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

Diff of /code/trunk/pcre_exec.c

Parent Directory | Revision Log | View Patch Patch

-revision 835 by ph10,
Wed Dec 28 16:10:09 2011 UTC
+revision 1092 by chpe,
Tue Oct 16 15:55:55 2012 UTC
 Line 6
  and semantics are as close as possible to those of the Perl 5 language.
                         Written by Philip Hazel
-            Copyright (c) 1997-2011 University of Cambridge
+            Copyright (c) 1997-2012 University of Cambridge
  -----------------------------------------------------------------------------
  Redistribution and use in source and binary forms, with or without
 Line 37 
 POSSIBILITY OF SUCH DAMAGE.
  -----------------------------------------------------------------------------
  */
  /* This module contains pcre_exec(), the externally visible function that does
  pattern matching using an NFA algorithm, trying to mimic Perl as closely as
  possible. There are also some static supporting functions. */
-Line 82 
 negative to avoid the external error cod
+Line 81 
 negative to avoid the external error cod
  #define MATCH_SKIP_ARG     (-993)
  #define MATCH_THEN         (-992)
- /* This is a convenience macro for code that occurs many times. */
- #define MRRETURN(ra) \
-   { \
-   md->mark = markptr; \
-   RRETURN(ra); \
-   }
  /* Maximum number of ints of offset to save on the stack for recursive calls.
  If the offset vector is bigger, malloc is used. This should be a multiple of 3,
  because the offset vector is always a multiple of 3 long. */
-Line 121 
 Returns:     nothing
+Line 112 
 Returns:     nothing
  */
  static void
- 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)
  {
  unsigned int c;
  if (is_subject && length > md->end_subject - p) length = md->end_subject - p;
-Line 148 
 Arguments:
+Line 139 
 Arguments:
    md          points to match data block
    caseless    TRUE if caseless
- Returns:      < 0 if not matched, otherwise the number of subject bytes matched
+ Returns:      >= 0 the number of subject bytes matched
+               -1 no match
+               -2 partial match; always given if at end subject
  */
  static int
- 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,
    BOOL caseless)
  {
- USPTR eptr_start = eptr;
+ PCRE_PUCHAR eptr_start = eptr;
- register USPTR p = md->start_subject + md->offset_vector[offset];
+ register PCRE_PUCHAR p = md->start_subject + md->offset_vector[offset];
  #ifdef PCRE_DEBUG
  if (eptr >= md->end_subject)
-Line 171 
 pchars(p, length, FALSE, md);
+Line 164 
 pchars(p, length, FALSE, md);
  printf("\n");
  #endif
- /* Always fail if reference not set (and not JavaScript compatible). */
+ /* Always fail if reference not set (and not JavaScript compatible - in that
+ case the length is passed as zero). */
  if (length < 0) return -1;
-Line 181 
 ASCII characters. */
+Line 175 
 ASCII characters. */
  if (caseless)
    {
- #ifdef SUPPORT_UTF8
+ #ifdef SUPPORT_UTF
  #ifdef SUPPORT_UCP
-   if (md->utf8)
+   if (md->utf)
      {
      /* Match characters up to the end of the reference. NOTE: the number of
-     bytes matched may differ, because there are some characters whose upper and
+     data units matched may differ, because in UTF-8 there are some characters
-     lower case versions code as different numbers of bytes. For example, U+023A
+     whose upper and lower case versions code have different numbers of bytes.
-     (2 bytes in UTF-8) is the upper case version of U+2C65 (3 bytes in UTF-8);
+     For example, U+023A (2 bytes in UTF-8) is the upper case version of U+2C65
-     a sequence of 3 of the former uses 6 bytes, as does a sequence of two of
+     (3 bytes in UTF-8); a sequence of 3 of the former uses 6 bytes, as does a
-     the latter. It is important, therefore, to check the length along the
+     sequence of two of the latter. It is important, therefore, to check the
-     reference, not along the subject (earlier code did this wrong). */
+     length along the reference, not along the subject (earlier code did this
+     wrong). */
-     USPTR endptr = p + length;
+     PCRE_PUCHAR endptr = p + length;
      while (p < endptr)
        {
-       int c, d;
+       pcre_uint32 c, d;
-       if (eptr >= md->end_subject) return -1;
+       const ucd_record *ur;
+       if (eptr >= md->end_subject) return -2;   /* Partial match */
        GETCHARINC(c, eptr);
        GETCHARINC(d, p);
-       if (c != d && c != UCD_OTHERCASE(d)) return -1;
+       ur = GET_UCD(d);
+       if (c != d && c != d + ur->other_case)
+         {
+         const pcre_uint32 *pp = PRIV(ucd_caseless_sets) + ur->caseset;
+         for (;;)
+           {
+           if (c < *pp) return -1;
+           if (c == *pp++) break;
+           }
+         }
        }
      }
    else
-Line 210 
 if (caseless)
+Line 215 
 if (caseless)
    /* The same code works when not in UTF-8 mode and in UTF-8 mode when there
    is no UCP support. */
      {
-     if (eptr + length > md->end_subject) return -1;
      while (length-- > 0)
-       { if (md->lcc[*p++] != md->lcc[*eptr++]) return -1; }
+       {
+       if (eptr >= md->end_subject) return -2;   /* Partial match */
+       if (TABLE_GET(*p, md->lcc, *p) != TABLE_GET(*eptr, md->lcc, *eptr)) return -1;
+       p++;
+       eptr++;
+       }
      }
    }
-Line 221 
 are in UTF-8 mode. */
+Line 230 
 are in UTF-8 mode. */
  else
    {
-   if (eptr + length > md->end_subject) return -1;
+   while (length-- > 0)
-   while (length-- > 0) if (*p++ != *eptr++) return -1;
+     {
+     if (eptr >= md->end_subject) return -2;   /* Partial match */
+     if (*p++ != *eptr++) return -1;
+     }
    }
- return eptr - eptr_start;
+ return (int)(eptr - eptr_start);
  }
-Line 290 
 actually used in this definition. */
+Line 302 
 actually used in this definition. */
  #define RMATCH(ra,rb,rc,rd,re,rw) \
    { \
    printf("match() called in line %d\n", __LINE__); \
-   rrc = match(ra,rb,mstart,markptr,rc,rd,re,rdepth+1); \
+   rrc = match(ra,rb,mstart,rc,rd,re,rdepth+1); \
    printf("to line %d\n", __LINE__); \
    }
  #define RRETURN(ra) \
-Line 300 
 actually used in this definition. */
+Line 312 
 actually used in this definition. */
    }
  #else
  #define RMATCH(ra,rb,rc,rd,re,rw) \
-   rrc = match(ra,rb,mstart,markptr,rc,rd,re,rdepth+1)
+   rrc = match(ra,rb,mstart,rc,rd,re,rdepth+1)
  #define RRETURN(ra) return ra
  #endif
-Line 315 
 argument of match(), which never changes
+Line 327 
 argument of match(), which never changes
  #define RMATCH(ra,rb,rc,rd,re,rw)\
    {\
-   heapframe *newframe = (heapframe *)(pcre_stack_malloc)(sizeof(heapframe));\
+   heapframe *newframe = frame->Xnextframe;\
-   if (newframe == NULL) RRETURN(PCRE_ERROR_NOMEMORY);\
+   if (newframe == NULL)\
-   frame->Xwhere = rw; \
+     {\
+     newframe = (heapframe *)(PUBL(stack_malloc))(sizeof(heapframe));\
+     if (newframe == NULL) RRETURN(PCRE_ERROR_NOMEMORY);\
+     newframe->Xnextframe = NULL;\
+     frame->Xnextframe = newframe;\
+     }\
+   frame->Xwhere = rw;\
    newframe->Xeptr = ra;\
    newframe->Xecode = rb;\
    newframe->Xmstart = mstart;\
-   newframe->Xmarkptr = markptr;\
    newframe->Xoffset_top = rc;\
    newframe->Xeptrb = re;\
    newframe->Xrdepth = frame->Xrdepth + 1;\
-Line 337 
 argument of match(), which never changes
+Line 354 
 argument of match(), which never changes
    {\
    heapframe *oldframe = frame;\
    frame = oldframe->Xprevframe;\
-   (pcre_stack_free)(oldframe);\
    if (frame != NULL)\
      {\
      rrc = ra;\
-Line 351 
 argument of match(), which never changes
+Line 367 
 argument of match(), which never changes
  typedef struct heapframe {
    struct heapframe *Xprevframe;
+   struct heapframe *Xnextframe;
    /* Function arguments that may change */
-   USPTR Xeptr;
+   PCRE_PUCHAR Xeptr;
-   const uschar *Xecode;
+   const pcre_uchar *Xecode;
-   USPTR Xmstart;
+   PCRE_PUCHAR Xmstart;
-   USPTR Xmarkptr;
    int Xoffset_top;
    eptrblock *Xeptrb;
    unsigned int Xrdepth;
    /* Function local variables */
-   USPTR Xcallpat;
+   PCRE_PUCHAR Xcallpat;
- #ifdef SUPPORT_UTF8
+ #ifdef SUPPORT_UTF
-   USPTR Xcharptr;
+   PCRE_PUCHAR Xcharptr;
  #endif
-   USPTR Xdata;
+   PCRE_PUCHAR Xdata;
-   USPTR Xnext;
+   PCRE_PUCHAR Xnext;
-   USPTR Xpp;
+   PCRE_PUCHAR Xpp;
-   USPTR Xprev;
+   PCRE_PUCHAR Xprev;
-   USPTR Xsaved_eptr;
+   PCRE_PUCHAR Xsaved_eptr;
    recursion_info Xnew_recursive;
-Line 385 
 typedef struct heapframe {
+Line 401 
 typedef struct heapframe {
    int Xprop_value;
    int Xprop_fail_result;
    int Xoclength;
-   uschar Xocchars[8];
+   pcre_uchar Xocchars[6];
  #endif
    int Xcodelink;
-Line 427 
 returns a negative (error) response, the
+Line 443 
 returns a negative (error) response, the
  same response. */
  /* These macros pack up tests that are used for partial matching, and which
- 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
  at the end of the subject and also past the start of the subject (i.e.
  something has been matched). For hard partial matching, we then return
  immediately. The second one is used when we already know we are past the end of
-Line 438 
 the subject. */
+Line 454 
 the subject. */
        eptr > md->start_used_ptr) \
      { \
      md->hitend = TRUE; \
-     if (md->partial > 1) MRRETURN(PCRE_ERROR_PARTIAL); \
+     if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL); \
      }
  #define SCHECK_PARTIAL()\
    if (md->partial != 0 && eptr > md->start_used_ptr) \
      { \
      md->hitend = TRUE; \
-     if (md->partial > 1) MRRETURN(PCRE_ERROR_PARTIAL); \
+     if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL); \
      }
  /* Performance note: It might be tempting to extract commonly used fields from
- 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
  performance. Tests using gcc on a SPARC disproved this; in the first case, it
  made performance worse.
-Line 459 
 Arguments:
+Line 475 
 Arguments:
     ecode       pointer to current position in compiled code
     mstart      pointer to the current match start position (can be modified
                   by encountering \K)
-    markptr     pointer to the most recent MARK name, or NULL
     offset_top  current top pointer
     md          pointer to "static" info for the match
     eptrb       pointer to chain of blocks containing eptr at start of
-Line 474 
 Returns:       MATCH_MATCH if matched
+Line 489 
 Returns:       MATCH_MATCH if matched
  */
  static int
- match(REGISTER USPTR eptr, REGISTER const uschar *ecode, USPTR mstart,
+ match(REGISTER PCRE_PUCHAR eptr, REGISTER const pcre_uchar *ecode,
-   const uschar *markptr, int offset_top, match_data *md, eptrblock *eptrb,
+   PCRE_PUCHAR mstart, int offset_top, match_data *md, eptrblock *eptrb,
    unsigned int rdepth)
  {
  /* These variables do not need to be preserved over recursion in this function,
-Line 484 
 so they can be ordinary variables in all
+Line 499 
 so they can be ordinary variables in all
  register int  rrc;         /* Returns from recursive calls */
  register int  i;           /* Used for loops not involving calls to RMATCH() */
- register unsigned int c;   /* Character values not kept over RMATCH() calls */
+ register pcre_uint32 c;    /* Character values not kept over RMATCH() calls */
- register BOOL utf8;        /* Local copy of UTF-8 flag for speed */
+ register BOOL utf;         /* Local copy of UTF flag for speed */
  BOOL minimize, possessive; /* Quantifier options */
  BOOL caseless;
  int condcode;
  /* When recursion is not being used, all "local" variables that have to be
- preserved over calls to RMATCH() are part of a "frame" which is obtained from
+ preserved over calls to RMATCH() are part of a "frame". We set up the top-level
- heap storage. Set up the top-level frame here; others are obtained from the
+ frame on the stack here; subsequent instantiations are obtained from the heap
- heap whenever RMATCH() does a "recursion". See the macro definitions above. */
+ whenever RMATCH() does a "recursion". See the macro definitions above. Putting
+ the top-level on the stack rather than malloc-ing them all gives a performance
+ boost in many cases where there is not much "recursion". */
  #ifdef NO_RECURSE
- heapframe *frame = (heapframe *)(pcre_stack_malloc)(sizeof(heapframe));
+ heapframe *frame = (heapframe *)md->match_frames_base;
- if (frame == NULL) RRETURN(PCRE_ERROR_NOMEMORY);
- frame->Xprevframe = NULL;            /* Marks the top level */
  /* Copy in the original argument variables */
  frame->Xeptr = eptr;
  frame->Xecode = ecode;
  frame->Xmstart = mstart;
- frame->Xmarkptr = markptr;
  frame->Xoffset_top = offset_top;
  frame->Xeptrb = eptrb;
  frame->Xrdepth = rdepth;
-Line 520 
 HEAP_RECURSE:
+Line 534 
 HEAP_RECURSE:
  #define eptr               frame->Xeptr
  #define ecode              frame->Xecode
  #define mstart             frame->Xmstart
- #define markptr            frame->Xmarkptr
  #define offset_top         frame->Xoffset_top
  #define eptrb              frame->Xeptrb
  #define rdepth             frame->Xrdepth
  /* Ditto for the local variables */
- #ifdef SUPPORT_UTF8
+ #ifdef SUPPORT_UTF
  #define charptr            frame->Xcharptr
  #endif
  #define callpat            frame->Xcallpat
-Line 585 
 declarations can be cut out in a block.
+Line 598 
 declarations can be cut out in a block.
  below are for variables that do not have to be preserved over a recursive call
  to RMATCH(). */
- #ifdef SUPPORT_UTF8
+ #ifdef SUPPORT_UTF
- const uschar *charptr;
+ const pcre_uchar *charptr;
  #endif
- const uschar *callpat;
+ const pcre_uchar *callpat;
- const uschar *data;
+ const pcre_uchar *data;
- const uschar *next;
+ const pcre_uchar *next;
- USPTR         pp;
+ PCRE_PUCHAR       pp;
- const uschar *prev;
+ const pcre_uchar *prev;
- USPTR         saved_eptr;
+ PCRE_PUCHAR       saved_eptr;
  recursion_info new_recursive;
-Line 606 
 int prop_type;
+Line 619 
 int prop_type;
  int prop_value;
  int prop_fail_result;
  int oclength;
- uschar occhars[8];
+ pcre_uchar occhars[6];
  #endif
  int codelink;
-Line 616 
 int max;
+Line 629 
 int max;
  int min;
  int number;
  int offset;
- int op;
+ pcre_uchar op;
  int save_capture_last;
  int save_offset1, save_offset2, save_offset3;
  int stacksave[REC_STACK_SAVE_MAX];
  eptrblock newptrb;
+ /* There is a special fudge for calling match() in a way that causes it to
+ measure the size of its basic stack frame when the stack is being used for
+ recursion. The second argument (ecode) being NULL triggers this behaviour. It
+ cannot normally ever be NULL. The return is the negated value of the frame
+ size. */
+ if (ecode == NULL)
+   {
+   if (rdepth == 0)
+     return match((PCRE_PUCHAR)&rdepth, NULL, NULL, 0, NULL, NULL, 1);
+   else
+     {
+     int len = (char *)&rdepth - (char *)eptr;
+     return (len > 0)? -len : len;
+     }
+   }
  #endif     /* NO_RECURSE */
  /* To save space on the stack and in the heap frame, I have doubled up on some
-Line 634 
 the alternative names that are used. */
+Line 664 
 the alternative names that are used. */
  #define code_offset   codelink
  #define condassert    condition
  #define matched_once  prev_is_word
+ #define foc           number
+ #define save_mark     data
  /* These statements are here to stop the compiler complaining about unitialized
  variables. */
-Line 659 
 defined). However, RMATCH isn't like a f
+Line 691 
 defined). However, RMATCH isn't like a f
  complicated macro. It has to be used in one particular way. This shouldn't,
  however, impact performance when true recursion is being used. */
- #ifdef SUPPORT_UTF8
+ #ifdef SUPPORT_UTF
- utf8 = md->utf8;       /* Local copy of the flag */
+ utf = md->utf;       /* Local copy of the flag */
  #else
- utf8 = FALSE;
+ utf = FALSE;
  #endif
  /* First check that we haven't called match() too many times, or that we
-Line 701 
 for (;;)
+Line 733 
 for (;;)
    switch(op)
      {
      case OP_MARK:
-     markptr = ecode + 2;
+     md->nomatch_mark = ecode + 2;
-     RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode] + ecode[1], offset_top, md,
+     md->mark = NULL;    /* In case previously set by assertion */
+     RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode] + ecode[1], offset_top, md,
        eptrb, RM55);
+     if ((rrc == MATCH_MATCH || rrc == MATCH_ACCEPT) &&
+          md->mark == NULL) md->mark = ecode + 2;
      /* A return of MATCH_SKIP_ARG means that matching failed at SKIP with an
      argument, and we must check whether that argument matches this MARK's
-Line 712 
 for (;;)
+Line 747 
 for (;;)
      position and return MATCH_SKIP. Otherwise, pass back the return code
      unaltered. */
-     if (rrc == MATCH_SKIP_ARG &&
+     else if (rrc == MATCH_SKIP_ARG &&
-         strcmp((char *)markptr, (char *)(md->start_match_ptr)) == 0)
+         STRCMP_UC_UC(ecode + 2, md->start_match_ptr) == 0)
        {
        md->start_match_ptr = eptr;
        RRETURN(MATCH_SKIP);
        }
-     if (md->mark == NULL) md->mark = markptr;
      RRETURN(rrc);
      case OP_FAIL:
-     MRRETURN(MATCH_NOMATCH);
+     RRETURN(MATCH_NOMATCH);
      /* COMMIT overrides PRUNE, SKIP, and THEN */
      case OP_COMMIT:
-     RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,
+     RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md,
        eptrb, RM52);
      if (rrc != MATCH_NOMATCH && rrc != MATCH_PRUNE &&
          rrc != MATCH_SKIP && rrc != MATCH_SKIP_ARG &&
          rrc != MATCH_THEN)
        RRETURN(rrc);
-     MRRETURN(MATCH_COMMIT);
+     RRETURN(MATCH_COMMIT);
      /* PRUNE overrides THEN */
      case OP_PRUNE:
-     RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,
+     RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md,
        eptrb, RM51);
      if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN) RRETURN(rrc);
-     MRRETURN(MATCH_PRUNE);
+     RRETURN(MATCH_PRUNE);
      case OP_PRUNE_ARG:
-     RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode] + ecode[1], offset_top, md,
+     md->nomatch_mark = ecode + 2;
+     md->mark = NULL;    /* In case previously set by assertion */
+     RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode] + ecode[1], offset_top, md,
        eptrb, RM56);
+     if ((rrc == MATCH_MATCH || rrc == MATCH_ACCEPT) &&
+          md->mark == NULL) md->mark = ecode + 2;
      if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN) RRETURN(rrc);
-     md->mark = ecode + 2;
      RRETURN(MATCH_PRUNE);
      /* SKIP overrides PRUNE and THEN */
      case OP_SKIP:
-     RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,
+     RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md,
        eptrb, RM53);
      if (rrc != MATCH_NOMATCH && rrc != MATCH_PRUNE && rrc != MATCH_THEN)
        RRETURN(rrc);
      md->start_match_ptr = eptr;   /* Pass back current position */
-     MRRETURN(MATCH_SKIP);
+     RRETURN(MATCH_SKIP);
+     /* Note that, for Perl compatibility, SKIP with an argument does NOT set
+     nomatch_mark. There is a flag that disables this opcode when re-matching a
+     pattern that ended with a SKIP for which there was not a matching MARK. */
      case OP_SKIP_ARG:
-     RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode] + ecode[1], offset_top, md,
+     if (md->ignore_skip_arg)
+       {
+       ecode += PRIV(OP_lengths)[*ecode] + ecode[1];
+       break;
+       }
+     RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode] + ecode[1], offset_top, md,
        eptrb, RM57);
      if (rrc != MATCH_NOMATCH && rrc != MATCH_PRUNE && rrc != MATCH_THEN)
        RRETURN(rrc);
      /* Pass back the current skip name by overloading md->start_match_ptr and
      returning the special MATCH_SKIP_ARG return code. This will either be
-     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
-     as PRUNE. */
+     with the md->ignore_skip_arg flag set. */
      md->start_match_ptr = ecode + 2;
      RRETURN(MATCH_SKIP_ARG);
-Line 780 
 for (;;)
+Line 825 
 for (;;)
      match pointer to do this. */
      case OP_THEN:
-     RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,
+     RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md,
        eptrb, RM54);
      if (rrc != MATCH_NOMATCH) RRETURN(rrc);
      md->start_match_ptr = ecode;
-     MRRETURN(MATCH_THEN);
+     RRETURN(MATCH_THEN);
      case OP_THEN_ARG:
-     RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode] + ecode[1], offset_top,
+     md->nomatch_mark = ecode + 2;
+     md->mark = NULL;    /* In case previously set by assertion */
+     RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode] + ecode[1], offset_top,
        md, eptrb, RM58);
+     if ((rrc == MATCH_MATCH || rrc == MATCH_ACCEPT) &&
+          md->mark == NULL) md->mark = ecode + 2;
      if (rrc != MATCH_NOMATCH) RRETURN(rrc);
      md->start_match_ptr = ecode;
-     md->mark = ecode + 2;
      RRETURN(MATCH_THEN);
      /* Handle an atomic group that does not contain any capturing parentheses.
-Line 810 
 for (;;)
+Line 858 
 for (;;)
      case OP_ONCE_NC:
      prev = ecode;
      saved_eptr = eptr;
+     save_mark = md->mark;
      do
        {
        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, eptrb, RM64);
        if (rrc == MATCH_MATCH)  /* Note: _not_ MATCH_ACCEPT */
          {
          mstart = md->start_match_ptr;
-         markptr = md->mark;
          break;
          }
        if (rrc == MATCH_THEN)
-Line 829 
 for (;;)
+Line 877 
 for (;;)
        if (rrc != MATCH_NOMATCH) RRETURN(rrc);
        ecode += GET(ecode,1);
+       md->mark = save_mark;
        }
      while (*ecode == OP_ALT);
-Line 868 
 for (;;)
+Line 917 
 for (;;)
        }
      else  /* OP_KETRMAX */
        {
-       md->match_function_type = MATCH_CBEGROUP;
        RMATCH(eptr, prev, offset_top, md, eptrb, RM66);
        if (rrc != MATCH_NOMATCH) RRETURN(rrc);
        ecode += 1 + LINK_SIZE;
-Line 908 
 for (;;)
+Line 956 
 for (;;)
        save_offset2 = md->offset_vector[offset+1];
        save_offset3 = md->offset_vector[md->offset_end - number];
        save_capture_last = md->capture_last;
+       save_mark = md->mark;
        DPRINTF(("saving %d %d %d\n", save_offset1, save_offset2, save_offset3));
        md->offset_vector[md->offset_end - number] =
-Line 916 
 for (;;)
+Line 965 
 for (;;)
        for (;;)
          {
          if (op >= OP_SBRA) md->match_function_type = MATCH_CBEGROUP;
-         RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,
+         RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md,
            eptrb, RM1);
          if (rrc == MATCH_ONCE) break;  /* Backing up through an atomic group */
-Line 944 
 for (;;)
+Line 993 
 for (;;)
          if (rrc != MATCH_NOMATCH) RRETURN(rrc);
          md->capture_last = save_capture_last;
          ecode += GET(ecode, 1);
+         md->mark = save_mark;
          if (*ecode != OP_ALT) break;
          }
-Line 954 
 for (;;)
+Line 1004 
 for (;;)
        /* At this point, rrc will be one of MATCH_ONCE or MATCH_NOMATCH. */
-       if (md->mark == NULL) md->mark = markptr;
        RRETURN(rrc);
        }
-Line 996 
 for (;;)
+Line 1045 
 for (;;)
      for (;;)
        {
-       if (op >= OP_SBRA || op == OP_ONCE) md->match_function_type = MATCH_CBEGROUP;
+       if (op >= OP_SBRA || op == OP_ONCE)
+         md->match_function_type = MATCH_CBEGROUP;
        /* If this is not a possibly empty group, and there are no (*THEN)s in
        the pattern, and this is the final alternative, optimize as described
-Line 1004 
 for (;;)
+Line 1054 
 for (;;)
        else if (!md->hasthen && ecode[GET(ecode, 1)] != OP_ALT)
          {
-         ecode += _pcre_OP_lengths[*ecode];
+         ecode += PRIV(OP_lengths)[*ecode];
          goto TAIL_RECURSE;
          }
        /* In all other cases, we have to make another call to match(). */
-       RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md, eptrb,
+       save_mark = md->mark;
+       RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md, eptrb,
          RM2);
        /* See comment in the code for capturing groups above about handling
-Line 1028 
 for (;;)
+Line 1079 
 for (;;)
          {
          if (rrc == MATCH_ONCE)
            {
-           const uschar *scode = ecode;
+           const pcre_uchar *scode = ecode;
            if (*scode != OP_ONCE)           /* If not at start, find it */
              {
              while (*scode == OP_ALT) scode += GET(scode, 1);
-Line 1039 
 for (;;)
+Line 1090 
 for (;;)
          RRETURN(rrc);
          }
        ecode += GET(ecode, 1);
+       md->mark = save_mark;
        if (*ecode != OP_ALT) break;
        }
-     if (md->mark == NULL) md->mark = markptr;
      RRETURN(MATCH_NOMATCH);
      /* Handle possessive capturing brackets with an unlimited repeat. We come
-Line 1071 
 for (;;)
+Line 1122 
 for (;;)
      if (offset < md->offset_max)
        {
        matched_once = FALSE;
-       code_offset = ecode - md->start_code;
+       code_offset = (int)(ecode - md->start_code);
        save_offset1 = md->offset_vector[offset];
        save_offset2 = md->offset_vector[offset+1];
-Line 1094 
 for (;;)
+Line 1145 
 for (;;)
          md->offset_vector[md->offset_end - number] =
            (int)(eptr - md->start_subject);
          if (op >= OP_SBRA) md->match_function_type = MATCH_CBEGROUP;
-         RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,
+         RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md,
            eptrb, RM63);
          if (rrc == MATCH_KETRPOS)
            {
-Line 1130 
 for (;;)
+Line 1181 
 for (;;)
          md->offset_vector[md->offset_end - number] = save_offset3;
          }
-       if (md->mark == NULL) md->mark = markptr;
        if (allow_zero || matched_once)
          {
          ecode += 1 + LINK_SIZE;
-Line 1162 
 for (;;)
+Line 1212 
 for (;;)
      POSSESSIVE_NON_CAPTURE:
      matched_once = FALSE;
-     code_offset = ecode - md->start_code;
+     code_offset = (int)(ecode - md->start_code);
      for (;;)
        {
        if (op >= OP_SBRA) md->match_function_type = MATCH_CBEGROUP;
-       RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,
+       RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md,
          eptrb, RM48);
        if (rrc == MATCH_KETRPOS)
          {
-Line 1217 
 for (;;)
+Line 1267 
 for (;;)
      if (ecode[LINK_SIZE+1] == OP_CALLOUT)
        {
-       if (pcre_callout != NULL)
+       if (PUBL(callout) != NULL)
          {
-         pcre_callout_block cb;
+         PUBL(callout_block) cb;
          cb.version          = 2;   /* Version 1 of the callout block */
          cb.callout_number   = ecode[LINK_SIZE+2];
          cb.offset_vector    = md->offset_vector;
+ #if defined COMPILE_PCRE8
          cb.subject          = (PCRE_SPTR)md->start_subject;
+ #elif defined COMPILE_PCRE16
+         cb.subject          = (PCRE_SPTR16)md->start_subject;
+ #elif defined COMPILE_PCRE32
+         cb.subject          = (PCRE_SPTR32)md->start_subject;
+ #endif
          cb.subject_length   = (int)(md->end_subject - md->start_subject);
          cb.start_match      = (int)(mstart - md->start_subject);
          cb.current_position = (int)(eptr - md->start_subject);
-Line 1232 
 for (;;)
+Line 1288 
 for (;;)
          cb.capture_top      = offset_top/2;
          cb.capture_last     = md->capture_last;
          cb.callout_data     = md->callout_data;
-         cb.mark             = markptr;
+         cb.mark             = md->nomatch_mark;
-         if ((rrc = (*pcre_callout)(&cb)) > 0) MRRETURN(MATCH_NOMATCH);
+         if ((rrc = (*PUBL(callout))(&cb)) > 0) RRETURN(MATCH_NOMATCH);
          if (rrc < 0) RRETURN(rrc);
          }
-       ecode += _pcre_OP_lengths[OP_CALLOUT];
+       ecode += PRIV(OP_lengths)[OP_CALLOUT];
        }
      condcode = ecode[LINK_SIZE+1];
-Line 1262 
 for (;;)
+Line 1318 
 for (;;)
          if (!condition && condcode == OP_NRREF)
            {
-           uschar *slotA = md->name_table;
+           pcre_uchar *slotA = md->name_table;
            for (i = 0; i < md->name_count; i++)
              {
              if (GET2(slotA, 0) == recno) break;
-Line 1275 
 for (;;)
+Line 1331 
 for (;;)
            if (i < md->name_count)
              {
-             uschar *slotB = slotA;
+             pcre_uchar *slotB = slotA;
              while (slotB > md->name_table)
                {
                slotB -= md->name_entry_size;
-               if (strcmp((char *)slotA + 2, (char *)slotB + 2) == 0)
+               if (STRCMP_UC_UC(slotA + IMM2_SIZE, slotB + IMM2_SIZE) == 0)
                  {
                  condition = GET2(slotB, 0) == md->recursive->group_num;
                  if (condition) break;
-Line 1295 
 for (;;)
+Line 1351 
 for (;;)
                for (i++; i < md->name_count; i++)
                  {
                  slotB += md->name_entry_size;
-                 if (strcmp((char *)slotA + 2, (char *)slotB + 2) == 0)
+                 if (STRCMP_UC_UC(slotA + IMM2_SIZE, slotB + IMM2_SIZE) == 0)
                    {
                    condition = GET2(slotB, 0) == md->recursive->group_num;
                    if (condition) break;
-Line 1308 
 for (;;)
+Line 1364 
 for (;;)
          /* Chose branch according to the condition */
-         ecode += condition? 3 : GET(ecode, 1);
+         ecode += condition? 1 + IMM2_SIZE : GET(ecode, 1);
          }
        }
-Line 1325 
 for (;;)
+Line 1381 
 for (;;)
        if (!condition && condcode == OP_NCREF)
          {
          int refno = offset >> 1;
-         uschar *slotA = md->name_table;
+         pcre_uchar *slotA = md->name_table;
          for (i = 0; i < md->name_count; i++)
            {
-Line 1339 
 for (;;)
+Line 1395 
 for (;;)
          if (i < md->name_count)
            {
-           uschar *slotB = slotA;
+           pcre_uchar *slotB = slotA;
            while (slotB > md->name_table)
              {
              slotB -= md->name_entry_size;
-             if (strcmp((char *)slotA + 2, (char *)slotB + 2) == 0)
+             if (STRCMP_UC_UC(slotA + IMM2_SIZE, slotB + IMM2_SIZE) == 0)
                {
                offset = GET2(slotB, 0) << 1;
                condition = offset < offset_top &&
-Line 1361 
 for (;;)
+Line 1417 
 for (;;)
              for (i++; i < md->name_count; i++)
                {
                slotB += md->name_entry_size;
-               if (strcmp((char *)slotA + 2, (char *)slotB + 2) == 0)
+               if (STRCMP_UC_UC(slotA + IMM2_SIZE, slotB + IMM2_SIZE) == 0)
                  {
                  offset = GET2(slotB, 0) << 1;
                  condition = offset < offset_top &&
-Line 1376 
 for (;;)
+Line 1432 
 for (;;)
        /* Chose branch according to the condition */
-       ecode += condition? 3 : GET(ecode, 1);
+       ecode += condition? 1 + IMM2_SIZE : GET(ecode, 1);
        }
      else if (condcode == OP_DEF)     /* DEFINE - always false */
-Line 1468 
 for (;;)
+Line 1524 
 for (;;)
        md->offset_vector[offset+1] = (int)(eptr - md->start_subject);
        if (offset_top <= offset) offset_top = offset + 2;
        }
-     ecode += 3;
+     ecode += 1 + IMM2_SIZE;
      break;
-Line 1488 
 for (;;)
+Line 1544 
 for (;;)
           (md->notempty ||
             (md->notempty_atstart &&
               mstart == md->start_subject + md->start_offset)))
-       MRRETURN(MATCH_NOMATCH);
+       RRETURN(MATCH_NOMATCH);
      /* Otherwise, we have a match. */
-Line 1497 
 for (;;)
+Line 1553 
 for (;;)
      md->start_match_ptr = mstart;       /* and the start (\K can modify) */
      /* For some reason, the macros don't work properly if an expression is
-     given as the argument to MRRETURN when the heap is in use. */
+     given as the argument to RRETURN when the heap is in use. */
      rrc = (op == OP_END)? MATCH_MATCH : MATCH_ACCEPT;
-     MRRETURN(rrc);
+     RRETURN(rrc);
      /* Assertion brackets. Check the alternative branches in turn - the
      matching won't pass the KET for an assertion. If any one branch matches,
-Line 1515 
 for (;;)
+Line 1571 
 for (;;)
      case OP_ASSERT:
      case OP_ASSERTBACK:
+     save_mark = md->mark;
      if (md->match_function_type == MATCH_CONDASSERT)
        {
        condassert = TRUE;
-Line 1528 
 for (;;)
+Line 1585 
 for (;;)
        if (rrc == MATCH_MATCH || rrc == MATCH_ACCEPT)
          {
          mstart = md->start_match_ptr;   /* In case \K reset it */
-         markptr = md->mark;
          break;
          }
+       md->mark = save_mark;
-       /* PCRE does not allow THEN to escape beyond an assertion; it is treated
+       /* A COMMIT failure must fail the entire assertion, without trying any
-       as NOMATCH. */
+       subsequent branches. */
+       if (rrc == MATCH_COMMIT) RRETURN(MATCH_NOMATCH);
+       /* PCRE does not allow THEN to escape beyond an assertion; it
+       is treated as NOMATCH. */
        if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN) RRETURN(rrc);
        ecode += GET(ecode, 1);
        }
      while (*ecode == OP_ALT);
-     if (*ecode == OP_KET) MRRETURN(MATCH_NOMATCH);
+     if (*ecode == OP_KET) RRETURN(MATCH_NOMATCH);
      /* If checking an assertion for a condition, return MATCH_MATCH. */
-Line 1560 
 for (;;)
+Line 1622 
 for (;;)
      case OP_ASSERT_NOT:
      case OP_ASSERTBACK_NOT:
+     save_mark = md->mark;
      if (md->match_function_type == MATCH_CONDASSERT)
        {
        condassert = TRUE;
-Line 1570 
 for (;;)
+Line 1633 
 for (;;)
      do
        {
        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, NULL, RM5);
-       if (rrc == MATCH_MATCH || rrc == MATCH_ACCEPT) MRRETURN(MATCH_NOMATCH);
+       md->mark = save_mark;
+       if (rrc == MATCH_MATCH || rrc == MATCH_ACCEPT) RRETURN(MATCH_NOMATCH);
        if (rrc == MATCH_SKIP || rrc == MATCH_PRUNE || rrc == MATCH_COMMIT)
          {
          do ecode += GET(ecode,1); while (*ecode == OP_ALT);
-Line 1596 
 for (;;)
+Line 1660 
 for (;;)
      back a number of characters, not bytes. */
      case OP_REVERSE:
- #ifdef SUPPORT_UTF8
+ #ifdef SUPPORT_UTF
-     if (utf8)
+     if (utf)
        {
        i = GET(ecode, 1);
        while (i-- > 0)
          {
          eptr--;
-         if (eptr < md->start_subject) MRRETURN(MATCH_NOMATCH);
+         if (eptr < md->start_subject) RRETURN(MATCH_NOMATCH);
          BACKCHAR(eptr);
          }
        }
-Line 1614 
 for (;;)
+Line 1678 
 for (;;)
        {
        eptr -= GET(ecode, 1);
-       if (eptr < md->start_subject) MRRETURN(MATCH_NOMATCH);
+       if (eptr < md->start_subject) RRETURN(MATCH_NOMATCH);
        }
      /* Save the earliest consulted character, then skip to next op code */
-Line 1628 
 for (;;)
+Line 1692 
 for (;;)
      function is able to force a failure. */
      case OP_CALLOUT:
-     if (pcre_callout != NULL)
+     if (PUBL(callout) != NULL)
        {
-       pcre_callout_block cb;
+       PUBL(callout_block) cb;
        cb.version          = 2;   /* Version 1 of the callout block */
        cb.callout_number   = ecode[1];
        cb.offset_vector    = md->offset_vector;
+ #if defined COMPILE_PCRE8
        cb.subject          = (PCRE_SPTR)md->start_subject;
+ #elif defined COMPILE_PCRE16
+       cb.subject          = (PCRE_SPTR16)md->start_subject;
+ #elif defined COMPILE_PCRE32
+       cb.subject          = (PCRE_SPTR32)md->start_subject;
+ #endif
        cb.subject_length   = (int)(md->end_subject - md->start_subject);
        cb.start_match      = (int)(mstart - md->start_subject);
        cb.current_position = (int)(eptr - md->start_subject);
-Line 1643 
 for (;;)
+Line 1713 
 for (;;)
        cb.capture_top      = offset_top/2;
        cb.capture_last     = md->capture_last;
        cb.callout_data     = md->callout_data;
-       cb.mark             = markptr;
+       cb.mark             = md->nomatch_mark;
-       if ((rrc = (*pcre_callout)(&cb)) > 0) MRRETURN(MATCH_NOMATCH);
+       if ((rrc = (*PUBL(callout))(&cb)) > 0) RRETURN(MATCH_NOMATCH);
        if (rrc < 0) RRETURN(rrc);
        }
      ecode += 2 + 2*LINK_SIZE;
-Line 1703 
 for (;;)
+Line 1773 
 for (;;)
        else
          {
          new_recursive.offset_save =
-           (int *)(pcre_malloc)(new_recursive.saved_max * sizeof(int));
+           (int *)(PUBL(malloc))(new_recursive.saved_max * sizeof(int));
          if (new_recursive.offset_save == NULL) RRETURN(PCRE_ERROR_NOMEMORY);
          }
        memcpy(new_recursive.offset_save, md->offset_vector,
-Line 1718 
 for (;;)
+Line 1788 
 for (;;)
        do
          {
          if (cbegroup) md->match_function_type = MATCH_CBEGROUP;
-         RMATCH(eptr, callpat + _pcre_OP_lengths[*callpat], offset_top,
+         RMATCH(eptr, callpat + PRIV(OP_lengths)[*callpat], offset_top,
            md, eptrb, RM6);
          memcpy(md->offset_vector, new_recursive.offset_save,
              new_recursive.saved_max * sizeof(int));
-Line 1727 
 for (;;)
+Line 1797 
 for (;;)
            {
            DPRINTF(("Recursion matched\n"));
            if (new_recursive.offset_save != stacksave)
-             (pcre_free)(new_recursive.offset_save);
+             (PUBL(free))(new_recursive.offset_save);
            /* Set where we got to in the subject, and reset the start in case
            it was changed by \K. This *is* propagated back out of a recursion,
-Line 1738 
 for (;;)
+Line 1808 
 for (;;)
            goto RECURSION_MATCHED;        /* Exit loop; end processing */
            }
-         /* PCRE does not allow THEN to escape beyond a recursion; it is treated
+         /* PCRE does not allow THEN or COMMIT to escape beyond a recursion; it
-         as NOMATCH. */
+         is treated as NOMATCH. */
-         else if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN)
+         else if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN &&
+                  rrc != MATCH_COMMIT)
            {
            DPRINTF(("Recursion gave error %d\n", rrc));
            if (new_recursive.offset_save != stacksave)
-             (pcre_free)(new_recursive.offset_save);
+             (PUBL(free))(new_recursive.offset_save);
            RRETURN(rrc);
            }
-Line 1757 
 for (;;)
+Line 1828 
 for (;;)
        DPRINTF(("Recursion didn't match\n"));
        md->recursive = new_recursive.prevrec;
        if (new_recursive.offset_save != stacksave)
-         (pcre_free)(new_recursive.offset_save);
+         (PUBL(free))(new_recursive.offset_save);
-       MRRETURN(MATCH_NOMATCH);
+       RRETURN(MATCH_NOMATCH);
        }
      RECURSION_MATCHED:
-Line 1838 
 for (;;)
+Line 1909 
 for (;;)
        md->end_match_ptr = eptr;      /* For ONCE_NC */
        md->end_offset_top = offset_top;
        md->start_match_ptr = mstart;
-       MRRETURN(MATCH_MATCH);         /* Sets md->mark */
+       RRETURN(MATCH_MATCH);         /* Sets md->mark */
        }
      /* For capturing groups we have to check the group number back at the start
-Line 1952 
 for (;;)
+Line 2023 
 for (;;)
          }
        if (*prev >= OP_SBRA)    /* Could match an empty string */
          {
-         md->match_function_type = MATCH_CBEGROUP;
          RMATCH(eptr, prev, offset_top, md, eptrb, RM50);
          RRETURN(rrc);
          }
-Line 1961 
 for (;;)
+Line 2031 
 for (;;)
        }
      else  /* OP_KETRMAX */
        {
-       if (*prev >= OP_SBRA) md->match_function_type = MATCH_CBEGROUP;
        RMATCH(eptr, prev, offset_top, md, eptrb, RM13);
        if (rrc == MATCH_ONCE && md->once_target == prev) rrc = MATCH_NOMATCH;
        if (rrc != MATCH_NOMATCH) RRETURN(rrc);
-Line 1980 
 for (;;)
+Line 2049 
 for (;;)
      /* Not multiline mode: start of subject assertion, unless notbol. */
      case OP_CIRC:
-     if (md->notbol && eptr == md->start_subject) MRRETURN(MATCH_NOMATCH);
+     if (md->notbol && eptr == md->start_subject) RRETURN(MATCH_NOMATCH);
      /* Start of subject assertion */
      case OP_SOD:
-     if (eptr != md->start_subject) MRRETURN(MATCH_NOMATCH);
+     if (eptr != md->start_subject) RRETURN(MATCH_NOMATCH);
      ecode++;
      break;
      /* Multiline mode: start of subject unless notbol, or after any newline. */
      case OP_CIRCM:
-     if (md->notbol && eptr == md->start_subject) MRRETURN(MATCH_NOMATCH);
+     if (md->notbol && eptr == md->start_subject) RRETURN(MATCH_NOMATCH);
      if (eptr != md->start_subject &&
          (eptr == md->end_subject || !WAS_NEWLINE(eptr)))
-       MRRETURN(MATCH_NOMATCH);
+       RRETURN(MATCH_NOMATCH);
      ecode++;
      break;
      /* Start of match assertion */
      case OP_SOM:
-     if (eptr != md->start_subject + md->start_offset) MRRETURN(MATCH_NOMATCH);
+     if (eptr != md->start_subject + md->start_offset) RRETURN(MATCH_NOMATCH);
      ecode++;
      break;
-Line 2018 
 for (;;)
+Line 2087 
 for (;;)
      case OP_DOLLM:
      if (eptr < md->end_subject)
-       { if (!IS_NEWLINE(eptr)) MRRETURN(MATCH_NOMATCH); }
+       {
+       if (!IS_NEWLINE(eptr))
+         {
+         if (md->partial != 0 &&
+             eptr + 1 >= md->end_subject &&
+             NLBLOCK->nltype == NLTYPE_FIXED &&
+             NLBLOCK->nllen == 2 &&
+             *eptr == NLBLOCK->nl[0])
+           {
+           md->hitend = TRUE;
+           if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
+           }
+         RRETURN(MATCH_NOMATCH);
+         }
+       }
      else
        {
-       if (md->noteol) MRRETURN(MATCH_NOMATCH);
+       if (md->noteol) RRETURN(MATCH_NOMATCH);
        SCHECK_PARTIAL();
        }
      ecode++;
-Line 2031 
 for (;;)
+Line 2114 
 for (;;)
      subject unless noteol is set. */
      case OP_DOLL:
-     if (md->noteol) MRRETURN(MATCH_NOMATCH);
+     if (md->noteol) RRETURN(MATCH_NOMATCH);
      if (!md->endonly) goto ASSERT_NL_OR_EOS;
      /* ... else fall through for endonly */
-Line 2039 
 for (;;)
+Line 2122 
 for (;;)
      /* End of subject assertion (\z) */
      case OP_EOD:
-     if (eptr < md->end_subject) MRRETURN(MATCH_NOMATCH);
+     if (eptr < md->end_subject) RRETURN(MATCH_NOMATCH);
      SCHECK_PARTIAL();
      ecode++;
      break;
-Line 2050 
 for (;;)
+Line 2133 
 for (;;)
      ASSERT_NL_OR_EOS:
      if (eptr < md->end_subject &&
          (!IS_NEWLINE(eptr) || eptr != md->end_subject - md->nllen))
-       MRRETURN(MATCH_NOMATCH);
+       {
+       if (md->partial != 0 &&
+           eptr + 1 >= md->end_subject &&
+           NLBLOCK->nltype == NLTYPE_FIXED &&
+           NLBLOCK->nllen == 2 &&
+           *eptr == NLBLOCK->nl[0])
+         {
+         md->hitend = TRUE;
+         if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
+         }
+       RRETURN(MATCH_NOMATCH);
+       }
      /* Either at end of string or \n before end. */
-Line 2069 
 for (;;)
+Line 2163 
 for (;;)
        be "non-word" characters. Remember the earliest consulted character for
        partial matching. */
- #ifdef SUPPORT_UTF8
+ #ifdef SUPPORT_UTF
-       if (utf8)
+       if (utf)
          {
          /* Get status of previous character */
          if (eptr == md->start_subject) prev_is_word = FALSE; else
            {
-           USPTR lastptr = eptr - 1;
+           PCRE_PUCHAR lastptr = eptr - 1;
-           while((*lastptr & 0xc0) == 0x80) lastptr--;
+           BACKCHAR(lastptr);
            if (lastptr < md->start_used_ptr) md->start_used_ptr = lastptr;
            GETCHAR(c, lastptr);
  #ifdef SUPPORT_UCP
-Line 2142 
 for (;;)
+Line 2236 
 for (;;)
              }
            else
  #endif
-           prev_is_word = ((md->ctypes[eptr[-1]] & ctype_word) != 0);
+           prev_is_word = MAX_255(eptr[-1])
+             && ((md->ctypes[eptr[-1]] & ctype_word) != 0);
            }
          /* Get status of next character */
-Line 2165 
 for (;;)
+Line 2260 
 for (;;)
            }
          else
  #endif
-         cur_is_word = ((md->ctypes[*eptr] & ctype_word) != 0);
+         cur_is_word = MAX_255(*eptr)
+           && ((md->ctypes[*eptr] & ctype_word) != 0);
          }
        /* Now see if the situation is what we want */
        if ((*ecode++ == OP_WORD_BOUNDARY)?
             cur_is_word == prev_is_word : cur_is_word != prev_is_word)
-         MRRETURN(MATCH_NOMATCH);
+         RRETURN(MATCH_NOMATCH);
        }
      break;
-     /* Match a single character type; inline for speed */
+     /* Match any single character type except newline; have to take care with
+     CRLF newlines and partial matching. */
      case OP_ANY:
-     if (IS_NEWLINE(eptr)) MRRETURN(MATCH_NOMATCH);
+     if (IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH);
+     if (md->partial != 0 &&
+         eptr + 1 >= md->end_subject &&
+         NLBLOCK->nltype == NLTYPE_FIXED &&
+         NLBLOCK->nllen == 2 &&
+         *eptr == NLBLOCK->nl[0])
+       {
+       md->hitend = TRUE;
+       if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
+       }
      /* Fall through */
+     /* Match any single character whatsoever. */
      case OP_ALLANY:
      if (eptr >= md->end_subject)   /* DO NOT merge the eptr++ here; it must */
        {                            /* not be updated before SCHECK_PARTIAL. */
        SCHECK_PARTIAL();
-       MRRETURN(MATCH_NOMATCH);
+       RRETURN(MATCH_NOMATCH);
        }
      eptr++;
-     if (utf8) while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;
+ #ifdef SUPPORT_UTF
+     if (utf) ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);
+ #endif
      ecode++;
      break;
-Line 2200 
 for (;;)
+Line 2311 
 for (;;)
      if (eptr >= md->end_subject)   /* DO NOT merge the eptr++ here; it must */
        {                            /* not be updated before SCHECK_PARTIAL. */
        SCHECK_PARTIAL();
-       MRRETURN(MATCH_NOMATCH);
+       RRETURN(MATCH_NOMATCH);
        }
      eptr++;
      ecode++;
-Line 2210 
 for (;;)
+Line 2321 
 for (;;)
      if (eptr >= md->end_subject)
        {
        SCHECK_PARTIAL();
-       MRRETURN(MATCH_NOMATCH);
+       RRETURN(MATCH_NOMATCH);
        }
      GETCHARINCTEST(c, eptr);
      if (
- #ifdef SUPPORT_UTF8
+ #if defined SUPPORT_UTF || !(defined COMPILE_PCRE8)
         c < 256 &&
  #endif
         (md->ctypes[c] & ctype_digit) != 0
         )
-       MRRETURN(MATCH_NOMATCH);
+       RRETURN(MATCH_NOMATCH);
      ecode++;
      break;
-Line 2227 
 for (;;)
+Line 2338 
 for (;;)
      if (eptr >= md->end_subject)
        {
        SCHECK_PARTIAL();
-       MRRETURN(MATCH_NOMATCH);
+       RRETURN(MATCH_NOMATCH);
        }
      GETCHARINCTEST(c, eptr);
      if (
- #ifdef SUPPORT_UTF8
+ #if defined SUPPORT_UTF || !(defined COMPILE_PCRE8)
-        c >= 256 ||
+        c > 255 ||
  #endif
         (md->ctypes[c] & ctype_digit) == 0
         )
-       MRRETURN(MATCH_NOMATCH);
+       RRETURN(MATCH_NOMATCH);
      ecode++;
      break;
-Line 2244 
 for (;;)
+Line 2355 
 for (;;)
      if (eptr >= md->end_subject)
        {
        SCHECK_PARTIAL();
-       MRRETURN(MATCH_NOMATCH);
+       RRETURN(MATCH_NOMATCH);
        }
      GETCHARINCTEST(c, eptr);
      if (
- #ifdef SUPPORT_UTF8
+ #if defined SUPPORT_UTF || !(defined COMPILE_PCRE8)
         c < 256 &&
  #endif
         (md->ctypes[c] & ctype_space) != 0
         )
-       MRRETURN(MATCH_NOMATCH);
+       RRETURN(MATCH_NOMATCH);
      ecode++;
      break;
-Line 2261 
 for (;;)
+Line 2372 
 for (;;)
      if (eptr >= md->end_subject)
        {
        SCHECK_PARTIAL();
-       MRRETURN(MATCH_NOMATCH);
+       RRETURN(MATCH_NOMATCH);
        }
      GETCHARINCTEST(c, eptr);
      if (
- #ifdef SUPPORT_UTF8
+ #if defined SUPPORT_UTF || !(defined COMPILE_PCRE8)
-        c >= 256 ||
+        c > 255 ||
  #endif
         (md->ctypes[c] & ctype_space) == 0
         )
-       MRRETURN(MATCH_NOMATCH);
+       RRETURN(MATCH_NOMATCH);
      ecode++;
      break;
-Line 2278 
 for (;;)
+Line 2389 
 for (;;)
      if (eptr >= md->end_subject)
        {
        SCHECK_PARTIAL();
-       MRRETURN(MATCH_NOMATCH);
+       RRETURN(MATCH_NOMATCH);
        }
      GETCHARINCTEST(c, eptr);
      if (
- #ifdef SUPPORT_UTF8
+ #if defined SUPPORT_UTF || !(defined COMPILE_PCRE8)
         c < 256 &&
  #endif
         (md->ctypes[c] & ctype_word) != 0
         )
-       MRRETURN(MATCH_NOMATCH);
+       RRETURN(MATCH_NOMATCH);
      ecode++;
      break;
-Line 2295 
 for (;;)
+Line 2406 
 for (;;)
      if (eptr >= md->end_subject)
        {
        SCHECK_PARTIAL();
-       MRRETURN(MATCH_NOMATCH);
+       RRETURN(MATCH_NOMATCH);
        }
      GETCHARINCTEST(c, eptr);
      if (
- #ifdef SUPPORT_UTF8
+ #if defined SUPPORT_UTF || !(defined COMPILE_PCRE8)
-        c >= 256 ||
+        c > 255 ||
  #endif
         (md->ctypes[c] & ctype_word) == 0
         )
-       MRRETURN(MATCH_NOMATCH);
+       RRETURN(MATCH_NOMATCH);
      ecode++;
      break;
-Line 2312 
 for (;;)
+Line 2423 
 for (;;)
      if (eptr >= md->end_subject)
        {
        SCHECK_PARTIAL();
-       MRRETURN(MATCH_NOMATCH);
+       RRETURN(MATCH_NOMATCH);
        }
      GETCHARINCTEST(c, eptr);
      switch(c)
        {
-       default: MRRETURN(MATCH_NOMATCH);
+       default: RRETURN(MATCH_NOMATCH);
-       case 0x000d:
+       case CHAR_CR:
-       if (eptr < md->end_subject && *eptr == 0x0a) eptr++;
+       if (eptr >= md->end_subject)
+         {
+         SCHECK_PARTIAL();
+         }
+       else if (*eptr == CHAR_LF) eptr++;
        break;
-       case 0x000a:
+       case CHAR_LF:
        break;
-       case 0x000b:
+       case CHAR_VT:
-       case 0x000c:
+       case CHAR_FF:
-       case 0x0085:
+       case CHAR_NEL:
+ #ifndef EBCDIC
        case 0x2028:
        case 0x2029:
-       if (md->bsr_anycrlf) MRRETURN(MATCH_NOMATCH);
+ #endif  /* Not EBCDIC */
+       if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);
        break;
        }
      ecode++;
-Line 2341 
 for (;;)
+Line 2458 
 for (;;)
      if (eptr >= md->end_subject)
        {
        SCHECK_PARTIAL();
-       MRRETURN(MATCH_NOMATCH);
+       RRETURN(MATCH_NOMATCH);
        }
      GETCHARINCTEST(c, eptr);
      switch(c)
        {
+       HSPACE_CASES: RRETURN(MATCH_NOMATCH);  /* Byte and multibyte cases */
        default: break;
-       case 0x09:      /* HT */
-       case 0x20:      /* SPACE */
-       case 0xa0:      /* NBSP */
-       case 0x1680:    /* OGHAM SPACE MARK */
-       case 0x180e:    /* MONGOLIAN VOWEL SEPARATOR */
-       case 0x2000:    /* EN QUAD */
-       case 0x2001:    /* EM QUAD */
-       case 0x2002:    /* EN SPACE */
-       case 0x2003:    /* EM SPACE */
-       case 0x2004:    /* THREE-PER-EM SPACE */
-       case 0x2005:    /* FOUR-PER-EM SPACE */
-       case 0x2006:    /* SIX-PER-EM SPACE */
-       case 0x2007:    /* FIGURE SPACE */
-       case 0x2008:    /* PUNCTUATION SPACE */
-       case 0x2009:    /* THIN SPACE */
-       case 0x200A:    /* HAIR SPACE */
-       case 0x202f:    /* NARROW NO-BREAK SPACE */
-       case 0x205f:    /* MEDIUM MATHEMATICAL SPACE */
-       case 0x3000:    /* IDEOGRAPHIC SPACE */
-       MRRETURN(MATCH_NOMATCH);
        }
      ecode++;
      break;
-Line 2375 
 for (;;)
+Line 2473 
 for (;;)
      if (eptr >= md->end_subject)
        {
        SCHECK_PARTIAL();
-       MRRETURN(MATCH_NOMATCH);
+       RRETURN(MATCH_NOMATCH);
        }
      GETCHARINCTEST(c, eptr);
      switch(c)
        {
-       default: MRRETURN(MATCH_NOMATCH);
+       HSPACE_CASES: break;  /* Byte and multibyte cases */
-       case 0x09:      /* HT */
+       default: RRETURN(MATCH_NOMATCH);
-       case 0x20:      /* SPACE */
-       case 0xa0:      /* NBSP */
-       case 0x1680:    /* OGHAM SPACE MARK */
-       case 0x180e:    /* MONGOLIAN VOWEL SEPARATOR */
-       case 0x2000:    /* EN QUAD */
-       case 0x2001:    /* EM QUAD */
-       case 0x2002:    /* EN SPACE */
-       case 0x2003:    /* EM SPACE */
-       case 0x2004:    /* THREE-PER-EM SPACE */
-       case 0x2005:    /* FOUR-PER-EM SPACE */
-       case 0x2006:    /* SIX-PER-EM SPACE */
-       case 0x2007:    /* FIGURE SPACE */
-       case 0x2008:    /* PUNCTUATION SPACE */
-       case 0x2009:    /* THIN SPACE */
-       case 0x200A:    /* HAIR SPACE */
-       case 0x202f:    /* NARROW NO-BREAK SPACE */
-       case 0x205f:    /* MEDIUM MATHEMATICAL SPACE */
-       case 0x3000:    /* IDEOGRAPHIC SPACE */
-       break;
        }
      ecode++;
      break;
-Line 2409 
 for (;;)
+Line 2488 
 for (;;)
      if (eptr >= md->end_subject)
        {
        SCHECK_PARTIAL();
-       MRRETURN(MATCH_NOMATCH);
+       RRETURN(MATCH_NOMATCH);
        }
      GETCHARINCTEST(c, eptr);
      switch(c)
        {
+       VSPACE_CASES: RRETURN(MATCH_NOMATCH);
        default: break;
-       case 0x0a:      /* LF */
-       case 0x0b:      /* VT */
-       case 0x0c:      /* FF */
-       case 0x0d:      /* CR */
-       case 0x85:      /* NEL */
-       case 0x2028:    /* LINE SEPARATOR */
-       case 0x2029:    /* PARAGRAPH SEPARATOR */
-       MRRETURN(MATCH_NOMATCH);
        }
      ecode++;
      break;
-Line 2431 
 for (;;)
+Line 2503 
 for (;;)
      if (eptr >= md->end_subject)
        {
        SCHECK_PARTIAL();
-       MRRETURN(MATCH_NOMATCH);
+       RRETURN(MATCH_NOMATCH);
        }
      GETCHARINCTEST(c, eptr);
      switch(c)
        {
-       default: MRRETURN(MATCH_NOMATCH);
+       VSPACE_CASES: break;
-       case 0x0a:      /* LF */
+       default: RRETURN(MATCH_NOMATCH);
-       case 0x0b:      /* VT */
-       case 0x0c:      /* FF */
-       case 0x0d:      /* CR */
-       case 0x85:      /* NEL */
-       case 0x2028:    /* LINE SEPARATOR */
-       case 0x2029:    /* PARAGRAPH SEPARATOR */
-       break;
        }
      ecode++;
      break;
-Line 2458 
 for (;;)
+Line 2523 
 for (;;)
      if (eptr >= md->end_subject)
        {
        SCHECK_PARTIAL();
-       MRRETURN(MATCH_NOMATCH);
+       RRETURN(MATCH_NOMATCH);
        }
      GETCHARINCTEST(c, eptr);
        {
+       const pcre_uint32 *cp;
        const ucd_record *prop = GET_UCD(c);
        switch(ecode[1])
          {
          case PT_ANY:
-         if (op == OP_NOTPROP) MRRETURN(MATCH_NOMATCH);
+         if (op == OP_NOTPROP) RRETURN(MATCH_NOMATCH);
          break;
          case PT_LAMP:
          if ((prop->chartype == ucp_Lu ||
               prop->chartype == ucp_Ll ||
               prop->chartype == ucp_Lt) == (op == OP_NOTPROP))
-           MRRETURN(MATCH_NOMATCH);
+           RRETURN(MATCH_NOMATCH);
          break;
          case PT_GC:
-         if ((ecode[2] != _pcre_ucp_gentype[prop->chartype]) == (op == OP_PROP))
+         if ((ecode[2] != PRIV(ucp_gentype)[prop->chartype]) == (op == OP_PROP))
-           MRRETURN(MATCH_NOMATCH);
+           RRETURN(MATCH_NOMATCH);
          break;
          case PT_PC:
          if ((ecode[2] != prop->chartype) == (op == OP_PROP))
-           MRRETURN(MATCH_NOMATCH);
+           RRETURN(MATCH_NOMATCH);
          break;
          case PT_SC:
          if ((ecode[2] != prop->script) == (op == OP_PROP))
-           MRRETURN(MATCH_NOMATCH);
+           RRETURN(MATCH_NOMATCH);
          break;
          /* These are specials */
          case PT_ALNUM:
-         if ((_pcre_ucp_gentype[prop->chartype] == ucp_L ||
+         if ((PRIV(ucp_gentype)[prop->chartype] == ucp_L ||
-              _pcre_ucp_gentype[prop->chartype] == ucp_N) == (op == OP_NOTPROP))
+              PRIV(ucp_gentype)[prop->chartype] == ucp_N) == (op == OP_NOTPROP))
-           MRRETURN(MATCH_NOMATCH);
+           RRETURN(MATCH_NOMATCH);
          break;
          case PT_SPACE:    /* Perl space */
-         if ((_pcre_ucp_gentype[prop->chartype] == ucp_Z ||
+         if ((PRIV(ucp_gentype)[prop->chartype] == ucp_Z ||
               c == CHAR_HT || c == CHAR_NL || c == CHAR_FF || c == CHAR_CR)
                 == (op == OP_NOTPROP))
-           MRRETURN(MATCH_NOMATCH);
+           RRETURN(MATCH_NOMATCH);
          break;
          case PT_PXSPACE:  /* POSIX space */
-         if ((_pcre_ucp_gentype[prop->chartype] == ucp_Z ||
+         if ((PRIV(ucp_gentype)[prop->chartype] == ucp_Z ||
               c == CHAR_HT || c == CHAR_NL || c == CHAR_VT ||
               c == CHAR_FF || c == CHAR_CR)
                 == (op == OP_NOTPROP))
-           MRRETURN(MATCH_NOMATCH);
+           RRETURN(MATCH_NOMATCH);
          break;
          case PT_WORD:
-         if ((_pcre_ucp_gentype[prop->chartype] == ucp_L ||
+         if ((PRIV(ucp_gentype)[prop->chartype] == ucp_L ||
-              _pcre_ucp_gentype[prop->chartype] == ucp_N ||
+              PRIV(ucp_gentype)[prop->chartype] == ucp_N ||
               c == CHAR_UNDERSCORE) == (op == OP_NOTPROP))
-           MRRETURN(MATCH_NOMATCH);
+           RRETURN(MATCH_NOMATCH);
+         break;
+         case PT_CLIST:
+         cp = PRIV(ucd_caseless_sets) + prop->caseset;
+         for (;;)
+           {
+           if (c < *cp)
+             { if (op == OP_PROP) { RRETURN(MATCH_NOMATCH); } else break; }
+           if (c == *cp++)
+             { if (op == OP_PROP) break; else { RRETURN(MATCH_NOMATCH); } }
+           }
          break;
          /* This should never occur */
-Line 2539 
 for (;;)
+Line 2616 
 for (;;)
      if (eptr >= md->end_subject)
        {
        SCHECK_PARTIAL();
-       MRRETURN(MATCH_NOMATCH);
+       RRETURN(MATCH_NOMATCH);
        }
-     GETCHARINCTEST(c, eptr);
+     else
-     if (UCD_CATEGORY(c) == ucp_M) MRRETURN(MATCH_NOMATCH);
-     while (eptr < md->end_subject)
        {
-       int len = 1;
+       int lgb, rgb;
-       if (!utf8) c = *eptr; else { GETCHARLEN(c, eptr, len); }
+       GETCHARINCTEST(c, eptr);
-       if (UCD_CATEGORY(c) != ucp_M) break;
+       lgb = UCD_GRAPHBREAK(c);
-       eptr += len;
+       while (eptr < md->end_subject)
+         {
+         int len = 1;
+         if (!utf) c = *eptr; else { GETCHARLEN(c, eptr, len); }
+         rgb = UCD_GRAPHBREAK(c);
+         if ((PRIV(ucp_gbtable)[lgb] & (1 << rgb)) == 0) break;
+         lgb = rgb;
+         eptr += len;
+         }
        }
+     CHECK_PARTIAL();
      ecode++;
      break;
- #endif
+ #endif  /* SUPPORT_UCP */
      /* Match a back reference, possibly repeatedly. Look past the end of the
-Line 2567 
 for (;;)
+Line 2651 
 for (;;)
      case OP_REFI:
      caseless = op == OP_REFI;
      offset = GET2(ecode, 1) << 1;               /* Doubled ref number */
-     ecode += 3;
+     ecode += 1 + IMM2_SIZE;
      /* If the reference is unset, there are two possibilities:
-Line 2607 
 for (;;)
+Line 2691 
 for (;;)
        case OP_CRMINRANGE:
        minimize = (*ecode == OP_CRMINRANGE);
        min = GET2(ecode, 1);
-       max = GET2(ecode, 3);
+       max = GET2(ecode, 1 + IMM2_SIZE);
        if (max == 0) max = INT_MAX;
-       ecode += 5;
+       ecode += 1 + 2 * IMM2_SIZE;
        break;
        default:               /* No repeat follows */
        if ((length = match_ref(offset, eptr, length, md, caseless)) < 0)
          {
+         if (length == -2) eptr = md->end_subject;   /* Partial match */
          CHECK_PARTIAL();
-         MRRETURN(MATCH_NOMATCH);
+         RRETURN(MATCH_NOMATCH);
          }
        eptr += length;
        continue;              /* With the main loop */
        }
      /* Handle repeated back references. If the length of the reference is
-     zero, just continue with the main loop. */
+     zero, just continue with the main loop. If the length is negative, it
+     means the reference is unset in non-Java-compatible mode. If the minimum is
+     zero, we can continue at the same level without recursion. For any other
+     minimum, carrying on will result in NOMATCH. */
      if (length == 0) continue;
+     if (length < 0 && min == 0) continue;
      /* First, ensure the minimum number of matches are present. We get back
      the length of the reference string explicitly rather than passing the
-Line 2636 
 for (;;)
+Line 2725 
 for (;;)
        int slength;
        if ((slength = match_ref(offset, eptr, length, md, caseless)) < 0)
          {
+         if (slength == -2) eptr = md->end_subject;   /* Partial match */
          CHECK_PARTIAL();
-         MRRETURN(MATCH_NOMATCH);
+         RRETURN(MATCH_NOMATCH);
          }
        eptr += slength;
        }
-Line 2656 
 for (;;)
+Line 2746 
 for (;;)
          int slength;
          RMATCH(eptr, ecode, offset_top, md, eptrb, RM14);
          if (rrc != MATCH_NOMATCH) RRETURN(rrc);
-         if (fi >= max) MRRETURN(MATCH_NOMATCH);
+         if (fi >= max) RRETURN(MATCH_NOMATCH);
          if ((slength = match_ref(offset, eptr, length, md, caseless)) < 0)
            {
+           if (slength == -2) eptr = md->end_subject;   /* Partial match */
            CHECK_PARTIAL();
-           MRRETURN(MATCH_NOMATCH);
+           RRETURN(MATCH_NOMATCH);
            }
          eptr += slength;
          }
-Line 2677 
 for (;;)
+Line 2768 
 for (;;)
          int slength;
          if ((slength = match_ref(offset, eptr, length, md, caseless)) < 0)
            {
-           CHECK_PARTIAL();
+           /* Can't use CHECK_PARTIAL because we don't want to update eptr in
+           the soft partial matching case. */
+           if (slength == -2 && md->partial != 0 &&
+               md->end_subject > md->start_used_ptr)
+             {
+             md->hitend = TRUE;
+             if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
+             }
            break;
            }
          eptr += slength;
          }
        while (eptr >= pp)
          {
          RMATCH(eptr, ecode, offset_top, md, eptrb, RM15);
          if (rrc != MATCH_NOMATCH) RRETURN(rrc);
          eptr -= length;
          }
-       MRRETURN(MATCH_NOMATCH);
+       RRETURN(MATCH_NOMATCH);
        }
      /* Control never gets here */
-Line 2706 
 for (;;)
+Line 2806 
 for (;;)
      case OP_NCLASS:
      case OP_CLASS:
        {
+       /* The data variable is saved across frames, so the byte map needs to
+       be stored there. */
+ #define BYTE_MAP ((pcre_uint8 *)data)
        data = ecode + 1;                /* Save for matching */
-       ecode += 33;                     /* Advance past the item */
+       ecode += 1 + (32 / sizeof(pcre_uchar)); /* Advance past the item */
        switch (*ecode)
          {
-Line 2728 
 for (;;)
+Line 2831 
 for (;;)
          case OP_CRMINRANGE:
          minimize = (*ecode == OP_CRMINRANGE);
          min = GET2(ecode, 1);
-         max = GET2(ecode, 3);
+         max = GET2(ecode, 1 + IMM2_SIZE);
          if (max == 0) max = INT_MAX;
-         ecode += 5;
+         ecode += 1 + 2 * IMM2_SIZE;
          break;
          default:               /* No repeat follows */
-Line 2740 
 for (;;)
+Line 2843 
 for (;;)
        /* First, ensure the minimum number of matches are present. */
- #ifdef SUPPORT_UTF8
+ #ifdef SUPPORT_UTF
-       /* UTF-8 mode */
+       if (utf)
-       if (utf8)
          {
          for (i = 1; i <= min; i++)
            {
            if (eptr >= md->end_subject)
              {
              SCHECK_PARTIAL();
-             MRRETURN(MATCH_NOMATCH);
+             RRETURN(MATCH_NOMATCH);
              }
            GETCHARINC(c, eptr);
            if (c > 255)
              {
-             if (op == OP_CLASS) MRRETURN(MATCH_NOMATCH);
+             if (op == OP_CLASS) RRETURN(MATCH_NOMATCH);
              }
            else
-             {
+             if ((BYTE_MAP[c/8] & (1 << (c&7))) == 0) RRETURN(MATCH_NOMATCH);
-             if ((data[c/8] & (1 << (c&7))) == 0) MRRETURN(MATCH_NOMATCH);
-             }
            }
          }
        else
  #endif
-       /* Not UTF-8 mode */
+       /* Not UTF mode */
          {
          for (i = 1; i <= min; i++)
            {
            if (eptr >= md->end_subject)
              {
              SCHECK_PARTIAL();
-             MRRETURN(MATCH_NOMATCH);
+             RRETURN(MATCH_NOMATCH);
              }
            c = *eptr++;
-           if ((data[c/8] & (1 << (c&7))) == 0) MRRETURN(MATCH_NOMATCH);
+ #ifndef COMPILE_PCRE8
+           if (c > 255)
+             {
+             if (op == OP_CLASS) RRETURN(MATCH_NOMATCH);
+             }
+           else
+ #endif
+             if ((BYTE_MAP[c/8] & (1 << (c&7))) == 0) RRETURN(MATCH_NOMATCH);
            }
          }
-Line 2788 
 for (;;)
+Line 2895 
 for (;;)
        if (minimize)
          {
- #ifdef SUPPORT_UTF8
+ #ifdef SUPPORT_UTF
-         /* UTF-8 mode */
+         if (utf)
-         if (utf8)
            {
            for (fi = min;; fi++)
              {
              RMATCH(eptr, ecode, offset_top, md, eptrb, RM16);
              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
-             if (fi >= max) MRRETURN(MATCH_NOMATCH);
+             if (fi >= max) RRETURN(MATCH_NOMATCH);
              if (eptr >= md->end_subject)
                {
                SCHECK_PARTIAL();
-               MRRETURN(MATCH_NOMATCH);
+               RRETURN(MATCH_NOMATCH);
                }
              GETCHARINC(c, eptr);
              if (c > 255)
                {
-               if (op == OP_CLASS) MRRETURN(MATCH_NOMATCH);
+               if (op == OP_CLASS) RRETURN(MATCH_NOMATCH);
                }
              else
-               {
+               if ((BYTE_MAP[c/8] & (1 << (c&7))) == 0) RRETURN(MATCH_NOMATCH);
-               if ((data[c/8] & (1 << (c&7))) == 0) MRRETURN(MATCH_NOMATCH);
-               }
              }
            }
          else
  #endif
-         /* Not UTF-8 mode */
+         /* Not UTF mode */
            {
            for (fi = min;; fi++)
              {
              RMATCH(eptr, ecode, offset_top, md, eptrb, RM17);
              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
-             if (fi >= max) MRRETURN(MATCH_NOMATCH);
+             if (fi >= max) RRETURN(MATCH_NOMATCH);
              if (eptr >= md->end_subject)
                {
                SCHECK_PARTIAL();
-               MRRETURN(MATCH_NOMATCH);
+               RRETURN(MATCH_NOMATCH);
                }
              c = *eptr++;
-             if ((data[c/8] & (1 << (c&7))) == 0) MRRETURN(MATCH_NOMATCH);
+ #ifndef COMPILE_PCRE8
+             if (c > 255)
+               {
+               if (op == OP_CLASS) RRETURN(MATCH_NOMATCH);
+               }
+             else
+ #endif
+               if ((BYTE_MAP[c/8] & (1 << (c&7))) == 0) RRETURN(MATCH_NOMATCH);
              }
            }
          /* Control never gets here */
-Line 2840 
 for (;;)
+Line 2951 
 for (;;)
          {
          pp = eptr;
- #ifdef SUPPORT_UTF8
+ #ifdef SUPPORT_UTF
-         /* UTF-8 mode */
+         if (utf)
-         if (utf8)
            {
            for (i = min; i < max; i++)
              {
-Line 2858 
 for (;;)
+Line 2968 
 for (;;)
                if (op == OP_CLASS) break;
                }
              else
-               {
+               if ((BYTE_MAP[c/8] & (1 << (c&7))) == 0) break;
-               if ((data[c/8] & (1 << (c&7))) == 0) break;
-               }
              eptr += len;
              }
            for (;;)
-Line 2873 
 for (;;)
+Line 2981 
 for (;;)
            }
          else
  #endif
-           /* Not UTF-8 mode */
+           /* Not UTF mode */
            {
            for (i = min; i < max; i++)
              {
-Line 2883 
 for (;;)
+Line 2991 
 for (;;)
                break;
                }
              c = *eptr;
-             if ((data[c/8] & (1 << (c&7))) == 0) break;
+ #ifndef COMPILE_PCRE8
+             if (c > 255)
+               {
+               if (op == OP_CLASS) break;
+               }
+             else
+ #endif
+               if ((BYTE_MAP[c/8] & (1 << (c&7))) == 0) break;
              eptr++;
              }
            while (eptr >= pp)
-Line 2894 
 for (;;)
+Line 3009 
 for (;;)
              }
            }
-         MRRETURN(MATCH_NOMATCH);
+         RRETURN(MATCH_NOMATCH);
          }
+ #undef BYTE_MAP
        }
      /* Control never gets here */
-Line 2904 
 for (;;)
+Line 3020 
 for (;;)
      when UTF-8 mode mode is supported. Nevertheless, we may not be in UTF-8
      mode, because Unicode properties are supported in non-UTF-8 mode. */
- #ifdef SUPPORT_UTF8
+ #if defined SUPPORT_UTF || !defined COMPILE_PCRE8
      case OP_XCLASS:
        {
        data = ecode + 1 + LINK_SIZE;                /* Save for matching */
-Line 2929 
 for (;;)
+Line 3045 
 for (;;)
          case OP_CRMINRANGE:
          minimize = (*ecode == OP_CRMINRANGE);
          min = GET2(ecode, 1);
-         max = GET2(ecode, 3);
+         max = GET2(ecode, 1 + IMM2_SIZE);
          if (max == 0) max = INT_MAX;
-         ecode += 5;
+         ecode += 1 + 2 * IMM2_SIZE;
          break;
          default:               /* No repeat follows */
-Line 2946 
 for (;;)
+Line 3062 
 for (;;)
          if (eptr >= md->end_subject)
            {
            SCHECK_PARTIAL();
-           MRRETURN(MATCH_NOMATCH);
+           RRETURN(MATCH_NOMATCH);
            }
          GETCHARINCTEST(c, eptr);
-         if (!_pcre_xclass(c, data)) MRRETURN(MATCH_NOMATCH);
+         if (!PRIV(xclass)(c, data, utf)) RRETURN(MATCH_NOMATCH);
          }
        /* If max == min we can continue with the main loop without the
-Line 2966 
 for (;;)
+Line 3082 
 for (;;)
            {
            RMATCH(eptr, ecode, offset_top, md, eptrb, RM20);
            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
-           if (fi >= max) MRRETURN(MATCH_NOMATCH);
+           if (fi >= max) RRETURN(MATCH_NOMATCH);
            if (eptr >= md->end_subject)
              {
              SCHECK_PARTIAL();
-             MRRETURN(MATCH_NOMATCH);
+             RRETURN(MATCH_NOMATCH);
              }
            GETCHARINCTEST(c, eptr);
-           if (!_pcre_xclass(c, data)) MRRETURN(MATCH_NOMATCH);
+           if (!PRIV(xclass)(c, data, utf)) RRETURN(MATCH_NOMATCH);
            }
          /* Control never gets here */
          }
-Line 2991 
 for (;;)
+Line 3107 
 for (;;)
              SCHECK_PARTIAL();
              break;
              }
+ #ifdef SUPPORT_UTF
            GETCHARLENTEST(c, eptr, len);
-           if (!_pcre_xclass(c, data)) break;
+ #else
+           c = *eptr;
+ #endif
+           if (!PRIV(xclass)(c, data, utf)) break;
            eptr += len;
            }
          for(;;)
-Line 3000 
 for (;;)
+Line 3120 
 for (;;)
            RMATCH(eptr, ecode, offset_top, md, eptrb, RM21);
            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
            if (eptr-- == pp) break;        /* Stop if tried at original pos */
-           if (utf8) BACKCHAR(eptr);
+ #ifdef SUPPORT_UTF
+           if (utf) BACKCHAR(eptr);
+ #endif
            }
-         MRRETURN(MATCH_NOMATCH);
+         RRETURN(MATCH_NOMATCH);
          }
        /* Control never gets here */
-Line 3012 
 for (;;)
+Line 3134 
 for (;;)
      /* Match a single character, casefully */
      case OP_CHAR:
- #ifdef SUPPORT_UTF8
+ #ifdef SUPPORT_UTF
-     if (utf8)
+     if (utf)
        {
        length = 1;
        ecode++;
-Line 3021 
 for (;;)
+Line 3143 
 for (;;)
        if (length > md->end_subject - eptr)
          {
          CHECK_PARTIAL();             /* Not SCHECK_PARTIAL() */
-         MRRETURN(MATCH_NOMATCH);
+         RRETURN(MATCH_NOMATCH);
          }
-       while (length-- > 0) if (*ecode++ != *eptr++) MRRETURN(MATCH_NOMATCH);
+       while (length-- > 0) if (*ecode++ != *eptr++) RRETURN(MATCH_NOMATCH);
        }
      else
  #endif
+     /* Not UTF mode */
-     /* Non-UTF-8 mode */
        {
        if (md->end_subject - eptr < 1)
          {
          SCHECK_PARTIAL();            /* This one can use SCHECK_PARTIAL() */
-         MRRETURN(MATCH_NOMATCH);
+         RRETURN(MATCH_NOMATCH);
          }
-       if (ecode[1] != *eptr++) MRRETURN(MATCH_NOMATCH);
+       if (ecode[1] != *eptr++) RRETURN(MATCH_NOMATCH);
        ecode += 2;
        }
      break;
-     /* Match a single character, caselessly */
+     /* Match a single character, caselessly. If we are at the end of the
+     subject, give up immediately. */
      case OP_CHARI:
- #ifdef SUPPORT_UTF8
+     if (eptr >= md->end_subject)
-     if (utf8)
+       {
+       SCHECK_PARTIAL();
+       RRETURN(MATCH_NOMATCH);
+       }
+ #ifdef SUPPORT_UTF
+     if (utf)
        {
        length = 1;
        ecode++;
        GETCHARLEN(fc, ecode, length);
-       if (length > md->end_subject - eptr)
-         {
-         CHECK_PARTIAL();             /* Not SCHECK_PARTIAL() */
-         MRRETURN(MATCH_NOMATCH);
-         }
        /* If the pattern character's value is < 128, we have only one byte, and
-       can use the fast lookup table. */
+       we know that its other case must also be one byte long, so we can use the
+       fast lookup table. We know that there is at least one byte left in the
+       subject. */
        if (fc < 128)
          {
-         if (md->lcc[*ecode++] != md->lcc[*eptr++]) MRRETURN(MATCH_NOMATCH);
+         if (md->lcc[fc]
+             != TABLE_GET(*eptr, md->lcc, *eptr)) RRETURN(MATCH_NOMATCH);
+         ecode++;
+         eptr++;
          }
-       /* Otherwise we must pick up the subject character */
+       /* Otherwise we must pick up the subject character. Note that we cannot
+       use the value of "length" to check for sufficient bytes left, because the
+       other case of the character may have more or fewer bytes.  */
        else
          {
-         unsigned int dc;
+         pcre_uint32 dc;
          GETCHARINC(dc, eptr);
          ecode += length;
-Line 3080 
 for (;;)
+Line 3209 
 for (;;)
  #ifdef SUPPORT_UCP
            if (dc != UCD_OTHERCASE(fc))
  #endif
-             MRRETURN(MATCH_NOMATCH);
+             RRETURN(MATCH_NOMATCH);
            }
          }
        }
      else
- #endif   /* SUPPORT_UTF8 */
+ #endif   /* SUPPORT_UTF */
-     /* Non-UTF-8 mode */
+     /* Not UTF mode */
        {
-       if (md->end_subject - eptr < 1)
+       if (TABLE_GET(ecode[1], md->lcc, ecode[1])
-         {
+           != TABLE_GET(*eptr, md->lcc, *eptr)) RRETURN(MATCH_NOMATCH);
-         SCHECK_PARTIAL();            /* This one can use SCHECK_PARTIAL() */
+       eptr++;
-         MRRETURN(MATCH_NOMATCH);
-         }
-       if (md->lcc[ecode[1]] != md->lcc[*eptr++]) MRRETURN(MATCH_NOMATCH);
        ecode += 2;
        }
      break;
-Line 3104 
 for (;;)
+Line 3230 
 for (;;)
      case OP_EXACT:
      case OP_EXACTI:
      min = max = GET2(ecode, 1);
-     ecode += 3;
+     ecode += 1 + IMM2_SIZE;
      goto REPEATCHAR;
      case OP_POSUPTO:
-Line 3119 
 for (;;)
+Line 3245 
 for (;;)
      min = 0;
      max = GET2(ecode, 1);
      minimize = *ecode == OP_MINUPTO || *ecode == OP_MINUPTOI;
-     ecode += 3;
+     ecode += 1 + IMM2_SIZE;
      goto REPEATCHAR;
      case OP_POSSTAR:
-Line 3167 
 for (;;)
+Line 3293 
 for (;;)
      /* Common code for all repeated single-character matches. */
      REPEATCHAR:
- #ifdef SUPPORT_UTF8
+ #ifdef SUPPORT_UTF
-     if (utf8)
+     if (utf)
        {
        length = 1;
        charptr = ecode;
-Line 3184 
 for (;;)
+Line 3310 
 for (;;)
          unsigned int othercase;
          if (op >= OP_STARI &&     /* Caseless */
              (othercase = UCD_OTHERCASE(fc)) != fc)
-           oclength = _pcre_ord2utf8(othercase, occhars);
+           oclength = PRIV(ord2utf)(othercase, occhars);
          else oclength = 0;
  #endif  /* SUPPORT_UCP */
          for (i = 1; i <= min; i++)
            {
            if (eptr <= md->end_subject - length &&
-             memcmp(eptr, charptr, length) == 0) eptr += length;
+             memcmp(eptr, charptr, IN_UCHARS(length)) == 0) eptr += length;
  #ifdef SUPPORT_UCP
            else if (oclength > 0 &&
                     eptr <= md->end_subject - oclength &&
-                    memcmp(eptr, occhars, oclength) == 0) eptr += oclength;
+                    memcmp(eptr, occhars, IN_UCHARS(oclength)) == 0) eptr += oclength;
  #endif  /* SUPPORT_UCP */
            else
              {
              CHECK_PARTIAL();
-             MRRETURN(MATCH_NOMATCH);
+             RRETURN(MATCH_NOMATCH);
              }
            }
-Line 3212 
 for (;;)
+Line 3338 
 for (;;)
              {
              RMATCH(eptr, ecode, offset_top, md, eptrb, RM22);
              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
-             if (fi >= max) MRRETURN(MATCH_NOMATCH);
+             if (fi >= max) RRETURN(MATCH_NOMATCH);
              if (eptr <= md->end_subject - length &&
-               memcmp(eptr, charptr, length) == 0) eptr += length;
+               memcmp(eptr, charptr, IN_UCHARS(length)) == 0) eptr += length;
  #ifdef SUPPORT_UCP
              else if (oclength > 0 &&
                       eptr <= md->end_subject - oclength &&
-                      memcmp(eptr, occhars, oclength) == 0) eptr += oclength;
+                      memcmp(eptr, occhars, IN_UCHARS(oclength)) == 0) eptr += oclength;
  #endif  /* SUPPORT_UCP */
              else
                {
                CHECK_PARTIAL();
-               MRRETURN(MATCH_NOMATCH);
+               RRETURN(MATCH_NOMATCH);
                }
              }
            /* Control never gets here */
-Line 3235 
 for (;;)
+Line 3361 
 for (;;)
            for (i = min; i < max; i++)
              {
              if (eptr <= md->end_subject - length &&
-                 memcmp(eptr, charptr, length) == 0) eptr += length;
+                 memcmp(eptr, charptr, IN_UCHARS(length)) == 0) eptr += length;
  #ifdef SUPPORT_UCP
              else if (oclength > 0 &&
                       eptr <= md->end_subject - oclength &&
-                      memcmp(eptr, occhars, oclength) == 0) eptr += oclength;
+                      memcmp(eptr, occhars, IN_UCHARS(oclength)) == 0) eptr += oclength;
  #endif  /* SUPPORT_UCP */
              else
                {
-Line 3254 
 for (;;)
+Line 3380 
 for (;;)
              {
              RMATCH(eptr, ecode, offset_top, md, eptrb, RM23);
              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
-             if (eptr == pp) { MRRETURN(MATCH_NOMATCH); }
+             if (eptr == pp) { RRETURN(MATCH_NOMATCH); }
  #ifdef SUPPORT_UCP
              eptr--;
              BACKCHAR(eptr);
-Line 3271 
 for (;;)
+Line 3397 
 for (;;)
        value of fc will always be < 128. */
        }
      else
- #endif  /* SUPPORT_UTF8 */
+ #endif  /* SUPPORT_UTF */
+       /* When not in UTF-8 mode, load a single-byte character. */
+       fc = *ecode++;
-     /* 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
+     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
      caseful cases, for speed, since matching characters is likely to be quite
      common. First, ensure the minimum number of matches are present. If min =
      max, continue at the same level without recursing. Otherwise, if
-Line 3287 
 for (;;)
+Line 3411 
 for (;;)
      maximizing, find the maximum number of characters and work backwards. */
      DPRINTF(("matching %c{%d,%d} against subject %.*s\n", fc, min, max,
-       max, eptr));
+       max, (char *)eptr));
      if (op >= OP_STARI)  /* Caseless */
        {
-       fc = md->lcc[fc];
+ #ifdef COMPILE_PCRE8
+       /* fc must be < 128 if UTF is enabled. */
+       foc = md->fcc[fc];
+ #else
+ #ifdef SUPPORT_UTF
+ #ifdef SUPPORT_UCP
+       if (utf && fc > 127)
+         foc = UCD_OTHERCASE(fc);
+ #else
+       if (utf && fc > 127)
+         foc = fc;
+ #endif /* SUPPORT_UCP */
+       else
+ #endif /* SUPPORT_UTF */
+         foc = TABLE_GET(fc, md->fcc, fc);
+ #endif /* COMPILE_PCRE8 */
        for (i = 1; i <= min; i++)
          {
          if (eptr >= md->end_subject)
            {
            SCHECK_PARTIAL();
-           MRRETURN(MATCH_NOMATCH);
+           RRETURN(MATCH_NOMATCH);
            }
-         if (fc != md->lcc[*eptr++]) MRRETURN(MATCH_NOMATCH);
+         if (fc != *eptr && foc != *eptr) RRETURN(MATCH_NOMATCH);
+         eptr++;
          }
        if (min == max) continue;
        if (minimize)
-Line 3308 
 for (;;)
+Line 3449 
 for (;;)
            {
            RMATCH(eptr, ecode, offset_top, md, eptrb, RM24);
            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
-           if (fi >= max) MRRETURN(MATCH_NOMATCH);
+           if (fi >= max) RRETURN(MATCH_NOMATCH);
            if (eptr >= md->end_subject)
              {
              SCHECK_PARTIAL();
-             MRRETURN(MATCH_NOMATCH);
+             RRETURN(MATCH_NOMATCH);
              }
-           if (fc != md->lcc[*eptr++]) MRRETURN(MATCH_NOMATCH);
+           if (fc != *eptr && foc != *eptr) RRETURN(MATCH_NOMATCH);
+           eptr++;
            }
          /* Control never gets here */
          }
-Line 3328 
 for (;;)
+Line 3470 
 for (;;)
              SCHECK_PARTIAL();
              break;
              }
-           if (fc != md->lcc[*eptr]) break;
+           if (fc != *eptr && foc != *eptr) break;
            eptr++;
            }
-Line 3340 
 for (;;)
+Line 3482 
 for (;;)
            eptr--;
            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
            }
-         MRRETURN(MATCH_NOMATCH);
+         RRETURN(MATCH_NOMATCH);
          }
        /* Control never gets here */
        }
-Line 3354 
 for (;;)
+Line 3496 
 for (;;)
          if (eptr >= md->end_subject)
            {
            SCHECK_PARTIAL();
-           MRRETURN(MATCH_NOMATCH);
+           RRETURN(MATCH_NOMATCH);
            }
-         if (fc != *eptr++) MRRETURN(MATCH_NOMATCH);
+         if (fc != *eptr++) RRETURN(MATCH_NOMATCH);
          }
        if (min == max) continue;
-Line 3367 
 for (;;)
+Line 3509 
 for (;;)
            {
            RMATCH(eptr, ecode, offset_top, md, eptrb, RM26);
            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
-           if (fi >= max) MRRETURN(MATCH_NOMATCH);
+           if (fi >= max) RRETURN(MATCH_NOMATCH);
            if (eptr >= md->end_subject)
              {
              SCHECK_PARTIAL();
-             MRRETURN(MATCH_NOMATCH);
+             RRETURN(MATCH_NOMATCH);
              }
-           if (fc != *eptr++) MRRETURN(MATCH_NOMATCH);
+           if (fc != *eptr++) RRETURN(MATCH_NOMATCH);
            }
          /* Control never gets here */
          }
-Line 3398 
 for (;;)
+Line 3540 
 for (;;)
            eptr--;
            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
            }
-         MRRETURN(MATCH_NOMATCH);
+         RRETURN(MATCH_NOMATCH);
          }
        }
      /* Control never gets here */
-Line 3411 
 for (;;)
+Line 3553 
 for (;;)
      if (eptr >= md->end_subject)
        {
        SCHECK_PARTIAL();
-       MRRETURN(MATCH_NOMATCH);
+       RRETURN(MATCH_NOMATCH);
        }
-     ecode++;
+ #ifdef SUPPORT_UTF
-     GETCHARINCTEST(c, eptr);
+     if (utf)
-     if (op == OP_NOTI)         /* The caseless case */
        {
- #ifdef SUPPORT_UTF8
+       register pcre_uint32 ch, och;
-       if (c < 256)
- #endif
+       ecode++;
-       c = md->lcc[c];
+       GETCHARINC(ch, ecode);
-       if (md->lcc[*ecode++] == c) MRRETURN(MATCH_NOMATCH);
+       GETCHARINC(c, eptr);
+       if (op == OP_NOT)
+         {
+         if (ch == c) RRETURN(MATCH_NOMATCH);
+         }
+       else
+         {
+ #ifdef SUPPORT_UCP
+         if (ch > 127)
+           och = UCD_OTHERCASE(ch);
+ #else
+         if (ch > 127)
+           och = ch;
+ #endif /* SUPPORT_UCP */
+         else
+           och = TABLE_GET(ch, md->fcc, ch);
+         if (ch == c || och == c) RRETURN(MATCH_NOMATCH);
+         }
        }
-     else    /* Caseful */
+     else
+ #endif
        {
-       if (*ecode++ == c) MRRETURN(MATCH_NOMATCH);
+       register pcre_uint32 ch = ecode[1];
+       c = *eptr++;
+       if (ch == c || (op == OP_NOTI && TABLE_GET(ch, md->fcc, ch) == c))
+         RRETURN(MATCH_NOMATCH);
+       ecode += 2;
        }
      break;
-Line 3439 
 for (;;)
+Line 3603 
 for (;;)
      case OP_NOTEXACT:
      case OP_NOTEXACTI:
      min = max = GET2(ecode, 1);
-     ecode += 3;
+     ecode += 1 + IMM2_SIZE;
      goto REPEATNOTCHAR;
      case OP_NOTUPTO:
-Line 3449 
 for (;;)
+Line 3613 
 for (;;)
      min = 0;
      max = GET2(ecode, 1);
      minimize = *ecode == OP_NOTMINUPTO || *ecode == OP_NOTMINUPTOI;
-     ecode += 3;
+     ecode += 1 + IMM2_SIZE;
      goto REPEATNOTCHAR;
      case OP_NOTPOSSTAR:
-Line 3481 
 for (;;)
+Line 3645 
 for (;;)
      possessive = TRUE;
      min = 0;
      max = GET2(ecode, 1);
-     ecode += 3;
+     ecode += 1 + IMM2_SIZE;
      goto REPEATNOTCHAR;
      case OP_NOTSTAR:
-Line 3505 
 for (;;)
+Line 3669 
 for (;;)
      /* Common code for all repeated single-byte matches. */
      REPEATNOTCHAR:
-     fc = *ecode++;
+     GETCHARINCTEST(fc, ecode);
      /* The code is duplicated for the caseless and caseful cases, for speed,
      since matching characters is likely to be quite common. First, ensure the
-Line 3516 
 for (;;)
+Line 3680 
 for (;;)
      characters and work backwards. */
      DPRINTF(("negative matching %c{%d,%d} against subject %.*s\n", fc, min, max,
-       max, eptr));
+       max, (char *)eptr));
      if (op >= OP_NOTSTARI)     /* Caseless */
        {
-       fc = md->lcc[fc];
+ #ifdef SUPPORT_UTF
+ #ifdef SUPPORT_UCP
+       if (utf && fc > 127)
+         foc = UCD_OTHERCASE(fc);
+ #else
+       if (utf && fc > 127)
+         foc = fc;
+ #endif /* SUPPORT_UCP */
+       else
+ #endif /* SUPPORT_UTF */
+         foc = TABLE_GET(fc, md->fcc, fc);
- #ifdef SUPPORT_UTF8
+ #ifdef SUPPORT_UTF
-       /* UTF-8 mode */
+       if (utf)
-       if (utf8)
          {
          register unsigned int d;
          for (i = 1; i <= min; i++)
-Line 3532 
 for (;;)
+Line 3705 
 for (;;)
            if (eptr >= md->end_subject)
              {
              SCHECK_PARTIAL();
-             MRRETURN(MATCH_NOMATCH);
+             RRETURN(MATCH_NOMATCH);
              }
            GETCHARINC(d, eptr);
-           if (d < 256) d = md->lcc[d];
+           if (fc == d || (unsigned int)foc == d) RRETURN(MATCH_NOMATCH);
-           if (fc == d) MRRETURN(MATCH_NOMATCH);
            }
          }
        else
  #endif
+       /* Not UTF mode */
-       /* Not UTF-8 mode */
          {
          for (i = 1; i <= min; i++)
            {
            if (eptr >= md->end_subject)
              {
              SCHECK_PARTIAL();
-             MRRETURN(MATCH_NOMATCH);
+             RRETURN(MATCH_NOMATCH);
              }
-           if (fc == md->lcc[*eptr++]) MRRETURN(MATCH_NOMATCH);
+           if (fc == *eptr || foc == *eptr) RRETURN(MATCH_NOMATCH);
+           eptr++;
            }
          }
-Line 3559 
 for (;;)
+Line 3731 
 for (;;)
        if (minimize)
          {
- #ifdef SUPPORT_UTF8
+ #ifdef SUPPORT_UTF
-         /* UTF-8 mode */
+         if (utf)
-         if (utf8)
            {
            register unsigned int d;
            for (fi = min;; fi++)
              {
              RMATCH(eptr, ecode, offset_top, md, eptrb, RM28);
              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
-             if (fi >= max) MRRETURN(MATCH_NOMATCH);
+             if (fi >= max) RRETURN(MATCH_NOMATCH);
              if (eptr >= md->end_subject)
                {
                SCHECK_PARTIAL();
-               MRRETURN(MATCH_NOMATCH);
+               RRETURN(MATCH_NOMATCH);
                }
              GETCHARINC(d, eptr);
-             if (d < 256) d = md->lcc[d];
+             if (fc == d || (unsigned int)foc == d) RRETURN(MATCH_NOMATCH);
-             if (fc == d) MRRETURN(MATCH_NOMATCH);
              }
            }
          else
  #endif
-         /* Not UTF-8 mode */
+         /* Not UTF mode */
            {
            for (fi = min;; fi++)
              {
              RMATCH(eptr, ecode, offset_top, md, eptrb, RM29);
              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
-             if (fi >= max) MRRETURN(MATCH_NOMATCH);
+             if (fi >= max) RRETURN(MATCH_NOMATCH);
              if (eptr >= md->end_subject)
                {
                SCHECK_PARTIAL();
-               MRRETURN(MATCH_NOMATCH);
+               RRETURN(MATCH_NOMATCH);
                }
-             if (fc == md->lcc[*eptr++]) MRRETURN(MATCH_NOMATCH);
+             if (fc == *eptr || foc == *eptr) RRETURN(MATCH_NOMATCH);
+             eptr++;
              }
            }
          /* Control never gets here */
-Line 3605 
 for (;;)
+Line 3776 
 for (;;)
          {
          pp = eptr;
- #ifdef SUPPORT_UTF8
+ #ifdef SUPPORT_UTF
-         /* UTF-8 mode */
+         if (utf)
-         if (utf8)
            {
            register unsigned int d;
            for (i = min; i < max; i++)
-Line 3619 
 for (;;)
+Line 3789 
 for (;;)
                break;
                }
              GETCHARLEN(d, eptr, len);
-             if (d < 256) d = md->lcc[d];
+             if (fc == d || (unsigned int)foc == d) break;
-             if (fc == d) break;
              eptr += len;
              }
-         if (possessive) continue;
+           if (possessive) continue;
-         for(;;)
+           for(;;)
              {
              RMATCH(eptr, ecode, offset_top, md, eptrb, RM30);
              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
-Line 3634 
 for (;;)
+Line 3803 
 for (;;)
            }
          else
  #endif
-         /* Not UTF-8 mode */
+         /* Not UTF mode */
            {
            for (i = min; i < max; i++)
              {
-Line 3643 
 for (;;)
+Line 3812 
 for (;;)
                SCHECK_PARTIAL();
                break;
                }
-             if (fc == md->lcc[*eptr]) break;
+             if (fc == *eptr || foc == *eptr) break;
              eptr++;
              }
            if (possessive) continue;
-Line 3655 
 for (;;)
+Line 3824 
 for (;;)
              }
            }
-         MRRETURN(MATCH_NOMATCH);
+         RRETURN(MATCH_NOMATCH);
          }
        /* Control never gets here */
        }
-Line 3664 
 for (;;)
+Line 3833 
 for (;;)
      else
        {
- #ifdef SUPPORT_UTF8
+ #ifdef SUPPORT_UTF
-       /* UTF-8 mode */
+       if (utf)
-       if (utf8)
          {
          register unsigned int d;
          for (i = 1; i <= min; i++)
-Line 3674 
 for (;;)
+Line 3842 
 for (;;)
            if (eptr >= md->end_subject)
              {
              SCHECK_PARTIAL();
-             MRRETURN(MATCH_NOMATCH);
+             RRETURN(MATCH_NOMATCH);
              }
            GETCHARINC(d, eptr);
-           if (fc == d) MRRETURN(MATCH_NOMATCH);
+           if (fc == d) RRETURN(MATCH_NOMATCH);
            }
          }
        else
  #endif
-       /* Not UTF-8 mode */
+       /* Not UTF mode */
          {
          for (i = 1; i <= min; i++)
            {
            if (eptr >= md->end_subject)
              {
              SCHECK_PARTIAL();
-             MRRETURN(MATCH_NOMATCH);
+             RRETURN(MATCH_NOMATCH);
              }
-           if (fc == *eptr++) MRRETURN(MATCH_NOMATCH);
+           if (fc == *eptr++) RRETURN(MATCH_NOMATCH);
            }
          }
-Line 3699 
 for (;;)
+Line 3867 
 for (;;)
        if (minimize)
          {
- #ifdef SUPPORT_UTF8
+ #ifdef SUPPORT_UTF
-         /* UTF-8 mode */
+         if (utf)
-         if (utf8)
            {
            register unsigned int d;
            for (fi = min;; fi++)
              {
              RMATCH(eptr, ecode, offset_top, md, eptrb, RM32);
              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
-             if (fi >= max) MRRETURN(MATCH_NOMATCH);
+             if (fi >= max) RRETURN(MATCH_NOMATCH);
              if (eptr >= md->end_subject)
                {
                SCHECK_PARTIAL();
-               MRRETURN(MATCH_NOMATCH);
+               RRETURN(MATCH_NOMATCH);
                }
              GETCHARINC(d, eptr);
-             if (fc == d) MRRETURN(MATCH_NOMATCH);
+             if (fc == d) RRETURN(MATCH_NOMATCH);
              }
            }
          else
  #endif
-         /* Not UTF-8 mode */
+         /* Not UTF mode */
            {
            for (fi = min;; fi++)
              {
              RMATCH(eptr, ecode, offset_top, md, eptrb, RM33);
              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
-             if (fi >= max) MRRETURN(MATCH_NOMATCH);
+             if (fi >= max) RRETURN(MATCH_NOMATCH);
              if (eptr >= md->end_subject)
                {
                SCHECK_PARTIAL();
-               MRRETURN(MATCH_NOMATCH);
+               RRETURN(MATCH_NOMATCH);
                }
-             if (fc == *eptr++) MRRETURN(MATCH_NOMATCH);
+             if (fc == *eptr++) RRETURN(MATCH_NOMATCH);
              }
            }
          /* Control never gets here */
-Line 3744 
 for (;;)
+Line 3911 
 for (;;)
          {
          pp = eptr;
- #ifdef SUPPORT_UTF8
+ #ifdef SUPPORT_UTF
-         /* UTF-8 mode */
+         if (utf)
-         if (utf8)
            {
            register unsigned int d;
            for (i = min; i < max; i++)
-Line 3772 
 for (;;)
+Line 3938 
 for (;;)
            }
          else
  #endif
-         /* Not UTF-8 mode */
+         /* Not UTF mode */
            {
            for (i = min; i < max; i++)
              {
-Line 3793 
 for (;;)
+Line 3959 
 for (;;)
              }
            }
-         MRRETURN(MATCH_NOMATCH);
+         RRETURN(MATCH_NOMATCH);
          }
        }
      /* Control never gets here */
-Line 3805 
 for (;;)
+Line 3971 
 for (;;)
      case OP_TYPEEXACT:
      min = max = GET2(ecode, 1);
      minimize = TRUE;
-     ecode += 3;
+     ecode += 1 + IMM2_SIZE;
      goto REPEATTYPE;
      case OP_TYPEUPTO:
-Line 3813 
 for (;;)
+Line 3979 
 for (;;)
      min = 0;
      max = GET2(ecode, 1);
      minimize = *ecode == OP_TYPEMINUPTO;
-     ecode += 3;
+     ecode += 1 + IMM2_SIZE;
      goto REPEATTYPE;
      case OP_TYPEPOSSTAR:
-Line 3841 
 for (;;)
+Line 4007 
 for (;;)
      possessive = TRUE;
      min = 0;
      max = GET2(ecode, 1);
-     ecode += 3;
+     ecode += 1 + IMM2_SIZE;
      goto REPEATTYPE;
      case OP_TYPESTAR:
-Line 3887 
 for (;;)
+Line 4053 
 for (;;)
          switch(prop_type)
            {
            case PT_ANY:
-           if (prop_fail_result) MRRETURN(MATCH_NOMATCH);
+           if (prop_fail_result) RRETURN(MATCH_NOMATCH);
            for (i = 1; i <= min; i++)
              {
              if (eptr >= md->end_subject)
                {
                SCHECK_PARTIAL();
-               MRRETURN(MATCH_NOMATCH);
+               RRETURN(MATCH_NOMATCH);
                }
              GETCHARINCTEST(c, eptr);
              }
-Line 3906 
 for (;;)
+Line 4072 
 for (;;)
              if (eptr >= md->end_subject)
                {
                SCHECK_PARTIAL();
-               MRRETURN(MATCH_NOMATCH);
+               RRETURN(MATCH_NOMATCH);
                }
              GETCHARINCTEST(c, eptr);
              chartype = UCD_CHARTYPE(c);
              if ((chartype == ucp_Lu ||
                   chartype == ucp_Ll ||
                   chartype == ucp_Lt) == prop_fail_result)
-               MRRETURN(MATCH_NOMATCH);
+               RRETURN(MATCH_NOMATCH);
              }
            break;
-Line 3923 
 for (;;)
+Line 4089 
 for (;;)
              if (eptr >= md->end_subject)
                {
                SCHECK_PARTIAL();
-               MRRETURN(MATCH_NOMATCH);
+               RRETURN(MATCH_NOMATCH);
                }
              GETCHARINCTEST(c, eptr);
              if ((UCD_CATEGORY(c) == prop_value) == prop_fail_result)
-               MRRETURN(MATCH_NOMATCH);
+               RRETURN(MATCH_NOMATCH);
              }
            break;
-Line 3937 
 for (;;)
+Line 4103 
 for (;;)
              if (eptr >= md->end_subject)
                {
                SCHECK_PARTIAL();
-               MRRETURN(MATCH_NOMATCH);
+               RRETURN(MATCH_NOMATCH);
                }
              GETCHARINCTEST(c, eptr);
              if ((UCD_CHARTYPE(c) == prop_value) == prop_fail_result)
-               MRRETURN(MATCH_NOMATCH);
+               RRETURN(MATCH_NOMATCH);
              }
            break;
-Line 3951 
 for (;;)
+Line 4117 
 for (;;)
              if (eptr >= md->end_subject)
                {
                SCHECK_PARTIAL();
-               MRRETURN(MATCH_NOMATCH);
+               RRETURN(MATCH_NOMATCH);
                }
              GETCHARINCTEST(c, eptr);
              if ((UCD_SCRIPT(c) == prop_value) == prop_fail_result)
-               MRRETURN(MATCH_NOMATCH);
+               RRETURN(MATCH_NOMATCH);
              }
            break;
-Line 3966 
 for (;;)
+Line 4132 
 for (;;)
              if (eptr >= md->end_subject)
                {
                SCHECK_PARTIAL();
-               MRRETURN(MATCH_NOMATCH);
+               RRETURN(MATCH_NOMATCH);
                }
              GETCHARINCTEST(c, eptr);
              category = UCD_CATEGORY(c);
              if ((category == ucp_L || category == ucp_N) == prop_fail_result)
-               MRRETURN(MATCH_NOMATCH);
+               RRETURN(MATCH_NOMATCH);
              }
            break;
-Line 3981 
 for (;;)
+Line 4147 
 for (;;)
              if (eptr >= md->end_subject)
                {
                SCHECK_PARTIAL();
-               MRRETURN(MATCH_NOMATCH);
+               RRETURN(MATCH_NOMATCH);
                }
              GETCHARINCTEST(c, eptr);
              if ((UCD_CATEGORY(c) == ucp_Z || c == CHAR_HT || c == CHAR_NL ||
                   c == CHAR_FF || c == CHAR_CR)
                     == prop_fail_result)
-               MRRETURN(MATCH_NOMATCH);
+               RRETURN(MATCH_NOMATCH);
              }
            break;
-Line 3997 
 for (;;)
+Line 4163 
 for (;;)
              if (eptr >= md->end_subject)
                {
                SCHECK_PARTIAL();
-               MRRETURN(MATCH_NOMATCH);
+               RRETURN(MATCH_NOMATCH);
                }
              GETCHARINCTEST(c, eptr);
              if ((UCD_CATEGORY(c) == ucp_Z || c == CHAR_HT || c == CHAR_NL ||
                   c == CHAR_VT || c == CHAR_FF || c == CHAR_CR)
                     == prop_fail_result)
-               MRRETURN(MATCH_NOMATCH);
+               RRETURN(MATCH_NOMATCH);
              }
            break;
-Line 4014 
 for (;;)
+Line 4180 
 for (;;)
              if (eptr >= md->end_subject)
                {
                SCHECK_PARTIAL();
-               MRRETURN(MATCH_NOMATCH);
+               RRETURN(MATCH_NOMATCH);
                }
              GETCHARINCTEST(c, eptr);
              category = UCD_CATEGORY(c);
              if ((category == ucp_L || category == ucp_N || c == CHAR_UNDERSCORE)
                     == prop_fail_result)
-               MRRETURN(MATCH_NOMATCH);
+               RRETURN(MATCH_NOMATCH);
              }
            break;
+           case PT_CLIST:
+           for (i = 1; i <= min; i++)
+             {
+             const pcre_uint32 *cp;
+             if (eptr >= md->end_subject)
+               {
+               SCHECK_PARTIAL();
+               RRETURN(MATCH_NOMATCH);
+               }
+             GETCHARINCTEST(c, eptr);
+             cp = PRIV(ucd_caseless_sets) + UCD_CASESET(c);
+             for (;;)
+               {
+               if (c < *cp)
+                 { if (prop_fail_result) break; else { RRETURN(MATCH_NOMATCH); } }
+               if (c == *cp++)
+                 { if (prop_fail_result) { RRETURN(MATCH_NOMATCH); } else break; }
+               }
+             }
+           break;
            /* This should not occur */
            default:
-Line 4041 
 for (;;)
+Line 4228 
 for (;;)
            if (eptr >= md->end_subject)
              {
              SCHECK_PARTIAL();
-             MRRETURN(MATCH_NOMATCH);
+             RRETURN(MATCH_NOMATCH);
              }
-           GETCHARINCTEST(c, eptr);
+           else
-           if (UCD_CATEGORY(c) == ucp_M) MRRETURN(MATCH_NOMATCH);
-           while (eptr < md->end_subject)
              {
-             int len = 1;
+             int lgb, rgb;
-             if (!utf8) c = *eptr; else { GETCHARLEN(c, eptr, len); }
+             GETCHARINCTEST(c, eptr);
-             if (UCD_CATEGORY(c) != ucp_M) break;
+             lgb = UCD_GRAPHBREAK(c);
-             eptr += len;
+            while (eptr < md->end_subject)
+               {
+               int len = 1;
+               if (!utf) c = *eptr; else { GETCHARLEN(c, eptr, len); }
+               rgb = UCD_GRAPHBREAK(c);
+               if ((PRIV(ucp_gbtable)[lgb] & (1 << rgb)) == 0) break;
+               lgb = rgb;
+               eptr += len;
+               }
              }
+           CHECK_PARTIAL();
            }
          }
-Line 4060 
 for (;;)
+Line 4254 
 for (;;)
  /* Handle all other cases when the coding is UTF-8 */
- #ifdef SUPPORT_UTF8
+ #ifdef SUPPORT_UTF
-       if (utf8) switch(ctype)
+       if (utf) switch(ctype)
          {
          case OP_ANY:
          for (i = 1; i <= min; i++)
-Line 4069 
 for (;;)
+Line 4263 
 for (;;)
            if (eptr >= md->end_subject)
              {
              SCHECK_PARTIAL();
-             MRRETURN(MATCH_NOMATCH);
+             RRETURN(MATCH_NOMATCH);
+             }
+           if (IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH);
+           if (md->partial != 0 &&
+               eptr + 1 >= md->end_subject &&
+               NLBLOCK->nltype == NLTYPE_FIXED &&
+               NLBLOCK->nllen == 2 &&
+               *eptr == NLBLOCK->nl[0])
+             {
+             md->hitend = TRUE;
+             if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
              }
-           if (IS_NEWLINE(eptr)) MRRETURN(MATCH_NOMATCH);
            eptr++;
-           while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;
+           ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);
            }
          break;
-Line 4083 
 for (;;)
+Line 4286 
 for (;;)
            if (eptr >= md->end_subject)
              {
              SCHECK_PARTIAL();
-             MRRETURN(MATCH_NOMATCH);
+             RRETURN(MATCH_NOMATCH);
              }
            eptr++;
-           while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;
+           ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);
            }
          break;
          case OP_ANYBYTE:
-         if (eptr > md->end_subject - min) MRRETURN(MATCH_NOMATCH);
+         if (eptr > md->end_subject - min) RRETURN(MATCH_NOMATCH);
          eptr += min;
          break;
-Line 4101 
 for (;;)
+Line 4304 
 for (;;)
            if (eptr >= md->end_subject)
              {
              SCHECK_PARTIAL();
-             MRRETURN(MATCH_NOMATCH);
+             RRETURN(MATCH_NOMATCH);
              }
            GETCHARINC(c, eptr);
            switch(c)
              {
-             default: MRRETURN(MATCH_NOMATCH);
+             default: RRETURN(MATCH_NOMATCH);
-             case 0x000d:
+             case CHAR_CR:
-             if (eptr < md->end_subject && *eptr == 0x0a) eptr++;
+             if (eptr < md->end_subject && *eptr == CHAR_LF) eptr++;
              break;
-             case 0x000a:
+             case CHAR_LF:
              break;
-             case 0x000b:
+             case CHAR_VT:
-             case 0x000c:
+             case CHAR_FF:
-             case 0x0085:
+             case CHAR_NEL:
+ #ifndef EBCDIC
              case 0x2028:
              case 0x2029:
-             if (md->bsr_anycrlf) MRRETURN(MATCH_NOMATCH);
+ #endif  /* Not EBCDIC */
+             if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);
              break;
              }
            }
-Line 4132 
 for (;;)
+Line 4337 
 for (;;)
            if (eptr >= md->end_subject)
              {
              SCHECK_PARTIAL();
-             MRRETURN(MATCH_NOMATCH);
+             RRETURN(MATCH_NOMATCH);
              }
            GETCHARINC(c, eptr);
            switch(c)
              {
+             HSPACE_CASES: RRETURN(MATCH_NOMATCH);  /* Byte and multibyte cases */
              default: break;
-             case 0x09:      /* HT */
-             case 0x20:      /* SPACE */
-             case 0xa0:      /* NBSP */
-             case 0x1680:    /* OGHAM SPACE MARK */
-             case 0x180e:    /* MONGOLIAN VOWEL SEPARATOR */
-             case 0x2000:    /* EN QUAD */
-             case 0x2001:    /* EM QUAD */
-             case 0x2002:    /* EN SPACE */
-             case 0x2003:    /* EM SPACE */
-             case 0x2004:    /* THREE-PER-EM SPACE */
-             case 0x2005:    /* FOUR-PER-EM SPACE */
-             case 0x2006:    /* SIX-PER-EM SPACE */
-             case 0x2007:    /* FIGURE SPACE */
-             case 0x2008:    /* PUNCTUATION SPACE */
-             case 0x2009:    /* THIN SPACE */
-             case 0x200A:    /* HAIR SPACE */
-             case 0x202f:    /* NARROW NO-BREAK SPACE */
-             case 0x205f:    /* MEDIUM MATHEMATICAL SPACE */
-             case 0x3000:    /* IDEOGRAPHIC SPACE */
-             MRRETURN(MATCH_NOMATCH);
              }
            }
          break;
-Line 4168 
 for (;;)
+Line 4354 
 for (;;)
            if (eptr >= md->end_subject)
              {
              SCHECK_PARTIAL();
-             MRRETURN(MATCH_NOMATCH);
+             RRETURN(MATCH_NOMATCH);
              }
            GETCHARINC(c, eptr);
            switch(c)
              {
-             default: MRRETURN(MATCH_NOMATCH);
+             HSPACE_CASES: break;  /* Byte and multibyte cases */
-             case 0x09:      /* HT */
+             default: RRETURN(MATCH_NOMATCH);
-             case 0x20:      /* SPACE */
-             case 0xa0:      /* NBSP */
-             case 0x1680:    /* OGHAM SPACE MARK */
-             case 0x180e:    /* MONGOLIAN VOWEL SEPARATOR */
-             case 0x2000:    /* EN QUAD */
-             case 0x2001:    /* EM QUAD */
-             case 0x2002:    /* EN SPACE */
-             case 0x2003:    /* EM SPACE */
-             case 0x2004:    /* THREE-PER-EM SPACE */
-             case 0x2005:    /* FOUR-PER-EM SPACE */
-             case 0x2006:    /* SIX-PER-EM SPACE */
-             case 0x2007:    /* FIGURE SPACE */
-             case 0x2008:    /* PUNCTUATION SPACE */
-             case 0x2009:    /* THIN SPACE */
-             case 0x200A:    /* HAIR SPACE */
-             case 0x202f:    /* NARROW NO-BREAK SPACE */
-             case 0x205f:    /* MEDIUM MATHEMATICAL SPACE */
-             case 0x3000:    /* IDEOGRAPHIC SPACE */
-             break;
              }
            }
          break;
-Line 4204 
 for (;;)
+Line 4371 
 for (;;)
            if (eptr >= md->end_subject)
              {
              SCHECK_PARTIAL();
-             MRRETURN(MATCH_NOMATCH);
+             RRETURN(MATCH_NOMATCH);
              }
            GETCHARINC(c, eptr);
            switch(c)
              {
+             VSPACE_CASES: RRETURN(MATCH_NOMATCH);
              default: break;
-             case 0x0a:      /* LF */
-             case 0x0b:      /* VT */
-             case 0x0c:      /* FF */
-             case 0x0d:      /* CR */
-             case 0x85:      /* NEL */
-             case 0x2028:    /* LINE SEPARATOR */
-             case 0x2029:    /* PARAGRAPH SEPARATOR */
-             MRRETURN(MATCH_NOMATCH);
              }
            }
          break;
-Line 4228 
 for (;;)
+Line 4388 
 for (;;)
            if (eptr >= md->end_subject)
              {
              SCHECK_PARTIAL();
-             MRRETURN(MATCH_NOMATCH);
+             RRETURN(MATCH_NOMATCH);
              }
            GETCHARINC(c, eptr);
            switch(c)
              {
-             default: MRRETURN(MATCH_NOMATCH);
+             VSPACE_CASES: break;
-             case 0x0a:      /* LF */
+             default: RRETURN(MATCH_NOMATCH);
-             case 0x0b:      /* VT */
-             case 0x0c:      /* FF */
-             case 0x0d:      /* CR */
-             case 0x85:      /* NEL */
-             case 0x2028:    /* LINE SEPARATOR */
-             case 0x2029:    /* PARAGRAPH SEPARATOR */
-             break;
              }
            }
          break;
-Line 4252 
 for (;;)
+Line 4405 
 for (;;)
            if (eptr >= md->end_subject)
              {
              SCHECK_PARTIAL();
-             MRRETURN(MATCH_NOMATCH);
+             RRETURN(MATCH_NOMATCH);
              }
            GETCHARINC(c, eptr);
            if (c < 128 && (md->ctypes[c] & ctype_digit) != 0)
-             MRRETURN(MATCH_NOMATCH);
+             RRETURN(MATCH_NOMATCH);
            }
          break;
-Line 4266 
 for (;;)
+Line 4419 
 for (;;)
            if (eptr >= md->end_subject)
              {
              SCHECK_PARTIAL();
-             MRRETURN(MATCH_NOMATCH);
+             RRETURN(MATCH_NOMATCH);
              }
-           if (*eptr >= 128 || (md->ctypes[*eptr++] & ctype_digit) == 0)
+           if (*eptr >= 128 || (md->ctypes[*eptr] & ctype_digit) == 0)
-             MRRETURN(MATCH_NOMATCH);
+             RRETURN(MATCH_NOMATCH);
+           eptr++;
            /* No need to skip more bytes - we know it's a 1-byte character */
            }
          break;
-Line 4280 
 for (;;)
+Line 4434 
 for (;;)
            if (eptr >= md->end_subject)
              {
              SCHECK_PARTIAL();
-             MRRETURN(MATCH_NOMATCH);
+             RRETURN(MATCH_NOMATCH);
              }
            if (*eptr < 128 && (md->ctypes[*eptr] & ctype_space) != 0)
-             MRRETURN(MATCH_NOMATCH);
+             RRETURN(MATCH_NOMATCH);
-           while (++eptr < md->end_subject && (*eptr & 0xc0) == 0x80);
+           eptr++;
+           ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);
            }
          break;
-Line 4294 
 for (;;)
+Line 4449 
 for (;;)
            if (eptr >= md->end_subject)
              {
              SCHECK_PARTIAL();
-             MRRETURN(MATCH_NOMATCH);
+             RRETURN(MATCH_NOMATCH);
              }
-           if (*eptr >= 128 || (md->ctypes[*eptr++] & ctype_space) == 0)
+           if (*eptr >= 128 || (md->ctypes[*eptr] & ctype_space) == 0)
-             MRRETURN(MATCH_NOMATCH);
+             RRETURN(MATCH_NOMATCH);
+           eptr++;
            /* No need to skip more bytes - we know it's a 1-byte character */
            }
          break;
-Line 4308 
 for (;;)
+Line 4464 
 for (;;)
            if (eptr >= md->end_subject)
              {
              SCHECK_PARTIAL();
-             MRRETURN(MATCH_NOMATCH);
+             RRETURN(MATCH_NOMATCH);
              }
            if (*eptr < 128 && (md->ctypes[*eptr] & ctype_word) != 0)
-             MRRETURN(MATCH_NOMATCH);
+             RRETURN(MATCH_NOMATCH);
-           while (++eptr < md->end_subject && (*eptr & 0xc0) == 0x80);
+           eptr++;
+           ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);
            }
          break;
-Line 4322 
 for (;;)
+Line 4479 
 for (;;)
            if (eptr >= md->end_subject)
              {
              SCHECK_PARTIAL();
-             MRRETURN(MATCH_NOMATCH);
+             RRETURN(MATCH_NOMATCH);
              }
-           if (*eptr >= 128 || (md->ctypes[*eptr++] & ctype_word) == 0)
+           if (*eptr >= 128 || (md->ctypes[*eptr] & ctype_word) == 0)
-             MRRETURN(MATCH_NOMATCH);
+             RRETURN(MATCH_NOMATCH);
+           eptr++;
            /* No need to skip more bytes - we know it's a 1-byte character */
            }
          break;
-Line 4335 
 for (;;)
+Line 4493 
 for (;;)
          }  /* End switch(ctype) */
        else
- #endif     /* SUPPORT_UTF8 */
+ #endif     /* SUPPORT_UTF */
        /* Code for the non-UTF-8 case for minimum matching of operators other
        than OP_PROP and OP_NOTPROP. */
-Line 4348 
 for (;;)
+Line 4506 
 for (;;)
            if (eptr >= md->end_subject)
              {
              SCHECK_PARTIAL();
-             MRRETURN(MATCH_NOMATCH);
+             RRETURN(MATCH_NOMATCH);
+             }
+           if (IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH);
+           if (md->partial != 0 &&
+               eptr + 1 >= md->end_subject &&
+               NLBLOCK->nltype == NLTYPE_FIXED &&
+               NLBLOCK->nllen == 2 &&
+               *eptr == NLBLOCK->nl[0])
+             {
+             md->hitend = TRUE;
+             if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL);
              }
-           if (IS_NEWLINE(eptr)) MRRETURN(MATCH_NOMATCH);
            eptr++;
            }
          break;
-Line 4359 
 for (;;)
+Line 4526 
 for (;;)
          if (eptr > md->end_subject - min)
            {
            SCHECK_PARTIAL();
-           MRRETURN(MATCH_NOMATCH);
+           RRETURN(MATCH_NOMATCH);
            }
          eptr += min;
          break;
-Line 4368 
 for (;;)
+Line 4535 
 for (;;)
          if (eptr > md->end_subject - min)
            {
            SCHECK_PARTIAL();
-           MRRETURN(MATCH_NOMATCH);
+           RRETURN(MATCH_NOMATCH);
            }
          eptr += min;
          break;
-Line 4379 
 for (;;)
+Line 4546 
 for (;;)
            if (eptr >= md->end_subject)
              {
              SCHECK_PARTIAL();
-             MRRETURN(MATCH_NOMATCH);
+             RRETURN(MATCH_NOMATCH);
              }
            switch(*eptr++)
              {
-             default: MRRETURN(MATCH_NOMATCH);
+             default: RRETURN(MATCH_NOMATCH);
-             case 0x000d:
+             case CHAR_CR:
-             if (eptr < md->end_subject && *eptr == 0x0a) eptr++;
+             if (eptr < md->end_subject && *eptr == CHAR_LF) eptr++;
              break;
-             case 0x000a:
+             case CHAR_LF:
              break;
-             case 0x000b:
+             case CHAR_VT:
-             case 0x000c:
+             case CHAR_FF:
-             case 0x0085:
+             case CHAR_NEL:
-             if (md->bsr_anycrlf) MRRETURN(MATCH_NOMATCH);
+ #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
+             case 0x2028:
+             case 0x2029:
+ #endif
+             if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);
              break;
              }
            }
-Line 4407 
 for (;;)
+Line 4578 
 for (;;)
            if (eptr >= md->end_subject)
              {
              SCHECK_PARTIAL();
-             MRRETURN(MATCH_NOMATCH);
+             RRETURN(MATCH_NOMATCH);
              }
            switch(*eptr++)
              {
              default: break;
-             case 0x09:      /* HT */
+             HSPACE_BYTE_CASES:
-             case 0x20:      /* SPACE */
+ #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
-             case 0xa0:      /* NBSP */
+             HSPACE_MULTIBYTE_CASES:
-             MRRETURN(MATCH_NOMATCH);
+ #endif
+             RRETURN(MATCH_NOMATCH);
              }
            }
          break;
-Line 4426 
 for (;;)
+Line 4598 
 for (;;)
            if (eptr >= md->end_subject)
              {
              SCHECK_PARTIAL();
-             MRRETURN(MATCH_NOMATCH);
+             RRETURN(MATCH_NOMATCH);
              }
            switch(*eptr++)
              {
-             default: MRRETURN(MATCH_NOMATCH);
+             default: RRETURN(MATCH_NOMATCH);
-             case 0x09:      /* HT */
+             HSPACE_BYTE_CASES:
-             case 0x20:      /* SPACE */
+ #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
-             case 0xa0:      /* NBSP */
+             HSPACE_MULTIBYTE_CASES:
+ #endif
              break;
              }
            }
-Line 4445 
 for (;;)
+Line 4618 
 for (;;)
            if (eptr >= md->end_subject)
              {
              SCHECK_PARTIAL();
-             MRRETURN(MATCH_NOMATCH);
+             RRETURN(MATCH_NOMATCH);
              }
            switch(*eptr++)
              {
+             VSPACE_BYTE_CASES:
+ #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
+             VSPACE_MULTIBYTE_CASES:
+ #endif
+             RRETURN(MATCH_NOMATCH);
              default: break;
-             case 0x0a:      /* LF */
-             case 0x0b:      /* VT */
-             case 0x0c:      /* FF */
-             case 0x0d:      /* CR */
-             case 0x85:      /* NEL */
-             MRRETURN(MATCH_NOMATCH);
              }
            }
          break;
-Line 4466 
 for (;;)
+Line 4638 
 for (;;)
            if (eptr >= md->end_subject)
              {
              SCHECK_PARTIAL();
-             MRRETURN(MATCH_NOMATCH);
+             RRETURN(MATCH_NOMATCH);
              }
            switch(*eptr++)
              {
-             default: MRRETURN(MATCH_NOMATCH);
+             default: RRETURN(MATCH_NOMATCH);
-             case 0x0a:      /* LF */
+             VSPACE_BYTE_CASES:
-             case 0x0b:      /* VT */
+ #if defined COMPILE_PCRE16 || defined COMPILE_PCRE32
-             case 0x0c:      /* FF */
+             VSPACE_MULTIBYTE_CASES:
-             case 0x0d:      /* CR */
+ #endif
-             case 0x85:      /* NEL */
              break;
              }
            }
-Line 4487 
 for (;;)
+Line 4658 
 for (;;)
            if (eptr >= md->end_subject)
              {
              SCHECK_PARTIAL();
-             MRRETURN(MATCH_NOMATCH);
+             RRETURN(MATCH_NOMATCH);
              }
-           if ((md->ctypes[*eptr++] & ctype_digit) != 0) MRRETURN(MATCH_NOMATCH);
+           if (MAX_255(*eptr) && (md->ctypes[*eptr] & ctype_digit) != 0)
+             RRETURN(MATCH_NOMATCH);
+           eptr++;
            }
          break;
-Line 4499 
 for (;;)
+Line 4672 
 for (;;)
            if (eptr >= md->end_subject)
              {
              SCHECK_PARTIAL();
-             MRRETURN(MATCH_NOMATCH);
+             RRETURN(MATCH_NOMATCH);
              }
-           if ((md->ctypes[*eptr++] & ctype_digit) == 0) MRRETURN(MATCH_NOMATCH);
+           if (!MAX_255(*eptr) || (md->ctypes[*eptr] & ctype_digit) == 0)
+             RRETURN(MATCH_NOMATCH);
+           eptr++;
            }
          break;
-Line 4511 
 for (;;)
+Line 4686 
 for (;;)
            if (eptr >= md->end_subject)
              {
              SCHECK_PARTIAL();
-             MRRETURN(MATCH_NOMATCH);
+             RRETURN(MATCH_NOMATCH);
              }
-           if ((md->ctypes[*eptr++] & ctype_space) != 0) MRRETURN(MATCH_NOMATCH);
+           if (MAX_255(*eptr) && (md->ctypes[*eptr] & ctype_space) != 0)
+             RRETURN(MATCH_NOMATCH);
+           eptr++;
            }
          break;
-Line 4523 
 for (;;)
+Line 4700 
 for (;;)
            if (eptr >= md->end_subject)
              {
              SCHECK_PARTIAL();
-             MRRETURN(MATCH_NOMATCH);
+             RRETURN(MATCH_NOMATCH);
              }
-           if ((md->ctypes[*eptr++] & ctype_space) == 0) MRRETURN(MATCH_NOMATCH);
+           if (!MAX_255(*eptr) || (md->ctypes[*eptr] & ctype_space) == 0)
+             RRETURN(MATCH_NOMATCH);
+           eptr++;
            }
          break;
-Line 4535 
 for (;;)
+Line 4714 
 for (;;)
            if (eptr >= md->end_subject)
              {
              SCHECK_PARTIAL();
-             MRRETURN(MATCH_NOMATCH);
+             RRETURN(MATCH_NOMATCH);
              }
-           if ((md->ctypes[*eptr++] & ctype_word) != 0)
+           if (MAX_255(*eptr) && (md->ctypes[*eptr] & ctype_word) != 0)
-             MRRETURN(MATCH_NOMATCH);
+             RRETURN(MATCH_NOMATCH);
+           eptr++;
            }
          break;
-Line 4548 
 for (;;)
+Line 4728 
 for (;;)
            if (eptr >= md->end_subject)
              {
              SCHECK_PARTIAL();
-             MRRETURN(MATCH_NOMATCH);
+             RRETURN(MATCH_NOMATCH);
              }
-           if ((md->ctypes[*eptr++] & ctype_word) == 0)
+           if (!MAX_255(*eptr) || (md->ctypes[*eptr] & ctype_word) == 0)
-             MRRETURN(MATCH_NOMATCH);
+             RRETURN(MATCH_NOMATCH);
+           eptr++;
            }
          break;
-Line 4580 
 for (;;)
+Line 4761 
 for (;;)
              {
              RMATCH(eptr, ecode, offset_top, md, eptrb, RM36);
              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
-             if (fi >= max) MRRETURN(MATCH_NOMATCH);
+             if (fi >= max) RRETURN(MATCH_NOMATCH);
              if (eptr >= md->end_subject)
                {
                SCHECK_PARTIAL();
-               MRRETURN(MATCH_NOMATCH);
+               RRETURN(MATCH_NOMATCH);
                }
              GETCHARINCTEST(c, eptr);
-             if (prop_fail_result) MRRETURN(MATCH_NOMATCH);
+             if (prop_fail_result) RRETURN(MATCH_NOMATCH);
              }
            /* Control never gets here */
-Line 4597 
 for (;;)
+Line 4778 
 for (;;)
              int chartype;
              RMATCH(eptr, ecode, offset_top, md, eptrb, RM37);
              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
-             if (fi >= max) MRRETURN(MATCH_NOMATCH);
+             if (fi >= max) RRETURN(MATCH_NOMATCH);
              if (eptr >= md->end_subject)
                {
                SCHECK_PARTIAL();
-               MRRETURN(MATCH_NOMATCH);
+               RRETURN(MATCH_NOMATCH);
                }
              GETCHARINCTEST(c, eptr);
              chartype = UCD_CHARTYPE(c);
              if ((chartype == ucp_Lu ||
                   chartype == ucp_Ll ||
                   chartype == ucp_Lt) == prop_fail_result)
-               MRRETURN(MATCH_NOMATCH);
+               RRETURN(MATCH_NOMATCH);
              }
            /* Control never gets here */
-Line 4617 
 for (;;)
+Line 4798 
 for (;;)
              {
              RMATCH(eptr, ecode, offset_top, md, eptrb, RM38);
              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
-             if (fi >= max) MRRETURN(MATCH_NOMATCH);
+             if (fi >= max) RRETURN(MATCH_NOMATCH);
              if (eptr >= md->end_subject)
                {
                SCHECK_PARTIAL();
-               MRRETURN(MATCH_NOMATCH);
+               RRETURN(MATCH_NOMATCH);
                }
              GETCHARINCTEST(c, eptr);
              if ((UCD_CATEGORY(c) == prop_value) == prop_fail_result)
-               MRRETURN(MATCH_NOMATCH);
+               RRETURN(MATCH_NOMATCH);
              }
            /* Control never gets here */
-Line 4634 
 for (;;)
+Line 4815 
 for (;;)
              {
              RMATCH(eptr, ecode, offset_top, md, eptrb, RM39);
              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
-             if (fi >= max) MRRETURN(MATCH_NOMATCH);
+             if (fi >= max) RRETURN(MATCH_NOMATCH);
              if (eptr >= md->end_subject)
                {
                SCHECK_PARTIAL();
-               MRRETURN(MATCH_NOMATCH);
+               RRETURN(MATCH_NOMATCH);
                }
              GETCHARINCTEST(c, eptr);
              if ((UCD_CHARTYPE(c) == prop_value) == prop_fail_result)
-               MRRETURN(MATCH_NOMATCH);
+               RRETURN(MATCH_NOMATCH);
              }
            /* Control never gets here */
-Line 4651 
 for (;;)
+Line 4832 
 for (;;)
              {
              RMATCH(eptr, ecode, offset_top, md, eptrb, RM40);
              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
-             if (fi >= max) MRRETURN(MATCH_NOMATCH);
+             if (fi >= max) RRETURN(MATCH_NOMATCH);
              if (eptr >= md->end_subject)
                {
                SCHECK_PARTIAL();
-               MRRETURN(MATCH_NOMATCH);
+               RRETURN(MATCH_NOMATCH);
                }
              GETCHARINCTEST(c, eptr);
              if ((UCD_SCRIPT(c) == prop_value) == prop_fail_result)
-               MRRETURN(MATCH_NOMATCH);
+               RRETURN(MATCH_NOMATCH);
              }
            /* Control never gets here */
-Line 4669 
 for (;;)
+Line 4850 
 for (;;)
              int category;
              RMATCH(eptr, ecode, offset_top, md, eptrb, RM59);
              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
-             if (fi >= max) MRRETURN(MATCH_NOMATCH);
+             if (fi >= max) RRETURN(MATCH_NOMATCH);
              if (eptr >= md->end_subject)
                {
                SCHECK_PARTIAL();
-               MRRETURN(MATCH_NOMATCH);
+               RRETURN(MATCH_NOMATCH);
                }
              GETCHARINCTEST(c, eptr);
              category = UCD_CATEGORY(c);
              if ((category == ucp_L || category == ucp_N) == prop_fail_result)
-               MRRETURN(MATCH_NOMATCH);
+               RRETURN(MATCH_NOMATCH);
              }
            /* Control never gets here */
-Line 4687 
 for (;;)
+Line 4868 
 for (;;)
              {
              RMATCH(eptr, ecode, offset_top, md, eptrb, RM60);
              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
-             if (fi >= max) MRRETURN(MATCH_NOMATCH);
+             if (fi >= max) RRETURN(MATCH_NOMATCH);
              if (eptr >= md->end_subject)
                {
                SCHECK_PARTIAL();
-               MRRETURN(MATCH_NOMATCH);
+               RRETURN(MATCH_NOMATCH);
                }
              GETCHARINCTEST(c, eptr);
              if ((UCD_CATEGORY(c) == ucp_Z || c == CHAR_HT || c == CHAR_NL ||
                   c == CHAR_FF || c == CHAR_CR)
                     == prop_fail_result)
-               MRRETURN(MATCH_NOMATCH);
+               RRETURN(MATCH_NOMATCH);
              }
            /* Control never gets here */
-Line 4706 
 for (;;)
+Line 4887 
 for (;;)
              {
              RMATCH(eptr, ecode, offset_top, md, eptrb, RM61);
              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
-             if (fi >= max) MRRETURN(MATCH_NOMATCH);
+             if (fi >= max) RRETURN(MATCH_NOMATCH);
              if (eptr >= md->end_subject)
                {
                SCHECK_PARTIAL();
-               MRRETURN(MATCH_NOMATCH);
+               RRETURN(MATCH_NOMATCH);
                }
              GETCHARINCTEST(c, eptr);
              if ((UCD_CATEGORY(c) == ucp_Z || c == CHAR_HT || c == CHAR_NL ||
                   c == CHAR_VT || c == CHAR_FF || c == CHAR_CR)
                     == prop_fail_result)
-               MRRETURN(MATCH_NOMATCH);
+               RRETURN(MATCH_NOMATCH);
              }
            /* Control never gets here */
-Line 4726 
 for (;;)
+Line 4907 
 for (;;)
              int category;
              RMATCH(eptr, ecode, offset_top, md, eptrb, RM62);
              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
-             if (fi >= max) MRRETURN(MATCH_NOMATCH);
+             if (fi >= max) RRETURN(MATCH_NOMATCH);
              if (eptr >= md->end_subject)
                {
                SCHECK_PARTIAL();
-               MRRETURN(MATCH_NOMATCH);
+               RRETURN(MATCH_NOMATCH);
                }
              GETCHARINCTEST(c, eptr);
              category = UCD_CATEGORY(c);
-Line 4738 
 for (;;)
+Line 4919 
 for (;;)
                   category == ucp_N ||
                   c == CHAR_UNDERSCORE)
                     == prop_fail_result)
-               MRRETURN(MATCH_NOMATCH);
+               RRETURN(MATCH_NOMATCH);
              }
            /* Control never gets here */
-           /* This should never occur */
+           case PT_CLIST:
+           for (fi = min;; fi++)
+             {
+             const pcre_uint32 *cp;
+             RMATCH(eptr, ecode, offset_top, md, eptrb, RM62);
+             if (rrc != MATCH_NOMATCH) RRETURN(rrc);
+             if (fi >= max) RRETURN(MATCH_NOMATCH);
+             if (eptr >= md->end_subject)
+               {
+               SCHECK_PARTIAL();
+               RRETURN(MATCH_NOMATCH);
+               }
+             GETCHARINCTEST(c, eptr);
+             cp = PRIV(ucd_caseless_sets) + UCD_CASESET(c);
+             for (;;)
+               {
+               if (c < *cp)
+                 { if (prop_fail_result) break; else { RRETURN(MATCH_NOMATCH); } }
+               if (c == *cp++)