/[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 1274 by ph10, Fri Mar 8 11:35:41 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 715  for (;;) Line 784  for (;;)
784      /* COMMIT overrides PRUNE, SKIP, and THEN */      /* COMMIT overrides PRUNE, SKIP, and THEN */
785    
786      case OP_COMMIT:      case OP_COMMIT:
787      RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,      RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md,
788        eptrb, RM52);        eptrb, RM52);
789      if (rrc != MATCH_NOMATCH && rrc != MATCH_PRUNE &&      if (rrc != MATCH_NOMATCH && rrc != MATCH_PRUNE &&
790          rrc != MATCH_SKIP && rrc != MATCH_SKIP_ARG &&          rrc != MATCH_SKIP && rrc != MATCH_SKIP_ARG &&
# Line 726  for (;;) Line 795  for (;;)
795      /* PRUNE overrides THEN */      /* PRUNE overrides THEN */
796    
797      case OP_PRUNE:      case OP_PRUNE:
798      RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,      RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md,
799        eptrb, RM51);        eptrb, RM51);
800      if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN) RRETURN(rrc);      if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN) RRETURN(rrc);
801      RRETURN(MATCH_PRUNE);      RRETURN(MATCH_PRUNE);
# Line 734  for (;;) Line 803  for (;;)
803      case OP_PRUNE_ARG:      case OP_PRUNE_ARG:
804      md->nomatch_mark = ecode + 2;      md->nomatch_mark = ecode + 2;
805      md->mark = NULL;    /* In case previously set by assertion */      md->mark = NULL;    /* In case previously set by assertion */
806      RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode] + ecode[1], offset_top, md,      RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode] + ecode[1], offset_top, md,
807        eptrb, RM56);        eptrb, RM56);
808      if ((rrc == MATCH_MATCH || rrc == MATCH_ACCEPT) &&      if ((rrc == MATCH_MATCH || rrc == MATCH_ACCEPT) &&
809           md->mark == NULL) md->mark = ecode + 2;           md->mark == NULL) md->mark = ecode + 2;
# Line 744  for (;;) Line 813  for (;;)
813      /* SKIP overrides PRUNE and THEN */      /* SKIP overrides PRUNE and THEN */
814    
815      case OP_SKIP:      case OP_SKIP:
816      RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,      RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md,
817        eptrb, RM53);        eptrb, RM53);
818      if (rrc != MATCH_NOMATCH && rrc != MATCH_PRUNE && rrc != MATCH_THEN)      if (rrc != MATCH_NOMATCH && rrc != MATCH_PRUNE && rrc != MATCH_THEN)
819        RRETURN(rrc);        RRETURN(rrc);
# Line 752  for (;;) Line 821  for (;;)
821      RRETURN(MATCH_SKIP);      RRETURN(MATCH_SKIP);
822    
823      /* Note that, for Perl compatibility, SKIP with an argument does NOT set      /* Note that, for Perl compatibility, SKIP with an argument does NOT set
824      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
825      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
826        that failed and any that preceed it (either they also failed, or were not
827        triggered). To do this, we maintain a count of executed SKIP_ARGs. If a
828        SKIP_ARG gets to top level, the match is re-run with md->ignore_skip_arg
829        set to the count of the one that failed. */
830    
831      case OP_SKIP_ARG:      case OP_SKIP_ARG:
832      if (md->ignore_skip_arg)      md->skip_arg_count++;
833        if (md->skip_arg_count <= md->ignore_skip_arg)
834        {        {
835        ecode += _pcre_OP_lengths[*ecode] + ecode[1];        ecode += PRIV(OP_lengths)[*ecode] + ecode[1];
836        break;        break;
837        }        }
838      RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode] + ecode[1], offset_top, md,      RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode] + ecode[1], offset_top, md,
839        eptrb, RM57);        eptrb, RM57);
840      if (rrc != MATCH_NOMATCH && rrc != MATCH_PRUNE && rrc != MATCH_THEN)      if (rrc != MATCH_NOMATCH && rrc != MATCH_PRUNE && rrc != MATCH_THEN)
841        RRETURN(rrc);        RRETURN(rrc);
842    
843      /* 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
844      returning the special MATCH_SKIP_ARG return code. This will either be      returning the special MATCH_SKIP_ARG return code. This will either be
845      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
846      with the md->ignore_skip_arg flag set. */      with md->ignore_skip_arg set to the value of md->skip_arg_count. */
847    
848      md->start_match_ptr = ecode + 2;      md->start_match_ptr = ecode + 2;
849      RRETURN(MATCH_SKIP_ARG);      RRETURN(MATCH_SKIP_ARG);
# Line 779  for (;;) Line 853  for (;;)
853      match pointer to do this. */      match pointer to do this. */
854    
855      case OP_THEN:      case OP_THEN:
856      RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,      RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md,
857        eptrb, RM54);        eptrb, RM54);
858      if (rrc != MATCH_NOMATCH) RRETURN(rrc);      if (rrc != MATCH_NOMATCH) RRETURN(rrc);
859      md->start_match_ptr = ecode;      md->start_match_ptr = ecode;
# Line 788  for (;;) Line 862  for (;;)
862      case OP_THEN_ARG:      case OP_THEN_ARG:
863      md->nomatch_mark = ecode + 2;      md->nomatch_mark = ecode + 2;
864      md->mark = NULL;    /* In case previously set by assertion */      md->mark = NULL;    /* In case previously set by assertion */
865      RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode] + ecode[1], offset_top,      RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode] + ecode[1], offset_top,
866        md, eptrb, RM58);        md, eptrb, RM58);
867      if ((rrc == MATCH_MATCH || rrc == MATCH_ACCEPT) &&      if ((rrc == MATCH_MATCH || rrc == MATCH_ACCEPT) &&
868           md->mark == NULL) md->mark = ecode + 2;           md->mark == NULL) md->mark = ecode + 2;
# Line 812  for (;;) Line 886  for (;;)
886      case OP_ONCE_NC:      case OP_ONCE_NC:
887      prev = ecode;      prev = ecode;
888      saved_eptr = eptr;      saved_eptr = eptr;
889        save_mark = md->mark;
890      do      do
891        {        {
892        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 905  for (;;)
905    
906        if (rrc != MATCH_NOMATCH) RRETURN(rrc);        if (rrc != MATCH_NOMATCH) RRETURN(rrc);
907        ecode += GET(ecode,1);        ecode += GET(ecode,1);
908          md->mark = save_mark;
909        }        }
910      while (*ecode == OP_ALT);      while (*ecode == OP_ALT);
911    
# Line 869  for (;;) Line 945  for (;;)
945        }        }
946      else  /* OP_KETRMAX */      else  /* OP_KETRMAX */
947        {        {
       md->match_function_type = MATCH_CBEGROUP;  
948        RMATCH(eptr, prev, offset_top, md, eptrb, RM66);        RMATCH(eptr, prev, offset_top, md, eptrb, RM66);
949        if (rrc != MATCH_NOMATCH) RRETURN(rrc);        if (rrc != MATCH_NOMATCH) RRETURN(rrc);
950        ecode += 1 + LINK_SIZE;        ecode += 1 + LINK_SIZE;
# Line 909  for (;;) Line 984  for (;;)
984        save_offset2 = md->offset_vector[offset+1];        save_offset2 = md->offset_vector[offset+1];
985        save_offset3 = md->offset_vector[md->offset_end - number];        save_offset3 = md->offset_vector[md->offset_end - number];
986        save_capture_last = md->capture_last;        save_capture_last = md->capture_last;
987          save_mark = md->mark;
988    
989        DPRINTF(("saving %d %d %d\n", save_offset1, save_offset2, save_offset3));        DPRINTF(("saving %d %d %d\n", save_offset1, save_offset2, save_offset3));
990        md->offset_vector[md->offset_end - number] =        md->offset_vector[md->offset_end - number] =
# Line 917  for (;;) Line 993  for (;;)
993        for (;;)        for (;;)
994          {          {
995          if (op >= OP_SBRA) md->match_function_type = MATCH_CBEGROUP;          if (op >= OP_SBRA) md->match_function_type = MATCH_CBEGROUP;
996          RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,          RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md,
997            eptrb, RM1);            eptrb, RM1);
998          if (rrc == MATCH_ONCE) break;  /* Backing up through an atomic group */          if (rrc == MATCH_ONCE) break;  /* Backing up through an atomic group */
999    
# Line 945  for (;;) Line 1021  for (;;)
1021          if (rrc != MATCH_NOMATCH) RRETURN(rrc);          if (rrc != MATCH_NOMATCH) RRETURN(rrc);
1022          md->capture_last = save_capture_last;          md->capture_last = save_capture_last;
1023          ecode += GET(ecode, 1);          ecode += GET(ecode, 1);
1024            md->mark = save_mark;
1025          if (*ecode != OP_ALT) break;          if (*ecode != OP_ALT) break;
1026          }          }
1027    
# Line 996  for (;;) Line 1073  for (;;)
1073    
1074      for (;;)      for (;;)
1075        {        {
1076        if (op >= OP_SBRA || op == OP_ONCE) md->match_function_type = MATCH_CBEGROUP;        if (op >= OP_SBRA || op == OP_ONCE)
1077            md->match_function_type = MATCH_CBEGROUP;
1078    
1079        /* 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
1080        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 1082  for (;;)
1082    
1083        else if (!md->hasthen && ecode[GET(ecode, 1)] != OP_ALT)        else if (!md->hasthen && ecode[GET(ecode, 1)] != OP_ALT)
1084          {          {
1085          ecode += _pcre_OP_lengths[*ecode];          ecode += PRIV(OP_lengths)[*ecode];
1086          goto TAIL_RECURSE;          goto TAIL_RECURSE;
1087          }          }
1088    
1089        /* In all other cases, we have to make another call to match(). */        /* In all other cases, we have to make another call to match(). */
1090    
1091        RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md, eptrb,        save_mark = md->mark;
1092          save_capture_last = md->capture_last;
1093          RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md, eptrb,
1094          RM2);          RM2);
1095    
1096        /* 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 1108  for (;;)
1108          {          {
1109          if (rrc == MATCH_ONCE)          if (rrc == MATCH_ONCE)
1110            {            {
1111            const uschar *scode = ecode;            const pcre_uchar *scode = ecode;
1112            if (*scode != OP_ONCE)           /* If not at start, find it */            if (*scode != OP_ONCE)           /* If not at start, find it */
1113              {              {
1114              while (*scode == OP_ALT) scode += GET(scode, 1);              while (*scode == OP_ALT) scode += GET(scode, 1);
# Line 1039  for (;;) Line 1119  for (;;)
1119          RRETURN(rrc);          RRETURN(rrc);
1120          }          }
1121        ecode += GET(ecode, 1);        ecode += GET(ecode, 1);
1122          md->mark = save_mark;
1123        if (*ecode != OP_ALT) break;        if (*ecode != OP_ALT) break;
1124          md->capture_last = save_capture_last;
1125        }        }
1126    
1127      RRETURN(MATCH_NOMATCH);      RRETURN(MATCH_NOMATCH);
# Line 1093  for (;;) Line 1175  for (;;)
1175          md->offset_vector[md->offset_end - number] =          md->offset_vector[md->offset_end - number] =
1176            (int)(eptr - md->start_subject);            (int)(eptr - md->start_subject);
1177          if (op >= OP_SBRA) md->match_function_type = MATCH_CBEGROUP;          if (op >= OP_SBRA) md->match_function_type = MATCH_CBEGROUP;
1178          RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,          RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md,
1179            eptrb, RM63);            eptrb, RM63);
1180          if (rrc == MATCH_KETRPOS)          if (rrc == MATCH_KETRPOS)
1181            {            {
# Line 1161  for (;;) Line 1243  for (;;)
1243      POSSESSIVE_NON_CAPTURE:      POSSESSIVE_NON_CAPTURE:
1244      matched_once = FALSE;      matched_once = FALSE;
1245      code_offset = (int)(ecode - md->start_code);      code_offset = (int)(ecode - md->start_code);
1246        save_capture_last = md->capture_last;
1247    
1248      for (;;)      for (;;)
1249        {        {
1250        if (op >= OP_SBRA) md->match_function_type = MATCH_CBEGROUP;        if (op >= OP_SBRA) md->match_function_type = MATCH_CBEGROUP;
1251        RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,        RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md,
1252          eptrb, RM48);          eptrb, RM48);
1253        if (rrc == MATCH_KETRPOS)        if (rrc == MATCH_KETRPOS)
1254          {          {
# Line 1190  for (;;) Line 1273  for (;;)
1273        if (rrc != MATCH_NOMATCH) RRETURN(rrc);        if (rrc != MATCH_NOMATCH) RRETURN(rrc);
1274        ecode += GET(ecode, 1);        ecode += GET(ecode, 1);
1275        if (*ecode != OP_ALT) break;        if (*ecode != OP_ALT) break;
1276          md->capture_last = save_capture_last;
1277        }        }
1278    
1279      if (matched_once || allow_zero)      if (matched_once || allow_zero)
# Line 1215  for (;;) Line 1299  for (;;)
1299    
1300      if (ecode[LINK_SIZE+1] == OP_CALLOUT)      if (ecode[LINK_SIZE+1] == OP_CALLOUT)
1301        {        {
1302        if (pcre_callout != NULL)        if (PUBL(callout) != NULL)
1303          {          {
1304          pcre_callout_block cb;          PUBL(callout_block) cb;
1305          cb.version          = 2;   /* Version 1 of the callout block */          cb.version          = 2;   /* Version 1 of the callout block */
1306          cb.callout_number   = ecode[LINK_SIZE+2];          cb.callout_number   = ecode[LINK_SIZE+2];
1307          cb.offset_vector    = md->offset_vector;          cb.offset_vector    = md->offset_vector;
1308    #if defined COMPILE_PCRE8
1309          cb.subject          = (PCRE_SPTR)md->start_subject;          cb.subject          = (PCRE_SPTR)md->start_subject;
1310    #elif defined COMPILE_PCRE16
1311            cb.subject          = (PCRE_SPTR16)md->start_subject;
1312    #elif defined COMPILE_PCRE32
1313            cb.subject          = (PCRE_SPTR32)md->start_subject;
1314    #endif
1315          cb.subject_length   = (int)(md->end_subject - md->start_subject);          cb.subject_length   = (int)(md->end_subject - md->start_subject);
1316          cb.start_match      = (int)(mstart - md->start_subject);          cb.start_match      = (int)(mstart - md->start_subject);
1317          cb.current_position = (int)(eptr - md->start_subject);          cb.current_position = (int)(eptr - md->start_subject);
1318          cb.pattern_position = GET(ecode, LINK_SIZE + 3);          cb.pattern_position = GET(ecode, LINK_SIZE + 3);
1319          cb.next_item_length = GET(ecode, 3 + 2*LINK_SIZE);          cb.next_item_length = GET(ecode, 3 + 2*LINK_SIZE);
1320          cb.capture_top      = offset_top/2;          cb.capture_top      = offset_top/2;
1321          cb.capture_last     = md->capture_last;          cb.capture_last     = md->capture_last & CAPLMASK;
1322            /* Internal change requires this for API compatibility. */
1323            if (cb.capture_last == 0) cb.capture_last = -1;
1324          cb.callout_data     = md->callout_data;          cb.callout_data     = md->callout_data;
1325          cb.mark             = md->nomatch_mark;          cb.mark             = md->nomatch_mark;
1326          if ((rrc = (*pcre_callout)(&cb)) > 0) RRETURN(MATCH_NOMATCH);          if ((rrc = (*PUBL(callout))(&cb)) > 0) RRETURN(MATCH_NOMATCH);
1327          if (rrc < 0) RRETURN(rrc);          if (rrc < 0) RRETURN(rrc);
1328          }          }
1329        ecode += _pcre_OP_lengths[OP_CALLOUT];        ecode += PRIV(OP_lengths)[OP_CALLOUT];
1330          codelink -= PRIV(OP_lengths)[OP_CALLOUT];
1331        }        }
1332    
1333      condcode = ecode[LINK_SIZE+1];      condcode = ecode[LINK_SIZE+1];
# Line 1250  for (;;) Line 1343  for (;;)
1343          }          }
1344        else        else
1345          {          {
1346          int recno = GET2(ecode, LINK_SIZE + 2);   /* Recursion group number*/          unsigned int recno = GET2(ecode, LINK_SIZE + 2);   /* Recursion group number*/
1347          condition = (recno == RREF_ANY || recno == md->recursive->group_num);          condition = (recno == RREF_ANY || recno == md->recursive->group_num);
1348    
1349          /* 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 1353  for (;;)
1353    
1354          if (!condition && condcode == OP_NRREF)          if (!condition && condcode == OP_NRREF)
1355            {            {
1356            uschar *slotA = md->name_table;            pcre_uchar *slotA = md->name_table;
1357            for (i = 0; i < md->name_count; i++)            for (i = 0; i < md->name_count; i++)
1358              {              {
1359              if (GET2(slotA, 0) == recno) break;              if (GET2(slotA, 0) == recno) break;
# Line 1273  for (;;) Line 1366  for (;;)
1366    
1367            if (i < md->name_count)            if (i < md->name_count)
1368              {              {
1369              uschar *slotB = slotA;              pcre_uchar *slotB = slotA;
1370              while (slotB > md->name_table)              while (slotB > md->name_table)
1371                {                {
1372                slotB -= md->name_entry_size;                slotB -= md->name_entry_size;
1373                if (strcmp((char *)slotA + 2, (char *)slotB + 2) == 0)                if (STRCMP_UC_UC(slotA + IMM2_SIZE, slotB + IMM2_SIZE) == 0)
1374                  {                  {
1375                  condition = GET2(slotB, 0) == md->recursive->group_num;                  condition = GET2(slotB, 0) == md->recursive->group_num;
1376                  if (condition) break;                  if (condition) break;
# Line 1293  for (;;) Line 1386  for (;;)
1386                for (i++; i < md->name_count; i++)                for (i++; i < md->name_count; i++)
1387                  {                  {
1388                  slotB += md->name_entry_size;                  slotB += md->name_entry_size;
1389                  if (strcmp((char *)slotA + 2, (char *)slotB + 2) == 0)                  if (STRCMP_UC_UC(slotA + IMM2_SIZE, slotB + IMM2_SIZE) == 0)
1390                    {                    {
1391                    condition = GET2(slotB, 0) == md->recursive->group_num;                    condition = GET2(slotB, 0) == md->recursive->group_num;
1392                    if (condition) break;                    if (condition) break;
# Line 1306  for (;;) Line 1399  for (;;)
1399    
1400          /* Chose branch according to the condition */          /* Chose branch according to the condition */
1401    
1402          ecode += condition? 3 : GET(ecode, 1);          ecode += condition? 1 + IMM2_SIZE : GET(ecode, 1);
1403          }          }
1404        }        }
1405    
# Line 1322  for (;;) Line 1415  for (;;)
1415    
1416        if (!condition && condcode == OP_NCREF)        if (!condition && condcode == OP_NCREF)
1417          {          {
1418          int refno = offset >> 1;          unsigned int refno = offset >> 1;
1419          uschar *slotA = md->name_table;          pcre_uchar *slotA = md->name_table;
1420    
1421          for (i = 0; i < md->name_count; i++)          for (i = 0; i < md->name_count; i++)
1422            {            {
# Line 1337  for (;;) Line 1430  for (;;)
1430    
1431          if (i < md->name_count)          if (i < md->name_count)
1432            {            {
1433            uschar *slotB = slotA;            pcre_uchar *slotB = slotA;
1434            while (slotB > md->name_table)            while (slotB > md->name_table)
1435              {              {
1436              slotB -= md->name_entry_size;              slotB -= md->name_entry_size;
1437              if (strcmp((char *)slotA + 2, (char *)slotB + 2) == 0)              if (STRCMP_UC_UC(slotA + IMM2_SIZE, slotB + IMM2_SIZE) == 0)
1438                {                {
1439                offset = GET2(slotB, 0) << 1;                offset = GET2(slotB, 0) << 1;
1440                condition = offset < offset_top &&                condition = offset < offset_top &&
# Line 1359  for (;;) Line 1452  for (;;)
1452              for (i++; i < md->name_count; i++)              for (i++; i < md->name_count; i++)
1453                {                {
1454                slotB += md->name_entry_size;                slotB += md->name_entry_size;
1455                if (strcmp((char *)slotA + 2, (char *)slotB + 2) == 0)                if (STRCMP_UC_UC(slotA + IMM2_SIZE, slotB + IMM2_SIZE) == 0)
1456                  {                  {
1457                  offset = GET2(slotB, 0) << 1;                  offset = GET2(slotB, 0) << 1;
1458                  condition = offset < offset_top &&                  condition = offset < offset_top &&
# Line 1374  for (;;) Line 1467  for (;;)
1467    
1468        /* Chose branch according to the condition */        /* Chose branch according to the condition */
1469    
1470        ecode += condition? 3 : GET(ecode, 1);        ecode += condition? 1 + IMM2_SIZE : GET(ecode, 1);
1471        }        }
1472    
1473      else if (condcode == OP_DEF)     /* DEFINE - always false */      else if (condcode == OP_DEF)     /* DEFINE - always false */
# Line 1450  for (;;) Line 1543  for (;;)
1543      to close any currently open capturing brackets. */      to close any currently open capturing brackets. */
1544    
1545      case OP_CLOSE:      case OP_CLOSE:
1546      number = GET2(ecode, 1);      number = GET2(ecode, 1);   /* Must be less than 65536 */
1547      offset = number << 1;      offset = number << 1;
1548    
1549  #ifdef PCRE_DEBUG  #ifdef PCRE_DEBUG
# Line 1458  for (;;) Line 1551  for (;;)
1551        printf("\n");        printf("\n");
1552  #endif  #endif
1553    
1554      md->capture_last = number;      md->capture_last = (md->capture_last & OVFLMASK) | number;
1555      if (offset >= md->offset_max) md->offset_overflow = TRUE; else      if (offset >= md->offset_max) md->capture_last |= OVFLBIT; else
1556        {        {
1557        md->offset_vector[offset] =        md->offset_vector[offset] =
1558          md->offset_vector[md->offset_end - number];          md->offset_vector[md->offset_end - number];
1559        md->offset_vector[offset+1] = (int)(eptr - md->start_subject);        md->offset_vector[offset+1] = (int)(eptr - md->start_subject);
1560        if (offset_top <= offset) offset_top = offset + 2;        if (offset_top <= offset) offset_top = offset + 2;
1561        }        }
1562      ecode += 3;      ecode += 1 + IMM2_SIZE;
1563      break;      break;
1564    
1565    
# Line 1513  for (;;) Line 1606  for (;;)
1606    
1607      case OP_ASSERT:      case OP_ASSERT:
1608      case OP_ASSERTBACK:      case OP_ASSERTBACK:
1609        save_mark = md->mark;
1610      if (md->match_function_type == MATCH_CONDASSERT)      if (md->match_function_type == MATCH_CONDASSERT)
1611        {        {
1612        condassert = TRUE;        condassert = TRUE;
# Line 1528  for (;;) Line 1622  for (;;)
1622          mstart = md->start_match_ptr;   /* In case \K reset it */          mstart = md->start_match_ptr;   /* In case \K reset it */
1623          break;          break;
1624          }          }
1625          md->mark = save_mark;
1626    
1627        /* PCRE does not allow THEN to escape beyond an assertion; it is treated        /* A COMMIT failure must fail the entire assertion, without trying any
1628        as NOMATCH. */        subsequent branches. */
1629    
1630          if (rrc == MATCH_COMMIT) RRETURN(MATCH_NOMATCH);
1631    
1632          /* PCRE does not allow THEN to escape beyond an assertion; it
1633          is treated as NOMATCH. */
1634    
1635        if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN) RRETURN(rrc);        if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN) RRETURN(rrc);
1636        ecode += GET(ecode, 1);        ecode += GET(ecode, 1);
# Line 1557  for (;;) Line 1657  for (;;)
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 1567  for (;;) Line 1668  for (;;)
1668      do      do
1669        {        {
1670        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, NULL, RM5);        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, NULL, RM5);
1671          md->mark = save_mark;
1672        if (rrc == MATCH_MATCH || rrc == MATCH_ACCEPT) RRETURN(MATCH_NOMATCH);        if (rrc == MATCH_MATCH || rrc == MATCH_ACCEPT) RRETURN(MATCH_NOMATCH);
1673        if (rrc == MATCH_SKIP || rrc == MATCH_PRUNE || rrc == MATCH_COMMIT)        if (rrc == MATCH_SKIP || rrc == MATCH_PRUNE || rrc == MATCH_COMMIT)
1674          {          {
# Line 1593  for (;;) Line 1695  for (;;)
1695      back a number of characters, not bytes. */      back a number of characters, not bytes. */
1696    
1697      case OP_REVERSE:      case OP_REVERSE:
1698  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
1699      if (utf8)      if (utf)
1700        {        {
1701        i = GET(ecode, 1);        i = GET(ecode, 1);
1702        while (i-- > 0)        while (i-- > 0)
# Line 1625  for (;;) Line 1727  for (;;)
1727      function is able to force a failure. */      function is able to force a failure. */
1728    
1729      case OP_CALLOUT:      case OP_CALLOUT:
1730      if (pcre_callout != NULL)      if (PUBL(callout) != NULL)
1731        {        {
1732        pcre_callout_block cb;        PUBL(callout_block) cb;
1733        cb.version          = 2;   /* Version 1 of the callout block */        cb.version          = 2;   /* Version 1 of the callout block */
1734        cb.callout_number   = ecode[1];        cb.callout_number   = ecode[1];
1735        cb.offset_vector    = md->offset_vector;        cb.offset_vector    = md->offset_vector;
1736    #if defined COMPILE_PCRE8
1737        cb.subject          = (PCRE_SPTR)md->start_subject;        cb.subject          = (PCRE_SPTR)md->start_subject;
1738    #elif defined COMPILE_PCRE16
1739          cb.subject          = (PCRE_SPTR16)md->start_subject;
1740    #elif defined COMPILE_PCRE32
1741          cb.subject          = (PCRE_SPTR32)md->start_subject;
1742    #endif
1743        cb.subject_length   = (int)(md->end_subject - md->start_subject);        cb.subject_length   = (int)(md->end_subject - md->start_subject);
1744        cb.start_match      = (int)(mstart - md->start_subject);        cb.start_match      = (int)(mstart - md->start_subject);
1745        cb.current_position = (int)(eptr - md->start_subject);        cb.current_position = (int)(eptr - md->start_subject);
1746        cb.pattern_position = GET(ecode, 2);        cb.pattern_position = GET(ecode, 2);
1747        cb.next_item_length = GET(ecode, 2 + LINK_SIZE);        cb.next_item_length = GET(ecode, 2 + LINK_SIZE);
1748        cb.capture_top      = offset_top/2;        cb.capture_top      = offset_top/2;
1749        cb.capture_last     = md->capture_last;        cb.capture_last     = md->capture_last & CAPLMASK;
1750          /* Internal change requires this for API compatibility. */
1751          if (cb.capture_last == 0) cb.capture_last = -1;
1752        cb.callout_data     = md->callout_data;        cb.callout_data     = md->callout_data;
1753        cb.mark             = md->nomatch_mark;        cb.mark             = md->nomatch_mark;
1754        if ((rrc = (*pcre_callout)(&cb)) > 0) RRETURN(MATCH_NOMATCH);        if ((rrc = (*PUBL(callout))(&cb)) > 0) RRETURN(MATCH_NOMATCH);
1755        if (rrc < 0) RRETURN(rrc);        if (rrc < 0) RRETURN(rrc);
1756        }        }
1757      ecode += 2 + 2*LINK_SIZE;      ecode += 2 + 2*LINK_SIZE;
# Line 1667  for (;;) Line 1777  for (;;)
1777      case OP_RECURSE:      case OP_RECURSE:
1778        {        {
1779        recursion_info *ri;        recursion_info *ri;
1780        int recno;        unsigned int recno;
1781    
1782        callpat = md->start_code + GET(ecode, 1);        callpat = md->start_code + GET(ecode, 1);
1783        recno = (callpat == md->start_code)? 0 :        recno = (callpat == md->start_code)? 0 :
# Line 1684  for (;;) Line 1794  for (;;)
1794        /* Add to "recursing stack" */        /* Add to "recursing stack" */
1795    
1796        new_recursive.group_num = recno;        new_recursive.group_num = recno;
1797          new_recursive.saved_capture_last = md->capture_last;
1798        new_recursive.subject_position = eptr;        new_recursive.subject_position = eptr;
1799        new_recursive.prevrec = md->recursive;        new_recursive.prevrec = md->recursive;
1800        md->recursive = &new_recursive;        md->recursive = &new_recursive;
# Line 1700  for (;;) Line 1811  for (;;)
1811        else        else
1812          {          {
1813          new_recursive.offset_save =          new_recursive.offset_save =
1814            (int *)(pcre_malloc)(new_recursive.saved_max * sizeof(int));            (int *)(PUBL(malloc))(new_recursive.saved_max * sizeof(int));
1815          if (new_recursive.offset_save == NULL) RRETURN(PCRE_ERROR_NOMEMORY);          if (new_recursive.offset_save == NULL) RRETURN(PCRE_ERROR_NOMEMORY);
1816          }          }
1817        memcpy(new_recursive.offset_save, md->offset_vector,        memcpy(new_recursive.offset_save, md->offset_vector,
1818              new_recursive.saved_max * sizeof(int));              new_recursive.saved_max * sizeof(int));
1819    
1820        /* OK, now we can do the recursion. After processing each alternative,        /* OK, now we can do the recursion. After processing each alternative,
1821        restore the offset data. If there were nested recursions, md->recursive        restore the offset data and the last captured value. If there were nested
1822        might be changed, so reset it before looping. */        recursions, md->recursive might be changed, so reset it before looping.
1823          */
1824    
1825        DPRINTF(("Recursing into group %d\n", new_recursive.group_num));        DPRINTF(("Recursing into group %d\n", new_recursive.group_num));
1826        cbegroup = (*callpat >= OP_SBRA);        cbegroup = (*callpat >= OP_SBRA);
1827        do        do
1828          {          {
1829          if (cbegroup) md->match_function_type = MATCH_CBEGROUP;          if (cbegroup) md->match_function_type = MATCH_CBEGROUP;
1830          RMATCH(eptr, callpat + _pcre_OP_lengths[*callpat], offset_top,          RMATCH(eptr, callpat + PRIV(OP_lengths)[*callpat], offset_top,
1831            md, eptrb, RM6);            md, eptrb, RM6);
1832          memcpy(md->offset_vector, new_recursive.offset_save,          memcpy(md->offset_vector, new_recursive.offset_save,
1833              new_recursive.saved_max * sizeof(int));              new_recursive.saved_max * sizeof(int));
1834            md->capture_last = new_recursive.saved_capture_last;
1835          md->recursive = new_recursive.prevrec;          md->recursive = new_recursive.prevrec;
1836          if (rrc == MATCH_MATCH || rrc == MATCH_ACCEPT)          if (rrc == MATCH_MATCH || rrc == MATCH_ACCEPT)
1837            {            {
1838            DPRINTF(("Recursion matched\n"));            DPRINTF(("Recursion matched\n"));
1839            if (new_recursive.offset_save != stacksave)            if (new_recursive.offset_save != stacksave)
1840              (pcre_free)(new_recursive.offset_save);              (PUBL(free))(new_recursive.offset_save);
1841    
1842            /* 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
1843            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 1848  for (;;)
1848            goto RECURSION_MATCHED;        /* Exit loop; end processing */            goto RECURSION_MATCHED;        /* Exit loop; end processing */
1849            }            }
1850    
1851          /* 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
1852          as NOMATCH. */          recursion; they are treated as NOMATCH. These codes are defined in a
1853            range that can be tested for. Any other return code is an error. */
1854    
1855          else if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN)          else if (rrc != MATCH_NOMATCH &&
1856                     (rrc < MATCH_BACKTRACK_MIN || rrc > MATCH_BACKTRACK_MAX))
1857            {            {
1858            DPRINTF(("Recursion gave error %d\n", rrc));            DPRINTF(("Recursion gave error %d\n", rrc));
1859            if (new_recursive.offset_save != stacksave)            if (new_recursive.offset_save != stacksave)
1860              (pcre_free)(new_recursive.offset_save);              (PUBL(free))(new_recursive.offset_save);
1861            RRETURN(rrc);            RRETURN(rrc);
1862            }            }
1863    
# Line 1754  for (;;) Line 1869  for (;;)
1869        DPRINTF(("Recursion didn't match\n"));        DPRINTF(("Recursion didn't match\n"));
1870        md->recursive = new_recursive.prevrec;        md->recursive = new_recursive.prevrec;
1871        if (new_recursive.offset_save != stacksave)        if (new_recursive.offset_save != stacksave)
1872          (pcre_free)(new_recursive.offset_save);          (PUBL(free))(new_recursive.offset_save);
1873        RRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
1874        }        }
1875    
# Line 1868  for (;;) Line 1983  for (;;)
1983    
1984        /* Deal with capturing */        /* Deal with capturing */
1985    
1986        md->capture_last = number;        md->capture_last = (md->capture_last & OVFLMASK) | number;
1987        if (offset >= md->offset_max) md->offset_overflow = TRUE; else        if (offset >= md->offset_max) md->capture_last |= OVFLBIT; else
1988          {          {
1989          /* If offset is greater than offset_top, it means that we are          /* If offset is greater than offset_top, it means that we are
1990          "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 2064  for (;;)
2064          }          }
2065        if (*prev >= OP_SBRA)    /* Could match an empty string */        if (*prev >= OP_SBRA)    /* Could match an empty string */
2066          {          {
         md->match_function_type = MATCH_CBEGROUP;  
2067          RMATCH(eptr, prev, offset_top, md, eptrb, RM50);          RMATCH(eptr, prev, offset_top, md, eptrb, RM50);
2068          RRETURN(rrc);          RRETURN(rrc);
2069          }          }
# Line 1958  for (;;) Line 2072  for (;;)
2072        }        }
2073      else  /* OP_KETRMAX */      else  /* OP_KETRMAX */
2074        {        {
       if (*prev >= OP_SBRA) md->match_function_type = MATCH_CBEGROUP;  
2075        RMATCH(eptr, prev, offset_top, md, eptrb, RM13);        RMATCH(eptr, prev, offset_top, md, eptrb, RM13);
2076        if (rrc == MATCH_ONCE && md->once_target == prev) rrc = MATCH_NOMATCH;        if (rrc == MATCH_ONCE && md->once_target == prev) rrc = MATCH_NOMATCH;
2077        if (rrc != MATCH_NOMATCH) RRETURN(rrc);        if (rrc != MATCH_NOMATCH) RRETURN(rrc);
# Line 2015  for (;;) Line 2128  for (;;)
2128    
2129      case OP_DOLLM:      case OP_DOLLM:
2130      if (eptr < md->end_subject)      if (eptr < md->end_subject)
2131        { if (!IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH); }        {
2132          if (!IS_NEWLINE(eptr))
2133            {
2134            if (md->partial != 0 &&
2135                eptr + 1 >= md->end_subject &&
2136                NLBLOCK->nltype == NLTYPE_FIXED &&
2137                NLBLOCK->nllen == 2 &&
2138                RAWUCHARTEST(eptr) == NLBLOCK->nl[0])
2139              {
2140              md->hitend = TRUE;
2141              if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
2142              }
2143            RRETURN(MATCH_NOMATCH);
2144            }
2145          }
2146      else      else
2147        {        {
2148        if (md->noteol) RRETURN(MATCH_NOMATCH);        if (md->noteol) RRETURN(MATCH_NOMATCH);
# Line 2047  for (;;) Line 2174  for (;;)
2174      ASSERT_NL_OR_EOS:      ASSERT_NL_OR_EOS:
2175      if (eptr < md->end_subject &&      if (eptr < md->end_subject &&
2176          (!IS_NEWLINE(eptr) || eptr != md->end_subject - md->nllen))          (!IS_NEWLINE(eptr) || eptr != md->end_subject - md->nllen))
2177          {
2178          if (md->partial != 0 &&
2179              eptr + 1 >= md->end_subject &&
2180              NLBLOCK->nltype == NLTYPE_FIXED &&
2181              NLBLOCK->nllen == 2 &&
2182              RAWUCHARTEST(eptr) == NLBLOCK->nl[0])
2183            {
2184            md->hitend = TRUE;
2185            if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
2186            }
2187        RRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2188          }
2189    
2190      /* Either at end of string or \n before end. */      /* Either at end of string or \n before end. */
2191    
# Line 2066  for (;;) Line 2204  for (;;)
2204        be "non-word" characters. Remember the earliest consulted character for        be "non-word" characters. Remember the earliest consulted character for
2205        partial matching. */        partial matching. */
2206    
2207  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
2208        if (utf8)        if (utf)
2209          {          {
2210          /* Get status of previous character */          /* Get status of previous character */
2211    
2212          if (eptr == md->start_subject) prev_is_word = FALSE; else          if (eptr == md->start_subject) prev_is_word = FALSE; else
2213            {            {
2214            USPTR lastptr = eptr - 1;            PCRE_PUCHAR lastptr = eptr - 1;
2215            while((*lastptr & 0xc0) == 0x80) lastptr--;            BACKCHAR(lastptr);
2216            if (lastptr < md->start_used_ptr) md->start_used_ptr = lastptr;            if (lastptr < md->start_used_ptr) md->start_used_ptr = lastptr;
2217            GETCHAR(c, lastptr);            GETCHAR(c, lastptr);
2218  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
# Line 2139  for (;;) Line 2277  for (;;)
2277              }              }
2278            else            else
2279  #endif  #endif
2280            prev_is_word = ((md->ctypes[eptr[-1]] & ctype_word) != 0);            prev_is_word = MAX_255(eptr[-1])
2281                && ((md->ctypes[eptr[-1]] & ctype_word) != 0);
2282            }            }
2283    
2284          /* Get status of next character */          /* Get status of next character */
# Line 2162  for (;;) Line 2301  for (;;)
2301            }            }
2302          else          else
2303  #endif  #endif
2304          cur_is_word = ((md->ctypes[*eptr] & ctype_word) != 0);          cur_is_word = MAX_255(*eptr)
2305              && ((md->ctypes[*eptr] & ctype_word) != 0);
2306          }          }
2307    
2308        /* Now see if the situation is what we want */        /* Now see if the situation is what we want */
# Line 2173  for (;;) Line 2313  for (;;)
2313        }        }
2314      break;      break;
2315    
2316      /* Match a single character type; inline for speed */      /* Match any single character type except newline; have to take care with
2317        CRLF newlines and partial matching. */
2318    
2319      case OP_ANY:      case OP_ANY:
2320      if (IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH);      if (IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH);
2321        if (md->partial != 0 &&
2322            eptr + 1 >= md->end_subject &&
2323            NLBLOCK->nltype == NLTYPE_FIXED &&
2324            NLBLOCK->nllen == 2 &&
2325            RAWUCHARTEST(eptr) == NLBLOCK->nl[0])
2326          {
2327          md->hitend = TRUE;
2328          if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
2329          }
2330    
2331      /* Fall through */      /* Fall through */
2332    
2333        /* Match any single character whatsoever. */
2334    
2335      case OP_ALLANY:      case OP_ALLANY:
2336      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 */
2337        {                            /* not be updated before SCHECK_PARTIAL. */        {                            /* not be updated before SCHECK_PARTIAL. */
# Line 2186  for (;;) Line 2339  for (;;)
2339        RRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2340        }        }
2341      eptr++;      eptr++;
2342      if (utf8) while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;  #ifdef SUPPORT_UTF
2343        if (utf) ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);
2344    #endif
2345      ecode++;      ecode++;
2346      break;      break;
2347    
# Line 2211  for (;;) Line 2366  for (;;)
2366        }        }
2367      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2368      if (      if (
2369  #ifdef SUPPORT_UTF8  #if defined SUPPORT_UTF || !(defined COMPILE_PCRE8)
2370         c < 256 &&         c < 256 &&
2371  #endif  #endif
2372         (md->ctypes[c] & ctype_digit) != 0         (md->ctypes[c] & ctype_digit) != 0
# Line 2228  for (;;) Line 2383  for (;;)
2383        }        }
2384      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2385      if (      if (
2386  #ifdef SUPPORT_UTF8  #if defined SUPPORT_UTF || !(defined COMPILE_PCRE8)
2387         c >= 256 ||         c > 255 ||
2388  #endif  #endif
2389         (md->ctypes[c] & ctype_digit) == 0         (md->ctypes[c] & ctype_digit) == 0
2390         )         )
# Line 2245  for (;;) Line 2400  for (;;)
2400        }        }
2401      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2402      if (      if (
2403  #ifdef SUPPORT_UTF8  #if defined SUPPORT_UTF || !(defined COMPILE_PCRE8)
2404         c < 256 &&         c < 256 &&
2405  #endif  #endif
2406         (md->ctypes[c] & ctype_space) != 0         (md->ctypes[c] & ctype_space) != 0
# Line 2262  for (;;) Line 2417  for (;;)
2417        }        }
2418      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2419      if (      if (
2420  #ifdef SUPPORT_UTF8  #if defined SUPPORT_UTF || !(defined COMPILE_PCRE8)
2421         c >= 256 ||         c > 255 ||
2422  #endif  #endif
2423         (md->ctypes[c] & ctype_space) == 0         (md->ctypes[c] & ctype_space) == 0
2424         )         )
# Line 2279  for (;;) Line 2434  for (;;)
2434        }        }
2435      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2436      if (      if (
2437  #ifdef SUPPORT_UTF8  #if defined SUPPORT_UTF || !(defined COMPILE_PCRE8)
2438         c < 256 &&         c < 256 &&
2439  #endif  #endif
2440         (md->ctypes[c] & ctype_word) != 0         (md->ctypes[c] & ctype_word) != 0
# Line 2296  for (;;) Line 2451  for (;;)
2451        }        }
2452      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2453      if (      if (
2454  #ifdef SUPPORT_UTF8  #if defined SUPPORT_UTF || !(defined COMPILE_PCRE8)
2455         c >= 256 ||         c > 255 ||
2456  #endif  #endif
2457         (md->ctypes[c] & ctype_word) == 0         (md->ctypes[c] & ctype_word) == 0
2458         )         )
# Line 2316  for (;;) Line 2471  for (;;)
2471        {        {
2472        default: RRETURN(MATCH_NOMATCH);        default: RRETURN(MATCH_NOMATCH);
2473    
2474        case 0x000d:        case CHAR_CR:
2475        if (eptr < md->end_subject && *eptr == 0x0a) eptr++;        if (eptr >= md->end_subject)
2476            {
2477            SCHECK_PARTIAL();
2478            }
2479          else if (RAWUCHARTEST(eptr) == CHAR_LF) eptr++;
2480        break;        break;
2481    
2482        case 0x000a:        case CHAR_LF:
2483        break;        break;
2484    
2485        case 0x000b:        case CHAR_VT:
2486        case 0x000c:        case CHAR_FF:
2487        case 0x0085:        case CHAR_NEL:
2488    #ifndef EBCDIC
2489        case 0x2028:        case 0x2028:
2490        case 0x2029:        case 0x2029:
2491    #endif  /* Not EBCDIC */
2492        if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);        if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);
2493        break;        break;
2494        }        }
# Line 2343  for (;;) Line 2504  for (;;)
2504      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2505      switch(c)      switch(c)
2506        {        {
2507          HSPACE_CASES: RRETURN(MATCH_NOMATCH);  /* Byte and multibyte cases */
2508        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);  
2509        }        }
2510      ecode++;      ecode++;
2511      break;      break;
# Line 2377  for (;;) Line 2519  for (;;)
2519      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2520      switch(c)      switch(c)
2521        {        {
2522          HSPACE_CASES: break;  /* Byte and multibyte cases */
2523        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;  
2524        }        }
2525      ecode++;      ecode++;
2526      break;      break;
# Line 2411  for (;;) Line 2534  for (;;)
2534      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2535      switch(c)      switch(c)
2536        {        {
2537          VSPACE_CASES: RRETURN(MATCH_NOMATCH);
2538        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);  
2539        }        }
2540      ecode++;      ecode++;
2541      break;      break;
# Line 2433  for (;;) Line 2549  for (;;)
2549      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2550      switch(c)      switch(c)
2551        {        {
2552          VSPACE_CASES: break;
2553        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;  
2554        }        }
2555      ecode++;      ecode++;
2556      break;      break;
# Line 2459  for (;;) Line 2568  for (;;)
2568        }        }
2569      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2570        {        {
2571          const pcre_uint32 *cp;
2572        const ucd_record *prop = GET_UCD(c);        const ucd_record *prop = GET_UCD(c);
2573    
2574        switch(ecode[1])        switch(ecode[1])
# Line 2475  for (;;) Line 2585  for (;;)
2585          break;          break;
2586    
2587          case PT_GC:          case PT_GC:
2588          if ((ecode[2] != _pcre_ucp_gentype[prop->chartype]) == (op == OP_PROP))          if ((ecode[2] != PRIV(ucp_gentype)[prop->chartype]) == (op == OP_PROP))
2589            RRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
2590          break;          break;
2591    
# Line 2492  for (;;) Line 2602  for (;;)
2602          /* These are specials */          /* These are specials */
2603    
2604          case PT_ALNUM:          case PT_ALNUM:
2605          if ((_pcre_ucp_gentype[prop->chartype] == ucp_L ||          if ((PRIV(ucp_gentype)[prop->chartype] == ucp_L ||
2606               _pcre_ucp_gentype[prop->chartype] == ucp_N) == (op == OP_NOTPROP))               PRIV(ucp_gentype)[prop->chartype] == ucp_N) == (op == OP_NOTPROP))
2607            RRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
2608          break;          break;
2609    
2610          case PT_SPACE:    /* Perl space */          case PT_SPACE:    /* Perl space */
2611          if ((_pcre_ucp_gentype[prop->chartype] == ucp_Z ||          if ((PRIV(ucp_gentype)[prop->chartype] == ucp_Z ||
2612               c == CHAR_HT || c == CHAR_NL || c == CHAR_FF || c == CHAR_CR)               c == CHAR_HT || c == CHAR_NL || c == CHAR_FF || c == CHAR_CR)
2613                 == (op == OP_NOTPROP))                 == (op == OP_NOTPROP))
2614            RRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
2615          break;          break;
2616    
2617          case PT_PXSPACE:  /* POSIX space */          case PT_PXSPACE:  /* POSIX space */
2618          if ((_pcre_ucp_gentype[prop->chartype] == ucp_Z ||          if ((PRIV(ucp_gentype)[prop->chartype] == ucp_Z ||
2619               c == CHAR_HT || c == CHAR_NL || c == CHAR_VT ||               c == CHAR_HT || c == CHAR_NL || c == CHAR_VT ||
2620               c == CHAR_FF || c == CHAR_CR)               c == CHAR_FF || c == CHAR_CR)
2621                 == (op == OP_NOTPROP))                 == (op == OP_NOTPROP))
# Line 2513  for (;;) Line 2623  for (;;)
2623          break;          break;
2624    
2625          case PT_WORD:          case PT_WORD:
2626          if ((_pcre_ucp_gentype[prop->chartype] == ucp_L ||          if ((PRIV(ucp_gentype)[prop->chartype] == ucp_L ||
2627               _pcre_ucp_gentype[prop->chartype] == ucp_N ||               PRIV(ucp_gentype)[prop->chartype] == ucp_N ||
2628               c == CHAR_UNDERSCORE) == (op == OP_NOTPROP))               c == CHAR_UNDERSCORE) == (op == OP_NOTPROP))
2629            RRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
2630          break;          break;
2631    
2632            case PT_CLIST:
2633            cp = PRIV(ucd_caseless_sets) + ecode[2];
2634            for (;;)
2635              {
2636              if (c < *cp)
2637                { if (op == OP_PROP) { RRETURN(MATCH_NOMATCH); } else break; }
2638              if (c == *cp++)
2639                { if (op == OP_PROP) break; else { RRETURN(MATCH_NOMATCH); } }
2640              }
2641            break;
2642    
2643            case PT_UCNC:
2644            if ((c == CHAR_DOLLAR_SIGN || c == CHAR_COMMERCIAL_AT ||
2645                 c == CHAR_GRAVE_ACCENT || (c >= 0xa0 && c <= 0xd7ff) ||
2646                 c >= 0xe000) == (op == OP_NOTPROP))
2647              RRETURN(MATCH_NOMATCH);
2648            break;
2649    
2650          /* This should never occur */          /* This should never occur */
2651    
2652          default:          default:
# Line 2538  for (;;) Line 2666  for (;;)
2666        SCHECK_PARTIAL();        SCHECK_PARTIAL();
2667        RRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2668        }        }
2669      GETCHARINCTEST(c, eptr);      else
     if (UCD_CATEGORY(c) == ucp_M) RRETURN(MATCH_NOMATCH);  
     while (eptr < md->end_subject)  
2670        {        {
2671        int len = 1;        int lgb, rgb;
2672        if (!utf8) c = *eptr; else { GETCHARLEN(c, eptr, len); }        GETCHARINCTEST(c, eptr);
2673        if (UCD_CATEGORY(c) != ucp_M) break;        lgb = UCD_GRAPHBREAK(c);
2674        eptr += len;        while (eptr < md->end_subject)
2675            {
2676            int len = 1;
2677            if (!utf) c = *eptr; else { GETCHARLEN(c, eptr, len); }
2678            rgb = UCD_GRAPHBREAK(c);
2679            if ((PRIV(ucp_gbtable)[lgb] & (1 << rgb)) == 0) break;
2680            lgb = rgb;
2681            eptr += len;
2682            }
2683        }        }
2684        CHECK_PARTIAL();
2685      ecode++;      ecode++;
2686      break;      break;
2687  #endif  #endif  /* SUPPORT_UCP */
2688    
2689    
2690      /* 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 2699  for (;;)
2699      case OP_REFI:      case OP_REFI:
2700      caseless = op == OP_REFI;      caseless = op == OP_REFI;
2701      offset = GET2(ecode, 1) << 1;               /* Doubled ref number */      offset = GET2(ecode, 1) << 1;               /* Doubled ref number */
2702      ecode += 3;      ecode += 1 + IMM2_SIZE;
2703    
2704      /* If the reference is unset, there are two possibilities:      /* If the reference is unset, there are two possibilities:
2705    
# Line 2604  for (;;) Line 2739  for (;;)
2739        case OP_CRMINRANGE:        case OP_CRMINRANGE:
2740        minimize = (*ecode == OP_CRMINRANGE);        minimize = (*ecode == OP_CRMINRANGE);
2741        min = GET2(ecode, 1);        min = GET2(ecode, 1);
2742        max = GET2(ecode, 3);        max = GET2(ecode, 1 + IMM2_SIZE);
2743        if (max == 0) max = INT_MAX;        if (max == 0) max = INT_MAX;
2744        ecode += 5;        ecode += 1 + 2 * IMM2_SIZE;
2745        break;        break;
2746    
2747        default:               /* No repeat follows */        default:               /* No repeat follows */
2748        if ((length = match_ref(offset, eptr, length, md, caseless)) < 0)        if ((length = match_ref(offset, eptr, length, md, caseless)) < 0)
2749          {          {
2750            if (length == -2) eptr = md->end_subject;   /* Partial match */
2751          CHECK_PARTIAL();          CHECK_PARTIAL();
2752          RRETURN(MATCH_NOMATCH);          RRETURN(MATCH_NOMATCH);
2753          }          }
# Line 2620  for (;;) Line 2756  for (;;)
2756        }        }
2757    
2758      /* Handle repeated back references. If the length of the reference is      /* Handle repeated back references. If the length of the reference is
2759      zero, just continue with the main loop. */      zero, just continue with the main loop. If the length is negative, it
2760        means the reference is unset in non-Java-compatible mode. If the minimum is
2761        zero, we can continue at the same level without recursion. For any other
2762        minimum, carrying on will result in NOMATCH. */
2763    
2764      if (length == 0) continue;      if (length == 0) continue;
2765        if (length < 0 && min == 0) continue;
2766    
2767      /* First, ensure the minimum number of matches are present. We get back      /* First, ensure the minimum number of matches are present. We get back
2768      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 2773  for (;;)
2773        int slength;        int slength;
2774        if ((slength = match_ref(offset, eptr, length, md, caseless)) < 0)        if ((slength = match_ref(offset, eptr, length, md, caseless)) < 0)
2775          {          {
2776            if (slength == -2) eptr = md->end_subject;   /* Partial match */
2777          CHECK_PARTIAL();          CHECK_PARTIAL();
2778          RRETURN(MATCH_NOMATCH);          RRETURN(MATCH_NOMATCH);
2779          }          }
# Line 2656  for (;;) Line 2797  for (;;)
2797          if (fi >= max) RRETURN(MATCH_NOMATCH);          if (fi >= max) RRETURN(MATCH_NOMATCH);
2798          if ((slength = match_ref(offset, eptr, length, md, caseless)) < 0)          if ((slength = match_ref(offset, eptr, length, md, caseless)) < 0)
2799            {            {
2800              if (slength == -2) eptr = md->end_subject;   /* Partial match */
2801            CHECK_PARTIAL();            CHECK_PARTIAL();
2802            RRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
2803            }            }
# Line 2674  for (;;) Line 2816  for (;;)
2816          int slength;          int slength;
2817          if ((slength = match_ref(offset, eptr, length, md, caseless)) < 0)          if ((slength = match_ref(offset, eptr, length, md, caseless)) < 0)
2818            {            {
2819            CHECK_PARTIAL();            /* Can't use CHECK_PARTIAL because we don't want to update eptr in
2820              the soft partial matching case. */
2821    
2822              if (slength == -2 && md->partial != 0 &&
2823                  md->end_subject > md->start_used_ptr)
2824                {
2825                md->hitend = TRUE;
2826                if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
2827                }
2828            break;            break;
2829            }            }
2830          eptr += slength;          eptr += slength;
2831          }          }
2832    
2833        while (eptr >= pp)        while (eptr >= pp)
2834          {          {
2835          RMATCH(eptr, ecode, offset_top, md, eptrb, RM15);          RMATCH(eptr, ecode, offset_top, md, eptrb, RM15);
# Line 2703  for (;;) Line 2854  for (;;)
2854      case OP_NCLASS:      case OP_NCLASS:
2855      case OP_CLASS:      case OP_CLASS:
2856        {        {
2857          /* The data variable is saved across frames, so the byte map needs to
2858          be stored there. */
2859    #define BYTE_MAP ((pcre_uint8 *)data)
2860        data = ecode + 1;                /* Save for matching */        data = ecode + 1;                /* Save for matching */
2861        ecode += 33;                     /* Advance past the item */        ecode += 1 + (32 / sizeof(pcre_uchar)); /* Advance past the item */
2862    
2863        switch (*ecode)        switch (*ecode)
2864          {          {
# Line 2725  for (;;) Line 2879  for (;;)
2879          case OP_CRMINRANGE:          case OP_CRMINRANGE:
2880          minimize = (*ecode == OP_CRMINRANGE);          minimize = (*ecode == OP_CRMINRANGE);
2881          min = GET2(ecode, 1);          min = GET2(ecode, 1);
2882          max = GET2(ecode, 3);          max = GET2(ecode, 1 + IMM2_SIZE);
2883          if (max == 0) max = INT_MAX;          if (max == 0) max = INT_MAX;
2884          ecode += 5;          ecode += 1 + 2 * IMM2_SIZE;
2885          break;          break;
2886    
2887          default:               /* No repeat follows */          default:               /* No repeat follows */
# Line 2737  for (;;) Line 2891  for (;;)
2891    
2892        /* First, ensure the minimum number of matches are present. */        /* First, ensure the minimum number of matches are present. */
2893    
2894  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
2895        /* UTF-8 mode */        if (utf)
       if (utf8)  
2896          {          {
2897          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
2898            {            {
# Line 2754  for (;;) Line 2907  for (;;)
2907              if (op == OP_CLASS) RRETURN(MATCH_NOMATCH);              if (op == OP_CLASS) RRETURN(MATCH_NOMATCH);
2908              }              }
2909            else            else
2910              {              if ((BYTE_MAP[c/8] & (1 << (c&7))) == 0) RRETURN(MATCH_NOMATCH);
             if ((data[c/8] & (1 << (c&7))) == 0) RRETURN(MATCH_NOMATCH);  
             }  
2911            }            }
2912          }          }
2913        else        else
2914  #endif  #endif
2915        /* Not UTF-8 mode */        /* Not UTF mode */
2916          {          {
2917          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
2918            {            {
# Line 2771  for (;;) Line 2922  for (;;)
2922              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
2923              }              }
2924            c = *eptr++;            c = *eptr++;
2925            if ((data[c/8] & (1 << (c&7))) == 0) RRETURN(MATCH_NOMATCH);  #ifndef COMPILE_PCRE8
2926              if (c > 255)
2927                {
2928                if (op == OP_CLASS) RRETURN(MATCH_NOMATCH);
2929                }
2930              else
2931    #endif
2932                if ((BYTE_MAP[c/8] & (1 << (c&7))) == 0) RRETURN(MATCH_NOMATCH);
2933            }            }
2934          }          }
2935    
# Line 2785  for (;;) Line 2943  for (;;)
2943    
2944        if (minimize)        if (minimize)
2945          {          {
2946  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
2947          /* UTF-8 mode */          if (utf)
         if (utf8)  
2948            {            {
2949            for (fi = min;; fi++)            for (fi = min;; fi++)
2950              {              {
# Line 2805  for (;;) Line 2962  for (;;)
2962                if (op == OP_CLASS) RRETURN(MATCH_NOMATCH);                if (op == OP_CLASS) RRETURN(MATCH_NOMATCH);
2963                }                }
2964              else              else
2965                {                if ((BYTE_MAP[c/8] & (1 << (c&7))) == 0) RRETURN(MATCH_NOMATCH);
               if ((data[c/8] & (1 << (c&7))) == 0) RRETURN(MATCH_NOMATCH);  
               }  
2966              }              }
2967            }            }
2968          else          else
2969  #endif  #endif
2970          /* Not UTF-8 mode */          /* Not UTF mode */
2971            {            {
2972            for (fi = min;; fi++)            for (fi = min;; fi++)
2973              {              {
# Line 2825  for (;;) Line 2980  for (;;)
2980                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
2981                }                }
2982              c = *eptr++;              c = *eptr++;
2983              if ((data[c/8] & (1 << (c&7))) == 0) RRETURN(MATCH_NOMATCH);  #ifndef COMPILE_PCRE8
2984                if (c > 255)
2985                  {
2986                  if (op == OP_CLASS) RRETURN(MATCH_NOMATCH);
2987                  }
2988                else
2989    #endif
2990                  if ((BYTE_MAP[c/8] & (1 << (c&7))) == 0) RRETURN(MATCH_NOMATCH);
2991              }              }
2992            }            }
2993          /* Control never gets here */          /* Control never gets here */
# Line 2837  for (;;) Line 2999  for (;;)
2999          {          {
3000          pp = eptr;          pp = eptr;
3001    
3002  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
3003          /* UTF-8 mode */          if (utf)
         if (utf8)  
3004            {            {
3005            for (i = min; i < max; i++)            for (i = min; i < max; i++)
3006              {              {
# Line 2855  for (;;) Line 3016  for (;;)
3016                if (op == OP_CLASS) break;                if (op == OP_CLASS) break;
3017                }                }
3018              else              else
3019                {                if ((BYTE_MAP[c/8] & (1 << (c&7))) == 0) break;
               if ((data[c/8] & (1 << (c&7))) == 0) break;  
               }  
3020              eptr += len;              eptr += len;
3021              }              }
3022            for (;;)            for (;;)
# Line 2870  for (;;) Line 3029  for (;;)
3029            }            }
3030          else          else
3031  #endif  #endif
3032            /* Not UTF-8 mode */            /* Not UTF mode */
3033            {            {
3034            for (i = min; i < max; i++)            for (i = min; i < max; i++)
3035              {              {
# Line 2880  for (;;) Line 3039  for (;;)
3039                break;                break;
3040                }                }
3041              c = *eptr;              c = *eptr;
3042              if ((data[c/8] & (1 << (c&7))) == 0) break;  #ifndef COMPILE_PCRE8
3043                if (c > 255)
3044                  {
3045                  if (op == OP_CLASS) break;
3046                  }
3047                else
3048    #endif
3049                  if ((BYTE_MAP[c/8] & (1 << (c&7))) == 0) break;
3050              eptr++;              eptr++;
3051              }              }
3052            while (eptr >= pp)            while (eptr >= pp)
# Line 2893  for (;;) Line 3059  for (;;)
3059    
3060          RRETURN(MATCH_NOMATCH);          RRETURN(MATCH_NOMATCH);
3061          }          }
3062    #undef BYTE_MAP
3063        }        }
3064      /* Control never gets here */      /* Control never gets here */
3065    
# Line 2901  for (;;) Line 3068  for (;;)
3068      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
3069      mode, because Unicode properties are supported in non-UTF-8 mode. */      mode, because Unicode properties are supported in non-UTF-8 mode. */
3070    
3071  #ifdef SUPPORT_UTF8  #if defined SUPPORT_UTF || !defined COMPILE_PCRE8
3072      case OP_XCLASS:      case OP_XCLASS:
3073        {        {
3074        data = ecode + 1 + LINK_SIZE;                /* Save for matching */        data = ecode + 1 + LINK_SIZE;                /* Save for matching */
# Line 2926  for (;;) Line 3093  for (;;)
3093          case OP_CRMINRANGE:          case OP_CRMINRANGE:
3094          minimize = (*ecode == OP_CRMINRANGE);          minimize = (*ecode == OP_CRMINRANGE);
3095          min = GET2(ecode, 1);          min = GET2(ecode, 1);
3096          max = GET2(ecode, 3);          max = GET2(ecode, 1 + IMM2_SIZE);
3097          if (max == 0) max = INT_MAX;          if (max == 0) max = INT_MAX;
3098          ecode += 5;          ecode += 1 + 2 * IMM2_SIZE;
3099          break;          break;
3100    
3101          default:               /* No repeat follows */          default:               /* No repeat follows */
# Line 2946  for (;;) Line 3113  for (;;)
3113            RRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
3114            }            }
3115          GETCHARINCTEST(c, eptr);          GETCHARINCTEST(c, eptr);
3116          if (!_pcre_xclass(c, data)) RRETURN(MATCH_NOMATCH);          if (!PRIV(xclass)(c, data, utf)) RRETURN(MATCH_NOMATCH);
3117          }          }
3118    
3119        /* 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 3137  for (;;)
3137              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
3138              }              }
3139            GETCHARINCTEST(c, eptr);            GETCHARINCTEST(c, eptr);
3140            if (!_pcre_xclass(c, data)) RRETURN(MATCH_NOMATCH);            if (!PRIV(xclass)(c, data, utf)) RRETURN(MATCH_NOMATCH);
3141            }            }
3142          /* Control never gets here */          /* Control never gets here */
3143          }          }
# Line 2988  for (;;) Line 3155  for (;;)
3155              SCHECK_PARTIAL();              SCHECK_PARTIAL();
3156              break;              break;
3157              }              }
3158    #ifdef SUPPORT_UTF
3159            GETCHARLENTEST(c, eptr, len);            GETCHARLENTEST(c, eptr, len);
3160            if (!_pcre_xclass(c, data)) break;  #else
3161              c = *eptr;
3162    #endif
3163              if (!PRIV(xclass)(c, data, utf)) break;
3164            eptr += len;            eptr += len;
3165            }            }
3166          for(;;)          for(;;)
# Line 2997  for (;;) Line 3168  for (;;)
3168            RMATCH(eptr, ecode, offset_top, md, eptrb, RM21);            RMATCH(eptr, ecode, offset_top, md, eptrb, RM21);
3169            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3170            if (eptr-- == pp) break;        /* Stop if tried at original pos */            if (eptr-- == pp) break;        /* Stop if tried at original pos */
3171            if (utf8) BACKCHAR(eptr);  #ifdef SUPPORT_UTF
3172              if (utf) BACKCHAR(eptr);
3173    #endif
3174            }            }
3175          RRETURN(MATCH_NOMATCH);          RRETURN(MATCH_NOMATCH);
3176          }          }
# Line 3009  for (;;) Line 3182  for (;;)
3182      /* Match a single character, casefully */      /* Match a single character, casefully */
3183    
3184      case OP_CHAR:      case OP_CHAR:
3185  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
3186      if (utf8)      if (utf)
3187        {        {
3188        length = 1;        length = 1;
3189        ecode++;        ecode++;
# Line 3020  for (;;) Line 3193  for (;;)
3193          CHECK_PARTIAL();             /* Not SCHECK_PARTIAL() */          CHECK_PARTIAL();             /* Not SCHECK_PARTIAL() */
3194          RRETURN(MATCH_NOMATCH);          RRETURN(MATCH_NOMATCH);
3195          }          }
3196        while (length-- > 0) if (*ecode++ != *eptr++) RRETURN(MATCH_NOMATCH);        while (length-- > 0) if (*ecode++ != RAWUCHARINC(eptr)) RRETURN(MATCH_NOMATCH);
3197        }        }
3198      else      else
3199  #endif  #endif
3200        /* Not UTF mode */
     /* Non-UTF-8 mode */  
3201        {        {
3202        if (md->end_subject - eptr < 1)        if (md->end_subject - eptr < 1)
3203          {          {
# Line 3047  for (;;) Line 3219  for (;;)
3219        RRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
3220        }        }
3221    
3222  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
3223      if (utf8)      if (utf)
3224        {        {
3225        length = 1;        length = 1;
3226        ecode++;        ecode++;
# Line 3061  for (;;) Line 3233  for (;;)
3233    
3234        if (fc < 128)        if (fc < 128)
3235          {          {
3236          if (md->lcc[*ecode++] != md->lcc[*eptr++]) RRETURN(MATCH_NOMATCH);          pcre_uint32 cc = RAWUCHAR(eptr);
3237            if (md->lcc[fc] != TABLE_GET(cc, md->lcc, cc)) RRETURN(MATCH_NOMATCH);
3238            ecode++;
3239            eptr++;
3240          }          }
3241    
3242        /* 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 3245  for (;;)
3245    
3246        else        else
3247          {          {
3248          unsigned int dc;          pcre_uint32 dc;
3249          GETCHARINC(dc, eptr);          GETCHARINC(dc, eptr);
3250          ecode += length;          ecode += length;
3251    
# Line 3087  for (;;) Line 3262  for (;;)
3262          }          }
3263        }        }
3264      else      else
3265  #endif   /* SUPPORT_UTF8 */  #endif   /* SUPPORT_UTF */
3266    
3267      /* Non-UTF-8 mode */      /* Not UTF mode */
3268        {        {
3269        if (md->lcc[ecode[1]] != md->lcc[*eptr++]) RRETURN(MATCH_NOMATCH);        if (TABLE_GET(ecode[1], md->lcc, ecode[1])
3270              != TABLE_GET(*eptr, md->lcc, *eptr)) RRETURN(MATCH_NOMATCH);
3271          eptr++;
3272        ecode += 2;        ecode += 2;
3273        }        }
3274      break;      break;
# Line 3101  for (;;) Line 3278  for (;;)
3278      case OP_EXACT:      case OP_EXACT:
3279      case OP_EXACTI:      case OP_EXACTI:
3280      min = max = GET2(ecode, 1);      min = max = GET2(ecode, 1);
3281      ecode += 3;      ecode += 1 + IMM2_SIZE;
3282      goto REPEATCHAR;      goto REPEATCHAR;
3283    
3284      case OP_POSUPTO:      case OP_POSUPTO:
# Line 3116  for (;;) Line 3293  for (;;)
3293      min = 0;      min = 0;
3294      max = GET2(ecode, 1);      max = GET2(ecode, 1);
3295      minimize = *ecode == OP_MINUPTO || *ecode == OP_MINUPTOI;      minimize = *ecode == OP_MINUPTO || *ecode == OP_MINUPTOI;
3296      ecode += 3;      ecode += 1 + IMM2_SIZE;
3297      goto REPEATCHAR;      goto REPEATCHAR;
3298    
3299      case OP_POSSTAR:      case OP_POSSTAR:
# Line 3164  for (;;) Line 3341  for (;;)
3341      /* Common code for all repeated single-character matches. */      /* Common code for all repeated single-character matches. */
3342    
3343      REPEATCHAR:      REPEATCHAR:
3344  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
3345      if (utf8)      if (utf)
3346        {        {
3347        length = 1;        length = 1;
3348        charptr = ecode;        charptr = ecode;
# Line 3178  for (;;) Line 3355  for (;;)
3355        if (length > 1)        if (length > 1)
3356          {          {
3357  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
3358          unsigned int othercase;          pcre_uint32 othercase;
3359          if (op >= OP_STARI &&     /* Caseless */          if (op >= OP_STARI &&     /* Caseless */
3360              (othercase = UCD_OTHERCASE(fc)) != fc)              (othercase = UCD_OTHERCASE(fc)) != fc)
3361            oclength = _pcre_ord2utf8(othercase, occhars);            oclength = PRIV(ord2utf)(othercase, occhars);
3362          else oclength = 0;          else oclength = 0;
3363  #endif  /* SUPPORT_UCP */  #endif  /* SUPPORT_UCP */
3364    
3365          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
3366            {            {
3367            if (eptr <= md->end_subject - length &&            if (eptr <= md->end_subject - length &&
3368              memcmp(eptr, charptr, length) == 0) eptr += length;              memcmp(eptr, charptr, IN_UCHARS(length)) == 0) eptr += length;
3369  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
3370            else if (oclength > 0 &&            else if (oclength > 0 &&
3371                     eptr <= md->end_subject - oclength &&                     eptr <= md->end_subject - oclength &&
3372                     memcmp(eptr, occhars, oclength) == 0) eptr += oclength;                     memcmp(eptr, occhars, IN_UCHARS(oclength)) == 0) eptr += oclength;
3373  #endif  /* SUPPORT_UCP */  #endif  /* SUPPORT_UCP */
3374            else            else
3375              {              {
# Line 3211  for (;;) Line 3388  for (;;)
3388              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3389              if (fi >= max) RRETURN(MATCH_NOMATCH);              if (fi >= max) RRETURN(MATCH_NOMATCH);
3390              if (eptr <= md->end_subject - length &&              if (eptr <= md->end_subject - length &&
3391                memcmp(eptr, charptr, length) == 0) eptr += length;                memcmp(eptr, charptr, IN_UCHARS(length)) == 0) eptr += length;
3392  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
3393              else if (oclength > 0 &&              else if (oclength > 0 &&
3394                       eptr <= md->end_subject - oclength &&                       eptr <= md->end_subject - oclength &&
3395                       memcmp(eptr, occhars, oclength) == 0) eptr += oclength;                       memcmp(eptr, occhars, IN_UCHARS(oclength)) == 0) eptr += oclength;
3396  #endif  /* SUPPORT_UCP */  #endif  /* SUPPORT_UCP */
3397              else              else
3398                {                {
# Line 3232  for (;;) Line 3409  for (;;)
3409            for (i = min; i < max; i++)            for (i = min; i < max; i++)
3410              {              {
3411              if (eptr <= md->end_subject - length &&              if (eptr <= md->end_subject - length &&
3412                  memcmp(eptr, charptr, length) == 0) eptr += length;                  memcmp(eptr, charptr, IN_UCHARS(length)) == 0) eptr += length;
3413  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
3414              else if (oclength > 0 &&              else if (oclength > 0 &&
3415                       eptr <= md->end_subject - oclength &&                       eptr <= md->end_subject - oclength &&
3416                       memcmp(eptr, occhars, oclength) == 0) eptr += oclength;                       memcmp(eptr, occhars, IN_UCHARS(oclength)) == 0) eptr += oclength;
3417  #endif  /* SUPPORT_UCP */  #endif  /* SUPPORT_UCP */
3418              else              else
3419                {                {
# Line 3268  for (;;) Line 3445  for (;;)
3445        value of fc will always be < 128. */        value of fc will always be < 128. */
3446        }        }
3447      else      else
3448  #endif  /* SUPPORT_UTF8 */  #endif  /* SUPPORT_UTF */
3449          /* When not in UTF-8 mode, load a single-byte character. */
3450          fc = *ecode++;
3451    
3452      /* When not in UTF-8 mode, load a single-byte character. */      /* The value of fc at this point is always one character, though we may
3453        or may not be in UTF mode. The code is duplicated for the caseless and
     fc = *ecode++;  
   
     /* 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  
3454      caseful cases, for speed, since matching characters is likely to be quite      caseful cases, for speed, since matching characters is likely to be quite
3455      common. First, ensure the minimum number of matches are present. If min =      common. First, ensure the minimum number of matches are present. If min =
3456      max, continue at the same level without recursing. Otherwise, if      max, continue at the same level without recursing. Otherwise, if
# Line 3284  for (;;) Line 3459  for (;;)
3459      maximizing, find the maximum number of characters and work backwards. */      maximizing, find the maximum number of characters and work backwards. */
3460    
3461      DPRINTF(("matching %c{%d,%d} against subject %.*s\n", fc, min, max,      DPRINTF(("matching %c{%d,%d} against subject %.*s\n", fc, min, max,
3462        max, eptr));        max, (char *)eptr));
3463    
3464      if (op >= OP_STARI)  /* Caseless */      if (op >= OP_STARI)  /* Caseless */
3465        {        {
3466        fc = md->lcc[fc];  #ifdef COMPILE_PCRE8
3467          /* fc must be < 128 if UTF is enabled. */
3468          foc = md->fcc[fc];
3469    #else
3470    #ifdef SUPPORT_UTF
3471    #ifdef SUPPORT_UCP
3472          if (utf && fc > 127)
3473            foc = UCD_OTHERCASE(fc);
3474    #else
3475          if (utf && fc > 127)
3476            foc = fc;
3477    #endif /* SUPPORT_UCP */
3478          else
3479    #endif /* SUPPORT_UTF */
3480            foc = TABLE_GET(fc, md->fcc, fc);
3481    #endif /* COMPILE_PCRE8 */
3482    
3483        for (i = 1; i <= min; i++)        for (i = 1; i <= min; i++)
3484          {          {
3485            pcre_uint32 cc;                 /* Faster than pcre_uchar */
3486          if (eptr >= md->end_subject)          if (eptr >= md->end_subject)
3487            {            {
3488            SCHECK_PARTIAL();            SCHECK_PARTIAL();
3489            RRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
3490            }            }
3491          if (fc != md->lcc[*eptr++]) RRETURN(MATCH_NOMATCH);          cc = RAWUCHARTEST(eptr);
3492            if (fc != cc && foc != cc) RRETURN(MATCH_NOMATCH);
3493            eptr++;
3494          }          }
3495        if (min == max) continue;        if (min == max) continue;
3496        if (minimize)        if (minimize)
3497          {          {
3498          for (fi = min;; fi++)          for (fi = min;; fi++)
3499            {            {
3500              pcre_uint32 cc;               /* Faster than pcre_uchar */
3501            RMATCH(eptr, ecode, offset_top, md, eptrb, RM24);            RMATCH(eptr, ecode, offset_top, md, eptrb, RM24);
3502            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3503            if (fi >= max) RRETURN(MATCH_NOMATCH);            if (fi >= max) RRETURN(MATCH_NOMATCH);
# Line 3311  for (;;) Line 3506  for (;;)
3506              SCHECK_PARTIAL();              SCHECK_PARTIAL();
3507              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
3508              }              }
3509            if (fc != md->lcc[*eptr++]) RRETURN(MATCH_NOMATCH);            cc = RAWUCHARTEST(eptr);
3510              if (fc != cc && foc != cc) RRETURN(MATCH_NOMATCH);
3511              eptr++;
3512            }            }
3513          /* Control never gets here */          /* Control never gets here */
3514          }          }
# Line 3320  for (;;) Line 3517  for (;;)
3517          pp = eptr;          pp = eptr;
3518          for (i = min; i < max; i++)          for (i = min; i < max; i++)
3519            {            {
3520              pcre_uint32 cc;               /* Faster than pcre_uchar */
3521            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
3522              {              {
3523              SCHECK_PARTIAL();              SCHECK_PARTIAL();
3524              break;              break;
3525              }              }
3526            if (fc != md->lcc[*eptr]) break;            cc = RAWUCHARTEST(eptr);
3527              if (fc != cc && foc != cc) break;
3528            eptr++;            eptr++;
3529            }            }
3530    
# Line 3353  for (;;) Line 3552  for (;;)
3552            SCHECK_PARTIAL();            SCHECK_PARTIAL();
3553            RRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
3554            }            }
3555          if (fc != *eptr++) RRETURN(MATCH_NOMATCH);          if (fc != RAWUCHARINCTEST(eptr)) RRETURN(MATCH_NOMATCH);
3556          }          }
3557    
3558        if (min == max) continue;        if (min == max) continue;
# Line 3370  for (;;) Line 3569  for (;;)
3569              SCHECK_PARTIAL();              SCHECK_PARTIAL();
3570              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
3571              }              }
3572            if (fc != *eptr++) RRETURN(MATCH_NOMATCH);            if (fc != RAWUCHARINCTEST(eptr)) RRETURN(MATCH_NOMATCH);
3573            }            }
3574          /* Control never gets here */          /* Control never gets here */
3575          }          }
# Line 3384  for (;;) Line 3583  for (;;)
3583              SCHECK_PARTIAL();              SCHECK_PARTIAL();
3584              break;              break;
3585              }              }
3586            if (fc != *eptr) break;            if (fc != RAWUCHARTEST(eptr)) break;
3587            eptr++;            eptr++;
3588            }            }
3589          if (possessive) continue;          if (possessive) continue;
# Line 3410  for (;;) Line 3609  for (;;)
3609        SCHECK_PARTIAL();        SCHECK_PARTIAL();
3610        RRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
3611        }        }
3612      ecode++;  #ifdef SUPPORT_UTF
3613      GETCHARINCTEST(c, eptr);      if (utf)
     if (op == OP_NOTI)         /* The caseless case */  
3614        {        {
3615  #ifdef SUPPORT_UTF8        register pcre_uint32 ch, och;
3616        if (c < 256)  
3617  #endif        ecode++;
3618        c = md->lcc[c];        GETCHARINC(ch, ecode);
3619        if (md->lcc[*ecode++] == c) RRETURN(MATCH_NOMATCH);        GETCHARINC(c, eptr);
3620    
3621          if (op == OP_NOT)
3622            {
3623            if (ch == c) RRETURN(MATCH_NOMATCH);
3624            }
3625          else
3626            {
3627    #ifdef SUPPORT_UCP
3628            if (ch > 127)
3629              och = UCD_OTHERCASE(ch);
3630    #else
3631            if (ch > 127)
3632              och = ch;
3633    #endif /* SUPPORT_UCP */
3634            else
3635              och = TABLE_GET(ch, md->fcc, ch);
3636            if (ch == c || och == c) RRETURN(MATCH_NOMATCH);
3637            }
3638        }        }
3639      else    /* Caseful */      else
3640    #endif
3641        {        {
3642        if (*ecode++ == c) RRETURN(MATCH_NOMATCH);        register pcre_uint32 ch = ecode[1];
3643          c = *eptr++;
3644          if (ch == c || (op == OP_NOTI && TABLE_GET(ch, md->fcc, ch) == c))
3645            RRETURN(MATCH_NOMATCH);
3646          ecode += 2;
3647        }        }
3648      break;      break;
3649    
# Line 3436  for (;;) Line 3657  for (;;)
3657      case OP_NOTEXACT:      case OP_NOTEXACT:
3658      case OP_NOTEXACTI:      case OP_NOTEXACTI:
3659      min = max = GET2(ecode, 1);      min = max = GET2(ecode, 1);
3660      ecode += 3;      ecode += 1 + IMM2_SIZE;
3661      goto REPEATNOTCHAR;      goto REPEATNOTCHAR;
3662    
3663      case OP_NOTUPTO:      case OP_NOTUPTO:
# Line 3446  for (;;) Line 3667  for (;;)
3667      min = 0;      min = 0;
3668      max = GET2(ecode, 1);      max = GET2(ecode, 1);
3669      minimize = *ecode == OP_NOTMINUPTO || *ecode == OP_NOTMINUPTOI;      minimize = *ecode == OP_NOTMINUPTO || *ecode == OP_NOTMINUPTOI;
3670      ecode += 3;      ecode += 1 + IMM2_SIZE;
3671      goto REPEATNOTCHAR;      goto REPEATNOTCHAR;
3672    
3673      case OP_NOTPOSSTAR:      case OP_NOTPOSSTAR:
# Line 3478  for (;;) Line 3699  for (;;)
3699      possessive = TRUE;      possessive = TRUE;
3700      min = 0;      min = 0;
3701      max = GET2(ecode, 1);      max = GET2(ecode, 1);
3702      ecode += 3;      ecode += 1 + IMM2_SIZE;
3703      goto REPEATNOTCHAR;      goto REPEATNOTCHAR;
3704    
3705      case OP_NOTSTAR:      case OP_NOTSTAR:
# Line 3502  for (;;) Line 3723  for (;;)
3723      /* Common code for all repeated single-byte matches. */      /* Common code for all repeated single-byte matches. */
3724    
3725      REPEATNOTCHAR:      REPEATNOTCHAR:
3726      fc = *ecode++;      GETCHARINCTEST(fc, ecode);
3727    
3728      /* The code is duplicated for the caseless and caseful cases, for speed,      /* The code is duplicated for the caseless and caseful cases, for speed,
3729      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 3734  for (;;)
3734      characters and work backwards. */      characters and work backwards. */
3735    
3736      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,
3737        max, eptr));        max, (char *)eptr));
3738    
3739      if (op >= OP_NOTSTARI)     /* Caseless */      if (op >= OP_NOTSTARI)     /* Caseless */
3740        {        {
3741        fc = md->lcc[fc];  #ifdef SUPPORT_UTF
3742    #ifdef SUPPORT_UCP
3743          if (utf && fc > 127)
3744            foc = UCD_OTHERCASE(fc);
3745    #else
3746          if (utf && fc > 127)
3747            foc = fc;
3748    #endif /* SUPPORT_UCP */
3749          else
3750    #endif /* SUPPORT_UTF */
3751            foc = TABLE_GET(fc, md->fcc, fc);
3752    
3753  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
3754        /* UTF-8 mode */        if (utf)
       if (utf8)  
3755          {          {
3756          register unsigned int d;          register pcre_uint32 d;
3757          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
3758            {            {
3759            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
# Line 3532  for (;;) Line 3762  for (;;)
3762              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
3763              }              }
3764            GETCHARINC(d, eptr);            GETCHARINC(d, eptr);
3765            if (d < 256) d = md->lcc[d];            if (fc == d || (unsigned int)foc == d) RRETURN(MATCH_NOMATCH);
           if (fc == d) RRETURN(MATCH_NOMATCH);  
3766            }            }
3767          }          }
3768        else        else
3769  #endif  #endif
3770          /* Not UTF mode */
       /* Not UTF-8 mode */  
3771          {          {
3772          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
3773            {            {
# Line 3548  for (;;) Line 3776  for (;;)
3776              SCHECK_PARTIAL();              SCHECK_PARTIAL();
3777              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
3778              }              }
3779            if (fc == md->lcc[*eptr++]) RRETURN(MATCH_NOMATCH);            if (fc == *eptr || foc == *eptr) RRETURN(MATCH_NOMATCH);
3780              eptr++;
3781            }            }
3782          }          }
3783    
# Line 3556  for (;;) Line 3785  for (;;)
3785    
3786        if (minimize)        if (minimize)
3787          {          {
3788  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
3789          /* UTF-8 mode */          if (utf)
         if (utf8)  
3790            {            {
3791            register unsigned int d;            register pcre_uint32 d;
3792            for (fi = min;; fi++)            for (fi = min;; fi++)
3793              {              {
3794              RMATCH(eptr, ecode, offset_top, md, eptrb, RM28);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM28);
# Line 3572  for (;;) Line 3800  for (;;)
3800                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
3801                }                }
3802              GETCHARINC(d, eptr);              GETCHARINC(d, eptr);
3803              if (d < 256) d = md->lcc[d];              if (fc == d || (unsigned int)foc == d) RRETURN(MATCH_NOMATCH);
             if (fc == d) RRETURN(MATCH_NOMATCH);  
3804              }              }
3805            }            }
3806          else          else
3807  #endif  #endif
3808          /* Not UTF-8 mode */          /* Not UTF mode */
3809            {            {
3810            for (fi = min;; fi++)            for (fi = min;; fi++)
3811              {              {
# Line 3590  for (;;) Line 3817  for (;;)
3817                SCHECK_PARTIAL();                SCHECK_PARTIAL();
3818                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
3819                }                }
3820              if (fc == md->lcc[*eptr++]) RRETURN(MATCH_NOMATCH);              if (fc == *eptr || foc == *eptr) RRETURN(MATCH_NOMATCH);
3821                eptr++;
3822              }              }
3823            }            }
3824          /* Control never gets here */          /* Control never gets here */
# Line 3602  for (;;) Line 3830  for (;;)
3830          {          {
3831          pp = eptr;          pp = eptr;
3832    
3833  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
3834          /* UTF-8 mode */          if (utf)
         if (utf8)  
3835            {            {
3836            register unsigned int d;            register pcre_uint32 d;
3837            for (i = min; i < max; i++)            for (i = min; i < max; i++)
3838              {              {
3839              int len = 1;              int len = 1;
# Line 3616  for (;;) Line 3843  for (;;)
3843                break;                break;
3844                }                }
3845              GETCHARLEN(d, eptr, len);              GETCHARLEN(d, eptr, len);
3846              if (d < 256) d = md->lcc[d];              if (fc == d || (unsigned int)foc == d) break;
             if (fc == d) break;  
3847              eptr += len;              eptr += len;
3848              }              }
3849          if (possessive) continue;            if (possessive) continue;
3850          for(;;)            for(;;)
3851              {              {
3852              RMATCH(eptr, ecode, offset_top, md, eptrb, RM30);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM30);
3853              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
# Line 3631  for (;;) Line 3857  for (;;)
3857            }            }
3858          else          else
3859  #endif  #endif
3860          /* Not UTF-8 mode */          /* Not UTF mode */
3861            {            {
3862            for (i = min; i < max; i++)            for (i = min; i < max; i++)
3863              {              {
# Line 3640  for (;;) Line 3866  for (;;)
3866                SCHECK_PARTIAL();                SCHECK_PARTIAL();
3867                break;                break;
3868                }                }
3869              if (fc == md->lcc[*eptr]) break;              if (fc == *eptr || foc == *eptr) break;
3870              eptr++;              eptr++;
3871              }              }
3872            if (possessive) continue;            if (possessive) continue;
# Line 3661  for (;;) Line 3887  for (;;)
3887    
3888      else      else
3889        {        {
3890  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
3891        /* UTF-8 mode */        if (utf)
       if (utf8)  
3892          {          {
3893          register unsigned int d;          register pcre_uint32 d;
3894          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
3895            {            {
3896            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
# Line 3679  for (;;) Line 3904  for (;;)
3904          }          }
3905        else        else
3906  #endif  #endif
3907        /* Not UTF-8 mode */        /* Not UTF mode */
3908          {          {
3909          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
3910            {            {
# Line 3696  for (;;) Line 3921  for (;;)
3921    
3922        if (minimize)        if (minimize)
3923          {          {
3924  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
3925          /* UTF-8 mode */          if (utf)
         if (utf8)  
3926            {            {
3927            register unsigned int d;            register pcre_uint32 d;
3928            for (fi = min;; fi++)            for (fi = min;; fi++)
3929              {              {
3930              RMATCH(eptr, ecode, offset_top, md, eptrb, RM32);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM32);
# Line 3717  for (;;) Line 3941  for (;;)
3941            }            }
3942          else          else
3943  #endif  #endif
3944          /* Not UTF-8 mode */          /* Not UTF mode */
3945            {            {
3946            for (fi = min;; fi++)            for (fi = min;; fi++)
3947              {              {
# Line 3741  for (;;) Line 3965  for (;;)
3965          {          {
3966          pp = eptr;          pp = eptr;
3967    
3968  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
3969          /* UTF-8 mode */          if (utf)
         if (utf8)  
3970            {            {
3971            register unsigned int d;            register pcre_uint32 d;
3972            for (i = min; i < max; i++)            for (i = min; i < max; i++)
3973              {              {
3974              int len = 1;              int len = 1;
# Line 3769  for (;;) Line 3992  for (;;)
3992            }            }
3993          else          else
3994  #endif  #endif
3995          /* Not UTF-8 mode */          /* Not UTF mode */
3996            {            {
3997            for (i = min; i < max; i++)            for (i = min; i < max; i++)
3998              {              {
# Line 3802  for (;;) Line 4025  for (;;)
4025      case OP_TYPEEXACT:      case OP_TYPEEXACT:
4026      min = max = GET2(ecode, 1);      min = max = GET2(ecode, 1);
4027      minimize = TRUE;      minimize = TRUE;
4028      ecode += 3;      ecode += 1 + IMM2_SIZE;
4029      goto REPEATTYPE;      goto REPEATTYPE;
4030    
4031      case OP_TYPEUPTO:      case OP_TYPEUPTO:
# Line 3810  for (;;) Line 4033  for (;;)
4033      min = 0;      min = 0;
4034      max = GET2(ecode, 1);      max = GET2(ecode, 1);
4035      minimize = *ecode == OP_TYPEMINUPTO;      minimize = *ecode == OP_TYPEMINUPTO;
4036      ecode += 3;      ecode += 1 + IMM2_SIZE;
4037      goto REPEATTYPE;      goto REPEATTYPE;
4038    
4039      case OP_TYPEPOSSTAR:      case OP_TYPEPOSSTAR:
# Line 3838  for (;;) Line 4061  for (;;)
4061      possessive = TRUE;      possessive = TRUE;
4062      min = 0;      min = 0;
4063      max = GET2(ecode, 1);      max = GET2(ecode, 1);
4064      ecode += 3;      ecode += 1 + IMM2_SIZE;
4065      goto REPEATTYPE;      goto REPEATTYPE;
4066    
4067      case OP_TYPESTAR:      case OP_TYPESTAR:
# Line 4021  for (;;) Line 4244  for (;;)
4244              }              }
4245            break;            break;
4246    
4247              case PT_CLIST:
4248              for (i = 1; i <= min; i++)
4249                {
4250                const pcre_uint32 *cp;
4251                if (eptr >= md->end_subject)
4252                  {
4253                  SCHECK_PARTIAL();
4254                  RRETURN(MATCH_NOMATCH);
4255                  }
4256                GETCHARINCTEST(c, eptr);
4257                cp = PRIV(ucd_caseless_sets) + prop_value;
4258                for (;;)
4259                  {
4260                  if (c < *cp)
4261                    { if (prop_fail_result) break; else { RRETURN(MATCH_NOMATCH); } }
4262                  if (c == *cp++)
4263                    { if (prop_fail_result) { RRETURN(MATCH_NOMATCH); } else break; }
4264                  }
4265                }
4266              break;
4267    
4268              case PT_UCNC:
4269              for (i = 1; i <= min; i++)
4270                {
4271                if (eptr >= md->end_subject)
4272                  {
4273                  SCHECK_PARTIAL();
4274                  RRETURN(MATCH_NOMATCH);
4275                  }
4276                GETCHARINCTEST(c, eptr);
4277                if ((c == CHAR_DOLLAR_SIGN || c == CHAR_COMMERCIAL_AT ||
4278                     c == CHAR_GRAVE_ACCENT || (c >= 0xa0 && c <= 0xd7ff) ||
4279                     c >= 0xe000) == prop_fail_result)
4280                  RRETURN(MATCH_NOMATCH);
4281                }
4282              break;
4283    
4284            /* This should not occur */            /* This should not occur */
4285    
4286            default:            default:
# Line 4040  for (;;) Line 4300  for (;;)
4300              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4301              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4302              }              }
4303            GETCHARINCTEST(c, eptr);            else
           if (UCD_CATEGORY(c) == ucp_M) RRETURN(MATCH_NOMATCH);  
           while (eptr < md->end_subject)  
4304              {              {
4305              int len = 1;              int lgb, rgb;
4306              if (!utf8) c = *eptr; else { GETCHARLEN(c, eptr, len); }              GETCHARINCTEST(c, eptr);
4307              if (UCD_CATEGORY(c) != ucp_M) break;              lgb = UCD_GRAPHBREAK(c);
4308              eptr += len;             while (eptr < md->end_subject)
4309                  {
4310                  int len = 1;
4311                  if (!utf) c = *eptr; else { GETCHARLEN(c, eptr, len); }
4312                  rgb = UCD_GRAPHBREAK(c);
4313                  if ((PRIV(ucp_gbtable)[lgb] & (1 << rgb)) == 0) break;
4314                  lgb = rgb;
4315                  eptr += len;
4316                  }
4317              }              }
4318              CHECK_PARTIAL();
4319            }            }
4320          }          }
4321    
# Line 4057  for (;;) Line 4324  for (;;)
4324    
4325  /* Handle all other cases when the coding is UTF-8 */  /* Handle all other cases when the coding is UTF-8 */
4326    
4327  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
4328        if (utf8) switch(ctype)        if (utf) switch(ctype)
4329          {          {
4330          case OP_ANY:          case OP_ANY:
4331          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
# Line 4069  for (;;) Line 4336  for (;;)
4336              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4337              }              }
4338            if (IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH);            if (IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH);
4339              if (md->partial != 0 &&
4340                  eptr + 1 >= md->end_subject &&
4341                  NLBLOCK->nltype == NLTYPE_FIXED &&
4342                  NLBLOCK->nllen == 2 &&
4343                  RAWUCHAR(eptr) == NLBLOCK->nl[0])
4344                {
4345                md->hitend = TRUE;
4346                if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
4347                }
4348            eptr++;            eptr++;
4349            while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;            ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);
4350            }            }
4351          break;          break;
4352    
# Line 4083  for (;;) Line 4359  for (;;)
4359              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4360              }              }
4361            eptr++;            eptr++;
4362            while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;            ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);
4363            }            }
4364          break;          break;
4365    
# Line 4105  for (;;) Line 4381  for (;;)
4381              {              {
4382              default: RRETURN(MATCH_NOMATCH);              default: RRETURN(MATCH_NOMATCH);
4383    
4384              case 0x000d:              case CHAR_CR:
4385              if (eptr < md->end_subject && *eptr == 0x0a) eptr++;              if (eptr < md->end_subject && RAWUCHAR(eptr) == CHAR_LF) eptr++;
4386              break;              break;
4387    
4388              case 0x000a:              case CHAR_LF:
4389              break;              break;
4390    
4391              case 0x000b:              case CHAR_VT:
4392              case 0x000c:              case CHAR_FF:
4393              case 0x0085:              case CHAR_NEL:
4394    #ifndef EBCDIC
4395              case 0x2028:              case 0x2028:
4396              case 0x2029:              case 0x2029:
4397    #endif  /* Not EBCDIC */
4398              if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);              if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);
4399              break;              break;
4400              }              }
# Line 4134  for (;;) Line 4412  for (;;)
4412            GETCHARINC(c, eptr);            GETCHARINC(c, eptr);
4413            switch(c)            switch(c)
4414              {              {
4415                HSPACE_CASES: RRETURN(MATCH_NOMATCH);  /* Byte and multibyte cases */
4416              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);  
4417              }              }
4418            }            }
4419          break;          break;
# Line 4170  for (;;) Line 4429  for (;;)
4429            GETCHARINC(c, eptr);            GETCHARINC(c, eptr);
4430            switch(c)            switch(c)
4431              {              {
4432                HSPACE_CASES: break;  /* Byte and multibyte cases */
4433              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;  
4434              }              }
4435            }            }
4436          break;          break;
# Line 4206  for (;;) Line 4446  for (;;)
4446            GETCHARINC(c, eptr);            GETCHARINC(c, eptr);
4447            switch(c)            switch(c)
4448              {              {
4449                VSPACE_CASES: RRETURN(MATCH_NOMATCH);
4450              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);  
4451              }              }
4452            }            }
4453          break;          break;
# Line 4230  for (;;) Line 4463  for (;;)
4463            GETCHARINC(c, eptr);            GETCHARINC(c, eptr);
4464            switch(c)            switch(c)
4465              {              {
4466                VSPACE_CASES: break;
4467              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;  
4468              }              }
4469            }            }
4470          break;          break;
# Line 4260  for (;;) Line 4486  for (;;)
4486          case OP_DIGIT:          case OP_DIGIT:
4487          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
4488            {            {
4489              pcre_uint32 cc;
4490            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
4491              {              {
4492              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4493              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4494              }              }
4495            if (*eptr >= 128 || (md->ctypes[*eptr++] & ctype_digit) == 0)            cc = RAWUCHAR(eptr);
4496              if (cc >= 128 || (md->ctypes[cc] & ctype_digit) == 0)
4497              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4498              eptr++;
4499            /* No need to skip more bytes - we know it's a 1-byte character */            /* No need to skip more bytes - we know it's a 1-byte character */
4500            }            }
4501          break;          break;
# Line 4274  for (;;) Line 4503  for (;;)
4503          case OP_NOT_WHITESPACE:          case OP_NOT_WHITESPACE:
4504          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
4505            {            {
4506              pcre_uint32 cc;
4507            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
4508              {              {
4509              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4510              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4511              }              }
4512            if (*eptr < 128 && (md->ctypes[*eptr] & ctype_space) != 0)            cc = RAWUCHAR(eptr);
4513              if (cc < 128 && (md->ctypes[cc] & ctype_space) != 0)
4514              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4515            while (++eptr < md->end_subject && (*eptr & 0xc0) == 0x80);            eptr++;
4516              ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);
4517            }            }
4518          break;          break;
4519    
4520          case OP_WHITESPACE:          case OP_WHITESPACE:
4521          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
4522            {            {
4523              pcre_uint32 cc;
4524            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
4525              {              {
4526              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4527              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4528              }              }
4529            if (*eptr >= 128 || (md->ctypes[*eptr++] & ctype_space) == 0)            cc = RAWUCHAR(eptr);
4530              if (cc >= 128 || (md->ctypes[cc] & ctype_space) == 0)
4531              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4532              eptr++;
4533            /* No need to skip more bytes - we know it's a 1-byte character */            /* No need to skip more bytes - we know it's a 1-byte character */
4534            }            }
4535          break;          break;
# Line 4302  for (;;) Line 4537  for (;;)
4537          case OP_NOT_WORDCHAR:          case OP_NOT_WORDCHAR:
4538          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
4539            {            {
4540              pcre_uint32 cc;
4541            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
4542              {              {
4543              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4544              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4545              }              }
4546            if (*eptr < 128 && (md->ctypes[*eptr] & ctype_word) != 0)            cc = RAWUCHAR(eptr);
4547              if (cc < 128 && (md->ctypes[cc] & ctype_word) != 0)
4548              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4549            while (++eptr < md->end_subject && (*eptr & 0xc0) == 0x80);            eptr++;
4550              ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);
4551            }            }
4552          break;          break;
4553    
4554          case OP_WORDCHAR:          case OP_WORDCHAR:
4555          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
4556            {            {
4557              pcre_uint32 cc;
4558            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
4559              {              {
4560              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4561              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4562              }              }
4563            if (*eptr >= 128 || (md->ctypes[*eptr++] & ctype_word) == 0)            cc = RAWUCHAR(eptr);
4564              if (cc >= 128 || (md->ctypes[cc] & ctype_word) == 0)
4565              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4566              eptr++;
4567            /* 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 */
4568            }            }
4569          break;          break;
# Line 4332  for (;;) Line 4573  for (;;)
4573          }  /* End switch(ctype) */          }  /* End switch(ctype) */
4574    
4575        else        else
4576  #endif     /* SUPPORT_UTF8 */  #endif     /* SUPPORT_UTF */
4577    
4578        /* 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
4579        than OP_PROP and OP_NOTPROP. */        than OP_PROP and OP_NOTPROP. */
# Line 4348  for (;;) Line 4589  for (;;)
4589              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4590              }              }
4591            if (IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH);            if (IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH);
4592              if (md->partial != 0 &&
4593                  eptr + 1 >= md->end_subject &&
4594                  NLBLOCK->nltype == NLTYPE_FIXED &&
4595                  NLBLOCK->nllen == 2 &&
4596                  *eptr == NLBLOCK->nl[0])
4597                {
4598                md->hitend = TRUE;
4599                if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
4600                }
4601            eptr++;            eptr++;
4602            }            }
4603          break;          break;
# Line 4382  for (;;) Line 4632  for (;;)
4632              {              {
4633              default: RRETURN(MATCH_NOMATCH);              default: RRETURN(MATCH_NOMATCH);
4634    
4635              case 0x000d:              case CHAR_CR:
4636              if (eptr < md->end_subject && *eptr == 0x0a) eptr++;              if (eptr < md->end_subject && *eptr == CHAR_LF) eptr++;
4637              break;              break;
4638    
4639              case 0x000a:              case CHAR_LF:
4640              break;              break;
4641    
4642              case 0x000b:              case CHAR_VT:
4643              case 0x000c:              case CHAR_FF:
4644              case 0x0085:              case CHAR_NEL:
4645    #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
4646                case 0x2028:
4647                case 0x2029:
4648    #endif
4649              if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);              if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);
4650              break;              break;
4651              }              }
# Line 4409  for (;;) Line 4663  for (;;)
4663            switch(*eptr++)            switch(*eptr++)
4664              {              {
4665              default: break;              default: break;
4666              case 0x09:      /* HT */              HSPACE_BYTE_CASES:
4667              case 0x20:      /* SPACE */  #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
4668              case 0xa0:      /* NBSP */              HSPACE_MULTIBYTE_CASES:
4669    #endif
4670              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4671              }              }
4672            }            }
# Line 4428  for (;;) Line 4683  for (;;)
4683            switch(*eptr++)            switch(*eptr++)
4684              {              {
4685              default: RRETURN(MATCH_NOMATCH);              default: RRETURN(MATCH_NOMATCH);
4686              case 0x09:      /* HT */              HSPACE_BYTE_CASES:
4687              case 0x20:      /* SPACE */  #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
4688              case 0xa0:      /* NBSP */              HSPACE_MULTIBYTE_CASES:
4689    #endif
4690              break;              break;
4691              }              }
4692            }            }
# Line 4446  for (;;) Line 4702  for (;;)
4702              }              }
4703            switch(*eptr++)            switch(*eptr++)
4704              {              {
4705              default: break;              VSPACE_BYTE_CASES:
4706              case 0x0a:      /* LF */  #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
4707              case 0x0b:      /* VT */              VSPACE_MULTIBYTE_CASES:
4708              case 0x0c:      /* FF */  #endif
             case 0x0d:      /* CR */  
             case 0x85:      /* NEL */  
4709              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4710                default: break;
4711              }              }
4712            }            }
4713          break;          break;
# Line 4468  for (;;) Line 4723  for (;;)
4723            switch(*eptr++)            switch(*eptr++)
4724              {              {
4725              default: RRETURN(MATCH_NOMATCH);              default: RRETURN(MATCH_NOMATCH);
4726              case 0x0a:      /* LF */              VSPACE_BYTE_CASES:
4727              case 0x0b:      /* VT */  #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
4728              case 0x0c:      /* FF */              VSPACE_MULTIBYTE_CASES:
4729              case 0x0d:      /* CR */  #endif
             case 0x85:      /* NEL */  
4730              break;              break;
4731              }              }
4732            }            }
# Line 4486  for (;;) Line 4740  for (;;)
4740              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4741              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4742              }              }
4743            if ((md->ctypes[*eptr++] & ctype_digit) != 0) RRETURN(MATCH_NOMATCH);            if (MAX_255(*eptr) && (md->ctypes[*eptr] & ctype_digit) != 0)
4744                RRETURN(MATCH_NOMATCH);
4745              eptr++;
4746            }            }
4747          break;          break;
4748    
# Line 4498  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 4510  for (;;) Line 4768  for (;;)
4768              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4769              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4770              }              }
4771            if ((md->ctypes[*eptr++] & ctype_space) != 0) RRETURN(MATCH_NOMATCH);            if (MAX_255(*eptr) && (md->ctypes[*eptr] & ctype_space) != 0)
4772                RRETURN(MATCH_NOMATCH);
4773              eptr++;
4774            }            }
4775          break;          break;
4776    
# Line 4522  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 4534  for (;;) Line 4796  for (;;)
4796              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4797              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4798              }              }
4799            if ((md->ctypes[*eptr++] & ctype_word) != 0)            if (MAX_255(*eptr) && (md->ctypes[*eptr] & ctype_word) != 0)
4800              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4801              eptr++;
4802            }            }
4803          break;          break;
4804    
# Line 4547  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 4739  for (;;) Line 5003  for (;;)
5003              }              }
5004            /* Control never gets here */            /* Control never gets here */
5005    
5006            /* This should never occur */            case PT_CLIST:
5007              for (fi = min;; fi++)
5008                {
5009                const pcre_uint32 *cp;
5010                RMATCH(eptr, ecode, offset_top, md, eptrb, RM67);
5011                if (rrc != MATCH_NOMATCH) RRETURN(rrc);
5012                if (fi >= max) RRETURN(MATCH_NOMATCH);
5013                if (eptr >= md->end_subject)
5014                  {
5015                  SCHECK_PARTIAL();
5016                  RRETURN(MATCH_NOMATCH);
5017                  }
5018                GETCHARINCTEST(c, eptr);
5019                cp = PRIV(ucd_caseless_sets) + prop_value;
5020                for (;;)
5021                  {
5022                  if (c < *cp)
5023                    { if (prop_fail_result) break; else { RRETURN(MATCH_NOMATCH); } }
5024                  if (c == *cp++)
5025                    { if (prop_fail_result) { RRETURN(MATCH_NOMATCH); } else break; }
5026                  }
5027                }
5028              /* Control never gets here */
5029    
5030              case PT_UCNC:
5031              for (fi = min;; fi++)
5032                {
5033                RMATCH(eptr, ecode, offset_top, md, eptrb, RM68);
5034                if (rrc != MATCH_NOMATCH) RRETURN(rrc);
5035                if (fi >= max) RRETURN(MATCH_NOMATCH);
5036                if (eptr >= md->end_subject)
5037                  {
5038                  SCHECK_PARTIAL();
5039                  RRETURN(MATCH_NOMATCH);
5040                  }
5041                GETCHARINCTEST(c, eptr);
5042                if ((c == CHAR_DOLLAR_SIGN || c == CHAR_COMMERCIAL_AT ||
5043                     c == CHAR_GRAVE_ACCENT || (c >= 0xa0 && c <= 0xd7ff) ||
5044                     c >= 0xe000) == prop_fail_result)
5045                  RRETURN(MATCH_NOMATCH);
5046                }
5047              /* Control never gets here */
5048    
5049              /* This should never occur */
5050            default:            default:
5051            RRETURN(PCRE_ERROR_INTERNAL);            RRETURN(PCRE_ERROR_INTERNAL);
5052            }            }
# Line 4761  for (;;) Line 5067  for (;;)
5067              SCHECK_PARTIAL();              SCHECK_PARTIAL();
5068              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
5069              }              }
5070            GETCHARINCTEST(c, eptr);            else
           if (UCD_CATEGORY(c) == ucp_M) RRETURN(MATCH_NOMATCH);  
           while (eptr < md->end_subject)  
5071              {              {
5072              int len = 1;              int lgb, rgb;
5073              if (!utf8) c = *eptr; else { GETCHARLEN(c, eptr, len); }              GETCHARINCTEST(c, eptr);
5074              if (UCD_CATEGORY(c) != ucp_M) break;              lgb = UCD_GRAPHBREAK(c);
5075              eptr += len;              while (eptr < md->end_subject)
5076                  {
5077                  int len = 1;
5078                  if (!utf) c = *eptr; else { GETCHARLEN(c, eptr, len); }
5079                  rgb = UCD_GRAPHBREAK(c);
5080                  if ((PRIV(ucp_gbtable)[lgb] & (1 << rgb)) == 0) break;
5081                  lgb = rgb;
5082                  eptr += len;
5083                  }
5084              }              }
5085              CHECK_PARTIAL();
5086            }            }
5087          }          }
5088        else        else
5089  #endif     /* SUPPORT_UCP */  #endif     /* SUPPORT_UCP */
5090    
5091  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
5092        /* UTF-8 mode */        if (utf)
       if (utf8)  
5093          {          {
5094          for (fi = min;; fi++)          for (fi = min;; fi++)
5095            {            {
# Line 4794  for (;;) Line 5106  for (;;)
5106            GETCHARINC(c, eptr);            GETCHARINC(c, eptr);
5107            switch(ctype)            switch(ctype)
5108              {              {
5109              case OP_ANY:        /* This is the non-NL case */              case OP_ANY:               /* This is the non-NL case */
5110                if (md->partial != 0 &&    /* Take care with CRLF partial */
5111                    eptr >= md->end_subject &&
5112                    NLBLOCK->nltype == NLTYPE_FIXED &&
5113                    NLBLOCK->nllen == 2 &&
5114                    c == NLBLOCK->nl[0])
5115                  {
5116                  md->hitend = TRUE;
5117                  if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
5118                  }
5119                break;
5120    
5121              case OP_ALLANY:              case OP_ALLANY:
5122              case OP_ANYBYTE:              case OP_ANYBYTE:
5123              break;              break;
# Line 4803  for (;;) Line 5126  for (;;)
5126              switch(c)              switch(c)
5127                {                {
5128                default: RRETURN(MATCH_NOMATCH);                default: RRETURN(MATCH_NOMATCH);
5129                case 0x000d:                case CHAR_CR:
5130                if (eptr < md->end_subject && *eptr == 0x0a) eptr++;                if (eptr < md->end_subject && RAWUCHAR(eptr) == CHAR_LF) eptr++;
5131                break;                break;
5132                case 0x000a:  
5133                  case CHAR_LF:
5134                break;                break;
5135    
5136                case 0x000b:                case CHAR_VT:
5137                case 0x000c:                case CHAR_FF:
5138                case 0x0085:                case CHAR_NEL:
5139    #ifndef EBCDIC
5140                case 0x2028:                case 0x2028:
5141                case 0x2029:                case 0x2029:
5142    #endif  /* Not EBCDIC */
5143                if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);                if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);
5144                break;                break;
5145                }                }
# Line 4822  for (;;) Line 5148  for (;;)
5148              case OP_NOT_HSPACE:              case OP_NOT_HSPACE:
5149              switch(c)              switch(c)
5150                {                {
5151                  HSPACE_CASES: RRETURN(MATCH_NOMATCH);
5152                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);  
5153                }                }
5154              break;              break;
5155    
5156              case OP_HSPACE:              case OP_HSPACE:
5157              switch(c)              switch(c)
5158                {                {
5159                  HSPACE_CASES: break;
5160                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;  
5161                }                }
5162              break;              break;
5163    
5164              case OP_NOT_VSPACE:              case OP_NOT_VSPACE:
5165              switch(c)              switch(c)
5166                {                {
5167                  VSPACE_CASES: RRETURN(MATCH_NOMATCH);
5168                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);  
5169                }                }
5170              break;              break;
5171    
5172              case OP_VSPACE:              case OP_VSPACE:
5173              switch(c)              switch(c)
5174                {                {
5175                  VSPACE_CASES: break;
5176                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;  
5177                }                }
5178              break;              break;
5179    
# Line 4919  for (;;) Line 5193  for (;;)
5193              break;              break;
5194    
5195              case OP_WHITESPACE:              case OP_WHITESPACE:
5196              if  (c >= 256 || (md->ctypes[c] & ctype_space) == 0)              if (c >= 256 || (md->ctypes[c] & ctype_space) == 0)
5197                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
5198              break;              break;
5199    
# Line 4940  for (;;) Line 5214  for (;;)
5214          }          }
5215        else        else
5216  #endif  #endif
5217        /* Not UTF-8 mode */        /* Not UTF mode */
5218          {          {
5219          for (fi = min;; fi++)          for (fi = min;; fi++)
5220            {            {
# Line 4957  for (;;) Line 5231  for (;;)
5231            c = *eptr++;            c = *eptr++;
5232            switch(ctype)            switch(ctype)
5233              {              {
5234              case OP_ANY:     /* This is the non-NL case */              case OP_ANY:               /* This is the non-NL case */
5235                if (md->partial != 0 &&    /* Take care with CRLF partial */
5236                    eptr >= md->end_subject &&
5237                    NLBLOCK->nltype == NLTYPE_FIXED &&
5238                    NLBLOCK->nllen == 2 &&
5239                    c == NLBLOCK->nl[0])
5240                  {
5241                  md->hitend = TRUE;
5242                  if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
5243                  }
5244                break;
5245    
5246              case OP_ALLANY:              case OP_ALLANY:
5247              case OP_ANYBYTE:              case OP_ANYBYTE:
5248              break;              break;
# Line 4966  for (;;) Line 5251  for (;;)
5251              switch(c)              switch(c)
5252                {                {
5253                default: RRETURN(MATCH_NOMATCH);                default: RRETURN(MATCH_NOMATCH);
5254                case 0x000d:                case CHAR_CR:
5255                if (eptr < md->end_subject && *eptr == 0x0a) eptr++;                if (eptr < md->end_subject && *eptr == CHAR_LF) eptr++;
5256                break;                break;
5257    
5258                case 0x000a:                case CHAR_LF:
5259                break;                break;
5260    
5261                case 0x000b:                case CHAR_VT:
5262                case 0x000c:                case CHAR_FF:
5263                case 0x0085:                case CHAR_NEL:
5264    #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
5265                  case 0x2028:
5266                  case 0x2029:
5267    #endif
5268                if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);                if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);
5269                break;                break;
5270                }                }
# Line 4985  for (;;) Line 5274  for (;;)
5274              switch(c)              switch(c)
5275                {                {
5276                default: break;                default: break;
5277                case 0x09:      /* HT */                HSPACE_BYTE_CASES:
5278                case 0x20:      /* SPACE */  #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
5279                case 0xa0:      /* NBSP */                HSPACE_MULTIBYTE_CASES:
5280    #endif
5281                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
5282                }                }
5283              break;              break;
# Line 4996  for (;;) Line 5286  for (;;)
5286              switch(c)              switch(c)
5287                {                {
5288                default: RRETURN(MATCH_NOMATCH);                default: RRETURN(MATCH_NOMATCH);
5289                case 0x09:      /* HT */                HSPACE_BYTE_CASES:
5290                case 0x20:      /* SPACE */  #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
5291                case 0xa0:      /* NBSP */                HSPACE_MULTIBYTE_CASES:
5292    #endif
5293                break;                break;
5294                }                }
5295              break;              break;
# Line 5007  for (;;) Line 5298  for (;;)
5298              switch(c)              switch(c)
5299                {                {
5300                default: break;                default: break;
5301                case 0x0a:      /* LF */                VSPACE_BYTE_CASES:
5302                case 0x0b:      /* VT */  #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
5303                case 0x0c:      /* FF */                VSPACE_MULTIBYTE_CASES:
5304                case 0x0d:      /* CR */  #endif
               case 0x85:      /* NEL */  
5305                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
5306                }                }
5307              break;              break;
# Line 5020  for (;;) Line 5310  for (;;)
5310              switch(c)              switch(c)
5311                {                {
5312                default: RRETURN(MATCH_NOMATCH);                default: RRETURN(MATCH_NOMATCH);
5313                case 0x0a:      /* LF */                VSPACE_BYTE_CASES:
5314                case 0x0b:      /* VT */  #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
5315                case 0x0c:      /* FF */                VSPACE_MULTIBYTE_CASES:
5316                case 0x0d:      /* CR */  #endif
               case 0x85:      /* NEL */  
5317                break;                break;
5318                }                }
5319              break;              break;
5320    
5321              case OP_NOT_DIGIT:              case OP_NOT_DIGIT:
5322              if ((md->ctypes[c] & ctype_digit) != 0) RRETURN(MATCH_NOMATCH);              if (MAX_255(c) && (md->ctypes[c] & ctype_digit) != 0) RRETURN(MATCH_NOMATCH);
5323              break;              break;
5324    
5325              case OP_DIGIT:              case OP_DIGIT:
5326              if ((md->ctypes[c] & ctype_digit) == 0) RRETURN(MATCH_NOMATCH);              if (!MAX_255(c) || (md->ctypes[c] & ctype_digit) == 0) RRETURN(MATCH_NOMATCH);
5327              break;              break;
5328    
5329              case OP_NOT_WHITESPACE:              case OP_NOT_WHITESPACE:
5330              if ((md->ctypes[c] & ctype_space) != 0) RRETURN(MATCH_NOMATCH);              if (MAX_255(c) && (md->ctypes[c] & ctype_space) != 0) RRETURN(MATCH_NOMATCH);
5331              break;              break;
5332    
5333              case OP_WHITESPACE:              case OP_WHITESPACE:
5334              if  ((md->ctypes[c] & ctype_space) == 0) RRETURN(MATCH_NOMATCH);              if (!MAX_255(c) || (md->ctypes[c] & ctype_space) == 0) RRETURN(MATCH_NOMATCH);
5335              break;              break;
5336    
5337              case OP_NOT_WORDCHAR:              case OP_NOT_WORDCHAR:
5338              if ((md->ctypes[c] & ctype_word) != 0) RRETURN(MATCH_NOMATCH);              if (MAX_255(c) && (md->ctypes[c] & ctype_word) != 0) RRETURN(MATCH_NOMATCH);
5339              break;              break;
5340    
5341              case OP_WORDCHAR:              case OP_WORDCHAR:
5342              if ((md->ctypes[c] & ctype_word) == 0) RRETURN(MATCH_NOMATCH);              if (!MAX_255(c) || (md->ctypes[c] & ctype_word) == 0) RRETURN(MATCH_NOMATCH);
5343              break;              break;
5344    
5345              default:              default:
# Line 5227  for (;;) Line 5516  for (;;)
5516              }              }
5517            break;            break;
5518    
5519              case PT_CLIST:
5520              for (i = min; i < max; i++)
5521                {
5522                const pcre_uint32 *cp;
5523                int len = 1;
5524                if (eptr >= md->end_subject)
5525                  {
5526                  SCHECK_PARTIAL();
5527                  break;
5528                  }
5529                GETCHARLENTEST(c, eptr, len);
5530                cp = PRIV(ucd_caseless_sets) + prop_value;
5531                for (;;)
5532                  {
5533                  if (c < *cp)
5534                    { if (prop_fail_result) break; else goto GOT_MAX; }
5535                  if (c == *cp++)
5536                    { if (prop_fail_result) goto GOT_MAX; else break; }
5537                  }
5538                eptr += len;
5539                }
5540              GOT_MAX:
5541              break;
5542    
5543              case PT_UCNC:
5544              for (i = min; i < max; i++)
5545                {
5546                int len = 1;
5547                if (eptr >= md->end_subject)
5548                  {
5549                  SCHECK_PARTIAL();
5550                  break;
5551                  }
5552                GETCHARLENTEST(c, eptr, len);
5553                if ((c == CHAR_DOLLAR_SIGN || c == CHAR_COMMERCIAL_AT ||
5554                     c == CHAR_GRAVE_ACCENT || (c >= 0xa0 && c <= 0xd7ff) ||
5555                     c >= 0xe000) == prop_fail_result)
5556                  break;
5557                eptr += len;
5558                }
5559              break;
5560    
5561            default:            default:
5562            RRETURN(PCRE_ERROR_INTERNAL);            RRETURN(PCRE_ERROR_INTERNAL);
5563            }            }
# Line 5239  for (;;) Line 5570  for (;;)
5570            RMATCH(eptr, ecode, offset_top, md, eptrb, RM44);            RMATCH(eptr, ecode, offset_top, md, eptrb, RM44);
5571            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
5572            if (eptr-- == pp) break;        /* Stop if tried at original pos */            if (eptr-- == pp) break;        /* Stop if tried at original pos */
5573            if (utf8) BACKCHAR(eptr);            if (utf) BACKCHAR(eptr);
5574            }            }
5575          }          }
5576    
# Line 5250  for (;;) Line 5581  for (;;)
5581          {          {
5582          for (i = min; i < max; i++)          for (i = min; i < max; i++)
5583            {            {
           int len = 1;  
5584            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
5585              {              {
5586              SCHECK_PARTIAL();              SCHECK_PARTIAL();
5587              break;              break;
5588              }              }
5589            if (!utf8) c = *eptr; else { GETCHARLEN(c, eptr, len); }            else
           if (UCD_CATEGORY(c) == ucp_M) break;  
           eptr += len;  
           while (eptr < md->end_subject)  
5590              {              {
5591              len = 1;              int lgb, rgb;
5592              if (!utf8) c = *eptr; else { GETCHARLEN(c, eptr, len); }              GETCHARINCTEST(c, eptr);
5593              if (UCD_CATEGORY(c) != ucp_M) break;              lgb = UCD_GRAPHBREAK(c);
5594              eptr += len;              while (eptr < md->end_subject)
5595                  {
5596                  int len = 1;
5597                  if (!utf) c = *eptr; else { GETCHARLEN(c, eptr, len); }
5598                  rgb = UCD_GRAPHBREAK(c);
5599                  if ((PRIV(ucp_gbtable)[lgb] & (1 << rgb)) == 0) break;
5600                  lgb = rgb;
5601                  eptr += len;
5602                  }
5603              }              }
5604              CHECK_PARTIAL();
5605            }            }
5606    
5607          /* eptr is now past the end of the maximum run */          /* eptr is now past the end of the maximum run */
# Line 5279  for (;;) Line 5615  for (;;)
5615            if (eptr-- == pp) break;        /* Stop if tried at original pos */            if (eptr-- == pp) break;        /* Stop if tried at original pos */
5616            for (;;)                        /* Move back over one extended */            for (;;)                        /* Move back over one extended */
5617              {              {
5618              if (!utf8) c = *eptr; else              if (!utf) c = *eptr; else
5619                {                {
5620                BACKCHAR(eptr);                BACKCHAR(eptr);
5621                GETCHAR(c, eptr);                GETCHAR(c, eptr);
# Line 5293  for (;;) Line 5629  for (;;)
5629        else        else
5630  #endif   /* SUPPORT_UCP */  #endif   /* SUPPORT_UCP */
5631    
5632  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
5633        /* UTF-8 mode */        if (utf)
   
       if (utf8)  
5634          {          {
5635          switch(ctype)          switch(ctype)
5636            {            {
# Line 5311  for (;;) Line 5645  for (;;)
5645                  break;                  break;
5646                  }                  }
5647                if (IS_NEWLINE(eptr)) break;                if (IS_NEWLINE(eptr)) break;
5648                  if (md->partial != 0 &&    /* Take care with CRLF partial */
5649                      eptr + 1 >= md->end_subject &&
5650                      NLBLOCK->nltype == NLTYPE_FIXED &&
5651                      NLBLOCK->nllen == 2 &&
5652                      RAWUCHAR(eptr) == NLBLOCK->nl[0])
5653                    {
5654                    md->hitend = TRUE;
5655                    if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
5656                    }
5657                eptr++;                eptr++;
5658                while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;                ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);
5659                }                }
5660              }              }
5661    
# Line 5328  for (;;) Line 5671  for (;;)
5671                  break;                  break;
5672                  }                  }
5673                if (IS_NEWLINE(eptr)) break;                if (IS_NEWLINE(eptr)) break;
5674                  if (md->partial != 0 &&    /* Take care with CRLF partial */
5675                      eptr + 1 >= md->end_subject &&
5676                      NLBLOCK->nltype == NLTYPE_FIXED &&
5677                      NLBLOCK->nllen == 2 &&
5678                      RAWUCHAR(eptr) == NLBLOCK->nl[0])
5679                    {
5680                    md->hitend = TRUE;
5681                    if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
5682                    }
5683                eptr++;                eptr++;