/[pcre]/code/trunk/pcre_exec.c
ViewVC logotype

Diff of /code/trunk/pcre_exec.c

Parent Directory Parent Directory | Revision Log Revision Log | View Patch Patch

revision 788 by ph10, Tue Dec 6 15:38:01 2011 UTC revision 1376 by ph10, Sat Oct 12 18:02:11 2013 UTC
# Line 6  Line 6 
6  and semantics are as close as possible to those of the Perl 5 language.  and semantics are as close as possible to those of the Perl 5 language.
7    
8                         Written by Philip Hazel                         Written by Philip Hazel
9             Copyright (c) 1997-2011 University of Cambridge             Copyright (c) 1997-2013 University of Cambridge
10    
11  -----------------------------------------------------------------------------  -----------------------------------------------------------------------------
12  Redistribution and use in source and binary forms, with or without  Redistribution and use in source and binary forms, with or without
# Line 37  POSSIBILITY OF SUCH DAMAGE. Line 37  POSSIBILITY OF SUCH DAMAGE.
37  -----------------------------------------------------------------------------  -----------------------------------------------------------------------------
38  */  */
39    
   
40  /* This module contains pcre_exec(), the externally visible function that does  /* This module contains pcre_exec(), the externally visible function that does
41  pattern matching using an NFA algorithm, trying to mimic Perl as closely as  pattern matching using an NFA algorithm, trying to mimic Perl as closely as
42  possible. There are also some static supporting functions. */  possible. There are also some static supporting functions. */
# Line 57  possible. There are also some static sup Line 56  possible. There are also some static sup
56  #undef min  #undef min
57  #undef max  #undef max
58    
59    /* The md->capture_last field uses the lower 16 bits for the last captured
60    substring (which can never be greater than 65535) and a bit in the top half
61    to mean "capture vector overflowed". This odd way of doing things was
62    implemented when it was realized that preserving and restoring the overflow bit
63    whenever the last capture number was saved/restored made for a neater
64    interface, and doing it this way saved on (a) another variable, which would
65    have increased the stack frame size (a big NO-NO in PCRE) and (b) another
66    separate set of save/restore instructions. The following defines are used in
67    implementing this. */
68    
69    #define CAPLMASK    0x0000ffff    /* The bits used for last_capture */
70    #define OVFLMASK    0xffff0000    /* The bits used for the overflow flag */
71    #define OVFLBIT     0x00010000    /* The bit that is set for overflow */
72    
73  /* Values for setting in md->match_function_type to indicate two special types  /* Values for setting in md->match_function_type to indicate two special types
74  of call to match(). We do it this way to save on using another stack variable,  of call to match(). We do it this way to save on using another stack variable,
75  as stack usage is to be discouraged. */  as stack usage is to be discouraged. */
# Line 74  defined PCRE_ERROR_xxx codes, which are Line 87  defined PCRE_ERROR_xxx codes, which are
87  negative to avoid the external error codes. */  negative to avoid the external error codes. */
88    
89  #define MATCH_ACCEPT       (-999)  #define MATCH_ACCEPT       (-999)
90  #define MATCH_COMMIT       (-998)  #define MATCH_KETRPOS      (-998)
91  #define MATCH_KETRPOS      (-997)  #define MATCH_ONCE         (-997)
92  #define MATCH_ONCE         (-996)  /* The next 5 must be kept together and in sequence so that a test that checks
93    for any one of them can use a range. */
94    #define MATCH_COMMIT       (-996)
95  #define MATCH_PRUNE        (-995)  #define MATCH_PRUNE        (-995)
96  #define MATCH_SKIP         (-994)  #define MATCH_SKIP         (-994)
97  #define MATCH_SKIP_ARG     (-993)  #define MATCH_SKIP_ARG     (-993)
98  #define MATCH_THEN         (-992)  #define MATCH_THEN         (-992)
99    #define MATCH_BACKTRACK_MAX MATCH_THEN
100    #define MATCH_BACKTRACK_MIN MATCH_COMMIT
101    
102  /* Maximum number of ints of offset to save on the stack for recursive calls.  /* Maximum number of ints of offset to save on the stack for recursive calls.
103  If the offset vector is bigger, malloc is used. This should be a multiple of 3,  If the offset vector is bigger, malloc is used. This should be a multiple of 3,
# Line 93  because the offset vector is always a mu Line 110  because the offset vector is always a mu
110  static const char rep_min[] = { 0, 0, 1, 1, 0, 0 };  static const char rep_min[] = { 0, 0, 1, 1, 0, 0 };
111  static const char rep_max[] = { 0, 0, 0, 0, 1, 1 };  static const char rep_max[] = { 0, 0, 0, 0, 1, 1 };
112    
   
   
113  #ifdef PCRE_DEBUG  #ifdef PCRE_DEBUG
114  /*************************************************  /*************************************************
115  *        Debugging function to print chars       *  *        Debugging function to print chars       *
# Line 113  Returns:     nothing Line 128  Returns:     nothing
128  */  */
129    
130  static void  static void
131  pchars(const uschar *p, int length, BOOL is_subject, match_data *md)  pchars(const pcre_uchar *p, int length, BOOL is_subject, match_data *md)
132  {  {
133  unsigned int c;  pcre_uint32 c;
134    BOOL utf = md->utf;
135  if (is_subject && length > md->end_subject - p) length = md->end_subject - p;  if (is_subject && length > md->end_subject - p) length = md->end_subject - p;
136  while (length-- > 0)  while (length-- > 0)
137    if (isprint(c = *(p++))) printf("%c", c); else printf("\\x%02x", c);    if (isprint(c = RAWUCHARINCTEST(p))) printf("%c", (char)c); else printf("\\x{%02x}", c);
138  }  }
139  #endif  #endif
140    
# Line 140  Arguments: Line 156  Arguments:
156    md          points to match data block    md          points to match data block
157    caseless    TRUE if caseless    caseless    TRUE if caseless
158    
159  Returns:      < 0 if not matched, otherwise the number of subject bytes matched  Returns:      >= 0 the number of subject bytes matched
160                  -1 no match
161                  -2 partial match; always given if at end subject
162  */  */
163    
164  static int  static int
165  match_ref(int offset, register USPTR eptr, int length, match_data *md,  match_ref(int offset, register PCRE_PUCHAR eptr, int length, match_data *md,
166    BOOL caseless)    BOOL caseless)
167  {  {
168  USPTR eptr_start = eptr;  PCRE_PUCHAR eptr_start = eptr;
169  register USPTR p = md->start_subject + md->offset_vector[offset];  register PCRE_PUCHAR p = md->start_subject + md->offset_vector[offset];
170    #ifdef SUPPORT_UTF
171    BOOL utf = md->utf;
172    #endif
173    
174  #ifdef PCRE_DEBUG  #ifdef PCRE_DEBUG
175  if (eptr >= md->end_subject)  if (eptr >= md->end_subject)
# Line 163  pchars(p, length, FALSE, md); Line 184  pchars(p, length, FALSE, md);
184  printf("\n");  printf("\n");
185  #endif  #endif
186    
187  /* Always fail if reference not set (and not JavaScript compatible). */  /* Always fail if reference not set (and not JavaScript compatible - in that
188    case the length is passed as zero). */
189    
190  if (length < 0) return -1;  if (length < 0) return -1;
191    
# Line 173  ASCII characters. */ Line 195  ASCII characters. */
195    
196  if (caseless)  if (caseless)
197    {    {
198  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
199  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
200    if (md->utf8)    if (utf)
201      {      {
202      /* Match characters up to the end of the reference. NOTE: the number of      /* Match characters up to the end of the reference. NOTE: the number of
203      bytes matched may differ, because there are some characters whose upper and      data units matched may differ, because in UTF-8 there are some characters
204      lower case versions code as different numbers of bytes. For example, U+023A      whose upper and lower case versions code have different numbers of bytes.
205      (2 bytes in UTF-8) is the upper case version of U+2C65 (3 bytes in UTF-8);      For example, U+023A (2 bytes in UTF-8) is the upper case version of U+2C65
206      a sequence of 3 of the former uses 6 bytes, as does a sequence of two of      (3 bytes in UTF-8); a sequence of 3 of the former uses 6 bytes, as does a
207      the latter. It is important, therefore, to check the length along the      sequence of two of the latter. It is important, therefore, to check the
208      reference, not along the subject (earlier code did this wrong). */      length along the reference, not along the subject (earlier code did this
209        wrong). */
210    
211      USPTR endptr = p + length;      PCRE_PUCHAR endptr = p + length;
212      while (p < endptr)      while (p < endptr)
213        {        {
214        int c, d;        pcre_uint32 c, d;
215        if (eptr >= md->end_subject) return -1;        const ucd_record *ur;
216          if (eptr >= md->end_subject) return -2;   /* Partial match */
217        GETCHARINC(c, eptr);        GETCHARINC(c, eptr);
218        GETCHARINC(d, p);        GETCHARINC(d, p);
219        if (c != d && c != UCD_OTHERCASE(d)) return -1;        ur = GET_UCD(d);
220          if (c != d && c != d + ur->other_case)
221            {
222            const pcre_uint32 *pp = PRIV(ucd_caseless_sets) + ur->caseset;
223            for (;;)
224              {
225              if (c < *pp) return -1;
226              if (c == *pp++) break;
227              }
228            }
229        }        }
230      }      }
231    else    else
# Line 202  if (caseless) Line 235  if (caseless)
235    /* The same code works when not in UTF-8 mode and in UTF-8 mode when there    /* The same code works when not in UTF-8 mode and in UTF-8 mode when there
236    is no UCP support. */    is no UCP support. */
237      {      {
     if (eptr + length > md->end_subject) return -1;  
238      while (length-- > 0)      while (length-- > 0)
239        { if (md->lcc[*p++] != md->lcc[*eptr++]) return -1; }        {
240          pcre_uint32 cc, cp;
241          if (eptr >= md->end_subject) return -2;   /* Partial match */
242          cc = RAWUCHARTEST(eptr);
243          cp = RAWUCHARTEST(p);
244          if (TABLE_GET(cp, md->lcc, cp) != TABLE_GET(cc, md->lcc, cc)) return -1;
245          p++;
246          eptr++;
247          }
248      }      }
249    }    }
250    
# Line 213  are in UTF-8 mode. */ Line 253  are in UTF-8 mode. */
253    
254  else  else
255    {    {
256    if (eptr + length > md->end_subject) return -1;    while (length-- > 0)
257    while (length-- > 0) if (*p++ != *eptr++) return -1;      {
258        if (eptr >= md->end_subject) return -2;   /* Partial match */
259        if (RAWUCHARINCTEST(p) != RAWUCHARINCTEST(eptr)) return -1;
260        }
261    }    }
262    
263  return (int)(eptr - eptr_start);  return (int)(eptr - eptr_start);
# Line 269  enum { RM1=1, RM2,  RM3,  RM4,  RM5,  RM Line 312  enum { RM1=1, RM2,  RM3,  RM4,  RM5,  RM
312         RM31,  RM32, RM33, RM34, RM35, RM36, RM37, RM38, RM39, RM40,         RM31,  RM32, RM33, RM34, RM35, RM36, RM37, RM38, RM39, RM40,
313         RM41,  RM42, RM43, RM44, RM45, RM46, RM47, RM48, RM49, RM50,         RM41,  RM42, RM43, RM44, RM45, RM46, RM47, RM48, RM49, RM50,
314         RM51,  RM52, RM53, RM54, RM55, RM56, RM57, RM58, RM59, RM60,         RM51,  RM52, RM53, RM54, RM55, RM56, RM57, RM58, RM59, RM60,
315         RM61,  RM62, RM63, RM64, RM65, RM66 };         RM61,  RM62, RM63, RM64, RM65, RM66, RM67, RM68 };
316    
317  /* These versions of the macros use the stack, as normal. There are debugging  /* These versions of the macros use the stack, as normal. There are debugging
318  versions and production versions. Note that the "rw" argument of RMATCH isn't  versions and production versions. Note that the "rw" argument of RMATCH isn't
# Line 287  actually used in this definition. */ Line 330  actually used in this definition. */
330    }    }
331  #define RRETURN(ra) \  #define RRETURN(ra) \
332    { \    { \
333    printf("match() returned %d from line %d ", ra, __LINE__); \    printf("match() returned %d from line %d\n", ra, __LINE__); \
334    return ra; \    return ra; \
335    }    }
336  #else  #else
# Line 307  argument of match(), which never changes Line 350  argument of match(), which never changes
350    
351  #define RMATCH(ra,rb,rc,rd,re,rw)\  #define RMATCH(ra,rb,rc,rd,re,rw)\
352    {\    {\
353    heapframe *newframe = (heapframe *)(pcre_stack_malloc)(sizeof(heapframe));\    heapframe *newframe = frame->Xnextframe;\
354    if (newframe == NULL) RRETURN(PCRE_ERROR_NOMEMORY);\    if (newframe == NULL)\
355    frame->Xwhere = rw; \      {\
356        newframe = (heapframe *)(PUBL(stack_malloc))(sizeof(heapframe));\
357        if (newframe == NULL) RRETURN(PCRE_ERROR_NOMEMORY);\
358        newframe->Xnextframe = NULL;\
359        frame->Xnextframe = newframe;\
360        }\
361      frame->Xwhere = rw;\
362    newframe->Xeptr = ra;\    newframe->Xeptr = ra;\
363    newframe->Xecode = rb;\    newframe->Xecode = rb;\
364    newframe->Xmstart = mstart;\    newframe->Xmstart = mstart;\
# Line 328  argument of match(), which never changes Line 377  argument of match(), which never changes
377    {\    {\
378    heapframe *oldframe = frame;\    heapframe *oldframe = frame;\
379    frame = oldframe->Xprevframe;\    frame = oldframe->Xprevframe;\
   (pcre_stack_free)(oldframe);\  
380    if (frame != NULL)\    if (frame != NULL)\
381      {\      {\
382      rrc = ra;\      rrc = ra;\
# Line 342  argument of match(), which never changes Line 390  argument of match(), which never changes
390    
391  typedef struct heapframe {  typedef struct heapframe {
392    struct heapframe *Xprevframe;    struct heapframe *Xprevframe;
393      struct heapframe *Xnextframe;
394    
395    /* Function arguments that may change */    /* Function arguments that may change */
396    
397    USPTR Xeptr;    PCRE_PUCHAR Xeptr;
398    const uschar *Xecode;    const pcre_uchar *Xecode;
399    USPTR Xmstart;    PCRE_PUCHAR Xmstart;
400    int Xoffset_top;    int Xoffset_top;
401    eptrblock *Xeptrb;    eptrblock *Xeptrb;
402    unsigned int Xrdepth;    unsigned int Xrdepth;
403    
404    /* Function local variables */    /* Function local variables */
405    
406    USPTR Xcallpat;    PCRE_PUCHAR Xcallpat;
407  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
408    USPTR Xcharptr;    PCRE_PUCHAR Xcharptr;
409  #endif  #endif
410    USPTR Xdata;    PCRE_PUCHAR Xdata;
411    USPTR Xnext;    PCRE_PUCHAR Xnext;
412    USPTR Xpp;    PCRE_PUCHAR Xpp;
413    USPTR Xprev;    PCRE_PUCHAR Xprev;
414    USPTR Xsaved_eptr;    PCRE_PUCHAR Xsaved_eptr;
415    
416    recursion_info Xnew_recursive;    recursion_info Xnew_recursive;
417    
# Line 372  typedef struct heapframe { Line 421  typedef struct heapframe {
421    
422  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
423    int Xprop_type;    int Xprop_type;
424    int Xprop_value;    unsigned int Xprop_value;
425    int Xprop_fail_result;    int Xprop_fail_result;
426    int Xoclength;    int Xoclength;
427    uschar Xocchars[8];    pcre_uchar Xocchars[6];
428  #endif  #endif
429    
430    int Xcodelink;    int Xcodelink;
# Line 385  typedef struct heapframe { Line 434  typedef struct heapframe {
434    int Xlength;    int Xlength;
435    int Xmax;    int Xmax;
436    int Xmin;    int Xmin;
437    int Xnumber;    unsigned int Xnumber;
438    int Xoffset;    int Xoffset;
439    int Xop;    unsigned int Xop;
440    int Xsave_capture_last;    pcre_int32 Xsave_capture_last;
441    int Xsave_offset1, Xsave_offset2, Xsave_offset3;    int Xsave_offset1, Xsave_offset2, Xsave_offset3;
442    int Xstacksave[REC_STACK_SAVE_MAX];    int Xstacksave[REC_STACK_SAVE_MAX];
443    
# Line 440  the subject. */ Line 489  the subject. */
489    
490    
491  /* Performance note: It might be tempting to extract commonly used fields from  /* Performance note: It might be tempting to extract commonly used fields from
492  the md structure (e.g. utf8, end_subject) into individual variables to improve  the md structure (e.g. utf, end_subject) into individual variables to improve
493  performance. Tests using gcc on a SPARC disproved this; in the first case, it  performance. Tests using gcc on a SPARC disproved this; in the first case, it
494  made performance worse.  made performance worse.
495    
# Line 463  Returns:       MATCH_MATCH if matched Line 512  Returns:       MATCH_MATCH if matched
512  */  */
513    
514  static int  static int
515  match(REGISTER USPTR eptr, REGISTER const uschar *ecode, USPTR mstart,  match(REGISTER PCRE_PUCHAR eptr, REGISTER const pcre_uchar *ecode,
516    int offset_top, match_data *md, eptrblock *eptrb, unsigned int rdepth)    PCRE_PUCHAR mstart, int offset_top, match_data *md, eptrblock *eptrb,
517      unsigned int rdepth)
518  {  {
519  /* These variables do not need to be preserved over recursion in this function,  /* These variables do not need to be preserved over recursion in this function,
520  so they can be ordinary variables in all cases. Mark some of them with  so they can be ordinary variables in all cases. Mark some of them with
# Line 472  so they can be ordinary variables in all Line 522  so they can be ordinary variables in all
522    
523  register int  rrc;         /* Returns from recursive calls */  register int  rrc;         /* Returns from recursive calls */
524  register int  i;           /* Used for loops not involving calls to RMATCH() */  register int  i;           /* Used for loops not involving calls to RMATCH() */
525  register unsigned int c;   /* Character values not kept over RMATCH() calls */  register pcre_uint32 c;    /* Character values not kept over RMATCH() calls */
526  register BOOL utf8;        /* Local copy of UTF-8 flag for speed */  register BOOL utf;         /* Local copy of UTF flag for speed */
527    
528  BOOL minimize, possessive; /* Quantifier options */  BOOL minimize, possessive; /* Quantifier options */
529  BOOL caseless;  BOOL caseless;
530  int condcode;  int condcode;
531    
532  /* When recursion is not being used, all "local" variables that have to be  /* When recursion is not being used, all "local" variables that have to be
533  preserved over calls to RMATCH() are part of a "frame" which is obtained from  preserved over calls to RMATCH() are part of a "frame". We set up the top-level
534  heap storage. Set up the top-level frame here; others are obtained from the  frame on the stack here; subsequent instantiations are obtained from the heap
535  heap whenever RMATCH() does a "recursion". See the macro definitions above. */  whenever RMATCH() does a "recursion". See the macro definitions above. Putting
536    the top-level on the stack rather than malloc-ing them all gives a performance
537    boost in many cases where there is not much "recursion". */
538    
539  #ifdef NO_RECURSE  #ifdef NO_RECURSE
540  heapframe *frame = (heapframe *)(pcre_stack_malloc)(sizeof(heapframe));  heapframe *frame = (heapframe *)md->match_frames_base;
 if (frame == NULL) RRETURN(PCRE_ERROR_NOMEMORY);  
 frame->Xprevframe = NULL;            /* Marks the top level */  
541    
542  /* Copy in the original argument variables */  /* Copy in the original argument variables */
543    
# Line 513  HEAP_RECURSE: Line 563  HEAP_RECURSE:
563    
564  /* Ditto for the local variables */  /* Ditto for the local variables */
565    
566  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
567  #define charptr            frame->Xcharptr  #define charptr            frame->Xcharptr
568  #endif  #endif
569  #define callpat            frame->Xcallpat  #define callpat            frame->Xcallpat
# Line 571  declarations can be cut out in a block. Line 621  declarations can be cut out in a block.
621  below are for variables that do not have to be preserved over a recursive call  below are for variables that do not have to be preserved over a recursive call
622  to RMATCH(). */  to RMATCH(). */
623    
624  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
625  const uschar *charptr;  const pcre_uchar *charptr;
626  #endif  #endif
627  const uschar *callpat;  const pcre_uchar *callpat;
628  const uschar *data;  const pcre_uchar *data;
629  const uschar *next;  const pcre_uchar *next;
630  USPTR         pp;  PCRE_PUCHAR       pp;
631  const uschar *prev;  const pcre_uchar *prev;
632  USPTR         saved_eptr;  PCRE_PUCHAR       saved_eptr;
633    
634  recursion_info new_recursive;  recursion_info new_recursive;
635    
# Line 589  BOOL prev_is_word; Line 639  BOOL prev_is_word;
639    
640  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
641  int prop_type;  int prop_type;
642  int prop_value;  unsigned int prop_value;
643  int prop_fail_result;  int prop_fail_result;
644  int oclength;  int oclength;
645  uschar occhars[8];  pcre_uchar occhars[6];
646  #endif  #endif
647    
648  int codelink;  int codelink;
# Line 600  int ctype; Line 650  int ctype;
650  int length;  int length;
651  int max;  int max;
652  int min;  int min;
653  int number;  unsigned int number;
654  int offset;  int offset;
655  int op;  unsigned int op;
656  int save_capture_last;  pcre_int32 save_capture_last;
657  int save_offset1, save_offset2, save_offset3;  int save_offset1, save_offset2, save_offset3;
658  int stacksave[REC_STACK_SAVE_MAX];  int stacksave[REC_STACK_SAVE_MAX];
659    
660  eptrblock newptrb;  eptrblock newptrb;
661    
662    /* There is a special fudge for calling match() in a way that causes it to
663    measure the size of its basic stack frame when the stack is being used for
664    recursion. The second argument (ecode) being NULL triggers this behaviour. It
665    cannot normally ever be NULL. The return is the negated value of the frame
666    size. */
667    
668    if (ecode == NULL)
669      {
670      if (rdepth == 0)
671        return match((PCRE_PUCHAR)&rdepth, NULL, NULL, 0, NULL, NULL, 1);
672      else
673        {
674        int len = (char *)&rdepth - (char *)eptr;
675        return (len > 0)? -len : len;
676        }
677      }
678  #endif     /* NO_RECURSE */  #endif     /* NO_RECURSE */
679    
680  /* To save space on the stack and in the heap frame, I have doubled up on some  /* To save space on the stack and in the heap frame, I have doubled up on some
# Line 620  the alternative names that are used. */ Line 687  the alternative names that are used. */
687  #define code_offset   codelink  #define code_offset   codelink
688  #define condassert    condition  #define condassert    condition
689  #define matched_once  prev_is_word  #define matched_once  prev_is_word
690    #define foc           number
691    #define save_mark     data
692    
693  /* These statements are here to stop the compiler complaining about unitialized  /* These statements are here to stop the compiler complaining about unitialized
694  variables. */  variables. */
# Line 645  defined). However, RMATCH isn't like a f Line 714  defined). However, RMATCH isn't like a f
714  complicated macro. It has to be used in one particular way. This shouldn't,  complicated macro. It has to be used in one particular way. This shouldn't,
715  however, impact performance when true recursion is being used. */  however, impact performance when true recursion is being used. */
716    
717  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
718  utf8 = md->utf8;       /* Local copy of the flag */  utf = md->utf;       /* Local copy of the flag */
719  #else  #else
720  utf8 = FALSE;  utf = FALSE;
721  #endif  #endif
722    
723  /* First check that we haven't called match() too many times, or that we  /* First check that we haven't called match() too many times, or that we
# Line 689  for (;;) Line 758  for (;;)
758      case OP_MARK:      case OP_MARK:
759      md->nomatch_mark = ecode + 2;      md->nomatch_mark = ecode + 2;
760      md->mark = NULL;    /* In case previously set by assertion */      md->mark = NULL;    /* In case previously set by assertion */
761      RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode] + ecode[1], offset_top, md,      RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode] + ecode[1], offset_top, md,
762        eptrb, RM55);        eptrb, RM55);
763      if ((rrc == MATCH_MATCH || rrc == MATCH_ACCEPT) &&      if ((rrc == MATCH_MATCH || rrc == MATCH_ACCEPT) &&
764           md->mark == NULL) md->mark = ecode + 2;           md->mark == NULL) md->mark = ecode + 2;
# Line 702  for (;;) Line 771  for (;;)
771      unaltered. */      unaltered. */
772    
773      else if (rrc == MATCH_SKIP_ARG &&      else if (rrc == MATCH_SKIP_ARG &&
774          strcmp((char *)(ecode + 2), (char *)(md->start_match_ptr)) == 0)          STRCMP_UC_UC_TEST(ecode + 2, md->start_match_ptr) == 0)
775        {        {
776        md->start_match_ptr = eptr;        md->start_match_ptr = eptr;
777        RRETURN(MATCH_SKIP);        RRETURN(MATCH_SKIP);
# Line 712  for (;;) Line 781  for (;;)
781      case OP_FAIL:      case OP_FAIL:
782      RRETURN(MATCH_NOMATCH);      RRETURN(MATCH_NOMATCH);
783    
     /* COMMIT overrides PRUNE, SKIP, and THEN */  
   
784      case OP_COMMIT:      case OP_COMMIT:
785      RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,      RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md,
786        eptrb, RM52);        eptrb, RM52);
787      if (rrc != MATCH_NOMATCH && rrc != MATCH_PRUNE &&      if (rrc != MATCH_NOMATCH) RRETURN(rrc);
         rrc != MATCH_SKIP && rrc != MATCH_SKIP_ARG &&  
         rrc != MATCH_THEN)  
       RRETURN(rrc);  
788      RRETURN(MATCH_COMMIT);      RRETURN(MATCH_COMMIT);
789    
     /* PRUNE overrides THEN */  
   
790      case OP_PRUNE:      case OP_PRUNE:
791      RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,      RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md,
792        eptrb, RM51);        eptrb, RM51);
793      if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN) RRETURN(rrc);      if (rrc != MATCH_NOMATCH) RRETURN(rrc);
794      RRETURN(MATCH_PRUNE);      RRETURN(MATCH_PRUNE);
795    
796      case OP_PRUNE_ARG:      case OP_PRUNE_ARG:
797      md->nomatch_mark = ecode + 2;      md->nomatch_mark = ecode + 2;
798      md->mark = NULL;    /* In case previously set by assertion */      md->mark = NULL;    /* In case previously set by assertion */
799      RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode] + ecode[1], offset_top, md,      RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode] + ecode[1], offset_top, md,
800        eptrb, RM56);        eptrb, RM56);
801      if ((rrc == MATCH_MATCH || rrc == MATCH_ACCEPT) &&      if ((rrc == MATCH_MATCH || rrc == MATCH_ACCEPT) &&
802           md->mark == NULL) md->mark = ecode + 2;           md->mark == NULL) md->mark = ecode + 2;
803      if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN) RRETURN(rrc);      if (rrc != MATCH_NOMATCH) RRETURN(rrc);
804      RRETURN(MATCH_PRUNE);      RRETURN(MATCH_PRUNE);
805    
     /* SKIP overrides PRUNE and THEN */  
   
806      case OP_SKIP:      case OP_SKIP:
807      RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,      RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md,
808        eptrb, RM53);        eptrb, RM53);
809      if (rrc != MATCH_NOMATCH && rrc != MATCH_PRUNE && rrc != MATCH_THEN)      if (rrc != MATCH_NOMATCH) RRETURN(rrc);
       RRETURN(rrc);  
810      md->start_match_ptr = eptr;   /* Pass back current position */      md->start_match_ptr = eptr;   /* Pass back current position */
811      RRETURN(MATCH_SKIP);      RRETURN(MATCH_SKIP);
812    
813      /* Note that, for Perl compatibility, SKIP with an argument does NOT set      /* Note that, for Perl compatibility, SKIP with an argument does NOT set
814      nomatch_mark. There is a flag that disables this opcode when re-matching a      nomatch_mark. When a pattern match ends with a SKIP_ARG for which there was
815      pattern that ended with a SKIP for which there was not a matching MARK. */      not a matching mark, we have to re-run the match, ignoring the SKIP_ARG
816        that failed and any that precede 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);
       RRETURN(rrc);  
831    
832      /* Pass back the current skip name by overloading md->start_match_ptr and      /* Pass back the current skip name by overloading md->start_match_ptr and
833      returning the special MATCH_SKIP_ARG return code. This will either be      returning the special MATCH_SKIP_ARG return code. This will either be
834      caught by a matching MARK, or get to the top, where it causes a rematch      caught by a matching MARK, or get to the top, where it causes a rematch
835      with the md->ignore_skip_arg flag set. */      with md->ignore_skip_arg set to the value of md->skip_arg_count. */
836    
837      md->start_match_ptr = ecode + 2;      md->start_match_ptr = ecode + 2;
838      RRETURN(MATCH_SKIP_ARG);      RRETURN(MATCH_SKIP_ARG);
# Line 779  for (;;) Line 842  for (;;)
842      match pointer to do this. */      match pointer to do this. */
843    
844      case OP_THEN:      case OP_THEN:
845      RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,      RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md,
846        eptrb, RM54);        eptrb, RM54);
847      if (rrc != MATCH_NOMATCH) RRETURN(rrc);      if (rrc != MATCH_NOMATCH) RRETURN(rrc);
848      md->start_match_ptr = ecode;      md->start_match_ptr = ecode;
# Line 788  for (;;) Line 851  for (;;)
851      case OP_THEN_ARG:      case OP_THEN_ARG:
852      md->nomatch_mark = ecode + 2;      md->nomatch_mark = ecode + 2;
853      md->mark = NULL;    /* In case previously set by assertion */      md->mark = NULL;    /* In case previously set by assertion */
854      RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode] + ecode[1], offset_top,      RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode] + ecode[1], offset_top,
855        md, eptrb, RM58);        md, eptrb, RM58);
856      if ((rrc == MATCH_MATCH || rrc == MATCH_ACCEPT) &&      if ((rrc == MATCH_MATCH || rrc == MATCH_ACCEPT) &&
857           md->mark == NULL) md->mark = ecode + 2;           md->mark == NULL) md->mark = ecode + 2;
# Line 812  for (;;) Line 875  for (;;)
875      case OP_ONCE_NC:      case OP_ONCE_NC:
876      prev = ecode;      prev = ecode;
877      saved_eptr = eptr;      saved_eptr = eptr;
878        save_mark = md->mark;
879      do      do
880        {        {
881        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, eptrb, RM64);        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, eptrb, RM64);
# Line 830  for (;;) Line 894  for (;;)
894    
895        if (rrc != MATCH_NOMATCH) RRETURN(rrc);        if (rrc != MATCH_NOMATCH) RRETURN(rrc);
896        ecode += GET(ecode,1);        ecode += GET(ecode,1);
897          md->mark = save_mark;
898        }        }
899      while (*ecode == OP_ALT);      while (*ecode == OP_ALT);
900    
# Line 869  for (;;) Line 934  for (;;)
934        }        }
935      else  /* OP_KETRMAX */      else  /* OP_KETRMAX */
936        {        {
       md->match_function_type = MATCH_CBEGROUP;  
937        RMATCH(eptr, prev, offset_top, md, eptrb, RM66);        RMATCH(eptr, prev, offset_top, md, eptrb, RM66);
938        if (rrc != MATCH_NOMATCH) RRETURN(rrc);        if (rrc != MATCH_NOMATCH) RRETURN(rrc);
939        ecode += 1 + LINK_SIZE;        ecode += 1 + LINK_SIZE;
# Line 909  for (;;) Line 973  for (;;)
973        save_offset2 = md->offset_vector[offset+1];        save_offset2 = md->offset_vector[offset+1];
974        save_offset3 = md->offset_vector[md->offset_end - number];        save_offset3 = md->offset_vector[md->offset_end - number];
975        save_capture_last = md->capture_last;        save_capture_last = md->capture_last;
976          save_mark = md->mark;
977    
978        DPRINTF(("saving %d %d %d\n", save_offset1, save_offset2, save_offset3));        DPRINTF(("saving %d %d %d\n", save_offset1, save_offset2, save_offset3));
979        md->offset_vector[md->offset_end - number] =        md->offset_vector[md->offset_end - number] =
# Line 917  for (;;) Line 982  for (;;)
982        for (;;)        for (;;)
983          {          {
984          if (op >= OP_SBRA) md->match_function_type = MATCH_CBEGROUP;          if (op >= OP_SBRA) md->match_function_type = MATCH_CBEGROUP;
985          RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,          RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md,
986            eptrb, RM1);            eptrb, RM1);
987          if (rrc == MATCH_ONCE) break;  /* Backing up through an atomic group */          if (rrc == MATCH_ONCE) break;  /* Backing up through an atomic group */
988    
# Line 945  for (;;) Line 1010  for (;;)
1010          if (rrc != MATCH_NOMATCH) RRETURN(rrc);          if (rrc != MATCH_NOMATCH) RRETURN(rrc);
1011          md->capture_last = save_capture_last;          md->capture_last = save_capture_last;
1012          ecode += GET(ecode, 1);          ecode += GET(ecode, 1);
1013            md->mark = save_mark;
1014          if (*ecode != OP_ALT) break;          if (*ecode != OP_ALT) break;
1015          }          }
1016    
# Line 996  for (;;) Line 1062  for (;;)
1062    
1063      for (;;)      for (;;)
1064        {        {
1065        if (op >= OP_SBRA || op == OP_ONCE) md->match_function_type = MATCH_CBEGROUP;        if (op >= OP_SBRA || op == OP_ONCE)
1066            md->match_function_type = MATCH_CBEGROUP;
1067    
1068        /* If this is not a possibly empty group, and there are no (*THEN)s in        /* If this is not a possibly empty group, and there are no (*THEN)s in
1069        the pattern, and this is the final alternative, optimize as described        the pattern, and this is the final alternative, optimize as described
# Line 1004  for (;;) Line 1071  for (;;)
1071    
1072        else if (!md->hasthen && ecode[GET(ecode, 1)] != OP_ALT)        else if (!md->hasthen && ecode[GET(ecode, 1)] != OP_ALT)
1073          {          {
1074          ecode += _pcre_OP_lengths[*ecode];          ecode += PRIV(OP_lengths)[*ecode];
1075          goto TAIL_RECURSE;          goto TAIL_RECURSE;
1076          }          }
1077    
1078        /* In all other cases, we have to make another call to match(). */        /* In all other cases, we have to make another call to match(). */
1079    
1080        RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md, eptrb,        save_mark = md->mark;
1081          save_capture_last = md->capture_last;
1082          RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md, eptrb,
1083          RM2);          RM2);
1084    
1085        /* See comment in the code for capturing groups above about handling        /* See comment in the code for capturing groups above about handling
# Line 1028  for (;;) Line 1097  for (;;)
1097          {          {
1098          if (rrc == MATCH_ONCE)          if (rrc == MATCH_ONCE)
1099            {            {
1100            const uschar *scode = ecode;            const pcre_uchar *scode = ecode;
1101            if (*scode != OP_ONCE)           /* If not at start, find it */            if (*scode != OP_ONCE)           /* If not at start, find it */
1102              {              {
1103              while (*scode == OP_ALT) scode += GET(scode, 1);              while (*scode == OP_ALT) scode += GET(scode, 1);
# Line 1039  for (;;) Line 1108  for (;;)
1108          RRETURN(rrc);          RRETURN(rrc);
1109          }          }
1110        ecode += GET(ecode, 1);        ecode += GET(ecode, 1);
1111          md->mark = save_mark;
1112        if (*ecode != OP_ALT) break;        if (*ecode != OP_ALT) break;
1113          md->capture_last = save_capture_last;
1114        }        }
1115    
1116      RRETURN(MATCH_NOMATCH);      RRETURN(MATCH_NOMATCH);
# Line 1093  for (;;) Line 1164  for (;;)
1164          md->offset_vector[md->offset_end - number] =          md->offset_vector[md->offset_end - number] =
1165            (int)(eptr - md->start_subject);            (int)(eptr - md->start_subject);
1166          if (op >= OP_SBRA) md->match_function_type = MATCH_CBEGROUP;          if (op >= OP_SBRA) md->match_function_type = MATCH_CBEGROUP;
1167          RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,          RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md,
1168            eptrb, RM63);            eptrb, RM63);
1169          if (rrc == MATCH_KETRPOS)          if (rrc == MATCH_KETRPOS)
1170            {            {
# Line 1161  for (;;) Line 1232  for (;;)
1232      POSSESSIVE_NON_CAPTURE:      POSSESSIVE_NON_CAPTURE:
1233      matched_once = FALSE;      matched_once = FALSE;
1234      code_offset = (int)(ecode - md->start_code);      code_offset = (int)(ecode - md->start_code);
1235        save_capture_last = md->capture_last;
1236    
1237      for (;;)      for (;;)
1238        {        {
1239        if (op >= OP_SBRA) md->match_function_type = MATCH_CBEGROUP;        if (op >= OP_SBRA) md->match_function_type = MATCH_CBEGROUP;
1240        RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,        RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md,
1241          eptrb, RM48);          eptrb, RM48);
1242        if (rrc == MATCH_KETRPOS)        if (rrc == MATCH_KETRPOS)
1243          {          {
# Line 1190  for (;;) Line 1262  for (;;)
1262        if (rrc != MATCH_NOMATCH) RRETURN(rrc);        if (rrc != MATCH_NOMATCH) RRETURN(rrc);
1263        ecode += GET(ecode, 1);        ecode += GET(ecode, 1);
1264        if (*ecode != OP_ALT) break;        if (*ecode != OP_ALT) break;
1265          md->capture_last = save_capture_last;
1266        }        }
1267    
1268      if (matched_once || allow_zero)      if (matched_once || allow_zero)
# Line 1201  for (;;) Line 1274  for (;;)
1274    
1275      /* Control never reaches here. */      /* Control never reaches here. */
1276    
1277      /* Conditional group: compilation checked that there are no more than      /* Conditional group: compilation checked that there are no more than two
1278      two branches. If the condition is false, skipping the first branch takes us      branches. If the condition is false, skipping the first branch takes us
1279      past the end if there is only one branch, but that's OK because that is      past the end of the item if there is only one branch, but that's exactly
1280      exactly what going to the ket would do. */      what we want. */
1281    
1282      case OP_COND:      case OP_COND:
1283      case OP_SCOND:      case OP_SCOND:
1284      codelink = GET(ecode, 1);  
1285        /* The variable codelink will be added to ecode when the condition is
1286        false, to get to the second branch. Setting it to the offset to the ALT
1287        or KET, then incrementing ecode achieves this effect. We now have ecode
1288        pointing to the condition or callout. */
1289    
1290        codelink = GET(ecode, 1);   /* Offset to the second branch */
1291        ecode += 1 + LINK_SIZE;     /* From this opcode */
1292    
1293      /* Because of the way auto-callout works during compile, a callout item is      /* Because of the way auto-callout works during compile, a callout item is
1294      inserted between OP_COND and an assertion condition. */      inserted between OP_COND and an assertion condition. */
1295    
1296      if (ecode[LINK_SIZE+1] == OP_CALLOUT)      if (*ecode == OP_CALLOUT)
1297        {        {
1298        if (pcre_callout != NULL)        if (PUBL(callout) != NULL)
1299          {          {
1300          pcre_callout_block cb;          PUBL(callout_block) cb;
1301          cb.version          = 2;   /* Version 1 of the callout block */          cb.version          = 2;   /* Version 1 of the callout block */
1302          cb.callout_number   = ecode[LINK_SIZE+2];          cb.callout_number   = ecode[1];
1303          cb.offset_vector    = md->offset_vector;          cb.offset_vector    = md->offset_vector;
1304    #if defined COMPILE_PCRE8
1305          cb.subject          = (PCRE_SPTR)md->start_subject;          cb.subject          = (PCRE_SPTR)md->start_subject;
1306    #elif defined COMPILE_PCRE16
1307            cb.subject          = (PCRE_SPTR16)md->start_subject;
1308    #elif defined COMPILE_PCRE32
1309            cb.subject          = (PCRE_SPTR32)md->start_subject;
1310    #endif
1311          cb.subject_length   = (int)(md->end_subject - md->start_subject);          cb.subject_length   = (int)(md->end_subject - md->start_subject);
1312          cb.start_match      = (int)(mstart - md->start_subject);          cb.start_match      = (int)(mstart - md->start_subject);
1313          cb.current_position = (int)(eptr - md->start_subject);          cb.current_position = (int)(eptr - md->start_subject);
1314          cb.pattern_position = GET(ecode, LINK_SIZE + 3);          cb.pattern_position = GET(ecode, 2);
1315          cb.next_item_length = GET(ecode, 3 + 2*LINK_SIZE);          cb.next_item_length = GET(ecode, 2 + LINK_SIZE);
1316          cb.capture_top      = offset_top/2;          cb.capture_top      = offset_top/2;
1317          cb.capture_last     = md->capture_last;          cb.capture_last     = md->capture_last & CAPLMASK;
1318            /* Internal change requires this for API compatibility. */
1319            if (cb.capture_last == 0) cb.capture_last = -1;
1320          cb.callout_data     = md->callout_data;          cb.callout_data     = md->callout_data;
1321          cb.mark             = md->nomatch_mark;          cb.mark             = md->nomatch_mark;
1322          if ((rrc = (*pcre_callout)(&cb)) > 0) RRETURN(MATCH_NOMATCH);          if ((rrc = (*PUBL(callout))(&cb)) > 0) RRETURN(MATCH_NOMATCH);
1323          if (rrc < 0) RRETURN(rrc);          if (rrc < 0) RRETURN(rrc);
1324          }          }
       ecode += _pcre_OP_lengths[OP_CALLOUT];  
       }  
1325    
1326      condcode = ecode[LINK_SIZE+1];        /* Advance ecode past the callout, so it now points to the condition. We
1327          must adjust codelink so that the value of ecode+codelink is unchanged. */
1328    
1329      /* Now see what the actual condition is */        ecode += PRIV(OP_lengths)[OP_CALLOUT];
1330          codelink -= PRIV(OP_lengths)[OP_CALLOUT];
1331          }
1332    
1333        /* Test the various possible conditions */
1334    
1335      if (condcode == OP_RREF || condcode == OP_NRREF)    /* Recursion test */      condition = FALSE;
1336        switch(condcode = *ecode)
1337        {        {
1338        if (md->recursive == NULL)                /* Not recursing => FALSE */        case OP_RREF:         /* Numbered group recursion test */
1339          {        if (md->recursive != NULL)     /* Not recursing => FALSE */
         condition = FALSE;  
         ecode += GET(ecode, 1);  
         }  
       else  
1340          {          {
1341          int recno = GET2(ecode, LINK_SIZE + 2);   /* Recursion group number*/          unsigned int recno = GET2(ecode, 1);   /* Recursion group number*/
1342          condition = (recno == RREF_ANY || recno == md->recursive->group_num);          condition = (recno == RREF_ANY || recno == md->recursive->group_num);
1343            }
1344          break;
1345    
1346          /* If the test is for recursion into a specific subpattern, and it is        case OP_DNRREF:       /* Duplicate named group recursion test */
1347          false, but the test was set up by name, scan the table to see if the        if (md->recursive != NULL)
1348          name refers to any other numbers, and test them. The condition is true          {
1349          if any one is set. */          int count = GET2(ecode, 1 + IMM2_SIZE);
1350            pcre_uchar *slot = md->name_table + GET2(ecode, 1) * md->name_entry_size;
1351          if (!condition && condcode == OP_NRREF)          while (count-- > 0)
1352            {            {
1353            uschar *slotA = md->name_table;            unsigned int recno = GET2(slot, 0);
1354            for (i = 0; i < md->name_count; i++)            condition = recno == md->recursive->group_num;
1355              {            if (condition) break;
1356              if (GET2(slotA, 0) == recno) break;            slot += md->name_entry_size;
             slotA += md->name_entry_size;  
             }  
   
           /* Found a name for the number - there can be only one; duplicate  
           names for different numbers are allowed, but not vice versa. First  
           scan down for duplicates. */  
   
           if (i < md->name_count)  
             {  
             uschar *slotB = slotA;  
             while (slotB > md->name_table)  
               {  
               slotB -= md->name_entry_size;  
               if (strcmp((char *)slotA + 2, (char *)slotB + 2) == 0)  
                 {  
                 condition = GET2(slotB, 0) == md->recursive->group_num;  
                 if (condition) break;  
                 }  
               else break;  
               }  
   
             /* Scan up for duplicates */  
   
             if (!condition)  
               {  
               slotB = slotA;  
               for (i++; i < md->name_count; i++)  
                 {  
                 slotB += md->name_entry_size;  
                 if (strcmp((char *)slotA + 2, (char *)slotB + 2) == 0)  
                   {  
                   condition = GET2(slotB, 0) == md->recursive->group_num;  
                   if (condition) break;  
                   }  
                 else break;  
                 }  
               }  
             }  
1357            }            }
   
         /* Chose branch according to the condition */  
   
         ecode += condition? 3 : GET(ecode, 1);  
1358          }          }
1359        }        break;
1360    
1361      else if (condcode == OP_CREF || condcode == OP_NCREF)  /* Group used test */        case OP_CREF:         /* Numbered group used test */
1362        {        offset = GET2(ecode, 1) << 1;  /* Doubled ref number */
       offset = GET2(ecode, LINK_SIZE+2) << 1;  /* Doubled ref number */  
1363        condition = offset < offset_top && md->offset_vector[offset] >= 0;        condition = offset < offset_top && md->offset_vector[offset] >= 0;
1364          break;
1365    
1366        /* If the numbered capture is unset, but the reference was by name,        case OP_DNCREF:      /* Duplicate named group used test */
       scan the table to see if the name refers to any other numbers, and test  
       them. The condition is true if any one is set. This is tediously similar  
       to the code above, but not close enough to try to amalgamate. */  
   
       if (!condition && condcode == OP_NCREF)  
1367          {          {
1368          int refno = offset >> 1;          int count = GET2(ecode, 1 + IMM2_SIZE);
1369          uschar *slotA = md->name_table;          pcre_uchar *slot = md->name_table + GET2(ecode, 1) * md->name_entry_size;
1370            while (count-- > 0)
1371          for (i = 0; i < md->name_count; i++)            {
1372            {            offset = GET2(slot, 0) << 1;
1373            if (GET2(slotA, 0) == refno) break;            condition = offset < offset_top && md->offset_vector[offset] >= 0;
1374            slotA += md->name_entry_size;            if (condition) break;
1375            }            slot += md->name_entry_size;
   
         /* Found a name for the number - there can be only one; duplicate names  
         for different numbers are allowed, but not vice versa. First scan down  
         for duplicates. */  
   
         if (i < md->name_count)  
           {  
           uschar *slotB = slotA;  
           while (slotB > md->name_table)  
             {  
             slotB -= md->name_entry_size;  
             if (strcmp((char *)slotA + 2, (char *)slotB + 2) == 0)  
               {  
               offset = GET2(slotB, 0) << 1;  
               condition = offset < offset_top &&  
                 md->offset_vector[offset] >= 0;  
               if (condition) break;  
               }  
             else break;  
             }  
   
           /* Scan up for duplicates */  
   
           if (!condition)  
             {  
             slotB = slotA;  
             for (i++; i < md->name_count; i++)  
               {  
               slotB += md->name_entry_size;  
               if (strcmp((char *)slotA + 2, (char *)slotB + 2) == 0)  
                 {  
                 offset = GET2(slotB, 0) << 1;  
                 condition = offset < offset_top &&  
                   md->offset_vector[offset] >= 0;  
                 if (condition) break;  
                 }  
               else break;  
               }  
             }  
1376            }            }
1377          }          }
1378          break;
1379    
1380        /* Chose branch according to the condition */        case OP_DEF:     /* DEFINE - always false */
1381          break;
       ecode += condition? 3 : GET(ecode, 1);  
       }  
   
     else if (condcode == OP_DEF)     /* DEFINE - always false */  
       {  
       condition = FALSE;  
       ecode += GET(ecode, 1);  
       }  
1382    
1383      /* The condition is an assertion. Call match() to evaluate it - setting        /* The condition is an assertion. Call match() to evaluate it - setting
1384      md->match_function_type to MATCH_CONDASSERT causes it to stop at the end of        md->match_function_type to MATCH_CONDASSERT causes it to stop at the end
1385      an assertion. */        of an assertion. */
1386    
1387      else        default:
       {  
1388        md->match_function_type = MATCH_CONDASSERT;        md->match_function_type = MATCH_CONDASSERT;
1389        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, NULL, RM3);        RMATCH(eptr, ecode, offset_top, md, NULL, RM3);
1390        if (rrc == MATCH_MATCH)        if (rrc == MATCH_MATCH)
1391          {          {
1392          if (md->end_offset_top > offset_top)          if (md->end_offset_top > offset_top)
1393            offset_top = md->end_offset_top;  /* Captures may have happened */            offset_top = md->end_offset_top;  /* Captures may have happened */
1394          condition = TRUE;          condition = TRUE;
1395          ecode += 1 + LINK_SIZE + GET(ecode, LINK_SIZE + 2);  
1396            /* Advance ecode past the assertion to the start of the first branch,
1397            but adjust it so that the general choosing code below works. */
1398    
1399            ecode += GET(ecode, 1);
1400          while (*ecode == OP_ALT) ecode += GET(ecode, 1);          while (*ecode == OP_ALT) ecode += GET(ecode, 1);
1401            ecode += 1 + LINK_SIZE - PRIV(OP_lengths)[condcode];
1402          }          }
1403    
1404        /* PCRE doesn't allow the effect of (*THEN) to escape beyond an        /* PCRE doesn't allow the effect of (*THEN) to escape beyond an
1405        assertion; it is therefore treated as NOMATCH. */        assertion; it is therefore treated as NOMATCH. Any other return is an
1406          error. */
1407    
1408        else if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN)        else if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN)
1409          {          {
1410          RRETURN(rrc);         /* Need braces because of following else */          RRETURN(rrc);         /* Need braces because of following else */
1411          }          }
1412        else        break;
         {  
         condition = FALSE;  
         ecode += codelink;  
         }  
1413        }        }
1414    
1415      /* We are now at the branch that is to be obeyed. As there is only one, can      /* Choose branch according to the condition */
1416      use tail recursion to avoid using another stack frame, except when there is  
1417      unlimited repeat of a possibly empty group. In the latter case, a recursive      ecode += condition? PRIV(OP_lengths)[condcode] : codelink;
     call to match() is always required, unless the second alternative doesn't  
     exist, in which case we can just plough on. Note that, for compatibility  
     with Perl, the | in a conditional group is NOT treated as creating two  
     alternatives. If a THEN is encountered in the branch, it propagates out to  
     the enclosing alternative (unless nested in a deeper set of alternatives,  
     of course). */  
1418    
1419      if (condition || *ecode == OP_ALT)      /* We are now at the branch that is to be obeyed. As there is only one, we
1420        can use tail recursion to avoid using another stack frame, except when
1421        there is unlimited repeat of a possibly empty group. In the latter case, a
1422        recursive call to match() is always required, unless the second alternative
1423        doesn't exist, in which case we can just plough on. Note that, for
1424        compatibility with Perl, the | in a conditional group is NOT treated as
1425        creating two alternatives. If a THEN is encountered in the branch, it
1426        propagates out to the enclosing alternative (unless nested in a deeper set
1427        of alternatives, of course). */
1428    
1429        if (condition || ecode[-(1+LINK_SIZE)] == OP_ALT)
1430        {        {
1431        if (op != OP_SCOND)        if (op != OP_SCOND)
1432          {          {
         ecode += 1 + LINK_SIZE;  
1433          goto TAIL_RECURSE;          goto TAIL_RECURSE;
1434          }          }
1435    
1436        md->match_function_type = MATCH_CBEGROUP;        md->match_function_type = MATCH_CBEGROUP;
1437        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, eptrb, RM49);        RMATCH(eptr, ecode, offset_top, md, eptrb, RM49);
1438        RRETURN(rrc);        RRETURN(rrc);
1439        }        }
1440    
# Line 1441  for (;;) Line 1442  for (;;)
1442    
1443      else      else
1444        {        {
       ecode += 1 + LINK_SIZE;  
1445        }        }
1446      break;      break;
1447    
# Line 1450  for (;;) Line 1450  for (;;)
1450      to close any currently open capturing brackets. */      to close any currently open capturing brackets. */
1451    
1452      case OP_CLOSE:      case OP_CLOSE:
1453      number = GET2(ecode, 1);      number = GET2(ecode, 1);   /* Must be less than 65536 */
1454      offset = number << 1;      offset = number << 1;
1455    
1456  #ifdef PCRE_DEBUG  #ifdef PCRE_DEBUG
# Line 1458  for (;;) Line 1458  for (;;)
1458        printf("\n");        printf("\n");
1459  #endif  #endif
1460    
1461      md->capture_last = number;      md->capture_last = (md->capture_last & OVFLMASK) | number;
1462      if (offset >= md->offset_max) md->offset_overflow = TRUE; else      if (offset >= md->offset_max) md->capture_last |= OVFLBIT; else
1463        {        {
1464        md->offset_vector[offset] =        md->offset_vector[offset] =
1465          md->offset_vector[md->offset_end - number];          md->offset_vector[md->offset_end - number];
1466        md->offset_vector[offset+1] = (int)(eptr - md->start_subject);        md->offset_vector[offset+1] = (int)(eptr - md->start_subject);
1467        if (offset_top <= offset) offset_top = offset + 2;        if (offset_top <= offset) offset_top = offset + 2;
1468        }        }
1469      ecode += 3;      ecode += 1 + IMM2_SIZE;
1470      break;      break;
1471    
1472    
# Line 1513  for (;;) Line 1513  for (;;)
1513    
1514      case OP_ASSERT:      case OP_ASSERT:
1515      case OP_ASSERTBACK:      case OP_ASSERTBACK:
1516        save_mark = md->mark;
1517      if (md->match_function_type == MATCH_CONDASSERT)      if (md->match_function_type == MATCH_CONDASSERT)
1518        {        {
1519        condassert = TRUE;        condassert = TRUE;
# Line 1520  for (;;) Line 1521  for (;;)
1521        }        }
1522      else condassert = FALSE;      else condassert = FALSE;
1523    
1524        /* Loop for each branch */
1525    
1526      do      do
1527        {        {
1528        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, NULL, RM4);        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, NULL, RM4);
1529    
1530          /* A match means that the assertion is true; break out of the loop
1531          that matches its alternatives. */
1532    
1533        if (rrc == MATCH_MATCH || rrc == MATCH_ACCEPT)        if (rrc == MATCH_MATCH || rrc == MATCH_ACCEPT)
1534          {          {
1535          mstart = md->start_match_ptr;   /* In case \K reset it */          mstart = md->start_match_ptr;   /* In case \K reset it */
1536          break;          break;
1537          }          }
1538    
1539        /* PCRE does not allow THEN to escape beyond an assertion; it is treated        /* If not matched, restore the previous mark setting. */
1540        as NOMATCH. */  
1541          md->mark = save_mark;
1542    
1543          /* See comment in the code for capturing groups above about handling
1544          THEN. */
1545    
1546          if (rrc == MATCH_THEN)
1547            {
1548            next = ecode + GET(ecode,1);
1549            if (md->start_match_ptr < next &&
1550                (*ecode == OP_ALT || *next == OP_ALT))
1551              rrc = MATCH_NOMATCH;
1552            }
1553    
1554          /* Anything other than NOMATCH causes the entire assertion to fail,
1555          passing back the return code. This includes COMMIT, SKIP, PRUNE and an
1556          uncaptured THEN, which means they take their normal effect. This
1557          consistent approach does not always have exactly the same effect as in
1558          Perl. */
1559    
1560        if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN) RRETURN(rrc);        if (rrc != MATCH_NOMATCH) RRETURN(rrc);
1561        ecode += GET(ecode, 1);        ecode += GET(ecode, 1);
1562        }        }
1563      while (*ecode == OP_ALT);      while (*ecode == OP_ALT);   /* Continue for next alternative */
1564    
1565        /* If we have tried all the alternative branches, the assertion has
1566        failed. If not, we broke out after a match. */
1567    
1568      if (*ecode == OP_KET) RRETURN(MATCH_NOMATCH);      if (*ecode == OP_KET) RRETURN(MATCH_NOMATCH);
1569    
# Line 1543  for (;;) Line 1571  for (;;)
1571    
1572      if (condassert) RRETURN(MATCH_MATCH);      if (condassert) RRETURN(MATCH_MATCH);
1573    
1574      /* Continue from after the assertion, updating the offsets high water      /* Continue from after a successful assertion, updating the offsets high
1575      mark, since extracts may have been taken during the assertion. */      water mark, since extracts may have been taken during the assertion. */
1576    
1577      do ecode += GET(ecode,1); while (*ecode == OP_ALT);      do ecode += GET(ecode,1); while (*ecode == OP_ALT);
1578      ecode += 1 + LINK_SIZE;      ecode += 1 + LINK_SIZE;
1579      offset_top = md->end_offset_top;      offset_top = md->end_offset_top;
1580      continue;      continue;
1581    
1582      /* Negative assertion: all branches must fail to match. Encountering SKIP,      /* Negative assertion: all branches must fail to match for the assertion to
1583      PRUNE, or COMMIT means we must assume failure without checking subsequent      succeed. */
     branches. */  
1584    
1585      case OP_ASSERT_NOT:      case OP_ASSERT_NOT:
1586      case OP_ASSERTBACK_NOT:      case OP_ASSERTBACK_NOT:
1587        save_mark = md->mark;
1588      if (md->match_function_type == MATCH_CONDASSERT)      if (md->match_function_type == MATCH_CONDASSERT)
1589        {        {
1590        condassert = TRUE;        condassert = TRUE;
# Line 1564  for (;;) Line 1592  for (;;)
1592        }        }
1593      else condassert = FALSE;      else condassert = FALSE;
1594    
1595        /* Loop for each alternative branch. */
1596    
1597      do      do
1598        {        {
1599        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, NULL, RM5);        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, NULL, RM5);
1600        if (rrc == MATCH_MATCH || rrc == MATCH_ACCEPT) RRETURN(MATCH_NOMATCH);        md->mark = save_mark;   /* Always restore the mark setting */
1601        if (rrc == MATCH_SKIP || rrc == MATCH_PRUNE || rrc == MATCH_COMMIT)  
1602          switch(rrc)
1603          {          {
1604          do ecode += GET(ecode,1); while (*ecode == OP_ALT);          case MATCH_MATCH:            /* A successful match means */
1605            case MATCH_ACCEPT:           /* the assertion has failed. */
1606            RRETURN(MATCH_NOMATCH);
1607    
1608            case MATCH_NOMATCH:          /* Carry on with next branch */
1609          break;          break;
1610    
1611            /* See comment in the code for capturing groups above about handling
1612            THEN. */
1613    
1614            case MATCH_THEN:
1615            next = ecode + GET(ecode,1);
1616            if (md->start_match_ptr < next &&
1617                (*ecode == OP_ALT || *next == OP_ALT))
1618              {
1619              rrc = MATCH_NOMATCH;
1620              break;
1621              }
1622            /* Otherwise fall through. */
1623    
1624            /* COMMIT, SKIP, PRUNE, and an uncaptured THEN cause the whole
1625            assertion to fail to match, without considering any more alternatives.
1626            Failing to match means the assertion is true. This is a consistent
1627            approach, but does not always have the same effect as in Perl. */
1628    
1629            case MATCH_COMMIT:
1630            case MATCH_SKIP:
1631            case MATCH_SKIP_ARG:
1632            case MATCH_PRUNE:
1633            do ecode += GET(ecode,1); while (*ecode == OP_ALT);
1634            goto NEG_ASSERT_TRUE;   /* Break out of alternation loop */
1635    
1636            /* Anything else is an error */
1637    
1638            default:
1639            RRETURN(rrc);
1640          }          }
1641    
1642        /* PCRE does not allow THEN to escape beyond an assertion; it is treated        /* Continue with next branch */
       as NOMATCH. */  
1643    
       if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN) RRETURN(rrc);  
1644        ecode += GET(ecode,1);        ecode += GET(ecode,1);
1645        }        }
1646      while (*ecode == OP_ALT);      while (*ecode == OP_ALT);
1647    
1648      if (condassert) RRETURN(MATCH_MATCH);  /* Condition assertion */      /* All branches in the assertion failed to match. */
1649    
1650      ecode += 1 + LINK_SIZE;      NEG_ASSERT_TRUE:
1651        if (condassert) RRETURN(MATCH_MATCH);  /* Condition assertion */
1652        ecode += 1 + LINK_SIZE;                /* Continue with current branch */
1653      continue;      continue;
1654    
1655      /* 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 1658  for (;;)
1658      back a number of characters, not bytes. */      back a number of characters, not bytes. */
1659    
1660      case OP_REVERSE:      case OP_REVERSE:
1661  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
1662      if (utf8)      if (utf)
1663        {        {
1664        i = GET(ecode, 1);        i = GET(ecode, 1);
1665        while (i-- > 0)        while (i-- > 0)
# Line 1625  for (;;) Line 1690  for (;;)
1690      function is able to force a failure. */      function is able to force a failure. */
1691    
1692      case OP_CALLOUT:      case OP_CALLOUT:
1693      if (pcre_callout != NULL)      if (PUBL(callout) != NULL)
1694        {        {
1695        pcre_callout_block cb;        PUBL(callout_block) cb;
1696        cb.version          = 2;   /* Version 1 of the callout block */        cb.version          = 2;   /* Version 1 of the callout block */
1697        cb.callout_number   = ecode[1];        cb.callout_number   = ecode[1];
1698        cb.offset_vector    = md->offset_vector;        cb.offset_vector    = md->offset_vector;
1699    #if defined COMPILE_PCRE8
1700        cb.subject          = (PCRE_SPTR)md->start_subject;        cb.subject          = (PCRE_SPTR)md->start_subject;
1701    #elif defined COMPILE_PCRE16
1702          cb.subject          = (PCRE_SPTR16)md->start_subject;
1703    #elif defined COMPILE_PCRE32
1704          cb.subject          = (PCRE_SPTR32)md->start_subject;
1705    #endif
1706        cb.subject_length   = (int)(md->end_subject - md->start_subject);        cb.subject_length   = (int)(md->end_subject - md->start_subject);
1707        cb.start_match      = (int)(mstart - md->start_subject);        cb.start_match      = (int)(mstart - md->start_subject);
1708        cb.current_position = (int)(eptr - md->start_subject);        cb.current_position = (int)(eptr - md->start_subject);
1709        cb.pattern_position = GET(ecode, 2);        cb.pattern_position = GET(ecode, 2);
1710        cb.next_item_length = GET(ecode, 2 + LINK_SIZE);        cb.next_item_length = GET(ecode, 2 + LINK_SIZE);
1711        cb.capture_top      = offset_top/2;        cb.capture_top      = offset_top/2;
1712        cb.capture_last     = md->capture_last;        cb.capture_last     = md->capture_last & CAPLMASK;
1713          /* Internal change requires this for API compatibility. */
1714          if (cb.capture_last == 0) cb.capture_last = -1;
1715        cb.callout_data     = md->callout_data;        cb.callout_data     = md->callout_data;
1716        cb.mark             = md->nomatch_mark;        cb.mark             = md->nomatch_mark;
1717        if ((rrc = (*pcre_callout)(&cb)) > 0) RRETURN(MATCH_NOMATCH);        if ((rrc = (*PUBL(callout))(&cb)) > 0) RRETURN(MATCH_NOMATCH);
1718        if (rrc < 0) RRETURN(rrc);        if (rrc < 0) RRETURN(rrc);
1719        }        }
1720      ecode += 2 + 2*LINK_SIZE;      ecode += 2 + 2*LINK_SIZE;
# Line 1667  for (;;) Line 1740  for (;;)
1740      case OP_RECURSE:      case OP_RECURSE:
1741        {        {
1742        recursion_info *ri;        recursion_info *ri;
1743        int recno;        unsigned int recno;
1744    
1745        callpat = md->start_code + GET(ecode, 1);        callpat = md->start_code + GET(ecode, 1);
1746        recno = (callpat == md->start_code)? 0 :        recno = (callpat == md->start_code)? 0 :
# Line 1684  for (;;) Line 1757  for (;;)
1757        /* Add to "recursing stack" */        /* Add to "recursing stack" */
1758    
1759        new_recursive.group_num = recno;        new_recursive.group_num = recno;
1760          new_recursive.saved_capture_last = md->capture_last;
1761        new_recursive.subject_position = eptr;        new_recursive.subject_position = eptr;
1762        new_recursive.prevrec = md->recursive;        new_recursive.prevrec = md->recursive;
1763        md->recursive = &new_recursive;        md->recursive = &new_recursive;
# Line 1700  for (;;) Line 1774  for (;;)
1774        else        else
1775          {          {
1776          new_recursive.offset_save =          new_recursive.offset_save =
1777            (int *)(pcre_malloc)(new_recursive.saved_max * sizeof(int));            (int *)(PUBL(malloc))(new_recursive.saved_max * sizeof(int));
1778          if (new_recursive.offset_save == NULL) RRETURN(PCRE_ERROR_NOMEMORY);          if (new_recursive.offset_save == NULL) RRETURN(PCRE_ERROR_NOMEMORY);
1779          }          }
1780        memcpy(new_recursive.offset_save, md->offset_vector,        memcpy(new_recursive.offset_save, md->offset_vector,
1781              new_recursive.saved_max * sizeof(int));              new_recursive.saved_max * sizeof(int));
1782    
1783        /* OK, now we can do the recursion. After processing each alternative,        /* OK, now we can do the recursion. After processing each alternative,
1784        restore the offset data. If there were nested recursions, md->recursive        restore the offset data and the last captured value. If there were nested
1785        might be changed, so reset it before looping. */        recursions, md->recursive might be changed, so reset it before looping.
1786          */
1787    
1788        DPRINTF(("Recursing into group %d\n", new_recursive.group_num));        DPRINTF(("Recursing into group %d\n", new_recursive.group_num));
1789        cbegroup = (*callpat >= OP_SBRA);        cbegroup = (*callpat >= OP_SBRA);
1790        do        do
1791          {          {
1792          if (cbegroup) md->match_function_type = MATCH_CBEGROUP;          if (cbegroup) md->match_function_type = MATCH_CBEGROUP;
1793          RMATCH(eptr, callpat + _pcre_OP_lengths[*callpat], offset_top,          RMATCH(eptr, callpat + PRIV(OP_lengths)[*callpat], offset_top,
1794            md, eptrb, RM6);            md, eptrb, RM6);
1795          memcpy(md->offset_vector, new_recursive.offset_save,          memcpy(md->offset_vector, new_recursive.offset_save,
1796              new_recursive.saved_max * sizeof(int));              new_recursive.saved_max * sizeof(int));
1797            md->capture_last = new_recursive.saved_capture_last;
1798          md->recursive = new_recursive.prevrec;          md->recursive = new_recursive.prevrec;
1799          if (rrc == MATCH_MATCH || rrc == MATCH_ACCEPT)          if (rrc == MATCH_MATCH || rrc == MATCH_ACCEPT)
1800            {            {
1801            DPRINTF(("Recursion matched\n"));            DPRINTF(("Recursion matched\n"));
1802            if (new_recursive.offset_save != stacksave)            if (new_recursive.offset_save != stacksave)
1803              (pcre_free)(new_recursive.offset_save);              (PUBL(free))(new_recursive.offset_save);
1804    
1805            /* 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
1806            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 1811  for (;;)
1811            goto RECURSION_MATCHED;        /* Exit loop; end processing */            goto RECURSION_MATCHED;        /* Exit loop; end processing */
1812            }            }
1813    
1814          /* 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
1815          as NOMATCH. */          recursion; they cause a NOMATCH for the entire recursion. These codes
1816            are defined in a range that can be tested for. */
1817    
1818            if (rrc >= MATCH_BACKTRACK_MIN && rrc <= MATCH_BACKTRACK_MAX)
1819              RRETURN(MATCH_NOMATCH);
1820    
1821            /* Any return code other than NOMATCH is an error. */
1822    
1823          else if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN)          if (rrc != MATCH_NOMATCH)
1824            {            {
1825            DPRINTF(("Recursion gave error %d\n", rrc));            DPRINTF(("Recursion gave error %d\n", rrc));
1826            if (new_recursive.offset_save != stacksave)            if (new_recursive.offset_save != stacksave)
1827              (pcre_free)(new_recursive.offset_save);              (PUBL(free))(new_recursive.offset_save);
1828            RRETURN(rrc);            RRETURN(rrc);
1829            }            }
1830    
# Line 1754  for (;;) Line 1836  for (;;)
1836        DPRINTF(("Recursion didn't match\n"));        DPRINTF(("Recursion didn't match\n"));
1837        md->recursive = new_recursive.prevrec;        md->recursive = new_recursive.prevrec;
1838        if (new_recursive.offset_save != stacksave)        if (new_recursive.offset_save != stacksave)
1839          (pcre_free)(new_recursive.offset_save);          (PUBL(free))(new_recursive.offset_save);
1840        RRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
1841        }        }
1842    
# Line 1868  for (;;) Line 1950  for (;;)
1950    
1951        /* Deal with capturing */        /* Deal with capturing */
1952    
1953        md->capture_last = number;        md->capture_last = (md->capture_last & OVFLMASK) | number;
1954        if (offset >= md->offset_max) md->offset_overflow = TRUE; else        if (offset >= md->offset_max) md->capture_last |= OVFLBIT; else
1955          {          {
1956          /* If offset is greater than offset_top, it means that we are          /* If offset is greater than offset_top, it means that we are
1957          "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 2031  for (;;)
2031          }          }
2032        if (*prev >= OP_SBRA)    /* Could match an empty string */        if (*prev >= OP_SBRA)    /* Could match an empty string */
2033          {          {
         md->match_function_type = MATCH_CBEGROUP;  
2034          RMATCH(eptr, prev, offset_top, md, eptrb, RM50);          RMATCH(eptr, prev, offset_top, md, eptrb, RM50);
2035          RRETURN(rrc);          RRETURN(rrc);
2036          }          }
# Line 1958  for (;;) Line 2039  for (;;)
2039        }        }
2040      else  /* OP_KETRMAX */      else  /* OP_KETRMAX */
2041        {        {
       if (*prev >= OP_SBRA) md->match_function_type = MATCH_CBEGROUP;  
2042        RMATCH(eptr, prev, offset_top, md, eptrb, RM13);        RMATCH(eptr, prev, offset_top, md, eptrb, RM13);
2043        if (rrc == MATCH_ONCE && md->once_target == prev) rrc = MATCH_NOMATCH;        if (rrc == MATCH_ONCE && md->once_target == prev) rrc = MATCH_NOMATCH;
2044        if (rrc != MATCH_NOMATCH) RRETURN(rrc);        if (rrc != MATCH_NOMATCH) RRETURN(rrc);
# Line 2015  for (;;) Line 2095  for (;;)
2095    
2096      case OP_DOLLM:      case OP_DOLLM:
2097      if (eptr < md->end_subject)      if (eptr < md->end_subject)
2098        { if (!IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH); }        {
2099          if (!IS_NEWLINE(eptr))
2100            {
2101            if (md->partial != 0 &&
2102                eptr + 1 >= md->end_subject &&
2103                NLBLOCK->nltype == NLTYPE_FIXED &&
2104                NLBLOCK->nllen == 2 &&
2105                RAWUCHARTEST(eptr) == NLBLOCK->nl[0])
2106              {
2107              md->hitend = TRUE;
2108              if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
2109              }
2110            RRETURN(MATCH_NOMATCH);
2111            }
2112          }
2113      else      else
2114        {        {
2115        if (md->noteol) RRETURN(MATCH_NOMATCH);        if (md->noteol) RRETURN(MATCH_NOMATCH);
# Line 2047  for (;;) Line 2141  for (;;)
2141      ASSERT_NL_OR_EOS:      ASSERT_NL_OR_EOS:
2142      if (eptr < md->end_subject &&      if (eptr < md->end_subject &&
2143          (!IS_NEWLINE(eptr) || eptr != md->end_subject - md->nllen))          (!IS_NEWLINE(eptr) || eptr != md->end_subject - md->nllen))
2144          {
2145          if (md->partial != 0 &&
2146              eptr + 1 >= md->end_subject &&
2147              NLBLOCK->nltype == NLTYPE_FIXED &&
2148              NLBLOCK->nllen == 2 &&
2149              RAWUCHARTEST(eptr) == NLBLOCK->nl[0])
2150            {
2151            md->hitend = TRUE;
2152            if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
2153            }
2154        RRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2155          }
2156    
2157      /* Either at end of string or \n before end. */      /* Either at end of string or \n before end. */
2158    
# Line 2066  for (;;) Line 2171  for (;;)
2171        be "non-word" characters. Remember the earliest consulted character for        be "non-word" characters. Remember the earliest consulted character for
2172        partial matching. */        partial matching. */
2173    
2174  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
2175        if (utf8)        if (utf)
2176          {          {
2177          /* Get status of previous character */          /* Get status of previous character */
2178    
2179          if (eptr == md->start_subject) prev_is_word = FALSE; else          if (eptr == md->start_subject) prev_is_word = FALSE; else
2180            {            {
2181            USPTR lastptr = eptr - 1;            PCRE_PUCHAR lastptr = eptr - 1;
2182            while((*lastptr & 0xc0) == 0x80) lastptr--;            BACKCHAR(lastptr);
2183            if (lastptr < md->start_used_ptr) md->start_used_ptr = lastptr;            if (lastptr < md->start_used_ptr) md->start_used_ptr = lastptr;
2184            GETCHAR(c, lastptr);            GETCHAR(c, lastptr);
2185  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
# Line 2139  for (;;) Line 2244  for (;;)
2244              }              }
2245            else            else
2246  #endif  #endif
2247            prev_is_word = ((md->ctypes[eptr[-1]] & ctype_word) != 0);            prev_is_word = MAX_255(eptr[-1])
2248                && ((md->ctypes[eptr[-1]] & ctype_word) != 0);
2249            }            }
2250    
2251          /* Get status of next character */          /* Get status of next character */
# Line 2162  for (;;) Line 2268  for (;;)
2268            }            }
2269          else          else
2270  #endif  #endif
2271          cur_is_word = ((md->ctypes[*eptr] & ctype_word) != 0);          cur_is_word = MAX_255(*eptr)
2272              && ((md->ctypes[*eptr] & ctype_word) != 0);
2273          }          }
2274    
2275        /* Now see if the situation is what we want */        /* Now see if the situation is what we want */
# Line 2173  for (;;) Line 2280  for (;;)
2280        }        }
2281      break;      break;
2282    
2283      /* Match a single character type; inline for speed */      /* Match any single character type except newline; have to take care with
2284        CRLF newlines and partial matching. */
2285    
2286      case OP_ANY:      case OP_ANY:
2287      if (IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH);      if (IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH);
2288        if (md->partial != 0 &&
2289            eptr + 1 >= md->end_subject &&
2290            NLBLOCK->nltype == NLTYPE_FIXED &&
2291            NLBLOCK->nllen == 2 &&
2292            RAWUCHARTEST(eptr) == NLBLOCK->nl[0])
2293          {
2294          md->hitend = TRUE;
2295          if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
2296          }
2297    
2298      /* Fall through */      /* Fall through */
2299    
2300        /* Match any single character whatsoever. */
2301    
2302      case OP_ALLANY:      case OP_ALLANY:
2303      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 */
2304        {                            /* not be updated before SCHECK_PARTIAL. */        {                            /* not be updated before SCHECK_PARTIAL. */
# Line 2186  for (;;) Line 2306  for (;;)
2306        RRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2307        }        }
2308      eptr++;      eptr++;
2309      if (utf8) while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;  #ifdef SUPPORT_UTF
2310        if (utf) ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);
2311    #endif
2312      ecode++;      ecode++;
2313      break;      break;
2314    
# Line 2211  for (;;) Line 2333  for (;;)
2333        }        }
2334      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2335      if (      if (
2336  #ifdef SUPPORT_UTF8  #if defined SUPPORT_UTF || !(defined COMPILE_PCRE8)
2337         c < 256 &&         c < 256 &&
2338  #endif  #endif
2339         (md->ctypes[c] & ctype_digit) != 0         (md->ctypes[c] & ctype_digit) != 0
# Line 2228  for (;;) Line 2350  for (;;)
2350        }        }
2351      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2352      if (      if (
2353  #ifdef SUPPORT_UTF8  #if defined SUPPORT_UTF || !(defined COMPILE_PCRE8)
2354         c >= 256 ||         c > 255 ||
2355  #endif  #endif
2356         (md->ctypes[c] & ctype_digit) == 0         (md->ctypes[c] & ctype_digit) == 0
2357         )         )
# Line 2245  for (;;) Line 2367  for (;;)
2367        }        }
2368      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2369      if (      if (
2370  #ifdef SUPPORT_UTF8  #if defined SUPPORT_UTF || !(defined COMPILE_PCRE8)
2371         c < 256 &&         c < 256 &&
2372  #endif  #endif
2373         (md->ctypes[c] & ctype_space) != 0         (md->ctypes[c] & ctype_space) != 0
# Line 2262  for (;;) Line 2384  for (;;)
2384        }        }
2385      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2386      if (      if (
2387  #ifdef SUPPORT_UTF8  #if defined SUPPORT_UTF || !(defined COMPILE_PCRE8)
2388         c >= 256 ||         c > 255 ||
2389  #endif  #endif
2390         (md->ctypes[c] & ctype_space) == 0         (md->ctypes[c] & ctype_space) == 0
2391         )         )
# Line 2279  for (;;) Line 2401  for (;;)
2401        }        }
2402      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2403      if (      if (
2404  #ifdef SUPPORT_UTF8  #if defined SUPPORT_UTF || !(defined COMPILE_PCRE8)
2405         c < 256 &&         c < 256 &&
2406  #endif  #endif
2407         (md->ctypes[c] & ctype_word) != 0         (md->ctypes[c] & ctype_word) != 0
# Line 2296  for (;;) Line 2418  for (;;)
2418        }        }
2419      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2420      if (      if (
2421  #ifdef SUPPORT_UTF8  #if defined SUPPORT_UTF || !(defined COMPILE_PCRE8)
2422         c >= 256 ||         c > 255 ||
2423  #endif  #endif
2424         (md->ctypes[c] & ctype_word) == 0         (md->ctypes[c] & ctype_word) == 0
2425         )         )
# Line 2316  for (;;) Line 2438  for (;;)
2438        {        {
2439        default: RRETURN(MATCH_NOMATCH);        default: RRETURN(MATCH_NOMATCH);
2440    
2441        case 0x000d:        case CHAR_CR:
2442        if (eptr < md->end_subject && *eptr == 0x0a) eptr++;        if (eptr >= md->end_subject)
2443            {
2444            SCHECK_PARTIAL();
2445            }
2446          else if (RAWUCHARTEST(eptr) == CHAR_LF) eptr++;
2447        break;        break;
2448    
2449        case 0x000a:        case CHAR_LF:
2450        break;        break;
2451    
2452        case 0x000b:        case CHAR_VT:
2453        case 0x000c:        case CHAR_FF:
2454        case 0x0085:        case CHAR_NEL:
2455    #ifndef EBCDIC
2456        case 0x2028:        case 0x2028:
2457        case 0x2029:        case 0x2029:
2458    #endif  /* Not EBCDIC */
2459        if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);        if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);
2460        break;        break;
2461        }        }
# Line 2343  for (;;) Line 2471  for (;;)
2471      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2472      switch(c)      switch(c)
2473        {        {
2474          HSPACE_CASES: RRETURN(MATCH_NOMATCH);  /* Byte and multibyte cases */
2475        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);  
2476        }        }
2477      ecode++;      ecode++;
2478      break;      break;
# Line 2377  for (;;) Line 2486  for (;;)
2486      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2487      switch(c)      switch(c)
2488        {        {
2489          HSPACE_CASES: break;  /* Byte and multibyte cases */
2490        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;  
2491        }        }
2492      ecode++;      ecode++;
2493      break;      break;
# Line 2411  for (;;) Line 2501  for (;;)
2501      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2502      switch(c)      switch(c)
2503        {        {
2504          VSPACE_CASES: RRETURN(MATCH_NOMATCH);
2505        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);  
2506        }        }
2507      ecode++;      ecode++;
2508      break;      break;
# Line 2433  for (;;) Line 2516  for (;;)
2516      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2517      switch(c)      switch(c)
2518        {        {
2519          VSPACE_CASES: break;
2520        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;  
2521        }        }
2522      ecode++;      ecode++;
2523      break;      break;
# Line 2459  for (;;) Line 2535  for (;;)
2535        }        }
2536      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2537        {        {
2538          const pcre_uint32 *cp;
2539        const ucd_record *prop = GET_UCD(c);        const ucd_record *prop = GET_UCD(c);
2540    
2541        switch(ecode[1])        switch(ecode[1])
# Line 2475  for (;;) Line 2552  for (;;)
2552          break;          break;
2553    
2554          case PT_GC:          case PT_GC:
2555          if ((ecode[2] != _pcre_ucp_gentype[prop->chartype]) == (op == OP_PROP))          if ((ecode[2] != PRIV(ucp_gentype)[prop->chartype]) == (op == OP_PROP))
2556            RRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
2557          break;          break;
2558    
# Line 2492  for (;;) Line 2569  for (;;)
2569          /* These are specials */          /* These are specials */
2570    
2571          case PT_ALNUM:          case PT_ALNUM:
2572          if ((_pcre_ucp_gentype[prop->chartype] == ucp_L ||          if ((PRIV(ucp_gentype)[prop->chartype] == ucp_L ||
2573               _pcre_ucp_gentype[prop->chartype] == ucp_N) == (op == OP_NOTPROP))               PRIV(ucp_gentype)[prop->chartype] == ucp_N) == (op == OP_NOTPROP))
2574            RRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
2575          break;          break;
2576    
2577          case PT_SPACE:    /* Perl space */          /* Perl space used to exclude VT, but from Perl 5.18 it is included,
2578          if ((_pcre_ucp_gentype[prop->chartype] == ucp_Z ||          which means that Perl space and POSIX space are now identical. PCRE
2579               c == CHAR_HT || c == CHAR_NL || c == CHAR_FF || c == CHAR_CR)          was changed at release 8.34. */
                == (op == OP_NOTPROP))  
           RRETURN(MATCH_NOMATCH);  
         break;  
2580    
2581            case PT_SPACE:    /* Perl space */
2582          case PT_PXSPACE:  /* POSIX space */          case PT_PXSPACE:  /* POSIX space */
2583          if ((_pcre_ucp_gentype[prop->chartype] == ucp_Z ||          switch(c)
2584               c == CHAR_HT || c == CHAR_NL || c == CHAR_VT ||            {
2585               c == CHAR_FF || c == CHAR_CR)            HSPACE_CASES:
2586                 == (op == OP_NOTPROP))            VSPACE_CASES:
2587            RRETURN(MATCH_NOMATCH);            if (op == OP_NOTPROP) RRETURN(MATCH_NOMATCH);
2588              break;
2589    
2590              default:
2591              if ((PRIV(ucp_gentype)[prop->chartype] == ucp_Z) ==
2592                (op == OP_NOTPROP)) RRETURN(MATCH_NOMATCH);
2593              break;
2594              }
2595          break;          break;
2596    
2597          case PT_WORD:          case PT_WORD:
2598          if ((_pcre_ucp_gentype[prop->chartype] == ucp_L ||          if ((PRIV(ucp_gentype)[prop->chartype] == ucp_L ||
2599               _pcre_ucp_gentype[prop->chartype] == ucp_N ||               PRIV(ucp_gentype)[prop->chartype] == ucp_N ||
2600               c == CHAR_UNDERSCORE) == (op == OP_NOTPROP))               c == CHAR_UNDERSCORE) == (op == OP_NOTPROP))
2601            RRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
2602          break;          break;
2603    
2604            case PT_CLIST:
2605            cp = PRIV(ucd_caseless_sets) + ecode[2];
2606            for (;;)
2607              {
2608              if (c < *cp)
2609                { if (op == OP_PROP) { RRETURN(MATCH_NOMATCH); } else break; }
2610              if (c == *cp++)
2611                { if (op == OP_PROP) break; else { RRETURN(MATCH_NOMATCH); } }
2612              }
2613            break;
2614    
2615            case PT_UCNC:
2616            if ((c == CHAR_DOLLAR_SIGN || c == CHAR_COMMERCIAL_AT ||
2617                 c == CHAR_GRAVE_ACCENT || (c >= 0xa0 && c <= 0xd7ff) ||
2618                 c >= 0xe000) == (op == OP_NOTPROP))
2619              RRETURN(MATCH_NOMATCH);
2620            break;
2621    
2622          /* This should never occur */          /* This should never occur */
2623    
2624          default:          default:
# Line 2538  for (;;) Line 2638  for (;;)
2638        SCHECK_PARTIAL();        SCHECK_PARTIAL();
2639        RRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2640        }        }
2641      GETCHARINCTEST(c, eptr);      else
     if (UCD_CATEGORY(c) == ucp_M) RRETURN(MATCH_NOMATCH);  
     while (eptr < md->end_subject)  
2642        {        {
2643        int len = 1;        int lgb, rgb;
2644        if (!utf8) c = *eptr; else { GETCHARLEN(c, eptr, len); }        GETCHARINCTEST(c, eptr);
2645        if (UCD_CATEGORY(c) != ucp_M) break;        lgb = UCD_GRAPHBREAK(c);
2646        eptr += len;        while (eptr < md->end_subject)
2647            {
2648            int len = 1;
2649            if (!utf) c = *eptr; else { GETCHARLEN(c, eptr, len); }
2650            rgb = UCD_GRAPHBREAK(c);
2651            if ((PRIV(ucp_gbtable)[lgb] & (1 << rgb)) == 0) break;
2652            lgb = rgb;
2653            eptr += len;
2654            }
2655        }        }
2656        CHECK_PARTIAL();
2657      ecode++;      ecode++;
2658      break;      break;
2659  #endif  #endif  /* SUPPORT_UCP */
2660    
2661    
2662      /* 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 2558  for (;;) Line 2665  for (;;)
2665      similar code to character type repeats - written out again for speed.      similar code to character type repeats - written out again for speed.
2666      However, if the referenced string is the empty string, always treat      However, if the referenced string is the empty string, always treat
2667      it as matched, any number of times (otherwise there could be infinite      it as matched, any number of times (otherwise there could be infinite
2668      loops). */      loops). If the reference is unset, there are two possibilities:
   
     case OP_REF:  
     case OP_REFI:  
     caseless = op == OP_REFI;  
     offset = GET2(ecode, 1) << 1;               /* Doubled ref number */  
     ecode += 3;  
   
     /* If the reference is unset, there are two possibilities:  
2669    
2670      (a) In the default, Perl-compatible state, set the length negative;      (a) In the default, Perl-compatible state, set the length negative;
2671      this ensures that every attempt at a match fails. We can't just fail      this ensures that every attempt at a match fails. We can't just fail
# Line 2576  for (;;) Line 2675  for (;;)
2675      so that the back reference matches an empty string.      so that the back reference matches an empty string.
2676    
2677      Otherwise, set the length to the length of what was matched by the      Otherwise, set the length to the length of what was matched by the
2678      referenced subpattern. */      referenced subpattern.
2679    
2680        The OP_REF and OP_REFI opcodes are used for a reference to a numbered group
2681        or to a non-duplicated named group. For a duplicated named group, OP_DNREF
2682        and OP_DNREFI are used. In this case we must scan the list of groups to
2683        which the name refers, and use the first one that is set. */
2684    
2685        case OP_DNREF:
2686        case OP_DNREFI:
2687        caseless = op == OP_DNREFI;
2688          {
2689          int count = GET2(ecode, 1+IMM2_SIZE);
2690          pcre_uchar *slot = md->name_table + GET2(ecode, 1) * md->name_entry_size;
2691          ecode += 1 + 2*IMM2_SIZE;
2692    
2693          while (count-- > 0)
2694            {
2695            offset = GET2(slot, 0) << 1;
2696            if (offset < offset_top && md->offset_vector[offset] >= 0) break;
2697            slot += md->name_entry_size;
2698            }
2699          if (count < 0)
2700            length = (md->jscript_compat)? 0 : -1;
2701          else
2702            length = md->offset_vector[offset+1] - md->offset_vector[offset];
2703          }
2704        goto REF_REPEAT;
2705    
2706        case OP_REF:
2707        case OP_REFI:
2708        caseless = op == OP_REFI;
2709        offset = GET2(ecode, 1) << 1;               /* Doubled ref number */
2710        ecode += 1 + IMM2_SIZE;
2711      if (offset >= offset_top || md->offset_vector[offset] < 0)      if (offset >= offset_top || md->offset_vector[offset] < 0)
2712        length = (md->jscript_compat)? 0 : -1;        length = (md->jscript_compat)? 0 : -1;
2713      else      else
# Line 2585  for (;;) Line 2715  for (;;)
2715    
2716      /* Set up for repetition, or handle the non-repeated case */      /* Set up for repetition, or handle the non-repeated case */
2717    
2718        REF_REPEAT:
2719      switch (*ecode)      switch (*ecode)
2720        {        {
2721        case OP_CRSTAR:        case OP_CRSTAR:
# Line 2604  for (;;) Line 2735  for (;;)
2735        case OP_CRMINRANGE:        case OP_CRMINRANGE:
2736        minimize = (*ecode == OP_CRMINRANGE);        minimize = (*ecode == OP_CRMINRANGE);
2737        min = GET2(ecode, 1);        min = GET2(ecode, 1);
2738        max = GET2(ecode, 3);        max = GET2(ecode, 1 + IMM2_SIZE);
2739        if (max == 0) max = INT_MAX;        if (max == 0) max = INT_MAX;
2740        ecode += 5;        ecode += 1 + 2 * IMM2_SIZE;
2741        break;        break;
2742    
2743        default:               /* No repeat follows */        default:               /* No repeat follows */
2744        if ((length = match_ref(offset, eptr, length, md, caseless)) < 0)        if ((length = match_ref(offset, eptr, length, md, caseless)) < 0)
2745          {          {
2746            if (length == -2) eptr = md->end_subject;   /* Partial match */
2747          CHECK_PARTIAL();          CHECK_PARTIAL();
2748          RRETURN(MATCH_NOMATCH);          RRETURN(MATCH_NOMATCH);
2749          }          }
# Line 2620  for (;;) Line 2752  for (;;)
2752        }        }
2753    
2754      /* Handle repeated back references. If the length of the reference is      /* Handle repeated back references. If the length of the reference is
2755      zero, just continue with the main loop. */      zero, just continue with the main loop. If the length is negative, it
2756        means the reference is unset in non-Java-compatible mode. If the minimum is
2757        zero, we can continue at the same level without recursion. For any other
2758        minimum, carrying on will result in NOMATCH. */
2759    
2760      if (length == 0) continue;      if (length == 0) continue;
2761        if (length < 0 && min == 0) continue;
2762    
2763      /* First, ensure the minimum number of matches are present. We get back      /* First, ensure the minimum number of matches are present. We get back
2764      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 2769  for (;;)
2769        int slength;        int slength;
2770        if ((slength = match_ref(offset, eptr, length, md, caseless)) < 0)        if ((slength = match_ref(offset, eptr, length, md, caseless)) < 0)
2771          {          {
2772            if (slength == -2) eptr = md->end_subject;   /* Partial match */
2773          CHECK_PARTIAL();          CHECK_PARTIAL();
2774          RRETURN(MATCH_NOMATCH);          RRETURN(MATCH_NOMATCH);
2775          }          }
# Line 2656  for (;;) Line 2793  for (;;)
2793          if (fi >= max) RRETURN(MATCH_NOMATCH);          if (fi >= max) RRETURN(MATCH_NOMATCH);
2794          if ((slength = match_ref(offset, eptr, length, md, caseless)) < 0)          if ((slength = match_ref(offset, eptr, length, md, caseless)) < 0)
2795            {            {
2796              if (slength == -2) eptr = md->end_subject;   /* Partial match */
2797            CHECK_PARTIAL();            CHECK_PARTIAL();
2798            RRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
2799            }            }
# Line 2674  for (;;) Line 2812  for (;;)
2812          int slength;          int slength;
2813          if ((slength = match_ref(offset, eptr, length, md, caseless)) < 0)          if ((slength = match_ref(offset, eptr, length, md, caseless)) < 0)
2814            {            {
2815            CHECK_PARTIAL();            /* Can't use CHECK_PARTIAL because we don't want to update eptr in
2816              the soft partial matching case. */
2817    
2818              if (slength == -2 && md->partial != 0 &&
2819                  md->end_subject > md->start_used_ptr)
2820                {
2821                md->hitend = TRUE;
2822                if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
2823                }
2824            break;            break;
2825            }            }
2826          eptr += slength;          eptr += slength;
2827          }          }
2828    
2829        while (eptr >= pp)        while (eptr >= pp)
2830          {          {
2831          RMATCH(eptr, ecode, offset_top, md, eptrb, RM15);          RMATCH(eptr, ecode, offset_top, md, eptrb, RM15);
# Line 2703  for (;;) Line 2850  for (;;)
2850      case OP_NCLASS:      case OP_NCLASS:
2851      case OP_CLASS:      case OP_CLASS:
2852        {        {
2853          /* The data variable is saved across frames, so the byte map needs to
2854          be stored there. */
2855    #define BYTE_MAP ((pcre_uint8 *)data)
2856        data = ecode + 1;                /* Save for matching */        data = ecode + 1;                /* Save for matching */
2857        ecode += 33;                     /* Advance past the item */        ecode += 1 + (32 / sizeof(pcre_uchar)); /* Advance past the item */
2858    
2859        switch (*ecode)        switch (*ecode)
2860          {          {
# Line 2725  for (;;) Line 2875  for (;;)
2875          case OP_CRMINRANGE:          case OP_CRMINRANGE:
2876          minimize = (*ecode == OP_CRMINRANGE);          minimize = (*ecode == OP_CRMINRANGE);
2877          min = GET2(ecode, 1);          min = GET2(ecode, 1);
2878          max = GET2(ecode, 3);          max = GET2(ecode, 1 + IMM2_SIZE);
2879          if (max == 0) max = INT_MAX;          if (max == 0) max = INT_MAX;
2880          ecode += 5;          ecode += 1 + 2 * IMM2_SIZE;
2881          break;          break;
2882    
2883          default:               /* No repeat follows */          default:               /* No repeat follows */
# Line 2737  for (;;) Line 2887  for (;;)
2887    
2888        /* First, ensure the minimum number of matches are present. */        /* First, ensure the minimum number of matches are present. */
2889    
2890  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
2891        /* UTF-8 mode */        if (utf)
       if (utf8)  
2892          {          {
2893          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
2894            {            {
# Line 2754  for (;;) Line 2903  for (;;)
2903              if (op == OP_CLASS) RRETURN(MATCH_NOMATCH);              if (op == OP_CLASS) RRETURN(MATCH_NOMATCH);
2904              }              }
2905            else            else
2906              {              if ((BYTE_MAP[c/8] & (1 << (c&7))) == 0) RRETURN(MATCH_NOMATCH);
             if ((data[c/8] & (1 << (c&7))) == 0) RRETURN(MATCH_NOMATCH);  
             }  
2907            }            }
2908          }          }
2909        else        else
2910  #endif  #endif
2911        /* Not UTF-8 mode */        /* Not UTF mode */
2912          {          {
2913          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
2914            {            {
# Line 2771  for (;;) Line 2918  for (;;)
2918              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
2919              }              }
2920            c = *eptr++;            c = *eptr++;
2921            if ((data[c/8] & (1 << (c&7))) == 0) RRETURN(MATCH_NOMATCH);  #ifndef COMPILE_PCRE8
2922              if (c > 255)
2923                {
2924                if (op == OP_CLASS) RRETURN(MATCH_NOMATCH);
2925                }
2926              else
2927    #endif
2928                if ((BYTE_MAP[c/8] & (1 << (c&7))) == 0) RRETURN(MATCH_NOMATCH);
2929            }            }
2930          }          }
2931    
# Line 2785  for (;;) Line 2939  for (;;)
2939    
2940        if (minimize)        if (minimize)
2941          {          {
2942  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
2943          /* UTF-8 mode */          if (utf)
         if (utf8)  
2944            {            {
2945            for (fi = min;; fi++)            for (fi = min;; fi++)
2946              {              {
# Line 2805  for (;;) Line 2958  for (;;)
2958                if (op == OP_CLASS) RRETURN(MATCH_NOMATCH);                if (op == OP_CLASS) RRETURN(MATCH_NOMATCH);
2959                }                }
2960              else              else
2961                {                if ((BYTE_MAP[c/8] & (1 << (c&7))) == 0) RRETURN(MATCH_NOMATCH);
               if ((data[c/8] & (1 << (c&7))) == 0) RRETURN(MATCH_NOMATCH);  
               }  
2962              }              }
2963            }            }
2964          else          else
2965  #endif  #endif
2966          /* Not UTF-8 mode */          /* Not UTF mode */
2967            {            {
2968            for (fi = min;; fi++)            for (fi = min;; fi++)
2969              {              {
# Line 2825  for (;;) Line 2976  for (;;)
2976                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
2977                }                }
2978              c = *eptr++;              c = *eptr++;
2979              if ((data[c/8] & (1 << (c&7))) == 0) RRETURN(MATCH_NOMATCH);  #ifndef COMPILE_PCRE8
2980                if (c > 255)
2981                  {
2982                  if (op == OP_CLASS) RRETURN(MATCH_NOMATCH);
2983                  }
2984                else
2985    #endif
2986                  if ((BYTE_MAP[c/8] & (1 << (c&7))) == 0) RRETURN(MATCH_NOMATCH);
2987              }              }
2988            }            }
2989          /* Control never gets here */          /* Control never gets here */
# Line 2837  for (;;) Line 2995  for (;;)
2995          {          {
2996          pp = eptr;          pp = eptr;
2997    
2998  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
2999          /* UTF-8 mode */          if (utf)
         if (utf8)  
3000            {            {
3001            for (i = min; i < max; i++)            for (i = min; i < max; i++)
3002              {              {
# Line 2855  for (;;) Line 3012  for (;;)
3012                if (op == OP_CLASS) break;                if (op == OP_CLASS) break;
3013                }                }
3014              else              else
3015                {                if ((BYTE_MAP[c/8] & (1 << (c&7))) == 0) break;
               if ((data[c/8] & (1 << (c&7))) == 0) break;  
               }  
3016              eptr += len;              eptr += len;
3017              }              }
3018            for (;;)            for (;;)
# Line 2870  for (;;) Line 3025  for (;;)
3025            }            }
3026          else          else
3027  #endif  #endif
3028            /* Not UTF-8 mode */            /* Not UTF mode */
3029            {            {
3030            for (i = min; i < max; i++)            for (i = min; i < max; i++)
3031              {              {
# Line 2880  for (;;) Line 3035  for (;;)
3035                break;                break;
3036                }                }
3037              c = *eptr;              c = *eptr;
3038              if ((data[c/8] & (1 << (c&7))) == 0) break;  #ifndef COMPILE_PCRE8
3039                if (c > 255)
3040                  {
3041                  if (op == OP_CLASS) break;
3042                  }
3043                else
3044    #endif
3045                  if ((BYTE_MAP[c/8] & (1 << (c&7))) == 0) break;
3046              eptr++;              eptr++;
3047              }              }
3048            while (eptr >= pp)            while (eptr >= pp)
# Line 2893  for (;;) Line 3055  for (;;)
3055    
3056          RRETURN(MATCH_NOMATCH);          RRETURN(MATCH_NOMATCH);
3057          }          }
3058    #undef BYTE_MAP
3059        }        }
3060      /* Control never gets here */      /* Control never gets here */
3061    
# Line 2901  for (;;) Line 3064  for (;;)
3064      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
3065      mode, because Unicode properties are supported in non-UTF-8 mode. */      mode, because Unicode properties are supported in non-UTF-8 mode. */
3066    
3067  #ifdef SUPPORT_UTF8  #if defined SUPPORT_UTF || !defined COMPILE_PCRE8
3068      case OP_XCLASS:      case OP_XCLASS:
3069        {        {
3070        data = ecode + 1 + LINK_SIZE;                /* Save for matching */        data = ecode + 1 + LINK_SIZE;                /* Save for matching */
# Line 2926  for (;;) Line 3089  for (;;)
3089          case OP_CRMINRANGE:          case OP_CRMINRANGE:
3090          minimize = (*ecode == OP_CRMINRANGE);          minimize = (*ecode == OP_CRMINRANGE);
3091          min = GET2(ecode, 1);          min = GET2(ecode, 1);
3092          max = GET2(ecode, 3);          max = GET2(ecode, 1 + IMM2_SIZE);
3093          if (max == 0) max = INT_MAX;          if (max == 0) max = INT_MAX;
3094          ecode += 5;          ecode += 1 + 2 * IMM2_SIZE;
3095          break;          break;
3096    
3097          default:               /* No repeat follows */          default:               /* No repeat follows */
# Line 2946  for (;;) Line 3109  for (;;)
3109            RRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
3110            }            }
3111          GETCHARINCTEST(c, eptr);          GETCHARINCTEST(c, eptr);
3112          if (!_pcre_xclass(c, data)) RRETURN(MATCH_NOMATCH);          if (!PRIV(xclass)(c, data, utf)) RRETURN(MATCH_NOMATCH);
3113          }          }
3114    
3115        /* 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 3133  for (;;)
3133              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
3134              }              }
3135            GETCHARINCTEST(c, eptr);            GETCHARINCTEST(c, eptr);
3136            if (!_pcre_xclass(c, data)) RRETURN(MATCH_NOMATCH);            if (!PRIV(xclass)(c, data, utf)) RRETURN(MATCH_NOMATCH);
3137            }            }
3138          /* Control never gets here */          /* Control never gets here */
3139          }          }
# Line 2988  for (;;) Line 3151  for (;;)
3151              SCHECK_PARTIAL();              SCHECK_PARTIAL();
3152              break;              break;
3153              }              }
3154    #ifdef SUPPORT_UTF
3155            GETCHARLENTEST(c, eptr, len);            GETCHARLENTEST(c, eptr, len);
3156            if (!_pcre_xclass(c, data)) break;  #else
3157            eptr += len;            c = *eptr;
3158            }  #endif
3159              if (!PRIV(xclass)(c, data, utf)) break;
3160              eptr += len;
3161              }
3162          for(;;)          for(;;)
3163            {            {
3164            RMATCH(eptr, ecode, offset_top, md, eptrb, RM21);            RMATCH(eptr, ecode, offset_top, md, eptrb, RM21);
3165            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3166            if (eptr-- == pp) break;        /* Stop if tried at original pos */            if (eptr-- == pp) break;        /* Stop if tried at original pos */
3167            if (utf8) BACKCHAR(eptr);  #ifdef SUPPORT_UTF
3168              if (utf) BACKCHAR(eptr);
3169    #endif
3170            }            }
3171          RRETURN(MATCH_NOMATCH);          RRETURN(MATCH_NOMATCH);
3172          }          }
# Line 3009  for (;;) Line 3178  for (;;)
3178      /* Match a single character, casefully */      /* Match a single character, casefully */
3179    
3180      case OP_CHAR:      case OP_CHAR:
3181  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
3182      if (utf8)      if (utf)
3183        {        {
3184        length = 1;        length = 1;
3185        ecode++;        ecode++;
# Line 3020  for (;;) Line 3189  for (;;)
3189          CHECK_PARTIAL();             /* Not SCHECK_PARTIAL() */          CHECK_PARTIAL();             /* Not SCHECK_PARTIAL() */
3190          RRETURN(MATCH_NOMATCH);          RRETURN(MATCH_NOMATCH);
3191          }          }
3192        while (length-- > 0) if (*ecode++ != *eptr++) RRETURN(MATCH_NOMATCH);        while (length-- > 0) if (*ecode++ != RAWUCHARINC(eptr)) RRETURN(MATCH_NOMATCH);
3193        }        }
3194      else      else
3195  #endif  #endif
3196        /* Not UTF mode */
     /* Non-UTF-8 mode */  
3197        {        {
3198        if (md->end_subject - eptr < 1)        if (md->end_subject - eptr < 1)
3199          {          {
# Line 3047  for (;;) Line 3215  for (;;)
3215        RRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
3216        }        }
3217    
3218  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
3219      if (utf8)      if (utf)
3220        {        {
3221        length = 1;        length = 1;
3222        ecode++;        ecode++;
# Line 3061  for (;;) Line 3229  for (;;)
3229    
3230        if (fc < 128)        if (fc < 128)
3231          {          {
3232          if (md->lcc[*ecode++] != md->lcc[*eptr++]) RRETURN(MATCH_NOMATCH);          pcre_uint32 cc = RAWUCHAR(eptr);
3233            if (md->lcc[fc] != TABLE_GET(cc, md->lcc, cc)) RRETURN(MATCH_NOMATCH);
3234            ecode++;
3235            eptr++;
3236          }          }
3237    
3238        /* Otherwise we must pick up the subject character. Note that we cannot        /* Otherwise we must pick up the subject character. Note that we cannot
# Line 3070  for (;;) Line 3241  for (;;)
3241    
3242        else        else
3243          {          {
3244          unsigned int dc;          pcre_uint32 dc;
3245          GETCHARINC(dc, eptr);          GETCHARINC(dc, eptr);
3246          ecode += length;          ecode += length;
3247    
# Line 3087  for (;;) Line 3258  for (;;)
3258          }          }
3259        }        }
3260      else      else
3261  #endif   /* SUPPORT_UTF8 */  #endif   /* SUPPORT_UTF */
3262    
3263      /* Non-UTF-8 mode */      /* Not UTF mode */
3264        {        {
3265        if (md->lcc[ecode[1]] != md->lcc[*eptr++]) RRETURN(MATCH_NOMATCH);        if (TABLE_GET(ecode[1], md->lcc, ecode[1])
3266              != TABLE_GET(*eptr, md->lcc, *eptr)) RRETURN(MATCH_NOMATCH);
3267          eptr++;
3268        ecode += 2;        ecode += 2;
3269        }        }
3270      break;      break;
# Line 3101  for (;;) Line 3274  for (;;)
3274      case OP_EXACT:      case OP_EXACT:
3275      case OP_EXACTI:      case OP_EXACTI:
3276      min = max = GET2(ecode, 1);      min = max = GET2(ecode, 1);
3277      ecode += 3;      ecode += 1 + IMM2_SIZE;
3278      goto REPEATCHAR;      goto REPEATCHAR;
3279    
3280      case OP_POSUPTO:      case OP_POSUPTO:
# Line 3116  for (;;) Line 3289  for (;;)
3289      min = 0;      min = 0;
3290      max = GET2(ecode, 1);      max = GET2(ecode, 1);
3291      minimize = *ecode == OP_MINUPTO || *ecode == OP_MINUPTOI;      minimize = *ecode == OP_MINUPTO || *ecode == OP_MINUPTOI;
3292      ecode += 3;      ecode += 1 + IMM2_SIZE;
3293      goto REPEATCHAR;      goto REPEATCHAR;
3294    
3295      case OP_POSSTAR:      case OP_POSSTAR:
# Line 3161  for (;;) Line 3334  for (;;)
3334      max = rep_max[c];                 /* zero for max => infinity */      max = rep_max[c];                 /* zero for max => infinity */
3335      if (max == 0) max = INT_MAX;      if (max == 0) max = INT_MAX;
3336    
3337      /* Common code for all repeated single-character matches. */      /* Common code for all repeated single-character matches. We first check
3338        for the minimum number of characters. If the minimum equals the maximum, we
3339        are done. Otherwise, if minimizing, check the rest of the pattern for a
3340        match; if there isn't one, advance up to the maximum, one character at a
3341        time.
3342    
3343        If maximizing, advance up to the maximum number of matching characters,
3344        until eptr is past the end of the maximum run. If possessive, we are
3345        then done (no backing up). Otherwise, match at this position; anything
3346        other than no match is immediately returned. For nomatch, back up one
3347        character, unless we are matching \R and the last thing matched was
3348        \r\n, in which case, back up two bytes. When we reach the first optional
3349        character position, we can save stack by doing a tail recurse.
3350    
3351        The various UTF/non-UTF and caseful/caseless cases are handled separately,
3352        for speed. */
3353    
3354      REPEATCHAR:      REPEATCHAR:
3355  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
3356      if (utf8)      if (utf)
3357        {        {
3358        length = 1;        length = 1;
3359        charptr = ecode;        charptr = ecode;
# Line 3178  for (;;) Line 3366  for (;;)
3366        if (length > 1)        if (length > 1)
3367          {          {
3368  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
3369          unsigned int othercase;          pcre_uint32 othercase;
3370          if (op >= OP_STARI &&     /* Caseless */          if (op >= OP_STARI &&     /* Caseless */
3371              (othercase = UCD_OTHERCASE(fc)) != fc)              (othercase = UCD_OTHERCASE(fc)) != fc)
3372            oclength = _pcre_ord2utf8(othercase, occhars);            oclength = PRIV(ord2utf)(othercase, occhars);
3373          else oclength = 0;          else oclength = 0;
3374  #endif  /* SUPPORT_UCP */  #endif  /* SUPPORT_UCP */
3375    
3376          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
3377            {            {
3378            if (eptr <= md->end_subject - length &&            if (eptr <= md->end_subject - length &&
3379              memcmp(eptr, charptr, length) == 0) eptr += length;              memcmp(eptr, charptr, IN_UCHARS(length)) == 0) eptr += length;
3380  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
3381            else if (oclength > 0 &&            else if (oclength > 0 &&
3382                     eptr <= md->end_subject - oclength &&                     eptr <= md->end_subject - oclength &&
3383                     memcmp(eptr, occhars, oclength) == 0) eptr += oclength;                     memcmp(eptr, occhars, IN_UCHARS(oclength)) == 0) eptr += oclength;
3384  #endif  /* SUPPORT_UCP */  #endif  /* SUPPORT_UCP */
3385            else            else
3386              {              {
# Line 3211  for (;;) Line 3399  for (;;)
3399              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3400              if (fi >= max) RRETURN(MATCH_NOMATCH);              if (fi >= max) RRETURN(MATCH_NOMATCH);
3401              if (eptr <= md->end_subject - length &&              if (eptr <= md->end_subject - length &&
3402                memcmp(eptr, charptr, length) == 0) eptr += length;                memcmp(eptr, charptr, IN_UCHARS(length)) == 0) eptr += length;
3403  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
3404              else if (oclength > 0 &&              else if (oclength > 0 &&
3405                       eptr <= md->end_subject - oclength &&                       eptr <= md->end_subject - oclength &&
3406                       memcmp(eptr, occhars, oclength) == 0) eptr += oclength;                       memcmp(eptr, occhars, IN_UCHARS(oclength)) == 0) eptr += oclength;
3407  #endif  /* SUPPORT_UCP */  #endif  /* SUPPORT_UCP */
3408              else              else
3409                {                {
# Line 3232  for (;;) Line 3420  for (;;)
3420            for (i = min; i < max; i++)            for (i = min; i < max; i++)
3421              {              {
3422              if (eptr <= md->end_subject - length &&              if (eptr <= md->end_subject - length &&
3423                  memcmp(eptr, charptr, length) == 0) eptr += length;                  memcmp(eptr, charptr, IN_UCHARS(length)) == 0) eptr += length;
3424  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
3425              else if (oclength > 0 &&              else if (oclength > 0 &&
3426                       eptr <= md->end_subject - oclength &&                       eptr <= md->end_subject - oclength &&
3427                       memcmp(eptr, occhars, oclength) == 0) eptr += oclength;                       memcmp(eptr, occhars, IN_UCHARS(oclength)) == 0) eptr += oclength;
3428  #endif  /* SUPPORT_UCP */  #endif  /* SUPPORT_UCP */
3429              else              else
3430                {                {
# Line 3245  for (;;) Line 3433  for (;;)
3433                }                }
3434              }              }
3435    
3436            if (possessive) continue;            if (possessive) continue;    /* No backtracking */
   
3437            for(;;)            for(;;)
3438              {              {
3439                if (eptr == pp) goto TAIL_RECURSE;
3440              RMATCH(eptr, ecode, offset_top, md, eptrb, RM23);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM23);
3441              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
             if (eptr == pp) { RRETURN(MATCH_NOMATCH); }  
3442  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
3443              eptr--;              eptr--;
3444              BACKCHAR(eptr);              BACKCHAR(eptr);
# Line 3268  for (;;) Line 3455  for (;;)
3455        value of fc will always be < 128. */        value of fc will always be < 128. */
3456        }        }
3457      else      else
3458  #endif  /* SUPPORT_UTF8 */  #endif  /* SUPPORT_UTF */
3459          /* When not in UTF-8 mode, load a single-byte character. */
3460      /* When not in UTF-8 mode, load a single-byte character. */        fc = *ecode++;
3461    
3462      fc = *ecode++;      /* The value of fc at this point is always one character, though we may
3463        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  
3464      caseful cases, for speed, since matching characters is likely to be quite      caseful cases, for speed, since matching characters is likely to be quite
3465      common. First, ensure the minimum number of matches are present. If min =      common. First, ensure the minimum number of matches are present. If min =
3466      max, continue at the same level without recursing. Otherwise, if      max, continue at the same level without recursing. Otherwise, if
# Line 3284  for (;;) Line 3469  for (;;)
3469      maximizing, find the maximum number of characters and work backwards. */      maximizing, find the maximum number of characters and work backwards. */
3470    
3471      DPRINTF(("matching %c{%d,%d} against subject %.*s\n", fc, min, max,      DPRINTF(("matching %c{%d,%d} against subject %.*s\n", fc, min, max,
3472        max, eptr));        max, (char *)eptr));
3473    
3474      if (op >= OP_STARI)  /* Caseless */      if (op >= OP_STARI)  /* Caseless */
3475        {        {
3476        fc = md->lcc[fc];  #ifdef COMPILE_PCRE8
3477          /* fc must be < 128 if UTF is enabled. */
3478          foc = md->fcc[fc];
3479    #else
3480    #ifdef SUPPORT_UTF
3481    #ifdef SUPPORT_UCP
3482          if (utf && fc > 127)
3483            foc = UCD_OTHERCASE(fc);
3484    #else
3485          if (utf && fc > 127)
3486            foc = fc;
3487    #endif /* SUPPORT_UCP */
3488          else
3489    #endif /* SUPPORT_UTF */
3490            foc = TABLE_GET(fc, md->fcc, fc);
3491    #endif /* COMPILE_PCRE8 */
3492    
3493        for (i = 1; i <= min; i++)        for (i = 1; i <= min; i++)
3494          {          {
3495            pcre_uint32 cc;                 /* Faster than pcre_uchar */
3496          if (eptr >= md->end_subject)          if (eptr >= md->end_subject)
3497            {            {
3498            SCHECK_PARTIAL();            SCHECK_PARTIAL();
3499            RRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
3500            }            }
3501          if (fc != md->lcc[*eptr++]) RRETURN(MATCH_NOMATCH);          cc = RAWUCHARTEST(eptr);
3502            if (fc != cc && foc != cc) RRETURN(MATCH_NOMATCH);
3503            eptr++;
3504          }          }
3505        if (min == max) continue;        if (min == max) continue;
3506        if (minimize)        if (minimize)
3507          {          {
3508          for (fi = min;; fi++)          for (fi = min;; fi++)
3509            {            {
3510              pcre_uint32 cc;               /* Faster than pcre_uchar */
3511            RMATCH(eptr, ecode, offset_top, md, eptrb, RM24);            RMATCH(eptr, ecode, offset_top, md, eptrb, RM24);
3512            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3513            if (fi >= max) RRETURN(MATCH_NOMATCH);            if (fi >= max) RRETURN(MATCH_NOMATCH);
# Line 3311  for (;;) Line 3516  for (;;)
3516              SCHECK_PARTIAL();              SCHECK_PARTIAL();
3517              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
3518              }              }
3519            if (fc != md->lcc[*eptr++]) RRETURN(MATCH_NOMATCH);            cc = RAWUCHARTEST(eptr);
3520              if (fc != cc && foc != cc) RRETURN(MATCH_NOMATCH);
3521              eptr++;
3522            }            }
3523          /* Control never gets here */          /* Control never gets here */
3524          }          }
# Line 3320  for (;;) Line 3527  for (;;)
3527          pp = eptr;          pp = eptr;
3528          for (i = min; i < max; i++)          for (i = min; i < max; i++)
3529            {            {
3530              pcre_uint32 cc;               /* Faster than pcre_uchar */
3531            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
3532              {              {
3533              SCHECK_PARTIAL();              SCHECK_PARTIAL();
3534              break;              break;
3535              }              }
3536            if (fc != md->lcc[*eptr]) break;            cc = RAWUCHARTEST(eptr);
3537              if (fc != cc && foc != cc) break;
3538            eptr++;            eptr++;
3539            }            }
3540            if (possessive) continue;       /* No backtracking */
3541          if (possessive) continue;          for (;;)
   
         while (eptr >= pp)  
3542            {            {
3543              if (eptr == pp) goto TAIL_RECURSE;
3544            RMATCH(eptr, ecode, offset_top, md, eptrb, RM25);            RMATCH(eptr, ecode, offset_top, md, eptrb, RM25);
3545            eptr--;            eptr--;
3546            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3547            }            }
3548          RRETURN(MATCH_NOMATCH);          /* Control never gets here */
3549          }          }
       /* Control never gets here */  
3550        }        }
3551    
3552      /* Caseful comparisons (includes all multi-byte characters) */      /* Caseful comparisons (includes all multi-byte characters) */
# Line 3353  for (;;) Line 3560  for (;;)
3560            SCHECK_PARTIAL();            SCHECK_PARTIAL();
3561            RRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
3562            }            }
3563          if (fc != *eptr++) RRETURN(MATCH_NOMATCH);          if (fc != RAWUCHARINCTEST(eptr)) RRETURN(MATCH_NOMATCH);
3564          }          }
3565    
3566        if (min == max) continue;        if (min == max) continue;
# Line 3370  for (;;) Line 3577  for (;;)
3577              SCHECK_PARTIAL();              SCHECK_PARTIAL();
3578              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
3579              }              }
3580            if (fc != *eptr++) RRETURN(MATCH_NOMATCH);            if (fc != RAWUCHARINCTEST(eptr)) RRETURN(MATCH_NOMATCH);
3581            }            }
3582          /* Control never gets here */          /* Control never gets here */
3583          }          }
# Line 3384  for (;;) Line 3591  for (;;)
3591              SCHECK_PARTIAL();              SCHECK_PARTIAL();
3592              break;              break;
3593              }              }
3594            if (fc != *eptr) break;            if (fc != RAWUCHARTEST(eptr)) break;
3595            eptr++;            eptr++;
3596            }            }
3597          if (possessive) continue;          if (possessive) continue;    /* No backtracking */
3598            for (;;)
         while (eptr >= pp)  
3599            {            {
3600              if (eptr == pp) goto TAIL_RECURSE;
3601            RMATCH(eptr, ecode, offset_top, md, eptrb, RM27);            RMATCH(eptr, ecode, offset_top, md, eptrb, RM27);
3602            eptr--;            eptr--;
3603            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3604            }            }
3605          RRETURN(MATCH_NOMATCH);          /* Control never gets here */
3606          }          }
3607        }        }
3608      /* Control never gets here */      /* Control never gets here */
# Line 3410  for (;;) Line 3617  for (;;)
3617        SCHECK_PARTIAL();        SCHECK_PARTIAL();
3618        RRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
3619        }        }
3620      ecode++;  #ifdef SUPPORT_UTF
3621      GETCHARINCTEST(c, eptr);      if (utf)
     if (op == OP_NOTI)         /* The caseless case */  
3622        {        {
3623  #ifdef SUPPORT_UTF8        register pcre_uint32 ch, och;
3624        if (c < 256)  
3625  #endif        ecode++;
3626        c = md->lcc[c];        GETCHARINC(ch, ecode);
3627        if (md->lcc[*ecode++] == c) RRETURN(MATCH_NOMATCH);        GETCHARINC(c, eptr);
3628    
3629          if (op == OP_NOT)
3630            {
3631            if (ch == c) RRETURN(MATCH_NOMATCH);
3632            }
3633          else
3634            {
3635    #ifdef SUPPORT_UCP
3636            if (ch > 127)
3637              och = UCD_OTHERCASE(ch);
3638    #else
3639            if (ch > 127)
3640              och = ch;
3641    #endif /* SUPPORT_UCP */
3642            else
3643              och = TABLE_GET(ch, md->fcc, ch);
3644            if (ch == c || och == c) RRETURN(MATCH_NOMATCH);
3645            }
3646        }        }
3647      else    /* Caseful */      else
3648    #endif
3649        {        {
3650        if (*ecode++ == c) RRETURN(MATCH_NOMATCH);        register pcre_uint32 ch = ecode[1];
3651          c = *eptr++;
3652          if (ch == c || (op == OP_NOTI && TABLE_GET(ch, md->fcc, ch) == c))
3653            RRETURN(MATCH_NOMATCH);
3654          ecode += 2;
3655        }        }
3656      break;      break;
3657    
# Line 3436  for (;;) Line 3665  for (;;)
3665      case OP_NOTEXACT:      case OP_NOTEXACT:
3666      case OP_NOTEXACTI:      case OP_NOTEXACTI:
3667      min = max = GET2(ecode, 1);      min = max = GET2(ecode, 1);
3668      ecode += 3;      ecode += 1 + IMM2_SIZE;
3669      goto REPEATNOTCHAR;      goto REPEATNOTCHAR;
3670    
3671      case OP_NOTUPTO:      case OP_NOTUPTO:
# Line 3446  for (;;) Line 3675  for (;;)
3675      min = 0;      min = 0;
3676      max = GET2(ecode, 1);      max = GET2(ecode, 1);
3677      minimize = *ecode == OP_NOTMINUPTO || *ecode == OP_NOTMINUPTOI;      minimize = *ecode == OP_NOTMINUPTO || *ecode == OP_NOTMINUPTOI;
3678      ecode += 3;      ecode += 1 + IMM2_SIZE;
3679      goto REPEATNOTCHAR;      goto REPEATNOTCHAR;
3680    
3681      case OP_NOTPOSSTAR:      case OP_NOTPOSSTAR:
# Line 3478  for (;;) Line 3707  for (;;)
3707      possessive = TRUE;      possessive = TRUE;
3708      min = 0;      min = 0;
3709      max = GET2(ecode, 1);      max = GET2(ecode, 1);
3710      ecode += 3;      ecode += 1 + IMM2_SIZE;
3711      goto REPEATNOTCHAR;      goto REPEATNOTCHAR;
3712    
3713      case OP_NOTSTAR:      case OP_NOTSTAR:
# Line 3502  for (;;) Line 3731  for (;;)
3731      /* Common code for all repeated single-byte matches. */      /* Common code for all repeated single-byte matches. */
3732    
3733      REPEATNOTCHAR:      REPEATNOTCHAR:
3734      fc = *ecode++;      GETCHARINCTEST(fc, ecode);
3735    
3736      /* The code is duplicated for the caseless and caseful cases, for speed,      /* The code is duplicated for the caseless and caseful cases, for speed,
3737      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 3742  for (;;)
3742      characters and work backwards. */      characters and work backwards. */
3743    
3744      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,
3745        max, eptr));        max, (char *)eptr));
3746    
3747      if (op >= OP_NOTSTARI)     /* Caseless */      if (op >= OP_NOTSTARI)     /* Caseless */
3748        {        {
3749        fc = md->lcc[fc];  #ifdef SUPPORT_UTF
3750    #ifdef SUPPORT_UCP
3751          if (utf && fc > 127)
3752            foc = UCD_OTHERCASE(fc);
3753    #else
3754          if (utf && fc > 127)
3755            foc = fc;
3756    #endif /* SUPPORT_UCP */
3757          else
3758    #endif /* SUPPORT_UTF */
3759            foc = TABLE_GET(fc, md->fcc, fc);
3760    
3761  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
3762        /* UTF-8 mode */        if (utf)
       if (utf8)  
3763          {          {
3764          register unsigned int d;          register pcre_uint32 d;
3765          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
3766            {            {
3767            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
# Line 3532  for (;;) Line 3770  for (;;)
3770              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
3771              }              }
3772            GETCHARINC(d, eptr);            GETCHARINC(d, eptr);
3773            if (d < 256) d = md->lcc[d];            if (fc == d || (unsigned int)foc == d) RRETURN(MATCH_NOMATCH);
           if (fc == d) RRETURN(MATCH_NOMATCH);  
3774            }            }
3775          }          }
3776        else        else
3777  #endif  #endif  /* SUPPORT_UTF */
3778          /* Not UTF mode */
       /* Not UTF-8 mode */  
3779          {          {
3780          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
3781            {            {
# Line 3548  for (;;) Line 3784  for (;;)
3784              SCHECK_PARTIAL();              SCHECK_PARTIAL();
3785              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
3786              }              }
3787            if (fc == md->lcc[*eptr++]) RRETURN(MATCH_NOMATCH);            if (fc == *eptr || foc == *eptr) RRETURN(MATCH_NOMATCH);
3788              eptr++;
3789            }            }
3790          }          }
3791    
# Line 3556  for (;;) Line 3793  for (;;)
3793    
3794        if (minimize)        if (minimize)
3795          {          {
3796  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
3797          /* UTF-8 mode */          if (utf)
         if (utf8)  
3798            {            {
3799            register unsigned int d;            register pcre_uint32 d;
3800            for (fi = min;; fi++)            for (fi = min;; fi++)
3801              {              {
3802              RMATCH(eptr, ecode, offset_top, md, eptrb, RM28);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM28);
# Line 3572  for (;;) Line 3808  for (;;)
3808                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
3809                }                }
3810              GETCHARINC(d, eptr);              GETCHARINC(d, eptr);
3811              if (d < 256) d = md->lcc[d];              if (fc == d || (unsigned int)foc == d) RRETURN(MATCH_NOMATCH);
             if (fc == d) RRETURN(MATCH_NOMATCH);  
3812              }              }
3813            }            }
3814          else          else
3815  #endif  #endif  /*SUPPORT_UTF */
3816          /* Not UTF-8 mode */          /* Not UTF mode */
3817            {            {
3818            for (fi = min;; fi++)            for (fi = min;; fi++)
3819              {              {
# Line 3590  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          /* Control never gets here */          /* Control never gets here */
# Line 3602  for (;;) Line 3838  for (;;)
3838          {          {
3839          pp = eptr;          pp = eptr;
3840    
3841  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
3842          /* UTF-8 mode */          if (utf)
         if (utf8)  
3843            {            {
3844            register unsigned int d;            register pcre_uint32 d;
3845            for (i = min; i < max; i++)            for (i = min; i < max; i++)
3846              {              {
3847              int len = 1;              int len = 1;
# Line 3616  for (;;) Line 3851  for (;;)
3851                break;                break;
3852                }                }
3853              GETCHARLEN(d, eptr, len);              GETCHARLEN(d, eptr, len);
3854              if (d < 256) d = md->lcc[d];              if (fc == d || (unsigned int)foc == d) break;
             if (fc == d) break;  
3855              eptr += len;              eptr += len;
3856              }              }
3857          if (possessive) continue;            if (possessive) continue;    /* No backtracking */
3858          for(;;)            for(;;)
3859              {              {
3860                if (eptr == pp) goto TAIL_RECURSE;
3861              RMATCH(eptr, ecode, offset_top, md, eptrb, RM30);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM30);
3862              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3863              if (eptr-- == pp) break;        /* Stop if tried at original pos */              eptr--;
3864              BACKCHAR(eptr);              BACKCHAR(eptr);
3865              }              }
3866            }            }
3867          else          else
3868  #endif  #endif  /* SUPPORT_UTF */
3869          /* Not UTF-8 mode */          /* Not UTF mode */
3870            {            {
3871            for (i = min; i < max; i++)            for (i = min; i < max; i++)
3872              {              {
# Line 3640  for (;;) Line 3875  for (;;)
3875                SCHECK_PARTIAL();                SCHECK_PARTIAL();
3876                break;                break;
3877                }                }
3878              if (fc == md->lcc[*eptr]) break;              if (fc == *eptr || foc == *eptr) break;
3879              eptr++;              eptr++;
3880              }              }
3881            if (possessive) continue;            if (possessive) continue;    /* No backtracking */
3882            while (eptr >= pp)            for (;;)
3883              {              {
3884                if (eptr == pp) goto TAIL_RECURSE;
3885              RMATCH(eptr, ecode, offset_top, md, eptrb, RM31);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM31);
3886              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3887              eptr--;              eptr--;
3888              }              }
3889            }            }
3890            /* Control never gets here */
         RRETURN(MATCH_NOMATCH);  
3891          }          }
       /* Control never gets here */  
3892        }        }
3893    
3894      /* Caseful comparisons */      /* Caseful comparisons */
3895    
3896      else      else
3897        {        {
3898  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
3899        /* UTF-8 mode */        if (utf)
       if (utf8)  
3900          {          {
3901          register unsigned int d;          register pcre_uint32 d;
3902          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
3903            {            {
3904            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
# Line 3679  for (;;) Line 3912  for (;;)
3912          }          }
3913        else        else
3914  #endif  #endif
3915        /* Not UTF-8 mode */        /* Not UTF mode */
3916          {          {
3917          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
3918            {            {
# Line 3696  for (;;) Line 3929  for (;;)
3929    
3930        if (minimize)        if (minimize)
3931          {          {
3932  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
3933          /* UTF-8 mode */          if (utf)
         if (utf8)  
3934            {            {
3935            register unsigned int d;            register pcre_uint32 d;
3936            for (fi = min;; fi++)            for (fi = min;; fi++)
3937              {              {
3938              RMATCH(eptr, ecode, offset_top, md, eptrb, RM32);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM32);
# Line 3717  for (;;) Line 3949  for (;;)
3949            }            }
3950          else          else
3951  #endif  #endif
3952          /* Not UTF-8 mode */          /* Not UTF mode */
3953            {            {
3954            for (fi = min;; fi++)            for (fi = min;; fi++)
3955              {              {
# Line 3741  for (;;) Line 3973  for (;;)
3973          {          {
3974          pp = eptr;          pp = eptr;
3975    
3976  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
3977          /* UTF-8 mode */          if (utf)
         if (utf8)  
3978            {            {
3979            register unsigned int d;            register pcre_uint32 d;
3980            for (i = min; i < max; i++)            for (i = min; i < max; i++)
3981              {              {
3982              int len = 1;              int len = 1;
# Line 3758  for (;;) Line 3989  for (;;)
3989              if (fc == d) break;              if (fc == d) break;
3990              eptr += len;              eptr += len;
3991              }              }
3992            if (possessive) continue;            if (possessive) continue;    /* No backtracking */
3993            for(;;)            for(;;)
3994              {              {
3995                if (eptr == pp) goto TAIL_RECURSE;
3996              RMATCH(eptr, ecode, offset_top, md, eptrb, RM34);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM34);
3997              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3998              if (eptr-- == pp) break;        /* Stop if tried at original pos */              eptr--;
3999              BACKCHAR(eptr);              BACKCHAR(eptr);
4000              }              }
4001            }            }
4002          else          else
4003  #endif  #endif
4004          /* Not UTF-8 mode */          /* Not UTF mode */
4005            {            {
4006            for (i = min; i < max; i++)            for (i = min; i < max; i++)
4007              {              {
# Line 3781  for (;;) Line 4013  for (;;)
4013              if (fc == *eptr) break;              if (fc == *eptr) break;
4014              eptr++;              eptr++;
4015              }              }
4016            if (possessive) continue;            if (possessive) continue;    /* No backtracking */
4017            while (eptr >= pp)            for (;;)
4018              {              {
4019                if (eptr == pp) goto TAIL_RECURSE;
4020              RMATCH(eptr, ecode, offset_top, md, eptrb, RM35);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM35);
4021              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
4022              eptr--;              eptr--;
4023              }              }
4024            }            }
4025            /* Control never gets here */
         RRETURN(MATCH_NOMATCH);  
4026          }          }
4027        }        }
4028      /* Control never gets here */      /* Control never gets here */
# Line 3802  for (;;) Line 4034  for (;;)
4034      case OP_TYPEEXACT:      case OP_TYPEEXACT:
4035      min = max = GET2(ecode, 1);      min = max = GET2(ecode, 1);
4036      minimize = TRUE;      minimize = TRUE;
4037      ecode += 3;      ecode += 1 + IMM2_SIZE;
4038      goto REPEATTYPE;      goto REPEATTYPE;
4039    
4040      case OP_TYPEUPTO:      case OP_TYPEUPTO:
# Line 3810  for (;;) Line 4042  for (;;)
4042      min = 0;      min = 0;
4043      max = GET2(ecode, 1);      max = GET2(ecode, 1);
4044      minimize = *ecode == OP_TYPEMINUPTO;      minimize = *ecode == OP_TYPEMINUPTO;
4045      ecode += 3;      ecode += 1 + IMM2_SIZE;
4046      goto REPEATTYPE;      goto REPEATTYPE;
4047    
4048      case OP_TYPEPOSSTAR:      case OP_TYPEPOSSTAR:
# Line 3838  for (;;) Line 4070  for (;;)
4070      possessive = TRUE;      possessive = TRUE;
4071      min = 0;      min = 0;
4072      max = GET2(ecode, 1);      max = GET2(ecode, 1);
4073      ecode += 3;      ecode += 1 + IMM2_SIZE;
4074      goto REPEATTYPE;      goto REPEATTYPE;
4075    
4076      case OP_TYPESTAR:      case OP_TYPESTAR:
# Line 3972  for (;;) Line 4204  for (;;)
4204              }              }
4205            break;            break;
4206    
4207              /* Perl space used to exclude VT, but from Perl 5.18 it is included,
4208              which means that Perl space and POSIX space are now identical. PCRE
4209              was changed at release 8.34. */
4210    
4211            case PT_SPACE:    /* Perl space */            case PT_SPACE:    /* Perl space */
4212              case PT_PXSPACE:  /* POSIX space */
4213            for (i = 1; i <= min; i++)            for (i = 1; i <= min; i++)
4214              {              {
4215              if (eptr >= md->end_subject)              if (eptr >= md->end_subject)
# Line 3981  for (;;) Line 4218  for (;;)
4218                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
4219                }                }
4220              GETCHARINCTEST(c, eptr);              GETCHARINCTEST(c, eptr);
4221              if ((UCD_CATEGORY(c) == ucp_Z || c == CHAR_HT || c == CHAR_NL ||              switch(c)
4222                   c == CHAR_FF || c == CHAR_CR)                {
4223                     == prop_fail_result)                HSPACE_CASES:
4224                RRETURN(MATCH_NOMATCH);                VSPACE_CASES:
4225                  if (prop_fail_result) RRETURN(MATCH_NOMATCH);
4226                  break;
4227    
4228                  default:
4229                  if ((UCD_CATEGORY(c) == ucp_Z) == prop_fail_result)
4230                    RRETURN(MATCH_NOMATCH);
4231                  break;
4232                  }
4233              }              }
4234            break;            break;
4235    
4236            case PT_PXSPACE:  /* POSIX space */            case PT_WORD:
4237            for (i = 1; i <= min; i++)            for (i = 1; i <= min; i++)
4238              {              {
4239                int category;
4240              if (eptr >= md->end_subject)              if (eptr >= md->end_subject)
4241                {                {
4242                SCHECK_PARTIAL();                SCHECK_PARTIAL();
4243                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
4244                }                }
4245              GETCHARINCTEST(c, eptr);              GETCHARINCTEST(c, eptr);
4246              if ((UCD_CATEGORY(c) == ucp_Z || c == CHAR_HT || c == CHAR_NL ||              category = UCD_CATEGORY(c);
4247                   c == CHAR_VT || c == CHAR_FF || c == CHAR_CR)              if ((category == ucp_L || category == ucp_N || c == CHAR_UNDERSCORE)
4248                     == prop_fail_result)                     == prop_fail_result)
4249                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
4250              }              }
4251            break;            break;
4252    
4253            case PT_WORD:            case PT_CLIST:
4254            for (i = 1; i <= min; i++)            for (i = 1; i <= min; i++)
4255              {              {
4256              int category;              const pcre_uint32 *cp;
4257              if (eptr >= md->end_subject)              if (eptr >= md->end_subject)
4258                {                {
4259                SCHECK_PARTIAL();                SCHECK_PARTIAL();
4260                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
4261                }                }
4262              GETCHARINCTEST(c, eptr);              GETCHARINCTEST(c, eptr);
4263              category = UCD_CATEGORY(c);              cp = PRIV(ucd_caseless_sets) + prop_value;
4264              if ((category == ucp_L || category == ucp_N || c == CHAR_UNDERSCORE)              for (;;)
4265                     == prop_fail_result)                {
4266                  if (c < *cp)
4267                    { if (prop_fail_result) break; else { RRETURN(MATCH_NOMATCH); } }
4268                  if (c == *cp++)
4269                    { if (prop_fail_result) { RRETURN(MATCH_NOMATCH); } else break; }
4270                  }
4271                }
4272              break;
4273    
4274              case PT_UCNC:
4275              for (i = 1; i <= min; i++)
4276                {
4277                if (eptr >= md->end_subject)
4278                  {
4279                  SCHECK_PARTIAL();
4280                  RRETURN(MATCH_NOMATCH);
4281                  }
4282                GETCHARINCTEST(c, eptr);
4283                if ((c == CHAR_DOLLAR_SIGN || c == CHAR_COMMERCIAL_AT ||
4284                     c == CHAR_GRAVE_ACCENT || (c >= 0xa0 && c <= 0xd7ff) ||
4285                     c >= 0xe000) == prop_fail_result)
4286                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
4287              }              }
4288            break;            break;
# Line 4040  for (;;) Line 4306  for (;;)
4306              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4307              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4308              }              }
4309            GETCHARINCTEST(c, eptr);            else
           if (UCD_CATEGORY(c) == ucp_M) RRETURN(MATCH_NOMATCH);  
           while (eptr < md->end_subject)  
4310              {              {
4311              int len = 1;              int lgb, rgb;
4312              if (!utf8) c = *eptr; else { GETCHARLEN(c, eptr, len); }              GETCHARINCTEST(c, eptr);
4313              if (UCD_CATEGORY(c) != ucp_M) break;              lgb = UCD_GRAPHBREAK(c);
4314              eptr += len;             while (eptr < md->end_subject)
4315                  {
4316                  int len = 1;
4317                  if (!utf) c = *eptr; else { GETCHARLEN(c, eptr, len); }
4318                  rgb = UCD_GRAPHBREAK(c);
4319                  if ((PRIV(ucp_gbtable)[lgb] & (1 << rgb)) == 0) break;
4320                  lgb = rgb;
4321                  eptr += len;
4322                  }
4323              }              }
4324              CHECK_PARTIAL();
4325            }            }
4326          }          }
4327    
# Line 4057  for (;;) Line 4330  for (;;)
4330    
4331  /* Handle all other cases when the coding is UTF-8 */  /* Handle all other cases when the coding is UTF-8 */
4332    
4333  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
4334        if (utf8) switch(ctype)        if (utf) switch(ctype)
4335          {          {
4336          case OP_ANY:          case OP_ANY:
4337          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
# Line 4069  for (;;) Line 4342  for (;;)
4342              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4343              }              }
4344            if (IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH);            if (IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH);
4345              if (md->partial != 0 &&
4346                  eptr + 1 >= md->end_subject &&
4347                  NLBLOCK->nltype == NLTYPE_FIXED &&
4348                  NLBLOCK->nllen == 2 &&
4349                  RAWUCHAR(eptr) == NLBLOCK->nl[0])
4350                {
4351                md->hitend = TRUE;
4352                if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
4353                }
4354            eptr++;            eptr++;
4355            while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;            ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);
4356            }            }
4357          break;          break;
4358    
# Line 4083  for (;;) Line 4365  for (;;)
4365              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4366              }              }
4367            eptr++;            eptr++;
4368            while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;            ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);
4369            }            }
4370          break;          break;
4371    
# Line 4105  for (;;) Line 4387  for (;;)
4387              {              {
4388              default: RRETURN(MATCH_NOMATCH);              default: RRETURN(MATCH_NOMATCH);
4389    
4390              case 0x000d:              case CHAR_CR:
4391              if (eptr < md->end_subject && *eptr == 0x0a) eptr++;              if (eptr < md->end_subject && RAWUCHAR(eptr) == CHAR_LF) eptr++;
4392              break;              break;
4393    
4394              case 0x000a:              case CHAR_LF:
4395              break;              break;
4396    
4397              case 0x000b:              case CHAR_VT:
4398              case 0x000c:              case CHAR_FF:
4399              case 0x0085:              case CHAR_NEL:
4400    #ifndef EBCDIC
4401              case 0x2028:              case 0x2028:
4402              case 0x2029:              case 0x2029:
4403    #endif  /* Not EBCDIC */
4404              if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);              if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);
4405              break;              break;
4406              }              }
# Line 4134  for (;;) Line 4418  for (;;)
4418            GETCHARINC(c, eptr);            GETCHARINC(c, eptr);
4419            switch(c)            switch(c)
4420              {              {
4421                HSPACE_CASES: RRETURN(MATCH_NOMATCH);  /* Byte and multibyte cases */
4422              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);  
4423              }              }
4424            }            }
4425          break;          break;
# Line 4170  for (;;) Line 4435  for (;;)
4435            GETCHARINC(c, eptr);            GETCHARINC(c, eptr);
4436            switch(c)            switch(c)
4437              {              {
4438                HSPACE_CASES: break;  /* Byte and multibyte cases */
4439              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;  
4440              }              }
4441            }            }
4442          break;          break;
# Line 4206  for (;;) Line 4452  for (;;)
4452            GETCHARINC(c, eptr);            GETCHARINC(c, eptr);
4453            switch(c)            switch(c)
4454              {              {
4455                VSPACE_CASES: RRETURN(MATCH_NOMATCH);
4456              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);  
4457              }              }
4458            }            }
4459          break;          break;
# Line 4230  for (;;) Line 4469  for (;;)
4469            GETCHARINC(c, eptr);            GETCHARINC(c, eptr);
4470            switch(c)            switch(c)
4471              {              {
4472                VSPACE_CASES: break;
4473              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;  
4474              }              }
4475            }            }
4476          break;          break;
# Line 4260  for (;;) Line 4492  for (;;)
4492          case OP_DIGIT:          case OP_DIGIT:
4493          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
4494            {            {
4495              pcre_uint32 cc;
4496            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
4497              {              {
4498              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4499              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4500              }              }
4501            if (*eptr >= 128 || (md->ctypes[*eptr++] & ctype_digit) == 0)            cc = RAWUCHAR(eptr);
4502              if (cc >= 128 || (md->ctypes[cc] & ctype_digit) == 0)
4503              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4504              eptr++;
4505            /* 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 */
4506            }            }
4507          break;          break;
# Line 4274  for (;;) Line 4509  for (;;)
4509          case OP_NOT_WHITESPACE:          case OP_NOT_WHITESPACE:
4510          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
4511            {            {
4512              pcre_uint32 cc;
4513            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
4514              {              {
4515              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4516              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4517              }              }
4518            if (*eptr < 128 && (md->ctypes[*eptr] & ctype_space) != 0)            cc = RAWUCHAR(eptr);
4519              if (cc < 128 && (md->ctypes[cc] & ctype_space) != 0)
4520              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4521            while (++eptr < md->end_subject && (*eptr & 0xc0) == 0x80);            eptr++;
4522              ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);
4523            }            }
4524          break;          break;
4525    
4526          case OP_WHITESPACE:          case OP_WHITESPACE:
4527          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
4528            {            {
4529              pcre_uint32 cc;
4530            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
4531              {              {
4532              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4533              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4534              }              }
4535            if (*eptr >= 128 || (md->ctypes[*eptr++] & ctype_space) == 0)            cc = RAWUCHAR(eptr);
4536              if (cc >= 128 || (md->ctypes[cc] & ctype_space) == 0)
4537              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4538              eptr++;
4539            /* 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 */
4540            }            }
4541          break;          break;
# Line 4302  for (;;) Line 4543  for (;;)
4543          case OP_NOT_WORDCHAR:          case OP_NOT_WORDCHAR:
4544          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
4545            {            {
4546              pcre_uint32 cc;
4547            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
4548              {              {
4549              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4550              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4551              }              }
4552            if (*eptr < 128 && (md->ctypes[*eptr] & ctype_word) != 0)            cc = RAWUCHAR(eptr);
4553              if (cc < 128 && (md->ctypes[cc] & ctype_word) != 0)
4554              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4555            while (++eptr < md->end_subject && (*eptr & 0xc0) == 0x80);            eptr++;
4556              ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);
4557            }            }
4558          break;          break;
4559    
4560          case OP_WORDCHAR:          case OP_WORDCHAR:
4561          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
4562            {            {
4563              pcre_uint32 cc;
4564            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
4565              {              {
4566              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4567              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4568              }              }
4569            if (*eptr >= 128 || (md->ctypes[*eptr++] & ctype_word) == 0)            cc = RAWUCHAR(eptr);
4570              if (cc >= 128 || (md->ctypes[cc] & ctype_word) == 0)
4571              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4572              eptr++;
4573            /* 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 */
4574            }            }
4575          break;          break;
# Line 4332  for (;;) Line 4579  for (;;)
4579          }  /* End switch(ctype) */          }  /* End switch(ctype) */
4580    
4581        else        else
4582  #endif     /* SUPPORT_UTF8 */  #endif     /* SUPPORT_UTF */
4583    
4584        /* 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
4585        than OP_PROP and OP_NOTPROP. */        than OP_PROP and OP_NOTPROP. */
# Line 4348  for (;;) Line 4595  for (;;)
4595              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4596              }              }
4597            if (IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH);            if (IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH);
4598              if (md->partial != 0 &&
4599                  eptr + 1 >= md->end_subject &&
4600                  NLBLOCK->nltype == NLTYPE_FIXED &&
4601                  NLBLOCK->nllen == 2 &&
4602                  *eptr == NLBLOCK->nl[0])
4603                {
4604                md->hitend = TRUE;
4605                if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
4606                }
4607            eptr++;            eptr++;
4608            }            }
4609          break;          break;
# Line 4382  for (;;) Line 4638  for (;;)
4638              {              {
4639              default: RRETURN(MATCH_NOMATCH);              default: RRETURN(MATCH_NOMATCH);
4640    
4641              case 0x000d:              case CHAR_CR:
4642              if (eptr < md->end_subject && *eptr == 0x0a) eptr++;              if (eptr < md->end_subject && *eptr == CHAR_LF) eptr++;
4643              break;              break;
4644    
4645              case 0x000a:              case CHAR_LF:
4646              break;              break;
4647    
4648              case 0x000b:              case CHAR_VT:
4649              case 0x000c:              case CHAR_FF:
4650              case 0x0085:              case CHAR_NEL:
4651    #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
4652                case 0x2028:
4653                case 0x2029:
4654    #endif
4655              if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);              if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);
4656              break;              break;
4657              }              }
# Line 4409  for (;;) Line 4669  for (;;)
4669            switch(*eptr++)            switch(*eptr++)
4670              {              {
4671              default: break;              default: break;
4672              case 0x09:      /* HT */              HSPACE_BYTE_CASES:
4673              case 0x20:      /* SPACE */  #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
4674              case 0xa0:      /* NBSP */              HSPACE_MULTIBYTE_CASES:
4675    #endif
4676              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4677              }              }
4678            }            }
# Line 4428  for (;;) Line 4689  for (;;)
4689            switch(*eptr++)            switch(*eptr++)
4690              {              {
4691              default: RRETURN(MATCH_NOMATCH);              default: RRETURN(MATCH_NOMATCH);
4692              case 0x09:      /* HT */              HSPACE_BYTE_CASES:
4693              case 0x20:      /* SPACE */  #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
4694              case 0xa0:      /* NBSP */              HSPACE_MULTIBYTE_CASES:
4695    #endif
4696              break;              break;
4697              }              }
4698            }            }
# Line 4446  for (;;) Line 4708  for (;;)
4708              }              }
4709            switch(*eptr++)            switch(*eptr++)
4710              {              {
4711              default: break;              VSPACE_BYTE_CASES:
4712              case 0x0a:      /* LF */  #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
4713              case 0x0b:      /* VT */              VSPACE_MULTIBYTE_CASES:
4714              case 0x0c:      /* FF */  #endif
             case 0x0d:      /* CR */  
             case 0x85:      /* NEL */  
4715              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4716                default: break;
4717              }              }
4718            }            }
4719          break;          break;
# Line 4468  for (;;) Line 4729  for (;;)
4729            switch(*eptr++)            switch(*eptr++)
4730              {              {
4731              default: RRETURN(MATCH_NOMATCH);              default: RRETURN(MATCH_NOMATCH);
4732              case 0x0a:      /* LF */              VSPACE_BYTE_CASES:
4733              case 0x0b:      /* VT */  #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
4734              case 0x0c:      /* FF */              VSPACE_MULTIBYTE_CASES:
4735              case 0x0d:      /* CR */  #endif
             case 0x85:      /* NEL */  
4736              break;              break;
4737              }              }
4738            }            }
# Line 4486  for (;;) Line 4746  for (;;)
4746              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4747              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4748              }              }
4749            if ((md->ctypes[*eptr++] & ctype_digit) != 0) RRETURN(MATCH_NOMATCH);            if (MAX_255(*eptr) && (md->ctypes[*eptr] & ctype_digit) != 0)
4750                RRETURN(MATCH_NOMATCH);
4751              eptr++;
4752            }            }
4753          break;          break;
4754    
# Line 4498  for (;;) Line 4760  for (;;)
4760              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4761              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4762              }              }
4763            if ((md->ctypes[*eptr++] & ctype_digit) == 0) RRETURN(MATCH_NOMATCH);            if (!MAX_255(*eptr) || (md->ctypes[*eptr] & ctype_digit) == 0)
4764                RRETURN(MATCH_NOMATCH);
4765              eptr++;
4766            }            }
4767          break;          break;
4768    
# Line 4510  for (;;) Line 4774  for (;;)
4774              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4775              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4776              }              }
4777            if ((md->ctypes[*eptr++] & ctype_space) != 0) RRETURN(MATCH_NOMATCH);            if (MAX_255(*eptr) && (md->ctypes[*eptr] & ctype_space) != 0)
4778                RRETURN(MATCH_NOMATCH);
4779              eptr++;
4780            }            }
4781          break;          break;
4782    
# Line 4522  for (;;) Line 4788  for (;;)
4788              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4789              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4790              }              }
4791            if ((md->ctypes[*eptr++] & ctype_space) == 0) RRETURN(MATCH_NOMATCH);            if (!MAX_255(*eptr) || (md->ctypes[*eptr] & ctype_space) == 0)
4792                RRETURN(MATCH_NOMATCH);
4793              eptr++;
4794            }            }
4795          break;          break;
4796    
# Line 4534  for (;;) Line 4802  for (;;)
4802              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4803              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4804              }              }
4805            if ((md->ctypes[*eptr++] & ctype_word) != 0)            if (MAX_255(*eptr) && (md->ctypes[*eptr] & ctype_word) != 0)
4806              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4807              eptr++;
4808            }            }
4809          break;          break;
4810    
# Line 4547  for (;;) Line 4816  for (;;)
4816              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4817              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4818              }              }
4819            if ((md->ctypes[*eptr++] & ctype_word) == 0)            if (!MAX_255(*eptr) || (md->ctypes[*eptr] & ctype_word) == 0)
4820              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4821              eptr++;
4822            }            }
4823          break;          break;
4824    
# Line 4679  for (;;) Line 4949  for (;;)
4949              }              }
4950            /* Control never gets here */            /* Control never gets here */
4951    
4952              /* Perl space used to exclude VT, but from Perl 5.18 it is included,
4953              which means that Perl space and POSIX space are now identical. PCRE
4954              was changed at release 8.34. */
4955    
4956            case PT_SPACE:    /* Perl space */            case PT_SPACE:    /* Perl space */
4957              case PT_PXSPACE:  /* POSIX space */
4958            for (fi = min;; fi++)            for (fi = min;; fi++)
4959              {              {
4960              RMATCH(eptr, ecode, offset_top, md, eptrb, RM60);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM61);
4961              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
4962              if (fi >= max) RRETURN(MATCH_NOMATCH);              if (fi >= max) RRETURN(MATCH_NOMATCH);
4963              if (eptr >= md->end_subject)              if (eptr >= md->end_subject)
# Line 4691  for (;;) Line 4966  for (;;)
4966                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
4967                }                }
4968              GETCHARINCTEST(c, eptr);              GETCHARINCTEST(c, eptr);
4969              if ((UCD_CATEGORY(c) == ucp_Z || c == CHAR_HT || c == CHAR_NL ||              switch(c)
4970                   c == CHAR_FF || c == CHAR_CR)                {
4971                     == prop_fail_result)                HSPACE_CASES:
4972                RRETURN(MATCH_NOMATCH);                VSPACE_CASES:
4973                  if (prop_fail_result) RRETURN(MATCH_NOMATCH);
4974                  break;
4975    
4976                  default:
4977                  if ((UCD_CATEGORY(c) == ucp_Z) == prop_fail_result)
4978                    RRETURN(MATCH_NOMATCH);
4979                  break;
4980                  }
4981              }              }
4982            /* Control never gets here */            /* Control never gets here */
4983    
4984            case PT_PXSPACE:  /* POSIX space */            case PT_WORD:
4985            for (fi = min;; fi++)            for (fi = min;; fi++)
4986              {              {
4987              RMATCH(eptr, ecode, offset_top, md, eptrb, RM61);              int category;
4988                RMATCH(eptr, ecode, offset_top, md, eptrb, RM62);
4989              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
4990              if (fi >= max) RRETURN(MATCH_NOMATCH);              if (fi >= max) RRETURN(MATCH_NOMATCH);
4991              if (eptr >= md->end_subject)              if (eptr >= md->end_subject)
# Line 4710  for (;;) Line 4994  for (;;)
4994                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
4995                }                }
4996              GETCHARINCTEST(c, eptr);              GETCHARINCTEST(c, eptr);
4997              if ((UCD_CATEGORY(c) == ucp_Z || c == CHAR_HT || c == CHAR_NL ||              category = UCD_CATEGORY(c);
4998                   c == CHAR_VT || c == CHAR_FF || c == CHAR_CR)              if ((category == ucp_L ||
4999                     category == ucp_N ||
5000                     c == CHAR_UNDERSCORE)
5001                     == prop_fail_result)                     == prop_fail_result)
5002                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
5003              }              }
5004            /* Control never gets here */            /* Control never gets here */
5005    
5006            case PT_WORD:            case PT_CLIST:
5007            for (fi = min;; fi++)            for (fi = min;; fi++)
5008              {              {
5009              int category;              const pcre_uint32 *cp;
5010              RMATCH(eptr, ecode, offset_top, md, eptrb, RM62);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM67);
5011              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
5012              if (fi >= max) RRETURN(MATCH_NOMATCH);              if (fi >= max) RRETURN(MATCH_NOMATCH);
5013              if (eptr >= md->end_subject)              if (eptr >= md->end_subject)
# Line 4730  for (;;) Line 5016  for (;;)
5016                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
5017                }                }
5018              GETCHARINCTEST(c, eptr);              GETCHARINCTEST(c, eptr);
5019              category = UCD_CATEGORY(c);              cp = PRIV(ucd_caseless_sets) + prop_value;
5020              if ((category == ucp_L ||              for (;;)
5021                   category == ucp_N ||                {
5022                   c == CHAR_UNDERSCORE)                if (c < *cp)
5023                     == prop_fail_result)                  { if (prop_fail_result) break; else { RRETURN(MATCH_NOMATCH); } }
5024                  if (c == *cp++)
5025                    { if (prop_fail_result) { RRETURN(MATCH_NOMATCH); } else break; }
5026                  }
5027                }
5028              /* Control never gets here */
5029    
5030              case PT_UCNC:
5031              for (fi = min;; fi++)
5032                {
5033                RMATCH(eptr, ecode, offset_top, md, eptrb, RM68);
5034                if (rrc != MATCH_NOMATCH) RRETURN(rrc);
5035                if (fi >= max) RRETURN(MATCH_NOMATCH);
5036                if (eptr >= md->end_subject)
5037                  {
5038                  SCHECK_PARTIAL();
5039                  RRETURN(MATCH_NOMATCH);
5040                  }
5041                GETCHARINCTEST(c, eptr);
5042                if ((c == CHAR_DOLLAR_SIGN || c == CHAR_COMMERCIAL_AT ||
5043                     c == CHAR_GRAVE_ACCENT || (c >= 0xa0 && c <= 0xd7ff) ||
5044                     c >= 0xe000) == prop_fail_result)
5045                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
5046              }              }
5047            /* Control never gets here */            /* Control never gets here */
5048    
5049            /* This should never occur */            /* This should never occur */
   
5050            default:            default:
5051            RRETURN(PCRE_ERROR_INTERNAL);            RRETURN(PCRE_ERROR_INTERNAL);
5052            }            }
# Line 4761  for (;;) Line 5067  for (;;)
5067              SCHECK_PARTIAL();              SCHECK_PARTIAL();
5068              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
5069              }              }
5070            GETCHARINCTEST(c, eptr);            else
           if (UCD_CATEGORY(c) == ucp_M) RRETURN(MATCH_NOMATCH);  
           while (eptr < md->end_subject)  
5071              {              {
5072              int len = 1;              int lgb, rgb;
5073              if (!utf8) c = *eptr; else { GETCHARLEN(c, eptr, len); }              GETCHARINCTEST(c, eptr);
5074              if (UCD_CATEGORY(c) != ucp_M) break;              lgb = UCD_GRAPHBREAK(c);
5075              eptr += len;              while (eptr < md->end_subject)
5076                  {
5077                  int len = 1;
5078                  if (!utf) c = *eptr; else { GETCHARLEN(c, eptr, len); }
5079                  rgb = UCD_GRAPHBREAK(c);
5080                  if ((PRIV(ucp_gbtable)[lgb] & (1 << rgb)) == 0) break;
5081                  lgb = rgb;
5082                  eptr += len;
5083                  }
5084              }              }
5085              CHECK_PARTIAL();
5086            }            }
5087          }          }
5088        else        else
5089  #endif     /* SUPPORT_UCP */  #endif     /* SUPPORT_UCP */
5090    
5091  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
5092        /* UTF-8 mode */        if (utf)
       if (utf8)  
5093          {          {
5094          for (fi = min;; fi++)          for (fi = min;; fi++)
5095            {            {
# Line 4794  for (;;) Line 5106  for (;;)
5106            GETCHARINC(c, eptr);            GETCHARINC(c, eptr);
5107            switch(ctype)            switch(ctype)
5108              {              {
5109              case OP_ANY:        /* This is the non-NL case */              case OP_ANY:               /* This is the non-NL case */
5110                if (md->partial != 0 &&    /* Take care with CRLF partial */
5111                    eptr >= md->end_subject &&
5112                    NLBLOCK->nltype == NLTYPE_FIXED &&
5113                    NLBLOCK->nllen == 2 &&
5114                    c == NLBLOCK->nl[0])
5115                  {
5116                  md->hitend = TRUE;
5117                  if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
5118                  }
5119                break;
5120    
5121              case OP_ALLANY:              case OP_ALLANY:
5122              case OP_ANYBYTE:              case OP_ANYBYTE:
5123              break;              break;
# Line 4803  for (;;) Line 5126  for (;;)
5126              switch(c)              switch(c)
5127                {                {
5128                default: RRETURN(MATCH_NOMATCH);                default: RRETURN(MATCH_NOMATCH);
5129                case 0x000d:                case CHAR_CR:
5130                if (eptr < md->end_subject && *eptr == 0x0a) eptr++;                if (eptr < md->end_subject && RAWUCHAR(eptr) == CHAR_LF) eptr++;
5131                break;                break;
5132                case 0x000a:  
5133                  case CHAR_LF:
5134                break;                break;
5135