/[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 788 by ph10, Tue Dec 6 15:38:01 2011 UTC revision 1011 by ph10, Sat Aug 25 11:36:15 2012 UTC
# Line 6  Line 6 
6  and semantics are as close as possible to those of the Perl 5 language.  and semantics are as close as possible to those of the Perl 5 language.
7    
8                         Written by Philip Hazel                         Written by Philip Hazel
9             Copyright (c) 1997-2011 University of Cambridge             Copyright (c) 1997-2012 University of Cambridge
10    
11  -----------------------------------------------------------------------------  -----------------------------------------------------------------------------
12  Redistribution and use in source and binary forms, with or without  Redistribution and use in source and binary forms, with or without
# Line 37  POSSIBILITY OF SUCH DAMAGE. Line 37  POSSIBILITY OF SUCH DAMAGE.
37  -----------------------------------------------------------------------------  -----------------------------------------------------------------------------
38  */  */
39    
   
40  /* This module contains pcre_exec(), the externally visible function that does  /* This module contains pcre_exec(), the externally visible function that does
41  pattern matching using an NFA algorithm, trying to mimic Perl as closely as  pattern matching using an NFA algorithm, trying to mimic Perl as closely as
42  possible. There are also some static supporting functions. */  possible. There are also some static supporting functions. */
# Line 113  Returns:     nothing Line 112  Returns:     nothing
112  */  */
113    
114  static void  static void
115  pchars(const uschar *p, int length, BOOL is_subject, match_data *md)  pchars(const pcre_uchar *p, int length, BOOL is_subject, match_data *md)
116  {  {
117  unsigned int c;  unsigned int c;
118  if (is_subject && length > md->end_subject - p) length = md->end_subject - p;  if (is_subject && length > md->end_subject - p) length = md->end_subject - p;
# Line 140  Arguments: Line 139  Arguments:
139    md          points to match data block    md          points to match data block
140    caseless    TRUE if caseless    caseless    TRUE if caseless
141    
142  Returns:      < 0 if not matched, otherwise the number of subject bytes matched  Returns:      >= 0 the number of subject bytes matched
143                  -1 no match
144                  -2 partial match; always given if at end subject
145  */  */
146    
147  static int  static int
148  match_ref(int offset, register USPTR eptr, int length, match_data *md,  match_ref(int offset, register PCRE_PUCHAR eptr, int length, match_data *md,
149    BOOL caseless)    BOOL caseless)
150  {  {
151  USPTR eptr_start = eptr;  PCRE_PUCHAR eptr_start = eptr;
152  register USPTR p = md->start_subject + md->offset_vector[offset];  register PCRE_PUCHAR p = md->start_subject + md->offset_vector[offset];
153    
154  #ifdef PCRE_DEBUG  #ifdef PCRE_DEBUG
155  if (eptr >= md->end_subject)  if (eptr >= md->end_subject)
# Line 163  pchars(p, length, FALSE, md); Line 164  pchars(p, length, FALSE, md);
164  printf("\n");  printf("\n");
165  #endif  #endif
166    
167  /* Always fail if reference not set (and not JavaScript compatible). */  /* Always fail if reference not set (and not JavaScript compatible - in that
168    case the length is passed as zero). */
169    
170  if (length < 0) return -1;  if (length < 0) return -1;
171    
# Line 173  ASCII characters. */ Line 175  ASCII characters. */
175    
176  if (caseless)  if (caseless)
177    {    {
178  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
179  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
180    if (md->utf8)    if (md->utf)
181      {      {
182      /* Match characters up to the end of the reference. NOTE: the number of      /* Match characters up to the end of the reference. NOTE: the number of
183      bytes matched may differ, because there are some characters whose upper and      bytes matched may differ, because there are some characters whose upper and
# Line 185  if (caseless) Line 187  if (caseless)
187      the latter. It is important, therefore, to check the length along the      the latter. It is important, therefore, to check the length along the
188      reference, not along the subject (earlier code did this wrong). */      reference, not along the subject (earlier code did this wrong). */
189    
190      USPTR endptr = p + length;      PCRE_PUCHAR endptr = p + length;
191      while (p < endptr)      while (p < endptr)
192        {        {
193        int c, d;        int c, d;
194        if (eptr >= md->end_subject) return -1;        if (eptr >= md->end_subject) return -2;   /* Partial match */
195        GETCHARINC(c, eptr);        GETCHARINC(c, eptr);
196        GETCHARINC(d, p);        GETCHARINC(d, p);
197        if (c != d && c != UCD_OTHERCASE(d)) return -1;        if (c != d && c != UCD_OTHERCASE(d)) return -1;
# Line 202  if (caseless) Line 204  if (caseless)
204    /* The same code works when not in UTF-8 mode and in UTF-8 mode when there    /* The same code works when not in UTF-8 mode and in UTF-8 mode when there
205    is no UCP support. */    is no UCP support. */
206      {      {
     if (eptr + length > md->end_subject) return -1;  
207      while (length-- > 0)      while (length-- > 0)
208        { if (md->lcc[*p++] != md->lcc[*eptr++]) return -1; }        {
209          if (eptr >= md->end_subject) return -2;   /* Partial match */
210          if (TABLE_GET(*p, md->lcc, *p) != TABLE_GET(*eptr, md->lcc, *eptr)) return -1;
211          p++;
212          eptr++;
213          }
214      }      }
215    }    }
216    
# Line 213  are in UTF-8 mode. */ Line 219  are in UTF-8 mode. */
219    
220  else  else
221    {    {
222    if (eptr + length > md->end_subject) return -1;    while (length-- > 0)
223    while (length-- > 0) if (*p++ != *eptr++) return -1;      {
224        if (eptr >= md->end_subject) return -2;   /* Partial match */
225        if (*p++ != *eptr++) return -1;
226        }
227    }    }
228    
229  return (int)(eptr - eptr_start);  return (int)(eptr - eptr_start);
# Line 307  argument of match(), which never changes Line 316  argument of match(), which never changes
316    
317  #define RMATCH(ra,rb,rc,rd,re,rw)\  #define RMATCH(ra,rb,rc,rd,re,rw)\
318    {\    {\
319    heapframe *newframe = (heapframe *)(pcre_stack_malloc)(sizeof(heapframe));\    heapframe *newframe = frame->Xnextframe;\
320    if (newframe == NULL) RRETURN(PCRE_ERROR_NOMEMORY);\    if (newframe == NULL)\
321    frame->Xwhere = rw; \      {\
322        newframe = (heapframe *)(PUBL(stack_malloc))(sizeof(heapframe));\
323        if (newframe == NULL) RRETURN(PCRE_ERROR_NOMEMORY);\
324        newframe->Xnextframe = NULL;\
325        frame->Xnextframe = newframe;\
326        }\
327      frame->Xwhere = rw;\
328    newframe->Xeptr = ra;\    newframe->Xeptr = ra;\
329    newframe->Xecode = rb;\    newframe->Xecode = rb;\
330    newframe->Xmstart = mstart;\    newframe->Xmstart = mstart;\
# Line 328  argument of match(), which never changes Line 343  argument of match(), which never changes
343    {\    {\
344    heapframe *oldframe = frame;\    heapframe *oldframe = frame;\
345    frame = oldframe->Xprevframe;\    frame = oldframe->Xprevframe;\
   (pcre_stack_free)(oldframe);\  
346    if (frame != NULL)\    if (frame != NULL)\
347      {\      {\
348      rrc = ra;\      rrc = ra;\
# Line 342  argument of match(), which never changes Line 356  argument of match(), which never changes
356    
357  typedef struct heapframe {  typedef struct heapframe {
358    struct heapframe *Xprevframe;    struct heapframe *Xprevframe;
359      struct heapframe *Xnextframe;
360    
361    /* Function arguments that may change */    /* Function arguments that may change */
362    
363    USPTR Xeptr;    PCRE_PUCHAR Xeptr;
364    const uschar *Xecode;    const pcre_uchar *Xecode;
365    USPTR Xmstart;    PCRE_PUCHAR Xmstart;
366    int Xoffset_top;    int Xoffset_top;
367    eptrblock *Xeptrb;    eptrblock *Xeptrb;
368    unsigned int Xrdepth;    unsigned int Xrdepth;
369    
370    /* Function local variables */    /* Function local variables */
371    
372    USPTR Xcallpat;    PCRE_PUCHAR Xcallpat;
373  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
374    USPTR Xcharptr;    PCRE_PUCHAR Xcharptr;
375  #endif  #endif
376    USPTR Xdata;    PCRE_PUCHAR Xdata;
377    USPTR Xnext;    PCRE_PUCHAR Xnext;
378    USPTR Xpp;    PCRE_PUCHAR Xpp;
379    USPTR Xprev;    PCRE_PUCHAR Xprev;
380    USPTR Xsaved_eptr;    PCRE_PUCHAR Xsaved_eptr;
381    
382    recursion_info Xnew_recursive;    recursion_info Xnew_recursive;
383    
# Line 375  typedef struct heapframe { Line 390  typedef struct heapframe {
390    int Xprop_value;    int Xprop_value;
391    int Xprop_fail_result;    int Xprop_fail_result;
392    int Xoclength;    int Xoclength;
393    uschar Xocchars[8];    pcre_uchar Xocchars[6];
394  #endif  #endif
395    
396    int Xcodelink;    int Xcodelink;
# Line 440  the subject. */ Line 455  the subject. */
455    
456    
457  /* Performance note: It might be tempting to extract commonly used fields from  /* Performance note: It might be tempting to extract commonly used fields from
458  the md structure (e.g. utf8, end_subject) into individual variables to improve  the md structure (e.g. utf, end_subject) into individual variables to improve
459  performance. Tests using gcc on a SPARC disproved this; in the first case, it  performance. Tests using gcc on a SPARC disproved this; in the first case, it
460  made performance worse.  made performance worse.
461    
# Line 463  Returns:       MATCH_MATCH if matched Line 478  Returns:       MATCH_MATCH if matched
478  */  */
479    
480  static int  static int
481  match(REGISTER USPTR eptr, REGISTER const uschar *ecode, USPTR mstart,  match(REGISTER PCRE_PUCHAR eptr, REGISTER const pcre_uchar *ecode,
482    int offset_top, match_data *md, eptrblock *eptrb, unsigned int rdepth)    PCRE_PUCHAR mstart, int offset_top, match_data *md, eptrblock *eptrb,
483      unsigned int rdepth)
484  {  {
485  /* These variables do not need to be preserved over recursion in this function,  /* These variables do not need to be preserved over recursion in this function,
486  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 489  so they can be ordinary variables in all
489  register int  rrc;         /* Returns from recursive calls */  register int  rrc;         /* Returns from recursive calls */
490  register int  i;           /* Used for loops not involving calls to RMATCH() */  register int  i;           /* Used for loops not involving calls to RMATCH() */
491  register unsigned int c;   /* Character values not kept over RMATCH() calls */  register unsigned int c;   /* Character values not kept over RMATCH() calls */
492  register BOOL utf8;        /* Local copy of UTF-8 flag for speed */  register BOOL utf;         /* Local copy of UTF flag for speed */
493    
494  BOOL minimize, possessive; /* Quantifier options */  BOOL minimize, possessive; /* Quantifier options */
495  BOOL caseless;  BOOL caseless;
496  int condcode;  int condcode;
497    
498  /* When recursion is not being used, all "local" variables that have to be  /* When recursion is not being used, all "local" variables that have to be
499  preserved over calls to RMATCH() are part of a "frame" which is obtained from  preserved over calls to RMATCH() are part of a "frame". We set up the top-level
500  heap storage. Set up the top-level frame here; others are obtained from the  frame on the stack here; subsequent instantiations are obtained from the heap
501  heap whenever RMATCH() does a "recursion". See the macro definitions above. */  whenever RMATCH() does a "recursion". See the macro definitions above. Putting
502    the top-level on the stack rather than malloc-ing them all gives a performance
503    boost in many cases where there is not much "recursion". */
504    
505  #ifdef NO_RECURSE  #ifdef NO_RECURSE
506  heapframe *frame = (heapframe *)(pcre_stack_malloc)(sizeof(heapframe));  heapframe *frame = (heapframe *)md->match_frames_base;
 if (frame == NULL) RRETURN(PCRE_ERROR_NOMEMORY);  
 frame->Xprevframe = NULL;            /* Marks the top level */  
507    
508  /* Copy in the original argument variables */  /* Copy in the original argument variables */
509    
# Line 513  HEAP_RECURSE: Line 529  HEAP_RECURSE:
529    
530  /* Ditto for the local variables */  /* Ditto for the local variables */
531    
532  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
533  #define charptr            frame->Xcharptr  #define charptr            frame->Xcharptr
534  #endif  #endif
535  #define callpat            frame->Xcallpat  #define callpat            frame->Xcallpat
# Line 571  declarations can be cut out in a block. Line 587  declarations can be cut out in a block.
587  below are for variables that do not have to be preserved over a recursive call  below are for variables that do not have to be preserved over a recursive call
588  to RMATCH(). */  to RMATCH(). */
589    
590  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
591  const uschar *charptr;  const pcre_uchar *charptr;
592  #endif  #endif
593  const uschar *callpat;  const pcre_uchar *callpat;
594  const uschar *data;  const pcre_uchar *data;
595  const uschar *next;  const pcre_uchar *next;
596  USPTR         pp;  PCRE_PUCHAR       pp;
597  const uschar *prev;  const pcre_uchar *prev;
598  USPTR         saved_eptr;  PCRE_PUCHAR       saved_eptr;
599    
600  recursion_info new_recursive;  recursion_info new_recursive;
601    
# Line 592  int prop_type; Line 608  int prop_type;
608  int prop_value;  int prop_value;
609  int prop_fail_result;  int prop_fail_result;
610  int oclength;  int oclength;
611  uschar occhars[8];  pcre_uchar occhars[6];
612  #endif  #endif
613    
614  int codelink;  int codelink;
# Line 608  int save_offset1, save_offset2, save_off Line 624  int save_offset1, save_offset2, save_off
624  int stacksave[REC_STACK_SAVE_MAX];  int stacksave[REC_STACK_SAVE_MAX];
625    
626  eptrblock newptrb;  eptrblock newptrb;
627    
628    /* There is a special fudge for calling match() in a way that causes it to
629    measure the size of its basic stack frame when the stack is being used for
630    recursion. The second argument (ecode) being NULL triggers this behaviour. It
631    cannot normally ever be NULL. The return is the negated value of the frame
632    size. */
633    
634    if (ecode == NULL)
635      {
636      if (rdepth == 0)
637        return match((PCRE_PUCHAR)&rdepth, NULL, NULL, 0, NULL, NULL, 1);
638      else
639        {
640        int len = (char *)&rdepth - (char *)eptr;
641        return (len > 0)? -len : len;
642        }
643      }
644  #endif     /* NO_RECURSE */  #endif     /* NO_RECURSE */
645    
646  /* To save space on the stack and in the heap frame, I have doubled up on some  /* To save space on the stack and in the heap frame, I have doubled up on some
# Line 620  the alternative names that are used. */ Line 653  the alternative names that are used. */
653  #define code_offset   codelink  #define code_offset   codelink
654  #define condassert    condition  #define condassert    condition
655  #define matched_once  prev_is_word  #define matched_once  prev_is_word
656    #define foc           number
657    #define save_mark     data
658    
659  /* These statements are here to stop the compiler complaining about unitialized  /* These statements are here to stop the compiler complaining about unitialized
660  variables. */  variables. */
# Line 645  defined). However, RMATCH isn't like a f Line 680  defined). However, RMATCH isn't like a f
680  complicated macro. It has to be used in one particular way. This shouldn't,  complicated macro. It has to be used in one particular way. This shouldn't,
681  however, impact performance when true recursion is being used. */  however, impact performance when true recursion is being used. */
682    
683  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
684  utf8 = md->utf8;       /* Local copy of the flag */  utf = md->utf;       /* Local copy of the flag */
685  #else  #else
686  utf8 = FALSE;  utf = FALSE;
687  #endif  #endif
688    
689  /* First check that we haven't called match() too many times, or that we  /* First check that we haven't called match() too many times, or that we
# Line 689  for (;;) Line 724  for (;;)
724      case OP_MARK:      case OP_MARK:
725      md->nomatch_mark = ecode + 2;      md->nomatch_mark = ecode + 2;
726      md->mark = NULL;    /* In case previously set by assertion */      md->mark = NULL;    /* In case previously set by assertion */
727      RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode] + ecode[1], offset_top, md,      RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode] + ecode[1], offset_top, md,
728        eptrb, RM55);        eptrb, RM55);
729      if ((rrc == MATCH_MATCH || rrc == MATCH_ACCEPT) &&      if ((rrc == MATCH_MATCH || rrc == MATCH_ACCEPT) &&
730           md->mark == NULL) md->mark = ecode + 2;           md->mark == NULL) md->mark = ecode + 2;
# Line 702  for (;;) Line 737  for (;;)
737      unaltered. */      unaltered. */
738    
739      else if (rrc == MATCH_SKIP_ARG &&      else if (rrc == MATCH_SKIP_ARG &&
740          strcmp((char *)(ecode + 2), (char *)(md->start_match_ptr)) == 0)          STRCMP_UC_UC(ecode + 2, md->start_match_ptr) == 0)
741        {        {
742        md->start_match_ptr = eptr;        md->start_match_ptr = eptr;
743        RRETURN(MATCH_SKIP);        RRETURN(MATCH_SKIP);
# Line 715  for (;;) Line 750  for (;;)
750      /* COMMIT overrides PRUNE, SKIP, and THEN */      /* COMMIT overrides PRUNE, SKIP, and THEN */
751    
752      case OP_COMMIT:      case OP_COMMIT:
753      RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,      RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md,
754        eptrb, RM52);        eptrb, RM52);
755      if (rrc != MATCH_NOMATCH && rrc != MATCH_PRUNE &&      if (rrc != MATCH_NOMATCH && rrc != MATCH_PRUNE &&
756          rrc != MATCH_SKIP && rrc != MATCH_SKIP_ARG &&          rrc != MATCH_SKIP && rrc != MATCH_SKIP_ARG &&
# Line 726  for (;;) Line 761  for (;;)
761      /* PRUNE overrides THEN */      /* PRUNE overrides THEN */
762    
763      case OP_PRUNE:      case OP_PRUNE:
764      RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,      RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md,
765        eptrb, RM51);        eptrb, RM51);
766      if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN) RRETURN(rrc);      if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN) RRETURN(rrc);
767      RRETURN(MATCH_PRUNE);      RRETURN(MATCH_PRUNE);
# Line 734  for (;;) Line 769  for (;;)
769      case OP_PRUNE_ARG:      case OP_PRUNE_ARG:
770      md->nomatch_mark = ecode + 2;      md->nomatch_mark = ecode + 2;
771      md->mark = NULL;    /* In case previously set by assertion */      md->mark = NULL;    /* In case previously set by assertion */
772      RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode] + ecode[1], offset_top, md,      RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode] + ecode[1], offset_top, md,
773        eptrb, RM56);        eptrb, RM56);
774      if ((rrc == MATCH_MATCH || rrc == MATCH_ACCEPT) &&      if ((rrc == MATCH_MATCH || rrc == MATCH_ACCEPT) &&
775           md->mark == NULL) md->mark = ecode + 2;           md->mark == NULL) md->mark = ecode + 2;
# Line 744  for (;;) Line 779  for (;;)
779      /* SKIP overrides PRUNE and THEN */      /* SKIP overrides PRUNE and THEN */
780    
781      case OP_SKIP:      case OP_SKIP:
782      RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,      RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md,
783        eptrb, RM53);        eptrb, RM53);
784      if (rrc != MATCH_NOMATCH && rrc != MATCH_PRUNE && rrc != MATCH_THEN)      if (rrc != MATCH_NOMATCH && rrc != MATCH_PRUNE && rrc != MATCH_THEN)
785        RRETURN(rrc);        RRETURN(rrc);
# Line 758  for (;;) Line 793  for (;;)
793      case OP_SKIP_ARG:      case OP_SKIP_ARG:
794      if (md->ignore_skip_arg)      if (md->ignore_skip_arg)
795        {        {
796        ecode += _pcre_OP_lengths[*ecode] + ecode[1];        ecode += PRIV(OP_lengths)[*ecode] + ecode[1];
797        break;        break;
798        }        }
799      RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode] + ecode[1], offset_top, md,      RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode] + ecode[1], offset_top, md,
800        eptrb, RM57);        eptrb, RM57);
801      if (rrc != MATCH_NOMATCH && rrc != MATCH_PRUNE && rrc != MATCH_THEN)      if (rrc != MATCH_NOMATCH && rrc != MATCH_PRUNE && rrc != MATCH_THEN)
802        RRETURN(rrc);        RRETURN(rrc);
# Line 779  for (;;) Line 814  for (;;)
814      match pointer to do this. */      match pointer to do this. */
815    
816      case OP_THEN:      case OP_THEN:
817      RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,      RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md,
818        eptrb, RM54);        eptrb, RM54);
819      if (rrc != MATCH_NOMATCH) RRETURN(rrc);      if (rrc != MATCH_NOMATCH) RRETURN(rrc);
820      md->start_match_ptr = ecode;      md->start_match_ptr = ecode;
# Line 788  for (;;) Line 823  for (;;)
823      case OP_THEN_ARG:      case OP_THEN_ARG:
824      md->nomatch_mark = ecode + 2;      md->nomatch_mark = ecode + 2;
825      md->mark = NULL;    /* In case previously set by assertion */      md->mark = NULL;    /* In case previously set by assertion */
826      RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode] + ecode[1], offset_top,      RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode] + ecode[1], offset_top,
827        md, eptrb, RM58);        md, eptrb, RM58);
828      if ((rrc == MATCH_MATCH || rrc == MATCH_ACCEPT) &&      if ((rrc == MATCH_MATCH || rrc == MATCH_ACCEPT) &&
829           md->mark == NULL) md->mark = ecode + 2;           md->mark == NULL) md->mark = ecode + 2;
# Line 812  for (;;) Line 847  for (;;)
847      case OP_ONCE_NC:      case OP_ONCE_NC:
848      prev = ecode;      prev = ecode;
849      saved_eptr = eptr;      saved_eptr = eptr;
850        save_mark = md->mark;
851      do      do
852        {        {
853        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, eptrb, RM64);        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, eptrb, RM64);
# Line 830  for (;;) Line 866  for (;;)
866    
867        if (rrc != MATCH_NOMATCH) RRETURN(rrc);        if (rrc != MATCH_NOMATCH) RRETURN(rrc);
868        ecode += GET(ecode,1);        ecode += GET(ecode,1);
869          md->mark = save_mark;
870        }        }
871      while (*ecode == OP_ALT);      while (*ecode == OP_ALT);
872    
# Line 869  for (;;) Line 906  for (;;)
906        }        }
907      else  /* OP_KETRMAX */      else  /* OP_KETRMAX */
908        {        {
       md->match_function_type = MATCH_CBEGROUP;  
909        RMATCH(eptr, prev, offset_top, md, eptrb, RM66);        RMATCH(eptr, prev, offset_top, md, eptrb, RM66);
910        if (rrc != MATCH_NOMATCH) RRETURN(rrc);        if (rrc != MATCH_NOMATCH) RRETURN(rrc);
911        ecode += 1 + LINK_SIZE;        ecode += 1 + LINK_SIZE;
# Line 909  for (;;) Line 945  for (;;)
945        save_offset2 = md->offset_vector[offset+1];        save_offset2 = md->offset_vector[offset+1];
946        save_offset3 = md->offset_vector[md->offset_end - number];        save_offset3 = md->offset_vector[md->offset_end - number];
947        save_capture_last = md->capture_last;        save_capture_last = md->capture_last;
948          save_mark = md->mark;
949    
950        DPRINTF(("saving %d %d %d\n", save_offset1, save_offset2, save_offset3));        DPRINTF(("saving %d %d %d\n", save_offset1, save_offset2, save_offset3));
951        md->offset_vector[md->offset_end - number] =        md->offset_vector[md->offset_end - number] =
# Line 917  for (;;) Line 954  for (;;)
954        for (;;)        for (;;)
955          {          {
956          if (op >= OP_SBRA) md->match_function_type = MATCH_CBEGROUP;          if (op >= OP_SBRA) md->match_function_type = MATCH_CBEGROUP;
957          RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,          RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md,
958            eptrb, RM1);            eptrb, RM1);
959          if (rrc == MATCH_ONCE) break;  /* Backing up through an atomic group */          if (rrc == MATCH_ONCE) break;  /* Backing up through an atomic group */
960    
# Line 945  for (;;) Line 982  for (;;)
982          if (rrc != MATCH_NOMATCH) RRETURN(rrc);          if (rrc != MATCH_NOMATCH) RRETURN(rrc);
983          md->capture_last = save_capture_last;          md->capture_last = save_capture_last;
984          ecode += GET(ecode, 1);          ecode += GET(ecode, 1);
985            md->mark = save_mark;
986          if (*ecode != OP_ALT) break;          if (*ecode != OP_ALT) break;
987          }          }
988    
# Line 996  for (;;) Line 1034  for (;;)
1034    
1035      for (;;)      for (;;)
1036        {        {
1037        if (op >= OP_SBRA || op == OP_ONCE) md->match_function_type = MATCH_CBEGROUP;        if (op >= OP_SBRA || op == OP_ONCE)
1038            md->match_function_type = MATCH_CBEGROUP;
1039    
1040        /* If this is not a possibly empty group, and there are no (*THEN)s in        /* If this is not a possibly empty group, and there are no (*THEN)s in
1041        the pattern, and this is the final alternative, optimize as described        the pattern, and this is the final alternative, optimize as described
# Line 1004  for (;;) Line 1043  for (;;)
1043    
1044        else if (!md->hasthen && ecode[GET(ecode, 1)] != OP_ALT)        else if (!md->hasthen && ecode[GET(ecode, 1)] != OP_ALT)
1045          {          {
1046          ecode += _pcre_OP_lengths[*ecode];          ecode += PRIV(OP_lengths)[*ecode];
1047          goto TAIL_RECURSE;          goto TAIL_RECURSE;
1048          }          }
1049    
1050        /* In all other cases, we have to make another call to match(). */        /* In all other cases, we have to make another call to match(). */
1051    
1052        RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md, eptrb,        save_mark = md->mark;
1053          RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md, eptrb,
1054          RM2);          RM2);
1055    
1056        /* See comment in the code for capturing groups above about handling        /* See comment in the code for capturing groups above about handling
# Line 1028  for (;;) Line 1068  for (;;)
1068          {          {
1069          if (rrc == MATCH_ONCE)          if (rrc == MATCH_ONCE)
1070            {            {
1071            const uschar *scode = ecode;            const pcre_uchar *scode = ecode;
1072            if (*scode != OP_ONCE)           /* If not at start, find it */            if (*scode != OP_ONCE)           /* If not at start, find it */
1073              {              {
1074              while (*scode == OP_ALT) scode += GET(scode, 1);              while (*scode == OP_ALT) scode += GET(scode, 1);
# Line 1039  for (;;) Line 1079  for (;;)
1079          RRETURN(rrc);          RRETURN(rrc);
1080          }          }
1081        ecode += GET(ecode, 1);        ecode += GET(ecode, 1);
1082          md->mark = save_mark;
1083        if (*ecode != OP_ALT) break;        if (*ecode != OP_ALT) break;
1084        }        }
1085    
# Line 1093  for (;;) Line 1134  for (;;)
1134          md->offset_vector[md->offset_end - number] =          md->offset_vector[md->offset_end - number] =
1135            (int)(eptr - md->start_subject);            (int)(eptr - md->start_subject);
1136          if (op >= OP_SBRA) md->match_function_type = MATCH_CBEGROUP;          if (op >= OP_SBRA) md->match_function_type = MATCH_CBEGROUP;
1137          RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,          RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md,
1138            eptrb, RM63);            eptrb, RM63);
1139          if (rrc == MATCH_KETRPOS)          if (rrc == MATCH_KETRPOS)
1140            {            {
# Line 1165  for (;;) Line 1206  for (;;)
1206      for (;;)      for (;;)
1207        {        {
1208        if (op >= OP_SBRA) md->match_function_type = MATCH_CBEGROUP;        if (op >= OP_SBRA) md->match_function_type = MATCH_CBEGROUP;
1209        RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,        RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md,
1210          eptrb, RM48);          eptrb, RM48);
1211        if (rrc == MATCH_KETRPOS)        if (rrc == MATCH_KETRPOS)
1212          {          {
# Line 1215  for (;;) Line 1256  for (;;)
1256    
1257      if (ecode[LINK_SIZE+1] == OP_CALLOUT)      if (ecode[LINK_SIZE+1] == OP_CALLOUT)
1258        {        {
1259        if (pcre_callout != NULL)        if (PUBL(callout) != NULL)
1260          {          {
1261          pcre_callout_block cb;          PUBL(callout_block) cb;
1262          cb.version          = 2;   /* Version 1 of the callout block */          cb.version          = 2;   /* Version 1 of the callout block */
1263          cb.callout_number   = ecode[LINK_SIZE+2];          cb.callout_number   = ecode[LINK_SIZE+2];
1264          cb.offset_vector    = md->offset_vector;          cb.offset_vector    = md->offset_vector;
1265    #ifdef COMPILE_PCRE8
1266          cb.subject          = (PCRE_SPTR)md->start_subject;          cb.subject          = (PCRE_SPTR)md->start_subject;
1267    #else
1268            cb.subject          = (PCRE_SPTR16)md->start_subject;
1269    #endif
1270          cb.subject_length   = (int)(md->end_subject - md->start_subject);          cb.subject_length   = (int)(md->end_subject - md->start_subject);
1271          cb.start_match      = (int)(mstart - md->start_subject);          cb.start_match      = (int)(mstart - md->start_subject);
1272          cb.current_position = (int)(eptr - md->start_subject);          cb.current_position = (int)(eptr - md->start_subject);
# Line 1231  for (;;) Line 1276  for (;;)
1276          cb.capture_last     = md->capture_last;          cb.capture_last     = md->capture_last;
1277          cb.callout_data     = md->callout_data;          cb.callout_data     = md->callout_data;
1278          cb.mark             = md->nomatch_mark;          cb.mark             = md->nomatch_mark;
1279          if ((rrc = (*pcre_callout)(&cb)) > 0) RRETURN(MATCH_NOMATCH);          if ((rrc = (*PUBL(callout))(&cb)) > 0) RRETURN(MATCH_NOMATCH);
1280          if (rrc < 0) RRETURN(rrc);          if (rrc < 0) RRETURN(rrc);
1281          }          }
1282        ecode += _pcre_OP_lengths[OP_CALLOUT];        ecode += PRIV(OP_lengths)[OP_CALLOUT];
1283        }        }
1284    
1285      condcode = ecode[LINK_SIZE+1];      condcode = ecode[LINK_SIZE+1];
# Line 1260  for (;;) Line 1305  for (;;)
1305    
1306          if (!condition && condcode == OP_NRREF)          if (!condition && condcode == OP_NRREF)
1307            {            {
1308            uschar *slotA = md->name_table;            pcre_uchar *slotA = md->name_table;
1309            for (i = 0; i < md->name_count; i++)            for (i = 0; i < md->name_count; i++)
1310              {              {
1311              if (GET2(slotA, 0) == recno) break;              if (GET2(slotA, 0) == recno) break;
# Line 1273  for (;;) Line 1318  for (;;)
1318    
1319            if (i < md->name_count)            if (i < md->name_count)
1320              {              {
1321              uschar *slotB = slotA;              pcre_uchar *slotB = slotA;
1322              while (slotB > md->name_table)              while (slotB > md->name_table)
1323                {                {
1324                slotB -= md->name_entry_size;                slotB -= md->name_entry_size;
1325                if (strcmp((char *)slotA + 2, (char *)slotB + 2) == 0)                if (STRCMP_UC_UC(slotA + IMM2_SIZE, slotB + IMM2_SIZE) == 0)
1326                  {                  {
1327                  condition = GET2(slotB, 0) == md->recursive->group_num;                  condition = GET2(slotB, 0) == md->recursive->group_num;
1328                  if (condition) break;                  if (condition) break;
# Line 1293  for (;;) Line 1338  for (;;)
1338                for (i++; i < md->name_count; i++)                for (i++; i < md->name_count; i++)
1339                  {                  {
1340                  slotB += md->name_entry_size;                  slotB += md->name_entry_size;
1341                  if (strcmp((char *)slotA + 2, (char *)slotB + 2) == 0)                  if (STRCMP_UC_UC(slotA + IMM2_SIZE, slotB + IMM2_SIZE) == 0)
1342                    {                    {
1343                    condition = GET2(slotB, 0) == md->recursive->group_num;                    condition = GET2(slotB, 0) == md->recursive->group_num;
1344                    if (condition) break;                    if (condition) break;
# Line 1306  for (;;) Line 1351  for (;;)
1351    
1352          /* Chose branch according to the condition */          /* Chose branch according to the condition */
1353    
1354          ecode += condition? 3 : GET(ecode, 1);          ecode += condition? 1 + IMM2_SIZE : GET(ecode, 1);
1355          }          }
1356        }        }
1357    
# Line 1323  for (;;) Line 1368  for (;;)
1368        if (!condition && condcode == OP_NCREF)        if (!condition && condcode == OP_NCREF)
1369          {          {
1370          int refno = offset >> 1;          int refno = offset >> 1;
1371          uschar *slotA = md->name_table;          pcre_uchar *slotA = md->name_table;
1372    
1373          for (i = 0; i < md->name_count; i++)          for (i = 0; i < md->name_count; i++)
1374            {            {
# Line 1337  for (;;) Line 1382  for (;;)
1382    
1383          if (i < md->name_count)          if (i < md->name_count)
1384            {            {
1385            uschar *slotB = slotA;            pcre_uchar *slotB = slotA;
1386            while (slotB > md->name_table)            while (slotB > md->name_table)
1387              {              {
1388              slotB -= md->name_entry_size;              slotB -= md->name_entry_size;
1389              if (strcmp((char *)slotA + 2, (char *)slotB + 2) == 0)              if (STRCMP_UC_UC(slotA + IMM2_SIZE, slotB + IMM2_SIZE) == 0)
1390                {                {
1391                offset = GET2(slotB, 0) << 1;                offset = GET2(slotB, 0) << 1;
1392                condition = offset < offset_top &&                condition = offset < offset_top &&
# Line 1359  for (;;) Line 1404  for (;;)
1404              for (i++; i < md->name_count; i++)              for (i++; i < md->name_count; i++)
1405                {                {
1406                slotB += md->name_entry_size;                slotB += md->name_entry_size;
1407                if (strcmp((char *)slotA + 2, (char *)slotB + 2) == 0)                if (STRCMP_UC_UC(slotA + IMM2_SIZE, slotB + IMM2_SIZE) == 0)
1408                  {                  {
1409                  offset = GET2(slotB, 0) << 1;                  offset = GET2(slotB, 0) << 1;
1410                  condition = offset < offset_top &&                  condition = offset < offset_top &&
# Line 1374  for (;;) Line 1419  for (;;)
1419    
1420        /* Chose branch according to the condition */        /* Chose branch according to the condition */
1421    
1422        ecode += condition? 3 : GET(ecode, 1);        ecode += condition? 1 + IMM2_SIZE : GET(ecode, 1);
1423        }        }
1424    
1425      else if (condcode == OP_DEF)     /* DEFINE - always false */      else if (condcode == OP_DEF)     /* DEFINE - always false */
# Line 1466  for (;;) Line 1511  for (;;)
1511        md->offset_vector[offset+1] = (int)(eptr - md->start_subject);        md->offset_vector[offset+1] = (int)(eptr - md->start_subject);
1512        if (offset_top <= offset) offset_top = offset + 2;        if (offset_top <= offset) offset_top = offset + 2;
1513        }        }
1514      ecode += 3;      ecode += 1 + IMM2_SIZE;
1515      break;      break;
1516    
1517    
# Line 1513  for (;;) Line 1558  for (;;)
1558    
1559      case OP_ASSERT:      case OP_ASSERT:
1560      case OP_ASSERTBACK:      case OP_ASSERTBACK:
1561        save_mark = md->mark;
1562      if (md->match_function_type == MATCH_CONDASSERT)      if (md->match_function_type == MATCH_CONDASSERT)
1563        {        {
1564        condassert = TRUE;        condassert = TRUE;
# Line 1528  for (;;) Line 1574  for (;;)
1574          mstart = md->start_match_ptr;   /* In case \K reset it */          mstart = md->start_match_ptr;   /* In case \K reset it */
1575          break;          break;
1576          }          }
1577          md->mark = save_mark;
1578    
1579        /* PCRE does not allow THEN to escape beyond an assertion; it is treated        /* A COMMIT failure must fail the entire assertion, without trying any
1580        as NOMATCH. */        subsequent branches. */
1581    
1582          if (rrc == MATCH_COMMIT) RRETURN(MATCH_NOMATCH);
1583    
1584          /* PCRE does not allow THEN to escape beyond an assertion; it
1585          is treated as NOMATCH. */
1586    
1587        if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN) RRETURN(rrc);        if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN) RRETURN(rrc);
1588        ecode += GET(ecode, 1);        ecode += GET(ecode, 1);
# Line 1557  for (;;) Line 1609  for (;;)
1609    
1610      case OP_ASSERT_NOT:      case OP_ASSERT_NOT:
1611      case OP_ASSERTBACK_NOT:      case OP_ASSERTBACK_NOT:
1612        save_mark = md->mark;
1613      if (md->match_function_type == MATCH_CONDASSERT)      if (md->match_function_type == MATCH_CONDASSERT)
1614        {        {
1615        condassert = TRUE;        condassert = TRUE;
# Line 1567  for (;;) Line 1620  for (;;)
1620      do      do
1621        {        {
1622        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, NULL, RM5);        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, NULL, RM5);
1623          md->mark = save_mark;
1624        if (rrc == MATCH_MATCH || rrc == MATCH_ACCEPT) RRETURN(MATCH_NOMATCH);        if (rrc == MATCH_MATCH || rrc == MATCH_ACCEPT) RRETURN(MATCH_NOMATCH);
1625        if (rrc == MATCH_SKIP || rrc == MATCH_PRUNE || rrc == MATCH_COMMIT)        if (rrc == MATCH_SKIP || rrc == MATCH_PRUNE || rrc == MATCH_COMMIT)
1626          {          {
# Line 1593  for (;;) Line 1647  for (;;)
1647      back a number of characters, not bytes. */      back a number of characters, not bytes. */
1648    
1649      case OP_REVERSE:      case OP_REVERSE:
1650  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
1651      if (utf8)      if (utf)
1652        {        {
1653        i = GET(ecode, 1);        i = GET(ecode, 1);
1654        while (i-- > 0)        while (i-- > 0)
# Line 1625  for (;;) Line 1679  for (;;)
1679      function is able to force a failure. */      function is able to force a failure. */
1680    
1681      case OP_CALLOUT:      case OP_CALLOUT:
1682      if (pcre_callout != NULL)      if (PUBL(callout) != NULL)
1683        {        {
1684        pcre_callout_block cb;        PUBL(callout_block) cb;
1685        cb.version          = 2;   /* Version 1 of the callout block */        cb.version          = 2;   /* Version 1 of the callout block */
1686        cb.callout_number   = ecode[1];        cb.callout_number   = ecode[1];
1687        cb.offset_vector    = md->offset_vector;        cb.offset_vector    = md->offset_vector;
1688    #ifdef COMPILE_PCRE8
1689        cb.subject          = (PCRE_SPTR)md->start_subject;        cb.subject          = (PCRE_SPTR)md->start_subject;
1690    #else
1691          cb.subject          = (PCRE_SPTR16)md->start_subject;
1692    #endif
1693        cb.subject_length   = (int)(md->end_subject - md->start_subject);        cb.subject_length   = (int)(md->end_subject - md->start_subject);
1694        cb.start_match      = (int)(mstart - md->start_subject);        cb.start_match      = (int)(mstart - md->start_subject);
1695        cb.current_position = (int)(eptr - md->start_subject);        cb.current_position = (int)(eptr - md->start_subject);
# Line 1641  for (;;) Line 1699  for (;;)
1699        cb.capture_last     = md->capture_last;        cb.capture_last     = md->capture_last;
1700        cb.callout_data     = md->callout_data;        cb.callout_data     = md->callout_data;
1701        cb.mark             = md->nomatch_mark;        cb.mark             = md->nomatch_mark;
1702        if ((rrc = (*pcre_callout)(&cb)) > 0) RRETURN(MATCH_NOMATCH);        if ((rrc = (*PUBL(callout))(&cb)) > 0) RRETURN(MATCH_NOMATCH);
1703        if (rrc < 0) RRETURN(rrc);        if (rrc < 0) RRETURN(rrc);
1704        }        }
1705      ecode += 2 + 2*LINK_SIZE;      ecode += 2 + 2*LINK_SIZE;
# Line 1700  for (;;) Line 1758  for (;;)
1758        else        else
1759          {          {
1760          new_recursive.offset_save =          new_recursive.offset_save =
1761            (int *)(pcre_malloc)(new_recursive.saved_max * sizeof(int));            (int *)(PUBL(malloc))(new_recursive.saved_max * sizeof(int));
1762          if (new_recursive.offset_save == NULL) RRETURN(PCRE_ERROR_NOMEMORY);          if (new_recursive.offset_save == NULL) RRETURN(PCRE_ERROR_NOMEMORY);
1763          }          }
1764        memcpy(new_recursive.offset_save, md->offset_vector,        memcpy(new_recursive.offset_save, md->offset_vector,
# Line 1715  for (;;) Line 1773  for (;;)
1773        do        do
1774          {          {
1775          if (cbegroup) md->match_function_type = MATCH_CBEGROUP;          if (cbegroup) md->match_function_type = MATCH_CBEGROUP;
1776          RMATCH(eptr, callpat + _pcre_OP_lengths[*callpat], offset_top,          RMATCH(eptr, callpat + PRIV(OP_lengths)[*callpat], offset_top,
1777            md, eptrb, RM6);            md, eptrb, RM6);
1778          memcpy(md->offset_vector, new_recursive.offset_save,          memcpy(md->offset_vector, new_recursive.offset_save,
1779              new_recursive.saved_max * sizeof(int));              new_recursive.saved_max * sizeof(int));
# Line 1724  for (;;) Line 1782  for (;;)
1782            {            {
1783            DPRINTF(("Recursion matched\n"));            DPRINTF(("Recursion matched\n"));
1784            if (new_recursive.offset_save != stacksave)            if (new_recursive.offset_save != stacksave)
1785              (pcre_free)(new_recursive.offset_save);              (PUBL(free))(new_recursive.offset_save);
1786    
1787            /* Set where we got to in the subject, and reset the start in case            /* Set where we got to in the subject, and reset the start in case
1788            it was changed by \K. This *is* propagated back out of a recursion,            it was changed by \K. This *is* propagated back out of a recursion,
# Line 1735  for (;;) Line 1793  for (;;)
1793            goto RECURSION_MATCHED;        /* Exit loop; end processing */            goto RECURSION_MATCHED;        /* Exit loop; end processing */
1794            }            }
1795    
1796          /* PCRE does not allow THEN to escape beyond a recursion; it is treated          /* PCRE does not allow THEN or COMMIT to escape beyond a recursion; it
1797          as NOMATCH. */          is treated as NOMATCH. */
1798    
1799          else if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN)          else if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN &&
1800                     rrc != MATCH_COMMIT)
1801            {            {
1802            DPRINTF(("Recursion gave error %d\n", rrc));            DPRINTF(("Recursion gave error %d\n", rrc));
1803            if (new_recursive.offset_save != stacksave)            if (new_recursive.offset_save != stacksave)
1804              (pcre_free)(new_recursive.offset_save);              (PUBL(free))(new_recursive.offset_save);
1805            RRETURN(rrc);            RRETURN(rrc);
1806            }            }
1807    
# Line 1754  for (;;) Line 1813  for (;;)
1813        DPRINTF(("Recursion didn't match\n"));        DPRINTF(("Recursion didn't match\n"));
1814        md->recursive = new_recursive.prevrec;        md->recursive = new_recursive.prevrec;
1815        if (new_recursive.offset_save != stacksave)        if (new_recursive.offset_save != stacksave)
1816          (pcre_free)(new_recursive.offset_save);          (PUBL(free))(new_recursive.offset_save);
1817        RRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
1818        }        }
1819    
# Line 1949  for (;;) Line 2008  for (;;)
2008          }          }
2009        if (*prev >= OP_SBRA)    /* Could match an empty string */        if (*prev >= OP_SBRA)    /* Could match an empty string */
2010          {          {
         md->match_function_type = MATCH_CBEGROUP;  
2011          RMATCH(eptr, prev, offset_top, md, eptrb, RM50);          RMATCH(eptr, prev, offset_top, md, eptrb, RM50);
2012          RRETURN(rrc);          RRETURN(rrc);
2013          }          }
# Line 1958  for (;;) Line 2016  for (;;)
2016        }        }
2017      else  /* OP_KETRMAX */      else  /* OP_KETRMAX */
2018        {        {
       if (*prev >= OP_SBRA) md->match_function_type = MATCH_CBEGROUP;  
2019        RMATCH(eptr, prev, offset_top, md, eptrb, RM13);        RMATCH(eptr, prev, offset_top, md, eptrb, RM13);
2020        if (rrc == MATCH_ONCE && md->once_target == prev) rrc = MATCH_NOMATCH;        if (rrc == MATCH_ONCE && md->once_target == prev) rrc = MATCH_NOMATCH;
2021        if (rrc != MATCH_NOMATCH) RRETURN(rrc);        if (rrc != MATCH_NOMATCH) RRETURN(rrc);
# Line 2015  for (;;) Line 2072  for (;;)
2072    
2073      case OP_DOLLM:      case OP_DOLLM:
2074      if (eptr < md->end_subject)      if (eptr < md->end_subject)
2075        { if (!IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH); }        {
2076          if (!IS_NEWLINE(eptr))
2077            {
2078            if (md->partial != 0 &&
2079                eptr + 1 >= md->end_subject &&
2080                NLBLOCK->nltype == NLTYPE_FIXED &&
2081                NLBLOCK->nllen == 2 &&
2082                *eptr == NLBLOCK->nl[0])
2083              {
2084              md->hitend = TRUE;
2085              if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
2086              }
2087            RRETURN(MATCH_NOMATCH);
2088            }
2089          }
2090      else      else
2091        {        {
2092        if (md->noteol) RRETURN(MATCH_NOMATCH);        if (md->noteol) RRETURN(MATCH_NOMATCH);
# Line 2047  for (;;) Line 2118  for (;;)
2118      ASSERT_NL_OR_EOS:      ASSERT_NL_OR_EOS:
2119      if (eptr < md->end_subject &&      if (eptr < md->end_subject &&
2120          (!IS_NEWLINE(eptr) || eptr != md->end_subject - md->nllen))          (!IS_NEWLINE(eptr) || eptr != md->end_subject - md->nllen))
2121          {
2122          if (md->partial != 0 &&
2123              eptr + 1 >= md->end_subject &&
2124              NLBLOCK->nltype == NLTYPE_FIXED &&
2125              NLBLOCK->nllen == 2 &&
2126              *eptr == NLBLOCK->nl[0])
2127            {
2128            md->hitend = TRUE;
2129            if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
2130            }
2131        RRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2132          }
2133    
2134      /* Either at end of string or \n before end. */      /* Either at end of string or \n before end. */
2135    
# Line 2066  for (;;) Line 2148  for (;;)
2148        be "non-word" characters. Remember the earliest consulted character for        be "non-word" characters. Remember the earliest consulted character for
2149        partial matching. */        partial matching. */
2150    
2151  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
2152        if (utf8)        if (utf)
2153          {          {
2154          /* Get status of previous character */          /* Get status of previous character */
2155    
2156          if (eptr == md->start_subject) prev_is_word = FALSE; else          if (eptr == md->start_subject) prev_is_word = FALSE; else
2157            {            {
2158            USPTR lastptr = eptr - 1;            PCRE_PUCHAR lastptr = eptr - 1;
2159            while((*lastptr & 0xc0) == 0x80) lastptr--;            BACKCHAR(lastptr);
2160            if (lastptr < md->start_used_ptr) md->start_used_ptr = lastptr;            if (lastptr < md->start_used_ptr) md->start_used_ptr = lastptr;
2161            GETCHAR(c, lastptr);            GETCHAR(c, lastptr);
2162  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
# Line 2139  for (;;) Line 2221  for (;;)
2221              }              }
2222            else            else
2223  #endif  #endif
2224            prev_is_word = ((md->ctypes[eptr[-1]] & ctype_word) != 0);            prev_is_word = MAX_255(eptr[-1])
2225                && ((md->ctypes[eptr[-1]] & ctype_word) != 0);
2226            }            }
2227    
2228          /* Get status of next character */          /* Get status of next character */
# Line 2162  for (;;) Line 2245  for (;;)
2245            }            }
2246          else          else
2247  #endif  #endif
2248          cur_is_word = ((md->ctypes[*eptr] & ctype_word) != 0);          cur_is_word = MAX_255(*eptr)
2249              && ((md->ctypes[*eptr] & ctype_word) != 0);
2250          }          }
2251    
2252        /* Now see if the situation is what we want */        /* Now see if the situation is what we want */
# Line 2173  for (;;) Line 2257  for (;;)
2257        }        }
2258      break;      break;
2259    
2260      /* Match a single character type; inline for speed */      /* Match any single character type except newline; have to take care with
2261        CRLF newlines and partial matching. */
2262    
2263      case OP_ANY:      case OP_ANY:
2264      if (IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH);      if (IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH);
2265        if (md->partial != 0 &&
2266            eptr + 1 >= md->end_subject &&
2267            NLBLOCK->nltype == NLTYPE_FIXED &&
2268            NLBLOCK->nllen == 2 &&
2269            *eptr == NLBLOCK->nl[0])
2270          {
2271          md->hitend = TRUE;
2272          if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
2273          }
2274    
2275      /* Fall through */      /* Fall through */
2276    
2277        /* Match any single character whatsoever. */
2278    
2279      case OP_ALLANY:      case OP_ALLANY:
2280      if (eptr >= md->end_subject)   /* DO NOT merge the eptr++ here; it must */      if (eptr >= md->end_subject)   /* DO NOT merge the eptr++ here; it must */
2281        {                            /* not be updated before SCHECK_PARTIAL. */        {                            /* not be updated before SCHECK_PARTIAL. */
# Line 2186  for (;;) Line 2283  for (;;)
2283        RRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2284        }        }
2285      eptr++;      eptr++;
2286      if (utf8) while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;  #ifdef SUPPORT_UTF
2287        if (utf) ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);
2288    #endif
2289      ecode++;      ecode++;
2290      break;      break;
2291    
# Line 2211  for (;;) Line 2310  for (;;)
2310        }        }
2311      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2312      if (      if (
2313  #ifdef SUPPORT_UTF8  #if defined SUPPORT_UTF || !(defined COMPILE_PCRE8)
2314         c < 256 &&         c < 256 &&
2315  #endif  #endif
2316         (md->ctypes[c] & ctype_digit) != 0         (md->ctypes[c] & ctype_digit) != 0
# Line 2228  for (;;) Line 2327  for (;;)
2327        }        }
2328      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2329      if (      if (
2330  #ifdef SUPPORT_UTF8  #if defined SUPPORT_UTF || !(defined COMPILE_PCRE8)
2331         c >= 256 ||         c > 255 ||
2332  #endif  #endif
2333         (md->ctypes[c] & ctype_digit) == 0         (md->ctypes[c] & ctype_digit) == 0
2334         )         )
# Line 2245  for (;;) Line 2344  for (;;)
2344        }        }
2345      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2346      if (      if (
2347  #ifdef SUPPORT_UTF8  #if defined SUPPORT_UTF || !(defined COMPILE_PCRE8)
2348         c < 256 &&         c < 256 &&
2349  #endif  #endif
2350         (md->ctypes[c] & ctype_space) != 0         (md->ctypes[c] & ctype_space) != 0
# Line 2262  for (;;) Line 2361  for (;;)
2361        }        }
2362      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2363      if (      if (
2364  #ifdef SUPPORT_UTF8  #if defined SUPPORT_UTF || !(defined COMPILE_PCRE8)
2365         c >= 256 ||         c > 255 ||
2366  #endif  #endif
2367         (md->ctypes[c] & ctype_space) == 0         (md->ctypes[c] & ctype_space) == 0
2368         )         )
# Line 2279  for (;;) Line 2378  for (;;)
2378        }        }
2379      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2380      if (      if (
2381  #ifdef SUPPORT_UTF8  #if defined SUPPORT_UTF || !(defined COMPILE_PCRE8)
2382         c < 256 &&         c < 256 &&
2383  #endif  #endif
2384         (md->ctypes[c] & ctype_word) != 0         (md->ctypes[c] & ctype_word) != 0
# Line 2296  for (;;) Line 2395  for (;;)
2395        }        }
2396      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2397      if (      if (
2398  #ifdef SUPPORT_UTF8  #if defined SUPPORT_UTF || !(defined COMPILE_PCRE8)
2399         c >= 256 ||         c > 255 ||
2400  #endif  #endif
2401         (md->ctypes[c] & ctype_word) == 0         (md->ctypes[c] & ctype_word) == 0
2402         )         )
# Line 2317  for (;;) Line 2416  for (;;)
2416        default: RRETURN(MATCH_NOMATCH);        default: RRETURN(MATCH_NOMATCH);
2417    
2418        case 0x000d:        case 0x000d:
2419        if (eptr < md->end_subject && *eptr == 0x0a) eptr++;        if (eptr >= md->end_subject)
2420            {
2421            SCHECK_PARTIAL();
2422            }
2423          else if (*eptr == 0x0a) eptr++;
2424        break;        break;
2425    
2426        case 0x000a:        case 0x000a:
# Line 2475  for (;;) Line 2578  for (;;)
2578          break;          break;
2579    
2580          case PT_GC:          case PT_GC:
2581          if ((ecode[2] != _pcre_ucp_gentype[prop->chartype]) == (op == OP_PROP))          if ((ecode[2] != PRIV(ucp_gentype)[prop->chartype]) == (op == OP_PROP))
2582            RRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
2583          break;          break;
2584    
# Line 2492  for (;;) Line 2595  for (;;)
2595          /* These are specials */          /* These are specials */
2596    
2597          case PT_ALNUM:          case PT_ALNUM:
2598          if ((_pcre_ucp_gentype[prop->chartype] == ucp_L ||          if ((PRIV(ucp_gentype)[prop->chartype] == ucp_L ||
2599               _pcre_ucp_gentype[prop->chartype] == ucp_N) == (op == OP_NOTPROP))               PRIV(ucp_gentype)[prop->chartype] == ucp_N) == (op == OP_NOTPROP))
2600            RRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
2601          break;          break;
2602    
2603          case PT_SPACE:    /* Perl space */          case PT_SPACE:    /* Perl space */
2604          if ((_pcre_ucp_gentype[prop->chartype] == ucp_Z ||          if ((PRIV(ucp_gentype)[prop->chartype] == ucp_Z ||
2605               c == CHAR_HT || c == CHAR_NL || c == CHAR_FF || c == CHAR_CR)               c == CHAR_HT || c == CHAR_NL || c == CHAR_FF || c == CHAR_CR)
2606                 == (op == OP_NOTPROP))                 == (op == OP_NOTPROP))
2607            RRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
2608          break;          break;
2609    
2610          case PT_PXSPACE:  /* POSIX space */          case PT_PXSPACE:  /* POSIX space */
2611          if ((_pcre_ucp_gentype[prop->chartype] == ucp_Z ||          if ((PRIV(ucp_gentype)[prop->chartype] == ucp_Z ||
2612               c == CHAR_HT || c == CHAR_NL || c == CHAR_VT ||               c == CHAR_HT || c == CHAR_NL || c == CHAR_VT ||
2613               c == CHAR_FF || c == CHAR_CR)               c == CHAR_FF || c == CHAR_CR)
2614                 == (op == OP_NOTPROP))                 == (op == OP_NOTPROP))
# Line 2513  for (;;) Line 2616  for (;;)
2616          break;          break;
2617    
2618          case PT_WORD:          case PT_WORD:
2619          if ((_pcre_ucp_gentype[prop->chartype] == ucp_L ||          if ((PRIV(ucp_gentype)[prop->chartype] == ucp_L ||
2620               _pcre_ucp_gentype[prop->chartype] == ucp_N ||               PRIV(ucp_gentype)[prop->chartype] == ucp_N ||
2621               c == CHAR_UNDERSCORE) == (op == OP_NOTPROP))               c == CHAR_UNDERSCORE) == (op == OP_NOTPROP))
2622            RRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
2623          break;          break;
# Line 2538  for (;;) Line 2641  for (;;)
2641        SCHECK_PARTIAL();        SCHECK_PARTIAL();
2642        RRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2643        }        }
2644      GETCHARINCTEST(c, eptr);      else
2645      if (UCD_CATEGORY(c) == ucp_M) RRETURN(MATCH_NOMATCH);        {
2646      while (eptr < md->end_subject)        int lgb, rgb;
2647        {        GETCHARINCTEST(c, eptr);
2648        int len = 1;        lgb = UCD_GRAPHBREAK(c);
2649        if (!utf8) c = *eptr; else { GETCHARLEN(c, eptr, len); }        while (eptr < md->end_subject)
2650        if (UCD_CATEGORY(c) != ucp_M) break;          {
2651        eptr += len;          int len = 1;
2652            if (!utf) c = *eptr; else { GETCHARLEN(c, eptr, len); }
2653            rgb = UCD_GRAPHBREAK(c);
2654            if (PRIV(ucp_gbtable)[lgb * ucp_gbCount + rgb] == 0) break;
2655            lgb = rgb;
2656            eptr += len;
2657            }
2658        }        }
2659        CHECK_PARTIAL();
2660      ecode++;      ecode++;
2661      break;      break;
2662  #endif  #endif
# Line 2564  for (;;) Line 2674  for (;;)
2674      case OP_REFI:      case OP_REFI:
2675      caseless = op == OP_REFI;      caseless = op == OP_REFI;
2676      offset = GET2(ecode, 1) << 1;               /* Doubled ref number */      offset = GET2(ecode, 1) << 1;               /* Doubled ref number */
2677      ecode += 3;      ecode += 1 + IMM2_SIZE;
2678    
2679      /* If the reference is unset, there are two possibilities:      /* If the reference is unset, there are two possibilities:
2680    
# Line 2604  for (;;) Line 2714  for (;;)
2714        case OP_CRMINRANGE:        case OP_CRMINRANGE:
2715        minimize = (*ecode == OP_CRMINRANGE);        minimize = (*ecode == OP_CRMINRANGE);
2716        min = GET2(ecode, 1);        min = GET2(ecode, 1);
2717        max = GET2(ecode, 3);        max = GET2(ecode, 1 + IMM2_SIZE);
2718        if (max == 0) max = INT_MAX;        if (max == 0) max = INT_MAX;
2719        ecode += 5;        ecode += 1 + 2 * IMM2_SIZE;
2720        break;        break;
2721    
2722        default:               /* No repeat follows */        default:               /* No repeat follows */
2723        if ((length = match_ref(offset, eptr, length, md, caseless)) < 0)        if ((length = match_ref(offset, eptr, length, md, caseless)) < 0)
2724          {          {
2725            if (length == -2) eptr = md->end_subject;   /* Partial match */
2726          CHECK_PARTIAL();          CHECK_PARTIAL();
2727          RRETURN(MATCH_NOMATCH);          RRETURN(MATCH_NOMATCH);
2728          }          }
# Line 2620  for (;;) Line 2731  for (;;)
2731        }        }
2732    
2733      /* Handle repeated back references. If the length of the reference is      /* Handle repeated back references. If the length of the reference is
2734      zero, just continue with the main loop. */      zero, just continue with the main loop. If the length is negative, it
2735        means the reference is unset in non-Java-compatible mode. If the minimum is
2736        zero, we can continue at the same level without recursion. For any other
2737        minimum, carrying on will result in NOMATCH. */
2738    
2739      if (length == 0) continue;      if (length == 0) continue;
2740        if (length < 0 && min == 0) continue;
2741    
2742      /* First, ensure the minimum number of matches are present. We get back      /* First, ensure the minimum number of matches are present. We get back
2743      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 2748  for (;;)
2748        int slength;        int slength;
2749        if ((slength = match_ref(offset, eptr, length, md, caseless)) < 0)        if ((slength = match_ref(offset, eptr, length, md, caseless)) < 0)
2750          {          {
2751            if (slength == -2) eptr = md->end_subject;   /* Partial match */
2752          CHECK_PARTIAL();          CHECK_PARTIAL();
2753          RRETURN(MATCH_NOMATCH);          RRETURN(MATCH_NOMATCH);
2754          }          }
# Line 2656  for (;;) Line 2772  for (;;)
2772          if (fi >= max) RRETURN(MATCH_NOMATCH);          if (fi >= max) RRETURN(MATCH_NOMATCH);
2773          if ((slength = match_ref(offset, eptr, length, md, caseless)) < 0)          if ((slength = match_ref(offset, eptr, length, md, caseless)) < 0)
2774            {            {
2775              if (slength == -2) eptr = md->end_subject;   /* Partial match */
2776            CHECK_PARTIAL();            CHECK_PARTIAL();
2777            RRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
2778            }            }
# Line 2674  for (;;) Line 2791  for (;;)
2791          int slength;          int slength;
2792          if ((slength = match_ref(offset, eptr, length, md, caseless)) < 0)          if ((slength = match_ref(offset, eptr, length, md, caseless)) < 0)
2793            {            {
2794            CHECK_PARTIAL();            /* Can't use CHECK_PARTIAL because we don't want to update eptr in
2795              the soft partial matching case. */
2796    
2797              if (slength == -2 && md->partial != 0 &&
2798                  md->end_subject > md->start_used_ptr)
2799                {
2800                md->hitend = TRUE;
2801                if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
2802                }
2803            break;            break;
2804            }            }
2805          eptr += slength;          eptr += slength;
2806          }          }
2807    
2808        while (eptr >= pp)        while (eptr >= pp)
2809          {          {
2810          RMATCH(eptr, ecode, offset_top, md, eptrb, RM15);          RMATCH(eptr, ecode, offset_top, md, eptrb, RM15);
# Line 2703  for (;;) Line 2829  for (;;)
2829      case OP_NCLASS:      case OP_NCLASS:
2830      case OP_CLASS:      case OP_CLASS:
2831        {        {
2832          /* The data variable is saved across frames, so the byte map needs to
2833          be stored there. */
2834    #define BYTE_MAP ((pcre_uint8 *)data)
2835        data = ecode + 1;                /* Save for matching */        data = ecode + 1;                /* Save for matching */
2836        ecode += 33;                     /* Advance past the item */        ecode += 1 + (32 / sizeof(pcre_uchar)); /* Advance past the item */
2837    
2838        switch (*ecode)        switch (*ecode)
2839          {          {
# Line 2725  for (;;) Line 2854  for (;;)
2854          case OP_CRMINRANGE:          case OP_CRMINRANGE:
2855          minimize = (*ecode == OP_CRMINRANGE);          minimize = (*ecode == OP_CRMINRANGE);
2856          min = GET2(ecode, 1);          min = GET2(ecode, 1);
2857          max = GET2(ecode, 3);          max = GET2(ecode, 1 + IMM2_SIZE);
2858          if (max == 0) max = INT_MAX;          if (max == 0) max = INT_MAX;
2859          ecode += 5;          ecode += 1 + 2 * IMM2_SIZE;
2860          break;          break;
2861    
2862          default:               /* No repeat follows */          default:               /* No repeat follows */
# Line 2737  for (;;) Line 2866  for (;;)
2866    
2867        /* First, ensure the minimum number of matches are present. */        /* First, ensure the minimum number of matches are present. */
2868    
2869  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
2870        /* UTF-8 mode */        if (utf)
       if (utf8)  
2871          {          {
2872          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
2873            {            {
# Line 2754  for (;;) Line 2882  for (;;)
2882              if (op == OP_CLASS) RRETURN(MATCH_NOMATCH);              if (op == OP_CLASS) RRETURN(MATCH_NOMATCH);
2883              }              }
2884            else            else
2885              {              if ((BYTE_MAP[c/8] & (1 << (c&7))) == 0) RRETURN(MATCH_NOMATCH);
             if ((data[c/8] & (1 << (c&7))) == 0) RRETURN(MATCH_NOMATCH);  
             }  
2886            }            }
2887          }          }
2888        else        else
2889  #endif  #endif
2890        /* Not UTF-8 mode */        /* Not UTF mode */
2891          {          {
2892          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
2893            {            {
# Line 2771  for (;;) Line 2897  for (;;)
2897              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
2898              }              }
2899            c = *eptr++;            c = *eptr++;
2900            if ((data[c/8] & (1 << (c&7))) == 0) RRETURN(MATCH_NOMATCH);  #ifndef COMPILE_PCRE8
2901              if (c > 255)
2902                {
2903                if (op == OP_CLASS) RRETURN(MATCH_NOMATCH);
2904                }
2905              else
2906    #endif
2907                if ((BYTE_MAP[c/8] & (1 << (c&7))) == 0) RRETURN(MATCH_NOMATCH);
2908            }            }
2909          }          }
2910    
# Line 2785  for (;;) Line 2918  for (;;)
2918    
2919        if (minimize)        if (minimize)
2920          {          {
2921  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
2922          /* UTF-8 mode */          if (utf)
         if (utf8)  
2923            {            {
2924            for (fi = min;; fi++)            for (fi = min;; fi++)
2925              {              {
# Line 2805  for (;;) Line 2937  for (;;)
2937                if (op == OP_CLASS) RRETURN(MATCH_NOMATCH);                if (op == OP_CLASS) RRETURN(MATCH_NOMATCH);
2938                }                }
2939              else              else
2940                {                if ((BYTE_MAP[c/8] & (1 << (c&7))) == 0) RRETURN(MATCH_NOMATCH);
               if ((data[c/8] & (1 << (c&7))) == 0) RRETURN(MATCH_NOMATCH);  
               }  
2941              }              }
2942            }            }
2943          else          else
2944  #endif  #endif
2945          /* Not UTF-8 mode */          /* Not UTF mode */
2946            {            {
2947            for (fi = min;; fi++)            for (fi = min;; fi++)
2948              {              {
# Line 2825  for (;;) Line 2955  for (;;)
2955                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
2956                }                }
2957              c = *eptr++;              c = *eptr++;
2958              if ((data[c/8] & (1 << (c&7))) == 0) RRETURN(MATCH_NOMATCH);  #ifndef COMPILE_PCRE8
2959                if (c > 255)
2960                  {
2961                  if (op == OP_CLASS) RRETURN(MATCH_NOMATCH);
2962                  }
2963                else
2964    #endif
2965                  if ((BYTE_MAP[c/8] & (1 << (c&7))) == 0) RRETURN(MATCH_NOMATCH);
2966              }              }
2967            }            }
2968          /* Control never gets here */          /* Control never gets here */
# Line 2837  for (;;) Line 2974  for (;;)
2974          {          {
2975          pp = eptr;          pp = eptr;
2976    
2977  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
2978          /* UTF-8 mode */          if (utf)
         if (utf8)  
2979            {            {
2980            for (i = min; i < max; i++)            for (i = min; i < max; i++)
2981              {              {
# Line 2855  for (;;) Line 2991  for (;;)
2991                if (op == OP_CLASS) break;                if (op == OP_CLASS) break;
2992                }                }
2993              else              else
2994                {                if ((BYTE_MAP[c/8] & (1 << (c&7))) == 0) break;
               if ((data[c/8] & (1 << (c&7))) == 0) break;  
               }  
2995              eptr += len;              eptr += len;
2996              }              }
2997            for (;;)            for (;;)
# Line 2870  for (;;) Line 3004  for (;;)
3004            }            }
3005          else          else
3006  #endif  #endif
3007            /* Not UTF-8 mode */            /* Not UTF mode */
3008            {            {
3009            for (i = min; i < max; i++)            for (i = min; i < max; i++)
3010              {              {
# Line 2880  for (;;) Line 3014  for (;;)
3014                break;                break;
3015                }                }
3016              c = *eptr;              c = *eptr;
3017              if ((data[c/8] & (1 << (c&7))) == 0) break;  #ifndef COMPILE_PCRE8
3018                if (c > 255)
3019                  {
3020                  if (op == OP_CLASS) break;
3021                  }
3022                else
3023    #endif
3024                  if ((BYTE_MAP[c/8] & (1 << (c&7))) == 0) break;
3025              eptr++;              eptr++;
3026              }              }
3027            while (eptr >= pp)            while (eptr >= pp)
# Line 2893  for (;;) Line 3034  for (;;)
3034    
3035          RRETURN(MATCH_NOMATCH);          RRETURN(MATCH_NOMATCH);
3036          }          }
3037    #undef BYTE_MAP
3038        }        }
3039      /* Control never gets here */      /* Control never gets here */
3040    
# Line 2901  for (;;) Line 3043  for (;;)
3043      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
3044      mode, because Unicode properties are supported in non-UTF-8 mode. */      mode, because Unicode properties are supported in non-UTF-8 mode. */
3045    
3046  #ifdef SUPPORT_UTF8  #if defined SUPPORT_UTF || !defined COMPILE_PCRE8
3047      case OP_XCLASS:      case OP_XCLASS:
3048        {        {
3049        data = ecode + 1 + LINK_SIZE;                /* Save for matching */        data = ecode + 1 + LINK_SIZE;                /* Save for matching */
# Line 2926  for (;;) Line 3068  for (;;)
3068          case OP_CRMINRANGE:          case OP_CRMINRANGE:
3069          minimize = (*ecode == OP_CRMINRANGE);          minimize = (*ecode == OP_CRMINRANGE);
3070          min = GET2(ecode, 1);          min = GET2(ecode, 1);
3071          max = GET2(ecode, 3);          max = GET2(ecode, 1 + IMM2_SIZE);
3072          if (max == 0) max = INT_MAX;          if (max == 0) max = INT_MAX;
3073          ecode += 5;          ecode += 1 + 2 * IMM2_SIZE;
3074          break;          break;
3075    
3076          default:               /* No repeat follows */          default:               /* No repeat follows */
# Line 2946  for (;;) Line 3088  for (;;)
3088            RRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
3089            }            }
3090          GETCHARINCTEST(c, eptr);          GETCHARINCTEST(c, eptr);
3091          if (!_pcre_xclass(c, data)) RRETURN(MATCH_NOMATCH);          if (!PRIV(xclass)(c, data, utf)) RRETURN(MATCH_NOMATCH);
3092          }          }
3093    
3094        /* 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 3112  for (;;)
3112              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
3113              }              }
3114            GETCHARINCTEST(c, eptr);            GETCHARINCTEST(c, eptr);
3115            if (!_pcre_xclass(c, data)) RRETURN(MATCH_NOMATCH);            if (!PRIV(xclass)(c, data, utf)) RRETURN(MATCH_NOMATCH);
3116            }            }
3117          /* Control never gets here */          /* Control never gets here */
3118          }          }
# Line 2988  for (;;) Line 3130  for (;;)
3130              SCHECK_PARTIAL();              SCHECK_PARTIAL();
3131              break;              break;
3132              }              }
3133    #ifdef SUPPORT_UTF
3134            GETCHARLENTEST(c, eptr, len);            GETCHARLENTEST(c, eptr, len);
3135            if (!_pcre_xclass(c, data)) break;  #else
3136              c = *eptr;
3137    #endif
3138              if (!PRIV(xclass)(c, data, utf)) break;
3139            eptr += len;            eptr += len;
3140            }            }
3141          for(;;)          for(;;)
# Line 2997  for (;;) Line 3143  for (;;)
3143            RMATCH(eptr, ecode, offset_top, md, eptrb, RM21);            RMATCH(eptr, ecode, offset_top, md, eptrb, RM21);
3144            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3145            if (eptr-- == pp) break;        /* Stop if tried at original pos */            if (eptr-- == pp) break;        /* Stop if tried at original pos */
3146            if (utf8) BACKCHAR(eptr);  #ifdef SUPPORT_UTF
3147              if (utf) BACKCHAR(eptr);
3148    #endif
3149            }            }
3150          RRETURN(MATCH_NOMATCH);          RRETURN(MATCH_NOMATCH);
3151          }          }
# Line 3009  for (;;) Line 3157  for (;;)
3157      /* Match a single character, casefully */      /* Match a single character, casefully */
3158    
3159      case OP_CHAR:      case OP_CHAR:
3160  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
3161      if (utf8)      if (utf)
3162        {        {
3163        length = 1;        length = 1;
3164        ecode++;        ecode++;
# Line 3024  for (;;) Line 3172  for (;;)
3172        }        }
3173      else      else
3174  #endif  #endif
3175        /* Not UTF mode */
     /* Non-UTF-8 mode */  
3176        {        {
3177        if (md->end_subject - eptr < 1)        if (md->end_subject - eptr < 1)
3178          {          {
# Line 3047  for (;;) Line 3194  for (;;)
3194        RRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
3195        }        }
3196    
3197  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
3198      if (utf8)      if (utf)
3199        {        {
3200        length = 1;        length = 1;
3201        ecode++;        ecode++;
# Line 3061  for (;;) Line 3208  for (;;)
3208    
3209        if (fc < 128)        if (fc < 128)
3210          {          {
3211          if (md->lcc[*ecode++] != md->lcc[*eptr++]) RRETURN(MATCH_NOMATCH);          if (md->lcc[fc]
3212                != TABLE_GET(*eptr, md->lcc, *eptr)) RRETURN(MATCH_NOMATCH);
3213            ecode++;
3214            eptr++;
3215          }          }
3216    
3217        /* Otherwise we must pick up the subject character. Note that we cannot        /* Otherwise we must pick up the subject character. Note that we cannot
# Line 3087  for (;;) Line 3237  for (;;)
3237          }          }
3238        }        }
3239      else      else
3240  #endif   /* SUPPORT_UTF8 */  #endif   /* SUPPORT_UTF */
3241    
3242      /* Non-UTF-8 mode */      /* Not UTF mode */
3243        {        {
3244        if (md->lcc[ecode[1]] != md->lcc[*eptr++]) RRETURN(MATCH_NOMATCH);        if (TABLE_GET(ecode[1], md->lcc, ecode[1])
3245              != TABLE_GET(*eptr, md->lcc, *eptr)) RRETURN(MATCH_NOMATCH);
3246          eptr++;
3247        ecode += 2;        ecode += 2;
3248        }        }
3249      break;      break;
# Line 3101  for (;;) Line 3253  for (;;)
3253      case OP_EXACT:      case OP_EXACT:
3254      case OP_EXACTI:      case OP_EXACTI:
3255      min = max = GET2(ecode, 1);      min = max = GET2(ecode, 1);
3256      ecode += 3;      ecode += 1 + IMM2_SIZE;
3257      goto REPEATCHAR;      goto REPEATCHAR;
3258    
3259      case OP_POSUPTO:      case OP_POSUPTO:
# Line 3116  for (;;) Line 3268  for (;;)
3268      min = 0;      min = 0;
3269      max = GET2(ecode, 1);      max = GET2(ecode, 1);
3270      minimize = *ecode == OP_MINUPTO || *ecode == OP_MINUPTOI;      minimize = *ecode == OP_MINUPTO || *ecode == OP_MINUPTOI;
3271      ecode += 3;      ecode += 1 + IMM2_SIZE;
3272      goto REPEATCHAR;      goto REPEATCHAR;
3273    
3274      case OP_POSSTAR:      case OP_POSSTAR:
# Line 3164  for (;;) Line 3316  for (;;)
3316      /* Common code for all repeated single-character matches. */      /* Common code for all repeated single-character matches. */
3317    
3318      REPEATCHAR:      REPEATCHAR:
3319  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
3320      if (utf8)      if (utf)
3321        {        {
3322        length = 1;        length = 1;
3323        charptr = ecode;        charptr = ecode;
# Line 3181  for (;;) Line 3333  for (;;)
3333          unsigned int othercase;          unsigned int othercase;
3334          if (op >= OP_STARI &&     /* Caseless */          if (op >= OP_STARI &&     /* Caseless */
3335              (othercase = UCD_OTHERCASE(fc)) != fc)              (othercase = UCD_OTHERCASE(fc)) != fc)
3336            oclength = _pcre_ord2utf8(othercase, occhars);            oclength = PRIV(ord2utf)(othercase, occhars);
3337          else oclength = 0;          else oclength = 0;
3338  #endif  /* SUPPORT_UCP */  #endif  /* SUPPORT_UCP */
3339    
3340          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
3341            {            {
3342            if (eptr <= md->end_subject - length &&            if (eptr <= md->end_subject - length &&
3343              memcmp(eptr, charptr, length) == 0) eptr += length;              memcmp(eptr, charptr, IN_UCHARS(length)) == 0) eptr += length;
3344  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
3345            else if (oclength > 0 &&            else if (oclength > 0 &&
3346                     eptr <= md->end_subject - oclength &&                     eptr <= md->end_subject - oclength &&
3347                     memcmp(eptr, occhars, oclength) == 0) eptr += oclength;                     memcmp(eptr, occhars, IN_UCHARS(oclength)) == 0) eptr += oclength;
3348  #endif  /* SUPPORT_UCP */  #endif  /* SUPPORT_UCP */
3349            else            else
3350              {              {
# Line 3211  for (;;) Line 3363  for (;;)
3363              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3364              if (fi >= max) RRETURN(MATCH_NOMATCH);              if (fi >= max) RRETURN(MATCH_NOMATCH);
3365              if (eptr <= md->end_subject - length &&              if (eptr <= md->end_subject - length &&
3366                memcmp(eptr, charptr, length) == 0) eptr += length;                memcmp(eptr, charptr, IN_UCHARS(length)) == 0) eptr += length;
3367  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
3368              else if (oclength > 0 &&              else if (oclength > 0 &&
3369                       eptr <= md->end_subject - oclength &&                       eptr <= md->end_subject - oclength &&
3370                       memcmp(eptr, occhars, oclength) == 0) eptr += oclength;                       memcmp(eptr, occhars, IN_UCHARS(oclength)) == 0) eptr += oclength;
3371  #endif  /* SUPPORT_UCP */  #endif  /* SUPPORT_UCP */
3372              else              else
3373                {                {
# Line 3232  for (;;) Line 3384  for (;;)
3384            for (i = min; i < max; i++)            for (i = min; i < max; i++)
3385              {              {
3386              if (eptr <= md->end_subject - length &&              if (eptr <= md->end_subject - length &&
3387                  memcmp(eptr, charptr, length) == 0) eptr += length;                  memcmp(eptr, charptr, IN_UCHARS(length)) == 0) eptr += length;
3388  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
3389              else if (oclength > 0 &&              else if (oclength > 0 &&
3390                       eptr <= md->end_subject - oclength &&                       eptr <= md->end_subject - oclength &&
3391                       memcmp(eptr, occhars, oclength) == 0) eptr += oclength;                       memcmp(eptr, occhars, IN_UCHARS(oclength)) == 0) eptr += oclength;
3392  #endif  /* SUPPORT_UCP */  #endif  /* SUPPORT_UCP */
3393              else              else
3394                {                {
# Line 3268  for (;;) Line 3420  for (;;)
3420        value of fc will always be < 128. */        value of fc will always be < 128. */
3421        }        }
3422      else      else
3423  #endif  /* SUPPORT_UTF8 */  #endif  /* SUPPORT_UTF */
3424          /* When not in UTF-8 mode, load a single-byte character. */
3425          fc = *ecode++;
3426    
3427      /* When not in UTF-8 mode, load a single-byte character. */      /* The value of fc at this point is always one character, though we may
3428        or may not be in UTF mode. The code is duplicated for the caseless and
     fc = *ecode++;  
   
     /* 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  
3429      caseful cases, for speed, since matching characters is likely to be quite      caseful cases, for speed, since matching characters is likely to be quite
3430      common. First, ensure the minimum number of matches are present. If min =      common. First, ensure the minimum number of matches are present. If min =
3431      max, continue at the same level without recursing. Otherwise, if      max, continue at the same level without recursing. Otherwise, if
# Line 3284  for (;;) Line 3434  for (;;)
3434      maximizing, find the maximum number of characters and work backwards. */      maximizing, find the maximum number of characters and work backwards. */
3435    
3436      DPRINTF(("matching %c{%d,%d} against subject %.*s\n", fc, min, max,      DPRINTF(("matching %c{%d,%d} against subject %.*s\n", fc, min, max,
3437        max, eptr));        max, (char *)eptr));
3438    
3439      if (op >= OP_STARI)  /* Caseless */      if (op >= OP_STARI)  /* Caseless */
3440        {        {
3441        fc = md->lcc[fc];  #ifdef COMPILE_PCRE8
3442          /* fc must be < 128 if UTF is enabled. */
3443          foc = md->fcc[fc];
3444    #else
3445    #ifdef SUPPORT_UTF
3446    #ifdef SUPPORT_UCP
3447          if (utf && fc > 127)
3448            foc = UCD_OTHERCASE(fc);
3449    #else
3450          if (utf && fc > 127)
3451            foc = fc;
3452    #endif /* SUPPORT_UCP */
3453          else
3454    #endif /* SUPPORT_UTF */
3455            foc = TABLE_GET(fc, md->fcc, fc);
3456    #endif /* COMPILE_PCRE8 */
3457    
3458        for (i = 1; i <= min; i++)        for (i = 1; i <= min; i++)
3459          {          {
3460          if (eptr >= md->end_subject)          if (eptr >= md->end_subject)
# Line 3296  for (;;) Line 3462  for (;;)
3462            SCHECK_PARTIAL();            SCHECK_PARTIAL();
3463            RRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
3464            }            }
3465          if (fc != md->lcc[*eptr++]) RRETURN(MATCH_NOMATCH);          if (fc != *eptr && foc != *eptr) RRETURN(MATCH_NOMATCH);
3466            eptr++;
3467          }          }
3468        if (min == max) continue;        if (min == max) continue;
3469        if (minimize)        if (minimize)
# Line 3311  for (;;) Line 3478  for (;;)
3478              SCHECK_PARTIAL();              SCHECK_PARTIAL();
3479              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
3480              }              }
3481            if (fc != md->lcc[*eptr++]) RRETURN(MATCH_NOMATCH);            if (fc != *eptr && foc != *eptr) RRETURN(MATCH_NOMATCH);
3482              eptr++;
3483            }            }
3484          /* Control never gets here */          /* Control never gets here */
3485          }          }
# Line 3325  for (;;) Line 3493  for (;;)
3493              SCHECK_PARTIAL();              SCHECK_PARTIAL();
3494              break;              break;
3495              }              }
3496            if (fc != md->lcc[*eptr]) break;            if (fc != *eptr && foc != *eptr) break;
3497            eptr++;            eptr++;
3498            }            }
3499    
# Line 3410  for (;;) Line 3578  for (;;)
3578        SCHECK_PARTIAL();        SCHECK_PARTIAL();
3579        RRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
3580        }        }
3581      ecode++;  #ifdef SUPPORT_UTF
3582      GETCHARINCTEST(c, eptr);      if (utf)
     if (op == OP_NOTI)         /* The caseless case */  
3583        {        {
3584  #ifdef SUPPORT_UTF8        register unsigned int ch, och;
3585        if (c < 256)  
3586  #endif        ecode++;
3587        c = md->lcc[c];        GETCHARINC(ch, ecode);
3588        if (md->lcc[*ecode++] == c) RRETURN(MATCH_NOMATCH);        GETCHARINC(c, eptr);
3589    
3590          if (op == OP_NOT)
3591            {
3592            if (ch == c) RRETURN(MATCH_NOMATCH);
3593            }
3594          else
3595            {
3596    #ifdef SUPPORT_UCP
3597            if (ch > 127)
3598              och = UCD_OTHERCASE(ch);
3599    #else
3600            if (ch > 127)
3601              och = ch;
3602    #endif /* SUPPORT_UCP */
3603            else
3604              och = TABLE_GET(ch, md->fcc, ch);
3605            if (ch == c || och == c) RRETURN(MATCH_NOMATCH);
3606            }
3607        }        }
3608      else    /* Caseful */      else
3609    #endif
3610        {        {
3611        if (*ecode++ == c) RRETURN(MATCH_NOMATCH);        register unsigned int ch = ecode[1];
3612          c = *eptr++;
3613          if (ch == c || (op == OP_NOTI && TABLE_GET(ch, md->fcc, ch) == c))
3614            RRETURN(MATCH_NOMATCH);
3615          ecode += 2;
3616        }        }
3617      break;      break;
3618    
# Line 3436  for (;;) Line 3626  for (;;)
3626      case OP_NOTEXACT:      case OP_NOTEXACT:
3627      case OP_NOTEXACTI:      case OP_NOTEXACTI:
3628      min = max = GET2(ecode, 1);      min = max = GET2(ecode, 1);
3629      ecode += 3;      ecode += 1 + IMM2_SIZE;
3630      goto REPEATNOTCHAR;      goto REPEATNOTCHAR;
3631    
3632      case OP_NOTUPTO:      case OP_NOTUPTO:
# Line 3446  for (;;) Line 3636  for (;;)
3636      min = 0;      min = 0;
3637      max = GET2(ecode, 1);      max = GET2(ecode, 1);
3638      minimize = *ecode == OP_NOTMINUPTO || *ecode == OP_NOTMINUPTOI;      minimize = *ecode == OP_NOTMINUPTO || *ecode == OP_NOTMINUPTOI;
3639      ecode += 3;      ecode += 1 + IMM2_SIZE;
3640      goto REPEATNOTCHAR;      goto REPEATNOTCHAR;
3641    
3642      case OP_NOTPOSSTAR:      case OP_NOTPOSSTAR:
# Line 3478  for (;;) Line 3668  for (;;)
3668      possessive = TRUE;      possessive = TRUE;
3669      min = 0;      min = 0;
3670      max = GET2(ecode, 1);      max = GET2(ecode, 1);
3671      ecode += 3;      ecode += 1 + IMM2_SIZE;
3672      goto REPEATNOTCHAR;      goto REPEATNOTCHAR;
3673    
3674      case OP_NOTSTAR:      case OP_NOTSTAR:
# Line 3502  for (;;) Line 3692  for (;;)
3692      /* Common code for all repeated single-byte matches. */      /* Common code for all repeated single-byte matches. */
3693    
3694      REPEATNOTCHAR:      REPEATNOTCHAR:
3695      fc = *ecode++;      GETCHARINCTEST(fc, ecode);
3696    
3697      /* The code is duplicated for the caseless and caseful cases, for speed,      /* The code is duplicated for the caseless and caseful cases, for speed,
3698      since matching characters is likely to be quite common. First, ensure the      since matching characters is likely to be quite common. First, ensure the
# Line 3513  for (;;) Line 3703  for (;;)
3703      characters and work backwards. */      characters and work backwards. */
3704    
3705      DPRINTF(("negative matching %c{%d,%d} against subject %.*s\n", fc, min, max,      DPRINTF(("negative matching %c{%d,%d} against subject %.*s\n", fc, min, max,
3706        max, eptr));        max, (char *)eptr));
3707    
3708      if (op >= OP_NOTSTARI)     /* Caseless */      if (op >= OP_NOTSTARI)     /* Caseless */
3709        {        {
3710        fc = md->lcc[fc];  #ifdef SUPPORT_UTF
3711    #ifdef SUPPORT_UCP
3712          if (utf && fc > 127)
3713            foc = UCD_OTHERCASE(fc);
3714    #else
3715          if (utf && fc > 127)
3716            foc = fc;
3717    #endif /* SUPPORT_UCP */
3718          else
3719    #endif /* SUPPORT_UTF */
3720            foc = TABLE_GET(fc, md->fcc, fc);
3721    
3722  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
3723        /* UTF-8 mode */        if (utf)
       if (utf8)  
3724          {          {
3725          register unsigned int d;          register unsigned int d;
3726          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
# Line 3532  for (;;) Line 3731  for (;;)
3731              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
3732              }              }
3733            GETCHARINC(d, eptr);            GETCHARINC(d, eptr);
3734            if (d < 256) d = md->lcc[d];            if (fc == d || (unsigned int)foc == d) RRETURN(MATCH_NOMATCH);
           if (fc == d) RRETURN(MATCH_NOMATCH);  
3735            }            }
3736          }          }
3737        else        else
3738  #endif  #endif
3739          /* Not UTF mode */
       /* Not UTF-8 mode */  
3740          {          {
3741          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
3742            {            {
# Line 3548  for (;;) Line 3745  for (;;)
3745              SCHECK_PARTIAL();              SCHECK_PARTIAL();
3746              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
3747              }              }
3748            if (fc == md->lcc[*eptr++]) RRETURN(MATCH_NOMATCH);            if (fc == *eptr || foc == *eptr) RRETURN(MATCH_NOMATCH);
3749              eptr++;
3750            }            }
3751          }          }
3752    
# Line 3556  for (;;) Line 3754  for (;;)
3754    
3755        if (minimize)        if (minimize)
3756          {          {
3757  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
3758          /* UTF-8 mode */          if (utf)
         if (utf8)  
3759            {            {
3760            register unsigned int d;            register unsigned int d;
3761            for (fi = min;; fi++)            for (fi = min;; fi++)
# Line 3572  for (;;) Line 3769  for (;;)
3769                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
3770                }                }
3771              GETCHARINC(d, eptr);              GETCHARINC(d, eptr);
3772              if (d < 256) d = md->lcc[d];              if (fc == d || (unsigned int)foc == d) RRETURN(MATCH_NOMATCH);
             if (fc == d) RRETURN(MATCH_NOMATCH);  
3773              }              }
3774            }            }
3775          else          else
3776  #endif  #endif
3777          /* Not UTF-8 mode */          /* Not UTF mode */
3778            {            {
3779            for (fi = min;; fi++)            for (fi = min;; fi++)
3780              {              {
# Line 3590  for (;;) Line 3786  for (;;)
3786                SCHECK_PARTIAL();                SCHECK_PARTIAL();
3787                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
3788                }                }
3789              if (fc == md->lcc[*eptr++]) RRETURN(MATCH_NOMATCH);              if (fc == *eptr || foc == *eptr) RRETURN(MATCH_NOMATCH);
3790                eptr++;
3791              }              }
3792            }            }
3793          /* Control never gets here */          /* Control never gets here */
# Line 3602  for (;;) Line 3799  for (;;)
3799          {          {
3800          pp = eptr;          pp = eptr;
3801    
3802  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
3803          /* UTF-8 mode */          if (utf)
         if (utf8)  
3804            {            {
3805            register unsigned int d;            register unsigned int d;
3806            for (i = min; i < max; i++)            for (i = min; i < max; i++)
# Line 3616  for (;;) Line 3812  for (;;)
3812                break;                break;
3813                }                }
3814              GETCHARLEN(d, eptr, len);              GETCHARLEN(d, eptr, len);
3815              if (d < 256) d = md->lcc[d];              if (fc == d || (unsigned int)foc == d) break;
             if (fc == d) break;  
3816              eptr += len;              eptr += len;
3817              }              }
3818          if (possessive) continue;            if (possessive) continue;
3819          for(;;)            for(;;)
3820              {              {
3821              RMATCH(eptr, ecode, offset_top, md, eptrb, RM30);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM30);
3822              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
# Line 3631  for (;;) Line 3826  for (;;)
3826            }            }
3827          else          else
3828  #endif  #endif
3829          /* Not UTF-8 mode */          /* Not UTF mode */
3830            {            {
3831            for (i = min; i < max; i++)            for (i = min; i < max; i++)
3832              {              {
# Line 3640  for (;;) Line 3835  for (;;)
3835                SCHECK_PARTIAL();                SCHECK_PARTIAL();
3836                break;                break;
3837                }                }
3838              if (fc == md->lcc[*eptr]) break;              if (fc == *eptr || foc == *eptr) break;
3839              eptr++;              eptr++;
3840              }              }
3841            if (possessive) continue;            if (possessive) continue;
# Line 3661  for (;;) Line 3856  for (;;)
3856    
3857      else      else
3858        {        {
3859  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
3860        /* UTF-8 mode */        if (utf)
       if (utf8)  
3861          {          {
3862          register unsigned int d;          register unsigned int d;
3863          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
# Line 3679  for (;;) Line 3873  for (;;)
3873          }          }
3874        else        else
3875  #endif  #endif
3876        /* Not UTF-8 mode */        /* Not UTF mode */
3877          {          {
3878          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
3879            {            {
# Line 3696  for (;;) Line 3890  for (;;)
3890    
3891        if (minimize)        if (minimize)
3892          {          {
3893  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
3894          /* UTF-8 mode */          if (utf)
         if (utf8)  
3895            {            {
3896            register unsigned int d;            register unsigned int d;
3897            for (fi = min;; fi++)            for (fi = min;; fi++)
# Line 3717  for (;;) Line 3910  for (;;)
3910            }            }
3911          else          else
3912  #endif  #endif
3913          /* Not UTF-8 mode */          /* Not UTF mode */
3914            {            {
3915            for (fi = min;; fi++)            for (fi = min;; fi++)
3916              {              {
# Line 3741  for (;;) Line 3934  for (;;)
3934          {          {
3935          pp = eptr;          pp = eptr;
3936    
3937  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
3938          /* UTF-8 mode */          if (utf)
         if (utf8)  
3939            {            {
3940            register unsigned int d;            register unsigned int d;
3941            for (i = min; i < max; i++)            for (i = min; i < max; i++)
# Line 3769  for (;;) Line 3961  for (;;)
3961            }            }
3962          else          else
3963  #endif  #endif
3964          /* Not UTF-8 mode */          /* Not UTF mode */
3965            {            {
3966            for (i = min; i < max; i++)            for (i = min; i < max; i++)
3967              {              {
# Line 3802  for (;;) Line 3994  for (;;)
3994      case OP_TYPEEXACT:      case OP_TYPEEXACT:
3995      min = max = GET2(ecode, 1);      min = max = GET2(ecode, 1);
3996      minimize = TRUE;      minimize = TRUE;
3997      ecode += 3;      ecode += 1 + IMM2_SIZE;
3998      goto REPEATTYPE;      goto REPEATTYPE;
3999    
4000      case OP_TYPEUPTO:      case OP_TYPEUPTO:
# Line 3810  for (;;) Line 4002  for (;;)
4002      min = 0;      min = 0;
4003      max = GET2(ecode, 1);      max = GET2(ecode, 1);
4004      minimize = *ecode == OP_TYPEMINUPTO;      minimize = *ecode == OP_TYPEMINUPTO;
4005      ecode += 3;      ecode += 1 + IMM2_SIZE;
4006      goto REPEATTYPE;      goto REPEATTYPE;
4007    
4008      case OP_TYPEPOSSTAR:      case OP_TYPEPOSSTAR:
# Line 3838  for (;;) Line 4030  for (;;)
4030      possessive = TRUE;      possessive = TRUE;
4031      min = 0;      min = 0;
4032      max = GET2(ecode, 1);      max = GET2(ecode, 1);
4033      ecode += 3;      ecode += 1 + IMM2_SIZE;
4034      goto REPEATTYPE;      goto REPEATTYPE;
4035    
4036      case OP_TYPESTAR:      case OP_TYPESTAR:
# Line 4040  for (;;) Line 4232  for (;;)
4232              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4233              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4234              }              }
4235            GETCHARINCTEST(c, eptr);            else
4236            if (UCD_CATEGORY(c) == ucp_M) RRETURN(MATCH_NOMATCH);              {
4237            while (eptr < md->end_subject)              int lgb, rgb;
4238              {              GETCHARINCTEST(c, eptr);
4239              int len = 1;              lgb = UCD_GRAPHBREAK(c);
4240              if (!utf8) c = *eptr; else { GETCHARLEN(c, eptr, len); }             while (eptr < md->end_subject)
4241              if (UCD_CATEGORY(c) != ucp_M) break;                {
4242              eptr += len;                int len = 1;
4243                  if (!utf) c = *eptr; else { GETCHARLEN(c, eptr, len); }
4244                  rgb = UCD_GRAPHBREAK(c);
4245                  if (PRIV(ucp_gbtable)[lgb * ucp_gbCount + rgb] == 0) break;
4246                  lgb = rgb;
4247                  eptr += len;
4248                  }
4249              }              }
4250              CHECK_PARTIAL();
4251            }            }
4252          }          }
4253    
# Line 4057  for (;;) Line 4256  for (;;)
4256    
4257  /* Handle all other cases when the coding is UTF-8 */  /* Handle all other cases when the coding is UTF-8 */
4258    
4259  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
4260        if (utf8) switch(ctype)        if (utf) switch(ctype)
4261          {          {
4262          case OP_ANY:          case OP_ANY:
4263          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
# Line 4069  for (;;) Line 4268  for (;;)
4268              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4269              }              }
4270            if (IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH);            if (IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH);
4271              if (md->partial != 0 &&
4272                  eptr + 1 >= md->end_subject &&
4273                  NLBLOCK->nltype == NLTYPE_FIXED &&
4274                  NLBLOCK->nllen == 2 &&
4275                  *eptr == NLBLOCK->nl[0])
4276                {
4277                md->hitend = TRUE;
4278                if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
4279                }
4280            eptr++;            eptr++;
4281            while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;            ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);
4282            }            }
4283          break;          break;
4284    
# Line 4083  for (;;) Line 4291  for (;;)
4291              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4292              }              }
4293            eptr++;            eptr++;
4294            while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;            ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);
4295            }            }
4296          break;          break;
4297    
# Line 4265  for (;;) Line 4473  for (;;)
4473              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4474              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4475              }              }
4476            if (*eptr >= 128 || (md->ctypes[*eptr++] & ctype_digit) == 0)            if (*eptr >= 128 || (md->ctypes[*eptr] & ctype_digit) == 0)
4477              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4478              eptr++;
4479            /* 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 */
4480            }            }
4481          break;          break;
# Line 4281  for (;;) Line 4490  for (;;)
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            while (++eptr < md->end_subject && (*eptr & 0xc0) == 0x80);            eptr++;
4494              ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);
4495            }            }
4496          break;          break;
4497    
# Line 4293  for (;;) Line 4503  for (;;)
4503              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4504              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4505              }              }
4506            if (*eptr >= 128 || (md->ctypes[*eptr++] & ctype_space) == 0)            if (*eptr >= 128 || (md->ctypes[*eptr] & ctype_space) == 0)
4507              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4508              eptr++;
4509            /* 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 */
4510            }            }
4511          break;          break;
# Line 4309  for (;;) Line 4520  for (;;)
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            while (++eptr < md->end_subject && (*eptr & 0xc0) == 0x80);            eptr++;
4524              ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);
4525            }            }
4526          break;          break;
4527    
# Line 4321  for (;;) Line 4533  for (;;)
4533              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4534              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4535              }              }
4536            if (*eptr >= 128 || (md->ctypes[*eptr++] & ctype_word) == 0)            if (*eptr >= 128 || (md->ctypes[*eptr] & ctype_word) == 0)
4537              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4538              eptr++;
4539            /* 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 */
4540            }            }
4541          break;          break;
# Line 4332  for (;;) Line 4545  for (;;)
4545          }  /* End switch(ctype) */          }  /* End switch(ctype) */
4546    
4547        else        else
4548  #endif     /* SUPPORT_UTF8 */  #endif     /* SUPPORT_UTF */
4549    
4550        /* 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
4551        than OP_PROP and OP_NOTPROP. */        than OP_PROP and OP_NOTPROP. */
# Line 4348  for (;;) Line 4561  for (;;)
4561              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4562              }              }
4563            if (IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH);            if (IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH);
4564              if (md->partial != 0 &&
4565                  eptr + 1 >= md->end_subject &&
4566                  NLBLOCK->nltype == NLTYPE_FIXED &&
4567                  NLBLOCK->nllen == 2 &&
4568                  *eptr == NLBLOCK->nl[0])
4569                {
4570                md->hitend = TRUE;
4571                if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
4572                }
4573            eptr++;            eptr++;
4574            }            }
4575          break;          break;
# Line 4392  for (;;) Line 4614  for (;;)
4614              case 0x000b:              case 0x000b:
4615              case 0x000c:              case 0x000c:
4616              case 0x0085:              case 0x0085:
4617    #ifdef COMPILE_PCRE16
4618                case 0x2028:
4619                case 0x2029:
4620    #endif
4621              if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);              if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);
4622              break;              break;
4623              }              }
# Line 4412  for (;;) Line 4638  for (;;)
4638              case 0x09:      /* HT */              case 0x09:      /* HT */
4639              case 0x20:      /* SPACE */              case 0x20:      /* SPACE */
4640              case 0xa0:      /* NBSP */              case 0xa0:      /* NBSP */
4641    #ifdef COMPILE_PCRE16
4642                case 0x1680:    /* OGHAM SPACE MARK */
4643                case 0x180e:    /* MONGOLIAN VOWEL SEPARATOR */
4644                case 0x2000:    /* EN QUAD */
4645                case 0x2001:    /* EM QUAD */
4646                case 0x2002:    /* EN SPACE */
4647                case 0x2003:    /* EM SPACE */
4648                case 0x2004:    /* THREE-PER-EM SPACE */
4649                case 0x2005:    /* FOUR-PER-EM SPACE */
4650                case 0x2006:    /* SIX-PER-EM SPACE */
4651                case 0x2007:    /* FIGURE SPACE */
4652                case 0x2008:    /* PUNCTUATION SPACE */
4653                case 0x2009:    /* THIN SPACE */
4654                case 0x200A:    /* HAIR SPACE */
4655                case 0x202f:    /* NARROW NO-BREAK SPACE */
4656                case 0x205f:    /* MEDIUM MATHEMATICAL SPACE */
4657                case 0x3000:    /* IDEOGRAPHIC SPACE */
4658    #endif
4659              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4660              }              }
4661            }            }
# Line 4431  for (;;) Line 4675  for (;;)
4675              case 0x09:      /* HT */              case 0x09:      /* HT */
4676              case 0x20:      /* SPACE */              case 0x20:      /* SPACE */
4677              case 0xa0:      /* NBSP */              case 0xa0:      /* NBSP */
4678    #ifdef COMPILE_PCRE16
4679                case 0x1680:    /* OGHAM SPACE MARK */
4680                case 0x180e:    /* MONGOLIAN VOWEL SEPARATOR */
4681                case 0x2000:    /* EN QUAD */
4682                case 0x2001:    /* EM QUAD */
4683                case 0x2002:    /* EN SPACE */
4684                case 0x2003:    /* EM SPACE */
4685                case 0x2004:    /* THREE-PER-EM SPACE */
4686                case 0x2005:    /* FOUR-PER-EM SPACE */
4687                case 0x2006:    /* SIX-PER-EM SPACE */
4688                case 0x2007:    /* FIGURE SPACE */
4689                case 0x2008:    /* PUNCTUATION SPACE */
4690                case 0x2009:    /* THIN SPACE */
4691                case 0x200A:    /* HAIR SPACE */
4692                case 0x202f:    /* NARROW NO-BREAK SPACE */
4693                case 0x205f:    /* MEDIUM MATHEMATICAL SPACE */
4694                case 0x3000:    /* IDEOGRAPHIC SPACE */
4695    #endif
4696              break;              break;
4697              }              }
4698            }            }
# Line 4452  for (;;) Line 4714  for (;;)
4714              case 0x0c:      /* FF */              case 0x0c:      /* FF */
4715              case 0x0d:      /* CR */              case 0x0d:      /* CR */
4716              case 0x85:      /* NEL */              case 0x85:      /* NEL */
4717    #ifdef COMPILE_PCRE16
4718                case 0x2028:    /* LINE SEPARATOR */
4719                case 0x2029:    /* PARAGRAPH SEPARATOR */
4720    #endif
4721              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4722              }              }
4723            }            }
# Line 4473  for (;;) Line 4739  for (;;)
4739              case 0x0c:      /* FF */              case 0x0c:      /* FF */
4740              case 0x0d:      /* CR */              case 0x0d:      /* CR */
4741              case 0x85:      /* NEL */              case 0x85:      /* NEL */
4742    #ifdef COMPILE_PCRE16
4743                case 0x2028:    /* LINE SEPARATOR */
4744                case 0x2029:    /* PARAGRAPH SEPARATOR */
4745    #endif
4746              break;              break;
4747              }              }
4748            }            }
# Line 4486  for (;;) Line 4756  for (;;)
4756              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4757              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4758              }              }
4759            if ((md->ctypes[*eptr++] & ctype_digit) != 0) RRETURN(MATCH_NOMATCH);            if (MAX_255(*eptr) && (md->ctypes[*eptr] & ctype_digit) != 0)
4760                RRETURN(MATCH_NOMATCH);
4761              eptr++;
4762            }            }
4763          break;          break;
4764    
# Line 4498  for (;;) Line 4770  for (;;)
4770              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4771              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4772              }              }
4773            if ((md->ctypes[*eptr++] & ctype_digit) == 0) RRETURN(MATCH_NOMATCH);            if (!MAX_255(*eptr) || (md->ctypes[*eptr] & ctype_digit) == 0)
4774                RRETURN(MATCH_NOMATCH);
4775              eptr++;
4776            }            }
4777          break;          break;
4778    
# Line 4510  for (;;) Line 4784  for (;;)
4784              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4785              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4786              }              }
4787            if ((md->ctypes[*eptr++] & ctype_space) != 0) RRETURN(MATCH_NOMATCH);            if (MAX_255(*eptr) && (md->ctypes[*eptr] & ctype_space) != 0)
4788                RRETURN(MATCH_NOMATCH);
4789              eptr++;
4790            }            }
4791          break;          break;
4792    
# Line 4522  for (;;) Line 4798  for (;;)
4798              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4799              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4800              }              }
4801            if ((md->ctypes[*eptr++] & ctype_space) == 0) RRETURN(MATCH_NOMATCH);            if (!MAX_255(*eptr) || (md->ctypes[*eptr] & ctype_space) == 0)
4802                RRETURN(MATCH_NOMATCH);
4803              eptr++;
4804            }            }
4805          break;          break;
4806    
# Line 4534  for (;;) Line 4812  for (;;)
4812              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4813              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4814              }              }
4815            if ((md->ctypes[*eptr++] & ctype_word) != 0)            if (MAX_255(*eptr) && (md->ctypes[*eptr] & ctype_word) != 0)
4816              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4817              eptr++;
4818            }            }
4819          break;          break;
4820    
# Line 4547  for (;;) Line 4826  for (;;)
4826              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4827              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4828              }              }
4829            if ((md->ctypes[*eptr++] & ctype_word) == 0)            if (!MAX_255(*eptr) || (md->ctypes[*eptr] & ctype_word) == 0)
4830              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4831              eptr++;
4832            }            }
4833          break;          break;
4834    
# Line 4761  for (;;) Line 5041  for (;;)
5041              SCHECK_PARTIAL();              SCHECK_PARTIAL();
5042              RRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
5043              }              }
5044            GETCHARINCTEST(c, eptr);            else
5045            if (UCD_CATEGORY(c) == ucp_M) RRETURN(MATCH_NOMATCH);              {
5046            while (eptr < md->end_subject)              int lgb, rgb;
5047              {              GETCHARINCTEST(c, eptr);
5048              int len = 1;              lgb = UCD_GRAPHBREAK(c);
5049              if (!utf8) c = *eptr; else { GETCHARLEN(c, eptr, len); }              while (eptr < md->end_subject)
5050              if (UCD_CATEGORY(c) != ucp_M) break;                {
5051              eptr += len;                int len = 1;
5052                  if (!utf) c = *eptr; else { GETCHARLEN(c, eptr, len); }
5053                  rgb = UCD_GRAPHBREAK(c);
5054                  if (PRIV(ucp_gbtable)[lgb * ucp_gbCount + rgb] == 0) break;
5055                  lgb = rgb;
5056                  eptr += len;
5057                  }
5058              }              }
5059              CHECK_PARTIAL();
5060            }            }
5061          }          }
5062        else        else
5063  #endif     /* SUPPORT_UCP */  #endif     /* SUPPORT_UCP */
5064    
5065  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
5066        /* UTF-8 mode */        if (utf)
       if (utf8)  
5067          {          {
5068          for (fi = min;; fi++)          for (fi = min;; fi++)
5069            {            {
# Line 4794  for (;;) Line 5080  for (;;)
5080            GETCHARINC(c, eptr);            GETCHARINC(c, eptr);
5081            switch(ctype)            switch(ctype)
5082              {              {
5083              case OP_ANY:        /* This is the non-NL case */              case OP_ANY:               /* This is the non-NL case */
5084                if (md->partial != 0 &&    /* Take care with CRLF partial */
5085                    eptr >= md->end_subject &&
5086                    NLBLOCK->nltype == NLTYPE_FIXED &&
5087                    NLBLOCK->nllen == 2 &&
5088                    c == NLBLOCK->nl[0])
5089                  {
5090                  md->hitend = TRUE;
5091                  if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
5092                  }
5093                break;
5094    
5095              case OP_ALLANY:              case OP_ALLANY:
5096              case OP_ANYBYTE:              case OP_ANYBYTE:
5097              break;              break;
# Line 4919  for (;;) Line 5216  for (;;)
5216              break;              break;
5217    
5218              case OP_WHITESPACE:              case OP_WHITESPACE:
5219              if  (c >= 256 || (md->ctypes[c] & ctype_space) == 0)              if (c >= 256 || (md->ctypes[c] & ctype_space) == 0)
5220                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
5221              break;              break;
5222    
# Line 4940  for (;;) Line 5237  for (;;)
5237          }          }
5238        else        else
5239  #endif  #endif
5240        /* Not UTF-8 mode */        /* Not UTF mode */
5241          {          {
5242          for (fi = min;; fi++)          for (fi = min;; fi++)
5243            {            {
# Line 4957  for (;;) Line 5254  for (;;)
5254            c = *eptr++;            c = *eptr++;
5255            switch(ctype)            switch(ctype)
5256              {              {
5257              case OP_ANY:     /* This is the non-NL case */              case OP_ANY:               /* This is the non-NL case */
5258                if (md->partial != 0 &&    /* Take care with CRLF partial */
5259                    eptr >= md->end_subject &&
5260                    NLBLOCK->nltype == NLTYPE_FIXED &&
5261                    NLBLOCK->nllen == 2 &&
5262                    c == NLBLOCK->nl[0])
5263                  {
5264                  md->hitend = TRUE;
5265                  if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
5266                  }
5267                break;
5268    
5269              case OP_ALLANY:              case OP_ALLANY:
5270              case OP_ANYBYTE:              case OP_ANYBYTE:
5271              break;              break;
# Line 4976  for (;;) Line 5284  for (;;)
5284                case 0x000b:                case 0x000b:
5285                case 0x000c:                case 0x000c:
5286                case 0x0085:                case 0x0085:
5287    #ifdef COMPILE_PCRE16
5288                  case 0x2028:
5289                  case 0x2029:
5290    #endif
5291                if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);                if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);
5292                break;                break;
5293                }                }
# Line 4988  for (;;) Line 5300  for (;;)
5300                case 0x09:      /* HT */                case 0x09:      /* HT */
5301                case 0x20:      /* SPACE */                case 0x20:      /* SPACE */
5302                case 0xa0:      /* NBSP */                case 0xa0:      /* NBSP */
5303    #ifdef COMPILE_PCRE16
5304                  case 0x1680:    /* OGHAM SPACE MARK */
5305                  case 0x180e:    /* MONGOLIAN VOWEL SEPARATOR */
5306                  case 0x2000:    /* EN QUAD */
5307                  case 0x2001:    /* EM QUAD */
5308                  case 0x2002:    /* EN SPACE */
5309                  case 0x2003:    /* EM SPACE */
5310                  case 0x2004:    /* THREE-PER-EM SPACE */
5311                  case 0x2005:    /* FOUR-PER-EM SPACE */
5312                  case 0x2006:    /* SIX-PER-EM SPACE */
5313                  case 0x2007:    /* FIGURE SPACE */
5314                  case 0x2008:    /* PUNCTUATION SPACE */
5315                  case 0x2009:    /* THIN SPACE */
5316                  case 0x200A:    /* HAIR SPACE */
5317                  case 0x202f:    /* NARROW NO-BREAK SPACE */
5318                  case 0x205f:    /* MEDIUM MATHEMATICAL SPACE */
5319                  case 0x3000:    /* IDEOGRAPHIC SPACE */
5320    #endif
5321                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
5322                }                }
5323              break;              break;
# Line 4999  for (;;) Line 5329  for (;;)
5329                case 0x09:      /* HT */                case 0x09:      /* HT */
5330                case 0x20:      /* SPACE */                case 0x20:      /* SPACE */
5331                case 0xa0:      /* NBSP */                case 0xa0:      /* NBSP */
5332    #ifdef COMPILE_PCRE16
5333                  case 0x1680:    /* OGHAM SPACE MARK */
5334                  case 0x180e:    /* MONGOLIAN VOWEL SEPARATOR */
5335                  case 0x2000:    /* EN QUAD */
5336                  case 0x2001:    /* EM QUAD */
5337                  case 0x2002:    /* EN SPACE */
5338                  case 0x2003:    /* EM SPACE */
5339                  case 0x2004:    /* THREE-PER-EM SPACE */
5340                  case 0x2005:    /* FOUR-PER-EM SPACE */
5341                  case 0x2006:    /* SIX-PER-EM SPACE */
5342                  case 0x2007:    /* FIGURE SPACE */
5343                  case 0x2008:    /* PUNCTUATION SPACE */
5344                  case 0x2009:    /* THIN SPACE */
5345                  case 0x200A:    /* HAIR SPACE */
5346                  case 0x202f:    /* NARROW NO-BREAK SPACE */
5347                  case 0x205f:    /* MEDIUM MATHEMATICAL SPACE */
5348                  case 0x3000:    /* IDEOGRAPHIC SPACE */
5349    #endif
5350                break;                break;
5351                }                }
5352              break;              break;
# Line 5012  for (;;) Line 5360  for (;;)
5360                case 0x0c:      /* FF */                case 0x0c:      /* FF */
5361                case 0x0d:      /* CR */                case 0x0d:      /* CR */
5362                case 0x85:      /* NEL */                case 0x85:      /* NEL */
5363    #ifdef COMPILE_PCRE16
5364                  case 0x2028:    /* LINE SEPARATOR */
5365                  case 0x2029:    /* PARAGRAPH SEPARATOR */
5366    #endif
5367                RRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
5368                }                }
5369              break;              break;
# Line 5025  for (;;) Line 5377  for (;;)
5377                case 0x0c:      /* FF */                case 0x0c:      /* FF */
5378                case 0x0d:      /* CR */                case 0x0d:      /* CR */
5379                case 0x85:      /* NEL */                case 0x85:      /* NEL */
5380    #ifdef COMPILE_PCRE16
5381                  case 0x2028:    /* LINE SEPARATOR */
5382                  case 0x2029:    /* PARAGRAPH SEPARATOR */
5383    #endif
5384                break;                break;
5385                }                }
5386              break;              break;
5387    
5388              case OP_NOT_DIGIT:              case OP_NOT_DIGIT:
5389              if ((md->ctypes[c] & ctype_digit) != 0) RRETURN(MATCH_NOMATCH);              if (MAX_255(c) && (md->ctypes[c] & ctype_digit) != 0) RRETURN(MATCH_NOMATCH);
5390              break;              break;
5391    
5392              case OP_DIGIT:              case OP_DIGIT:
5393              if ((md->ctypes[c] & ctype_digit) == 0) RRETURN(MATCH_NOMATCH);              if (!MAX_255(c) || (md->ctypes[c] & ctype_digit) == 0) RRETURN(MATCH_NOMATCH);
5394              break;              break;
5395    
5396              case OP_NOT_WHITESPACE:              case OP_NOT_WHITESPACE:
5397              if ((md->ctypes[c] & ctype_space) != 0) RRETURN(MATCH_NOMATCH);              if (MAX_255(c) && (md->ctypes[c] & ctype_space) != 0) RRETURN(MATCH_NOMATCH);
5398              break;              break;
5399    
5400              case OP_WHITESPACE:              case OP_WHITESPACE:
5401              if  ((md->ctypes[c] & ctype_space) == 0) RRETURN(MATCH_NOMATCH);              if (!MAX_255(c) || (md->ctypes[c] & ctype_space) == 0) RRETURN(MATCH_NOMATCH);
5402              break;              break;
5403    
5404              case OP_NOT_WORDCHAR:              case OP_NOT_WORDCHAR:
5405              if ((md->ctypes[c] & ctype_word) != 0) RRETURN(MATCH_NOMATCH);              if (MAX_255(c) && (md->ctypes[c] & ctype_word) != 0) RRETURN(MATCH_NOMATCH);
5406              break;              break;
5407    
5408              case OP_WORDCHAR:              case OP_WORDCHAR:
5409              if ((md->ctypes[c] & ctype_word) == 0) RRETURN(MATCH_NOMATCH);              if (!MAX_255(c) || (md->ctypes[c] & ctype_word) == 0) RRETURN(MATCH_NOMATCH);
5410              break;              break;
5411    
5412              default:              default:
# Line 5239  for (;;) Line 5595  for (;;)
5595            RMATCH(eptr, ecode, offset_top, md, eptrb, RM44);            RMATCH(eptr, ecode, offset_top, md, eptrb, RM44);
5596            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
5597            if (eptr-- == pp) break;        /* Stop if tried at original pos */            if (eptr-- == pp) break;        /* Stop if tried at original pos */
5598            if (utf8) BACKCHAR(eptr);            if (utf) BACKCHAR(eptr);
5599            }            }
5600          }          }
5601    
# Line 5250  for (;;) Line 5606  for (;;)
5606          {          {
5607          for (i = min; i < max; i++)          for (i = min; i < max; i++)
5608            {            {
           int len = 1;  
5609            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
5610              {              {
5611              SCHECK_PARTIAL();              SCHECK_PARTIAL();
5612              break;              break;
5613              }              }
5614            if (!utf8) c = *eptr; else { GETCHARLEN(c, eptr, len); }            else
5615            if (UCD_CATEGORY(c) == ucp_M) break;              {
5616            eptr += len;              int lgb, rgb;
5617            while (eptr < md->end_subject)              GETCHARINCTEST(c, eptr);
5618              {              lgb = UCD_GRAPHBREAK(c);
5619              len = 1;              while (eptr < md->end_subject)
5620              if (!utf8) c = *eptr; else { GETCHARLEN(c, eptr, len); }                {
5621              if (UCD_CATEGORY(c) != ucp_M) break;                int len = 1;
5622              eptr += len;                if (!utf) c = *eptr; else { GETCHARLEN(c, eptr, len); }
5623                  rgb = UCD_GRAPHBREAK(c);
5624                  if (PRIV(ucp_gbtable)[lgb * ucp_gbCount + rgb] == 0) break;
5625                  lgb = rgb;
5626                  eptr += len;
5627                  }
5628              }              }
5629              CHECK_PARTIAL();
5630            }            }
5631    
5632          /* eptr is now past the end of the maximum run */          /* eptr is now past the end of the maximum run */
# Line 5279  for (;;) Line 5640  for (;;)
5640            if (eptr-- == pp) break;        /* Stop if tried at original pos */            if (eptr-- == pp) break;        /* Stop if tried at original pos */
5641            for (;;)                        /* Move back over one extended */            for (;;)                        /* Move back over one extended */
5642              {              {
5643              if (!utf8) c = *eptr; else              if (!utf) c = *eptr; else
5644                {                {
5645                BACKCHAR(eptr);                BACKCHAR(eptr);
5646                GETCHAR(c, eptr);                GETCHAR(c, eptr);
# Line 5293  for (;;) Line 5654  for (;;)
5654        else        else
5655  #endif   /* SUPPORT_UCP */  #endif   /* SUPPORT_UCP */
5656    
5657  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
5658        /* UTF-8 mode */        if (utf)
   
       if (utf8)  
5659          {          {
5660          switch(ctype)          switch(ctype)
5661            {            {
# Line 5311  for (;;) Line 5670  for (;;)
5670                  break;                  break;
5671                  }                  }
5672                if (IS_NEWLINE(eptr)) break;                if (IS_NEWLINE(eptr)) break;
5673                  if (md->partial != 0 &&    /* Take care with CRLF partial */
5674                      eptr + 1 >= md->end_subject &&
5675                      NLBLOCK->nltype == NLTYPE_FIXED &&
5676                      NLBLOCK->nllen == 2 &&
5677                      *eptr == NLBLOCK->nl[0])
5678                    {
5679                    md->hitend = TRUE;
5680                    if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
5681                    }
5682                eptr++;                eptr++;
5683                while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;                ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);
5684                }                }
5685              }              }
5686    
# Line 5328  for (;;) Line 5696  for (;;)
5696                  break;                  break;
5697                  }                  }
5698                if (IS_NEWLINE(eptr)) break;                if (IS_NEWLINE(eptr)) break;
5699                  if (md->partial != 0 &&    /* Take care with CRLF partial */
5700                      eptr + 1 >= md->end_subject &&
5701                      NLBLOCK->nltype == NLTYPE_FIXED &&
5702                      NLBLOCK->nllen == 2 &&
5703                      *eptr == NLBLOCK->nl[0])
5704                    {
5705                    md->hitend = TRUE;
5706                    if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
5707                    }
5708                eptr++;                eptr++;
5709                while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;                ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);
5710                }                }
5711              }              }
5712            break;            break;
# Line 5345  for (;;) Line 5722  for (;;)
5722                  break;                  break;
5723                  }                  }
5724                eptr++;                eptr++;
5725                while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;                ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);
5726                }                }
5727              }              }
5728            else            else
# Line 5578  for (;;) Line 5955  for (;;)
5955            }            }
5956          }          }
5957        else        else
5958  #endif  /* SUPPORT_UTF8 */  #endif  /* SUPPORT_UTF */
5959          /* Not UTF mode */
       /* Not UTF-8 mode */  
5960          {          {
5961          switch(ctype)          switch(ctype)
5962            {            {
# Line 5593  for (;;) Line 5969  for (;;)
5969                break;                break;
5970                }                }
5971              if (IS_NEWLINE(eptr)) break;              if (IS_NEWLINE(eptr)) break;
5972                if (md->partial != 0 &&    /* Take care with CRLF partial */
5973                    eptr + 1 >= md->end_subject &&
5974                    NLBLOCK->nltype == NLTYPE_FIXED &&
5975                    NLBLOCK->nllen == 2 &&
5976                    *eptr == NLBLOCK->nl[0])
5977                  {
5978                  md->hitend = TRUE;
5979                  if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
5980                  }
5981              eptr++;              eptr++;
5982              }              }
5983            break;            break;
# Line 5624  for (;;) Line 6009  for (;;)
6009                }                }
6010              else              else
6011                {                {
6012                if (c != 0x000a &&                if (c != 0x000a && (md->bsr_anycrlf ||
6013                    (md->bsr_anycrlf ||                  (c != 0x000b && c != 0x000c && c != 0x0085
6014                      (c != 0x000b && c != 0x000c && c != 0x0085)))  #ifdef COMPILE_PCRE16
6015                  break;                  && c != 0x2028 && c != 0x2029
6016    #endif
6017                    ))) break;
6018                eptr++;                eptr++;
6019                }                }
6020              }              }
# Line 5642  for (;;) Line 6029  for (;;)
6029                break;                break;
6030                }                }
6031              c = *eptr;              c = *eptr;
6032              if (c == 0x09 || c == 0x20 || c == 0xa0) break;              if (c == 0x09 || c == 0x20 || c == 0xa0
6033    #ifdef COMPILE_PCRE16
6034                  || c == 0x1680 || c == 0x180e || (c >= 0x2000 && c <= 0x200A)
6035                  || c == 0x202f || c == 0x205f || c == 0x3000
6036    #endif
6037                  ) break;
6038              eptr++;              eptr++;
6039              }              }
6040            break;            break;
# Line 5656  for (;;) Line 6048  for (;;)
6048                break;                break;
6049                }                }
6050              c = *eptr;              c = *eptr;
6051              if (c != 0x09 && c != 0x20 && c != 0xa0) break;              if (c != 0x09 && c != 0x20 && c != 0xa0
6052    #ifdef COMPILE_PCRE16
6053                  && c != 0x1680 && c != 0x180e && (c < 0x2000 || c > 0x200A)
6054                  && c != 0x202f && c != 0x205f && c != 0x3000
6055    #endif
6056                  ) break;
6057              eptr++;              eptr++;
6058              }              }
6059            break;            break;
# Line 5670  for (;;) Line 6067  for (;;)
6067                break;                break;
6068                }                }
6069              c = *eptr;              c = *eptr;
6070              if (c == 0x0a || c == 0x0b || c == 0x0c || c == 0x0d || c == 0x85)              if (c == 0x0a || c == 0x0b || c == 0x0c || c == 0x0d || c == 0x85
6071                break;  #ifdef COMPILE_PCRE16
6072                  || c == 0x2028 || c == 0x2029
6073    #endif
6074                  ) break;
6075              eptr++;              eptr++;
6076              }              }
6077            break;            break;
# Line 5685  for (;;) Line 6085  for (;;)
6085                break;                break;
6086                }                }
6087              c = *eptr;              c = *eptr;
6088              if (c != 0x0a && c != 0x0b && c != 0x0c && c != 0x0d && c != 0x85)              if (c != 0x0a && c != 0x0b && c != 0x0c && c != 0x0d && c != 0x85
6089                break;  #ifdef COMPILE_PCRE16
6090                  && c != 0x2028 && c != 0x2029
6091    #endif
6092                  ) break;
6093              eptr++;              eptr++;
6094              }              }
6095            break;            break;
# Line 5699  for (;;) Line 6102  for (;;)
6102                SCHECK_PARTIAL();                SCHECK_PARTIAL();
6103                break;                break;
6104                }                }
6105              if ((md->ctypes[*eptr] & ctype_digit) != 0) break;              if (MAX_255(*eptr) && (md->ctypes[*eptr] & ctype_digit) != 0) break;
6106              eptr++;              eptr++;
6107              }              }
6108            break;            break;
# Line 5712  for (;;) Line 6115  for (;;)
6115                SCHECK_PARTIAL();                SCHECK_PARTIAL();
6116                break;                break;
6117                }                }
6118              if ((md->ctypes[*eptr] & ctype_digit) == 0) break;              if (!MAX_255(*eptr) || (md->ctypes[*eptr] & ctype_digit) == 0) break;
6119              eptr++;              eptr++;
6120              }              }
6121            break;            break;
# Line 5725  for (;;) Line 6128  for (;;)
6128                SCHECK_PARTIAL();                SCHECK_PARTIAL();
6129                break;                break;
6130                }                }
6131              if ((md->ctypes[*eptr] & ctype_space) != 0) break;              if (MAX_255(*eptr) && (md->ctypes[*eptr] & ctype_space) != 0) break;
6132              eptr++;              eptr++;
6133              }              }
6134            break;            break;
# Line 5738  for (;;) Line 6141  for (;;)
6141                SCHECK_PARTIAL();                SCHECK_PARTIAL();
6142                break;                break;
6143                }                }
6144              if ((md->ctypes[*eptr] & ctype_space) == 0) break;              if (!MAX_255(*eptr) || (md->ctypes[*eptr] & ctype_space) == 0) break;
6145              eptr++;              eptr++;
6146              }              }
6147            break;            break;
# Line 5751  for (;;) Line 6154  for (;;)
6154                SCHECK_PARTIAL();                SCHECK_PARTIAL();
6155                break;                break;
6156                }                }
6157              if ((md->ctypes[*eptr] & ctype_word) != 0) break;              if (MAX_255(*eptr) && (md->ctypes[*eptr] & ctype_word) != 0) break;
6158              eptr++;              eptr++;
6159              }              }
6160            break;            break;
# Line 5764  for (;;) Line 6167  for (;;)
6167                SCHECK_PARTIAL();                SCHECK_PARTIAL();
6168                break;                break;
6169                }                }
6170              if ((md->ctypes[*eptr] & ctype_word) == 0) break;              if (!MAX_255(*eptr) || (md->ctypes[*eptr] & ctype_word) == 0) break;
6171              eptr++;              eptr++;
6172              }              }
6173            break;            break;
# Line 5827  switch (frame->Xwhere) Line 6230  switch (frame->Xwhere)
6230    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)
6231    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)
6232    LBL(65) LBL(66)    LBL(65) LBL(66)
6233  #ifdef SUPPORT_UTF8  #if defined SUPPORT_UTF || !defined COMPILE_PCRE8
6234    LBL(16) LBL(18) LBL(20) LBL(21) LBL(22) LBL(23) LBL(28) LBL(30)    LBL(21)
6235    #endif
6236    #ifdef SUPPORT_UTF
6237      LBL(16) LBL(18) LBL(20)
6238      LBL(22) LBL(23) LBL(28) LBL(30)
6239    LBL(32) LBL(34) LBL(42) LBL(46)    LBL(32) LBL(34) LBL(42) LBL(46)
6240  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
6241    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)
6242    LBL(59) LBL(60) LBL(61) LBL(62)    LBL(59) LBL(60) LBL(61) LBL(62)
6243  #endif  /* SUPPORT_UCP */  #endif  /* SUPPORT_UCP */
6244  #endif  /* SUPPORT_UTF8 */  #endif  /* SUPPORT_UTF */
6245    default:    default:
6246    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));
6247    
6248    printf("+++jump error in pcre match: label %d non-existent\n", frame->Xwhere);
6249    
6250    return PCRE_ERROR_INTERNAL;    return PCRE_ERROR_INTERNAL;
6251    }    }
6252  #undef LBL  #undef LBL
# Line 5898  Undefine all the macros that were define Line 6308  Undefine all the macros that were define
6308  ***************************************************************************/  ***************************************************************************/
6309    
6310    
6311    #ifdef NO_RECURSE
6312    /*************************************************
6313    *          Release allocated heap frames         *
6314    *************************************************/
6315    
6316    /* This function releases all the allocated frames. The base frame is on the
6317    machine stack, and so must not be freed.
6318    
6319    Argument: the address of the base frame
6320    Returns:  nothing
6321    */
6322    
6323    static void
6324    release_match_heapframes (heapframe *frame_base)
6325    {
6326    heapframe *nextframe = frame_base->Xnextframe;
6327    while (nextframe != NULL)
6328      {
6329      heapframe *oldframe = nextframe;
6330      nextframe = nextframe->Xnextframe;
6331      (PUBL(stack_free))(oldframe);
6332      }
6333    }
6334    #endif
6335    
6336    
6337  /*************************************************  /*************************************************
6338  *         Execute a Regular Expression           *  *         Execute a Regular Expression           *
# Line 5923  Returns:          > 0 => success; value Line 6358  Returns:          > 0 => success; value
6358                   < -1 => some kind of unexpected problem                   < -1 => some kind of unexpected problem
6359  */  */
6360    
6361    #ifdef COMPILE_PCRE8
6362  PCRE_EXP_DEFN int PCRE_CALL_CONVENTION  PCRE_EXP_DEFN int PCRE_CALL_CONVENTION
6363  pcre_exec(const pcre *argument_re, const pcre_extra *extra_data,  pcre_exec(const pcre *argument_re, const pcre_extra *extra_data,
6364    PCRE_SPTR subject, int length, int start_offset, int options, int *offsets,    PCRE_SPTR subject, int length, int start_offset, int options, int *offsets,
6365    int offsetcount)    int offsetcount)
6366    #else
6367    PCRE_EXP_DEFN int PCRE_CALL_CONVENTION
6368    pcre16_exec(const pcre16 *argument_re, const pcre16_extra *extra_data,
6369      PCRE_SPTR16 subject, int length, int start_offset, int options, int *offsets,
6370      int offsetcount)
6371    #endif
6372  {  {
6373  int rc, ocount, arg_offset_max;  int rc, ocount, arg_offset_max;
 int first_byte = -1;  
 int req_byte = -1;  
 int req_byte2 = -1;  
6374  int newline;  int newline;
6375  BOOL using_temporary_offsets = FALSE;  BOOL using_temporary_offsets = FALSE;
6376  BOOL anchored;  BOOL anchored;
6377  BOOL startline;  BOOL startline;
6378  BOOL firstline;  BOOL firstline;
6379  BOOL first_byte_caseless = FALSE;  BOOL utf;
6380  BOOL req_byte_caseless = FALSE;  BOOL has_first_char = FALSE;
6381  BOOL utf8;  BOOL has_req_char = FALSE;
6382    pcre_uchar first_char = 0;
6383    pcre_uchar first_char2 = 0;
6384    pcre_uchar req_char = 0;
6385    pcre_uchar req_char2 = 0;
6386  match_data match_block;  match_data match_block;
6387  match_data *md = &match_block;  match_data *md = &match_block;
6388  const uschar *tables;  const pcre_uint8 *tables;
6389  const uschar *start_bits = NULL;  const pcre_uint8 *start_bits = NULL;
6390  USPTR start_match = (USPTR)subject + start_offset;  PCRE_PUCHAR start_match = (PCRE_PUCHAR)subject + start_offset;
6391  USPTR end_subject;  PCRE_PUCHAR end_subject;
6392  USPTR start_partial = NULL;  PCRE_PUCHAR start_partial = NULL;
6393  USPTR req_byte_ptr = start_match - 1;  PCRE_PUCHAR req_char_ptr = start_match - 1;
6394    
 pcre_study_data internal_study;  
6395  const pcre_study_data *study;  const pcre_study_data *study;
6396    const REAL_PCRE *re = (const REAL_PCRE *)argument_re;
6397    
6398  real_pcre internal_re;  #ifdef NO_RECURSE
6399  const real_pcre *external_re = (const real_pcre *)argument_re;  heapframe frame_zero;
6400  const real_pcre *re = external_re;  frame_zero.Xprevframe = NULL;            /* Marks the top level */
6401    frame_zero.Xnextframe = NULL;            /* None are allocated yet */
6402    md->match_frames_base = &frame_zero;
6403    #endif
6404    
6405    /* Check for the special magic call that measures the size of the stack used
6406    per recursive call of match(). Without the funny casting for sizeof, a Windows
6407    compiler gave this error: "unary minus operator applied to unsigned type,
6408    result still unsigned". Hopefully the cast fixes that. */
6409    
6410    if (re == NULL && extra_data == NULL && subject == NULL && length == -999 &&
6411        start_offset == -999)
6412    #ifdef NO_RECURSE
6413      return -((int)sizeof(heapframe));
6414    #else
6415      return match(NULL, NULL, NULL, 0, NULL, NULL, 0);
6416    #endif
6417    
6418  /* Plausibility checks */  /* Plausibility checks */
6419    
6420  if ((options & ~PUBLIC_EXEC_OPTIONS) != 0) return PCRE_ERROR_BADOPTION;  if ((options & ~PUBLIC_EXEC_OPTIONS) != 0) return PCRE_ERROR_BADOPTION;
6421  if (re == NULL || subject == NULL ||  if (re == NULL || subject == NULL || (offsets == NULL && offsetcount > 0))
6422     (offsets == NULL && offsetcount > 0)) return PCRE_ERROR_NULL;    return PCRE_ERROR_NULL;
6423  if (offsetcount < 0) return PCRE_ERROR_BADCOUNT;  if (offsetcount < 0) return PCRE_ERROR_BADCOUNT;
6424  if (start_offset < 0 || start_offset > length) return PCRE_ERROR_BADOFFSET;  if (start_offset < 0 || start_offset > length) return PCRE_ERROR_BADOFFSET;
6425    
6426    /* Check that the first field in the block is the magic number. If it is not,
6427    return with PCRE_ERROR_BADMAGIC. However, if the magic number is equal to
6428    REVERSED_MAGIC_NUMBER we return with PCRE_ERROR_BADENDIANNESS, which
6429    means that the pattern is likely compiled with different endianness. */
6430    
6431    if (re->magic_number != MAGIC_NUMBER)
6432      return re->magic_number == REVERSED_MAGIC_NUMBER?
6433        PCRE_ERROR_BADENDIANNESS:PCRE_ERROR_BADMAGIC;
6434    if ((re->flags & PCRE_MODE) == 0) return PCRE_ERROR_BADMODE;
6435    
6436  /* 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
6437  follows immediately afterwards. Other values in the md block are used only  follows immediately afterwards. Other values in the md block are used only
6438  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,
6439  so they are set up later. */  so they are set up later. */
6440    
6441  utf8 = md->utf8 = (re->options & PCRE_UTF8) != 0;  /* PCRE_UTF16 has the same value as PCRE_UTF8. */
6442    utf = md->utf = (re->options & PCRE_UTF8) != 0;
6443  md->partial = ((options & PCRE_PARTIAL_HARD) != 0)? 2 :  md->partial = ((options & PCRE_PARTIAL_HARD) != 0)? 2 :
6444                ((options & PCRE_PARTIAL_SOFT) != 0)? 1 : 0;                ((options & PCRE_PARTIAL_SOFT) != 0)? 1 : 0;
6445    
6446  /* 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
6447  code for an invalid string if a results vector is available. */  code for an invalid string if a results vector is available. */
6448    
6449  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
6450  if (utf8 && (options & PCRE_NO_UTF8_CHECK) == 0)  if (utf && (options & PCRE_NO_UTF8_CHECK) == 0)
6451    {    {
6452    int erroroffset;    int erroroffset;
6453    int errorcode = _pcre_valid_utf8((USPTR)subject, length, &erroroffset);    int errorcode = PRIV(valid_utf)((PCRE_PUCHAR)subject, length, &erroroffset);
6454    if (errorcode != 0)    if (errorcode != 0)
6455      {      {
6456      if (offsetcount >= 2)      if (offsetcount >= 2)
# Line 5988  if (utf8 && (options & PCRE_NO_UTF8_CHEC Line 6458  if (utf8 && (options & PCRE_NO_UTF8_CHEC
6458        offsets[0] = erroroffset;        offsets[0] = erroroffset;
6459        offsets[1] = errorcode;        offsets[1] = errorcode;
6460        }        }
6461    #ifdef COMPILE_PCRE16
6462        return (errorcode <= PCRE_UTF16_ERR1 && md->partial > 1)?
6463          PCRE_ERROR_SHORTUTF16 : PCRE_ERROR_BADUTF16;
6464    #else
6465      return (errorcode <= PCRE_UTF8_ERR5 && md->partial > 1)?      return (errorcode <= PCRE_UTF8_ERR5 && md->partial > 1)?
6466        PCRE_ERROR_SHORTUTF8 : PCRE_ERROR_BADUTF8;        PCRE_ERROR_SHORTUTF8 : PCRE_ERROR_BADUTF8;
6467    #endif
6468      }      }
6469    
6470    /* 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. */
6471    if (start_offset > 0 && start_offset < length &&    if (start_offset > 0 && start_offset < length &&
6472        (((USPTR)subject)[start_offset] & 0xc0) == 0x80)        NOT_FIRSTCHAR(((PCRE_PUCHAR)subject)[start_offset]))
6473      return PCRE_ERROR_BADUTF8_OFFSET;      return PCRE_ERROR_BADUTF8_OFFSET;
6474    }    }
6475  #endif  #endif
# Line 6002  if (utf8 && (options & PCRE_NO_UTF8_CHEC Line 6477  if (utf8 && (options & PCRE_NO_UTF8_CHEC
6477  /* If the pattern was successfully studied with JIT support, run the JIT  /* If the pattern was successfully studied with JIT support, run the JIT
6478  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
6479  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
6480  an unsupported flag is set. In particular, JIT does not support partial  an unsupported flag is set. */
 matching. */  
6481    
6482  #ifdef SUPPORT_JIT  #ifdef SUPPORT_JIT
6483  if (extra_data != NULL  if (extra_data != NULL
6484      && (extra_data->flags & PCRE_EXTRA_EXECUTABLE_JIT) != 0      && (extra_data->flags & (PCRE_EXTRA_EXECUTABLE_JIT |
6485                                 PCRE_EXTRA_TABLES)) == PCRE_EXTRA_EXECUTABLE_JIT
6486      && extra_data->executable_jit != NULL      && extra_data->executable_jit != NULL
     && (extra_data->flags & PCRE_EXTRA_TABLES) == 0  
6487      && (options & ~(PCRE_NO_UTF8_CHECK | PCRE_NOTBOL | PCRE_NOTEOL |      && (options & ~(PCRE_NO_UTF8_CHECK | PCRE_NOTBOL | PCRE_NOTEOL |
6488                      PCRE_NOTEMPTY | PCRE_NOTEMPTY_ATSTART)) == 0)                      PCRE_NOTEMPTY | PCRE_NOTEMPTY_ATSTART |
6489    return _pcre_jit_exec(re, extra_data->executable_jit, subject, length,                      PCRE_PARTIAL_SOFT | PCRE_PARTIAL_HARD)) == 0)
6490      start_offset, options, ((extra_data->flags & PCRE_EXTRA_MATCH_LIMIT) == 0)    {
6491      ? MATCH_LIMIT : extra_data->match_limit, offsets, offsetcount);    rc = PRIV(jit_exec)(re, extra_data, (const pcre_uchar *)subject, length,
6492           start_offset, options, offsets, offsetcount);
6493    
6494      /* PCRE_ERROR_NULL means that the selected normal or partial matching
6495      mode is not compiled. In this case we simply fallback to interpreter. */
6496    
6497      if (rc != PCRE_ERROR_NULL) return rc;
6498      }
6499  #endif  #endif
6500    
6501  /* 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
6502  numbers associated with a given name, for condition testing. */  numbers associated with a given name, for condition testing. */
6503    
6504  md->name_table = (uschar *)re + re->name_table_offset;  md->name_table = (pcre_uchar *)re + re->name_table_offset;
6505  md->name_count = re->name_count;  md->name_count = re->name_count;
6506  md->name_entry_size = re->name_entry_size;  md->name_entry_size = re->name_entry_size;
6507    
# Line 6034  md->callout_data = NULL; Line 6515  md->callout_data = NULL;
6515    
6516  /* The table pointer is always in native byte order. */  /* The table pointer is always in native byte order. */
6517    
6518  tables = external_re->tables;  tables = re->tables;
6519    
6520  if (extra_data != NULL)  if (extra_data != NULL)
6521    {    {
# Line 6054  if (extra_data != NULL) Line 6535  if (extra_data != NULL)
6535  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
6536  in other programs later. */  in other programs later. */
6537    
6538  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;  
   }  
6539    
6540  /* Set up other data */  /* Set up other data */
6541    
# Line 6076  firstline = (re->options & PCRE_FIRSTLIN Line 6545  firstline = (re->options & PCRE_FIRSTLIN
6545    
6546  /* 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. */
6547    
6548  md->start_code = (const uschar *)external_re + re->name_table_offset +  md->start_code = (const pcre_uchar *)re + re->name_table_offset +
6549    re->name_count * re->name_entry_size;    re->name_count * re->name_entry_size;
6550    
6551  md->start_subject = (USPTR)subject;  md->start_subject = (PCRE_PUCHAR)subject;
6552  md->start_offset = start_offset;  md->start_offset = start_offset;
6553  md->end_subject = md->start_subject + length;  md->end_subject = md->start_subject + length;
6554  end_subject = md->end_subject;  end_subject = md->end_subject;
# Line 6104  md->recursive = NULL; Line 6573  md->recursive = NULL;
6573  md->hasthen = (re->flags & PCRE_HASTHEN) != 0;  md->hasthen = (re->flags & PCRE_HASTHEN) != 0;
6574    
6575  md->lcc = tables + lcc_offset;  md->lcc = tables + lcc_offset;
6576    md->fcc = tables + fcc_offset;
6577  md->ctypes = tables + ctypes_offset;  md->ctypes = tables + ctypes_offset;
6578    
6579  /* Handle different \R options. */  /* Handle different \R options. */
# Line 6190  arg_offset_max = (2*ocount)/3; Line 6660  arg_offset_max = (2*ocount)/3;
6660  if (re->top_backref > 0 && re->top_backref >= ocount/3)  if (re->top_backref > 0 && re->top_backref >= ocount/3)
6661    {    {
6662    ocount = re->top_backref * 3 + 3;    ocount = re->top_backref * 3 + 3;
6663    md->offset_vector = (int *)(pcre_malloc)(ocount * sizeof(int));    md->offset_vector = (int *)(PUBL(malloc))(ocount * sizeof(int));
6664    if (md->offset_vector == NULL) return PCRE_ERROR_NOMEMORY;    if (md->offset_vector == NULL) return PCRE_ERROR_NOMEMORY;
6665    using_temporary_offsets = TRUE;    using_temporary_offsets = TRUE;
6666    DPRINTF(("Got memory to hold back references\n"));    DPRINTF(("Got memory to hold back references\n"));
# Line 6217  if (md->offset_vector != NULL) Line 6687  if (md->offset_vector != NULL)
6687    md->offset_vector[0] = md->offset_vector[1] = -1;    md->offset_vector[0] = md->offset_vector[1] = -1;
6688    }    }
6689    
6690  /* 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
6691  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
6692  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
6693  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 6697  if (!anchored)
6697    {    {
6698    if ((re->flags & PCRE_FIRSTSET) != 0)    if ((re->flags & PCRE_FIRSTSET) != 0)
6699      {      {
6700      first_byte = re->first_byte & 255;      has_first_char = TRUE;
6701      if ((first_byte_caseless = ((re->first_byte & REQ_CASELESS) != 0)) == TRUE)      first_char = first_char2 = (pcre_uchar)(re->first_char);
6702        first_byte = md->lcc[first_byte];      if ((re->flags & PCRE_FCH_CASELESS) != 0)
6703          {
6704          first_char2 = TABLE_GET(first_char, md->fcc, first_char);
6705    #if defined SUPPORT_UCP && !(defined COMPILE_PCRE8)
6706          if (utf && first_char > 127)
6707            first_char2 = UCD_OTHERCASE(first_char);
6708    #endif
6709          }
6710      }      }
6711    else    else
6712      if (!startline && study != NULL &&      if (!startline && study != NULL &&
# Line 6242  character" set. */ Line 6719  character" set. */
6719    
6720  if ((re->flags & PCRE_REQCHSET) != 0)  if ((re->flags & PCRE_REQCHSET) != 0)
6721    {    {
6722    req_byte = re->req_byte & 255;    has_req_char = TRUE;
6723    req_byte_caseless = (re->req_byte & REQ_CASELESS) != 0;    req_char = req_char2 = (pcre_uchar)(re->req_char);
6724    req_byte2 = (tables + fcc_offset)[req_byte];  /* case flipped */    if ((re->flags & PCRE_RCH_CASELESS) != 0)
6725        {
6726        req_char2 = TABLE_GET(req_char, md->fcc, req_char);
6727    #if defined SUPPORT_UCP && !(defined COMPILE_PCRE8)
6728        if (utf && req_char > 127)
6729          req_char2 = UCD_OTHERCASE(req_char);
6730    #endif
6731        }
6732    }    }
6733    
6734    
   
   
6735  /* ==========================================================================*/  /* ==========================================================================*/
6736    
6737  /* 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 6739  the loop runs just once. */
6739    
6740  for(;;)  for(;;)
6741    {    {
6742    USPTR save_end_subject = end_subject;    PCRE_PUCHAR save_end_subject = end_subject;
6743    USPTR new_start_match;    PCRE_PUCHAR new_start_match;
6744    
6745    /* 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
6746    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 6750  for(;;)
6750    
6751    if (firstline)    if (firstline)
6752      {      {
6753      USPTR t = start_match;      PCRE_PUCHAR t = start_match;
6754  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
6755      if (utf8)      if (utf)
6756        {        {
6757        while (t < md->end_subject && !IS_NEWLINE(t))        while (t < md->end_subject && !IS_NEWLINE(t))
6758          {          {
6759          t++;          t++;
6760          while (t < end_subject && (*t & 0xc0) == 0x80) t++;          ACROSSCHAR(t < end_subject, *t, t++);
6761          }          }
6762        }        }
6763      else      else
# Line 6292  for(;;) Line 6774  for(;;)
6774    
6775    if (((options | re->options) & PCRE_NO_START_OPTIMIZE) == 0)    if (((options | re->options) & PCRE_NO_START_OPTIMIZE) == 0)
6776      {      {
6777      /* Advance to a unique first byte if there is one. */      /* Advance to a unique first char if there is one. */
6778    
6779      if (first_byte >= 0)      if (has_first_char)
6780        {        {
6781        if (first_byte_caseless)        if (first_char != first_char2)
6782          while (start_match < end_subject && md->lcc[*start_match] != first_byte)          while (start_match < end_subject &&
6783                *start_match != first_char && *start_match != first_char2)
6784            start_match++;            start_match++;
6785        else        else
6786          while (start_match < end_subject && *start_match != first_byte)          while (start_match < end_subject && *start_match != first_char)
6787            start_match++;            start_match++;
6788        }        }
6789    
# Line 6310  for(;;) Line 6793  for(;;)
6793        {        {
6794        if (start_match > md->start_subject + start_offset)        if (start_match > md->start_subject + start_offset)
6795          {          {
6796  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
6797          if (utf8)          if (utf)
6798            {            {
6799            while (start_match < end_subject && !WAS_NEWLINE(start_match))            while (start_match < end_subject && !WAS_NEWLINE(start_match))
6800              {              {
6801              start_match++;              start_match++;
6802              while(start_match < end_subject && (*start_match & 0xc0) == 0x80)              ACROSSCHAR(start_match < end_subject, *start_match,
6803                start_match++;                start_match++);
6804              }              }
6805            }            }
6806          else          else
# Line 6344  for(;;) Line 6827  for(;;)
6827        while (start_match < end_subject)        while (start_match < end_subject)
6828          {          {
6829          register unsigned int c = *start_match;          register unsigned int c = *start_match;
6830    #ifndef COMPILE_PCRE8
6831            if (c > 255) c = 255;
6832    #endif
6833          if ((start_bits[c/8] & (1 << (c&7))) == 0)          if ((start_bits[c/8] & (1 << (c&7))) == 0)
6834            {            {
6835            start_match++;            start_match++;
6836  #ifdef SUPPORT_UTF8  #if defined SUPPORT_UTF && defined COMPILE_PCRE8
6837            if (utf8)            /* In non 8-bit mode, the iteration will stop for
6838              while(start_match < end_subject && (*start_match & 0xc0) == 0x80)            characters > 255 at the beginning or not stop at all. */
6839                start_match++;            if (utf)
6840                ACROSSCHAR(start_match < end_subject, *start_match,
6841                  start_match++);
6842  #endif  #endif
6843            }            }
6844          else break;          else break;
# Line 6379  for(;;) Line 6867  for(;;)
6867        break;        break;
6868        }        }
6869    
6870      /* 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
6871      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
6872      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.
6873      This optimization can save a huge amount of backtracking in patterns with      This optimization can save a huge amount of backtracking in patterns with
6874      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 6881  for(;;)
6881      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
6882      long. */      long. */
6883    
6884      if (req_byte >= 0 && end_subject - start_match < REQ_BYTE_MAX)      if (has_req_char && end_subject - start_match < REQ_BYTE_MAX)
6885        {        {
6886        register USPTR p = start_match + ((first_byte >= 0)? 1 : 0);        register PCRE_PUCHAR p = start_match + (has_first_char? 1:0);
6887    
6888        /* 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
6889        place we found it at last time. */        place we found it at last time. */
6890    
6891        if (p > req_byte_ptr)        if (p > req_char_ptr)
6892          {          {
6893          if (req_byte_caseless)          if (req_char != req_char2)
6894            {            {
6895            while (p < end_subject)            while (p < end_subject)
6896              {              {
6897              register int pp = *p++;              register int pp = *p++;
6898              if (pp == req_byte || pp == req_byte2) { p--; break; }              if (pp == req_char || pp == req_char2) { p--; break; }
6899              }              }
6900            }            }
6901          else          else
6902            {            {
6903            while (p < end_subject)            while (p < end_subject)
6904              {              {
6905              if (*p++ == req_byte) { p--; break; }              if (*p++ == req_char) { p--; break; }
6906              }              }
6907            }            }
6908    
# Line 6431  for(;;) Line 6919  for(;;)
6919          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
6920          the start hasn't passed this character yet. */          the start hasn't passed this character yet. */
6921    
6922          req_byte_ptr = p;          req_char_ptr = p;
6923          }          }
6924        }        }
6925      }      }
# Line 6486  for(;;) Line 6974  for(;;)
6974      case MATCH_THEN:      case MATCH_THEN:
6975      md->ignore_skip_arg = FALSE;      md->ignore_skip_arg = FALSE;
6976      new_start_match = start_match + 1;      new_start_match = start_match + 1;
6977  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
6978      if (utf8)      if (utf)
6979        while(new_start_match < end_subject && (*new_start_match & 0xc0) == 0x80)        ACROSSCHAR(new_start_match < end_subject, *new_start_match,
6980          new_start_match++;          new_start_match++);
6981  #endif  #endif
6982      break;      break;
6983    
# Line 6527  for(;;) Line 7015  for(;;)
7015    
7016    /* 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
7017    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
7018    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
7019      normal matching start_match will aways be greater than the first position at
7020      this stage, but a failed *SKIP can cause a return at the same point, which is
7021      why the first test exists. */
7022    
7023    if (start_match[-1] == CHAR_CR &&    if (start_match > (PCRE_PUCHAR)subject + start_offset &&
7024          start_match[-1] == CHAR_CR &&
7025        start_match < end_subject &&        start_match < end_subject &&
7026        *start_match == CHAR_NL &&        *start_match == CHAR_NL &&
7027        (re->flags & PCRE_HASCRORLF) == 0 &&        (re->flags & PCRE_HASCRORLF) == 0 &&
# Line 6575  if (rc == MATCH_MATCH || rc == MATCH_ACC Line 7067  if (rc == MATCH_MATCH || rc == MATCH_ACC
7067        }        }
7068      if (md->end_offset_top > arg_offset_max) md->offset_overflow = TRUE;      if (md->end_offset_top > arg_offset_max) md->offset_overflow = TRUE;
7069      DPRINTF(("Freeing temporary memory\n"));      DPRINTF(("Freeing temporary memory\n"));
7070      (pcre_free)(md->offset_vector);      (PUBL(free))(md->offset_vector);
7071      }      }
7072    
7073    /* Set the return code to the number of captured strings, or 0 if there were    /* Set the return code to the number of captured strings, or 0 if there were
# Line 6597  if (rc == MATCH_MATCH || rc == MATCH_ACC Line 7089  if (rc == MATCH_MATCH || rc == MATCH_ACC
7089      {      {
7090      register int *iptr, *iend;      register int *iptr, *iend;
7091      int resetcount = 2 + re->top_bracket * 2;      int resetcount = 2 + re->top_bracket * 2;
7092      if (resetcount > offsetcount) resetcount = ocount;      if (resetcount > offsetcount) resetcount = offsetcount;
7093      iptr = offsets + md->end_offset_top;      iptr = offsets + md->end_offset_top;
7094      iend = offsets + resetcount;      iend = offsets + resetcount;
7095      while (iptr < iend) *iptr++ = -1;      while (iptr < iend) *iptr++ = -1;
# Line 6616  if (rc == MATCH_MATCH || rc == MATCH_ACC Line 7108  if (rc == MATCH_MATCH || rc == MATCH_ACC
7108    /* Return MARK data if requested */    /* Return MARK data if requested */
7109    
7110    if (extra_data != NULL && (extra_data->flags & PCRE_EXTRA_MARK) != 0)    if (extra_data != NULL && (extra_data->flags & PCRE_EXTRA_MARK) != 0)
7111      *(extra_data->mark) = (unsigned char *)(md->mark);      *(extra_data->mark) = (pcre_uchar *)md->mark;
7112    DPRINTF((">>>> returning %d\n", rc));    DPRINTF((">>>> returning %d\n", rc));
7113    #ifdef NO_RECURSE
7114      release_match_heapframes(&frame_zero);
7115    #endif
7116    return rc;    return rc;
7117    }    }
7118    
# Line 6627  attempt has failed at all permitted star Line 7122  attempt has failed at all permitted star
7122  if (using_temporary_offsets)  if (using_temporary_offsets)
7123    {    {
7124    DPRINTF(("Freeing temporary memory\n"));    DPRINTF(("Freeing temporary memory\n"));
7125    (pcre_free)(md->offset_vector);    (PUBL(free))(md->offset_vector);
7126    }    }
7127    
7128  /* 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 7130  if (using_temporary_offsets)
7130  if (rc != MATCH_NOMATCH && rc != PCRE_ERROR_PARTIAL)  if (rc != MATCH_NOMATCH && rc != PCRE_ERROR_PARTIAL)
7131    {    {
7132    DPRINTF((">>>> error: returning %d\n", rc));    DPRINTF((">>>> error: returning %d\n", rc));
7133    #ifdef NO_RECURSE
7134      release_match_heapframes(&frame_zero);
7135    #endif
7136    return rc;    return rc;
7137    }    }
7138    
# Line 6646  if (start_partial != NULL) Line 7144  if (start_partial != NULL)
7144    md->mark = NULL;    md->mark = NULL;
7145    if (offsetcount > 1)    if (offsetcount > 1)
7146      {      {
7147      offsets[0] = (int)(start_partial - (USPTR)subject);      offsets[0] = (int)(start_partial - (PCRE_PUCHAR)subject);
7148      offsets[1] = (int)(end_subject - (USPTR)subject);      offsets[1] = (int)(end_subject - (PCRE_PUCHAR)subject);
7149      }      }
7150    rc = PCRE_ERROR_PARTIAL;    rc = PCRE_ERROR_PARTIAL;
7151    }    }
# Line 6663  else Line 7161  else
7161  /* Return the MARK data if it has been requested. */  /* Return the MARK data if it has been requested. */
7162    
7163  if (extra_data != NULL && (extra_data->flags & PCRE_EXTRA_MARK) != 0)  if (extra_data != NULL && (extra_data->flags & PCRE_EXTRA_MARK) != 0)
7164    *(extra_data->mark) = (unsigned char *)(md->nomatch_mark);    *(extra_data->mark) = (pcre_uchar *)md->nomatch_mark;
7165    #ifdef NO_RECURSE
7166      release_match_heapframes(&frame_zero);
7167    #endif
7168  return rc;  return rc;
7169  }  }
7170    

Legend:
Removed from v.788  
changed lines
  Added in v.1011

  ViewVC Help
Powered by ViewVC 1.1.5