/[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 933 by ph10, Sat Feb 25 12:18:23 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 1534  for (;;) Line 1581  for (;;)
1581    
1582        if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN) RRETURN(rrc);        if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN) RRETURN(rrc);
1583        ecode += GET(ecode, 1);        ecode += GET(ecode, 1);
1584          md->mark = save_mark;
1585        }        }
1586      while (*ecode == OP_ALT);      while (*ecode == OP_ALT);
1587    
# Line 1557  for (;;) Line 1605  for (;;)
1605    
1606      case OP_ASSERT_NOT:      case OP_ASSERT_NOT:
1607      case OP_ASSERTBACK_NOT:      case OP_ASSERTBACK_NOT:
1608        save_mark = md->mark;
1609      if (md->match_function_type == MATCH_CONDASSERT)      if (md->match_function_type == MATCH_CONDASSERT)
1610        {        {
1611        condassert = TRUE;        condassert = TRUE;
# Line 1567  for (;;) Line 1616  for (;;)
1616      do      do
1617        {        {
1618        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, NULL, RM5);        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, NULL, RM5);
1619          md->mark = save_mark;
1620        if (rrc == MATCH_MATCH || rrc == MATCH_ACCEPT) RRETURN(MATCH_NOMATCH);        if (rrc == MATCH_MATCH || rrc == MATCH_ACCEPT) RRETURN(MATCH_NOMATCH);
1621        if (rrc == MATCH_SKIP || rrc == MATCH_PRUNE || rrc == MATCH_COMMIT)        if (rrc == MATCH_SKIP || rrc == MATCH_PRUNE || rrc == MATCH_COMMIT)
1622          {          {
# Line 1593  for (;;) Line 1643  for (;;)
1643      back a number of characters, not bytes. */      back a number of characters, not bytes. */
1644    
1645      case OP_REVERSE:      case OP_REVERSE:
1646  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
1647      if (utf8)      if (utf)
1648        {        {
1649        i = GET(ecode, 1);        i = GET(ecode, 1);
1650        while (i-- > 0)        while (i-- > 0)
# Line 1625  for (;;) Line 1675  for (;;)
1675      function is able to force a failure. */      function is able to force a failure. */
1676    
1677      case OP_CALLOUT:      case OP_CALLOUT:
1678      if (pcre_callout != NULL)      if (PUBL(callout) != NULL)
1679        {        {
1680        pcre_callout_block cb;        PUBL(callout_block) cb;
1681        cb.version          = 2;   /* Version 1 of the callout block */        cb.version          = 2;   /* Version 1 of the callout block */
1682        cb.callout_number   = ecode[1];        cb.callout_number   = ecode[1];
1683        cb.offset_vector    = md->offset_vector;        cb.offset_vector    = md->offset_vector;
1684    #ifdef COMPILE_PCRE8
1685        cb.subject          = (PCRE_SPTR)md->start_subject;        cb.subject          = (PCRE_SPTR)md->start_subject;
1686    #else
1687          cb.subject          = (PCRE_SPTR16)md->start_subject;
1688    #endif
1689        cb.subject_length   = (int)(md->end_subject - md->start_subject);        cb.subject_length   = (int)(md->end_subject - md->start_subject);
1690        cb.start_match      = (int)(mstart - md->start_subject);        cb.start_match      = (int)(mstart - md->start_subject);
1691        cb.current_position = (int)(eptr - md->start_subject);        cb.current_position = (int)(eptr - md->start_subject);
# Line 1641  for (;;) Line 1695  for (;;)
1695        cb.capture_last     = md->capture_last;        cb.capture_last     = md->capture_last;
1696        cb.callout_data     = md->callout_data;        cb.callout_data     = md->callout_data;
1697        cb.mark             = md->nomatch_mark;        cb.mark             = md->nomatch_mark;
1698        if ((rrc = (*pcre_callout)(&cb)) > 0) RRETURN(MATCH_NOMATCH);        if ((rrc = (*PUBL(callout))(&cb)) > 0) RRETURN(MATCH_NOMATCH);
1699        if (rrc < 0) RRETURN(rrc);        if (rrc < 0) RRETURN(rrc);
1700        }        }
1701      ecode += 2 + 2*LINK_SIZE;      ecode += 2 + 2*LINK_SIZE;
# Line 1700  for (;;) Line 1754  for (;;)
1754        else        else
1755          {          {
1756          new_recursive.offset_save =          new_recursive.offset_save =
1757            (int *)(pcre_malloc)(new_recursive.saved_max * sizeof(int));            (int *)(PUBL(malloc))(new_recursive.saved_max * sizeof(int));
1758          if (new_recursive.offset_save == NULL) RRETURN(PCRE_ERROR_NOMEMORY);          if (new_recursive.offset_save == NULL) RRETURN(PCRE_ERROR_NOMEMORY);
1759          }          }
1760        memcpy(new_recursive.offset_save, md->offset_vector,        memcpy(new_recursive.offset_save, md->offset_vector,
# Line 1715  for (;;) Line 1769  for (;;)
1769        do        do
1770          {          {
1771          if (cbegroup) md->match_function_type = MATCH_CBEGROUP;          if (cbegroup) md->match_function_type = MATCH_CBEGROUP;
1772          RMATCH(eptr, callpat + _pcre_OP_lengths[*callpat], offset_top,          RMATCH(eptr, callpat + PRIV(OP_lengths)[*callpat], offset_top,
1773            md, eptrb, RM6);            md, eptrb, RM6);
1774          memcpy(md->offset_vector, new_recursive.offset_save,          memcpy(md->offset_vector, new_recursive.offset_save,
1775              new_recursive.saved_max * sizeof(int));              new_recursive.saved_max * sizeof(int));
# Line 1724  for (;;) Line 1778  for (;;)
1778            {            {
1779            DPRINTF(("Recursion matched\n"));            DPRINTF(("Recursion matched\n"));
1780            if (new_recursive.offset_save != stacksave)            if (new_recursive.offset_save != stacksave)
1781              (pcre_free)(new_recursive.offset_save);              (PUBL(free))(new_recursive.offset_save);
1782    
1783            /* 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
1784            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 1742  for (;;) Line 1796  for (;;)
1796            {            {
1797            DPRINTF(("Recursion gave error %d\n", rrc));            DPRINTF(("Recursion gave error %d\n", rrc));
1798            if (new_recursive.offset_save != stacksave)            if (new_recursive.offset_save != stacksave)
1799              (pcre_free)(new_recursive.offset_save);              (PUBL(free))(new_recursive.offset_save);
1800            RRETURN(rrc);            RRETURN(rrc);
1801            }            }
1802    
# Line 1754  for (;;) Line 1808  for (;;)
1808        DPRINTF(("Recursion didn't match\n"));        DPRINTF(("Recursion didn't match\n"));
1809        md->recursive = new_recursive.prevrec;        md->recursive = new_recursive.prevrec;
1810        if (new_recursive.offset_save != stacksave)        if (new_recursive.offset_save != stacksave)
1811          (pcre_free)(new_recursive.offset_save);          (PUBL(free))(new_recursive.offset_save);
1812        RRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
1813        }        }
1814    
# Line 2015  for (;;) Line 2069  for (;;)
2069    
2070      case OP_DOLLM:      case OP_DOLLM:
2071      if (eptr < md->end_subject)      if (eptr < md->end_subject)
2072        { if (!IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH); }        {
2073          if (!IS_NEWLINE(eptr))
2074            {
2075            if (md->partial != 0 &&
2076                eptr + 1 >= md->end_subject &&
2077                NLBLOCK->nltype == NLTYPE_FIXED &&
2078                NLBLOCK->nllen == 2 &&
2079                *eptr == NLBLOCK->nl[0])
2080              {
2081              md->hitend = TRUE;
2082              if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
2083              }
2084            RRETURN(MATCH_NOMATCH);
2085            }
2086          }
2087      else      else
2088        {        {
2089        if (md->noteol) RRETURN(MATCH_NOMATCH);        if (md->noteol) RRETURN(MATCH_NOMATCH);
# Line 2047  for (;;) Line 2115  for (;;)
2115      ASSERT_NL_OR_EOS:      ASSERT_NL_OR_EOS:
2116      if (eptr < md->end_subject &&      if (eptr < md->end_subject &&
2117          (!IS_NEWLINE(eptr) || eptr != md->end_subject - md->nllen))          (!IS_NEWLINE(eptr) || eptr != md->end_subject - md->nllen))
2118          {
2119          if (md->partial != 0 &&
2120              eptr + 1 >= md->end_subject &&
2121              NLBLOCK->nltype == NLTYPE_FIXED &&
2122              NLBLOCK->nllen == 2 &&
2123              *eptr == NLBLOCK->nl[0])
2124            {
2125            md->hitend = TRUE;
2126            if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
2127            }
2128        RRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2129          }
2130    
2131      /* Either at end of string or \n before end. */      /* Either at end of string or \n before end. */
2132    
# Line 2066  for (;;) Line 2145  for (;;)
2145        be "non-word" characters. Remember the earliest consulted character for        be "non-word" characters. Remember the earliest consulted character for
2146        partial matching. */        partial matching. */
2147    
2148  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
2149        if (utf8)        if (utf)
2150          {          {
2151          /* Get status of previous character */          /* Get status of previous character */
2152    
2153          if (eptr == md->start_subject) prev_is_word = FALSE; else          if (eptr == md->start_subject) prev_is_word = FALSE; else
2154            {            {
2155            USPTR lastptr = eptr - 1;            PCRE_PUCHAR lastptr = eptr - 1;
2156            while((*lastptr & 0xc0) == 0x80) lastptr--;            BACKCHAR(lastptr);
2157            if (lastptr < md->start_used_ptr) md->start_used_ptr = lastptr;            if (lastptr < md->start_used_ptr) md->start_used_ptr = lastptr;
2158            GETCHAR(c, lastptr);            GETCHAR(c, lastptr);
2159  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
# Line 2139  for (;;) Line 2218  for (;;)
2218              }              }
2219            else            else
2220  #endif  #endif
2221            prev_is_word = ((md->ctypes[eptr[-1]] & ctype_word) != 0);            prev_is_word = MAX_255(eptr[-1])
2222                && ((md->ctypes[eptr[-1]] & ctype_word) != 0);
2223            }            }
2224    
2225          /* Get status of next character */          /* Get status of next character */
# Line 2162  for (;;) Line 2242  for (;;)
2242            }            }
2243          else          else
2244  #endif  #endif
2245          cur_is_word = ((md->ctypes[*eptr] & ctype_word) != 0);          cur_is_word = MAX_255(*eptr)
2246              && ((md->ctypes[*eptr] & ctype_word) != 0);
2247          }          }
2248    
2249        /* Now see if the situation is what we want */        /* Now see if the situation is what we want */
# Line 2173  for (;;) Line 2254  for (;;)
2254        }        }
2255      break;      break;
2256    
2257      /* Match a single character type; inline for speed */      /* Match any single character type except newline; have to take care with
2258        CRLF newlines and partial matching. */
2259    
2260      case OP_ANY:      case OP_ANY:
2261      if (IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH);      if (IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH);
2262        if (md->partial != 0 &&
2263            eptr + 1 >= md->end_subject &&
2264            NLBLOCK->nltype == NLTYPE_FIXED &&
2265            NLBLOCK->nllen == 2 &&
2266            *eptr == NLBLOCK->nl[0])
2267          {
2268          md->hitend = TRUE;
2269          if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
2270          }
2271    
2272      /* Fall through */      /* Fall through */
2273    
2274        /* Match any single character whatsoever. */
2275    
2276      case OP_ALLANY:      case OP_ALLANY:
2277      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 */
2278        {                            /* not be updated before SCHECK_PARTIAL. */        {                            /* not be updated before SCHECK_PARTIAL. */
# Line 2186  for (;;) Line 2280  for (;;)
2280        RRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2281        }        }
2282      eptr++;      eptr++;
2283      if (utf8) while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;  #ifdef SUPPORT_UTF
2284        if (utf) ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);
2285    #endif
2286      ecode++;      ecode++;
2287      break;      break;
2288    
# Line 2211  for (;;) Line 2307  for (;;)
2307        }        }
2308      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2309      if (      if (
2310  #ifdef SUPPORT_UTF8  #if defined SUPPORT_UTF || !(defined COMPILE_PCRE8)
2311         c < 256 &&         c < 256 &&
2312  #endif  #endif
2313         (md->ctypes[c] & ctype_digit) != 0         (md->ctypes[c] & ctype_digit) != 0
# Line 2228  for (;;) Line 2324  for (;;)
2324        }        }
2325      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2326      if (      if (
2327  #ifdef SUPPORT_UTF8  #if defined SUPPORT_UTF || !(defined COMPILE_PCRE8)
2328         c >= 256 ||         c > 255 ||
2329  #endif  #endif
2330         (md->ctypes[c] & ctype_digit) == 0         (md->ctypes[c] & ctype_digit) == 0
2331         )         )
# Line 2245  for (;;) Line 2341  for (;;)
2341        }        }
2342      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2343      if (      if (
2344  #ifdef SUPPORT_UTF8  #if defined SUPPORT_UTF || !(defined COMPILE_PCRE8)
2345         c < 256 &&         c < 256 &&
2346  #endif  #endif
2347         (md->ctypes[c] & ctype_space) != 0         (md->ctypes[c] & ctype_space) != 0
# Line 2262  for (;;) Line 2358  for (;;)
2358        }        }
2359      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2360      if (      if (
2361  #ifdef SUPPORT_UTF8  #if defined SUPPORT_UTF || !(defined COMPILE_PCRE8)
2362         c >= 256 ||         c > 255 ||
2363  #endif  #endif
2364         (md->ctypes[c] & ctype_space) == 0         (md->ctypes[c] & ctype_space) == 0
2365         )         )
# Line 2279  for (;;) Line 2375  for (;;)
2375        }        }
2376      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2377      if (      if (
2378  #ifdef SUPPORT_UTF8  #if defined SUPPORT_UTF || !(defined COMPILE_PCRE8)
2379         c < 256 &&         c < 256 &&
2380  #endif  #endif
2381         (md->ctypes[c] & ctype_word) != 0         (md->ctypes[c] & ctype_word) != 0
# Line 2296  for (;;) Line 2392  for (;;)
2392        }        }
2393      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2394      if (      if (
2395  #ifdef SUPPORT_UTF8  #if defined SUPPORT_UTF || !(defined COMPILE_PCRE8)
2396         c >= 256 ||         c > 255 ||
2397  #endif  #endif
2398         (md->ctypes[c] & ctype_word) == 0         (md->ctypes[c] & ctype_word) == 0
2399         )         )
# Line 2317  for (;;) Line 2413  for (;;)
2413        default: RRETURN(MATCH_NOMATCH);        default: RRETURN(MATCH_NOMATCH);
2414    
2415        case 0x000d:        case 0x000d:
2416        if (eptr < md->end_subject && *eptr == 0x0a) eptr++;        if (eptr >= md->end_subject)
2417            {
2418            SCHECK_PARTIAL();
2419            }
2420          else if (*eptr == 0x0a) eptr++;
2421        break;        break;
2422    
2423        case 0x000a:        case 0x000a:
# Line 2475  for (;;) Line 2575  for (;;)
2575          break;          break;
2576    
2577          case PT_GC:          case PT_GC:
2578          if ((ecode[2] != _pcre_ucp_gentype[prop->chartype]) == (op == OP_PROP))          if ((ecode[2] != PRIV(ucp_gentype)[prop->chartype]) == (op == OP_PROP))
2579            RRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
2580          break;          break;
2581    
# Line 2492  for (;;) Line 2592  for (;;)
2592          /* These are specials */          /* These are specials */
2593    
2594          case PT_ALNUM:          case PT_ALNUM:
2595          if ((_pcre_ucp_gentype[prop->chartype] == ucp_L ||          if ((PRIV(ucp_gentype)[prop->chartype] == ucp_L ||
2596               _pcre_ucp_gentype[prop->chartype] == ucp_N) == (op == OP_NOTPROP))               PRIV(ucp_gentype)[prop->chartype] == ucp_N) == (op == OP_NOTPROP))
2597            RRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
2598          break;          break;
2599    
2600          case PT_SPACE:    /* Perl space */          case PT_SPACE:    /* Perl space */
2601          if ((_pcre_ucp_gentype[prop->chartype] == ucp_Z ||          if ((PRIV(ucp_gentype)[prop->chartype] == ucp_Z ||
2602               c == CHAR_HT || c == CHAR_NL || c == CHAR_FF || c == CHAR_CR)               c == CHAR_HT || c == CHAR_NL || c == CHAR_FF || c == CHAR_CR)
2603                 == (op == OP_NOTPROP))                 == (op == OP_NOTPROP))
2604            RRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
2605          break;          break;
2606    
2607          case PT_PXSPACE:  /* POSIX space */          case PT_PXSPACE:  /* POSIX space */
2608          if ((_pcre_ucp_gentype[prop->chartype] == ucp_Z ||          if ((PRIV(ucp_gentype)[prop->chartype] == ucp_Z ||
2609               c == CHAR_HT || c == CHAR_NL || c == CHAR_VT ||               c == CHAR_HT || c == CHAR_NL || c == CHAR_VT ||
2610               c == CHAR_FF || c == CHAR_CR)               c == CHAR_FF || c == CHAR_CR)
2611                 == (op == OP_NOTPROP))                 == (op == OP_NOTPROP))
# Line 2513  for (;;) Line 2613  for (;;)
2613          break;          break;
2614    
2615          case PT_WORD:          case PT_WORD:
2616          if ((_pcre_ucp_gentype[prop->chartype] == ucp_L ||          if ((PRIV(ucp_gentype)[prop->chartype] == ucp_L ||
2617               _pcre_ucp_gentype[prop->chartype] == ucp_N ||               PRIV(ucp_gentype)[prop->chartype] == ucp_N ||
2618               c == CHAR_UNDERSCORE) == (op == OP_NOTPROP))               c == CHAR_UNDERSCORE) == (op == OP_NOTPROP))
2619            RRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
2620          break;          break;
# Line 2543  for (;;) Line 2643  for (;;)
2643      while (eptr < md->end_subject)      while (eptr < md->end_subject)
2644        {        {
2645        int len = 1;        int len = 1;
2646        if (!utf8) c = *eptr; else { GETCHARLEN(c, eptr, len); }        if (!utf) c = *eptr; else { GETCHARLEN(c, eptr, len); }
2647        if (UCD_CATEGORY(c) != ucp_M) break;        if (UCD_CATEGORY(c) != ucp_M) break;
2648        eptr += len;        eptr += len;
2649        }        }
2650        CHECK_PARTIAL();
2651      ecode++;      ecode++;
2652      break;      break;
2653  #endif  #endif
# Line 2564  for (;;) Line 2665  for (;;)
2665      case OP_REFI:      case OP_REFI:
2666      caseless = op == OP_REFI;      caseless = op == OP_REFI;
2667      offset = GET2(ecode, 1) << 1;               /* Doubled ref number */      offset = GET2(ecode, 1) << 1;               /* Doubled ref number */
2668      ecode += 3;      ecode += 1 + IMM2_SIZE;
2669    
2670      /* If the reference is unset, there are two possibilities:      /* If the reference is unset, there are two possibilities:
2671    
# Line 2604  for (;;) Line 2705  for (;;)
2705        case OP_CRMINRANGE:        case OP_CRMINRANGE:
2706        minimize = (*ecode == OP_CRMINRANGE);        minimize = (*ecode == OP_CRMINRANGE);
2707        min = GET2(ecode, 1);        min = GET2(ecode, 1);
2708        max = GET2(ecode, 3);        max = GET2(ecode, 1 + IMM2_SIZE);
2709        if (max == 0) max = INT_MAX;        if (max == 0) max = INT_MAX;
2710        ecode += 5;        ecode += 1 + 2 * IMM2_SIZE;
2711        break;        break;
2712    
2713        default:               /* No repeat follows */        default:               /* No repeat follows */
2714        if ((length = match_ref(offset, eptr, length, md, caseless)) < 0)        if ((length = match_ref(offset, eptr, length, md, caseless)) < 0)
2715          {          {
2716            if (length == -2) eptr = md->end_subject;   /* Partial match */
2717          CHECK_PARTIAL();          CHECK_PARTIAL();
2718          RRETURN(MATCH_NOMATCH);          RRETURN(MATCH_NOMATCH);
2719          }          }
# Line 2620  for (;;) Line 2722  for (;;)
2722        }        }
2723    
2724      /* Handle repeated back references. If the length of the reference is      /* Handle repeated back references. If the length of the reference is
2725      zero, just continue with the main loop. */      zero, just continue with the main loop. If the length is negative, it
2726        means the reference is unset in non-Java-compatible mode. If the minimum is
2727        zero, we can continue at the same level without recursion. For any other
2728        minimum, carrying on will result in NOMATCH. */
2729    
2730      if (length == 0) continue;      if (length == 0) continue;
2731        if (length < 0 && min == 0) continue;
2732    
2733      /* First, ensure the minimum number of matches are present. We get back      /* First, ensure the minimum number of matches are present. We get back
2734      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 2739  for (;;)
2739        int slength;        int slength;
2740        if ((slength = match_ref(offset, eptr, length, md, caseless)) < 0)        if ((slength = match_ref(offset, eptr, length, md, caseless)) < 0)
2741          {          {
2742            if (slength == -2) eptr = md->end_subject;   /* Partial match */
2743          CHECK_PARTIAL();          CHECK_PARTIAL();
2744          RRETURN(MATCH_NOMATCH);          RRETURN(MATCH_NOMATCH);
2745          }          }
# Line 2656  for (;;) Line 2763  for (;;)
2763          if (fi >= max) RRETURN(MATCH_NOMATCH);          if (fi >= max) RRETURN(MATCH_NOMATCH);
2764          if ((slength = match_ref(offset, eptr, length, md, caseless)) < 0)          if ((slength = match_ref(offset, eptr, length, md, caseless)) < 0)
2765            {            {
2766              if (slength == -2) eptr = md->end_subject;   /* Partial match */
2767            CHECK_PARTIAL();            CHECK_PARTIAL();
2768            RRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
2769            }            }
# Line 2674  for (;;) Line 2782  for (;;)
2782          int slength;          int slength;
2783          if ((slength = match_ref(offset, eptr, length, md, caseless)) < 0)          if ((slength = match_ref(offset, eptr, length, md, caseless)) < 0)
2784            {            {
2785            CHECK_PARTIAL();            /* Can't use CHECK_PARTIAL because we don't want to update eptr in
2786              the soft partial matching case. */
2787    
2788              if (slength == -2 && md->partial != 0 &&
2789                  md->end_subject > md->start_used_ptr)
2790                {
2791                md->hitend = TRUE;
2792                if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
2793                }
2794            break;            break;
2795            }            }
2796          eptr += slength;          eptr += slength;
2797          }          }
2798    
2799        while (eptr >= pp)        while (eptr >= pp)
2800          {          {
2801          RMATCH(eptr, ecode, offset_top, md, eptrb, RM15);          RMATCH(eptr, ecode, offset_top, md, eptrb, RM15);
# Line 2703  for (;;) Line 2820  for (;;)
2820      case OP_NCLASS:      case OP_NCLASS:
2821      case OP_CLASS:      case OP_CLASS:
2822        {        {
2823          /* The data variable is saved across frames, so the byte map needs to
2824          be stored there. */
2825    #define BYTE_MAP ((pcre_uint8 *)data)
2826        data = ecode + 1;                /* Save for matching */        data = ecode + 1;                /* Save for matching */
2827        ecode += 33;                     /* Advance past the item */        ecode += 1 + (32 / sizeof(pcre_uchar)); /* Advance past the item */
2828    
2829        switch (*ecode)        switch (*ecode)
2830          {          {
# Line 2725  for (;;) Line 2845  for (;;)
2845          case OP_CRMINRANGE:          case OP_CRMINRANGE:
2846          minimize = (*ecode == OP_CRMINRANGE);          minimize = (*ecode == OP_CRMINRANGE);
2847          min = GET2(ecode, 1);          min = GET2(ecode, 1);
2848          max = GET2(ecode, 3);          max = GET2(ecode, 1 + IMM2_SIZE);
2849          if (max == 0) max = INT_MAX;          if (max == 0) max = INT_MAX;
2850          ecode += 5;          ecode += 1 + 2 * IMM2_SIZE;
2851          break;          break;
2852    
2853          default:               /* No repeat follows */          default:               /* No repeat follows */
# Line 2737  for (;;) Line 2857  for (;;)
2857    
2858        /* First, ensure the minimum number of matches are present. */        /* First, ensure the minimum number of matches are present. */
2859    
2860  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
2861        /* UTF-8 mode */        if (utf)
       if (utf8)  
2862          {          {
2863          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
2864            {            {
# Line 2754  for (;;) Line 2873  for (;;)
2873              if (op == OP_CLASS) RRETURN(MATCH_NOMATCH);              if (op == OP_CLASS) RRETURN(MATCH_NOMATCH);
2874              }              }
2875            else            else
2876              {              if ((BYTE_MAP[c/8] & (1 << (c&7))) == 0) RRETURN(MATCH_NOMATCH);
             if ((data[c/8] & (1 << (c&7))) == 0) RRETURN(MATCH_NOMATCH);  
             }  
2877            }            }
2878          }          }
2879        else        else
2880  #endif  #endif
2881        /* Not UTF-8 mode */        /* Not UTF mode */
2882          {          {
2883          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
2884            {            {
# Line 2771  for (;;) Line 2888  for (;;)
2888              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
2889              }              }
2890            c = *eptr++;            c = *eptr++;
2891            if ((data[c/8] & (1 << (c&7))) == 0) RRETURN(MATCH_NOMATCH);  #ifndef COMPILE_PCRE8
2892              if (c > 255)
2893                {
2894                if (op == OP_CLASS) RRETURN(MATCH_NOMATCH);
2895                }
2896              else
2897    #endif
2898                if ((BYTE_MAP[c/8] & (1 << (c&7))) == 0) RRETURN(MATCH_NOMATCH);
2899            }            }
2900          }          }
2901    
# Line 2785  for (;;) Line 2909  for (;;)
2909    
2910        if (minimize)        if (minimize)
2911          {          {
2912  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
2913          /* UTF-8 mode */          if (utf)
         if (utf8)  
2914            {            {
2915            for (fi = min;; fi++)            for (fi = min;; fi++)
2916              {              {
# Line 2805  for (;;) Line 2928  for (;;)
2928                if (op == OP_CLASS) RRETURN(MATCH_NOMATCH);                if (op == OP_CLASS) RRETURN(MATCH_NOMATCH);
2929                }                }
2930              else              else
2931                {                if ((BYTE_MAP[c/8] & (1 << (c&7))) == 0) RRETURN(MATCH_NOMATCH);
               if ((data[c/8] & (1 << (c&7))) == 0) RRETURN(MATCH_NOMATCH);  
               }  
2932              }              }
2933            }            }
2934          else          else
2935  #endif  #endif
2936          /* Not UTF-8 mode */          /* Not UTF mode */
2937            {            {
2938            for (fi = min;; fi++)            for (fi = min;; fi++)
2939              {              {
# Line 2825  for (;;) Line 2946  for (;;)
2946                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
2947                }                }
2948              c = *eptr++;              c = *eptr++;
2949              if ((data[c/8] & (1 << (c&7))) == 0) RRETURN(MATCH_NOMATCH);  #ifndef COMPILE_PCRE8
2950                if (c > 255)
2951                  {
2952                  if (op == OP_CLASS) RRETURN(MATCH_NOMATCH);
2953                  }
2954                else
2955    #endif
2956                  if ((BYTE_MAP[c/8] & (1 << (c&7))) == 0) RRETURN(MATCH_NOMATCH);
2957              }              }
2958            }            }
2959          /* Control never gets here */          /* Control never gets here */
# Line 2837  for (;;) Line 2965  for (;;)
2965          {          {
2966          pp = eptr;          pp = eptr;
2967    
2968  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
2969          /* UTF-8 mode */          if (utf)
         if (utf8)  
2970            {            {
2971            for (i = min; i < max; i++)            for (i = min; i < max; i++)
2972              {              {
# Line 2855  for (;;) Line 2982  for (;;)
2982                if (op == OP_CLASS) break;                if (op == OP_CLASS) break;
2983                }                }
2984              else              else
2985                {                if ((BYTE_MAP[c/8] & (1 << (c&7))) == 0) break;
               if ((data[c/8] & (1 << (c&7))) == 0) break;  
               }  
2986              eptr += len;              eptr += len;
2987              }              }
2988            for (;;)            for (;;)
# Line 2870  for (;;) Line 2995  for (;;)
2995            }            }
2996          else          else
2997  #endif  #endif
2998            /* Not UTF-8 mode */            /* Not UTF mode */
2999            {            {
3000            for (i = min; i < max; i++)            for (i = min; i < max; i++)
3001              {              {
# Line 2880  for (;;) Line 3005  for (;;)
3005                break;                break;
3006                }                }
3007              c = *eptr;              c = *eptr;
3008              if ((data[c/8] & (1 << (c&7))) == 0) break;  #ifndef COMPILE_PCRE8
3009                if (c > 255)
3010                  {
3011                  if (op == OP_CLASS) break;
3012                  }
3013                else
3014    #endif
3015                  if ((BYTE_MAP[c/8] & (1 << (c&7))) == 0) break;
3016              eptr++;              eptr++;
3017              }              }
3018            while (eptr >= pp)            while (eptr >= pp)
# Line 2893  for (;;) Line 3025  for (;;)
3025    
3026          RRETURN(MATCH_NOMATCH);          RRETURN(MATCH_NOMATCH);
3027          }          }
3028    #undef BYTE_MAP
3029        }        }
3030      /* Control never gets here */      /* Control never gets here */
3031    
# Line 2901  for (;;) Line 3034  for (;;)
3034      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
3035      mode, because Unicode properties are supported in non-UTF-8 mode. */      mode, because Unicode properties are supported in non-UTF-8 mode. */
3036    
3037  #ifdef SUPPORT_UTF8  #if defined SUPPORT_UTF || !defined COMPILE_PCRE8
3038      case OP_XCLASS:      case OP_XCLASS:
3039        {        {
3040        data = ecode + 1 + LINK_SIZE;                /* Save for matching */        data = ecode + 1 + LINK_SIZE;                /* Save for matching */
# Line 2926  for (;;) Line 3059  for (;;)
3059          case OP_CRMINRANGE:          case OP_CRMINRANGE:
3060          minimize = (*ecode == OP_CRMINRANGE);          minimize = (*ecode == OP_CRMINRANGE);
3061          min = GET2(ecode, 1);          min = GET2(ecode, 1);
3062          max = GET2(ecode, 3);          max = GET2(ecode, 1 + IMM2_SIZE);
3063          if (max == 0) max = INT_MAX;          if (max == 0) max = INT_MAX;
3064          ecode += 5;          ecode += 1 + 2 * IMM2_SIZE;
3065          break;          break;
3066    
3067          default:               /* No repeat follows */          default:               /* No repeat follows */
# Line 2946  for (;;) Line 3079  for (;;)
3079            RRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
3080            }            }
3081          GETCHARINCTEST(c, eptr);          GETCHARINCTEST(c, eptr);
3082          if (!_pcre_xclass(c, data)) RRETURN(MATCH_NOMATCH);          if (!PRIV(xclass)(c, data, utf)) RRETURN(MATCH_NOMATCH);
3083          }          }
3084    
3085        /* 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 3103  for (;;)
3103              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
3104              }              }
3105            GETCHARINCTEST(c, eptr);            GETCHARINCTEST(c, eptr);
3106            if (!_pcre_xclass(c, data)) RRETURN(MATCH_NOMATCH);            if (!PRIV(xclass)(c, data, utf)) RRETURN(MATCH_NOMATCH);
3107            }            }
3108          /* Control never gets here */          /* Control never gets here */
3109          }          }
# Line 2988  for (;;) Line 3121  for (;;)
3121              SCHECK_PARTIAL();              SCHECK_PARTIAL();
3122              break;              break;
3123              }              }
3124    #ifdef SUPPORT_UTF
3125            GETCHARLENTEST(c, eptr, len);            GETCHARLENTEST(c, eptr, len);
3126            if (!_pcre_xclass(c, data)) break;  #else
3127              c = *eptr;
3128    #endif
3129              if (!PRIV(xclass)(c, data, utf)) break;
3130            eptr += len;            eptr += len;
3131            }            }
3132          for(;;)          for(;;)
# Line 2997  for (;;) Line 3134  for (;;)
3134            RMATCH(eptr, ecode, offset_top, md, eptrb, RM21);            RMATCH(eptr, ecode, offset_top, md, eptrb, RM21);
3135            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3136            if (eptr-- == pp) break;        /* Stop if tried at original pos */            if (eptr-- == pp) break;        /* Stop if tried at original pos */
3137            if (utf8) BACKCHAR(eptr);  #ifdef SUPPORT_UTF
3138              if (utf) BACKCHAR(eptr);
3139    #endif
3140            }            }
3141          RRETURN(MATCH_NOMATCH);          RRETURN(MATCH_NOMATCH);
3142          }          }
# Line 3009  for (;;) Line 3148  for (;;)
3148      /* Match a single character, casefully */      /* Match a single character, casefully */
3149    
3150      case OP_CHAR:      case OP_CHAR:
3151  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
3152      if (utf8)      if (utf)
3153        {        {
3154        length = 1;        length = 1;
3155        ecode++;        ecode++;
# Line 3024  for (;;) Line 3163  for (;;)
3163        }        }
3164      else      else
3165  #endif  #endif
3166        /* Not UTF mode */
     /* Non-UTF-8 mode */  
3167        {        {
3168        if (md->end_subject - eptr < 1)        if (md->end_subject - eptr < 1)
3169          {          {
# Line 3037  for (;;) Line 3175  for (;;)
3175        }        }
3176      break;      break;
3177    
3178      /* Match a single character, caselessly */      /* Match a single character, caselessly. If we are at the end of the
3179        subject, give up immediately. */
3180    
3181      case OP_CHARI:      case OP_CHARI:
3182  #ifdef SUPPORT_UTF8      if (eptr >= md->end_subject)
3183      if (utf8)        {
3184          SCHECK_PARTIAL();
3185          RRETURN(MATCH_NOMATCH);
3186          }
3187    
3188    #ifdef SUPPORT_UTF
3189        if (utf)
3190        {        {
3191        length = 1;        length = 1;
3192        ecode++;        ecode++;
3193        GETCHARLEN(fc, ecode, length);        GETCHARLEN(fc, ecode, length);
3194    
       if (length > md->end_subject - eptr)  
         {  
         CHECK_PARTIAL();             /* Not SCHECK_PARTIAL() */  
         RRETURN(MATCH_NOMATCH);  
         }  
   
3195        /* 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
3196        can use the fast lookup table. */        we know that its other case must also be one byte long, so we can use the
3197          fast lookup table. We know that there is at least one byte left in the
3198          subject. */
3199    
3200        if (fc < 128)        if (fc < 128)
3201          {          {
3202          if (md->lcc[*ecode++] != md->lcc[*eptr++]) RRETURN(MATCH_NOMATCH);          if (md->lcc[fc]
3203                != TABLE_GET(*eptr, md->lcc, *eptr)) RRETURN(MATCH_NOMATCH);
3204            ecode++;
3205            eptr++;
3206          }          }
3207    
3208        /* Otherwise we must pick up the subject character */        /* Otherwise we must pick up the subject character. Note that we cannot
3209          use the value of "length" to check for sufficient bytes left, because the
3210          other case of the character may have more or fewer bytes.  */
3211    
3212        else        else
3213          {          {
# Line 3082  for (;;) Line 3228  for (;;)
3228          }          }
3229        }        }
3230      else      else
3231  #endif   /* SUPPORT_UTF8 */  #endif   /* SUPPORT_UTF */
3232    
3233      /* Non-UTF-8 mode */      /* Not UTF mode */
3234        {        {
3235        if (md->end_subject - eptr < 1)        if (TABLE_GET(ecode[1], md->lcc, ecode[1])
3236          {            != TABLE_GET(*eptr, md->lcc, *eptr)) RRETURN(MATCH_NOMATCH);
3237          SCHECK_PARTIAL();            /* This one can use SCHECK_PARTIAL() */        eptr++;
         RRETURN(MATCH_NOMATCH);  
         }  
       if (md->lcc[ecode[1]] != md->lcc[*eptr++]) RRETURN(MATCH_NOMATCH);  
3238        ecode += 2;        ecode += 2;
3239        }        }
3240      break;      break;
# Line 3101  for (;;) Line 3244  for (;;)
3244      case OP_EXACT:      case OP_EXACT:
3245      case OP_EXACTI:      case OP_EXACTI:
3246      min = max = GET2(ecode, 1);      min = max = GET2(ecode, 1);
3247      ecode += 3;      ecode += 1 + IMM2_SIZE;
3248      goto REPEATCHAR;      goto REPEATCHAR;
3249    
3250      case OP_POSUPTO:      case OP_POSUPTO:
# Line 3116  for (;;) Line 3259  for (;;)
3259      min = 0;      min = 0;
3260      max = GET2(ecode, 1);      max = GET2(ecode, 1);
3261      minimize = *ecode == OP_MINUPTO || *ecode == OP_MINUPTOI;      minimize = *ecode == OP_MINUPTO || *ecode == OP_MINUPTOI;
3262      ecode += 3;      ecode += 1 + IMM2_SIZE;
3263      goto REPEATCHAR;      goto REPEATCHAR;
3264    
3265      case OP_POSSTAR:      case OP_POSSTAR:
# Line 3164  for (;;) Line 3307  for (;;)
3307      /* Common code for all repeated single-character matches. */      /* Common code for all repeated single-character matches. */
3308    
3309      REPEATCHAR:      REPEATCHAR:
3310  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
3311      if (utf8)      if (utf)
3312        {        {
3313        length = 1;        length = 1;
3314        charptr = ecode;        charptr = ecode;
# Line 3181  for (;;) Line 3324  for (;;)
3324          unsigned int othercase;          unsigned int othercase;
3325          if (op >= OP_STARI &&     /* Caseless */          if (op >= OP_STARI &&     /* Caseless */
3326              (othercase = UCD_OTHERCASE(fc)) != fc)              (othercase = UCD_OTHERCASE(fc)) != fc)
3327            oclength = _pcre_ord2utf8(othercase, occhars);            oclength = PRIV(ord2utf)(othercase, occhars);
3328          else oclength = 0;          else oclength = 0;
3329  #endif  /* SUPPORT_UCP */  #endif  /* SUPPORT_UCP */
3330    
3331          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
3332            {            {
3333            if (eptr <= md->end_subject - length &&            if (eptr <= md->end_subject - length &&
3334              memcmp(eptr, charptr, length) == 0) eptr += length;              memcmp(eptr, charptr, IN_UCHARS(length)) == 0) eptr += length;
3335  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
3336            else if (oclength > 0 &&            else if (oclength > 0 &&
3337                     eptr <= md->end_subject - oclength &&                     eptr <= md->end_subject - oclength &&
3338                     memcmp(eptr, occhars, oclength) == 0) eptr += oclength;                     memcmp(eptr, occhars, IN_UCHARS(oclength)) == 0) eptr += oclength;
3339  #endif  /* SUPPORT_UCP */  #endif  /* SUPPORT_UCP */
3340            else            else
3341              {              {
# Line 3211  for (;;) Line 3354  for (;;)
3354              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3355              if (fi >= max) RRETURN(MATCH_NOMATCH);              if (fi >= max) RRETURN(MATCH_NOMATCH);
3356              if (eptr <= md->end_subject - length &&              if (eptr <= md->end_subject - length &&
3357                memcmp(eptr, charptr, length) == 0) eptr += length;                memcmp(eptr, charptr, IN_UCHARS(length)) == 0) eptr += length;
3358  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
3359              else if (oclength > 0 &&              else if (oclength > 0 &&
3360                       eptr <= md->end_subject - oclength &&                       eptr <= md->end_subject - oclength &&
3361                       memcmp(eptr, occhars, oclength) == 0) eptr += oclength;                       memcmp(eptr, occhars, IN_UCHARS(oclength)) == 0) eptr += oclength;
3362  #endif  /* SUPPORT_UCP */  #endif  /* SUPPORT_UCP */
3363              else              else
3364                {                {
# Line 3232  for (;;) Line 3375  for (;;)
3375            for (i = min; i < max; i++)            for (i = min; i < max; i++)
3376              {              {
3377              if (eptr <= md->end_subject - length &&              if (eptr <= md->end_subject - length &&
3378                  memcmp(eptr, charptr, length) == 0) eptr += length;                  memcmp(eptr, charptr, IN_UCHARS(length)) == 0) eptr += length;
3379  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
3380              else if (oclength > 0 &&              else if (oclength > 0 &&
3381                       eptr <= md->end_subject - oclength &&                       eptr <= md->end_subject - oclength &&
3382                       memcmp(eptr, occhars, oclength) == 0) eptr += oclength;                       memcmp(eptr, occhars, IN_UCHARS(oclength)) == 0) eptr += oclength;
3383  #endif  /* SUPPORT_UCP */  #endif  /* SUPPORT_UCP */
3384              else              else
3385                {                {
# Line 3268  for (;;) Line 3411  for (;;)
3411        value of fc will always be < 128. */        value of fc will always be < 128. */
3412        }        }
3413      else      else
3414  #endif  /* SUPPORT_UTF8 */  #endif  /* SUPPORT_UTF */
3415          /* When not in UTF-8 mode, load a single-byte character. */
3416      /* When not in UTF-8 mode, load a single-byte character. */        fc = *ecode++;
3417    
3418      fc = *ecode++;      /* The value of fc at this point is always one character, though we may
3419        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  
3420      caseful cases, for speed, since matching characters is likely to be quite      caseful cases, for speed, since matching characters is likely to be quite
3421      common. First, ensure the minimum number of matches are present. If min =      common. First, ensure the minimum number of matches are present. If min =
3422      max, continue at the same level without recursing. Otherwise, if      max, continue at the same level without recursing. Otherwise, if
# Line 3288  for (;;) Line 3429  for (;;)
3429    
3430      if (op >= OP_STARI)  /* Caseless */      if (op >= OP_STARI)  /* Caseless */
3431        {        {
3432        fc = md->lcc[fc];  #ifdef COMPILE_PCRE8
3433          /* fc must be < 128 if UTF is enabled. */
3434          foc = md->fcc[fc];
3435    #else
3436    #ifdef SUPPORT_UTF
3437    #ifdef SUPPORT_UCP
3438          if (utf && fc > 127)
3439            foc = UCD_OTHERCASE(fc);
3440    #else
3441          if (utf && fc > 127)
3442            foc = fc;
3443    #endif /* SUPPORT_UCP */
3444          else
3445    #endif /* SUPPORT_UTF */
3446            foc = TABLE_GET(fc, md->fcc, fc);
3447    #endif /* COMPILE_PCRE8 */
3448    
3449        for (i = 1; i <= min; i++)        for (i = 1; i <= min; i++)
3450          {          {
3451          if (eptr >= md->end_subject)          if (eptr >= md->end_subject)
# Line 3296  for (;;) Line 3453  for (;;)
3453            SCHECK_PARTIAL();            SCHECK_PARTIAL();
3454            RRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
3455            }            }
3456          if (fc != md->lcc[*eptr++]) RRETURN(MATCH_NOMATCH);          if (fc != *eptr && foc != *eptr) RRETURN(MATCH_NOMATCH);
3457            eptr++;
3458          }          }
3459        if (min == max) continue;        if (min == max) continue;
3460        if (minimize)        if (minimize)
# Line 3311  for (;;) Line 3469  for (;;)
3469              SCHECK_PARTIAL();              SCHECK_PARTIAL();
3470              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
3471              }              }
3472            if (fc != md->lcc[*eptr++]) RRETURN(MATCH_NOMATCH);            if (fc != *eptr && foc != *eptr) RRETURN(MATCH_NOMATCH);
3473              eptr++;
3474            }            }
3475          /* Control never gets here */          /* Control never gets here */
3476          }          }
# Line 3325  for (;;) Line 3484  for (;;)
3484              SCHECK_PARTIAL();              SCHECK_PARTIAL();
3485              break;              break;
3486              }              }
3487            if (fc != md->lcc[*eptr]) break;            if (fc != *eptr && foc != *eptr) break;
3488            eptr++;            eptr++;
3489            }            }
3490    
# Line 3410  for (;;) Line 3569  for (;;)
3569        SCHECK_PARTIAL();        SCHECK_PARTIAL();
3570        RRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
3571        }        }
3572      ecode++;  #ifdef SUPPORT_UTF
3573      GETCHARINCTEST(c, eptr);      if (utf)
     if (op == OP_NOTI)         /* The caseless case */  
3574        {        {
3575  #ifdef SUPPORT_UTF8        register unsigned int ch, och;
3576        if (c < 256)  
3577  #endif        ecode++;
3578        c = md->lcc[c];        GETCHARINC(ch, ecode);
3579        if (md->lcc[*ecode++] == c) RRETURN(MATCH_NOMATCH);        GETCHARINC(c, eptr);
3580    
3581          if (op == OP_NOT)
3582            {
3583            if (ch == c) RRETURN(MATCH_NOMATCH);
3584            }
3585          else
3586            {
3587    #ifdef SUPPORT_UCP
3588            if (ch > 127)
3589              och = UCD_OTHERCASE(ch);
3590    #else
3591            if (ch > 127)
3592              och = ch;
3593    #endif /* SUPPORT_UCP */
3594            else
3595              och = TABLE_GET(ch, md->fcc, ch);
3596            if (ch == c || och == c) RRETURN(MATCH_NOMATCH);
3597            }
3598        }        }
3599      else    /* Caseful */      else
3600    #endif
3601        {        {
3602        if (*ecode++ == c) RRETURN(MATCH_NOMATCH);        register unsigned int ch = ecode[1];
3603          c = *eptr++;
3604          if (ch == c || (op == OP_NOTI && TABLE_GET(ch, md->fcc, ch) == c))
3605            RRETURN(MATCH_NOMATCH);
3606          ecode += 2;
3607        }        }
3608      break;      break;
3609    
# Line 3436  for (;;) Line 3617  for (;;)
3617      case OP_NOTEXACT:      case OP_NOTEXACT:
3618      case OP_NOTEXACTI:      case OP_NOTEXACTI:
3619      min = max = GET2(ecode, 1);      min = max = GET2(ecode, 1);
3620      ecode += 3;      ecode += 1 + IMM2_SIZE;
3621      goto REPEATNOTCHAR;      goto REPEATNOTCHAR;
3622    
3623      case OP_NOTUPTO:      case OP_NOTUPTO:
# Line 3446  for (;;) Line 3627  for (;;)
3627      min = 0;      min = 0;
3628      max = GET2(ecode, 1);      max = GET2(ecode, 1);
3629      minimize = *ecode == OP_NOTMINUPTO || *ecode == OP_NOTMINUPTOI;      minimize = *ecode == OP_NOTMINUPTO || *ecode == OP_NOTMINUPTOI;
3630      ecode += 3;      ecode += 1 + IMM2_SIZE;
3631      goto REPEATNOTCHAR;      goto REPEATNOTCHAR;
3632    
3633      case OP_NOTPOSSTAR:      case OP_NOTPOSSTAR:
# Line 3478  for (;;) Line 3659  for (;;)
3659      possessive = TRUE;      possessive = TRUE;
3660      min = 0;      min = 0;
3661      max = GET2(ecode, 1);      max = GET2(ecode, 1);
3662      ecode += 3;      ecode += 1 + IMM2_SIZE;
3663      goto REPEATNOTCHAR;      goto REPEATNOTCHAR;
3664    
3665      case OP_NOTSTAR:      case OP_NOTSTAR:
# Line 3502  for (;;) Line 3683  for (;;)
3683      /* Common code for all repeated single-byte matches. */      /* Common code for all repeated single-byte matches. */
3684    
3685      REPEATNOTCHAR:      REPEATNOTCHAR:
3686      fc = *ecode++;      GETCHARINCTEST(fc, ecode);
3687    
3688      /* The code is duplicated for the caseless and caseful cases, for speed,      /* The code is duplicated for the caseless and caseful cases, for speed,
3689      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 3698  for (;;)
3698    
3699      if (op >= OP_NOTSTARI)     /* Caseless */      if (op >= OP_NOTSTARI)     /* Caseless */
3700        {        {
3701        fc = md->lcc[fc];  #ifdef SUPPORT_UTF
3702    #ifdef SUPPORT_UCP
3703          if (utf && fc > 127)
3704            foc = UCD_OTHERCASE(fc);
3705    #else
3706          if (utf && fc > 127)
3707            foc = fc;
3708    #endif /* SUPPORT_UCP */
3709          else
3710    #endif /* SUPPORT_UTF */
3711            foc = TABLE_GET(fc, md->fcc, fc);
3712    
3713  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
3714        /* UTF-8 mode */        if (utf)
       if (utf8)  
3715          {          {
3716          register unsigned int d;          register unsigned int d;
3717          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
# Line 3532  for (;;) Line 3722  for (;;)
3722              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
3723              }              }
3724            GETCHARINC(d, eptr);            GETCHARINC(d, eptr);
3725            if (d < 256) d = md->lcc[d];            if (fc == d || (unsigned int)foc == d) RRETURN(MATCH_NOMATCH);
           if (fc == d) RRETURN(MATCH_NOMATCH);  
3726            }            }
3727          }          }
3728        else        else
3729  #endif  #endif
3730          /* Not UTF mode */
       /* Not UTF-8 mode */  
3731          {          {
3732          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
3733            {            {
# Line 3548  for (;;) Line 3736  for (;;)
3736              SCHECK_PARTIAL();              SCHECK_PARTIAL();
3737              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
3738              }              }
3739            if (fc == md->lcc[*eptr++]) RRETURN(MATCH_NOMATCH);            if (fc == *eptr || foc == *eptr) RRETURN(MATCH_NOMATCH);
3740              eptr++;
3741            }            }
3742          }          }
3743    
# Line 3556  for (;;) Line 3745  for (;;)
3745    
3746        if (minimize)        if (minimize)
3747          {          {
3748  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
3749          /* UTF-8 mode */          if (utf)
         if (utf8)  
3750            {            {
3751            register unsigned int d;            register unsigned int d;
3752            for (fi = min;; fi++)            for (fi = min;; fi++)
# Line 3572  for (;;) Line 3760  for (;;)
3760                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
3761                }                }
3762              GETCHARINC(d, eptr);              GETCHARINC(d, eptr);
3763              if (d < 256) d = md->lcc[d];              if (fc == d || (unsigned int)foc == d) RRETURN(MATCH_NOMATCH);
             if (fc == d) RRETURN(MATCH_NOMATCH);  
3764              }              }
3765            }            }
3766          else          else
3767  #endif  #endif
3768          /* Not UTF-8 mode */          /* Not UTF mode */
3769            {            {
3770            for (fi = min;; fi++)            for (fi = min;; fi++)
3771              {              {
# Line 3590  for (;;) Line 3777  for (;;)
3777                SCHECK_PARTIAL();                SCHECK_PARTIAL();
3778                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
3779                }                }
3780              if (fc == md->lcc[*eptr++]) RRETURN(MATCH_NOMATCH);              if (fc == *eptr || foc == *eptr) RRETURN(MATCH_NOMATCH);
3781                eptr++;
3782              }              }
3783            }            }
3784          /* Control never gets here */          /* Control never gets here */
# Line 3602  for (;;) Line 3790  for (;;)
3790          {          {
3791          pp = eptr;          pp = eptr;
3792    
3793  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
3794          /* UTF-8 mode */          if (utf)
         if (utf8)  
3795            {            {
3796            register unsigned int d;            register unsigned int d;
3797            for (i = min; i < max; i++)            for (i = min; i < max; i++)
# Line 3616  for (;;) Line 3803  for (;;)
3803                break;                break;
3804                }                }
3805              GETCHARLEN(d, eptr, len);              GETCHARLEN(d, eptr, len);
3806              if (d < 256) d = md->lcc[d];              if (fc == d || (unsigned int)foc == d) break;
             if (fc == d) break;  
3807              eptr += len;              eptr += len;
3808              }              }
3809          if (possessive) continue;            if (possessive) continue;
3810          for(;;)            for(;;)
3811              {              {
3812              RMATCH(eptr, ecode, offset_top, md, eptrb, RM30);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM30);
3813              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
# Line 3631  for (;;) Line 3817  for (;;)
3817            }            }
3818          else          else
3819  #endif  #endif
3820          /* Not UTF-8 mode */          /* Not UTF mode */
3821            {            {
3822            for (i = min; i < max; i++)            for (i = min; i < max; i++)
3823              {              {
# Line 3640  for (;;) Line 3826  for (;;)
3826                SCHECK_PARTIAL();                SCHECK_PARTIAL();
3827                break;                break;
3828                }                }
3829              if (fc == md->lcc[*eptr]) break;              if (fc == *eptr || foc == *eptr) break;
3830              eptr++;              eptr++;
3831              }              }
3832            if (possessive) continue;            if (possessive) continue;
# Line 3661  for (;;) Line 3847  for (;;)
3847    
3848      else      else
3849        {        {
3850  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
3851        /* UTF-8 mode */        if (utf)
       if (utf8)  
3852          {          {
3853          register unsigned int d;          register unsigned int d;
3854          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
# Line 3679  for (;;) Line 3864  for (;;)
3864          }          }
3865        else        else
3866  #endif  #endif
3867        /* Not UTF-8 mode */        /* Not UTF mode */
3868          {          {
3869          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
3870            {            {
# Line 3696  for (;;) Line 3881  for (;;)
3881    
3882        if (minimize)        if (minimize)
3883          {          {
3884  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
3885          /* UTF-8 mode */          if (utf)
         if (utf8)  
3886            {            {
3887            register unsigned int d;            register unsigned int d;
3888            for (fi = min;; fi++)            for (fi = min;; fi++)
# Line 3717  for (;;) Line 3901  for (;;)
3901            }            }
3902          else          else
3903  #endif  #endif
3904          /* Not UTF-8 mode */          /* Not UTF mode */
3905            {            {
3906            for (fi = min;; fi++)            for (fi = min;; fi++)
3907              {              {
# Line 3741  for (;;) Line 3925  for (;;)
3925          {          {
3926          pp = eptr;          pp = eptr;
3927    
3928  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
3929          /* UTF-8 mode */          if (utf)
         if (utf8)  
3930            {            {
3931            register unsigned int d;            register unsigned int d;
3932            for (i = min; i < max; i++)            for (i = min; i < max; i++)
# Line 3769  for (;;) Line 3952  for (;;)
3952            }            }
3953          else          else
3954  #endif  #endif
3955          /* Not UTF-8 mode */          /* Not UTF mode */
3956            {            {
3957            for (i = min; i < max; i++)            for (i = min; i < max; i++)
3958              {              {
# Line 3802  for (;;) Line 3985  for (;;)
3985      case OP_TYPEEXACT:      case OP_TYPEEXACT:
3986      min = max = GET2(ecode, 1);      min = max = GET2(ecode, 1);
3987      minimize = TRUE;      minimize = TRUE;
3988      ecode += 3;      ecode += 1 + IMM2_SIZE;
3989      goto REPEATTYPE;      goto REPEATTYPE;
3990    
3991      case OP_TYPEUPTO:      case OP_TYPEUPTO:
# Line 3810  for (;;) Line 3993  for (;;)
3993      min = 0;      min = 0;
3994      max = GET2(ecode, 1);      max = GET2(ecode, 1);
3995      minimize = *ecode == OP_TYPEMINUPTO;      minimize = *ecode == OP_TYPEMINUPTO;
3996      ecode += 3;      ecode += 1 + IMM2_SIZE;
3997      goto REPEATTYPE;      goto REPEATTYPE;
3998    
3999      case OP_TYPEPOSSTAR:      case OP_TYPEPOSSTAR:
# Line 3838  for (;;) Line 4021  for (;;)
4021      possessive = TRUE;      possessive = TRUE;
4022      min = 0;      min = 0;
4023      max = GET2(ecode, 1);      max = GET2(ecode, 1);
4024      ecode += 3;      ecode += 1 + IMM2_SIZE;
4025      goto REPEATTYPE;      goto REPEATTYPE;
4026    
4027      case OP_TYPESTAR:      case OP_TYPESTAR:
# Line 4045  for (;;) Line 4228  for (;;)
4228            while (eptr < md->end_subject)            while (eptr < md->end_subject)
4229              {              {
4230              int len = 1;              int len = 1;
4231              if (!utf8) c = *eptr; else { GETCHARLEN(c, eptr, len); }              if (!utf) c = *eptr; else { GETCHARLEN(c, eptr, len); }
4232              if (UCD_CATEGORY(c) != ucp_M) break;              if (UCD_CATEGORY(c) != ucp_M) break;
4233              eptr += len;              eptr += len;
4234              }              }
4235              CHECK_PARTIAL();
4236            }            }
4237          }          }
4238    
# Line 4057  for (;;) Line 4241  for (;;)
4241    
4242  /* Handle all other cases when the coding is UTF-8 */  /* Handle all other cases when the coding is UTF-8 */
4243    
4244  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
4245        if (utf8) switch(ctype)        if (utf) switch(ctype)
4246          {          {
4247          case OP_ANY:          case OP_ANY:
4248          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
# Line 4069  for (;;) Line 4253  for (;;)
4253              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4254              }              }
4255            if (IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH);            if (IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH);
4256              if (md->partial != 0 &&
4257                  eptr + 1 >= md->end_subject &&
4258                  NLBLOCK->nltype == NLTYPE_FIXED &&
4259                  NLBLOCK->nllen == 2 &&
4260                  *eptr == NLBLOCK->nl[0])
4261                {
4262                md->hitend = TRUE;
4263                if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
4264                }
4265            eptr++;            eptr++;
4266            while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;            ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);
4267            }            }
4268          break;          break;
4269    
# Line 4083  for (;;) Line 4276  for (;;)
4276              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4277              }              }
4278            eptr++;            eptr++;
4279            while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;            ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);
4280            }            }
4281          break;          break;
4282    
# Line 4265  for (;;) Line 4458  for (;;)
4458              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4459              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4460              }              }
4461            if (*eptr >= 128 || (md->ctypes[*eptr++] & ctype_digit) == 0)            if (*eptr >= 128 || (md->ctypes[*eptr] & ctype_digit) == 0)
4462              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4463              eptr++;
4464            /* 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 */
4465            }            }
4466          break;          break;
# Line 4281  for (;;) Line 4475  for (;;)
4475              }              }
4476            if (*eptr < 128 && (md->ctypes[*eptr] & ctype_space) != 0)            if (*eptr < 128 && (md->ctypes[*eptr] & ctype_space) != 0)
4477              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4478            while (++eptr < md->end_subject && (*eptr & 0xc0) == 0x80);            eptr++;
4479              ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);
4480            }            }
4481          break;          break;
4482    
# Line 4293  for (;;) Line 4488  for (;;)
4488              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4489              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4490              }              }
4491            if (*eptr >= 128 || (md->ctypes[*eptr++] & ctype_space) == 0)            if (*eptr >= 128 || (md->ctypes[*eptr] & ctype_space) == 0)
4492              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4493              eptr++;
4494            /* 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 */
4495            }            }
4496          break;          break;
# Line 4309  for (;;) Line 4505  for (;;)
4505              }              }
4506            if (*eptr < 128 && (md->ctypes[*eptr] & ctype_word) != 0)            if (*eptr < 128 && (md->ctypes[*eptr] & ctype_word) != 0)
4507              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4508            while (++eptr < md->end_subject && (*eptr & 0xc0) == 0x80);            eptr++;
4509              ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);
4510            }            }
4511          break;          break;
4512    
# Line 4321  for (;;) Line 4518  for (;;)
4518              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4519              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4520              }              }
4521            if (*eptr >= 128 || (md->ctypes[*eptr++] & ctype_word) == 0)            if (*eptr >= 128 || (md->ctypes[*eptr] & ctype_word) == 0)
4522              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4523              eptr++;
4524            /* 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 */
4525            }            }
4526          break;          break;
# Line 4332  for (;;) Line 4530  for (;;)
4530          }  /* End switch(ctype) */          }  /* End switch(ctype) */
4531    
4532        else        else
4533  #endif     /* SUPPORT_UTF8 */  #endif     /* SUPPORT_UTF */
4534    
4535        /* 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
4536        than OP_PROP and OP_NOTPROP. */        than OP_PROP and OP_NOTPROP. */
# Line 4348  for (;;) Line 4546  for (;;)
4546              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4547              }              }
4548            if (IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH);            if (IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH);
4549              if (md->partial != 0 &&
4550                  eptr + 1 >= md->end_subject &&
4551                  NLBLOCK->nltype == NLTYPE_FIXED &&
4552                  NLBLOCK->nllen == 2 &&
4553                  *eptr == NLBLOCK->nl[0])
4554                {
4555                md->hitend = TRUE;
4556                if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
4557                }
4558            eptr++;            eptr++;
4559            }            }
4560          break;          break;
# Line 4392  for (;;) Line 4599  for (;;)
4599              case 0x000b:              case 0x000b:
4600              case 0x000c:              case 0x000c:
4601              case 0x0085:              case 0x0085:
4602    #ifdef COMPILE_PCRE16
4603                case 0x2028:
4604                case 0x2029:
4605    #endif
4606              if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);              if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);
4607              break;              break;
4608              }              }
# Line 4412  for (;;) Line 4623  for (;;)
4623              case 0x09:      /* HT */              case 0x09:      /* HT */
4624              case 0x20:      /* SPACE */              case 0x20:      /* SPACE */
4625              case 0xa0:      /* NBSP */              case 0xa0:      /* NBSP */
4626    #ifdef COMPILE_PCRE16
4627                case 0x1680:    /* OGHAM SPACE MARK */
4628                case 0x180e:    /* MONGOLIAN VOWEL SEPARATOR */
4629                case 0x2000:    /* EN QUAD */
4630                case 0x2001:    /* EM QUAD */
4631                case 0x2002:    /* EN SPACE */
4632                case 0x2003:    /* EM SPACE */
4633                case 0x2004:    /* THREE-PER-EM SPACE */
4634                case 0x2005:    /* FOUR-PER-EM SPACE */
4635                case 0x2006:    /* SIX-PER-EM SPACE */
4636                case 0x2007:    /* FIGURE SPACE */
4637                case 0x2008:    /* PUNCTUATION SPACE */
4638                case 0x2009:    /* THIN SPACE */
4639                case 0x200A:    /* HAIR SPACE */
4640                case 0x202f:    /* NARROW NO-BREAK SPACE */
4641                case 0x205f:    /* MEDIUM MATHEMATICAL SPACE */
4642                case 0x3000:    /* IDEOGRAPHIC SPACE */
4643    #endif
4644              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4645              }              }
4646            }            }
# Line 4431  for (;;) Line 4660  for (;;)
4660              case 0x09:      /* HT */              case 0x09:      /* HT */
4661              case 0x20:      /* SPACE */              case 0x20:      /* SPACE */
4662              case 0xa0:      /* NBSP */              case 0xa0:      /* NBSP */
4663    #ifdef COMPILE_PCRE16
4664                case 0x1680:    /* OGHAM SPACE MARK */
4665                case 0x180e:    /* MONGOLIAN VOWEL SEPARATOR */
4666                case 0x2000:    /* EN QUAD */
4667                case 0x2001:    /* EM QUAD */
4668                case 0x2002:    /* EN SPACE */
4669                case 0x2003:    /* EM SPACE */
4670                case 0x2004:    /* THREE-PER-EM SPACE */
4671                case 0x2005:    /* FOUR-PER-EM SPACE */
4672                case 0x2006:    /* SIX-PER-EM SPACE */
4673                case 0x2007:    /* FIGURE SPACE */
4674                case 0x2008:    /* PUNCTUATION SPACE */
4675                case 0x2009:    /* THIN SPACE */
4676                case 0x200A:    /* HAIR SPACE */
4677                case 0x202f:    /* NARROW NO-BREAK SPACE */
4678                case 0x205f:    /* MEDIUM MATHEMATICAL SPACE */
4679                case 0x3000:    /* IDEOGRAPHIC SPACE */
4680    #endif
4681              break;              break;
4682              }              }
4683            }            }
# Line 4452  for (;;) Line 4699  for (;;)
4699              case 0x0c:      /* FF */              case 0x0c:      /* FF */
4700              case 0x0d:      /* CR */              case 0x0d:      /* CR */
4701              case 0x85:      /* NEL */              case 0x85:      /* NEL */
4702    #ifdef COMPILE_PCRE16
4703                case 0x2028:    /* LINE SEPARATOR */
4704                case 0x2029:    /* PARAGRAPH SEPARATOR */
4705    #endif
4706              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4707              }              }
4708            }            }
# Line 4473  for (;;) Line 4724  for (;;)
4724              case 0x0c:      /* FF */              case 0x0c:      /* FF */
4725              case 0x0d:      /* CR */              case 0x0d:      /* CR */
4726              case 0x85:      /* NEL */              case 0x85:      /* NEL */
4727    #ifdef COMPILE_PCRE16
4728                case 0x2028:    /* LINE SEPARATOR */
4729                case 0x2029:    /* PARAGRAPH SEPARATOR */
4730    #endif
4731              break;              break;
4732              }              }
4733            }            }
# Line 4486  for (;;) Line 4741  for (;;)
4741              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4742              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4743              }              }
4744            if ((md->ctypes[*eptr++] & ctype_digit) != 0) RRETURN(MATCH_NOMATCH);            if (MAX_255(*eptr) && (md->ctypes[*eptr] & ctype_digit) != 0)
4745                RRETURN(MATCH_NOMATCH);
4746              eptr++;
4747            }            }
4748          break;          break;
4749    
# Line 4498  for (;;) Line 4755  for (;;)
4755              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4756              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4757              }              }
4758            if ((md->ctypes[*eptr++] & ctype_digit) == 0) RRETURN(MATCH_NOMATCH);            if (!MAX_255(*eptr) || (md->ctypes[*eptr] & ctype_digit) == 0)
4759                RRETURN(MATCH_NOMATCH);
4760              eptr++;
4761            }            }
4762          break;          break;
4763    
# Line 4510  for (;;) Line 4769  for (;;)
4769              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4770              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4771              }              }
4772            if ((md->ctypes[*eptr++] & ctype_space) != 0) RRETURN(MATCH_NOMATCH);            if (MAX_255(*eptr) && (md->ctypes[*eptr] & ctype_space) != 0)
4773                RRETURN(MATCH_NOMATCH);
4774              eptr++;
4775            }            }
4776          break;          break;
4777    
# Line 4522  for (;;) Line 4783  for (;;)
4783              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4784              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4785              }              }
4786            if ((md->ctypes[*eptr++] & ctype_space) == 0) RRETURN(MATCH_NOMATCH);            if (!MAX_255(*eptr) || (md->ctypes[*eptr] & ctype_space) == 0)
4787                RRETURN(MATCH_NOMATCH);
4788              eptr++;
4789            }            }
4790          break;          break;
4791    
# Line 4534  for (;;) Line 4797  for (;;)
4797              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4798              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4799              }              }
4800            if ((md->ctypes[*eptr++] & ctype_word) != 0)            if (MAX_255(*eptr) && (md->ctypes[*eptr] & ctype_word) != 0)
4801              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4802              eptr++;
4803            }            }
4804          break;          break;
4805    
# Line 4547  for (;;) Line 4811  for (;;)
4811              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4812              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4813              }              }
4814            if ((md->ctypes[*eptr++] & ctype_word) == 0)            if (!MAX_255(*eptr) || (md->ctypes[*eptr] & ctype_word) == 0)
4815              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4816              eptr++;
4817            }            }
4818          break;          break;
4819    
# Line 4766  for (;;) Line 5031  for (;;)
5031            while (eptr < md->end_subject)            while (eptr < md->end_subject)
5032              {              {
5033              int len = 1;              int len = 1;
5034              if (!utf8) c = *eptr; else { GETCHARLEN(c, eptr, len); }              if (!utf) c = *eptr; else { GETCHARLEN(c, eptr, len); }
5035              if (UCD_CATEGORY(c) != ucp_M) break;              if (UCD_CATEGORY(c) != ucp_M) break;
5036              eptr += len;              eptr += len;
5037              }              }
5038              CHECK_PARTIAL();
5039            }            }
5040          }          }
5041        else        else
5042  #endif     /* SUPPORT_UCP */  #endif     /* SUPPORT_UCP */
5043    
5044  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
5045        /* UTF-8 mode */        if (utf)
       if (utf8)  
5046          {          {
5047          for (fi = min;; fi++)          for (fi = min;; fi++)
5048            {            {
# Line 4794  for (;;) Line 5059  for (;;)
5059            GETCHARINC(c, eptr);            GETCHARINC(c, eptr);
5060            switch(ctype)            switch(ctype)
5061              {              {
5062              case OP_ANY:        /* This is the non-NL case */              case OP_ANY:               /* This is the non-NL case */
5063                if (md->partial != 0 &&    /* Take care with CRLF partial */
5064                    eptr >= md->end_subject &&
5065                    NLBLOCK->nltype == NLTYPE_FIXED &&
5066                    NLBLOCK->nllen == 2 &&
5067                    c == NLBLOCK->nl[0])
5068                  {
5069                  md->hitend = TRUE;
5070                  if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
5071                  }
5072                break;
5073    
5074              case OP_ALLANY:              case OP_ALLANY:
5075              case OP_ANYBYTE:              case OP_ANYBYTE:
5076              break;              break;
# Line 4919  for (;;) Line 5195  for (;;)
5195              break;              break;
5196    
5197              case OP_WHITESPACE:              case OP_WHITESPACE:
5198              if  (c >= 256 || (md->ctypes[c] & ctype_space) == 0)              if (c >= 256 || (md->ctypes[c] & ctype_space) == 0)
5199                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
5200              break;              break;
5201    
# Line 4940  for (;;) Line 5216  for (;;)
5216          }          }
5217        else        else
5218  #endif  #endif
5219        /* Not UTF-8 mode */        /* Not UTF mode */
5220          {          {
5221          for (fi = min;; fi++)          for (fi = min;; fi++)
5222            {            {
# Line 4957  for (;;) Line 5233  for (;;)
5233            c = *eptr++;            c = *eptr++;
5234            switch(ctype)            switch(ctype)
5235              {              {
5236              case OP_ANY:     /* This is the non-NL case */              case OP_ANY:               /* This is the non-NL case */
5237                if (md->partial != 0 &&    /* Take care with CRLF partial */
5238                    eptr >= md->end_subject &&
5239                    NLBLOCK->nltype == NLTYPE_FIXED &&
5240                    NLBLOCK->nllen == 2 &&
5241                    c == NLBLOCK->nl[0])
5242                  {
5243                  md->hitend = TRUE;
5244                  if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
5245                  }
5246                break;
5247    
5248              case OP_ALLANY:              case OP_ALLANY:
5249              case OP_ANYBYTE:              case OP_ANYBYTE:
5250              break;              break;
# Line 4976  for (;;) Line 5263  for (;;)
5263                case 0x000b:                case 0x000b:
5264                case 0x000c:                case 0x000c:
5265                case 0x0085:                case 0x0085:
5266    #ifdef COMPILE_PCRE16
5267                  case 0x2028:
5268                  case 0x2029:
5269    #endif
5270                if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);                if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);
5271                break;                break;
5272                }                }
# Line 4988  for (;;) Line 5279  for (;;)
5279                case 0x09:      /* HT */                case 0x09:      /* HT */
5280                case 0x20:      /* SPACE */                case 0x20:      /* SPACE */
5281                case 0xa0:      /* NBSP */                case 0xa0:      /* NBSP */
5282    #ifdef COMPILE_PCRE16
5283                  case 0x1680:    /* OGHAM SPACE MARK */
5284                  case 0x180e:    /* MONGOLIAN VOWEL SEPARATOR */
5285                  case 0x2000:    /* EN QUAD */
5286                  case 0x2001:    /* EM QUAD */
5287                  case 0x2002:    /* EN SPACE */
5288                  case 0x2003:    /* EM SPACE */
5289                  case 0x2004:    /* THREE-PER-EM SPACE */
5290                  case 0x2005:    /* FOUR-PER-EM SPACE */
5291                  case 0x2006:    /* SIX-PER-EM SPACE */
5292                  case 0x2007:    /* FIGURE SPACE */
5293                  case 0x2008:    /* PUNCTUATION SPACE */
5294                  case 0x2009:    /* THIN SPACE */
5295                  case 0x200A:    /* HAIR SPACE */
5296                  case 0x202f:    /* NARROW NO-BREAK SPACE */
5297                  case 0x205f:    /* MEDIUM MATHEMATICAL SPACE */
5298                  case 0x3000:    /* IDEOGRAPHIC SPACE */
5299    #endif
5300                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
5301                }                }
5302              break;              break;
# Line 4999  for (;;) Line 5308  for (;;)
5308                case 0x09:      /* HT */                case 0x09:      /* HT */
5309                case 0x20:      /* SPACE */                case 0x20:      /* SPACE */
5310                case 0xa0:      /* NBSP */                case 0xa0:      /* NBSP */
5311    #ifdef COMPILE_PCRE16
5312                  case 0x1680:    /* OGHAM SPACE MARK */
5313                  case 0x180e:    /* MONGOLIAN VOWEL SEPARATOR */
5314                  case 0x2000:    /* EN QUAD */
5315                  case 0x2001:    /* EM QUAD */
5316                  case 0x2002:    /* EN SPACE */
5317                  case 0x2003:    /* EM SPACE */
5318                  case 0x2004:    /* THREE-PER-EM SPACE */
5319                  case 0x2005:    /* FOUR-PER-EM SPACE */
5320                  case 0x2006:    /* SIX-PER-EM SPACE */
5321                  case 0x2007:    /* FIGURE SPACE */
5322                  case 0x2008:    /* PUNCTUATION SPACE */
5323                  case 0x2009:    /* THIN SPACE */
5324                  case 0x200A:    /* HAIR SPACE */
5325                  case 0x202f:    /* NARROW NO-BREAK SPACE */
5326                  case 0x205f:    /* MEDIUM MATHEMATICAL SPACE */
5327                  case 0x3000:    /* IDEOGRAPHIC SPACE */
5328    #endif
5329                break;                break;
5330                }                }
5331              break;              break;
# Line 5012  for (;;) Line 5339  for (;;)
5339                case 0x0c:      /* FF */                case 0x0c:      /* FF */
5340                case 0x0d:      /* CR */                case 0x0d:      /* CR */
5341                case 0x85:      /* NEL */                case 0x85:      /* NEL */
5342    #ifdef COMPILE_PCRE16
5343                  case 0x2028:    /* LINE SEPARATOR */
5344                  case 0x2029:    /* PARAGRAPH SEPARATOR */
5345    #endif
5346                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
5347                }                }
5348              break;              break;
# Line 5025  for (;;) Line 5356  for (;;)
5356                case 0x0c:      /* FF */                case 0x0c:      /* FF */
5357                case 0x0d:      /* CR */                case 0x0d:      /* CR */
5358                case 0x85:      /* NEL */                case 0x85:      /* NEL */
5359    #ifdef COMPILE_PCRE16
5360                  case 0x2028:    /* LINE SEPARATOR */
5361                  case 0x2029:    /* PARAGRAPH SEPARATOR */
5362    #endif
5363                break;                break;
5364                }                }
5365              break;              break;
5366    
5367              case OP_NOT_DIGIT:              case OP_NOT_DIGIT:
5368              if ((md->ctypes[c] & ctype_digit) != 0) RRETURN(MATCH_NOMATCH);              if (MAX_255(c) && (md->ctypes[c] & ctype_digit) != 0) RRETURN(MATCH_NOMATCH);
5369              break;              break;
5370    
5371              case OP_DIGIT:              case OP_DIGIT:
5372              if ((md->ctypes[c] & ctype_digit) == 0) RRETURN(MATCH_NOMATCH);              if (!MAX_255(c) || (md->ctypes[c] & ctype_digit) == 0) RRETURN(MATCH_NOMATCH);
5373              break;              break;
5374    
5375              case OP_NOT_WHITESPACE:              case OP_NOT_WHITESPACE:
5376              if ((md->ctypes[c] & ctype_space) != 0) RRETURN(MATCH_NOMATCH);              if (MAX_255(c) && (md->ctypes[c] & ctype_space) != 0) RRETURN(MATCH_NOMATCH);
5377              break;              break;
5378    
5379              case OP_WHITESPACE:              case OP_WHITESPACE:
5380              if  ((md->ctypes[c] & ctype_space) == 0) RRETURN(MATCH_NOMATCH);              if (!MAX_255(c) || (md->ctypes[c] & ctype_space) == 0) RRETURN(MATCH_NOMATCH);
5381              break;              break;
5382    
5383              case OP_NOT_WORDCHAR:              case OP_NOT_WORDCHAR:
5384              if ((md->ctypes[c] & ctype_word) != 0) RRETURN(MATCH_NOMATCH);              if (MAX_255(c) && (md->ctypes[c] & ctype_word) != 0) RRETURN(MATCH_NOMATCH);
5385              break;              break;
5386    
5387              case OP_WORDCHAR:              case OP_WORDCHAR:
5388              if ((md->ctypes[c] & ctype_word) == 0) RRETURN(MATCH_NOMATCH);              if (!MAX_255(c) || (md->ctypes[c] & ctype_word) == 0) RRETURN(MATCH_NOMATCH);
5389              break;              break;
5390    
5391              default:              default:
# Line 5239  for (;;) Line 5574  for (;;)
5574            RMATCH(eptr, ecode, offset_top, md, eptrb, RM44);            RMATCH(eptr, ecode, offset_top, md, eptrb, RM44);
5575            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
5576            if (eptr-- == pp) break;        /* Stop if tried at original pos */            if (eptr-- == pp) break;        /* Stop if tried at original pos */
5577            if (utf8) BACKCHAR(eptr);            if (utf) BACKCHAR(eptr);
5578            }            }
5579          }          }
5580    
# Line 5256  for (;;) Line 5591  for (;;)
5591              SCHECK_PARTIAL();              SCHECK_PARTIAL();
5592              break;              break;
5593              }              }
5594            if (!utf8) c = *eptr; else { GETCHARLEN(c, eptr, len); }            if (!utf) c = *eptr; else { GETCHARLEN(c, eptr, len); }
5595            if (UCD_CATEGORY(c) == ucp_M) break;            if (UCD_CATEGORY(c) == ucp_M) break;
5596            eptr += len;            eptr += len;
5597            while (eptr < md->end_subject)            while (eptr < md->end_subject)
5598              {              {
5599              len = 1;              len = 1;
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              }              }
5604              CHECK_PARTIAL();
5605            }            }
5606    
5607          /* eptr is now past the end of the maximum run */          /* eptr is now past the end of the maximum run */
# Line 5279  for (;;) Line 5615  for (;;)
5615            if (eptr-- == pp) break;        /* Stop if tried at original pos */            if (eptr-- == pp) break;        /* Stop if tried at original pos */
5616            for (;;)                        /* Move back over one extended */            for (;;)                        /* Move back over one extended */
5617              {              {
5618              if (!utf8) c = *eptr; else              if (!utf) c = *eptr; else
5619                {                {
5620                BACKCHAR(eptr);                BACKCHAR(eptr);
5621                GETCHAR(c, eptr);                GETCHAR(c, eptr);
# Line 5293  for (;;) Line 5629  for (;;)
5629        else        else
5630  #endif   /* SUPPORT_UCP */  #endif   /* SUPPORT_UCP */
5631    
5632  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
5633        /* UTF-8 mode */        if (utf)
   
       if (utf8)  
5634          {          {
5635          switch(ctype)          switch(ctype)
5636            {            {
# Line 5311  for (;;) Line 5645  for (;;)
5645                  break;                  break;
5646                  }                  }
5647                if (IS_NEWLINE(eptr)) break;                if (IS_NEWLINE(eptr)) break;
5648                  if (md->partial != 0 &&    /* Take care with CRLF partial */
5649                      eptr + 1 >= md->end_subject &&
5650                      NLBLOCK->nltype == NLTYPE_FIXED &&
5651                      NLBLOCK->nllen == 2 &&
5652                      *eptr == NLBLOCK->nl[0])
5653                    {
5654                    md->hitend = TRUE;
5655                    if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
5656                    }
5657                eptr++;                eptr++;
5658                while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;                ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);
5659                }                }
5660              }              }
5661    
# Line 5328  for (;;) Line 5671  for (;;)
5671                  break;                  break;
5672                  }                  }
5673                if (IS_NEWLINE(eptr)) break;                if (IS_NEWLINE(eptr)) break;
5674                  if (md->partial != 0 &&    /* Take care with CRLF partial */
5675                      eptr + 1 >= md->end_subject &&
5676                      NLBLOCK->nltype == NLTYPE_FIXED &&
5677                      NLBLOCK->nllen == 2 &&
5678                      *eptr == NLBLOCK->nl[0])
5679                    {
5680                    md->hitend = TRUE;
5681                    if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
5682                    }
5683                eptr++;                eptr++;
5684                while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;                ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);
5685                }                }
5686              }              }
5687            break;            break;
# Line 5345  for (;;) Line 5697  for (;;)
5697                  break;                  break;
5698                  }                  }
5699                eptr++;                eptr++;
5700                while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;                ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);
5701                }                }
5702              }              }
5703            else            else
# Line 5578  for (;;) Line 5930  for (;;)
5930            }            }
5931          }          }
5932        else        else
5933  #endif  /* SUPPORT_UTF8 */  #endif  /* SUPPORT_UTF */
5934          /* Not UTF mode */
       /* Not UTF-8 mode */  
5935          {          {
5936          switch(ctype)          switch(ctype)
5937            {            {
# Line 5593  for (;;) Line 5944  for (;;)
5944                break;                break;
5945                }                }
5946              if (IS_NEWLINE(eptr)) break;              if (IS_NEWLINE(eptr)) break;
5947                if (md->partial != 0 &&    /* Take care with CRLF partial */
5948                    eptr + 1 >= md->end_subject &&
5949                    NLBLOCK->nltype == NLTYPE_FIXED &&
5950                    NLBLOCK->nllen == 2 &&
5951                    *eptr == NLBLOCK->nl[0])
5952                  {
5953                  md->hitend = TRUE;
5954                  if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
5955                  }
5956              eptr++;              eptr++;
5957              }              }
5958            break;            break;
# Line 5624  for (;;) Line 5984  for (;;)
5984                }                }
5985              else              else
5986                {                {
5987                if (c != 0x000a &&                if (c != 0x000a && (md->bsr_anycrlf ||
5988                    (md->bsr_anycrlf ||                  (c != 0x000b && c != 0x000c && c != 0x0085
5989                      (c != 0x000b && c != 0x000c && c != 0x0085)))  #ifdef COMPILE_PCRE16
5990                  break;                  && c != 0x2028 && c != 0x2029
5991    #endif
5992                    ))) break;
5993                eptr++;                eptr++;
5994                }                }
5995              }              }
# Line 5642  for (;;) Line 6004  for (;;)
6004                break;                break;
6005                }                }
6006              c = *eptr;              c = *eptr;
6007              if (c == 0x09 || c == 0x20 || c == 0xa0) break;              if (c == 0x09 || c == 0x20 || c == 0xa0
6008    #ifdef COMPILE_PCRE16
6009                  || c == 0x1680 || c == 0x180e || (c >= 0x2000 && c <= 0x200A)
6010                  || c == 0x202f || c == 0x205f || c == 0x3000
6011    #endif
6012                  ) break;
6013              eptr++;              eptr++;
6014              }              }
6015            break;            break;
# Line 5656  for (;;) Line 6023  for (;;)
6023                break;                break;
6024                }                }
6025              c = *eptr;              c = *eptr;
6026              if (c != 0x09 && c != 0x20 && c != 0xa0) break;              if (c != 0x09 && c != 0x20 && c != 0xa0
6027    #ifdef COMPILE_PCRE16
6028                  && c != 0x1680 && c != 0x180e && (c < 0x2000 || c > 0x200A)
6029                  && c != 0x202f && c != 0x205f && c != 0x3000
6030    #endif
6031                  ) break;
6032              eptr++;              eptr++;
6033              }              }
6034            break;            break;
# Line 5670  for (;;) Line 6042  for (;;)
6042                break;                break;
6043                }                }
6044              c = *eptr;              c = *eptr;
6045              if (c == 0x0a || c == 0x0b || c == 0x0c || c == 0x0d || c == 0x85)              if (c == 0x0a || c == 0x0b || c == 0x0c || c == 0x0d || c == 0x85
6046                break;  #ifdef COMPILE_PCRE16
6047                  || c == 0x2028 || c == 0x2029
6048    #endif
6049                  ) break;
6050              eptr++;              eptr++;
6051              }              }
6052            break;            break;
# Line 5685  for (;;) Line 6060  for (;;)
6060                break;                break;
6061                }                }
6062              c = *eptr;              c = *eptr;
6063              if (c != 0x0a && c != 0x0b && c != 0x0c && c != 0x0d && c != 0x85)              if (c != 0x0a && c != 0x0b && c != 0x0c && c != 0x0d && c != 0x85
6064                break;  #ifdef COMPILE_PCRE16
6065                  && c != 0x2028 && c != 0x2029
6066    #endif
6067                  ) break;
6068              eptr++;              eptr++;
6069              }              }
6070            break;            break;
# Line 5699  for (;;) Line 6077  for (;;)
6077                SCHECK_PARTIAL();                SCHECK_PARTIAL();
6078                break;                break;
6079                }                }
6080              if ((md->ctypes[*eptr] & ctype_digit) != 0) break;              if (MAX_255(*eptr) && (md->ctypes[*eptr] & ctype_digit) != 0) break;
6081              eptr++;              eptr++;
6082              }              }
6083            break;            break;
# Line 5712  for (;;) Line 6090  for (;;)
6090                SCHECK_PARTIAL();                SCHECK_PARTIAL();
6091                break;                break;
6092                }                }
6093              if ((md->ctypes[*eptr] & ctype_digit) == 0) break;              if (!MAX_255(*eptr) || (md->ctypes[*eptr] & ctype_digit) == 0) break;
6094              eptr++;              eptr++;
6095              }              }
6096            break;            break;
# Line 5725  for (;;) Line 6103  for (;;)
6103                SCHECK_PARTIAL();                SCHECK_PARTIAL();
6104                break;                break;
6105                }                }
6106              if ((md->ctypes[*eptr] & ctype_space) != 0) break;              if (MAX_255(*eptr) && (md->ctypes[*eptr] & ctype_space) != 0) break;
6107              eptr++;              eptr++;
6108              }              }
6109            break;            break;
# Line 5738  for (;;) Line 6116  for (;;)
6116                SCHECK_PARTIAL();                SCHECK_PARTIAL();
6117                break;                break;
6118                }                }
6119              if ((md->ctypes[*eptr] & ctype_space) == 0) break;              if (!MAX_255(*eptr) || (md->ctypes[*eptr] & ctype_space) == 0) break;
6120              eptr++;              eptr++;
6121              }              }
6122            break;            break;
# Line 5751  for (;;) Line 6129  for (;;)
6129                SCHECK_PARTIAL();                SCHECK_PARTIAL();
6130                break;                break;
6131                }                }
6132              if ((md->ctypes[*eptr] & ctype_word) != 0) break;              if (MAX_255(*eptr) && (md->ctypes[*eptr] & ctype_word) != 0) break;
6133              eptr++;              eptr++;
6134              }              }
6135            break;            break;
# Line 5764  for (;;) Line 6142  for (;;)
6142                SCHECK_PARTIAL();                SCHECK_PARTIAL();
6143                break;                break;
6144                }                }
6145              if ((md->ctypes[*eptr] & ctype_word) == 0) break;              if (!MAX_255(*eptr) || (md->ctypes[*eptr] & ctype_word) == 0) break;
6146              eptr++;              eptr++;
6147              }              }
6148            break;            break;
# Line 5827  switch (frame->Xwhere) Line 6205  switch (frame->Xwhere)
6205    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)
6206    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)
6207    LBL(65) LBL(66)    LBL(65) LBL(66)
6208  #ifdef SUPPORT_UTF8  #if defined SUPPORT_UTF || !defined COMPILE_PCRE8
6209    LBL(16) LBL(18) LBL(20) LBL(21) LBL(22) LBL(23) LBL(28) LBL(30)    LBL(21)
6210    #endif
6211    #ifdef SUPPORT_UTF
6212      LBL(16) LBL(18) LBL(20)
6213      LBL(22) LBL(23) LBL(28) LBL(30)
6214    LBL(32) LBL(34) LBL(42) LBL(46)    LBL(32) LBL(34) LBL(42) LBL(46)
6215  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
6216    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)
6217    LBL(59) LBL(60) LBL(61) LBL(62)    LBL(59) LBL(60) LBL(61) LBL(62)
6218  #endif  /* SUPPORT_UCP */  #endif  /* SUPPORT_UCP */
6219  #endif  /* SUPPORT_UTF8 */  #endif  /* SUPPORT_UTF */
6220    default:    default:
6221    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));
6222    
6223    printf("+++jump error in pcre match: label %d non-existent\n", frame->Xwhere);
6224    
6225    return PCRE_ERROR_INTERNAL;    return PCRE_ERROR_INTERNAL;
6226    }    }
6227  #undef LBL  #undef LBL
# Line 5898  Undefine all the macros that were define Line 6283  Undefine all the macros that were define
6283  ***************************************************************************/  ***************************************************************************/
6284    
6285    
6286    #ifdef NO_RECURSE
6287    /*************************************************
6288    *          Release allocated heap frames         *
6289    *************************************************/
6290    
6291    /* This function releases all the allocated frames. The base frame is on the
6292    machine stack, and so must not be freed.
6293    
6294    Argument: the address of the base frame
6295    Returns:  nothing
6296    */
6297    
6298    static void
6299    release_match_heapframes (heapframe *frame_base)
6300    {
6301    heapframe *nextframe = frame_base->Xnextframe;
6302    while (nextframe != NULL)
6303      {
6304      heapframe *oldframe = nextframe;
6305      nextframe = nextframe->Xnextframe;
6306      (PUBL(stack_free))(oldframe);
6307      }
6308    }
6309    #endif
6310    
6311    
6312  /*************************************************  /*************************************************
6313  *         Execute a Regular Expression           *  *         Execute a Regular Expression           *
# Line 5923  Returns:          > 0 => success; value Line 6333  Returns:          > 0 => success; value
6333                   < -1 => some kind of unexpected problem                   < -1 => some kind of unexpected problem
6334  */  */
6335    
6336    #ifdef COMPILE_PCRE8
6337  PCRE_EXP_DEFN int PCRE_CALL_CONVENTION  PCRE_EXP_DEFN int PCRE_CALL_CONVENTION
6338  pcre_exec(const pcre *argument_re, const pcre_extra *extra_data,  pcre_exec(const pcre *argument_re, const pcre_extra *extra_data,
6339    PCRE_SPTR subject, int length, int start_offset, int options, int *offsets,    PCRE_SPTR subject, int length, int start_offset, int options, int *offsets,
6340    int offsetcount)    int offsetcount)
6341    #else
6342    PCRE_EXP_DEFN int PCRE_CALL_CONVENTION
6343    pcre16_exec(const pcre16 *argument_re, const pcre16_extra *extra_data,
6344      PCRE_SPTR16 subject, int length, int start_offset, int options, int *offsets,
6345      int offsetcount)
6346    #endif
6347  {  {
6348  int rc, ocount, arg_offset_max;  int rc, ocount, arg_offset_max;
 int first_byte = -1;  
 int req_byte = -1;  
 int req_byte2 = -1;  
6349  int newline;  int newline;
6350  BOOL using_temporary_offsets = FALSE;  BOOL using_temporary_offsets = FALSE;
6351  BOOL anchored;  BOOL anchored;
6352  BOOL startline;  BOOL startline;
6353  BOOL firstline;  BOOL firstline;
6354  BOOL first_byte_caseless = FALSE;  BOOL utf;
6355  BOOL req_byte_caseless = FALSE;  BOOL has_first_char = FALSE;
6356  BOOL utf8;  BOOL has_req_char = FALSE;
6357    pcre_uchar first_char = 0;
6358    pcre_uchar first_char2 = 0;
6359    pcre_uchar req_char = 0;
6360    pcre_uchar req_char2 = 0;
6361  match_data match_block;  match_data match_block;
6362  match_data *md = &match_block;  match_data *md = &match_block;
6363  const uschar *tables;  const pcre_uint8 *tables;
6364  const uschar *start_bits = NULL;  const pcre_uint8 *start_bits = NULL;
6365  USPTR start_match = (USPTR)subject + start_offset;  PCRE_PUCHAR start_match = (PCRE_PUCHAR)subject + start_offset;
6366  USPTR end_subject;  PCRE_PUCHAR end_subject;
6367  USPTR start_partial = NULL;  PCRE_PUCHAR start_partial = NULL;
6368  USPTR req_byte_ptr = start_match - 1;  PCRE_PUCHAR req_char_ptr = start_match - 1;
6369    
 pcre_study_data internal_study;  
6370  const pcre_study_data *study;  const pcre_study_data *study;
6371    const REAL_PCRE *re = (const REAL_PCRE *)argument_re;
6372    
6373    #ifdef NO_RECURSE
6374    heapframe frame_zero;
6375    frame_zero.Xprevframe = NULL;            /* Marks the top level */
6376    frame_zero.Xnextframe = NULL;            /* None are allocated yet */
6377    md->match_frames_base = &frame_zero;
6378    #endif
6379    
6380    /* Check for the special magic call that measures the size of the stack used
6381    per recursive call of match(). Without the funny casting for sizeof, a Windows
6382    compiler gave this error: "unary minus operator applied to unsigned type,
6383    result still unsigned". Hopefully the cast fixes that. */
6384    
6385  real_pcre internal_re;  if (re == NULL && extra_data == NULL && subject == NULL && length == -999 &&
6386  const real_pcre *external_re = (const real_pcre *)argument_re;      start_offset == -999)
6387  const real_pcre *re = external_re;  #ifdef NO_RECURSE
6388      return -((int)sizeof(heapframe));
6389    #else
6390      return match(NULL, NULL, NULL, 0, NULL, NULL, 0);
6391    #endif
6392    
6393  /* Plausibility checks */  /* Plausibility checks */
6394    
6395  if ((options & ~PUBLIC_EXEC_OPTIONS) != 0) return PCRE_ERROR_BADOPTION;  if ((options & ~PUBLIC_EXEC_OPTIONS) != 0) return PCRE_ERROR_BADOPTION;
6396  if (re == NULL || subject == NULL ||  if (re == NULL || subject == NULL || (offsets == NULL && offsetcount > 0))
6397     (offsets == NULL && offsetcount > 0)) return PCRE_ERROR_NULL;    return PCRE_ERROR_NULL;
6398  if (offsetcount < 0) return PCRE_ERROR_BADCOUNT;  if (offsetcount < 0) return PCRE_ERROR_BADCOUNT;
6399  if (start_offset < 0 || start_offset > length) return PCRE_ERROR_BADOFFSET;  if (start_offset < 0 || start_offset > length) return PCRE_ERROR_BADOFFSET;
6400    
6401    /* Check that the first field in the block is the magic number. If it is not,
6402    return with PCRE_ERROR_BADMAGIC. However, if the magic number is equal to
6403    REVERSED_MAGIC_NUMBER we return with PCRE_ERROR_BADENDIANNESS, which
6404    means that the pattern is likely compiled with different endianness. */
6405    
6406    if (re->magic_number != MAGIC_NUMBER)
6407      return re->magic_number == REVERSED_MAGIC_NUMBER?
6408        PCRE_ERROR_BADENDIANNESS:PCRE_ERROR_BADMAGIC;
6409    if ((re->flags & PCRE_MODE) == 0) return PCRE_ERROR_BADMODE;
6410    
6411  /* 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
6412  follows immediately afterwards. Other values in the md block are used only  follows immediately afterwards. Other values in the md block are used only
6413  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,
6414  so they are set up later. */  so they are set up later. */
6415    
6416  utf8 = md->utf8 = (re->options & PCRE_UTF8) != 0;  /* PCRE_UTF16 has the same value as PCRE_UTF8. */
6417    utf = md->utf = (re->options & PCRE_UTF8) != 0;
6418  md->partial = ((options & PCRE_PARTIAL_HARD) != 0)? 2 :  md->partial = ((options & PCRE_PARTIAL_HARD) != 0)? 2 :
6419                ((options & PCRE_PARTIAL_SOFT) != 0)? 1 : 0;                ((options & PCRE_PARTIAL_SOFT) != 0)? 1 : 0;
6420    
6421  /* 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
6422  code for an invalid string if a results vector is available. */  code for an invalid string if a results vector is available. */
6423    
6424  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
6425  if (utf8 && (options & PCRE_NO_UTF8_CHECK) == 0)  if (utf && (options & PCRE_NO_UTF8_CHECK) == 0)
6426    {    {
6427    int erroroffset;    int erroroffset;
6428    int errorcode = _pcre_valid_utf8((USPTR)subject, length, &erroroffset);    int errorcode = PRIV(valid_utf)((PCRE_PUCHAR)subject, length, &erroroffset);
6429    if (errorcode != 0)    if (errorcode != 0)
6430      {      {
6431      if (offsetcount >= 2)      if (offsetcount >= 2)
# Line 5988  if (utf8 && (options & PCRE_NO_UTF8_CHEC Line 6433  if (utf8 && (options & PCRE_NO_UTF8_CHEC
6433        offsets[0] = erroroffset;        offsets[0] = erroroffset;
6434        offsets[1] = errorcode;        offsets[1] = errorcode;
6435        }        }
6436    #ifdef COMPILE_PCRE16
6437        return (errorcode <= PCRE_UTF16_ERR1 && md->partial > 1)?
6438          PCRE_ERROR_SHORTUTF16 : PCRE_ERROR_BADUTF16;
6439    #else
6440      return (errorcode <= PCRE_UTF8_ERR5 && md->partial > 1)?      return (errorcode <= PCRE_UTF8_ERR5 && md->partial > 1)?
6441        PCRE_ERROR_SHORTUTF8 : PCRE_ERROR_BADUTF8;        PCRE_ERROR_SHORTUTF8 : PCRE_ERROR_BADUTF8;
6442    #endif
6443      }      }
6444    
6445    /* 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. */
6446    if (start_offset > 0 && start_offset < length &&    if (start_offset > 0 && start_offset < length &&
6447        (((USPTR)subject)[start_offset] & 0xc0) == 0x80)        NOT_FIRSTCHAR(((PCRE_PUCHAR)subject)[start_offset]))
6448      return PCRE_ERROR_BADUTF8_OFFSET;      return PCRE_ERROR_BADUTF8_OFFSET;
6449    }    }
6450  #endif  #endif
# Line 6002  if (utf8 && (options & PCRE_NO_UTF8_CHEC Line 6452  if (utf8 && (options & PCRE_NO_UTF8_CHEC
6452  /* If the pattern was successfully studied with JIT support, run the JIT  /* If the pattern was successfully studied with JIT support, run the JIT
6453  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
6454  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
6455  an unsupported flag is set. In particular, JIT does not support partial  an unsupported flag is set. */
 matching. */  
6456    
6457  #ifdef SUPPORT_JIT  #ifdef SUPPORT_JIT
6458  if (extra_data != NULL  if (extra_data != NULL
6459      && (extra_data->flags & PCRE_EXTRA_EXECUTABLE_JIT) != 0      && (extra_data->flags & (PCRE_EXTRA_EXECUTABLE_JIT |
6460                                 PCRE_EXTRA_TABLES)) == PCRE_EXTRA_EXECUTABLE_JIT
6461      && extra_data->executable_jit != NULL      && extra_data->executable_jit != NULL
     && (extra_data->flags & PCRE_EXTRA_TABLES) == 0  
6462      && (options & ~(PCRE_NO_UTF8_CHECK | PCRE_NOTBOL | PCRE_NOTEOL |      && (options & ~(PCRE_NO_UTF8_CHECK | PCRE_NOTBOL | PCRE_NOTEOL |
6463                      PCRE_NOTEMPTY | PCRE_NOTEMPTY_ATSTART)) == 0)                      PCRE_NOTEMPTY | PCRE_NOTEMPTY_ATSTART |
6464    return _pcre_jit_exec(re, extra_data->executable_jit, subject, length,                      PCRE_PARTIAL_SOFT | PCRE_PARTIAL_HARD)) == 0)
6465      start_offset, options, ((extra_data->flags & PCRE_EXTRA_MATCH_LIMIT) == 0)    {
6466      ? MATCH_LIMIT : extra_data->match_limit, offsets, offsetcount);    rc = PRIV(jit_exec)(re, extra_data->executable_jit,
6467        (const pcre_uchar *)subject, length, start_offset, options,
6468        ((extra_data->flags & PCRE_EXTRA_MATCH_LIMIT) == 0)
6469        ? MATCH_LIMIT : extra_data->match_limit, offsets, offsetcount,
6470        ((extra_data->flags & PCRE_EXTRA_MARK) != 0) ? extra_data->mark : NULL);
6471    
6472      /* PCRE_ERROR_NULL means that the selected normal or partial matching
6473      mode is not compiled. In this case we simply fallback to interpreter. */
6474    
6475      if (rc != PCRE_ERROR_NULL) return rc;
6476      }
6477  #endif  #endif
6478    
6479  /* 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
6480  numbers associated with a given name, for condition testing. */  numbers associated with a given name, for condition testing. */
6481    
6482  md->name_table = (uschar *)re + re->name_table_offset;  md->name_table = (pcre_uchar *)re + re->name_table_offset;
6483  md->name_count = re->name_count;  md->name_count = re->name_count;
6484  md->name_entry_size = re->name_entry_size;  md->name_entry_size = re->name_entry_size;
6485    
# Line 6034  md->callout_data = NULL; Line 6493  md->callout_data = NULL;
6493    
6494  /* The table pointer is always in native byte order. */  /* The table pointer is always in native byte order. */
6495    
6496  tables = external_re->tables;  tables = re->tables;
6497    
6498  if (extra_data != NULL)  if (extra_data != NULL)
6499    {    {
# Line 6054  if (extra_data != NULL) Line 6513  if (extra_data != NULL)
6513  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
6514  in other programs later. */  in other programs later. */
6515    
6516  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;  
   }  
6517    
6518  /* Set up other data */  /* Set up other data */
6519    
# Line 6076  firstline = (re->options & PCRE_FIRSTLIN Line 6523  firstline = (re->options & PCRE_FIRSTLIN
6523    
6524  /* 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. */
6525    
6526  md->start_code = (const uschar *)external_re + re->name_table_offset +  md->start_code = (const pcre_uchar *)re + re->name_table_offset +
6527    re->name_count * re->name_entry_size;    re->name_count * re->name_entry_size;
6528    
6529  md->start_subject = (USPTR)subject;  md->start_subject = (PCRE_PUCHAR)subject;
6530  md->start_offset = start_offset;  md->start_offset = start_offset;
6531  md->end_subject = md->start_subject + length;  md->end_subject = md->start_subject + length;
6532  end_subject = md->end_subject;  end_subject = md->end_subject;
# Line 6104  md->recursive = NULL; Line 6551  md->recursive = NULL;
6551  md->hasthen = (re->flags & PCRE_HASTHEN) != 0;  md->hasthen = (re->flags & PCRE_HASTHEN) != 0;
6552    
6553  md->lcc = tables + lcc_offset;  md->lcc = tables + lcc_offset;
6554    md->fcc = tables + fcc_offset;
6555  md->ctypes = tables + ctypes_offset;  md->ctypes = tables + ctypes_offset;
6556    
6557  /* Handle different \R options. */  /* Handle different \R options. */
# Line 6190  arg_offset_max = (2*ocount)/3; Line 6638  arg_offset_max = (2*ocount)/3;
6638  if (re->top_backref > 0 && re->top_backref >= ocount/3)  if (re->top_backref > 0 && re->top_backref >= ocount/3)
6639    {    {
6640    ocount = re->top_backref * 3 + 3;    ocount = re->top_backref * 3 + 3;
6641    md->offset_vector = (int *)(pcre_malloc)(ocount * sizeof(int));    md->offset_vector = (int *)(PUBL(malloc))(ocount * sizeof(int));
6642    if (md->offset_vector == NULL) return PCRE_ERROR_NOMEMORY;    if (md->offset_vector == NULL) return PCRE_ERROR_NOMEMORY;
6643    using_temporary_offsets = TRUE;    using_temporary_offsets = TRUE;
6644    DPRINTF(("Got memory to hold back references\n"));    DPRINTF(("Got memory to hold back references\n"));
# Line 6217  if (md->offset_vector != NULL) Line 6665  if (md->offset_vector != NULL)
6665    md->offset_vector[0] = md->offset_vector[1] = -1;    md->offset_vector[0] = md->offset_vector[1] = -1;
6666    }    }
6667    
6668  /* 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
6669  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
6670  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
6671  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 6675  if (!anchored)
6675    {    {
6676    if ((re->flags & PCRE_FIRSTSET) != 0)    if ((re->flags & PCRE_FIRSTSET) != 0)
6677      {      {
6678      first_byte = re->first_byte & 255;      has_first_char = TRUE;
6679      if ((first_byte_caseless = ((re->first_byte & REQ_CASELESS) != 0)) == TRUE)      first_char = first_char2 = (pcre_uchar)(re->first_char);
6680        first_byte = md->lcc[first_byte];      if ((re->flags & PCRE_FCH_CASELESS) != 0)
6681          {
6682          first_char2 = TABLE_GET(first_char, md->fcc, first_char);
6683    #if defined SUPPORT_UCP && !(defined COMPILE_PCRE8)
6684          if (utf && first_char > 127)
6685            first_char2 = UCD_OTHERCASE(first_char);
6686    #endif
6687          }
6688      }      }
6689    else    else
6690      if (!startline && study != NULL &&      if (!startline && study != NULL &&
# Line 6242  character" set. */ Line 6697  character" set. */
6697    
6698  if ((re->flags & PCRE_REQCHSET) != 0)  if ((re->flags & PCRE_REQCHSET) != 0)
6699    {    {
6700    req_byte = re->req_byte & 255;    has_req_char = TRUE;
6701    req_byte_caseless = (re->req_byte & REQ_CASELESS) != 0;    req_char = req_char2 = (pcre_uchar)(re->req_char);
6702    req_byte2 = (tables + fcc_offset)[req_byte];  /* case flipped */    if ((re->flags & PCRE_RCH_CASELESS) != 0)
6703        {
6704        req_char2 = TABLE_GET(req_char, md->fcc, req_char);
6705    #if defined SUPPORT_UCP && !(defined COMPILE_PCRE8)
6706        if (utf && req_char > 127)
6707          req_char2 = UCD_OTHERCASE(req_char);
6708    #endif
6709        }
6710    }    }
6711    
6712    
   
   
6713  /* ==========================================================================*/  /* ==========================================================================*/
6714    
6715  /* 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 6717  the loop runs just once. */
6717    
6718  for(;;)  for(;;)
6719    {    {
6720    USPTR save_end_subject = end_subject;    PCRE_PUCHAR save_end_subject = end_subject;
6721    USPTR new_start_match;    PCRE_PUCHAR new_start_match;
6722    
6723    /* 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
6724    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 6728  for(;;)
6728    
6729    if (firstline)    if (firstline)
6730      {      {
6731      USPTR t = start_match;      PCRE_PUCHAR t = start_match;
6732  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
6733      if (utf8)      if (utf)
6734        {        {
6735        while (t < md->end_subject && !IS_NEWLINE(t))        while (t < md->end_subject && !IS_NEWLINE(t))
6736          {          {
6737          t++;          t++;
6738          while (t < end_subject && (*t & 0xc0) == 0x80) t++;          ACROSSCHAR(t < end_subject, *t, t++);
6739          }          }
6740        }        }
6741      else      else
# Line 6292  for(;;) Line 6752  for(;;)
6752    
6753    if (((options | re->options) & PCRE_NO_START_OPTIMIZE) == 0)    if (((options | re->options) & PCRE_NO_START_OPTIMIZE) == 0)
6754      {      {
6755      /* Advance to a unique first byte if there is one. */      /* Advance to a unique first char if there is one. */
6756    
6757      if (first_byte >= 0)      if (has_first_char)
6758        {        {
6759        if (first_byte_caseless)        if (first_char != first_char2)
6760          while (start_match < end_subject && md->lcc[*start_match] != first_byte)          while (start_match < end_subject &&
6761                *start_match != first_char && *start_match != first_char2)
6762            start_match++;            start_match++;
6763        else        else
6764          while (start_match < end_subject && *start_match != first_byte)          while (start_match < end_subject && *start_match != first_char)
6765            start_match++;            start_match++;
6766        }        }
6767    
# Line 6310  for(;;) Line 6771  for(;;)
6771        {        {
6772        if (start_match > md->start_subject + start_offset)        if (start_match > md->start_subject + start_offset)
6773          {          {
6774  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
6775          if (utf8)          if (utf)
6776            {            {
6777            while (start_match < end_subject && !WAS_NEWLINE(start_match))            while (start_match < end_subject && !WAS_NEWLINE(start_match))
6778              {              {
6779              start_match++;              start_match++;
6780              while(start_match < end_subject && (*start_match & 0xc0) == 0x80)              ACROSSCHAR(start_match < end_subject, *start_match,
6781                start_match++;                start_match++);
6782              }              }
6783            }            }
6784          else          else
# Line 6344  for(;;) Line 6805  for(;;)
6805        while (start_match < end_subject)        while (start_match < end_subject)
6806          {          {
6807          register unsigned int c = *start_match;          register unsigned int c = *start_match;
6808    #ifndef COMPILE_PCRE8
6809            if (c > 255) c = 255;
6810    #endif
6811          if ((start_bits[c/8] & (1 << (c&7))) == 0)          if ((start_bits[c/8] & (1 << (c&7))) == 0)
6812            {            {
6813            start_match++;            start_match++;
6814  #ifdef SUPPORT_UTF8  #if defined SUPPORT_UTF && defined COMPILE_PCRE8
6815            if (utf8)            /* In non 8-bit mode, the iteration will stop for
6816              while(start_match < end_subject && (*start_match & 0xc0) == 0x80)            characters > 255 at the beginning or not stop at all. */
6817                start_match++;            if (utf)
6818                ACROSSCHAR(start_match < end_subject, *start_match,
6819                  start_match++);
6820  #endif  #endif
6821            }            }
6822          else break;          else break;
# Line 6379  for(;;) Line 6845  for(;;)
6845        break;        break;
6846        }        }
6847    
6848      /* 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
6849      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
6850      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.
6851      This optimization can save a huge amount of backtracking in patterns with      This optimization can save a huge amount of backtracking in patterns with
6852      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 6859  for(;;)
6859      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
6860      long. */      long. */
6861    
6862      if (req_byte >= 0 && end_subject - start_match < REQ_BYTE_MAX)      if (has_req_char && end_subject - start_match < REQ_BYTE_MAX)
6863        {        {
6864        register USPTR p = start_match + ((first_byte >= 0)? 1 : 0);        register PCRE_PUCHAR p = start_match + (has_first_char? 1:0);
6865    
6866        /* 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
6867        place we found it at last time. */        place we found it at last time. */
6868    
6869        if (p > req_byte_ptr)        if (p > req_char_ptr)
6870          {          {
6871          if (req_byte_caseless)          if (req_char != req_char2)
6872            {            {
6873            while (p < end_subject)            while (p < end_subject)
6874              {              {
6875              register int pp = *p++;              register int pp = *p++;
6876              if (pp == req_byte || pp == req_byte2) { p--; break; }              if (pp == req_char || pp == req_char2) { p--; break; }
6877              }              }
6878            }            }
6879          else          else
6880            {            {
6881            while (p < end_subject)            while (p < end_subject)
6882              {              {
6883              if (*p++ == req_byte) { p--; break; }              if (*p++ == req_char) { p--; break; }
6884              }              }
6885            }            }
6886    
# Line 6431  for(;;) Line 6897  for(;;)
6897          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
6898          the start hasn't passed this character yet. */          the start hasn't passed this character yet. */
6899    
6900          req_byte_ptr = p;          req_char_ptr = p;
6901          }          }
6902        }        }
6903      }      }
# Line 6456  for(;;) Line 6922  for(;;)
6922    switch(rc)    switch(rc)
6923      {      {
6924      /* 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
6925      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
6926      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
6927      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
6928      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
6929      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. */
6930    
6931      case MATCH_SKIP_ARG:      case MATCH_SKIP_ARG:
6932      new_start_match = start_match;      new_start_match = start_match;
6933      md->ignore_skip_arg = TRUE;      md->ignore_skip_arg = TRUE;
6934      break;      break;
6935    
6936      /* 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
6937      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 6952  for(;;)
6952      case MATCH_THEN:      case MATCH_THEN:
6953      md->ignore_skip_arg = FALSE;      md->ignore_skip_arg = FALSE;
6954      new_start_match = start_match + 1;      new_start_match = start_match + 1;
6955  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
6956      if (utf8)      if (utf)
6957        while(new_start_match < end_subject && (*new_start_match & 0xc0) == 0x80)        ACROSSCHAR(new_start_match < end_subject, *new_start_match,
6958          new_start_match++;          new_start_match++);
6959  #endif  #endif
6960      break;      break;
6961    
# Line 6527  for(;;) Line 6993  for(;;)
6993    
6994    /* 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
6995    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
6996    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
6997      normal matching start_match will aways be greater than the first position at
6998      this stage, but a failed *SKIP can cause a return at the same point, which is
6999      why the first test exists. */
7000    
7001    if (start_match[-1] == CHAR_CR &&    if (start_match > (PCRE_PUCHAR)subject + start_offset &&
7002          start_match[-1] == CHAR_CR &&
7003        start_match < end_subject &&        start_match < end_subject &&
7004        *start_match == CHAR_NL &&        *start_match == CHAR_NL &&
7005        (re->flags & PCRE_HASCRORLF) == 0 &&        (re->flags & PCRE_HASCRORLF) == 0 &&
# Line 6575  if (rc == MATCH_MATCH || rc == MATCH_ACC Line 7045  if (rc == MATCH_MATCH || rc == MATCH_ACC
7045        }        }
7046      if (md->end_offset_top > arg_offset_max) md->offset_overflow = TRUE;      if (md->end_offset_top > arg_offset_max) md->offset_overflow = TRUE;
7047      DPRINTF(("Freeing temporary memory\n"));      DPRINTF(("Freeing temporary memory\n"));
7048      (pcre_free)(md->offset_vector);      (PUBL(free))(md->offset_vector);
7049      }      }
7050    
7051    /* 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 6614  if (rc == MATCH_MATCH || rc == MATCH_ACC Line 7084  if (rc == MATCH_MATCH || rc == MATCH_ACC
7084      }      }
7085    
7086    /* Return MARK data if requested */    /* Return MARK data if requested */
7087    
7088    if (extra_data != NULL && (extra_data->flags & PCRE_EXTRA_MARK) != 0)    if (extra_data != NULL && (extra_data->flags & PCRE_EXTRA_MARK) != 0)
7089      *(extra_data->mark) = (unsigned char *)(md->mark);      *(extra_data->mark) = (pcre_uchar *)md->mark;
7090    DPRINTF((">>>> returning %d\n", rc));    DPRINTF((">>>> returning %d\n", rc));
7091    #ifdef NO_RECURSE
7092      release_match_heapframes(&frame_zero);
7093    #endif
7094    return rc;    return rc;
7095    }    }
7096    
# Line 6627  attempt has failed at all permitted star Line 7100  attempt has failed at all permitted star
7100  if (using_temporary_offsets)  if (using_temporary_offsets)
7101    {    {
7102    DPRINTF(("Freeing temporary memory\n"));    DPRINTF(("Freeing temporary memory\n"));
7103    (pcre_free)(md->offset_vector);    (PUBL(free))(md->offset_vector);
7104    }    }
7105    
7106  /* 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 7108  if (using_temporary_offsets)
7108  if (rc != MATCH_NOMATCH && rc != PCRE_ERROR_PARTIAL)  if (rc != MATCH_NOMATCH && rc != PCRE_ERROR_PARTIAL)
7109    {    {
7110    DPRINTF((">>>> error: returning %d\n", rc));    DPRINTF((">>>> error: returning %d\n", rc));
7111    #ifdef NO_RECURSE
7112      release_match_heapframes(&frame_zero);
7113    #endif
7114    return rc;    return rc;
7115    }    }
7116    
# Line 6646  if (start_partial != NULL) Line 7122  if (start_partial != NULL)
7122    md->mark = NULL;    md->mark = NULL;
7123    if (offsetcount > 1)    if (offsetcount > 1)
7124      {      {
7125      offsets[0] = (int)(start_partial - (USPTR)subject);      offsets[0] = (int)(start_partial - (PCRE_PUCHAR)subject);
7126      offsets[1] = (int)(end_subject - (USPTR)subject);      offsets[1] = (int)(end_subject - (PCRE_PUCHAR)subject);
7127      }      }
7128    rc = PCRE_ERROR_PARTIAL;    rc = PCRE_ERROR_PARTIAL;
7129    }    }
# Line 6663  else Line 7139  else
7139  /* Return the MARK data if it has been requested. */  /* Return the MARK data if it has been requested. */
7140    
7141  if (extra_data != NULL && (extra_data->flags & PCRE_EXTRA_MARK) != 0)  if (extra_data != NULL && (extra_data->flags & PCRE_EXTRA_MARK) != 0)
7142    *(extra_data->mark) = (unsigned char *)(md->nomatch_mark);    *(extra_data->mark) = (pcre_uchar *)md->nomatch_mark;
7143    #ifdef NO_RECURSE
7144      release_match_heapframes(&frame_zero);
7145    #endif
7146  return rc;  return rc;
7147  }  }
7148    

Legend:
Removed from v.771  
changed lines
  Added in v.933

  ViewVC Help
Powered by ViewVC 1.1.5