/[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 1302 by ph10, Wed Mar 27 11:13:36 2013 UTC
# Line 6  Line 6 
6  and semantics are as close as possible to those of the Perl 5 language.  and semantics are as close as possible to those of the Perl 5 language.
7    
8                         Written by Philip Hazel                         Written by Philip Hazel
9             Copyright (c) 1997-2011 University of Cambridge             Copyright (c) 1997-2013 University of Cambridge
10    
11  -----------------------------------------------------------------------------  -----------------------------------------------------------------------------
12  Redistribution and use in source and binary forms, with or without  Redistribution and use in source and binary forms, with or without
# Line 37  POSSIBILITY OF SUCH DAMAGE. Line 37  POSSIBILITY OF SUCH DAMAGE.
37  -----------------------------------------------------------------------------  -----------------------------------------------------------------------------
38  */  */
39    
   
40  /* This module contains pcre_exec(), the externally visible function that does  /* This module contains pcre_exec(), the externally visible function that does
41  pattern matching using an NFA algorithm, trying to mimic Perl as closely as  pattern matching using an NFA algorithm, trying to mimic Perl as closely as
42  possible. There are also some static supporting functions. */  possible. There are also some static supporting functions. */
# Line 57  possible. There are also some static sup Line 56  possible. There are also some static sup
56  #undef min  #undef min
57  #undef max  #undef max
58    
59    /* The md->capture_last field uses the lower 16 bits for the last captured
60    substring (which can never be greater than 65535) and a bit in the top half
61    to mean "capture vector overflowed". This odd way of doing things was
62    implemented when it was realized that preserving and restoring the overflow bit
63    whenever the last capture number was saved/restored made for a neater
64    interface, and doing it this way saved on (a) another variable, which would
65    have increased the stack frame size (a big NO-NO in PCRE) and (b) another
66    separate set of save/restore instructions. The following defines are used in
67    implementing this. */
68    
69    #define CAPLMASK    0x0000ffff    /* The bits used for last_capture */
70    #define OVFLMASK    0xffff0000    /* The bits used for the overflow flag */
71    #define OVFLBIT     0x00010000    /* The bit that is set for overflow */
72    
73  /* Values for setting in md->match_function_type to indicate two special types  /* Values for setting in md->match_function_type to indicate two special types
74  of call to match(). We do it this way to save on using another stack variable,  of call to match(). We do it this way to save on using another stack variable,
75  as stack usage is to be discouraged. */  as stack usage is to be discouraged. */
# Line 74  defined PCRE_ERROR_xxx codes, which are Line 87  defined PCRE_ERROR_xxx codes, which are
87  negative to avoid the external error codes. */  negative to avoid the external error codes. */
88    
89  #define MATCH_ACCEPT       (-999)  #define MATCH_ACCEPT       (-999)
90  #define MATCH_COMMIT       (-998)  #define MATCH_KETRPOS      (-998)
91  #define MATCH_KETRPOS      (-997)  #define MATCH_ONCE         (-997)
92  #define MATCH_ONCE         (-996)  /* The next 5 must be kept together and in sequence so that a test that checks
93    for any one of them can use a range. */
94    #define MATCH_COMMIT       (-996)
95  #define MATCH_PRUNE        (-995)  #define MATCH_PRUNE        (-995)
96  #define MATCH_SKIP         (-994)  #define MATCH_SKIP         (-994)
97  #define MATCH_SKIP_ARG     (-993)  #define MATCH_SKIP_ARG     (-993)
98  #define MATCH_THEN         (-992)  #define MATCH_THEN         (-992)
99    #define MATCH_BACKTRACK_MAX MATCH_THEN
100    #define MATCH_BACKTRACK_MIN MATCH_COMMIT
101    
102  /* Maximum number of ints of offset to save on the stack for recursive calls.  /* Maximum number of ints of offset to save on the stack for recursive calls.
103  If the offset vector is bigger, malloc is used. This should be a multiple of 3,  If the offset vector is bigger, malloc is used. This should be a multiple of 3,
# Line 93  because the offset vector is always a mu Line 110  because the offset vector is always a mu
110  static const char rep_min[] = { 0, 0, 1, 1, 0, 0 };  static const char rep_min[] = { 0, 0, 1, 1, 0, 0 };
111  static const char rep_max[] = { 0, 0, 0, 0, 1, 1 };  static const char rep_max[] = { 0, 0, 0, 0, 1, 1 };
112    
   
   
113  #ifdef PCRE_DEBUG  #ifdef PCRE_DEBUG
114  /*************************************************  /*************************************************
115  *        Debugging function to print chars       *  *        Debugging function to print chars       *
# Line 113  Returns:     nothing Line 128  Returns:     nothing
128  */  */
129    
130  static void  static void
131  pchars(const uschar *p, int length, BOOL is_subject, match_data *md)  pchars(const pcre_uchar *p, int length, BOOL is_subject, match_data *md)
132  {  {
133  unsigned int c;  pcre_uint32 c;
134    BOOL utf = md->utf;
135  if (is_subject && length > md->end_subject - p) length = md->end_subject - p;  if (is_subject && length > md->end_subject - p) length = md->end_subject - p;
136  while (length-- > 0)  while (length-- > 0)
137    if (isprint(c = *(p++))) printf("%c", c); else printf("\\x%02x", c);    if (isprint(c = RAWUCHARINCTEST(p))) printf("%c", (char)c); else printf("\\x{%02x}", c);
138  }  }
139  #endif  #endif
140    
# Line 140  Arguments: Line 156  Arguments:
156    md          points to match data block    md          points to match data block
157    caseless    TRUE if caseless    caseless    TRUE if caseless
158    
159  Returns:      < 0 if not matched, otherwise the number of subject bytes matched  Returns:      >= 0 the number of subject bytes matched
160                  -1 no match
161                  -2 partial match; always given if at end subject
162  */  */
163    
164  static int  static int
165  match_ref(int offset, register USPTR eptr, int length, match_data *md,  match_ref(int offset, register PCRE_PUCHAR eptr, int length, match_data *md,
166    BOOL caseless)    BOOL caseless)
167  {  {
168  USPTR eptr_start = eptr;  PCRE_PUCHAR eptr_start = eptr;
169  register USPTR p = md->start_subject + md->offset_vector[offset];  register PCRE_PUCHAR p = md->start_subject + md->offset_vector[offset];
170    #ifdef SUPPORT_UTF
171    BOOL utf = md->utf;
172    #endif
173    
174  #ifdef PCRE_DEBUG  #ifdef PCRE_DEBUG
175  if (eptr >= md->end_subject)  if (eptr >= md->end_subject)
# Line 163  pchars(p, length, FALSE, md); Line 184  pchars(p, length, FALSE, md);
184  printf("\n");  printf("\n");
185  #endif  #endif
186    
187  /* Always fail if reference not set (and not JavaScript compatible). */  /* Always fail if reference not set (and not JavaScript compatible - in that
188    case the length is passed as zero). */
189    
190  if (length < 0) return -1;  if (length < 0) return -1;
191    
# Line 173  ASCII characters. */ Line 195  ASCII characters. */
195    
196  if (caseless)  if (caseless)
197    {    {
198  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
199  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
200    if (md->utf8)    if (utf)
201      {      {
202      /* Match characters up to the end of the reference. NOTE: the number of      /* Match characters up to the end of the reference. NOTE: the number of
203      bytes matched may differ, because there are some characters whose upper and      data units matched may differ, because in UTF-8 there are some characters
204      lower case versions code as different numbers of bytes. For example, U+023A      whose upper and lower case versions code have different numbers of bytes.
205      (2 bytes in UTF-8) is the upper case version of U+2C65 (3 bytes in UTF-8);      For example, U+023A (2 bytes in UTF-8) is the upper case version of U+2C65
206      a sequence of 3 of the former uses 6 bytes, as does a sequence of two of      (3 bytes in UTF-8); a sequence of 3 of the former uses 6 bytes, as does a
207      the latter. It is important, therefore, to check the length along the      sequence of two of the latter. It is important, therefore, to check the
208      reference, not along the subject (earlier code did this wrong). */      length along the reference, not along the subject (earlier code did this
209        wrong). */
210    
211      USPTR endptr = p + length;      PCRE_PUCHAR endptr = p + length;
212      while (p < endptr)      while (p < endptr)
213        {        {
214        int c, d;        pcre_uint32 c, d;
215        if (eptr >= md->end_subject) return -1;        const ucd_record *ur;
216          if (eptr >= md->end_subject) return -2;   /* Partial match */
217        GETCHARINC(c, eptr);        GETCHARINC(c, eptr);
218        GETCHARINC(d, p);        GETCHARINC(d, p);
219        if (c != d && c != UCD_OTHERCASE(d)) return -1;        ur = GET_UCD(d);
220          if (c != d && c != d + ur->other_case)
221            {
222            const pcre_uint32 *pp = PRIV(ucd_caseless_sets) + ur->caseset;
223            for (;;)
224              {
225              if (c < *pp) return -1;
226              if (c == *pp++) break;
227              }
228            }
229        }        }
230      }      }
231    else    else
# Line 202  if (caseless) Line 235  if (caseless)
235    /* The same code works when not in UTF-8 mode and in UTF-8 mode when there    /* The same code works when not in UTF-8 mode and in UTF-8 mode when there
236    is no UCP support. */    is no UCP support. */
237      {      {
     if (eptr + length > md->end_subject) return -1;  
238      while (length-- > 0)      while (length-- > 0)
239        { if (md->lcc[*p++] != md->lcc[*eptr++]) return -1; }        {
240          pcre_uint32 cc, cp;
241          if (eptr >= md->end_subject) return -2;   /* Partial match */
242          cc = RAWUCHARTEST(eptr);
243          cp = RAWUCHARTEST(p);
244          if (TABLE_GET(cp, md->lcc, cp) != TABLE_GET(cc, md->lcc, cc)) return -1;
245          p++;
246          eptr++;
247          }
248      }      }
249    }    }
250    
# Line 213  are in UTF-8 mode. */ Line 253  are in UTF-8 mode. */
253    
254  else  else
255    {    {
256    if (eptr + length > md->end_subject) return -1;    while (length-- > 0)
257    while (length-- > 0) if (*p++ != *eptr++) return -1;      {
258        if (eptr >= md->end_subject) return -2;   /* Partial match */
259        if (RAWUCHARINCTEST(p) != RAWUCHARINCTEST(eptr)) return -1;
260        }
261    }    }
262    
263  return eptr - eptr_start;  return (int)(eptr - eptr_start);
264  }  }
265    
266    
# Line 269  enum { RM1=1, RM2,  RM3,  RM4,  RM5,  RM Line 312  enum { RM1=1, RM2,  RM3,  RM4,  RM5,  RM
312         RM31,  RM32, RM33, RM34, RM35, RM36, RM37, RM38, RM39, RM40,         RM31,  RM32, RM33, RM34, RM35, RM36, RM37, RM38, RM39, RM40,
313         RM41,  RM42, RM43, RM44, RM45, RM46, RM47, RM48, RM49, RM50,         RM41,  RM42, RM43, RM44, RM45, RM46, RM47, RM48, RM49, RM50,
314         RM51,  RM52, RM53, RM54, RM55, RM56, RM57, RM58, RM59, RM60,         RM51,  RM52, RM53, RM54, RM55, RM56, RM57, RM58, RM59, RM60,
315         RM61,  RM62, RM63, RM64, RM65, RM66 };         RM61,  RM62, RM63, RM64, RM65, RM66, RM67, RM68 };
316    
317  /* These versions of the macros use the stack, as normal. There are debugging  /* These versions of the macros use the stack, as normal. There are debugging
318  versions and production versions. Note that the "rw" argument of RMATCH isn't  versions and production versions. Note that the "rw" argument of RMATCH isn't
# Line 287  actually used in this definition. */ Line 330  actually used in this definition. */
330    }    }
331  #define RRETURN(ra) \  #define RRETURN(ra) \
332    { \    { \
333    printf("match() returned %d from line %d ", ra, __LINE__); \    printf("match() returned %d from line %d\n", ra, __LINE__); \
334    return ra; \    return ra; \
335    }    }
336  #else  #else
# Line 307  argument of match(), which never changes Line 350  argument of match(), which never changes
350    
351  #define RMATCH(ra,rb,rc,rd,re,rw)\  #define RMATCH(ra,rb,rc,rd,re,rw)\
352    {\    {\
353    heapframe *newframe = (heapframe *)(pcre_stack_malloc)(sizeof(heapframe));\    heapframe *newframe = frame->Xnextframe;\
354    if (newframe == NULL) RRETURN(PCRE_ERROR_NOMEMORY);\    if (newframe == NULL)\
355    frame->Xwhere = rw; \      {\
356        newframe = (heapframe *)(PUBL(stack_malloc))(sizeof(heapframe));\
357        if (newframe == NULL) RRETURN(PCRE_ERROR_NOMEMORY);\
358        newframe->Xnextframe = NULL;\
359        frame->Xnextframe = newframe;\
360        }\
361      frame->Xwhere = rw;\
362    newframe->Xeptr = ra;\    newframe->Xeptr = ra;\
363    newframe->Xecode = rb;\    newframe->Xecode = rb;\
364    newframe->Xmstart = mstart;\    newframe->Xmstart = mstart;\
# Line 328  argument of match(), which never changes Line 377  argument of match(), which never changes
377    {\    {\
378    heapframe *oldframe = frame;\    heapframe *oldframe = frame;\
379    frame = oldframe->Xprevframe;\    frame = oldframe->Xprevframe;\
   (pcre_stack_free)(oldframe);\  
380    if (frame != NULL)\    if (frame != NULL)\
381      {\      {\
382      rrc = ra;\      rrc = ra;\
# Line 342  argument of match(), which never changes Line 390  argument of match(), which never changes
390    
391  typedef struct heapframe {  typedef struct heapframe {
392    struct heapframe *Xprevframe;    struct heapframe *Xprevframe;
393      struct heapframe *Xnextframe;
394    
395    /* Function arguments that may change */    /* Function arguments that may change */
396    
397    USPTR Xeptr;    PCRE_PUCHAR Xeptr;
398    const uschar *Xecode;    const pcre_uchar *Xecode;
399    USPTR Xmstart;    PCRE_PUCHAR Xmstart;
400    int Xoffset_top;    int Xoffset_top;
401    eptrblock *Xeptrb;    eptrblock *Xeptrb;
402    unsigned int Xrdepth;    unsigned int Xrdepth;
403    
404    /* Function local variables */    /* Function local variables */
405    
406    USPTR Xcallpat;    PCRE_PUCHAR Xcallpat;
407  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
408    USPTR Xcharptr;    PCRE_PUCHAR Xcharptr;
409  #endif  #endif
410    USPTR Xdata;    PCRE_PUCHAR Xdata;
411    USPTR Xnext;    PCRE_PUCHAR Xnext;
412    USPTR Xpp;    PCRE_PUCHAR Xpp;
413    USPTR Xprev;    PCRE_PUCHAR Xprev;
414    USPTR Xsaved_eptr;    PCRE_PUCHAR Xsaved_eptr;
415    
416    recursion_info Xnew_recursive;    recursion_info Xnew_recursive;
417    
# Line 372  typedef struct heapframe { Line 421  typedef struct heapframe {
421    
422  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
423    int Xprop_type;    int Xprop_type;
424    int Xprop_value;    unsigned int Xprop_value;
425    int Xprop_fail_result;    int Xprop_fail_result;
426    int Xoclength;    int Xoclength;
427    uschar Xocchars[8];    pcre_uchar Xocchars[6];
428  #endif  #endif
429    
430    int Xcodelink;    int Xcodelink;
# Line 385  typedef struct heapframe { Line 434  typedef struct heapframe {
434    int Xlength;    int Xlength;
435    int Xmax;    int Xmax;
436    int Xmin;    int Xmin;
437    int Xnumber;    unsigned int Xnumber;
438    int Xoffset;    int Xoffset;
439    int Xop;    unsigned int Xop;
440    int Xsave_capture_last;    pcre_int32 Xsave_capture_last;
441    int Xsave_offset1, Xsave_offset2, Xsave_offset3;    int Xsave_offset1, Xsave_offset2, Xsave_offset3;
442    int Xstacksave[REC_STACK_SAVE_MAX];    int Xstacksave[REC_STACK_SAVE_MAX];
443    
# Line 417  returns a negative (error) response, the Line 466  returns a negative (error) response, the
466  same response. */  same response. */
467    
468  /* These macros pack up tests that are used for partial matching, and which  /* These macros pack up tests that are used for partial matching, and which
469  appears several times in the code. We set the "hit end" flag if the pointer is  appear several times in the code. We set the "hit end" flag if the pointer is
470  at the end of the subject and also past the start of the subject (i.e.  at the end of the subject and also past the start of the subject (i.e.
471  something has been matched). For hard partial matching, we then return  something has been matched). For hard partial matching, we then return
472  immediately. The second one is used when we already know we are past the end of  immediately. The second one is used when we already know we are past the end of
# Line 440  the subject. */ Line 489  the subject. */
489    
490    
491  /* Performance note: It might be tempting to extract commonly used fields from  /* Performance note: It might be tempting to extract commonly used fields from
492  the md structure (e.g. utf8, end_subject) into individual variables to improve  the md structure (e.g. utf, end_subject) into individual variables to improve
493  performance. Tests using gcc on a SPARC disproved this; in the first case, it  performance. Tests using gcc on a SPARC disproved this; in the first case, it
494  made performance worse.  made performance worse.
495    
# Line 463  Returns:       MATCH_MATCH if matched Line 512  Returns:       MATCH_MATCH if matched
512  */  */
513    
514  static int  static int
515  match(REGISTER USPTR eptr, REGISTER const uschar *ecode, USPTR mstart,  match(REGISTER PCRE_PUCHAR eptr, REGISTER const pcre_uchar *ecode,
516    int offset_top, match_data *md, eptrblock *eptrb, unsigned int rdepth)    PCRE_PUCHAR mstart, int offset_top, match_data *md, eptrblock *eptrb,
517      unsigned int rdepth)
518  {  {
519  /* These variables do not need to be preserved over recursion in this function,  /* These variables do not need to be preserved over recursion in this function,
520  so they can be ordinary variables in all cases. Mark some of them with  so they can be ordinary variables in all cases. Mark some of them with
# Line 472  so they can be ordinary variables in all Line 522  so they can be ordinary variables in all
522    
523  register int  rrc;         /* Returns from recursive calls */  register int  rrc;         /* Returns from recursive calls */
524  register int  i;           /* Used for loops not involving calls to RMATCH() */  register int  i;           /* Used for loops not involving calls to RMATCH() */
525  register unsigned int c;   /* Character values not kept over RMATCH() calls */  register pcre_uint32 c;    /* Character values not kept over RMATCH() calls */
526  register BOOL utf8;        /* Local copy of UTF-8 flag for speed */  register BOOL utf;         /* Local copy of UTF flag for speed */
527    
528  BOOL minimize, possessive; /* Quantifier options */  BOOL minimize, possessive; /* Quantifier options */
529  BOOL caseless;  BOOL caseless;
530  int condcode;  int condcode;
531    
532  /* When recursion is not being used, all "local" variables that have to be  /* When recursion is not being used, all "local" variables that have to be
533  preserved over calls to RMATCH() are part of a "frame" which is obtained from  preserved over calls to RMATCH() are part of a "frame". We set up the top-level
534  heap storage. Set up the top-level frame here; others are obtained from the  frame on the stack here; subsequent instantiations are obtained from the heap
535  heap whenever RMATCH() does a "recursion". See the macro definitions above. */  whenever RMATCH() does a "recursion". See the macro definitions above. Putting
536    the top-level on the stack rather than malloc-ing them all gives a performance
537    boost in many cases where there is not much "recursion". */
538    
539  #ifdef NO_RECURSE  #ifdef NO_RECURSE
540  heapframe *frame = (heapframe *)(pcre_stack_malloc)(sizeof(heapframe));  heapframe *frame = (heapframe *)md->match_frames_base;
 if (frame == NULL) RRETURN(PCRE_ERROR_NOMEMORY);  
 frame->Xprevframe = NULL;            /* Marks the top level */  
541    
542  /* Copy in the original argument variables */  /* Copy in the original argument variables */
543    
# Line 513  HEAP_RECURSE: Line 563  HEAP_RECURSE:
563    
564  /* Ditto for the local variables */  /* Ditto for the local variables */
565    
566  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
567  #define charptr            frame->Xcharptr  #define charptr            frame->Xcharptr
568  #endif  #endif
569  #define callpat            frame->Xcallpat  #define callpat            frame->Xcallpat
# Line 571  declarations can be cut out in a block. Line 621  declarations can be cut out in a block.
621  below are for variables that do not have to be preserved over a recursive call  below are for variables that do not have to be preserved over a recursive call
622  to RMATCH(). */  to RMATCH(). */
623    
624  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
625  const uschar *charptr;  const pcre_uchar *charptr;
626  #endif  #endif
627  const uschar *callpat;  const pcre_uchar *callpat;
628  const uschar *data;  const pcre_uchar *data;
629  const uschar *next;  const pcre_uchar *next;
630  USPTR         pp;  PCRE_PUCHAR       pp;
631  const uschar *prev;  const pcre_uchar *prev;
632  USPTR         saved_eptr;  PCRE_PUCHAR       saved_eptr;
633    
634  recursion_info new_recursive;  recursion_info new_recursive;
635    
# Line 589  BOOL prev_is_word; Line 639  BOOL prev_is_word;
639    
640  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
641  int prop_type;  int prop_type;
642  int prop_value;  unsigned int prop_value;
643  int prop_fail_result;  int prop_fail_result;
644  int oclength;  int oclength;
645  uschar occhars[8];  pcre_uchar occhars[6];
646  #endif  #endif
647    
648  int codelink;  int codelink;
# Line 600  int ctype; Line 650  int ctype;
650  int length;  int length;
651  int max;  int max;
652  int min;  int min;
653  int number;  unsigned int number;
654  int offset;  int offset;
655  int op;  unsigned int op;
656  int save_capture_last;  pcre_int32 save_capture_last;
657  int save_offset1, save_offset2, save_offset3;  int save_offset1, save_offset2, save_offset3;
658  int stacksave[REC_STACK_SAVE_MAX];  int stacksave[REC_STACK_SAVE_MAX];
659    
660  eptrblock newptrb;  eptrblock newptrb;
661    
662    /* There is a special fudge for calling match() in a way that causes it to
663    measure the size of its basic stack frame when the stack is being used for
664    recursion. The second argument (ecode) being NULL triggers this behaviour. It
665    cannot normally ever be NULL. The return is the negated value of the frame
666    size. */
667    
668    if (ecode == NULL)
669      {
670      if (rdepth == 0)
671        return match((PCRE_PUCHAR)&rdepth, NULL, NULL, 0, NULL, NULL, 1);
672      else
673        {
674        int len = (char *)&rdepth - (char *)eptr;
675        return (len > 0)? -len : len;
676        }
677      }
678  #endif     /* NO_RECURSE */  #endif     /* NO_RECURSE */
679    
680  /* To save space on the stack and in the heap frame, I have doubled up on some  /* To save space on the stack and in the heap frame, I have doubled up on some
# Line 620  the alternative names that are used. */ Line 687  the alternative names that are used. */
687  #define code_offset   codelink  #define code_offset   codelink
688  #define condassert    condition  #define condassert    condition
689  #define matched_once  prev_is_word  #define matched_once  prev_is_word
690    #define foc           number
691    #define save_mark     data
692    
693  /* These statements are here to stop the compiler complaining about unitialized  /* These statements are here to stop the compiler complaining about unitialized
694  variables. */  variables. */
# Line 645  defined). However, RMATCH isn't like a f Line 714  defined). However, RMATCH isn't like a f
714  complicated macro. It has to be used in one particular way. This shouldn't,  complicated macro. It has to be used in one particular way. This shouldn't,
715  however, impact performance when true recursion is being used. */  however, impact performance when true recursion is being used. */
716    
717  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
718  utf8 = md->utf8;       /* Local copy of the flag */  utf = md->utf;       /* Local copy of the flag */
719  #else  #else
720  utf8 = FALSE;  utf = FALSE;
721  #endif  #endif
722    
723  /* First check that we haven't called match() too many times, or that we  /* First check that we haven't called match() too many times, or that we
# Line 689  for (;;) Line 758  for (;;)
758      case OP_MARK:      case OP_MARK:
759      md->nomatch_mark = ecode + 2;      md->nomatch_mark = ecode + 2;
760      md->mark = NULL;    /* In case previously set by assertion */      md->mark = NULL;    /* In case previously set by assertion */
761      RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode] + ecode[1], offset_top, md,      RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode] + ecode[1], offset_top, md,
762        eptrb, RM55);        eptrb, RM55);
763      if ((rrc == MATCH_MATCH || rrc == MATCH_ACCEPT) &&      if ((rrc == MATCH_MATCH || rrc == MATCH_ACCEPT) &&
764           md->mark == NULL) md->mark = ecode + 2;           md->mark == NULL) md->mark = ecode + 2;
# Line 702  for (;;) Line 771  for (;;)
771      unaltered. */      unaltered. */
772    
773      else if (rrc == MATCH_SKIP_ARG &&      else if (rrc == MATCH_SKIP_ARG &&
774          strcmp((char *)(ecode + 2), (char *)(md->start_match_ptr)) == 0)          STRCMP_UC_UC_TEST(ecode + 2, md->start_match_ptr) == 0)
775        {        {
776        md->start_match_ptr = eptr;        md->start_match_ptr = eptr;
777        RRETURN(MATCH_SKIP);        RRETURN(MATCH_SKIP);
# Line 712  for (;;) Line 781  for (;;)
781      case OP_FAIL:      case OP_FAIL:
782      RRETURN(MATCH_NOMATCH);      RRETURN(MATCH_NOMATCH);
783    
     /* COMMIT overrides PRUNE, SKIP, and THEN */  
   
784      case OP_COMMIT:      case OP_COMMIT:
785      RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,      RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md,
786        eptrb, RM52);        eptrb, RM52);
787      if (rrc != MATCH_NOMATCH && rrc != MATCH_PRUNE &&      if (rrc != MATCH_NOMATCH) RRETURN(rrc);
         rrc != MATCH_SKIP && rrc != MATCH_SKIP_ARG &&  
         rrc != MATCH_THEN)  
       RRETURN(rrc);  
788      RRETURN(MATCH_COMMIT);      RRETURN(MATCH_COMMIT);
789    
     /* PRUNE overrides THEN */  
   
790      case OP_PRUNE:      case OP_PRUNE:
791      RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,      RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md,
792        eptrb, RM51);        eptrb, RM51);
793      if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN) RRETURN(rrc);      if (rrc != MATCH_NOMATCH) RRETURN(rrc);
794      RRETURN(MATCH_PRUNE);      RRETURN(MATCH_PRUNE);
795    
796      case OP_PRUNE_ARG:      case OP_PRUNE_ARG:
797      md->nomatch_mark = ecode + 2;      md->nomatch_mark = ecode + 2;
798      md->mark = NULL;    /* In case previously set by assertion */      md->mark = NULL;    /* In case previously set by assertion */
799      RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode] + ecode[1], offset_top, md,      RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode] + ecode[1], offset_top, md,
800        eptrb, RM56);        eptrb, RM56);
801      if ((rrc == MATCH_MATCH || rrc == MATCH_ACCEPT) &&      if ((rrc == MATCH_MATCH || rrc == MATCH_ACCEPT) &&
802           md->mark == NULL) md->mark = ecode + 2;           md->mark == NULL) md->mark = ecode + 2;
803      if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN) RRETURN(rrc);      if (rrc != MATCH_NOMATCH) RRETURN(rrc);
804      RRETURN(MATCH_PRUNE);      RRETURN(MATCH_PRUNE);
805    
     /* SKIP overrides PRUNE and THEN */  
   
806      case OP_SKIP:      case OP_SKIP:
807      RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,      RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md,
808        eptrb, RM53);        eptrb, RM53);
809      if (rrc != MATCH_NOMATCH && rrc != MATCH_PRUNE && rrc != MATCH_THEN)      if (rrc != MATCH_NOMATCH) RRETURN(rrc);
       RRETURN(rrc);  
810      md->start_match_ptr = eptr;   /* Pass back current position */      md->start_match_ptr = eptr;   /* Pass back current position */
811      RRETURN(MATCH_SKIP);      RRETURN(MATCH_SKIP);
812    
813      /* Note that, for Perl compatibility, SKIP with an argument does NOT set      /* Note that, for Perl compatibility, SKIP with an argument does NOT set
814      nomatch_mark. There is a flag that disables this opcode when re-matching a      nomatch_mark. When a pattern match ends with a SKIP_ARG for which there was
815      pattern that ended with a SKIP for which there was not a matching MARK. */      not a matching mark, we have to re-run the match, ignoring the SKIP_ARG
816        that failed and any that preceed it (either they also failed, or were not
817        triggered). To do this, we maintain a count of executed SKIP_ARGs. If a
818        SKIP_ARG gets to top level, the match is re-run with md->ignore_skip_arg
819        set to the count of the one that failed. */
820    
821      case OP_SKIP_ARG:      case OP_SKIP_ARG:
822      if (md->ignore_skip_arg)      md->skip_arg_count++;
823        if (md->skip_arg_count <= md->ignore_skip_arg)
824        {        {
825        ecode += _pcre_OP_lengths[*ecode] + ecode[1];        ecode += PRIV(OP_lengths)[*ecode] + ecode[1];
826        break;        break;
827        }        }
828      RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode] + ecode[1], offset_top, md,      RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode] + ecode[1], offset_top, md,
829        eptrb, RM57);        eptrb, RM57);
830      if (rrc != MATCH_NOMATCH && rrc != MATCH_PRUNE && rrc != MATCH_THEN)      if (rrc != MATCH_NOMATCH) RRETURN(rrc);
831        RRETURN(rrc);  
   
832      /* Pass back the current skip name by overloading md->start_match_ptr and      /* Pass back the current skip name by overloading md->start_match_ptr and
833      returning the special MATCH_SKIP_ARG return code. This will either be      returning the special MATCH_SKIP_ARG return code. This will either be
834      caught by a matching MARK, or get to the top, where it causes a rematch      caught by a matching MARK, or get to the top, where it causes a rematch
835      with the md->ignore_skip_arg flag set. */      with md->ignore_skip_arg set to the value of md->skip_arg_count. */
836    
837      md->start_match_ptr = ecode + 2;      md->start_match_ptr = ecode + 2;
838      RRETURN(MATCH_SKIP_ARG);      RRETURN(MATCH_SKIP_ARG);
# Line 779  for (;;) Line 842  for (;;)
842      match pointer to do this. */      match pointer to do this. */
843    
844      case OP_THEN:      case OP_THEN:
845      RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,      RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md,
846        eptrb, RM54);        eptrb, RM54);
847      if (rrc != MATCH_NOMATCH) RRETURN(rrc);      if (rrc != MATCH_NOMATCH) RRETURN(rrc);
848      md->start_match_ptr = ecode;      md->start_match_ptr = ecode;
# Line 788  for (;;) Line 851  for (;;)
851      case OP_THEN_ARG:      case OP_THEN_ARG:
852      md->nomatch_mark = ecode + 2;      md->nomatch_mark = ecode + 2;
853      md->mark = NULL;    /* In case previously set by assertion */      md->mark = NULL;    /* In case previously set by assertion */
854      RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode] + ecode[1], offset_top,      RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode] + ecode[1], offset_top,
855        md, eptrb, RM58);        md, eptrb, RM58);
856      if ((rrc == MATCH_MATCH || rrc == MATCH_ACCEPT) &&      if ((rrc == MATCH_MATCH || rrc == MATCH_ACCEPT) &&
857           md->mark == NULL) md->mark = ecode + 2;           md->mark == NULL) md->mark = ecode + 2;
# Line 812  for (;;) Line 875  for (;;)
875      case OP_ONCE_NC:      case OP_ONCE_NC:
876      prev = ecode;      prev = ecode;
877      saved_eptr = eptr;      saved_eptr = eptr;
878        save_mark = md->mark;
879      do      do
880        {        {
881        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, eptrb, RM64);        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, eptrb, RM64);
# Line 830  for (;;) Line 894  for (;;)
894    
895        if (rrc != MATCH_NOMATCH) RRETURN(rrc);        if (rrc != MATCH_NOMATCH) RRETURN(rrc);
896        ecode += GET(ecode,1);        ecode += GET(ecode,1);
897          md->mark = save_mark;
898        }        }
899      while (*ecode == OP_ALT);      while (*ecode == OP_ALT);
900    
# Line 869  for (;;) Line 934  for (;;)
934        }        }
935      else  /* OP_KETRMAX */      else  /* OP_KETRMAX */
936        {        {
       md->match_function_type = MATCH_CBEGROUP;  
937        RMATCH(eptr, prev, offset_top, md, eptrb, RM66);        RMATCH(eptr, prev, offset_top, md, eptrb, RM66);
938        if (rrc != MATCH_NOMATCH) RRETURN(rrc);        if (rrc != MATCH_NOMATCH) RRETURN(rrc);
939        ecode += 1 + LINK_SIZE;        ecode += 1 + LINK_SIZE;
# Line 909  for (;;) Line 973  for (;;)
973        save_offset2 = md->offset_vector[offset+1];        save_offset2 = md->offset_vector[offset+1];
974        save_offset3 = md->offset_vector[md->offset_end - number];        save_offset3 = md->offset_vector[md->offset_end - number];
975        save_capture_last = md->capture_last;        save_capture_last = md->capture_last;
976          save_mark = md->mark;
977    
978        DPRINTF(("saving %d %d %d\n", save_offset1, save_offset2, save_offset3));        DPRINTF(("saving %d %d %d\n", save_offset1, save_offset2, save_offset3));
979        md->offset_vector[md->offset_end - number] =        md->offset_vector[md->offset_end - number] =
# Line 917  for (;;) Line 982  for (;;)
982        for (;;)        for (;;)
983          {          {
984          if (op >= OP_SBRA) md->match_function_type = MATCH_CBEGROUP;          if (op >= OP_SBRA) md->match_function_type = MATCH_CBEGROUP;
985          RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,          RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md,
986            eptrb, RM1);            eptrb, RM1);
987          if (rrc == MATCH_ONCE) break;  /* Backing up through an atomic group */          if (rrc == MATCH_ONCE) break;  /* Backing up through an atomic group */
988    
# Line 945  for (;;) Line 1010  for (;;)
1010          if (rrc != MATCH_NOMATCH) RRETURN(rrc);          if (rrc != MATCH_NOMATCH) RRETURN(rrc);
1011          md->capture_last = save_capture_last;          md->capture_last = save_capture_last;
1012          ecode += GET(ecode, 1);          ecode += GET(ecode, 1);
1013            md->mark = save_mark;
1014          if (*ecode != OP_ALT) break;          if (*ecode != OP_ALT) break;
1015          }          }
1016    
# Line 996  for (;;) Line 1062  for (;;)
1062    
1063      for (;;)      for (;;)
1064        {        {
1065        if (op >= OP_SBRA || op == OP_ONCE) md->match_function_type = MATCH_CBEGROUP;        if (op >= OP_SBRA || op == OP_ONCE)
1066            md->match_function_type = MATCH_CBEGROUP;
1067    
1068        /* If this is not a possibly empty group, and there are no (*THEN)s in        /* If this is not a possibly empty group, and there are no (*THEN)s in
1069        the pattern, and this is the final alternative, optimize as described        the pattern, and this is the final alternative, optimize as described
# Line 1004  for (;;) Line 1071  for (;;)
1071    
1072        else if (!md->hasthen && ecode[GET(ecode, 1)] != OP_ALT)        else if (!md->hasthen && ecode[GET(ecode, 1)] != OP_ALT)
1073          {          {
1074          ecode += _pcre_OP_lengths[*ecode];          ecode += PRIV(OP_lengths)[*ecode];
1075          goto TAIL_RECURSE;          goto TAIL_RECURSE;
1076          }          }
1077    
1078        /* In all other cases, we have to make another call to match(). */        /* In all other cases, we have to make another call to match(). */
1079    
1080        RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md, eptrb,        save_mark = md->mark;
1081          save_capture_last = md->capture_last;
1082          RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md, eptrb,
1083          RM2);          RM2);
1084    
1085        /* See comment in the code for capturing groups above about handling        /* See comment in the code for capturing groups above about handling
# Line 1028  for (;;) Line 1097  for (;;)
1097          {          {
1098          if (rrc == MATCH_ONCE)          if (rrc == MATCH_ONCE)
1099            {            {
1100            const uschar *scode = ecode;            const pcre_uchar *scode = ecode;
1101            if (*scode != OP_ONCE)           /* If not at start, find it */            if (*scode != OP_ONCE)           /* If not at start, find it */
1102              {              {
1103              while (*scode == OP_ALT) scode += GET(scode, 1);              while (*scode == OP_ALT) scode += GET(scode, 1);
# Line 1039  for (;;) Line 1108  for (;;)
1108          RRETURN(rrc);          RRETURN(rrc);
1109          }          }
1110        ecode += GET(ecode, 1);        ecode += GET(ecode, 1);
1111          md->mark = save_mark;
1112        if (*ecode != OP_ALT) break;        if (*ecode != OP_ALT) break;
1113          md->capture_last = save_capture_last;
1114        }        }
1115    
1116      RRETURN(MATCH_NOMATCH);      RRETURN(MATCH_NOMATCH);
# Line 1070  for (;;) Line 1141  for (;;)
1141      if (offset < md->offset_max)      if (offset < md->offset_max)
1142        {        {
1143        matched_once = FALSE;        matched_once = FALSE;
1144        code_offset = ecode - md->start_code;        code_offset = (int)(ecode - md->start_code);
1145    
1146        save_offset1 = md->offset_vector[offset];        save_offset1 = md->offset_vector[offset];
1147        save_offset2 = md->offset_vector[offset+1];        save_offset2 = md->offset_vector[offset+1];
# Line 1093  for (;;) Line 1164  for (;;)
1164          md->offset_vector[md->offset_end - number] =          md->offset_vector[md->offset_end - number] =
1165            (int)(eptr - md->start_subject);            (int)(eptr - md->start_subject);
1166          if (op >= OP_SBRA) md->match_function_type = MATCH_CBEGROUP;          if (op >= OP_SBRA) md->match_function_type = MATCH_CBEGROUP;
1167          RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,          RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md,
1168            eptrb, RM63);            eptrb, RM63);
1169          if (rrc == MATCH_KETRPOS)          if (rrc == MATCH_KETRPOS)
1170            {            {
# Line 1160  for (;;) Line 1231  for (;;)
1231    
1232      POSSESSIVE_NON_CAPTURE:      POSSESSIVE_NON_CAPTURE:
1233      matched_once = FALSE;      matched_once = FALSE;
1234      code_offset = ecode - md->start_code;      code_offset = (int)(ecode - md->start_code);
1235        save_capture_last = md->capture_last;
1236    
1237      for (;;)      for (;;)
1238        {        {
1239        if (op >= OP_SBRA) md->match_function_type = MATCH_CBEGROUP;        if (op >= OP_SBRA) md->match_function_type = MATCH_CBEGROUP;
1240        RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,        RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md,
1241          eptrb, RM48);          eptrb, RM48);
1242        if (rrc == MATCH_KETRPOS)        if (rrc == MATCH_KETRPOS)
1243          {          {
# Line 1190  for (;;) Line 1262  for (;;)
1262        if (rrc != MATCH_NOMATCH) RRETURN(rrc);        if (rrc != MATCH_NOMATCH) RRETURN(rrc);
1263        ecode += GET(ecode, 1);        ecode += GET(ecode, 1);
1264        if (*ecode != OP_ALT) break;        if (*ecode != OP_ALT) break;
1265          md->capture_last = save_capture_last;
1266        }        }
1267    
1268      if (matched_once || allow_zero)      if (matched_once || allow_zero)
# Line 1215  for (;;) Line 1288  for (;;)
1288    
1289      if (ecode[LINK_SIZE+1] == OP_CALLOUT)      if (ecode[LINK_SIZE+1] == OP_CALLOUT)
1290        {        {
1291        if (pcre_callout != NULL)        if (PUBL(callout) != NULL)
1292          {          {
1293          pcre_callout_block cb;          PUBL(callout_block) cb;
1294          cb.version          = 2;   /* Version 1 of the callout block */          cb.version          = 2;   /* Version 1 of the callout block */
1295          cb.callout_number   = ecode[LINK_SIZE+2];          cb.callout_number   = ecode[LINK_SIZE+2];
1296          cb.offset_vector    = md->offset_vector;          cb.offset_vector    = md->offset_vector;
1297    #if defined COMPILE_PCRE8
1298          cb.subject          = (PCRE_SPTR)md->start_subject;          cb.subject          = (PCRE_SPTR)md->start_subject;
1299    #elif defined COMPILE_PCRE16
1300            cb.subject          = (PCRE_SPTR16)md->start_subject;
1301    #elif defined COMPILE_PCRE32
1302            cb.subject          = (PCRE_SPTR32)md->start_subject;
1303    #endif
1304          cb.subject_length   = (int)(md->end_subject - md->start_subject);          cb.subject_length   = (int)(md->end_subject - md->start_subject);
1305          cb.start_match      = (int)(mstart - md->start_subject);          cb.start_match      = (int)(mstart - md->start_subject);
1306          cb.current_position = (int)(eptr - md->start_subject);          cb.current_position = (int)(eptr - md->start_subject);
1307          cb.pattern_position = GET(ecode, LINK_SIZE + 3);          cb.pattern_position = GET(ecode, LINK_SIZE + 3);
1308          cb.next_item_length = GET(ecode, 3 + 2*LINK_SIZE);          cb.next_item_length = GET(ecode, 3 + 2*LINK_SIZE);
1309          cb.capture_top      = offset_top/2;          cb.capture_top      = offset_top/2;
1310          cb.capture_last     = md->capture_last;          cb.capture_last     = md->capture_last & CAPLMASK;
1311            /* Internal change requires this for API compatibility. */
1312            if (cb.capture_last == 0) cb.capture_last = -1;
1313          cb.callout_data     = md->callout_data;          cb.callout_data     = md->callout_data;
1314          cb.mark             = md->nomatch_mark;          cb.mark             = md->nomatch_mark;
1315          if ((rrc = (*pcre_callout)(&cb)) > 0) RRETURN(MATCH_NOMATCH);          if ((rrc = (*PUBL(callout))(&cb)) > 0) RRETURN(MATCH_NOMATCH);
1316          if (rrc < 0) RRETURN(rrc);          if (rrc < 0) RRETURN(rrc);
1317          }          }
1318        ecode += _pcre_OP_lengths[OP_CALLOUT];        ecode += PRIV(OP_lengths)[OP_CALLOUT];
1319          codelink -= PRIV(OP_lengths)[OP_CALLOUT];
1320        }        }
1321    
1322      condcode = ecode[LINK_SIZE+1];      condcode = ecode[LINK_SIZE+1];
# Line 1250  for (;;) Line 1332  for (;;)
1332          }          }
1333        else        else
1334          {          {
1335          int recno = GET2(ecode, LINK_SIZE + 2);   /* Recursion group number*/          unsigned int recno = GET2(ecode, LINK_SIZE + 2);   /* Recursion group number*/
1336          condition = (recno == RREF_ANY || recno == md->recursive->group_num);          condition = (recno == RREF_ANY || recno == md->recursive->group_num);
1337    
1338          /* If the test is for recursion into a specific subpattern, and it is          /* If the test is for recursion into a specific subpattern, and it is
# Line 1260  for (;;) Line 1342  for (;;)
1342    
1343          if (!condition && condcode == OP_NRREF)          if (!condition && condcode == OP_NRREF)
1344            {            {
1345            uschar *slotA = md->name_table;            pcre_uchar *slotA = md->name_table;
1346            for (i = 0; i < md->name_count; i++)            for (i = 0; i < md->name_count; i++)
1347              {              {
1348              if (GET2(slotA, 0) == recno) break;              if (GET2(slotA, 0) == recno) break;
# Line 1273  for (;;) Line 1355  for (;;)
1355    
1356            if (i < md->name_count)            if (i < md->name_count)
1357              {              {
1358              uschar *slotB = slotA;              pcre_uchar *slotB = slotA;
1359              while (slotB > md->name_table)              while (slotB > md->name_table)
1360                {                {
1361                slotB -= md->name_entry_size;                slotB -= md->name_entry_size;
1362                if (strcmp((char *)slotA + 2, (char *)slotB + 2) == 0)                if (STRCMP_UC_UC(slotA + IMM2_SIZE, slotB + IMM2_SIZE) == 0)
1363                  {                  {
1364                  condition = GET2(slotB, 0) == md->recursive->group_num;                  condition = GET2(slotB, 0) == md->recursive->group_num;
1365                  if (condition) break;                  if (condition) break;
# Line 1293  for (;;) Line 1375  for (;;)
1375                for (i++; i < md->name_count; i++)                for (i++; i < md->name_count; i++)
1376                  {                  {
1377                  slotB += md->name_entry_size;                  slotB += md->name_entry_size;
1378                  if (strcmp((char *)slotA + 2, (char *)slotB + 2) == 0)                  if (STRCMP_UC_UC(slotA + IMM2_SIZE, slotB + IMM2_SIZE) == 0)
1379                    {                    {
1380                    condition = GET2(slotB, 0) == md->recursive->group_num;                    condition = GET2(slotB, 0) == md->recursive->group_num;
1381                    if (condition) break;                    if (condition) break;
# Line 1306  for (;;) Line 1388  for (;;)
1388    
1389          /* Chose branch according to the condition */          /* Chose branch according to the condition */
1390    
1391          ecode += condition? 3 : GET(ecode, 1);          ecode += condition? 1 + IMM2_SIZE : GET(ecode, 1);
1392          }          }
1393        }        }
1394    
# Line 1322  for (;;) Line 1404  for (;;)
1404    
1405        if (!condition && condcode == OP_NCREF)        if (!condition && condcode == OP_NCREF)
1406          {          {
1407          int refno = offset >> 1;          unsigned int refno = offset >> 1;
1408          uschar *slotA = md->name_table;          pcre_uchar *slotA = md->name_table;
1409    
1410          for (i = 0; i < md->name_count; i++)          for (i = 0; i < md->name_count; i++)
1411            {            {
# Line 1337  for (;;) Line 1419  for (;;)
1419    
1420          if (i < md->name_count)          if (i < md->name_count)
1421            {            {
1422            uschar *slotB = slotA;            pcre_uchar *slotB = slotA;
1423            while (slotB > md->name_table)            while (slotB > md->name_table)
1424              {              {
1425              slotB -= md->name_entry_size;              slotB -= md->name_entry_size;
1426              if (strcmp((char *)slotA + 2, (char *)slotB + 2) == 0)              if (STRCMP_UC_UC(slotA + IMM2_SIZE, slotB + IMM2_SIZE) == 0)
1427                {                {
1428                offset = GET2(slotB, 0) << 1;                offset = GET2(slotB, 0) << 1;
1429                condition = offset < offset_top &&                condition = offset < offset_top &&
# Line 1359  for (;;) Line 1441  for (;;)
1441              for (i++; i < md->name_count; i++)              for (i++; i < md->name_count; i++)
1442                {                {
1443                slotB += md->name_entry_size;                slotB += md->name_entry_size;
1444                if (strcmp((char *)slotA + 2, (char *)slotB + 2) == 0)                if (STRCMP_UC_UC(slotA + IMM2_SIZE, slotB + IMM2_SIZE) == 0)
1445                  {                  {
1446                  offset = GET2(slotB, 0) << 1;                  offset = GET2(slotB, 0) << 1;
1447                  condition = offset < offset_top &&                  condition = offset < offset_top &&
# Line 1374  for (;;) Line 1456  for (;;)
1456    
1457        /* Chose branch according to the condition */        /* Chose branch according to the condition */
1458    
1459        ecode += condition? 3 : GET(ecode, 1);        ecode += condition? 1 + IMM2_SIZE : GET(ecode, 1);
1460        }        }
1461    
1462      else if (condcode == OP_DEF)     /* DEFINE - always false */      else if (condcode == OP_DEF)     /* DEFINE - always false */
# Line 1450  for (;;) Line 1532  for (;;)
1532      to close any currently open capturing brackets. */      to close any currently open capturing brackets. */
1533    
1534      case OP_CLOSE:      case OP_CLOSE:
1535      number = GET2(ecode, 1);      number = GET2(ecode, 1);   /* Must be less than 65536 */
1536      offset = number << 1;      offset = number << 1;
1537    
1538  #ifdef PCRE_DEBUG  #ifdef PCRE_DEBUG
# Line 1458  for (;;) Line 1540  for (;;)
1540        printf("\n");        printf("\n");
1541  #endif  #endif
1542    
1543      md->capture_last = number;      md->capture_last = (md->capture_last & OVFLMASK) | number;
1544      if (offset >= md->offset_max) md->offset_overflow = TRUE; else      if (offset >= md->offset_max) md->capture_last |= OVFLBIT; else
1545        {        {
1546        md->offset_vector[offset] =        md->offset_vector[offset] =
1547          md->offset_vector[md->offset_end - number];          md->offset_vector[md->offset_end - number];
1548        md->offset_vector[offset+1] = (int)(eptr - md->start_subject);        md->offset_vector[offset+1] = (int)(eptr - md->start_subject);
1549        if (offset_top <= offset) offset_top = offset + 2;        if (offset_top <= offset) offset_top = offset + 2;
1550        }        }
1551      ecode += 3;      ecode += 1 + IMM2_SIZE;
1552      break;      break;
1553    
1554    
# Line 1513  for (;;) Line 1595  for (;;)
1595    
1596      case OP_ASSERT:      case OP_ASSERT:
1597      case OP_ASSERTBACK:      case OP_ASSERTBACK:
1598        save_mark = md->mark;
1599      if (md->match_function_type == MATCH_CONDASSERT)      if (md->match_function_type == MATCH_CONDASSERT)
1600        {        {
1601        condassert = TRUE;        condassert = TRUE;
# Line 1520  for (;;) Line 1603  for (;;)
1603        }        }
1604      else condassert = FALSE;      else condassert = FALSE;
1605    
1606        /* Loop for each branch */
1607    
1608      do      do
1609        {        {
1610        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, NULL, RM4);        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, NULL, RM4);
1611    
1612          /* A match means that the assertion is true; break out of the loop
1613          that matches its alternatives. */
1614    
1615        if (rrc == MATCH_MATCH || rrc == MATCH_ACCEPT)        if (rrc == MATCH_MATCH || rrc == MATCH_ACCEPT)
1616          {          {
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    
1621          /* If not matched, restore the previous mark setting. */
1622    
1623          md->mark = save_mark;
1624    
1625        /* PCRE does not allow THEN to escape beyond an assertion; it is treated        /* See comment in the code for capturing groups above about handling
1626        as NOMATCH. */        THEN. */
1627    
1628        if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN) RRETURN(rrc);        if (rrc == MATCH_THEN)
1629            {
1630            next = ecode + GET(ecode,1);
1631            if (md->start_match_ptr < next &&
1632                (*ecode == OP_ALT || *next == OP_ALT))
1633              rrc = MATCH_NOMATCH;
1634            }
1635    
1636          /* Anything other than NOMATCH causes the entire assertion to fail,
1637          passing back the return code. This includes COMMIT, SKIP, PRUNE and an
1638          uncaptured THEN, which means they take their normal effect. This
1639          consistent approach does not always have exactly the same effect as in
1640          Perl. */
1641    
1642          if (rrc != MATCH_NOMATCH) RRETURN(rrc);
1643        ecode += GET(ecode, 1);        ecode += GET(ecode, 1);
1644        }        }
1645      while (*ecode == OP_ALT);      while (*ecode == OP_ALT);   /* Continue for next alternative */
1646    
1647        /* If we have tried all the alternative branches, the assertion has
1648        failed. If not, we broke out after a match. */
1649    
1650      if (*ecode == OP_KET) RRETURN(MATCH_NOMATCH);      if (*ecode == OP_KET) RRETURN(MATCH_NOMATCH);
1651    
# Line 1543  for (;;) Line 1653  for (;;)
1653    
1654      if (condassert) RRETURN(MATCH_MATCH);      if (condassert) RRETURN(MATCH_MATCH);
1655    
1656      /* Continue from after the assertion, updating the offsets high water      /* Continue from after a successful assertion, updating the offsets high
1657      mark, since extracts may have been taken during the assertion. */      water mark, since extracts may have been taken during the assertion. */
1658    
1659      do ecode += GET(ecode,1); while (*ecode == OP_ALT);      do ecode += GET(ecode,1); while (*ecode == OP_ALT);
1660      ecode += 1 + LINK_SIZE;      ecode += 1 + LINK_SIZE;
1661      offset_top = md->end_offset_top;      offset_top = md->end_offset_top;
1662      continue;      continue;
1663    
1664      /* Negative assertion: all branches must fail to match. Encountering SKIP,      /* Negative assertion: all branches must fail to match for the assertion to
1665      PRUNE, or COMMIT means we must assume failure without checking subsequent      succeed. */
     branches. */  
1666    
1667      case OP_ASSERT_NOT:      case OP_ASSERT_NOT:
1668      case OP_ASSERTBACK_NOT:      case OP_ASSERTBACK_NOT:
1669        save_mark = md->mark;
1670      if (md->match_function_type == MATCH_CONDASSERT)      if (md->match_function_type == MATCH_CONDASSERT)
1671        {        {
1672        condassert = TRUE;        condassert = TRUE;
# Line 1564  for (;;) Line 1674  for (;;)
1674        }        }
1675      else condassert = FALSE;      else condassert = FALSE;
1676    
1677        /* Loop for each alternative branch. */
1678    
1679      do      do
1680        {        {
1681        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, NULL, RM5);        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, NULL, RM5);
1682        if (rrc == MATCH_MATCH || rrc == MATCH_ACCEPT) RRETURN(MATCH_NOMATCH);        md->mark = save_mark;   /* Always restore the mark setting */
1683        if (rrc == MATCH_SKIP || rrc == MATCH_PRUNE || rrc == MATCH_COMMIT)  
1684          switch(rrc)
1685          {          {
1686          do ecode += GET(ecode,1); while (*ecode == OP_ALT);          case MATCH_MATCH:            /* A successful match means */
1687          break;          case MATCH_ACCEPT:           /* the assertion has failed. */
1688          }          RRETURN(MATCH_NOMATCH);
1689    
1690            case MATCH_NOMATCH:          /* Carry on with next branch */
1691            break;
1692    
1693        /* PCRE does not allow THEN to escape beyond an assertion; it is treated          /* See comment in the code for capturing groups above about handling
1694        as NOMATCH. */          THEN. */
1695    
1696        if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN) RRETURN(rrc);          case MATCH_THEN:
1697            next = ecode + GET(ecode,1);
1698            if (md->start_match_ptr < next &&
1699                (*ecode == OP_ALT || *next == OP_ALT))
1700              {
1701              rrc = MATCH_NOMATCH;
1702              break;
1703              }
1704            /* Otherwise fall through. */
1705    
1706            /* COMMIT, SKIP, PRUNE, and an uncaptured THEN cause the whole
1707            assertion to fail to match, without considering any more alternatives.
1708            Failing to match means the assertion is true. This is a consistent
1709            approach, but does not always have the same effect as in Perl. */
1710    
1711            case MATCH_COMMIT:
1712            case MATCH_SKIP:
1713            case MATCH_SKIP_ARG:
1714            case MATCH_PRUNE:
1715            do ecode += GET(ecode,1); while (*ecode == OP_ALT);
1716            goto NEG_ASSERT_TRUE;   /* Break out of alternation loop */
1717    
1718            /* Anything else is an error */
1719    
1720            default:
1721            RRETURN(rrc);
1722            }
1723    
1724          /* Continue with next branch */
1725    
1726        ecode += GET(ecode,1);        ecode += GET(ecode,1);
1727        }        }
1728      while (*ecode == OP_ALT);      while (*ecode == OP_ALT);
1729    
1730        /* All branches in the assertion failed to match. */
1731    
1732        NEG_ASSERT_TRUE:
1733      if (condassert) RRETURN(MATCH_MATCH);  /* Condition assertion */      if (condassert) RRETURN(MATCH_MATCH);  /* Condition assertion */
1734        ecode += 1 + LINK_SIZE;                /* Continue with current branch */
     ecode += 1 + LINK_SIZE;  
1735      continue;      continue;
1736    
1737      /* Move the subject pointer back. This occurs only at the start of      /* Move the subject pointer back. This occurs only at the start of
# Line 1593  for (;;) Line 1740  for (;;)
1740      back a number of characters, not bytes. */      back a number of characters, not bytes. */
1741    
1742      case OP_REVERSE:      case OP_REVERSE:
1743  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
1744      if (utf8)      if (utf)
1745        {        {
1746        i = GET(ecode, 1);        i = GET(ecode, 1);
1747        while (i-- > 0)        while (i-- > 0)
# Line 1625  for (;;) Line 1772  for (;;)
1772      function is able to force a failure. */      function is able to force a failure. */
1773    
1774      case OP_CALLOUT:      case OP_CALLOUT:
1775      if (pcre_callout != NULL)      if (PUBL(callout) != NULL)
1776        {        {
1777        pcre_callout_block cb;        PUBL(callout_block) cb;
1778        cb.version          = 2;   /* Version 1 of the callout block */        cb.version          = 2;   /* Version 1 of the callout block */
1779        cb.callout_number   = ecode[1];        cb.callout_number   = ecode[1];
1780        cb.offset_vector    = md->offset_vector;        cb.offset_vector    = md->offset_vector;
1781    #if defined COMPILE_PCRE8
1782        cb.subject          = (PCRE_SPTR)md->start_subject;        cb.subject          = (PCRE_SPTR)md->start_subject;
1783    #elif defined COMPILE_PCRE16
1784          cb.subject          = (PCRE_SPTR16)md->start_subject;
1785    #elif defined COMPILE_PCRE32
1786          cb.subject          = (PCRE_SPTR32)md->start_subject;
1787    #endif
1788        cb.subject_length   = (int)(md->end_subject - md->start_subject);        cb.subject_length   = (int)(md->end_subject - md->start_subject);
1789        cb.start_match      = (int)(mstart - md->start_subject);        cb.start_match      = (int)(mstart - md->start_subject);
1790        cb.current_position = (int)(eptr - md->start_subject);        cb.current_position = (int)(eptr - md->start_subject);
1791        cb.pattern_position = GET(ecode, 2);        cb.pattern_position = GET(ecode, 2);
1792        cb.next_item_length = GET(ecode, 2 + LINK_SIZE);        cb.next_item_length = GET(ecode, 2 + LINK_SIZE);
1793        cb.capture_top      = offset_top/2;        cb.capture_top      = offset_top/2;
1794        cb.capture_last     = md->capture_last;        cb.capture_last     = md->capture_last & CAPLMASK;
1795          /* Internal change requires this for API compatibility. */
1796          if (cb.capture_last == 0) cb.capture_last = -1;
1797        cb.callout_data     = md->callout_data;        cb.callout_data     = md->callout_data;
1798        cb.mark             = md->nomatch_mark;        cb.mark             = md->nomatch_mark;
1799        if ((rrc = (*pcre_callout)(&cb)) > 0) RRETURN(MATCH_NOMATCH);        if ((rrc = (*PUBL(callout))(&cb)) > 0) RRETURN(MATCH_NOMATCH);
1800        if (rrc < 0) RRETURN(rrc);        if (rrc < 0) RRETURN(rrc);
1801        }        }
1802      ecode += 2 + 2*LINK_SIZE;      ecode += 2 + 2*LINK_SIZE;
# Line 1667  for (;;) Line 1822  for (;;)
1822      case OP_RECURSE:      case OP_RECURSE:
1823        {        {
1824        recursion_info *ri;        recursion_info *ri;
1825        int recno;        unsigned int recno;
1826    
1827        callpat = md->start_code + GET(ecode, 1);        callpat = md->start_code + GET(ecode, 1);
1828        recno = (callpat == md->start_code)? 0 :        recno = (callpat == md->start_code)? 0 :
# Line 1684  for (;;) Line 1839  for (;;)
1839        /* Add to "recursing stack" */        /* Add to "recursing stack" */
1840    
1841        new_recursive.group_num = recno;        new_recursive.group_num = recno;
1842          new_recursive.saved_capture_last = md->capture_last;
1843        new_recursive.subject_position = eptr;        new_recursive.subject_position = eptr;
1844        new_recursive.prevrec = md->recursive;        new_recursive.prevrec = md->recursive;
1845        md->recursive = &new_recursive;        md->recursive = &new_recursive;
# Line 1700  for (;;) Line 1856  for (;;)
1856        else        else
1857          {          {
1858          new_recursive.offset_save =          new_recursive.offset_save =
1859            (int *)(pcre_malloc)(new_recursive.saved_max * sizeof(int));            (int *)(PUBL(malloc))(new_recursive.saved_max * sizeof(int));
1860          if (new_recursive.offset_save == NULL) RRETURN(PCRE_ERROR_NOMEMORY);          if (new_recursive.offset_save == NULL) RRETURN(PCRE_ERROR_NOMEMORY);
1861          }          }
1862        memcpy(new_recursive.offset_save, md->offset_vector,        memcpy(new_recursive.offset_save, md->offset_vector,
1863              new_recursive.saved_max * sizeof(int));              new_recursive.saved_max * sizeof(int));
1864    
1865        /* OK, now we can do the recursion. After processing each alternative,        /* OK, now we can do the recursion. After processing each alternative,
1866        restore the offset data. If there were nested recursions, md->recursive        restore the offset data and the last captured value. If there were nested
1867        might be changed, so reset it before looping. */        recursions, md->recursive might be changed, so reset it before looping.
1868          */
1869    
1870        DPRINTF(("Recursing into group %d\n", new_recursive.group_num));        DPRINTF(("Recursing into group %d\n", new_recursive.group_num));
1871        cbegroup = (*callpat >= OP_SBRA);        cbegroup = (*callpat >= OP_SBRA);
1872        do        do
1873          {          {
1874          if (cbegroup) md->match_function_type = MATCH_CBEGROUP;          if (cbegroup) md->match_function_type = MATCH_CBEGROUP;
1875          RMATCH(eptr, callpat + _pcre_OP_lengths[*callpat], offset_top,          RMATCH(eptr, callpat + PRIV(OP_lengths)[*callpat], offset_top,
1876            md, eptrb, RM6);            md, eptrb, RM6);
1877          memcpy(md->offset_vector, new_recursive.offset_save,          memcpy(md->offset_vector, new_recursive.offset_save,
1878              new_recursive.saved_max * sizeof(int));              new_recursive.saved_max * sizeof(int));
1879            md->capture_last = new_recursive.saved_capture_last;
1880          md->recursive = new_recursive.prevrec;          md->recursive = new_recursive.prevrec;
1881          if (rrc == MATCH_MATCH || rrc == MATCH_ACCEPT)          if (rrc == MATCH_MATCH || rrc == MATCH_ACCEPT)
1882            {            {
1883            DPRINTF(("Recursion matched\n"));            DPRINTF(("Recursion matched\n"));
1884            if (new_recursive.offset_save != stacksave)            if (new_recursive.offset_save != stacksave)
1885              (pcre_free)(new_recursive.offset_save);              (PUBL(free))(new_recursive.offset_save);
1886    
1887            /* 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
1888            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 1893  for (;;)
1893            goto RECURSION_MATCHED;        /* Exit loop; end processing */            goto RECURSION_MATCHED;        /* Exit loop; end processing */
1894            }            }
1895    
1896          /* 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
1897          as NOMATCH. */          recursion; they cause a NOMATCH for the entire recursion. These codes
1898            are defined in a range that can be tested for. */
1899    
1900            if (rrc >= MATCH_BACKTRACK_MIN && rrc <= MATCH_BACKTRACK_MAX)
1901              RRETURN(MATCH_NOMATCH);
1902    
1903            /* Any return code other than NOMATCH is an error. */
1904    
1905          else if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN)          if (rrc != MATCH_NOMATCH)
1906            {            {
1907            DPRINTF(("Recursion gave error %d\n", rrc));            DPRINTF(("Recursion gave error %d\n", rrc));
1908            if (new_recursive.offset_save != stacksave)            if (new_recursive.offset_save != stacksave)
1909              (pcre_free)(new_recursive.offset_save);              (PUBL(free))(new_recursive.offset_save);
1910            RRETURN(rrc);            RRETURN(rrc);
1911            }            }
1912    
# Line 1754  for (;;) Line 1918  for (;;)
1918        DPRINTF(("Recursion didn't match\n"));        DPRINTF(("Recursion didn't match\n"));
1919        md->recursive = new_recursive.prevrec;        md->recursive = new_recursive.prevrec;
1920        if (new_recursive.offset_save != stacksave)        if (new_recursive.offset_save != stacksave)
1921          (pcre_free)(new_recursive.offset_save);          (PUBL(free))(new_recursive.offset_save);
1922        RRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
1923        }        }
1924    
# Line 1868  for (;;) Line 2032  for (;;)
2032    
2033        /* Deal with capturing */        /* Deal with capturing */
2034    
2035        md->capture_last = number;        md->capture_last = (md->capture_last & OVFLMASK) | number;
2036        if (offset >= md->offset_max) md->offset_overflow = TRUE; else        if (offset >= md->offset_max) md->capture_last |= OVFLBIT; else
2037          {          {
2038          /* If offset is greater than offset_top, it means that we are          /* If offset is greater than offset_top, it means that we are
2039          "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 2113  for (;;)
2113          }          }
2114        if (*prev >= OP_SBRA)    /* Could match an empty string */        if (*prev >= OP_SBRA)    /* Could match an empty string */
2115          {          {
         md->match_function_type = MATCH_CBEGROUP;  
2116          RMATCH(eptr, prev, offset_top, md, eptrb, RM50);          RMATCH(eptr, prev, offset_top, md, eptrb, RM50);
2117          RRETURN(rrc);          RRETURN(rrc);
2118          }          }
# Line 1958  for (;;) Line 2121  for (;;)
2121        }        }
2122      else  /* OP_KETRMAX */      else  /* OP_KETRMAX */
2123        {        {
       if (*prev >= OP_SBRA) md->match_function_type = MATCH_CBEGROUP;  
2124        RMATCH(eptr, prev, offset_top, md, eptrb, RM13);        RMATCH(eptr, prev, offset_top, md, eptrb, RM13);
2125        if (rrc == MATCH_ONCE && md->once_target == prev) rrc = MATCH_NOMATCH;        if (rrc == MATCH_ONCE && md->once_target == prev) rrc = MATCH_NOMATCH;
2126        if (rrc != MATCH_NOMATCH) RRETURN(rrc);        if (rrc != MATCH_NOMATCH) RRETURN(rrc);
# Line 2015  for (;;) Line 2177  for (;;)
2177    
2178      case OP_DOLLM:      case OP_DOLLM:
2179      if (eptr < md->end_subject)      if (eptr < md->end_subject)
2180        { if (!IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH); }        {
2181          if (!IS_NEWLINE(eptr))
2182            {
2183            if (md->partial != 0 &&
2184                eptr + 1 >= md->end_subject &&
2185                NLBLOCK->nltype == NLTYPE_FIXED &&
2186                NLBLOCK->nllen == 2 &&
2187                RAWUCHARTEST(eptr) == NLBLOCK->nl[0])
2188              {
2189              md->hitend = TRUE;
2190              if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
2191              }
2192            RRETURN(MATCH_NOMATCH);
2193            }
2194          }
2195      else      else
2196        {        {
2197        if (md->noteol) RRETURN(MATCH_NOMATCH);        if (md->noteol) RRETURN(MATCH_NOMATCH);
# Line 2047  for (;;) Line 2223  for (;;)
2223      ASSERT_NL_OR_EOS:      ASSERT_NL_OR_EOS:
2224      if (eptr < md->end_subject &&      if (eptr < md->end_subject &&
2225          (!IS_NEWLINE(eptr) || eptr != md->end_subject - md->nllen))          (!IS_NEWLINE(eptr) || eptr != md->end_subject - md->nllen))
2226          {
2227          if (md->partial != 0 &&
2228              eptr + 1 >= md->end_subject &&
2229              NLBLOCK->nltype == NLTYPE_FIXED &&
2230              NLBLOCK->nllen == 2 &&
2231              RAWUCHARTEST(eptr) == NLBLOCK->nl[0])
2232            {
2233            md->hitend = TRUE;
2234            if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
2235            }
2236        RRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2237          }
2238    
2239      /* Either at end of string or \n before end. */      /* Either at end of string or \n before end. */
2240    
# Line 2066  for (;;) Line 2253  for (;;)
2253        be "non-word" characters. Remember the earliest consulted character for        be "non-word" characters. Remember the earliest consulted character for
2254        partial matching. */        partial matching. */
2255    
2256  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
2257        if (utf8)        if (utf)
2258          {          {
2259          /* Get status of previous character */          /* Get status of previous character */
2260    
2261          if (eptr == md->start_subject) prev_is_word = FALSE; else          if (eptr == md->start_subject) prev_is_word = FALSE; else
2262            {            {
2263            USPTR lastptr = eptr - 1;            PCRE_PUCHAR lastptr = eptr - 1;
2264            while((*lastptr & 0xc0) == 0x80) lastptr--;            BACKCHAR(lastptr);
2265            if (lastptr < md->start_used_ptr) md->start_used_ptr = lastptr;            if (lastptr < md->start_used_ptr) md->start_used_ptr = lastptr;
2266            GETCHAR(c, lastptr);            GETCHAR(c, lastptr);
2267  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
# Line 2139  for (;;) Line 2326  for (;;)
2326              }              }
2327            else            else
2328  #endif  #endif
2329            prev_is_word = ((md->ctypes[eptr[-1]] & ctype_word) != 0);            prev_is_word = MAX_255(eptr[-1])
2330                && ((md->ctypes[eptr[-1]] & ctype_word) != 0);
2331            }            }
2332    
2333          /* Get status of next character */          /* Get status of next character */
# Line 2162  for (;;) Line 2350  for (;;)
2350            }            }
2351          else          else
2352  #endif  #endif
2353          cur_is_word = ((md->ctypes[*eptr] & ctype_word) != 0);          cur_is_word = MAX_255(*eptr)
2354              && ((md->ctypes[*eptr] & ctype_word) != 0);
2355          }          }
2356    
2357        /* Now see if the situation is what we want */        /* Now see if the situation is what we want */
# Line 2173  for (;;) Line 2362  for (;;)
2362        }        }
2363      break;      break;
2364    
2365      /* Match a single character type; inline for speed */      /* Match any single character type except newline; have to take care with
2366        CRLF newlines and partial matching. */
2367    
2368      case OP_ANY:      case OP_ANY:
2369      if (IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH);      if (IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH);
2370        if (md->partial != 0 &&
2371            eptr + 1 >= md->end_subject &&
2372            NLBLOCK->nltype == NLTYPE_FIXED &&
2373            NLBLOCK->nllen == 2 &&
2374            RAWUCHARTEST(eptr) == NLBLOCK->nl[0])
2375          {
2376          md->hitend = TRUE;
2377          if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
2378          }
2379    
2380      /* Fall through */      /* Fall through */
2381    
2382        /* Match any single character whatsoever. */
2383    
2384      case OP_ALLANY:      case OP_ALLANY:
2385      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 */
2386        {                            /* not be updated before SCHECK_PARTIAL. */        {                            /* not be updated before SCHECK_PARTIAL. */
# Line 2186  for (;;) Line 2388  for (;;)
2388        RRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2389        }        }
2390      eptr++;      eptr++;
2391      if (utf8) while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;  #ifdef SUPPORT_UTF
2392        if (utf) ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);
2393    #endif
2394      ecode++;      ecode++;
2395      break;      break;
2396    
# Line 2211  for (;;) Line 2415  for (;;)
2415        }        }
2416      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2417      if (      if (
2418  #ifdef SUPPORT_UTF8  #if defined SUPPORT_UTF || !(defined COMPILE_PCRE8)
2419         c < 256 &&         c < 256 &&
2420  #endif  #endif
2421         (md->ctypes[c] & ctype_digit) != 0         (md->ctypes[c] & ctype_digit) != 0
# Line 2228  for (;;) Line 2432  for (;;)
2432        }        }
2433      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2434      if (      if (
2435  #ifdef SUPPORT_UTF8  #if defined SUPPORT_UTF || !(defined COMPILE_PCRE8)
2436         c >= 256 ||         c > 255 ||
2437  #endif  #endif
2438         (md->ctypes[c] & ctype_digit) == 0         (md->ctypes[c] & ctype_digit) == 0
2439         )         )
# Line 2245  for (;;) Line 2449  for (;;)
2449        }        }
2450      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2451      if (      if (
2452  #ifdef SUPPORT_UTF8  #if defined SUPPORT_UTF || !(defined COMPILE_PCRE8)
2453         c < 256 &&         c < 256 &&
2454  #endif  #endif
2455         (md->ctypes[c] & ctype_space) != 0         (md->ctypes[c] & ctype_space) != 0
# Line 2262  for (;;) Line 2466  for (;;)
2466        }        }
2467      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2468      if (      if (
2469  #ifdef SUPPORT_UTF8  #if defined SUPPORT_UTF || !(defined COMPILE_PCRE8)
2470         c >= 256 ||         c > 255 ||
2471  #endif  #endif
2472         (md->ctypes[c] & ctype_space) == 0         (md->ctypes[c] & ctype_space) == 0
2473         )         )
# Line 2279  for (;;) Line 2483  for (;;)
2483        }        }
2484      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2485      if (      if (
2486  #ifdef SUPPORT_UTF8  #if defined SUPPORT_UTF || !(defined COMPILE_PCRE8)
2487         c < 256 &&         c < 256 &&
2488  #endif  #endif
2489         (md->ctypes[c] & ctype_word) != 0         (md->ctypes[c] & ctype_word) != 0
# Line 2296  for (;;) Line 2500  for (;;)
2500        }        }
2501      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2502      if (      if (
2503  #ifdef SUPPORT_UTF8  #if defined SUPPORT_UTF || !(defined COMPILE_PCRE8)
2504         c >= 256 ||         c > 255 ||
2505  #endif  #endif
2506         (md->ctypes[c] & ctype_word) == 0         (md->ctypes[c] & ctype_word) == 0
2507         )         )
# Line 2316  for (;;) Line 2520  for (;;)
2520        {        {
2521        default: RRETURN(MATCH_NOMATCH);        default: RRETURN(MATCH_NOMATCH);
2522    
2523        case 0x000d:        case CHAR_CR:
2524        if (eptr < md->end_subject && *eptr == 0x0a) eptr++;        if (eptr >= md->end_subject)
2525            {
2526            SCHECK_PARTIAL();
2527            }
2528          else if (RAWUCHARTEST(eptr) == CHAR_LF) eptr++;
2529        break;        break;
2530    
2531        case 0x000a:        case CHAR_LF:
2532        break;        break;
2533    
2534        case 0x000b:        case CHAR_VT:
2535        case 0x000c:        case CHAR_FF:
2536        case 0x0085:        case CHAR_NEL:
2537    #ifndef EBCDIC
2538        case 0x2028:        case 0x2028:
2539        case 0x2029:        case 0x2029:
2540    #endif  /* Not EBCDIC */
2541        if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);        if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);
2542        break;        break;
2543        }        }
# Line 2343  for (;;) Line 2553  for (;;)
2553      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2554      switch(c)      switch(c)
2555        {        {
2556          HSPACE_CASES: RRETURN(MATCH_NOMATCH);  /* Byte and multibyte cases */
2557        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);  
2558        }        }
2559      ecode++;      ecode++;
2560      break;      break;
# Line 2377  for (;;) Line 2568  for (;;)
2568      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2569      switch(c)      switch(c)
2570        {        {
2571          HSPACE_CASES: break;  /* Byte and multibyte cases */
2572        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;  
2573        }        }
2574      ecode++;      ecode++;
2575      break;      break;
# Line 2411  for (;;) Line 2583  for (;;)
2583      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2584      switch(c)      switch(c)
2585        {        {
2586          VSPACE_CASES: RRETURN(MATCH_NOMATCH);
2587        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);  
2588        }        }
2589      ecode++;      ecode++;
2590      break;      break;
# Line 2433  for (;;) Line 2598  for (;;)
2598      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2599      switch(c)      switch(c)
2600        {        {
2601          VSPACE_CASES: break;
2602        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;  
2603        }        }
2604      ecode++;      ecode++;
2605      break;      break;
# Line 2459  for (;;) Line 2617  for (;;)
2617        }        }
2618      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2619        {        {
2620          const pcre_uint32 *cp;
2621        const ucd_record *prop = GET_UCD(c);        const ucd_record *prop = GET_UCD(c);
2622    
2623        switch(ecode[1])        switch(ecode[1])
# Line 2475  for (;;) Line 2634  for (;;)
2634          break;          break;
2635    
2636          case PT_GC:          case PT_GC:
2637          if ((ecode[2] != _pcre_ucp_gentype[prop->chartype]) == (op == OP_PROP))          if ((ecode[2] != PRIV(ucp_gentype)[prop->chartype]) == (op == OP_PROP))
2638            RRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
2639          break;          break;
2640    
# Line 2492  for (;;) Line 2651  for (;;)
2651          /* These are specials */          /* These are specials */
2652    
2653          case PT_ALNUM:          case PT_ALNUM:
2654          if ((_pcre_ucp_gentype[prop->chartype] == ucp_L ||          if ((PRIV(ucp_gentype)[prop->chartype] == ucp_L ||
2655               _pcre_ucp_gentype[prop->chartype] == ucp_N) == (op == OP_NOTPROP))               PRIV(ucp_gentype)[prop->chartype] == ucp_N) == (op == OP_NOTPROP))
2656            RRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
2657          break;          break;
2658    
2659          case PT_SPACE:    /* Perl space */          case PT_SPACE:    /* Perl space */
2660          if ((_pcre_ucp_gentype[prop->chartype] == ucp_Z ||          if ((PRIV(ucp_gentype)[prop->chartype] == ucp_Z ||
2661               c == CHAR_HT || c == CHAR_NL || c == CHAR_FF || c == CHAR_CR)               c == CHAR_HT || c == CHAR_NL || c == CHAR_FF || c == CHAR_CR)
2662                 == (op == OP_NOTPROP))                 == (op == OP_NOTPROP))
2663            RRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
2664          break;          break;
2665    
2666          case PT_PXSPACE:  /* POSIX space */          case PT_PXSPACE:  /* POSIX space */
2667          if ((_pcre_ucp_gentype[prop->chartype] == ucp_Z ||          if ((PRIV(ucp_gentype)[prop->chartype] == ucp_Z ||
2668               c == CHAR_HT || c == CHAR_NL || c == CHAR_VT ||               c == CHAR_HT || c == CHAR_NL || c == CHAR_VT ||
2669               c == CHAR_FF || c == CHAR_CR)               c == CHAR_FF || c == CHAR_CR)
2670                 == (op == OP_NOTPROP))                 == (op == OP_NOTPROP))
# Line 2513  for (;;) Line 2672  for (;;)
2672          break;          break;
2673    
2674          case PT_WORD:          case PT_WORD:
2675          if ((_pcre_ucp_gentype[prop->chartype] == ucp_L ||          if ((PRIV(ucp_gentype)[prop->chartype] == ucp_L ||
2676               _pcre_ucp_gentype[prop->chartype] == ucp_N ||               PRIV(ucp_gentype)[prop->chartype] == ucp_N ||
2677               c == CHAR_UNDERSCORE) == (op == OP_NOTPROP))               c == CHAR_UNDERSCORE) == (op == OP_NOTPROP))
2678            RRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
2679          break;          break;
2680    
2681            case PT_CLIST:
2682            cp = PRIV(ucd_caseless_sets) + ecode[2];
2683            for (;;)
2684              {
2685              if (c < *cp)
2686                { if (op == OP_PROP) { RRETURN(MATCH_NOMATCH); } else break; }
2687              if (c == *cp++)
2688                { if (op == OP_PROP) break; else { RRETURN(MATCH_NOMATCH); } }
2689              }
2690            break;
2691    
2692            case PT_UCNC:
2693            if ((c == CHAR_DOLLAR_SIGN || c == CHAR_COMMERCIAL_AT ||
2694                 c == CHAR_GRAVE_ACCENT || (c >= 0xa0 && c <= 0xd7ff) ||
2695                 c >= 0xe000) == (op == OP_NOTPROP))
2696              RRETURN(MATCH_NOMATCH);
2697            break;
2698    
2699          /* This should never occur */          /* This should never occur */
2700    
2701          default:          default:
# Line 2538  for (;;) Line 2715  for (;;)
2715        SCHECK_PARTIAL();        SCHECK_PARTIAL();
2716        RRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2717        }        }
2718      GETCHARINCTEST(c, eptr);      else
     if (UCD_CATEGORY(c) == ucp_M) RRETURN(MATCH_NOMATCH);  
     while (eptr < md->end_subject)  
2719        {        {
2720        int len = 1;        int lgb, rgb;
2721        if (!utf8) c = *eptr; else { GETCHARLEN(c, eptr, len); }        GETCHARINCTEST(c, eptr);
2722        if (UCD_CATEGORY(c) != ucp_M) break;        lgb = UCD_GRAPHBREAK(c);
2723        eptr += len;        while (eptr < md->end_subject)
2724            {
2725            int len = 1;
2726            if (!utf) c = *eptr; else { GETCHARLEN(c, eptr, len); }
2727            rgb = UCD_GRAPHBREAK(c);
2728            if ((PRIV(ucp_gbtable)[lgb] & (1 << rgb)) == 0) break;
2729            lgb = rgb;
2730            eptr += len;
2731            }
2732        }        }
2733        CHECK_PARTIAL();
2734      ecode++;      ecode++;
2735      break;      break;
2736  #endif  #endif  /* SUPPORT_UCP */
2737    
2738    
2739      /* 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 2748  for (;;)
2748      case OP_REFI:      case OP_REFI:
2749      caseless = op == OP_REFI;      caseless = op == OP_REFI;
2750      offset = GET2(ecode, 1) << 1;               /* Doubled ref number */      offset = GET2(ecode, 1) << 1;               /* Doubled ref number */
2751      ecode += 3;      ecode += 1 + IMM2_SIZE;
2752    
2753      /* If the reference is unset, there are two possibilities:      /* If the reference is unset, there are two possibilities:
2754    
# Line 2604  for (;;) Line 2788  for (;;)
2788        case OP_CRMINRANGE:        case OP_CRMINRANGE:
2789        minimize = (*ecode == OP_CRMINRANGE);        minimize = (*ecode == OP_CRMINRANGE);
2790        min = GET2(ecode, 1);        min = GET2(ecode, 1);
2791        max = GET2(ecode, 3);        max = GET2(ecode, 1 + IMM2_SIZE);
2792        if (max == 0) max = INT_MAX;        if (max == 0) max = INT_MAX;
2793        ecode += 5;        ecode += 1 + 2 * IMM2_SIZE;
2794        break;        break;
2795    
2796        default:               /* No repeat follows */        default:               /* No repeat follows */
2797        if ((length = match_ref(offset, eptr, length, md, caseless)) < 0)        if ((length = match_ref(offset, eptr, length, md, caseless)) < 0)
2798          {          {
2799            if (length == -2) eptr = md->end_subject;   /* Partial match */
2800          CHECK_PARTIAL();          CHECK_PARTIAL();
2801          RRETURN(MATCH_NOMATCH);          RRETURN(MATCH_NOMATCH);
2802          }          }
# Line 2620  for (;;) Line 2805  for (;;)
2805        }        }
2806    
2807      /* Handle repeated back references. If the length of the reference is      /* Handle repeated back references. If the length of the reference is
2808      zero, just continue with the main loop. */      zero, just continue with the main loop. If the length is negative, it
2809        means the reference is unset in non-Java-compatible mode. If the minimum is
2810        zero, we can continue at the same level without recursion. For any other
2811        minimum, carrying on will result in NOMATCH. */
2812    
2813      if (length == 0) continue;      if (length == 0) continue;
2814        if (length < 0 && min == 0) continue;
2815    
2816      /* First, ensure the minimum number of matches are present. We get back      /* First, ensure the minimum number of matches are present. We get back
2817      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 2822  for (;;)
2822        int slength;        int slength;
2823        if ((slength = match_ref(offset, eptr, length, md, caseless)) < 0)        if ((slength = match_ref(offset, eptr, length, md, caseless)) < 0)
2824          {          {
2825            if (slength == -2) eptr = md->end_subject;   /* Partial match */
2826          CHECK_PARTIAL();          CHECK_PARTIAL();
2827          RRETURN(MATCH_NOMATCH);          RRETURN(MATCH_NOMATCH);
2828          }          }
# Line 2656  for (;;) Line 2846  for (;;)
2846          if (fi >= max) RRETURN(MATCH_NOMATCH);          if (fi >= max) RRETURN(MATCH_NOMATCH);
2847          if ((slength = match_ref(offset, eptr, length, md, caseless)) < 0)          if ((slength = match_ref(offset, eptr, length, md, caseless)) < 0)
2848            {            {
2849              if (slength == -2) eptr = md->end_subject;   /* Partial match */
2850            CHECK_PARTIAL();            CHECK_PARTIAL();
2851            RRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
2852            }            }
# Line 2674  for (;;) Line 2865  for (;;)
2865          int slength;          int slength;
2866          if ((slength = match_ref(offset, eptr, length, md, caseless)) < 0)          if ((slength = match_ref(offset, eptr, length, md, caseless)) < 0)
2867            {            {
2868            CHECK_PARTIAL();            /* Can't use CHECK_PARTIAL because we don't want to update eptr in
2869              the soft partial matching case. */
2870    
2871              if (slength == -2 && md->partial != 0 &&
2872                  md->end_subject > md->start_used_ptr)
2873                {
2874                md->hitend = TRUE;
2875                if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
2876                }
2877            break;            break;
2878            }            }
2879          eptr += slength;          eptr += slength;
2880          }          }
2881    
2882        while (eptr >= pp)        while (eptr >= pp)
2883          {          {
2884          RMATCH(eptr, ecode, offset_top, md, eptrb, RM15);          RMATCH(eptr, ecode, offset_top, md, eptrb, RM15);
# Line 2703  for (;;) Line 2903  for (;;)
2903      case OP_NCLASS:      case OP_NCLASS:
2904      case OP_CLASS:      case OP_CLASS:
2905        {        {
2906          /* The data variable is saved across frames, so the byte map needs to
2907          be stored there. */
2908    #define BYTE_MAP ((pcre_uint8 *)data)
2909        data = ecode + 1;                /* Save for matching */        data = ecode + 1;                /* Save for matching */
2910        ecode += 33;                     /* Advance past the item */        ecode += 1 + (32 / sizeof(pcre_uchar)); /* Advance past the item */
2911    
2912        switch (*ecode)        switch (*ecode)
2913          {          {
# Line 2725  for (;;) Line 2928  for (;;)
2928          case OP_CRMINRANGE:          case OP_CRMINRANGE:
2929          minimize = (*ecode == OP_CRMINRANGE);          minimize = (*ecode == OP_CRMINRANGE);
2930          min = GET2(ecode, 1);          min = GET2(ecode, 1);
2931          max = GET2(ecode, 3);          max = GET2(ecode, 1 + IMM2_SIZE);
2932          if (max == 0) max = INT_MAX;          if (max == 0) max = INT_MAX;
2933          ecode += 5;          ecode += 1 + 2 * IMM2_SIZE;
2934          break;          break;
2935    
2936          default:               /* No repeat follows */          default:               /* No repeat follows */
# Line 2737  for (;;) Line 2940  for (;;)
2940    
2941        /* First, ensure the minimum number of matches are present. */        /* First, ensure the minimum number of matches are present. */
2942    
2943  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
2944        /* UTF-8 mode */        if (utf)
       if (utf8)  
2945          {          {
2946          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
2947            {            {
# Line 2754  for (;;) Line 2956  for (;;)
2956              if (op == OP_CLASS) RRETURN(MATCH_NOMATCH);              if (op == OP_CLASS) RRETURN(MATCH_NOMATCH);
2957              }              }
2958            else            else
2959              {              if ((BYTE_MAP[c/8] & (1 << (c&7))) == 0) RRETURN(MATCH_NOMATCH);
             if ((data[c/8] & (1 << (c&7))) == 0) RRETURN(MATCH_NOMATCH);  
             }  
2960            }            }
2961          }          }
2962        else        else
2963  #endif  #endif
2964        /* Not UTF-8 mode */        /* Not UTF mode */
2965          {          {
2966          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
2967            {            {
# Line 2771  for (;;) Line 2971  for (;;)
2971              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
2972              }              }
2973            c = *eptr++;            c = *eptr++;
2974            if ((data[c/8] & (1 << (c&7))) == 0) RRETURN(MATCH_NOMATCH);  #ifndef COMPILE_PCRE8
2975              if (c > 255)
2976                {
2977                if (op == OP_CLASS) RRETURN(MATCH_NOMATCH);
2978                }
2979              else
2980    #endif
2981                if ((BYTE_MAP[c/8] & (1 << (c&7))) == 0) RRETURN(MATCH_NOMATCH);
2982            }            }
2983          }          }
2984    
# Line 2785  for (;;) Line 2992  for (;;)
2992    
2993        if (minimize)        if (minimize)
2994          {          {
2995  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
2996          /* UTF-8 mode */          if (utf)
         if (utf8)  
2997            {            {
2998            for (fi = min;; fi++)            for (fi = min;; fi++)
2999              {              {
# Line 2805  for (;;) Line 3011  for (;;)
3011                if (op == OP_CLASS) RRETURN(MATCH_NOMATCH);                if (op == OP_CLASS) RRETURN(MATCH_NOMATCH);
3012                }                }
3013              else              else
3014                {                if ((BYTE_MAP[c/8] & (1 << (c&7))) == 0) RRETURN(MATCH_NOMATCH);
               if ((data[c/8] & (1 << (c&7))) == 0) RRETURN(MATCH_NOMATCH);  
               }  
3015              }              }
3016            }            }
3017          else          else
3018  #endif  #endif
3019          /* Not UTF-8 mode */          /* Not UTF mode */
3020            {            {
3021            for (fi = min;; fi++)            for (fi = min;; fi++)
3022              {              {
# Line 2825  for (;;) Line 3029  for (;;)
3029                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
3030                }                }
3031              c = *eptr++;              c = *eptr++;
3032              if ((data[c/8] & (1 << (c&7))) == 0) RRETURN(MATCH_NOMATCH);  #ifndef COMPILE_PCRE8
3033                if (c > 255)
3034                  {
3035                  if (op == OP_CLASS) RRETURN(MATCH_NOMATCH);
3036                  }
3037                else
3038    #endif
3039                  if ((BYTE_MAP[c/8] & (1 << (c&7))) == 0) RRETURN(MATCH_NOMATCH);
3040              }              }
3041            }            }
3042          /* Control never gets here */          /* Control never gets here */
# Line 2837  for (;;) Line 3048  for (;;)
3048          {          {
3049          pp = eptr;          pp = eptr;
3050    
3051  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
3052          /* UTF-8 mode */          if (utf)
         if (utf8)  
3053            {            {
3054            for (i = min; i < max; i++)            for (i = min; i < max; i++)
3055              {              {
# Line 2855  for (;;) Line 3065  for (;;)
3065                if (op == OP_CLASS) break;                if (op == OP_CLASS) break;
3066                }                }
3067              else              else
3068                {                if ((BYTE_MAP[c/8] & (1 << (c&7))) == 0) break;
               if ((data[c/8] & (1 << (c&7))) == 0) break;  
               }  
3069              eptr += len;              eptr += len;
3070              }              }
3071            for (;;)            for (;;)
# Line 2870  for (;;) Line 3078  for (;;)
3078            }            }
3079          else          else
3080  #endif  #endif
3081            /* Not UTF-8 mode */            /* Not UTF mode */
3082            {            {
3083            for (i = min; i < max; i++)            for (i = min; i < max; i++)
3084              {              {
# Line 2880  for (;;) Line 3088  for (;;)
3088                break;                break;
3089                }                }
3090              c = *eptr;              c = *eptr;
3091              if ((data[c/8] & (1 << (c&7))) == 0) break;  #ifndef COMPILE_PCRE8
3092                if (c > 255)
3093                  {
3094                  if (op == OP_CLASS) break;
3095                  }
3096                else
3097    #endif
3098                  if ((BYTE_MAP[c/8] & (1 << (c&7))) == 0) break;
3099              eptr++;              eptr++;
3100              }              }
3101            while (eptr >= pp)            while (eptr >= pp)
# Line 2893  for (;;) Line 3108  for (;;)
3108    
3109          RRETURN(MATCH_NOMATCH);          RRETURN(MATCH_NOMATCH);
3110          }          }
3111    #undef BYTE_MAP
3112        }        }
3113      /* Control never gets here */      /* Control never gets here */
3114    
# Line 2901  for (;;) Line 3117  for (;;)
3117      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
3118      mode, because Unicode properties are supported in non-UTF-8 mode. */      mode, because Unicode properties are supported in non-UTF-8 mode. */
3119    
3120  #ifdef SUPPORT_UTF8  #if defined SUPPORT_UTF || !defined COMPILE_PCRE8
3121      case OP_XCLASS:      case OP_XCLASS:
3122        {        {
3123        data = ecode + 1 + LINK_SIZE;                /* Save for matching */        data = ecode + 1 + LINK_SIZE;                /* Save for matching */
# Line 2926  for (;;) Line 3142  for (;;)
3142          case OP_CRMINRANGE:          case OP_CRMINRANGE:
3143          minimize = (*ecode == OP_CRMINRANGE);          minimize = (*ecode == OP_CRMINRANGE);
3144          min = GET2(ecode, 1);          min = GET2(ecode, 1);
3145          max = GET2(ecode, 3);          max = GET2(ecode, 1 + IMM2_SIZE);
3146          if (max == 0) max = INT_MAX;          if (max == 0) max = INT_MAX;
3147          ecode += 5;          ecode += 1 + 2 * IMM2_SIZE;
3148          break;          break;
3149    
3150          default:               /* No repeat follows */          default:               /* No repeat follows */
# Line 2946  for (;;) Line 3162  for (;;)
3162            RRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
3163            }            }
3164          GETCHARINCTEST(c, eptr);          GETCHARINCTEST(c, eptr);
3165          if (!_pcre_xclass(c, data)) RRETURN(MATCH_NOMATCH);          if (!PRIV(xclass)(c, data, utf)) RRETURN(MATCH_NOMATCH);
3166          }          }
3167    
3168        /* 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 3186  for (;;)
3186              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
3187              }              }
3188            GETCHARINCTEST(c, eptr);            GETCHARINCTEST(c, eptr);
3189            if (!_pcre_xclass(c, data)) RRETURN(MATCH_NOMATCH);            if (!PRIV(xclass)(c, data, utf)) RRETURN(MATCH_NOMATCH);
3190            }            }
3191          /* Control never gets here */          /* Control never gets here */
3192          }          }
# Line 2988  for (;;) Line 3204  for (;;)
3204              SCHECK_PARTIAL();              SCHECK_PARTIAL();
3205              break;              break;
3206              }              }
3207    #ifdef SUPPORT_UTF
3208            GETCHARLENTEST(c, eptr, len);            GETCHARLENTEST(c, eptr, len);
3209            if (!_pcre_xclass(c, data)) break;  #else
3210              c = *eptr;
3211    #endif
3212              if (!PRIV(xclass)(c, data, utf)) break;
3213            eptr += len;            eptr += len;
3214            }            }
3215          for(;;)          for(;;)
# Line 2997  for (;;) Line 3217  for (;;)
3217            RMATCH(eptr, ecode, offset_top, md, eptrb, RM21);            RMATCH(eptr, ecode, offset_top, md, eptrb, RM21);
3218            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3219            if (eptr-- == pp) break;        /* Stop if tried at original pos */            if (eptr-- == pp) break;        /* Stop if tried at original pos */
3220            if (utf8) BACKCHAR(eptr);  #ifdef SUPPORT_UTF
3221              if (utf) BACKCHAR(eptr);
3222    #endif
3223            }            }
3224          RRETURN(MATCH_NOMATCH);          RRETURN(MATCH_NOMATCH);
3225          }          }
# Line 3009  for (;;) Line 3231  for (;;)
3231      /* Match a single character, casefully */      /* Match a single character, casefully */
3232    
3233      case OP_CHAR:      case OP_CHAR:
3234  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
3235      if (utf8)      if (utf)
3236        {        {
3237        length = 1;        length = 1;
3238        ecode++;        ecode++;
# Line 3020  for (;;) Line 3242  for (;;)
3242          CHECK_PARTIAL();             /* Not SCHECK_PARTIAL() */          CHECK_PARTIAL();             /* Not SCHECK_PARTIAL() */
3243          RRETURN(MATCH_NOMATCH);          RRETURN(MATCH_NOMATCH);
3244          }          }
3245        while (length-- > 0) if (*ecode++ != *eptr++) RRETURN(MATCH_NOMATCH);        while (length-- > 0) if (*ecode++ != RAWUCHARINC(eptr)) RRETURN(MATCH_NOMATCH);
3246        }        }
3247      else      else
3248  #endif  #endif
3249        /* Not UTF mode */
     /* Non-UTF-8 mode */  
3250        {        {
3251        if (md->end_subject - eptr < 1)        if (md->end_subject - eptr < 1)
3252          {          {
# Line 3037  for (;;) Line 3258  for (;;)
3258        }        }
3259      break;      break;
3260    
3261      /* Match a single character, caselessly */      /* Match a single character, caselessly. If we are at the end of the
3262        subject, give up immediately. */
3263    
3264      case OP_CHARI:      case OP_CHARI:
3265  #ifdef SUPPORT_UTF8      if (eptr >= md->end_subject)
3266      if (utf8)        {
3267          SCHECK_PARTIAL();
3268          RRETURN(MATCH_NOMATCH);
3269          }
3270    
3271    #ifdef SUPPORT_UTF
3272        if (utf)
3273        {        {
3274        length = 1;        length = 1;
3275        ecode++;        ecode++;
3276        GETCHARLEN(fc, ecode, length);        GETCHARLEN(fc, ecode, length);
3277    
       if (length > md->end_subject - eptr)  
         {  
         CHECK_PARTIAL();             /* Not SCHECK_PARTIAL() */  
         RRETURN(MATCH_NOMATCH);  
         }  
   
3278        /* 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
3279        can use the fast lookup table. */        we know that its other case must also be one byte long, so we can use the
3280          fast lookup table. We know that there is at least one byte left in the
3281          subject. */
3282    
3283        if (fc < 128)        if (fc < 128)
3284          {          {
3285          if (md->lcc[*ecode++] != md->lcc[*eptr++]) RRETURN(MATCH_NOMATCH);          pcre_uint32 cc = RAWUCHAR(eptr);
3286            if (md->lcc[fc] != TABLE_GET(cc, md->lcc, cc)) RRETURN(MATCH_NOMATCH);
3287            ecode++;
3288            eptr++;
3289          }          }
3290    
3291        /* Otherwise we must pick up the subject character */        /* Otherwise we must pick up the subject character. Note that we cannot
3292          use the value of "length" to check for sufficient bytes left, because the
3293          other case of the character may have more or fewer bytes.  */
3294    
3295        else        else
3296          {          {
3297          unsigned int dc;          pcre_uint32 dc;
3298          GETCHARINC(dc, eptr);          GETCHARINC(dc, eptr);
3299          ecode += length;          ecode += length;
3300    
# Line 3082  for (;;) Line 3311  for (;;)
3311          }          }
3312        }        }
3313      else      else
3314  #endif   /* SUPPORT_UTF8 */  #endif   /* SUPPORT_UTF */
3315    
3316      /* Non-UTF-8 mode */      /* Not UTF mode */
3317        {        {
3318        if (md->end_subject - eptr < 1)        if (TABLE_GET(ecode[1], md->lcc, ecode[1])
3319          {            != TABLE_GET(*eptr, md->lcc, *eptr)) RRETURN(MATCH_NOMATCH);
3320          SCHECK_PARTIAL();            /* This one can use SCHECK_PARTIAL() */        eptr++;
         RRETURN(MATCH_NOMATCH);  
         }  
       if (md->lcc[ecode[1]] != md->lcc[*eptr++]) RRETURN(MATCH_NOMATCH);  
3321        ecode += 2;        ecode += 2;
3322        }        }
3323      break;      break;
# Line 3101  for (;;) Line 3327  for (;;)
3327      case OP_EXACT:      case OP_EXACT:
3328      case OP_EXACTI:      case OP_EXACTI:
3329      min = max = GET2(ecode, 1);      min = max = GET2(ecode, 1);
3330      ecode += 3;      ecode += 1 + IMM2_SIZE;
3331      goto REPEATCHAR;      goto REPEATCHAR;
3332    
3333      case OP_POSUPTO:      case OP_POSUPTO:
# Line 3116  for (;;) Line 3342  for (;;)
3342      min = 0;      min = 0;
3343      max = GET2(ecode, 1);      max = GET2(ecode, 1);
3344      minimize = *ecode == OP_MINUPTO || *ecode == OP_MINUPTOI;      minimize = *ecode == OP_MINUPTO || *ecode == OP_MINUPTOI;
3345      ecode += 3;      ecode += 1 + IMM2_SIZE;
3346      goto REPEATCHAR;      goto REPEATCHAR;
3347    
3348      case OP_POSSTAR:      case OP_POSSTAR:
# Line 3164  for (;;) Line 3390  for (;;)
3390      /* Common code for all repeated single-character matches. */      /* Common code for all repeated single-character matches. */
3391    
3392      REPEATCHAR:      REPEATCHAR:
3393  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
3394      if (utf8)      if (utf)
3395        {        {
3396        length = 1;        length = 1;
3397        charptr = ecode;        charptr = ecode;
# Line 3178  for (;;) Line 3404  for (;;)
3404        if (length > 1)        if (length > 1)
3405          {          {
3406  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
3407          unsigned int othercase;          pcre_uint32 othercase;
3408          if (op >= OP_STARI &&     /* Caseless */          if (op >= OP_STARI &&     /* Caseless */
3409              (othercase = UCD_OTHERCASE(fc)) != fc)              (othercase = UCD_OTHERCASE(fc)) != fc)
3410            oclength = _pcre_ord2utf8(othercase, occhars);            oclength = PRIV(ord2utf)(othercase, occhars);
3411          else oclength = 0;          else oclength = 0;
3412  #endif  /* SUPPORT_UCP */  #endif  /* SUPPORT_UCP */
3413    
3414          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
3415            {            {
3416            if (eptr <= md->end_subject - length &&            if (eptr <= md->end_subject - length &&
3417              memcmp(eptr, charptr, length) == 0) eptr += length;              memcmp(eptr, charptr, IN_UCHARS(length)) == 0) eptr += length;
3418  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
3419            else if (oclength > 0 &&            else if (oclength > 0 &&
3420                     eptr <= md->end_subject - oclength &&                     eptr <= md->end_subject - oclength &&
3421                     memcmp(eptr, occhars, oclength) == 0) eptr += oclength;                     memcmp(eptr, occhars, IN_UCHARS(oclength)) == 0) eptr += oclength;
3422  #endif  /* SUPPORT_UCP */  #endif  /* SUPPORT_UCP */
3423            else            else
3424              {              {
# Line 3211  for (;;) Line 3437  for (;;)
3437              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3438              if (fi >= max) RRETURN(MATCH_NOMATCH);              if (fi >= max) RRETURN(MATCH_NOMATCH);
3439              if (eptr <= md->end_subject - length &&              if (eptr <= md->end_subject - length &&
3440                memcmp(eptr, charptr, length) == 0) eptr += length;                memcmp(eptr, charptr, IN_UCHARS(length)) == 0) eptr += length;
3441  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
3442              else if (oclength > 0 &&              else if (oclength > 0 &&
3443                       eptr <= md->end_subject - oclength &&                       eptr <= md->end_subject - oclength &&
3444                       memcmp(eptr, occhars, oclength) == 0) eptr += oclength;                       memcmp(eptr, occhars, IN_UCHARS(oclength)) == 0) eptr += oclength;
3445  #endif  /* SUPPORT_UCP */  #endif  /* SUPPORT_UCP */
3446              else              else
3447                {                {
# Line 3232  for (;;) Line 3458  for (;;)
3458            for (i = min; i < max; i++)            for (i = min; i < max; i++)
3459              {              {
3460              if (eptr <= md->end_subject - length &&              if (eptr <= md->end_subject - length &&
3461                  memcmp(eptr, charptr, length) == 0) eptr += length;                  memcmp(eptr, charptr, IN_UCHARS(length)) == 0) eptr += length;
3462  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
3463              else if (oclength > 0 &&              else if (oclength > 0 &&
3464                       eptr <= md->end_subject - oclength &&                       eptr <= md->end_subject - oclength &&
3465                       memcmp(eptr, occhars, oclength) == 0) eptr += oclength;                       memcmp(eptr, occhars, IN_UCHARS(oclength)) == 0) eptr += oclength;
3466  #endif  /* SUPPORT_UCP */  #endif  /* SUPPORT_UCP */
3467              else              else
3468                {                {
# Line 3268  for (;;) Line 3494  for (;;)
3494        value of fc will always be < 128. */        value of fc will always be < 128. */
3495        }        }
3496      else      else
3497  #endif  /* SUPPORT_UTF8 */  #endif  /* SUPPORT_UTF */
3498          /* When not in UTF-8 mode, load a single-byte character. */
3499      /* When not in UTF-8 mode, load a single-byte character. */        fc = *ecode++;
3500    
3501      fc = *ecode++;      /* The value of fc at this point is always one character, though we may
3502        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  
3503      caseful cases, for speed, since matching characters is likely to be quite      caseful cases, for speed, since matching characters is likely to be quite
3504      common. First, ensure the minimum number of matches are present. If min =      common. First, ensure the minimum number of matches are present. If min =
3505      max, continue at the same level without recursing. Otherwise, if      max, continue at the same level without recursing. Otherwise, if
# Line 3284  for (;;) Line 3508  for (;;)
3508      maximizing, find the maximum number of characters and work backwards. */      maximizing, find the maximum number of characters and work backwards. */
3509    
3510      DPRINTF(("matching %c{%d,%d} against subject %.*s\n", fc, min, max,      DPRINTF(("matching %c{%d,%d} against subject %.*s\n", fc, min, max,
3511        max, eptr));        max, (char *)eptr));
3512    
3513      if (op >= OP_STARI)  /* Caseless */      if (op >= OP_STARI)  /* Caseless */
3514        {        {
3515        fc = md->lcc[fc];  #ifdef COMPILE_PCRE8
3516          /* fc must be < 128 if UTF is enabled. */
3517          foc = md->fcc[fc];
3518    #else
3519    #ifdef SUPPORT_UTF
3520    #ifdef SUPPORT_UCP
3521          if (utf && fc > 127)
3522            foc = UCD_OTHERCASE(fc);
3523    #else
3524          if (utf && fc > 127)
3525            foc = fc;
3526    #endif /* SUPPORT_UCP */
3527          else
3528    #endif /* SUPPORT_UTF */
3529            foc = TABLE_GET(fc, md->fcc, fc);
3530    #endif /* COMPILE_PCRE8 */
3531    
3532        for (i = 1; i <= min; i++)        for (i = 1; i <= min; i++)
3533          {          {
3534            pcre_uint32 cc;                 /* Faster than pcre_uchar */
3535          if (eptr >= md->end_subject)          if (eptr >= md->end_subject)
3536            {            {
3537            SCHECK_PARTIAL();            SCHECK_PARTIAL();
3538            RRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
3539            }            }
3540          if (fc != md->lcc[*eptr++]) RRETURN(MATCH_NOMATCH);          cc = RAWUCHARTEST(eptr);
3541            if (fc != cc && foc != cc) RRETURN(MATCH_NOMATCH);
3542            eptr++;
3543          }          }
3544        if (min == max) continue;        if (min == max) continue;
3545        if (minimize)        if (minimize)
3546          {          {
3547          for (fi = min;; fi++)          for (fi = min;; fi++)
3548            {            {
3549              pcre_uint32 cc;               /* Faster than pcre_uchar */
3550            RMATCH(eptr, ecode, offset_top, md, eptrb, RM24);            RMATCH(eptr, ecode, offset_top, md, eptrb, RM24);
3551            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3552            if (fi >= max) RRETURN(MATCH_NOMATCH);            if (fi >= max) RRETURN(MATCH_NOMATCH);
# Line 3311  for (;;) Line 3555  for (;;)
3555              SCHECK_PARTIAL();              SCHECK_PARTIAL();
3556              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
3557              }              }
3558            if (fc != md->lcc[*eptr++]) RRETURN(MATCH_NOMATCH);            cc = RAWUCHARTEST(eptr);
3559              if (fc != cc && foc != cc) RRETURN(MATCH_NOMATCH);
3560              eptr++;
3561            }            }
3562          /* Control never gets here */          /* Control never gets here */
3563          }          }
# Line 3320  for (;;) Line 3566  for (;;)
3566          pp = eptr;          pp = eptr;
3567          for (i = min; i < max; i++)          for (i = min; i < max; i++)
3568            {            {
3569              pcre_uint32 cc;               /* Faster than pcre_uchar */
3570            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
3571              {              {
3572              SCHECK_PARTIAL();              SCHECK_PARTIAL();
3573              break;              break;
3574              }              }
3575            if (fc != md->lcc[*eptr]) break;            cc = RAWUCHARTEST(eptr);
3576              if (fc != cc && foc != cc) break;
3577            eptr++;            eptr++;
3578            }            }
3579    
# Line 3353  for (;;) Line 3601  for (;;)
3601            SCHECK_PARTIAL();            SCHECK_PARTIAL();
3602            RRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
3603            }            }
3604          if (fc != *eptr++) RRETURN(MATCH_NOMATCH);          if (fc != RAWUCHARINCTEST(eptr)) RRETURN(MATCH_NOMATCH);
3605          }          }
3606    
3607        if (min == max) continue;        if (min == max) continue;
# Line 3370  for (;;) Line 3618  for (;;)
3618              SCHECK_PARTIAL();              SCHECK_PARTIAL();
3619              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
3620              }              }
3621            if (fc != *eptr++) RRETURN(MATCH_NOMATCH);            if (fc != RAWUCHARINCTEST(eptr)) RRETURN(MATCH_NOMATCH);
3622            }            }
3623          /* Control never gets here */          /* Control never gets here */
3624          }          }
# Line 3384  for (;;) Line 3632  for (;;)
3632              SCHECK_PARTIAL();              SCHECK_PARTIAL();
3633              break;              break;
3634              }              }
3635            if (fc != *eptr) break;            if (fc != RAWUCHARTEST(eptr)) break;
3636            eptr++;            eptr++;
3637            }            }
3638          if (possessive) continue;          if (possessive) continue;
# Line 3410  for (;;) Line 3658  for (;;)
3658        SCHECK_PARTIAL();        SCHECK_PARTIAL();
3659        RRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
3660        }        }
3661      ecode++;  #ifdef SUPPORT_UTF
3662      GETCHARINCTEST(c, eptr);      if (utf)
     if (op == OP_NOTI)         /* The caseless case */  
3663        {        {
3664  #ifdef SUPPORT_UTF8        register pcre_uint32 ch, och;
3665        if (c < 256)  
3666  #endif        ecode++;
3667        c = md->lcc[c];        GETCHARINC(ch, ecode);
3668        if (md->lcc[*ecode++] == c) RRETURN(MATCH_NOMATCH);        GETCHARINC(c, eptr);
3669    
3670          if (op == OP_NOT)
3671            {
3672            if (ch == c) RRETURN(MATCH_NOMATCH);
3673            }
3674          else
3675            {
3676    #ifdef SUPPORT_UCP
3677            if (ch > 127)
3678              och = UCD_OTHERCASE(ch);
3679    #else
3680            if (ch > 127)
3681              och = ch;
3682    #endif /* SUPPORT_UCP */
3683            else
3684              och = TABLE_GET(ch, md->fcc, ch);
3685            if (ch == c || och == c) RRETURN(MATCH_NOMATCH);
3686            }
3687        }        }
3688      else    /* Caseful */      else
3689    #endif
3690        {        {
3691        if (*ecode++ == c) RRETURN(MATCH_NOMATCH);        register pcre_uint32 ch = ecode[1];
3692          c = *eptr++;
3693          if (ch == c || (op == OP_NOTI && TABLE_GET(ch, md->fcc, ch) == c))
3694            RRETURN(MATCH_NOMATCH);
3695          ecode += 2;
3696        }        }
3697      break;      break;
3698    
# Line 3436  for (;;) Line 3706  for (;;)
3706      case OP_NOTEXACT:      case OP_NOTEXACT:
3707      case OP_NOTEXACTI:      case OP_NOTEXACTI:
3708      min = max = GET2(ecode, 1);      min = max = GET2(ecode, 1);
3709      ecode += 3;      ecode += 1 + IMM2_SIZE;
3710      goto REPEATNOTCHAR;      goto REPEATNOTCHAR;
3711    
3712      case OP_NOTUPTO:      case OP_NOTUPTO:
# Line 3446  for (;;) Line 3716  for (;;)
3716      min = 0;      min = 0;
3717      max = GET2(ecode, 1);      max = GET2(ecode, 1);
3718      minimize = *ecode == OP_NOTMINUPTO || *ecode == OP_NOTMINUPTOI;      minimize = *ecode == OP_NOTMINUPTO || *ecode == OP_NOTMINUPTOI;
3719      ecode += 3;      ecode += 1 + IMM2_SIZE;
3720      goto REPEATNOTCHAR;      goto REPEATNOTCHAR;
3721    
3722      case OP_NOTPOSSTAR:      case OP_NOTPOSSTAR:
# Line 3478  for (;;) Line 3748  for (;;)
3748      possessive = TRUE;      possessive = TRUE;
3749      min = 0;      min = 0;
3750      max = GET2(ecode, 1);      max = GET2(ecode, 1);
3751      ecode += 3;      ecode += 1 + IMM2_SIZE;
3752      goto REPEATNOTCHAR;      goto REPEATNOTCHAR;
3753    
3754      case OP_NOTSTAR:      case OP_NOTSTAR:
# Line 3502  for (;;) Line 3772  for (;;)
3772      /* Common code for all repeated single-byte matches. */      /* Common code for all repeated single-byte matches. */
3773    
3774      REPEATNOTCHAR:      REPEATNOTCHAR:
3775      fc = *ecode++;      GETCHARINCTEST(fc, ecode);
3776    
3777      /* The code is duplicated for the caseless and caseful cases, for speed,      /* The code is duplicated for the caseless and caseful cases, for speed,
3778      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 3783  for (;;)
3783      characters and work backwards. */      characters and work backwards. */
3784    
3785      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,
3786        max, eptr));        max, (char *)eptr));
3787    
3788      if (op >= OP_NOTSTARI)     /* Caseless */      if (op >= OP_NOTSTARI)     /* Caseless */
3789        {        {
3790        fc = md->lcc[fc];  #ifdef SUPPORT_UTF
3791    #ifdef SUPPORT_UCP
3792          if (utf && fc > 127)
3793            foc = UCD_OTHERCASE(fc);
3794    #else
3795          if (utf && fc > 127)
3796            foc = fc;
3797    #endif /* SUPPORT_UCP */
3798          else
3799    #endif /* SUPPORT_UTF */
3800            foc = TABLE_GET(fc, md->fcc, fc);
3801    
3802  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
3803        /* UTF-8 mode */        if (utf)
       if (utf8)  
3804          {          {
3805          register unsigned int d;          register pcre_uint32 d;
3806          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
3807            {            {
3808            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
# Line 3532  for (;;) Line 3811  for (;;)
3811              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
3812              }              }
3813            GETCHARINC(d, eptr);            GETCHARINC(d, eptr);
3814            if (d < 256) d = md->lcc[d];            if (fc == d || (unsigned int)foc == d) RRETURN(MATCH_NOMATCH);
           if (fc == d) RRETURN(MATCH_NOMATCH);  
3815            }            }
3816          }          }
3817        else        else
3818  #endif  #endif
3819          /* Not UTF mode */
       /* Not UTF-8 mode */  
3820          {          {
3821          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
3822            {            {
# Line 3548  for (;;) Line 3825  for (;;)
3825              SCHECK_PARTIAL();              SCHECK_PARTIAL();
3826              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
3827              }              }
3828            if (fc == md->lcc[*eptr++]) RRETURN(MATCH_NOMATCH);            if (fc == *eptr || foc == *eptr) RRETURN(MATCH_NOMATCH);
3829              eptr++;
3830            }            }
3831          }          }
3832    
# Line 3556  for (;;) Line 3834  for (;;)
3834    
3835        if (minimize)        if (minimize)
3836          {          {
3837  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
3838          /* UTF-8 mode */          if (utf)
         if (utf8)  
3839            {            {
3840            register unsigned int d;            register pcre_uint32 d;
3841            for (fi = min;; fi++)            for (fi = min;; fi++)
3842              {              {
3843              RMATCH(eptr, ecode, offset_top, md, eptrb, RM28);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM28);
# Line 3572  for (;;) Line 3849  for (;;)
3849                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
3850                }                }
3851              GETCHARINC(d, eptr);              GETCHARINC(d, eptr);
3852              if (d < 256) d = md->lcc[d];              if (fc == d || (unsigned int)foc == d) RRETURN(MATCH_NOMATCH);
             if (fc == d) RRETURN(MATCH_NOMATCH);  
3853              }              }
3854            }            }
3855          else          else
3856  #endif  #endif
3857          /* Not UTF-8 mode */          /* Not UTF mode */
3858            {            {
3859            for (fi = min;; fi++)            for (fi = min;; fi++)
3860              {              {
# Line 3590  for (;;) Line 3866  for (;;)
3866                SCHECK_PARTIAL();                SCHECK_PARTIAL();
3867                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
3868                }                }
3869              if (fc == md->lcc[*eptr++]) RRETURN(MATCH_NOMATCH);              if (fc == *eptr || foc == *eptr) RRETURN(MATCH_NOMATCH);
3870                eptr++;
3871              }              }
3872            }            }
3873          /* Control never gets here */          /* Control never gets here */
# Line 3602  for (;;) Line 3879  for (;;)
3879          {          {
3880          pp = eptr;          pp = eptr;
3881    
3882  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
3883          /* UTF-8 mode */          if (utf)
         if (utf8)  
3884            {            {
3885            register unsigned int d;            register pcre_uint32 d;
3886            for (i = min; i < max; i++)            for (i = min; i < max; i++)
3887              {              {
3888              int len = 1;              int len = 1;
# Line 3616  for (;;) Line 3892  for (;;)
3892                break;                break;
3893                }                }
3894              GETCHARLEN(d, eptr, len);              GETCHARLEN(d, eptr, len);
3895              if (d < 256) d = md->lcc[d];              if (fc == d || (unsigned int)foc == d) break;
             if (fc == d) break;  
3896              eptr += len;              eptr += len;
3897              }              }
3898          if (possessive) continue;            if (possessive) continue;
3899          for(;;)            for(;;)
3900              {              {
3901              RMATCH(eptr, ecode, offset_top, md, eptrb, RM30);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM30);
3902              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
# Line 3631  for (;;) Line 3906  for (;;)
3906            }            }
3907          else          else
3908  #endif  #endif
3909          /* Not UTF-8 mode */          /* Not UTF mode */
3910            {            {
3911            for (i = min; i < max; i++)            for (i = min; i < max; i++)
3912              {              {
# Line 3640  for (;;) Line 3915  for (;;)
3915                SCHECK_PARTIAL();                SCHECK_PARTIAL();
3916                break;                break;
3917                }                }
3918              if (fc == md->lcc[*eptr]) break;              if (fc == *eptr || foc == *eptr) break;
3919              eptr++;              eptr++;
3920              }              }
3921            if (possessive) continue;            if (possessive) continue;
# Line 3661  for (;;) Line 3936  for (;;)
3936    
3937      else      else
3938        {        {
3939  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
3940        /* UTF-8 mode */        if (utf)
       if (utf8)  
3941          {          {
3942          register unsigned int d;          register pcre_uint32 d;
3943          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
3944            {            {
3945            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
# Line 3679  for (;;) Line 3953  for (;;)
3953          }          }
3954        else        else
3955  #endif  #endif
3956        /* Not UTF-8 mode */        /* Not UTF mode */
3957          {          {
3958          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
3959            {            {
# Line 3696  for (;;) Line 3970  for (;;)
3970    
3971        if (minimize)        if (minimize)
3972          {          {
3973  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
3974          /* UTF-8 mode */          if (utf)
         if (utf8)  
3975            {            {
3976            register unsigned int d;            register pcre_uint32 d;
3977            for (fi = min;; fi++)            for (fi = min;; fi++)
3978              {              {
3979              RMATCH(eptr, ecode, offset_top, md, eptrb, RM32);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM32);
# Line 3717  for (;;) Line 3990  for (;;)
3990            }            }
3991          else          else
3992  #endif  #endif
3993          /* Not UTF-8 mode */          /* Not UTF mode */
3994            {            {
3995            for (fi = min;; fi++)            for (fi = min;; fi++)
3996              {              {
# Line 3741  for (;;) Line 4014  for (;;)
4014          {          {
4015          pp = eptr;          pp = eptr;
4016    
4017  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
4018          /* UTF-8 mode */          if (utf)
         if (utf8)  
4019            {            {
4020            register unsigned int d;            register pcre_uint32 d;
4021            for (i = min; i < max; i++)            for (i = min; i < max; i++)
4022              {              {
4023              int len = 1;              int len = 1;
# Line 3769  for (;;) Line 4041  for (;;)
4041            }            }
4042          else          else
4043  #endif  #endif
4044          /* Not UTF-8 mode */          /* Not UTF mode */
4045            {            {
4046            for (i = min; i < max; i++)            for (i = min; i < max; i++)
4047              {              {
# Line 3802  for (;;) Line 4074  for (;;)
4074      case OP_TYPEEXACT:      case OP_TYPEEXACT:
4075      min = max = GET2(ecode, 1);      min = max = GET2(ecode, 1);
4076      minimize = TRUE;      minimize = TRUE;
4077      ecode += 3;      ecode += 1 + IMM2_SIZE;
4078      goto REPEATTYPE;      goto REPEATTYPE;
4079    
4080      case OP_TYPEUPTO:      case OP_TYPEUPTO:
# Line 3810  for (;;) Line 4082  for (;;)
4082      min = 0;      min = 0;
4083      max = GET2(ecode, 1);      max = GET2(ecode, 1);
4084      minimize = *ecode == OP_TYPEMINUPTO;      minimize = *ecode == OP_TYPEMINUPTO;
4085      ecode += 3;      ecode += 1 + IMM2_SIZE;
4086      goto REPEATTYPE;      goto REPEATTYPE;
4087    
4088      case OP_TYPEPOSSTAR:      case OP_TYPEPOSSTAR:
# Line 3838  for (;;) Line 4110  for (;;)
4110      possessive = TRUE;      possessive = TRUE;
4111      min = 0;      min = 0;
4112      max = GET2(ecode, 1);      max = GET2(ecode, 1);
4113      ecode += 3;      ecode += 1 + IMM2_SIZE;
4114      goto REPEATTYPE;      goto REPEATTYPE;
4115    
4116      case OP_TYPESTAR:      case OP_TYPESTAR:
# Line 4021  for (;;) Line 4293  for (;;)
4293              }              }
4294            break;            break;
4295    
4296              case PT_CLIST:
4297              for (i = 1; i <= min; i++)
4298                {
4299                const pcre_uint32 *cp;
4300                if (eptr >= md->end_subject)
4301                  {
4302                  SCHECK_PARTIAL();
4303                  RRETURN(MATCH_NOMATCH);
4304                  }
4305                GETCHARINCTEST(c, eptr);
4306                cp = PRIV(ucd_caseless_sets) + prop_value;
4307                for (;;)
4308                  {
4309                  if (c < *cp)
4310                    { if (prop_fail_result) break; else { RRETURN(MATCH_NOMATCH); } }
4311                  if (c == *cp++)
4312                    { if (prop_fail_result) { RRETURN(MATCH_NOMATCH); } else break; }
4313                  }
4314                }
4315              break;
4316    
4317              case PT_UCNC:
4318              for (i = 1; i <= min; i++)
4319                {
4320                if (eptr >= md->end_subject)
4321                  {
4322                  SCHECK_PARTIAL();
4323                  RRETURN(MATCH_NOMATCH);
4324                  }
4325                GETCHARINCTEST(c, eptr);
4326                if ((c == CHAR_DOLLAR_SIGN || c == CHAR_COMMERCIAL_AT ||
4327                     c == CHAR_GRAVE_ACCENT || (c >= 0xa0 && c <= 0xd7ff) ||
4328                     c >= 0xe000) == prop_fail_result)
4329                  RRETURN(MATCH_NOMATCH);
4330                }
4331              break;
4332    
4333            /* This should not occur */            /* This should not occur */
4334    
4335            default:            default:
# Line 4040  for (;;) Line 4349  for (;;)
4349              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4350              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4351              }              }
4352            GETCHARINCTEST(c, eptr);            else
           if (UCD_CATEGORY(c) == ucp_M) RRETURN(MATCH_NOMATCH);  
           while (eptr < md->end_subject)  
4353              {              {
4354              int len = 1;              int lgb, rgb;
4355              if (!utf8) c = *eptr; else { GETCHARLEN(c, eptr, len); }              GETCHARINCTEST(c, eptr);
4356              if (UCD_CATEGORY(c) != ucp_M) break;              lgb = UCD_GRAPHBREAK(c);
4357              eptr += len;             while (eptr < md->end_subject)
4358                  {
4359                  int len = 1;
4360                  if (!utf) c = *eptr; else { GETCHARLEN(c, eptr, len); }
4361                  rgb = UCD_GRAPHBREAK(c);
4362                  if ((PRIV(ucp_gbtable)[lgb] & (1 << rgb)) == 0) break;
4363                  lgb = rgb;
4364                  eptr += len;
4365                  }
4366              }              }
4367              CHECK_PARTIAL();
4368            }            }
4369          }          }
4370    
# Line 4057  for (;;) Line 4373  for (;;)
4373    
4374  /* Handle all other cases when the coding is UTF-8 */  /* Handle all other cases when the coding is UTF-8 */
4375    
4376  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
4377        if (utf8) switch(ctype)        if (utf) switch(ctype)
4378          {          {
4379          case OP_ANY:          case OP_ANY:
4380          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
# Line 4069  for (;;) Line 4385  for (;;)
4385              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4386              }              }
4387            if (IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH);            if (IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH);
4388              if (md->partial != 0 &&
4389                  eptr + 1 >= md->end_subject &&
4390                  NLBLOCK->nltype == NLTYPE_FIXED &&
4391                  NLBLOCK->nllen == 2 &&
4392                  RAWUCHAR(eptr) == NLBLOCK->nl[0])
4393                {
4394                md->hitend = TRUE;
4395                if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
4396                }
4397            eptr++;            eptr++;
4398            while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;            ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);
4399            }            }
4400          break;          break;
4401    
# Line 4083  for (;;) Line 4408  for (;;)
4408              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4409              }              }
4410            eptr++;            eptr++;
4411            while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;            ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);
4412            }            }
4413          break;          break;
4414    
# Line 4105  for (;;) Line 4430  for (;;)
4430              {              {
4431              default: RRETURN(MATCH_NOMATCH);              default: RRETURN(MATCH_NOMATCH);
4432    
4433              case 0x000d:              case CHAR_CR:
4434              if (eptr < md->end_subject && *eptr == 0x0a) eptr++;              if (eptr < md->end_subject && RAWUCHAR(eptr) == CHAR_LF) eptr++;
4435              break;              break;
4436    
4437              case 0x000a:              case CHAR_LF:
4438              break;              break;
4439    
4440              case 0x000b:              case CHAR_VT:
4441              case 0x000c:              case CHAR_FF:
4442              case 0x0085:              case CHAR_NEL:
4443    #ifndef EBCDIC
4444              case 0x2028:              case 0x2028:
4445              case 0x2029:              case 0x2029:
4446    #endif  /* Not EBCDIC */
4447              if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);              if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);
4448              break;              break;
4449              }              }
# Line 4134  for (;;) Line 4461  for (;;)
4461            GETCHARINC(c, eptr);            GETCHARINC(c, eptr);
4462            switch(c)            switch(c)
4463              {              {
4464                HSPACE_CASES: RRETURN(MATCH_NOMATCH);  /* Byte and multibyte cases */
4465              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);  
4466              }              }
4467            }            }
4468          break;          break;
# Line 4170  for (;;) Line 4478  for (;;)
4478            GETCHARINC(c, eptr);            GETCHARINC(c, eptr);
4479            switch(c)            switch(c)
4480              {              {
4481                HSPACE_CASES: break;  /* Byte and multibyte cases */
4482              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;  
4483              }              }
4484            }            }
4485          break;          break;
# Line 4206  for (;;) Line 4495  for (;;)
4495            GETCHARINC(c, eptr);            GETCHARINC(c, eptr);
4496            switch(c)            switch(c)
4497              {              {
4498                VSPACE_CASES: RRETURN(MATCH_NOMATCH);
4499              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);  
4500              }              }
4501            }            }
4502          break;          break;
# Line 4230  for (;;) Line 4512  for (;;)
4512            GETCHARINC(c, eptr);            GETCHARINC(c, eptr);
4513            switch(c)            switch(c)
4514              {              {
4515                VSPACE_CASES: break;
4516              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;  
4517              }              }
4518            }            }
4519          break;          break;
# Line 4260  for (;;) Line 4535  for (;;)
4535          case OP_DIGIT:          case OP_DIGIT:
4536          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
4537            {            {
4538              pcre_uint32 cc;
4539            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
4540              {              {
4541              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4542              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4543              }              }
4544            if (*eptr >= 128 || (md->ctypes[*eptr++] & ctype_digit) == 0)            cc = RAWUCHAR(eptr);
4545              if (cc >= 128 || (md->ctypes[cc] & ctype_digit) == 0)
4546              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4547              eptr++;
4548            /* 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 */
4549            }            }
4550          break;          break;
# Line 4274  for (;;) Line 4552  for (;;)
4552          case OP_NOT_WHITESPACE:          case OP_NOT_WHITESPACE:
4553          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
4554            {            {
4555              pcre_uint32 cc;
4556            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
4557              {              {
4558              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4559              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4560              }              }
4561            if (*eptr < 128 && (md->ctypes[*eptr] & ctype_space) != 0)            cc = RAWUCHAR(eptr);
4562              if (cc < 128 && (md->ctypes[cc] & ctype_space) != 0)
4563              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4564            while (++eptr < md->end_subject && (*eptr & 0xc0) == 0x80);            eptr++;
4565              ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);
4566            }            }
4567          break;          break;
4568    
4569          case OP_WHITESPACE:          case OP_WHITESPACE:
4570          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
4571            {            {
4572              pcre_uint32 cc;
4573            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
4574              {              {
4575              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4576              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4577              }              }
4578            if (*eptr >= 128 || (md->ctypes[*eptr++] & ctype_space) == 0)            cc = RAWUCHAR(eptr);
4579              if (cc >= 128 || (md->ctypes[cc] & ctype_space) == 0)
4580              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4581              eptr++;
4582            /* 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 */
4583            }            }
4584          break;          break;
# Line 4302  for (;;) Line 4586  for (;;)
4586          case OP_NOT_WORDCHAR:          case OP_NOT_WORDCHAR:
4587          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
4588            {            {
4589              pcre_uint32 cc;
4590            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
4591              {              {
4592              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4593              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4594              }              }
4595            if (*eptr < 128 && (md->ctypes[*eptr] & ctype_word) != 0)            cc = RAWUCHAR(eptr);
4596              if (cc < 128 && (md->ctypes[cc] & ctype_word) != 0)
4597              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4598            while (++eptr < md->end_subject && (*eptr & 0xc0) == 0x80);            eptr++;
4599              ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);
4600            }            }
4601          break;          break;
4602    
4603          case OP_WORDCHAR:          case OP_WORDCHAR:
4604          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
4605            {            {
4606              pcre_uint32 cc;
4607            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
4608              {              {
4609              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4610              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4611              }              }
4612            if (*eptr >= 128 || (md->ctypes[*eptr++] & ctype_word) == 0)            cc = RAWUCHAR(eptr);
4613              if (cc >= 128 || (md->ctypes[cc] & ctype_word) == 0)
4614              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4615              eptr++;
4616            /* 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 */
4617            }            }
4618          break;          break;
# Line 4332  for (;;) Line 4622  for (;;)
4622          }  /* End switch(ctype) */          }  /* End switch(ctype) */
4623    
4624        else        else
4625  #endif     /* SUPPORT_UTF8 */  #endif     /* SUPPORT_UTF */
4626    
4627        /* 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
4628        than OP_PROP and OP_NOTPROP. */        than OP_PROP and OP_NOTPROP. */
# Line 4348  for (;;) Line 4638  for (;;)
4638              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4639              }              }
4640            if (IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH);            if (IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH);
4641              if (md->partial != 0 &&
4642                  eptr + 1 >= md->end_subject &&
4643                  NLBLOCK->nltype == NLTYPE_FIXED &&
4644                  NLBLOCK->nllen == 2 &&
4645                  *eptr == NLBLOCK->nl[0])
4646                {
4647                md->hitend = TRUE;
4648                if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
4649                }
4650            eptr++;            eptr++;
4651            }            }
4652          break;          break;
# Line 4382  for (;;) Line 4681  for (;;)
4681              {              {
4682              default: RRETURN(MATCH_NOMATCH);              default: RRETURN(MATCH_NOMATCH);
4683    
4684              case 0x000d:              case CHAR_CR:
4685              if (eptr < md->end_subject && *eptr == 0x0a) eptr++;              if (eptr < md->end_subject && *eptr == CHAR_LF) eptr++;
4686              break;              break;
4687    
4688              case 0x000a:              case CHAR_LF:
4689              break;              break;
4690    
4691              case 0x000b:              case CHAR_VT:
4692              case 0x000c:              case CHAR_FF:
4693              case 0x0085:              case CHAR_NEL:
4694    #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
4695                case 0x2028:
4696                case 0x2029:
4697    #endif
4698              if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);              if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);
4699              break;              break;
4700              }              }
# Line 4409  for (;;) Line 4712  for (;;)
4712            switch(*eptr++)            switch(*eptr++)
4713              {              {
4714              default: break;              default: break;
4715              case 0x09:      /* HT */              HSPACE_BYTE_CASES:
4716              case 0x20:      /* SPACE */  #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
4717              case 0xa0:      /* NBSP */              HSPACE_MULTIBYTE_CASES:
4718    #endif
4719              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4720              }              }
4721            }            }
# Line 4428  for (;;) Line 4732  for (;;)
4732            switch(*eptr++)            switch(*eptr++)
4733              {              {
4734              default: RRETURN(MATCH_NOMATCH);              default: RRETURN(MATCH_NOMATCH);
4735              case 0x09:      /* HT */              HSPACE_BYTE_CASES:
4736              case 0x20:      /* SPACE */  #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
4737              case 0xa0:      /* NBSP */              HSPACE_MULTIBYTE_CASES:
4738    #endif
4739              break;              break;
4740              }              }
4741            }            }
# Line 4446  for (;;) Line 4751  for (;;)
4751              }              }
4752            switch(*eptr++)            switch(*eptr++)
4753              {              {
4754              default: break;              VSPACE_BYTE_CASES:
4755              case 0x0a:      /* LF */  #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
4756              case 0x0b:      /* VT */              VSPACE_MULTIBYTE_CASES:
4757              case 0x0c:      /* FF */  #endif
             case 0x0d:      /* CR */  
             case 0x85:      /* NEL */  
4758              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4759                default: break;
4760              }              }
4761            }            }
4762          break;          break;
# Line 4468  for (;;) Line 4772  for (;;)
4772            switch(*eptr++)            switch(*eptr++)
4773              {              {
4774              default: RRETURN(MATCH_NOMATCH);              default: RRETURN(MATCH_NOMATCH);
4775              case 0x0a:      /* LF */              VSPACE_BYTE_CASES:
4776              case 0x0b:      /* VT */  #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
4777              case 0x0c:      /* FF */              VSPACE_MULTIBYTE_CASES:
4778              case 0x0d:      /* CR */  #endif
             case 0x85:      /* NEL */  
4779              break;              break;
4780              }              }
4781            }            }
# Line 4486  for (;;) Line 4789  for (;;)
4789              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4790              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4791              }              }
4792            if ((md->ctypes[*eptr++] & ctype_digit) != 0) RRETURN(MATCH_NOMATCH);            if (MAX_255(*eptr) && (md->ctypes[*eptr] & ctype_digit) != 0)
4793                RRETURN(MATCH_NOMATCH);
4794              eptr++;
4795            }            }
4796          break;          break;
4797    
# Line 4498  for (;;) Line 4803  for (;;)
4803              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4804              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4805              }              }
4806            if ((md->ctypes[*eptr++] & ctype_digit) == 0) RRETURN(MATCH_NOMATCH);            if (!MAX_255(*eptr) || (md->ctypes[*eptr] & ctype_digit) == 0)
4807                RRETURN(MATCH_NOMATCH);
4808              eptr++;
4809            }            }
4810          break;          break;
4811    
# Line 4510  for (;;) Line 4817  for (;;)
4817              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4818              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4819              }              }
4820            if ((md->ctypes[*eptr++] & ctype_space) != 0) RRETURN(MATCH_NOMATCH);            if (MAX_255(*eptr) && (md->ctypes[*eptr] & ctype_space) != 0)
4821                RRETURN(MATCH_NOMATCH);
4822              eptr++;
4823            }            }
4824          break;          break;
4825    
# Line 4522  for (;;) Line 4831  for (;;)
4831              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4832              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4833              }              }
4834            if ((md->ctypes[*eptr++] & ctype_space) == 0) RRETURN(MATCH_NOMATCH);            if (!MAX_255(*eptr) || (md->ctypes[*eptr] & ctype_space) == 0)
4835                RRETURN(MATCH_NOMATCH);
4836              eptr++;
4837            }            }
4838          break;          break;
4839    
# Line 4534  for (;;) Line 4845  for (;;)
4845              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4846              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4847              }              }
4848            if ((md->ctypes[*eptr++] & ctype_word) != 0)            if (MAX_255(*eptr) && (md->ctypes[*eptr] & ctype_word) != 0)
4849              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4850              eptr++;
4851            }            }
4852          break;          break;
4853    
# Line 4547  for (;;) Line 4859  for (;;)
4859              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4860              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4861              }              }
4862            if ((md->ctypes[*eptr++] & ctype_word) == 0)            if (!MAX_255(*eptr) || (md->ctypes[*eptr] & ctype_word) == 0)
4863              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4864              eptr++;
4865            }            }
4866          break;          break;
4867    
# Line 4739  for (;;) Line 5052  for (;;)
5052              }              }
5053            /* Control never gets here */            /* Control never gets here */
5054    
5055            /* This should never occur */            case PT_CLIST:
5056              for (fi = min;; fi++)
5057                {
5058                const pcre_uint32 *cp;
5059                RMATCH(eptr, ecode, offset_top, md, eptrb, RM67);
5060                if (rrc != MATCH_NOMATCH) RRETURN(rrc);
5061                if (fi >= max) RRETURN(MATCH_NOMATCH);
5062                if (eptr >= md->end_subject)
5063                  {
5064                  SCHECK_PARTIAL();
5065                  RRETURN(MATCH_NOMATCH);
5066                  }
5067                GETCHARINCTEST(c, eptr);
5068                cp = PRIV(ucd_caseless_sets) + prop_value;
5069                for (;;)
5070                  {
5071                  if (c < *cp)
5072                    { if (prop_fail_result) break; else { RRETURN(MATCH_NOMATCH); } }
5073                  if (c == *cp++)
5074                    { if (prop_fail_result) { RRETURN(MATCH_NOMATCH); } else break; }
5075                  }
5076                }
5077              /* Control never gets here */
5078    
5079              case PT_UCNC:
5080              for (fi = min;; fi++)
5081                {
5082                RMATCH(eptr, ecode, offset_top, md, eptrb, RM68);
5083                if (rrc != MATCH_NOMATCH) RRETURN(rrc);
5084                if (fi >= max) RRETURN(MATCH_NOMATCH);
5085                if (eptr >= md->end_subject)
5086                  {
5087                  SCHECK_PARTIAL();
5088                  RRETURN(MATCH_NOMATCH);
5089                  }
5090                GETCHARINCTEST(c, eptr);
5091                if ((c == CHAR_DOLLAR_SIGN || c == CHAR_COMMERCIAL_AT ||
5092                     c == CHAR_GRAVE_ACCENT || (c >= 0xa0 && c <= 0xd7ff) ||
5093                     c >= 0xe000) == prop_fail_result)
5094                  RRETURN(MATCH_NOMATCH);
5095                }
5096              /* Control never gets here */
5097    
5098              /* This should never occur */
5099            default:            default:
5100            RRETURN(PCRE_ERROR_INTERNAL);            RRETURN(PCRE_ERROR_INTERNAL);
5101            }            }
# Line 4761  for (;;) Line 5116  for (;;)
5116              SCHECK_PARTIAL();              SCHECK_PARTIAL();
5117              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
5118              }              }
5119            GETCHARINCTEST(c, eptr);            else
           if (UCD_CATEGORY(c) == ucp_M) RRETURN(MATCH_NOMATCH);  
           while (eptr < md->end_subject)  
5120              {              {
5121              int len = 1;              int lgb, rgb;
5122              if (!utf8) c = *eptr; else { GETCHARLEN(c, eptr, len); }              GETCHARINCTEST(c, eptr);
5123              if (UCD_CATEGORY(c) != ucp_M) break;              lgb = UCD_GRAPHBREAK(c);
5124              eptr += len;              while (eptr < md->end_subject)
5125                  {
5126                  int len = 1;
5127                  if (!utf) c = *eptr; else { GETCHARLEN(c, eptr, len); }
5128                  rgb = UCD_GRAPHBREAK(c);
5129                  if ((PRIV(ucp_gbtable)[lgb] & (1 << rgb)) == 0) break;
5130                  lgb = rgb;
5131                  eptr += len;
5132                  }
5133              }              }
5134              CHECK_PARTIAL();
5135            }            }
5136          }          }
5137        else        else
5138  #endif     /* SUPPORT_UCP */  #endif     /* SUPPORT_UCP */
5139    
5140  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
5141        /* UTF-8 mode */        if (utf)
       if (utf8)  
5142          {          {
5143          for (fi = min;; fi++)          for (fi = min;; fi++)
5144            {            {
# Line 4794  for (;;) Line 5155  for (;;)
5155            GETCHARINC(c, eptr);            GETCHARINC(c, eptr);
5156            switch(ctype)            switch(ctype)
5157              {              {
5158              case OP_ANY:        /* This is the non-NL case */              case OP_ANY:               /* This is the non-NL case */
5159                if (md->partial != 0 &&    /* Take care with CRLF partial */
5160                    eptr >= md->end_subject &&
5161                    NLBLOCK->nltype == NLTYPE_FIXED &&
5162                    NLBLOCK->nllen == 2 &&
5163                    c == NLBLOCK->nl[0])
5164                  {
5165                  md->hitend = TRUE;
5166                  if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
5167                  }
5168                break;
5169    
5170              case OP_ALLANY:              case OP_ALLANY:
5171              case OP_ANYBYTE:              case OP_ANYBYTE:
5172              break;              break;
# Line 4803  for (;;) Line 5175  for (;;)
5175              switch(c)              switch(c)
5176                {                {
5177                default: RRETURN(MATCH_NOMATCH);                default: RRETURN(MATCH_NOMATCH);
5178                case 0x000d:                case CHAR_CR:
5179                if (eptr < md->end_subject && *eptr == 0x0a) eptr++;                if (eptr < md->end_subject && RAWUCHAR(eptr) == CHAR_LF) eptr++;
5180                break;                break;
5181                case 0x000a:  
5182                  case CHAR_LF:
5183                break;                break;
5184    
5185                case 0x000b:                case CHAR_VT:
5186                case 0x000c:                case CHAR_FF:
5187                case 0x0085:                case CHAR_NEL:
5188    #ifndef EBCDIC
5189                case 0x2028:                case 0x2028:
5190                case 0x2029:                case 0x2029:
5191    #endif  /* Not EBCDIC */
5192                if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);                if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);
5193                break;                break;
5194                }                }
# Line 4822  for (;;) Line 5197  for (;;)
5197              case OP_NOT_HSPACE:              case OP_NOT_HSPACE:
5198              switch(c)              switch(c)
5199                {                {
5200                  HSPACE_CASES: RRETURN(MATCH_NOMATCH);
5201                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);  
5202                }                }
5203              break;              break;
5204    
5205              case OP_HSPACE:              case OP_HSPACE:
5206              switch(c)              switch(c)
5207                {                {
5208                  HSPACE_CASES: break;
5209                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;  
5210                }                }
5211              break;              break;
5212    
5213              case OP_NOT_VSPACE:              case OP_NOT_VSPACE:
5214              switch(c)              switch(c)
5215                {                {
5216                  VSPACE_CASES: RRETURN(MATCH_NOMATCH);
5217                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);  
5218                }                }
5219              break;              break;
5220    
5221              case OP_VSPACE:              case OP_VSPACE:
5222              switch(c)              switch(c)
5223                {                {
5224                  VSPACE_CASES: break;
5225                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;  
5226                }                }
5227              break;              break;
5228    
# Line 4919  for (;;) Line 5242  for (;;)
5242              break;              break;
5243    
5244              case OP_WHITESPACE:              case OP_WHITESPACE:
5245              if  (c >= 256 || (md->ctypes[c] & ctype_space) == 0)              if (c >= 256 || (md->ctypes[c] & ctype_space) == 0)
5246                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
5247              break;              break;
5248    
# Line 4940  for (;;) Line 5263  for (;;)
5263          }          }
5264        else        else
5265  #endif  #endif
5266        /* Not UTF-8 mode */        /* Not UTF mode */
5267          {          {
5268          for (fi = min;; fi++)          for (fi = min;; fi++)
5269            {            {
# Line 4957  for (;;) Line 5280  for (;;)
5280            c = *eptr++;            c = *eptr++;
5281            switch(ctype)            switch(ctype)
5282              {              {
5283              case OP_ANY:     /* This is the non-NL case */              case OP_ANY:               /* This is the non-NL case */
5284                if (md->partial != 0 &&    /* Take care with CRLF partial */
5285                    eptr >= md->end_subject &&
5286                    NLBLOCK->nltype == NLTYPE_FIXED &&
5287                    NLBLOCK->nllen == 2 &&
5288                    c == NLBLOCK->nl[0])
5289                  {
5290                  md->hitend = TRUE;
5291                  if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
5292                  }
5293                break;
5294    
5295              case OP_ALLANY:              case OP_ALLANY:
5296              case OP_ANYBYTE:              case OP_ANYBYTE:
5297              break;              break;
# Line 4966  for (;;) Line 5300  for (;;)
5300              switch(c)              switch(c)
5301                {                {
5302                default: RRETURN(MATCH_NOMATCH);                default: RRETURN(MATCH_NOMATCH);
5303                case 0x000d:                case CHAR_CR:
5304                if (eptr < md->end_subject && *eptr == 0x0a) eptr++;                if (eptr < md->end_subject && *eptr == CHAR_LF) eptr++;
5305                break;                break;
5306    
5307                case 0x000a:                case CHAR_LF:
5308                break;                break;
5309    
5310                case 0x000b:                case CHAR_VT:
5311                case 0x000c:                case CHAR_FF:
5312                case 0x0085:                case CHAR_NEL:
5313    #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
5314                  case 0x2028:
5315                  case 0x2029:
5316    #endif
5317                if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);                if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);
5318                break;                break;
5319                }                }
# Line 4985  for (;;) Line 5323  for (;;)
5323              switch(c)              switch(c)
5324                {                {
5325                default: break;                default: break;
5326                case 0x09:      /* HT */                HSPACE_BYTE_CASES:
5327                case 0x20:      /* SPACE */  #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
5328                case 0xa0:      /* NBSP */                HSPACE_MULTIBYTE_CASES:
5329    #endif
5330                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
5331                }                }
5332              break;              break;
# Line 4996  for (;;) Line 5335  for (;;)
5335              switch(c)              switch(c)
5336                {                {
5337                default: RRETURN(MATCH_NOMATCH);                default: RRETURN(MATCH_NOMATCH);
5338                case 0x09:      /* HT */                HSPACE_BYTE_CASES:
5339                case 0x20:      /* SPACE */  #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
5340                case 0xa0:      /* NBSP */                HSPACE_MULTIBYTE_CASES:
5341    #endif
5342                break;                break;
5343                }                }
5344              break;              break;
# Line 5007  for (;;) Line 5347  for (;;)
5347              switch(c)              switch(c)
5348                {                {
5349                default: break;                default: break;
5350                case 0x0a:      /* LF */                VSPACE_BYTE_CASES:
5351                case 0x0b:      /* VT */  #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
5352                case 0x0c:      /* FF */                VSPACE_MULTIBYTE_CASES:
5353                case 0x0d:      /* CR */  #endif
               case 0x85:      /* NEL */  
5354                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
5355                }                }
5356              break;              break;
# Line 5020  for (;;) Line 5359  for (;;)
5359              switch(c)              switch(c)
5360                {                {
5361                default: RRETURN(MATCH_NOMATCH);                default: RRETURN(MATCH_NOMATCH);
5362                case 0x0a:      /* LF */                VSPACE_BYTE_CASES:
5363                case 0x0b:      /* VT */  #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
5364                case 0x0c:      /* FF */                VSPACE_MULTIBYTE_CASES:
5365                case 0x0d:      /* CR */  #endif
               case 0x85:      /* NEL */  
5366                break;                break;
5367                }                }
5368              break;              break;
5369    
5370              case OP_NOT_DIGIT:              case OP_NOT_DIGIT:
5371              if ((md->ctypes[c] & ctype_digit) != 0) RRETURN(MATCH_NOMATCH);              if (MAX_255(c) && (md->ctypes[c] & ctype_digit) != 0) RRETURN(MATCH_NOMATCH);
5372              break;              break;
5373    
5374              case OP_DIGIT:              case OP_DIGIT:
5375              if ((md->ctypes[c] & ctype_digit) == 0) RRETURN(MATCH_NOMATCH);              if (!MAX_255(c) || (md->ctypes[c] & ctype_digit) == 0) RRETURN(MATCH_NOMATCH);
5376              break;              break;
5377    
5378              case OP_NOT_WHITESPACE:              case OP_NOT_WHITESPACE:
5379              if ((md->ctypes[c] & ctype_space) != 0) RRETURN(MATCH_NOMATCH);              if (MAX_255(c) && (md->ctypes[c] & ctype_space) != 0) RRETURN(MATCH_NOMATCH);
5380              break;              break;
5381    
5382              case OP_WHITESPACE:              case OP_WHITESPACE:
5383              if  ((md->ctypes[c] & ctype_space) == 0) RRETURN(MATCH_NOMATCH);              if (!MAX_255(c) || (md->ctypes[c] & ctype_space) == 0) RRETURN(MATCH_NOMATCH);
5384              break;              break;
5385    
5386              case OP_NOT_WORDCHAR:              case OP_NOT_WORDCHAR:
5387              if ((md->ctypes[c] & ctype_word) != 0) RRETURN(MATCH_NOMATCH);              if (MAX_255(c) && (md->ctypes[c] & ctype_word) != 0) RRETURN(MATCH_NOMATCH);
5388              break;              break;
5389    
5390              case OP_WORDCHAR:              case OP_WORDCHAR:
5391              if ((md->ctypes[c] & ctype_word) == 0) RRETURN(MATCH_NOMATCH);              if (!MAX_255(c) || (md->ctypes[c] & ctype_word) == 0) RRETURN(MATCH_NOMATCH);
5392              break;              break;
5393    
5394              default:              default:
# Line 5227  for (;;) Line 5565  for (;;)
5565              }              }
5566            break;            break;
5567    
5568              case PT_CLIST:
5569              for (i = min; i < max; i++)
5570                {
5571                const pcre_uint32 *cp;
5572                int len = 1;
5573                if (eptr >= md->end_subject)
5574                  {
5575                  SCHECK_PARTIAL();
5576                  break;
5577                  }
5578                GETCHARLENTEST(c, eptr, len);
5579                cp = PRIV(ucd_caseless_sets) + prop_value;
5580                for (;;)
5581                  {
5582                  if (c < *cp)
5583                    { if (prop_fail_result) break; else goto GOT_MAX; }
5584                  if (c == *cp++)
5585                    { if (prop_fail_result) goto GOT_MAX; else break; }
5586                  }
5587                eptr += len;
5588                }
5589              GOT_MAX:
5590              break;
5591    
5592              case PT_UCNC:
5593              for (i = min; i < max; i++)
5594                {
5595                int len = 1;
5596                if (eptr >= md->end_subject)
5597                  {
5598                  SCHECK_PARTIAL();
5599                  break;
5600                  }
5601                GETCHARLENTEST(c, eptr, len);
5602                if ((c == CHAR_DOLLAR_SIGN || c == CHAR_COMMERCIAL_AT ||
5603                     c == CHAR_GRAVE_ACCENT || (c >= 0xa0 && c <= 0xd7ff) ||
5604                     c >= 0xe000) == prop_fail_result)
5605                  break;
5606                eptr += len;
5607                }
5608              break;
5609    
5610            default:            default:
5611            RRETURN(PCRE_ERROR_INTERNAL);            RRETURN(PCRE_ERROR_INTERNAL);
5612            }            }
# Line 5239  for (;;) Line 5619  for (;;)
5619            RMATCH(eptr, ecode, offset_top, md, eptrb, RM44);            RMATCH(eptr, ecode, offset_top, md, eptrb, RM44);
5620            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
5621            if (eptr-- == pp) break;        /* Stop if tried at original pos */            if (eptr-- == pp) break;        /* Stop if tried at original pos */
5622            if (utf8) BACKCHAR(eptr);            if (utf) BACKCHAR(eptr);
5623            }            }
5624          }          }
5625    
# Line 5250  for (;;) Line 5630  for (;;)
5630          {          {
5631          for (i = min; i < max; i++)          for (i = min; i < max; i++)
5632            {            {
           int len = 1;  
5633            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
5634              {              {
5635              SCHECK_PARTIAL();              SCHECK_PARTIAL();
5636              break;              break;
5637              }              }
5638            if (!utf8) c = *eptr; else { GETCHARLEN(c, eptr, len); }            else
           if (UCD_CATEGORY(c) == ucp_M) break;  
           eptr += len;  
           while (eptr < md->end_subject)  
5639              {              {
5640              len = 1;              int lgb, rgb;
5641              if (!utf8) c = *eptr; else { GETCHARLEN(c, eptr, len); }              GETCHARINCTEST(c, eptr);
5642              if (UCD_CATEGORY(c) != ucp_M) break;              lgb = UCD_GRAPHBREAK(c);
5643              eptr += len;              while (eptr < md->end_subject)
5644                  {
5645                  int len = 1;
5646                  if (!utf) c = *eptr; else { GETCHARLEN(c, eptr, len); }
5647                  rgb = UCD_GRAPHBREAK(c);
5648                  if ((PRIV(ucp_gbtable)[lgb] & (1 << rgb)) == 0) break;
5649                  lgb = rgb;
5650                  eptr += len;
5651                  }
5652              }              }
5653              CHECK_PARTIAL();
5654            }            }
5655    
5656          /* eptr is now past the end of the maximum run */          /* eptr is now past the end of the maximum run */
# Line 5279  for (;;) Line 5664  for (;;)
5664            if (eptr-- == pp) break;        /* Stop if tried at original pos */            if (eptr-- == pp) break;        /* Stop if tried at original pos */
5665            for (;;)                        /* Move back over one extended */            for (;;)                        /* Move back over one extended */
5666              {              {
5667              if (!utf8) c = *eptr; else              if (!utf) c = *eptr; else
5668                {                {
5669                BACKCHAR(eptr);                BACKCHAR(eptr);
5670                GETCHAR(c, eptr);                GETCHAR(c, eptr);
# Line 5293  for (;;) Line 5678  for (;;)
5678        else        else
5679  #endif   /* SUPPORT_UCP */  #endif   /* SUPPORT_UCP */
5680    
5681  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
5682        /* UTF-8 mode */        if (utf)
   
       if (utf8)  
5683          {          {
5684          switch(ctype)          switch(ctype)
5685            {            {
# Line 5311  for (;;) Line 5694  for (;;)
5694                  break;                  break;
5695                  }                  }
5696                if (IS_NEWLINE(eptr)) break;                if (IS_NEWLINE(eptr)) break;
5697                  if (md->partial != 0 &&    /* Take care with CRLF partial */
5698                      eptr + 1 >= md->end_subject &&
5699                      NLBLOCK->nltype == NLTYPE_FIXED &&
5700                      NLBLOCK->nllen == 2 &&
5701                      RAWUCHAR(eptr) == NLBLOCK->nl[0])
5702                    {
5703                    md->hitend = TRUE;
5704                    if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
5705                    }
5706                eptr++;                eptr++;
5707                while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;                ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);
5708                }                }
5709              }              }
5710    
# Line 5328  for (;;) Line 5720  for (;;)
5720                  break;                  break;
5721                  }                  }
5722                if (IS_NEWLINE(eptr)) break;                if (IS_NEWLINE(eptr)) break;