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

Diff of /code/trunk/pcre_exec.c

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

revision 771 by ph10, Tue Nov 29 15:34:12 2011 UTC revision 1412 by ph10, Sun Dec 15 17:01:46 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 90  because the offset vector is always a mu Line 107  because the offset vector is always a mu
107    
108  /* Min and max values for the common repeats; for the maxima, 0 => infinity */  /* Min and max values for the common repeats; for the maxima, 0 => infinity */
109    
110  static const char rep_min[] = { 0, 0, 1, 1, 0, 0 };  static const char rep_min[] = { 0, 0, 1, 1, 0, 0, 0, 0, 0, 1, 0, };
111  static const char rep_max[] = { 0, 0, 0, 0, 1, 1 };  static const char rep_max[] = { 0, 0, 0, 0, 1, 1, 0, 0, 0, 0, 1, };
   
   
112    
113  #ifdef PCRE_DEBUG  #ifdef PCRE_DEBUG
114  /*************************************************  /*************************************************
# 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    #if defined SUPPORT_UTF && defined SUPPORT_UCP
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  #if defined SUPPORT_UTF && defined SUPPORT_UCP
199  #ifdef SUPPORT_UCP    if (utf)
   if (md->utf8)  
200      {      {
201      /* 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
202      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
203      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.
204      (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
205      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
206      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
207      reference, not along the subject (earlier code did this wrong). */      length along the reference, not along the subject (earlier code did this
208        wrong). */
209    
210      USPTR endptr = p + length;      PCRE_PUCHAR endptr = p + length;
211      while (p < endptr)      while (p < endptr)
212        {        {
213        int c, d;        pcre_uint32 c, d;
214        if (eptr >= md->end_subject) return -1;        const ucd_record *ur;
215          if (eptr >= md->end_subject) return -2;   /* Partial match */
216        GETCHARINC(c, eptr);        GETCHARINC(c, eptr);
217        GETCHARINC(d, p);        GETCHARINC(d, p);
218        if (c != d && c != UCD_OTHERCASE(d)) return -1;        ur = GET_UCD(d);
219          if (c != d && c != d + ur->other_case)
220            {
221            const pcre_uint32 *pp = PRIV(ucd_caseless_sets) + ur->caseset;
222            for (;;)
223              {
224              if (c < *pp) return -1;
225              if (c == *pp++) break;
226              }
227            }
228        }        }
229      }      }
230    else    else
231  #endif  #endif
 #endif  
232    
233    /* 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
234    is no UCP support. */    is no UCP support. */
235      {      {
     if (eptr + length > md->end_subject) return -1;  
236      while (length-- > 0)      while (length-- > 0)
237        { if (md->lcc[*p++] != md->lcc[*eptr++]) return -1; }        {
238          pcre_uint32 cc, cp;
239          if (eptr >= md->end_subject) return -2;   /* Partial match */
240          cc = RAWUCHARTEST(eptr);
241          cp = RAWUCHARTEST(p);
242          if (TABLE_GET(cp, md->lcc, cp) != TABLE_GET(cc, md->lcc, cc)) return -1;
243          p++;
244          eptr++;
245          }
246      }      }
247    }    }
248    
# Line 213  are in UTF-8 mode. */ Line 251  are in UTF-8 mode. */
251    
252  else  else
253    {    {
254    if (eptr + length > md->end_subject) return -1;    while (length-- > 0)
255    while (length-- > 0) if (*p++ != *eptr++) return -1;      {
256        if (eptr >= md->end_subject) return -2;   /* Partial match */
257        if (RAWUCHARINCTEST(p) != RAWUCHARINCTEST(eptr)) return -1;
258        }
259    }    }
260    
261  return eptr - eptr_start;  return (int)(eptr - eptr_start);
262  }  }
263    
264    
# Line 269  enum { RM1=1, RM2,  RM3,  RM4,  RM5,  RM Line 310  enum { RM1=1, RM2,  RM3,  RM4,  RM5,  RM
310         RM31,  RM32, RM33, RM34, RM35, RM36, RM37, RM38, RM39, RM40,         RM31,  RM32, RM33, RM34, RM35, RM36, RM37, RM38, RM39, RM40,
311         RM41,  RM42, RM43, RM44, RM45, RM46, RM47, RM48, RM49, RM50,         RM41,  RM42, RM43, RM44, RM45, RM46, RM47, RM48, RM49, RM50,
312         RM51,  RM52, RM53, RM54, RM55, RM56, RM57, RM58, RM59, RM60,         RM51,  RM52, RM53, RM54, RM55, RM56, RM57, RM58, RM59, RM60,
313         RM61,  RM62, RM63, RM64, RM65, RM66 };         RM61,  RM62, RM63, RM64, RM65, RM66, RM67 };
314    
315  /* 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
316  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 328  actually used in this definition. */
328    }    }
329  #define RRETURN(ra) \  #define RRETURN(ra) \
330    { \    { \
331    printf("match() returned %d from line %d ", ra, __LINE__); \    printf("match() returned %d from line %d\n", ra, __LINE__); \
332    return ra; \    return ra; \
333    }    }
334  #else  #else
# Line 307  argument of match(), which never changes Line 348  argument of match(), which never changes
348    
349  #define RMATCH(ra,rb,rc,rd,re,rw)\  #define RMATCH(ra,rb,rc,rd,re,rw)\
350    {\    {\
351    heapframe *newframe = (heapframe *)(pcre_stack_malloc)(sizeof(heapframe));\    heapframe *newframe = frame->Xnextframe;\
352    if (newframe == NULL) RRETURN(PCRE_ERROR_NOMEMORY);\    if (newframe == NULL)\
353    frame->Xwhere = rw; \      {\
354        newframe = (heapframe *)(PUBL(stack_malloc))(sizeof(heapframe));\
355        if (newframe == NULL) RRETURN(PCRE_ERROR_NOMEMORY);\
356        newframe->Xnextframe = NULL;\
357        frame->Xnextframe = newframe;\
358        }\
359      frame->Xwhere = rw;\
360    newframe->Xeptr = ra;\    newframe->Xeptr = ra;\
361    newframe->Xecode = rb;\    newframe->Xecode = rb;\
362    newframe->Xmstart = mstart;\    newframe->Xmstart = mstart;\
# Line 328  argument of match(), which never changes Line 375  argument of match(), which never changes
375    {\    {\
376    heapframe *oldframe = frame;\    heapframe *oldframe = frame;\
377    frame = oldframe->Xprevframe;\    frame = oldframe->Xprevframe;\
   (pcre_stack_free)(oldframe);\  
378    if (frame != NULL)\    if (frame != NULL)\
379      {\      {\
380      rrc = ra;\      rrc = ra;\
# Line 342  argument of match(), which never changes Line 388  argument of match(), which never changes
388    
389  typedef struct heapframe {  typedef struct heapframe {
390    struct heapframe *Xprevframe;    struct heapframe *Xprevframe;
391      struct heapframe *Xnextframe;
392    
393    /* Function arguments that may change */    /* Function arguments that may change */
394    
395    USPTR Xeptr;    PCRE_PUCHAR Xeptr;
396    const uschar *Xecode;    const pcre_uchar *Xecode;
397    USPTR Xmstart;    PCRE_PUCHAR Xmstart;
398    int Xoffset_top;    int Xoffset_top;
399    eptrblock *Xeptrb;    eptrblock *Xeptrb;
400    unsigned int Xrdepth;    unsigned int Xrdepth;
401    
402    /* Function local variables */    /* Function local variables */
403    
404    USPTR Xcallpat;    PCRE_PUCHAR Xcallpat;
405  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
406    USPTR Xcharptr;    PCRE_PUCHAR Xcharptr;
407  #endif  #endif
408    USPTR Xdata;    PCRE_PUCHAR Xdata;
409    USPTR Xnext;    PCRE_PUCHAR Xnext;
410    USPTR Xpp;    PCRE_PUCHAR Xpp;
411    USPTR Xprev;    PCRE_PUCHAR Xprev;
412    USPTR Xsaved_eptr;    PCRE_PUCHAR Xsaved_eptr;
413    
414    recursion_info Xnew_recursive;    recursion_info Xnew_recursive;
415    
# Line 372  typedef struct heapframe { Line 419  typedef struct heapframe {
419    
420  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
421    int Xprop_type;    int Xprop_type;
422    int Xprop_value;    unsigned int Xprop_value;
423    int Xprop_fail_result;    int Xprop_fail_result;
424    int Xoclength;    int Xoclength;
425    uschar Xocchars[8];    pcre_uchar Xocchars[6];
426  #endif  #endif
427    
428    int Xcodelink;    int Xcodelink;
# Line 385  typedef struct heapframe { Line 432  typedef struct heapframe {
432    int Xlength;    int Xlength;
433    int Xmax;    int Xmax;
434    int Xmin;    int Xmin;
435    int Xnumber;    unsigned int Xnumber;
436    int Xoffset;    int Xoffset;
437    int Xop;    unsigned int Xop;
438    int Xsave_capture_last;    pcre_int32 Xsave_capture_last;
439    int Xsave_offset1, Xsave_offset2, Xsave_offset3;    int Xsave_offset1, Xsave_offset2, Xsave_offset3;
440    int Xstacksave[REC_STACK_SAVE_MAX];    int Xstacksave[REC_STACK_SAVE_MAX];
441    
# Line 417  returns a negative (error) response, the Line 464  returns a negative (error) response, the
464  same response. */  same response. */
465    
466  /* These macros pack up tests that are used for partial matching, and which  /* These macros pack up tests that are used for partial matching, and which
467  appears several times in the code. We set the "hit end" flag if the pointer is  appear several times in the code. We set the "hit end" flag if the pointer is
468  at the end of the subject and also past the start of the subject (i.e.  at the end of the subject and also past the start of the subject (i.e.
469  something has been matched). For hard partial matching, we then return  something has been matched). For hard partial matching, we then return
470  immediately. The second one is used when we already know we are past the end of  immediately. The second one is used when we already know we are past the end of
# Line 440  the subject. */ Line 487  the subject. */
487    
488    
489  /* Performance note: It might be tempting to extract commonly used fields from  /* Performance note: It might be tempting to extract commonly used fields from
490  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
491  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
492  made performance worse.  made performance worse.
493    
# Line 463  Returns:       MATCH_MATCH if matched Line 510  Returns:       MATCH_MATCH if matched
510  */  */
511    
512  static int  static int
513  match(REGISTER USPTR eptr, REGISTER const uschar *ecode, USPTR mstart,  match(REGISTER PCRE_PUCHAR eptr, REGISTER const pcre_uchar *ecode,
514    int offset_top, match_data *md, eptrblock *eptrb, unsigned int rdepth)    PCRE_PUCHAR mstart, int offset_top, match_data *md, eptrblock *eptrb,
515      unsigned int rdepth)
516  {  {
517  /* 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,
518  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 520  so they can be ordinary variables in all
520    
521  register int  rrc;         /* Returns from recursive calls */  register int  rrc;         /* Returns from recursive calls */
522  register int  i;           /* Used for loops not involving calls to RMATCH() */  register int  i;           /* Used for loops not involving calls to RMATCH() */
523  register unsigned int c;   /* Character values not kept over RMATCH() calls */  register pcre_uint32 c;    /* Character values not kept over RMATCH() calls */
524  register BOOL utf8;        /* Local copy of UTF-8 flag for speed */  register BOOL utf;         /* Local copy of UTF flag for speed */
525    
526  BOOL minimize, possessive; /* Quantifier options */  BOOL minimize, possessive; /* Quantifier options */
527  BOOL caseless;  BOOL caseless;
528  int condcode;  int condcode;
529    
530  /* 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
531  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
532  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
533  heap whenever RMATCH() does a "recursion". See the macro definitions above. */  whenever RMATCH() does a "recursion". See the macro definitions above. Putting
534    the top-level on the stack rather than malloc-ing them all gives a performance
535    boost in many cases where there is not much "recursion". */
536    
537  #ifdef NO_RECURSE  #ifdef NO_RECURSE
538  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 */  
539    
540  /* Copy in the original argument variables */  /* Copy in the original argument variables */
541    
# Line 513  HEAP_RECURSE: Line 561  HEAP_RECURSE:
561    
562  /* Ditto for the local variables */  /* Ditto for the local variables */
563    
564  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
565  #define charptr            frame->Xcharptr  #define charptr            frame->Xcharptr
566  #endif  #endif
567  #define callpat            frame->Xcallpat  #define callpat            frame->Xcallpat
# Line 571  declarations can be cut out in a block. Line 619  declarations can be cut out in a block.
619  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
620  to RMATCH(). */  to RMATCH(). */
621    
622  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
623  const uschar *charptr;  const pcre_uchar *charptr;
624  #endif  #endif
625  const uschar *callpat;  const pcre_uchar *callpat;
626  const uschar *data;  const pcre_uchar *data;
627  const uschar *next;  const pcre_uchar *next;
628  USPTR         pp;  PCRE_PUCHAR       pp;
629  const uschar *prev;  const pcre_uchar *prev;
630  USPTR         saved_eptr;  PCRE_PUCHAR       saved_eptr;
631    
632  recursion_info new_recursive;  recursion_info new_recursive;
633    
# Line 589  BOOL prev_is_word; Line 637  BOOL prev_is_word;
637    
638  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
639  int prop_type;  int prop_type;
640  int prop_value;  unsigned int prop_value;
641  int prop_fail_result;  int prop_fail_result;
642  int oclength;  int oclength;
643  uschar occhars[8];  pcre_uchar occhars[6];
644  #endif  #endif
645    
646  int codelink;  int codelink;
# Line 600  int ctype; Line 648  int ctype;
648  int length;  int length;
649  int max;  int max;
650  int min;  int min;
651  int number;  unsigned int number;
652  int offset;  int offset;
653  int op;  unsigned int op;
654  int save_capture_last;  pcre_int32 save_capture_last;
655  int save_offset1, save_offset2, save_offset3;  int save_offset1, save_offset2, save_offset3;
656  int stacksave[REC_STACK_SAVE_MAX];  int stacksave[REC_STACK_SAVE_MAX];
657    
658  eptrblock newptrb;  eptrblock newptrb;
659    
660    /* There is a special fudge for calling match() in a way that causes it to
661    measure the size of its basic stack frame when the stack is being used for
662    recursion. The second argument (ecode) being NULL triggers this behaviour. It
663    cannot normally ever be NULL. The return is the negated value of the frame
664    size. */
665    
666    if (ecode == NULL)
667      {
668      if (rdepth == 0)
669        return match((PCRE_PUCHAR)&rdepth, NULL, NULL, 0, NULL, NULL, 1);
670      else
671        {
672        int len = (char *)&rdepth - (char *)eptr;
673        return (len > 0)? -len : len;
674        }
675      }
676  #endif     /* NO_RECURSE */  #endif     /* NO_RECURSE */
677    
678  /* 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 685  the alternative names that are used. */
685  #define code_offset   codelink  #define code_offset   codelink
686  #define condassert    condition  #define condassert    condition
687  #define matched_once  prev_is_word  #define matched_once  prev_is_word
688    #define foc           number
689    #define save_mark     data
690    
691  /* These statements are here to stop the compiler complaining about unitialized  /* These statements are here to stop the compiler complaining about unitialized
692  variables. */  variables. */
# Line 645  defined). However, RMATCH isn't like a f Line 712  defined). However, RMATCH isn't like a f
712  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,
713  however, impact performance when true recursion is being used. */  however, impact performance when true recursion is being used. */
714    
715  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
716  utf8 = md->utf8;       /* Local copy of the flag */  utf = md->utf;       /* Local copy of the flag */
717  #else  #else
718  utf8 = FALSE;  utf = FALSE;
719  #endif  #endif
720    
721  /* 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 756  for (;;)
756      case OP_MARK:      case OP_MARK:
757      md->nomatch_mark = ecode + 2;      md->nomatch_mark = ecode + 2;
758      md->mark = NULL;    /* In case previously set by assertion */      md->mark = NULL;    /* In case previously set by assertion */
759      RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode] + ecode[1], offset_top, md,      RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode] + ecode[1], offset_top, md,
760        eptrb, RM55);        eptrb, RM55);
761      if ((rrc == MATCH_MATCH || rrc == MATCH_ACCEPT) &&      if ((rrc == MATCH_MATCH || rrc == MATCH_ACCEPT) &&
762           md->mark == NULL) md->mark = ecode + 2;           md->mark == NULL) md->mark = ecode + 2;
# Line 702  for (;;) Line 769  for (;;)
769      unaltered. */      unaltered. */
770    
771      else if (rrc == MATCH_SKIP_ARG &&      else if (rrc == MATCH_SKIP_ARG &&
772          strcmp((char *)(ecode + 2), (char *)(md->start_match_ptr)) == 0)          STRCMP_UC_UC_TEST(ecode + 2, md->start_match_ptr) == 0)
773        {        {
774        md->start_match_ptr = eptr;        md->start_match_ptr = eptr;
775        RRETURN(MATCH_SKIP);        RRETURN(MATCH_SKIP);
# Line 712  for (;;) Line 779  for (;;)
779      case OP_FAIL:      case OP_FAIL:
780      RRETURN(MATCH_NOMATCH);      RRETURN(MATCH_NOMATCH);
781    
     /* COMMIT overrides PRUNE, SKIP, and THEN */  
   
782      case OP_COMMIT:      case OP_COMMIT:
783      RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,      RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md,
784        eptrb, RM52);        eptrb, RM52);
785      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);  
786      RRETURN(MATCH_COMMIT);      RRETURN(MATCH_COMMIT);
787    
     /* PRUNE overrides THEN */  
   
788      case OP_PRUNE:      case OP_PRUNE:
789      RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,      RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md,
790        eptrb, RM51);        eptrb, RM51);
791      if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN) RRETURN(rrc);      if (rrc != MATCH_NOMATCH) RRETURN(rrc);
792      RRETURN(MATCH_PRUNE);      RRETURN(MATCH_PRUNE);
793    
794      case OP_PRUNE_ARG:      case OP_PRUNE_ARG:
795      md->nomatch_mark = ecode + 2;      md->nomatch_mark = ecode + 2;
796      md->mark = NULL;    /* In case previously set by assertion */      md->mark = NULL;    /* In case previously set by assertion */
797      RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode] + ecode[1], offset_top, md,      RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode] + ecode[1], offset_top, md,
798        eptrb, RM56);        eptrb, RM56);
799      if ((rrc == MATCH_MATCH || rrc == MATCH_ACCEPT) &&      if ((rrc == MATCH_MATCH || rrc == MATCH_ACCEPT) &&
800           md->mark == NULL) md->mark = ecode + 2;           md->mark == NULL) md->mark = ecode + 2;
801      if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN) RRETURN(rrc);      if (rrc != MATCH_NOMATCH) RRETURN(rrc);
802      RRETURN(MATCH_PRUNE);      RRETURN(MATCH_PRUNE);
803    
     /* SKIP overrides PRUNE and THEN */  
   
804      case OP_SKIP:      case OP_SKIP:
805      RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,      RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md,
806        eptrb, RM53);        eptrb, RM53);
807      if (rrc != MATCH_NOMATCH && rrc != MATCH_PRUNE && rrc != MATCH_THEN)      if (rrc != MATCH_NOMATCH) RRETURN(rrc);
       RRETURN(rrc);  
808      md->start_match_ptr = eptr;   /* Pass back current position */      md->start_match_ptr = eptr;   /* Pass back current position */
809      RRETURN(MATCH_SKIP);      RRETURN(MATCH_SKIP);
810    
811      /* Note that, for Perl compatibility, SKIP with an argument does NOT set      /* Note that, for Perl compatibility, SKIP with an argument does NOT set
812      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
813      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
814        that failed and any that precede it (either they also failed, or were not
815        triggered). To do this, we maintain a count of executed SKIP_ARGs. If a
816        SKIP_ARG gets to top level, the match is re-run with md->ignore_skip_arg
817        set to the count of the one that failed. */
818    
819      case OP_SKIP_ARG:      case OP_SKIP_ARG:
820      if (md->ignore_skip_arg)      md->skip_arg_count++;
821        if (md->skip_arg_count <= md->ignore_skip_arg)
822        {        {
823        ecode += _pcre_OP_lengths[*ecode] + ecode[1];        ecode += PRIV(OP_lengths)[*ecode] + ecode[1];
824        break;        break;
825        }        }
826      RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode] + ecode[1], offset_top, md,      RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode] + ecode[1], offset_top, md,
827        eptrb, RM57);        eptrb, RM57);
828      if (rrc != MATCH_NOMATCH && rrc != MATCH_PRUNE && rrc != MATCH_THEN)      if (rrc != MATCH_NOMATCH) RRETURN(rrc);
       RRETURN(rrc);  
829    
830      /* 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
831      returning the special MATCH_SKIP_ARG return code. This will either be      returning the special MATCH_SKIP_ARG return code. This will either be
832      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
833      with the md->ignore_skip_arg flag set. */      with md->ignore_skip_arg set to the value of md->skip_arg_count. */
834    
835      md->start_match_ptr = ecode + 2;      md->start_match_ptr = ecode + 2;
836      RRETURN(MATCH_SKIP_ARG);      RRETURN(MATCH_SKIP_ARG);
# Line 779  for (;;) Line 840  for (;;)
840      match pointer to do this. */      match pointer to do this. */
841    
842      case OP_THEN:      case OP_THEN:
843      RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,      RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md,
844        eptrb, RM54);        eptrb, RM54);
845      if (rrc != MATCH_NOMATCH) RRETURN(rrc);      if (rrc != MATCH_NOMATCH) RRETURN(rrc);
846      md->start_match_ptr = ecode;      md->start_match_ptr = ecode;
# Line 788  for (;;) Line 849  for (;;)
849      case OP_THEN_ARG:      case OP_THEN_ARG:
850      md->nomatch_mark = ecode + 2;      md->nomatch_mark = ecode + 2;
851      md->mark = NULL;    /* In case previously set by assertion */      md->mark = NULL;    /* In case previously set by assertion */
852      RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode] + ecode[1], offset_top,      RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode] + ecode[1], offset_top,
853        md, eptrb, RM58);        md, eptrb, RM58);
854      if ((rrc == MATCH_MATCH || rrc == MATCH_ACCEPT) &&      if ((rrc == MATCH_MATCH || rrc == MATCH_ACCEPT) &&
855           md->mark == NULL) md->mark = ecode + 2;           md->mark == NULL) md->mark = ecode + 2;
# Line 812  for (;;) Line 873  for (;;)
873      case OP_ONCE_NC:      case OP_ONCE_NC:
874      prev = ecode;      prev = ecode;
875      saved_eptr = eptr;      saved_eptr = eptr;
876        save_mark = md->mark;
877      do      do
878        {        {
879        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 892  for (;;)
892    
893        if (rrc != MATCH_NOMATCH) RRETURN(rrc);        if (rrc != MATCH_NOMATCH) RRETURN(rrc);
894        ecode += GET(ecode,1);        ecode += GET(ecode,1);
895          md->mark = save_mark;
896        }        }
897      while (*ecode == OP_ALT);      while (*ecode == OP_ALT);
898    
# Line 869  for (;;) Line 932  for (;;)
932        }        }
933      else  /* OP_KETRMAX */      else  /* OP_KETRMAX */
934        {        {
       md->match_function_type = MATCH_CBEGROUP;  
935        RMATCH(eptr, prev, offset_top, md, eptrb, RM66);        RMATCH(eptr, prev, offset_top, md, eptrb, RM66);
936        if (rrc != MATCH_NOMATCH) RRETURN(rrc);        if (rrc != MATCH_NOMATCH) RRETURN(rrc);
937        ecode += 1 + LINK_SIZE;        ecode += 1 + LINK_SIZE;
# Line 909  for (;;) Line 971  for (;;)
971        save_offset2 = md->offset_vector[offset+1];        save_offset2 = md->offset_vector[offset+1];
972        save_offset3 = md->offset_vector[md->offset_end - number];        save_offset3 = md->offset_vector[md->offset_end - number];
973        save_capture_last = md->capture_last;        save_capture_last = md->capture_last;
974          save_mark = md->mark;
975    
976        DPRINTF(("saving %d %d %d\n", save_offset1, save_offset2, save_offset3));        DPRINTF(("saving %d %d %d\n", save_offset1, save_offset2, save_offset3));
977        md->offset_vector[md->offset_end - number] =        md->offset_vector[md->offset_end - number] =
# Line 917  for (;;) Line 980  for (;;)
980        for (;;)        for (;;)
981          {          {
982          if (op >= OP_SBRA) md->match_function_type = MATCH_CBEGROUP;          if (op >= OP_SBRA) md->match_function_type = MATCH_CBEGROUP;
983          RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,          RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md,
984            eptrb, RM1);            eptrb, RM1);
985          if (rrc == MATCH_ONCE) break;  /* Backing up through an atomic group */          if (rrc == MATCH_ONCE) break;  /* Backing up through an atomic group */
986    
# Line 945  for (;;) Line 1008  for (;;)
1008          if (rrc != MATCH_NOMATCH) RRETURN(rrc);          if (rrc != MATCH_NOMATCH) RRETURN(rrc);
1009          md->capture_last = save_capture_last;          md->capture_last = save_capture_last;
1010          ecode += GET(ecode, 1);          ecode += GET(ecode, 1);
1011            md->mark = save_mark;
1012          if (*ecode != OP_ALT) break;          if (*ecode != OP_ALT) break;
1013          }          }
1014    
# Line 996  for (;;) Line 1060  for (;;)
1060    
1061      for (;;)      for (;;)
1062        {        {
1063        if (op >= OP_SBRA || op == OP_ONCE) md->match_function_type = MATCH_CBEGROUP;        if (op >= OP_SBRA || op == OP_ONCE)
1064            md->match_function_type = MATCH_CBEGROUP;
1065    
1066        /* 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
1067        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 1069  for (;;)
1069    
1070        else if (!md->hasthen && ecode[GET(ecode, 1)] != OP_ALT)        else if (!md->hasthen && ecode[GET(ecode, 1)] != OP_ALT)
1071          {          {
1072          ecode += _pcre_OP_lengths[*ecode];          ecode += PRIV(OP_lengths)[*ecode];
1073          goto TAIL_RECURSE;          goto TAIL_RECURSE;
1074          }          }
1075    
1076        /* In all other cases, we have to make another call to match(). */        /* In all other cases, we have to make another call to match(). */
1077    
1078        RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md, eptrb,        save_mark = md->mark;
1079          save_capture_last = md->capture_last;
1080          RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md, eptrb,
1081          RM2);          RM2);
1082    
1083        /* 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 1095  for (;;)
1095          {          {
1096          if (rrc == MATCH_ONCE)          if (rrc == MATCH_ONCE)
1097            {            {
1098            const uschar *scode = ecode;            const pcre_uchar *scode = ecode;
1099            if (*scode != OP_ONCE)           /* If not at start, find it */            if (*scode != OP_ONCE)           /* If not at start, find it */
1100              {              {
1101              while (*scode == OP_ALT) scode += GET(scode, 1);              while (*scode == OP_ALT) scode += GET(scode, 1);
# Line 1039  for (;;) Line 1106  for (;;)
1106          RRETURN(rrc);          RRETURN(rrc);
1107          }          }
1108        ecode += GET(ecode, 1);        ecode += GET(ecode, 1);
1109          md->mark = save_mark;
1110        if (*ecode != OP_ALT) break;        if (*ecode != OP_ALT) break;
1111          md->capture_last = save_capture_last;
1112        }        }
1113    
1114      RRETURN(MATCH_NOMATCH);      RRETURN(MATCH_NOMATCH);
# Line 1070  for (;;) Line 1139  for (;;)
1139      if (offset < md->offset_max)      if (offset < md->offset_max)
1140        {        {
1141        matched_once = FALSE;        matched_once = FALSE;
1142        code_offset = ecode - md->start_code;        code_offset = (int)(ecode - md->start_code);
1143    
1144        save_offset1 = md->offset_vector[offset];        save_offset1 = md->offset_vector[offset];
1145        save_offset2 = md->offset_vector[offset+1];        save_offset2 = md->offset_vector[offset+1];
# Line 1093  for (;;) Line 1162  for (;;)
1162          md->offset_vector[md->offset_end - number] =          md->offset_vector[md->offset_end - number] =
1163            (int)(eptr - md->start_subject);            (int)(eptr - md->start_subject);
1164          if (op >= OP_SBRA) md->match_function_type = MATCH_CBEGROUP;          if (op >= OP_SBRA) md->match_function_type = MATCH_CBEGROUP;
1165          RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,          RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md,
1166            eptrb, RM63);            eptrb, RM63);
1167          if (rrc == MATCH_KETRPOS)          if (rrc == MATCH_KETRPOS)
1168            {            {
# Line 1102  for (;;) Line 1171  for (;;)
1171            ecode = md->start_code + code_offset;            ecode = md->start_code + code_offset;
1172            save_capture_last = md->capture_last;            save_capture_last = md->capture_last;
1173            matched_once = TRUE;            matched_once = TRUE;
1174              mstart = md->start_match_ptr;    /* In case \K changed it */
1175            continue;            continue;
1176            }            }
1177    
# Line 1160  for (;;) Line 1230  for (;;)
1230    
1231      POSSESSIVE_NON_CAPTURE:      POSSESSIVE_NON_CAPTURE:
1232      matched_once = FALSE;      matched_once = FALSE;
1233      code_offset = ecode - md->start_code;      code_offset = (int)(ecode - md->start_code);
1234        save_capture_last = md->capture_last;
1235    
1236      for (;;)      for (;;)
1237        {        {
1238        if (op >= OP_SBRA) md->match_function_type = MATCH_CBEGROUP;        if (op >= OP_SBRA) md->match_function_type = MATCH_CBEGROUP;
1239        RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,        RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md,
1240          eptrb, RM48);          eptrb, RM48);
1241        if (rrc == MATCH_KETRPOS)        if (rrc == MATCH_KETRPOS)
1242          {          {
# Line 1173  for (;;) Line 1244  for (;;)
1244          eptr = md->end_match_ptr;          eptr = md->end_match_ptr;
1245          ecode = md->start_code + code_offset;          ecode = md->start_code + code_offset;
1246          matched_once = TRUE;          matched_once = TRUE;
1247            mstart = md->start_match_ptr;   /* In case \K reset it */
1248          continue;          continue;
1249          }          }
1250    
# 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          ecode += PRIV(OP_lengths)[OP_CALLOUT];
1330          codelink -= PRIV(OP_lengths)[OP_CALLOUT];
1331          }
1332    
1333      /* Now see what the actual condition is */      /* 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;
1418      call to match() is always required, unless the second alternative doesn't  
1419      exist, in which case we can just plough on. Note that, for compatibility      /* We are now at the branch that is to be obeyed. As there is only one, we
1420      with Perl, the | in a conditional group is NOT treated as creating two      can use tail recursion to avoid using another stack frame, except when
1421      alternatives. If a THEN is encountered in the branch, it propagates out to      there is unlimited repeat of a possibly empty group. In the latter case, a
1422      the enclosing alternative (unless nested in a deeper set of alternatives,      recursive call to match() is always required, unless the second alternative
1423      of course). */      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 == OP_ALT)      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 1925  for (;;) Line 2007  for (;;)
2007    
2008      if (*ecode == OP_KETRPOS)      if (*ecode == OP_KETRPOS)
2009        {        {
2010          md->start_match_ptr = mstart;    /* In case \K reset it */
2011        md->end_match_ptr = eptr;        md->end_match_ptr = eptr;
2012        md->end_offset_top = offset_top;        md->end_offset_top = offset_top;
2013        RRETURN(MATCH_KETRPOS);        RRETURN(MATCH_KETRPOS);
# Line 1949  for (;;) Line 2032  for (;;)
2032          }          }
2033        if (*prev >= OP_SBRA)    /* Could match an empty string */        if (*prev >= OP_SBRA)    /* Could match an empty string */
2034          {          {
         md->match_function_type = MATCH_CBEGROUP;  
2035          RMATCH(eptr, prev, offset_top, md, eptrb, RM50);          RMATCH(eptr, prev, offset_top, md, eptrb, RM50);
2036          RRETURN(rrc);          RRETURN(rrc);
2037          }          }
# Line 1958  for (;;) Line 2040  for (;;)
2040        }        }
2041      else  /* OP_KETRMAX */      else  /* OP_KETRMAX */
2042        {        {
       if (*prev >= OP_SBRA) md->match_function_type = MATCH_CBEGROUP;  
2043        RMATCH(eptr, prev, offset_top, md, eptrb, RM13);        RMATCH(eptr, prev, offset_top, md, eptrb, RM13);
2044        if (rrc == MATCH_ONCE && md->once_target == prev) rrc = MATCH_NOMATCH;        if (rrc == MATCH_ONCE && md->once_target == prev) rrc = MATCH_NOMATCH;
2045        if (rrc != MATCH_NOMATCH) RRETURN(rrc);        if (rrc != MATCH_NOMATCH) RRETURN(rrc);
# Line 2015  for (;;) Line 2096  for (;;)
2096    
2097      case OP_DOLLM:      case OP_DOLLM:
2098      if (eptr < md->end_subject)      if (eptr < md->end_subject)
2099        { if (!IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH); }        {
2100          if (!IS_NEWLINE(eptr))
2101            {
2102            if (md->partial != 0 &&
2103                eptr + 1 >= md->end_subject &&
2104                NLBLOCK->nltype == NLTYPE_FIXED &&
2105                NLBLOCK->nllen == 2 &&
2106                RAWUCHARTEST(eptr) == NLBLOCK->nl[0])
2107              {
2108              md->hitend = TRUE;
2109              if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
2110              }
2111            RRETURN(MATCH_NOMATCH);
2112            }
2113          }
2114      else      else
2115        {        {
2116        if (md->noteol) RRETURN(MATCH_NOMATCH);        if (md->noteol) RRETURN(MATCH_NOMATCH);
# Line 2047  for (;;) Line 2142  for (;;)
2142      ASSERT_NL_OR_EOS:      ASSERT_NL_OR_EOS:
2143      if (eptr < md->end_subject &&      if (eptr < md->end_subject &&
2144          (!IS_NEWLINE(eptr) || eptr != md->end_subject - md->nllen))          (!IS_NEWLINE(eptr) || eptr != md->end_subject - md->nllen))
2145          {
2146          if (md->partial != 0 &&
2147              eptr + 1 >= md->end_subject &&
2148              NLBLOCK->nltype == NLTYPE_FIXED &&
2149              NLBLOCK->nllen == 2 &&
2150              RAWUCHARTEST(eptr) == NLBLOCK->nl[0])
2151            {
2152            md->hitend = TRUE;
2153            if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
2154            }
2155        RRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2156          }
2157    
2158      /* Either at end of string or \n before end. */      /* Either at end of string or \n before end. */
2159    
# Line 2066  for (;;) Line 2172  for (;;)
2172        be "non-word" characters. Remember the earliest consulted character for        be "non-word" characters. Remember the earliest consulted character for
2173        partial matching. */        partial matching. */
2174    
2175  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
2176        if (utf8)        if (utf)
2177          {          {
2178          /* Get status of previous character */          /* Get status of previous character */
2179    
2180          if (eptr == md->start_subject) prev_is_word = FALSE; else          if (eptr == md->start_subject) prev_is_word = FALSE; else
2181            {            {
2182            USPTR lastptr = eptr - 1;            PCRE_PUCHAR lastptr = eptr - 1;
2183            while((*lastptr & 0xc0) == 0x80) lastptr--;            BACKCHAR(lastptr);
2184            if (lastptr < md->start_used_ptr) md->start_used_ptr = lastptr;            if (lastptr < md->start_used_ptr) md->start_used_ptr = lastptr;
2185            GETCHAR(c, lastptr);            GETCHAR(c, lastptr);
2186  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
# Line 2139  for (;;) Line 2245  for (;;)
2245              }              }
2246            else            else
2247  #endif  #endif
2248            prev_is_word = ((md->ctypes[eptr[-1]] & ctype_word) != 0);            prev_is_word = MAX_255(eptr[-1])
2249                && ((md->ctypes[eptr[-1]] & ctype_word) != 0);
2250            }            }
2251    
2252          /* Get status of next character */          /* Get status of next character */
# Line 2162  for (;;) Line 2269  for (;;)
2269            }            }
2270          else          else
2271  #endif  #endif
2272          cur_is_word = ((md->ctypes[*eptr] & ctype_word) != 0);          cur_is_word = MAX_255(*eptr)
2273              && ((md->ctypes[*eptr] & ctype_word) != 0);
2274          }          }
2275    
2276        /* Now see if the situation is what we want */        /* Now see if the situation is what we want */
# Line 2173  for (;;) Line 2281  for (;;)
2281        }        }
2282      break;      break;
2283    
2284      /* Match a single character type; inline for speed */      /* Match any single character type except newline; have to take care with
2285        CRLF newlines and partial matching. */
2286    
2287      case OP_ANY:      case OP_ANY:
2288      if (IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH);      if (IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH);
2289        if (md->partial != 0 &&
2290            eptr + 1 >= md->end_subject &&
2291            NLBLOCK->nltype == NLTYPE_FIXED &&
2292            NLBLOCK->nllen == 2 &&
2293            RAWUCHARTEST(eptr) == NLBLOCK->nl[0])
2294          {
2295          md->hitend = TRUE;
2296          if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
2297          }
2298    
2299      /* Fall through */      /* Fall through */
2300    
2301        /* Match any single character whatsoever. */
2302    
2303      case OP_ALLANY:      case OP_ALLANY:
2304      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 */
2305        {                            /* not be updated before SCHECK_PARTIAL. */        {                            /* not be updated before SCHECK_PARTIAL. */
# Line 2186  for (;;) Line 2307  for (;;)
2307        RRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2308        }        }
2309      eptr++;      eptr++;
2310      if (utf8) while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;  #ifdef SUPPORT_UTF
2311        if (utf) ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);
2312    #endif
2313      ecode++;      ecode++;
2314      break;      break;
2315    
# Line 2211  for (;;) Line 2334  for (;;)
2334        }        }
2335      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2336      if (      if (
2337  #ifdef SUPPORT_UTF8  #if defined SUPPORT_UTF || !(defined COMPILE_PCRE8)
2338         c < 256 &&         c < 256 &&
2339  #endif  #endif
2340         (md->ctypes[c] & ctype_digit) != 0         (md->ctypes[c] & ctype_digit) != 0
# Line 2228  for (;;) Line 2351  for (;;)
2351        }        }
2352      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2353      if (      if (
2354  #ifdef SUPPORT_UTF8  #if defined SUPPORT_UTF || !(defined COMPILE_PCRE8)
2355         c >= 256 ||         c > 255 ||
2356  #endif  #endif
2357         (md->ctypes[c] & ctype_digit) == 0         (md->ctypes[c] & ctype_digit) == 0
2358         )         )
# Line 2245  for (;;) Line 2368  for (;;)
2368        }        }
2369      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2370      if (      if (
2371  #ifdef SUPPORT_UTF8  #if defined SUPPORT_UTF || !(defined COMPILE_PCRE8)
2372         c < 256 &&         c < 256 &&
2373  #endif  #endif
2374         (md->ctypes[c] & ctype_space) != 0         (md->ctypes[c] & ctype_space) != 0
# Line 2262  for (;;) Line 2385  for (;;)
2385        }        }
2386      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2387      if (      if (
2388  #ifdef SUPPORT_UTF8  #if defined SUPPORT_UTF || !(defined COMPILE_PCRE8)
2389         c >= 256 ||         c > 255 ||
2390  #endif  #endif
2391         (md->ctypes[c] & ctype_space) == 0         (md->ctypes[c] & ctype_space) == 0
2392         )         )
# Line 2279  for (;;) Line 2402  for (;;)
2402        }        }
2403      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2404      if (      if (
2405  #ifdef SUPPORT_UTF8  #if defined SUPPORT_UTF || !(defined COMPILE_PCRE8)
2406         c < 256 &&         c < 256 &&
2407  #endif  #endif
2408         (md->ctypes[c] & ctype_word) != 0         (md->ctypes[c] & ctype_word) != 0
# Line 2296  for (;;) Line 2419  for (;;)
2419        }        }
2420      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2421      if (      if (
2422  #ifdef SUPPORT_UTF8  #if defined SUPPORT_UTF || !(defined COMPILE_PCRE8)
2423         c >= 256 ||         c > 255 ||
2424  #endif  #endif
2425         (md->ctypes[c] & ctype_word) == 0         (md->ctypes[c] & ctype_word) == 0
2426         )         )
# Line 2316  for (;;) Line 2439  for (;;)
2439        {        {
2440        default: RRETURN(MATCH_NOMATCH);        default: RRETURN(MATCH_NOMATCH);
2441    
2442        case 0x000d:        case CHAR_CR:
2443        if (eptr < md->end_subject && *eptr == 0x0a) eptr++;        if (eptr >= md->end_subject)
2444            {
2445            SCHECK_PARTIAL();
2446            }
2447          else if (RAWUCHARTEST(eptr) == CHAR_LF) eptr++;
2448        break;        break;
2449    
2450        case 0x000a:        case CHAR_LF:
2451        break;        break;
2452    
2453        case 0x000b:        case CHAR_VT:
2454        case 0x000c:        case CHAR_FF:
2455        case 0x0085:        case CHAR_NEL:
2456    #ifndef EBCDIC
2457        case 0x2028:        case 0x2028:
2458        case 0x2029:        case 0x2029:
2459    #endif  /* Not EBCDIC */
2460        if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);        if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);
2461        break;        break;
2462        }        }
# Line 2343  for (;;) Line 2472  for (;;)
2472      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2473      switch(c)      switch(c)
2474        {        {
2475          HSPACE_CASES: RRETURN(MATCH_NOMATCH);  /* Byte and multibyte cases */
2476        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);  
2477        }        }
2478      ecode++;      ecode++;
2479      break;      break;
# Line 2377  for (;;) Line 2487  for (;;)
2487      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2488      switch(c)      switch(c)
2489        {        {
2490          HSPACE_CASES: break;  /* Byte and multibyte cases */
2491        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;  
2492        }        }
2493      ecode++;      ecode++;
2494      break;      break;
# Line 2411  for (;;) Line 2502  for (;;)
2502      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2503      switch(c)      switch(c)
2504        {        {
2505          VSPACE_CASES: RRETURN(MATCH_NOMATCH);
2506        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);  
2507        }        }
2508      ecode++;      ecode++;
2509      break;      break;
# Line 2433  for (;;) Line 2517  for (;;)
2517      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2518      switch(c)      switch(c)
2519        {        {
2520          VSPACE_CASES: break;
2521        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;  
2522        }        }
2523      ecode++;      ecode++;
2524      break;      break;
# Line 2459  for (;;) Line 2536  for (;;)
2536        }        }
2537      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2538        {        {
2539          const pcre_uint32 *cp;
2540        const ucd_record *prop = GET_UCD(c);        const ucd_record *prop = GET_UCD(c);
2541    
2542        switch(ecode[1])        switch(ecode[1])
# Line 2475  for (;;) Line 2553  for (;;)
2553          break;          break;
2554    
2555          case PT_GC:          case PT_GC:
2556          if ((ecode[2] != _pcre_ucp_gentype[prop->chartype]) == (op == OP_PROP))          if ((ecode[2] != PRIV(ucp_gentype)[prop->chartype]) == (op == OP_PROP))
2557            RRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
2558          break;          break;
2559    
# Line 2492  for (;;) Line 2570  for (;;)
2570          /* These are specials */          /* These are specials */
2571    
2572          case PT_ALNUM:          case PT_ALNUM:
2573          if ((_pcre_ucp_gentype[prop->chartype] == ucp_L ||          if ((PRIV(ucp_gentype)[prop->chartype] == ucp_L ||
2574               _pcre_ucp_gentype[prop->chartype] == ucp_N) == (op == OP_NOTPROP))               PRIV(ucp_gentype)[prop->chartype] == ucp_N) == (op == OP_NOTPROP))
2575            RRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
2576          break;          break;
2577    
2578          case PT_SPACE:    /* Perl space */          /* Perl space used to exclude VT, but from Perl 5.18 it is included,
2579          if ((_pcre_ucp_gentype[prop->chartype] == ucp_Z ||          which means that Perl space and POSIX space are now identical. PCRE
2580               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;  
2581    
2582            case PT_SPACE:    /* Perl space */
2583          case PT_PXSPACE:  /* POSIX space */          case PT_PXSPACE:  /* POSIX space */
2584          if ((_pcre_ucp_gentype[prop->chartype] == ucp_Z ||          switch(c)
2585               c == CHAR_HT || c == CHAR_NL || c == CHAR_VT ||            {
2586               c == CHAR_FF || c == CHAR_CR)            HSPACE_CASES:
2587                 == (op == OP_NOTPROP))            VSPACE_CASES:
2588            RRETURN(MATCH_NOMATCH);            if (op == OP_NOTPROP) RRETURN(MATCH_NOMATCH);
2589              break;
2590    
2591              default:
2592              if ((PRIV(ucp_gentype)[prop->chartype] == ucp_Z) ==
2593                (op == OP_NOTPROP)) RRETURN(MATCH_NOMATCH);
2594              break;
2595              }
2596          break;          break;
2597    
2598          case PT_WORD:          case PT_WORD:
2599          if ((_pcre_ucp_gentype[prop->chartype] == ucp_L ||          if ((PRIV(ucp_gentype)[prop->chartype] == ucp_L ||
2600               _pcre_ucp_gentype[prop->chartype] == ucp_N ||               PRIV(ucp_gentype)[prop->chartype] == ucp_N ||
2601               c == CHAR_UNDERSCORE) == (op == OP_NOTPROP))               c == CHAR_UNDERSCORE) == (op == OP_NOTPROP))
2602            RRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
2603          break;          break;
2604    
2605            case PT_CLIST:
2606            cp = PRIV(ucd_caseless_sets) + ecode[2];
2607            for (;;)
2608              {
2609              if (c < *cp)
2610                { if (op == OP_PROP) { RRETURN(MATCH_NOMATCH); } else break; }
2611              if (c == *cp++)
2612                { if (op == OP_PROP) break; else { RRETURN(MATCH_NOMATCH); } }
2613              }
2614            break;
2615    
2616            case PT_UCNC:
2617            if ((c == CHAR_DOLLAR_SIGN || c == CHAR_COMMERCIAL_AT ||
2618                 c == CHAR_GRAVE_ACCENT || (c >= 0xa0 && c <= 0xd7ff) ||
2619                 c >= 0xe000) == (op == OP_NOTPROP))
2620              RRETURN(MATCH_NOMATCH);
2621            break;
2622    
2623          /* This should never occur */          /* This should never occur */
2624    
2625          default:          default:
# Line 2538  for (;;) Line 2639  for (;;)
2639        SCHECK_PARTIAL();        SCHECK_PARTIAL();
2640        RRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2641        }        }
2642      GETCHARINCTEST(c, eptr);      else
     if (UCD_CATEGORY(c) == ucp_M) RRETURN(MATCH_NOMATCH);  
     while (eptr < md->end_subject)  
2643        {        {
2644        int len = 1;        int lgb, rgb;
2645        if (!utf8) c = *eptr; else { GETCHARLEN(c, eptr, len); }        GETCHARINCTEST(c, eptr);
2646        if (UCD_CATEGORY(c) != ucp_M) break;        lgb = UCD_GRAPHBREAK(c);
2647        eptr += len;        while (eptr < md->end_subject)
2648            {
2649            int len = 1;
2650            if (!utf) c = *eptr; else { GETCHARLEN(c, eptr, len); }
2651            rgb = UCD_GRAPHBREAK(c);
2652            if ((PRIV(ucp_gbtable)[lgb] & (1 << rgb)) == 0) break;
2653            lgb = rgb;
2654            eptr += len;
2655            }
2656        }        }
2657        CHECK_PARTIAL();
2658      ecode++;      ecode++;
2659      break;      break;
2660  #endif  #endif  /* SUPPORT_UCP */
2661    
2662    
2663      /* 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 2666  for (;;)
2666      similar code to character type repeats - written out again for speed.      similar code to character type repeats - written out again for speed.
2667      However, if the referenced string is the empty string, always treat      However, if the referenced string is the empty string, always treat
2668      it as matched, any number of times (otherwise there could be infinite      it as matched, any number of times (otherwise there could be infinite
2669      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:  
2670    
2671      (a) In the default, Perl-compatible state, set the length negative;      (a) In the default, Perl-compatible state, set the length negative;
2672      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 2676  for (;;)
2676      so that the back reference matches an empty string.      so that the back reference matches an empty string.
2677    
2678      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
2679      referenced subpattern. */      referenced subpattern.
2680    
2681        The OP_REF and OP_REFI opcodes are used for a reference to a numbered group
2682        or to a non-duplicated named group. For a duplicated named group, OP_DNREF
2683        and OP_DNREFI are used. In this case we must scan the list of groups to
2684        which the name refers, and use the first one that is set. */
2685    
2686        case OP_DNREF:
2687        case OP_DNREFI:
2688        caseless = op == OP_DNREFI;
2689          {
2690          int count = GET2(ecode, 1+IMM2_SIZE);
2691          pcre_uchar *slot = md->name_table + GET2(ecode, 1) * md->name_entry_size;
2692          ecode += 1 + 2*IMM2_SIZE;
2693    
2694          while (count-- > 0)
2695            {
2696            offset = GET2(slot, 0) << 1;
2697            if (offset < offset_top && md->offset_vector[offset] >= 0) break;
2698            slot += md->name_entry_size;
2699            }
2700          if (count < 0)
2701            length = (md->jscript_compat)? 0 : -1;
2702          else
2703            length = md->offset_vector[offset+1] - md->offset_vector[offset];
2704          }
2705        goto REF_REPEAT;
2706    
2707        case OP_REF:
2708        case OP_REFI:
2709        caseless = op == OP_REFI;
2710        offset = GET2(ecode, 1) << 1;               /* Doubled ref number */
2711        ecode += 1 + IMM2_SIZE;
2712      if (offset >= offset_top || md->offset_vector[offset] < 0)      if (offset >= offset_top || md->offset_vector[offset] < 0)
2713        length = (md->jscript_compat)? 0 : -1;        length = (md->jscript_compat)? 0 : -1;
2714      else      else
# Line 2585  for (;;) Line 2716  for (;;)
2716    
2717      /* Set up for repetition, or handle the non-repeated case */      /* Set up for repetition, or handle the non-repeated case */
2718    
2719        REF_REPEAT:
2720      switch (*ecode)      switch (*ecode)
2721        {        {
2722        case OP_CRSTAR:        case OP_CRSTAR:
# Line 2604  for (;;) Line 2736  for (;;)
2736        case OP_CRMINRANGE:        case OP_CRMINRANGE:
2737        minimize = (*ecode == OP_CRMINRANGE);        minimize = (*ecode == OP_CRMINRANGE);
2738        min = GET2(ecode, 1);        min = GET2(ecode, 1);
2739        max = GET2(ecode, 3);        max = GET2(ecode, 1 + IMM2_SIZE);
2740        if (max == 0) max = INT_MAX;        if (max == 0) max = INT_MAX;
2741        ecode += 5;        ecode += 1 + 2 * IMM2_SIZE;
2742        break;        break;
2743    
2744        default:               /* No repeat follows */        default:               /* No repeat follows */
2745        if ((length = match_ref(offset, eptr, length, md, caseless)) < 0)        if ((length = match_ref(offset, eptr, length, md, caseless)) < 0)
2746          {          {
2747            if (length == -2) eptr = md->end_subject;   /* Partial match */
2748          CHECK_PARTIAL();          CHECK_PARTIAL();
2749          RRETURN(MATCH_NOMATCH);          RRETURN(MATCH_NOMATCH);
2750          }          }
# Line 2620  for (;;) Line 2753  for (;;)
2753        }        }
2754    
2755      /* Handle repeated back references. If the length of the reference is      /* Handle repeated back references. If the length of the reference is
2756      zero, just continue with the main loop. */      zero, just continue with the main loop. If the length is negative, it
2757        means the reference is unset in non-Java-compatible mode. If the minimum is
2758        zero, we can continue at the same level without recursion. For any other
2759        minimum, carrying on will result in NOMATCH. */
2760    
2761      if (length == 0) continue;      if (length == 0) continue;
2762        if (length < 0 && min == 0) continue;
2763    
2764      /* First, ensure the minimum number of matches are present. We get back      /* First, ensure the minimum number of matches are present. We get back
2765      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 2770  for (;;)
2770        int slength;        int slength;
2771        if ((slength = match_ref(offset, eptr, length, md, caseless)) < 0)        if ((slength = match_ref(offset, eptr, length, md, caseless)) < 0)
2772          {          {
2773            if (slength == -2) eptr = md->end_subject;   /* Partial match */
2774          CHECK_PARTIAL();          CHECK_PARTIAL();
2775          RRETURN(MATCH_NOMATCH);          RRETURN(MATCH_NOMATCH);
2776          }          }
# Line 2656  for (;;) Line 2794  for (;;)
2794          if (fi >= max) RRETURN(MATCH_NOMATCH);          if (fi >= max) RRETURN(MATCH_NOMATCH);
2795          if ((slength = match_ref(offset, eptr, length, md, caseless)) < 0)          if ((slength = match_ref(offset, eptr, length, md, caseless)) < 0)
2796            {            {
2797              if (slength == -2) eptr = md->end_subject;   /* Partial match */
2798            CHECK_PARTIAL();            CHECK_PARTIAL();
2799            RRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
2800            }            }
# Line 2674  for (;;) Line 2813  for (;;)
2813          int slength;          int slength;
2814          if ((slength = match_ref(offset, eptr, length, md, caseless)) < 0)          if ((slength = match_ref(offset, eptr, length, md, caseless)) < 0)
2815            {            {
2816            CHECK_PARTIAL();            /* Can't use CHECK_PARTIAL because we don't want to update eptr in
2817              the soft partial matching case. */
2818    
2819              if (slength == -2 && md->partial != 0 &&
2820                  md->end_subject > md->start_used_ptr)
2821                {
2822                md->hitend = TRUE;
2823                if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
2824                }
2825            break;            break;
2826            }            }
2827          eptr += slength;          eptr += slength;
2828          }          }
2829    
2830        while (eptr >= pp)        while (eptr >= pp)
2831          {          {
2832          RMATCH(eptr, ecode, offset_top, md, eptrb, RM15);          RMATCH(eptr, ecode, offset_top, md, eptrb, RM15);
# Line 2703  for (;;) Line 2851  for (;;)
2851      case OP_NCLASS:      case OP_NCLASS:
2852      case OP_CLASS:      case OP_CLASS:
2853        {        {
2854          /* The data variable is saved across frames, so the byte map needs to
2855          be stored there. */
2856    #define BYTE_MAP ((pcre_uint8 *)data)
2857        data = ecode + 1;                /* Save for matching */        data = ecode + 1;                /* Save for matching */
2858        ecode += 33;                     /* Advance past the item */        ecode += 1 + (32 / sizeof(pcre_uchar)); /* Advance past the item */
2859    
2860        switch (*ecode)        switch (*ecode)
2861          {          {
# Line 2714  for (;;) Line 2865  for (;;)
2865          case OP_CRMINPLUS:          case OP_CRMINPLUS:
2866          case OP_CRQUERY:          case OP_CRQUERY:
2867          case OP_CRMINQUERY:          case OP_CRMINQUERY:
2868            case OP_CRPOSSTAR:
2869            case OP_CRPOSPLUS:
2870            case OP_CRPOSQUERY:
2871          c = *ecode++ - OP_CRSTAR;          c = *ecode++ - OP_CRSTAR;
2872          minimize = (c & 1) != 0;          if (c < OP_CRPOSSTAR - OP_CRSTAR) minimize = (c & 1) != 0;
2873            else possessive = TRUE;
2874          min = rep_min[c];                 /* Pick up values from tables; */          min = rep_min[c];                 /* Pick up values from tables; */
2875          max = rep_max[c];                 /* zero for max => infinity */          max = rep_max[c];                 /* zero for max => infinity */
2876          if (max == 0) max = INT_MAX;          if (max == 0) max = INT_MAX;
# Line 2723  for (;;) Line 2878  for (;;)
2878    
2879          case OP_CRRANGE:          case OP_CRRANGE:
2880          case OP_CRMINRANGE:          case OP_CRMINRANGE:
2881            case OP_CRPOSRANGE:
2882          minimize = (*ecode == OP_CRMINRANGE);          minimize = (*ecode == OP_CRMINRANGE);
2883            possessive = (*ecode == OP_CRPOSRANGE);
2884          min = GET2(ecode, 1);          min = GET2(ecode, 1);
2885          max = GET2(ecode, 3);          max = GET2(ecode, 1 + IMM2_SIZE);
2886          if (max == 0) max = INT_MAX;          if (max == 0) max = INT_MAX;
2887          ecode += 5;          ecode += 1 + 2 * IMM2_SIZE;
2888          break;          break;
2889    
2890          default:               /* No repeat follows */          default:               /* No repeat follows */
# Line 2737  for (;;) Line 2894  for (;;)
2894    
2895        /* First, ensure the minimum number of matches are present. */        /* First, ensure the minimum number of matches are present. */
2896    
2897  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
2898        /* UTF-8 mode */        if (utf)
       if (utf8)  
2899          {          {
2900          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
2901            {            {
# Line 2754  for (;;) Line 2910  for (;;)
2910              if (op == OP_CLASS) RRETURN(MATCH_NOMATCH);              if (op == OP_CLASS) RRETURN(MATCH_NOMATCH);
2911              }              }
2912            else            else
2913              {              if ((BYTE_MAP[c/8] & (1 << (c&7))) == 0) RRETURN(MATCH_NOMATCH);
             if ((data[c/8] & (1 << (c&7))) == 0) RRETURN(MATCH_NOMATCH);  
             }  
2914            }            }
2915          }          }
2916        else        else
2917  #endif  #endif
2918        /* Not UTF-8 mode */        /* Not UTF mode */
2919          {          {
2920          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
2921            {            {
# Line 2771  for (;;) Line 2925  for (;;)
2925              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
2926              }              }
2927            c = *eptr++;            c = *eptr++;
2928            if ((data[c/8] & (1 << (c&7))) == 0) RRETURN(MATCH_NOMATCH);  #ifndef COMPILE_PCRE8
2929              if (c > 255)
2930                {
2931                if (op == OP_CLASS) RRETURN(MATCH_NOMATCH);
2932                }
2933              else
2934    #endif
2935                if ((BYTE_MAP[c/8] & (1 << (c&7))) == 0) RRETURN(MATCH_NOMATCH);
2936            }            }
2937          }          }
2938    
# Line 2785  for (;;) Line 2946  for (;;)
2946    
2947        if (minimize)        if (minimize)
2948          {          {
2949  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
2950          /* UTF-8 mode */          if (utf)
         if (utf8)  
2951            {            {
2952            for (fi = min;; fi++)            for (fi = min;; fi++)
2953              {              {
# Line 2805  for (;;) Line 2965  for (;;)
2965                if (op == OP_CLASS) RRETURN(MATCH_NOMATCH);                if (op == OP_CLASS) RRETURN(MATCH_NOMATCH);
2966                }                }
2967              else              else
2968                {                if ((BYTE_MAP[c/8] & (1 << (c&7))) == 0) RRETURN(MATCH_NOMATCH);
               if ((data[c/8] & (1 << (c&7))) == 0) RRETURN(MATCH_NOMATCH);  
               }  
2969              }              }
2970            }            }
2971          else          else
2972  #endif  #endif
2973          /* Not UTF-8 mode */          /* Not UTF mode */
2974            {            {
2975            for (fi = min;; fi++)            for (fi = min;; fi++)
2976              {              {
# Line 2825  for (;;) Line 2983  for (;;)
2983                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
2984                }                }
2985              c = *eptr++;              c = *eptr++;
2986              if ((data[c/8] & (1 << (c&7))) == 0) RRETURN(MATCH_NOMATCH);  #ifndef COMPILE_PCRE8
2987                if (c > 255)
2988                  {
2989                  if (op == OP_CLASS) RRETURN(MATCH_NOMATCH);
2990                  }
2991                else
2992    #endif
2993                  if ((BYTE_MAP[c/8] & (1 << (c&7))) == 0) RRETURN(MATCH_NOMATCH);
2994              }              }
2995            }            }
2996          /* Control never gets here */          /* Control never gets here */
# Line 2837  for (;;) Line 3002  for (;;)
3002          {          {
3003          pp = eptr;          pp = eptr;
3004    
3005  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
3006          /* UTF-8 mode */          if (utf)
         if (utf8)  
3007            {            {
3008            for (i = min; i < max; i++)            for (i = min; i < max; i++)
3009              {              {
# Line 2855  for (;;) Line 3019  for (;;)
3019                if (op == OP_CLASS) break;                if (op == OP_CLASS) break;
3020                }                }
3021              else              else
3022                {                if ((BYTE_MAP[c/8] & (1 << (c&7))) == 0) break;
               if ((data[c/8] & (1 << (c&7))) == 0) break;  
               }  
3023              eptr += len;              eptr += len;
3024              }              }
3025    
3026              if (possessive) continue;    /* No backtracking */
3027    
3028            for (;;)            for (;;)
3029              {              {
3030              RMATCH(eptr, ecode, offset_top, md, eptrb, RM18);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM18);
# Line 2870  for (;;) Line 3035  for (;;)
3035            }            }
3036          else          else
3037  #endif  #endif
3038            /* Not UTF-8 mode */            /* Not UTF mode */
3039            {            {
3040            for (i = min; i < max; i++)            for (i = min; i < max; i++)
3041              {              {
# Line 2880  for (;;) Line 3045  for (;;)
3045                break;                break;
3046                }                }
3047              c = *eptr;              c = *eptr;
3048              if ((data[c/8] & (1 << (c&7))) == 0) break;  #ifndef COMPILE_PCRE8
3049              eptr++;              if (c > 255)
3050                  {
3051                  if (op == OP_CLASS) break;
3052                  }
3053                else
3054    #endif
3055                  if ((BYTE_MAP[c/8] & (1 << (c&7))) == 0) break;
3056                eptr++;
3057              }              }
3058    
3059              if (possessive) continue;    /* No backtracking */
3060    
3061            while (eptr >= pp)            while (eptr >= pp)
3062              {              {
3063              RMATCH(eptr, ecode, offset_top, md, eptrb, RM19);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM19);
# Line 2893  for (;;) Line 3068  for (;;)
3068    
3069          RRETURN(MATCH_NOMATCH);          RRETURN(MATCH_NOMATCH);
3070          }          }
3071    #undef BYTE_MAP
3072        }        }
3073      /* Control never gets here */      /* Control never gets here */
3074    
3075    
3076      /* Match an extended character class. This opcode is encountered only      /* Match an extended character class. In the 8-bit library, this opcode is
3077      when UTF-8 mode mode is supported. Nevertheless, we may not be in UTF-8      encountered only when UTF-8 mode mode is supported. In the 16-bit and
3078      mode, because Unicode properties are supported in non-UTF-8 mode. */      32-bit libraries, codepoints greater than 255 may be encountered even when
3079        UTF is not supported. */
3080    
3081  #ifdef SUPPORT_UTF8  #if defined SUPPORT_UTF || !defined COMPILE_PCRE8
3082      case OP_XCLASS:      case OP_XCLASS:
3083        {        {
3084        data = ecode + 1 + LINK_SIZE;                /* Save for matching */        data = ecode + 1 + LINK_SIZE;                /* Save for matching */
# Line 2915  for (;;) Line 3092  for (;;)
3092          case OP_CRMINPLUS:          case OP_CRMINPLUS:
3093          case OP_CRQUERY:          case OP_CRQUERY:
3094          case OP_CRMINQUERY:          case OP_CRMINQUERY:
3095            case OP_CRPOSSTAR:
3096            case OP_CRPOSPLUS:
3097            case OP_CRPOSQUERY:
3098          c = *ecode++ - OP_CRSTAR;          c = *ecode++ - OP_CRSTAR;
3099          minimize = (c & 1) != 0;          if (c < OP_CRPOSSTAR - OP_CRSTAR) minimize = (c & 1) != 0;
3100            else possessive = TRUE;
3101          min = rep_min[c];                 /* Pick up values from tables; */          min = rep_min[c];                 /* Pick up values from tables; */
3102          max = rep_max[c];                 /* zero for max => infinity */          max = rep_max[c];                 /* zero for max => infinity */
3103          if (max == 0) max = INT_MAX;          if (max == 0) max = INT_MAX;
# Line 2924  for (;;) Line 3105  for (;;)
3105    
3106          case OP_CRRANGE:          case OP_CRRANGE:
3107          case OP_CRMINRANGE:          case OP_CRMINRANGE:
3108            case OP_CRPOSRANGE:
3109          minimize = (*ecode == OP_CRMINRANGE);          minimize = (*ecode == OP_CRMINRANGE);
3110            possessive = (*ecode == OP_CRPOSRANGE);
3111          min = GET2(ecode, 1);          min = GET2(ecode, 1);
3112          max = GET2(ecode, 3);          max = GET2(ecode, 1 + IMM2_SIZE);
3113          if (max == 0) max = INT_MAX;          if (max == 0) max = INT_MAX;
3114          ecode += 5;          ecode += 1 + 2 * IMM2_SIZE;
3115          break;          break;
3116    
3117          default:               /* No repeat follows */          default:               /* No repeat follows */
# Line 2946  for (;;) Line 3129  for (;;)
3129            RRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
3130            }            }
3131          GETCHARINCTEST(c, eptr);          GETCHARINCTEST(c, eptr);
3132          if (!_pcre_xclass(c, data)) RRETURN(MATCH_NOMATCH);          if (!PRIV(xclass)(c, data, utf)) RRETURN(MATCH_NOMATCH);
3133          }          }
3134    
3135        /* 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 3153  for (;;)
3153              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
3154              }              }
3155            GETCHARINCTEST(c, eptr);            GETCHARINCTEST(c, eptr);
3156            if (!_pcre_xclass(c, data)) RRETURN(MATCH_NOMATCH);            if (!PRIV(xclass)(c, data, utf)) RRETURN(MATCH_NOMATCH);
3157            }            }
3158          /* Control never gets here */          /* Control never gets here */
3159          }          }
# Line 2988  for (;;) Line 3171  for (;;)
3171              SCHECK_PARTIAL();              SCHECK_PARTIAL();
3172              break;              break;
3173              }              }
3174    #ifdef SUPPORT_UTF
3175            GETCHARLENTEST(c, eptr, len);            GETCHARLENTEST(c, eptr, len);
3176            if (!_pcre_xclass(c, data)) break;  #else
3177              c = *eptr;
3178    #endif
3179              if (!PRIV(xclass)(c, data, utf)) break;
3180            eptr += len;            eptr += len;
3181            }            }
3182    
3183            if (possessive) continue;    /* No backtracking */
3184    
3185          for(;;)          for(;;)
3186            {            {
3187            RMATCH(eptr, ecode, offset_top, md, eptrb, RM21);            RMATCH(eptr, ecode, offset_top, md, eptrb, RM21);
3188            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3189            if (eptr-- == pp) break;        /* Stop if tried at original pos */            if (eptr-- == pp) break;        /* Stop if tried at original pos */
3190            if (utf8) BACKCHAR(eptr);  #ifdef SUPPORT_UTF
3191              if (utf) BACKCHAR(eptr);
3192    #endif
3193            }            }
3194          RRETURN(MATCH_NOMATCH);          RRETURN(MATCH_NOMATCH);
3195          }          }
# Line 3009  for (;;) Line 3201  for (;;)
3201      /* Match a single character, casefully */      /* Match a single character, casefully */
3202    
3203      case OP_CHAR:      case OP_CHAR:
3204  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
3205      if (utf8)      if (utf)
3206        {        {
3207        length = 1;        length = 1;
3208        ecode++;        ecode++;
# Line 3020  for (;;) Line 3212  for (;;)
3212          CHECK_PARTIAL();             /* Not SCHECK_PARTIAL() */          CHECK_PARTIAL();             /* Not SCHECK_PARTIAL() */
3213          RRETURN(MATCH_NOMATCH);          RRETURN(MATCH_NOMATCH);
3214          }          }
3215        while (length-- > 0) if (*ecode++ != *eptr++) RRETURN(MATCH_NOMATCH);        while (length-- > 0) if (*ecode++ != RAWUCHARINC(eptr)) RRETURN(MATCH_NOMATCH);
3216        }        }
3217      else      else
3218  #endif  #endif
3219        /* Not UTF mode */
     /* Non-UTF-8 mode */  
3220        {        {
3221        if (md->end_subject - eptr < 1)        if (md->end_subject - eptr < 1)
3222          {          {
# Line 3037  for (;;) Line 3228  for (;;)
3228        }        }
3229      break;      break;
3230    
3231      /* Match a single character, caselessly */      /* Match a single character, caselessly. If we are at the end of the
3232        subject, give up immediately. */
3233    
3234      case OP_CHARI:      case OP_CHARI:
3235  #ifdef SUPPORT_UTF8      if (eptr >= md->end_subject)
3236      if (utf8)        {
3237          SCHECK_PARTIAL();
3238          RRETURN(MATCH_NOMATCH);
3239          }
3240    
3241    #ifdef SUPPORT_UTF
3242        if (utf)
3243        {        {
3244        length = 1;        length = 1;
3245        ecode++;        ecode++;
3246        GETCHARLEN(fc, ecode, length);        GETCHARLEN(fc, ecode, length);
3247    
       if (length > md->end_subject - eptr)  
         {  
         CHECK_PARTIAL();             /* Not SCHECK_PARTIAL() */  
         RRETURN(MATCH_NOMATCH);  
         }  
   
3248        /* If the pattern character's value is < 128, we have only one byte, and        /* If the pattern character's value is < 128, we have only one byte, and
3249        can use the fast lookup table. */        we know that its other case must also be one byte long, so we can use the
3250          fast lookup table. We know that there is at least one byte left in the
3251          subject. */
3252    
3253        if (fc < 128)        if (fc < 128)
3254          {          {
3255          if (md->lcc[*ecode++] != md->lcc[*eptr++]) RRETURN(MATCH_NOMATCH);          pcre_uint32 cc = RAWUCHAR(eptr);
3256            if (md->lcc[fc] != TABLE_GET(cc, md->lcc, cc)) RRETURN(MATCH_NOMATCH);
3257            ecode++;
3258            eptr++;
3259          }          }
3260    
3261        /* Otherwise we must pick up the subject character */        /* Otherwise we must pick up the subject character. Note that we cannot
3262          use the value of "length" to check for sufficient bytes left, because the
3263          other case of the character may have more or fewer bytes.  */
3264    
3265        else        else
3266          {          {
3267          unsigned int dc;          pcre_uint32 dc;
3268          GETCHARINC(dc, eptr);          GETCHARINC(dc, eptr);
3269          ecode += length;          ecode += length;
3270    
# Line 3082  for (;;) Line 3281  for (;;)
3281          }          }
3282        }        }
3283      else      else
3284  #endif   /* SUPPORT_UTF8 */  #endif   /* SUPPORT_UTF */
3285    
3286      /* Non-UTF-8 mode */      /* Not UTF mode */
3287        {        {
3288        if (md->end_subject - eptr < 1)        if (TABLE_GET(ecode[1], md->lcc, ecode[1])
3289          {            != TABLE_GET(*eptr, md->lcc, *eptr)) RRETURN(MATCH_NOMATCH);
3290          SCHECK_PARTIAL();            /* This one can use SCHECK_PARTIAL() */        eptr++;
         RRETURN(MATCH_NOMATCH);  
         }  
       if (md->lcc[ecode[1]] != md->lcc[*eptr++]) RRETURN(MATCH_NOMATCH);  
3291        ecode += 2;        ecode += 2;
3292        }        }
3293      break;      break;
# Line 3101  for (;;) Line 3297  for (;;)
3297      case OP_EXACT:      case OP_EXACT:
3298      case OP_EXACTI:      case OP_EXACTI:
3299      min = max = GET2(ecode, 1);      min = max = GET2(ecode, 1);
3300      ecode += 3;      ecode += 1 + IMM2_SIZE;
3301      goto REPEATCHAR;      goto REPEATCHAR;
3302    
3303      case OP_POSUPTO:      case OP_POSUPTO:
# Line 3116  for (;;) Line 3312  for (;;)
3312      min = 0;      min = 0;
3313      max = GET2(ecode, 1);      max = GET2(ecode, 1);
3314      minimize = *ecode == OP_MINUPTO || *ecode == OP_MINUPTOI;      minimize = *ecode == OP_MINUPTO || *ecode == OP_MINUPTOI;
3315      ecode += 3;      ecode += 1 + IMM2_SIZE;
3316      goto REPEATCHAR;      goto REPEATCHAR;
3317    
3318      case OP_POSSTAR:      case OP_POSSTAR:
# Line 3161  for (;;) Line 3357  for (;;)
3357      max = rep_max[c];                 /* zero for max => infinity */      max = rep_max[c];                 /* zero for max => infinity */
3358      if (max == 0) max = INT_MAX;      if (max == 0) max = INT_MAX;
3359    
3360      /* Common code for all repeated single-character matches. */      /* Common code for all repeated single-character matches. We first check
3361        for the minimum number of characters. If the minimum equals the maximum, we
3362        are done. Otherwise, if minimizing, check the rest of the pattern for a
3363        match; if there isn't one, advance up to the maximum, one character at a
3364        time.
3365    
3366        If maximizing, advance up to the maximum number of matching characters,
3367        until eptr is past the end of the maximum run. If possessive, we are
3368        then done (no backing up). Otherwise, match at this position; anything
3369        other than no match is immediately returned. For nomatch, back up one
3370        character, unless we are matching \R and the last thing matched was
3371        \r\n, in which case, back up two bytes. When we reach the first optional
3372        character position, we can save stack by doing a tail recurse.
3373    
3374        The various UTF/non-UTF and caseful/caseless cases are handled separately,
3375        for speed. */
3376    
3377      REPEATCHAR:      REPEATCHAR:
3378  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
3379      if (utf8)      if (utf)
3380        {        {
3381        length = 1;        length = 1;
3382        charptr = ecode;        charptr = ecode;
# Line 3178  for (;;) Line 3389  for (;;)
3389        if (length > 1)        if (length > 1)
3390          {          {
3391  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
3392          unsigned int othercase;          pcre_uint32 othercase;
3393          if (op >= OP_STARI &&     /* Caseless */          if (op >= OP_STARI &&     /* Caseless */
3394              (othercase = UCD_OTHERCASE(fc)) != fc)              (othercase = UCD_OTHERCASE(fc)) != fc)
3395            oclength = _pcre_ord2utf8(othercase, occhars);            oclength = PRIV(ord2utf)(othercase, occhars);
3396          else oclength = 0;          else oclength = 0;
3397  #endif  /* SUPPORT_UCP */  #endif  /* SUPPORT_UCP */
3398    
3399          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
3400            {            {
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 3211  for (;;) Line 3422  for (;;)
3422              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3423              if (fi >= max) RRETURN(MATCH_NOMATCH);              if (fi >= max) RRETURN(MATCH_NOMATCH);
3424              if (eptr <= md->end_subject - length &&              if (eptr <= md->end_subject - length &&
3425                memcmp(eptr, charptr, length) == 0) eptr += length;                memcmp(eptr, charptr, IN_UCHARS(length)) == 0) eptr += length;
3426  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
3427              else if (oclength > 0 &&              else if (oclength > 0 &&
3428                       eptr <= md->end_subject - oclength &&                       eptr <= md->end_subject - oclength &&
3429                       memcmp(eptr, occhars, oclength) == 0) eptr += oclength;                       memcmp(eptr, occhars, IN_UCHARS(oclength)) == 0) eptr += oclength;
3430  #endif  /* SUPPORT_UCP */  #endif  /* SUPPORT_UCP */
3431              else              else
3432                {                {
# Line 3232  for (;;) Line 3443  for (;;)
3443            for (i = min; i < max; i++)            for (i = min; i < max; i++)
3444              {              {
3445              if (eptr <= md->end_subject - length &&              if (eptr <= md->end_subject - length &&
3446                  memcmp(eptr, charptr, length) == 0) eptr += length;                  memcmp(eptr, charptr, IN_UCHARS(length)) == 0) eptr += length;
3447  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
3448              else if (oclength > 0 &&              else if (oclength > 0 &&
3449                       eptr <= md->end_subject - oclength &&                       eptr <= md->end_subject - oclength &&
3450                       memcmp(eptr, occhars, oclength) == 0) eptr += oclength;                       memcmp(eptr, occhars, IN_UCHARS(oclength)) == 0) eptr += oclength;
3451  #endif  /* SUPPORT_UCP */  #endif  /* SUPPORT_UCP */
3452              else              else
3453                {                {
# Line 3245  for (;;) Line 3456  for (;;)
3456                }                }
3457              }              }
3458    
3459            if (possessive) continue;            if (possessive) continue;    /* No backtracking */
   
3460            for(;;)            for(;;)
3461              {              {
3462                if (eptr == pp) goto TAIL_RECURSE;
3463              RMATCH(eptr, ecode, offset_top, md, eptrb, RM23);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM23);
3464              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
             if (eptr == pp) { RRETURN(MATCH_NOMATCH); }  
3465  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
3466              eptr--;              eptr--;
3467              BACKCHAR(eptr);              BACKCHAR(eptr);
# Line 3268  for (;;) Line 3478  for (;;)
3478        value of fc will always be < 128. */        value of fc will always be < 128. */
3479        }        }
3480      else      else
3481  #endif  /* SUPPORT_UTF8 */  #endif  /* SUPPORT_UTF */
3482          /* When not in UTF-8 mode, load a single-byte character. */
3483      /* When not in UTF-8 mode, load a single-byte character. */        fc = *ecode++;
   
     fc = *ecode++;  
3484    
3485      /* The value of fc at this point is always less than 256, though we may or      /* The value of fc at this point is always one character, though we may
3486      may not be in UTF-8 mode. The code is duplicated for the caseless and      or may not be in UTF mode. The code is duplicated for the caseless and
3487      caseful cases, for speed, since matching characters is likely to be quite      caseful cases, for speed, since matching characters is likely to be quite
3488      common. First, ensure the minimum number of matches are present. If min =      common. First, ensure the minimum number of matches are present. If min =
3489      max, continue at the same level without recursing. Otherwise, if      max, continue at the same level without recursing. Otherwise, if
# Line 3284  for (;;) Line 3492  for (;;)
3492      maximizing, find the maximum number of characters and work backwards. */      maximizing, find the maximum number of characters and work backwards. */
3493    
3494      DPRINTF(("matching %c{%d,%d} against subject %.*s\n", fc, min, max,      DPRINTF(("matching %c{%d,%d} against subject %.*s\n", fc, min, max,
3495        max, eptr));        max, (char *)eptr));
3496    
3497      if (op >= OP_STARI)  /* Caseless */      if (op >= OP_STARI)  /* Caseless */
3498        {        {
3499        fc = md->lcc[fc];  #ifdef COMPILE_PCRE8
3500          /* fc must be < 128 if UTF is enabled. */
3501          foc = md->fcc[fc];
3502    #else
3503    #ifdef SUPPORT_UTF
3504    #ifdef SUPPORT_UCP
3505          if (utf && fc > 127)
3506            foc = UCD_OTHERCASE(fc);
3507    #else
3508          if (utf && fc > 127)
3509            foc = fc;
3510    #endif /* SUPPORT_UCP */
3511          else
3512    #endif /* SUPPORT_UTF */
3513            foc = TABLE_GET(fc, md->fcc, fc);
3514    #endif /* COMPILE_PCRE8 */
3515    
3516        for (i = 1; i <= min; i++)        for (i = 1; i <= min; i++)
3517          {          {
3518            pcre_uint32 cc;                 /* Faster than pcre_uchar */
3519          if (eptr >= md->end_subject)          if (eptr >= md->end_subject)
3520            {            {
3521            SCHECK_PARTIAL();            SCHECK_PARTIAL();
3522            RRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
3523            }            }
3524          if (fc != md->lcc[*eptr++]) RRETURN(MATCH_NOMATCH);          cc = RAWUCHARTEST(eptr);
3525            if (fc != cc && foc != cc) RRETURN(MATCH_NOMATCH);
3526            eptr++;
3527          }          }
3528        if (min == max) continue;        if (min == max) continue;
3529        if (minimize)        if (minimize)
3530          {          {
3531          for (fi = min;; fi++)          for (fi = min;; fi++)
3532            {            {
3533              pcre_uint32 cc;               /* Faster than pcre_uchar */
3534            RMATCH(eptr, ecode, offset_top, md, eptrb, RM24);            RMATCH(eptr, ecode, offset_top, md, eptrb, RM24);
3535            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3536            if (fi >= max) RRETURN(MATCH_NOMATCH);            if (fi >= max) RRETURN(MATCH_NOMATCH);
# Line 3311  for (;;) Line 3539  for (;;)
3539              SCHECK_PARTIAL();              SCHECK_PARTIAL();
3540              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
3541              }              }
3542            if (fc != md->lcc[*eptr++]) RRETURN(MATCH_NOMATCH);            cc = RAWUCHARTEST(eptr);
3543              if (fc != cc && foc != cc) RRETURN(MATCH_NOMATCH);
3544              eptr++;
3545            }            }
3546          /* Control never gets here */          /* Control never gets here */
3547          }          }
# Line 3320  for (;;) Line 3550  for (;;)
3550          pp = eptr;          pp = eptr;
3551          for (i = min; i < max; i++)          for (i = min; i < max; i++)
3552            {            {
3553              pcre_uint32 cc;               /* Faster than pcre_uchar */
3554            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
3555              {              {
3556              SCHECK_PARTIAL();              SCHECK_PARTIAL();
3557              break;              break;
3558              }              }
3559            if (fc != md->lcc[*eptr]) break;            cc = RAWUCHARTEST(eptr);
3560              if (fc != cc && foc != cc) break;
3561            eptr++;            eptr++;
3562            }            }
3563            if (possessive) continue;       /* No backtracking */
3564          if (possessive) continue;          for (;;)
   
         while (eptr >= pp)  
3565            {            {
3566              if (eptr == pp) goto TAIL_RECURSE;
3567            RMATCH(eptr, ecode, offset_top, md, eptrb, RM25);            RMATCH(eptr, ecode, offset_top, md, eptrb, RM25);
3568            eptr--;            eptr--;
3569            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3570            }            }
3571          RRETURN(MATCH_NOMATCH);          /* Control never gets here */
3572          }          }
       /* Control never gets here */  
3573        }        }
3574    
3575      /* Caseful comparisons (includes all multi-byte characters) */      /* Caseful comparisons (includes all multi-byte characters) */
# Line 3353  for (;;) Line 3583  for (;;)
3583            SCHECK_PARTIAL();            SCHECK_PARTIAL();
3584            RRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
3585            }            }
3586          if (fc != *eptr++) RRETURN(MATCH_NOMATCH);          if (fc != RAWUCHARINCTEST(eptr)) RRETURN(MATCH_NOMATCH);
3587          }          }
3588    
3589        if (min == max) continue;        if (min == max) continue;
# Line 3370  for (;;) Line 3600  for (;;)
3600              SCHECK_PARTIAL();              SCHECK_PARTIAL();
3601              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
3602              }              }
3603            if (fc != *eptr++) RRETURN(MATCH_NOMATCH);            if (fc != RAWUCHARINCTEST(eptr)) RRETURN(MATCH_NOMATCH);
3604            }            }
3605          /* Control never gets here */          /* Control never gets here */
3606          }          }
# Line 3384  for (;;) Line 3614  for (;;)
3614              SCHECK_PARTIAL();              SCHECK_PARTIAL();
3615              break;              break;
3616              }              }
3617            if (fc != *eptr) break;            if (fc != RAWUCHARTEST(eptr)) break;
3618            eptr++;            eptr++;
3619            }            }
3620          if (possessive) continue;          if (possessive) continue;    /* No backtracking */
3621            for (;;)
         while (eptr >= pp)  
3622            {            {
3623              if (eptr == pp) goto TAIL_RECURSE;
3624            RMATCH(eptr, ecode, offset_top, md, eptrb, RM27);            RMATCH(eptr, ecode, offset_top, md, eptrb, RM27);
3625            eptr--;            eptr--;
3626            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3627            }            }
3628          RRETURN(MATCH_NOMATCH);          /* Control never gets here */
3629          }          }
3630        }        }
3631      /* Control never gets here */      /* Control never gets here */
# Line 3410  for (;;) Line 3640  for (;;)
3640        SCHECK_PARTIAL();        SCHECK_PARTIAL();
3641        RRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
3642        }        }
3643      ecode++;  #ifdef SUPPORT_UTF
3644      GETCHARINCTEST(c, eptr);      if (utf)
     if (op == OP_NOTI)         /* The caseless case */  
3645        {        {
3646  #ifdef SUPPORT_UTF8        register pcre_uint32 ch, och;
3647        if (c < 256)  
3648  #endif        ecode++;
3649        c = md->lcc[c];        GETCHARINC(ch, ecode);
3650        if (md->lcc[*ecode++] == c) RRETURN(MATCH_NOMATCH);        GETCHARINC(c, eptr);
3651    
3652          if (op == OP_NOT)
3653            {
3654            if (ch == c) RRETURN(MATCH_NOMATCH);
3655            }
3656          else
3657            {
3658    #ifdef SUPPORT_UCP
3659            if (ch > 127)
3660              och = UCD_OTHERCASE(ch);
3661    #else
3662            if (ch > 127)
3663              och = ch;
3664    #endif /* SUPPORT_UCP */
3665            else
3666              och = TABLE_GET(ch, md->fcc, ch);
3667            if (ch == c || och == c) RRETURN(MATCH_NOMATCH);
3668            }
3669        }        }
3670      else    /* Caseful */      else
3671    #endif
3672        {        {
3673        if (*ecode++ == c) RRETURN(MATCH_NOMATCH);        register pcre_uint32 ch = ecode[1];
3674          c = *eptr++;
3675          if (ch == c || (op == OP_NOTI && TABLE_GET(ch, md->fcc, ch) == c))
3676            RRETURN(MATCH_NOMATCH);
3677          ecode += 2;
3678        }        }
3679      break;      break;
3680    
# Line 3436  for (;;) Line 3688  for (;;)
3688      case OP_NOTEXACT:      case OP_NOTEXACT:
3689      case OP_NOTEXACTI:      case OP_NOTEXACTI:
3690      min = max = GET2(ecode, 1);      min = max = GET2(ecode, 1);
3691      ecode += 3;      ecode += 1 + IMM2_SIZE;
3692      goto REPEATNOTCHAR;      goto REPEATNOTCHAR;
3693    
3694      case OP_NOTUPTO:      case OP_NOTUPTO:
# Line 3446  for (;;) Line 3698  for (;;)
3698      min = 0;      min = 0;
3699      max = GET2(ecode, 1);      max = GET2(ecode, 1);
3700      minimize = *ecode == OP_NOTMINUPTO || *ecode == OP_NOTMINUPTOI;      minimize = *ecode == OP_NOTMINUPTO || *ecode == OP_NOTMINUPTOI;
3701      ecode += 3;      ecode += 1 + IMM2_SIZE;
3702      goto REPEATNOTCHAR;      goto REPEATNOTCHAR;
3703    
3704      case OP_NOTPOSSTAR:      case OP_NOTPOSSTAR:
# Line 3478  for (;;) Line 3730  for (;;)
3730      possessive = TRUE;      possessive = TRUE;
3731      min = 0;      min = 0;
3732      max = GET2(ecode, 1);      max = GET2(ecode, 1);
3733      ecode += 3;      ecode += 1 + IMM2_SIZE;
3734      goto REPEATNOTCHAR;      goto REPEATNOTCHAR;
3735    
3736      case OP_NOTSTAR:      case OP_NOTSTAR:
# Line 3502  for (;;) Line 3754  for (;;)
3754      /* Common code for all repeated single-byte matches. */      /* Common code for all repeated single-byte matches. */
3755    
3756      REPEATNOTCHAR:      REPEATNOTCHAR:
3757      fc = *ecode++;      GETCHARINCTEST(fc, ecode);
3758    
3759      /* The code is duplicated for the caseless and caseful cases, for speed,      /* The code is duplicated for the caseless and caseful cases, for speed,
3760      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 3765  for (;;)
3765      characters and work backwards. */      characters and work backwards. */
3766    
3767      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,
3768        max, eptr));        max, (char *)eptr));
3769    
3770      if (op >= OP_NOTSTARI)     /* Caseless */      if (op >= OP_NOTSTARI)     /* Caseless */
3771        {        {
3772        fc = md->lcc[fc];  #ifdef SUPPORT_UTF
3773    #ifdef SUPPORT_UCP
3774          if (utf && fc > 127)
3775            foc = UCD_OTHERCASE(fc);
3776    #else
3777          if (utf && fc > 127)
3778            foc = fc;
3779    #endif /* SUPPORT_UCP */
3780          else
3781    #endif /* SUPPORT_UTF */
3782            foc = TABLE_GET(fc, md->fcc, fc);
3783    
3784  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
3785        /* UTF-8 mode */        if (utf)
       if (utf8)  
3786          {          {
3787          register unsigned int d;          register pcre_uint32 d;
3788          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
3789            {            {
3790            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
# Line 3532  for (;;) Line 3793  for (;;)
3793              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
3794              }              }
3795            GETCHARINC(d, eptr);            GETCHARINC(d, eptr);
3796            if (d < 256) d = md->lcc[d];            if (fc == d || (unsigned int)foc == d) RRETURN(MATCH_NOMATCH);
           if (fc == d) RRETURN(MATCH_NOMATCH);  
3797            }            }
3798          }          }
3799        else        else
3800  #endif  #endif  /* SUPPORT_UTF */
3801          /* Not UTF mode */
       /* Not UTF-8 mode */  
3802          {          {
3803          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
3804            {            {
# Line 3548  for (;;) Line 3807  for (;;)
3807              SCHECK_PARTIAL();              SCHECK_PARTIAL();
3808              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
3809              }              }
3810            if (fc == md->lcc[*eptr++]) RRETURN(MATCH_NOMATCH);            if (fc == *eptr || foc == *eptr) RRETURN(MATCH_NOMATCH);
3811              eptr++;
3812            }            }
3813          }          }
3814    
# Line 3556  for (;;) Line 3816  for (;;)
3816    
3817        if (minimize)        if (minimize)
3818          {          {
3819  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
3820          /* UTF-8 mode */          if (utf)
         if (utf8)  
3821            {            {
3822            register unsigned int d;            register pcre_uint32 d;
3823            for (fi = min;; fi++)            for (fi = min;; fi++)
3824              {              {
3825              RMATCH(eptr, ecode, offset_top, md, eptrb, RM28);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM28);
# Line 3572  for (;;) Line 3831  for (;;)
3831                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
3832                }                }
3833              GETCHARINC(d, eptr);              GETCHARINC(d, eptr);
3834              if (d < 256) d = md->lcc[d];              if (fc == d || (unsigned int)foc == d) RRETURN(MATCH_NOMATCH);
             if (fc == d) RRETURN(MATCH_NOMATCH);  
3835              }              }
3836            }            }
3837          else          else
3838  #endif  #endif  /*SUPPORT_UTF */
3839          /* Not UTF-8 mode */          /* Not UTF mode */
3840            {            {
3841            for (fi = min;; fi++)            for (fi = min;; fi++)
3842              {              {
# Line 3590  for (;;) Line 3848  for (;;)
3848                SCHECK_PARTIAL();                SCHECK_PARTIAL();
3849                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
3850                }                }
3851              if (fc == md->lcc[*eptr++]) RRETURN(MATCH_NOMATCH);              if (fc == *eptr || foc == *eptr) RRETURN(MATCH_NOMATCH);
3852                eptr++;
3853              }              }
3854            }            }
3855          /* Control never gets here */          /* Control never gets here */
# Line 3602  for (;;) Line 3861  for (;;)
3861          {          {
3862          pp = eptr;          pp = eptr;
3863    
3864  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
3865          /* UTF-8 mode */          if (utf)
         if (utf8)  
3866            {            {
3867            register unsigned int d;            register pcre_uint32 d;
3868            for (i = min; i < max; i++)            for (i = min; i < max; i++)
3869              {              {
3870              int len = 1;              int len = 1;
# Line 3616  for (;;) Line 3874  for (;;)
3874                break;                break;
3875                }                }
3876              GETCHARLEN(d, eptr, len);              GETCHARLEN(d, eptr, len);
3877              if (d < 256) d = md->lcc[d];              if (fc == d || (unsigned int)foc == d) break;
             if (fc == d) break;  
3878              eptr += len;              eptr += len;
3879              }              }
3880          if (possessive) continue;            if (possessive) continue;    /* No backtracking */
3881          for(;;)            for(;;)
3882              {              {
3883                if (eptr == pp) goto TAIL_RECURSE;
3884              RMATCH(eptr, ecode, offset_top, md, eptrb, RM30);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM30);
3885              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3886              if (eptr-- == pp) break;        /* Stop if tried at original pos */              eptr--;
3887              BACKCHAR(eptr);              BACKCHAR(eptr);
3888              }              }
3889            }            }
3890          else          else
3891  #endif  #endif  /* SUPPORT_UTF */
3892          /* Not UTF-8 mode */          /* Not UTF mode */
3893            {            {
3894            for (i = min; i < max; i++)            for (i = min; i < max; i++)
3895              {              {
# Line 3640  for (;;) Line 3898  for (;;)
3898                SCHECK_PARTIAL();                SCHECK_PARTIAL();
3899                break;                break;
3900                }                }
3901              if (fc == md->lcc[*eptr]) break;              if (fc == *eptr || foc == *eptr) break;
3902              eptr++;              eptr++;
3903              }              }
3904            if (possessive) continue;            if (possessive) continue;    /* No backtracking */
3905            while (eptr >= pp)            for (;;)
3906              {              {
3907                if (eptr == pp) goto TAIL_RECURSE;
3908              RMATCH(eptr, ecode, offset_top, md, eptrb, RM31);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM31);
3909              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3910              eptr--;              eptr--;
3911              }              }
3912            }            }
3913            /* Control never gets here */
         RRETURN(MATCH_NOMATCH);  
3914          }          }
       /* Control never gets here */  
3915        }        }
3916    
3917      /* Caseful comparisons */      /* Caseful comparisons */
3918    
3919      else      else
3920        {        {
3921  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
3922        /* UTF-8 mode */        if (utf)
       if (utf8)  
3923          {          {
3924          register unsigned int d;          register pcre_uint32 d;
3925          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
3926            {            {
3927            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
# Line 3679  for (;;) Line 3935  for (;;)
3935          }          }
3936        else        else
3937  #endif  #endif
3938        /* Not UTF-8 mode */        /* Not UTF mode */
3939          {          {
3940          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
3941            {            {
# Line 3696  for (;;) Line 3952  for (;;)
3952    
3953        if (minimize)        if (minimize)
3954          {          {
3955  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
3956          /* UTF-8 mode */          if (utf)
         if (utf8)  
3957            {            {
3958            register unsigned int d;            register pcre_uint32 d;
3959            for (fi = min;; fi++)            for (fi = min;; fi++)
3960              {              {
3961              RMATCH(eptr, ecode, offset_top, md, eptrb, RM32);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM32);
# Line 3717  for (;;) Line 3972  for (;;)
3972            }            }
3973          else          else
3974  #endif  #endif
3975          /* Not UTF-8 mode */          /* Not UTF mode */
3976            {            {
3977            for (fi = min;; fi++)            for (fi = min;; fi++)
3978              {              {
# Line 3741  for (;;) Line 3996  for (;;)
3996          {          {
3997          pp = eptr;          pp = eptr;
3998    
3999  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
4000          /* UTF-8 mode */          if (utf)
         if (utf8)  
4001            {            {
4002            register unsigned int d;            register pcre_uint32 d;
4003            for (i = min; i < max; i++)            for (i = min; i < max; i++)
4004              {              {
4005              int len = 1;              int len = 1;
# Line 3758  for (;;) Line 4012  for (;;)
4012              if (fc == d) break;              if (fc == d) break;
4013              eptr += len;              eptr += len;
4014              }              }
4015            if (possessive) continue;            if (possessive) continue;    /* No backtracking */
4016            for(;;)            for(;;)
4017              {              {
4018                if (eptr == pp) goto TAIL_RECURSE;
4019              RMATCH(eptr, ecode, offset_top, md, eptrb, RM34);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM34);
4020              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
4021              if (eptr-- == pp) break;        /* Stop if tried at original pos */              eptr--;
4022              BACKCHAR(eptr);              BACKCHAR(eptr);
4023              }              }
4024            }            }
4025          else          else
4026  #endif  #endif
4027          /* Not UTF-8 mode */          /* Not UTF mode */
4028            {            {
4029            for (i = min; i < max; i++)            for (i = min; i < max; i++)
4030              {              {
# Line 3781  for (;;) Line 4036  for (;;)
4036              if (fc == *eptr) break;              if (fc == *eptr) break;
4037              eptr++;              eptr++;
4038              }              }
4039            if (possessive) continue;            if (possessive) continue;    /* No backtracking */
4040            while (eptr >= pp)            for (;;)
4041              {              {
4042                if (eptr == pp) goto TAIL_RECURSE;
4043              RMATCH(eptr, ecode, offset_top, md, eptrb, RM35);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM35);
4044              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
4045              eptr--;              eptr--;
4046              }              }
4047            }            }
4048            /* Control never gets here */
         RRETURN(MATCH_NOMATCH);  
4049          }          }
4050        }        }
4051      /* Control never gets here */      /* Control never gets here */
# Line 3802  for (;;) Line 4057  for (;;)
4057      case OP_TYPEEXACT:      case OP_TYPEEXACT:
4058      min = max = GET2(ecode, 1);      min = max = GET2(ecode, 1);
4059      minimize = TRUE;      minimize = TRUE;
4060      ecode += 3;      ecode += 1 + IMM2_SIZE;
4061      goto REPEATTYPE;      goto REPEATTYPE;
4062    
4063      case OP_TYPEUPTO:      case OP_TYPEUPTO:
# Line 3810  for (;;) Line 4065  for (;;)
4065      min = 0;      min = 0;
4066      max = GET2(ecode, 1);      max = GET2(ecode, 1);
4067      minimize = *ecode == OP_TYPEMINUPTO;      minimize = *ecode == OP_TYPEMINUPTO;
4068      ecode += 3;      ecode += 1 + IMM2_SIZE;
4069      goto REPEATTYPE;      goto REPEATTYPE;
4070    
4071      case OP_TYPEPOSSTAR:      case OP_TYPEPOSSTAR:
# Line 3838  for (;;) Line 4093  for (;;)
4093      possessive = TRUE;      possessive = TRUE;
4094      min = 0;      min = 0;
4095      max = GET2(ecode, 1);      max = GET2(ecode, 1);
4096      ecode += 3;      ecode += 1 + IMM2_SIZE;
4097      goto REPEATTYPE;      goto REPEATTYPE;
4098    
4099      case OP_TYPESTAR:      case OP_TYPESTAR:
# Line 3972  for (;;) Line 4227  for (;;)
4227              }              }
4228            break;            break;
4229    
4230              /* Perl space used to exclude VT, but from Perl 5.18 it is included,
4231              which means that Perl space and POSIX space are now identical. PCRE
4232              was changed at release 8.34. */
4233    
4234            case PT_SPACE:    /* Perl space */            case PT_SPACE:    /* Perl space */
4235              case PT_PXSPACE:  /* POSIX space */
4236            for (i = 1; i <= min; i++)            for (i = 1; i <= min; i++)
4237              {              {
4238              if (eptr >= md->end_subject)              if (eptr >= md->end_subject)
# Line 3981  for (;;) Line 4241  for (;;)
4241                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
4242                }                }
4243              GETCHARINCTEST(c, eptr);              GETCHARINCTEST(c, eptr);
4244              if ((UCD_CATEGORY(c) == ucp_Z || c == CHAR_HT || c == CHAR_NL ||              switch(c)
4245                   c == CHAR_FF || c == CHAR_CR)                {
4246                     == prop_fail_result)                HSPACE_CASES:
4247                RRETURN(MATCH_NOMATCH);                VSPACE_CASES:
4248                  if (prop_fail_result) RRETURN(MATCH_NOMATCH);
4249                  break;
4250    
4251                  default:
4252                  if ((UCD_CATEGORY(c) == ucp_Z) == prop_fail_result)
4253                    RRETURN(MATCH_NOMATCH);
4254                  break;
4255                  }
4256              }              }
4257            break;            break;
4258    
4259            case PT_PXSPACE:  /* POSIX space */            case PT_WORD:
4260            for (i = 1; i <= min; i++)            for (i = 1; i <= min; i++)
4261              {              {
4262                int category;
4263              if (eptr >= md->end_subject)              if (eptr >= md->end_subject)
4264                {                {
4265                SCHECK_PARTIAL();                SCHECK_PARTIAL();
4266                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
4267                }                }
4268              GETCHARINCTEST(c, eptr);              GETCHARINCTEST(c, eptr);
4269              if ((UCD_CATEGORY(c) == ucp_Z || c == CHAR_HT || c == CHAR_NL ||              category = UCD_CATEGORY(c);
4270                   c == CHAR_VT || c == CHAR_FF || c == CHAR_CR)              if ((category == ucp_L || category == ucp_N || c == CHAR_UNDERSCORE)
4271                     == prop_fail_result)                     == prop_fail_result)
4272                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
4273              }              }
4274            break;            break;
4275    
4276            case PT_WORD:            case PT_CLIST:
4277            for (i = 1; i <= min; i++)            for (i = 1; i <= min; i++)
4278              {              {
4279              int category;              const pcre_uint32 *cp;
4280              if (eptr >= md->end_subject)              if (eptr >= md->end_subject)
4281                {                {
4282                SCHECK_PARTIAL();                SCHECK_PARTIAL();
4283                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
4284                }                }
4285              GETCHARINCTEST(c, eptr);              GETCHARINCTEST(c, eptr);
4286              category = UCD_CATEGORY(c);              cp = PRIV(ucd_caseless_sets) + prop_value;
4287              if ((category == ucp_L || category == ucp_N || c == CHAR_UNDERSCORE)              for (;;)
4288                     == prop_fail_result)                {
4289                  if (c < *cp)
4290                    { if (prop_fail_result) break; else { RRETURN(MATCH_NOMATCH); } }
4291                  if (c == *cp++)
4292                    { if (prop_fail_result) { RRETURN(MATCH_NOMATCH); } else break; }
4293                  }
4294                }
4295              break;
4296    
4297              case PT_UCNC:
4298              for (i = 1; i <= min; i++)
4299                {
4300                if (eptr >= md->end_subject)
4301                  {
4302                  SCHECK_PARTIAL();
4303                  RRETURN(MATCH_NOMATCH);
4304                  }
4305                GETCHARINCTEST(c, eptr);
4306                if ((c == CHAR_DOLLAR_SIGN || c == CHAR_COMMERCIAL_AT ||
4307                     c == CHAR_GRAVE_ACCENT || (c >= 0xa0 && c <= 0xd7ff) ||
4308                     c >= 0xe000) == prop_fail_result)
4309                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
4310              }              }
4311            break;            break;
# Line 4040  for (;;) Line 4329  for (;;)
4329              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4330              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4331              }              }
4332            GETCHARINCTEST(c, eptr);            else
           if (UCD_CATEGORY(c) == ucp_M) RRETURN(MATCH_NOMATCH);  
           while (eptr < md->end_subject)  
4333              {              {
4334              int len = 1;              int lgb, rgb;
4335              if (!utf8) c = *eptr; else { GETCHARLEN(c, eptr, len); }              GETCHARINCTEST(c, eptr);
4336              if (UCD_CATEGORY(c) != ucp_M) break;              lgb = UCD_GRAPHBREAK(c);
4337              eptr += len;             while (eptr < md->end_subject)
4338                  {
4339                  int len = 1;
4340                  if (!utf) c = *eptr; else { GETCHARLEN(c, eptr, len); }
4341                  rgb = UCD_GRAPHBREAK(c);
4342                  if ((PRIV(ucp_gbtable)[lgb] & (1 << rgb)) == 0) break;
4343                  lgb = rgb;
4344                  eptr += len;
4345                  }
4346              }              }
4347              CHECK_PARTIAL();
4348            }            }
4349          }          }
4350    
# Line 4057  for (;;) Line 4353  for (;;)
4353    
4354  /* Handle all other cases when the coding is UTF-8 */  /* Handle all other cases when the coding is UTF-8 */
4355    
4356  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
4357        if (utf8) switch(ctype)        if (utf) switch(ctype)
4358          {          {
4359          case OP_ANY:          case OP_ANY:
4360          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
# Line 4069  for (;;) Line 4365  for (;;)
4365              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4366              }              }
4367            if (IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH);            if (IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH);
4368              if (md->partial != 0 &&
4369                  eptr + 1 >= md->end_subject &&
4370                  NLBLOCK->nltype == NLTYPE_FIXED &&
4371                  NLBLOCK->nllen == 2 &&
4372                  RAWUCHAR(eptr) == NLBLOCK->nl[0])
4373                {
4374                md->hitend = TRUE;
4375                if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
4376                }
4377            eptr++;            eptr++;
4378            while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;            ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);
4379            }            }
4380          break;          break;
4381    
# Line 4083  for (;;) Line 4388  for (;;)
4388              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4389              }              }
4390            eptr++;            eptr++;
4391            while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;            ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);
4392            }            }
4393          break;          break;
4394    
# Line 4105  for (;;) Line 4410  for (;;)
4410              {              {
4411              default: RRETURN(MATCH_NOMATCH);              default: RRETURN(MATCH_NOMATCH);
4412    
4413              case 0x000d:              case CHAR_CR:
4414              if (eptr < md->end_subject && *eptr == 0x0a) eptr++;              if (eptr < md->end_subject && RAWUCHAR(eptr) == CHAR_LF) eptr++;
4415              break;              break;
4416    
4417              case 0x000a:              case CHAR_LF:
4418              break;              break;
4419    
4420              case 0x000b:              case CHAR_VT:
4421              case 0x000c:              case CHAR_FF:
4422              case 0x0085:              case CHAR_NEL:
4423    #ifndef EBCDIC
4424              case 0x2028:              case 0x2028:
4425              case 0x2029:              case 0x2029:
4426    #endif  /* Not EBCDIC */
4427              if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);              if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);
4428              break;              break;
4429              }              }
# Line 4134  for (;;) Line 4441  for (;;)
4441            GETCHARINC(c, eptr);            GETCHARINC(c, eptr);
4442            switch(c)            switch(c)
4443              {              {
4444                HSPACE_CASES: RRETURN(MATCH_NOMATCH);  /* Byte and multibyte cases */
4445              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);  
4446              }              }
4447            }            }
4448          break;          break;
# Line 4170  for (;;) Line 4458  for (;;)
4458            GETCHARINC(c, eptr);            GETCHARINC(c, eptr);
4459            switch(c)            switch(c)
4460              {              {
4461                HSPACE_CASES: break;  /* Byte and multibyte cases */
4462              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;  
4463              }              }
4464            }            }
4465          break;          break;
# Line 4206  for (;;) Line 4475  for (;;)
4475            GETCHARINC(c, eptr);            GETCHARINC(c, eptr);
4476            switch(c)            switch(c)
4477              {              {
4478                VSPACE_CASES: RRETURN(MATCH_NOMATCH);
4479              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);  
4480              }              }
4481            }            }
4482          break;          break;
# Line 4230  for (;;) Line 4492  for (;;)
4492            GETCHARINC(c, eptr);            GETCHARINC(c, eptr);
4493            switch(c)            switch(c)
4494              {              {
4495                VSPACE_CASES: break;
4496              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;  
4497              }              }
4498            }            }
4499          break;          break;
# Line 4260  for (;;) Line 4515  for (;;)
4515          case OP_DIGIT:          case OP_DIGIT:
4516          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
4517            {            {
4518              pcre_uint32 cc;
4519            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
4520              {              {
4521              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4522              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4523              }              }
4524            if (*eptr >= 128 || (md->ctypes[*eptr++] & ctype_digit) == 0)            cc = RAWUCHAR(eptr);
4525              if (cc >= 128 || (md->ctypes[cc] & ctype_digit) == 0)
4526              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4527              eptr++;
4528            /* No need to skip more bytes - we know it's a 1-byte character */            /* No need to skip more bytes - we know it's a 1-byte character */
4529            }            }
4530          break;          break;
# Line 4274  for (;;) Line 4532  for (;;)
4532          case OP_NOT_WHITESPACE:          case OP_NOT_WHITESPACE:
4533          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
4534            {            {
4535              pcre_uint32 cc;
4536            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
4537              {              {
4538              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4539              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4540              }              }
4541            if (*eptr < 128 && (md->ctypes[*eptr] & ctype_space) != 0)            cc = RAWUCHAR(eptr);
4542              if (cc < 128 && (md->ctypes[cc] & ctype_space) != 0)
4543              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4544            while (++eptr < md->end_subject && (*eptr & 0xc0) == 0x80);            eptr++;
4545              ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);
4546            }            }
4547          break;          break;
4548    
4549          case OP_WHITESPACE:          case OP_WHITESPACE:
4550          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
4551            {            {
4552              pcre_uint32 cc;
4553            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
4554              {              {
4555              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4556              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4557              }              }
4558            if (*eptr >= 128 || (md->ctypes[*eptr++] & ctype_space) == 0)            cc = RAWUCHAR(eptr);
4559              if (cc >= 128 || (md->ctypes[cc] & ctype_space) == 0)
4560              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4561              eptr++;
4562            /* No need to skip more bytes - we know it's a 1-byte character */            /* No need to skip more bytes - we know it's a 1-byte character */
4563            }            }
4564          break;          break;
# Line 4302  for (;;) Line 4566  for (;;)
4566          case OP_NOT_WORDCHAR:          case OP_NOT_WORDCHAR:
4567          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
4568            {            {
4569              pcre_uint32 cc;
4570            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
4571              {              {
4572              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4573              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4574              }              }
4575            if (*eptr < 128 && (md->ctypes[*eptr] & ctype_word) != 0)            cc = RAWUCHAR(eptr);
4576              if (cc < 128 && (md->ctypes[cc] & ctype_word) != 0)
4577              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4578            while (++eptr < md->end_subject && (*eptr & 0xc0) == 0x80);            eptr++;
4579              ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);
4580            }            }
4581          break;          break;
4582    
4583          case OP_WORDCHAR:          case OP_WORDCHAR:
4584          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
4585            {            {
4586              pcre_uint32 cc;
4587            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
4588              {              {
4589              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4590              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4591              }              }
4592            if (*eptr >= 128 || (md->ctypes[*eptr++] & ctype_word) == 0)            cc = RAWUCHAR(eptr);
4593              if (cc >= 128 || (md->ctypes[cc] & ctype_word) == 0)
4594              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4595              eptr++;
4596            /* 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 */
4597            }            }
4598          break;          break;
# Line 4332  for (;;) Line 4602  for (;;)
4602          }  /* End switch(ctype) */          }  /* End switch(ctype) */
4603    
4604        else        else
4605  #endif     /* SUPPORT_UTF8 */  #endif     /* SUPPORT_UTF */
4606    
4607        /* 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
4608        than OP_PROP and OP_NOTPROP. */        than OP_PROP and OP_NOTPROP. */
# Line 4348  for (;;) Line 4618  for (;;)
4618              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4619              }              }
4620            if (IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH);            if (IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH);
4621              if (md->partial != 0 &&
4622                  eptr + 1 >= md->end_subject &&
4623                  NLBLOCK->nltype == NLTYPE_FIXED &&
4624                  NLBLOCK->nllen == 2 &&
4625                  *eptr == NLBLOCK->nl[0])
4626                {
4627                md->hitend = TRUE;
4628                if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
4629                }
4630            eptr++;            eptr++;
4631            }            }
4632          break;          break;
# Line 4382  for (;;) Line 4661  for (;;)
4661              {              {
4662              default: RRETURN(MATCH_NOMATCH);              default: RRETURN(MATCH_NOMATCH);
4663    
4664              case 0x000d:              case CHAR_CR:
4665              if (eptr < md->end_subject && *eptr == 0x0a) eptr++;              if (eptr < md->end_subject && *eptr == CHAR_LF) eptr++;
4666              break;              break;
4667    
4668              case 0x000a:              case CHAR_LF:
4669              break;              break;
4670    
4671              case 0x000b:              case CHAR_VT:
4672              case 0x000c:              case CHAR_FF:
4673              case 0x0085:              case CHAR_NEL:
4674    #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
4675                case 0x2028:
4676                case 0x2029:
4677    #endif
4678              if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);              if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);
4679              break;              break;
4680              }              }
# Line 4409  for (;;) Line 4692  for (;;)
4692            switch(*eptr++)            switch(*eptr++)
4693              {              {
4694              default: break;              default: break;
4695              case 0x09:      /* HT */              HSPACE_BYTE_CASES:
4696              case 0x20:      /* SPACE */  #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
4697              case 0xa0:      /* NBSP */              HSPACE_MULTIBYTE_CASES:
4698    #endif
4699              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4700              }              }
4701            }            }
# Line 4428  for (;;) Line 4712  for (;;)
4712            switch(*eptr++)            switch(*eptr++)
4713              {              {
4714              default: RRETURN(MATCH_NOMATCH);              default: RRETURN(MATCH_NOMATCH);
4715              case 0x09:      /* HT */              HSPACE_BYTE_CASES:
4716              case 0x20:      /* SPACE */  #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
4717              case 0xa0:      /* NBSP */              HSPACE_MULTIBYTE_CASES:
4718    #endif
4719              break;              break;
4720              }              }
4721            }            }
# Line 4446  for (;;) Line 4731  for (;;)
4731              }              }
4732            switch(*eptr++)            switch(*eptr++)
4733              {              {
4734              default: break;              VSPACE_BYTE_CASES:
4735              case 0x0a:      /* LF */  #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
4736              case 0x0b:      /* VT */              VSPACE_MULTIBYTE_CASES:
4737              case 0x0c:      /* FF */  #endif
             case 0x0d:      /* CR */  
             case 0x85:      /* NEL */  
4738              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4739                default: break;
4740              }              }
4741            }            }
4742          break;          break;
# Line 4468  for (;;) Line 4752  for (;;)
4752            switch(*eptr++)            switch(*eptr++)
4753              {              {
4754              default: RRETURN(MATCH_NOMATCH);              default: RRETURN(MATCH_NOMATCH);
4755              case 0x0a:      /* LF */              VSPACE_BYTE_CASES:
4756              case 0x0b:      /* VT */  #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
4757              case 0x0c:      /* FF */              VSPACE_MULTIBYTE_CASES:
4758              case 0x0d:      /* CR */  #endif
             case 0x85:      /* NEL */  
4759              break;              break;
4760              }              }
4761            }            }
# Line 4486  for (;;) Line 4769  for (;;)
4769              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4770              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4771              }              }
4772            if ((md->ctypes[*eptr++] & ctype_digit) != 0) RRETURN(MATCH_NOMATCH);            if (MAX_255(*eptr) && (md->ctypes[*eptr] & ctype_digit) != 0)
4773                RRETURN(MATCH_NOMATCH);
4774              eptr++;
4775            }            }
4776          break;          break;
4777    
# Line 4498  for (;;) Line 4783  for (;;)
4783              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4784              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4785              }              }
4786            if ((md->ctypes[*eptr++] & ctype_digit) == 0) RRETURN(MATCH_NOMATCH);            if (!MAX_255(*eptr) || (md->ctypes[*eptr] & ctype_digit) == 0)
4787                RRETURN(MATCH_NOMATCH);
4788              eptr++;
4789            }            }
4790          break;          break;
4791    
# Line 4510  for (;;) Line 4797  for (;;)
4797              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4798              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4799              }              }
4800            if ((md->ctypes[*eptr++] & ctype_space) != 0) RRETURN(MATCH_NOMATCH);            if (MAX_255(*eptr) && (md->ctypes[*eptr] & ctype_space) != 0)
4801                RRETURN(MATCH_NOMATCH);
4802              eptr++;
4803            }            }
4804          break;          break;
4805    
# Line 4522  for (;;) Line 4811  for (;;)
4811              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4812              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4813              }              }
4814            if ((md->ctypes[*eptr++] & ctype_space) == 0) RRETURN(MATCH_NOMATCH);            if (!MAX_255(*eptr) || (md->ctypes[*eptr] & ctype_space) == 0)
4815                RRETURN(MATCH_NOMATCH);
4816              eptr++;
4817            }            }
4818          break;          break;
4819    
# Line 4534  for (;;) Line 4825  for (;;)
4825              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4826              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4827              }              }
4828            if ((md->ctypes[*eptr++] & ctype_word) != 0)            if (MAX_255(*eptr) && (md->ctypes[*eptr] & ctype_word) != 0)
4829              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4830              eptr++;
4831            }            }
4832          break;          break;
4833    
# Line 4547  for (;;) Line 4839  for (;;)
4839              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4840              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4841              }              }
4842            if ((md->ctypes[*eptr++] & ctype_word) == 0)            if (!MAX_255(*eptr) || (md->ctypes[*eptr] & ctype_word) == 0)
4843              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4844              eptr++;
4845            }            }
4846          break;          break;
4847    
# Line 4679  for (;;) Line 4972  for (;;)
4972              }              }
4973            /* Control never gets here */            /* Control never gets here */
4974    
4975              /* Perl space used to exclude VT, but from Perl 5.18 it is included,
4976              which means that Perl space and POSIX space are now identical. PCRE
4977              was changed at release 8.34. */
4978    
4979            case PT_SPACE:    /* Perl space */            case PT_SPACE:    /* Perl space */
4980              case PT_PXSPACE:  /* POSIX space */
4981            for (fi = min;; fi++)            for (fi = min;; fi++)
4982              {              {
4983              RMATCH(eptr, ecode, offset_top, md, eptrb, RM60);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM61);
4984              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
4985              if (fi >= max) RRETURN(MATCH_NOMATCH);              if (fi >= max) RRETURN(MATCH_NOMATCH);
4986              if (eptr >= md->end_subject)              if (eptr >= md->end_subject)
# Line 4691  for (;;) Line 4989  for (;;)
4989                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
4990                }                }
4991              GETCHARINCTEST(c, eptr);              GETCHARINCTEST(c, eptr);
4992              if ((UCD_CATEGORY(c) == ucp_Z || c == CHAR_HT || c == CHAR_NL ||              switch(c)
4993                   c == CHAR_FF || c == CHAR_CR)                {
4994                     == prop_fail_result)                HSPACE_CASES:
4995                RRETURN(MATCH_NOMATCH);                VSPACE_CASES:
4996                  if (prop_fail_result) RRETURN(MATCH_NOMATCH);
4997                  break;
4998    
4999                  default:
5000                  if ((UCD_CATEGORY(c) == ucp_Z) == prop_fail_result)
5001                    RRETURN(MATCH_NOMATCH);
5002                  break;
5003                  }
5004              }              }
5005            /* Control never gets here */            /* Control never gets here */
5006    
5007            case PT_PXSPACE:  /* POSIX space */            case PT_WORD:
5008            for (fi = min;; fi++)            for (fi = min;; fi++)
5009              {              {
5010              RMATCH(eptr, ecode, offset_top, md, eptrb, RM61);              int category;
5011                RMATCH(eptr, ecode, offset_top, md, eptrb, RM62);
5012              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
5013              if (fi >= max) RRETURN(MATCH_NOMATCH);              if (fi >= max) RRETURN(MATCH_NOMATCH);
5014              if (eptr >= md->end_subject)              if (eptr >= md->end_subject)
# Line 4710  for (;;) Line 5017  for (;;)
5017                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
5018                }                }
5019              GETCHARINCTEST(c, eptr);              GETCHARINCTEST(c, eptr);
5020              if ((UCD_CATEGORY(c) == ucp_Z || c == CHAR_HT || c == CHAR_NL ||              category = UCD_CATEGORY(c);
5021                   c == CHAR_VT || c == CHAR_FF || c == CHAR_CR)              if ((category == ucp_L ||
5022                     category == ucp_N ||
5023                     c == CHAR_UNDERSCORE)
5024                     == prop_fail_result)                     == prop_fail_result)
5025                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
5026              }              }