/[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 728 by ph10, Mon Oct 10 16:01:03 2011 UTC revision 1033 by ph10, Mon Sep 10 11:02:48 2012 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) 1997-2012 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 82  negative to avoid the external error cod Line 81  negative to avoid the external error cod
81  #define MATCH_SKIP_ARG     (-993)  #define MATCH_SKIP_ARG     (-993)
82  #define MATCH_THEN         (-992)  #define MATCH_THEN         (-992)
83    
 /* This is a convenience macro for code that occurs many times. */  
   
 #define MRRETURN(ra) \  
   { \  
   md->mark = markptr; \  
   RRETURN(ra); \  
   }  
   
84  /* 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.
85  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,
86  because the offset vector is always a multiple of 3 long. */  because the offset vector is always a multiple of 3 long. */
# Line 121  Returns:     nothing Line 112  Returns:     nothing
112  */  */
113    
114  static void  static void
115  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)
116  {  {
117  unsigned int c;  unsigned int c;
118  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;
# Line 148  Arguments: Line 139  Arguments:
139    md          points to match data block    md          points to match data block
140    caseless    TRUE if caseless    caseless    TRUE if caseless
141    
142  Returns:      < 0 if not matched, otherwise the number of subject bytes matched  Returns:      >= 0 the number of subject bytes matched
143                  -1 no match
144                  -2 partial match; always given if at end subject
145  */  */
146    
147  static int  static int
148  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,
149    BOOL caseless)    BOOL caseless)
150  {  {
151  USPTR eptr_start = eptr;  PCRE_PUCHAR eptr_start = eptr;
152  register USPTR p = md->start_subject + md->offset_vector[offset];  register PCRE_PUCHAR p = md->start_subject + md->offset_vector[offset];
153    
154  #ifdef PCRE_DEBUG  #ifdef PCRE_DEBUG
155  if (eptr >= md->end_subject)  if (eptr >= md->end_subject)
# Line 171  pchars(p, length, FALSE, md); Line 164  pchars(p, length, FALSE, md);
164  printf("\n");  printf("\n");
165  #endif  #endif
166    
167  /* Always fail if reference not set (and not JavaScript compatible). */  /* Always fail if reference not set (and not JavaScript compatible - in that
168    case the length is passed as zero). */
169    
170  if (length < 0) return -1;  if (length < 0) return -1;
171    
# Line 181  ASCII characters. */ Line 175  ASCII characters. */
175    
176  if (caseless)  if (caseless)
177    {    {
178  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
179  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
180    if (md->utf8)    if (md->utf)
181      {      {
182      /* 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
183      bytes matched may differ, because there are some characters whose upper and      bytes matched may differ, because there are some characters whose upper and
# Line 193  if (caseless) Line 187  if (caseless)
187      the latter. It is important, therefore, to check the length along the      the latter. It is important, therefore, to check the length along the
188      reference, not along the subject (earlier code did this wrong). */      reference, not along the subject (earlier code did this wrong). */
189    
190      USPTR endptr = p + length;      PCRE_PUCHAR endptr = p + length;
191      while (p < endptr)      while (p < endptr)
192        {        {
193        int c, d;        int c, d;
194        if (eptr >= md->end_subject) return -1;        if (eptr >= md->end_subject) return -2;   /* Partial match */
195        GETCHARINC(c, eptr);        GETCHARINC(c, eptr);
196        GETCHARINC(d, p);        GETCHARINC(d, p);
197        if (c != d && c != UCD_OTHERCASE(d)) return -1;        if (c != d && c != UCD_OTHERCASE(d)) return -1;
# Line 210  if (caseless) Line 204  if (caseless)
204    /* 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
205    is no UCP support. */    is no UCP support. */
206      {      {
     if (eptr + length > md->end_subject) return -1;  
207      while (length-- > 0)      while (length-- > 0)
208        { if (md->lcc[*p++] != md->lcc[*eptr++]) return -1; }        {
209          if (eptr >= md->end_subject) return -2;   /* Partial match */
210          if (TABLE_GET(*p, md->lcc, *p) != TABLE_GET(*eptr, md->lcc, *eptr)) return -1;
211          p++;
212          eptr++;
213          }
214      }      }
215    }    }
216    
# Line 221  are in UTF-8 mode. */ Line 219  are in UTF-8 mode. */
219    
220  else  else
221    {    {
222    if (eptr + length > md->end_subject) return -1;    while (length-- > 0)
223    while (length-- > 0) if (*p++ != *eptr++) return -1;      {
224        if (eptr >= md->end_subject) return -2;   /* Partial match */
225        if (*p++ != *eptr++) return -1;
226        }
227    }    }
228    
229  return eptr - eptr_start;  return (int)(eptr - eptr_start);
230  }  }
231    
232    
# Line 290  actually used in this definition. */ Line 291  actually used in this definition. */
291  #define RMATCH(ra,rb,rc,rd,re,rw) \  #define RMATCH(ra,rb,rc,rd,re,rw) \
292    { \    { \
293    printf("match() called in line %d\n", __LINE__); \    printf("match() called in line %d\n", __LINE__); \
294    rrc = match(ra,rb,mstart,markptr,rc,rd,re,rdepth+1); \    rrc = match(ra,rb,mstart,rc,rd,re,rdepth+1); \
295    printf("to line %d\n", __LINE__); \    printf("to line %d\n", __LINE__); \
296    }    }
297  #define RRETURN(ra) \  #define RRETURN(ra) \
# Line 300  actually used in this definition. */ Line 301  actually used in this definition. */
301    }    }
302  #else  #else
303  #define RMATCH(ra,rb,rc,rd,re,rw) \  #define RMATCH(ra,rb,rc,rd,re,rw) \
304    rrc = match(ra,rb,mstart,markptr,rc,rd,re,rdepth+1)    rrc = match(ra,rb,mstart,rc,rd,re,rdepth+1)
305  #define RRETURN(ra) return ra  #define RRETURN(ra) return ra
306  #endif  #endif
307    
# Line 315  argument of match(), which never changes Line 316  argument of match(), which never changes
316    
317  #define RMATCH(ra,rb,rc,rd,re,rw)\  #define RMATCH(ra,rb,rc,rd,re,rw)\
318    {\    {\
319    heapframe *newframe = (heapframe *)(pcre_stack_malloc)(sizeof(heapframe));\    heapframe *newframe = frame->Xnextframe;\
320    if (newframe == NULL) RRETURN(PCRE_ERROR_NOMEMORY);\    if (newframe == NULL)\
321    frame->Xwhere = rw; \      {\
322        newframe = (heapframe *)(PUBL(stack_malloc))(sizeof(heapframe));\
323        if (newframe == NULL) RRETURN(PCRE_ERROR_NOMEMORY);\
324        newframe->Xnextframe = NULL;\
325        frame->Xnextframe = newframe;\
326        }\
327      frame->Xwhere = rw;\
328    newframe->Xeptr = ra;\    newframe->Xeptr = ra;\
329    newframe->Xecode = rb;\    newframe->Xecode = rb;\
330    newframe->Xmstart = mstart;\    newframe->Xmstart = mstart;\
   newframe->Xmarkptr = markptr;\  
331    newframe->Xoffset_top = rc;\    newframe->Xoffset_top = rc;\
332    newframe->Xeptrb = re;\    newframe->Xeptrb = re;\
333    newframe->Xrdepth = frame->Xrdepth + 1;\    newframe->Xrdepth = frame->Xrdepth + 1;\
# Line 337  argument of match(), which never changes Line 343  argument of match(), which never changes
343    {\    {\
344    heapframe *oldframe = frame;\    heapframe *oldframe = frame;\
345    frame = oldframe->Xprevframe;\    frame = oldframe->Xprevframe;\
   (pcre_stack_free)(oldframe);\  
346    if (frame != NULL)\    if (frame != NULL)\
347      {\      {\
348      rrc = ra;\      rrc = ra;\
# Line 351  argument of match(), which never changes Line 356  argument of match(), which never changes
356    
357  typedef struct heapframe {  typedef struct heapframe {
358    struct heapframe *Xprevframe;    struct heapframe *Xprevframe;
359      struct heapframe *Xnextframe;
360    
361    /* Function arguments that may change */    /* Function arguments that may change */
362    
363    USPTR Xeptr;    PCRE_PUCHAR Xeptr;
364    const uschar *Xecode;    const pcre_uchar *Xecode;
365    USPTR Xmstart;    PCRE_PUCHAR Xmstart;
   USPTR Xmarkptr;  
366    int Xoffset_top;    int Xoffset_top;
367    eptrblock *Xeptrb;    eptrblock *Xeptrb;
368    unsigned int Xrdepth;    unsigned int Xrdepth;
369    
370    /* Function local variables */    /* Function local variables */
371    
372    USPTR Xcallpat;    PCRE_PUCHAR Xcallpat;
373  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
374    USPTR Xcharptr;    PCRE_PUCHAR Xcharptr;
375  #endif  #endif
376    USPTR Xdata;    PCRE_PUCHAR Xdata;
377    USPTR Xnext;    PCRE_PUCHAR Xnext;
378    USPTR Xpp;    PCRE_PUCHAR Xpp;
379    USPTR Xprev;    PCRE_PUCHAR Xprev;
380    USPTR Xsaved_eptr;    PCRE_PUCHAR Xsaved_eptr;
381    
382    recursion_info Xnew_recursive;    recursion_info Xnew_recursive;
383    
# Line 385  typedef struct heapframe { Line 390  typedef struct heapframe {
390    int Xprop_value;    int Xprop_value;
391    int Xprop_fail_result;    int Xprop_fail_result;
392    int Xoclength;    int Xoclength;
393    uschar Xocchars[8];    pcre_uchar Xocchars[6];
394  #endif  #endif
395    
396    int Xcodelink;    int Xcodelink;
# Line 427  returns a negative (error) response, the Line 432  returns a negative (error) response, the
432  same response. */  same response. */
433    
434  /* 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
435  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
436  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.
437  something has been matched). For hard partial matching, we then return  something has been matched). For hard partial matching, we then return
438  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 438  the subject. */ Line 443  the subject. */
443        eptr > md->start_used_ptr) \        eptr > md->start_used_ptr) \
444      { \      { \
445      md->hitend = TRUE; \      md->hitend = TRUE; \
446      if (md->partial > 1) MRRETURN(PCRE_ERROR_PARTIAL); \      if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL); \
447      }      }
448    
449  #define SCHECK_PARTIAL()\  #define SCHECK_PARTIAL()\
450    if (md->partial != 0 && eptr > md->start_used_ptr) \    if (md->partial != 0 && eptr > md->start_used_ptr) \
451      { \      { \
452      md->hitend = TRUE; \      md->hitend = TRUE; \
453      if (md->partial > 1) MRRETURN(PCRE_ERROR_PARTIAL); \      if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL); \
454      }      }
455    
456    
457  /* Performance note: It might be tempting to extract commonly used fields from  /* Performance note: It might be tempting to extract commonly used fields from
458  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
459  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
460  made performance worse.  made performance worse.
461    
# Line 459  Arguments: Line 464  Arguments:
464     ecode       pointer to current position in compiled code     ecode       pointer to current position in compiled code
465     mstart      pointer to the current match start position (can be modified     mstart      pointer to the current match start position (can be modified
466                   by encountering \K)                   by encountering \K)
    markptr     pointer to the most recent MARK name, or NULL  
467     offset_top  current top pointer     offset_top  current top pointer
468     md          pointer to "static" info for the match     md          pointer to "static" info for the match
469     eptrb       pointer to chain of blocks containing eptr at start of     eptrb       pointer to chain of blocks containing eptr at start of
# Line 474  Returns:       MATCH_MATCH if matched Line 478  Returns:       MATCH_MATCH if matched
478  */  */
479    
480  static int  static int
481  match(REGISTER USPTR eptr, REGISTER const uschar *ecode, USPTR mstart,  match(REGISTER PCRE_PUCHAR eptr, REGISTER const pcre_uchar *ecode,
482    const uschar *markptr, int offset_top, match_data *md, eptrblock *eptrb,    PCRE_PUCHAR mstart, int offset_top, match_data *md, eptrblock *eptrb,
483    unsigned int rdepth)    unsigned int rdepth)
484  {  {
485  /* 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,
# Line 485  so they can be ordinary variables in all Line 489  so they can be ordinary variables in all
489  register int  rrc;         /* Returns from recursive calls */  register int  rrc;         /* Returns from recursive calls */
490  register int  i;           /* Used for loops not involving calls to RMATCH() */  register int  i;           /* Used for loops not involving calls to RMATCH() */
491  register unsigned int c;   /* Character values not kept over RMATCH() calls */  register unsigned int c;   /* Character values not kept over RMATCH() calls */
492  register BOOL utf8;        /* Local copy of UTF-8 flag for speed */  register BOOL utf;         /* Local copy of UTF flag for speed */
493    
494  BOOL minimize, possessive; /* Quantifier options */  BOOL minimize, possessive; /* Quantifier options */
495  BOOL caseless;  BOOL caseless;
496  int condcode;  int condcode;
497    
498  /* 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
499  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
500  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
501  heap whenever RMATCH() does a "recursion". See the macro definitions above. */  whenever RMATCH() does a "recursion". See the macro definitions above. Putting
502    the top-level on the stack rather than malloc-ing them all gives a performance
503    boost in many cases where there is not much "recursion". */
504    
505  #ifdef NO_RECURSE  #ifdef NO_RECURSE
506  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 */  
507    
508  /* Copy in the original argument variables */  /* Copy in the original argument variables */
509    
510  frame->Xeptr = eptr;  frame->Xeptr = eptr;
511  frame->Xecode = ecode;  frame->Xecode = ecode;
512  frame->Xmstart = mstart;  frame->Xmstart = mstart;
 frame->Xmarkptr = markptr;  
513  frame->Xoffset_top = offset_top;  frame->Xoffset_top = offset_top;
514  frame->Xeptrb = eptrb;  frame->Xeptrb = eptrb;
515  frame->Xrdepth = rdepth;  frame->Xrdepth = rdepth;
# Line 520  HEAP_RECURSE: Line 523  HEAP_RECURSE:
523  #define eptr               frame->Xeptr  #define eptr               frame->Xeptr
524  #define ecode              frame->Xecode  #define ecode              frame->Xecode
525  #define mstart             frame->Xmstart  #define mstart             frame->Xmstart
 #define markptr            frame->Xmarkptr  
526  #define offset_top         frame->Xoffset_top  #define offset_top         frame->Xoffset_top
527  #define eptrb              frame->Xeptrb  #define eptrb              frame->Xeptrb
528  #define rdepth             frame->Xrdepth  #define rdepth             frame->Xrdepth
529    
530  /* Ditto for the local variables */  /* Ditto for the local variables */
531    
532  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
533  #define charptr            frame->Xcharptr  #define charptr            frame->Xcharptr
534  #endif  #endif
535  #define callpat            frame->Xcallpat  #define callpat            frame->Xcallpat
# Line 585  declarations can be cut out in a block. Line 587  declarations can be cut out in a block.
587  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
588  to RMATCH(). */  to RMATCH(). */
589    
590  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
591  const uschar *charptr;  const pcre_uchar *charptr;
592  #endif  #endif
593  const uschar *callpat;  const pcre_uchar *callpat;
594  const uschar *data;  const pcre_uchar *data;
595  const uschar *next;  const pcre_uchar *next;
596  USPTR         pp;  PCRE_PUCHAR       pp;
597  const uschar *prev;  const pcre_uchar *prev;
598  USPTR         saved_eptr;  PCRE_PUCHAR       saved_eptr;
599    
600  recursion_info new_recursive;  recursion_info new_recursive;
601    
# Line 606  int prop_type; Line 608  int prop_type;
608  int prop_value;  int prop_value;
609  int prop_fail_result;  int prop_fail_result;
610  int oclength;  int oclength;
611  uschar occhars[8];  pcre_uchar occhars[6];
612  #endif  #endif
613    
614  int codelink;  int codelink;
# Line 622  int save_offset1, save_offset2, save_off Line 624  int save_offset1, save_offset2, save_off
624  int stacksave[REC_STACK_SAVE_MAX];  int stacksave[REC_STACK_SAVE_MAX];
625    
626  eptrblock newptrb;  eptrblock newptrb;
627    
628    /* There is a special fudge for calling match() in a way that causes it to
629    measure the size of its basic stack frame when the stack is being used for
630    recursion. The second argument (ecode) being NULL triggers this behaviour. It
631    cannot normally ever be NULL. The return is the negated value of the frame
632    size. */
633    
634    if (ecode == NULL)
635      {
636      if (rdepth == 0)
637        return match((PCRE_PUCHAR)&rdepth, NULL, NULL, 0, NULL, NULL, 1);
638      else
639        {
640        int len = (char *)&rdepth - (char *)eptr;
641        return (len > 0)? -len : len;
642        }
643      }
644  #endif     /* NO_RECURSE */  #endif     /* NO_RECURSE */
645    
646  /* 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 634  the alternative names that are used. */ Line 653  the alternative names that are used. */
653  #define code_offset   codelink  #define code_offset   codelink
654  #define condassert    condition  #define condassert    condition
655  #define matched_once  prev_is_word  #define matched_once  prev_is_word
656    #define foc           number
657    #define save_mark     data
658    
659  /* These statements are here to stop the compiler complaining about unitialized  /* These statements are here to stop the compiler complaining about unitialized
660  variables. */  variables. */
# Line 659  defined). However, RMATCH isn't like a f Line 680  defined). However, RMATCH isn't like a f
680  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,
681  however, impact performance when true recursion is being used. */  however, impact performance when true recursion is being used. */
682    
683  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
684  utf8 = md->utf8;       /* Local copy of the flag */  utf = md->utf;       /* Local copy of the flag */
685  #else  #else
686  utf8 = FALSE;  utf = FALSE;
687  #endif  #endif
688    
689  /* 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 701  for (;;) Line 722  for (;;)
722    switch(op)    switch(op)
723      {      {
724      case OP_MARK:      case OP_MARK:
725      markptr = ecode + 2;      md->nomatch_mark = ecode + 2;
726      RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode] + ecode[1], offset_top, md,      md->mark = NULL;    /* In case previously set by assertion */
727        RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode] + ecode[1], offset_top, md,
728        eptrb, RM55);        eptrb, RM55);
729        if ((rrc == MATCH_MATCH || rrc == MATCH_ACCEPT) &&
730             md->mark == NULL) md->mark = ecode + 2;
731    
732      /* A return of MATCH_SKIP_ARG means that matching failed at SKIP with an      /* A return of MATCH_SKIP_ARG means that matching failed at SKIP with an
733      argument, and we must check whether that argument matches this MARK's      argument, and we must check whether that argument matches this MARK's
# Line 712  for (;;) Line 736  for (;;)
736      position and return MATCH_SKIP. Otherwise, pass back the return code      position and return MATCH_SKIP. Otherwise, pass back the return code
737      unaltered. */      unaltered. */
738    
739      if (rrc == MATCH_SKIP_ARG &&      else if (rrc == MATCH_SKIP_ARG &&
740          strcmp((char *)markptr, (char *)(md->start_match_ptr)) == 0)          STRCMP_UC_UC(ecode + 2, md->start_match_ptr) == 0)
741        {        {
742        md->start_match_ptr = eptr;        md->start_match_ptr = eptr;
743        RRETURN(MATCH_SKIP);        RRETURN(MATCH_SKIP);
744        }        }
   
     if (md->mark == NULL) md->mark = markptr;  
745      RRETURN(rrc);      RRETURN(rrc);
746    
747      case OP_FAIL:      case OP_FAIL:
748      MRRETURN(MATCH_NOMATCH);      RRETURN(MATCH_NOMATCH);
749    
750      /* COMMIT overrides PRUNE, SKIP, and THEN */      /* COMMIT overrides PRUNE, SKIP, and THEN */
751    
752      case OP_COMMIT:      case OP_COMMIT:
753      RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,      RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md,
754        eptrb, RM52);        eptrb, RM52);
755      if (rrc != MATCH_NOMATCH && rrc != MATCH_PRUNE &&      if (rrc != MATCH_NOMATCH && rrc != MATCH_PRUNE &&
756          rrc != MATCH_SKIP && rrc != MATCH_SKIP_ARG &&          rrc != MATCH_SKIP && rrc != MATCH_SKIP_ARG &&
757          rrc != MATCH_THEN)          rrc != MATCH_THEN)
758        RRETURN(rrc);        RRETURN(rrc);
759      MRRETURN(MATCH_COMMIT);      RRETURN(MATCH_COMMIT);
760    
761      /* PRUNE overrides THEN */      /* PRUNE overrides THEN */
762    
763      case OP_PRUNE:      case OP_PRUNE:
764      RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,      RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md,
765        eptrb, RM51);        eptrb, RM51);
766      if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN) RRETURN(rrc);      if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN) RRETURN(rrc);
767      MRRETURN(MATCH_PRUNE);      RRETURN(MATCH_PRUNE);
768    
769      case OP_PRUNE_ARG:      case OP_PRUNE_ARG:
770      RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode] + ecode[1], offset_top, md,      md->nomatch_mark = ecode + 2;
771        md->mark = NULL;    /* In case previously set by assertion */
772        RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode] + ecode[1], offset_top, md,
773        eptrb, RM56);        eptrb, RM56);
774        if ((rrc == MATCH_MATCH || rrc == MATCH_ACCEPT) &&
775             md->mark == NULL) md->mark = ecode + 2;
776      if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN) RRETURN(rrc);      if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN) RRETURN(rrc);
     md->mark = ecode + 2;  
777      RRETURN(MATCH_PRUNE);      RRETURN(MATCH_PRUNE);
778    
779      /* SKIP overrides PRUNE and THEN */      /* SKIP overrides PRUNE and THEN */
780    
781      case OP_SKIP:      case OP_SKIP:
782      RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,      RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md,
783        eptrb, RM53);        eptrb, RM53);
784      if (rrc != MATCH_NOMATCH && rrc != MATCH_PRUNE && rrc != MATCH_THEN)      if (rrc != MATCH_NOMATCH && rrc != MATCH_PRUNE && rrc != MATCH_THEN)
785        RRETURN(rrc);        RRETURN(rrc);
786      md->start_match_ptr = eptr;   /* Pass back current position */      md->start_match_ptr = eptr;   /* Pass back current position */
787      MRRETURN(MATCH_SKIP);      RRETURN(MATCH_SKIP);
788    
789        /* Note that, for Perl compatibility, SKIP with an argument does NOT set
790        nomatch_mark. There is a flag that disables this opcode when re-matching a
791        pattern that ended with a SKIP for which there was not a matching MARK. */
792    
793      case OP_SKIP_ARG:      case OP_SKIP_ARG:
794      RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode] + ecode[1], offset_top, md,      if (md->ignore_skip_arg)
795          {
796          ecode += PRIV(OP_lengths)[*ecode] + ecode[1];
797          break;
798          }
799        RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode] + ecode[1], offset_top, md,
800        eptrb, RM57);        eptrb, RM57);
801      if (rrc != MATCH_NOMATCH && rrc != MATCH_PRUNE && rrc != MATCH_THEN)      if (rrc != MATCH_NOMATCH && rrc != MATCH_PRUNE && rrc != MATCH_THEN)
802        RRETURN(rrc);        RRETURN(rrc);
803    
804      /* 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
805      returning the special MATCH_SKIP_ARG return code. This will either be      returning the special MATCH_SKIP_ARG return code. This will either be
806      caught by a matching MARK, or get to the top, where it is treated the same      caught by a matching MARK, or get to the top, where it causes a rematch
807      as PRUNE. */      with the md->ignore_skip_arg flag set. */
808    
809      md->start_match_ptr = ecode + 2;      md->start_match_ptr = ecode + 2;
810      RRETURN(MATCH_SKIP_ARG);      RRETURN(MATCH_SKIP_ARG);
# Line 780  for (;;) Line 814  for (;;)
814      match pointer to do this. */      match pointer to do this. */
815    
816      case OP_THEN:      case OP_THEN:
817      RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,      RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md,
818        eptrb, RM54);        eptrb, RM54);
819      if (rrc != MATCH_NOMATCH) RRETURN(rrc);      if (rrc != MATCH_NOMATCH) RRETURN(rrc);
820      md->start_match_ptr = ecode;      md->start_match_ptr = ecode;
821      MRRETURN(MATCH_THEN);      RRETURN(MATCH_THEN);
822    
823      case OP_THEN_ARG:      case OP_THEN_ARG:
824      RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode] + ecode[1], offset_top,      md->nomatch_mark = ecode + 2;
825        md->mark = NULL;    /* In case previously set by assertion */
826        RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode] + ecode[1], offset_top,
827        md, eptrb, RM58);        md, eptrb, RM58);
828        if ((rrc == MATCH_MATCH || rrc == MATCH_ACCEPT) &&
829             md->mark == NULL) md->mark = ecode + 2;
830      if (rrc != MATCH_NOMATCH) RRETURN(rrc);      if (rrc != MATCH_NOMATCH) RRETURN(rrc);
831      md->start_match_ptr = ecode;      md->start_match_ptr = ecode;
     md->mark = ecode + 2;  
832      RRETURN(MATCH_THEN);      RRETURN(MATCH_THEN);
833    
834      /* Handle an atomic group that does not contain any capturing parentheses.      /* Handle an atomic group that does not contain any capturing parentheses.
835      This can be handled like an assertion. Prior to 8.13, all atomic groups      This can be handled like an assertion. Prior to 8.13, all atomic groups
836      were handled this way. In 8.13, the code was changed as below for ONCE, so      were handled this way. In 8.13, the code was changed as below for ONCE, so
837      that backups pass through the group and thereby reset captured values.      that backups pass through the group and thereby reset captured values.
838      However, this uses a lot more stack, so in 8.20, atomic groups that do not      However, this uses a lot more stack, so in 8.20, atomic groups that do not
839      contain any captures generate OP_ONCE_NC, which can be handled in the old,      contain any captures generate OP_ONCE_NC, which can be handled in the old,
840      less stack intensive way.      less stack intensive way.
841    
842      Check the alternative branches in turn - the matching won't pass the KET      Check the alternative branches in turn - the matching won't pass the KET
# Line 810  for (;;) Line 847  for (;;)
847      case OP_ONCE_NC:      case OP_ONCE_NC:
848      prev = ecode;      prev = ecode;
849      saved_eptr = eptr;      saved_eptr = eptr;
850        save_mark = md->mark;
851      do      do
852        {        {
853        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, eptrb, RM64);        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, eptrb, RM64);
# Line 821  for (;;) Line 859  for (;;)
859        if (rrc == MATCH_THEN)        if (rrc == MATCH_THEN)
860          {          {
861          next = ecode + GET(ecode,1);          next = ecode + GET(ecode,1);
862          if (md->start_match_ptr < next &&          if (md->start_match_ptr < next &&
863              (*ecode == OP_ALT || *next == OP_ALT))              (*ecode == OP_ALT || *next == OP_ALT))
864            rrc = MATCH_NOMATCH;            rrc = MATCH_NOMATCH;
865          }          }
866    
867        if (rrc != MATCH_NOMATCH) RRETURN(rrc);        if (rrc != MATCH_NOMATCH) RRETURN(rrc);
868        ecode += GET(ecode,1);        ecode += GET(ecode,1);
869          md->mark = save_mark;
870        }        }
871      while (*ecode == OP_ALT);      while (*ecode == OP_ALT);
872    
# Line 867  for (;;) Line 906  for (;;)
906        }        }
907      else  /* OP_KETRMAX */      else  /* OP_KETRMAX */
908        {        {
       md->match_function_type = MATCH_CBEGROUP;  
909        RMATCH(eptr, prev, offset_top, md, eptrb, RM66);        RMATCH(eptr, prev, offset_top, md, eptrb, RM66);
910        if (rrc != MATCH_NOMATCH) RRETURN(rrc);        if (rrc != MATCH_NOMATCH) RRETURN(rrc);
911        ecode += 1 + LINK_SIZE;        ecode += 1 + LINK_SIZE;
# Line 907  for (;;) Line 945  for (;;)
945        save_offset2 = md->offset_vector[offset+1];        save_offset2 = md->offset_vector[offset+1];
946        save_offset3 = md->offset_vector[md->offset_end - number];        save_offset3 = md->offset_vector[md->offset_end - number];
947        save_capture_last = md->capture_last;        save_capture_last = md->capture_last;
948          save_mark = md->mark;
949    
950        DPRINTF(("saving %d %d %d\n", save_offset1, save_offset2, save_offset3));        DPRINTF(("saving %d %d %d\n", save_offset1, save_offset2, save_offset3));
951        md->offset_vector[md->offset_end - number] =        md->offset_vector[md->offset_end - number] =
# Line 915  for (;;) Line 954  for (;;)
954        for (;;)        for (;;)
955          {          {
956          if (op >= OP_SBRA) md->match_function_type = MATCH_CBEGROUP;          if (op >= OP_SBRA) md->match_function_type = MATCH_CBEGROUP;
957          RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,          RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md,
958            eptrb, RM1);            eptrb, RM1);
959          if (rrc == MATCH_ONCE) break;  /* Backing up through an atomic group */          if (rrc == MATCH_ONCE) break;  /* Backing up through an atomic group */
960    
961          /* If we backed up to a THEN, check whether it is within the current          /* If we backed up to a THEN, check whether it is within the current
962          branch by comparing the address of the THEN that is passed back with          branch by comparing the address of the THEN that is passed back with
963          the end of the branch. If it is within the current branch, and the          the end of the branch. If it is within the current branch, and the
964          branch is one of two or more alternatives (it either starts or ends          branch is one of two or more alternatives (it either starts or ends
965          with OP_ALT), we have reached the limit of THEN's action, so convert          with OP_ALT), we have reached the limit of THEN's action, so convert
966          the return code to NOMATCH, which will cause normal backtracking to          the return code to NOMATCH, which will cause normal backtracking to
967          happen from now on. Otherwise, THEN is passed back to an outer          happen from now on. Otherwise, THEN is passed back to an outer
968          alternative. This implements Perl's treatment of parenthesized groups,          alternative. This implements Perl's treatment of parenthesized groups,
969          where a group not containing | does not affect the current alternative,          where a group not containing | does not affect the current alternative,
970          that is, (X) is NOT the same as (X|(*F)). */          that is, (X) is NOT the same as (X|(*F)). */
971    
972          if (rrc == MATCH_THEN)          if (rrc == MATCH_THEN)
973            {            {
974            next = ecode + GET(ecode,1);            next = ecode + GET(ecode,1);
975            if (md->start_match_ptr < next &&            if (md->start_match_ptr < next &&
976                (*ecode == OP_ALT || *next == OP_ALT))                (*ecode == OP_ALT || *next == OP_ALT))
977              rrc = MATCH_NOMATCH;              rrc = MATCH_NOMATCH;
978            }            }
979    
980          /* Anything other than NOMATCH is passed back. */          /* Anything other than NOMATCH is passed back. */
981    
982          if (rrc != MATCH_NOMATCH) RRETURN(rrc);          if (rrc != MATCH_NOMATCH) RRETURN(rrc);
983          md->capture_last = save_capture_last;          md->capture_last = save_capture_last;
984          ecode += GET(ecode, 1);          ecode += GET(ecode, 1);
985            md->mark = save_mark;
986          if (*ecode != OP_ALT) break;          if (*ecode != OP_ALT) break;
987          }          }
988    
# Line 953  for (;;) Line 993  for (;;)
993    
994        /* At this point, rrc will be one of MATCH_ONCE or MATCH_NOMATCH. */        /* At this point, rrc will be one of MATCH_ONCE or MATCH_NOMATCH. */
995    
       if (md->mark == NULL) md->mark = markptr;  
996        RRETURN(rrc);        RRETURN(rrc);
997        }        }
998    
# Line 995  for (;;) Line 1034  for (;;)
1034    
1035      for (;;)      for (;;)
1036        {        {
1037        if (op >= OP_SBRA || op == OP_ONCE) md->match_function_type = MATCH_CBEGROUP;        if (op >= OP_SBRA || op == OP_ONCE)
1038            md->match_function_type = MATCH_CBEGROUP;
1039    
1040        /* 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
1041        the pattern, and this is the final alternative, optimize as described        the pattern, and this is the final alternative, optimize as described
# Line 1003  for (;;) Line 1043  for (;;)
1043    
1044        else if (!md->hasthen && ecode[GET(ecode, 1)] != OP_ALT)        else if (!md->hasthen && ecode[GET(ecode, 1)] != OP_ALT)
1045          {          {
1046          ecode += _pcre_OP_lengths[*ecode];          ecode += PRIV(OP_lengths)[*ecode];
1047          goto TAIL_RECURSE;          goto TAIL_RECURSE;
1048          }          }
1049    
1050        /* In all other cases, we have to make another call to match(). */        /* In all other cases, we have to make another call to match(). */
1051    
1052        RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md, eptrb,        save_mark = md->mark;
1053          RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md, eptrb,
1054          RM2);          RM2);
1055    
1056        /* See comment in the code for capturing groups above about handling        /* See comment in the code for capturing groups above about handling
1057        THEN. */        THEN. */
1058    
1059        if (rrc == MATCH_THEN)        if (rrc == MATCH_THEN)
1060          {          {
1061          next = ecode + GET(ecode,1);          next = ecode + GET(ecode,1);
1062          if (md->start_match_ptr < next &&          if (md->start_match_ptr < next &&
1063              (*ecode == OP_ALT || *next == OP_ALT))              (*ecode == OP_ALT || *next == OP_ALT))
1064            rrc = MATCH_NOMATCH;            rrc = MATCH_NOMATCH;
1065          }          }
1066    
1067        if (rrc != MATCH_NOMATCH)        if (rrc != MATCH_NOMATCH)
1068          {          {
1069          if (rrc == MATCH_ONCE)          if (rrc == MATCH_ONCE)
1070            {            {
1071            const uschar *scode = ecode;            const pcre_uchar *scode = ecode;
1072            if (*scode != OP_ONCE)           /* If not at start, find it */            if (*scode != OP_ONCE)           /* If not at start, find it */
1073              {              {
1074              while (*scode == OP_ALT) scode += GET(scode, 1);              while (*scode == OP_ALT) scode += GET(scode, 1);
# Line 1038  for (;;) Line 1079  for (;;)
1079          RRETURN(rrc);          RRETURN(rrc);
1080          }          }
1081        ecode += GET(ecode, 1);        ecode += GET(ecode, 1);
1082          md->mark = save_mark;
1083        if (*ecode != OP_ALT) break;        if (*ecode != OP_ALT) break;
1084        }        }
1085    
     if (md->mark == NULL) md->mark = markptr;  
1086      RRETURN(MATCH_NOMATCH);      RRETURN(MATCH_NOMATCH);
1087    
1088      /* Handle possessive capturing brackets with an unlimited repeat. We come      /* Handle possessive capturing brackets with an unlimited repeat. We come
# Line 1070  for (;;) Line 1111  for (;;)
1111      if (offset < md->offset_max)      if (offset < md->offset_max)
1112        {        {
1113        matched_once = FALSE;        matched_once = FALSE;
1114        code_offset = ecode - md->start_code;        code_offset = (int)(ecode - md->start_code);
1115    
1116        save_offset1 = md->offset_vector[offset];        save_offset1 = md->offset_vector[offset];
1117        save_offset2 = md->offset_vector[offset+1];        save_offset2 = md->offset_vector[offset+1];
# Line 1093  for (;;) Line 1134  for (;;)
1134          md->offset_vector[md->offset_end - number] =          md->offset_vector[md->offset_end - number] =
1135            (int)(eptr - md->start_subject);            (int)(eptr - md->start_subject);
1136          if (op >= OP_SBRA) md->match_function_type = MATCH_CBEGROUP;          if (op >= OP_SBRA) md->match_function_type = MATCH_CBEGROUP;
1137          RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,          RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md,
1138            eptrb, RM63);            eptrb, RM63);
1139          if (rrc == MATCH_KETRPOS)          if (rrc == MATCH_KETRPOS)
1140            {            {
# Line 1104  for (;;) Line 1145  for (;;)
1145            matched_once = TRUE;            matched_once = TRUE;
1146            continue;            continue;
1147            }            }
1148    
1149          /* See comment in the code for capturing groups above about handling          /* See comment in the code for capturing groups above about handling
1150          THEN. */          THEN. */
1151    
1152          if (rrc == MATCH_THEN)          if (rrc == MATCH_THEN)
1153            {            {
1154            next = ecode + GET(ecode,1);            next = ecode + GET(ecode,1);
1155            if (md->start_match_ptr < next &&            if (md->start_match_ptr < next &&
1156                (*ecode == OP_ALT || *next == OP_ALT))                (*ecode == OP_ALT || *next == OP_ALT))
1157              rrc = MATCH_NOMATCH;              rrc = MATCH_NOMATCH;
1158            }            }
1159    
1160          if (rrc != MATCH_NOMATCH) RRETURN(rrc);          if (rrc != MATCH_NOMATCH) RRETURN(rrc);
1161          md->capture_last = save_capture_last;          md->capture_last = save_capture_last;
# Line 1129  for (;;) Line 1170  for (;;)
1170          md->offset_vector[md->offset_end - number] = save_offset3;          md->offset_vector[md->offset_end - number] = save_offset3;
1171          }          }
1172    
       if (md->mark == NULL) md->mark = markptr;  
1173        if (allow_zero || matched_once)        if (allow_zero || matched_once)
1174          {          {
1175          ecode += 1 + LINK_SIZE;          ecode += 1 + LINK_SIZE;
# Line 1161  for (;;) Line 1201  for (;;)
1201    
1202      POSSESSIVE_NON_CAPTURE:      POSSESSIVE_NON_CAPTURE:
1203      matched_once = FALSE;      matched_once = FALSE;
1204      code_offset = ecode - md->start_code;      code_offset = (int)(ecode - md->start_code);
1205    
1206      for (;;)      for (;;)
1207        {        {
1208        if (op >= OP_SBRA) md->match_function_type = MATCH_CBEGROUP;        if (op >= OP_SBRA) md->match_function_type = MATCH_CBEGROUP;
1209        RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,        RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md,
1210          eptrb, RM48);          eptrb, RM48);
1211        if (rrc == MATCH_KETRPOS)        if (rrc == MATCH_KETRPOS)
1212          {          {
# Line 1176  for (;;) Line 1216  for (;;)
1216          matched_once = TRUE;          matched_once = TRUE;
1217          continue;          continue;
1218          }          }
1219    
1220        /* See comment in the code for capturing groups above about handling        /* See comment in the code for capturing groups above about handling
1221        THEN. */        THEN. */
1222    
1223        if (rrc == MATCH_THEN)        if (rrc == MATCH_THEN)
1224          {          {
1225          next = ecode + GET(ecode,1);          next = ecode + GET(ecode,1);
1226          if (md->start_match_ptr < next &&          if (md->start_match_ptr < next &&
1227              (*ecode == OP_ALT || *next == OP_ALT))              (*ecode == OP_ALT || *next == OP_ALT))
1228            rrc = MATCH_NOMATCH;            rrc = MATCH_NOMATCH;
1229          }          }
1230    
1231        if (rrc != MATCH_NOMATCH) RRETURN(rrc);        if (rrc != MATCH_NOMATCH) RRETURN(rrc);
1232        ecode += GET(ecode, 1);        ecode += GET(ecode, 1);
# Line 1216  for (;;) Line 1256  for (;;)
1256    
1257      if (ecode[LINK_SIZE+1] == OP_CALLOUT)      if (ecode[LINK_SIZE+1] == OP_CALLOUT)
1258        {        {
1259        if (pcre_callout != NULL)        if (PUBL(callout) != NULL)
1260          {          {
1261          pcre_callout_block cb;          PUBL(callout_block) cb;
1262          cb.version          = 2;   /* Version 1 of the callout block */          cb.version          = 2;   /* Version 1 of the callout block */
1263          cb.callout_number   = ecode[LINK_SIZE+2];          cb.callout_number   = ecode[LINK_SIZE+2];
1264          cb.offset_vector    = md->offset_vector;          cb.offset_vector    = md->offset_vector;
1265    #ifdef COMPILE_PCRE8
1266          cb.subject          = (PCRE_SPTR)md->start_subject;          cb.subject          = (PCRE_SPTR)md->start_subject;
1267    #else
1268            cb.subject          = (PCRE_SPTR16)md->start_subject;
1269    #endif
1270          cb.subject_length   = (int)(md->end_subject - md->start_subject);          cb.subject_length   = (int)(md->end_subject - md->start_subject);
1271          cb.start_match      = (int)(mstart - md->start_subject);          cb.start_match      = (int)(mstart - md->start_subject);
1272          cb.current_position = (int)(eptr - md->start_subject);          cb.current_position = (int)(eptr - md->start_subject);
# Line 1231  for (;;) Line 1275  for (;;)
1275          cb.capture_top      = offset_top/2;          cb.capture_top      = offset_top/2;
1276          cb.capture_last     = md->capture_last;          cb.capture_last     = md->capture_last;
1277          cb.callout_data     = md->callout_data;          cb.callout_data     = md->callout_data;
1278          cb.mark             = markptr;          cb.mark             = md->nomatch_mark;
1279          if ((rrc = (*pcre_callout)(&cb)) > 0) MRRETURN(MATCH_NOMATCH);          if ((rrc = (*PUBL(callout))(&cb)) > 0) RRETURN(MATCH_NOMATCH);
1280          if (rrc < 0) RRETURN(rrc);          if (rrc < 0) RRETURN(rrc);
1281          }          }
1282        ecode += _pcre_OP_lengths[OP_CALLOUT];        ecode += PRIV(OP_lengths)[OP_CALLOUT];
1283        }        }
1284    
1285      condcode = ecode[LINK_SIZE+1];      condcode = ecode[LINK_SIZE+1];
# Line 1252  for (;;) Line 1296  for (;;)
1296        else        else
1297          {          {
1298          int recno = GET2(ecode, LINK_SIZE + 2);   /* Recursion group number*/          int recno = GET2(ecode, LINK_SIZE + 2);   /* Recursion group number*/
1299          condition =  (recno == RREF_ANY || recno == md->recursive->group_num);          condition = (recno == RREF_ANY || recno == md->recursive->group_num);
1300    
1301          /* If the test is for recursion into a specific subpattern, and it is          /* If the test is for recursion into a specific subpattern, and it is
1302          false, but the test was set up by name, scan the table to see if the          false, but the test was set up by name, scan the table to see if the
1303          name refers to any other numbers, and test them. The condition is true          name refers to any other numbers, and test them. The condition is true
1304          if any one is set. */          if any one is set. */
1305    
1306          if (!condition && condcode == OP_NRREF && recno != RREF_ANY)          if (!condition && condcode == OP_NRREF)
1307            {            {
1308            uschar *slotA = md->name_table;            pcre_uchar *slotA = md->name_table;
1309            for (i = 0; i < md->name_count; i++)            for (i = 0; i < md->name_count; i++)
1310              {              {
1311              if (GET2(slotA, 0) == recno) break;              if (GET2(slotA, 0) == recno) break;
# Line 1274  for (;;) Line 1318  for (;;)
1318    
1319            if (i < md->name_count)            if (i < md->name_count)
1320              {              {
1321              uschar *slotB = slotA;              pcre_uchar *slotB = slotA;
1322              while (slotB > md->name_table)              while (slotB > md->name_table)
1323                {                {
1324                slotB -= md->name_entry_size;                slotB -= md->name_entry_size;
1325                if (strcmp((char *)slotA + 2, (char *)slotB + 2) == 0)                if (STRCMP_UC_UC(slotA + IMM2_SIZE, slotB + IMM2_SIZE) == 0)
1326                  {                  {
1327                  condition = GET2(slotB, 0) == md->recursive->group_num;                  condition = GET2(slotB, 0) == md->recursive->group_num;
1328                  if (condition) break;                  if (condition) break;
# Line 1294  for (;;) Line 1338  for (;;)
1338                for (i++; i < md->name_count; i++)                for (i++; i < md->name_count; i++)
1339                  {                  {
1340                  slotB += md->name_entry_size;                  slotB += md->name_entry_size;
1341                  if (strcmp((char *)slotA + 2, (char *)slotB + 2) == 0)                  if (STRCMP_UC_UC(slotA + IMM2_SIZE, slotB + IMM2_SIZE) == 0)
1342                    {                    {
1343                    condition = GET2(slotB, 0) == md->recursive->group_num;                    condition = GET2(slotB, 0) == md->recursive->group_num;
1344                    if (condition) break;                    if (condition) break;
# Line 1307  for (;;) Line 1351  for (;;)
1351    
1352          /* Chose branch according to the condition */          /* Chose branch according to the condition */
1353    
1354          ecode += condition? 3 : GET(ecode, 1);          ecode += condition? 1 + IMM2_SIZE : GET(ecode, 1);
1355          }          }
1356        }        }
1357    
# Line 1324  for (;;) Line 1368  for (;;)
1368        if (!condition && condcode == OP_NCREF)        if (!condition && condcode == OP_NCREF)
1369          {          {
1370          int refno = offset >> 1;          int refno = offset >> 1;
1371          uschar *slotA = md->name_table;          pcre_uchar *slotA = md->name_table;
1372    
1373          for (i = 0; i < md->name_count; i++)          for (i = 0; i < md->name_count; i++)
1374            {            {
# Line 1338  for (;;) Line 1382  for (;;)
1382    
1383          if (i < md->name_count)          if (i < md->name_count)
1384            {            {
1385            uschar *slotB = slotA;            pcre_uchar *slotB = slotA;
1386            while (slotB > md->name_table)            while (slotB > md->name_table)
1387              {              {
1388              slotB -= md->name_entry_size;              slotB -= md->name_entry_size;
1389              if (strcmp((char *)slotA + 2, (char *)slotB + 2) == 0)              if (STRCMP_UC_UC(slotA + IMM2_SIZE, slotB + IMM2_SIZE) == 0)
1390                {                {
1391                offset = GET2(slotB, 0) << 1;                offset = GET2(slotB, 0) << 1;
1392                condition = offset < offset_top &&                condition = offset < offset_top &&
# Line 1360  for (;;) Line 1404  for (;;)
1404              for (i++; i < md->name_count; i++)              for (i++; i < md->name_count; i++)
1405                {                {
1406                slotB += md->name_entry_size;                slotB += md->name_entry_size;
1407                if (strcmp((char *)slotA + 2, (char *)slotB + 2) == 0)                if (STRCMP_UC_UC(slotA + IMM2_SIZE, slotB + IMM2_SIZE) == 0)
1408                  {                  {
1409                  offset = GET2(slotB, 0) << 1;                  offset = GET2(slotB, 0) << 1;
1410                  condition = offset < offset_top &&                  condition = offset < offset_top &&
# Line 1375  for (;;) Line 1419  for (;;)
1419    
1420        /* Chose branch according to the condition */        /* Chose branch according to the condition */
1421    
1422        ecode += condition? 3 : GET(ecode, 1);        ecode += condition? 1 + IMM2_SIZE : GET(ecode, 1);
1423        }        }
1424    
1425      else if (condcode == OP_DEF)     /* DEFINE - always false */      else if (condcode == OP_DEF)     /* DEFINE - always false */
# Line 1400  for (;;) Line 1444  for (;;)
1444          ecode += 1 + LINK_SIZE + GET(ecode, LINK_SIZE + 2);          ecode += 1 + LINK_SIZE + GET(ecode, LINK_SIZE + 2);
1445          while (*ecode == OP_ALT) ecode += GET(ecode, 1);          while (*ecode == OP_ALT) ecode += GET(ecode, 1);
1446          }          }
1447    
1448        /* PCRE doesn't allow the effect of (*THEN) to escape beyond an        /* PCRE doesn't allow the effect of (*THEN) to escape beyond an
1449        assertion; it is therefore treated as NOMATCH. */        assertion; it is therefore treated as NOMATCH. */
1450    
1451        else if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN)        else if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN)
1452          {          {
1453          RRETURN(rrc);         /* Need braces because of following else */          RRETURN(rrc);         /* Need braces because of following else */
1454          }          }
# Line 1432  for (;;) Line 1476  for (;;)
1476          ecode += 1 + LINK_SIZE;          ecode += 1 + LINK_SIZE;
1477          goto TAIL_RECURSE;          goto TAIL_RECURSE;
1478          }          }
1479    
1480        md->match_function_type = MATCH_CBEGROUP;        md->match_function_type = MATCH_CBEGROUP;
1481        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, eptrb, RM49);        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, eptrb, RM49);
1482        RRETURN(rrc);        RRETURN(rrc);
# Line 1467  for (;;) Line 1511  for (;;)
1511        md->offset_vector[offset+1] = (int)(eptr - md->start_subject);        md->offset_vector[offset+1] = (int)(eptr - md->start_subject);
1512        if (offset_top <= offset) offset_top = offset + 2;        if (offset_top <= offset) offset_top = offset + 2;
1513        }        }
1514      ecode += 3;      ecode += 1 + IMM2_SIZE;
1515      break;      break;
1516    
1517    
# Line 1487  for (;;) Line 1531  for (;;)
1531           (md->notempty ||           (md->notempty ||
1532             (md->notempty_atstart &&             (md->notempty_atstart &&
1533               mstart == md->start_subject + md->start_offset)))               mstart == md->start_subject + md->start_offset)))
1534        MRRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
1535    
1536      /* Otherwise, we have a match. */      /* Otherwise, we have a match. */
1537    
# Line 1496  for (;;) Line 1540  for (;;)
1540      md->start_match_ptr = mstart;       /* and the start (\K can modify) */      md->start_match_ptr = mstart;       /* and the start (\K can modify) */
1541    
1542      /* For some reason, the macros don't work properly if an expression is      /* For some reason, the macros don't work properly if an expression is
1543      given as the argument to MRRETURN when the heap is in use. */      given as the argument to RRETURN when the heap is in use. */
1544    
1545      rrc = (op == OP_END)? MATCH_MATCH : MATCH_ACCEPT;      rrc = (op == OP_END)? MATCH_MATCH : MATCH_ACCEPT;
1546      MRRETURN(rrc);      RRETURN(rrc);
1547    
1548      /* Assertion brackets. Check the alternative branches in turn - the      /* Assertion brackets. Check the alternative branches in turn - the
1549      matching won't pass the KET for an assertion. If any one branch matches,      matching won't pass the KET for an assertion. If any one branch matches,
# Line 1514  for (;;) Line 1558  for (;;)
1558    
1559      case OP_ASSERT:      case OP_ASSERT:
1560      case OP_ASSERTBACK:      case OP_ASSERTBACK:
1561        save_mark = md->mark;
1562      if (md->match_function_type == MATCH_CONDASSERT)      if (md->match_function_type == MATCH_CONDASSERT)
1563        {        {
1564        condassert = TRUE;        condassert = TRUE;
# Line 1527  for (;;) Line 1572  for (;;)
1572        if (rrc == MATCH_MATCH || rrc == MATCH_ACCEPT)        if (rrc == MATCH_MATCH || rrc == MATCH_ACCEPT)
1573          {          {
1574          mstart = md->start_match_ptr;   /* In case \K reset it */          mstart = md->start_match_ptr;   /* In case \K reset it */
         markptr = md->mark;  
1575          break;          break;
1576          }          }
1577          md->mark = save_mark;
1578        /* PCRE does not allow THEN to escape beyond an assertion; it is treated  
1579        as NOMATCH. */        /* A COMMIT failure must fail the entire assertion, without trying any
1580          subsequent branches. */
1581    
1582          if (rrc == MATCH_COMMIT) RRETURN(MATCH_NOMATCH);
1583    
1584          /* PCRE does not allow THEN to escape beyond an assertion; it
1585          is treated as NOMATCH. */
1586    
1587        if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN) RRETURN(rrc);        if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN) RRETURN(rrc);
1588        ecode += GET(ecode, 1);        ecode += GET(ecode, 1);
1589        }        }
1590      while (*ecode == OP_ALT);      while (*ecode == OP_ALT);
1591    
1592      if (*ecode == OP_KET) MRRETURN(MATCH_NOMATCH);      if (*ecode == OP_KET) RRETURN(MATCH_NOMATCH);
1593    
1594      /* If checking an assertion for a condition, return MATCH_MATCH. */      /* If checking an assertion for a condition, return MATCH_MATCH. */
1595    
# Line 1559  for (;;) Line 1609  for (;;)
1609    
1610      case OP_ASSERT_NOT:      case OP_ASSERT_NOT:
1611      case OP_ASSERTBACK_NOT:      case OP_ASSERTBACK_NOT:
1612        save_mark = md->mark;
1613      if (md->match_function_type == MATCH_CONDASSERT)      if (md->match_function_type == MATCH_CONDASSERT)
1614        {        {
1615        condassert = TRUE;        condassert = TRUE;
# Line 1569  for (;;) Line 1620  for (;;)
1620      do      do
1621        {        {
1622        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, NULL, RM5);        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, NULL, RM5);
1623        if (rrc == MATCH_MATCH || rrc == MATCH_ACCEPT) MRRETURN(MATCH_NOMATCH);        md->mark = save_mark;
1624          if (rrc == MATCH_MATCH || rrc == MATCH_ACCEPT) RRETURN(MATCH_NOMATCH);
1625        if (rrc == MATCH_SKIP || rrc == MATCH_PRUNE || rrc == MATCH_COMMIT)        if (rrc == MATCH_SKIP || rrc == MATCH_PRUNE || rrc == MATCH_COMMIT)
1626          {          {
1627          do ecode += GET(ecode,1); while (*ecode == OP_ALT);          do ecode += GET(ecode,1); while (*ecode == OP_ALT);
1628          break;          break;
1629          }          }
1630    
1631        /* PCRE does not allow THEN to escape beyond an assertion; it is treated        /* PCRE does not allow THEN to escape beyond an assertion; it is treated
1632        as NOMATCH. */        as NOMATCH. */
1633    
1634        if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN) RRETURN(rrc);        if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN) RRETURN(rrc);
# Line 1595  for (;;) Line 1647  for (;;)
1647      back a number of characters, not bytes. */      back a number of characters, not bytes. */
1648    
1649      case OP_REVERSE:      case OP_REVERSE:
1650  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
1651      if (utf8)      if (utf)
1652        {        {
1653        i = GET(ecode, 1);        i = GET(ecode, 1);
1654        while (i-- > 0)        while (i-- > 0)
1655          {          {
1656          eptr--;          eptr--;
1657          if (eptr < md->start_subject) MRRETURN(MATCH_NOMATCH);          if (eptr < md->start_subject) RRETURN(MATCH_NOMATCH);
1658          BACKCHAR(eptr);          BACKCHAR(eptr);
1659          }          }
1660        }        }
# Line 1613  for (;;) Line 1665  for (;;)
1665    
1666        {        {
1667        eptr -= GET(ecode, 1);        eptr -= GET(ecode, 1);
1668        if (eptr < md->start_subject) MRRETURN(MATCH_NOMATCH);        if (eptr < md->start_subject) RRETURN(MATCH_NOMATCH);
1669        }        }
1670    
1671      /* Save the earliest consulted character, then skip to next op code */      /* Save the earliest consulted character, then skip to next op code */
# Line 1627  for (;;) Line 1679  for (;;)
1679      function is able to force a failure. */      function is able to force a failure. */
1680    
1681      case OP_CALLOUT:      case OP_CALLOUT:
1682      if (pcre_callout != NULL)      if (PUBL(callout) != NULL)
1683        {        {
1684        pcre_callout_block cb;        PUBL(callout_block) cb;
1685        cb.version          = 2;   /* Version 1 of the callout block */        cb.version          = 2;   /* Version 1 of the callout block */
1686        cb.callout_number   = ecode[1];        cb.callout_number   = ecode[1];
1687        cb.offset_vector    = md->offset_vector;        cb.offset_vector    = md->offset_vector;
1688    #ifdef COMPILE_PCRE8
1689        cb.subject          = (PCRE_SPTR)md->start_subject;        cb.subject          = (PCRE_SPTR)md->start_subject;
1690    #else
1691          cb.subject          = (PCRE_SPTR16)md->start_subject;
1692    #endif
1693        cb.subject_length   = (int)(md->end_subject - md->start_subject);        cb.subject_length   = (int)(md->end_subject - md->start_subject);
1694        cb.start_match      = (int)(mstart - md->start_subject);        cb.start_match      = (int)(mstart - md->start_subject);
1695        cb.current_position = (int)(eptr - md->start_subject);        cb.current_position = (int)(eptr - md->start_subject);
# Line 1642  for (;;) Line 1698  for (;;)
1698        cb.capture_top      = offset_top/2;        cb.capture_top      = offset_top/2;
1699        cb.capture_last     = md->capture_last;        cb.capture_last     = md->capture_last;
1700        cb.callout_data     = md->callout_data;        cb.callout_data     = md->callout_data;
1701        cb.mark             = markptr;        cb.mark             = md->nomatch_mark;
1702        if ((rrc = (*pcre_callout)(&cb)) > 0) MRRETURN(MATCH_NOMATCH);        if ((rrc = (*PUBL(callout))(&cb)) > 0) RRETURN(MATCH_NOMATCH);
1703        if (rrc < 0) RRETURN(rrc);        if (rrc < 0) RRETURN(rrc);
1704        }        }
1705      ecode += 2 + 2*LINK_SIZE;      ecode += 2 + 2*LINK_SIZE;
# Line 1702  for (;;) Line 1758  for (;;)
1758        else        else
1759          {          {
1760          new_recursive.offset_save =          new_recursive.offset_save =
1761            (int *)(pcre_malloc)(new_recursive.saved_max * sizeof(int));            (int *)(PUBL(malloc))(new_recursive.saved_max * sizeof(int));
1762          if (new_recursive.offset_save == NULL) RRETURN(PCRE_ERROR_NOMEMORY);          if (new_recursive.offset_save == NULL) RRETURN(PCRE_ERROR_NOMEMORY);
1763          }          }
1764        memcpy(new_recursive.offset_save, md->offset_vector,        memcpy(new_recursive.offset_save, md->offset_vector,
# Line 1717  for (;;) Line 1773  for (;;)
1773        do        do
1774          {          {
1775          if (cbegroup) md->match_function_type = MATCH_CBEGROUP;          if (cbegroup) md->match_function_type = MATCH_CBEGROUP;
1776          RMATCH(eptr, callpat + _pcre_OP_lengths[*callpat], offset_top,          RMATCH(eptr, callpat + PRIV(OP_lengths)[*callpat], offset_top,
1777            md, eptrb, RM6);            md, eptrb, RM6);
1778          memcpy(md->offset_vector, new_recursive.offset_save,          memcpy(md->offset_vector, new_recursive.offset_save,
1779              new_recursive.saved_max * sizeof(int));              new_recursive.saved_max * sizeof(int));
# Line 1726  for (;;) Line 1782  for (;;)
1782            {            {
1783            DPRINTF(("Recursion matched\n"));            DPRINTF(("Recursion matched\n"));
1784            if (new_recursive.offset_save != stacksave)            if (new_recursive.offset_save != stacksave)
1785              (pcre_free)(new_recursive.offset_save);              (PUBL(free))(new_recursive.offset_save);
1786    
1787            /* 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
1788            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 1737  for (;;) Line 1793  for (;;)
1793            goto RECURSION_MATCHED;        /* Exit loop; end processing */            goto RECURSION_MATCHED;        /* Exit loop; end processing */
1794            }            }
1795    
1796          /* PCRE does not allow THEN to escape beyond a recursion; it is treated          /* PCRE does not allow THEN or COMMIT to escape beyond a recursion; it
1797          as NOMATCH. */          is treated as NOMATCH. */
1798    
1799          else if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN)          else if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN &&
1800                     rrc != MATCH_COMMIT)
1801            {            {
1802            DPRINTF(("Recursion gave error %d\n", rrc));            DPRINTF(("Recursion gave error %d\n", rrc));
1803            if (new_recursive.offset_save != stacksave)            if (new_recursive.offset_save != stacksave)
1804              (pcre_free)(new_recursive.offset_save);              (PUBL(free))(new_recursive.offset_save);
1805            RRETURN(rrc);            RRETURN(rrc);
1806            }            }
1807    
# Line 1756  for (;;) Line 1813  for (;;)
1813        DPRINTF(("Recursion didn't match\n"));        DPRINTF(("Recursion didn't match\n"));
1814        md->recursive = new_recursive.prevrec;        md->recursive = new_recursive.prevrec;
1815        if (new_recursive.offset_save != stacksave)        if (new_recursive.offset_save != stacksave)
1816          (pcre_free)(new_recursive.offset_save);          (PUBL(free))(new_recursive.offset_save);
1817        MRRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
1818        }        }
1819    
1820      RECURSION_MATCHED:      RECURSION_MATCHED:
# Line 1826  for (;;) Line 1883  for (;;)
1883        }        }
1884      else saved_eptr = NULL;      else saved_eptr = NULL;
1885    
1886      /* If we are at the end of an assertion group or a non-capturing atomic      /* If we are at the end of an assertion group or a non-capturing atomic
1887      group, stop matching and return MATCH_MATCH, but record the current high      group, stop matching and return MATCH_MATCH, but record the current high
1888      water mark for use by positive assertions. We also need to record the match      water mark for use by positive assertions. We also need to record the match
1889      start in case it was changed by \K. */      start in case it was changed by \K. */
1890    
1891      if ((*prev >= OP_ASSERT && *prev <= OP_ASSERTBACK_NOT) ||      if ((*prev >= OP_ASSERT && *prev <= OP_ASSERTBACK_NOT) ||
1892           *prev == OP_ONCE_NC)           *prev == OP_ONCE_NC)
1893        {        {
1894        md->end_match_ptr = eptr;      /* For ONCE_NC */        md->end_match_ptr = eptr;      /* For ONCE_NC */
1895        md->end_offset_top = offset_top;        md->end_offset_top = offset_top;
1896        md->start_match_ptr = mstart;        md->start_match_ptr = mstart;
1897        MRRETURN(MATCH_MATCH);         /* Sets md->mark */        RRETURN(MATCH_MATCH);         /* Sets md->mark */
1898        }        }
1899    
1900      /* For capturing groups we have to check the group number back at the start      /* For capturing groups we have to check the group number back at the start
# Line 1951  for (;;) Line 2008  for (;;)
2008          }          }
2009        if (*prev >= OP_SBRA)    /* Could match an empty string */        if (*prev >= OP_SBRA)    /* Could match an empty string */
2010          {          {
         md->match_function_type = MATCH_CBEGROUP;  
2011          RMATCH(eptr, prev, offset_top, md, eptrb, RM50);          RMATCH(eptr, prev, offset_top, md, eptrb, RM50);
2012          RRETURN(rrc);          RRETURN(rrc);
2013          }          }
# Line 1960  for (;;) Line 2016  for (;;)
2016        }        }
2017      else  /* OP_KETRMAX */      else  /* OP_KETRMAX */
2018        {        {
       if (*prev >= OP_SBRA) md->match_function_type = MATCH_CBEGROUP;  
2019        RMATCH(eptr, prev, offset_top, md, eptrb, RM13);        RMATCH(eptr, prev, offset_top, md, eptrb, RM13);
2020        if (rrc == MATCH_ONCE && md->once_target == prev) rrc = MATCH_NOMATCH;        if (rrc == MATCH_ONCE && md->once_target == prev) rrc = MATCH_NOMATCH;
2021        if (rrc != MATCH_NOMATCH) RRETURN(rrc);        if (rrc != MATCH_NOMATCH) RRETURN(rrc);
# Line 1979  for (;;) Line 2034  for (;;)
2034      /* Not multiline mode: start of subject assertion, unless notbol. */      /* Not multiline mode: start of subject assertion, unless notbol. */
2035    
2036      case OP_CIRC:      case OP_CIRC:
2037      if (md->notbol && eptr == md->start_subject) MRRETURN(MATCH_NOMATCH);      if (md->notbol && eptr == md->start_subject) RRETURN(MATCH_NOMATCH);
2038    
2039      /* Start of subject assertion */      /* Start of subject assertion */
2040    
2041      case OP_SOD:      case OP_SOD:
2042      if (eptr != md->start_subject) MRRETURN(MATCH_NOMATCH);      if (eptr != md->start_subject) RRETURN(MATCH_NOMATCH);
2043      ecode++;      ecode++;
2044      break;      break;
2045    
2046      /* Multiline mode: start of subject unless notbol, or after any newline. */      /* Multiline mode: start of subject unless notbol, or after any newline. */
2047    
2048      case OP_CIRCM:      case OP_CIRCM:
2049      if (md->notbol && eptr == md->start_subject) MRRETURN(MATCH_NOMATCH);      if (md->notbol && eptr == md->start_subject) RRETURN(MATCH_NOMATCH);
2050      if (eptr != md->start_subject &&      if (eptr != md->start_subject &&
2051          (eptr == md->end_subject || !WAS_NEWLINE(eptr)))          (eptr == md->end_subject || !WAS_NEWLINE(eptr)))
2052        MRRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2053      ecode++;      ecode++;
2054      break;      break;
2055    
2056      /* Start of match assertion */      /* Start of match assertion */
2057    
2058      case OP_SOM:      case OP_SOM:
2059      if (eptr != md->start_subject + md->start_offset) MRRETURN(MATCH_NOMATCH);      if (eptr != md->start_subject + md->start_offset) RRETURN(MATCH_NOMATCH);
2060      ecode++;      ecode++;
2061      break;      break;
2062    
# Line 2017  for (;;) Line 2072  for (;;)
2072    
2073      case OP_DOLLM:      case OP_DOLLM:
2074      if (eptr < md->end_subject)      if (eptr < md->end_subject)
2075        { if (!IS_NEWLINE(eptr)) MRRETURN(MATCH_NOMATCH); }        {
2076          if (!IS_NEWLINE(eptr))
2077            {
2078            if (md->partial != 0 &&
2079                eptr + 1 >= md->end_subject &&
2080                NLBLOCK->nltype == NLTYPE_FIXED &&
2081                NLBLOCK->nllen == 2 &&
2082                *eptr == NLBLOCK->nl[0])
2083              {
2084              md->hitend = TRUE;
2085              if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
2086              }
2087            RRETURN(MATCH_NOMATCH);
2088            }
2089          }
2090      else      else
2091        {        {
2092        if (md->noteol) MRRETURN(MATCH_NOMATCH);        if (md->noteol) RRETURN(MATCH_NOMATCH);
2093        SCHECK_PARTIAL();        SCHECK_PARTIAL();
2094        }        }
2095      ecode++;      ecode++;
# Line 2030  for (;;) Line 2099  for (;;)
2099      subject unless noteol is set. */      subject unless noteol is set. */
2100    
2101      case OP_DOLL:      case OP_DOLL:
2102      if (md->noteol) MRRETURN(MATCH_NOMATCH);      if (md->noteol) RRETURN(MATCH_NOMATCH);
2103      if (!md->endonly) goto ASSERT_NL_OR_EOS;      if (!md->endonly) goto ASSERT_NL_OR_EOS;
2104    
2105      /* ... else fall through for endonly */      /* ... else fall through for endonly */
# Line 2038  for (;;) Line 2107  for (;;)
2107      /* End of subject assertion (\z) */      /* End of subject assertion (\z) */
2108    
2109      case OP_EOD:      case OP_EOD:
2110      if (eptr < md->end_subject) MRRETURN(MATCH_NOMATCH);      if (eptr < md->end_subject) RRETURN(MATCH_NOMATCH);
2111      SCHECK_PARTIAL();      SCHECK_PARTIAL();
2112      ecode++;      ecode++;
2113      break;      break;
# Line 2049  for (;;) Line 2118  for (;;)
2118      ASSERT_NL_OR_EOS:      ASSERT_NL_OR_EOS:
2119      if (eptr < md->end_subject &&      if (eptr < md->end_subject &&
2120          (!IS_NEWLINE(eptr) || eptr != md->end_subject - md->nllen))          (!IS_NEWLINE(eptr) || eptr != md->end_subject - md->nllen))
2121        MRRETURN(MATCH_NOMATCH);        {
2122          if (md->partial != 0 &&
2123              eptr + 1 >= md->end_subject &&
2124              NLBLOCK->nltype == NLTYPE_FIXED &&
2125              NLBLOCK->nllen == 2 &&
2126              *eptr == NLBLOCK->nl[0])
2127            {
2128            md->hitend = TRUE;
2129            if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
2130            }
2131          RRETURN(MATCH_NOMATCH);
2132          }
2133    
2134      /* Either at end of string or \n before end. */      /* Either at end of string or \n before end. */
2135    
# Line 2068  for (;;) Line 2148  for (;;)
2148        be "non-word" characters. Remember the earliest consulted character for        be "non-word" characters. Remember the earliest consulted character for
2149        partial matching. */        partial matching. */
2150    
2151  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
2152        if (utf8)        if (utf)
2153          {          {
2154          /* Get status of previous character */          /* Get status of previous character */
2155    
2156          if (eptr == md->start_subject) prev_is_word = FALSE; else          if (eptr == md->start_subject) prev_is_word = FALSE; else
2157            {            {
2158            USPTR lastptr = eptr - 1;            PCRE_PUCHAR lastptr = eptr - 1;
2159            while((*lastptr & 0xc0) == 0x80) lastptr--;            BACKCHAR(lastptr);
2160            if (lastptr < md->start_used_ptr) md->start_used_ptr = lastptr;            if (lastptr < md->start_used_ptr) md->start_used_ptr = lastptr;
2161            GETCHAR(c, lastptr);            GETCHAR(c, lastptr);
2162  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
# Line 2141  for (;;) Line 2221  for (;;)
2221              }              }
2222            else            else
2223  #endif  #endif
2224            prev_is_word = ((md->ctypes[eptr[-1]] & ctype_word) != 0);            prev_is_word = MAX_255(eptr[-1])
2225                && ((md->ctypes[eptr[-1]] & ctype_word) != 0);
2226            }            }
2227    
2228          /* Get status of next character */          /* Get status of next character */
# Line 2164  for (;;) Line 2245  for (;;)
2245            }            }
2246          else          else
2247  #endif  #endif
2248          cur_is_word = ((md->ctypes[*eptr] & ctype_word) != 0);          cur_is_word = MAX_255(*eptr)
2249              && ((md->ctypes[*eptr] & ctype_word) != 0);
2250          }          }
2251    
2252        /* Now see if the situation is what we want */        /* Now see if the situation is what we want */
2253    
2254        if ((*ecode++ == OP_WORD_BOUNDARY)?        if ((*ecode++ == OP_WORD_BOUNDARY)?
2255             cur_is_word == prev_is_word : cur_is_word != prev_is_word)             cur_is_word == prev_is_word : cur_is_word != prev_is_word)
2256          MRRETURN(MATCH_NOMATCH);          RRETURN(MATCH_NOMATCH);
2257        }        }
2258      break;      break;
2259    
2260      /* Match a single character type; inline for speed */      /* Match any single character type except newline; have to take care with
2261        CRLF newlines and partial matching. */
2262    
2263      case OP_ANY:      case OP_ANY:
2264      if (IS_NEWLINE(eptr)) MRRETURN(MATCH_NOMATCH);      if (IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH);
2265        if (md->partial != 0 &&
2266            eptr + 1 >= md->end_subject &&
2267            NLBLOCK->nltype == NLTYPE_FIXED &&
2268            NLBLOCK->nllen == 2 &&
2269            *eptr == NLBLOCK->nl[0])
2270          {
2271          md->hitend = TRUE;
2272          if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
2273          }
2274    
2275      /* Fall through */      /* Fall through */
2276    
2277        /* Match any single character whatsoever. */
2278    
2279      case OP_ALLANY:      case OP_ALLANY:
2280      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 */
2281        {                            /* not be updated before SCHECK_PARTIAL. */        {                            /* not be updated before SCHECK_PARTIAL. */
2282        SCHECK_PARTIAL();        SCHECK_PARTIAL();
2283        MRRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2284        }        }
2285      eptr++;      eptr++;
2286      if (utf8) while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;  #ifdef SUPPORT_UTF
2287        if (utf) ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);
2288    #endif
2289      ecode++;      ecode++;
2290      break;      break;
2291    
# Line 2199  for (;;) Line 2296  for (;;)
2296      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 */
2297        {                            /* not be updated before SCHECK_PARTIAL. */        {                            /* not be updated before SCHECK_PARTIAL. */
2298        SCHECK_PARTIAL();        SCHECK_PARTIAL();
2299        MRRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2300        }        }
2301      eptr++;      eptr++;
2302      ecode++;      ecode++;
# Line 2209  for (;;) Line 2306  for (;;)
2306      if (eptr >= md->end_subject)      if (eptr >= md->end_subject)
2307        {        {
2308        SCHECK_PARTIAL();        SCHECK_PARTIAL();
2309        MRRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2310        }        }
2311      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2312      if (      if (
2313  #ifdef SUPPORT_UTF8  #if defined SUPPORT_UTF || !(defined COMPILE_PCRE8)
2314         c < 256 &&         c < 256 &&
2315  #endif  #endif
2316         (md->ctypes[c] & ctype_digit) != 0         (md->ctypes[c] & ctype_digit) != 0
2317         )         )
2318        MRRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2319      ecode++;      ecode++;
2320      break;      break;
2321    
# Line 2226  for (;;) Line 2323  for (;;)
2323      if (eptr >= md->end_subject)      if (eptr >= md->end_subject)
2324        {        {
2325        SCHECK_PARTIAL();        SCHECK_PARTIAL();
2326        MRRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2327        }        }
2328      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2329      if (      if (
2330  #ifdef SUPPORT_UTF8  #if defined SUPPORT_UTF || !(defined COMPILE_PCRE8)
2331         c >= 256 ||         c > 255 ||
2332  #endif  #endif
2333         (md->ctypes[c] & ctype_digit) == 0         (md->ctypes[c] & ctype_digit) == 0
2334         )         )
2335        MRRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2336      ecode++;      ecode++;
2337      break;      break;
2338    
# Line 2243  for (;;) Line 2340  for (;;)
2340      if (eptr >= md->end_subject)      if (eptr >= md->end_subject)
2341        {        {
2342        SCHECK_PARTIAL();        SCHECK_PARTIAL();
2343        MRRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2344        }        }
2345      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2346      if (      if (
2347  #ifdef SUPPORT_UTF8  #if defined SUPPORT_UTF || !(defined COMPILE_PCRE8)
2348         c < 256 &&         c < 256 &&
2349  #endif  #endif
2350         (md->ctypes[c] & ctype_space) != 0         (md->ctypes[c] & ctype_space) != 0
2351         )         )
2352        MRRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2353      ecode++;      ecode++;
2354      break;      break;
2355    
# Line 2260  for (;;) Line 2357  for (;;)
2357      if (eptr >= md->end_subject)      if (eptr >= md->end_subject)
2358        {        {
2359        SCHECK_PARTIAL();        SCHECK_PARTIAL();
2360        MRRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2361        }        }
2362      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2363      if (      if (
2364  #ifdef SUPPORT_UTF8  #if defined SUPPORT_UTF || !(defined COMPILE_PCRE8)
2365         c >= 256 ||         c > 255 ||
2366  #endif  #endif
2367         (md->ctypes[c] & ctype_space) == 0         (md->ctypes[c] & ctype_space) == 0
2368         )         )
2369        MRRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2370      ecode++;      ecode++;
2371      break;      break;
2372    
# Line 2277  for (;;) Line 2374  for (;;)
2374      if (eptr >= md->end_subject)      if (eptr >= md->end_subject)
2375        {        {
2376        SCHECK_PARTIAL();        SCHECK_PARTIAL();
2377        MRRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2378        }        }
2379      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2380      if (      if (
2381  #ifdef SUPPORT_UTF8  #if defined SUPPORT_UTF || !(defined COMPILE_PCRE8)
2382         c < 256 &&         c < 256 &&
2383  #endif  #endif
2384         (md->ctypes[c] & ctype_word) != 0         (md->ctypes[c] & ctype_word) != 0
2385         )         )
2386        MRRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2387      ecode++;      ecode++;
2388      break;      break;
2389    
# Line 2294  for (;;) Line 2391  for (;;)
2391      if (eptr >= md->end_subject)      if (eptr >= md->end_subject)
2392        {        {
2393        SCHECK_PARTIAL();        SCHECK_PARTIAL();
2394        MRRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2395        }        }
2396      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2397      if (      if (
2398  #ifdef SUPPORT_UTF8  #if defined SUPPORT_UTF || !(defined COMPILE_PCRE8)
2399         c >= 256 ||         c > 255 ||
2400  #endif  #endif
2401         (md->ctypes[c] & ctype_word) == 0         (md->ctypes[c] & ctype_word) == 0
2402         )         )
2403        MRRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2404      ecode++;      ecode++;
2405      break;      break;
2406    
# Line 2311  for (;;) Line 2408  for (;;)
2408      if (eptr >= md->end_subject)      if (eptr >= md->end_subject)
2409        {        {
2410        SCHECK_PARTIAL();        SCHECK_PARTIAL();
2411        MRRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2412        }        }
2413      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2414      switch(c)      switch(c)
2415        {        {
2416        default: MRRETURN(MATCH_NOMATCH);        default: RRETURN(MATCH_NOMATCH);
2417    
2418        case 0x000d:        case CHAR_CR:
2419        if (eptr < md->end_subject && *eptr == 0x0a) eptr++;        if (eptr >= md->end_subject)
2420            {
2421            SCHECK_PARTIAL();
2422            }
2423          else if (*eptr == CHAR_LF) eptr++;
2424        break;        break;
2425    
2426        case 0x000a:        case CHAR_LF:
2427        break;        break;
2428    
2429        case 0x000b:        case CHAR_VT:
2430        case 0x000c:        case CHAR_FF:
2431        case 0x0085:        case CHAR_NEL:
2432    #ifndef EBCDIC
2433        case 0x2028:        case 0x2028:
2434        case 0x2029:        case 0x2029:
2435        if (md->bsr_anycrlf) MRRETURN(MATCH_NOMATCH);  #endif  /* Not EBCDIC */
2436          if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);
2437        break;        break;
2438        }        }
2439      ecode++;      ecode++;
# Line 2340  for (;;) Line 2443  for (;;)
2443      if (eptr >= md->end_subject)      if (eptr >= md->end_subject)
2444        {        {
2445        SCHECK_PARTIAL();        SCHECK_PARTIAL();
2446        MRRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2447        }        }
2448      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2449      switch(c)      switch(c)
2450        {        {
2451        default: break;        default: break;
2452        case 0x09:      /* HT */        case CHAR_HT:
2453        case 0x20:      /* SPACE */        case CHAR_SPACE:
2454    #ifndef EBCDIC
2455        case 0xa0:      /* NBSP */        case 0xa0:      /* NBSP */
2456        case 0x1680:    /* OGHAM SPACE MARK */        case 0x1680:    /* OGHAM SPACE MARK */
2457        case 0x180e:    /* MONGOLIAN VOWEL SEPARATOR */        case 0x180e:    /* MONGOLIAN VOWEL SEPARATOR */
# Line 2365  for (;;) Line 2469  for (;;)
2469        case 0x202f:    /* NARROW NO-BREAK SPACE */        case 0x202f:    /* NARROW NO-BREAK SPACE */
2470        case 0x205f:    /* MEDIUM MATHEMATICAL SPACE */        case 0x205f:    /* MEDIUM MATHEMATICAL SPACE */
2471        case 0x3000:    /* IDEOGRAPHIC SPACE */        case 0x3000:    /* IDEOGRAPHIC SPACE */
2472        MRRETURN(MATCH_NOMATCH);  #endif  /* Not EBCDIC */
2473          RRETURN(MATCH_NOMATCH);
2474        }        }
2475      ecode++;      ecode++;
2476      break;      break;
# Line 2374  for (;;) Line 2479  for (;;)
2479      if (eptr >= md->end_subject)      if (eptr >= md->end_subject)
2480        {        {
2481        SCHECK_PARTIAL();        SCHECK_PARTIAL();
2482        MRRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2483        }        }
2484      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2485      switch(c)      switch(c)
2486        {        {
2487        default: MRRETURN(MATCH_NOMATCH);        default: RRETURN(MATCH_NOMATCH);
2488        case 0x09:      /* HT */        case CHAR_HT:
2489        case 0x20:      /* SPACE */        case CHAR_SPACE:
2490    #ifndef EBCDIC
2491        case 0xa0:      /* NBSP */        case 0xa0:      /* NBSP */
2492        case 0x1680:    /* OGHAM SPACE MARK */        case 0x1680:    /* OGHAM SPACE MARK */
2493        case 0x180e:    /* MONGOLIAN VOWEL SEPARATOR */        case 0x180e:    /* MONGOLIAN VOWEL SEPARATOR */
# Line 2399  for (;;) Line 2505  for (;;)
2505        case 0x202f:    /* NARROW NO-BREAK SPACE */        case 0x202f:    /* NARROW NO-BREAK SPACE */
2506        case 0x205f:    /* MEDIUM MATHEMATICAL SPACE */        case 0x205f:    /* MEDIUM MATHEMATICAL SPACE */
2507        case 0x3000:    /* IDEOGRAPHIC SPACE */        case 0x3000:    /* IDEOGRAPHIC SPACE */
2508    #endif  /* Not EBCDIC */
2509        break;        break;
2510        }        }
2511      ecode++;      ecode++;
# Line 2408  for (;;) Line 2515  for (;;)
2515      if (eptr >= md->end_subject)      if (eptr >= md->end_subject)
2516        {        {
2517        SCHECK_PARTIAL();        SCHECK_PARTIAL();
2518        MRRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2519        }        }
2520      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2521      switch(c)      switch(c)
2522        {        {
2523        default: break;        default: break;
2524        case 0x0a:      /* LF */        case CHAR_LF:
2525        case 0x0b:      /* VT */        case CHAR_VT:
2526        case 0x0c:      /* FF */        case CHAR_FF:
2527        case 0x0d:      /* CR */        case CHAR_CR:
2528        case 0x85:      /* NEL */        case CHAR_NEL:
2529    #ifndef EBCDIC
2530        case 0x2028:    /* LINE SEPARATOR */        case 0x2028:    /* LINE SEPARATOR */
2531        case 0x2029:    /* PARAGRAPH SEPARATOR */        case 0x2029:    /* PARAGRAPH SEPARATOR */
2532        MRRETURN(MATCH_NOMATCH);  #endif  /* Not EBCDIC */
2533          RRETURN(MATCH_NOMATCH);
2534        }        }
2535      ecode++;      ecode++;
2536      break;      break;
# Line 2430  for (;;) Line 2539  for (;;)
2539      if (eptr >= md->end_subject)      if (eptr >= md->end_subject)
2540        {        {
2541        SCHECK_PARTIAL();        SCHECK_PARTIAL();
2542        MRRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2543        }        }
2544      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2545      switch(c)      switch(c)
2546        {        {
2547        default: MRRETURN(MATCH_NOMATCH);        default: RRETURN(MATCH_NOMATCH);
2548        case 0x0a:      /* LF */        case CHAR_LF:
2549        case 0x0b:      /* VT */        case CHAR_VT:
2550        case 0x0c:      /* FF */        case CHAR_FF:
2551        case 0x0d:      /* CR */        case CHAR_CR:
2552        case 0x85:      /* NEL */        case CHAR_NEL:
2553    #ifndef EBCDIC
2554        case 0x2028:    /* LINE SEPARATOR */        case 0x2028:    /* LINE SEPARATOR */
2555        case 0x2029:    /* PARAGRAPH SEPARATOR */        case 0x2029:    /* PARAGRAPH SEPARATOR */
2556    #endif  /* Not EBCDIC */
2557        break;        break;
2558        }        }
2559      ecode++;      ecode++;
# Line 2457  for (;;) Line 2568  for (;;)
2568      if (eptr >= md->end_subject)      if (eptr >= md->end_subject)
2569        {        {
2570        SCHECK_PARTIAL();        SCHECK_PARTIAL();
2571        MRRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2572        }        }
2573      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2574        {        {
# Line 2466  for (;;) Line 2577  for (;;)
2577        switch(ecode[1])        switch(ecode[1])
2578          {          {
2579          case PT_ANY:          case PT_ANY:
2580          if (op == OP_NOTPROP) MRRETURN(MATCH_NOMATCH);          if (op == OP_NOTPROP) RRETURN(MATCH_NOMATCH);
2581          break;          break;
2582    
2583          case PT_LAMP:          case PT_LAMP:
2584          if ((prop->chartype == ucp_Lu ||          if ((prop->chartype == ucp_Lu ||
2585               prop->chartype == ucp_Ll ||               prop->chartype == ucp_Ll ||
2586               prop->chartype == ucp_Lt) == (op == OP_NOTPROP))               prop->chartype == ucp_Lt) == (op == OP_NOTPROP))
2587            MRRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
2588          break;          break;
2589    
2590          case PT_GC:          case PT_GC:
2591          if ((ecode[2] != _pcre_ucp_gentype[prop->chartype]) == (op == OP_PROP))          if ((ecode[2] != PRIV(ucp_gentype)[prop->chartype]) == (op == OP_PROP))
2592            MRRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
2593          break;          break;
2594    
2595          case PT_PC:          case PT_PC:
2596          if ((ecode[2] != prop->chartype) == (op == OP_PROP))          if ((ecode[2] != prop->chartype) == (op == OP_PROP))
2597            MRRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
2598          break;          break;
2599    
2600          case PT_SC:          case PT_SC:
2601          if ((ecode[2] != prop->script) == (op == OP_PROP))          if ((ecode[2] != prop->script) == (op == OP_PROP))
2602            MRRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
2603          break;          break;
2604    
2605          /* These are specials */          /* These are specials */
2606    
2607          case PT_ALNUM:          case PT_ALNUM:
2608          if ((_pcre_ucp_gentype[prop->chartype] == ucp_L ||          if ((PRIV(ucp_gentype)[prop->chartype] == ucp_L ||
2609               _pcre_ucp_gentype[prop->chartype] == ucp_N) == (op == OP_NOTPROP))               PRIV(ucp_gentype)[prop->chartype] == ucp_N) == (op == OP_NOTPROP))
2610            MRRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
2611          break;          break;
2612    
2613          case PT_SPACE:    /* Perl space */          case PT_SPACE:    /* Perl space */
2614          if ((_pcre_ucp_gentype[prop->chartype] == ucp_Z ||          if ((PRIV(ucp_gentype)[prop->chartype] == ucp_Z ||
2615               c == CHAR_HT || c == CHAR_NL || c == CHAR_FF || c == CHAR_CR)               c == CHAR_HT || c == CHAR_NL || c == CHAR_FF || c == CHAR_CR)
2616                 == (op == OP_NOTPROP))                 == (op == OP_NOTPROP))
2617            MRRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
2618          break;          break;
2619    
2620          case PT_PXSPACE:  /* POSIX space */          case PT_PXSPACE:  /* POSIX space */
2621          if ((_pcre_ucp_gentype[prop->chartype] == ucp_Z ||          if ((PRIV(ucp_gentype)[prop->chartype] == ucp_Z ||
2622               c == CHAR_HT || c == CHAR_NL || c == CHAR_VT ||               c == CHAR_HT || c == CHAR_NL || c == CHAR_VT ||
2623               c == CHAR_FF || c == CHAR_CR)               c == CHAR_FF || c == CHAR_CR)
2624                 == (op == OP_NOTPROP))                 == (op == OP_NOTPROP))
2625            MRRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
2626          break;          break;
2627    
2628          case PT_WORD:          case PT_WORD:
2629          if ((_pcre_ucp_gentype[prop->chartype] == ucp_L ||          if ((PRIV(ucp_gentype)[prop->chartype] == ucp_L ||
2630               _pcre_ucp_gentype[prop->chartype] == ucp_N ||               PRIV(ucp_gentype)[prop->chartype] == ucp_N ||
2631               c == CHAR_UNDERSCORE) == (op == OP_NOTPROP))               c == CHAR_UNDERSCORE) == (op == OP_NOTPROP))
2632            MRRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
2633          break;          break;
2634    
2635          /* This should never occur */          /* This should never occur */
# Line 2538  for (;;) Line 2649  for (;;)
2649      if (eptr >= md->end_subject)      if (eptr >= md->end_subject)
2650        {        {
2651        SCHECK_PARTIAL();        SCHECK_PARTIAL();
2652        MRRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2653        }        }
2654      GETCHARINCTEST(c, eptr);      else
2655      if (UCD_CATEGORY(c) == ucp_M) MRRETURN(MATCH_NOMATCH);        {
2656      while (eptr < md->end_subject)        int lgb, rgb;
2657        {        GETCHARINCTEST(c, eptr);
2658        int len = 1;        lgb = UCD_GRAPHBREAK(c);
2659        if (!utf8) c = *eptr; else { GETCHARLEN(c, eptr, len); }        while (eptr < md->end_subject)
2660        if (UCD_CATEGORY(c) != ucp_M) break;          {
2661        eptr += len;          int len = 1;
2662            if (!utf) c = *eptr; else { GETCHARLEN(c, eptr, len); }
2663            rgb = UCD_GRAPHBREAK(c);
2664            if ((PRIV(ucp_gbtable)[lgb] & (1 << rgb)) == 0) break;
2665            lgb = rgb;
2666            eptr += len;
2667            }
2668        }        }
2669        CHECK_PARTIAL();
2670      ecode++;      ecode++;
2671      break;      break;
2672  #endif  #endif
# Line 2566  for (;;) Line 2684  for (;;)
2684      case OP_REFI:      case OP_REFI:
2685      caseless = op == OP_REFI;      caseless = op == OP_REFI;
2686      offset = GET2(ecode, 1) << 1;               /* Doubled ref number */      offset = GET2(ecode, 1) << 1;               /* Doubled ref number */
2687      ecode += 3;      ecode += 1 + IMM2_SIZE;
2688    
2689      /* If the reference is unset, there are two possibilities:      /* If the reference is unset, there are two possibilities:
2690    
# Line 2606  for (;;) Line 2724  for (;;)
2724        case OP_CRMINRANGE:        case OP_CRMINRANGE:
2725        minimize = (*ecode == OP_CRMINRANGE);        minimize = (*ecode == OP_CRMINRANGE);
2726        min = GET2(ecode, 1);        min = GET2(ecode, 1);
2727        max = GET2(ecode, 3);        max = GET2(ecode, 1 + IMM2_SIZE);
2728        if (max == 0) max = INT_MAX;        if (max == 0) max = INT_MAX;
2729        ecode += 5;        ecode += 1 + 2 * IMM2_SIZE;
2730        break;        break;
2731    
2732        default:               /* No repeat follows */        default:               /* No repeat follows */
2733        if ((length = match_ref(offset, eptr, length, md, caseless)) < 0)        if ((length = match_ref(offset, eptr, length, md, caseless)) < 0)
2734          {          {
2735            if (length == -2) eptr = md->end_subject;   /* Partial match */
2736          CHECK_PARTIAL();          CHECK_PARTIAL();
2737          MRRETURN(MATCH_NOMATCH);          RRETURN(MATCH_NOMATCH);
2738          }          }
2739        eptr += length;        eptr += length;
2740        continue;              /* With the main loop */        continue;              /* With the main loop */
2741        }        }
2742    
2743      /* Handle repeated back references. If the length of the reference is      /* Handle repeated back references. If the length of the reference is
2744      zero, just continue with the main loop. */      zero, just continue with the main loop. If the length is negative, it
2745        means the reference is unset in non-Java-compatible mode. If the minimum is
2746        zero, we can continue at the same level without recursion. For any other
2747        minimum, carrying on will result in NOMATCH. */
2748    
2749      if (length == 0) continue;      if (length == 0) continue;
2750        if (length < 0 && min == 0) continue;
2751    
2752      /* First, ensure the minimum number of matches are present. We get back      /* First, ensure the minimum number of matches are present. We get back
2753      the length of the reference string explicitly rather than passing the      the length of the reference string explicitly rather than passing the
# Line 2635  for (;;) Line 2758  for (;;)
2758        int slength;        int slength;
2759        if ((slength = match_ref(offset, eptr, length, md, caseless)) < 0)        if ((slength = match_ref(offset, eptr, length, md, caseless)) < 0)
2760          {          {
2761            if (slength == -2) eptr = md->end_subject;   /* Partial match */
2762          CHECK_PARTIAL();          CHECK_PARTIAL();
2763          MRRETURN(MATCH_NOMATCH);          RRETURN(MATCH_NOMATCH);
2764          }          }
2765        eptr += slength;        eptr += slength;
2766        }        }
# Line 2655  for (;;) Line 2779  for (;;)
2779          int slength;          int slength;
2780          RMATCH(eptr, ecode, offset_top, md, eptrb, RM14);          RMATCH(eptr, ecode, offset_top, md, eptrb, RM14);
2781          if (rrc != MATCH_NOMATCH) RRETURN(rrc);          if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2782          if (fi >= max) MRRETURN(MATCH_NOMATCH);          if (fi >= max) RRETURN(MATCH_NOMATCH);
2783          if ((slength = match_ref(offset, eptr, length, md, caseless)) < 0)          if ((slength = match_ref(offset, eptr, length, md, caseless)) < 0)
2784            {            {
2785              if (slength == -2) eptr = md->end_subject;   /* Partial match */
2786            CHECK_PARTIAL();            CHECK_PARTIAL();
2787            MRRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
2788            }            }
2789          eptr += slength;          eptr += slength;
2790          }          }
# Line 2676  for (;;) Line 2801  for (;;)
2801          int slength;          int slength;
2802          if ((slength = match_ref(offset, eptr, length, md, caseless)) < 0)          if ((slength = match_ref(offset, eptr, length, md, caseless)) < 0)
2803            {            {
2804            CHECK_PARTIAL();            /* Can't use CHECK_PARTIAL because we don't want to update eptr in
2805              the soft partial matching case. */
2806    
2807              if (slength == -2 && md->partial != 0 &&
2808                  md->end_subject > md->start_used_ptr)
2809                {
2810                md->hitend = TRUE;
2811                if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
2812                }
2813            break;            break;
2814            }            }
2815          eptr += slength;          eptr += slength;
2816          }          }
2817    
2818        while (eptr >= pp)        while (eptr >= pp)
2819          {          {
2820          RMATCH(eptr, ecode, offset_top, md, eptrb, RM15);          RMATCH(eptr, ecode, offset_top, md, eptrb, RM15);
2821          if (rrc != MATCH_NOMATCH) RRETURN(rrc);          if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2822          eptr -= length;          eptr -= length;
2823          }          }
2824        MRRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2825        }        }
2826      /* Control never gets here */      /* Control never gets here */
2827    
# Line 2705  for (;;) Line 2839  for (;;)
2839      case OP_NCLASS:      case OP_NCLASS:
2840      case OP_CLASS:      case OP_CLASS:
2841        {        {
2842          /* The data variable is saved across frames, so the byte map needs to
2843          be stored there. */
2844    #define BYTE_MAP ((pcre_uint8 *)data)
2845        data = ecode + 1;                /* Save for matching */        data = ecode + 1;                /* Save for matching */
2846        ecode += 33;                     /* Advance past the item */        ecode += 1 + (32 / sizeof(pcre_uchar)); /* Advance past the item */
2847    
2848        switch (*ecode)        switch (*ecode)
2849          {          {
# Line 2727  for (;;) Line 2864  for (;;)
2864          case OP_CRMINRANGE:          case OP_CRMINRANGE:
2865          minimize = (*ecode == OP_CRMINRANGE);          minimize = (*ecode == OP_CRMINRANGE);
2866          min = GET2(ecode, 1);          min = GET2(ecode, 1);
2867          max = GET2(ecode, 3);          max = GET2(ecode, 1 + IMM2_SIZE);
2868          if (max == 0) max = INT_MAX;          if (max == 0) max = INT_MAX;
2869          ecode += 5;          ecode += 1 + 2 * IMM2_SIZE;
2870          break;          break;
2871    
2872          default:               /* No repeat follows */          default:               /* No repeat follows */
# Line 2739  for (;;) Line 2876  for (;;)
2876    
2877        /* First, ensure the minimum number of matches are present. */        /* First, ensure the minimum number of matches are present. */
2878    
2879  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
2880        /* UTF-8 mode */        if (utf)
       if (utf8)  
2881          {          {
2882          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
2883            {            {
2884            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
2885              {              {
2886              SCHECK_PARTIAL();              SCHECK_PARTIAL();
2887              MRRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
2888              }              }
2889            GETCHARINC(c, eptr);            GETCHARINC(c, eptr);
2890            if (c > 255)            if (c > 255)
2891              {              {
2892              if (op == OP_CLASS) MRRETURN(MATCH_NOMATCH);              if (op == OP_CLASS) RRETURN(MATCH_NOMATCH);
2893              }              }
2894            else            else
2895              {              if ((BYTE_MAP[c/8] & (1 << (c&7))) == 0) RRETURN(MATCH_NOMATCH);
             if ((data[c/8] & (1 << (c&7))) == 0) MRRETURN(MATCH_NOMATCH);  
             }  
2896            }            }
2897          }          }
2898        else        else
2899  #endif  #endif
2900        /* Not UTF-8 mode */        /* Not UTF mode */
2901          {          {
2902          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
2903            {            {
2904            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
2905              {              {
2906              SCHECK_PARTIAL();              SCHECK_PARTIAL();
2907              MRRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
2908              }              }
2909            c = *eptr++;            c = *eptr++;
2910            if ((data[c/8] & (1 << (c&7))) == 0) MRRETURN(MATCH_NOMATCH);  #ifndef COMPILE_PCRE8
2911              if (c > 255)
2912                {
2913                if (op == OP_CLASS) RRETURN(MATCH_NOMATCH);
2914                }
2915              else
2916    #endif
2917                if ((BYTE_MAP[c/8] & (1 << (c&7))) == 0) RRETURN(MATCH_NOMATCH);
2918            }            }
2919          }          }
2920    
# Line 2787  for (;;) Line 2928  for (;;)
2928    
2929        if (minimize)        if (minimize)
2930          {          {
2931  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
2932          /* UTF-8 mode */          if (utf)
         if (utf8)  
2933            {            {
2934            for (fi = min;; fi++)            for (fi = min;; fi++)
2935              {              {
2936              RMATCH(eptr, ecode, offset_top, md, eptrb, RM16);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM16);
2937              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2938              if (fi >= max) MRRETURN(MATCH_NOMATCH);              if (fi >= max) RRETURN(MATCH_NOMATCH);
2939              if (eptr >= md->end_subject)              if (eptr >= md->end_subject)
2940                {                {
2941                SCHECK_PARTIAL();                SCHECK_PARTIAL();
2942                MRRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
2943                }                }
2944              GETCHARINC(c, eptr);              GETCHARINC(c, eptr);
2945              if (c > 255)              if (c > 255)
2946                {                {
2947                if (op == OP_CLASS) MRRETURN(MATCH_NOMATCH);                if (op == OP_CLASS) RRETURN(MATCH_NOMATCH);
2948                }                }
2949              else              else
2950                {                if ((BYTE_MAP[c/8] & (1 << (c&7))) == 0) RRETURN(MATCH_NOMATCH);
               if ((data[c/8] & (1 << (c&7))) == 0) MRRETURN(MATCH_NOMATCH);  
               }  
2951              }              }
2952            }            }
2953          else          else
2954  #endif  #endif
2955          /* Not UTF-8 mode */          /* Not UTF mode */
2956            {            {
2957            for (fi = min;; fi++)            for (fi = min;; fi++)
2958              {              {
2959              RMATCH(eptr, ecode, offset_top, md, eptrb, RM17);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM17);
2960              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2961              if (fi >= max) MRRETURN(MATCH_NOMATCH);              if (fi >= max) RRETURN(MATCH_NOMATCH);
2962              if (eptr >= md->end_subject)              if (eptr >= md->end_subject)
2963                {                {
2964                SCHECK_PARTIAL();                SCHECK_PARTIAL();
2965                MRRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
2966                }                }
2967              c = *eptr++;              c = *eptr++;
2968              if ((data[c/8] & (1 << (c&7))) == 0) MRRETURN(MATCH_NOMATCH);  #ifndef COMPILE_PCRE8
2969                if (c > 255)
2970                  {
2971                  if (op == OP_CLASS) RRETURN(MATCH_NOMATCH);
2972                  }
2973                else
2974    #endif
2975                  if ((BYTE_MAP[c/8] & (1 << (c&7))) == 0) RRETURN(MATCH_NOMATCH);
2976              }              }
2977            }            }
2978          /* Control never gets here */          /* Control never gets here */
# Line 2839  for (;;) Line 2984  for (;;)
2984          {          {
2985          pp = eptr;          pp = eptr;
2986    
2987  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
2988          /* UTF-8 mode */          if (utf)
         if (utf8)  
2989            {            {
2990            for (i = min; i < max; i++)            for (i = min; i < max; i++)
2991              {              {
# Line 2857  for (;;) Line 3001  for (;;)
3001                if (op == OP_CLASS) break;                if (op == OP_CLASS) break;
3002                }                }
3003              else              else
3004                {                if ((BYTE_MAP[c/8] & (1 << (c&7))) == 0) break;
               if ((data[c/8] & (1 << (c&7))) == 0) break;  
               }  
3005              eptr += len;              eptr += len;
3006              }              }
3007            for (;;)            for (;;)
# Line 2872  for (;;) Line 3014  for (;;)
3014            }            }
3015          else          else
3016  #endif  #endif
3017            /* Not UTF-8 mode */            /* Not UTF mode */
3018            {            {
3019            for (i = min; i < max; i++)            for (i = min; i < max; i++)
3020              {              {
# Line 2882  for (;;) Line 3024  for (;;)
3024                break;                break;
3025                }                }
3026              c = *eptr;              c = *eptr;
3027              if ((data[c/8] & (1 << (c&7))) == 0) break;  #ifndef COMPILE_PCRE8
3028                if (c > 255)
3029                  {
3030                  if (op == OP_CLASS) break;
3031                  }
3032                else
3033    #endif
3034                  if ((BYTE_MAP[c/8] & (1 << (c&7))) == 0) break;
3035              eptr++;              eptr++;
3036              }              }
3037            while (eptr >= pp)            while (eptr >= pp)
# Line 2893  for (;;) Line 3042  for (;;)
3042              }              }
3043            }            }
3044    
3045          MRRETURN(MATCH_NOMATCH);          RRETURN(MATCH_NOMATCH);
3046          }          }
3047    #undef BYTE_MAP
3048        }        }
3049      /* Control never gets here */      /* Control never gets here */
3050    
# Line 2903  for (;;) Line 3053  for (;;)
3053      when UTF-8 mode mode is supported. Nevertheless, we may not be in UTF-8      when UTF-8 mode mode is supported. Nevertheless, we may not be in UTF-8
3054      mode, because Unicode properties are supported in non-UTF-8 mode. */      mode, because Unicode properties are supported in non-UTF-8 mode. */
3055    
3056  #ifdef SUPPORT_UTF8  #if defined SUPPORT_UTF || !defined COMPILE_PCRE8
3057      case OP_XCLASS:      case OP_XCLASS:
3058        {        {
3059        data = ecode + 1 + LINK_SIZE;                /* Save for matching */        data = ecode + 1 + LINK_SIZE;                /* Save for matching */
# Line 2928  for (;;) Line 3078  for (;;)
3078          case OP_CRMINRANGE:          case OP_CRMINRANGE:
3079          minimize = (*ecode == OP_CRMINRANGE);          minimize = (*ecode == OP_CRMINRANGE);
3080          min = GET2(ecode, 1);          min = GET2(ecode, 1);
3081          max = GET2(ecode, 3);          max = GET2(ecode, 1 + IMM2_SIZE);
3082          if (max == 0) max = INT_MAX;          if (max == 0) max = INT_MAX;
3083          ecode += 5;          ecode += 1 + 2 * IMM2_SIZE;
3084          break;          break;
3085    
3086          default:               /* No repeat follows */          default:               /* No repeat follows */
# Line 2945  for (;;) Line 3095  for (;;)
3095          if (eptr >= md->end_subject)          if (eptr >= md->end_subject)
3096            {            {
3097            SCHECK_PARTIAL();            SCHECK_PARTIAL();
3098            MRRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
3099            }            }
3100          GETCHARINCTEST(c, eptr);          GETCHARINCTEST(c, eptr);
3101          if (!_pcre_xclass(c, data)) MRRETURN(MATCH_NOMATCH);          if (!PRIV(xclass)(c, data, utf)) RRETURN(MATCH_NOMATCH);
3102          }          }
3103    
3104        /* 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 2965  for (;;) Line 3115  for (;;)
3115            {            {
3116            RMATCH(eptr, ecode, offset_top, md, eptrb, RM20);            RMATCH(eptr, ecode, offset_top, md, eptrb, RM20);
3117            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3118            if (fi >= max) MRRETURN(MATCH_NOMATCH);            if (fi >= max) RRETURN(MATCH_NOMATCH);
3119            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
3120              {              {
3121              SCHECK_PARTIAL();              SCHECK_PARTIAL();
3122              MRRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
3123              }              }
3124            GETCHARINCTEST(c, eptr);            GETCHARINCTEST(c, eptr);
3125            if (!_pcre_xclass(c, data)) MRRETURN(MATCH_NOMATCH);            if (!PRIV(xclass)(c, data, utf)) RRETURN(MATCH_NOMATCH);
3126            }            }
3127          /* Control never gets here */          /* Control never gets here */
3128          }          }
# Line 2990  for (;;) Line 3140  for (;;)
3140              SCHECK_PARTIAL();              SCHECK_PARTIAL();
3141              break;              break;
3142              }              }
3143    #ifdef SUPPORT_UTF
3144            GETCHARLENTEST(c, eptr, len);            GETCHARLENTEST(c, eptr, len);
3145            if (!_pcre_xclass(c, data)) break;  #else
3146              c = *eptr;
3147    #endif
3148              if (!PRIV(xclass)(c, data, utf)) break;
3149            eptr += len;            eptr += len;
3150            }            }
3151          for(;;)          for(;;)
# Line 2999  for (;;) Line 3153  for (;;)
3153            RMATCH(eptr, ecode, offset_top, md, eptrb, RM21);            RMATCH(eptr, ecode, offset_top, md, eptrb, RM21);
3154            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3155            if (eptr-- == pp) break;        /* Stop if tried at original pos */            if (eptr-- == pp) break;        /* Stop if tried at original pos */
3156            if (utf8) BACKCHAR(eptr);  #ifdef SUPPORT_UTF
3157              if (utf) BACKCHAR(eptr);
3158    #endif
3159            }            }
3160          MRRETURN(MATCH_NOMATCH);          RRETURN(MATCH_NOMATCH);
3161          }          }
3162    
3163        /* Control never gets here */        /* Control never gets here */
# Line 3011  for (;;) Line 3167  for (;;)
3167      /* Match a single character, casefully */      /* Match a single character, casefully */
3168    
3169      case OP_CHAR:      case OP_CHAR:
3170  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
3171      if (utf8)      if (utf)
3172        {        {
3173        length = 1;        length = 1;
3174        ecode++;        ecode++;
# Line 3020  for (;;) Line 3176  for (;;)
3176        if (length > md->end_subject - eptr)        if (length > md->end_subject - eptr)
3177          {          {
3178          CHECK_PARTIAL();             /* Not SCHECK_PARTIAL() */          CHECK_PARTIAL();             /* Not SCHECK_PARTIAL() */
3179          MRRETURN(MATCH_NOMATCH);          RRETURN(MATCH_NOMATCH);
3180          }          }
3181        while (length-- > 0) if (*ecode++ != *eptr++) MRRETURN(MATCH_NOMATCH);        while (length-- > 0) if (*ecode++ != *eptr++) RRETURN(MATCH_NOMATCH);
3182        }        }
3183      else      else
3184  #endif  #endif
3185        /* Not UTF mode */
     /* Non-UTF-8 mode */  
3186        {        {
3187        if (md->end_subject - eptr < 1)        if (md->end_subject - eptr < 1)
3188          {          {
3189          SCHECK_PARTIAL();            /* This one can use SCHECK_PARTIAL() */          SCHECK_PARTIAL();            /* This one can use SCHECK_PARTIAL() */
3190          MRRETURN(MATCH_NOMATCH);          RRETURN(MATCH_NOMATCH);
3191          }          }
3192        if (ecode[1] != *eptr++) MRRETURN(MATCH_NOMATCH);        if (ecode[1] != *eptr++) RRETURN(MATCH_NOMATCH);
3193        ecode += 2;        ecode += 2;
3194        }        }
3195      break;      break;
3196    
3197      /* Match a single character, caselessly */      /* Match a single character, caselessly. If we are at the end of the
3198        subject, give up immediately. */
3199    
3200      case OP_CHARI:      case OP_CHARI:
3201  #ifdef SUPPORT_UTF8      if (eptr >= md->end_subject)
3202      if (utf8)        {
3203          SCHECK_PARTIAL();
3204          RRETURN(MATCH_NOMATCH);
3205          }
3206    
3207    #ifdef SUPPORT_UTF
3208        if (utf)
3209        {        {
3210        length = 1;        length = 1;
3211        ecode++;        ecode++;
3212        GETCHARLEN(fc, ecode, length);        GETCHARLEN(fc, ecode, length);
3213    
       if (length > md->end_subject - eptr)  
         {  
         CHECK_PARTIAL();             /* Not SCHECK_PARTIAL() */  
         MRRETURN(MATCH_NOMATCH);  
         }  
   
3214        /* 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
3215        can use the fast lookup table. */        we know that its other case must also be one byte long, so we can use the
3216          fast lookup table. We know that there is at least one byte left in the
3217          subject. */
3218    
3219        if (fc < 128)        if (fc < 128)
3220          {          {
3221          if (md->lcc[*ecode++] != md->lcc[*eptr++]) MRRETURN(MATCH_NOMATCH);          if (md->lcc[fc]
3222                != TABLE_GET(*eptr, md->lcc, *eptr)) RRETURN(MATCH_NOMATCH);
3223            ecode++;
3224            eptr++;
3225          }          }
3226    
3227        /* Otherwise we must pick up the subject character */        /* Otherwise we must pick up the subject character. Note that we cannot
3228          use the value of "length" to check for sufficient bytes left, because the
3229          other case of the character may have more or fewer bytes.  */
3230    
3231        else        else
3232          {          {
# Line 3079  for (;;) Line 3242  for (;;)
3242  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
3243            if (dc != UCD_OTHERCASE(fc))            if (dc != UCD_OTHERCASE(fc))
3244  #endif  #endif
3245              MRRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
3246            }            }
3247          }          }
3248        }        }
3249      else      else
3250  #endif   /* SUPPORT_UTF8 */  #endif   /* SUPPORT_UTF */
3251    
3252      /* Non-UTF-8 mode */      /* Not UTF mode */
3253        {        {
3254        if (md->end_subject - eptr < 1)        if (TABLE_GET(ecode[1], md->lcc, ecode[1])
3255          {            != TABLE_GET(*eptr, md->lcc, *eptr)) RRETURN(MATCH_NOMATCH);
3256          SCHECK_PARTIAL();            /* This one can use SCHECK_PARTIAL() */        eptr++;
         MRRETURN(MATCH_NOMATCH);  
         }  
       if (md->lcc[ecode[1]] != md->lcc[*eptr++]) MRRETURN(MATCH_NOMATCH);  
3257        ecode += 2;        ecode += 2;
3258        }        }
3259      break;      break;
# Line 3103  for (;;) Line 3263  for (;;)
3263      case OP_EXACT:      case OP_EXACT:
3264      case OP_EXACTI:      case OP_EXACTI:
3265      min = max = GET2(ecode, 1);      min = max = GET2(ecode, 1);
3266      ecode += 3;      ecode += 1 + IMM2_SIZE;
3267      goto REPEATCHAR;      goto REPEATCHAR;
3268    
3269      case OP_POSUPTO:      case OP_POSUPTO:
# Line 3118  for (;;) Line 3278  for (;;)
3278      min = 0;      min = 0;
3279      max = GET2(ecode, 1);      max = GET2(ecode, 1);
3280      minimize = *ecode == OP_MINUPTO || *ecode == OP_MINUPTOI;      minimize = *ecode == OP_MINUPTO || *ecode == OP_MINUPTOI;
3281      ecode += 3;      ecode += 1 + IMM2_SIZE;
3282      goto REPEATCHAR;      goto REPEATCHAR;
3283    
3284      case OP_POSSTAR:      case OP_POSSTAR:
# Line 3166  for (;;) Line 3326  for (;;)
3326      /* Common code for all repeated single-character matches. */      /* Common code for all repeated single-character matches. */
3327    
3328      REPEATCHAR:      REPEATCHAR:
3329  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
3330      if (utf8)      if (utf)
3331        {        {
3332        length = 1;        length = 1;
3333        charptr = ecode;        charptr = ecode;
# Line 3183  for (;;) Line 3343  for (;;)
3343          unsigned int othercase;          unsigned int othercase;
3344          if (op >= OP_STARI &&     /* Caseless */          if (op >= OP_STARI &&     /* Caseless */
3345              (othercase = UCD_OTHERCASE(fc)) != fc)              (othercase = UCD_OTHERCASE(fc)) != fc)
3346            oclength = _pcre_ord2utf8(othercase, occhars);            oclength = PRIV(ord2utf)(othercase, occhars);
3347          else oclength = 0;          else oclength = 0;
3348  #endif  /* SUPPORT_UCP */  #endif  /* SUPPORT_UCP */
3349    
3350          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
3351            {            {
3352            if (eptr <= md->end_subject - length &&            if (eptr <= md->end_subject - length &&
3353              memcmp(eptr, charptr, length) == 0) eptr += length;              memcmp(eptr, charptr, IN_UCHARS(length)) == 0) eptr += length;
3354  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
3355            else if (oclength > 0 &&            else if (oclength > 0 &&
3356                     eptr <= md->end_subject - oclength &&                     eptr <= md->end_subject - oclength &&
3357                     memcmp(eptr, occhars, oclength) == 0) eptr += oclength;                     memcmp(eptr, occhars, IN_UCHARS(oclength)) == 0) eptr += oclength;
3358  #endif  /* SUPPORT_UCP */  #endif  /* SUPPORT_UCP */
3359            else            else
3360              {              {
3361              CHECK_PARTIAL();              CHECK_PARTIAL();
3362              MRRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
3363              }              }
3364            }            }
3365    
# Line 3211  for (;;) Line 3371  for (;;)
3371              {              {
3372              RMATCH(eptr, ecode, offset_top, md, eptrb, RM22);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM22);
3373              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3374              if (fi >= max) MRRETURN(MATCH_NOMATCH);              if (fi >= max) RRETURN(MATCH_NOMATCH);
3375              if (eptr <= md->end_subject - length &&              if (eptr <= md->end_subject - length &&
3376                memcmp(eptr, charptr, length) == 0) eptr += length;                memcmp(eptr, charptr, IN_UCHARS(length)) == 0) eptr += length;
3377  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
3378              else if (oclength > 0 &&              else if (oclength > 0 &&
3379                       eptr <= md->end_subject - oclength &&                       eptr <= md->end_subject - oclength &&
3380                       memcmp(eptr, occhars, oclength) == 0) eptr += oclength;                       memcmp(eptr, occhars, IN_UCHARS(oclength)) == 0) eptr += oclength;
3381  #endif  /* SUPPORT_UCP */  #endif  /* SUPPORT_UCP */
3382              else              else
3383                {                {
3384                CHECK_PARTIAL();                CHECK_PARTIAL();
3385                MRRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
3386                }                }
3387              }              }
3388            /* Control never gets here */            /* Control never gets here */
# Line 3234  for (;;) Line 3394  for (;;)
3394            for (i = min; i < max; i++)            for (i = min; i < max; i++)
3395              {              {
3396              if (eptr <= md->end_subject - length &&              if (eptr <= md->end_subject - length &&
3397                  memcmp(eptr, charptr, length) == 0) eptr += length;                  memcmp(eptr, charptr, IN_UCHARS(length)) == 0) eptr += length;
3398  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
3399              else if (oclength > 0 &&              else if (oclength > 0 &&
3400                       eptr <= md->end_subject - oclength &&                       eptr <= md->end_subject - oclength &&
3401                       memcmp(eptr, occhars, oclength) == 0) eptr += oclength;                       memcmp(eptr, occhars, IN_UCHARS(oclength)) == 0) eptr += oclength;
3402  #endif  /* SUPPORT_UCP */  #endif  /* SUPPORT_UCP */
3403              else              else
3404                {                {
# Line 3253  for (;;) Line 3413  for (;;)
3413              {              {
3414              RMATCH(eptr, ecode, offset_top, md, eptrb, RM23);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM23);
3415              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3416              if (eptr == pp) { MRRETURN(MATCH_NOMATCH); }              if (eptr == pp) { RRETURN(MATCH_NOMATCH); }
3417  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
3418              eptr--;              eptr--;
3419              BACKCHAR(eptr);              BACKCHAR(eptr);
# Line 3270  for (;;) Line 3430  for (;;)
3430        value of fc will always be < 128. */        value of fc will always be < 128. */
3431        }        }
3432      else      else
3433  #endif  /* SUPPORT_UTF8 */  #endif  /* SUPPORT_UTF */
3434          /* When not in UTF-8 mode, load a single-byte character. */
3435      /* When not in UTF-8 mode, load a single-byte character. */        fc = *ecode++;
3436    
3437      fc = *ecode++;      /* The value of fc at this point is always one character, though we may
3438        or may not be in UTF mode. The code is duplicated for the caseless and
     /* The value of fc at this point is always less than 256, though we may or  
     may not be in UTF-8 mode. The code is duplicated for the caseless and  
3439      caseful cases, for speed, since matching characters is likely to be quite      caseful cases, for speed, since matching characters is likely to be quite
3440      common. First, ensure the minimum number of matches are present. If min =      common. First, ensure the minimum number of matches are present. If min =
3441      max, continue at the same level without recursing. Otherwise, if      max, continue at the same level without recursing. Otherwise, if
# Line 3286  for (;;) Line 3444  for (;;)
3444      maximizing, find the maximum number of characters and work backwards. */      maximizing, find the maximum number of characters and work backwards. */
3445    
3446      DPRINTF(("matching %c{%d,%d} against subject %.*s\n", fc, min, max,      DPRINTF(("matching %c{%d,%d} against subject %.*s\n", fc, min, max,
3447        max, eptr));        max, (char *)eptr));
3448    
3449      if (op >= OP_STARI)  /* Caseless */      if (op >= OP_STARI)  /* Caseless */
3450        {        {
3451        fc = md->lcc[fc];  #ifdef COMPILE_PCRE8
3452          /* fc must be < 128 if UTF is enabled. */
3453          foc = md->fcc[fc];
3454    #else
3455    #ifdef SUPPORT_UTF
3456    #ifdef SUPPORT_UCP
3457          if (utf && fc > 127)
3458            foc = UCD_OTHERCASE(fc);
3459    #else
3460          if (utf && fc > 127)
3461            foc = fc;
3462    #endif /* SUPPORT_UCP */
3463          else
3464    #endif /* SUPPORT_UTF */
3465            foc = TABLE_GET(fc, md->fcc, fc);
3466    #endif /* COMPILE_PCRE8 */
3467    
3468        for (i = 1; i <= min; i++)        for (i = 1; i <= min; i++)
3469          {          {
3470          if (eptr >= md->end_subject)          if (eptr >= md->end_subject)
3471            {            {
3472            SCHECK_PARTIAL();            SCHECK_PARTIAL();
3473            MRRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
3474            }            }
3475          if (fc != md->lcc[*eptr++]) MRRETURN(MATCH_NOMATCH);          if (fc != *eptr && foc != *eptr) RRETURN(MATCH_NOMATCH);
3476            eptr++;
3477          }          }
3478        if (min == max) continue;        if (min == max) continue;
3479        if (minimize)        if (minimize)
# Line 3307  for (;;) Line 3482  for (;;)
3482            {            {
3483            RMATCH(eptr, ecode, offset_top, md, eptrb, RM24);            RMATCH(eptr, ecode, offset_top, md, eptrb, RM24);
3484            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3485            if (fi >= max) MRRETURN(MATCH_NOMATCH);            if (fi >= max) RRETURN(MATCH_NOMATCH);
3486            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
3487              {              {
3488              SCHECK_PARTIAL();              SCHECK_PARTIAL();
3489              MRRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
3490              }              }
3491            if (fc != md->lcc[*eptr++]) MRRETURN(MATCH_NOMATCH);            if (fc != *eptr && foc != *eptr) RRETURN(MATCH_NOMATCH);
3492              eptr++;
3493            }            }
3494          /* Control never gets here */          /* Control never gets here */
3495          }          }
# Line 3327  for (;;) Line 3503  for (;;)
3503              SCHECK_PARTIAL();              SCHECK_PARTIAL();
3504              break;              break;
3505              }              }
3506            if (fc != md->lcc[*eptr]) break;            if (fc != *eptr && foc != *eptr) break;
3507            eptr++;            eptr++;
3508            }            }
3509    
# Line 3339  for (;;) Line 3515  for (;;)
3515            eptr--;            eptr--;
3516            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3517            }            }
3518          MRRETURN(MATCH_NOMATCH);          RRETURN(MATCH_NOMATCH);
3519          }          }
3520        /* Control never gets here */        /* Control never gets here */
3521        }        }
# Line 3353  for (;;) Line 3529  for (;;)
3529          if (eptr >= md->end_subject)          if (eptr >= md->end_subject)
3530            {            {
3531            SCHECK_PARTIAL();            SCHECK_PARTIAL();
3532            MRRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
3533            }            }
3534          if (fc != *eptr++) MRRETURN(MATCH_NOMATCH);          if (fc != *eptr++) RRETURN(MATCH_NOMATCH);
3535          }          }
3536    
3537        if (min == max) continue;        if (min == max) continue;
# Line 3366  for (;;) Line 3542  for (;;)
3542            {            {
3543            RMATCH(eptr, ecode, offset_top, md, eptrb, RM26);            RMATCH(eptr, ecode, offset_top, md, eptrb, RM26);
3544            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3545            if (fi >= max) MRRETURN(MATCH_NOMATCH);            if (fi >= max) RRETURN(MATCH_NOMATCH);
3546            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
3547              {              {
3548              SCHECK_PARTIAL();              SCHECK_PARTIAL();
3549              MRRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
3550              }              }
3551            if (fc != *eptr++) MRRETURN(MATCH_NOMATCH);            if (fc != *eptr++) RRETURN(MATCH_NOMATCH);
3552            }            }
3553          /* Control never gets here */          /* Control never gets here */
3554          }          }
# Line 3397  for (;;) Line 3573  for (;;)
3573            eptr--;            eptr--;
3574            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3575            }            }
3576          MRRETURN(MATCH_NOMATCH);          RRETURN(MATCH_NOMATCH);
3577          }          }
3578        }        }
3579      /* Control never gets here */      /* Control never gets here */
# Line 3410  for (;;) Line 3586  for (;;)
3586      if (eptr >= md->end_subject)      if (eptr >= md->end_subject)
3587        {        {
3588        SCHECK_PARTIAL();        SCHECK_PARTIAL();
3589        MRRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
3590        }        }
3591      ecode++;  #ifdef SUPPORT_UTF
3592      GETCHARINCTEST(c, eptr);      if (utf)
     if (op == OP_NOTI)         /* The caseless case */  
3593        {        {
3594  #ifdef SUPPORT_UTF8        register unsigned int ch, och;
3595        if (c < 256)  
3596  #endif        ecode++;
3597        c = md->lcc[c];        GETCHARINC(ch, ecode);
3598        if (md->lcc[*ecode++] == c) MRRETURN(MATCH_NOMATCH);        GETCHARINC(c, eptr);
3599    
3600          if (op == OP_NOT)
3601            {
3602            if (ch == c) RRETURN(MATCH_NOMATCH);
3603            }
3604          else
3605            {
3606    #ifdef SUPPORT_UCP
3607            if (ch > 127)
3608              och = UCD_OTHERCASE(ch);
3609    #else
3610            if (ch > 127)
3611              och = ch;
3612    #endif /* SUPPORT_UCP */
3613            else
3614              och = TABLE_GET(ch, md->fcc, ch);
3615            if (ch == c || och == c) RRETURN(MATCH_NOMATCH);
3616            }
3617        }        }
3618      else    /* Caseful */      else
3619    #endif
3620        {        {
3621        if (*ecode++ == c) MRRETURN(MATCH_NOMATCH);        register unsigned int ch = ecode[1];
3622          c = *eptr++;
3623          if (ch == c || (op == OP_NOTI && TABLE_GET(ch, md->fcc, ch) == c))
3624            RRETURN(MATCH_NOMATCH);
3625          ecode += 2;
3626        }        }
3627      break;      break;
3628    
# Line 3438  for (;;) Line 3636  for (;;)
3636      case OP_NOTEXACT:      case OP_NOTEXACT:
3637      case OP_NOTEXACTI:      case OP_NOTEXACTI:
3638      min = max = GET2(ecode, 1);      min = max = GET2(ecode, 1);
3639      ecode += 3;      ecode += 1 + IMM2_SIZE;
3640      goto REPEATNOTCHAR;      goto REPEATNOTCHAR;
3641    
3642      case OP_NOTUPTO:      case OP_NOTUPTO:
# Line 3448  for (;;) Line 3646  for (;;)
3646      min = 0;      min = 0;
3647      max = GET2(ecode, 1);      max = GET2(ecode, 1);
3648      minimize = *ecode == OP_NOTMINUPTO || *ecode == OP_NOTMINUPTOI;      minimize = *ecode == OP_NOTMINUPTO || *ecode == OP_NOTMINUPTOI;
3649      ecode += 3;      ecode += 1 + IMM2_SIZE;
3650      goto REPEATNOTCHAR;      goto REPEATNOTCHAR;
3651    
3652      case OP_NOTPOSSTAR:      case OP_NOTPOSSTAR:
# Line 3480  for (;;) Line 3678  for (;;)
3678      possessive = TRUE;      possessive = TRUE;
3679      min = 0;      min = 0;
3680      max = GET2(ecode, 1);      max = GET2(ecode, 1);
3681      ecode += 3;      ecode += 1 + IMM2_SIZE;
3682      goto REPEATNOTCHAR;      goto REPEATNOTCHAR;
3683    
3684      case OP_NOTSTAR:      case OP_NOTSTAR:
# Line 3504  for (;;) Line 3702  for (;;)
3702      /* Common code for all repeated single-byte matches. */      /* Common code for all repeated single-byte matches. */
3703    
3704      REPEATNOTCHAR:      REPEATNOTCHAR:
3705      fc = *ecode++;      GETCHARINCTEST(fc, ecode);
3706    
3707      /* The code is duplicated for the caseless and caseful cases, for speed,      /* The code is duplicated for the caseless and caseful cases, for speed,
3708      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 3515  for (;;) Line 3713  for (;;)
3713      characters and work backwards. */      characters and work backwards. */
3714    
3715      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,
3716        max, eptr));        max, (char *)eptr));
3717    
3718      if (op >= OP_NOTSTARI)     /* Caseless */      if (op >= OP_NOTSTARI)     /* Caseless */
3719        {        {
3720        fc = md->lcc[fc];  #ifdef SUPPORT_UTF
3721    #ifdef SUPPORT_UCP
3722          if (utf && fc > 127)
3723            foc = UCD_OTHERCASE(fc);
3724    #else
3725          if (utf && fc > 127)
3726            foc = fc;
3727    #endif /* SUPPORT_UCP */
3728          else
3729    #endif /* SUPPORT_UTF */
3730            foc = TABLE_GET(fc, md->fcc, fc);
3731    
3732  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
3733        /* UTF-8 mode */        if (utf)
       if (utf8)  
3734          {          {
3735          register unsigned int d;          register unsigned int d;
3736          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
# Line 3531  for (;;) Line 3738  for (;;)
3738            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
3739              {              {
3740              SCHECK_PARTIAL();              SCHECK_PARTIAL();
3741              MRRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
3742              }              }
3743            GETCHARINC(d, eptr);            GETCHARINC(d, eptr);
3744            if (d < 256) d = md->lcc[d];            if (fc == d || (unsigned int)foc == d) RRETURN(MATCH_NOMATCH);
           if (fc == d) MRRETURN(MATCH_NOMATCH);  
3745            }            }
3746          }          }
3747        else        else
3748  #endif  #endif
3749          /* Not UTF mode */
       /* Not UTF-8 mode */  
3750          {          {
3751          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
3752            {            {
3753            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
3754              {              {
3755              SCHECK_PARTIAL();              SCHECK_PARTIAL();
3756              MRRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
3757              }              }
3758            if (fc == md->lcc[*eptr++]) MRRETURN(MATCH_NOMATCH);            if (fc == *eptr || foc == *eptr) RRETURN(MATCH_NOMATCH);
3759              eptr++;
3760            }            }
3761          }          }
3762    
# Line 3558  for (;;) Line 3764  for (;;)
3764    
3765        if (minimize)        if (minimize)
3766          {          {
3767  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
3768          /* UTF-8 mode */          if (utf)
         if (utf8)  
3769            {            {
3770            register unsigned int d;            register unsigned int d;
3771            for (fi = min;; fi++)            for (fi = min;; fi++)
3772              {              {
3773              RMATCH(eptr, ecode, offset_top, md, eptrb, RM28);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM28);
3774              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3775              if (fi >= max) MRRETURN(MATCH_NOMATCH);              if (fi >= max) RRETURN(MATCH_NOMATCH);
3776              if (eptr >= md->end_subject)              if (eptr >= md->end_subject)
3777                {                {
3778                SCHECK_PARTIAL();                SCHECK_PARTIAL();
3779                MRRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
3780                }                }
3781              GETCHARINC(d, eptr);              GETCHARINC(d, eptr);
3782              if (d < 256) d = md->lcc[d];              if (fc == d || (unsigned int)foc == d) RRETURN(MATCH_NOMATCH);
             if (fc == d) MRRETURN(MATCH_NOMATCH);  
3783              }              }
3784            }            }
3785          else          else
3786  #endif  #endif
3787          /* Not UTF-8 mode */          /* Not UTF mode */
3788            {            {
3789            for (fi = min;; fi++)            for (fi = min;; fi++)
3790              {              {
3791              RMATCH(eptr, ecode, offset_top, md, eptrb, RM29);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM29);
3792              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3793              if (fi >= max) MRRETURN(MATCH_NOMATCH);              if (fi >= max) RRETURN(MATCH_NOMATCH);
3794              if (eptr >= md->end_subject)              if (eptr >= md->end_subject)
3795                {                {
3796                SCHECK_PARTIAL();                SCHECK_PARTIAL();
3797                MRRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
3798                }                }
3799              if (fc == md->lcc[*eptr++]) MRRETURN(MATCH_NOMATCH);              if (fc == *eptr || foc == *eptr) RRETURN(MATCH_NOMATCH);
3800                eptr++;
3801              }              }
3802            }            }
3803          /* Control never gets here */          /* Control never gets here */
# Line 3604  for (;;) Line 3809  for (;;)
3809          {          {
3810          pp = eptr;          pp = eptr;
3811    
3812  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
3813          /* UTF-8 mode */          if (utf)
         if (utf8)  
3814            {            {
3815            register unsigned int d;            register unsigned int d;
3816            for (i = min; i < max; i++)            for (i = min; i < max; i++)
# Line 3618  for (;;) Line 3822  for (;;)
3822                break;                break;
3823                }                }
3824              GETCHARLEN(d, eptr, len);              GETCHARLEN(d, eptr, len);
3825              if (d < 256) d = md->lcc[d];              if (fc == d || (unsigned int)foc == d) break;
             if (fc == d) break;  
3826              eptr += len;              eptr += len;
3827              }              }
3828          if (possessive) continue;            if (possessive) continue;
3829          for(;;)            for(;;)
3830              {              {
3831              RMATCH(eptr, ecode, offset_top, md, eptrb, RM30);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM30);
3832              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
# Line 3633  for (;;) Line 3836  for (;;)
3836            }            }
3837          else          else
3838  #endif  #endif
3839          /* Not UTF-8 mode */          /* Not UTF mode */
3840            {            {
3841            for (i = min; i < max; i++)            for (i = min; i < max; i++)
3842              {              {
# Line 3642  for (;;) Line 3845  for (;;)
3845                SCHECK_PARTIAL();                SCHECK_PARTIAL();
3846                break;                break;
3847                }                }
3848              if (fc == md->lcc[*eptr]) break;              if (fc == *eptr || foc == *eptr) break;
3849              eptr++;              eptr++;
3850              }              }
3851            if (possessive) continue;            if (possessive) continue;
# Line 3654  for (;;) Line 3857  for (;;)
3857              }              }
3858            }            }
3859    
3860          MRRETURN(MATCH_NOMATCH);          RRETURN(MATCH_NOMATCH);
3861          }          }
3862        /* Control never gets here */        /* Control never gets here */
3863        }        }
# Line 3663  for (;;) Line 3866  for (;;)
3866    
3867      else      else
3868        {        {
3869  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
3870        /* UTF-8 mode */        if (utf)
       if (utf8)  
3871          {          {
3872          register unsigned int d;          register unsigned int d;
3873          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
# Line 3673  for (;;) Line 3875  for (;;)
3875            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
3876              {              {
3877              SCHECK_PARTIAL();              SCHECK_PARTIAL();
3878              MRRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
3879              }              }
3880            GETCHARINC(d, eptr);            GETCHARINC(d, eptr);
3881            if (fc == d) MRRETURN(MATCH_NOMATCH);            if (fc == d) RRETURN(MATCH_NOMATCH);
3882            }            }
3883          }          }
3884        else        else
3885  #endif  #endif
3886        /* Not UTF-8 mode */        /* Not UTF mode */
3887          {          {
3888          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
3889            {            {
3890            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
3891              {              {
3892              SCHECK_PARTIAL();              SCHECK_PARTIAL();
3893              MRRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
3894              }              }
3895            if (fc == *eptr++) MRRETURN(MATCH_NOMATCH);            if (fc == *eptr++) RRETURN(MATCH_NOMATCH);
3896            }            }
3897          }          }
3898    
# Line 3698  for (;;) Line 3900  for (;;)
3900    
3901        if (minimize)        if (minimize)
3902          {          {
3903  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
3904          /* UTF-8 mode */          if (utf)
         if (utf8)  
3905            {            {
3906            register unsigned int d;            register unsigned int d;
3907            for (fi = min;; fi++)            for (fi = min;; fi++)
3908              {              {
3909              RMATCH(eptr, ecode, offset_top, md, eptrb, RM32);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM32);
3910              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3911              if (fi >= max) MRRETURN(MATCH_NOMATCH);              if (fi >= max) RRETURN(MATCH_NOMATCH);
3912              if (eptr >= md->end_subject)              if (eptr >= md->end_subject)
3913                {                {
3914                SCHECK_PARTIAL();                SCHECK_PARTIAL();
3915                MRRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
3916                }                }
3917              GETCHARINC(d, eptr);              GETCHARINC(d, eptr);
3918              if (fc == d) MRRETURN(MATCH_NOMATCH);              if (fc == d) RRETURN(MATCH_NOMATCH);
3919              }              }
3920            }            }
3921          else          else
3922  #endif  #endif
3923          /* Not UTF-8 mode */          /* Not UTF mode */
3924            {            {
3925            for (fi = min;; fi++)            for (fi = min;; fi++)
3926              {              {
3927              RMATCH(eptr, ecode, offset_top, md, eptrb, RM33);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM33);
3928              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3929              if (fi >= max) MRRETURN(MATCH_NOMATCH);              if (fi >= max) RRETURN(MATCH_NOMATCH);
3930              if (eptr >= md->end_subject)              if (eptr >= md->end_subject)
3931                {                {
3932                SCHECK_PARTIAL();                SCHECK_PARTIAL();
3933                MRRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
3934                }                }
3935              if (fc == *eptr++) MRRETURN(MATCH_NOMATCH);              if (fc == *eptr++) RRETURN(MATCH_NOMATCH);
3936              }              }
3937            }            }
3938          /* Control never gets here */          /* Control never gets here */
# Line 3743  for (;;) Line 3944  for (;;)
3944          {          {
3945          pp = eptr;          pp = eptr;
3946    
3947  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
3948          /* UTF-8 mode */          if (utf)
         if (utf8)  
3949            {            {
3950            register unsigned int d;            register unsigned int d;
3951            for (i = min; i < max; i++)            for (i = min; i < max; i++)
# Line 3771  for (;;) Line 3971  for (;;)
3971            }            }
3972          else          else
3973  #endif  #endif
3974          /* Not UTF-8 mode */          /* Not UTF mode */
3975            {            {
3976            for (i = min; i < max; i++)            for (i = min; i < max; i++)
3977              {              {
# Line 3792  for (;;) Line 3992  for (;;)
3992              }              }
3993            }            }
3994    
3995          MRRETURN(MATCH_NOMATCH);          RRETURN(MATCH_NOMATCH);
3996          }          }
3997        }        }
3998      /* Control never gets here */      /* Control never gets here */
# Line 3804  for (;;) Line 4004  for (;;)
4004      case OP_TYPEEXACT:      case OP_TYPEEXACT:
4005      min = max = GET2(ecode, 1);      min = max = GET2(ecode, 1);
4006      minimize = TRUE;      minimize = TRUE;
4007      ecode += 3;      ecode += 1 + IMM2_SIZE;
4008      goto REPEATTYPE;      goto REPEATTYPE;
4009    
4010      case OP_TYPEUPTO:      case OP_TYPEUPTO:
# Line 3812  for (;;) Line 4012  for (;;)
4012      min = 0;      min = 0;
4013      max = GET2(ecode, 1);      max = GET2(ecode, 1);
4014      minimize = *ecode == OP_TYPEMINUPTO;      minimize = *ecode == OP_TYPEMINUPTO;
4015      ecode += 3;      ecode += 1 + IMM2_SIZE;
4016      goto REPEATTYPE;      goto REPEATTYPE;
4017    
4018      case OP_TYPEPOSSTAR:      case OP_TYPEPOSSTAR:
# Line 3840  for (;;) Line 4040  for (;;)
4040      possessive = TRUE;      possessive = TRUE;
4041      min = 0;      min = 0;
4042      max = GET2(ecode, 1);      max = GET2(ecode, 1);
4043      ecode += 3;      ecode += 1 + IMM2_SIZE;
4044      goto REPEATTYPE;      goto REPEATTYPE;
4045    
4046      case OP_TYPESTAR:      case OP_TYPESTAR:
# Line 3886  for (;;) Line 4086  for (;;)
4086          switch(prop_type)          switch(prop_type)
4087            {            {
4088            case PT_ANY:            case PT_ANY:
4089            if (prop_fail_result) MRRETURN(MATCH_NOMATCH);            if (prop_fail_result) RRETURN(MATCH_NOMATCH);
4090            for (i = 1; i <= min; i++)            for (i = 1; i <= min; i++)
4091              {              {
4092              if (eptr >= md->end_subject)              if (eptr >= md->end_subject)
4093                {                {
4094                SCHECK_PARTIAL();                SCHECK_PARTIAL();
4095                MRRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
4096                }                }
4097              GETCHARINCTEST(c, eptr);              GETCHARINCTEST(c, eptr);
4098              }              }
# Line 3905  for (;;) Line 4105  for (;;)
4105              if (eptr >= md->end_subject)              if (eptr >= md->end_subject)
4106                {                {
4107                SCHECK_PARTIAL();                SCHECK_PARTIAL();
4108                MRRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
4109                }                }
4110              GETCHARINCTEST(c, eptr);              GETCHARINCTEST(c, eptr);
4111              chartype = UCD_CHARTYPE(c);              chartype = UCD_CHARTYPE(c);
4112              if ((chartype == ucp_Lu ||              if ((chartype == ucp_Lu ||
4113                   chartype == ucp_Ll ||                   chartype == ucp_Ll ||
4114                   chartype == ucp_Lt) == prop_fail_result)                   chartype == ucp_Lt) == prop_fail_result)
4115                MRRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
4116              }              }
4117            break;            break;
4118    
# Line 3922  for (;;) Line 4122  for (;;)
4122              if (eptr >= md->end_subject)              if (eptr >= md->end_subject)
4123                {                {
4124                SCHECK_PARTIAL();                SCHECK_PARTIAL();
4125                MRRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
4126                }                }
4127              GETCHARINCTEST(c, eptr);              GETCHARINCTEST(c, eptr);
4128              if ((UCD_CATEGORY(c) == prop_value) == prop_fail_result)              if ((UCD_CATEGORY(c) == prop_value) == prop_fail_result)
4129                MRRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
4130              }              }
4131            break;            break;
4132    
# Line 3936  for (;;) Line 4136  for (;;)
4136              if (eptr >= md->end_subject)              if (eptr >= md->end_subject)
4137                {                {
4138                SCHECK_PARTIAL();                SCHECK_PARTIAL();
4139                MRRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
4140                }                }
4141              GETCHARINCTEST(c, eptr);              GETCHARINCTEST(c, eptr);
4142              if ((UCD_CHARTYPE(c) == prop_value) == prop_fail_result)              if ((UCD_CHARTYPE(c) == prop_value) == prop_fail_result)
4143                MRRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
4144              }              }
4145            break;            break;
4146    
# Line 3950  for (;;) Line 4150  for (;;)
4150              if (eptr >= md->end_subject)              if (eptr >= md->end_subject)
4151                {                {
4152                SCHECK_PARTIAL();                SCHECK_PARTIAL();
4153                MRRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
4154                }                }
4155              GETCHARINCTEST(c, eptr);              GETCHARINCTEST(c, eptr);
4156              if ((UCD_SCRIPT(c) == prop_value) == prop_fail_result)              if ((UCD_SCRIPT(c) == prop_value) == prop_fail_result)
4157                MRRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
4158              }              }
4159            break;            break;
4160    
# Line 3965  for (;;) Line 4165  for (;;)
4165              if (eptr >= md->end_subject)              if (eptr >= md->end_subject)
4166                {                {
4167                SCHECK_PARTIAL();                SCHECK_PARTIAL();
4168                MRRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
4169                }                }
4170              GETCHARINCTEST(c, eptr);              GETCHARINCTEST(c, eptr);
4171              category = UCD_CATEGORY(c);              category = UCD_CATEGORY(c);
4172              if ((category == ucp_L || category == ucp_N) == prop_fail_result)              if ((category == ucp_L || category == ucp_N) == prop_fail_result)
4173                MRRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
4174              }              }
4175            break;            break;
4176    
# Line 3980  for (;;) Line 4180  for (;;)
4180              if (eptr >= md->end_subject)              if (eptr >= md->end_subject)
4181                {                {
4182                SCHECK_PARTIAL();                SCHECK_PARTIAL();
4183                MRRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
4184                }                }
4185              GETCHARINCTEST(c, eptr);              GETCHARINCTEST(c, eptr);
4186              if ((UCD_CATEGORY(c) == ucp_Z || c == CHAR_HT || c == CHAR_NL ||              if ((UCD_CATEGORY(c) == ucp_Z || c == CHAR_HT || c == CHAR_NL ||
4187                   c == CHAR_FF || c == CHAR_CR)                   c == CHAR_FF || c == CHAR_CR)
4188                     == prop_fail_result)                     == prop_fail_result)
4189                MRRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
4190              }              }
4191            break;            break;
4192    
# Line 3996  for (;;) Line 4196  for (;;)
4196              if (eptr >= md->end_subject)              if (eptr >= md->end_subject)
4197                {                {
4198                SCHECK_PARTIAL();                SCHECK_PARTIAL();
4199                MRRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
4200                }                }
4201              GETCHARINCTEST(c, eptr);              GETCHARINCTEST(c, eptr);
4202              if ((UCD_CATEGORY(c) == ucp_Z || c == CHAR_HT || c == CHAR_NL ||              if ((UCD_CATEGORY(c) == ucp_Z || c == CHAR_HT || c == CHAR_NL ||
4203                   c == CHAR_VT || c == CHAR_FF || c == CHAR_CR)                   c == CHAR_VT || c == CHAR_FF || c == CHAR_CR)
4204                     == prop_fail_result)                     == prop_fail_result)
4205                MRRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
4206              }              }
4207            break;            break;
4208    
# Line 4013  for (;;) Line 4213  for (;;)
4213              if (eptr >= md->end_subject)              if (eptr >= md->end_subject)
4214                {                {
4215                SCHECK_PARTIAL();                SCHECK_PARTIAL();
4216                MRRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
4217                }                }
4218              GETCHARINCTEST(c, eptr);              GETCHARINCTEST(c, eptr);
4219              category = UCD_CATEGORY(c);              category = UCD_CATEGORY(c);
4220              if ((category == ucp_L || category == ucp_N || c == CHAR_UNDERSCORE)              if ((category == ucp_L || category == ucp_N || c == CHAR_UNDERSCORE)
4221                     == prop_fail_result)                     == prop_fail_result)
4222                MRRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
4223              }              }
4224            break;            break;
4225    
# Line 4040  for (;;) Line 4240  for (;;)
4240            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
4241              {              {
4242              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4243              MRRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4244              }              }
4245            GETCHARINCTEST(c, eptr);            else
4246            if (UCD_CATEGORY(c) == ucp_M) MRRETURN(MATCH_NOMATCH);              {
4247            while (eptr < md->end_subject)              int lgb, rgb;
4248              {              GETCHARINCTEST(c, eptr);
4249              int len = 1;              lgb = UCD_GRAPHBREAK(c);
4250              if (!utf8) c = *eptr; else { GETCHARLEN(c, eptr, len); }             while (eptr < md->end_subject)
4251              if (UCD_CATEGORY(c) != ucp_M) break;                {
4252              eptr += len;                int len = 1;
4253                  if (!utf) c = *eptr; else { GETCHARLEN(c, eptr, len); }
4254                  rgb = UCD_GRAPHBREAK(c);
4255                  if ((PRIV(ucp_gbtable)[lgb] & (1 << rgb)) == 0) break;
4256                  lgb = rgb;
4257                  eptr += len;
4258                  }
4259              }              }
4260              CHECK_PARTIAL();
4261            }            }
4262          }          }
4263    
# Line 4059  for (;;) Line 4266  for (;;)
4266    
4267  /* Handle all other cases when the coding is UTF-8 */  /* Handle all other cases when the coding is UTF-8 */
4268    
4269  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
4270        if (utf8) switch(ctype)        if (utf) switch(ctype)
4271          {          {
4272          case OP_ANY:          case OP_ANY:
4273          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
# Line 4068  for (;;) Line 4275  for (;;)
4275            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
4276              {              {
4277              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4278              MRRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4279                }
4280              if (IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH);
4281              if (md->partial != 0 &&
4282                  eptr + 1 >= md->end_subject &&
4283                  NLBLOCK->nltype == NLTYPE_FIXED &&
4284                  NLBLOCK->nllen == 2 &&
4285                  *eptr == NLBLOCK->nl[0])
4286                {
4287                md->hitend = TRUE;
4288                if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
4289              }              }
           if (IS_NEWLINE(eptr)) MRRETURN(MATCH_NOMATCH);  
4290            eptr++;            eptr++;
4291            while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;            ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);
4292            }            }
4293          break;          break;
4294    
# Line 4082  for (;;) Line 4298  for (;;)
4298            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
4299              {              {
4300              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4301              MRRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4302              }              }
4303            eptr++;            eptr++;
4304            while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;            ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);
4305            }            }
4306          break;          break;
4307    
4308          case OP_ANYBYTE:          case OP_ANYBYTE:
4309          if (eptr > md->end_subject - min) MRRETURN(MATCH_NOMATCH);          if (eptr > md->end_subject - min) RRETURN(MATCH_NOMATCH);
4310          eptr += min;          eptr += min;
4311          break;          break;
4312    
# Line 4100  for (;;) Line 4316  for (;;)
4316            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
4317              {              {
4318              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4319              MRRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4320              }              }
4321            GETCHARINC(c, eptr);            GETCHARINC(c, eptr);
4322            switch(c)            switch(c)
4323              {              {
4324              default: MRRETURN(MATCH_NOMATCH);              default: RRETURN(MATCH_NOMATCH);
4325    
4326              case 0x000d:              case CHAR_CR:
4327              if (eptr < md->end_subject && *eptr == 0x0a) eptr++;              if (eptr < md->end_subject && *eptr == CHAR_LF) eptr++;
4328              break;              break;
4329    
4330              case 0x000a:              case CHAR_LF:
4331              break;              break;
4332    
4333              case 0x000b:              case CHAR_VT:
4334              case 0x000c:              case CHAR_FF:
4335              case 0x0085:              case CHAR_NEL:
4336    #ifndef EBCDIC
4337              case 0x2028:              case 0x2028:
4338              case 0x2029:              case 0x2029:
4339              if (md->bsr_anycrlf) MRRETURN(MATCH_NOMATCH);  #endif  /* Not EBCDIC */
4340                if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);
4341              break;              break;
4342              }              }
4343            }            }
# Line 4131  for (;;) Line 4349  for (;;)
4349            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
4350              {              {
4351              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4352              MRRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4353              }              }
4354            GETCHARINC(c, eptr);            GETCHARINC(c, eptr);
4355            switch(c)            switch(c)
4356              {              {
4357              default: break;              default: break;
4358              case 0x09:      /* HT */              case CHAR_HT:
4359              case 0x20:      /* SPACE */              case CHAR_SPACE:
4360    #ifndef EBCDIC
4361              case 0xa0:      /* NBSP */              case 0xa0:      /* NBSP */
4362              case 0x1680:    /* OGHAM SPACE MARK */              case 0x1680:    /* OGHAM SPACE MARK */
4363              case 0x180e:    /* MONGOLIAN VOWEL SEPARATOR */              case 0x180e:    /* MONGOLIAN VOWEL SEPARATOR */
# Line 4156  for (;;) Line 4375  for (;;)
4375              case 0x202f:    /* NARROW NO-BREAK SPACE */              case 0x202f:    /* NARROW NO-BREAK SPACE */
4376              case 0x205f:    /* MEDIUM MATHEMATICAL SPACE */              case 0x205f:    /* MEDIUM MATHEMATICAL SPACE */
4377              case 0x3000:    /* IDEOGRAPHIC SPACE */              case 0x3000:    /* IDEOGRAPHIC SPACE */
4378              MRRETURN(MATCH_NOMATCH);  #endif  /* Not EBCDIC */
4379                RRETURN(MATCH_NOMATCH);
4380              }              }
4381            }            }
4382          break;          break;
# Line 4167  for (;;) Line 4387  for (;;)
4387            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
4388              {              {
4389              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4390              MRRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4391              }              }
4392            GETCHARINC(c, eptr);            GETCHARINC(c, eptr);
4393            switch(c)            switch(c)
4394              {              {
4395              default: MRRETURN(MATCH_NOMATCH);              default: RRETURN(MATCH_NOMATCH);
4396              case 0x09:      /* HT */              case CHAR_HT:
4397              case 0x20:      /* SPACE */              case CHAR_SPACE:
4398    #ifndef EBCDIC
4399              case 0xa0:      /* NBSP */              case 0xa0:      /* NBSP */
4400              case 0x1680:    /* OGHAM SPACE MARK */              case 0x1680:    /* OGHAM SPACE MARK */
4401              case 0x180e:    /* MONGOLIAN VOWEL SEPARATOR */              case 0x180e:    /* MONGOLIAN VOWEL SEPARATOR */
# Line 4192  for (;;) Line 4413  for (;;)
4413              case 0x202f:    /* NARROW NO-BREAK SPACE */              case 0x202f:    /* NARROW NO-BREAK SPACE */
4414              case 0x205f:    /* MEDIUM MATHEMATICAL SPACE */              case 0x205f:    /* MEDIUM MATHEMATICAL SPACE */
4415              case 0x3000:    /* IDEOGRAPHIC SPACE */              case 0x3000:    /* IDEOGRAPHIC SPACE */
4416    #endif
4417              break;              break;
4418              }              }
4419            }            }
# Line 4203  for (;;) Line 4425  for (;;)
4425            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
4426              {              {
4427              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4428              MRRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4429              }              }
4430            GETCHARINC(c, eptr);            GETCHARINC(c, eptr);
4431            switch(c)            switch(c)
4432              {              {
4433              default: break;              default: break;
4434              case 0x0a:      /* LF */              case CHAR_LF:
4435              case 0x0b:      /* VT */              case CHAR_VT:
4436              case 0x0c:      /* FF */              case CHAR_FF:
4437              case 0x0d:      /* CR */              case CHAR_CR:
4438              case 0x85:      /* NEL */              case CHAR_NEL:
4439    #ifndef EBCDIC
4440              case 0x2028:    /* LINE SEPARATOR */              case 0x2028:    /* LINE SEPARATOR */
4441              case 0x2029:    /* PARAGRAPH SEPARATOR */              case 0x2029:    /* PARAGRAPH SEPARATOR */
4442              MRRETURN(MATCH_NOMATCH);  #endif
4443                RRETURN(MATCH_NOMATCH);
4444              }              }
4445            }            }
4446          break;          break;
# Line 4227  for (;;) Line 4451  for (;;)
4451            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
4452              {              {
4453              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4454              MRRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4455              }              }
4456            GETCHARINC(c, eptr);            GETCHARINC(c, eptr);
4457            switch(c)            switch(c)
4458              {              {
4459              default: MRRETURN(MATCH_NOMATCH);              default: RRETURN(MATCH_NOMATCH);
4460              case 0x0a:      /* LF */              case CHAR_LF:
4461              case 0x0b:      /* VT */              case CHAR_VT:
4462              case 0x0c:      /* FF */              case CHAR_FF:
4463              case 0x0d:      /* CR */              case CHAR_CR:
4464              case 0x85:      /* NEL */              case CHAR_NEL:
4465    #ifndef EBCDIC
4466              case 0x2028:    /* LINE SEPARATOR */              case 0x2028:    /* LINE SEPARATOR */
4467              case 0x2029:    /* PARAGRAPH SEPARATOR */              case 0x2029:    /* PARAGRAPH SEPARATOR */
4468    #endif
4469              break;              break;
4470              }              }
4471            }            }
# Line 4251  for (;;) Line 4477  for (;;)
4477            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
4478              {              {
4479              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4480              MRRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4481              }              }
4482            GETCHARINC(c, eptr);            GETCHARINC(c, eptr);
4483            if (c < 128 && (md->ctypes[c] & ctype_digit) != 0)            if (c < 128 && (md->ctypes[c] & ctype_digit) != 0)
4484              MRRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4485            }            }
4486          break;          break;
4487    
# Line 4265  for (;;) Line 4491  for (;;)
4491            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
4492              {              {
4493              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4494              MRRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4495              }              }
4496            if (*eptr >= 128 || (md->ctypes[*eptr++] & ctype_digit) == 0)            if (*eptr >= 128 || (md->ctypes[*eptr] & ctype_digit) == 0)
4497              MRRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4498              eptr++;
4499            /* 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 */
4500            }            }
4501          break;          break;
# Line 4279  for (;;) Line 4506  for (;;)
4506            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
4507              {              {
4508              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4509              MRRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4510              }              }
4511            if (*eptr < 128 && (md->ctypes[*eptr] & ctype_space) != 0)            if (*eptr < 128 && (md->ctypes[*eptr] & ctype_space) != 0)
4512              MRRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4513            while (++eptr < md->end_subject && (*eptr & 0xc0) == 0x80);            eptr++;
4514              ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);
4515            }            }
4516          break;          break;
4517    
# Line 4293  for (;;) Line 4521  for (;;)
4521            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
4522              {              {
4523              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4524              MRRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4525              }              }
4526            if (*eptr >= 128 || (md->ctypes[*eptr++] & ctype_space) == 0)            if (*eptr >= 128 || (md->ctypes[*eptr] & ctype_space) == 0)
4527              MRRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4528              eptr++;
4529            /* 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 */
4530            }            }
4531          break;          break;
# Line 4307  for (;;) Line 4536  for (;;)
4536            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
4537              {              {
4538              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4539              MRRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4540              }              }
4541            if (*eptr < 128 && (md->ctypes[*eptr] & ctype_word) != 0)            if (*eptr < 128 && (md->ctypes[*eptr] & ctype_word) != 0)
4542              MRRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4543            while (++eptr < md->end_subject && (*eptr & 0xc0) == 0x80);            eptr++;
4544              ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);
4545            }            }
4546          break;          break;
4547    
# Line 4321  for (;;) Line 4551  for (;;)
4551            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
4552              {              {
4553              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4554              MRRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4555              }              }
4556            if (*eptr >= 128 || (md->ctypes[*eptr++] & ctype_word) == 0)            if (*eptr >= 128 || (md->ctypes[*eptr] & ctype_word) == 0)
4557              MRRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4558              eptr++;
4559            /* 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 */
4560            }            }
4561          break;          break;
# Line 4334  for (;;) Line 4565  for (;;)
4565          }  /* End switch(ctype) */          }  /* End switch(ctype) */
4566    
4567        else        else
4568  #endif     /* SUPPORT_UTF8 */  #endif     /* SUPPORT_UTF */
4569    
4570        /* 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
4571        than OP_PROP and OP_NOTPROP. */        than OP_PROP and OP_NOTPROP. */
# Line 4347  for (;;) Line 4578  for (;;)
4578            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
4579              {              {
4580              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4581              MRRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4582                }
4583              if (IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH);
4584              if (md->partial != 0 &&
4585                  eptr + 1 >= md->end_subject &&
4586                  NLBLOCK->nltype == NLTYPE_FIXED &&
4587                  NLBLOCK->nllen == 2 &&
4588                  *eptr == NLBLOCK->nl[0])
4589                {
4590                md->hitend = TRUE;
4591                if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
4592              }              }
           if (IS_NEWLINE(eptr)) MRRETURN(MATCH_NOMATCH);  
4593            eptr++;            eptr++;
4594            }            }
4595          break;          break;
# Line 4358  for (;;) Line 4598  for (;;)
4598          if (eptr > md->end_subject - min)          if (eptr > md->end_subject - min)
4599            {            {
4600            SCHECK_PARTIAL();            SCHECK_PARTIAL();
4601            MRRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
4602            }            }
4603          eptr += min;          eptr += min;
4604          break;          break;
# Line 4367  for (;;) Line 4607  for (;;)
4607          if (eptr > md->end_subject - min)          if (eptr > md->end_subject - min)
4608            {            {
4609            SCHECK_PARTIAL();