/[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 975 by ph10, Sat Jun 2 11:03:06 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 113  Returns:     nothing Line 113  Returns:     nothing
113  */  */
114    
115  static void  static void
116  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)
117  {  {
118  unsigned int c;  unsigned int c;
119  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 140  Arguments:
140    md          points to match data block    md          points to match data block
141    caseless    TRUE if caseless    caseless    TRUE if caseless
142    
143  Returns:      < 0 if not matched, otherwise the number of subject bytes matched  Returns:      >= 0 the number of subject bytes matched
144                  -1 no match
145                  -2 partial match; always given if at end subject
146  */  */
147    
148  static int  static int
149  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,
150    BOOL caseless)    BOOL caseless)
151  {  {
152  USPTR eptr_start = eptr;  PCRE_PUCHAR eptr_start = eptr;
153  register USPTR p = md->start_subject + md->offset_vector[offset];  register PCRE_PUCHAR p = md->start_subject + md->offset_vector[offset];
154    
155  #ifdef PCRE_DEBUG  #ifdef PCRE_DEBUG
156  if (eptr >= md->end_subject)  if (eptr >= md->end_subject)
# Line 163  pchars(p, length, FALSE, md); Line 165  pchars(p, length, FALSE, md);
165  printf("\n");  printf("\n");
166  #endif  #endif
167    
168  /* Always fail if reference not set (and not JavaScript compatible). */  /* Always fail if reference not set (and not JavaScript compatible - in that
169    case the length is passed as zero). */
170    
171  if (length < 0) return -1;  if (length < 0) return -1;
172    
# Line 173  ASCII characters. */ Line 176  ASCII characters. */
176    
177  if (caseless)  if (caseless)
178    {    {
179  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
180  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
181    if (md->utf8)    if (md->utf)
182      {      {
183      /* 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
184      bytes matched may differ, because there are some characters whose upper and      bytes matched may differ, because there are some characters whose upper and
# Line 185  if (caseless) Line 188  if (caseless)
188      the latter. It is important, therefore, to check the length along the      the latter. It is important, therefore, to check the length along the
189      reference, not along the subject (earlier code did this wrong). */      reference, not along the subject (earlier code did this wrong). */
190    
191      USPTR endptr = p + length;      PCRE_PUCHAR endptr = p + length;
192      while (p < endptr)      while (p < endptr)
193        {        {
194        int c, d;        int c, d;
195        if (eptr >= md->end_subject) return -1;        if (eptr >= md->end_subject) return -2;   /* Partial match */
196        GETCHARINC(c, eptr);        GETCHARINC(c, eptr);
197        GETCHARINC(d, p);        GETCHARINC(d, p);
198        if (c != d && c != UCD_OTHERCASE(d)) return -1;        if (c != d && c != UCD_OTHERCASE(d)) return -1;
# Line 202  if (caseless) Line 205  if (caseless)
205    /* 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
206    is no UCP support. */    is no UCP support. */
207      {      {
     if (eptr + length > md->end_subject) return -1;  
208      while (length-- > 0)      while (length-- > 0)
209        { if (md->lcc[*p++] != md->lcc[*eptr++]) return -1; }        {
210          if (eptr >= md->end_subject) return -2;   /* Partial match */
211          if (TABLE_GET(*p, md->lcc, *p) != TABLE_GET(*eptr, md->lcc, *eptr)) return -1;
212          p++;
213          eptr++;
214          }
215      }      }
216    }    }
217    
# Line 213  are in UTF-8 mode. */ Line 220  are in UTF-8 mode. */
220    
221  else  else
222    {    {
223    if (eptr + length > md->end_subject) return -1;    while (length-- > 0)
224    while (length-- > 0) if (*p++ != *eptr++) return -1;      {
225        if (eptr >= md->end_subject) return -2;   /* Partial match */
226        if (*p++ != *eptr++) return -1;
227        }
228    }    }
229    
230  return eptr - eptr_start;  return (int)(eptr - eptr_start);
231  }  }
232    
233    
# Line 307  argument of match(), which never changes Line 317  argument of match(), which never changes
317    
318  #define RMATCH(ra,rb,rc,rd,re,rw)\  #define RMATCH(ra,rb,rc,rd,re,rw)\
319    {\    {\
320    heapframe *newframe = (heapframe *)(pcre_stack_malloc)(sizeof(heapframe));\    heapframe *newframe = frame->Xnextframe;\
321    if (newframe == NULL) RRETURN(PCRE_ERROR_NOMEMORY);\    if (newframe == NULL)\
322    frame->Xwhere = rw; \      {\
323        newframe = (heapframe *)(PUBL(stack_malloc))(sizeof(heapframe));\
324        if (newframe == NULL) RRETURN(PCRE_ERROR_NOMEMORY);\
325        newframe->Xnextframe = NULL;\
326        frame->Xnextframe = newframe;\
327        }\
328      frame->Xwhere = rw;\
329    newframe->Xeptr = ra;\    newframe->Xeptr = ra;\
330    newframe->Xecode = rb;\    newframe->Xecode = rb;\
331    newframe->Xmstart = mstart;\    newframe->Xmstart = mstart;\
# Line 328  argument of match(), which never changes Line 344  argument of match(), which never changes
344    {\    {\
345    heapframe *oldframe = frame;\    heapframe *oldframe = frame;\
346    frame = oldframe->Xprevframe;\    frame = oldframe->Xprevframe;\
   (pcre_stack_free)(oldframe);\  
347    if (frame != NULL)\    if (frame != NULL)\
348      {\      {\
349      rrc = ra;\      rrc = ra;\
# Line 342  argument of match(), which never changes Line 357  argument of match(), which never changes
357    
358  typedef struct heapframe {  typedef struct heapframe {
359    struct heapframe *Xprevframe;    struct heapframe *Xprevframe;
360      struct heapframe *Xnextframe;
361    
362    /* Function arguments that may change */    /* Function arguments that may change */
363    
364    USPTR Xeptr;    PCRE_PUCHAR Xeptr;
365    const uschar *Xecode;    const pcre_uchar *Xecode;
366    USPTR Xmstart;    PCRE_PUCHAR Xmstart;
367    int Xoffset_top;    int Xoffset_top;
368    eptrblock *Xeptrb;    eptrblock *Xeptrb;
369    unsigned int Xrdepth;    unsigned int Xrdepth;
370    
371    /* Function local variables */    /* Function local variables */
372    
373    USPTR Xcallpat;    PCRE_PUCHAR Xcallpat;
374  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
375    USPTR Xcharptr;    PCRE_PUCHAR Xcharptr;
376  #endif  #endif
377    USPTR Xdata;    PCRE_PUCHAR Xdata;
378    USPTR Xnext;    PCRE_PUCHAR Xnext;
379    USPTR Xpp;    PCRE_PUCHAR Xpp;
380    USPTR Xprev;    PCRE_PUCHAR Xprev;
381    USPTR Xsaved_eptr;    PCRE_PUCHAR Xsaved_eptr;
382    
383    recursion_info Xnew_recursive;    recursion_info Xnew_recursive;
384    
# Line 375  typedef struct heapframe { Line 391  typedef struct heapframe {
391    int Xprop_value;    int Xprop_value;
392    int Xprop_fail_result;    int Xprop_fail_result;
393    int Xoclength;    int Xoclength;
394    uschar Xocchars[8];    pcre_uchar Xocchars[6];
395  #endif  #endif
396    
397    int Xcodelink;    int Xcodelink;
# Line 417  returns a negative (error) response, the Line 433  returns a negative (error) response, the
433  same response. */  same response. */
434    
435  /* 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
436  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
437  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.
438  something has been matched). For hard partial matching, we then return  something has been matched). For hard partial matching, we then return
439  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 456  the subject. */
456    
457    
458  /* Performance note: It might be tempting to extract commonly used fields from  /* Performance note: It might be tempting to extract commonly used fields from
459  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
460  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
461  made performance worse.  made performance worse.
462    
# Line 463  Returns:       MATCH_MATCH if matched Line 479  Returns:       MATCH_MATCH if matched
479  */  */
480    
481  static int  static int
482  match(REGISTER USPTR eptr, REGISTER const uschar *ecode, USPTR mstart,  match(REGISTER PCRE_PUCHAR eptr, REGISTER const pcre_uchar *ecode,
483    int offset_top, match_data *md, eptrblock *eptrb, unsigned int rdepth)    PCRE_PUCHAR mstart, int offset_top, match_data *md, eptrblock *eptrb,
484      unsigned int rdepth)
485  {  {
486  /* 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,
487  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 473  so they can be ordinary variables in all Line 490  so they can be ordinary variables in all
490  register int  rrc;         /* Returns from recursive calls */  register int  rrc;         /* Returns from recursive calls */
491  register int  i;           /* Used for loops not involving calls to RMATCH() */  register int  i;           /* Used for loops not involving calls to RMATCH() */
492  register unsigned int c;   /* Character values not kept over RMATCH() calls */  register unsigned int c;   /* Character values not kept over RMATCH() calls */
493  register BOOL utf8;        /* Local copy of UTF-8 flag for speed */  register BOOL utf;         /* Local copy of UTF flag for speed */
494    
495  BOOL minimize, possessive; /* Quantifier options */  BOOL minimize, possessive; /* Quantifier options */
496  BOOL caseless;  BOOL caseless;
497  int condcode;  int condcode;
498    
499  /* 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
500  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
501  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
502  heap whenever RMATCH() does a "recursion". See the macro definitions above. */  whenever RMATCH() does a "recursion". See the macro definitions above. Putting
503    the top-level on the stack rather than malloc-ing them all gives a performance
504    boost in many cases where there is not much "recursion". */
505    
506  #ifdef NO_RECURSE  #ifdef NO_RECURSE
507  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 */  
508    
509  /* Copy in the original argument variables */  /* Copy in the original argument variables */
510    
# Line 513  HEAP_RECURSE: Line 530  HEAP_RECURSE:
530    
531  /* Ditto for the local variables */  /* Ditto for the local variables */
532    
533  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
534  #define charptr            frame->Xcharptr  #define charptr            frame->Xcharptr
535  #endif  #endif
536  #define callpat            frame->Xcallpat  #define callpat            frame->Xcallpat
# Line 571  declarations can be cut out in a block. Line 588  declarations can be cut out in a block.
588  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
589  to RMATCH(). */  to RMATCH(). */
590    
591  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
592  const uschar *charptr;  const pcre_uchar *charptr;
593  #endif  #endif
594  const uschar *callpat;  const pcre_uchar *callpat;
595  const uschar *data;  const pcre_uchar *data;
596  const uschar *next;  const pcre_uchar *next;
597  USPTR         pp;  PCRE_PUCHAR       pp;
598  const uschar *prev;  const pcre_uchar *prev;
599  USPTR         saved_eptr;  PCRE_PUCHAR       saved_eptr;
600    
601  recursion_info new_recursive;  recursion_info new_recursive;
602    
# Line 592  int prop_type; Line 609  int prop_type;
609  int prop_value;  int prop_value;
610  int prop_fail_result;  int prop_fail_result;
611  int oclength;  int oclength;
612  uschar occhars[8];  pcre_uchar occhars[6];
613  #endif  #endif
614    
615  int codelink;  int codelink;
# Line 608  int save_offset1, save_offset2, save_off Line 625  int save_offset1, save_offset2, save_off
625  int stacksave[REC_STACK_SAVE_MAX];  int stacksave[REC_STACK_SAVE_MAX];
626    
627  eptrblock newptrb;  eptrblock newptrb;
628    
629    /* There is a special fudge for calling match() in a way that causes it to
630    measure the size of its basic stack frame when the stack is being used for
631    recursion. The second argument (ecode) being NULL triggers this behaviour. It
632    cannot normally ever be NULL. The return is the negated value of the frame
633    size. */
634    
635    if (ecode == NULL)
636      {
637      if (rdepth == 0)
638        return match((PCRE_PUCHAR)&rdepth, NULL, NULL, 0, NULL, NULL, 1);
639      else
640        {
641        int len = (char *)&rdepth - (char *)eptr;
642        return (len > 0)? -len : len;
643        }
644      }
645  #endif     /* NO_RECURSE */  #endif     /* NO_RECURSE */
646    
647  /* 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 654  the alternative names that are used. */
654  #define code_offset   codelink  #define code_offset   codelink
655  #define condassert    condition  #define condassert    condition
656  #define matched_once  prev_is_word  #define matched_once  prev_is_word
657    #define foc           number
658    #define save_mark     data
659    
660  /* These statements are here to stop the compiler complaining about unitialized  /* These statements are here to stop the compiler complaining about unitialized
661  variables. */  variables. */
# Line 645  defined). However, RMATCH isn't like a f Line 681  defined). However, RMATCH isn't like a f
681  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,
682  however, impact performance when true recursion is being used. */  however, impact performance when true recursion is being used. */
683    
684  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
685  utf8 = md->utf8;       /* Local copy of the flag */  utf = md->utf;       /* Local copy of the flag */
686  #else  #else
687  utf8 = FALSE;  utf = FALSE;
688  #endif  #endif
689    
690  /* 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 725  for (;;)
725      case OP_MARK:      case OP_MARK:
726      md->nomatch_mark = ecode + 2;      md->nomatch_mark = ecode + 2;
727      md->mark = NULL;    /* In case previously set by assertion */      md->mark = NULL;    /* In case previously set by assertion */
728      RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode] + ecode[1], offset_top, md,      RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode] + ecode[1], offset_top, md,
729        eptrb, RM55);        eptrb, RM55);
730      if ((rrc == MATCH_MATCH || rrc == MATCH_ACCEPT) &&      if ((rrc == MATCH_MATCH || rrc == MATCH_ACCEPT) &&
731           md->mark == NULL) md->mark = ecode + 2;           md->mark == NULL) md->mark = ecode + 2;
# Line 702  for (;;) Line 738  for (;;)
738      unaltered. */      unaltered. */
739    
740      else if (rrc == MATCH_SKIP_ARG &&      else if (rrc == MATCH_SKIP_ARG &&
741          strcmp((char *)(ecode + 2), (char *)(md->start_match_ptr)) == 0)          STRCMP_UC_UC(ecode + 2, md->start_match_ptr) == 0)
742        {        {
743        md->start_match_ptr = eptr;        md->start_match_ptr = eptr;
744        RRETURN(MATCH_SKIP);        RRETURN(MATCH_SKIP);
# Line 715  for (;;) Line 751  for (;;)
751      /* COMMIT overrides PRUNE, SKIP, and THEN */      /* COMMIT overrides PRUNE, SKIP, and THEN */
752    
753      case OP_COMMIT:      case OP_COMMIT:
754      RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,      RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md,
755        eptrb, RM52);        eptrb, RM52);
756      if (rrc != MATCH_NOMATCH && rrc != MATCH_PRUNE &&      if (rrc != MATCH_NOMATCH && rrc != MATCH_PRUNE &&
757          rrc != MATCH_SKIP && rrc != MATCH_SKIP_ARG &&          rrc != MATCH_SKIP && rrc != MATCH_SKIP_ARG &&
# Line 726  for (;;) Line 762  for (;;)
762      /* PRUNE overrides THEN */      /* PRUNE overrides THEN */
763    
764      case OP_PRUNE:      case OP_PRUNE:
765      RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,      RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md,
766        eptrb, RM51);        eptrb, RM51);
767      if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN) RRETURN(rrc);      if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN) RRETURN(rrc);
768      RRETURN(MATCH_PRUNE);      RRETURN(MATCH_PRUNE);
# Line 734  for (;;) Line 770  for (;;)
770      case OP_PRUNE_ARG:      case OP_PRUNE_ARG:
771      md->nomatch_mark = ecode + 2;      md->nomatch_mark = ecode + 2;
772      md->mark = NULL;    /* In case previously set by assertion */      md->mark = NULL;    /* In case previously set by assertion */
773      RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode] + ecode[1], offset_top, md,      RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode] + ecode[1], offset_top, md,
774        eptrb, RM56);        eptrb, RM56);
775      if ((rrc == MATCH_MATCH || rrc == MATCH_ACCEPT) &&      if ((rrc == MATCH_MATCH || rrc == MATCH_ACCEPT) &&
776           md->mark == NULL) md->mark = ecode + 2;           md->mark == NULL) md->mark = ecode + 2;
# Line 744  for (;;) Line 780  for (;;)
780      /* SKIP overrides PRUNE and THEN */      /* SKIP overrides PRUNE and THEN */
781    
782      case OP_SKIP:      case OP_SKIP:
783      RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,      RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md,
784        eptrb, RM53);        eptrb, RM53);
785      if (rrc != MATCH_NOMATCH && rrc != MATCH_PRUNE && rrc != MATCH_THEN)      if (rrc != MATCH_NOMATCH && rrc != MATCH_PRUNE && rrc != MATCH_THEN)
786        RRETURN(rrc);        RRETURN(rrc);
# Line 752  for (;;) Line 788  for (;;)
788      RRETURN(MATCH_SKIP);      RRETURN(MATCH_SKIP);
789    
790      /* Note that, for Perl compatibility, SKIP with an argument does NOT set      /* Note that, for Perl compatibility, SKIP with an argument does NOT set
791      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
792      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. */
793    
794      case OP_SKIP_ARG:      case OP_SKIP_ARG:
795      if (md->ignore_skip_arg)      if (md->ignore_skip_arg)
796        {        {
797        ecode += _pcre_OP_lengths[*ecode] + ecode[1];        ecode += PRIV(OP_lengths)[*ecode] + ecode[1];
798        break;        break;
799        }        }
800      RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode] + ecode[1], offset_top, md,      RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode] + ecode[1], offset_top, md,
801        eptrb, RM57);        eptrb, RM57);
802      if (rrc != MATCH_NOMATCH && rrc != MATCH_PRUNE && rrc != MATCH_THEN)      if (rrc != MATCH_NOMATCH && rrc != MATCH_PRUNE && rrc != MATCH_THEN)
803        RRETURN(rrc);        RRETURN(rrc);
804    
805      /* 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
806      returning the special MATCH_SKIP_ARG return code. This will either be      returning the special MATCH_SKIP_ARG return code. This will either be
807      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
808      with the md->ignore_skip_arg flag set. */      with the md->ignore_skip_arg flag set. */
809    
810      md->start_match_ptr = ecode + 2;      md->start_match_ptr = ecode + 2;
# Line 779  for (;;) Line 815  for (;;)
815      match pointer to do this. */      match pointer to do this. */
816    
817      case OP_THEN:      case OP_THEN:
818      RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,      RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md,
819        eptrb, RM54);        eptrb, RM54);
820      if (rrc != MATCH_NOMATCH) RRETURN(rrc);      if (rrc != MATCH_NOMATCH) RRETURN(rrc);
821      md->start_match_ptr = ecode;      md->start_match_ptr = ecode;
# Line 788  for (;;) Line 824  for (;;)
824      case OP_THEN_ARG:      case OP_THEN_ARG:
825      md->nomatch_mark = ecode + 2;      md->nomatch_mark = ecode + 2;
826      md->mark = NULL;    /* In case previously set by assertion */      md->mark = NULL;    /* In case previously set by assertion */
827      RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode] + ecode[1], offset_top,      RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode] + ecode[1], offset_top,
828        md, eptrb, RM58);        md, eptrb, RM58);
829      if ((rrc == MATCH_MATCH || rrc == MATCH_ACCEPT) &&      if ((rrc == MATCH_MATCH || rrc == MATCH_ACCEPT) &&
830           md->mark == NULL) md->mark = ecode + 2;           md->mark == NULL) md->mark = ecode + 2;
# Line 812  for (;;) Line 848  for (;;)
848      case OP_ONCE_NC:      case OP_ONCE_NC:
849      prev = ecode;      prev = ecode;
850      saved_eptr = eptr;      saved_eptr = eptr;
851        save_mark = md->mark;
852      do      do
853        {        {
854        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 867  for (;;)
867    
868        if (rrc != MATCH_NOMATCH) RRETURN(rrc);        if (rrc != MATCH_NOMATCH) RRETURN(rrc);
869        ecode += GET(ecode,1);        ecode += GET(ecode,1);
870          md->mark = save_mark;
871        }        }
872      while (*ecode == OP_ALT);      while (*ecode == OP_ALT);
873    
# Line 909  for (;;) Line 947  for (;;)
947        save_offset2 = md->offset_vector[offset+1];        save_offset2 = md->offset_vector[offset+1];
948        save_offset3 = md->offset_vector[md->offset_end - number];        save_offset3 = md->offset_vector[md->offset_end - number];
949        save_capture_last = md->capture_last;        save_capture_last = md->capture_last;
950          save_mark = md->mark;
951    
952        DPRINTF(("saving %d %d %d\n", save_offset1, save_offset2, save_offset3));        DPRINTF(("saving %d %d %d\n", save_offset1, save_offset2, save_offset3));
953        md->offset_vector[md->offset_end - number] =        md->offset_vector[md->offset_end - number] =
# Line 917  for (;;) Line 956  for (;;)
956        for (;;)        for (;;)
957          {          {
958          if (op >= OP_SBRA) md->match_function_type = MATCH_CBEGROUP;          if (op >= OP_SBRA) md->match_function_type = MATCH_CBEGROUP;
959          RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,          RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md,
960            eptrb, RM1);            eptrb, RM1);
961          if (rrc == MATCH_ONCE) break;  /* Backing up through an atomic group */          if (rrc == MATCH_ONCE) break;  /* Backing up through an atomic group */
962    
# Line 945  for (;;) Line 984  for (;;)
984          if (rrc != MATCH_NOMATCH) RRETURN(rrc);          if (rrc != MATCH_NOMATCH) RRETURN(rrc);
985          md->capture_last = save_capture_last;          md->capture_last = save_capture_last;
986          ecode += GET(ecode, 1);          ecode += GET(ecode, 1);
987            md->mark = save_mark;
988          if (*ecode != OP_ALT) break;          if (*ecode != OP_ALT) break;
989          }          }
990    
# Line 1004  for (;;) Line 1044  for (;;)
1044    
1045        else if (!md->hasthen && ecode[GET(ecode, 1)] != OP_ALT)        else if (!md->hasthen && ecode[GET(ecode, 1)] != OP_ALT)
1046          {          {
1047          ecode += _pcre_OP_lengths[*ecode];          ecode += PRIV(OP_lengths)[*ecode];
1048          goto TAIL_RECURSE;          goto TAIL_RECURSE;
1049          }          }
1050    
1051        /* In all other cases, we have to make another call to match(). */        /* In all other cases, we have to make another call to match(). */
1052    
1053        RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md, eptrb,        save_mark = md->mark;
1054          RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md, eptrb,
1055          RM2);          RM2);
1056    
1057        /* 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 1069  for (;;)
1069          {          {
1070          if (rrc == MATCH_ONCE)          if (rrc == MATCH_ONCE)
1071            {            {
1072            const uschar *scode = ecode;            const pcre_uchar *scode = ecode;
1073            if (*scode != OP_ONCE)           /* If not at start, find it */            if (*scode != OP_ONCE)           /* If not at start, find it */
1074              {              {
1075              while (*scode == OP_ALT) scode += GET(scode, 1);              while (*scode == OP_ALT) scode += GET(scode, 1);
# Line 1039  for (;;) Line 1080  for (;;)
1080          RRETURN(rrc);          RRETURN(rrc);
1081          }          }
1082        ecode += GET(ecode, 1);        ecode += GET(ecode, 1);
1083          md->mark = save_mark;
1084        if (*ecode != OP_ALT) break;        if (*ecode != OP_ALT) break;
1085        }        }
1086    
# Line 1070  for (;;) Line 1112  for (;;)
1112      if (offset < md->offset_max)      if (offset < md->offset_max)
1113        {        {
1114        matched_once = FALSE;        matched_once = FALSE;
1115        code_offset = ecode - md->start_code;        code_offset = (int)(ecode - md->start_code);
1116    
1117        save_offset1 = md->offset_vector[offset];        save_offset1 = md->offset_vector[offset];
1118        save_offset2 = md->offset_vector[offset+1];        save_offset2 = md->offset_vector[offset+1];
# Line 1093  for (;;) Line 1135  for (;;)
1135          md->offset_vector[md->offset_end - number] =          md->offset_vector[md->offset_end - number] =
1136            (int)(eptr - md->start_subject);            (int)(eptr - md->start_subject);
1137          if (op >= OP_SBRA) md->match_function_type = MATCH_CBEGROUP;          if (op >= OP_SBRA) md->match_function_type = MATCH_CBEGROUP;
1138          RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,          RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md,
1139            eptrb, RM63);            eptrb, RM63);
1140          if (rrc == MATCH_KETRPOS)          if (rrc == MATCH_KETRPOS)
1141            {            {
# Line 1160  for (;;) Line 1202  for (;;)
1202    
1203      POSSESSIVE_NON_CAPTURE:      POSSESSIVE_NON_CAPTURE:
1204      matched_once = FALSE;      matched_once = FALSE;
1205      code_offset = ecode - md->start_code;      code_offset = (int)(ecode - md->start_code);
1206    
1207      for (;;)      for (;;)
1208        {        {
1209        if (op >= OP_SBRA) md->match_function_type = MATCH_CBEGROUP;        if (op >= OP_SBRA) md->match_function_type = MATCH_CBEGROUP;
1210        RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,        RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md,
1211          eptrb, RM48);          eptrb, RM48);
1212        if (rrc == MATCH_KETRPOS)        if (rrc == MATCH_KETRPOS)
1213          {          {
# Line 1215  for (;;) Line 1257  for (;;)
1257    
1258      if (ecode[LINK_SIZE+1] == OP_CALLOUT)      if (ecode[LINK_SIZE+1] == OP_CALLOUT)
1259        {        {
1260        if (pcre_callout != NULL)        if (PUBL(callout) != NULL)
1261          {          {
1262          pcre_callout_block cb;          PUBL(callout_block) cb;
1263          cb.version          = 2;   /* Version 1 of the callout block */          cb.version          = 2;   /* Version 1 of the callout block */
1264          cb.callout_number   = ecode[LINK_SIZE+2];          cb.callout_number   = ecode[LINK_SIZE+2];
1265          cb.offset_vector    = md->offset_vector;          cb.offset_vector    = md->offset_vector;
1266    #ifdef COMPILE_PCRE8
1267          cb.subject          = (PCRE_SPTR)md->start_subject;          cb.subject          = (PCRE_SPTR)md->start_subject;
1268    #else
1269            cb.subject          = (PCRE_SPTR16)md->start_subject;
1270    #endif
1271          cb.subject_length   = (int)(md->end_subject - md->start_subject);          cb.subject_length   = (int)(md->end_subject - md->start_subject);
1272          cb.start_match      = (int)(mstart - md->start_subject);          cb.start_match      = (int)(mstart - md->start_subject);
1273          cb.current_position = (int)(eptr - md->start_subject);          cb.current_position = (int)(eptr - md->start_subject);
# Line 1231  for (;;) Line 1277  for (;;)
1277          cb.capture_last     = md->capture_last;          cb.capture_last     = md->capture_last;
1278          cb.callout_data     = md->callout_data;          cb.callout_data     = md->callout_data;
1279          cb.mark             = md->nomatch_mark;          cb.mark             = md->nomatch_mark;
1280          if ((rrc = (*pcre_callout)(&cb)) > 0) RRETURN(MATCH_NOMATCH);          if ((rrc = (*PUBL(callout))(&cb)) > 0) RRETURN(MATCH_NOMATCH);
1281          if (rrc < 0) RRETURN(rrc);          if (rrc < 0) RRETURN(rrc);
1282          }          }
1283        ecode += _pcre_OP_lengths[OP_CALLOUT];        ecode += PRIV(OP_lengths)[OP_CALLOUT];
1284        }        }
1285    
1286      condcode = ecode[LINK_SIZE+1];      condcode = ecode[LINK_SIZE+1];
# Line 1260  for (;;) Line 1306  for (;;)
1306    
1307          if (!condition && condcode == OP_NRREF)          if (!condition && condcode == OP_NRREF)
1308            {            {
1309            uschar *slotA = md->name_table;            pcre_uchar *slotA = md->name_table;
1310            for (i = 0; i < md->name_count; i++)            for (i = 0; i < md->name_count; i++)
1311              {              {
1312              if (GET2(slotA, 0) == recno) break;              if (GET2(slotA, 0) == recno) break;
# Line 1273  for (;;) Line 1319  for (;;)
1319    
1320            if (i < md->name_count)            if (i < md->name_count)
1321              {              {
1322              uschar *slotB = slotA;              pcre_uchar *slotB = slotA;
1323              while (slotB > md->name_table)              while (slotB > md->name_table)
1324                {                {
1325                slotB -= md->name_entry_size;                slotB -= md->name_entry_size;
1326                if (strcmp((char *)slotA + 2, (char *)slotB + 2) == 0)                if (STRCMP_UC_UC(slotA + IMM2_SIZE, slotB + IMM2_SIZE) == 0)
1327                  {                  {
1328                  condition = GET2(slotB, 0) == md->recursive->group_num;                  condition = GET2(slotB, 0) == md->recursive->group_num;
1329                  if (condition) break;                  if (condition) break;
# Line 1293  for (;;) Line 1339  for (;;)
1339                for (i++; i < md->name_count; i++)                for (i++; i < md->name_count; i++)
1340                  {                  {
1341                  slotB += md->name_entry_size;                  slotB += md->name_entry_size;
1342                  if (strcmp((char *)slotA + 2, (char *)slotB + 2) == 0)                  if (STRCMP_UC_UC(slotA + IMM2_SIZE, slotB + IMM2_SIZE) == 0)
1343                    {                    {
1344                    condition = GET2(slotB, 0) == md->recursive->group_num;                    condition = GET2(slotB, 0) == md->recursive->group_num;
1345                    if (condition) break;                    if (condition) break;
# Line 1306  for (;;) Line 1352  for (;;)
1352    
1353          /* Chose branch according to the condition */          /* Chose branch according to the condition */
1354    
1355          ecode += condition? 3 : GET(ecode, 1);          ecode += condition? 1 + IMM2_SIZE : GET(ecode, 1);
1356          }          }
1357        }        }
1358    
# Line 1323  for (;;) Line 1369  for (;;)
1369        if (!condition && condcode == OP_NCREF)        if (!condition && condcode == OP_NCREF)
1370          {          {
1371          int refno = offset >> 1;          int refno = offset >> 1;
1372          uschar *slotA = md->name_table;          pcre_uchar *slotA = md->name_table;
1373    
1374          for (i = 0; i < md->name_count; i++)          for (i = 0; i < md->name_count; i++)
1375            {            {
# Line 1337  for (;;) Line 1383  for (;;)
1383    
1384          if (i < md->name_count)          if (i < md->name_count)
1385            {            {
1386            uschar *slotB = slotA;            pcre_uchar *slotB = slotA;
1387            while (slotB > md->name_table)            while (slotB > md->name_table)
1388              {              {
1389              slotB -= md->name_entry_size;              slotB -= md->name_entry_size;
1390              if (strcmp((char *)slotA + 2, (char *)slotB + 2) == 0)              if (STRCMP_UC_UC(slotA + IMM2_SIZE, slotB + IMM2_SIZE) == 0)
1391                {                {
1392                offset = GET2(slotB, 0) << 1;                offset = GET2(slotB, 0) << 1;
1393                condition = offset < offset_top &&                condition = offset < offset_top &&
# Line 1359  for (;;) Line 1405  for (;;)
1405              for (i++; i < md->name_count; i++)              for (i++; i < md->name_count; i++)
1406                {                {
1407                slotB += md->name_entry_size;                slotB += md->name_entry_size;
1408                if (strcmp((char *)slotA + 2, (char *)slotB + 2) == 0)                if (STRCMP_UC_UC(slotA + IMM2_SIZE, slotB + IMM2_SIZE) == 0)
1409                  {                  {
1410                  offset = GET2(slotB, 0) << 1;                  offset = GET2(slotB, 0) << 1;
1411                  condition = offset < offset_top &&                  condition = offset < offset_top &&
# Line 1374  for (;;) Line 1420  for (;;)
1420    
1421        /* Chose branch according to the condition */        /* Chose branch according to the condition */
1422    
1423        ecode += condition? 3 : GET(ecode, 1);        ecode += condition? 1 + IMM2_SIZE : GET(ecode, 1);
1424        }        }
1425    
1426      else if (condcode == OP_DEF)     /* DEFINE - always false */      else if (condcode == OP_DEF)     /* DEFINE - always false */
# Line 1466  for (;;) Line 1512  for (;;)
1512        md->offset_vector[offset+1] = (int)(eptr - md->start_subject);        md->offset_vector[offset+1] = (int)(eptr - md->start_subject);
1513        if (offset_top <= offset) offset_top = offset + 2;        if (offset_top <= offset) offset_top = offset + 2;
1514        }        }
1515      ecode += 3;      ecode += 1 + IMM2_SIZE;
1516      break;      break;
1517    
1518    
# Line 1513  for (;;) Line 1559  for (;;)
1559    
1560      case OP_ASSERT:      case OP_ASSERT:
1561      case OP_ASSERTBACK:      case OP_ASSERTBACK:
1562        save_mark = md->mark;
1563      if (md->match_function_type == MATCH_CONDASSERT)      if (md->match_function_type == MATCH_CONDASSERT)
1564        {        {
1565        condassert = TRUE;        condassert = TRUE;
# Line 1528  for (;;) Line 1575  for (;;)
1575          mstart = md->start_match_ptr;   /* In case \K reset it */          mstart = md->start_match_ptr;   /* In case \K reset it */
1576          break;          break;
1577          }          }
1578          md->mark = save_mark;
1579    
1580        /* PCRE does not allow THEN to escape beyond an assertion; it is treated        /* A COMMIT failure must fail the entire assertion, without trying any
1581        as NOMATCH. */        subsequent branches. */
1582    
1583          if (rrc == MATCH_COMMIT) RRETURN(MATCH_NOMATCH);
1584    
1585          /* PCRE does not allow THEN to escape beyond an assertion; it
1586          is treated as NOMATCH. */
1587    
1588        if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN) RRETURN(rrc);        if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN) RRETURN(rrc);
1589        ecode += GET(ecode, 1);        ecode += GET(ecode, 1);
# Line 1557  for (;;) Line 1610  for (;;)
1610    
1611      case OP_ASSERT_NOT:      case OP_ASSERT_NOT:
1612      case OP_ASSERTBACK_NOT:      case OP_ASSERTBACK_NOT:
1613        save_mark = md->mark;
1614      if (md->match_function_type == MATCH_CONDASSERT)      if (md->match_function_type == MATCH_CONDASSERT)
1615        {        {
1616        condassert = TRUE;        condassert = TRUE;
# Line 1567  for (;;) Line 1621  for (;;)
1621      do      do
1622        {        {
1623        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, NULL, RM5);        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, NULL, RM5);
1624          md->mark = save_mark;
1625        if (rrc == MATCH_MATCH || rrc == MATCH_ACCEPT) RRETURN(MATCH_NOMATCH);        if (rrc == MATCH_MATCH || rrc == MATCH_ACCEPT) RRETURN(MATCH_NOMATCH);
1626        if (rrc == MATCH_SKIP || rrc == MATCH_PRUNE || rrc == MATCH_COMMIT)        if (rrc == MATCH_SKIP || rrc == MATCH_PRUNE || rrc == MATCH_COMMIT)
1627          {          {
# Line 1593  for (;;) Line 1648  for (;;)
1648      back a number of characters, not bytes. */      back a number of characters, not bytes. */
1649    
1650      case OP_REVERSE:      case OP_REVERSE:
1651  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
1652      if (utf8)      if (utf)
1653        {        {
1654        i = GET(ecode, 1);        i = GET(ecode, 1);
1655        while (i-- > 0)        while (i-- > 0)
# Line 1625  for (;;) Line 1680  for (;;)
1680      function is able to force a failure. */      function is able to force a failure. */
1681    
1682      case OP_CALLOUT:      case OP_CALLOUT:
1683      if (pcre_callout != NULL)      if (PUBL(callout) != NULL)
1684        {        {
1685        pcre_callout_block cb;        PUBL(callout_block) cb;
1686        cb.version          = 2;   /* Version 1 of the callout block */        cb.version          = 2;   /* Version 1 of the callout block */
1687        cb.callout_number   = ecode[1];        cb.callout_number   = ecode[1];
1688        cb.offset_vector    = md->offset_vector;        cb.offset_vector    = md->offset_vector;
1689    #ifdef COMPILE_PCRE8
1690        cb.subject          = (PCRE_SPTR)md->start_subject;        cb.subject          = (PCRE_SPTR)md->start_subject;
1691    #else
1692          cb.subject          = (PCRE_SPTR16)md->start_subject;
1693    #endif
1694        cb.subject_length   = (int)(md->end_subject - md->start_subject);        cb.subject_length   = (int)(md->end_subject - md->start_subject);
1695        cb.start_match      = (int)(mstart - md->start_subject);        cb.start_match      = (int)(mstart - md->start_subject);
1696        cb.current_position = (int)(eptr - md->start_subject);        cb.current_position = (int)(eptr - md->start_subject);
# Line 1641  for (;;) Line 1700  for (;;)
1700        cb.capture_last     = md->capture_last;        cb.capture_last     = md->capture_last;
1701        cb.callout_data     = md->callout_data;        cb.callout_data     = md->callout_data;
1702        cb.mark             = md->nomatch_mark;        cb.mark             = md->nomatch_mark;
1703        if ((rrc = (*pcre_callout)(&cb)) > 0) RRETURN(MATCH_NOMATCH);        if ((rrc = (*PUBL(callout))(&cb)) > 0) RRETURN(MATCH_NOMATCH);
1704        if (rrc < 0) RRETURN(rrc);        if (rrc < 0) RRETURN(rrc);
1705        }        }
1706      ecode += 2 + 2*LINK_SIZE;      ecode += 2 + 2*LINK_SIZE;
# Line 1700  for (;;) Line 1759  for (;;)
1759        else        else
1760          {          {
1761          new_recursive.offset_save =          new_recursive.offset_save =
1762            (int *)(pcre_malloc)(new_recursive.saved_max * sizeof(int));            (int *)(PUBL(malloc))(new_recursive.saved_max * sizeof(int));
1763          if (new_recursive.offset_save == NULL) RRETURN(PCRE_ERROR_NOMEMORY);          if (new_recursive.offset_save == NULL) RRETURN(PCRE_ERROR_NOMEMORY);
1764          }          }
1765        memcpy(new_recursive.offset_save, md->offset_vector,        memcpy(new_recursive.offset_save, md->offset_vector,
# Line 1715  for (;;) Line 1774  for (;;)
1774        do        do
1775          {          {
1776          if (cbegroup) md->match_function_type = MATCH_CBEGROUP;          if (cbegroup) md->match_function_type = MATCH_CBEGROUP;
1777          RMATCH(eptr, callpat + _pcre_OP_lengths[*callpat], offset_top,          RMATCH(eptr, callpat + PRIV(OP_lengths)[*callpat], offset_top,
1778            md, eptrb, RM6);            md, eptrb, RM6);
1779          memcpy(md->offset_vector, new_recursive.offset_save,          memcpy(md->offset_vector, new_recursive.offset_save,
1780              new_recursive.saved_max * sizeof(int));              new_recursive.saved_max * sizeof(int));
# Line 1724  for (;;) Line 1783  for (;;)
1783            {            {
1784            DPRINTF(("Recursion matched\n"));            DPRINTF(("Recursion matched\n"));
1785            if (new_recursive.offset_save != stacksave)            if (new_recursive.offset_save != stacksave)
1786              (pcre_free)(new_recursive.offset_save);              (PUBL(free))(new_recursive.offset_save);
1787    
1788            /* 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
1789            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 1794  for (;;)
1794            goto RECURSION_MATCHED;        /* Exit loop; end processing */            goto RECURSION_MATCHED;        /* Exit loop; end processing */
1795            }            }
1796    
1797          /* 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
1798          as NOMATCH. */          is treated as NOMATCH. */
1799    
1800          else if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN)          else if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN &&
1801                     rrc != MATCH_COMMIT)
1802            {            {
1803            DPRINTF(("Recursion gave error %d\n", rrc));            DPRINTF(("Recursion gave error %d\n", rrc));
1804            if (new_recursive.offset_save != stacksave)            if (new_recursive.offset_save != stacksave)
1805              (pcre_free)(new_recursive.offset_save);              (PUBL(free))(new_recursive.offset_save);
1806            RRETURN(rrc);            RRETURN(rrc);
1807            }            }
1808    
# Line 1754  for (;;) Line 1814  for (;;)
1814        DPRINTF(("Recursion didn't match\n"));        DPRINTF(("Recursion didn't match\n"));
1815        md->recursive = new_recursive.prevrec;        md->recursive = new_recursive.prevrec;
1816        if (new_recursive.offset_save != stacksave)        if (new_recursive.offset_save != stacksave)
1817          (pcre_free)(new_recursive.offset_save);          (PUBL(free))(new_recursive.offset_save);
1818        RRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
1819        }        }
1820    
# Line 2015  for (;;) Line 2075  for (;;)
2075    
2076      case OP_DOLLM:      case OP_DOLLM:
2077      if (eptr < md->end_subject)      if (eptr < md->end_subject)
2078        { if (!IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH); }        {
2079          if (!IS_NEWLINE(eptr))
2080            {
2081            if (md->partial != 0 &&
2082                eptr + 1 >= md->end_subject &&
2083                NLBLOCK->nltype == NLTYPE_FIXED &&
2084                NLBLOCK->nllen == 2 &&
2085                *eptr == NLBLOCK->nl[0])
2086              {
2087              md->hitend = TRUE;
2088              if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
2089              }
2090            RRETURN(MATCH_NOMATCH);
2091            }
2092          }
2093      else      else
2094        {        {
2095        if (md->noteol) RRETURN(MATCH_NOMATCH);        if (md->noteol) RRETURN(MATCH_NOMATCH);
# Line 2047  for (;;) Line 2121  for (;;)
2121      ASSERT_NL_OR_EOS:      ASSERT_NL_OR_EOS:
2122      if (eptr < md->end_subject &&      if (eptr < md->end_subject &&
2123          (!IS_NEWLINE(eptr) || eptr != md->end_subject - md->nllen))          (!IS_NEWLINE(eptr) || eptr != md->end_subject - md->nllen))
2124          {
2125          if (md->partial != 0 &&
2126              eptr + 1 >= md->end_subject &&
2127              NLBLOCK->nltype == NLTYPE_FIXED &&
2128              NLBLOCK->nllen == 2 &&
2129              *eptr == NLBLOCK->nl[0])
2130            {
2131            md->hitend = TRUE;
2132            if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
2133            }
2134        RRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2135          }
2136    
2137      /* Either at end of string or \n before end. */      /* Either at end of string or \n before end. */
2138    
# Line 2066  for (;;) Line 2151  for (;;)
2151        be "non-word" characters. Remember the earliest consulted character for        be "non-word" characters. Remember the earliest consulted character for
2152        partial matching. */        partial matching. */
2153    
2154  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
2155        if (utf8)        if (utf)
2156          {          {
2157          /* Get status of previous character */          /* Get status of previous character */
2158    
2159          if (eptr == md->start_subject) prev_is_word = FALSE; else          if (eptr == md->start_subject) prev_is_word = FALSE; else
2160            {            {
2161            USPTR lastptr = eptr - 1;            PCRE_PUCHAR lastptr = eptr - 1;
2162            while((*lastptr & 0xc0) == 0x80) lastptr--;            BACKCHAR(lastptr);
2163            if (lastptr < md->start_used_ptr) md->start_used_ptr = lastptr;            if (lastptr < md->start_used_ptr) md->start_used_ptr = lastptr;
2164            GETCHAR(c, lastptr);            GETCHAR(c, lastptr);
2165  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
# Line 2139  for (;;) Line 2224  for (;;)
2224              }              }
2225            else            else
2226  #endif  #endif
2227            prev_is_word = ((md->ctypes[eptr[-1]] & ctype_word) != 0);            prev_is_word = MAX_255(eptr[-1])
2228                && ((md->ctypes[eptr[-1]] & ctype_word) != 0);
2229            }            }
2230    
2231          /* Get status of next character */          /* Get status of next character */
# Line 2162  for (;;) Line 2248  for (;;)
2248            }            }
2249          else          else
2250  #endif  #endif
2251          cur_is_word = ((md->ctypes[*eptr] & ctype_word) != 0);          cur_is_word = MAX_255(*eptr)
2252              && ((md->ctypes[*eptr] & ctype_word) != 0);
2253          }          }
2254    
2255        /* Now see if the situation is what we want */        /* Now see if the situation is what we want */
# Line 2173  for (;;) Line 2260  for (;;)
2260        }        }
2261      break;      break;
2262    
2263      /* Match a single character type; inline for speed */      /* Match any single character type except newline; have to take care with
2264        CRLF newlines and partial matching. */
2265    
2266      case OP_ANY:      case OP_ANY:
2267      if (IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH);      if (IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH);
2268        if (md->partial != 0 &&
2269            eptr + 1 >= md->end_subject &&
2270            NLBLOCK->nltype == NLTYPE_FIXED &&
2271            NLBLOCK->nllen == 2 &&
2272            *eptr == NLBLOCK->nl[0])
2273          {
2274          md->hitend = TRUE;
2275          if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
2276          }
2277    
2278      /* Fall through */      /* Fall through */
2279    
2280        /* Match any single character whatsoever. */
2281    
2282      case OP_ALLANY:      case OP_ALLANY:
2283      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 */
2284        {                            /* not be updated before SCHECK_PARTIAL. */        {                            /* not be updated before SCHECK_PARTIAL. */
# Line 2186  for (;;) Line 2286  for (;;)
2286        RRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2287        }        }
2288      eptr++;      eptr++;
2289      if (utf8) while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;  #ifdef SUPPORT_UTF
2290        if (utf) ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);
2291    #endif
2292      ecode++;      ecode++;
2293      break;      break;
2294    
# Line 2211  for (;;) Line 2313  for (;;)
2313        }        }
2314      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2315      if (      if (
2316  #ifdef SUPPORT_UTF8  #if defined SUPPORT_UTF || !(defined COMPILE_PCRE8)
2317         c < 256 &&         c < 256 &&
2318  #endif  #endif
2319         (md->ctypes[c] & ctype_digit) != 0         (md->ctypes[c] & ctype_digit) != 0
# Line 2228  for (;;) Line 2330  for (;;)
2330        }        }
2331      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2332      if (      if (
2333  #ifdef SUPPORT_UTF8  #if defined SUPPORT_UTF || !(defined COMPILE_PCRE8)
2334         c >= 256 ||         c > 255 ||
2335  #endif  #endif
2336         (md->ctypes[c] & ctype_digit) == 0         (md->ctypes[c] & ctype_digit) == 0
2337         )         )
# Line 2245  for (;;) Line 2347  for (;;)
2347        }        }
2348      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2349      if (      if (
2350  #ifdef SUPPORT_UTF8  #if defined SUPPORT_UTF || !(defined COMPILE_PCRE8)
2351         c < 256 &&         c < 256 &&
2352  #endif  #endif
2353         (md->ctypes[c] & ctype_space) != 0         (md->ctypes[c] & ctype_space) != 0
# Line 2262  for (;;) Line 2364  for (;;)
2364        }        }
2365      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2366      if (      if (
2367  #ifdef SUPPORT_UTF8  #if defined SUPPORT_UTF || !(defined COMPILE_PCRE8)
2368         c >= 256 ||         c > 255 ||
2369  #endif  #endif
2370         (md->ctypes[c] & ctype_space) == 0         (md->ctypes[c] & ctype_space) == 0
2371         )         )
# Line 2279  for (;;) Line 2381  for (;;)
2381        }        }
2382      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2383      if (      if (
2384  #ifdef SUPPORT_UTF8  #if defined SUPPORT_UTF || !(defined COMPILE_PCRE8)
2385         c < 256 &&         c < 256 &&
2386  #endif  #endif
2387         (md->ctypes[c] & ctype_word) != 0         (md->ctypes[c] & ctype_word) != 0
# Line 2296  for (;;) Line 2398  for (;;)
2398        }        }
2399      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2400      if (      if (
2401  #ifdef SUPPORT_UTF8  #if defined SUPPORT_UTF || !(defined COMPILE_PCRE8)
2402         c >= 256 ||         c > 255 ||
2403  #endif  #endif
2404         (md->ctypes[c] & ctype_word) == 0         (md->ctypes[c] & ctype_word) == 0
2405         )         )
# Line 2317  for (;;) Line 2419  for (;;)
2419        default: RRETURN(MATCH_NOMATCH);        default: RRETURN(MATCH_NOMATCH);
2420    
2421        case 0x000d:        case 0x000d:
2422        if (eptr < md->end_subject && *eptr == 0x0a) eptr++;        if (eptr >= md->end_subject)
2423            {
2424            SCHECK_PARTIAL();
2425            }
2426          else if (*eptr == 0x0a) eptr++;
2427        break;        break;
2428    
2429        case 0x000a:        case 0x000a:
# Line 2475  for (;;) Line 2581  for (;;)
2581          break;          break;
2582    
2583          case PT_GC:          case PT_GC:
2584          if ((ecode[2] != _pcre_ucp_gentype[prop->chartype]) == (op == OP_PROP))          if ((ecode[2] != PRIV(ucp_gentype)[prop->chartype]) == (op == OP_PROP))
2585            RRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
2586          break;          break;
2587    
# Line 2492  for (;;) Line 2598  for (;;)
2598          /* These are specials */          /* These are specials */
2599    
2600          case PT_ALNUM:          case PT_ALNUM:
2601          if ((_pcre_ucp_gentype[prop->chartype] == ucp_L ||          if ((PRIV(ucp_gentype)[prop->chartype] == ucp_L ||
2602               _pcre_ucp_gentype[prop->chartype] == ucp_N) == (op == OP_NOTPROP))               PRIV(ucp_gentype)[prop->chartype] == ucp_N) == (op == OP_NOTPROP))
2603            RRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
2604          break;          break;
2605    
2606          case PT_SPACE:    /* Perl space */          case PT_SPACE:    /* Perl space */
2607          if ((_pcre_ucp_gentype[prop->chartype] == ucp_Z ||          if ((PRIV(ucp_gentype)[prop->chartype] == ucp_Z ||
2608               c == CHAR_HT || c == CHAR_NL || c == CHAR_FF || c == CHAR_CR)               c == CHAR_HT || c == CHAR_NL || c == CHAR_FF || c == CHAR_CR)
2609                 == (op == OP_NOTPROP))                 == (op == OP_NOTPROP))
2610            RRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
2611          break;          break;
2612    
2613          case PT_PXSPACE:  /* POSIX space */          case PT_PXSPACE:  /* POSIX space */
2614          if ((_pcre_ucp_gentype[prop->chartype] == ucp_Z ||          if ((PRIV(ucp_gentype)[prop->chartype] == ucp_Z ||
2615               c == CHAR_HT || c == CHAR_NL || c == CHAR_VT ||               c == CHAR_HT || c == CHAR_NL || c == CHAR_VT ||
2616               c == CHAR_FF || c == CHAR_CR)               c == CHAR_FF || c == CHAR_CR)
2617                 == (op == OP_NOTPROP))                 == (op == OP_NOTPROP))
# Line 2513  for (;;) Line 2619  for (;;)
2619          break;          break;
2620    
2621          case PT_WORD:          case PT_WORD:
2622          if ((_pcre_ucp_gentype[prop->chartype] == ucp_L ||          if ((PRIV(ucp_gentype)[prop->chartype] == ucp_L ||
2623               _pcre_ucp_gentype[prop->chartype] == ucp_N ||               PRIV(ucp_gentype)[prop->chartype] == ucp_N ||
2624               c == CHAR_UNDERSCORE) == (op == OP_NOTPROP))               c == CHAR_UNDERSCORE) == (op == OP_NOTPROP))
2625            RRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
2626          break;          break;
# Line 2543  for (;;) Line 2649  for (;;)
2649      while (eptr < md->end_subject)      while (eptr < md->end_subject)
2650        {        {
2651        int len = 1;        int len = 1;
2652        if (!utf8) c = *eptr; else { GETCHARLEN(c, eptr, len); }        if (!utf) c = *eptr; else { GETCHARLEN(c, eptr, len); }
2653        if (UCD_CATEGORY(c) != ucp_M) break;        if (UCD_CATEGORY(c) != ucp_M) break;
2654        eptr += len;        eptr += len;
2655        }        }
2656        CHECK_PARTIAL();
2657      ecode++;      ecode++;
2658      break;      break;
2659  #endif  #endif
# Line 2564  for (;;) Line 2671  for (;;)
2671      case OP_REFI:      case OP_REFI:
2672      caseless = op == OP_REFI;      caseless = op == OP_REFI;
2673      offset = GET2(ecode, 1) << 1;               /* Doubled ref number */      offset = GET2(ecode, 1) << 1;               /* Doubled ref number */
2674      ecode += 3;      ecode += 1 + IMM2_SIZE;
2675    
2676      /* If the reference is unset, there are two possibilities:      /* If the reference is unset, there are two possibilities:
2677    
# Line 2604  for (;;) Line 2711  for (;;)
2711        case OP_CRMINRANGE:        case OP_CRMINRANGE:
2712        minimize = (*ecode == OP_CRMINRANGE);        minimize = (*ecode == OP_CRMINRANGE);
2713        min = GET2(ecode, 1);        min = GET2(ecode, 1);
2714        max = GET2(ecode, 3);        max = GET2(ecode, 1 + IMM2_SIZE);
2715        if (max == 0) max = INT_MAX;        if (max == 0) max = INT_MAX;
2716        ecode += 5;        ecode += 1 + 2 * IMM2_SIZE;
2717        break;        break;
2718    
2719        default:               /* No repeat follows */        default:               /* No repeat follows */
2720        if ((length = match_ref(offset, eptr, length, md, caseless)) < 0)        if ((length = match_ref(offset, eptr, length, md, caseless)) < 0)
2721          {          {
2722            if (length == -2) eptr = md->end_subject;   /* Partial match */
2723          CHECK_PARTIAL();          CHECK_PARTIAL();
2724          RRETURN(MATCH_NOMATCH);          RRETURN(MATCH_NOMATCH);
2725          }          }
# Line 2620  for (;;) Line 2728  for (;;)
2728        }        }
2729    
2730      /* Handle repeated back references. If the length of the reference is      /* Handle repeated back references. If the length of the reference is
2731      zero, just continue with the main loop. */      zero, just continue with the main loop. If the length is negative, it
2732        means the reference is unset in non-Java-compatible mode. If the minimum is
2733        zero, we can continue at the same level without recursion. For any other
2734        minimum, carrying on will result in NOMATCH. */
2735    
2736      if (length == 0) continue;      if (length == 0) continue;
2737        if (length < 0 && min == 0) continue;
2738    
2739      /* First, ensure the minimum number of matches are present. We get back      /* First, ensure the minimum number of matches are present. We get back
2740      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 2745  for (;;)
2745        int slength;        int slength;
2746        if ((slength = match_ref(offset, eptr, length, md, caseless)) < 0)        if ((slength = match_ref(offset, eptr, length, md, caseless)) < 0)
2747          {          {
2748            if (slength == -2) eptr = md->end_subject;   /* Partial match */
2749          CHECK_PARTIAL();          CHECK_PARTIAL();
2750          RRETURN(MATCH_NOMATCH);          RRETURN(MATCH_NOMATCH);
2751          }          }
# Line 2656  for (;;) Line 2769  for (;;)
2769          if (fi >= max) RRETURN(MATCH_NOMATCH);          if (fi >= max) RRETURN(MATCH_NOMATCH);
2770          if ((slength = match_ref(offset, eptr, length, md, caseless)) < 0)          if ((slength = match_ref(offset, eptr, length, md, caseless)) < 0)
2771            {            {
2772              if (slength == -2) eptr = md->end_subject;   /* Partial match */
2773            CHECK_PARTIAL();            CHECK_PARTIAL();
2774            RRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
2775            }            }
# Line 2674  for (;;) Line 2788  for (;;)
2788          int slength;          int slength;
2789          if ((slength = match_ref(offset, eptr, length, md, caseless)) < 0)          if ((slength = match_ref(offset, eptr, length, md, caseless)) < 0)
2790            {            {
2791            CHECK_PARTIAL();            /* Can't use CHECK_PARTIAL because we don't want to update eptr in
2792              the soft partial matching case. */
2793    
2794              if (slength == -2 && md->partial != 0 &&
2795                  md->end_subject > md->start_used_ptr)
2796                {
2797                md->hitend = TRUE;
2798                if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
2799                }
2800            break;            break;
2801            }            }
2802          eptr += slength;          eptr += slength;
2803          }          }
2804    
2805        while (eptr >= pp)        while (eptr >= pp)
2806          {          {
2807          RMATCH(eptr, ecode, offset_top, md, eptrb, RM15);          RMATCH(eptr, ecode, offset_top, md, eptrb, RM15);
# Line 2703  for (;;) Line 2826  for (;;)
2826      case OP_NCLASS:      case OP_NCLASS:
2827      case OP_CLASS:      case OP_CLASS:
2828        {        {
2829          /* The data variable is saved across frames, so the byte map needs to
2830          be stored there. */
2831    #define BYTE_MAP ((pcre_uint8 *)data)
2832        data = ecode + 1;                /* Save for matching */        data = ecode + 1;                /* Save for matching */
2833        ecode += 33;                     /* Advance past the item */        ecode += 1 + (32 / sizeof(pcre_uchar)); /* Advance past the item */
2834    
2835        switch (*ecode)        switch (*ecode)
2836          {          {
# Line 2725  for (;;) Line 2851  for (;;)
2851          case OP_CRMINRANGE:          case OP_CRMINRANGE:
2852          minimize = (*ecode == OP_CRMINRANGE);          minimize = (*ecode == OP_CRMINRANGE);
2853          min = GET2(ecode, 1);          min = GET2(ecode, 1);
2854          max = GET2(ecode, 3);          max = GET2(ecode, 1 + IMM2_SIZE);
2855          if (max == 0) max = INT_MAX;          if (max == 0) max = INT_MAX;
2856          ecode += 5;          ecode += 1 + 2 * IMM2_SIZE;
2857          break;          break;
2858    
2859          default:               /* No repeat follows */          default:               /* No repeat follows */
# Line 2737  for (;;) Line 2863  for (;;)
2863    
2864        /* First, ensure the minimum number of matches are present. */        /* First, ensure the minimum number of matches are present. */
2865    
2866  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
2867        /* UTF-8 mode */        if (utf)
       if (utf8)  
2868          {          {
2869          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
2870            {            {
# Line 2754  for (;;) Line 2879  for (;;)
2879              if (op == OP_CLASS) RRETURN(MATCH_NOMATCH);              if (op == OP_CLASS) RRETURN(MATCH_NOMATCH);
2880              }              }
2881            else            else
2882              {              if ((BYTE_MAP[c/8] & (1 << (c&7))) == 0) RRETURN(MATCH_NOMATCH);
             if ((data[c/8] & (1 << (c&7))) == 0) RRETURN(MATCH_NOMATCH);  
             }  
2883            }            }
2884          }          }
2885        else        else
2886  #endif  #endif
2887        /* Not UTF-8 mode */        /* Not UTF mode */
2888          {          {
2889          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
2890            {            {
# Line 2771  for (;;) Line 2894  for (;;)
2894              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
2895              }              }
2896            c = *eptr++;            c = *eptr++;
2897            if ((data[c/8] & (1 << (c&7))) == 0) RRETURN(MATCH_NOMATCH);  #ifndef COMPILE_PCRE8
2898              if (c > 255)
2899                {
2900                if (op == OP_CLASS) RRETURN(MATCH_NOMATCH);
2901                }
2902              else
2903    #endif
2904                if ((BYTE_MAP[c/8] & (1 << (c&7))) == 0) RRETURN(MATCH_NOMATCH);
2905            }            }
2906          }          }
2907    
# Line 2785  for (;;) Line 2915  for (;;)
2915    
2916        if (minimize)        if (minimize)
2917          {          {
2918  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
2919          /* UTF-8 mode */          if (utf)
         if (utf8)  
2920            {            {
2921            for (fi = min;; fi++)            for (fi = min;; fi++)
2922              {              {
# Line 2805  for (;;) Line 2934  for (;;)
2934                if (op == OP_CLASS) RRETURN(MATCH_NOMATCH);                if (op == OP_CLASS) RRETURN(MATCH_NOMATCH);
2935                }                }
2936              else              else
2937                {                if ((BYTE_MAP[c/8] & (1 << (c&7))) == 0) RRETURN(MATCH_NOMATCH);
               if ((data[c/8] & (1 << (c&7))) == 0) RRETURN(MATCH_NOMATCH);  
               }  
2938              }              }
2939            }            }
2940          else          else
2941  #endif  #endif
2942          /* Not UTF-8 mode */          /* Not UTF mode */
2943            {            {
2944            for (fi = min;; fi++)            for (fi = min;; fi++)
2945              {              {
# Line 2825  for (;;) Line 2952  for (;;)
2952                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
2953                }                }
2954              c = *eptr++;              c = *eptr++;
2955              if ((data[c/8] & (1 << (c&7))) == 0) RRETURN(MATCH_NOMATCH);  #ifndef COMPILE_PCRE8
2956                if (c > 255)
2957                  {
2958                  if (op == OP_CLASS) RRETURN(MATCH_NOMATCH);
2959                  }
2960                else
2961    #endif
2962                  if ((BYTE_MAP[c/8] & (1 << (c&7))) == 0) RRETURN(MATCH_NOMATCH);
2963              }              }
2964            }            }
2965          /* Control never gets here */          /* Control never gets here */
# Line 2837  for (;;) Line 2971  for (;;)
2971          {          {
2972          pp = eptr;          pp = eptr;
2973    
2974  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
2975          /* UTF-8 mode */          if (utf)
         if (utf8)  
2976            {            {
2977            for (i = min; i < max; i++)            for (i = min; i < max; i++)
2978              {              {
# Line 2855  for (;;) Line 2988  for (;;)
2988                if (op == OP_CLASS) break;                if (op == OP_CLASS) break;
2989                }                }
2990              else              else
2991                {                if ((BYTE_MAP[c/8] & (1 << (c&7))) == 0) break;
               if ((data[c/8] & (1 << (c&7))) == 0) break;  
               }  
2992              eptr += len;              eptr += len;
2993              }              }
2994            for (;;)            for (;;)
# Line 2870  for (;;) Line 3001  for (;;)
3001            }            }
3002          else          else
3003  #endif  #endif
3004            /* Not UTF-8 mode */            /* Not UTF mode */
3005            {            {
3006            for (i = min; i < max; i++)            for (i = min; i < max; i++)
3007              {              {
# Line 2880  for (;;) Line 3011  for (;;)
3011                break;                break;
3012                }                }
3013              c = *eptr;              c = *eptr;
3014              if ((data[c/8] & (1 << (c&7))) == 0) break;  #ifndef COMPILE_PCRE8
3015                if (c > 255)
3016                  {
3017                  if (op == OP_CLASS) break;
3018                  }
3019                else
3020    #endif
3021                  if ((BYTE_MAP[c/8] & (1 << (c&7))) == 0) break;
3022              eptr++;              eptr++;
3023              }              }
3024            while (eptr >= pp)            while (eptr >= pp)
# Line 2893  for (;;) Line 3031  for (;;)
3031    
3032          RRETURN(MATCH_NOMATCH);          RRETURN(MATCH_NOMATCH);
3033          }          }
3034    #undef BYTE_MAP
3035        }        }
3036      /* Control never gets here */      /* Control never gets here */
3037    
# Line 2901  for (;;) Line 3040  for (;;)
3040      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
3041      mode, because Unicode properties are supported in non-UTF-8 mode. */      mode, because Unicode properties are supported in non-UTF-8 mode. */
3042    
3043  #ifdef SUPPORT_UTF8  #if defined SUPPORT_UTF || !defined COMPILE_PCRE8
3044      case OP_XCLASS:      case OP_XCLASS:
3045        {        {
3046        data = ecode + 1 + LINK_SIZE;                /* Save for matching */        data = ecode + 1 + LINK_SIZE;                /* Save for matching */
# Line 2926  for (;;) Line 3065  for (;;)
3065          case OP_CRMINRANGE:          case OP_CRMINRANGE:
3066          minimize = (*ecode == OP_CRMINRANGE);          minimize = (*ecode == OP_CRMINRANGE);
3067          min = GET2(ecode, 1);          min = GET2(ecode, 1);
3068          max = GET2(ecode, 3);          max = GET2(ecode, 1 + IMM2_SIZE);
3069          if (max == 0) max = INT_MAX;          if (max == 0) max = INT_MAX;
3070          ecode += 5;          ecode += 1 + 2 * IMM2_SIZE;
3071          break;          break;
3072    
3073          default:               /* No repeat follows */          default:               /* No repeat follows */
# Line 2946  for (;;) Line 3085  for (;;)
3085            RRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
3086            }            }
3087          GETCHARINCTEST(c, eptr);          GETCHARINCTEST(c, eptr);
3088          if (!_pcre_xclass(c, data)) RRETURN(MATCH_NOMATCH);          if (!PRIV(xclass)(c, data, utf)) RRETURN(MATCH_NOMATCH);
3089          }          }
3090    
3091        /* 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 3109  for (;;)
3109              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
3110              }              }
3111            GETCHARINCTEST(c, eptr);            GETCHARINCTEST(c, eptr);
3112            if (!_pcre_xclass(c, data)) RRETURN(MATCH_NOMATCH);            if (!PRIV(xclass)(c, data, utf)) RRETURN(MATCH_NOMATCH);
3113            }            }
3114          /* Control never gets here */          /* Control never gets here */
3115          }          }
# Line 2988  for (;;) Line 3127  for (;;)
3127              SCHECK_PARTIAL();              SCHECK_PARTIAL();
3128              break;              break;
3129              }              }
3130    #ifdef SUPPORT_UTF
3131            GETCHARLENTEST(c, eptr, len);            GETCHARLENTEST(c, eptr, len);
3132            if (!_pcre_xclass(c, data)) break;  #else
3133              c = *eptr;
3134    #endif
3135              if (!PRIV(xclass)(c, data, utf)) break;
3136            eptr += len;            eptr += len;
3137            }            }
3138          for(;;)          for(;;)
# Line 2997  for (;;) Line 3140  for (;;)
3140            RMATCH(eptr, ecode, offset_top, md, eptrb, RM21);            RMATCH(eptr, ecode, offset_top, md, eptrb, RM21);
3141            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3142            if (eptr-- == pp) break;        /* Stop if tried at original pos */            if (eptr-- == pp) break;        /* Stop if tried at original pos */
3143            if (utf8) BACKCHAR(eptr);  #ifdef SUPPORT_UTF
3144              if (utf) BACKCHAR(eptr);
3145    #endif
3146            }            }
3147          RRETURN(MATCH_NOMATCH);          RRETURN(MATCH_NOMATCH);
3148          }          }
# Line 3009  for (;;) Line 3154  for (;;)
3154      /* Match a single character, casefully */      /* Match a single character, casefully */
3155    
3156      case OP_CHAR:      case OP_CHAR:
3157  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
3158      if (utf8)      if (utf)
3159        {        {
3160        length = 1;        length = 1;
3161        ecode++;        ecode++;
# Line 3024  for (;;) Line 3169  for (;;)
3169        }        }
3170      else      else
3171  #endif  #endif
3172        /* Not UTF mode */
     /* Non-UTF-8 mode */  
3173        {        {
3174        if (md->end_subject - eptr < 1)        if (md->end_subject - eptr < 1)
3175          {          {
# Line 3037  for (;;) Line 3181  for (;;)
3181        }        }
3182      break;      break;
3183    
3184      /* Match a single character, caselessly */      /* Match a single character, caselessly. If we are at the end of the
3185        subject, give up immediately. */
3186    
3187      case OP_CHARI:      case OP_CHARI:
3188  #ifdef SUPPORT_UTF8      if (eptr >= md->end_subject)
3189      if (utf8)        {
3190          SCHECK_PARTIAL();
3191          RRETURN(MATCH_NOMATCH);
3192          }
3193    
3194    #ifdef SUPPORT_UTF
3195        if (utf)
3196        {        {
3197        length = 1;        length = 1;
3198        ecode++;        ecode++;
3199        GETCHARLEN(fc, ecode, length);        GETCHARLEN(fc, ecode, length);
3200    
       if (length > md->end_subject - eptr)  
         {  
         CHECK_PARTIAL();             /* Not SCHECK_PARTIAL() */  
         RRETURN(MATCH_NOMATCH);  
         }  
   
3201        /* 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
3202        can use the fast lookup table. */        we know that its other case must also be one byte long, so we can use the
3203          fast lookup table. We know that there is at least one byte left in the
3204          subject. */
3205    
3206        if (fc < 128)        if (fc < 128)
3207          {          {
3208          if (md->lcc[*ecode++] != md->lcc[*eptr++]) RRETURN(MATCH_NOMATCH);          if (md->lcc[fc]
3209                != TABLE_GET(*eptr, md->lcc, *eptr)) RRETURN(MATCH_NOMATCH);
3210            ecode++;
3211            eptr++;
3212          }          }
3213    
3214        /* Otherwise we must pick up the subject character */        /* Otherwise we must pick up the subject character. Note that we cannot
3215          use the value of "length" to check for sufficient bytes left, because the
3216          other case of the character may have more or fewer bytes.  */
3217    
3218        else        else
3219          {          {
# Line 3082  for (;;) Line 3234  for (;;)
3234          }          }
3235        }        }
3236      else      else
3237  #endif   /* SUPPORT_UTF8 */  #endif   /* SUPPORT_UTF */
3238    
3239      /* Non-UTF-8 mode */      /* Not UTF mode */
3240        {        {
3241        if (md->end_subject - eptr < 1)        if (TABLE_GET(ecode[1], md->lcc, ecode[1])
3242          {            != TABLE_GET(*eptr, md->lcc, *eptr)) RRETURN(MATCH_NOMATCH);
3243          SCHECK_PARTIAL();            /* This one can use SCHECK_PARTIAL() */        eptr++;
         RRETURN(MATCH_NOMATCH);  
         }  
       if (md->lcc[ecode[1]] != md->lcc[*eptr++]) RRETURN(MATCH_NOMATCH);  
3244        ecode += 2;        ecode += 2;
3245        }        }
3246      break;      break;
# Line 3101  for (;;) Line 3250  for (;;)
3250      case OP_EXACT:      case OP_EXACT:
3251      case OP_EXACTI:      case OP_EXACTI:
3252      min = max = GET2(ecode, 1);      min = max = GET2(ecode, 1);
3253      ecode += 3;      ecode += 1 + IMM2_SIZE;
3254      goto REPEATCHAR;      goto REPEATCHAR;
3255    
3256      case OP_POSUPTO:      case OP_POSUPTO:
# Line 3116  for (;;) Line 3265  for (;;)
3265      min = 0;      min = 0;
3266      max = GET2(ecode, 1);      max = GET2(ecode, 1);
3267      minimize = *ecode == OP_MINUPTO || *ecode == OP_MINUPTOI;      minimize = *ecode == OP_MINUPTO || *ecode == OP_MINUPTOI;
3268      ecode += 3;      ecode += 1 + IMM2_SIZE;
3269      goto REPEATCHAR;      goto REPEATCHAR;
3270    
3271      case OP_POSSTAR:      case OP_POSSTAR:
# Line 3164  for (;;) Line 3313  for (;;)
3313      /* Common code for all repeated single-character matches. */      /* Common code for all repeated single-character matches. */
3314    
3315      REPEATCHAR:      REPEATCHAR:
3316  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
3317      if (utf8)      if (utf)
3318        {        {
3319        length = 1;        length = 1;
3320        charptr = ecode;        charptr = ecode;
# Line 3181  for (;;) Line 3330  for (;;)
3330          unsigned int othercase;          unsigned int othercase;
3331          if (op >= OP_STARI &&     /* Caseless */          if (op >= OP_STARI &&     /* Caseless */
3332              (othercase = UCD_OTHERCASE(fc)) != fc)              (othercase = UCD_OTHERCASE(fc)) != fc)
3333            oclength = _pcre_ord2utf8(othercase, occhars);            oclength = PRIV(ord2utf)(othercase, occhars);
3334          else oclength = 0;          else oclength = 0;
3335  #endif  /* SUPPORT_UCP */  #endif  /* SUPPORT_UCP */
3336    
3337          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
3338            {            {
3339            if (eptr <= md->end_subject - length &&            if (eptr <= md->end_subject - length &&
3340              memcmp(eptr, charptr, length) == 0) eptr += length;              memcmp(eptr, charptr, IN_UCHARS(length)) == 0) eptr += length;
3341  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
3342            else if (oclength > 0 &&            else if (oclength > 0 &&
3343                     eptr <= md->end_subject - oclength &&                     eptr <= md->end_subject - oclength &&
3344                     memcmp(eptr, occhars, oclength) == 0) eptr += oclength;                     memcmp(eptr, occhars, IN_UCHARS(oclength)) == 0) eptr += oclength;
3345  #endif  /* SUPPORT_UCP */  #endif  /* SUPPORT_UCP */
3346            else            else
3347              {              {
# Line 3211  for (;;) Line 3360  for (;;)
3360              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3361              if (fi >= max) RRETURN(MATCH_NOMATCH);              if (fi >= max) RRETURN(MATCH_NOMATCH);
3362              if (eptr <= md->end_subject - length &&              if (eptr <= md->end_subject - length &&
3363                memcmp(eptr, charptr, length) == 0) eptr += length;                memcmp(eptr, charptr, IN_UCHARS(length)) == 0) eptr += length;
3364  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
3365              else if (oclength > 0 &&              else if (oclength > 0 &&
3366                       eptr <= md->end_subject - oclength &&                       eptr <= md->end_subject - oclength &&
3367                       memcmp(eptr, occhars, oclength) == 0) eptr += oclength;                       memcmp(eptr, occhars, IN_UCHARS(oclength)) == 0) eptr += oclength;
3368  #endif  /* SUPPORT_UCP */  #endif  /* SUPPORT_UCP */
3369              else              else
3370                {                {
# Line 3232  for (;;) Line 3381  for (;;)
3381            for (i = min; i < max; i++)            for (i = min; i < max; i++)
3382              {              {
3383              if (eptr <= md->end_subject - length &&              if (eptr <= md->end_subject - length &&
3384                  memcmp(eptr, charptr, length) == 0) eptr += length;                  memcmp(eptr, charptr, IN_UCHARS(length)) == 0) eptr += length;
3385  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
3386              else if (oclength > 0 &&              else if (oclength > 0 &&
3387                       eptr <= md->end_subject - oclength &&                       eptr <= md->end_subject - oclength &&
3388                       memcmp(eptr, occhars, oclength) == 0) eptr += oclength;                       memcmp(eptr, occhars, IN_UCHARS(oclength)) == 0) eptr += oclength;
3389  #endif  /* SUPPORT_UCP */  #endif  /* SUPPORT_UCP */
3390              else              else
3391                {                {
# Line 3268  for (;;) Line 3417  for (;;)
3417        value of fc will always be < 128. */        value of fc will always be < 128. */
3418        }        }
3419      else      else
3420  #endif  /* SUPPORT_UTF8 */  #endif  /* SUPPORT_UTF */
3421          /* When not in UTF-8 mode, load a single-byte character. */
3422      /* When not in UTF-8 mode, load a single-byte character. */        fc = *ecode++;
3423    
3424      fc = *ecode++;      /* The value of fc at this point is always one character, though we may
3425        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  
3426      caseful cases, for speed, since matching characters is likely to be quite      caseful cases, for speed, since matching characters is likely to be quite
3427      common. First, ensure the minimum number of matches are present. If min =      common. First, ensure the minimum number of matches are present. If min =
3428      max, continue at the same level without recursing. Otherwise, if      max, continue at the same level without recursing. Otherwise, if
# Line 3288  for (;;) Line 3435  for (;;)
3435    
3436      if (op >= OP_STARI)  /* Caseless */      if (op >= OP_STARI)  /* Caseless */
3437        {        {
3438        fc = md->lcc[fc];  #ifdef COMPILE_PCRE8
3439          /* fc must be < 128 if UTF is enabled. */
3440          foc = md->fcc[fc];
3441    #else
3442    #ifdef SUPPORT_UTF
3443    #ifdef SUPPORT_UCP
3444          if (utf && fc > 127)
3445            foc = UCD_OTHERCASE(fc);
3446    #else
3447          if (utf && fc > 127)
3448            foc = fc;
3449    #endif /* SUPPORT_UCP */
3450          else
3451    #endif /* SUPPORT_UTF */
3452            foc = TABLE_GET(fc, md->fcc, fc);
3453    #endif /* COMPILE_PCRE8 */
3454    
3455        for (i = 1; i <= min; i++)        for (i = 1; i <= min; i++)
3456          {          {
3457          if (eptr >= md->end_subject)          if (eptr >= md->end_subject)
# Line 3296  for (;;) Line 3459  for (;;)
3459            SCHECK_PARTIAL();            SCHECK_PARTIAL();
3460            RRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
3461            }            }
3462          if (fc != md->lcc[*eptr++]) RRETURN(MATCH_NOMATCH);          if (fc != *eptr && foc != *eptr) RRETURN(MATCH_NOMATCH);
3463            eptr++;
3464          }          }
3465        if (min == max) continue;        if (min == max) continue;
3466        if (minimize)        if (minimize)
# Line 3311  for (;;) Line 3475  for (;;)
3475              SCHECK_PARTIAL();              SCHECK_PARTIAL();
3476              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
3477              }              }
3478            if (fc != md->lcc[*eptr++]) RRETURN(MATCH_NOMATCH);            if (fc != *eptr && foc != *eptr) RRETURN(MATCH_NOMATCH);
3479              eptr++;
3480            }            }
3481          /* Control never gets here */          /* Control never gets here */
3482          }          }
# Line 3325  for (;;) Line 3490  for (;;)
3490              SCHECK_PARTIAL();              SCHECK_PARTIAL();
3491              break;              break;
3492              }              }
3493            if (fc != md->lcc[*eptr]) break;            if (fc != *eptr && foc != *eptr) break;
3494            eptr++;            eptr++;
3495            }            }
3496    
# Line 3410  for (;;) Line 3575  for (;;)
3575        SCHECK_PARTIAL();        SCHECK_PARTIAL();
3576        RRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
3577        }        }
3578      ecode++;  #ifdef SUPPORT_UTF
3579      GETCHARINCTEST(c, eptr);      if (utf)
     if (op == OP_NOTI)         /* The caseless case */  
3580        {        {
3581  #ifdef SUPPORT_UTF8        register unsigned int ch, och;
3582        if (c < 256)  
3583  #endif        ecode++;
3584        c = md->lcc[c];        GETCHARINC(ch, ecode);
3585        if (md->lcc[*ecode++] == c) RRETURN(MATCH_NOMATCH);        GETCHARINC(c, eptr);
3586    
3587          if (op == OP_NOT)
3588            {
3589            if (ch == c) RRETURN(MATCH_NOMATCH);
3590            }
3591          else
3592            {
3593    #ifdef SUPPORT_UCP
3594            if (ch > 127)
3595              och = UCD_OTHERCASE(ch);
3596    #else
3597            if (ch > 127)
3598              och = ch;
3599    #endif /* SUPPORT_UCP */
3600            else
3601              och = TABLE_GET(ch, md->fcc, ch);
3602            if (ch == c || och == c) RRETURN(MATCH_NOMATCH);
3603            }
3604        }        }
3605      else    /* Caseful */      else
3606    #endif
3607        {        {
3608        if (*ecode++ == c) RRETURN(MATCH_NOMATCH);        register unsigned int ch = ecode[1];
3609          c = *eptr++;
3610          if (ch == c || (op == OP_NOTI && TABLE_GET(ch, md->fcc, ch) == c))
3611            RRETURN(MATCH_NOMATCH);
3612          ecode += 2;
3613        }        }
3614      break;      break;
3615    
# Line 3436  for (;;) Line 3623  for (;;)
3623      case OP_NOTEXACT:      case OP_NOTEXACT:
3624      case OP_NOTEXACTI:      case OP_NOTEXACTI:
3625      min = max = GET2(ecode, 1);      min = max = GET2(ecode, 1);
3626      ecode += 3;      ecode += 1 + IMM2_SIZE;
3627      goto REPEATNOTCHAR;      goto REPEATNOTCHAR;
3628    
3629      case OP_NOTUPTO:      case OP_NOTUPTO:
# Line 3446  for (;;) Line 3633  for (;;)
3633      min = 0;      min = 0;
3634      max = GET2(ecode, 1);      max = GET2(ecode, 1);
3635      minimize = *ecode == OP_NOTMINUPTO || *ecode == OP_NOTMINUPTOI;      minimize = *ecode == OP_NOTMINUPTO || *ecode == OP_NOTMINUPTOI;
3636      ecode += 3;      ecode += 1 + IMM2_SIZE;
3637      goto REPEATNOTCHAR;      goto REPEATNOTCHAR;
3638    
3639      case OP_NOTPOSSTAR:      case OP_NOTPOSSTAR:
# Line 3478  for (;;) Line 3665  for (;;)
3665      possessive = TRUE;      possessive = TRUE;
3666      min = 0;      min = 0;
3667      max = GET2(ecode, 1);      max = GET2(ecode, 1);
3668      ecode += 3;      ecode += 1 + IMM2_SIZE;
3669      goto REPEATNOTCHAR;      goto REPEATNOTCHAR;
3670    
3671      case OP_NOTSTAR:      case OP_NOTSTAR:
# Line 3502  for (;;) Line 3689  for (;;)
3689      /* Common code for all repeated single-byte matches. */      /* Common code for all repeated single-byte matches. */
3690    
3691      REPEATNOTCHAR:      REPEATNOTCHAR:
3692      fc = *ecode++;      GETCHARINCTEST(fc, ecode);
3693    
3694      /* The code is duplicated for the caseless and caseful cases, for speed,      /* The code is duplicated for the caseless and caseful cases, for speed,
3695      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 3517  for (;;) Line 3704  for (;;)
3704    
3705      if (op >= OP_NOTSTARI)     /* Caseless */      if (op >= OP_NOTSTARI)     /* Caseless */
3706        {        {
3707        fc = md->lcc[fc];  #ifdef SUPPORT_UTF
3708    #ifdef SUPPORT_UCP
3709          if (utf && fc > 127)
3710            foc = UCD_OTHERCASE(fc);
3711    #else
3712          if (utf && fc > 127)
3713            foc = fc;
3714    #endif /* SUPPORT_UCP */
3715          else
3716    #endif /* SUPPORT_UTF */
3717            foc = TABLE_GET(fc, md->fcc, fc);
3718    
3719  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
3720        /* UTF-8 mode */        if (utf)
       if (utf8)  
3721          {          {
3722          register unsigned int d;          register unsigned int d;
3723          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
# Line 3532  for (;;) Line 3728  for (;;)
3728              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
3729              }              }
3730            GETCHARINC(d, eptr);            GETCHARINC(d, eptr);
3731            if (d < 256) d = md->lcc[d];            if (fc == d || (unsigned int)foc == d) RRETURN(MATCH_NOMATCH);
           if (fc == d) RRETURN(MATCH_NOMATCH);  
3732            }            }
3733          }          }
3734        else        else
3735  #endif  #endif
3736          /* Not UTF mode */
       /* Not UTF-8 mode */  
3737          {          {
3738          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
3739            {            {
# Line 3548  for (;;) Line 3742  for (;;)
3742              SCHECK_PARTIAL();              SCHECK_PARTIAL();
3743              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
3744              }              }
3745            if (fc == md->lcc[*eptr++]) RRETURN(MATCH_NOMATCH);            if (fc == *eptr || foc == *eptr) RRETURN(MATCH_NOMATCH);
3746              eptr++;
3747            }            }
3748          }          }
3749    
# Line 3556  for (;;) Line 3751  for (;;)
3751    
3752        if (minimize)        if (minimize)
3753          {          {
3754  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
3755          /* UTF-8 mode */          if (utf)
         if (utf8)  
3756            {            {
3757            register unsigned int d;            register unsigned int d;
3758            for (fi = min;; fi++)            for (fi = min;; fi++)
# Line 3572  for (;;) Line 3766  for (;;)
3766                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
3767                }                }
3768              GETCHARINC(d, eptr);              GETCHARINC(d, eptr);
3769              if (d < 256) d = md->lcc[d];              if (fc == d || (unsigned int)foc == d) RRETURN(MATCH_NOMATCH);
             if (fc == d) RRETURN(MATCH_NOMATCH);  
3770              }              }
3771            }            }
3772          else          else
3773  #endif  #endif
3774          /* Not UTF-8 mode */          /* Not UTF mode */
3775            {            {
3776            for (fi = min;; fi++)            for (fi = min;; fi++)
3777              {              {
# Line 3590  for (;;) Line 3783  for (;;)
3783                SCHECK_PARTIAL();                SCHECK_PARTIAL();
3784                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
3785                }                }
3786              if (fc == md->lcc[*eptr++]) RRETURN(MATCH_NOMATCH);              if (fc == *eptr || foc == *eptr) RRETURN(MATCH_NOMATCH);
3787                eptr++;
3788              }              }
3789            }            }
3790          /* Control never gets here */          /* Control never gets here */
# Line 3602  for (;;) Line 3796  for (;;)
3796          {          {
3797          pp = eptr;          pp = eptr;
3798    
3799  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
3800          /* UTF-8 mode */          if (utf)
         if (utf8)  
3801            {            {
3802            register unsigned int d;            register unsigned int d;
3803            for (i = min; i < max; i++)            for (i = min; i < max; i++)
# Line 3616  for (;;) Line 3809  for (;;)
3809                break;                break;
3810                }                }
3811              GETCHARLEN(d, eptr, len);              GETCHARLEN(d, eptr, len);
3812              if (d < 256) d = md->lcc[d];              if (fc == d || (unsigned int)foc == d) break;
             if (fc == d) break;  
3813              eptr += len;              eptr += len;
3814              }              }
3815          if (possessive) continue;            if (possessive) continue;
3816          for(;;)            for(;;)
3817              {              {
3818              RMATCH(eptr, ecode, offset_top, md, eptrb, RM30);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM30);
3819              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
# Line 3631  for (;;) Line 3823  for (;;)
3823            }            }
3824          else          else
3825  #endif  #endif
3826          /* Not UTF-8 mode */          /* Not UTF mode */
3827            {            {
3828            for (i = min; i < max; i++)            for (i = min; i < max; i++)
3829              {              {
# Line 3640  for (;;) Line 3832  for (;;)
3832                SCHECK_PARTIAL();                SCHECK_PARTIAL();
3833                break;                break;
3834                }                }
3835              if (fc == md->lcc[*eptr]) break;              if (fc == *eptr || foc == *eptr) break;
3836              eptr++;              eptr++;
3837              }              }
3838            if (possessive) continue;            if (possessive) continue;
# Line 3661  for (;;) Line 3853  for (;;)
3853    
3854      else      else
3855        {        {
3856  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
3857        /* UTF-8 mode */        if (utf)
       if (utf8)  
3858          {          {
3859          register unsigned int d;          register unsigned int d;
3860          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
# Line 3679  for (;;) Line 3870  for (;;)
3870          }          }
3871        else        else
3872  #endif  #endif
3873        /* Not UTF-8 mode */        /* Not UTF mode */
3874          {          {
3875          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
3876            {            {
# Line 3696  for (;;) Line 3887  for (;;)
3887    
3888        if (minimize)        if (minimize)
3889          {          {
3890  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
3891          /* UTF-8 mode */          if (utf)
         if (utf8)  
3892            {            {
3893            register unsigned int d;            register unsigned int d;
3894            for (fi = min;; fi++)            for (fi = min;; fi++)
# Line 3717  for (;;) Line 3907  for (;;)
3907            }            }
3908          else          else
3909  #endif  #endif
3910          /* Not UTF-8 mode */          /* Not UTF mode */
3911            {            {
3912            for (fi = min;; fi++)            for (fi = min;; fi++)
3913              {              {
# Line 3741  for (;;) Line 3931  for (;;)
3931          {          {
3932          pp = eptr;          pp = eptr;
3933    
3934  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
3935          /* UTF-8 mode */          if (utf)
         if (utf8)  
3936            {            {
3937            register unsigned int d;            register unsigned int d;
3938            for (i = min; i < max; i++)            for (i = min; i < max; i++)
# Line 3769  for (;;) Line 3958  for (;;)
3958            }            }
3959          else          else
3960  #endif  #endif
3961          /* Not UTF-8 mode */          /* Not UTF mode */
3962            {            {
3963            for (i = min; i < max; i++)            for (i = min; i < max; i++)
3964              {              {
# Line 3802  for (;;) Line 3991  for (;;)
3991      case OP_TYPEEXACT:      case OP_TYPEEXACT:
3992      min = max = GET2(ecode, 1);      min = max = GET2(ecode, 1);
3993      minimize = TRUE;      minimize = TRUE;
3994      ecode += 3;      ecode += 1 + IMM2_SIZE;
3995      goto REPEATTYPE;      goto REPEATTYPE;
3996    
3997      case OP_TYPEUPTO:      case OP_TYPEUPTO:
# Line 3810  for (;;) Line 3999  for (;;)
3999      min = 0;      min = 0;
4000      max = GET2(ecode, 1);      max = GET2(ecode, 1);
4001      minimize = *ecode == OP_TYPEMINUPTO;      minimize = *ecode == OP_TYPEMINUPTO;
4002      ecode += 3;      ecode += 1 + IMM2_SIZE;
4003      goto REPEATTYPE;      goto REPEATTYPE;
4004    
4005      case OP_TYPEPOSSTAR:      case OP_TYPEPOSSTAR:
# Line 3838  for (;;) Line 4027  for (;;)
4027      possessive = TRUE;      possessive = TRUE;
4028      min = 0;      min = 0;
4029      max = GET2(ecode, 1);      max = GET2(ecode, 1);
4030      ecode += 3;      ecode += 1 + IMM2_SIZE;
4031      goto REPEATTYPE;      goto REPEATTYPE;
4032    
4033      case OP_TYPESTAR:      case OP_TYPESTAR:
# Line 4045  for (;;) Line 4234  for (;;)
4234            while (eptr < md->end_subject)            while (eptr < md->end_subject)
4235              {              {
4236              int len = 1;              int len = 1;
4237              if (!utf8) c = *eptr; else { GETCHARLEN(c, eptr, len); }              if (!utf) c = *eptr; else { GETCHARLEN(c, eptr, len); }
4238              if (UCD_CATEGORY(c) != ucp_M) break;              if (UCD_CATEGORY(c) != ucp_M) break;
4239              eptr += len;              eptr += len;
4240              }              }
4241              CHECK_PARTIAL();
4242            }            }
4243          }          }
4244    
# Line 4057  for (;;) Line 4247  for (;;)
4247    
4248  /* Handle all other cases when the coding is UTF-8 */  /* Handle all other cases when the coding is UTF-8 */
4249    
4250  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
4251        if (utf8) switch(ctype)        if (utf) switch(ctype)
4252          {          {
4253          case OP_ANY:          case OP_ANY:
4254          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
# Line 4069  for (;;) Line 4259  for (;;)
4259              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4260              }              }
4261            if (IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH);            if (IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH);
4262              if (md->partial != 0 &&
4263                  eptr + 1 >= md->end_subject &&
4264                  NLBLOCK->nltype == NLTYPE_FIXED &&
4265                  NLBLOCK->nllen == 2 &&
4266                  *eptr == NLBLOCK->nl[0])
4267                {
4268                md->hitend = TRUE;
4269                if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
4270                }
4271            eptr++;            eptr++;
4272            while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;            ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);
4273            }            }
4274          break;          break;
4275    
# Line 4083  for (;;) Line 4282  for (;;)
4282              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4283              }              }
4284            eptr++;            eptr++;
4285            while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;            ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);
4286            }            }
4287          break;          break;
4288    
# Line 4265  for (;;) Line 4464  for (;;)
4464              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4465              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4466              }              }
4467            if (*eptr >= 128 || (md->ctypes[*eptr++] & ctype_digit) == 0)            if (*eptr >= 128 || (md->ctypes[*eptr] & ctype_digit) == 0)
4468              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4469              eptr++;
4470            /* 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 */
4471            }            }
4472          break;          break;
# Line 4281  for (;;) Line 4481  for (;;)
4481              }              }
4482            if (*eptr < 128 && (md->ctypes[*eptr] & ctype_space) != 0)            if (*eptr < 128 && (md->ctypes[*eptr] & ctype_space) != 0)
4483              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4484            while (++eptr < md->end_subject && (*eptr & 0xc0) == 0x80);            eptr++;
4485              ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);
4486            }            }
4487          break;          break;
4488    
# Line 4293  for (;;) Line 4494  for (;;)
4494              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4495              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4496              }              }
4497            if (*eptr >= 128 || (md->ctypes[*eptr++] & ctype_space) == 0)            if (*eptr >= 128 || (md->ctypes[*eptr] & ctype_space) == 0)
4498              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4499              eptr++;
4500            /* 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 */
4501            }            }
4502          break;          break;
# Line 4309  for (;;) Line 4511  for (;;)
4511              }              }
4512            if (*eptr < 128 && (md->ctypes[*eptr] & ctype_word) != 0)            if (*eptr < 128 && (md->ctypes[*eptr] & ctype_word) != 0)
4513              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4514            while (++eptr < md->end_subject && (*eptr & 0xc0) == 0x80);            eptr++;
4515              ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);
4516            }            }
4517          break;          break;
4518    
# Line 4321  for (;;) Line 4524  for (;;)
4524              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4525              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4526              }              }
4527            if (*eptr >= 128 || (md->ctypes[*eptr++] & ctype_word) == 0)            if (*eptr >= 128 || (md->ctypes[*eptr] & ctype_word) == 0)
4528              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4529              eptr++;
4530            /* 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 */
4531            }            }
4532          break;          break;
# Line 4332  for (;;) Line 4536  for (;;)
4536          }  /* End switch(ctype) */          }  /* End switch(ctype) */
4537    
4538        else        else
4539  #endif     /* SUPPORT_UTF8 */  #endif     /* SUPPORT_UTF */
4540    
4541        /* 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
4542        than OP_PROP and OP_NOTPROP. */        than OP_PROP and OP_NOTPROP. */
# Line 4348  for (;;) Line 4552  for (;;)
4552              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4553              }              }
4554            if (IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH);            if (IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH);
4555              if (md->partial != 0 &&
4556                  eptr + 1 >= md->end_subject &&
4557                  NLBLOCK->nltype == NLTYPE_FIXED &&
4558                  NLBLOCK->nllen == 2 &&
4559                  *eptr == NLBLOCK->nl[0])
4560                {
4561                md->hitend = TRUE;
4562                if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
4563                }
4564            eptr++;            eptr++;
4565            }            }
4566          break;          break;
# Line 4392  for (;;) Line 4605  for (;;)
4605              case 0x000b:              case 0x000b:
4606              case 0x000c:              case 0x000c:
4607              case 0x0085:              case 0x0085:
4608    #ifdef COMPILE_PCRE16
4609                case 0x2028:
4610                case 0x2029:
4611    #endif
4612              if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);              if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);
4613              break;              break;
4614              }              }
# Line 4412  for (;;) Line 4629  for (;;)
4629              case 0x09:      /* HT */              case 0x09:      /* HT */
4630              case 0x20:      /* SPACE */              case 0x20:      /* SPACE */
4631              case 0xa0:      /* NBSP */              case 0xa0:      /* NBSP */
4632    #ifdef COMPILE_PCRE16
4633                case 0x1680:    /* OGHAM SPACE MARK */
4634                case 0x180e:    /* MONGOLIAN VOWEL SEPARATOR */
4635                case 0x2000:    /* EN QUAD */
4636                case 0x2001:    /* EM QUAD */
4637                case 0x2002:    /* EN SPACE */
4638                case 0x2003:    /* EM SPACE */
4639                case 0x2004:    /* THREE-PER-EM SPACE */
4640                case 0x2005:    /* FOUR-PER-EM SPACE */
4641                case 0x2006:    /* SIX-PER-EM SPACE */
4642                case 0x2007:    /* FIGURE SPACE */
4643                case 0x2008:    /* PUNCTUATION SPACE */
4644                case 0x2009:    /* THIN SPACE */
4645                case 0x200A:    /* HAIR SPACE */
4646                case 0x202f:    /* NARROW NO-BREAK SPACE */
4647                case 0x205f:    /* MEDIUM MATHEMATICAL SPACE */
4648                case 0x3000:    /* IDEOGRAPHIC SPACE */
4649    #endif
4650              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4651              }              }
4652            }            }
# Line 4431  for (;;) Line 4666  for (;;)
4666              case 0x09:      /* HT */              case 0x09:      /* HT */
4667              case 0x20:      /* SPACE */              case 0x20:      /* SPACE */
4668              case 0xa0:      /* NBSP */              case 0xa0:      /* NBSP */
4669    #ifdef COMPILE_PCRE16
4670                case 0x1680:    /* OGHAM SPACE MARK */
4671                case 0x180e:    /* MONGOLIAN VOWEL SEPARATOR */
4672                case 0x2000:    /* EN QUAD */
4673                case 0x2001:    /* EM QUAD */
4674                case 0x2002:    /* EN SPACE */
4675                case 0x2003:    /* EM SPACE */
4676                case 0x2004:    /* THREE-PER-EM SPACE */
4677                case 0x2005:    /* FOUR-PER-EM SPACE */
4678                case 0x2006:    /* SIX-PER-EM SPACE */
4679                case 0x2007:    /* FIGURE SPACE */
4680                case 0x2008:    /* PUNCTUATION SPACE */
4681                case 0x2009:    /* THIN SPACE */
4682                case 0x200A:    /* HAIR SPACE */
4683                case 0x202f:    /* NARROW NO-BREAK SPACE */
4684                case 0x205f:    /* MEDIUM MATHEMATICAL SPACE */
4685                case 0x3000:    /* IDEOGRAPHIC SPACE */
4686    #endif
4687              break;              break;
4688              }              }
4689            }            }
# Line 4452  for (;;) Line 4705  for (;;)
4705              case 0x0c:      /* FF */              case 0x0c:      /* FF */
4706              case 0x0d:      /* CR */              case 0x0d:      /* CR */
4707              case 0x85:      /* NEL */              case 0x85:      /* NEL */
4708    #ifdef COMPILE_PCRE16
4709                case 0x2028:    /* LINE SEPARATOR */
4710                case 0x2029:    /* PARAGRAPH SEPARATOR */
4711    #endif
4712              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4713              }              }
4714            }            }
# Line 4473  for (;;) Line 4730  for (;;)
4730              case 0x0c:      /* FF */              case 0x0c:      /* FF */
4731              case 0x0d:      /* CR */              case 0x0d:      /* CR */
4732              case 0x85:      /* NEL */              case 0x85:      /* NEL */
4733    #ifdef COMPILE_PCRE16
4734                case 0x2028:    /* LINE SEPARATOR */
4735                case 0x2029:    /* PARAGRAPH SEPARATOR */
4736    #endif
4737              break;              break;
4738              }              }
4739            }            }
# Line 4486  for (;;) Line 4747  for (;;)
4747              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4748              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4749              }              }
4750            if ((md->ctypes[*eptr++] & ctype_digit) != 0) RRETURN(MATCH_NOMATCH);            if (MAX_255(*eptr) && (md->ctypes[*eptr] & ctype_digit) != 0)
4751                RRETURN(MATCH_NOMATCH);
4752              eptr++;
4753            }            }
4754          break;          break;
4755    
# Line 4498  for (;;) Line 4761  for (;;)
4761              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4762              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4763              }              }
4764            if ((md->ctypes[*eptr++] & ctype_digit) == 0) RRETURN(MATCH_NOMATCH);            if (!MAX_255(*eptr) || (md->ctypes[*eptr] & ctype_digit) == 0)
4765                RRETURN(MATCH_NOMATCH);
4766              eptr++;
4767            }            }
4768          break;          break;
4769    
# Line 4510  for (;;) Line 4775  for (;;)
4775              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4776              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4777              }              }
4778            if ((md->ctypes[*eptr++] & ctype_space) != 0) RRETURN(MATCH_NOMATCH);            if (MAX_255(*eptr) && (md->ctypes[*eptr] & ctype_space) != 0)
4779                RRETURN(MATCH_NOMATCH);
4780              eptr++;
4781            }            }
4782          break;          break;
4783    
# Line 4522  for (;;) Line 4789  for (;;)
4789              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4790              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4791              }              }
4792            if ((md->ctypes[*eptr++] & ctype_space) == 0) RRETURN(MATCH_NOMATCH);            if (!MAX_255(*eptr) || (md->ctypes[*eptr] & ctype_space) == 0)
4793                RRETURN(MATCH_NOMATCH);
4794              eptr++;
4795            }            }
4796          break;          break;
4797    
# Line 4534  for (;;) Line 4803  for (;;)
4803              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4804              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4805              }              }
4806            if ((md->ctypes[*eptr++] & ctype_word) != 0)            if (MAX_255(*eptr) && (md->ctypes[*eptr] & ctype_word) != 0)
4807              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4808              eptr++;
4809            }            }
4810          break;          break;
4811    
# Line 4547  for (;;) Line 4817  for (;;)
4817              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4818              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4819              }              }
4820            if ((md->ctypes[*eptr++] & ctype_word) == 0)            if (!MAX_255(*eptr) || (md->ctypes[*eptr] & ctype_word) == 0)
4821              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4822              eptr++;
4823            }            }
4824          break;          break;
4825    
# Line 4766  for (;;) Line 5037  for (;;)
5037            while (eptr < md->end_subject)            while (eptr < md->end_subject)
5038              {              {
5039              int len = 1;              int len = 1;
5040              if (!utf8) c = *eptr; else { GETCHARLEN(c, eptr, len); }              if (!utf) c = *eptr; else { GETCHARLEN(c, eptr, len); }
5041              if (UCD_CATEGORY(c) != ucp_M) break;              if (UCD_CATEGORY(c) != ucp_M) break;
5042              eptr += len;              eptr += len;
5043              }              }
5044              CHECK_PARTIAL();
5045            }            }
5046          }          }
5047        else        else
5048  #endif     /* SUPPORT_UCP */  #endif     /* SUPPORT_UCP */
5049    
5050  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
5051        /* UTF-8 mode */        if (utf)
       if (utf8)  
5052          {          {
5053          for (fi = min;; fi++)          for (fi = min;; fi++)
5054            {            {
# Line 4794  for (;;) Line 5065  for (;;)
5065            GETCHARINC(c, eptr);            GETCHARINC(c, eptr);
5066            switch(ctype)            switch(ctype)
5067              {              {
5068              case OP_ANY:        /* This is the non-NL case */              case OP_ANY:               /* This is the non-NL case */
5069                if (md->partial != 0 &&    /* Take care with CRLF partial */
5070                    eptr >= md->end_subject &&
5071                    NLBLOCK->nltype == NLTYPE_FIXED &&
5072                    NLBLOCK->nllen == 2 &&
5073                    c == NLBLOCK->nl[0])
5074                  {
5075                  md->hitend = TRUE;
5076                  if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
5077                  }
5078                break;
5079    
5080              case OP_ALLANY:              case OP_ALLANY:
5081              case OP_ANYBYTE:              case OP_ANYBYTE:
5082              break;              break;
# Line 4919  for (;;) Line 5201  for (;;)
5201              break;              break;
5202    
5203              case OP_WHITESPACE:              case OP_WHITESPACE:
5204              if  (c >= 256 || (md->ctypes[c] & ctype_space) == 0)              if (c >= 256 || (md->ctypes[c] & ctype_space) == 0)
5205                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
5206              break;              break;
5207    
# Line 4940  for (;;) Line 5222  for (;;)
5222          }          }
5223        else        else
5224  #endif  #endif
5225        /* Not UTF-8 mode */        /* Not UTF mode */
5226          {          {
5227          for (fi = min;; fi++)          for (fi = min;; fi++)
5228            {            {
# Line 4957  for (;;) Line 5239  for (;;)
5239            c = *eptr++;            c = *eptr++;
5240            switch(ctype)            switch(ctype)
5241              {              {
5242              case OP_ANY:     /* This is the non-NL case */              case OP_ANY:               /* This is the non-NL case */
5243                if (md->partial != 0 &&    /* Take care with CRLF partial */
5244                    eptr >= md->end_subject &&
5245                    NLBLOCK->nltype == NLTYPE_FIXED &&
5246                    NLBLOCK->nllen == 2 &&
5247                    c == NLBLOCK->nl[0])
5248                  {
5249                  md->hitend = TRUE;
5250                  if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
5251                  }
5252                break;
5253    
5254              case OP_ALLANY:              case OP_ALLANY:
5255              case OP_ANYBYTE:              case OP_ANYBYTE:
5256              break;              break;
# Line 4976  for (;;) Line 5269  for (;;)
5269                case 0x000b:                case 0x000b:
5270                case 0x000c:                case 0x000c:
5271                case 0x0085:                case 0x0085:
5272    #ifdef COMPILE_PCRE16
5273                  case 0x2028:
5274                  case 0x2029:
5275    #endif
5276                if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);                if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);
5277                break;                break;
5278                }                }
# Line 4988  for (;;) Line 5285  for (;;)
5285                case 0x09:      /* HT */                case 0x09:      /* HT */
5286                case 0x20:      /* SPACE */                case 0x20:      /* SPACE */
5287                case 0xa0:      /* NBSP */                case 0xa0:      /* NBSP */
5288    #ifdef COMPILE_PCRE16
5289                  case 0x1680:    /* OGHAM SPACE MARK */
5290                  case 0x180e:    /* MONGOLIAN VOWEL SEPARATOR */
5291                  case 0x2000:    /* EN QUAD */
5292                  case 0x2001:    /* EM QUAD */
5293                  case 0x2002:    /* EN SPACE */
5294                  case 0x2003:    /* EM SPACE */
5295                  case 0x2004:    /* THREE-PER-EM SPACE */
5296                  case 0x2005:    /* FOUR-PER-EM SPACE */
5297                  case 0x2006:    /* SIX-PER-EM SPACE */
5298                  case 0x2007:    /* FIGURE SPACE */
5299                  case 0x2008:    /* PUNCTUATION SPACE */
5300                  case 0x2009:    /* THIN SPACE */
5301                  case 0x200A:    /* HAIR SPACE */
5302                  case 0x202f:    /* NARROW NO-BREAK SPACE */
5303                  case 0x205f:    /* MEDIUM MATHEMATICAL SPACE */
5304                  case 0x3000:    /* IDEOGRAPHIC SPACE */
5305    #endif
5306                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
5307                }                }
5308              break;              break;
# Line 4999  for (;;) Line 5314  for (;;)
5314                case 0x09:      /* HT */                case 0x09:      /* HT */
5315                case 0x20:      /* SPACE */                case 0x20:      /* SPACE */
5316                case 0xa0:      /* NBSP */                case 0xa0:      /* NBSP */
5317    #ifdef COMPILE_PCRE16
5318                  case 0x1680:    /* OGHAM SPACE MARK */
5319                  case 0x180e:    /* MONGOLIAN VOWEL SEPARATOR */
5320                  case 0x2000:    /* EN QUAD */
5321                  case 0x2001:    /* EM QUAD */
5322                  case 0x2002:    /* EN SPACE */
5323                  case 0x2003:    /* EM SPACE */
5324                  case 0x2004:    /* THREE-PER-EM SPACE */
5325                  case 0x2005:    /* FOUR-PER-EM SPACE */
5326                  case 0x2006:    /* SIX-PER-EM SPACE */
5327                  case 0x2007:    /* FIGURE SPACE */
5328                  case 0x2008:    /* PUNCTUATION SPACE */
5329                  case 0x2009:    /* THIN SPACE */
5330                  case 0x200A:    /* HAIR SPACE */
5331                  case 0x202f:    /* NARROW NO-BREAK SPACE */
5332                  case 0x205f:    /* MEDIUM MATHEMATICAL SPACE */
5333                  case 0x3000:    /* IDEOGRAPHIC SPACE */
5334    #endif
5335                break;                break;
5336                }                }
5337              break;              break;
# Line 5012  for (;;) Line 5345  for (;;)
5345                case 0x0c:      /* FF */                case 0x0c:      /* FF */
5346                case 0x0d:      /* CR */                case 0x0d:      /* CR */
5347                case 0x85:      /* NEL */                case 0x85:      /* NEL */
5348    #ifdef COMPILE_PCRE16
5349                  case 0x2028:    /* LINE SEPARATOR */
5350                  case 0x2029:    /* PARAGRAPH SEPARATOR */
5351    #endif
5352                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
5353                }                }
5354              break;              break;
# Line 5025  for (;;) Line 5362  for (;;)
5362                case 0x0c:      /* FF */                case 0x0c:      /* FF */
5363                case 0x0d:      /* CR */                case 0x0d:      /* CR */
5364                case 0x85:      /* NEL */                case 0x85:      /* NEL */
5365    #ifdef COMPILE_PCRE16
5366                  case 0x2028:    /* LINE SEPARATOR */
5367                  case 0x2029:    /* PARAGRAPH SEPARATOR */
5368    #endif
5369                break;                break;
5370                }                }
5371              break;              break;
5372    
5373              case OP_NOT_DIGIT:              case OP_NOT_DIGIT:
5374              if ((md->ctypes[c] & ctype_digit) != 0) RRETURN(MATCH_NOMATCH);              if (MAX_255(c) && (md->ctypes[c] & ctype_digit) != 0) RRETURN(MATCH_NOMATCH);
5375              break;              break;
5376    
5377              case OP_DIGIT:              case OP_DIGIT:
5378              if ((md->ctypes[c] & ctype_digit) == 0) RRETURN(MATCH_NOMATCH);              if (!MAX_255(c) || (md->ctypes[c] & ctype_digit) == 0) RRETURN(MATCH_NOMATCH);
5379              break;              break;
5380    
5381              case OP_NOT_WHITESPACE:              case OP_NOT_WHITESPACE:
5382              if ((md->ctypes[c] & ctype_space) != 0) RRETURN(MATCH_NOMATCH);              if (MAX_255(c) && (md->ctypes[c] & ctype_space) != 0) RRETURN(MATCH_NOMATCH);
5383              break;              break;
5384    
5385              case OP_WHITESPACE:              case OP_WHITESPACE:
5386              if  ((md->ctypes[c] & ctype_space) == 0) RRETURN(MATCH_NOMATCH);              if (!MAX_255(c) || (md->ctypes[c] & ctype_space) == 0) RRETURN(MATCH_NOMATCH);
5387              break;              break;
5388    
5389              case OP_NOT_WORDCHAR:              case OP_NOT_WORDCHAR:
5390              if ((md->ctypes[c] & ctype_word) != 0) RRETURN(MATCH_NOMATCH);              if (MAX_255(c) && (md->ctypes[c] & ctype_word) != 0) RRETURN(MATCH_NOMATCH);
5391              break;              break;
5392    
5393              case OP_WORDCHAR:              case OP_WORDCHAR:
5394              if ((md->ctypes[c] & ctype_word) == 0) RRETURN(MATCH_NOMATCH);              if (!MAX_255(c) || (md->ctypes[c] & ctype_word) == 0) RRETURN(MATCH_NOMATCH);
5395              break;              break;
5396    
5397              default:              default:
# Line 5239  for (;;) Line 5580  for (;;)
5580            RMATCH(eptr, ecode, offset_top, md, eptrb, RM44);            RMATCH(eptr, ecode, offset_top, md, eptrb, RM44);
5581            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
5582            if (eptr-- == pp) break;        /* Stop if tried at original pos */            if (eptr-- == pp) break;        /* Stop if tried at original pos */
5583            if (utf8) BACKCHAR(eptr);            if (utf) BACKCHAR(eptr);
5584            }            }
5585          }          }
5586    
# Line 5256  for (;;) Line 5597  for (;;)
5597              SCHECK_PARTIAL();              SCHECK_PARTIAL();
5598              break;              break;
5599              }              }
5600            if (!utf8) c = *eptr; else { GETCHARLEN(c, eptr, len); }            if (!utf) c = *eptr; else { GETCHARLEN(c, eptr, len); }
5601            if (UCD_CATEGORY(c) == ucp_M) break;            if (UCD_CATEGORY(c) == ucp_M) break;
5602            eptr += len;            eptr += len;
5603            while (eptr < md->end_subject)            while (eptr < md->end_subject)
5604              {              {
5605              len = 1;              len = 1;
5606              if (!utf8) c = *eptr; else { GETCHARLEN(c, eptr, len); }              if (!utf) c = *eptr; else { GETCHARLEN(c, eptr, len); }
5607              if (UCD_CATEGORY(c) != ucp_M) break;              if (UCD_CATEGORY(c) != ucp_M) break;
5608              eptr += len;              eptr += len;
5609              }              }
5610              CHECK_PARTIAL();
5611            }            }
5612    
5613          /* eptr is now past the end of the maximum run */          /* eptr is now past the end of the maximum run */
# Line 5279  for (;;) Line 5621  for (;;)
5621            if (eptr-- == pp) break;        /* Stop if tried at original pos */            if (eptr-- == pp) break;        /* Stop if tried at original pos */
5622            for (;;)                        /* Move back over one extended */            for (;;)                        /* Move back over one extended */
5623              {              {
5624              if (!utf8) c = *eptr; else              if (!utf) c = *eptr; else
5625                {                {
5626                BACKCHAR(eptr);                BACKCHAR(eptr);
5627                GETCHAR(c, eptr);                GETCHAR(c, eptr);
# Line 5293  for (;;) Line 5635  for (;;)
5635        else        else
5636  #endif   /* SUPPORT_UCP */  #endif   /* SUPPORT_UCP */
5637    
5638  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
5639        /* UTF-8 mode */        if (utf)
   
       if (utf8)  
5640          {          {
5641          switch(ctype)          switch(ctype)
5642            {            {
# Line 5311  for (;;) Line 5651  for (;;)
5651                  break;                  break;
5652                  }                  }
5653                if (IS_NEWLINE(eptr)) break;                if (IS_NEWLINE(eptr)) break;
5654                  if (md->partial != 0 &&    /* Take care with CRLF partial */
5655                      eptr + 1 >= md->end_subject &&
5656                      NLBLOCK->nltype == NLTYPE_FIXED &&
5657                      NLBLOCK->nllen == 2 &&
5658                      *eptr == NLBLOCK->nl[0])
5659                    {
5660                    md->hitend = TRUE;
5661                    if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
5662                    }
5663                eptr++;                eptr++;
5664                while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;                ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);
5665                }                }
5666              }              }
5667    
# Line 5328  for (;;) Line 5677  for (;;)
5677                  break;                  break;
5678                  }                  }
5679                if (IS_NEWLINE(eptr)) break;                if (IS_NEWLINE(eptr)) break;
5680                  if (md->partial != 0 &&    /* Take care with CRLF partial */
5681                      eptr + 1 >= md->end_subject &&
5682                      NLBLOCK->nltype == NLTYPE_FIXED &&
5683                      NLBLOCK->nllen == 2 &&
5684                      *eptr == NLBLOCK->nl[0])
5685                    {
5686                    md->hitend = TRUE;
5687                    if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
5688                    }
5689                eptr++;                eptr++;
5690                while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;                ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);
5691                }                }
5692              }              }
5693            break;            break;
# Line 5345  for (;;) Line 5703  for (;;)
5703                  break;                  break;
5704                  }                  }
5705                eptr++;                eptr++;
5706                while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;                ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);
5707                }                }
5708              }              }
5709            else            else
# Line 5578  for (;;) Line 5936  for (;;)
5936            }            }
5937          }          }
5938        else        else
5939  #endif  /* SUPPORT_UTF8 */  #endif  /* SUPPORT_UTF */
5940          /* Not UTF mode */
       /* Not UTF-8 mode */  
5941          {          {
5942          switch(ctype)          switch(ctype)
5943            {            {
# Line 5593  for (;;) Line 5950  for (;;)
5950                break;                break;
5951                }                }
5952              if (IS_NEWLINE(eptr)) break;              if (IS_NEWLINE(eptr)) break;
5953                if (md->partial != 0 &&    /* Take care with CRLF partial */
5954                    eptr + 1 >= md->end_subject &&
5955                    NLBLOCK->nltype == NLTYPE_FIXED &&
5956                    NLBLOCK->nllen == 2 &&
5957                    *eptr == NLBLOCK->nl[0])
5958                  {
5959                  md->hitend = TRUE;
5960                  if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
5961                  }
5962              eptr++;              eptr++;
5963              }              }
5964            break;            break;
# Line 5624  for (;;) Line 5990  for (;;)
5990                }                }
5991              else              else
5992                {                {
5993                if (c != 0x000a &&                if (c != 0x000a && (md->bsr_anycrlf ||
5994                    (md->bsr_anycrlf ||                  (c != 0x000b && c != 0x000c && c != 0x0085
5995                      (c != 0x000b && c != 0x000c && c != 0x0085)))  #ifdef COMPILE_PCRE16
5996                  break;                  && c != 0x2028 && c != 0x2029
5997    #endif
5998                    ))) break;
5999                eptr++;                eptr++;
6000                }                }
6001              }              }
# Line 5642  for (;;) Line 6010  for (;;)
6010                break;                break;
6011                }                }
6012              c = *eptr;              c = *eptr;
6013              if (c == 0x09 || c == 0x20 || c == 0xa0) break;              if (c == 0x09 || c == 0x20 || c == 0xa0
6014    #ifdef COMPILE_PCRE16
6015                  || c == 0x1680 || c == 0x180e || (c >= 0x2000 && c <= 0x200A)
6016                  || c == 0x202f || c == 0x205f || c == 0x3000
6017    #endif
6018                  ) break;
6019              eptr++;              eptr++;
6020              }              }
6021            break;            break;
# Line 5656  for (;;) Line 6029  for (;;)
6029                break;                break;
6030                }                }
6031              c = *eptr;              c = *eptr;
6032              if (c != 0x09 && c != 0x20 && c != 0xa0) break;              if (c != 0x09 && c != 0x20 && c != 0xa0
6033    #ifdef COMPILE_PCRE16
6034                  && c != 0x1680 && c != 0x180e && (c < 0x2000 || c > 0x200A)
6035                  && c != 0x202f && c != 0x205f && c != 0x3000
6036    #endif
6037                  ) break;
6038              eptr++;              eptr++;
6039              }              }
6040            break;            break;
# Line 5670  for (;;) Line 6048  for (;;)
6048                break;                break;
6049                }                }
6050              c = *eptr;              c = *eptr;
6051              if (c == 0x0a || c == 0x0b || c == 0x0c || c == 0x0d || c == 0x85)              if (c == 0x0a || c == 0x0b || c == 0x0c || c == 0x0d || c == 0x85
6052                break;  #ifdef COMPILE_PCRE16
6053                  || c == 0x2028 || c == 0x2029
6054    #endif
6055                  ) break;
6056              eptr++;              eptr++;
6057              }              }
6058            break;            break;
# Line 5685  for (;;) Line 6066  for (;;)
6066                break;                break;
6067                }                }
6068              c = *eptr;              c = *eptr;
6069              if (c != 0x0a && c != 0x0b && c != 0x0c && c != 0x0d && c != 0x85)              if (c != 0x0a && c != 0x0b && c != 0x0c && c != 0x0d && c != 0x85
6070                break;  #ifdef COMPILE_PCRE16
6071                  && c != 0x2028 && c != 0x2029
6072    #endif
6073                  ) break;
6074              eptr++;              eptr++;
6075              }              }
6076            break;            break;
# Line 5699  for (;;) Line 6083  for (;;)
6083                SCHECK_PARTIAL();                SCHECK_PARTIAL();
6084                break;                break;
6085                }                }
6086              if ((md->ctypes[*eptr] & ctype_digit) != 0) break;              if (MAX_255(*eptr) && (md->ctypes[*eptr] & ctype_digit) != 0) break;
6087              eptr++;              eptr++;
6088              }              }
6089            break;            break;
# Line 5712  for (;;) Line 6096  for (;;)
6096                SCHECK_PARTIAL();                SCHECK_PARTIAL();
6097                break;                break;
6098                }                }
6099              if ((md->ctypes[*eptr] & ctype_digit) == 0) break;              if (!MAX_255(*eptr) || (md->ctypes[*eptr] & ctype_digit) == 0) break;
6100              eptr++;              eptr++;
6101              }              }
6102            break;            break;
# Line 5725  for (;;) Line 6109  for (;;)
6109                SCHECK_PARTIAL();                SCHECK_PARTIAL();
6110                break;                break;
6111                }                }
6112              if ((md->ctypes[*eptr] & ctype_space) != 0) break;              if (MAX_255(*eptr) && (md->ctypes[*eptr] & ctype_space) != 0) break;
6113              eptr++;              eptr++;
6114              }              }
6115            break;            break;
# Line 5738  for (;;) Line 6122  for (;;)
6122                SCHECK_PARTIAL();                SCHECK_PARTIAL();
6123                break;                break;
6124                }                }
6125              if ((md->ctypes[*eptr] & ctype_space) == 0) break;              if (!MAX_255(*eptr) || (md->ctypes[*eptr] & ctype_space) == 0) break;
6126              eptr++;              eptr++;
6127              }              }
6128            break;            break;
# Line 5751  for (;;) Line 6135  for (;;)
6135                SCHECK_PARTIAL();                SCHECK_PARTIAL();
6136                break;                break;
6137                }                }
6138              if ((md->ctypes[*eptr] & ctype_word) != 0) break;              if (MAX_255(*eptr) && (md->ctypes[*eptr] & ctype_word) != 0) break;
6139              eptr++;              eptr++;
6140              }              }
6141            break;            break;
# Line 5764  for (;;) Line 6148  for (;;)
6148                SCHECK_PARTIAL();                SCHECK_PARTIAL();
6149                break;                break;
6150                }                }
6151              if ((md->ctypes[*eptr] & ctype_word) == 0) break;              if (!MAX_255(*eptr) || (md->ctypes[*eptr] & ctype_word) == 0) break;
6152              eptr++;              eptr++;
6153              }              }
6154            break;            break;
# Line 5827  switch (frame->Xwhere) Line 6211  switch (frame->Xwhere)
6211    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)
6212    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)
6213    LBL(65) LBL(66)    LBL(65) LBL(66)
6214  #ifdef SUPPORT_UTF8  #if defined SUPPORT_UTF || !defined COMPILE_PCRE8
6215    LBL(16) LBL(18) LBL(20) LBL(21) LBL(22) LBL(23) LBL(28) LBL(30)    LBL(21)
6216    #endif
6217    #ifdef SUPPORT_UTF
6218      LBL(16) LBL(18) LBL(20)
6219      LBL(22) LBL(23) LBL(28) LBL(30)
6220    LBL(32) LBL(34) LBL(42) LBL(46)    LBL(32) LBL(34) LBL(42) LBL(46)
6221  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
6222    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)
6223    LBL(59) LBL(60) LBL(61) LBL(62)    LBL(59) LBL(60) LBL(61) LBL(62)
6224  #endif  /* SUPPORT_UCP */  #endif  /* SUPPORT_UCP */
6225  #endif  /* SUPPORT_UTF8 */  #endif  /* SUPPORT_UTF */
6226    default:    default:
6227    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));
6228    
6229    printf("+++jump error in pcre match: label %d non-existent\n", frame->Xwhere);
6230    
6231    return PCRE_ERROR_INTERNAL;    return PCRE_ERROR_INTERNAL;
6232    }    }
6233  #undef LBL  #undef LBL
# Line 5898  Undefine all the macros that were define Line 6289  Undefine all the macros that were define
6289  ***************************************************************************/  ***************************************************************************/
6290    
6291    
6292    #ifdef NO_RECURSE
6293    /*************************************************
6294    *          Release allocated heap frames         *
6295    *************************************************/
6296    
6297    /* This function releases all the allocated frames. The base frame is on the
6298    machine stack, and so must not be freed.
6299    
6300    Argument: the address of the base frame
6301    Returns:  nothing
6302    */
6303    
6304    static void
6305    release_match_heapframes (heapframe *frame_base)
6306    {
6307    heapframe *nextframe = frame_base->Xnextframe;
6308    while (nextframe != NULL)
6309      {
6310      heapframe *oldframe = nextframe;
6311      nextframe = nextframe->Xnextframe;
6312      (PUBL(stack_free))(oldframe);
6313      }
6314    }
6315    #endif
6316    
6317    
6318  /*************************************************  /*************************************************
6319  *         Execute a Regular Expression           *  *         Execute a Regular Expression           *
# Line 5923  Returns:          > 0 => success; value Line 6339  Returns:          > 0 => success; value
6339                   < -1 => some kind of unexpected problem                   < -1 => some kind of unexpected problem
6340  */  */
6341    
6342    #ifdef COMPILE_PCRE8
6343  PCRE_EXP_DEFN int PCRE_CALL_CONVENTION  PCRE_EXP_DEFN int PCRE_CALL_CONVENTION
6344  pcre_exec(const pcre *argument_re, const pcre_extra *extra_data,  pcre_exec(const pcre *argument_re, const pcre_extra *extra_data,
6345    PCRE_SPTR subject, int length, int start_offset, int options, int *offsets,    PCRE_SPTR subject, int length, int start_offset, int options, int *offsets,
6346    int offsetcount)    int offsetcount)
6347    #else
6348    PCRE_EXP_DEFN int PCRE_CALL_CONVENTION
6349    pcre16_exec(const pcre16 *argument_re, const pcre16_extra *extra_data,
6350      PCRE_SPTR16 subject, int length, int start_offset, int options, int *offsets,
6351      int offsetcount)
6352    #endif
6353  {  {
6354  int rc, ocount, arg_offset_max;  int rc, ocount, arg_offset_max;
 int first_byte = -1;  
 int req_byte = -1;  
 int req_byte2 = -1;  
6355  int newline;  int newline;
6356  BOOL using_temporary_offsets = FALSE;  BOOL using_temporary_offsets = FALSE;
6357  BOOL anchored;  BOOL anchored;
6358  BOOL startline;  BOOL startline;
6359  BOOL firstline;  BOOL firstline;
6360  BOOL first_byte_caseless = FALSE;  BOOL utf;
6361  BOOL req_byte_caseless = FALSE;  BOOL has_first_char = FALSE;
6362  BOOL utf8;  BOOL has_req_char = FALSE;
6363    pcre_uchar first_char = 0;
6364    pcre_uchar first_char2 = 0;
6365    pcre_uchar req_char = 0;
6366    pcre_uchar req_char2 = 0;
6367  match_data match_block;  match_data match_block;
6368  match_data *md = &match_block;  match_data *md = &match_block;
6369  const uschar *tables;  const pcre_uint8 *tables;
6370  const uschar *start_bits = NULL;  const pcre_uint8 *start_bits = NULL;
6371  USPTR start_match = (USPTR)subject + start_offset;  PCRE_PUCHAR start_match = (PCRE_PUCHAR)subject + start_offset;
6372  USPTR end_subject;  PCRE_PUCHAR end_subject;
6373  USPTR start_partial = NULL;  PCRE_PUCHAR start_partial = NULL;
6374  USPTR req_byte_ptr = start_match - 1;  PCRE_PUCHAR req_char_ptr = start_match - 1;
6375    
 pcre_study_data internal_study;  
6376  const pcre_study_data *study;  const pcre_study_data *study;
6377    const REAL_PCRE *re = (const REAL_PCRE *)argument_re;
6378    
6379    #ifdef NO_RECURSE
6380    heapframe frame_zero;
6381    frame_zero.Xprevframe = NULL;            /* Marks the top level */
6382    frame_zero.Xnextframe = NULL;            /* None are allocated yet */
6383    md->match_frames_base = &frame_zero;
6384    #endif
6385    
6386    /* Check for the special magic call that measures the size of the stack used
6387    per recursive call of match(). Without the funny casting for sizeof, a Windows
6388    compiler gave this error: "unary minus operator applied to unsigned type,
6389    result still unsigned". Hopefully the cast fixes that. */
6390    
6391  real_pcre internal_re;  if (re == NULL && extra_data == NULL && subject == NULL && length == -999 &&
6392  const real_pcre *external_re = (const real_pcre *)argument_re;      start_offset == -999)
6393  const real_pcre *re = external_re;  #ifdef NO_RECURSE
6394      return -((int)sizeof(heapframe));
6395    #else
6396      return match(NULL, NULL, NULL, 0, NULL, NULL, 0);
6397    #endif
6398    
6399  /* Plausibility checks */  /* Plausibility checks */
6400    
6401  if ((options & ~PUBLIC_EXEC_OPTIONS) != 0) return PCRE_ERROR_BADOPTION;  if ((options & ~PUBLIC_EXEC_OPTIONS) != 0) return PCRE_ERROR_BADOPTION;
6402  if (re == NULL || subject == NULL ||  if (re == NULL || subject == NULL || (offsets == NULL && offsetcount > 0))
6403     (offsets == NULL && offsetcount > 0)) return PCRE_ERROR_NULL;    return PCRE_ERROR_NULL;
6404  if (offsetcount < 0) return PCRE_ERROR_BADCOUNT;  if (offsetcount < 0) return PCRE_ERROR_BADCOUNT;
6405  if (start_offset < 0 || start_offset > length) return PCRE_ERROR_BADOFFSET;  if (start_offset < 0 || start_offset > length) return PCRE_ERROR_BADOFFSET;
6406    
6407    /* Check that the first field in the block is the magic number. If it is not,
6408    return with PCRE_ERROR_BADMAGIC. However, if the magic number is equal to
6409    REVERSED_MAGIC_NUMBER we return with PCRE_ERROR_BADENDIANNESS, which
6410    means that the pattern is likely compiled with different endianness. */
6411    
6412    if (re->magic_number != MAGIC_NUMBER)
6413      return re->magic_number == REVERSED_MAGIC_NUMBER?
6414        PCRE_ERROR_BADENDIANNESS:PCRE_ERROR_BADMAGIC;
6415    if ((re->flags & PCRE_MODE) == 0) return PCRE_ERROR_BADMODE;
6416    
6417  /* 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
6418  follows immediately afterwards. Other values in the md block are used only  follows immediately afterwards. Other values in the md block are used only
6419  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,
6420  so they are set up later. */  so they are set up later. */
6421    
6422  utf8 = md->utf8 = (re->options & PCRE_UTF8) != 0;  /* PCRE_UTF16 has the same value as PCRE_UTF8. */
6423    utf = md->utf = (re->options & PCRE_UTF8) != 0;
6424  md->partial = ((options & PCRE_PARTIAL_HARD) != 0)? 2 :  md->partial = ((options & PCRE_PARTIAL_HARD) != 0)? 2 :
6425                ((options & PCRE_PARTIAL_SOFT) != 0)? 1 : 0;                ((options & PCRE_PARTIAL_SOFT) != 0)? 1 : 0;
6426    
6427  /* 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
6428  code for an invalid string if a results vector is available. */  code for an invalid string if a results vector is available. */
6429    
6430  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
6431  if (utf8 && (options & PCRE_NO_UTF8_CHECK) == 0)  if (utf && (options & PCRE_NO_UTF8_CHECK) == 0)
6432    {    {
6433    int erroroffset;    int erroroffset;
6434    int errorcode = _pcre_valid_utf8((USPTR)subject, length, &erroroffset);    int errorcode = PRIV(valid_utf)((PCRE_PUCHAR)subject, length, &erroroffset);
6435    if (errorcode != 0)    if (errorcode != 0)
6436      {      {
6437      if (offsetcount >= 2)      if (offsetcount >= 2)
# Line 5988  if (utf8 && (options & PCRE_NO_UTF8_CHEC Line 6439  if (utf8 && (options & PCRE_NO_UTF8_CHEC
6439        offsets[0] = erroroffset;        offsets[0] = erroroffset;
6440        offsets[1] = errorcode;        offsets[1] = errorcode;
6441        }        }
6442    #ifdef COMPILE_PCRE16
6443        return (errorcode <= PCRE_UTF16_ERR1 && md->partial > 1)?
6444          PCRE_ERROR_SHORTUTF16 : PCRE_ERROR_BADUTF16;
6445    #else
6446      return (errorcode <= PCRE_UTF8_ERR5 && md->partial > 1)?      return (errorcode <= PCRE_UTF8_ERR5 && md->partial > 1)?
6447        PCRE_ERROR_SHORTUTF8 : PCRE_ERROR_BADUTF8;        PCRE_ERROR_SHORTUTF8 : PCRE_ERROR_BADUTF8;
6448    #endif
6449      }      }
6450    
6451    /* 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. */
6452    if (start_offset > 0 && start_offset < length &&    if (start_offset > 0 && start_offset < length &&
6453        (((USPTR)subject)[start_offset] & 0xc0) == 0x80)        NOT_FIRSTCHAR(((PCRE_PUCHAR)subject)[start_offset]))
6454      return PCRE_ERROR_BADUTF8_OFFSET;      return PCRE_ERROR_BADUTF8_OFFSET;
6455    }    }
6456  #endif  #endif
# Line 6002  if (utf8 && (options & PCRE_NO_UTF8_CHEC Line 6458  if (utf8 && (options & PCRE_NO_UTF8_CHEC
6458  /* If the pattern was successfully studied with JIT support, run the JIT  /* If the pattern was successfully studied with JIT support, run the JIT
6459  executable instead of the rest of this function. Most options must be set at  executable instead of the rest of this function. Most options must be set at
6460  compile time for the JIT code to be usable. Fallback to the normal code path if  compile time for the JIT code to be usable. Fallback to the normal code path if
6461  an unsupported flag is set. In particular, JIT does not support partial  an unsupported flag is set. */
 matching. */  
6462    
6463  #ifdef SUPPORT_JIT  #ifdef SUPPORT_JIT
6464  if (extra_data != NULL  if (extra_data != NULL
6465      && (extra_data->flags & PCRE_EXTRA_EXECUTABLE_JIT) != 0      && (extra_data->flags & (PCRE_EXTRA_EXECUTABLE_JIT |
6466                                 PCRE_EXTRA_TABLES)) == PCRE_EXTRA_EXECUTABLE_JIT
6467      && extra_data->executable_jit != NULL      && extra_data->executable_jit != NULL
     && (extra_data->flags & PCRE_EXTRA_TABLES) == 0  
6468      && (options & ~(PCRE_NO_UTF8_CHECK | PCRE_NOTBOL | PCRE_NOTEOL |      && (options & ~(PCRE_NO_UTF8_CHECK | PCRE_NOTBOL | PCRE_NOTEOL |
6469                      PCRE_NOTEMPTY | PCRE_NOTEMPTY_ATSTART)) == 0)                      PCRE_NOTEMPTY | PCRE_NOTEMPTY_ATSTART |
6470    return _pcre_jit_exec(re, extra_data->executable_jit, subject, length,                      PCRE_PARTIAL_SOFT | PCRE_PARTIAL_HARD)) == 0)
6471      start_offset, options, ((extra_data->flags & PCRE_EXTRA_MATCH_LIMIT) == 0)    {
6472      ? MATCH_LIMIT : extra_data->match_limit, offsets, offsetcount);    rc = PRIV(jit_exec)(re, extra_data, (const pcre_uchar *)subject, length,
6473           start_offset, options, offsets, offsetcount);
6474    
6475      /* PCRE_ERROR_NULL means that the selected normal or partial matching
6476      mode is not compiled. In this case we simply fallback to interpreter. */
6477    
6478      if (rc != PCRE_ERROR_NULL) return rc;
6479      }
6480  #endif  #endif
6481    
6482  /* Carry on with non-JIT matching. This information is for finding all the  /* Carry on with non-JIT matching. This information is for finding all the
6483  numbers associated with a given name, for condition testing. */  numbers associated with a given name, for condition testing. */
6484    
6485  md->name_table = (uschar *)re + re->name_table_offset;  md->name_table = (pcre_uchar *)re + re->name_table_offset;
6486  md->name_count = re->name_count;  md->name_count = re->name_count;
6487  md->name_entry_size = re->name_entry_size;  md->name_entry_size = re->name_entry_size;
6488    
# Line 6034  md->callout_data = NULL; Line 6496  md->callout_data = NULL;
6496    
6497  /* The table pointer is always in native byte order. */  /* The table pointer is always in native byte order. */
6498    
6499  tables = external_re->tables;  tables = re->tables;
6500    
6501  if (extra_data != NULL)  if (extra_data != NULL)
6502    {    {
# Line 6054  if (extra_data != NULL) Line 6516  if (extra_data != NULL)
6516  is a feature that makes it possible to save compiled regex and re-use them  is a feature that makes it possible to save compiled regex and re-use them
6517  in other programs later. */  in other programs later. */
6518    
6519  if (tables == NULL) tables = _pcre_default_tables;  if (tables == NULL) tables = PRIV(default_tables);
   
 /* Check that the first field in the block is the magic number. If it is not,  
 test for a regex that was compiled on a host of opposite endianness. If this is  
 the case, flipped values are put in internal_re and internal_study if there was  
 study data too. */  
   
 if (re->magic_number != MAGIC_NUMBER)  
   {  
   re = _pcre_try_flipped(re, &internal_re, study, &internal_study);  
   if (re == NULL) return PCRE_ERROR_BADMAGIC;  
   if (study != NULL) study = &internal_study;  
   }  
6520    
6521  /* Set up other data */  /* Set up other data */
6522    
# Line 6076  firstline = (re->options & PCRE_FIRSTLIN Line 6526  firstline = (re->options & PCRE_FIRSTLIN
6526    
6527  /* The code starts after the real_pcre block and the capture name table. */  /* The code starts after the real_pcre block and the capture name table. */
6528    
6529  md->start_code = (const uschar *)external_re + re->name_table_offset +  md->start_code = (const pcre_uchar *)re + re->name_table_offset +
6530    re->name_count * re->name_entry_size;    re->name_count * re->name_entry_size;
6531    
6532  md->start_subject = (USPTR)subject;  md->start_subject = (PCRE_PUCHAR)subject;
6533  md->start_offset = start_offset;  md->start_offset = start_offset;
6534  md->end_subject = md->start_subject + length;  md->end_subject = md->start_subject + length;
6535  end_subject = md->end_subject;  end_subject = md->end_subject;
# Line 6104  md->recursive = NULL; Line 6554  md->recursive = NULL;
6554  md->hasthen = (re->flags & PCRE_HASTHEN) != 0;  md->hasthen = (re->flags & PCRE_HASTHEN) != 0;
6555    
6556  md->lcc = tables + lcc_offset;  md->lcc = tables + lcc_offset;
6557    md->fcc = tables + fcc_offset;
6558  md->ctypes = tables + ctypes_offset;  md->ctypes = tables + ctypes_offset;
6559    
6560  /* Handle different \R options. */  /* Handle different \R options. */
# Line 6190  arg_offset_max = (2*ocount)/3; Line 6641  arg_offset_max = (2*ocount)/3;
6641  if (re->top_backref > 0 && re->top_backref >= ocount/3)  if (re->top_backref > 0 && re->top_backref >= ocount/3)
6642    {    {
6643    ocount = re->top_backref * 3 + 3;    ocount = re->top_backref * 3 + 3;
6644    md->offset_vector = (int *)(pcre_malloc)(ocount * sizeof(int));    md->offset_vector = (int *)(PUBL(malloc))(ocount * sizeof(int));
6645    if (md->offset_vector == NULL) return PCRE_ERROR_NOMEMORY;    if (md->offset_vector == NULL) return PCRE_ERROR_NOMEMORY;
6646    using_temporary_offsets = TRUE;    using_temporary_offsets = TRUE;
6647    DPRINTF(("Got memory to hold back references\n"));    DPRINTF(("Got memory to hold back references\n"));
# Line 6217  if (md->offset_vector != NULL) Line 6668  if (md->offset_vector != NULL)
6668    md->offset_vector[0] = md->offset_vector[1] = -1;    md->offset_vector[0] = md->offset_vector[1] = -1;
6669    }    }
6670    
6671  /* Set up the first character to match, if available. The first_byte value is  /* Set up the first character to match, if available. The first_char value is
6672  never set for an anchored regular expression, but the anchoring may be forced  never set for an anchored regular expression, but the anchoring may be forced
6673  at run time, so we have to test for anchoring. The first char may be unset for  at run time, so we have to test for anchoring. The first char may be unset for
6674  an unanchored pattern, of course. If there's no first char and the pattern was  an unanchored pattern, of course. If there's no first char and the pattern was
# Line 6227  if (!anchored) Line 6678  if (!anchored)
6678    {    {
6679    if ((re->flags & PCRE_FIRSTSET) != 0)    if ((re->flags & PCRE_FIRSTSET) != 0)
6680      {      {
6681      first_byte = re->first_byte & 255;      has_first_char = TRUE;
6682      if ((first_byte_caseless = ((re->first_byte & REQ_CASELESS) != 0)) == TRUE)      first_char = first_char2 = (pcre_uchar)(re->first_char);
6683        first_byte = md->lcc[first_byte];      if ((re->flags & PCRE_FCH_CASELESS) != 0)
6684          {
6685          first_char2 = TABLE_GET(first_char, md->fcc, first_char);
6686    #if defined SUPPORT_UCP && !(defined COMPILE_PCRE8)
6687          if (utf && first_char > 127)
6688            first_char2 = UCD_OTHERCASE(first_char);
6689    #endif
6690          }
6691      }      }
6692    else    else
6693      if (!startline && study != NULL &&      if (!startline && study != NULL &&
# Line 6242  character" set. */ Line 6700  character" set. */
6700    
6701  if ((re->flags & PCRE_REQCHSET) != 0)  if ((re->flags & PCRE_REQCHSET) != 0)
6702    {    {
6703    req_byte = re->req_byte & 255;    has_req_char = TRUE;
6704    req_byte_caseless = (re->req_byte & REQ_CASELESS) != 0;    req_char = req_char2 = (pcre_uchar)(re->req_char);
6705    req_byte2 = (tables + fcc_offset)[req_byte];  /* case flipped */    if ((re->flags & PCRE_RCH_CASELESS) != 0)
6706        {
6707        req_char2 = TABLE_GET(req_char, md->fcc, req_char);
6708    #if defined SUPPORT_UCP && !(defined COMPILE_PCRE8)
6709        if (utf && req_char > 127)
6710          req_char2 = UCD_OTHERCASE(req_char);
6711    #endif
6712        }
6713    }    }
6714    
6715    
   
   
6716  /* ==========================================================================*/  /* ==========================================================================*/
6717    
6718  /* Loop for handling unanchored repeated matching attempts; for anchored regexs  /* Loop for handling unanchored repeated matching attempts; for anchored regexs
# Line 6257  the loop runs just once. */ Line 6720  the loop runs just once. */
6720    
6721  for(;;)  for(;;)
6722    {    {
6723    USPTR save_end_subject = end_subject;    PCRE_PUCHAR save_end_subject = end_subject;
6724    USPTR new_start_match;    PCRE_PUCHAR new_start_match;
6725    
6726    /* If firstline is TRUE, the start of the match is constrained to the first    /* If firstline is TRUE, the start of the match is constrained to the first
6727    line of a multiline string. That is, the match must be before or at the first    line of a multiline string. That is, the match must be before or at the first
# Line 6268  for(;;) Line 6731  for(;;)
6731    
6732    if (firstline)    if (firstline)
6733      {      {
6734      USPTR t = start_match;      PCRE_PUCHAR t = start_match;
6735  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
6736      if (utf8)      if (utf)
6737        {        {
6738        while (t < md->end_subject && !IS_NEWLINE(t))        while (t < md->end_subject && !IS_NEWLINE(t))
6739          {          {
6740          t++;          t++;
6741          while (t < end_subject && (*t & 0xc0) == 0x80) t++;          ACROSSCHAR(t < end_subject, *t, t++);
6742          }          }
6743        }        }
6744      else      else
# Line 6292  for(;;) Line 6755  for(;;)
6755    
6756    if (((options | re->options) & PCRE_NO_START_OPTIMIZE) == 0)    if (((options | re->options) & PCRE_NO_START_OPTIMIZE) == 0)
6757      {      {
6758      /* Advance to a unique first byte if there is one. */      /* Advance to a unique first char if there is one. */
6759    
6760      if (first_byte >= 0)      if (has_first_char)
6761        {        {
6762        if (first_byte_caseless)        if (first_char != first_char2)
6763          while (start_match < end_subject && md->lcc[*start_match] != first_byte)          while (start_match < end_subject &&
6764                *start_match != first_char && *start_match != first_char2)
6765            start_match++;            start_match++;
6766        else        else
6767          while (start_match < end_subject && *start_match != first_byte)          while (start_match < end_subject && *start_match != first_char)
6768            start_match++;            start_match++;
6769        }        }
6770    
# Line 6310  for(;;) Line 6774  for(;;)
6774        {        {
6775        if (start_match > md->start_subject + start_offset)        if (start_match > md->start_subject + start_offset)
6776          {          {
6777  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
6778          if (utf8)          if (utf)
6779            {            {
6780            while (start_match < end_subject && !WAS_NEWLINE(start_match))            while (start_match < end_subject && !WAS_NEWLINE(start_match))
6781              {              {
6782              start_match++;              start_match++;
6783              while(start_match < end_subject && (*start_match & 0xc0) == 0x80)              ACROSSCHAR(start_match < end_subject, *start_match,
6784                start_match++;                start_match++);
6785              }              }
6786            }            }
6787          else          else
# Line 6344  for(;;) Line 6808  for(;;)
6808        while (start_match < end_subject)        while (start_match < end_subject)
6809          {          {
6810          register unsigned int c = *start_match;          register unsigned int c = *start_match;
6811    #ifndef COMPILE_PCRE8
6812            if (c > 255) c = 255;
6813    #endif
6814          if ((start_bits[c/8] & (1 << (c&7))) == 0)          if ((start_bits[c/8] & (1 << (c&7))) == 0)
6815            {            {
6816            start_match++;            start_match++;
6817  #ifdef SUPPORT_UTF8  #if defined SUPPORT_UTF && defined COMPILE_PCRE8
6818            if (utf8)            /* In non 8-bit mode, the iteration will stop for
6819              while(start_match < end_subject && (*start_match & 0xc0) == 0x80)            characters > 255 at the beginning or not stop at all. */
6820                start_match++;            if (utf)
6821                ACROSSCHAR(start_match < end_subject, *start_match,
6822                  start_match++);
6823  #endif  #endif
6824            }            }
6825          else break;          else break;
# Line 6379  for(;;) Line 6848  for(;;)
6848        break;        break;
6849        }        }
6850    
6851      /* If req_byte is set, we know that that character must appear in the      /* If req_char is set, we know that that character must appear in the
6852      subject for the match to succeed. If the first character is set, req_byte      subject for the match to succeed. If the first character is set, req_char
6853      must be later in the subject; otherwise the test starts at the match point.      must be later in the subject; otherwise the test starts at the match point.
6854      This optimization can save a huge amount of backtracking in patterns with      This optimization can save a huge amount of backtracking in patterns with
6855      nested unlimited repeats that aren't going to match. Writing separate code      nested unlimited repeats that aren't going to match. Writing separate code
# Line 6393  for(;;) Line 6862  for(;;)
6862      32-megabyte string... so we don't do this when the string is sufficiently      32-megabyte string... so we don't do this when the string is sufficiently
6863      long. */      long. */
6864    
6865      if (req_byte >= 0 && end_subject - start_match < REQ_BYTE_MAX)      if (has_req_char && end_subject - start_match < REQ_BYTE_MAX)
6866        {        {
6867        register USPTR p = start_match + ((first_byte >= 0)? 1 : 0);        register PCRE_PUCHAR p = start_match + (has_first_char? 1:0);
6868    
6869        /* We don't need to repeat the search if we haven't yet reached the        /* We don't need to repeat the search if we haven't yet reached the
6870        place we found it at last time. */        place we found it at last time. */
6871    
6872        if (p > req_byte_ptr)        if (p > req_char_ptr)
6873          {          {
6874          if (req_byte_caseless)          if (req_char != req_char2)
6875            {            {
6876            while (p < end_subject)            while (p < end_subject)
6877              {              {
6878              register int pp = *p++;              register int pp = *p++;
6879              if (pp == req_byte || pp == req_byte2) { p--; break; }              if (pp == req_char || pp == req_char2) { p--; break; }
6880              }              }
6881            }            }
6882          else          else
6883            {            {
6884            while (p < end_subject)            while (p < end_subject)
6885              {              {
6886              if (*p++ == req_byte) { p--; break; }              if (*p++ == req_char) { p--; break; }
6887              }              }
6888            }            }
6889    
# Line 6431  for(;;) Line 6900  for(;;)
6900          found it, so that we don't search again next time round the loop if          found it, so that we don't search again next time round the loop if
6901          the start hasn't passed this character yet. */          the start hasn't passed this character yet. */
6902    
6903          req_byte_ptr = p;          req_char_ptr = p;
6904          }          }
6905        }        }
6906      }      }
# Line 6456  for(;;) Line 6925  for(;;)
6925    switch(rc)    switch(rc)
6926      {      {
6927      /* If MATCH_SKIP_ARG reaches this level it means that a MARK that matched      /* If MATCH_SKIP_ARG reaches this level it means that a MARK that matched
6928      the SKIP's arg was not found. In this circumstance, Perl ignores the SKIP      the SKIP's arg was not found. In this circumstance, Perl ignores the SKIP
6929      entirely. The only way we can do that is to re-do the match at the same      entirely. The only way we can do that is to re-do the match at the same
6930      point, with a flag to force SKIP with an argument to be ignored. Just      point, with a flag to force SKIP with an argument to be ignored. Just
6931      treating this case as NOMATCH does not work because it does not check other      treating this case as NOMATCH does not work because it does not check other
6932      alternatives in patterns such as A(*SKIP:A)B|AC when the subject is AC. */      alternatives in patterns such as A(*SKIP:A)B|AC when the subject is AC. */
6933    
6934      case MATCH_SKIP_ARG:      case MATCH_SKIP_ARG:
6935      new_start_match = start_match;      new_start_match = start_match;
6936      md->ignore_skip_arg = TRUE;      md->ignore_skip_arg = TRUE;
6937      break;      break;
6938    
6939      /* SKIP passes back the next starting point explicitly, but if it is the      /* SKIP passes back the next starting point explicitly, but if it is the
6940      same as the match we have just done, treat it as NOMATCH. */      same as the match we have just done, treat it as NOMATCH. */
# Line 6486  for(;;) Line 6955  for(;;)
6955      case MATCH_THEN:      case MATCH_THEN:
6956      md->ignore_skip_arg = FALSE;      md->ignore_skip_arg = FALSE;
6957      new_start_match = start_match + 1;      new_start_match = start_match + 1;
6958  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
6959      if (utf8)      if (utf)
6960        while(new_start_match < end_subject && (*new_start_match & 0xc0) == 0x80)        ACROSSCHAR(new_start_match < end_subject, *new_start_match,
6961          new_start_match++;          new_start_match++);
6962  #endif  #endif
6963      break;      break;
6964    
# Line 6527  for(;;) Line 6996  for(;;)
6996    
6997    /* If we have just passed a CR and we are now at a LF, and the pattern does    /* If we have just passed a CR and we are now at a LF, and the pattern does
6998    not contain any explicit matches for \r or \n, and the newline option is CRLF    not contain any explicit matches for \r or \n, and the newline option is CRLF
6999    or ANY or ANYCRLF, advance the match position by one more character. */    or ANY or ANYCRLF, advance the match position by one more character. In
7000      normal matching start_match will aways be greater than the first position at
7001      this stage, but a failed *SKIP can cause a return at the same point, which is
7002      why the first test exists. */
7003    
7004    if (start_match[-1] == CHAR_CR &&    if (start_match > (PCRE_PUCHAR)subject + start_offset &&
7005          start_match[-1] == CHAR_CR &&
7006        start_match < end_subject &&        start_match < end_subject &&
7007        *start_match == CHAR_NL &&        *start_match == CHAR_NL &&
7008        (re->flags & PCRE_HASCRORLF) == 0 &&        (re->flags & PCRE_HASCRORLF) == 0 &&
# Line 6575  if (rc == MATCH_MATCH || rc == MATCH_ACC Line 7048  if (rc == MATCH_MATCH || rc == MATCH_ACC
7048        }        }
7049      if (md->end_offset_top > arg_offset_max) md->offset_overflow = TRUE;      if (md->end_offset_top > arg_offset_max) md->offset_overflow = TRUE;
7050      DPRINTF(("Freeing temporary memory\n"));      DPRINTF(("Freeing temporary memory\n"));
7051      (pcre_free)(md->offset_vector);      (PUBL(free))(md->offset_vector);
7052      }      }
7053    
7054    /* Set the return code to the number of captured strings, or 0 if there were    /* Set the return code to the number of captured strings, or 0 if there were
# Line 6597  if (rc == MATCH_MATCH || rc == MATCH_ACC Line 7070  if (rc == MATCH_MATCH || rc == MATCH_ACC
7070      {      {
7071      register int *iptr, *iend;      register int *iptr, *iend;
7072      int resetcount = 2 + re->top_bracket * 2;      int resetcount = 2 + re->top_bracket * 2;
7073      if (resetcount > offsetcount) resetcount = ocount;      if (resetcount > offsetcount) resetcount = offsetcount;
7074      iptr = offsets + md->end_offset_top;      iptr = offsets + md->end_offset_top;
7075      iend = offsets + resetcount;      iend = offsets + resetcount;
7076      while (iptr < iend) *iptr++ = -1;      while (iptr < iend) *iptr++ = -1;
# Line 6614  if (rc == MATCH_MATCH || rc == MATCH_ACC Line 7087  if (rc == MATCH_MATCH || rc == MATCH_ACC
7087      }      }
7088    
7089    /* Return MARK data if requested */    /* Return MARK data if requested */
7090    
7091    if (extra_data != NULL && (extra_data->flags & PCRE_EXTRA_MARK) != 0)    if (extra_data != NULL && (extra_data->flags & PCRE_EXTRA_MARK) != 0)
7092      *(extra_data->mark) = (unsigned char *)(md->mark);      *(extra_data->mark) = (pcre_uchar *)md->mark;
7093    DPRINTF((">>>> returning %d\n", rc));    DPRINTF((">>>> returning %d\n", rc));
7094    #ifdef NO_RECURSE
7095      release_match_heapframes(&frame_zero);
7096    #endif
7097    return rc;    return rc;
7098    }    }
7099    
# Line 6627  attempt has failed at all permitted star Line 7103  attempt has failed at all permitted star
7103  if (using_temporary_offsets)  if (using_temporary_offsets)
7104    {    {
7105    DPRINTF(("Freeing temporary memory\n"));    DPRINTF(("Freeing temporary memory\n"));
7106    (pcre_free)(md->offset_vector);    (PUBL(free))(md->offset_vector);
7107    }    }
7108    
7109  /* For anything other than nomatch or partial match, just return the code. */  /* For anything other than nomatch or partial match, just return the code. */
# Line 6635  if (using_temporary_offsets) Line 7111  if (using_temporary_offsets)
7111  if (rc != MATCH_NOMATCH && rc != PCRE_ERROR_PARTIAL)  if (rc != MATCH_NOMATCH && rc != PCRE_ERROR_PARTIAL)
7112    {    {
7113    DPRINTF((">>>> error: returning %d\n", rc));    DPRINTF((">>>> error: returning %d\n", rc));
7114    #ifdef NO_RECURSE
7115      release_match_heapframes(&frame_zero);
7116    #endif
7117    return rc;    return rc;
7118    }    }
7119    
# Line 6646  if (start_partial != NULL) Line 7125  if (start_partial != NULL)
7125    md->mark = NULL;    md->mark = NULL;
7126    if (offsetcount > 1)    if (offsetcount > 1)
7127      {      {
7128      offsets[0] = (int)(start_partial - (USPTR)subject);      offsets[0] = (int)(start_partial - (PCRE_PUCHAR)subject);
7129      offsets[1] = (int)(end_subject - (USPTR)subject);      offsets[1] = (int)(end_subject - (PCRE_PUCHAR)subject);
7130      }      }
7131    rc = PCRE_ERROR_PARTIAL;    rc = PCRE_ERROR_PARTIAL;
7132    }    }
# Line 6663  else Line 7142  else
7142  /* Return the MARK data if it has been requested. */  /* Return the MARK data if it has been requested. */
7143    
7144  if (extra_data != NULL && (extra_data->flags & PCRE_EXTRA_MARK) != 0)  if (extra_data != NULL && (extra_data->flags & PCRE_EXTRA_MARK) != 0)
7145    *(extra_data->mark) = (unsigned char *)(md->nomatch_mark);    *(extra_data->mark) = (pcre_uchar *)md->nomatch_mark;
7146    #ifdef NO_RECURSE
7147      release_match_heapframes(&frame_zero);
7148    #endif
7149  return rc;  return rc;
7150  }  }
7151    

Legend:
Removed from v.771  
changed lines
  Added in v.975

  ViewVC Help
Powered by ViewVC 1.1.5