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

Diff of /code/trunk/pcre_exec.c

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

revision 771 by ph10, Tue Nov 29 15:34:12 2011 UTC revision 1084 by chpe, Tue Oct 16 15:55:28 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 113  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 140  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 163  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 173  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      data units matched may differ, because in UTF-8 there are some characters
184      lower case versions code as different numbers of bytes. For example, U+023A      whose upper and lower case versions code have different numbers of bytes.
185      (2 bytes in UTF-8) is the upper case version of U+2C65 (3 bytes in UTF-8);      For example, U+023A (2 bytes in UTF-8) is the upper case version of U+2C65
186      a sequence of 3 of the former uses 6 bytes, as does a sequence of two of      (3 bytes in UTF-8); a sequence of 3 of the former uses 6 bytes, as does a
187      the latter. It is important, therefore, to check the length along the      sequence of two of the latter. It is important, therefore, to check the
188      reference, not along the subject (earlier code did this wrong). */      length along the reference, not along the subject (earlier code did this
189        wrong). */
190    
191      USPTR endptr = p + length;      PCRE_PUCHAR endptr = p + length;
192      while (p < endptr)      while (p < endptr)
193        {        {
194        int c, d;        pcre_uint32 c, d;
195        if (eptr >= md->end_subject) return -1;        const ucd_record *ur;
196          if (eptr >= md->end_subject) return -2;   /* Partial match */
197        GETCHARINC(c, eptr);        GETCHARINC(c, eptr);
198        GETCHARINC(d, p);        GETCHARINC(d, p);
199        if (c != d && c != UCD_OTHERCASE(d)) return -1;        ur = GET_UCD(d);
200          if (c != d && c != d + ur->other_case)
201            {
202            const pcre_uint32 *pp = PRIV(ucd_caseless_sets) + ur->caseset;
203            for (;;)
204              {
205              if (c < *pp) return -1;
206              if (c == *pp++) break;
207              }
208            }
209        }        }
210      }      }
211    else    else
# Line 202  if (caseless) Line 215  if (caseless)
215    /* 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
216    is no UCP support. */    is no UCP support. */
217      {      {
     if (eptr + length > md->end_subject) return -1;  
218      while (length-- > 0)      while (length-- > 0)
219        { if (md->lcc[*p++] != md->lcc[*eptr++]) return -1; }        {
220          if (eptr >= md->end_subject) return -2;   /* Partial match */
221          if (TABLE_GET(*p, md->lcc, *p) != TABLE_GET(*eptr, md->lcc, *eptr)) return -1;
222          p++;
223          eptr++;
224          }
225      }      }
226    }    }
227    
# Line 213  are in UTF-8 mode. */ Line 230  are in UTF-8 mode. */
230    
231  else  else
232    {    {
233    if (eptr + length > md->end_subject) return -1;    while (length-- > 0)
234    while (length-- > 0) if (*p++ != *eptr++) return -1;      {
235        if (eptr >= md->end_subject) return -2;   /* Partial match */
236        if (*p++ != *eptr++) return -1;
237        }
238    }    }
239    
240  return eptr - eptr_start;  return (int)(eptr - eptr_start);
241  }  }
242    
243    
# Line 307  argument of match(), which never changes Line 327  argument of match(), which never changes
327    
328  #define RMATCH(ra,rb,rc,rd,re,rw)\  #define RMATCH(ra,rb,rc,rd,re,rw)\
329    {\    {\
330    heapframe *newframe = (heapframe *)(pcre_stack_malloc)(sizeof(heapframe));\    heapframe *newframe = frame->Xnextframe;\
331    if (newframe == NULL) RRETURN(PCRE_ERROR_NOMEMORY);\    if (newframe == NULL)\
332    frame->Xwhere = rw; \      {\
333        newframe = (heapframe *)(PUBL(stack_malloc))(sizeof(heapframe));\
334        if (newframe == NULL) RRETURN(PCRE_ERROR_NOMEMORY);\
335        newframe->Xnextframe = NULL;\
336        frame->Xnextframe = newframe;\
337        }\
338      frame->Xwhere = rw;\
339    newframe->Xeptr = ra;\    newframe->Xeptr = ra;\
340    newframe->Xecode = rb;\    newframe->Xecode = rb;\
341    newframe->Xmstart = mstart;\    newframe->Xmstart = mstart;\
# Line 328  argument of match(), which never changes Line 354  argument of match(), which never changes
354    {\    {\
355    heapframe *oldframe = frame;\    heapframe *oldframe = frame;\
356    frame = oldframe->Xprevframe;\    frame = oldframe->Xprevframe;\
   (pcre_stack_free)(oldframe);\  
357    if (frame != NULL)\    if (frame != NULL)\
358      {\      {\
359      rrc = ra;\      rrc = ra;\
# Line 342  argument of match(), which never changes Line 367  argument of match(), which never changes
367    
368  typedef struct heapframe {  typedef struct heapframe {
369    struct heapframe *Xprevframe;    struct heapframe *Xprevframe;
370      struct heapframe *Xnextframe;
371    
372    /* Function arguments that may change */    /* Function arguments that may change */
373    
374    USPTR Xeptr;    PCRE_PUCHAR Xeptr;
375    const uschar *Xecode;    const pcre_uchar *Xecode;
376    USPTR Xmstart;    PCRE_PUCHAR Xmstart;
377    int Xoffset_top;    int Xoffset_top;
378    eptrblock *Xeptrb;    eptrblock *Xeptrb;
379    unsigned int Xrdepth;    unsigned int Xrdepth;
380    
381    /* Function local variables */    /* Function local variables */
382    
383    USPTR Xcallpat;    PCRE_PUCHAR Xcallpat;
384  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
385    USPTR Xcharptr;    PCRE_PUCHAR Xcharptr;
386  #endif  #endif
387    USPTR Xdata;    PCRE_PUCHAR Xdata;
388    USPTR Xnext;    PCRE_PUCHAR Xnext;
389    USPTR Xpp;    PCRE_PUCHAR Xpp;
390    USPTR Xprev;    PCRE_PUCHAR Xprev;
391    USPTR Xsaved_eptr;    PCRE_PUCHAR Xsaved_eptr;
392    
393    recursion_info Xnew_recursive;    recursion_info Xnew_recursive;
394    
# Line 375  typedef struct heapframe { Line 401  typedef struct heapframe {
401    int Xprop_value;    int Xprop_value;
402    int Xprop_fail_result;    int Xprop_fail_result;
403    int Xoclength;    int Xoclength;
404    uschar Xocchars[8];    pcre_uchar Xocchars[6];
405  #endif  #endif
406    
407    int Xcodelink;    int Xcodelink;
# Line 417  returns a negative (error) response, the Line 443  returns a negative (error) response, the
443  same response. */  same response. */
444    
445  /* 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
446  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
447  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.
448  something has been matched). For hard partial matching, we then return  something has been matched). For hard partial matching, we then return
449  immediately. The second one is used when we already know we are past the end of  immediately. The second one is used when we already know we are past the end of
# Line 440  the subject. */ Line 466  the subject. */
466    
467    
468  /* Performance note: It might be tempting to extract commonly used fields from  /* Performance note: It might be tempting to extract commonly used fields from
469  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
470  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
471  made performance worse.  made performance worse.
472    
# Line 463  Returns:       MATCH_MATCH if matched Line 489  Returns:       MATCH_MATCH if matched
489  */  */
490    
491  static int  static int
492  match(REGISTER USPTR eptr, REGISTER const uschar *ecode, USPTR mstart,  match(REGISTER PCRE_PUCHAR eptr, REGISTER const pcre_uchar *ecode,
493    int offset_top, match_data *md, eptrblock *eptrb, unsigned int rdepth)    PCRE_PUCHAR mstart, int offset_top, match_data *md, eptrblock *eptrb,
494      unsigned int rdepth)
495  {  {
496  /* 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,
497  so they can be ordinary variables in all cases. Mark some of them with  so they can be ordinary variables in all cases. Mark some of them with
# Line 472  so they can be ordinary variables in all Line 499  so they can be ordinary variables in all
499    
500  register int  rrc;         /* Returns from recursive calls */  register int  rrc;         /* Returns from recursive calls */
501  register int  i;           /* Used for loops not involving calls to RMATCH() */  register int  i;           /* Used for loops not involving calls to RMATCH() */
502  register unsigned int c;   /* Character values not kept over RMATCH() calls */  register pcre_uint32 c;    /* Character values not kept over RMATCH() calls */
503  register BOOL utf8;        /* Local copy of UTF-8 flag for speed */  register BOOL utf;         /* Local copy of UTF flag for speed */
504    
505  BOOL minimize, possessive; /* Quantifier options */  BOOL minimize, possessive; /* Quantifier options */
506  BOOL caseless;  BOOL caseless;
507  int condcode;  int condcode;
508    
509  /* 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
510  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
511  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
512  heap whenever RMATCH() does a "recursion". See the macro definitions above. */  whenever RMATCH() does a "recursion". See the macro definitions above. Putting
513    the top-level on the stack rather than malloc-ing them all gives a performance
514    boost in many cases where there is not much "recursion". */
515    
516  #ifdef NO_RECURSE  #ifdef NO_RECURSE
517  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 */  
518    
519  /* Copy in the original argument variables */  /* Copy in the original argument variables */
520    
# Line 513  HEAP_RECURSE: Line 540  HEAP_RECURSE:
540    
541  /* Ditto for the local variables */  /* Ditto for the local variables */
542    
543  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
544  #define charptr            frame->Xcharptr  #define charptr            frame->Xcharptr
545  #endif  #endif
546  #define callpat            frame->Xcallpat  #define callpat            frame->Xcallpat
# Line 571  declarations can be cut out in a block. Line 598  declarations can be cut out in a block.
598  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
599  to RMATCH(). */  to RMATCH(). */
600    
601  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
602  const uschar *charptr;  const pcre_uchar *charptr;
603  #endif  #endif
604  const uschar *callpat;  const pcre_uchar *callpat;
605  const uschar *data;  const pcre_uchar *data;
606  const uschar *next;  const pcre_uchar *next;
607  USPTR         pp;  PCRE_PUCHAR       pp;
608  const uschar *prev;  const pcre_uchar *prev;
609  USPTR         saved_eptr;  PCRE_PUCHAR       saved_eptr;
610    
611  recursion_info new_recursive;  recursion_info new_recursive;
612    
# Line 592  int prop_type; Line 619  int prop_type;
619  int prop_value;  int prop_value;
620  int prop_fail_result;  int prop_fail_result;
621  int oclength;  int oclength;
622  uschar occhars[8];  pcre_uchar occhars[6];
623  #endif  #endif
624    
625  int codelink;  int codelink;
# Line 602  int max; Line 629  int max;
629  int min;  int min;
630  int number;  int number;
631  int offset;  int offset;
632  int op;  pcre_uchar op;
633  int save_capture_last;  int save_capture_last;
634  int save_offset1, save_offset2, save_offset3;  int save_offset1, save_offset2, save_offset3;
635  int stacksave[REC_STACK_SAVE_MAX];  int stacksave[REC_STACK_SAVE_MAX];
636    
637  eptrblock newptrb;  eptrblock newptrb;
638    
639    /* There is a special fudge for calling match() in a way that causes it to
640    measure the size of its basic stack frame when the stack is being used for
641    recursion. The second argument (ecode) being NULL triggers this behaviour. It
642    cannot normally ever be NULL. The return is the negated value of the frame
643    size. */
644    
645    if (ecode == NULL)
646      {
647      if (rdepth == 0)
648        return match((PCRE_PUCHAR)&rdepth, NULL, NULL, 0, NULL, NULL, 1);
649      else
650        {
651        int len = (char *)&rdepth - (char *)eptr;
652        return (len > 0)? -len : len;
653        }
654      }
655  #endif     /* NO_RECURSE */  #endif     /* NO_RECURSE */
656    
657  /* To save space on the stack and in the heap frame, I have doubled up on some  /* To save space on the stack and in the heap frame, I have doubled up on some
# Line 620  the alternative names that are used. */ Line 664  the alternative names that are used. */
664  #define code_offset   codelink  #define code_offset   codelink
665  #define condassert    condition  #define condassert    condition
666  #define matched_once  prev_is_word  #define matched_once  prev_is_word
667    #define foc           number
668    #define save_mark     data
669    
670  /* These statements are here to stop the compiler complaining about unitialized  /* These statements are here to stop the compiler complaining about unitialized
671  variables. */  variables. */
# Line 645  defined). However, RMATCH isn't like a f Line 691  defined). However, RMATCH isn't like a f
691  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,
692  however, impact performance when true recursion is being used. */  however, impact performance when true recursion is being used. */
693    
694  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
695  utf8 = md->utf8;       /* Local copy of the flag */  utf = md->utf;       /* Local copy of the flag */
696  #else  #else
697  utf8 = FALSE;  utf = FALSE;
698  #endif  #endif
699    
700  /* First check that we haven't called match() too many times, or that we  /* First check that we haven't called match() too many times, or that we
# Line 689  for (;;) Line 735  for (;;)
735      case OP_MARK:      case OP_MARK:
736      md->nomatch_mark = ecode + 2;      md->nomatch_mark = ecode + 2;
737      md->mark = NULL;    /* In case previously set by assertion */      md->mark = NULL;    /* In case previously set by assertion */
738      RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode] + ecode[1], offset_top, md,      RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode] + ecode[1], offset_top, md,
739        eptrb, RM55);        eptrb, RM55);
740      if ((rrc == MATCH_MATCH || rrc == MATCH_ACCEPT) &&      if ((rrc == MATCH_MATCH || rrc == MATCH_ACCEPT) &&
741           md->mark == NULL) md->mark = ecode + 2;           md->mark == NULL) md->mark = ecode + 2;
# Line 702  for (;;) Line 748  for (;;)
748      unaltered. */      unaltered. */
749    
750      else if (rrc == MATCH_SKIP_ARG &&      else if (rrc == MATCH_SKIP_ARG &&
751          strcmp((char *)(ecode + 2), (char *)(md->start_match_ptr)) == 0)          STRCMP_UC_UC(ecode + 2, md->start_match_ptr) == 0)
752        {        {
753        md->start_match_ptr = eptr;        md->start_match_ptr = eptr;
754        RRETURN(MATCH_SKIP);        RRETURN(MATCH_SKIP);
# Line 715  for (;;) Line 761  for (;;)
761      /* COMMIT overrides PRUNE, SKIP, and THEN */      /* COMMIT overrides PRUNE, SKIP, and THEN */
762    
763      case OP_COMMIT:      case OP_COMMIT:
764      RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,      RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md,
765        eptrb, RM52);        eptrb, RM52);
766      if (rrc != MATCH_NOMATCH && rrc != MATCH_PRUNE &&      if (rrc != MATCH_NOMATCH && rrc != MATCH_PRUNE &&
767          rrc != MATCH_SKIP && rrc != MATCH_SKIP_ARG &&          rrc != MATCH_SKIP && rrc != MATCH_SKIP_ARG &&
# Line 726  for (;;) Line 772  for (;;)
772      /* PRUNE overrides THEN */      /* PRUNE overrides THEN */
773    
774      case OP_PRUNE:      case OP_PRUNE:
775      RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,      RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md,
776        eptrb, RM51);        eptrb, RM51);
777      if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN) RRETURN(rrc);      if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN) RRETURN(rrc);
778      RRETURN(MATCH_PRUNE);      RRETURN(MATCH_PRUNE);
# Line 734  for (;;) Line 780  for (;;)
780      case OP_PRUNE_ARG:      case OP_PRUNE_ARG:
781      md->nomatch_mark = ecode + 2;      md->nomatch_mark = ecode + 2;
782      md->mark = NULL;    /* In case previously set by assertion */      md->mark = NULL;    /* In case previously set by assertion */
783      RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode] + ecode[1], offset_top, md,      RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode] + ecode[1], offset_top, md,
784        eptrb, RM56);        eptrb, RM56);
785      if ((rrc == MATCH_MATCH || rrc == MATCH_ACCEPT) &&      if ((rrc == MATCH_MATCH || rrc == MATCH_ACCEPT) &&
786           md->mark == NULL) md->mark = ecode + 2;           md->mark == NULL) md->mark = ecode + 2;
# Line 744  for (;;) Line 790  for (;;)
790      /* SKIP overrides PRUNE and THEN */      /* SKIP overrides PRUNE and THEN */
791    
792      case OP_SKIP:      case OP_SKIP:
793      RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,      RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md,
794        eptrb, RM53);        eptrb, RM53);
795      if (rrc != MATCH_NOMATCH && rrc != MATCH_PRUNE && rrc != MATCH_THEN)      if (rrc != MATCH_NOMATCH && rrc != MATCH_PRUNE && rrc != MATCH_THEN)
796        RRETURN(rrc);        RRETURN(rrc);
# Line 752  for (;;) Line 798  for (;;)
798      RRETURN(MATCH_SKIP);      RRETURN(MATCH_SKIP);
799    
800      /* Note that, for Perl compatibility, SKIP with an argument does NOT set      /* Note that, for Perl compatibility, SKIP with an argument does NOT set
801      nomatch_mark. There is a flag that disables this opcode when re-matching a      nomatch_mark. There is a flag that disables this opcode when re-matching a
802      pattern that ended with a SKIP for which there was not a matching MARK. */      pattern that ended with a SKIP for which there was not a matching MARK. */
803    
804      case OP_SKIP_ARG:      case OP_SKIP_ARG:
805      if (md->ignore_skip_arg)      if (md->ignore_skip_arg)
806        {        {
807        ecode += _pcre_OP_lengths[*ecode] + ecode[1];        ecode += PRIV(OP_lengths)[*ecode] + ecode[1];
808        break;        break;
809        }        }
810      RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode] + ecode[1], offset_top, md,      RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode] + ecode[1], offset_top, md,
811        eptrb, RM57);        eptrb, RM57);
812      if (rrc != MATCH_NOMATCH && rrc != MATCH_PRUNE && rrc != MATCH_THEN)      if (rrc != MATCH_NOMATCH && rrc != MATCH_PRUNE && rrc != MATCH_THEN)
813        RRETURN(rrc);        RRETURN(rrc);
814    
815      /* 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
816      returning the special MATCH_SKIP_ARG return code. This will either be      returning the special MATCH_SKIP_ARG return code. This will either be
817      caught by a matching MARK, or get to the top, where it causes a rematch      caught by a matching MARK, or get to the top, where it causes a rematch
818      with the md->ignore_skip_arg flag set. */      with the md->ignore_skip_arg flag set. */
819    
820      md->start_match_ptr = ecode + 2;      md->start_match_ptr = ecode + 2;
# Line 779  for (;;) Line 825  for (;;)
825      match pointer to do this. */      match pointer to do this. */
826    
827      case OP_THEN:      case OP_THEN:
828      RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,      RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md,
829        eptrb, RM54);        eptrb, RM54);
830      if (rrc != MATCH_NOMATCH) RRETURN(rrc);      if (rrc != MATCH_NOMATCH) RRETURN(rrc);
831      md->start_match_ptr = ecode;      md->start_match_ptr = ecode;
# Line 788  for (;;) Line 834  for (;;)
834      case OP_THEN_ARG:      case OP_THEN_ARG:
835      md->nomatch_mark = ecode + 2;      md->nomatch_mark = ecode + 2;
836      md->mark = NULL;    /* In case previously set by assertion */      md->mark = NULL;    /* In case previously set by assertion */
837      RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode] + ecode[1], offset_top,      RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode] + ecode[1], offset_top,
838        md, eptrb, RM58);        md, eptrb, RM58);
839      if ((rrc == MATCH_MATCH || rrc == MATCH_ACCEPT) &&      if ((rrc == MATCH_MATCH || rrc == MATCH_ACCEPT) &&
840           md->mark == NULL) md->mark = ecode + 2;           md->mark == NULL) md->mark = ecode + 2;
# Line 812  for (;;) Line 858  for (;;)
858      case OP_ONCE_NC:      case OP_ONCE_NC:
859      prev = ecode;      prev = ecode;
860      saved_eptr = eptr;      saved_eptr = eptr;
861        save_mark = md->mark;
862      do      do
863        {        {
864        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, eptrb, RM64);        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, eptrb, RM64);
# Line 830  for (;;) Line 877  for (;;)
877    
878        if (rrc != MATCH_NOMATCH) RRETURN(rrc);        if (rrc != MATCH_NOMATCH) RRETURN(rrc);
879        ecode += GET(ecode,1);        ecode += GET(ecode,1);
880          md->mark = save_mark;
881        }        }
882      while (*ecode == OP_ALT);      while (*ecode == OP_ALT);
883    
# Line 869  for (;;) Line 917  for (;;)
917        }        }
918      else  /* OP_KETRMAX */      else  /* OP_KETRMAX */
919        {        {
       md->match_function_type = MATCH_CBEGROUP;  
920        RMATCH(eptr, prev, offset_top, md, eptrb, RM66);        RMATCH(eptr, prev, offset_top, md, eptrb, RM66);
921        if (rrc != MATCH_NOMATCH) RRETURN(rrc);        if (rrc != MATCH_NOMATCH) RRETURN(rrc);
922        ecode += 1 + LINK_SIZE;        ecode += 1 + LINK_SIZE;
# Line 909  for (;;) Line 956  for (;;)
956        save_offset2 = md->offset_vector[offset+1];        save_offset2 = md->offset_vector[offset+1];
957        save_offset3 = md->offset_vector[md->offset_end - number];        save_offset3 = md->offset_vector[md->offset_end - number];
958        save_capture_last = md->capture_last;        save_capture_last = md->capture_last;
959          save_mark = md->mark;
960    
961        DPRINTF(("saving %d %d %d\n", save_offset1, save_offset2, save_offset3));        DPRINTF(("saving %d %d %d\n", save_offset1, save_offset2, save_offset3));
962        md->offset_vector[md->offset_end - number] =        md->offset_vector[md->offset_end - number] =
# Line 917  for (;;) Line 965  for (;;)
965        for (;;)        for (;;)
966          {          {
967          if (op >= OP_SBRA) md->match_function_type = MATCH_CBEGROUP;          if (op >= OP_SBRA) md->match_function_type = MATCH_CBEGROUP;
968          RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,          RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md,
969            eptrb, RM1);            eptrb, RM1);
970          if (rrc == MATCH_ONCE) break;  /* Backing up through an atomic group */          if (rrc == MATCH_ONCE) break;  /* Backing up through an atomic group */
971    
# Line 945  for (;;) Line 993  for (;;)
993          if (rrc != MATCH_NOMATCH) RRETURN(rrc);          if (rrc != MATCH_NOMATCH) RRETURN(rrc);
994          md->capture_last = save_capture_last;          md->capture_last = save_capture_last;
995          ecode += GET(ecode, 1);          ecode += GET(ecode, 1);
996            md->mark = save_mark;
997          if (*ecode != OP_ALT) break;          if (*ecode != OP_ALT) break;
998          }          }
999    
# Line 996  for (;;) Line 1045  for (;;)
1045    
1046      for (;;)      for (;;)
1047        {        {
1048        if (op >= OP_SBRA || op == OP_ONCE) md->match_function_type = MATCH_CBEGROUP;        if (op >= OP_SBRA || op == OP_ONCE)
1049            md->match_function_type = MATCH_CBEGROUP;
1050    
1051        /* 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
1052        the pattern, and this is the final alternative, optimize as described        the pattern, and this is the final alternative, optimize as described
# Line 1004  for (;;) Line 1054  for (;;)
1054    
1055        else if (!md->hasthen && ecode[GET(ecode, 1)] != OP_ALT)        else if (!md->hasthen && ecode[GET(ecode, 1)] != OP_ALT)
1056          {          {
1057          ecode += _pcre_OP_lengths[*ecode];          ecode += PRIV(OP_lengths)[*ecode];
1058          goto TAIL_RECURSE;          goto TAIL_RECURSE;
1059          }          }
1060    
1061        /* In all other cases, we have to make another call to match(). */        /* In all other cases, we have to make another call to match(). */
1062    
1063        RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md, eptrb,        save_mark = md->mark;
1064          RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md, eptrb,
1065          RM2);          RM2);
1066    
1067        /* See comment in the code for capturing groups above about handling        /* See comment in the code for capturing groups above about handling
# Line 1028  for (;;) Line 1079  for (;;)
1079          {          {
1080          if (rrc == MATCH_ONCE)          if (rrc == MATCH_ONCE)
1081            {            {
1082            const uschar *scode = ecode;            const pcre_uchar *scode = ecode;
1083            if (*scode != OP_ONCE)           /* If not at start, find it */            if (*scode != OP_ONCE)           /* If not at start, find it */
1084              {              {
1085              while (*scode == OP_ALT) scode += GET(scode, 1);              while (*scode == OP_ALT) scode += GET(scode, 1);
# Line 1039  for (;;) Line 1090  for (;;)
1090          RRETURN(rrc);          RRETURN(rrc);
1091          }          }
1092        ecode += GET(ecode, 1);        ecode += GET(ecode, 1);
1093          md->mark = save_mark;
1094        if (*ecode != OP_ALT) break;        if (*ecode != OP_ALT) break;
1095        }        }
1096    
# Line 1070  for (;;) Line 1122  for (;;)
1122      if (offset < md->offset_max)      if (offset < md->offset_max)
1123        {        {
1124        matched_once = FALSE;        matched_once = FALSE;
1125        code_offset = ecode - md->start_code;        code_offset = (int)(ecode - md->start_code);
1126    
1127        save_offset1 = md->offset_vector[offset];        save_offset1 = md->offset_vector[offset];
1128        save_offset2 = md->offset_vector[offset+1];        save_offset2 = md->offset_vector[offset+1];
# Line 1093  for (;;) Line 1145  for (;;)
1145          md->offset_vector[md->offset_end - number] =          md->offset_vector[md->offset_end - number] =
1146            (int)(eptr - md->start_subject);            (int)(eptr - md->start_subject);
1147          if (op >= OP_SBRA) md->match_function_type = MATCH_CBEGROUP;          if (op >= OP_SBRA) md->match_function_type = MATCH_CBEGROUP;
1148          RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,          RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md,
1149            eptrb, RM63);            eptrb, RM63);
1150          if (rrc == MATCH_KETRPOS)          if (rrc == MATCH_KETRPOS)
1151            {            {
# Line 1160  for (;;) Line 1212  for (;;)
1212    
1213      POSSESSIVE_NON_CAPTURE:      POSSESSIVE_NON_CAPTURE:
1214      matched_once = FALSE;      matched_once = FALSE;
1215      code_offset = ecode - md->start_code;      code_offset = (int)(ecode - md->start_code);
1216    
1217      for (;;)      for (;;)
1218        {        {
1219        if (op >= OP_SBRA) md->match_function_type = MATCH_CBEGROUP;        if (op >= OP_SBRA) md->match_function_type = MATCH_CBEGROUP;
1220        RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,        RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md,
1221          eptrb, RM48);          eptrb, RM48);
1222        if (rrc == MATCH_KETRPOS)        if (rrc == MATCH_KETRPOS)
1223          {          {
# Line 1215  for (;;) Line 1267  for (;;)
1267    
1268      if (ecode[LINK_SIZE+1] == OP_CALLOUT)      if (ecode[LINK_SIZE+1] == OP_CALLOUT)
1269        {        {
1270        if (pcre_callout != NULL)        if (PUBL(callout) != NULL)
1271          {          {
1272          pcre_callout_block cb;          PUBL(callout_block) cb;
1273          cb.version          = 2;   /* Version 1 of the callout block */          cb.version          = 2;   /* Version 1 of the callout block */
1274          cb.callout_number   = ecode[LINK_SIZE+2];          cb.callout_number   = ecode[LINK_SIZE+2];
1275          cb.offset_vector    = md->offset_vector;          cb.offset_vector    = md->offset_vector;
1276    #if defined COMPILE_PCRE8
1277          cb.subject          = (PCRE_SPTR)md->start_subject;          cb.subject          = (PCRE_SPTR)md->start_subject;
1278    #elif defined COMPILE_PCRE16
1279            cb.subject          = (PCRE_SPTR16)md->start_subject;
1280    #elif defined COMPILE_PCRE32
1281            cb.subject          = (PCRE_SPTR32)md->start_subject;
1282    #endif
1283          cb.subject_length   = (int)(md->end_subject - md->start_subject);          cb.subject_length   = (int)(md->end_subject - md->start_subject);
1284          cb.start_match      = (int)(mstart - md->start_subject);          cb.start_match      = (int)(mstart - md->start_subject);
1285          cb.current_position = (int)(eptr - md->start_subject);          cb.current_position = (int)(eptr - md->start_subject);
# Line 1231  for (;;) Line 1289  for (;;)
1289          cb.capture_last     = md->capture_last;          cb.capture_last     = md->capture_last;
1290          cb.callout_data     = md->callout_data;          cb.callout_data     = md->callout_data;
1291          cb.mark             = md->nomatch_mark;          cb.mark             = md->nomatch_mark;
1292          if ((rrc = (*pcre_callout)(&cb)) > 0) RRETURN(MATCH_NOMATCH);          if ((rrc = (*PUBL(callout))(&cb)) > 0) RRETURN(MATCH_NOMATCH);
1293          if (rrc < 0) RRETURN(rrc);          if (rrc < 0) RRETURN(rrc);
1294          }          }
1295        ecode += _pcre_OP_lengths[OP_CALLOUT];        ecode += PRIV(OP_lengths)[OP_CALLOUT];
1296        }        }
1297    
1298      condcode = ecode[LINK_SIZE+1];      condcode = ecode[LINK_SIZE+1];
# Line 1260  for (;;) Line 1318  for (;;)
1318    
1319          if (!condition && condcode == OP_NRREF)          if (!condition && condcode == OP_NRREF)
1320            {            {
1321            uschar *slotA = md->name_table;            pcre_uchar *slotA = md->name_table;
1322            for (i = 0; i < md->name_count; i++)            for (i = 0; i < md->name_count; i++)
1323              {              {
1324              if (GET2(slotA, 0) == recno) break;              if (GET2(slotA, 0) == recno) break;
# Line 1273  for (;;) Line 1331  for (;;)
1331    
1332            if (i < md->name_count)            if (i < md->name_count)
1333              {              {
1334              uschar *slotB = slotA;              pcre_uchar *slotB = slotA;
1335              while (slotB > md->name_table)              while (slotB > md->name_table)
1336                {                {
1337                slotB -= md->name_entry_size;                slotB -= md->name_entry_size;
1338                if (strcmp((char *)slotA + 2, (char *)slotB + 2) == 0)                if (STRCMP_UC_UC(slotA + IMM2_SIZE, slotB + IMM2_SIZE) == 0)
1339                  {                  {
1340                  condition = GET2(slotB, 0) == md->recursive->group_num;                  condition = GET2(slotB, 0) == md->recursive->group_num;
1341                  if (condition) break;                  if (condition) break;
# Line 1293  for (;;) Line 1351  for (;;)
1351                for (i++; i < md->name_count; i++)                for (i++; i < md->name_count; i++)
1352                  {                  {
1353                  slotB += md->name_entry_size;                  slotB += md->name_entry_size;
1354                  if (strcmp((char *)slotA + 2, (char *)slotB + 2) == 0)                  if (STRCMP_UC_UC(slotA + IMM2_SIZE, slotB + IMM2_SIZE) == 0)
1355                    {                    {
1356                    condition = GET2(slotB, 0) == md->recursive->group_num;                    condition = GET2(slotB, 0) == md->recursive->group_num;
1357                    if (condition) break;                    if (condition) break;
# Line 1306  for (;;) Line 1364  for (;;)
1364    
1365          /* Chose branch according to the condition */          /* Chose branch according to the condition */
1366    
1367          ecode += condition? 3 : GET(ecode, 1);          ecode += condition? 1 + IMM2_SIZE : GET(ecode, 1);
1368          }          }
1369        }        }
1370    
# Line 1323  for (;;) Line 1381  for (;;)
1381        if (!condition && condcode == OP_NCREF)        if (!condition && condcode == OP_NCREF)
1382          {          {
1383          int refno = offset >> 1;          int refno = offset >> 1;
1384          uschar *slotA = md->name_table;          pcre_uchar *slotA = md->name_table;
1385    
1386          for (i = 0; i < md->name_count; i++)          for (i = 0; i < md->name_count; i++)
1387            {            {
# Line 1337  for (;;) Line 1395  for (;;)
1395    
1396          if (i < md->name_count)          if (i < md->name_count)
1397            {            {
1398            uschar *slotB = slotA;            pcre_uchar *slotB = slotA;
1399            while (slotB > md->name_table)            while (slotB > md->name_table)
1400              {              {
1401              slotB -= md->name_entry_size;              slotB -= md->name_entry_size;
1402              if (strcmp((char *)slotA + 2, (char *)slotB + 2) == 0)              if (STRCMP_UC_UC(slotA + IMM2_SIZE, slotB + IMM2_SIZE) == 0)
1403                {                {
1404                offset = GET2(slotB, 0) << 1;                offset = GET2(slotB, 0) << 1;
1405                condition = offset < offset_top &&                condition = offset < offset_top &&
# Line 1359  for (;;) Line 1417  for (;;)
1417              for (i++; i < md->name_count; i++)              for (i++; i < md->name_count; i++)
1418                {                {
1419                slotB += md->name_entry_size;                slotB += md->name_entry_size;
1420                if (strcmp((char *)slotA + 2, (char *)slotB + 2) == 0)                if (STRCMP_UC_UC(slotA + IMM2_SIZE, slotB + IMM2_SIZE) == 0)
1421                  {                  {
1422                  offset = GET2(slotB, 0) << 1;                  offset = GET2(slotB, 0) << 1;
1423                  condition = offset < offset_top &&                  condition = offset < offset_top &&
# Line 1374  for (;;) Line 1432  for (;;)
1432    
1433        /* Chose branch according to the condition */        /* Chose branch according to the condition */
1434    
1435        ecode += condition? 3 : GET(ecode, 1);        ecode += condition? 1 + IMM2_SIZE : GET(ecode, 1);
1436        }        }
1437    
1438      else if (condcode == OP_DEF)     /* DEFINE - always false */      else if (condcode == OP_DEF)     /* DEFINE - always false */
# Line 1466  for (;;) Line 1524  for (;;)
1524        md->offset_vector[offset+1] = (int)(eptr - md->start_subject);        md->offset_vector[offset+1] = (int)(eptr - md->start_subject);
1525        if (offset_top <= offset) offset_top = offset + 2;        if (offset_top <= offset) offset_top = offset + 2;
1526        }        }
1527      ecode += 3;      ecode += 1 + IMM2_SIZE;
1528      break;      break;
1529    
1530    
# Line 1513  for (;;) Line 1571  for (;;)
1571    
1572      case OP_ASSERT:      case OP_ASSERT:
1573      case OP_ASSERTBACK:      case OP_ASSERTBACK:
1574        save_mark = md->mark;
1575      if (md->match_function_type == MATCH_CONDASSERT)      if (md->match_function_type == MATCH_CONDASSERT)
1576        {        {
1577        condassert = TRUE;        condassert = TRUE;
# Line 1528  for (;;) Line 1587  for (;;)
1587          mstart = md->start_match_ptr;   /* In case \K reset it */          mstart = md->start_match_ptr;   /* In case \K reset it */
1588          break;          break;
1589          }          }
1590          md->mark = save_mark;
1591    
1592        /* PCRE does not allow THEN to escape beyond an assertion; it is treated        /* A COMMIT failure must fail the entire assertion, without trying any
1593        as NOMATCH. */        subsequent branches. */
1594    
1595          if (rrc == MATCH_COMMIT) RRETURN(MATCH_NOMATCH);
1596    
1597          /* PCRE does not allow THEN to escape beyond an assertion; it
1598          is treated as NOMATCH. */
1599    
1600        if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN) RRETURN(rrc);        if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN) RRETURN(rrc);
1601        ecode += GET(ecode, 1);        ecode += GET(ecode, 1);
# Line 1557  for (;;) Line 1622  for (;;)
1622    
1623      case OP_ASSERT_NOT:      case OP_ASSERT_NOT:
1624      case OP_ASSERTBACK_NOT:      case OP_ASSERTBACK_NOT:
1625        save_mark = md->mark;
1626      if (md->match_function_type == MATCH_CONDASSERT)      if (md->match_function_type == MATCH_CONDASSERT)
1627        {        {
1628        condassert = TRUE;        condassert = TRUE;
# Line 1567  for (;;) Line 1633  for (;;)
1633      do      do
1634        {        {
1635        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, NULL, RM5);        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, NULL, RM5);
1636          md->mark = save_mark;
1637        if (rrc == MATCH_MATCH || rrc == MATCH_ACCEPT) RRETURN(MATCH_NOMATCH);        if (rrc == MATCH_MATCH || rrc == MATCH_ACCEPT) RRETURN(MATCH_NOMATCH);
1638        if (rrc == MATCH_SKIP || rrc == MATCH_PRUNE || rrc == MATCH_COMMIT)        if (rrc == MATCH_SKIP || rrc == MATCH_PRUNE || rrc == MATCH_COMMIT)
1639          {          {
# Line 1593  for (;;) Line 1660  for (;;)
1660      back a number of characters, not bytes. */      back a number of characters, not bytes. */
1661    
1662      case OP_REVERSE:      case OP_REVERSE:
1663  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
1664      if (utf8)      if (utf)
1665        {        {
1666        i = GET(ecode, 1);        i = GET(ecode, 1);
1667        while (i-- > 0)        while (i-- > 0)
# Line 1625  for (;;) Line 1692  for (;;)
1692      function is able to force a failure. */      function is able to force a failure. */
1693    
1694      case OP_CALLOUT:      case OP_CALLOUT:
1695      if (pcre_callout != NULL)      if (PUBL(callout) != NULL)
1696        {        {
1697        pcre_callout_block cb;        PUBL(callout_block) cb;
1698        cb.version          = 2;   /* Version 1 of the callout block */        cb.version          = 2;   /* Version 1 of the callout block */
1699        cb.callout_number   = ecode[1];        cb.callout_number   = ecode[1];
1700        cb.offset_vector    = md->offset_vector;        cb.offset_vector    = md->offset_vector;
1701    #if defined COMPILE_PCRE8
1702        cb.subject          = (PCRE_SPTR)md->start_subject;        cb.subject          = (PCRE_SPTR)md->start_subject;
1703    #elif defined COMPILE_PCRE16
1704          cb.subject          = (PCRE_SPTR16)md->start_subject;
1705    #elif defined COMPILE_PCRE32
1706          cb.subject          = (PCRE_SPTR32)md->start_subject;
1707    #endif
1708        cb.subject_length   = (int)(md->end_subject - md->start_subject);        cb.subject_length   = (int)(md->end_subject - md->start_subject);
1709        cb.start_match      = (int)(mstart - md->start_subject);        cb.start_match      = (int)(mstart - md->start_subject);
1710        cb.current_position = (int)(eptr - md->start_subject);        cb.current_position = (int)(eptr - md->start_subject);
# Line 1641  for (;;) Line 1714  for (;;)
1714        cb.capture_last     = md->capture_last;        cb.capture_last     = md->capture_last;
1715        cb.callout_data     = md->callout_data;        cb.callout_data     = md->callout_data;
1716        cb.mark             = md->nomatch_mark;        cb.mark             = md->nomatch_mark;
1717        if ((rrc = (*pcre_callout)(&cb)) > 0) RRETURN(MATCH_NOMATCH);        if ((rrc = (*PUBL(callout))(&cb)) > 0) RRETURN(MATCH_NOMATCH);
1718        if (rrc < 0) RRETURN(rrc);        if (rrc < 0) RRETURN(rrc);
1719        }        }
1720      ecode += 2 + 2*LINK_SIZE;      ecode += 2 + 2*LINK_SIZE;
# Line 1700  for (;;) Line 1773  for (;;)
1773        else        else
1774          {          {
1775          new_recursive.offset_save =          new_recursive.offset_save =
1776            (int *)(pcre_malloc)(new_recursive.saved_max * sizeof(int));            (int *)(PUBL(malloc))(new_recursive.saved_max * sizeof(int));
1777          if (new_recursive.offset_save == NULL) RRETURN(PCRE_ERROR_NOMEMORY);          if (new_recursive.offset_save == NULL) RRETURN(PCRE_ERROR_NOMEMORY);
1778          }          }
1779        memcpy(new_recursive.offset_save, md->offset_vector,        memcpy(new_recursive.offset_save, md->offset_vector,
# Line 1715  for (;;) Line 1788  for (;;)
1788        do        do
1789          {          {
1790          if (cbegroup) md->match_function_type = MATCH_CBEGROUP;          if (cbegroup) md->match_function_type = MATCH_CBEGROUP;
1791          RMATCH(eptr, callpat + _pcre_OP_lengths[*callpat], offset_top,          RMATCH(eptr, callpat + PRIV(OP_lengths)[*callpat], offset_top,
1792            md, eptrb, RM6);            md, eptrb, RM6);
1793          memcpy(md->offset_vector, new_recursive.offset_save,          memcpy(md->offset_vector, new_recursive.offset_save,
1794              new_recursive.saved_max * sizeof(int));              new_recursive.saved_max * sizeof(int));
# Line 1724  for (;;) Line 1797  for (;;)
1797            {            {
1798            DPRINTF(("Recursion matched\n"));            DPRINTF(("Recursion matched\n"));
1799            if (new_recursive.offset_save != stacksave)            if (new_recursive.offset_save != stacksave)
1800              (pcre_free)(new_recursive.offset_save);              (PUBL(free))(new_recursive.offset_save);
1801    
1802            /* 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
1803            it was changed by \K. This *is* propagated back out of a recursion,            it was changed by \K. This *is* propagated back out of a recursion,
# Line 1735  for (;;) Line 1808  for (;;)
1808            goto RECURSION_MATCHED;        /* Exit loop; end processing */            goto RECURSION_MATCHED;        /* Exit loop; end processing */
1809            }            }
1810    
1811          /* 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
1812          as NOMATCH. */          is treated as NOMATCH. */
1813    
1814          else if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN)          else if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN &&
1815                     rrc != MATCH_COMMIT)
1816            {            {
1817            DPRINTF(("Recursion gave error %d\n", rrc));            DPRINTF(("Recursion gave error %d\n", rrc));
1818            if (new_recursive.offset_save != stacksave)            if (new_recursive.offset_save != stacksave)
1819              (pcre_free)(new_recursive.offset_save);              (PUBL(free))(new_recursive.offset_save);
1820            RRETURN(rrc);            RRETURN(rrc);
1821            }            }
1822    
# Line 1754  for (;;) Line 1828  for (;;)
1828        DPRINTF(("Recursion didn't match\n"));        DPRINTF(("Recursion didn't match\n"));
1829        md->recursive = new_recursive.prevrec;        md->recursive = new_recursive.prevrec;
1830        if (new_recursive.offset_save != stacksave)        if (new_recursive.offset_save != stacksave)
1831          (pcre_free)(new_recursive.offset_save);          (PUBL(free))(new_recursive.offset_save);
1832        RRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
1833        }        }
1834    
# Line 1949  for (;;) Line 2023  for (;;)
2023          }          }
2024        if (*prev >= OP_SBRA)    /* Could match an empty string */        if (*prev >= OP_SBRA)    /* Could match an empty string */
2025          {          {
         md->match_function_type = MATCH_CBEGROUP;  
2026          RMATCH(eptr, prev, offset_top, md, eptrb, RM50);          RMATCH(eptr, prev, offset_top, md, eptrb, RM50);
2027          RRETURN(rrc);          RRETURN(rrc);
2028          }          }
# Line 1958  for (;;) Line 2031  for (;;)
2031        }        }
2032      else  /* OP_KETRMAX */      else  /* OP_KETRMAX */
2033        {        {
       if (*prev >= OP_SBRA) md->match_function_type = MATCH_CBEGROUP;  
2034        RMATCH(eptr, prev, offset_top, md, eptrb, RM13);        RMATCH(eptr, prev, offset_top, md, eptrb, RM13);
2035        if (rrc == MATCH_ONCE && md->once_target == prev) rrc = MATCH_NOMATCH;        if (rrc == MATCH_ONCE && md->once_target == prev) rrc = MATCH_NOMATCH;
2036        if (rrc != MATCH_NOMATCH) RRETURN(rrc);        if (rrc != MATCH_NOMATCH) RRETURN(rrc);
# Line 2015  for (;;) Line 2087  for (;;)
2087    
2088      case OP_DOLLM:      case OP_DOLLM:
2089      if (eptr < md->end_subject)      if (eptr < md->end_subject)
2090        { if (!IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH); }        {
2091          if (!IS_NEWLINE(eptr))
2092            {
2093            if (md->partial != 0 &&
2094                eptr + 1 >= md->end_subject &&
2095                NLBLOCK->nltype == NLTYPE_FIXED &&
2096                NLBLOCK->nllen == 2 &&
2097                *eptr == NLBLOCK->nl[0])
2098              {
2099              md->hitend = TRUE;
2100              if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
2101              }
2102            RRETURN(MATCH_NOMATCH);
2103            }
2104          }
2105      else      else
2106        {        {
2107        if (md->noteol) RRETURN(MATCH_NOMATCH);        if (md->noteol) RRETURN(MATCH_NOMATCH);
# Line 2047  for (;;) Line 2133  for (;;)
2133      ASSERT_NL_OR_EOS:      ASSERT_NL_OR_EOS:
2134      if (eptr < md->end_subject &&      if (eptr < md->end_subject &&
2135          (!IS_NEWLINE(eptr) || eptr != md->end_subject - md->nllen))          (!IS_NEWLINE(eptr) || eptr != md->end_subject - md->nllen))
2136          {
2137          if (md->partial != 0 &&
2138              eptr + 1 >= md->end_subject &&
2139              NLBLOCK->nltype == NLTYPE_FIXED &&
2140              NLBLOCK->nllen == 2 &&
2141              *eptr == NLBLOCK->nl[0])
2142            {
2143            md->hitend = TRUE;
2144            if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
2145            }
2146        RRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2147          }
2148    
2149      /* Either at end of string or \n before end. */      /* Either at end of string or \n before end. */
2150    
# Line 2066  for (;;) Line 2163  for (;;)
2163        be "non-word" characters. Remember the earliest consulted character for        be "non-word" characters. Remember the earliest consulted character for
2164        partial matching. */        partial matching. */
2165    
2166  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
2167        if (utf8)        if (utf)
2168          {          {
2169          /* Get status of previous character */          /* Get status of previous character */
2170    
2171          if (eptr == md->start_subject) prev_is_word = FALSE; else          if (eptr == md->start_subject) prev_is_word = FALSE; else
2172            {            {
2173            USPTR lastptr = eptr - 1;            PCRE_PUCHAR lastptr = eptr - 1;
2174            while((*lastptr & 0xc0) == 0x80) lastptr--;            BACKCHAR(lastptr);
2175            if (lastptr < md->start_used_ptr) md->start_used_ptr = lastptr;            if (lastptr < md->start_used_ptr) md->start_used_ptr = lastptr;
2176            GETCHAR(c, lastptr);            GETCHAR(c, lastptr);
2177  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
# Line 2139  for (;;) Line 2236  for (;;)
2236              }              }
2237            else            else
2238  #endif  #endif
2239            prev_is_word = ((md->ctypes[eptr[-1]] & ctype_word) != 0);            prev_is_word = MAX_255(eptr[-1])
2240                && ((md->ctypes[eptr[-1]] & ctype_word) != 0);
2241            }            }
2242    
2243          /* Get status of next character */          /* Get status of next character */
# Line 2162  for (;;) Line 2260  for (;;)
2260            }            }
2261          else          else
2262  #endif  #endif
2263          cur_is_word = ((md->ctypes[*eptr] & ctype_word) != 0);          cur_is_word = MAX_255(*eptr)
2264              && ((md->ctypes[*eptr] & ctype_word) != 0);
2265          }          }
2266    
2267        /* Now see if the situation is what we want */        /* Now see if the situation is what we want */
# Line 2173  for (;;) Line 2272  for (;;)
2272        }        }
2273      break;      break;
2274    
2275      /* Match a single character type; inline for speed */      /* Match any single character type except newline; have to take care with
2276        CRLF newlines and partial matching. */
2277    
2278      case OP_ANY:      case OP_ANY:
2279      if (IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH);      if (IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH);
2280        if (md->partial != 0 &&
2281            eptr + 1 >= md->end_subject &&
2282            NLBLOCK->nltype == NLTYPE_FIXED &&
2283            NLBLOCK->nllen == 2 &&
2284            *eptr == NLBLOCK->nl[0])
2285          {
2286          md->hitend = TRUE;
2287          if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
2288          }
2289    
2290      /* Fall through */      /* Fall through */
2291    
2292        /* Match any single character whatsoever. */
2293    
2294      case OP_ALLANY:      case OP_ALLANY:
2295      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 */
2296        {                            /* not be updated before SCHECK_PARTIAL. */        {                            /* not be updated before SCHECK_PARTIAL. */
# Line 2186  for (;;) Line 2298  for (;;)
2298        RRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2299        }        }
2300      eptr++;      eptr++;
2301      if (utf8) while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;  #ifdef SUPPORT_UTF
2302        if (utf) ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);
2303    #endif
2304      ecode++;      ecode++;
2305      break;      break;
2306    
# Line 2211  for (;;) Line 2325  for (;;)
2325        }        }
2326      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2327      if (      if (
2328  #ifdef SUPPORT_UTF8  #if defined SUPPORT_UTF || !(defined COMPILE_PCRE8)
2329         c < 256 &&         c < 256 &&
2330  #endif  #endif
2331         (md->ctypes[c] & ctype_digit) != 0         (md->ctypes[c] & ctype_digit) != 0
# Line 2228  for (;;) Line 2342  for (;;)
2342        }        }
2343      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2344      if (      if (
2345  #ifdef SUPPORT_UTF8  #if defined SUPPORT_UTF || !(defined COMPILE_PCRE8)
2346         c >= 256 ||         c > 255 ||
2347  #endif  #endif
2348         (md->ctypes[c] & ctype_digit) == 0         (md->ctypes[c] & ctype_digit) == 0
2349         )         )
# Line 2245  for (;;) Line 2359  for (;;)
2359        }        }
2360      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2361      if (      if (
2362  #ifdef SUPPORT_UTF8  #if defined SUPPORT_UTF || !(defined COMPILE_PCRE8)
2363         c < 256 &&         c < 256 &&
2364  #endif  #endif
2365         (md->ctypes[c] & ctype_space) != 0         (md->ctypes[c] & ctype_space) != 0
# Line 2262  for (;;) Line 2376  for (;;)
2376        }        }
2377      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2378      if (      if (
2379  #ifdef SUPPORT_UTF8  #if defined SUPPORT_UTF || !(defined COMPILE_PCRE8)
2380         c >= 256 ||         c > 255 ||
2381  #endif  #endif
2382         (md->ctypes[c] & ctype_space) == 0         (md->ctypes[c] & ctype_space) == 0
2383         )         )
# Line 2279  for (;;) Line 2393  for (;;)
2393        }        }
2394      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2395      if (      if (
2396  #ifdef SUPPORT_UTF8  #if defined SUPPORT_UTF || !(defined COMPILE_PCRE8)
2397         c < 256 &&         c < 256 &&
2398  #endif  #endif
2399         (md->ctypes[c] & ctype_word) != 0         (md->ctypes[c] & ctype_word) != 0
# Line 2296  for (;;) Line 2410  for (;;)
2410        }        }
2411      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2412      if (      if (
2413  #ifdef SUPPORT_UTF8  #if defined SUPPORT_UTF || !(defined COMPILE_PCRE8)
2414         c >= 256 ||         c > 255 ||
2415  #endif  #endif
2416         (md->ctypes[c] & ctype_word) == 0         (md->ctypes[c] & ctype_word) == 0
2417         )         )
# Line 2316  for (;;) Line 2430  for (;;)
2430        {        {
2431        default: RRETURN(MATCH_NOMATCH);        default: RRETURN(MATCH_NOMATCH);
2432    
2433        case 0x000d:        case CHAR_CR:
2434        if (eptr < md->end_subject && *eptr == 0x0a) eptr++;        if (eptr >= md->end_subject)
2435            {
2436            SCHECK_PARTIAL();
2437            }
2438          else if (*eptr == CHAR_LF) eptr++;
2439        break;        break;
2440    
2441        case 0x000a:        case CHAR_LF:
2442        break;        break;
2443    
2444        case 0x000b:        case CHAR_VT:
2445        case 0x000c:        case CHAR_FF:
2446        case 0x0085:        case CHAR_NEL:
2447    #ifndef EBCDIC
2448        case 0x2028:        case 0x2028:
2449        case 0x2029:        case 0x2029:
2450    #endif  /* Not EBCDIC */
2451        if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);        if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);
2452        break;        break;
2453        }        }
# Line 2343  for (;;) Line 2463  for (;;)
2463      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2464      switch(c)      switch(c)
2465        {        {
2466          HSPACE_CASES: RRETURN(MATCH_NOMATCH);  /* Byte and multibyte cases */
2467        default: break;        default: break;
       case 0x09:      /* HT */  
       case 0x20:      /* SPACE */  
       case 0xa0:      /* NBSP */  
       case 0x1680:    /* OGHAM SPACE MARK */  
       case 0x180e:    /* MONGOLIAN VOWEL SEPARATOR */  
       case 0x2000:    /* EN QUAD */  
       case 0x2001:    /* EM QUAD */  
       case 0x2002:    /* EN SPACE */  
       case 0x2003:    /* EM SPACE */  
       case 0x2004:    /* THREE-PER-EM SPACE */  
       case 0x2005:    /* FOUR-PER-EM SPACE */  
       case 0x2006:    /* SIX-PER-EM SPACE */  
       case 0x2007:    /* FIGURE SPACE */  
       case 0x2008:    /* PUNCTUATION SPACE */  
       case 0x2009:    /* THIN SPACE */  
       case 0x200A:    /* HAIR SPACE */  
       case 0x202f:    /* NARROW NO-BREAK SPACE */  
       case 0x205f:    /* MEDIUM MATHEMATICAL SPACE */  
       case 0x3000:    /* IDEOGRAPHIC SPACE */  
       RRETURN(MATCH_NOMATCH);  
2468        }        }
2469      ecode++;      ecode++;
2470      break;      break;
# Line 2377  for (;;) Line 2478  for (;;)
2478      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2479      switch(c)      switch(c)
2480        {        {
2481          HSPACE_CASES: break;  /* Byte and multibyte cases */
2482        default: RRETURN(MATCH_NOMATCH);        default: RRETURN(MATCH_NOMATCH);
       case 0x09:      /* HT */  
       case 0x20:      /* SPACE */  
       case 0xa0:      /* NBSP */  
       case 0x1680:    /* OGHAM SPACE MARK */  
       case 0x180e:    /* MONGOLIAN VOWEL SEPARATOR */  
       case 0x2000:    /* EN QUAD */  
       case 0x2001:    /* EM QUAD */  
       case 0x2002:    /* EN SPACE */  
       case 0x2003:    /* EM SPACE */  
       case 0x2004:    /* THREE-PER-EM SPACE */  
       case 0x2005:    /* FOUR-PER-EM SPACE */  
       case 0x2006:    /* SIX-PER-EM SPACE */  
       case 0x2007:    /* FIGURE SPACE */  
       case 0x2008:    /* PUNCTUATION SPACE */  
       case 0x2009:    /* THIN SPACE */  
       case 0x200A:    /* HAIR SPACE */  
       case 0x202f:    /* NARROW NO-BREAK SPACE */  
       case 0x205f:    /* MEDIUM MATHEMATICAL SPACE */  
       case 0x3000:    /* IDEOGRAPHIC SPACE */  
       break;  
2483        }        }
2484      ecode++;      ecode++;
2485      break;      break;
# Line 2411  for (;;) Line 2493  for (;;)
2493      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2494      switch(c)      switch(c)
2495        {        {
2496          VSPACE_CASES: RRETURN(MATCH_NOMATCH);
2497        default: break;        default: break;
       case 0x0a:      /* LF */  
       case 0x0b:      /* VT */  
       case 0x0c:      /* FF */  
       case 0x0d:      /* CR */  
       case 0x85:      /* NEL */  
       case 0x2028:    /* LINE SEPARATOR */  
       case 0x2029:    /* PARAGRAPH SEPARATOR */  
       RRETURN(MATCH_NOMATCH);  
2498        }        }
2499      ecode++;      ecode++;
2500      break;      break;
# Line 2433  for (;;) Line 2508  for (;;)
2508      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2509      switch(c)      switch(c)
2510        {        {
2511          VSPACE_CASES: break;
2512        default: RRETURN(MATCH_NOMATCH);        default: RRETURN(MATCH_NOMATCH);
       case 0x0a:      /* LF */  
       case 0x0b:      /* VT */  
       case 0x0c:      /* FF */  
       case 0x0d:      /* CR */  
       case 0x85:      /* NEL */  
       case 0x2028:    /* LINE SEPARATOR */  
       case 0x2029:    /* PARAGRAPH SEPARATOR */  
       break;  
2513        }        }
2514      ecode++;      ecode++;
2515      break;      break;
# Line 2459  for (;;) Line 2527  for (;;)
2527        }        }
2528      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2529        {        {
2530          const pcre_uint32 *cp;
2531        const ucd_record *prop = GET_UCD(c);        const ucd_record *prop = GET_UCD(c);
2532    
2533        switch(ecode[1])        switch(ecode[1])
# Line 2475  for (;;) Line 2544  for (;;)
2544          break;          break;
2545    
2546          case PT_GC:          case PT_GC:
2547          if ((ecode[2] != _pcre_ucp_gentype[prop->chartype]) == (op == OP_PROP))          if ((ecode[2] != PRIV(ucp_gentype)[prop->chartype]) == (op == OP_PROP))
2548            RRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
2549          break;          break;
2550    
# Line 2492  for (;;) Line 2561  for (;;)
2561          /* These are specials */          /* These are specials */
2562    
2563          case PT_ALNUM:          case PT_ALNUM:
2564          if ((_pcre_ucp_gentype[prop->chartype] == ucp_L ||          if ((PRIV(ucp_gentype)[prop->chartype] == ucp_L ||
2565               _pcre_ucp_gentype[prop->chartype] == ucp_N) == (op == OP_NOTPROP))               PRIV(ucp_gentype)[prop->chartype] == ucp_N) == (op == OP_NOTPROP))
2566            RRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
2567          break;          break;
2568    
2569          case PT_SPACE:    /* Perl space */          case PT_SPACE:    /* Perl space */
2570          if ((_pcre_ucp_gentype[prop->chartype] == ucp_Z ||          if ((PRIV(ucp_gentype)[prop->chartype] == ucp_Z ||
2571               c == CHAR_HT || c == CHAR_NL || c == CHAR_FF || c == CHAR_CR)               c == CHAR_HT || c == CHAR_NL || c == CHAR_FF || c == CHAR_CR)
2572                 == (op == OP_NOTPROP))                 == (op == OP_NOTPROP))
2573            RRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
2574          break;          break;
2575    
2576          case PT_PXSPACE:  /* POSIX space */          case PT_PXSPACE:  /* POSIX space */
2577          if ((_pcre_ucp_gentype[prop->chartype] == ucp_Z ||          if ((PRIV(ucp_gentype)[prop->chartype] == ucp_Z ||
2578               c == CHAR_HT || c == CHAR_NL || c == CHAR_VT ||               c == CHAR_HT || c == CHAR_NL || c == CHAR_VT ||
2579               c == CHAR_FF || c == CHAR_CR)               c == CHAR_FF || c == CHAR_CR)
2580                 == (op == OP_NOTPROP))                 == (op == OP_NOTPROP))
# Line 2513  for (;;) Line 2582  for (;;)
2582          break;          break;
2583    
2584          case PT_WORD:          case PT_WORD:
2585          if ((_pcre_ucp_gentype[prop->chartype] == ucp_L ||          if ((PRIV(ucp_gentype)[prop->chartype] == ucp_L ||
2586               _pcre_ucp_gentype[prop->chartype] == ucp_N ||               PRIV(ucp_gentype)[prop->chartype] == ucp_N ||
2587               c == CHAR_UNDERSCORE) == (op == OP_NOTPROP))               c == CHAR_UNDERSCORE) == (op == OP_NOTPROP))
2588            RRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
2589          break;          break;
2590    
2591            case PT_CLIST:
2592            cp = PRIV(ucd_caseless_sets) + prop->caseset;
2593            for (;;)
2594              {
2595              if (c < *cp)
2596                { if (op == OP_PROP) { RRETURN(MATCH_NOMATCH); } else break; }
2597              if (c == *cp++)
2598                { if (op == OP_PROP) break; else { RRETURN(MATCH_NOMATCH); } }
2599              }
2600            break;
2601    
2602          /* This should never occur */          /* This should never occur */
2603    
2604          default:          default:
# Line 2538  for (;;) Line 2618  for (;;)
2618        SCHECK_PARTIAL();        SCHECK_PARTIAL();
2619        RRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2620        }        }
2621      GETCHARINCTEST(c, eptr);      else
     if (UCD_CATEGORY(c) == ucp_M) RRETURN(MATCH_NOMATCH);  
     while (eptr < md->end_subject)  
2622        {        {
2623        int len = 1;        int lgb, rgb;
2624        if (!utf8) c = *eptr; else { GETCHARLEN(c, eptr, len); }        GETCHARINCTEST(c, eptr);
2625        if (UCD_CATEGORY(c) != ucp_M) break;        lgb = UCD_GRAPHBREAK(c);
2626        eptr += len;        while (eptr < md->end_subject)
2627            {
2628            int len = 1;
2629            if (!utf) c = *eptr; else { GETCHARLEN(c, eptr, len); }
2630            rgb = UCD_GRAPHBREAK(c);
2631            if ((PRIV(ucp_gbtable)[lgb] & (1 << rgb)) == 0) break;
2632            lgb = rgb;
2633            eptr += len;
2634            }
2635        }        }
2636        CHECK_PARTIAL();
2637      ecode++;      ecode++;
2638      break;      break;
2639  #endif  #endif  /* SUPPORT_UCP */
2640    
2641    
2642      /* Match a back reference, possibly repeatedly. Look past the end of the      /* Match a back reference, possibly repeatedly. Look past the end of the
# Line 2564  for (;;) Line 2651  for (;;)
2651      case OP_REFI:      case OP_REFI:
2652      caseless = op == OP_REFI;      caseless = op == OP_REFI;
2653      offset = GET2(ecode, 1) << 1;               /* Doubled ref number */      offset = GET2(ecode, 1) << 1;               /* Doubled ref number */
2654      ecode += 3;      ecode += 1 + IMM2_SIZE;
2655    
2656      /* If the reference is unset, there are two possibilities:      /* If the reference is unset, there are two possibilities:
2657    
# Line 2604  for (;;) Line 2691  for (;;)
2691        case OP_CRMINRANGE:        case OP_CRMINRANGE:
2692        minimize = (*ecode == OP_CRMINRANGE);        minimize = (*ecode == OP_CRMINRANGE);
2693        min = GET2(ecode, 1);        min = GET2(ecode, 1);
2694        max = GET2(ecode, 3);        max = GET2(ecode, 1 + IMM2_SIZE);
2695        if (max == 0) max = INT_MAX;        if (max == 0) max = INT_MAX;
2696        ecode += 5;        ecode += 1 + 2 * IMM2_SIZE;
2697        break;        break;
2698    
2699        default:               /* No repeat follows */        default:               /* No repeat follows */
2700        if ((length = match_ref(offset, eptr, length, md, caseless)) < 0)        if ((length = match_ref(offset, eptr, length, md, caseless)) < 0)
2701          {          {
2702            if (length == -2) eptr = md->end_subject;   /* Partial match */
2703          CHECK_PARTIAL();          CHECK_PARTIAL();
2704          RRETURN(MATCH_NOMATCH);          RRETURN(MATCH_NOMATCH);
2705          }          }
# Line 2620  for (;;) Line 2708  for (;;)
2708        }        }
2709    
2710      /* Handle repeated back references. If the length of the reference is      /* Handle repeated back references. If the length of the reference is
2711      zero, just continue with the main loop. */      zero, just continue with the main loop. If the length is negative, it
2712        means the reference is unset in non-Java-compatible mode. If the minimum is
2713        zero, we can continue at the same level without recursion. For any other
2714        minimum, carrying on will result in NOMATCH. */
2715    
2716      if (length == 0) continue;      if (length == 0) continue;
2717        if (length < 0 && min == 0) continue;
2718    
2719      /* First, ensure the minimum number of matches are present. We get back      /* First, ensure the minimum number of matches are present. We get back
2720      the length of the reference string explicitly rather than passing the      the length of the reference string explicitly rather than passing the
# Line 2633  for (;;) Line 2725  for (;;)
2725        int slength;        int slength;
2726        if ((slength = match_ref(offset, eptr, length, md, caseless)) < 0)        if ((slength = match_ref(offset, eptr, length, md, caseless)) < 0)
2727          {          {
2728            if (slength == -2) eptr = md->end_subject;   /* Partial match */
2729          CHECK_PARTIAL();          CHECK_PARTIAL();
2730          RRETURN(MATCH_NOMATCH);          RRETURN(MATCH_NOMATCH);
2731          }          }
# Line 2656  for (;;) Line 2749  for (;;)
2749          if (fi >= max) RRETURN(MATCH_NOMATCH);          if (fi >= max) RRETURN(MATCH_NOMATCH);
2750          if ((slength = match_ref(offset, eptr, length, md, caseless)) < 0)          if ((slength = match_ref(offset, eptr, length, md, caseless)) < 0)
2751            {            {
2752              if (slength == -2) eptr = md->end_subject;   /* Partial match */
2753            CHECK_PARTIAL();            CHECK_PARTIAL();
2754            RRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
2755            }            }
# Line 2674  for (;;) Line 2768  for (;;)
2768          int slength;          int slength;
2769          if ((slength = match_ref(offset, eptr, length, md, caseless)) < 0)          if ((slength = match_ref(offset, eptr, length, md, caseless)) < 0)
2770            {            {
2771            CHECK_PARTIAL();            /* Can't use CHECK_PARTIAL because we don't want to update eptr in
2772              the soft partial matching case. */
2773    
2774              if (slength == -2 && md->partial != 0 &&
2775                  md->end_subject > md->start_used_ptr)
2776                {
2777                md->hitend = TRUE;
2778                if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
2779                }
2780            break;            break;
2781            }            }
2782          eptr += slength;          eptr += slength;
2783          }          }
2784    
2785        while (eptr >= pp)        while (eptr >= pp)
2786          {          {
2787          RMATCH(eptr, ecode, offset_top, md, eptrb, RM15);          RMATCH(eptr, ecode, offset_top, md, eptrb, RM15);
# Line 2703  for (;;) Line 2806  for (;;)
2806      case OP_NCLASS:      case OP_NCLASS:
2807      case OP_CLASS:      case OP_CLASS:
2808        {        {
2809          /* The data variable is saved across frames, so the byte map needs to
2810          be stored there. */
2811    #define BYTE_MAP ((pcre_uint8 *)data)
2812        data = ecode + 1;                /* Save for matching */        data = ecode + 1;                /* Save for matching */
2813        ecode += 33;                     /* Advance past the item */        ecode += 1 + (32 / sizeof(pcre_uchar)); /* Advance past the item */
2814    
2815        switch (*ecode)        switch (*ecode)
2816          {          {
# Line 2725  for (;;) Line 2831  for (;;)
2831          case OP_CRMINRANGE:          case OP_CRMINRANGE:
2832          minimize = (*ecode == OP_CRMINRANGE);          minimize = (*ecode == OP_CRMINRANGE);
2833          min = GET2(ecode, 1);          min = GET2(ecode, 1);
2834          max = GET2(ecode, 3);          max = GET2(ecode, 1 + IMM2_SIZE);
2835          if (max == 0) max = INT_MAX;          if (max == 0) max = INT_MAX;
2836          ecode += 5;          ecode += 1 + 2 * IMM2_SIZE;
2837          break;          break;
2838    
2839          default:               /* No repeat follows */          default:               /* No repeat follows */
# Line 2737  for (;;) Line 2843  for (;;)
2843    
2844        /* First, ensure the minimum number of matches are present. */        /* First, ensure the minimum number of matches are present. */
2845    
2846  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
2847        /* UTF-8 mode */        if (utf)
       if (utf8)  
2848          {          {
2849          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
2850            {            {
# Line 2754  for (;;) Line 2859  for (;;)
2859              if (op == OP_CLASS) RRETURN(MATCH_NOMATCH);              if (op == OP_CLASS) RRETURN(MATCH_NOMATCH);
2860              }              }
2861            else            else
2862              {              if ((BYTE_MAP[c/8] & (1 << (c&7))) == 0) RRETURN(MATCH_NOMATCH);
             if ((data[c/8] & (1 << (c&7))) == 0) RRETURN(MATCH_NOMATCH);  
             }  
2863            }            }
2864          }          }
2865        else        else
2866  #endif  #endif
2867        /* Not UTF-8 mode */        /* Not UTF mode */
2868          {          {
2869          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
2870            {            {
# Line 2771  for (;;) Line 2874  for (;;)
2874              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
2875              }              }
2876            c = *eptr++;            c = *eptr++;
2877            if ((data[c/8] & (1 << (c&7))) == 0) RRETURN(MATCH_NOMATCH);  #ifndef COMPILE_PCRE8
2878              if (c > 255)
2879                {
2880                if (op == OP_CLASS) RRETURN(MATCH_NOMATCH);
2881                }
2882              else
2883    #endif
2884                if ((BYTE_MAP[c/8] & (1 << (c&7))) == 0) RRETURN(MATCH_NOMATCH);
2885            }            }
2886          }          }
2887    
# Line 2785  for (;;) Line 2895  for (;;)
2895    
2896        if (minimize)        if (minimize)
2897          {          {
2898  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
2899          /* UTF-8 mode */          if (utf)
         if (utf8)  
2900            {            {
2901            for (fi = min;; fi++)            for (fi = min;; fi++)
2902              {              {
# Line 2805  for (;;) Line 2914  for (;;)
2914                if (op == OP_CLASS) RRETURN(MATCH_NOMATCH);                if (op == OP_CLASS) RRETURN(MATCH_NOMATCH);
2915                }                }
2916              else              else
2917                {                if ((BYTE_MAP[c/8] & (1 << (c&7))) == 0) RRETURN(MATCH_NOMATCH);
               if ((data[c/8] & (1 << (c&7))) == 0) RRETURN(MATCH_NOMATCH);  
               }  
2918              }              }
2919            }            }
2920          else          else
2921  #endif  #endif
2922          /* Not UTF-8 mode */          /* Not UTF mode */
2923            {            {
2924            for (fi = min;; fi++)            for (fi = min;; fi++)
2925              {              {
# Line 2825  for (;;) Line 2932  for (;;)
2932                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
2933                }                }
2934              c = *eptr++;              c = *eptr++;
2935              if ((data[c/8] & (1 << (c&7))) == 0) RRETURN(MATCH_NOMATCH);  #ifndef COMPILE_PCRE8
2936                if (c > 255)
2937                  {
2938                  if (op == OP_CLASS) RRETURN(MATCH_NOMATCH);
2939                  }
2940                else
2941    #endif
2942                  if ((BYTE_MAP[c/8] & (1 << (c&7))) == 0) RRETURN(MATCH_NOMATCH);
2943              }              }
2944            }            }
2945          /* Control never gets here */          /* Control never gets here */
# Line 2837  for (;;) Line 2951  for (;;)
2951          {          {
2952          pp = eptr;          pp = eptr;
2953    
2954  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
2955          /* UTF-8 mode */          if (utf)
         if (utf8)  
2956            {            {
2957            for (i = min; i < max; i++)            for (i = min; i < max; i++)
2958              {              {
# Line 2855  for (;;) Line 2968  for (;;)
2968                if (op == OP_CLASS) break;                if (op == OP_CLASS) break;
2969                }                }
2970              else              else
2971                {                if ((BYTE_MAP[c/8] & (1 << (c&7))) == 0) break;
               if ((data[c/8] & (1 << (c&7))) == 0) break;  
               }  
2972              eptr += len;              eptr += len;
2973              }              }
2974            for (;;)            for (;;)
# Line 2870  for (;;) Line 2981  for (;;)
2981            }            }
2982          else          else
2983  #endif  #endif
2984            /* Not UTF-8 mode */            /* Not UTF mode */
2985            {            {
2986            for (i = min; i < max; i++)            for (i = min; i < max; i++)
2987              {              {
# Line 2880  for (;;) Line 2991  for (;;)
2991                break;                break;
2992                }                }
2993              c = *eptr;              c = *eptr;
2994              if ((data[c/8] & (1 << (c&7))) == 0) break;  #ifndef COMPILE_PCRE8
2995                if (c > 255)
2996                  {
2997                  if (op == OP_CLASS) break;
2998                  }
2999                else
3000    #endif
3001                  if ((BYTE_MAP[c/8] & (1 << (c&7))) == 0) break;
3002              eptr++;              eptr++;
3003              }              }
3004            while (eptr >= pp)            while (eptr >= pp)
# Line 2893  for (;;) Line 3011  for (;;)
3011    
3012          RRETURN(MATCH_NOMATCH);          RRETURN(MATCH_NOMATCH);
3013          }          }
3014    #undef BYTE_MAP
3015        }        }
3016      /* Control never gets here */      /* Control never gets here */
3017    
# Line 2901  for (;;) Line 3020  for (;;)
3020      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
3021      mode, because Unicode properties are supported in non-UTF-8 mode. */      mode, because Unicode properties are supported in non-UTF-8 mode. */
3022    
3023  #ifdef SUPPORT_UTF8  #if defined SUPPORT_UTF || !defined COMPILE_PCRE8
3024      case OP_XCLASS:      case OP_XCLASS:
3025        {        {
3026        data = ecode + 1 + LINK_SIZE;                /* Save for matching */        data = ecode + 1 + LINK_SIZE;                /* Save for matching */
# Line 2926  for (;;) Line 3045  for (;;)
3045          case OP_CRMINRANGE:          case OP_CRMINRANGE:
3046          minimize = (*ecode == OP_CRMINRANGE);          minimize = (*ecode == OP_CRMINRANGE);
3047          min = GET2(ecode, 1);          min = GET2(ecode, 1);
3048          max = GET2(ecode, 3);          max = GET2(ecode, 1 + IMM2_SIZE);
3049          if (max == 0) max = INT_MAX;          if (max == 0) max = INT_MAX;
3050          ecode += 5;          ecode += 1 + 2 * IMM2_SIZE;
3051          break;          break;
3052    
3053          default:               /* No repeat follows */          default:               /* No repeat follows */
# Line 2946  for (;;) Line 3065  for (;;)
3065            RRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
3066            }            }
3067          GETCHARINCTEST(c, eptr);          GETCHARINCTEST(c, eptr);
3068          if (!_pcre_xclass(c, data)) RRETURN(MATCH_NOMATCH);          if (!PRIV(xclass)(c, data, utf)) RRETURN(MATCH_NOMATCH);
3069          }          }
3070    
3071        /* If max == min we can continue with the main loop without the        /* If max == min we can continue with the main loop without the
# Line 2970  for (;;) Line 3089  for (;;)
3089              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
3090              }              }
3091            GETCHARINCTEST(c, eptr);            GETCHARINCTEST(c, eptr);
3092            if (!_pcre_xclass(c, data)) RRETURN(MATCH_NOMATCH);            if (!PRIV(xclass)(c, data, utf)) RRETURN(MATCH_NOMATCH);
3093            }            }
3094          /* Control never gets here */          /* Control never gets here */
3095          }          }
# Line 2988  for (;;) Line 3107  for (;;)
3107              SCHECK_PARTIAL();              SCHECK_PARTIAL();
3108              break;              break;
3109              }              }
3110    #ifdef SUPPORT_UTF
3111            GETCHARLENTEST(c, eptr, len);            GETCHARLENTEST(c, eptr, len);
3112            if (!_pcre_xclass(c, data)) break;  #else
3113              c = *eptr;
3114    #endif
3115              if (!PRIV(xclass)(c, data, utf)) break;
3116            eptr += len;            eptr += len;
3117            }            }
3118          for(;;)          for(;;)
# Line 2997  for (;;) Line 3120  for (;;)
3120            RMATCH(eptr, ecode, offset_top, md, eptrb, RM21);            RMATCH(eptr, ecode, offset_top, md, eptrb, RM21);
3121            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3122            if (eptr-- == pp) break;        /* Stop if tried at original pos */            if (eptr-- == pp) break;        /* Stop if tried at original pos */
3123            if (utf8) BACKCHAR(eptr);  #ifdef SUPPORT_UTF
3124              if (utf) BACKCHAR(eptr);
3125    #endif
3126            }            }
3127          RRETURN(MATCH_NOMATCH);          RRETURN(MATCH_NOMATCH);
3128          }          }
# Line 3009  for (;;) Line 3134  for (;;)
3134      /* Match a single character, casefully */      /* Match a single character, casefully */
3135    
3136      case OP_CHAR:      case OP_CHAR:
3137  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
3138      if (utf8)      if (utf)
3139        {        {
3140        length = 1;        length = 1;
3141        ecode++;        ecode++;
# Line 3024  for (;;) Line 3149  for (;;)
3149        }        }
3150      else      else
3151  #endif  #endif
3152        /* Not UTF mode */
     /* Non-UTF-8 mode */  
3153        {        {
3154        if (md->end_subject - eptr < 1)        if (md->end_subject - eptr < 1)
3155          {          {
# Line 3037  for (;;) Line 3161  for (;;)
3161        }        }
3162      break;      break;
3163    
3164      /* Match a single character, caselessly */      /* Match a single character, caselessly. If we are at the end of the
3165        subject, give up immediately. */
3166    
3167      case OP_CHARI:      case OP_CHARI:
3168  #ifdef SUPPORT_UTF8      if (eptr >= md->end_subject)
3169      if (utf8)        {
3170          SCHECK_PARTIAL();
3171          RRETURN(MATCH_NOMATCH);
3172          }
3173    
3174    #ifdef SUPPORT_UTF
3175        if (utf)
3176        {        {
3177        length = 1;        length = 1;
3178        ecode++;        ecode++;
3179        GETCHARLEN(fc, ecode, length);        GETCHARLEN(fc, ecode, length);
3180    
       if (length > md->end_subject - eptr)  
         {  
         CHECK_PARTIAL();             /* Not SCHECK_PARTIAL() */  
         RRETURN(MATCH_NOMATCH);  
         }  
   
3181        /* 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
3182        can use the fast lookup table. */        we know that its other case must also be one byte long, so we can use the
3183          fast lookup table. We know that there is at least one byte left in the
3184          subject. */
3185    
3186        if (fc < 128)        if (fc < 128)
3187          {          {
3188          if (md->lcc[*ecode++] != md->lcc[*eptr++]) RRETURN(MATCH_NOMATCH);          if (md->lcc[fc]
3189                != TABLE_GET(*eptr, md->lcc, *eptr)) RRETURN(MATCH_NOMATCH);
3190            ecode++;
3191            eptr++;
3192          }          }
3193    
3194        /* Otherwise we must pick up the subject character */        /* Otherwise we must pick up the subject character. Note that we cannot
3195          use the value of "length" to check for sufficient bytes left, because the
3196          other case of the character may have more or fewer bytes.  */
3197    
3198        else        else
3199          {          {
3200          unsigned int dc;          pcre_uint32 dc;
3201          GETCHARINC(dc, eptr);          GETCHARINC(dc, eptr);
3202          ecode += length;          ecode += length;
3203    
# Line 3082  for (;;) Line 3214  for (;;)
3214          }          }
3215        }        }
3216      else      else
3217  #endif   /* SUPPORT_UTF8 */  #endif   /* SUPPORT_UTF */
3218    
3219      /* Non-UTF-8 mode */      /* Not UTF mode */
3220        {        {
3221        if (md->end_subject - eptr < 1)        if (TABLE_GET(ecode[1], md->lcc, ecode[1])
3222          {            != TABLE_GET(*eptr, md->lcc, *eptr)) RRETURN(MATCH_NOMATCH);
3223          SCHECK_PARTIAL();            /* This one can use SCHECK_PARTIAL() */        eptr++;
         RRETURN(MATCH_NOMATCH);  
         }  
       if (md->lcc[ecode[1]] != md->lcc[*eptr++]) RRETURN(MATCH_NOMATCH);  
3224        ecode += 2;        ecode += 2;
3225        }        }
3226      break;      break;
# Line 3101  for (;;) Line 3230  for (;;)
3230      case OP_EXACT:      case OP_EXACT:
3231      case OP_EXACTI:      case OP_EXACTI:
3232      min = max = GET2(ecode, 1);      min = max = GET2(ecode, 1);
3233      ecode += 3;      ecode += 1 + IMM2_SIZE;
3234      goto REPEATCHAR;      goto REPEATCHAR;
3235    
3236      case OP_POSUPTO:      case OP_POSUPTO:
# Line 3116  for (;;) Line 3245  for (;;)
3245      min = 0;      min = 0;
3246      max = GET2(ecode, 1);      max = GET2(ecode, 1);
3247      minimize = *ecode == OP_MINUPTO || *ecode == OP_MINUPTOI;      minimize = *ecode == OP_MINUPTO || *ecode == OP_MINUPTOI;
3248      ecode += 3;      ecode += 1 + IMM2_SIZE;
3249      goto REPEATCHAR;      goto REPEATCHAR;
3250    
3251      case OP_POSSTAR:      case OP_POSSTAR:
# Line 3164  for (;;) Line 3293  for (;;)
3293      /* Common code for all repeated single-character matches. */      /* Common code for all repeated single-character matches. */
3294    
3295      REPEATCHAR:      REPEATCHAR:
3296  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
3297      if (utf8)      if (utf)
3298        {        {
3299        length = 1;        length = 1;
3300        charptr = ecode;        charptr = ecode;
# Line 3181  for (;;) Line 3310  for (;;)
3310          unsigned int othercase;          unsigned int othercase;
3311          if (op >= OP_STARI &&     /* Caseless */          if (op >= OP_STARI &&     /* Caseless */
3312              (othercase = UCD_OTHERCASE(fc)) != fc)              (othercase = UCD_OTHERCASE(fc)) != fc)
3313            oclength = _pcre_ord2utf8(othercase, occhars);            oclength = PRIV(ord2utf)(othercase, occhars);
3314          else oclength = 0;          else oclength = 0;
3315  #endif  /* SUPPORT_UCP */  #endif  /* SUPPORT_UCP */
3316    
3317          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
3318            {            {
3319            if (eptr <= md->end_subject - length &&            if (eptr <= md->end_subject - length &&
3320              memcmp(eptr, charptr, length) == 0) eptr += length;              memcmp(eptr, charptr, IN_UCHARS(length)) == 0) eptr += length;
3321  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
3322            else if (oclength > 0 &&            else if (oclength > 0 &&
3323                     eptr <= md->end_subject - oclength &&                     eptr <= md->end_subject - oclength &&
3324                     memcmp(eptr, occhars, oclength) == 0) eptr += oclength;                     memcmp(eptr, occhars, IN_UCHARS(oclength)) == 0) eptr += oclength;
3325  #endif  /* SUPPORT_UCP */  #endif  /* SUPPORT_UCP */
3326            else            else
3327              {              {
# Line 3211  for (;;) Line 3340  for (;;)
3340              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3341              if (fi >= max) RRETURN(MATCH_NOMATCH);              if (fi >= max) RRETURN(MATCH_NOMATCH);
3342              if (eptr <= md->end_subject - length &&              if (eptr <= md->end_subject - length &&
3343                memcmp(eptr, charptr, length) == 0) eptr += length;                memcmp(eptr, charptr, IN_UCHARS(length)) == 0) eptr += length;
3344  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
3345              else if (oclength > 0 &&              else if (oclength > 0 &&
3346                       eptr <= md->end_subject - oclength &&                       eptr <= md->end_subject - oclength &&
3347                       memcmp(eptr, occhars, oclength) == 0) eptr += oclength;                       memcmp(eptr, occhars, IN_UCHARS(oclength)) == 0) eptr += oclength;
3348  #endif  /* SUPPORT_UCP */  #endif  /* SUPPORT_UCP */
3349              else              else
3350                {                {
# Line 3232  for (;;) Line 3361  for (;;)
3361            for (i = min; i < max; i++)            for (i = min; i < max; i++)
3362              {              {
3363              if (eptr <= md->end_subject - length &&              if (eptr <= md->end_subject - length &&
3364                  memcmp(eptr, charptr, length) == 0) eptr += length;                  memcmp(eptr, charptr, IN_UCHARS(length)) == 0) eptr += length;
3365  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
3366              else if (oclength > 0 &&              else if (oclength > 0 &&
3367                       eptr <= md->end_subject - oclength &&                       eptr <= md->end_subject - oclength &&
3368                       memcmp(eptr, occhars, oclength) == 0) eptr += oclength;                       memcmp(eptr, occhars, IN_UCHARS(oclength)) == 0) eptr += oclength;
3369  #endif  /* SUPPORT_UCP */  #endif  /* SUPPORT_UCP */
3370              else              else
3371                {                {
# Line 3268  for (;;) Line 3397  for (;;)
3397        value of fc will always be < 128. */        value of fc will always be < 128. */
3398        }        }
3399      else      else
3400  #endif  /* SUPPORT_UTF8 */  #endif  /* SUPPORT_UTF */
3401          /* When not in UTF-8 mode, load a single-byte character. */
3402      /* When not in UTF-8 mode, load a single-byte character. */        fc = *ecode++;
3403    
3404      fc = *ecode++;      /* The value of fc at this point is always one character, though we may
3405        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  
3406      caseful cases, for speed, since matching characters is likely to be quite      caseful cases, for speed, since matching characters is likely to be quite
3407      common. First, ensure the minimum number of matches are present. If min =      common. First, ensure the minimum number of matches are present. If min =
3408      max, continue at the same level without recursing. Otherwise, if      max, continue at the same level without recursing. Otherwise, if
# Line 3284  for (;;) Line 3411  for (;;)
3411      maximizing, find the maximum number of characters and work backwards. */      maximizing, find the maximum number of characters and work backwards. */
3412    
3413      DPRINTF(("matching %c{%d,%d} against subject %.*s\n", fc, min, max,      DPRINTF(("matching %c{%d,%d} against subject %.*s\n", fc, min, max,
3414        max, eptr));        max, (char *)eptr));
3415    
3416      if (op >= OP_STARI)  /* Caseless */      if (op >= OP_STARI)  /* Caseless */
3417        {        {
3418        fc = md->lcc[fc];  #ifdef COMPILE_PCRE8
3419          /* fc must be < 128 if UTF is enabled. */
3420          foc = md->fcc[fc];
3421    #else
3422    #ifdef SUPPORT_UTF
3423    #ifdef SUPPORT_UCP
3424          if (utf && fc > 127)
3425            foc = UCD_OTHERCASE(fc);
3426    #else
3427          if (utf && fc > 127)
3428            foc = fc;
3429    #endif /* SUPPORT_UCP */
3430          else
3431    #endif /* SUPPORT_UTF */
3432            foc = TABLE_GET(fc, md->fcc, fc);
3433    #endif /* COMPILE_PCRE8 */
3434    
3435        for (i = 1; i <= min; i++)        for (i = 1; i <= min; i++)
3436          {          {
3437          if (eptr >= md->end_subject)          if (eptr >= md->end_subject)
# Line 3296  for (;;) Line 3439  for (;;)
3439            SCHECK_PARTIAL();            SCHECK_PARTIAL();
3440            RRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
3441            }            }
3442          if (fc != md->lcc[*eptr++]) RRETURN(MATCH_NOMATCH);          if (fc != *eptr && foc != *eptr) RRETURN(MATCH_NOMATCH);
3443            eptr++;
3444          }          }
3445        if (min == max) continue;        if (min == max) continue;
3446        if (minimize)        if (minimize)
# Line 3311  for (;;) Line 3455  for (;;)
3455              SCHECK_PARTIAL();              SCHECK_PARTIAL();
3456              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
3457              }              }
3458            if (fc != md->lcc[*eptr++]) RRETURN(MATCH_NOMATCH);            if (fc != *eptr && foc != *eptr) RRETURN(MATCH_NOMATCH);
3459              eptr++;
3460            }            }
3461          /* Control never gets here */          /* Control never gets here */
3462          }          }
# Line 3325  for (;;) Line 3470  for (;;)
3470              SCHECK_PARTIAL();              SCHECK_PARTIAL();
3471              break;              break;
3472              }              }
3473            if (fc != md->lcc[*eptr]) break;            if (fc != *eptr && foc != *eptr) break;
3474            eptr++;            eptr++;
3475            }            }
3476    
# Line 3410  for (;;) Line 3555  for (;;)
3555        SCHECK_PARTIAL();        SCHECK_PARTIAL();
3556        RRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
3557        }        }
3558      ecode++;  #ifdef SUPPORT_UTF
3559      GETCHARINCTEST(c, eptr);      if (utf)
     if (op == OP_NOTI)         /* The caseless case */  
3560        {        {
3561  #ifdef SUPPORT_UTF8        register unsigned int ch, och;
3562        if (c < 256)  
3563  #endif        ecode++;
3564        c = md->lcc[c];        GETCHARINC(ch, ecode);
3565        if (md->lcc[*ecode++] == c) RRETURN(MATCH_NOMATCH);        GETCHARINC(c, eptr);
3566    
3567          if (op == OP_NOT)
3568            {
3569            if (ch == c) RRETURN(MATCH_NOMATCH);
3570            }
3571          else
3572            {
3573    #ifdef SUPPORT_UCP
3574            if (ch > 127)
3575              och = UCD_OTHERCASE(ch);
3576    #else
3577            if (ch > 127)
3578              och = ch;
3579    #endif /* SUPPORT_UCP */
3580            else
3581              och = TABLE_GET(ch, md->fcc, ch);
3582            if (ch == c || och == c) RRETURN(MATCH_NOMATCH);
3583            }
3584        }        }
3585      else    /* Caseful */      else
3586    #endif
3587        {        {
3588        if (*ecode++ == c) RRETURN(MATCH_NOMATCH);        register unsigned int ch = ecode[1];
3589          c = *eptr++;
3590          if (ch == c || (op == OP_NOTI && TABLE_GET(ch, md->fcc, ch) == c))
3591            RRETURN(MATCH_NOMATCH);
3592          ecode += 2;
3593        }        }
3594      break;      break;
3595    
# Line 3436  for (;;) Line 3603  for (;;)
3603      case OP_NOTEXACT:      case OP_NOTEXACT:
3604      case OP_NOTEXACTI:      case OP_NOTEXACTI:
3605      min = max = GET2(ecode, 1);      min = max = GET2(ecode, 1);
3606      ecode += 3;      ecode += 1 + IMM2_SIZE;
3607      goto REPEATNOTCHAR;      goto REPEATNOTCHAR;
3608    
3609      case OP_NOTUPTO:      case OP_NOTUPTO:
# Line 3446  for (;;) Line 3613  for (;;)
3613      min = 0;      min = 0;
3614      max = GET2(ecode, 1);      max = GET2(ecode, 1);
3615      minimize = *ecode == OP_NOTMINUPTO || *ecode == OP_NOTMINUPTOI;      minimize = *ecode == OP_NOTMINUPTO || *ecode == OP_NOTMINUPTOI;
3616      ecode += 3;      ecode += 1 + IMM2_SIZE;
3617      goto REPEATNOTCHAR;      goto REPEATNOTCHAR;
3618    
3619      case OP_NOTPOSSTAR:      case OP_NOTPOSSTAR:
# Line 3478  for (;;) Line 3645  for (;;)
3645      possessive = TRUE;      possessive = TRUE;
3646      min = 0;      min = 0;
3647      max = GET2(ecode, 1);      max = GET2(ecode, 1);
3648      ecode += 3;      ecode += 1 + IMM2_SIZE;
3649      goto REPEATNOTCHAR;      goto REPEATNOTCHAR;
3650    
3651      case OP_NOTSTAR:      case OP_NOTSTAR:
# Line 3502  for (;;) Line 3669  for (;;)
3669      /* Common code for all repeated single-byte matches. */      /* Common code for all repeated single-byte matches. */
3670    
3671      REPEATNOTCHAR:      REPEATNOTCHAR:
3672      fc = *ecode++;      GETCHARINCTEST(fc, ecode);
3673    
3674      /* The code is duplicated for the caseless and caseful cases, for speed,      /* The code is duplicated for the caseless and caseful cases, for speed,
3675      since matching characters is likely to be quite common. First, ensure the      since matching characters is likely to be quite common. First, ensure the
# Line 3513  for (;;) Line 3680  for (;;)
3680      characters and work backwards. */      characters and work backwards. */
3681    
3682      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,
3683        max, eptr));        max, (char *)eptr));
3684    
3685      if (op >= OP_NOTSTARI)     /* Caseless */      if (op >= OP_NOTSTARI)     /* Caseless */
3686        {        {
3687        fc = md->lcc[fc];  #ifdef SUPPORT_UTF
3688    #ifdef SUPPORT_UCP
3689          if (utf && fc > 127)
3690            foc = UCD_OTHERCASE(fc);
3691    #else
3692          if (utf && fc > 127)
3693            foc = fc;
3694    #endif /* SUPPORT_UCP */
3695          else
3696    #endif /* SUPPORT_UTF */
3697            foc = TABLE_GET(fc, md->fcc, fc);
3698    
3699  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
3700        /* UTF-8 mode */        if (utf)
       if (utf8)  
3701          {          {
3702          register unsigned int d;          register unsigned int d;
3703          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
# Line 3532  for (;;) Line 3708  for (;;)
3708              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
3709              }              }
3710            GETCHARINC(d, eptr);            GETCHARINC(d, eptr);
3711            if (d < 256) d = md->lcc[d];            if (fc == d || (unsigned int)foc == d) RRETURN(MATCH_NOMATCH);
           if (fc == d) RRETURN(MATCH_NOMATCH);  
3712            }            }
3713          }          }
3714        else        else
3715  #endif  #endif
3716          /* Not UTF mode */
       /* Not UTF-8 mode */  
3717          {          {
3718          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
3719            {            {
# Line 3548  for (;;) Line 3722  for (;;)
3722              SCHECK_PARTIAL();              SCHECK_PARTIAL();
3723              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
3724              }              }
3725            if (fc == md->lcc[*eptr++]) RRETURN(MATCH_NOMATCH);            if (fc == *eptr || foc == *eptr) RRETURN(MATCH_NOMATCH);
3726              eptr++;
3727            }            }
3728          }          }
3729    
# Line 3556  for (;;) Line 3731  for (;;)
3731    
3732        if (minimize)        if (minimize)
3733          {          {
3734  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
3735          /* UTF-8 mode */          if (utf)
         if (utf8)  
3736            {            {
3737            register unsigned int d;            register unsigned int d;
3738            for (fi = min;; fi++)            for (fi = min;; fi++)
# Line 3572  for (;;) Line 3746  for (;;)
3746                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
3747                }                }
3748              GETCHARINC(d, eptr);              GETCHARINC(d, eptr);
3749              if (d < 256) d = md->lcc[d];              if (fc == d || (unsigned int)foc == d) RRETURN(MATCH_NOMATCH);
             if (fc == d) RRETURN(MATCH_NOMATCH);  
3750              }              }
3751            }            }
3752          else          else
3753  #endif  #endif
3754          /* Not UTF-8 mode */          /* Not UTF mode */
3755            {            {
3756            for (fi = min;; fi++)            for (fi = min;; fi++)
3757              {              {
# Line 3590  for (;;) Line 3763  for (;;)
3763                SCHECK_PARTIAL();                SCHECK_PARTIAL();
3764                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
3765                }                }
3766              if (fc == md->lcc[*eptr++]) RRETURN(MATCH_NOMATCH);              if (fc == *eptr || foc == *eptr) RRETURN(MATCH_NOMATCH);
3767                eptr++;
3768              }              }
3769            }            }
3770          /* Control never gets here */          /* Control never gets here */
# Line 3602  for (;;) Line 3776  for (;;)
3776          {          {
3777          pp = eptr;          pp = eptr;
3778    
3779  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
3780          /* UTF-8 mode */          if (utf)
         if (utf8)  
3781            {            {
3782            register unsigned int d;            register unsigned int d;
3783            for (i = min; i < max; i++)            for (i = min; i < max; i++)
# Line 3616  for (;;) Line 3789  for (;;)
3789                break;                break;
3790                }                }
3791              GETCHARLEN(d, eptr, len);              GETCHARLEN(d, eptr, len);
3792              if (d < 256) d = md->lcc[d];              if (fc == d || (unsigned int)foc == d) break;
             if (fc == d) break;  
3793              eptr += len;              eptr += len;
3794              }              }
3795          if (possessive) continue;            if (possessive) continue;
3796          for(;;)            for(;;)
3797              {              {
3798              RMATCH(eptr, ecode, offset_top, md, eptrb, RM30);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM30);
3799              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
# Line 3631  for (;;) Line 3803  for (;;)
3803            }            }
3804          else          else
3805  #endif  #endif
3806          /* Not UTF-8 mode */          /* Not UTF mode */
3807            {            {
3808            for (i = min; i < max; i++)            for (i = min; i < max; i++)
3809              {              {
# Line 3640  for (;;) Line 3812  for (;;)
3812                SCHECK_PARTIAL();                SCHECK_PARTIAL();
3813                break;                break;
3814                }                }
3815              if (fc == md->lcc[*eptr]) break;              if (fc == *eptr || foc == *eptr) break;
3816              eptr++;              eptr++;
3817              }              }
3818            if (possessive) continue;            if (possessive) continue;
# Line 3661  for (;;) Line 3833  for (;;)
3833    
3834      else      else
3835        {        {
3836  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
3837        /* UTF-8 mode */        if (utf)
       if (utf8)  
3838          {          {
3839          register unsigned int d;          register unsigned int d;
3840          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
# Line 3679  for (;;) Line 3850  for (;;)
3850          }          }
3851        else        else
3852  #endif  #endif
3853        /* Not UTF-8 mode */        /* Not UTF mode */
3854          {          {
3855          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
3856            {            {
# Line 3696  for (;;) Line 3867  for (;;)
3867    
3868        if (minimize)        if (minimize)
3869          {          {
3870  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
3871          /* UTF-8 mode */          if (utf)
         if (utf8)  
3872            {            {
3873            register unsigned int d;            register unsigned int d;
3874            for (fi = min;; fi++)            for (fi = min;; fi++)
# Line 3717  for (;;) Line 3887  for (;;)
3887            }            }
3888          else          else
3889  #endif  #endif
3890          /* Not UTF-8 mode */          /* Not UTF mode */
3891            {            {
3892            for (fi = min;; fi++)            for (fi = min;; fi++)
3893              {              {
# Line 3741  for (;;) Line 3911  for (;;)
3911          {          {
3912          pp = eptr;          pp = eptr;
3913    
3914  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
3915          /* UTF-8 mode */          if (utf)
         if (utf8)  
3916            {            {
3917            register unsigned int d;            register unsigned int d;
3918            for (i = min; i < max; i++)            for (i = min; i < max; i++)
# Line 3769  for (;;) Line 3938  for (;;)
3938            }            }
3939          else          else
3940  #endif  #endif
3941          /* Not UTF-8 mode */          /* Not UTF mode */
3942            {            {
3943            for (i = min; i < max; i++)            for (i = min; i < max; i++)
3944              {              {
# Line 3802  for (;;) Line 3971  for (;;)
3971      case OP_TYPEEXACT:      case OP_TYPEEXACT:
3972      min = max = GET2(ecode, 1);      min = max = GET2(ecode, 1);
3973      minimize = TRUE;      minimize = TRUE;
3974      ecode += 3;      ecode += 1 + IMM2_SIZE;
3975      goto REPEATTYPE;      goto REPEATTYPE;
3976    
3977      case OP_TYPEUPTO:      case OP_TYPEUPTO:
# Line 3810  for (;;) Line 3979  for (;;)
3979      min = 0;      min = 0;
3980      max = GET2(ecode, 1);      max = GET2(ecode, 1);
3981      minimize = *ecode == OP_TYPEMINUPTO;      minimize = *ecode == OP_TYPEMINUPTO;
3982      ecode += 3;      ecode += 1 + IMM2_SIZE;
3983      goto REPEATTYPE;      goto REPEATTYPE;
3984    
3985      case OP_TYPEPOSSTAR:      case OP_TYPEPOSSTAR:
# Line 3838  for (;;) Line 4007  for (;;)
4007      possessive = TRUE;      possessive = TRUE;
4008      min = 0;      min = 0;
4009      max = GET2(ecode, 1);      max = GET2(ecode, 1);
4010      ecode += 3;      ecode += 1 + IMM2_SIZE;
4011      goto REPEATTYPE;      goto REPEATTYPE;
4012    
4013      case OP_TYPESTAR:      case OP_TYPESTAR:
# Line 4020  for (;;) Line 4189  for (;;)
4189                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
4190              }              }
4191            break;            break;
4192    
4193              case PT_CLIST:
4194              for (i = 1; i <= min; i++)
4195                {
4196                const pcre_uint32 *cp;
4197                if (eptr >= md->end_subject)
4198                  {
4199                  SCHECK_PARTIAL();
4200                  RRETURN(MATCH_NOMATCH);
4201                  }
4202                GETCHARINCTEST(c, eptr);
4203                cp = PRIV(ucd_caseless_sets) + UCD_CASESET(c);
4204                for (;;)
4205                  {
4206                  if (c < *cp)
4207                    { if (prop_fail_result) break; else { RRETURN(MATCH_NOMATCH); } }
4208                  if (c == *cp++)
4209                    { if (prop_fail_result) { RRETURN(MATCH_NOMATCH); } else break; }
4210                  }
4211                }
4212              break;
4213    
4214            /* This should not occur */            /* This should not occur */
4215    
4216            default:            default:
# Line 4040  for (;;) Line 4230  for (;;)
4230              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4231              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4232              }              }
4233            GETCHARINCTEST(c, eptr);            else
           if (UCD_CATEGORY(c) == ucp_M) RRETURN(MATCH_NOMATCH);  
           while (eptr < md->end_subject)  
4234              {              {
4235              int len = 1;              int lgb, rgb;
4236              if (!utf8) c = *eptr; else { GETCHARLEN(c, eptr, len); }              GETCHARINCTEST(c, eptr);
4237              if (UCD_CATEGORY(c) != ucp_M) break;              lgb = UCD_GRAPHBREAK(c);
4238              eptr += len;             while (eptr < md->end_subject)
4239                  {
4240                  int len = 1;
4241                  if (!utf) c = *eptr; else { GETCHARLEN(c, eptr, len); }
4242                  rgb = UCD_GRAPHBREAK(c);
4243                  if ((PRIV(ucp_gbtable)[lgb] & (1 << rgb)) == 0) break;
4244                  lgb = rgb;
4245                  eptr += len;
4246                  }
4247              }              }
4248              CHECK_PARTIAL();
4249            }            }
4250          }          }
4251    
# Line 4057  for (;;) Line 4254  for (;;)
4254    
4255  /* Handle all other cases when the coding is UTF-8 */  /* Handle all other cases when the coding is UTF-8 */
4256    
4257  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
4258        if (utf8) switch(ctype)        if (utf) switch(ctype)
4259          {          {
4260          case OP_ANY:          case OP_ANY:
4261          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
# Line 4069  for (;;) Line 4266  for (;;)
4266              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4267              }              }
4268            if (IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH);            if (IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH);
4269              if (md->partial != 0 &&
4270                  eptr + 1 >= md->end_subject &&
4271                  NLBLOCK->nltype == NLTYPE_FIXED &&
4272                  NLBLOCK->nllen == 2 &&
4273                  *eptr == NLBLOCK->nl[0])
4274                {
4275                md->hitend = TRUE;
4276                if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
4277                }
4278            eptr++;            eptr++;
4279            while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;            ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);
4280            }            }
4281          break;          break;
4282    
# Line 4083  for (;;) Line 4289  for (;;)
4289              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4290              }              }
4291            eptr++;            eptr++;
4292            while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;            ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);
4293            }            }
4294          break;          break;
4295    
# Line 4105  for (;;) Line 4311  for (;;)
4311              {              {
4312              default: RRETURN(MATCH_NOMATCH);              default: RRETURN(MATCH_NOMATCH);
4313    
4314              case 0x000d:              case CHAR_CR:
4315              if (eptr < md->end_subject && *eptr == 0x0a) eptr++;              if (eptr < md->end_subject && *eptr == CHAR_LF) eptr++;
4316              break;              break;
4317    
4318              case 0x000a:              case CHAR_LF:
4319              break;              break;
4320    
4321              case 0x000b:              case CHAR_VT:
4322              case 0x000c:              case CHAR_FF:
4323              case 0x0085:              case CHAR_NEL:
4324    #ifndef EBCDIC
4325              case 0x2028:              case 0x2028:
4326              case 0x2029:              case 0x2029:
4327    #endif  /* Not EBCDIC */
4328              if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);              if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);
4329              break;              break;
4330              }              }
# Line 4134  for (;;) Line 4342  for (;;)
4342            GETCHARINC(c, eptr);            GETCHARINC(c, eptr);
4343            switch(c)            switch(c)
4344              {              {
4345                HSPACE_CASES: RRETURN(MATCH_NOMATCH);  /* Byte and multibyte cases */
4346              default: break;              default: break;
             case 0x09:      /* HT */  
             case 0x20:      /* SPACE */  
             case 0xa0:      /* NBSP */  
             case 0x1680:    /* OGHAM SPACE MARK */  
             case 0x180e:    /* MONGOLIAN VOWEL SEPARATOR */  
             case 0x2000:    /* EN QUAD */  
             case 0x2001:    /* EM QUAD */  
             case 0x2002:    /* EN SPACE */  
             case 0x2003:    /* EM SPACE */  
             case 0x2004:    /* THREE-PER-EM SPACE */  
             case 0x2005:    /* FOUR-PER-EM SPACE */  
             case 0x2006:    /* SIX-PER-EM SPACE */  
             case 0x2007:    /* FIGURE SPACE */  
             case 0x2008:    /* PUNCTUATION SPACE */  
             case 0x2009:    /* THIN SPACE */  
             case 0x200A:    /* HAIR SPACE */  
             case 0x202f:    /* NARROW NO-BREAK SPACE */  
             case 0x205f:    /* MEDIUM MATHEMATICAL SPACE */  
             case 0x3000:    /* IDEOGRAPHIC SPACE */  
             RRETURN(MATCH_NOMATCH);  
4347              }              }
4348            }            }
4349          break;          break;
# Line 4170  for (;;) Line 4359  for (;;)
4359            GETCHARINC(c, eptr);            GETCHARINC(c, eptr);
4360            switch(c)            switch(c)
4361              {              {
4362                HSPACE_CASES: break;  /* Byte and multibyte cases */
4363              default: RRETURN(MATCH_NOMATCH);              default: RRETURN(MATCH_NOMATCH);
             case 0x09:      /* HT */  
             case 0x20:      /* SPACE */  
             case 0xa0:      /* NBSP */  
             case 0x1680:    /* OGHAM SPACE MARK */  
             case 0x180e:    /* MONGOLIAN VOWEL SEPARATOR */  
             case 0x2000:    /* EN QUAD */  
             case 0x2001:    /* EM QUAD */  
             case 0x2002:    /* EN SPACE */  
             case 0x2003:    /* EM SPACE */  
             case 0x2004:    /* THREE-PER-EM SPACE */  
             case 0x2005:    /* FOUR-PER-EM SPACE */  
             case 0x2006:    /* SIX-PER-EM SPACE */  
             case 0x2007:    /* FIGURE SPACE */  
             case 0x2008:    /* PUNCTUATION SPACE */  
             case 0x2009:    /* THIN SPACE */  
             case 0x200A:    /* HAIR SPACE */  
             case 0x202f:    /* NARROW NO-BREAK SPACE */  
             case 0x205f:    /* MEDIUM MATHEMATICAL SPACE */  
             case 0x3000:    /* IDEOGRAPHIC SPACE */  
             break;  
4364              }              }
4365            }            }
4366          break;          break;
# Line 4206  for (;;) Line 4376  for (;;)
4376            GETCHARINC(c, eptr);            GETCHARINC(c, eptr);
4377            switch(c)            switch(c)
4378              {              {
4379                VSPACE_CASES: RRETURN(MATCH_NOMATCH);
4380              default: break;              default: break;
             case 0x0a:      /* LF */  
             case 0x0b:      /* VT */  
             case 0x0c:      /* FF */  
             case 0x0d:      /* CR */  
             case 0x85:      /* NEL */  
             case 0x2028:    /* LINE SEPARATOR */  
             case 0x2029:    /* PARAGRAPH SEPARATOR */  
             RRETURN(MATCH_NOMATCH);  
4381              }              }
4382            }            }
4383          break;          break;
# Line 4230  for (;;) Line 4393  for (;;)
4393            GETCHARINC(c, eptr);            GETCHARINC(c, eptr);
4394            switch(c)            switch(c)
4395              {              {
4396                VSPACE_CASES: break;
4397              default: RRETURN(MATCH_NOMATCH);              default: RRETURN(MATCH_NOMATCH);
             case 0x0a:      /* LF */  
             case 0x0b:      /* VT */  
             case 0x0c:      /* FF */  
             case 0x0d:      /* CR */  
             case 0x85:      /* NEL */  
             case 0x2028:    /* LINE SEPARATOR */  
             case 0x2029:    /* PARAGRAPH SEPARATOR */  
             break;  
4398              }              }
4399            }            }
4400          break;          break;
# Line 4265  for (;;) Line 4421  for (;;)
4421              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4422              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4423              }              }
4424            if (*eptr >= 128 || (md->ctypes[*eptr++] & ctype_digit) == 0)            if (*eptr >= 128 || (md->ctypes[*eptr] & ctype_digit) == 0)
4425              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4426              eptr++;
4427            /* 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 */
4428            }            }
4429          break;          break;
# Line 4281  for (;;) Line 4438  for (;;)
4438              }              }
4439            if (*eptr < 128 && (md->ctypes[*eptr] & ctype_space) != 0)            if (*eptr < 128 && (md->ctypes[*eptr] & ctype_space) != 0)
4440              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4441            while (++eptr < md->end_subject && (*eptr & 0xc0) == 0x80);            eptr++;
4442              ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);
4443            }            }
4444          break;          break;
4445    
# Line 4293  for (;;) Line 4451  for (;;)
4451              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4452              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4453              }              }
4454            if (*eptr >= 128 || (md->ctypes[*eptr++] & ctype_space) == 0)            if (*eptr >= 128 || (md->ctypes[*eptr] & ctype_space) == 0)
4455              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4456              eptr++;
4457            /* 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 */
4458            }            }
4459          break;          break;
# Line 4309  for (;;) Line 4468  for (;;)
4468              }              }
4469            if (*eptr < 128 && (md->ctypes[*eptr] & ctype_word) != 0)            if (*eptr < 128 && (md->ctypes[*eptr] & ctype_word) != 0)
4470              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4471            while (++eptr < md->end_subject && (*eptr & 0xc0) == 0x80);            eptr++;
4472              ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);
4473            }            }
4474          break;          break;
4475    
# Line 4321  for (;;) Line 4481  for (;;)
4481              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4482              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4483              }              }
4484            if (*eptr >= 128 || (md->ctypes[*eptr++] & ctype_word) == 0)            if (*eptr >= 128 || (md->ctypes[*eptr] & ctype_word) == 0)
4485              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4486              eptr++;
4487            /* 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 */
4488            }            }
4489          break;          break;
# Line 4332  for (;;) Line 4493  for (;;)
4493          }  /* End switch(ctype) */          }  /* End switch(ctype) */
4494    
4495        else        else
4496  #endif     /* SUPPORT_UTF8 */  #endif     /* SUPPORT_UTF */
4497    
4498        /* 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
4499        than OP_PROP and OP_NOTPROP. */        than OP_PROP and OP_NOTPROP. */
# Line 4348  for (;;) Line 4509  for (;;)
4509              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4510              }              }
4511            if (IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH);            if (IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH);
4512              if (md->partial != 0 &&
4513                  eptr + 1 >= md->end_subject &&
4514                  NLBLOCK->nltype == NLTYPE_FIXED &&
4515                  NLBLOCK->nllen == 2 &&
4516                  *eptr == NLBLOCK->nl[0])
4517                {
4518                md->hitend = TRUE;
4519                if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
4520                }
4521            eptr++;            eptr++;
4522            }            }
4523          break;          break;
# Line 4382  for (;;) Line 4552  for (;;)
4552              {              {
4553              default: RRETURN(MATCH_NOMATCH);              default: RRETURN(MATCH_NOMATCH);
4554    
4555              case 0x000d:              case CHAR_CR:
4556              if (eptr < md->end_subject && *eptr == 0x0a) eptr++;              if (eptr < md->end_subject && *eptr == CHAR_LF) eptr++;
4557              break;              break;
4558    
4559              case 0x000a:              case CHAR_LF:
4560              break;              break;
4561    
4562              case 0x000b:              case CHAR_VT:
4563              case 0x000c:              case CHAR_FF:
4564              case 0x0085:              case CHAR_NEL:
4565    #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
4566                case 0x2028:
4567                case 0x2029:
4568    #endif
4569              if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);              if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);
4570              break;              break;
4571              }              }
# Line 4409  for (;;) Line 4583  for (;;)
4583            switch(*eptr++)            switch(*eptr++)
4584              {              {
4585              default: break;              default: break;
4586              case 0x09:      /* HT */              HSPACE_BYTE_CASES:
4587              case 0x20:      /* SPACE */  #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
4588              case 0xa0:      /* NBSP */              HSPACE_MULTIBYTE_CASES:
4589    #endif
4590              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4591              }              }
4592            }            }
# Line 4428  for (;;) Line 4603  for (;;)
4603            switch(*eptr++)            switch(*eptr++)
4604              {              {
4605              default: RRETURN(MATCH_NOMATCH);              default: RRETURN(MATCH_NOMATCH);
4606              case 0x09:      /* HT */              HSPACE_BYTE_CASES:
4607              case 0x20:      /* SPACE */  #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
4608              case 0xa0:      /* NBSP */              HSPACE_MULTIBYTE_CASES:
4609    #endif
4610              break;              break;
4611              }              }
4612            }            }
# Line 4446  for (;;) Line 4622  for (;;)
4622              }              }
4623            switch(*eptr++)            switch(*eptr++)
4624              {              {
4625              default: break;              VSPACE_BYTE_CASES:
4626              case 0x0a:      /* LF */  #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
4627              case 0x0b:      /* VT */              VSPACE_MULTIBYTE_CASES:
4628              case 0x0c:      /* FF */  #endif
             case 0x0d:      /* CR */  
             case 0x85:      /* NEL */  
4629              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4630                default: break;
4631              }              }
4632            }            }
4633          break;          break;
# Line 4468  for (;;) Line 4643  for (;;)
4643            switch(*eptr++)            switch(*eptr++)
4644              {              {
4645              default: RRETURN(MATCH_NOMATCH);              default: RRETURN(MATCH_NOMATCH);
4646              case 0x0a:      /* LF */              VSPACE_BYTE_CASES:
4647              case 0x0b:      /* VT */  #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
4648              case 0x0c:      /* FF */              VSPACE_MULTIBYTE_CASES:
4649              case 0x0d:      /* CR */  #endif
             case 0x85:      /* NEL */  
4650              break;              break;
4651              }              }
4652            }            }
# Line 4486  for (;;) Line 4660  for (;;)
4660              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4661              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4662              }              }
4663            if ((md->ctypes[*eptr++] & ctype_digit) != 0) RRETURN(MATCH_NOMATCH);            if (MAX_255(*eptr) && (md->ctypes[*eptr] & ctype_digit) != 0)
4664                RRETURN(MATCH_NOMATCH);
4665              eptr++;
4666            }            }
4667          break;          break;
4668    
# Line 4498  for (;;) Line 4674  for (;;)
4674              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4675              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4676              }              }
4677            if ((md->ctypes[*eptr++] & ctype_digit) == 0) RRETURN(MATCH_NOMATCH);            if (!MAX_255(*eptr) || (md->ctypes[*eptr] & ctype_digit) == 0)
4678                RRETURN(MATCH_NOMATCH);
4679              eptr++;
4680            }            }
4681          break;          break;
4682    
# Line 4510  for (;;) Line 4688  for (;;)
4688              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4689              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4690              }              }
4691            if ((md->ctypes[*eptr++] & ctype_space) != 0) RRETURN(MATCH_NOMATCH);            if (MAX_255(*eptr) && (md->ctypes[*eptr] & ctype_space) != 0)
4692                RRETURN(MATCH_NOMATCH);
4693              eptr++;
4694            }            }
4695          break;          break;
4696    
# Line 4522  for (;;) Line 4702  for (;;)
4702              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4703              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4704              }              }
4705            if ((md->ctypes[*eptr++] & ctype_space) == 0) RRETURN(MATCH_NOMATCH);            if (!MAX_255(*eptr) || (md->ctypes[*eptr] & ctype_space) == 0)
4706                RRETURN(MATCH_NOMATCH);
4707              eptr++;
4708            }            }
4709          break;          break;
4710    
# Line 4534  for (;;) Line 4716  for (;;)
4716              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4717              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4718              }              }
4719            if ((md->ctypes[*eptr++] & ctype_word) != 0)            if (MAX_255(*eptr) && (md->ctypes[*eptr] & ctype_word) != 0)
4720              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4721              eptr++;
4722            }            }
4723          break;          break;
4724    
# Line 4547  for (;;) Line 4730  for (;;)
4730              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4731              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4732              }              }
4733            if ((md->ctypes[*eptr++] & ctype_word) == 0)            if (!MAX_255(*eptr) || (md->ctypes[*eptr] & ctype_word) == 0)
4734              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4735              eptr++;
4736            }            }
4737          break;          break;
4738    
# Line 4739  for (;;) Line 4923  for (;;)
4923              }              }
4924            /* Control never gets here */            /* Control never gets here */
4925    
4926            /* This should never occur */            case PT_CLIST:
4927              for (fi = min;; fi++)
4928                {
4929                const pcre_uint32 *cp;
4930                RMATCH(eptr, ecode, offset_top, md, eptrb, RM62);
4931                if (rrc != MATCH_NOMATCH) RRETURN(rrc);
4932                if (fi >= max) RRETURN(MATCH_NOMATCH);
4933                if (eptr >= md->end_subject)
4934                  {
4935                  SCHECK_PARTIAL();
4936                  RRETURN(MATCH_NOMATCH);
4937                  }
4938                GETCHARINCTEST(c, eptr);
4939                cp = PRIV(ucd_caseless_sets) + UCD_CASESET(c);
4940                for (;;)
4941                  {
4942                  if (c < *cp)
4943                    { if (prop_fail_result) break; else { RRETURN(MATCH_NOMATCH); } }
4944                  if (c == *cp++)
4945                    { if (prop_fail_result) { RRETURN(MATCH_NOMATCH); } else break; }
4946                  }
4947                }
4948              /* Control never gets here */
4949    
4950              /* This should never occur */
4951            default:            default:
4952            RRETURN(PCRE_ERROR_INTERNAL);            RRETURN(PCRE_ERROR_INTERNAL);
4953            }            }
# Line 4761  for (;;) Line 4968  for (;;)
4968              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4969              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4970              }              }
4971            GETCHARINCTEST(c, eptr);            else
           if (UCD_CATEGORY(c) == ucp_M) RRETURN(MATCH_NOMATCH);  
           while (eptr < md->end_subject)  
4972              {              {
4973              int len = 1;              int lgb, rgb;
4974              if (!utf8) c = *eptr; else { GETCHARLEN(c, eptr, len); }              GETCHARINCTEST(c, eptr);
4975              if (UCD_CATEGORY(c) != ucp_M) break;              lgb = UCD_GRAPHBREAK(c);
4976              eptr += len;              while (eptr < md->end_subject)
4977                  {
4978                  int len = 1;
4979                  if (!utf) c = *eptr; else { GETCHARLEN(c, eptr, len); }
4980                  rgb = UCD_GRAPHBREAK(c);
4981                  if ((PRIV(ucp_gbtable)[lgb] & (1 << rgb)) == 0) break;
4982                  lgb = rgb;
4983                  eptr += len;
4984                  }
4985              }              }
4986              CHECK_PARTIAL();
4987            }            }
4988          }          }
4989        else        else
4990  #endif     /* SUPPORT_UCP */  #endif     /* SUPPORT_UCP */
4991    
4992  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
4993        /* UTF-8 mode */        if (utf)
       if (utf8)  
4994          {          {
4995          for (fi = min;; fi++)          for (fi = min;; fi++)
4996            {            {
# Line 4794  for (;;) Line 5007  for (;;)
5007            GETCHARINC(c, eptr);            GETCHARINC(c, eptr);
5008            switch(ctype)            switch(ctype)
5009              {              {
5010              case OP_ANY:        /* This is the non-NL case */              case OP_ANY:               /* This is the non-NL case */
5011                if (md->partial != 0 &&    /* Take care with CRLF partial */
5012                    eptr >= md->end_subject &&
5013                    NLBLOCK->nltype == NLTYPE_FIXED &&
5014                    NLBLOCK->nllen == 2 &&
5015                    c == NLBLOCK->nl[0])
5016                  {
5017                  md->hitend = TRUE;
5018                  if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
5019                  }
5020                break;
5021    
5022              case OP_ALLANY:              case OP_ALLANY:
5023              case OP_ANYBYTE:              case OP_ANYBYTE:
5024              break;              break;
# Line 4803  for (;;) Line 5027  for (;;)
5027              switch(c)              switch(c)
5028                {                {
5029                default: RRETURN(MATCH_NOMATCH);                default: RRETURN(MATCH_NOMATCH);
5030                case 0x000d:                case CHAR_CR:
5031                if (eptr < md->end_subject && *eptr == 0x0a) eptr++;                if (eptr < md->end_subject && *eptr == CHAR_LF) eptr++;
5032                break;                break;
5033                case 0x000a:  
5034                  case CHAR_LF:
5035                break;                break;
5036    
5037                case 0x000b:                case CHAR_VT:
5038                case 0x000c:                case CHAR_FF:
5039                case 0x0085:                case CHAR_NEL:
5040    #ifndef EBCDIC
5041                case 0x2028:                case 0x2028:
5042                case 0x2029:                case 0x2029:
5043    #endif  /* Not EBCDIC */
5044                if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);                if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);
5045                break;                break;
5046                }                }
# Line 4822  for (;;) Line 5049  for (;;)
5049              case OP_NOT_HSPACE:              case OP_NOT_HSPACE:
5050              switch(c)              switch(c)
5051                {                {
5052                  HSPACE_CASES: RRETURN(MATCH_NOMATCH);
5053                default: break;                default: break;
               case 0x09:      /* HT */  
               case 0x20:      /* SPACE */  
               case 0xa0:      /* NBSP */  
               case 0x1680:    /* OGHAM SPACE MARK */  
               case 0x180e:    /* MONGOLIAN VOWEL SEPARATOR */  
               case 0x2000:    /* EN QUAD */  
               case 0x2001:    /* EM QUAD */  
               case 0x2002:    /* EN SPACE */  
               case 0x2003:    /* EM SPACE */  
               case 0x2004:    /* THREE-PER-EM SPACE */  
               case 0x2005:    /* FOUR-PER-EM SPACE */  
               case 0x2006:    /* SIX-PER-EM SPACE */  
               case 0x2007:    /* FIGURE SPACE */  
               case 0x2008:    /* PUNCTUATION SPACE */  
               case 0x2009:    /* THIN SPACE */  
               case 0x200A:    /* HAIR SPACE */  
               case 0x202f:    /* NARROW NO-BREAK SPACE */  
               case 0x205f:    /* MEDIUM MATHEMATICAL SPACE */  
               case 0x3000:    /* IDEOGRAPHIC SPACE */  
               RRETURN(MATCH_NOMATCH);  
5054                }                }
5055              break;              break;
5056    
5057              case OP_HSPACE:              case OP_HSPACE:
5058              switch(c)              switch(c)
5059                {                {
5060                  HSPACE_CASES: break;
5061                default: RRETURN(MATCH_NOMATCH);                default: RRETURN(MATCH_NOMATCH);
               case 0x09:      /* HT */  
               case 0x20:      /* SPACE */  
               case 0xa0:      /* NBSP */  
               case 0x1680:    /* OGHAM SPACE MARK */  
               case 0x180e:    /* MONGOLIAN VOWEL SEPARATOR */  
               case 0x2000:    /* EN QUAD */  
               case 0x2001:    /* EM QUAD */  
               case 0x2002:    /* EN SPACE */  
               case 0x2003:    /* EM SPACE */  
               case 0x2004:    /* THREE-PER-EM SPACE */  
               case 0x2005:    /* FOUR-PER-EM SPACE */  
               case 0x2006:    /* SIX-PER-EM SPACE */  
               case 0x2007:    /* FIGURE SPACE */  
               case 0x2008:    /* PUNCTUATION SPACE */  
               case 0x2009:    /* THIN SPACE */  
               case 0x200A:    /* HAIR SPACE */  
               case 0x202f:    /* NARROW NO-BREAK SPACE */  
               case 0x205f:    /* MEDIUM MATHEMATICAL SPACE */  
               case 0x3000:    /* IDEOGRAPHIC SPACE */  
               break;  
5062                }                }
5063              break;              break;
5064    
5065              case OP_NOT_VSPACE:              case OP_NOT_VSPACE:
5066              switch(c)              switch(c)
5067                {                {
5068                  VSPACE_CASES: RRETURN(MATCH_NOMATCH);
5069                default: break;                default: break;
               case 0x0a:      /* LF */  
               case 0x0b:      /* VT */  
               case 0x0c:      /* FF */  
               case 0x0d:      /* CR */  
               case 0x85:      /* NEL */  
               case 0x2028:    /* LINE SEPARATOR */  
               case 0x2029:    /* PARAGRAPH SEPARATOR */  
               RRETURN(MATCH_NOMATCH);  
5070                }                }
5071              break;              break;
5072    
5073              case OP_VSPACE:              case OP_VSPACE:
5074              switch(c)              switch(c)
5075                {                {
5076                  VSPACE_CASES: break;
5077                default: RRETURN(MATCH_NOMATCH);                default: RRETURN(MATCH_NOMATCH);
               case 0x0a:      /* LF */  
               case 0x0b:      /* VT */  
               case 0x0c:      /* FF */  
               case 0x0d:      /* CR */  
               case 0x85:      /* NEL */  
               case 0x2028:    /* LINE SEPARATOR */  
               case 0x2029:    /* PARAGRAPH SEPARATOR */  
               break;  
5078                }                }
5079              break;              break;
5080    
# Line 4919  for (;;) Line 5094  for (;;)
5094              break;              break;
5095    
5096              case OP_WHITESPACE:              case OP_WHITESPACE:
5097              if  (c >= 256 || (md->ctypes[c] & ctype_space) == 0)              if (c >= 256 || (md->ctypes[c] & ctype_space) == 0)
5098                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
5099              break;              break;
5100    
# Line 4940  for (;;) Line 5115  for (;;)
5115          }          }
5116        else        else
5117  #endif  #endif
5118        /* Not UTF-8 mode */        /* Not UTF mode */
5119          {          {
5120          for (fi = min;; fi++)          for (fi = min;; fi++)
5121            {            {
# Line 4957  for (;;) Line 5132  for (;;)
5132            c = *eptr++;            c = *eptr++;
5133            switch(ctype)            switch(ctype)
5134              {              {
5135              case OP_ANY:     /* This is the non-NL case */              case OP_ANY:               /* This is the non-NL case */
5136                if (md->partial != 0 &&    /* Take care with CRLF partial */
5137                    eptr >= md->end_subject &&
5138                    NLBLOCK->nltype == NLTYPE_FIXED &&
5139                    NLBLOCK->nllen == 2 &&
5140                    c == NLBLOCK->nl[0])
5141                  {
5142                  md->hitend = TRUE;
5143                  if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
5144                  }
5145                break;
5146    
5147              case OP_ALLANY:              case OP_ALLANY:
5148              case OP_ANYBYTE:              case OP_ANYBYTE:
5149              break;              break;
# Line 4966  for (;;) Line 5152  for (;;)
5152              switch(c)              switch(c)
5153                {                {
5154                default: RRETURN(MATCH_NOMATCH);                default: RRETURN(MATCH_NOMATCH);
5155                case 0x000d:                case CHAR_CR:
5156                if (eptr < md->end_subject && *eptr == 0x0a) eptr++;                if (eptr < md->end_subject && *eptr == CHAR_LF) eptr++;
5157                break;                break;
5158    
5159                case 0x000a:                case CHAR_LF:
5160                break;                break;
5161    
5162                case 0x000b:                case CHAR_VT:
5163                case 0x000c:                case CHAR_FF:
5164                case 0x0085:                case CHAR_NEL:
5165    #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
5166                  case 0x2028:
5167                  case 0x2029:
5168    #endif
5169                if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);                if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);
5170                break;                break;
5171                }                }
# Line 4985  for (;;) Line 5175  for (;;)
5175              switch(c)              switch(c)
5176                {                {
5177                default: break;                default: break;
5178                case 0x09:      /* HT */                HSPACE_BYTE_CASES:
5179                case 0x20:      /* SPACE */  #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
5180                case 0xa0:      /* NBSP */                HSPACE_MULTIBYTE_CASES:
5181    #endif
5182                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
5183                }                }
5184              break;              break;
# Line 4996  for (;;) Line 5187  for (;;)
5187              switch(c)              switch(c)
5188                {                {
5189                default: RRETURN(MATCH_NOMATCH);                default: RRETURN(MATCH_NOMATCH);
5190                case 0x09:      /* HT */                HSPACE_BYTE_CASES:
5191                case 0x20:      /* SPACE */  #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
5192                case 0xa0:      /* NBSP */                HSPACE_MULTIBYTE_CASES:
5193    #endif
5194                break;                break;
5195                }                }
5196              break;              break;
# Line 5007  for (;;) Line 5199  for (;;)
5199              switch(c)              switch(c)
5200                {                {
5201                default: break;                default: break;
5202                case 0x0a:      /* LF */                VSPACE_BYTE_CASES:
5203                case 0x0b:      /* VT */  #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
5204                case 0x0c:      /* FF */                VSPACE_MULTIBYTE_CASES:
5205                case 0x0d:      /* CR */  #endif
               case 0x85:      /* NEL */  
5206                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
5207                }                }
5208              break;              break;
# Line 5020  for (;;) Line 5211  for (;;)
5211              switch(c)              switch(c)
5212                {                {
5213                default: RRETURN(MATCH_NOMATCH);                default: RRETURN(MATCH_NOMATCH);
5214                case 0x0a:      /* LF */                VSPACE_BYTE_CASES:
5215                case 0x0b:      /* VT */  #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
5216                case 0x0c:      /* FF */                VSPACE_MULTIBYTE_CASES:
5217                case 0x0d:      /* CR */  #endif
               case 0x85:      /* NEL */  
5218                break;                break;
5219                }                }
5220              break;              break;
5221    
5222              case OP_NOT_DIGIT:              case OP_NOT_DIGIT:
5223              if ((md->ctypes[c] & ctype_digit) != 0) RRETURN(MATCH_NOMATCH);              if (MAX_255(c) && (md->ctypes[c] & ctype_digit) != 0) RRETURN(MATCH_NOMATCH);
5224              break;              break;
5225    
5226              case OP_DIGIT:              case OP_DIGIT:
5227              if ((md->ctypes[c] & ctype_digit) == 0) RRETURN(MATCH_NOMATCH);              if (!MAX_255(c) || (md->ctypes[c] & ctype_digit) == 0) RRETURN(MATCH_NOMATCH);
5228              break;              break;
5229    
5230              case OP_NOT_WHITESPACE:              case OP_NOT_WHITESPACE:
5231              if ((md->ctypes[c] & ctype_space) != 0) RRETURN(MATCH_NOMATCH);              if (MAX_255(c) && (md->ctypes[c] & ctype_space) != 0) RRETURN(MATCH_NOMATCH);
5232              break;              break;
5233    
5234              case OP_WHITESPACE:              case OP_WHITESPACE:
5235              if  ((md->ctypes[c] & ctype_space) == 0) RRETURN(MATCH_NOMATCH);              if (!MAX_255(c) || (md->ctypes[c] & ctype_space) == 0) RRETURN(MATCH_NOMATCH);
5236              break;              break;
5237    
5238              case OP_NOT_WORDCHAR:              case OP_NOT_WORDCHAR:
5239              if ((md->ctypes[c] & ctype_word) != 0) RRETURN(MATCH_NOMATCH);              if (MAX_255(c) && (md->ctypes[c] & ctype_word) != 0) RRETURN(MATCH_NOMATCH);
5240              break;              break;
5241    
5242              case OP_WORDCHAR:              case OP_WORDCHAR:
5243              if ((md->ctypes[c] & ctype_word) == 0) RRETURN(MATCH_NOMATCH);              if (!MAX_255(c) || (md->ctypes[c] & ctype_word) == 0) RRETURN(MATCH_NOMATCH);
5244              break;              break;
5245    
5246              default:              default:
# Line 5226  for (;;) Line 5416  for (;;)
5416              eptr+= len;              eptr+= len;
5417              }              }
5418            break;            break;
5419    
5420              case PT_CLIST:
5421              for (i = min; i < max; i++)
5422                {
5423                const pcre_uint32 *cp;
5424                int len = 1;
5425                if (eptr >= md->end_subject)
5426                  {
5427                  SCHECK_PARTIAL();
5428                  break;
5429                  }
5430                GETCHARLENTEST(c, eptr, len);
5431                cp = PRIV(ucd_caseless_sets) + UCD_CASESET(c);
5432                for (;;)
5433                  {
5434                  if (c < *cp)
5435                    { if (prop_fail_result) break; else goto GOT_MAX; }
5436                  if (c == *cp++)
5437                    { if (prop_fail_result) goto GOT_MAX; else break; }
5438                  }
5439                eptr += len;
5440                }
5441              GOT_MAX:
5442              break;
5443    
5444            default:            default:
5445            RRETURN(PCRE_ERROR_INTERNAL);            RRETURN(PCRE_ERROR_INTERNAL);
# Line 5239  for (;;) Line 5453  for (;;)
5453            RMATCH(eptr, ecode, offset_top, md, eptrb, RM44);            RMATCH(eptr, ecode, offset_top, md, eptrb, RM44);
5454            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
5455            if (eptr-- == pp) break;        /* Stop if tried at original pos */            if (eptr-- == pp) break;        /* Stop if tried at original pos */
5456            if (utf8) BACKCHAR(eptr);            if (utf) BACKCHAR(eptr);
5457            }            }
5458          }          }
5459    
# Line 5250  for (;;) Line 5464  for (;;)
5464          {          {
5465          for (i = min; i < max; i++)          for (i = min; i < max; i++)
5466            {            {
           int len = 1;  
5467            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
5468              {              {
5469              SCHECK_PARTIAL();              SCHECK_PARTIAL();
5470              break;              break;
5471              }              }
5472            if (!utf8) c = *eptr; else { GETCHARLEN(c, eptr, len); }            else
           if (UCD_CATEGORY(c) == ucp_M) break;  
           eptr += len;  
           while (eptr < md->end_subject)  
5473              {              {
5474              len = 1;              int lgb, rgb;
5475              if (!utf8) c = *eptr; else { GETCHARLEN(c, eptr, len); }              GETCHARINCTEST(c, eptr);
5476              if (UCD_CATEGORY(c) != ucp_M) break;              lgb = UCD_GRAPHBREAK(c);
5477              eptr += len;              while (eptr < md->end_subject)
5478                  {
5479                  int len = 1;
5480                  if (!utf) c = *eptr; else { GETCHARLEN(c, eptr, len); }
5481                  rgb = UCD_GRAPHBREAK(c);
5482                  if ((PRIV(ucp_gbtable)[lgb] & (1 << rgb)) == 0) break;
5483                  lgb = rgb;
5484                  eptr += len;
5485                  }
5486              }              }
5487              CHECK_PARTIAL();
5488            }            }
5489    
5490          /* eptr is now past the end of the maximum run */          /* eptr is now past the end of the maximum run */
# Line 5279  for (;;) Line 5498  for (;;)
5498            if (eptr-- == pp) break;        /* Stop if tried at original pos */            if (eptr-- == pp) break;        /* Stop if tried at original pos */
5499            for (;;)                        /* Move back over one extended */            for (;;)                        /* Move back over one extended */
5500              {              {
5501              if (!utf8) c = *eptr; else              if (!utf) c = *eptr; else
5502                {                {
5503                BACKCHAR(eptr);                BACKCHAR(eptr);
5504                GETCHAR(c, eptr);                GETCHAR(c, eptr);
# Line 5293  for (;;) Line 5512  for (;;)
5512        else        else
5513  #endif   /* SUPPORT_UCP */  #endif   /* SUPPORT_UCP */
5514    
5515  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
5516        /* UTF-8 mode */        if (utf)
   
       if (utf8)  
5517          {          {
5518          switch(ctype)          switch(ctype)
5519            {            {
# Line 5311  for (;;) Line 5528  for (;;)
5528                  break;                  break;
5529                  }                  }
5530                if (IS_NEWLINE(eptr)) break;                if (IS_NEWLINE(eptr)) break;
5531                  if (md->partial != 0 &&    /* Take care with CRLF partial */
5532                      eptr + 1 >= md->end_subject &&
5533                      NLBLOCK->nltype == NLTYPE_FIXED &&
5534                      NLBLOCK->nllen == 2 &&
5535                      *eptr == NLBLOCK->nl[0])
5536                    {
5537                    md->hitend = TRUE;
5538                    if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
5539                    }
5540                eptr++;                eptr++;
5541                while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;                ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);
5542                }                }
5543              }              }
5544    
# Line 5328  for (;;) Line 5554  for (;;)
5554                  break;                  break;
5555                  }                  }
5556                if (IS_NEWLINE(eptr)) break;                if (IS_NEWLINE(eptr)) break;
5557                  if (md->partial != 0 &&    /* Take care with CRLF partial */
5558                      eptr + 1 >= md->end_subject &&
5559                      NLBLOCK->nltype == NLTYPE_FIXED &&
5560                      NLBLOCK->nllen == 2 &&
5561                      *eptr == NLBLOCK->nl[0])
5562                    {
5563                    md->hitend = TRUE;
5564                    if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
5565                    }
5566                eptr++;                eptr++;
5567                while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;                ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);
5568                }                }
5569              }              }
5570            break;            break;
# Line 5345  for (;;) Line 5580  for (;;)
5580                  break;                  break;
5581                  }                  }
5582                eptr++;                eptr++;
5583                while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;                ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);
5584                }                }
5585              }              }
5586            else            else
# Line 5377  for (;;) Line 5612  for (;;)
5612                break;                break;
5613                }                }
5614              GETCHARLEN(c, eptr, len);              GETCHARLEN(c, eptr, len);
5615              if (c == 0x000d)              if (c == CHAR_CR)
5616                {                {
5617                if (++eptr >= md->end_subject) break;                if (++eptr >= md->end_subject) break;
5618                if (*eptr == 0x000a) eptr++;                if (*eptr == CHAR_LF) eptr++;
5619                }                }
5620              else              else
5621                {                {
5622                if (c != 0x000a &&                if (c != CHAR_LF &&
5623                    (md->bsr_anycrlf ||                    (md->bsr_anycrlf ||
5624                     (c != 0x000b && c != 0x000c &&                     (c != CHAR_VT && c != CHAR_FF && c != CHAR_NEL
5625                      c != 0x0085 && c != 0x2028 && c != 0x2029)))  #ifndef EBCDIC
5626                        && c != 0x2028 && c != 0x2029
5627    #endif  /* Not EBCDIC */
5628                        )))
5629                  break;                  break;
5630                eptr += len;                eptr += len;
5631                }                }
# Line 5408  for (;;) Line 5646  for (;;)
5646              GETCHARLEN(c, eptr, len);              GETCHARLEN(c, eptr, len);
5647              switch(c)              switch(c)
5648                {                {
5649                  HSPACE_CASES: gotspace = TRUE; break;
5650                default: gotspace = FALSE; break;                default: gotspace = FALSE; break;
               case 0x09:      /* HT */  
               case 0x20:      /* SPACE */  
               case 0xa0:      /* NBSP */  
               case 0x1680:    /* OGHAM SPACE MARK */  
               case 0x180e:    /* MONGOLIAN VOWEL SEPARATOR */  
               case 0x2000:    /* EN QUAD */  
               case 0x2001:    /* EM QUAD */  
               case 0x2002:    /* EN SPACE */  
               case 0x2003:    /* EM SPACE */  
               case 0x2004:    /* THREE-PER-EM SPACE */  
               case 0x2005:    /* FOUR-PER-EM SPACE */  
               case 0x2006:    /* SIX-PER-EM SPACE */  
               case 0x2007:    /* FIGURE SPACE */  
               case 0x2008:    /* PUNCTUATION SPACE */  
               case 0x2009:    /* THIN SPACE */  
               case 0x200A:    /* HAIR SPACE */  
               case 0x202f:    /* NARROW NO-BREAK SPACE */  
               case 0x205f:    /* MEDIUM MATHEMATICAL SPACE */  
               case 0x3000:    /* IDEOGRAPHIC SPACE */  
               gotspace = TRUE;  
               break;  
5651                }                }
5652              if (gotspace == (ctype == OP_NOT_HSPACE)) break;              if (gotspace == (ctype == OP_NOT_HSPACE)) break;
5653              eptr += len;              eptr += len;
# Line 5450  for (;;) Line 5668  for (;;)
5668              GETCHARLEN(c, eptr, len);              GETCHARLEN(c, eptr, len);
5669              switch(c)              switch(c)
5670                {                {
5671                  VSPACE_CASES: gotspace = TRUE; break;
5672                default: gotspace = FALSE; break;                default: gotspace = FALSE; break;
               case 0x0a:      /* LF */  
               case 0x0b:      /* VT */  
               case 0x0c:      /* FF */  
               case 0x0d:      /* CR */  
               case 0x85:      /* NEL */  
               case 0x2028:    /* LINE SEPARATOR */  
               case 0x2029:    /* PARAGRAPH SEPARATOR */  
               gotspace = TRUE;  
               break;  
5673                }                }
5674              if (gotspace == (ctype == OP_NOT_VSPACE)) break;              if (gotspace == (ctype == OP_NOT_VSPACE)) break;
5675              eptr += len;              eptr += len;
# Line 5573  for (;;) Line 5783  for (;;)
5783            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
5784            if (eptr-- == pp) break;        /* Stop if tried at original pos */            if (eptr-- == pp) break;        /* Stop if tried at original pos */
5785            BACKCHAR(eptr);            BACKCHAR(eptr);
5786            if (ctype == OP_ANYNL && eptr > pp  && *eptr == '\n' &&            if (ctype == OP_ANYNL && eptr > pp  && *eptr == CHAR_NL &&
5787                eptr[-1] == '\r') eptr--;                eptr[-1] == CHAR_CR) eptr--;
5788            }            }
5789          }          }
5790        else        else
5791  #endif  /* SUPPORT_UTF8 */  #endif  /* SUPPORT_UTF */
5792          /* Not UTF mode */
       /* Not UTF-8 mode */  
5793          {          {
5794          switch(ctype)          switch(ctype)
5795            {            {
# Line 5593  for (;;) Line 5802  for (;;)
5802                break;                break;
5803                }                }
5804              if (IS_NEWLINE(eptr)) break;              if (IS_NEWLINE(eptr)) break;
5805                if (md->partial != 0 &&    /* Take care with CRLF partial */
5806                    eptr + 1 >= md->end_subject &&
5807                    NLBLOCK->nltype == NLTYPE_FIXED &&
5808                    NLBLOCK->nllen == 2 &&
5809                    *eptr == NLBLOCK->nl[0])
5810                  {
5811                  md->hitend = TRUE;
5812                  if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
5813                  }
5814              eptr++;              eptr++;
5815              }              }
5816            break;            break;
# Line 5617  for (;;) Line 5835  for (;;)
5835                break;                break;
5836                }                }
5837              c = *eptr;              c = *eptr;
5838              if (c == 0x000d)              if (c == CHAR_CR)
5839                {                {
5840                if (++eptr >= md->end_subject) break;                if (++eptr >= md->end_subject) break;
5841                if (*eptr == 0x000a) eptr++;                if (*eptr == CHAR_LF) eptr++;
5842                }                }
5843              else              else
5844                {                {
5845                if (c != 0x000a &&                if (c != CHAR_LF && (md->bsr_anycrlf ||
5846                    (md->bsr_anycrlf ||                   (c != CHAR_VT && c != CHAR_FF && c != CHAR_NEL
5847                      (c != 0x000b && c != 0x000c && c != 0x0085)))  #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
5848                  break;                   && c != 0x2028 && c != 0x2029
5849    #endif
5850                     ))) break;
5851                eptr++;                eptr++;
5852                }                }
5853              }              }
# Line 5641  for (;;) Line 5861  for (;;)
5861                SCHECK_PARTIAL();                SCHECK_PARTIAL();
5862                break;                break;
5863                }                }
5864              c = *eptr;              switch(*eptr)
5865              if (c == 0x09 || c == 0x20 || c == 0xa0) break;                {
5866              eptr++;                default: eptr++; break;
5867                  HSPACE_BYTE_CASES:
5868    #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
5869                  HSPACE_MULTIBYTE_CASES:
5870    #endif
5871                  goto ENDLOOP00;
5872                  }
5873              }              }
5874              ENDLOOP00:
5875            break;            break;
5876    
5877            case OP_HSPACE:            case OP_HSPACE:
# Line 5655  for (;;) Line 5882  for (;;)
5882                SCHECK_PARTIAL();                SCHECK_PARTIAL();
5883                break;                break;
5884                }                }
5885              c = *eptr;              switch(*eptr)
5886              if (c != 0x09 && c != 0x20 && c != 0xa0) break;                {
5887              eptr++;                default: goto ENDLOOP01;
5888                  HSPACE_BYTE_CASES:
5889    #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
5890                  HSPACE_MULTIBYTE_CASES:
5891    #endif
5892                  eptr++; break;
5893                  }
5894              }              }
5895              ENDLOOP01:
5896            break;            break;
5897    
5898            case OP_NOT_VSPACE:            case OP_NOT_VSPACE:
# Line 5669  for (;;) Line 5903  for (;;)
5903                SCHECK_PARTIAL();                SCHECK_PARTIAL();
5904                break;                break;
5905                }                }
5906              c = *eptr;              switch(*eptr)
5907              if (c == 0x0a || c == 0x0b || c == 0x0c || c == 0x0d || c == 0x85)                {
5908                break;                default: eptr++; break;
5909              eptr++;                VSPACE_BYTE_CASES:
5910    #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
5911                  VSPACE_MULTIBYTE_CASES:
5912    #endif
5913                  goto ENDLOOP02;
5914                  }
5915              }              }
5916              ENDLOOP02:
5917            break;            break;
5918    
5919            case OP_VSPACE:            case OP_VSPACE:
# Line 5684  for (;;) Line 5924  for (;;)
5924                SCHECK_PARTIAL();                SCHECK_PARTIAL();
5925                break;                break;
5926                }                }
5927              c = *eptr;              switch(*eptr)
5928              if (c != 0x0a && c != 0x0b && c != 0x0c && c != 0x0d && c != 0x85)                {
5929                break;                default: goto ENDLOOP03;
5930              eptr++;                VSPACE_BYTE_CASES:
5931    #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
5932                  VSPACE_MULTIBYTE_CASES:
5933    #endif
5934                  eptr++; break;
5935                  }
5936              }              }
5937              ENDLOOP03:
5938            break;            break;
5939    
5940            case OP_NOT_DIGIT:            case OP_NOT_DIGIT:
# Line 5699  for (;;) Line 5945  for (;;)
5945                SCHECK_PARTIAL();                SCHECK_PARTIAL();
5946                break;                break;
5947                }                }
5948              if ((md->ctypes[*eptr] & ctype_digit) != 0) break;              if (MAX_255(*eptr) && (md->ctypes[*eptr] & ctype_digit) != 0) break;
5949              eptr++;              eptr++;
5950              }              }
5951            break;            break;
# Line 5712  for (;;) Line 5958  for (;;)
5958                SCHECK_PARTIAL();                SCHECK_PARTIAL();
5959                break;                break;
5960                }                }
5961              if ((md->ctypes[*eptr] & ctype_digit) == 0) break;              if (!MAX_255(*eptr) || (md->ctypes[*eptr] & ctype_digit) == 0) break;
5962              eptr++;              eptr++;
5963              }              }
5964            break;            break;
# Line 5725  for (;;) Line 5971  for (;;)
5971                SCHECK_PARTIAL();                SCHECK_PARTIAL();
5972                break;                break;
5973                }                }
5974              if ((md->ctypes[*eptr] & ctype_space) != 0) break;              if (MAX_255(*eptr) && (md->ctypes[*eptr] & ctype_space) != 0) break;
5975              eptr++;              eptr++;
5976              }              }
5977            break;            break;
# Line 5738  for (;;) Line 5984  for (;;)
5984                SCHECK_PARTIAL();                SCHECK_PARTIAL();
5985                break;                break;
5986                }                }
5987              if ((md->ctypes[*eptr] & ctype_space) == 0) break;              if (!MAX_255(*eptr) || (md->ctypes[*eptr] & ctype_space) == 0) break;
5988              eptr++;              eptr++;
5989              }              }
5990            break;            break;
# Line 5751  for (;;) Line 5997  for (;;)
5997                SCHECK_PARTIAL();                SCHECK_PARTIAL();
5998                break;                break;
5999                }                }
6000              if ((md->ctypes[*eptr] & ctype_word) != 0) break;              if (MAX_255(*eptr) && (md->ctypes[*eptr] & ctype_word) != 0) break;
6001              eptr++;              eptr++;
6002              }              }
6003            break;            break;
# Line 5764  for (;;) Line 6010  for (;;)
6010                SCHECK_PARTIAL();                SCHECK_PARTIAL();
6011                break;                break;
6012                }                }
6013              if ((md->ctypes[*eptr] & ctype_word) == 0) break;              if (!MAX_255(*eptr) || (md->ctypes[*eptr] & ctype_word) == 0) break;
6014              eptr++;              eptr++;
6015              }              }
6016            break;            break;
# Line 5785  for (;;) Line 6031  for (;;)
6031            RMATCH(eptr, ecode, offset_top, md, eptrb, RM47);            RMATCH(eptr, ecode, offset_top, md, eptrb, RM47);
6032            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
6033            eptr--;            eptr--;
6034            if (ctype == OP_ANYNL && eptr > pp  && *eptr == '\n' &&            if (ctype == OP_ANYNL && eptr > pp  && *eptr == CHAR_LF &&
6035                eptr[-1] == '\r') eptr--;                eptr[-1] == CHAR_CR) eptr--;
6036            }            }
6037          }          }
6038    
# Line 5827  switch (frame->Xwhere) Line 6073  switch (frame->Xwhere)
6073    LBL(35) LBL(43) LBL(47) LBL(48) LBL(49) LBL(50) LBL(51) LBL(52)    LBL(35) LBL(43) LBL(47) LBL(48) LBL(49) LBL(50) LBL(51) LBL(52)
6074    LBL(53) LBL(54) LBL(55) LBL(56) LBL(57) LBL(58) LBL(63) LBL(64)    LBL(53) LBL(54) LBL(55) LBL(56) LBL(57) LBL(58) LBL(63) LBL(64)
6075    LBL(65) LBL(66)    LBL(65) LBL(66)
6076  #ifdef SUPPORT_UTF8  #if defined SUPPORT_UTF || !defined COMPILE_PCRE8
6077    LBL(16) LBL(18) LBL(20) LBL(21) LBL(22) LBL(23) LBL(28) LBL(30)    LBL(21)
6078    #endif
6079    #ifdef SUPPORT_UTF
6080      LBL(16) LBL(18) LBL(20)
6081      LBL(22) LBL(23) LBL(28) LBL(30)
6082    LBL(32) LBL(34) LBL(42) LBL(46)    LBL(32) LBL(34) LBL(42) LBL(46)
6083  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
6084    LBL(36) LBL(37) LBL(38) LBL(39) LBL(40) LBL(41) LBL(44) LBL(45)    LBL(36) LBL(37) LBL(38) LBL(39) LBL(40) LBL(41) LBL(44) LBL(45)
6085    LBL(59) LBL(60) LBL(61) LBL(62)    LBL(59) LBL(60) LBL(61) LBL(62)
6086  #endif  /* SUPPORT_UCP */  #endif  /* SUPPORT_UCP */
6087  #endif  /* SUPPORT_UTF8 */  #endif  /* SUPPORT_UTF */
6088    default:    default:
6089    DPRINTF(("jump error in pcre match: label %d non-existent\n", frame->Xwhere));    DPRINTF(("jump error in pcre match: label %d non-existent\n", frame->Xwhere));
6090    
6091    printf("+++jump error in pcre match: label %d non-existent\n", frame->Xwhere);
6092    
6093    return PCRE_ERROR_INTERNAL;    return PCRE_ERROR_INTERNAL;
6094    }    }
6095  #undef LBL  #undef LBL
# Line 5898  Undefine all the macros that were define Line 6151  Undefine all the macros that were define
6151  ***************************************************************************/  ***************************************************************************/
6152    
6153    
6154    #ifdef NO_RECURSE
6155    /*************************************************
6156    *          Release allocated heap frames         *
6157    *************************************************/
6158    
6159    /* This function releases all the allocated frames. The base frame is on the
6160    machine stack, and so must not be freed.
6161    
6162    Argument: the address of the base frame
6163    Returns:  nothing
6164    */
6165    
6166    static void
6167    release_match_heapframes (heapframe *frame_base)
6168    {
6169    heapframe *nextframe = frame_base->Xnextframe;
6170    while (nextframe != NULL)
6171      {
6172      heapframe *oldframe = nextframe;
6173      nextframe = nextframe->Xnextframe;
6174      (PUBL(stack_free))(oldframe);
6175      }
6176    }
6177    #endif
6178    
6179    
6180  /*************************************************  /*************************************************
6181  *         Execute a Regular Expression           *  *         Execute a Regular Expression           *
# Line 5923  Returns:          > 0 => success; value Line 6201  Returns:          > 0 => success; value
6201                   < -1 => some kind of unexpected problem                   < -1 => some kind of unexpected problem
6202  */  */
6203    
6204    #if defined COMPILE_PCRE8
6205  PCRE_EXP_DEFN int PCRE_CALL_CONVENTION  PCRE_EXP_DEFN int PCRE_CALL_CONVENTION
6206  pcre_exec(const pcre *argument_re, const pcre_extra *extra_data,  pcre_exec(const pcre *argument_re, const pcre_extra *extra_data,
6207    PCRE_SPTR subject, int length, int start_offset, int options, int *offsets,    PCRE_SPTR subject, int length, int start_offset, int options, int *offsets,
6208    int offsetcount)    int offsetcount)
6209    #elif defined COMPILE_PCRE16
6210    PCRE_EXP_DEFN int PCRE_CALL_CONVENTION
6211    pcre16_exec(const pcre16 *argument_re, const pcre16_extra *extra_data,
6212      PCRE_SPTR16 subject, int length, int start_offset, int options, int *offsets,
6213      int offsetcount)
6214    #elif defined COMPILE_PCRE32
6215    PCRE_EXP_DEFN int PCRE_CALL_CONVENTION
6216    pcre32_exec(const pcre32 *argument_re, const pcre32_extra *extra_data,
6217      PCRE_SPTR32 subject, int length, int start_offset, int options, int *offsets,
6218      int offsetcount)
6219    #endif
6220  {  {
6221  int rc, ocount, arg_offset_max;  int rc, ocount, arg_offset_max;
 int first_byte = -1;  
 int req_byte = -1;  
 int req_byte2 = -1;  
6222  int newline;  int newline;
6223  BOOL using_temporary_offsets = FALSE;  BOOL using_temporary_offsets = FALSE;
6224  BOOL anchored;  BOOL anchored;
6225  BOOL startline;  BOOL startline;
6226  BOOL firstline;  BOOL firstline;
6227  BOOL first_byte_caseless = FALSE;  BOOL utf;
6228  BOOL req_byte_caseless = FALSE;  BOOL has_first_char = FALSE;
6229  BOOL utf8;  BOOL has_req_char = FALSE;
6230    pcre_uchar first_char = 0;
6231    pcre_uchar first_char2 = 0;
6232    pcre_uchar req_char = 0;
6233    pcre_uchar req_char2 = 0;
6234  match_data match_block;  match_data match_block;
6235  match_data *md = &match_block;  match_data *md = &match_block;
6236  const uschar *tables;  const pcre_uint8 *tables;
6237  const uschar *start_bits = NULL;  const pcre_uint8 *start_bits = NULL;
6238  USPTR start_match = (USPTR)subject + start_offset;  PCRE_PUCHAR start_match = (PCRE_PUCHAR)subject + start_offset;
6239  USPTR end_subject;  PCRE_PUCHAR end_subject;
6240  USPTR start_partial = NULL;  PCRE_PUCHAR start_partial = NULL;
6241  USPTR req_byte_ptr = start_match - 1;  PCRE_PUCHAR req_char_ptr = start_match - 1;
6242    
 pcre_study_data internal_study;  
6243  const pcre_study_data *study;  const pcre_study_data *study;
6244    const REAL_PCRE *re = (const REAL_PCRE *)argument_re;
6245    
6246  real_pcre internal_re;  #ifdef NO_RECURSE
6247  const real_pcre *external_re = (const real_pcre *)argument_re;  heapframe frame_zero;
6248  const real_pcre *re = external_re;  frame_zero.Xprevframe = NULL;            /* Marks the top level */
6249    frame_zero.Xnextframe = NULL;            /* None are allocated yet */
6250    md->match_frames_base = &frame_zero;
6251    #endif
6252    
6253    /* Check for the special magic call that measures the size of the stack used
6254    per recursive call of match(). Without the funny casting for sizeof, a Windows
6255    compiler gave this error: "unary minus operator applied to unsigned type,
6256    result still unsigned". Hopefully the cast fixes that. */
6257    
6258    if (re == NULL && extra_data == NULL && subject == NULL && length == -999 &&
6259        start_offset == -999)
6260    #ifdef NO_RECURSE
6261      return -((int)sizeof(heapframe));
6262    #else
6263      return match(NULL, NULL, NULL, 0, NULL, NULL, 0);
6264    #endif
6265    
6266  /* Plausibility checks */  /* Plausibility checks */
6267    
6268  if ((options & ~PUBLIC_EXEC_OPTIONS) != 0) return PCRE_ERROR_BADOPTION;  if ((options & ~PUBLIC_EXEC_OPTIONS) != 0) return PCRE_ERROR_BADOPTION;
6269  if (re == NULL || subject == NULL ||  if (re == NULL || subject == NULL || (offsets == NULL && offsetcount > 0))
6270     (offsets == NULL && offsetcount > 0)) return PCRE_ERROR_NULL;    return PCRE_ERROR_NULL;
6271  if (offsetcount < 0) return PCRE_ERROR_BADCOUNT;  if (offsetcount < 0) return PCRE_ERROR_BADCOUNT;
6272  if (start_offset < 0 || start_offset > length) return PCRE_ERROR_BADOFFSET;  if (start_offset < 0 || start_offset > length) return PCRE_ERROR_BADOFFSET;
6273    
6274    /* Check that the first field in the block is the magic number. If it is not,
6275    return with PCRE_ERROR_BADMAGIC. However, if the magic number is equal to
6276    REVERSED_MAGIC_NUMBER we return with PCRE_ERROR_BADENDIANNESS, which
6277    means that the pattern is likely compiled with different endianness. */
6278    
6279    if (re->magic_number != MAGIC_NUMBER)
6280      return re->magic_number == REVERSED_MAGIC_NUMBER?
6281        PCRE_ERROR_BADENDIANNESS:PCRE_ERROR_BADMAGIC;
6282    if ((re->flags & PCRE_MODE) == 0) return PCRE_ERROR_BADMODE;
6283    
6284  /* These two settings are used in the code for checking a UTF-8 string that  /* These two settings are used in the code for checking a UTF-8 string that
6285  follows immediately afterwards. Other values in the md block are used only  follows immediately afterwards. Other values in the md block are used only
6286  during "normal" pcre_exec() processing, not when the JIT support is in use,  during "normal" pcre_exec() processing, not when the JIT support is in use,
6287  so they are set up later. */  so they are set up later. */
6288    
6289  utf8 = md->utf8 = (re->options & PCRE_UTF8) != 0;  /* PCRE_UTF16 has the same value as PCRE_UTF8. */
6290    utf = md->utf = (re->options & PCRE_UTF8) != 0;
6291  md->partial = ((options & PCRE_PARTIAL_HARD) != 0)? 2 :  md->partial = ((options & PCRE_PARTIAL_HARD) != 0)? 2 :
6292                ((options & PCRE_PARTIAL_SOFT) != 0)? 1 : 0;                ((options & PCRE_PARTIAL_SOFT) != 0)? 1 : 0;
6293    
6294  /* Check a UTF-8 string if required. Pass back the character offset and error  /* Check a UTF-8 string if required. Pass back the character offset and error
6295  code for an invalid string if a results vector is available. */  code for an invalid string if a results vector is available. */
6296    
6297  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
6298  if (utf8 && (options & PCRE_NO_UTF8_CHECK) == 0)  if (utf && (options & PCRE_NO_UTF8_CHECK) == 0)
6299    {    {
6300    int erroroffset;    int erroroffset;
6301    int errorcode = _pcre_valid_utf8((USPTR)subject, length, &erroroffset);    int errorcode = PRIV(valid_utf)((PCRE_PUCHAR)subject, length, &erroroffset);
6302    if (errorcode != 0)    if (errorcode != 0)
6303      {      {
6304      if (offsetcount >= 2)      if (offsetcount >= 2)
# Line 5988  if (utf8 && (options & PCRE_NO_UTF8_CHEC Line 6306  if (utf8 && (options & PCRE_NO_UTF8_CHEC
6306        offsets[0] = erroroffset;        offsets[0] = erroroffset;
6307        offsets[1] = errorcode;        offsets[1] = errorcode;
6308        }        }
6309    #if defined COMPILE_PCRE8
6310      return (errorcode <= PCRE_UTF8_ERR5 && md->partial > 1)?      return (errorcode <= PCRE_UTF8_ERR5 && md->partial > 1)?
6311        PCRE_ERROR_SHORTUTF8 : PCRE_ERROR_BADUTF8;        PCRE_ERROR_SHORTUTF8 : PCRE_ERROR_BADUTF8;
6312    #elif defined COMPILE_PCRE16
6313        return (errorcode <= PCRE_UTF16_ERR1 && md->partial > 1)?
6314          PCRE_ERROR_SHORTUTF16 : PCRE_ERROR_BADUTF16;
6315    #elif defined COMPILE_PCRE32
6316        return PCRE_ERROR_BADUTF32;
6317    #endif
6318      }      }
6319    #if defined COMPILE_PCRE8 || defined COMPILE_PCRE16
6320    /* Check that a start_offset points to the start of a UTF-8 character. */    /* Check that a start_offset points to the start of a UTF character. */