pcre2/src/pcre2_match.c

/*************************************************
*      Perl-Compatible Regular Expressions       *
*************************************************/

/* PCRE is a library of functions to support regular expressions whose syntax
and semantics are as close as possible to those of the Perl 5 language.

                       Written by Philip Hazel
     Original API code Copyright (c) 1997-2012 University of Cambridge
         New API code Copyright (c) 2016 University of Cambridge

-----------------------------------------------------------------------------
Redistribution and use in source and binary forms, with or without
modification, are permitted provided that the following conditions are met:

    * Redistributions of source code must retain the above copyright notice,
      this list of conditions and the following disclaimer.

    * Redistributions in binary form must reproduce the above copyright
      notice, this list of conditions and the following disclaimer in the
      documentation and/or other materials provided with the distribution.

    * Neither the name of the University of Cambridge nor the names of its
      contributors may be used to endorse or promote products derived from
      this software without specific prior written permission.

THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS"
AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE
LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
POSSIBILITY OF SUCH DAMAGE.
-----------------------------------------------------------------------------
*/


#ifdef HAVE_CONFIG_H
#include "config.h"
#endif

#define NLBLOCK mb             /* Block containing newline information */
#define PSSTART start_subject  /* Field containing processed string start */
#define PSEND   end_subject    /* Field containing processed string end */

#include "pcre2_internal.h"

/* Masks for identifying the public options that are permitted at match time.
*/

#define PUBLIC_MATCH_OPTIONS \
  (PCRE2_ANCHORED|PCRE2_NOTBOL|PCRE2_NOTEOL|PCRE2_NOTEMPTY| \
   PCRE2_NOTEMPTY_ATSTART|PCRE2_NO_UTF_CHECK|PCRE2_PARTIAL_HARD| \
   PCRE2_PARTIAL_SOFT|PCRE2_NO_JIT)

#define PUBLIC_JIT_MATCH_OPTIONS \
   (PCRE2_NO_UTF_CHECK|PCRE2_NOTBOL|PCRE2_NOTEOL|PCRE2_NOTEMPTY|\
    PCRE2_NOTEMPTY_ATSTART|PCRE2_PARTIAL_SOFT|PCRE2_PARTIAL_HARD)

/* The mb->capture_last field uses the lower 16 bits for the last captured
substring (which can never be greater than 65535) and a bit in the top half
to mean "capture vector overflowed". This odd way of doing things was
implemented when it was realized that preserving and restoring the overflow bit
whenever the last capture number was saved/restored made for a neater
interface, and doing it this way saved on (a) another variable, which would
have increased the stack frame size (a big NO-NO in PCRE) and (b) another
separate set of save/restore instructions. The following defines are used in
implementing this. */

#define CAPLMASK    0x0000ffff    /* The bits used for last_capture */
#define OVFLMASK    0xffff0000    /* The bits used for the overflow flag */
#define OVFLBIT     0x00010000    /* The bit that is set for overflow */

/* Bits for setting in mb->match_function_type to indicate two special types
of call to match(). We do it this way to save on using another stack variable,
as stack usage is to be discouraged. */

#define MATCH_CONDASSERT     1  /* Called to check a condition assertion */
#define MATCH_CBEGROUP       2  /* Could-be-empty unlimited repeat group */

/* Non-error returns from the match() function. Error returns are externally
defined PCRE2_ERROR_xxx codes, which are all negative. */

#define MATCH_MATCH        1
#define MATCH_NOMATCH      0

/* Special internal returns from the match() function. Make them sufficiently
negative to avoid the external error codes. */

#define MATCH_ACCEPT       (-999)
#define MATCH_KETRPOS      (-998)
#define MATCH_ONCE         (-997)
/* The next 5 must be kept together and in sequence so that a test that checks
for any one of them can use a range. */
#define MATCH_COMMIT       (-996)
#define MATCH_PRUNE        (-995)
#define MATCH_SKIP         (-994)
#define MATCH_SKIP_ARG     (-993)
#define MATCH_THEN         (-992)
#define MATCH_BACKTRACK_MAX MATCH_THEN
#define MATCH_BACKTRACK_MIN MATCH_COMMIT

/* Min and max values for the common repeats; for the maxima, 0 => infinity */

static const char rep_min[] = { 0, 0, 1, 1, 0, 0, 0, 0, 0, 1, 0, };
static const char rep_max[] = { 0, 0, 0, 0, 1, 1, 0, 0, 0, 0, 1, };

/* Maximum number of ovector elements that can be saved on the system stack
when processing OP_RECURSE in non-HEAP_MATCH_RECURSE mode. If the ovector is
bigger, malloc() is used. This value should be a multiple of 3, because the
ovector length is always a multiple of 3. */

#define OP_RECURSE_STACK_SAVE_MAX 45



/*************************************************
*          Match a back-reference                *
*************************************************/

/* This function is called only when it is known that the offset lies within
the offsets that have so far been used in the match. Note that in caseless
UTF-8 mode, the number of subject bytes matched may be different to the number
of reference bytes. (In theory this could also happen in UTF-16 mode, but it
seems unlikely.)

Arguments:
  offset      index into the offset vector
  offset_top  top of the used offset vector
  eptr        pointer into the subject
  mb          points to match block
  caseless    TRUE if caseless
  lengthptr   pointer for returning the length matched

Returns:      = 0 sucessful match; number of code units matched is set
              < 0 no match
              > 0 partial match
*/

static int
match_ref(PCRE2_SIZE offset, PCRE2_SIZE offset_top, PCRE2_SPTR eptr,
  match_block *mb, BOOL caseless, PCRE2_SIZE *lengthptr)
{
#if defined SUPPORT_UNICODE
BOOL utf = (mb->poptions & PCRE2_UTF) != 0;
#endif

PCRE2_SPTR p;
PCRE2_SIZE length;
PCRE2_SPTR eptr_start = eptr;

/* Deal with an unset group. The default is no match, but there is an option to
match an empty string. */

if (offset >= offset_top || mb->ovector[offset] == PCRE2_UNSET)
  {
  if ((mb->poptions & PCRE2_MATCH_UNSET_BACKREF) != 0)
    {
    *lengthptr = 0;
    return 0;      /* Match */
    }
  else return -1;  /* No match */
  }

/* Separate the caseless and UTF cases for speed. */

p = mb->start_subject + mb->ovector[offset];
length = mb->ovector[offset+1] - mb->ovector[offset];

if (caseless)
  {
#if defined SUPPORT_UNICODE
  if (utf)
    {
    /* Match characters up to the end of the reference. NOTE: the number of
    code units matched may differ, because in UTF-8 there are some characters
    whose upper and lower case versions code have different numbers of bytes.
    For example, U+023A (2 bytes in UTF-8) is the upper case version of U+2C65
    (3 bytes in UTF-8); a sequence of 3 of the former uses 6 bytes, as does a
    sequence of two of the latter. It is important, therefore, to check the
    length along the reference, not along the subject (earlier code did this
    wrong). */

    PCRE2_SPTR endptr = p + length;
    while (p < endptr)
      {
      uint32_t c, d;
      const ucd_record *ur;
      if (eptr >= mb->end_subject) return 1;   /* Partial match */
      GETCHARINC(c, eptr);
      GETCHARINC(d, p);
      ur = GET_UCD(d);
      if (c != d && c != (uint32_t)((int)d + ur->other_case))
        {
        const uint32_t *pp = PRIV(ucd_caseless_sets) + ur->caseset;
        for (;;)
          {
          if (c < *pp) return -1;  /* No match */
          if (c == *pp++) break;
          }
        }
      }
    }
  else
#endif

    /* Not in UTF mode */

    {
    for (; length > 0; length--)
      {
      uint32_t cc, cp;
      if (eptr >= mb->end_subject) return 1;   /* Partial match */
      cc = UCHAR21TEST(eptr);
      cp = UCHAR21TEST(p);
      if (TABLE_GET(cp, mb->lcc, cp) != TABLE_GET(cc, mb->lcc, cc))
        return -1;  /* No match */
      p++;
      eptr++;
      }
    }
  }

/* In the caseful case, we can just compare the code units, whether or not we
are in UTF mode. */

else
  {
  for (; length > 0; length--)
    {
    if (eptr >= mb->end_subject) return 1;   /* Partial match */
    if (UCHAR21INCTEST(p) != UCHAR21INCTEST(eptr)) return -1;  /*No match */
    }
  }

*lengthptr = eptr - eptr_start;
return 0;  /* Match */
}



/***************************************************************************
****************************************************************************
                   RECURSION IN THE match() FUNCTION

The match() function is highly recursive, though not every recursive call
increases the recursion depth. Nevertheless, some regular expressions can cause
it to recurse to a great depth. I was writing for Unix, so I just let it call
itself recursively. This uses the stack for saving everything that has to be
saved for a recursive call. On Unix, the stack can be large, and this works
fine.

It turns out that on some non-Unix-like systems there are problems with
programs that use a lot of stack. (This despite the fact that every last chip
has oodles of memory these days, and techniques for extending the stack have
been known for decades.) So....

There is a fudge, triggered by defining HEAP_MATCH_RECURSE, which avoids
recursive calls by keeping local variables that need to be preserved in blocks
of memory on the heap instead instead of on the stack. Macros are used to
achieve this so that the actual code doesn't look very different to what it
always used to.

The original heap-recursive code used longjmp(). However, it seems that this
can be very slow on some operating systems. Following a suggestion from Stan
Switzer, the use of longjmp() has been abolished, at the cost of having to
provide a unique number for each call to RMATCH. There is no way of generating
a sequence of numbers at compile time in C. I have given them names, to make
them stand out more clearly.

Crude tests on x86 Linux show a small speedup of around 5-8%. However, on
FreeBSD, avoiding longjmp() more than halves the time taken to run the standard
tests. Furthermore, not using longjmp() means that local dynamic variables
don't have indeterminate values; this has meant that the frame size can be
reduced because the result can be "passed back" by straight setting of the
variable instead of being passed in the frame.
****************************************************************************
***************************************************************************/

/* Numbers for RMATCH calls. When this list is changed, the code at HEAP_RETURN
below must be updated in sync.  */

enum { RM1=1, RM2,  RM3,  RM4,  RM5,  RM6,  RM7,  RM8,  RM9,  RM10,
       RM11,  RM12, RM13, RM14, RM15, RM16, RM17, RM18, RM19, RM20,
       RM21,  RM22, RM23, RM24, RM25, RM26, RM27, RM28, RM29, RM30,
       RM31,  RM32, RM33, RM34, RM35, RM36, RM37, RM38, RM39, RM40,
       RM41,  RM42, RM43, RM44, RM45, RM46, RM47, RM48, RM49, RM50,
       RM51,  RM52, RM53, RM54, RM55, RM56, RM57, RM58, RM59, RM60,
       RM61,  RM62, RM63, RM64, RM65, RM66, RM67, RM68 };

/* These versions of the macros use the stack, as normal. Note that the "rw"
argument of RMATCH isn't actually used in this definition. */

#ifndef HEAP_MATCH_RECURSE
#define RMATCH(ra,rb,rc,rd,re,rw) \
  rrc = match(ra,rb,mstart,rc,rd,re,rdepth+1)
#define RRETURN(ra) return ra
#else

/* These versions of the macros manage a private stack on the heap. Note that
the "rd" argument of RMATCH isn't actually used in this definition. It's the mb
argument of match(), which never changes. */

#define RMATCH(ra,rb,rc,rd,re,rw)\
  {\
  heapframe *newframe = frame->Xnextframe;\
  if (newframe == NULL)\
    {\
    newframe = (heapframe *)(mb->stack_memctl.malloc)\
      (sizeof(heapframe), mb->stack_memctl.memory_data);\
    if (newframe == NULL) RRETURN(PCRE2_ERROR_NOMEMORY);\
    newframe->Xnextframe = NULL;\
    frame->Xnextframe = newframe;\
    }\
  frame->Xwhere = rw;\
  newframe->Xeptr = ra;\
  newframe->Xecode = rb;\
  newframe->Xmstart = mstart;\
  newframe->Xoffset_top = rc;\
  newframe->Xeptrb = re;\
  newframe->Xrdepth = frame->Xrdepth + 1;\
  newframe->Xprevframe = frame;\
  frame = newframe;\
  goto HEAP_RECURSE;\
  L_##rw:;\
  }

#define RRETURN(ra)\
  {\
  heapframe *oldframe = frame;\
  frame = oldframe->Xprevframe;\
  if (frame != NULL)\
    {\
    rrc = ra;\
    goto HEAP_RETURN;\
    }\
  return ra;\
  }


/* Structure for remembering the local variables in a private frame. Arrange it
so as to minimize the number of holes. */

typedef struct heapframe {
  struct heapframe *Xprevframe;
  struct heapframe *Xnextframe;

#ifdef SUPPORT_UNICODE
  PCRE2_SPTR Xcharptr;
#endif
  PCRE2_SPTR Xeptr;
  PCRE2_SPTR Xecode;
  PCRE2_SPTR Xmstart;
  PCRE2_SPTR Xcallpat;
  PCRE2_SPTR Xdata;
  PCRE2_SPTR Xnext_ecode;
  PCRE2_SPTR Xpp;
  PCRE2_SPTR Xprev;
  PCRE2_SPTR Xsaved_eptr;

  eptrblock *Xeptrb;

  PCRE2_SIZE Xlength;
  PCRE2_SIZE Xoffset;
  PCRE2_SIZE Xoffset_top;
  PCRE2_SIZE Xsave_offset1, Xsave_offset2, Xsave_offset3;

  uint32_t Xfc;
  uint32_t Xnumber;
  uint32_t Xrdepth;
  uint32_t Xop;
  uint32_t Xsave_capture_last;

#ifdef SUPPORT_UNICODE
  uint32_t Xprop_value;
  int Xprop_type;
  int Xprop_fail_result;
  int Xoclength;
#endif

  int Xcodelink;
  int Xctype;
  int Xfi;
  int Xmax;
  int Xmin;
  int Xwhere;    /* Where to jump back to */

  BOOL Xcondition;
  BOOL Xcur_is_word;
  BOOL Xprev_is_word;

  eptrblock Xnewptrb;
  recursion_info Xnew_recursive;

#ifdef SUPPORT_UNICODE
  PCRE2_UCHAR Xocchars[6];
#endif
} heapframe;

#endif


/***************************************************************************
***************************************************************************/


/* When HEAP_MATCH_RECURSE is not defined, the match() function implements
backtrack points by calling itself recursively in all but one case. The one
special case is when processing OP_RECURSE, which specifies recursion in the
pattern. The entire ovector must be saved and restored while processing
OP_RECURSE. If the ovector is small enough, instead of calling match()
directly, op_recurse_ovecsave() is called. This function uses the system stack
to save the ovector while calling match() to process the pattern recursion. */

#ifndef HEAP_MATCH_RECURSE

/* We need a prototype for match() because it is mutually recursive with
op_recurse_ovecsave(). */

static int
match(PCRE2_SPTR eptr, PCRE2_SPTR ecode, PCRE2_SPTR mstart,
  PCRE2_SIZE offset_top, match_block *mb, eptrblock *eptrb, uint32_t rdepth);


/*************************************************
*      Process OP_RECURSE, stacking ovector      *
*************************************************/

/* When this function is called, mb->recursive has already been updated to
point to a new recursion data block, and all its fields other than ovec_save
have been set.

This function exists so that the local vector variable ovecsave is no longer
defined in the match() function, as it was in PCRE1. It is used only when there
is recursion in the pattern, so it wastes a lot of stack to have it defined for
every call of match(). We now use this function as an indirect way of calling
match() only in the case when ovecsave is needed. (David Wheeler used to say
"All problems in computer science can be solved by another level of
indirection.")

HOWEVER: when this file is compiled by gcc in an optimizing mode, because this
function is called only once, and only from within match(), gcc will "inline"
it - that is, move it inside match() - and this completely negates its reason
for existence. Therefore, we mark it as non-inline when gcc is in use.

Arguments:
  eptr        pointer to current character in subject
  callpat     the recursion point in the pattern
  mstart      pointer to the current match start position (can be modified
                by encountering \K)
  offset_top  current top pointer (highest ovector offset used + 1)
  mb          pointer to "static" info block for the match
  eptrb       pointer to chain of blocks containing eptr at start of
                brackets - for testing for empty matches
  rdepth      the recursion depth

Returns:      a match() return code
*/

static int
#if defined(__GNUC__) && !defined(__INTEL_COMPILER)
__attribute__ ((noinline))
#endif
op_recurse_ovecsave(PCRE2_SPTR eptr, PCRE2_SPTR callpat,
  PCRE2_SPTR mstart, PCRE2_SIZE offset_top, match_block *mb, eptrblock *eptrb,
  uint32_t rdepth)
{
int rrc;
BOOL cbegroup = *callpat >= OP_SBRA;
recursion_info *new_recursive = mb->recursive;
PCRE2_SIZE ovecsave[OP_RECURSE_STACK_SAVE_MAX];

/* Save the ovector */

new_recursive->ovec_save = ovecsave;
memcpy(ovecsave, mb->ovector, mb->offset_end * sizeof(PCRE2_SIZE));

/* Do the recursion. After processing each alternative, restore the ovector
data and the last captured value. */

do
  {
  if (cbegroup) mb->match_function_type |= MATCH_CBEGROUP;
  rrc = match(eptr, callpat + PRIV(OP_lengths)[*callpat], mstart, offset_top,
    mb, eptrb, rdepth + 1);
  memcpy(mb->ovector, new_recursive->ovec_save,
      mb->offset_end * sizeof(PCRE2_SIZE));
  mb->capture_last = new_recursive->saved_capture_last;
  if (rrc == MATCH_MATCH || rrc == MATCH_ACCEPT) return rrc;

  /* PCRE does not allow THEN, SKIP, PRUNE or COMMIT to escape beyond a
  recursion; they cause a NOMATCH for the entire recursion. These codes
  are defined in a range that can be tested for. */

  if (rrc >= MATCH_BACKTRACK_MIN && rrc <= MATCH_BACKTRACK_MAX)
    return MATCH_NOMATCH;

  /* Any return code other than NOMATCH is an error. Otherwise, advance to the
  next alternative or to the end of the recursing subpattern. If there were
  nested recursions, mb->recursive might be changed, so reset it before
  looping. */

  if (rrc != MATCH_NOMATCH) return rrc;
  mb->recursive = new_recursive;
  callpat += GET(callpat, 1);
  }
while (*callpat == OP_ALT);  /* Loop for the alternatives */

/* None of the alternatives matched. */

return MATCH_NOMATCH;
}
#endif  /* HEAP_MATCH_RECURSE */



/*************************************************
*         Match from current position            *
*************************************************/

/* This function is called recursively in many circumstances. Whenever it
returns a negative (error) response, the outer incarnation must also return the
same response. */

/* These macros pack up tests that are used for partial matching, and which
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 earliest inspected character (i.e.
something has been matched, even if not part of the actual matched string). For
hard partial matching, we then return immediately. The second one is used when
we already know we are past the end of the subject. */

#define CHECK_PARTIAL()\
  if (mb->partial != 0 && eptr >= mb->end_subject && \
      eptr > mb->start_used_ptr) \
    { \
    mb->hitend = TRUE; \
    if (mb->partial > 1) RRETURN(PCRE2_ERROR_PARTIAL); \
    }

#define SCHECK_PARTIAL()\
  if (mb->partial != 0 && eptr > mb->start_used_ptr) \
    { \
    mb->hitend = TRUE; \
    if (mb->partial > 1) RRETURN(PCRE2_ERROR_PARTIAL); \
    }


/* Performance note: It might be tempting to extract commonly used fields from
the mb 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.

Arguments:
   eptr        pointer to current character in subject
   ecode       pointer to current position in compiled code
   mstart      pointer to the current match start position (can be modified
                 by encountering \K)
   offset_top  current top pointer (highest ovector offset used + 1)
   mb          pointer to "static" info block for the match
   eptrb       pointer to chain of blocks containing eptr at start of
                 brackets - for testing for empty matches
   rdepth      the recursion depth

Returns:       MATCH_MATCH if matched            )  these values are >= 0
               MATCH_NOMATCH if failed to match  )
               a negative MATCH_xxx value for PRUNE, SKIP, etc
               a negative PCRE2_ERROR_xxx value if aborted by an error condition
                 (e.g. stopped by repeated call or recursion limit)
*/

static int
match(PCRE2_SPTR eptr, PCRE2_SPTR ecode, PCRE2_SPTR mstart,
  PCRE2_SIZE offset_top, match_block *mb, eptrblock *eptrb, uint32_t rdepth)
{
/* These variables do not need to be preserved over recursion in this function,
so they can be ordinary variables in all cases. Mark some of them with
"register" because they are used a lot in loops. */

int  rrc;         /* Returns from recursive calls */
int  i;           /* Used for loops not involving calls to RMATCH() */
uint32_t c;       /* Character values not kept over RMATCH() calls */
BOOL utf;         /* Local copy of UTF flag for speed */

BOOL minimize, possessive; /* Quantifier options */
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". We set up the top-level
frame on the stack here; subsequent instantiations are obtained from the heap
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 HEAP_MATCH_RECURSE
heapframe *frame = (heapframe *)mb->match_frames_base;

/* Copy in the original argument variables */

frame->Xeptr = eptr;
frame->Xecode = ecode;
frame->Xmstart = mstart;
frame->Xoffset_top = offset_top;
frame->Xeptrb = eptrb;
frame->Xrdepth = rdepth;

/* This is where control jumps back to to effect "recursion" */

HEAP_RECURSE:

/* Macros make the argument variables come from the current frame */

#define eptr               frame->Xeptr
#define ecode              frame->Xecode
#define mstart             frame->Xmstart
#define offset_top         frame->Xoffset_top
#define eptrb              frame->Xeptrb
#define rdepth             frame->Xrdepth

/* Ditto for the local variables */

#ifdef SUPPORT_UNICODE
#define charptr            frame->Xcharptr
#define prop_value         frame->Xprop_value
#define prop_type          frame->Xprop_type
#define prop_fail_result   frame->Xprop_fail_result
#define oclength           frame->Xoclength
#define occhars            frame->Xocchars
#endif


#define callpat            frame->Xcallpat
#define codelink           frame->Xcodelink
#define data               frame->Xdata
#define next_ecode         frame->Xnext_ecode
#define pp                 frame->Xpp
#define prev               frame->Xprev
#define saved_eptr         frame->Xsaved_eptr

#define new_recursive      frame->Xnew_recursive

#define ctype              frame->Xctype
#define fc                 frame->Xfc
#define fi                 frame->Xfi
#define length             frame->Xlength
#define max                frame->Xmax
#define min                frame->Xmin
#define number             frame->Xnumber
#define offset             frame->Xoffset
#define op                 frame->Xop
#define save_capture_last  frame->Xsave_capture_last
#define save_offset1       frame->Xsave_offset1
#define save_offset2       frame->Xsave_offset2
#define save_offset3       frame->Xsave_offset3

#define condition          frame->Xcondition
#define cur_is_word        frame->Xcur_is_word
#define prev_is_word       frame->Xprev_is_word

#define newptrb            frame->Xnewptrb

/* When normal stack-based recursion is being used for match(), local variables
are allocated on the stack and get preserved during recursion in the usual way.
In this environment, fi and i, and fc and c, can be the same variables. */

#else         /* HEAP_MATCH_RECURSE not defined */
#define fi i
#define fc c

/* Many of the following variables are used only in small blocks of the code.
My normal style of coding would have declared them within each of those blocks.
However, in order to accommodate the version of this code that uses an external
"stack" implemented on the heap, it is easier to declare them all here, so the
declarations can be cut out in a block. The only declarations within blocks
below are for variables that do not have to be preserved over a recursive call
to RMATCH(). */

#ifdef SUPPORT_UNICODE
PCRE2_SPTR charptr;
#endif
PCRE2_SPTR callpat;
PCRE2_SPTR data;
PCRE2_SPTR next_ecode;
PCRE2_SPTR pp;
PCRE2_SPTR prev;
PCRE2_SPTR saved_eptr;

PCRE2_SIZE length;
PCRE2_SIZE offset;
PCRE2_SIZE save_offset1, save_offset2, save_offset3;

uint32_t number;
uint32_t op;
uint32_t save_capture_last;

#ifdef SUPPORT_UNICODE
uint32_t prop_value;
int prop_type;
int prop_fail_result;
int oclength;
PCRE2_UCHAR occhars[6];
#endif

int codelink;
int ctype;
int max;
int min;

BOOL condition;
BOOL cur_is_word;
BOOL prev_is_word;

eptrblock newptrb;
recursion_info new_recursive;
#endif  /* HEAP_MATCH_RECURSE not defined */

/* To save space on the stack and in the heap frame, I have doubled up on some
of the local variables that are used only in localised parts of the code, but
still need to be preserved over recursive calls of match(). These macros define
the alternative names that are used. */

#define allow_zero      cur_is_word
#define caseless        cur_is_word
#define cbegroup        condition
#define code_offset     codelink
#define condassert      condition
#define foc             number
#define matched_once    prev_is_word
#define save_mark       data

/* These statements are here to stop the compiler complaining about unitialized
variables. */

#ifdef SUPPORT_UNICODE
prop_value = 0;
prop_fail_result = 0;
#endif


/* This label is used for tail recursion, which is used in a few cases even
when HEAP_MATCH_RECURSE is not defined, in order to reduce the amount of stack
that is used. Thanks to Ian Taylor for noticing this possibility and sending
the original patch. */

TAIL_RECURSE:

/* OK, now we can get on with the real code of the function. Recursive calls
are specified by the macro RMATCH and RRETURN is used to return. When
HEAP_MATCH_RECURSE is *not* defined, these just turn into a recursive call to
match() and a "return", respectively. However, RMATCH isn't like a function
call because it's quite a 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_UNICODE
utf = (mb->poptions & PCRE2_UTF) != 0;
#else
utf = FALSE;
#endif

/* First check that we haven't called match() too many times, or that we
haven't exceeded the recursive call limit. */

if (mb->match_call_count++ >= mb->match_limit) RRETURN(PCRE2_ERROR_MATCHLIMIT);
if (rdepth >= mb->match_limit_recursion) RRETURN(PCRE2_ERROR_RECURSIONLIMIT);

/* At the start of a group with an unlimited repeat that may match an empty
string, the variable mb->match_function_type contains the MATCH_CBEGROUP bit.
It is done this way to save having to use another function argument, which
would take up space on the stack. See also MATCH_CONDASSERT below.

When MATCH_CBEGROUP is set, add the current subject pointer to the chain of
such remembered pointers, to be checked when we hit the closing ket, in order
to break infinite loops that match no characters. When match() is called in
other circumstances, don't add to the chain. The MATCH_CBEGROUP feature must
NOT be used with tail recursion, because the memory block that is used is on
the stack, so a new one may be required for each match(). */

if ((mb->match_function_type & MATCH_CBEGROUP) != 0)
  {
  newptrb.epb_saved_eptr = eptr;
  newptrb.epb_prev = eptrb;
  eptrb = &newptrb;
  mb->match_function_type &= ~MATCH_CBEGROUP;
  }

/* Now, at last, we can start processing the opcodes. */

for (;;)
  {
  minimize = possessive = FALSE;
  op = *ecode;

  switch(op)
    {
    case OP_MARK:
    mb->nomatch_mark = ecode + 2;
    mb->mark = NULL;    /* In case previously set by assertion */
    RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode] + ecode[1], offset_top, mb,
      eptrb, RM55);
    if ((rrc == MATCH_MATCH || rrc == MATCH_ACCEPT) &&
         mb->mark == NULL) mb->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
    argument. It is passed back in mb->start_match_ptr (an overloading of that
    variable). If it does match, we reset that variable to the current subject
    position and return MATCH_SKIP. Otherwise, pass back the return code
    unaltered. */

    else if (rrc == MATCH_SKIP_ARG &&
        PRIV(strcmp)(ecode + 2, mb->start_match_ptr) == 0)
      {
      mb->start_match_ptr = eptr;
      RRETURN(MATCH_SKIP);
      }
    RRETURN(rrc);

    case OP_FAIL:
    RRETURN(MATCH_NOMATCH);

    case OP_COMMIT:
    RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, mb,
      eptrb, RM52);
    if (rrc != MATCH_NOMATCH) RRETURN(rrc);
    RRETURN(MATCH_COMMIT);

    case OP_PRUNE:
    RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, mb,
      eptrb, RM51);
    if (rrc != MATCH_NOMATCH) RRETURN(rrc);
    RRETURN(MATCH_PRUNE);

    case OP_PRUNE_ARG:
    mb->nomatch_mark = ecode + 2;
    mb->mark = NULL;    /* In case previously set by assertion */
    RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode] + ecode[1], offset_top, mb,
      eptrb, RM56);
    if ((rrc == MATCH_MATCH || rrc == MATCH_ACCEPT) &&
         mb->mark == NULL) mb->mark = ecode + 2;
    if (rrc != MATCH_NOMATCH) RRETURN(rrc);
    RRETURN(MATCH_PRUNE);

    case OP_SKIP:
    RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, mb,
      eptrb, RM53);
    if (rrc != MATCH_NOMATCH) RRETURN(rrc);
    mb->start_match_ptr = eptr;   /* Pass back current position */
    RRETURN(MATCH_SKIP);

    /* Note that, for Perl compatibility, SKIP with an argument does NOT set
    nomatch_mark. When a pattern match ends with a SKIP_ARG for which there was
    not a matching mark, we have to re-run the match, ignoring the SKIP_ARG
    that failed and any that precede it (either they also failed, or were not
    triggered). To do this, we maintain a count of executed SKIP_ARGs. If a
    SKIP_ARG gets to top level, the match is re-run with mb->ignore_skip_arg
    set to the count of the one that failed. */

    case OP_SKIP_ARG:
    mb->skip_arg_count++;
    if (mb->skip_arg_count <= mb->ignore_skip_arg)
      {
      ecode += PRIV(OP_lengths)[*ecode] + ecode[1];
      break;
      }
    RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode] + ecode[1], offset_top, mb,
      eptrb, RM57);
    if (rrc != MATCH_NOMATCH) RRETURN(rrc);

    /* Pass back the current skip name by overloading mb->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 causes a rematch
    with mb->ignore_skip_arg set to the value of mb->skip_arg_count. */

    mb->start_match_ptr = ecode + 2;
    RRETURN(MATCH_SKIP_ARG);

    /* For THEN (and THEN_ARG) we pass back the address of the opcode, so that
    the branch in which it occurs can be determined. Overload the start of
    match pointer to do this. */

    case OP_THEN:
    RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, mb,
      eptrb, RM54);
    if (rrc != MATCH_NOMATCH) RRETURN(rrc);
    mb->start_match_ptr = ecode;
    RRETURN(MATCH_THEN);

    case OP_THEN_ARG:
    mb->nomatch_mark = ecode + 2;
    mb->mark = NULL;    /* In case previously set by assertion */
    RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode] + ecode[1], offset_top,
      mb, eptrb, RM58);
    if ((rrc == MATCH_MATCH || rrc == MATCH_ACCEPT) &&
         mb->mark == NULL) mb->mark = ecode + 2;
    if (rrc != MATCH_NOMATCH) RRETURN(rrc);
    mb->start_match_ptr = ecode;
    RRETURN(MATCH_THEN);

    /* Handle an atomic group that does not contain any capturing parentheses.
    This can be handled like an assertion. Prior to 8.13, all atomic groups
    were handled this way. In 8.13, the code was changed as below for ONCE, so
    that backups pass through the group and thereby reset captured values.
    However, this uses a lot more stack, so in 8.20, atomic groups that do not
    contain any captures generate OP_ONCE_NC, which can be handled in the old,
    less stack intensive way.

    Check the alternative branches in turn - the matching won't pass the KET
    for this kind of subpattern. If any one branch matches, we carry on as at
    the end of a normal bracket, leaving the subject pointer, but resetting
    the start-of-match value in case it was changed by \K. */

    case OP_ONCE_NC:
    prev = ecode;
    saved_eptr = eptr;
    save_mark = mb->mark;
    do
      {
      RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, mb, eptrb, RM64);
      if (rrc == MATCH_MATCH)  /* Note: _not_ MATCH_ACCEPT */
        {
        mstart = mb->start_match_ptr;
        break;
        }
      if (rrc == MATCH_THEN)
        {
        next_ecode = ecode + GET(ecode,1);
        if (mb->start_match_ptr < next_ecode &&
            (*ecode == OP_ALT || *next_ecode == OP_ALT))
          rrc = MATCH_NOMATCH;
        }

      if (rrc != MATCH_NOMATCH) RRETURN(rrc);
      ecode += GET(ecode,1);
      mb->mark = save_mark;
      }
    while (*ecode == OP_ALT);

    /* If hit the end of the group (which could be repeated), fail */

    if (*ecode != OP_ONCE_NC && *ecode != OP_ALT) RRETURN(MATCH_NOMATCH);

    /* Continue as from after the group, updating the offsets high water
    mark, since extracts may have been taken. */

    do ecode += GET(ecode, 1); while (*ecode == OP_ALT);

    offset_top = mb->end_offset_top;
    eptr = mb->end_match_ptr;

    /* For a non-repeating ket, just continue at this level. This also
    happens for a repeating ket if no characters were matched in the group.
    This is the forcible breaking of infinite loops as implemented in Perl
    5.005. */

    if (*ecode == OP_KET || eptr == saved_eptr)
      {
      ecode += 1+LINK_SIZE;
      break;
      }

    /* The repeating kets try the rest of the pattern or restart from the
    preceding bracket, in the appropriate order. The second "call" of match()
    uses tail recursion, to avoid using another stack frame. */

    if (*ecode == OP_KETRMIN)
      {
      RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, mb, eptrb, RM65);
      if (rrc != MATCH_NOMATCH) RRETURN(rrc);
      ecode = prev;
      goto TAIL_RECURSE;
      }
    else  /* OP_KETRMAX */
      {
      RMATCH(eptr, prev, offset_top, mb, eptrb, RM66);
      if (rrc != MATCH_NOMATCH) RRETURN(rrc);
      ecode += 1 + LINK_SIZE;
      goto TAIL_RECURSE;
      }
    /* Control never gets here */

    /* Handle a capturing bracket, other than those that are possessive with an
    unlimited repeat. If there is space in the offset vector, save the current
    subject position in the working slot at the top of the vector. We mustn't
    change the current values of the data slot, because they may be set from a
    previous iteration of this group, and be referred to by a reference inside
    the group. A failure to match might occur after the group has succeeded,
    if something later on doesn't match. For this reason, we need to restore
    the working value and also the values of the final offsets, in case they
    were set by a previous iteration of the same bracket.

    If there isn't enough space in the offset vector, treat this as if it were
    a non-capturing bracket. Don't worry about setting the flag for the error
    case here; that is handled in the code for KET. */

    case OP_CBRA:
    case OP_SCBRA:
    number = GET2(ecode, 1+LINK_SIZE);
    offset = number << 1;

    if (offset < mb->offset_max)
      {
      save_offset1 = mb->ovector[offset];
      save_offset2 = mb->ovector[offset+1];
      save_offset3 = mb->ovector[mb->offset_end - number];
      save_capture_last = mb->capture_last;
      save_mark = mb->mark;

      mb->ovector[mb->offset_end - number] = eptr - mb->start_subject;

      for (;;)
        {
        if (op >= OP_SBRA) mb->match_function_type |= MATCH_CBEGROUP;
        RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, mb,
          eptrb, RM1);
        if (rrc == MATCH_ONCE) break;  /* Backing up through an atomic group */

        /* If we backed up to a THEN, check whether it is within the current
        branch by comparing the address of the THEN that is passed back with
        the end of the branch. If it is within the current branch, and the
        branch is one of two or more alternatives (it either starts or ends
        with OP_ALT), we have reached the limit of THEN's action, so convert
        the return code to NOMATCH, which will cause normal backtracking to
        happen from now on. Otherwise, THEN is passed back to an outer
        alternative. This implements Perl's treatment of parenthesized groups,
        where a group not containing | does not affect the current alternative,
        that is, (X) is NOT the same as (X|(*F)). */

        if (rrc == MATCH_THEN)
          {
          next_ecode = ecode + GET(ecode,1);
          if (mb->start_match_ptr < next_ecode &&
              (*ecode == OP_ALT || *next_ecode == OP_ALT))
            rrc = MATCH_NOMATCH;
          }

        /* Anything other than NOMATCH is passed back. */

        if (rrc != MATCH_NOMATCH) RRETURN(rrc);
        mb->capture_last = save_capture_last;
        ecode += GET(ecode, 1);
        mb->mark = save_mark;
        if (*ecode != OP_ALT) break;
        }

      mb->ovector[offset] = save_offset1;
      mb->ovector[offset+1] = save_offset2;
      mb->ovector[mb->offset_end - number] = save_offset3;

      /* At this point, rrc will be one of MATCH_ONCE or MATCH_NOMATCH. */

      RRETURN(rrc);
      }

    /* FALL THROUGH ... Insufficient room for saving captured contents. Treat
    as a non-capturing bracket. */

    /* VVVVVVVVVVVVVVVVVVVVVVVVV */
    /* VVVVVVVVVVVVVVVVVVVVVVVVV */

    /* Non-capturing or atomic group, except for possessive with unlimited
    repeat and ONCE group with no captures. Loop for all the alternatives.

    When we get to the final alternative within the brackets, we used to return
    the result of a recursive call to match() whatever happened so it was
    possible to reduce stack usage by turning this into a tail recursion,
    except in the case of a possibly empty group. However, now that there is
    the possiblity of (*THEN) occurring in the final alternative, this
    optimization is no longer always possible.

    We can optimize if we know there are no (*THEN)s in the pattern; at present
    this is the best that can be done.

    MATCH_ONCE is returned when the end of an atomic group is successfully
    reached, but subsequent matching fails. It passes back up the tree (causing
    captured values to be reset) until the original atomic group level is
    reached. This is tested by comparing mb->once_target with the start of the
    group. At this point, the return is converted into MATCH_NOMATCH so that
    previous backup points can be taken. */

    case OP_ONCE:
    case OP_BRA:
    case OP_SBRA:

    for (;;)
      {
      if (op >= OP_SBRA || op == OP_ONCE)
        mb->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
      above. */

      else if (!mb->hasthen && ecode[GET(ecode, 1)] != OP_ALT)
        {
        ecode += PRIV(OP_lengths)[*ecode];
        goto TAIL_RECURSE;
        }

      /* In all other cases, we have to make another call to match(). */

      save_mark = mb->mark;
      save_capture_last = mb->capture_last;
      RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, mb, eptrb,
        RM2);

      /* See comment in the code for capturing groups above about handling
      THEN. */

      if (rrc == MATCH_THEN)
        {
        next_ecode = ecode + GET(ecode,1);
        if (mb->start_match_ptr < next_ecode &&
            (*ecode == OP_ALT || *next_ecode == OP_ALT))
          rrc = MATCH_NOMATCH;
        }

      if (rrc != MATCH_NOMATCH)
        {
        if (rrc == MATCH_ONCE)
          {
          PCRE2_SPTR scode = ecode;
          if (*scode != OP_ONCE)           /* If not at start, find it */
            {
            while (*scode == OP_ALT) scode += GET(scode, 1);
            scode -= GET(scode, 1);
            }
          if (mb->once_target == scode) rrc = MATCH_NOMATCH;
          }
        RRETURN(rrc);
        }
      ecode += GET(ecode, 1);
      mb->mark = save_mark;
      if (*ecode != OP_ALT) break;
      mb->capture_last = save_capture_last;
      }

    RRETURN(MATCH_NOMATCH);

    /* Handle possessive capturing brackets with an unlimited repeat. We come
    here from BRAZERO with allow_zero set TRUE. The ovector values are
    handled similarly to the normal case above. However, the matching is
    different. The end of these brackets will always be OP_KETRPOS, which
    returns MATCH_KETRPOS without going further in the pattern. By this means
    we can handle the group by iteration rather than recursion, thereby
    reducing the amount of stack needed. If the ovector is too small for
    capturing, treat as non-capturing. */

    case OP_CBRAPOS:
    case OP_SCBRAPOS:
    allow_zero = FALSE;

    POSSESSIVE_CAPTURE:
    number = GET2(ecode, 1+LINK_SIZE);
    offset = number << 1;
    if (offset >= mb->offset_max) goto POSSESSIVE_NON_CAPTURE;

    matched_once = FALSE;
    code_offset = (int)(ecode - mb->start_code);

    save_offset1 = mb->ovector[offset];
    save_offset2 = mb->ovector[offset+1];
    save_offset3 = mb->ovector[mb->offset_end - number];
    save_capture_last = mb->capture_last;

    /* Each time round the loop, save the current subject position for use
    when the group matches. For MATCH_MATCH, the group has matched, so we
    restart it with a new subject starting position, remembering that we had
    at least one match. For MATCH_NOMATCH, carry on with the alternatives, as
    usual. If we haven't matched any alternatives in any iteration, check to
    see if a previous iteration matched. If so, the group has matched;
    continue from afterwards. Otherwise it has failed; restore the previous
    capture values before returning NOMATCH. */

    for (;;)
      {
      mb->ovector[mb->offset_end - number] = eptr - mb->start_subject;
      if (op >= OP_SBRA) mb->match_function_type |= MATCH_CBEGROUP;
      RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, mb,
        eptrb, RM63);
      if (rrc == MATCH_KETRPOS)
        {
        offset_top = mb->end_offset_top;
        ecode = mb->start_code + code_offset;
        save_capture_last = mb->capture_last;
        matched_once = TRUE;
        mstart = mb->start_match_ptr;    /* In case \K changed it */
        if (eptr == mb->end_match_ptr)   /* Matched an empty string */
          {
          do ecode += GET(ecode, 1); while (*ecode == OP_ALT);
          break;
          }
        eptr = mb->end_match_ptr;
        continue;
        }

      /* See comment in the code for capturing groups above about handling
      THEN. */

      if (rrc == MATCH_THEN)
        {
        next_ecode = ecode + GET(ecode,1);
        if (mb->start_match_ptr < next_ecode &&
            (*ecode == OP_ALT || *next_ecode == OP_ALT))
          rrc = MATCH_NOMATCH;
        }

      if (rrc != MATCH_NOMATCH) RRETURN(rrc);
      mb->capture_last = save_capture_last;
      ecode += GET(ecode, 1);
      if (*ecode != OP_ALT) break;
      }

    if (!matched_once)
      {
      mb->ovector[offset] = save_offset1;
      mb->ovector[offset+1] = save_offset2;
      mb->ovector[mb->offset_end - number] = save_offset3;
      }

    if (allow_zero || matched_once)
      {
      ecode += 1 + LINK_SIZE;
      break;
      }
    RRETURN(MATCH_NOMATCH);

    /* Non-capturing possessive bracket with unlimited repeat. We come here
    from BRAZERO with allow_zero = TRUE. The code is similar to the above,
    without the capturing complication. It is written out separately for speed
    and cleanliness. */

    case OP_BRAPOS:
    case OP_SBRAPOS:
    allow_zero = FALSE;

    POSSESSIVE_NON_CAPTURE:
    matched_once = FALSE;
    code_offset = (int)(ecode - mb->start_code);
    save_capture_last = mb->capture_last;

    for (;;)
      {
      if (op >= OP_SBRA) mb->match_function_type |= MATCH_CBEGROUP;
      RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, mb,
        eptrb, RM48);
      if (rrc == MATCH_KETRPOS)
        {
        offset_top = mb->end_offset_top;
        ecode = mb->start_code + code_offset;
        matched_once = TRUE;
        mstart = mb->start_match_ptr;   /* In case \K reset it */
        if (eptr == mb->end_match_ptr)  /* Matched an empty string */
          {
          do ecode += GET(ecode, 1); while (*ecode == OP_ALT);
          break;
          }
        eptr = mb->end_match_ptr;
        continue;
        }

      /* See comment in the code for capturing groups above about handling
      THEN. */

      if (rrc == MATCH_THEN)
        {
        next_ecode = ecode + GET(ecode,1);
        if (mb->start_match_ptr < next_ecode &&
            (*ecode == OP_ALT || *next_ecode == OP_ALT))
          rrc = MATCH_NOMATCH;
        }

      if (rrc != MATCH_NOMATCH) RRETURN(rrc);
      ecode += GET(ecode, 1);
      if (*ecode != OP_ALT) break;
      mb->capture_last = save_capture_last;
      }

    if (matched_once || allow_zero)
      {
      ecode += 1 + LINK_SIZE;
      break;
      }
    RRETURN(MATCH_NOMATCH);

    /* Control never reaches here. */

    /* Conditional group: compilation checked that there are no more than two
    branches. If the condition is false, skipping the first branch takes us
    past the end of the item if there is only one branch, but that's exactly
    what we want. */

    case OP_COND:
    case OP_SCOND:

    /* The variable codelink will be added to ecode when the condition is
    false, to get to the second branch. Setting it to the offset to the ALT
    or KET, then incrementing ecode achieves this effect. We now have ecode
    pointing to the condition or callout. */

    codelink = GET(ecode, 1);   /* Offset to the second branch */
    ecode += 1 + LINK_SIZE;     /* From this opcode */

    /* Because of the way auto-callout works during compile, a callout item is
    inserted between OP_COND and an assertion condition. */

    if (*ecode == OP_CALLOUT || *ecode == OP_CALLOUT_STR)
      {
      unsigned int callout_length = (*ecode == OP_CALLOUT)
          ? PRIV(OP_lengths)[OP_CALLOUT] : GET(ecode, 1 + 2*LINK_SIZE);

      if (mb->callout != NULL)
        {
        pcre2_callout_block cb;
        cb.version          = 1;
        cb.capture_top      = (uint32_t)offset_top/2;
        cb.capture_last     = mb->capture_last & CAPLMASK;
        cb.offset_vector    = mb->ovector;
        cb.mark             = mb->nomatch_mark;
        cb.subject          = mb->start_subject;
        cb.subject_length   = (PCRE2_SIZE)(mb->end_subject - mb->start_subject);
        cb.start_match      = (PCRE2_SIZE)(mstart - mb->start_subject);
        cb.current_position = (PCRE2_SIZE)(eptr - mb->start_subject);
        cb.pattern_position = GET(ecode, 1);
        cb.next_item_length = GET(ecode, 1 + LINK_SIZE);

        if (*ecode == OP_CALLOUT)
          {
          cb.callout_number = ecode[1 + 2*LINK_SIZE];
          cb.callout_string_offset = 0;
          cb.callout_string = NULL;
          cb.callout_string_length = 0;
          }
        else
          {
          cb.callout_number = 0;
          cb.callout_string_offset = GET(ecode, 1 + 3*LINK_SIZE);
          cb.callout_string = ecode + (1 + 4*LINK_SIZE) + 1;
          cb.callout_string_length =
            callout_length - (1 + 4*LINK_SIZE) - 2;
          }

        if ((rrc = mb->callout(&cb, mb->callout_data)) > 0)
          RRETURN(MATCH_NOMATCH);
        if (rrc < 0) RRETURN(rrc);
        }

      /* Advance ecode past the callout, so it now points to the condition. We
      must adjust codelink so that the value of ecode+codelink is unchanged. */

      ecode += callout_length;
      codelink -= callout_length;
      }

    /* Test the various possible conditions */

    condition = FALSE;
    switch(condcode = *ecode)
      {
      case OP_RREF:                  /* Numbered group recursion test */
      if (mb->recursive != NULL)     /* Not recursing => FALSE */
        {
        uint32_t recno = GET2(ecode, 1);   /* Recursion group number*/
        condition = (recno == RREF_ANY || recno == mb->recursive->group_num);
        }
      break;

      case OP_DNRREF:       /* Duplicate named group recursion test */
      if (mb->recursive != NULL)
        {
        int count = GET2(ecode, 1 + IMM2_SIZE);
        PCRE2_SPTR slot = mb->name_table + GET2(ecode, 1) * mb->name_entry_size;
        while (count-- > 0)
          {
          uint32_t recno = GET2(slot, 0);
          condition = recno == mb->recursive->group_num;
          if (condition) break;
          slot += mb->name_entry_size;
          }
        }
      break;

      case OP_CREF:                  /* Numbered group used test */
      offset = GET2(ecode, 1) << 1;  /* Doubled ref number */
      condition = offset < offset_top &&
        mb->ovector[offset] != PCRE2_UNSET;
      break;

      case OP_DNCREF:      /* Duplicate named group used test */
        {
        int count = GET2(ecode, 1 + IMM2_SIZE);
        PCRE2_SPTR slot = mb->name_table + GET2(ecode, 1) * mb->name_entry_size;
        while (count-- > 0)
          {
          offset = GET2(slot, 0) << 1;
          condition = offset < offset_top &&
            mb->ovector[offset] != PCRE2_UNSET;
          if (condition) break;
          slot += mb->name_entry_size;
          }
        }
      break;

      case OP_FALSE:
      case OP_FAIL:   /* The assertion (?!) becomes OP_FAIL */
      break;

      case OP_TRUE:
      condition = TRUE;
      break;

      /* The condition is an assertion. Call match() to evaluate it - setting
      the MATCH_CONDASSERT bit in mb->match_function_type causes it to stop at
      the end of an assertion. */

      default:
      mb->match_function_type |= MATCH_CONDASSERT;
      RMATCH(eptr, ecode, offset_top, mb, NULL, RM3);
      if (rrc == MATCH_MATCH)
        {
        if (mb->end_offset_top > offset_top)
          offset_top = mb->end_offset_top;  /* Captures may have happened */
        condition = TRUE;

        /* Advance ecode past the assertion to the start of the first branch,
        but adjust it so that the general choosing code below works. If the
        assertion has a quantifier that allows zero repeats we must skip over
        the BRAZERO. This is a lunatic thing to do, but somebody did! */

        if (*ecode == OP_BRAZERO) ecode++;
        ecode += GET(ecode, 1);
        while (*ecode == OP_ALT) ecode += GET(ecode, 1);
        ecode += 1 + LINK_SIZE - PRIV(OP_lengths)[condcode];
        }

      /* PCRE doesn't allow the effect of (*THEN) to escape beyond an
      assertion; it is therefore treated as NOMATCH. Any other return is an
      error. */

      else if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN)
        {
        RRETURN(rrc);         /* Need braces because of following else */
        }
      break;
      }

    /* Choose branch according to the condition */

    ecode += condition? PRIV(OP_lengths)[condcode] : codelink;

    /* We are now at the branch that is to be obeyed. As there is only one, we
    can use tail recursion to avoid using another stack frame, except when
    there is unlimited repeat of a possibly empty group. In the latter case, a
    recursive call to match() is always required, unless the second alternative
    doesn't exist, in which case we can just plough on. Note that, for
    compatibility with Perl, the | in a conditional group is NOT treated as
    creating two alternatives. If a THEN is encountered in the branch, it
    propagates out to the enclosing alternative (unless nested in a deeper set
    of alternatives, of course). */

    if (condition || ecode[-(1+LINK_SIZE)] == OP_ALT)
      {
      if (op != OP_SCOND)
        {
        goto TAIL_RECURSE;
        }

      mb->match_function_type |= MATCH_CBEGROUP;
      RMATCH(eptr, ecode, offset_top, mb, eptrb, RM49);
      RRETURN(rrc);
      }

     /* Condition false & no alternative; continue after the group. */

    else
      {
      }
    break;


    /* Before OP_ACCEPT there may be any number of OP_CLOSE opcodes,
    to close any currently open capturing brackets. */

    case OP_CLOSE:
    number = GET2(ecode, 1);   /* Must be less than 65536 */
    offset = number << 1;
    mb->capture_last = (mb->capture_last & OVFLMASK) | number;
    if (offset >= mb->offset_max) mb->capture_last |= OVFLBIT; else
      {
      mb->ovector[offset] =
        mb->ovector[mb->offset_end - number];
      mb->ovector[offset+1] = eptr - mb->start_subject;

      /* If this group is at or above the current highwater mark, ensure that
      any groups between the current high water mark and this group are marked
      unset and then update the high water mark. */

      if (offset >= offset_top)
        {
        PCRE2_SIZE *iptr = mb->ovector + offset_top;
        PCRE2_SIZE *iend = mb->ovector + offset;
        while (iptr < iend) *iptr++ = PCRE2_UNSET;
        offset_top = offset + 2;
        }
      }
    ecode += 1 + IMM2_SIZE;
    break;


    /* End of the pattern, either real or forced. In an assertion ACCEPT,
    update the last used pointer. */

    case OP_ASSERT_ACCEPT:
    if (eptr > mb->last_used_ptr) mb->last_used_ptr = eptr;

    case OP_ACCEPT:
    case OP_END:

    /* If we have matched an empty string, fail if not in an assertion and not
    in a recursion if either PCRE2_NOTEMPTY is set, or if PCRE2_NOTEMPTY_ATSTART
    is set and we have matched at the start of the subject. In both cases,
    backtracking will then try other alternatives, if any. */

    if (eptr == mstart && op != OP_ASSERT_ACCEPT &&
         mb->recursive == NULL &&
         ((mb->moptions & PCRE2_NOTEMPTY) != 0 ||
           ((mb->moptions & PCRE2_NOTEMPTY_ATSTART) != 0 &&
             mstart == mb->start_subject + mb->start_offset)))
      RRETURN(MATCH_NOMATCH);

    /* Otherwise, we have a match. */

    mb->end_match_ptr = eptr;           /* Record where we ended */
    mb->end_offset_top = offset_top;    /* and how many extracts were taken */
    mb->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 RRETURN when the heap is in use. */

    rrc = (op == OP_END)? MATCH_MATCH : MATCH_ACCEPT;
    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,
    the assertion is true. Lookbehind assertions have an OP_REVERSE item at the
    start of each branch to move the current point backwards, so the code at
    this level is identical to the lookahead case. When the assertion is part
    of a condition, we want to return immediately afterwards. The caller of
    this incarnation of the match() function will have set MATCH_CONDASSERT in
    mb->match_function type, and one of these opcodes will be the first opcode
    that is processed. We use a local variable that is preserved over calls to
    match() to remember this case. */

    case OP_ASSERT:
    case OP_ASSERTBACK:
    save_mark = mb->mark;
    if ((mb->match_function_type & MATCH_CONDASSERT) != 0)
      {
      condassert = TRUE;
      mb->match_function_type &= ~MATCH_CONDASSERT;
      }
    else condassert = FALSE;

    /* Loop for each branch */

    do
      {
      RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, mb, NULL, RM4);

      /* A match means that the assertion is true; break out of the loop
      that matches its alternatives. */

      if (rrc == MATCH_MATCH || rrc == MATCH_ACCEPT)
        {
        mstart = mb->start_match_ptr;   /* In case \K reset it */
        break;
        }

      /* If not matched, restore the previous mark setting. */

      mb->mark = save_mark;

      /* See comment in the code for capturing groups above about handling
      THEN. */

      if (rrc == MATCH_THEN)
        {
        next_ecode = ecode + GET(ecode,1);
        if (mb->start_match_ptr < next_ecode &&
            (*ecode == OP_ALT || *next_ecode == OP_ALT))
          rrc = MATCH_NOMATCH;
        }

      /* Anything other than NOMATCH causes the entire assertion to fail,
      passing back the return code. This includes COMMIT, SKIP, PRUNE and an
      uncaptured THEN, which means they take their normal effect. This
      consistent approach does not always have exactly the same effect as in
      Perl. */

      if (rrc != MATCH_NOMATCH) RRETURN(rrc);
      ecode += GET(ecode, 1);
      }
    while (*ecode == OP_ALT);   /* Continue for next alternative */

    /* If we have tried all the alternative branches, the assertion has
    failed. If not, we broke out after a match. */

    if (*ecode == OP_KET) RRETURN(MATCH_NOMATCH);

    /* If checking an assertion for a condition, return MATCH_MATCH. */

    if (condassert) RRETURN(MATCH_MATCH);

    /* Continue from after a successful assertion, updating the offsets high
    water mark, since extracts may have been taken during the assertion. */

    do ecode += GET(ecode,1); while (*ecode == OP_ALT);
    ecode += 1 + LINK_SIZE;
    offset_top = mb->end_offset_top;
    continue;

    /* Negative assertion: all branches must fail to match for the assertion to
    succeed. */

    case OP_ASSERT_NOT:
    case OP_ASSERTBACK_NOT:
    save_mark = mb->mark;
    if ((mb->match_function_type & MATCH_CONDASSERT) != 0)
      {
      condassert = TRUE;
      mb->match_function_type &= ~MATCH_CONDASSERT;
      }
    else condassert = FALSE;

    /* Loop for each alternative branch. */

    do
      {
      RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, mb, NULL, RM5);
      mb->mark = save_mark;   /* Always restore the mark setting */

      switch(rrc)
        {
        case MATCH_MATCH:            /* A successful match means */
        case MATCH_ACCEPT:           /* the assertion has failed. */
        RRETURN(MATCH_NOMATCH);

        case MATCH_NOMATCH:          /* Carry on with next branch */
        break;

        /* See comment in the code for capturing groups above about handling
        THEN. */

        case MATCH_THEN:
        next_ecode = ecode + GET(ecode,1);
        if (mb->start_match_ptr < next_ecode &&
            (*ecode == OP_ALT || *next_ecode == OP_ALT))
          {
          rrc = MATCH_NOMATCH;
          break;
          }
        /* Otherwise fall through. */

        /* COMMIT, SKIP, PRUNE, and an uncaptured THEN cause the whole
        assertion to fail to match, without considering any more alternatives.
        Failing to match means the assertion is true. This is a consistent
        approach, but does not always have the same effect as in Perl. */

        case MATCH_COMMIT:
        case MATCH_SKIP:
        case MATCH_SKIP_ARG:
        case MATCH_PRUNE:
        do ecode += GET(ecode,1); while (*ecode == OP_ALT);
        goto NEG_ASSERT_TRUE;   /* Break out of alternation loop */

        /* Anything else is an error */

        default:
        RRETURN(rrc);
        }

      /* Continue with next branch */

      ecode += GET(ecode,1);
      }
    while (*ecode == OP_ALT);

    /* All branches in the assertion failed to match. */

    NEG_ASSERT_TRUE:
    if (condassert) RRETURN(MATCH_MATCH);  /* Condition assertion */
    ecode += 1 + LINK_SIZE;                /* Continue with current branch */
    continue;

    /* Move the subject pointer back. This occurs only at the start of
    each branch of a lookbehind assertion. If we are too close to the start to
    move back, this match function fails. When working with UTF-8 we move
    back a number of characters, not bytes. */

    case OP_REVERSE:
    i = GET(ecode, 1);
#ifdef SUPPORT_UNICODE
    if (utf)
      {
      while (i-- > 0)
        {
        if (eptr <= mb->start_subject) RRETURN(MATCH_NOMATCH);
        eptr--;
        BACKCHAR(eptr);
        }
      }
    else
#endif

    /* No UTF-8 support, or not in UTF-8 mode: count is byte count */

      {
      if (i > eptr - mb->start_subject) RRETURN(MATCH_NOMATCH);
      eptr -= i;
      }

    /* Save the earliest consulted character, then skip to next op code */

    if (eptr < mb->start_used_ptr) mb->start_used_ptr = eptr;
    ecode += 1 + LINK_SIZE;
    break;

    /* The callout item calls an external function, if one is provided, passing
    details of the match so far. This is mainly for debugging, though the
    function is able to force a failure. */

    case OP_CALLOUT:
    case OP_CALLOUT_STR:
      {
      unsigned int callout_length = (*ecode == OP_CALLOUT)
          ? PRIV(OP_lengths)[OP_CALLOUT] : GET(ecode, 1 + 2*LINK_SIZE);

      if (mb->callout != NULL)
        {
        pcre2_callout_block cb;
        cb.version          = 1;
        cb.callout_number   = ecode[LINK_SIZE + 1];
        cb.capture_top      = (uint32_t)offset_top/2;
        cb.capture_last     = mb->capture_last & CAPLMASK;
        cb.offset_vector    = mb->ovector;
        cb.mark             = mb->nomatch_mark;
        cb.subject          = mb->start_subject;
        cb.subject_length   = (PCRE2_SIZE)(mb->end_subject - mb->start_subject);
        cb.start_match      = (PCRE2_SIZE)(mstart - mb->start_subject);
        cb.current_position = (PCRE2_SIZE)(eptr - mb->start_subject);
        cb.pattern_position = GET(ecode, 1);
        cb.next_item_length = GET(ecode, 1 + LINK_SIZE);

        if (*ecode == OP_CALLOUT)
          {
          cb.callout_number = ecode[1 + 2*LINK_SIZE];
          cb.callout_string_offset = 0;
          cb.callout_string = NULL;
          cb.callout_string_length = 0;
          }
        else
          {
          cb.callout_number = 0;
          cb.callout_string_offset = GET(ecode, 1 + 3*LINK_SIZE);
          cb.callout_string = ecode + (1 + 4*LINK_SIZE) + 1;
          cb.callout_string_length =
            callout_length - (1 + 4*LINK_SIZE) - 2;
          }

        if ((rrc = mb->callout(&cb, mb->callout_data)) > 0)
          RRETURN(MATCH_NOMATCH);
        if (rrc < 0) RRETURN(rrc);
        }
      ecode += callout_length;
      }
    break;

    /* Recursion either matches the current regex, or some subexpression. The
    offset data is the offset to the starting bracket from the start of the
    whole pattern. (This is so that it works from duplicated subpatterns.)

    The state of the capturing groups is preserved over recursion, and
    re-instated afterwards. We don't know how many are started and not yet
    finished (offset_top records the completed total) so we just have to save
    all the potential data. There may be up to 65535 such values, which is too
    large to put on the stack, but using malloc for small numbers seems
    expensive. As a compromise, the stack is used when there are no more than
    OP_RECURSE_STACK_SAVE_MAX values to store; otherwise malloc is used.

    There are also other values that have to be saved. We use a chained
    sequence of blocks that actually live on the stack. Thanks to Robin Houston
    for the original version of this logic. It has, however, been hacked around
    a lot, so he is not to blame for the current way it works. */

    case OP_RECURSE:
      {
      ovecsave_frame *fr;
      recursion_info *ri;
      uint32_t recno;

      callpat = mb->start_code + GET(ecode, 1);
      recno = (callpat == mb->start_code)? 0 : GET2(callpat, 1 + LINK_SIZE);

      /* Check for repeating a pattern recursion without advancing the subject
      pointer. This should catch convoluted mutual recursions. (Some simple
      cases are caught at compile time.) */

      for (ri = mb->recursive; ri != NULL; ri = ri->prevrec)
        if (recno == ri->group_num && eptr == ri->subject_position)
          RRETURN(PCRE2_ERROR_RECURSELOOP);

      /* Add to "recursing stack" */

      new_recursive.group_num = recno;
      new_recursive.saved_capture_last = mb->capture_last;
      new_recursive.subject_position = eptr;
      new_recursive.prevrec = mb->recursive;
      mb->recursive = &new_recursive;

      /* Where to continue from afterwards */

      ecode += 1 + LINK_SIZE;

      /* When we are using the system stack for match() recursion we can call a
      function that uses the system stack for preserving the ovector while
      processing the pattern recursion, but only if the ovector is small
      enough. */

#ifndef HEAP_MATCH_RECURSE
      if (mb->offset_end <= OP_RECURSE_STACK_SAVE_MAX)
        {
        rrc = op_recurse_ovecsave(eptr, callpat, mstart, offset_top, mb,
          eptrb, rdepth);
        mb->recursive = new_recursive.prevrec;
        if (rrc != MATCH_MATCH && rrc != MATCH_ACCEPT) RRETURN(rrc);

        /* 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,
        for Perl compatibility. */

        eptr = mb->end_match_ptr;
        mstart = mb->start_match_ptr;
        break;   /* End of processing OP_RECURSE */
        }
#endif
      /* If the ovector is too big, or if we are using the heap for match()
      recursion, we have to use the heap for saving the ovector. Used ovecsave
      frames are kept on a chain and re-used. This makes a small improvement in
      execution time on Linux. */

      if (mb->ovecsave_chain != NULL)
        {
        new_recursive.ovec_save = mb->ovecsave_chain->saved_ovec;
        mb->ovecsave_chain = mb->ovecsave_chain->next;
        }
      else
        {
        fr = (ovecsave_frame *)(mb->memctl.malloc(sizeof(ovecsave_frame *) +
          mb->offset_end * sizeof(PCRE2_SIZE), mb->memctl.memory_data));
        if (fr == NULL) RRETURN(PCRE2_ERROR_NOMEMORY);
        new_recursive.ovec_save = fr->saved_ovec;
        }

      memcpy(new_recursive.ovec_save, mb->ovector,
        mb->offset_end * sizeof(PCRE2_SIZE));

      /* Do the recursion. After processing each alternative, restore the
      ovector data and the last captured value. This code has the same overall
      logic as the code in the op_recurse_ovecsave() function, but is adapted
      to use RMATCH/RRETURN and to release the heap block containing the saved
      ovector. */

      cbegroup = (*callpat >= OP_SBRA);
      do
        {
        if (cbegroup) mb->match_function_type |= MATCH_CBEGROUP;
        RMATCH(eptr, callpat + PRIV(OP_lengths)[*callpat], offset_top,
          mb, eptrb, RM6);
        memcpy(mb->ovector, new_recursive.ovec_save,
            mb->offset_end * sizeof(PCRE2_SIZE));
        mb->capture_last = new_recursive.saved_capture_last;
        mb->recursive = new_recursive.prevrec;

        if (rrc == MATCH_MATCH || rrc == MATCH_ACCEPT)
          {
          fr = (ovecsave_frame *)
            ((uint8_t *)new_recursive.ovec_save - sizeof(ovecsave_frame *));
          fr->next = mb->ovecsave_chain;
          mb->ovecsave_chain = fr;

          /* 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,
          for Perl compatibility. */

          eptr = mb->end_match_ptr;
          mstart = mb->start_match_ptr;
          goto RECURSION_MATCHED;        /* Exit loop; end processing */
          }

        /* PCRE does not allow THEN, SKIP, PRUNE or COMMIT to escape beyond a
        recursion; they cause a NOMATCH for the entire recursion. These codes
        are defined in a range that can be tested for. */

        if (rrc >= MATCH_BACKTRACK_MIN && rrc <= MATCH_BACKTRACK_MAX)
          {
          rrc = MATCH_NOMATCH;
          goto RECURSION_RETURN;
          }

        /* Any return code other than NOMATCH is an error. */

        if (rrc != MATCH_NOMATCH) goto RECURSION_RETURN;
        mb->recursive = &new_recursive;
        callpat += GET(callpat, 1);
        }
      while (*callpat == OP_ALT);

      RECURSION_RETURN:
      mb->recursive = new_recursive.prevrec;
      fr = (ovecsave_frame *)
        ((uint8_t *)new_recursive.ovec_save - sizeof(ovecsave_frame *));
      fr->next = mb->ovecsave_chain;
      mb->ovecsave_chain = fr;
      RRETURN(rrc);
      }

    RECURSION_MATCHED:
    break;

    /* An alternation is the end of a branch; scan along to find the end of the
    bracketed group and go to there. */

    case OP_ALT:
    do ecode += GET(ecode,1); while (*ecode == OP_ALT);
    break;

    /* BRAZERO, BRAMINZERO and SKIPZERO occur just before a bracket group,
    indicating that it may occur zero times. It may repeat infinitely, or not
    at all - i.e. it could be ()* or ()? or even (){0} in the pattern. Brackets
    with fixed upper repeat limits are compiled as a number of copies, with the
    optional ones preceded by BRAZERO or BRAMINZERO. */

    case OP_BRAZERO:
    next_ecode = ecode + 1;
    RMATCH(eptr, next_ecode, offset_top, mb, eptrb, RM10);
    if (rrc != MATCH_NOMATCH) RRETURN(rrc);
    do next_ecode += GET(next_ecode, 1); while (*next_ecode == OP_ALT);
    ecode = next_ecode + 1 + LINK_SIZE;
    break;

    case OP_BRAMINZERO:
    next_ecode = ecode + 1;
    do next_ecode += GET(next_ecode, 1); while (*next_ecode == OP_ALT);
    RMATCH(eptr, next_ecode + 1+LINK_SIZE, offset_top, mb, eptrb, RM11);
    if (rrc != MATCH_NOMATCH) RRETURN(rrc);
    ecode++;
    break;

    case OP_SKIPZERO:
    next_ecode = ecode+1;
    do next_ecode += GET(next_ecode,1); while (*next_ecode == OP_ALT);
    ecode = next_ecode + 1 + LINK_SIZE;
    break;

    /* BRAPOSZERO occurs before a possessive bracket group. Don't do anything
    here; just jump to the group, with allow_zero set TRUE. */

    case OP_BRAPOSZERO:
    op = *(++ecode);
    allow_zero = TRUE;
    if (op == OP_CBRAPOS || op == OP_SCBRAPOS) goto POSSESSIVE_CAPTURE;
      goto POSSESSIVE_NON_CAPTURE;

    /* End of a group, repeated or non-repeating. */

    case OP_KET:
    case OP_KETRMIN:
    case OP_KETRMAX:
    case OP_KETRPOS:
    prev = ecode - GET(ecode, 1);

    /* If this was a group that remembered the subject start, in order to break
    infinite repeats of empty string matches, retrieve the subject start from
    the chain. Otherwise, set it NULL. */

    if (*prev >= OP_SBRA || *prev == OP_ONCE)
      {
      saved_eptr = eptrb->epb_saved_eptr;   /* Value at start of group */
      eptrb = eptrb->epb_prev;              /* Backup to previous group */
      }
    else saved_eptr = NULL;

    /* If we are at the end of an assertion group or a non-capturing atomic
    group, stop matching and return MATCH_MATCH, but record the current high
    water mark for use by positive assertions. We also need to record the match
    start in case it was changed by \K. */

    if ((*prev >= OP_ASSERT && *prev <= OP_ASSERTBACK_NOT) ||
         *prev == OP_ONCE_NC)
      {
      mb->end_match_ptr = eptr;      /* For ONCE_NC */
      mb->end_offset_top = offset_top;
      mb->start_match_ptr = mstart;
      if (eptr > mb->last_used_ptr) mb->last_used_ptr = eptr;
      RRETURN(MATCH_MATCH);         /* Sets mb->mark */
      }

    /* For capturing groups we have to check the group number back at the start
    and if necessary complete handling an extraction by setting the offsets and
    bumping the high water mark. Whole-pattern recursion is coded as a recurse
    into group 0, so it won't be picked up here. Instead, we catch it when the
    OP_END is reached. Other recursion is handled here. We just have to record
    the current subject position and start match pointer and give a MATCH
    return. */

    if (*prev == OP_CBRA || *prev == OP_SCBRA ||
        *prev == OP_CBRAPOS || *prev == OP_SCBRAPOS)
      {
      number = GET2(prev, 1+LINK_SIZE);
      offset = number << 1;

      /* Handle a recursively called group. */

      if (mb->recursive != NULL && mb->recursive->group_num == number)
        {
        mb->end_match_ptr = eptr;
        mb->start_match_ptr = mstart;
        if (eptr > mb->last_used_ptr) mb->last_used_ptr = eptr;
        RRETURN(MATCH_MATCH);
        }

      /* Deal with capturing */

      mb->capture_last = (mb->capture_last & OVFLMASK) | number;
      if (offset >= mb->offset_max) mb->capture_last |= OVFLBIT; else
        {
        /* If offset is greater than offset_top, it means that we are
        "skipping" a capturing group, and that group's offsets must be marked
        unset. In earlier versions of PCRE, all the offsets were unset at the
        start of matching, but this doesn't work because atomic groups and
        assertions can cause a value to be set that should later be unset.
        Example: matching /(?>(a))b|(a)c/ against "ac". This sets group 1 as
        part of the atomic group, but this is not on the final matching path,
        so must be unset when 2 is set. (If there is no group 2, there is no
        problem, because offset_top will then be 2, indicating no capture.) */

        if (offset > offset_top)
          {
          PCRE2_SIZE *iptr = mb->ovector + offset_top;
          PCRE2_SIZE *iend = mb->ovector + offset;
          while (iptr < iend) *iptr++ = PCRE2_UNSET;
          }

        /* Now make the extraction */

        mb->ovector[offset] = mb->ovector[mb->offset_end - number];
        mb->ovector[offset+1] = eptr - mb->start_subject;
        if (offset_top <= offset) offset_top = offset + 2;
        }
      }

    /* OP_KETRPOS is a possessive repeating ket. Remember the current position,
    and return the MATCH_KETRPOS. This makes it possible to do the repeats one
    at a time from the outer level, thus saving stack. This must precede the
    empty string test - in this case that test is done at the outer level. */

    if (*ecode == OP_KETRPOS)
      {
      mb->start_match_ptr = mstart;    /* In case \K reset it */
      mb->end_match_ptr = eptr;
      mb->end_offset_top = offset_top;
      if (eptr > mb->last_used_ptr) mb->last_used_ptr = eptr;
      RRETURN(MATCH_KETRPOS);
      }

    /* For an ordinary non-repeating ket, just continue at this level. This
    also happens for a repeating ket if no characters were matched in the
    group. This is the forcible breaking of infinite loops as implemented in
    Perl 5.005. For a non-repeating atomic group that includes captures,
    establish a backup point by processing the rest of the pattern at a lower
    level. If this results in a NOMATCH return, pass MATCH_ONCE back to the
    original OP_ONCE level, thereby bypassing intermediate backup points, but
    resetting any captures that happened along the way. */

    if (*ecode == OP_KET || eptr == saved_eptr)
      {
      if (*prev == OP_ONCE)
        {
        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, mb, eptrb, RM12);
        if (rrc != MATCH_NOMATCH) RRETURN(rrc);
        mb->once_target = prev;  /* Level at which to change to MATCH_NOMATCH */
        RRETURN(MATCH_ONCE);
        }
      ecode += 1 + LINK_SIZE;    /* Carry on at this level */
      break;
      }

    /* The normal repeating kets try the rest of the pattern or restart from
    the preceding bracket, in the appropriate order. In the second case, we can
    use tail recursion to avoid using another stack frame, unless we have an
    an atomic group or an unlimited repeat of a group that can match an empty
    string. */

    if (*ecode == OP_KETRMIN)
      {
      RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, mb, eptrb, RM7);
      if (rrc != MATCH_NOMATCH) RRETURN(rrc);
      if (*prev == OP_ONCE)
        {
        RMATCH(eptr, prev, offset_top, mb, eptrb, RM8);
        if (rrc != MATCH_NOMATCH) RRETURN(rrc);
        mb->once_target = prev;  /* Level at which to change to MATCH_NOMATCH */
        RRETURN(MATCH_ONCE);
        }
      if (*prev >= OP_SBRA)    /* Could match an empty string */
        {
        RMATCH(eptr, prev, offset_top, mb, eptrb, RM50);
        RRETURN(rrc);
        }
      ecode = prev;
      goto TAIL_RECURSE;
      }
    else  /* OP_KETRMAX */
      {
      RMATCH(eptr, prev, offset_top, mb, eptrb, RM13);
      if (rrc == MATCH_ONCE && mb->once_target == prev) rrc = MATCH_NOMATCH;
      if (rrc != MATCH_NOMATCH) RRETURN(rrc);
      if (*prev == OP_ONCE)
        {
        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, mb, eptrb, RM9);
        if (rrc != MATCH_NOMATCH) RRETURN(rrc);
        mb->once_target = prev;
        RRETURN(MATCH_ONCE);
        }
      ecode += 1 + LINK_SIZE;
      goto TAIL_RECURSE;
      }
    /* Control never gets here */

    /* Not multiline mode: start of subject assertion, unless notbol. */

    case OP_CIRC:
    if ((mb->moptions & PCRE2_NOTBOL) != 0 && eptr == mb->start_subject)
      RRETURN(MATCH_NOMATCH);

    /* Start of subject assertion */

    case OP_SOD:
    if (eptr != mb->start_subject) RRETURN(MATCH_NOMATCH);
    ecode++;
    break;

    /* Multiline mode: start of subject unless notbol, or after any newline
    except for one at the very end, unless PCRE2_ALT_CIRCUMFLEX is set. */

    case OP_CIRCM:
    if ((mb->moptions & PCRE2_NOTBOL) != 0 && eptr == mb->start_subject)
      RRETURN(MATCH_NOMATCH);
    if (eptr != mb->start_subject &&
        ((eptr == mb->end_subject &&
           (mb->poptions & PCRE2_ALT_CIRCUMFLEX) == 0) ||
         !WAS_NEWLINE(eptr)))
      RRETURN(MATCH_NOMATCH);
    ecode++;
    break;

    /* Start of match assertion */

    case OP_SOM:
    if (eptr != mb->start_subject + mb->start_offset) RRETURN(MATCH_NOMATCH);
    ecode++;
    break;

    /* Reset the start of match point */

    case OP_SET_SOM:
    mstart = eptr;
    ecode++;
    break;

    /* Multiline mode: assert before any newline, or before end of subject
    unless noteol is set. */

    case OP_DOLLM:
    if (eptr < mb->end_subject)
      {
      if (!IS_NEWLINE(eptr))
        {
        if (mb->partial != 0 &&
            eptr + 1 >= mb->end_subject &&
            NLBLOCK->nltype == NLTYPE_FIXED &&
            NLBLOCK->nllen == 2 &&
            UCHAR21TEST(eptr) == NLBLOCK->nl[0])
          {
          mb->hitend = TRUE;
          if (mb->partial > 1) RRETURN(PCRE2_ERROR_PARTIAL);
          }
        RRETURN(MATCH_NOMATCH);
        }
      }
    else
      {
      if ((mb->moptions & PCRE2_NOTEOL) != 0) RRETURN(MATCH_NOMATCH);
      SCHECK_PARTIAL();
      }
    ecode++;
    break;

    /* Not multiline mode: assert before a terminating newline or before end of
    subject unless noteol is set. */

    case OP_DOLL:
    if ((mb->moptions & PCRE2_NOTEOL) != 0) RRETURN(MATCH_NOMATCH);
    if ((mb->poptions & PCRE2_DOLLAR_ENDONLY) == 0) goto ASSERT_NL_OR_EOS;

    /* ... else fall through for endonly */

    /* End of subject assertion (\z) */

    case OP_EOD:
    if (eptr < mb->end_subject) RRETURN(MATCH_NOMATCH);
    SCHECK_PARTIAL();
    ecode++;
    break;

    /* End of subject or ending \n assertion (\Z) */

    case OP_EODN:
    ASSERT_NL_OR_EOS:
    if (eptr < mb->end_subject &&
        (!IS_NEWLINE(eptr) || eptr != mb->end_subject - mb->nllen))
      {
      if (mb->partial != 0 &&
          eptr + 1 >= mb->end_subject &&
          NLBLOCK->nltype == NLTYPE_FIXED &&
          NLBLOCK->nllen == 2 &&
          UCHAR21TEST(eptr) == NLBLOCK->nl[0])
        {
        mb->hitend = TRUE;
        if (mb->partial > 1) RRETURN(PCRE2_ERROR_PARTIAL);
        }
      RRETURN(MATCH_NOMATCH);
      }

    /* Either at end of string or \n before end. */

    SCHECK_PARTIAL();
    ecode++;
    break;

    /* Word boundary assertions */

    case OP_NOT_WORD_BOUNDARY:
    case OP_WORD_BOUNDARY:
      {

      /* Find out if the previous and current characters are "word" characters.
      It takes a bit more work in UTF-8 mode. Characters > 255 are assumed to
      be "non-word" characters. Remember the earliest consulted character for
      partial matching. */

#ifdef SUPPORT_UNICODE
      if (utf)
        {
        /* Get status of previous character */

        if (eptr == mb->start_subject) prev_is_word = FALSE; else
          {
          PCRE2_SPTR lastptr = eptr - 1;
          BACKCHAR(lastptr);
          if (lastptr < mb->start_used_ptr) mb->start_used_ptr = lastptr;
          GETCHAR(c, lastptr);
          if ((mb->poptions & PCRE2_UCP) != 0)
            {
            if (c == '_') prev_is_word = TRUE; else
              {
              int cat = UCD_CATEGORY(c);
              prev_is_word = (cat == ucp_L || cat == ucp_N);
              }
            }
          else
          prev_is_word = c < 256 && (mb->ctypes[c] & ctype_word) != 0;
          }

        /* Get status of next character */

        if (eptr >= mb->end_subject)
          {
          SCHECK_PARTIAL();
          cur_is_word = FALSE;
          }
        else
          {
          PCRE2_SPTR nextptr = eptr + 1;
          FORWARDCHARTEST(nextptr, mb->end_subject);
          if (nextptr > mb->last_used_ptr) mb->last_used_ptr = nextptr;
          GETCHAR(c, eptr);
          if ((mb->poptions & PCRE2_UCP) != 0)
            {
            if (c == '_') cur_is_word = TRUE; else
              {
              int cat = UCD_CATEGORY(c);
              cur_is_word = (cat == ucp_L || cat == ucp_N);
              }
            }
          else
          cur_is_word = c < 256 && (mb->ctypes[c] & ctype_word) != 0;
          }
        }
      else
#endif  /* SUPPORT UTF */

      /* Not in UTF-8 mode, but we may still have PCRE2_UCP set, and for
      consistency with the behaviour of \w we do use it in this case. */

        {
        /* Get status of previous character */

        if (eptr == mb->start_subject) prev_is_word = FALSE; else
          {
          if (eptr <= mb->start_used_ptr) mb->start_used_ptr = eptr - 1;
#ifdef SUPPORT_UNICODE
          if ((mb->poptions & PCRE2_UCP) != 0)
            {
            c = eptr[-1];
            if (c == '_') prev_is_word = TRUE; else
              {
              int cat = UCD_CATEGORY(c);
              prev_is_word = (cat == ucp_L || cat == ucp_N);
              }
            }
          else
#endif
          prev_is_word = MAX_255(eptr[-1])
            && ((mb->ctypes[eptr[-1]] & ctype_word) != 0);
          }

        /* Get status of next character */

        if (eptr >= mb->end_subject)
          {
          SCHECK_PARTIAL();
          cur_is_word = FALSE;
          }
        else
          {
          if (eptr >= mb->last_used_ptr) mb->last_used_ptr = eptr + 1;
#ifdef SUPPORT_UNICODE
          if ((mb->poptions & PCRE2_UCP) != 0)
            {
            c = *eptr;
            if (c == '_') cur_is_word = TRUE; else
              {
              int cat = UCD_CATEGORY(c);
              cur_is_word = (cat == ucp_L || cat == ucp_N);
              }
            }
          else
#endif
          cur_is_word = MAX_255(*eptr)
            && ((mb->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)
        RRETURN(MATCH_NOMATCH);
      }
    break;

    /* Match any single character type except newline; have to take care with
    CRLF newlines and partial matching. */

    case OP_ANY:
    if (IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH);
    if (mb->partial != 0 &&
        eptr == mb->end_subject - 1 &&
        NLBLOCK->nltype == NLTYPE_FIXED &&
        NLBLOCK->nllen == 2 &&
        UCHAR21TEST(eptr) == NLBLOCK->nl[0])
      {
      mb->hitend = TRUE;
      if (mb->partial > 1) RRETURN(PCRE2_ERROR_PARTIAL);
      }

    /* Fall through */

    /* Match any single character whatsoever. */

    case OP_ALLANY:
    if (eptr >= mb->end_subject)   /* DO NOT merge the eptr++ here; it must */
      {                            /* not be updated before SCHECK_PARTIAL. */
      SCHECK_PARTIAL();
      RRETURN(MATCH_NOMATCH);
      }
    eptr++;
#ifdef SUPPORT_UNICODE
    if (utf) ACROSSCHAR(eptr < mb->end_subject, *eptr, eptr++);
#endif
    ecode++;
    break;

    /* Match a single code unit, even in UTF-8 mode. This opcode really does
    match any code unit, even newline. (It really should be called ANYCODEUNIT,
    of course - the byte name is from pre-16 bit days.) */

    case OP_ANYBYTE:
    if (eptr >= mb->end_subject)   /* DO NOT merge the eptr++ here; it must */
      {                            /* not be updated before SCHECK_PARTIAL. */
      SCHECK_PARTIAL();
      RRETURN(MATCH_NOMATCH);
      }
    eptr++;
    ecode++;
    break;

    case OP_NOT_DIGIT:
    if (eptr >= mb->end_subject)
      {
      SCHECK_PARTIAL();
      RRETURN(MATCH_NOMATCH);
      }
    GETCHARINCTEST(c, eptr);
    if (
#ifdef SUPPORT_WIDE_CHARS
       c < 256 &&
#endif
       (mb->ctypes[c] & ctype_digit) != 0
       )
      RRETURN(MATCH_NOMATCH);
    ecode++;
    break;

    case OP_DIGIT:
    if (eptr >= mb->end_subject)
      {
      SCHECK_PARTIAL();
      RRETURN(MATCH_NOMATCH);
      }
    GETCHARINCTEST(c, eptr);
    if (
#ifdef SUPPORT_WIDE_CHARS
       c > 255 ||
#endif
       (mb->ctypes[c] & ctype_digit) == 0
       )
      RRETURN(MATCH_NOMATCH);
    ecode++;
    break;

    case OP_NOT_WHITESPACE:
    if (eptr >= mb->end_subject)
      {
      SCHECK_PARTIAL();
      RRETURN(MATCH_NOMATCH);
      }
    GETCHARINCTEST(c, eptr);
    if (
#ifdef SUPPORT_WIDE_CHARS
       c < 256 &&
#endif
       (mb->ctypes[c] & ctype_space) != 0
       )
      RRETURN(MATCH_NOMATCH);
    ecode++;
    break;

    case OP_WHITESPACE:
    if (eptr >= mb->end_subject)
      {
      SCHECK_PARTIAL();
      RRETURN(MATCH_NOMATCH);
      }
    GETCHARINCTEST(c, eptr);
    if (
#ifdef SUPPORT_WIDE_CHARS
       c > 255 ||
#endif
       (mb->ctypes[c] & ctype_space) == 0
       )
      RRETURN(MATCH_NOMATCH);
    ecode++;
    break;

    case OP_NOT_WORDCHAR:
    if (eptr >= mb->end_subject)
      {
      SCHECK_PARTIAL();
      RRETURN(MATCH_NOMATCH);
      }
    GETCHARINCTEST(c, eptr);
    if (
#ifdef SUPPORT_WIDE_CHARS
       c < 256 &&
#endif
       (mb->ctypes[c] & ctype_word) != 0
       )
      RRETURN(MATCH_NOMATCH);
    ecode++;
    break;

    case OP_WORDCHAR:
    if (eptr >= mb->end_subject)
      {
      SCHECK_PARTIAL();
      RRETURN(MATCH_NOMATCH);
      }
    GETCHARINCTEST(c, eptr);
    if (
#ifdef SUPPORT_WIDE_CHARS
       c > 255 ||
#endif
       (mb->ctypes[c] & ctype_word) == 0
       )
      RRETURN(MATCH_NOMATCH);
    ecode++;
    break;

    case OP_ANYNL:
    if (eptr >= mb->end_subject)
      {
      SCHECK_PARTIAL();
      RRETURN(MATCH_NOMATCH);
      }
    GETCHARINCTEST(c, eptr);
    switch(c)
      {
      default: RRETURN(MATCH_NOMATCH);

      case CHAR_CR:
      if (eptr >= mb->end_subject)
        {
        SCHECK_PARTIAL();
        }
      else if (UCHAR21TEST(eptr) == CHAR_LF) eptr++;
      break;

      case CHAR_LF:
      break;

      case CHAR_VT:
      case CHAR_FF:
      case CHAR_NEL:
#ifndef EBCDIC
      case 0x2028:
      case 0x2029:
#endif  /* Not EBCDIC */
      if (mb->bsr_convention == PCRE2_BSR_ANYCRLF) RRETURN(MATCH_NOMATCH);
      break;
      }
    ecode++;
    break;

    case OP_NOT_HSPACE:
    if (eptr >= mb->end_subject)
      {
      SCHECK_PARTIAL();
      RRETURN(MATCH_NOMATCH);
      }
    GETCHARINCTEST(c, eptr);
    switch(c)
      {
      HSPACE_CASES: RRETURN(MATCH_NOMATCH);  /* Byte and multibyte cases */
      default: break;
      }
    ecode++;
    break;

    case OP_HSPACE:
    if (eptr >= mb->end_subject)
      {
      SCHECK_PARTIAL();
      RRETURN(MATCH_NOMATCH);
      }
    GETCHARINCTEST(c, eptr);
    switch(c)
      {
      HSPACE_CASES: break;  /* Byte and multibyte cases */
      default: RRETURN(MATCH_NOMATCH);
      }
    ecode++;
    break;

    case OP_NOT_VSPACE:
    if (eptr >= mb->end_subject)
      {
      SCHECK_PARTIAL();
      RRETURN(MATCH_NOMATCH);
      }
    GETCHARINCTEST(c, eptr);
    switch(c)
      {
      VSPACE_CASES: RRETURN(MATCH_NOMATCH);
      default: break;
      }
    ecode++;
    break;

    case OP_VSPACE:
    if (eptr >= mb->end_subject)
      {
      SCHECK_PARTIAL();
      RRETURN(MATCH_NOMATCH);
      }
    GETCHARINCTEST(c, eptr);
    switch(c)
      {
      VSPACE_CASES: break;
      default: RRETURN(MATCH_NOMATCH);
      }
    ecode++;
    break;

#ifdef SUPPORT_UNICODE
    /* Check the next character by Unicode property. We will get here only
    if the support is in the binary; otherwise a compile-time error occurs. */

    case OP_PROP:
    case OP_NOTPROP:
    if (eptr >= mb->end_subject)
      {
      SCHECK_PARTIAL();
      RRETURN(MATCH_NOMATCH);
      }
    GETCHARINCTEST(c, eptr);
      {
      const uint32_t *cp;
      const ucd_record *prop = GET_UCD(c);

      switch(ecode[1])
        {
        case PT_ANY:
        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))
          RRETURN(MATCH_NOMATCH);
        break;

        case PT_GC:
        if ((ecode[2] != PRIV(ucp_gentype)[prop->chartype]) == (op == OP_PROP))
          RRETURN(MATCH_NOMATCH);
        break;

        case PT_PC:
        if ((ecode[2] != prop->chartype) == (op == OP_PROP))
          RRETURN(MATCH_NOMATCH);
        break;

        case PT_SC:
        if ((ecode[2] != prop->script) == (op == OP_PROP))
          RRETURN(MATCH_NOMATCH);
        break;

        /* These are specials */

        case PT_ALNUM:
        if ((PRIV(ucp_gentype)[prop->chartype] == ucp_L ||
             PRIV(ucp_gentype)[prop->chartype] == ucp_N) == (op == OP_NOTPROP))
          RRETURN(MATCH_NOMATCH);
        break;

        /* Perl space used to exclude VT, but from Perl 5.18 it is included,
        which means that Perl space and POSIX space are now identical. PCRE
        was changed at release 8.34. */

        case PT_SPACE:    /* Perl space */
        case PT_PXSPACE:  /* POSIX space */
        switch(c)
          {
          HSPACE_CASES:
          VSPACE_CASES:
          if (op == OP_NOTPROP) RRETURN(MATCH_NOMATCH);
          break;

          default:
          if ((PRIV(ucp_gentype)[prop->chartype] == ucp_Z) ==
            (op == OP_NOTPROP)) RRETURN(MATCH_NOMATCH);
          break;
          }
        break;

        case PT_WORD:
        if ((PRIV(ucp_gentype)[prop->chartype] == ucp_L ||
             PRIV(ucp_gentype)[prop->chartype] == ucp_N ||
             c == CHAR_UNDERSCORE) == (op == OP_NOTPROP))
          RRETURN(MATCH_NOMATCH);
        break;

        case PT_CLIST:
        cp = PRIV(ucd_caseless_sets) + ecode[2];
        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;

        case PT_UCNC:
        if ((c == CHAR_DOLLAR_SIGN || c == CHAR_COMMERCIAL_AT ||
             c == CHAR_GRAVE_ACCENT || (c >= 0xa0 && c <= 0xd7ff) ||
             c >= 0xe000) == (op == OP_NOTPROP))
          RRETURN(MATCH_NOMATCH);
        break;

        /* This should never occur */

        default:
        RRETURN(PCRE2_ERROR_INTERNAL);
        }

      ecode += 3;
      }
    break;

    /* Match an extended Unicode sequence. We will get here only if the support
    is in the binary; otherwise a compile-time error occurs. */

    case OP_EXTUNI:
    if (eptr >= mb->end_subject)
      {
      SCHECK_PARTIAL();
      RRETURN(MATCH_NOMATCH);
      }
    else
      {
      int lgb, rgb;
      GETCHARINCTEST(c, eptr);
      lgb = UCD_GRAPHBREAK(c);
      while (eptr < mb->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  /* SUPPORT_UNICODE */


    /* Match a back reference, possibly repeatedly. Look past the end of the
    item to see if there is repeat information following.

    The OP_REF and OP_REFI opcodes are used for a reference to a numbered group
    or to a non-duplicated named group. For a duplicated named group, OP_DNREF
    and OP_DNREFI are used. In this case we must scan the list of groups to
    which the name refers, and use the first one that is set. */

    case OP_DNREF:
    case OP_DNREFI:
    caseless = op == OP_DNREFI;
      {
      int count = GET2(ecode, 1+IMM2_SIZE);
      PCRE2_SPTR slot = mb->name_table + GET2(ecode, 1) * mb->name_entry_size;
      ecode += 1 + 2*IMM2_SIZE;

      /* Initializing 'offset' avoids a compiler warning in the REF_REPEAT
      code. */

      offset = 0;
      while (count-- > 0)
        {
        offset = GET2(slot, 0) << 1;
        if (offset < offset_top && mb->ovector[offset] != PCRE2_UNSET) break;
        slot += mb->name_entry_size;
        }
      }
    goto REF_REPEAT;

    case OP_REF:
    case OP_REFI:
    caseless = op == OP_REFI;
    offset = GET2(ecode, 1) << 1;               /* Doubled ref number */
    ecode += 1 + IMM2_SIZE;

    /* Set up for repetition, or handle the non-repeated case */

    REF_REPEAT:
    switch (*ecode)
      {
      case OP_CRSTAR:
      case OP_CRMINSTAR:
      case OP_CRPLUS:
      case OP_CRMINPLUS:
      case OP_CRQUERY:
      case OP_CRMINQUERY:
      c = *ecode++ - OP_CRSTAR;
      minimize = (c & 1) != 0;
      min = rep_min[c];                 /* Pick up values from tables; */
      max = rep_max[c];                 /* zero for max => infinity */
      if (max == 0) max = INT_MAX;
      break;

      case OP_CRRANGE:
      case OP_CRMINRANGE:
      minimize = (*ecode == OP_CRMINRANGE);
      min = GET2(ecode, 1);
      max = GET2(ecode, 1 + IMM2_SIZE);
      if (max == 0) max = INT_MAX;
      ecode += 1 + 2 * IMM2_SIZE;
      break;

      default:                  /* No repeat follows */
        {
        int rc = match_ref(offset, offset_top, eptr, mb, caseless, &length);
        if (rc != 0)
          {
          if (rc > 0) eptr = mb->end_subject;   /* Partial match */
          CHECK_PARTIAL();
          RRETURN(MATCH_NOMATCH);
          }
        }
      eptr += length;
      continue;              /* With the main loop */
      }

    /* Handle repeated back references. If a set group has length zero, just
    continue with the main loop, because it matches however many times. For an
    unset reference, if the minimum is zero, we can also just continue. We an
    also continue if PCRE2_MATCH_UNSET_BACKREF is set, because this makes unset
    group be have as a zero-length group. For any other unset cases, carrying
    on will result in NOMATCH. */

    if (offset < offset_top && mb->ovector[offset] != PCRE2_UNSET)
      {
      if (mb->ovector[offset] == mb->ovector[offset + 1]) continue;
      }
    else  /* Group is not set */
      {
      if (min == 0 || (mb->poptions & PCRE2_MATCH_UNSET_BACKREF) != 0)
        continue;
      }

    /* First, ensure the minimum number of matches are present. */

    for (i = 1; i <= min; i++)
      {
      PCRE2_SIZE slength;
      int rc = match_ref(offset, offset_top, eptr, mb, caseless, &slength);
      if (rc != 0)
        {
        if (rc > 0) eptr = mb->end_subject;   /* Partial match */
        CHECK_PARTIAL();
        RRETURN(MATCH_NOMATCH);
        }
      eptr += slength;
      }

    /* If min = max, continue at the same level without recursion.
    They are not both allowed to be zero. */

    if (min == max) continue;

    /* If minimizing, keep trying and advancing the pointer */

    if (minimize)
      {
      for (fi = min;; fi++)
        {
        int rc;
        PCRE2_SIZE slength;
        RMATCH(eptr, ecode, offset_top, mb, eptrb, RM14);
        if (rrc != MATCH_NOMATCH) RRETURN(rrc);
        if (fi >= max) RRETURN(MATCH_NOMATCH);
        rc = match_ref(offset, offset_top, eptr, mb, caseless, &slength);
        if (rc != 0)
          {
          if (rc > 0) eptr = mb->end_subject;   /* Partial match */
          CHECK_PARTIAL();
          RRETURN(MATCH_NOMATCH);
          }
        eptr += slength;
        }
      /* Control never gets here */
      }

    /* If maximizing, find the longest string and work backwards, as long as
    the matched lengths for each iteration are the same. */

    else
      {
      BOOL samelengths = TRUE;
      pp = eptr;
      length = mb->ovector[offset+1] - mb->ovector[offset];

      for (i = min; i < max; i++)
        {
        PCRE2_SIZE slength;
        int rc = match_ref(offset, offset_top, eptr, mb, caseless, &slength);

        if (rc != 0)
          {
          /* Can't use CHECK_PARTIAL because we don't want to update eptr in
          the soft partial matching case. */

          if (rc > 0 && mb->partial != 0 &&
              mb->end_subject > mb->start_used_ptr)
            {
            mb->hitend = TRUE;
            if (mb->partial > 1) RRETURN(PCRE2_ERROR_PARTIAL);
            }
          break;
          }

        if (slength != length) samelengths = FALSE;
        eptr += slength;
        }

      /* If the length matched for each repetition is the same as the length of
      the captured group, we can easily work backwards. This is the normal
      case. However, in caseless UTF-8 mode there are pairs of case-equivalent
      characters whose lengths (in terms of code units) differ. However, this
      is very rare, so we handle it by re-matching fewer and fewer times. */

      if (samelengths)
        {
        while (eptr >= pp)
          {
          RMATCH(eptr, ecode, offset_top, mb, eptrb, RM15);
          if (rrc != MATCH_NOMATCH) RRETURN(rrc);
          eptr -= length;
          }
        }

      /* The rare case of non-matching lengths. Re-scan the repetition for each
      iteration. We know that match_ref() will succeed every time. */

      else
        {
        max = i;
        for (;;)
          {
          RMATCH(eptr, ecode, offset_top, mb, eptrb, RM68);
          if (rrc != MATCH_NOMATCH) RRETURN(rrc);
          if (eptr == pp) break;  /* Failed after minimal repetition */
          eptr = pp;
          max--;
          for (i = min; i < max; i++)
            {
            PCRE2_SIZE slength;
            (void)match_ref(offset, offset_top, eptr, mb, caseless, &slength);
            eptr += slength;
            }
          }
        }

      RRETURN(MATCH_NOMATCH);
      }
    /* Control never gets here */

    /* Match a bit-mapped character class, possibly repeatedly. This op code is
    used when all the characters in the class have values in the range 0-255,
    and either the matching is caseful, or the characters are in the range
    0-127 when UTF-8 processing is enabled. The only difference between
    OP_CLASS and OP_NCLASS occurs when a data character outside the range is
    encountered.

    First, look past the end of the item to see if there is repeat information
    following. Then obey similar code to character type repeats - written out
    again for speed. */

    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 ((uint8_t *)data)
      data = ecode + 1;                /* Save for matching */
      ecode += 1 + (32 / sizeof(PCRE2_UCHAR)); /* Advance past the item */

      switch (*ecode)
        {
        case OP_CRSTAR:
        case OP_CRMINSTAR:
        case OP_CRPLUS:
        case OP_CRMINPLUS:
        case OP_CRQUERY:
        case OP_CRMINQUERY:
        case OP_CRPOSSTAR:
        case OP_CRPOSPLUS:
        case OP_CRPOSQUERY:
        c = *ecode++ - OP_CRSTAR;
        if (c < OP_CRPOSSTAR - OP_CRSTAR) minimize = (c & 1) != 0;
        else possessive = TRUE;
        min = rep_min[c];                 /* Pick up values from tables; */
        max = rep_max[c];                 /* zero for max => infinity */
        if (max == 0) max = INT_MAX;
        break;

        case OP_CRRANGE:
        case OP_CRMINRANGE:
        case OP_CRPOSRANGE:
        minimize = (*ecode == OP_CRMINRANGE);
        possessive = (*ecode == OP_CRPOSRANGE);
        min = GET2(ecode, 1);
        max = GET2(ecode, 1 + IMM2_SIZE);
        if (max == 0) max = INT_MAX;
        ecode += 1 + 2 * IMM2_SIZE;
        break;

        default:               /* No repeat follows */
        min = max = 1;
        break;
        }

      /* First, ensure the minimum number of matches are present. */

#ifdef SUPPORT_UNICODE
      if (utf)
        {
        for (i = 1; i <= min; i++)
          {
          if (eptr >= mb->end_subject)
            {
            SCHECK_PARTIAL();
            RRETURN(MATCH_NOMATCH);
            }
          GETCHARINC(c, eptr);
          if (c > 255)
            {
            if (op == OP_CLASS) RRETURN(MATCH_NOMATCH);
            }
          else
            if ((BYTE_MAP[c/8] & (1 << (c&7))) == 0) RRETURN(MATCH_NOMATCH);
          }
        }
      else
#endif
      /* Not UTF mode */
        {
        for (i = 1; i <= min; i++)
          {
          if (eptr >= mb->end_subject)
            {
            SCHECK_PARTIAL();
            RRETURN(MATCH_NOMATCH);
            }
          c = *eptr++;
#if PCRE2_CODE_UNIT_WIDTH != 8
          if (c > 255)
            {
            if (op == OP_CLASS) RRETURN(MATCH_NOMATCH);
            }
          else
#endif
            if ((BYTE_MAP[c/8] & (1 << (c&7))) == 0) RRETURN(MATCH_NOMATCH);
          }
        }

      /* If max == min we can continue with the main loop without the
      need to recurse. */

      if (min == max) continue;

      /* If minimizing, keep testing the rest of the expression and advancing
      the pointer while it matches the class. */

      if (minimize)
        {
#ifdef SUPPORT_UNICODE
        if (utf)
          {
          for (fi = min;; fi++)
            {
            RMATCH(eptr, ecode, offset_top, mb, eptrb, RM16);
            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
            if (fi >= max) RRETURN(MATCH_NOMATCH);
            if (eptr >= mb->end_subject)
              {
              SCHECK_PARTIAL();
              RRETURN(MATCH_NOMATCH);
              }
            GETCHARINC(c, eptr);
            if (c > 255)
              {
              if (op == OP_CLASS) RRETURN(MATCH_NOMATCH);
              }
            else
              if ((BYTE_MAP[c/8] & (1 << (c&7))) == 0) RRETURN(MATCH_NOMATCH);
            }
          }
        else
#endif
        /* Not UTF mode */
          {
          for (fi = min;; fi++)
            {
            RMATCH(eptr, ecode, offset_top, mb, eptrb, RM17);
            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
            if (fi >= max) RRETURN(MATCH_NOMATCH);
            if (eptr >= mb->end_subject)
              {
              SCHECK_PARTIAL();
              RRETURN(MATCH_NOMATCH);
              }
            c = *eptr++;
#if PCRE2_CODE_UNIT_WIDTH != 8
            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 */
        }

      /* If maximizing, find the longest possible run, then work backwards. */

      else
        {
        pp = eptr;

#ifdef SUPPORT_UNICODE
        if (utf)
          {
          for (i = min; i < max; i++)
            {
            int len = 1;
            if (eptr >= mb->end_subject)
              {
              SCHECK_PARTIAL();
              break;
              }
            GETCHARLEN(c, eptr, len);
            if (c > 255)
              {
              if (op == OP_CLASS) break;
              }
            else
              if ((BYTE_MAP[c/8] & (1 << (c&7))) == 0) break;
            eptr += len;
            }

          if (possessive) continue;    /* No backtracking */

          for (;;)
            {
            RMATCH(eptr, ecode, offset_top, mb, eptrb, RM18);
            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
            if (eptr-- == pp) break;        /* Stop if tried at original pos */
            BACKCHAR(eptr);
            }
          }
        else
#endif
          /* Not UTF mode */
          {
          for (i = min; i < max; i++)
            {
            if (eptr >= mb->end_subject)
              {
              SCHECK_PARTIAL();
              break;
              }
            c = *eptr;
#if PCRE2_CODE_UNIT_WIDTH != 8
            if (c > 255)
              {
              if (op == OP_CLASS) break;
              }
            else
#endif
              if ((BYTE_MAP[c/8] & (1 << (c&7))) == 0) break;
            eptr++;
            }

          if (possessive) continue;    /* No backtracking */

          while (eptr >= pp)
            {
            RMATCH(eptr, ecode, offset_top, mb, eptrb, RM19);
            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
            eptr--;
            }
          }

        RRETURN(MATCH_NOMATCH);
        }
#undef BYTE_MAP
      }
    /* Control never gets here */


    /* Match an extended character class. In the 8-bit library, this opcode is
    encountered only when UTF-8 mode mode is supported. In the 16-bit and
    32-bit libraries, codepoints greater than 255 may be encountered even when
    UTF is not supported. */

#ifdef SUPPORT_WIDE_CHARS
    case OP_XCLASS:
      {
      data = ecode + 1 + LINK_SIZE;                /* Save for matching */
      ecode += GET(ecode, 1);                      /* Advance past the item */

      switch (*ecode)
        {
        case OP_CRSTAR:
        case OP_CRMINSTAR:
        case OP_CRPLUS:
        case OP_CRMINPLUS:
        case OP_CRQUERY:
        case OP_CRMINQUERY:
        case OP_CRPOSSTAR:
        case OP_CRPOSPLUS:
        case OP_CRPOSQUERY:
        c = *ecode++ - OP_CRSTAR;
        if (c < OP_CRPOSSTAR - OP_CRSTAR) minimize = (c & 1) != 0;
        else possessive = TRUE;
        min = rep_min[c];                 /* Pick up values from tables; */
        max = rep_max[c];                 /* zero for max => infinity */
        if (max == 0) max = INT_MAX;
        break;

        case OP_CRRANGE:
        case OP_CRMINRANGE:
        case OP_CRPOSRANGE:
        minimize = (*ecode == OP_CRMINRANGE);
        possessive = (*ecode == OP_CRPOSRANGE);
        min = GET2(ecode, 1);
        max = GET2(ecode, 1 + IMM2_SIZE);
        if (max == 0) max = INT_MAX;
        ecode += 1 + 2 * IMM2_SIZE;
        break;

        default:               /* No repeat follows */
        min = max = 1;
        break;
        }

      /* First, ensure the minimum number of matches are present. */

      for (i = 1; i <= min; i++)
        {
        if (eptr >= mb->end_subject)
          {
          SCHECK_PARTIAL();
          RRETURN(MATCH_NOMATCH);
          }
        GETCHARINCTEST(c, eptr);
        if (!PRIV(xclass)(c, data, utf)) RRETURN(MATCH_NOMATCH);
        }

      /* If max == min we can continue with the main loop without the
      need to recurse. */

      if (min == max) continue;

      /* If minimizing, keep testing the rest of the expression and advancing
      the pointer while it matches the class. */

      if (minimize)
        {
        for (fi = min;; fi++)
          {
          RMATCH(eptr, ecode, offset_top, mb, eptrb, RM20);
          if (rrc != MATCH_NOMATCH) RRETURN(rrc);
          if (fi >= max) RRETURN(MATCH_NOMATCH);
          if (eptr >= mb->end_subject)
            {
            SCHECK_PARTIAL();
            RRETURN(MATCH_NOMATCH);
            }
          GETCHARINCTEST(c, eptr);
          if (!PRIV(xclass)(c, data, utf)) RRETURN(MATCH_NOMATCH);
          }
        /* Control never gets here */
        }

      /* If maximizing, find the longest possible run, then work backwards. */

      else
        {
        pp = eptr;
        for (i = min; i < max; i++)
          {
          int len = 1;
          if (eptr >= mb->end_subject)
            {
            SCHECK_PARTIAL();
            break;
            }
#ifdef SUPPORT_UNICODE
          GETCHARLENTEST(c, eptr, len);
#else
          c = *eptr;
#endif
          if (!PRIV(xclass)(c, data, utf)) break;
          eptr += len;
          }

        if (possessive) continue;    /* No backtracking */

        for(;;)
          {
          RMATCH(eptr, ecode, offset_top, mb, eptrb, RM21);
          if (rrc != MATCH_NOMATCH) RRETURN(rrc);
          if (eptr-- == pp) break;        /* Stop if tried at original pos */
#ifdef SUPPORT_UNICODE
          if (utf) BACKCHAR(eptr);
#endif
          }
        RRETURN(MATCH_NOMATCH);
        }

      /* Control never gets here */
      }
#endif    /* End of XCLASS */

    /* Match a single character, casefully */

    case OP_CHAR:
#ifdef SUPPORT_UNICODE
    if (utf)
      {
      length = 1;
      ecode++;
      GETCHARLEN(fc, ecode, length);
      if (length > (PCRE2_SIZE)(mb->end_subject - eptr))
        {
        CHECK_PARTIAL();             /* Not SCHECK_PARTIAL() */
        RRETURN(MATCH_NOMATCH);
        }
      for (; length > 0; length--)
        {
        if (*ecode++ != UCHAR21INC(eptr)) RRETURN(MATCH_NOMATCH);
        }
      }
    else
#endif
    /* Not UTF mode */
      {
      if (mb->end_subject - eptr < 1)
        {
        SCHECK_PARTIAL();            /* This one can use SCHECK_PARTIAL() */
        RRETURN(MATCH_NOMATCH);
        }
      if (ecode[1] != *eptr++) RRETURN(MATCH_NOMATCH);
      ecode += 2;
      }
    break;

    /* Match a single character, caselessly. If we are at the end of the
    subject, give up immediately. */

    case OP_CHARI:
    if (eptr >= mb->end_subject)
      {
      SCHECK_PARTIAL();
      RRETURN(MATCH_NOMATCH);
      }

#ifdef SUPPORT_UNICODE
    if (utf)
      {
      length = 1;
      ecode++;
      GETCHARLEN(fc, ecode, length);

      /* If the pattern character's value is < 128, we have only one byte, and
      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)
        {
        uint32_t cc = UCHAR21(eptr);
        if (mb->lcc[fc] != TABLE_GET(cc, mb->lcc, cc)) RRETURN(MATCH_NOMATCH);
        ecode++;
        eptr++;
        }

      /* 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
        {
        uint32_t dc;
        GETCHARINC(dc, eptr);
        ecode += length;

        /* If we have Unicode property support, we can use it to test the other
        case of the character, if there is one. */

        if (fc != dc)
          {
#ifdef SUPPORT_UNICODE
          if (dc != UCD_OTHERCASE(fc))
#endif
            RRETURN(MATCH_NOMATCH);
          }
        }
      }
    else
#endif   /* SUPPORT_UNICODE */

    /* Not UTF mode */
      {
      if (TABLE_GET(ecode[1], mb->lcc, ecode[1])
          != TABLE_GET(*eptr, mb->lcc, *eptr)) RRETURN(MATCH_NOMATCH);
      eptr++;
      ecode += 2;
      }
    break;

    /* Match a single character repeatedly. */

    case OP_EXACT:
    case OP_EXACTI:
    min = max = GET2(ecode, 1);
    ecode += 1 + IMM2_SIZE;
    goto REPEATCHAR;

    case OP_POSUPTO:
    case OP_POSUPTOI:
    possessive = TRUE;
    /* Fall through */

    case OP_UPTO:
    case OP_UPTOI:
    case OP_MINUPTO:
    case OP_MINUPTOI:
    min = 0;
    max = GET2(ecode, 1);
    minimize = *ecode == OP_MINUPTO || *ecode == OP_MINUPTOI;
    ecode += 1 + IMM2_SIZE;
    goto REPEATCHAR;

    case OP_POSSTAR:
    case OP_POSSTARI:
    possessive = TRUE;
    min = 0;
    max = INT_MAX;
    ecode++;
    goto REPEATCHAR;

    case OP_POSPLUS:
    case OP_POSPLUSI:
    possessive = TRUE;
    min = 1;
    max = INT_MAX;
    ecode++;
    goto REPEATCHAR;

    case OP_POSQUERY:
    case OP_POSQUERYI:
    possessive = TRUE;
    min = 0;
    max = 1;
    ecode++;
    goto REPEATCHAR;

    case OP_STAR:
    case OP_STARI:
    case OP_MINSTAR:
    case OP_MINSTARI:
    case OP_PLUS:
    case OP_PLUSI:
    case OP_MINPLUS:
    case OP_MINPLUSI:
    case OP_QUERY:
    case OP_QUERYI:
    case OP_MINQUERY:
    case OP_MINQUERYI:
    c = *ecode++ - ((op < OP_STARI)? OP_STAR : OP_STARI);
    minimize = (c & 1) != 0;
    min = rep_min[c];                 /* Pick up values from tables; */
    max = rep_max[c];                 /* zero for max => infinity */
    if (max == 0) max = INT_MAX;

    /* Common code for all repeated single-character matches. We first check
    for the minimum number of characters. If the minimum equals the maximum, we
    are done. Otherwise, if minimizing, check the rest of the pattern for a
    match; if there isn't one, advance up to the maximum, one character at a
    time.

    If maximizing, advance up to the maximum number of matching characters,
    until eptr is past the end of the maximum run. If possessive, we are
    then done (no backing up). Otherwise, match at this position; anything
    other than no match is immediately returned. For nomatch, back up one
    character, unless we are matching \R and the last thing matched was
    \r\n, in which case, back up two bytes. When we reach the first optional
    character position, we can save stack by doing a tail recurse.

    The various UTF/non-UTF and caseful/caseless cases are handled separately,
    for speed. */

    REPEATCHAR:
#ifdef SUPPORT_UNICODE
    if (utf)
      {
      length = 1;
      charptr = ecode;
      GETCHARLEN(fc, ecode, length);
      ecode += length;

      /* Handle multibyte character matching specially here. There is
      support for caseless matching if UCP support is present. */

      if (length > 1)
        {
        uint32_t othercase;
        if (op >= OP_STARI &&     /* Caseless */
            (othercase = UCD_OTHERCASE(fc)) != fc)
          oclength = PRIV(ord2utf)(othercase, occhars);
        else oclength = 0;

        for (i = 1; i <= min; i++)
          {
          if (eptr <= mb->end_subject - length &&
            memcmp(eptr, charptr, CU2BYTES(length)) == 0) eptr += length;
          else if (oclength > 0 &&
                   eptr <= mb->end_subject - oclength &&
                   memcmp(eptr, occhars, CU2BYTES(oclength)) == 0) eptr += oclength;
          else
            {
            CHECK_PARTIAL();
            RRETURN(MATCH_NOMATCH);
            }
          }

        if (min == max) continue;

        if (minimize)
          {
          for (fi = min;; fi++)
            {
            RMATCH(eptr, ecode, offset_top, mb, eptrb, RM22);
            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
            if (fi >= max) RRETURN(MATCH_NOMATCH);
            if (eptr <= mb->end_subject - length &&
              memcmp(eptr, charptr, CU2BYTES(length)) == 0) eptr += length;
            else if (oclength > 0 &&
                     eptr <= mb->end_subject - oclength &&
                     memcmp(eptr, occhars, CU2BYTES(oclength)) == 0) eptr += oclength;
            else
              {
              CHECK_PARTIAL();
              RRETURN(MATCH_NOMATCH);
              }
            }
          /* Control never gets here */
          }

        else  /* Maximize */
          {
          pp = eptr;
          for (i = min; i < max; i++)
            {
            if (eptr <= mb->end_subject - length &&
                memcmp(eptr, charptr, CU2BYTES(length)) == 0) eptr += length;
            else if (oclength > 0 &&
                     eptr <= mb->end_subject - oclength &&
                     memcmp(eptr, occhars, CU2BYTES(oclength)) == 0) eptr += oclength;
            else
              {
              CHECK_PARTIAL();
              break;
              }
            }

          if (possessive) continue;    /* No backtracking */

          /* After \C in UTF mode, pp might be in the middle of a Unicode
          character. Use <= pp to ensure backtracking doesn't go too far. */

          for(;;)
            {
            if (eptr <= pp) goto TAIL_RECURSE;
            RMATCH(eptr, ecode, offset_top, mb, eptrb, RM23);
            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
            eptr--;
            BACKCHAR(eptr);
            }
          }
        /* Control never gets here */
        }

      /* If the length of a UTF-8 character is 1, we fall through here, and
      obey the code as for non-UTF-8 characters below, though in this case the
      value of fc will always be < 128. */
      }
    else
#endif  /* SUPPORT_UNICODE */

      /* When not in UTF-8 mode, load a single-byte character. */
      fc = *ecode++;

    /* 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
    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
    minimizing, keep trying the rest of the expression and advancing one
    matching character if failing, up to the maximum. Alternatively, if
    maximizing, find the maximum number of characters and work backwards. */

    if (op >= OP_STARI)  /* Caseless */
      {
#if PCRE2_CODE_UNIT_WIDTH == 8
      /* fc must be < 128 if UTF is enabled. */
      foc = mb->fcc[fc];
#else
#ifdef SUPPORT_UNICODE
      if (utf && fc > 127)
        foc = UCD_OTHERCASE(fc);
      else
#endif /* SUPPORT_UNICODE */
        foc = TABLE_GET(fc, mb->fcc, fc);
#endif /* PCRE2_CODE_UNIT_WIDTH == 8 */

      for (i = 1; i <= min; i++)
        {
        uint32_t cc;                 /* Faster than PCRE2_UCHAR */
        if (eptr >= mb->end_subject)
          {
          SCHECK_PARTIAL();
          RRETURN(MATCH_NOMATCH);
          }
        cc = UCHAR21TEST(eptr);
        if (fc != cc && foc != cc) RRETURN(MATCH_NOMATCH);
        eptr++;
        }
      if (min == max) continue;
      if (minimize)
        {
        for (fi = min;; fi++)
          {
          uint32_t cc;               /* Faster than PCRE2_UCHAR */
          RMATCH(eptr, ecode, offset_top, mb, eptrb, RM24);
          if (rrc != MATCH_NOMATCH) RRETURN(rrc);
          if (fi >= max) RRETURN(MATCH_NOMATCH);
          if (eptr >= mb->end_subject)
            {
            SCHECK_PARTIAL();
            RRETURN(MATCH_NOMATCH);
            }
          cc = UCHAR21TEST(eptr);
          if (fc != cc && foc != cc) RRETURN(MATCH_NOMATCH);
          eptr++;
          }
        /* Control never gets here */
        }
      else  /* Maximize */
        {
        pp = eptr;
        for (i = min; i < max; i++)
          {
          uint32_t cc;               /* Faster than PCRE2_UCHAR */
          if (eptr >= mb->end_subject)
            {
            SCHECK_PARTIAL();
            break;
            }
          cc = UCHAR21TEST(eptr);
          if (fc != cc && foc != cc) break;
          eptr++;
          }
        if (possessive) continue;       /* No backtracking */
        for (;;)
          {
          if (eptr == pp) goto TAIL_RECURSE;
          RMATCH(eptr, ecode, offset_top, mb, eptrb, RM25);
          eptr--;
          if (rrc != MATCH_NOMATCH) RRETURN(rrc);
          }
        /* Control never gets here */
        }
      }

    /* Caseful comparisons (includes all multi-byte characters) */

    else
      {
      for (i = 1; i <= min; i++)
        {
        if (eptr >= mb->end_subject)
          {
          SCHECK_PARTIAL();
          RRETURN(MATCH_NOMATCH);
          }
        if (fc != UCHAR21INCTEST(eptr)) RRETURN(MATCH_NOMATCH);
        }

      if (min == max) continue;

      if (minimize)
        {
        for (fi = min;; fi++)
          {
          RMATCH(eptr, ecode, offset_top, mb, eptrb, RM26);
          if (rrc != MATCH_NOMATCH) RRETURN(rrc);
          if (fi >= max) RRETURN(MATCH_NOMATCH);
          if (eptr >= mb->end_subject)
            {
            SCHECK_PARTIAL();
            RRETURN(MATCH_NOMATCH);
            }
          if (fc != UCHAR21INCTEST(eptr)) RRETURN(MATCH_NOMATCH);
          }
        /* Control never gets here */
        }
      else  /* Maximize */
        {
        pp = eptr;
        for (i = min; i < max; i++)
          {
          if (eptr >= mb->end_subject)
            {
            SCHECK_PARTIAL();
            break;
            }
          if (fc != UCHAR21TEST(eptr)) break;
          eptr++;
          }
        if (possessive) continue;    /* No backtracking */
        for (;;)
          {
          if (eptr == pp) goto TAIL_RECURSE;
          RMATCH(eptr, ecode, offset_top, mb, eptrb, RM27);
          eptr--;
          if (rrc != MATCH_NOMATCH) RRETURN(rrc);
          }
        /* Control never gets here */
        }
      }
    /* Control never gets here */

    /* Match a negated single one-byte character. The character we are
    checking can be multibyte. */

    case OP_NOT:
    case OP_NOTI:
    if (eptr >= mb->end_subject)
      {
      SCHECK_PARTIAL();
      RRETURN(MATCH_NOMATCH);
      }
#ifdef SUPPORT_UNICODE
    if (utf)
      {
      uint32_t ch, och;

      ecode++;
      GETCHARINC(ch, ecode);
      GETCHARINC(c, eptr);

      if (op == OP_NOT)
        {
        if (ch == c) RRETURN(MATCH_NOMATCH);
        }
      else
        {
        if (ch > 127)
          och = UCD_OTHERCASE(ch);
        else
          och = TABLE_GET(ch, mb->fcc, ch);
        if (ch == c || och == c) RRETURN(MATCH_NOMATCH);
        }
      }
    else
#endif  /* SUPPORT_UNICODE */
      {
      uint32_t ch = ecode[1];
      c = *eptr++;
      if (ch == c || (op == OP_NOTI && TABLE_GET(ch, mb->fcc, ch) == c))
        RRETURN(MATCH_NOMATCH);
      ecode += 2;
      }
    break;

    /* Match a negated single one-byte character repeatedly. This is almost a
    repeat of the code for a repeated single character, but I haven't found a
    nice way of commoning these up that doesn't require a test of the
    positive/negative option for each character match. Maybe that wouldn't add
    very much to the time taken, but character matching *is* what this is all
    about... */

    case OP_NOTEXACT:
    case OP_NOTEXACTI:
    min = max = GET2(ecode, 1);
    ecode += 1 + IMM2_SIZE;
    goto REPEATNOTCHAR;

    case OP_NOTUPTO:
    case OP_NOTUPTOI:
    case OP_NOTMINUPTO:
    case OP_NOTMINUPTOI:
    min = 0;
    max = GET2(ecode, 1);
    minimize = *ecode == OP_NOTMINUPTO || *ecode == OP_NOTMINUPTOI;
    ecode += 1 + IMM2_SIZE;
    goto REPEATNOTCHAR;

    case OP_NOTPOSSTAR:
    case OP_NOTPOSSTARI:
    possessive = TRUE;
    min = 0;
    max = INT_MAX;
    ecode++;
    goto REPEATNOTCHAR;

    case OP_NOTPOSPLUS:
    case OP_NOTPOSPLUSI:
    possessive = TRUE;
    min = 1;
    max = INT_MAX;
    ecode++;
    goto REPEATNOTCHAR;

    case OP_NOTPOSQUERY:
    case OP_NOTPOSQUERYI:
    possessive = TRUE;
    min = 0;
    max = 1;
    ecode++;
    goto REPEATNOTCHAR;

    case OP_NOTPOSUPTO:
    case OP_NOTPOSUPTOI:
    possessive = TRUE;
    min = 0;
    max = GET2(ecode, 1);
    ecode += 1 + IMM2_SIZE;
    goto REPEATNOTCHAR;

    case OP_NOTSTAR:
    case OP_NOTSTARI:
    case OP_NOTMINSTAR:
    case OP_NOTMINSTARI:
    case OP_NOTPLUS:
    case OP_NOTPLUSI:
    case OP_NOTMINPLUS:
    case OP_NOTMINPLUSI:
    case OP_NOTQUERY:
    case OP_NOTQUERYI:
    case OP_NOTMINQUERY:
    case OP_NOTMINQUERYI:
    c = *ecode++ - ((op >= OP_NOTSTARI)? OP_NOTSTARI: OP_NOTSTAR);
    minimize = (c & 1) != 0;
    min = rep_min[c];                 /* Pick up values from tables; */
    max = rep_max[c];                 /* zero for max => infinity */
    if (max == 0) max = INT_MAX;

    /* Common code for all repeated single-byte matches. */

    REPEATNOTCHAR:
    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
    minimum number of matches are present. If min = max, continue at the same
    level without recursing. Otherwise, if minimizing, keep trying the rest of
    the expression and advancing one matching character if failing, up to the
    maximum. Alternatively, if maximizing, find the maximum number of
    characters and work backwards. */

    if (op >= OP_NOTSTARI)     /* Caseless */
      {
#ifdef SUPPORT_UNICODE
      if (utf && fc > 127)
        foc = UCD_OTHERCASE(fc);
      else
#endif /* SUPPORT_UNICODE */
        foc = TABLE_GET(fc, mb->fcc, fc);

#ifdef SUPPORT_UNICODE
      if (utf)
        {
        uint32_t d;
        for (i = 1; i <= min; i++)
          {
          if (eptr >= mb->end_subject)
            {
            SCHECK_PARTIAL();
            RRETURN(MATCH_NOMATCH);
            }
          GETCHARINC(d, eptr);
          if (fc == d || (uint32_t)foc == d) RRETURN(MATCH_NOMATCH);
          }
        }
      else
#endif  /* SUPPORT_UNICODE */
      /* Not UTF mode */
        {
        for (i = 1; i <= min; i++)
          {
          if (eptr >= mb->end_subject)
            {
            SCHECK_PARTIAL();
            RRETURN(MATCH_NOMATCH);
            }
          if (fc == *eptr || foc == *eptr) RRETURN(MATCH_NOMATCH);
          eptr++;
          }
        }

      if (min == max) continue;

      if (minimize)
        {
#ifdef SUPPORT_UNICODE
        if (utf)
          {
          uint32_t d;
          for (fi = min;; fi++)
            {
            RMATCH(eptr, ecode, offset_top, mb, eptrb, RM28);
            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
            if (fi >= max) RRETURN(MATCH_NOMATCH);
            if (eptr >= mb->end_subject)
              {
              SCHECK_PARTIAL();
              RRETURN(MATCH_NOMATCH);
              }
            GETCHARINC(d, eptr);
            if (fc == d || (uint32_t)foc == d) RRETURN(MATCH_NOMATCH);
            }
          }
        else
#endif  /*SUPPORT_UNICODE */
        /* Not UTF mode */
          {
          for (fi = min;; fi++)
            {
            RMATCH(eptr, ecode, offset_top, mb, eptrb, RM29);
            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
            if (fi >= max) RRETURN(MATCH_NOMATCH);
            if (eptr >= mb->end_subject)
              {
              SCHECK_PARTIAL();
              RRETURN(MATCH_NOMATCH);
              }
            if (fc == *eptr || foc == *eptr) RRETURN(MATCH_NOMATCH);
            eptr++;
            }
          }
        /* Control never gets here */
        }

      /* Maximize case */

      else
        {
        pp = eptr;

#ifdef SUPPORT_UNICODE
        if (utf)
          {
          uint32_t d;
          for (i = min; i < max; i++)
            {
            int len = 1;
            if (eptr >= mb->end_subject)
              {
              SCHECK_PARTIAL();
              break;
              }
            GETCHARLEN(d, eptr, len);
            if (fc == d || (uint32_t)foc == d) break;
            eptr += len;
            }
          if (possessive) continue;    /* No backtracking */

          /* After \C in UTF mode, pp might be in the middle of a Unicode
          character. Use <= pp to ensure backtracking doesn't go too far. */

          for(;;)
            {
            if (eptr <= pp) goto TAIL_RECURSE;
            RMATCH(eptr, ecode, offset_top, mb, eptrb, RM30);
            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
            eptr--;
            BACKCHAR(eptr);
            }
          }
        else
#endif  /* SUPPORT_UNICODE */
        /* Not UTF mode */
          {
          for (i = min; i < max; i++)
            {
            if (eptr >= mb->end_subject)
              {
              SCHECK_PARTIAL();
              break;
              }
            if (fc == *eptr || foc == *eptr) break;
            eptr++;
            }
          if (possessive) continue;    /* No backtracking */
          for (;;)
            {
            if (eptr == pp) goto TAIL_RECURSE;
            RMATCH(eptr, ecode, offset_top, mb, eptrb, RM31);
            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
            eptr--;
            }
          }
        /* Control never gets here */
        }
      }

    /* Caseful comparisons */

    else
      {
#ifdef SUPPORT_UNICODE
      if (utf)
        {
        uint32_t d;
        for (i = 1; i <= min; i++)
          {
          if (eptr >= mb->end_subject)
            {
            SCHECK_PARTIAL();
            RRETURN(MATCH_NOMATCH);
            }
          GETCHARINC(d, eptr);
          if (fc == d) RRETURN(MATCH_NOMATCH);
          }
        }
      else
#endif
      /* Not UTF mode */
        {
        for (i = 1; i <= min; i++)
          {
          if (eptr >= mb->end_subject)
            {
            SCHECK_PARTIAL();
            RRETURN(MATCH_NOMATCH);
            }
          if (fc == *eptr++) RRETURN(MATCH_NOMATCH);
          }
        }

      if (min == max) continue;

      if (minimize)
        {
#ifdef SUPPORT_UNICODE
        if (utf)
          {
          uint32_t d;
          for (fi = min;; fi++)
            {
            RMATCH(eptr, ecode, offset_top, mb, eptrb, RM32);
            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
            if (fi >= max) RRETURN(MATCH_NOMATCH);
            if (eptr >= mb->end_subject)
              {
              SCHECK_PARTIAL();
              RRETURN(MATCH_NOMATCH);
              }
            GETCHARINC(d, eptr);
            if (fc == d) RRETURN(MATCH_NOMATCH);
            }
          }
        else
#endif
        /* Not UTF mode */
          {
          for (fi = min;; fi++)
            {
            RMATCH(eptr, ecode, offset_top, mb, eptrb, RM33);
            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
            if (fi >= max) RRETURN(MATCH_NOMATCH);
            if (eptr >= mb->end_subject)
              {
              SCHECK_PARTIAL();
              RRETURN(MATCH_NOMATCH);
              }
            if (fc == *eptr++) RRETURN(MATCH_NOMATCH);
            }
          }
        /* Control never gets here */
        }

      /* Maximize case */

      else
        {
        pp = eptr;

#ifdef SUPPORT_UNICODE
        if (utf)
          {
          uint32_t d;
          for (i = min; i < max; i++)
            {
            int len = 1;
            if (eptr >= mb->end_subject)
              {
              SCHECK_PARTIAL();
              break;
              }
            GETCHARLEN(d, eptr, len);
            if (fc == d) break;
            eptr += len;
            }
          if (possessive) continue;    /* No backtracking */

          /* After \C in UTF mode, pp might be in the middle of a Unicode
          character. Use <= pp to ensure backtracking doesn't go too far. */

          for(;;)
            {
            if (eptr <= pp) goto TAIL_RECURSE;
            RMATCH(eptr, ecode, offset_top, mb, eptrb, RM34);
            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
            eptr--;
            BACKCHAR(eptr);
            }
          }
        else
#endif
        /* Not UTF mode */
          {
          for (i = min; i < max; i++)
            {
            if (eptr >= mb->end_subject)
              {
              SCHECK_PARTIAL();
              break;
              }
            if (fc == *eptr) break;
            eptr++;
            }
          if (possessive) continue;    /* No backtracking */
          for (;;)
            {
            if (eptr == pp) goto TAIL_RECURSE;
            RMATCH(eptr, ecode, offset_top, mb, eptrb, RM35);
            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
            eptr--;
            }
          }
        /* Control never gets here */
        }
      }
    /* Control never gets here */

    /* Match a single character type repeatedly; several different opcodes
    share code. This is very similar to the code for single characters, but we
    repeat it in the interests of efficiency. */

    case OP_TYPEEXACT:
    min = max = GET2(ecode, 1);
    minimize = TRUE;
    ecode += 1 + IMM2_SIZE;
    goto REPEATTYPE;

    case OP_TYPEUPTO:
    case OP_TYPEMINUPTO:
    min = 0;
    max = GET2(ecode, 1);
    minimize = *ecode == OP_TYPEMINUPTO;
    ecode += 1 + IMM2_SIZE;
    goto REPEATTYPE;

    case OP_TYPEPOSSTAR:
    possessive = TRUE;
    min = 0;
    max = INT_MAX;
    ecode++;
    goto REPEATTYPE;

    case OP_TYPEPOSPLUS:
    possessive = TRUE;
    min = 1;
    max = INT_MAX;
    ecode++;
    goto REPEATTYPE;

    case OP_TYPEPOSQUERY:
    possessive = TRUE;
    min = 0;
    max = 1;
    ecode++;
    goto REPEATTYPE;

    case OP_TYPEPOSUPTO:
    possessive = TRUE;
    min = 0;
    max = GET2(ecode, 1);
    ecode += 1 + IMM2_SIZE;
    goto REPEATTYPE;

    case OP_TYPESTAR:
    case OP_TYPEMINSTAR:
    case OP_TYPEPLUS:
    case OP_TYPEMINPLUS:
    case OP_TYPEQUERY:
    case OP_TYPEMINQUERY:
    c = *ecode++ - OP_TYPESTAR;
    minimize = (c & 1) != 0;
    min = rep_min[c];                 /* Pick up values from tables; */
    max = rep_max[c];                 /* zero for max => infinity */
    if (max == 0) max = INT_MAX;

    /* Common code for all repeated single character type matches. Note that
    in UTF-8 mode, '.' matches a character of any length, but for the other
    character types, the valid characters are all one-byte long. */

    REPEATTYPE:
    ctype = *ecode++;      /* Code for the character type */

#ifdef SUPPORT_UNICODE
    if (ctype == OP_PROP || ctype == OP_NOTPROP)
      {
      prop_fail_result = ctype == OP_NOTPROP;
      prop_type = *ecode++;
      prop_value = *ecode++;
      }
    else prop_type = -1;
#endif

    /* First, ensure the minimum number of matches are present. Use inline
    code for maximizing the speed, and do the type test once at the start
    (i.e. keep it out of the loop). Separate the UTF-8 code completely as that
    is tidier. Also separate the UCP code, which can be the same for both UTF-8
    and single-bytes. */

    if (min > 0)
      {
#ifdef SUPPORT_UNICODE
      if (prop_type >= 0)
        {
        switch(prop_type)
          {
          case PT_ANY:
          if (prop_fail_result) RRETURN(MATCH_NOMATCH);
          for (i = 1; i <= min; i++)
            {
            if (eptr >= mb->end_subject)
              {
              SCHECK_PARTIAL();
              RRETURN(MATCH_NOMATCH);
              }
            GETCHARINCTEST(c, eptr);
            }
          break;

          case PT_LAMP:
          for (i = 1; i <= min; i++)
            {
            int chartype;
            if (eptr >= mb->end_subject)
              {
              SCHECK_PARTIAL();
              RRETURN(MATCH_NOMATCH);
              }
            GETCHARINCTEST(c, eptr);
            chartype = UCD_CHARTYPE(c);
            if ((chartype == ucp_Lu ||
                 chartype == ucp_Ll ||
                 chartype == ucp_Lt) == prop_fail_result)
              RRETURN(MATCH_NOMATCH);
            }
          break;

          case PT_GC:
          for (i = 1; i <= min; i++)
            {
            if (eptr >= mb->end_subject)
              {
              SCHECK_PARTIAL();
              RRETURN(MATCH_NOMATCH);
              }
            GETCHARINCTEST(c, eptr);
            if ((UCD_CATEGORY(c) == prop_value) == prop_fail_result)
              RRETURN(MATCH_NOMATCH);
            }
          break;

          case PT_PC:
          for (i = 1; i <= min; i++)
            {
            if (eptr >= mb->end_subject)
              {
              SCHECK_PARTIAL();
              RRETURN(MATCH_NOMATCH);
              }
            GETCHARINCTEST(c, eptr);
            if ((UCD_CHARTYPE(c) == prop_value) == prop_fail_result)
              RRETURN(MATCH_NOMATCH);
            }
          break;

          case PT_SC:
          for (i = 1; i <= min; i++)
            {
            if (eptr >= mb->end_subject)
              {
              SCHECK_PARTIAL();
              RRETURN(MATCH_NOMATCH);
              }
            GETCHARINCTEST(c, eptr);
            if ((UCD_SCRIPT(c) == prop_value) == prop_fail_result)
              RRETURN(MATCH_NOMATCH);
            }
          break;

          case PT_ALNUM:
          for (i = 1; i <= min; i++)
            {
            int category;
            if (eptr >= mb->end_subject)
              {
              SCHECK_PARTIAL();
              RRETURN(MATCH_NOMATCH);
              }
            GETCHARINCTEST(c, eptr);
            category = UCD_CATEGORY(c);
            if ((category == ucp_L || category == ucp_N) == prop_fail_result)
              RRETURN(MATCH_NOMATCH);
            }
          break;

          /* Perl space used to exclude VT, but from Perl 5.18 it is included,
          which means that Perl space and POSIX space are now identical. PCRE
          was changed at release 8.34. */

          case PT_SPACE:    /* Perl space */
          case PT_PXSPACE:  /* POSIX space */
          for (i = 1; i <= min; i++)
            {
            if (eptr >= mb->end_subject)
              {
              SCHECK_PARTIAL();
              RRETURN(MATCH_NOMATCH);
              }
            GETCHARINCTEST(c, eptr);
            switch(c)
              {
              HSPACE_CASES:
              VSPACE_CASES:
              if (prop_fail_result) RRETURN(MATCH_NOMATCH);
              break;

              default:
              if ((UCD_CATEGORY(c) == ucp_Z) == prop_fail_result)
                RRETURN(MATCH_NOMATCH);
              break;
              }
            }
          break;

          case PT_WORD:
          for (i = 1; i <= min; i++)
            {
            int category;
            if (eptr >= mb->end_subject)
              {
              SCHECK_PARTIAL();
              RRETURN(MATCH_NOMATCH);
              }
            GETCHARINCTEST(c, eptr);
            category = UCD_CATEGORY(c);
            if ((category == ucp_L || category == ucp_N || c == CHAR_UNDERSCORE)
                   == prop_fail_result)
              RRETURN(MATCH_NOMATCH);
            }
          break;

          case PT_CLIST:
          for (i = 1; i <= min; i++)
            {
            const uint32_t *cp;
            if (eptr >= mb->end_subject)
              {
              SCHECK_PARTIAL();
              RRETURN(MATCH_NOMATCH);
              }
            GETCHARINCTEST(c, eptr);
            cp = PRIV(ucd_caseless_sets) + prop_value;
            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;

          case PT_UCNC:
          for (i = 1; i <= min; i++)
            {
            if (eptr >= mb->end_subject)
              {
              SCHECK_PARTIAL();
              RRETURN(MATCH_NOMATCH);
              }
            GETCHARINCTEST(c, eptr);
            if ((c == CHAR_DOLLAR_SIGN || c == CHAR_COMMERCIAL_AT ||
                 c == CHAR_GRAVE_ACCENT || (c >= 0xa0 && c <= 0xd7ff) ||
                 c >= 0xe000) == prop_fail_result)
              RRETURN(MATCH_NOMATCH);
            }
          break;

          /* This should not occur */

          default:
          RRETURN(PCRE2_ERROR_INTERNAL);
          }
        }

      /* Match extended Unicode sequences. We will get here only if the
      support is in the binary; otherwise a compile-time error occurs. */

      else if (ctype == OP_EXTUNI)
        {
        for (i = 1; i <= min; i++)
          {
          if (eptr >= mb->end_subject)
            {
            SCHECK_PARTIAL();
            RRETURN(MATCH_NOMATCH);
            }
          else
            {
            int lgb, rgb;
            GETCHARINCTEST(c, eptr);
            lgb = UCD_GRAPHBREAK(c);
           while (eptr < mb->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();
          }
        }

      else
#endif     /* SUPPORT_UNICODE */

/* Handle all other cases when the coding is UTF-8 */

#ifdef SUPPORT_UNICODE
      if (utf) switch(ctype)
        {
        case OP_ANY:
        for (i = 1; i <= min; i++)
          {
          if (eptr >= mb->end_subject)
            {
            SCHECK_PARTIAL();
            RRETURN(MATCH_NOMATCH);
            }
          if (IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH);
          if (mb->partial != 0 &&
              eptr + 1 >= mb->end_subject &&
              NLBLOCK->nltype == NLTYPE_FIXED &&
              NLBLOCK->nllen == 2 &&
              UCHAR21(eptr) == NLBLOCK->nl[0])
            {
            mb->hitend = TRUE;
            if (mb->partial > 1) RRETURN(PCRE2_ERROR_PARTIAL);
            }
          eptr++;
          ACROSSCHAR(eptr < mb->end_subject, *eptr, eptr++);
          }
        break;

        case OP_ALLANY:
        for (i = 1; i <= min; i++)
          {
          if (eptr >= mb->end_subject)
            {
            SCHECK_PARTIAL();
            RRETURN(MATCH_NOMATCH);
            }
          eptr++;
          ACROSSCHAR(eptr < mb->end_subject, *eptr, eptr++);
          }
        break;

        case OP_ANYBYTE:
        if (eptr > mb->end_subject - min) RRETURN(MATCH_NOMATCH);
        eptr += min;
        break;

        case OP_ANYNL:
        for (i = 1; i <= min; i++)
          {
          if (eptr >= mb->end_subject)
            {
            SCHECK_PARTIAL();
            RRETURN(MATCH_NOMATCH);
            }
          GETCHARINC(c, eptr);
          switch(c)
            {
            default: RRETURN(MATCH_NOMATCH);

            case CHAR_CR:
            if (eptr < mb->end_subject && UCHAR21(eptr) == CHAR_LF) eptr++;
            break;

            case CHAR_LF:
            break;

            case CHAR_VT:
            case CHAR_FF:
            case CHAR_NEL:
#ifndef EBCDIC
            case 0x2028:
            case 0x2029:
#endif  /* Not EBCDIC */
            if (mb->bsr_convention == PCRE2_BSR_ANYCRLF) RRETURN(MATCH_NOMATCH);
            break;
            }
          }
        break;

        case OP_NOT_HSPACE:
        for (i = 1; i <= min; i++)
          {
          if (eptr >= mb->end_subject)
            {
            SCHECK_PARTIAL();
            RRETURN(MATCH_NOMATCH);
            }
          GETCHARINC(c, eptr);
          switch(c)
            {
            HSPACE_CASES: RRETURN(MATCH_NOMATCH);  /* Byte and multibyte cases */
            default: break;
            }
          }
        break;

        case OP_HSPACE:
        for (i = 1; i <= min; i++)
          {
          if (eptr >= mb->end_subject)
            {
            SCHECK_PARTIAL();
            RRETURN(MATCH_NOMATCH);
            }
          GETCHARINC(c, eptr);
          switch(c)
            {
            HSPACE_CASES: break;  /* Byte and multibyte cases */
            default: RRETURN(MATCH_NOMATCH);
            }
          }
        break;

        case OP_NOT_VSPACE:
        for (i = 1; i <= min; i++)
          {
          if (eptr >= mb->end_subject)
            {
            SCHECK_PARTIAL();
            RRETURN(MATCH_NOMATCH);
            }
          GETCHARINC(c, eptr);
          switch(c)
            {
            VSPACE_CASES: RRETURN(MATCH_NOMATCH);
            default: break;
            }
          }
        break;

        case OP_VSPACE:
        for (i = 1; i <= min; i++)
          {
          if (eptr >= mb->end_subject)
            {
            SCHECK_PARTIAL();
            RRETURN(MATCH_NOMATCH);
            }
          GETCHARINC(c, eptr);
          switch(c)
            {
            VSPACE_CASES: break;
            default: RRETURN(MATCH_NOMATCH);
            }
          }
        break;

        case OP_NOT_DIGIT:
        for (i = 1; i <= min; i++)
          {
          if (eptr >= mb->end_subject)
            {
            SCHECK_PARTIAL();
            RRETURN(MATCH_NOMATCH);
            }
          GETCHARINC(c, eptr);
          if (c < 128 && (mb->ctypes[c] & ctype_digit) != 0)
            RRETURN(MATCH_NOMATCH);
          }
        break;

        case OP_DIGIT:
        for (i = 1; i <= min; i++)
          {
          uint32_t cc;
          if (eptr >= mb->end_subject)
            {
            SCHECK_PARTIAL();
            RRETURN(MATCH_NOMATCH);
            }
          cc = UCHAR21(eptr);
          if (cc >= 128 || (mb->ctypes[cc] & ctype_digit) == 0)
            RRETURN(MATCH_NOMATCH);
          eptr++;
          /* No need to skip more bytes - we know it's a 1-byte character */
          }
        break;

        case OP_NOT_WHITESPACE:
        for (i = 1; i <= min; i++)
          {
          uint32_t cc;
          if (eptr >= mb->end_subject)
            {
            SCHECK_PARTIAL();
            RRETURN(MATCH_NOMATCH);
            }
          cc = UCHAR21(eptr);
          if (cc < 128 && (mb->ctypes[cc] & ctype_space) != 0)
            RRETURN(MATCH_NOMATCH);
          eptr++;
          ACROSSCHAR(eptr < mb->end_subject, *eptr, eptr++);
          }
        break;

        case OP_WHITESPACE:
        for (i = 1; i <= min; i++)
          {
          uint32_t cc;
          if (eptr >= mb->end_subject)
            {
            SCHECK_PARTIAL();
            RRETURN(MATCH_NOMATCH);
            }
          cc = UCHAR21(eptr);
          if (cc >= 128 || (mb->ctypes[cc] & ctype_space) == 0)
            RRETURN(MATCH_NOMATCH);
          eptr++;
          /* No need to skip more bytes - we know it's a 1-byte character */
          }
        break;

        case OP_NOT_WORDCHAR:
        for (i = 1; i <= min; i++)
          {
          uint32_t cc;
          if (eptr >= mb->end_subject)
            {
            SCHECK_PARTIAL();
            RRETURN(MATCH_NOMATCH);
            }
          cc = UCHAR21(eptr);
          if (cc < 128 && (mb->ctypes[cc] & ctype_word) != 0)
            RRETURN(MATCH_NOMATCH);
          eptr++;
          ACROSSCHAR(eptr < mb->end_subject, *eptr, eptr++);
          }
        break;

        case OP_WORDCHAR:
        for (i = 1; i <= min; i++)
          {
          uint32_t cc;
          if (eptr >= mb->end_subject)
            {
            SCHECK_PARTIAL();
            RRETURN(MATCH_NOMATCH);
            }
          cc = UCHAR21(eptr);
          if (cc >= 128 || (mb->ctypes[cc] & ctype_word) == 0)
            RRETURN(MATCH_NOMATCH);
          eptr++;
          /* No need to skip more bytes - we know it's a 1-byte character */
          }
        break;

        default:
        RRETURN(PCRE2_ERROR_INTERNAL);
        }  /* End switch(ctype) */

      else
#endif     /* SUPPORT_UNICODE */

      /* Code for the non-UTF-8 case for minimum matching of operators other
      than OP_PROP and OP_NOTPROP. */

      switch(ctype)
        {
        case OP_ANY:
        for (i = 1; i <= min; i++)
          {
          if (eptr >= mb->end_subject)
            {
            SCHECK_PARTIAL();
            RRETURN(MATCH_NOMATCH);
            }
          if (IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH);
          if (mb->partial != 0 &&
              eptr + 1 >= mb->end_subject &&
              NLBLOCK->nltype == NLTYPE_FIXED &&
              NLBLOCK->nllen == 2 &&
              *eptr == NLBLOCK->nl[0])
            {
            mb->hitend = TRUE;
            if (mb->partial > 1) RRETURN(PCRE2_ERROR_PARTIAL);
            }
          eptr++;
          }
        break;

        case OP_ALLANY:
        if (eptr > mb->end_subject - min)
          {
          SCHECK_PARTIAL();
          RRETURN(MATCH_NOMATCH);
          }
        eptr += min;
        break;

        case OP_ANYBYTE:
        if (eptr > mb->end_subject - min)
          {
          SCHECK_PARTIAL();
          RRETURN(MATCH_NOMATCH);
          }
        eptr += min;
        break;

        case OP_ANYNL:
        for (i = 1; i <= min; i++)
          {
          if (eptr >= mb->end_subject)
            {
            SCHECK_PARTIAL();
            RRETURN(MATCH_NOMATCH);
            }
          switch(*eptr++)
            {
            default: RRETURN(MATCH_NOMATCH);

            case CHAR_CR:
            if (eptr < mb->end_subject && *eptr == CHAR_LF) eptr++;
            break;

            case CHAR_LF:
            break;

            case CHAR_VT:
            case CHAR_FF:
            case CHAR_NEL:
#if PCRE2_CODE_UNIT_WIDTH != 8
            case 0x2028:
            case 0x2029:
#endif
            if (mb->bsr_convention == PCRE2_BSR_ANYCRLF) RRETURN(MATCH_NOMATCH);
            break;
            }
          }
        break;

        case OP_NOT_HSPACE:
        for (i = 1; i <= min; i++)
          {
          if (eptr >= mb->end_subject)
            {
            SCHECK_PARTIAL();
            RRETURN(MATCH_NOMATCH);
            }
          switch(*eptr++)
            {
            default: break;
            HSPACE_BYTE_CASES:
#if PCRE2_CODE_UNIT_WIDTH != 8
            HSPACE_MULTIBYTE_CASES:
#endif
            RRETURN(MATCH_NOMATCH);
            }
          }
        break;

        case OP_HSPACE:
        for (i = 1; i <= min; i++)
          {
          if (eptr >= mb->end_subject)
            {
            SCHECK_PARTIAL();
            RRETURN(MATCH_NOMATCH);
            }
          switch(*eptr++)
            {
            default: RRETURN(MATCH_NOMATCH);
            HSPACE_BYTE_CASES:
#if PCRE2_CODE_UNIT_WIDTH != 8
            HSPACE_MULTIBYTE_CASES:
#endif
            break;
            }
          }
        break;

        case OP_NOT_VSPACE:
        for (i = 1; i <= min; i++)
          {
          if (eptr >= mb->end_subject)
            {
            SCHECK_PARTIAL();
            RRETURN(MATCH_NOMATCH);
            }
          switch(*eptr++)
            {
            VSPACE_BYTE_CASES:
#if PCRE2_CODE_UNIT_WIDTH != 8
            VSPACE_MULTIBYTE_CASES:
#endif
            RRETURN(MATCH_NOMATCH);
            default: break;
            }
          }
        break;

        case OP_VSPACE:
        for (i = 1; i <= min; i++)
          {
          if (eptr >= mb->end_subject)
            {
            SCHECK_PARTIAL();
            RRETURN(MATCH_NOMATCH);
            }
          switch(*eptr++)
            {
            default: RRETURN(MATCH_NOMATCH);
            VSPACE_BYTE_CASES:
#if PCRE2_CODE_UNIT_WIDTH != 8
            VSPACE_MULTIBYTE_CASES:
#endif
            break;
            }
          }
        break;

        case OP_NOT_DIGIT:
        for (i = 1; i <= min; i++)
          {
          if (eptr >= mb->end_subject)
            {
            SCHECK_PARTIAL();
            RRETURN(MATCH_NOMATCH);
            }
          if (MAX_255(*eptr) && (mb->ctypes[*eptr] & ctype_digit) != 0)
            RRETURN(MATCH_NOMATCH);
          eptr++;
          }
        break;

        case OP_DIGIT:
        for (i = 1; i <= min; i++)
          {
          if (eptr >= mb->end_subject)
            {
            SCHECK_PARTIAL();
            RRETURN(MATCH_NOMATCH);
            }
          if (!MAX_255(*eptr) || (mb->ctypes[*eptr] & ctype_digit) == 0)
            RRETURN(MATCH_NOMATCH);
          eptr++;
          }
        break;

        case OP_NOT_WHITESPACE:
        for (i = 1; i <= min; i++)
          {
          if (eptr >= mb->end_subject)
            {
            SCHECK_PARTIAL();
            RRETURN(MATCH_NOMATCH);
            }
          if (MAX_255(*eptr) && (mb->ctypes[*eptr] & ctype_space) != 0)
            RRETURN(MATCH_NOMATCH);
          eptr++;
          }
        break;

        case OP_WHITESPACE:
        for (i = 1; i <= min; i++)
          {
          if (eptr >= mb->end_subject)
            {
            SCHECK_PARTIAL();
            RRETURN(MATCH_NOMATCH);
            }
          if (!MAX_255(*eptr) || (mb->ctypes[*eptr] & ctype_space) == 0)
            RRETURN(MATCH_NOMATCH);
          eptr++;
          }
        break;

        case OP_NOT_WORDCHAR:
        for (i = 1; i <= min; i++)
          {
          if (eptr >= mb->end_subject)
            {
            SCHECK_PARTIAL();
            RRETURN(MATCH_NOMATCH);
            }
          if (MAX_255(*eptr) && (mb->ctypes[*eptr] & ctype_word) != 0)
            RRETURN(MATCH_NOMATCH);
          eptr++;
          }
        break;

        case OP_WORDCHAR:
        for (i = 1; i <= min; i++)
          {
          if (eptr >= mb->end_subject)
            {
            SCHECK_PARTIAL();
            RRETURN(MATCH_NOMATCH);
            }
          if (!MAX_255(*eptr) || (mb->ctypes[*eptr] & ctype_word) == 0)
            RRETURN(MATCH_NOMATCH);
          eptr++;
          }
        break;

        default:
        RRETURN(PCRE2_ERROR_INTERNAL);
        }
      }

    /* If min = max, continue at the same level without recursing */

    if (min == max) continue;

    /* If minimizing, we have to test the rest of the pattern before each
    subsequent match. Again, separate the UTF-8 case for speed, and also
    separate the UCP cases. */

    if (minimize)
      {
#ifdef SUPPORT_UNICODE
      if (prop_type >= 0)
        {
        switch(prop_type)
          {
          case PT_ANY:
          for (fi = min;; fi++)
            {
            RMATCH(eptr, ecode, offset_top, mb, eptrb, RM36);
            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
            if (fi >= max) RRETURN(MATCH_NOMATCH);
            if (eptr >= mb->end_subject)
              {
              SCHECK_PARTIAL();
              RRETURN(MATCH_NOMATCH);
              }
            GETCHARINCTEST(c, eptr);
            if (prop_fail_result) RRETURN(MATCH_NOMATCH);
            }
          /* Control never gets here */

          case PT_LAMP:
          for (fi = min;; fi++)
            {
            int chartype;
            RMATCH(eptr, ecode, offset_top, mb, eptrb, RM37);
            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
            if (fi >= max) RRETURN(MATCH_NOMATCH);
            if (eptr >= mb->end_subject)
              {
              SCHECK_PARTIAL();
              RRETURN(MATCH_NOMATCH);
              }
            GETCHARINCTEST(c, eptr);
            chartype = UCD_CHARTYPE(c);
            if ((chartype == ucp_Lu ||
                 chartype == ucp_Ll ||
                 chartype == ucp_Lt) == prop_fail_result)
              RRETURN(MATCH_NOMATCH);
            }
          /* Control never gets here */

          case PT_GC:
          for (fi = min;; fi++)
            {
            RMATCH(eptr, ecode, offset_top, mb, eptrb, RM38);
            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
            if (fi >= max) RRETURN(MATCH_NOMATCH);
            if (eptr >= mb->end_subject)
              {
              SCHECK_PARTIAL();
              RRETURN(MATCH_NOMATCH);
              }
            GETCHARINCTEST(c, eptr);
            if ((UCD_CATEGORY(c) == prop_value) == prop_fail_result)
              RRETURN(MATCH_NOMATCH);
            }
          /* Control never gets here */

          case PT_PC:
          for (fi = min;; fi++)
            {
            RMATCH(eptr, ecode, offset_top, mb, eptrb, RM39);
            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
            if (fi >= max) RRETURN(MATCH_NOMATCH);
            if (eptr >= mb->end_subject)
              {
              SCHECK_PARTIAL();
              RRETURN(MATCH_NOMATCH);
              }
            GETCHARINCTEST(c, eptr);
            if ((UCD_CHARTYPE(c) == prop_value) == prop_fail_result)
              RRETURN(MATCH_NOMATCH);
            }
          /* Control never gets here */

          case PT_SC:
          for (fi = min;; fi++)
            {
            RMATCH(eptr, ecode, offset_top, mb, eptrb, RM40);
            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
            if (fi >= max) RRETURN(MATCH_NOMATCH);
            if (eptr >= mb->end_subject)
              {
              SCHECK_PARTIAL();
              RRETURN(MATCH_NOMATCH);
              }
            GETCHARINCTEST(c, eptr);
            if ((UCD_SCRIPT(c) == prop_value) == prop_fail_result)
              RRETURN(MATCH_NOMATCH);
            }
          /* Control never gets here */

          case PT_ALNUM:
          for (fi = min;; fi++)
            {
            int category;
            RMATCH(eptr, ecode, offset_top, mb, eptrb, RM59);
            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
            if (fi >= max) RRETURN(MATCH_NOMATCH);
            if (eptr >= mb->end_subject)
              {
              SCHECK_PARTIAL();
              RRETURN(MATCH_NOMATCH);
              }
            GETCHARINCTEST(c, eptr);
            category = UCD_CATEGORY(c);
            if ((category == ucp_L || category == ucp_N) == prop_fail_result)
              RRETURN(MATCH_NOMATCH);
            }
          /* Control never gets here */

          /* Perl space used to exclude VT, but from Perl 5.18 it is included,
          which means that Perl space and POSIX space are now identical. PCRE
          was changed at release 8.34. */

          case PT_SPACE:    /* Perl space */
          case PT_PXSPACE:  /* POSIX space */
          for (fi = min;; fi++)
            {
            RMATCH(eptr, ecode, offset_top, mb, eptrb, RM61);
            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
            if (fi >= max) RRETURN(MATCH_NOMATCH);
            if (eptr >= mb->end_subject)
              {
              SCHECK_PARTIAL();
              RRETURN(MATCH_NOMATCH);
              }
            GETCHARINCTEST(c, eptr);
            switch(c)
              {
              HSPACE_CASES:
              VSPACE_CASES:
              if (prop_fail_result) RRETURN(MATCH_NOMATCH);
              break;

              default:
              if ((UCD_CATEGORY(c) == ucp_Z) == prop_fail_result)
                RRETURN(MATCH_NOMATCH);
              break;
              }
            }
          /* Control never gets here */

          case PT_WORD:
          for (fi = min;; fi++)
            {
            int category;
            RMATCH(eptr, ecode, offset_top, mb, eptrb, RM62);
            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
            if (fi >= max) RRETURN(MATCH_NOMATCH);
            if (eptr >= mb->end_subject)
              {
              SCHECK_PARTIAL();
              RRETURN(MATCH_NOMATCH);
              }
            GETCHARINCTEST(c, eptr);
            category = UCD_CATEGORY(c);
            if ((category == ucp_L ||
                 category == ucp_N ||
                 c == CHAR_UNDERSCORE)
                   == prop_fail_result)
              RRETURN(MATCH_NOMATCH);
            }
          /* Control never gets here */

          case PT_CLIST:
          for (fi = min;; fi++)
            {
            const uint32_t *cp;
            RMATCH(eptr, ecode, offset_top, mb, eptrb, RM67);
            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
            if (fi >= max) RRETURN(MATCH_NOMATCH);
            if (eptr >= mb->end_subject)
              {
              SCHECK_PARTIAL();
              RRETURN(MATCH_NOMATCH);
              }
            GETCHARINCTEST(c, eptr);
            cp = PRIV(ucd_caseless_sets) + prop_value;
            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; }
              }
            }
          /* Control never gets here */

          case PT_UCNC:
          for (fi = min;; fi++)
            {
            RMATCH(eptr, ecode, offset_top, mb, eptrb, RM60);
            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
            if (fi >= max) RRETURN(MATCH_NOMATCH);
            if (eptr >= mb->end_subject)
              {
              SCHECK_PARTIAL();
              RRETURN(MATCH_NOMATCH);
              }
            GETCHARINCTEST(c, eptr);
            if ((c == CHAR_DOLLAR_SIGN || c == CHAR_COMMERCIAL_AT ||
                 c == CHAR_GRAVE_ACCENT || (c >= 0xa0 && c <= 0xd7ff) ||
                 c >= 0xe000) == prop_fail_result)
              RRETURN(MATCH_NOMATCH);
            }
          /* Control never gets here */

          /* This should never occur */
          default:
          RRETURN(PCRE2_ERROR_INTERNAL);
          }
        }

      /* Match extended Unicode sequences. We will get here only if the
      support is in the binary; otherwise a compile-time error occurs. */

      else if (ctype == OP_EXTUNI)
        {
        for (fi = min;; fi++)
          {
          RMATCH(eptr, ecode, offset_top, mb, eptrb, RM41);
          if (rrc != MATCH_NOMATCH) RRETURN(rrc);
          if (fi >= max) RRETURN(MATCH_NOMATCH);
          if (eptr >= mb->end_subject)
            {
            SCHECK_PARTIAL();
            RRETURN(MATCH_NOMATCH);
            }
          else
            {
            int lgb, rgb;
            GETCHARINCTEST(c, eptr);
            lgb = UCD_GRAPHBREAK(c);
            while (eptr < mb->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();
          }
        }
      else
#endif     /* SUPPORT_UNICODE */

#ifdef SUPPORT_UNICODE
      if (utf)
        {
        for (fi = min;; fi++)
          {
          RMATCH(eptr, ecode, offset_top, mb, eptrb, RM42);
          if (rrc != MATCH_NOMATCH) RRETURN(rrc);
          if (fi >= max) RRETURN(MATCH_NOMATCH);
          if (eptr >= mb->end_subject)
            {
            SCHECK_PARTIAL();
            RRETURN(MATCH_NOMATCH);
            }
          if (ctype == OP_ANY && IS_NEWLINE(eptr))
            RRETURN(MATCH_NOMATCH);
          GETCHARINC(c, eptr);
          switch(ctype)
            {
            case OP_ANY:               /* This is the non-NL case */
            if (mb->partial != 0 &&    /* Take care with CRLF partial */
                eptr >= mb->end_subject &&
                NLBLOCK->nltype == NLTYPE_FIXED &&
                NLBLOCK->nllen == 2 &&
                c == NLBLOCK->nl[0])
              {
              mb->hitend = TRUE;
              if (mb->partial > 1) RRETURN(PCRE2_ERROR_PARTIAL);
              }
            break;

            case OP_ALLANY:
            case OP_ANYBYTE:
            break;

            case OP_ANYNL:
            switch(c)
              {
              default: RRETURN(MATCH_NOMATCH);
              case CHAR_CR:
              if (eptr < mb->end_subject && UCHAR21(eptr) == CHAR_LF) eptr++;
              break;

              case CHAR_LF:
              break;

              case CHAR_VT:
              case CHAR_FF:
              case CHAR_NEL:
#ifndef EBCDIC
              case 0x2028:
              case 0x2029:
#endif  /* Not EBCDIC */
              if (mb->bsr_convention == PCRE2_BSR_ANYCRLF) RRETURN(MATCH_NOMATCH);
              break;
              }
            break;

            case OP_NOT_HSPACE:
            switch(c)
              {
              HSPACE_CASES: RRETURN(MATCH_NOMATCH);
              default: break;
              }
            break;

            case OP_HSPACE:
            switch(c)
              {
              HSPACE_CASES: break;
              default: RRETURN(MATCH_NOMATCH);
              }
            break;

            case OP_NOT_VSPACE:
            switch(c)
              {
              VSPACE_CASES: RRETURN(MATCH_NOMATCH);
              default: break;
              }
            break;

            case OP_VSPACE:
            switch(c)
              {
              VSPACE_CASES: break;
              default: RRETURN(MATCH_NOMATCH);
              }
            break;

            case OP_NOT_DIGIT:
            if (c < 256 && (mb->ctypes[c] & ctype_digit) != 0)
              RRETURN(MATCH_NOMATCH);
            break;

            case OP_DIGIT:
            if (c >= 256 || (mb->ctypes[c] & ctype_digit) == 0)
              RRETURN(MATCH_NOMATCH);
            break;

            case OP_NOT_WHITESPACE:
            if (c < 256 && (mb->ctypes[c] & ctype_space) != 0)
              RRETURN(MATCH_NOMATCH);
            break;

            case OP_WHITESPACE:
            if (c >= 256 || (mb->ctypes[c] & ctype_space) == 0)
              RRETURN(MATCH_NOMATCH);
            break;

            case OP_NOT_WORDCHAR:
            if (c < 256 && (mb->ctypes[c] & ctype_word) != 0)
              RRETURN(MATCH_NOMATCH);
            break;

            case OP_WORDCHAR:
            if (c >= 256 || (mb->ctypes[c] & ctype_word) == 0)
              RRETURN(MATCH_NOMATCH);
            break;

            default:
            RRETURN(PCRE2_ERROR_INTERNAL);
            }
          }
        }
      else
#endif
      /* Not UTF mode */
        {
        for (fi = min;; fi++)
          {
          RMATCH(eptr, ecode, offset_top, mb, eptrb, RM43);
          if (rrc != MATCH_NOMATCH) RRETURN(rrc);
          if (fi >= max) RRETURN(MATCH_NOMATCH);
          if (eptr >= mb->end_subject)
            {
            SCHECK_PARTIAL();
            RRETURN(MATCH_NOMATCH);
            }
          if (ctype == OP_ANY && IS_NEWLINE(eptr))
            RRETURN(MATCH_NOMATCH);
          c = *eptr++;
          switch(ctype)
            {
            case OP_ANY:               /* This is the non-NL case */
            if (mb->partial != 0 &&    /* Take care with CRLF partial */
                eptr >= mb->end_subject &&
                NLBLOCK->nltype == NLTYPE_FIXED &&
                NLBLOCK->nllen == 2 &&
                c == NLBLOCK->nl[0])
              {
              mb->hitend = TRUE;
              if (mb->partial > 1) RRETURN(PCRE2_ERROR_PARTIAL);
              }
            break;

            case OP_ALLANY:
            case OP_ANYBYTE:
            break;

            case OP_ANYNL:
            switch(c)
              {
              default: RRETURN(MATCH_NOMATCH);
              case CHAR_CR:
              if (eptr < mb->end_subject && *eptr == CHAR_LF) eptr++;
              break;

              case CHAR_LF:
              break;

              case CHAR_VT:
              case CHAR_FF:
              case CHAR_NEL:
#if PCRE2_CODE_UNIT_WIDTH != 8
              case 0x2028:
              case 0x2029:
#endif
              if (mb->bsr_convention == PCRE2_BSR_ANYCRLF) RRETURN(MATCH_NOMATCH);
              break;
              }
            break;

            case OP_NOT_HSPACE:
            switch(c)
              {
              default: break;
              HSPACE_BYTE_CASES:
#if PCRE2_CODE_UNIT_WIDTH != 8
              HSPACE_MULTIBYTE_CASES:
#endif
              RRETURN(MATCH_NOMATCH);
              }
            break;

            case OP_HSPACE:
            switch(c)
              {
              default: RRETURN(MATCH_NOMATCH);
              HSPACE_BYTE_CASES:
#if PCRE2_CODE_UNIT_WIDTH != 8
              HSPACE_MULTIBYTE_CASES:
#endif
              break;
              }
            break;

            case OP_NOT_VSPACE:
            switch(c)
              {
              default: break;
              VSPACE_BYTE_CASES:
#if PCRE2_CODE_UNIT_WIDTH != 8
              VSPACE_MULTIBYTE_CASES:
#endif
              RRETURN(MATCH_NOMATCH);
              }
            break;

            case OP_VSPACE:
            switch(c)
              {
              default: RRETURN(MATCH_NOMATCH);
              VSPACE_BYTE_CASES:
#if PCRE2_CODE_UNIT_WIDTH != 8
              VSPACE_MULTIBYTE_CASES:
#endif
              break;
              }
            break;

            case OP_NOT_DIGIT:
            if (MAX_255(c) && (mb->ctypes[c] & ctype_digit) != 0) RRETURN(MATCH_NOMATCH);
            break;

            case OP_DIGIT:
            if (!MAX_255(c) || (mb->ctypes[c] & ctype_digit) == 0) RRETURN(MATCH_NOMATCH);
            break;

            case OP_NOT_WHITESPACE:
            if (MAX_255(c) && (mb->ctypes[c] & ctype_space) != 0) RRETURN(MATCH_NOMATCH);
            break;

            case OP_WHITESPACE:
            if (!MAX_255(c) || (mb->ctypes[c] & ctype_space) == 0) RRETURN(MATCH_NOMATCH);
            break;

            case OP_NOT_WORDCHAR:
            if (MAX_255(c) && (mb->ctypes[c] & ctype_word) != 0) RRETURN(MATCH_NOMATCH);
            break;

            case OP_WORDCHAR:
            if (!MAX_255(c) || (mb->ctypes[c] & ctype_word) == 0) RRETURN(MATCH_NOMATCH);
            break;

            default:
            RRETURN(PCRE2_ERROR_INTERNAL);
            }
          }
        }
      /* Control never gets here */
      }

    /* If maximizing, it is worth using inline code for speed, doing the type
    test once at the start (i.e. keep it out of the loop). Again, keep the
    UTF-8 and UCP stuff separate. */

    else
      {
      pp = eptr;  /* Remember where we started */

#ifdef SUPPORT_UNICODE
      if (prop_type >= 0)
        {
        switch(prop_type)
          {
          case PT_ANY:
          for (i = min; i < max; i++)
            {
            int len = 1;
            if (eptr >= mb->end_subject)
              {
              SCHECK_PARTIAL();
              break;
              }
            GETCHARLENTEST(c, eptr, len);
            if (prop_fail_result) break;
            eptr+= len;
            }
          break;

          case PT_LAMP:
          for (i = min; i < max; i++)
            {
            int chartype;
            int len = 1;
            if (eptr >= mb->end_subject)
              {
              SCHECK_PARTIAL();
              break;
              }
            GETCHARLENTEST(c, eptr, len);
            chartype = UCD_CHARTYPE(c);
            if ((chartype == ucp_Lu ||
                 chartype == ucp_Ll ||
                 chartype == ucp_Lt) == prop_fail_result)
              break;
            eptr+= len;
            }
          break;

          case PT_GC:
          for (i = min; i < max; i++)
            {
            int len = 1;
            if (eptr >= mb->end_subject)
              {
              SCHECK_PARTIAL();
              break;
              }
            GETCHARLENTEST(c, eptr, len);
            if ((UCD_CATEGORY(c) == prop_value) == prop_fail_result) break;
            eptr+= len;
            }
          break;

          case PT_PC:
          for (i = min; i < max; i++)
            {
            int len = 1;
            if (eptr >= mb->end_subject)
              {
              SCHECK_PARTIAL();
              break;
              }
            GETCHARLENTEST(c, eptr, len);
            if ((UCD_CHARTYPE(c) == prop_value) == prop_fail_result) break;
            eptr+= len;
            }
          break;

          case PT_SC:
          for (i = min; i < max; i++)
            {
            int len = 1;
            if (eptr >= mb->end_subject)
              {
              SCHECK_PARTIAL();
              break;
              }
            GETCHARLENTEST(c, eptr, len);
            if ((UCD_SCRIPT(c) == prop_value) == prop_fail_result) break;
            eptr+= len;
            }
          break;

          case PT_ALNUM:
          for (i = min; i < max; i++)
            {
            int category;
            int len = 1;
            if (eptr >= mb->end_subject)
              {
              SCHECK_PARTIAL();
              break;
              }
            GETCHARLENTEST(c, eptr, len);
            category = UCD_CATEGORY(c);
            if ((category == ucp_L || category == ucp_N) == prop_fail_result)
              break;
            eptr+= len;
            }
          break;

          /* Perl space used to exclude VT, but from Perl 5.18 it is included,
          which means that Perl space and POSIX space are now identical. PCRE
          was changed at release 8.34. */

          case PT_SPACE:    /* Perl space */
          case PT_PXSPACE:  /* POSIX space */
          for (i = min; i < max; i++)
            {
            int len = 1;
            if (eptr >= mb->end_subject)
              {
              SCHECK_PARTIAL();
              break;
              }
            GETCHARLENTEST(c, eptr, len);
            switch(c)
              {
              HSPACE_CASES:
              VSPACE_CASES:
              if (prop_fail_result) goto ENDLOOP99;  /* Break the loop */
              break;

              default:
              if ((UCD_CATEGORY(c) == ucp_Z) == prop_fail_result)
                goto ENDLOOP99;   /* Break the loop */
              break;
              }
            eptr+= len;
            }
          ENDLOOP99:
          break;

          case PT_WORD:
          for (i = min; i < max; i++)
            {
            int category;
            int len = 1;
            if (eptr >= mb->end_subject)
              {
              SCHECK_PARTIAL();
              break;
              }
            GETCHARLENTEST(c, eptr, len);
            category = UCD_CATEGORY(c);
            if ((category == ucp_L || category == ucp_N ||
                 c == CHAR_UNDERSCORE) == prop_fail_result)
              break;
            eptr+= len;
            }
          break;

          case PT_CLIST:
          for (i = min; i < max; i++)
            {
            const uint32_t *cp;
            int len = 1;
            if (eptr >= mb->end_subject)
              {
              SCHECK_PARTIAL();
              break;
              }
            GETCHARLENTEST(c, eptr, len);
            cp = PRIV(ucd_caseless_sets) + prop_value;
            for (;;)
              {
              if (c < *cp)
                { if (prop_fail_result) break; else goto GOT_MAX; }
              if (c == *cp++)
                { if (prop_fail_result) goto GOT_MAX; else break; }
              }
            eptr += len;
            }
          GOT_MAX:
          break;

          case PT_UCNC:
          for (i = min; i < max; i++)
            {
            int len = 1;
            if (eptr >= mb->end_subject)
              {
              SCHECK_PARTIAL();
              break;
              }
            GETCHARLENTEST(c, eptr, len);
            if ((c == CHAR_DOLLAR_SIGN || c == CHAR_COMMERCIAL_AT ||
                 c == CHAR_GRAVE_ACCENT || (c >= 0xa0 && c <= 0xd7ff) ||
                 c >= 0xe000) == prop_fail_result)
              break;
            eptr += len;
            }
          break;

          default:
          RRETURN(PCRE2_ERROR_INTERNAL);
          }

        /* eptr is now past the end of the maximum run */

        if (possessive) continue;    /* No backtracking */

        /* After \C in UTF mode, pp might be in the middle of a Unicode
        character. Use <= pp to ensure backtracking doesn't go too far. */

        for(;;)
          {
          if (eptr <= pp) goto TAIL_RECURSE;
          RMATCH(eptr, ecode, offset_top, mb, eptrb, RM44);
          if (rrc != MATCH_NOMATCH) RRETURN(rrc);
          eptr--;
          if (utf) BACKCHAR(eptr);
          }
        }

      /* Match extended Unicode grapheme clusters. We will get here only if the
      support is in the binary; otherwise a compile-time error occurs. */

      else if (ctype == OP_EXTUNI)
        {
        for (i = min; i < max; i++)
          {
          if (eptr >= mb->end_subject)
            {
            SCHECK_PARTIAL();
            break;
            }
          else
            {
            int lgb, rgb;
            GETCHARINCTEST(c, eptr);
            lgb = UCD_GRAPHBREAK(c);
            while (eptr < mb->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();
          }

        /* eptr is now past the end of the maximum run */

        if (possessive) continue;    /* No backtracking */

        /* We use <= pp rather than == pp to detect the start of the run while
        backtracking because the use of \C in UTF mode can cause BACKCHAR to
        move back past pp. This is just palliative; the use of \C in UTF mode
        is fraught with danger. */

        for(;;)
          {
          int lgb, rgb;
          PCRE2_SPTR fptr;

          if (eptr <= pp) goto TAIL_RECURSE;   /* At start of char run */
          RMATCH(eptr, ecode, offset_top, mb, eptrb, RM45);
          if (rrc != MATCH_NOMATCH) RRETURN(rrc);

          /* Backtracking over an extended grapheme cluster involves inspecting
          the previous two characters (if present) to see if a break is
          permitted between them. */

          eptr--;
          if (!utf) c = *eptr; else
            {
            BACKCHAR(eptr);
            GETCHAR(c, eptr);
            }
          rgb = UCD_GRAPHBREAK(c);

          for (;;)
            {
            if (eptr <= pp) goto TAIL_RECURSE;   /* At start of char run */
            fptr = eptr - 1;
            if (!utf) c = *fptr; else
              {
              BACKCHAR(fptr);
              GETCHAR(c, fptr);
              }
            lgb = UCD_GRAPHBREAK(c);
            if ((PRIV(ucp_gbtable)[lgb] & (1 << rgb)) == 0) break;
            eptr = fptr;
            rgb = lgb;
            }
          }
        }

      else
#endif   /* SUPPORT_UNICODE */

#ifdef SUPPORT_UNICODE
      if (utf)
        {
        switch(ctype)
          {
          case OP_ANY:
          for (i = min; i < max; i++)
            {
            if (eptr >= mb->end_subject)
              {
              SCHECK_PARTIAL();
              break;
              }
            if (IS_NEWLINE(eptr)) break;
            if (mb->partial != 0 &&    /* Take care with CRLF partial */
                eptr + 1 >= mb->end_subject &&
                NLBLOCK->nltype == NLTYPE_FIXED &&
                NLBLOCK->nllen == 2 &&
                UCHAR21(eptr) == NLBLOCK->nl[0])
              {
              mb->hitend = TRUE;
              if (mb->partial > 1) RRETURN(PCRE2_ERROR_PARTIAL);
              }
            eptr++;
            ACROSSCHAR(eptr < mb->end_subject, *eptr, eptr++);
            }
          break;

          case OP_ALLANY:
          if (max < INT_MAX)
            {
            for (i = min; i < max; i++)
              {
              if (eptr >= mb->end_subject)
                {
                SCHECK_PARTIAL();
                break;
                }
              eptr++;
              ACROSSCHAR(eptr < mb->end_subject, *eptr, eptr++);
              }
            }
          else
            {
            eptr = mb->end_subject;   /* Unlimited UTF-8 repeat */
            SCHECK_PARTIAL();
            }
          break;

          /* The byte case is the same as non-UTF8 */

          case OP_ANYBYTE:
          c = max - min;
          if (c > (uint32_t)(mb->end_subject - eptr))
            {
            eptr = mb->end_subject;
            SCHECK_PARTIAL();
            }
          else eptr += c;
          break;

          case OP_ANYNL:
          for (i = min; i < max; i++)
            {
            int len = 1;
            if (eptr >= mb->end_subject)
              {
              SCHECK_PARTIAL();
              break;
              }
            GETCHARLEN(c, eptr, len);
            if (c == CHAR_CR)
              {
              if (++eptr >= mb->end_subject) break;
              if (UCHAR21(eptr) == CHAR_LF) eptr++;
              }
            else
              {
              if (c != CHAR_LF &&
                  (mb->bsr_convention == PCRE2_BSR_ANYCRLF ||
                   (c != CHAR_VT && c != CHAR_FF && c != CHAR_NEL
#ifndef EBCDIC
                    && c != 0x2028 && c != 0x2029
#endif  /* Not EBCDIC */
                    )))
                break;
              eptr += len;
              }
            }
          break;

          case OP_NOT_HSPACE:
          case OP_HSPACE:
          for (i = min; i < max; i++)
            {
            BOOL gotspace;
            int len = 1;
            if (eptr >= mb->end_subject)
              {
              SCHECK_PARTIAL();
              break;
              }
            GETCHARLEN(c, eptr, len);
            switch(c)
              {
              HSPACE_CASES: gotspace = TRUE; break;
              default: gotspace = FALSE; break;
              }
            if (gotspace == (ctype == OP_NOT_HSPACE)) break;
            eptr += len;
            }
          break;

          case OP_NOT_VSPACE:
          case OP_VSPACE:
          for (i = min; i < max; i++)
            {
            BOOL gotspace;
            int len = 1;
            if (eptr >= mb->end_subject)
              {
              SCHECK_PARTIAL();
              break;
              }
            GETCHARLEN(c, eptr, len);
            switch(c)
              {
              VSPACE_CASES: gotspace = TRUE; break;
              default: gotspace = FALSE; break;
              }
            if (gotspace == (ctype == OP_NOT_VSPACE)) break;
            eptr += len;
            }
          break;

          case OP_NOT_DIGIT:
          for (i = min; i < max; i++)
            {
            int len = 1;
            if (eptr >= mb->end_subject)
              {
              SCHECK_PARTIAL();
              break;
              }
            GETCHARLEN(c, eptr, len);
            if (c < 256 && (mb->ctypes[c] & ctype_digit) != 0) break;
            eptr+= len;
            }
          break;

          case OP_DIGIT:
          for (i = min; i < max; i++)
            {
            int len = 1;
            if (eptr >= mb->end_subject)
              {
              SCHECK_PARTIAL();
              break;
              }
            GETCHARLEN(c, eptr, len);
            if (c >= 256 ||(mb->ctypes[c] & ctype_digit) == 0) break;
            eptr+= len;
            }
          break;

          case OP_NOT_WHITESPACE:
          for (i = min; i < max; i++)
            {
            int len = 1;
            if (eptr >= mb->end_subject)
              {
              SCHECK_PARTIAL();
              break;
              }
            GETCHARLEN(c, eptr, len);
            if (c < 256 && (mb->ctypes[c] & ctype_space) != 0) break;
            eptr+= len;
            }
          break;

          case OP_WHITESPACE:
          for (i = min; i < max; i++)
            {
            int len = 1;
            if (eptr >= mb->end_subject)
              {
              SCHECK_PARTIAL();
              break;
              }
            GETCHARLEN(c, eptr, len);
            if (c >= 256 ||(mb->ctypes[c] & ctype_space) == 0) break;
            eptr+= len;
            }
          break;

          case OP_NOT_WORDCHAR:
          for (i = min; i < max; i++)
            {
            int len = 1;
            if (eptr >= mb->end_subject)
              {
              SCHECK_PARTIAL();
              break;
              }
            GETCHARLEN(c, eptr, len);
            if (c < 256 && (mb->ctypes[c] & ctype_word) != 0) break;
            eptr+= len;
            }
          break;

          case OP_WORDCHAR:
          for (i = min; i < max; i++)
            {
            int len = 1;
            if (eptr >= mb->end_subject)
              {
              SCHECK_PARTIAL();
              break;
              }
            GETCHARLEN(c, eptr, len);
            if (c >= 256 || (mb->ctypes[c] & ctype_word) == 0) break;
            eptr+= len;
            }
          break;

          default:
          RRETURN(PCRE2_ERROR_INTERNAL);
          }

        if (possessive) continue;    /* No backtracking */

        /* After \C in UTF mode, pp might be in the middle of a Unicode
        character. Use <= pp to ensure backtracking doesn't go too far. */

        for(;;)
          {
          if (eptr <= pp) goto TAIL_RECURSE;
          RMATCH(eptr, ecode, offset_top, mb, eptrb, RM46);
          if (rrc != MATCH_NOMATCH) RRETURN(rrc);
          eptr--;
          BACKCHAR(eptr);
          if (ctype == OP_ANYNL && eptr > pp  && UCHAR21(eptr) == CHAR_NL &&
              UCHAR21(eptr - 1) == CHAR_CR) eptr--;
          }
        }
      else
#endif  /* SUPPORT_UNICODE */
      /* Not UTF mode */
        {
        switch(ctype)
          {
          case OP_ANY:
          for (i = min; i < max; i++)
            {
            if (eptr >= mb->end_subject)
              {
              SCHECK_PARTIAL();
              break;
              }
            if (IS_NEWLINE(eptr)) break;
            if (mb->partial != 0 &&    /* Take care with CRLF partial */
                eptr + 1 >= mb->end_subject &&
                NLBLOCK->nltype == NLTYPE_FIXED &&
                NLBLOCK->nllen == 2 &&
                *eptr == NLBLOCK->nl[0])
              {
              mb->hitend = TRUE;
              if (mb->partial > 1) RRETURN(PCRE2_ERROR_PARTIAL);
              }
            eptr++;
            }
          break;

          case OP_ALLANY:
          case OP_ANYBYTE:
          c = max - min;
          if (c > (uint32_t)(mb->end_subject - eptr))
            {
            eptr = mb->end_subject;
            SCHECK_PARTIAL();
            }
          else eptr += c;
          break;

          case OP_ANYNL:
          for (i = min; i < max; i++)
            {
            if (eptr >= mb->end_subject)
              {
              SCHECK_PARTIAL();
              break;
              }
            c = *eptr;
            if (c == CHAR_CR)
              {
              if (++eptr >= mb->end_subject) break;
              if (*eptr == CHAR_LF) eptr++;
              }
            else
              {
              if (c != CHAR_LF && (mb->bsr_convention == PCRE2_BSR_ANYCRLF ||
                 (c != CHAR_VT && c != CHAR_FF && c != CHAR_NEL
#if PCRE2_CODE_UNIT_WIDTH != 8
                 && c != 0x2028 && c != 0x2029
#endif
                 ))) break;
              eptr++;
              }
            }
          break;

          case OP_NOT_HSPACE:
          for (i = min; i < max; i++)
            {
            if (eptr >= mb->end_subject)
              {
              SCHECK_PARTIAL();
              break;
              }
            switch(*eptr)
              {
              default: eptr++; break;
              HSPACE_BYTE_CASES:
#if PCRE2_CODE_UNIT_WIDTH != 8
              HSPACE_MULTIBYTE_CASES:
#endif
              goto ENDLOOP00;
              }
            }
          ENDLOOP00:
          break;

          case OP_HSPACE:
          for (i = min; i < max; i++)
            {
            if (eptr >= mb->end_subject)
              {
              SCHECK_PARTIAL();
              break;
              }
            switch(*eptr)
              {
              default: goto ENDLOOP01;
              HSPACE_BYTE_CASES:
#if PCRE2_CODE_UNIT_WIDTH != 8
              HSPACE_MULTIBYTE_CASES:
#endif
              eptr++; break;
              }
            }
          ENDLOOP01:
          break;

          case OP_NOT_VSPACE:
          for (i = min; i < max; i++)
            {
            if (eptr >= mb->end_subject)
              {
              SCHECK_PARTIAL();
              break;
              }
            switch(*eptr)
              {
              default: eptr++; break;
              VSPACE_BYTE_CASES:
#if PCRE2_CODE_UNIT_WIDTH != 8
              VSPACE_MULTIBYTE_CASES:
#endif
              goto ENDLOOP02;
              }
            }
          ENDLOOP02:
          break;

          case OP_VSPACE:
          for (i = min; i < max; i++)
            {
            if (eptr >= mb->end_subject)
              {
              SCHECK_PARTIAL();
              break;
              }
            switch(*eptr)
              {
              default: goto ENDLOOP03;
              VSPACE_BYTE_CASES:
#if PCRE2_CODE_UNIT_WIDTH != 8
              VSPACE_MULTIBYTE_CASES:
#endif
              eptr++; break;
              }
            }
          ENDLOOP03:
          break;

          case OP_NOT_DIGIT:
          for (i = min; i < max; i++)
            {
            if (eptr >= mb->end_subject)
              {
              SCHECK_PARTIAL();
              break;
              }
            if (MAX_255(*eptr) && (mb->ctypes[*eptr] & ctype_digit) != 0) break;
            eptr++;
            }
          break;

          case OP_DIGIT:
          for (i = min; i < max; i++)
            {
            if (eptr >= mb->end_subject)
              {
              SCHECK_PARTIAL();
              break;
              }
            if (!MAX_255(*eptr) || (mb->ctypes[*eptr] & ctype_digit) == 0) break;
            eptr++;
            }
          break;

          case OP_NOT_WHITESPACE:
          for (i = min; i < max; i++)
            {
            if (eptr >= mb->end_subject)
              {
              SCHECK_PARTIAL();
              break;
              }
            if (MAX_255(*eptr) && (mb->ctypes[*eptr] & ctype_space) != 0) break;
            eptr++;
            }
          break;

          case OP_WHITESPACE:
          for (i = min; i < max; i++)
            {
            if (eptr >= mb->end_subject)
              {
              SCHECK_PARTIAL();
              break;
              }
            if (!MAX_255(*eptr) || (mb->ctypes[*eptr] & ctype_space) == 0) break;
            eptr++;
            }
          break;

          case OP_NOT_WORDCHAR:
          for (i = min; i < max; i++)
            {
            if (eptr >= mb->end_subject)
              {
              SCHECK_PARTIAL();
              break;
              }
            if (MAX_255(*eptr) && (mb->ctypes[*eptr] & ctype_word) != 0) break;
            eptr++;
            }
          break;

          case OP_WORDCHAR:
          for (i = min; i < max; i++)
            {
            if (eptr >= mb->end_subject)
              {
              SCHECK_PARTIAL();
              break;
              }
            if (!MAX_255(*eptr) || (mb->ctypes[*eptr] & ctype_word) == 0) break;
            eptr++;
            }
          break;

          default:
          RRETURN(PCRE2_ERROR_INTERNAL);
          }

        if (possessive) continue;    /* No backtracking */
        for (;;)
          {
          if (eptr == pp) goto TAIL_RECURSE;
          RMATCH(eptr, ecode, offset_top, mb, eptrb, RM47);
          if (rrc != MATCH_NOMATCH) RRETURN(rrc);
          eptr--;
          if (ctype == OP_ANYNL && eptr > pp  && *eptr == CHAR_LF &&
              eptr[-1] == CHAR_CR) eptr--;
          }
        }

      /* Control never gets here */
      }

    /* There's been some horrible disaster. Arrival here can only mean there is
    something seriously wrong in the code above or the OP_xxx definitions. */

    default:
    RRETURN(PCRE2_ERROR_INTERNAL);
    }

  /* Do not stick any code in here without much thought; it is assumed
  that "continue" in the code above comes out to here to repeat the main
  loop. */

  }             /* End of main loop */
/* Control never reaches here */


/* When compiling to use the heap rather than the stack for recursive calls to
match(), the RRETURN() macro jumps here. The number that is saved in
frame->Xwhere indicates which label we actually want to return to. */

#ifdef HEAP_MATCH_RECURSE
#define LBL(val) case val: goto L_RM##val;
HEAP_RETURN:
switch (frame->Xwhere)
  {
  LBL( 1) LBL( 2) LBL( 3) LBL( 4) LBL( 5) LBL( 6) LBL( 7) LBL( 8)
  LBL( 9) LBL(10) LBL(11) LBL(12) LBL(13) LBL(14) LBL(15) LBL(17)
  LBL(19) LBL(24) LBL(25) LBL(26) LBL(27) LBL(29) LBL(31) LBL(33)
  LBL(35) LBL(43) LBL(47) LBL(48) LBL(49) LBL(50) LBL(51) LBL(52)
  LBL(53) LBL(54) LBL(55) LBL(56) LBL(57) LBL(58) LBL(63) LBL(64)
  LBL(65) LBL(66) LBL(68)
#ifdef SUPPORT_WIDE_CHARS
  LBL(20) LBL(21)
#endif
#ifdef SUPPORT_UNICODE
  LBL(16) LBL(18)
  LBL(22) LBL(23) LBL(28) LBL(30)
  LBL(32) LBL(34) LBL(42) LBL(46)
  LBL(36) LBL(37) LBL(38) LBL(39) LBL(40) LBL(41) LBL(44) LBL(45)
  LBL(59) LBL(60) LBL(61) LBL(62) LBL(67)
#endif  /* SUPPORT_UNICODE */
  default:
  return PCRE2_ERROR_INTERNAL;
  }
#undef LBL
#endif  /* HEAP_MATCH_RECURSE */
}


/***************************************************************************
****************************************************************************
                   RECURSION IN THE match() FUNCTION

Undefine all the macros that were defined above to handle this. */

#ifdef HEAP_MATCH_RECURSE
#undef eptr
#undef ecode
#undef mstart
#undef offset_top
#undef eptrb
#undef flags

#undef callpat
#undef charptr
#undef data
#undef next_ecode
#undef pp
#undef prev
#undef saved_eptr

#undef new_recursive

#undef cur_is_word
#undef condition
#undef prev_is_word

#undef ctype
#undef length
#undef max
#undef min
#undef number
#undef offset
#undef op
#undef save_capture_last
#undef save_offset1
#undef save_offset2
#undef save_offset3

#undef newptrb
#endif  /* HEAP_MATCH_RECURSE */

/* These two are defined as macros in both cases */

#undef fc
#undef fi

/***************************************************************************
***************************************************************************/


#ifdef HEAP_MATCH_RECURSE
/*************************************************
*          Release allocated heap frames         *
*************************************************/

/* This function releases all the allocated frames. The base frame is on the
machine stack, and so must not be freed.

Argument:
  frame_base    the address of the base frame
  mb            the match block

Returns:  nothing
*/

static void
release_match_heapframes (heapframe *frame_base, match_block *mb)
{
heapframe *nextframe = frame_base->Xnextframe;
while (nextframe != NULL)
  {
  heapframe *oldframe = nextframe;
  nextframe = nextframe->Xnextframe;
  mb->stack_memctl.free(oldframe, mb->stack_memctl.memory_data);
  }
}
#endif  /* HEAP_MATCH_RECURSE */



/*************************************************
*           Match a Regular Expression           *
*************************************************/

/* This function applies a compiled pattern to a subject string and picks out
portions of the string if it matches. Two elements in the vector are set for
each substring: the offsets to the start and end of the substring.

Arguments:
  code            points to the compiled expression
  subject         points to the subject string
  length          length of subject string (may contain binary zeros)
  start_offset    where to start in the subject string
  options         option bits
  match_data      points to a match_data block
  mcontext        points a PCRE2 context

Returns:          > 0 => success; value is the number of ovector pairs filled
                  = 0 => success, but ovector is not big enough
                   -1 => failed to match (PCRE2_ERROR_NOMATCH)
                   -2 => partial match (PCRE2_ERROR_PARTIAL)
                 < -2 => some kind of unexpected problem
*/

PCRE2_EXP_DEFN int PCRE2_CALL_CONVENTION
pcre2_match(const pcre2_code *code, PCRE2_SPTR subject, PCRE2_SIZE length,
  PCRE2_SIZE start_offset, uint32_t options, pcre2_match_data *match_data,
  pcre2_match_context *mcontext)
{
int rc;
int ocount;

const uint8_t *start_bits = NULL;

const pcre2_real_code *re = (const pcre2_real_code *)code;

BOOL anchored;
BOOL firstline;
BOOL has_first_cu = FALSE;
BOOL has_req_cu = FALSE;
BOOL startline;
BOOL using_temporary_offsets = FALSE;
BOOL utf;

PCRE2_UCHAR first_cu = 0;
PCRE2_UCHAR first_cu2 = 0;
PCRE2_UCHAR req_cu = 0;
PCRE2_UCHAR req_cu2 = 0;

PCRE2_SPTR bumpalong_limit;
PCRE2_SPTR end_subject;
PCRE2_SPTR start_match = subject + start_offset;
PCRE2_SPTR req_cu_ptr = start_match - 1;
PCRE2_SPTR start_partial = NULL;
PCRE2_SPTR match_partial = NULL;

/* We need to have mb pointing to a match block, because the IS_NEWLINE macro
is used below, and it expects NLBLOCK to be defined as a pointer. */

match_block actual_match_block;
match_block *mb = &actual_match_block;

#ifdef HEAP_MATCH_RECURSE
heapframe frame_zero;
frame_zero.Xprevframe = NULL;            /* Marks the top level */
frame_zero.Xnextframe = NULL;            /* None are allocated yet */
mb->match_frames_base = &frame_zero;
#endif

/* A length equal to PCRE2_ZERO_TERMINATED implies a zero-terminated
subject string. */

if (length == PCRE2_ZERO_TERMINATED) length = PRIV(strlen)(subject);
end_subject = subject + length;

/* Plausibility checks */

if ((options & ~PUBLIC_MATCH_OPTIONS) != 0) return PCRE2_ERROR_BADOPTION;
if (code == NULL || subject == NULL || match_data == NULL)
  return PCRE2_ERROR_NULL;
if (start_offset > length) return PCRE2_ERROR_BADOFFSET;

/* Check that the first field in the block is the magic number. */

if (re->magic_number != MAGIC_NUMBER) return PCRE2_ERROR_BADMAGIC;

/* Check the code unit width. */

if ((re->flags & PCRE2_MODE_MASK) != PCRE2_CODE_UNIT_WIDTH/8)
  return PCRE2_ERROR_BADMODE;

/* PCRE2_NOTEMPTY and PCRE2_NOTEMPTY_ATSTART are match-time flags in the
options variable for this function. Users of PCRE2 who are not calling the
function directly would like to have a way of setting these flags, in the same
way that they can set pcre2_compile() flags like PCRE2_NO_AUTOPOSSESS with
constructions like (*NO_AUTOPOSSESS). To enable this, (*NOTEMPTY) and
(*NOTEMPTY_ATSTART) set bits in the pattern's "flag" function which can now be
transferred to the options for this function. The bits are guaranteed to be
adjacent, but do not have the same values. This bit of Boolean trickery assumes
that the match-time bits are not more significant than the flag bits. If by
accident this is not the case, a compile-time division by zero error will
occur. */

#define FF (PCRE2_NOTEMPTY_SET|PCRE2_NE_ATST_SET)
#define OO (PCRE2_NOTEMPTY|PCRE2_NOTEMPTY_ATSTART)
options |= (re->flags & FF) / ((FF & (~FF+1)) / (OO & (~OO+1)));
#undef FF
#undef OO

/* A NULL match context means "use a default context" */

if (mcontext == NULL)
  mcontext = (pcre2_match_context *)(&PRIV(default_match_context));

/* These two settings are used in the code for checking a UTF string that
follows immediately afterwards. Other values in the mb block are used only
during interpretive pcre_match() processing, not when the JIT support is in
use, so they are set up later. */

utf = (re->overall_options & PCRE2_UTF) != 0;
mb->partial = ((options & PCRE2_PARTIAL_HARD) != 0)? 2 :
              ((options & PCRE2_PARTIAL_SOFT) != 0)? 1 : 0;

/* Check a UTF string for validity if required. For 8-bit and 16-bit strings,
we must also check that a starting offset does not point into the middle of a
multiunit character. We check only the portion of the subject that is going to
be inspected during matching - from the offset minus the maximum back reference
to the given length. This saves time when a small part of a large subject is
being matched by the use of a starting offset. Note that the maximum lookbehind
is a number of characters, not code units. */

#ifdef SUPPORT_UNICODE
if (utf && (options & PCRE2_NO_UTF_CHECK) == 0)
  {
  PCRE2_SPTR check_subject = start_match;  /* start_match includes offset */

  if (start_offset > 0)
    {
#if PCRE2_CODE_UNIT_WIDTH != 32
    unsigned int i;
    if (start_match < end_subject && NOT_FIRSTCU(*start_match))
      return PCRE2_ERROR_BADUTFOFFSET;
    for (i = re->max_lookbehind; i > 0 && check_subject > subject; i--)
      {
      check_subject--;
      while (check_subject > subject &&
#if PCRE2_CODE_UNIT_WIDTH == 8
      (*check_subject & 0xc0) == 0x80)
#else  /* 16-bit */
      (*check_subject & 0xfc00) == 0xdc00)
#endif /* PCRE2_CODE_UNIT_WIDTH == 8 */
        check_subject--;
      }
#else
    /* In the 32-bit library, one code unit equals one character. However,
    we cannot just subtract the lookbehind and then compare pointers, because
    a very large lookbehind could create an invalid pointer. */

    if (start_offset >= re->max_lookbehind)
      check_subject -= re->max_lookbehind;
    else
      check_subject = subject;
#endif  /* PCRE2_CODE_UNIT_WIDTH != 32 */
    }

  /* Validate the relevant portion of the subject. After an error, adjust the
  offset to be an absolute offset in the whole string. */

  match_data->rc = PRIV(valid_utf)(check_subject,
    length - (check_subject - subject), &(match_data->startchar));
  if (match_data->rc != 0)
    {
    match_data->startchar += check_subject - subject;
    return match_data->rc;
    }
  }
#endif  /* SUPPORT_UNICODE */

/* It is an error to set an offset limit without setting the flag at compile
time. */

if (mcontext->offset_limit != PCRE2_UNSET &&
     (re->overall_options & PCRE2_USE_OFFSET_LIMIT) == 0)
  return PCRE2_ERROR_BADOFFSETLIMIT;

/* If the pattern was successfully studied with JIT support, run the JIT
executable instead of the rest of this function. Most options must be set at
compile time for the JIT code to be usable. Fallback to the normal code path if
an unsupported option is set or if JIT returns BADOPTION (which means that the
selected normal or partial matching mode was not compiled). */

#ifdef SUPPORT_JIT
if (re->executable_jit != NULL && (options & ~PUBLIC_JIT_MATCH_OPTIONS) == 0)
  {
  rc = pcre2_jit_match(code, subject, length, start_offset, options,
    match_data, mcontext);
  if (rc != PCRE2_ERROR_JIT_BADOPTION) return rc;
  }
#endif

/* Carry on with non-JIT matching. */

anchored = ((re->overall_options | options) & PCRE2_ANCHORED) != 0;
firstline = (re->overall_options & PCRE2_FIRSTLINE) != 0;
startline = (re->flags & PCRE2_STARTLINE) != 0;
bumpalong_limit =  (mcontext->offset_limit == PCRE2_UNSET)?
  end_subject : subject + mcontext->offset_limit;

/* Fill in the fields in the match block. */

mb->callout = mcontext->callout;
mb->callout_data = mcontext->callout_data;
mb->memctl = mcontext->memctl;
#ifdef HEAP_MATCH_RECURSE
mb->stack_memctl = mcontext->stack_memctl;
#endif

mb->start_subject = subject;
mb->start_offset = start_offset;
mb->end_subject = end_subject;
mb->hasthen = (re->flags & PCRE2_HASTHEN) != 0;

mb->moptions = options;                 /* Match options */
mb->poptions = re->overall_options;     /* Pattern options */

mb->ignore_skip_arg = 0;
mb->mark = mb->nomatch_mark = NULL;     /* In case never set */
mb->recursive = NULL;                   /* No recursion at top level */
mb->ovecsave_chain = NULL;              /* No ovecsave blocks yet */
mb->hitend = FALSE;

/* The name table is needed for finding all the numbers associated with a
given name, for condition testing. The code follows the name table. */

mb->name_table = (PCRE2_UCHAR *)((uint8_t *)re + sizeof(pcre2_real_code));
mb->name_count = re->name_count;
mb->name_entry_size = re->name_entry_size;
mb->start_code = mb->name_table + re->name_count * re->name_entry_size;

/* Limits set in the pattern override the match context only if they are
smaller. */

mb->match_limit = (mcontext->match_limit < re->limit_match)?
                  mcontext->match_limit : re->limit_match;
mb->match_limit_recursion = (mcontext->recursion_limit < re->limit_recursion)?
                            mcontext->recursion_limit : re->limit_recursion;

/* Pointers to the individual character tables */

mb->lcc = re->tables + lcc_offset;
mb->fcc = re->tables + fcc_offset;
mb->ctypes = re->tables + ctypes_offset;

/* Process the \R and newline settings. */

mb->bsr_convention = re->bsr_convention;
mb->nltype = NLTYPE_FIXED;
switch(re->newline_convention)
  {
  case PCRE2_NEWLINE_CR:
  mb->nllen = 1;
  mb->nl[0] = CHAR_CR;
  break;

  case PCRE2_NEWLINE_LF:
  mb->nllen = 1;
  mb->nl[0] = CHAR_NL;
  break;

  case PCRE2_NEWLINE_CRLF:
  mb->nllen = 2;
  mb->nl[0] = CHAR_CR;
  mb->nl[1] = CHAR_NL;
  break;

  case PCRE2_NEWLINE_ANY:
  mb->nltype = NLTYPE_ANY;
  break;

  case PCRE2_NEWLINE_ANYCRLF:
  mb->nltype = NLTYPE_ANYCRLF;
  break;

  default: return PCRE2_ERROR_INTERNAL;
  }

/* If the expression has got more back references than the offsets supplied can
hold, we get a temporary chunk of memory to use during the matching. Otherwise,
we can use the vector supplied. The size of the ovector is three times the
value in the oveccount field. Two-thirds of it is pairs for storing matching
offsets, and the top third is working space. */

if (re->top_backref >= match_data->oveccount)
  {
  ocount = re->top_backref * 3 + 3;
  mb->ovector = (PCRE2_SIZE *)(mb->memctl.malloc(ocount * sizeof(PCRE2_SIZE),
    mb->memctl.memory_data));
  if (mb->ovector == NULL) return PCRE2_ERROR_NOMEMORY;
  using_temporary_offsets = TRUE;
  }
else
  {
  ocount = 3 * match_data->oveccount;
  mb->ovector = match_data->ovector;
  }

mb->offset_end = ocount;
mb->offset_max = (2*ocount)/3;

/* Reset the working variable associated with each extraction. These should
never be used unless previously set, but they get saved and restored, and so we
initialize them to avoid reading uninitialized locations. Also, unset the
offsets for the matched string. This is really just for tidiness with callouts,
in case they inspect these fields. */

if (ocount > 0)
  {
  PCRE2_SIZE *iptr = mb->ovector + ocount;
  PCRE2_SIZE *iend = iptr - re->top_bracket;
  if (iend < mb->ovector + 2) iend = mb->ovector + 2;
  while (--iptr >= iend) *iptr = PCRE2_UNSET;
  mb->ovector[0] = mb->ovector[1] = PCRE2_UNSET;
  }

/* Set up the first code unit to match, if available. The first_codeunit value
is never set for an anchored regular expression, but the anchoring may be
forced at run time, so we have to test for anchoring. The first code unit may
be unset for an unanchored pattern, of course. If there's no first code unit
there may be a bitmap of possible first characters. */

if (!anchored)
  {
  if ((re->flags & PCRE2_FIRSTSET) != 0)
    {
    has_first_cu = TRUE;
    first_cu = first_cu2 = (PCRE2_UCHAR)(re->first_codeunit);
    if ((re->flags & PCRE2_FIRSTCASELESS) != 0)
      {
      first_cu2 = TABLE_GET(first_cu, mb->fcc, first_cu);
#if defined SUPPORT_UNICODE && PCRE2_CODE_UNIT_WIDTH != 8
      if (utf && first_cu > 127) first_cu2 = UCD_OTHERCASE(first_cu);
#endif
      }
    }
  else
    if (!startline && (re->flags & PCRE2_FIRSTMAPSET) != 0)
      start_bits = re->start_bitmap;
  }

/* For anchored or unanchored matches, there may be a "last known required
character" set. */

if ((re->flags & PCRE2_LASTSET) != 0)
  {
  has_req_cu = TRUE;
  req_cu = req_cu2 = (PCRE2_UCHAR)(re->last_codeunit);
  if ((re->flags & PCRE2_LASTCASELESS) != 0)
    {
    req_cu2 = TABLE_GET(req_cu, mb->fcc, req_cu);
#if defined SUPPORT_UNICODE && PCRE2_CODE_UNIT_WIDTH != 8
    if (utf && req_cu > 127) req_cu2 = UCD_OTHERCASE(req_cu);
#endif
    }
  }


/* ==========================================================================*/

/* Loop for handling unanchored repeated matching attempts; for anchored regexs
the loop runs just once. */

for(;;)
  {
  PCRE2_SPTR new_start_match;
  mb->capture_last = 0;

  /* ----------------- Start of match optimizations ---------------- */

  /* There are some optimizations that avoid running the match if a known
  starting point is not found, or if a known later code unit is not present.
  However, there is an option (settable at compile time) that disables these,
  for testing and for ensuring that all callouts do actually occur. */

  if ((re->overall_options & PCRE2_NO_START_OPTIMIZE) == 0)
    {
    PCRE2_SPTR save_end_subject = end_subject;

    /* If firstline is TRUE, the start of the match is constrained to the first
    line of a multiline string. That is, the match must be before or at the
    first newline. Implement this by temporarily adjusting end_subject so that
    we stop the optimization scans at a newline. If the match fails at the
    newline, later code breaks this loop. */

    if (firstline)
      {
      PCRE2_SPTR t = start_match;
#ifdef SUPPORT_UNICODE
      if (utf)
        {
        while (t < mb->end_subject && !IS_NEWLINE(t))
          {
          t++;
          ACROSSCHAR(t < end_subject, *t, t++);
          }
        }
      else
#endif
      while (t < mb->end_subject && !IS_NEWLINE(t)) t++;
      end_subject = t;
      }

    /* Advance to a unique first code unit if there is one. In 8-bit mode, the
    use of memchr() gives a big speed up. */

    if (has_first_cu)
      {
      PCRE2_UCHAR smc;
      if (first_cu != first_cu2)
        while (start_match < end_subject &&
          (smc = UCHAR21TEST(start_match)) != first_cu && smc != first_cu2)
          start_match++;
      else
        {
#if PCRE2_CODE_UNIT_WIDTH != 8
        while (start_match < end_subject && UCHAR21TEST(start_match) != first_cu)
          start_match++;
#else
        start_match = memchr(start_match, first_cu, end_subject - start_match);
        if (start_match == NULL) start_match = end_subject;
#endif
        }
      }

    /* Or to just after a linebreak for a multiline match */

    else if (startline)
      {
      if (start_match > mb->start_subject + start_offset)
        {
#ifdef SUPPORT_UNICODE
        if (utf)
          {
          while (start_match < end_subject && !WAS_NEWLINE(start_match))
            {
            start_match++;
            ACROSSCHAR(start_match < end_subject, *start_match,
              start_match++);
            }
          }
        else
#endif
        while (start_match < end_subject && !WAS_NEWLINE(start_match))
          start_match++;

        /* If we have just passed a CR and the newline option is ANY or
        ANYCRLF, and we are now at a LF, advance the match position by one more
        code unit. */

        if (start_match[-1] == CHAR_CR &&
             (mb->nltype == NLTYPE_ANY || mb->nltype == NLTYPE_ANYCRLF) &&
             start_match < end_subject &&
             UCHAR21TEST(start_match) == CHAR_NL)
          start_match++;
        }
      }

    /* Or to a non-unique first code unit if any have been identified. The
    bitmap contains only 256 bits. When code units are 16 or 32 bits wide, all
    code units greater than 254 set the 255 bit. */

    else if (start_bits != NULL)
      {
      while (start_match < end_subject)
        {
        uint32_t c = UCHAR21TEST(start_match);
#if PCRE2_CODE_UNIT_WIDTH != 8
        if (c > 255) c = 255;
#endif
        if ((start_bits[c/8] & (1 << (c&7))) != 0) break;
        start_match++;
        }
      }

    /* Restore fudged end_subject */

    end_subject = save_end_subject;

    /* The following two optimizations are disabled for partial matching. */

    if (!mb->partial)
      {
      /* The minimum matching length is a lower bound; no actual string of that
      length may actually match the pattern. Although the value is, strictly,
      in characters, we treat it as code units to avoid spending too much time
      in this optimization. */

      if (end_subject - start_match < re->minlength)
        {
        rc = MATCH_NOMATCH;
        break;
        }

      /* If req_cu is set, we know that that code unit must appear in the
      subject for the match to succeed. If the first code unit is set, req_cu
      must be later in the subject; otherwise the test starts at the match
      point. This optimization can save a huge amount of backtracking in
      patterns with nested unlimited repeats that aren't going to match.
      Writing separate code for cased/caseless versions makes it go faster, as
      does using an autoincrement and backing off on a match.

      HOWEVER: when the subject string is very, very long, searching to its end
      can take a long time, and give bad performance on quite ordinary
      patterns. This showed up when somebody was matching something like
      /^\d+C/ on a 32-megabyte string... so we don't do this when the string is
      sufficiently long. */

      if (has_req_cu && end_subject - start_match < REQ_CU_MAX)
        {
        PCRE2_SPTR p = start_match + (has_first_cu? 1:0);

        /* We don't need to repeat the search if we haven't yet reached the
        place we found it at last time. */

        if (p > req_cu_ptr)
          {
          if (req_cu != req_cu2)
            {
            while (p < end_subject)
              {
              uint32_t pp = UCHAR21INCTEST(p);
              if (pp == req_cu || pp == req_cu2) { p--; break; }
              }
            }
          else
            {
            while (p < end_subject)
              {
              if (UCHAR21INCTEST(p) == req_cu) { p--; break; }
              }
            }

          /* If we can't find the required code unit, break the matching loop,
          forcing a match failure. */

          if (p >= end_subject)
            {
            rc = MATCH_NOMATCH;
            break;
            }

          /* If we have found the required code unit, save the point where we
          found it, so that we don't search again next time round the loop if
          the start hasn't passed this code unit yet. */

          req_cu_ptr = p;
          }
        }
      }
    }

  /* ------------ End of start of match optimizations ------------ */

  /* Give no match if we have passed the bumpalong limit. */

  if (start_match > bumpalong_limit)
    {
    rc = MATCH_NOMATCH;
    break;
    }

  /* OK, we can now run the match. If "hitend" is set afterwards, remember the
  first starting point for which a partial match was found. */

  mb->start_match_ptr = start_match;
  mb->start_used_ptr = start_match;
  mb->last_used_ptr = start_match;
  mb->match_call_count = 0;
  mb->match_function_type = 0;
  mb->end_offset_top = 0;
  mb->skip_arg_count = 0;
  rc = match(start_match, mb->start_code, start_match, 2, mb, NULL, 0);

  if (mb->hitend && start_partial == NULL)
    {
    start_partial = mb->start_used_ptr;
    match_partial = start_match;
    }

  switch(rc)
    {
    /* If MATCH_SKIP_ARG reaches this level it means that a MARK that matched
    the SKIP's arg was not found. In this circumstance, Perl ignores the SKIP
    entirely. The only way we can do that is to re-do the match at the same
    point, with a flag to force SKIP with an argument to be ignored. Just
    treating this case as NOMATCH does not work because it does not check other
    alternatives in patterns such as A(*SKIP:A)B|AC when the subject is AC. */

    case MATCH_SKIP_ARG:
    new_start_match = start_match;
    mb->ignore_skip_arg = mb->skip_arg_count;
    break;

    /* SKIP passes back the next starting point explicitly, but if it is no
    greater than the match we have just done, treat it as NOMATCH. */

    case MATCH_SKIP:
    if (mb->start_match_ptr > start_match)
      {
      new_start_match = mb->start_match_ptr;
      break;
      }
    /* Fall through */

    /* NOMATCH and PRUNE advance by one character. THEN at this level acts
    exactly like PRUNE. Unset ignore SKIP-with-argument. */

    case MATCH_NOMATCH:
    case MATCH_PRUNE:
    case MATCH_THEN:
    mb->ignore_skip_arg = 0;
    new_start_match = start_match + 1;
#ifdef SUPPORT_UNICODE
    if (utf)
      ACROSSCHAR(new_start_match < end_subject, *new_start_match,
        new_start_match++);
#endif
    break;

    /* COMMIT disables the bumpalong, but otherwise behaves as NOMATCH. */

    case MATCH_COMMIT:
    rc = MATCH_NOMATCH;
    goto ENDLOOP;

    /* Any other return is either a match, or some kind of error. */

    default:
    goto ENDLOOP;
    }

  /* Control reaches here for the various types of "no match at this point"
  result. Reset the code to MATCH_NOMATCH for subsequent checking. */

  rc = MATCH_NOMATCH;

  /* If PCRE2_FIRSTLINE is set, the match must happen before or at the first
  newline in the subject (though it may continue over the newline). Therefore,
  if we have just failed to match, starting at a newline, do not continue. */

  if (firstline && IS_NEWLINE(start_match)) break;

  /* Advance to new matching position */

  start_match = new_start_match;

  /* Break the loop if the pattern is anchored or if we have passed the end of
  the subject. */

  if (anchored || start_match > end_subject) break;

  /* If we have just passed a CR and we are now at a LF, and the pattern does
  not contain any explicit matches for \r or \n, and the newline option is CRLF
  or ANY or ANYCRLF, advance the match position by one more code unit. In
  normal matching start_match will aways be greater than the first position at
  this stage, but a failed *SKIP can cause a return at the same point, which is
  why the first test exists. */

  if (start_match > subject + start_offset &&
      start_match[-1] == CHAR_CR &&
      start_match < end_subject &&
      *start_match == CHAR_NL &&
      (re->flags & PCRE2_HASCRORLF) == 0 &&
        (mb->nltype == NLTYPE_ANY ||
         mb->nltype == NLTYPE_ANYCRLF ||
         mb->nllen == 2))
    start_match++;

  mb->mark = NULL;   /* Reset for start of next match attempt */
  }                  /* End of for(;;) "bumpalong" loop */

/* ==========================================================================*/

/* When we reach here, one of the stopping conditions is true:

(1) The match succeeded, either completely, or partially;

(2) The pattern is anchored or the match was failed by (*COMMIT);

(3) We are past the end of the subject or the bumpalong limit;

(4) PCRE2_FIRSTLINE is set and we have failed to match at a newline, because
    this option requests that a match occur at or before the first newline in
    the subject.

(5) Some kind of error occurred.

*/

ENDLOOP:

#ifdef HEAP_MATCH_RECURSE
release_match_heapframes(&frame_zero, mb);
#endif

/* Release any frames that were saved from recursions. */

while (mb->ovecsave_chain != NULL)
  {
  ovecsave_frame *this = mb->ovecsave_chain;
  mb->ovecsave_chain = this->next;
  mb->memctl.free(this, mb->memctl.memory_data);
  }

/* Fill in fields that are always returned in the match data. */

match_data->code = re;
match_data->subject = subject;
match_data->mark = mb->mark;
match_data->matchedby = PCRE2_MATCHEDBY_INTERPRETER;

/* Handle a fully successful match. */

if (rc == MATCH_MATCH || rc == MATCH_ACCEPT)
  {
  uint32_t arg_offset_max = 2 * match_data->oveccount;

  /* When the offset vector is big enough to deal with any backreferences,
  captured substring offsets will already be set up. In the case where we had
  to get some local memory to hold offsets for backreference processing, copy
  those that we can. In this case there need not be overflow if certain parts
  of the pattern were not used, even though there are more capturing
  parentheses than vector slots. */

  if (using_temporary_offsets)
    {
    if (arg_offset_max >= 4)
      {
      memcpy(match_data->ovector + 2, mb->ovector + 2,
        (arg_offset_max - 2) * sizeof(PCRE2_SIZE));
      }
    if (mb->end_offset_top > arg_offset_max) mb->capture_last |= OVFLBIT;
    mb->memctl.free(mb->ovector, mb->memctl.memory_data);
    }

  /* Set the return code to the number of captured strings, or 0 if there were
  too many to fit into the ovector. */

  match_data->rc = ((mb->capture_last & OVFLBIT) != 0)?
    0 : (int)mb->end_offset_top/2;

  /* If there is space in the offset vector, set any pairs that follow the
  highest-numbered captured string but are less than the number of capturing
  groups in the pattern (and are within the ovector) to PCRE2_UNSET. It is
  documented that this happens. In earlier versions, the whole set of potential
  capturing offsets was initialized each time round the loop, but this is
  handled differently now. "Gaps" are set to PCRE2_UNSET dynamically instead
  (this fixed a bug). Thus, it is only those at the end that need setting here.
  We can't just mark them all unset at the start of the whole thing because
  they may get set in one branch that is not the final matching branch. */

  if (mb->end_offset_top/2 <= re->top_bracket)
    {
    PCRE2_SIZE *iptr, *iend;
    int resetcount = re->top_bracket + 1;
    if (resetcount > match_data->oveccount) resetcount = match_data->oveccount;
    iptr = match_data->ovector + mb->end_offset_top;
    iend = match_data->ovector + 2 * resetcount;
    while (iptr < iend) *iptr++ = PCRE2_UNSET;
    }

  /* If there is space, set up the whole thing as substring 0. The value of
  mb->start_match_ptr might be modified if \K was encountered on the success
  matching path. */

  if (match_data->oveccount < 1) rc = 0; else
    {
    match_data->ovector[0] = mb->start_match_ptr - mb->start_subject;
    match_data->ovector[1] = mb->end_match_ptr - mb->start_subject;
    }

  /* Set the remaining returned values */

  match_data->startchar = start_match - subject;
  match_data->leftchar = mb->start_used_ptr - subject;
  match_data->rightchar = ((mb->last_used_ptr > mb->end_match_ptr)?
    mb->last_used_ptr : mb->end_match_ptr) - subject;
  return match_data->rc;
  }

/* Control gets here if there has been a partial match, an error, or if the
overall match attempt has failed at all permitted starting positions. Any mark
data is in the nomatch_mark field. */

match_data->mark = mb->nomatch_mark;

/* For anything other than nomatch or partial match, just return the code. */

if (rc != MATCH_NOMATCH && rc != PCRE2_ERROR_PARTIAL)
  match_data->rc = rc;

/* Else handle a partial match. */

else if (match_partial != NULL)
  {
  if (match_data->oveccount > 0)
    {
    match_data->ovector[0] = match_partial - subject;
    match_data->ovector[1] = end_subject - subject;
    }
  match_data->startchar = match_partial - subject;
  match_data->leftchar = start_partial - subject;
  match_data->rightchar = end_subject - subject;
  match_data->rc = PCRE2_ERROR_PARTIAL;
  }

/* Else this is the classic nomatch case. */

else match_data->rc = PCRE2_ERROR_NOMATCH;

/* Free any temporary offsets. */

if (using_temporary_offsets)
  mb->memctl.free(mb->ovector, mb->memctl.memory_data);
return match_data->rc;
}

/* End of pcre2_match.c */