/[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 1271 by ph10, Wed Mar 6 16:50:38 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 715  for (;;) Line 784  for (;;)
784      /* COMMIT overrides PRUNE, SKIP, and THEN */      /* COMMIT overrides PRUNE, SKIP, and THEN */
785    
786      case OP_COMMIT:      case OP_COMMIT:
787      RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,      RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md,
788        eptrb, RM52);        eptrb, RM52);
789      if (rrc != MATCH_NOMATCH && rrc != MATCH_PRUNE &&      if (rrc != MATCH_NOMATCH && rrc != MATCH_PRUNE &&
790          rrc != MATCH_SKIP && rrc != MATCH_SKIP_ARG &&          rrc != MATCH_SKIP && rrc != MATCH_SKIP_ARG &&
# Line 726  for (;;) Line 795  for (;;)
795      /* PRUNE overrides THEN */      /* PRUNE overrides THEN */
796    
797      case OP_PRUNE:      case OP_PRUNE:
798      RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,      RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md,
799        eptrb, RM51);        eptrb, RM51);
800      if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN) RRETURN(rrc);      if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN) RRETURN(rrc);
801      RRETURN(MATCH_PRUNE);      RRETURN(MATCH_PRUNE);
# Line 734  for (;;) Line 803  for (;;)
803      case OP_PRUNE_ARG:      case OP_PRUNE_ARG:
804      md->nomatch_mark = ecode + 2;      md->nomatch_mark = ecode + 2;
805      md->mark = NULL;    /* In case previously set by assertion */      md->mark = NULL;    /* In case previously set by assertion */
806      RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode] + ecode[1], offset_top, md,      RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode] + ecode[1], offset_top, md,
807        eptrb, RM56);        eptrb, RM56);
808      if ((rrc == MATCH_MATCH || rrc == MATCH_ACCEPT) &&      if ((rrc == MATCH_MATCH || rrc == MATCH_ACCEPT) &&
809           md->mark == NULL) md->mark = ecode + 2;           md->mark == NULL) md->mark = ecode + 2;
# Line 744  for (;;) Line 813  for (;;)
813      /* SKIP overrides PRUNE and THEN */      /* SKIP overrides PRUNE and THEN */
814    
815      case OP_SKIP:      case OP_SKIP:
816      RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,      RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md,
817        eptrb, RM53);        eptrb, RM53);
818      if (rrc != MATCH_NOMATCH && rrc != MATCH_PRUNE && rrc != MATCH_THEN)      if (rrc != MATCH_NOMATCH && rrc != MATCH_PRUNE && rrc != MATCH_THEN)
819        RRETURN(rrc);        RRETURN(rrc);
# Line 752  for (;;) Line 821  for (;;)
821      RRETURN(MATCH_SKIP);      RRETURN(MATCH_SKIP);
822    
823      /* Note that, for Perl compatibility, SKIP with an argument does NOT set      /* Note that, for Perl compatibility, SKIP with an argument does NOT set
824      nomatch_mark. There is a flag that disables this opcode when re-matching a      nomatch_mark. There is a flag that disables this opcode when re-matching a
825      pattern that ended with a SKIP for which there was not a matching MARK. */      pattern that ended with a SKIP for which there was not a matching MARK. */
826    
827      case OP_SKIP_ARG:      case OP_SKIP_ARG:
828      if (md->ignore_skip_arg)      if (md->ignore_skip_arg)
829        {        {
830        ecode += _pcre_OP_lengths[*ecode] + ecode[1];        ecode += PRIV(OP_lengths)[*ecode] + ecode[1];
831        break;        break;
832        }        }
833      RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode] + ecode[1], offset_top, md,      RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode] + ecode[1], offset_top, md,
834        eptrb, RM57);        eptrb, RM57);
835      if (rrc != MATCH_NOMATCH && rrc != MATCH_PRUNE && rrc != MATCH_THEN)      if (rrc != MATCH_NOMATCH && rrc != MATCH_PRUNE && rrc != MATCH_THEN)
836        RRETURN(rrc);        RRETURN(rrc);
837    
838      /* 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
839      returning the special MATCH_SKIP_ARG return code. This will either be      returning the special MATCH_SKIP_ARG return code. This will either be
840      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
841      with the md->ignore_skip_arg flag set. */      with the md->ignore_skip_arg flag set. */
842    
843      md->start_match_ptr = ecode + 2;      md->start_match_ptr = ecode + 2;
# Line 779  for (;;) Line 848  for (;;)
848      match pointer to do this. */      match pointer to do this. */
849    
850      case OP_THEN:      case OP_THEN:
851      RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,      RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md,
852        eptrb, RM54);        eptrb, RM54);
853      if (rrc != MATCH_NOMATCH) RRETURN(rrc);      if (rrc != MATCH_NOMATCH) RRETURN(rrc);
854      md->start_match_ptr = ecode;      md->start_match_ptr = ecode;
# Line 788  for (;;) Line 857  for (;;)
857      case OP_THEN_ARG:      case OP_THEN_ARG:
858      md->nomatch_mark = ecode + 2;      md->nomatch_mark = ecode + 2;
859      md->mark = NULL;    /* In case previously set by assertion */      md->mark = NULL;    /* In case previously set by assertion */
860      RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode] + ecode[1], offset_top,      RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode] + ecode[1], offset_top,
861        md, eptrb, RM58);        md, eptrb, RM58);
862      if ((rrc == MATCH_MATCH || rrc == MATCH_ACCEPT) &&      if ((rrc == MATCH_MATCH || rrc == MATCH_ACCEPT) &&
863           md->mark == NULL) md->mark = ecode + 2;           md->mark == NULL) md->mark = ecode + 2;
# Line 812  for (;;) Line 881  for (;;)
881      case OP_ONCE_NC:      case OP_ONCE_NC:
882      prev = ecode;      prev = ecode;
883      saved_eptr = eptr;      saved_eptr = eptr;
884        save_mark = md->mark;
885      do      do
886        {        {
887        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 900  for (;;)
900    
901        if (rrc != MATCH_NOMATCH) RRETURN(rrc);        if (rrc != MATCH_NOMATCH) RRETURN(rrc);
902        ecode += GET(ecode,1);        ecode += GET(ecode,1);
903          md->mark = save_mark;
904        }        }
905      while (*ecode == OP_ALT);      while (*ecode == OP_ALT);
906    
# Line 869  for (;;) Line 940  for (;;)
940        }        }
941      else  /* OP_KETRMAX */      else  /* OP_KETRMAX */
942        {        {
       md->match_function_type = MATCH_CBEGROUP;  
943        RMATCH(eptr, prev, offset_top, md, eptrb, RM66);        RMATCH(eptr, prev, offset_top, md, eptrb, RM66);
944        if (rrc != MATCH_NOMATCH) RRETURN(rrc);        if (rrc != MATCH_NOMATCH) RRETURN(rrc);
945        ecode += 1 + LINK_SIZE;        ecode += 1 + LINK_SIZE;
# Line 909  for (;;) Line 979  for (;;)
979        save_offset2 = md->offset_vector[offset+1];        save_offset2 = md->offset_vector[offset+1];
980        save_offset3 = md->offset_vector[md->offset_end - number];        save_offset3 = md->offset_vector[md->offset_end - number];
981        save_capture_last = md->capture_last;        save_capture_last = md->capture_last;
982          save_mark = md->mark;
983    
984        DPRINTF(("saving %d %d %d\n", save_offset1, save_offset2, save_offset3));        DPRINTF(("saving %d %d %d\n", save_offset1, save_offset2, save_offset3));
985        md->offset_vector[md->offset_end - number] =        md->offset_vector[md->offset_end - number] =
# Line 917  for (;;) Line 988  for (;;)
988        for (;;)        for (;;)
989          {          {
990          if (op >= OP_SBRA) md->match_function_type = MATCH_CBEGROUP;          if (op >= OP_SBRA) md->match_function_type = MATCH_CBEGROUP;
991          RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,          RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md,
992            eptrb, RM1);            eptrb, RM1);
993          if (rrc == MATCH_ONCE) break;  /* Backing up through an atomic group */          if (rrc == MATCH_ONCE) break;  /* Backing up through an atomic group */
994    
# Line 945  for (;;) Line 1016  for (;;)
1016          if (rrc != MATCH_NOMATCH) RRETURN(rrc);          if (rrc != MATCH_NOMATCH) RRETURN(rrc);
1017          md->capture_last = save_capture_last;          md->capture_last = save_capture_last;
1018          ecode += GET(ecode, 1);          ecode += GET(ecode, 1);
1019            md->mark = save_mark;
1020          if (*ecode != OP_ALT) break;          if (*ecode != OP_ALT) break;
1021          }          }
1022    
# Line 996  for (;;) Line 1068  for (;;)
1068    
1069      for (;;)      for (;;)
1070        {        {
1071        if (op >= OP_SBRA || op == OP_ONCE) md->match_function_type = MATCH_CBEGROUP;        if (op >= OP_SBRA || op == OP_ONCE)
1072            md->match_function_type = MATCH_CBEGROUP;
1073    
1074        /* 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
1075        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 1077  for (;;)
1077    
1078        else if (!md->hasthen && ecode[GET(ecode, 1)] != OP_ALT)        else if (!md->hasthen && ecode[GET(ecode, 1)] != OP_ALT)
1079          {          {
1080          ecode += _pcre_OP_lengths[*ecode];          ecode += PRIV(OP_lengths)[*ecode];
1081          goto TAIL_RECURSE;          goto TAIL_RECURSE;
1082          }          }
1083    
1084        /* In all other cases, we have to make another call to match(). */        /* In all other cases, we have to make another call to match(). */
1085    
1086        RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md, eptrb,        save_mark = md->mark;
1087          save_capture_last = md->capture_last;
1088          RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md, eptrb,
1089          RM2);          RM2);
1090    
1091        /* 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 1103  for (;;)
1103          {          {
1104          if (rrc == MATCH_ONCE)          if (rrc == MATCH_ONCE)
1105            {            {
1106            const uschar *scode = ecode;            const pcre_uchar *scode = ecode;
1107            if (*scode != OP_ONCE)           /* If not at start, find it */            if (*scode != OP_ONCE)           /* If not at start, find it */
1108              {              {
1109              while (*scode == OP_ALT) scode += GET(scode, 1);              while (*scode == OP_ALT) scode += GET(scode, 1);
# Line 1039  for (;;) Line 1114  for (;;)
1114          RRETURN(rrc);          RRETURN(rrc);
1115          }          }
1116        ecode += GET(ecode, 1);        ecode += GET(ecode, 1);
1117          md->mark = save_mark;
1118        if (*ecode != OP_ALT) break;        if (*ecode != OP_ALT) break;
1119          md->capture_last = save_capture_last;
1120        }        }
1121    
1122      RRETURN(MATCH_NOMATCH);      RRETURN(MATCH_NOMATCH);
# Line 1070  for (;;) Line 1147  for (;;)
1147      if (offset < md->offset_max)      if (offset < md->offset_max)
1148        {        {
1149        matched_once = FALSE;        matched_once = FALSE;
1150        code_offset = ecode - md->start_code;        code_offset = (int)(ecode - md->start_code);
1151    
1152        save_offset1 = md->offset_vector[offset];        save_offset1 = md->offset_vector[offset];
1153        save_offset2 = md->offset_vector[offset+1];        save_offset2 = md->offset_vector[offset+1];
# Line 1093  for (;;) Line 1170  for (;;)
1170          md->offset_vector[md->offset_end - number] =          md->offset_vector[md->offset_end - number] =
1171            (int)(eptr - md->start_subject);            (int)(eptr - md->start_subject);
1172          if (op >= OP_SBRA) md->match_function_type = MATCH_CBEGROUP;          if (op >= OP_SBRA) md->match_function_type = MATCH_CBEGROUP;
1173          RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,          RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md,
1174            eptrb, RM63);            eptrb, RM63);
1175          if (rrc == MATCH_KETRPOS)          if (rrc == MATCH_KETRPOS)
1176            {            {
# Line 1160  for (;;) Line 1237  for (;;)
1237    
1238      POSSESSIVE_NON_CAPTURE:      POSSESSIVE_NON_CAPTURE:
1239      matched_once = FALSE;      matched_once = FALSE;
1240      code_offset = ecode - md->start_code;      code_offset = (int)(ecode - md->start_code);
1241        save_capture_last = md->capture_last;
1242    
1243      for (;;)      for (;;)
1244        {        {
1245        if (op >= OP_SBRA) md->match_function_type = MATCH_CBEGROUP;        if (op >= OP_SBRA) md->match_function_type = MATCH_CBEGROUP;
1246        RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,        RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md,
1247          eptrb, RM48);          eptrb, RM48);
1248        if (rrc == MATCH_KETRPOS)        if (rrc == MATCH_KETRPOS)
1249          {          {
# Line 1190  for (;;) Line 1268  for (;;)
1268        if (rrc != MATCH_NOMATCH) RRETURN(rrc);        if (rrc != MATCH_NOMATCH) RRETURN(rrc);
1269        ecode += GET(ecode, 1);        ecode += GET(ecode, 1);
1270        if (*ecode != OP_ALT) break;        if (*ecode != OP_ALT) break;
1271          md->capture_last = save_capture_last;
1272        }        }
1273    
1274      if (matched_once || allow_zero)      if (matched_once || allow_zero)
# Line 1215  for (;;) Line 1294  for (;;)
1294    
1295      if (ecode[LINK_SIZE+1] == OP_CALLOUT)      if (ecode[LINK_SIZE+1] == OP_CALLOUT)
1296        {        {
1297        if (pcre_callout != NULL)        if (PUBL(callout) != NULL)
1298          {          {
1299          pcre_callout_block cb;          PUBL(callout_block) cb;
1300          cb.version          = 2;   /* Version 1 of the callout block */          cb.version          = 2;   /* Version 1 of the callout block */
1301          cb.callout_number   = ecode[LINK_SIZE+2];          cb.callout_number   = ecode[LINK_SIZE+2];
1302          cb.offset_vector    = md->offset_vector;          cb.offset_vector    = md->offset_vector;
1303    #if defined COMPILE_PCRE8
1304          cb.subject          = (PCRE_SPTR)md->start_subject;          cb.subject          = (PCRE_SPTR)md->start_subject;
1305    #elif defined COMPILE_PCRE16
1306            cb.subject          = (PCRE_SPTR16)md->start_subject;
1307    #elif defined COMPILE_PCRE32
1308            cb.subject          = (PCRE_SPTR32)md->start_subject;
1309    #endif
1310          cb.subject_length   = (int)(md->end_subject - md->start_subject);          cb.subject_length   = (int)(md->end_subject - md->start_subject);
1311          cb.start_match      = (int)(mstart - md->start_subject);          cb.start_match      = (int)(mstart - md->start_subject);
1312          cb.current_position = (int)(eptr - md->start_subject);          cb.current_position = (int)(eptr - md->start_subject);
1313          cb.pattern_position = GET(ecode, LINK_SIZE + 3);          cb.pattern_position = GET(ecode, LINK_SIZE + 3);
1314          cb.next_item_length = GET(ecode, 3 + 2*LINK_SIZE);          cb.next_item_length = GET(ecode, 3 + 2*LINK_SIZE);
1315          cb.capture_top      = offset_top/2;          cb.capture_top      = offset_top/2;
1316          cb.capture_last     = md->capture_last;          cb.capture_last     = md->capture_last & CAPLMASK;
1317            /* Internal change requires this for API compatibility. */
1318            if (cb.capture_last == 0) cb.capture_last = -1;
1319          cb.callout_data     = md->callout_data;          cb.callout_data     = md->callout_data;
1320          cb.mark             = md->nomatch_mark;          cb.mark             = md->nomatch_mark;
1321          if ((rrc = (*pcre_callout)(&cb)) > 0) RRETURN(MATCH_NOMATCH);          if ((rrc = (*PUBL(callout))(&cb)) > 0) RRETURN(MATCH_NOMATCH);
1322          if (rrc < 0) RRETURN(rrc);          if (rrc < 0) RRETURN(rrc);
1323          }          }
1324        ecode += _pcre_OP_lengths[OP_CALLOUT];        ecode += PRIV(OP_lengths)[OP_CALLOUT];
1325          codelink -= PRIV(OP_lengths)[OP_CALLOUT];
1326        }        }
1327    
1328      condcode = ecode[LINK_SIZE+1];      condcode = ecode[LINK_SIZE+1];
# Line 1250  for (;;) Line 1338  for (;;)
1338          }          }
1339        else        else
1340          {          {
1341          int recno = GET2(ecode, LINK_SIZE + 2);   /* Recursion group number*/          unsigned int recno = GET2(ecode, LINK_SIZE + 2);   /* Recursion group number*/
1342          condition = (recno == RREF_ANY || recno == md->recursive->group_num);          condition = (recno == RREF_ANY || recno == md->recursive->group_num);
1343    
1344          /* 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 1348  for (;;)
1348    
1349          if (!condition && condcode == OP_NRREF)          if (!condition && condcode == OP_NRREF)
1350            {            {
1351            uschar *slotA = md->name_table;            pcre_uchar *slotA = md->name_table;
1352            for (i = 0; i < md->name_count; i++)            for (i = 0; i < md->name_count; i++)
1353              {              {
1354              if (GET2(slotA, 0) == recno) break;              if (GET2(slotA, 0) == recno) break;
# Line 1273  for (;;) Line 1361  for (;;)
1361    
1362            if (i < md->name_count)            if (i < md->name_count)
1363              {              {
1364              uschar *slotB = slotA;              pcre_uchar *slotB = slotA;
1365              while (slotB > md->name_table)              while (slotB > md->name_table)
1366                {                {
1367                slotB -= md->name_entry_size;                slotB -= md->name_entry_size;
1368                if (strcmp((char *)slotA + 2, (char *)slotB + 2) == 0)                if (STRCMP_UC_UC(slotA + IMM2_SIZE, slotB + IMM2_SIZE) == 0)
1369                  {                  {
1370                  condition = GET2(slotB, 0) == md->recursive->group_num;                  condition = GET2(slotB, 0) == md->recursive->group_num;
1371                  if (condition) break;                  if (condition) break;
# Line 1293  for (;;) Line 1381  for (;;)
1381                for (i++; i < md->name_count; i++)                for (i++; i < md->name_count; i++)
1382                  {                  {
1383                  slotB += md->name_entry_size;                  slotB += md->name_entry_size;
1384                  if (strcmp((char *)slotA + 2, (char *)slotB + 2) == 0)                  if (STRCMP_UC_UC(slotA + IMM2_SIZE, slotB + IMM2_SIZE) == 0)
1385                    {                    {
1386                    condition = GET2(slotB, 0) == md->recursive->group_num;                    condition = GET2(slotB, 0) == md->recursive->group_num;
1387                    if (condition) break;                    if (condition) break;
# Line 1306  for (;;) Line 1394  for (;;)
1394    
1395          /* Chose branch according to the condition */          /* Chose branch according to the condition */
1396    
1397          ecode += condition? 3 : GET(ecode, 1);          ecode += condition? 1 + IMM2_SIZE : GET(ecode, 1);
1398          }          }
1399        }        }
1400    
# Line 1322  for (;;) Line 1410  for (;;)
1410    
1411        if (!condition && condcode == OP_NCREF)        if (!condition && condcode == OP_NCREF)
1412          {          {
1413          int refno = offset >> 1;          unsigned int refno = offset >> 1;
1414          uschar *slotA = md->name_table;          pcre_uchar *slotA = md->name_table;
1415    
1416          for (i = 0; i < md->name_count; i++)          for (i = 0; i < md->name_count; i++)
1417            {            {
# Line 1337  for (;;) Line 1425  for (;;)
1425    
1426          if (i < md->name_count)          if (i < md->name_count)
1427            {            {
1428            uschar *slotB = slotA;            pcre_uchar *slotB = slotA;
1429            while (slotB > md->name_table)            while (slotB > md->name_table)
1430              {              {
1431              slotB -= md->name_entry_size;              slotB -= md->name_entry_size;
1432              if (strcmp((char *)slotA + 2, (char *)slotB + 2) == 0)              if (STRCMP_UC_UC(slotA + IMM2_SIZE, slotB + IMM2_SIZE) == 0)
1433                {                {
1434                offset = GET2(slotB, 0) << 1;                offset = GET2(slotB, 0) << 1;
1435                condition = offset < offset_top &&                condition = offset < offset_top &&
# Line 1359  for (;;) Line 1447  for (;;)
1447              for (i++; i < md->name_count; i++)              for (i++; i < md->name_count; i++)
1448                {                {
1449                slotB += md->name_entry_size;                slotB += md->name_entry_size;
1450                if (strcmp((char *)slotA + 2, (char *)slotB + 2) == 0)                if (STRCMP_UC_UC(slotA + IMM2_SIZE, slotB + IMM2_SIZE) == 0)
1451                  {                  {
1452                  offset = GET2(slotB, 0) << 1;                  offset = GET2(slotB, 0) << 1;
1453                  condition = offset < offset_top &&                  condition = offset < offset_top &&
# Line 1374  for (;;) Line 1462  for (;;)
1462    
1463        /* Chose branch according to the condition */        /* Chose branch according to the condition */
1464    
1465        ecode += condition? 3 : GET(ecode, 1);        ecode += condition? 1 + IMM2_SIZE : GET(ecode, 1);
1466        }        }
1467    
1468      else if (condcode == OP_DEF)     /* DEFINE - always false */      else if (condcode == OP_DEF)     /* DEFINE - always false */
# Line 1450  for (;;) Line 1538  for (;;)
1538      to close any currently open capturing brackets. */      to close any currently open capturing brackets. */
1539    
1540      case OP_CLOSE:      case OP_CLOSE:
1541      number = GET2(ecode, 1);      number = GET2(ecode, 1);   /* Must be less than 65536 */
1542      offset = number << 1;      offset = number << 1;
1543    
1544  #ifdef PCRE_DEBUG  #ifdef PCRE_DEBUG
# Line 1458  for (;;) Line 1546  for (;;)
1546        printf("\n");        printf("\n");
1547  #endif  #endif
1548    
1549      md->capture_last = number;      md->capture_last = (md->capture_last & OVFLMASK) | number;
1550      if (offset >= md->offset_max) md->offset_overflow = TRUE; else      if (offset >= md->offset_max) md->capture_last |= OVFLBIT; else
1551        {        {
1552        md->offset_vector[offset] =        md->offset_vector[offset] =
1553          md->offset_vector[md->offset_end - number];          md->offset_vector[md->offset_end - number];
1554        md->offset_vector[offset+1] = (int)(eptr - md->start_subject);        md->offset_vector[offset+1] = (int)(eptr - md->start_subject);
1555        if (offset_top <= offset) offset_top = offset + 2;        if (offset_top <= offset) offset_top = offset + 2;
1556        }        }
1557      ecode += 3;      ecode += 1 + IMM2_SIZE;
1558      break;      break;
1559    
1560    
# Line 1513  for (;;) Line 1601  for (;;)
1601    
1602      case OP_ASSERT:      case OP_ASSERT:
1603      case OP_ASSERTBACK:      case OP_ASSERTBACK:
1604        save_mark = md->mark;
1605      if (md->match_function_type == MATCH_CONDASSERT)      if (md->match_function_type == MATCH_CONDASSERT)
1606        {        {
1607        condassert = TRUE;        condassert = TRUE;
# Line 1528  for (;;) Line 1617  for (;;)
1617          mstart = md->start_match_ptr;   /* In case \K reset it */          mstart = md->start_match_ptr;   /* In case \K reset it */
1618          break;          break;
1619          }          }
1620          md->mark = save_mark;
1621    
1622        /* PCRE does not allow THEN to escape beyond an assertion; it is treated        /* A COMMIT failure must fail the entire assertion, without trying any
1623        as NOMATCH. */        subsequent branches. */
1624    
1625          if (rrc == MATCH_COMMIT) RRETURN(MATCH_NOMATCH);
1626    
1627          /* PCRE does not allow THEN to escape beyond an assertion; it
1628          is treated as NOMATCH. */
1629    
1630        if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN) RRETURN(rrc);        if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN) RRETURN(rrc);
1631        ecode += GET(ecode, 1);        ecode += GET(ecode, 1);
# Line 1557  for (;;) Line 1652  for (;;)
1652    
1653      case OP_ASSERT_NOT:      case OP_ASSERT_NOT:
1654      case OP_ASSERTBACK_NOT:      case OP_ASSERTBACK_NOT:
1655        save_mark = md->mark;
1656      if (md->match_function_type == MATCH_CONDASSERT)      if (md->match_function_type == MATCH_CONDASSERT)
1657        {        {
1658        condassert = TRUE;        condassert = TRUE;
# Line 1567  for (;;) Line 1663  for (;;)
1663      do      do
1664        {        {
1665        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, NULL, RM5);        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, NULL, RM5);
1666          md->mark = save_mark;
1667        if (rrc == MATCH_MATCH || rrc == MATCH_ACCEPT) RRETURN(MATCH_NOMATCH);        if (rrc == MATCH_MATCH || rrc == MATCH_ACCEPT) RRETURN(MATCH_NOMATCH);
1668        if (rrc == MATCH_SKIP || rrc == MATCH_PRUNE || rrc == MATCH_COMMIT)        if (rrc == MATCH_SKIP || rrc == MATCH_PRUNE || rrc == MATCH_COMMIT)
1669          {          {
# Line 1593  for (;;) Line 1690  for (;;)
1690      back a number of characters, not bytes. */      back a number of characters, not bytes. */
1691    
1692      case OP_REVERSE:      case OP_REVERSE:
1693  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
1694      if (utf8)      if (utf)
1695        {        {
1696        i = GET(ecode, 1);        i = GET(ecode, 1);
1697        while (i-- > 0)        while (i-- > 0)
# Line 1625  for (;;) Line 1722  for (;;)
1722      function is able to force a failure. */      function is able to force a failure. */
1723    
1724      case OP_CALLOUT:      case OP_CALLOUT:
1725      if (pcre_callout != NULL)      if (PUBL(callout) != NULL)
1726        {        {
1727        pcre_callout_block cb;        PUBL(callout_block) cb;
1728        cb.version          = 2;   /* Version 1 of the callout block */        cb.version          = 2;   /* Version 1 of the callout block */
1729        cb.callout_number   = ecode[1];        cb.callout_number   = ecode[1];
1730        cb.offset_vector    = md->offset_vector;        cb.offset_vector    = md->offset_vector;
1731    #if defined COMPILE_PCRE8
1732        cb.subject          = (PCRE_SPTR)md->start_subject;        cb.subject          = (PCRE_SPTR)md->start_subject;
1733    #elif defined COMPILE_PCRE16
1734          cb.subject          = (PCRE_SPTR16)md->start_subject;
1735    #elif defined COMPILE_PCRE32
1736          cb.subject          = (PCRE_SPTR32)md->start_subject;
1737    #endif
1738        cb.subject_length   = (int)(md->end_subject - md->start_subject);        cb.subject_length   = (int)(md->end_subject - md->start_subject);
1739        cb.start_match      = (int)(mstart - md->start_subject);        cb.start_match      = (int)(mstart - md->start_subject);
1740        cb.current_position = (int)(eptr - md->start_subject);        cb.current_position = (int)(eptr - md->start_subject);
1741        cb.pattern_position = GET(ecode, 2);        cb.pattern_position = GET(ecode, 2);
1742        cb.next_item_length = GET(ecode, 2 + LINK_SIZE);        cb.next_item_length = GET(ecode, 2 + LINK_SIZE);
1743        cb.capture_top      = offset_top/2;        cb.capture_top      = offset_top/2;
1744        cb.capture_last     = md->capture_last;        cb.capture_last     = md->capture_last & CAPLMASK;
1745          /* Internal change requires this for API compatibility. */
1746          if (cb.capture_last == 0) cb.capture_last = -1;
1747        cb.callout_data     = md->callout_data;        cb.callout_data     = md->callout_data;
1748        cb.mark             = md->nomatch_mark;        cb.mark             = md->nomatch_mark;
1749        if ((rrc = (*pcre_callout)(&cb)) > 0) RRETURN(MATCH_NOMATCH);        if ((rrc = (*PUBL(callout))(&cb)) > 0) RRETURN(MATCH_NOMATCH);
1750        if (rrc < 0) RRETURN(rrc);        if (rrc < 0) RRETURN(rrc);
1751        }        }
1752      ecode += 2 + 2*LINK_SIZE;      ecode += 2 + 2*LINK_SIZE;
# Line 1667  for (;;) Line 1772  for (;;)
1772      case OP_RECURSE:      case OP_RECURSE:
1773        {        {
1774        recursion_info *ri;        recursion_info *ri;
1775        int recno;        unsigned int recno;
1776    
1777        callpat = md->start_code + GET(ecode, 1);        callpat = md->start_code + GET(ecode, 1);
1778        recno = (callpat == md->start_code)? 0 :        recno = (callpat == md->start_code)? 0 :
# Line 1684  for (;;) Line 1789  for (;;)
1789        /* Add to "recursing stack" */        /* Add to "recursing stack" */
1790    
1791        new_recursive.group_num = recno;        new_recursive.group_num = recno;
1792          new_recursive.saved_capture_last = md->capture_last;
1793        new_recursive.subject_position = eptr;        new_recursive.subject_position = eptr;
1794        new_recursive.prevrec = md->recursive;        new_recursive.prevrec = md->recursive;
1795        md->recursive = &new_recursive;        md->recursive = &new_recursive;
# Line 1700  for (;;) Line 1806  for (;;)
1806        else        else
1807          {          {
1808          new_recursive.offset_save =          new_recursive.offset_save =
1809            (int *)(pcre_malloc)(new_recursive.saved_max * sizeof(int));            (int *)(PUBL(malloc))(new_recursive.saved_max * sizeof(int));
1810          if (new_recursive.offset_save == NULL) RRETURN(PCRE_ERROR_NOMEMORY);          if (new_recursive.offset_save == NULL) RRETURN(PCRE_ERROR_NOMEMORY);
1811          }          }
1812        memcpy(new_recursive.offset_save, md->offset_vector,        memcpy(new_recursive.offset_save, md->offset_vector,
1813              new_recursive.saved_max * sizeof(int));              new_recursive.saved_max * sizeof(int));
1814    
1815        /* OK, now we can do the recursion. After processing each alternative,        /* OK, now we can do the recursion. After processing each alternative,
1816        restore the offset data. If there were nested recursions, md->recursive        restore the offset data and the last captured value. If there were nested
1817        might be changed, so reset it before looping. */        recursions, md->recursive might be changed, so reset it before looping.
1818          */
1819    
1820        DPRINTF(("Recursing into group %d\n", new_recursive.group_num));        DPRINTF(("Recursing into group %d\n", new_recursive.group_num));
1821        cbegroup = (*callpat >= OP_SBRA);        cbegroup = (*callpat >= OP_SBRA);
1822        do        do
1823          {          {
1824          if (cbegroup) md->match_function_type = MATCH_CBEGROUP;          if (cbegroup) md->match_function_type = MATCH_CBEGROUP;
1825          RMATCH(eptr, callpat + _pcre_OP_lengths[*callpat], offset_top,          RMATCH(eptr, callpat + PRIV(OP_lengths)[*callpat], offset_top,
1826            md, eptrb, RM6);            md, eptrb, RM6);
1827          memcpy(md->offset_vector, new_recursive.offset_save,          memcpy(md->offset_vector, new_recursive.offset_save,
1828              new_recursive.saved_max * sizeof(int));              new_recursive.saved_max * sizeof(int));
1829            md->capture_last = new_recursive.saved_capture_last;
1830          md->recursive = new_recursive.prevrec;          md->recursive = new_recursive.prevrec;
1831          if (rrc == MATCH_MATCH || rrc == MATCH_ACCEPT)          if (rrc == MATCH_MATCH || rrc == MATCH_ACCEPT)
1832            {            {
1833            DPRINTF(("Recursion matched\n"));            DPRINTF(("Recursion matched\n"));
1834            if (new_recursive.offset_save != stacksave)            if (new_recursive.offset_save != stacksave)
1835              (pcre_free)(new_recursive.offset_save);              (PUBL(free))(new_recursive.offset_save);
1836    
1837            /* 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
1838            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 1843  for (;;)
1843            goto RECURSION_MATCHED;        /* Exit loop; end processing */            goto RECURSION_MATCHED;        /* Exit loop; end processing */
1844            }            }
1845    
1846          /* 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
1847          as NOMATCH. */          recursion; they are treated as NOMATCH. These codes are defined in a
1848            range that can be tested for. Any other return code is an error. */
1849    
1850          else if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN)          else if (rrc != MATCH_NOMATCH &&
1851                     (rrc < MATCH_BACKTRACK_MIN || rrc > MATCH_BACKTRACK_MAX))
1852            {            {
1853            DPRINTF(("Recursion gave error %d\n", rrc));            DPRINTF(("Recursion gave error %d\n", rrc));
1854            if (new_recursive.offset_save != stacksave)            if (new_recursive.offset_save != stacksave)
1855              (pcre_free)(new_recursive.offset_save);              (PUBL(free))(new_recursive.offset_save);
1856            RRETURN(rrc);            RRETURN(rrc);
1857            }            }
1858    
# Line 1754  for (;;) Line 1864  for (;;)
1864        DPRINTF(("Recursion didn't match\n"));        DPRINTF(("Recursion didn't match\n"));
1865        md->recursive = new_recursive.prevrec;        md->recursive = new_recursive.prevrec;
1866        if (new_recursive.offset_save != stacksave)        if (new_recursive.offset_save != stacksave)
1867          (pcre_free)(new_recursive.offset_save);          (PUBL(free))(new_recursive.offset_save);
1868        RRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
1869        }        }
1870    
# Line 1868  for (;;) Line 1978  for (;;)
1978    
1979        /* Deal with capturing */        /* Deal with capturing */
1980    
1981        md->capture_last = number;        md->capture_last = (md->capture_last & OVFLMASK) | number;
1982        if (offset >= md->offset_max) md->offset_overflow = TRUE; else        if (offset >= md->offset_max) md->capture_last |= OVFLBIT; else
1983          {          {
1984          /* If offset is greater than offset_top, it means that we are          /* If offset is greater than offset_top, it means that we are
1985          "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 2059  for (;;)
2059          }          }
2060        if (*prev >= OP_SBRA)    /* Could match an empty string */        if (*prev >= OP_SBRA)    /* Could match an empty string */
2061          {          {
         md->match_function_type = MATCH_CBEGROUP;  
2062          RMATCH(eptr, prev, offset_top, md, eptrb, RM50);          RMATCH(eptr, prev, offset_top, md, eptrb, RM50);
2063          RRETURN(rrc);          RRETURN(rrc);
2064          }          }
# Line 1958  for (;;) Line 2067  for (;;)
2067        }        }
2068      else  /* OP_KETRMAX */      else  /* OP_KETRMAX */
2069        {        {
       if (*prev >= OP_SBRA) md->match_function_type = MATCH_CBEGROUP;  
2070        RMATCH(eptr, prev, offset_top, md, eptrb, RM13);        RMATCH(eptr, prev, offset_top, md, eptrb, RM13);
2071        if (rrc == MATCH_ONCE && md->once_target == prev) rrc = MATCH_NOMATCH;        if (rrc == MATCH_ONCE && md->once_target == prev) rrc = MATCH_NOMATCH;
2072        if (rrc != MATCH_NOMATCH) RRETURN(rrc);        if (rrc != MATCH_NOMATCH) RRETURN(rrc);
# Line 2015  for (;;) Line 2123  for (;;)
2123    
2124      case OP_DOLLM:      case OP_DOLLM:
2125      if (eptr < md->end_subject)      if (eptr < md->end_subject)
2126        { if (!IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH); }        {
2127          if (!IS_NEWLINE(eptr))
2128            {
2129            if (md->partial != 0 &&
2130                eptr + 1 >= md->end_subject &&
2131                NLBLOCK->nltype == NLTYPE_FIXED &&
2132                NLBLOCK->nllen == 2 &&
2133                RAWUCHARTEST(eptr) == NLBLOCK->nl[0])
2134              {
2135              md->hitend = TRUE;
2136              if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
2137              }
2138            RRETURN(MATCH_NOMATCH);
2139            }
2140          }
2141      else      else
2142        {        {
2143        if (md->noteol) RRETURN(MATCH_NOMATCH);        if (md->noteol) RRETURN(MATCH_NOMATCH);
# Line 2047  for (;;) Line 2169  for (;;)
2169      ASSERT_NL_OR_EOS:      ASSERT_NL_OR_EOS:
2170      if (eptr < md->end_subject &&      if (eptr < md->end_subject &&
2171          (!IS_NEWLINE(eptr) || eptr != md->end_subject - md->nllen))          (!IS_NEWLINE(eptr) || eptr != md->end_subject - md->nllen))
2172          {
2173          if (md->partial != 0 &&
2174              eptr + 1 >= md->end_subject &&
2175              NLBLOCK->nltype == NLTYPE_FIXED &&
2176              NLBLOCK->nllen == 2 &&
2177              RAWUCHARTEST(eptr) == NLBLOCK->nl[0])
2178            {
2179            md->hitend = TRUE;
2180            if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
2181            }
2182        RRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2183          }
2184    
2185      /* Either at end of string or \n before end. */      /* Either at end of string or \n before end. */
2186    
# Line 2066  for (;;) Line 2199  for (;;)
2199        be "non-word" characters. Remember the earliest consulted character for        be "non-word" characters. Remember the earliest consulted character for
2200        partial matching. */        partial matching. */
2201    
2202  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
2203        if (utf8)        if (utf)
2204          {          {
2205          /* Get status of previous character */          /* Get status of previous character */
2206    
2207          if (eptr == md->start_subject) prev_is_word = FALSE; else          if (eptr == md->start_subject) prev_is_word = FALSE; else
2208            {            {
2209            USPTR lastptr = eptr - 1;            PCRE_PUCHAR lastptr = eptr - 1;
2210            while((*lastptr & 0xc0) == 0x80) lastptr--;            BACKCHAR(lastptr);
2211            if (lastptr < md->start_used_ptr) md->start_used_ptr = lastptr;            if (lastptr < md->start_used_ptr) md->start_used_ptr = lastptr;
2212            GETCHAR(c, lastptr);            GETCHAR(c, lastptr);
2213  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
# Line 2139  for (;;) Line 2272  for (;;)
2272              }              }
2273            else            else
2274  #endif  #endif
2275            prev_is_word = ((md->ctypes[eptr[-1]] & ctype_word) != 0);            prev_is_word = MAX_255(eptr[-1])
2276                && ((md->ctypes[eptr[-1]] & ctype_word) != 0);
2277            }            }
2278    
2279          /* Get status of next character */          /* Get status of next character */
# Line 2162  for (;;) Line 2296  for (;;)
2296            }            }
2297          else          else
2298  #endif  #endif
2299          cur_is_word = ((md->ctypes[*eptr] & ctype_word) != 0);          cur_is_word = MAX_255(*eptr)
2300              && ((md->ctypes[*eptr] & ctype_word) != 0);
2301          }          }
2302    
2303        /* Now see if the situation is what we want */        /* Now see if the situation is what we want */
# Line 2173  for (;;) Line 2308  for (;;)
2308        }        }
2309      break;      break;
2310    
2311      /* Match a single character type; inline for speed */      /* Match any single character type except newline; have to take care with
2312        CRLF newlines and partial matching. */
2313    
2314      case OP_ANY:      case OP_ANY:
2315      if (IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH);      if (IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH);
2316        if (md->partial != 0 &&
2317            eptr + 1 >= md->end_subject &&
2318            NLBLOCK->nltype == NLTYPE_FIXED &&
2319            NLBLOCK->nllen == 2 &&
2320            RAWUCHARTEST(eptr) == NLBLOCK->nl[0])
2321          {
2322          md->hitend = TRUE;
2323          if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
2324          }
2325    
2326      /* Fall through */      /* Fall through */
2327    
2328        /* Match any single character whatsoever. */
2329    
2330      case OP_ALLANY:      case OP_ALLANY:
2331      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 */
2332        {                            /* not be updated before SCHECK_PARTIAL. */        {                            /* not be updated before SCHECK_PARTIAL. */
# Line 2186  for (;;) Line 2334  for (;;)
2334        RRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2335        }        }
2336      eptr++;      eptr++;
2337      if (utf8) while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;  #ifdef SUPPORT_UTF
2338        if (utf) ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);
2339    #endif
2340      ecode++;      ecode++;
2341      break;      break;
2342    
# Line 2211  for (;;) Line 2361  for (;;)
2361        }        }
2362      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2363      if (      if (
2364  #ifdef SUPPORT_UTF8  #if defined SUPPORT_UTF || !(defined COMPILE_PCRE8)
2365         c < 256 &&         c < 256 &&
2366  #endif  #endif
2367         (md->ctypes[c] & ctype_digit) != 0         (md->ctypes[c] & ctype_digit) != 0
# Line 2228  for (;;) Line 2378  for (;;)
2378        }        }
2379      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2380      if (      if (
2381  #ifdef SUPPORT_UTF8  #if defined SUPPORT_UTF || !(defined COMPILE_PCRE8)
2382         c >= 256 ||         c > 255 ||
2383  #endif  #endif
2384         (md->ctypes[c] & ctype_digit) == 0         (md->ctypes[c] & ctype_digit) == 0
2385         )         )
# Line 2245  for (;;) Line 2395  for (;;)
2395        }        }
2396      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2397      if (      if (
2398  #ifdef SUPPORT_UTF8  #if defined SUPPORT_UTF || !(defined COMPILE_PCRE8)
2399         c < 256 &&         c < 256 &&
2400  #endif  #endif
2401         (md->ctypes[c] & ctype_space) != 0         (md->ctypes[c] & ctype_space) != 0
# Line 2262  for (;;) Line 2412  for (;;)
2412        }        }
2413      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2414      if (      if (
2415  #ifdef SUPPORT_UTF8  #if defined SUPPORT_UTF || !(defined COMPILE_PCRE8)
2416         c >= 256 ||         c > 255 ||
2417  #endif  #endif
2418         (md->ctypes[c] & ctype_space) == 0         (md->ctypes[c] & ctype_space) == 0
2419         )         )
# Line 2279  for (;;) Line 2429  for (;;)
2429        }        }
2430      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2431      if (      if (
2432  #ifdef SUPPORT_UTF8  #if defined SUPPORT_UTF || !(defined COMPILE_PCRE8)
2433         c < 256 &&         c < 256 &&
2434  #endif  #endif
2435         (md->ctypes[c] & ctype_word) != 0         (md->ctypes[c] & ctype_word) != 0
# Line 2296  for (;;) Line 2446  for (;;)
2446        }        }
2447      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2448      if (      if (
2449  #ifdef SUPPORT_UTF8  #if defined SUPPORT_UTF || !(defined COMPILE_PCRE8)
2450         c >= 256 ||         c > 255 ||
2451  #endif  #endif
2452         (md->ctypes[c] & ctype_word) == 0         (md->ctypes[c] & ctype_word) == 0
2453         )         )
# Line 2316  for (;;) Line 2466  for (;;)
2466        {        {
2467        default: RRETURN(MATCH_NOMATCH);        default: RRETURN(MATCH_NOMATCH);
2468    
2469        case 0x000d:        case CHAR_CR:
2470        if (eptr < md->end_subject && *eptr == 0x0a) eptr++;        if (eptr >= md->end_subject)
2471            {
2472            SCHECK_PARTIAL();
2473            }
2474          else if (RAWUCHARTEST(eptr) == CHAR_LF) eptr++;
2475        break;        break;
2476    
2477        case 0x000a:        case CHAR_LF:
2478        break;        break;
2479    
2480        case 0x000b:        case CHAR_VT:
2481        case 0x000c:        case CHAR_FF:
2482        case 0x0085:        case CHAR_NEL:
2483    #ifndef EBCDIC
2484        case 0x2028:        case 0x2028:
2485        case 0x2029:        case 0x2029:
2486    #endif  /* Not EBCDIC */
2487        if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);        if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);
2488        break;        break;
2489        }        }
# Line 2343  for (;;) Line 2499  for (;;)
2499      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2500      switch(c)      switch(c)
2501        {        {
2502          HSPACE_CASES: RRETURN(MATCH_NOMATCH);  /* Byte and multibyte cases */
2503        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);  
2504        }        }
2505      ecode++;      ecode++;
2506      break;      break;
# Line 2377  for (;;) Line 2514  for (;;)
2514      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2515      switch(c)      switch(c)
2516        {        {
2517          HSPACE_CASES: break;  /* Byte and multibyte cases */
2518        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;  
2519        }        }
2520      ecode++;      ecode++;
2521      break;      break;
# Line 2411  for (;;) Line 2529  for (;;)
2529      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2530      switch(c)      switch(c)
2531        {        {
2532          VSPACE_CASES: RRETURN(MATCH_NOMATCH);
2533        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);  
2534        }        }
2535      ecode++;      ecode++;
2536      break;      break;
# Line 2433  for (;;) Line 2544  for (;;)
2544      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2545      switch(c)      switch(c)
2546        {        {
2547          VSPACE_CASES: break;
2548        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;  
2549        }        }
2550      ecode++;      ecode++;
2551      break;      break;
# Line 2459  for (;;) Line 2563  for (;;)
2563        }        }
2564      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2565        {        {
2566          const pcre_uint32 *cp;
2567        const ucd_record *prop = GET_UCD(c);        const ucd_record *prop = GET_UCD(c);
2568    
2569        switch(ecode[1])        switch(ecode[1])
# Line 2475  for (;;) Line 2580  for (;;)
2580          break;          break;
2581    
2582          case PT_GC:          case PT_GC:
2583          if ((ecode[2] != _pcre_ucp_gentype[prop->chartype]) == (op == OP_PROP))          if ((ecode[2] != PRIV(ucp_gentype)[prop->chartype]) == (op == OP_PROP))
2584            RRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
2585          break;          break;
2586    
# Line 2492  for (;;) Line 2597  for (;;)
2597          /* These are specials */          /* These are specials */
2598    
2599          case PT_ALNUM:          case PT_ALNUM:
2600          if ((_pcre_ucp_gentype[prop->chartype] == ucp_L ||          if ((PRIV(ucp_gentype)[prop->chartype] == ucp_L ||
2601               _pcre_ucp_gentype[prop->chartype] == ucp_N) == (op == OP_NOTPROP))               PRIV(ucp_gentype)[prop->chartype] == ucp_N) == (op == OP_NOTPROP))
2602            RRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
2603          break;          break;
2604    
2605          case PT_SPACE:    /* Perl space */          case PT_SPACE:    /* Perl space */
2606          if ((_pcre_ucp_gentype[prop->chartype] == ucp_Z ||          if ((PRIV(ucp_gentype)[prop->chartype] == ucp_Z ||
2607               c == CHAR_HT || c == CHAR_NL || c == CHAR_FF || c == CHAR_CR)               c == CHAR_HT || c == CHAR_NL || c == CHAR_FF || c == CHAR_CR)
2608                 == (op == OP_NOTPROP))                 == (op == OP_NOTPROP))
2609            RRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
2610          break;          break;
2611    
2612          case PT_PXSPACE:  /* POSIX space */          case PT_PXSPACE:  /* POSIX space */
2613          if ((_pcre_ucp_gentype[prop->chartype] == ucp_Z ||          if ((PRIV(ucp_gentype)[prop->chartype] == ucp_Z ||
2614               c == CHAR_HT || c == CHAR_NL || c == CHAR_VT ||               c == CHAR_HT || c == CHAR_NL || c == CHAR_VT ||
2615               c == CHAR_FF || c == CHAR_CR)               c == CHAR_FF || c == CHAR_CR)
2616                 == (op == OP_NOTPROP))                 == (op == OP_NOTPROP))
# Line 2513  for (;;) Line 2618  for (;;)
2618          break;          break;
2619    
2620          case PT_WORD:          case PT_WORD:
2621          if ((_pcre_ucp_gentype[prop->chartype] == ucp_L ||          if ((PRIV(ucp_gentype)[prop->chartype] == ucp_L ||
2622               _pcre_ucp_gentype[prop->chartype] == ucp_N ||               PRIV(ucp_gentype)[prop->chartype] == ucp_N ||
2623               c == CHAR_UNDERSCORE) == (op == OP_NOTPROP))               c == CHAR_UNDERSCORE) == (op == OP_NOTPROP))
2624            RRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
2625          break;          break;
2626    
2627            case PT_CLIST:
2628            cp = PRIV(ucd_caseless_sets) + ecode[2];
2629            for (;;)
2630              {
2631              if (c < *cp)
2632                { if (op == OP_PROP) { RRETURN(MATCH_NOMATCH); } else break; }
2633              if (c == *cp++)
2634                { if (op == OP_PROP) break; else { RRETURN(MATCH_NOMATCH); } }
2635              }
2636            break;
2637    
2638            case PT_UCNC:
2639            if ((c == CHAR_DOLLAR_SIGN || c == CHAR_COMMERCIAL_AT ||
2640                 c == CHAR_GRAVE_ACCENT || (c >= 0xa0 && c <= 0xd7ff) ||
2641                 c >= 0xe000) == (op == OP_NOTPROP))
2642              RRETURN(MATCH_NOMATCH);
2643            break;
2644    
2645          /* This should never occur */          /* This should never occur */
2646    
2647          default:          default:
# Line 2538  for (;;) Line 2661  for (;;)
2661        SCHECK_PARTIAL();        SCHECK_PARTIAL();
2662        RRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2663        }        }
2664      GETCHARINCTEST(c, eptr);      else
     if (UCD_CATEGORY(c) == ucp_M) RRETURN(MATCH_NOMATCH);  
     while (eptr < md->end_subject)  
2665        {        {
2666        int len = 1;        int lgb, rgb;
2667        if (!utf8) c = *eptr; else { GETCHARLEN(c, eptr, len); }        GETCHARINCTEST(c, eptr);
2668        if (UCD_CATEGORY(c) != ucp_M) break;        lgb = UCD_GRAPHBREAK(c);
2669        eptr += len;        while (eptr < md->end_subject)
2670            {
2671            int len = 1;
2672            if (!utf) c = *eptr; else { GETCHARLEN(c, eptr, len); }
2673            rgb = UCD_GRAPHBREAK(c);
2674            if ((PRIV(ucp_gbtable)[lgb] & (1 << rgb)) == 0) break;
2675            lgb = rgb;
2676            eptr += len;
2677            }
2678        }        }
2679        CHECK_PARTIAL();
2680      ecode++;      ecode++;
2681      break;      break;
2682  #endif  #endif  /* SUPPORT_UCP */
2683    
2684    
2685      /* 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 2694  for (;;)
2694      case OP_REFI:      case OP_REFI:
2695      caseless = op == OP_REFI;      caseless = op == OP_REFI;
2696      offset = GET2(ecode, 1) << 1;               /* Doubled ref number */      offset = GET2(ecode, 1) << 1;               /* Doubled ref number */
2697      ecode += 3;      ecode += 1 + IMM2_SIZE;
2698    
2699      /* If the reference is unset, there are two possibilities:      /* If the reference is unset, there are two possibilities:
2700    
# Line 2604  for (;;) Line 2734  for (;;)
2734        case OP_CRMINRANGE:        case OP_CRMINRANGE:
2735        minimize = (*ecode == OP_CRMINRANGE);        minimize = (*ecode == OP_CRMINRANGE);
2736        min = GET2(ecode, 1);        min = GET2(ecode, 1);
2737        max = GET2(ecode, 3);        max = GET2(ecode, 1 + IMM2_SIZE);
2738        if (max == 0) max = INT_MAX;        if (max == 0) max = INT_MAX;
2739        ecode += 5;        ecode += 1 + 2 * IMM2_SIZE;
2740        break;        break;
2741    
2742        default:               /* No repeat follows */        default:               /* No repeat follows */
2743        if ((length = match_ref(offset, eptr, length, md, caseless)) < 0)        if ((length = match_ref(offset, eptr, length, md, caseless)) < 0)
2744          {          {
2745            if (length == -2) eptr = md->end_subject;   /* Partial match */
2746          CHECK_PARTIAL();          CHECK_PARTIAL();
2747          RRETURN(MATCH_NOMATCH);          RRETURN(MATCH_NOMATCH);
2748          }          }
# Line 2620  for (;;) Line 2751  for (;;)
2751        }        }
2752    
2753      /* Handle repeated back references. If the length of the reference is      /* Handle repeated back references. If the length of the reference is
2754      zero, just continue with the main loop. */      zero, just continue with the main loop. If the length is negative, it
2755        means the reference is unset in non-Java-compatible mode. If the minimum is
2756        zero, we can continue at the same level without recursion. For any other
2757        minimum, carrying on will result in NOMATCH. */
2758    
2759      if (length == 0) continue;      if (length == 0) continue;
2760        if (length < 0 && min == 0) continue;
2761    
2762      /* First, ensure the minimum number of matches are present. We get back      /* First, ensure the minimum number of matches are present. We get back
2763      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 2768  for (;;)
2768        int slength;        int slength;
2769        if ((slength = match_ref(offset, eptr, length, md, caseless)) < 0)        if ((slength = match_ref(offset, eptr, length, md, caseless)) < 0)
2770          {          {
2771            if (slength == -2) eptr = md->end_subject;   /* Partial match */
2772          CHECK_PARTIAL();          CHECK_PARTIAL();
2773          RRETURN(MATCH_NOMATCH);          RRETURN(MATCH_NOMATCH);
2774          }          }
# Line 2656  for (;;) Line 2792  for (;;)
2792          if (fi >= max) RRETURN(MATCH_NOMATCH);          if (fi >= max) RRETURN(MATCH_NOMATCH);
2793          if ((slength = match_ref(offset, eptr, length, md, caseless)) < 0)          if ((slength = match_ref(offset, eptr, length, md, caseless)) < 0)
2794            {            {
2795              if (slength == -2) eptr = md->end_subject;   /* Partial match */
2796            CHECK_PARTIAL();            CHECK_PARTIAL();
2797            RRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
2798            }            }
# Line 2674  for (;;) Line 2811  for (;;)
2811          int slength;          int slength;
2812          if ((slength = match_ref(offset, eptr, length, md, caseless)) < 0)          if ((slength = match_ref(offset, eptr, length, md, caseless)) < 0)
2813            {            {
2814            CHECK_PARTIAL();            /* Can't use CHECK_PARTIAL because we don't want to update eptr in
2815              the soft partial matching case. */
2816    
2817              if (slength == -2 && md->partial != 0 &&
2818                  md->end_subject > md->start_used_ptr)
2819                {
2820                md->hitend = TRUE;
2821                if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
2822                }
2823            break;            break;
2824            }            }
2825          eptr += slength;          eptr += slength;
2826          }          }
2827    
2828        while (eptr >= pp)        while (eptr >= pp)
2829          {          {
2830          RMATCH(eptr, ecode, offset_top, md, eptrb, RM15);          RMATCH(eptr, ecode, offset_top, md, eptrb, RM15);
# Line 2703  for (;;) Line 2849  for (;;)
2849      case OP_NCLASS:      case OP_NCLASS:
2850      case OP_CLASS:      case OP_CLASS:
2851        {        {
2852          /* The data variable is saved across frames, so the byte map needs to
2853          be stored there. */
2854    #define BYTE_MAP ((pcre_uint8 *)data)
2855        data = ecode + 1;                /* Save for matching */        data = ecode + 1;                /* Save for matching */
2856        ecode += 33;                     /* Advance past the item */        ecode += 1 + (32 / sizeof(pcre_uchar)); /* Advance past the item */
2857    
2858        switch (*ecode)        switch (*ecode)
2859          {          {
# Line 2725  for (;;) Line 2874  for (;;)
2874          case OP_CRMINRANGE:          case OP_CRMINRANGE:
2875          minimize = (*ecode == OP_CRMINRANGE);          minimize = (*ecode == OP_CRMINRANGE);
2876          min = GET2(ecode, 1);          min = GET2(ecode, 1);
2877          max = GET2(ecode, 3);          max = GET2(ecode, 1 + IMM2_SIZE);
2878          if (max == 0) max = INT_MAX;          if (max == 0) max = INT_MAX;
2879          ecode += 5;          ecode += 1 + 2 * IMM2_SIZE;
2880          break;          break;
2881    
2882          default:               /* No repeat follows */          default:               /* No repeat follows */
# Line 2737  for (;;) Line 2886  for (;;)
2886    
2887        /* First, ensure the minimum number of matches are present. */        /* First, ensure the minimum number of matches are present. */
2888    
2889  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
2890        /* UTF-8 mode */        if (utf)
       if (utf8)  
2891          {          {
2892          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
2893            {            {
# Line 2754  for (;;) Line 2902  for (;;)
2902              if (op == OP_CLASS) RRETURN(MATCH_NOMATCH);              if (op == OP_CLASS) RRETURN(MATCH_NOMATCH);
2903              }              }
2904            else            else
2905              {              if ((BYTE_MAP[c/8] & (1 << (c&7))) == 0) RRETURN(MATCH_NOMATCH);
             if ((data[c/8] & (1 << (c&7))) == 0) RRETURN(MATCH_NOMATCH);  
             }  
2906            }            }
2907          }          }
2908        else        else
2909  #endif  #endif
2910        /* Not UTF-8 mode */        /* Not UTF mode */
2911          {          {
2912          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
2913            {            {
# Line 2771  for (;;) Line 2917  for (;;)
2917              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
2918              }              }
2919            c = *eptr++;            c = *eptr++;
2920            if ((data[c/8] & (1 << (c&7))) == 0) RRETURN(MATCH_NOMATCH);  #ifndef COMPILE_PCRE8
2921              if (c > 255)
2922                {
2923                if (op == OP_CLASS) RRETURN(MATCH_NOMATCH);
2924                }
2925              else
2926    #endif
2927                if ((BYTE_MAP[c/8] & (1 << (c&7))) == 0) RRETURN(MATCH_NOMATCH);
2928            }            }
2929          }          }
2930    
# Line 2785  for (;;) Line 2938  for (;;)
2938    
2939        if (minimize)        if (minimize)
2940          {          {
2941  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
2942          /* UTF-8 mode */          if (utf)
         if (utf8)  
2943            {            {
2944            for (fi = min;; fi++)            for (fi = min;; fi++)
2945              {              {
# Line 2805  for (;;) Line 2957  for (;;)
2957                if (op == OP_CLASS) RRETURN(MATCH_NOMATCH);                if (op == OP_CLASS) RRETURN(MATCH_NOMATCH);
2958                }                }
2959              else              else
2960                {                if ((BYTE_MAP[c/8] & (1 << (c&7))) == 0) RRETURN(MATCH_NOMATCH);
               if ((data[c/8] & (1 << (c&7))) == 0) RRETURN(MATCH_NOMATCH);  
               }  
2961              }              }
2962            }            }
2963          else          else
2964  #endif  #endif
2965          /* Not UTF-8 mode */          /* Not UTF mode */
2966            {            {
2967            for (fi = min;; fi++)            for (fi = min;; fi++)
2968              {              {
# Line 2825  for (;;) Line 2975  for (;;)
2975                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
2976                }                }
2977              c = *eptr++;              c = *eptr++;
2978              if ((data[c/8] & (1 << (c&7))) == 0) RRETURN(MATCH_NOMATCH);  #ifndef COMPILE_PCRE8
2979                if (c > 255)
2980                  {
2981                  if (op == OP_CLASS) RRETURN(MATCH_NOMATCH);
2982                  }
2983                else
2984    #endif
2985                  if ((BYTE_MAP[c/8] & (1 << (c&7))) == 0) RRETURN(MATCH_NOMATCH);
2986              }              }
2987            }            }
2988          /* Control never gets here */          /* Control never gets here */
# Line 2837  for (;;) Line 2994  for (;;)
2994          {          {
2995          pp = eptr;          pp = eptr;
2996    
2997  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
2998          /* UTF-8 mode */          if (utf)
         if (utf8)  
2999            {            {
3000            for (i = min; i < max; i++)            for (i = min; i < max; i++)
3001              {              {
# Line 2855  for (;;) Line 3011  for (;;)
3011                if (op == OP_CLASS) break;                if (op == OP_CLASS) break;
3012                }                }
3013              else              else
3014                {                if ((BYTE_MAP[c/8] & (1 << (c&7))) == 0) break;
               if ((data[c/8] & (1 << (c&7))) == 0) break;  
               }  
3015              eptr += len;              eptr += len;
3016              }              }
3017            for (;;)            for (;;)
# Line 2870  for (;;) Line 3024  for (;;)
3024            }            }
3025          else          else
3026  #endif  #endif
3027            /* Not UTF-8 mode */            /* Not UTF mode */
3028            {            {
3029            for (i = min; i < max; i++)            for (i = min; i < max; i++)
3030              {              {
# Line 2880  for (;;) Line 3034  for (;;)
3034                break;                break;
3035                }                }
3036              c = *eptr;              c = *eptr;
3037              if ((data[c/8] & (1 << (c&7))) == 0) break;  #ifndef COMPILE_PCRE8
3038                if (c > 255)
3039                  {
3040                  if (op == OP_CLASS) break;
3041                  }
3042                else
3043    #endif
3044                  if ((BYTE_MAP[c/8] & (1 << (c&7))) == 0) break;
3045              eptr++;              eptr++;
3046              }              }
3047            while (eptr >= pp)            while (eptr >= pp)
# Line 2893  for (;;) Line 3054  for (;;)
3054    
3055          RRETURN(MATCH_NOMATCH);          RRETURN(MATCH_NOMATCH);
3056          }          }
3057    #undef BYTE_MAP
3058        }        }
3059      /* Control never gets here */      /* Control never gets here */
3060    
# Line 2901  for (;;) Line 3063  for (;;)
3063      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
3064      mode, because Unicode properties are supported in non-UTF-8 mode. */      mode, because Unicode properties are supported in non-UTF-8 mode. */
3065    
3066  #ifdef SUPPORT_UTF8  #if defined SUPPORT_UTF || !defined COMPILE_PCRE8
3067      case OP_XCLASS:      case OP_XCLASS:
3068        {        {
3069        data = ecode + 1 + LINK_SIZE;                /* Save for matching */        data = ecode + 1 + LINK_SIZE;                /* Save for matching */
# Line 2926  for (;;) Line 3088  for (;;)
3088          case OP_CRMINRANGE:          case OP_CRMINRANGE:
3089          minimize = (*ecode == OP_CRMINRANGE);          minimize = (*ecode == OP_CRMINRANGE);
3090          min = GET2(ecode, 1);          min = GET2(ecode, 1);
3091          max = GET2(ecode, 3);          max = GET2(ecode, 1 + IMM2_SIZE);
3092          if (max == 0) max = INT_MAX;          if (max == 0) max = INT_MAX;
3093          ecode += 5;          ecode += 1 + 2 * IMM2_SIZE;
3094          break;          break;
3095    
3096          default:               /* No repeat follows */          default:               /* No repeat follows */
# Line 2946  for (;;) Line 3108  for (;;)
3108            RRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
3109            }            }
3110          GETCHARINCTEST(c, eptr);          GETCHARINCTEST(c, eptr);
3111          if (!_pcre_xclass(c, data)) RRETURN(MATCH_NOMATCH);          if (!PRIV(xclass)(c, data, utf)) RRETURN(MATCH_NOMATCH);
3112          }          }
3113    
3114        /* 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 3132  for (;;)
3132              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
3133              }              }
3134            GETCHARINCTEST(c, eptr);            GETCHARINCTEST(c, eptr);
3135            if (!_pcre_xclass(c, data)) RRETURN(MATCH_NOMATCH);            if (!PRIV(xclass)(c, data, utf)) RRETURN(MATCH_NOMATCH);
3136            }            }
3137          /* Control never gets here */          /* Control never gets here */
3138          }          }
# Line 2988  for (;;) Line 3150  for (;;)
3150              SCHECK_PARTIAL();              SCHECK_PARTIAL();
3151              break;              break;
3152              }              }
3153    #ifdef SUPPORT_UTF
3154            GETCHARLENTEST(c, eptr, len);            GETCHARLENTEST(c, eptr, len);
3155            if (!_pcre_xclass(c, data)) break;  #else
3156              c = *eptr;
3157    #endif
3158              if (!PRIV(xclass)(c, data, utf)) break;
3159            eptr += len;            eptr += len;
3160            }            }
3161          for(;;)          for(;;)
# Line 2997  for (;;) Line 3163  for (;;)
3163            RMATCH(eptr, ecode, offset_top, md, eptrb, RM21);            RMATCH(eptr, ecode, offset_top, md, eptrb, RM21);
3164            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3165            if (eptr-- == pp) break;        /* Stop if tried at original pos */            if (eptr-- == pp) break;        /* Stop if tried at original pos */
3166            if (utf8) BACKCHAR(eptr);  #ifdef SUPPORT_UTF
3167              if (utf) BACKCHAR(eptr);
3168    #endif
3169            }            }
3170          RRETURN(MATCH_NOMATCH);          RRETURN(MATCH_NOMATCH);
3171          }          }
# Line 3009  for (;;) Line 3177  for (;;)
3177      /* Match a single character, casefully */      /* Match a single character, casefully */
3178    
3179      case OP_CHAR:      case OP_CHAR:
3180  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
3181      if (utf8)      if (utf)
3182        {        {
3183        length = 1;        length = 1;
3184        ecode++;        ecode++;
# Line 3020  for (;;) Line 3188  for (;;)
3188          CHECK_PARTIAL();             /* Not SCHECK_PARTIAL() */          CHECK_PARTIAL();             /* Not SCHECK_PARTIAL() */
3189          RRETURN(MATCH_NOMATCH);          RRETURN(MATCH_NOMATCH);
3190          }          }
3191        while (length-- > 0) if (*ecode++ != *eptr++) RRETURN(MATCH_NOMATCH);        while (length-- > 0) if (*ecode++ != RAWUCHARINC(eptr)) RRETURN(MATCH_NOMATCH);
3192        }        }
3193      else      else
3194  #endif  #endif
3195        /* Not UTF mode */
     /* Non-UTF-8 mode */  
3196        {        {
3197        if (md->end_subject - eptr < 1)        if (md->end_subject - eptr < 1)
3198          {          {
# Line 3037  for (;;) Line 3204  for (;;)
3204        }        }
3205      break;      break;
3206    
3207      /* Match a single character, caselessly */      /* Match a single character, caselessly. If we are at the end of the
3208        subject, give up immediately. */
3209    
3210      case OP_CHARI:      case OP_CHARI:
3211  #ifdef SUPPORT_UTF8      if (eptr >= md->end_subject)
3212      if (utf8)        {
3213          SCHECK_PARTIAL();
3214          RRETURN(MATCH_NOMATCH);
3215          }
3216    
3217    #ifdef SUPPORT_UTF
3218        if (utf)
3219        {        {
3220        length = 1;        length = 1;
3221        ecode++;        ecode++;
3222        GETCHARLEN(fc, ecode, length);        GETCHARLEN(fc, ecode, length);
3223    
       if (length > md->end_subject - eptr)  
         {  
         CHECK_PARTIAL();             /* Not SCHECK_PARTIAL() */  
         RRETURN(MATCH_NOMATCH);  
         }  
   
3224        /* 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
3225        can use the fast lookup table. */        we know that its other case must also be one byte long, so we can use the
3226          fast lookup table. We know that there is at least one byte left in the
3227          subject. */
3228    
3229        if (fc < 128)        if (fc < 128)
3230          {          {
3231          if (md->lcc[*ecode++] != md->lcc[*eptr++]) RRETURN(MATCH_NOMATCH);          pcre_uint32 cc = RAWUCHAR(eptr);
3232            if (md->lcc[fc] != TABLE_GET(cc, md->lcc, cc)) RRETURN(MATCH_NOMATCH);
3233            ecode++;
3234            eptr++;
3235          }          }
3236    
3237        /* Otherwise we must pick up the subject character */        /* Otherwise we must pick up the subject character. Note that we cannot
3238          use the value of "length" to check for sufficient bytes left, because the
3239          other case of the character may have more or fewer bytes.  */
3240    
3241        else        else
3242          {          {
3243          unsigned int dc;          pcre_uint32 dc;
3244          GETCHARINC(dc, eptr);          GETCHARINC(dc, eptr);
3245          ecode += length;          ecode += length;
3246    
# Line 3082  for (;;) Line 3257  for (;;)
3257          }          }
3258        }        }
3259      else      else
3260  #endif   /* SUPPORT_UTF8 */  #endif   /* SUPPORT_UTF */
3261    
3262      /* Non-UTF-8 mode */      /* Not UTF mode */
3263        {        {
3264        if (md->end_subject - eptr < 1)        if (TABLE_GET(ecode[1], md->lcc, ecode[1])
3265          {            != TABLE_GET(*eptr, md->lcc, *eptr)) RRETURN(MATCH_NOMATCH);
3266          SCHECK_PARTIAL();            /* This one can use SCHECK_PARTIAL() */        eptr++;
         RRETURN(MATCH_NOMATCH);  
         }  
       if (md->lcc[ecode[1]] != md->lcc[*eptr++]) RRETURN(MATCH_NOMATCH);  
3267        ecode += 2;        ecode += 2;
3268        }        }
3269      break;      break;
# Line 3101  for (;;) Line 3273  for (;;)
3273      case OP_EXACT:      case OP_EXACT:
3274      case OP_EXACTI:      case OP_EXACTI:
3275      min = max = GET2(ecode, 1);      min = max = GET2(ecode, 1);
3276      ecode += 3;      ecode += 1 + IMM2_SIZE;
3277      goto REPEATCHAR;      goto REPEATCHAR;
3278    
3279      case OP_POSUPTO:      case OP_POSUPTO:
# Line 3116  for (;;) Line 3288  for (;;)
3288      min = 0;      min = 0;
3289      max = GET2(ecode, 1);      max = GET2(ecode, 1);
3290      minimize = *ecode == OP_MINUPTO || *ecode == OP_MINUPTOI;      minimize = *ecode == OP_MINUPTO || *ecode == OP_MINUPTOI;
3291      ecode += 3;      ecode += 1 + IMM2_SIZE;
3292      goto REPEATCHAR;      goto REPEATCHAR;
3293    
3294      case OP_POSSTAR:      case OP_POSSTAR:
# Line 3164  for (;;) Line 3336  for (;;)
3336      /* Common code for all repeated single-character matches. */      /* Common code for all repeated single-character matches. */
3337    
3338      REPEATCHAR:      REPEATCHAR:
3339  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
3340      if (utf8)      if (utf)
3341        {        {
3342        length = 1;        length = 1;
3343        charptr = ecode;        charptr = ecode;
# Line 3178  for (;;) Line 3350  for (;;)
3350        if (length > 1)        if (length > 1)
3351          {          {
3352  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
3353          unsigned int othercase;          pcre_uint32 othercase;
3354          if (op >= OP_STARI &&     /* Caseless */          if (op >= OP_STARI &&     /* Caseless */
3355              (othercase = UCD_OTHERCASE(fc)) != fc)              (othercase = UCD_OTHERCASE(fc)) != fc)
3356            oclength = _pcre_ord2utf8(othercase, occhars);            oclength = PRIV(ord2utf)(othercase, occhars);
3357          else oclength = 0;          else oclength = 0;
3358  #endif  /* SUPPORT_UCP */  #endif  /* SUPPORT_UCP */
3359    
3360          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
3361            {            {
3362            if (eptr <= md->end_subject - length &&            if (eptr <= md->end_subject - length &&
3363              memcmp(eptr, charptr, length) == 0) eptr += length;              memcmp(eptr, charptr, IN_UCHARS(length)) == 0) eptr += length;
3364  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
3365            else if (oclength > 0 &&            else if (oclength > 0 &&
3366                     eptr <= md->end_subject - oclength &&                     eptr <= md->end_subject - oclength &&
3367                     memcmp(eptr, occhars, oclength) == 0) eptr += oclength;                     memcmp(eptr, occhars, IN_UCHARS(oclength)) == 0) eptr += oclength;
3368  #endif  /* SUPPORT_UCP */  #endif  /* SUPPORT_UCP */
3369            else            else
3370              {              {
# Line 3211  for (;;) Line 3383  for (;;)
3383              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3384              if (fi >= max) RRETURN(MATCH_NOMATCH);              if (fi >= max) RRETURN(MATCH_NOMATCH);
3385              if (eptr <= md->end_subject - length &&              if (eptr <= md->end_subject - length &&
3386                memcmp(eptr, charptr, length) == 0) eptr += length;                memcmp(eptr, charptr, IN_UCHARS(length)) == 0) eptr += length;
3387  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
3388              else if (oclength > 0 &&              else if (oclength > 0 &&
3389                       eptr <= md->end_subject - oclength &&                       eptr <= md->end_subject - oclength &&
3390                       memcmp(eptr, occhars, oclength) == 0) eptr += oclength;                       memcmp(eptr, occhars, IN_UCHARS(oclength)) == 0) eptr += oclength;
3391  #endif  /* SUPPORT_UCP */  #endif  /* SUPPORT_UCP */
3392              else              else
3393                {                {
# Line 3232  for (;;) Line 3404  for (;;)
3404            for (i = min; i < max; i++)            for (i = min; i < max; i++)
3405              {              {
3406              if (eptr <= md->end_subject - length &&              if (eptr <= md->end_subject - length &&
3407                  memcmp(eptr, charptr, length) == 0) eptr += length;                  memcmp(eptr, charptr, IN_UCHARS(length)) == 0) eptr += length;
3408  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
3409              else if (oclength > 0 &&              else if (oclength > 0 &&
3410                       eptr <= md->end_subject - oclength &&                       eptr <= md->end_subject - oclength &&
3411                       memcmp(eptr, occhars, oclength) == 0) eptr += oclength;                       memcmp(eptr, occhars, IN_UCHARS(oclength)) == 0) eptr += oclength;
3412  #endif  /* SUPPORT_UCP */  #endif  /* SUPPORT_UCP */
3413              else              else
3414                {                {
# Line 3268  for (;;) Line 3440  for (;;)
3440        value of fc will always be < 128. */        value of fc will always be < 128. */
3441        }        }
3442      else      else
3443  #endif  /* SUPPORT_UTF8 */  #endif  /* SUPPORT_UTF */
3444          /* When not in UTF-8 mode, load a single-byte character. */
3445      /* When not in UTF-8 mode, load a single-byte character. */        fc = *ecode++;
3446    
3447      fc = *ecode++;      /* The value of fc at this point is always one character, though we may
3448        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  
3449      caseful cases, for speed, since matching characters is likely to be quite      caseful cases, for speed, since matching characters is likely to be quite
3450      common. First, ensure the minimum number of matches are present. If min =      common. First, ensure the minimum number of matches are present. If min =
3451      max, continue at the same level without recursing. Otherwise, if      max, continue at the same level without recursing. Otherwise, if
# Line 3284  for (;;) Line 3454  for (;;)
3454      maximizing, find the maximum number of characters and work backwards. */      maximizing, find the maximum number of characters and work backwards. */
3455    
3456      DPRINTF(("matching %c{%d,%d} against subject %.*s\n", fc, min, max,      DPRINTF(("matching %c{%d,%d} against subject %.*s\n", fc, min, max,
3457        max, eptr));        max, (char *)eptr));
3458    
3459      if (op >= OP_STARI)  /* Caseless */      if (op >= OP_STARI)  /* Caseless */
3460        {        {
3461        fc = md->lcc[fc];  #ifdef COMPILE_PCRE8
3462          /* fc must be < 128 if UTF is enabled. */
3463          foc = md->fcc[fc];
3464    #else
3465    #ifdef SUPPORT_UTF
3466    #ifdef SUPPORT_UCP
3467          if (utf && fc > 127)
3468            foc = UCD_OTHERCASE(fc);
3469    #else
3470          if (utf && fc > 127)
3471            foc = fc;
3472    #endif /* SUPPORT_UCP */
3473          else
3474    #endif /* SUPPORT_UTF */
3475            foc = TABLE_GET(fc, md->fcc, fc);
3476    #endif /* COMPILE_PCRE8 */
3477    
3478        for (i = 1; i <= min; i++)        for (i = 1; i <= min; i++)
3479          {          {
3480            pcre_uint32 cc;                 /* Faster than pcre_uchar */
3481          if (eptr >= md->end_subject)          if (eptr >= md->end_subject)
3482            {            {
3483            SCHECK_PARTIAL();            SCHECK_PARTIAL();
3484            RRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
3485            }            }
3486          if (fc != md->lcc[*eptr++]) RRETURN(MATCH_NOMATCH);          cc = RAWUCHARTEST(eptr);
3487            if (fc != cc && foc != cc) RRETURN(MATCH_NOMATCH);
3488            eptr++;
3489          }          }
3490        if (min == max) continue;        if (min == max) continue;
3491        if (minimize)        if (minimize)
3492          {          {
3493          for (fi = min;; fi++)          for (fi = min;; fi++)
3494            {            {
3495              pcre_uint32 cc;               /* Faster than pcre_uchar */
3496            RMATCH(eptr, ecode, offset_top, md, eptrb, RM24);            RMATCH(eptr, ecode, offset_top, md, eptrb, RM24);
3497            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3498            if (fi >= max) RRETURN(MATCH_NOMATCH);            if (fi >= max) RRETURN(MATCH_NOMATCH);
# Line 3311  for (;;) Line 3501  for (;;)
3501              SCHECK_PARTIAL();              SCHECK_PARTIAL();
3502              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
3503              }              }
3504            if (fc != md->lcc[*eptr++]) RRETURN(MATCH_NOMATCH);            cc = RAWUCHARTEST(eptr);
3505              if (fc != cc && foc != cc) RRETURN(MATCH_NOMATCH);
3506              eptr++;
3507            }            }
3508          /* Control never gets here */          /* Control never gets here */
3509          }          }
# Line 3320  for (;;) Line 3512  for (;;)
3512          pp = eptr;          pp = eptr;
3513          for (i = min; i < max; i++)          for (i = min; i < max; i++)
3514            {            {
3515              pcre_uint32 cc;               /* Faster than pcre_uchar */
3516            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
3517              {              {
3518              SCHECK_PARTIAL();              SCHECK_PARTIAL();
3519              break;              break;
3520              }              }
3521            if (fc != md->lcc[*eptr]) break;            cc = RAWUCHARTEST(eptr);
3522              if (fc != cc && foc != cc) break;
3523            eptr++;            eptr++;
3524            }            }
3525    
# Line 3353  for (;;) Line 3547  for (;;)
3547            SCHECK_PARTIAL();            SCHECK_PARTIAL();
3548            RRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
3549            }            }
3550          if (fc != *eptr++) RRETURN(MATCH_NOMATCH);          if (fc != RAWUCHARINCTEST(eptr)) RRETURN(MATCH_NOMATCH);
3551          }          }
3552    
3553        if (min == max) continue;        if (min == max) continue;
# Line 3370  for (;;) Line 3564  for (;;)
3564              SCHECK_PARTIAL();              SCHECK_PARTIAL();
3565              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
3566              }              }
3567            if (fc != *eptr++) RRETURN(MATCH_NOMATCH);            if (fc != RAWUCHARINCTEST(eptr)) RRETURN(MATCH_NOMATCH);
3568            }            }
3569          /* Control never gets here */          /* Control never gets here */
3570          }          }
# Line 3384  for (;;) Line 3578  for (;;)
3578              SCHECK_PARTIAL();              SCHECK_PARTIAL();
3579              break;              break;
3580              }              }
3581            if (fc != *eptr) break;            if (fc != RAWUCHARTEST(eptr)) break;
3582            eptr++;            eptr++;
3583            }            }
3584          if (possessive) continue;          if (possessive) continue;
# Line 3410  for (;;) Line 3604  for (;;)
3604        SCHECK_PARTIAL();        SCHECK_PARTIAL();
3605        RRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
3606        }        }
3607      ecode++;  #ifdef SUPPORT_UTF
3608      GETCHARINCTEST(c, eptr);      if (utf)
     if (op == OP_NOTI)         /* The caseless case */  
3609        {        {
3610  #ifdef SUPPORT_UTF8        register pcre_uint32 ch, och;
3611        if (c < 256)  
3612  #endif        ecode++;
3613        c = md->lcc[c];        GETCHARINC(ch, ecode);
3614        if (md->lcc[*ecode++] == c) RRETURN(MATCH_NOMATCH);        GETCHARINC(c, eptr);
3615    
3616          if (op == OP_NOT)
3617            {
3618            if (ch == c) RRETURN(MATCH_NOMATCH);
3619            }
3620          else
3621            {
3622    #ifdef SUPPORT_UCP
3623            if (ch > 127)
3624              och = UCD_OTHERCASE(ch);
3625    #else
3626            if (ch > 127)
3627              och = ch;
3628    #endif /* SUPPORT_UCP */
3629            else
3630              och = TABLE_GET(ch, md->fcc, ch);
3631            if (ch == c || och == c) RRETURN(MATCH_NOMATCH);
3632            }
3633        }        }
3634      else    /* Caseful */      else
3635    #endif
3636        {        {
3637        if (*ecode++ == c) RRETURN(MATCH_NOMATCH);        register pcre_uint32 ch = ecode[1];
3638          c = *eptr++;
3639          if (ch == c || (op == OP_NOTI && TABLE_GET(ch, md->fcc, ch) == c))
3640            RRETURN(MATCH_NOMATCH);
3641          ecode += 2;
3642        }        }
3643      break;      break;
3644    
# Line 3436  for (;;) Line 3652  for (;;)
3652      case OP_NOTEXACT:      case OP_NOTEXACT:
3653      case OP_NOTEXACTI:      case OP_NOTEXACTI:
3654      min = max = GET2(ecode, 1);      min = max = GET2(ecode, 1);
3655      ecode += 3;      ecode += 1 + IMM2_SIZE;
3656      goto REPEATNOTCHAR;      goto REPEATNOTCHAR;
3657    
3658      case OP_NOTUPTO:      case OP_NOTUPTO:
# Line 3446  for (;;) Line 3662  for (;;)
3662      min = 0;      min = 0;
3663      max = GET2(ecode, 1);      max = GET2(ecode, 1);
3664      minimize = *ecode == OP_NOTMINUPTO || *ecode == OP_NOTMINUPTOI;      minimize = *ecode == OP_NOTMINUPTO || *ecode == OP_NOTMINUPTOI;
3665      ecode += 3;      ecode += 1 + IMM2_SIZE;
3666      goto REPEATNOTCHAR;      goto REPEATNOTCHAR;
3667    
3668      case OP_NOTPOSSTAR:      case OP_NOTPOSSTAR:
# Line 3478  for (;;) Line 3694  for (;;)
3694      possessive = TRUE;      possessive = TRUE;
3695      min = 0;      min = 0;
3696      max = GET2(ecode, 1);      max = GET2(ecode, 1);
3697      ecode += 3;      ecode += 1 + IMM2_SIZE;
3698      goto REPEATNOTCHAR;      goto REPEATNOTCHAR;
3699    
3700      case OP_NOTSTAR:      case OP_NOTSTAR:
# Line 3502  for (;;) Line 3718  for (;;)
3718      /* Common code for all repeated single-byte matches. */      /* Common code for all repeated single-byte matches. */
3719    
3720      REPEATNOTCHAR:      REPEATNOTCHAR:
3721      fc = *ecode++;      GETCHARINCTEST(fc, ecode);
3722    
3723      /* The code is duplicated for the caseless and caseful cases, for speed,      /* The code is duplicated for the caseless and caseful cases, for speed,
3724      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 3729  for (;;)
3729      characters and work backwards. */      characters and work backwards. */
3730    
3731      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,
3732        max, eptr));        max, (char *)eptr));
3733    
3734      if (op >= OP_NOTSTARI)     /* Caseless */      if (op >= OP_NOTSTARI)     /* Caseless */
3735        {        {
3736        fc = md->lcc[fc];  #ifdef SUPPORT_UTF
3737    #ifdef SUPPORT_UCP
3738          if (utf && fc > 127)
3739            foc = UCD_OTHERCASE(fc);
3740    #else
3741          if (utf && fc > 127)
3742            foc = fc;
3743    #endif /* SUPPORT_UCP */
3744          else
3745    #endif /* SUPPORT_UTF */
3746            foc = TABLE_GET(fc, md->fcc, fc);
3747    
3748  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
3749        /* UTF-8 mode */        if (utf)
       if (utf8)  
3750          {          {
3751          register unsigned int d;          register pcre_uint32 d;
3752          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
3753            {            {
3754            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
# Line 3532  for (;;) Line 3757  for (;;)
3757              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
3758              }              }
3759            GETCHARINC(d, eptr);            GETCHARINC(d, eptr);
3760            if (d < 256) d = md->lcc[d];            if (fc == d || (unsigned int)foc == d) RRETURN(MATCH_NOMATCH);
           if (fc == d) RRETURN(MATCH_NOMATCH);  
3761            }            }
3762          }          }
3763        else        else
3764  #endif  #endif
3765          /* Not UTF mode */
       /* Not UTF-8 mode */  
3766          {          {
3767          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
3768            {            {
# Line 3548  for (;;) Line 3771  for (;;)
3771              SCHECK_PARTIAL();              SCHECK_PARTIAL();
3772              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
3773              }              }
3774            if (fc == md->lcc[*eptr++]) RRETURN(MATCH_NOMATCH);            if (fc == *eptr || foc == *eptr) RRETURN(MATCH_NOMATCH);
3775              eptr++;
3776            }            }
3777          }          }
3778    
# Line 3556  for (;;) Line 3780  for (;;)
3780    
3781        if (minimize)        if (minimize)
3782          {          {
3783  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
3784          /* UTF-8 mode */          if (utf)
         if (utf8)  
3785            {            {
3786            register unsigned int d;            register pcre_uint32 d;
3787            for (fi = min;; fi++)            for (fi = min;; fi++)
3788              {              {
3789              RMATCH(eptr, ecode, offset_top, md, eptrb, RM28);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM28);
# Line 3572  for (;;) Line 3795  for (;;)
3795                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
3796                }                }
3797              GETCHARINC(d, eptr);              GETCHARINC(d, eptr);
3798              if (d < 256) d = md->lcc[d];              if (fc == d || (unsigned int)foc == d) RRETURN(MATCH_NOMATCH);
             if (fc == d) RRETURN(MATCH_NOMATCH);  
3799              }              }
3800            }            }
3801          else          else
3802  #endif  #endif
3803          /* Not UTF-8 mode */          /* Not UTF mode */
3804            {            {
3805            for (fi = min;; fi++)            for (fi = min;; fi++)
3806              {              {
# Line 3590  for (;;) Line 3812  for (;;)
3812                SCHECK_PARTIAL();                SCHECK_PARTIAL();
3813                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
3814                }                }
3815              if (fc == md->lcc[*eptr++]) RRETURN(MATCH_NOMATCH);              if (fc == *eptr || foc == *eptr) RRETURN(MATCH_NOMATCH);
3816                eptr++;
3817              }              }
3818            }            }
3819          /* Control never gets here */          /* Control never gets here */
# Line 3602  for (;;) Line 3825  for (;;)
3825          {          {
3826          pp = eptr;          pp = eptr;
3827    
3828  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
3829          /* UTF-8 mode */          if (utf)
         if (utf8)  
3830            {            {
3831            register unsigned int d;            register pcre_uint32 d;
3832            for (i = min; i < max; i++)            for (i = min; i < max; i++)
3833              {              {
3834              int len = 1;              int len = 1;
# Line 3616  for (;;) Line 3838  for (;;)
3838                break;                break;
3839                }                }
3840              GETCHARLEN(d, eptr, len);              GETCHARLEN(d, eptr, len);
3841              if (d < 256) d = md->lcc[d];              if (fc == d || (unsigned int)foc == d) break;
             if (fc == d) break;  
3842              eptr += len;              eptr += len;
3843              }              }
3844          if (possessive) continue;            if (possessive) continue;
3845          for(;;)            for(;;)
3846              {              {
3847              RMATCH(eptr, ecode, offset_top, md, eptrb, RM30);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM30);
3848              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
# Line 3631  for (;;) Line 3852  for (;;)
3852            }            }
3853          else          else
3854  #endif  #endif
3855          /* Not UTF-8 mode */          /* Not UTF mode */
3856            {            {
3857            for (i = min; i < max; i++)            for (i = min; i < max; i++)
3858              {              {
# Line 3640  for (;;) Line 3861  for (;;)
3861                SCHECK_PARTIAL();                SCHECK_PARTIAL();
3862                break;                break;
3863                }                }
3864              if (fc == md->lcc[*eptr]) break;              if (fc == *eptr || foc == *eptr) break;
3865              eptr++;              eptr++;
3866              }              }
3867            if (possessive) continue;            if (possessive) continue;
# Line 3661  for (;;) Line 3882  for (;;)
3882    
3883      else      else
3884        {        {
3885  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
3886        /* UTF-8 mode */        if (utf)
       if (utf8)  
3887          {          {
3888          register unsigned int d;          register pcre_uint32 d;
3889          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
3890            {            {
3891            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
# Line 3679  for (;;) Line 3899  for (;;)
3899          }          }
3900        else        else
3901  #endif  #endif
3902        /* Not UTF-8 mode */        /* Not UTF mode */
3903          {          {
3904          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
3905            {            {
# Line 3696  for (;;) Line 3916  for (;;)
3916    
3917        if (minimize)        if (minimize)
3918          {          {
3919  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
3920          /* UTF-8 mode */          if (utf)
         if (utf8)  
3921            {            {
3922            register unsigned int d;            register pcre_uint32 d;
3923            for (fi = min;; fi++)            for (fi = min;; fi++)
3924              {              {
3925              RMATCH(eptr, ecode, offset_top, md, eptrb, RM32);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM32);
# Line 3717  for (;;) Line 3936  for (;;)
3936            }            }
3937          else          else
3938  #endif  #endif
3939          /* Not UTF-8 mode */          /* Not UTF mode */
3940            {            {
3941            for (fi = min;; fi++)            for (fi = min;; fi++)
3942              {              {
# Line 3741  for (;;) Line 3960  for (;;)
3960          {          {
3961          pp = eptr;          pp = eptr;
3962    
3963  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
3964          /* UTF-8 mode */          if (utf)
         if (utf8)  
3965            {            {
3966            register unsigned int d;            register pcre_uint32 d;
3967            for (i = min; i < max; i++)            for (i = min; i < max; i++)
3968              {              {
3969              int len = 1;              int len = 1;
# Line 3769  for (;;) Line 3987  for (;;)
3987            }            }
3988          else          else
3989  #endif  #endif
3990          /* Not UTF-8 mode */          /* Not UTF mode */
3991            {            {
3992            for (i = min; i < max; i++)            for (i = min; i < max; i++)
3993              {              {
# Line 3802  for (;;) Line 4020  for (;;)
4020      case OP_TYPEEXACT:      case OP_TYPEEXACT:
4021      min = max = GET2(ecode, 1);      min = max = GET2(ecode, 1);
4022      minimize = TRUE;      minimize = TRUE;
4023      ecode += 3;      ecode += 1 + IMM2_SIZE;
4024      goto REPEATTYPE;      goto REPEATTYPE;
4025    
4026      case OP_TYPEUPTO:      case OP_TYPEUPTO:
# Line 3810  for (;;) Line 4028  for (;;)
4028      min = 0;      min = 0;
4029      max = GET2(ecode, 1);      max = GET2(ecode, 1);
4030      minimize = *ecode == OP_TYPEMINUPTO;      minimize = *ecode == OP_TYPEMINUPTO;
4031      ecode += 3;      ecode += 1 + IMM2_SIZE;
4032      goto REPEATTYPE;      goto REPEATTYPE;
4033    
4034      case OP_TYPEPOSSTAR:      case OP_TYPEPOSSTAR:
# Line 3838  for (;;) Line 4056  for (;;)
4056      possessive = TRUE;      possessive = TRUE;
4057      min = 0;      min = 0;
4058      max = GET2(ecode, 1);      max = GET2(ecode, 1);
4059      ecode += 3;      ecode += 1 + IMM2_SIZE;
4060      goto REPEATTYPE;      goto REPEATTYPE;
4061    
4062      case OP_TYPESTAR:      case OP_TYPESTAR:
# Line 4021  for (;;) Line 4239  for (;;)
4239              }              }
4240            break;            break;
4241    
4242              case PT_CLIST:
4243              for (i = 1; i <= min; i++)
4244                {
4245                const pcre_uint32 *cp;
4246                if (eptr >= md->end_subject)
4247                  {
4248                  SCHECK_PARTIAL();
4249                  RRETURN(MATCH_NOMATCH);
4250                  }
4251                GETCHARINCTEST(c, eptr);
4252                cp = PRIV(ucd_caseless_sets) + prop_value;
4253                for (;;)
4254                  {
4255                  if (c < *cp)
4256                    { if (prop_fail_result) break; else { RRETURN(MATCH_NOMATCH); } }
4257                  if (c == *cp++)
4258                    { if (prop_fail_result) { RRETURN(MATCH_NOMATCH); } else break; }
4259                  }
4260                }
4261              break;
4262    
4263              case PT_UCNC:
4264              for (i = 1; i <= min; i++)
4265                {
4266                if (eptr >= md->end_subject)
4267                  {
4268                  SCHECK_PARTIAL();
4269                  RRETURN(MATCH_NOMATCH);
4270                  }
4271                GETCHARINCTEST(c, eptr);
4272                if ((c == CHAR_DOLLAR_SIGN || c == CHAR_COMMERCIAL_AT ||
4273                     c == CHAR_GRAVE_ACCENT || (c >= 0xa0 && c <= 0xd7ff) ||
4274                     c >= 0xe000) == prop_fail_result)
4275                  RRETURN(MATCH_NOMATCH);
4276                }
4277              break;
4278    
4279            /* This should not occur */            /* This should not occur */
4280    
4281            default:            default:
# Line 4040  for (;;) Line 4295  for (;;)
4295              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4296              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4297              }              }
4298            GETCHARINCTEST(c, eptr);            else
           if (UCD_CATEGORY(c) == ucp_M) RRETURN(MATCH_NOMATCH);  
           while (eptr < md->end_subject)  
4299              {              {
4300              int len = 1;              int lgb, rgb;
4301              if (!utf8) c = *eptr; else { GETCHARLEN(c, eptr, len); }              GETCHARINCTEST(c, eptr);
4302              if (UCD_CATEGORY(c) != ucp_M) break;              lgb = UCD_GRAPHBREAK(c);
4303              eptr += len;             while (eptr < md->end_subject)
4304                  {
4305                  int len = 1;
4306                  if (!utf) c = *eptr; else { GETCHARLEN(c, eptr, len); }
4307                  rgb = UCD_GRAPHBREAK(c);
4308                  if ((PRIV(ucp_gbtable)[lgb] & (1 << rgb)) == 0) break;
4309                  lgb = rgb;
4310                  eptr += len;
4311                  }
4312              }              }
4313              CHECK_PARTIAL();
4314            }            }
4315          }          }
4316    
# Line 4057  for (;;) Line 4319  for (;;)
4319    
4320  /* Handle all other cases when the coding is UTF-8 */  /* Handle all other cases when the coding is UTF-8 */
4321    
4322  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
4323        if (utf8) switch(ctype)        if (utf) switch(ctype)
4324          {          {
4325          case OP_ANY:          case OP_ANY:
4326          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
# Line 4069  for (;;) Line 4331  for (;;)
4331              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4332              }              }
4333            if (IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH);            if (IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH);
4334              if (md->partial != 0 &&
4335                  eptr + 1 >= md->end_subject &&
4336                  NLBLOCK->nltype == NLTYPE_FIXED &&
4337                  NLBLOCK->nllen == 2 &&
4338                  RAWUCHAR(eptr) == NLBLOCK->nl[0])
4339                {
4340                md->hitend = TRUE;
4341                if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
4342                }
4343            eptr++;            eptr++;
4344            while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;            ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);
4345            }            }
4346          break;          break;
4347    
# Line 4083  for (;;) Line 4354  for (;;)
4354              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4355              }              }
4356            eptr++;            eptr++;
4357            while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;            ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);
4358            }            }
4359          break;          break;
4360    
# Line 4105  for (;;) Line 4376  for (;;)
4376              {              {
4377              default: RRETURN(MATCH_NOMATCH);              default: RRETURN(MATCH_NOMATCH);
4378    
4379              case 0x000d:              case CHAR_CR:
4380              if (eptr < md->end_subject && *eptr == 0x0a) eptr++;              if (eptr < md->end_subject && RAWUCHAR(eptr) == CHAR_LF) eptr++;
4381              break;              break;
4382    
4383              case 0x000a:              case CHAR_LF:
4384              break;              break;
4385    
4386              case 0x000b:              case CHAR_VT:
4387              case 0x000c:              case CHAR_FF:
4388              case 0x0085:              case CHAR_NEL:
4389    #ifndef EBCDIC
4390              case 0x2028:              case 0x2028:
4391              case 0x2029:              case 0x2029:
4392    #endif  /* Not EBCDIC */
4393              if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);              if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);
4394              break;              break;
4395              }              }
# Line 4134  for (;;) Line 4407  for (;;)
4407            GETCHARINC(c, eptr);            GETCHARINC(c, eptr);
4408            switch(c)            switch(c)
4409              {              {
4410                HSPACE_CASES: RRETURN(MATCH_NOMATCH);  /* Byte and multibyte cases */
4411              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);  
4412              }              }
4413            }            }
4414          break;          break;
# Line 4170  for (;;) Line 4424  for (;;)
4424            GETCHARINC(c, eptr);            GETCHARINC(c, eptr);
4425            switch(c)            switch(c)
4426              {              {
4427                HSPACE_CASES: break;  /* Byte and multibyte cases */
4428              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;  
4429              }              }
4430            }            }
4431          break;          break;
# Line 4206  for (;;) Line 4441  for (;;)
4441            GETCHARINC(c, eptr);            GETCHARINC(c, eptr);
4442            switch(c)            switch(c)
4443              {              {
4444                VSPACE_CASES: RRETURN(MATCH_NOMATCH);
4445              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);  
4446              }              }
4447            }            }
4448          break;          break;
# Line 4230  for (;;) Line 4458  for (;;)
4458            GETCHARINC(c, eptr);            GETCHARINC(c, eptr);
4459            switch(c)            switch(c)
4460              {              {
4461                VSPACE_CASES: break;
4462              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;  
4463              }              }
4464            }            }
4465          break;          break;
# Line 4260  for (;;) Line 4481  for (;;)
4481          case OP_DIGIT:          case OP_DIGIT:
4482          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
4483            {            {
4484              pcre_uint32 cc;
4485            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
4486              {              {
4487              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4488              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4489              }              }
4490            if (*eptr >= 128 || (md->ctypes[*eptr++] & ctype_digit) == 0)            cc = RAWUCHAR(eptr);
4491              if (cc >= 128 || (md->ctypes[cc] & ctype_digit) == 0)
4492              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4493              eptr++;
4494            /* 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 */
4495            }            }
4496          break;          break;
# Line 4274  for (;;) Line 4498  for (;;)
4498          case OP_NOT_WHITESPACE:          case OP_NOT_WHITESPACE:
4499          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
4500            {            {
4501              pcre_uint32 cc;
4502            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
4503              {              {
4504              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4505              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4506              }              }
4507            if (*eptr < 128 && (md->ctypes[*eptr] & ctype_space) != 0)            cc = RAWUCHAR(eptr);
4508              if (cc < 128 && (md->ctypes[cc] & ctype_space) != 0)
4509              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4510            while (++eptr < md->end_subject && (*eptr & 0xc0) == 0x80);            eptr++;
4511              ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);
4512            }            }
4513          break;          break;
4514    
4515          case OP_WHITESPACE:          case OP_WHITESPACE:
4516          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
4517            {            {
4518              pcre_uint32 cc;
4519            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
4520              {              {
4521              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4522              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4523              }              }
4524            if (*eptr >= 128 || (md->ctypes[*eptr++] & ctype_space) == 0)            cc = RAWUCHAR(eptr);
4525              if (cc >= 128 || (md->ctypes[cc] & ctype_space) == 0)
4526              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4527              eptr++;
4528            /* 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 */
4529            }            }
4530          break;          break;
# Line 4302  for (;;) Line 4532  for (;;)
4532          case OP_NOT_WORDCHAR:          case OP_NOT_WORDCHAR:
4533          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
4534            {            {
4535              pcre_uint32 cc;
4536            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
4537              {              {
4538              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4539              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4540              }              }
4541            if (*eptr < 128 && (md->ctypes[*eptr] & ctype_word) != 0)            cc = RAWUCHAR(eptr);
4542              if (cc < 128 && (md->ctypes[cc] & ctype_word) != 0)
4543              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4544            while (++eptr < md->end_subject && (*eptr & 0xc0) == 0x80);            eptr++;
4545              ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);
4546            }            }
4547          break;          break;
4548    
4549          case OP_WORDCHAR:          case OP_WORDCHAR:
4550          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
4551            {            {
4552              pcre_uint32 cc;
4553            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
4554              {              {
4555              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4556              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4557              }              }
4558            if (*eptr >= 128 || (md->ctypes[*eptr++] & ctype_word) == 0)            cc = RAWUCHAR(eptr);
4559              if (cc >= 128 || (md->ctypes[cc] & ctype_word) == 0)
4560              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4561              eptr++;
4562            /* 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 */
4563            }            }
4564          break;          break;
# Line 4332  for (;;) Line 4568  for (;;)
4568          }  /* End switch(ctype) */          }  /* End switch(ctype) */
4569    
4570        else        else
4571  #endif     /* SUPPORT_UTF8 */  #endif     /* SUPPORT_UTF */
4572    
4573        /* 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
4574        than OP_PROP and OP_NOTPROP. */        than OP_PROP and OP_NOTPROP. */
# Line 4348  for (;;) Line 4584  for (;;)
4584              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4585              }              }
4586            if (IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH);            if (IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH);
4587              if (md->partial != 0 &&
4588                  eptr + 1 >= md->end_subject &&
4589                  NLBLOCK->nltype == NLTYPE_FIXED &&
4590                  NLBLOCK->nllen == 2 &&
4591                  *eptr == NLBLOCK->nl[0])
4592                {
4593                md->hitend = TRUE;
4594                if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
4595                }
4596            eptr++;            eptr++;
4597            }            }
4598          break;          break;
# Line 4382  for (;;) Line 4627  for (;;)
4627              {              {
4628              default: RRETURN(MATCH_NOMATCH);              default: RRETURN(MATCH_NOMATCH);
4629    
4630              case 0x000d:              case CHAR_CR:
4631              if (eptr < md->end_subject && *eptr == 0x0a) eptr++;              if (eptr < md->end_subject && *eptr == CHAR_LF) eptr++;
4632              break;              break;
4633    
4634              case 0x000a:              case CHAR_LF:
4635              break;              break;
4636    
4637              case 0x000b:              case CHAR_VT:
4638              case 0x000c:              case CHAR_FF:
4639              case 0x0085:              case CHAR_NEL:
4640    #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
4641                case 0x2028:
4642                case 0x2029:
4643    #endif
4644              if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);              if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);
4645              break;              break;
4646              }              }
# Line 4409  for (;;) Line 4658  for (;;)
4658            switch(*eptr++)            switch(*eptr++)
4659              {              {
4660              default: break;              default: break;
4661              case 0x09:      /* HT */              HSPACE_BYTE_CASES:
4662              case 0x20:      /* SPACE */  #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
4663              case 0xa0:      /* NBSP */              HSPACE_MULTIBYTE_CASES:
4664    #endif
4665              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4666              }              }
4667            }            }
# Line 4428  for (;;) Line 4678  for (;;)
4678            switch(*eptr++)            switch(*eptr++)
4679              {              {
4680              default: RRETURN(MATCH_NOMATCH);              default: RRETURN(MATCH_NOMATCH);
4681              case 0x09:      /* HT */              HSPACE_BYTE_CASES:
4682              case 0x20:      /* SPACE */  #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
4683              case 0xa0:      /* NBSP */              HSPACE_MULTIBYTE_CASES:
4684    #endif
4685              break;              break;
4686              }              }
4687            }            }
# Line 4446  for (;;) Line 4697  for (;;)
4697              }              }
4698            switch(*eptr++)            switch(*eptr++)
4699              {              {
4700              default: break;              VSPACE_BYTE_CASES:
4701              case 0x0a:      /* LF */  #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
4702              case 0x0b:      /* VT */              VSPACE_MULTIBYTE_CASES:
4703              case 0x0c:      /* FF */  #endif
             case 0x0d:      /* CR */  
             case 0x85:      /* NEL */  
4704              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4705                default: break;
4706              }              }
4707            }            }
4708          break;          break;
# Line 4468  for (;;) Line 4718  for (;;)
4718            switch(*eptr++)            switch(*eptr++)
4719              {              {
4720              default: RRETURN(MATCH_NOMATCH);              default: RRETURN(MATCH_NOMATCH);
4721              case 0x0a:      /* LF */              VSPACE_BYTE_CASES:
4722              case 0x0b:      /* VT */  #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
4723              case 0x0c:      /* FF */              VSPACE_MULTIBYTE_CASES:
4724              case 0x0d:      /* CR */  #endif
             case 0x85:      /* NEL */  
4725              break;              break;
4726              }              }
4727            }            }
# Line 4486  for (;;) Line 4735  for (;;)
4735              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4736              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4737              }              }
4738            if ((md->ctypes[*eptr++] & ctype_digit) != 0) RRETURN(MATCH_NOMATCH);            if (MAX_255(*eptr) && (md->ctypes[*eptr] & ctype_digit) != 0)
4739                RRETURN(MATCH_NOMATCH);
4740              eptr++;
4741            }            }
4742          break;          break;
4743    
# Line 4498  for (;;) Line 4749  for (;;)
4749              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4750              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4751              }              }
4752            if ((md->ctypes[*eptr++] & ctype_digit) == 0) RRETURN(MATCH_NOMATCH);            if (!MAX_255(*eptr) || (md->ctypes[*eptr] & ctype_digit) == 0)
4753                RRETURN(MATCH_NOMATCH);
4754              eptr++;
4755            }            }
4756          break;          break;
4757    
# Line 4510  for (;;) Line 4763  for (;;)
4763              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4764              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4765              }              }
4766            if ((md->ctypes[*eptr++] & ctype_space) != 0) RRETURN(MATCH_NOMATCH);            if (MAX_255(*eptr) && (md->ctypes[*eptr] & ctype_space) != 0)
4767                RRETURN(MATCH_NOMATCH);
4768              eptr++;
4769            }            }
4770          break;          break;
4771    
# Line 4522  for (;;) Line 4777  for (;;)
4777              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4778              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4779              }              }
4780            if ((md->ctypes[*eptr++] & ctype_space) == 0) RRETURN(MATCH_NOMATCH);            if (!MAX_255(*eptr) || (md->ctypes[*eptr] & ctype_space) == 0)
4781                RRETURN(MATCH_NOMATCH);
4782              eptr++;
4783            }            }
4784          break;          break;
4785    
# Line 4534  for (;;) Line 4791  for (;;)
4791              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4792              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4793              }              }
4794            if ((md->ctypes[*eptr++] & ctype_word) != 0)            if (MAX_255(*eptr) && (md->ctypes[*eptr] & ctype_word) != 0)
4795              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4796              eptr++;
4797            }            }
4798          break;          break;
4799    
# Line 4547  for (;;) Line 4805  for (;;)
4805              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4806              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4807              }              }
4808            if ((md->ctypes[*eptr++] & ctype_word) == 0)            if (!MAX_255(*eptr) || (md->ctypes[*eptr] & ctype_word) == 0)
4809              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4810              eptr++;
4811            }            }
4812          break;          break;
4813    
# Line 4739  for (;;) Line 4998  for (;;)
4998              }              }
4999            /* Control never gets here */            /* Control never gets here */
5000    
5001            /* This should never occur */            case PT_CLIST:
5002              for (fi = min;; fi++)
5003                {
5004                const pcre_uint32 *cp;
5005                RMATCH(eptr, ecode, offset_top, md, eptrb, RM67);
5006                if (rrc != MATCH_NOMATCH) RRETURN(rrc);
5007                if (fi >= max) RRETURN(MATCH_NOMATCH);
5008                if (eptr >= md->end_subject)
5009                  {
5010                  SCHECK_PARTIAL();
5011                  RRETURN(MATCH_NOMATCH);
5012                  }
5013                GETCHARINCTEST(c, eptr);
5014                cp = PRIV(ucd_caseless_sets) + prop_value;
5015                for (;;)
5016                  {
5017                  if (c < *cp)
5018                    { if (prop_fail_result) break; else { RRETURN(MATCH_NOMATCH); } }
5019                  if (c == *cp++)
5020                    { if (prop_fail_result) { RRETURN(MATCH_NOMATCH); } else break; }
5021                  }
5022                }
5023              /* Control never gets here */
5024    
5025              case PT_UCNC:
5026              for (fi = min;; fi++)
5027                {
5028                RMATCH(eptr, ecode, offset_top, md, eptrb, RM68);
5029                if (rrc != MATCH_NOMATCH) RRETURN(rrc);
5030                if (fi >= max) RRETURN(MATCH_NOMATCH);
5031                if (eptr >= md->end_subject)
5032                  {
5033                  SCHECK_PARTIAL();
5034                  RRETURN(MATCH_NOMATCH);
5035                  }
5036                GETCHARINCTEST(c, eptr);
5037                if ((c == CHAR_DOLLAR_SIGN || c == CHAR_COMMERCIAL_AT ||
5038                     c == CHAR_GRAVE_ACCENT || (c >= 0xa0 && c <= 0xd7ff) ||
5039                     c >= 0xe000) == prop_fail_result)
5040                  RRETURN(MATCH_NOMATCH);
5041                }
5042              /* Control never gets here */
5043    
5044              /* This should never occur */
5045            default:            default:
5046            RRETURN(PCRE_ERROR_INTERNAL);            RRETURN(PCRE_ERROR_INTERNAL);
5047            }            }
# Line 4761  for (;;) Line 5062  for (;;)
5062              SCHECK_PARTIAL();              SCHECK_PARTIAL();
5063              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
5064              }              }
5065            GETCHARINCTEST(c, eptr);            else
           if (UCD_CATEGORY(c) == ucp_M) RRETURN(MATCH_NOMATCH);  
           while (eptr < md->end_subject)  
5066              {              {
5067              int len = 1;              int lgb, rgb;
5068              if (!utf8) c = *eptr; else { GETCHARLEN(c, eptr, len); }              GETCHARINCTEST(c, eptr);
5069              if (UCD_CATEGORY(c) != ucp_M) break;              lgb = UCD_GRAPHBREAK(c);
5070              eptr += len;              while (eptr < md->end_subject)
5071                  {
5072                  int len = 1;
5073                  if (!utf) c = *eptr; else { GETCHARLEN(c, eptr, len); }
5074                  rgb = UCD_GRAPHBREAK(c);
5075                  if ((PRIV(ucp_gbtable)[lgb] & (1 << rgb)) == 0) break;
5076                  lgb = rgb;
5077                  eptr += len;
5078                  }
5079              }              }
5080              CHECK_PARTIAL();
5081            }            }
5082          }          }
5083        else        else
5084  #endif     /* SUPPORT_UCP */  #endif     /* SUPPORT_UCP */
5085    
5086  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
5087        /* UTF-8 mode */        if (utf)
       if (utf8)  
5088          {          {
5089          for (fi = min;; fi++)          for (fi = min;; fi++)
5090            {            {
# Line 4794  for (;;) Line 5101  for (;;)
5101            GETCHARINC(c, eptr);            GETCHARINC(c, eptr);
5102            switch(ctype)            switch(ctype)
5103              {              {
5104              case OP_ANY:        /* This is the non-NL case */              case OP_ANY:               /* This is the non-NL case */
5105                if (md->partial != 0 &&    /* Take care with CRLF partial */
5106                    eptr >= md->end_subject &&
5107                    NLBLOCK->nltype == NLTYPE_FIXED &&
5108                    NLBLOCK->nllen == 2 &&
5109                    c == NLBLOCK->nl[0])
5110                  {
5111                  md->hitend = TRUE;
5112                  if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
5113                  }
5114                break;
5115    
5116              case OP_ALLANY:              case OP_ALLANY:
5117              case OP_ANYBYTE:              case OP_ANYBYTE:
5118              break;              break;
# Line 4803  for (;;) Line 5121  for (;;)
5121              switch(c)              switch(c)
5122                {                {
5123                default: RRETURN(MATCH_NOMATCH);                default: RRETURN(MATCH_NOMATCH);
5124                case 0x000d:                case CHAR_CR:
5125                if (eptr < md->end_subject && *eptr == 0x0a) eptr++;                if (eptr < md->end_subject && RAWUCHAR(eptr) == CHAR_LF) eptr++;
5126                break;                break;
5127                case 0x000a:  
5128                  case CHAR_LF:
5129                break;                break;
5130    
5131                case 0x000b:                case CHAR_VT:
5132                case 0x000c:                case CHAR_FF:
5133                case 0x0085:                case CHAR_NEL:
5134    #ifndef EBCDIC
5135                case 0x2028:                case 0x2028:
5136                case 0x2029:                case 0x2029:
5137    #endif  /* Not EBCDIC */
5138                if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);                if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);
5139                break;                break;
5140                }                }
# Line 4822  for (;;) Line 5143  for (;;)
5143              case OP_NOT_HSPACE:              case OP_NOT_HSPACE:
5144              switch(c)              switch(c)
5145                {                {
5146                  HSPACE_CASES: RRETURN(MATCH_NOMATCH);
5147                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);  
5148                }                }
5149              break;              break;
5150    
5151              case OP_HSPACE:              case OP_HSPACE:
5152              switch(c)              switch(c)
5153                {                {
5154                  HSPACE_CASES: break;
5155                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;  
5156                }                }
5157              break;              break;
5158    
5159              case OP_NOT_VSPACE:              case OP_NOT_VSPACE:
5160              switch(c)              switch(c)
5161                {                {
5162                  VSPACE_CASES: RRETURN(MATCH_NOMATCH);
5163                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);  
5164                }                }
5165              break;              break;
5166    
5167              case OP_VSPACE:              case OP_VSPACE:
5168              switch(c)              switch(c)
5169                {                {
5170                  VSPACE_CASES: break;
5171                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;  
5172                }                }
5173              break;              break;
5174    
# Line 4919  for (;;) Line 5188  for (;;)
5188              break;              break;
5189    
5190              case OP_WHITESPACE:              case OP_WHITESPACE:
5191              if  (c >= 256 || (md->ctypes[c] & ctype_space) == 0)              if (c >= 256 || (md->ctypes[c] & ctype_space) == 0)
5192                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
5193              break;              break;
5194    
# Line 4940  for (;;) Line 5209  for (;;)
5209          }          }
5210        else        else
5211  #endif  #endif
5212        /* Not UTF-8 mode */        /* Not UTF mode */
5213          {          {
5214          for (fi = min;; fi++)          for (fi = min;; fi++)
5215            {            {
# Line 4957  for (;;) Line 5226  for (;;)
5226            c = *eptr++;            c = *eptr++;
5227            switch(ctype)            switch(ctype)
5228              {              {
5229              case OP_ANY:     /* This is the non-NL case */              case OP_ANY:               /* This is the non-NL case */
5230                if (md->partial != 0 &&    /* Take care with CRLF partial */
5231                    eptr >= md->end_subject &&
5232                    NLBLOCK->nltype == NLTYPE_FIXED &&
5233                    NLBLOCK->nllen == 2 &&
5234                    c == NLBLOCK->nl[0])
5235                  {
5236                  md->hitend = TRUE;
5237                  if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
5238                  }
5239                break;
5240    
5241              case OP_ALLANY:              case OP_ALLANY:
5242              case OP_ANYBYTE:              case OP_ANYBYTE:
5243              break;              break;
# Line 4966  for (;;) Line 5246  for (;;)
5246              switch(c)              switch(c)
5247                {                {
5248                default: RRETURN(MATCH_NOMATCH);                default: RRETURN(MATCH_NOMATCH);
5249                case 0x000d:                case CHAR_CR:
5250                if (eptr < md->end_subject && *eptr == 0x0a) eptr++;                if (eptr < md->end_subject && *eptr == CHAR_LF) eptr++;
5251                break;                break;
5252    
5253                case 0x000a:                case CHAR_LF:
5254                break;                break;
5255    
5256                case 0x000b:                case CHAR_VT:
5257                case 0x000c:                case CHAR_FF:
5258                case 0x0085:                case CHAR_NEL:
5259    #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
5260                  case 0x2028:
5261                  case 0x2029:
5262    #endif
5263                if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);                if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);
5264                break;                break;
5265                }                }
# Line 4985  for (;;) Line 5269  for (;;)
5269              switch(c)              switch(c)
5270                {                {
5271                default: break;                default: break;
5272                case 0x09:      /* HT */                HSPACE_BYTE_CASES:
5273                case 0x20:      /* SPACE */  #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
5274                case 0xa0:      /* NBSP */                HSPACE_MULTIBYTE_CASES:
5275    #endif
5276                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
5277                }                }
5278              break;              break;
# Line 4996  for (;;) Line 5281  for (;;)
5281              switch(c)              switch(c)
5282                {                {
5283                default: RRETURN(MATCH_NOMATCH);                default: RRETURN(MATCH_NOMATCH);
5284                case 0x09:      /* HT */                HSPACE_BYTE_CASES:
5285                case 0x20:      /* SPACE */  #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
5286                case 0xa0:      /* NBSP */                HSPACE_MULTIBYTE_CASES:
5287    #endif
5288                break;                break;
5289                }                }
5290              break;              break;
# Line 5007  for (;;) Line 5293  for (;;)
5293              switch(c)              switch(c)
5294                {                {
5295                default: break;                default: break;
5296                case 0x0a:      /* LF */                VSPACE_BYTE_CASES:
5297                case 0x0b:      /* VT */  #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
5298                case 0x0c:      /* FF */                VSPACE_MULTIBYTE_CASES:
5299                case 0x0d:      /* CR */  #endif
               case 0x85:      /* NEL */  
5300                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
5301                }                }
5302              break;              break;
# Line 5020  for (;;) Line 5305  for (;;)
5305              switch(c)              switch(c)
5306                {                {
5307                default: RRETURN(MATCH_NOMATCH);                default: RRETURN(MATCH_NOMATCH);
5308                case 0x0a:      /* LF */                VSPACE_BYTE_CASES:
5309                case 0x0b:      /* VT */  #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
5310                case 0x0c:      /* FF */                VSPACE_MULTIBYTE_CASES:
5311                case 0x0d:      /* CR */  #endif
               case 0x85:      /* NEL */  
5312                break;                break;
5313                }                }
5314              break;              break;
5315    
5316              case OP_NOT_DIGIT:              case OP_NOT_DIGIT:
5317              if ((md->ctypes[c] & ctype_digit) != 0) RRETURN(MATCH_NOMATCH);              if (MAX_255(c) && (md->ctypes[c] & ctype_digit) != 0) RRETURN(MATCH_NOMATCH);
5318              break;              break;
5319    
5320              case OP_DIGIT:              case OP_DIGIT:
5321              if ((md->ctypes[c] & ctype_digit) == 0) RRETURN(MATCH_NOMATCH);              if (!MAX_255(c) || (md->ctypes[c] & ctype_digit) == 0) RRETURN(MATCH_NOMATCH);
5322              break;              break;
5323    
5324              case OP_NOT_WHITESPACE:              case OP_NOT_WHITESPACE:
5325              if ((md->ctypes[c] & ctype_space) != 0) RRETURN(MATCH_NOMATCH);              if (MAX_255(c) && (md->ctypes[c] & ctype_space) != 0) RRETURN(MATCH_NOMATCH);
5326              break;              break;
5327    
5328              case OP_WHITESPACE:              case OP_WHITESPACE:
5329              if  ((md->ctypes[c] & ctype_space) == 0) RRETURN(MATCH_NOMATCH);              if (!MAX_255(c) || (md->ctypes[c] & ctype_space) == 0) RRETURN(MATCH_NOMATCH);
5330              break;              break;
5331    
5332              case OP_NOT_WORDCHAR:              case OP_NOT_WORDCHAR:
5333              if ((md->ctypes[c] & ctype_word) != 0) RRETURN(MATCH_NOMATCH);              if (MAX_255(c) && (md->ctypes[c] & ctype_word) != 0) RRETURN(MATCH_NOMATCH);
5334              break;              break;
5335    
5336              case OP_WORDCHAR:              case OP_WORDCHAR:
5337              if ((md->ctypes[c] & ctype_word) == 0) RRETURN(MATCH_NOMATCH);              if (!MAX_255(c) || (md->ctypes[c] & ctype_word) == 0) RRETURN(MATCH_NOMATCH);
5338              break;              break;
5339    
5340              default:              default:
# Line 5227  for (;;) Line 5511  for (;;)
5511              }              }
5512            break;            break;
5513    
5514              case PT_CLIST:
5515              for (i = min; i < max; i++)
5516                {
5517                const pcre_uint32 *cp;
5518                int len = 1;
5519                if (eptr >= md->end_subject)
5520                  {
5521                  SCHECK_PARTIAL();
5522                  break;
5523                  }
5524                GETCHARLENTEST(c, eptr, len);
5525                cp = PRIV(ucd_caseless_sets) + prop_value;
5526                for (;;)
5527                  {
5528                  if (c < *cp)
5529                    { if (prop_fail_result) break; else goto GOT_MAX; }
5530                  if (c == *cp++)
5531                    { if (prop_fail_result) goto GOT_MAX; else break; }
5532                  }
5533                eptr += len;
5534                }
5535              GOT_MAX:
5536              break;
5537    
5538              case PT_UCNC:
5539              for (i = min; i < max; i++)
5540                {
5541                int len = 1;
5542                if (eptr >= md->end_subject)
5543                  {
5544                  SCHECK_PARTIAL();
5545                  break;
5546                  }
5547                GETCHARLENTEST(c, eptr, len);
5548                if ((c == CHAR_DOLLAR_SIGN || c == CHAR_COMMERCIAL_AT ||
5549                     c == CHAR_GRAVE_ACCENT || (c >= 0xa0 && c <= 0xd7ff) ||
5550                     c >= 0xe000) == prop_fail_result)
5551                  break;
5552                eptr += len;
5553                }
5554              break;
5555    
5556            default:            default:
5557            RRETURN(PCRE_ERROR_INTERNAL);            RRETURN(PCRE_ERROR_INTERNAL);
5558            }            }
# Line 5239  for (;;) Line 5565  for (;;)
5565            RMATCH(eptr, ecode, offset_top, md, eptrb, RM44);            RMATCH(eptr, ecode, offset_top, md, eptrb, RM44);
5566            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
5567            if (eptr-- == pp) break;        /* Stop if tried at original pos */            if (eptr-- == pp) break;        /* Stop if tried at original pos */
5568            if (utf8) BACKCHAR(eptr);            if (utf) BACKCHAR(eptr);
5569            }            }
5570          }          }
5571    
# Line 5250  for (;;) Line 5576  for (;;)
5576          {          {
5577          for (i = min; i < max; i++)          for (i = min; i < max; i++)
5578            {            {
           int len = 1;  
5579            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
5580              {              {
5581              SCHECK_PARTIAL();              SCHECK_PARTIAL();
5582              break;              break;
5583              }              }
5584            if (!utf8) c = *eptr; else { GETCHARLEN(c, eptr, len); }            else
           if (UCD_CATEGORY(c) == ucp_M) break;  
           eptr += len;  
           while (eptr < md->end_subject)  
5585              {              {
5586              len = 1;              int lgb, rgb;
5587              if (!utf8) c = *eptr; else { GETCHARLEN(c, eptr, len); }              GETCHARINCTEST(c, eptr);
5588              if (UCD_CATEGORY(c) != ucp_M) break;              lgb = UCD_GRAPHBREAK(c);
5589              eptr += len;              while (eptr < md->end_subject)
5590                  {
5591                  int len = 1;
5592                  if (!utf) c = *eptr; else { GETCHARLEN(c, eptr, len); }
5593                  rgb = UCD_GRAPHBREAK(c);
5594                  if ((PRIV(ucp_gbtable)[lgb] & (1 << rgb)) == 0) break;
5595                  lgb = rgb;
5596                  eptr += len;
5597                  }
5598              }              }
5599              CHECK_PARTIAL();
5600            }            }
5601    
5602          /* eptr is now past the end of the maximum run */          /* eptr is now past the end of the maximum run */
# Line 5279  for (;;) Line 5610  for (;;)
5610            if (eptr-- == pp) break;        /* Stop if tried at original pos */            if (eptr-- == pp) break;        /* Stop if tried at original pos */
5611            for (;;)                        /* Move back over one extended */            for (;;)                        /* Move back over one extended */
5612              {              {
5613              if (!utf8) c = *eptr; else              if (!utf) c = *eptr; else
5614                {                {
5615                BACKCHAR(eptr);                BACKCHAR(eptr);
5616                GETCHAR(c, eptr);                GETCHAR(c, eptr);
# Line 5293  for (;;) Line 5624  for (;;)
5624        else        else
5625  #endif   /* SUPPORT_UCP */  #endif   /* SUPPORT_UCP */
5626    
5627  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
5628        /* UTF-8 mode */        if (utf)
   
       if (utf8)  
5629          {          {
5630          switch(ctype)          switch(ctype)
5631            {            {
# Line 5311  for (;;) Line 5640  for (;;)
5640                  break;                  break;
5641                  }                  }
5642                if (IS_NEWLINE(eptr)) break;                if (IS_NEWLINE(eptr)) break;
5643                  if (md->partial != 0 &&    /* Take care with CRLF partial */
5644                      eptr + 1 >= md->end_subject &&
5645                      NLBLOCK->nltype == NLTYPE_FIXED &&
5646                      NLBLOCK->nllen == 2 &&
5647                      RAWUCHAR(eptr) == NLBLOCK->nl[0])
5648                    {
5649                    md->hitend = TRUE;
5650                    if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
5651                    }
5652                eptr++;                eptr++;
5653                while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;                ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);
5654                }                }
5655              }              }
5656    
# Line 5328  for (;;) Line 5666  for (;;)
5666                  break;                  break;
5667                  }                  }
5668                if (IS_NEWLINE(eptr)) break;                if (IS_NEWLINE(eptr)) break;
5669                  if (md->partial != 0 &&    /* Take care with CRLF partial */
5670                      eptr + 1 >= md->end_subject &&
5671                      NLBLOCK->nltype == NLTYPE_FIXED &&
5672                      NLBLOCK->nllen == 2 &&
5673                      RAWUCHAR(eptr) == NLBLOCK->nl[0])
5674                    {
5675                    md->hitend = TRUE;
5676                    if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
5677                    }
5678                eptr++;                eptr++;
5679                while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;                ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);
5680                }                }
5681              }              }
5682            break;            break;
# Line 5345  for (;;) Line 5692  for (;;)
5692                  break;                  break;
5693                  }                  }
5694                eptr++;                eptr++;
5695                while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;                ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);
5696                }                }
5697              }              }
5698            else            else
# Line 5377  for (;;) Line 5724  for (;;)
5724                break;                break;
5725                }                }
5726              GETCHARLEN(c, eptr, len);              GETCHARLEN(c, eptr, len);
5727              if (c == 0x000d)              if (c == CHAR_CR)
5728                {                {
5729                if (++eptr >= md->end_subject) break;                if (++eptr >= md->end_subject) break;
5730                if (*eptr == 0x000a) eptr++;                if (RAWUCHAR(eptr) == CHAR_LF) eptr++;
5731                }                }
5732              else              else
5733                {                {
5734                if (c != 0x000a &&                if (c != CHAR_LF &&
5735                    (md->bsr_anycrlf ||                    (md->bsr_anycrlf ||
5736                     (c != 0x000b && c != 0x000c &&                     (c != CHAR_VT && c != CHAR_FF && c != CHAR_NEL
5737                      c != 0x0085 && c != 0x2028 && c != 0x2029)))  #ifndef EBCDIC
5738                        && c != 0x2028 && c != 0x2029
5739    #endif  /* Not EBCDIC */
5740                        )))
5741                  break;                  break;
5742                eptr += len;                eptr += len;
5743                }                }
# Line 5408  for (;;) Line 5758  for (;;)
5758              GETCHARLEN(c, eptr, len);              GETCHARLEN(c, eptr, len);
5759              switch(c)              switch(c)
5760                {                {
5761                  HSPACE_CASES: gotspace = TRUE; break;
5762                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;  
5763                }                }
5764              if (gotspace == (ctype == OP_NOT_HSPACE)) break;              if (gotspace == (ctype == OP_NOT_HSPACE)) break;
5765              eptr += len;              eptr += len;
# Line 5450  for (;;) Line 5780  for (;;)
5780              GETCHARLEN(c, eptr, len);              GETCHARLEN(c, eptr, len);
5781              switch(c)              switch(c)
5782                {                {
5783                  VSPACE_CASES: gotspace = TRUE; break;
5784                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;  
5785                }                }
5786              if (gotspace == (ctype == OP_NOT_VSPACE)) break;              if (gotspace == (ctype == OP_NOT_VSPACE)) break;
5787              eptr += len;              eptr += len;
# Line 5573  for (;;) Line 5895  for (;;)
5895            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
5896            if (eptr-- == pp) break;        /* Stop if tried at original pos */            if (eptr-- == pp) break;        /* Stop if tried at original pos */
5897            BACKCHAR(eptr);            BACKCHAR(eptr);
5898            if (ctype == OP_ANYNL && eptr > pp  && *eptr == '\n' &&            if (ctype == OP_ANYNL && eptr > pp  && RAWUCHAR(eptr) == CHAR_NL &&
5899                eptr[-1] == '\r') eptr--;                RAWUCHAR(eptr - 1) == CHAR_CR) eptr--;
5900            }            }
5901          }          }
5902        else        else
5903  #endif  /* SUPPORT_UTF8 */  #endif  /* SUPPORT_UTF */
5904          /* Not UTF mode */
       /* Not UTF-8 mode */  
5905          {          {
5906          switch(ctype)          switch(ctype)
5907            {            {
# Line 5593  for (;;) Line 5914  for (;;)
5914                break;                break;
5915                }                }
5916              if (IS_NEWLINE(eptr)) break;              if (IS_NEWLINE(eptr)) break;
5917                if (md->partial != 0 &&    /* Take care with CRLF partial */
5918                    eptr + 1 >= md->end_subject &&
5919                    NLBLOCK->nltype == NLTYPE_FIXED &&
5920                    NLBLOCK->nllen == 2 &&
5921