/[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 1302 by ph10, Wed Mar 27 11:13:36 2013 UTC
# Line 6  Line 6 
6  and semantics are as close as possible to those of the Perl 5 language.  and semantics are as close as possible to those of the Perl 5 language.
7    
8                         Written by Philip Hazel                         Written by Philip Hazel
9             Copyright (c) 1997-2012 University of Cambridge             Copyright (c) 1997-2013 University of Cambridge
10    
11  -----------------------------------------------------------------------------  -----------------------------------------------------------------------------
12  Redistribution and use in source and binary forms, with or without  Redistribution and use in source and binary forms, with or without
# Line 37  POSSIBILITY OF SUCH DAMAGE. Line 37  POSSIBILITY OF SUCH DAMAGE.
37  -----------------------------------------------------------------------------  -----------------------------------------------------------------------------
38  */  */
39    
   
40  /* This module contains pcre_exec(), the externally visible function that does  /* This module contains pcre_exec(), the externally visible function that does
41  pattern matching using an NFA algorithm, trying to mimic Perl as closely as  pattern matching using an NFA algorithm, trying to mimic Perl as closely as
42  possible. There are also some static supporting functions. */  possible. There are also some static supporting functions. */
# Line 57  possible. There are also some static sup Line 56  possible. There are also some static sup
56  #undef min  #undef min
57  #undef max  #undef max
58    
59    /* The md->capture_last field uses the lower 16 bits for the last captured
60    substring (which can never be greater than 65535) and a bit in the top half
61    to mean "capture vector overflowed". This odd way of doing things was
62    implemented when it was realized that preserving and restoring the overflow bit
63    whenever the last capture number was saved/restored made for a neater
64    interface, and doing it this way saved on (a) another variable, which would
65    have increased the stack frame size (a big NO-NO in PCRE) and (b) another
66    separate set of save/restore instructions. The following defines are used in
67    implementing this. */
68    
69    #define CAPLMASK    0x0000ffff    /* The bits used for last_capture */
70    #define OVFLMASK    0xffff0000    /* The bits used for the overflow flag */
71    #define OVFLBIT     0x00010000    /* The bit that is set for overflow */
72    
73  /* Values for setting in md->match_function_type to indicate two special types  /* Values for setting in md->match_function_type to indicate two special types
74  of call to match(). We do it this way to save on using another stack variable,  of call to match(). We do it this way to save on using another stack variable,
75  as stack usage is to be discouraged. */  as stack usage is to be discouraged. */
# Line 74  defined PCRE_ERROR_xxx codes, which are Line 87  defined PCRE_ERROR_xxx codes, which are
87  negative to avoid the external error codes. */  negative to avoid the external error codes. */
88    
89  #define MATCH_ACCEPT       (-999)  #define MATCH_ACCEPT       (-999)
90  #define MATCH_COMMIT       (-998)  #define MATCH_KETRPOS      (-998)
91  #define MATCH_KETRPOS      (-997)  #define MATCH_ONCE         (-997)
92  #define MATCH_ONCE         (-996)  /* The next 5 must be kept together and in sequence so that a test that checks
93    for any one of them can use a range. */
94    #define MATCH_COMMIT       (-996)
95  #define MATCH_PRUNE        (-995)  #define MATCH_PRUNE        (-995)
96  #define MATCH_SKIP         (-994)  #define MATCH_SKIP         (-994)
97  #define MATCH_SKIP_ARG     (-993)  #define MATCH_SKIP_ARG     (-993)
98  #define MATCH_THEN         (-992)  #define MATCH_THEN         (-992)
99    #define MATCH_BACKTRACK_MAX MATCH_THEN
100    #define MATCH_BACKTRACK_MIN MATCH_COMMIT
101    
102  /* Maximum number of ints of offset to save on the stack for recursive calls.  /* Maximum number of ints of offset to save on the stack for recursive calls.
103  If the offset vector is bigger, malloc is used. This should be a multiple of 3,  If the offset vector is bigger, malloc is used. This should be a multiple of 3,
# Line 93  because the offset vector is always a mu Line 110  because the offset vector is always a mu
110  static const char rep_min[] = { 0, 0, 1, 1, 0, 0 };  static const char rep_min[] = { 0, 0, 1, 1, 0, 0 };
111  static const char rep_max[] = { 0, 0, 0, 0, 1, 1 };  static const char rep_max[] = { 0, 0, 0, 0, 1, 1 };
112    
   
   
113  #ifdef PCRE_DEBUG  #ifdef PCRE_DEBUG
114  /*************************************************  /*************************************************
115  *        Debugging function to print chars       *  *        Debugging function to print chars       *
# Line 115  Returns:     nothing Line 130  Returns:     nothing
130  static void  static void
131  pchars(const pcre_uchar *p, int length, BOOL is_subject, match_data *md)  pchars(const pcre_uchar *p, int length, BOOL is_subject, match_data *md)
132  {  {
133  unsigned int c;  pcre_uint32 c;
134    BOOL utf = md->utf;
135  if (is_subject && length > md->end_subject - p) length = md->end_subject - p;  if (is_subject && length > md->end_subject - p) length = md->end_subject - p;
136  while (length-- > 0)  while (length-- > 0)
137    if (isprint(c = *(p++))) printf("%c", c); else printf("\\x%02x", c);    if (isprint(c = RAWUCHARINCTEST(p))) printf("%c", (char)c); else printf("\\x{%02x}", c);
138  }  }
139  #endif  #endif
140    
# Line 140  Arguments: Line 156  Arguments:
156    md          points to match data block    md          points to match data block
157    caseless    TRUE if caseless    caseless    TRUE if caseless
158    
159  Returns:      < 0 if not matched, otherwise the number of subject bytes matched  Returns:      >= 0 the number of subject bytes matched
160                  -1 no match
161                  -2 partial match; always given if at end subject
162  */  */
163    
164  static int  static int
# Line 149  match_ref(int offset, register PCRE_PUCH Line 167  match_ref(int offset, register PCRE_PUCH
167  {  {
168  PCRE_PUCHAR eptr_start = eptr;  PCRE_PUCHAR eptr_start = eptr;
169  register PCRE_PUCHAR p = md->start_subject + md->offset_vector[offset];  register PCRE_PUCHAR p = md->start_subject + md->offset_vector[offset];
170    #ifdef SUPPORT_UTF
171    BOOL utf = md->utf;
172    #endif
173    
174  #ifdef PCRE_DEBUG  #ifdef PCRE_DEBUG
175  if (eptr >= md->end_subject)  if (eptr >= md->end_subject)
# Line 163  pchars(p, length, FALSE, md); Line 184  pchars(p, length, FALSE, md);
184  printf("\n");  printf("\n");
185  #endif  #endif
186    
187  /* Always fail if reference not set (and not JavaScript compatible). */  /* Always fail if reference not set (and not JavaScript compatible - in that
188    case the length is passed as zero). */
189    
190  if (length < 0) return -1;  if (length < 0) return -1;
191    
# Line 175  if (caseless) Line 197  if (caseless)
197    {    {
198  #ifdef SUPPORT_UTF  #ifdef SUPPORT_UTF
199  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
200    if (md->utf)    if (utf)
201      {      {
202      /* Match characters up to the end of the reference. NOTE: the number of      /* Match characters up to the end of the reference. NOTE: the number of
203      bytes matched may differ, because there are some characters whose upper and      data units matched may differ, because in UTF-8 there are some characters
204      lower case versions code as different numbers of bytes. For example, U+023A      whose upper and lower case versions code have different numbers of bytes.
205      (2 bytes in UTF-8) is the upper case version of U+2C65 (3 bytes in UTF-8);      For example, U+023A (2 bytes in UTF-8) is the upper case version of U+2C65
206      a sequence of 3 of the former uses 6 bytes, as does a sequence of two of      (3 bytes in UTF-8); a sequence of 3 of the former uses 6 bytes, as does a
207      the latter. It is important, therefore, to check the length along the      sequence of two of the latter. It is important, therefore, to check the
208      reference, not along the subject (earlier code did this wrong). */      length along the reference, not along the subject (earlier code did this
209        wrong). */
210    
211      PCRE_PUCHAR endptr = p + length;      PCRE_PUCHAR endptr = p + length;
212      while (p < endptr)      while (p < endptr)
213        {        {
214        int c, d;        pcre_uint32 c, d;
215        if (eptr >= md->end_subject) return -1;        const ucd_record *ur;
216          if (eptr >= md->end_subject) return -2;   /* Partial match */
217        GETCHARINC(c, eptr);        GETCHARINC(c, eptr);
218        GETCHARINC(d, p);        GETCHARINC(d, p);
219        if (c != d && c != UCD_OTHERCASE(d)) return -1;        ur = GET_UCD(d);
220          if (c != d && c != d + ur->other_case)
221            {
222            const pcre_uint32 *pp = PRIV(ucd_caseless_sets) + ur->caseset;
223            for (;;)
224              {
225              if (c < *pp) return -1;
226              if (c == *pp++) break;
227              }
228            }
229        }        }
230      }      }
231    else    else
# Line 202  if (caseless) Line 235  if (caseless)
235    /* The same code works when not in UTF-8 mode and in UTF-8 mode when there    /* The same code works when not in UTF-8 mode and in UTF-8 mode when there
236    is no UCP support. */    is no UCP support. */
237      {      {
     if (eptr + length > md->end_subject) return -1;  
238      while (length-- > 0)      while (length-- > 0)
239        {        {
240        if (TABLE_GET(*p, md->lcc, *p) != TABLE_GET(*eptr, md->lcc, *eptr)) return -1;        pcre_uint32 cc, cp;
241          if (eptr >= md->end_subject) return -2;   /* Partial match */
242          cc = RAWUCHARTEST(eptr);
243          cp = RAWUCHARTEST(p);
244          if (TABLE_GET(cp, md->lcc, cp) != TABLE_GET(cc, md->lcc, cc)) return -1;
245        p++;        p++;
246        eptr++;        eptr++;
247        }        }
# Line 217  are in UTF-8 mode. */ Line 253  are in UTF-8 mode. */
253    
254  else  else
255    {    {
256    if (eptr + length > md->end_subject) return -1;    while (length-- > 0)
257    while (length-- > 0) if (*p++ != *eptr++) return -1;      {
258        if (eptr >= md->end_subject) return -2;   /* Partial match */
259        if (RAWUCHARINCTEST(p) != RAWUCHARINCTEST(eptr)) return -1;
260        }
261    }    }
262    
263  return (int)(eptr - eptr_start);  return (int)(eptr - eptr_start);
# Line 273  enum { RM1=1, RM2,  RM3,  RM4,  RM5,  RM Line 312  enum { RM1=1, RM2,  RM3,  RM4,  RM5,  RM
312         RM31,  RM32, RM33, RM34, RM35, RM36, RM37, RM38, RM39, RM40,         RM31,  RM32, RM33, RM34, RM35, RM36, RM37, RM38, RM39, RM40,
313         RM41,  RM42, RM43, RM44, RM45, RM46, RM47, RM48, RM49, RM50,         RM41,  RM42, RM43, RM44, RM45, RM46, RM47, RM48, RM49, RM50,
314         RM51,  RM52, RM53, RM54, RM55, RM56, RM57, RM58, RM59, RM60,         RM51,  RM52, RM53, RM54, RM55, RM56, RM57, RM58, RM59, RM60,
315         RM61,  RM62, RM63, RM64, RM65, RM66 };         RM61,  RM62, RM63, RM64, RM65, RM66, RM67, RM68 };
316    
317  /* These versions of the macros use the stack, as normal. There are debugging  /* These versions of the macros use the stack, as normal. There are debugging
318  versions and production versions. Note that the "rw" argument of RMATCH isn't  versions and production versions. Note that the "rw" argument of RMATCH isn't
# Line 291  actually used in this definition. */ Line 330  actually used in this definition. */
330    }    }
331  #define RRETURN(ra) \  #define RRETURN(ra) \
332    { \    { \
333    printf("match() returned %d from line %d ", ra, __LINE__); \    printf("match() returned %d from line %d\n", ra, __LINE__); \
334    return ra; \    return ra; \
335    }    }
336  #else  #else
# Line 311  argument of match(), which never changes Line 350  argument of match(), which never changes
350    
351  #define RMATCH(ra,rb,rc,rd,re,rw)\  #define RMATCH(ra,rb,rc,rd,re,rw)\
352    {\    {\
353    heapframe *newframe = (heapframe *)(PUBL(stack_malloc))(sizeof(heapframe));\    heapframe *newframe = frame->Xnextframe;\
354    if (newframe == NULL) RRETURN(PCRE_ERROR_NOMEMORY);\    if (newframe == NULL)\
355    frame->Xwhere = rw; \      {\
356        newframe = (heapframe *)(PUBL(stack_malloc))(sizeof(heapframe));\
357        if (newframe == NULL) RRETURN(PCRE_ERROR_NOMEMORY);\
358        newframe->Xnextframe = NULL;\
359        frame->Xnextframe = newframe;\
360        }\
361      frame->Xwhere = rw;\
362    newframe->Xeptr = ra;\    newframe->Xeptr = ra;\
363    newframe->Xecode = rb;\    newframe->Xecode = rb;\
364    newframe->Xmstart = mstart;\    newframe->Xmstart = mstart;\
# Line 332  argument of match(), which never changes Line 377  argument of match(), which never changes
377    {\    {\
378    heapframe *oldframe = frame;\    heapframe *oldframe = frame;\
379    frame = oldframe->Xprevframe;\    frame = oldframe->Xprevframe;\
   if (oldframe != &frame_zero) (PUBL(stack_free))(oldframe);\  
380    if (frame != NULL)\    if (frame != NULL)\
381      {\      {\
382      rrc = ra;\      rrc = ra;\
# Line 346  argument of match(), which never changes Line 390  argument of match(), which never changes
390    
391  typedef struct heapframe {  typedef struct heapframe {
392    struct heapframe *Xprevframe;    struct heapframe *Xprevframe;
393      struct heapframe *Xnextframe;
394    
395    /* Function arguments that may change */    /* Function arguments that may change */
396    
# Line 376  typedef struct heapframe { Line 421  typedef struct heapframe {
421    
422  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
423    int Xprop_type;    int Xprop_type;
424    int Xprop_value;    unsigned int Xprop_value;
425    int Xprop_fail_result;    int Xprop_fail_result;
426    int Xoclength;    int Xoclength;
427    pcre_uchar Xocchars[6];    pcre_uchar Xocchars[6];
# Line 389  typedef struct heapframe { Line 434  typedef struct heapframe {
434    int Xlength;    int Xlength;
435    int Xmax;    int Xmax;
436    int Xmin;    int Xmin;
437    int Xnumber;    unsigned int Xnumber;
438    int Xoffset;    int Xoffset;
439    int Xop;    unsigned int Xop;
440    int Xsave_capture_last;    pcre_int32 Xsave_capture_last;
441    int Xsave_offset1, Xsave_offset2, Xsave_offset3;    int Xsave_offset1, Xsave_offset2, Xsave_offset3;
442    int Xstacksave[REC_STACK_SAVE_MAX];    int Xstacksave[REC_STACK_SAVE_MAX];
443    
# Line 477  so they can be ordinary variables in all Line 522  so they can be ordinary variables in all
522    
523  register int  rrc;         /* Returns from recursive calls */  register int  rrc;         /* Returns from recursive calls */
524  register int  i;           /* Used for loops not involving calls to RMATCH() */  register int  i;           /* Used for loops not involving calls to RMATCH() */
525  register unsigned int c;   /* Character values not kept over RMATCH() calls */  register pcre_uint32 c;    /* Character values not kept over RMATCH() calls */
526  register BOOL utf;         /* Local copy of UTF flag for speed */  register BOOL utf;         /* Local copy of UTF flag for speed */
527    
528  BOOL minimize, possessive; /* Quantifier options */  BOOL minimize, possessive; /* Quantifier options */
# Line 487  int condcode; Line 532  int condcode;
532  /* When recursion is not being used, all "local" variables that have to be  /* When recursion is not being used, all "local" variables that have to be
533  preserved over calls to RMATCH() are part of a "frame". We set up the top-level  preserved over calls to RMATCH() are part of a "frame". We set up the top-level
534  frame on the stack here; subsequent instantiations are obtained from the heap  frame on the stack here; subsequent instantiations are obtained from the heap
535  whenever RMATCH() does a "recursion". See the macro definitions above. Putting  whenever RMATCH() does a "recursion". See the macro definitions above. Putting
536  the top-level on the stack rather than malloc-ing them all gives a performance  the top-level on the stack rather than malloc-ing them all gives a performance
537  boost in many cases where there is not much "recursion". */  boost in many cases where there is not much "recursion". */
538    
539  #ifdef NO_RECURSE  #ifdef NO_RECURSE
540  heapframe frame_zero;  heapframe *frame = (heapframe *)md->match_frames_base;
 heapframe *frame = &frame_zero;  
 frame->Xprevframe = NULL;            /* Marks the top level */  
541    
542  /* Copy in the original argument variables */  /* Copy in the original argument variables */
543    
# Line 596  BOOL prev_is_word; Line 639  BOOL prev_is_word;
639    
640  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
641  int prop_type;  int prop_type;
642  int prop_value;  unsigned int prop_value;
643  int prop_fail_result;  int prop_fail_result;
644  int oclength;  int oclength;
645  pcre_uchar occhars[6];  pcre_uchar occhars[6];
# Line 607  int ctype; Line 650  int ctype;
650  int length;  int length;
651  int max;  int max;
652  int min;  int min;
653  int number;  unsigned int number;
654  int offset;  int offset;
655  int op;  unsigned int op;
656  int save_capture_last;  pcre_int32 save_capture_last;
657  int save_offset1, save_offset2, save_offset3;  int save_offset1, save_offset2, save_offset3;
658  int stacksave[REC_STACK_SAVE_MAX];  int stacksave[REC_STACK_SAVE_MAX];
659    
660  eptrblock newptrb;  eptrblock newptrb;
661    
662  /* There is a special fudge for calling match() in a way that causes it to  /* There is a special fudge for calling match() in a way that causes it to
663  measure the size of its basic stack frame when the stack is being used for  measure the size of its basic stack frame when the stack is being used for
664  recursion. The second argument (ecode) being NULL triggers this behaviour. It  recursion. The second argument (ecode) being NULL triggers this behaviour. It
665  cannot normally every be NULL. The return is the negated value of the frame  cannot normally ever be NULL. The return is the negated value of the frame
666  size. */  size. */
667    
668  if (ecode == NULL)  if (ecode == NULL)
# Line 631  if (ecode == NULL) Line 674  if (ecode == NULL)
674      int len = (char *)&rdepth - (char *)eptr;      int len = (char *)&rdepth - (char *)eptr;
675      return (len > 0)? -len : len;      return (len > 0)? -len : len;
676      }      }
677    }    }
678  #endif     /* NO_RECURSE */  #endif     /* NO_RECURSE */
679    
680  /* To save space on the stack and in the heap frame, I have doubled up on some  /* To save space on the stack and in the heap frame, I have doubled up on some
# Line 728  for (;;) Line 771  for (;;)
771      unaltered. */      unaltered. */
772    
773      else if (rrc == MATCH_SKIP_ARG &&      else if (rrc == MATCH_SKIP_ARG &&
774          STRCMP_UC_UC(ecode + 2, md->start_match_ptr) == 0)          STRCMP_UC_UC_TEST(ecode + 2, md->start_match_ptr) == 0)
775        {        {
776        md->start_match_ptr = eptr;        md->start_match_ptr = eptr;
777        RRETURN(MATCH_SKIP);        RRETURN(MATCH_SKIP);
# Line 738  for (;;) Line 781  for (;;)
781      case OP_FAIL:      case OP_FAIL:
782      RRETURN(MATCH_NOMATCH);      RRETURN(MATCH_NOMATCH);
783    
     /* COMMIT overrides PRUNE, SKIP, and THEN */  
   
784      case OP_COMMIT:      case OP_COMMIT:
785      RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md,      RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md,
786        eptrb, RM52);        eptrb, RM52);
787      if (rrc != MATCH_NOMATCH && rrc != MATCH_PRUNE &&      if (rrc != MATCH_NOMATCH) RRETURN(rrc);
         rrc != MATCH_SKIP && rrc != MATCH_SKIP_ARG &&  
         rrc != MATCH_THEN)  
       RRETURN(rrc);  
788      RRETURN(MATCH_COMMIT);      RRETURN(MATCH_COMMIT);
789    
     /* PRUNE overrides THEN */  
   
790      case OP_PRUNE:      case OP_PRUNE:
791      RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md,      RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md,
792        eptrb, RM51);        eptrb, RM51);
793      if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN) RRETURN(rrc);      if (rrc != MATCH_NOMATCH) RRETURN(rrc);
794      RRETURN(MATCH_PRUNE);      RRETURN(MATCH_PRUNE);
795    
796      case OP_PRUNE_ARG:      case OP_PRUNE_ARG:
# Line 764  for (;;) Line 800  for (;;)
800        eptrb, RM56);        eptrb, RM56);
801      if ((rrc == MATCH_MATCH || rrc == MATCH_ACCEPT) &&      if ((rrc == MATCH_MATCH || rrc == MATCH_ACCEPT) &&
802           md->mark == NULL) md->mark = ecode + 2;           md->mark == NULL) md->mark = ecode + 2;
803      if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN) RRETURN(rrc);      if (rrc != MATCH_NOMATCH) RRETURN(rrc);
804      RRETURN(MATCH_PRUNE);      RRETURN(MATCH_PRUNE);
805    
     /* SKIP overrides PRUNE and THEN */  
   
806      case OP_SKIP:      case OP_SKIP:
807      RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md,      RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md,
808        eptrb, RM53);        eptrb, RM53);
809      if (rrc != MATCH_NOMATCH && rrc != MATCH_PRUNE && rrc != MATCH_THEN)      if (rrc != MATCH_NOMATCH) RRETURN(rrc);
       RRETURN(rrc);  
810      md->start_match_ptr = eptr;   /* Pass back current position */      md->start_match_ptr = eptr;   /* Pass back current position */
811      RRETURN(MATCH_SKIP);      RRETURN(MATCH_SKIP);
812    
813      /* Note that, for Perl compatibility, SKIP with an argument does NOT set      /* Note that, for Perl compatibility, SKIP with an argument does NOT set
814      nomatch_mark. There is a flag that disables this opcode when re-matching a      nomatch_mark. When a pattern match ends with a SKIP_ARG for which there was
815      pattern that ended with a SKIP for which there was not a matching MARK. */      not a matching mark, we have to re-run the match, ignoring the SKIP_ARG
816        that failed and any that preceed it (either they also failed, or were not
817        triggered). To do this, we maintain a count of executed SKIP_ARGs. If a
818        SKIP_ARG gets to top level, the match is re-run with md->ignore_skip_arg
819        set to the count of the one that failed. */
820    
821      case OP_SKIP_ARG:      case OP_SKIP_ARG:
822      if (md->ignore_skip_arg)      md->skip_arg_count++;
823        if (md->skip_arg_count <= md->ignore_skip_arg)
824        {        {
825        ecode += PRIV(OP_lengths)[*ecode] + ecode[1];        ecode += PRIV(OP_lengths)[*ecode] + ecode[1];
826        break;        break;
827        }        }
828      RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode] + ecode[1], offset_top, md,      RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode] + ecode[1], offset_top, md,
829        eptrb, RM57);        eptrb, RM57);
830      if (rrc != MATCH_NOMATCH && rrc != MATCH_PRUNE && rrc != MATCH_THEN)      if (rrc != MATCH_NOMATCH) RRETURN(rrc);
831        RRETURN(rrc);  
   
832      /* Pass back the current skip name by overloading md->start_match_ptr and      /* Pass back the current skip name by overloading md->start_match_ptr and
833      returning the special MATCH_SKIP_ARG return code. This will either be      returning the special MATCH_SKIP_ARG return code. This will either be
834      caught by a matching MARK, or get to the top, where it causes a rematch      caught by a matching MARK, or get to the top, where it causes a rematch
835      with the md->ignore_skip_arg flag set. */      with md->ignore_skip_arg set to the value of md->skip_arg_count. */
836    
837      md->start_match_ptr = ecode + 2;      md->start_match_ptr = ecode + 2;
838      RRETURN(MATCH_SKIP_ARG);      RRETURN(MATCH_SKIP_ARG);
# Line 838  for (;;) Line 875  for (;;)
875      case OP_ONCE_NC:      case OP_ONCE_NC:
876      prev = ecode;      prev = ecode;
877      saved_eptr = eptr;      saved_eptr = eptr;
878      save_mark = md->mark;      save_mark = md->mark;
879      do      do
880        {        {
881        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, eptrb, RM64);        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, eptrb, RM64);
# Line 857  for (;;) Line 894  for (;;)
894    
895        if (rrc != MATCH_NOMATCH) RRETURN(rrc);        if (rrc != MATCH_NOMATCH) RRETURN(rrc);
896        ecode += GET(ecode,1);        ecode += GET(ecode,1);
897        md->mark = save_mark;        md->mark = save_mark;
898        }        }
899      while (*ecode == OP_ALT);      while (*ecode == OP_ALT);
900    
# Line 897  for (;;) Line 934  for (;;)
934        }        }
935      else  /* OP_KETRMAX */      else  /* OP_KETRMAX */
936        {        {
       md->match_function_type = MATCH_CBEGROUP;  
937        RMATCH(eptr, prev, offset_top, md, eptrb, RM66);        RMATCH(eptr, prev, offset_top, md, eptrb, RM66);
938        if (rrc != MATCH_NOMATCH) RRETURN(rrc);        if (rrc != MATCH_NOMATCH) RRETURN(rrc);
939        ecode += 1 + LINK_SIZE;        ecode += 1 + LINK_SIZE;
# Line 937  for (;;) Line 973  for (;;)
973        save_offset2 = md->offset_vector[offset+1];        save_offset2 = md->offset_vector[offset+1];
974        save_offset3 = md->offset_vector[md->offset_end - number];        save_offset3 = md->offset_vector[md->offset_end - number];
975        save_capture_last = md->capture_last;        save_capture_last = md->capture_last;
976        save_mark = md->mark;        save_mark = md->mark;
977    
978        DPRINTF(("saving %d %d %d\n", save_offset1, save_offset2, save_offset3));        DPRINTF(("saving %d %d %d\n", save_offset1, save_offset2, save_offset3));
979        md->offset_vector[md->offset_end - number] =        md->offset_vector[md->offset_end - number] =
# Line 1026  for (;;) Line 1062  for (;;)
1062    
1063      for (;;)      for (;;)
1064        {        {
1065        if (op >= OP_SBRA || op == OP_ONCE) md->match_function_type = MATCH_CBEGROUP;        if (op >= OP_SBRA || op == OP_ONCE)
1066            md->match_function_type = MATCH_CBEGROUP;
1067    
1068        /* If this is not a possibly empty group, and there are no (*THEN)s in        /* If this is not a possibly empty group, and there are no (*THEN)s in
1069        the pattern, and this is the final alternative, optimize as described        the pattern, and this is the final alternative, optimize as described
# Line 1041  for (;;) Line 1078  for (;;)
1078        /* In all other cases, we have to make another call to match(). */        /* In all other cases, we have to make another call to match(). */
1079    
1080        save_mark = md->mark;        save_mark = md->mark;
1081          save_capture_last = md->capture_last;
1082        RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md, eptrb,        RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md, eptrb,
1083          RM2);          RM2);
1084    
1085        /* See comment in the code for capturing groups above about handling        /* See comment in the code for capturing groups above about handling
1086        THEN. */        THEN. */
1087    
# Line 1070  for (;;) Line 1108  for (;;)
1108          RRETURN(rrc);          RRETURN(rrc);
1109          }          }
1110        ecode += GET(ecode, 1);        ecode += GET(ecode, 1);
1111        md->mark = save_mark;        md->mark = save_mark;
1112        if (*ecode != OP_ALT) break;        if (*ecode != OP_ALT) break;
1113          md->capture_last = save_capture_last;
1114        }        }
1115    
1116      RRETURN(MATCH_NOMATCH);      RRETURN(MATCH_NOMATCH);
# Line 1193  for (;;) Line 1232  for (;;)
1232      POSSESSIVE_NON_CAPTURE:      POSSESSIVE_NON_CAPTURE:
1233      matched_once = FALSE;      matched_once = FALSE;
1234      code_offset = (int)(ecode - md->start_code);      code_offset = (int)(ecode - md->start_code);
1235        save_capture_last = md->capture_last;
1236    
1237      for (;;)      for (;;)
1238        {        {
# Line 1222  for (;;) Line 1262  for (;;)
1262        if (rrc != MATCH_NOMATCH) RRETURN(rrc);        if (rrc != MATCH_NOMATCH) RRETURN(rrc);
1263        ecode += GET(ecode, 1);        ecode += GET(ecode, 1);
1264        if (*ecode != OP_ALT) break;        if (*ecode != OP_ALT) break;
1265          md->capture_last = save_capture_last;
1266        }        }
1267    
1268      if (matched_once || allow_zero)      if (matched_once || allow_zero)
# Line 1253  for (;;) Line 1294  for (;;)
1294          cb.version          = 2;   /* Version 1 of the callout block */          cb.version          = 2;   /* Version 1 of the callout block */
1295          cb.callout_number   = ecode[LINK_SIZE+2];          cb.callout_number   = ecode[LINK_SIZE+2];
1296          cb.offset_vector    = md->offset_vector;          cb.offset_vector    = md->offset_vector;
1297  #ifdef COMPILE_PCRE8  #if defined COMPILE_PCRE8
1298          cb.subject          = (PCRE_SPTR)md->start_subject;          cb.subject          = (PCRE_SPTR)md->start_subject;
1299  #else  #elif defined COMPILE_PCRE16
1300          cb.subject          = (PCRE_SPTR16)md->start_subject;          cb.subject          = (PCRE_SPTR16)md->start_subject;
1301    #elif defined COMPILE_PCRE32
1302            cb.subject          = (PCRE_SPTR32)md->start_subject;
1303  #endif  #endif
1304          cb.subject_length   = (int)(md->end_subject - md->start_subject);          cb.subject_length   = (int)(md->end_subject - md->start_subject);
1305          cb.start_match      = (int)(mstart - md->start_subject);          cb.start_match      = (int)(mstart - md->start_subject);
# Line 1264  for (;;) Line 1307  for (;;)
1307          cb.pattern_position = GET(ecode, LINK_SIZE + 3);          cb.pattern_position = GET(ecode, LINK_SIZE + 3);
1308          cb.next_item_length = GET(ecode, 3 + 2*LINK_SIZE);          cb.next_item_length = GET(ecode, 3 + 2*LINK_SIZE);
1309          cb.capture_top      = offset_top/2;          cb.capture_top      = offset_top/2;
1310          cb.capture_last     = md->capture_last;          cb.capture_last     = md->capture_last & CAPLMASK;
1311            /* Internal change requires this for API compatibility. */
1312            if (cb.capture_last == 0) cb.capture_last = -1;
1313          cb.callout_data     = md->callout_data;          cb.callout_data     = md->callout_data;
1314          cb.mark             = md->nomatch_mark;          cb.mark             = md->nomatch_mark;
1315          if ((rrc = (*PUBL(callout))(&cb)) > 0) RRETURN(MATCH_NOMATCH);          if ((rrc = (*PUBL(callout))(&cb)) > 0) RRETURN(MATCH_NOMATCH);
1316          if (rrc < 0) RRETURN(rrc);          if (rrc < 0) RRETURN(rrc);
1317          }          }
1318        ecode += PRIV(OP_lengths)[OP_CALLOUT];        ecode += PRIV(OP_lengths)[OP_CALLOUT];
1319          codelink -= PRIV(OP_lengths)[OP_CALLOUT];
1320        }        }
1321    
1322      condcode = ecode[LINK_SIZE+1];      condcode = ecode[LINK_SIZE+1];
# Line 1286  for (;;) Line 1332  for (;;)
1332          }          }
1333        else        else
1334          {          {
1335          int recno = GET2(ecode, LINK_SIZE + 2);   /* Recursion group number*/          unsigned int recno = GET2(ecode, LINK_SIZE + 2);   /* Recursion group number*/
1336          condition = (recno == RREF_ANY || recno == md->recursive->group_num);          condition = (recno == RREF_ANY || recno == md->recursive->group_num);
1337    
1338          /* If the test is for recursion into a specific subpattern, and it is          /* If the test is for recursion into a specific subpattern, and it is
# Line 1358  for (;;) Line 1404  for (;;)
1404    
1405        if (!condition && condcode == OP_NCREF)        if (!condition && condcode == OP_NCREF)
1406          {          {
1407          int refno = offset >> 1;          unsigned int refno = offset >> 1;
1408          pcre_uchar *slotA = md->name_table;          pcre_uchar *slotA = md->name_table;
1409    
1410          for (i = 0; i < md->name_count; i++)          for (i = 0; i < md->name_count; i++)
# Line 1486  for (;;) Line 1532  for (;;)
1532      to close any currently open capturing brackets. */      to close any currently open capturing brackets. */
1533    
1534      case OP_CLOSE:      case OP_CLOSE:
1535      number = GET2(ecode, 1);      number = GET2(ecode, 1);   /* Must be less than 65536 */
1536      offset = number << 1;      offset = number << 1;
1537    
1538  #ifdef PCRE_DEBUG  #ifdef PCRE_DEBUG
# Line 1494  for (;;) Line 1540  for (;;)
1540        printf("\n");        printf("\n");
1541  #endif  #endif
1542    
1543      md->capture_last = number;      md->capture_last = (md->capture_last & OVFLMASK) | number;
1544      if (offset >= md->offset_max) md->offset_overflow = TRUE; else      if (offset >= md->offset_max) md->capture_last |= OVFLBIT; else
1545        {        {
1546        md->offset_vector[offset] =        md->offset_vector[offset] =
1547          md->offset_vector[md->offset_end - number];          md->offset_vector[md->offset_end - number];
# Line 1549  for (;;) Line 1595  for (;;)
1595    
1596      case OP_ASSERT:      case OP_ASSERT:
1597      case OP_ASSERTBACK:      case OP_ASSERTBACK:
1598      save_mark = md->mark;      save_mark = md->mark;
1599      if (md->match_function_type == MATCH_CONDASSERT)      if (md->match_function_type == MATCH_CONDASSERT)
1600        {        {
1601        condassert = TRUE;        condassert = TRUE;
# Line 1557  for (;;) Line 1603  for (;;)
1603        }        }
1604      else condassert = FALSE;      else condassert = FALSE;
1605    
1606        /* Loop for each branch */
1607    
1608      do      do
1609        {        {
1610        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, NULL, RM4);        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, NULL, RM4);
1611    
1612          /* A match means that the assertion is true; break out of the loop
1613          that matches its alternatives. */
1614    
1615        if (rrc == MATCH_MATCH || rrc == MATCH_ACCEPT)        if (rrc == MATCH_MATCH || rrc == MATCH_ACCEPT)
1616          {          {
1617          mstart = md->start_match_ptr;   /* In case \K reset it */          mstart = md->start_match_ptr;   /* In case \K reset it */
1618          break;          break;
1619          }          }
1620    
1621          /* If not matched, restore the previous mark setting. */
1622    
1623          md->mark = save_mark;
1624    
1625          /* See comment in the code for capturing groups above about handling
1626          THEN. */
1627    
1628        /* PCRE does not allow THEN to escape beyond an assertion; it is treated        if (rrc == MATCH_THEN)
1629        as NOMATCH. */          {
1630            next = ecode + GET(ecode,1);
1631            if (md->start_match_ptr < next &&
1632                (*ecode == OP_ALT || *next == OP_ALT))
1633              rrc = MATCH_NOMATCH;
1634            }
1635    
1636          /* Anything other than NOMATCH causes the entire assertion to fail,
1637          passing back the return code. This includes COMMIT, SKIP, PRUNE and an
1638          uncaptured THEN, which means they take their normal effect. This
1639          consistent approach does not always have exactly the same effect as in
1640          Perl. */
1641    
1642        if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN) RRETURN(rrc);        if (rrc != MATCH_NOMATCH) RRETURN(rrc);
1643        ecode += GET(ecode, 1);        ecode += GET(ecode, 1);
       md->mark = save_mark;  
1644        }        }
1645      while (*ecode == OP_ALT);      while (*ecode == OP_ALT);   /* Continue for next alternative */
1646    
1647        /* If we have tried all the alternative branches, the assertion has
1648        failed. If not, we broke out after a match. */
1649    
1650      if (*ecode == OP_KET) RRETURN(MATCH_NOMATCH);      if (*ecode == OP_KET) RRETURN(MATCH_NOMATCH);
1651    
# Line 1581  for (;;) Line 1653  for (;;)
1653    
1654      if (condassert) RRETURN(MATCH_MATCH);      if (condassert) RRETURN(MATCH_MATCH);
1655    
1656      /* Continue from after the assertion, updating the offsets high water      /* Continue from after a successful assertion, updating the offsets high
1657      mark, since extracts may have been taken during the assertion. */      water mark, since extracts may have been taken during the assertion. */
1658    
1659      do ecode += GET(ecode,1); while (*ecode == OP_ALT);      do ecode += GET(ecode,1); while (*ecode == OP_ALT);
1660      ecode += 1 + LINK_SIZE;      ecode += 1 + LINK_SIZE;
1661      offset_top = md->end_offset_top;      offset_top = md->end_offset_top;
1662      continue;      continue;
1663    
1664      /* Negative assertion: all branches must fail to match. Encountering SKIP,      /* Negative assertion: all branches must fail to match for the assertion to
1665      PRUNE, or COMMIT means we must assume failure without checking subsequent      succeed. */
     branches. */  
1666    
1667      case OP_ASSERT_NOT:      case OP_ASSERT_NOT:
1668      case OP_ASSERTBACK_NOT:      case OP_ASSERTBACK_NOT:
1669      save_mark = md->mark;      save_mark = md->mark;
1670      if (md->match_function_type == MATCH_CONDASSERT)      if (md->match_function_type == MATCH_CONDASSERT)
1671        {        {
1672        condassert = TRUE;        condassert = TRUE;
# Line 1603  for (;;) Line 1674  for (;;)
1674        }        }
1675      else condassert = FALSE;      else condassert = FALSE;
1676    
1677        /* Loop for each alternative branch. */
1678    
1679      do      do
1680        {        {
1681        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, NULL, RM5);        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, NULL, RM5);
1682        md->mark = save_mark;        md->mark = save_mark;   /* Always restore the mark setting */
1683        if (rrc == MATCH_MATCH || rrc == MATCH_ACCEPT) RRETURN(MATCH_NOMATCH);  
1684        if (rrc == MATCH_SKIP || rrc == MATCH_PRUNE || rrc == MATCH_COMMIT)        switch(rrc)
1685          {          {
1686          do ecode += GET(ecode,1); while (*ecode == OP_ALT);          case MATCH_MATCH:            /* A successful match means */
1687          break;          case MATCH_ACCEPT:           /* the assertion has failed. */
1688          }          RRETURN(MATCH_NOMATCH);
1689    
1690            case MATCH_NOMATCH:          /* Carry on with next branch */
1691            break;
1692    
1693        /* PCRE does not allow THEN to escape beyond an assertion; it is treated          /* See comment in the code for capturing groups above about handling
1694        as NOMATCH. */          THEN. */
1695    
1696        if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN) RRETURN(rrc);          case MATCH_THEN:
1697            next = ecode + GET(ecode,1);
1698            if (md->start_match_ptr < next &&
1699                (*ecode == OP_ALT || *next == OP_ALT))
1700              {
1701              rrc = MATCH_NOMATCH;
1702              break;
1703              }
1704            /* Otherwise fall through. */
1705    
1706            /* COMMIT, SKIP, PRUNE, and an uncaptured THEN cause the whole
1707            assertion to fail to match, without considering any more alternatives.
1708            Failing to match means the assertion is true. This is a consistent
1709            approach, but does not always have the same effect as in Perl. */
1710    
1711            case MATCH_COMMIT:
1712            case MATCH_SKIP:
1713            case MATCH_SKIP_ARG:
1714            case MATCH_PRUNE:
1715            do ecode += GET(ecode,1); while (*ecode == OP_ALT);
1716            goto NEG_ASSERT_TRUE;   /* Break out of alternation loop */
1717    
1718            /* Anything else is an error */
1719    
1720            default:
1721            RRETURN(rrc);
1722            }
1723    
1724          /* Continue with next branch */
1725    
1726        ecode += GET(ecode,1);        ecode += GET(ecode,1);
1727        }        }
1728      while (*ecode == OP_ALT);      while (*ecode == OP_ALT);
1729    
1730        /* All branches in the assertion failed to match. */
1731    
1732        NEG_ASSERT_TRUE:
1733      if (condassert) RRETURN(MATCH_MATCH);  /* Condition assertion */      if (condassert) RRETURN(MATCH_MATCH);  /* Condition assertion */
1734        ecode += 1 + LINK_SIZE;                /* Continue with current branch */
     ecode += 1 + LINK_SIZE;  
1735      continue;      continue;
1736    
1737      /* Move the subject pointer back. This occurs only at the start of      /* Move the subject pointer back. This occurs only at the start of
# Line 1671  for (;;) Line 1778  for (;;)
1778        cb.version          = 2;   /* Version 1 of the callout block */        cb.version          = 2;   /* Version 1 of the callout block */
1779        cb.callout_number   = ecode[1];        cb.callout_number   = ecode[1];
1780        cb.offset_vector    = md->offset_vector;        cb.offset_vector    = md->offset_vector;
1781  #ifdef COMPILE_PCRE8  #if defined COMPILE_PCRE8
1782        cb.subject          = (PCRE_SPTR)md->start_subject;        cb.subject          = (PCRE_SPTR)md->start_subject;
1783  #else  #elif defined COMPILE_PCRE16
1784        cb.subject          = (PCRE_SPTR16)md->start_subject;        cb.subject          = (PCRE_SPTR16)md->start_subject;
1785    #elif defined COMPILE_PCRE32
1786          cb.subject          = (PCRE_SPTR32)md->start_subject;
1787  #endif  #endif
1788        cb.subject_length   = (int)(md->end_subject - md->start_subject);        cb.subject_length   = (int)(md->end_subject - md->start_subject);
1789        cb.start_match      = (int)(mstart - md->start_subject);        cb.start_match      = (int)(mstart - md->start_subject);
# Line 1682  for (;;) Line 1791  for (;;)
1791        cb.pattern_position = GET(ecode, 2);        cb.pattern_position = GET(ecode, 2);
1792        cb.next_item_length = GET(ecode, 2 + LINK_SIZE);        cb.next_item_length = GET(ecode, 2 + LINK_SIZE);
1793        cb.capture_top      = offset_top/2;        cb.capture_top      = offset_top/2;
1794        cb.capture_last     = md->capture_last;        cb.capture_last     = md->capture_last & CAPLMASK;
1795          /* Internal change requires this for API compatibility. */
1796          if (cb.capture_last == 0) cb.capture_last = -1;
1797        cb.callout_data     = md->callout_data;        cb.callout_data     = md->callout_data;
1798        cb.mark             = md->nomatch_mark;        cb.mark             = md->nomatch_mark;
1799        if ((rrc = (*PUBL(callout))(&cb)) > 0) RRETURN(MATCH_NOMATCH);        if ((rrc = (*PUBL(callout))(&cb)) > 0) RRETURN(MATCH_NOMATCH);
# Line 1711  for (;;) Line 1822  for (;;)
1822      case OP_RECURSE:      case OP_RECURSE:
1823        {        {
1824        recursion_info *ri;        recursion_info *ri;
1825        int recno;        unsigned int recno;
1826    
1827        callpat = md->start_code + GET(ecode, 1);        callpat = md->start_code + GET(ecode, 1);
1828        recno = (callpat == md->start_code)? 0 :        recno = (callpat == md->start_code)? 0 :
# Line 1728  for (;;) Line 1839  for (;;)
1839        /* Add to "recursing stack" */        /* Add to "recursing stack" */
1840    
1841        new_recursive.group_num = recno;        new_recursive.group_num = recno;
1842          new_recursive.saved_capture_last = md->capture_last;
1843        new_recursive.subject_position = eptr;        new_recursive.subject_position = eptr;
1844        new_recursive.prevrec = md->recursive;        new_recursive.prevrec = md->recursive;
1845        md->recursive = &new_recursive;        md->recursive = &new_recursive;
# Line 1751  for (;;) Line 1863  for (;;)
1863              new_recursive.saved_max * sizeof(int));              new_recursive.saved_max * sizeof(int));
1864    
1865        /* OK, now we can do the recursion. After processing each alternative,        /* OK, now we can do the recursion. After processing each alternative,
1866        restore the offset data. If there were nested recursions, md->recursive        restore the offset data and the last captured value. If there were nested
1867        might be changed, so reset it before looping. */        recursions, md->recursive might be changed, so reset it before looping.
1868          */
1869    
1870        DPRINTF(("Recursing into group %d\n", new_recursive.group_num));        DPRINTF(("Recursing into group %d\n", new_recursive.group_num));
1871        cbegroup = (*callpat >= OP_SBRA);        cbegroup = (*callpat >= OP_SBRA);
# Line 1763  for (;;) Line 1876  for (;;)
1876            md, eptrb, RM6);            md, eptrb, RM6);
1877          memcpy(md->offset_vector, new_recursive.offset_save,          memcpy(md->offset_vector, new_recursive.offset_save,
1878              new_recursive.saved_max * sizeof(int));              new_recursive.saved_max * sizeof(int));
1879            md->capture_last = new_recursive.saved_capture_last;
1880          md->recursive = new_recursive.prevrec;          md->recursive = new_recursive.prevrec;
1881          if (rrc == MATCH_MATCH || rrc == MATCH_ACCEPT)          if (rrc == MATCH_MATCH || rrc == MATCH_ACCEPT)
1882            {            {
# Line 1779  for (;;) Line 1893  for (;;)
1893            goto RECURSION_MATCHED;        /* Exit loop; end processing */            goto RECURSION_MATCHED;        /* Exit loop; end processing */
1894            }            }
1895    
1896          /* PCRE does not allow THEN to escape beyond a recursion; it is treated          /* PCRE does not allow THEN, SKIP, PRUNE or COMMIT to escape beyond a
1897          as NOMATCH. */          recursion; they cause a NOMATCH for the entire recursion. These codes
1898            are defined in a range that can be tested for. */
1899    
1900            if (rrc >= MATCH_BACKTRACK_MIN && rrc <= MATCH_BACKTRACK_MAX)
1901              RRETURN(MATCH_NOMATCH);
1902    
1903            /* Any return code other than NOMATCH is an error. */
1904    
1905          else if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN)          if (rrc != MATCH_NOMATCH)
1906            {            {
1907            DPRINTF(("Recursion gave error %d\n", rrc));            DPRINTF(("Recursion gave error %d\n", rrc));
1908            if (new_recursive.offset_save != stacksave)            if (new_recursive.offset_save != stacksave)
# Line 1912  for (;;) Line 2032  for (;;)
2032    
2033        /* Deal with capturing */        /* Deal with capturing */
2034    
2035        md->capture_last = number;        md->capture_last = (md->capture_last & OVFLMASK) | number;
2036        if (offset >= md->offset_max) md->offset_overflow = TRUE; else        if (offset >= md->offset_max) md->capture_last |= OVFLBIT; else
2037          {          {
2038          /* If offset is greater than offset_top, it means that we are          /* If offset is greater than offset_top, it means that we are
2039          "skipping" a capturing group, and that group's offsets must be marked          "skipping" a capturing group, and that group's offsets must be marked
# Line 1993  for (;;) Line 2113  for (;;)
2113          }          }
2114        if (*prev >= OP_SBRA)    /* Could match an empty string */        if (*prev >= OP_SBRA)    /* Could match an empty string */
2115          {          {
         md->match_function_type = MATCH_CBEGROUP;  
2116          RMATCH(eptr, prev, offset_top, md, eptrb, RM50);          RMATCH(eptr, prev, offset_top, md, eptrb, RM50);
2117          RRETURN(rrc);          RRETURN(rrc);
2118          }          }
# Line 2002  for (;;) Line 2121  for (;;)
2121        }        }
2122      else  /* OP_KETRMAX */      else  /* OP_KETRMAX */
2123        {        {
       if (*prev >= OP_SBRA) md->match_function_type = MATCH_CBEGROUP;  
2124        RMATCH(eptr, prev, offset_top, md, eptrb, RM13);        RMATCH(eptr, prev, offset_top, md, eptrb, RM13);
2125        if (rrc == MATCH_ONCE && md->once_target == prev) rrc = MATCH_NOMATCH;        if (rrc == MATCH_ONCE && md->once_target == prev) rrc = MATCH_NOMATCH;
2126        if (rrc != MATCH_NOMATCH) RRETURN(rrc);        if (rrc != MATCH_NOMATCH) RRETURN(rrc);
# Line 2059  for (;;) Line 2177  for (;;)
2177    
2178      case OP_DOLLM:      case OP_DOLLM:
2179      if (eptr < md->end_subject)      if (eptr < md->end_subject)
2180        { if (!IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH); }        {
2181          if (!IS_NEWLINE(eptr))
2182            {
2183            if (md->partial != 0 &&
2184                eptr + 1 >= md->end_subject &&
2185                NLBLOCK->nltype == NLTYPE_FIXED &&
2186                NLBLOCK->nllen == 2 &&
2187                RAWUCHARTEST(eptr) == NLBLOCK->nl[0])
2188              {
2189              md->hitend = TRUE;
2190              if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
2191              }
2192            RRETURN(MATCH_NOMATCH);
2193            }
2194          }
2195      else      else
2196        {        {
2197        if (md->noteol) RRETURN(MATCH_NOMATCH);        if (md->noteol) RRETURN(MATCH_NOMATCH);
# Line 2091  for (;;) Line 2223  for (;;)
2223      ASSERT_NL_OR_EOS:      ASSERT_NL_OR_EOS:
2224      if (eptr < md->end_subject &&      if (eptr < md->end_subject &&
2225          (!IS_NEWLINE(eptr) || eptr != md->end_subject - md->nllen))          (!IS_NEWLINE(eptr) || eptr != md->end_subject - md->nllen))
2226          {
2227          if (md->partial != 0 &&
2228              eptr + 1 >= md->end_subject &&
2229              NLBLOCK->nltype == NLTYPE_FIXED &&
2230              NLBLOCK->nllen == 2 &&
2231              RAWUCHARTEST(eptr) == NLBLOCK->nl[0])
2232            {
2233            md->hitend = TRUE;
2234            if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
2235            }
2236        RRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2237          }
2238    
2239      /* Either at end of string or \n before end. */      /* Either at end of string or \n before end. */
2240    
# Line 2219  for (;;) Line 2362  for (;;)
2362        }        }
2363      break;      break;
2364    
2365      /* Match a single character type; inline for speed */      /* Match any single character type except newline; have to take care with
2366        CRLF newlines and partial matching. */
2367    
2368      case OP_ANY:      case OP_ANY:
2369      if (IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH);      if (IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH);
2370        if (md->partial != 0 &&
2371            eptr + 1 >= md->end_subject &&
2372            NLBLOCK->nltype == NLTYPE_FIXED &&
2373            NLBLOCK->nllen == 2 &&
2374            RAWUCHARTEST(eptr) == NLBLOCK->nl[0])
2375          {
2376          md->hitend = TRUE;
2377          if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
2378          }
2379    
2380      /* Fall through */      /* Fall through */
2381    
2382        /* Match any single character whatsoever. */
2383    
2384      case OP_ALLANY:      case OP_ALLANY:
2385      if (eptr >= md->end_subject)   /* DO NOT merge the eptr++ here; it must */      if (eptr >= md->end_subject)   /* DO NOT merge the eptr++ here; it must */
2386        {                            /* not be updated before SCHECK_PARTIAL. */        {                            /* not be updated before SCHECK_PARTIAL. */
# Line 2364  for (;;) Line 2520  for (;;)
2520        {        {
2521        default: RRETURN(MATCH_NOMATCH);        default: RRETURN(MATCH_NOMATCH);
2522    
2523        case 0x000d:        case CHAR_CR:
2524        if (eptr < md->end_subject && *eptr == 0x0a) eptr++;        if (eptr >= md->end_subject)
2525            {
2526            SCHECK_PARTIAL();
2527            }
2528          else if (RAWUCHARTEST(eptr) == CHAR_LF) eptr++;
2529        break;        break;
2530    
2531        case 0x000a:        case CHAR_LF:
2532        break;        break;
2533    
2534        case 0x000b:        case CHAR_VT:
2535        case 0x000c:        case CHAR_FF:
2536        case 0x0085:        case CHAR_NEL:
2537    #ifndef EBCDIC
2538        case 0x2028:        case 0x2028:
2539        case 0x2029:        case 0x2029:
2540    #endif  /* Not EBCDIC */
2541        if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);        if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);
2542        break;        break;
2543        }        }
# Line 2391  for (;;) Line 2553  for (;;)
2553      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2554      switch(c)      switch(c)
2555        {        {
2556          HSPACE_CASES: RRETURN(MATCH_NOMATCH);  /* Byte and multibyte cases */
2557        default: break;        default: break;
       case 0x09:      /* HT */  
       case 0x20:      /* SPACE */  
       case 0xa0:      /* NBSP */  
       case 0x1680:    /* OGHAM SPACE MARK */  
       case 0x180e:    /* MONGOLIAN VOWEL SEPARATOR */  
       case 0x2000:    /* EN QUAD */  
       case 0x2001:    /* EM QUAD */  
       case 0x2002:    /* EN SPACE */  
       case 0x2003:    /* EM SPACE */  
       case 0x2004:    /* THREE-PER-EM SPACE */  
       case 0x2005:    /* FOUR-PER-EM SPACE */  
       case 0x2006:    /* SIX-PER-EM SPACE */  
       case 0x2007:    /* FIGURE SPACE */  
       case 0x2008:    /* PUNCTUATION SPACE */  
       case 0x2009:    /* THIN SPACE */  
       case 0x200A:    /* HAIR SPACE */  
       case 0x202f:    /* NARROW NO-BREAK SPACE */  
       case 0x205f:    /* MEDIUM MATHEMATICAL SPACE */  
       case 0x3000:    /* IDEOGRAPHIC SPACE */  
       RRETURN(MATCH_NOMATCH);  
2558        }        }
2559      ecode++;      ecode++;
2560      break;      break;
# Line 2425  for (;;) Line 2568  for (;;)
2568      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2569      switch(c)      switch(c)
2570        {        {
2571          HSPACE_CASES: break;  /* Byte and multibyte cases */
2572        default: RRETURN(MATCH_NOMATCH);        default: RRETURN(MATCH_NOMATCH);
       case 0x09:      /* HT */  
       case 0x20:      /* SPACE */  
       case 0xa0:      /* NBSP */  
       case 0x1680:    /* OGHAM SPACE MARK */  
       case 0x180e:    /* MONGOLIAN VOWEL SEPARATOR */  
       case 0x2000:    /* EN QUAD */  
       case 0x2001:    /* EM QUAD */  
       case 0x2002:    /* EN SPACE */  
       case 0x2003:    /* EM SPACE */  
       case 0x2004:    /* THREE-PER-EM SPACE */  
       case 0x2005:    /* FOUR-PER-EM SPACE */  
       case 0x2006:    /* SIX-PER-EM SPACE */  
       case 0x2007:    /* FIGURE SPACE */  
       case 0x2008:    /* PUNCTUATION SPACE */  
       case 0x2009:    /* THIN SPACE */  
       case 0x200A:    /* HAIR SPACE */  
       case 0x202f:    /* NARROW NO-BREAK SPACE */  
       case 0x205f:    /* MEDIUM MATHEMATICAL SPACE */  
       case 0x3000:    /* IDEOGRAPHIC SPACE */  
       break;  
2573        }        }
2574      ecode++;      ecode++;
2575      break;      break;
# Line 2459  for (;;) Line 2583  for (;;)
2583      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2584      switch(c)      switch(c)
2585        {        {
2586          VSPACE_CASES: RRETURN(MATCH_NOMATCH);
2587        default: break;        default: break;
       case 0x0a:      /* LF */  
       case 0x0b:      /* VT */  
       case 0x0c:      /* FF */  
       case 0x0d:      /* CR */  
       case 0x85:      /* NEL */  
       case 0x2028:    /* LINE SEPARATOR */  
       case 0x2029:    /* PARAGRAPH SEPARATOR */  
       RRETURN(MATCH_NOMATCH);  
2588        }        }
2589      ecode++;      ecode++;
2590      break;      break;
# Line 2481  for (;;) Line 2598  for (;;)
2598      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2599      switch(c)      switch(c)
2600        {        {
2601          VSPACE_CASES: break;
2602        default: RRETURN(MATCH_NOMATCH);        default: RRETURN(MATCH_NOMATCH);
       case 0x0a:      /* LF */  
       case 0x0b:      /* VT */  
       case 0x0c:      /* FF */  
       case 0x0d:      /* CR */  
       case 0x85:      /* NEL */  
       case 0x2028:    /* LINE SEPARATOR */  
       case 0x2029:    /* PARAGRAPH SEPARATOR */  
       break;  
2603        }        }
2604      ecode++;      ecode++;
2605      break;      break;
# Line 2507  for (;;) Line 2617  for (;;)
2617        }        }
2618      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2619        {        {
2620          const pcre_uint32 *cp;
2621        const ucd_record *prop = GET_UCD(c);        const ucd_record *prop = GET_UCD(c);
2622    
2623        switch(ecode[1])        switch(ecode[1])
# Line 2567  for (;;) Line 2678  for (;;)
2678            RRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
2679          break;          break;
2680    
2681            case PT_CLIST:
2682            cp = PRIV(ucd_caseless_sets) + ecode[2];
2683            for (;;)
2684              {
2685              if (c < *cp)
2686                { if (op == OP_PROP) { RRETURN(MATCH_NOMATCH); } else break; }
2687              if (c == *cp++)
2688                { if (op == OP_PROP) break; else { RRETURN(MATCH_NOMATCH); } }
2689              }
2690            break;
2691    
2692            case PT_UCNC:
2693            if ((c == CHAR_DOLLAR_SIGN || c == CHAR_COMMERCIAL_AT ||
2694                 c == CHAR_GRAVE_ACCENT || (c >= 0xa0 && c <= 0xd7ff) ||
2695                 c >= 0xe000) == (op == OP_NOTPROP))
2696              RRETURN(MATCH_NOMATCH);
2697            break;
2698    
2699          /* This should never occur */          /* This should never occur */
2700    
2701          default:          default:
# Line 2586  for (;;) Line 2715  for (;;)
2715        SCHECK_PARTIAL();        SCHECK_PARTIAL();
2716        RRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2717        }        }
2718      GETCHARINCTEST(c, eptr);      else
     if (UCD_CATEGORY(c) == ucp_M) RRETURN(MATCH_NOMATCH);  
     while (eptr < md->end_subject)  
2719        {        {
2720        int len = 1;        int lgb, rgb;
2721        if (!utf) c = *eptr; else { GETCHARLEN(c, eptr, len); }        GETCHARINCTEST(c, eptr);
2722        if (UCD_CATEGORY(c) != ucp_M) break;        lgb = UCD_GRAPHBREAK(c);
2723        eptr += len;        while (eptr < md->end_subject)
2724            {
2725            int len = 1;
2726            if (!utf) c = *eptr; else { GETCHARLEN(c, eptr, len); }
2727            rgb = UCD_GRAPHBREAK(c);
2728            if ((PRIV(ucp_gbtable)[lgb] & (1 << rgb)) == 0) break;
2729            lgb = rgb;
2730            eptr += len;
2731            }
2732        }        }
2733        CHECK_PARTIAL();
2734      ecode++;      ecode++;
2735      break;      break;
2736  #endif  #endif  /* SUPPORT_UCP */
2737    
2738    
2739      /* Match a back reference, possibly repeatedly. Look past the end of the      /* Match a back reference, possibly repeatedly. Look past the end of the
# Line 2660  for (;;) Line 2796  for (;;)
2796        default:               /* No repeat follows */        default:               /* No repeat follows */
2797        if ((length = match_ref(offset, eptr, length, md, caseless)) < 0)        if ((length = match_ref(offset, eptr, length, md, caseless)) < 0)
2798          {          {
2799            if (length == -2) eptr = md->end_subject;   /* Partial match */
2800          CHECK_PARTIAL();          CHECK_PARTIAL();
2801          RRETURN(MATCH_NOMATCH);          RRETURN(MATCH_NOMATCH);
2802          }          }
# Line 2685  for (;;) Line 2822  for (;;)
2822        int slength;        int slength;
2823        if ((slength = match_ref(offset, eptr, length, md, caseless)) < 0)        if ((slength = match_ref(offset, eptr, length, md, caseless)) < 0)
2824          {          {
2825            if (slength == -2) eptr = md->end_subject;   /* Partial match */
2826          CHECK_PARTIAL();          CHECK_PARTIAL();
2827          RRETURN(MATCH_NOMATCH);          RRETURN(MATCH_NOMATCH);
2828          }          }
# Line 2708  for (;;) Line 2846  for (;;)
2846          if (fi >= max) RRETURN(MATCH_NOMATCH);          if (fi >= max) RRETURN(MATCH_NOMATCH);
2847          if ((slength = match_ref(offset, eptr, length, md, caseless)) < 0)          if ((slength = match_ref(offset, eptr, length, md, caseless)) < 0)
2848            {            {
2849              if (slength == -2) eptr = md->end_subject;   /* Partial match */
2850            CHECK_PARTIAL();            CHECK_PARTIAL();
2851            RRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
2852            }            }
# Line 2726  for (;;) Line 2865  for (;;)
2865          int slength;          int slength;
2866          if ((slength = match_ref(offset, eptr, length, md, caseless)) < 0)          if ((slength = match_ref(offset, eptr, length, md, caseless)) < 0)
2867            {            {
2868            CHECK_PARTIAL();            /* Can't use CHECK_PARTIAL because we don't want to update eptr in
2869              the soft partial matching case. */
2870    
2871              if (slength == -2 && md->partial != 0 &&
2872                  md->end_subject > md->start_used_ptr)
2873                {
2874                md->hitend = TRUE;
2875                if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
2876                }
2877            break;            break;
2878            }            }
2879          eptr += slength;          eptr += slength;
2880          }          }
2881    
2882        while (eptr >= pp)        while (eptr >= pp)
2883          {          {
2884          RMATCH(eptr, ecode, offset_top, md, eptrb, RM15);          RMATCH(eptr, ecode, offset_top, md, eptrb, RM15);
# Line 3094  for (;;) Line 3242  for (;;)
3242          CHECK_PARTIAL();             /* Not SCHECK_PARTIAL() */          CHECK_PARTIAL();             /* Not SCHECK_PARTIAL() */
3243          RRETURN(MATCH_NOMATCH);          RRETURN(MATCH_NOMATCH);
3244          }          }
3245        while (length-- > 0) if (*ecode++ != *eptr++) RRETURN(MATCH_NOMATCH);        while (length-- > 0) if (*ecode++ != RAWUCHARINC(eptr)) RRETURN(MATCH_NOMATCH);
3246        }        }
3247      else      else
3248  #endif  #endif
# Line 3134  for (;;) Line 3282  for (;;)
3282    
3283        if (fc < 128)        if (fc < 128)
3284          {          {
3285          if (md->lcc[fc]          pcre_uint32 cc = RAWUCHAR(eptr);
3286              != TABLE_GET(*eptr, md->lcc, *eptr)) RRETURN(MATCH_NOMATCH);          if (md->lcc[fc] != TABLE_GET(cc, md->lcc, cc)) RRETURN(MATCH_NOMATCH);
3287          ecode++;          ecode++;
3288          eptr++;          eptr++;
3289          }          }
# Line 3146  for (;;) Line 3294  for (;;)
3294    
3295        else        else
3296          {          {
3297          unsigned int dc;          pcre_uint32 dc;
3298          GETCHARINC(dc, eptr);          GETCHARINC(dc, eptr);
3299          ecode += length;          ecode += length;
3300    
# Line 3256  for (;;) Line 3404  for (;;)
3404        if (length > 1)        if (length > 1)
3405          {          {
3406  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
3407          unsigned int othercase;          pcre_uint32 othercase;
3408          if (op >= OP_STARI &&     /* Caseless */          if (op >= OP_STARI &&     /* Caseless */
3409              (othercase = UCD_OTHERCASE(fc)) != fc)              (othercase = UCD_OTHERCASE(fc)) != fc)
3410            oclength = PRIV(ord2utf)(othercase, occhars);            oclength = PRIV(ord2utf)(othercase, occhars);
# Line 3360  for (;;) Line 3508  for (;;)
3508      maximizing, find the maximum number of characters and work backwards. */      maximizing, find the maximum number of characters and work backwards. */
3509    
3510      DPRINTF(("matching %c{%d,%d} against subject %.*s\n", fc, min, max,      DPRINTF(("matching %c{%d,%d} against subject %.*s\n", fc, min, max,
3511        max, eptr));        max, (char *)eptr));
3512    
3513      if (op >= OP_STARI)  /* Caseless */      if (op >= OP_STARI)  /* Caseless */
3514        {        {
# Line 3383  for (;;) Line 3531  for (;;)
3531    
3532        for (i = 1; i <= min; i++)        for (i = 1; i <= min; 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            RRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
3539            }            }
3540          if (fc != *eptr && foc != *eptr) RRETURN(MATCH_NOMATCH);          cc = RAWUCHARTEST(eptr);
3541            if (fc != cc && foc != cc) RRETURN(MATCH_NOMATCH);
3542          eptr++;          eptr++;
3543          }          }
3544        if (min == max) continue;        if (min == max) continue;
# Line 3396  for (;;) Line 3546  for (;;)
3546          {          {
3547          for (fi = min;; fi++)          for (fi = min;; fi++)
3548            {            {
3549              pcre_uint32 cc;               /* Faster than pcre_uchar */
3550            RMATCH(eptr, ecode, offset_top, md, eptrb, RM24);            RMATCH(eptr, ecode, offset_top, md, eptrb, RM24);
3551            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3552            if (fi >= max) RRETURN(MATCH_NOMATCH);            if (fi >= max) RRETURN(MATCH_NOMATCH);
# Line 3404  for (;;) Line 3555  for (;;)
3555              SCHECK_PARTIAL();              SCHECK_PARTIAL();
3556              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
3557              }              }
3558            if (fc != *eptr && foc != *eptr) RRETURN(MATCH_NOMATCH);            cc = RAWUCHARTEST(eptr);
3559              if (fc != cc && foc != cc) RRETURN(MATCH_NOMATCH);
3560            eptr++;            eptr++;
3561            }            }
3562          /* Control never gets here */          /* Control never gets here */
# Line 3414  for (;;) Line 3566  for (;;)
3566          pp = eptr;          pp = eptr;
3567          for (i = min; i < max; i++)          for (i = min; i < max; i++)
3568            {            {
3569              pcre_uint32 cc;               /* Faster than pcre_uchar */
3570            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
3571              {              {
3572              SCHECK_PARTIAL();              SCHECK_PARTIAL();
3573              break;              break;
3574              }              }
3575            if (fc != *eptr && foc != *eptr) break;            cc = RAWUCHARTEST(eptr);
3576              if (fc != cc && foc != cc) break;
3577            eptr++;            eptr++;
3578            }            }
3579    
# Line 3447  for (;;) Line 3601  for (;;)
3601            SCHECK_PARTIAL();            SCHECK_PARTIAL();
3602            RRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
3603            }            }
3604          if (fc != *eptr++) RRETURN(MATCH_NOMATCH);          if (fc != RAWUCHARINCTEST(eptr)) RRETURN(MATCH_NOMATCH);
3605          }          }
3606    
3607        if (min == max) continue;        if (min == max) continue;
# Line 3464  for (;;) Line 3618  for (;;)
3618              SCHECK_PARTIAL();              SCHECK_PARTIAL();
3619              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
3620              }              }
3621            if (fc != *eptr++) RRETURN(MATCH_NOMATCH);            if (fc != RAWUCHARINCTEST(eptr)) RRETURN(MATCH_NOMATCH);
3622            }            }
3623          /* Control never gets here */          /* Control never gets here */
3624          }          }
# Line 3478  for (;;) Line 3632  for (;;)
3632              SCHECK_PARTIAL();              SCHECK_PARTIAL();
3633              break;              break;
3634              }              }
3635            if (fc != *eptr) break;            if (fc != RAWUCHARTEST(eptr)) break;
3636            eptr++;            eptr++;
3637            }            }
3638          if (possessive) continue;          if (possessive) continue;
# Line 3504  for (;;) Line 3658  for (;;)
3658        SCHECK_PARTIAL();        SCHECK_PARTIAL();
3659        RRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
3660        }        }
     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  
3661  #ifdef SUPPORT_UTF  #ifdef SUPPORT_UTF
3662        if (utf)
3663          {
3664          register pcre_uint32 ch, och;
3665    
3666          ecode++;
3667          GETCHARINC(ch, ecode);
3668          GETCHARINC(c, eptr);
3669    
3670          if (op == OP_NOT)
3671            {
3672            if (ch == c) RRETURN(MATCH_NOMATCH);
3673            }
3674          else
3675            {
3676  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
3677        if (utf && ch > 127)          if (ch > 127)
3678          och = UCD_OTHERCASE(ch);            och = UCD_OTHERCASE(ch);
3679  #else  #else
3680        if (utf && ch > 127)          if (ch > 127)
3681          och = ch;            och = ch;
3682  #endif /* SUPPORT_UCP */  #endif /* SUPPORT_UCP */
3683        else          else
3684  #endif /* SUPPORT_UTF */            och = TABLE_GET(ch, md->fcc, ch);
3685          och = TABLE_GET(ch, md->fcc, ch);          if (ch == c || och == c) RRETURN(MATCH_NOMATCH);
3686  #endif /* COMPILE_PCRE8 */          }
       if (ch == c || och == c) RRETURN(MATCH_NOMATCH);  
3687        }        }
3688      else    /* Caseful */      else
3689    #endif
3690        {        {
3691        if (*ecode++ == c) RRETURN(MATCH_NOMATCH);        register pcre_uint32 ch = ecode[1];
3692          c = *eptr++;
3693          if (ch == c || (op == OP_NOTI && TABLE_GET(ch, md->fcc, ch) == c))
3694            RRETURN(MATCH_NOMATCH);
3695          ecode += 2;
3696        }        }
3697      break;      break;
3698    
# Line 3610  for (;;) Line 3772  for (;;)
3772      /* Common code for all repeated single-byte matches. */      /* Common code for all repeated single-byte matches. */
3773    
3774      REPEATNOTCHAR:      REPEATNOTCHAR:
3775      fc = *ecode++;      GETCHARINCTEST(fc, ecode);
3776    
3777      /* The code is duplicated for the caseless and caseful cases, for speed,      /* The code is duplicated for the caseless and caseful cases, for speed,
3778      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 3783  for (;;)
3783      characters and work backwards. */      characters and work backwards. */
3784    
3785      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,
3786        max, eptr));        max, (char *)eptr));
3787    
3788      if (op >= OP_NOTSTARI)     /* Caseless */      if (op >= OP_NOTSTARI)     /* Caseless */
3789        {        {
 #ifdef COMPILE_PCRE8  
       /* fc must be < 128 if UTF is enabled. */  
       foc = md->fcc[fc];  
 #else  
3790  #ifdef SUPPORT_UTF  #ifdef SUPPORT_UTF
3791  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
3792        if (utf && fc > 127)        if (utf && fc > 127)
# Line 3640  for (;;) Line 3798  for (;;)
3798        else        else
3799  #endif /* SUPPORT_UTF */  #endif /* SUPPORT_UTF */
3800          foc = TABLE_GET(fc, md->fcc, fc);          foc = TABLE_GET(fc, md->fcc, fc);
 #endif /* COMPILE_PCRE8 */  
3801    
3802  #ifdef SUPPORT_UTF  #ifdef SUPPORT_UTF
3803        if (utf)        if (utf)
3804          {          {
3805          register unsigned int d;          register pcre_uint32 d;
3806          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
3807            {            {
3808            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
# Line 3654  for (;;) Line 3811  for (;;)
3811              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
3812              }              }
3813            GETCHARINC(d, eptr);            GETCHARINC(d, eptr);
3814            if (fc == d || foc == d) RRETURN(MATCH_NOMATCH);            if (fc == d || (unsigned int)foc == d) RRETURN(MATCH_NOMATCH);
3815            }            }
3816          }          }
3817        else        else
# Line 3680  for (;;) Line 3837  for (;;)
3837  #ifdef SUPPORT_UTF  #ifdef SUPPORT_UTF
3838          if (utf)          if (utf)
3839            {            {
3840            register unsigned int d;            register pcre_uint32 d;
3841            for (fi = min;; fi++)            for (fi = min;; fi++)
3842              {              {
3843              RMATCH(eptr, ecode, offset_top, md, eptrb, RM28);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM28);
# Line 3692  for (;;) Line 3849  for (;;)
3849                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
3850                }                }
3851              GETCHARINC(d, eptr);              GETCHARINC(d, eptr);
3852              if (fc == d || foc == d) RRETURN(MATCH_NOMATCH);              if (fc == d || (unsigned int)foc == d) RRETURN(MATCH_NOMATCH);
3853              }              }
3854            }            }
3855          else          else
# Line 3725  for (;;) Line 3882  for (;;)
3882  #ifdef SUPPORT_UTF  #ifdef SUPPORT_UTF
3883          if (utf)          if (utf)
3884            {            {
3885            register unsigned int d;            register pcre_uint32 d;
3886            for (i = min; i < max; i++)            for (i = min; i < max; i++)
3887              {              {
3888              int len = 1;              int len = 1;
# Line 3735  for (;;) Line 3892  for (;;)
3892                break;                break;
3893                }                }
3894              GETCHARLEN(d, eptr, len);              GETCHARLEN(d, eptr, len);
3895              if (fc == d || foc == d) break;              if (fc == d || (unsigned int)foc == d) break;
3896              eptr += len;              eptr += len;
3897              }              }
3898            if (possessive) continue;            if (possessive) continue;
# Line 3782  for (;;) Line 3939  for (;;)
3939  #ifdef SUPPORT_UTF  #ifdef SUPPORT_UTF
3940        if (utf)        if (utf)
3941          {          {
3942          register unsigned int d;          register pcre_uint32 d;
3943          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
3944            {            {
3945            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
# Line 3816  for (;;) Line 3973  for (;;)
3973  #ifdef SUPPORT_UTF  #ifdef SUPPORT_UTF
3974          if (utf)          if (utf)
3975            {            {
3976            register unsigned int d;            register pcre_uint32 d;
3977            for (fi = min;; fi++)            for (fi = min;; fi++)
3978              {              {
3979              RMATCH(eptr, ecode, offset_top, md, eptrb, RM32);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM32);
# Line 3860  for (;;) Line 4017  for (;;)
4017  #ifdef SUPPORT_UTF  #ifdef SUPPORT_UTF
4018          if (utf)          if (utf)
4019            {            {
4020            register unsigned int d;            register pcre_uint32 d;
4021            for (i = min; i < max; i++)            for (i = min; i < max; i++)
4022              {              {
4023              int len = 1;              int len = 1;
# Line 4136  for (;;) Line 4293  for (;;)
4293              }              }
4294            break;            break;
4295    
4296              case PT_CLIST:
4297              for (i = 1; i <= min; i++)
4298                {
4299                const pcre_uint32 *cp;
4300                if (eptr >= md->end_subject)
4301                  {
4302                  SCHECK_PARTIAL();
4303                  RRETURN(MATCH_NOMATCH);
4304                  }
4305                GETCHARINCTEST(c, eptr);
4306                cp = PRIV(ucd_caseless_sets) + prop_value;
4307                for (;;)
4308                  {
4309                  if (c < *cp)
4310                    { if (prop_fail_result) break; else { RRETURN(MATCH_NOMATCH); } }
4311                  if (c == *cp++)
4312                    { if (prop_fail_result) { RRETURN(MATCH_NOMATCH); } else break; }
4313                  }
4314                }
4315              break;
4316    
4317              case PT_UCNC:
4318              for (i = 1; i <= min; i++)
4319                {
4320                if (eptr >= md->end_subject)
4321                  {
4322                  SCHECK_PARTIAL();
4323                  RRETURN(MATCH_NOMATCH);
4324                  }
4325                GETCHARINCTEST(c, eptr);
4326                if ((c == CHAR_DOLLAR_SIGN || c == CHAR_COMMERCIAL_AT ||
4327                     c == CHAR_GRAVE_ACCENT || (c >= 0xa0 && c <= 0xd7ff) ||
4328                     c >= 0xe000) == prop_fail_result)
4329                  RRETURN(MATCH_NOMATCH);
4330                }
4331              break;
4332    
4333            /* This should not occur */            /* This should not occur */
4334    
4335            default:            default:
# Line 4155  for (;;) Line 4349  for (;;)
4349              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4350              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4351              }              }
4352            GETCHARINCTEST(c, eptr);            else
           if (UCD_CATEGORY(c) == ucp_M) RRETURN(MATCH_NOMATCH);  
           while (eptr < md->end_subject)  
4353              {              {
4354              int len = 1;              int lgb, rgb;
4355              if (!utf) c = *eptr; else { GETCHARLEN(c, eptr, len); }              GETCHARINCTEST(c, eptr);
4356              if (UCD_CATEGORY(c) != ucp_M) break;              lgb = UCD_GRAPHBREAK(c);
4357              eptr += len;             while (eptr < md->end_subject)
4358                  {
4359                  int len = 1;
4360                  if (!utf) c = *eptr; else { GETCHARLEN(c, eptr, len); }
4361                  rgb = UCD_GRAPHBREAK(c);
4362                  if ((PRIV(ucp_gbtable)[lgb] & (1 << rgb)) == 0) break;
4363                  lgb = rgb;
4364                  eptr += len;
4365                  }
4366              }              }
4367              CHECK_PARTIAL();
4368            }            }
4369          }          }
4370    
# Line 4184  for (;;) Line 4385  for (;;)
4385              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4386              }              }
4387            if (IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH);            if (IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH);
4388              if (md->partial != 0 &&
4389                  eptr + 1 >= md->end_subject &&
4390                  NLBLOCK->nltype == NLTYPE_FIXED &&
4391                  NLBLOCK->nllen == 2 &&
4392                  RAWUCHAR(eptr) == NLBLOCK->nl[0])
4393                {
4394                md->hitend = TRUE;
4395                if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
4396                }
4397            eptr++;            eptr++;
4398            ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);            ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);
4399            }            }
# Line 4220  for (;;) Line 4430  for (;;)
4430              {              {
4431              default: RRETURN(MATCH_NOMATCH);              default: RRETURN(MATCH_NOMATCH);
4432    
4433              case 0x000d:              case CHAR_CR:
4434              if (eptr < md->end_subject && *eptr == 0x0a) eptr++;              if (eptr < md->end_subject && RAWUCHAR(eptr) == CHAR_LF) eptr++;
4435              break;              break;
4436    
4437              case 0x000a:              case CHAR_LF:
4438              break;              break;
4439    
4440              case 0x000b:              case CHAR_VT:
4441              case 0x000c:              case CHAR_FF:
4442              case 0x0085:              case CHAR_NEL:
4443    #ifndef EBCDIC
4444              case 0x2028:              case 0x2028:
4445              case 0x2029:              case 0x2029:
4446    #endif  /* Not EBCDIC */
4447              if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);              if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);
4448              break;              break;
4449              }              }
# Line 4249  for (;;) Line 4461  for (;;)
4461            GETCHARINC(c, eptr);            GETCHARINC(c, eptr);
4462            switch(c)            switch(c)
4463              {              {
4464                HSPACE_CASES: RRETURN(MATCH_NOMATCH);  /* Byte and multibyte cases */
4465              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);  
4466              }              }
4467            }            }
4468          break;          break;
# Line 4285  for (;;) Line 4478  for (;;)
4478            GETCHARINC(c, eptr);            GETCHARINC(c, eptr);
4479            switch(c)            switch(c)
4480              {              {
4481                HSPACE_CASES: break;  /* Byte and multibyte cases */
4482              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;  
4483              }              }
4484            }            }
4485          break;          break;
# Line 4321  for (;;) Line 4495  for (;;)
4495            GETCHARINC(c, eptr);            GETCHARINC(c, eptr);
4496            switch(c)            switch(c)
4497              {              {
4498                VSPACE_CASES: RRETURN(MATCH_NOMATCH);
4499              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);  
4500              }              }
4501            }            }
4502          break;          break;
# Line 4345  for (;;) Line 4512  for (;;)
4512            GETCHARINC(c, eptr);            GETCHARINC(c, eptr);
4513            switch(c)            switch(c)
4514              {              {
4515                VSPACE_CASES: break;
4516              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;  
4517              }              }
4518            }            }
4519          break;          break;
# Line 4375  for (;;) Line 4535  for (;;)
4535          case OP_DIGIT:          case OP_DIGIT:
4536          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
4537            {            {
4538              pcre_uint32 cc;
4539            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
4540              {              {
4541              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4542              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4543              }              }
4544            if (*eptr >= 128 || (md->ctypes[*eptr] & ctype_digit) == 0)            cc = RAWUCHAR(eptr);
4545              if (cc >= 128 || (md->ctypes[cc] & ctype_digit) == 0)
4546              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4547            eptr++;            eptr++;
4548            /* 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 4552  for (;;)
4552          case OP_NOT_WHITESPACE:          case OP_NOT_WHITESPACE:
4553          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
4554            {            {
4555              pcre_uint32 cc;
4556            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
4557              {              {
4558              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4559              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4560              }              }
4561            if (*eptr < 128 && (md->ctypes[*eptr] & ctype_space) != 0)            cc = RAWUCHAR(eptr);
4562              if (cc < 128 && (md->ctypes[cc] & ctype_space) != 0)
4563              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4564            eptr++;            eptr++;
4565            ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);            ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);
# Line 4405  for (;;) Line 4569  for (;;)
4569          case OP_WHITESPACE:          case OP_WHITESPACE:
4570          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
4571            {            {
4572              pcre_uint32 cc;
4573            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
4574              {              {
4575              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4576              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4577              }              }
4578            if (*eptr >= 128 || (md->ctypes[*eptr] & ctype_space) == 0)            cc = RAWUCHAR(eptr);
4579              if (cc >= 128 || (md->ctypes[cc] & ctype_space) == 0)
4580              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4581            eptr++;            eptr++;
4582            /* 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 4586  for (;;)
4586          case OP_NOT_WORDCHAR:          case OP_NOT_WORDCHAR:
4587          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
4588            {            {
4589              pcre_uint32 cc;
4590            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
4591              {              {
4592              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4593              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4594              }              }
4595            if (*eptr < 128 && (md->ctypes[*eptr] & ctype_word) != 0)            cc = RAWUCHAR(eptr);
4596              if (cc < 128 && (md->ctypes[cc] & ctype_word) != 0)
4597              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4598            eptr++;            eptr++;
4599            ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);            ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);
# Line 4435  for (;;) Line 4603  for (;;)
4603          case OP_WORDCHAR:          case OP_WORDCHAR:
4604          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
4605            {            {
4606              pcre_uint32 cc;
4607            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
4608              {              {
4609              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4610              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4611              }              }
4612            if (*eptr >= 128 || (md->ctypes[*eptr] & ctype_word) == 0)            cc = RAWUCHAR(eptr);
4613              if (cc >= 128 || (md->ctypes[cc] & ctype_word) == 0)
4614              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4615            eptr++;            eptr++;
4616            /* 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 4638  for (;;)
4638              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4639              }              }
4640            if (IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH);            if (IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH);
4641              if (md->partial != 0 &&
4642                  eptr + 1 >= md->end_subject &&
4643                  NLBLOCK->nltype == NLTYPE_FIXED &&
4644                  NLBLOCK->nllen == 2 &&
4645                  *eptr == NLBLOCK->nl[0])
4646                {
4647                md->hitend = TRUE;
4648                if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
4649                }
4650            eptr++;            eptr++;
4651            }            }
4652          break;          break;
# Line 4502  for (;;) Line 4681  for (;;)
4681              {              {
4682              default: RRETURN(MATCH_NOMATCH);              default: RRETURN(MATCH_NOMATCH);
4683    
4684              case 0x000d:              case CHAR_CR:
4685              if (eptr < md->end_subject && *eptr == 0x0a) eptr++;              if (eptr < md->end_subject && *eptr == CHAR_LF) eptr++;
4686              break;              break;
4687    
4688              case 0x000a:              case CHAR_LF:
4689              break;              break;
4690    
4691              case 0x000b:              case CHAR_VT:
4692              case 0x000c:              case CHAR_FF:
4693              case 0x0085:              case CHAR_NEL:
4694  #ifdef COMPILE_PCRE16  #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
4695              case 0x2028:              case 0x2028:
4696              case 0x2029:              case 0x2029:
4697  #endif  #endif
# Line 4533  for (;;) Line 4712  for (;;)
4712            switch(*eptr++)            switch(*eptr++)
4713              {              {
4714              default: break;              default: break;
4715              case 0x09:      /* HT */              HSPACE_BYTE_CASES:
4716              case 0x20:      /* SPACE */  #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
4717              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 */  
4718  #endif  #endif
4719              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4720              }              }
# Line 4570  for (;;) Line 4732  for (;;)
4732            switch(*eptr++)            switch(*eptr++)
4733              {              {
4734              default: RRETURN(MATCH_NOMATCH);              default: RRETURN(MATCH_NOMATCH);
4735              case 0x09:      /* HT */              HSPACE_BYTE_CASES:
4736              case 0x20:      /* SPACE */  #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
4737              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 */  
4738  #endif  #endif
4739              break;              break;
4740              }              }
# Line 4606  for (;;) Line 4751  for (;;)
4751              }              }
4752            switch(*eptr++)            switch(*eptr++)
4753              {              {
4754              default: break;              VSPACE_BYTE_CASES:
4755              case 0x0a:      /* LF */  #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
4756              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 */  
4757  #endif  #endif
4758              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4759                default: break;
4760              }              }
4761            }            }
4762          break;          break;
# Line 4632  for (;;) Line 4772  for (;;)
4772            switch(*eptr++)            switch(*eptr++)
4773              {              {
4774              default: RRETURN(MATCH_NOMATCH);              default: RRETURN(MATCH_NOMATCH);
4775              case 0x0a:      /* LF */              VSPACE_BYTE_CASES:
4776              case 0x0b:      /* VT */  #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
4777              case 0x0c:      /* FF */              VSPACE_MULTIBYTE_CASES:
             case 0x0d:      /* CR */  
             case 0x85:      /* NEL */  
 #ifdef COMPILE_PCRE16  
             case 0x2028:    /* LINE SEPARATOR */  
             case 0x2029:    /* PARAGRAPH SEPARATOR */  
4778  #endif  #endif
4779              break;              break;
4780              }              }
# Line 4917  for (;;) Line 5052  for (;;)
5052              }              }
5053            /* Control never gets here */            /* Control never gets here */
5054    
5055            /* This should never occur */            case PT_CLIST:
5056              for (fi = min;; fi++)
5057                {
5058                const pcre_uint32 *cp;
5059                RMATCH(eptr, ecode, offset_top, md, eptrb, RM67);
5060                if (rrc != MATCH_NOMATCH) RRETURN(rrc);
5061                if (fi >= max) RRETURN(MATCH_NOMATCH);
5062                if (eptr >= md->end_subject)
5063                  {
5064                  SCHECK_PARTIAL();
5065                  RRETURN(MATCH_NOMATCH);
5066                  }
5067                GETCHARINCTEST(c, eptr);
5068                cp = PRIV(ucd_caseless_sets) + prop_value;
5069                for (;;)
5070                  {
5071                  if (c < *cp)
5072                    { if (prop_fail_result) break; else { RRETURN(MATCH_NOMATCH); } }
5073                  if (c == *cp++)
5074                    { if (prop_fail_result) { RRETURN(MATCH_NOMATCH); } else break; }
5075                  }
5076                }
5077              /* Control never gets here */
5078    
5079              case PT_UCNC:
5080              for (fi = min;; fi++)
5081                {
5082                RMATCH(eptr, ecode, offset_top, md, eptrb, RM68);
5083                if (rrc != MATCH_NOMATCH) RRETURN(rrc);
5084                if (fi >= max) RRETURN(MATCH_NOMATCH);
5085                if (eptr >= md->end_subject)
5086                  {
5087                  SCHECK_PARTIAL();
5088                  RRETURN(MATCH_NOMATCH);
5089                  }
5090                GETCHARINCTEST(c, eptr);
5091                if ((c == CHAR_DOLLAR_SIGN || c == CHAR_COMMERCIAL_AT ||
5092                     c == CHAR_GRAVE_ACCENT || (c >= 0xa0 && c <= 0xd7ff) ||
5093                     c >= 0xe000) == prop_fail_result)
5094                  RRETURN(MATCH_NOMATCH);
5095                }
5096              /* Control never gets here */
5097    
5098              /* This should never occur */
5099            default:            default:
5100            RRETURN(PCRE_ERROR_INTERNAL);            RRETURN(PCRE_ERROR_INTERNAL);
5101            }            }
# Line 4939  for (;;) Line 5116  for (;;)
5116              SCHECK_PARTIAL();              SCHECK_PARTIAL();
5117              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
5118              }              }
5119            GETCHARINCTEST(c, eptr);            else
           if (UCD_CATEGORY(c) == ucp_M) RRETURN(MATCH_NOMATCH);  
           while (eptr < md->end_subject)  
5120              {              {
5121              int len = 1;              int lgb, rgb;
5122              if (!utf) c = *eptr; else { GETCHARLEN(c, eptr, len); }              GETCHARINCTEST(c, eptr);
5123              if (UCD_CATEGORY(c) != ucp_M) break;              lgb = UCD_GRAPHBREAK(c);
5124              eptr += len;              while (eptr < md->end_subject)
5125                  {
5126                  int len = 1;
5127                  if (!utf) c = *eptr; else { GETCHARLEN(c, eptr, len); }
5128                  rgb = UCD_GRAPHBREAK(c);
5129                  if ((PRIV(ucp_gbtable)[lgb] & (1 << rgb)) == 0) break;
5130                  lgb = rgb;
5131                  eptr += len;
5132                  }
5133              }              }
5134              CHECK_PARTIAL();
5135            }            }
5136          }          }
5137        else        else
# Line 4971  for (;;) Line 5155  for (;;)
5155            GETCHARINC(c, eptr);            GETCHARINC(c, eptr);
5156            switch(ctype)            switch(ctype)
5157              {              {
5158              case OP_ANY:        /* This is the non-NL case */              case OP_ANY:               /* This is the non-NL case */
5159                if (md->partial != 0 &&    /* Take care with CRLF partial */
5160                    eptr >= md->end_subject &&
5161                    NLBLOCK->nltype == NLTYPE_FIXED &&
5162                    NLBLOCK->nllen == 2 &&
5163                    c == NLBLOCK->nl[0])
5164                  {
5165                  md->hitend = TRUE;
5166                  if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
5167                  }
5168                break;
5169    
5170              case OP_ALLANY:              case OP_ALLANY:
5171              case OP_ANYBYTE:              case OP_ANYBYTE:
5172              break;              break;
# Line 4980  for (;;) Line 5175  for (;;)
5175              switch(c)              switch(c)
5176                {                {
5177                default: RRETURN(MATCH_NOMATCH);                default: RRETURN(MATCH_NOMATCH);
5178                case 0x000d:                case CHAR_CR:
5179                if (eptr < md->end_subject && *eptr == 0x0a) eptr++;                if (eptr < md->end_subject && RAWUCHAR(eptr) == CHAR_LF) eptr++;
5180                break;                break;
5181                case 0x000a:  
5182                  case CHAR_LF:
5183                break;                break;
5184    
5185                case 0x000b:                case CHAR_VT:
5186                case 0x000c:                case CHAR_FF:
5187                case 0x0085:                case CHAR_NEL:
5188    #ifndef EBCDIC
5189                case 0x2028:                case 0x2028:
5190                case 0x2029:                case 0x2029:
5191    #endif  /* Not EBCDIC */
5192                if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);                if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);
5193                break;                break;
5194                }                }
# Line 4999  for (;;) Line 5197  for (;;)
5197              case OP_NOT_HSPACE:              case OP_NOT_HSPACE:
5198              switch(c)              switch(c)
5199                {                {
5200                  HSPACE_CASES: RRETURN(MATCH_NOMATCH);
5201                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);  
5202                }                }
5203              break;              break;
5204    
5205              case OP_HSPACE:              case OP_HSPACE:
5206              switch(c)              switch(c)
5207                {                {
5208                  HSPACE_CASES: break;
5209                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;  
5210                }                }
5211              break;              break;
5212    
5213              case OP_NOT_VSPACE:              case OP_NOT_VSPACE:
5214              switch(c)              switch(c)
5215                {                {
5216                  VSPACE_CASES: RRETURN(MATCH_NOMATCH);
5217                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);  
5218                }                }
5219              break;              break;
5220    
5221              case OP_VSPACE:              case OP_VSPACE:
5222              switch(c)              switch(c)
5223                {                {
5224                  VSPACE_CASES: break;
5225                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;  
5226                }                }
5227              break;              break;
5228    
# Line 5134  for (;;) Line 5280  for (;;)
5280            c = *eptr++;            c = *eptr++;
5281            switch(ctype)            switch(ctype)
5282              {              {
5283              case OP_ANY:     /* This is the non-NL case */              case OP_ANY:               /* This is the non-NL case */
5284                if (md->partial != 0 &&    /* Take care with CRLF partial */
5285                    eptr >= md->end_subject &&
5286                    NLBLOCK->nltype == NLTYPE_FIXED &&
5287                    NLBLOCK->nllen == 2 &&
5288                    c == NLBLOCK->nl[0])
5289                  {
5290                  md->hitend = TRUE;
5291                  if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
5292                  }
5293                break;
5294    
5295              case OP_ALLANY:              case OP_ALLANY:
5296              case OP_ANYBYTE:              case OP_ANYBYTE:
5297              break;              break;
# Line 5143  for (;;) Line 5300  for (;;)
5300              switch(c)              switch(c)
5301                {                {
5302                default: RRETURN(MATCH_NOMATCH);                default: RRETURN(MATCH_NOMATCH);
5303                case 0x000d:                case CHAR_CR:
5304                if (eptr < md->end_subject && *eptr == 0x0a) eptr++;                if (eptr < md->end_subject && *eptr == CHAR_LF) eptr++;
5305                break;                break;
5306    
5307                case 0x000a:                case CHAR_LF:
5308                break;                break;
5309    
5310                case 0x000b:                case CHAR_VT:
5311                case 0x000c:                case CHAR_FF:
5312                case 0x0085:                case CHAR_NEL:
5313  #ifdef COMPILE_PCRE16  #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
5314                case 0x2028:                case 0x2028:
5315                case 0x2029:                case 0x2029:
5316  #endif  #endif
# Line 5166  for (;;) Line 5323  for (;;)
5323              switch(c)              switch(c)
5324                {                {
5325                default: break;                default: break;
5326                case 0x09:      /* HT */                HSPACE_BYTE_CASES:
5327                case 0x20:      /* SPACE */  #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
5328                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 */  
5329  #endif  #endif
5330                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
5331                }                }
# Line 5195  for (;;) Line 5335  for (;;)
5335              switch(c)              switch(c)
5336                {                {
5337                default: RRETURN(MATCH_NOMATCH);                default: RRETURN(MATCH_NOMATCH);
5338                case 0x09:      /* HT */                HSPACE_BYTE_CASES:
5339                case 0x20:      /* SPACE */  #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
5340                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 */  
5341  #endif  #endif
5342                break;                break;
5343                }                }
# Line 5224  for (;;) Line 5347  for (;;)
5347              switch(c)              switch(c)
5348                {                {
5349                default: break;                default: break;
5350                case 0x0a:      /* LF */                VSPACE_BYTE_CASES:
5351                case 0x0b:      /* VT */  #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
5352                case 0x0c:      /* FF */                VSPACE_MULTIBYTE_CASES:
               case 0x0d:      /* CR */  
               case 0x85:      /* NEL */  
 #ifdef COMPILE_PCRE16  
               case 0x2028:    /* LINE SEPARATOR */  
               case 0x2029:    /* PARAGRAPH SEPARATOR */  
5353  #endif  #endif
5354                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
5355                }                }
# Line 5241  for (;;) Line 5359  for (;;)
5359              switch(c)              switch(c)
5360                {                {
5361                default: RRETURN(MATCH_NOMATCH);                default: RRETURN(MATCH_NOMATCH);
5362                case 0x0a:      /* LF */                VSPACE_BYTE_CASES:
5363                case 0x0b:      /* VT */  #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
5364                case 0x0c:      /* FF */                VSPACE_MULTIBYTE_CASES:
               case 0x0d:      /* CR */  
               case 0x85:      /* NEL */  
 #ifdef COMPILE_PCRE16  
               case 0x2028:    /* LINE SEPARATOR */  
               case 0x2029:    /* PARAGRAPH SEPARATOR */  
5365  #endif  #endif
5366                break;                break;
5367                }                }
# Line 5452  for (;;) Line 5565  for (;;)
5565              }              }
5566            break;            break;
5567    
5568              case PT_CLIST:
5569              for (i = min; i < max; i++)
5570                {
5571                const pcre_uint32 *cp;
5572                int len = 1;
5573                if (eptr >= md->end_subject)
5574                  {
5575                  SCHECK_PARTIAL();
5576                  break;
5577                  }
5578                GETCHARLENTEST(c, eptr, len);
5579                cp = PRIV(ucd_caseless_sets) + prop_value;
5580                for (;;)
5581                  {
5582                  if (c < *cp)
5583                    { if (prop_fail_result) break; else goto GOT_MAX; }
5584                  if (c == *cp++)
5585                    { if (prop_fail_result) goto GOT_MAX; else break; }
5586                  }
5587                eptr += len;
5588                }
5589              GOT_MAX:
5590              break;
5591    
5592              case PT_UCNC:
5593              for (i = min; i < max; i++)
5594                {
5595                int len = 1;
5596                if (eptr >= md->end_subject)
5597                  {
5598                  SCHECK_PARTIAL();
5599                  break;
5600                  }
5601                GETCHARLENTEST(c, eptr, len);
5602                if ((c == CHAR_DOLLAR_SIGN || c == CHAR_COMMERCIAL_AT ||
5603                     c == CHAR_GRAVE_ACCENT || (c >= 0xa0 && c <= 0xd7ff) ||
5604                     c >= 0xe000) == prop_fail_result)
5605                  break;
5606                eptr += len;
5607                }
5608              break;
5609    
5610            default:            default:
5611            RRETURN(PCRE_ERROR_INTERNAL);            RRETURN(PCRE_ERROR_INTERNAL);
5612            }            }
# Line 5475  for (;;) Line 5630  for (;;)
5630          {          {
5631          for (i = min; i < max; i++)          for (i = min; i < max; i++)
5632            {            {
           int len = 1;  
5633            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
5634              {              {
5635              SCHECK_PARTIAL();              SCHECK_PARTIAL();
5636              break;              break;
5637              }              }
5638            if (!utf) c = *eptr; else { GETCHARLEN(c, eptr, len); }            else
           if (UCD_CATEGORY(c) == ucp_M) break;  
           eptr += len;  
           while (eptr < md->end_subject)  
5639              {              {
5640              len = 1;              int lgb, rgb;
5641              if (!utf) c = *eptr; else { GETCHARLEN(c, eptr, len); }              GETCHARINCTEST(c, eptr);
5642              if (UCD_CATEGORY(c) != ucp_M) break;              lgb = UCD_GRAPHBREAK(c);
5643              eptr += len;              while (eptr < md->end_subject)
5644                  {
5645                  int len = 1;
5646                  if (!utf) c = *eptr; else { GETCHARLEN(c, eptr, len); }
5647                  rgb = UCD_GRAPHBREAK(c);
5648                  if ((PRIV(ucp_gbtable)[lgb] & (1 << rgb)) == 0) break;
5649                  lgb = rgb;
5650                  eptr += len;
5651                  }
5652              }              }
5653              CHECK_PARTIAL();
5654            }            }
5655    
5656          /* eptr is now past the end of the maximum run */          /* eptr is now past the end of the maximum run */
# Line 5534  for (;;) Line 5694  for (;;)
5694                  break;                  break;
5695                  }                  }
5696                if (IS_NEWLINE(eptr)) break;                if (IS_NEWLINE(eptr)) break;
5697                  if (md->partial != 0 &&    /* Take care with CRLF partial */
5698                      eptr + 1 >= md->end_subject &&
5699                      NLBLOCK->nltype == NLTYPE_FIXED &&
5700                      NLBLOCK->nllen == 2 &&
5701                      RAWUCHAR(eptr) == NLBLOCK->nl[0])
5702                    {
5703                    md->hitend = TRUE;
5704                    if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
5705                    }
5706                eptr++;                eptr++;
5707                ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);                ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);
5708                }                }
# Line 5551  for (;;) Line 5720  for (;;)
5720                  break;                  break;
5721                  }                  }
5722                if (IS_NEWLINE(eptr)) break;                if (IS_NEWLINE(eptr)) break;
5723                  if (md->partial != 0 &&    /* Take care with CRLF partial */
5724                      eptr + 1 >= md->end_subject &&
5725                      NLBLOCK->nltype == NLTYPE_FIXED &&
5726                      NLBLOCK->nllen == 2 &&
5727                      RAWUCHAR(eptr) == NLBLOCK->nl[0])
5728                    {
5729                    md->hitend = TRUE;
5730                    if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
5731                    }
5732                eptr++;                eptr++;
5733                ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);                ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);
5734                }                }
# Line 5600  for (;;) Line 5778  for (;;)
5778                break;                break;
5779                }                }
5780              GETCHARLEN(c, eptr, len);              GETCHARLEN(c, eptr, len);
5781              if (c == 0x000d)              if (c == CHAR_CR)
5782                {                {
5783                if (++eptr >= md->end_subject) break;                if (++eptr >= md->end_subject) break;
5784                if (*eptr == 0x000a) eptr++;                if (RAWUCHAR(eptr) == CHAR_LF) eptr++;
5785                }                }
5786              else              else
5787                {                {
5788                if (c != 0x000a &&                if (c != CHAR_LF &&
5789                    (md->bsr_anycrlf ||                    (md->bsr_anycrlf ||
5790                     (c != 0x000b && c != 0x000c &&                     (c != CHAR_VT && c != CHAR_FF && c != CHAR_NEL
5791                      c != 0x0085 && c != 0x2028 && c != 0x2029)))  #ifndef EBCDIC
5792                        && c != 0x2028 && c != 0x2029
5793    #endif  /* Not EBCDIC */
5794                        )))
5795                  break;                  break;
5796                eptr += len;                eptr += len;
5797                }                }
# Line 5631  for (;;) Line 5812  for (;;)
5812              GETCHARLEN(c, eptr, len);              GETCHARLEN(c, eptr, len);
5813              switch(c)              switch(c)
5814                {                {
5815                  HSPACE_CASES: gotspace = TRUE; break;
5816                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;  
5817                }                }
5818              if (gotspace == (ctype == OP_NOT_HSPACE)) break;              if (gotspace == (ctype == OP_NOT_HSPACE)) break;
5819              eptr += len;              eptr += len;
# Line 5673  for (;;) Line 5834  for (;;)
5834              GETCHARLEN(c, eptr, len);              GETCHARLEN(c, eptr, len);
5835              switch(c)              switch(c)
5836                {                {
5837                  VSPACE_CASES: gotspace = TRUE; break;
5838                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;  
5839                }                }
5840              if (gotspace == (ctype == OP_NOT_VSPACE)) break;              if (gotspace == (ctype == OP_NOT_VSPACE)) break;
5841              eptr += len;              eptr += len;
# Line 5796  for (;;) Line 5949  for (;;)
5949            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
5950            if (eptr-- == pp) break;        /* Stop if tried at original pos */            if (eptr-- == pp) break;        /* Stop if tried at original pos */
5951            BACKCHAR(eptr);            BACKCHAR(eptr);
5952            if (ctype == OP_ANYNL && eptr > pp  && *eptr == '\n' &&            if (ctype == OP_ANYNL && eptr > pp  && RAWUCHAR(eptr) == CHAR_NL &&
5953                eptr[-1] == '\r') eptr--;                RAWUCHAR(eptr - 1) == CHAR_CR) eptr--;
5954            }            }
5955          }          }
5956        else        else
# Line 5815  for (;;) Line 5968  for (;;)
5968                break;                break;
5969                }                }
5970              if (IS_NEWLINE(eptr)) break;              if (IS_NEWLINE(eptr)) break;
5971                if (md->partial != 0 &&    /* Take care with CRLF partial */
5972                    eptr + 1 >= md->end_subject &&
5973                    NLBLOCK->nltype == NLTYPE_FIXED &&
5974                    NLBLOCK->nllen == 2 &&
5975                    *eptr == NLBLOCK->nl[0])
5976                  {
5977                  md->hitend = TRUE;
5978                  if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
5979                  }
5980              eptr++;              eptr++;
5981              }              }
5982            break;            break;
# Line 5839  for (;;) Line 6001  for (;;)
6001                break;                break;
6002                }                }
6003              c = *eptr;              c = *eptr;
6004              if (c == 0x000d)              if (c == CHAR_CR)
6005                {                {
6006                if (++eptr >= md->end_subject) break;                if (++eptr >= md->end_subject) break;
6007                if (*eptr == 0x000a) eptr++;                if (*eptr == CHAR_LF) eptr++;
6008                }                }
6009              else              else
6010                {                {
6011                if (c != 0x000a && (md->bsr_anycrlf ||                if (c != CHAR_LF && (md->bsr_anycrlf ||
6012                  (c != 0x000b && c != 0x000c && c != 0x0085                   (c != CHAR_VT && c != CHAR_FF && c != CHAR_NEL
6013  #ifdef COMPILE_PCRE16  #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
6014                  && c != 0x2028 && c != 0x2029                   && c != 0x2028 && c != 0x2029
6015  #endif  #endif
6016                  ))) break;                   ))) break;
6017                eptr++;                eptr++;
6018                }                }
6019              }              }
# Line 5865  for (;;) Line 6027  for (;;)
6027                SCHECK_PARTIAL();                SCHECK_PARTIAL();
6028                break;                break;
6029                }                }
6030              c = *eptr;              switch(*eptr)
6031              if (c == 0x09 || c == 0x20 || c == 0xa0                {
6032  #ifdef COMPILE_PCRE16                default: eptr++; break;
6033                || c == 0x1680 || c == 0x180e || (c >= 0x2000 && c <= 0x200A)                HSPACE_BYTE_CASES:
6034                || c == 0x202f || c == 0x205f || c == 0x3000  #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
6035                  HSPACE_MULTIBYTE_CASES:
6036  #endif  #endif
6037                ) break;                goto ENDLOOP00;
6038              eptr++;                }
6039              }              }
6040              ENDLOOP00:
6041            break;            break;
6042    
6043            case OP_HSPACE:            case OP_HSPACE:
# Line 5884  for (;;) Line 6048  for (;;)
6048                SCHECK_PARTIAL();                SCHECK_PARTIAL();
6049                break;                break;
6050                }                }
6051              c = *eptr;              switch(*eptr)
6052              if (c != 0x09 && c != 0x20 && c != 0xa0                {
6053  #ifdef COMPILE_PCRE16                default: goto ENDLOOP01;
6054                && c != 0x1680 && c != 0x180e && (c < 0x2000 || c > 0x200A)                HSPACE_BYTE_CASES:
6055                && c != 0x202f && c != 0x205f && c != 0x3000  #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
6056                  HSPACE_MULTIBYTE_CASES:
6057  #endif  #endif
6058                ) break;                eptr++; break;
6059              eptr++;                }
6060              }              }
6061              ENDLOOP01:
6062            break;            break;
6063    
6064            case OP_NOT_VSPACE:            case OP_NOT_VSPACE:
# Line 5903  for (;;) Line 6069  for (;;)
6069                SCHECK_PARTIAL();                SCHECK_PARTIAL();
6070                break;                break;
6071                }                }
6072              c = *eptr;              switch(*eptr)
6073              if (c == 0x0a || c == 0x0b || c == 0x0c || c == 0x0d || c == 0x85                {
6074  #ifdef COMPILE_PCRE16                default: eptr++; break;
6075                || c == 0x2028 || c == 0x2029                VSPACE_BYTE_CASES:
6076    #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
6077                  VSPACE_MULTIBYTE_CASES:
6078  #endif  #endif
6079                ) break;                goto ENDLOOP02;
6080              eptr++;                }
6081              }              }
6082              ENDLOOP02:
6083            break;            break;
6084    
6085            case OP_VSPACE:            case OP_VSPACE:
# Line 5921  for (;;) Line 6090  for (;;)
6090                SCHECK_PARTIAL();                SCHECK_PARTIAL();
6091                break;                break;
6092                }                }
6093              c = *eptr;              switch(*eptr)
6094              if (c != 0x0a && c != 0x0b && c != 0x0c && c != 0x0d && c != 0x85                {
6095  #ifdef COMPILE_PCRE16                default: goto ENDLOOP03;
6096                && c != 0x2028 && c != 0x2029                VSPACE_BYTE_CASES:
6097    #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
6098                  VSPACE_MULTIBYTE_CASES:
6099  #endif  #endif
6100                ) break;                eptr++; break;
6101              eptr++;                }
6102              }              }
6103              ENDLOOP03:
6104            break;            break;
6105    
6106            case OP_NOT_DIGIT:            case OP_NOT_DIGIT:
# Line 6025  for (;;) Line 6197  for (;;)
6197            RMATCH(eptr, ecode, offset_top, md, eptrb, RM47);            RMATCH(eptr, ecode, offset_top, md, eptrb, RM47);
6198            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
6199            eptr--;            eptr--;
6200            if (ctype == OP_ANYNL && eptr > pp  && *eptr == '\n' &&            if (ctype == OP_ANYNL && eptr > pp  && *eptr == CHAR_LF &&
6201                eptr[-1] == '\r') eptr--;                eptr[-1] == CHAR_CR) eptr--;
6202            }            }
6203          }          }
6204    
# Line 6076  switch (frame->Xwhere) Line 6248  switch (frame->Xwhere)
6248    LBL(32) LBL(34) LBL(42) LBL(46)    LBL(32) LBL(34) LBL(42) LBL(46)
6249  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
6250    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)
6251    LBL(59) LBL(60) LBL(61) LBL(62)    LBL(59) LBL(60) LBL(61) LBL(62) LBL(67) LBL(68)
6252  #endif  /* SUPPORT_UCP */  #endif  /* SUPPORT_UCP */
6253  #endif  /* SUPPORT_UTF */  #endif  /* SUPPORT_UTF */
6254    default:    default:
6255    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);  
   
6256    return PCRE_ERROR_INTERNAL;    return PCRE_ERROR_INTERNAL;
6257    }    }
6258  #undef LBL  #undef LBL
# Line 6145  Undefine all the macros that were define Line 6314  Undefine all the macros that were define
6314  ***************************************************************************/  ***************************************************************************/
6315    
6316    
6317    #ifdef NO_RECURSE
6318    /*************************************************
6319    *          Release allocated heap frames         *
6320    *************************************************/
6321    
6322    /* This function releases all the allocated frames. The base frame is on the
6323    machine stack, and so must not be freed.
6324    
6325    Argument: the address of the base frame
6326    Returns:  nothing
6327    */
6328    
6329    static void
6330    release_match_heapframes (heapframe *frame_base)
6331    {
6332    heapframe *nextframe = frame_base->Xnextframe;
6333    while (nextframe != NULL)
6334      {
6335      heapframe *oldframe = nextframe;
6336      nextframe = nextframe->Xnextframe;
6337      (PUBL(stack_free))(oldframe);
6338      }
6339    }
6340    #endif
6341    
6342    
6343  /*************************************************  /*************************************************
6344  *         Execute a Regular Expression           *  *         Execute a Regular Expression           *
# Line 6170  Returns:          > 0 => success; value Line 6364  Returns:          > 0 => success; value
6364                   < -1 => some kind of unexpected problem                   < -1 => some kind of unexpected problem
6365  */  */
6366    
6367  #ifdef COMPILE_PCRE8  #if defined COMPILE_PCRE8
6368  PCRE_EXP_DEFN int PCRE_CALL_CONVENTION  PCRE_EXP_DEFN int PCRE_CALL_CONVENTION
6369  pcre_exec(const pcre *argument_re, const pcre_extra *extra_data,  pcre_exec(const pcre *argument_re, const pcre_extra *extra_data,
6370    PCRE_SPTR subject, int length, int start_offset, int options, int *offsets,    PCRE_SPTR subject, int length, int start_offset, int options, int *offsets,
6371    int offsetcount)    int offsetcount)
6372  #else  #elif defined COMPILE_PCRE16
6373  PCRE_EXP_DEFN int PCRE_CALL_CONVENTION  PCRE_EXP_DEFN int PCRE_CALL_CONVENTION
6374  pcre16_exec(const pcre16 *argument_re, const pcre16_extra *extra_data,  pcre16_exec(const pcre16 *argument_re, const pcre16_extra *extra_data,
6375    PCRE_SPTR16 subject, int length, int start_offset, int options, int *offsets,    PCRE_SPTR16 subject, int length, int start_offset, int options, int *offsets,
6376    int offsetcount)    int offsetcount)
6377    #elif defined COMPILE_PCRE32
6378    PCRE_EXP_DEFN int PCRE_CALL_CONVENTION
6379    pcre32_exec(const pcre32 *argument_re, const pcre32_extra *extra_data,
6380      PCRE_SPTR32 subject, int length, int start_offset, int options, int *offsets,
6381      int offsetcount)
6382  #endif  #endif
6383  {  {
6384  int rc, ocount, arg_offset_max;  int rc, ocount, arg_offset_max;
# Line 6202  const pcre_uint8 *start_bits = NULL; Line 6401  const pcre_uint8 *start_bits = NULL;
6401  PCRE_PUCHAR start_match = (PCRE_PUCHAR)subject + start_offset;  PCRE_PUCHAR start_match = (PCRE_PUCHAR)subject + start_offset;
6402  PCRE_PUCHAR end_subject;  PCRE_PUCHAR end_subject;
6403  PCRE_PUCHAR start_partial = NULL;  PCRE_PUCHAR start_partial = NULL;
6404    PCRE_PUCHAR match_partial;
6405  PCRE_PUCHAR req_char_ptr = start_match - 1;  PCRE_PUCHAR req_char_ptr = start_match - 1;
6406    
6407  const pcre_study_data *study;  const pcre_study_data *study;
6408  const REAL_PCRE *re = (const REAL_PCRE *)argument_re;  const REAL_PCRE *re = (const REAL_PCRE *)argument_re;
6409    
6410  /* Check for the special magic call that measures the size of the stack used  #ifdef NO_RECURSE
6411  per recursive call of match(). */  heapframe frame_zero;
6412    frame_zero.Xprevframe = NULL;            /* Marks the top level */
6413    frame_zero.Xnextframe = NULL;            /* None are allocated yet */
6414    md->match_frames_base = &frame_zero;
6415    #endif
6416    
6417    /* Check for the special magic call that measures the size of the stack used
6418    per recursive call of match(). Without the funny casting for sizeof, a Windows
6419    compiler gave this error: "unary minus operator applied to unsigned type,
6420    result still unsigned". Hopefully the cast fixes that. */
6421    
6422  if (re == NULL && extra_data == NULL && subject == NULL && length == -1)  if (re == NULL && extra_data == NULL && subject == NULL && length == -999 &&
6423        start_offset == -999)
6424  #ifdef NO_RECURSE  #ifdef NO_RECURSE
6425    return -sizeof(heapframe);    return -((int)sizeof(heapframe));
6426  #else  #else
6427    return match(NULL, NULL, NULL, 0, NULL, NULL, 0);    return match(NULL, NULL, NULL, 0, NULL, NULL, 0);
6428  #endif  #endif
6429    
6430  /* Plausibility checks */  /* Plausibility checks */
6431    
6432  if ((options & ~PUBLIC_EXEC_OPTIONS) != 0) return PCRE_ERROR_BADOPTION;  if ((options & ~PUBLIC_EXEC_OPTIONS) != 0) return PCRE_ERROR_BADOPTION;
6433  if (re == NULL || subject == NULL || (offsets == NULL && offsetcount > 0))  if (re == NULL || subject == NULL || (offsets == NULL && offsetcount > 0))
6434    return PCRE_ERROR_NULL;    return PCRE_ERROR_NULL;
6435  if (offsetcount < 0) return PCRE_ERROR_BADCOUNT;  if (offsetcount < 0) return PCRE_ERROR_BADCOUNT;
6436    if (length < 0) return PCRE_ERROR_BADLENGTH;
6437  if (start_offset < 0 || start_offset > length) return PCRE_ERROR_BADOFFSET;  if (start_offset < 0 || start_offset > length) return PCRE_ERROR_BADOFFSET;
6438    
6439  /* 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 6471  if (utf && (options & PCRE_NO_UTF8_CHECK
6471        offsets[0] = erroroffset;        offsets[0] = erroroffset;
6472        offsets[1] = errorcode;        offsets[1] = errorcode;
6473        }        }
6474  #ifdef COMPILE_PCRE16  #if defined COMPILE_PCRE8
     return (errorcode <= PCRE_UTF16_ERR1 && md->partial > 1)?  
       PCRE_ERROR_SHORTUTF16 : PCRE_ERROR_BADUTF16;  
 #else  
6475      return (errorcode <= PCRE_UTF8_ERR5 && md->partial > 1)?      return (errorcode <= PCRE_UTF8_ERR5 && md->partial > 1)?
6476        PCRE_ERROR_SHORTUTF8 : PCRE_ERROR_BADUTF8;        PCRE_ERROR_SHORTUTF8 : PCRE_ERROR_BADUTF8;
6477    #elif defined COMPILE_PCRE16
6478        return (errorcode <= PCRE_UTF16_ERR1 && md->partial > 1)?
6479          PCRE_ERROR_SHORTUTF16 : PCRE_ERROR_BADUTF16;
6480    #elif defined COMPILE_PCRE32
6481        return PCRE_ERROR_BADUTF32;
6482  #endif  #endif
6483      }      }
6484    #if defined COMPILE_PCRE8 || defined COMPILE_PCRE16
6485    /* 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. */
6486    if (start_offset > 0 && start_offset < length &&    if (start_offset > 0 && start_offset < length &&
6487        NOT_FIRSTCHAR(((PCRE_PUCHAR)subject)[start_offset]))        NOT_FIRSTCHAR(((PCRE_PUCHAR)subject)[start_offset]))
6488      return PCRE_ERROR_BADUTF8_OFFSET;      return PCRE_ERROR_BADUTF8_OFFSET;
6489    #endif
6490    }    }
6491  #endif  #endif
6492    
6493  /* If the pattern was successfully studied with JIT support, run the JIT  /* If the pattern was successfully studied with JIT support, run the JIT
6494  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
6495  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
6496  an unsupported flag is set. In particular, JIT does not support partial  an unsupported flag is set. */
 matching. */  
6497    
6498  #ifdef SUPPORT_JIT  #ifdef SUPPORT_JIT
6499  if (extra_data != NULL  if (extra_data != NULL
6500      && (extra_data->flags & PCRE_EXTRA_EXECUTABLE_JIT) != 0      && (extra_data->flags & (PCRE_EXTRA_EXECUTABLE_JIT |
6501                                 PCRE_EXTRA_TABLES)) == PCRE_EXTRA_EXECUTABLE_JIT
6502      && extra_data->executable_jit != NULL      && extra_data->executable_jit != NULL
6503      && (extra_data->flags & PCRE_EXTRA_TABLES) == 0      && (options & ~PUBLIC_JIT_EXEC_OPTIONS) == 0)
6504      && (options & ~(PCRE_NO_UTF8_CHECK | PCRE_NOTBOL | PCRE_NOTEOL |    {
6505                      PCRE_NOTEMPTY | PCRE_NOTEMPTY_ATSTART)) == 0)    rc = PRIV(jit_exec)(extra_data, (const pcre_uchar *)subject, length,
6506    return PRIV(jit_exec)(re, extra_data->executable_jit,         start_offset, options, offsets, offsetcount);
6507      (const pcre_uchar *)subject, length, start_offset, options,  
6508      ((extra_data->flags & PCRE_EXTRA_MATCH_LIMIT) == 0)    /* PCRE_ERROR_NULL means that the selected normal or partial matching
6509      ? MATCH_LIMIT : extra_data->match_limit, offsets, offsetcount);    mode is not compiled. In this case we simply fallback to interpreter. */
6510    
6511      if (rc != PCRE_ERROR_JIT_BADOPTION) return rc;
6512      }
6513  #endif  #endif
6514    
6515  /* 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 6570  end_subject = md->end_subject;
6570  md->endonly = (re->options & PCRE_DOLLAR_ENDONLY) != 0;  md->endonly = (re->options & PCRE_DOLLAR_ENDONLY) != 0;
6571  md->use_ucp = (re->options & PCRE_UCP) != 0;  md->use_ucp = (re->options & PCRE_UCP) != 0;
6572  md->jscript_compat = (re->options & PCRE_JAVASCRIPT_COMPAT) != 0;  md->jscript_compat = (re->options & PCRE_JAVASCRIPT_COMPAT) != 0;
6573  md->ignore_skip_arg = FALSE;  md->ignore_skip_arg = 0;
6574    
6575  /* Some options are unpacked into BOOL variables in the hope that testing  /* Some options are unpacked into BOOL variables in the hope that testing
6576  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 6680  if (re->top_backref > 0 && re->top_backr
6680    DPRINTF(("Got memory to hold back references\n"));    DPRINTF(("Got memory to hold back references\n"));
6681    }    }
6682  else md->offset_vector = offsets;  else md->offset_vector = offsets;
   
6683  md->offset_end = ocount;  md->offset_end = ocount;
6684  md->offset_max = (2*ocount)/3;  md->offset_max = (2*ocount)/3;
6685  md->offset_overflow = FALSE;  md->capture_last = 0;
 md->capture_last = -1;  
6686    
6687  /* Reset the working variable associated with each extraction. These should  /* Reset the working variable associated with each extraction. These should
6688  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 6710  if (!anchored)
6710    if ((re->flags & PCRE_FIRSTSET) != 0)    if ((re->flags & PCRE_FIRSTSET) != 0)
6711      {      {
6712      has_first_char = TRUE;      has_first_char = TRUE;
6713      first_char = first_char2 = re->first_char;      first_char = first_char2 = (pcre_uchar)(re->first_char);
6714      if ((re->flags & PCRE_FCH_CASELESS) != 0)      if ((re->flags & PCRE_FCH_CASELESS) != 0)
6715        {        {
6716        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 6732  character" set. */
6732  if ((re->flags & PCRE_REQCHSET) != 0)  if ((re->flags & PCRE_REQCHSET) != 0)
6733    {    {
6734    has_req_char = TRUE;    has_req_char = TRUE;
6735    req_char = req_char2 = re->req_char;    req_char = req_char2 = (pcre_uchar)(re->req_char);
6736    if ((re->flags & PCRE_RCH_CASELESS) != 0)    if ((re->flags & PCRE_RCH_CASELESS) != 0)
6737      {      {
6738      req_char2 = TABLE_GET(req_char, md->fcc, req_char);      req_char2 = TABLE_GET(req_char, md->fcc, req_char);
# Line 6575  for(;;) Line 6790  for(;;)
6790    
6791      if (has_first_char)      if (has_first_char)
6792        {        {
6793          pcre_uchar smc;
6794    
6795        if (first_char != first_char2)        if (first_char != first_char2)
6796          while (start_match < end_subject &&          while (start_match < end_subject &&
6797              *start_match != first_char && *start_match != first_char2)            (smc = RAWUCHARTEST(start_match)) != first_char && smc != first_char2)
6798            start_match++;            start_match++;
6799        else        else
6800          while (start_match < end_subject && *start_match != first_char)          while (start_match < end_subject && RAWUCHARTEST(start_match) != first_char)
6801            start_match++;            start_match++;
6802        }        }
6803    
# Line 6612  for(;;) Line 6829  for(;;)
6829          if (start_match[-1] == CHAR_CR &&          if (start_match[-1] == CHAR_CR &&
6830               (md->nltype == NLTYPE_ANY || md->nltype == NLTYPE_ANYCRLF) &&               (md->nltype == NLTYPE_ANY || md->nltype == NLTYPE_ANYCRLF) &&
6831               start_match < end_subject &&               start_match < end_subject &&
6832               *start_match == CHAR_NL)               RAWUCHARTEST(start_match) == CHAR_NL)
6833            start_match++;            start_match++;
6834          }          }
6835        }        }
# Line 6623  for(;;) Line 6840  for(;;)
6840        {        {
6841        while (start_match < end_subject)        while (start_match < end_subject)
6842          {          {
6843          register unsigned int c = *start_match;          register pcre_uint32 c = RAWUCHARTEST(start_match);
6844  #ifndef COMPILE_PCRE8  #ifndef COMPILE_PCRE8
6845          if (c > 255) c = 255;          if (c > 255) c = 255;
6846  #endif  #endif
# Line 6691  for(;;) Line 6908  for(;;)
6908            {            {
6909            while (p < end_subject)            while (p < end_subject)
6910              {              {
6911              register int pp = *p++;              register pcre_uint32 pp = RAWUCHARINCTEST(p);
6912              if (pp == req_char || pp == req_char2) { p--; break; }              if (pp == req_char || pp == req_char2) { p--; break; }
6913              }              }
6914            }            }
# Line 6699  for(;;) Line 6916  for(;;)
6916            {            {
6917            while (p < end_subject)            while (p < end_subject)
6918              {              {
6919              if (*p++ == req_char) { p--; break; }              if (RAWUCHARINCTEST(p) == req_char) { p--; break; }
6920              }              }
6921            }            }
6922    
# Line 6735  for(;;) Line 6952  for(;;)
6952    md->match_call_count = 0;    md->match_call_count = 0;
6953    md->match_function_type = 0;    md->match_function_type = 0;
6954    md->end_offset_top = 0;    md->end_offset_top = 0;
6955      md->skip_arg_count = 0;
6956    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);
6957    if (md->hitend && start_partial == NULL) start_partial = md->start_used_ptr;    if (md->hitend && start_partial == NULL)
6958        {
6959        start_partial = md->start_used_ptr;
6960        match_partial = start_match;
6961        }
6962    
6963    switch(rc)    switch(rc)
6964      {      {
# Line 6749  for(;;) Line 6971  for(;;)
6971    
6972      case MATCH_SKIP_ARG:      case MATCH_SKIP_ARG:
6973      new_start_match = start_match;      new_start_match = start_match;
6974      md->ignore_skip_arg = TRUE;      md->ignore_skip_arg = md->skip_arg_count;
6975      break;      break;
6976    
6977      /* 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
6978      same as the match we have just done, treat it as NOMATCH. */      greater than the match we have just done, treat it as NOMATCH. */
6979    
6980      case MATCH_SKIP:      case MATCH_SKIP:
6981      if (md->start_match_ptr != start_match)      if (md->start_match_ptr > start_match)
6982        {        {
6983        new_start_match = md->start_match_ptr;        new_start_match = md->start_match_ptr;
6984        break;        break;
# Line 6764  for(;;) Line 6986  for(;;)
6986      /* Fall through */      /* Fall through */
6987    
6988      /* NOMATCH and PRUNE advance by one character. THEN at this level acts      /* NOMATCH and PRUNE advance by one character. THEN at this level acts
6989      exactly like PRUNE. Unset the ignore SKIP-with-argument flag. */      exactly like PRUNE. Unset ignore SKIP-with-argument. */
6990    
6991      case MATCH_NOMATCH:      case MATCH_NOMATCH:
6992      case MATCH_PRUNE:      case MATCH_PRUNE:
6993      case MATCH_THEN:      case MATCH_THEN:
6994      md->ignore_skip_arg = FALSE;      md->ignore_skip_arg = 0;
6995      new_start_match = start_match + 1;      new_start_match = start_match + 1;
6996  #ifdef SUPPORT_UTF  #ifdef SUPPORT_UTF
6997      if (utf)      if (utf)
# Line 6862  if (rc == MATCH_MATCH || rc == MATCH_ACC Line 7084  if (rc == MATCH_MATCH || rc == MATCH_ACC
7084          (arg_offset_max - 2) * sizeof(int));          (arg_offset_max - 2) * sizeof(int));
7085        DPRINTF(("Copied offsets from temporary memory\n"));        DPRINTF(("Copied offsets from temporary memory\n"));
7086        }        }
7087      if (md->end_offset_top > arg_offset_max) md->offset_overflow = TRUE;      if (md->end_offset_top > arg_offset_max) md->capture_last |= OVFLBIT;
7088      DPRINTF(("Freeing temporary memory\n"));      DPRINTF(("Freeing temporary memory\n"));
7089      (PUBL(free))(md->offset_vector);      (PUBL(free))(md->offset_vector);
7090      }      }
# Line 6870  if (rc == MATCH_MATCH || rc == MATCH_ACC Line 7092  if (rc == MATCH_MATCH || rc == MATCH_ACC
7092    /* 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
7093    too many to fit into the vector. */    too many to fit into the vector. */
7094    
7095    rc = (md->offset_overflow && md->end_offset_top >= arg_offset_max)?    rc = ((md->capture_last & OVFLBIT) != 0 &&
7096             md->end_offset_top >= arg_offset_max)?
7097      0 : md->end_offset_top/2;      0 : md->end_offset_top/2;
7098    
7099    /* 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 7109  if (rc == MATCH_MATCH || rc == MATCH_ACC
7109      {      {
7110      register int *iptr, *iend;      register int *iptr, *iend;
7111      int resetcount = 2 + re->top_bracket * 2;      int resetcount = 2 + re->top_bracket * 2;
7112      if (resetcount > offsetcount) resetcount = ocount;      if (resetcount > offsetcount) resetcount = offsetcount;
7113      iptr = offsets + md->end_offset_top;      iptr = offsets + md->end_offset_top;
7114      iend = offsets + resetcount;      iend = offsets + resetcount;
7115      while (iptr < iend) *iptr++ = -1;      while (iptr < iend) *iptr++ = -1;
# Line 6907  if (rc == MATCH_MATCH || rc == MATCH_ACC Line 7130  if (rc == MATCH_MATCH || rc == MATCH_ACC
7130    if (extra_data != NULL && (extra_data->flags & PCRE_EXTRA_MARK) != 0)    if (extra_data != NULL && (extra_data->flags & PCRE_EXTRA_MARK) != 0)
7131      *(extra_data->mark) = (pcre_uchar *)md->mark;      *(extra_data->mark) = (pcre_uchar *)md->mark;
7132    DPRINTF((">>>> returning %d\n", rc));    DPRINTF((">>>> returning %d\n", rc));
7133    #ifdef NO_RECURSE
7134      release_match_heapframes(&frame_zero);
7135    #endif
7136    return rc;    return rc;
7137    }    }
7138    
# Line 6924  if (using_temporary_offsets) Line 7150  if (using_temporary_offsets)
7150  if (rc != MATCH_NOMATCH && rc != PCRE_ERROR_PARTIAL)  if (rc != MATCH_NOMATCH && rc != PCRE_ERROR_PARTIAL)
7151    {    {
7152    DPRINTF((">>>> error: returning %d\n", rc));    DPRINTF((">>>> error: returning %d\n", rc));
7153    #ifdef NO_RECURSE
7154      release_match_heapframes(&frame_zero);
7155    #endif
7156    return rc;    return rc;
7157    }    }
7158    
# Line 6937  if (start_partial != NULL) Line 7166  if (start_partial != NULL)
7166      {      {
7167      offsets[0] = (int)(start_partial - (PCRE_PUCHAR)subject);      offsets[0] = (int)(start_partial - (PCRE_PUCHAR)subject);
7168      offsets[1] = (int)(end_subject - (PCRE_PUCHAR)subject);      offsets[1] = (int)(end_subject - (PCRE_PUCHAR)subject);
7169        if (offsetcount > 2)
7170          offsets[2] = (int)(match_partial - (PCRE_PUCHAR)subject);
7171      }      }
7172    rc = PCRE_ERROR_PARTIAL;    rc = PCRE_ERROR_PARTIAL;
7173    }    }
# Line 6953  else Line 7184  else
7184    
7185  if (extra_data != NULL && (extra_data->flags & PCRE_EXTRA_MARK) != 0)  if (extra_data != NULL && (extra_data->flags & PCRE_EXTRA_MARK) != 0)
7186    *(extra_data->mark) = (pcre_uchar *)md->nomatch_mark;    *(extra_data->mark) = (pcre_uchar *)md->nomatch_mark;
7187    #ifdef NO_RECURSE
7188      release_match_heapframes(&frame_zero);
7189    #endif
7190  return rc;  return rc;
7191  }  }
7192    

Legend:
Removed from v.895  
changed lines
  Added in v.1302

  ViewVC Help
Powered by ViewVC 1.1.5