /* Marks the top level */  
541    
542  /* Copy in the original argument variables */  /* Copy in the original argument variables */
543    
# Line 596  BOOL prev_is_word; Line 639  BOOL prev_is_word;
639    
640  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
641  int prop_type;  int prop_type;
642  int prop_value;  unsigned int prop_value;
643  int prop_fail_result;  int prop_fail_result;
644  int oclength;  int oclength;
645  pcre_uchar occhars[6];  pcre_uchar occhars[6];
# Line 607  int ctype; Line 650  int ctype;
650  int length;  int length;
651  int max;  int max;
652  int min;  int min;
653  int number;  unsigned int number;
654  int offset;  int offset;
655  int op;  unsigned int op;
656  int save_capture_last;  pcre_int32 save_capture_last;
657  int save_offset1, save_offset2, save_offset3;  int save_offset1, save_offset2, save_offset3;
658  int stacksave[REC_STACK_SAVE_MAX];  int stacksave[REC_STACK_SAVE_MAX];
659    
660  eptrblock newptrb;  eptrblock newptrb;
661    
662  /* There is a special fudge for calling match() in a way that causes it to  /* There is a special fudge for calling match() in a way that causes it to
663  measure the size of its basic stack frame when the stack is being used for  measure the size of its basic stack frame when the stack is being used for
664  recursion. The second argument (ecode) being NULL triggers this behaviour. It  recursion. The second argument (ecode) being NULL triggers this behaviour. It
665  cannot normally every be NULL. The return is the negated value of the frame  cannot normally ever be NULL. The return is the negated value of the frame
666  size. */  size. */
667    
668  if (ecode == NULL)  if (ecode == NULL)
# Line 631  if (ecode == NULL) Line 674  if (ecode == NULL)
674      int len = (char *)&rdepth - (char *)eptr;      int len = (char *)&rdepth - (char *)eptr;
675      return (len > 0)? -len : len;      return (len > 0)? -len : len;
676      }      }
677    }    }
678  #endif     /* NO_RECURSE */  #endif     /* NO_RECURSE */
679    
680  /* To save space on the stack and in the heap frame, I have doubled up on some  /* To save space on the stack and in the heap frame, I have doubled up on some
# Line 728  for (;;) Line 771  for (;;)
771      unaltered. */      unaltered. */
772    
773      else if (rrc == MATCH_SKIP_ARG &&      else if (rrc == MATCH_SKIP_ARG &&
774          STRCMP_UC_UC(ecode + 2, md->start_match_ptr) == 0)          STRCMP_UC_UC_TEST(ecode + 2, md->start_match_ptr) == 0)
775        {        {
776        md->start_match_ptr = eptr;        md->start_match_ptr = eptr;
777        RRETURN(MATCH_SKIP);        RRETURN(MATCH_SKIP);
# Line 738  for (;;) Line 781  for (;;)
781      case OP_FAIL:      case OP_FAIL:
782      RRETURN(MATCH_NOMATCH);      RRETURN(MATCH_NOMATCH);
783    
     /* COMMIT overrides PRUNE, SKIP, and THEN */  
   
784      case OP_COMMIT:      case OP_COMMIT:
785      RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md,      RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md,
786        eptrb, RM52);        eptrb, RM52);
787      if (rrc != MATCH_NOMATCH && rrc != MATCH_PRUNE &&      if (rrc != MATCH_NOMATCH) RRETURN(rrc);
         rrc != MATCH_SKIP && rrc != MATCH_SKIP_ARG &&  
         rrc != MATCH_THEN)  
       RRETURN(rrc);  
788      RRETURN(MATCH_COMMIT);      RRETURN(MATCH_COMMIT);
789    
     /* PRUNE overrides THEN */  
   
790      case OP_PRUNE:      case OP_PRUNE:
791      RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md,      RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md,
792        eptrb, RM51);        eptrb, RM51);
793      if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN) RRETURN(rrc);      if (rrc != MATCH_NOMATCH) RRETURN(rrc);
794      RRETURN(MATCH_PRUNE);      RRETURN(MATCH_PRUNE);
795    
796      case OP_PRUNE_ARG:      case OP_PRUNE_ARG:
# Line 764  for (;;) Line 800  for (;;)
800        eptrb, RM56);        eptrb, RM56);
801      if ((rrc == MATCH_MATCH || rrc == MATCH_ACCEPT) &&      if ((rrc == MATCH_MATCH || rrc == MATCH_ACCEPT) &&
802           md->mark == NULL) md->mark = ecode + 2;           md->mark == NULL) md->mark = ecode + 2;
803      if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN) RRETURN(rrc);      if (rrc != MATCH_NOMATCH) RRETURN(rrc);
804      RRETURN(MATCH_PRUNE);      RRETURN(MATCH_PRUNE);
805    
     /* SKIP overrides PRUNE and THEN */  
   
806      case OP_SKIP:      case OP_SKIP:
807      RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md,      RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md,
808        eptrb, RM53);        eptrb, RM53);
809      if (rrc != MATCH_NOMATCH && rrc != MATCH_PRUNE && rrc != MATCH_THEN)      if (rrc != MATCH_NOMATCH) RRETURN(rrc);
       RRETURN(rrc);  
810      md->start_match_ptr = eptr;   /* Pass back current position */      md->start_match_ptr = eptr;   /* Pass back current position */
811      RRETURN(MATCH_SKIP);      RRETURN(MATCH_SKIP);
812    
813      /* Note that, for Perl compatibility, SKIP with an argument does NOT set      /* Note that, for Perl compatibility, SKIP with an argument does NOT set
814      nomatch_mark. There is a flag that disables this opcode when re-matching a      nomatch_mark. When a pattern match ends with a SKIP_ARG for which there was
815      pattern that ended with a SKIP for which there was not a matching MARK. */      not a matching mark, we have to re-run the match, ignoring the SKIP_ARG
816        that failed and any that preceed it (either they also failed, or were not
817        triggered). To do this, we maintain a count of executed SKIP_ARGs. If a
818        SKIP_ARG gets to top level, the match is re-run with md->ignore_skip_arg
819        set to the count of the one that failed. */
820    
821      case OP_SKIP_ARG:      case OP_SKIP_ARG:
822      if (md->ignore_skip_arg)      md->skip_arg_count++;
823        if (md->skip_arg_count <= md->ignore_skip_arg)
824        {        {
825        ecode += PRIV(OP_lengths)[*ecode] + ecode[1];        ecode += PRIV(OP_lengths)[*ecode] + ecode[1];
826        break;        break;
827        }        }
828      RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode] + ecode[1], offset_top, md,      RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode] + ecode[1], offset_top, md,
829        eptrb, RM57);        eptrb, RM57);
830      if (rrc != MATCH_NOMATCH && rrc != MATCH_PRUNE && rrc != MATCH_THEN)      if (rrc != MATCH_NOMATCH) RRETURN(rrc);
831        RRETURN(rrc);  
   
832      /* Pass back the current skip name by overloading md->start_match_ptr and      /* Pass back the current skip name by overloading md->start_match_ptr and
833      returning the special MATCH_SKIP_ARG return code. This will either be      returning the special MATCH_SKIP_ARG return code. This will either be
834      caught by a matching MARK, or get to the top, where it causes a rematch      caught by a matching MARK, or get to the top, where it causes a rematch
835      with the md->ignore_skip_arg flag set. */      with md->ignore_skip_arg set to the value of md->skip_arg_count. */
836    
837      md->start_match_ptr = ecode + 2;      md->start_match_ptr = ecode + 2;
838      RRETURN(MATCH_SKIP_ARG);      RRETURN(MATCH_SKIP_ARG);
# Line 838  for (;;) Line 875  for (;;)
875      case OP_ONCE_NC:      case OP_ONCE_NC:
876      prev = ecode;      prev = ecode;
877      saved_eptr = eptr;      saved_eptr = eptr;
878      save_mark = md->mark;      save_mark = md->mark;
879      do      do
880        {        {
881        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, eptrb, RM64);        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, eptrb, RM64);
# Line 857  for (;;) Line 894  for (;;)
894    
895        if (rrc != MATCH_NOMATCH) RRETURN(rrc);        if (rrc != MATCH_NOMATCH) RRETURN(rrc);
896        ecode += GET(ecode,1);        ecode += GET(ecode,1);
897        md->mark = save_mark;        md->mark = save_mark;
898        }        }
899      while (*ecode == OP_ALT);      while (*ecode == OP_ALT);
900    
# Line 897  for (;;) Line 934  for (;;)
934        }        }
935      else  /* OP_KETRMAX */      else  /* OP_KETRMAX */
936        {        {
       md->match_function_type = MATCH_CBEGROUP;  
937        RMATCH(eptr, prev, offset_top, md, eptrb, RM66);        RMATCH(eptr, prev, offset_top, md, eptrb, RM66);
938        if (rrc != MATCH_NOMATCH) RRETURN(rrc);        if (rrc != MATCH_NOMATCH) RRETURN(rrc);
939        ecode += 1 + LINK_SIZE;        ecode += 1 + LINK_SIZE;
# Line 937  for (;;) Line 973  for (;;)
973        save_offset2 = md->offset_vector[offset+1];        save_offset2 = md->offset_vector[offset+1];
974        save_offset3 = md->offset_vector[md->offset_end - number];        save_offset3 = md->offset_vector[md->offset_end - number];
975        save_capture_last = md->capture_last;        save_capture_last = md->capture_last;
976        save_mark = md->mark;        save_mark = md->mark;
977    
978        DPRINTF(("saving %d %d %d\n", save_offset1, save_offset2, save_offset3));        DPRINTF(("saving %d %d %d\n", save_offset1, save_offset2, save_offset3));
979        md->offset_vector[md->offset_end - number] =        md->offset_vector[md->offset_end - number] =
# Line 1026  for (;;) Line 1062  for (;;)
1062    
1063      for (;;)      for (;;)
1064        {        {
1065        if (op >= OP_SBRA || op == OP_ONCE) md->match_function_type = MATCH_CBEGROUP;        if (op >= OP_SBRA || op == OP_ONCE)
1066            md->match_function_type = MATCH_CBEGROUP;
1067    
1068        /* If this is not a possibly empty group, and there are no (*THEN)s in        /* If this is not a possibly empty group, and there are no (*THEN)s in
1069        the pattern, and this is the final alternative, optimize as described        the pattern, and this is the final alternative, optimize as described
# Line 1041  for (;;) Line 1078  for (;;)
1078        /* In all other cases, we have to make another call to match(). */        /* In all other cases, we have to make another call to match(). */
1079    
1080        save_mark = md->mark;        save_mark = md->mark;
1081          save_capture_last = md->capture_last;
1082        RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md, eptrb,        RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md, eptrb,
1083          RM2);          RM2);
1084    
1085        /* See comment in the code for capturing groups above about handling        /* See comment in the code for capturing groups above about handling
1086        THEN. */        THEN. */
1087    
# Line 1070  for (;;) Line 1108  for (;;)
1108          RRETURN(rrc);          RRETURN(rrc);
1109          }          }
1110        ecode += GET(ecode, 1);        ecode += GET(ecode, 1);
1111        md->mark = save_mark;        md->mark = save_mark;
1112        if (*ecode != OP_ALT) break;        if (*ecode != OP_ALT) break;
1113          md->capture_last = save_capture_last;
1114        }        }
1115    
1116      RRETURN(MATCH_NOMATCH);      RRETURN(MATCH_NOMATCH);
# Line 1193  for (;;) Line 1232  for (;;)
1232      POSSESSIVE_NON_CAPTURE:      POSSESSIVE_NON_CAPTURE:
1233      matched_once = FALSE;      matched_once = FALSE;
1234      code_offset = (int)(ecode - md->start_code);      code_offset = (int)(ecode - md->start_code);
1235        save_capture_last = md->capture_last;
1236    
1237      for (;;)      for (;;)
1238        {        {
# Line 1222  for (;;) Line 1262  for (;;)
1262        if (rrc != MATCH_NOMATCH) RRETURN(rrc);        if (rrc != MATCH_NOMATCH) RRETURN(rrc);
1263        ecode += GET(ecode, 1);        ecode += GET(ecode, 1);
1264        if (*ecode != OP_ALT) break;        if (*ecode != OP_ALT) break;
1265          md->capture_last = save_capture_last;
1266        }        }
1267    
1268      if (matched_once || allow_zero)      if (matched_once || allow_zero)
# Line 1253  for (;;) Line 1294  for (;;)
1294          cb.version          = 2;   /* Version 1 of the callout block */          cb.version          = 2;   /* Version 1 of the callout block */
1295          cb.callout_number   = ecode[LINK_SIZE+2];          cb.callout_number   = ecode[LINK_SIZE+2];
1296          cb.offset_vector    = md->offset_vector;          cb.offset_vector    = md->offset_vector;
1297  #ifdef COMPILE_PCRE8  #if defined COMPILE_PCRE8
1298          cb.subject          = (PCRE_SPTR)md->start_subject;          cb.subject          = (PCRE_SPTR)md->start_subject;
1299  #else  #elif defined COMPILE_PCRE16
1300          cb.subject          = (PCRE_SPTR16)md->start_subject;          cb.subject          = (PCRE_SPTR16)md->start_subject;
1301    #elif defined COMPILE_PCRE32
1302            cb.subject          = (PCRE_SPTR32)md->start_subject;
1303  #endif  #endif
1304          cb.subject_length   = (int)(md->end_subject - md->start_subject);          cb.subject_length   = (int)(md->end_subject - md->start_subject);
1305          cb.start_match      = (int)(mstart - md->start_subject);          cb.start_match      = (int)(mstart - md->start_subject);
# Line 1264  for (;;) Line 1307  for (;;)
1307          cb.pattern_position = GET(ecode, LINK_SIZE + 3);          cb.pattern_position = GET(ecode, LINK_SIZE + 3);
1308          cb.next_item_length = GET(ecode, 3 + 2*LINK_SIZE);          cb.next_item_length = GET(ecode, 3 + 2*LINK_SIZE);
1309          cb.capture_top      = offset_top/2;          cb.capture_top      = offset_top/2;
1310          cb.capture_last     = md->capture_last;          cb.capture_last     = md->capture_last & CAPLMASK;
1311            /* Internal change requires this for API compatibility. */
1312            if (cb.capture_last == 0) cb.capture_last = -1;
1313          cb.callout_data     = md->callout_data;          cb.callout_data     = md->callout_data;
1314          cb.mark             = md->nomatch_mark;          cb.mark             = md->nomatch_mark;
1315          if ((rrc = (*PUBL(callout))(&cb)) > 0) RRETURN(MATCH_NOMATCH);          if ((rrc = (*PUBL(callout))(&cb)) > 0) RRETURN(MATCH_NOMATCH);
1316          if (rrc < 0) RRETURN(rrc);          if (rrc < 0) RRETURN(rrc);
1317          }          }
1318        ecode += PRIV(OP_lengths)[OP_CALLOUT];        ecode += PRIV(OP_lengths)[OP_CALLOUT];
1319          codelink -= PRIV(OP_lengths)[OP_CALLOUT];
1320        }        }
1321    
1322      condcode = ecode[LINK_SIZE+1];      condcode = ecode[LINK_SIZE+1];
# Line 1286  for (;;) Line 1332  for (;;)
1332          }          }
1333        else        else
1334          {          {
1335          int recno = GET2(ecode, LINK_SIZE + 2);   /* Recursion group number*/          unsigned int recno = GET2(ecode, LINK_SIZE + 2);   /* Recursion group number*/
1336          condition = (recno == RREF_ANY || recno == md->recursive->group_num);          condition = (recno == RREF_ANY || recno == md->recursive->group_num);
1337    
1338          /* If the test is for recursion into a specific subpattern, and it is          /* If the test is for recursion into a specific subpattern, and it is
# Line 1358  for (;;) Line 1404  for (;;)
1404    
1405        if (!condition && condcode == OP_NCREF)        if (!condition && condcode == OP_NCREF)
1406          {          {
1407          int refno = offset >> 1;          unsigned int refno = offset >> 1;
1408          pcre_uchar *slotA = md->name_table;          pcre_uchar *slotA = md->name_table;
1409    
1410          for (i = 0; i < md->name_count; i++)          for (i = 0; i < md->name_count; i++)
# Line 1486  for (;;) Line 1532  for (;;)
1532      to close any currently open capturing brackets. */      to close any currently open capturing brackets. */
1533    
1534      case OP_CLOSE:      case OP_CLOSE:
1535      number = GET2(ecode, 1);      number = GET2(ecode, 1);   /* Must be less than 65536 */
1536      offset = number << 1;      offset = number << 1;
1537    
1538  #ifdef PCRE_DEBUG  #ifdef PCRE_DEBUG
# Line 1494  for (;;) Line 1540  for (;;)
1540        printf("\n");        printf("\n");
1541  #endif  #endif
1542    
1543      md->capture_last = number;      md->capture_last = (md->capture_last & OVFLMASK) | number;
1544      if (offset >= md->offset_max) md->offset_overflow = TRUE; else      if (offset >= md->offset_max) md->capture_last |= OVFLBIT; else
1545        {        {
1546        md->offset_vector[offset] =        md->offset_vector[offset] =
1547          md->offset_vector[md->offset_end - number];          md->offset_vector[md->offset_end - number];
# Line 1549  for (;;) Line 1595  for (;;)
1595    
1596      case OP_ASSERT:      case OP_ASSERT:
1597      case OP_ASSERTBACK:      case OP_ASSERTBACK:
1598      save_mark = md->mark;      save_mark = md->mark;
1599      if (md->match_function_type == MATCH_CONDASSERT)      if (md->match_function_type == MATCH_CONDASSERT)
1600        {        {
1601        condassert = TRUE;        condassert = TRUE;
# Line 1557  for (;;) Line 1603  for (;;)
1603        }        }
1604      else condassert = FALSE;      else condassert = FALSE;
1605    
1606        /* Loop for each branch */
1607    
1608      do      do
1609        {        {
1610        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, NULL, RM4);        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, NULL, RM4);
# Line 1565  for (;;) Line 1613  for (;;)
1613          mstart = md->start_match_ptr;   /* In case \K reset it */          mstart = md->start_match_ptr;   /* In case \K reset it */
1614          break;          break;
1615          }          }
1616          md->mark = save_mark;
1617    
1618          /* See comment in the code for capturing groups above about handling
1619          THEN. */
1620    
1621        /* PCRE does not allow THEN to escape beyond an assertion; it is treated        if (rrc == MATCH_THEN)
1622        as NOMATCH. */          {
1623            next = ecode + GET(ecode,1);
1624            if (md->start_match_ptr < next &&
1625                (*ecode == OP_ALT || *next == OP_ALT))
1626              rrc = MATCH_NOMATCH;
1627            }
1628    
1629          /* Anything other than NOMATCH causes the assertion to fail. This
1630          includes COMMIT, SKIP, and PRUNE. However, this consistent approach does
1631          not always have exactly the same effect as in Perl. */
1632    
1633        if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN) RRETURN(rrc);        if (rrc != MATCH_NOMATCH) RRETURN(rrc);
1634        ecode += GET(ecode, 1);        ecode += GET(ecode, 1);
       md->mark = save_mark;  
1635        }        }
1636      while (*ecode == OP_ALT);      while (*ecode == OP_ALT);
1637    
1638        /* If we have tried all the alternative branches, the assertion has
1639        failed. */
1640    
1641      if (*ecode == OP_KET) RRETURN(MATCH_NOMATCH);      if (*ecode == OP_KET) RRETURN(MATCH_NOMATCH);
1642    
# Line 1581  for (;;) Line 1644  for (;;)
1644    
1645      if (condassert) RRETURN(MATCH_MATCH);      if (condassert) RRETURN(MATCH_MATCH);
1646    
1647      /* Continue from after the assertion, updating the offsets high water      /* Continue from after a successful assertion, updating the offsets high
1648      mark, since extracts may have been taken during the assertion. */      water mark, since extracts may have been taken during the assertion. */
1649    
1650      do ecode += GET(ecode,1); while (*ecode == OP_ALT);      do ecode += GET(ecode,1); while (*ecode == OP_ALT);
1651      ecode += 1 + LINK_SIZE;      ecode += 1 + LINK_SIZE;
1652      offset_top = md->end_offset_top;      offset_top = md->end_offset_top;
1653      continue;      continue;
1654    
1655      /* Negative assertion: all branches must fail to match. Encountering SKIP,      /* Negative assertion: all branches must fail to match for the assertion to
1656      PRUNE, or COMMIT means we must assume failure without checking subsequent      succeed. */
     branches. */  
1657    
1658      case OP_ASSERT_NOT:      case OP_ASSERT_NOT:
1659      case OP_ASSERTBACK_NOT:      case OP_ASSERTBACK_NOT:
1660      save_mark = md->mark;      save_mark = md->mark;
1661      if (md->match_function_type == MATCH_CONDASSERT)      if (md->match_function_type == MATCH_CONDASSERT)
1662        {        {
1663        condassert = TRUE;        condassert = TRUE;
# Line 1603  for (;;) Line 1665  for (;;)
1665        }        }
1666      else condassert = FALSE;      else condassert = FALSE;
1667    
1668        /* Loop for each alternative branch. */
1669    
1670      do      do
1671        {        {
1672        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, NULL, RM5);        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, NULL, RM5);
1673        md->mark = save_mark;        md->mark = save_mark;
1674    
1675          /* A successful match means the assertion has failed. */
1676    
1677        if (rrc == MATCH_MATCH || rrc == MATCH_ACCEPT) RRETURN(MATCH_NOMATCH);        if (rrc == MATCH_MATCH || rrc == MATCH_ACCEPT) RRETURN(MATCH_NOMATCH);
1678        if (rrc == MATCH_SKIP || rrc == MATCH_PRUNE || rrc == MATCH_COMMIT)  
1679          /* See comment in the code for capturing groups above about handling
1680          THEN. */
1681    
1682          if (rrc == MATCH_THEN)
1683          {          {
1684          do ecode += GET(ecode,1); while (*ecode == OP_ALT);          next = ecode + GET(ecode,1);
1685          break;          if (md->start_match_ptr < next &&
1686                (*ecode == OP_ALT || *next == OP_ALT))
1687              rrc = MATCH_NOMATCH;
1688          }          }
1689    
1690          /* No match on a branch means we must carry on and try the next branch.
1691          Anything else, in particular, SKIP, PRUNE, etc. causes a failure in the
1692          enclosing branch. This is a consistent approach, but does not always have
1693          the same effect as in Perl. */
1694    
1695        /* PCRE does not allow THEN to escape beyond an assertion; it is treated        if (rrc != MATCH_NOMATCH) RRETURN(rrc);
       as NOMATCH. */  
   
       if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN) RRETURN(rrc);  
1696        ecode += GET(ecode,1);        ecode += GET(ecode,1);
1697        }        }
1698      while (*ecode == OP_ALT);      while (*ecode == OP_ALT);
1699    
1700        /* All branches in the assertion failed to match. */
1701    
1702      if (condassert) RRETURN(MATCH_MATCH);  /* Condition assertion */      if (condassert) RRETURN(MATCH_MATCH);  /* Condition assertion */
1703        ecode += 1 + LINK_SIZE;                /* Continue with current branch */
     ecode += 1 + LINK_SIZE;  
1704      continue;      continue;
1705    
1706      /* Move the subject pointer back. This occurs only at the start of      /* Move the subject pointer back. This occurs only at the start of
# Line 1671  for (;;) Line 1747  for (;;)
1747        cb.version          = 2;   /* Version 1 of the callout block */        cb.version          = 2;   /* Version 1 of the callout block */
1748        cb.callout_number   = ecode[1];        cb.callout_number   = ecode[1];
1749        cb.offset_vector    = md->offset_vector;        cb.offset_vector    = md->offset_vector;
1750  #ifdef COMPILE_PCRE8  #if defined COMPILE_PCRE8
1751        cb.subject          = (PCRE_SPTR)md->start_subject;        cb.subject          = (PCRE_SPTR)md->start_subject;
1752  #else  #elif defined COMPILE_PCRE16
1753        cb.subject          = (PCRE_SPTR16)md->start_subject;        cb.subject          = (PCRE_SPTR16)md->start_subject;
1754    #elif defined COMPILE_PCRE32
1755          cb.subject          = (PCRE_SPTR32)md->start_subject;
1756  #endif  #endif
1757        cb.subject_length   = (int)(md->end_subject - md->start_subject);        cb.subject_length   = (int)(md->end_subject - md->start_subject);
1758        cb.start_match      = (int)(mstart - md->start_subject);        cb.start_match      = (int)(mstart - md->start_subject);
# Line 1682  for (;;) Line 1760  for (;;)
1760        cb.pattern_position = GET(ecode, 2);        cb.pattern_position = GET(ecode, 2);
1761        cb.next_item_length = GET(ecode, 2 + LINK_SIZE);        cb.next_item_length = GET(ecode, 2 + LINK_SIZE);
1762        cb.capture_top      = offset_top/2;        cb.capture_top      = offset_top/2;
1763        cb.capture_last     = md->capture_last;        cb.capture_last     = md->capture_last & CAPLMASK;
1764          /* Internal change requires this for API compatibility. */
1765          if (cb.capture_last == 0) cb.capture_last = -1;
1766        cb.callout_data     = md->callout_data;        cb.callout_data     = md->callout_data;
1767        cb.mark             = md->nomatch_mark;        cb.mark             = md->nomatch_mark;
1768        if ((rrc = (*PUBL(callout))(&cb)) > 0) RRETURN(MATCH_NOMATCH);        if ((rrc = (*PUBL(callout))(&cb)) > 0) RRETURN(MATCH_NOMATCH);
# Line 1711  for (;;) Line 1791  for (;;)
1791      case OP_RECURSE:      case OP_RECURSE:
1792        {        {
1793        recursion_info *ri;        recursion_info *ri;
1794        int recno;        unsigned int recno;
1795    
1796        callpat = md->start_code + GET(ecode, 1);        callpat = md->start_code + GET(ecode, 1);
1797        recno = (callpat == md->start_code)? 0 :        recno = (callpat == md->start_code)? 0 :
# Line 1728  for (;;) Line 1808  for (;;)
1808        /* Add to "recursing stack" */        /* Add to "recursing stack" */
1809    
1810        new_recursive.group_num = recno;        new_recursive.group_num = recno;
1811          new_recursive.saved_capture_last = md->capture_last;
1812        new_recursive.subject_position = eptr;        new_recursive.subject_position = eptr;
1813        new_recursive.prevrec = md->recursive;        new_recursive.prevrec = md->recursive;
1814        md->recursive = &new_recursive;        md->recursive = &new_recursive;
# Line 1751  for (;;) Line 1832  for (;;)
1832              new_recursive.saved_max * sizeof(int));              new_recursive.saved_max * sizeof(int));
1833    
1834        /* OK, now we can do the recursion. After processing each alternative,        /* OK, now we can do the recursion. After processing each alternative,
1835        restore the offset data. If there were nested recursions, md->recursive        restore the offset data and the last captured value. If there were nested
1836        might be changed, so reset it before looping. */        recursions, md->recursive might be changed, so reset it before looping.
1837          */
1838    
1839        DPRINTF(("Recursing into group %d\n", new_recursive.group_num));        DPRINTF(("Recursing into group %d\n", new_recursive.group_num));
1840        cbegroup = (*callpat >= OP_SBRA);        cbegroup = (*callpat >= OP_SBRA);
# Line 1763  for (;;) Line 1845  for (;;)
1845            md, eptrb, RM6);            md, eptrb, RM6);
1846          memcpy(md->offset_vector, new_recursive.offset_save,          memcpy(md->offset_vector, new_recursive.offset_save,
1847              new_recursive.saved_max * sizeof(int));              new_recursive.saved_max * sizeof(int));
1848            md->capture_last = new_recursive.saved_capture_last;
1849          md->recursive = new_recursive.prevrec;          md->recursive = new_recursive.prevrec;
1850          if (rrc == MATCH_MATCH || rrc == MATCH_ACCEPT)          if (rrc == MATCH_MATCH || rrc == MATCH_ACCEPT)
1851            {            {
# Line 1779  for (;;) Line 1862  for (;;)
1862            goto RECURSION_MATCHED;        /* Exit loop; end processing */            goto RECURSION_MATCHED;        /* Exit loop; end processing */
1863            }            }
1864    
1865          /* 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
1866          as NOMATCH. */          recursion; they are treated as NOMATCH. These codes are defined in a
1867            range that can be tested for. Any other return code is an error. */
1868    
1869          else if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN)          else if (rrc != MATCH_NOMATCH &&
1870                     (rrc < MATCH_BACKTRACK_MIN || rrc > MATCH_BACKTRACK_MAX))
1871            {            {
1872            DPRINTF(("Recursion gave error %d\n", rrc));            DPRINTF(("Recursion gave error %d\n", rrc));
1873            if (new_recursive.offset_save != stacksave)            if (new_recursive.offset_save != stacksave)
# Line 1912  for (;;) Line 1997  for (;;)
1997    
1998        /* Deal with capturing */        /* Deal with capturing */
1999    
2000        md->capture_last = number;        md->capture_last = (md->capture_last & OVFLMASK) | number;
2001        if (offset >= md->offset_max) md->offset_overflow = TRUE; else        if (offset >= md->offset_max) md->capture_last |= OVFLBIT; else
2002          {          {
2003          /* If offset is greater than offset_top, it means that we are          /* If offset is greater than offset_top, it means that we are
2004          "skipping" a capturing group, and that group's offsets must be marked          "skipping" a capturing group, and that group's offsets must be marked
# Line 1993  for (;;) Line 2078  for (;;)
2078          }          }
2079        if (*prev >= OP_SBRA)    /* Could match an empty string */        if (*prev >= OP_SBRA)    /* Could match an empty string */
2080          {          {
         md->match_function_type = MATCH_CBEGROUP;  
2081          RMATCH(eptr, prev, offset_top, md, eptrb, RM50);          RMATCH(eptr, prev, offset_top, md, eptrb, RM50);
2082          RRETURN(rrc);          RRETURN(rrc);
2083          }          }
# Line 2002  for (;;) Line 2086  for (;;)
2086        }        }
2087      else  /* OP_KETRMAX */      else  /* OP_KETRMAX */
2088        {        {
       if (*prev >= OP_SBRA) md->match_function_type = MATCH_CBEGROUP;  
2089        RMATCH(eptr, prev, offset_top, md, eptrb, RM13);        RMATCH(eptr, prev, offset_top, md, eptrb, RM13);
2090        if (rrc == MATCH_ONCE && md->once_target == prev) rrc = MATCH_NOMATCH;        if (rrc == MATCH_ONCE && md->once_target == prev) rrc = MATCH_NOMATCH;
2091        if (rrc != MATCH_NOMATCH) RRETURN(rrc);        if (rrc != MATCH_NOMATCH) RRETURN(rrc);
# Line 2059  for (;;) Line 2142  for (;;)
2142    
2143      case OP_DOLLM:      case OP_DOLLM:
2144      if (eptr < md->end_subject)      if (eptr < md->end_subject)
2145        { if (!IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH); }        {
2146          if (!IS_NEWLINE(eptr))
2147            {
2148            if (md->partial != 0 &&
2149                eptr + 1 >= md->end_subject &&
2150                NLBLOCK->nltype == NLTYPE_FIXED &&
2151                NLBLOCK->nllen == 2 &&
2152                RAWUCHARTEST(eptr) == NLBLOCK->nl[0])
2153              {
2154              md->hitend = TRUE;
2155              if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
2156              }
2157            RRETURN(MATCH_NOMATCH);
2158            }
2159          }
2160      else      else
2161        {        {
2162        if (md->noteol) RRETURN(MATCH_NOMATCH);        if (md->noteol) RRETURN(MATCH_NOMATCH);
# Line 2091  for (;;) Line 2188  for (;;)
2188      ASSERT_NL_OR_EOS:      ASSERT_NL_OR_EOS:
2189      if (eptr < md->end_subject &&      if (eptr < md->end_subject &&
2190          (!IS_NEWLINE(eptr) || eptr != md->end_subject - md->nllen))          (!IS_NEWLINE(eptr) || eptr != md->end_subject - md->nllen))
2191          {
2192          if (md->partial != 0 &&
2193              eptr + 1 >= md->end_subject &&
2194              NLBLOCK->nltype == NLTYPE_FIXED &&
2195              NLBLOCK->nllen == 2 &&
2196              RAWUCHARTEST(eptr) == NLBLOCK->nl[0])
2197            {
2198            md->hitend = TRUE;
2199            if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
2200            }
2201        RRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2202          }
2203    
2204      /* Either at end of string or \n before end. */      /* Either at end of string or \n before end. */
2205    
# Line 2219  for (;;) Line 2327  for (;;)
2327        }        }
2328      break;      break;
2329    
2330      /* Match a single character type; inline for speed */      /* Match any single character type except newline; have to take care with
2331        CRLF newlines and partial matching. */
2332    
2333      case OP_ANY:      case OP_ANY:
2334      if (IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH);      if (IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH);
2335        if (md->partial != 0 &&
2336            eptr + 1 >= md->end_subject &&
2337            NLBLOCK->nltype == NLTYPE_FIXED &&
2338            NLBLOCK->nllen == 2 &&
2339            RAWUCHARTEST(eptr) == NLBLOCK->nl[0])
2340          {
2341          md->hitend = TRUE;
2342          if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
2343          }
2344    
2345      /* Fall through */      /* Fall through */
2346    
2347        /* Match any single character whatsoever. */
2348    
2349      case OP_ALLANY:      case OP_ALLANY:
2350      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 */
2351        {                            /* not be updated before SCHECK_PARTIAL. */        {                            /* not be updated before SCHECK_PARTIAL. */
# Line 2364  for (;;) Line 2485  for (;;)
2485        {        {
2486        default: RRETURN(MATCH_NOMATCH);        default: RRETURN(MATCH_NOMATCH);
2487    
2488        case 0x000d:        case CHAR_CR:
2489        if (eptr < md->end_subject && *eptr == 0x0a) eptr++;        if (eptr >= md->end_subject)
2490            {
2491            SCHECK_PARTIAL();
2492            }
2493          else if (RAWUCHARTEST(eptr) == CHAR_LF) eptr++;
2494        break;        break;
2495    
2496        case 0x000a:        case CHAR_LF:
2497        break;        break;
2498    
2499        case 0x000b:        case CHAR_VT:
2500        case 0x000c:        case CHAR_FF:
2501        case 0x0085:        case CHAR_NEL:
2502    #ifndef EBCDIC
2503        case 0x2028:        case 0x2028:
2504        case 0x2029:        case 0x2029:
2505    #endif  /* Not EBCDIC */
2506        if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);        if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);
2507        break;        break;
2508        }        }
# Line 2391  for (;;) Line 2518  for (;;)
2518      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2519      switch(c)      switch(c)
2520        {        {
2521          HSPACE_CASES: RRETURN(MATCH_NOMATCH);  /* Byte and multibyte cases */
2522        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);  
2523        }        }
2524      ecode++;      ecode++;
2525      break;      break;
# Line 2425  for (;;) Line 2533  for (;;)
2533      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2534      switch(c)      switch(c)
2535        {        {
2536          HSPACE_CASES: break;  /* Byte and multibyte cases */
2537        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;  
2538        }        }
2539      ecode++;      ecode++;
2540      break;      break;
# Line 2459  for (;;) Line 2548  for (;;)
2548      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2549      switch(c)      switch(c)
2550        {        {
2551          VSPACE_CASES: RRETURN(MATCH_NOMATCH);
2552        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);  
2553        }        }
2554      ecode++;      ecode++;
2555      break;      break;
# Line 2481  for (;;) Line 2563  for (;;)
2563      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2564      switch(c)      switch(c)
2565        {        {
2566          VSPACE_CASES: break;
2567        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;  
2568        }        }
2569      ecode++;      ecode++;
2570      break;      break;
# Line 2507  for (;;) Line 2582  for (;;)
2582        }        }
2583      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2584        {        {
2585          const pcre_uint32 *cp;
2586        const ucd_record *prop = GET_UCD(c);        const ucd_record *prop = GET_UCD(c);
2587    
2588        switch(ecode[1])        switch(ecode[1])
# Line 2567  for (;;) Line 2643  for (;;)
2643            RRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
2644          break;          break;
2645    
2646            case PT_CLIST:
2647            cp = PRIV(ucd_caseless_sets) + ecode[2];
2648            for (;;)
2649              {
2650              if (c < *cp)
2651                { if (op == OP_PROP) { RRETURN(MATCH_NOMATCH); } else break; }
2652              if (c == *cp++)
2653                { if (op == OP_PROP) break; else { RRETURN(MATCH_NOMATCH); } }
2654              }
2655            break;
2656    
2657            case PT_UCNC:
2658            if ((c == CHAR_DOLLAR_SIGN || c == CHAR_COMMERCIAL_AT ||
2659                 c == CHAR_GRAVE_ACCENT || (c >= 0xa0 && c <= 0xd7ff) ||
2660                 c >= 0xe000) == (op == OP_NOTPROP))
2661              RRETURN(MATCH_NOMATCH);
2662            break;
2663    
2664          /* This should never occur */          /* This should never occur */
2665    
2666          default:          default:
# Line 2586  for (;;) Line 2680  for (;;)
2680        SCHECK_PARTIAL();        SCHECK_PARTIAL();
2681        RRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2682        }        }
2683      GETCHARINCTEST(c, eptr);      else
     if (UCD_CATEGORY(c) == ucp_M) RRETURN(MATCH_NOMATCH);  
     while (eptr < md->end_subject)  
2684        {        {
2685        int len = 1;        int lgb, rgb;
2686        if (!utf) c = *eptr; else { GETCHARLEN(c, eptr, len); }        GETCHARINCTEST(c, eptr);
2687        if (UCD_CATEGORY(c) != ucp_M) break;        lgb = UCD_GRAPHBREAK(c);
2688        eptr += len;        while (eptr < md->end_subject)
2689            {
2690            int len = 1;
2691            if (!utf) c = *eptr; else { GETCHARLEN(c, eptr, len); }
2692            rgb = UCD_GRAPHBREAK(c);
2693            if ((PRIV(ucp_gbtable)[lgb] & (1 << rgb)) == 0) break;
2694            lgb = rgb;
2695            eptr += len;
2696            }
2697        }        }
2698        CHECK_PARTIAL();
2699      ecode++;      ecode++;
2700      break;      break;
2701  #endif  #endif  /* SUPPORT_UCP */
2702    
2703    
2704      /* Match a back reference, possibly repeatedly. Look past the end of the      /* Match a back reference, possibly repeatedly. Look past the end of the
# Line 2660  for (;;) Line 2761  for (;;)
2761        default:               /* No repeat follows */        default:               /* No repeat follows */
2762        if ((length = match_ref(offset, eptr, length, md, caseless)) < 0)        if ((length = match_ref(offset, eptr, length, md, caseless)) < 0)
2763          {          {
2764            if (length == -2) eptr = md->end_subject;   /* Partial match */
2765          CHECK_PARTIAL();          CHECK_PARTIAL();
2766          RRETURN(MATCH_NOMATCH);          RRETURN(MATCH_NOMATCH);
2767          }          }
# Line 2685  for (;;) Line 2787  for (;;)
2787        int slength;        int slength;
2788        if ((slength = match_ref(offset, eptr, length, md, caseless)) < 0)        if ((slength = match_ref(offset, eptr, length, md, caseless)) < 0)
2789          {          {
2790            if (slength == -2) eptr = md->end_subject;   /* Partial match */
2791          CHECK_PARTIAL();          CHECK_PARTIAL();
2792          RRETURN(MATCH_NOMATCH);          RRETURN(MATCH_NOMATCH);
2793          }          }
# Line 2708  for (;;) Line 2811  for (;;)
2811          if (fi >= max) RRETURN(MATCH_NOMATCH);          if (fi >= max) RRETURN(MATCH_NOMATCH);
2812          if ((slength = match_ref(offset, eptr, length, md, caseless)) < 0)          if ((slength = match_ref(offset, eptr, length, md, caseless)) < 0)
2813            {            {
2814              if (slength == -2) eptr = md->end_subject;   /* Partial match */
2815            CHECK_PARTIAL();            CHECK_PARTIAL();
2816            RRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
2817            }            }
# Line 2726  for (;;) Line 2830  for (;;)
2830          int slength;          int slength;
2831          if ((slength = match_ref(offset, eptr, length, md, caseless)) < 0)          if ((slength = match_ref(offset, eptr, length, md, caseless)) < 0)
2832            {            {
2833            CHECK_PARTIAL();            /* Can't use CHECK_PARTIAL because we don't want to update eptr in
2834              the soft partial matching case. */
2835    
2836              if (slength == -2 && md->partial != 0 &&
2837                  md->end_subject > md->start_used_ptr)
2838                {
2839                md->hitend = TRUE;
2840                if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
2841                }
2842            break;            break;
2843            }            }
2844          eptr += slength;          eptr += slength;
2845          }          }
2846    
2847        while (eptr >= pp)        while (eptr >= pp)
2848          {          {
2849          RMATCH(eptr, ecode, offset_top, md, eptrb, RM15);          RMATCH(eptr, ecode, offset_top, md, eptrb, RM15);
# Line 3094  for (;;) Line 3207  for (;;)
3207          CHECK_PARTIAL();             /* Not SCHECK_PARTIAL() */          CHECK_PARTIAL();             /* Not SCHECK_PARTIAL() */
3208          RRETURN(MATCH_NOMATCH);          RRETURN(MATCH_NOMATCH);
3209          }          }
3210        while (length-- > 0) if (*ecode++ != *eptr++) RRETURN(MATCH_NOMATCH);        while (length-- > 0) if (*ecode++ != RAWUCHARINC(eptr)) RRETURN(MATCH_NOMATCH);
3211        }        }
3212      else      else
3213  #endif  #endif
# Line 3134  for (;;) Line 3247  for (;;)
3247    
3248        if (fc < 128)        if (fc < 128)
3249          {          {
3250          if (md->lcc[fc]          pcre_uint32 cc = RAWUCHAR(eptr);
3251              != TABLE_GET(*eptr, md->lcc, *eptr)) RRETURN(MATCH_NOMATCH);          if (md->lcc[fc] != TABLE_GET(cc, md->lcc, cc)) RRETURN(MATCH_NOMATCH);
3252          ecode++;          ecode++;
3253          eptr++;          eptr++;
3254          }          }
# Line 3146  for (;;) Line 3259  for (;;)
3259    
3260        else        else
3261          {          {
3262          unsigned int dc;          pcre_uint32 dc;
3263          GETCHARINC(dc, eptr);          GETCHARINC(dc, eptr);
3264          ecode += length;          ecode += length;
3265    
# Line 3256  for (;;) Line 3369  for (;;)
3369        if (length > 1)        if (length > 1)
3370          {          {
3371  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
3372          unsigned int othercase;          pcre_uint32 othercase;
3373          if (op >= OP_STARI &&     /* Caseless */          if (op >= OP_STARI &&     /* Caseless */
3374              (othercase = UCD_OTHERCASE(fc)) != fc)              (othercase = UCD_OTHERCASE(fc)) != fc)
3375            oclength = PRIV(ord2utf)(othercase, occhars);            oclength = PRIV(ord2utf)(othercase, occhars);
# Line 3360  for (;;) Line 3473  for (;;)
3473      maximizing, find the maximum number of characters and work backwards. */      maximizing, find the maximum number of characters and work backwards. */
3474    
3475      DPRINTF(("matching %c{%d,%d} against subject %.*s\n", fc, min, max,      DPRINTF(("matching %c{%d,%d} against subject %.*s\n", fc, min, max,
3476        max, eptr));        max, (char *)eptr));
3477    
3478      if (op >= OP_STARI)  /* Caseless */      if (op >= OP_STARI)  /* Caseless */
3479        {        {
# Line 3383  for (;;) Line 3496  for (;;)
3496    
3497        for (i = 1; i <= min; i++)        for (i = 1; i <= min; i++)
3498          {          {
3499            pcre_uint32 cc;                 /* Faster than pcre_uchar */
3500          if (eptr >= md->end_subject)          if (eptr >= md->end_subject)
3501            {            {
3502            SCHECK_PARTIAL();            SCHECK_PARTIAL();
3503            RRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
3504            }            }
3505          if (fc != *eptr && foc != *eptr) RRETURN(MATCH_NOMATCH);          cc = RAWUCHARTEST(eptr);
3506            if (fc != cc && foc != cc) RRETURN(MATCH_NOMATCH);
3507          eptr++;          eptr++;
3508          }          }
3509        if (min == max) continue;        if (min == max) continue;
# Line 3396  for (;;) Line 3511  for (;;)
3511          {          {
3512          for (fi = min;; fi++)          for (fi = min;; fi++)
3513            {            {
3514              pcre_uint32 cc;               /* Faster than pcre_uchar */
3515            RMATCH(eptr, ecode, offset_top, md, eptrb, RM24);            RMATCH(eptr, ecode, offset_top, md, eptrb, RM24);
3516            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3517            if (fi >= max) RRETURN(MATCH_NOMATCH);            if (fi >= max) RRETURN(MATCH_NOMATCH);
# Line 3404  for (;;) Line 3520  for (;;)
3520              SCHECK_PARTIAL();              SCHECK_PARTIAL();
3521              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
3522              }              }
3523            if (fc != *eptr && foc != *eptr) RRETURN(MATCH_NOMATCH);            cc = RAWUCHARTEST(eptr);
3524              if (fc != cc && foc != cc) RRETURN(MATCH_NOMATCH);
3525            eptr++;            eptr++;
3526            }            }
3527          /* Control never gets here */          /* Control never gets here */
# Line 3414  for (;;) Line 3531  for (;;)
3531          pp = eptr;          pp = eptr;
3532          for (i = min; i < max; i++)          for (i = min; i < max; i++)
3533            {            {
3534              pcre_uint32 cc;               /* Faster than pcre_uchar */
3535            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
3536              {              {
3537              SCHECK_PARTIAL();              SCHECK_PARTIAL();
3538              break;              break;
3539              }              }
3540            if (fc != *eptr && foc != *eptr) break;            cc = RAWUCHARTEST(eptr);
3541              if (fc != cc && foc != cc) break;
3542            eptr++;            eptr++;
3543            }            }
3544    
# Line 3447  for (;;) Line 3566  for (;;)
3566            SCHECK_PARTIAL();            SCHECK_PARTIAL();
3567            RRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
3568            }            }
3569          if (fc != *eptr++) RRETURN(MATCH_NOMATCH);          if (fc != RAWUCHARINCTEST(eptr)) RRETURN(MATCH_NOMATCH);
3570          }          }
3571    
3572        if (min == max) continue;        if (min == max) continue;
# Line 3464  for (;;) Line 3583  for (;;)
3583              SCHECK_PARTIAL();              SCHECK_PARTIAL();
3584              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
3585              }              }
3586            if (fc != *eptr++) RRETURN(MATCH_NOMATCH);            if (fc != RAWUCHARINCTEST(eptr)) RRETURN(MATCH_NOMATCH);
3587            }            }
3588          /* Control never gets here */          /* Control never gets here */
3589          }          }
# Line 3478  for (;;) Line 3597  for (;;)
3597              SCHECK_PARTIAL();              SCHECK_PARTIAL();
3598              break;              break;
3599              }              }
3600            if (fc != *eptr) break;            if (fc != RAWUCHARTEST(eptr)) break;
3601            eptr++;            eptr++;
3602            }            }
3603          if (possessive) continue;          if (possessive) continue;
# Line 3504  for (;;) Line 3623  for (;;)
3623        SCHECK_PARTIAL();        SCHECK_PARTIAL();
3624        RRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
3625        }        }
     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  
3626  #ifdef SUPPORT_UTF  #ifdef SUPPORT_UTF
3627        if (utf)
3628          {
3629          register pcre_uint32 ch, och;
3630    
3631          ecode++;
3632          GETCHARINC(ch, ecode);
3633          GETCHARINC(c, eptr);
3634    
3635          if (op == OP_NOT)
3636            {
3637            if (ch == c) RRETURN(MATCH_NOMATCH);
3638            }
3639          else
3640            {
3641  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
3642        if (utf && ch > 127)          if (ch > 127)
3643          och = UCD_OTHERCASE(ch);            och = UCD_OTHERCASE(ch);
3644  #else  #else
3645        if (utf && ch > 127)          if (ch > 127)
3646          och = ch;            och = ch;
3647  #endif /* SUPPORT_UCP */  #endif /* SUPPORT_UCP */
3648        else          else
3649  #endif /* SUPPORT_UTF */            och = TABLE_GET(ch, md->fcc, ch);
3650          och = TABLE_GET(ch, md->fcc, ch);          if (ch == c || och == c) RRETURN(MATCH_NOMATCH);
3651  #endif /* COMPILE_PCRE8 */          }
       if (ch == c || och == c) RRETURN(MATCH_NOMATCH);  
3652        }        }
3653      else    /* Caseful */      else
3654    #endif
3655        {        {
3656        if (*ecode++ == c) RRETURN(MATCH_NOMATCH);        register pcre_uint32 ch = ecode[1];
3657          c = *eptr++;
3658          if (ch == c || (op == OP_NOTI && TABLE_GET(ch, md->fcc, ch) == c))
3659            RRETURN(MATCH_NOMATCH);
3660          ecode += 2;
3661        }        }
3662      break;      break;
3663    
# Line 3610  for (;;) Line 3737  for (;;)
3737      /* Common code for all repeated single-byte matches. */      /* Common code for all repeated single-byte matches. */
3738    
3739      REPEATNOTCHAR:      REPEATNOTCHAR:
3740      fc = *ecode++;      GETCHARINCTEST(fc, ecode);
3741    
3742      /* The code is duplicated for the caseless and caseful cases, for speed,      /* The code is duplicated for the caseless and caseful cases, for speed,
3743      since matching characters is likely to be quite common. First, ensure the      since matching characters is likely to be quite common. First, ensure the
# Line 3621  for (;;) Line 3748  for (;;)
3748      characters and work backwards. */      characters and work backwards. */
3749    
3750      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,
3751        max, eptr));        max, (char *)eptr));
3752    
3753      if (op >= OP_NOTSTARI)     /* Caseless */      if (op >= OP_NOTSTARI)     /* Caseless */
3754        {        {
 #ifdef COMPILE_PCRE8  
       /* fc must be < 128 if UTF is enabled. */  
       foc = md->fcc[fc];  
 #else  
3755  #ifdef SUPPORT_UTF  #ifdef SUPPORT_UTF
3756  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
3757        if (utf && fc > 127)        if (utf && fc > 127)
# Line 3640  for (;;) Line 3763  for (;;)
3763        else        else
3764  #endif /* SUPPORT_UTF */  #endif /* SUPPORT_UTF */
3765          foc = TABLE_GET(fc, md->fcc, fc);          foc = TABLE_GET(fc, md->fcc, fc);
 #endif /* COMPILE_PCRE8 */  
3766    
3767  #ifdef SUPPORT_UTF  #ifdef SUPPORT_UTF
3768        if (utf)        if (utf)
3769          {          {
3770          register unsigned int d;          register pcre_uint32 d;
3771          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
3772            {            {
3773            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
# Line 3654  for (;;) Line 3776  for (;;)
3776              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
3777              }              }
3778            GETCHARINC(d, eptr);            GETCHARINC(d, eptr);
3779            if (fc == d || foc == d) RRETURN(MATCH_NOMATCH);            if (fc == d || (unsigned int)foc == d) RRETURN(MATCH_NOMATCH);
3780            }            }
3781          }          }
3782        else        else
# Line 3680  for (;;) Line 3802  for (;;)
3802  #ifdef SUPPORT_UTF  #ifdef SUPPORT_UTF
3803          if (utf)          if (utf)
3804            {            {
3805            register unsigned int d;            register pcre_uint32 d;
3806            for (fi = min;; fi++)            for (fi = min;; fi++)
3807              {              {
3808              RMATCH(eptr, ecode, offset_top, md, eptrb, RM28);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM28);
# Line 3692  for (;;) Line 3814  for (;;)
3814                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
3815                }                }
3816              GETCHARINC(d, eptr);              GETCHARINC(d, eptr);
3817              if (fc == d || foc == d) RRETURN(MATCH_NOMATCH);              if (fc == d || (unsigned int)foc == d) RRETURN(MATCH_NOMATCH);
3818              }              }
3819            }            }
3820          else          else
# Line 3725  for (;;) Line 3847  for (;;)
3847  #ifdef SUPPORT_UTF  #ifdef SUPPORT_UTF
3848          if (utf)          if (utf)
3849            {            {
3850            register unsigned int d;            register pcre_uint32 d;
3851            for (i = min; i < max; i++)            for (i = min; i < max; i++)
3852              {              {
3853              int len = 1;              int len = 1;
# Line 3735  for (;;) Line 3857  for (;;)
3857                break;                break;
3858                }                }
3859              GETCHARLEN(d, eptr, len);              GETCHARLEN(d, eptr, len);
3860              if (fc == d || foc == d) break;              if (fc == d || (unsigned int)foc == d) break;
3861              eptr += len;              eptr += len;
3862              }              }
3863            if (possessive) continue;            if (possessive) continue;
# Line 3782  for (;;) Line 3904  for (;;)
3904  #ifdef SUPPORT_UTF  #ifdef SUPPORT_UTF
3905        if (utf)        if (utf)
3906          {          {
3907          register unsigned int d;          register pcre_uint32 d;
3908          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
3909            {            {
3910            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
# Line 3816  for (;;) Line 3938  for (;;)
3938  #ifdef SUPPORT_UTF  #ifdef SUPPORT_UTF
3939          if (utf)          if (utf)
3940            {            {
3941            register unsigned int d;            register pcre_uint32 d;
3942            for (fi = min;; fi++)            for (fi = min;; fi++)
3943              {              {
3944              RMATCH(eptr, ecode, offset_top, md, eptrb, RM32);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM32);
# Line 3860  for (;;) Line 3982  for (;;)
3982  #ifdef SUPPORT_UTF  #ifdef SUPPORT_UTF
3983          if (utf)          if (utf)
3984            {            {
3985            register unsigned int d;            register pcre_uint32 d;
3986            for (i = min; i < max; i++)            for (i = min; i < max; i++)
3987              {              {
3988              int len = 1;              int len = 1;
# Line 4136  for (;;) Line 4258  for (;;)
4258              }              }
4259            break;            break;
4260    
4261              case PT_CLIST:
4262              for (i = 1; i <= min; i++)
4263                {
4264                const pcre_uint32 *cp;
4265                if (eptr >= md->end_subject)
4266                  {
4267                  SCHECK_PARTIAL();
4268                  RRETURN(MATCH_NOMATCH);
4269                  }
4270                GETCHARINCTEST(c, eptr);
4271                cp = PRIV(ucd_caseless_sets) + prop_value;
4272                for (;;)
4273                  {
4274                  if (c < *cp)
4275                    { if (prop_fail_result) break; else { RRETURN(MATCH_NOMATCH); } }
4276                  if (c == *cp++)
4277                    { if (prop_fail_result) { RRETURN(MATCH_NOMATCH); } else break; }
4278                  }
4279                }
4280              break;
4281    
4282              case PT_UCNC:
4283              for (i = 1; i <= min; i++)
4284                {
4285                if (eptr >= md->end_subject)
4286                  {
4287                  SCHECK_PARTIAL();
4288                  RRETURN(MATCH_NOMATCH);
4289                  }
4290                GETCHARINCTEST(c, eptr);
4291                if ((c == CHAR_DOLLAR_SIGN || c == CHAR_COMMERCIAL_AT ||
4292                     c == CHAR_GRAVE_ACCENT || (c >= 0xa0 && c <= 0xd7ff) ||
4293                     c >= 0xe000) == prop_fail_result)
4294                  RRETURN(MATCH_NOMATCH);
4295                }
4296              break;
4297    
4298            /* This should not occur */            /* This should not occur */
4299    
4300            default:            default:
# Line 4155  for (;;) Line 4314  for (;;)
4314              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4315              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4316              }              }
4317            GETCHARINCTEST(c, eptr);            else
           if (UCD_CATEGORY(c) == ucp_M) RRETURN(MATCH_NOMATCH);  
           while (eptr < md->end_subject)  
4318              {              {
4319              int len = 1;              int lgb, rgb;
4320              if (!utf) c = *eptr; else { GETCHARLEN(c, eptr, len); }              GETCHARINCTEST(c, eptr);
4321              if (UCD_CATEGORY(c) != ucp_M) break;              lgb = UCD_GRAPHBREAK(c);
4322              eptr += len;             while (eptr < md->end_subject)
4323                  {
4324                  int len = 1;
4325                  if (!utf) c = *eptr; else { GETCHARLEN(c, eptr, len); }
4326                  rgb = UCD_GRAPHBREAK(c);
4327                  if ((PRIV(ucp_gbtable)[lgb] & (1 << rgb)) == 0) break;
4328                  lgb = rgb;
4329                  eptr += len;
4330                  }
4331              }              }
4332              CHECK_PARTIAL();
4333            }            }
4334          }          }
4335    
# Line 4184  for (;;) Line 4350  for (;;)
4350              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4351              }              }
4352            if (IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH);            if (IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH);
4353              if (md->partial != 0 &&
4354                  eptr + 1 >= md->end_subject &&
4355                  NLBLOCK->nltype == NLTYPE_FIXED &&
4356                  NLBLOCK->nllen == 2 &&
4357                  RAWUCHAR(eptr) == NLBLOCK->nl[0])
4358                {
4359                md->hitend = TRUE;
4360                if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
4361                }
4362            eptr++;            eptr++;
4363            ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);            ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);
4364            }            }
# Line 4220  for (;;) Line 4395  for (;;)
4395              {              {
4396              default: RRETURN(MATCH_NOMATCH);              default: RRETURN(MATCH_NOMATCH);
4397    
4398              case 0x000d:              case CHAR_CR:
4399              if (eptr < md->end_subject && *eptr == 0x0a) eptr++;              if (eptr < md->end_subject && RAWUCHAR(eptr) == CHAR_LF) eptr++;
4400              break;              break;
4401    
4402              case 0x000a:              case CHAR_LF:
4403              break;              break;
4404    
4405              case 0x000b:              case CHAR_VT:
4406              case 0x000c:              case CHAR_FF:
4407              case 0x0085:              case CHAR_NEL:
4408    #ifndef EBCDIC
4409              case 0x2028:              case 0x2028:
4410              case 0x2029:              case 0x2029:
4411    #endif  /* Not EBCDIC */
4412              if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);              if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);
4413              break;              break;
4414              }              }
# Line 4249  for (;;) Line 4426  for (;;)
4426            GETCHARINC(c, eptr);            GETCHARINC(c, eptr);
4427            switch(c)            switch(c)
4428              {              {
4429                HSPACE_CASES: RRETURN(MATCH_NOMATCH);  /* Byte and multibyte cases */
4430              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);  
4431              }              }
4432            }            }
4433          break;          break;
# Line 4285  for (;;) Line 4443  for (;;)
4443            GETCHARINC(c, eptr);            GETCHARINC(c, eptr);
4444            switch(c)            switch(c)
4445              {              {
4446                HSPACE_CASES: break;  /* Byte and multibyte cases */
4447              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;  
4448              }              }
4449            }            }
4450          break;          break;
# Line 4321  for (;;) Line 4460  for (;;)
4460            GETCHARINC(c, eptr);            GETCHARINC(c, eptr);
4461            switch(c)            switch(c)
4462              {              {
4463                VSPACE_CASES: RRETURN(MATCH_NOMATCH);
4464              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);  
4465              }              }
4466            }            }
4467          break;          break;
# Line 4345  for (;;) Line 4477  for (;;)
4477            GETCHARINC(c, eptr);            GETCHARINC(c, eptr);
4478            switch(c)            switch(c)
4479              {              {
4480                VSPACE_CASES: break;
4481              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;  
4482              }              }
4483            }            }
4484          break;          break;
# Line 4375  for (;;) Line 4500  for (;;)
4500          case OP_DIGIT:          case OP_DIGIT:
4501          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
4502            {            {
4503              pcre_uint32 cc;
4504            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
4505              {              {
4506              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4507              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4508              }              }
4509            if (*eptr >= 128 || (md->ctypes[*eptr] & ctype_digit) == 0)            cc = RAWUCHAR(eptr);
4510              if (cc >= 128 || (md->ctypes[cc] & ctype_digit) == 0)
4511              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4512            eptr++;            eptr++;
4513            /* No need to skip more bytes - we know it's a 1-byte character */            /* No need to skip more bytes - we know it's a 1-byte character */
# Line 4390  for (;;) Line 4517  for (;;)
4517          case OP_NOT_WHITESPACE:          case OP_NOT_WHITESPACE:
4518          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
4519            {            {
4520              pcre_uint32 cc;
4521            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
4522              {              {
4523              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4524              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4525              }              }
4526            if (*eptr < 128 && (md->ctypes[*eptr] & ctype_space) != 0)            cc = RAWUCHAR(eptr);
4527              if (cc < 128 && (md->ctypes[cc] & ctype_space) != 0)
4528              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4529            eptr++;            eptr++;
4530            ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);            ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);
# Line 4405  for (;;) Line 4534  for (;;)
4534          case OP_WHITESPACE:          case OP_WHITESPACE:
4535          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
4536            {            {
4537              pcre_uint32 cc;
4538            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
4539              {              {
4540              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4541              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4542              }              }
4543            if (*eptr >= 128 || (md->ctypes[*eptr] & ctype_space) == 0)            cc = RAWUCHAR(eptr);
4544              if (cc >= 128 || (md->ctypes[cc] & ctype_space) == 0)
4545              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4546            eptr++;            eptr++;
4547            /* No need to skip more bytes - we know it's a 1-byte character */            /* No need to skip more bytes - we know it's a 1-byte character */
# Line 4420  for (;;) Line 4551  for (;;)
4551          case OP_NOT_WORDCHAR:          case OP_NOT_WORDCHAR:
4552          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
4553            {            {
4554              pcre_uint32 cc;
4555            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
4556              {              {
4557              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4558              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4559              }              }
4560            if (*eptr < 128 && (md->ctypes[*eptr] & ctype_word) != 0)            cc = RAWUCHAR(eptr);
4561              if (cc < 128 && (md->ctypes[cc] & ctype_word) != 0)
4562              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4563            eptr++;            eptr++;
4564            ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);            ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);
# Line 4435  for (;;) Line 4568  for (;;)
4568          case OP_WORDCHAR:          case OP_WORDCHAR:
4569          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
4570            {            {
4571              pcre_uint32 cc;
4572            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
4573              {              {
4574              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4575              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4576              }              }
4577            if (*eptr >= 128 || (md->ctypes[*eptr] & ctype_word) == 0)            cc = RAWUCHAR(eptr);
4578              if (cc >= 128 || (md->ctypes[cc] & ctype_word) == 0)
4579              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4580            eptr++;            eptr++;
4581            /* No need to skip more bytes - we know it's a 1-byte character */            /* No need to skip more bytes - we know it's a 1-byte character */
# Line 4468  for (;;) Line 4603  for (;;)
4603              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4604              }              }
4605            if (IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH);            if (IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH);
4606              if (md->partial != 0 &&
4607                  eptr + 1 >= md->end_subject &&
4608                  NLBLOCK->nltype == NLTYPE_FIXED &&
4609                  NLBLOCK->nllen == 2 &&
4610                  *eptr == NLBLOCK->nl[0])
4611                {
4612                md->hitend = TRUE;
4613                if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
4614                }
4615            eptr++;            eptr++;
4616            }            }
4617          break;          break;
# Line 4502  for (;;) Line 4646  for (;;)
4646              {              {
4647              default: RRETURN(MATCH_NOMATCH);              default: RRETURN(MATCH_NOMATCH);
4648    
4649              case 0x000d:              case CHAR_CR:
4650              if (eptr < md->end_subject && *eptr == 0x0a) eptr++;              if (eptr < md->end_subject && *eptr == CHAR_LF) eptr++;
4651              break;              break;
4652    
4653              case 0x000a:              case CHAR_LF:
4654              break;              break;
4655    
4656              case 0x000b:              case CHAR_VT:
4657              case 0x000c:              case CHAR_FF:
4658              case 0x0085:              case CHAR_NEL:
4659  #ifdef COMPILE_PCRE16  #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
4660              case 0x2028:              case 0x2028:
4661              case 0x2029:              case 0x2029:
4662  #endif  #endif
# Line 4533  for (;;) Line 4677  for (;;)
4677            switch(*eptr++)            switch(*eptr++)
4678              {              {
4679              default: break;              default: break;
4680              case 0x09:      /* HT */              HSPACE_BYTE_CASES:
4681              case 0x20:      /* SPACE */  #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
4682              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 */  
4683  #endif  #endif
4684              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4685              }              }
# Line 4570  for (;;) Line 4697  for (;;)
4697            switch(*eptr++)            switch(*eptr++)
4698              {              {
4699              default: RRETURN(MATCH_NOMATCH);              default: RRETURN(MATCH_NOMATCH);
4700              case 0x09:      /* HT */              HSPACE_BYTE_CASES:
4701              case 0x20:      /* SPACE */  #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
4702              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 */  
4703  #endif  #endif
4704              break;              break;
4705              }              }
# Line 4606  for (;;) Line 4716  for (;;)
4716              }              }
4717            switch(*eptr++)            switch(*eptr++)
4718              {              {
4719              default: break;              VSPACE_BYTE_CASES:
4720              case 0x0a:      /* LF */  #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
4721              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 */  
4722  #endif  #endif
4723              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4724                default: break;
4725              }              }
4726            }            }
4727          break;          break;
# Line 4632  for (;;) Line 4737  for (;;)
4737            switch(*eptr++)            switch(*eptr++)
4738              {              {
4739              default: RRETURN(MATCH_NOMATCH);              default: RRETURN(MATCH_NOMATCH);
4740              case 0x0a:      /* LF */              VSPACE_BYTE_CASES:
4741              case 0x0b:      /* VT */  #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
4742              case 0x0c:      /* FF */              VSPACE_MULTIBYTE_CASES:
             case 0x0d:      /* CR */  
             case 0x85:      /* NEL */  
 #ifdef COMPILE_PCRE16  
             case 0x2028:    /* LINE SEPARATOR */  
             case 0x2029:    /* PARAGRAPH SEPARATOR */  
4743  #endif  #endif
4744              break;              break;
4745              }              }
# Line 4917  for (;;) Line 5017  for (;;)
5017              }              }
5018            /* Control never gets here */            /* Control never gets here */
5019    
5020            /* This should never occur */            case PT_CLIST:
5021              for (fi = min;; fi++)
5022                {
5023                const pcre_uint32 *cp;
5024                RMATCH(eptr, ecode, offset_top, md, eptrb, RM67);
5025                if (rrc != MATCH_NOMATCH) RRETURN(rrc);
5026                if (fi >= max) RRETURN(MATCH_NOMATCH);
5027                if (eptr >= md->end_subject)
5028                  {
5029                  SCHECK_PARTIAL();
5030                  RRETURN(MATCH_NOMATCH);
5031                  }
5032                GETCHARINCTEST(c, eptr);
5033                cp = PRIV(ucd_caseless_sets) + prop_value;
5034                for (;;)
5035                  {
5036                  if (c < *cp)
5037                    { if (prop_fail_result) break; else { RRETURN(MATCH_NOMATCH); } }
5038                  if (c == *cp++)
5039                    { if (prop_fail_result) { RRETURN(MATCH_NOMATCH); } else break; }
5040                  }
5041                }
5042              /* Control never gets here */
5043    
5044              case PT_UCNC:
5045              for (fi = min;; fi++)
5046                {
5047                RMATCH(eptr, ecode, offset_top, md, eptrb, RM68);
5048                if (rrc != MATCH_NOMATCH) RRETURN(rrc);
5049                if (fi >= max) RRETURN(MATCH_NOMATCH);
5050                if (eptr >= md->end_subject)
5051                  {
5052                  SCHECK_PARTIAL();
5053                  RRETURN(MATCH_NOMATCH);
5054                  }
5055                GETCHARINCTEST(c, eptr);
5056                if ((c == CHAR_DOLLAR_SIGN || c == CHAR_COMMERCIAL_AT ||
5057                     c == CHAR_GRAVE_ACCENT || (c >= 0xa0 && c <= 0xd7ff) ||
5058                     c >= 0xe000) == prop_fail_result)
5059                  RRETURN(MATCH_NOMATCH);
5060                }
5061              /* Control never gets here */
5062    
5063              /* This should never occur */
5064            default:            default:
5065            RRETURN(PCRE_ERROR_INTERNAL);            RRETURN(PCRE_ERROR_INTERNAL);
5066            }            }
# Line 4939  for (;;) Line 5081  for (;;)
5081              SCHECK_PARTIAL();              SCHECK_PARTIAL();
5082              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
5083              }              }
5084            GETCHARINCTEST(c, eptr);            else
           if (UCD_CATEGORY(c) == ucp_M) RRETURN(MATCH_NOMATCH);  
           while (eptr < md->end_subject)  
5085              {              {
5086              int len = 1;              int lgb, rgb;
5087              if (!utf) c = *eptr; else { GETCHARLEN(c, eptr, len); }              GETCHARINCTEST(c, eptr);
5088              if (UCD_CATEGORY(c) != ucp_M) break;              lgb = UCD_GRAPHBREAK(c);
5089              eptr += len;              while (eptr < md->end_subject)
5090                  {
5091                  int len = 1;
5092                  if (!utf) c = *eptr; else { GETCHARLEN(c, eptr, len); }
5093                  rgb = UCD_GRAPHBREAK(c);
5094                  if ((PRIV(ucp_gbtable)[lgb] & (1 << rgb)) == 0) break;
5095                  lgb = rgb;
5096                  eptr += len;
5097                  }
5098              }              }
5099              CHECK_PARTIAL();
5100            }            }
5101          }          }
5102        else        else
# Line 4971  for (;;) Line 5120  for (;;)
5120            GETCHARINC(c, eptr);            GETCHARINC(c, eptr);
5121            switch(ctype)            switch(ctype)
5122              {              {
5123              case OP_ANY:        /* This is the non-NL case */              case OP_ANY:               /* This is the non-NL case */
5124                if (md->partial != 0 &&    /* Take care with CRLF partial */
5125                    eptr >= md->end_subject &&
5126                    NLBLOCK->nltype == NLTYPE_FIXED &&
5127                    NLBLOCK->nllen == 2 &&
5128                    c == NLBLOCK->nl[0])
5129                  {
5130                  md->hitend = TRUE;
5131                  if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
5132                  }
5133                break;
5134    
5135              case OP_ALLANY:              case OP_ALLANY:
5136              case OP_ANYBYTE:              case OP_ANYBYTE:
5137              break;              break;
# Line 4980  for (;;) Line 5140  for (;;)
5140              switch(c)              switch(c)
5141                {                {
5142                default: RRETURN(MATCH_NOMATCH);                default: RRETURN(MATCH_NOMATCH);
5143                case 0x000d:                case CHAR_CR:
5144                if (eptr < md->end_subject && *eptr == 0x0a) eptr++;                if (eptr < md->end_subject && RAWUCHAR(eptr) == CHAR_LF) eptr++;
5145                break;                break;
5146                case 0x000a:  
5147                  case CHAR_LF:
5148                break;                break;
5149    
5150                case 0x000b:                case CHAR_VT:
5151                case 0x000c:                case CHAR_FF:
5152                case 0x0085:                case CHAR_NEL:
5153    #ifndef EBCDIC
5154                case 0x2028:                case 0x2028:
5155                case 0x2029:                case 0x2029:
5156    #endif  /* Not EBCDIC */
5157                if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);                if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);
5158                break;                break;
5159                }                }
# Line 4999  for (;;) Line 5162  for (;;)
5162              case OP_NOT_HSPACE:              case OP_NOT_HSPACE:
5163              switch(c)              switch(c)
5164                {                {
5165                  HSPACE_CASES: RRETURN(MATCH_NOMATCH);
5166                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);  
5167                }                }
5168              break;              break;
5169    
5170              case OP_HSPACE:              case OP_HSPACE:
5171              switch(c)              switch(c)
5172                {                {
5173                  HSPACE_CASES: break;
5174                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;  
5175                }                }
5176              break;              break;
5177    
5178              case OP_NOT_VSPACE:              case OP_NOT_VSPACE:
5179              switch(c)              switch(c)
5180                {                {
5181                  VSPACE_CASES: RRETURN(MATCH_NOMATCH);
5182                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);  
5183                }                }
5184              break;              break;
5185    
5186              case OP_VSPACE:              case OP_VSPACE:
5187              switch(c)              switch(c)
5188                {                {
5189                  VSPACE_CASES: break;
5190                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;  
5191                }                }
5192              break;              break;
5193    
# Line 5134  for (;;) Line 5245  for (;;)
5245            c = *eptr++;            c = *eptr++;
5246            switch(ctype)            switch(ctype)
5247              {              {
5248              case OP_ANY:     /* This is the non-NL case */              case OP_ANY:               /* This is the non-NL case */
5249                if (md->partial != 0 &&    /* Take care with CRLF partial */
5250                    eptr >= md->end_subject &&
5251                    NLBLOCK->nltype == NLTYPE_FIXED &&
5252                    NLBLOCK->nllen == 2 &&
5253                    c == NLBLOCK->nl[0])
5254                  {
5255                  md->hitend = TRUE;
5256                  if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
5257                  }
5258                break;
5259    
5260              case OP_ALLANY:              case OP_ALLANY:
5261              case OP_ANYBYTE:              case OP_ANYBYTE:
5262              break;              break;
# Line 5143  for (;;) Line 5265  for (;;)
5265              switch(c)              switch(c)
5266                {                {
5267                default: RRETURN(MATCH_NOMATCH);                default: RRETURN(MATCH_NOMATCH);
5268                case 0x000d:                case CHAR_CR:
5269                if (eptr < md->end_subject && *eptr == 0x0a) eptr++;                if (eptr < md->end_subject && *eptr == CHAR_LF) eptr++;
5270                break;                break;
5271    
5272                case 0x000a:                case CHAR_LF:
5273                break;                break;
5274    
5275                case 0x000b:                case CHAR_VT:
5276                case 0x000c:                case CHAR_FF:
5277                case 0x0085:                case CHAR_NEL:
5278  #ifdef COMPILE_PCRE16  #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
5279                case 0x2028:                case 0x2028:
5280                case 0x2029:                case 0x2029:
5281  #endif  #endif
# Line 5166  for (;;) Line 5288  for (;;)
5288              switch(c)              switch(c)
5289                {                {
5290                default: break;                default: break;
5291                case 0x09:      /* HT */                HSPACE_BYTE_CASES:
5292                case 0x20:      /* SPACE */  #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
5293                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 */  
5294  #endif  #endif
5295                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
5296                }                }
# Line 5195  for (;;) Line 5300  for (;;)
5300              switch(c)              switch(c)
5301                {                {
5302                default: RRETURN(MATCH_NOMATCH);                default: RRETURN(MATCH_NOMATCH);
5303                case 0x09:      /* HT */                HSPACE_BYTE_CASES:
5304                case 0x20:      /* SPACE */  #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
5305                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 */  
5306  #endif  #endif
5307                break;                break;
5308                }                }
# Line 5224  for (;;) Line 5312  for (;;)
5312              switch(c)              switch(c)
5313                {                {
5314                default: break;                default: break;
5315                case 0x0a:      /* LF */                VSPACE_BYTE_CASES:
5316                case 0x0b:      /* VT */  #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
5317                case 0x0c:      /* FF */                VSPACE_MULTIBYTE_CASES:
               case 0x0d:      /* CR */  
               case 0x85:      /* NEL */  
 #ifdef COMPILE_PCRE16  
               case 0x2028:    /* LINE SEPARATOR */  
               case 0x2029:    /* PARAGRAPH SEPARATOR */  
5318  #endif  #endif
5319                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
5320                }                }
# Line 5241  for (;;) Line 5324  for (;;)
5324              switch(c)              switch(c)
5325                {                {
5326                default: RRETURN(MATCH_NOMATCH);                default: RRETURN(MATCH_NOMATCH);
5327                case 0x0a:      /* LF */                VSPACE_BYTE_CASES:
5328                case 0x0b:      /* VT */  #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
5329                case 0x0c:      /* FF */                VSPACE_MULTIBYTE_CASES:
               case 0x0d:      /* CR */  
               case 0x85:      /* NEL */  
 #ifdef COMPILE_PCRE16  
               case 0x2028:    /* LINE SEPARATOR */  
               case 0x2029:    /* PARAGRAPH SEPARATOR */  
5330  #endif  #endif
5331                break;                break;
5332                }                }
# Line 5452  for (;;) Line 5530  for (;;)
5530              }              }
5531            break;            break;
5532    
5533              case PT_CLIST:
5534              for (i = min; i < max; i++)
5535                {
5536                const pcre_uint32 *cp;
5537                int len = 1;
5538                if (eptr >= md->end_subject)
5539                  {
5540                  SCHECK_PARTIAL();
5541                  break;
5542                  }
5543                GETCHARLENTEST(c, eptr, len);
5544                cp = PRIV(ucd_caseless_sets) + prop_value;
5545                for (;;)
5546                  {
5547                  if (c < *cp)
5548                    { if (prop_fail_result) break; else goto GOT_MAX; }
5549                  if (c == *cp++)
5550                    { if (prop_fail_result) goto GOT_MAX; else break; }
5551                  }
5552                eptr += len;
5553                }
5554              GOT_MAX:
5555              break;
5556    
5557              case PT_UCNC:
5558              for (i = min; i < max; i++)
5559                {
5560                int len = 1;
5561                if (eptr >= md->end_subject)
5562                  {
5563                  SCHECK_PARTIAL();
5564                  break;
5565                  }
5566                GETCHARLENTEST(c, eptr, len);
5567                if ((c == CHAR_DOLLAR_SIGN || c == CHAR_COMMERCIAL_AT ||
5568                     c == CHAR_GRAVE_ACCENT || (c >= 0xa0 && c <= 0xd7ff) ||
5569                     c >= 0xe000) == prop_fail_result)
5570                  break;
5571                eptr += len;
5572                }
5573              break;
5574    
5575            default:            default:
5576            RRETURN(PCRE_ERROR_INTERNAL);            RRETURN(PCRE_ERROR_INTERNAL);
5577            }            }
# Line 5475  for (;;) Line 5595  for (;;)
5595          {          {
5596          for (i = min; i < max; i++)          for (i = min; i < max; i++)
5597            {            {
           int len = 1;  
5598            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
5599              {              {
5600              SCHECK_PARTIAL();              SCHECK_PARTIAL();
5601              break;              break;
5602              }              }
5603            if (!utf) c = *eptr; else { GETCHARLEN(c, eptr, len); }            else
           if (UCD_CATEGORY(c) == ucp_M) break;  
           eptr += len;  
           while (eptr < md->end_subject)  
5604              {              {
5605              len = 1;              int lgb, rgb;
5606              if (!utf) c = *eptr; else { GETCHARLEN(c, eptr, len); }              GETCHARINCTEST(c, eptr);
5607              if (UCD_CATEGORY(c) != ucp_M) break;              lgb = UCD_GRAPHBREAK(c);
5608              eptr += len;              while (eptr < md->end_subject)
5609                  {
5610                  int len = 1;
5611                  if (!utf) c = *eptr; else { GETCHARLEN(c, eptr, len); }
5612                  rgb = UCD_GRAPHBREAK(c);
5613                  if ((PRIV(ucp_gbtable)[lgb] & (1 << rgb)) == 0) break;
5614                  lgb = rgb;
5615                  eptr += len;
5616                  }
5617              }              }
5618              CHECK_PARTIAL();
5619            }            }
5620    
5621          /* eptr is now past the end of the maximum run */          /* eptr is now past the end of the maximum run */
# Line 5534  for (;;) Line 5659  for (;;)
5659                  break;                  break;
5660                  }                  }
5661                if (IS_NEWLINE(eptr)) break;                if (IS_NEWLINE(eptr)) break;
5662                  if (md->partial != 0 &&    /* Take care with CRLF partial */
5663                      eptr + 1 >= md->end_subject &&
5664                      NLBLOCK->nltype == NLTYPE_FIXED &&
5665                      NLBLOCK->nllen == 2 &&
5666                      RAWUCHAR(eptr) == NLBLOCK->nl[0])
5667                    {
5668                    md->hitend = TRUE;
5669                    if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
5670                    }
5671                eptr++;                eptr++;
5672                ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);                ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);
5673                }                }
# Line 5551  for (;;) Line 5685  for (;;)
5685                  break;                  break;
5686                  }                  }
5687                if (IS_NEWLINE(eptr)) break;                if (IS_NEWLINE(eptr)) break;
5688                  if (md->partial != 0 &&    /* Take care with CRLF partial */
5689                      eptr + 1 >= md->end_subject &&
5690                      NLBLOCK->nltype == NLTYPE_FIXED &&
5691                      NLBLOCK->nllen == 2 &&
5692                      RAWUCHAR(eptr) == NLBLOCK->nl[0])
5693                    {
5694                    md->hitend = TRUE;
5695                    if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
5696                    }
5697                eptr++;                eptr++;
5698                ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);                ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);
5699                }                }
# Line 5600  for (;;) Line 5743  for (;;)
5743                break;                break;
5744                }                }
5745              GETCHARLEN(c, eptr, len);              GETCHARLEN(c, eptr, len);
5746              if (c == 0x000d)              if (c == CHAR_CR)
5747                {                {
5748                if (++eptr >= md->end_subject) break;                if (++eptr >= md->end_subject) break;
5749                if (*eptr == 0x000a) eptr++;                if (RAWUCHAR(eptr) == CHAR_LF) eptr++;
5750                }                }
5751              else              else
5752                {                {
5753                if (c != 0x000a &&                if (c != CHAR_LF &&
5754                    (md->bsr_anycrlf ||                    (md->bsr_anycrlf ||
5755                     (c != 0x000b && c != 0x000c &&                     (c != CHAR_VT && c != CHAR_FF && c != CHAR_NEL
5756                      c != 0x0085 && c != 0x2028 && c != 0x2029)))  #ifndef EBCDIC
5757                        && c != 0x2028 && c != 0x2029
5758    #endif  /* Not EBCDIC */
5759                        )))
5760                  break;                  break;
5761                eptr += len;                eptr += len;
5762                }                }
# Line 5631  for (;;) Line 5777  for (;;)
5777              GETCHARLEN(c, eptr, len);              GETCHARLEN(c, eptr, len);
5778              switch(c)              switch(c)
5779                {                {
5780                  HSPACE_CASES: gotspace = TRUE; break;
5781                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;  
5782                }                }
5783              if (gotspace == (ctype == OP_NOT_HSPACE)) break;              if (gotspace == (ctype == OP_NOT_HSPACE)) break;
5784              eptr += len;              eptr += len;
# Line 5673  for (;;) Line 5799  for (;;)
5799              GETCHARLEN(c, eptr, len);              GETCHARLEN(c, eptr, len);
5800              switch(c)              switch(c)
5801                {                {
5802                  VSPACE_CASES: gotspace = TRUE; break;
5803                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;  
5804                }                }
5805              if (gotspace == (ctype == OP_NOT_VSPACE)) break;              if (gotspace == (ctype == OP_NOT_VSPACE)) break;
5806              eptr += len;              eptr += len;
# Line 5796  for (;;) Line 5914  for (;;)
5914            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
5915            if (eptr-- == pp) break;        /* Stop if tried at original pos */            if (eptr-- == pp) break;        /* Stop if tried at original pos */
5916            BACKCHAR(eptr);            BACKCHAR(eptr);
5917            if (ctype == OP_ANYNL && eptr > pp  && *eptr == '\n' &&            if (ctype == OP_ANYNL && eptr > pp  && RAWUCHAR(eptr) == CHAR_NL &&
5918                eptr[-1] == '\r') eptr--;                RAWUCHAR(eptr - 1) == CHAR_CR) eptr--;
5919            }            }
5920          }          }
5921        else        else
# Line 5815  for (;;) Line 5933  for (;;)
5933                break;                break;
5934                }                }
5935              if (IS_NEWLINE(eptr)) break;              if (IS_NEWLINE(eptr)) break;
5936                if (md->partial != 0 &&    /* Take care with CRLF partial */
5937                    eptr + 1 >= md->end_subject &&
5938                    NLBLOCK->nltype == NLTYPE_FIXED &&
5939                    NLBLOCK->nllen == 2 &&
5940                    *eptr == NLBLOCK->nl[0])
5941                  {
5942                  md->hitend = TRUE;
5943                  if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
5944                  }
5945              eptr++;              eptr++;
5946              }              }
5947            break;            break;
# Line 5839  for (;;) Line 5966  for (;;)
5966                break;                break;
5967                }                }
5968              c = *eptr;              c = *eptr;
5969              if (c == 0x000d)              if (c == CHAR_CR)
5970                {                {
5971                if (++eptr >= md->end_subject) break;                if (++eptr >= md->end_subject) break;
5972                if (*eptr == 0x000a) eptr++;                if (*eptr == CHAR_LF) eptr++;
5973                }                }
5974              else              else
5975                {                {
5976                if (c != 0x000a && (md->bsr_anycrlf ||                if (c != CHAR_LF && (md->bsr_anycrlf ||
5977                  (c != 0x000b && c != 0x000c && c != 0x0085                   (c != CHAR_VT && c != CHAR_FF && c != CHAR_NEL
5978  #ifdef COMPILE_PCRE16  #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
5979                  && c != 0x2028 && c != 0x2029                   && c != 0x2028 && c != 0x2029
5980  #endif  #endif
5981                  ))) break;                   ))) break;
5982                eptr++;                eptr++;
5983                }                }
5984              }              }
# Line 5865  for (;;) Line 5992  for (;;)
5992                SCHECK_PARTIAL();                SCHECK_PARTIAL();
5993                break;                break;
5994                }                }
5995              c = *eptr;              switch(*eptr)
5996              if (c == 0x09 || c == 0x20 || c == 0xa0                {
5997  #ifdef COMPILE_PCRE16                default: eptr++; break;
5998                || c == 0x1680 || c == 0x180e || (c >= 0x2000 && c <= 0x200A)                HSPACE_BYTE_CASES:
5999                || c == 0x202f || c == 0x205f || c == 0x3000  #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
6000                  HSPACE_MULTIBYTE_CASES:
6001  #endif  #endif
6002                ) break;                goto ENDLOOP00;
6003              eptr++;                }
6004              }              }
6005              ENDLOOP00:
6006            break;            break;
6007    
6008            case OP_HSPACE:            case OP_HSPACE:
# Line 5884  for (;;) Line 6013  for (;;)
6013                SCHECK_PARTIAL();                SCHECK_PARTIAL();
6014                break;                break;
6015                }                }
6016              c = *eptr;              switch(*eptr)
6017              if (c != 0x09 && c != 0x20 && c != 0xa0                {
6018  #ifdef COMPILE_PCRE16                default: goto ENDLOOP01;
6019                && c != 0x1680 && c != 0x180e && (c < 0x2000 || c > 0x200A)                HSPACE_BYTE_CASES:
6020                && c != 0x202f && c != 0x205f && c != 0x3000  #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
6021                  HSPACE_MULTIBYTE_CASES:
6022  #endif  #endif
6023                ) break;                eptr++; break;
6024              eptr++;                }
6025              }              }
6026              ENDLOOP01:
6027            break;            break;
6028    
6029            case OP_NOT_VSPACE:            case OP_NOT_VSPACE:
# Line 5903  for (;;) Line 6034  for (;;)
6034                SCHECK_PARTIAL();                SCHECK_PARTIAL();
6035                break;                break;
6036                }                }
6037              c = *eptr;              switch(*eptr)
6038              if (c == 0x0a || c == 0x0b || c == 0x0c || c == 0x0d || c == 0x85                {
6039  #ifdef COMPILE_PCRE16                default: eptr++; break;
6040                || c == 0x2028 || c == 0x2029                VSPACE_BYTE_CASES:
6041    #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
6042                  VSPACE_MULTIBYTE_CASES:
6043  #endif  #endif
6044                ) break;                goto ENDLOOP02;
6045              eptr++;                }
6046              }              }
6047              ENDLOOP02:
6048            break;            break;
6049    
6050            case OP_VSPACE:            case OP_VSPACE:
# Line 5921  for (;;) Line 6055  for (;;)
6055                SCHECK_PARTIAL();                SCHECK_PARTIAL();
6056                break;                break;
6057                }                }
6058              c = *eptr;              switch(*eptr)
6059              if (c != 0x0a && c != 0x0b && c != 0x0c && c != 0x0d && c != 0x85                {
6060  #ifdef COMPILE_PCRE16                default: goto ENDLOOP03;
6061                && c != 0x2028 && c != 0x2029                VSPACE_BYTE_CASES:
6062    #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
6063                  VSPACE_MULTIBYTE_CASES:
6064  #endif  #endif
6065                ) break;                eptr++; break;
6066              eptr++;                }
6067              }              }
6068              ENDLOOP03:
6069            break;            break;
6070    
6071            case OP_NOT_DIGIT:            case OP_NOT_DIGIT:
# Line 6025  for (;;) Line 6162  for (;;)
6162            RMATCH(eptr, ecode, offset_top, md, eptrb, RM47);            RMATCH(eptr, ecode, offset_top, md, eptrb, RM47);
6163            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
6164            eptr--;            eptr--;
6165            if (ctype == OP_ANYNL && eptr > pp  && *eptr == '\n' &&            if (ctype == OP_ANYNL && eptr > pp  && *eptr == CHAR_LF &&
6166                eptr[-1] == '\r') eptr--;                eptr[-1] == CHAR_CR) eptr--;
6167            }            }
6168          }          }
6169    
# Line 6076  switch (frame->Xwhere) Line 6213  switch (frame->Xwhere)
6213    LBL(32) LBL(34) LBL(42) LBL(46)    LBL(32) LBL(34) LBL(42) LBL(46)
6214  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
6215    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)
6216    LBL(59) LBL(60) LBL(61) LBL(62)    LBL(59) LBL(60) LBL(61) LBL(62) LBL(67) LBL(68)
6217  #endif  /* SUPPORT_UCP */  #endif  /* SUPPORT_UCP */
6218  #endif  /* SUPPORT_UTF */  #endif  /* SUPPORT_UTF */
6219    default:    default:
6220    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);  
   
6221    return PCRE_ERROR_INTERNAL;    return PCRE_ERROR_INTERNAL;
6222    }    }
6223  #undef LBL  #undef LBL
# Line 6145  Undefine all the macros that were define Line 6279  Undefine all the macros that were define
6279  ***************************************************************************/  ***************************************************************************/
6280    
6281    
6282    #ifdef NO_RECURSE
6283    /*************************************************
6284    *          Release allocated heap frames         *
6285    *************************************************/
6286    
6287    /* This function releases all the allocated frames. The base frame is on the
6288    machine stack, and so must not be freed.
6289    
6290    Argument: the address of the base frame
6291    Returns:  nothing
6292    */
6293    
6294    static void
6295    release_match_heapframes (heapframe *frame_base)
6296    {
6297    heapframe *nextframe = frame_base->Xnextframe;
6298    while (nextframe != NULL)
6299      {
6300      heapframe *oldframe = nextframe;
6301      nextframe = nextframe->Xnextframe;
6302      (PUBL(stack_free))(oldframe);
6303      }
6304    }
6305    #endif
6306    
6307    
6308  /*************************************************  /*************************************************
6309  *         Execute a Regular Expression           *  *         Execute a Regular Expression           *
# Line 6170  Returns:          > 0 => success; value Line 6329  Returns:          > 0 => success; value
6329                   < -1 => some kind of unexpected problem                   < -1 => some kind of unexpected problem
6330  */  */
6331    
6332  #ifdef COMPILE_PCRE8  #if defined COMPILE_PCRE8
6333  PCRE_EXP_DEFN int PCRE_CALL_CONVENTION  PCRE_EXP_DEFN int PCRE_CALL_CONVENTION
6334  pcre_exec(const pcre *argument_re, const pcre_extra *extra_data,  pcre_exec(const pcre *argument_re, const pcre_extra *extra_data,
6335    PCRE_SPTR subject, int length, int start_offset, int options, int *offsets,    PCRE_SPTR subject, int length, int start_offset, int options, int *offsets,
6336    int offsetcount)    int offsetcount)
6337  #else  #elif defined COMPILE_PCRE16
6338  PCRE_EXP_DEFN int PCRE_CALL_CONVENTION  PCRE_EXP_DEFN int PCRE_CALL_CONVENTION
6339  pcre16_exec(const pcre16 *argument_re, const pcre16_extra *extra_data,  pcre16_exec(const pcre16 *argument_re, const pcre16_extra *extra_data,
6340    PCRE_SPTR16 subject, int length, int start_offset, int options, int *offsets,    PCRE_SPTR16 subject, int length, int start_offset, int options, int *offsets,
6341    int offsetcount)    int offsetcount)
6342    #elif defined COMPILE_PCRE32
6343    PCRE_EXP_DEFN int PCRE_CALL_CONVENTION
6344    pcre32_exec(const pcre32 *argument_re, const pcre32_extra *extra_data,
6345      PCRE_SPTR32 subject, int length, int start_offset, int options, int *offsets,
6346      int offsetcount)
6347  #endif  #endif
6348  {  {
6349  int rc, ocount, arg_offset_max;  int rc, ocount, arg_offset_max;
# Line 6202  const pcre_uint8 *start_bits = NULL; Line 6366  const pcre_uint8 *start_bits = NULL;
6366  PCRE_PUCHAR start_match = (PCRE_PUCHAR)subject + start_offset;  PCRE_PUCHAR start_match = (PCRE_PUCHAR)subject + start_offset;
6367  PCRE_PUCHAR end_subject;  PCRE_PUCHAR end_subject;
6368  PCRE_PUCHAR start_partial = NULL;  PCRE_PUCHAR start_partial = NULL;
6369    PCRE_PUCHAR match_partial;
6370  PCRE_PUCHAR req_char_ptr = start_match - 1;  PCRE_PUCHAR req_char_ptr = start_match - 1;
6371    
6372  const pcre_study_data *study;  const pcre_study_data *study;
6373  const REAL_PCRE *re = (const REAL_PCRE *)argument_re;  const REAL_PCRE *re = (const REAL_PCRE *)argument_re;
6374    
6375  /* Check for the special magic call that measures the size of the stack used  #ifdef NO_RECURSE
6376  per recursive call of match(). */  heapframe frame_zero;
6377    frame_zero.Xprevframe = NULL;            /* Marks the top level */
6378    frame_zero.Xnextframe = NULL;            /* None are allocated yet */
6379    md->match_frames_base = &frame_zero;
6380    #endif
6381    
6382    /* Check for the special magic call that measures the size of the stack used
6383    per recursive call of match(). Without the funny casting for sizeof, a Windows
6384    compiler gave this error: "unary minus operator applied to unsigned type,
6385    result still unsigned". Hopefully the cast fixes that. */
6386    
6387  if (re == NULL && extra_data == NULL && subject == NULL && length == -1)  if (re == NULL && extra_data == NULL && subject == NULL && length == -999 &&
6388        start_offset == -999)
6389  #ifdef NO_RECURSE  #ifdef NO_RECURSE
6390    return -sizeof(heapframe);    return -((int)sizeof(heapframe));
6391  #else  #else
6392    return match(NULL, NULL, NULL, 0, NULL, NULL, 0);    return match(NULL, NULL, NULL, 0, NULL, NULL, 0);
6393  #endif  #endif
6394    
6395  /* Plausibility checks */  /* Plausibility checks */
6396    
6397  if ((options & ~PUBLIC_EXEC_OPTIONS) != 0) return PCRE_ERROR_BADOPTION;  if ((options & ~PUBLIC_EXEC_OPTIONS) != 0) return PCRE_ERROR_BADOPTION;
6398  if (re == NULL || subject == NULL || (offsets == NULL && offsetcount > 0))  if (re == NULL || subject == NULL || (offsets == NULL && offsetcount > 0))
6399    return PCRE_ERROR_NULL;    return PCRE_ERROR_NULL;
6400  if (offsetcount < 0) return PCRE_ERROR_BADCOUNT;  if (offsetcount < 0) return PCRE_ERROR_BADCOUNT;
6401    if (length < 0) return PCRE_ERROR_BADLENGTH;
6402  if (start_offset < 0 || start_offset > length) return PCRE_ERROR_BADOFFSET;  if (start_offset < 0 || start_offset > length) return PCRE_ERROR_BADOFFSET;
6403    
6404  /* Check that the first field in the block is the magic number. If it is not,  /* Check that the first field in the block is the magic number. If it is not,
# Line 6260  if (utf && (options & PCRE_NO_UTF8_CHECK Line 6436  if (utf && (options & PCRE_NO_UTF8_CHECK
6436        offsets[0] = erroroffset;        offsets[0] = erroroffset;
6437        offsets[1] = errorcode;        offsets[1] = errorcode;
6438        }        }
6439  #ifdef COMPILE_PCRE16  #if defined COMPILE_PCRE8
     return (errorcode <= PCRE_UTF16_ERR1 && md->partial > 1)?  
       PCRE_ERROR_SHORTUTF16 : PCRE_ERROR_BADUTF16;  
 #else  
6440      return (errorcode <= PCRE_UTF8_ERR5 && md->partial > 1)?      return (errorcode <= PCRE_UTF8_ERR5 && md->partial > 1)?
6441        PCRE_ERROR_SHORTUTF8 : PCRE_ERROR_BADUTF8;        PCRE_ERROR_SHORTUTF8 : PCRE_ERROR_BADUTF8;
6442    #elif defined COMPILE_PCRE16
6443        return (errorcode <= PCRE_UTF16_ERR1 && md->partial > 1)?
6444          PCRE_ERROR_SHORTUTF16 : PCRE_ERROR_BADUTF16;
6445    #elif defined COMPILE_PCRE32
6446        return PCRE_ERROR_BADUTF32;
6447  #endif  #endif
6448      }      }
6449    #if defined COMPILE_PCRE8 || defined COMPILE_PCRE16
6450    /* 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. */
6451    if (start_offset > 0 && start_offset < length &&    if (start_offset > 0 && start_offset < length &&
6452        NOT_FIRSTCHAR(((PCRE_PUCHAR)subject)[start_offset]))        NOT_FIRSTCHAR(((PCRE_PUCHAR)subject)[start_offset]))
6453      return PCRE_ERROR_BADUTF8_OFFSET;      return PCRE_ERROR_BADUTF8_OFFSET;
6454    #endif
6455    }    }
6456  #endif  #endif
6457    
6458  /* If the pattern was successfully studied with JIT support, run the JIT  /* If the pattern was successfully studied with JIT support, run the JIT
6459  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
6460  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
6461  an unsupported flag is set. In particular, JIT does not support partial  an unsupported flag is set. */
 matching. */  
6462    
6463  #ifdef SUPPORT_JIT  #ifdef SUPPORT_JIT
6464  if (extra_data != NULL  if (extra_data != NULL
6465      && (extra_data->flags & PCRE_EXTRA_EXECUTABLE_JIT) != 0      && (extra_data->flags & (PCRE_EXTRA_EXECUTABLE_JIT |
6466                                 PCRE_EXTRA_TABLES)) == PCRE_EXTRA_EXECUTABLE_JIT
6467      && extra_data->executable_jit != NULL      && extra_data->executable_jit != NULL
6468      && (extra_data->flags & PCRE_EXTRA_TABLES) == 0      && (options & ~PUBLIC_JIT_EXEC_OPTIONS) == 0)
6469      && (options & ~(PCRE_NO_UTF8_CHECK | PCRE_NOTBOL | PCRE_NOTEOL |    {
6470                      PCRE_NOTEMPTY | PCRE_NOTEMPTY_ATSTART)) == 0)    rc = PRIV(jit_exec)(extra_data, (const pcre_uchar *)subject, length,
6471    return PRIV(jit_exec)(re, extra_data->executable_jit,         start_offset, options, offsets, offsetcount);
6472      (const pcre_uchar *)subject, length, start_offset, options,  
6473      ((extra_data->flags & PCRE_EXTRA_MATCH_LIMIT) == 0)    /* PCRE_ERROR_NULL means that the selected normal or partial matching
6474      ? MATCH_LIMIT : extra_data->match_limit, offsets, offsetcount);    mode is not compiled. In this case we simply fallback to interpreter. */
6475    
6476      if (rc != PCRE_ERROR_JIT_BADOPTION) return rc;
6477      }
6478  #endif  #endif
6479    
6480  /* 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 6353  end_subject = md->end_subject; Line 6535  end_subject = md->end_subject;
6535  md->endonly = (re->options & PCRE_DOLLAR_ENDONLY) != 0;  md->endonly = (re->options & PCRE_DOLLAR_ENDONLY) != 0;
6536  md->use_ucp = (re->options & PCRE_UCP) != 0;  md->use_ucp = (re->options & PCRE_UCP) != 0;
6537  md->jscript_compat = (re->options & PCRE_JAVASCRIPT_COMPAT) != 0;  md->jscript_compat = (re->options & PCRE_JAVASCRIPT_COMPAT) != 0;
6538  md->ignore_skip_arg = FALSE;  md->ignore_skip_arg = 0;
6539    
6540  /* Some options are unpacked into BOOL variables in the hope that testing  /* Some options are unpacked into BOOL variables in the hope that testing
6541  them will be faster than individual option bits. */  them will be faster than individual option bits. */
# Line 6463  if (re->top_backref > 0 && re->top_backr Line 6645  if (re->top_backref > 0 && re->top_backr
6645    DPRINTF(("Got memory to hold back references\n"));    DPRINTF(("Got memory to hold back references\n"));
6646    }    }
6647  else md->offset_vector = offsets;  else md->offset_vector = offsets;
   
6648  md->offset_end = ocount;  md->offset_end = ocount;
6649  md->offset_max = (2*ocount)/3;  md->offset_max = (2*ocount)/3;
6650  md->offset_overflow = FALSE;  md->capture_last = 0;
 md->capture_last = -1;  
6651    
6652  /* Reset the working variable associated with each extraction. These should  /* Reset the working variable associated with each extraction. These should
6653  never be used unless previously set, but they get saved and restored, and so we  never be used unless previously set, but they get saved and restored, and so we
# Line 6495  if (!anchored) Line 6675  if (!anchored)
6675    if ((re->flags & PCRE_FIRSTSET) != 0)    if ((re->flags & PCRE_FIRSTSET) != 0)
6676      {      {
6677      has_first_char = TRUE;      has_first_char = TRUE;
6678      first_char = first_char2 = re->first_char;      first_char = first_char2 = (pcre_uchar)(re->first_char);
6679      if ((re->flags & PCRE_FCH_CASELESS) != 0)      if ((re->flags & PCRE_FCH_CASELESS) != 0)
6680        {        {
6681        first_char2 = TABLE_GET(first_char, md->fcc, first_char);        first_char2 = TABLE_GET(first_char, md->fcc, first_char);
# Line 6517  character" set. */ Line 6697  character" set. */
6697  if ((re->flags & PCRE_REQCHSET) != 0)  if ((re->flags & PCRE_REQCHSET) != 0)
6698    {    {
6699    has_req_char = TRUE;    has_req_char = TRUE;
6700    req_char = req_char2 = re->req_char;    req_char = req_char2 = (pcre_uchar)(re->req_char);
6701    if ((re->flags & PCRE_RCH_CASELESS) != 0)    if ((re->flags & PCRE_RCH_CASELESS) != 0)
6702      {      {
6703      req_char2 = TABLE_GET(req_char, md->fcc, req_char);      req_char2 = TABLE_GET(req_char, md->fcc, req_char);
# Line 6575  for(;;) Line 6755  for(;;)
6755    
6756      if (has_first_char)      if (has_first_char)
6757        {        {
6758          pcre_uchar smc;
6759    
6760        if (first_char != first_char2)        if (first_char != first_char2)
6761          while (start_match < end_subject &&          while (start_match < end_subject &&
6762              *start_match != first_char && *start_match != first_char2)            (smc = RAWUCHARTEST(start_match)) != first_char && smc != first_char2)
6763            start_match++;            start_match++;
6764        else        else
6765          while (start_match < end_subject && *start_match != first_char)          while (start_match < end_subject && RAWUCHARTEST(start_match) != first_char)
6766            start_match++;            start_match++;
6767        }        }
6768    
# Line 6612  for(;;) Line 6794  for(;;)
6794          if (start_match[-1] == CHAR_CR &&          if (start_match[-1] == CHAR_CR &&
6795               (md->nltype == NLTYPE_ANY || md->nltype == NLTYPE_ANYCRLF) &&               (md->nltype == NLTYPE_ANY || md->nltype == NLTYPE_ANYCRLF) &&
6796               start_match < end_subject &&               start_match < end_subject &&
6797               *start_match == CHAR_NL)               RAWUCHARTEST(start_match) == CHAR_NL)
6798            start_match++;            start_match++;
6799          }          }
6800        }        }
# Line 6623  for(;;) Line 6805  for(;;)
6805        {        {
6806        while (start_match < end_subject)        while (start_match < end_subject)
6807          {          {
6808          register unsigned int c = *start_match;          register pcre_uint32 c = RAWUCHARTEST(start_match);
6809  #ifndef COMPILE_PCRE8  #ifndef COMPILE_PCRE8
6810          if (c > 255) c = 255;          if (c > 255) c = 255;
6811  #endif  #endif
# Line 6691  for(;;) Line 6873  for(;;)
6873            {            {
6874            while (p < end_subject)            while (p < end_subject)
6875              {              {
6876              register int pp = *p++;              register pcre_uint32 pp = RAWUCHARINCTEST(p);
6877              if (pp == req_char || pp == req_char2) { p--; break; }              if (pp == req_char || pp == req_char2) { p--; break; }
6878              }              }
6879            }            }
# Line 6699  for(;;) Line 6881  for(;;)
6881            {            {
6882            while (p < end_subject)            while (p < end_subject)
6883              {              {
6884              if (*p++ == req_char) { p--; break; }              if (RAWUCHARINCTEST(p) == req_char) { p--; break; }
6885              }              }
6886            }            }
6887    
# Line 6735  for(;;) Line 6917  for(;;)
6917    md->match_call_count = 0;    md->match_call_count = 0;
6918    md->match_function_type = 0;    md->match_function_type = 0;
6919    md->end_offset_top = 0;    md->end_offset_top = 0;
6920      md->skip_arg_count = 0;
6921    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);
6922    if (md->hitend && start_partial == NULL) start_partial = md->start_used_ptr;    if (md->hitend && start_partial == NULL)
6923        {
6924        start_partial = md->start_used_ptr;
6925        match_partial = start_match;
6926        }
6927    
6928    switch(rc)    switch(rc)
6929      {      {
# Line 6749  for(;;) Line 6936  for(;;)
6936    
6937      case MATCH_SKIP_ARG:      case MATCH_SKIP_ARG:
6938      new_start_match = start_match;      new_start_match = start_match;
6939      md->ignore_skip_arg = TRUE;      md->ignore_skip_arg = md->skip_arg_count;
6940      break;      break;
6941    
6942      /* 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
6943      same as the match we have just done, treat it as NOMATCH. */      greater than the match we have just done, treat it as NOMATCH. */
6944    
6945      case MATCH_SKIP:      case MATCH_SKIP:
6946      if (md->start_match_ptr != start_match)      if (md->start_match_ptr > start_match)
6947        {        {
6948        new_start_match = md->start_match_ptr;        new_start_match = md->start_match_ptr;
6949        break;        break;
# Line 6764  for(;;) Line 6951  for(;;)
6951      /* Fall through */      /* Fall through */
6952    
6953      /* NOMATCH and PRUNE advance by one character. THEN at this level acts      /* NOMATCH and PRUNE advance by one character. THEN at this level acts
6954      exactly like PRUNE. Unset the ignore SKIP-with-argument flag. */      exactly like PRUNE. Unset ignore SKIP-with-argument. */
6955    
6956      case MATCH_NOMATCH:      case MATCH_NOMATCH:
6957      case MATCH_PRUNE:      case MATCH_PRUNE:
6958      case MATCH_THEN:      case MATCH_THEN:
6959      md->ignore_skip_arg = FALSE;      md->ignore_skip_arg = 0;
6960      new_start_match = start_match + 1;      new_start_match = start_match + 1;
6961  #ifdef SUPPORT_UTF  #ifdef SUPPORT_UTF
6962      if (utf)      if (utf)
# Line 6862  if (rc == MATCH_MATCH || rc == MATCH_ACC Line 7049  if (rc == MATCH_MATCH || rc == MATCH_ACC
7049          (arg_offset_max - 2) * sizeof(int));          (arg_offset_max - 2) * sizeof(int));
7050        DPRINTF(("Copied offsets from temporary memory\n"));        DPRINTF(("Copied offsets from temporary memory\n"));
7051        }        }
7052      if (md->end_offset_top > arg_offset_max) md->offset_overflow = TRUE;      if (md->end_offset_top > arg_offset_max) md->capture_last |= OVFLBIT;
7053      DPRINTF(("Freeing temporary memory\n"));      DPRINTF(("Freeing temporary memory\n"));
7054      (PUBL(free))(md->offset_vector);      (PUBL(free))(md->offset_vector);
7055      }      }
# Line 6870  if (rc == MATCH_MATCH || rc == MATCH_ACC Line 7057  if (rc == MATCH_MATCH || rc == MATCH_ACC
7057    /* 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
7058    too many to fit into the vector. */    too many to fit into the vector. */
7059    
7060    rc = (md->offset_overflow && md->end_offset_top >= arg_offset_max)?    rc = ((md->capture_last & OVFLBIT) != 0 &&
7061             md->end_offset_top >= arg_offset_max)?
7062      0 : md->end_offset_top/2;      0 : md->end_offset_top/2;
7063    
7064    /* If there is space in the offset vector, set any unused pairs at the end of    /* If there is space in the offset vector, set any unused pairs at the end of
# Line 6886  if (rc == MATCH_MATCH || rc == MATCH_ACC Line 7074  if (rc == MATCH_MATCH || rc == MATCH_ACC
7074      {      {
7075      register int *iptr, *iend;      register int *iptr, *iend;
7076      int resetcount = 2 + re->top_bracket * 2;      int resetcount = 2 + re->top_bracket * 2;
7077      if (resetcount > offsetcount) resetcount = ocount;      if (resetcount > offsetcount) resetcount = offsetcount;
7078      iptr = offsets + md->end_offset_top;      iptr = offsets + md->end_offset_top;
7079      iend = offsets + resetcount;      iend = offsets + resetcount;
7080      while (iptr < iend) *iptr++ = -1;      while (iptr < iend) *iptr++ = -1;
# Line 6907  if (rc == MATCH_MATCH || rc == MATCH_ACC Line 7095  if (rc == MATCH_MATCH || rc == MATCH_ACC
7095    if (extra_data != NULL && (extra_data->flags & PCRE_EXTRA_MARK) != 0)    if (extra_data != NULL && (extra_data->flags & PCRE_EXTRA_MARK) != 0)
7096      *(extra_data->mark) = (pcre_uchar *)md->mark;      *(extra_data->mark) = (pcre_uchar *)md->mark;
7097    DPRINTF((">>>> returning %d\n", rc));    DPRINTF((">>>> returning %d\n", rc));
7098    #ifdef NO_RECURSE
7099      release_match_heapframes(&frame_zero);
7100    #endif
7101    return rc;    return rc;
7102    }    }
7103    
# Line 6924  if (using_temporary_offsets) Line 7115  if (using_temporary_offsets)
7115  if (rc != MATCH_NOMATCH && rc != PCRE_ERROR_PARTIAL)  if (rc != MATCH_NOMATCH && rc != PCRE_ERROR_PARTIAL)
7116    {    {
7117    DPRINTF((">>>> error: returning %d\n", rc));    DPRINTF((">>>> error: returning %d\n", rc));
7118    #ifdef NO_RECURSE
7119      release_match_heapframes(&frame_zero);
7120    #endif
7121    return rc;    return rc;
7122    }    }
7123    
# Line 6937  if (start_partial != NULL) Line 7131  if (start_partial != NULL)
7131      {      {
7132      offsets[0] = (int)(start_partial - (PCRE_PUCHAR)subject);      offsets[0] = (int)(start_partial - (PCRE_PUCHAR)subject);
7133      offsets[1] = (int)(end_subject - (PCRE_PUCHAR)subject);      offsets[1] = (int)(end_subject - (PCRE_PUCHAR)subject);
7134        if (offsetcount > 2)
7135          offsets[2] = (int)(match_partial - (PCRE_PUCHAR)subject);
7136      }      }
7137    rc = PCRE_ERROR_PARTIAL;    rc = PCRE_ERROR_PARTIAL;
7138    }    }
# Line 6953  else Line 7149  else
7149    
7150  if (extra_data != NULL && (extra_data->flags & PCRE_EXTRA_MARK) != 0)  if (extra_data != NULL && (extra_data->flags & PCRE_EXTRA_MARK) != 0)
7151    *(extra_data->mark) = (pcre_uchar *)md->nomatch_mark;    *(extra_data->mark) = (pcre_uchar *)md->nomatch_mark;
7152    #ifdef NO_RECURSE
7153      release_match_heapframes(&frame_zero);
7154    #endif
7155  return rc;  return rc;
7156  }  }
7157    

Legend:
Removed from v.895  
changed lines
  Added in v.1296

  ViewVC Help
Powered by ViewVC 1.1.5