/[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 1257 by ph10, Fri Feb 22 20:20:30 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 93  because the offset vector is always a mu Line 106  because the offset vector is always a mu
106  static const char rep_min[] = { 0, 0, 1, 1, 0, 0 };  static const char rep_min[] = { 0, 0, 1, 1, 0, 0 };
107  static const char rep_max[] = { 0, 0, 0, 0, 1, 1 };  static const char rep_max[] = { 0, 0, 0, 0, 1, 1 };
108    
   
   
109  #ifdef PCRE_DEBUG  #ifdef PCRE_DEBUG
110  /*************************************************  /*************************************************
111  *        Debugging function to print chars       *  *        Debugging function to print chars       *
# Line 115  Returns:     nothing Line 126  Returns:     nothing
126  static void  static void
127  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)
128  {  {
129  unsigned int c;  pcre_uint32 c;
130    BOOL utf = md->utf;
131  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;
132  while (length-- > 0)  while (length-- > 0)
133    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);
134  }  }
135  #endif  #endif
136    
# Line 140  Arguments: Line 152  Arguments:
152    md          points to match data block    md          points to match data block
153    caseless    TRUE if caseless    caseless    TRUE if caseless
154    
155  Returns:      < 0 if not matched, otherwise the number of subject bytes matched  Returns:      >= 0 the number of subject bytes matched
156                  -1 no match
157                  -2 partial match; always given if at end subject
158  */  */
159    
160  static int  static int
# Line 149  match_ref(int offset, register PCRE_PUCH Line 163  match_ref(int offset, register PCRE_PUCH
163  {  {
164  PCRE_PUCHAR eptr_start = eptr;  PCRE_PUCHAR eptr_start = eptr;
165  register PCRE_PUCHAR p = md->start_subject + md->offset_vector[offset];  register PCRE_PUCHAR p = md->start_subject + md->offset_vector[offset];
166    #ifdef SUPPORT_UTF
167    BOOL utf = md->utf;
168    #endif
169    
170  #ifdef PCRE_DEBUG  #ifdef PCRE_DEBUG
171  if (eptr >= md->end_subject)  if (eptr >= md->end_subject)
# Line 163  pchars(p, length, FALSE, md); Line 180  pchars(p, length, FALSE, md);
180  printf("\n");  printf("\n");
181  #endif  #endif
182    
183  /* Always fail if reference not set (and not JavaScript compatible). */  /* Always fail if reference not set (and not JavaScript compatible - in that
184    case the length is passed as zero). */
185    
186  if (length < 0) return -1;  if (length < 0) return -1;
187    
# Line 175  if (caseless) Line 193  if (caseless)
193    {    {
194  #ifdef SUPPORT_UTF  #ifdef SUPPORT_UTF
195  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
196    if (md->utf)    if (utf)
197      {      {
198      /* 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
199      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
200      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.
201      (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
202      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
203      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
204      reference, not along the subject (earlier code did this wrong). */      length along the reference, not along the subject (earlier code did this
205        wrong). */
206    
207      PCRE_PUCHAR endptr = p + length;      PCRE_PUCHAR endptr = p + length;
208      while (p < endptr)      while (p < endptr)
209        {        {
210        int c, d;        pcre_uint32 c, d;
211        if (eptr >= md->end_subject) return -1;        const ucd_record *ur;
212          if (eptr >= md->end_subject) return -2;   /* Partial match */
213        GETCHARINC(c, eptr);        GETCHARINC(c, eptr);
214        GETCHARINC(d, p);        GETCHARINC(d, p);
215        if (c != d && c != UCD_OTHERCASE(d)) return -1;        ur = GET_UCD(d);
216          if (c != d && c != d + ur->other_case)
217            {
218            const pcre_uint32 *pp = PRIV(ucd_caseless_sets) + ur->caseset;
219            for (;;)
220              {
221              if (c < *pp) return -1;
222              if (c == *pp++) break;
223              }
224            }
225        }        }
226      }      }
227    else    else
# Line 202  if (caseless) Line 231  if (caseless)
231    /* 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
232    is no UCP support. */    is no UCP support. */
233      {      {
     if (eptr + length > md->end_subject) return -1;  
234      while (length-- > 0)      while (length-- > 0)
235        {        {
236        if (TABLE_GET(*p, md->lcc, *p) != TABLE_GET(*eptr, md->lcc, *eptr)) return -1;        pcre_uint32 cc, cp;
237          if (eptr >= md->end_subject) return -2;   /* Partial match */
238          cc = RAWUCHARTEST(eptr);
239          cp = RAWUCHARTEST(p);
240          if (TABLE_GET(cp, md->lcc, cp) != TABLE_GET(cc, md->lcc, cc)) return -1;
241        p++;        p++;
242        eptr++;        eptr++;
243        }        }
# Line 217  are in UTF-8 mode. */ Line 249  are in UTF-8 mode. */
249    
250  else  else
251    {    {
252    if (eptr + length > md->end_subject) return -1;    while (length-- > 0)
253    while (length-- > 0) if (*p++ != *eptr++) return -1;      {
254        if (eptr >= md->end_subject) return -2;   /* Partial match */
255        if (RAWUCHARINCTEST(p) != RAWUCHARINCTEST(eptr)) return -1;
256        }
257    }    }
258    
259  return (int)(eptr - eptr_start);  return (int)(eptr - eptr_start);
# Line 273  enum { RM1=1, RM2,  RM3,  RM4,  RM5,  RM Line 308  enum { RM1=1, RM2,  RM3,  RM4,  RM5,  RM
308         RM31,  RM32, RM33, RM34, RM35, RM36, RM37, RM38, RM39, RM40,         RM31,  RM32, RM33, RM34, RM35, RM36, RM37, RM38, RM39, RM40,
309         RM41,  RM42, RM43, RM44, RM45, RM46, RM47, RM48, RM49, RM50,         RM41,  RM42, RM43, RM44, RM45, RM46, RM47, RM48, RM49, RM50,
310         RM51,  RM52, RM53, RM54, RM55, RM56, RM57, RM58, RM59, RM60,         RM51,  RM52, RM53, RM54, RM55, RM56, RM57, RM58, RM59, RM60,
311         RM61,  RM62, RM63, RM64, RM65, RM66 };         RM61,  RM62, RM63, RM64, RM65, RM66, RM67 };
312    
313  /* 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
314  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 326  actually used in this definition. */
326    }    }
327  #define RRETURN(ra) \  #define RRETURN(ra) \
328    { \    { \
329    printf("match() returned %d from line %d ", ra, __LINE__); \    printf("match() returned %d from line %d\n", ra, __LINE__); \
330    return ra; \    return ra; \
331    }    }
332  #else  #else
# Line 311  argument of match(), which never changes Line 346  argument of match(), which never changes
346    
347  #define RMATCH(ra,rb,rc,rd,re,rw)\  #define RMATCH(ra,rb,rc,rd,re,rw)\
348    {\    {\
349    heapframe *newframe = (heapframe *)(PUBL(stack_malloc))(sizeof(heapframe));\    heapframe *newframe = frame->Xnextframe;\
350    if (newframe == NULL) RRETURN(PCRE_ERROR_NOMEMORY);\    if (newframe == NULL)\
351    frame->Xwhere = rw; \      {\
352        newframe = (heapframe *)(PUBL(stack_malloc))(sizeof(heapframe));\
353        if (newframe == NULL) RRETURN(PCRE_ERROR_NOMEMORY);\
354        newframe->Xnextframe = NULL;\
355        frame->Xnextframe = newframe;\
356        }\
357      frame->Xwhere = rw;\
358    newframe->Xeptr = ra;\    newframe->Xeptr = ra;\
359    newframe->Xecode = rb;\    newframe->Xecode = rb;\
360    newframe->Xmstart = mstart;\    newframe->Xmstart = mstart;\
# Line 332  argument of match(), which never changes Line 373  argument of match(), which never changes
373    {\    {\
374    heapframe *oldframe = frame;\    heapframe *oldframe = frame;\
375    frame = oldframe->Xprevframe;\    frame = oldframe->Xprevframe;\
   if (oldframe != &frame_zero) (PUBL(stack_free))(oldframe);\  
376    if (frame != NULL)\    if (frame != NULL)\
377      {\      {\
378      rrc = ra;\      rrc = ra;\
# Line 346  argument of match(), which never changes Line 386  argument of match(), which never changes
386    
387  typedef struct heapframe {  typedef struct heapframe {
388    struct heapframe *Xprevframe;    struct heapframe *Xprevframe;
389      struct heapframe *Xnextframe;
390    
391    /* Function arguments that may change */    /* Function arguments that may change */
392    
# Line 376  typedef struct heapframe { Line 417  typedef struct heapframe {
417    
418  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
419    int Xprop_type;    int Xprop_type;
420    int Xprop_value;    unsigned int Xprop_value;
421    int Xprop_fail_result;    int Xprop_fail_result;
422    int Xoclength;    int Xoclength;
423    pcre_uchar Xocchars[6];    pcre_uchar Xocchars[6];
# Line 389  typedef struct heapframe { Line 430  typedef struct heapframe {
430    int Xlength;    int Xlength;
431    int Xmax;    int Xmax;
432    int Xmin;    int Xmin;
433    int Xnumber;    unsigned int Xnumber;
434    int Xoffset;    int Xoffset;
435    int Xop;    unsigned int Xop;
436    int Xsave_capture_last;    pcre_int32 Xsave_capture_last;
437    int Xsave_offset1, Xsave_offset2, Xsave_offset3;    int Xsave_offset1, Xsave_offset2, Xsave_offset3;
438    int Xstacksave[REC_STACK_SAVE_MAX];    int Xstacksave[REC_STACK_SAVE_MAX];
439    
# Line 477  so they can be ordinary variables in all Line 518  so they can be ordinary variables in all
518    
519  register int  rrc;         /* Returns from recursive calls */  register int  rrc;         /* Returns from recursive calls */
520  register int  i;           /* Used for loops not involving calls to RMATCH() */  register int  i;           /* Used for loops not involving calls to RMATCH() */
521  register unsigned int c;   /* Character values not kept over RMATCH() calls */  register pcre_uint32 c;    /* Character values not kept over RMATCH() calls */
522  register BOOL utf;         /* Local copy of UTF flag for speed */  register BOOL utf;         /* Local copy of UTF flag for speed */
523    
524  BOOL minimize, possessive; /* Quantifier options */  BOOL minimize, possessive; /* Quantifier options */
# Line 487  int condcode; Line 528  int condcode;
528  /* 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
529  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
530  frame on the stack here; subsequent instantiations are obtained from the heap  frame on the stack here; subsequent instantiations are obtained from the heap
531  whenever RMATCH() does a "recursion". See the macro definitions above. Putting  whenever RMATCH() does a "recursion". See the macro definitions above. Putting
532  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
533  boost in many cases where there is not much "recursion". */  boost in many cases where there is not much "recursion". */
534    
535  #ifdef NO_RECURSE  #ifdef NO_RECURSE
536  heapframe frame_zero;  heapframe *frame = (heapframe *)md->match_frames_base;
 heapframe *frame = &frame_zero;  
 frame->Xprevframe = NULL;            /* Marks the top level */  
537    
538  /* Copy in the original argument variables */  /* Copy in the original argument variables */
539    
# Line 596  BOOL prev_is_word; Line 635  BOOL prev_is_word;
635    
636  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
637  int prop_type;  int prop_type;
638  int prop_value;  unsigned int prop_value;
639  int prop_fail_result;  int prop_fail_result;
640  int oclength;  int oclength;
641  pcre_uchar occhars[6];  pcre_uchar occhars[6];
# Line 607  int ctype; Line 646  int ctype;
646  int length;  int length;
647  int max;  int max;
648  int min;  int min;
649  int number;  unsigned int number;
650  int offset;  int offset;
651  int op;  unsigned int op;
652  int save_capture_last;  pcre_int32 save_capture_last;
653  int save_offset1, save_offset2, save_offset3;  int save_offset1, save_offset2, save_offset3;
654  int stacksave[REC_STACK_SAVE_MAX];  int stacksave[REC_STACK_SAVE_MAX];
655    
656  eptrblock newptrb;  eptrblock newptrb;
657    
658  /* 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
659  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
660  recursion. The second argument (ecode) being NULL triggers this behaviour. It  recursion. The second argument (ecode) being NULL triggers this behaviour. It
661  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
662  size. */  size. */
663    
664  if (ecode == NULL)  if (ecode == NULL)
# Line 631  if (ecode == NULL) Line 670  if (ecode == NULL)
670      int len = (char *)&rdepth - (char *)eptr;      int len = (char *)&rdepth - (char *)eptr;
671      return (len > 0)? -len : len;      return (len > 0)? -len : len;
672      }      }
673    }    }
674  #endif     /* NO_RECURSE */  #endif     /* NO_RECURSE */
675    
676  /* 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 767  for (;;)
767      unaltered. */      unaltered. */
768    
769      else if (rrc == MATCH_SKIP_ARG &&      else if (rrc == MATCH_SKIP_ARG &&
770          STRCMP_UC_UC(ecode + 2, md->start_match_ptr) == 0)          STRCMP_UC_UC_TEST(ecode + 2, md->start_match_ptr) == 0)
771        {        {
772        md->start_match_ptr = eptr;        md->start_match_ptr = eptr;
773        RRETURN(MATCH_SKIP);        RRETURN(MATCH_SKIP);
# Line 838  for (;;) Line 877  for (;;)
877      case OP_ONCE_NC:      case OP_ONCE_NC:
878      prev = ecode;      prev = ecode;
879      saved_eptr = eptr;      saved_eptr = eptr;
880      save_mark = md->mark;      save_mark = md->mark;
881      do      do
882        {        {
883        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 896  for (;;)
896    
897        if (rrc != MATCH_NOMATCH) RRETURN(rrc);        if (rrc != MATCH_NOMATCH) RRETURN(rrc);
898        ecode += GET(ecode,1);        ecode += GET(ecode,1);
899        md->mark = save_mark;        md->mark = save_mark;
900        }        }
901      while (*ecode == OP_ALT);      while (*ecode == OP_ALT);
902    
# Line 897  for (;;) Line 936  for (;;)
936        }        }
937      else  /* OP_KETRMAX */      else  /* OP_KETRMAX */
938        {        {
       md->match_function_type = MATCH_CBEGROUP;  
939        RMATCH(eptr, prev, offset_top, md, eptrb, RM66);        RMATCH(eptr, prev, offset_top, md, eptrb, RM66);
940        if (rrc != MATCH_NOMATCH) RRETURN(rrc);        if (rrc != MATCH_NOMATCH) RRETURN(rrc);
941        ecode += 1 + LINK_SIZE;        ecode += 1 + LINK_SIZE;
# Line 937  for (;;) Line 975  for (;;)
975        save_offset2 = md->offset_vector[offset+1];        save_offset2 = md->offset_vector[offset+1];
976        save_offset3 = md->offset_vector[md->offset_end - number];        save_offset3 = md->offset_vector[md->offset_end - number];
977        save_capture_last = md->capture_last;        save_capture_last = md->capture_last;
978        save_mark = md->mark;        save_mark = md->mark;
979    
980        DPRINTF(("saving %d %d %d\n", save_offset1, save_offset2, save_offset3));        DPRINTF(("saving %d %d %d\n", save_offset1, save_offset2, save_offset3));
981        md->offset_vector[md->offset_end - number] =        md->offset_vector[md->offset_end - number] =
# Line 1026  for (;;) Line 1064  for (;;)
1064    
1065      for (;;)      for (;;)
1066        {        {
1067        if (op >= OP_SBRA || op == OP_ONCE) md->match_function_type = MATCH_CBEGROUP;        if (op >= OP_SBRA || op == OP_ONCE)
1068            md->match_function_type = MATCH_CBEGROUP;
1069    
1070        /* 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
1071        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 1080  for (;;)
1080        /* In all other cases, we have to make another call to match(). */        /* In all other cases, we have to make another call to match(). */
1081    
1082        save_mark = md->mark;        save_mark = md->mark;
1083          save_capture_last = md->capture_last;
1084        RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md, eptrb,        RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md, eptrb,
1085          RM2);          RM2);
1086    
1087        /* See comment in the code for capturing groups above about handling        /* See comment in the code for capturing groups above about handling
1088        THEN. */        THEN. */
1089    
# Line 1070  for (;;) Line 1110  for (;;)
1110          RRETURN(rrc);          RRETURN(rrc);
1111          }          }
1112        ecode += GET(ecode, 1);        ecode += GET(ecode, 1);
1113        md->mark = save_mark;        md->mark = save_mark;
1114        if (*ecode != OP_ALT) break;        if (*ecode != OP_ALT) break;
1115          md->capture_last = save_capture_last;
1116        }        }
1117    
1118      RRETURN(MATCH_NOMATCH);      RRETURN(MATCH_NOMATCH);
# Line 1193  for (;;) Line 1234  for (;;)
1234      POSSESSIVE_NON_CAPTURE:      POSSESSIVE_NON_CAPTURE:
1235      matched_once = FALSE;      matched_once = FALSE;
1236      code_offset = (int)(ecode - md->start_code);      code_offset = (int)(ecode - md->start_code);
1237        save_capture_last = md->capture_last;
1238    
1239      for (;;)      for (;;)
1240        {        {
# Line 1222  for (;;) Line 1264  for (;;)
1264        if (rrc != MATCH_NOMATCH) RRETURN(rrc);        if (rrc != MATCH_NOMATCH) RRETURN(rrc);
1265        ecode += GET(ecode, 1);        ecode += GET(ecode, 1);
1266        if (*ecode != OP_ALT) break;        if (*ecode != OP_ALT) break;
1267          md->capture_last = save_capture_last;
1268        }        }
1269    
1270      if (matched_once || allow_zero)      if (matched_once || allow_zero)
# Line 1253  for (;;) Line 1296  for (;;)
1296          cb.version          = 2;   /* Version 1 of the callout block */          cb.version          = 2;   /* Version 1 of the callout block */
1297          cb.callout_number   = ecode[LINK_SIZE+2];          cb.callout_number   = ecode[LINK_SIZE+2];
1298          cb.offset_vector    = md->offset_vector;          cb.offset_vector    = md->offset_vector;
1299  #ifdef COMPILE_PCRE8  #if defined COMPILE_PCRE8
1300          cb.subject          = (PCRE_SPTR)md->start_subject;          cb.subject          = (PCRE_SPTR)md->start_subject;
1301  #else  #elif defined COMPILE_PCRE16
1302          cb.subject          = (PCRE_SPTR16)md->start_subject;          cb.subject          = (PCRE_SPTR16)md->start_subject;
1303    #elif defined COMPILE_PCRE32
1304            cb.subject          = (PCRE_SPTR32)md->start_subject;
1305  #endif  #endif
1306          cb.subject_length   = (int)(md->end_subject - md->start_subject);          cb.subject_length   = (int)(md->end_subject - md->start_subject);
1307          cb.start_match      = (int)(mstart - md->start_subject);          cb.start_match      = (int)(mstart - md->start_subject);
# Line 1264  for (;;) Line 1309  for (;;)
1309          cb.pattern_position = GET(ecode, LINK_SIZE + 3);          cb.pattern_position = GET(ecode, LINK_SIZE + 3);
1310          cb.next_item_length = GET(ecode, 3 + 2*LINK_SIZE);          cb.next_item_length = GET(ecode, 3 + 2*LINK_SIZE);
1311          cb.capture_top      = offset_top/2;          cb.capture_top      = offset_top/2;
1312          cb.capture_last     = md->capture_last;          cb.capture_last     = md->capture_last & CAPLMASK;
1313            /* Internal change requires this for API compatibility. */
1314            if (cb.capture_last == 0) cb.capture_last = -1;
1315          cb.callout_data     = md->callout_data;          cb.callout_data     = md->callout_data;
1316          cb.mark             = md->nomatch_mark;          cb.mark             = md->nomatch_mark;
1317          if ((rrc = (*PUBL(callout))(&cb)) > 0) RRETURN(MATCH_NOMATCH);          if ((rrc = (*PUBL(callout))(&cb)) > 0) RRETURN(MATCH_NOMATCH);
# Line 1286  for (;;) Line 1333  for (;;)
1333          }          }
1334        else        else
1335          {          {
1336          int recno = GET2(ecode, LINK_SIZE + 2);   /* Recursion group number*/          unsigned int recno = GET2(ecode, LINK_SIZE + 2);   /* Recursion group number*/
1337          condition = (recno == RREF_ANY || recno == md->recursive->group_num);          condition = (recno == RREF_ANY || recno == md->recursive->group_num);
1338    
1339          /* 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 1405  for (;;)
1405    
1406        if (!condition && condcode == OP_NCREF)        if (!condition && condcode == OP_NCREF)
1407          {          {
1408          int refno = offset >> 1;          unsigned int refno = offset >> 1;
1409          pcre_uchar *slotA = md->name_table;          pcre_uchar *slotA = md->name_table;
1410    
1411          for (i = 0; i < md->name_count; i++)          for (i = 0; i < md->name_count; i++)
# Line 1486  for (;;) Line 1533  for (;;)
1533      to close any currently open capturing brackets. */      to close any currently open capturing brackets. */
1534    
1535      case OP_CLOSE:      case OP_CLOSE:
1536      number = GET2(ecode, 1);      number = GET2(ecode, 1);   /* Must be less than 65536 */
1537      offset = number << 1;      offset = number << 1;
1538    
1539  #ifdef PCRE_DEBUG  #ifdef PCRE_DEBUG
# Line 1494  for (;;) Line 1541  for (;;)
1541        printf("\n");        printf("\n");
1542  #endif  #endif
1543    
1544      md->capture_last = number;      md->capture_last = (md->capture_last & OVFLMASK) | number;
1545      if (offset >= md->offset_max) md->offset_overflow = TRUE; else      if (offset >= md->offset_max) md->capture_last |= OVFLBIT; else
1546        {        {
1547        md->offset_vector[offset] =        md->offset_vector[offset] =
1548          md->offset_vector[md->offset_end - number];          md->offset_vector[md->offset_end - number];
# Line 1549  for (;;) Line 1596  for (;;)
1596    
1597      case OP_ASSERT:      case OP_ASSERT:
1598      case OP_ASSERTBACK:      case OP_ASSERTBACK:
1599      save_mark = md->mark;      save_mark = md->mark;
1600      if (md->match_function_type == MATCH_CONDASSERT)      if (md->match_function_type == MATCH_CONDASSERT)
1601        {        {
1602        condassert = TRUE;        condassert = TRUE;
# Line 1565  for (;;) Line 1612  for (;;)
1612          mstart = md->start_match_ptr;   /* In case \K reset it */          mstart = md->start_match_ptr;   /* In case \K reset it */
1613          break;          break;
1614          }          }
1615          md->mark = save_mark;
1616    
1617        /* PCRE does not allow THEN to escape beyond an assertion; it is treated        /* A COMMIT failure must fail the entire assertion, without trying any
1618        as NOMATCH. */        subsequent branches. */
1619    
1620          if (rrc == MATCH_COMMIT) RRETURN(MATCH_NOMATCH);
1621    
1622          /* PCRE does not allow THEN to escape beyond an assertion; it
1623          is treated as NOMATCH. */
1624    
1625        if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN) RRETURN(rrc);        if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN) RRETURN(rrc);
1626        ecode += GET(ecode, 1);        ecode += GET(ecode, 1);
       md->mark = save_mark;  
1627        }        }
1628      while (*ecode == OP_ALT);      while (*ecode == OP_ALT);
1629    
# Line 1595  for (;;) Line 1647  for (;;)
1647    
1648      case OP_ASSERT_NOT:      case OP_ASSERT_NOT:
1649      case OP_ASSERTBACK_NOT:      case OP_ASSERTBACK_NOT:
1650      save_mark = md->mark;      save_mark = md->mark;
1651      if (md->match_function_type == MATCH_CONDASSERT)      if (md->match_function_type == MATCH_CONDASSERT)
1652        {        {
1653        condassert = TRUE;        condassert = TRUE;
# Line 1606  for (;;) Line 1658  for (;;)
1658      do      do
1659        {        {
1660        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, NULL, RM5);        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, NULL, RM5);
1661        md->mark = save_mark;        md->mark = save_mark;
1662        if (rrc == MATCH_MATCH || rrc == MATCH_ACCEPT) RRETURN(MATCH_NOMATCH);        if (rrc == MATCH_MATCH || rrc == MATCH_ACCEPT) RRETURN(MATCH_NOMATCH);
1663        if (rrc == MATCH_SKIP || rrc == MATCH_PRUNE || rrc == MATCH_COMMIT)        if (rrc == MATCH_SKIP || rrc == MATCH_PRUNE || rrc == MATCH_COMMIT)
1664          {          {
# Line 1671  for (;;) Line 1723  for (;;)
1723        cb.version          = 2;   /* Version 1 of the callout block */        cb.version          = 2;   /* Version 1 of the callout block */
1724        cb.callout_number   = ecode[1];        cb.callout_number   = ecode[1];
1725        cb.offset_vector    = md->offset_vector;        cb.offset_vector    = md->offset_vector;
1726  #ifdef COMPILE_PCRE8  #if defined COMPILE_PCRE8
1727        cb.subject          = (PCRE_SPTR)md->start_subject;        cb.subject          = (PCRE_SPTR)md->start_subject;
1728  #else  #elif defined COMPILE_PCRE16
1729        cb.subject          = (PCRE_SPTR16)md->start_subject;        cb.subject          = (PCRE_SPTR16)md->start_subject;
1730    #elif defined COMPILE_PCRE32
1731          cb.subject          = (PCRE_SPTR32)md->start_subject;
1732  #endif  #endif
1733        cb.subject_length   = (int)(md->end_subject - md->start_subject);        cb.subject_length   = (int)(md->end_subject - md->start_subject);
1734        cb.start_match      = (int)(mstart - md->start_subject);        cb.start_match      = (int)(mstart - md->start_subject);
# Line 1682  for (;;) Line 1736  for (;;)
1736        cb.pattern_position = GET(ecode, 2);        cb.pattern_position = GET(ecode, 2);
1737        cb.next_item_length = GET(ecode, 2 + LINK_SIZE);        cb.next_item_length = GET(ecode, 2 + LINK_SIZE);
1738        cb.capture_top      = offset_top/2;        cb.capture_top      = offset_top/2;
1739        cb.capture_last     = md->capture_last;        cb.capture_last     = md->capture_last & CAPLMASK;
1740          /* Internal change requires this for API compatibility. */
1741          if (cb.capture_last == 0) cb.capture_last = -1;
1742        cb.callout_data     = md->callout_data;        cb.callout_data     = md->callout_data;
1743        cb.mark             = md->nomatch_mark;        cb.mark             = md->nomatch_mark;
1744        if ((rrc = (*PUBL(callout))(&cb)) > 0) RRETURN(MATCH_NOMATCH);        if ((rrc = (*PUBL(callout))(&cb)) > 0) RRETURN(MATCH_NOMATCH);
# Line 1711  for (;;) Line 1767  for (;;)
1767      case OP_RECURSE:      case OP_RECURSE:
1768        {        {
1769        recursion_info *ri;        recursion_info *ri;
1770        int recno;        unsigned int recno;
1771    
1772        callpat = md->start_code + GET(ecode, 1);        callpat = md->start_code + GET(ecode, 1);
1773        recno = (callpat == md->start_code)? 0 :        recno = (callpat == md->start_code)? 0 :
# Line 1728  for (;;) Line 1784  for (;;)
1784        /* Add to "recursing stack" */        /* Add to "recursing stack" */
1785    
1786        new_recursive.group_num = recno;        new_recursive.group_num = recno;
1787          new_recursive.saved_capture_last = md->capture_last;
1788        new_recursive.subject_position = eptr;        new_recursive.subject_position = eptr;
1789        new_recursive.prevrec = md->recursive;        new_recursive.prevrec = md->recursive;
1790        md->recursive = &new_recursive;        md->recursive = &new_recursive;
# Line 1751  for (;;) Line 1808  for (;;)
1808              new_recursive.saved_max * sizeof(int));              new_recursive.saved_max * sizeof(int));
1809    
1810        /* OK, now we can do the recursion. After processing each alternative,        /* OK, now we can do the recursion. After processing each alternative,
1811        restore the offset data. If there were nested recursions, md->recursive        restore the offset data and the last captured value. If there were nested
1812        might be changed, so reset it before looping. */        recursions, md->recursive might be changed, so reset it before looping.
1813          */
1814    
1815        DPRINTF(("Recursing into group %d\n", new_recursive.group_num));        DPRINTF(("Recursing into group %d\n", new_recursive.group_num));
1816        cbegroup = (*callpat >= OP_SBRA);        cbegroup = (*callpat >= OP_SBRA);
# Line 1763  for (;;) Line 1821  for (;;)
1821            md, eptrb, RM6);            md, eptrb, RM6);
1822          memcpy(md->offset_vector, new_recursive.offset_save,          memcpy(md->offset_vector, new_recursive.offset_save,
1823              new_recursive.saved_max * sizeof(int));              new_recursive.saved_max * sizeof(int));
1824            md->capture_last = new_recursive.saved_capture_last;
1825          md->recursive = new_recursive.prevrec;          md->recursive = new_recursive.prevrec;
1826          if (rrc == MATCH_MATCH || rrc == MATCH_ACCEPT)          if (rrc == MATCH_MATCH || rrc == MATCH_ACCEPT)
1827            {            {
# Line 1779  for (;;) Line 1838  for (;;)
1838            goto RECURSION_MATCHED;        /* Exit loop; end processing */            goto RECURSION_MATCHED;        /* Exit loop; end processing */
1839            }            }
1840    
1841          /* PCRE does not allow THEN to escape beyond a recursion; it is treated          /* PCRE does not allow THEN or COMMIT to escape beyond a recursion; it
1842          as NOMATCH. */          is treated as NOMATCH. */
1843    
1844          else if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN)          else if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN &&
1845                     rrc != MATCH_COMMIT)
1846            {            {
1847            DPRINTF(("Recursion gave error %d\n", rrc));            DPRINTF(("Recursion gave error %d\n", rrc));
1848            if (new_recursive.offset_save != stacksave)            if (new_recursive.offset_save != stacksave)
# Line 1912  for (;;) Line 1972  for (;;)
1972    
1973        /* Deal with capturing */        /* Deal with capturing */
1974    
1975        md->capture_last = number;        md->capture_last = (md->capture_last & OVFLMASK) | number;
1976        if (offset >= md->offset_max) md->offset_overflow = TRUE; else        if (offset >= md->offset_max) md->capture_last |= OVFLBIT; else
1977          {          {
1978          /* If offset is greater than offset_top, it means that we are          /* If offset is greater than offset_top, it means that we are
1979          "skipping" a capturing group, and that group's offsets must be marked          "skipping" a capturing group, and that group's offsets must be marked
# Line 1993  for (;;) Line 2053  for (;;)
2053          }          }
2054        if (*prev >= OP_SBRA)    /* Could match an empty string */        if (*prev >= OP_SBRA)    /* Could match an empty string */
2055          {          {
         md->match_function_type = MATCH_CBEGROUP;  
2056          RMATCH(eptr, prev, offset_top, md, eptrb, RM50);          RMATCH(eptr, prev, offset_top, md, eptrb, RM50);
2057          RRETURN(rrc);          RRETURN(rrc);
2058          }          }
# Line 2002  for (;;) Line 2061  for (;;)
2061        }        }
2062      else  /* OP_KETRMAX */      else  /* OP_KETRMAX */
2063        {        {
       if (*prev >= OP_SBRA) md->match_function_type = MATCH_CBEGROUP;  
2064        RMATCH(eptr, prev, offset_top, md, eptrb, RM13);        RMATCH(eptr, prev, offset_top, md, eptrb, RM13);
2065        if (rrc == MATCH_ONCE && md->once_target == prev) rrc = MATCH_NOMATCH;        if (rrc == MATCH_ONCE && md->once_target == prev) rrc = MATCH_NOMATCH;
2066        if (rrc != MATCH_NOMATCH) RRETURN(rrc);        if (rrc != MATCH_NOMATCH) RRETURN(rrc);
# Line 2059  for (;;) Line 2117  for (;;)
2117    
2118      case OP_DOLLM:      case OP_DOLLM:
2119      if (eptr < md->end_subject)      if (eptr < md->end_subject)
2120        { if (!IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH); }        {
2121          if (!IS_NEWLINE(eptr))
2122            {
2123            if (md->partial != 0 &&
2124                eptr + 1 >= md->end_subject &&
2125                NLBLOCK->nltype == NLTYPE_FIXED &&
2126                NLBLOCK->nllen == 2 &&
2127                RAWUCHARTEST(eptr) == NLBLOCK->nl[0])
2128              {
2129              md->hitend = TRUE;
2130              if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
2131              }
2132            RRETURN(MATCH_NOMATCH);
2133            }
2134          }
2135      else      else
2136        {        {
2137        if (md->noteol) RRETURN(MATCH_NOMATCH);        if (md->noteol) RRETURN(MATCH_NOMATCH);
# Line 2091  for (;;) Line 2163  for (;;)
2163      ASSERT_NL_OR_EOS:      ASSERT_NL_OR_EOS:
2164      if (eptr < md->end_subject &&      if (eptr < md->end_subject &&
2165          (!IS_NEWLINE(eptr) || eptr != md->end_subject - md->nllen))          (!IS_NEWLINE(eptr) || eptr != md->end_subject - md->nllen))
2166          {
2167          if (md->partial != 0 &&
2168              eptr + 1 >= md->end_subject &&
2169              NLBLOCK->nltype == NLTYPE_FIXED &&
2170              NLBLOCK->nllen == 2 &&
2171              RAWUCHARTEST(eptr) == NLBLOCK->nl[0])
2172            {
2173            md->hitend = TRUE;
2174            if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
2175            }
2176        RRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2177          }
2178    
2179      /* Either at end of string or \n before end. */      /* Either at end of string or \n before end. */
2180    
# Line 2219  for (;;) Line 2302  for (;;)
2302        }        }
2303      break;      break;
2304    
2305      /* Match a single character type; inline for speed */      /* Match any single character type except newline; have to take care with
2306        CRLF newlines and partial matching. */
2307    
2308      case OP_ANY:      case OP_ANY:
2309      if (IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH);      if (IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH);
2310        if (md->partial != 0 &&
2311            eptr + 1 >= md->end_subject &&
2312            NLBLOCK->nltype == NLTYPE_FIXED &&
2313            NLBLOCK->nllen == 2 &&
2314            RAWUCHARTEST(eptr) == NLBLOCK->nl[0])
2315          {
2316          md->hitend = TRUE;
2317          if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
2318          }
2319    
2320      /* Fall through */      /* Fall through */
2321    
2322        /* Match any single character whatsoever. */
2323    
2324      case OP_ALLANY:      case OP_ALLANY:
2325      if (eptr >= md->end_subject)   /* DO NOT merge the eptr++ here; it must */      if (eptr >= md->end_subject)   /* DO NOT merge the eptr++ here; it must */
2326        {                            /* not be updated before SCHECK_PARTIAL. */        {                            /* not be updated before SCHECK_PARTIAL. */
# Line 2364  for (;;) Line 2460  for (;;)
2460        {        {
2461        default: RRETURN(MATCH_NOMATCH);        default: RRETURN(MATCH_NOMATCH);
2462    
2463        case 0x000d:        case CHAR_CR:
2464        if (eptr < md->end_subject && *eptr == 0x0a) eptr++;        if (eptr >= md->end_subject)
2465            {
2466            SCHECK_PARTIAL();
2467            }
2468          else if (RAWUCHARTEST(eptr) == CHAR_LF) eptr++;
2469        break;        break;
2470    
2471        case 0x000a:        case CHAR_LF:
2472        break;        break;
2473    
2474        case 0x000b:        case CHAR_VT:
2475        case 0x000c:        case CHAR_FF:
2476        case 0x0085:        case CHAR_NEL:
2477    #ifndef EBCDIC
2478        case 0x2028:        case 0x2028:
2479        case 0x2029:        case 0x2029:
2480    #endif  /* Not EBCDIC */
2481        if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);        if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);
2482        break;        break;
2483        }        }
# Line 2391  for (;;) Line 2493  for (;;)
2493      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2494      switch(c)      switch(c)
2495        {        {
2496          HSPACE_CASES: RRETURN(MATCH_NOMATCH);  /* Byte and multibyte cases */
2497        default: break;        default: break;
       case 0x09:      /* HT */  
       case 0x20:      /* SPACE */  
       case 0xa0:      /* NBSP */  
       case 0x1680:    /* OGHAM SPACE MARK */  
       case 0x180e:    /* MONGOLIAN VOWEL SEPARATOR */  
       case 0x2000:    /* EN QUAD */  
       case 0x2001:    /* EM QUAD */  
       case 0x2002:    /* EN SPACE */  
       case 0x2003:    /* EM SPACE */  
       case 0x2004:    /* THREE-PER-EM SPACE */  
       case 0x2005:    /* FOUR-PER-EM SPACE */  
       case 0x2006:    /* SIX-PER-EM SPACE */  
       case 0x2007:    /* FIGURE SPACE */  
       case 0x2008:    /* PUNCTUATION SPACE */  
       case 0x2009:    /* THIN SPACE */  
       case 0x200A:    /* HAIR SPACE */  
       case 0x202f:    /* NARROW NO-BREAK SPACE */  
       case 0x205f:    /* MEDIUM MATHEMATICAL SPACE */  
       case 0x3000:    /* IDEOGRAPHIC SPACE */  
       RRETURN(MATCH_NOMATCH);  
2498        }        }
2499      ecode++;      ecode++;
2500      break;      break;
# Line 2425  for (;;) Line 2508  for (;;)
2508      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2509      switch(c)      switch(c)
2510        {        {
2511          HSPACE_CASES: break;  /* Byte and multibyte cases */
2512        default: RRETURN(MATCH_NOMATCH);        default: RRETURN(MATCH_NOMATCH);
       case 0x09:      /* HT */  
       case 0x20:      /* SPACE */  
       case 0xa0:      /* NBSP */  
       case 0x1680:    /* OGHAM SPACE MARK */  
       case 0x180e:    /* MONGOLIAN VOWEL SEPARATOR */  
       case 0x2000:    /* EN QUAD */  
       case 0x2001:    /* EM QUAD */  
       case 0x2002:    /* EN SPACE */  
       case 0x2003:    /* EM SPACE */  
       case 0x2004:    /* THREE-PER-EM SPACE */  
       case 0x2005:    /* FOUR-PER-EM SPACE */  
       case 0x2006:    /* SIX-PER-EM SPACE */  
       case 0x2007:    /* FIGURE SPACE */  
       case 0x2008:    /* PUNCTUATION SPACE */  
       case 0x2009:    /* THIN SPACE */  
       case 0x200A:    /* HAIR SPACE */  
       case 0x202f:    /* NARROW NO-BREAK SPACE */  
       case 0x205f:    /* MEDIUM MATHEMATICAL SPACE */  
       case 0x3000:    /* IDEOGRAPHIC SPACE */  
       break;  
2513        }        }
2514      ecode++;      ecode++;
2515      break;      break;
# Line 2459  for (;;) Line 2523  for (;;)
2523      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2524      switch(c)      switch(c)
2525        {        {
2526          VSPACE_CASES: RRETURN(MATCH_NOMATCH);
2527        default: break;        default: break;
       case 0x0a:      /* LF */  
       case 0x0b:      /* VT */  
       case 0x0c:      /* FF */  
       case 0x0d:      /* CR */  
       case 0x85:      /* NEL */  
       case 0x2028:    /* LINE SEPARATOR */  
       case 0x2029:    /* PARAGRAPH SEPARATOR */  
       RRETURN(MATCH_NOMATCH);  
2528        }        }
2529      ecode++;      ecode++;
2530      break;      break;
# Line 2481  for (;;) Line 2538  for (;;)
2538      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2539      switch(c)      switch(c)
2540        {        {
2541          VSPACE_CASES: break;
2542        default: RRETURN(MATCH_NOMATCH);        default: RRETURN(MATCH_NOMATCH);
       case 0x0a:      /* LF */  
       case 0x0b:      /* VT */  
       case 0x0c:      /* FF */  
       case 0x0d:      /* CR */  
       case 0x85:      /* NEL */  
       case 0x2028:    /* LINE SEPARATOR */  
       case 0x2029:    /* PARAGRAPH SEPARATOR */  
       break;  
2543        }        }
2544      ecode++;      ecode++;
2545      break;      break;
# Line 2507  for (;;) Line 2557  for (;;)
2557        }        }
2558      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2559        {        {
2560          const pcre_uint32 *cp;
2561        const ucd_record *prop = GET_UCD(c);        const ucd_record *prop = GET_UCD(c);
2562    
2563        switch(ecode[1])        switch(ecode[1])
# Line 2567  for (;;) Line 2618  for (;;)
2618            RRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
2619          break;          break;
2620    
2621            case PT_CLIST:
2622            cp = PRIV(ucd_caseless_sets) + ecode[2];
2623            for (;;)
2624              {
2625              if (c < *cp)
2626                { if (op == OP_PROP) { RRETURN(MATCH_NOMATCH); } else break; }
2627              if (c == *cp++)
2628                { if (op == OP_PROP) break; else { RRETURN(MATCH_NOMATCH); } }
2629              }
2630            break;
2631    
2632          /* This should never occur */          /* This should never occur */
2633    
2634          default:          default:
# Line 2586  for (;;) Line 2648  for (;;)
2648        SCHECK_PARTIAL();        SCHECK_PARTIAL();
2649        RRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2650        }        }
2651      GETCHARINCTEST(c, eptr);      else
     if (UCD_CATEGORY(c) == ucp_M) RRETURN(MATCH_NOMATCH);  
     while (eptr < md->end_subject)  
2652        {        {
2653        int len = 1;        int lgb, rgb;
2654        if (!utf) c = *eptr; else { GETCHARLEN(c, eptr, len); }        GETCHARINCTEST(c, eptr);
2655        if (UCD_CATEGORY(c) != ucp_M) break;        lgb = UCD_GRAPHBREAK(c);
2656        eptr += len;        while (eptr < md->end_subject)
2657            {
2658            int len = 1;
2659            if (!utf) c = *eptr; else { GETCHARLEN(c, eptr, len); }
2660            rgb = UCD_GRAPHBREAK(c);
2661            if ((PRIV(ucp_gbtable)[lgb] & (1 << rgb)) == 0) break;
2662            lgb = rgb;
2663            eptr += len;
2664            }
2665        }        }
2666        CHECK_PARTIAL();
2667      ecode++;      ecode++;
2668      break;      break;
2669  #endif  #endif  /* SUPPORT_UCP */
2670    
2671    
2672      /* 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 2729  for (;;)
2729        default:               /* No repeat follows */        default:               /* No repeat follows */
2730        if ((length = match_ref(offset, eptr, length, md, caseless)) < 0)        if ((length = match_ref(offset, eptr, length, md, caseless)) < 0)
2731          {          {
2732            if (length == -2) eptr = md->end_subject;   /* Partial match */
2733          CHECK_PARTIAL();          CHECK_PARTIAL();
2734          RRETURN(MATCH_NOMATCH);          RRETURN(MATCH_NOMATCH);
2735          }          }
# Line 2685  for (;;) Line 2755  for (;;)
2755        int slength;        int slength;
2756        if ((slength = match_ref(offset, eptr, length, md, caseless)) < 0)        if ((slength = match_ref(offset, eptr, length, md, caseless)) < 0)
2757          {          {
2758            if (slength == -2) eptr = md->end_subject;   /* Partial match */
2759          CHECK_PARTIAL();          CHECK_PARTIAL();
2760          RRETURN(MATCH_NOMATCH);          RRETURN(MATCH_NOMATCH);
2761          }          }
# Line 2708  for (;;) Line 2779  for (;;)
2779          if (fi >= max) RRETURN(MATCH_NOMATCH);          if (fi >= max) RRETURN(MATCH_NOMATCH);
2780          if ((slength = match_ref(offset, eptr, length, md, caseless)) < 0)          if ((slength = match_ref(offset, eptr, length, md, caseless)) < 0)
2781            {            {
2782              if (slength == -2) eptr = md->end_subject;   /* Partial match */
2783            CHECK_PARTIAL();            CHECK_PARTIAL();
2784            RRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
2785            }            }
# Line 2726  for (;;) Line 2798  for (;;)
2798          int slength;          int slength;
2799          if ((slength = match_ref(offset, eptr, length, md, caseless)) < 0)          if ((slength = match_ref(offset, eptr, length, md, caseless)) < 0)
2800            {            {
2801            CHECK_PARTIAL();            /* Can't use CHECK_PARTIAL because we don't want to update eptr in
2802              the soft partial matching case. */
2803    
2804              if (slength == -2 && md->partial != 0 &&
2805                  md->end_subject > md->start_used_ptr)
2806                {
2807                md->hitend = TRUE;
2808                if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
2809                }
2810            break;            break;
2811            }            }
2812          eptr += slength;          eptr += slength;
2813          }          }
2814    
2815        while (eptr >= pp)        while (eptr >= pp)
2816          {          {
2817          RMATCH(eptr, ecode, offset_top, md, eptrb, RM15);          RMATCH(eptr, ecode, offset_top, md, eptrb, RM15);
# Line 3094  for (;;) Line 3175  for (;;)
3175          CHECK_PARTIAL();             /* Not SCHECK_PARTIAL() */          CHECK_PARTIAL();             /* Not SCHECK_PARTIAL() */
3176          RRETURN(MATCH_NOMATCH);          RRETURN(MATCH_NOMATCH);
3177          }          }
3178        while (length-- > 0) if (*ecode++ != *eptr++) RRETURN(MATCH_NOMATCH);        while (length-- > 0) if (*ecode++ != RAWUCHARINC(eptr)) RRETURN(MATCH_NOMATCH);
3179        }        }
3180      else      else
3181  #endif  #endif
# Line 3134  for (;;) Line 3215  for (;;)
3215    
3216        if (fc < 128)        if (fc < 128)
3217          {          {
3218          if (md->lcc[fc]          pcre_uint32 cc = RAWUCHAR(eptr);
3219              != TABLE_GET(*eptr, md->lcc, *eptr)) RRETURN(MATCH_NOMATCH);          if (md->lcc[fc] != TABLE_GET(cc, md->lcc, cc)) RRETURN(MATCH_NOMATCH);
3220          ecode++;          ecode++;
3221          eptr++;          eptr++;
3222          }          }
# Line 3146  for (;;) Line 3227  for (;;)
3227    
3228        else        else
3229          {          {
3230          unsigned int dc;          pcre_uint32 dc;
3231          GETCHARINC(dc, eptr);          GETCHARINC(dc, eptr);
3232          ecode += length;          ecode += length;
3233    
# Line 3256  for (;;) Line 3337  for (;;)
3337        if (length > 1)        if (length > 1)
3338          {          {
3339  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
3340          unsigned int othercase;          pcre_uint32 othercase;
3341          if (op >= OP_STARI &&     /* Caseless */          if (op >= OP_STARI &&     /* Caseless */
3342              (othercase = UCD_OTHERCASE(fc)) != fc)              (othercase = UCD_OTHERCASE(fc)) != fc)
3343            oclength = PRIV(ord2utf)(othercase, occhars);            oclength = PRIV(ord2utf)(othercase, occhars);
# Line 3360  for (;;) Line 3441  for (;;)
3441      maximizing, find the maximum number of characters and work backwards. */      maximizing, find the maximum number of characters and work backwards. */
3442    
3443      DPRINTF(("matching %c{%d,%d} against subject %.*s\n", fc, min, max,      DPRINTF(("matching %c{%d,%d} against subject %.*s\n", fc, min, max,
3444        max, eptr));        max, (char *)eptr));
3445    
3446      if (op >= OP_STARI)  /* Caseless */      if (op >= OP_STARI)  /* Caseless */
3447        {        {
# Line 3383  for (;;) Line 3464  for (;;)
3464    
3465        for (i = 1; i <= min; i++)        for (i = 1; i <= min; i++)
3466          {          {
3467            pcre_uint32 cc;                 /* Faster than pcre_uchar */
3468          if (eptr >= md->end_subject)          if (eptr >= md->end_subject)
3469            {            {
3470            SCHECK_PARTIAL();            SCHECK_PARTIAL();
3471            RRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
3472            }            }
3473          if (fc != *eptr && foc != *eptr) RRETURN(MATCH_NOMATCH);          cc = RAWUCHARTEST(eptr);
3474            if (fc != cc && foc != cc) RRETURN(MATCH_NOMATCH);
3475          eptr++;          eptr++;
3476          }          }
3477        if (min == max) continue;        if (min == max) continue;
# Line 3396  for (;;) Line 3479  for (;;)
3479          {          {
3480          for (fi = min;; fi++)          for (fi = min;; fi++)
3481            {            {
3482              pcre_uint32 cc;               /* Faster than pcre_uchar */
3483            RMATCH(eptr, ecode, offset_top, md, eptrb, RM24);            RMATCH(eptr, ecode, offset_top, md, eptrb, RM24);
3484            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3485            if (fi >= max) RRETURN(MATCH_NOMATCH);            if (fi >= max) RRETURN(MATCH_NOMATCH);
# Line 3404  for (;;) Line 3488  for (;;)
3488              SCHECK_PARTIAL();              SCHECK_PARTIAL();
3489              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
3490              }              }
3491            if (fc != *eptr && foc != *eptr) RRETURN(MATCH_NOMATCH);            cc = RAWUCHARTEST(eptr);
3492              if (fc != cc && foc != cc) RRETURN(MATCH_NOMATCH);
3493            eptr++;            eptr++;
3494            }            }
3495          /* Control never gets here */          /* Control never gets here */
# Line 3414  for (;;) Line 3499  for (;;)
3499          pp = eptr;          pp = eptr;
3500          for (i = min; i < max; i++)          for (i = min; i < max; i++)
3501            {            {
3502              pcre_uint32 cc;               /* Faster than pcre_uchar */
3503            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
3504              {              {
3505              SCHECK_PARTIAL();              SCHECK_PARTIAL();
3506              break;              break;
3507              }              }
3508            if (fc != *eptr && foc != *eptr) break;            cc = RAWUCHARTEST(eptr);
3509              if (fc != cc && foc != cc) break;
3510            eptr++;            eptr++;
3511            }            }
3512    
# Line 3447  for (;;) Line 3534  for (;;)
3534            SCHECK_PARTIAL();            SCHECK_PARTIAL();
3535            RRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
3536            }            }
3537          if (fc != *eptr++) RRETURN(MATCH_NOMATCH);          if (fc != RAWUCHARINCTEST(eptr)) RRETURN(MATCH_NOMATCH);
3538          }          }
3539    
3540        if (min == max) continue;        if (min == max) continue;
# Line 3464  for (;;) Line 3551  for (;;)
3551              SCHECK_PARTIAL();              SCHECK_PARTIAL();
3552              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
3553              }              }
3554            if (fc != *eptr++) RRETURN(MATCH_NOMATCH);            if (fc != RAWUCHARINCTEST(eptr)) RRETURN(MATCH_NOMATCH);
3555            }            }
3556          /* Control never gets here */          /* Control never gets here */
3557          }          }
# Line 3478  for (;;) Line 3565  for (;;)
3565              SCHECK_PARTIAL();              SCHECK_PARTIAL();
3566              break;              break;
3567              }              }
3568            if (fc != *eptr) break;            if (fc != RAWUCHARTEST(eptr)) break;
3569            eptr++;            eptr++;
3570            }            }
3571          if (possessive) continue;          if (possessive) continue;
# Line 3504  for (;;) Line 3591  for (;;)
3591        SCHECK_PARTIAL();        SCHECK_PARTIAL();
3592        RRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
3593        }        }
     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  
3594  #ifdef SUPPORT_UTF  #ifdef SUPPORT_UTF
3595        if (utf)
3596          {
3597          register pcre_uint32 ch, och;
3598    
3599          ecode++;
3600          GETCHARINC(ch, ecode);
3601          GETCHARINC(c, eptr);
3602    
3603          if (op == OP_NOT)
3604            {
3605            if (ch == c) RRETURN(MATCH_NOMATCH);
3606            }
3607          else
3608            {
3609  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
3610        if (utf && ch > 127)          if (ch > 127)
3611          och = UCD_OTHERCASE(ch);            och = UCD_OTHERCASE(ch);
3612  #else  #else
3613        if (utf && ch > 127)          if (ch > 127)
3614          och = ch;            och = ch;
3615  #endif /* SUPPORT_UCP */  #endif /* SUPPORT_UCP */
3616        else          else
3617  #endif /* SUPPORT_UTF */            och = TABLE_GET(ch, md->fcc, ch);
3618          och = TABLE_GET(ch, md->fcc, ch);          if (ch == c || och == c) RRETURN(MATCH_NOMATCH);
3619  #endif /* COMPILE_PCRE8 */          }
       if (ch == c || och == c) RRETURN(MATCH_NOMATCH);  
3620        }        }
3621      else    /* Caseful */      else
3622    #endif
3623        {        {
3624        if (*ecode++ == c) RRETURN(MATCH_NOMATCH);        register pcre_uint32 ch = ecode[1];
3625          c = *eptr++;
3626          if (ch == c || (op == OP_NOTI && TABLE_GET(ch, md->fcc, ch) == c))
3627            RRETURN(MATCH_NOMATCH);
3628          ecode += 2;
3629        }        }
3630      break;      break;
3631    
# Line 3610  for (;;) Line 3705  for (;;)
3705      /* Common code for all repeated single-byte matches. */      /* Common code for all repeated single-byte matches. */
3706    
3707      REPEATNOTCHAR:      REPEATNOTCHAR:
3708      fc = *ecode++;      GETCHARINCTEST(fc, ecode);
3709    
3710      /* The code is duplicated for the caseless and caseful cases, for speed,      /* The code is duplicated for the caseless and caseful cases, for speed,
3711      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 3716  for (;;)
3716      characters and work backwards. */      characters and work backwards. */
3717    
3718      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,
3719        max, eptr));        max, (char *)eptr));
3720    
3721      if (op >= OP_NOTSTARI)     /* Caseless */      if (op >= OP_NOTSTARI)     /* Caseless */
3722        {        {
 #ifdef COMPILE_PCRE8  
       /* fc must be < 128 if UTF is enabled. */  
       foc = md->fcc[fc];  
 #else  
3723  #ifdef SUPPORT_UTF  #ifdef SUPPORT_UTF
3724  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
3725        if (utf && fc > 127)        if (utf && fc > 127)
# Line 3640  for (;;) Line 3731  for (;;)
3731        else        else
3732  #endif /* SUPPORT_UTF */  #endif /* SUPPORT_UTF */
3733          foc = TABLE_GET(fc, md->fcc, fc);          foc = TABLE_GET(fc, md->fcc, fc);
 #endif /* COMPILE_PCRE8 */  
3734    
3735  #ifdef SUPPORT_UTF  #ifdef SUPPORT_UTF
3736        if (utf)        if (utf)
3737          {          {
3738          register unsigned int d;          register pcre_uint32 d;
3739          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
3740            {            {
3741            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
# Line 3654  for (;;) Line 3744  for (;;)
3744              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
3745              }              }
3746            GETCHARINC(d, eptr);            GETCHARINC(d, eptr);
3747            if (fc == d || foc == d) RRETURN(MATCH_NOMATCH);            if (fc == d || (unsigned int)foc == d) RRETURN(MATCH_NOMATCH);
3748            }            }
3749          }          }
3750        else        else
# Line 3680  for (;;) Line 3770  for (;;)
3770  #ifdef SUPPORT_UTF  #ifdef SUPPORT_UTF
3771          if (utf)          if (utf)
3772            {            {
3773            register unsigned int d;            register pcre_uint32 d;
3774            for (fi = min;; fi++)            for (fi = min;; fi++)
3775              {              {
3776              RMATCH(eptr, ecode, offset_top, md, eptrb, RM28);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM28);
# Line 3692  for (;;) Line 3782  for (;;)
3782                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
3783                }                }
3784              GETCHARINC(d, eptr);              GETCHARINC(d, eptr);
3785              if (fc == d || foc == d) RRETURN(MATCH_NOMATCH);              if (fc == d || (unsigned int)foc == d) RRETURN(MATCH_NOMATCH);
3786              }              }
3787            }            }
3788          else          else
# Line 3725  for (;;) Line 3815  for (;;)
3815  #ifdef SUPPORT_UTF  #ifdef SUPPORT_UTF
3816          if (utf)          if (utf)
3817            {            {
3818            register unsigned int d;            register pcre_uint32 d;
3819            for (i = min; i < max; i++)            for (i = min; i < max; i++)
3820              {              {
3821              int len = 1;              int len = 1;
# Line 3735  for (;;) Line 3825  for (;;)
3825                break;                break;
3826                }                }
3827              GETCHARLEN(d, eptr, len);              GETCHARLEN(d, eptr, len);
3828              if (fc == d || foc == d) break;              if (fc == d || (unsigned int)foc == d) break;
3829              eptr += len;              eptr += len;
3830              }              }
3831            if (possessive) continue;            if (possessive) continue;
# Line 3782  for (;;) Line 3872  for (;;)
3872  #ifdef SUPPORT_UTF  #ifdef SUPPORT_UTF
3873        if (utf)        if (utf)
3874          {          {
3875          register unsigned int d;          register pcre_uint32 d;
3876          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
3877            {            {
3878            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
# Line 3816  for (;;) Line 3906  for (;;)
3906  #ifdef SUPPORT_UTF  #ifdef SUPPORT_UTF
3907          if (utf)          if (utf)
3908            {            {
3909            register unsigned int d;            register pcre_uint32 d;
3910            for (fi = min;; fi++)            for (fi = min;; fi++)
3911              {              {
3912              RMATCH(eptr, ecode, offset_top, md, eptrb, RM32);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM32);
# Line 3860  for (;;) Line 3950  for (;;)
3950  #ifdef SUPPORT_UTF  #ifdef SUPPORT_UTF
3951          if (utf)          if (utf)
3952            {            {
3953            register unsigned int d;            register pcre_uint32 d;
3954            for (i = min; i < max; i++)            for (i = min; i < max; i++)
3955              {              {
3956              int len = 1;              int len = 1;
# Line 4136  for (;;) Line 4226  for (;;)
4226              }              }
4227            break;            break;
4228    
4229              case PT_CLIST:
4230              for (i = 1; i <= min; i++)
4231                {
4232                const pcre_uint32 *cp;
4233                if (eptr >= md->end_subject)
4234                  {
4235                  SCHECK_PARTIAL();
4236                  RRETURN(MATCH_NOMATCH);
4237                  }
4238                GETCHARINCTEST(c, eptr);
4239                cp = PRIV(ucd_caseless_sets) + prop_value;
4240                for (;;)
4241                  {
4242                  if (c < *cp)
4243                    { if (prop_fail_result) break; else { RRETURN(MATCH_NOMATCH); } }
4244                  if (c == *cp++)
4245                    { if (prop_fail_result) { RRETURN(MATCH_NOMATCH); } else break; }
4246                  }
4247                }
4248              break;
4249    
4250            /* This should not occur */            /* This should not occur */
4251    
4252            default:            default:
# Line 4155  for (;;) Line 4266  for (;;)
4266              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4267              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4268              }              }
4269            GETCHARINCTEST(c, eptr);            else
           if (UCD_CATEGORY(c) == ucp_M) RRETURN(MATCH_NOMATCH);  
           while (eptr < md->end_subject)  
4270              {              {
4271              int len = 1;              int lgb, rgb;
4272              if (!utf) c = *eptr; else { GETCHARLEN(c, eptr, len); }              GETCHARINCTEST(c, eptr);
4273              if (UCD_CATEGORY(c) != ucp_M) break;              lgb = UCD_GRAPHBREAK(c);
4274              eptr += len;             while (eptr < md->end_subject)
4275                  {
4276                  int len = 1;
4277                  if (!utf) c = *eptr; else { GETCHARLEN(c, eptr, len); }
4278                  rgb = UCD_GRAPHBREAK(c);
4279                  if ((PRIV(ucp_gbtable)[lgb] & (1 << rgb)) == 0) break;
4280                  lgb = rgb;
4281                  eptr += len;
4282                  }
4283              }              }
4284              CHECK_PARTIAL();
4285            }            }
4286          }          }
4287    
# Line 4184  for (;;) Line 4302  for (;;)
4302              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4303              }              }
4304            if (IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH);            if (IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH);
4305              if (md->partial != 0 &&
4306                  eptr + 1 >= md->end_subject &&
4307                  NLBLOCK->nltype == NLTYPE_FIXED &&
4308                  NLBLOCK->nllen == 2 &&
4309                  RAWUCHAR(eptr) == NLBLOCK->nl[0])
4310                {
4311                md->hitend = TRUE;
4312                if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
4313                }
4314            eptr++;            eptr++;
4315            ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);            ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);
4316            }            }
# Line 4220  for (;;) Line 4347  for (;;)
4347              {              {
4348              default: RRETURN(MATCH_NOMATCH);              default: RRETURN(MATCH_NOMATCH);
4349    
4350              case 0x000d:              case CHAR_CR:
4351              if (eptr < md->end_subject && *eptr == 0x0a) eptr++;              if (eptr < md->end_subject && RAWUCHAR(eptr) == CHAR_LF) eptr++;
4352              break;              break;
4353    
4354              case 0x000a:              case CHAR_LF:
4355              break;              break;
4356    
4357              case 0x000b:              case CHAR_VT:
4358              case 0x000c:              case CHAR_FF:
4359              case 0x0085:              case CHAR_NEL:
4360    #ifndef EBCDIC
4361              case 0x2028:              case 0x2028:
4362              case 0x2029:              case 0x2029:
4363    #endif  /* Not EBCDIC */
4364              if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);              if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);
4365              break;              break;
4366              }              }
# Line 4249  for (;;) Line 4378  for (;;)
4378            GETCHARINC(c, eptr);            GETCHARINC(c, eptr);
4379            switch(c)            switch(c)
4380              {              {
4381                HSPACE_CASES: RRETURN(MATCH_NOMATCH);  /* Byte and multibyte cases */
4382              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);  
4383              }              }
4384            }            }
4385          break;          break;
# Line 4285  for (;;) Line 4395  for (;;)
4395            GETCHARINC(c, eptr);            GETCHARINC(c, eptr);
4396            switch(c)            switch(c)
4397              {              {
4398                HSPACE_CASES: break;  /* Byte and multibyte cases */
4399              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;  
4400              }              }
4401            }            }
4402          break;          break;
# Line 4321  for (;;) Line 4412  for (;;)
4412            GETCHARINC(c, eptr);            GETCHARINC(c, eptr);
4413            switch(c)            switch(c)
4414              {              {
4415                VSPACE_CASES: RRETURN(MATCH_NOMATCH);
4416              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);  
4417              }              }
4418            }            }
4419          break;          break;
# Line 4345  for (;;) Line 4429  for (;;)
4429            GETCHARINC(c, eptr);            GETCHARINC(c, eptr);
4430            switch(c)            switch(c)
4431              {              {
4432                VSPACE_CASES: break;
4433              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;  
4434              }              }
4435            }            }
4436          break;          break;
# Line 4375  for (;;) Line 4452  for (;;)
4452          case OP_DIGIT:          case OP_DIGIT:
4453          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
4454            {            {
4455              pcre_uint32 cc;
4456            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
4457              {              {
4458              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4459              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4460              }              }
4461            if (*eptr >= 128 || (md->ctypes[*eptr] & ctype_digit) == 0)            cc = RAWUCHAR(eptr);
4462              if (cc >= 128 || (md->ctypes[cc] & ctype_digit) == 0)
4463              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4464            eptr++;            eptr++;
4465            /* 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 4469  for (;;)
4469          case OP_NOT_WHITESPACE:          case OP_NOT_WHITESPACE:
4470          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
4471            {            {
4472              pcre_uint32 cc;
4473            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
4474              {              {
4475              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4476              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4477              }              }
4478            if (*eptr < 128 && (md->ctypes[*eptr] & ctype_space) != 0)            cc = RAWUCHAR(eptr);
4479              if (cc < 128 && (md->ctypes[cc] & ctype_space) != 0)
4480              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4481            eptr++;            eptr++;
4482            ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);            ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);
# Line 4405  for (;;) Line 4486  for (;;)
4486          case OP_WHITESPACE:          case OP_WHITESPACE:
4487          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
4488            {            {
4489              pcre_uint32 cc;
4490            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
4491              {              {
4492              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4493              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4494              }              }
4495            if (*eptr >= 128 || (md->ctypes[*eptr] & ctype_space) == 0)            cc = RAWUCHAR(eptr);
4496              if (cc >= 128 || (md->ctypes[cc] & ctype_space) == 0)
4497              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4498            eptr++;            eptr++;
4499            /* 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 4503  for (;;)
4503          case OP_NOT_WORDCHAR:          case OP_NOT_WORDCHAR:
4504          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
4505            {            {
4506              pcre_uint32 cc;
4507            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
4508              {              {
4509              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4510              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4511              }              }
4512            if (*eptr < 128 && (md->ctypes[*eptr] & ctype_word) != 0)            cc = RAWUCHAR(eptr);
4513              if (cc < 128 && (md->ctypes[cc] & ctype_word) != 0)
4514              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4515            eptr++;            eptr++;
4516            ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);            ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);
# Line 4435  for (;;) Line 4520  for (;;)
4520          case OP_WORDCHAR:          case OP_WORDCHAR:
4521          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
4522            {            {
4523              pcre_uint32 cc;
4524            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
4525              {              {
4526              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4527              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4528              }              }
4529            if (*eptr >= 128 || (md->ctypes[*eptr] & ctype_word) == 0)            cc = RAWUCHAR(eptr);
4530              if (cc >= 128 || (md->ctypes[cc] & ctype_word) == 0)
4531              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4532            eptr++;            eptr++;
4533            /* 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 4555  for (;;)
4555              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4556              }              }
4557            if (IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH);            if (IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH);
4558              if (md->partial != 0 &&
4559                  eptr + 1 >= md->end_subject &&
4560                  NLBLOCK->nltype == NLTYPE_FIXED &&
4561                  NLBLOCK->nllen == 2 &&
4562                  *eptr == NLBLOCK->nl[0])
4563                {
4564                md->hitend = TRUE;
4565                if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
4566                }
4567            eptr++;            eptr++;
4568            }            }
4569          break;          break;
# Line 4502  for (;;) Line 4598  for (;;)
4598              {              {
4599              default: RRETURN(MATCH_NOMATCH);              default: RRETURN(MATCH_NOMATCH);
4600    
4601              case 0x000d:              case CHAR_CR:
4602              if (eptr < md->end_subject && *eptr == 0x0a) eptr++;              if (eptr < md->end_subject && *eptr == CHAR_LF) eptr++;
4603              break;              break;
4604    
4605              case 0x000a:              case CHAR_LF:
4606              break;              break;
4607    
4608              case 0x000b:              case CHAR_VT:
4609              case 0x000c:              case CHAR_FF:
4610              case 0x0085:              case CHAR_NEL:
4611  #ifdef COMPILE_PCRE16  #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
4612              case 0x2028:              case 0x2028:
4613              case 0x2029:              case 0x2029:
4614  #endif  #endif
# Line 4533  for (;;) Line 4629  for (;;)
4629            switch(*eptr++)            switch(*eptr++)
4630              {              {
4631              default: break;              default: break;
4632              case 0x09:      /* HT */              HSPACE_BYTE_CASES:
4633              case 0x20:      /* SPACE */  #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
4634              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 */  
4635  #endif  #endif
4636              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4637              }              }
# Line 4570  for (;;) Line 4649  for (;;)
4649            switch(*eptr++)            switch(*eptr++)
4650              {              {
4651              default: RRETURN(MATCH_NOMATCH);              default: RRETURN(MATCH_NOMATCH);
4652              case 0x09:      /* HT */              HSPACE_BYTE_CASES:
4653              case 0x20:      /* SPACE */  #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
4654              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 */  
4655  #endif  #endif
4656              break;              break;
4657              }              }
# Line 4606  for (;;) Line 4668  for (;;)
4668              }              }
4669            switch(*eptr++)            switch(*eptr++)
4670              {              {
4671              default: break;              VSPACE_BYTE_CASES:
4672              case 0x0a:      /* LF */  #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
4673              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 */  
4674  #endif  #endif
4675              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4676                default: break;
4677              }              }
4678            }            }
4679          break;          break;
# Line 4632  for (;;) Line 4689  for (;;)
4689            switch(*eptr++)            switch(*eptr++)
4690              {              {
4691              default: RRETURN(MATCH_NOMATCH);              default: RRETURN(MATCH_NOMATCH);
4692              case 0x0a:      /* LF */              VSPACE_BYTE_CASES:
4693              case 0x0b:      /* VT */  #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
4694              case 0x0c:      /* FF */              VSPACE_MULTIBYTE_CASES:
             case 0x0d:      /* CR */  
             case 0x85:      /* NEL */  
 #ifdef COMPILE_PCRE16  
             case 0x2028:    /* LINE SEPARATOR */  
             case 0x2029:    /* PARAGRAPH SEPARATOR */  
4695  #endif  #endif
4696              break;              break;
4697              }              }
# Line 4917  for (;;) Line 4969  for (;;)
4969              }              }
4970            /* Control never gets here */            /* Control never gets here */
4971    
4972            /* This should never occur */            case PT_CLIST:
4973              for (fi = min;; fi++)
4974                {
4975                const pcre_uint32 *cp;
4976                RMATCH(eptr, ecode, offset_top, md, eptrb, RM67);
4977                if (rrc != MATCH_NOMATCH) RRETURN(rrc);
4978                if (fi >= max) RRETURN(MATCH_NOMATCH);
4979                if (eptr >= md->end_subject)
4980                  {
4981                  SCHECK_PARTIAL();
4982                  RRETURN(MATCH_NOMATCH);
4983                  }
4984                GETCHARINCTEST(c, eptr);
4985                cp = PRIV(ucd_caseless_sets) + prop_value;
4986                for (;;)
4987                  {
4988                  if (c < *cp)
4989                    { if (prop_fail_result) break; else { RRETURN(MATCH_NOMATCH); } }
4990                  if (c == *cp++)
4991                    { if (prop_fail_result) { RRETURN(MATCH_NOMATCH); } else break; }
4992                  }
4993                }
4994              /* Control never gets here */
4995    
4996              /* This should never occur */
4997            default:            default:
4998            RRETURN(PCRE_ERROR_INTERNAL);            RRETURN(PCRE_ERROR_INTERNAL);
4999            }            }
# Line 4939  for (;;) Line 5014  for (;;)
5014              SCHECK_PARTIAL();              SCHECK_PARTIAL();
5015              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
5016              }              }
5017            GETCHARINCTEST(c, eptr);            else
           if (UCD_CATEGORY(c) == ucp_M) RRETURN(MATCH_NOMATCH);  
           while (eptr < md->end_subject)  
5018              {              {
5019              int len = 1;              int lgb, rgb;
5020              if (!utf) c = *eptr; else { GETCHARLEN(c, eptr, len); }              GETCHARINCTEST(c, eptr);
5021              if (UCD_CATEGORY(c) != ucp_M) break;              lgb = UCD_GRAPHBREAK(c);
5022              eptr += len;              while (eptr < md->end_subject)
5023                  {
5024                  int len = 1;
5025                  if (!utf) c = *eptr; else { GETCHARLEN(c, eptr, len); }
5026                  rgb = UCD_GRAPHBREAK(c);
5027                  if ((PRIV(ucp_gbtable)[lgb] & (1 << rgb)) == 0) break;
5028                  lgb = rgb;
5029                  eptr += len;
5030                  }
5031              }              }
5032              CHECK_PARTIAL();
5033            }            }
5034          }          }
5035        else        else
# Line 4971  for (;;) Line 5053  for (;;)
5053            GETCHARINC(c, eptr);            GETCHARINC(c, eptr);
5054            switch(ctype)            switch(ctype)
5055              {              {
5056              case OP_ANY:        /* This is the non-NL case */              case OP_ANY:               /* This is the non-NL case */
5057                if (md->partial != 0 &&    /* Take care with CRLF partial */
5058                    eptr >= md->end_subject &&
5059                    NLBLOCK->nltype == NLTYPE_FIXED &&
5060                    NLBLOCK->nllen == 2 &&
5061                    c == NLBLOCK->nl[0])
5062                  {
5063                  md->hitend = TRUE;
5064                  if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
5065                  }
5066                break;
5067    
5068              case OP_ALLANY:              case OP_ALLANY:
5069              case OP_ANYBYTE:              case OP_ANYBYTE:
5070              break;              break;
# Line 4980  for (;;) Line 5073  for (;;)
5073              switch(c)              switch(c)
5074                {                {
5075                default: RRETURN(MATCH_NOMATCH);                default: RRETURN(MATCH_NOMATCH);
5076                case 0x000d:                case CHAR_CR:
5077                if (eptr < md->end_subject && *eptr == 0x0a) eptr++;                if (eptr < md->end_subject && RAWUCHAR(eptr) == CHAR_LF) eptr++;
5078                break;                break;
5079                case 0x000a:  
5080                  case CHAR_LF:
5081                break;                break;
5082    
5083                case 0x000b:                case CHAR_VT:
5084                case 0x000c:                case CHAR_FF:
5085                case 0x0085:                case CHAR_NEL:
5086    #ifndef EBCDIC
5087                case 0x2028:                case 0x2028:
5088                case 0x2029:                case 0x2029:
5089    #endif  /* Not EBCDIC */
5090                if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);                if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);
5091                break;                break;
5092                }                }
# Line 4999  for (;;) Line 5095  for (;;)
5095              case OP_NOT_HSPACE:              case OP_NOT_HSPACE:
5096              switch(c)              switch(c)
5097                {                {
5098                  HSPACE_CASES: RRETURN(MATCH_NOMATCH);
5099                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);  
5100                }                }
5101              break;              break;
5102    
5103              case OP_HSPACE:              case OP_HSPACE:
5104              switch(c)              switch(c)
5105                {                {
5106                  HSPACE_CASES: break;
5107                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;  
5108                }                }
5109              break;              break;
5110    
5111              case OP_NOT_VSPACE:              case OP_NOT_VSPACE:
5112              switch(c)              switch(c)
5113                {                {
5114                  VSPACE_CASES: RRETURN(MATCH_NOMATCH);
5115                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);  
5116                }                }
5117              break;              break;
5118    
5119              case OP_VSPACE:              case OP_VSPACE:
5120              switch(c)              switch(c)
5121                {                {
5122                  VSPACE_CASES: break;
5123                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;  
5124                }                }
5125              break;              break;
5126    
# Line 5134  for (;;) Line 5178  for (;;)
5178            c = *eptr++;            c = *eptr++;
5179            switch(ctype)            switch(ctype)
5180              {              {
5181              case OP_ANY:     /* This is the non-NL case */              case OP_ANY:               /* This is the non-NL case */
5182                if (md->partial != 0 &&    /* Take care with CRLF partial */
5183                    eptr >= md->end_subject &&
5184                    NLBLOCK->nltype == NLTYPE_FIXED &&
5185                    NLBLOCK->nllen == 2 &&
5186                    c == NLBLOCK->nl[0])
5187                  {
5188                  md->hitend = TRUE;
5189                  if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
5190                  }
5191                break;
5192    
5193              case OP_ALLANY:              case OP_ALLANY:
5194              case OP_ANYBYTE:              case OP_ANYBYTE:
5195              break;              break;
# Line 5143  for (;;) Line 5198  for (;;)
5198              switch(c)              switch(c)
5199                {                {
5200                default: RRETURN(MATCH_NOMATCH);                default: RRETURN(MATCH_NOMATCH);
5201                case 0x000d:                case CHAR_CR:
5202                if (eptr < md->end_subject && *eptr == 0x0a) eptr++;                if (eptr < md->end_subject && *eptr == CHAR_LF) eptr++;
5203                break;                break;
5204    
5205                case 0x000a:                case CHAR_LF:
5206                break;                break;
5207    
5208                case 0x000b:                case CHAR_VT:
5209                case 0x000c:                case CHAR_FF:
5210                case 0x0085:                case CHAR_NEL:
5211  #ifdef COMPILE_PCRE16  #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
5212                case 0x2028:                case 0x2028:
5213                case 0x2029:                case 0x2029:
5214  #endif  #endif
# Line 5166  for (;;) Line 5221  for (;;)
5221              switch(c)              switch(c)
5222                {                {
5223                default: break;                default: break;
5224                case 0x09:      /* HT */                HSPACE_BYTE_CASES:
5225                case 0x20:      /* SPACE */  #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
5226                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 */  
5227  #endif  #endif
5228                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
5229                }                }
# Line 5195  for (;;) Line 5233  for (;;)
5233              switch(c)              switch(c)
5234                {                {
5235                default: RRETURN(MATCH_NOMATCH);                default: RRETURN(MATCH_NOMATCH);
5236                case 0x09:      /* HT */                HSPACE_BYTE_CASES:
5237                case 0x20:      /* SPACE */  #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
5238                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 */  
5239  #endif  #endif
5240                break;                break;
5241                }                }
# Line 5224  for (;;) Line 5245  for (;;)
5245              switch(c)              switch(c)
5246                {                {
5247                default: break;                default: break;
5248                case 0x0a:      /* LF */                VSPACE_BYTE_CASES:
5249                case 0x0b:      /* VT */  #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
5250                case 0x0c:      /* FF */                VSPACE_MULTIBYTE_CASES:
               case 0x0d:      /* CR */  
               case 0x85:      /* NEL */  
 #ifdef COMPILE_PCRE16  
               case 0x2028:    /* LINE SEPARATOR */  
               case 0x2029:    /* PARAGRAPH SEPARATOR */  
5251  #endif  #endif
5252                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
5253                }                }
# Line 5241  for (;;) Line 5257  for (;;)
5257              switch(c)              switch(c)
5258                {                {
5259                default: RRETURN(MATCH_NOMATCH);                default: RRETURN(MATCH_NOMATCH);
5260                case 0x0a:      /* LF */                VSPACE_BYTE_CASES:
5261                case 0x0b:      /* VT */  #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
5262                case 0x0c:      /* FF */                VSPACE_MULTIBYTE_CASES:
               case 0x0d:      /* CR */  
               case 0x85:      /* NEL */  
 #ifdef COMPILE_PCRE16  
               case 0x2028:    /* LINE SEPARATOR */  
               case 0x2029:    /* PARAGRAPH SEPARATOR */  
5263  #endif  #endif
5264                break;                break;
5265                }                }
# Line 5452  for (;;) Line 5463  for (;;)
5463              }              }
5464            break;            break;
5465    
5466              case PT_CLIST:
5467              for (i = min; i < max; i++)
5468                {
5469                const pcre_uint32 *cp;
5470                int len = 1;
5471                if (eptr >= md->end_subject)
5472                  {
5473                  SCHECK_PARTIAL();
5474                  break;
5475                  }
5476                GETCHARLENTEST(c, eptr, len);
5477                cp = PRIV(ucd_caseless_sets) + prop_value;
5478                for (;;)
5479                  {
5480                  if (c < *cp)
5481                    { if (prop_fail_result) break; else goto GOT_MAX; }
5482                  if (c == *cp++)
5483                    { if (prop_fail_result) goto GOT_MAX; else break; }
5484                  }
5485                eptr += len;
5486                }
5487              GOT_MAX:
5488              break;
5489    
5490            default:            default:
5491            RRETURN(PCRE_ERROR_INTERNAL);            RRETURN(PCRE_ERROR_INTERNAL);
5492            }            }
# Line 5475  for (;;) Line 5510  for (;;)
5510          {          {
5511          for (i = min; i < max; i++)          for (i = min; i < max; i++)
5512            {            {
           int len = 1;  
5513            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
5514              {              {
5515              SCHECK_PARTIAL();              SCHECK_PARTIAL();
5516              break;              break;
5517              }              }
5518            if (!utf) c = *eptr; else { GETCHARLEN(c, eptr, len); }            else
           if (UCD_CATEGORY(c) == ucp_M) break;  
           eptr += len;  
           while (eptr < md->end_subject)  
5519              {              {
5520              len = 1;              int lgb, rgb;
5521              if (!utf) c = *eptr; else { GETCHARLEN(c, eptr, len); }              GETCHARINCTEST(c, eptr);
5522              if (UCD_CATEGORY(c) != ucp_M) break;              lgb = UCD_GRAPHBREAK(c);
5523              eptr += len;              while (eptr < md->end_subject)
5524                  {
5525                  int len = 1;
5526                  if (!utf) c = *eptr; else { GETCHARLEN(c, eptr, len); }
5527                  rgb = UCD_GRAPHBREAK(c);
5528                  if ((PRIV(ucp_gbtable)[lgb] & (1 << rgb)) == 0) break;
5529                  lgb = rgb;
5530                  eptr += len;
5531                  }
5532              }              }
5533              CHECK_PARTIAL();
5534            }            }
5535    
5536          /* eptr is now past the end of the maximum run */          /* eptr is now past the end of the maximum run */
# Line 5534  for (;;) Line 5574  for (;;)
5574                  break;                  break;
5575                  }                  }
5576                if (IS_NEWLINE(eptr)) break;                if (IS_NEWLINE(eptr)) break;
5577                  if (md->partial != 0 &&    /* Take care with CRLF partial */
5578                      eptr + 1 >= md->end_subject &&
5579                      NLBLOCK->nltype == NLTYPE_FIXED &&
5580                      NLBLOCK->nllen == 2 &&
5581                      RAWUCHAR(eptr) == NLBLOCK->nl[0])
5582                    {
5583                    md->hitend = TRUE;
5584                    if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
5585                    }
5586                eptr++;                eptr++;
5587                ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);                ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);
5588                }                }
# Line 5551  for (;;) Line 5600  for (;;)
5600                  break;                  break;
5601                  }                  }
5602                if (IS_NEWLINE(eptr)) break;                if (IS_NEWLINE(eptr)) break;
5603                  if (md->partial != 0 &&    /* Take care with CRLF partial */
5604                      eptr + 1 >= md->end_subject &&
5605                      NLBLOCK->nltype == NLTYPE_FIXED &&
5606                      NLBLOCK->nllen == 2 &&
5607                      RAWUCHAR(eptr) == NLBLOCK->nl[0])
5608                    {
5609                    md->hitend = TRUE;
5610                    if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
5611                    }
5612                eptr++;                eptr++;
5613                ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);                ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);
5614                }                }
# Line 5600  for (;;) Line 5658  for (;;)
5658                break;                break;
5659                }                }
5660              GETCHARLEN(c, eptr, len);              GETCHARLEN(c, eptr, len);
5661              if (c == 0x000d)              if (c == CHAR_CR)
5662                {                {
5663                if (++eptr >= md->end_subject) break;                if (++eptr >= md->end_subject) break;
5664                if (*eptr == 0x000a) eptr++;                if (RAWUCHAR(eptr) == CHAR_LF) eptr++;
5665                }                }
5666              else              else
5667                {                {
5668                if (c != 0x000a &&                if (c != CHAR_LF &&
5669                    (md->bsr_anycrlf ||                    (md->bsr_anycrlf ||
5670                     (c != 0x000b && c != 0x000c &&                     (c != CHAR_VT && c != CHAR_FF && c != CHAR_NEL
5671                      c != 0x0085 && c != 0x2028 && c != 0x2029)))  #ifndef EBCDIC
5672                        && c != 0x2028 && c != 0x2029
5673    #endif  /* Not EBCDIC */
5674                        )))
5675                  break;                  break;
5676                eptr += len;                eptr += len;
5677                }                }
# Line 5631  for (;;) Line 5692  for (;;)
5692              GETCHARLEN(c, eptr, len);              GETCHARLEN(c, eptr, len);
5693              switch(c)              switch(c)
5694                {                {
5695                  HSPACE_CASES: gotspace = TRUE; break;
5696                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;  
5697                }                }
5698              if (gotspace == (ctype == OP_NOT_HSPACE)) break;              if (gotspace == (ctype == OP_NOT_HSPACE)) break;
5699              eptr += len;              eptr += len;
# Line 5673  for (;;) Line 5714  for (;;)
5714              GETCHARLEN(c, eptr, len);              GETCHARLEN(c, eptr, len);
5715              switch(c)              switch(c)
5716                {                {
5717                  VSPACE_CASES: gotspace = TRUE; break;
5718                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;  
5719                }                }
5720              if (gotspace == (ctype == OP_NOT_VSPACE)) break;              if (gotspace == (ctype == OP_NOT_VSPACE)) break;
5721              eptr += len;              eptr += len;
# Line 5796  for (;;) Line 5829  for (;;)
5829            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
5830            if (eptr-- == pp) break;        /* Stop if tried at original pos */            if (eptr-- == pp) break;        /* Stop if tried at original pos */
5831            BACKCHAR(eptr);            BACKCHAR(eptr);
5832            if (ctype == OP_ANYNL && eptr > pp  && *eptr == '\n' &&            if (ctype == OP_ANYNL && eptr > pp  && RAWUCHAR(eptr) == CHAR_NL &&
5833                eptr[-1] == '\r') eptr--;                RAWUCHAR(eptr - 1) == CHAR_CR) eptr--;
5834            }            }
5835          }          }
5836        else        else
# Line 5815  for (;;) Line 5848  for (;;)
5848                break;                break;
5849                }                }
5850              if (IS_NEWLINE(eptr)) break;              if (IS_NEWLINE(eptr)) break;
5851                if (md->partial != 0 &&    /* Take care with CRLF partial */
5852                    eptr + 1 >= md->end_subject &&
5853                    NLBLOCK->nltype == NLTYPE_FIXED &&
5854                    NLBLOCK->nllen == 2 &&
5855                    *eptr == NLBLOCK->nl[0])
5856                  {
5857                  md->hitend = TRUE;
5858                  if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
5859                  }
5860              eptr++;              eptr++;
5861              }              }
5862            break;            break;
# Line 5839  for (;;) Line 5881  for (;;)
5881                break;                break;
5882                }                }
5883              c = *eptr;              c = *eptr;
5884              if (c == 0x000d)              if (c == CHAR_CR)
5885                {                {
5886                if (++eptr >= md->end_subject) break;                if (++eptr >= md->end_subject) break;
5887                if (*eptr == 0x000a) eptr++;                if (*eptr == CHAR_LF) eptr++;
5888                }                }
5889              else              else
5890                {                {
5891                if (c != 0x000a && (md->bsr_anycrlf ||                if (c != CHAR_LF && (md->bsr_anycrlf ||
5892                  (c != 0x000b && c != 0x000c && c != 0x0085                   (c != CHAR_VT && c != CHAR_FF && c != CHAR_NEL
5893  #ifdef COMPILE_PCRE16  #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
5894                  && c != 0x2028 && c != 0x2029                   && c != 0x2028 && c != 0x2029
5895  #endif  #endif
5896                  ))) break;                   ))) break;
5897                eptr++;                eptr++;
5898                }                }
5899              }              }
# Line 5865  for (;;) Line 5907  for (;;)
5907                SCHECK_PARTIAL();                SCHECK_PARTIAL();
5908                break;                break;
5909                }                }
5910              c = *eptr;              switch(*eptr)
5911              if (c == 0x09 || c == 0x20 || c == 0xa0                {
5912  #ifdef COMPILE_PCRE16                default: eptr++; break;
5913                || c == 0x1680 || c == 0x180e || (c >= 0x2000 && c <= 0x200A)                HSPACE_BYTE_CASES:
5914                || c == 0x202f || c == 0x205f || c == 0x3000  #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
5915                  HSPACE_MULTIBYTE_CASES:
5916  #endif  #endif
5917                ) break;                goto ENDLOOP00;
5918              eptr++;                }
5919              }              }
5920              ENDLOOP00:
5921            break;            break;
5922    
5923            case OP_HSPACE:            case OP_HSPACE:
# Line 5884  for (;;) Line 5928  for (;;)
5928                SCHECK_PARTIAL();                SCHECK_PARTIAL();
5929                break;                break;
5930                }                }
5931              c = *eptr;              switch(*eptr)
5932              if (c != 0x09 && c != 0x20 && c != 0xa0                {
5933  #ifdef COMPILE_PCRE16                default: goto ENDLOOP01;
5934                && c != 0x1680 && c != 0x180e && (c < 0x2000 || c > 0x200A)                HSPACE_BYTE_CASES:
5935                && c != 0x202f && c != 0x205f && c != 0x3000  #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
5936                  HSPACE_MULTIBYTE_CASES:
5937  #endif  #endif
5938                ) break;                eptr++; break;
5939              eptr++;                }
5940              }              }
5941              ENDLOOP01:
5942            break;            break;
5943    
5944            case OP_NOT_VSPACE:            case OP_NOT_VSPACE:
# Line 5903  for (;;) Line 5949  for (;;)
5949                SCHECK_PARTIAL();                SCHECK_PARTIAL();
5950                break;                break;
5951                }                }
5952              c = *eptr;              switch(*eptr)
5953              if (c == 0x0a || c == 0x0b || c == 0x0c || c == 0x0d || c == 0x85                {
5954  #ifdef COMPILE_PCRE16                default: eptr++; break;
5955                || c == 0x2028 || c == 0x2029                VSPACE_BYTE_CASES:
5956    #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
5957                  VSPACE_MULTIBYTE_CASES:
5958  #endif  #endif
5959                ) break;                goto ENDLOOP02;
5960              eptr++;                }
5961              }              }
5962              ENDLOOP02:
5963            break;            break;
5964    
5965            case OP_VSPACE:            case OP_VSPACE:
# Line 5921  for (;;) Line 5970  for (;;)
5970                SCHECK_PARTIAL();                SCHECK_PARTIAL();
5971                break;                break;
5972                }                }
5973              c = *eptr;              switch(*eptr)
5974              if (c != 0x0a && c != 0x0b && c != 0x0c && c != 0x0d && c != 0x85                {
5975  #ifdef COMPILE_PCRE16                default: goto ENDLOOP03;
5976                && c != 0x2028 && c != 0x2029                VSPACE_BYTE_CASES:
5977    #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
5978                  VSPACE_MULTIBYTE_CASES:
5979  #endif  #endif
5980                ) break;                eptr++; break;
5981              eptr++;                }
5982              }              }
5983              ENDLOOP03:
5984            break;            break;
5985    
5986            case OP_NOT_DIGIT:            case OP_NOT_DIGIT:
# Line 6025  for (;;) Line 6077  for (;;)
6077            RMATCH(eptr, ecode, offset_top, md, eptrb, RM47);            RMATCH(eptr, ecode, offset_top, md, eptrb, RM47);
6078            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
6079            eptr--;            eptr--;
6080            if (ctype == OP_ANYNL && eptr > pp  && *eptr == '\n' &&            if (ctype == OP_ANYNL && eptr > pp  && *eptr == CHAR_LF &&
6081                eptr[-1] == '\r') eptr--;                eptr[-1] == CHAR_CR) eptr--;
6082            }            }
6083          }          }
6084    
# Line 6076  switch (frame->Xwhere) Line 6128  switch (frame->Xwhere)
6128    LBL(32) LBL(34) LBL(42) LBL(46)    LBL(32) LBL(34) LBL(42) LBL(46)
6129  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
6130    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)
6131    LBL(59) LBL(60) LBL(61) LBL(62)    LBL(59) LBL(60) LBL(61) LBL(62) LBL(67)
6132  #endif  /* SUPPORT_UCP */  #endif  /* SUPPORT_UCP */
6133  #endif  /* SUPPORT_UTF */  #endif  /* SUPPORT_UTF */
6134    default:    default:
6135    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);  
   
6136    return PCRE_ERROR_INTERNAL;    return PCRE_ERROR_INTERNAL;
6137    }    }
6138  #undef LBL  #undef LBL
# Line 6145  Undefine all the macros that were define Line 6194  Undefine all the macros that were define
6194  ***************************************************************************/  ***************************************************************************/
6195    
6196    
6197    #ifdef NO_RECURSE
6198    /*************************************************
6199    *          Release allocated heap frames         *
6200    *************************************************/
6201    
6202    /* This function releases all the allocated frames. The base frame is on the
6203    machine stack, and so must not be freed.
6204    
6205    Argument: the address of the base frame
6206    Returns:  nothing
6207    */
6208    
6209    static void
6210    release_match_heapframes (heapframe *frame_base)
6211    {
6212    heapframe *nextframe = frame_base->Xnextframe;
6213    while (nextframe != NULL)
6214      {
6215      heapframe *oldframe = nextframe;
6216      nextframe = nextframe->Xnextframe;
6217      (PUBL(stack_free))(oldframe);
6218      }
6219    }
6220    #endif
6221    
6222    
6223  /*************************************************  /*************************************************
6224  *         Execute a Regular Expression           *  *         Execute a Regular Expression           *
# Line 6170  Returns:          > 0 => success; value Line 6244  Returns:          > 0 => success; value
6244                   < -1 => some kind of unexpected problem                   < -1 => some kind of unexpected problem
6245  */  */
6246    
6247  #ifdef COMPILE_PCRE8  #if defined COMPILE_PCRE8
6248  PCRE_EXP_DEFN int PCRE_CALL_CONVENTION  PCRE_EXP_DEFN int PCRE_CALL_CONVENTION
6249  pcre_exec(const pcre *argument_re, const pcre_extra *extra_data,  pcre_exec(const pcre *argument_re, const pcre_extra *extra_data,
6250    PCRE_SPTR subject, int length, int start_offset, int options, int *offsets,    PCRE_SPTR subject, int length, int start_offset, int options, int *offsets,
6251    int offsetcount)    int offsetcount)
6252  #else  #elif defined COMPILE_PCRE16
6253  PCRE_EXP_DEFN int PCRE_CALL_CONVENTION  PCRE_EXP_DEFN int PCRE_CALL_CONVENTION
6254  pcre16_exec(const pcre16 *argument_re, const pcre16_extra *extra_data,  pcre16_exec(const pcre16 *argument_re, const pcre16_extra *extra_data,
6255    PCRE_SPTR16 subject, int length, int start_offset, int options, int *offsets,    PCRE_SPTR16 subject, int length, int start_offset, int options, int *offsets,
6256    int offsetcount)    int offsetcount)
6257    #elif defined COMPILE_PCRE32
6258    PCRE_EXP_DEFN int PCRE_CALL_CONVENTION
6259    pcre32_exec(const pcre32 *argument_re, const pcre32_extra *extra_data,
6260      PCRE_SPTR32 subject, int length, int start_offset, int options, int *offsets,
6261      int offsetcount)
6262  #endif  #endif
6263  {  {
6264  int rc, ocount, arg_offset_max;  int rc, ocount, arg_offset_max;
# Line 6202  const pcre_uint8 *start_bits = NULL; Line 6281  const pcre_uint8 *start_bits = NULL;
6281  PCRE_PUCHAR start_match = (PCRE_PUCHAR)subject + start_offset;  PCRE_PUCHAR start_match = (PCRE_PUCHAR)subject + start_offset;
6282  PCRE_PUCHAR end_subject;  PCRE_PUCHAR end_subject;
6283  PCRE_PUCHAR start_partial = NULL;  PCRE_PUCHAR start_partial = NULL;
6284    PCRE_PUCHAR match_partial;
6285  PCRE_PUCHAR req_char_ptr = start_match - 1;  PCRE_PUCHAR req_char_ptr = start_match - 1;
6286    
6287  const pcre_study_data *study;  const pcre_study_data *study;
6288  const REAL_PCRE *re = (const REAL_PCRE *)argument_re;  const REAL_PCRE *re = (const REAL_PCRE *)argument_re;
6289    
6290  /* Check for the special magic call that measures the size of the stack used  #ifdef NO_RECURSE
6291  per recursive call of match(). */  heapframe frame_zero;
6292    frame_zero.Xprevframe = NULL;            /* Marks the top level */
6293    frame_zero.Xnextframe = NULL;            /* None are allocated yet */
6294    md->match_frames_base = &frame_zero;
6295    #endif
6296    
6297    /* Check for the special magic call that measures the size of the stack used
6298    per recursive call of match(). Without the funny casting for sizeof, a Windows
6299    compiler gave this error: "unary minus operator applied to unsigned type,
6300    result still unsigned". Hopefully the cast fixes that. */
6301    
6302  if (re == NULL && extra_data == NULL && subject == NULL && length == -1)  if (re == NULL && extra_data == NULL && subject == NULL && length == -999 &&
6303        start_offset == -999)
6304  #ifdef NO_RECURSE  #ifdef NO_RECURSE
6305    return -sizeof(heapframe);    return -((int)sizeof(heapframe));
6306  #else  #else
6307    return match(NULL, NULL, NULL, 0, NULL, NULL, 0);    return match(NULL, NULL, NULL, 0, NULL, NULL, 0);
6308  #endif  #endif
6309    
6310  /* Plausibility checks */  /* Plausibility checks */
6311    
6312  if ((options & ~PUBLIC_EXEC_OPTIONS) != 0) return PCRE_ERROR_BADOPTION;  if ((options & ~PUBLIC_EXEC_OPTIONS) != 0) return PCRE_ERROR_BADOPTION;
6313  if (re == NULL || subject == NULL || (offsets == NULL && offsetcount > 0))  if (re == NULL || subject == NULL || (offsets == NULL && offsetcount > 0))
6314    return PCRE_ERROR_NULL;    return PCRE_ERROR_NULL;
6315  if (offsetcount < 0) return PCRE_ERROR_BADCOUNT;  if (offsetcount < 0) return PCRE_ERROR_BADCOUNT;
6316    if (length < 0) return PCRE_ERROR_BADLENGTH;
6317  if (start_offset < 0 || start_offset > length) return PCRE_ERROR_BADOFFSET;  if (start_offset < 0 || start_offset > length) return PCRE_ERROR_BADOFFSET;
6318    
6319  /* 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 6351  if (utf && (options & PCRE_NO_UTF8_CHECK
6351        offsets[0] = erroroffset;        offsets[0] = erroroffset;
6352        offsets[1] = errorcode;        offsets[1] = errorcode;
6353        }        }
6354  #ifdef COMPILE_PCRE16  #if defined COMPILE_PCRE8
     return (errorcode <= PCRE_UTF16_ERR1 && md->partial > 1)?  
       PCRE_ERROR_SHORTUTF16 : PCRE_ERROR_BADUTF16;  
 #else  
6355      return (errorcode <= PCRE_UTF8_ERR5 && md->partial > 1)?      return (errorcode <= PCRE_UTF8_ERR5 && md->partial > 1)?
6356        PCRE_ERROR_SHORTUTF8 : PCRE_ERROR_BADUTF8;        PCRE_ERROR_SHORTUTF8 : PCRE_ERROR_BADUTF8;
6357    #elif defined COMPILE_PCRE16
6358        return (errorcode <= PCRE_UTF16_ERR1 && md->partial > 1)?
6359          PCRE_ERROR_SHORTUTF16 : PCRE_ERROR_BADUTF16;
6360    #elif defined COMPILE_PCRE32
6361        return PCRE_ERROR_BADUTF32;
6362  #endif  #endif
6363      }      }
6364    #if defined COMPILE_PCRE8 || defined COMPILE_PCRE16
6365    /* 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. */
6366    if (start_offset > 0 && start_offset < length &&    if (start_offset > 0 && start_offset < length &&
6367        NOT_FIRSTCHAR(((PCRE_PUCHAR)subject)[start_offset]))        NOT_FIRSTCHAR(((PCRE_PUCHAR)subject)[start_offset]))
6368      return PCRE_ERROR_BADUTF8_OFFSET;      return PCRE_ERROR_BADUTF8_OFFSET;
6369    #endif
6370    }    }
6371  #endif  #endif
6372    
6373  /* If the pattern was successfully studied with JIT support, run the JIT  /* If the pattern was successfully studied with JIT support, run the JIT
6374  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
6375  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
6376  an unsupported flag is set. In particular, JIT does not support partial  an unsupported flag is set. */
 matching. */  
6377    
6378  #ifdef SUPPORT_JIT  #ifdef SUPPORT_JIT
6379  if (extra_data != NULL  if (extra_data != NULL
6380      && (extra_data->flags & PCRE_EXTRA_EXECUTABLE_JIT) != 0      && (extra_data->flags & (PCRE_EXTRA_EXECUTABLE_JIT |
6381                                 PCRE_EXTRA_TABLES)) == PCRE_EXTRA_EXECUTABLE_JIT
6382      && extra_data->executable_jit != NULL      && extra_data->executable_jit != NULL
6383      && (extra_data->flags & PCRE_EXTRA_TABLES) == 0      && (options & ~PUBLIC_JIT_EXEC_OPTIONS) == 0)
6384      && (options & ~(PCRE_NO_UTF8_CHECK | PCRE_NOTBOL | PCRE_NOTEOL |    {
6385                      PCRE_NOTEMPTY | PCRE_NOTEMPTY_ATSTART)) == 0)    rc = PRIV(jit_exec)(extra_data, (const pcre_uchar *)subject, length,
6386    return PRIV(jit_exec)(re, extra_data->executable_jit,         start_offset, options, offsets, offsetcount);
6387      (const pcre_uchar *)subject, length, start_offset, options,  
6388      ((extra_data->flags & PCRE_EXTRA_MATCH_LIMIT) == 0)    /* PCRE_ERROR_NULL means that the selected normal or partial matching
6389      ? MATCH_LIMIT : extra_data->match_limit, offsets, offsetcount);    mode is not compiled. In this case we simply fallback to interpreter. */
6390    
6391      if (rc != PCRE_ERROR_JIT_BADOPTION) return rc;
6392      }
6393  #endif  #endif
6394    
6395  /* 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 6463  if (re->top_backref > 0 && re->top_backr Line 6560  if (re->top_backref > 0 && re->top_backr
6560    DPRINTF(("Got memory to hold back references\n"));    DPRINTF(("Got memory to hold back references\n"));
6561    }    }
6562  else md->offset_vector = offsets;  else md->offset_vector = offsets;
   
6563  md->offset_end = ocount;  md->offset_end = ocount;
6564  md->offset_max = (2*ocount)/3;  md->offset_max = (2*ocount)/3;
6565  md->offset_overflow = FALSE;  md->capture_last = 0;
 md->capture_last = -1;  
6566    
6567  /* Reset the working variable associated with each extraction. These should  /* Reset the working variable associated with each extraction. These should
6568  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 6590  if (!anchored)
6590    if ((re->flags & PCRE_FIRSTSET) != 0)    if ((re->flags & PCRE_FIRSTSET) != 0)
6591      {      {
6592      has_first_char = TRUE;      has_first_char = TRUE;
6593      first_char = first_char2 = re->first_char;      first_char = first_char2 = (pcre_uchar)(re->first_char);
6594      if ((re->flags & PCRE_FCH_CASELESS) != 0)      if ((re->flags & PCRE_FCH_CASELESS) != 0)
6595        {        {
6596        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 6612  character" set. */
6612  if ((re->flags & PCRE_REQCHSET) != 0)  if ((re->flags & PCRE_REQCHSET) != 0)
6613    {    {
6614    has_req_char = TRUE;    has_req_char = TRUE;
6615    req_char = req_char2 = re->req_char;    req_char = req_char2 = (pcre_uchar)(re->req_char);
6616    if ((re->flags & PCRE_RCH_CASELESS) != 0)    if ((re->flags & PCRE_RCH_CASELESS) != 0)
6617      {      {
6618      req_char2 = TABLE_GET(req_char, md->fcc, req_char);      req_char2 = TABLE_GET(req_char, md->fcc, req_char);
# Line 6575  for(;;) Line 6670  for(;;)
6670    
6671      if (has_first_char)      if (has_first_char)
6672        {        {
6673          pcre_uchar smc;
6674    
6675        if (first_char != first_char2)        if (first_char != first_char2)
6676          while (start_match < end_subject &&          while (start_match < end_subject &&
6677              *start_match != first_char && *start_match != first_char2)            (smc = RAWUCHARTEST(start_match)) != first_char && smc != first_char2)
6678            start_match++;            start_match++;
6679        else        else
6680          while (start_match < end_subject && *start_match != first_char)          while (start_match < end_subject && RAWUCHARTEST(start_match) != first_char)
6681            start_match++;            start_match++;
6682        }        }
6683    
# Line 6612  for(;;) Line 6709  for(;;)
6709          if (start_match[-1] == CHAR_CR &&          if (start_match[-1] == CHAR_CR &&
6710               (md->nltype == NLTYPE_ANY || md->nltype == NLTYPE_ANYCRLF) &&               (md->nltype == NLTYPE_ANY || md->nltype == NLTYPE_ANYCRLF) &&
6711               start_match < end_subject &&               start_match < end_subject &&
6712               *start_match == CHAR_NL)               RAWUCHARTEST(start_match) == CHAR_NL)
6713            start_match++;            start_match++;
6714          }          }
6715        }        }
# Line 6623  for(;;) Line 6720  for(;;)
6720        {        {
6721        while (start_match < end_subject)        while (start_match < end_subject)
6722          {          {
6723          register unsigned int c = *start_match;          register pcre_uint32 c = RAWUCHARTEST(start_match);
6724  #ifndef COMPILE_PCRE8  #ifndef COMPILE_PCRE8
6725          if (c > 255) c = 255;          if (c > 255) c = 255;
6726  #endif  #endif
# Line 6691  for(;;) Line 6788  for(;;)
6788            {            {
6789            while (p < end_subject)            while (p < end_subject)
6790              {              {
6791              register int pp = *p++;              register pcre_uint32 pp = RAWUCHARINCTEST(p);
6792              if (pp == req_char || pp == req_char2) { p--; break; }              if (pp == req_char || pp == req_char2) { p--; break; }
6793              }              }
6794            }            }
# Line 6699  for(;;) Line 6796  for(;;)
6796            {            {
6797            while (p < end_subject)            while (p < end_subject)
6798              {              {
6799              if (*p++ == req_char) { p--; break; }              if (RAWUCHARINCTEST(p) == req_char) { p--; break; }
6800              }              }
6801            }            }
6802    
# Line 6736  for(;;) Line 6833  for(;;)
6833    md->match_function_type = 0;    md->match_function_type = 0;
6834    md->end_offset_top = 0;    md->end_offset_top = 0;
6835    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);
6836    if (md->hitend && start_partial == NULL) start_partial = md->start_used_ptr;    if (md->hitend && start_partial == NULL)
6837        {
6838        start_partial = md->start_used_ptr;
6839        match_partial = start_match;
6840        }
6841    
6842    switch(rc)    switch(rc)
6843      {      {
# Line 6862  if (rc == MATCH_MATCH || rc == MATCH_ACC Line 6963  if (rc == MATCH_MATCH || rc == MATCH_ACC
6963          (arg_offset_max - 2) * sizeof(int));          (arg_offset_max - 2) * sizeof(int));
6964        DPRINTF(("Copied offsets from temporary memory\n"));        DPRINTF(("Copied offsets from temporary memory\n"));
6965        }        }
6966      if (md->end_offset_top > arg_offset_max) md->offset_overflow = TRUE;      if (md->end_offset_top > arg_offset_max) md->capture_last |= OVFLBIT;
6967      DPRINTF(("Freeing temporary memory\n"));      DPRINTF(("Freeing temporary memory\n"));
6968      (PUBL(free))(md->offset_vector);      (PUBL(free))(md->offset_vector);
6969      }      }
# Line 6870  if (rc == MATCH_MATCH || rc == MATCH_ACC Line 6971  if (rc == MATCH_MATCH || rc == MATCH_ACC
6971    /* 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
6972    too many to fit into the vector. */    too many to fit into the vector. */
6973    
6974    rc = (md->offset_overflow && md->end_offset_top >= arg_offset_max)?    rc = ((md->capture_last & OVFLBIT) != 0 &&
6975             md->end_offset_top >= arg_offset_max)?
6976      0 : md->end_offset_top/2;      0 : md->end_offset_top/2;
6977    
6978    /* 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 6988  if (rc == MATCH_MATCH || rc == MATCH_ACC
6988      {      {
6989      register int *iptr, *iend;      register int *iptr, *iend;
6990      int resetcount = 2 + re->top_bracket * 2;      int resetcount = 2 + re->top_bracket * 2;
6991      if (resetcount > offsetcount) resetcount = ocount;      if (resetcount > offsetcount) resetcount = offsetcount;
6992      iptr = offsets + md->end_offset_top;      iptr = offsets + md->end_offset_top;
6993      iend = offsets + resetcount;      iend = offsets + resetcount;
6994      while (iptr < iend) *iptr++ = -1;      while (iptr < iend) *iptr++ = -1;
# Line 6907  if (rc == MATCH_MATCH || rc == MATCH_ACC Line 7009  if (rc == MATCH_MATCH || rc == MATCH_ACC
7009    if (extra_data != NULL && (extra_data->flags & PCRE_EXTRA_MARK) != 0)    if (extra_data != NULL && (extra_data->flags & PCRE_EXTRA_MARK) != 0)
7010      *(extra_data->mark) = (pcre_uchar *)md->mark;      *(extra_data->mark) = (pcre_uchar *)md->mark;
7011    DPRINTF((">>>> returning %d\n", rc));    DPRINTF((">>>> returning %d\n", rc));
7012    #ifdef NO_RECURSE
7013      release_match_heapframes(&frame_zero);
7014    #endif
7015    return rc;    return rc;
7016    }    }
7017    
# Line 6924  if (using_temporary_offsets) Line 7029  if (using_temporary_offsets)
7029  if (rc != MATCH_NOMATCH && rc != PCRE_ERROR_PARTIAL)  if (rc != MATCH_NOMATCH && rc != PCRE_ERROR_PARTIAL)
7030    {    {
7031    DPRINTF((">>>> error: returning %d\n", rc));    DPRINTF((">>>> error: returning %d\n", rc));
7032    #ifdef NO_RECURSE
7033      release_match_heapframes(&frame_zero);
7034    #endif
7035    return rc;    return rc;
7036    }    }
7037    
# Line 6937  if (start_partial != NULL) Line 7045  if (start_partial != NULL)
7045      {      {
7046      offsets[0] = (int)(start_partial - (PCRE_PUCHAR)subject);      offsets[0] = (int)(start_partial - (PCRE_PUCHAR)subject);
7047      offsets[1] = (int)(end_subject - (PCRE_PUCHAR)subject);      offsets[1] = (int)(end_subject - (PCRE_PUCHAR)subject);
7048        if (offsetcount > 2)
7049          offsets[2] = (int)(match_partial - (PCRE_PUCHAR)subject);
7050      }      }
7051    rc = PCRE_ERROR_PARTIAL;    rc = PCRE_ERROR_PARTIAL;
7052    }    }
# Line 6953  else Line 7063  else
7063    
7064  if (extra_data != NULL && (extra_data->flags & PCRE_EXTRA_MARK) != 0)  if (extra_data != NULL && (extra_data->flags & PCRE_EXTRA_MARK) != 0)
7065    *(extra_data->mark) = (pcre_uchar *)md->nomatch_mark;    *(extra_data->mark) = (pcre_uchar *)md->nomatch_mark;
7066    #ifdef NO_RECURSE
7067      release_match_heapframes(&frame_zero);
7068    #endif
7069  return rc;  return rc;
7070  }  }
7071    

Legend:
Removed from v.895  
changed lines
  Added in v.1257

  ViewVC Help
Powered by ViewVC 1.1.5