/[pcre]/code/branches/pcre16/pcre_exec.c
ViewVC logotype

Diff of /code/branches/pcre16/pcre_exec.c

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

code/trunk/pcre_exec.c revision 723 by ph10, Sat Oct 8 15:55:23 2011 UTC code/branches/pcre16/pcre_exec.c revision 804 by zherczeg, Wed Dec 14 11:18:01 2011 UTC
# Line 82  negative to avoid the external error cod Line 82  negative to avoid the external error cod
82  #define MATCH_SKIP_ARG     (-993)  #define MATCH_SKIP_ARG     (-993)
83  #define MATCH_THEN         (-992)  #define MATCH_THEN         (-992)
84    
 /* This is a convenience macro for code that occurs many times. */  
   
 #define MRRETURN(ra) \  
   { \  
   md->mark = markptr; \  
   RRETURN(ra); \  
   }  
   
85  /* Maximum number of ints of offset to save on the stack for recursive calls.  /* Maximum number of ints of offset to save on the stack for recursive calls.
86  If the offset vector is bigger, malloc is used. This should be a multiple of 3,  If the offset vector is bigger, malloc is used. This should be a multiple of 3,
87  because the offset vector is always a multiple of 3 long. */  because the offset vector is always a multiple of 3 long. */
# Line 121  Returns:     nothing Line 113  Returns:     nothing
113  */  */
114    
115  static void  static void
116  pchars(const uschar *p, int length, BOOL is_subject, match_data *md)  pchars(const pcre_uchar *p, int length, BOOL is_subject, match_data *md)
117  {  {
118  unsigned int c;  unsigned int c;
119  if (is_subject && length > md->end_subject - p) length = md->end_subject - p;  if (is_subject && length > md->end_subject - p) length = md->end_subject - p;
# Line 152  Returns:      < 0 if not matched, otherw Line 144  Returns:      < 0 if not matched, otherw
144  */  */
145    
146  static int  static int
147  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,
148    BOOL caseless)    BOOL caseless)
149  {  {
150  USPTR eptr_start = eptr;  PCRE_PUCHAR eptr_start = eptr;
151  register USPTR p = md->start_subject + md->offset_vector[offset];  register PCRE_PUCHAR p = md->start_subject + md->offset_vector[offset];
152    
153  #ifdef PCRE_DEBUG  #ifdef PCRE_DEBUG
154  if (eptr >= md->end_subject)  if (eptr >= md->end_subject)
# Line 181  ASCII characters. */ Line 173  ASCII characters. */
173    
174  if (caseless)  if (caseless)
175    {    {
176  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
177  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
178    if (md->utf8)    if (md->utf)
179      {      {
180      /* 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
181      bytes matched may differ, because there are some characters whose upper and      bytes matched may differ, because there are some characters whose upper and
# Line 193  if (caseless) Line 185  if (caseless)
185      the latter. It is important, therefore, to check the length along the      the latter. It is important, therefore, to check the length along the
186      reference, not along the subject (earlier code did this wrong). */      reference, not along the subject (earlier code did this wrong). */
187    
188      USPTR endptr = p + length;      PCRE_PUCHAR endptr = p + length;
189      while (p < endptr)      while (p < endptr)
190        {        {
191        int c, d;        int c, d;
# Line 225  else Line 217  else
217    while (length-- > 0) if (*p++ != *eptr++) return -1;    while (length-- > 0) if (*p++ != *eptr++) return -1;
218    }    }
219    
220  return eptr - eptr_start;  return (int)(eptr - eptr_start);
221  }  }
222    
223    
# Line 290  actually used in this definition. */ Line 282  actually used in this definition. */
282  #define RMATCH(ra,rb,rc,rd,re,rw) \  #define RMATCH(ra,rb,rc,rd,re,rw) \
283    { \    { \
284    printf("match() called in line %d\n", __LINE__); \    printf("match() called in line %d\n", __LINE__); \
285    rrc = match(ra,rb,mstart,markptr,rc,rd,re,rdepth+1); \    rrc = match(ra,rb,mstart,rc,rd,re,rdepth+1); \
286    printf("to line %d\n", __LINE__); \    printf("to line %d\n", __LINE__); \
287    }    }
288  #define RRETURN(ra) \  #define RRETURN(ra) \
# Line 300  actually used in this definition. */ Line 292  actually used in this definition. */
292    }    }
293  #else  #else
294  #define RMATCH(ra,rb,rc,rd,re,rw) \  #define RMATCH(ra,rb,rc,rd,re,rw) \
295    rrc = match(ra,rb,mstart,markptr,rc,rd,re,rdepth+1)    rrc = match(ra,rb,mstart,rc,rd,re,rdepth+1)
296  #define RRETURN(ra) return ra  #define RRETURN(ra) return ra
297  #endif  #endif
298    
# Line 315  argument of match(), which never changes Line 307  argument of match(), which never changes
307    
308  #define RMATCH(ra,rb,rc,rd,re,rw)\  #define RMATCH(ra,rb,rc,rd,re,rw)\
309    {\    {\
310    heapframe *newframe = (heapframe *)(pcre_stack_malloc)(sizeof(heapframe));\    heapframe *newframe = (heapframe *)(PUBL(stack_malloc))(sizeof(heapframe));\
311    if (newframe == NULL) RRETURN(PCRE_ERROR_NOMEMORY);\    if (newframe == NULL) RRETURN(PCRE_ERROR_NOMEMORY);\
312    frame->Xwhere = rw; \    frame->Xwhere = rw; \
313    newframe->Xeptr = ra;\    newframe->Xeptr = ra;\
314    newframe->Xecode = rb;\    newframe->Xecode = rb;\
315    newframe->Xmstart = mstart;\    newframe->Xmstart = mstart;\
   newframe->Xmarkptr = markptr;\  
316    newframe->Xoffset_top = rc;\    newframe->Xoffset_top = rc;\
317    newframe->Xeptrb = re;\    newframe->Xeptrb = re;\
318    newframe->Xrdepth = frame->Xrdepth + 1;\    newframe->Xrdepth = frame->Xrdepth + 1;\
# Line 337  argument of match(), which never changes Line 328  argument of match(), which never changes
328    {\    {\
329    heapframe *oldframe = frame;\    heapframe *oldframe = frame;\
330    frame = oldframe->Xprevframe;\    frame = oldframe->Xprevframe;\
331    (pcre_stack_free)(oldframe);\    (PUBL(stack_free))(oldframe);\
332    if (frame != NULL)\    if (frame != NULL)\
333      {\      {\
334      rrc = ra;\      rrc = ra;\
# Line 354  typedef struct heapframe { Line 345  typedef struct heapframe {
345    
346    /* Function arguments that may change */    /* Function arguments that may change */
347    
348    USPTR Xeptr;    PCRE_PUCHAR Xeptr;
349    const uschar *Xecode;    const pcre_uchar *Xecode;
350    USPTR Xmstart;    PCRE_PUCHAR Xmstart;
   USPTR Xmarkptr;  
351    int Xoffset_top;    int Xoffset_top;
352    eptrblock *Xeptrb;    eptrblock *Xeptrb;
353    unsigned int Xrdepth;    unsigned int Xrdepth;
354    
355    /* Function local variables */    /* Function local variables */
356    
357    USPTR Xcallpat;    PCRE_PUCHAR Xcallpat;
358  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
359    USPTR Xcharptr;    PCRE_PUCHAR Xcharptr;
360  #endif  #endif
361    USPTR Xdata;    PCRE_PUCHAR Xdata;
362    USPTR Xnext;    PCRE_PUCHAR Xnext;
363    USPTR Xpp;    PCRE_PUCHAR Xpp;
364    USPTR Xprev;    PCRE_PUCHAR Xprev;
365    USPTR Xsaved_eptr;    PCRE_PUCHAR Xsaved_eptr;
366    
367    recursion_info Xnew_recursive;    recursion_info Xnew_recursive;
368    
# Line 385  typedef struct heapframe { Line 375  typedef struct heapframe {
375    int Xprop_value;    int Xprop_value;
376    int Xprop_fail_result;    int Xprop_fail_result;
377    int Xoclength;    int Xoclength;
378    uschar Xocchars[8];    pcre_uchar Xocchars[6];
379  #endif  #endif
380    
381    int Xcodelink;    int Xcodelink;
# Line 427  returns a negative (error) response, the Line 417  returns a negative (error) response, the
417  same response. */  same response. */
418    
419  /* These macros pack up tests that are used for partial matching, and which  /* These macros pack up tests that are used for partial matching, and which
420  appears several times in the code. We set the "hit end" flag if the pointer is  appear several times in the code. We set the "hit end" flag if the pointer is
421  at the end of the subject and also past the start of the subject (i.e.  at the end of the subject and also past the start of the subject (i.e.
422  something has been matched). For hard partial matching, we then return  something has been matched). For hard partial matching, we then return
423  immediately. The second one is used when we already know we are past the end of  immediately. The second one is used when we already know we are past the end of
# Line 438  the subject. */ Line 428  the subject. */
428        eptr > md->start_used_ptr) \        eptr > md->start_used_ptr) \
429      { \      { \
430      md->hitend = TRUE; \      md->hitend = TRUE; \
431      if (md->partial > 1) MRRETURN(PCRE_ERROR_PARTIAL); \      if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL); \
432      }      }
433    
434  #define SCHECK_PARTIAL()\  #define SCHECK_PARTIAL()\
435    if (md->partial != 0 && eptr > md->start_used_ptr) \    if (md->partial != 0 && eptr > md->start_used_ptr) \
436      { \      { \
437      md->hitend = TRUE; \      md->hitend = TRUE; \
438      if (md->partial > 1) MRRETURN(PCRE_ERROR_PARTIAL); \      if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL); \
439      }      }
440    
441    
442  /* Performance note: It might be tempting to extract commonly used fields from  /* Performance note: It might be tempting to extract commonly used fields from
443  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
444  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
445  made performance worse.  made performance worse.
446    
# Line 459  Arguments: Line 449  Arguments:
449     ecode       pointer to current position in compiled code     ecode       pointer to current position in compiled code
450     mstart      pointer to the current match start position (can be modified     mstart      pointer to the current match start position (can be modified
451                   by encountering \K)                   by encountering \K)
    markptr     pointer to the most recent MARK name, or NULL  
452     offset_top  current top pointer     offset_top  current top pointer
453     md          pointer to "static" info for the match     md          pointer to "static" info for the match
454     eptrb       pointer to chain of blocks containing eptr at start of     eptrb       pointer to chain of blocks containing eptr at start of
# Line 474  Returns:       MATCH_MATCH if matched Line 463  Returns:       MATCH_MATCH if matched
463  */  */
464    
465  static int  static int
466  match(REGISTER USPTR eptr, REGISTER const uschar *ecode, USPTR mstart,  match(REGISTER PCRE_PUCHAR eptr, REGISTER const pcre_uchar *ecode,
467    const uschar *markptr, int offset_top, match_data *md, eptrblock *eptrb,    PCRE_PUCHAR mstart, int offset_top, match_data *md, eptrblock *eptrb,
468    unsigned int rdepth)    unsigned int rdepth)
469  {  {
470  /* These variables do not need to be preserved over recursion in this function,  /* These variables do not need to be preserved over recursion in this function,
# Line 485  so they can be ordinary variables in all Line 474  so they can be ordinary variables in all
474  register int  rrc;         /* Returns from recursive calls */  register int  rrc;         /* Returns from recursive calls */
475  register int  i;           /* Used for loops not involving calls to RMATCH() */  register int  i;           /* Used for loops not involving calls to RMATCH() */
476  register unsigned int c;   /* Character values not kept over RMATCH() calls */  register unsigned int c;   /* Character values not kept over RMATCH() calls */
477  register BOOL utf8;        /* Local copy of UTF-8 flag for speed */  register BOOL utf;         /* Local copy of UTF flag for speed */
478    
479  BOOL minimize, possessive; /* Quantifier options */  BOOL minimize, possessive; /* Quantifier options */
480  BOOL caseless;  BOOL caseless;
# Line 497  heap storage. Set up the top-level frame Line 486  heap storage. Set up the top-level frame
486  heap whenever RMATCH() does a "recursion". See the macro definitions above. */  heap whenever RMATCH() does a "recursion". See the macro definitions above. */
487    
488  #ifdef NO_RECURSE  #ifdef NO_RECURSE
489  heapframe *frame = (heapframe *)(pcre_stack_malloc)(sizeof(heapframe));  heapframe *frame = (heapframe *)(PUBL(stack_malloc))(sizeof(heapframe));
490  if (frame == NULL) RRETURN(PCRE_ERROR_NOMEMORY);  if (frame == NULL) RRETURN(PCRE_ERROR_NOMEMORY);
491  frame->Xprevframe = NULL;            /* Marks the top level */  frame->Xprevframe = NULL;            /* Marks the top level */
492    
# Line 506  frame->Xprevframe = NULL;            /* Line 495  frame->Xprevframe = NULL;            /*
495  frame->Xeptr = eptr;  frame->Xeptr = eptr;
496  frame->Xecode = ecode;  frame->Xecode = ecode;
497  frame->Xmstart = mstart;  frame->Xmstart = mstart;
 frame->Xmarkptr = markptr;  
498  frame->Xoffset_top = offset_top;  frame->Xoffset_top = offset_top;
499  frame->Xeptrb = eptrb;  frame->Xeptrb = eptrb;
500  frame->Xrdepth = rdepth;  frame->Xrdepth = rdepth;
# Line 520  HEAP_RECURSE: Line 508  HEAP_RECURSE:
508  #define eptr               frame->Xeptr  #define eptr               frame->Xeptr
509  #define ecode              frame->Xecode  #define ecode              frame->Xecode
510  #define mstart             frame->Xmstart  #define mstart             frame->Xmstart
 #define markptr            frame->Xmarkptr  
511  #define offset_top         frame->Xoffset_top  #define offset_top         frame->Xoffset_top
512  #define eptrb              frame->Xeptrb  #define eptrb              frame->Xeptrb
513  #define rdepth             frame->Xrdepth  #define rdepth             frame->Xrdepth
514    
515  /* Ditto for the local variables */  /* Ditto for the local variables */
516    
517  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
518  #define charptr            frame->Xcharptr  #define charptr            frame->Xcharptr
519  #endif  #endif
520  #define callpat            frame->Xcallpat  #define callpat            frame->Xcallpat
# Line 585  declarations can be cut out in a block. Line 572  declarations can be cut out in a block.
572  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
573  to RMATCH(). */  to RMATCH(). */
574    
575  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
576  const uschar *charptr;  const pcre_uchar *charptr;
577  #endif  #endif
578  const uschar *callpat;  const pcre_uchar *callpat;
579  const uschar *data;  const pcre_uchar *data;
580  const uschar *next;  const pcre_uchar *next;
581  USPTR         pp;  PCRE_PUCHAR       pp;
582  const uschar *prev;  const pcre_uchar *prev;
583  USPTR         saved_eptr;  PCRE_PUCHAR       saved_eptr;
584    
585  recursion_info new_recursive;  recursion_info new_recursive;
586    
# Line 606  int prop_type; Line 593  int prop_type;
593  int prop_value;  int prop_value;
594  int prop_fail_result;  int prop_fail_result;
595  int oclength;  int oclength;
596  uschar occhars[8];  pcre_uchar occhars[6];
597  #endif  #endif
598    
599  int codelink;  int codelink;
# Line 634  the alternative names that are used. */ Line 621  the alternative names that are used. */
621  #define code_offset   codelink  #define code_offset   codelink
622  #define condassert    condition  #define condassert    condition
623  #define matched_once  prev_is_word  #define matched_once  prev_is_word
624    #define foc           number
625    
626  /* These statements are here to stop the compiler complaining about unitialized  /* These statements are here to stop the compiler complaining about unitialized
627  variables. */  variables. */
# Line 659  defined). However, RMATCH isn't like a f Line 647  defined). However, RMATCH isn't like a f
647  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,
648  however, impact performance when true recursion is being used. */  however, impact performance when true recursion is being used. */
649    
650  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
651  utf8 = md->utf8;       /* Local copy of the flag */  utf = md->utf;       /* Local copy of the flag */
652  #else  #else
653  utf8 = FALSE;  utf = FALSE;
654  #endif  #endif
655    
656  /* First check that we haven't called match() too many times, or that we  /* First check that we haven't called match() too many times, or that we
# Line 701  for (;;) Line 689  for (;;)
689    switch(op)    switch(op)
690      {      {
691      case OP_MARK:      case OP_MARK:
692      markptr = ecode + 2;      md->nomatch_mark = ecode + 2;
693      RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode] + ecode[1], offset_top, md,      md->mark = NULL;    /* In case previously set by assertion */
694        RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode] + ecode[1], offset_top, md,
695        eptrb, RM55);        eptrb, RM55);
696        if ((rrc == MATCH_MATCH || rrc == MATCH_ACCEPT) &&
697             md->mark == NULL) md->mark = ecode + 2;
698    
699      /* A return of MATCH_SKIP_ARG means that matching failed at SKIP with an      /* A return of MATCH_SKIP_ARG means that matching failed at SKIP with an
700      argument, and we must check whether that argument matches this MARK's      argument, and we must check whether that argument matches this MARK's
# Line 712  for (;;) Line 703  for (;;)
703      position and return MATCH_SKIP. Otherwise, pass back the return code      position and return MATCH_SKIP. Otherwise, pass back the return code
704      unaltered. */      unaltered. */
705    
706      if (rrc == MATCH_SKIP_ARG &&      else if (rrc == MATCH_SKIP_ARG &&
707          strcmp((char *)markptr, (char *)(md->start_match_ptr)) == 0)          STRCMP_UC_UC(ecode + 2, md->start_match_ptr) == 0)
708        {        {
709        md->start_match_ptr = eptr;        md->start_match_ptr = eptr;
710        RRETURN(MATCH_SKIP);        RRETURN(MATCH_SKIP);
711        }        }
   
     if (md->mark == NULL) md->mark = markptr;  
712      RRETURN(rrc);      RRETURN(rrc);
713    
714      case OP_FAIL:      case OP_FAIL:
715      MRRETURN(MATCH_NOMATCH);      RRETURN(MATCH_NOMATCH);
716    
717      /* COMMIT overrides PRUNE, SKIP, and THEN */      /* COMMIT overrides PRUNE, SKIP, and THEN */
718    
719      case OP_COMMIT:      case OP_COMMIT:
720      RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,      RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md,
721        eptrb, RM52);        eptrb, RM52);
722      if (rrc != MATCH_NOMATCH && rrc != MATCH_PRUNE &&      if (rrc != MATCH_NOMATCH && rrc != MATCH_PRUNE &&
723          rrc != MATCH_SKIP && rrc != MATCH_SKIP_ARG &&          rrc != MATCH_SKIP && rrc != MATCH_SKIP_ARG &&
724          rrc != MATCH_THEN)          rrc != MATCH_THEN)
725        RRETURN(rrc);        RRETURN(rrc);
726      MRRETURN(MATCH_COMMIT);      RRETURN(MATCH_COMMIT);
727    
728      /* PRUNE overrides THEN */      /* PRUNE overrides THEN */
729    
730      case OP_PRUNE:      case OP_PRUNE:
731      RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,      RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md,
732        eptrb, RM51);        eptrb, RM51);
733      if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN) RRETURN(rrc);      if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN) RRETURN(rrc);
734      MRRETURN(MATCH_PRUNE);      RRETURN(MATCH_PRUNE);
735    
736      case OP_PRUNE_ARG:      case OP_PRUNE_ARG:
737      RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode] + ecode[1], offset_top, md,      md->nomatch_mark = ecode + 2;
738        md->mark = NULL;    /* In case previously set by assertion */
739        RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode] + ecode[1], offset_top, md,
740        eptrb, RM56);        eptrb, RM56);
741        if ((rrc == MATCH_MATCH || rrc == MATCH_ACCEPT) &&
742             md->mark == NULL) md->mark = ecode + 2;
743      if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN) RRETURN(rrc);      if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN) RRETURN(rrc);
     md->mark = ecode + 2;  
744      RRETURN(MATCH_PRUNE);      RRETURN(MATCH_PRUNE);
745    
746      /* SKIP overrides PRUNE and THEN */      /* SKIP overrides PRUNE and THEN */
747    
748      case OP_SKIP:      case OP_SKIP:
749      RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,      RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md,
750        eptrb, RM53);        eptrb, RM53);
751      if (rrc != MATCH_NOMATCH && rrc != MATCH_PRUNE && rrc != MATCH_THEN)      if (rrc != MATCH_NOMATCH && rrc != MATCH_PRUNE && rrc != MATCH_THEN)
752        RRETURN(rrc);        RRETURN(rrc);
753      md->start_match_ptr = eptr;   /* Pass back current position */      md->start_match_ptr = eptr;   /* Pass back current position */
754      MRRETURN(MATCH_SKIP);      RRETURN(MATCH_SKIP);
755    
756        /* Note that, for Perl compatibility, SKIP with an argument does NOT set
757        nomatch_mark. There is a flag that disables this opcode when re-matching a
758        pattern that ended with a SKIP for which there was not a matching MARK. */
759    
760      case OP_SKIP_ARG:      case OP_SKIP_ARG:
761      RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode] + ecode[1], offset_top, md,      if (md->ignore_skip_arg)
762          {
763          ecode += PRIV(OP_lengths)[*ecode] + ecode[1];
764          break;
765          }
766        RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode] + ecode[1], offset_top, md,
767        eptrb, RM57);        eptrb, RM57);
768      if (rrc != MATCH_NOMATCH && rrc != MATCH_PRUNE && rrc != MATCH_THEN)      if (rrc != MATCH_NOMATCH && rrc != MATCH_PRUNE && rrc != MATCH_THEN)
769        RRETURN(rrc);        RRETURN(rrc);
770    
771      /* Pass back the current skip name by overloading md->start_match_ptr and      /* Pass back the current skip name by overloading md->start_match_ptr and
772      returning the special MATCH_SKIP_ARG return code. This will either be      returning the special MATCH_SKIP_ARG return code. This will either be
773      caught by a matching MARK, or get to the top, where it is treated the same      caught by a matching MARK, or get to the top, where it causes a rematch
774      as PRUNE. */      with the md->ignore_skip_arg flag set. */
775    
776      md->start_match_ptr = ecode + 2;      md->start_match_ptr = ecode + 2;
777      RRETURN(MATCH_SKIP_ARG);      RRETURN(MATCH_SKIP_ARG);
# Line 780  for (;;) Line 781  for (;;)
781      match pointer to do this. */      match pointer to do this. */
782    
783      case OP_THEN:      case OP_THEN:
784      RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,      RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md,
785        eptrb, RM54);        eptrb, RM54);
786      if (rrc != MATCH_NOMATCH) RRETURN(rrc);      if (rrc != MATCH_NOMATCH) RRETURN(rrc);
787      md->start_match_ptr = ecode;      md->start_match_ptr = ecode;
788      MRRETURN(MATCH_THEN);      RRETURN(MATCH_THEN);
789    
790      case OP_THEN_ARG:      case OP_THEN_ARG:
791      RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode] + ecode[1], offset_top,      md->nomatch_mark = ecode + 2;
792        md->mark = NULL;    /* In case previously set by assertion */
793        RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode] + ecode[1], offset_top,
794        md, eptrb, RM58);        md, eptrb, RM58);
795        if ((rrc == MATCH_MATCH || rrc == MATCH_ACCEPT) &&
796             md->mark == NULL) md->mark = ecode + 2;
797      if (rrc != MATCH_NOMATCH) RRETURN(rrc);      if (rrc != MATCH_NOMATCH) RRETURN(rrc);
798      md->start_match_ptr = ecode;      md->start_match_ptr = ecode;
     md->mark = ecode + 2;  
799      RRETURN(MATCH_THEN);      RRETURN(MATCH_THEN);
800    
801      /* Handle an atomic group that does not contain any capturing parentheses.      /* Handle an atomic group that does not contain any capturing parentheses.
802      This can be handled like an assertion. Prior to 8.13, all atomic groups      This can be handled like an assertion. Prior to 8.13, all atomic groups
803      were handled this way. In 8.13, the code was changed as below for ONCE, so      were handled this way. In 8.13, the code was changed as below for ONCE, so
804      that backups pass through the group and thereby reset captured values.      that backups pass through the group and thereby reset captured values.
805      However, this uses a lot more stack, so in 8.20, atomic groups that do not      However, this uses a lot more stack, so in 8.20, atomic groups that do not
806      contain any captures generate OP_ONCE_NC, which can be handled in the old,      contain any captures generate OP_ONCE_NC, which can be handled in the old,
807      less stack intensive way.      less stack intensive way.
808    
809      Check the alternative branches in turn - the matching won't pass the KET      Check the alternative branches in turn - the matching won't pass the KET
# Line 821  for (;;) Line 825  for (;;)
825        if (rrc == MATCH_THEN)        if (rrc == MATCH_THEN)
826          {          {
827          next = ecode + GET(ecode,1);          next = ecode + GET(ecode,1);
828          if (md->start_match_ptr < next &&          if (md->start_match_ptr < next &&
829              (*ecode == OP_ALT || *next == OP_ALT))              (*ecode == OP_ALT || *next == OP_ALT))
830            rrc = MATCH_NOMATCH;            rrc = MATCH_NOMATCH;
831          }          }
832    
833        if (rrc != MATCH_NOMATCH) RRETURN(rrc);        if (rrc != MATCH_NOMATCH) RRETURN(rrc);
834        ecode += GET(ecode,1);        ecode += GET(ecode,1);
835        }        }
# Line 867  for (;;) Line 871  for (;;)
871        }        }
872      else  /* OP_KETRMAX */      else  /* OP_KETRMAX */
873        {        {
874        md->match_function_type = MATCH_CBEGROUP;        md->match_function_type = MATCH_CBEGROUP;
875        RMATCH(eptr, prev, offset_top, md, eptrb, RM66);        RMATCH(eptr, prev, offset_top, md, eptrb, RM66);
876        if (rrc != MATCH_NOMATCH) RRETURN(rrc);        if (rrc != MATCH_NOMATCH) RRETURN(rrc);
877        ecode += 1 + LINK_SIZE;        ecode += 1 + LINK_SIZE;
# Line 915  for (;;) Line 919  for (;;)
919        for (;;)        for (;;)
920          {          {
921          if (op >= OP_SBRA) md->match_function_type = MATCH_CBEGROUP;          if (op >= OP_SBRA) md->match_function_type = MATCH_CBEGROUP;
922          RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,          RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md,
923            eptrb, RM1);            eptrb, RM1);
924          if (rrc == MATCH_ONCE) break;  /* Backing up through an atomic group */          if (rrc == MATCH_ONCE) break;  /* Backing up through an atomic group */
925    
926          /* If we backed up to a THEN, check whether it is within the current          /* If we backed up to a THEN, check whether it is within the current
927          branch by comparing the address of the THEN that is passed back with          branch by comparing the address of the THEN that is passed back with
928          the end of the branch. If it is within the current branch, and the          the end of the branch. If it is within the current branch, and the
929          branch is one of two or more alternatives (it either starts or ends          branch is one of two or more alternatives (it either starts or ends
930          with OP_ALT), we have reached the limit of THEN's action, so convert          with OP_ALT), we have reached the limit of THEN's action, so convert
931          the return code to NOMATCH, which will cause normal backtracking to          the return code to NOMATCH, which will cause normal backtracking to
932          happen from now on. Otherwise, THEN is passed back to an outer          happen from now on. Otherwise, THEN is passed back to an outer
933          alternative. This implements Perl's treatment of parenthesized groups,          alternative. This implements Perl's treatment of parenthesized groups,
934          where a group not containing | does not affect the current alternative,          where a group not containing | does not affect the current alternative,
935          that is, (X) is NOT the same as (X|(*F)). */          that is, (X) is NOT the same as (X|(*F)). */
936    
937          if (rrc == MATCH_THEN)          if (rrc == MATCH_THEN)
938            {            {
939            next = ecode + GET(ecode,1);            next = ecode + GET(ecode,1);
940            if (md->start_match_ptr < next &&            if (md->start_match_ptr < next &&
941                (*ecode == OP_ALT || *next == OP_ALT))                (*ecode == OP_ALT || *next == OP_ALT))
942              rrc = MATCH_NOMATCH;              rrc = MATCH_NOMATCH;
943            }            }
944    
945          /* Anything other than NOMATCH is passed back. */          /* Anything other than NOMATCH is passed back. */
946    
947          if (rrc != MATCH_NOMATCH) RRETURN(rrc);          if (rrc != MATCH_NOMATCH) RRETURN(rrc);
# Line 953  for (;;) Line 957  for (;;)
957    
958        /* At this point, rrc will be one of MATCH_ONCE or MATCH_NOMATCH. */        /* At this point, rrc will be one of MATCH_ONCE or MATCH_NOMATCH. */
959    
       if (md->mark == NULL) md->mark = markptr;  
960        RRETURN(rrc);        RRETURN(rrc);
961        }        }
962    
# Line 1003  for (;;) Line 1006  for (;;)
1006    
1007        else if (!md->hasthen && ecode[GET(ecode, 1)] != OP_ALT)        else if (!md->hasthen && ecode[GET(ecode, 1)] != OP_ALT)
1008          {          {
1009          ecode += _pcre_OP_lengths[*ecode];          ecode += PRIV(OP_lengths)[*ecode];
1010          goto TAIL_RECURSE;          goto TAIL_RECURSE;
1011          }          }
1012    
1013        /* In all other cases, we have to make another call to match(). */        /* In all other cases, we have to make another call to match(). */
1014    
1015        RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md, eptrb,        RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md, eptrb,
1016          RM2);          RM2);
1017    
1018        /* See comment in the code for capturing groups above about handling        /* See comment in the code for capturing groups above about handling
1019        THEN. */        THEN. */
1020    
1021        if (rrc == MATCH_THEN)        if (rrc == MATCH_THEN)
1022          {          {
1023          next = ecode + GET(ecode,1);          next = ecode + GET(ecode,1);
1024          if (md->start_match_ptr < next &&          if (md->start_match_ptr < next &&
1025              (*ecode == OP_ALT || *next == OP_ALT))              (*ecode == OP_ALT || *next == OP_ALT))
1026            rrc = MATCH_NOMATCH;            rrc = MATCH_NOMATCH;
1027          }          }
1028    
1029        if (rrc != MATCH_NOMATCH)        if (rrc != MATCH_NOMATCH)
1030          {          {
1031          if (rrc == MATCH_ONCE)          if (rrc == MATCH_ONCE)
1032            {            {
1033            const uschar *scode = ecode;            const pcre_uchar *scode = ecode;
1034            if (*scode != OP_ONCE)           /* If not at start, find it */            if (*scode != OP_ONCE)           /* If not at start, find it */
1035              {              {
1036              while (*scode == OP_ALT) scode += GET(scode, 1);              while (*scode == OP_ALT) scode += GET(scode, 1);
# Line 1040  for (;;) Line 1043  for (;;)
1043        ecode += GET(ecode, 1);        ecode += GET(ecode, 1);
1044        if (*ecode != OP_ALT) break;        if (*ecode != OP_ALT) break;
1045        }        }
1046    
     if (md->mark == NULL) md->mark = markptr;  
1047      RRETURN(MATCH_NOMATCH);      RRETURN(MATCH_NOMATCH);
1048    
1049      /* Handle possessive capturing brackets with an unlimited repeat. We come      /* Handle possessive capturing brackets with an unlimited repeat. We come
# Line 1070  for (;;) Line 1072  for (;;)
1072      if (offset < md->offset_max)      if (offset < md->offset_max)
1073        {        {
1074        matched_once = FALSE;        matched_once = FALSE;
1075        code_offset = ecode - md->start_code;        code_offset = (int)(ecode - md->start_code);
1076    
1077        save_offset1 = md->offset_vector[offset];        save_offset1 = md->offset_vector[offset];
1078        save_offset2 = md->offset_vector[offset+1];        save_offset2 = md->offset_vector[offset+1];
# Line 1093  for (;;) Line 1095  for (;;)
1095          md->offset_vector[md->offset_end - number] =          md->offset_vector[md->offset_end - number] =
1096            (int)(eptr - md->start_subject);            (int)(eptr - md->start_subject);
1097          if (op >= OP_SBRA) md->match_function_type = MATCH_CBEGROUP;          if (op >= OP_SBRA) md->match_function_type = MATCH_CBEGROUP;
1098          RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,          RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md,
1099            eptrb, RM63);            eptrb, RM63);
1100          if (rrc == MATCH_KETRPOS)          if (rrc == MATCH_KETRPOS)
1101            {            {
# Line 1104  for (;;) Line 1106  for (;;)
1106            matched_once = TRUE;            matched_once = TRUE;
1107            continue;            continue;
1108            }            }
1109    
1110          /* See comment in the code for capturing groups above about handling          /* See comment in the code for capturing groups above about handling
1111          THEN. */          THEN. */
1112    
1113          if (rrc == MATCH_THEN)          if (rrc == MATCH_THEN)
1114            {            {
1115            next = ecode + GET(ecode,1);            next = ecode + GET(ecode,1);
1116            if (md->start_match_ptr < next &&            if (md->start_match_ptr < next &&
1117                (*ecode == OP_ALT || *next == OP_ALT))                (*ecode == OP_ALT || *next == OP_ALT))
1118              rrc = MATCH_NOMATCH;              rrc = MATCH_NOMATCH;
1119            }            }
1120    
1121          if (rrc != MATCH_NOMATCH) RRETURN(rrc);          if (rrc != MATCH_NOMATCH) RRETURN(rrc);
1122          md->capture_last = save_capture_last;          md->capture_last = save_capture_last;
# Line 1129  for (;;) Line 1131  for (;;)
1131          md->offset_vector[md->offset_end - number] = save_offset3;          md->offset_vector[md->offset_end - number] = save_offset3;
1132          }          }
1133    
       if (md->mark == NULL) md->mark = markptr;  
1134        if (allow_zero || matched_once)        if (allow_zero || matched_once)
1135          {          {
1136          ecode += 1 + LINK_SIZE;          ecode += 1 + LINK_SIZE;
# Line 1161  for (;;) Line 1162  for (;;)
1162    
1163      POSSESSIVE_NON_CAPTURE:      POSSESSIVE_NON_CAPTURE:
1164      matched_once = FALSE;      matched_once = FALSE;
1165      code_offset = ecode - md->start_code;      code_offset = (int)(ecode - md->start_code);
1166    
1167      for (;;)      for (;;)
1168        {        {
1169        if (op >= OP_SBRA) md->match_function_type = MATCH_CBEGROUP;        if (op >= OP_SBRA) md->match_function_type = MATCH_CBEGROUP;
1170        RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,        RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md,
1171          eptrb, RM48);          eptrb, RM48);
1172        if (rrc == MATCH_KETRPOS)        if (rrc == MATCH_KETRPOS)
1173          {          {
# Line 1176  for (;;) Line 1177  for (;;)
1177          matched_once = TRUE;          matched_once = TRUE;
1178          continue;          continue;
1179          }          }
1180    
1181        /* See comment in the code for capturing groups above about handling        /* See comment in the code for capturing groups above about handling
1182        THEN. */        THEN. */
1183    
1184        if (rrc == MATCH_THEN)        if (rrc == MATCH_THEN)
1185          {          {
1186          next = ecode + GET(ecode,1);          next = ecode + GET(ecode,1);
1187          if (md->start_match_ptr < next &&          if (md->start_match_ptr < next &&
1188              (*ecode == OP_ALT || *next == OP_ALT))              (*ecode == OP_ALT || *next == OP_ALT))
1189            rrc = MATCH_NOMATCH;            rrc = MATCH_NOMATCH;
1190          }          }
1191    
1192        if (rrc != MATCH_NOMATCH) RRETURN(rrc);        if (rrc != MATCH_NOMATCH) RRETURN(rrc);
1193        ecode += GET(ecode, 1);        ecode += GET(ecode, 1);
# Line 1216  for (;;) Line 1217  for (;;)
1217    
1218      if (ecode[LINK_SIZE+1] == OP_CALLOUT)      if (ecode[LINK_SIZE+1] == OP_CALLOUT)
1219        {        {
1220        if (pcre_callout != NULL)        if (PUBL(callout) != NULL)
1221          {          {
1222          pcre_callout_block cb;          pcre_callout_block cb;
1223          cb.version          = 2;   /* Version 1 of the callout block */          cb.version          = 2;   /* Version 1 of the callout block */
# Line 1231  for (;;) Line 1232  for (;;)
1232          cb.capture_top      = offset_top/2;          cb.capture_top      = offset_top/2;
1233          cb.capture_last     = md->capture_last;          cb.capture_last     = md->capture_last;
1234          cb.callout_data     = md->callout_data;          cb.callout_data     = md->callout_data;
1235          cb.mark             = markptr;          cb.mark             = md->nomatch_mark;
1236          if ((rrc = (*pcre_callout)(&cb)) > 0) MRRETURN(MATCH_NOMATCH);          if ((rrc = (*PUBL(callout))(&cb)) > 0) RRETURN(MATCH_NOMATCH);
1237          if (rrc < 0) RRETURN(rrc);          if (rrc < 0) RRETURN(rrc);
1238          }          }
1239        ecode += _pcre_OP_lengths[OP_CALLOUT];        ecode += PRIV(OP_lengths)[OP_CALLOUT];
1240        }        }
1241    
1242      condcode = ecode[LINK_SIZE+1];      condcode = ecode[LINK_SIZE+1];
# Line 1252  for (;;) Line 1253  for (;;)
1253        else        else
1254          {          {
1255          int recno = GET2(ecode, LINK_SIZE + 2);   /* Recursion group number*/          int recno = GET2(ecode, LINK_SIZE + 2);   /* Recursion group number*/
1256          condition =  (recno == RREF_ANY || recno == md->recursive->group_num);          condition = (recno == RREF_ANY || recno == md->recursive->group_num);
1257    
1258          /* If the test is for recursion into a specific subpattern, and it is          /* If the test is for recursion into a specific subpattern, and it is
1259          false, but the test was set up by name, scan the table to see if the          false, but the test was set up by name, scan the table to see if the
1260          name refers to any other numbers, and test them. The condition is true          name refers to any other numbers, and test them. The condition is true
1261          if any one is set. */          if any one is set. */
1262    
1263          if (!condition && condcode == OP_NRREF && recno != RREF_ANY)          if (!condition && condcode == OP_NRREF)
1264            {            {
1265            uschar *slotA = md->name_table;            pcre_uchar *slotA = md->name_table;
1266            for (i = 0; i < md->name_count; i++)            for (i = 0; i < md->name_count; i++)
1267              {              {
1268              if (GET2(slotA, 0) == recno) break;              if (GET2(slotA, 0) == recno) break;
# Line 1274  for (;;) Line 1275  for (;;)
1275    
1276            if (i < md->name_count)            if (i < md->name_count)
1277              {              {
1278              uschar *slotB = slotA;              pcre_uchar *slotB = slotA;
1279              while (slotB > md->name_table)              while (slotB > md->name_table)
1280                {                {
1281                slotB -= md->name_entry_size;                slotB -= md->name_entry_size;
1282                if (strcmp((char *)slotA + 2, (char *)slotB + 2) == 0)                if (STRCMP_UC_UC(slotA + IMM2_SIZE, slotB + IMM2_SIZE) == 0)
1283                  {                  {
1284                  condition = GET2(slotB, 0) == md->recursive->group_num;                  condition = GET2(slotB, 0) == md->recursive->group_num;
1285                  if (condition) break;                  if (condition) break;
# Line 1294  for (;;) Line 1295  for (;;)
1295                for (i++; i < md->name_count; i++)                for (i++; i < md->name_count; i++)
1296                  {                  {
1297                  slotB += md->name_entry_size;                  slotB += md->name_entry_size;
1298                  if (strcmp((char *)slotA + 2, (char *)slotB + 2) == 0)                  if (STRCMP_UC_UC(slotA + IMM2_SIZE, slotB + IMM2_SIZE) == 0)
1299                    {                    {
1300                    condition = GET2(slotB, 0) == md->recursive->group_num;                    condition = GET2(slotB, 0) == md->recursive->group_num;
1301                    if (condition) break;                    if (condition) break;
# Line 1307  for (;;) Line 1308  for (;;)
1308    
1309          /* Chose branch according to the condition */          /* Chose branch according to the condition */
1310    
1311          ecode += condition? 3 : GET(ecode, 1);          ecode += condition? 1 + IMM2_SIZE : GET(ecode, 1);
1312          }          }
1313        }        }
1314    
# Line 1324  for (;;) Line 1325  for (;;)
1325        if (!condition && condcode == OP_NCREF)        if (!condition && condcode == OP_NCREF)
1326          {          {
1327          int refno = offset >> 1;          int refno = offset >> 1;
1328          uschar *slotA = md->name_table;          pcre_uchar *slotA = md->name_table;
1329    
1330          for (i = 0; i < md->name_count; i++)          for (i = 0; i < md->name_count; i++)
1331            {            {
# Line 1338  for (;;) Line 1339  for (;;)
1339    
1340          if (i < md->name_count)          if (i < md->name_count)
1341            {            {
1342            uschar *slotB = slotA;            pcre_uchar *slotB = slotA;
1343            while (slotB > md->name_table)            while (slotB > md->name_table)
1344              {              {
1345              slotB -= md->name_entry_size;              slotB -= md->name_entry_size;
1346              if (strcmp((char *)slotA + 2, (char *)slotB + 2) == 0)              if (STRCMP_UC_UC(slotA + IMM2_SIZE, slotB + IMM2_SIZE) == 0)
1347                {                {
1348                offset = GET2(slotB, 0) << 1;                offset = GET2(slotB, 0) << 1;
1349                condition = offset < offset_top &&                condition = offset < offset_top &&
# Line 1360  for (;;) Line 1361  for (;;)
1361              for (i++; i < md->name_count; i++)              for (i++; i < md->name_count; i++)
1362                {                {
1363                slotB += md->name_entry_size;                slotB += md->name_entry_size;
1364                if (strcmp((char *)slotA + 2, (char *)slotB + 2) == 0)                if (STRCMP_UC_UC(slotA + IMM2_SIZE, slotB + IMM2_SIZE) == 0)
1365                  {                  {
1366                  offset = GET2(slotB, 0) << 1;                  offset = GET2(slotB, 0) << 1;
1367                  condition = offset < offset_top &&                  condition = offset < offset_top &&
# Line 1375  for (;;) Line 1376  for (;;)
1376    
1377        /* Chose branch according to the condition */        /* Chose branch according to the condition */
1378    
1379        ecode += condition? 3 : GET(ecode, 1);        ecode += condition? 1 + IMM2_SIZE : GET(ecode, 1);
1380        }        }
1381    
1382      else if (condcode == OP_DEF)     /* DEFINE - always false */      else if (condcode == OP_DEF)     /* DEFINE - always false */
# Line 1400  for (;;) Line 1401  for (;;)
1401          ecode += 1 + LINK_SIZE + GET(ecode, LINK_SIZE + 2);          ecode += 1 + LINK_SIZE + GET(ecode, LINK_SIZE + 2);
1402          while (*ecode == OP_ALT) ecode += GET(ecode, 1);          while (*ecode == OP_ALT) ecode += GET(ecode, 1);
1403          }          }
1404    
1405        /* PCRE doesn't allow the effect of (*THEN) to escape beyond an        /* PCRE doesn't allow the effect of (*THEN) to escape beyond an
1406        assertion; it is therefore treated as NOMATCH. */        assertion; it is therefore treated as NOMATCH. */
1407    
1408        else if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN)        else if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN)
1409          {          {
1410          RRETURN(rrc);         /* Need braces because of following else */          RRETURN(rrc);         /* Need braces because of following else */
1411          }          }
# Line 1432  for (;;) Line 1433  for (;;)
1433          ecode += 1 + LINK_SIZE;          ecode += 1 + LINK_SIZE;
1434          goto TAIL_RECURSE;          goto TAIL_RECURSE;
1435          }          }
1436    
1437        md->match_function_type = MATCH_CBEGROUP;        md->match_function_type = MATCH_CBEGROUP;
1438        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, eptrb, RM49);        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, eptrb, RM49);
1439        RRETURN(rrc);        RRETURN(rrc);
# Line 1467  for (;;) Line 1468  for (;;)
1468        md->offset_vector[offset+1] = (int)(eptr - md->start_subject);        md->offset_vector[offset+1] = (int)(eptr - md->start_subject);
1469        if (offset_top <= offset) offset_top = offset + 2;        if (offset_top <= offset) offset_top = offset + 2;
1470        }        }
1471      ecode += 3;      ecode += 1 + IMM2_SIZE;
1472      break;      break;
1473    
1474    
# Line 1487  for (;;) Line 1488  for (;;)
1488           (md->notempty ||           (md->notempty ||
1489             (md->notempty_atstart &&             (md->notempty_atstart &&
1490               mstart == md->start_subject + md->start_offset)))               mstart == md->start_subject + md->start_offset)))
1491        MRRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
1492    
1493      /* Otherwise, we have a match. */      /* Otherwise, we have a match. */
1494    
# Line 1496  for (;;) Line 1497  for (;;)
1497      md->start_match_ptr = mstart;       /* and the start (\K can modify) */      md->start_match_ptr = mstart;       /* and the start (\K can modify) */
1498    
1499      /* For some reason, the macros don't work properly if an expression is      /* For some reason, the macros don't work properly if an expression is
1500      given as the argument to MRRETURN when the heap is in use. */      given as the argument to RRETURN when the heap is in use. */
1501    
1502      rrc = (op == OP_END)? MATCH_MATCH : MATCH_ACCEPT;      rrc = (op == OP_END)? MATCH_MATCH : MATCH_ACCEPT;
1503      MRRETURN(rrc);      RRETURN(rrc);
1504    
1505      /* Assertion brackets. Check the alternative branches in turn - the      /* Assertion brackets. Check the alternative branches in turn - the
1506      matching won't pass the KET for an assertion. If any one branch matches,      matching won't pass the KET for an assertion. If any one branch matches,
# Line 1527  for (;;) Line 1528  for (;;)
1528        if (rrc == MATCH_MATCH || rrc == MATCH_ACCEPT)        if (rrc == MATCH_MATCH || rrc == MATCH_ACCEPT)
1529          {          {
1530          mstart = md->start_match_ptr;   /* In case \K reset it */          mstart = md->start_match_ptr;   /* In case \K reset it */
         markptr = md->mark;  
1531          break;          break;
1532          }          }
1533    
1534        /* PCRE does not allow THEN to escape beyond an assertion; it is treated        /* PCRE does not allow THEN to escape beyond an assertion; it is treated
1535        as NOMATCH. */        as NOMATCH. */
1536    
1537        if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN) RRETURN(rrc);        if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN) RRETURN(rrc);
1538        ecode += GET(ecode, 1);        ecode += GET(ecode, 1);
1539        }        }
1540      while (*ecode == OP_ALT);      while (*ecode == OP_ALT);
1541    
1542      if (*ecode == OP_KET) MRRETURN(MATCH_NOMATCH);      if (*ecode == OP_KET) RRETURN(MATCH_NOMATCH);
1543    
1544      /* If checking an assertion for a condition, return MATCH_MATCH. */      /* If checking an assertion for a condition, return MATCH_MATCH. */
1545    
# Line 1569  for (;;) Line 1569  for (;;)
1569      do      do
1570        {        {
1571        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, NULL, RM5);        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, NULL, RM5);
1572        if (rrc == MATCH_MATCH || rrc == MATCH_ACCEPT) MRRETURN(MATCH_NOMATCH);        if (rrc == MATCH_MATCH || rrc == MATCH_ACCEPT) RRETURN(MATCH_NOMATCH);
1573        if (rrc == MATCH_SKIP || rrc == MATCH_PRUNE || rrc == MATCH_COMMIT)        if (rrc == MATCH_SKIP || rrc == MATCH_PRUNE || rrc == MATCH_COMMIT)
1574          {          {
1575          do ecode += GET(ecode,1); while (*ecode == OP_ALT);          do ecode += GET(ecode,1); while (*ecode == OP_ALT);
1576          break;          break;
1577          }          }
1578    
1579        /* PCRE does not allow THEN to escape beyond an assertion; it is treated        /* PCRE does not allow THEN to escape beyond an assertion; it is treated
1580        as NOMATCH. */        as NOMATCH. */
1581    
1582        if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN) RRETURN(rrc);        if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN) RRETURN(rrc);
# Line 1595  for (;;) Line 1595  for (;;)
1595      back a number of characters, not bytes. */      back a number of characters, not bytes. */
1596    
1597      case OP_REVERSE:      case OP_REVERSE:
1598  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
1599      if (utf8)      if (utf)
1600        {        {
1601        i = GET(ecode, 1);        i = GET(ecode, 1);
1602        while (i-- > 0)        while (i-- > 0)
1603          {          {
1604          eptr--;          eptr--;
1605          if (eptr < md->start_subject) MRRETURN(MATCH_NOMATCH);          if (eptr < md->start_subject) RRETURN(MATCH_NOMATCH);
1606          BACKCHAR(eptr);          BACKCHAR(eptr);
1607          }          }
1608        }        }
# Line 1613  for (;;) Line 1613  for (;;)
1613    
1614        {        {
1615        eptr -= GET(ecode, 1);        eptr -= GET(ecode, 1);
1616        if (eptr < md->start_subject) MRRETURN(MATCH_NOMATCH);        if (eptr < md->start_subject) RRETURN(MATCH_NOMATCH);
1617        }        }
1618    
1619      /* Save the earliest consulted character, then skip to next op code */      /* Save the earliest consulted character, then skip to next op code */
# Line 1627  for (;;) Line 1627  for (;;)
1627      function is able to force a failure. */      function is able to force a failure. */
1628    
1629      case OP_CALLOUT:      case OP_CALLOUT:
1630      if (pcre_callout != NULL)      if (PUBL(callout) != NULL)
1631        {        {
1632        pcre_callout_block cb;        pcre_callout_block cb;
1633        cb.version          = 2;   /* Version 1 of the callout block */        cb.version          = 2;   /* Version 1 of the callout block */
# Line 1642  for (;;) Line 1642  for (;;)
1642        cb.capture_top      = offset_top/2;        cb.capture_top      = offset_top/2;
1643        cb.capture_last     = md->capture_last;        cb.capture_last     = md->capture_last;
1644        cb.callout_data     = md->callout_data;        cb.callout_data     = md->callout_data;
1645        cb.mark             = markptr;        cb.mark             = md->nomatch_mark;
1646        if ((rrc = (*pcre_callout)(&cb)) > 0) MRRETURN(MATCH_NOMATCH);        if ((rrc = (*PUBL(callout))(&cb)) > 0) RRETURN(MATCH_NOMATCH);
1647        if (rrc < 0) RRETURN(rrc);        if (rrc < 0) RRETURN(rrc);
1648        }        }
1649      ecode += 2 + 2*LINK_SIZE;      ecode += 2 + 2*LINK_SIZE;
# Line 1702  for (;;) Line 1702  for (;;)
1702        else        else
1703          {          {
1704          new_recursive.offset_save =          new_recursive.offset_save =
1705            (int *)(pcre_malloc)(new_recursive.saved_max * sizeof(int));            (int *)(PUBL(malloc))(new_recursive.saved_max * sizeof(int));
1706          if (new_recursive.offset_save == NULL) RRETURN(PCRE_ERROR_NOMEMORY);          if (new_recursive.offset_save == NULL) RRETURN(PCRE_ERROR_NOMEMORY);
1707          }          }
1708        memcpy(new_recursive.offset_save, md->offset_vector,        memcpy(new_recursive.offset_save, md->offset_vector,
# Line 1717  for (;;) Line 1717  for (;;)
1717        do        do
1718          {          {
1719          if (cbegroup) md->match_function_type = MATCH_CBEGROUP;          if (cbegroup) md->match_function_type = MATCH_CBEGROUP;
1720          RMATCH(eptr, callpat + _pcre_OP_lengths[*callpat], offset_top,          RMATCH(eptr, callpat + PRIV(OP_lengths)[*callpat], offset_top,
1721            md, eptrb, RM6);            md, eptrb, RM6);
1722          memcpy(md->offset_vector, new_recursive.offset_save,          memcpy(md->offset_vector, new_recursive.offset_save,
1723              new_recursive.saved_max * sizeof(int));              new_recursive.saved_max * sizeof(int));
# Line 1726  for (;;) Line 1726  for (;;)
1726            {            {
1727            DPRINTF(("Recursion matched\n"));            DPRINTF(("Recursion matched\n"));
1728            if (new_recursive.offset_save != stacksave)            if (new_recursive.offset_save != stacksave)
1729              (pcre_free)(new_recursive.offset_save);              (PUBL(free))(new_recursive.offset_save);
1730    
1731            /* 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
1732            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 1740  for (;;) Line 1740  for (;;)
1740          /* PCRE does not allow THEN to escape beyond a recursion; it is treated          /* PCRE does not allow THEN to escape beyond a recursion; it is treated
1741          as NOMATCH. */          as NOMATCH. */
1742    
1743          else if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN)          else if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN)
1744            {            {
1745            DPRINTF(("Recursion gave error %d\n", rrc));            DPRINTF(("Recursion gave error %d\n", rrc));
1746            if (new_recursive.offset_save != stacksave)            if (new_recursive.offset_save != stacksave)
1747              (pcre_free)(new_recursive.offset_save);              (PUBL(free))(new_recursive.offset_save);
1748            RRETURN(rrc);            RRETURN(rrc);
1749            }            }
1750    
# Line 1756  for (;;) Line 1756  for (;;)
1756        DPRINTF(("Recursion didn't match\n"));        DPRINTF(("Recursion didn't match\n"));
1757        md->recursive = new_recursive.prevrec;        md->recursive = new_recursive.prevrec;
1758        if (new_recursive.offset_save != stacksave)        if (new_recursive.offset_save != stacksave)
1759          (pcre_free)(new_recursive.offset_save);          (PUBL(free))(new_recursive.offset_save);
1760        MRRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
1761        }        }
1762    
1763      RECURSION_MATCHED:      RECURSION_MATCHED:
# Line 1826  for (;;) Line 1826  for (;;)
1826        }        }
1827      else saved_eptr = NULL;      else saved_eptr = NULL;
1828    
1829      /* If we are at the end of an assertion group or a non-capturing atomic      /* If we are at the end of an assertion group or a non-capturing atomic
1830      group, stop matching and return MATCH_MATCH, but record the current high      group, stop matching and return MATCH_MATCH, but record the current high
1831      water mark for use by positive assertions. We also need to record the match      water mark for use by positive assertions. We also need to record the match
1832      start in case it was changed by \K. */      start in case it was changed by \K. */
1833    
1834      if ((*prev >= OP_ASSERT && *prev <= OP_ASSERTBACK_NOT) ||      if ((*prev >= OP_ASSERT && *prev <= OP_ASSERTBACK_NOT) ||
1835           *prev == OP_ONCE_NC)           *prev == OP_ONCE_NC)
1836        {        {
1837        md->end_match_ptr = eptr;      /* For ONCE_NC */        md->end_match_ptr = eptr;      /* For ONCE_NC */
1838        md->end_offset_top = offset_top;        md->end_offset_top = offset_top;
1839        md->start_match_ptr = mstart;        md->start_match_ptr = mstart;
1840        MRRETURN(MATCH_MATCH);         /* Sets md->mark */        RRETURN(MATCH_MATCH);         /* Sets md->mark */
1841        }        }
1842    
1843      /* For capturing groups we have to check the group number back at the start      /* For capturing groups we have to check the group number back at the start
# Line 1979  for (;;) Line 1979  for (;;)
1979      /* Not multiline mode: start of subject assertion, unless notbol. */      /* Not multiline mode: start of subject assertion, unless notbol. */
1980    
1981      case OP_CIRC:      case OP_CIRC:
1982      if (md->notbol && eptr == md->start_subject) MRRETURN(MATCH_NOMATCH);      if (md->notbol && eptr == md->start_subject) RRETURN(MATCH_NOMATCH);
1983    
1984      /* Start of subject assertion */      /* Start of subject assertion */
1985    
1986      case OP_SOD:      case OP_SOD:
1987      if (eptr != md->start_subject) MRRETURN(MATCH_NOMATCH);      if (eptr != md->start_subject) RRETURN(MATCH_NOMATCH);
1988      ecode++;      ecode++;
1989      break;      break;
1990    
1991      /* Multiline mode: start of subject unless notbol, or after any newline. */      /* Multiline mode: start of subject unless notbol, or after any newline. */
1992    
1993      case OP_CIRCM:      case OP_CIRCM:
1994      if (md->notbol && eptr == md->start_subject) MRRETURN(MATCH_NOMATCH);      if (md->notbol && eptr == md->start_subject) RRETURN(MATCH_NOMATCH);
1995      if (eptr != md->start_subject &&      if (eptr != md->start_subject &&
1996          (eptr == md->end_subject || !WAS_NEWLINE(eptr)))          (eptr == md->end_subject || !WAS_NEWLINE(eptr)))
1997        MRRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
1998      ecode++;      ecode++;
1999      break;      break;
2000    
2001      /* Start of match assertion */      /* Start of match assertion */
2002    
2003      case OP_SOM:      case OP_SOM:
2004      if (eptr != md->start_subject + md->start_offset) MRRETURN(MATCH_NOMATCH);      if (eptr != md->start_subject + md->start_offset) RRETURN(MATCH_NOMATCH);
2005      ecode++;      ecode++;
2006      break;      break;
2007    
# Line 2017  for (;;) Line 2017  for (;;)
2017    
2018      case OP_DOLLM:      case OP_DOLLM:
2019      if (eptr < md->end_subject)      if (eptr < md->end_subject)
2020        { if (!IS_NEWLINE(eptr)) MRRETURN(MATCH_NOMATCH); }        { if (!IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH); }
2021      else      else
2022        {        {
2023        if (md->noteol) MRRETURN(MATCH_NOMATCH);        if (md->noteol) RRETURN(MATCH_NOMATCH);
2024        SCHECK_PARTIAL();        SCHECK_PARTIAL();
2025        }        }
2026      ecode++;      ecode++;
# Line 2030  for (;;) Line 2030  for (;;)
2030      subject unless noteol is set. */      subject unless noteol is set. */
2031    
2032      case OP_DOLL:      case OP_DOLL:
2033      if (md->noteol) MRRETURN(MATCH_NOMATCH);      if (md->noteol) RRETURN(MATCH_NOMATCH);
2034      if (!md->endonly) goto ASSERT_NL_OR_EOS;      if (!md->endonly) goto ASSERT_NL_OR_EOS;
2035    
2036      /* ... else fall through for endonly */      /* ... else fall through for endonly */
# Line 2038  for (;;) Line 2038  for (;;)
2038      /* End of subject assertion (\z) */      /* End of subject assertion (\z) */
2039    
2040      case OP_EOD:      case OP_EOD:
2041      if (eptr < md->end_subject) MRRETURN(MATCH_NOMATCH);      if (eptr < md->end_subject) RRETURN(MATCH_NOMATCH);
2042      SCHECK_PARTIAL();      SCHECK_PARTIAL();
2043      ecode++;      ecode++;
2044      break;      break;
# Line 2049  for (;;) Line 2049  for (;;)
2049      ASSERT_NL_OR_EOS:      ASSERT_NL_OR_EOS:
2050      if (eptr < md->end_subject &&      if (eptr < md->end_subject &&
2051          (!IS_NEWLINE(eptr) || eptr != md->end_subject - md->nllen))          (!IS_NEWLINE(eptr) || eptr != md->end_subject - md->nllen))
2052        MRRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2053    
2054      /* Either at end of string or \n before end. */      /* Either at end of string or \n before end. */
2055    
# Line 2068  for (;;) Line 2068  for (;;)
2068        be "non-word" characters. Remember the earliest consulted character for        be "non-word" characters. Remember the earliest consulted character for
2069        partial matching. */        partial matching. */
2070    
2071  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
2072        if (utf8)        if (utf)
2073          {          {
2074          /* Get status of previous character */          /* Get status of previous character */
2075    
2076          if (eptr == md->start_subject) prev_is_word = FALSE; else          if (eptr == md->start_subject) prev_is_word = FALSE; else
2077            {            {
2078            USPTR lastptr = eptr - 1;            PCRE_PUCHAR lastptr = eptr - 1;
2079            while((*lastptr & 0xc0) == 0x80) lastptr--;            BACKCHAR(lastptr);
2080            if (lastptr < md->start_used_ptr) md->start_used_ptr = lastptr;            if (lastptr < md->start_used_ptr) md->start_used_ptr = lastptr;
2081            GETCHAR(c, lastptr);            GETCHAR(c, lastptr);
2082  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
# Line 2141  for (;;) Line 2141  for (;;)
2141              }              }
2142            else            else
2143  #endif  #endif
2144            prev_is_word = ((md->ctypes[eptr[-1]] & ctype_word) != 0);            prev_is_word = MAX_255(eptr[-1])
2145                && ((md->ctypes[eptr[-1]] & ctype_word) != 0);
2146            }            }
2147    
2148          /* Get status of next character */          /* Get status of next character */
# Line 2164  for (;;) Line 2165  for (;;)
2165            }            }
2166          else          else
2167  #endif  #endif
2168          cur_is_word = ((md->ctypes[*eptr] & ctype_word) != 0);          cur_is_word = MAX_255(*eptr)
2169              && ((md->ctypes[*eptr] & ctype_word) != 0);
2170          }          }
2171    
2172        /* Now see if the situation is what we want */        /* Now see if the situation is what we want */
2173    
2174        if ((*ecode++ == OP_WORD_BOUNDARY)?        if ((*ecode++ == OP_WORD_BOUNDARY)?
2175             cur_is_word == prev_is_word : cur_is_word != prev_is_word)             cur_is_word == prev_is_word : cur_is_word != prev_is_word)
2176          MRRETURN(MATCH_NOMATCH);          RRETURN(MATCH_NOMATCH);
2177        }        }
2178      break;      break;
2179    
2180      /* Match a single character type; inline for speed */      /* Match a single character type; inline for speed */
2181    
2182      case OP_ANY:      case OP_ANY:
2183      if (IS_NEWLINE(eptr)) MRRETURN(MATCH_NOMATCH);      if (IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH);
2184      /* Fall through */      /* Fall through */
2185    
2186      case OP_ALLANY:      case OP_ALLANY:
2187      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 */
2188        {                            /* not be updated before SCHECK_PARTIAL. */        {                            /* not be updated before SCHECK_PARTIAL. */
2189        SCHECK_PARTIAL();        SCHECK_PARTIAL();
2190        MRRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2191        }        }
2192      eptr++;      eptr++;
2193      if (utf8) while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;  #ifdef SUPPORT_UTF
2194        if (utf) ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);
2195    #endif
2196      ecode++;      ecode++;
2197      break;      break;
2198    
# Line 2199  for (;;) Line 2203  for (;;)
2203      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 */
2204        {                            /* not be updated before SCHECK_PARTIAL. */        {                            /* not be updated before SCHECK_PARTIAL. */
2205        SCHECK_PARTIAL();        SCHECK_PARTIAL();
2206        MRRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2207        }        }
2208      eptr++;      eptr++;
2209      ecode++;      ecode++;
# Line 2209  for (;;) Line 2213  for (;;)
2213      if (eptr >= md->end_subject)      if (eptr >= md->end_subject)
2214        {        {
2215        SCHECK_PARTIAL();        SCHECK_PARTIAL();
2216        MRRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2217        }        }
2218      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2219      if (      if (
2220  #ifdef SUPPORT_UTF8  #if defined SUPPORT_UTF || !(defined COMPILE_PCRE8)
2221         c < 256 &&         c < 256 &&
2222  #endif  #endif
2223         (md->ctypes[c] & ctype_digit) != 0         (md->ctypes[c] & ctype_digit) != 0
2224         )         )
2225        MRRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2226      ecode++;      ecode++;
2227      break;      break;
2228    
# Line 2226  for (;;) Line 2230  for (;;)
2230      if (eptr >= md->end_subject)      if (eptr >= md->end_subject)
2231        {        {
2232        SCHECK_PARTIAL();        SCHECK_PARTIAL();
2233        MRRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2234        }        }
2235      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2236      if (      if (
2237  #ifdef SUPPORT_UTF8  #if defined SUPPORT_UTF || !(defined COMPILE_PCRE8)
2238         c >= 256 ||         c > 255 ||
2239  #endif  #endif
2240         (md->ctypes[c] & ctype_digit) == 0         (md->ctypes[c] & ctype_digit) == 0
2241         )         )
2242        MRRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2243      ecode++;      ecode++;
2244      break;      break;
2245    
# Line 2243  for (;;) Line 2247  for (;;)
2247      if (eptr >= md->end_subject)      if (eptr >= md->end_subject)
2248        {        {
2249        SCHECK_PARTIAL();        SCHECK_PARTIAL();
2250        MRRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2251        }        }
2252      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2253      if (      if (
2254  #ifdef SUPPORT_UTF8  #if defined SUPPORT_UTF || !(defined COMPILE_PCRE8)
2255         c < 256 &&         c < 256 &&
2256  #endif  #endif
2257         (md->ctypes[c] & ctype_space) != 0         (md->ctypes[c] & ctype_space) != 0
2258         )         )
2259        MRRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2260      ecode++;      ecode++;
2261      break;      break;
2262    
# Line 2260  for (;;) Line 2264  for (;;)
2264      if (eptr >= md->end_subject)      if (eptr >= md->end_subject)
2265        {        {
2266        SCHECK_PARTIAL();        SCHECK_PARTIAL();
2267        MRRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2268        }        }
2269      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2270      if (      if (
2271  #ifdef SUPPORT_UTF8  #if defined SUPPORT_UTF || !(defined COMPILE_PCRE8)
2272         c >= 256 ||         c > 255 ||
2273  #endif  #endif
2274         (md->ctypes[c] & ctype_space) == 0         (md->ctypes[c] & ctype_space) == 0
2275         )         )
2276        MRRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2277      ecode++;      ecode++;
2278      break;      break;
2279    
# Line 2277  for (;;) Line 2281  for (;;)
2281      if (eptr >= md->end_subject)      if (eptr >= md->end_subject)
2282        {        {
2283        SCHECK_PARTIAL();        SCHECK_PARTIAL();
2284        MRRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2285        }        }
2286      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2287      if (      if (
2288  #ifdef SUPPORT_UTF8  #if defined SUPPORT_UTF || !(defined COMPILE_PCRE8)
2289         c < 256 &&         c < 256 &&
2290  #endif  #endif
2291         (md->ctypes[c] & ctype_word) != 0         (md->ctypes[c] & ctype_word) != 0
2292         )         )
2293        MRRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2294      ecode++;      ecode++;
2295      break;      break;
2296    
# Line 2294  for (;;) Line 2298  for (;;)
2298      if (eptr >= md->end_subject)      if (eptr >= md->end_subject)
2299        {        {
2300        SCHECK_PARTIAL();        SCHECK_PARTIAL();
2301        MRRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2302        }        }
2303      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2304      if (      if (
2305  #ifdef SUPPORT_UTF8  #if defined SUPPORT_UTF || !(defined COMPILE_PCRE8)
2306         c >= 256 ||         c > 255 ||
2307  #endif  #endif
2308         (md->ctypes[c] & ctype_word) == 0         (md->ctypes[c] & ctype_word) == 0
2309         )         )
2310        MRRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2311      ecode++;      ecode++;
2312      break;      break;
2313    
# Line 2311  for (;;) Line 2315  for (;;)
2315      if (eptr >= md->end_subject)      if (eptr >= md->end_subject)
2316        {        {
2317        SCHECK_PARTIAL();        SCHECK_PARTIAL();
2318        MRRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2319        }        }
2320      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2321      switch(c)      switch(c)
2322        {        {
2323        default: MRRETURN(MATCH_NOMATCH);        default: RRETURN(MATCH_NOMATCH);
2324    
2325        case 0x000d:        case 0x000d:
2326        if (eptr < md->end_subject && *eptr == 0x0a) eptr++;        if (eptr < md->end_subject && *eptr == 0x0a) eptr++;
# Line 2330  for (;;) Line 2334  for (;;)
2334        case 0x0085:        case 0x0085:
2335        case 0x2028:        case 0x2028:
2336        case 0x2029:        case 0x2029:
2337        if (md->bsr_anycrlf) MRRETURN(MATCH_NOMATCH);        if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);
2338        break;        break;
2339        }        }
2340      ecode++;      ecode++;
# Line 2340  for (;;) Line 2344  for (;;)
2344      if (eptr >= md->end_subject)      if (eptr >= md->end_subject)
2345        {        {
2346        SCHECK_PARTIAL();        SCHECK_PARTIAL();
2347        MRRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2348        }        }
2349      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2350      switch(c)      switch(c)
# Line 2365  for (;;) Line 2369  for (;;)
2369        case 0x202f:    /* NARROW NO-BREAK SPACE */        case 0x202f:    /* NARROW NO-BREAK SPACE */
2370        case 0x205f:    /* MEDIUM MATHEMATICAL SPACE */        case 0x205f:    /* MEDIUM MATHEMATICAL SPACE */
2371        case 0x3000:    /* IDEOGRAPHIC SPACE */        case 0x3000:    /* IDEOGRAPHIC SPACE */
2372        MRRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2373        }        }
2374      ecode++;      ecode++;
2375      break;      break;
# Line 2374  for (;;) Line 2378  for (;;)
2378      if (eptr >= md->end_subject)      if (eptr >= md->end_subject)
2379        {        {
2380        SCHECK_PARTIAL();        SCHECK_PARTIAL();
2381        MRRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2382        }        }
2383      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2384      switch(c)      switch(c)
2385        {        {
2386        default: MRRETURN(MATCH_NOMATCH);        default: RRETURN(MATCH_NOMATCH);
2387        case 0x09:      /* HT */        case 0x09:      /* HT */
2388        case 0x20:      /* SPACE */        case 0x20:      /* SPACE */
2389        case 0xa0:      /* NBSP */        case 0xa0:      /* NBSP */
# Line 2408  for (;;) Line 2412  for (;;)
2412      if (eptr >= md->end_subject)      if (eptr >= md->end_subject)
2413        {        {
2414        SCHECK_PARTIAL();        SCHECK_PARTIAL();
2415        MRRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2416        }        }
2417      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2418      switch(c)      switch(c)
# Line 2421  for (;;) Line 2425  for (;;)
2425        case 0x85:      /* NEL */        case 0x85:      /* NEL */
2426        case 0x2028:    /* LINE SEPARATOR */        case 0x2028:    /* LINE SEPARATOR */
2427        case 0x2029:    /* PARAGRAPH SEPARATOR */        case 0x2029:    /* PARAGRAPH SEPARATOR */
2428        MRRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2429        }        }
2430      ecode++;      ecode++;
2431      break;      break;
# Line 2430  for (;;) Line 2434  for (;;)
2434      if (eptr >= md->end_subject)      if (eptr >= md->end_subject)
2435        {        {
2436        SCHECK_PARTIAL();        SCHECK_PARTIAL();
2437        MRRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2438        }        }
2439      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2440      switch(c)      switch(c)
2441        {        {
2442        default: MRRETURN(MATCH_NOMATCH);        default: RRETURN(MATCH_NOMATCH);
2443        case 0x0a:      /* LF */        case 0x0a:      /* LF */
2444        case 0x0b:      /* VT */        case 0x0b:      /* VT */
2445        case 0x0c:      /* FF */        case 0x0c:      /* FF */
# Line 2457  for (;;) Line 2461  for (;;)
2461      if (eptr >= md->end_subject)      if (eptr >= md->end_subject)
2462        {        {
2463        SCHECK_PARTIAL();        SCHECK_PARTIAL();
2464        MRRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2465        }        }
2466      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2467        {        {
# Line 2466  for (;;) Line 2470  for (;;)
2470        switch(ecode[1])        switch(ecode[1])
2471          {          {
2472          case PT_ANY:          case PT_ANY:
2473          if (op == OP_NOTPROP) MRRETURN(MATCH_NOMATCH);          if (op == OP_NOTPROP) RRETURN(MATCH_NOMATCH);
2474          break;          break;
2475    
2476          case PT_LAMP:          case PT_LAMP:
2477          if ((prop->chartype == ucp_Lu ||          if ((prop->chartype == ucp_Lu ||
2478               prop->chartype == ucp_Ll ||               prop->chartype == ucp_Ll ||
2479               prop->chartype == ucp_Lt) == (op == OP_NOTPROP))               prop->chartype == ucp_Lt) == (op == OP_NOTPROP))
2480            MRRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
2481          break;          break;
2482    
2483          case PT_GC:          case PT_GC:
2484          if ((ecode[2] != _pcre_ucp_gentype[prop->chartype]) == (op == OP_PROP))          if ((ecode[2] != PRIV(ucp_gentype)[prop->chartype]) == (op == OP_PROP))
2485            MRRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
2486          break;          break;
2487    
2488          case PT_PC:          case PT_PC:
2489          if ((ecode[2] != prop->chartype) == (op == OP_PROP))          if ((ecode[2] != prop->chartype) == (op == OP_PROP))
2490            MRRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
2491          break;          break;
2492    
2493          case PT_SC:          case PT_SC:
2494          if ((ecode[2] != prop->script) == (op == OP_PROP))          if ((ecode[2] != prop->script) == (op == OP_PROP))
2495            MRRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
2496          break;          break;
2497    
2498          /* These are specials */          /* These are specials */
2499    
2500          case PT_ALNUM:          case PT_ALNUM:
2501          if ((_pcre_ucp_gentype[prop->chartype] == ucp_L ||          if ((PRIV(ucp_gentype)[prop->chartype] == ucp_L ||
2502               _pcre_ucp_gentype[prop->chartype] == ucp_N) == (op == OP_NOTPROP))               PRIV(ucp_gentype)[prop->chartype] == ucp_N) == (op == OP_NOTPROP))
2503            MRRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
2504          break;          break;
2505    
2506          case PT_SPACE:    /* Perl space */          case PT_SPACE:    /* Perl space */
2507          if ((_pcre_ucp_gentype[prop->chartype] == ucp_Z ||          if ((PRIV(ucp_gentype)[prop->chartype] == ucp_Z ||
2508               c == CHAR_HT || c == CHAR_NL || c == CHAR_FF || c == CHAR_CR)               c == CHAR_HT || c == CHAR_NL || c == CHAR_FF || c == CHAR_CR)
2509                 == (op == OP_NOTPROP))                 == (op == OP_NOTPROP))
2510            MRRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
2511          break;          break;
2512    
2513          case PT_PXSPACE:  /* POSIX space */          case PT_PXSPACE:  /* POSIX space */
2514          if ((_pcre_ucp_gentype[prop->chartype] == ucp_Z ||          if ((PRIV(ucp_gentype)[prop->chartype] == ucp_Z ||
2515               c == CHAR_HT || c == CHAR_NL || c == CHAR_VT ||               c == CHAR_HT || c == CHAR_NL || c == CHAR_VT ||
2516               c == CHAR_FF || c == CHAR_CR)               c == CHAR_FF || c == CHAR_CR)
2517                 == (op == OP_NOTPROP))                 == (op == OP_NOTPROP))
2518            MRRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
2519          break;          break;
2520    
2521          case PT_WORD:          case PT_WORD:
2522          if ((_pcre_ucp_gentype[prop->chartype] == ucp_L ||          if ((PRIV(ucp_gentype)[prop->chartype] == ucp_L ||
2523               _pcre_ucp_gentype[prop->chartype] == ucp_N ||               PRIV(ucp_gentype)[prop->chartype] == ucp_N ||
2524               c == CHAR_UNDERSCORE) == (op == OP_NOTPROP))               c == CHAR_UNDERSCORE) == (op == OP_NOTPROP))
2525            MRRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
2526          break;          break;
2527    
2528          /* This should never occur */          /* This should never occur */
# Line 2538  for (;;) Line 2542  for (;;)
2542      if (eptr >= md->end_subject)      if (eptr >= md->end_subject)
2543        {        {
2544        SCHECK_PARTIAL();        SCHECK_PARTIAL();
2545        MRRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2546        }        }
2547      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2548      if (UCD_CATEGORY(c) == ucp_M) MRRETURN(MATCH_NOMATCH);      if (UCD_CATEGORY(c) == ucp_M) RRETURN(MATCH_NOMATCH);
2549      while (eptr < md->end_subject)      while (eptr < md->end_subject)
2550        {        {
2551        int len = 1;        int len = 1;
2552        if (!utf8) c = *eptr; else { GETCHARLEN(c, eptr, len); }        if (!utf) c = *eptr; else { GETCHARLEN(c, eptr, len); }
2553        if (UCD_CATEGORY(c) != ucp_M) break;        if (UCD_CATEGORY(c) != ucp_M) break;
2554        eptr += len;        eptr += len;
2555        }        }
# Line 2566  for (;;) Line 2570  for (;;)
2570      case OP_REFI:      case OP_REFI:
2571      caseless = op == OP_REFI;      caseless = op == OP_REFI;
2572      offset = GET2(ecode, 1) << 1;               /* Doubled ref number */      offset = GET2(ecode, 1) << 1;               /* Doubled ref number */
2573      ecode += 3;      ecode += 1 + IMM2_SIZE;
2574    
2575      /* If the reference is unset, there are two possibilities:      /* If the reference is unset, there are two possibilities:
2576    
# Line 2606  for (;;) Line 2610  for (;;)
2610        case OP_CRMINRANGE:        case OP_CRMINRANGE:
2611        minimize = (*ecode == OP_CRMINRANGE);        minimize = (*ecode == OP_CRMINRANGE);
2612        min = GET2(ecode, 1);        min = GET2(ecode, 1);
2613        max = GET2(ecode, 3);        max = GET2(ecode, 1 + IMM2_SIZE);
2614        if (max == 0) max = INT_MAX;        if (max == 0) max = INT_MAX;
2615        ecode += 5;        ecode += 1 + 2 * IMM2_SIZE;
2616        break;        break;
2617    
2618        default:               /* No repeat follows */        default:               /* No repeat follows */
2619        if ((length = match_ref(offset, eptr, length, md, caseless)) < 0)        if ((length = match_ref(offset, eptr, length, md, caseless)) < 0)
2620          {          {
2621          CHECK_PARTIAL();          CHECK_PARTIAL();
2622          MRRETURN(MATCH_NOMATCH);          RRETURN(MATCH_NOMATCH);
2623          }          }
2624        eptr += length;        eptr += length;
2625        continue;              /* With the main loop */        continue;              /* With the main loop */
# Line 2636  for (;;) Line 2640  for (;;)
2640        if ((slength = match_ref(offset, eptr, length, md, caseless)) < 0)        if ((slength = match_ref(offset, eptr, length, md, caseless)) < 0)
2641          {          {
2642          CHECK_PARTIAL();          CHECK_PARTIAL();
2643          MRRETURN(MATCH_NOMATCH);          RRETURN(MATCH_NOMATCH);
2644          }          }
2645        eptr += slength;        eptr += slength;
2646        }        }
# Line 2655  for (;;) Line 2659  for (;;)
2659          int slength;          int slength;
2660          RMATCH(eptr, ecode, offset_top, md, eptrb, RM14);          RMATCH(eptr, ecode, offset_top, md, eptrb, RM14);
2661          if (rrc != MATCH_NOMATCH) RRETURN(rrc);          if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2662          if (fi >= max) MRRETURN(MATCH_NOMATCH);          if (fi >= max) RRETURN(MATCH_NOMATCH);
2663          if ((slength = match_ref(offset, eptr, length, md, caseless)) < 0)          if ((slength = match_ref(offset, eptr, length, md, caseless)) < 0)
2664            {            {
2665            CHECK_PARTIAL();            CHECK_PARTIAL();
2666            MRRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
2667            }            }
2668          eptr += slength;          eptr += slength;
2669          }          }
# Line 2687  for (;;) Line 2691  for (;;)
2691          if (rrc != MATCH_NOMATCH) RRETURN(rrc);          if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2692          eptr -= length;          eptr -= length;
2693          }          }
2694        MRRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2695        }        }
2696      /* Control never gets here */      /* Control never gets here */
2697    
# Line 2705  for (;;) Line 2709  for (;;)
2709      case OP_NCLASS:      case OP_NCLASS:
2710      case OP_CLASS:      case OP_CLASS:
2711        {        {
2712          /* The data variable is saved across frames, so the byte map needs to
2713          be stored there. */
2714    #define BYTE_MAP ((pcre_uint8 *)data)
2715        data = ecode + 1;                /* Save for matching */        data = ecode + 1;                /* Save for matching */
2716        ecode += 33;                     /* Advance past the item */        ecode += 1 + (32 / sizeof(pcre_uchar)); /* Advance past the item */
2717    
2718        switch (*ecode)        switch (*ecode)
2719          {          {
# Line 2727  for (;;) Line 2734  for (;;)
2734          case OP_CRMINRANGE:          case OP_CRMINRANGE:
2735          minimize = (*ecode == OP_CRMINRANGE);          minimize = (*ecode == OP_CRMINRANGE);
2736          min = GET2(ecode, 1);          min = GET2(ecode, 1);
2737          max = GET2(ecode, 3);          max = GET2(ecode, 1 + IMM2_SIZE);
2738          if (max == 0) max = INT_MAX;          if (max == 0) max = INT_MAX;
2739          ecode += 5;          ecode += 1 + 2 * IMM2_SIZE;
2740          break;          break;
2741    
2742          default:               /* No repeat follows */          default:               /* No repeat follows */
# Line 2739  for (;;) Line 2746  for (;;)
2746    
2747        /* First, ensure the minimum number of matches are present. */        /* First, ensure the minimum number of matches are present. */
2748    
2749  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
2750        /* UTF-8 mode */        if (utf)
       if (utf8)  
2751          {          {
2752          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
2753            {            {
2754            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
2755              {              {
2756              SCHECK_PARTIAL();              SCHECK_PARTIAL();
2757              MRRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
2758              }              }
2759            GETCHARINC(c, eptr);            GETCHARINC(c, eptr);
2760            if (c > 255)            if (c > 255)
2761              {              {
2762              if (op == OP_CLASS) MRRETURN(MATCH_NOMATCH);              if (op == OP_CLASS) RRETURN(MATCH_NOMATCH);
2763              }              }
2764            else            else
2765              {              if ((BYTE_MAP[c/8] & (1 << (c&7))) == 0) RRETURN(MATCH_NOMATCH);
             if ((data[c/8] & (1 << (c&7))) == 0) MRRETURN(MATCH_NOMATCH);  
             }  
2766            }            }
2767          }          }
2768        else        else
2769  #endif  #endif
2770        /* Not UTF-8 mode */        /* Not UTF mode */
2771          {          {
2772          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
2773            {            {
2774            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
2775              {              {
2776              SCHECK_PARTIAL();              SCHECK_PARTIAL();
2777              MRRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
2778              }              }
2779            c = *eptr++;            c = *eptr++;
2780            if ((data[c/8] & (1 << (c&7))) == 0) MRRETURN(MATCH_NOMATCH);  #ifndef COMPILE_PCRE8
2781              if (c > 255)
2782                {
2783                if (op == OP_CLASS) RRETURN(MATCH_NOMATCH);
2784                }
2785              else
2786    #endif
2787                if ((BYTE_MAP[c/8] & (1 << (c&7))) == 0) RRETURN(MATCH_NOMATCH);
2788            }            }
2789          }          }
2790    
# Line 2787  for (;;) Line 2798  for (;;)
2798    
2799        if (minimize)        if (minimize)
2800          {          {
2801  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
2802          /* UTF-8 mode */          if (utf)
         if (utf8)  
2803            {            {
2804            for (fi = min;; fi++)            for (fi = min;; fi++)
2805              {              {
2806              RMATCH(eptr, ecode, offset_top, md, eptrb, RM16);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM16);
2807              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2808              if (fi >= max) MRRETURN(MATCH_NOMATCH);              if (fi >= max) RRETURN(MATCH_NOMATCH);
2809              if (eptr >= md->end_subject)              if (eptr >= md->end_subject)
2810                {                {
2811                SCHECK_PARTIAL();                SCHECK_PARTIAL();
2812                MRRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
2813                }                }
2814              GETCHARINC(c, eptr);              GETCHARINC(c, eptr);
2815              if (c > 255)              if (c > 255)
2816                {                {
2817                if (op == OP_CLASS) MRRETURN(MATCH_NOMATCH);                if (op == OP_CLASS) RRETURN(MATCH_NOMATCH);
2818                }                }
2819              else              else
2820                {                if ((BYTE_MAP[c/8] & (1 << (c&7))) == 0) RRETURN(MATCH_NOMATCH);
               if ((data[c/8] & (1 << (c&7))) == 0) MRRETURN(MATCH_NOMATCH);  
               }  
2821              }              }
2822            }            }
2823          else          else
2824  #endif  #endif
2825          /* Not UTF-8 mode */          /* Not UTF mode */
2826            {            {
2827            for (fi = min;; fi++)            for (fi = min;; fi++)
2828              {              {
2829              RMATCH(eptr, ecode, offset_top, md, eptrb, RM17);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM17);
2830              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2831              if (fi >= max) MRRETURN(MATCH_NOMATCH);              if (fi >= max) RRETURN(MATCH_NOMATCH);
2832              if (eptr >= md->end_subject)              if (eptr >= md->end_subject)
2833                {                {
2834                SCHECK_PARTIAL();                SCHECK_PARTIAL();
2835                MRRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
2836                }                }
2837              c = *eptr++;              c = *eptr++;
2838              if ((data[c/8] & (1 << (c&7))) == 0) MRRETURN(MATCH_NOMATCH);  #ifndef COMPILE_PCRE8
2839                if (c > 255)
2840                  {
2841                  if (op == OP_CLASS) RRETURN(MATCH_NOMATCH);
2842                  }
2843                else
2844    #endif
2845                  if ((BYTE_MAP[c/8] & (1 << (c&7))) == 0) RRETURN(MATCH_NOMATCH);
2846              }              }
2847            }            }
2848          /* Control never gets here */          /* Control never gets here */
# Line 2839  for (;;) Line 2854  for (;;)
2854          {          {
2855          pp = eptr;          pp = eptr;
2856    
2857  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
2858          /* UTF-8 mode */          if (utf)
         if (utf8)  
2859            {            {
2860            for (i = min; i < max; i++)            for (i = min; i < max; i++)
2861              {              {
# Line 2857  for (;;) Line 2871  for (;;)
2871                if (op == OP_CLASS) break;                if (op == OP_CLASS) break;
2872                }                }
2873              else              else
2874                {                if ((BYTE_MAP[c/8] & (1 << (c&7))) == 0) break;
               if ((data[c/8] & (1 << (c&7))) == 0) break;  
               }  
2875              eptr += len;              eptr += len;
2876              }              }
2877            for (;;)            for (;;)
# Line 2872  for (;;) Line 2884  for (;;)
2884            }            }
2885          else          else
2886  #endif  #endif
2887            /* Not UTF-8 mode */            /* Not UTF mode */
2888            {            {
2889            for (i = min; i < max; i++)            for (i = min; i < max; i++)
2890              {              {
# Line 2882  for (;;) Line 2894  for (;;)
2894                break;                break;
2895                }                }
2896              c = *eptr;              c = *eptr;
2897              if ((data[c/8] & (1 << (c&7))) == 0) break;  #ifndef COMPILE_PCRE8
2898                if (c > 255)
2899                  {
2900                  if (op == OP_CLASS) break;
2901                  }
2902                else
2903    #endif
2904                  if ((BYTE_MAP[c/8] & (1 << (c&7))) == 0) break;
2905              eptr++;              eptr++;
2906              }              }
2907            while (eptr >= pp)            while (eptr >= pp)
# Line 2893  for (;;) Line 2912  for (;;)
2912              }              }
2913            }            }
2914    
2915          MRRETURN(MATCH_NOMATCH);          RRETURN(MATCH_NOMATCH);
2916          }          }
2917    #undef BYTE_MAP
2918        }        }
2919      /* Control never gets here */      /* Control never gets here */
2920    
# Line 2903  for (;;) Line 2923  for (;;)
2923      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
2924      mode, because Unicode properties are supported in non-UTF-8 mode. */      mode, because Unicode properties are supported in non-UTF-8 mode. */
2925    
2926  #ifdef SUPPORT_UTF8  #if defined SUPPORT_UTF || !defined COMPILE_PCRE8
2927      case OP_XCLASS:      case OP_XCLASS:
2928        {        {
2929        data = ecode + 1 + LINK_SIZE;                /* Save for matching */        data = ecode + 1 + LINK_SIZE;                /* Save for matching */
# Line 2928  for (;;) Line 2948  for (;;)
2948          case OP_CRMINRANGE:          case OP_CRMINRANGE:
2949          minimize = (*ecode == OP_CRMINRANGE);          minimize = (*ecode == OP_CRMINRANGE);
2950          min = GET2(ecode, 1);          min = GET2(ecode, 1);
2951          max = GET2(ecode, 3);          max = GET2(ecode, 1 + IMM2_SIZE);
2952          if (max == 0) max = INT_MAX;          if (max == 0) max = INT_MAX;
2953          ecode += 5;          ecode += 1 + 2 * IMM2_SIZE;
2954          break;          break;
2955    
2956          default:               /* No repeat follows */          default:               /* No repeat follows */
# Line 2945  for (;;) Line 2965  for (;;)
2965          if (eptr >= md->end_subject)          if (eptr >= md->end_subject)
2966            {            {
2967            SCHECK_PARTIAL();            SCHECK_PARTIAL();
2968            MRRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
2969            }            }
2970          GETCHARINCTEST(c, eptr);          GETCHARINCTEST(c, eptr);
2971          if (!_pcre_xclass(c, data)) MRRETURN(MATCH_NOMATCH);          if (!PRIV(xclass)(c, data, utf)) RRETURN(MATCH_NOMATCH);
2972          }          }
2973    
2974        /* If max == min we can continue with the main loop without the        /* If max == min we can continue with the main loop without the
# Line 2965  for (;;) Line 2985  for (;;)
2985            {            {
2986            RMATCH(eptr, ecode, offset_top, md, eptrb, RM20);            RMATCH(eptr, ecode, offset_top, md, eptrb, RM20);
2987            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2988            if (fi >= max) MRRETURN(MATCH_NOMATCH);            if (fi >= max) RRETURN(MATCH_NOMATCH);
2989            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
2990              {              {
2991              SCHECK_PARTIAL();              SCHECK_PARTIAL();
2992              MRRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
2993              }              }
2994            GETCHARINCTEST(c, eptr);            GETCHARINCTEST(c, eptr);
2995            if (!_pcre_xclass(c, data)) MRRETURN(MATCH_NOMATCH);            if (!PRIV(xclass)(c, data, utf)) RRETURN(MATCH_NOMATCH);
2996            }            }
2997          /* Control never gets here */          /* Control never gets here */
2998          }          }
# Line 2990  for (;;) Line 3010  for (;;)
3010              SCHECK_PARTIAL();              SCHECK_PARTIAL();
3011              break;              break;
3012              }              }
3013    #ifdef SUPPORT_UTF
3014            GETCHARLENTEST(c, eptr, len);            GETCHARLENTEST(c, eptr, len);
3015            if (!_pcre_xclass(c, data)) break;  #else
3016              c = *eptr;
3017    #endif
3018              if (!PRIV(xclass)(c, data, utf)) break;
3019            eptr += len;            eptr += len;
3020            }            }
3021          for(;;)          for(;;)
# Line 2999  for (;;) Line 3023  for (;;)
3023            RMATCH(eptr, ecode, offset_top, md, eptrb, RM21);            RMATCH(eptr, ecode, offset_top, md, eptrb, RM21);
3024            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3025            if (eptr-- == pp) break;        /* Stop if tried at original pos */            if (eptr-- == pp) break;        /* Stop if tried at original pos */
3026            if (utf8) BACKCHAR(eptr);  #ifdef SUPPORT_UTF
3027              if (utf) BACKCHAR(eptr);
3028    #endif
3029            }            }
3030          MRRETURN(MATCH_NOMATCH);          RRETURN(MATCH_NOMATCH);
3031          }          }
3032    
3033        /* Control never gets here */        /* Control never gets here */
# Line 3011  for (;;) Line 3037  for (;;)
3037      /* Match a single character, casefully */      /* Match a single character, casefully */
3038    
3039      case OP_CHAR:      case OP_CHAR:
3040  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
3041      if (utf8)      if (utf)
3042        {        {
3043        length = 1;        length = 1;
3044        ecode++;        ecode++;
# Line 3020  for (;;) Line 3046  for (;;)
3046        if (length > md->end_subject - eptr)        if (length > md->end_subject - eptr)
3047          {          {
3048          CHECK_PARTIAL();             /* Not SCHECK_PARTIAL() */          CHECK_PARTIAL();             /* Not SCHECK_PARTIAL() */
3049          MRRETURN(MATCH_NOMATCH);          RRETURN(MATCH_NOMATCH);
3050          }          }
3051        while (length-- > 0) if (*ecode++ != *eptr++) MRRETURN(MATCH_NOMATCH);        while (length-- > 0) if (*ecode++ != *eptr++) RRETURN(MATCH_NOMATCH);
3052        }        }
3053      else      else
3054  #endif  #endif
3055        /* Not UTF mode */
     /* Non-UTF-8 mode */  
3056        {        {
3057        if (md->end_subject - eptr < 1)        if (md->end_subject - eptr < 1)
3058          {          {
3059          SCHECK_PARTIAL();            /* This one can use SCHECK_PARTIAL() */          SCHECK_PARTIAL();            /* This one can use SCHECK_PARTIAL() */
3060          MRRETURN(MATCH_NOMATCH);          RRETURN(MATCH_NOMATCH);
3061          }          }
3062        if (ecode[1] != *eptr++) MRRETURN(MATCH_NOMATCH);        if (ecode[1] != *eptr++) RRETURN(MATCH_NOMATCH);
3063        ecode += 2;        ecode += 2;
3064        }        }
3065      break;      break;
3066    
3067      /* Match a single character, caselessly */      /* Match a single character, caselessly. If we are at the end of the
3068        subject, give up immediately. */
3069    
3070      case OP_CHARI:      case OP_CHARI:
3071  #ifdef SUPPORT_UTF8      if (eptr >= md->end_subject)
3072      if (utf8)        {
3073          SCHECK_PARTIAL();
3074          RRETURN(MATCH_NOMATCH);
3075          }
3076    
3077    #ifdef SUPPORT_UTF
3078        if (utf)
3079        {        {
3080        length = 1;        length = 1;
3081        ecode++;        ecode++;
3082        GETCHARLEN(fc, ecode, length);        GETCHARLEN(fc, ecode, length);
3083    
       if (length > md->end_subject - eptr)  
         {  
         CHECK_PARTIAL();             /* Not SCHECK_PARTIAL() */  
         MRRETURN(MATCH_NOMATCH);  
         }  
   
3084        /* If the pattern character's value is < 128, we have only one byte, and        /* If the pattern character's value is < 128, we have only one byte, and
3085        can use the fast lookup table. */        we know that its other case must also be one byte long, so we can use the
3086          fast lookup table. We know that there is at least one byte left in the
3087          subject. */
3088    
3089        if (fc < 128)        if (fc < 128)
3090          {          {
3091          if (md->lcc[*ecode++] != md->lcc[*eptr++]) MRRETURN(MATCH_NOMATCH);          if (md->lcc[fc]
3092                != TABLE_GET(*eptr, md->lcc, *eptr)) RRETURN(MATCH_NOMATCH);
3093            ecode++;
3094            eptr++;
3095          }          }
3096    
3097        /* Otherwise we must pick up the subject character */        /* Otherwise we must pick up the subject character. Note that we cannot
3098          use the value of "length" to check for sufficient bytes left, because the
3099          other case of the character may have more or fewer bytes.  */
3100    
3101        else        else
3102          {          {
# Line 3079  for (;;) Line 3112  for (;;)
3112  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
3113            if (dc != UCD_OTHERCASE(fc))            if (dc != UCD_OTHERCASE(fc))
3114  #endif  #endif
3115              MRRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
3116            }            }
3117          }          }
3118        }        }
3119      else      else
3120  #endif   /* SUPPORT_UTF8 */  #endif   /* SUPPORT_UTF */
3121    
3122      /* Non-UTF-8 mode */      /* Not UTF mode */
3123        {        {
3124        if (md->end_subject - eptr < 1)        if (TABLE_GET(ecode[1], md->lcc, ecode[1])
3125          {            != TABLE_GET(*eptr, md->lcc, *eptr)) RRETURN(MATCH_NOMATCH);
3126          SCHECK_PARTIAL();            /* This one can use SCHECK_PARTIAL() */        eptr++;
         MRRETURN(MATCH_NOMATCH);  
         }  
       if (md->lcc[ecode[1]] != md->lcc[*eptr++]) MRRETURN(MATCH_NOMATCH);  
3127        ecode += 2;        ecode += 2;
3128        }        }
3129      break;      break;
# Line 3103  for (;;) Line 3133  for (;;)
3133      case OP_EXACT:      case OP_EXACT:
3134      case OP_EXACTI:      case OP_EXACTI:
3135      min = max = GET2(ecode, 1);      min = max = GET2(ecode, 1);
3136      ecode += 3;      ecode += 1 + IMM2_SIZE;
3137      goto REPEATCHAR;      goto REPEATCHAR;
3138    
3139      case OP_POSUPTO:      case OP_POSUPTO:
# Line 3118  for (;;) Line 3148  for (;;)
3148      min = 0;      min = 0;
3149      max = GET2(ecode, 1);      max = GET2(ecode, 1);
3150      minimize = *ecode == OP_MINUPTO || *ecode == OP_MINUPTOI;      minimize = *ecode == OP_MINUPTO || *ecode == OP_MINUPTOI;
3151      ecode += 3;      ecode += 1 + IMM2_SIZE;
3152      goto REPEATCHAR;      goto REPEATCHAR;
3153    
3154      case OP_POSSTAR:      case OP_POSSTAR:
# Line 3166  for (;;) Line 3196  for (;;)
3196      /* Common code for all repeated single-character matches. */      /* Common code for all repeated single-character matches. */
3197    
3198      REPEATCHAR:      REPEATCHAR:
3199  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
3200      if (utf8)      if (utf)
3201        {        {
3202        length = 1;        length = 1;
3203        charptr = ecode;        charptr = ecode;
# Line 3183  for (;;) Line 3213  for (;;)
3213          unsigned int othercase;          unsigned int othercase;
3214          if (op >= OP_STARI &&     /* Caseless */          if (op >= OP_STARI &&     /* Caseless */
3215              (othercase = UCD_OTHERCASE(fc)) != fc)              (othercase = UCD_OTHERCASE(fc)) != fc)
3216            oclength = _pcre_ord2utf8(othercase, occhars);            oclength = PRIV(ord2utf)(othercase, occhars);
3217          else oclength = 0;          else oclength = 0;
3218  #endif  /* SUPPORT_UCP */  #endif  /* SUPPORT_UCP */
3219    
3220          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
3221            {            {
3222            if (eptr <= md->end_subject - length &&            if (eptr <= md->end_subject - length &&
3223              memcmp(eptr, charptr, length) == 0) eptr += length;              memcmp(eptr, charptr, IN_UCHARS(length)) == 0) eptr += length;
3224  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
3225            else if (oclength > 0 &&            else if (oclength > 0 &&
3226                     eptr <= md->end_subject - oclength &&                     eptr <= md->end_subject - oclength &&
3227                     memcmp(eptr, occhars, oclength) == 0) eptr += oclength;                     memcmp(eptr, occhars, IN_UCHARS(oclength)) == 0) eptr += oclength;
3228  #endif  /* SUPPORT_UCP */  #endif  /* SUPPORT_UCP */
3229            else            else
3230              {              {
3231              CHECK_PARTIAL();              CHECK_PARTIAL();
3232              MRRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
3233              }              }
3234            }            }
3235    
# Line 3211  for (;;) Line 3241  for (;;)
3241              {              {
3242              RMATCH(eptr, ecode, offset_top, md, eptrb, RM22);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM22);
3243              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3244              if (fi >= max) MRRETURN(MATCH_NOMATCH);              if (fi >= max) RRETURN(MATCH_NOMATCH);
3245              if (eptr <= md->end_subject - length &&              if (eptr <= md->end_subject - length &&
3246                memcmp(eptr, charptr, length) == 0) eptr += length;                memcmp(eptr, charptr, IN_UCHARS(length)) == 0) eptr += length;
3247  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
3248              else if (oclength > 0 &&              else if (oclength > 0 &&
3249                       eptr <= md->end_subject - oclength &&                       eptr <= md->end_subject - oclength &&
3250                       memcmp(eptr, occhars, oclength) == 0) eptr += oclength;                       memcmp(eptr, occhars, IN_UCHARS(oclength)) == 0) eptr += oclength;
3251  #endif  /* SUPPORT_UCP */  #endif  /* SUPPORT_UCP */
3252              else              else
3253                {                {
3254                CHECK_PARTIAL();                CHECK_PARTIAL();
3255                MRRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
3256                }                }
3257              }              }
3258            /* Control never gets here */            /* Control never gets here */
# Line 3234  for (;;) Line 3264  for (;;)
3264            for (i = min; i < max; i++)            for (i = min; i < max; i++)
3265              {              {
3266              if (eptr <= md->end_subject - length &&              if (eptr <= md->end_subject - length &&
3267                  memcmp(eptr, charptr, length) == 0) eptr += length;                  memcmp(eptr, charptr, IN_UCHARS(length)) == 0) eptr += length;
3268  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
3269              else if (oclength > 0 &&              else if (oclength > 0 &&
3270                       eptr <= md->end_subject - oclength &&                       eptr <= md->end_subject - oclength &&
3271                       memcmp(eptr, occhars, oclength) == 0) eptr += oclength;                       memcmp(eptr, occhars, IN_UCHARS(oclength)) == 0) eptr += oclength;
3272  #endif  /* SUPPORT_UCP */  #endif  /* SUPPORT_UCP */
3273              else              else
3274                {                {
# Line 3253  for (;;) Line 3283  for (;;)
3283              {              {
3284              RMATCH(eptr, ecode, offset_top, md, eptrb, RM23);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM23);
3285              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3286              if (eptr == pp) { MRRETURN(MATCH_NOMATCH); }              if (eptr == pp) { RRETURN(MATCH_NOMATCH); }
3287  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
3288              eptr--;              eptr--;
3289              BACKCHAR(eptr);              BACKCHAR(eptr);
# Line 3270  for (;;) Line 3300  for (;;)
3300        value of fc will always be < 128. */        value of fc will always be < 128. */
3301        }        }
3302      else      else
3303  #endif  /* SUPPORT_UTF8 */  #endif  /* SUPPORT_UTF */
3304          /* When not in UTF-8 mode, load a single-byte character. */
3305          fc = *ecode++;
3306    
3307      /* 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
3308        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  
3309      caseful cases, for speed, since matching characters is likely to be quite      caseful cases, for speed, since matching characters is likely to be quite
3310      common. First, ensure the minimum number of matches are present. If min =      common. First, ensure the minimum number of matches are present. If min =
3311      max, continue at the same level without recursing. Otherwise, if      max, continue at the same level without recursing. Otherwise, if
# Line 3290  for (;;) Line 3318  for (;;)
3318    
3319      if (op >= OP_STARI)  /* Caseless */      if (op >= OP_STARI)  /* Caseless */
3320        {        {
3321        fc = md->lcc[fc];  #ifdef COMPILE_PCRE8
3322          /* fc must be < 128 if UTF is enabled. */
3323          foc = md->fcc[fc];
3324    #else
3325    #ifdef SUPPORT_UTF
3326    #ifdef SUPPORT_UCP
3327          if (utf && fc > 127)
3328            foc = UCD_OTHERCASE(fc);
3329    #else
3330          if (utf && fc > 127)
3331            foc = fc;
3332    #endif /* SUPPORT_UCP */
3333          else
3334    #endif /* SUPPORT_UTF */
3335            foc = TABLE_GET(fc, md->fcc, fc);
3336    #endif /* COMPILE_PCRE8 */
3337    
3338        for (i = 1; i <= min; i++)        for (i = 1; i <= min; i++)
3339          {          {
3340          if (eptr >= md->end_subject)          if (eptr >= md->end_subject)
3341            {            {
3342            SCHECK_PARTIAL();            SCHECK_PARTIAL();
3343            MRRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
3344            }            }
3345          if (fc != md->lcc[*eptr++]) MRRETURN(MATCH_NOMATCH);          if (fc != *eptr && foc != *eptr) RRETURN(MATCH_NOMATCH);
3346            eptr++;
3347          }          }
3348        if (min == max) continue;        if (min == max) continue;
3349        if (minimize)        if (minimize)
# Line 3307  for (;;) Line 3352  for (;;)
3352            {            {
3353            RMATCH(eptr, ecode, offset_top, md, eptrb, RM24);            RMATCH(eptr, ecode, offset_top, md, eptrb, RM24);
3354            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3355            if (fi >= max) MRRETURN(MATCH_NOMATCH);            if (fi >= max) RRETURN(MATCH_NOMATCH);
3356            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
3357              {              {
3358              SCHECK_PARTIAL();              SCHECK_PARTIAL();
3359              MRRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
3360              }              }
3361            if (fc != md->lcc[*eptr++]) MRRETURN(MATCH_NOMATCH);            if (fc != *eptr && foc != *eptr) RRETURN(MATCH_NOMATCH);
3362              eptr++;
3363            }            }
3364          /* Control never gets here */          /* Control never gets here */
3365          }          }
# Line 3327  for (;;) Line 3373  for (;;)
3373              SCHECK_PARTIAL();              SCHECK_PARTIAL();
3374              break;              break;
3375              }              }
3376            if (fc != md->lcc[*eptr]) break;            if (fc != *eptr && foc != *eptr) break;
3377            eptr++;            eptr++;
3378            }            }
3379    
# Line 3339  for (;;) Line 3385  for (;;)
3385            eptr--;            eptr--;
3386            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3387            }            }
3388          MRRETURN(MATCH_NOMATCH);          RRETURN(MATCH_NOMATCH);
3389          }          }
3390        /* Control never gets here */        /* Control never gets here */
3391        }        }
# Line 3353  for (;;) Line 3399  for (;;)
3399          if (eptr >= md->end_subject)          if (eptr >= md->end_subject)
3400            {            {
3401            SCHECK_PARTIAL();            SCHECK_PARTIAL();
3402            MRRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
3403            }            }
3404          if (fc != *eptr++) MRRETURN(MATCH_NOMATCH);          if (fc != *eptr++) RRETURN(MATCH_NOMATCH);
3405          }          }
3406    
3407        if (min == max) continue;        if (min == max) continue;
# Line 3366  for (;;) Line 3412  for (;;)
3412            {            {
3413            RMATCH(eptr, ecode, offset_top, md, eptrb, RM26);            RMATCH(eptr, ecode, offset_top, md, eptrb, RM26);
3414            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3415            if (fi >= max) MRRETURN(MATCH_NOMATCH);            if (fi >= max) RRETURN(MATCH_NOMATCH);
3416            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
3417              {              {
3418              SCHECK_PARTIAL();              SCHECK_PARTIAL();
3419              MRRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
3420              }              }
3421            if (fc != *eptr++) MRRETURN(MATCH_NOMATCH);            if (fc != *eptr++) RRETURN(MATCH_NOMATCH);
3422            }            }
3423          /* Control never gets here */          /* Control never gets here */
3424          }          }
# Line 3397  for (;;) Line 3443  for (;;)
3443            eptr--;            eptr--;
3444            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3445            }            }
3446          MRRETURN(MATCH_NOMATCH);          RRETURN(MATCH_NOMATCH);
3447          }          }
3448        }        }
3449      /* Control never gets here */      /* Control never gets here */
# Line 3410  for (;;) Line 3456  for (;;)
3456      if (eptr >= md->end_subject)      if (eptr >= md->end_subject)
3457        {        {
3458        SCHECK_PARTIAL();        SCHECK_PARTIAL();
3459        MRRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
3460        }        }
3461      ecode++;      ecode++;
3462      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
3463      if (op == OP_NOTI)         /* The caseless case */      if (op == OP_NOTI)         /* The caseless case */
3464        {        {
3465  #ifdef SUPPORT_UTF8        register int ch, och;
3466        if (c < 256)        ch = *ecode++;
3467  #endif  #ifdef COMPILE_PCRE8
3468        c = md->lcc[c];        /* ch must be < 128 if UTF is enabled. */
3469        if (md->lcc[*ecode++] == c) MRRETURN(MATCH_NOMATCH);        och = md->fcc[ch];
3470    #else
3471    #ifdef SUPPORT_UTF
3472    #ifdef SUPPORT_UCP
3473          if (utf && ch > 127)
3474            och = UCD_OTHERCASE(ch);
3475    #else
3476          if (utf && ch > 127)
3477            och = ch;
3478    #endif /* SUPPORT_UCP */
3479          else
3480    #endif /* SUPPORT_UTF */
3481            och = TABLE_GET(ch, md->fcc, ch);
3482    #endif /* COMPILE_PCRE8 */
3483          if (ch == c || och == c) RRETURN(MATCH_NOMATCH);
3484        }        }
3485      else    /* Caseful */      else    /* Caseful */
3486        {        {
3487        if (*ecode++ == c) MRRETURN(MATCH_NOMATCH);        if (*ecode++ == c) RRETURN(MATCH_NOMATCH);
3488        }        }
3489      break;      break;
3490    
# Line 3438  for (;;) Line 3498  for (;;)
3498      case OP_NOTEXACT:      case OP_NOTEXACT:
3499      case OP_NOTEXACTI:      case OP_NOTEXACTI:
3500      min = max = GET2(ecode, 1);      min = max = GET2(ecode, 1);
3501      ecode += 3;      ecode += 1 + IMM2_SIZE;
3502      goto REPEATNOTCHAR;      goto REPEATNOTCHAR;
3503    
3504      case OP_NOTUPTO:      case OP_NOTUPTO:
# Line 3448  for (;;) Line 3508  for (;;)
3508      min = 0;      min = 0;
3509      max = GET2(ecode, 1);      max = GET2(ecode, 1);
3510      minimize = *ecode == OP_NOTMINUPTO || *ecode == OP_NOTMINUPTOI;      minimize = *ecode == OP_NOTMINUPTO || *ecode == OP_NOTMINUPTOI;
3511      ecode += 3;      ecode += 1 + IMM2_SIZE;
3512      goto REPEATNOTCHAR;      goto REPEATNOTCHAR;
3513    
3514      case OP_NOTPOSSTAR:      case OP_NOTPOSSTAR:
# Line 3480  for (;;) Line 3540  for (;;)
3540      possessive = TRUE;      possessive = TRUE;
3541      min = 0;      min = 0;
3542      max = GET2(ecode, 1);      max = GET2(ecode, 1);
3543      ecode += 3;      ecode += 1 + IMM2_SIZE;
3544      goto REPEATNOTCHAR;      goto REPEATNOTCHAR;
3545    
3546      case OP_NOTSTAR:      case OP_NOTSTAR:
# Line 3519  for (;;) Line 3579  for (;;)
3579    
3580      if (op >= OP_NOTSTARI)     /* Caseless */      if (op >= OP_NOTSTARI)     /* Caseless */
3581        {        {
3582        fc = md->lcc[fc];  #ifdef COMPILE_PCRE8
3583          /* fc must be < 128 if UTF is enabled. */
3584          foc = md->fcc[fc];
3585    #else
3586    #ifdef SUPPORT_UTF
3587    #ifdef SUPPORT_UCP
3588          if (utf && fc > 127)
3589            foc = UCD_OTHERCASE(fc);
3590    #else
3591          if (utf && fc > 127)
3592            foc = fc;
3593    #endif /* SUPPORT_UCP */
3594          else
3595    #endif /* SUPPORT_UTF */
3596            foc = TABLE_GET(fc, md->fcc, fc);
3597    #endif /* COMPILE_PCRE8 */
3598    
3599  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
3600        /* UTF-8 mode */        if (utf)
       if (utf8)  
3601          {          {
3602          register unsigned int d;          register unsigned int d;
3603          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
# Line 3531  for (;;) Line 3605  for (;;)
3605            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
3606              {              {
3607              SCHECK_PARTIAL();              SCHECK_PARTIAL();
3608              MRRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
3609              }              }
3610            GETCHARINC(d, eptr);            GETCHARINC(d, eptr);
3611            if (d < 256) d = md->lcc[d];            if (fc == d || foc == d) RRETURN(MATCH_NOMATCH);
           if (fc == d) MRRETURN(MATCH_NOMATCH);  
3612            }            }
3613          }          }
3614        else        else
3615  #endif  #endif
3616          /* Not UTF mode */
       /* Not UTF-8 mode */  
3617          {          {
3618          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
3619            {            {
3620            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
3621              {              {
3622              SCHECK_PARTIAL();              SCHECK_PARTIAL();
3623              MRRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
3624              }              }
3625            if (fc == md->lcc[*eptr++]) MRRETURN(MATCH_NOMATCH);            if (fc == *eptr || foc == *eptr) RRETURN(MATCH_NOMATCH);
3626              eptr++;
3627            }            }
3628          }          }
3629    
# Line 3558  for (;;) Line 3631  for (;;)
3631    
3632        if (minimize)        if (minimize)
3633          {          {
3634  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
3635          /* UTF-8 mode */          if (utf)
         if (utf8)  
3636            {            {
3637            register unsigned int d;            register unsigned int d;
3638            for (fi = min;; fi++)            for (fi = min;; fi++)
3639              {              {
3640              RMATCH(eptr, ecode, offset_top, md, eptrb, RM28);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM28);
3641              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3642              if (fi >= max) MRRETURN(MATCH_NOMATCH);              if (fi >= max) RRETURN(MATCH_NOMATCH);
3643              if (eptr >= md->end_subject)              if (eptr >= md->end_subject)
3644                {                {
3645                SCHECK_PARTIAL();                SCHECK_PARTIAL();
3646                MRRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
3647                }                }
3648              GETCHARINC(d, eptr);              GETCHARINC(d, eptr);
3649              if (d < 256) d = md->lcc[d];              if (fc == d || foc == d) RRETURN(MATCH_NOMATCH);
             if (fc == d) MRRETURN(MATCH_NOMATCH);  
3650              }              }
3651            }            }
3652          else          else
3653  #endif  #endif
3654          /* Not UTF-8 mode */          /* Not UTF mode */
3655            {            {
3656            for (fi = min;; fi++)            for (fi = min;; fi++)
3657              {              {
3658              RMATCH(eptr, ecode, offset_top, md, eptrb, RM29);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM29);
3659              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3660              if (fi >= max) MRRETURN(MATCH_NOMATCH);              if (fi >= max) RRETURN(MATCH_NOMATCH);
3661              if (eptr >= md->end_subject)              if (eptr >= md->end_subject)
3662                {                {
3663                SCHECK_PARTIAL();                SCHECK_PARTIAL();
3664                MRRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
3665                }                }
3666              if (fc == md->lcc[*eptr++]) MRRETURN(MATCH_NOMATCH);              if (fc == *eptr || foc == *eptr) RRETURN(MATCH_NOMATCH);
3667                eptr++;
3668              }              }
3669            }            }
3670          /* Control never gets here */          /* Control never gets here */
# Line 3604  for (;;) Line 3676  for (;;)
3676          {          {
3677          pp = eptr;          pp = eptr;
3678    
3679  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
3680          /* UTF-8 mode */          if (utf)
         if (utf8)  
3681            {            {
3682            register unsigned int d;            register unsigned int d;
3683            for (i = min; i < max; i++)            for (i = min; i < max; i++)
# Line 3618  for (;;) Line 3689  for (;;)
3689                break;                break;
3690                }                }
3691              GETCHARLEN(d, eptr, len);              GETCHARLEN(d, eptr, len);
3692              if (d < 256) d = md->lcc[d];              if (fc == d || foc == d) break;
             if (fc == d) break;  
3693              eptr += len;              eptr += len;
3694              }              }
3695          if (possessive) continue;          if (possessive) continue;
# Line 3633  for (;;) Line 3703  for (;;)
3703            }            }
3704          else          else
3705  #endif  #endif
3706          /* Not UTF-8 mode */          /* Not UTF mode */
3707            {            {
3708            for (i = min; i < max; i++)            for (i = min; i < max; i++)
3709              {              {
# Line 3642  for (;;) Line 3712  for (;;)
3712                SCHECK_PARTIAL();                SCHECK_PARTIAL();
3713                break;                break;
3714                }                }
3715              if (fc == md->lcc[*eptr]) break;              if (fc == *eptr || foc == *eptr) break;
3716              eptr++;              eptr++;
3717              }              }
3718            if (possessive) continue;            if (possessive) continue;
# Line 3654  for (;;) Line 3724  for (;;)
3724              }              }
3725            }            }
3726    
3727          MRRETURN(MATCH_NOMATCH);          RRETURN(MATCH_NOMATCH);
3728          }          }
3729        /* Control never gets here */        /* Control never gets here */
3730        }        }
# Line 3663  for (;;) Line 3733  for (;;)
3733    
3734      else      else
3735        {        {
3736  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
3737        /* UTF-8 mode */        if (utf)
       if (utf8)  
3738          {          {
3739          register unsigned int d;          register unsigned int d;
3740          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
# Line 3673  for (;;) Line 3742  for (;;)
3742            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
3743              {              {
3744              SCHECK_PARTIAL();              SCHECK_PARTIAL();
3745              MRRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
3746              }              }
3747            GETCHARINC(d, eptr);            GETCHARINC(d, eptr);
3748            if (fc == d) MRRETURN(MATCH_NOMATCH);            if (fc == d) RRETURN(MATCH_NOMATCH);
3749            }            }
3750          }          }
3751        else        else
3752  #endif  #endif
3753        /* Not UTF-8 mode */        /* Not UTF mode */
3754          {          {
3755          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
3756            {            {
3757            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
3758              {              {
3759              SCHECK_PARTIAL();              SCHECK_PARTIAL();
3760              MRRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
3761              }              }
3762            if (fc == *eptr++) MRRETURN(MATCH_NOMATCH);            if (fc == *eptr++) RRETURN(MATCH_NOMATCH);
3763            }            }
3764          }          }
3765    
# Line 3698  for (;;) Line 3767  for (;;)
3767    
3768        if (minimize)        if (minimize)
3769          {          {
3770  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
3771          /* UTF-8 mode */          if (utf)
         if (utf8)  
3772            {            {
3773            register unsigned int d;            register unsigned int d;
3774            for (fi = min;; fi++)            for (fi = min;; fi++)
3775              {              {
3776              RMATCH(eptr, ecode, offset_top, md, eptrb, RM32);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM32);
3777              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3778              if (fi >= max) MRRETURN(MATCH_NOMATCH);              if (fi >= max) RRETURN(MATCH_NOMATCH);
3779              if (eptr >= md->end_subject)              if (eptr >= md->end_subject)
3780                {                {
3781                SCHECK_PARTIAL();                SCHECK_PARTIAL();
3782                MRRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
3783                }                }
3784              GETCHARINC(d, eptr);              GETCHARINC(d, eptr);
3785              if (fc == d) MRRETURN(MATCH_NOMATCH);              if (fc == d) RRETURN(MATCH_NOMATCH);
3786              }              }
3787            }            }
3788          else          else
3789  #endif  #endif
3790          /* Not UTF-8 mode */          /* Not UTF mode */
3791            {            {
3792            for (fi = min;; fi++)            for (fi = min;; fi++)
3793              {              {
3794              RMATCH(eptr, ecode, offset_top, md, eptrb, RM33);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM33);
3795              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3796              if (fi >= max) MRRETURN(MATCH_NOMATCH);              if (fi >= max) RRETURN(MATCH_NOMATCH);
3797              if (eptr >= md->end_subject)              if (eptr >= md->end_subject)
3798                {                {
3799                SCHECK_PARTIAL();                SCHECK_PARTIAL();
3800                MRRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
3801                }                }
3802              if (fc == *eptr++) MRRETURN(MATCH_NOMATCH);              if (fc == *eptr++) RRETURN(MATCH_NOMATCH);
3803              }              }
3804            }            }
3805          /* Control never gets here */          /* Control never gets here */
# Line 3743  for (;;) Line 3811  for (;;)
3811          {          {
3812          pp = eptr;          pp = eptr;
3813    
3814  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
3815          /* UTF-8 mode */          if (utf)
         if (utf8)  
3816            {            {
3817            register unsigned int d;            register unsigned int d;
3818            for (i = min; i < max; i++)            for (i = min; i < max; i++)
# Line 3771  for (;;) Line 3838  for (;;)
3838            }            }
3839          else          else
3840  #endif  #endif
3841          /* Not UTF-8 mode */          /* Not UTF mode */
3842            {            {
3843            for (i = min; i < max; i++)            for (i = min; i < max; i++)
3844              {              {
# Line 3792  for (;;) Line 3859  for (;;)
3859              }              }
3860            }            }
3861    
3862          MRRETURN(MATCH_NOMATCH);          RRETURN(MATCH_NOMATCH);
3863          }          }
3864        }        }
3865      /* Control never gets here */      /* Control never gets here */
# Line 3804  for (;;) Line 3871  for (;;)
3871      case OP_TYPEEXACT:      case OP_TYPEEXACT:
3872      min = max = GET2(ecode, 1);      min = max = GET2(ecode, 1);
3873      minimize = TRUE;      minimize = TRUE;
3874      ecode += 3;      ecode += 1 + IMM2_SIZE;
3875      goto REPEATTYPE;      goto REPEATTYPE;
3876    
3877      case OP_TYPEUPTO:      case OP_TYPEUPTO:
# Line 3812  for (;;) Line 3879  for (;;)
3879      min = 0;      min = 0;
3880      max = GET2(ecode, 1);      max = GET2(ecode, 1);
3881      minimize = *ecode == OP_TYPEMINUPTO;      minimize = *ecode == OP_TYPEMINUPTO;
3882      ecode += 3;      ecode += 1 + IMM2_SIZE;
3883      goto REPEATTYPE;      goto REPEATTYPE;
3884    
3885      case OP_TYPEPOSSTAR:      case OP_TYPEPOSSTAR:
# Line 3840  for (;;) Line 3907  for (;;)
3907      possessive = TRUE;      possessive = TRUE;
3908      min = 0;      min = 0;
3909      max = GET2(ecode, 1);      max = GET2(ecode, 1);
3910      ecode += 3;      ecode += 1 + IMM2_SIZE;
3911      goto REPEATTYPE;      goto REPEATTYPE;
3912    
3913      case OP_TYPESTAR:      case OP_TYPESTAR:
# Line 3886  for (;;) Line 3953  for (;;)
3953          switch(prop_type)          switch(prop_type)
3954            {            {
3955            case PT_ANY:            case PT_ANY:
3956            if (prop_fail_result) MRRETURN(MATCH_NOMATCH);            if (prop_fail_result) RRETURN(MATCH_NOMATCH);
3957            for (i = 1; i <= min; i++)            for (i = 1; i <= min; i++)
3958              {              {
3959              if (eptr >= md->end_subject)              if (eptr >= md->end_subject)
3960                {                {
3961                SCHECK_PARTIAL();                SCHECK_PARTIAL();
3962                MRRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
3963                }                }
3964              GETCHARINCTEST(c, eptr);              GETCHARINCTEST(c, eptr);
3965              }              }
# Line 3905  for (;;) Line 3972  for (;;)
3972              if (eptr >= md->end_subject)              if (eptr >= md->end_subject)
3973                {                {
3974                SCHECK_PARTIAL();                SCHECK_PARTIAL();
3975                MRRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
3976                }                }
3977              GETCHARINCTEST(c, eptr);              GETCHARINCTEST(c, eptr);
3978              chartype = UCD_CHARTYPE(c);              chartype = UCD_CHARTYPE(c);
3979              if ((chartype == ucp_Lu ||              if ((chartype == ucp_Lu ||
3980                   chartype == ucp_Ll ||                   chartype == ucp_Ll ||
3981                   chartype == ucp_Lt) == prop_fail_result)                   chartype == ucp_Lt) == prop_fail_result)
3982                MRRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
3983              }              }
3984            break;            break;
3985    
# Line 3922  for (;;) Line 3989  for (;;)
3989              if (eptr >= md->end_subject)              if (eptr >= md->end_subject)
3990                {                {
3991                SCHECK_PARTIAL();                SCHECK_PARTIAL();
3992                MRRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
3993                }                }
3994              GETCHARINCTEST(c, eptr);              GETCHARINCTEST(c, eptr);
3995              if ((UCD_CATEGORY(c) == prop_value) == prop_fail_result)              if ((UCD_CATEGORY(c) == prop_value) == prop_fail_result)
3996                MRRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
3997              }              }
3998            break;            break;
3999    
# Line 3936  for (;;) Line 4003  for (;;)
4003              if (eptr >= md->end_subject)              if (eptr >= md->end_subject)
4004                {                {
4005                SCHECK_PARTIAL();                SCHECK_PARTIAL();
4006                MRRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
4007                }                }
4008              GETCHARINCTEST(c, eptr);              GETCHARINCTEST(c, eptr);
4009              if ((UCD_CHARTYPE(c) == prop_value) == prop_fail_result)              if ((UCD_CHARTYPE(c) == prop_value) == prop_fail_result)
4010                MRRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
4011              }              }
4012            break;            break;
4013    
# Line 3950  for (;;) Line 4017  for (;;)
4017              if (eptr >= md->end_subject)              if (eptr >= md->end_subject)
4018                {                {
4019                SCHECK_PARTIAL();                SCHECK_PARTIAL();
4020                MRRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
4021                }                }
4022              GETCHARINCTEST(c, eptr);              GETCHARINCTEST(c, eptr);
4023              if ((UCD_SCRIPT(c) == prop_value) == prop_fail_result)              if ((UCD_SCRIPT(c) == prop_value) == prop_fail_result)
4024                MRRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
4025              }              }
4026            break;            break;
4027    
# Line 3965  for (;;) Line 4032  for (;;)
4032              if (eptr >= md->end_subject)              if (eptr >= md->end_subject)
4033                {                {
4034                SCHECK_PARTIAL();                SCHECK_PARTIAL();
4035                MRRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
4036                }                }
4037              GETCHARINCTEST(c, eptr);              GETCHARINCTEST(c, eptr);
4038              category = UCD_CATEGORY(c);              category = UCD_CATEGORY(c);
4039              if ((category == ucp_L || category == ucp_N) == prop_fail_result)              if ((category == ucp_L || category == ucp_N) == prop_fail_result)
4040                MRRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
4041              }              }
4042            break;            break;
4043    
# Line 3980  for (;;) Line 4047  for (;;)
4047              if (eptr >= md->end_subject)              if (eptr >= md->end_subject)
4048                {                {
4049                SCHECK_PARTIAL();                SCHECK_PARTIAL();
4050                MRRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
4051                }                }
4052              GETCHARINCTEST(c, eptr);              GETCHARINCTEST(c, eptr);
4053              if ((UCD_CATEGORY(c) == ucp_Z || c == CHAR_HT || c == CHAR_NL ||              if ((UCD_CATEGORY(c) == ucp_Z || c == CHAR_HT || c == CHAR_NL ||
4054                   c == CHAR_FF || c == CHAR_CR)                   c == CHAR_FF || c == CHAR_CR)
4055                     == prop_fail_result)                     == prop_fail_result)
4056                MRRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
4057              }              }
4058            break;            break;
4059    
# Line 3996  for (;;) Line 4063  for (;;)
4063              if (eptr >= md->end_subject)              if (eptr >= md->end_subject)
4064                {                {
4065                SCHECK_PARTIAL();                SCHECK_PARTIAL();
4066                MRRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
4067                }                }
4068              GETCHARINCTEST(c, eptr);              GETCHARINCTEST(c, eptr);
4069              if ((UCD_CATEGORY(c) == ucp_Z || c == CHAR_HT || c == CHAR_NL ||              if ((UCD_CATEGORY(c) == ucp_Z || c == CHAR_HT || c == CHAR_NL ||
4070                   c == CHAR_VT || c == CHAR_FF || c == CHAR_CR)                   c == CHAR_VT || c == CHAR_FF || c == CHAR_CR)
4071                     == prop_fail_result)                     == prop_fail_result)
4072                MRRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
4073              }              }
4074            break;            break;
4075    
# Line 4013  for (;;) Line 4080  for (;;)
4080              if (eptr >= md->end_subject)              if (eptr >= md->end_subject)
4081                {                {
4082                SCHECK_PARTIAL();                SCHECK_PARTIAL();
4083                MRRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
4084                }                }
4085              GETCHARINCTEST(c, eptr);              GETCHARINCTEST(c, eptr);
4086              category = UCD_CATEGORY(c);              category = UCD_CATEGORY(c);
4087              if ((category == ucp_L || category == ucp_N || c == CHAR_UNDERSCORE)              if ((category == ucp_L || category == ucp_N || c == CHAR_UNDERSCORE)
4088                     == prop_fail_result)                     == prop_fail_result)
4089                MRRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
4090              }              }
4091            break;            break;
4092    
# Line 4040  for (;;) Line 4107  for (;;)
4107            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
4108              {              {
4109              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4110              MRRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4111              }              }
4112            GETCHARINCTEST(c, eptr);            GETCHARINCTEST(c, eptr);
4113            if (UCD_CATEGORY(c) == ucp_M) MRRETURN(MATCH_NOMATCH);            if (UCD_CATEGORY(c) == ucp_M) RRETURN(MATCH_NOMATCH);
4114            while (eptr < md->end_subject)            while (eptr < md->end_subject)
4115              {              {
4116              int len = 1;              int len = 1;
4117              if (!utf8) c = *eptr; else { GETCHARLEN(c, eptr, len); }              if (!utf) c = *eptr; else { GETCHARLEN(c, eptr, len); }
4118              if (UCD_CATEGORY(c) != ucp_M) break;              if (UCD_CATEGORY(c) != ucp_M) break;
4119              eptr += len;              eptr += len;
4120              }              }
# Line 4059  for (;;) Line 4126  for (;;)
4126    
4127  /* Handle all other cases when the coding is UTF-8 */  /* Handle all other cases when the coding is UTF-8 */
4128    
4129  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
4130        if (utf8) switch(ctype)        if (utf) switch(ctype)
4131          {          {
4132          case OP_ANY:          case OP_ANY:
4133          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
# Line 4068  for (;;) Line 4135  for (;;)
4135            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
4136              {              {
4137              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4138              MRRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4139              }              }
4140            if (IS_NEWLINE(eptr)) MRRETURN(MATCH_NOMATCH);            if (IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH);
4141            eptr++;            eptr++;
4142            while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;            ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);
4143            }            }
4144          break;          break;
4145    
# Line 4082  for (;;) Line 4149  for (;;)
4149            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
4150              {              {
4151              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4152              MRRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4153              }              }
4154            eptr++;            eptr++;
4155            while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;            ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);
4156            }            }
4157          break;          break;
4158    
4159          case OP_ANYBYTE:          case OP_ANYBYTE:
4160          if (eptr > md->end_subject - min) MRRETURN(MATCH_NOMATCH);          if (eptr > md->end_subject - min) RRETURN(MATCH_NOMATCH);
4161          eptr += min;          eptr += min;
4162          break;          break;
4163    
# Line 4100  for (;;) Line 4167  for (;;)
4167            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
4168              {              {
4169              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4170              MRRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4171              }              }
4172            GETCHARINC(c, eptr);            GETCHARINC(c, eptr);
4173            switch(c)            switch(c)
4174              {              {
4175              default: MRRETURN(MATCH_NOMATCH);              default: RRETURN(MATCH_NOMATCH);
4176    
4177              case 0x000d:              case 0x000d:
4178              if (eptr < md->end_subject && *eptr == 0x0a) eptr++;              if (eptr < md->end_subject && *eptr == 0x0a) eptr++;
# Line 4119  for (;;) Line 4186  for (;;)
4186              case 0x0085:              case 0x0085:
4187              case 0x2028:              case 0x2028:
4188              case 0x2029:              case 0x2029:
4189              if (md->bsr_anycrlf) MRRETURN(MATCH_NOMATCH);              if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);
4190              break;              break;
4191              }              }
4192            }            }
# Line 4131  for (;;) Line 4198  for (;;)
4198            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
4199              {              {
4200              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4201              MRRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4202              }              }
4203            GETCHARINC(c, eptr);            GETCHARINC(c, eptr);
4204            switch(c)            switch(c)
# Line 4156  for (;;) Line 4223  for (;;)
4223              case 0x202f:    /* NARROW NO-BREAK SPACE */              case 0x202f:    /* NARROW NO-BREAK SPACE */
4224              case 0x205f:    /* MEDIUM MATHEMATICAL SPACE */              case 0x205f:    /* MEDIUM MATHEMATICAL SPACE */
4225              case 0x3000:    /* IDEOGRAPHIC SPACE */              case 0x3000:    /* IDEOGRAPHIC SPACE */
4226              MRRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4227              }              }
4228            }            }
4229          break;          break;
# Line 4167  for (;;) Line 4234  for (;;)
4234            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
4235              {              {
4236              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4237              MRRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4238              }              }
4239            GETCHARINC(c, eptr);            GETCHARINC(c, eptr);
4240            switch(c)            switch(c)
4241              {              {
4242              default: MRRETURN(MATCH_NOMATCH);              default: RRETURN(MATCH_NOMATCH);
4243              case 0x09:      /* HT */              case 0x09:      /* HT */
4244              case 0x20:      /* SPACE */              case 0x20:      /* SPACE */
4245              case 0xa0:      /* NBSP */              case 0xa0:      /* NBSP */
# Line 4203  for (;;) Line 4270  for (;;)
4270            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
4271              {              {
4272              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4273              MRRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4274              }              }
4275            GETCHARINC(c, eptr);            GETCHARINC(c, eptr);
4276            switch(c)            switch(c)
# Line 4216  for (;;) Line 4283  for (;;)
4283              case 0x85:      /* NEL */              case 0x85:      /* NEL */
4284              case 0x2028:    /* LINE SEPARATOR */              case 0x2028:    /* LINE SEPARATOR */
4285              case 0x2029:    /* PARAGRAPH SEPARATOR */              case 0x2029:    /* PARAGRAPH SEPARATOR */
4286              MRRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4287              }              }
4288            }            }
4289          break;          break;
# Line 4227  for (;;) Line 4294  for (;;)
4294            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
4295              {              {
4296              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4297              MRRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4298              }              }
4299            GETCHARINC(c, eptr);            GETCHARINC(c, eptr);
4300            switch(c)            switch(c)
4301              {              {
4302              default: MRRETURN(MATCH_NOMATCH);              default: RRETURN(MATCH_NOMATCH);
4303              case 0x0a:      /* LF */              case 0x0a:      /* LF */
4304              case 0x0b:      /* VT */              case 0x0b:      /* VT */
4305              case 0x0c:      /* FF */              case 0x0c:      /* FF */
# Line 4251  for (;;) Line 4318  for (;;)
4318            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
4319              {              {
4320              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4321              MRRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4322              }              }
4323            GETCHARINC(c, eptr);            GETCHARINC(c, eptr);
4324            if (c < 128 && (md->ctypes[c] & ctype_digit) != 0)            if (c < 128 && (md->ctypes[c] & ctype_digit) != 0)
4325              MRRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4326            }            }
4327          break;          break;
4328    
# Line 4265  for (;;) Line 4332  for (;;)
4332            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
4333              {              {
4334              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4335              MRRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4336              }              }
4337            if (*eptr >= 128 || (md->ctypes[*eptr++] & ctype_digit) == 0)            if (*eptr >= 128 || (md->ctypes[*eptr] & ctype_digit) == 0)
4338              MRRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4339              eptr++;
4340            /* 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 */
4341            }            }
4342          break;          break;
# Line 4279  for (;;) Line 4347  for (;;)
4347            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
4348              {              {
4349              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4350              MRRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4351              }              }
4352            if (*eptr < 128 && (md->ctypes[*eptr] & ctype_space) != 0)            if (*eptr < 128 && (md->ctypes[*eptr] & ctype_space) != 0)
4353              MRRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4354            while (++eptr < md->end_subject && (*eptr & 0xc0) == 0x80);            eptr++;
4355              ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);
4356            }            }
4357          break;          break;
4358    
# Line 4293  for (;;) Line 4362  for (;;)
4362            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
4363              {              {
4364              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4365              MRRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4366              }              }
4367            if (*eptr >= 128 || (md->ctypes[*eptr++] & ctype_space) == 0)            if (*eptr >= 128 || (md->ctypes[*eptr] & ctype_space) == 0)
4368              MRRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4369              eptr++;
4370            /* 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 */
4371            }            }
4372          break;          break;
# Line 4307  for (;;) Line 4377  for (;;)
4377            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
4378              {              {
4379              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4380              MRRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4381              }              }
4382            if (*eptr < 128 && (md->ctypes[*eptr] & ctype_word) != 0)            if (*eptr < 128 && (md->ctypes[*eptr] & ctype_word) != 0)
4383              MRRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4384            while (++eptr < md->end_subject && (*eptr & 0xc0) == 0x80);            eptr++;
4385              ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);
4386            }            }
4387          break;          break;
4388    
# Line 4321  for (;;) Line 4392  for (;;)
4392            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
4393              {              {
4394              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4395              MRRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4396              }              }
4397            if (*eptr >= 128 || (md->ctypes[*eptr++] & ctype_word) == 0)            if (*eptr >= 128 || (md->ctypes[*eptr] & ctype_word) == 0)
4398              MRRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4399              eptr++;
4400            /* 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 */
4401            }            }
4402          break;          break;
# Line 4334  for (;;) Line 4406  for (;;)
4406          }  /* End switch(ctype) */          }  /* End switch(ctype) */
4407    
4408        else        else
4409  #endif     /* SUPPORT_UTF8 */  #endif     /* SUPPORT_UTF */
4410    
4411        /* 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
4412        than OP_PROP and OP_NOTPROP. */        than OP_PROP and OP_NOTPROP. */
# Line 4347  for (;;) Line 4419  for (;;)
4419            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
4420              {              {
4421              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4422              MRRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4423              }              }
4424            if (IS_NEWLINE(eptr)) MRRETURN(MATCH_NOMATCH);            if (IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH);
4425            eptr++;            eptr++;
4426            }            }
4427          break;          break;
# Line 4358  for (;;) Line 4430  for (;;)
4430          if (eptr > md->end_subject - min)          if (eptr > md->end_subject - min)
4431            {            {
4432            SCHECK_PARTIAL();            SCHECK_PARTIAL();
4433            MRRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
4434            }            }
4435          eptr += min;          eptr += min;
4436          break;          break;
# Line 4367  for (;;) Line 4439  for (;;)
4439          if (eptr > md->end_subject - min)          if (eptr > md->end_subject - min)
4440            {            {
4441            SCHECK_PARTIAL();            SCHECK_PARTIAL();
4442            MRRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
4443            }            }
4444          eptr += min;          eptr += min;
4445          break;          break;
# Line 4378  for (;;) Line 4450  for (;;)
4450            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
4451              {              {
4452              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4453              MRRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4454              }              }
4455            switch(*eptr++)            switch(*eptr++)
4456              {              {
4457              default: MRRETURN(MATCH_NOMATCH);              default: RRETURN(MATCH_NOMATCH);
4458    
4459              case 0x000d:              case 0x000d:
4460              if (eptr < md->end_subject && *eptr == 0x0a) eptr++;              if (eptr < md->end_subject && *eptr == 0x0a) eptr++;
# Line 4394  for (;;) Line 4466  for (;;)
4466              case 0x000b:              case 0x000b:
4467              case 0x000c:              case 0x000c:
4468              case 0x0085:              case 0x0085:
4469              if (md->bsr_anycrlf) MRRETURN(MATCH_NOMATCH);              if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);
4470              break;              break;
4471              }              }
4472            }            }
# Line 4406  for (;;) Line 4478  for (;;)
4478            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
4479              {              {
4480              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4481              MRRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4482              }              }
4483            switch(*eptr++)            switch(*eptr++)
4484              {              {
# Line 4414  for (;;) Line 4486  for (;;)
4486              case 0x09:      /* HT */              case 0x09:      /* HT */
4487              case 0x20:      /* SPACE */              case 0x20:      /* SPACE */
4488              case 0xa0:      /* NBSP */              case 0xa0:      /* NBSP */
4489              MRRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4490              }              }
4491            }            }
4492          break;          break;
# Line 4425  for (;;) Line 4497  for (;;)
4497            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
4498              {              {
4499              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4500              MRRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4501              }              }
4502            switch(*eptr++)            switch(*eptr++)
4503              {              {
4504              default: MRRETURN(MATCH_NOMATCH);              default: RRETURN(MATCH_NOMATCH);
4505              case 0x09:      /* HT */              case 0x09:      /* HT */
4506              case 0x20:      /* SPACE */              case 0x20:      /* SPACE */
4507              case 0xa0:      /* NBSP */              case 0xa0:      /* NBSP */
# Line 4444  for (;;) Line 4516  for (;;)
4516            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
4517              {              {
4518              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4519              MRRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4520              }              }
4521            switch(*eptr++)            switch(*eptr++)
4522              {              {
# Line 4454  for (;;) Line 4526  for (;;)
4526              case 0x0c:      /* FF */              case 0x0c:      /* FF */
4527              case 0x0d:      /* CR */              case 0x0d:      /* CR */
4528              case 0x85:      /* NEL */              case 0x85:      /* NEL */
4529              MRRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4530              }              }
4531            }            }
4532          break;          break;
# Line 4465  for (;;) Line 4537  for (;;)
4537            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
4538              {              {
4539              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4540              MRRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4541              }              }
4542            switch(*eptr++)            switch(*eptr++)
4543              {              {
4544              default: MRRETURN(MATCH_NOMATCH);              default: RRETURN(MATCH_NOMATCH);
4545              case 0x0a:      /* LF */              case 0x0a:      /* LF */
4546              case 0x0b:      /* VT */              case 0x0b:      /* VT */
4547              case 0x0c:      /* FF */              case 0x0c:      /* FF */
# Line 4486  for (;;) Line 4558  for (;;)
4558            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
4559              {              {
4560              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4561              MRRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4562              }              }
4563            if ((md->ctypes[*eptr++] & ctype_digit) != 0) MRRETURN(MATCH_NOMATCH);            if (MAX_255(*eptr) && (md->ctypes[*eptr] & ctype_digit) != 0)
4564                RRETURN(MATCH_NOMATCH);
4565              eptr++;
4566            }            }
4567          break;          break;
4568    
# Line 4498  for (;;) Line 4572  for (;;)
4572            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
4573              {              {
4574              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4575              MRRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4576              }              }
4577            if ((md->ctypes[*eptr++] & ctype_digit) == 0) MRRETURN(MATCH_NOMATCH);            if (!MAX_255(*eptr) || (md->ctypes[*eptr] & ctype_digit) == 0)
4578                RRETURN(MATCH_NOMATCH);
4579              eptr++;
4580            }            }
4581          break;          break;
4582    
# Line 4510  for (;;) Line 4586  for (;;)
4586            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
4587              {              {
4588              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4589              MRRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4590              }              }
4591            if ((md->ctypes[*eptr++] & ctype_space) != 0) MRRETURN(MATCH_NOMATCH);            if (MAX_255(*eptr) && (md->ctypes[*eptr] & ctype_space) != 0)
4592                RRETURN(MATCH_NOMATCH);
4593              eptr++;
4594            }            }
4595          break;          break;
4596    
# Line 4522  for (;;) Line 4600  for (;;)
4600            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
4601              {              {
4602              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4603              MRRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4604              }              }
4605            if ((md->ctypes[*eptr++] & ctype_space) == 0) MRRETURN(MATCH_NOMATCH);            if (!MAX_255(*eptr) || (md->ctypes[*eptr] & ctype_space) == 0)
4606                RRETURN(MATCH_NOMATCH);
4607              eptr++;
4608            }            }
4609          break;          break;
4610    
# Line 4534  for (;;) Line 4614  for (;;)
4614            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
4615              {              {
4616              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4617              MRRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4618              }              }
4619            if ((md->ctypes[*eptr++] & ctype_word) != 0)            if (MAX_255(*eptr) && (md->ctypes[*eptr] & ctype_word) != 0)
4620              MRRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4621              eptr++;
4622            }            }
4623          break;          break;
4624    
# Line 4547  for (;;) Line 4628  for (;;)
4628            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
4629              {              {
4630              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4631              MRRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4632              }              }
4633            if ((md->ctypes[*eptr++] & ctype_word) == 0)            if (!MAX_255(*eptr) || (md->ctypes[*eptr] & ctype_word) == 0)
4634              MRRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4635              eptr++;
4636            }            }
4637          break;          break;
4638    
# Line 4579  for (;;) Line 4661  for (;;)
4661              {              {
4662              RMATCH(eptr, ecode, offset_top, md, eptrb, RM36);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM36);
4663              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
4664              if (fi >= max) MRRETURN(MATCH_NOMATCH);              if (fi >= max) RRETURN(MATCH_NOMATCH);
4665              if (eptr >= md->end_subject)              if (eptr >= md->end_subject)
4666                {                {
4667                SCHECK_PARTIAL();                SCHECK_PARTIAL();
4668                MRRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
4669                }                }
4670              GETCHARINCTEST(c, eptr);              GETCHARINCTEST(c, eptr);
4671              if (prop_fail_result) MRRETURN(MATCH_NOMATCH);              if (prop_fail_result) RRETURN(MATCH_NOMATCH);
4672              }              }
4673            /* Control never gets here */            /* Control never gets here */
4674    
# Line 4596  for (;;) Line 4678  for (;;)
4678              int chartype;              int chartype;
4679              RMATCH(eptr, ecode, offset_top, md, eptrb, RM37);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM37);
4680              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
4681              if (fi >= max) MRRETURN(MATCH_NOMATCH);              if (fi >= max) RRETURN(MATCH_NOMATCH);
4682              if (eptr >= md->end_subject)              if (eptr >= md->end_subject)
4683                {                {
4684                SCHECK_PARTIAL();                SCHECK_PARTIAL();
4685                MRRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
4686                }                }
4687              GETCHARINCTEST(c, eptr);              GETCHARINCTEST(c, eptr);
4688              chartype = UCD_CHARTYPE(c);              chartype = UCD_CHARTYPE(c);
4689              if ((chartype == ucp_Lu ||              if ((chartype == ucp_Lu ||
4690                   chartype == ucp_Ll ||                   chartype == ucp_Ll ||
4691                   chartype == ucp_Lt) == prop_fail_result)                   chartype == ucp_Lt) == prop_fail_result)
4692                MRRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
4693              }              }
4694            /* Control never gets here */            /* Control never gets here */
4695    
# Line 4616  for (;;) Line 4698  for (;;)
4698              {              {
4699              RMATCH(eptr, ecode, offset_top, md, eptrb, RM38);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM38);
4700              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
4701              if (fi >= max) MRRETURN(MATCH_NOMATCH);              if (fi >= max) RRETURN(MATCH_NOMATCH);
4702              if (eptr >= md->end_subject)              if (eptr >= md->end_subject)
4703                {                {
4704                SCHECK_PARTIAL();                SCHECK_PARTIAL();
4705                MRRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
4706                }                }
4707              GETCHARINCTEST(c, eptr);              GETCHARINCTEST(c, eptr);
4708              if ((UCD_CATEGORY(c) == prop_value) == prop_fail_result)              if ((UCD_CATEGORY(c) == prop_value) == prop_fail_result)
4709                MRRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
4710              }              }
4711            /* Control never gets here */            /* Control never gets here */
4712    
# Line 4633  for (;;) Line 4715  for (;;)
4715              {              {
4716              RMATCH(eptr, ecode, offset_top, md, eptrb, RM39);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM39);
4717              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
4718              if (fi >= max) MRRETURN(MATCH_NOMATCH);              if (fi >= max) RRETURN(MATCH_NOMATCH);
4719              if (eptr >= md->end_subject)              if (eptr >= md->end_subject)
4720                {                {
4721                SCHECK_PARTIAL();                SCHECK_PARTIAL();
4722                MRRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
4723                }                }
4724              GETCHARINCTEST(c, eptr);              GETCHARINCTEST(c, eptr);
4725              if ((UCD_CHARTYPE(c) == prop_value) == prop_fail_result)              if ((UCD_CHARTYPE(c) == prop_value) == prop_fail_result)
4726                MRRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
4727              }              }
4728            /* Control never gets here */            /* Control never gets here */
4729    
# Line 4650  for (;;) Line 4732  for (;;)
4732              {              {
4733              RMATCH(eptr, ecode, offset_top, md, eptrb, RM40);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM40);
4734              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
4735              if (fi >= max) MRRETURN(MATCH_NOMATCH);              if (fi >= max) RRETURN(MATCH_NOMATCH);
4736              if (eptr >= md->end_subject)              if (eptr >= md->end_subject)
4737                {                {
4738                SCHECK_PARTIAL();                SCHECK_PARTIAL();
4739                MRRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
4740                }                }
4741              GETCHARINCTEST(c, eptr);              GETCHARINCTEST(c, eptr);
4742              if ((UCD_SCRIPT(c) == prop_value) == prop_fail_result)              if ((UCD_SCRIPT(c) == prop_value) == prop_fail_result)
4743                MRRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
4744              }              }
4745            /* Control never gets here */            /* Control never gets here */
4746    
# Line 4668  for (;;) Line 4750  for (;;)
4750              int category;              int category;
4751              RMATCH(eptr, ecode, offset_top, md, eptrb, RM59);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM59);
4752              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
4753              if (fi >= max) MRRETURN(MATCH_NOMATCH);              if (fi >= max) RRETURN(MATCH_NOMATCH);
4754              if (eptr >= md->end_subject)              if (eptr >= md->end_subject)
4755                {                {
4756                SCHECK_PARTIAL();                SCHECK_PARTIAL();
4757                MRRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
4758                }                }
4759              GETCHARINCTEST(c, eptr);              GETCHARINCTEST(c, eptr);
4760              category = UCD_CATEGORY(c);              category = UCD_CATEGORY(c);
4761              if ((category == ucp_L || category == ucp_N) == prop_fail_result)              if ((category == ucp_L || category == ucp_N) == prop_fail_result)
4762                MRRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
4763              }              }
4764            /* Control never gets here */            /* Control never gets here */
4765    
# Line 4686  for (;;) Line 4768  for (;;)
4768              {              {
4769              RMATCH(eptr, ecode, offset_top, md, eptrb, RM60);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM60);
4770              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
4771              if (fi >= max) MRRETURN(MATCH_NOMATCH);              if (fi >= max) RRETURN(MATCH_NOMATCH);
4772              if (eptr >= md->end_subject)              if (eptr >= md->end_subject)
4773                {                {
4774                SCHECK_PARTIAL();                SCHECK_PARTIAL();
4775                MRRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
4776                }                }
4777              GETCHARINCTEST(c, eptr);              GETCHARINCTEST(c, eptr);
4778              if ((UCD_CATEGORY(c) == ucp_Z || c == CHAR_HT || c == CHAR_NL ||              if ((UCD_CATEGORY(c) == ucp_Z || c == CHAR_HT || c == CHAR_NL ||
4779                   c == CHAR_FF || c == CHAR_CR)                   c == CHAR_FF || c == CHAR_CR)
4780                     == prop_fail_result)                     == prop_fail_result)
4781                MRRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
4782              }              }
4783            /* Control never gets here */            /* Control never gets here */
4784    
# Line 4705  for (;;) Line 4787  for (;;)
4787              {              {
4788              RMATCH(eptr, ecode, offset_top, md, eptrb, RM61);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM61);
4789              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
4790              if (fi >= max) MRRETURN(MATCH_NOMATCH);              if (fi >= max) RRETURN(MATCH_NOMATCH);
4791              if (eptr >= md->end_subject)              if (eptr >= md->end_subject)
4792                {                {
4793                SCHECK_PARTIAL();                SCHECK_PARTIAL();
4794                MRRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
4795                }                }
4796              GETCHARINCTEST(c, eptr);              GETCHARINCTEST(c, eptr);
4797              if ((UCD_CATEGORY(c) == ucp_Z || c == CHAR_HT || c == CHAR_NL ||              if ((UCD_CATEGORY(c) == ucp_Z || c == CHAR_HT || c == CHAR_NL ||
4798                   c == CHAR_VT || c == CHAR_FF || c == CHAR_CR)                   c == CHAR_VT || c == CHAR_FF || c == CHAR_CR)
4799                     == prop_fail_result)                     == prop_fail_result)
4800                MRRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
4801              }              }
4802            /* Control never gets here */            /* Control never gets here */
4803    
# Line 4725  for (;;) Line 4807  for (;;)
4807              int category;              int category;
4808              RMATCH(eptr, ecode, offset_top, md, eptrb, RM62);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM62);
4809              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
4810              if (fi >= max) MRRETURN(MATCH_NOMATCH);              if (fi >= max) RRETURN(MATCH_NOMATCH);
4811              if (eptr >= md->end_subject)              if (eptr >= md->end_subject)
4812                {                {
4813                SCHECK_PARTIAL();                SCHECK_PARTIAL();
4814                MRRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
4815                }                }
4816              GETCHARINCTEST(c, eptr);              GETCHARINCTEST(c, eptr);
4817              category = UCD_CATEGORY(c);              category = UCD_CATEGORY(c);
# Line 4737  for (;;) Line 4819  for (;;)
4819                   category == ucp_N ||                   category == ucp_N ||
4820                   c == CHAR_UNDERSCORE)                   c == CHAR_UNDERSCORE)
4821                     == prop_fail_result)                     == prop_fail_result)
4822                MRRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
4823              }              }
4824            /* Control never gets here */            /* Control never gets here */
4825    
# Line 4757  for (;;) Line 4839  for (;;)
4839            {            {
4840            RMATCH(eptr, ecode, offset_top, md, eptrb, RM41);            RMATCH(eptr, ecode, offset_top, md, eptrb, RM41);
4841            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
4842            if (fi >= max) MRRETURN(MATCH_NOMATCH);            if (fi >= max) RRETURN(MATCH_NOMATCH);
4843            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
4844              {              {
4845              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4846              MRRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4847              }              }
4848            GETCHARINCTEST(c, eptr);            GETCHARINCTEST(c, eptr);
4849            if (UCD_CATEGORY(c) == ucp_M) MRRETURN(MATCH_NOMATCH);            if (UCD_CATEGORY(c) == ucp_M) RRETURN(MATCH_NOMATCH);
4850            while (eptr < md->end_subject)            while (eptr < md->end_subject)
4851              {              {
4852              int len = 1;              int len = 1;
4853              if (!utf8) c = *eptr; else { GETCHARLEN(c, eptr, len); }              if (!utf) c = *eptr; else { GETCHARLEN(c, eptr, len); }
4854              if (UCD_CATEGORY(c) != ucp_M) break;              if (UCD_CATEGORY(c) != ucp_M) break;
4855              eptr += len;              eptr += len;
4856              }              }
# Line 4777  for (;;) Line 4859  for (;;)
4859        else        else
4860  #endif     /* SUPPORT_UCP */  #endif     /* SUPPORT_UCP */
4861    
4862  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
4863        /* UTF-8 mode */        if (utf)
       if (utf8)  
4864          {          {
4865          for (fi = min;; fi++)          for (fi = min;; fi++)
4866            {            {
4867            RMATCH(eptr, ecode, offset_top, md, eptrb, RM42);            RMATCH(eptr, ecode, offset_top, md, eptrb, RM42);
4868            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
4869            if (fi >= max) MRRETURN(MATCH_NOMATCH);            if (fi >= max) RRETURN(MATCH_NOMATCH);
4870            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
4871              {              {
4872              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4873              MRRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4874              }              }
4875            if (ctype == OP_ANY && IS_NEWLINE(eptr))            if (ctype == OP_ANY && IS_NEWLINE(eptr))
4876              MRRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4877            GETCHARINC(c, eptr);            GETCHARINC(c, eptr);
4878            switch(ctype)            switch(ctype)
4879              {              {
# Line 4804  for (;;) Line 4885  for (;;)
4885              case OP_ANYNL:              case OP_ANYNL:
4886              switch(c)              switch(c)
4887                {                {
4888                default: MRRETURN(MATCH_NOMATCH);                default: RRETURN(MATCH_NOMATCH);
4889                case 0x000d:                case 0x000d:
4890                if (eptr < md->end_subject && *eptr == 0x0a) eptr++;                if (eptr < md->end_subject && *eptr == 0x0a) eptr++;
4891                break;                break;
# Line 4816  for (;;) Line 4897  for (;;)
4897                case 0x0085:                case 0x0085:
4898                case 0x2028:                case 0x2028:
4899                case 0x2029:                case 0x2029:
4900                if (md->bsr_anycrlf) MRRETURN(MATCH_NOMATCH);                if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);
4901                break;                break;
4902                }                }
4903              break;              break;
# Line 4844  for (;;) Line 4925  for (;;)
4925                case 0x202f:    /* NARROW NO-BREAK SPACE */                case 0x202f:    /* NARROW NO-BREAK SPACE */
4926                case 0x205f:    /* MEDIUM MATHEMATICAL SPACE */                case 0x205f:    /* MEDIUM MATHEMATICAL SPACE */
4927                case 0x3000:    /* IDEOGRAPHIC SPACE */                case 0x3000:    /* IDEOGRAPHIC SPACE */
4928                MRRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
4929                }                }
4930              break;              break;
4931    
4932              case OP_HSPACE:              case OP_HSPACE:
4933              switch(c)              switch(c)
4934                {                {
4935                default: MRRETURN(MATCH_NOMATCH);                default: RRETURN(MATCH_NOMATCH);
4936                case 0x09:      /* HT */                case 0x09:      /* HT */
4937                case 0x20:      /* SPACE */                case 0x20:      /* SPACE */
4938                case 0xa0:      /* NBSP */                case 0xa0:      /* NBSP */
# Line 4886  for (;;) Line 4967  for (;;)
4967                case 0x85:      /* NEL */                case 0x85:      /* NEL */
4968                case 0x2028:    /* LINE SEPARATOR */                case 0x2028:    /* LINE SEPARATOR */
4969                case 0x2029:    /* PARAGRAPH SEPARATOR */                case 0x2029:    /* PARAGRAPH SEPARATOR */
4970                MRRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
4971                }                }
4972              break;              break;
4973    
4974              case OP_VSPACE:              case OP_VSPACE:
4975              switch(c)              switch(c)
4976                {                {
4977                default: MRRETURN(MATCH_NOMATCH);                default: RRETURN(MATCH_NOMATCH);
4978                case 0x0a:      /* LF */                case 0x0a:      /* LF */
4979                case 0x0b:      /* VT */                case 0x0b:      /* VT */
4980                case 0x0c:      /* FF */                case 0x0c:      /* FF */
# Line 4907  for (;;) Line 4988  for (;;)
4988    
4989              case OP_NOT_DIGIT:              case OP_NOT_DIGIT:
4990              if (c < 256 && (md->ctypes[c] & ctype_digit) != 0)              if (c < 256 && (md->ctypes[c] & ctype_digit) != 0)
4991                MRRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
4992              break;              break;
4993    
4994              case OP_DIGIT:              case OP_DIGIT:
4995              if (c >= 256 || (md->ctypes[c] & ctype_digit) == 0)              if (c >= 256 || (md->ctypes[c] & ctype_digit) == 0)
4996                MRRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
4997              break;              break;
4998    
4999              case OP_NOT_WHITESPACE:              case OP_NOT_WHITESPACE:
5000              if (c < 256 && (md->ctypes[c] & ctype_space) != 0)              if (c < 256 && (md->ctypes[c] & ctype_space) != 0)
5001                MRRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
5002              break;              break;
5003    
5004              case OP_WHITESPACE:              case OP_WHITESPACE:
5005              if  (c >= 256 || (md->ctypes[c] & ctype_space) == 0)              if (c >= 256 || (md->ctypes[c] & ctype_space) == 0)
5006                MRRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
5007              break;              break;
5008    
5009              case OP_NOT_WORDCHAR:              case OP_NOT_WORDCHAR:
5010              if (c < 256 && (md->ctypes[c] & ctype_word) != 0)              if (c < 256 && (md->ctypes[c] & ctype_word) != 0)
5011                MRRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
5012              break;              break;
5013    
5014              case OP_WORDCHAR:              case OP_WORDCHAR:
5015              if (c >= 256 || (md->ctypes[c] & ctype_word) == 0)              if (c >= 256 || (md->ctypes[c] & ctype_word) == 0)
5016                MRRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
5017              break;              break;
5018    
5019              default:              default:
# Line 4942  for (;;) Line 5023  for (;;)
5023          }          }
5024        else        else
5025  #endif  #endif
5026        /* Not UTF-8 mode */        /* Not UTF mode */
5027          {          {
5028          for (fi = min;; fi++)          for (fi = min;; fi++)
5029            {            {
5030            RMATCH(eptr, ecode, offset_top, md, eptrb, RM43);            RMATCH(eptr, ecode, offset_top, md, eptrb, RM43);
5031            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
5032            if (fi >= max) MRRETURN(MATCH_NOMATCH);            if (fi >= max) RRETURN(MATCH_NOMATCH);
5033            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
5034              {              {
5035              SCHECK_PARTIAL();              SCHECK_PARTIAL();
5036              MRRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
5037              }              }
5038            if (ctype == OP_ANY && IS_NEWLINE(eptr))            if (ctype == OP_ANY && IS_NEWLINE(eptr))
5039              MRRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
5040            c = *eptr++;            c = *eptr++;
5041            switch(ctype)            switch(ctype)
5042              {              {
# Line 4967  for (;;) Line 5048  for (;;)
5048              case OP_ANYNL:              case OP_ANYNL:
5049              switch(c)              switch(c)
5050                {                {
5051                default: MRRETURN(MATCH_NOMATCH);                de