/[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 771 by ph10, Tue Nov 29 15:34:12 2011 UTC revision 1284 by ph10, Fri Mar 15 11:54:58 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 eptr - eptr_start;  return (int)(eptr - eptr_start);
264  }  }
265    
266    
# 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 417  returns a negative (error) response, the Line 466  returns a negative (error) response, the
466  same response. */  same response. */
467    
468  /* These macros pack up tests that are used for partial matching, and which  /* These macros pack up tests that are used for partial matching, and which
469  appears several times in the code. We set the "hit end" flag if the pointer is  appear several times in the code. We set the "hit end" flag if the pointer is
470  at the end of the subject and also past the start of the subject (i.e.  at the end of the subject and also past the start of the subject (i.e.
471  something has been matched). For hard partial matching, we then return  something has been matched). For hard partial matching, we then return
472  immediately. The second one is used when we already know we are past the end of  immediately. The second one is used when we already know we are past the end of
# Line 440  the subject. */ Line 489  the subject. */
489    
490    
491  /* Performance note: It might be tempting to extract commonly used fields from  /* Performance note: It might be tempting to extract commonly used fields from
492  the md structure (e.g. utf8, end_subject) into individual variables to improve  the md structure (e.g. utf, end_subject) into individual variables to improve
493  performance. Tests using gcc on a SPARC disproved this; in the first case, it  performance. Tests using gcc on a SPARC disproved this; in the first case, it
494  made performance worse.  made performance worse.
495    
# Line 463  Returns:       MATCH_MATCH if matched Line 512  Returns:       MATCH_MATCH if matched
512  */  */
513    
514  static int  static int
515  match(REGISTER USPTR eptr, REGISTER const uschar *ecode, USPTR mstart,  match(REGISTER PCRE_PUCHAR eptr, REGISTER const pcre_uchar *ecode,
516    int offset_top, match_data *md, eptrblock *eptrb, unsigned int rdepth)    PCRE_PUCHAR mstart, int offset_top, match_data *md, eptrblock *eptrb,
517      unsigned int rdepth)
518  {  {
519  /* These variables do not need to be preserved over recursion in this function,  /* These variables do not need to be preserved over recursion in this function,
520  so they can be ordinary variables in all cases. Mark some of them with  so they can be ordinary variables in all cases. Mark some of them with
# Line 472  so they can be ordinary variables in all Line 522  so they can be ordinary variables in all
522    
523  register int  rrc;         /* Returns from recursive calls */  register int  rrc;         /* Returns from recursive calls */
524  register int  i;           /* Used for loops not involving calls to RMATCH() */  register int  i;           /* Used for loops not involving calls to RMATCH() */
525  register unsigned int c;   /* Character values not kept over RMATCH() calls */  register pcre_uint32 c;    /* Character values not kept over RMATCH() calls */
526  register BOOL utf8;        /* Local copy of UTF-8 flag for speed */  register BOOL utf;         /* Local copy of UTF flag for speed */
527    
528  BOOL minimize, possessive; /* Quantifier options */  BOOL minimize, possessive; /* Quantifier options */
529  BOOL caseless;  BOOL caseless;
530  int condcode;  int condcode;
531    
532  /* When recursion is not being used, all "local" variables that have to be  /* When recursion is not being used, all "local" variables that have to be
533  preserved over calls to RMATCH() are part of a "frame" which is obtained from  preserved over calls to RMATCH() are part of a "frame". We set up the top-level
534  heap storage. Set up the top-level frame here; others are obtained from the  frame on the stack here; subsequent instantiations are obtained from the heap
535  heap whenever RMATCH() does a "recursion". See the macro definitions above. */  whenever RMATCH() does a "recursion". See the macro definitions above. Putting
536    the top-level on the stack rather than malloc-ing them all gives a performance
537    boost in many cases where there is not much "recursion". */
538    
539  #ifdef NO_RECURSE  #ifdef NO_RECURSE
540  heapframe *frame = (heapframe *)(pcre_stack_malloc)(sizeof(heapframe));  heapframe *frame = (heapframe *)md->match_frames_base;
 if (frame == NULL) RRETURN(PCRE_ERROR_NOMEMORY);  
 frame->Xprevframe = NULL;            /* Marks the top level */  
541    
542  /* Copy in the original argument variables */  /* Copy in the original argument variables */
543    
# Line 513  HEAP_RECURSE: Line 563  HEAP_RECURSE:
563    
564  /* Ditto for the local variables */  /* Ditto for the local variables */
565    
566  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
567  #define charptr            frame->Xcharptr  #define charptr            frame->Xcharptr
568  #endif  #endif
569  #define callpat            frame->Xcallpat  #define callpat            frame->Xcallpat
# Line 571  declarations can be cut out in a block. Line 621  declarations can be cut out in a block.
621  below are for variables that do not have to be preserved over a recursive call  below are for variables that do not have to be preserved over a recursive call
622  to RMATCH(). */  to RMATCH(). */
623    
624  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
625  const uschar *charptr;  const pcre_uchar *charptr;
626  #endif  #endif
627  const uschar *callpat;  const pcre_uchar *callpat;
628  const uschar *data;  const pcre_uchar *data;
629  const uschar *next;  const pcre_uchar *next;
630  USPTR         pp;  PCRE_PUCHAR       pp;
631  const uschar *prev;  const pcre_uchar *prev;
632  USPTR         saved_eptr;  PCRE_PUCHAR       saved_eptr;
633    
634  recursion_info new_recursive;  recursion_info new_recursive;
635    
# Line 589  BOOL prev_is_word; Line 639  BOOL prev_is_word;
639    
640  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
641  int prop_type;  int prop_type;
642  int prop_value;  unsigned int prop_value;
643  int prop_fail_result;  int prop_fail_result;
644  int oclength;  int oclength;
645  uschar occhars[8];  pcre_uchar occhars[6];
646  #endif  #endif
647    
648  int codelink;  int codelink;
# Line 600  int ctype; Line 650  int ctype;
650  int length;  int length;
651  int max;  int max;
652  int min;  int min;
653  int number;  unsigned int number;
654  int offset;  int offset;
655  int op;  unsigned int op;
656  int save_capture_last;  pcre_int32 save_capture_last;
657  int save_offset1, save_offset2, save_offset3;  int save_offset1, save_offset2, save_offset3;
658  int stacksave[REC_STACK_SAVE_MAX];  int stacksave[REC_STACK_SAVE_MAX];
659    
660  eptrblock newptrb;  eptrblock newptrb;
661    
662    /* There is a special fudge for calling match() in a way that causes it to
663    measure the size of its basic stack frame when the stack is being used for
664    recursion. The second argument (ecode) being NULL triggers this behaviour. It
665    cannot normally ever be NULL. The return is the negated value of the frame
666    size. */
667    
668    if (ecode == NULL)
669      {
670      if (rdepth == 0)
671        return match((PCRE_PUCHAR)&rdepth, NULL, NULL, 0, NULL, NULL, 1);
672      else
673        {
674        int len = (char *)&rdepth - (char *)eptr;
675        return (len > 0)? -len : len;
676        }
677      }
678  #endif     /* NO_RECURSE */  #endif     /* NO_RECURSE */
679    
680  /* To save space on the stack and in the heap frame, I have doubled up on some  /* To save space on the stack and in the heap frame, I have doubled up on some
# Line 620  the alternative names that are used. */ Line 687  the alternative names that are used. */
687  #define code_offset   codelink  #define code_offset   codelink
688  #define condassert    condition  #define condassert    condition
689  #define matched_once  prev_is_word  #define matched_once  prev_is_word
690    #define foc           number
691    #define save_mark     data
692    
693  /* These statements are here to stop the compiler complaining about unitialized  /* These statements are here to stop the compiler complaining about unitialized
694  variables. */  variables. */
# Line 645  defined). However, RMATCH isn't like a f Line 714  defined). However, RMATCH isn't like a f
714  complicated macro. It has to be used in one particular way. This shouldn't,  complicated macro. It has to be used in one particular way. This shouldn't,
715  however, impact performance when true recursion is being used. */  however, impact performance when true recursion is being used. */
716    
717  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
718  utf8 = md->utf8;       /* Local copy of the flag */  utf = md->utf;       /* Local copy of the flag */
719  #else  #else
720  utf8 = FALSE;  utf = FALSE;
721  #endif  #endif
722    
723  /* First check that we haven't called match() too many times, or that we  /* First check that we haven't called match() too many times, or that we
# Line 689  for (;;) Line 758  for (;;)
758      case OP_MARK:      case OP_MARK:
759      md->nomatch_mark = ecode + 2;      md->nomatch_mark = ecode + 2;
760      md->mark = NULL;    /* In case previously set by assertion */      md->mark = NULL;    /* In case previously set by assertion */
761      RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode] + ecode[1], offset_top, md,      RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode] + ecode[1], offset_top, md,
762        eptrb, RM55);        eptrb, RM55);
763      if ((rrc == MATCH_MATCH || rrc == MATCH_ACCEPT) &&      if ((rrc == MATCH_MATCH || rrc == MATCH_ACCEPT) &&
764           md->mark == NULL) md->mark = ecode + 2;           md->mark == NULL) md->mark = ecode + 2;
# Line 702  for (;;) Line 771  for (;;)
771      unaltered. */      unaltered. */
772    
773      else if (rrc == MATCH_SKIP_ARG &&      else if (rrc == MATCH_SKIP_ARG &&
774          strcmp((char *)(ecode + 2), (char *)(md->start_match_ptr)) == 0)          STRCMP_UC_UC_TEST(ecode + 2, md->start_match_ptr) == 0)
775        {        {
776        md->start_match_ptr = eptr;        md->start_match_ptr = eptr;
777        RRETURN(MATCH_SKIP);        RRETURN(MATCH_SKIP);
# Line 712  for (;;) Line 781  for (;;)
781      case OP_FAIL:      case OP_FAIL:
782      RRETURN(MATCH_NOMATCH);      RRETURN(MATCH_NOMATCH);
783    
     /* COMMIT overrides PRUNE, SKIP, and THEN */  
   
784      case OP_COMMIT:      case OP_COMMIT:
785      RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,      RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md,
786        eptrb, RM52);        eptrb, RM52);
787      if (rrc != MATCH_NOMATCH && rrc != MATCH_PRUNE &&      if (rrc != MATCH_NOMATCH) RRETURN(rrc);
         rrc != MATCH_SKIP && rrc != MATCH_SKIP_ARG &&  
         rrc != MATCH_THEN)  
       RRETURN(rrc);  
788      RRETURN(MATCH_COMMIT);      RRETURN(MATCH_COMMIT);
789    
     /* PRUNE overrides THEN */  
   
790      case OP_PRUNE:      case OP_PRUNE:
791      RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,      RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md,
792        eptrb, RM51);        eptrb, RM51);
793      if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN) RRETURN(rrc);      if (rrc != MATCH_NOMATCH) RRETURN(rrc);
794      RRETURN(MATCH_PRUNE);      RRETURN(MATCH_PRUNE);
795    
796      case OP_PRUNE_ARG:      case OP_PRUNE_ARG:
797      md->nomatch_mark = ecode + 2;      md->nomatch_mark = ecode + 2;
798      md->mark = NULL;    /* In case previously set by assertion */      md->mark = NULL;    /* In case previously set by assertion */
799      RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode] + ecode[1], offset_top, md,      RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode] + ecode[1], offset_top, md,
800        eptrb, RM56);        eptrb, RM56);
801      if ((rrc == MATCH_MATCH || rrc == MATCH_ACCEPT) &&      if ((rrc == MATCH_MATCH || rrc == MATCH_ACCEPT) &&
802           md->mark == NULL) md->mark = ecode + 2;           md->mark == NULL) md->mark = ecode + 2;
803      if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN) RRETURN(rrc);      if (rrc != MATCH_NOMATCH) RRETURN(rrc);
804      RRETURN(MATCH_PRUNE);      RRETURN(MATCH_PRUNE);
805    
     /* SKIP overrides PRUNE and THEN */  
   
806      case OP_SKIP:      case OP_SKIP:
807      RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,      RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md,
808        eptrb, RM53);        eptrb, RM53);
809      if (rrc != MATCH_NOMATCH && rrc != MATCH_PRUNE && rrc != MATCH_THEN)      if (rrc != MATCH_NOMATCH) RRETURN(rrc);
       RRETURN(rrc);  
810      md->start_match_ptr = eptr;   /* Pass back current position */      md->start_match_ptr = eptr;   /* Pass back current position */
811      RRETURN(MATCH_SKIP);      RRETURN(MATCH_SKIP);
812    
813      /* Note that, for Perl compatibility, SKIP with an argument does NOT set      /* Note that, for Perl compatibility, SKIP with an argument does NOT set
814      nomatch_mark. There is a flag that disables this opcode when re-matching a      nomatch_mark. When a pattern match ends with a SKIP_ARG for which there was
815      pattern that ended with a SKIP for which there was not a matching MARK. */      not a matching mark, we have to re-run the match, ignoring the SKIP_ARG
816        that failed and any that preceed it (either they also failed, or were not
817        triggered). To do this, we maintain a count of executed SKIP_ARGs. If a
818        SKIP_ARG gets to top level, the match is re-run with md->ignore_skip_arg
819        set to the count of the one that failed. */
820    
821      case OP_SKIP_ARG:      case OP_SKIP_ARG:
822      if (md->ignore_skip_arg)      md->skip_arg_count++;
823        if (md->skip_arg_count <= md->ignore_skip_arg)
824        {        {
825        ecode += _pcre_OP_lengths[*ecode] + ecode[1];        ecode += PRIV(OP_lengths)[*ecode] + ecode[1];
826        break;        break;
827        }        }
828      RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode] + ecode[1], offset_top, md,      RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode] + ecode[1], offset_top, md,
829        eptrb, RM57);        eptrb, RM57);
830      if (rrc != MATCH_NOMATCH && rrc != MATCH_PRUNE && rrc != MATCH_THEN)      if (rrc != MATCH_NOMATCH) RRETURN(rrc);
831        RRETURN(rrc);  
   
832      /* Pass back the current skip name by overloading md->start_match_ptr and      /* Pass back the current skip name by overloading md->start_match_ptr and
833      returning the special MATCH_SKIP_ARG return code. This will either be      returning the special MATCH_SKIP_ARG return code. This will either be
834      caught by a matching MARK, or get to the top, where it causes a rematch      caught by a matching MARK, or get to the top, where it causes a rematch
835      with the md->ignore_skip_arg flag set. */      with md->ignore_skip_arg set to the value of md->skip_arg_count. */
836    
837      md->start_match_ptr = ecode + 2;      md->start_match_ptr = ecode + 2;
838      RRETURN(MATCH_SKIP_ARG);      RRETURN(MATCH_SKIP_ARG);
# Line 779  for (;;) Line 842  for (;;)
842      match pointer to do this. */      match pointer to do this. */
843    
844      case OP_THEN:      case OP_THEN:
845      RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,      RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md,
846        eptrb, RM54);        eptrb, RM54);
847      if (rrc != MATCH_NOMATCH) RRETURN(rrc);      if (rrc != MATCH_NOMATCH) RRETURN(rrc);
848      md->start_match_ptr = ecode;      md->start_match_ptr = ecode;
# Line 788  for (;;) Line 851  for (;;)
851      case OP_THEN_ARG:      case OP_THEN_ARG:
852      md->nomatch_mark = ecode + 2;      md->nomatch_mark = ecode + 2;
853      md->mark = NULL;    /* In case previously set by assertion */      md->mark = NULL;    /* In case previously set by assertion */
854      RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode] + ecode[1], offset_top,      RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode] + ecode[1], offset_top,
855        md, eptrb, RM58);        md, eptrb, RM58);
856      if ((rrc == MATCH_MATCH || rrc == MATCH_ACCEPT) &&      if ((rrc == MATCH_MATCH || rrc == MATCH_ACCEPT) &&
857           md->mark == NULL) md->mark = ecode + 2;           md->mark == NULL) md->mark = ecode + 2;
# Line 812  for (;;) Line 875  for (;;)
875      case OP_ONCE_NC:      case OP_ONCE_NC:
876      prev = ecode;      prev = ecode;
877      saved_eptr = eptr;      saved_eptr = eptr;
878        save_mark = md->mark;
879      do      do
880        {        {
881        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, eptrb, RM64);        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, eptrb, RM64);
# Line 830  for (;;) Line 894  for (;;)
894    
895        if (rrc != MATCH_NOMATCH) RRETURN(rrc);        if (rrc != MATCH_NOMATCH) RRETURN(rrc);
896        ecode += GET(ecode,1);        ecode += GET(ecode,1);
897          md->mark = save_mark;
898        }        }
899      while (*ecode == OP_ALT);      while (*ecode == OP_ALT);
900    
# Line 869  for (;;) Line 934  for (;;)
934        }        }
935      else  /* OP_KETRMAX */      else  /* OP_KETRMAX */
936        {        {
       md->match_function_type = MATCH_CBEGROUP;  
937        RMATCH(eptr, prev, offset_top, md, eptrb, RM66);        RMATCH(eptr, prev, offset_top, md, eptrb, RM66);
938        if (rrc != MATCH_NOMATCH) RRETURN(rrc);        if (rrc != MATCH_NOMATCH) RRETURN(rrc);
939        ecode += 1 + LINK_SIZE;        ecode += 1 + LINK_SIZE;
# Line 909  for (;;) Line 973  for (;;)
973        save_offset2 = md->offset_vector[offset+1];        save_offset2 = md->offset_vector[offset+1];
974        save_offset3 = md->offset_vector[md->offset_end - number];        save_offset3 = md->offset_vector[md->offset_end - number];
975        save_capture_last = md->capture_last;        save_capture_last = md->capture_last;
976          save_mark = md->mark;
977    
978        DPRINTF(("saving %d %d %d\n", save_offset1, save_offset2, save_offset3));        DPRINTF(("saving %d %d %d\n", save_offset1, save_offset2, save_offset3));
979        md->offset_vector[md->offset_end - number] =        md->offset_vector[md->offset_end - number] =
# Line 917  for (;;) Line 982  for (;;)
982        for (;;)        for (;;)
983          {          {
984          if (op >= OP_SBRA) md->match_function_type = MATCH_CBEGROUP;          if (op >= OP_SBRA) md->match_function_type = MATCH_CBEGROUP;
985          RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,          RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md,
986            eptrb, RM1);            eptrb, RM1);
987          if (rrc == MATCH_ONCE) break;  /* Backing up through an atomic group */          if (rrc == MATCH_ONCE) break;  /* Backing up through an atomic group */
988    
# Line 945  for (;;) Line 1010  for (;;)
1010          if (rrc != MATCH_NOMATCH) RRETURN(rrc);          if (rrc != MATCH_NOMATCH) RRETURN(rrc);
1011          md->capture_last = save_capture_last;          md->capture_last = save_capture_last;
1012          ecode += GET(ecode, 1);          ecode += GET(ecode, 1);
1013            md->mark = save_mark;
1014          if (*ecode != OP_ALT) break;          if (*ecode != OP_ALT) break;
1015          }          }
1016    
# Line 996  for (;;) Line 1062  for (;;)
1062    
1063      for (;;)      for (;;)
1064        {        {
1065        if (op >= OP_SBRA || op == OP_ONCE) md->match_function_type = MATCH_CBEGROUP;        if (op >= OP_SBRA || op == OP_ONCE)
1066            md->match_function_type = MATCH_CBEGROUP;
1067    
1068        /* If this is not a possibly empty group, and there are no (*THEN)s in        /* If this is not a possibly empty group, and there are no (*THEN)s in
1069        the pattern, and this is the final alternative, optimize as described        the pattern, and this is the final alternative, optimize as described
# Line 1004  for (;;) Line 1071  for (;;)
1071    
1072        else if (!md->hasthen && ecode[GET(ecode, 1)] != OP_ALT)        else if (!md->hasthen && ecode[GET(ecode, 1)] != OP_ALT)
1073          {          {
1074          ecode += _pcre_OP_lengths[*ecode];          ecode += PRIV(OP_lengths)[*ecode];
1075          goto TAIL_RECURSE;          goto TAIL_RECURSE;
1076          }          }
1077    
1078        /* In all other cases, we have to make another call to match(). */        /* In all other cases, we have to make another call to match(). */
1079    
1080        RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md, eptrb,        save_mark = md->mark;
1081          save_capture_last = md->capture_last;
1082          RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md, eptrb,
1083          RM2);          RM2);
1084    
1085        /* See comment in the code for capturing groups above about handling        /* See comment in the code for capturing groups above about handling
# Line 1028  for (;;) Line 1097  for (;;)
1097          {          {
1098          if (rrc == MATCH_ONCE)          if (rrc == MATCH_ONCE)
1099            {            {
1100            const uschar *scode = ecode;            const pcre_uchar *scode = ecode;
1101            if (*scode != OP_ONCE)           /* If not at start, find it */            if (*scode != OP_ONCE)           /* If not at start, find it */
1102              {              {
1103              while (*scode == OP_ALT) scode += GET(scode, 1);              while (*scode == OP_ALT) scode += GET(scode, 1);
# Line 1039  for (;;) Line 1108  for (;;)
1108          RRETURN(rrc);          RRETURN(rrc);
1109          }          }
1110        ecode += GET(ecode, 1);        ecode += GET(ecode, 1);
1111          md->mark = save_mark;
1112        if (*ecode != OP_ALT) break;        if (*ecode != OP_ALT) break;
1113          md->capture_last = save_capture_last;
1114        }        }
1115    
1116      RRETURN(MATCH_NOMATCH);      RRETURN(MATCH_NOMATCH);
# Line 1070  for (;;) Line 1141  for (;;)
1141      if (offset < md->offset_max)      if (offset < md->offset_max)
1142        {        {
1143        matched_once = FALSE;        matched_once = FALSE;
1144        code_offset = ecode - md->start_code;        code_offset = (int)(ecode - md->start_code);
1145    
1146        save_offset1 = md->offset_vector[offset];        save_offset1 = md->offset_vector[offset];
1147        save_offset2 = md->offset_vector[offset+1];        save_offset2 = md->offset_vector[offset+1];
# Line 1093  for (;;) Line 1164  for (;;)
1164          md->offset_vector[md->offset_end - number] =          md->offset_vector[md->offset_end - number] =
1165            (int)(eptr - md->start_subject);            (int)(eptr - md->start_subject);
1166          if (op >= OP_SBRA) md->match_function_type = MATCH_CBEGROUP;          if (op >= OP_SBRA) md->match_function_type = MATCH_CBEGROUP;
1167          RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,          RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md,
1168            eptrb, RM63);            eptrb, RM63);
1169          if (rrc == MATCH_KETRPOS)          if (rrc == MATCH_KETRPOS)
1170            {            {
# Line 1160  for (;;) Line 1231  for (;;)
1231    
1232      POSSESSIVE_NON_CAPTURE:      POSSESSIVE_NON_CAPTURE:
1233      matched_once = FALSE;      matched_once = FALSE;
1234      code_offset = ecode - md->start_code;      code_offset = (int)(ecode - md->start_code);
1235        save_capture_last = md->capture_last;
1236    
1237      for (;;)      for (;;)
1238        {        {
1239        if (op >= OP_SBRA) md->match_function_type = MATCH_CBEGROUP;        if (op >= OP_SBRA) md->match_function_type = MATCH_CBEGROUP;
1240        RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,        RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md,
1241          eptrb, RM48);          eptrb, RM48);
1242        if (rrc == MATCH_KETRPOS)        if (rrc == MATCH_KETRPOS)
1243          {          {
# Line 1190  for (;;) Line 1262  for (;;)
1262        if (rrc != MATCH_NOMATCH) RRETURN(rrc);        if (rrc != MATCH_NOMATCH) RRETURN(rrc);
1263        ecode += GET(ecode, 1);        ecode += GET(ecode, 1);
1264        if (*ecode != OP_ALT) break;        if (*ecode != OP_ALT) break;
1265          md->capture_last = save_capture_last;
1266        }        }
1267    
1268      if (matched_once || allow_zero)      if (matched_once || allow_zero)
# Line 1215  for (;;) Line 1288  for (;;)
1288    
1289      if (ecode[LINK_SIZE+1] == OP_CALLOUT)      if (ecode[LINK_SIZE+1] == OP_CALLOUT)
1290        {        {
1291        if (pcre_callout != NULL)        if (PUBL(callout) != NULL)
1292          {          {
1293          pcre_callout_block cb;          PUBL(callout_block) cb;
1294          cb.version          = 2;   /* Version 1 of the callout block */          cb.version          = 2;   /* Version 1 of the callout block */
1295          cb.callout_number   = ecode[LINK_SIZE+2];          cb.callout_number   = ecode[LINK_SIZE+2];
1296          cb.offset_vector    = md->offset_vector;          cb.offset_vector    = md->offset_vector;
1297    #if defined COMPILE_PCRE8
1298          cb.subject          = (PCRE_SPTR)md->start_subject;          cb.subject          = (PCRE_SPTR)md->start_subject;
1299    #elif defined COMPILE_PCRE16
1300            cb.subject          = (PCRE_SPTR16)md->start_subject;
1301    #elif defined COMPILE_PCRE32
1302            cb.subject          = (PCRE_SPTR32)md->start_subject;
1303    #endif
1304          cb.subject_length   = (int)(md->end_subject - md->start_subject);          cb.subject_length   = (int)(md->end_subject - md->start_subject);
1305          cb.start_match      = (int)(mstart - md->start_subject);          cb.start_match      = (int)(mstart - md->start_subject);
1306          cb.current_position = (int)(eptr - md->start_subject);          cb.current_position = (int)(eptr - md->start_subject);
1307          cb.pattern_position = GET(ecode, LINK_SIZE + 3);          cb.pattern_position = GET(ecode, LINK_SIZE + 3);
1308          cb.next_item_length = GET(ecode, 3 + 2*LINK_SIZE);          cb.next_item_length = GET(ecode, 3 + 2*LINK_SIZE);
1309          cb.capture_top      = offset_top/2;          cb.capture_top      = offset_top/2;
1310          cb.capture_last     = md->capture_last;          cb.capture_last     = md->capture_last & CAPLMASK;
1311            /* Internal change requires this for API compatibility. */
1312            if (cb.capture_last == 0) cb.capture_last = -1;
1313          cb.callout_data     = md->callout_data;          cb.callout_data     = md->callout_data;
1314          cb.mark             = md->nomatch_mark;          cb.mark             = md->nomatch_mark;
1315          if ((rrc = (*pcre_callout)(&cb)) > 0) RRETURN(MATCH_NOMATCH);          if ((rrc = (*PUBL(callout))(&cb)) > 0) RRETURN(MATCH_NOMATCH);
1316          if (rrc < 0) RRETURN(rrc);          if (rrc < 0) RRETURN(rrc);
1317          }          }
1318        ecode += _pcre_OP_lengths[OP_CALLOUT];        ecode += PRIV(OP_lengths)[OP_CALLOUT];
1319          codelink -= PRIV(OP_lengths)[OP_CALLOUT];
1320        }        }
1321    
1322      condcode = ecode[LINK_SIZE+1];      condcode = ecode[LINK_SIZE+1];
# Line 1250  for (;;) Line 1332  for (;;)
1332          }          }
1333        else        else
1334          {          {
1335          int recno = GET2(ecode, LINK_SIZE + 2);   /* Recursion group number*/          unsigned int recno = GET2(ecode, LINK_SIZE + 2);   /* Recursion group number*/
1336          condition = (recno == RREF_ANY || recno == md->recursive->group_num);          condition = (recno == RREF_ANY || recno == md->recursive->group_num);
1337    
1338          /* If the test is for recursion into a specific subpattern, and it is          /* If the test is for recursion into a specific subpattern, and it is
# Line 1260  for (;;) Line 1342  for (;;)
1342    
1343          if (!condition && condcode == OP_NRREF)          if (!condition && condcode == OP_NRREF)
1344            {            {
1345            uschar *slotA = md->name_table;            pcre_uchar *slotA = md->name_table;
1346            for (i = 0; i < md->name_count; i++)            for (i = 0; i < md->name_count; i++)
1347              {              {
1348              if (GET2(slotA, 0) == recno) break;              if (GET2(slotA, 0) == recno) break;
# Line 1273  for (;;) Line 1355  for (;;)
1355    
1356            if (i < md->name_count)            if (i < md->name_count)
1357              {              {
1358              uschar *slotB = slotA;              pcre_uchar *slotB = slotA;
1359              while (slotB > md->name_table)              while (slotB > md->name_table)
1360                {                {
1361                slotB -= md->name_entry_size;                slotB -= md->name_entry_size;
1362                if (strcmp((char *)slotA + 2, (char *)slotB + 2) == 0)                if (STRCMP_UC_UC(slotA + IMM2_SIZE, slotB + IMM2_SIZE) == 0)
1363                  {                  {
1364                  condition = GET2(slotB, 0) == md->recursive->group_num;                  condition = GET2(slotB, 0) == md->recursive->group_num;
1365                  if (condition) break;                  if (condition) break;
# Line 1293  for (;;) Line 1375  for (;;)
1375                for (i++; i < md->name_count; i++)                for (i++; i < md->name_count; i++)
1376                  {                  {
1377                  slotB += md->name_entry_size;                  slotB += md->name_entry_size;
1378                  if (strcmp((char *)slotA + 2, (char *)slotB + 2) == 0)                  if (STRCMP_UC_UC(slotA + IMM2_SIZE, slotB + IMM2_SIZE) == 0)
1379                    {                    {
1380                    condition = GET2(slotB, 0) == md->recursive->group_num;                    condition = GET2(slotB, 0) == md->recursive->group_num;
1381                    if (condition) break;                    if (condition) break;
# Line 1306  for (;;) Line 1388  for (;;)
1388    
1389          /* Chose branch according to the condition */          /* Chose branch according to the condition */
1390    
1391          ecode += condition? 3 : GET(ecode, 1);          ecode += condition? 1 + IMM2_SIZE : GET(ecode, 1);
1392          }          }
1393        }        }
1394    
# Line 1322  for (;;) Line 1404  for (;;)
1404    
1405        if (!condition && condcode == OP_NCREF)        if (!condition && condcode == OP_NCREF)
1406          {          {
1407          int refno = offset >> 1;          unsigned int refno = offset >> 1;
1408          uschar *slotA = md->name_table;          pcre_uchar *slotA = md->name_table;
1409    
1410          for (i = 0; i < md->name_count; i++)          for (i = 0; i < md->name_count; i++)
1411            {            {
# Line 1337  for (;;) Line 1419  for (;;)
1419    
1420          if (i < md->name_count)          if (i < md->name_count)
1421            {            {
1422            uschar *slotB = slotA;            pcre_uchar *slotB = slotA;
1423            while (slotB > md->name_table)            while (slotB > md->name_table)
1424              {              {
1425              slotB -= md->name_entry_size;              slotB -= md->name_entry_size;
1426              if (strcmp((char *)slotA + 2, (char *)slotB + 2) == 0)              if (STRCMP_UC_UC(slotA + IMM2_SIZE, slotB + IMM2_SIZE) == 0)
1427                {                {
1428                offset = GET2(slotB, 0) << 1;                offset = GET2(slotB, 0) << 1;
1429                condition = offset < offset_top &&                condition = offset < offset_top &&
# Line 1359  for (;;) Line 1441  for (;;)
1441              for (i++; i < md->name_count; i++)              for (i++; i < md->name_count; i++)
1442                {                {
1443                slotB += md->name_entry_size;                slotB += md->name_entry_size;
1444                if (strcmp((char *)slotA + 2, (char *)slotB + 2) == 0)                if (STRCMP_UC_UC(slotA + IMM2_SIZE, slotB + IMM2_SIZE) == 0)
1445                  {                  {
1446                  offset = GET2(slotB, 0) << 1;                  offset = GET2(slotB, 0) << 1;
1447                  condition = offset < offset_top &&                  condition = offset < offset_top &&
# Line 1374  for (;;) Line 1456  for (;;)
1456    
1457        /* Chose branch according to the condition */        /* Chose branch according to the condition */
1458    
1459        ecode += condition? 3 : GET(ecode, 1);        ecode += condition? 1 + IMM2_SIZE : GET(ecode, 1);
1460        }        }
1461    
1462      else if (condcode == OP_DEF)     /* DEFINE - always false */      else if (condcode == OP_DEF)     /* DEFINE - always false */
# Line 1450  for (;;) Line 1532  for (;;)
1532      to close any currently open capturing brackets. */      to close any currently open capturing brackets. */
1533    
1534      case OP_CLOSE:      case OP_CLOSE:
1535      number = GET2(ecode, 1);      number = GET2(ecode, 1);   /* Must be less than 65536 */
1536      offset = number << 1;      offset = number << 1;
1537    
1538  #ifdef PCRE_DEBUG  #ifdef PCRE_DEBUG
# Line 1458  for (;;) Line 1540  for (;;)
1540        printf("\n");        printf("\n");
1541  #endif  #endif
1542    
1543      md->capture_last = number;      md->capture_last = (md->capture_last & OVFLMASK) | number;
1544      if (offset >= md->offset_max) md->offset_overflow = TRUE; else      if (offset >= md->offset_max) md->capture_last |= OVFLBIT; else
1545        {        {
1546        md->offset_vector[offset] =        md->offset_vector[offset] =
1547          md->offset_vector[md->offset_end - number];          md->offset_vector[md->offset_end - number];
1548        md->offset_vector[offset+1] = (int)(eptr - md->start_subject);        md->offset_vector[offset+1] = (int)(eptr - md->start_subject);
1549        if (offset_top <= offset) offset_top = offset + 2;        if (offset_top <= offset) offset_top = offset + 2;
1550        }        }
1551      ecode += 3;      ecode += 1 + IMM2_SIZE;
1552      break;      break;
1553    
1554    
# Line 1513  for (;;) Line 1595  for (;;)
1595    
1596      case OP_ASSERT:      case OP_ASSERT:
1597      case OP_ASSERTBACK:      case OP_ASSERTBACK:
1598        save_mark = md->mark;
1599      if (md->match_function_type == MATCH_CONDASSERT)      if (md->match_function_type == MATCH_CONDASSERT)
1600        {        {
1601        condassert = TRUE;        condassert = TRUE;
# Line 1528  for (;;) Line 1611  for (;;)
1611          mstart = md->start_match_ptr;   /* In case \K reset it */          mstart = md->start_match_ptr;   /* In case \K reset it */
1612          break;          break;
1613          }          }
1614          md->mark = save_mark;
1615    
1616        /* PCRE does not allow THEN to escape beyond an assertion; it is treated        /* A COMMIT failure must fail the entire assertion, without trying any
1617        as NOMATCH. */        subsequent branches. */
1618    
1619          if (rrc == MATCH_COMMIT) RRETURN(MATCH_NOMATCH);
1620    
1621          /* PCRE does not allow THEN to escape beyond an assertion; it
1622          is treated as NOMATCH. */
1623    
1624        if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN) RRETURN(rrc);        if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN) RRETURN(rrc);
1625        ecode += GET(ecode, 1);        ecode += GET(ecode, 1);
# Line 1557  for (;;) Line 1646  for (;;)
1646    
1647      case OP_ASSERT_NOT:      case OP_ASSERT_NOT:
1648      case OP_ASSERTBACK_NOT:      case OP_ASSERTBACK_NOT:
1649        save_mark = md->mark;
1650      if (md->match_function_type == MATCH_CONDASSERT)      if (md->match_function_type == MATCH_CONDASSERT)
1651        {        {
1652        condassert = TRUE;        condassert = TRUE;
# Line 1567  for (;;) Line 1657  for (;;)
1657      do      do
1658        {        {
1659        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, NULL, RM5);        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, NULL, RM5);
1660          md->mark = save_mark;
1661        if (rrc == MATCH_MATCH || rrc == MATCH_ACCEPT) RRETURN(MATCH_NOMATCH);        if (rrc == MATCH_MATCH || rrc == MATCH_ACCEPT) RRETURN(MATCH_NOMATCH);
1662        if (rrc == MATCH_SKIP || rrc == MATCH_PRUNE || rrc == MATCH_COMMIT)        if (rrc == MATCH_SKIP || rrc == MATCH_PRUNE || rrc == MATCH_COMMIT)
1663          {          {
# Line 1593  for (;;) Line 1684  for (;;)
1684      back a number of characters, not bytes. */      back a number of characters, not bytes. */
1685    
1686      case OP_REVERSE:      case OP_REVERSE:
1687  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
1688      if (utf8)      if (utf)
1689        {        {
1690        i = GET(ecode, 1);        i = GET(ecode, 1);
1691        while (i-- > 0)        while (i-- > 0)
# Line 1625  for (;;) Line 1716  for (;;)
1716      function is able to force a failure. */      function is able to force a failure. */
1717    
1718      case OP_CALLOUT:      case OP_CALLOUT:
1719      if (pcre_callout != NULL)      if (PUBL(callout) != NULL)
1720        {        {
1721        pcre_callout_block cb;        PUBL(callout_block) cb;
1722        cb.version          = 2;   /* Version 1 of the callout block */        cb.version          = 2;   /* Version 1 of the callout block */
1723        cb.callout_number   = ecode[1];        cb.callout_number   = ecode[1];
1724        cb.offset_vector    = md->offset_vector;        cb.offset_vector    = md->offset_vector;
1725    #if defined COMPILE_PCRE8
1726        cb.subject          = (PCRE_SPTR)md->start_subject;        cb.subject          = (PCRE_SPTR)md->start_subject;
1727    #elif defined COMPILE_PCRE16
1728          cb.subject          = (PCRE_SPTR16)md->start_subject;
1729    #elif defined COMPILE_PCRE32
1730          cb.subject          = (PCRE_SPTR32)md->start_subject;
1731    #endif
1732        cb.subject_length   = (int)(md->end_subject - md->start_subject);        cb.subject_length   = (int)(md->end_subject - md->start_subject);
1733        cb.start_match      = (int)(mstart - md->start_subject);        cb.start_match      = (int)(mstart - md->start_subject);
1734        cb.current_position = (int)(eptr - md->start_subject);        cb.current_position = (int)(eptr - md->start_subject);
1735        cb.pattern_position = GET(ecode, 2);        cb.pattern_position = GET(ecode, 2);
1736        cb.next_item_length = GET(ecode, 2 + LINK_SIZE);        cb.next_item_length = GET(ecode, 2 + LINK_SIZE);
1737        cb.capture_top      = offset_top/2;        cb.capture_top      = offset_top/2;
1738        cb.capture_last     = md->capture_last;        cb.capture_last     = md->capture_last & CAPLMASK;
1739          /* Internal change requires this for API compatibility. */
1740          if (cb.capture_last == 0) cb.capture_last = -1;
1741        cb.callout_data     = md->callout_data;        cb.callout_data     = md->callout_data;
1742        cb.mark             = md->nomatch_mark;        cb.mark             = md->nomatch_mark;
1743        if ((rrc = (*pcre_callout)(&cb)) > 0) RRETURN(MATCH_NOMATCH);        if ((rrc = (*PUBL(callout))(&cb)) > 0) RRETURN(MATCH_NOMATCH);
1744        if (rrc < 0) RRETURN(rrc);        if (rrc < 0) RRETURN(rrc);
1745        }        }
1746      ecode += 2 + 2*LINK_SIZE;      ecode += 2 + 2*LINK_SIZE;
# Line 1667  for (;;) Line 1766  for (;;)
1766      case OP_RECURSE:      case OP_RECURSE:
1767        {        {
1768        recursion_info *ri;        recursion_info *ri;
1769        int recno;        unsigned int recno;
1770    
1771        callpat = md->start_code + GET(ecode, 1);        callpat = md->start_code + GET(ecode, 1);
1772        recno = (callpat == md->start_code)? 0 :        recno = (callpat == md->start_code)? 0 :
# Line 1684  for (;;) Line 1783  for (;;)
1783        /* Add to "recursing stack" */        /* Add to "recursing stack" */
1784    
1785        new_recursive.group_num = recno;        new_recursive.group_num = recno;
1786          new_recursive.saved_capture_last = md->capture_last;
1787        new_recursive.subject_position = eptr;        new_recursive.subject_position = eptr;
1788        new_recursive.prevrec = md->recursive;        new_recursive.prevrec = md->recursive;
1789        md->recursive = &new_recursive;        md->recursive = &new_recursive;
# Line 1700  for (;;) Line 1800  for (;;)
1800        else        else
1801          {          {
1802          new_recursive.offset_save =          new_recursive.offset_save =
1803            (int *)(pcre_malloc)(new_recursive.saved_max * sizeof(int));            (int *)(PUBL(malloc))(new_recursive.saved_max * sizeof(int));
1804          if (new_recursive.offset_save == NULL) RRETURN(PCRE_ERROR_NOMEMORY);          if (new_recursive.offset_save == NULL) RRETURN(PCRE_ERROR_NOMEMORY);
1805          }          }
1806        memcpy(new_recursive.offset_save, md->offset_vector,        memcpy(new_recursive.offset_save, md->offset_vector,
1807              new_recursive.saved_max * sizeof(int));              new_recursive.saved_max * sizeof(int));
1808    
1809        /* OK, now we can do the recursion. After processing each alternative,        /* OK, now we can do the recursion. After processing each alternative,
1810        restore the offset data. If there were nested recursions, md->recursive        restore the offset data and the last captured value. If there were nested
1811        might be changed, so reset it before looping. */        recursions, md->recursive might be changed, so reset it before looping.
1812          */
1813    
1814        DPRINTF(("Recursing into group %d\n", new_recursive.group_num));        DPRINTF(("Recursing into group %d\n", new_recursive.group_num));
1815        cbegroup = (*callpat >= OP_SBRA);        cbegroup = (*callpat >= OP_SBRA);
1816        do        do
1817          {          {
1818          if (cbegroup) md->match_function_type = MATCH_CBEGROUP;          if (cbegroup) md->match_function_type = MATCH_CBEGROUP;
1819          RMATCH(eptr, callpat + _pcre_OP_lengths[*callpat], offset_top,          RMATCH(eptr, callpat + PRIV(OP_lengths)[*callpat], offset_top,
1820            md, eptrb, RM6);            md, eptrb, RM6);
1821          memcpy(md->offset_vector, new_recursive.offset_save,          memcpy(md->offset_vector, new_recursive.offset_save,
1822              new_recursive.saved_max * sizeof(int));              new_recursive.saved_max * sizeof(int));
1823            md->capture_last = new_recursive.saved_capture_last;
1824          md->recursive = new_recursive.prevrec;          md->recursive = new_recursive.prevrec;
1825          if (rrc == MATCH_MATCH || rrc == MATCH_ACCEPT)          if (rrc == MATCH_MATCH || rrc == MATCH_ACCEPT)
1826            {            {
1827            DPRINTF(("Recursion matched\n"));            DPRINTF(("Recursion matched\n"));
1828            if (new_recursive.offset_save != stacksave)            if (new_recursive.offset_save != stacksave)
1829              (pcre_free)(new_recursive.offset_save);              (PUBL(free))(new_recursive.offset_save);
1830    
1831            /* 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
1832            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 1837  for (;;)
1837            goto RECURSION_MATCHED;        /* Exit loop; end processing */            goto RECURSION_MATCHED;        /* Exit loop; end processing */
1838            }            }
1839    
1840          /* 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
1841          as NOMATCH. */          recursion; they are treated as NOMATCH. These codes are defined in a
1842            range that can be tested for. Any other return code is an error. */
1843    
1844          else if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN)          else if (rrc != MATCH_NOMATCH &&
1845                     (rrc < MATCH_BACKTRACK_MIN || rrc > MATCH_BACKTRACK_MAX))
1846            {            {
1847            DPRINTF(("Recursion gave error %d\n", rrc));            DPRINTF(("Recursion gave error %d\n", rrc));
1848            if (new_recursive.offset_save != stacksave)            if (new_recursive.offset_save != stacksave)
1849              (pcre_free)(new_recursive.offset_save);              (PUBL(free))(new_recursive.offset_save);
1850            RRETURN(rrc);            RRETURN(rrc);
1851            }            }
1852    
# Line 1754  for (;;) Line 1858  for (;;)
1858        DPRINTF(("Recursion didn't match\n"));        DPRINTF(("Recursion didn't match\n"));
1859        md->recursive = new_recursive.prevrec;        md->recursive = new_recursive.prevrec;
1860        if (new_recursive.offset_save != stacksave)        if (new_recursive.offset_save != stacksave)
1861          (pcre_free)(new_recursive.offset_save);          (PUBL(free))(new_recursive.offset_save);
1862        RRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
1863        }        }
1864    
# Line 1868  for (;;) Line 1972  for (;;)
1972    
1973        /* Deal with capturing */        /* Deal with capturing */
1974    
1975        md->capture_last = number;        md->capture_last = (md->capture_last & OVFLMASK) | number;
1976        if (offset >= md->offset_max) md->offset_overflow = TRUE; else        if (offset >= md->offset_max) md->capture_last |= OVFLBIT; else
1977          {          {
1978          /* If offset is greater than offset_top, it means that we are          /* If offset is greater than offset_top, it means that we are
1979          "skipping" a capturing group, and that group's offsets must be marked          "skipping" a capturing group, and that group's offsets must be marked
# Line 1949  for (;;) Line 2053  for (;;)
2053          }          }
2054        if (*prev >= OP_SBRA)    /* Could match an empty string */        if (*prev >= OP_SBRA)    /* Could match an empty string */
2055          {          {
         md->match_function_type = MATCH_CBEGROUP;  
2056          RMATCH(eptr, prev, offset_top, md, eptrb, RM50);          RMATCH(eptr, prev, offset_top, md, eptrb, RM50);
2057          RRETURN(rrc);          RRETURN(rrc);
2058          }          }
# Line 1958  for (;;) Line 2061  for (;;)
2061        }        }
2062      else  /* OP_KETRMAX */      else  /* OP_KETRMAX */
2063        {        {
       if (*prev >= OP_SBRA) md->match_function_type = MATCH_CBEGROUP;  
2064        RMATCH(eptr, prev, offset_top, md, eptrb, RM13);        RMATCH(eptr, prev, offset_top, md, eptrb, RM13);
2065        if (rrc == MATCH_ONCE && md->once_target == prev) rrc = MATCH_NOMATCH;        if (rrc == MATCH_ONCE && md->once_target == prev) rrc = MATCH_NOMATCH;
2066        if (rrc != MATCH_NOMATCH) RRETURN(rrc);        if (rrc != MATCH_NOMATCH) RRETURN(rrc);
# Line 2015  for (;;) Line 2117  for (;;)
2117    
2118      case OP_DOLLM:      case OP_DOLLM:
2119      if (eptr < md->end_subject)      if (eptr < md->end_subject)
2120        { if (!IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH); }        {
2121          if (!IS_NEWLINE(eptr))
2122            {
2123            if (md->partial != 0 &&
2124                eptr + 1 >= md->end_subject &&
2125                NLBLOCK->nltype == NLTYPE_FIXED &&
2126                NLBLOCK->nllen == 2 &&
2127                RAWUCHARTEST(eptr) == NLBLOCK->nl[0])
2128              {
2129              md->hitend = TRUE;
2130              if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
2131              }
2132            RRETURN(MATCH_NOMATCH);
2133            }
2134          }
2135      else      else
2136        {        {
2137        if (md->noteol) RRETURN(MATCH_NOMATCH);        if (md->noteol) RRETURN(MATCH_NOMATCH);
# Line 2047  for (;;) Line 2163  for (;;)
2163      ASSERT_NL_OR_EOS:      ASSERT_NL_OR_EOS:
2164      if (eptr < md->end_subject &&      if (eptr < md->end_subject &&
2165          (!IS_NEWLINE(eptr) || eptr != md->end_subject - md->nllen))          (!IS_NEWLINE(eptr) || eptr != md->end_subject - md->nllen))
2166          {
2167          if (md->partial != 0 &&
2168              eptr + 1 >= md->end_subject &&
2169              NLBLOCK->nltype == NLTYPE_FIXED &&
2170              NLBLOCK->nllen == 2 &&
2171              RAWUCHARTEST(eptr) == NLBLOCK->nl[0])
2172            {
2173            md->hitend = TRUE;
2174            if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
2175            }
2176        RRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2177          }
2178    
2179      /* Either at end of string or \n before end. */      /* Either at end of string or \n before end. */
2180    
# Line 2066  for (;;) Line 2193  for (;;)
2193        be "non-word" characters. Remember the earliest consulted character for        be "non-word" characters. Remember the earliest consulted character for
2194        partial matching. */        partial matching. */
2195    
2196  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
2197        if (utf8)        if (utf)
2198          {          {
2199          /* Get status of previous character */          /* Get status of previous character */
2200    
2201          if (eptr == md->start_subject) prev_is_word = FALSE; else          if (eptr == md->start_subject) prev_is_word = FALSE; else
2202            {            {
2203            USPTR lastptr = eptr - 1;            PCRE_PUCHAR lastptr = eptr - 1;
2204            while((*lastptr & 0xc0) == 0x80) lastptr--;            BACKCHAR(lastptr);
2205            if (lastptr < md->start_used_ptr) md->start_used_ptr = lastptr;            if (lastptr < md->start_used_ptr) md->start_used_ptr = lastptr;
2206            GETCHAR(c, lastptr);            GETCHAR(c, lastptr);
2207  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
# Line 2139  for (;;) Line 2266  for (;;)
2266              }              }
2267            else            else
2268  #endif  #endif
2269            prev_is_word = ((md->ctypes[eptr[-1]] & ctype_word) != 0);            prev_is_word = MAX_255(eptr[-1])
2270                && ((md->ctypes[eptr[-1]] & ctype_word) != 0);
2271            }            }
2272    
2273          /* Get status of next character */          /* Get status of next character */
# Line 2162  for (;;) Line 2290  for (;;)
2290            }            }
2291          else          else
2292  #endif  #endif
2293          cur_is_word = ((md->ctypes[*eptr] & ctype_word) != 0);          cur_is_word = MAX_255(*eptr)
2294              && ((md->ctypes[*eptr] & ctype_word) != 0);
2295          }          }
2296    
2297        /* Now see if the situation is what we want */        /* Now see if the situation is what we want */
# Line 2173  for (;;) Line 2302  for (;;)
2302        }        }
2303      break;      break;
2304    
2305      /* Match a single character type; inline for speed */      /* Match any single character type except newline; have to take care with
2306        CRLF newlines and partial matching. */
2307    
2308      case OP_ANY:      case OP_ANY:
2309      if (IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH);      if (IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH);
2310        if (md->partial != 0 &&
2311            eptr + 1 >= md->end_subject &&
2312            NLBLOCK->nltype == NLTYPE_FIXED &&
2313            NLBLOCK->nllen == 2 &&
2314            RAWUCHARTEST(eptr) == NLBLOCK->nl[0])
2315          {
2316          md->hitend = TRUE;
2317          if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
2318          }
2319    
2320      /* Fall through */      /* Fall through */
2321    
2322        /* Match any single character whatsoever. */
2323    
2324      case OP_ALLANY:      case OP_ALLANY:
2325      if (eptr >= md->end_subject)   /* DO NOT merge the eptr++ here; it must */      if (eptr >= md->end_subject)   /* DO NOT merge the eptr++ here; it must */
2326        {                            /* not be updated before SCHECK_PARTIAL. */        {                            /* not be updated before SCHECK_PARTIAL. */
# Line 2186  for (;;) Line 2328  for (;;)
2328        RRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2329        }        }
2330      eptr++;      eptr++;
2331      if (utf8) while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;  #ifdef SUPPORT_UTF
2332        if (utf) ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);
2333    #endif
2334      ecode++;      ecode++;
2335      break;      break;
2336    
# Line 2211  for (;;) Line 2355  for (;;)
2355        }        }
2356      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2357      if (      if (
2358  #ifdef SUPPORT_UTF8  #if defined SUPPORT_UTF || !(defined COMPILE_PCRE8)
2359         c < 256 &&         c < 256 &&
2360  #endif  #endif
2361         (md->ctypes[c] & ctype_digit) != 0         (md->ctypes[c] & ctype_digit) != 0
# Line 2228  for (;;) Line 2372  for (;;)
2372        }        }
2373      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2374      if (      if (
2375  #ifdef SUPPORT_UTF8  #if defined SUPPORT_UTF || !(defined COMPILE_PCRE8)
2376         c >= 256 ||         c > 255 ||
2377  #endif  #endif
2378         (md->ctypes[c] & ctype_digit) == 0         (md->ctypes[c] & ctype_digit) == 0
2379         )         )
# Line 2245  for (;;) Line 2389  for (;;)
2389        }        }
2390      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2391      if (      if (
2392  #ifdef SUPPORT_UTF8  #if defined SUPPORT_UTF || !(defined COMPILE_PCRE8)
2393         c < 256 &&         c < 256 &&
2394  #endif  #endif
2395         (md->ctypes[c] & ctype_space) != 0         (md->ctypes[c] & ctype_space) != 0
# Line 2262  for (;;) Line 2406  for (;;)
2406        }        }
2407      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2408      if (      if (
2409  #ifdef SUPPORT_UTF8  #if defined SUPPORT_UTF || !(defined COMPILE_PCRE8)
2410         c >= 256 ||         c > 255 ||
2411  #endif  #endif
2412         (md->ctypes[c] & ctype_space) == 0         (md->ctypes[c] & ctype_space) == 0
2413         )         )
# Line 2279  for (;;) Line 2423  for (;;)
2423        }        }
2424      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2425      if (      if (
2426  #ifdef SUPPORT_UTF8  #if defined SUPPORT_UTF || !(defined COMPILE_PCRE8)
2427         c < 256 &&         c < 256 &&
2428  #endif  #endif
2429         (md->ctypes[c] & ctype_word) != 0         (md->ctypes[c] & ctype_word) != 0
# Line 2296  for (;;) Line 2440  for (;;)
2440        }        }
2441      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2442      if (      if (
2443  #ifdef SUPPORT_UTF8  #if defined SUPPORT_UTF || !(defined COMPILE_PCRE8)
2444         c >= 256 ||         c > 255 ||
2445  #endif  #endif
2446         (md->ctypes[c] & ctype_word) == 0         (md->ctypes[c] & ctype_word) == 0
2447         )         )
# Line 2316  for (;;) Line 2460  for (;;)
2460        {        {
2461        default: RRETURN(MATCH_NOMATCH);        default: RRETURN(MATCH_NOMATCH);
2462    
2463        case 0x000d:        case CHAR_CR:
2464        if (eptr < md->end_subject && *eptr == 0x0a) eptr++;        if (eptr >= md->end_subject)
2465            {
2466            SCHECK_PARTIAL();
2467            }
2468          else if (RAWUCHARTEST(eptr) == CHAR_LF) eptr++;
2469        break;        break;
2470    
2471        case 0x000a:        case CHAR_LF:
2472        break;        break;
2473    
2474        case 0x000b:        case CHAR_VT:
2475        case 0x000c:        case CHAR_FF:
2476        case 0x0085:        case CHAR_NEL:
2477    #ifndef EBCDIC
2478        case 0x2028:        case 0x2028:
2479        case 0x2029:        case 0x2029:
2480    #endif  /* Not EBCDIC */
2481        if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);        if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);
2482        break;        break;
2483        }        }
# Line 2343  for (;;) Line 2493  for (;;)
2493      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2494      switch(c)      switch(c)
2495        {        {
2496          HSPACE_CASES: RRETURN(MATCH_NOMATCH);  /* Byte and multibyte cases */
2497        default: break;        default: break;
       case 0x09:      /* HT */  
       case 0x20:      /* SPACE */  
       case 0xa0:      /* NBSP */  
       case 0x1680:    /* OGHAM SPACE MARK */  
       case 0x180e:    /* MONGOLIAN VOWEL SEPARATOR */  
       case 0x2000:    /* EN QUAD */  
       case 0x2001:    /* EM QUAD */  
       case 0x2002:    /* EN SPACE */  
       case 0x2003:    /* EM SPACE */  
       case 0x2004:    /* THREE-PER-EM SPACE */  
       case 0x2005:    /* FOUR-PER-EM SPACE */  
       case 0x2006:    /* SIX-PER-EM SPACE */  
       case 0x2007:    /* FIGURE SPACE */  
       case 0x2008:    /* PUNCTUATION SPACE */  
       case 0x2009:    /* THIN SPACE */  
       case 0x200A:    /* HAIR SPACE */  
       case 0x202f:    /* NARROW NO-BREAK SPACE */  
       case 0x205f:    /* MEDIUM MATHEMATICAL SPACE */  
       case 0x3000:    /* IDEOGRAPHIC SPACE */  
       RRETURN(MATCH_NOMATCH);  
2498        }        }
2499      ecode++;      ecode++;
2500      break;      break;
# Line 2377  for (;;) Line 2508  for (;;)
2508      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2509      switch(c)      switch(c)
2510        {        {
2511          HSPACE_CASES: break;  /* Byte and multibyte cases */
2512        default: RRETURN(MATCH_NOMATCH);        default: RRETURN(MATCH_NOMATCH);
       case 0x09:      /* HT */  
       case 0x20:      /* SPACE */  
       case 0xa0:      /* NBSP */  
       case 0x1680:    /* OGHAM SPACE MARK */  
       case 0x180e:    /* MONGOLIAN VOWEL SEPARATOR */  
       case 0x2000:    /* EN QUAD */  
       case 0x2001:    /* EM QUAD */  
       case 0x2002:    /* EN SPACE */  
       case 0x2003:    /* EM SPACE */  
       case 0x2004:    /* THREE-PER-EM SPACE */  
       case 0x2005:    /* FOUR-PER-EM SPACE */  
       case 0x2006:    /* SIX-PER-EM SPACE */  
       case 0x2007:    /* FIGURE SPACE */  
       case 0x2008:    /* PUNCTUATION SPACE */  
       case 0x2009:    /* THIN SPACE */  
       case 0x200A:    /* HAIR SPACE */  
       case 0x202f:    /* NARROW NO-BREAK SPACE */  
       case 0x205f:    /* MEDIUM MATHEMATICAL SPACE */  
       case 0x3000:    /* IDEOGRAPHIC SPACE */  
       break;  
2513        }        }
2514      ecode++;      ecode++;
2515      break;      break;
# Line 2411  for (;;) Line 2523  for (;;)
2523      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2524      switch(c)      switch(c)
2525        {        {
2526          VSPACE_CASES: RRETURN(MATCH_NOMATCH);
2527        default: break;        default: break;
       case 0x0a:      /* LF */  
       case 0x0b:      /* VT */  
       case 0x0c:      /* FF */  
       case 0x0d:      /* CR */  
       case 0x85:      /* NEL */  
       case 0x2028:    /* LINE SEPARATOR */  
       case 0x2029:    /* PARAGRAPH SEPARATOR */  
       RRETURN(MATCH_NOMATCH);  
2528        }        }
2529      ecode++;      ecode++;
2530      break;      break;
# Line 2433  for (;;) Line 2538  for (;;)
2538      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2539      switch(c)      switch(c)
2540        {        {
2541          VSPACE_CASES: break;
2542        default: RRETURN(MATCH_NOMATCH);        default: RRETURN(MATCH_NOMATCH);
       case 0x0a:      /* LF */  
       case 0x0b:      /* VT */  
       case 0x0c:      /* FF */  
       case 0x0d:      /* CR */  
       case 0x85:      /* NEL */  
       case 0x2028:    /* LINE SEPARATOR */  
       case 0x2029:    /* PARAGRAPH SEPARATOR */  
       break;  
2543        }        }
2544      ecode++;      ecode++;
2545      break;      break;
# Line 2459  for (;;) Line 2557  for (;;)
2557        }        }
2558      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2559        {        {
2560          const pcre_uint32 *cp;
2561        const ucd_record *prop = GET_UCD(c);        const ucd_record *prop = GET_UCD(c);
2562    
2563        switch(ecode[1])        switch(ecode[1])
# Line 2475  for (;;) Line 2574  for (;;)
2574          break;          break;
2575    
2576          case PT_GC:          case PT_GC:
2577          if ((ecode[2] != _pcre_ucp_gentype[prop->chartype]) == (op == OP_PROP))          if ((ecode[2] != PRIV(ucp_gentype)[prop->chartype]) == (op == OP_PROP))
2578            RRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
2579          break;          break;
2580    
# Line 2492  for (;;) Line 2591  for (;;)
2591          /* These are specials */          /* These are specials */
2592    
2593          case PT_ALNUM:          case PT_ALNUM:
2594          if ((_pcre_ucp_gentype[prop->chartype] == ucp_L ||          if ((PRIV(ucp_gentype)[prop->chartype] == ucp_L ||
2595               _pcre_ucp_gentype[prop->chartype] == ucp_N) == (op == OP_NOTPROP))               PRIV(ucp_gentype)[prop->chartype] == ucp_N) == (op == OP_NOTPROP))
2596            RRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
2597          break;          break;
2598    
2599          case PT_SPACE:    /* Perl space */          case PT_SPACE:    /* Perl space */
2600          if ((_pcre_ucp_gentype[prop->chartype] == ucp_Z ||          if ((PRIV(ucp_gentype)[prop->chartype] == ucp_Z ||
2601               c == CHAR_HT || c == CHAR_NL || c == CHAR_FF || c == CHAR_CR)               c == CHAR_HT || c == CHAR_NL || c == CHAR_FF || c == CHAR_CR)
2602                 == (op == OP_NOTPROP))                 == (op == OP_NOTPROP))
2603            RRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
2604          break;          break;
2605    
2606          case PT_PXSPACE:  /* POSIX space */          case PT_PXSPACE:  /* POSIX space */
2607          if ((_pcre_ucp_gentype[prop->chartype] == ucp_Z ||          if ((PRIV(ucp_gentype)[prop->chartype] == ucp_Z ||
2608               c == CHAR_HT || c == CHAR_NL || c == CHAR_VT ||               c == CHAR_HT || c == CHAR_NL || c == CHAR_VT ||
2609               c == CHAR_FF || c == CHAR_CR)               c == CHAR_FF || c == CHAR_CR)
2610                 == (op == OP_NOTPROP))                 == (op == OP_NOTPROP))
# Line 2513  for (;;) Line 2612  for (;;)
2612          break;          break;
2613    
2614          case PT_WORD:          case PT_WORD:
2615          if ((_pcre_ucp_gentype[prop->chartype] == ucp_L ||          if ((PRIV(ucp_gentype)[prop->chartype] == ucp_L ||
2616               _pcre_ucp_gentype[prop->chartype] == ucp_N ||               PRIV(ucp_gentype)[prop->chartype] == ucp_N ||
2617               c == CHAR_UNDERSCORE) == (op == OP_NOTPROP))               c == CHAR_UNDERSCORE) == (op == OP_NOTPROP))
2618            RRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
2619          break;          break;
2620    
2621            case PT_CLIST:
2622            cp = PRIV(ucd_caseless_sets) + ecode[2];
2623            for (;;)
2624              {
2625              if (c < *cp)
2626                { if (op == OP_PROP) { RRETURN(MATCH_NOMATCH); } else break; }
2627              if (c == *cp++)
2628                { if (op == OP_PROP) break; else { RRETURN(MATCH_NOMATCH); } }
2629              }
2630            break;
2631    
2632            case PT_UCNC:
2633            if ((c == CHAR_DOLLAR_SIGN || c == CHAR_COMMERCIAL_AT ||
2634                 c == CHAR_GRAVE_ACCENT || (c >= 0xa0 && c <= 0xd7ff) ||
2635                 c >= 0xe000) == (op == OP_NOTPROP))
2636              RRETURN(MATCH_NOMATCH);
2637            break;
2638    
2639          /* This should never occur */          /* This should never occur */
2640    
2641          default:          default:
# Line 2538  for (;;) Line 2655  for (;;)
2655        SCHECK_PARTIAL();        SCHECK_PARTIAL();
2656        RRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2657        }        }
2658      GETCHARINCTEST(c, eptr);      else
     if (UCD_CATEGORY(c) == ucp_M) RRETURN(MATCH_NOMATCH);  
     while (eptr < md->end_subject)  
2659        {        {
2660        int len = 1;        int lgb, rgb;
2661        if (!utf8) c = *eptr; else { GETCHARLEN(c, eptr, len); }        GETCHARINCTEST(c, eptr);
2662        if (UCD_CATEGORY(c) != ucp_M) break;        lgb = UCD_GRAPHBREAK(c);
2663        eptr += len;        while (eptr < md->end_subject)
2664            {
2665            int len = 1;
2666            if (!utf) c = *eptr; else { GETCHARLEN(c, eptr, len); }
2667            rgb = UCD_GRAPHBREAK(c);
2668            if ((PRIV(ucp_gbtable)[lgb] & (1 << rgb)) == 0) break;
2669            lgb = rgb;
2670            eptr += len;
2671            }
2672        }        }
2673        CHECK_PARTIAL();
2674      ecode++;      ecode++;
2675      break;      break;
2676  #endif  #endif  /* SUPPORT_UCP */
2677    
2678    
2679      /* 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 2688  for (;;)
2688      case OP_REFI:      case OP_REFI:
2689      caseless = op == OP_REFI;      caseless = op == OP_REFI;
2690      offset = GET2(ecode, 1) << 1;               /* Doubled ref number */      offset = GET2(ecode, 1) << 1;               /* Doubled ref number */
2691      ecode += 3;      ecode += 1 + IMM2_SIZE;
2692    
2693      /* If the reference is unset, there are two possibilities:      /* If the reference is unset, there are two possibilities:
2694    
# Line 2604  for (;;) Line 2728  for (;;)
2728        case OP_CRMINRANGE:        case OP_CRMINRANGE:
2729        minimize = (*ecode == OP_CRMINRANGE);        minimize = (*ecode == OP_CRMINRANGE);
2730        min = GET2(ecode, 1);        min = GET2(ecode, 1);
2731        max = GET2(ecode, 3);        max = GET2(ecode, 1 + IMM2_SIZE);
2732        if (max == 0) max = INT_MAX;        if (max == 0) max = INT_MAX;
2733        ecode += 5;        ecode += 1 + 2 * IMM2_SIZE;
2734        break;        break;
2735    
2736        default:               /* No repeat follows */        default:               /* No repeat follows */
2737        if ((length = match_ref(offset, eptr, length, md, caseless)) < 0)        if ((length = match_ref(offset, eptr, length, md, caseless)) < 0)
2738          {          {
2739            if (length == -2) eptr = md->end_subject;   /* Partial match */
2740          CHECK_PARTIAL();          CHECK_PARTIAL();
2741          RRETURN(MATCH_NOMATCH);          RRETURN(MATCH_NOMATCH);
2742          }          }
# Line 2620  for (;;) Line 2745  for (;;)
2745        }        }
2746    
2747      /* Handle repeated back references. If the length of the reference is      /* Handle repeated back references. If the length of the reference is
2748      zero, just continue with the main loop. */      zero, just continue with the main loop. If the length is negative, it
2749        means the reference is unset in non-Java-compatible mode. If the minimum is
2750        zero, we can continue at the same level without recursion. For any other
2751        minimum, carrying on will result in NOMATCH. */
2752    
2753      if (length == 0) continue;      if (length == 0) continue;
2754        if (length < 0 && min == 0) continue;
2755    
2756      /* First, ensure the minimum number of matches are present. We get back      /* First, ensure the minimum number of matches are present. We get back
2757      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 2762  for (;;)
2762        int slength;        int slength;
2763        if ((slength = match_ref(offset, eptr, length, md, caseless)) < 0)        if ((slength = match_ref(offset, eptr, length, md, caseless)) < 0)
2764          {          {
2765            if (slength == -2) eptr = md->end_subject;   /* Partial match */
2766          CHECK_PARTIAL();          CHECK_PARTIAL();
2767          RRETURN(MATCH_NOMATCH);          RRETURN(MATCH_NOMATCH);
2768          }          }
# Line 2656  for (;;) Line 2786  for (;;)
2786          if (fi >= max) RRETURN(MATCH_NOMATCH);          if (fi >= max) RRETURN(MATCH_NOMATCH);
2787          if ((slength = match_ref(offset, eptr, length, md, caseless)) < 0)          if ((slength = match_ref(offset, eptr, length, md, caseless)) < 0)
2788            {            {
2789              if (slength == -2) eptr = md->end_subject;   /* Partial match */
2790            CHECK_PARTIAL();            CHECK_PARTIAL();
2791            RRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
2792            }            }
# Line 2674  for (;;) Line 2805  for (;;)
2805          int slength;          int slength;
2806          if ((slength = match_ref(offset, eptr, length, md, caseless)) < 0)          if ((slength = match_ref(offset, eptr, length, md, caseless)) < 0)
2807            {            {
2808            CHECK_PARTIAL();            /* Can't use CHECK_PARTIAL because we don't want to update eptr in
2809              the soft partial matching case. */
2810    
2811              if (slength == -2 && md->partial != 0 &&
2812                  md->end_subject > md->start_used_ptr)
2813                {
2814                md->hitend = TRUE;
2815                if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
2816                }
2817            break;            break;
2818            }            }
2819          eptr += slength;          eptr += slength;
2820          }          }
2821    
2822        while (eptr >= pp)        while (eptr >= pp)
2823          {          {
2824          RMATCH(eptr, ecode, offset_top, md, eptrb, RM15);          RMATCH(eptr, ecode, offset_top, md, eptrb, RM15);
# Line 2703  for (;;) Line 2843  for (;;)
2843      case OP_NCLASS:      case OP_NCLASS:
2844      case OP_CLASS:      case OP_CLASS:
2845        {        {
2846          /* The data variable is saved across frames, so the byte map needs to
2847          be stored there. */
2848    #define BYTE_MAP ((pcre_uint8 *)data)
2849        data = ecode + 1;                /* Save for matching */        data = ecode + 1;                /* Save for matching */
2850        ecode += 33;                     /* Advance past the item */        ecode += 1 + (32 / sizeof(pcre_uchar)); /* Advance past the item */
2851    
2852        switch (*ecode)        switch (*ecode)
2853          {          {
# Line 2725  for (;;) Line 2868  for (;;)
2868          case OP_CRMINRANGE:          case OP_CRMINRANGE:
2869          minimize = (*ecode == OP_CRMINRANGE);          minimize = (*ecode == OP_CRMINRANGE);
2870          min = GET2(ecode, 1);          min = GET2(ecode, 1);
2871          max = GET2(ecode, 3);          max = GET2(ecode, 1 + IMM2_SIZE);
2872          if (max == 0) max = INT_MAX;          if (max == 0) max = INT_MAX;
2873          ecode += 5;          ecode += 1 + 2 * IMM2_SIZE;
2874          break;          break;
2875    
2876          default:               /* No repeat follows */          default:               /* No repeat follows */
# Line 2737  for (;;) Line 2880  for (;;)
2880    
2881        /* First, ensure the minimum number of matches are present. */        /* First, ensure the minimum number of matches are present. */
2882    
2883  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
2884        /* UTF-8 mode */        if (utf)
       if (utf8)  
2885          {          {
2886          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
2887            {            {
# Line 2754  for (;;) Line 2896  for (;;)
2896              if (op == OP_CLASS) RRETURN(MATCH_NOMATCH);              if (op == OP_CLASS) RRETURN(MATCH_NOMATCH);
2897              }              }
2898            else            else
2899              {              if ((BYTE_MAP[c/8] & (1 << (c&7))) == 0) RRETURN(MATCH_NOMATCH);
             if ((data[c/8] & (1 << (c&7))) == 0) RRETURN(MATCH_NOMATCH);  
             }  
2900            }            }
2901          }          }
2902        else        else
2903  #endif  #endif
2904        /* Not UTF-8 mode */        /* Not UTF mode */
2905          {          {
2906          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
2907            {            {
# Line 2771  for (;;) Line 2911  for (;;)
2911              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
2912              }              }
2913            c = *eptr++;            c = *eptr++;
2914            if ((data[c/8] & (1 << (c&7))) == 0) RRETURN(MATCH_NOMATCH);  #ifndef COMPILE_PCRE8
2915              if (c > 255)
2916                {
2917                if (op == OP_CLASS) RRETURN(MATCH_NOMATCH);
2918                }
2919              else
2920    #endif
2921                if ((BYTE_MAP[c/8] & (1 << (c&7))) == 0) RRETURN(MATCH_NOMATCH);
2922            }            }
2923          }          }
2924    
# Line 2785  for (;;) Line 2932  for (;;)
2932    
2933        if (minimize)        if (minimize)
2934          {          {
2935  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
2936          /* UTF-8 mode */          if (utf)
         if (utf8)  
2937            {            {
2938            for (fi = min;; fi++)            for (fi = min;; fi++)
2939              {              {
# Line 2805  for (;;) Line 2951  for (;;)
2951                if (op == OP_CLASS) RRETURN(MATCH_NOMATCH);                if (op == OP_CLASS) RRETURN(MATCH_NOMATCH);
2952                }                }
2953              else              else
2954                {                if ((BYTE_MAP[c/8] & (1 << (c&7))) == 0) RRETURN(MATCH_NOMATCH);
               if ((data[c/8] & (1 << (c&7))) == 0) RRETURN(MATCH_NOMATCH);  
               }  
2955              }              }
2956            }            }
2957          else          else
2958  #endif  #endif
2959          /* Not UTF-8 mode */          /* Not UTF mode */
2960            {            {
2961            for (fi = min;; fi++)            for (fi = min;; fi++)
2962              {              {
# Line 2825  for (;;) Line 2969  for (;;)
2969                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
2970                }                }
2971              c = *eptr++;              c = *eptr++;
2972              if ((data[c/8] & (1 << (c&7))) == 0) RRETURN(MATCH_NOMATCH);  #ifndef COMPILE_PCRE8
2973                if (c > 255)
2974                  {
2975                  if (op == OP_CLASS) RRETURN(MATCH_NOMATCH);
2976                  }
2977                else
2978    #endif
2979                  if ((BYTE_MAP[c/8] & (1 << (c&7))) == 0) RRETURN(MATCH_NOMATCH);
2980              }              }
2981            }            }
2982          /* Control never gets here */          /* Control never gets here */
# Line 2837  for (;;) Line 2988  for (;;)
2988          {          {
2989          pp = eptr;          pp = eptr;
2990    
2991  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
2992          /* UTF-8 mode */          if (utf)
         if (utf8)  
2993            {            {
2994            for (i = min; i < max; i++)            for (i = min; i < max; i++)
2995              {              {
# Line 2855  for (;;) Line 3005  for (;;)
3005                if (op == OP_CLASS) break;                if (op == OP_CLASS) break;
3006                }                }
3007              else              else
3008                {                if ((BYTE_MAP[c/8] & (1 << (c&7))) == 0) break;
               if ((data[c/8] & (1 << (c&7))) == 0) break;  
               }  
3009              eptr += len;              eptr += len;
3010              }              }
3011            for (;;)            for (;;)
# Line 2870  for (;;) Line 3018  for (;;)
3018            }            }
3019          else          else
3020  #endif  #endif
3021            /* Not UTF-8 mode */            /* Not UTF mode */
3022            {            {
3023            for (i = min; i < max; i++)            for (i = min; i < max; i++)
3024              {              {
# Line 2880  for (;;) Line 3028  for (;;)
3028                break;                break;
3029                }                }
3030              c = *eptr;              c = *eptr;
3031              if ((data[c/8] & (1 << (c&7))) == 0) break;  #ifndef COMPILE_PCRE8
3032                if (c > 255)
3033                  {
3034                  if (op == OP_CLASS) break;
3035                  }
3036                else
3037    #endif
3038                  if ((BYTE_MAP[c/8] & (1 << (c&7))) == 0) break;
3039              eptr++;              eptr++;
3040              }              }
3041            while (eptr >= pp)            while (eptr >= pp)
# Line 2893  for (;;) Line 3048  for (;;)
3048    
3049          RRETURN(MATCH_NOMATCH);          RRETURN(MATCH_NOMATCH);
3050          }          }
3051    #undef BYTE_MAP
3052        }        }
3053      /* Control never gets here */      /* Control never gets here */
3054    
# Line 2901  for (;;) Line 3057  for (;;)
3057      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
3058      mode, because Unicode properties are supported in non-UTF-8 mode. */      mode, because Unicode properties are supported in non-UTF-8 mode. */
3059    
3060  #ifdef SUPPORT_UTF8  #if defined SUPPORT_UTF || !defined COMPILE_PCRE8
3061      case OP_XCLASS:      case OP_XCLASS:
3062        {        {
3063        data = ecode + 1 + LINK_SIZE;                /* Save for matching */        data = ecode + 1 + LINK_SIZE;                /* Save for matching */
# Line 2926  for (;;) Line 3082  for (;;)
3082          case OP_CRMINRANGE:          case OP_CRMINRANGE:
3083          minimize = (*ecode == OP_CRMINRANGE);          minimize = (*ecode == OP_CRMINRANGE);
3084          min = GET2(ecode, 1);          min = GET2(ecode, 1);
3085          max = GET2(ecode, 3);          max = GET2(ecode, 1 + IMM2_SIZE);
3086          if (max == 0) max = INT_MAX;          if (max == 0) max = INT_MAX;
3087          ecode += 5;          ecode += 1 + 2 * IMM2_SIZE;
3088          break;          break;
3089    
3090          default:               /* No repeat follows */          default:               /* No repeat follows */
# Line 2946  for (;;) Line 3102  for (;;)
3102            RRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
3103            }            }
3104          GETCHARINCTEST(c, eptr);          GETCHARINCTEST(c, eptr);
3105          if (!_pcre_xclass(c, data)) RRETURN(MATCH_NOMATCH);          if (!PRIV(xclass)(c, data, utf)) RRETURN(MATCH_NOMATCH);
3106          }          }
3107    
3108        /* 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 3126  for (;;)
3126              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
3127              }              }
3128            GETCHARINCTEST(c, eptr);            GETCHARINCTEST(c, eptr);
3129            if (!_pcre_xclass(c, data)) RRETURN(MATCH_NOMATCH);            if (!PRIV(xclass)(c, data, utf)) RRETURN(MATCH_NOMATCH);
3130            }            }
3131          /* Control never gets here */          /* Control never gets here */
3132          }          }
# Line 2988  for (;;) Line 3144  for (;;)
3144              SCHECK_PARTIAL();              SCHECK_PARTIAL();
3145              break;              break;
3146              }              }
3147    #ifdef SUPPORT_UTF
3148            GETCHARLENTEST(c, eptr, len);            GETCHARLENTEST(c, eptr, len);
3149            if (!_pcre_xclass(c, data)) break;  #else
3150              c = *eptr;
3151    #endif
3152              if (!PRIV(xclass)(c, data, utf)) break;
3153            eptr += len;            eptr += len;
3154            }            }
3155          for(;;)          for(;;)
# Line 2997  for (;;) Line 3157  for (;;)
3157            RMATCH(eptr, ecode, offset_top, md, eptrb, RM21);            RMATCH(eptr, ecode, offset_top, md, eptrb, RM21);
3158            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3159            if (eptr-- == pp) break;        /* Stop if tried at original pos */            if (eptr-- == pp) break;        /* Stop if tried at original pos */
3160            if (utf8) BACKCHAR(eptr);  #ifdef SUPPORT_UTF
3161              if (utf) BACKCHAR(eptr);
3162    #endif
3163            }            }
3164          RRETURN(MATCH_NOMATCH);          RRETURN(MATCH_NOMATCH);
3165          }          }
# Line 3009  for (;;) Line 3171  for (;;)
3171      /* Match a single character, casefully */      /* Match a single character, casefully */
3172    
3173      case OP_CHAR:      case OP_CHAR:
3174  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
3175      if (utf8)      if (utf)
3176        {        {
3177        length = 1;        length = 1;
3178        ecode++;        ecode++;
# Line 3020  for (;;) Line 3182  for (;;)
3182          CHECK_PARTIAL();             /* Not SCHECK_PARTIAL() */          CHECK_PARTIAL();             /* Not SCHECK_PARTIAL() */
3183          RRETURN(MATCH_NOMATCH);          RRETURN(MATCH_NOMATCH);
3184          }          }
3185        while (length-- > 0) if (*ecode++ != *eptr++) RRETURN(MATCH_NOMATCH);        while (length-- > 0) if (*ecode++ != RAWUCHARINC(eptr)) RRETURN(MATCH_NOMATCH);
3186        }        }
3187      else      else
3188  #endif  #endif
3189        /* Not UTF mode */
     /* Non-UTF-8 mode */  
3190        {        {
3191        if (md->end_subject - eptr < 1)        if (md->end_subject - eptr < 1)
3192          {          {
# Line 3037  for (;;) Line 3198  for (;;)
3198        }        }
3199      break;      break;
3200    
3201      /* Match a single character, caselessly */      /* Match a single character, caselessly. If we are at the end of the
3202        subject, give up immediately. */
3203    
3204      case OP_CHARI:      case OP_CHARI:
3205  #ifdef SUPPORT_UTF8      if (eptr >= md->end_subject)
3206      if (utf8)        {
3207          SCHECK_PARTIAL();
3208          RRETURN(MATCH_NOMATCH);
3209          }
3210    
3211    #ifdef SUPPORT_UTF
3212        if (utf)
3213        {        {
3214        length = 1;        length = 1;
3215        ecode++;        ecode++;
3216        GETCHARLEN(fc, ecode, length);        GETCHARLEN(fc, ecode, length);
3217    
       if (length > md->end_subject - eptr)  
         {  
         CHECK_PARTIAL();             /* Not SCHECK_PARTIAL() */  
         RRETURN(MATCH_NOMATCH);  
         }  
   
3218        /* If the pattern character's value is < 128, we have only one byte, and        /* If the pattern character's value is < 128, we have only one byte, and
3219        can use the fast lookup table. */        we know that its other case must also be one byte long, so we can use the
3220          fast lookup table. We know that there is at least one byte left in the
3221          subject. */
3222    
3223        if (fc < 128)        if (fc < 128)
3224          {          {
3225          if (md->lcc[*ecode++] != md->lcc[*eptr++]) RRETURN(MATCH_NOMATCH);          pcre_uint32 cc = RAWUCHAR(eptr);
3226            if (md->lcc[fc] != TABLE_GET(cc, md->lcc, cc)) RRETURN(MATCH_NOMATCH);
3227            ecode++;
3228            eptr++;
3229          }          }
3230    
3231        /* Otherwise we must pick up the subject character */        /* Otherwise we must pick up the subject character. Note that we cannot
3232          use the value of "length" to check for sufficient bytes left, because the
3233          other case of the character may have more or fewer bytes.  */
3234    
3235        else        else
3236          {          {
3237          unsigned int dc;          pcre_uint32 dc;
3238          GETCHARINC(dc, eptr);          GETCHARINC(dc, eptr);
3239          ecode += length;          ecode += length;
3240    
# Line 3082  for (;;) Line 3251  for (;;)
3251          }          }
3252        }        }
3253      else      else
3254  #endif   /* SUPPORT_UTF8 */  #endif   /* SUPPORT_UTF */
3255    
3256      /* Non-UTF-8 mode */      /* Not UTF mode */
3257        {        {
3258        if (md->end_subject - eptr < 1)        if (TABLE_GET(ecode[1], md->lcc, ecode[1])
3259          {            != TABLE_GET(*eptr, md->lcc, *eptr)) RRETURN(MATCH_NOMATCH);
3260          SCHECK_PARTIAL();            /* This one can use SCHECK_PARTIAL() */        eptr++;
         RRETURN(MATCH_NOMATCH);  
         }  
       if (md->lcc[ecode[1]] != md->lcc[*eptr++]) RRETURN(MATCH_NOMATCH);  
3261        ecode += 2;        ecode += 2;
3262        }        }
3263      break;      break;
# Line 3101  for (;;) Line 3267  for (;;)
3267      case OP_EXACT:      case OP_EXACT:
3268      case OP_EXACTI:      case OP_EXACTI:
3269      min = max = GET2(ecode, 1);      min = max = GET2(ecode, 1);
3270      ecode += 3;      ecode += 1 + IMM2_SIZE;
3271      goto REPEATCHAR;      goto REPEATCHAR;
3272    
3273      case OP_POSUPTO:      case OP_POSUPTO:
# Line 3116  for (;;) Line 3282  for (;;)
3282      min = 0;      min = 0;
3283      max = GET2(ecode, 1);      max = GET2(ecode, 1);
3284      minimize = *ecode == OP_MINUPTO || *ecode == OP_MINUPTOI;      minimize = *ecode == OP_MINUPTO || *ecode == OP_MINUPTOI;
3285      ecode += 3;      ecode += 1 + IMM2_SIZE;
3286      goto REPEATCHAR;      goto REPEATCHAR;
3287    
3288      case OP_POSSTAR:      case OP_POSSTAR:
# Line 3164  for (;;) Line 3330  for (;;)
3330      /* Common code for all repeated single-character matches. */      /* Common code for all repeated single-character matches. */
3331    
3332      REPEATCHAR:      REPEATCHAR:
3333  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
3334      if (utf8)      if (utf)
3335        {        {
3336        length = 1;        length = 1;
3337        charptr = ecode;        charptr = ecode;
# Line 3178  for (;;) Line 3344  for (;;)
3344        if (length > 1)        if (length > 1)
3345          {          {
3346  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
3347          unsigned int othercase;          pcre_uint32 othercase;
3348          if (op >= OP_STARI &&     /* Caseless */          if (op >= OP_STARI &&     /* Caseless */
3349              (othercase = UCD_OTHERCASE(fc)) != fc)              (othercase = UCD_OTHERCASE(fc)) != fc)
3350            oclength = _pcre_ord2utf8(othercase, occhars);            oclength = PRIV(ord2utf)(othercase, occhars);
3351          else oclength = 0;          else oclength = 0;
3352  #endif  /* SUPPORT_UCP */  #endif  /* SUPPORT_UCP */
3353    
3354          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
3355            {            {
3356            if (eptr <= md->end_subject - length &&            if (eptr <= md->end_subject - length &&
3357              memcmp(eptr, charptr, length) == 0) eptr += length;              memcmp(eptr, charptr, IN_UCHARS(length)) == 0) eptr += length;
3358  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
3359            else if (oclength > 0 &&            else if (oclength > 0 &&
3360                     eptr <= md->end_subject - oclength &&                     eptr <= md->end_subject - oclength &&
3361                     memcmp(eptr, occhars, oclength) == 0) eptr += oclength;                     memcmp(eptr, occhars, IN_UCHARS(oclength)) == 0) eptr += oclength;
3362  #endif  /* SUPPORT_UCP */  #endif  /* SUPPORT_UCP */
3363            else            else
3364              {              {
# Line 3211  for (;;) Line 3377  for (;;)
3377              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3378              if (fi >= max) RRETURN(MATCH_NOMATCH);              if (fi >= max) RRETURN(MATCH_NOMATCH);
3379              if (eptr <= md->end_subject - length &&              if (eptr <= md->end_subject - length &&
3380                memcmp(eptr, charptr, length) == 0) eptr += length;                memcmp(eptr, charptr, IN_UCHARS(length)) == 0) eptr += length;
3381  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
3382              else if (oclength > 0 &&              else if (oclength > 0 &&
3383                       eptr <= md->end_subject - oclength &&                       eptr <= md->end_subject - oclength &&
3384                       memcmp(eptr, occhars, oclength) == 0) eptr += oclength;                       memcmp(eptr, occhars, IN_UCHARS(oclength)) == 0) eptr += oclength;
3385  #endif  /* SUPPORT_UCP */  #endif  /* SUPPORT_UCP */
3386              else              else
3387                {                {
# Line 3232  for (;;) Line 3398  for (;;)
3398            for (i = min; i < max; i++)            for (i = min; i < max; i++)
3399              {              {
3400              if (eptr <= md->end_subject - length &&              if (eptr <= md->end_subject - length &&
3401                  memcmp(eptr, charptr, length) == 0) eptr += length;                  memcmp(eptr, charptr, IN_UCHARS(length)) == 0) eptr += length;
3402  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
3403              else if (oclength > 0 &&              else if (oclength > 0 &&
3404                       eptr <= md->end_subject - oclength &&                       eptr <= md->end_subject - oclength &&
3405                       memcmp(eptr, occhars, oclength) == 0) eptr += oclength;                       memcmp(eptr, occhars, IN_UCHARS(oclength)) == 0) eptr += oclength;
3406  #endif  /* SUPPORT_UCP */  #endif  /* SUPPORT_UCP */
3407              else              else
3408                {                {
# Line 3268  for (;;) Line 3434  for (;;)
3434        value of fc will always be < 128. */        value of fc will always be < 128. */
3435        }        }
3436      else      else
3437  #endif  /* SUPPORT_UTF8 */  #endif  /* SUPPORT_UTF */
3438          /* When not in UTF-8 mode, load a single-byte character. */
3439      /* When not in UTF-8 mode, load a single-byte character. */        fc = *ecode++;
3440    
3441      fc = *ecode++;      /* The value of fc at this point is always one character, though we may
3442        or may not be in UTF mode. The code is duplicated for the caseless and
     /* The value of fc at this point is always less than 256, though we may or  
     may not be in UTF-8 mode. The code is duplicated for the caseless and  
3443      caseful cases, for speed, since matching characters is likely to be quite      caseful cases, for speed, since matching characters is likely to be quite
3444      common. First, ensure the minimum number of matches are present. If min =      common. First, ensure the minimum number of matches are present. If min =
3445      max, continue at the same level without recursing. Otherwise, if      max, continue at the same level without recursing. Otherwise, if
# Line 3284  for (;;) Line 3448  for (;;)
3448      maximizing, find the maximum number of characters and work backwards. */      maximizing, find the maximum number of characters and work backwards. */
3449    
3450      DPRINTF(("matching %c{%d,%d} against subject %.*s\n", fc, min, max,      DPRINTF(("matching %c{%d,%d} against subject %.*s\n", fc, min, max,
3451        max, eptr));        max, (char *)eptr));
3452    
3453      if (op >= OP_STARI)  /* Caseless */      if (op >= OP_STARI)  /* Caseless */
3454        {        {
3455        fc = md->lcc[fc];  #ifdef COMPILE_PCRE8
3456          /* fc must be < 128 if UTF is enabled. */
3457          foc = md->fcc[fc];
3458    #else
3459    #ifdef SUPPORT_UTF
3460    #ifdef SUPPORT_UCP
3461          if (utf && fc > 127)
3462            foc = UCD_OTHERCASE(fc);
3463    #else
3464          if (utf && fc > 127)
3465            foc = fc;
3466    #endif /* SUPPORT_UCP */
3467          else
3468    #endif /* SUPPORT_UTF */
3469            foc = TABLE_GET(fc, md->fcc, fc);
3470    #endif /* COMPILE_PCRE8 */
3471    
3472        for (i = 1; i <= min; i++)        for (i = 1; i <= min; i++)
3473          {          {
3474            pcre_uint32 cc;                 /* Faster than pcre_uchar */
3475          if (eptr >= md->end_subject)          if (eptr >= md->end_subject)
3476            {            {
3477            SCHECK_PARTIAL();            SCHECK_PARTIAL();
3478            RRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
3479            }            }
3480          if (fc != md->lcc[*eptr++]) RRETURN(MATCH_NOMATCH);          cc = RAWUCHARTEST(eptr);
3481            if (fc != cc && foc != cc) RRETURN(MATCH_NOMATCH);
3482            eptr++;
3483          }          }
3484        if (min == max) continue;        if (min == max) continue;
3485        if (minimize)        if (minimize)
3486          {          {
3487          for (fi = min;; fi++)          for (fi = min;; fi++)
3488            {            {
3489              pcre_uint32 cc;               /* Faster than pcre_uchar */
3490            RMATCH(eptr, ecode, offset_top, md, eptrb, RM24);            RMATCH(eptr, ecode, offset_top, md, eptrb, RM24);
3491            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3492            if (fi >= max) RRETURN(MATCH_NOMATCH);            if (fi >= max) RRETURN(MATCH_NOMATCH);
# Line 3311  for (;;) Line 3495  for (;;)
3495              SCHECK_PARTIAL();              SCHECK_PARTIAL();
3496              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
3497              }              }
3498            if (fc != md->lcc[*eptr++]) RRETURN(MATCH_NOMATCH);            cc = RAWUCHARTEST(eptr);
3499              if (fc != cc && foc != cc) RRETURN(MATCH_NOMATCH);
3500              eptr++;
3501            }            }
3502          /* Control never gets here */          /* Control never gets here */
3503          }          }
# Line 3320  for (;;) Line 3506  for (;;)
3506          pp = eptr;          pp = eptr;
3507          for (i = min; i < max; i++)          for (i = min; i < max; i++)
3508            {            {
3509              pcre_uint32 cc;               /* Faster than pcre_uchar */
3510            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
3511              {              {
3512              SCHECK_PARTIAL();              SCHECK_PARTIAL();
3513              break;              break;
3514              }              }
3515            if (fc != md->lcc[*eptr]) break;            cc = RAWUCHARTEST(eptr);
3516              if (fc != cc && foc != cc) break;
3517            eptr++;            eptr++;
3518            }            }
3519    
# Line 3353  for (;;) Line 3541  for (;;)
3541            SCHECK_PARTIAL();            SCHECK_PARTIAL();
3542            RRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
3543            }            }
3544          if (fc != *eptr++) RRETURN(MATCH_NOMATCH);          if (fc != RAWUCHARINCTEST(eptr)) RRETURN(MATCH_NOMATCH);
3545          }          }
3546    
3547        if (min == max) continue;        if (min == max) continue;
# Line 3370  for (;;) Line 3558  for (;;)
3558              SCHECK_PARTIAL();              SCHECK_PARTIAL();
3559              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
3560              }              }
3561            if (fc != *eptr++) RRETURN(MATCH_NOMATCH);            if (fc != RAWUCHARINCTEST(eptr)) RRETURN(MATCH_NOMATCH);
3562            }            }
3563          /* Control never gets here */          /* Control never gets here */
3564          }          }
# Line 3384  for (;;) Line 3572  for (;;)
3572              SCHECK_PARTIAL();              SCHECK_PARTIAL();
3573              break;              break;
3574              }              }
3575            if (fc != *eptr) break;            if (fc != RAWUCHARTEST(eptr)) break;
3576            eptr++;            eptr++;
3577            }            }
3578          if (possessive) continue;          if (possessive) continue;
# Line 3410  for (;;) Line 3598  for (;;)
3598        SCHECK_PARTIAL();        SCHECK_PARTIAL();
3599        RRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
3600        }        }
3601      ecode++;  #ifdef SUPPORT_UTF
3602      GETCHARINCTEST(c, eptr);      if (utf)
     if (op == OP_NOTI)         /* The caseless case */  
3603        {        {
3604  #ifdef SUPPORT_UTF8        register pcre_uint32 ch, och;
3605        if (c < 256)  
3606  #endif        ecode++;
3607        c = md->lcc[c];        GETCHARINC(ch, ecode);
3608        if (md->lcc[*ecode++] == c) RRETURN(MATCH_NOMATCH);        GETCHARINC(c, eptr);
3609    
3610          if (op == OP_NOT)
3611            {
3612            if (ch == c) RRETURN(MATCH_NOMATCH);
3613            }
3614          else
3615            {
3616    #ifdef SUPPORT_UCP
3617            if (ch > 127)
3618              och = UCD_OTHERCASE(ch);
3619    #else
3620            if (ch > 127)
3621              och = ch;
3622    #endif /* SUPPORT_UCP */
3623            else
3624              och = TABLE_GET(ch, md->fcc, ch);
3625            if (ch == c || och == c) RRETURN(MATCH_NOMATCH);
3626            }
3627        }        }
3628      else    /* Caseful */      else
3629    #endif
3630        {        {
3631        if (*ecode++ == c) RRETURN(MATCH_NOMATCH);        register pcre_uint32 ch = ecode[1];
3632          c = *eptr++;
3633          if (ch == c || (op == OP_NOTI && TABLE_GET(ch, md->fcc, ch) == c))
3634            RRETURN(MATCH_NOMATCH);
3635          ecode += 2;
3636        }        }
3637      break;      break;
3638    
# Line 3436  for (;;) Line 3646  for (;;)
3646      case OP_NOTEXACT:      case OP_NOTEXACT:
3647      case OP_NOTEXACTI:      case OP_NOTEXACTI:
3648      min = max = GET2(ecode, 1);      min = max = GET2(ecode, 1);
3649      ecode += 3;      ecode += 1 + IMM2_SIZE;
3650      goto REPEATNOTCHAR;      goto REPEATNOTCHAR;
3651    
3652      case OP_NOTUPTO:      case OP_NOTUPTO:
# Line 3446  for (;;) Line 3656  for (;;)
3656      min = 0;      min = 0;
3657      max = GET2(ecode, 1);      max = GET2(ecode, 1);
3658      minimize = *ecode == OP_NOTMINUPTO || *ecode == OP_NOTMINUPTOI;      minimize = *ecode == OP_NOTMINUPTO || *ecode == OP_NOTMINUPTOI;
3659      ecode += 3;      ecode += 1 + IMM2_SIZE;
3660      goto REPEATNOTCHAR;      goto REPEATNOTCHAR;
3661    
3662      case OP_NOTPOSSTAR:      case OP_NOTPOSSTAR:
# Line 3478  for (;;) Line 3688  for (;;)
3688      possessive = TRUE;      possessive = TRUE;
3689      min = 0;      min = 0;
3690      max = GET2(ecode, 1);      max = GET2(ecode, 1);
3691      ecode += 3;      ecode += 1 + IMM2_SIZE;
3692      goto REPEATNOTCHAR;      goto REPEATNOTCHAR;
3693    
3694      case OP_NOTSTAR:      case OP_NOTSTAR:
# Line 3502  for (;;) Line 3712  for (;;)
3712      /* Common code for all repeated single-byte matches. */      /* Common code for all repeated single-byte matches. */
3713    
3714      REPEATNOTCHAR:      REPEATNOTCHAR:
3715      fc = *ecode++;      GETCHARINCTEST(fc, ecode);
3716    
3717      /* The code is duplicated for the caseless and caseful cases, for speed,      /* The code is duplicated for the caseless and caseful cases, for speed,
3718      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 3723  for (;;)
3723      characters and work backwards. */      characters and work backwards. */
3724    
3725      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,
3726        max, eptr));        max, (char *)eptr));
3727    
3728      if (op >= OP_NOTSTARI)     /* Caseless */      if (op >= OP_NOTSTARI)     /* Caseless */
3729        {        {
3730        fc = md->lcc[fc];  #ifdef SUPPORT_UTF
3731    #ifdef SUPPORT_UCP
3732          if (utf && fc > 127)
3733            foc = UCD_OTHERCASE(fc);
3734    #else
3735          if (utf && fc > 127)
3736            foc = fc;
3737    #endif /* SUPPORT_UCP */
3738          else
3739    #endif /* SUPPORT_UTF */
3740            foc = TABLE_GET(fc, md->fcc, fc);
3741    
3742  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
3743        /* UTF-8 mode */        if (utf)
       if (utf8)  
3744          {          {
3745          register unsigned int d;          register pcre_uint32 d;
3746          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
3747            {            {
3748            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
# Line 3532  for (;;) Line 3751  for (;;)
3751              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
3752              }              }
3753            GETCHARINC(d, eptr);            GETCHARINC(d, eptr);
3754            if (d < 256) d = md->lcc[d];            if (fc == d || (unsigned int)foc == d) RRETURN(MATCH_NOMATCH);
           if (fc == d) RRETURN(MATCH_NOMATCH);  
3755            }            }
3756          }          }
3757        else        else
3758  #endif  #endif
3759          /* Not UTF mode */
       /* Not UTF-8 mode */  
3760          {          {
3761          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
3762            {            {
# Line 3548  for (;;) Line 3765  for (;;)
3765              SCHECK_PARTIAL();              SCHECK_PARTIAL();
3766              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
3767              }              }
3768            if (fc == md->lcc[*eptr++]) RRETURN(MATCH_NOMATCH);            if (fc == *eptr || foc == *eptr) RRETURN(MATCH_NOMATCH);
3769              eptr++;
3770            }            }
3771          }          }
3772    
# Line 3556  for (;;) Line 3774  for (;;)
3774    
3775        if (minimize)        if (minimize)
3776          {          {
3777  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
3778          /* UTF-8 mode */          if (utf)
         if (utf8)  
3779            {            {
3780            register unsigned int d;            register pcre_uint32 d;
3781            for (fi = min;; fi++)            for (fi = min;; fi++)
3782              {              {
3783              RMATCH(eptr, ecode, offset_top, md, eptrb, RM28);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM28);
# Line 3572  for (;;) Line 3789  for (;;)
3789                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
3790                }                }
3791              GETCHARINC(d, eptr);              GETCHARINC(d, eptr);
3792              if (d < 256) d = md->lcc[d];              if (fc == d || (unsigned int)foc == d) RRETURN(MATCH_NOMATCH);
             if (fc == d) RRETURN(MATCH_NOMATCH);  
3793              }              }
3794            }            }
3795          else          else
3796  #endif  #endif
3797          /* Not UTF-8 mode */          /* Not UTF mode */
3798            {            {
3799            for (fi = min;; fi++)            for (fi = min;; fi++)
3800              {              {
# Line 3590  for (;;) Line 3806  for (;;)
3806                SCHECK_PARTIAL();                SCHECK_PARTIAL();
3807                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
3808                }                }
3809              if (fc == md->lcc[*eptr++]) RRETURN(MATCH_NOMATCH);              if (fc == *eptr || foc == *eptr) RRETURN(MATCH_NOMATCH);
3810                eptr++;
3811              }              }
3812            }            }
3813          /* Control never gets here */          /* Control never gets here */
# Line 3602  for (;;) Line 3819  for (;;)
3819          {          {
3820          pp = eptr;          pp = eptr;
3821    
3822  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
3823          /* UTF-8 mode */          if (utf)
         if (utf8)  
3824            {            {
3825            register unsigned int d;            register pcre_uint32 d;
3826            for (i = min; i < max; i++)            for (i = min; i < max; i++)
3827              {              {
3828              int len = 1;              int len = 1;
# Line 3616  for (;;) Line 3832  for (;;)
3832                break;                break;
3833                }                }
3834              GETCHARLEN(d, eptr, len);              GETCHARLEN(d, eptr, len);
3835              if (d < 256) d = md->lcc[d];              if (fc == d || (unsigned int)foc == d) break;
             if (fc == d) break;  
3836              eptr += len;              eptr += len;
3837              }              }
3838          if (possessive) continue;            if (possessive) continue;
3839          for(;;)            for(;;)
3840              {              {
3841              RMATCH(eptr, ecode, offset_top, md, eptrb, RM30);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM30);
3842              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
# Line 3631  for (;;) Line 3846  for (;;)
3846            }            }
3847          else          else
3848  #endif  #endif
3849          /* Not UTF-8 mode */          /* Not UTF mode */
3850            {            {
3851            for (i = min; i < max; i++)            for (i = min; i < max; i++)
3852              {              {
# Line 3640  for (;;) Line 3855  for (;;)
3855                SCHECK_PARTIAL();                SCHECK_PARTIAL();
3856                break;                break;
3857                }                }
3858              if (fc == md->lcc[*eptr]) break;              if (fc == *eptr || foc == *eptr) break;
3859              eptr++;              eptr++;
3860              }              }
3861            if (possessive) continue;            if (possessive) continue;
# Line 3661  for (;;) Line 3876  for (;;)
3876    
3877      else      else
3878        {        {
3879  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
3880        /* UTF-8 mode */        if (utf)
       if (utf8)  
3881          {          {
3882          register unsigned int d;          register pcre_uint32 d;
3883          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
3884            {            {
3885            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
# Line 3679  for (;;) Line 3893  for (;;)
3893          }          }
3894        else        else
3895  #endif  #endif
3896        /* Not UTF-8 mode */        /* Not UTF mode */
3897          {          {
3898          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
3899            {            {
# Line 3696  for (;;) Line 3910  for (;;)
3910    
3911        if (minimize)        if (minimize)
3912          {          {
3913  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
3914          /* UTF-8 mode */          if (utf)
         if (utf8)  
3915            {            {
3916            register unsigned int d;            register pcre_uint32 d;
3917            for (fi = min;; fi++)            for (fi = min;; fi++)
3918              {              {
3919              RMATCH(eptr, ecode, offset_top, md, eptrb, RM32);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM32);
# Line 3717  for (;;) Line 3930  for (;;)
3930            }            }
3931          else          else
3932  #endif  #endif
3933          /* Not UTF-8 mode */          /* Not UTF mode */
3934            {            {
3935            for (fi = min;; fi++)            for (fi = min;; fi++)
3936              {              {
# Line 3741  for (;;) Line 3954  for (;;)
3954          {          {
3955          pp = eptr;          pp = eptr;
3956    
3957  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
3958          /* UTF-8 mode */          if (utf)
         if (utf8)  
3959            {            {
3960            register unsigned int d;            register pcre_uint32 d;
3961            for (i = min; i < max; i++)            for (i = min; i < max; i++)
3962              {              {
3963              int len = 1;              int len = 1;
# Line 3769  for (;;) Line 3981  for (;;)
3981            }            }
3982          else          else
3983  #endif  #endif
3984          /* Not UTF-8 mode */          /* Not UTF mode */
3985            {            {
3986            for (i = min; i < max; i++)            for (i = min; i < max; i++)
3987              {              {
# Line 3802  for (;;) Line 4014  for (;;)
4014      case OP_TYPEEXACT:      case OP_TYPEEXACT:
4015      min = max = GET2(ecode, 1);      min = max = GET2(ecode, 1);
4016      minimize = TRUE;      minimize = TRUE;
4017      ecode += 3;      ecode += 1 + IMM2_SIZE;
4018      goto REPEATTYPE;      goto REPEATTYPE;
4019    
4020      case OP_TYPEUPTO:      case OP_TYPEUPTO:
# Line 3810  for (;;) Line 4022  for (;;)
4022      min = 0;      min = 0;
4023      max = GET2(ecode, 1);      max = GET2(ecode, 1);
4024      minimize = *ecode == OP_TYPEMINUPTO;      minimize = *ecode == OP_TYPEMINUPTO;
4025      ecode += 3;      ecode += 1 + IMM2_SIZE;
4026      goto REPEATTYPE;      goto REPEATTYPE;
4027    
4028      case OP_TYPEPOSSTAR:      case OP_TYPEPOSSTAR:
# Line 3838  for (;;) Line 4050  for (;;)
4050      possessive = TRUE;      possessive = TRUE;
4051      min = 0;      min = 0;
4052      max = GET2(ecode, 1);      max = GET2(ecode, 1);
4053      ecode += 3;      ecode += 1 + IMM2_SIZE;
4054      goto REPEATTYPE;      goto REPEATTYPE;
4055    
4056      case OP_TYPESTAR:      case OP_TYPESTAR:
# Line 4021  for (;;) Line 4233  for (;;)
4233              }              }
4234            break;            break;
4235    
4236              case PT_CLIST:
4237              for (i = 1; i <= min; i++)
4238                {
4239                const pcre_uint32 *cp;
4240                if (eptr >= md->end_subject)
4241                  {
4242                  SCHECK_PARTIAL();
4243                  RRETURN(MATCH_NOMATCH);
4244                  }
4245                GETCHARINCTEST(c, eptr);
4246                cp = PRIV(ucd_caseless_sets) + prop_value;
4247                for (;;)
4248                  {
4249                  if (c < *cp)
4250                    { if (prop_fail_result) break; else { RRETURN(MATCH_NOMATCH); } }
4251                  if (c == *cp++)
4252                    { if (prop_fail_result) { RRETURN(MATCH_NOMATCH); } else break; }
4253                  }
4254                }
4255              break;
4256    
4257              case PT_UCNC:
4258              for (i = 1; i <= min; i++)
4259                {
4260                if (eptr >= md->end_subject)
4261                  {
4262                  SCHECK_PARTIAL();
4263                  RRETURN(MATCH_NOMATCH);
4264                  }
4265                GETCHARINCTEST(c, eptr);
4266                if ((c == CHAR_DOLLAR_SIGN || c == CHAR_COMMERCIAL_AT ||
4267                     c == CHAR_GRAVE_ACCENT || (c >= 0xa0 && c <= 0xd7ff) ||
4268                     c >= 0xe000) == prop_fail_result)
4269                  RRETURN(MATCH_NOMATCH);
4270                }
4271              break;
4272    
4273            /* This should not occur */            /* This should not occur */
4274    
4275            default:            default:
# Line 4040  for (;;) Line 4289  for (;;)
4289              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4290              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4291              }              }
4292            GETCHARINCTEST(c, eptr);            else
           if (UCD_CATEGORY(c) == ucp_M) RRETURN(MATCH_NOMATCH);  
           while (eptr < md->end_subject)  
4293              {              {
4294              int len = 1;              int lgb, rgb;
4295              if (!utf8) c = *eptr; else { GETCHARLEN(c, eptr, len); }              GETCHARINCTEST(c, eptr);
4296              if (UCD_CATEGORY(c) != ucp_M) break;              lgb = UCD_GRAPHBREAK(c);
4297              eptr += len;             while (eptr < md->end_subject)
4298                  {
4299                  int len = 1;
4300                  if (!utf) c = *eptr; else { GETCHARLEN(c, eptr, len); }
4301                  rgb = UCD_GRAPHBREAK(c);
4302                  if ((PRIV(ucp_gbtable)[lgb] & (1 << rgb)) == 0) break;
4303                  lgb = rgb;
4304                  eptr += len;
4305                  }
4306              }              }
4307              CHECK_PARTIAL();
4308            }            }
4309          }          }
4310    
# Line 4057  for (;;) Line 4313  for (;;)
4313    
4314  /* Handle all other cases when the coding is UTF-8 */  /* Handle all other cases when the coding is UTF-8 */
4315    
4316  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
4317        if (utf8) switch(ctype)        if (utf) switch(ctype)
4318          {          {
4319          case OP_ANY:          case OP_ANY:
4320          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
# Line 4069  for (;;) Line 4325  for (;;)
4325              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4326              }              }
4327            if (IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH);            if (IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH);
4328              if (md->partial != 0 &&
4329                  eptr + 1 >= md->end_subject &&
4330                  NLBLOCK->nltype == NLTYPE_FIXED &&
4331                  NLBLOCK->nllen == 2 &&
4332                  RAWUCHAR(eptr) == NLBLOCK->nl[0])
4333                {
4334                md->hitend = TRUE;
4335                if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
4336                }
4337            eptr++;            eptr++;
4338            while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;            ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);
4339            }            }
4340          break;          break;
4341    
# Line 4083  for (;;) Line 4348  for (;;)
4348              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4349              }              }
4350            eptr++;            eptr++;
4351            while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;            ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);
4352            }            }
4353          break;          break;
4354    
# Line 4105  for (;;) Line 4370  for (;;)
4370              {              {
4371              default: RRETURN(MATCH_NOMATCH);              default: RRETURN(MATCH_NOMATCH);
4372    
4373              case 0x000d:              case CHAR_CR:
4374              if (eptr < md->end_subject && *eptr == 0x0a) eptr++;              if (eptr < md->end_subject && RAWUCHAR(eptr) == CHAR_LF) eptr++;
4375              break;              break;
4376    
4377              case 0x000a:              case CHAR_LF:
4378              break;              break;
4379    
4380              case 0x000b:              case CHAR_VT:
4381              case 0x000c:              case CHAR_FF:
4382              case 0x0085:              case CHAR_NEL:
4383    #ifndef EBCDIC
4384              case 0x2028:              case 0x2028:
4385              case 0x2029:              case 0x2029:
4386    #endif  /* Not EBCDIC */
4387              if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);              if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);
4388              break;              break;
4389              }              }
# Line 4134  for (;;) Line 4401  for (;;)
4401            GETCHARINC(c, eptr);            GETCHARINC(c, eptr);
4402            switch(c)            switch(c)
4403              {              {
4404                HSPACE_CASES: RRETURN(MATCH_NOMATCH);  /* Byte and multibyte cases */
4405              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);  
4406              }              }
4407            }            }
4408          break;          break;
# Line 4170  for (;;) Line 4418  for (;;)
4418            GETCHARINC(c, eptr);            GETCHARINC(c, eptr);
4419            switch(c)            switch(c)
4420              {              {
4421                HSPACE_CASES: break;  /* Byte and multibyte cases */
4422              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;  
4423              }              }
4424            }            }
4425          break;          break;
# Line 4206  for (;;) Line 4435  for (;;)
4435            GETCHARINC(c, eptr);            GETCHARINC(c, eptr);
4436            switch(c)            switch(c)
4437              {              {
4438                VSPACE_CASES: RRETURN(MATCH_NOMATCH);
4439              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);  
4440              }              }
4441            }            }
4442          break;          break;
# Line 4230  for (;;) Line 4452  for (;;)
4452            GETCHARINC(c, eptr);            GETCHARINC(c, eptr);
4453            switch(c)            switch(c)
4454              {              {
4455                VSPACE_CASES: break;
4456              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;  
4457              }              }
4458            }            }
4459          break;          break;
# Line 4260  for (;;) Line 4475  for (;;)
4475          case OP_DIGIT:          case OP_DIGIT:
4476          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
4477            {            {
4478              pcre_uint32 cc;
4479            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
4480              {              {
4481              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4482              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4483              }              }
4484            if (*eptr >= 128 || (md->ctypes[*eptr++] & ctype_digit) == 0)            cc = RAWUCHAR(eptr);
4485              if (cc >= 128 || (md->ctypes[cc] & ctype_digit) == 0)
4486              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4487              eptr++;
4488            /* 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 */
4489            }            }
4490          break;          break;
# Line 4274  for (;;) Line 4492  for (;;)
4492          case OP_NOT_WHITESPACE:          case OP_NOT_WHITESPACE:
4493          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
4494            {            {
4495              pcre_uint32 cc;
4496            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
4497              {              {
4498              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4499              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4500              }              }
4501            if (*eptr < 128 && (md->ctypes[*eptr] & ctype_space) != 0)            cc = RAWUCHAR(eptr);
4502              if (cc < 128 && (md->ctypes[cc] & ctype_space) != 0)
4503              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4504            while (++eptr < md->end_subject && (*eptr & 0xc0) == 0x80);            eptr++;
4505              ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);
4506            }            }
4507          break;          break;
4508    
4509          case OP_WHITESPACE:          case OP_WHITESPACE:
4510          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
4511            {            {
4512              pcre_uint32 cc;
4513            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
4514              {              {
4515              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4516              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4517              }              }
4518            if (*eptr >= 128 || (md->ctypes[*eptr++] & ctype_space) == 0)            cc = RAWUCHAR(eptr);
4519              if (cc >= 128 || (md->ctypes[cc] & ctype_space) == 0)
4520              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4521              eptr++;
4522            /* 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 */
4523            }            }
4524          break;          break;
# Line 4302  for (;;) Line 4526  for (;;)
4526          case OP_NOT_WORDCHAR:          case OP_NOT_WORDCHAR:
4527          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
4528            {            {
4529              pcre_uint32 cc;
4530            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
4531              {              {
4532              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4533              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4534              }              }
4535            if (*eptr < 128 && (md->ctypes[*eptr] & ctype_word) != 0)            cc = RAWUCHAR(eptr);
4536              if (cc < 128 && (md->ctypes[cc] & ctype_word) != 0)
4537              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4538            while (++eptr < md->end_subject && (*eptr & 0xc0) == 0x80);            eptr++;
4539              ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);
4540            }            }
4541          break;          break;
4542    
4543          case OP_WORDCHAR:          case OP_WORDCHAR:
4544          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
4545            {            {
4546              pcre_uint32 cc;
4547            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
4548              {              {
4549              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4550              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4551              }              }
4552            if (*eptr >= 128 || (md->ctypes[*eptr++] & ctype_word) == 0)            cc = RAWUCHAR(eptr);
4553              if (cc >= 128 || (md->ctypes[cc] & ctype_word) == 0)
4554              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4555              eptr++;
4556            /* 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 */
4557            }            }
4558          break;          break;
# Line 4332  for (;;) Line 4562  for (;;)
4562          }  /* End switch(ctype) */          }  /* End switch(ctype) */
4563    
4564        else        else
4565  #endif     /* SUPPORT_UTF8 */  #endif     /* SUPPORT_UTF */
4566    
4567        /* 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
4568        than OP_PROP and OP_NOTPROP. */        than OP_PROP and OP_NOTPROP. */
# Line 4348  for (;;) Line 4578  for (;;)
4578              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4579              }              }
4580            if (IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH);            if (IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH);
4581              if (md->partial != 0 &&
4582                  eptr + 1 >= md->end_subject &&
4583                  NLBLOCK->nltype == NLTYPE_FIXED &&
4584                  NLBLOCK->nllen == 2 &&
4585                  *eptr == NLBLOCK->nl[0])
4586                {
4587                md->hitend = TRUE;
4588                if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
4589                }
4590            eptr++;            eptr++;
4591            }            }
4592          break;          break;
# Line 4382  for (;;) Line 4621  for (;;)
4621              {              {
4622              default: RRETURN(MATCH_NOMATCH);              default: RRETURN(MATCH_NOMATCH);
4623    
4624              case 0x000d:              case CHAR_CR:
4625              if (eptr < md->end_subject && *eptr == 0x0a) eptr++;              if (eptr < md->end_subject && *eptr == CHAR_LF) eptr++;
4626              break;              break;
4627    
4628              case 0x000a:              case CHAR_LF:
4629              break;              break;
4630    
4631              case 0x000b:              case CHAR_VT:
4632              case 0x000c:              case CHAR_FF:
4633              case 0x0085:              case CHAR_NEL:
4634    #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
4635                case 0x2028:
4636                case 0x2029:
4637    #endif
4638              if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);              if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);
4639              break;              break;
4640              }              }
# Line 4409  for (;;) Line 4652  for (;;)
4652            switch(*eptr++)            switch(*eptr++)
4653              {              {
4654              default: break;              default: break;
4655              case 0x09:      /* HT */              HSPACE_BYTE_CASES:
4656              case 0x20:      /* SPACE */  #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
4657              case 0xa0:      /* NBSP */              HSPACE_MULTIBYTE_CASES:
4658    #endif
4659              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4660              }              }
4661            }            }
# Line 4428  for (;;) Line 4672  for (;;)
4672            switch(*eptr++)            switch(*eptr++)
4673              {              {
4674              default: RRETURN(MATCH_NOMATCH);              default: RRETURN(MATCH_NOMATCH);
4675              case 0x09:      /* HT */              HSPACE_BYTE_CASES:
4676              case 0x20:      /* SPACE */  #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
4677              case 0xa0:      /* NBSP */              HSPACE_MULTIBYTE_CASES:
4678    #endif
4679              break;              break;
4680              }              }
4681            }            }
# Line 4446  for (;;) Line 4691  for (;;)
4691              }              }
4692            switch(*eptr++)            switch(*eptr++)
4693              {              {
4694              default: break;              VSPACE_BYTE_CASES:
4695              case 0x0a:      /* LF */  #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
4696              case 0x0b:      /* VT */              VSPACE_MULTIBYTE_CASES:
4697              case 0x0c:      /* FF */  #endif
             case 0x0d:      /* CR */  
             case 0x85:      /* NEL */  
4698              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4699                default: break;
4700              }              }
4701            }            }
4702          break;          break;
# Line 4468  for (;;) Line 4712  for (;;)
4712            switch(*eptr++)            switch(*eptr++)
4713              {              {
4714              default: RRETURN(MATCH_NOMATCH);              default: RRETURN(MATCH_NOMATCH);
4715              case 0x0a:      /* LF */              VSPACE_BYTE_CASES:
4716              case 0x0b:      /* VT */  #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
4717              case 0x0c:      /* FF */              VSPACE_MULTIBYTE_CASES:
4718              case 0x0d:      /* CR */  #endif
             case 0x85:      /* NEL */  
4719              break;              break;
4720              }              }
4721            }            }
# Line 4486  for (;;) Line 4729  for (;;)
4729              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4730              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4731              }              }
4732            if ((md->ctypes[*eptr++] & ctype_digit) != 0) RRETURN(MATCH_NOMATCH);            if (MAX_255(*eptr) && (md->ctypes[*eptr] & ctype_digit) != 0)
4733                RRETURN(MATCH_NOMATCH);
4734              eptr++;
4735            }            }
4736          break;          break;
4737    
# Line 4498  for (;;) Line 4743  for (;;)
4743              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4744              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4745              }              }
4746            if ((md->ctypes[*eptr++] & ctype_digit) == 0) RRETURN(MATCH_NOMATCH);            if (!MAX_255(*eptr) || (md->ctypes[*eptr] & ctype_digit) == 0)
4747                RRETURN(MATCH_NOMATCH);
4748              eptr++;
4749            }            }
4750          break;          break;
4751    
# Line 4510  for (;;) Line 4757  for (;;)
4757              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4758              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4759              }              }
4760            if ((md->ctypes[*eptr++] & ctype_space) != 0) RRETURN(MATCH_NOMATCH);            if (MAX_255(*eptr) && (md->ctypes[*eptr] & ctype_space) != 0)
4761                RRETURN(MATCH_NOMATCH);
4762              eptr++;
4763            }            }
4764          break;          break;
4765    
# Line 4522  for (;;) Line 4771  for (;;)
4771              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4772              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4773              }              }
4774            if ((md->ctypes[*eptr++] & ctype_space) == 0) RRETURN(MATCH_NOMATCH);            if (!MAX_255(*eptr) || (md->ctypes[*eptr] & ctype_space) == 0)
4775                RRETURN(MATCH_NOMATCH);
4776              eptr++;
4777            }            }
4778          break;          break;
4779    
# Line 4534  for (;;) Line 4785  for (;;)
4785              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4786              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4787              }              }
4788            if ((md->ctypes[*eptr++] & ctype_word) != 0)            if (MAX_255(*eptr) && (md->ctypes[*eptr] & ctype_word) != 0)
4789              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4790              eptr++;
4791            }            }
4792          break;          break;
4793    
# Line 4547  for (;;) Line 4799  for (;;)
4799              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4800              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4801              }              }
4802            if ((md->ctypes[*eptr++] & ctype_word) == 0)            if (!MAX_255(*eptr) || (md->ctypes[*eptr] & ctype_word) == 0)
4803              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4804              eptr++;
4805            }            }
4806          break;          break;
4807    
# Line 4739  for (;;) Line 4992  for (;;)
4992              }              }
4993            /* Control never gets here */            /* Control never gets here */
4994    
4995            /* This should never occur */            case PT_CLIST:
4996              for (fi = min;; fi++)
4997                {
4998                const pcre_uint32 *cp;
4999                RMATCH(eptr, ecode, offset_top, md, eptrb, RM67);
5000                if (rrc != MATCH_NOMATCH) RRETURN(rrc);
5001                if (fi >= max) RRETURN(MATCH_NOMATCH);
5002                if (eptr >= md->end_subject)
5003                  {
5004                  SCHECK_PARTIAL();
5005                  RRETURN(MATCH_NOMATCH);
5006                  }
5007                GETCHARINCTEST(c, eptr);
5008                cp = PRIV(ucd_caseless_sets) + prop_value;
5009                for (;;)
5010                  {
5011                  if (c < *cp)
5012                    { if (prop_fail_result) break; else { RRETURN(MATCH_NOMATCH); } }
5013                  if (c == *cp++)
5014                    { if (prop_fail_result) { RRETURN(MATCH_NOMATCH); } else break; }
5015                  }
5016                }
5017              /* Control never gets here */
5018    
5019              case PT_UCNC:
5020              for (fi = min;; fi++)
5021                {
5022                RMATCH(eptr, ecode, offset_top, md, eptrb, RM68);
5023                if (rrc != MATCH_NOMATCH) RRETURN(rrc);
5024                if (fi >= max) RRETURN(MATCH_NOMATCH);
5025                if (eptr >= md->end_subject)
5026                  {
5027                  SCHECK_PARTIAL();
5028                  RRETURN(MATCH_NOMATCH);
5029                  }
5030                GETCHARINCTEST(c, eptr);
5031                if ((c == CHAR_DOLLAR_SIGN || c == CHAR_COMMERCIAL_AT ||
5032                     c == CHAR_GRAVE_ACCENT || (c >= 0xa0 && c <= 0xd7ff) ||
5033                     c >= 0xe000) == prop_fail_result)
5034                  RRETURN(MATCH_NOMATCH);
5035                }
5036              /* Control never gets here */
5037    
5038              /* This should never occur */
5039            default:            default:
5040            RRETURN(PCRE_ERROR_INTERNAL);            RRETURN(PCRE_ERROR_INTERNAL);
5041            }            }
# Line 4761  for (;;) Line 5056  for (;;)
5056              SCHECK_PARTIAL();              SCHECK_PARTIAL();
5057              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
5058              }              }
5059            GETCHARINCTEST(c, eptr);            else
           if (UCD_CATEGORY(c) == ucp_M) RRETURN(MATCH_NOMATCH);  
           while (eptr < md->end_subject)  
5060              {              {
5061              int len = 1;              int lgb, rgb;
5062              if (!utf8) c = *eptr; else { GETCHARLEN(c, eptr, len); }              GETCHARINCTEST(c, eptr);
5063              if (UCD_CATEGORY(c) != ucp_M) break;              lgb = UCD_GRAPHBREAK(c);
5064              eptr += len;              while (eptr < md->end_subject)
5065                  {
5066                  int len = 1;
5067                  if (!utf) c = *eptr; else { GETCHARLEN(c, eptr, len); }
5068                  rgb = UCD_GRAPHBREAK(c);
5069                  if ((PRIV(ucp_gbtable)[lgb] & (1 << rgb)) == 0) break;
5070                  lgb = rgb;
5071                  eptr += len;
5072                  }
5073              }              }
5074              CHECK_PARTIAL();
5075            }            }
5076          }          }
5077        else        else
5078  #endif     /* SUPPORT_UCP */  #endif     /* SUPPORT_UCP */
5079    
5080  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
5081        /* UTF-8 mode */        if (utf)
       if (utf8)  
5082          {          {
5083          for (fi = min;; fi++)          for (fi = min;; fi++)
5084            {            {
# Line 4794  for (;;) Line 5095  for (;;)
5095            GETCHARINC(c, eptr);            GETCHARINC(c, eptr);
5096            switch(ctype)            switch(ctype)
5097              {              {
5098              case OP_ANY:        /* This is the non-NL case */              case OP_ANY:               /* This is the non-NL case */
5099                if (md->partial != 0 &&    /* Take care with CRLF partial */
5100                    eptr >= md->end_subject &&
5101                    NLBLOCK->nltype == NLTYPE_FIXED &&
5102                    NLBLOCK->nllen == 2 &&
5103                    c == NLBLOCK->nl[0])
5104                  {
5105                  md->hitend = TRUE;
5106                  if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
5107                  }
5108                break;
5109    
5110              case OP_ALLANY:              case OP_ALLANY:
5111              case OP_ANYBYTE:              case OP_ANYBYTE:
5112              break;              break;
# Line 4803  for (;;) Line 5115  for (;;)
5115              switch(c)              switch(c)
5116                {                {
5117                default: RRETURN(MATCH_NOMATCH);                default: RRETURN(MATCH_NOMATCH);
5118                case 0x000d:                case CHAR_CR:
5119                if (eptr < md->end_subject && *eptr == 0x0a) eptr++;                if (eptr < md->end_subject && RAWUCHAR(eptr) == CHAR_LF) eptr++;
5120                break;                break;
5121                case 0x000a:  
5122                  case CHAR_LF:
5123                break;                break;
5124    
5125                case 0x000b:                case CHAR_VT:
5126                case 0x000c:                case CHAR_FF:
5127                case 0x0085:                case CHAR_NEL:
5128    #ifndef EBCDIC
5129                case 0x2028:                case 0x2028:
5130                case 0x2029:                case 0x2029:
5131    #endif  /* Not EBCDIC */
5132                if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);                if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);
5133                break;                break;
5134                }                }
# Line 4822  for (;;) Line 5137  for (;;)
5137              case OP_NOT_HSPACE:              case OP_NOT_HSPACE:
5138              switch(c)              switch(c)
5139                {                {
5140                  HSPACE_CASES: RRETURN(MATCH_NOMATCH);
5141                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);  
5142                }                }
5143              break;              break;
5144    
5145              case OP_HSPACE:              case OP_HSPACE:
5146              switch(c)              switch(c)
5147                {                {
5148                  HSPACE_CASES: break;
5149                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;  
5150                }                }
5151              break;              break;
5152    
5153              case OP_NOT_VSPACE:              case OP_NOT_VSPACE:
5154              switch(c)              switch(c)
5155                {                {
5156                  VSPACE_CASES: RRETURN(MATCH_NOMATCH);
5157                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);  
5158                }                }
5159              break;              break;
5160    
5161              case OP_VSPACE:              case OP_VSPACE:
5162              switch(c)              switch(c)
5163                {                {
5164                  VSPACE_CASES: break;
5165                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;  
5166                }                }
5167              break;              break;
5168    
# Line 4919  for (;;) Line 5182  for (;;)
5182              break;              break;
5183    
5184              case OP_WHITESPACE:              case OP_WHITESPACE:
5185              if  (c >= 256 || (md->ctypes[c] & ctype_space) == 0)              if (c >= 256 || (md->ctypes[c] & ctype_space) == 0)
5186                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
5187              break;              break;
5188    
# Line 4940  for (;;) Line 5203  for (;;)
5203          }          }
5204        else        else
5205  #endif  #endif
5206        /* Not UTF-8 mode */        /* Not UTF mode */
5207          {          {
5208          for (fi = min;; fi++)          for (fi = min;; fi++)
5209            {            {
# Line 4957  for (;;) Line 5220  for (;;)
5220            c = *eptr++;            c = *eptr++;
5221            switch(ctype)            switch(ctype)
5222              {              {
5223              case OP_ANY:     /* This is the non-NL case */              case OP_ANY:               /* This is the non-NL case */
5224                if (md->partial != 0 &&    /* Take care with CRLF partial */
5225                    eptr >= md->end_subject &&
5226                    NLBLOCK->nltype == NLTYPE_FIXED &&
5227                    NLBLOCK->nllen == 2 &&
5228                    c == NLBLOCK->nl[0])
5229                  {
5230                  md->hitend = TRUE;
5231                  if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
5232                  }
5233                break;
5234    
5235              case OP_ALLANY:              case OP_ALLANY:
5236              case OP_ANYBYTE:              case OP_ANYBYTE:
5237              break;              break;
# Line 4966  for (;;) Line 5240  for (;;)
5240              switch(c)              switch(c)
5241                {                {
5242                default: RRETURN(MATCH_NOMATCH);                default: RRETURN(MATCH_NOMATCH);
5243                case 0x000d:                case CHAR_CR:
5244                if (eptr < md->end_subject && *eptr == 0x0a) eptr++;                if (eptr < md->end_subject && *eptr == CHAR_LF) eptr++;
5245                break;                break;
5246    
5247                case 0x000a:                case CHAR_LF:
5248                break;                break;
5249    
5250                case 0x000b:                case CHAR_VT:
5251                case 0x000c:                case CHAR_FF:
5252                case 0x0085:                case CHAR_NEL:
5253    #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
5254                  case 0x2028:
5255                  case 0x2029:
5256    #endif
5257                if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);                if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);
5258                break;                break;
5259                }                }
# Line 4985  for (;;) Line 5263  for (;;)
5263              switch(c)              switch(c)
5264                {                {
5265                default: break;                default: break;
5266                case 0x09:      /* HT */                HSPACE_BYTE_CASES:
5267                case 0x20:      /* SPACE */  #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
5268                case 0xa0:      /* NBSP */                HSPACE_MULTIBYTE_CASES:
5269    #endif
5270                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
5271                }                }
5272              break;              break;
# Line 4996  for (;;) Line 5275  for (;;)
5275              switch(c)              switch(c)
5276                {                {
5277                default: RRETURN(MATCH_NOMATCH);                default: RRETURN(MATCH_NOMATCH);
5278                case 0x09:      /* HT */                HSPACE_BYTE_CASES:
5279                case 0x20:      /* SPACE */  #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
5280                case 0xa0:      /* NBSP */                HSPACE_MULTIBYTE_CASES:
5281    #endif
5282                break;                break;
5283                }                }
5284              break;              break;
# Line 5007  for (;;) Line 5287  for (;;)
5287              switch(c)              switch(c)
5288                {                {
5289                default: break;                default: break;
5290                case 0x0a:      /* LF */                VSPACE_BYTE_CASES:
5291                case 0x0b:      /* VT */  #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
5292                case 0x0c:      /* FF */                VSPACE_MULTIBYTE_CASES:
5293                case 0x0d:      /* CR */  #endif
               case 0x85:      /* NEL */  
5294                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
5295                }                }
5296              break;              break;
# Line 5020  for (;;) Line 5299  for (;;)
5299              switch(c)              switch(c)
5300                {                {
5301                default: RRETURN(MATCH_NOMATCH);                default: RRETURN(MATCH_NOMATCH);
5302                case 0x0a:      /* LF */                VSPACE_BYTE_CASES:
5303                case 0x0b:      /* VT */  #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
5304                case 0x0c:      /* FF */                VSPACE_MULTIBYTE_CASES:
5305                case 0x0d:      /* CR */  #endif
               case 0x85:      /* NEL */  
5306                break;                break;
5307                }                }
5308              break;              break;
5309    
5310              case OP_NOT_DIGIT:              case OP_NOT_DIGIT:
5311              if ((md->ctypes[c] & ctype_digit) != 0) RRETURN(MATCH_NOMATCH);              if (MAX_255(c) && (md->ctypes[c] & ctype_digit) != 0) RRETURN(MATCH_NOMATCH);
5312              break;              break;
5313    
5314              case OP_DIGIT:              case OP_DIGIT:
5315              if ((md->ctypes[c] & ctype_digit) == 0) RRETURN(MATCH_NOMATCH);              if (!MAX_255(c) || (md->ctypes[c] & ctype_digit) == 0) RRETURN(MATCH_NOMATCH);
5316              break;              break;
5317    
5318              case OP_NOT_WHITESPACE:              case OP_NOT_WHITESPACE:
5319              if ((md->ctypes[c] & ctype_space) != 0) RRETURN(MATCH_NOMATCH);              if (MAX_255(c) && (md->ctypes[c] & ctype_space) != 0) RRETURN(MATCH_NOMATCH);
5320              break;              break;
5321    
5322              case OP_WHITESPACE:              case OP_WHITESPACE:
5323              if  ((md->ctypes[c] & ctype_space) == 0) RRETURN(MATCH_NOMATCH);              if (!MAX_255(c) || (md->ctypes[c] & ctype_space) == 0) RRETURN(MATCH_NOMATCH);
5324              break;              break;
5325    
5326              case OP_NOT_WORDCHAR:              case OP_NOT_WORDCHAR:
5327              if ((md->ctypes[c] & ctype_word) != 0) RRETURN(MATCH_NOMATCH);              if (MAX_255(c) && (md->ctypes[c] & ctype_word) != 0) RRETURN(MATCH_NOMATCH);
5328              break;              break;
5329    
5330              case OP_WORDCHAR:              case OP_WORDCHAR:
5331              if ((md->ctypes[c] & ctype_word) == 0) RRETURN(MATCH_NOMATCH);              if (!MAX_255(c) || (md->ctypes[c] & ctype_word) == 0) RRETURN(MATCH_NOMATCH);
5332              break;              break;
5333    
5334              default:              default:
# Line 5227  for (;;) Line 5505  for (;;)
5505              }              }
5506            break;            break;
5507    
5508              case PT_CLIST:
5509              for (i = min; i < max; i++)
5510                {
5511                const pcre_uint32 *cp;
5512                int len = 1;
5513                if (eptr >= md->end_subject)
5514                  {
5515                  SCHECK_PARTIAL();
5516                  break;
5517                  }
5518                GETCHARLENTEST(c, eptr, len);
5519                cp = PRIV(ucd_caseless_sets) + prop_value;
5520                for (;;)
5521                  {
5522                  if (c < *cp)
5523                    { if (prop_fail_result) break; else goto GOT_MAX; }
5524                  if (c == *cp++)
5525                    { if (prop_fail_result) goto GOT_MAX; else break; }
5526                  }
5527                eptr += len;
5528                }
5529              GOT_MAX:
5530              break;
5531    
5532              case PT_UCNC:
5533              for (i = min; i < max; i++)
5534                {
5535                int len = 1;
5536                if (eptr >= md->end_subject)
5537                  {
5538                  SCHECK_PARTIAL();
5539                  break;
5540                  }
5541                GETCHARLENTEST(c, eptr, len);
5542                if ((c == CHAR_DOLLAR_SIGN || c == CHAR_COMMERCIAL_AT ||
5543                     c == CHAR_GRAVE_ACCENT || (c >= 0xa0 && c <= 0xd7ff) ||
5544                     c >= 0xe000) == prop_fail_result)
5545                  break;
5546                eptr += len;
5547                }
5548              break;
5549    
5550            default:            default:
5551            RRETURN(PCRE_ERROR_INTERNAL);            RRETURN(PCRE_ERROR_INTERNAL);
5552            }            }
# Line 5239  for (;;) Line 5559  for (;;)
5559            RMATCH(eptr, ecode, offset_top, md, eptrb, RM44);            RMATCH(eptr, ecode, offset_top, md, eptrb, RM44);
5560            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
5561            if (eptr-- == pp) break;        /* Stop if tried at original pos */            if (eptr-- == pp) break;        /* Stop if tried at original pos */
5562            if (utf8) BACKCHAR(eptr);            if (utf) BACKCHAR(eptr);
5563            }            }
5564          }          }
5565    
# Line 5250  for (;;) Line 5570  for (;;)
5570          {          {
5571          for (i = min; i < max; i++)          for (i = min; i < max; i++)
5572            {            {
           int len = 1;  
5573            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
5574              {              {
5575              SCHECK_PARTIAL();              SCHECK_PARTIAL();
5576              break;              break;
5577              }              }
5578            if (!utf8) c = *eptr; else { GETCHARLEN(c, eptr, len); }            else
           if (UCD_CATEGORY(c) == ucp_M) break;  
           eptr += len;  
           while (eptr < md->end_subject)  
5579              {              {
5580              len = 1;              int lgb, rgb;
5581              if (!utf8) c = *eptr; else { GETCHARLEN(c, eptr, len); }              GETCHARINCTEST(c, eptr);
5582              if (UCD_CATEGORY(c) != ucp_M) break;              lgb = UCD_GRAPHBREAK(c);
5583              eptr += len;              while (eptr < md->end_subject)
5584                  {
5585                  int len = 1;
5586                  if (!utf) c = *eptr; else { GETCHARLEN(c, eptr, len); }
5587                  rgb = UCD_GRAPHBREAK(c);
5588                  if ((PRIV(ucp_gbtable)[lgb] & (1 << rgb)) == 0) break;
5589                  lgb = rgb;
5590                  eptr += len;
5591                  }
5592              }              }
5593              CHECK_PARTIAL();
5594            }            }
5595    
5596          /* eptr is now past the end of the maximum run */          /* eptr is now past the end of the maximum run */
# Line 5279  for (;;) Line 5604  for (;;)
5604            if (eptr-- == pp) break;        /* Stop if tried at original pos */            if (eptr-- == pp) break;        /* Stop if tried at original pos */
5605            for (;;)                        /* Move back over one extended */            for (;;)                        /* Move back over one extended */
5606              {              {
5607              if (!utf8) c = *eptr; else              if (!utf) c = *eptr; else
5608                {                {
5609                BACKCHAR(eptr);                BACKCHAR(eptr);
5610                GETCHAR(c, eptr);                GETCHAR(c, eptr);
# Line 5293  for (;;) Line 5618  for (;;)
5618        else        else
5619  #endif   /* SUPPORT_UCP */  #endif   /* SUPPORT_UCP */
5620    
5621  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
5622        /* UTF-8 mode */        if (utf)
   
       if (utf8)  
5623          {          {
5624          switch(ctype)          switch(ctype)
5625            {            {
# Line 5311  for (;;) Line 5634  for (;;)
5634                  break;                  break;
5635                  }                  }
5636                if (IS_NEWLINE(eptr)) break;                if (IS_NEWLINE(eptr)) break;
5637                  if (md->partial != 0 &&    /* Take care with CRLF partial */
5638                      eptr + 1 >= md->end_subject &&
5639                      NLBLOCK->nltype == NLTYPE_FIXED &&
5640                      NLBLOCK->nllen == 2 &&
5641                      RAWUCHAR(eptr) == NLBLOCK->nl[0])
5642                    {
5643                    md->hitend = TRUE;
5644                    if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
5645                    }
5646                eptr++;                eptr++;
5647                while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;                ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);
5648                }                }
5649              }              }
5650    
# Line 5328  for (;;) Line 5660  for (;;)
5660                  break;                  break;
5661                  }                  }
5662                if (IS_NEWLINE(eptr)) break;                if (IS_NEWLINE(eptr)) break;
5663                  if (md->partial != 0 &&    /* Take care with CRLF partial */
5664                      eptr + 1 >= md->end_subject &&
5665                      NLBLOCK->nltype == NLTYPE_FIXED &&
5666                      NLBLOCK->nllen == 2 &&
5667                      RAWUCHAR(eptr) == NLBLOCK->nl[0])
5668                    {
5669                    md->hitend = TRUE;
5670                    if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
5671                    }
5672                eptr++;                eptr++;
5673                while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;                ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);
5674                }                }
5675              }              }
5676            break;            break;
# Line 5345  for (;;) Line 5686  for (;;)
5686                  break;                  break;
5687                  }                  }
5688                eptr++;                eptr++;
5689                while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;                ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);
5690                }                }
5691              }              }
5692            else            else
# Line 5377  for (;;) Line 5718  for (;;)
5718                break;                break;
5719                }                }
5720              GETCHARLEN(c, eptr, len);              GETCHARLEN(c, eptr, len);
5721              if (c == 0x000d)              if (c == CHAR_CR)
5722                {                {
5723                if (++eptr >= md->end_subject) break;                if (++eptr >= md->end_subject) break;
5724                if (*eptr == 0x000a) eptr++;                if (RAWUCHAR(eptr) == CHAR_LF) eptr++;
5725                }                }
5726              else              else
5727                {                {
5728                if (c != 0x000a &&                if (c != CHAR_LF &&
5729                    (md->bsr_anycrlf ||                    (md->bsr_anycrlf ||
5730                     (c != 0x000b && c != 0x000c &&                     (c != CHAR_VT && c != CHAR_FF && c != CHAR_NEL
5731                      c != 0x0085 && c != 0x2028 && c != 0x2029)))  #ifndef EBCDIC
5732                        && c != 0x2028 && c != 0x2029
5733    #endif  /* Not EBCDIC */
5734                        )))
5735                  break;                  break;
5736                eptr += len;                eptr += len;
5737                }                }
# Line 5408  for (;;) Line 5752  for (;;)
5752              GETCHARLEN(c, eptr, len);              GETCHARLEN(c, eptr, len);
5753              switch(c)              switch(c)
5754                {                {
5755                  HSPACE_CASES: gotspace = TRUE; break;
5756                default: gotspace = FALSE; break;                default: gotspace = FALSE; break;
               case 0x09:      /* HT */  
               case 0x20:      /* SPACE */  
               case 0xa0:      /* NBSP */  
               case 0x1680:    /* OGHAM SPACE MARK */  
               case 0x180e:    /* MONGOLIAN VOWEL SEPARATOR */  
               case 0x2000:    /* EN QUAD */  
               case 0x2001:    /* EM QUAD */  
               case 0x2002:    /* EN SPACE */  
               case 0x2003:    /* EM SPACE */  
               case 0x2004:    /* THREE-PER-EM SPACE */  
               case 0x2005:    /* FOUR-PER-EM SPACE */  
               case 0x2006:    /* SIX-PER-EM SPACE */  
               case 0x2007:    /* FIGURE SPACE */  
               case 0x2008:    /* PUNCTUATION SPACE */  
               case 0x2009:    /* THIN SPACE */  
               case 0x200A:    /* HAIR SPACE */  
               case 0x202f:    /* NARROW NO-BREAK SPACE */  
               case 0x205f:    /* MEDIUM MATHEMATICAL SPACE */  
               case 0x3000:    /* IDEOGRAPHIC SPACE */  
               gotspace = TRUE;  
               break;  
5757                }                }
5758              if (gotspace == (ctype == OP_NOT_HSPACE)) break;              if (gotspace == (ctype == OP_NOT_HSPACE)) break;
5759              eptr += len;              eptr += len;
# Line 5450  for (;;) Line 5774  for (;;)
5774              GETCHARLEN(c, eptr, len);              GETCHARLEN(c, eptr, len);
5775              switch(c)              switch(c)
5776                {                {
5777                  VSPACE_CASES: gotspace = TRUE; break;
5778                default: gotspace = FALSE; break;                default: gotspace = FALSE; break;
               case 0x0a:      /* LF */  
               case 0x0b:      /* VT */  
               case 0x0c:      /* FF */  
               case 0x0d:      /* CR */  
               case 0x85:      /* NEL */  
               case 0x2028:    /* LINE SEPARATOR */  
               case 0x2029:    /* PARAGRAPH SEPARATOR */  
               gotspace = TRUE;  
               break;  
5779                }                }
5780              if (gotspace == (ctype == OP_NOT_VSPACE)) break;              if (gotspace == (ctype == OP_NOT_VSPACE)) break;
5781              eptr += len;              eptr += len;
# Line 5573  for (;;) Line 5889  for (;;)
5889            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
5890            if (eptr-- == pp) break;        /* Stop if tried at original pos */            if (eptr-- == pp) break;        /* Stop if tried at original pos */
5891            BACKCHAR(eptr);            BACKCHAR(eptr);
5892            if (ctype == OP_ANYNL && eptr > pp  && *eptr == '\n' &&            if (ctype == OP_ANYNL && eptr > pp  && RAWUCHAR(eptr) == CHAR_NL &&
5893                eptr[-1] == '\r') eptr--;                RAWUCHAR(eptr - 1) == CHAR_CR) eptr--;
5894            }            }
5895          }          }
5896        else        else
5897  #endif  /* SUPPORT_UTF8 */  #endif  /* SUPPORT_UTF */
5898          /* Not UTF mode */
       /* Not UTF-8 mode */  
5899          {          {
5900          switch(ctype)          switch(ctype)
5901            {            {
# Line 5593  for (;;) Line 5908  for (;;)
5908                break;                break;
5909                }                }