/[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 788 by ph10, Tue Dec 6 15:38:01 2011 UTC revision 1296 by ph10, Tue Mar 19 16:29:12 2013 UTC
# Line 6  Line 6 
6  and semantics are as close as possible to those of the Perl 5 language.  and semantics are as close as possible to those of the Perl 5 language.
7    
8                         Written by Philip Hazel                         Written by Philip Hazel
9             Copyright (c) 1997-2011 University of Cambridge             Copyright (c) 1997-2013 University of Cambridge
10    
11  -----------------------------------------------------------------------------  -----------------------------------------------------------------------------
12  Redistribution and use in source and binary forms, with or without  Redistribution and use in source and binary forms, with or without
# Line 37  POSSIBILITY OF SUCH DAMAGE. Line 37  POSSIBILITY OF SUCH DAMAGE.
37  -----------------------------------------------------------------------------  -----------------------------------------------------------------------------
38  */  */
39    
   
40  /* This module contains pcre_exec(), the externally visible function that does  /* This module contains pcre_exec(), the externally visible function that does
41  pattern matching using an NFA algorithm, trying to mimic Perl as closely as  pattern matching using an NFA algorithm, trying to mimic Perl as closely as
42  possible. There are also some static supporting functions. */  possible. There are also some static supporting functions. */
# Line 57  possible. There are also some static sup Line 56  possible. There are also some static sup
56  #undef min  #undef min
57  #undef max  #undef max
58    
59    /* The md->capture_last field uses the lower 16 bits for the last captured
60    substring (which can never be greater than 65535) and a bit in the top half
61    to mean "capture vector overflowed". This odd way of doing things was
62    implemented when it was realized that preserving and restoring the overflow bit
63    whenever the last capture number was saved/restored made for a neater
64    interface, and doing it this way saved on (a) another variable, which would
65    have increased the stack frame size (a big NO-NO in PCRE) and (b) another
66    separate set of save/restore instructions. The following defines are used in
67    implementing this. */
68    
69    #define CAPLMASK    0x0000ffff    /* The bits used for last_capture */
70    #define OVFLMASK    0xffff0000    /* The bits used for the overflow flag */
71    #define OVFLBIT     0x00010000    /* The bit that is set for overflow */
72    
73  /* Values for setting in md->match_function_type to indicate two special types  /* Values for setting in md->match_function_type to indicate two special types
74  of call to match(). We do it this way to save on using another stack variable,  of call to match(). We do it this way to save on using another stack variable,
75  as stack usage is to be discouraged. */  as stack usage is to be discouraged. */
# Line 74  defined PCRE_ERROR_xxx codes, which are Line 87  defined PCRE_ERROR_xxx codes, which are
87  negative to avoid the external error codes. */  negative to avoid the external error codes. */
88    
89  #define MATCH_ACCEPT       (-999)  #define MATCH_ACCEPT       (-999)
90  #define MATCH_COMMIT       (-998)  #define MATCH_KETRPOS      (-998)
91  #define MATCH_KETRPOS      (-997)  #define MATCH_ONCE         (-997)
92  #define MATCH_ONCE         (-996)  /* The next 5 must be kept together and in sequence so that a test that checks
93    for any one of them can use a range. */
94    #define MATCH_COMMIT       (-996)
95  #define MATCH_PRUNE        (-995)  #define MATCH_PRUNE        (-995)
96  #define MATCH_SKIP         (-994)  #define MATCH_SKIP         (-994)
97  #define MATCH_SKIP_ARG     (-993)  #define MATCH_SKIP_ARG     (-993)
98  #define MATCH_THEN         (-992)  #define MATCH_THEN         (-992)
99    #define MATCH_BACKTRACK_MAX MATCH_THEN
100    #define MATCH_BACKTRACK_MIN MATCH_COMMIT
101    
102  /* Maximum number of ints of offset to save on the stack for recursive calls.  /* Maximum number of ints of offset to save on the stack for recursive calls.
103  If the offset vector is bigger, malloc is used. This should be a multiple of 3,  If the offset vector is bigger, malloc is used. This should be a multiple of 3,
# Line 93  because the offset vector is always a mu Line 110  because the offset vector is always a mu
110  static const char rep_min[] = { 0, 0, 1, 1, 0, 0 };  static const char rep_min[] = { 0, 0, 1, 1, 0, 0 };
111  static const char rep_max[] = { 0, 0, 0, 0, 1, 1 };  static const char rep_max[] = { 0, 0, 0, 0, 1, 1 };
112    
   
   
113  #ifdef PCRE_DEBUG  #ifdef PCRE_DEBUG
114  /*************************************************  /*************************************************
115  *        Debugging function to print chars       *  *        Debugging function to print chars       *
# Line 113  Returns:     nothing Line 128  Returns:     nothing
128  */  */
129    
130  static void  static void
131  pchars(const uschar *p, int length, BOOL is_subject, match_data *md)  pchars(const pcre_uchar *p, int length, BOOL is_subject, match_data *md)
132  {  {
133  unsigned int c;  pcre_uint32 c;
134    BOOL utf = md->utf;
135  if (is_subject && length > md->end_subject - p) length = md->end_subject - p;  if (is_subject && length > md->end_subject - p) length = md->end_subject - p;
136  while (length-- > 0)  while (length-- > 0)
137    if (isprint(c = *(p++))) printf("%c", c); else printf("\\x%02x", c);    if (isprint(c = RAWUCHARINCTEST(p))) printf("%c", (char)c); else printf("\\x{%02x}", c);
138  }  }
139  #endif  #endif
140    
# Line 140  Arguments: Line 156  Arguments:
156    md          points to match data block    md          points to match data block
157    caseless    TRUE if caseless    caseless    TRUE if caseless
158    
159  Returns:      < 0 if not matched, otherwise the number of subject bytes matched  Returns:      >= 0 the number of subject bytes matched
160                  -1 no match
161                  -2 partial match; always given if at end subject
162  */  */
163    
164  static int  static int
165  match_ref(int offset, register USPTR eptr, int length, match_data *md,  match_ref(int offset, register PCRE_PUCHAR eptr, int length, match_data *md,
166    BOOL caseless)    BOOL caseless)
167  {  {
168  USPTR eptr_start = eptr;  PCRE_PUCHAR eptr_start = eptr;
169  register USPTR p = md->start_subject + md->offset_vector[offset];  register PCRE_PUCHAR p = md->start_subject + md->offset_vector[offset];
170    #ifdef SUPPORT_UTF
171    BOOL utf = md->utf;
172    #endif
173    
174  #ifdef PCRE_DEBUG  #ifdef PCRE_DEBUG
175  if (eptr >= md->end_subject)  if (eptr >= md->end_subject)
# Line 163  pchars(p, length, FALSE, md); Line 184  pchars(p, length, FALSE, md);
184  printf("\n");  printf("\n");
185  #endif  #endif
186    
187  /* Always fail if reference not set (and not JavaScript compatible). */  /* Always fail if reference not set (and not JavaScript compatible - in that
188    case the length is passed as zero). */
189    
190  if (length < 0) return -1;  if (length < 0) return -1;
191    
# Line 173  ASCII characters. */ Line 195  ASCII characters. */
195    
196  if (caseless)  if (caseless)
197    {    {
198  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
199  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
200    if (md->utf8)    if (utf)
201      {      {
202      /* Match characters up to the end of the reference. NOTE: the number of      /* Match characters up to the end of the reference. NOTE: the number of
203      bytes matched may differ, because there are some characters whose upper and      data units matched may differ, because in UTF-8 there are some characters
204      lower case versions code as different numbers of bytes. For example, U+023A      whose upper and lower case versions code have different numbers of bytes.
205      (2 bytes in UTF-8) is the upper case version of U+2C65 (3 bytes in UTF-8);      For example, U+023A (2 bytes in UTF-8) is the upper case version of U+2C65
206      a sequence of 3 of the former uses 6 bytes, as does a sequence of two of      (3 bytes in UTF-8); a sequence of 3 of the former uses 6 bytes, as does a
207      the latter. It is important, therefore, to check the length along the      sequence of two of the latter. It is important, therefore, to check the
208      reference, not along the subject (earlier code did this wrong). */      length along the reference, not along the subject (earlier code did this
209        wrong). */
210    
211      USPTR endptr = p + length;      PCRE_PUCHAR endptr = p + length;
212      while (p < endptr)      while (p < endptr)
213        {        {
214        int c, d;        pcre_uint32 c, d;
215        if (eptr >= md->end_subject) return -1;        const ucd_record *ur;
216          if (eptr >= md->end_subject) return -2;   /* Partial match */
217        GETCHARINC(c, eptr);        GETCHARINC(c, eptr);
218        GETCHARINC(d, p);        GETCHARINC(d, p);
219        if (c != d && c != UCD_OTHERCASE(d)) return -1;        ur = GET_UCD(d);
220          if (c != d && c != d + ur->other_case)
221            {
222            const pcre_uint32 *pp = PRIV(ucd_caseless_sets) + ur->caseset;
223            for (;;)
224              {
225              if (c < *pp) return -1;
226              if (c == *pp++) break;
227              }
228            }
229        }        }
230      }      }
231    else    else
# Line 202  if (caseless) Line 235  if (caseless)
235    /* The same code works when not in UTF-8 mode and in UTF-8 mode when there    /* The same code works when not in UTF-8 mode and in UTF-8 mode when there
236    is no UCP support. */    is no UCP support. */
237      {      {
     if (eptr + length > md->end_subject) return -1;  
238      while (length-- > 0)      while (length-- > 0)
239        { if (md->lcc[*p++] != md->lcc[*eptr++]) return -1; }        {
240          pcre_uint32 cc, cp;
241          if (eptr >= md->end_subject) return -2;   /* Partial match */
242          cc = RAWUCHARTEST(eptr);
243          cp = RAWUCHARTEST(p);
244          if (TABLE_GET(cp, md->lcc, cp) != TABLE_GET(cc, md->lcc, cc)) return -1;
245          p++;
246          eptr++;
247          }
248      }      }
249    }    }
250    
# Line 213  are in UTF-8 mode. */ Line 253  are in UTF-8 mode. */
253    
254  else  else
255    {    {
256    if (eptr + length > md->end_subject) return -1;    while (length-- > 0)
257    while (length-- > 0) if (*p++ != *eptr++) return -1;      {
258        if (eptr >= md->end_subject) return -2;   /* Partial match */
259        if (RAWUCHARINCTEST(p) != RAWUCHARINCTEST(eptr)) return -1;
260        }
261    }    }
262    
263  return (int)(eptr - eptr_start);  return (int)(eptr - eptr_start);
# Line 269  enum { RM1=1, RM2,  RM3,  RM4,  RM5,  RM Line 312  enum { RM1=1, RM2,  RM3,  RM4,  RM5,  RM
312         RM31,  RM32, RM33, RM34, RM35, RM36, RM37, RM38, RM39, RM40,         RM31,  RM32, RM33, RM34, RM35, RM36, RM37, RM38, RM39, RM40,
313         RM41,  RM42, RM43, RM44, RM45, RM46, RM47, RM48, RM49, RM50,         RM41,  RM42, RM43, RM44, RM45, RM46, RM47, RM48, RM49, RM50,
314         RM51,  RM52, RM53, RM54, RM55, RM56, RM57, RM58, RM59, RM60,         RM51,  RM52, RM53, RM54, RM55, RM56, RM57, RM58, RM59, RM60,
315         RM61,  RM62, RM63, RM64, RM65, RM66 };         RM61,  RM62, RM63, RM64, RM65, RM66, RM67, RM68 };
316    
317  /* These versions of the macros use the stack, as normal. There are debugging  /* These versions of the macros use the stack, as normal. There are debugging
318  versions and production versions. Note that the "rw" argument of RMATCH isn't  versions and production versions. Note that the "rw" argument of RMATCH isn't
# Line 287  actually used in this definition. */ Line 330  actually used in this definition. */
330    }    }
331  #define RRETURN(ra) \  #define RRETURN(ra) \
332    { \    { \
333    printf("match() returned %d from line %d ", ra, __LINE__); \    printf("match() returned %d from line %d\n", ra, __LINE__); \
334    return ra; \    return ra; \
335    }    }
336  #else  #else
# Line 307  argument of match(), which never changes Line 350  argument of match(), which never changes
350    
351  #define RMATCH(ra,rb,rc,rd,re,rw)\  #define RMATCH(ra,rb,rc,rd,re,rw)\
352    {\    {\
353    heapframe *newframe = (heapframe *)(pcre_stack_malloc)(sizeof(heapframe));\    heapframe *newframe = frame->Xnextframe;\
354    if (newframe == NULL) RRETURN(PCRE_ERROR_NOMEMORY);\    if (newframe == NULL)\
355    frame->Xwhere = rw; \      {\
356        newframe = (heapframe *)(PUBL(stack_malloc))(sizeof(heapframe));\
357        if (newframe == NULL) RRETURN(PCRE_ERROR_NOMEMORY);\
358        newframe->Xnextframe = NULL;\
359        frame->Xnextframe = newframe;\
360        }\
361      frame->Xwhere = rw;\
362    newframe->Xeptr = ra;\    newframe->Xeptr = ra;\
363    newframe->Xecode = rb;\    newframe->Xecode = rb;\
364    newframe->Xmstart = mstart;\    newframe->Xmstart = mstart;\
# Line 328  argument of match(), which never changes Line 377  argument of match(), which never changes
377    {\    {\
378    heapframe *oldframe = frame;\    heapframe *oldframe = frame;\
379    frame = oldframe->Xprevframe;\    frame = oldframe->Xprevframe;\
   (pcre_stack_free)(oldframe);\  
380    if (frame != NULL)\    if (frame != NULL)\
381      {\      {\
382      rrc = ra;\      rrc = ra;\
# Line 342  argument of match(), which never changes Line 390  argument of match(), which never changes
390    
391  typedef struct heapframe {  typedef struct heapframe {
392    struct heapframe *Xprevframe;    struct heapframe *Xprevframe;
393      struct heapframe *Xnextframe;
394    
395    /* Function arguments that may change */    /* Function arguments that may change */
396    
397    USPTR Xeptr;    PCRE_PUCHAR Xeptr;
398    const uschar *Xecode;    const pcre_uchar *Xecode;
399    USPTR Xmstart;    PCRE_PUCHAR Xmstart;
400    int Xoffset_top;    int Xoffset_top;
401    eptrblock *Xeptrb;    eptrblock *Xeptrb;
402    unsigned int Xrdepth;    unsigned int Xrdepth;
403    
404    /* Function local variables */    /* Function local variables */
405    
406    USPTR Xcallpat;    PCRE_PUCHAR Xcallpat;
407  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
408    USPTR Xcharptr;    PCRE_PUCHAR Xcharptr;
409  #endif  #endif
410    USPTR Xdata;    PCRE_PUCHAR Xdata;
411    USPTR Xnext;    PCRE_PUCHAR Xnext;
412    USPTR Xpp;    PCRE_PUCHAR Xpp;
413    USPTR Xprev;    PCRE_PUCHAR Xprev;
414    USPTR Xsaved_eptr;    PCRE_PUCHAR Xsaved_eptr;
415    
416    recursion_info Xnew_recursive;    recursion_info Xnew_recursive;
417    
# Line 372  typedef struct heapframe { Line 421  typedef struct heapframe {
421    
422  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
423    int Xprop_type;    int Xprop_type;
424    int Xprop_value;    unsigned int Xprop_value;
425    int Xprop_fail_result;    int Xprop_fail_result;
426    int Xoclength;    int Xoclength;
427    uschar Xocchars[8];    pcre_uchar Xocchars[6];
428  #endif  #endif
429    
430    int Xcodelink;    int Xcodelink;
# Line 385  typedef struct heapframe { Line 434  typedef struct heapframe {
434    int Xlength;    int Xlength;
435    int Xmax;    int Xmax;
436    int Xmin;    int Xmin;
437    int Xnumber;    unsigned int Xnumber;
438    int Xoffset;    int Xoffset;
439    int Xop;    unsigned int Xop;
440    int Xsave_capture_last;    pcre_int32 Xsave_capture_last;
441    int Xsave_offset1, Xsave_offset2, Xsave_offset3;    int Xsave_offset1, Xsave_offset2, Xsave_offset3;
442    int Xstacksave[REC_STACK_SAVE_MAX];    int Xstacksave[REC_STACK_SAVE_MAX];
443    
# Line 440  the subject. */ Line 489  the subject. */
489    
490    
491  /* Performance note: It might be tempting to extract commonly used fields from  /* Performance note: It might be tempting to extract commonly used fields from
492  the md structure (e.g. utf8, end_subject) into individual variables to improve  the md structure (e.g. utf, end_subject) into individual variables to improve
493  performance. Tests using gcc on a SPARC disproved this; in the first case, it  performance. Tests using gcc on a SPARC disproved this; in the first case, it
494  made performance worse.  made performance worse.
495    
# Line 463  Returns:       MATCH_MATCH if matched Line 512  Returns:       MATCH_MATCH if matched
512  */  */
513    
514  static int  static int
515  match(REGISTER USPTR eptr, REGISTER const uschar *ecode, USPTR mstart,  match(REGISTER PCRE_PUCHAR eptr, REGISTER const pcre_uchar *ecode,
516    int offset_top, match_data *md, eptrblock *eptrb, unsigned int rdepth)    PCRE_PUCHAR mstart, int offset_top, match_data *md, eptrblock *eptrb,
517      unsigned int rdepth)
518  {  {
519  /* These variables do not need to be preserved over recursion in this function,  /* These variables do not need to be preserved over recursion in this function,
520  so they can be ordinary variables in all cases. Mark some of them with  so they can be ordinary variables in all cases. Mark some of them with
# Line 472  so they can be ordinary variables in all Line 522  so they can be ordinary variables in all
522    
523  register int  rrc;         /* Returns from recursive calls */  register int  rrc;         /* Returns from recursive calls */
524  register int  i;           /* Used for loops not involving calls to RMATCH() */  register int  i;           /* Used for loops not involving calls to RMATCH() */
525  register unsigned int c;   /* Character values not kept over RMATCH() calls */  register pcre_uint32 c;    /* Character values not kept over RMATCH() calls */
526  register BOOL utf8;        /* Local copy of UTF-8 flag for speed */  register BOOL utf;         /* Local copy of UTF flag for speed */
527    
528  BOOL minimize, possessive; /* Quantifier options */  BOOL minimize, possessive; /* Quantifier options */
529  BOOL caseless;  BOOL caseless;
530  int condcode;  int condcode;
531    
532  /* When recursion is not being used, all "local" variables that have to be  /* When recursion is not being used, all "local" variables that have to be
533  preserved over calls to RMATCH() are part of a "frame" which is obtained from  preserved over calls to RMATCH() are part of a "frame". We set up the top-level
534  heap storage. Set up the top-level frame here; others are obtained from the  frame on the stack here; subsequent instantiations are obtained from the heap
535  heap whenever RMATCH() does a "recursion". See the macro definitions above. */  whenever RMATCH() does a "recursion". See the macro definitions above. Putting
536    the top-level on the stack rather than malloc-ing them all gives a performance
537    boost in many cases where there is not much "recursion". */
538    
539  #ifdef NO_RECURSE  #ifdef NO_RECURSE
540  heapframe *frame = (heapframe *)(pcre_stack_malloc)(sizeof(heapframe));  heapframe *frame = (heapframe *)md->match_frames_base;
 if (frame == NULL) RRETURN(PCRE_ERROR_NOMEMORY);  
 frame->Xprevframe = NULL;            /* Marks the top level */  
541    
542  /* Copy in the original argument variables */  /* Copy in the original argument variables */
543    
# Line 513  HEAP_RECURSE: Line 563  HEAP_RECURSE:
563    
564  /* Ditto for the local variables */  /* Ditto for the local variables */
565    
566  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
567  #define charptr            frame->Xcharptr  #define charptr            frame->Xcharptr
568  #endif  #endif
569  #define callpat            frame->Xcallpat  #define callpat            frame->Xcallpat
# Line 571  declarations can be cut out in a block. Line 621  declarations can be cut out in a block.
621  below are for variables that do not have to be preserved over a recursive call  below are for variables that do not have to be preserved over a recursive call
622  to RMATCH(). */  to RMATCH(). */
623    
624  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
625  const uschar *charptr;  const pcre_uchar *charptr;
626  #endif  #endif
627  const uschar *callpat;  const pcre_uchar *callpat;
628  const uschar *data;  const pcre_uchar *data;
629  const uschar *next;  const pcre_uchar *next;
630  USPTR         pp;  PCRE_PUCHAR       pp;
631  const uschar *prev;  const pcre_uchar *prev;
632  USPTR         saved_eptr;  PCRE_PUCHAR       saved_eptr;
633    
634  recursion_info new_recursive;  recursion_info new_recursive;
635    
# Line 589  BOOL prev_is_word; Line 639  BOOL prev_is_word;
639    
640  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
641  int prop_type;  int prop_type;
642  int prop_value;  unsigned int prop_value;
643  int prop_fail_result;  int prop_fail_result;
644  int oclength;  int oclength;
645  uschar occhars[8];  pcre_uchar occhars[6];
646  #endif  #endif
647    
648  int codelink;  int codelink;
# Line 600  int ctype; Line 650  int ctype;
650  int length;  int length;
651  int max;  int max;
652  int min;  int min;
653  int number;  unsigned int number;
654  int offset;  int offset;
655  int op;  unsigned int op;
656  int save_capture_last;  pcre_int32 save_capture_last;
657  int save_offset1, save_offset2, save_offset3;  int save_offset1, save_offset2, save_offset3;
658  int stacksave[REC_STACK_SAVE_MAX];  int stacksave[REC_STACK_SAVE_MAX];
659    
660  eptrblock newptrb;  eptrblock newptrb;
661    
662    /* There is a special fudge for calling match() in a way that causes it to
663    measure the size of its basic stack frame when the stack is being used for
664    recursion. The second argument (ecode) being NULL triggers this behaviour. It
665    cannot normally ever be NULL. The return is the negated value of the frame
666    size. */
667    
668    if (ecode == NULL)
669      {
670      if (rdepth == 0)
671        return match((PCRE_PUCHAR)&rdepth, NULL, NULL, 0, NULL, NULL, 1);
672      else
673        {
674        int len = (char *)&rdepth - (char *)eptr;
675        return (len > 0)? -len : len;
676        }
677      }
678  #endif     /* NO_RECURSE */  #endif     /* NO_RECURSE */
679    
680  /* To save space on the stack and in the heap frame, I have doubled up on some  /* To save space on the stack and in the heap frame, I have doubled up on some
# Line 620  the alternative names that are used. */ Line 687  the alternative names that are used. */
687  #define code_offset   codelink  #define code_offset   codelink
688  #define condassert    condition  #define condassert    condition
689  #define matched_once  prev_is_word  #define matched_once  prev_is_word
690    #define foc           number
691    #define save_mark     data
692    
693  /* These statements are here to stop the compiler complaining about unitialized  /* These statements are here to stop the compiler complaining about unitialized
694  variables. */  variables. */
# Line 645  defined). However, RMATCH isn't like a f Line 714  defined). However, RMATCH isn't like a f
714  complicated macro. It has to be used in one particular way. This shouldn't,  complicated macro. It has to be used in one particular way. This shouldn't,
715  however, impact performance when true recursion is being used. */  however, impact performance when true recursion is being used. */
716    
717  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
718  utf8 = md->utf8;       /* Local copy of the flag */  utf = md->utf;       /* Local copy of the flag */
719  #else  #else
720  utf8 = FALSE;  utf = FALSE;
721  #endif  #endif
722    
723  /* First check that we haven't called match() too many times, or that we  /* First check that we haven't called match() too many times, or that we
# Line 689  for (;;) Line 758  for (;;)
758      case OP_MARK:      case OP_MARK:
759      md->nomatch_mark = ecode + 2;      md->nomatch_mark = ecode + 2;
760      md->mark = NULL;    /* In case previously set by assertion */      md->mark = NULL;    /* In case previously set by assertion */
761      RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode] + ecode[1], offset_top, md,      RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode] + ecode[1], offset_top, md,
762        eptrb, RM55);        eptrb, RM55);
763      if ((rrc == MATCH_MATCH || rrc == MATCH_ACCEPT) &&      if ((rrc == MATCH_MATCH || rrc == MATCH_ACCEPT) &&
764           md->mark == NULL) md->mark = ecode + 2;           md->mark == NULL) md->mark = ecode + 2;
# Line 702  for (;;) Line 771  for (;;)
771      unaltered. */      unaltered. */
772    
773      else if (rrc == MATCH_SKIP_ARG &&      else if (rrc == MATCH_SKIP_ARG &&
774          strcmp((char *)(ecode + 2), (char *)(md->start_match_ptr)) == 0)          STRCMP_UC_UC_TEST(ecode + 2, md->start_match_ptr) == 0)
775        {        {
776        md->start_match_ptr = eptr;        md->start_match_ptr = eptr;
777        RRETURN(MATCH_SKIP);        RRETURN(MATCH_SKIP);
# Line 712  for (;;) Line 781  for (;;)
781      case OP_FAIL:      case OP_FAIL:
782      RRETURN(MATCH_NOMATCH);      RRETURN(MATCH_NOMATCH);
783    
     /* COMMIT overrides PRUNE, SKIP, and THEN */  
   
784      case OP_COMMIT:      case OP_COMMIT:
785      RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,      RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md,
786        eptrb, RM52);        eptrb, RM52);
787      if (rrc != MATCH_NOMATCH && rrc != MATCH_PRUNE &&      if (rrc != MATCH_NOMATCH) RRETURN(rrc);
         rrc != MATCH_SKIP && rrc != MATCH_SKIP_ARG &&  
         rrc != MATCH_THEN)  
       RRETURN(rrc);  
788      RRETURN(MATCH_COMMIT);      RRETURN(MATCH_COMMIT);
789    
     /* PRUNE overrides THEN */  
   
790      case OP_PRUNE:      case OP_PRUNE:
791      RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,      RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md,
792        eptrb, RM51);        eptrb, RM51);
793      if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN) RRETURN(rrc);      if (rrc != MATCH_NOMATCH) RRETURN(rrc);
794      RRETURN(MATCH_PRUNE);      RRETURN(MATCH_PRUNE);
795    
796      case OP_PRUNE_ARG:      case OP_PRUNE_ARG:
797      md->nomatch_mark = ecode + 2;      md->nomatch_mark = ecode + 2;
798      md->mark = NULL;    /* In case previously set by assertion */      md->mark = NULL;    /* In case previously set by assertion */
799      RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode] + ecode[1], offset_top, md,      RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode] + ecode[1], offset_top, md,
800        eptrb, RM56);        eptrb, RM56);
801      if ((rrc == MATCH_MATCH || rrc == MATCH_ACCEPT) &&      if ((rrc == MATCH_MATCH || rrc == MATCH_ACCEPT) &&
802           md->mark == NULL) md->mark = ecode + 2;           md->mark == NULL) md->mark = ecode + 2;
803      if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN) RRETURN(rrc);      if (rrc != MATCH_NOMATCH) RRETURN(rrc);
804      RRETURN(MATCH_PRUNE);      RRETURN(MATCH_PRUNE);
805    
     /* SKIP overrides PRUNE and THEN */  
   
806      case OP_SKIP:      case OP_SKIP:
807      RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,      RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md,
808        eptrb, RM53);        eptrb, RM53);
809      if (rrc != MATCH_NOMATCH && rrc != MATCH_PRUNE && rrc != MATCH_THEN)      if (rrc != MATCH_NOMATCH) RRETURN(rrc);
       RRETURN(rrc);  
810      md->start_match_ptr = eptr;   /* Pass back current position */      md->start_match_ptr = eptr;   /* Pass back current position */
811      RRETURN(MATCH_SKIP);      RRETURN(MATCH_SKIP);
812    
813      /* Note that, for Perl compatibility, SKIP with an argument does NOT set      /* Note that, for Perl compatibility, SKIP with an argument does NOT set
814      nomatch_mark. There is a flag that disables this opcode when re-matching a      nomatch_mark. When a pattern match ends with a SKIP_ARG for which there was
815      pattern that ended with a SKIP for which there was not a matching MARK. */      not a matching mark, we have to re-run the match, ignoring the SKIP_ARG
816        that failed and any that preceed it (either they also failed, or were not
817        triggered). To do this, we maintain a count of executed SKIP_ARGs. If a
818        SKIP_ARG gets to top level, the match is re-run with md->ignore_skip_arg
819        set to the count of the one that failed. */
820    
821      case OP_SKIP_ARG:      case OP_SKIP_ARG:
822      if (md->ignore_skip_arg)      md->skip_arg_count++;
823        if (md->skip_arg_count <= md->ignore_skip_arg)
824        {        {
825        ecode += _pcre_OP_lengths[*ecode] + ecode[1];        ecode += PRIV(OP_lengths)[*ecode] + ecode[1];
826        break;        break;
827        }        }
828      RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode] + ecode[1], offset_top, md,      RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode] + ecode[1], offset_top, md,
829        eptrb, RM57);        eptrb, RM57);
830      if (rrc != MATCH_NOMATCH && rrc != MATCH_PRUNE && rrc != MATCH_THEN)      if (rrc != MATCH_NOMATCH) RRETURN(rrc);
831        RRETURN(rrc);  
   
832      /* Pass back the current skip name by overloading md->start_match_ptr and      /* Pass back the current skip name by overloading md->start_match_ptr and
833      returning the special MATCH_SKIP_ARG return code. This will either be      returning the special MATCH_SKIP_ARG return code. This will either be
834      caught by a matching MARK, or get to the top, where it causes a rematch      caught by a matching MARK, or get to the top, where it causes a rematch
835      with the md->ignore_skip_arg flag set. */      with md->ignore_skip_arg set to the value of md->skip_arg_count. */
836    
837      md->start_match_ptr = ecode + 2;      md->start_match_ptr = ecode + 2;
838      RRETURN(MATCH_SKIP_ARG);      RRETURN(MATCH_SKIP_ARG);
# Line 779  for (;;) Line 842  for (;;)
842      match pointer to do this. */      match pointer to do this. */
843    
844      case OP_THEN:      case OP_THEN:
845      RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,      RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md,
846        eptrb, RM54);        eptrb, RM54);
847      if (rrc != MATCH_NOMATCH) RRETURN(rrc);      if (rrc != MATCH_NOMATCH) RRETURN(rrc);
848      md->start_match_ptr = ecode;      md->start_match_ptr = ecode;
# Line 788  for (;;) Line 851  for (;;)
851      case OP_THEN_ARG:      case OP_THEN_ARG:
852      md->nomatch_mark = ecode + 2;      md->nomatch_mark = ecode + 2;
853      md->mark = NULL;    /* In case previously set by assertion */      md->mark = NULL;    /* In case previously set by assertion */
854      RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode] + ecode[1], offset_top,      RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode] + ecode[1], offset_top,
855        md, eptrb, RM58);        md, eptrb, RM58);
856      if ((rrc == MATCH_MATCH || rrc == MATCH_ACCEPT) &&      if ((rrc == MATCH_MATCH || rrc == MATCH_ACCEPT) &&
857           md->mark == NULL) md->mark = ecode + 2;           md->mark == NULL) md->mark = ecode + 2;
# Line 812  for (;;) Line 875  for (;;)
875      case OP_ONCE_NC:      case OP_ONCE_NC:
876      prev = ecode;      prev = ecode;
877      saved_eptr = eptr;      saved_eptr = eptr;
878        save_mark = md->mark;
879      do      do
880        {        {
881        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, eptrb, RM64);        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, eptrb, RM64);
# Line 830  for (;;) Line 894  for (;;)
894    
895        if (rrc != MATCH_NOMATCH) RRETURN(rrc);        if (rrc != MATCH_NOMATCH) RRETURN(rrc);
896        ecode += GET(ecode,1);        ecode += GET(ecode,1);
897          md->mark = save_mark;
898        }        }
899      while (*ecode == OP_ALT);      while (*ecode == OP_ALT);
900    
# Line 869  for (;;) Line 934  for (;;)
934        }        }
935      else  /* OP_KETRMAX */      else  /* OP_KETRMAX */
936        {        {
       md->match_function_type = MATCH_CBEGROUP;  
937        RMATCH(eptr, prev, offset_top, md, eptrb, RM66);        RMATCH(eptr, prev, offset_top, md, eptrb, RM66);
938        if (rrc != MATCH_NOMATCH) RRETURN(rrc);        if (rrc != MATCH_NOMATCH) RRETURN(rrc);
939        ecode += 1 + LINK_SIZE;        ecode += 1 + LINK_SIZE;
# Line 909  for (;;) Line 973  for (;;)
973        save_offset2 = md->offset_vector[offset+1];        save_offset2 = md->offset_vector[offset+1];
974        save_offset3 = md->offset_vector[md->offset_end - number];        save_offset3 = md->offset_vector[md->offset_end - number];
975        save_capture_last = md->capture_last;        save_capture_last = md->capture_last;
976          save_mark = md->mark;
977    
978        DPRINTF(("saving %d %d %d\n", save_offset1, save_offset2, save_offset3));        DPRINTF(("saving %d %d %d\n", save_offset1, save_offset2, save_offset3));
979        md->offset_vector[md->offset_end - number] =        md->offset_vector[md->offset_end - number] =
# Line 917  for (;;) Line 982  for (;;)
982        for (;;)        for (;;)
983          {          {
984          if (op >= OP_SBRA) md->match_function_type = MATCH_CBEGROUP;          if (op >= OP_SBRA) md->match_function_type = MATCH_CBEGROUP;
985          RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,          RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md,
986            eptrb, RM1);            eptrb, RM1);
987          if (rrc == MATCH_ONCE) break;  /* Backing up through an atomic group */          if (rrc == MATCH_ONCE) break;  /* Backing up through an atomic group */
988    
# Line 945  for (;;) Line 1010  for (;;)
1010          if (rrc != MATCH_NOMATCH) RRETURN(rrc);          if (rrc != MATCH_NOMATCH) RRETURN(rrc);
1011          md->capture_last = save_capture_last;          md->capture_last = save_capture_last;
1012          ecode += GET(ecode, 1);          ecode += GET(ecode, 1);
1013            md->mark = save_mark;
1014          if (*ecode != OP_ALT) break;          if (*ecode != OP_ALT) break;
1015          }          }
1016    
# Line 996  for (;;) Line 1062  for (;;)
1062    
1063      for (;;)      for (;;)
1064        {        {
1065        if (op >= OP_SBRA || op == OP_ONCE) md->match_function_type = MATCH_CBEGROUP;        if (op >= OP_SBRA || op == OP_ONCE)
1066            md->match_function_type = MATCH_CBEGROUP;
1067    
1068        /* If this is not a possibly empty group, and there are no (*THEN)s in        /* If this is not a possibly empty group, and there are no (*THEN)s in
1069        the pattern, and this is the final alternative, optimize as described        the pattern, and this is the final alternative, optimize as described
# Line 1004  for (;;) Line 1071  for (;;)
1071    
1072        else if (!md->hasthen && ecode[GET(ecode, 1)] != OP_ALT)        else if (!md->hasthen && ecode[GET(ecode, 1)] != OP_ALT)
1073          {          {
1074          ecode += _pcre_OP_lengths[*ecode];          ecode += PRIV(OP_lengths)[*ecode];
1075          goto TAIL_RECURSE;          goto TAIL_RECURSE;
1076          }          }
1077    
1078        /* In all other cases, we have to make another call to match(). */        /* In all other cases, we have to make another call to match(). */
1079    
1080        RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md, eptrb,        save_mark = md->mark;
1081          save_capture_last = md->capture_last;
1082          RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md, eptrb,
1083          RM2);          RM2);
1084    
1085        /* See comment in the code for capturing groups above about handling        /* See comment in the code for capturing groups above about handling
# Line 1028  for (;;) Line 1097  for (;;)
1097          {          {
1098          if (rrc == MATCH_ONCE)          if (rrc == MATCH_ONCE)
1099            {            {
1100            const uschar *scode = ecode;            const pcre_uchar *scode = ecode;
1101            if (*scode != OP_ONCE)           /* If not at start, find it */            if (*scode != OP_ONCE)           /* If not at start, find it */
1102              {              {
1103              while (*scode == OP_ALT) scode += GET(scode, 1);              while (*scode == OP_ALT) scode += GET(scode, 1);
# Line 1039  for (;;) Line 1108  for (;;)
1108          RRETURN(rrc);          RRETURN(rrc);
1109          }          }
1110        ecode += GET(ecode, 1);        ecode += GET(ecode, 1);
1111          md->mark = save_mark;
1112        if (*ecode != OP_ALT) break;        if (*ecode != OP_ALT) break;
1113          md->capture_last = save_capture_last;
1114        }        }
1115    
1116      RRETURN(MATCH_NOMATCH);      RRETURN(MATCH_NOMATCH);
# Line 1093  for (;;) Line 1164  for (;;)
1164          md->offset_vector[md->offset_end - number] =          md->offset_vector[md->offset_end - number] =
1165            (int)(eptr - md->start_subject);            (int)(eptr - md->start_subject);
1166          if (op >= OP_SBRA) md->match_function_type = MATCH_CBEGROUP;          if (op >= OP_SBRA) md->match_function_type = MATCH_CBEGROUP;
1167          RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,          RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md,
1168            eptrb, RM63);            eptrb, RM63);
1169          if (rrc == MATCH_KETRPOS)          if (rrc == MATCH_KETRPOS)
1170            {            {
# Line 1161  for (;;) Line 1232  for (;;)
1232      POSSESSIVE_NON_CAPTURE:      POSSESSIVE_NON_CAPTURE:
1233      matched_once = FALSE;      matched_once = FALSE;
1234      code_offset = (int)(ecode - md->start_code);      code_offset = (int)(ecode - md->start_code);
1235        save_capture_last = md->capture_last;
1236    
1237      for (;;)      for (;;)
1238        {        {
1239        if (op >= OP_SBRA) md->match_function_type = MATCH_CBEGROUP;        if (op >= OP_SBRA) md->match_function_type = MATCH_CBEGROUP;
1240        RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,        RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md,
1241          eptrb, RM48);          eptrb, RM48);
1242        if (rrc == MATCH_KETRPOS)        if (rrc == MATCH_KETRPOS)
1243          {          {
# Line 1190  for (;;) Line 1262  for (;;)
1262        if (rrc != MATCH_NOMATCH) RRETURN(rrc);        if (rrc != MATCH_NOMATCH) RRETURN(rrc);
1263        ecode += GET(ecode, 1);        ecode += GET(ecode, 1);
1264        if (*ecode != OP_ALT) break;        if (*ecode != OP_ALT) break;
1265          md->capture_last = save_capture_last;
1266        }        }
1267    
1268      if (matched_once || allow_zero)      if (matched_once || allow_zero)
# Line 1215  for (;;) Line 1288  for (;;)
1288    
1289      if (ecode[LINK_SIZE+1] == OP_CALLOUT)      if (ecode[LINK_SIZE+1] == OP_CALLOUT)
1290        {        {
1291        if (pcre_callout != NULL)        if (PUBL(callout) != NULL)
1292          {          {
1293          pcre_callout_block cb;          PUBL(callout_block) cb;
1294          cb.version          = 2;   /* Version 1 of the callout block */          cb.version          = 2;   /* Version 1 of the callout block */
1295          cb.callout_number   = ecode[LINK_SIZE+2];          cb.callout_number   = ecode[LINK_SIZE+2];
1296          cb.offset_vector    = md->offset_vector;          cb.offset_vector    = md->offset_vector;
1297    #if defined COMPILE_PCRE8
1298          cb.subject          = (PCRE_SPTR)md->start_subject;          cb.subject          = (PCRE_SPTR)md->start_subject;
1299    #elif defined COMPILE_PCRE16
1300            cb.subject          = (PCRE_SPTR16)md->start_subject;
1301    #elif defined COMPILE_PCRE32
1302            cb.subject          = (PCRE_SPTR32)md->start_subject;
1303    #endif
1304          cb.subject_length   = (int)(md->end_subject - md->start_subject);          cb.subject_length   = (int)(md->end_subject - md->start_subject);
1305          cb.start_match      = (int)(mstart - md->start_subject);          cb.start_match      = (int)(mstart - md->start_subject);
1306          cb.current_position = (int)(eptr - md->start_subject);          cb.current_position = (int)(eptr - md->start_subject);
1307          cb.pattern_position = GET(ecode, LINK_SIZE + 3);          cb.pattern_position = GET(ecode, LINK_SIZE + 3);
1308          cb.next_item_length = GET(ecode, 3 + 2*LINK_SIZE);          cb.next_item_length = GET(ecode, 3 + 2*LINK_SIZE);
1309          cb.capture_top      = offset_top/2;          cb.capture_top      = offset_top/2;
1310          cb.capture_last     = md->capture_last;          cb.capture_last     = md->capture_last & CAPLMASK;
1311            /* Internal change requires this for API compatibility. */
1312            if (cb.capture_last == 0) cb.capture_last = -1;
1313          cb.callout_data     = md->callout_data;          cb.callout_data     = md->callout_data;
1314          cb.mark             = md->nomatch_mark;          cb.mark             = md->nomatch_mark;
1315          if ((rrc = (*pcre_callout)(&cb)) > 0) RRETURN(MATCH_NOMATCH);          if ((rrc = (*PUBL(callout))(&cb)) > 0) RRETURN(MATCH_NOMATCH);
1316          if (rrc < 0) RRETURN(rrc);          if (rrc < 0) RRETURN(rrc);
1317          }          }
1318        ecode += _pcre_OP_lengths[OP_CALLOUT];        ecode += PRIV(OP_lengths)[OP_CALLOUT];
1319          codelink -= PRIV(OP_lengths)[OP_CALLOUT];
1320        }        }
1321    
1322      condcode = ecode[LINK_SIZE+1];      condcode = ecode[LINK_SIZE+1];
# Line 1250  for (;;) Line 1332  for (;;)
1332          }          }
1333        else        else
1334          {          {
1335          int recno = GET2(ecode, LINK_SIZE + 2);   /* Recursion group number*/          unsigned int recno = GET2(ecode, LINK_SIZE + 2);   /* Recursion group number*/
1336          condition = (recno == RREF_ANY || recno == md->recursive->group_num);          condition = (recno == RREF_ANY || recno == md->recursive->group_num);
1337    
1338          /* If the test is for recursion into a specific subpattern, and it is          /* If the test is for recursion into a specific subpattern, and it is
# Line 1260  for (;;) Line 1342  for (;;)
1342    
1343          if (!condition && condcode == OP_NRREF)          if (!condition && condcode == OP_NRREF)
1344            {            {
1345            uschar *slotA = md->name_table;            pcre_uchar *slotA = md->name_table;
1346            for (i = 0; i < md->name_count; i++)            for (i = 0; i < md->name_count; i++)
1347              {              {
1348              if (GET2(slotA, 0) == recno) break;              if (GET2(slotA, 0) == recno) break;
# Line 1273  for (;;) Line 1355  for (;;)
1355    
1356            if (i < md->name_count)            if (i < md->name_count)
1357              {              {
1358              uschar *slotB = slotA;              pcre_uchar *slotB = slotA;
1359              while (slotB > md->name_table)              while (slotB > md->name_table)
1360                {                {
1361                slotB -= md->name_entry_size;                slotB -= md->name_entry_size;
1362                if (strcmp((char *)slotA + 2, (char *)slotB + 2) == 0)                if (STRCMP_UC_UC(slotA + IMM2_SIZE, slotB + IMM2_SIZE) == 0)
1363                  {                  {
1364                  condition = GET2(slotB, 0) == md->recursive->group_num;                  condition = GET2(slotB, 0) == md->recursive->group_num;
1365                  if (condition) break;                  if (condition) break;
# Line 1293  for (;;) Line 1375  for (;;)
1375                for (i++; i < md->name_count; i++)                for (i++; i < md->name_count; i++)
1376                  {                  {
1377                  slotB += md->name_entry_size;                  slotB += md->name_entry_size;
1378                  if (strcmp((char *)slotA + 2, (char *)slotB + 2) == 0)                  if (STRCMP_UC_UC(slotA + IMM2_SIZE, slotB + IMM2_SIZE) == 0)
1379                    {                    {
1380                    condition = GET2(slotB, 0) == md->recursive->group_num;                    condition = GET2(slotB, 0) == md->recursive->group_num;
1381                    if (condition) break;                    if (condition) break;
# Line 1306  for (;;) Line 1388  for (;;)
1388    
1389          /* Chose branch according to the condition */          /* Chose branch according to the condition */
1390    
1391          ecode += condition? 3 : GET(ecode, 1);          ecode += condition? 1 + IMM2_SIZE : GET(ecode, 1);
1392          }          }
1393        }        }
1394    
# Line 1322  for (;;) Line 1404  for (;;)
1404    
1405        if (!condition && condcode == OP_NCREF)        if (!condition && condcode == OP_NCREF)
1406          {          {
1407          int refno = offset >> 1;          unsigned int refno = offset >> 1;
1408          uschar *slotA = md->name_table;          pcre_uchar *slotA = md->name_table;
1409    
1410          for (i = 0; i < md->name_count; i++)          for (i = 0; i < md->name_count; i++)
1411            {            {
# Line 1337  for (;;) Line 1419  for (;;)
1419    
1420          if (i < md->name_count)          if (i < md->name_count)
1421            {            {
1422            uschar *slotB = slotA;            pcre_uchar *slotB = slotA;
1423            while (slotB > md->name_table)            while (slotB > md->name_table)
1424              {              {
1425              slotB -= md->name_entry_size;              slotB -= md->name_entry_size;
1426              if (strcmp((char *)slotA + 2, (char *)slotB + 2) == 0)              if (STRCMP_UC_UC(slotA + IMM2_SIZE, slotB + IMM2_SIZE) == 0)
1427                {                {
1428                offset = GET2(slotB, 0) << 1;                offset = GET2(slotB, 0) << 1;
1429                condition = offset < offset_top &&                condition = offset < offset_top &&
# Line 1359  for (;;) Line 1441  for (;;)
1441              for (i++; i < md->name_count; i++)              for (i++; i < md->name_count; i++)
1442                {                {
1443                slotB += md->name_entry_size;                slotB += md->name_entry_size;
1444                if (strcmp((char *)slotA + 2, (char *)slotB + 2) == 0)                if (STRCMP_UC_UC(slotA + IMM2_SIZE, slotB + IMM2_SIZE) == 0)
1445                  {                  {
1446                  offset = GET2(slotB, 0) << 1;                  offset = GET2(slotB, 0) << 1;
1447                  condition = offset < offset_top &&                  condition = offset < offset_top &&
# Line 1374  for (;;) Line 1456  for (;;)
1456    
1457        /* Chose branch according to the condition */        /* Chose branch according to the condition */
1458    
1459        ecode += condition? 3 : GET(ecode, 1);        ecode += condition? 1 + IMM2_SIZE : GET(ecode, 1);
1460        }        }
1461    
1462      else if (condcode == OP_DEF)     /* DEFINE - always false */      else if (condcode == OP_DEF)     /* DEFINE - always false */
# Line 1450  for (;;) Line 1532  for (;;)
1532      to close any currently open capturing brackets. */      to close any currently open capturing brackets. */
1533    
1534      case OP_CLOSE:      case OP_CLOSE:
1535      number = GET2(ecode, 1);      number = GET2(ecode, 1);   /* Must be less than 65536 */
1536      offset = number << 1;      offset = number << 1;
1537    
1538  #ifdef PCRE_DEBUG  #ifdef PCRE_DEBUG
# Line 1458  for (;;) Line 1540  for (;;)
1540        printf("\n");        printf("\n");
1541  #endif  #endif
1542    
1543      md->capture_last = number;      md->capture_last = (md->capture_last & OVFLMASK) | number;
1544      if (offset >= md->offset_max) md->offset_overflow = TRUE; else      if (offset >= md->offset_max) md->capture_last |= OVFLBIT; else
1545        {        {
1546        md->offset_vector[offset] =        md->offset_vector[offset] =
1547          md->offset_vector[md->offset_end - number];          md->offset_vector[md->offset_end - number];
1548        md->offset_vector[offset+1] = (int)(eptr - md->start_subject);        md->offset_vector[offset+1] = (int)(eptr - md->start_subject);
1549        if (offset_top <= offset) offset_top = offset + 2;        if (offset_top <= offset) offset_top = offset + 2;
1550        }        }
1551      ecode += 3;      ecode += 1 + IMM2_SIZE;
1552      break;      break;
1553    
1554    
# Line 1513  for (;;) Line 1595  for (;;)
1595    
1596      case OP_ASSERT:      case OP_ASSERT:
1597      case OP_ASSERTBACK:      case OP_ASSERTBACK:
1598        save_mark = md->mark;
1599      if (md->match_function_type == MATCH_CONDASSERT)      if (md->match_function_type == MATCH_CONDASSERT)
1600        {        {
1601        condassert = TRUE;        condassert = TRUE;
# Line 1520  for (;;) Line 1603  for (;;)
1603        }        }
1604      else condassert = FALSE;      else condassert = FALSE;
1605    
1606        /* Loop for each branch */
1607    
1608      do      do
1609        {        {
1610        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, NULL, RM4);        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, NULL, RM4);
# Line 1528  for (;;) Line 1613  for (;;)
1613          mstart = md->start_match_ptr;   /* In case \K reset it */          mstart = md->start_match_ptr;   /* In case \K reset it */
1614          break;          break;
1615          }          }
1616          md->mark = save_mark;
1617    
1618        /* PCRE does not allow THEN to escape beyond an assertion; it is treated        /* See comment in the code for capturing groups above about handling
1619        as NOMATCH. */        THEN. */
1620    
1621        if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN) RRETURN(rrc);        if (rrc == MATCH_THEN)
1622            {
1623            next = ecode + GET(ecode,1);
1624            if (md->start_match_ptr < next &&
1625                (*ecode == OP_ALT || *next == OP_ALT))
1626              rrc = MATCH_NOMATCH;
1627            }
1628    
1629          /* Anything other than NOMATCH causes the assertion to fail. This
1630          includes COMMIT, SKIP, and PRUNE. However, this consistent approach does
1631          not always have exactly the same effect as in Perl. */
1632    
1633          if (rrc != MATCH_NOMATCH) RRETURN(rrc);
1634        ecode += GET(ecode, 1);        ecode += GET(ecode, 1);
1635        }        }
1636      while (*ecode == OP_ALT);      while (*ecode == OP_ALT);
1637    
1638        /* If we have tried all the alternative branches, the assertion has
1639        failed. */
1640    
1641      if (*ecode == OP_KET) RRETURN(MATCH_NOMATCH);      if (*ecode == OP_KET) RRETURN(MATCH_NOMATCH);
1642    
# Line 1543  for (;;) Line 1644  for (;;)
1644    
1645      if (condassert) RRETURN(MATCH_MATCH);      if (condassert) RRETURN(MATCH_MATCH);
1646    
1647      /* Continue from after the assertion, updating the offsets high water      /* Continue from after a successful assertion, updating the offsets high
1648      mark, since extracts may have been taken during the assertion. */      water mark, since extracts may have been taken during the assertion. */
1649    
1650      do ecode += GET(ecode,1); while (*ecode == OP_ALT);      do ecode += GET(ecode,1); while (*ecode == OP_ALT);
1651      ecode += 1 + LINK_SIZE;      ecode += 1 + LINK_SIZE;
1652      offset_top = md->end_offset_top;      offset_top = md->end_offset_top;
1653      continue;      continue;
1654    
1655      /* Negative assertion: all branches must fail to match. Encountering SKIP,      /* Negative assertion: all branches must fail to match for the assertion to
1656      PRUNE, or COMMIT means we must assume failure without checking subsequent      succeed. */
     branches. */  
1657    
1658      case OP_ASSERT_NOT:      case OP_ASSERT_NOT:
1659      case OP_ASSERTBACK_NOT:      case OP_ASSERTBACK_NOT:
1660        save_mark = md->mark;
1661      if (md->match_function_type == MATCH_CONDASSERT)      if (md->match_function_type == MATCH_CONDASSERT)
1662        {        {
1663        condassert = TRUE;        condassert = TRUE;
# Line 1564  for (;;) Line 1665  for (;;)
1665        }        }
1666      else condassert = FALSE;      else condassert = FALSE;
1667    
1668        /* Loop for each alternative branch. */
1669    
1670      do      do
1671        {        {
1672        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, NULL, RM5);        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, NULL, RM5);
1673          md->mark = save_mark;
1674    
1675          /* A successful match means the assertion has failed. */
1676    
1677        if (rrc == MATCH_MATCH || rrc == MATCH_ACCEPT) RRETURN(MATCH_NOMATCH);        if (rrc == MATCH_MATCH || rrc == MATCH_ACCEPT) RRETURN(MATCH_NOMATCH);
1678        if (rrc == MATCH_SKIP || rrc == MATCH_PRUNE || rrc == MATCH_COMMIT)  
1679          /* See comment in the code for capturing groups above about handling
1680          THEN. */
1681    
1682          if (rrc == MATCH_THEN)
1683          {          {
1684          do ecode += GET(ecode,1); while (*ecode == OP_ALT);          next = ecode + GET(ecode,1);
1685          break;          if (md->start_match_ptr < next &&
1686                (*ecode == OP_ALT || *next == OP_ALT))
1687              rrc = MATCH_NOMATCH;
1688          }          }
1689    
1690          /* No match on a branch means we must carry on and try the next branch.
1691          Anything else, in particular, SKIP, PRUNE, etc. causes a failure in the
1692          enclosing branch. This is a consistent approach, but does not always have
1693          the same effect as in Perl. */
1694    
1695        /* PCRE does not allow THEN to escape beyond an assertion; it is treated        if (rrc != MATCH_NOMATCH) RRETURN(rrc);
       as NOMATCH. */  
   
       if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN) RRETURN(rrc);  
1696        ecode += GET(ecode,1);        ecode += GET(ecode,1);
1697        }        }
1698      while (*ecode == OP_ALT);      while (*ecode == OP_ALT);
1699    
1700        /* All branches in the assertion failed to match. */
1701    
1702      if (condassert) RRETURN(MATCH_MATCH);  /* Condition assertion */      if (condassert) RRETURN(MATCH_MATCH);  /* Condition assertion */
1703        ecode += 1 + LINK_SIZE;                /* Continue with current branch */
     ecode += 1 + LINK_SIZE;  
1704      continue;      continue;
1705    
1706      /* Move the subject pointer back. This occurs only at the start of      /* Move the subject pointer back. This occurs only at the start of
# Line 1593  for (;;) Line 1709  for (;;)
1709      back a number of characters, not bytes. */      back a number of characters, not bytes. */
1710    
1711      case OP_REVERSE:      case OP_REVERSE:
1712  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
1713      if (utf8)      if (utf)
1714        {        {
1715        i = GET(ecode, 1);        i = GET(ecode, 1);
1716        while (i-- > 0)        while (i-- > 0)
# Line 1625  for (;;) Line 1741  for (;;)
1741      function is able to force a failure. */      function is able to force a failure. */
1742    
1743      case OP_CALLOUT:      case OP_CALLOUT:
1744      if (pcre_callout != NULL)      if (PUBL(callout) != NULL)
1745        {        {
1746        pcre_callout_block cb;        PUBL(callout_block) cb;
1747        cb.version          = 2;   /* Version 1 of the callout block */        cb.version          = 2;   /* Version 1 of the callout block */
1748        cb.callout_number   = ecode[1];        cb.callout_number   = ecode[1];
1749        cb.offset_vector    = md->offset_vector;        cb.offset_vector    = md->offset_vector;
1750    #if defined COMPILE_PCRE8
1751        cb.subject          = (PCRE_SPTR)md->start_subject;        cb.subject          = (PCRE_SPTR)md->start_subject;
1752    #elif defined COMPILE_PCRE16
1753          cb.subject          = (PCRE_SPTR16)md->start_subject;
1754    #elif defined COMPILE_PCRE32
1755          cb.subject          = (PCRE_SPTR32)md->start_subject;
1756    #endif
1757        cb.subject_length   = (int)(md->end_subject - md->start_subject);        cb.subject_length   = (int)(md->end_subject - md->start_subject);
1758        cb.start_match      = (int)(mstart - md->start_subject);        cb.start_match      = (int)(mstart - md->start_subject);
1759        cb.current_position = (int)(eptr - md->start_subject);        cb.current_position = (int)(eptr - md->start_subject);
1760        cb.pattern_position = GET(ecode, 2);        cb.pattern_position = GET(ecode, 2);
1761        cb.next_item_length = GET(ecode, 2 + LINK_SIZE);        cb.next_item_length = GET(ecode, 2 + LINK_SIZE);
1762        cb.capture_top      = offset_top/2;        cb.capture_top      = offset_top/2;
1763        cb.capture_last     = md->capture_last;        cb.capture_last     = md->capture_last & CAPLMASK;
1764          /* Internal change requires this for API compatibility. */
1765          if (cb.capture_last == 0) cb.capture_last = -1;
1766        cb.callout_data     = md->callout_data;        cb.callout_data     = md->callout_data;
1767        cb.mark             = md->nomatch_mark;        cb.mark             = md->nomatch_mark;
1768        if ((rrc = (*pcre_callout)(&cb)) > 0) RRETURN(MATCH_NOMATCH);        if ((rrc = (*PUBL(callout))(&cb)) > 0) RRETURN(MATCH_NOMATCH);
1769        if (rrc < 0) RRETURN(rrc);        if (rrc < 0) RRETURN(rrc);
1770        }        }
1771      ecode += 2 + 2*LINK_SIZE;      ecode += 2 + 2*LINK_SIZE;
# Line 1667  for (;;) Line 1791  for (;;)
1791      case OP_RECURSE:      case OP_RECURSE:
1792        {        {
1793        recursion_info *ri;        recursion_info *ri;
1794        int recno;        unsigned int recno;
1795    
1796        callpat = md->start_code + GET(ecode, 1);        callpat = md->start_code + GET(ecode, 1);
1797        recno = (callpat == md->start_code)? 0 :        recno = (callpat == md->start_code)? 0 :
# Line 1684  for (;;) Line 1808  for (;;)
1808        /* Add to "recursing stack" */        /* Add to "recursing stack" */
1809    
1810        new_recursive.group_num = recno;        new_recursive.group_num = recno;
1811          new_recursive.saved_capture_last = md->capture_last;
1812        new_recursive.subject_position = eptr;        new_recursive.subject_position = eptr;
1813        new_recursive.prevrec = md->recursive;        new_recursive.prevrec = md->recursive;
1814        md->recursive = &new_recursive;        md->recursive = &new_recursive;
# Line 1700  for (;;) Line 1825  for (;;)
1825        else        else
1826          {          {
1827          new_recursive.offset_save =          new_recursive.offset_save =
1828            (int *)(pcre_malloc)(new_recursive.saved_max * sizeof(int));            (int *)(PUBL(malloc))(new_recursive.saved_max * sizeof(int));
1829          if (new_recursive.offset_save == NULL) RRETURN(PCRE_ERROR_NOMEMORY);          if (new_recursive.offset_save == NULL) RRETURN(PCRE_ERROR_NOMEMORY);
1830          }          }
1831        memcpy(new_recursive.offset_save, md->offset_vector,        memcpy(new_recursive.offset_save, md->offset_vector,
1832              new_recursive.saved_max * sizeof(int));              new_recursive.saved_max * sizeof(int));
1833    
1834        /* OK, now we can do the recursion. After processing each alternative,        /* OK, now we can do the recursion. After processing each alternative,
1835        restore the offset data. If there were nested recursions, md->recursive        restore the offset data and the last captured value. If there were nested
1836        might be changed, so reset it before looping. */        recursions, md->recursive might be changed, so reset it before looping.
1837          */
1838    
1839        DPRINTF(("Recursing into group %d\n", new_recursive.group_num));        DPRINTF(("Recursing into group %d\n", new_recursive.group_num));
1840        cbegroup = (*callpat >= OP_SBRA);        cbegroup = (*callpat >= OP_SBRA);
1841        do        do
1842          {          {
1843          if (cbegroup) md->match_function_type = MATCH_CBEGROUP;          if (cbegroup) md->match_function_type = MATCH_CBEGROUP;
1844          RMATCH(eptr, callpat + _pcre_OP_lengths[*callpat], offset_top,          RMATCH(eptr, callpat + PRIV(OP_lengths)[*callpat], offset_top,
1845            md, eptrb, RM6);            md, eptrb, RM6);
1846          memcpy(md->offset_vector, new_recursive.offset_save,          memcpy(md->offset_vector, new_recursive.offset_save,
1847              new_recursive.saved_max * sizeof(int));              new_recursive.saved_max * sizeof(int));
1848            md->capture_last = new_recursive.saved_capture_last;
1849          md->recursive = new_recursive.prevrec;          md->recursive = new_recursive.prevrec;
1850          if (rrc == MATCH_MATCH || rrc == MATCH_ACCEPT)          if (rrc == MATCH_MATCH || rrc == MATCH_ACCEPT)
1851            {            {
1852            DPRINTF(("Recursion matched\n"));            DPRINTF(("Recursion matched\n"));
1853            if (new_recursive.offset_save != stacksave)            if (new_recursive.offset_save != stacksave)
1854              (pcre_free)(new_recursive.offset_save);              (PUBL(free))(new_recursive.offset_save);
1855    
1856            /* Set where we got to in the subject, and reset the start in case            /* Set where we got to in the subject, and reset the start in case
1857            it was changed by \K. This *is* propagated back out of a recursion,            it was changed by \K. This *is* propagated back out of a recursion,
# Line 1735  for (;;) Line 1862  for (;;)
1862            goto RECURSION_MATCHED;        /* Exit loop; end processing */            goto RECURSION_MATCHED;        /* Exit loop; end processing */
1863            }            }
1864    
1865          /* PCRE does not allow THEN to escape beyond a recursion; it is treated          /* PCRE does not allow THEN, SKIP, PRUNE or COMMIT to escape beyond a
1866          as NOMATCH. */          recursion; they are treated as NOMATCH. These codes are defined in a
1867            range that can be tested for. Any other return code is an error. */
1868    
1869          else if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN)          else if (rrc != MATCH_NOMATCH &&
1870                     (rrc < MATCH_BACKTRACK_MIN || rrc > MATCH_BACKTRACK_MAX))
1871            {            {
1872            DPRINTF(("Recursion gave error %d\n", rrc));            DPRINTF(("Recursion gave error %d\n", rrc));
1873            if (new_recursive.offset_save != stacksave)            if (new_recursive.offset_save != stacksave)
1874              (pcre_free)(new_recursive.offset_save);              (PUBL(free))(new_recursive.offset_save);
1875            RRETURN(rrc);            RRETURN(rrc);
1876            }            }
1877    
# Line 1754  for (;;) Line 1883  for (;;)
1883        DPRINTF(("Recursion didn't match\n"));        DPRINTF(("Recursion didn't match\n"));
1884        md->recursive = new_recursive.prevrec;        md->recursive = new_recursive.prevrec;
1885        if (new_recursive.offset_save != stacksave)        if (new_recursive.offset_save != stacksave)
1886          (pcre_free)(new_recursive.offset_save);          (PUBL(free))(new_recursive.offset_save);
1887        RRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
1888        }        }
1889    
# Line 1868  for (;;) Line 1997  for (;;)
1997    
1998        /* Deal with capturing */        /* Deal with capturing */
1999    
2000        md->capture_last = number;        md->capture_last = (md->capture_last & OVFLMASK) | number;
2001        if (offset >= md->offset_max) md->offset_overflow = TRUE; else        if (offset >= md->offset_max) md->capture_last |= OVFLBIT; else
2002          {          {
2003          /* If offset is greater than offset_top, it means that we are          /* If offset is greater than offset_top, it means that we are
2004          "skipping" a capturing group, and that group's offsets must be marked          "skipping" a capturing group, and that group's offsets must be marked
# Line 1949  for (;;) Line 2078  for (;;)
2078          }          }
2079        if (*prev >= OP_SBRA)    /* Could match an empty string */        if (*prev >= OP_SBRA)    /* Could match an empty string */
2080          {          {
         md->match_function_type = MATCH_CBEGROUP;  
2081          RMATCH(eptr, prev, offset_top, md, eptrb, RM50);          RMATCH(eptr, prev, offset_top, md, eptrb, RM50);
2082          RRETURN(rrc);          RRETURN(rrc);
2083          }          }
# Line 1958  for (;;) Line 2086  for (;;)
2086        }        }
2087      else  /* OP_KETRMAX */      else  /* OP_KETRMAX */
2088        {        {
       if (*prev >= OP_SBRA) md->match_function_type = MATCH_CBEGROUP;  
2089        RMATCH(eptr, prev, offset_top, md, eptrb, RM13);        RMATCH(eptr, prev, offset_top, md, eptrb, RM13);
2090        if (rrc == MATCH_ONCE && md->once_target == prev) rrc = MATCH_NOMATCH;        if (rrc == MATCH_ONCE && md->once_target == prev) rrc = MATCH_NOMATCH;
2091        if (rrc != MATCH_NOMATCH) RRETURN(rrc);        if (rrc != MATCH_NOMATCH) RRETURN(rrc);
# Line 2015  for (;;) Line 2142  for (;;)
2142    
2143      case OP_DOLLM:      case OP_DOLLM:
2144      if (eptr < md->end_subject)      if (eptr < md->end_subject)
2145        { if (!IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH); }        {
2146          if (!IS_NEWLINE(eptr))
2147            {
2148            if (md->partial != 0 &&
2149                eptr + 1 >= md->end_subject &&
2150                NLBLOCK->nltype == NLTYPE_FIXED &&
2151                NLBLOCK->nllen == 2 &&
2152                RAWUCHARTEST(eptr) == NLBLOCK->nl[0])
2153              {
2154              md->hitend = TRUE;
2155              if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
2156              }
2157            RRETURN(MATCH_NOMATCH);
2158            }
2159          }
2160      else      else
2161        {        {
2162        if (md->noteol) RRETURN(MATCH_NOMATCH);        if (md->noteol) RRETURN(MATCH_NOMATCH);
# Line 2047  for (;;) Line 2188  for (;;)
2188      ASSERT_NL_OR_EOS:      ASSERT_NL_OR_EOS:
2189      if (eptr < md->end_subject &&      if (eptr < md->end_subject &&
2190          (!IS_NEWLINE(eptr) || eptr != md->end_subject - md->nllen))          (!IS_NEWLINE(eptr) || eptr != md->end_subject - md->nllen))
2191          {
2192          if (md->partial != 0 &&
2193              eptr + 1 >= md->end_subject &&
2194              NLBLOCK->nltype == NLTYPE_FIXED &&
2195              NLBLOCK->nllen == 2 &&
2196              RAWUCHARTEST(eptr) == NLBLOCK->nl[0])
2197            {
2198            md->hitend = TRUE;
2199            if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
2200            }
2201        RRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2202          }
2203    
2204      /* Either at end of string or \n before end. */      /* Either at end of string or \n before end. */
2205    
# Line 2066  for (;;) Line 2218  for (;;)
2218        be "non-word" characters. Remember the earliest consulted character for        be "non-word" characters. Remember the earliest consulted character for
2219        partial matching. */        partial matching. */
2220    
2221  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
2222        if (utf8)        if (utf)
2223          {          {
2224          /* Get status of previous character */          /* Get status of previous character */
2225    
2226          if (eptr == md->start_subject) prev_is_word = FALSE; else          if (eptr == md->start_subject) prev_is_word = FALSE; else
2227            {            {
2228            USPTR lastptr = eptr - 1;            PCRE_PUCHAR lastptr = eptr - 1;
2229            while((*lastptr & 0xc0) == 0x80) lastptr--;            BACKCHAR(lastptr);
2230            if (lastptr < md->start_used_ptr) md->start_used_ptr = lastptr;            if (lastptr < md->start_used_ptr) md->start_used_ptr = lastptr;
2231            GETCHAR(c, lastptr);            GETCHAR(c, lastptr);
2232  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
# Line 2139  for (;;) Line 2291  for (;;)
2291              }              }
2292            else            else
2293  #endif  #endif
2294            prev_is_word = ((md->ctypes[eptr[-1]] & ctype_word) != 0);            prev_is_word = MAX_255(eptr[-1])
2295                && ((md->ctypes[eptr[-1]] & ctype_word) != 0);
2296            }            }
2297    
2298          /* Get status of next character */          /* Get status of next character */
# Line 2162  for (;;) Line 2315  for (;;)
2315            }            }
2316          else          else
2317  #endif  #endif
2318          cur_is_word = ((md->ctypes[*eptr] & ctype_word) != 0);          cur_is_word = MAX_255(*eptr)
2319              && ((md->ctypes[*eptr] & ctype_word) != 0);
2320          }          }
2321    
2322        /* Now see if the situation is what we want */        /* Now see if the situation is what we want */
# Line 2173  for (;;) Line 2327  for (;;)
2327        }        }
2328      break;      break;
2329    
2330      /* Match a single character type; inline for speed */      /* Match any single character type except newline; have to take care with
2331        CRLF newlines and partial matching. */
2332    
2333      case OP_ANY:      case OP_ANY:
2334      if (IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH);      if (IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH);
2335        if (md->partial != 0 &&
2336            eptr + 1 >= md->end_subject &&
2337            NLBLOCK->nltype == NLTYPE_FIXED &&
2338            NLBLOCK->nllen == 2 &&
2339            RAWUCHARTEST(eptr) == NLBLOCK->nl[0])
2340          {
2341          md->hitend = TRUE;
2342          if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
2343          }
2344    
2345      /* Fall through */      /* Fall through */
2346    
2347        /* Match any single character whatsoever. */
2348    
2349      case OP_ALLANY:      case OP_ALLANY:
2350      if (eptr >= md->end_subject)   /* DO NOT merge the eptr++ here; it must */      if (eptr >= md->end_subject)   /* DO NOT merge the eptr++ here; it must */
2351        {                            /* not be updated before SCHECK_PARTIAL. */        {                            /* not be updated before SCHECK_PARTIAL. */
# Line 2186  for (;;) Line 2353  for (;;)
2353        RRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2354        }        }
2355      eptr++;      eptr++;
2356      if (utf8) while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;  #ifdef SUPPORT_UTF
2357        if (utf) ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);
2358    #endif
2359      ecode++;      ecode++;
2360      break;      break;
2361    
# Line 2211  for (;;) Line 2380  for (;;)
2380        }        }
2381      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2382      if (      if (
2383  #ifdef SUPPORT_UTF8  #if defined SUPPORT_UTF || !(defined COMPILE_PCRE8)
2384         c < 256 &&         c < 256 &&
2385  #endif  #endif
2386         (md->ctypes[c] & ctype_digit) != 0         (md->ctypes[c] & ctype_digit) != 0
# Line 2228  for (;;) Line 2397  for (;;)
2397        }        }
2398      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2399      if (      if (
2400  #ifdef SUPPORT_UTF8  #if defined SUPPORT_UTF || !(defined COMPILE_PCRE8)
2401         c >= 256 ||         c > 255 ||
2402  #endif  #endif
2403         (md->ctypes[c] & ctype_digit) == 0         (md->ctypes[c] & ctype_digit) == 0
2404         )         )
# Line 2245  for (;;) Line 2414  for (;;)
2414        }        }
2415      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2416      if (      if (
2417  #ifdef SUPPORT_UTF8  #if defined SUPPORT_UTF || !(defined COMPILE_PCRE8)
2418         c < 256 &&         c < 256 &&
2419  #endif  #endif
2420         (md->ctypes[c] & ctype_space) != 0         (md->ctypes[c] & ctype_space) != 0
# Line 2262  for (;;) Line 2431  for (;;)
2431        }        }
2432      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2433      if (      if (
2434  #ifdef SUPPORT_UTF8  #if defined SUPPORT_UTF || !(defined COMPILE_PCRE8)
2435         c >= 256 ||         c > 255 ||
2436  #endif  #endif
2437         (md->ctypes[c] & ctype_space) == 0         (md->ctypes[c] & ctype_space) == 0
2438         )         )
# Line 2279  for (;;) Line 2448  for (;;)
2448        }        }
2449      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2450      if (      if (
2451  #ifdef SUPPORT_UTF8  #if defined SUPPORT_UTF || !(defined COMPILE_PCRE8)
2452         c < 256 &&         c < 256 &&
2453  #endif  #endif
2454         (md->ctypes[c] & ctype_word) != 0         (md->ctypes[c] & ctype_word) != 0
# Line 2296  for (;;) Line 2465  for (;;)
2465        }        }
2466      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2467      if (      if (
2468  #ifdef SUPPORT_UTF8  #if defined SUPPORT_UTF || !(defined COMPILE_PCRE8)
2469         c >= 256 ||         c > 255 ||
2470  #endif  #endif
2471         (md->ctypes[c] & ctype_word) == 0         (md->ctypes[c] & ctype_word) == 0
2472         )         )
# Line 2316  for (;;) Line 2485  for (;;)
2485        {        {
2486        default: RRETURN(MATCH_NOMATCH);        default: RRETURN(MATCH_NOMATCH);
2487    
2488        case 0x000d:        case CHAR_CR:
2489        if (eptr < md->end_subject && *eptr == 0x0a) eptr++;        if (eptr >= md->end_subject)
2490            {
2491            SCHECK_PARTIAL();
2492            }
2493          else if (RAWUCHARTEST(eptr) == CHAR_LF) eptr++;
2494        break;        break;
2495    
2496        case 0x000a:        case CHAR_LF:
2497        break;        break;
2498    
2499        case 0x000b:        case CHAR_VT:
2500        case 0x000c:        case CHAR_FF:
2501        case 0x0085:        case CHAR_NEL:
2502    #ifndef EBCDIC
2503        case 0x2028:        case 0x2028:
2504        case 0x2029:        case 0x2029:
2505    #endif  /* Not EBCDIC */
2506        if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);        if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);
2507        break;        break;
2508        }        }
# Line 2343  for (;;) Line 2518  for (;;)
2518      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2519      switch(c)      switch(c)
2520        {        {
2521          HSPACE_CASES: RRETURN(MATCH_NOMATCH);  /* Byte and multibyte cases */
2522        default: break;        default: break;
       case 0x09:      /* HT */  
       case 0x20:      /* SPACE */  
       case 0xa0:      /* NBSP */  
       case 0x1680:    /* OGHAM SPACE MARK */  
       case 0x180e:    /* MONGOLIAN VOWEL SEPARATOR */  
       case 0x2000:    /* EN QUAD */  
       case 0x2001:    /* EM QUAD */  
       case 0x2002:    /* EN SPACE */  
       case 0x2003:    /* EM SPACE */  
       case 0x2004:    /* THREE-PER-EM SPACE */  
       case 0x2005:    /* FOUR-PER-EM SPACE */  
       case 0x2006:    /* SIX-PER-EM SPACE */  
       case 0x2007:    /* FIGURE SPACE */  
       case 0x2008:    /* PUNCTUATION SPACE */  
       case 0x2009:    /* THIN SPACE */  
       case 0x200A:    /* HAIR SPACE */  
       case 0x202f:    /* NARROW NO-BREAK SPACE */  
       case 0x205f:    /* MEDIUM MATHEMATICAL SPACE */  
       case 0x3000:    /* IDEOGRAPHIC SPACE */  
       RRETURN(MATCH_NOMATCH);  
2523        }        }
2524      ecode++;      ecode++;
2525      break;      break;
# Line 2377  for (;;) Line 2533  for (;;)
2533      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2534      switch(c)      switch(c)
2535        {        {
2536          HSPACE_CASES: break;  /* Byte and multibyte cases */
2537        default: RRETURN(MATCH_NOMATCH);        default: RRETURN(MATCH_NOMATCH);
       case 0x09:      /* HT */  
       case 0x20:      /* SPACE */  
       case 0xa0:      /* NBSP */  
       case 0x1680:    /* OGHAM SPACE MARK */  
       case 0x180e:    /* MONGOLIAN VOWEL SEPARATOR */  
       case 0x2000:    /* EN QUAD */  
       case 0x2001:    /* EM QUAD */  
       case 0x2002:    /* EN SPACE */  
       case 0x2003:    /* EM SPACE */  
       case 0x2004:    /* THREE-PER-EM SPACE */  
       case 0x2005:    /* FOUR-PER-EM SPACE */  
       case 0x2006:    /* SIX-PER-EM SPACE */  
       case 0x2007:    /* FIGURE SPACE */  
       case 0x2008:    /* PUNCTUATION SPACE */  
       case 0x2009:    /* THIN SPACE */  
       case 0x200A:    /* HAIR SPACE */  
       case 0x202f:    /* NARROW NO-BREAK SPACE */  
       case 0x205f:    /* MEDIUM MATHEMATICAL SPACE */  
       case 0x3000:    /* IDEOGRAPHIC SPACE */  
       break;  
2538        }        }
2539      ecode++;      ecode++;
2540      break;      break;
# Line 2411  for (;;) Line 2548  for (;;)
2548      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2549      switch(c)      switch(c)
2550        {        {
2551          VSPACE_CASES: RRETURN(MATCH_NOMATCH);
2552        default: break;        default: break;
       case 0x0a:      /* LF */  
       case 0x0b:      /* VT */  
       case 0x0c:      /* FF */  
       case 0x0d:      /* CR */  
       case 0x85:      /* NEL */  
       case 0x2028:    /* LINE SEPARATOR */  
       case 0x2029:    /* PARAGRAPH SEPARATOR */  
       RRETURN(MATCH_NOMATCH);  
2553        }        }
2554      ecode++;      ecode++;
2555      break;      break;
# Line 2433  for (;;) Line 2563  for (;;)
2563      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2564      switch(c)      switch(c)
2565        {        {
2566          VSPACE_CASES: break;
2567        default: RRETURN(MATCH_NOMATCH);        default: RRETURN(MATCH_NOMATCH);
       case 0x0a:      /* LF */  
       case 0x0b:      /* VT */  
       case 0x0c:      /* FF */  
       case 0x0d:      /* CR */  
       case 0x85:      /* NEL */  
       case 0x2028:    /* LINE SEPARATOR */  
       case 0x2029:    /* PARAGRAPH SEPARATOR */  
       break;  
2568        }        }
2569      ecode++;      ecode++;
2570      break;      break;
# Line 2459  for (;;) Line 2582  for (;;)
2582        }        }
2583      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2584        {        {
2585          const pcre_uint32 *cp;
2586        const ucd_record *prop = GET_UCD(c);        const ucd_record *prop = GET_UCD(c);
2587    
2588        switch(ecode[1])        switch(ecode[1])
# Line 2475  for (;;) Line 2599  for (;;)
2599          break;          break;
2600    
2601          case PT_GC:          case PT_GC:
2602          if ((ecode[2] != _pcre_ucp_gentype[prop->chartype]) == (op == OP_PROP))          if ((ecode[2] != PRIV(ucp_gentype)[prop->chartype]) == (op == OP_PROP))
2603            RRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
2604          break;          break;
2605    
# Line 2492  for (;;) Line 2616  for (;;)
2616          /* These are specials */          /* These are specials */
2617    
2618          case PT_ALNUM:          case PT_ALNUM:
2619          if ((_pcre_ucp_gentype[prop->chartype] == ucp_L ||          if ((PRIV(ucp_gentype)[prop->chartype] == ucp_L ||
2620               _pcre_ucp_gentype[prop->chartype] == ucp_N) == (op == OP_NOTPROP))               PRIV(ucp_gentype)[prop->chartype] == ucp_N) == (op == OP_NOTPROP))
2621            RRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
2622          break;          break;
2623    
2624          case PT_SPACE:    /* Perl space */          case PT_SPACE:    /* Perl space */
2625          if ((_pcre_ucp_gentype[prop->chartype] == ucp_Z ||          if ((PRIV(ucp_gentype)[prop->chartype] == ucp_Z ||
2626               c == CHAR_HT || c == CHAR_NL || c == CHAR_FF || c == CHAR_CR)               c == CHAR_HT || c == CHAR_NL || c == CHAR_FF || c == CHAR_CR)
2627                 == (op == OP_NOTPROP))                 == (op == OP_NOTPROP))
2628            RRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
2629          break;          break;
2630    
2631          case PT_PXSPACE:  /* POSIX space */          case PT_PXSPACE:  /* POSIX space */
2632          if ((_pcre_ucp_gentype[prop->chartype] == ucp_Z ||          if ((PRIV(ucp_gentype)[prop->chartype] == ucp_Z ||
2633               c == CHAR_HT || c == CHAR_NL || c == CHAR_VT ||               c == CHAR_HT || c == CHAR_NL || c == CHAR_VT ||
2634               c == CHAR_FF || c == CHAR_CR)               c == CHAR_FF || c == CHAR_CR)
2635                 == (op == OP_NOTPROP))                 == (op == OP_NOTPROP))
# Line 2513  for (;;) Line 2637  for (;;)
2637          break;          break;
2638    
2639          case PT_WORD:          case PT_WORD:
2640          if ((_pcre_ucp_gentype[prop->chartype] == ucp_L ||          if ((PRIV(ucp_gentype)[prop->chartype] == ucp_L ||
2641               _pcre_ucp_gentype[prop->chartype] == ucp_N ||               PRIV(ucp_gentype)[prop->chartype] == ucp_N ||
2642               c == CHAR_UNDERSCORE) == (op == OP_NOTPROP))               c == CHAR_UNDERSCORE) == (op == OP_NOTPROP))
2643            RRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
2644          break;          break;
2645    
2646            case PT_CLIST:
2647            cp = PRIV(ucd_caseless_sets) + ecode[2];
2648            for (;;)
2649              {
2650              if (c < *cp)
2651                { if (op == OP_PROP) { RRETURN(MATCH_NOMATCH); } else break; }
2652              if (c == *cp++)
2653                { if (op == OP_PROP) break; else { RRETURN(MATCH_NOMATCH); } }
2654              }
2655            break;
2656    
2657            case PT_UCNC:
2658            if ((c == CHAR_DOLLAR_SIGN || c == CHAR_COMMERCIAL_AT ||
2659                 c == CHAR_GRAVE_ACCENT || (c >= 0xa0 && c <= 0xd7ff) ||
2660                 c >= 0xe000) == (op == OP_NOTPROP))
2661              RRETURN(MATCH_NOMATCH);
2662            break;
2663    
2664          /* This should never occur */          /* This should never occur */
2665    
2666          default:          default:
# Line 2538  for (;;) Line 2680  for (;;)
2680        SCHECK_PARTIAL();        SCHECK_PARTIAL();
2681        RRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2682        }        }
2683      GETCHARINCTEST(c, eptr);      else
     if (UCD_CATEGORY(c) == ucp_M) RRETURN(MATCH_NOMATCH);  
     while (eptr < md->end_subject)  
2684        {        {
2685        int len = 1;        int lgb, rgb;
2686        if (!utf8) c = *eptr; else { GETCHARLEN(c, eptr, len); }        GETCHARINCTEST(c, eptr);
2687        if (UCD_CATEGORY(c) != ucp_M) break;        lgb = UCD_GRAPHBREAK(c);
2688        eptr += len;        while (eptr < md->end_subject)
2689            {
2690            int len = 1;
2691            if (!utf) c = *eptr; else { GETCHARLEN(c, eptr, len); }
2692            rgb = UCD_GRAPHBREAK(c);
2693            if ((PRIV(ucp_gbtable)[lgb] & (1 << rgb)) == 0) break;
2694            lgb = rgb;
2695            eptr += len;
2696            }
2697        }        }
2698        CHECK_PARTIAL();
2699      ecode++;      ecode++;
2700      break;      break;
2701  #endif  #endif  /* SUPPORT_UCP */
2702    
2703    
2704      /* Match a back reference, possibly repeatedly. Look past the end of the      /* Match a back reference, possibly repeatedly. Look past the end of the
# Line 2564  for (;;) Line 2713  for (;;)
2713      case OP_REFI:      case OP_REFI:
2714      caseless = op == OP_REFI;      caseless = op == OP_REFI;
2715      offset = GET2(ecode, 1) << 1;               /* Doubled ref number */      offset = GET2(ecode, 1) << 1;               /* Doubled ref number */
2716      ecode += 3;      ecode += 1 + IMM2_SIZE;
2717    
2718      /* If the reference is unset, there are two possibilities:      /* If the reference is unset, there are two possibilities:
2719    
# Line 2604  for (;;) Line 2753  for (;;)
2753        case OP_CRMINRANGE:        case OP_CRMINRANGE:
2754        minimize = (*ecode == OP_CRMINRANGE);        minimize = (*ecode == OP_CRMINRANGE);
2755        min = GET2(ecode, 1);        min = GET2(ecode, 1);
2756        max = GET2(ecode, 3);        max = GET2(ecode, 1 + IMM2_SIZE);
2757        if (max == 0) max = INT_MAX;        if (max == 0) max = INT_MAX;
2758        ecode += 5;        ecode += 1 + 2 * IMM2_SIZE;
2759        break;        break;
2760    
2761        default:               /* No repeat follows */        default:               /* No repeat follows */
2762        if ((length = match_ref(offset, eptr, length, md, caseless)) < 0)        if ((length = match_ref(offset, eptr, length, md, caseless)) < 0)
2763          {          {
2764            if (length == -2) eptr = md->end_subject;   /* Partial match */
2765          CHECK_PARTIAL();          CHECK_PARTIAL();
2766          RRETURN(MATCH_NOMATCH);          RRETURN(MATCH_NOMATCH);
2767          }          }
# Line 2620  for (;;) Line 2770  for (;;)
2770        }        }
2771    
2772      /* Handle repeated back references. If the length of the reference is      /* Handle repeated back references. If the length of the reference is
2773      zero, just continue with the main loop. */      zero, just continue with the main loop. If the length is negative, it
2774        means the reference is unset in non-Java-compatible mode. If the minimum is
2775        zero, we can continue at the same level without recursion. For any other
2776        minimum, carrying on will result in NOMATCH. */
2777    
2778      if (length == 0) continue;      if (length == 0) continue;
2779        if (length < 0 && min == 0) continue;
2780    
2781      /* First, ensure the minimum number of matches are present. We get back      /* First, ensure the minimum number of matches are present. We get back
2782      the length of the reference string explicitly rather than passing the      the length of the reference string explicitly rather than passing the
# Line 2633  for (;;) Line 2787  for (;;)
2787        int slength;        int slength;
2788        if ((slength = match_ref(offset, eptr, length, md, caseless)) < 0)        if ((slength = match_ref(offset, eptr, length, md, caseless)) < 0)
2789          {          {
2790            if (slength == -2) eptr = md->end_subject;   /* Partial match */
2791          CHECK_PARTIAL();          CHECK_PARTIAL();
2792          RRETURN(MATCH_NOMATCH);          RRETURN(MATCH_NOMATCH);
2793          }          }
# Line 2656  for (;;) Line 2811  for (;;)
2811          if (fi >= max) RRETURN(MATCH_NOMATCH);          if (fi >= max) RRETURN(MATCH_NOMATCH);
2812          if ((slength = match_ref(offset, eptr, length, md, caseless)) < 0)          if ((slength = match_ref(offset, eptr, length, md, caseless)) < 0)
2813            {            {
2814              if (slength == -2) eptr = md->end_subject;   /* Partial match */
2815            CHECK_PARTIAL();            CHECK_PARTIAL();
2816            RRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
2817            }            }
# Line 2674  for (;;) Line 2830  for (;;)
2830          int slength;          int slength;
2831          if ((slength = match_ref(offset, eptr, length, md, caseless)) < 0)          if ((slength = match_ref(offset, eptr, length, md, caseless)) < 0)
2832            {            {
2833            CHECK_PARTIAL();            /* Can't use CHECK_PARTIAL because we don't want to update eptr in
2834              the soft partial matching case. */
2835    
2836              if (slength == -2 && md->partial != 0 &&
2837                  md->end_subject > md->start_used_ptr)
2838                {
2839                md->hitend = TRUE;
2840                if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
2841                }
2842            break;            break;
2843            }            }
2844          eptr += slength;          eptr += slength;
2845          }          }
2846    
2847        while (eptr >= pp)        while (eptr >= pp)
2848          {          {
2849          RMATCH(eptr, ecode, offset_top, md, eptrb, RM15);          RMATCH(eptr, ecode, offset_top, md, eptrb, RM15);
# Line 2703  for (;;) Line 2868  for (;;)
2868      case OP_NCLASS:      case OP_NCLASS:
2869      case OP_CLASS:      case OP_CLASS:
2870        {        {
2871          /* The data variable is saved across frames, so the byte map needs to
2872          be stored there. */
2873    #define BYTE_MAP ((pcre_uint8 *)data)
2874        data = ecode + 1;                /* Save for matching */        data = ecode + 1;                /* Save for matching */
2875        ecode += 33;                     /* Advance past the item */        ecode += 1 + (32 / sizeof(pcre_uchar)); /* Advance past the item */
2876    
2877        switch (*ecode)        switch (*ecode)
2878          {          {
# Line 2725  for (;;) Line 2893  for (;;)
2893          case OP_CRMINRANGE:          case OP_CRMINRANGE:
2894          minimize = (*ecode == OP_CRMINRANGE);          minimize = (*ecode == OP_CRMINRANGE);
2895          min = GET2(ecode, 1);          min = GET2(ecode, 1);
2896          max = GET2(ecode, 3);          max = GET2(ecode, 1 + IMM2_SIZE);
2897          if (max == 0) max = INT_MAX;          if (max == 0) max = INT_MAX;
2898          ecode += 5;          ecode += 1 + 2 * IMM2_SIZE;
2899          break;          break;
2900    
2901          default:               /* No repeat follows */          default:               /* No repeat follows */
# Line 2737  for (;;) Line 2905  for (;;)
2905    
2906        /* First, ensure the minimum number of matches are present. */        /* First, ensure the minimum number of matches are present. */
2907    
2908  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
2909        /* UTF-8 mode */        if (utf)
       if (utf8)  
2910          {          {
2911          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
2912            {            {
# Line 2754  for (;;) Line 2921  for (;;)
2921              if (op == OP_CLASS) RRETURN(MATCH_NOMATCH);              if (op == OP_CLASS) RRETURN(MATCH_NOMATCH);
2922              }              }
2923            else            else
2924              {              if ((BYTE_MAP[c/8] & (1 << (c&7))) == 0) RRETURN(MATCH_NOMATCH);
             if ((data[c/8] & (1 << (c&7))) == 0) RRETURN(MATCH_NOMATCH);  
             }  
2925            }            }
2926          }          }
2927        else        else
2928  #endif  #endif
2929        /* Not UTF-8 mode */        /* Not UTF mode */
2930          {          {
2931          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
2932            {            {
# Line 2771  for (;;) Line 2936  for (;;)
2936              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
2937              }              }
2938            c = *eptr++;            c = *eptr++;
2939            if ((data[c/8] & (1 << (c&7))) == 0) RRETURN(MATCH_NOMATCH);  #ifndef COMPILE_PCRE8
2940              if (c > 255)
2941                {
2942                if (op == OP_CLASS) RRETURN(MATCH_NOMATCH);
2943                }
2944              else
2945    #endif
2946                if ((BYTE_MAP[c/8] & (1 << (c&7))) == 0) RRETURN(MATCH_NOMATCH);
2947            }            }
2948          }          }
2949    
# Line 2785  for (;;) Line 2957  for (;;)
2957    
2958        if (minimize)        if (minimize)
2959          {          {
2960  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
2961          /* UTF-8 mode */          if (utf)
         if (utf8)  
2962            {            {
2963            for (fi = min;; fi++)            for (fi = min;; fi++)
2964              {              {
# Line 2805  for (;;) Line 2976  for (;;)
2976                if (op == OP_CLASS) RRETURN(MATCH_NOMATCH);                if (op == OP_CLASS) RRETURN(MATCH_NOMATCH);
2977                }                }
2978              else              else
2979                {                if ((BYTE_MAP[c/8] & (1 << (c&7))) == 0) RRETURN(MATCH_NOMATCH);
               if ((data[c/8] & (1 << (c&7))) == 0) RRETURN(MATCH_NOMATCH);  
               }  
2980              }              }
2981            }            }
2982          else          else
2983  #endif  #endif
2984          /* Not UTF-8 mode */          /* Not UTF mode */
2985            {            {
2986            for (fi = min;; fi++)            for (fi = min;; fi++)
2987              {              {
# Line 2825  for (;;) Line 2994  for (;;)
2994                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
2995                }                }
2996              c = *eptr++;              c = *eptr++;
2997              if ((data[c/8] & (1 << (c&7))) == 0) RRETURN(MATCH_NOMATCH);  #ifndef COMPILE_PCRE8
2998                if (c > 255)
2999                  {
3000                  if (op == OP_CLASS) RRETURN(MATCH_NOMATCH);
3001                  }
3002                else
3003    #endif
3004                  if ((BYTE_MAP[c/8] & (1 << (c&7))) == 0) RRETURN(MATCH_NOMATCH);
3005              }              }
3006            }            }
3007          /* Control never gets here */          /* Control never gets here */
# Line 2837  for (;;) Line 3013  for (;;)
3013          {          {
3014          pp = eptr;          pp = eptr;
3015    
3016  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
3017          /* UTF-8 mode */          if (utf)
         if (utf8)  
3018            {            {
3019            for (i = min; i < max; i++)            for (i = min; i < max; i++)
3020              {              {
# Line 2855  for (;;) Line 3030  for (;;)
3030                if (op == OP_CLASS) break;                if (op == OP_CLASS) break;
3031                }                }
3032              else              else
3033                {                if ((BYTE_MAP[c/8] & (1 << (c&7))) == 0) break;
               if ((data[c/8] & (1 << (c&7))) == 0) break;  
               }  
3034              eptr += len;              eptr += len;
3035              }              }
3036            for (;;)            for (;;)
# Line 2870  for (;;) Line 3043  for (;;)
3043            }            }
3044          else          else
3045  #endif  #endif
3046            /* Not UTF-8 mode */            /* Not UTF mode */
3047            {            {
3048            for (i = min; i < max; i++)            for (i = min; i < max; i++)
3049              {              {
# Line 2880  for (;;) Line 3053  for (;;)
3053                break;                break;
3054                }                }
3055              c = *eptr;              c = *eptr;
3056              if ((data[c/8] & (1 << (c&7))) == 0) break;  #ifndef COMPILE_PCRE8
3057                if (c > 255)
3058                  {
3059                  if (op == OP_CLASS) break;
3060                  }
3061                else
3062    #endif
3063                  if ((BYTE_MAP[c/8] & (1 << (c&7))) == 0) break;
3064              eptr++;              eptr++;
3065              }              }
3066            while (eptr >= pp)            while (eptr >= pp)
# Line 2893  for (;;) Line 3073  for (;;)
3073    
3074          RRETURN(MATCH_NOMATCH);          RRETURN(MATCH_NOMATCH);
3075          }          }
3076    #undef BYTE_MAP
3077        }        }
3078      /* Control never gets here */      /* Control never gets here */
3079    
# Line 2901  for (;;) Line 3082  for (;;)
3082      when UTF-8 mode mode is supported. Nevertheless, we may not be in UTF-8      when UTF-8 mode mode is supported. Nevertheless, we may not be in UTF-8
3083      mode, because Unicode properties are supported in non-UTF-8 mode. */      mode, because Unicode properties are supported in non-UTF-8 mode. */
3084    
3085  #ifdef SUPPORT_UTF8  #if defined SUPPORT_UTF || !defined COMPILE_PCRE8
3086      case OP_XCLASS:      case OP_XCLASS:
3087        {        {
3088        data = ecode + 1 + LINK_SIZE;                /* Save for matching */        data = ecode + 1 + LINK_SIZE;                /* Save for matching */
# Line 2926  for (;;) Line 3107  for (;;)
3107          case OP_CRMINRANGE:          case OP_CRMINRANGE:
3108          minimize = (*ecode == OP_CRMINRANGE);          minimize = (*ecode == OP_CRMINRANGE);
3109          min = GET2(ecode, 1);          min = GET2(ecode, 1);
3110          max = GET2(ecode, 3);          max = GET2(ecode, 1 + IMM2_SIZE);
3111          if (max == 0) max = INT_MAX;          if (max == 0) max = INT_MAX;
3112          ecode += 5;          ecode += 1 + 2 * IMM2_SIZE;
3113          break;          break;
3114    
3115          default:               /* No repeat follows */          default:               /* No repeat follows */
# Line 2946  for (;;) Line 3127  for (;;)
3127            RRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
3128            }            }
3129          GETCHARINCTEST(c, eptr);          GETCHARINCTEST(c, eptr);
3130          if (!_pcre_xclass(c, data)) RRETURN(MATCH_NOMATCH);          if (!PRIV(xclass)(c, data, utf)) RRETURN(MATCH_NOMATCH);
3131          }          }
3132    
3133        /* If max == min we can continue with the main loop without the        /* If max == min we can continue with the main loop without the
# Line 2970  for (;;) Line 3151  for (;;)
3151              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
3152              }              }
3153            GETCHARINCTEST(c, eptr);            GETCHARINCTEST(c, eptr);
3154            if (!_pcre_xclass(c, data)) RRETURN(MATCH_NOMATCH);            if (!PRIV(xclass)(c, data, utf)) RRETURN(MATCH_NOMATCH);
3155            }            }
3156          /* Control never gets here */          /* Control never gets here */
3157          }          }
# Line 2988  for (;;) Line 3169  for (;;)
3169              SCHECK_PARTIAL();              SCHECK_PARTIAL();
3170              break;              break;
3171              }              }
3172    #ifdef SUPPORT_UTF
3173            GETCHARLENTEST(c, eptr, len);            GETCHARLENTEST(c, eptr, len);
3174            if (!_pcre_xclass(c, data)) break;  #else
3175              c = *eptr;
3176    #endif
3177              if (!PRIV(xclass)(c, data, utf)) break;
3178            eptr += len;            eptr += len;
3179            }            }
3180          for(;;)          for(;;)
# Line 2997  for (;;) Line 3182  for (;;)
3182            RMATCH(eptr, ecode, offset_top, md, eptrb, RM21);            RMATCH(eptr, ecode, offset_top, md, eptrb, RM21);
3183            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3184            if (eptr-- == pp) break;        /* Stop if tried at original pos */            if (eptr-- == pp) break;        /* Stop if tried at original pos */
3185            if (utf8) BACKCHAR(eptr);  #ifdef SUPPORT_UTF
3186              if (utf) BACKCHAR(eptr);
3187    #endif
3188            }            }
3189          RRETURN(MATCH_NOMATCH);          RRETURN(MATCH_NOMATCH);
3190          }          }
# Line 3009  for (;;) Line 3196  for (;;)
3196      /* Match a single character, casefully */      /* Match a single character, casefully */
3197    
3198      case OP_CHAR:      case OP_CHAR:
3199  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
3200      if (utf8)      if (utf)
3201        {        {
3202        length = 1;        length = 1;
3203        ecode++;        ecode++;
# Line 3020  for (;;) Line 3207  for (;;)
3207          CHECK_PARTIAL();             /* Not SCHECK_PARTIAL() */          CHECK_PARTIAL();             /* Not SCHECK_PARTIAL() */
3208          RRETURN(MATCH_NOMATCH);          RRETURN(MATCH_NOMATCH);
3209          }          }
3210        while (length-- > 0) if (*ecode++ != *eptr++) RRETURN(MATCH_NOMATCH);        while (length-- > 0) if (*ecode++ != RAWUCHARINC(eptr)) RRETURN(MATCH_NOMATCH);
3211        }        }
3212      else      else
3213  #endif  #endif
3214        /* Not UTF mode */
     /* Non-UTF-8 mode */  
3215        {        {
3216        if (md->end_subject - eptr < 1)        if (md->end_subject - eptr < 1)
3217          {          {
# Line 3047  for (;;) Line 3233  for (;;)
3233        RRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
3234        }        }
3235    
3236  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
3237      if (utf8)      if (utf)
3238        {        {
3239        length = 1;        length = 1;
3240        ecode++;        ecode++;
# Line 3061  for (;;) Line 3247  for (;;)
3247    
3248        if (fc < 128)        if (fc < 128)
3249          {          {
3250          if (md->lcc[*ecode++] != md->lcc[*eptr++]) RRETURN(MATCH_NOMATCH);          pcre_uint32 cc = RAWUCHAR(eptr);
3251            if (md->lcc[fc] != TABLE_GET(cc, md->lcc, cc)) RRETURN(MATCH_NOMATCH);
3252            ecode++;
3253            eptr++;
3254          }          }
3255    
3256        /* Otherwise we must pick up the subject character. Note that we cannot        /* Otherwise we must pick up the subject character. Note that we cannot
# Line 3070  for (;;) Line 3259  for (;;)
3259    
3260        else        else
3261          {          {
3262          unsigned int dc;          pcre_uint32 dc;
3263          GETCHARINC(dc, eptr);          GETCHARINC(dc, eptr);
3264          ecode += length;          ecode += length;
3265    
# Line 3087  for (;;) Line 3276  for (;;)
3276          }          }
3277        }        }
3278      else      else
3279  #endif   /* SUPPORT_UTF8 */  #endif   /* SUPPORT_UTF */
3280    
3281      /* Non-UTF-8 mode */      /* Not UTF mode */
3282        {        {
3283        if (md->lcc[ecode[1]] != md->lcc[*eptr++]) RRETURN(MATCH_NOMATCH);        if (TABLE_GET(ecode[1], md->lcc, ecode[1])
3284              != TABLE_GET(*eptr, md->lcc, *eptr)) RRETURN(MATCH_NOMATCH);
3285          eptr++;
3286        ecode += 2;        ecode += 2;
3287        }        }
3288      break;      break;
# Line 3101  for (;;) Line 3292  for (;;)
3292      case OP_EXACT:      case OP_EXACT:
3293      case OP_EXACTI:      case OP_EXACTI:
3294      min = max = GET2(ecode, 1);      min = max = GET2(ecode, 1);
3295      ecode += 3;      ecode += 1 + IMM2_SIZE;
3296      goto REPEATCHAR;      goto REPEATCHAR;
3297    
3298      case OP_POSUPTO:      case OP_POSUPTO:
# Line 3116  for (;;) Line 3307  for (;;)
3307      min = 0;      min = 0;
3308      max = GET2(ecode, 1);      max = GET2(ecode, 1);
3309      minimize = *ecode == OP_MINUPTO || *ecode == OP_MINUPTOI;      minimize = *ecode == OP_MINUPTO || *ecode == OP_MINUPTOI;
3310      ecode += 3;      ecode += 1 + IMM2_SIZE;
3311      goto REPEATCHAR;      goto REPEATCHAR;
3312    
3313      case OP_POSSTAR:      case OP_POSSTAR:
# Line 3164  for (;;) Line 3355  for (;;)
3355      /* Common code for all repeated single-character matches. */      /* Common code for all repeated single-character matches. */
3356    
3357      REPEATCHAR:      REPEATCHAR:
3358  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
3359      if (utf8)      if (utf)
3360        {        {
3361        length = 1;        length = 1;
3362        charptr = ecode;        charptr = ecode;
# Line 3178  for (;;) Line 3369  for (;;)
3369        if (length > 1)        if (length > 1)
3370          {          {
3371  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
3372          unsigned int othercase;          pcre_uint32 othercase;
3373          if (op >= OP_STARI &&     /* Caseless */          if (op >= OP_STARI &&     /* Caseless */
3374              (othercase = UCD_OTHERCASE(fc)) != fc)              (othercase = UCD_OTHERCASE(fc)) != fc)
3375            oclength = _pcre_ord2utf8(othercase, occhars);            oclength = PRIV(ord2utf)(othercase, occhars);
3376          else oclength = 0;          else oclength = 0;
3377  #endif  /* SUPPORT_UCP */  #endif  /* SUPPORT_UCP */
3378    
3379          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
3380            {            {
3381            if (eptr <= md->end_subject - length &&            if (eptr <= md->end_subject - length &&
3382              memcmp(eptr, charptr, length) == 0) eptr += length;              memcmp(eptr, charptr, IN_UCHARS(length)) == 0) eptr += length;
3383  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
3384            else if (oclength > 0 &&            else if (oclength > 0 &&
3385                     eptr <= md->end_subject - oclength &&                     eptr <= md->end_subject - oclength &&
3386                     memcmp(eptr, occhars, oclength) == 0) eptr += oclength;                     memcmp(eptr, occhars, IN_UCHARS(oclength)) == 0) eptr += oclength;
3387  #endif  /* SUPPORT_UCP */  #endif  /* SUPPORT_UCP */
3388            else            else
3389              {              {
# Line 3211  for (;;) Line 3402  for (;;)
3402              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3403              if (fi >= max) RRETURN(MATCH_NOMATCH);              if (fi >= max) RRETURN(MATCH_NOMATCH);
3404              if (eptr <= md->end_subject - length &&              if (eptr <= md->end_subject - length &&
3405                memcmp(eptr, charptr, length) == 0) eptr += length;                memcmp(eptr, charptr, IN_UCHARS(length)) == 0) eptr += length;
3406  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
3407              else if (oclength > 0 &&              else if (oclength > 0 &&
3408                       eptr <= md->end_subject - oclength &&                       eptr <= md->end_subject - oclength &&
3409                       memcmp(eptr, occhars, oclength) == 0) eptr += oclength;                       memcmp(eptr, occhars, IN_UCHARS(oclength)) == 0) eptr += oclength;
3410  #endif  /* SUPPORT_UCP */  #endif  /* SUPPORT_UCP */
3411              else              else
3412                {                {
# Line 3232  for (;;) Line 3423  for (;;)
3423            for (i = min; i < max; i++)            for (i = min; i < max; i++)
3424              {              {
3425              if (eptr <= md->end_subject - length &&              if (eptr <= md->end_subject - length &&
3426                  memcmp(eptr, charptr, length) == 0) eptr += length;                  memcmp(eptr, charptr, IN_UCHARS(length)) == 0) eptr += length;
3427  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
3428              else if (oclength > 0 &&              else if (oclength > 0 &&
3429                       eptr <= md->end_subject - oclength &&                       eptr <= md->end_subject - oclength &&
3430                       memcmp(eptr, occhars, oclength) == 0) eptr += oclength;                       memcmp(eptr, occhars, IN_UCHARS(oclength)) == 0) eptr += oclength;
3431  #endif  /* SUPPORT_UCP */  #endif  /* SUPPORT_UCP */
3432              else              else
3433                {                {
# Line 3268  for (;;) Line 3459  for (;;)
3459        value of fc will always be < 128. */        value of fc will always be < 128. */
3460        }        }
3461      else      else
3462  #endif  /* SUPPORT_UTF8 */  #endif  /* SUPPORT_UTF */
3463          /* When not in UTF-8 mode, load a single-byte character. */
3464      /* When not in UTF-8 mode, load a single-byte character. */        fc = *ecode++;
3465    
3466      fc = *ecode++;      /* The value of fc at this point is always one character, though we may
3467        or may not be in UTF mode. The code is duplicated for the caseless and
     /* The value of fc at this point is always less than 256, though we may or  
     may not be in UTF-8 mode. The code is duplicated for the caseless and  
3468      caseful cases, for speed, since matching characters is likely to be quite      caseful cases, for speed, since matching characters is likely to be quite
3469      common. First, ensure the minimum number of matches are present. If min =      common. First, ensure the minimum number of matches are present. If min =
3470      max, continue at the same level without recursing. Otherwise, if      max, continue at the same level without recursing. Otherwise, if
# Line 3284  for (;;) Line 3473  for (;;)
3473      maximizing, find the maximum number of characters and work backwards. */      maximizing, find the maximum number of characters and work backwards. */
3474    
3475      DPRINTF(("matching %c{%d,%d} against subject %.*s\n", fc, min, max,      DPRINTF(("matching %c{%d,%d} against subject %.*s\n", fc, min, max,
3476        max, eptr));        max, (char *)eptr));
3477    
3478      if (op >= OP_STARI)  /* Caseless */      if (op >= OP_STARI)  /* Caseless */
3479        {        {
3480        fc = md->lcc[fc];  #ifdef COMPILE_PCRE8
3481          /* fc must be < 128 if UTF is enabled. */
3482          foc = md->fcc[fc];
3483    #else
3484    #ifdef SUPPORT_UTF
3485    #ifdef SUPPORT_UCP
3486          if (utf && fc > 127)
3487            foc = UCD_OTHERCASE(fc);
3488    #else
3489          if (utf && fc > 127)
3490            foc = fc;
3491    #endif /* SUPPORT_UCP */
3492          else
3493    #endif /* SUPPORT_UTF */
3494            foc = TABLE_GET(fc, md->fcc, fc);
3495    #endif /* COMPILE_PCRE8 */
3496    
3497        for (i = 1; i <= min; i++)        for (i = 1; i <= min; i++)
3498          {          {
3499            pcre_uint32 cc;                 /* Faster than pcre_uchar */
3500          if (eptr >= md->end_subject)          if (eptr >= md->end_subject)
3501            {            {
3502            SCHECK_PARTIAL();            SCHECK_PARTIAL();
3503            RRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
3504            }            }
3505          if (fc != md->lcc[*eptr++]) RRETURN(MATCH_NOMATCH);          cc = RAWUCHARTEST(eptr);
3506            if (fc != cc && foc != cc) RRETURN(MATCH_NOMATCH);
3507            eptr++;
3508          }          }
3509        if (min == max) continue;        if (min == max) continue;
3510        if (minimize)        if (minimize)
3511          {          {
3512          for (fi = min;; fi++)          for (fi = min;; fi++)
3513            {            {
3514              pcre_uint32 cc;               /* Faster than pcre_uchar */
3515            RMATCH(eptr, ecode, offset_top, md, eptrb, RM24);            RMATCH(eptr, ecode, offset_top, md, eptrb, RM24);
3516            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3517            if (fi >= max) RRETURN(MATCH_NOMATCH);            if (fi >= max) RRETURN(MATCH_NOMATCH);
# Line 3311  for (;;) Line 3520  for (;;)
3520              SCHECK_PARTIAL();              SCHECK_PARTIAL();
3521              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
3522              }              }
3523            if (fc != md->lcc[*eptr++]) RRETURN(MATCH_NOMATCH);            cc = RAWUCHARTEST(eptr);
3524              if (fc != cc && foc != cc) RRETURN(MATCH_NOMATCH);
3525              eptr++;
3526            }            }
3527          /* Control never gets here */          /* Control never gets here */
3528          }          }
# Line 3320  for (;;) Line 3531  for (;;)
3531          pp = eptr;          pp = eptr;
3532          for (i = min; i < max; i++)          for (i = min; i < max; i++)
3533            {            {
3534              pcre_uint32 cc;               /* Faster than pcre_uchar */
3535            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
3536              {              {
3537              SCHECK_PARTIAL();              SCHECK_PARTIAL();
3538              break;              break;
3539              }              }
3540            if (fc != md->lcc[*eptr]) break;            cc = RAWUCHARTEST(eptr);
3541              if (fc != cc && foc != cc) break;
3542            eptr++;            eptr++;
3543            }            }
3544    
# Line 3353  for (;;) Line 3566  for (;;)
3566            SCHECK_PARTIAL();            SCHECK_PARTIAL();
3567            RRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
3568            }            }
3569          if (fc != *eptr++) RRETURN(MATCH_NOMATCH);          if (fc != RAWUCHARINCTEST(eptr)) RRETURN(MATCH_NOMATCH);
3570          }          }
3571    
3572        if (min == max) continue;        if (min == max) continue;
# Line 3370  for (;;) Line 3583  for (;;)
3583              SCHECK_PARTIAL();              SCHECK_PARTIAL();
3584              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
3585              }              }
3586            if (fc != *eptr++) RRETURN(MATCH_NOMATCH);            if (fc != RAWUCHARINCTEST(eptr)) RRETURN(MATCH_NOMATCH);
3587            }            }
3588          /* Control never gets here */          /* Control never gets here */
3589          }          }
# Line 3384  for (;;) Line 3597  for (;;)
3597              SCHECK_PARTIAL();              SCHECK_PARTIAL();
3598              break;              break;
3599              }              }
3600            if (fc != *eptr) break;            if (fc != RAWUCHARTEST(eptr)) break;
3601            eptr++;            eptr++;
3602            }            }
3603          if (possessive) continue;          if (possessive) continue;
# Line 3410  for (;;) Line 3623  for (;;)
3623        SCHECK_PARTIAL();        SCHECK_PARTIAL();
3624        RRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
3625        }        }
3626      ecode++;  #ifdef SUPPORT_UTF
3627      GETCHARINCTEST(c, eptr);      if (utf)
     if (op == OP_NOTI)         /* The caseless case */  
3628        {        {
3629  #ifdef SUPPORT_UTF8        register pcre_uint32 ch, och;
3630        if (c < 256)  
3631  #endif        ecode++;
3632        c = md->lcc[c];        GETCHARINC(ch, ecode);
3633        if (md->lcc[*ecode++] == c) RRETURN(MATCH_NOMATCH);        GETCHARINC(c, eptr);
3634    
3635          if (op == OP_NOT)
3636            {
3637            if (ch == c) RRETURN(MATCH_NOMATCH);
3638            }
3639          else
3640            {
3641    #ifdef SUPPORT_UCP
3642            if (ch > 127)
3643              och = UCD_OTHERCASE(ch);
3644    #else
3645            if (ch > 127)
3646              och = ch;
3647    #endif /* SUPPORT_UCP */
3648            else
3649              och = TABLE_GET(ch, md->fcc, ch);
3650            if (ch == c || och == c) RRETURN(MATCH_NOMATCH);
3651            }
3652        }        }
3653      else    /* Caseful */      else
3654    #endif
3655        {        {
3656        if (*ecode++ == c) RRETURN(MATCH_NOMATCH);        register pcre_uint32 ch = ecode[1];
3657          c = *eptr++;
3658          if (ch == c || (op == OP_NOTI && TABLE_GET(ch, md->fcc, ch) == c))
3659            RRETURN(MATCH_NOMATCH);
3660          ecode += 2;
3661        }        }
3662      break;      break;
3663    
# Line 3436  for (;;) Line 3671  for (;;)
3671      case OP_NOTEXACT:      case OP_NOTEXACT:
3672      case OP_NOTEXACTI:      case OP_NOTEXACTI:
3673      min = max = GET2(ecode, 1);      min = max = GET2(ecode, 1);
3674      ecode += 3;      ecode += 1 + IMM2_SIZE;
3675      goto REPEATNOTCHAR;      goto REPEATNOTCHAR;
3676    
3677      case OP_NOTUPTO:      case OP_NOTUPTO:
# Line 3446  for (;;) Line 3681  for (;;)
3681      min = 0;      min = 0;
3682      max = GET2(ecode, 1);      max = GET2(ecode, 1);
3683      minimize = *ecode == OP_NOTMINUPTO || *ecode == OP_NOTMINUPTOI;      minimize = *ecode == OP_NOTMINUPTO || *ecode == OP_NOTMINUPTOI;
3684      ecode += 3;      ecode += 1 + IMM2_SIZE;
3685      goto REPEATNOTCHAR;      goto REPEATNOTCHAR;
3686    
3687      case OP_NOTPOSSTAR:      case OP_NOTPOSSTAR:
# Line 3478  for (;;) Line 3713  for (;;)
3713      possessive = TRUE;      possessive = TRUE;
3714      min = 0;      min = 0;
3715      max = GET2(ecode, 1);      max = GET2(ecode, 1);
3716      ecode += 3;      ecode += 1 + IMM2_SIZE;
3717      goto REPEATNOTCHAR;      goto REPEATNOTCHAR;
3718    
3719      case OP_NOTSTAR:      case OP_NOTSTAR:
# Line 3502  for (;;) Line 3737  for (;;)
3737      /* Common code for all repeated single-byte matches. */      /* Common code for all repeated single-byte matches. */
3738    
3739      REPEATNOTCHAR:      REPEATNOTCHAR:
3740      fc = *ecode++;      GETCHARINCTEST(fc, ecode);
3741    
3742      /* The code is duplicated for the caseless and caseful cases, for speed,      /* The code is duplicated for the caseless and caseful cases, for speed,
3743      since matching characters is likely to be quite common. First, ensure the      since matching characters is likely to be quite common. First, ensure the
# Line 3513  for (;;) Line 3748  for (;;)
3748      characters and work backwards. */      characters and work backwards. */
3749    
3750      DPRINTF(("negative matching %c{%d,%d} against subject %.*s\n", fc, min, max,      DPRINTF(("negative matching %c{%d,%d} against subject %.*s\n", fc, min, max,
3751        max, eptr));        max, (char *)eptr));
3752    
3753      if (op >= OP_NOTSTARI)     /* Caseless */      if (op >= OP_NOTSTARI)     /* Caseless */
3754        {        {
3755        fc = md->lcc[fc];  #ifdef SUPPORT_UTF
3756    #ifdef SUPPORT_UCP
3757          if (utf && fc > 127)
3758            foc = UCD_OTHERCASE(fc);
3759    #else
3760          if (utf && fc > 127)
3761            foc = fc;
3762    #endif /* SUPPORT_UCP */
3763          else
3764    #endif /* SUPPORT_UTF */
3765            foc = TABLE_GET(fc, md->fcc, fc);
3766    
3767  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
3768        /* UTF-8 mode */        if (utf)
       if (utf8)  
3769          {          {
3770          register unsigned int d;          register pcre_uint32 d;
3771          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
3772            {            {
3773            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
# Line 3532  for (;;) Line 3776  for (;;)
3776              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
3777              }              }
3778            GETCHARINC(d, eptr);            GETCHARINC(d, eptr);
3779            if (d < 256) d = md->lcc[d];            if (fc == d || (unsigned int)foc == d) RRETURN(MATCH_NOMATCH);
           if (fc == d) RRETURN(MATCH_NOMATCH);  
3780            }            }
3781          }          }
3782        else        else
3783  #endif  #endif
3784          /* Not UTF mode */
       /* Not UTF-8 mode */  
3785          {          {
3786          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
3787            {            {
# Line 3548  for (;;) Line 3790  for (;;)
3790              SCHECK_PARTIAL();              SCHECK_PARTIAL();
3791              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
3792              }              }
3793            if (fc == md->lcc[*eptr++]) RRETURN(MATCH_NOMATCH);            if (fc == *eptr || foc == *eptr) RRETURN(MATCH_NOMATCH);
3794              eptr++;
3795            }            }
3796          }          }
3797    
# Line 3556  for (;;) Line 3799  for (;;)
3799    
3800        if (minimize)        if (minimize)
3801          {          {
3802  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
3803          /* UTF-8 mode */          if (utf)
         if (utf8)  
3804            {            {
3805            register unsigned int d;            register pcre_uint32 d;
3806            for (fi = min;; fi++)            for (fi = min;; fi++)
3807              {              {
3808              RMATCH(eptr, ecode, offset_top, md, eptrb, RM28);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM28);
# Line 3572  for (;;) Line 3814  for (;;)
3814                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
3815                }                }
3816              GETCHARINC(d, eptr);              GETCHARINC(d, eptr);
3817              if (d < 256) d = md->lcc[d];              if (fc == d || (unsigned int)foc == d) RRETURN(MATCH_NOMATCH);
             if (fc == d) RRETURN(MATCH_NOMATCH);  
3818              }              }
3819            }            }
3820          else          else
3821  #endif  #endif
3822          /* Not UTF-8 mode */          /* Not UTF mode */
3823            {            {
3824            for (fi = min;; fi++)            for (fi = min;; fi++)
3825              {              {
# Line 3590  for (;;) Line 3831  for (;;)
3831                SCHECK_PARTIAL();                SCHECK_PARTIAL();
3832                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
3833                }                }
3834              if (fc == md->lcc[*eptr++]) RRETURN(MATCH_NOMATCH);              if (fc == *eptr || foc == *eptr) RRETURN(MATCH_NOMATCH);
3835                eptr++;
3836              }              }
3837            }            }
3838          /* Control never gets here */          /* Control never gets here */
# Line 3602  for (;;) Line 3844  for (;;)
3844          {          {
3845          pp = eptr;          pp = eptr;
3846    
3847  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
3848          /* UTF-8 mode */          if (utf)
         if (utf8)  
3849            {            {
3850            register unsigned int d;            register pcre_uint32 d;
3851            for (i = min; i < max; i++)            for (i = min; i < max; i++)
3852              {              {
3853              int len = 1;              int len = 1;
# Line 3616  for (;;) Line 3857  for (;;)
3857                break;                break;
3858                }                }
3859              GETCHARLEN(d, eptr, len);              GETCHARLEN(d, eptr, len);
3860              if (d < 256) d = md->lcc[d];              if (fc == d || (unsigned int)foc == d) break;
             if (fc == d) break;  
3861              eptr += len;              eptr += len;
3862              }              }
3863          if (possessive) continue;            if (possessive) continue;
3864          for(;;)            for(;;)
3865              {              {
3866              RMATCH(eptr, ecode, offset_top, md, eptrb, RM30);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM30);
3867              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
# Line 3631  for (;;) Line 3871  for (;;)
3871            }            }
3872          else          else
3873  #endif  #endif
3874          /* Not UTF-8 mode */          /* Not UTF mode */
3875            {            {
3876            for (i = min; i < max; i++)            for (i = min; i < max; i++)
3877              {              {
# Line 3640  for (;;) Line 3880  for (;;)
3880                SCHECK_PARTIAL();                SCHECK_PARTIAL();
3881                break;                break;
3882                }                }
3883              if (fc == md->lcc[*eptr]) break;              if (fc == *eptr || foc == *eptr) break;
3884              eptr++;              eptr++;
3885              }              }
3886            if (possessive) continue;            if (possessive) continue;
# Line 3661  for (;;) Line 3901  for (;;)
3901    
3902      else      else
3903        {        {
3904  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
3905        /* UTF-8 mode */        if (utf)
       if (utf8)  
3906          {          {
3907          register unsigned int d;          register pcre_uint32 d;
3908          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
3909            {            {
3910            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
# Line 3679  for (;;) Line 3918  for (;;)
3918          }          }
3919        else        else
3920  #endif  #endif
3921        /* Not UTF-8 mode */        /* Not UTF mode */
3922          {          {
3923          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
3924            {            {
# Line 3696  for (;;) Line 3935  for (;;)
3935    
3936        if (minimize)        if (minimize)
3937          {          {
3938  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
3939          /* UTF-8 mode */          if (utf)
         if (utf8)  
3940            {            {
3941            register unsigned int d;            register pcre_uint32 d;
3942            for (fi = min;; fi++)            for (fi = min;; fi++)
3943              {              {
3944              RMATCH(eptr, ecode, offset_top, md, eptrb, RM32);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM32);
# Line 3717  for (;;) Line 3955  for (;;)
3955            }            }
3956          else          else
3957  #endif  #endif
3958          /* Not UTF-8 mode */          /* Not UTF mode */
3959            {            {
3960            for (fi = min;; fi++)            for (fi = min;; fi++)
3961              {              {
# Line 3741  for (;;) Line 3979  for (;;)
3979          {          {
3980          pp = eptr;          pp = eptr;
3981    
3982  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
3983          /* UTF-8 mode */          if (utf)
         if (utf8)  
3984            {            {
3985            register unsigned int d;            register pcre_uint32 d;
3986            for (i = min; i < max; i++)            for (i = min; i < max; i++)
3987              {              {
3988              int len = 1;              int len = 1;
# Line 3769  for (;;) Line 4006  for (;;)
4006            }            }
4007          else          else
4008  #endif  #endif
4009          /* Not UTF-8 mode */          /* Not UTF mode */
4010            {            {
4011            for (i = min; i < max; i++)            for (i = min; i < max; i++)
4012              {              {
# Line 3802  for (;;) Line 4039  for (;;)
4039      case OP_TYPEEXACT:      case OP_TYPEEXACT:
4040      min = max = GET2(ecode, 1);      min = max = GET2(ecode, 1);
4041      minimize = TRUE;      minimize = TRUE;
4042      ecode += 3;      ecode += 1 + IMM2_SIZE;
4043      goto REPEATTYPE;      goto REPEATTYPE;
4044    
4045      case OP_TYPEUPTO:      case OP_TYPEUPTO:
# Line 3810  for (;;) Line 4047  for (;;)
4047      min = 0;      min = 0;
4048      max = GET2(ecode, 1);      max = GET2(ecode, 1);
4049      minimize = *ecode == OP_TYPEMINUPTO;      minimize = *ecode == OP_TYPEMINUPTO;
4050      ecode += 3;      ecode += 1 + IMM2_SIZE;
4051      goto REPEATTYPE;      goto REPEATTYPE;
4052    
4053      case OP_TYPEPOSSTAR:      case OP_TYPEPOSSTAR:
# Line 3838  for (;;) Line 4075  for (;;)
4075      possessive = TRUE;      possessive = TRUE;
4076      min = 0;      min = 0;
4077      max = GET2(ecode, 1);      max = GET2(ecode, 1);
4078      ecode += 3;      ecode += 1 + IMM2_SIZE;
4079      goto REPEATTYPE;      goto REPEATTYPE;
4080    
4081      case OP_TYPESTAR:      case OP_TYPESTAR:
# Line 4021  for (;;) Line 4258  for (;;)
4258              }              }
4259            break;            break;
4260    
4261              case PT_CLIST:
4262              for (i = 1; i <= min; i++)
4263                {
4264                const pcre_uint32 *cp;
4265                if (eptr >= md->end_subject)
4266                  {
4267                  SCHECK_PARTIAL();
4268                  RRETURN(MATCH_NOMATCH);
4269                  }
4270                GETCHARINCTEST(c, eptr);
4271                cp = PRIV(ucd_caseless_sets) + prop_value;
4272                for (;;)
4273                  {
4274                  if (c < *cp)
4275                    { if (prop_fail_result) break; else { RRETURN(MATCH_NOMATCH); } }
4276                  if (c == *cp++)
4277                    { if (prop_fail_result) { RRETURN(MATCH_NOMATCH); } else break; }
4278                  }
4279                }
4280              break;
4281    
4282              case PT_UCNC:
4283              for (i = 1; i <= min; i++)
4284                {
4285                if (eptr >= md->end_subject)
4286                  {
4287                  SCHECK_PARTIAL();
4288                  RRETURN(MATCH_NOMATCH);
4289                  }
4290                GETCHARINCTEST(c, eptr);
4291                if ((c == CHAR_DOLLAR_SIGN || c == CHAR_COMMERCIAL_AT ||
4292                     c == CHAR_GRAVE_ACCENT || (c >= 0xa0 && c <= 0xd7ff) ||
4293                     c >= 0xe000) == prop_fail_result)
4294                  RRETURN(MATCH_NOMATCH);
4295                }
4296              break;
4297    
4298            /* This should not occur */            /* This should not occur */
4299    
4300            default:            default:
# Line 4040  for (;;) Line 4314  for (;;)
4314              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4315              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4316              }              }
4317            GETCHARINCTEST(c, eptr);            else
           if (UCD_CATEGORY(c) == ucp_M) RRETURN(MATCH_NOMATCH);  
           while (eptr < md->end_subject)  
4318              {              {
4319              int len = 1;              int lgb, rgb;
4320              if (!utf8) c = *eptr; else { GETCHARLEN(c, eptr, len); }              GETCHARINCTEST(c, eptr);
4321              if (UCD_CATEGORY(c) != ucp_M) break;              lgb = UCD_GRAPHBREAK(c);
4322              eptr += len;             while (eptr < md->end_subject)
4323                  {
4324                  int len = 1;
4325                  if (!utf) c = *eptr; else { GETCHARLEN(c, eptr, len); }
4326                  rgb = UCD_GRAPHBREAK(c);
4327                  if ((PRIV(ucp_gbtable)[lgb] & (1 << rgb)) == 0) break;
4328                  lgb = rgb;
4329                  eptr += len;
4330                  }
4331              }              }
4332              CHECK_PARTIAL();
4333            }            }
4334          }          }
4335    
# Line 4057  for (;;) Line 4338  for (;;)
4338    
4339  /* Handle all other cases when the coding is UTF-8 */  /* Handle all other cases when the coding is UTF-8 */
4340    
4341  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
4342        if (utf8) switch(ctype)        if (utf) switch(ctype)
4343          {          {
4344          case OP_ANY:          case OP_ANY:
4345          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
# Line 4069  for (;;) Line 4350  for (;;)
4350              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4351              }              }
4352            if (IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH);            if (IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH);
4353              if (md->partial != 0 &&
4354                  eptr + 1 >= md->end_subject &&
4355                  NLBLOCK->nltype == NLTYPE_FIXED &&
4356                  NLBLOCK->nllen == 2 &&
4357                  RAWUCHAR(eptr) == NLBLOCK->nl[0])
4358                {
4359                md->hitend = TRUE;
4360                if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
4361                }
4362            eptr++;            eptr++;
4363            while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;            ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);
4364            }            }
4365          break;          break;
4366    
# Line 4083  for (;;) Line 4373  for (;;)
4373              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4374              }              }
4375            eptr++;            eptr++;
4376            while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;            ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);
4377            }            }
4378          break;          break;
4379    
# Line 4105  for (;;) Line 4395  for (;;)
4395              {              {
4396              default: RRETURN(MATCH_NOMATCH);              default: RRETURN(MATCH_NOMATCH);
4397    
4398              case 0x000d:              case CHAR_CR:
4399              if (eptr < md->end_subject && *eptr == 0x0a) eptr++;              if (eptr < md->end_subject && RAWUCHAR(eptr) == CHAR_LF) eptr++;
4400              break;              break;
4401    
4402              case 0x000a:              case CHAR_LF:
4403              break;              break;
4404    
4405              case 0x000b:              case CHAR_VT:
4406              case 0x000c:              case CHAR_FF:
4407              case 0x0085:              case CHAR_NEL:
4408    #ifndef EBCDIC
4409              case 0x2028:              case 0x2028:
4410              case 0x2029:              case 0x2029:
4411    #endif  /* Not EBCDIC */
4412              if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);              if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);
4413              break;              break;
4414              }              }
# Line 4134  for (;;) Line 4426  for (;;)
4426            GETCHARINC(c, eptr);            GETCHARINC(c, eptr);
4427            switch(c)            switch(c)
4428              {              {
4429                HSPACE_CASES: RRETURN(MATCH_NOMATCH);  /* Byte and multibyte cases */
4430              default: break;              default: break;
             case 0x09:      /* HT */  
             case 0x20:      /* SPACE */  
             case 0xa0:      /* NBSP */  
             case 0x1680:    /* OGHAM SPACE MARK */  
             case 0x180e:    /* MONGOLIAN VOWEL SEPARATOR */  
             case 0x2000:    /* EN QUAD */  
             case 0x2001:    /* EM QUAD */  
             case 0x2002:    /* EN SPACE */  
             case 0x2003:    /* EM SPACE */  
             case 0x2004:    /* THREE-PER-EM SPACE */  
             case 0x2005:    /* FOUR-PER-EM SPACE */  
             case 0x2006:    /* SIX-PER-EM SPACE */  
             case 0x2007:    /* FIGURE SPACE */  
             case 0x2008:    /* PUNCTUATION SPACE */  
             case 0x2009:    /* THIN SPACE */  
             case 0x200A:    /* HAIR SPACE */  
             case 0x202f:    /* NARROW NO-BREAK SPACE */  
             case 0x205f:    /* MEDIUM MATHEMATICAL SPACE */  
             case 0x3000:    /* IDEOGRAPHIC SPACE */  
             RRETURN(MATCH_NOMATCH);  
4431              }              }
4432            }            }
4433          break;          break;
# Line 4170  for (;;) Line 4443  for (;;)
4443            GETCHARINC(c, eptr);            GETCHARINC(c, eptr);
4444            switch(c)            switch(c)
4445              {              {
4446                HSPACE_CASES: break;  /* Byte and multibyte cases */
4447              default: RRETURN(MATCH_NOMATCH);              default: RRETURN(MATCH_NOMATCH);
             case 0x09:      /* HT */  
             case 0x20:      /* SPACE */  
             case 0xa0:      /* NBSP */  
             case 0x1680:    /* OGHAM SPACE MARK */  
             case 0x180e:    /* MONGOLIAN VOWEL SEPARATOR */  
             case 0x2000:    /* EN QUAD */  
             case 0x2001:    /* EM QUAD */  
             case 0x2002:    /* EN SPACE */  
             case 0x2003:    /* EM SPACE */  
             case 0x2004:    /* THREE-PER-EM SPACE */  
             case 0x2005:    /* FOUR-PER-EM SPACE */  
             case 0x2006:    /* SIX-PER-EM SPACE */  
             case 0x2007:    /* FIGURE SPACE */  
             case 0x2008:    /* PUNCTUATION SPACE */  
             case 0x2009:    /* THIN SPACE */  
             case 0x200A:    /* HAIR SPACE */  
             case 0x202f:    /* NARROW NO-BREAK SPACE */  
             case 0x205f:    /* MEDIUM MATHEMATICAL SPACE */  
             case 0x3000:    /* IDEOGRAPHIC SPACE */  
             break;  
4448              }              }
4449            }            }
4450          break;          break;
# Line 4206  for (;;) Line 4460  for (;;)
4460            GETCHARINC(c, eptr);            GETCHARINC(c, eptr);
4461            switch(c)            switch(c)
4462              {              {
4463                VSPACE_CASES: RRETURN(MATCH_NOMATCH);
4464              default: break;              default: break;
             case 0x0a:      /* LF */  
             case 0x0b:      /* VT */  
             case 0x0c:      /* FF */  
             case 0x0d:      /* CR */  
             case 0x85:      /* NEL */  
             case 0x2028:    /* LINE SEPARATOR */  
             case 0x2029:    /* PARAGRAPH SEPARATOR */  
             RRETURN(MATCH_NOMATCH);  
4465              }              }
4466            }            }
4467          break;          break;
# Line 4230  for (;;) Line 4477  for (;;)
4477            GETCHARINC(c, eptr);            GETCHARINC(c, eptr);
4478            switch(c)            switch(c)
4479              {              {
4480                VSPACE_CASES: break;
4481              default: RRETURN(MATCH_NOMATCH);              default: RRETURN(MATCH_NOMATCH);
             case 0x0a:      /* LF */  
             case 0x0b:      /* VT */  
             case 0x0c:      /* FF */  
             case 0x0d:      /* CR */  
             case 0x85:      /* NEL */  
             case 0x2028:    /* LINE SEPARATOR */  
             case 0x2029:    /* PARAGRAPH SEPARATOR */  
             break;  
4482              }              }
4483            }            }
4484          break;          break;
# Line 4260  for (;;) Line 4500  for (;;)
4500          case OP_DIGIT:          case OP_DIGIT:
4501          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
4502            {            {
4503              pcre_uint32 cc;
4504            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
4505              {              {
4506              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4507              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4508              }              }
4509            if (*eptr >= 128 || (md->ctypes[*eptr++] & ctype_digit) == 0)            cc = RAWUCHAR(eptr);
4510              if (cc >= 128 || (md->ctypes[cc] & ctype_digit) == 0)
4511              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4512              eptr++;
4513            /* No need to skip more bytes - we know it's a 1-byte character */            /* No need to skip more bytes - we know it's a 1-byte character */
4514            }            }
4515          break;          break;
# Line 4274  for (;;) Line 4517  for (;;)
4517          case OP_NOT_WHITESPACE:          case OP_NOT_WHITESPACE:
4518          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
4519            {            {
4520              pcre_uint32 cc;
4521            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
4522              {              {
4523              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4524              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4525              }              }
4526            if (*eptr < 128 && (md->ctypes[*eptr] & ctype_space) != 0)            cc = RAWUCHAR(eptr);
4527              if (cc < 128 && (md->ctypes[cc] & ctype_space) != 0)
4528              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4529            while (++eptr < md->end_subject && (*eptr & 0xc0) == 0x80);            eptr++;
4530              ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);
4531            }            }
4532          break;          break;
4533    
4534          case OP_WHITESPACE:          case OP_WHITESPACE:
4535          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
4536            {            {
4537              pcre_uint32 cc;
4538            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
4539              {              {
4540              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4541              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4542              }              }
4543            if (*eptr >= 128 || (md->ctypes[*eptr++] & ctype_space) == 0)            cc = RAWUCHAR(eptr);
4544              if (cc >= 128 || (md->ctypes[cc] & ctype_space) == 0)
4545              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4546              eptr++;
4547            /* No need to skip more bytes - we know it's a 1-byte character */            /* No need to skip more bytes - we know it's a 1-byte character */
4548            }            }
4549          break;          break;
# Line 4302  for (;;) Line 4551  for (;;)
4551          case OP_NOT_WORDCHAR:          case OP_NOT_WORDCHAR:
4552          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
4553            {            {
4554              pcre_uint32 cc;
4555            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
4556              {              {
4557              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4558              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4559              }              }
4560            if (*eptr < 128 && (md->ctypes[*eptr] & ctype_word) != 0)            cc = RAWUCHAR(eptr);
4561              if (cc < 128 && (md->ctypes[cc] & ctype_word) != 0)
4562              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4563            while (++eptr < md->end_subject && (*eptr & 0xc0) == 0x80);            eptr++;
4564              ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);
4565            }            }
4566          break;          break;
4567    
4568          case OP_WORDCHAR:          case OP_WORDCHAR:
4569          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
4570            {            {
4571              pcre_uint32 cc;
4572            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
4573              {              {
4574              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4575              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4576              }              }
4577            if (*eptr >= 128 || (md->ctypes[*eptr++] & ctype_word) == 0)            cc = RAWUCHAR(eptr);
4578              if (cc >= 128 || (md->ctypes[cc] & ctype_word) == 0)
4579              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4580              eptr++;
4581            /* No need to skip more bytes - we know it's a 1-byte character */            /* No need to skip more bytes - we know it's a 1-byte character */
4582            }            }
4583          break;          break;
# Line 4332  for (;;) Line 4587  for (;;)
4587          }  /* End switch(ctype) */          }  /* End switch(ctype) */
4588    
4589        else        else
4590  #endif     /* SUPPORT_UTF8 */  #endif     /* SUPPORT_UTF */
4591    
4592        /* Code for the non-UTF-8 case for minimum matching of operators other        /* Code for the non-UTF-8 case for minimum matching of operators other
4593        than OP_PROP and OP_NOTPROP. */        than OP_PROP and OP_NOTPROP. */
# Line 4348  for (;;) Line 4603  for (;;)
4603              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4604              }              }
4605            if (IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH);            if (IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH);
4606              if (md->partial != 0 &&
4607                  eptr + 1 >= md->end_subject &&
4608                  NLBLOCK->nltype == NLTYPE_FIXED &&
4609                  NLBLOCK->nllen == 2 &&
4610                  *eptr == NLBLOCK->nl[0])
4611                {
4612                md->hitend = TRUE;
4613                if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
4614                }
4615            eptr++;            eptr++;
4616            }            }
4617          break;          break;
# Line 4382  for (;;) Line 4646  for (;;)
4646              {              {
4647              default: RRETURN(MATCH_NOMATCH);              default: RRETURN(MATCH_NOMATCH);
4648    
4649              case 0x000d:              case CHAR_CR:
4650              if (eptr < md->end_subject && *eptr == 0x0a) eptr++;              if (eptr < md->end_subject && *eptr == CHAR_LF) eptr++;
4651              break;              break;
4652    
4653              case 0x000a:              case CHAR_LF:
4654              break;              break;
4655    
4656              case 0x000b:              case CHAR_VT:
4657              case 0x000c:              case CHAR_FF:
4658              case 0x0085:              case CHAR_NEL:
4659    #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
4660                case 0x2028:
4661                case 0x2029:
4662    #endif
4663              if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);              if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);
4664              break;              break;
4665              }              }
# Line 4409  for (;;) Line 4677  for (;;)
4677            switch(*eptr++)            switch(*eptr++)
4678              {              {
4679              default: break;              default: break;
4680              case 0x09:      /* HT */              HSPACE_BYTE_CASES:
4681              case 0x20:      /* SPACE */  #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
4682              case 0xa0:      /* NBSP */              HSPACE_MULTIBYTE_CASES:
4683    #endif
4684              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4685              }              }
4686            }            }
# Line 4428  for (;;) Line 4697  for (;;)
4697            switch(*eptr++)            switch(*eptr++)
4698              {              {
4699              default: RRETURN(MATCH_NOMATCH);              default: RRETURN(MATCH_NOMATCH);
4700              case 0x09:      /* HT */              HSPACE_BYTE_CASES:
4701              case 0x20:      /* SPACE */  #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
4702              case 0xa0:      /* NBSP */              HSPACE_MULTIBYTE_CASES:
4703    #endif
4704              break;              break;
4705              }              }
4706            }            }
# Line 4446  for (;;) Line 4716  for (;;)
4716              }              }
4717            switch(*eptr++)            switch(*eptr++)
4718              {              {
4719              default: break;              VSPACE_BYTE_CASES:
4720              case 0x0a:      /* LF */  #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
4721              case 0x0b:      /* VT */              VSPACE_MULTIBYTE_CASES:
4722              case 0x0c:      /* FF */  #endif
             case 0x0d:      /* CR */  
             case 0x85:      /* NEL */  
4723              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4724                default: break;
4725              }              }
4726            }            }
4727          break;          break;
# Line 4468  for (;;) Line 4737  for (;;)
4737            switch(*eptr++)            switch(*eptr++)
4738              {              {
4739              default: RRETURN(MATCH_NOMATCH);              default: RRETURN(MATCH_NOMATCH);
4740              case 0x0a:      /* LF */              VSPACE_BYTE_CASES:
4741              case 0x0b:      /* VT */  #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
4742              case 0x0c:      /* FF */              VSPACE_MULTIBYTE_CASES:
4743              case 0x0d:      /* CR */  #endif
             case 0x85:      /* NEL */  
4744              break;              break;
4745              }              }
4746            }            }
# Line 4486  for (;;) Line 4754  for (;;)
4754              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4755              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4756              }              }
4757            if ((md->ctypes[*eptr++] & ctype_digit) != 0) RRETURN(MATCH_NOMATCH);            if (MAX_255(*eptr) && (md->ctypes[*eptr] & ctype_digit) != 0)
4758                RRETURN(MATCH_NOMATCH);
4759              eptr++;
4760            }            }
4761          break;          break;
4762    
# Line 4498  for (;;) Line 4768  for (;;)
4768              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4769              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4770              }              }
4771            if ((md->ctypes[*eptr++] & ctype_digit) == 0) RRETURN(MATCH_NOMATCH);            if (!MAX_255(*eptr) || (md->ctypes[*eptr] & ctype_digit) == 0)
4772                RRETURN(MATCH_NOMATCH);
4773              eptr++;
4774            }            }
4775          break;          break;
4776    
# Line 4510  for (;;) Line 4782  for (;;)
4782              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4783              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4784              }              }
4785            if ((md->ctypes[*eptr++] & ctype_space) != 0) RRETURN(MATCH_NOMATCH);            if (MAX_255(*eptr) && (md->ctypes[*eptr] & ctype_space) != 0)
4786                RRETURN(MATCH_NOMATCH);
4787              eptr++;
4788            }            }
4789          break;          break;
4790    
# Line 4522  for (;;) Line 4796  for (;;)
4796              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4797              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4798              }              }
4799            if ((md->ctypes[*eptr++] & ctype_space) == 0) RRETURN(MATCH_NOMATCH);            if (!MAX_255(*eptr) || (md->ctypes[*eptr] & ctype_space) == 0)
4800                RRETURN(MATCH_NOMATCH);
4801              eptr++;
4802            }            }
4803          break;          break;
4804    
# Line 4534  for (;;) Line 4810  for (;;)
4810              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4811              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4812              }              }
4813            if ((md->ctypes[*eptr++] & ctype_word) != 0)            if (MAX_255(*eptr) && (md->ctypes[*eptr] & ctype_word) != 0)
4814              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4815              eptr++;
4816            }            }
4817          break;          break;
4818    
# Line 4547  for (;;) Line 4824  for (;;)
4824              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4825              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4826              }              }
4827            if ((md->ctypes[*eptr++] & ctype_word) == 0)            if (!MAX_255(*eptr) || (md->ctypes[*eptr] & ctype_word) == 0)
4828              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4829              eptr++;
4830            }            }
4831          break;          break;
4832    
# Line 4739  for (;;) Line 5017  for (;;)
5017              }              }
5018            /* Control never gets here */            /* Control never gets here */
5019    
5020            /* This should never occur */            case PT_CLIST:
5021              for (fi = min;; fi++)
5022                {
5023                const pcre_uint32 *cp;
5024                RMATCH(eptr, ecode, offset_top, md, eptrb, RM67);
5025                if (rrc != MATCH_NOMATCH) RRETURN(rrc);
5026                if (fi >= max) RRETURN(MATCH_NOMATCH);
5027                if (eptr >= md->end_subject)
5028                  {
5029                  SCHECK_PARTIAL();
5030                  RRETURN(MATCH_NOMATCH);
5031                  }
5032                GETCHARINCTEST(c, eptr);
5033                cp = PRIV(ucd_caseless_sets) + prop_value;
5034                for (;;)
5035                  {
5036                  if (c < *cp)
5037                    { if (prop_fail_result) break; else { RRETURN(MATCH_NOMATCH); } }
5038                  if (c == *cp++)
5039                    { if (prop_fail_result) { RRETURN(MATCH_NOMATCH); } else break; }
5040                  }
5041                }
5042              /* Control never gets here */
5043    
5044              case PT_UCNC:
5045              for (fi = min;; fi++)
5046                {
5047                RMATCH(eptr, ecode, offset_top, md, eptrb, RM68);
5048                if (rrc != MATCH_NOMATCH) RRETURN(rrc);
5049                if (fi >= max) RRETURN(MATCH_NOMATCH);
5050                if (eptr >= md->end_subject)
5051                  {
5052                  SCHECK_PARTIAL();
5053                  RRETURN(MATCH_NOMATCH);
5054                  }
5055                GETCHARINCTEST(c, eptr);
5056                if ((c == CHAR_DOLLAR_SIGN || c == CHAR_COMMERCIAL_AT ||
5057                     c == CHAR_GRAVE_ACCENT || (c >= 0xa0 && c <= 0xd7ff) ||
5058                     c >= 0xe000) == prop_fail_result)
5059                  RRETURN(MATCH_NOMATCH);
5060                }
5061              /* Control never gets here */
5062    
5063              /* This should never occur */
5064            default:            default:
5065            RRETURN(PCRE_ERROR_INTERNAL);            RRETURN(PCRE_ERROR_INTERNAL);
5066            }            }
# Line 4761  for (;;) Line 5081  for (;;)
5081              SCHECK_PARTIAL();              SCHECK_PARTIAL();
5082              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
5083              }              }
5084            GETCHARINCTEST(c, eptr);            else
           if (UCD_CATEGORY(c) == ucp_M) RRETURN(MATCH_NOMATCH);  
           while (eptr < md->end_subject)  
5085              {              {
5086              int len = 1;              int lgb, rgb;
5087              if (!utf8) c = *eptr; else { GETCHARLEN(c, eptr, len); }              GETCHARINCTEST(c, eptr);
5088              if (UCD_CATEGORY(c) != ucp_M) break;              lgb = UCD_GRAPHBREAK(c);
5089              eptr += len;              while (eptr < md->end_subject)
5090                  {
5091                  int len = 1;
5092                  if (!utf) c = *eptr; else { GETCHARLEN(c, eptr, len); }
5093                  rgb = UCD_GRAPHBREAK(c);
5094                  if ((PRIV(ucp_gbtable)[lgb] & (1 << rgb)) == 0) break;
5095                  lgb = rgb;
5096                  eptr += len;
5097                  }
5098              }              }
5099              CHECK_PARTIAL();
5100            }            }
5101          }          }
5102        else        else
5103  #endif     /* SUPPORT_UCP */  #endif     /* SUPPORT_UCP */
5104    
5105  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
5106        /* UTF-8 mode */        if (utf)
       if (utf8)  
5107          {          {
5108          for (fi = min;; fi++)          for (fi = min;; fi++)
5109            {            {
# Line 4794  for (;;) Line 5120  for (;;)
5120            GETCHARINC(c, eptr);            GETCHARINC(c, eptr);
5121            switch(ctype)            switch(ctype)
5122              {              {
5123              case OP_ANY:        /* This is the non-NL case */              case OP_ANY:               /* This is the non-NL case */
5124                if (md->partial != 0 &&    /* Take care with CRLF partial */
5125                    eptr >= md->end_subject &&
5126                    NLBLOCK->nltype == NLTYPE_FIXED &&
5127                    NLBLOCK->nllen == 2 &&
5128                    c == NLBLOCK->nl[0])
5129                  {
5130                  md->hitend = TRUE;
5131                  if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
5132                  }
5133                break;
5134    
5135              case OP_ALLANY:              case OP_ALLANY:
5136              case OP_ANYBYTE:              case OP_ANYBYTE:
5137              break;              break;
# Line 4803  for (;;) Line 5140  for (;;)
5140              switch(c)              switch(c)
5141                {                {
5142                default: RRETURN(MATCH_NOMATCH);                default: RRETURN(MATCH_NOMATCH);
5143                case 0x000d:                case CHAR_CR:
5144                if (eptr < md->end_subject && *eptr == 0x0a) eptr++;                if (eptr < md->end_subject && RAWUCHAR(eptr) == CHAR_LF) eptr++;
5145                break;                break;
5146                case 0x000a:  
5147                  case CHAR_LF:
5148                break;                break;
5149    
5150                case 0x000b:                case CHAR_VT:
5151                case 0x000c:                case CHAR_FF:
5152                case 0x0085:                case CHAR_NEL:
5153    #ifndef EBCDIC
5154                case 0x2028:                case 0x2028:
5155                case 0x2029:                case 0x2029:
5156    #endif  /* Not EBCDIC */
5157                if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);                if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);
5158                break;                break;
5159                }                }
# Line 4822  for (;;) Line 5162  for (;;)
5162              case OP_NOT_HSPACE:              case OP_NOT_HSPACE:
5163              switch(c)              switch(c)
5164                {                {
5165                  HSPACE_CASES: RRETURN(MATCH_NOMATCH);
5166                default: break;                default: break;
               case 0x09:      /* HT */  
               case 0x20:      /* SPACE */  
               case 0xa0:      /* NBSP */  
               case 0x1680:    /* OGHAM SPACE MARK */  
               case 0x180e:    /* MONGOLIAN VOWEL SEPARATOR */  
               case 0x2000:    /* EN QUAD */  
               case 0x2001:    /* EM QUAD */  
               case 0x2002:    /* EN SPACE */  
               case 0x2003:    /* EM SPACE */  
               case 0x2004:    /* THREE-PER-EM SPACE */  
               case 0x2005:    /* FOUR-PER-EM SPACE */  
               case 0x2006:    /* SIX-PER-EM SPACE */  
               case 0x2007:    /* FIGURE SPACE */  
               case 0x2008:    /* PUNCTUATION SPACE */  
               case 0x2009:    /* THIN SPACE */  
               case 0x200A:    /* HAIR SPACE */  
               case 0x202f:    /* NARROW NO-BREAK SPACE */  
               case 0x205f:    /* MEDIUM MATHEMATICAL SPACE */  
               case 0x3000:    /* IDEOGRAPHIC SPACE */  
               RRETURN(MATCH_NOMATCH);  
5167                }                }
5168              break;              break;
5169    
5170              case OP_HSPACE:              case OP_HSPACE:
5171              switch(c)              switch(c)
5172                {                {
5173                  HSPACE_CASES: break;
5174                default: RRETURN(MATCH_NOMATCH);                default: RRETURN(MATCH_NOMATCH);
               case 0x09:      /* HT */  
               case 0x20:      /* SPACE */  
               case 0xa0:      /* NBSP */  
               case 0x1680:    /* OGHAM SPACE MARK */  
               case 0x180e:    /* MONGOLIAN VOWEL SEPARATOR */  
               case 0x2000:    /* EN QUAD */  
               case 0x2001:    /* EM QUAD */  
               case 0x2002:    /* EN SPACE */  
               case 0x2003:    /* EM SPACE */  
               case 0x2004:    /* THREE-PER-EM SPACE */  
               case 0x2005:    /* FOUR-PER-EM SPACE */  
               case 0x2006:    /* SIX-PER-EM SPACE */  
               case 0x2007:    /* FIGURE SPACE */  
               case 0x2008:    /* PUNCTUATION SPACE */  
               case 0x2009:    /* THIN SPACE */  
               case 0x200A:    /* HAIR SPACE */  
               case 0x202f:    /* NARROW NO-BREAK SPACE */  
               case 0x205f:    /* MEDIUM MATHEMATICAL SPACE */  
               case 0x3000:    /* IDEOGRAPHIC SPACE */  
               break;  
5175                }                }
5176              break;              break;
5177    
5178              case OP_NOT_VSPACE:              case OP_NOT_VSPACE:
5179              switch(c)              switch(c)
5180                {                {
5181                  VSPACE_CASES: RRETURN(MATCH_NOMATCH);
5182                default: break;                default: break;
               case 0x0a:      /* LF */  
               case 0x0b:      /* VT */  
               case 0x0c:      /* FF */  
               case 0x0d:      /* CR */  
               case 0x85:      /* NEL */  
               case 0x2028:    /* LINE SEPARATOR */  
               case 0x2029:    /* PARAGRAPH SEPARATOR */  
               RRETURN(MATCH_NOMATCH);  
5183                }                }
5184              break;              break;
5185    
5186              case OP_VSPACE:              case OP_VSPACE:
5187              switch(c)              switch(c)
5188                {                {
5189                  VSPACE_CASES: break;
5190                default: RRETURN(MATCH_NOMATCH);                default: RRETURN(MATCH_NOMATCH);
               case 0x0a:      /* LF */  
               case 0x0b:      /* VT */  
               case 0x0c:      /* FF */  
               case 0x0d:      /* CR */  
               case 0x85:      /* NEL */  
               case 0x2028:    /* LINE SEPARATOR */  
               case 0x2029:    /* PARAGRAPH SEPARATOR */  
               break;  
5191                }                }
5192              break;              break;
5193    
# Line 4919  for (;;) Line 5207  for (;;)
5207              break;              break;
5208    
5209              case OP_WHITESPACE:              case OP_WHITESPACE:
5210              if  (c >= 256 || (md->ctypes[c] & ctype_space) == 0)              if (c >= 256 || (md->ctypes[c] & ctype_space) == 0)
5211                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
5212              break;              break;
5213    
# Line 4940  for (;;) Line 5228  for (;;)
5228          }          }
5229        else        else
5230  #endif  #endif
5231        /* Not UTF-8 mode */        /* Not UTF mode */
5232          {          {
5233          for (fi = min;; fi++)          for (fi = min;; fi++)
5234            {            {
# Line 4957  for (;;) Line 5245  for (;;)
5245            c = *eptr++;            c = *eptr++;
5246            switch(ctype)            switch(ctype)
5247              {              {
5248              case OP_ANY:     /* This is the non-NL case */              case OP_ANY:               /* This is the non-NL case */
5249                if (md->partial != 0 &&    /* Take care with CRLF partial */
5250                    eptr >= md->end_subject &&
5251                    NLBLOCK->nltype == NLTYPE_FIXED &&
5252                    NLBLOCK->nllen == 2 &&
5253                    c == NLBLOCK->nl[0])
5254                  {
5255                  md->hitend = TRUE;
5256                  if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
5257                  }
5258                break;
5259    
5260              case OP_ALLANY:              case OP_ALLANY:
5261              case OP_ANYBYTE:              case OP_ANYBYTE:
5262              break;              break;
# Line 4966  for (;;) Line 5265  for (;;)
5265              switch(c)              switch(c)
5266                {                {
5267                default: RRETURN(MATCH_NOMATCH);                default: RRETURN(MATCH_NOMATCH);
5268                case 0x000d:                case CHAR_CR:
5269                if (eptr < md->end_subject && *eptr == 0x0a) eptr++;                if (eptr < md->end_subject && *eptr == CHAR_LF) eptr++;
5270                break;                break;
5271    
5272                case 0x000a:                case CHAR_LF:
5273                break;                break;
5274    
5275                case 0x000b:                case CHAR_VT:
5276                case 0x000c:                case CHAR_FF:
5277                case 0x0085:                case CHAR_NEL:
5278    #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
5279                  case 0x2028:
5280                  case 0x2029:
5281    #endif
5282                if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);                if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);
5283                break;                break;
5284                }                }
# Line 4985  for (;;) Line 5288  for (;;)
5288              switch(c)              switch(c)
5289                {                {
5290                default: break;                default: break;
5291                case 0x09:      /* HT */                HSPACE_BYTE_CASES:
5292                case 0x20:      /* SPACE */  #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
5293                case 0xa0:      /* NBSP */                HSPACE_MULTIBYTE_CASES:
5294    #endif
5295                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
5296                }                }
5297              break;              break;
# Line 4996  for (;;) Line 5300  for (;;)
5300              switch(c)              switch(c)
5301                {                {
5302                default: RRETURN(MATCH_NOMATCH);                default: RRETURN(MATCH_NOMATCH);
5303                case 0x09:      /* HT */                HSPACE_BYTE_CASES:
5304                case 0x20:      /* SPACE */  #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
5305                case 0xa0:      /* NBSP */                HSPACE_MULTIBYTE_CASES:
5306    #endif
5307                break;                break;
5308                }                }
5309              break;              break;
# Line 5007  for (;;) Line 5312  for (;;)
5312              switch(c)              switch(c)
5313                {                {
5314                default: break;                default: break;
5315                case 0x0a:      /* LF */                VSPACE_BYTE_CASES:
5316                case 0x0b:      /* VT */  #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
5317                case 0x0c:      /* FF */                VSPACE_MULTIBYTE_CASES:
5318                case 0x0d:      /* CR */  #endif
               case 0x85:      /* NEL */  
5319                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
5320                }                }
5321              break;              break;
# Line 5020  for (;;) Line 5324  for (;;)
5324              switch(c)              switch(c)
5325                {                {
5326                default: RRETURN(MATCH_NOMATCH);                default: RRETURN(MATCH_NOMATCH);
5327                case 0x0a:      /* LF */                VSPACE_BYTE_CASES:
5328                case 0x0b:      /* VT */  #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
5329                case 0x0c:      /* FF */                VSPACE_MULTIBYTE_CASES:
5330                case 0x0d:      /* CR */  #endif
               case 0x85:      /* NEL */  
5331                break;                break;
5332                }                }
5333              break;              break;
5334    
5335              case OP_NOT_DIGIT:              case OP_NOT_DIGIT:
5336              if ((md->ctypes[c] & ctype_digit) != 0) RRETURN(MATCH_NOMATCH);              if (MAX_255(c) && (md->ctypes[c] & ctype_digit) != 0) RRETURN(MATCH_NOMATCH);
5337              break;              break;
5338    
5339              case OP_DIGIT:              case OP_DIGIT:
5340              if ((md->ctypes[c] & ctype_digit) == 0) RRETURN(MATCH_NOMATCH);              if (!MAX_255(c) || (md->ctypes[c] & ctype_digit) == 0) RRETURN(MATCH_NOMATCH);
5341              break;              break;
5342    
5343              case OP_NOT_WHITESPACE:              case OP_NOT_WHITESPACE:
5344              if ((md->ctypes[c] & ctype_space) != 0) RRETURN(MATCH_NOMATCH);              if (MAX_255(c) && (md->ctypes[c] & ctype_space) != 0) RRETURN(MATCH_NOMATCH);
5345              break;              break;
5346    
5347              case OP_WHITESPACE:              case OP_WHITESPACE:
5348              if  ((md->ctypes[c] & ctype_space) == 0) RRETURN(MATCH_NOMATCH);              if (!MAX_255(c) || (md->ctypes[c] & ctype_space) == 0) RRETURN(MATCH_NOMATCH);
5349              break;              break;
5350    
5351              case OP_NOT_WORDCHAR:              case OP_NOT_WORDCHAR:
5352              if ((md->ctypes[c] & ctype_word) != 0) RRETURN(MATCH_NOMATCH);              if (MAX_255(c) && (md->ctypes[c] & ctype_word) != 0) RRETURN(MATCH_NOMATCH);
5353              break;              break;
5354    
5355              case OP_WORDCHAR:              case OP_WORDCHAR:
5356              if ((md->ctypes[c] & ctype_word) == 0) RRETURN(MATCH_NOMATCH);              if (!MAX_255(c) || (md->ctypes[c] & ctype_word) == 0) RRETURN(MATCH_NOMATCH);
5357              break;              break;
5358    
5359              default:              default:
# Line 5227  for (;;) Line 5530  for (;;)
5530              }              }
5531            break;            break;
5532    
5533              case PT_CLIST:
5534              for (i = min; i < max; i++)
5535                {
5536                const pcre_uint32 *cp;
5537                int len = 1;
5538                if (eptr >= md->end_subject)
5539                  {
5540                  SCHECK_PARTIAL();
5541                  break;
5542                  }
5543                GETCHARLENTEST(c, eptr, len);
5544                cp = PRIV(ucd_caseless_sets) + prop_value;
5545                for (;;)
5546                  {
5547                  if (c < *cp)
5548                    { if (prop_fail_result) break; else goto GOT_MAX; }
5549                  if (c == *cp++)
5550                    { if (prop_fail_result) goto GOT_MAX; else break; }
5551                  }
5552                eptr += len;
5553                }
5554              GOT_MAX:
5555              break;
5556    
5557              case PT_UCNC:
5558              for (i = min; i < max; i++)
5559                {
5560                int len = 1;
5561                if (eptr >= md->end_subject)
5562                  {
5563                  SCHECK_PARTIAL();
5564                  break;
5565                  }
5566                GETCHARLENTEST(c, eptr, len);
5567                if ((c == CHAR_DOLLAR_SIGN || c == CHAR_COMMERCIAL_AT ||
5568                     c == CHAR_GRAVE_ACCENT || (c >= 0xa0 && c <= 0xd7ff) ||
5569                     c >= 0xe000) == prop_fail_result)
5570                  break;
5571                eptr += len;
5572                }
5573              break;
5574    
5575            default:            default:
5576            RRETURN(PCRE_ERROR_INTERNAL);            RRETURN(PCRE_ERROR_INTERNAL);
5577            }            }
# Line 5239  for (;;) Line 5584  for (;;)
5584            RMATCH(eptr, ecode, offset_top, md, eptrb, RM44);            RMATCH(eptr, ecode, offset_top, md, eptrb, RM44);
5585            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
5586            if (eptr-- == pp) break;        /* Stop if tried at original pos */            if (eptr-- == pp) break;        /* Stop if tried at original pos */
5587            if (utf8) BACKCHAR(eptr);            if (utf) BACKCHAR(eptr);
5588            }            }
5589          }          }
5590    
# Line 5250  for (;;) Line 5595  for (;;)
5595          {          {
5596          for (i = min; i < max; i++)          for (i = min; i < max; i++)
5597            {            {
           int len = 1;  
5598            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
5599              {              {
5600              SCHECK_PARTIAL();              SCHECK_PARTIAL();
5601              break;              break;
5602              }              }
5603            if (!utf8) c = *eptr; else { GETCHARLEN(c, eptr, len); }            else
           if (UCD_CATEGORY(c) == ucp_M) break;  
           eptr += len;  
           while (eptr < md->end_subject)