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

Diff of /code/trunk/pcre_exec.c

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

revision 609 by ph10, Wed Jun 15 18:09:23 2011 UTC revision 852 by zherczeg, Thu Jan 5 19:18:12 2012 UTC
# Line 6  Line 6 
6  and semantics are as close as possible to those of the Perl 5 language.  and semantics are as close as possible to those of the Perl 5 language.
7    
8                         Written by Philip Hazel                         Written by Philip Hazel
9             Copyright (c) 1997-2011 University of Cambridge             Copyright (c) 1997-2012 University of Cambridge
10    
11  -----------------------------------------------------------------------------  -----------------------------------------------------------------------------
12  Redistribution and use in source and binary forms, with or without  Redistribution and use in source and binary forms, with or without
# Line 57  possible. There are also some static sup Line 57  possible. There are also some static sup
57  #undef min  #undef min
58  #undef max  #undef max
59    
60  /* Values for setting in md->match_function_type to indicate two special types  /* Values for setting in md->match_function_type to indicate two special types
61  of call to match(). We do it this way to save on using another stack variable,  of call to match(). We do it this way to save on using another stack variable,
62  as stack usage is to be discouraged. */  as stack usage is to be discouraged. */
63    
64  #define MATCH_CONDASSERT     1  /* Called to check a condition assertion */  #define MATCH_CONDASSERT     1  /* Called to check a condition assertion */
# Line 76  negative to avoid the external error cod Line 76  negative to avoid the external error cod
76  #define MATCH_ACCEPT       (-999)  #define MATCH_ACCEPT       (-999)
77  #define MATCH_COMMIT       (-998)  #define MATCH_COMMIT       (-998)
78  #define MATCH_KETRPOS      (-997)  #define MATCH_KETRPOS      (-997)
79  #define MATCH_PRUNE        (-996)  #define MATCH_ONCE         (-996)
80  #define MATCH_SKIP         (-995)  #define MATCH_PRUNE        (-995)
81  #define MATCH_SKIP_ARG     (-994)  #define MATCH_SKIP         (-994)
82  #define MATCH_THEN         (-993)  #define MATCH_SKIP_ARG     (-993)
83    #define MATCH_THEN         (-992)
 /* This is a convenience macro for code that occurs many times. */  
   
 #define MRRETURN(ra) \  
   { \  
   md->mark = markptr; \  
   RRETURN(ra); \  
   }  
84    
85  /* Maximum number of ints of offset to save on the stack for recursive calls.  /* Maximum number of ints of offset to save on the stack for recursive calls.
86  If the offset vector is bigger, malloc is used. This should be a multiple of 3,  If the offset vector is bigger, malloc is used. This should be a multiple of 3,
# Line 120  Returns:     nothing Line 113  Returns:     nothing
113  */  */
114    
115  static void  static void
116  pchars(const uschar *p, int length, BOOL is_subject, match_data *md)  pchars(const pcre_uchar *p, int length, BOOL is_subject, match_data *md)
117  {  {
118  unsigned int c;  unsigned int c;
119  if (is_subject && length > md->end_subject - p) length = md->end_subject - p;  if (is_subject && length > md->end_subject - p) length = md->end_subject - p;
# Line 137  while (length-- > 0) Line 130  while (length-- > 0)
130    
131  /* Normally, if a back reference hasn't been set, the length that is passed is  /* Normally, if a back reference hasn't been set, the length that is passed is
132  negative, so the match always fails. However, in JavaScript compatibility mode,  negative, so the match always fails. However, in JavaScript compatibility mode,
133  the length passed is zero. Note that in caseless UTF-8 mode, the number of  the length passed is zero. Note that in caseless UTF-8 mode, the number of
134  subject bytes matched may be different to the number of reference bytes.  subject bytes matched may be different to the number of reference bytes.
135    
136  Arguments:  Arguments:
# Line 151  Returns:      < 0 if not matched, otherw Line 144  Returns:      < 0 if not matched, otherw
144  */  */
145    
146  static int  static int
147  match_ref(int offset, register USPTR eptr, int length, match_data *md,  match_ref(int offset, register PCRE_PUCHAR eptr, int length, match_data *md,
148    BOOL caseless)    BOOL caseless)
149  {  {
150  USPTR eptr_start = eptr;  PCRE_PUCHAR eptr_start = eptr;
151  register USPTR p = md->start_subject + md->offset_vector[offset];  register PCRE_PUCHAR p = md->start_subject + md->offset_vector[offset];
152    
153  #ifdef PCRE_DEBUG  #ifdef PCRE_DEBUG
154  if (eptr >= md->end_subject)  if (eptr >= md->end_subject)
# Line 180  ASCII characters. */ Line 173  ASCII characters. */
173    
174  if (caseless)  if (caseless)
175    {    {
176  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
177  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
178    if (md->utf8)    if (md->utf)
179      {      {
180      /* Match characters up to the end of the reference. NOTE: the number of      /* Match characters up to the end of the reference. NOTE: the number of
181      bytes matched may differ, because there are some characters whose upper and      bytes matched may differ, because there are some characters whose upper and
182      lower case versions code as different numbers of bytes. For example, U+023A      lower case versions code as different numbers of bytes. For example, U+023A
183      (2 bytes in UTF-8) is the upper case version of U+2C65 (3 bytes in UTF-8);      (2 bytes in UTF-8) is the upper case version of U+2C65 (3 bytes in UTF-8);
184      a sequence of 3 of the former uses 6 bytes, as does a sequence of two of      a sequence of 3 of the former uses 6 bytes, as does a sequence of two of
185      the latter. It is important, therefore, to check the length along the      the latter. It is important, therefore, to check the length along the
186      reference, not along the subject (earlier code did this wrong). */      reference, not along the subject (earlier code did this wrong). */
187    
188      USPTR endptr = p + length;      PCRE_PUCHAR endptr = p + length;
189      while (p < endptr)      while (p < endptr)
190        {        {
191        int c, d;        int c, d;
# Line 209  if (caseless) Line 202  if (caseless)
202    /* The same code works when not in UTF-8 mode and in UTF-8 mode when there    /* The same code works when not in UTF-8 mode and in UTF-8 mode when there
203    is no UCP support. */    is no UCP support. */
204      {      {
205      if (eptr + length > md->end_subject) return -1;      if (eptr + length > md->end_subject) return -1;
206      while (length-- > 0)      while (length-- > 0)
207        { if (md->lcc[*p++] != md->lcc[*eptr++]) return -1; }        {
208      }        if (TABLE_GET(*p, md->lcc, *p) != TABLE_GET(*eptr, md->lcc, *eptr)) return -1;
209          p++;
210          eptr++;
211          }
212        }
213    }    }
214    
215  /* In the caseful case, we can just compare the bytes, whether or not we  /* In the caseful case, we can just compare the bytes, whether or not we
216  are in UTF-8 mode. */  are in UTF-8 mode. */
217    
218  else  else
219    {    {
220    if (eptr + length > md->end_subject) return -1;    if (eptr + length > md->end_subject) return -1;
221    while (length-- > 0) if (*p++ != *eptr++) return -1;    while (length-- > 0) if (*p++ != *eptr++) return -1;
222    }    }
223    
224  return eptr - eptr_start;  return (int)(eptr - eptr_start);
225  }  }
226    
227    
# Line 276  enum { RM1=1, RM2,  RM3,  RM4,  RM5,  RM Line 273  enum { RM1=1, RM2,  RM3,  RM4,  RM5,  RM
273         RM31,  RM32, RM33, RM34, RM35, RM36, RM37, RM38, RM39, RM40,         RM31,  RM32, RM33, RM34, RM35, RM36, RM37, RM38, RM39, RM40,
274         RM41,  RM42, RM43, RM44, RM45, RM46, RM47, RM48, RM49, RM50,         RM41,  RM42, RM43, RM44, RM45, RM46, RM47, RM48, RM49, RM50,
275         RM51,  RM52, RM53, RM54, RM55, RM56, RM57, RM58, RM59, RM60,         RM51,  RM52, RM53, RM54, RM55, RM56, RM57, RM58, RM59, RM60,
276         RM61,  RM62, RM63};         RM61,  RM62, RM63, RM64, RM65, RM66 };
277    
278  /* These versions of the macros use the stack, as normal. There are debugging  /* These versions of the macros use the stack, as normal. There are debugging
279  versions and production versions. Note that the "rw" argument of RMATCH isn't  versions and production versions. Note that the "rw" argument of RMATCH isn't
# Line 289  actually used in this definition. */ Line 286  actually used in this definition. */
286  #define RMATCH(ra,rb,rc,rd,re,rw) \  #define RMATCH(ra,rb,rc,rd,re,rw) \
287    { \    { \
288    printf("match() called in line %d\n", __LINE__); \    printf("match() called in line %d\n", __LINE__); \
289    rrc = match(ra,rb,mstart,markptr,rc,rd,re,rdepth+1); \    rrc = match(ra,rb,mstart,rc,rd,re,rdepth+1); \
290    printf("to line %d\n", __LINE__); \    printf("to line %d\n", __LINE__); \
291    }    }
292  #define RRETURN(ra) \  #define RRETURN(ra) \
# Line 299  actually used in this definition. */ Line 296  actually used in this definition. */
296    }    }
297  #else  #else
298  #define RMATCH(ra,rb,rc,rd,re,rw) \  #define RMATCH(ra,rb,rc,rd,re,rw) \
299    rrc = match(ra,rb,mstart,markptr,rc,rd,re,rdepth+1)    rrc = match(ra,rb,mstart,rc,rd,re,rdepth+1)
300  #define RRETURN(ra) return ra  #define RRETURN(ra) return ra
301  #endif  #endif
302    
# Line 314  argument of match(), which never changes Line 311  argument of match(), which never changes
311    
312  #define RMATCH(ra,rb,rc,rd,re,rw)\  #define RMATCH(ra,rb,rc,rd,re,rw)\
313    {\    {\
314    heapframe *newframe = (heapframe *)(pcre_stack_malloc)(sizeof(heapframe));\    heapframe *newframe = (heapframe *)(PUBL(stack_malloc))(sizeof(heapframe));\
315    if (newframe == NULL) RRETURN(PCRE_ERROR_NOMEMORY);\    if (newframe == NULL) RRETURN(PCRE_ERROR_NOMEMORY);\
316    frame->Xwhere = rw; \    frame->Xwhere = rw; \
317    newframe->Xeptr = ra;\    newframe->Xeptr = ra;\
318    newframe->Xecode = rb;\    newframe->Xecode = rb;\
319    newframe->Xmstart = mstart;\    newframe->Xmstart = mstart;\
   newframe->Xmarkptr = markptr;\  
320    newframe->Xoffset_top = rc;\    newframe->Xoffset_top = rc;\
321    newframe->Xeptrb = re;\    newframe->Xeptrb = re;\
322    newframe->Xrdepth = frame->Xrdepth + 1;\    newframe->Xrdepth = frame->Xrdepth + 1;\
# Line 336  argument of match(), which never changes Line 332  argument of match(), which never changes
332    {\    {\
333    heapframe *oldframe = frame;\    heapframe *oldframe = frame;\
334    frame = oldframe->Xprevframe;\    frame = oldframe->Xprevframe;\
335    (pcre_stack_free)(oldframe);\    (PUBL(stack_free))(oldframe);\
336    if (frame != NULL)\    if (frame != NULL)\
337      {\      {\
338      rrc = ra;\      rrc = ra;\
# Line 353  typedef struct heapframe { Line 349  typedef struct heapframe {
349    
350    /* Function arguments that may change */    /* Function arguments that may change */
351    
352    USPTR Xeptr;    PCRE_PUCHAR Xeptr;
353    const uschar *Xecode;    const pcre_uchar *Xecode;
354    USPTR Xmstart;    PCRE_PUCHAR Xmstart;
   USPTR Xmarkptr;  
355    int Xoffset_top;    int Xoffset_top;
356    eptrblock *Xeptrb;    eptrblock *Xeptrb;
357    unsigned int Xrdepth;    unsigned int Xrdepth;
358    
359    /* Function local variables */    /* Function local variables */
360    
361    USPTR Xcallpat;    PCRE_PUCHAR Xcallpat;
362  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
363    USPTR Xcharptr;    PCRE_PUCHAR Xcharptr;
364  #endif  #endif
365    USPTR Xdata;    PCRE_PUCHAR Xdata;
366    USPTR Xnext;    PCRE_PUCHAR Xnext;
367    USPTR Xpp;    PCRE_PUCHAR Xpp;
368    USPTR Xprev;    PCRE_PUCHAR Xprev;
369    USPTR Xsaved_eptr;    PCRE_PUCHAR Xsaved_eptr;
370    
371    recursion_info Xnew_recursive;    recursion_info Xnew_recursive;
372    
# Line 383  typedef struct heapframe { Line 378  typedef struct heapframe {
378    int Xprop_type;    int Xprop_type;
379    int Xprop_value;    int Xprop_value;
380    int Xprop_fail_result;    int Xprop_fail_result;
   int Xprop_category;  
   int Xprop_chartype;  
   int Xprop_script;  
381    int Xoclength;    int Xoclength;
382    uschar Xocchars[8];    pcre_uchar Xocchars[6];
383  #endif  #endif
384    
385    int Xcodelink;    int Xcodelink;
# Line 429  returns a negative (error) response, the Line 421  returns a negative (error) response, the
421  same response. */  same response. */
422    
423  /* These macros pack up tests that are used for partial matching, and which  /* These macros pack up tests that are used for partial matching, and which
424  appears several times in the code. We set the "hit end" flag if the pointer is  appear several times in the code. We set the "hit end" flag if the pointer is
425  at the end of the subject and also past the start of the subject (i.e.  at the end of the subject and also past the start of the subject (i.e.
426  something has been matched). For hard partial matching, we then return  something has been matched). For hard partial matching, we then return
427  immediately. The second one is used when we already know we are past the end of  immediately. The second one is used when we already know we are past the end of
# Line 440  the subject. */ Line 432  the subject. */
432        eptr > md->start_used_ptr) \        eptr > md->start_used_ptr) \
433      { \      { \
434      md->hitend = TRUE; \      md->hitend = TRUE; \
435      if (md->partial > 1) MRRETURN(PCRE_ERROR_PARTIAL); \      if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL); \
436      }      }
437    
438  #define SCHECK_PARTIAL()\  #define SCHECK_PARTIAL()\
439    if (md->partial != 0 && eptr > md->start_used_ptr) \    if (md->partial != 0 && eptr > md->start_used_ptr) \
440      { \      { \
441      md->hitend = TRUE; \      md->hitend = TRUE; \
442      if (md->partial > 1) MRRETURN(PCRE_ERROR_PARTIAL); \      if (md->partial > 1) RRETURN(PCRE_ERROR_PARTIAL); \
443      }      }
444    
445    
446  /* Performance note: It might be tempting to extract commonly used fields from  /* Performance note: It might be tempting to extract commonly used fields from
447  the md structure (e.g. utf8, end_subject) into individual variables to improve  the md structure (e.g. utf, end_subject) into individual variables to improve
448  performance. Tests using gcc on a SPARC disproved this; in the first case, it  performance. Tests using gcc on a SPARC disproved this; in the first case, it
449  made performance worse.  made performance worse.
450    
# Line 461  Arguments: Line 453  Arguments:
453     ecode       pointer to current position in compiled code     ecode       pointer to current position in compiled code
454     mstart      pointer to the current match start position (can be modified     mstart      pointer to the current match start position (can be modified
455                   by encountering \K)                   by encountering \K)
    markptr     pointer to the most recent MARK name, or NULL  
456     offset_top  current top pointer     offset_top  current top pointer
457     md          pointer to "static" info for the match     md          pointer to "static" info for the match
458     eptrb       pointer to chain of blocks containing eptr at start of     eptrb       pointer to chain of blocks containing eptr at start of
# Line 476  Returns:       MATCH_MATCH if matched Line 467  Returns:       MATCH_MATCH if matched
467  */  */
468    
469  static int  static int
470  match(REGISTER USPTR eptr, REGISTER const uschar *ecode, USPTR mstart,  match(REGISTER PCRE_PUCHAR eptr, REGISTER const pcre_uchar *ecode,
471    const uschar *markptr, int offset_top, match_data *md, eptrblock *eptrb,    PCRE_PUCHAR mstart, int offset_top, match_data *md, eptrblock *eptrb,
472    unsigned int rdepth)    unsigned int rdepth)
473  {  {
474  /* These variables do not need to be preserved over recursion in this function,  /* These variables do not need to be preserved over recursion in this function,
# Line 487  so they can be ordinary variables in all Line 478  so they can be ordinary variables in all
478  register int  rrc;         /* Returns from recursive calls */  register int  rrc;         /* Returns from recursive calls */
479  register int  i;           /* Used for loops not involving calls to RMATCH() */  register int  i;           /* Used for loops not involving calls to RMATCH() */
480  register unsigned int c;   /* Character values not kept over RMATCH() calls */  register unsigned int c;   /* Character values not kept over RMATCH() calls */
481  register BOOL utf8;        /* Local copy of UTF-8 flag for speed */  register BOOL utf;         /* Local copy of UTF flag for speed */
482    
483  BOOL minimize, possessive; /* Quantifier options */  BOOL minimize, possessive; /* Quantifier options */
484  BOOL caseless;  BOOL caseless;
# Line 499  heap storage. Set up the top-level frame Line 490  heap storage. Set up the top-level frame
490  heap whenever RMATCH() does a "recursion". See the macro definitions above. */  heap whenever RMATCH() does a "recursion". See the macro definitions above. */
491    
492  #ifdef NO_RECURSE  #ifdef NO_RECURSE
493  heapframe *frame = (heapframe *)(pcre_stack_malloc)(sizeof(heapframe));  heapframe *frame = (heapframe *)(PUBL(stack_malloc))(sizeof(heapframe));
494  if (frame == NULL) RRETURN(PCRE_ERROR_NOMEMORY);  if (frame == NULL) RRETURN(PCRE_ERROR_NOMEMORY);
495  frame->Xprevframe = NULL;            /* Marks the top level */  frame->Xprevframe = NULL;            /* Marks the top level */
496    
# Line 508  frame->Xprevframe = NULL;            /* Line 499  frame->Xprevframe = NULL;            /*
499  frame->Xeptr = eptr;  frame->Xeptr = eptr;
500  frame->Xecode = ecode;  frame->Xecode = ecode;
501  frame->Xmstart = mstart;  frame->Xmstart = mstart;
 frame->Xmarkptr = markptr;  
502  frame->Xoffset_top = offset_top;  frame->Xoffset_top = offset_top;
503  frame->Xeptrb = eptrb;  frame->Xeptrb = eptrb;
504  frame->Xrdepth = rdepth;  frame->Xrdepth = rdepth;
# Line 522  HEAP_RECURSE: Line 512  HEAP_RECURSE:
512  #define eptr               frame->Xeptr  #define eptr               frame->Xeptr
513  #define ecode              frame->Xecode  #define ecode              frame->Xecode
514  #define mstart             frame->Xmstart  #define mstart             frame->Xmstart
 #define markptr            frame->Xmarkptr  
515  #define offset_top         frame->Xoffset_top  #define offset_top         frame->Xoffset_top
516  #define eptrb              frame->Xeptrb  #define eptrb              frame->Xeptrb
517  #define rdepth             frame->Xrdepth  #define rdepth             frame->Xrdepth
518    
519  /* Ditto for the local variables */  /* Ditto for the local variables */
520    
521  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
522  #define charptr            frame->Xcharptr  #define charptr            frame->Xcharptr
523  #endif  #endif
524  #define callpat            frame->Xcallpat  #define callpat            frame->Xcallpat
# Line 550  HEAP_RECURSE: Line 539  HEAP_RECURSE:
539  #define prop_type          frame->Xprop_type  #define prop_type          frame->Xprop_type
540  #define prop_value         frame->Xprop_value  #define prop_value         frame->Xprop_value
541  #define prop_fail_result   frame->Xprop_fail_result  #define prop_fail_result   frame->Xprop_fail_result
 #define prop_category      frame->Xprop_category  
 #define prop_chartype      frame->Xprop_chartype  
 #define prop_script        frame->Xprop_script  
542  #define oclength           frame->Xoclength  #define oclength           frame->Xoclength
543  #define occhars            frame->Xocchars  #define occhars            frame->Xocchars
544  #endif  #endif
# Line 590  declarations can be cut out in a block. Line 576  declarations can be cut out in a block.
576  below are for variables that do not have to be preserved over a recursive call  below are for variables that do not have to be preserved over a recursive call
577  to RMATCH(). */  to RMATCH(). */
578    
579  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
580  const uschar *charptr;  const pcre_uchar *charptr;
581  #endif  #endif
582  const uschar *callpat;  const pcre_uchar *callpat;
583  const uschar *data;  const pcre_uchar *data;
584  const uschar *next;  const pcre_uchar *next;
585  USPTR         pp;  PCRE_PUCHAR       pp;
586  const uschar *prev;  const pcre_uchar *prev;
587  USPTR         saved_eptr;  PCRE_PUCHAR       saved_eptr;
588    
589  recursion_info new_recursive;  recursion_info new_recursive;
590    
591  BOOL cur_is_word;  BOOL cur_is_word;
592  BOOL condition;  BOOL condition;
593  BOOL prev_is_word;  BOOL prev_is_word;
594    
# Line 610  BOOL prev_is_word; Line 596  BOOL prev_is_word;
596  int prop_type;  int prop_type;
597  int prop_value;  int prop_value;
598  int prop_fail_result;  int prop_fail_result;
 int prop_category;  
 int prop_chartype;  
 int prop_script;  
599  int oclength;  int oclength;
600  uschar occhars[8];  pcre_uchar occhars[6];
601  #endif  #endif
602    
603  int codelink;  int codelink;
# Line 632  int stacksave[REC_STACK_SAVE_MAX]; Line 615  int stacksave[REC_STACK_SAVE_MAX];
615  eptrblock newptrb;  eptrblock newptrb;
616  #endif     /* NO_RECURSE */  #endif     /* NO_RECURSE */
617    
618  /* To save space on the stack and in the heap frame, I have doubled up on some  /* To save space on the stack and in the heap frame, I have doubled up on some
619  of the local variables that are used only in localised parts of the code, but  of the local variables that are used only in localised parts of the code, but
620  still need to be preserved over recursive calls of match(). These macros define  still need to be preserved over recursive calls of match(). These macros define
621  the alternative names that are used. */  the alternative names that are used. */
622    
623  #define allow_zero    cur_is_word  #define allow_zero    cur_is_word
# Line 642  the alternative names that are used. */ Line 625  the alternative names that are used. */
625  #define code_offset   codelink  #define code_offset   codelink
626  #define condassert    condition  #define condassert    condition
627  #define matched_once  prev_is_word  #define matched_once  prev_is_word
628    #define foc           number
629    
630  /* These statements are here to stop the compiler complaining about unitialized  /* These statements are here to stop the compiler complaining about unitialized
631  variables. */  variables. */
# Line 667  defined). However, RMATCH isn't like a f Line 651  defined). However, RMATCH isn't like a f
651  complicated macro. It has to be used in one particular way. This shouldn't,  complicated macro. It has to be used in one particular way. This shouldn't,
652  however, impact performance when true recursion is being used. */  however, impact performance when true recursion is being used. */
653    
654  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
655  utf8 = md->utf8;       /* Local copy of the flag */  utf = md->utf;       /* Local copy of the flag */
656  #else  #else
657  utf8 = FALSE;  utf = FALSE;
658  #endif  #endif
659    
660  /* First check that we haven't called match() too many times, or that we  /* First check that we haven't called match() too many times, or that we
# Line 680  if (md->match_call_count++ >= md->match_ Line 664  if (md->match_call_count++ >= md->match_
664  if (rdepth >= md->match_limit_recursion) RRETURN(PCRE_ERROR_RECURSIONLIMIT);  if (rdepth >= md->match_limit_recursion) RRETURN(PCRE_ERROR_RECURSIONLIMIT);
665    
666  /* At the start of a group with an unlimited repeat that may match an empty  /* At the start of a group with an unlimited repeat that may match an empty
667  string, the variable md->match_function_type is set to MATCH_CBEGROUP. It is  string, the variable md->match_function_type is set to MATCH_CBEGROUP. It is
668  done this way to save having to use another function argument, which would take  done this way to save having to use another function argument, which would take
669  up space on the stack. See also MATCH_CONDASSERT below.  up space on the stack. See also MATCH_CONDASSERT below.
670    
671  When MATCH_CBEGROUP is set, add the current subject pointer to the chain of  When MATCH_CBEGROUP is set, add the current subject pointer to the chain of
# Line 705  for (;;) Line 689  for (;;)
689    {    {
690    minimize = possessive = FALSE;    minimize = possessive = FALSE;
691    op = *ecode;    op = *ecode;
692    
693    switch(op)    switch(op)
694      {      {
695      case OP_MARK:      case OP_MARK:
696      markptr = ecode + 2;      md->nomatch_mark = ecode + 2;
697      RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode] + ecode[1], offset_top, md,      md->mark = NULL;    /* In case previously set by assertion */
698        RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode] + ecode[1], offset_top, md,
699        eptrb, RM55);        eptrb, RM55);
700        if ((rrc == MATCH_MATCH || rrc == MATCH_ACCEPT) &&
701             md->mark == NULL) md->mark = ecode + 2;
702    
703      /* A return of MATCH_SKIP_ARG means that matching failed at SKIP with an      /* A return of MATCH_SKIP_ARG means that matching failed at SKIP with an
704      argument, and we must check whether that argument matches this MARK's      argument, and we must check whether that argument matches this MARK's
# Line 720  for (;;) Line 707  for (;;)
707      position and return MATCH_SKIP. Otherwise, pass back the return code      position and return MATCH_SKIP. Otherwise, pass back the return code
708      unaltered. */      unaltered. */
709    
710      if (rrc == MATCH_SKIP_ARG &&      else if (rrc == MATCH_SKIP_ARG &&
711          strcmp((char *)markptr, (char *)(md->start_match_ptr)) == 0)          STRCMP_UC_UC(ecode + 2, md->start_match_ptr) == 0)
712        {        {
713        md->start_match_ptr = eptr;        md->start_match_ptr = eptr;
714        RRETURN(MATCH_SKIP);        RRETURN(MATCH_SKIP);
715        }        }
   
     if (md->mark == NULL) md->mark = markptr;  
716      RRETURN(rrc);      RRETURN(rrc);
717    
718      case OP_FAIL:      case OP_FAIL:
719      MRRETURN(MATCH_NOMATCH);      RRETURN(MATCH_NOMATCH);
720    
721      /* COMMIT overrides PRUNE, SKIP, and THEN */      /* COMMIT overrides PRUNE, SKIP, and THEN */
722    
723      case OP_COMMIT:      case OP_COMMIT:
724      RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,      RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md,
725        eptrb, RM52);        eptrb, RM52);
726      if (rrc != MATCH_NOMATCH && rrc != MATCH_PRUNE &&      if (rrc != MATCH_NOMATCH && rrc != MATCH_PRUNE &&
727          rrc != MATCH_SKIP && rrc != MATCH_SKIP_ARG &&          rrc != MATCH_SKIP && rrc != MATCH_SKIP_ARG &&
728          rrc != MATCH_THEN)          rrc != MATCH_THEN)
729        RRETURN(rrc);        RRETURN(rrc);
730      MRRETURN(MATCH_COMMIT);      RRETURN(MATCH_COMMIT);
731    
732      /* PRUNE overrides THEN */      /* PRUNE overrides THEN */
733    
734      case OP_PRUNE:      case OP_PRUNE:
735      RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,      RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md,
736        eptrb, RM51);        eptrb, RM51);
737      if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN) RRETURN(rrc);      if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN) RRETURN(rrc);
738      MRRETURN(MATCH_PRUNE);      RRETURN(MATCH_PRUNE);
739    
740      case OP_PRUNE_ARG:      case OP_PRUNE_ARG:
741      RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode] + ecode[1], offset_top, md,      md->nomatch_mark = ecode + 2;
742        md->mark = NULL;    /* In case previously set by assertion */
743        RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode] + ecode[1], offset_top, md,
744        eptrb, RM56);        eptrb, RM56);
745        if ((rrc == MATCH_MATCH || rrc == MATCH_ACCEPT) &&
746             md->mark == NULL) md->mark = ecode + 2;
747      if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN) RRETURN(rrc);      if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN) RRETURN(rrc);
     md->mark = ecode + 2;  
748      RRETURN(MATCH_PRUNE);      RRETURN(MATCH_PRUNE);
749    
750      /* SKIP overrides PRUNE and THEN */      /* SKIP overrides PRUNE and THEN */
751    
752      case OP_SKIP:      case OP_SKIP:
753      RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,      RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md,
754        eptrb, RM53);        eptrb, RM53);
755      if (rrc != MATCH_NOMATCH && rrc != MATCH_PRUNE && rrc != MATCH_THEN)      if (rrc != MATCH_NOMATCH && rrc != MATCH_PRUNE && rrc != MATCH_THEN)
756        RRETURN(rrc);        RRETURN(rrc);
757      md->start_match_ptr = eptr;   /* Pass back current position */      md->start_match_ptr = eptr;   /* Pass back current position */
758      MRRETURN(MATCH_SKIP);      RRETURN(MATCH_SKIP);
759    
760        /* Note that, for Perl compatibility, SKIP with an argument does NOT set
761        nomatch_mark. There is a flag that disables this opcode when re-matching a
762        pattern that ended with a SKIP for which there was not a matching MARK. */
763    
764      case OP_SKIP_ARG:      case OP_SKIP_ARG:
765      RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode] + ecode[1], offset_top, md,      if (md->ignore_skip_arg)
766          {
767          ecode += PRIV(OP_lengths)[*ecode] + ecode[1];
768          break;
769          }
770        RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode] + ecode[1], offset_top, md,
771        eptrb, RM57);        eptrb, RM57);
772      if (rrc != MATCH_NOMATCH && rrc != MATCH_PRUNE && rrc != MATCH_THEN)      if (rrc != MATCH_NOMATCH && rrc != MATCH_PRUNE && rrc != MATCH_THEN)
773        RRETURN(rrc);        RRETURN(rrc);
774    
775      /* Pass back the current skip name by overloading md->start_match_ptr and      /* Pass back the current skip name by overloading md->start_match_ptr and
776      returning the special MATCH_SKIP_ARG return code. This will either be      returning the special MATCH_SKIP_ARG return code. This will either be
777      caught by a matching MARK, or get to the top, where it is treated the same      caught by a matching MARK, or get to the top, where it causes a rematch
778      as PRUNE. */      with the md->ignore_skip_arg flag set. */
779    
780      md->start_match_ptr = ecode + 2;      md->start_match_ptr = ecode + 2;
781      RRETURN(MATCH_SKIP_ARG);      RRETURN(MATCH_SKIP_ARG);
782    
783      /* For THEN (and THEN_ARG) we pass back the address of the bracket or      /* For THEN (and THEN_ARG) we pass back the address of the opcode, so that
784      the alt that is at the start of the current branch. This makes it possible      the branch in which it occurs can be determined. Overload the start of
785      to skip back past alternatives that precede the THEN within the current      match pointer to do this. */
     branch. */  
786    
787      case OP_THEN:      case OP_THEN:
788      RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,      RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md,
789        eptrb, RM54);        eptrb, RM54);
790      if (rrc != MATCH_NOMATCH) RRETURN(rrc);      if (rrc != MATCH_NOMATCH) RRETURN(rrc);
791      md->start_match_ptr = ecode - GET(ecode, 1);      md->start_match_ptr = ecode;
792      MRRETURN(MATCH_THEN);      RRETURN(MATCH_THEN);
793    
794      case OP_THEN_ARG:      case OP_THEN_ARG:
795      RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode] + ecode[1+LINK_SIZE],      md->nomatch_mark = ecode + 2;
796        offset_top, md, eptrb, RM58);      md->mark = NULL;    /* In case previously set by assertion */
797        RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode] + ecode[1], offset_top,
798          md, eptrb, RM58);
799        if ((rrc == MATCH_MATCH || rrc == MATCH_ACCEPT) &&
800             md->mark == NULL) md->mark = ecode + 2;
801      if (rrc != MATCH_NOMATCH) RRETURN(rrc);      if (rrc != MATCH_NOMATCH) RRETURN(rrc);
802      md->start_match_ptr = ecode - GET(ecode, 1);      md->start_match_ptr = ecode;
     md->mark = ecode + LINK_SIZE + 2;  
803      RRETURN(MATCH_THEN);      RRETURN(MATCH_THEN);
804    
805        /* Handle an atomic group that does not contain any capturing parentheses.
806        This can be handled like an assertion. Prior to 8.13, all atomic groups
807        were handled this way. In 8.13, the code was changed as below for ONCE, so
808        that backups pass through the group and thereby reset captured values.
809        However, this uses a lot more stack, so in 8.20, atomic groups that do not
810        contain any captures generate OP_ONCE_NC, which can be handled in the old,
811        less stack intensive way.
812    
813        Check the alternative branches in turn - the matching won't pass the KET
814        for this kind of subpattern. If any one branch matches, we carry on as at
815        the end of a normal bracket, leaving the subject pointer, but resetting
816        the start-of-match value in case it was changed by \K. */
817    
818        case OP_ONCE_NC:
819        prev = ecode;
820        saved_eptr = eptr;
821        do
822          {
823          RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, eptrb, RM64);
824          if (rrc == MATCH_MATCH)  /* Note: _not_ MATCH_ACCEPT */
825            {
826            mstart = md->start_match_ptr;
827            break;
828            }
829          if (rrc == MATCH_THEN)
830            {
831            next = ecode + GET(ecode,1);
832            if (md->start_match_ptr < next &&
833                (*ecode == OP_ALT || *next == OP_ALT))
834              rrc = MATCH_NOMATCH;
835            }
836    
837          if (rrc != MATCH_NOMATCH) RRETURN(rrc);
838          ecode += GET(ecode,1);
839          }
840        while (*ecode == OP_ALT);
841    
842        /* If hit the end of the group (which could be repeated), fail */
843    
844        if (*ecode != OP_ONCE_NC && *ecode != OP_ALT) RRETURN(MATCH_NOMATCH);
845    
846        /* Continue as from after the group, updating the offsets high water
847        mark, since extracts may have been taken. */
848    
849        do ecode += GET(ecode, 1); while (*ecode == OP_ALT);
850    
851        offset_top = md->end_offset_top;
852        eptr = md->end_match_ptr;
853    
854        /* For a non-repeating ket, just continue at this level. This also
855        happens for a repeating ket if no characters were matched in the group.
856        This is the forcible breaking of infinite loops as implemented in Perl
857        5.005. */
858    
859        if (*ecode == OP_KET || eptr == saved_eptr)
860          {
861          ecode += 1+LINK_SIZE;
862          break;
863          }
864    
865        /* The repeating kets try the rest of the pattern or restart from the
866        preceding bracket, in the appropriate order. The second "call" of match()
867        uses tail recursion, to avoid using another stack frame. */
868    
869        if (*ecode == OP_KETRMIN)
870          {
871          RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, eptrb, RM65);
872          if (rrc != MATCH_NOMATCH) RRETURN(rrc);
873          ecode = prev;
874          goto TAIL_RECURSE;
875          }
876        else  /* OP_KETRMAX */
877          {
878          md->match_function_type = MATCH_CBEGROUP;
879          RMATCH(eptr, prev, offset_top, md, eptrb, RM66);
880          if (rrc != MATCH_NOMATCH) RRETURN(rrc);
881          ecode += 1 + LINK_SIZE;
882          goto TAIL_RECURSE;
883          }
884        /* Control never gets here */
885    
886      /* Handle a capturing bracket, other than those that are possessive with an      /* Handle a capturing bracket, other than those that are possessive with an
887      unlimited repeat. If there is space in the offset vector, save the current      unlimited repeat. If there is space in the offset vector, save the current
888      subject position in the working slot at the top of the vector. We mustn't      subject position in the working slot at the top of the vector. We mustn't
889      change the current values of the data slot, because they may be set from a      change the current values of the data slot, because they may be set from a
890      previous iteration of this group, and be referred to by a reference inside      previous iteration of this group, and be referred to by a reference inside
891      the group. If we fail to match, we need to restore this value and also the      the group. A failure to match might occur after the group has succeeded,
892      values of the final offsets, in case they were set by a previous iteration      if something later on doesn't match. For this reason, we need to restore
893      of the same bracket.      the working value and also the values of the final offsets, in case they
894        were set by a previous iteration of the same bracket.
895    
896      If there isn't enough space in the offset vector, treat this as if it were      If there isn't enough space in the offset vector, treat this as if it were
897      a non-capturing bracket. Don't worry about setting the flag for the error      a non-capturing bracket. Don't worry about setting the flag for the error
# Line 820  for (;;) Line 901  for (;;)
901      case OP_SCBRA:      case OP_SCBRA:
902      number = GET2(ecode, 1+LINK_SIZE);      number = GET2(ecode, 1+LINK_SIZE);
903      offset = number << 1;      offset = number << 1;
904    
905  #ifdef PCRE_DEBUG  #ifdef PCRE_DEBUG
906      printf("start bracket %d\n", number);      printf("start bracket %d\n", number);
907      printf("subject=");      printf("subject=");
# Line 841  for (;;) Line 922  for (;;)
922    
923        for (;;)        for (;;)
924          {          {
925          if (op >= OP_SBRA) md->match_function_type = MATCH_CBEGROUP;          if (op >= OP_SBRA) md->match_function_type = MATCH_CBEGROUP;
926          RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,          RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md,
927            eptrb, RM1);            eptrb, RM1);
928          if (rrc != MATCH_NOMATCH &&          if (rrc == MATCH_ONCE) break;  /* Backing up through an atomic group */
929              (rrc != MATCH_THEN || md->start_match_ptr != ecode))  
930            RRETURN(rrc);          /* If we backed up to a THEN, check whether it is within the current
931            branch by comparing the address of the THEN that is passed back with
932            the end of the branch. If it is within the current branch, and the
933            branch is one of two or more alternatives (it either starts or ends
934            with OP_ALT), we have reached the limit of THEN's action, so convert
935            the return code to NOMATCH, which will cause normal backtracking to
936            happen from now on. Otherwise, THEN is passed back to an outer
937            alternative. This implements Perl's treatment of parenthesized groups,
938            where a group not containing | does not affect the current alternative,
939            that is, (X) is NOT the same as (X|(*F)). */
940    
941            if (rrc == MATCH_THEN)
942              {
943              next = ecode + GET(ecode,1);
944              if (md->start_match_ptr < next &&
945                  (*ecode == OP_ALT || *next == OP_ALT))
946                rrc = MATCH_NOMATCH;
947              }
948    
949            /* Anything other than NOMATCH is passed back. */
950    
951            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
952          md->capture_last = save_capture_last;          md->capture_last = save_capture_last;
953          ecode += GET(ecode, 1);          ecode += GET(ecode, 1);
954          if (*ecode != OP_ALT) break;          if (*ecode != OP_ALT) break;
955          }          }
956    
957        DPRINTF(("bracket %d failed\n", number));        DPRINTF(("bracket %d failed\n", number));
   
958        md->offset_vector[offset] = save_offset1;        md->offset_vector[offset] = save_offset1;
959        md->offset_vector[offset+1] = save_offset2;        md->offset_vector[offset+1] = save_offset2;
960        md->offset_vector[md->offset_end - number] = save_offset3;        md->offset_vector[md->offset_end - number] = save_offset3;
961    
962        if (rrc != MATCH_THEN && md->mark == NULL) md->mark = markptr;        /* At this point, rrc will be one of MATCH_ONCE or MATCH_NOMATCH. */
963        RRETURN(MATCH_NOMATCH);  
964          RRETURN(rrc);
965        }        }
966    
967      /* FALL THROUGH ... Insufficient room for saving captured contents. Treat      /* FALL THROUGH ... Insufficient room for saving captured contents. Treat
# Line 873  for (;;) Line 975  for (;;)
975      /* VVVVVVVVVVVVVVVVVVVVVVVVV */      /* VVVVVVVVVVVVVVVVVVVVVVVVV */
976      /* VVVVVVVVVVVVVVVVVVVVVVVVV */      /* VVVVVVVVVVVVVVVVVVVVVVVVV */
977    
978      /* Non-capturing bracket, except for possessive with unlimited repeat. Loop      /* Non-capturing or atomic group, except for possessive with unlimited
979      for all the alternatives. When we get to the final alternative within the      repeat and ONCE group with no captures. Loop for all the alternatives.
980      brackets, we used to return the result of a recursive call to match()  
981      whatever happened so it was possible to reduce stack usage by turning this      When we get to the final alternative within the brackets, we used to return
982      into a tail recursion, except in the case of a possibly empty group.      the result of a recursive call to match() whatever happened so it was
983      However, now that there is the possiblity of (*THEN) occurring in the final      possible to reduce stack usage by turning this into a tail recursion,
984      alternative, this optimization is no longer possible. */      except in the case of a possibly empty group. However, now that there is
985        the possiblity of (*THEN) occurring in the final alternative, this
986        optimization is no longer always possible.
987    
988        We can optimize if we know there are no (*THEN)s in the pattern; at present
989        this is the best that can be done.
990    
991        MATCH_ONCE is returned when the end of an atomic group is successfully
992        reached, but subsequent matching fails. It passes back up the tree (causing
993        captured values to be reset) until the original atomic group level is
994        reached. This is tested by comparing md->once_target with the start of the
995        group. At this point, the return is converted into MATCH_NOMATCH so that
996        previous backup points can be taken. */
997    
998        case OP_ONCE:
999      case OP_BRA:      case OP_BRA:
1000      case OP_SBRA:      case OP_SBRA:
1001      DPRINTF(("start non-capturing bracket\n"));      DPRINTF(("start non-capturing bracket\n"));
1002    
1003      for (;;)      for (;;)
1004        {        {
1005        if (op >= OP_SBRA) md->match_function_type = MATCH_CBEGROUP;        if (op >= OP_SBRA || op == OP_ONCE) md->match_function_type = MATCH_CBEGROUP;
1006        RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md, eptrb,  
1007          /* If this is not a possibly empty group, and there are no (*THEN)s in
1008          the pattern, and this is the final alternative, optimize as described
1009          above. */
1010    
1011          else if (!md->hasthen && ecode[GET(ecode, 1)] != OP_ALT)
1012            {
1013            ecode += PRIV(OP_lengths)[*ecode];
1014            goto TAIL_RECURSE;
1015            }
1016    
1017          /* In all other cases, we have to make another call to match(). */
1018    
1019          RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md, eptrb,
1020          RM2);          RM2);
1021        if (rrc != MATCH_NOMATCH &&  
1022            (rrc != MATCH_THEN || md->start_match_ptr != ecode))        /* See comment in the code for capturing groups above about handling
1023          THEN. */
1024    
1025          if (rrc == MATCH_THEN)
1026            {
1027            next = ecode + GET(ecode,1);
1028            if (md->start_match_ptr < next &&
1029                (*ecode == OP_ALT || *next == OP_ALT))
1030              rrc = MATCH_NOMATCH;
1031            }
1032    
1033          if (rrc != MATCH_NOMATCH)
1034            {
1035            if (rrc == MATCH_ONCE)
1036              {
1037              const pcre_uchar *scode = ecode;
1038              if (*scode != OP_ONCE)           /* If not at start, find it */
1039                {
1040                while (*scode == OP_ALT) scode += GET(scode, 1);
1041                scode -= GET(scode, 1);
1042                }
1043              if (md->once_target == scode) rrc = MATCH_NOMATCH;
1044              }
1045          RRETURN(rrc);          RRETURN(rrc);
1046            }
1047        ecode += GET(ecode, 1);        ecode += GET(ecode, 1);
1048        if (*ecode != OP_ALT) break;        if (*ecode != OP_ALT) break;
1049        }        }
1050    
     if (rrc != MATCH_THEN && md->mark == NULL) md->mark = markptr;  
1051      RRETURN(MATCH_NOMATCH);      RRETURN(MATCH_NOMATCH);
1052    
1053      /* Handle possessive capturing brackets with an unlimited repeat. We come      /* Handle possessive capturing brackets with an unlimited repeat. We come
1054      here from BRAZERO with allow_zero set TRUE. The offset_vector values are      here from BRAZERO with allow_zero set TRUE. The offset_vector values are
1055      handled similarly to the normal case above. However, the matching is      handled similarly to the normal case above. However, the matching is
1056      different. The end of these brackets will always be OP_KETRPOS, which      different. The end of these brackets will always be OP_KETRPOS, which
1057      returns MATCH_KETRPOS without going further in the pattern. By this means      returns MATCH_KETRPOS without going further in the pattern. By this means
1058      we can handle the group by iteration rather than recursion, thereby      we can handle the group by iteration rather than recursion, thereby
1059      reducing the amount of stack needed. */      reducing the amount of stack needed. */
1060    
1061      case OP_CBRAPOS:      case OP_CBRAPOS:
1062      case OP_SCBRAPOS:      case OP_SCBRAPOS:
1063      allow_zero = FALSE;      allow_zero = FALSE;
1064    
1065      POSSESSIVE_CAPTURE:      POSSESSIVE_CAPTURE:
1066      number = GET2(ecode, 1+LINK_SIZE);      number = GET2(ecode, 1+LINK_SIZE);
1067      offset = number << 1;      offset = number << 1;
# Line 925  for (;;) Line 1076  for (;;)
1076      if (offset < md->offset_max)      if (offset < md->offset_max)
1077        {        {
1078        matched_once = FALSE;        matched_once = FALSE;
1079        code_offset = ecode - md->start_code;        code_offset = (int)(ecode - md->start_code);
1080    
1081        save_offset1 = md->offset_vector[offset];        save_offset1 = md->offset_vector[offset];
1082        save_offset2 = md->offset_vector[offset+1];        save_offset2 = md->offset_vector[offset+1];
# Line 933  for (;;) Line 1084  for (;;)
1084        save_capture_last = md->capture_last;        save_capture_last = md->capture_last;
1085    
1086        DPRINTF(("saving %d %d %d\n", save_offset1, save_offset2, save_offset3));        DPRINTF(("saving %d %d %d\n", save_offset1, save_offset2, save_offset3));
1087    
1088        /* Each time round the loop, save the current subject position for use        /* Each time round the loop, save the current subject position for use
1089        when the group matches. For MATCH_MATCH, the group has matched, so we        when the group matches. For MATCH_MATCH, the group has matched, so we
1090        restart it with a new subject starting position, remembering that we had        restart it with a new subject starting position, remembering that we had
1091        at least one match. For MATCH_NOMATCH, carry on with the alternatives, as        at least one match. For MATCH_NOMATCH, carry on with the alternatives, as
1092        usual. If we haven't matched any alternatives in any iteration, check to        usual. If we haven't matched any alternatives in any iteration, check to
1093        see if a previous iteration matched. If so, the group has matched;        see if a previous iteration matched. If so, the group has matched;
1094        continue from afterwards. Otherwise it has failed; restore the previous        continue from afterwards. Otherwise it has failed; restore the previous
1095        capture values before returning NOMATCH. */        capture values before returning NOMATCH. */
1096    
1097        for (;;)        for (;;)
1098          {          {
1099          md->offset_vector[md->offset_end - number] =          md->offset_vector[md->offset_end - number] =
1100            (int)(eptr - md->start_subject);            (int)(eptr - md->start_subject);
1101          if (op >= OP_SBRA) md->match_function_type = MATCH_CBEGROUP;          if (op >= OP_SBRA) md->match_function_type = MATCH_CBEGROUP;
1102          RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,          RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md,
1103            eptrb, RM63);            eptrb, RM63);
1104          if (rrc == MATCH_KETRPOS)          if (rrc == MATCH_KETRPOS)
1105            {            {
1106            offset_top = md->end_offset_top;            offset_top = md->end_offset_top;
1107            eptr = md->end_match_ptr;            eptr = md->end_match_ptr;
1108            ecode = md->start_code + code_offset;            ecode = md->start_code + code_offset;
1109            save_capture_last = md->capture_last;            save_capture_last = md->capture_last;
1110            matched_once = TRUE;            matched_once = TRUE;
1111            continue;            continue;
1112            }            }
1113          if (rrc != MATCH_NOMATCH &&  
1114              (rrc != MATCH_THEN || md->start_match_ptr != ecode))          /* See comment in the code for capturing groups above about handling
1115            RRETURN(rrc);          THEN. */
1116    
1117            if (rrc == MATCH_THEN)
1118              {
1119              next = ecode + GET(ecode,1);
1120              if (md->start_match_ptr < next &&
1121                  (*ecode == OP_ALT || *next == OP_ALT))
1122                rrc = MATCH_NOMATCH;
1123              }
1124    
1125            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
1126          md->capture_last = save_capture_last;          md->capture_last = save_capture_last;
1127          ecode += GET(ecode, 1);          ecode += GET(ecode, 1);
1128          if (*ecode != OP_ALT) break;          if (*ecode != OP_ALT) break;
1129          }          }
1130    
1131        if (!matched_once)        if (!matched_once)
1132          {          {
1133          md->offset_vector[offset] = save_offset1;          md->offset_vector[offset] = save_offset1;
1134          md->offset_vector[offset+1] = save_offset2;          md->offset_vector[offset+1] = save_offset2;
1135          md->offset_vector[md->offset_end - number] = save_offset3;          md->offset_vector[md->offset_end - number] = save_offset3;
1136          }          }
1137    
       if (rrc != MATCH_THEN && md->mark == NULL) md->mark = markptr;  
1138        if (allow_zero || matched_once)        if (allow_zero || matched_once)
1139          {          {
1140          ecode += 1 + LINK_SIZE;          ecode += 1 + LINK_SIZE;
1141          break;          break;
1142          }          }
1143    
1144        RRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
1145        }        }
1146    
1147      /* FALL THROUGH ... Insufficient room for saving captured contents. Treat      /* FALL THROUGH ... Insufficient room for saving captured contents. Treat
1148      as a non-capturing bracket. */      as a non-capturing bracket. */
1149    
# Line 995  for (;;) Line 1155  for (;;)
1155      /* VVVVVVVVVVVVVVVVVVVVVVVVV */      /* VVVVVVVVVVVVVVVVVVVVVVVVV */
1156      /* VVVVVVVVVVVVVVVVVVVVVVVVV */      /* VVVVVVVVVVVVVVVVVVVVVVVVV */
1157    
1158      /* Non-capturing possessive bracket with unlimited repeat. We come here      /* Non-capturing possessive bracket with unlimited repeat. We come here
1159      from BRAZERO with allow_zero = TRUE. The code is similar to the above,      from BRAZERO with allow_zero = TRUE. The code is similar to the above,
1160      without the capturing complication. It is written out separately for speed      without the capturing complication. It is written out separately for speed
1161      and cleanliness. */      and cleanliness. */
1162    
1163      case OP_BRAPOS:      case OP_BRAPOS:
1164      case OP_SBRAPOS:      case OP_SBRAPOS:
1165      allow_zero = FALSE;      allow_zero = FALSE;
1166    
1167      POSSESSIVE_NON_CAPTURE:      POSSESSIVE_NON_CAPTURE:
1168      matched_once = FALSE;      matched_once = FALSE;
1169      code_offset = ecode - md->start_code;      code_offset = (int)(ecode - md->start_code);
1170    
1171      for (;;)      for (;;)
1172        {        {
1173        if (op >= OP_SBRA) md->match_function_type = MATCH_CBEGROUP;        if (op >= OP_SBRA) md->match_function_type = MATCH_CBEGROUP;
1174        RMATCH(eptr, ecode + _pcre_OP_lengths[*ecode], offset_top, md,        RMATCH(eptr, ecode + PRIV(OP_lengths)[*ecode], offset_top, md,
1175          eptrb, RM48);          eptrb, RM48);
1176        if (rrc == MATCH_KETRPOS)        if (rrc == MATCH_KETRPOS)
1177          {          {
1178            offset_top = md->end_offset_top;
1179          eptr = md->end_match_ptr;          eptr = md->end_match_ptr;
1180          ecode = md->start_code + code_offset;          ecode = md->start_code + code_offset;
1181          matched_once = TRUE;          matched_once = TRUE;
1182          continue;          continue;
1183          }          }
1184        if (rrc != MATCH_NOMATCH &&  
1185            (rrc != MATCH_THEN || md->start_match_ptr != ecode))        /* See comment in the code for capturing groups above about handling
1186          RRETURN(rrc);        THEN. */
1187    
1188          if (rrc == MATCH_THEN)
1189            {
1190            next = ecode + GET(ecode,1);
1191            if (md->start_match_ptr < next &&
1192                (*ecode == OP_ALT || *next == OP_ALT))
1193              rrc = MATCH_NOMATCH;
1194            }
1195    
1196          if (rrc != MATCH_NOMATCH) RRETURN(rrc);
1197        ecode += GET(ecode, 1);        ecode += GET(ecode, 1);
1198        if (*ecode != OP_ALT) break;        if (*ecode != OP_ALT) break;
1199        }        }
1200    
1201      if (matched_once || allow_zero)      if (matched_once || allow_zero)
1202        {        {
1203        ecode += 1 + LINK_SIZE;        ecode += 1 + LINK_SIZE;
1204        break;        break;
1205        }        }
1206      RRETURN(MATCH_NOMATCH);      RRETURN(MATCH_NOMATCH);
1207    
1208      /* Control never reaches here. */      /* Control never reaches here. */
# Line 1050  for (;;) Line 1221  for (;;)
1221    
1222      if (ecode[LINK_SIZE+1] == OP_CALLOUT)      if (ecode[LINK_SIZE+1] == OP_CALLOUT)
1223        {        {
1224        if (pcre_callout != NULL)        if (PUBL(callout) != NULL)
1225          {          {
1226          pcre_callout_block cb;          PUBL(callout_block) cb;
1227          cb.version          = 1;   /* Version 1 of the callout block */          cb.version          = 2;   /* Version 1 of the callout block */
1228          cb.callout_number   = ecode[LINK_SIZE+2];          cb.callout_number   = ecode[LINK_SIZE+2];
1229          cb.offset_vector    = md->offset_vector;          cb.offset_vector    = md->offset_vector;
1230    #ifdef COMPILE_PCRE8
1231          cb.subject          = (PCRE_SPTR)md->start_subject;          cb.subject          = (PCRE_SPTR)md->start_subject;
1232    #else
1233            cb.subject          = (PCRE_SPTR16)md->start_subject;
1234    #endif
1235          cb.subject_length   = (int)(md->end_subject - md->start_subject);          cb.subject_length   = (int)(md->end_subject - md->start_subject);
1236          cb.start_match      = (int)(mstart - md->start_subject);          cb.start_match      = (int)(mstart - md->start_subject);
1237          cb.current_position = (int)(eptr - md->start_subject);          cb.current_position = (int)(eptr - md->start_subject);
# Line 1065  for (;;) Line 1240  for (;;)
1240          cb.capture_top      = offset_top/2;          cb.capture_top      = offset_top/2;
1241          cb.capture_last     = md->capture_last;          cb.capture_last     = md->capture_last;
1242          cb.callout_data     = md->callout_data;          cb.callout_data     = md->callout_data;
1243          if ((rrc = (*pcre_callout)(&cb)) > 0) MRRETURN(MATCH_NOMATCH);          cb.mark             = md->nomatch_mark;
1244            if ((rrc = (*PUBL(callout))(&cb)) > 0) RRETURN(MATCH_NOMATCH);
1245          if (rrc < 0) RRETURN(rrc);          if (rrc < 0) RRETURN(rrc);
1246          }          }
1247        ecode += _pcre_OP_lengths[OP_CALLOUT];        ecode += PRIV(OP_lengths)[OP_CALLOUT];
1248        }        }
1249    
1250      condcode = ecode[LINK_SIZE+1];      condcode = ecode[LINK_SIZE+1];
# Line 1085  for (;;) Line 1261  for (;;)
1261        else        else
1262          {          {
1263          int recno = GET2(ecode, LINK_SIZE + 2);   /* Recursion group number*/          int recno = GET2(ecode, LINK_SIZE + 2);   /* Recursion group number*/
1264          condition =  (recno == RREF_ANY || recno == md->recursive->group_num);          condition = (recno == RREF_ANY || recno == md->recursive->group_num);
1265    
1266          /* If the test is for recursion into a specific subpattern, and it is          /* If the test is for recursion into a specific subpattern, and it is
1267          false, but the test was set up by name, scan the table to see if the          false, but the test was set up by name, scan the table to see if the
1268          name refers to any other numbers, and test them. The condition is true          name refers to any other numbers, and test them. The condition is true
1269          if any one is set. */          if any one is set. */
1270    
1271          if (!condition && condcode == OP_NRREF && recno != RREF_ANY)          if (!condition && condcode == OP_NRREF)
1272            {            {
1273            uschar *slotA = md->name_table;            pcre_uchar *slotA = md->name_table;
1274            for (i = 0; i < md->name_count; i++)            for (i = 0; i < md->name_count; i++)
1275              {              {
1276              if (GET2(slotA, 0) == recno) break;              if (GET2(slotA, 0) == recno) break;
# Line 1107  for (;;) Line 1283  for (;;)
1283    
1284            if (i < md->name_count)            if (i < md->name_count)
1285              {              {
1286              uschar *slotB = slotA;              pcre_uchar *slotB = slotA;
1287              while (slotB > md->name_table)              while (slotB > md->name_table)
1288                {                {
1289                slotB -= md->name_entry_size;                slotB -= md->name_entry_size;
1290                if (strcmp((char *)slotA + 2, (char *)slotB + 2) == 0)                if (STRCMP_UC_UC(slotA + IMM2_SIZE, slotB + IMM2_SIZE) == 0)
1291                  {                  {
1292                  condition = GET2(slotB, 0) == md->recursive->group_num;                  condition = GET2(slotB, 0) == md->recursive->group_num;
1293                  if (condition) break;                  if (condition) break;
# Line 1127  for (;;) Line 1303  for (;;)
1303                for (i++; i < md->name_count; i++)                for (i++; i < md->name_count; i++)
1304                  {                  {
1305                  slotB += md->name_entry_size;                  slotB += md->name_entry_size;
1306                  if (strcmp((char *)slotA + 2, (char *)slotB + 2) == 0)                  if (STRCMP_UC_UC(slotA + IMM2_SIZE, slotB + IMM2_SIZE) == 0)
1307                    {                    {
1308                    condition = GET2(slotB, 0) == md->recursive->group_num;                    condition = GET2(slotB, 0) == md->recursive->group_num;
1309                    if (condition) break;                    if (condition) break;
# Line 1140  for (;;) Line 1316  for (;;)
1316    
1317          /* Chose branch according to the condition */          /* Chose branch according to the condition */
1318    
1319          ecode += condition? 3 : GET(ecode, 1);          ecode += condition? 1 + IMM2_SIZE : GET(ecode, 1);
1320          }          }
1321        }        }
1322    
# Line 1157  for (;;) Line 1333  for (;;)
1333        if (!condition && condcode == OP_NCREF)        if (!condition && condcode == OP_NCREF)
1334          {          {
1335          int refno = offset >> 1;          int refno = offset >> 1;
1336          uschar *slotA = md->name_table;          pcre_uchar *slotA = md->name_table;
1337    
1338          for (i = 0; i < md->name_count; i++)          for (i = 0; i < md->name_count; i++)
1339            {            {
# Line 1171  for (;;) Line 1347  for (;;)
1347    
1348          if (i < md->name_count)          if (i < md->name_count)
1349            {            {
1350            uschar *slotB = slotA;            pcre_uchar *slotB = slotA;
1351            while (slotB > md->name_table)            while (slotB > md->name_table)
1352              {              {
1353              slotB -= md->name_entry_size;              slotB -= md->name_entry_size;
1354              if (strcmp((char *)slotA + 2, (char *)slotB + 2) == 0)              if (STRCMP_UC_UC(slotA + IMM2_SIZE, slotB + IMM2_SIZE) == 0)
1355                {                {
1356                offset = GET2(slotB, 0) << 1;                offset = GET2(slotB, 0) << 1;
1357                condition = offset < offset_top &&                condition = offset < offset_top &&
# Line 1193  for (;;) Line 1369  for (;;)
1369              for (i++; i < md->name_count; i++)              for (i++; i < md->name_count; i++)
1370                {                {
1371                slotB += md->name_entry_size;                slotB += md->name_entry_size;
1372                if (strcmp((char *)slotA + 2, (char *)slotB + 2) == 0)                if (STRCMP_UC_UC(slotA + IMM2_SIZE, slotB + IMM2_SIZE) == 0)
1373                  {                  {
1374                  offset = GET2(slotB, 0) << 1;                  offset = GET2(slotB, 0) << 1;
1375                  condition = offset < offset_top &&                  condition = offset < offset_top &&
# Line 1208  for (;;) Line 1384  for (;;)
1384    
1385        /* Chose branch according to the condition */        /* Chose branch according to the condition */
1386    
1387        ecode += condition? 3 : GET(ecode, 1);        ecode += condition? 1 + IMM2_SIZE : GET(ecode, 1);
1388        }        }
1389    
1390      else if (condcode == OP_DEF)     /* DEFINE - always false */      else if (condcode == OP_DEF)     /* DEFINE - always false */
# Line 1223  for (;;) Line 1399  for (;;)
1399    
1400      else      else
1401        {        {
1402        md->match_function_type = MATCH_CONDASSERT;        md->match_function_type = MATCH_CONDASSERT;
1403        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, NULL, RM3);        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, NULL, RM3);
1404        if (rrc == MATCH_MATCH)        if (rrc == MATCH_MATCH)
1405          {          {
1406            if (md->end_offset_top > offset_top)
1407              offset_top = md->end_offset_top;  /* Captures may have happened */
1408          condition = TRUE;          condition = TRUE;
1409          ecode += 1 + LINK_SIZE + GET(ecode, LINK_SIZE + 2);          ecode += 1 + LINK_SIZE + GET(ecode, LINK_SIZE + 2);
1410          while (*ecode == OP_ALT) ecode += GET(ecode, 1);          while (*ecode == OP_ALT) ecode += GET(ecode, 1);
1411          }          }
1412        else if (rrc != MATCH_NOMATCH &&  
1413                (rrc != MATCH_THEN || md->start_match_ptr != ecode))        /* PCRE doesn't allow the effect of (*THEN) to escape beyond an
1414          assertion; it is therefore treated as NOMATCH. */
1415    
1416          else if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN)
1417          {          {
1418          RRETURN(rrc);         /* Need braces because of following else */          RRETURN(rrc);         /* Need braces because of following else */
1419          }          }
# Line 1243  for (;;) Line 1424  for (;;)
1424          }          }
1425        }        }
1426    
1427      /* We are now at the branch that is to be obeyed. As there is only one,      /* We are now at the branch that is to be obeyed. As there is only one, can
1428      we used to use tail recursion to avoid using another stack frame, except      use tail recursion to avoid using another stack frame, except when there is
1429      when there was unlimited repeat of a possibly empty group. However, that      unlimited repeat of a possibly empty group. In the latter case, a recursive
1430      strategy no longer works because of the possibilty of (*THEN) being      call to match() is always required, unless the second alternative doesn't
1431      encountered in the branch. A recursive call to match() is always required,      exist, in which case we can just plough on. Note that, for compatibility
1432      unless the second alternative doesn't exist, in which case we can just      with Perl, the | in a conditional group is NOT treated as creating two
1433      plough on. */      alternatives. If a THEN is encountered in the branch, it propagates out to
1434        the enclosing alternative (unless nested in a deeper set of alternatives,
1435        of course). */
1436    
1437      if (condition || *ecode == OP_ALT)      if (condition || *ecode == OP_ALT)
1438        {        {
1439        if (op == OP_SCOND) md->match_function_type = MATCH_CBEGROUP;        if (op != OP_SCOND)
1440            {
1441            ecode += 1 + LINK_SIZE;
1442            goto TAIL_RECURSE;
1443            }
1444    
1445          md->match_function_type = MATCH_CBEGROUP;
1446        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, eptrb, RM49);        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, eptrb, RM49);
       if (rrc == MATCH_THEN && md->start_match_ptr == ecode)  
         rrc = MATCH_NOMATCH;  
1447        RRETURN(rrc);        RRETURN(rrc);
1448        }        }
1449      else                         /* Condition false & no alternative */  
1450         /* Condition false & no alternative; continue after the group. */
1451    
1452        else
1453        {        {
1454        ecode += 1 + LINK_SIZE;        ecode += 1 + LINK_SIZE;
1455        }        }
# Line 1286  for (;;) Line 1476  for (;;)
1476        md->offset_vector[offset+1] = (int)(eptr - md->start_subject);        md->offset_vector[offset+1] = (int)(eptr - md->start_subject);
1477        if (offset_top <= offset) offset_top = offset + 2;        if (offset_top <= offset) offset_top = offset + 2;
1478        }        }
1479      ecode += 3;      ecode += 1 + IMM2_SIZE;
1480      break;      break;
1481    
1482    
1483      /* End of the pattern, either real or forced. If we are in a recursion, we      /* End of the pattern, either real or forced. */
     should restore the offsets appropriately, and if it's a top-level  
     recursion, continue from after the call. */  
1484    
     case OP_ACCEPT:  
1485      case OP_END:      case OP_END:
1486      if (md->recursive != NULL)      case OP_ACCEPT:
1487        {      case OP_ASSERT_ACCEPT:
1488        recursion_info *rec = md->recursive;  
1489        md->recursive = rec->prevrec;      /* If we have matched an empty string, fail if not in an assertion and not
1490        memmove(md->offset_vector, rec->offset_save,      in a recursion if either PCRE_NOTEMPTY is set, or if PCRE_NOTEMPTY_ATSTART
1491          rec->saved_max * sizeof(int));      is set and we have matched at the start of the subject. In both cases,
1492        offset_top = rec->save_offset_top;      backtracking will then try other alternatives, if any. */
1493        if (rec->group_num == 0)  
1494          {      if (eptr == mstart && op != OP_ASSERT_ACCEPT &&
1495          ecode = rec->after_call;           md->recursive == NULL &&
1496          break;           (md->notempty ||
1497          }             (md->notempty_atstart &&
1498        }               mstart == md->start_subject + md->start_offset)))
1499          RRETURN(MATCH_NOMATCH);
     /* Otherwise, if we have matched an empty string, fail if PCRE_NOTEMPTY is  
     set, or if PCRE_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. */  
   
     else if (eptr == mstart &&  
         (md->notempty ||  
           (md->notempty_atstart &&  
             mstart == md->start_subject + md->start_offset)))  
       MRRETURN(MATCH_NOMATCH);  
1500    
1501      /* Otherwise, we have a match. */      /* Otherwise, we have a match. */
1502    
1503      md->end_match_ptr = eptr;           /* Record where we ended */      md->end_match_ptr = eptr;           /* Record where we ended */
1504      md->end_offset_top = offset_top;    /* and how many extracts were taken */      md->end_offset_top = offset_top;    /* and how many extracts were taken */
1505      md->start_match_ptr = mstart;       /* and the start (\K can modify) */      md->start_match_ptr = mstart;       /* and the start (\K can modify) */
1506    
1507      /* For some reason, the macros don't work properly if an expression is      /* For some reason, the macros don't work properly if an expression is
1508      given as the argument to MRRETURN when the heap is in use. */      given as the argument to RRETURN when the heap is in use. */
1509    
1510      rrc = (op == OP_END)? MATCH_MATCH : MATCH_ACCEPT;      rrc = (op == OP_END)? MATCH_MATCH : MATCH_ACCEPT;
1511      MRRETURN(rrc);      RRETURN(rrc);
1512    
1513      /* Assertion brackets. Check the alternative branches in turn - the      /* Assertion brackets. Check the alternative branches in turn - the
1514      matching won't pass the KET for an assertion. If any one branch matches,      matching won't pass the KET for an assertion. If any one branch matches,
1515      the assertion is true. Lookbehind assertions have an OP_REVERSE item at the      the assertion is true. Lookbehind assertions have an OP_REVERSE item at the
1516      start of each branch to move the current point backwards, so the code at      start of each branch to move the current point backwards, so the code at
1517      this level is identical to the lookahead case. When the assertion is part      this level is identical to the lookahead case. When the assertion is part
1518      of a condition, we want to return immediately afterwards. The caller of      of a condition, we want to return immediately afterwards. The caller of
1519      this incarnation of the match() function will have set MATCH_CONDASSERT in      this incarnation of the match() function will have set MATCH_CONDASSERT in
1520      md->match_function type, and one of these opcodes will be the first opcode      md->match_function type, and one of these opcodes will be the first opcode
1521      that is processed. We use a local variable that is preserved over calls to      that is processed. We use a local variable that is preserved over calls to
1522      match() to remember this case. */      match() to remember this case. */
1523    
1524      case OP_ASSERT:      case OP_ASSERT:
# Line 1351  for (;;) Line 1528  for (;;)
1528        condassert = TRUE;        condassert = TRUE;
1529        md->match_function_type = 0;        md->match_function_type = 0;
1530        }        }
1531      else condassert = FALSE;      else condassert = FALSE;
1532    
1533      do      do
1534        {        {
1535        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, NULL, RM4);        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, NULL, RM4);
# Line 1361  for (;;) Line 1538  for (;;)
1538          mstart = md->start_match_ptr;   /* In case \K reset it */          mstart = md->start_match_ptr;   /* In case \K reset it */
1539          break;          break;
1540          }          }
1541        if (rrc != MATCH_NOMATCH &&  
1542            (rrc != MATCH_THEN || md->start_match_ptr != ecode))        /* PCRE does not allow THEN to escape beyond an assertion; it is treated
1543          RRETURN(rrc);        as NOMATCH. */
1544    
1545          if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN) RRETURN(rrc);
1546        ecode += GET(ecode, 1);        ecode += GET(ecode, 1);
1547        }        }
1548      while (*ecode == OP_ALT);      while (*ecode == OP_ALT);
1549    
1550      if (*ecode == OP_KET) MRRETURN(MATCH_NOMATCH);      if (*ecode == OP_KET) RRETURN(MATCH_NOMATCH);
1551    
1552      /* If checking an assertion for a condition, return MATCH_MATCH. */      /* If checking an assertion for a condition, return MATCH_MATCH. */
1553    
# Line 1393  for (;;) Line 1572  for (;;)
1572        condassert = TRUE;        condassert = TRUE;
1573        md->match_function_type = 0;        md->match_function_type = 0;
1574        }        }
1575      else condassert = FALSE;      else condassert = FALSE;
1576    
1577      do      do
1578        {        {
1579        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, NULL, RM5);        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, NULL, RM5);
1580        if (rrc == MATCH_MATCH || rrc == MATCH_ACCEPT) MRRETURN(MATCH_NOMATCH);        if (rrc == MATCH_MATCH || rrc == MATCH_ACCEPT) RRETURN(MATCH_NOMATCH);
1581        if (rrc == MATCH_SKIP || rrc == MATCH_PRUNE || rrc == MATCH_COMMIT)        if (rrc == MATCH_SKIP || rrc == MATCH_PRUNE || rrc == MATCH_COMMIT)
1582          {          {
1583          do ecode += GET(ecode,1); while (*ecode == OP_ALT);          do ecode += GET(ecode,1); while (*ecode == OP_ALT);
1584          break;          break;
1585          }          }
1586        if (rrc != MATCH_NOMATCH &&  
1587            (rrc != MATCH_THEN || md->start_match_ptr != ecode))        /* PCRE does not allow THEN to escape beyond an assertion; it is treated
1588          RRETURN(rrc);        as NOMATCH. */
1589    
1590          if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN) RRETURN(rrc);
1591        ecode += GET(ecode,1);        ecode += GET(ecode,1);
1592        }        }
1593      while (*ecode == OP_ALT);      while (*ecode == OP_ALT);
1594    
1595      if (condassert) RRETURN(MATCH_MATCH);  /* Condition assertion */      if (condassert) RRETURN(MATCH_MATCH);  /* Condition assertion */
1596    
1597      ecode += 1 + LINK_SIZE;      ecode += 1 + LINK_SIZE;
1598      continue;      continue;
1599    
# Line 1422  for (;;) Line 1603  for (;;)
1603      back a number of characters, not bytes. */      back a number of characters, not bytes. */
1604    
1605      case OP_REVERSE:      case OP_REVERSE:
1606  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
1607      if (utf8)      if (utf)
1608        {        {
1609        i = GET(ecode, 1);        i = GET(ecode, 1);
1610        while (i-- > 0)        while (i-- > 0)
1611          {          {
1612          eptr--;          eptr--;
1613          if (eptr < md->start_subject) MRRETURN(MATCH_NOMATCH);          if (eptr < md->start_subject) RRETURN(MATCH_NOMATCH);
1614          BACKCHAR(eptr);          BACKCHAR(eptr);
1615          }          }
1616        }        }
# Line 1440  for (;;) Line 1621  for (;;)
1621    
1622        {        {
1623        eptr -= GET(ecode, 1);        eptr -= GET(ecode, 1);
1624        if (eptr < md->start_subject) MRRETURN(MATCH_NOMATCH);        if (eptr < md->start_subject) RRETURN(MATCH_NOMATCH);
1625        }        }
1626    
1627      /* Save the earliest consulted character, then skip to next op code */      /* Save the earliest consulted character, then skip to next op code */
# Line 1454  for (;;) Line 1635  for (;;)
1635      function is able to force a failure. */      function is able to force a failure. */
1636    
1637      case OP_CALLOUT:      case OP_CALLOUT:
1638      if (pcre_callout != NULL)      if (PUBL(callout) != NULL)
1639        {        {
1640        pcre_callout_block cb;        PUBL(callout_block) cb;
1641        cb.version          = 1;   /* Version 1 of the callout block */        cb.version          = 2;   /* Version 1 of the callout block */
1642        cb.callout_number   = ecode[1];        cb.callout_number   = ecode[1];
1643        cb.offset_vector    = md->offset_vector;        cb.offset_vector    = md->offset_vector;
1644    #ifdef COMPILE_PCRE8
1645        cb.subject          = (PCRE_SPTR)md->start_subject;        cb.subject          = (PCRE_SPTR)md->start_subject;
1646    #else
1647          cb.subject          = (PCRE_SPTR16)md->start_subject;
1648    #endif
1649        cb.subject_length   = (int)(md->end_subject - md->start_subject);        cb.subject_length   = (int)(md->end_subject - md->start_subject);
1650        cb.start_match      = (int)(mstart - md->start_subject);        cb.start_match      = (int)(mstart - md->start_subject);
1651        cb.current_position = (int)(eptr - md->start_subject);        cb.current_position = (int)(eptr - md->start_subject);
# Line 1469  for (;;) Line 1654  for (;;)
1654        cb.capture_top      = offset_top/2;        cb.capture_top      = offset_top/2;
1655        cb.capture_last     = md->capture_last;        cb.capture_last     = md->capture_last;
1656        cb.callout_data     = md->callout_data;        cb.callout_data     = md->callout_data;
1657        if ((rrc = (*pcre_callout)(&cb)) > 0) MRRETURN(MATCH_NOMATCH);        cb.mark             = md->nomatch_mark;
1658          if ((rrc = (*PUBL(callout))(&cb)) > 0) RRETURN(MATCH_NOMATCH);
1659        if (rrc < 0) RRETURN(rrc);        if (rrc < 0) RRETURN(rrc);
1660        }        }
1661      ecode += 2 + 2*LINK_SIZE;      ecode += 2 + 2*LINK_SIZE;
# Line 1479  for (;;) Line 1665  for (;;)
1665      offset data is the offset to the starting bracket from the start of the      offset data is the offset to the starting bracket from the start of the
1666      whole pattern. (This is so that it works from duplicated subpatterns.)      whole pattern. (This is so that it works from duplicated subpatterns.)
1667    
1668      If there are any capturing brackets started but not finished, we have to      The state of the capturing groups is preserved over recursion, and
1669      save their starting points and reinstate them after the recursion. However,      re-instated afterwards. We don't know how many are started and not yet
1670      we don't know how many such there are (offset_top records the completed      finished (offset_top records the completed total) so we just have to save
1671      total) so we just have to save all the potential data. There may be up to      all the potential data. There may be up to 65535 such values, which is too
1672      65535 such values, which is too large to put on the stack, but using malloc      large to put on the stack, but using malloc for small numbers seems
1673      for small numbers seems expensive. As a compromise, the stack is used when      expensive. As a compromise, the stack is used when there are no more than
1674      there are no more than REC_STACK_SAVE_MAX values to store; otherwise malloc      REC_STACK_SAVE_MAX values to store; otherwise malloc is used.
     is used. A problem is what to do if the malloc fails ... there is no way of  
     returning to the top level with an error. Save the top REC_STACK_SAVE_MAX  
     values on the stack, and accept that the rest may be wrong.  
1675    
1676      There are also other values that have to be saved. We use a chained      There are also other values that have to be saved. We use a chained
1677      sequence of blocks that actually live on the stack. Thanks to Robin Houston      sequence of blocks that actually live on the stack. Thanks to Robin Houston
1678      for the original version of this logic. */      for the original version of this logic. It has, however, been hacked around
1679        a lot, so he is not to blame for the current way it works. */
1680    
1681      case OP_RECURSE:      case OP_RECURSE:
1682        {        {
1683          recursion_info *ri;
1684          int recno;
1685    
1686        callpat = md->start_code + GET(ecode, 1);        callpat = md->start_code + GET(ecode, 1);
1687        new_recursive.group_num = (callpat == md->start_code)? 0 :        recno = (callpat == md->start_code)? 0 :
1688          GET2(callpat, 1 + LINK_SIZE);          GET2(callpat, 1 + LINK_SIZE);
1689    
1690          /* Check for repeating a recursion without advancing the subject pointer.
1691          This should catch convoluted mutual recursions. (Some simple cases are
1692          caught at compile time.) */
1693    
1694          for (ri = md->recursive; ri != NULL; ri = ri->prevrec)
1695            if (recno == ri->group_num && eptr == ri->subject_position)
1696              RRETURN(PCRE_ERROR_RECURSELOOP);
1697    
1698        /* Add to "recursing stack" */        /* Add to "recursing stack" */
1699    
1700          new_recursive.group_num = recno;
1701          new_recursive.subject_position = eptr;
1702        new_recursive.prevrec = md->recursive;        new_recursive.prevrec = md->recursive;
1703        md->recursive = &new_recursive;        md->recursive = &new_recursive;
1704    
1705        /* Find where to continue from afterwards */        /* Where to continue from afterwards */
1706    
1707        ecode += 1 + LINK_SIZE;        ecode += 1 + LINK_SIZE;
       new_recursive.after_call = ecode;  
1708    
1709        /* Now save the offset data. */        /* Now save the offset data */
1710    
1711        new_recursive.saved_max = md->offset_end;        new_recursive.saved_max = md->offset_end;
1712        if (new_recursive.saved_max <= REC_STACK_SAVE_MAX)        if (new_recursive.saved_max <= REC_STACK_SAVE_MAX)
# Line 1518  for (;;) Line 1714  for (;;)
1714        else        else
1715          {          {
1716          new_recursive.offset_save =          new_recursive.offset_save =
1717            (int *)(pcre_malloc)(new_recursive.saved_max * sizeof(int));            (int *)(PUBL(malloc))(new_recursive.saved_max * sizeof(int));
1718          if (new_recursive.offset_save == NULL) RRETURN(PCRE_ERROR_NOMEMORY);          if (new_recursive.offset_save == NULL) RRETURN(PCRE_ERROR_NOMEMORY);
1719          }          }
   
1720        memcpy(new_recursive.offset_save, md->offset_vector,        memcpy(new_recursive.offset_save, md->offset_vector,
1721              new_recursive.saved_max * sizeof(int));              new_recursive.saved_max * sizeof(int));
1722        new_recursive.save_offset_top = offset_top;  
1723          /* OK, now we can do the recursion. After processing each alternative,
1724        /* OK, now we can do the recursion. For each top-level alternative we        restore the offset data. If there were nested recursions, md->recursive
1725        restore the offset and recursion data. */        might be changed, so reset it before looping. */
1726    
1727        DPRINTF(("Recursing into group %d\n", new_recursive.group_num));        DPRINTF(("Recursing into group %d\n", new_recursive.group_num));
1728        cbegroup = (*callpat >= OP_SBRA);        cbegroup = (*callpat >= OP_SBRA);
1729        do        do
1730          {          {
1731          if (cbegroup) md->match_function_type = MATCH_CBEGROUP;          if (cbegroup) md->match_function_type = MATCH_CBEGROUP;
1732          RMATCH(eptr, callpat + _pcre_OP_lengths[*callpat], offset_top,          RMATCH(eptr, callpat + PRIV(OP_lengths)[*callpat], offset_top,
1733            md, eptrb, RM6);            md, eptrb, RM6);
1734            memcpy(md->offset_vector, new_recursive.offset_save,
1735                new_recursive.saved_max * sizeof(int));
1736            md->recursive = new_recursive.prevrec;
1737          if (rrc == MATCH_MATCH || rrc == MATCH_ACCEPT)          if (rrc == MATCH_MATCH || rrc == MATCH_ACCEPT)
1738            {            {
1739            DPRINTF(("Recursion matched\n"));            DPRINTF(("Recursion matched\n"));
           md->recursive = new_recursive.prevrec;  
1740            if (new_recursive.offset_save != stacksave)            if (new_recursive.offset_save != stacksave)
1741              (pcre_free)(new_recursive.offset_save);              (PUBL(free))(new_recursive.offset_save);
1742            MRRETURN(MATCH_MATCH);  
1743              /* Set where we got to in the subject, and reset the start in case
1744              it was changed by \K. This *is* propagated back out of a recursion,
1745              for Perl compatibility. */
1746    
1747              eptr = md->end_match_ptr;
1748              mstart = md->start_match_ptr;
1749              goto RECURSION_MATCHED;        /* Exit loop; end processing */
1750            }            }
1751          else if (rrc != MATCH_NOMATCH &&  
1752                  (rrc != MATCH_THEN || md->start_match_ptr != ecode))          /* PCRE does not allow THEN to escape beyond a recursion; it is treated
1753            as NOMATCH. */
1754    
1755            else if (rrc != MATCH_NOMATCH && rrc != MATCH_THEN)
1756            {            {
1757            DPRINTF(("Recursion gave error %d\n", rrc));            DPRINTF(("Recursion gave error %d\n", rrc));
1758            if (new_recursive.offset_save != stacksave)            if (new_recursive.offset_save != stacksave)
1759              (pcre_free)(new_recursive.offset_save);              (PUBL(free))(new_recursive.offset_save);
1760            RRETURN(rrc);            RRETURN(rrc);
1761            }            }
1762    
1763          md->recursive = &new_recursive;          md->recursive = &new_recursive;
         memcpy(md->offset_vector, new_recursive.offset_save,  
             new_recursive.saved_max * sizeof(int));  
1764          callpat += GET(callpat, 1);          callpat += GET(callpat, 1);
1765          }          }
1766        while (*callpat == OP_ALT);        while (*callpat == OP_ALT);
# Line 1563  for (;;) Line 1768  for (;;)
1768        DPRINTF(("Recursion didn't match\n"));        DPRINTF(("Recursion didn't match\n"));
1769        md->recursive = new_recursive.prevrec;        md->recursive = new_recursive.prevrec;
1770        if (new_recursive.offset_save != stacksave)        if (new_recursive.offset_save != stacksave)
1771          (pcre_free)(new_recursive.offset_save);          (PUBL(free))(new_recursive.offset_save);
1772        MRRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
       }  
     /* Control never reaches here */  
   
     /* "Once" brackets are like assertion brackets except that after a match,  
     the point in the subject string is not moved back. Thus there can never be  
     a move back into the brackets. Friedl calls these "atomic" subpatterns.  
     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:  
     prev = ecode;  
     saved_eptr = eptr;  
   
     do  
       {  
       RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, eptrb, RM7);  
       if (rrc == MATCH_MATCH)  /* Note: _not_ MATCH_ACCEPT */  
         {  
         mstart = md->start_match_ptr;  
         break;  
         }  
       if (rrc != MATCH_NOMATCH &&  
           (rrc != MATCH_THEN || md->start_match_ptr != ecode))  
         RRETURN(rrc);  
       ecode += GET(ecode,1);  
       }  
     while (*ecode == OP_ALT);  
   
     /* If hit the end of the group (which could be repeated), fail */  
   
     if (*ecode != OP_ONCE && *ecode != OP_ALT) RRETURN(MATCH_NOMATCH);  
   
     /* Continue as from after the assertion, updating the offsets high water  
     mark, since extracts may have been taken. */  
   
     do ecode += GET(ecode, 1); while (*ecode == OP_ALT);  
   
     offset_top = md->end_offset_top;  
     eptr = md->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 there is an options reset, it will get obeyed in the normal  
     course of events. */  
   
     if (*ecode == OP_KET || eptr == saved_eptr)  
       {  
       ecode += 1+LINK_SIZE;  
       break;  
1773        }        }
1774    
1775      /* The repeating kets try the rest of the pattern or restart from the      RECURSION_MATCHED:
1776      preceding bracket, in the appropriate order. The second "call" of match()      break;
     uses tail recursion, to avoid using another stack frame. */  
   
     if (*ecode == OP_KETRMIN)  
       {  
       RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, eptrb, RM8);  
       if (rrc != MATCH_NOMATCH) RRETURN(rrc);  
       ecode = prev;  
       goto TAIL_RECURSE;  
       }  
     else  /* OP_KETRMAX */  
       {  
       md->match_function_type = MATCH_CBEGROUP;  
       RMATCH(eptr, prev, offset_top, md, eptrb, RM9);  
       if (rrc != MATCH_NOMATCH) RRETURN(rrc);  
       ecode += 1 + LINK_SIZE;  
       goto TAIL_RECURSE;  
       }  
     /* Control never gets here */  
1777    
1778      /* An alternation is the end of a branch; scan along to find the end of the      /* An alternation is the end of a branch; scan along to find the end of the
1779      bracketed group and go to there. */      bracketed group and go to there. */
# Line 1652  for (;;) Line 1787  for (;;)
1787      at all - i.e. it could be ()* or ()? or even (){0} in the pattern. Brackets      at all - i.e. it could be ()* or ()? or even (){0} in the pattern. Brackets
1788      with fixed upper repeat limits are compiled as a number of copies, with the      with fixed upper repeat limits are compiled as a number of copies, with the
1789      optional ones preceded by BRAZERO or BRAMINZERO. */      optional ones preceded by BRAZERO or BRAMINZERO. */
1790    
1791      case OP_BRAZERO:      case OP_BRAZERO:
1792      next = ecode + 1;      next = ecode + 1;
1793      RMATCH(eptr, next, offset_top, md, eptrb, RM10);      RMATCH(eptr, next, offset_top, md, eptrb, RM10);
# Line 1660  for (;;) Line 1795  for (;;)
1795      do next += GET(next, 1); while (*next == OP_ALT);      do next += GET(next, 1); while (*next == OP_ALT);
1796      ecode = next + 1 + LINK_SIZE;      ecode = next + 1 + LINK_SIZE;
1797      break;      break;
1798    
1799      case OP_BRAMINZERO:      case OP_BRAMINZERO:
1800      next = ecode + 1;      next = ecode + 1;
1801      do next += GET(next, 1); while (*next == OP_ALT);      do next += GET(next, 1); while (*next == OP_ALT);
# Line 1674  for (;;) Line 1809  for (;;)
1809      do next += GET(next,1); while (*next == OP_ALT);      do next += GET(next,1); while (*next == OP_ALT);
1810      ecode = next + 1 + LINK_SIZE;      ecode = next + 1 + LINK_SIZE;
1811      break;      break;
1812    
1813      /* BRAPOSZERO occurs before a possessive bracket group. Don't do anything      /* BRAPOSZERO occurs before a possessive bracket group. Don't do anything
1814      here; just jump to the group, with allow_zero set TRUE. */      here; just jump to the group, with allow_zero set TRUE. */
1815    
1816      case OP_BRAPOSZERO:      case OP_BRAPOSZERO:
1817      op = *(++ecode);      op = *(++ecode);
1818      allow_zero = TRUE;      allow_zero = TRUE;
1819      if (op == OP_CBRAPOS || op == OP_SCBRAPOS) goto POSSESSIVE_CAPTURE;      if (op == OP_CBRAPOS || op == OP_SCBRAPOS) goto POSSESSIVE_CAPTURE;
1820        goto POSSESSIVE_NON_CAPTURE;        goto POSSESSIVE_NON_CAPTURE;
# Line 1689  for (;;) Line 1824  for (;;)
1824      case OP_KET:      case OP_KET:
1825      case OP_KETRMIN:      case OP_KETRMIN:
1826      case OP_KETRMAX:      case OP_KETRMAX:
1827      case OP_KETRPOS:      case OP_KETRPOS:
1828      prev = ecode - GET(ecode, 1);      prev = ecode - GET(ecode, 1);
1829    
1830      /* If this was a group that remembered the subject start, in order to break      /* If this was a group that remembered the subject start, in order to break
1831      infinite repeats of empty string matches, retrieve the subject start from      infinite repeats of empty string matches, retrieve the subject start from
1832      the chain. Otherwise, set it NULL. */      the chain. Otherwise, set it NULL. */
1833    
1834      if (*prev >= OP_SBRA)      if (*prev >= OP_SBRA || *prev == OP_ONCE)
1835        {        {
1836        saved_eptr = eptrb->epb_saved_eptr;   /* Value at start of group */        saved_eptr = eptrb->epb_saved_eptr;   /* Value at start of group */
1837        eptrb = eptrb->epb_prev;              /* Backup to previous group */        eptrb = eptrb->epb_prev;              /* Backup to previous group */
1838        }        }
1839      else saved_eptr = NULL;      else saved_eptr = NULL;
1840    
1841      /* If we are at the end of an assertion group or an atomic group, stop      /* If we are at the end of an assertion group or a non-capturing atomic
1842      matching and return MATCH_MATCH, but record the current high water mark for      group, stop matching and return MATCH_MATCH, but record the current high
1843      use by positive assertions. We also need to record the match start in case      water mark for use by positive assertions. We also need to record the match
1844      it was changed by \K. */      start in case it was changed by \K. */
1845    
1846      if (*prev == OP_ASSERT || *prev == OP_ASSERT_NOT ||      if ((*prev >= OP_ASSERT && *prev <= OP_ASSERTBACK_NOT) ||
1847          *prev == OP_ASSERTBACK || *prev == OP_ASSERTBACK_NOT ||           *prev == OP_ONCE_NC)
         *prev == OP_ONCE)  
1848        {        {
1849        md->end_match_ptr = eptr;      /* For ONCE */        md->end_match_ptr = eptr;      /* For ONCE_NC */
1850        md->end_offset_top = offset_top;        md->end_offset_top = offset_top;
1851        md->start_match_ptr = mstart;        md->start_match_ptr = mstart;
1852        MRRETURN(MATCH_MATCH);        RRETURN(MATCH_MATCH);         /* Sets md->mark */
1853        }        }
1854    
1855      /* For capturing groups we have to check the group number back at the start      /* For capturing groups we have to check the group number back at the start
1856      and if necessary complete handling an extraction by setting the offsets and      and if necessary complete handling an extraction by setting the offsets and
1857      bumping the high water mark. Note that whole-pattern recursion is coded as      bumping the high water mark. Whole-pattern recursion is coded as a recurse
1858      a recurse into group 0, so it won't be picked up here. Instead, we catch it      into group 0, so it won't be picked up here. Instead, we catch it when the
1859      when the OP_END is reached. Other recursion is handled here. */      OP_END is reached. Other recursion is handled here. We just have to record
1860        the current subject position and start match pointer and give a MATCH
1861        return. */
1862    
1863      if (*prev == OP_CBRA || *prev == OP_SCBRA ||      if (*prev == OP_CBRA || *prev == OP_SCBRA ||
1864          *prev == OP_CBRAPOS || *prev == OP_SCBRAPOS)          *prev == OP_CBRAPOS || *prev == OP_SCBRAPOS)
# Line 1735  for (;;) Line 1871  for (;;)
1871        printf("\n");        printf("\n");
1872  #endif  #endif
1873    
1874          /* Handle a recursively called group. */
1875    
1876          if (md->recursive != NULL && md->recursive->group_num == number)
1877            {
1878            md->end_match_ptr = eptr;
1879            md->start_match_ptr = mstart;
1880            RRETURN(MATCH_MATCH);
1881            }
1882    
1883          /* Deal with capturing */
1884    
1885        md->capture_last = number;        md->capture_last = number;
1886        if (offset >= md->offset_max) md->offset_overflow = TRUE; else        if (offset >= md->offset_max) md->offset_overflow = TRUE; else
1887          {          {
1888            /* If offset is greater than offset_top, it means that we are
1889            "skipping" a capturing group, and that group's offsets must be marked
1890            unset. In earlier versions of PCRE, all the offsets were unset at the
1891            start of matching, but this doesn't work because atomic groups and
1892            assertions can cause a value to be set that should later be unset.
1893            Example: matching /(?>(a))b|(a)c/ against "ac". This sets group 1 as
1894            part of the atomic group, but this is not on the final matching path,
1895            so must be unset when 2 is set. (If there is no group 2, there is no
1896            problem, because offset_top will then be 2, indicating no capture.) */
1897    
1898            if (offset > offset_top)
1899              {
1900              register int *iptr = md->offset_vector + offset_top;
1901              register int *iend = md->offset_vector + offset;
1902              while (iptr < iend) *iptr++ = -1;
1903              }
1904    
1905            /* Now make the extraction */
1906    
1907          md->offset_vector[offset] =          md->offset_vector[offset] =
1908            md->offset_vector[md->offset_end - number];            md->offset_vector[md->offset_end - number];
1909          md->offset_vector[offset+1] = (int)(eptr - md->start_subject);          md->offset_vector[offset+1] = (int)(eptr - md->start_subject);
1910          if (offset_top <= offset) offset_top = offset + 2;          if (offset_top <= offset) offset_top = offset + 2;
1911          }          }
   
       /* Handle a recursively called group. Restore the offsets  
       appropriately and continue from after the call. */  
   
       if (md->recursive != NULL && md->recursive->group_num == number)  
         {  
         recursion_info *rec = md->recursive;  
         DPRINTF(("Recursion (%d) succeeded - continuing\n", number));  
         md->recursive = rec->prevrec;  
         memcpy(md->offset_vector, rec->offset_save,  
           rec->saved_max * sizeof(int));  
         offset_top = rec->save_offset_top;  
         ecode = rec->after_call;  
         break;  
         }  
1912        }        }
1913    
1914      /* For a non-repeating ket, just continue at this level. This also      /* For an ordinary non-repeating ket, just continue at this level. This
1915      happens for a repeating ket if no characters were matched in the group.      also happens for a repeating ket if no characters were matched in the
1916      This is the forcible breaking of infinite loops as implemented in Perl      group. This is the forcible breaking of infinite loops as implemented in
1917      5.005. If there is an options reset, it will get obeyed in the normal      Perl 5.005. For a non-repeating atomic group that includes captures,
1918      course of events. */      establish a backup point by processing the rest of the pattern at a lower
1919        level. If this results in a NOMATCH return, pass MATCH_ONCE back to the
1920        original OP_ONCE level, thereby bypassing intermediate backup points, but
1921        resetting any captures that happened along the way. */
1922    
1923      if (*ecode == OP_KET || eptr == saved_eptr)      if (*ecode == OP_KET || eptr == saved_eptr)
1924        {        {
1925        ecode += 1 + LINK_SIZE;        if (*prev == OP_ONCE)
1926            {
1927            RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, eptrb, RM12);
1928            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
1929            md->once_target = prev;  /* Level at which to change to MATCH_NOMATCH */
1930            RRETURN(MATCH_ONCE);
1931            }
1932          ecode += 1 + LINK_SIZE;    /* Carry on at this level */
1933        break;        break;
1934        }        }
1935    
1936      /* OP_KETRPOS is a possessive repeating ket. Remember the current position,      /* OP_KETRPOS is a possessive repeating ket. Remember the current position,
1937      and return the MATCH_KETRPOS. This makes it possible to do the repeats one      and return the MATCH_KETRPOS. This makes it possible to do the repeats one
1938      at a time from the outer level, thus saving stack. */      at a time from the outer level, thus saving stack. */
1939    
1940      if (*ecode == OP_KETRPOS)      if (*ecode == OP_KETRPOS)
1941        {        {
1942        md->end_match_ptr = eptr;        md->end_match_ptr = eptr;
1943        md->end_offset_top = offset_top;        md->end_offset_top = offset_top;
1944        RRETURN(MATCH_KETRPOS);        RRETURN(MATCH_KETRPOS);
1945        }        }
1946    
1947      /* The normal repeating kets try the rest of the pattern or restart from      /* The normal repeating kets try the rest of the pattern or restart from
1948      the preceding bracket, in the appropriate order. In the second case, we can      the preceding bracket, in the appropriate order. In the second case, we can
1949      use tail recursion to avoid using another stack frame, unless we have an      use tail recursion to avoid using another stack frame, unless we have an
1950      unlimited repeat of a group that can match an empty string. */      an atomic group or an unlimited repeat of a group that can match an empty
1951        string. */
1952    
1953      if (*ecode == OP_KETRMIN)      if (*ecode == OP_KETRMIN)
1954        {        {
1955        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, eptrb, RM12);        RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, eptrb, RM7);
1956        if (rrc != MATCH_NOMATCH) RRETURN(rrc);        if (rrc != MATCH_NOMATCH) RRETURN(rrc);
1957          if (*prev == OP_ONCE)
1958            {
1959            RMATCH(eptr, prev, offset_top, md, eptrb, RM8);
1960            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
1961            md->once_target = prev;  /* Level at which to change to MATCH_NOMATCH */
1962            RRETURN(MATCH_ONCE);
1963            }
1964        if (*prev >= OP_SBRA)    /* Could match an empty string */        if (*prev >= OP_SBRA)    /* Could match an empty string */
1965          {          {
1966          md->match_function_type = MATCH_CBEGROUP;          md->match_function_type = MATCH_CBEGROUP;
1967          RMATCH(eptr, prev, offset_top, md, eptrb, RM50);          RMATCH(eptr, prev, offset_top, md, eptrb, RM50);
1968          RRETURN(rrc);          RRETURN(rrc);
1969          }          }
# Line 1803  for (;;) Line 1972  for (;;)
1972        }        }
1973      else  /* OP_KETRMAX */      else  /* OP_KETRMAX */
1974        {        {
1975        if (*prev >= OP_SBRA) md->match_function_type = MATCH_CBEGROUP;        if (*prev >= OP_SBRA) md->match_function_type = MATCH_CBEGROUP;
1976        RMATCH(eptr, prev, offset_top, md, eptrb, RM13);        RMATCH(eptr, prev, offset_top, md, eptrb, RM13);
1977          if (rrc == MATCH_ONCE && md->once_target == prev) rrc = MATCH_NOMATCH;
1978        if (rrc != MATCH_NOMATCH) RRETURN(rrc);        if (rrc != MATCH_NOMATCH) RRETURN(rrc);
1979          if (*prev == OP_ONCE)
1980            {
1981            RMATCH(eptr, ecode + 1 + LINK_SIZE, offset_top, md, eptrb, RM9);
1982            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
1983            md->once_target = prev;
1984            RRETURN(MATCH_ONCE);
1985            }
1986        ecode += 1 + LINK_SIZE;        ecode += 1 + LINK_SIZE;
1987        goto TAIL_RECURSE;        goto TAIL_RECURSE;
1988        }        }
# Line 1814  for (;;) Line 1991  for (;;)
1991      /* Not multiline mode: start of subject assertion, unless notbol. */      /* Not multiline mode: start of subject assertion, unless notbol. */
1992    
1993      case OP_CIRC:      case OP_CIRC:
1994      if (md->notbol && eptr == md->start_subject) MRRETURN(MATCH_NOMATCH);      if (md->notbol && eptr == md->start_subject) RRETURN(MATCH_NOMATCH);
1995    
1996      /* Start of subject assertion */      /* Start of subject assertion */
1997    
1998      case OP_SOD:      case OP_SOD:
1999      if (eptr != md->start_subject) MRRETURN(MATCH_NOMATCH);      if (eptr != md->start_subject) RRETURN(MATCH_NOMATCH);
2000      ecode++;      ecode++;
2001      break;      break;
2002    
2003      /* Multiline mode: start of subject unless notbol, or after any newline. */      /* Multiline mode: start of subject unless notbol, or after any newline. */
2004    
2005      case OP_CIRCM:      case OP_CIRCM:
2006      if (md->notbol && eptr == md->start_subject) MRRETURN(MATCH_NOMATCH);      if (md->notbol && eptr == md->start_subject) RRETURN(MATCH_NOMATCH);
2007      if (eptr != md->start_subject &&      if (eptr != md->start_subject &&
2008          (eptr == md->end_subject || !WAS_NEWLINE(eptr)))          (eptr == md->end_subject || !WAS_NEWLINE(eptr)))
2009        MRRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2010      ecode++;      ecode++;
2011      break;      break;
2012    
2013      /* Start of match assertion */      /* Start of match assertion */
2014    
2015      case OP_SOM:      case OP_SOM:
2016      if (eptr != md->start_subject + md->start_offset) MRRETURN(MATCH_NOMATCH);      if (eptr != md->start_subject + md->start_offset) RRETURN(MATCH_NOMATCH);
2017      ecode++;      ecode++;
2018      break;      break;
2019    
# Line 1852  for (;;) Line 2029  for (;;)
2029    
2030      case OP_DOLLM:      case OP_DOLLM:
2031      if (eptr < md->end_subject)      if (eptr < md->end_subject)
2032        { if (!IS_NEWLINE(eptr)) MRRETURN(MATCH_NOMATCH); }        { if (!IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH); }
2033      else      else
2034        {        {
2035        if (md->noteol) MRRETURN(MATCH_NOMATCH);        if (md->noteol) RRETURN(MATCH_NOMATCH);
2036        SCHECK_PARTIAL();        SCHECK_PARTIAL();
2037        }        }
2038      ecode++;      ecode++;
2039      break;      break;
2040    
2041      /* Not multiline mode: assert before a terminating newline or before end of      /* Not multiline mode: assert before a terminating newline or before end of
2042      subject unless noteol is set. */      subject unless noteol is set. */
2043    
2044      case OP_DOLL:      case OP_DOLL:
2045      if (md->noteol) MRRETURN(MATCH_NOMATCH);      if (md->noteol) RRETURN(MATCH_NOMATCH);
2046      if (!md->endonly) goto ASSERT_NL_OR_EOS;      if (!md->endonly) goto ASSERT_NL_OR_EOS;
2047    
2048      /* ... else fall through for endonly */      /* ... else fall through for endonly */
# Line 1873  for (;;) Line 2050  for (;;)
2050      /* End of subject assertion (\z) */      /* End of subject assertion (\z) */
2051    
2052      case OP_EOD:      case OP_EOD:
2053      if (eptr < md->end_subject) MRRETURN(MATCH_NOMATCH);      if (eptr < md->end_subject) RRETURN(MATCH_NOMATCH);
2054      SCHECK_PARTIAL();      SCHECK_PARTIAL();
2055      ecode++;      ecode++;
2056      break;      break;
# Line 1884  for (;;) Line 2061  for (;;)
2061      ASSERT_NL_OR_EOS:      ASSERT_NL_OR_EOS:
2062      if (eptr < md->end_subject &&      if (eptr < md->end_subject &&
2063          (!IS_NEWLINE(eptr) || eptr != md->end_subject - md->nllen))          (!IS_NEWLINE(eptr) || eptr != md->end_subject - md->nllen))
2064        MRRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2065    
2066      /* Either at end of string or \n before end. */      /* Either at end of string or \n before end. */
2067    
# Line 1903  for (;;) Line 2080  for (;;)
2080        be "non-word" characters. Remember the earliest consulted character for        be "non-word" characters. Remember the earliest consulted character for
2081        partial matching. */        partial matching. */
2082    
2083  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
2084        if (utf8)        if (utf)
2085          {          {
2086          /* Get status of previous character */          /* Get status of previous character */
2087    
2088          if (eptr == md->start_subject) prev_is_word = FALSE; else          if (eptr == md->start_subject) prev_is_word = FALSE; else
2089            {            {
2090            USPTR lastptr = eptr - 1;            PCRE_PUCHAR lastptr = eptr - 1;
2091            while((*lastptr & 0xc0) == 0x80) lastptr--;            BACKCHAR(lastptr);
2092            if (lastptr < md->start_used_ptr) md->start_used_ptr = lastptr;            if (lastptr < md->start_used_ptr) md->start_used_ptr = lastptr;
2093            GETCHAR(c, lastptr);            GETCHAR(c, lastptr);
2094  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
# Line 1976  for (;;) Line 2153  for (;;)
2153              }              }
2154            else            else
2155  #endif  #endif
2156            prev_is_word = ((md->ctypes[eptr[-1]] & ctype_word) != 0);            prev_is_word = MAX_255(eptr[-1])
2157                && ((md->ctypes[eptr[-1]] & ctype_word) != 0);
2158            }            }
2159    
2160          /* Get status of next character */          /* Get status of next character */
# Line 1999  for (;;) Line 2177  for (;;)
2177            }            }
2178          else          else
2179  #endif  #endif
2180          cur_is_word = ((md->ctypes[*eptr] & ctype_word) != 0);          cur_is_word = MAX_255(*eptr)
2181              && ((md->ctypes[*eptr] & ctype_word) != 0);
2182          }          }
2183    
2184        /* Now see if the situation is what we want */        /* Now see if the situation is what we want */
2185    
2186        if ((*ecode++ == OP_WORD_BOUNDARY)?        if ((*ecode++ == OP_WORD_BOUNDARY)?
2187             cur_is_word == prev_is_word : cur_is_word != prev_is_word)             cur_is_word == prev_is_word : cur_is_word != prev_is_word)
2188          MRRETURN(MATCH_NOMATCH);          RRETURN(MATCH_NOMATCH);
2189        }        }
2190      break;      break;
2191    
2192      /* Match a single character type; inline for speed */      /* Match a single character type; inline for speed */
2193    
2194      case OP_ANY:      case OP_ANY:
2195      if (IS_NEWLINE(eptr)) MRRETURN(MATCH_NOMATCH);      if (IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH);
2196      /* Fall through */      /* Fall through */
2197    
2198      case OP_ALLANY:      case OP_ALLANY:
2199      if (eptr++ >= md->end_subject)      if (eptr >= md->end_subject)   /* DO NOT merge the eptr++ here; it must */
2200        {        {                            /* not be updated before SCHECK_PARTIAL. */
2201        SCHECK_PARTIAL();        SCHECK_PARTIAL();
2202        MRRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2203        }        }
2204      if (utf8) while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;      eptr++;
2205    #ifdef SUPPORT_UTF
2206        if (utf) ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);
2207    #endif
2208      ecode++;      ecode++;
2209      break;      break;
2210    
# Line 2030  for (;;) Line 2212  for (;;)
2212      any byte, even newline, independent of the setting of PCRE_DOTALL. */      any byte, even newline, independent of the setting of PCRE_DOTALL. */
2213    
2214      case OP_ANYBYTE:      case OP_ANYBYTE:
2215      if (eptr++ >= md->end_subject)      if (eptr >= md->end_subject)   /* DO NOT merge the eptr++ here; it must */
2216        {        {                            /* not be updated before SCHECK_PARTIAL. */
2217        SCHECK_PARTIAL();        SCHECK_PARTIAL();
2218        MRRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2219        }        }
2220        eptr++;
2221      ecode++;      ecode++;
2222      break;      break;
2223    
# Line 2042  for (;;) Line 2225  for (;;)
2225      if (eptr >= md->end_subject)      if (eptr >= md->end_subject)
2226        {        {
2227        SCHECK_PARTIAL();        SCHECK_PARTIAL();
2228        MRRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2229        }        }
2230      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2231      if (      if (
2232  #ifdef SUPPORT_UTF8  #if defined SUPPORT_UTF || !(defined COMPILE_PCRE8)
2233         c < 256 &&         c < 256 &&
2234  #endif  #endif
2235         (md->ctypes[c] & ctype_digit) != 0         (md->ctypes[c] & ctype_digit) != 0
2236         )         )
2237        MRRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2238      ecode++;      ecode++;
2239      break;      break;
2240    
# Line 2059  for (;;) Line 2242  for (;;)
2242      if (eptr >= md->end_subject)      if (eptr >= md->end_subject)
2243        {        {
2244        SCHECK_PARTIAL();        SCHECK_PARTIAL();
2245        MRRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2246        }        }
2247      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2248      if (      if (
2249  #ifdef SUPPORT_UTF8  #if defined SUPPORT_UTF || !(defined COMPILE_PCRE8)
2250         c >= 256 ||         c > 255 ||
2251  #endif  #endif
2252         (md->ctypes[c] & ctype_digit) == 0         (md->ctypes[c] & ctype_digit) == 0
2253         )         )
2254        MRRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2255      ecode++;      ecode++;
2256      break;      break;
2257    
# Line 2076  for (;;) Line 2259  for (;;)
2259      if (eptr >= md->end_subject)      if (eptr >= md->end_subject)
2260        {        {
2261        SCHECK_PARTIAL();        SCHECK_PARTIAL();
2262        MRRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2263        }        }
2264      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2265      if (      if (
2266  #ifdef SUPPORT_UTF8  #if defined SUPPORT_UTF || !(defined COMPILE_PCRE8)
2267         c < 256 &&         c < 256 &&
2268  #endif  #endif
2269         (md->ctypes[c] & ctype_space) != 0         (md->ctypes[c] & ctype_space) != 0
2270         )         )
2271        MRRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2272      ecode++;      ecode++;
2273      break;      break;
2274    
# Line 2093  for (;;) Line 2276  for (;;)
2276      if (eptr >= md->end_subject)      if (eptr >= md->end_subject)
2277        {        {
2278        SCHECK_PARTIAL();        SCHECK_PARTIAL();
2279        MRRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2280        }        }
2281      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2282      if (      if (
2283  #ifdef SUPPORT_UTF8  #if defined SUPPORT_UTF || !(defined COMPILE_PCRE8)
2284         c >= 256 ||         c > 255 ||
2285  #endif  #endif
2286         (md->ctypes[c] & ctype_space) == 0         (md->ctypes[c] & ctype_space) == 0
2287         )         )
2288        MRRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2289      ecode++;      ecode++;
2290      break;      break;
2291    
# Line 2110  for (;;) Line 2293  for (;;)
2293      if (eptr >= md->end_subject)      if (eptr >= md->end_subject)
2294        {        {
2295        SCHECK_PARTIAL();        SCHECK_PARTIAL();
2296        MRRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2297        }        }
2298      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2299      if (      if (
2300  #ifdef SUPPORT_UTF8  #if defined SUPPORT_UTF || !(defined COMPILE_PCRE8)
2301         c < 256 &&         c < 256 &&
2302  #endif  #endif
2303         (md->ctypes[c] & ctype_word) != 0         (md->ctypes[c] & ctype_word) != 0
2304         )         )
2305        MRRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2306      ecode++;      ecode++;
2307      break;      break;
2308    
# Line 2127  for (;;) Line 2310  for (;;)
2310      if (eptr >= md->end_subject)      if (eptr >= md->end_subject)
2311        {        {
2312        SCHECK_PARTIAL();        SCHECK_PARTIAL();
2313        MRRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2314        }        }
2315      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2316      if (      if (
2317  #ifdef SUPPORT_UTF8  #if defined SUPPORT_UTF || !(defined COMPILE_PCRE8)
2318         c >= 256 ||         c > 255 ||
2319  #endif  #endif
2320         (md->ctypes[c] & ctype_word) == 0         (md->ctypes[c] & ctype_word) == 0
2321         )         )
2322        MRRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2323      ecode++;      ecode++;
2324      break;      break;
2325    
# Line 2144  for (;;) Line 2327  for (;;)
2327      if (eptr >= md->end_subject)      if (eptr >= md->end_subject)
2328        {        {
2329        SCHECK_PARTIAL();        SCHECK_PARTIAL();
2330        MRRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2331        }        }
2332      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2333      switch(c)      switch(c)
2334        {        {
2335        default: MRRETURN(MATCH_NOMATCH);        default: RRETURN(MATCH_NOMATCH);
2336    
2337        case 0x000d:        case 0x000d:
2338        if (eptr < md->end_subject && *eptr == 0x0a) eptr++;        if (eptr < md->end_subject && *eptr == 0x0a) eptr++;
2339        break;        break;
# Line 2163  for (;;) Line 2346  for (;;)
2346        case 0x0085:        case 0x0085:
2347        case 0x2028:        case 0x2028:
2348        case 0x2029:        case 0x2029:
2349        if (md->bsr_anycrlf) MRRETURN(MATCH_NOMATCH);        if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);
2350        break;        break;
2351        }        }
2352      ecode++;      ecode++;
# Line 2173  for (;;) Line 2356  for (;;)
2356      if (eptr >= md->end_subject)      if (eptr >= md->end_subject)
2357        {        {
2358        SCHECK_PARTIAL();        SCHECK_PARTIAL();
2359        MRRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2360        }        }
2361      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2362      switch(c)      switch(c)
# Line 2198  for (;;) Line 2381  for (;;)
2381        case 0x202f:    /* NARROW NO-BREAK SPACE */        case 0x202f:    /* NARROW NO-BREAK SPACE */
2382        case 0x205f:    /* MEDIUM MATHEMATICAL SPACE */        case 0x205f:    /* MEDIUM MATHEMATICAL SPACE */
2383        case 0x3000:    /* IDEOGRAPHIC SPACE */        case 0x3000:    /* IDEOGRAPHIC SPACE */
2384        MRRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2385        }        }
2386      ecode++;      ecode++;
2387      break;      break;
# Line 2207  for (;;) Line 2390  for (;;)
2390      if (eptr >= md->end_subject)      if (eptr >= md->end_subject)
2391        {        {
2392        SCHECK_PARTIAL();        SCHECK_PARTIAL();
2393        MRRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2394        }        }
2395      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2396      switch(c)      switch(c)
2397        {        {
2398        default: MRRETURN(MATCH_NOMATCH);        default: RRETURN(MATCH_NOMATCH);
2399        case 0x09:      /* HT */        case 0x09:      /* HT */
2400        case 0x20:      /* SPACE */        case 0x20:      /* SPACE */
2401        case 0xa0:      /* NBSP */        case 0xa0:      /* NBSP */
# Line 2241  for (;;) Line 2424  for (;;)
2424      if (eptr >= md->end_subject)      if (eptr >= md->end_subject)
2425        {        {
2426        SCHECK_PARTIAL();        SCHECK_PARTIAL();
2427        MRRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2428        }        }
2429      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2430      switch(c)      switch(c)
# Line 2254  for (;;) Line 2437  for (;;)
2437        case 0x85:      /* NEL */        case 0x85:      /* NEL */
2438        case 0x2028:    /* LINE SEPARATOR */        case 0x2028:    /* LINE SEPARATOR */
2439        case 0x2029:    /* PARAGRAPH SEPARATOR */        case 0x2029:    /* PARAGRAPH SEPARATOR */
2440        MRRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2441        }        }
2442      ecode++;      ecode++;
2443      break;      break;
# Line 2263  for (;;) Line 2446  for (;;)
2446      if (eptr >= md->end_subject)      if (eptr >= md->end_subject)
2447        {        {
2448        SCHECK_PARTIAL();        SCHECK_PARTIAL();
2449        MRRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2450        }        }
2451      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2452      switch(c)      switch(c)
2453        {        {
2454        default: MRRETURN(MATCH_NOMATCH);        default: RRETURN(MATCH_NOMATCH);
2455        case 0x0a:      /* LF */        case 0x0a:      /* LF */
2456        case 0x0b:      /* VT */        case 0x0b:      /* VT */
2457        case 0x0c:      /* FF */        case 0x0c:      /* FF */
# Line 2290  for (;;) Line 2473  for (;;)
2473      if (eptr >= md->end_subject)      if (eptr >= md->end_subject)
2474        {        {
2475        SCHECK_PARTIAL();        SCHECK_PARTIAL();
2476        MRRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2477        }        }
2478      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2479        {        {
# Line 2299  for (;;) Line 2482  for (;;)
2482        switch(ecode[1])        switch(ecode[1])
2483          {          {
2484          case PT_ANY:          case PT_ANY:
2485          if (op == OP_NOTPROP) MRRETURN(MATCH_NOMATCH);          if (op == OP_NOTPROP) RRETURN(MATCH_NOMATCH);
2486          break;          break;
2487    
2488          case PT_LAMP:          case PT_LAMP:
2489          if ((prop->chartype == ucp_Lu ||          if ((prop->chartype == ucp_Lu ||
2490               prop->chartype == ucp_Ll ||               prop->chartype == ucp_Ll ||
2491               prop->chartype == ucp_Lt) == (op == OP_NOTPROP))               prop->chartype == ucp_Lt) == (op == OP_NOTPROP))
2492            MRRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
2493          break;          break;
2494    
2495          case PT_GC:          case PT_GC:
2496          if ((ecode[2] != _pcre_ucp_gentype[prop->chartype]) == (op == OP_PROP))          if ((ecode[2] != PRIV(ucp_gentype)[prop->chartype]) == (op == OP_PROP))
2497            MRRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
2498          break;          break;
2499    
2500          case PT_PC:          case PT_PC:
2501          if ((ecode[2] != prop->chartype) == (op == OP_PROP))          if ((ecode[2] != prop->chartype) == (op == OP_PROP))
2502            MRRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
2503          break;          break;
2504    
2505          case PT_SC:          case PT_SC:
2506          if ((ecode[2] != prop->script) == (op == OP_PROP))          if ((ecode[2] != prop->script) == (op == OP_PROP))
2507            MRRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
2508          break;          break;
2509    
2510          /* These are specials */          /* These are specials */
2511    
2512          case PT_ALNUM:          case PT_ALNUM:
2513          if ((_pcre_ucp_gentype[prop->chartype] == ucp_L ||          if ((PRIV(ucp_gentype)[prop->chartype] == ucp_L ||
2514               _pcre_ucp_gentype[prop->chartype] == ucp_N) == (op == OP_NOTPROP))               PRIV(ucp_gentype)[prop->chartype] == ucp_N) == (op == OP_NOTPROP))
2515            MRRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
2516          break;          break;
2517    
2518          case PT_SPACE:    /* Perl space */          case PT_SPACE:    /* Perl space */
2519          if ((_pcre_ucp_gentype[prop->chartype] == ucp_Z ||          if ((PRIV(ucp_gentype)[prop->chartype] == ucp_Z ||
2520               c == CHAR_HT || c == CHAR_NL || c == CHAR_FF || c == CHAR_CR)               c == CHAR_HT || c == CHAR_NL || c == CHAR_FF || c == CHAR_CR)
2521                 == (op == OP_NOTPROP))                 == (op == OP_NOTPROP))
2522            MRRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
2523          break;          break;
2524    
2525          case PT_PXSPACE:  /* POSIX space */          case PT_PXSPACE:  /* POSIX space */
2526          if ((_pcre_ucp_gentype[prop->chartype] == ucp_Z ||          if ((PRIV(ucp_gentype)[prop->chartype] == ucp_Z ||
2527               c == CHAR_HT || c == CHAR_NL || c == CHAR_VT ||               c == CHAR_HT || c == CHAR_NL || c == CHAR_VT ||
2528               c == CHAR_FF || c == CHAR_CR)               c == CHAR_FF || c == CHAR_CR)
2529                 == (op == OP_NOTPROP))                 == (op == OP_NOTPROP))
2530            MRRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
2531          break;          break;
2532    
2533          case PT_WORD:          case PT_WORD:
2534          if ((_pcre_ucp_gentype[prop->chartype] == ucp_L ||          if ((PRIV(ucp_gentype)[prop->chartype] == ucp_L ||
2535               _pcre_ucp_gentype[prop->chartype] == ucp_N ||               PRIV(ucp_gentype)[prop->chartype] == ucp_N ||
2536               c == CHAR_UNDERSCORE) == (op == OP_NOTPROP))               c == CHAR_UNDERSCORE) == (op == OP_NOTPROP))
2537            MRRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
2538          break;          break;
2539    
2540          /* This should never occur */          /* This should never occur */
# Line 2371  for (;;) Line 2554  for (;;)
2554      if (eptr >= md->end_subject)      if (eptr >= md->end_subject)
2555        {        {
2556        SCHECK_PARTIAL();        SCHECK_PARTIAL();
2557        MRRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2558        }        }
2559      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
2560        if (UCD_CATEGORY(c) == ucp_M) RRETURN(MATCH_NOMATCH);
2561        while (eptr < md->end_subject)
2562        {        {
2563        int category = UCD_CATEGORY(c);        int len = 1;
2564        if (category == ucp_M) MRRETURN(MATCH_NOMATCH);        if (!utf) c = *eptr; else { GETCHARLEN(c, eptr, len); }
2565        while (eptr < md->end_subject)        if (UCD_CATEGORY(c) != ucp_M) break;
2566          {        eptr += len;
         int len = 1;  
         if (!utf8) c = *eptr; else  
           {  
           GETCHARLEN(c, eptr, len);  
           }  
         category = UCD_CATEGORY(c);  
         if (category != ucp_M) break;  
         eptr += len;  
         }  
2567        }        }
2568      ecode++;      ecode++;
2569      break;      break;
# Line 2403  for (;;) Line 2579  for (;;)
2579      loops). */      loops). */
2580    
2581      case OP_REF:      case OP_REF:
2582      case OP_REFI:      case OP_REFI:
2583      caseless = op == OP_REFI;      caseless = op == OP_REFI;
2584      offset = GET2(ecode, 1) << 1;               /* Doubled ref number */      offset = GET2(ecode, 1) << 1;               /* Doubled ref number */
2585      ecode += 3;      ecode += 1 + IMM2_SIZE;
2586    
2587      /* If the reference is unset, there are two possibilities:      /* If the reference is unset, there are two possibilities:
2588    
# Line 2446  for (;;) Line 2622  for (;;)
2622        case OP_CRMINRANGE:        case OP_CRMINRANGE:
2623        minimize = (*ecode == OP_CRMINRANGE);        minimize = (*ecode == OP_CRMINRANGE);
2624        min = GET2(ecode, 1);        min = GET2(ecode, 1);
2625        max = GET2(ecode, 3);        max = GET2(ecode, 1 + IMM2_SIZE);
2626        if (max == 0) max = INT_MAX;        if (max == 0) max = INT_MAX;
2627        ecode += 5;        ecode += 1 + 2 * IMM2_SIZE;
2628        break;        break;
2629    
2630        default:               /* No repeat follows */        default:               /* No repeat follows */
2631        if ((length = match_ref(offset, eptr, length, md, caseless)) < 0)        if ((length = match_ref(offset, eptr, length, md, caseless)) < 0)
2632          {          {
2633          CHECK_PARTIAL();          CHECK_PARTIAL();
2634          MRRETURN(MATCH_NOMATCH);          RRETURN(MATCH_NOMATCH);
2635          }          }
2636        eptr += length;        eptr += length;
2637        continue;              /* With the main loop */        continue;              /* With the main loop */
2638        }        }
2639    
2640      /* Handle repeated back references. If the length of the reference is      /* Handle repeated back references. If the length of the reference is
2641      zero, just continue with the main loop. */      zero, just continue with the main loop. If the length is negative, it
2642        means the reference is unset in non-Java-compatible mode. If the minimum is
2643        zero, we can continue at the same level without recursion. For any other
2644        minimum, carrying on will result in NOMATCH. */
2645    
2646      if (length == 0) continue;      if (length == 0) continue;
2647        if (length < 0 && min == 0) continue;
2648    
2649      /* First, ensure the minimum number of matches are present. We get back      /* First, ensure the minimum number of matches are present. We get back
2650      the length of the reference string explicitly rather than passing the      the length of the reference string explicitly rather than passing the
# Line 2472  for (;;) Line 2652  for (;;)
2652    
2653      for (i = 1; i <= min; i++)      for (i = 1; i <= min; i++)
2654        {        {
2655        int slength;        int slength;
2656        if ((slength = match_ref(offset, eptr, length, md, caseless)) < 0)        if ((slength = match_ref(offset, eptr, length, md, caseless)) < 0)
2657          {          {
2658          CHECK_PARTIAL();          CHECK_PARTIAL();
2659          MRRETURN(MATCH_NOMATCH);          RRETURN(MATCH_NOMATCH);
2660          }          }
2661        eptr += slength;        eptr += slength;
2662        }        }
# Line 2492  for (;;) Line 2672  for (;;)
2672        {        {
2673        for (fi = min;; fi++)        for (fi = min;; fi++)
2674          {          {
2675          int slength;          int slength;
2676          RMATCH(eptr, ecode, offset_top, md, eptrb, RM14);          RMATCH(eptr, ecode, offset_top, md, eptrb, RM14);
2677          if (rrc != MATCH_NOMATCH) RRETURN(rrc);          if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2678          if (fi >= max) MRRETURN(MATCH_NOMATCH);          if (fi >= max) RRETURN(MATCH_NOMATCH);
2679          if ((slength = match_ref(offset, eptr, length, md, caseless)) < 0)          if ((slength = match_ref(offset, eptr, length, md, caseless)) < 0)
2680            {            {
2681            CHECK_PARTIAL();            CHECK_PARTIAL();
2682            MRRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
2683            }            }
2684          eptr += slength;          eptr += slength;
2685          }          }
# Line 2513  for (;;) Line 2693  for (;;)
2693        pp = eptr;        pp = eptr;
2694        for (i = min; i < max; i++)        for (i = min; i < max; i++)
2695          {          {
2696          int slength;          int slength;
2697          if ((slength = match_ref(offset, eptr, length, md, caseless)) < 0)          if ((slength = match_ref(offset, eptr, length, md, caseless)) < 0)
2698            {            {
2699            CHECK_PARTIAL();            CHECK_PARTIAL();
# Line 2527  for (;;) Line 2707  for (;;)
2707          if (rrc != MATCH_NOMATCH) RRETURN(rrc);          if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2708          eptr -= length;          eptr -= length;
2709          }          }
2710        MRRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
2711        }        }
2712      /* Control never gets here */      /* Control never gets here */
2713    
# Line 2545  for (;;) Line 2725  for (;;)
2725      case OP_NCLASS:      case OP_NCLASS:
2726      case OP_CLASS:      case OP_CLASS:
2727        {        {
2728          /* The data variable is saved across frames, so the byte map needs to
2729          be stored there. */
2730    #define BYTE_MAP ((pcre_uint8 *)data)
2731        data = ecode + 1;                /* Save for matching */        data = ecode + 1;                /* Save for matching */
2732        ecode += 33;                     /* Advance past the item */        ecode += 1 + (32 / sizeof(pcre_uchar)); /* Advance past the item */
2733    
2734        switch (*ecode)        switch (*ecode)
2735          {          {
# Line 2567  for (;;) Line 2750  for (;;)
2750          case OP_CRMINRANGE:          case OP_CRMINRANGE:
2751          minimize = (*ecode == OP_CRMINRANGE);          minimize = (*ecode == OP_CRMINRANGE);
2752          min = GET2(ecode, 1);          min = GET2(ecode, 1);
2753          max = GET2(ecode, 3);          max = GET2(ecode, 1 + IMM2_SIZE);
2754          if (max == 0) max = INT_MAX;          if (max == 0) max = INT_MAX;
2755          ecode += 5;          ecode += 1 + 2 * IMM2_SIZE;
2756          break;          break;
2757    
2758          default:               /* No repeat follows */          default:               /* No repeat follows */
# Line 2579  for (;;) Line 2762  for (;;)
2762    
2763        /* First, ensure the minimum number of matches are present. */        /* First, ensure the minimum number of matches are present. */
2764    
2765  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
2766        /* UTF-8 mode */        if (utf)
       if (utf8)  
2767          {          {
2768          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
2769            {            {
2770            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
2771              {              {
2772              SCHECK_PARTIAL();              SCHECK_PARTIAL();
2773              MRRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
2774              }              }
2775            GETCHARINC(c, eptr);            GETCHARINC(c, eptr);
2776            if (c > 255)            if (c > 255)
2777              {              {
2778              if (op == OP_CLASS) MRRETURN(MATCH_NOMATCH);              if (op == OP_CLASS) RRETURN(MATCH_NOMATCH);
2779              }              }
2780            else            else
2781              {              if ((BYTE_MAP[c/8] & (1 << (c&7))) == 0) RRETURN(MATCH_NOMATCH);
             if ((data[c/8] & (1 << (c&7))) == 0) MRRETURN(MATCH_NOMATCH);  
             }  
2782            }            }
2783          }          }
2784        else        else
2785  #endif  #endif
2786        /* Not UTF-8 mode */        /* Not UTF mode */
2787          {          {
2788          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
2789            {            {
2790            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
2791              {              {
2792              SCHECK_PARTIAL();              SCHECK_PARTIAL();
2793              MRRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
2794              }              }
2795            c = *eptr++;            c = *eptr++;
2796            if ((data[c/8] & (1 << (c&7))) == 0) MRRETURN(MATCH_NOMATCH);  #ifndef COMPILE_PCRE8
2797              if (c > 255)
2798                {
2799                if (op == OP_CLASS) RRETURN(MATCH_NOMATCH);
2800                }
2801              else
2802    #endif
2803                if ((BYTE_MAP[c/8] & (1 << (c&7))) == 0) RRETURN(MATCH_NOMATCH);
2804            }            }
2805          }          }
2806    
# Line 2627  for (;;) Line 2814  for (;;)
2814    
2815        if (minimize)        if (minimize)
2816          {          {
2817  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
2818          /* UTF-8 mode */          if (utf)
         if (utf8)  
2819            {            {
2820            for (fi = min;; fi++)            for (fi = min;; fi++)
2821              {              {
2822              RMATCH(eptr, ecode, offset_top, md, eptrb, RM16);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM16);
2823              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2824              if (fi >= max) MRRETURN(MATCH_NOMATCH);              if (fi >= max) RRETURN(MATCH_NOMATCH);
2825              if (eptr >= md->end_subject)              if (eptr >= md->end_subject)
2826                {                {
2827                SCHECK_PARTIAL();                SCHECK_PARTIAL();
2828                MRRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
2829                }                }
2830              GETCHARINC(c, eptr);              GETCHARINC(c, eptr);
2831              if (c > 255)              if (c > 255)
2832                {                {
2833                if (op == OP_CLASS) MRRETURN(MATCH_NOMATCH);                if (op == OP_CLASS) RRETURN(MATCH_NOMATCH);
2834                }                }
2835              else              else
2836                {                if ((BYTE_MAP[c/8] & (1 << (c&7))) == 0) RRETURN(MATCH_NOMATCH);
               if ((data[c/8] & (1 << (c&7))) == 0) MRRETURN(MATCH_NOMATCH);  
               }  
2837              }              }
2838            }            }
2839          else          else
2840  #endif  #endif
2841          /* Not UTF-8 mode */          /* Not UTF mode */
2842            {            {
2843            for (fi = min;; fi++)            for (fi = min;; fi++)
2844              {              {
2845              RMATCH(eptr, ecode, offset_top, md, eptrb, RM17);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM17);
2846              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
2847              if (fi >= max) MRRETURN(MATCH_NOMATCH);              if (fi >= max) RRETURN(MATCH_NOMATCH);
2848              if (eptr >= md->end_subject)              if (eptr >= md->end_subject)
2849                {                {
2850                SCHECK_PARTIAL();                SCHECK_PARTIAL();
2851                MRRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
2852                }                }
2853              c = *eptr++;              c = *eptr++;
2854              if ((data[c/8] & (1 << (c&7))) == 0) MRRETURN(MATCH_NOMATCH);  #ifndef COMPILE_PCRE8
2855                if (c > 255)
2856                  {
2857                  if (op == OP_CLASS) RRETURN(MATCH_NOMATCH);
2858                  }
2859                else
2860    #endif
2861                  if ((BYTE_MAP[c/8] & (1 << (c&7))) == 0) RRETURN(MATCH_NOMATCH);
2862              }              }
2863            }            }
2864          /* Control never gets here */          /* Control never gets here */
# Line 2679  for (;;) Line 2870  for (;;)
2870          {          {
2871          pp = eptr;          pp = eptr;
2872    
2873  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
2874          /* UTF-8 mode */          if (utf)
         if (utf8)  
2875            {            {
2876            for (i = min; i < max; i++)            for (i = min; i < max; i++)
2877              {              {
# Line 2697  for (;;) Line 2887  for (;;)
2887                if (op == OP_CLASS) break;                if (op == OP_CLASS) break;
2888                }                }
2889              else              else
2890                {                if ((BYTE_MAP[c/8] & (1 << (c&7))) == 0) break;
               if ((data[c/8] & (1 << (c&7))) == 0) break;  
               }  
2891              eptr += len;              eptr += len;
2892              }              }
2893            for (;;)            for (;;)
# Line 2712  for (;;) Line 2900  for (;;)
2900            }            }
2901          else          else
2902  #endif  #endif
2903            /* Not UTF-8 mode */            /* Not UTF mode */
2904            {            {
2905            for (i = min; i < max; i++)            for (i = min; i < max; i++)
2906              {              {
# Line 2722  for (;;) Line 2910  for (;;)
2910                break;                break;
2911                }                }
2912              c = *eptr;              c = *eptr;
2913              if ((data[c/8] & (1 << (c&7))) == 0) break;  #ifndef COMPILE_PCRE8
2914                if (c > 255)
2915                  {
2916                  if (op == OP_CLASS) break;
2917                  }
2918                else
2919    #endif
2920                  if ((BYTE_MAP[c/8] & (1 << (c&7))) == 0) break;
2921              eptr++;              eptr++;
2922              }              }
2923            while (eptr >= pp)            while (eptr >= pp)
# Line 2733  for (;;) Line 2928  for (;;)
2928              }              }
2929            }            }
2930    
2931          MRRETURN(MATCH_NOMATCH);          RRETURN(MATCH_NOMATCH);
2932          }          }
2933    #undef BYTE_MAP
2934        }        }
2935      /* Control never gets here */      /* Control never gets here */
2936    
# Line 2743  for (;;) Line 2939  for (;;)
2939      when UTF-8 mode mode is supported. Nevertheless, we may not be in UTF-8      when UTF-8 mode mode is supported. Nevertheless, we may not be in UTF-8
2940      mode, because Unicode properties are supported in non-UTF-8 mode. */      mode, because Unicode properties are supported in non-UTF-8 mode. */
2941    
2942  #ifdef SUPPORT_UTF8  #if defined SUPPORT_UTF || !defined COMPILE_PCRE8
2943      case OP_XCLASS:      case OP_XCLASS:
2944        {        {
2945        data = ecode + 1 + LINK_SIZE;                /* Save for matching */        data = ecode + 1 + LINK_SIZE;                /* Save for matching */
# Line 2768  for (;;) Line 2964  for (;;)
2964          case OP_CRMINRANGE:          case OP_CRMINRANGE:
2965          minimize = (*ecode == OP_CRMINRANGE);          minimize = (*ecode == OP_CRMINRANGE);
2966          min = GET2(ecode, 1);          min = GET2(ecode, 1);
2967          max = GET2(ecode, 3);          max = GET2(ecode, 1 + IMM2_SIZE);
2968          if (max == 0) max = INT_MAX;          if (max == 0) max = INT_MAX;
2969          ecode += 5;          ecode += 1 + 2 * IMM2_SIZE;
2970          break;          break;
2971    
2972          default:               /* No repeat follows */          default:               /* No repeat follows */
# Line 2785  for (;;) Line 2981  for (;;)
2981          if (eptr >= md->end_subject)          if (eptr >= md->end_subject)
2982            {            {
2983            SCHECK_PARTIAL();            SCHECK_PARTIAL();
2984            MRRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
2985            }            }
2986          GETCHARINCTEST(c, eptr);          GETCHARINCTEST(c, eptr);
2987          if (!_pcre_xclass(c, data)) MRRETURN(MATCH_NOMATCH);          if (!PRIV(xclass)(c, data, utf)) RRETURN(MATCH_NOMATCH);
2988          }          }
2989    
2990        /* If max == min we can continue with the main loop without the        /* If max == min we can continue with the main loop without the
# Line 2805  for (;;) Line 3001  for (;;)
3001            {            {
3002            RMATCH(eptr, ecode, offset_top, md, eptrb, RM20);            RMATCH(eptr, ecode, offset_top, md, eptrb, RM20);
3003            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3004            if (fi >= max) MRRETURN(MATCH_NOMATCH);            if (fi >= max) RRETURN(MATCH_NOMATCH);
3005            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
3006              {              {
3007              SCHECK_PARTIAL();              SCHECK_PARTIAL();
3008              MRRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
3009              }              }
3010            GETCHARINCTEST(c, eptr);            GETCHARINCTEST(c, eptr);
3011            if (!_pcre_xclass(c, data)) MRRETURN(MATCH_NOMATCH);            if (!PRIV(xclass)(c, data, utf)) RRETURN(MATCH_NOMATCH);
3012            }            }
3013          /* Control never gets here */          /* Control never gets here */
3014          }          }
# Line 2830  for (;;) Line 3026  for (;;)
3026              SCHECK_PARTIAL();              SCHECK_PARTIAL();
3027              break;              break;
3028              }              }
3029    #ifdef SUPPORT_UTF
3030            GETCHARLENTEST(c, eptr, len);            GETCHARLENTEST(c, eptr, len);
3031            if (!_pcre_xclass(c, data)) break;  #else
3032              c = *eptr;
3033    #endif
3034              if (!PRIV(xclass)(c, data, utf)) break;
3035            eptr += len;            eptr += len;
3036            }            }
3037          for(;;)          for(;;)
# Line 2839  for (;;) Line 3039  for (;;)
3039            RMATCH(eptr, ecode, offset_top, md, eptrb, RM21);            RMATCH(eptr, ecode, offset_top, md, eptrb, RM21);
3040            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3041            if (eptr-- == pp) break;        /* Stop if tried at original pos */            if (eptr-- == pp) break;        /* Stop if tried at original pos */
3042            if (utf8) BACKCHAR(eptr);  #ifdef SUPPORT_UTF
3043              if (utf) BACKCHAR(eptr);
3044    #endif
3045            }            }
3046          MRRETURN(MATCH_NOMATCH);          RRETURN(MATCH_NOMATCH);
3047          }          }
3048    
3049        /* Control never gets here */        /* Control never gets here */
# Line 2851  for (;;) Line 3053  for (;;)
3053      /* Match a single character, casefully */      /* Match a single character, casefully */
3054    
3055      case OP_CHAR:      case OP_CHAR:
3056  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
3057      if (utf8)      if (utf)
3058        {        {
3059        length = 1;        length = 1;
3060        ecode++;        ecode++;
# Line 2860  for (;;) Line 3062  for (;;)
3062        if (length > md->end_subject - eptr)        if (length > md->end_subject - eptr)
3063          {          {
3064          CHECK_PARTIAL();             /* Not SCHECK_PARTIAL() */          CHECK_PARTIAL();             /* Not SCHECK_PARTIAL() */
3065          MRRETURN(MATCH_NOMATCH);          RRETURN(MATCH_NOMATCH);
3066          }          }
3067        while (length-- > 0) if (*ecode++ != *eptr++) MRRETURN(MATCH_NOMATCH);        while (length-- > 0) if (*ecode++ != *eptr++) RRETURN(MATCH_NOMATCH);
3068        }        }
3069      else      else
3070  #endif  #endif
3071        /* Not UTF mode */
     /* Non-UTF-8 mode */  
3072        {        {
3073        if (md->end_subject - eptr < 1)        if (md->end_subject - eptr < 1)
3074          {          {
3075          SCHECK_PARTIAL();            /* This one can use SCHECK_PARTIAL() */          SCHECK_PARTIAL();            /* This one can use SCHECK_PARTIAL() */
3076          MRRETURN(MATCH_NOMATCH);          RRETURN(MATCH_NOMATCH);
3077          }          }
3078        if (ecode[1] != *eptr++) MRRETURN(MATCH_NOMATCH);        if (ecode[1] != *eptr++) RRETURN(MATCH_NOMATCH);
3079        ecode += 2;        ecode += 2;
3080        }        }
3081      break;      break;
3082    
3083      /* Match a single character, caselessly */      /* Match a single character, caselessly. If we are at the end of the
3084        subject, give up immediately. */
3085    
3086        case OP_CHARI:
3087        if (eptr >= md->end_subject)
3088          {
3089          SCHECK_PARTIAL();
3090          RRETURN(MATCH_NOMATCH);
3091          }
3092    
3093      case OP_CHARI:  #ifdef SUPPORT_UTF
3094  #ifdef SUPPORT_UTF8      if (utf)
     if (utf8)  
3095        {        {
3096        length = 1;        length = 1;
3097        ecode++;        ecode++;
3098        GETCHARLEN(fc, ecode, length);        GETCHARLEN(fc, ecode, length);
3099    
       if (length > md->end_subject - eptr)  
         {  
         CHECK_PARTIAL();             /* Not SCHECK_PARTIAL() */  
         MRRETURN(MATCH_NOMATCH);  
         }  
   
3100        /* If the pattern character's value is < 128, we have only one byte, and        /* If the pattern character's value is < 128, we have only one byte, and
3101        can use the fast lookup table. */        we know that its other case must also be one byte long, so we can use the
3102          fast lookup table. We know that there is at least one byte left in the
3103          subject. */
3104    
3105        if (fc < 128)        if (fc < 128)
3106          {          {
3107          if (md->lcc[*ecode++] != md->lcc[*eptr++]) MRRETURN(MATCH_NOMATCH);          if (md->lcc[fc]
3108                != TABLE_GET(*eptr, md->lcc, *eptr)) RRETURN(MATCH_NOMATCH);
3109            ecode++;
3110            eptr++;
3111          }          }
3112    
3113        /* Otherwise we must pick up the subject character */        /* Otherwise we must pick up the subject character. Note that we cannot
3114          use the value of "length" to check for sufficient bytes left, because the
3115          other case of the character may have more or fewer bytes.  */
3116    
3117        else        else
3118          {          {
# Line 2919  for (;;) Line 3128  for (;;)
3128  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
3129            if (dc != UCD_OTHERCASE(fc))            if (dc != UCD_OTHERCASE(fc))
3130  #endif  #endif
3131              MRRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
3132            }            }
3133          }          }
3134        }        }
3135      else      else
3136  #endif   /* SUPPORT_UTF8 */  #endif   /* SUPPORT_UTF */
3137    
3138      /* Non-UTF-8 mode */      /* Not UTF mode */
3139        {        {
3140        if (md->end_subject - eptr < 1)        if (TABLE_GET(ecode[1], md->lcc, ecode[1])
3141          {            != TABLE_GET(*eptr, md->lcc, *eptr)) RRETURN(MATCH_NOMATCH);
3142          SCHECK_PARTIAL();            /* This one can use SCHECK_PARTIAL() */        eptr++;
         MRRETURN(MATCH_NOMATCH);  
         }  
       if (md->lcc[ecode[1]] != md->lcc[*eptr++]) MRRETURN(MATCH_NOMATCH);  
3143        ecode += 2;        ecode += 2;
3144        }        }
3145      break;      break;
# Line 2943  for (;;) Line 3149  for (;;)
3149      case OP_EXACT:      case OP_EXACT:
3150      case OP_EXACTI:      case OP_EXACTI:
3151      min = max = GET2(ecode, 1);      min = max = GET2(ecode, 1);
3152      ecode += 3;      ecode += 1 + IMM2_SIZE;
3153      goto REPEATCHAR;      goto REPEATCHAR;
3154    
3155      case OP_POSUPTO:      case OP_POSUPTO:
# Line 2958  for (;;) Line 3164  for (;;)
3164      min = 0;      min = 0;
3165      max = GET2(ecode, 1);      max = GET2(ecode, 1);
3166      minimize = *ecode == OP_MINUPTO || *ecode == OP_MINUPTOI;      minimize = *ecode == OP_MINUPTO || *ecode == OP_MINUPTOI;
3167      ecode += 3;      ecode += 1 + IMM2_SIZE;
3168      goto REPEATCHAR;      goto REPEATCHAR;
3169    
3170      case OP_POSSTAR:      case OP_POSSTAR:
# Line 3006  for (;;) Line 3212  for (;;)
3212      /* Common code for all repeated single-character matches. */      /* Common code for all repeated single-character matches. */
3213    
3214      REPEATCHAR:      REPEATCHAR:
3215  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
3216      if (utf8)      if (utf)
3217        {        {
3218        length = 1;        length = 1;
3219        charptr = ecode;        charptr = ecode;
# Line 3023  for (;;) Line 3229  for (;;)
3229          unsigned int othercase;          unsigned int othercase;
3230          if (op >= OP_STARI &&     /* Caseless */          if (op >= OP_STARI &&     /* Caseless */
3231              (othercase = UCD_OTHERCASE(fc)) != fc)              (othercase = UCD_OTHERCASE(fc)) != fc)
3232            oclength = _pcre_ord2utf8(othercase, occhars);            oclength = PRIV(ord2utf)(othercase, occhars);
3233          else oclength = 0;          else oclength = 0;
3234  #endif  /* SUPPORT_UCP */  #endif  /* SUPPORT_UCP */
3235    
3236          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
3237            {            {
3238            if (eptr <= md->end_subject - length &&            if (eptr <= md->end_subject - length &&
3239              memcmp(eptr, charptr, length) == 0) eptr += length;              memcmp(eptr, charptr, IN_UCHARS(length)) == 0) eptr += length;
3240  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
3241            else if (oclength > 0 &&            else if (oclength > 0 &&
3242                     eptr <= md->end_subject - oclength &&                     eptr <= md->end_subject - oclength &&
3243                     memcmp(eptr, occhars, oclength) == 0) eptr += oclength;                     memcmp(eptr, occhars, IN_UCHARS(oclength)) == 0) eptr += oclength;
3244  #endif  /* SUPPORT_UCP */  #endif  /* SUPPORT_UCP */
3245            else            else
3246              {              {
3247              CHECK_PARTIAL();              CHECK_PARTIAL();
3248              MRRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
3249              }              }
3250            }            }
3251    
# Line 3051  for (;;) Line 3257  for (;;)
3257              {              {
3258              RMATCH(eptr, ecode, offset_top, md, eptrb, RM22);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM22);
3259              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3260              if (fi >= max) MRRETURN(MATCH_NOMATCH);              if (fi >= max) RRETURN(MATCH_NOMATCH);
3261              if (eptr <= md->end_subject - length &&              if (eptr <= md->end_subject - length &&
3262                memcmp(eptr, charptr, length) == 0) eptr += length;                memcmp(eptr, charptr, IN_UCHARS(length)) == 0) eptr += length;
3263  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
3264              else if (oclength > 0 &&              else if (oclength > 0 &&
3265                       eptr <= md->end_subject - oclength &&                       eptr <= md->end_subject - oclength &&
3266                       memcmp(eptr, occhars, oclength) == 0) eptr += oclength;                       memcmp(eptr, occhars, IN_UCHARS(oclength)) == 0) eptr += oclength;
3267  #endif  /* SUPPORT_UCP */  #endif  /* SUPPORT_UCP */
3268              else              else
3269                {                {
3270                CHECK_PARTIAL();                CHECK_PARTIAL();
3271                MRRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
3272                }                }
3273              }              }
3274            /* Control never gets here */            /* Control never gets here */
# Line 3074  for (;;) Line 3280  for (;;)
3280            for (i = min; i < max; i++)            for (i = min; i < max; i++)
3281              {              {
3282              if (eptr <= md->end_subject - length &&              if (eptr <= md->end_subject - length &&
3283                  memcmp(eptr, charptr, length) == 0) eptr += length;                  memcmp(eptr, charptr, IN_UCHARS(length)) == 0) eptr += length;
3284  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
3285              else if (oclength > 0 &&              else if (oclength > 0 &&
3286                       eptr <= md->end_subject - oclength &&                       eptr <= md->end_subject - oclength &&
3287                       memcmp(eptr, occhars, oclength) == 0) eptr += oclength;                       memcmp(eptr, occhars, IN_UCHARS(oclength)) == 0) eptr += oclength;
3288  #endif  /* SUPPORT_UCP */  #endif  /* SUPPORT_UCP */
3289              else              else
3290                {                {
# Line 3093  for (;;) Line 3299  for (;;)
3299              {              {
3300              RMATCH(eptr, ecode, offset_top, md, eptrb, RM23);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM23);
3301              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3302              if (eptr == pp) { MRRETURN(MATCH_NOMATCH); }              if (eptr == pp) { RRETURN(MATCH_NOMATCH); }
3303  #ifdef SUPPORT_UCP  #ifdef SUPPORT_UCP
3304              eptr--;              eptr--;
3305              BACKCHAR(eptr);              BACKCHAR(eptr);
# Line 3110  for (;;) Line 3316  for (;;)
3316        value of fc will always be < 128. */        value of fc will always be < 128. */
3317        }        }
3318      else      else
3319  #endif  /* SUPPORT_UTF8 */  #endif  /* SUPPORT_UTF */
3320          /* When not in UTF-8 mode, load a single-byte character. */
3321      /* When not in UTF-8 mode, load a single-byte character. */        fc = *ecode++;
3322    
3323      fc = *ecode++;      /* The value of fc at this point is always one character, though we may
3324        or may not be in UTF mode. The code is duplicated for the caseless and
     /* The value of fc at this point is always less than 256, though we may or  
     may not be in UTF-8 mode. The code is duplicated for the caseless and  
3325      caseful cases, for speed, since matching characters is likely to be quite      caseful cases, for speed, since matching characters is likely to be quite
3326      common. First, ensure the minimum number of matches are present. If min =      common. First, ensure the minimum number of matches are present. If min =
3327      max, continue at the same level without recursing. Otherwise, if      max, continue at the same level without recursing. Otherwise, if
# Line 3130  for (;;) Line 3334  for (;;)
3334    
3335      if (op >= OP_STARI)  /* Caseless */      if (op >= OP_STARI)  /* Caseless */
3336        {        {
3337        fc = md->lcc[fc];  #ifdef COMPILE_PCRE8
3338          /* fc must be < 128 if UTF is enabled. */
3339          foc = md->fcc[fc];
3340    #else
3341    #ifdef SUPPORT_UTF
3342    #ifdef SUPPORT_UCP
3343          if (utf && fc > 127)
3344            foc = UCD_OTHERCASE(fc);
3345    #else
3346          if (utf && fc > 127)
3347            foc = fc;
3348    #endif /* SUPPORT_UCP */
3349          else
3350    #endif /* SUPPORT_UTF */
3351            foc = TABLE_GET(fc, md->fcc, fc);
3352    #endif /* COMPILE_PCRE8 */
3353    
3354        for (i = 1; i <= min; i++)        for (i = 1; i <= min; i++)
3355          {          {
3356          if (eptr >= md->end_subject)          if (eptr >= md->end_subject)
3357            {            {
3358            SCHECK_PARTIAL();            SCHECK_PARTIAL();
3359            MRRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
3360            }            }
3361          if (fc != md->lcc[*eptr++]) MRRETURN(MATCH_NOMATCH);          if (fc != *eptr && foc != *eptr) RRETURN(MATCH_NOMATCH);
3362            eptr++;
3363          }          }
3364        if (min == max) continue;        if (min == max) continue;
3365        if (minimize)        if (minimize)
# Line 3147  for (;;) Line 3368  for (;;)
3368            {            {
3369            RMATCH(eptr, ecode, offset_top, md, eptrb, RM24);            RMATCH(eptr, ecode, offset_top, md, eptrb, RM24);
3370            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3371            if (fi >= max) MRRETURN(MATCH_NOMATCH);            if (fi >= max) RRETURN(MATCH_NOMATCH);
3372            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
3373              {              {
3374              SCHECK_PARTIAL();              SCHECK_PARTIAL();
3375              MRRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
3376              }              }
3377            if (fc != md->lcc[*eptr++]) MRRETURN(MATCH_NOMATCH);            if (fc != *eptr && foc != *eptr) RRETURN(MATCH_NOMATCH);
3378              eptr++;
3379            }            }
3380          /* Control never gets here */          /* Control never gets here */
3381          }          }
# Line 3167  for (;;) Line 3389  for (;;)
3389              SCHECK_PARTIAL();              SCHECK_PARTIAL();
3390              break;              break;
3391              }              }
3392            if (fc != md->lcc[*eptr]) break;            if (fc != *eptr && foc != *eptr) break;
3393            eptr++;            eptr++;
3394            }            }
3395    
# Line 3179  for (;;) Line 3401  for (;;)
3401            eptr--;            eptr--;
3402            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3403            }            }
3404          MRRETURN(MATCH_NOMATCH);          RRETURN(MATCH_NOMATCH);
3405          }          }
3406        /* Control never gets here */        /* Control never gets here */
3407        }        }
# Line 3193  for (;;) Line 3415  for (;;)
3415          if (eptr >= md->end_subject)          if (eptr >= md->end_subject)
3416            {            {
3417            SCHECK_PARTIAL();            SCHECK_PARTIAL();
3418            MRRETURN(MATCH_NOMATCH);            RRETURN(MATCH_NOMATCH);
3419            }            }
3420          if (fc != *eptr++) MRRETURN(MATCH_NOMATCH);          if (fc != *eptr++) RRETURN(MATCH_NOMATCH);
3421          }          }
3422    
3423        if (min == max) continue;        if (min == max) continue;
# Line 3206  for (;;) Line 3428  for (;;)
3428            {            {
3429            RMATCH(eptr, ecode, offset_top, md, eptrb, RM26);            RMATCH(eptr, ecode, offset_top, md, eptrb, RM26);
3430            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3431            if (fi >= max) MRRETURN(MATCH_NOMATCH);            if (fi >= max) RRETURN(MATCH_NOMATCH);
3432            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
3433              {              {
3434              SCHECK_PARTIAL();              SCHECK_PARTIAL();
3435              MRRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
3436              }              }
3437            if (fc != *eptr++) MRRETURN(MATCH_NOMATCH);            if (fc != *eptr++) RRETURN(MATCH_NOMATCH);
3438            }            }
3439          /* Control never gets here */          /* Control never gets here */
3440          }          }
# Line 3237  for (;;) Line 3459  for (;;)
3459            eptr--;            eptr--;
3460            if (rrc != MATCH_NOMATCH) RRETURN(rrc);            if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3461            }            }
3462          MRRETURN(MATCH_NOMATCH);          RRETURN(MATCH_NOMATCH);
3463          }          }
3464        }        }
3465      /* Control never gets here */      /* Control never gets here */
# Line 3246  for (;;) Line 3468  for (;;)
3468      checking can be multibyte. */      checking can be multibyte. */
3469    
3470      case OP_NOT:      case OP_NOT:
3471      case OP_NOTI:      case OP_NOTI:
3472      if (eptr >= md->end_subject)      if (eptr >= md->end_subject)
3473        {        {
3474        SCHECK_PARTIAL();        SCHECK_PARTIAL();
3475        MRRETURN(MATCH_NOMATCH);        RRETURN(MATCH_NOMATCH);
3476        }        }
3477      ecode++;      ecode++;
3478      GETCHARINCTEST(c, eptr);      GETCHARINCTEST(c, eptr);
3479      if (op == OP_NOTI)         /* The caseless case */      if (op == OP_NOTI)         /* The caseless case */
3480        {        {
3481  #ifdef SUPPORT_UTF8        register int ch, och;
3482        if (c < 256)        ch = *ecode++;
3483  #endif  #ifdef COMPILE_PCRE8
3484        c = md->lcc[c];        /* ch must be < 128 if UTF is enabled. */
3485        if (md->lcc[*ecode++] == c) MRRETURN(MATCH_NOMATCH);        och = md->fcc[ch];
3486    #else
3487    #ifdef SUPPORT_UTF
3488    #ifdef SUPPORT_UCP
3489          if (utf && ch > 127)
3490            och = UCD_OTHERCASE(ch);
3491    #else
3492          if (utf && ch > 127)
3493            och = ch;
3494    #endif /* SUPPORT_UCP */
3495          else
3496    #endif /* SUPPORT_UTF */
3497            och = TABLE_GET(ch, md->fcc, ch);
3498    #endif /* COMPILE_PCRE8 */
3499          if (ch == c || och == c) RRETURN(MATCH_NOMATCH);
3500        }        }
3501      else    /* Caseful */      else    /* Caseful */
3502        {        {
3503        if (*ecode++ == c) MRRETURN(MATCH_NOMATCH);        if (*ecode++ == c) RRETURN(MATCH_NOMATCH);
3504        }        }
3505      break;      break;
3506    
# Line 3278  for (;;) Line 3514  for (;;)
3514      case OP_NOTEXACT:      case OP_NOTEXACT:
3515      case OP_NOTEXACTI:      case OP_NOTEXACTI:
3516      min = max = GET2(ecode, 1);      min = max = GET2(ecode, 1);
3517      ecode += 3;      ecode += 1 + IMM2_SIZE;
3518      goto REPEATNOTCHAR;      goto REPEATNOTCHAR;
3519    
3520      case OP_NOTUPTO:      case OP_NOTUPTO:
# Line 3288  for (;;) Line 3524  for (;;)
3524      min = 0;      min = 0;
3525      max = GET2(ecode, 1);      max = GET2(ecode, 1);
3526      minimize = *ecode == OP_NOTMINUPTO || *ecode == OP_NOTMINUPTOI;      minimize = *ecode == OP_NOTMINUPTO || *ecode == OP_NOTMINUPTOI;
3527      ecode += 3;      ecode += 1 + IMM2_SIZE;
3528      goto REPEATNOTCHAR;      goto REPEATNOTCHAR;
3529    
3530      case OP_NOTPOSSTAR:      case OP_NOTPOSSTAR:
# Line 3320  for (;;) Line 3556  for (;;)
3556      possessive = TRUE;      possessive = TRUE;
3557      min = 0;      min = 0;
3558      max = GET2(ecode, 1);      max = GET2(ecode, 1);
3559      ecode += 3;      ecode += 1 + IMM2_SIZE;
3560      goto REPEATNOTCHAR;      goto REPEATNOTCHAR;
3561    
3562      case OP_NOTSTAR:      case OP_NOTSTAR:
# Line 3359  for (;;) Line 3595  for (;;)
3595    
3596      if (op >= OP_NOTSTARI)     /* Caseless */      if (op >= OP_NOTSTARI)     /* Caseless */
3597        {        {
3598        fc = md->lcc[fc];  #ifdef COMPILE_PCRE8
3599          /* fc must be < 128 if UTF is enabled. */
3600          foc = md->fcc[fc];
3601    #else
3602    #ifdef SUPPORT_UTF
3603    #ifdef SUPPORT_UCP
3604          if (utf && fc > 127)
3605            foc = UCD_OTHERCASE(fc);
3606    #else
3607          if (utf && fc > 127)
3608            foc = fc;
3609    #endif /* SUPPORT_UCP */
3610          else
3611    #endif /* SUPPORT_UTF */
3612            foc = TABLE_GET(fc, md->fcc, fc);
3613    #endif /* COMPILE_PCRE8 */
3614    
3615  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
3616        /* UTF-8 mode */        if (utf)
       if (utf8)  
3617          {          {
3618          register unsigned int d;          register unsigned int d;
3619          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
# Line 3371  for (;;) Line 3621  for (;;)
3621            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
3622              {              {
3623              SCHECK_PARTIAL();              SCHECK_PARTIAL();
3624              MRRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
3625              }              }
3626            GETCHARINC(d, eptr);            GETCHARINC(d, eptr);
3627            if (d < 256) d = md->lcc[d];            if (fc == d || foc == d) RRETURN(MATCH_NOMATCH);
           if (fc == d) MRRETURN(MATCH_NOMATCH);  
3628            }            }
3629          }          }
3630        else        else
3631  #endif  #endif
3632          /* Not UTF mode */
       /* Not UTF-8 mode */  
3633          {          {
3634          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
3635            {            {
3636            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
3637              {              {
3638              SCHECK_PARTIAL();              SCHECK_PARTIAL();
3639              MRRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
3640              }              }
3641            if (fc == md->lcc[*eptr++]) MRRETURN(MATCH_NOMATCH);            if (fc == *eptr || foc == *eptr) RRETURN(MATCH_NOMATCH);
3642              eptr++;
3643            }            }
3644          }          }
3645    
# Line 3398  for (;;) Line 3647  for (;;)
3647    
3648        if (minimize)        if (minimize)
3649          {          {
3650  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
3651          /* UTF-8 mode */          if (utf)
         if (utf8)  
3652            {            {
3653            register unsigned int d;            register unsigned int d;
3654            for (fi = min;; fi++)            for (fi = min;; fi++)
3655              {              {
3656              RMATCH(eptr, ecode, offset_top, md, eptrb, RM28);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM28);
3657              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3658              if (fi >= max) MRRETURN(MATCH_NOMATCH);              if (fi >= max) RRETURN(MATCH_NOMATCH);
3659              if (eptr >= md->end_subject)              if (eptr >= md->end_subject)
3660                {                {
3661                SCHECK_PARTIAL();                SCHECK_PARTIAL();
3662                MRRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
3663                }                }
3664              GETCHARINC(d, eptr);              GETCHARINC(d, eptr);
3665              if (d < 256) d = md->lcc[d];              if (fc == d || foc == d) RRETURN(MATCH_NOMATCH);
             if (fc == d) MRRETURN(MATCH_NOMATCH);  
3666              }              }
3667            }            }
3668          else          else
3669  #endif  #endif
3670          /* Not UTF-8 mode */          /* Not UTF mode */
3671            {            {
3672            for (fi = min;; fi++)            for (fi = min;; fi++)
3673              {              {
3674              RMATCH(eptr, ecode, offset_top, md, eptrb, RM29);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM29);
3675              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3676              if (fi >= max) MRRETURN(MATCH_NOMATCH);              if (fi >= max) RRETURN(MATCH_NOMATCH);
3677              if (eptr >= md->end_subject)              if (eptr >= md->end_subject)
3678                {                {
3679                SCHECK_PARTIAL();                SCHECK_PARTIAL();
3680                MRRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
3681                }                }
3682              if (fc == md->lcc[*eptr++]) MRRETURN(MATCH_NOMATCH);              if (fc == *eptr || foc == *eptr) RRETURN(MATCH_NOMATCH);
3683                eptr++;
3684              }              }
3685            }            }
3686          /* Control never gets here */          /* Control never gets here */
# Line 3444  for (;;) Line 3692  for (;;)
3692          {          {
3693          pp = eptr;          pp = eptr;
3694    
3695  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
3696          /* UTF-8 mode */          if (utf)
         if (utf8)  
3697            {            {
3698            register unsigned int d;            register unsigned int d;
3699            for (i = min; i < max; i++)            for (i = min; i < max; i++)
# Line 3458  for (;;) Line 3705  for (;;)
3705                break;                break;
3706                }                }
3707              GETCHARLEN(d, eptr, len);              GETCHARLEN(d, eptr, len);
3708              if (d < 256) d = md->lcc[d];              if (fc == d || foc == d) break;
             if (fc == d) break;  
3709              eptr += len;              eptr += len;
3710              }              }
3711          if (possessive) continue;          if (possessive) continue;
# Line 3473  for (;;) Line 3719  for (;;)
3719            }            }
3720          else          else
3721  #endif  #endif
3722          /* Not UTF-8 mode */          /* Not UTF mode */
3723            {            {
3724            for (i = min; i < max; i++)            for (i = min; i < max; i++)
3725              {              {
# Line 3482  for (;;) Line 3728  for (;;)
3728                SCHECK_PARTIAL();                SCHECK_PARTIAL();
3729                break;                break;
3730                }                }
3731              if (fc == md->lcc[*eptr]) break;              if (fc == *eptr || foc == *eptr) break;
3732              eptr++;              eptr++;
3733              }              }
3734            if (possessive) continue;            if (possessive) continue;
# Line 3494  for (;;) Line 3740  for (;;)
3740              }              }
3741            }            }
3742    
3743          MRRETURN(MATCH_NOMATCH);          RRETURN(MATCH_NOMATCH);
3744          }          }
3745        /* Control never gets here */        /* Control never gets here */
3746        }        }
# Line 3503  for (;;) Line 3749  for (;;)
3749    
3750      else      else
3751        {        {
3752  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
3753        /* UTF-8 mode */        if (utf)
       if (utf8)  
3754          {          {
3755          register unsigned int d;          register unsigned int d;
3756          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
# Line 3513  for (;;) Line 3758  for (;;)
3758            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
3759              {              {
3760              SCHECK_PARTIAL();              SCHECK_PARTIAL();
3761              MRRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
3762              }              }
3763            GETCHARINC(d, eptr);            GETCHARINC(d, eptr);
3764            if (fc == d) MRRETURN(MATCH_NOMATCH);            if (fc == d) RRETURN(MATCH_NOMATCH);
3765            }            }
3766          }          }
3767        else        else
3768  #endif  #endif
3769        /* Not UTF-8 mode */        /* Not UTF mode */
3770          {          {
3771          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
3772            {            {
3773            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
3774              {              {
3775              SCHECK_PARTIAL();              SCHECK_PARTIAL();
3776              MRRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
3777              }              }
3778            if (fc == *eptr++) MRRETURN(MATCH_NOMATCH);            if (fc == *eptr++) RRETURN(MATCH_NOMATCH);
3779            }            }
3780          }          }
3781    
# Line 3538  for (;;) Line 3783  for (;;)
3783    
3784        if (minimize)        if (minimize)
3785          {          {
3786  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
3787          /* UTF-8 mode */          if (utf)
         if (utf8)  
3788            {            {
3789            register unsigned int d;            register unsigned int d;
3790            for (fi = min;; fi++)            for (fi = min;; fi++)
3791              {              {
3792              RMATCH(eptr, ecode, offset_top, md, eptrb, RM32);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM32);
3793              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3794              if (fi >= max) MRRETURN(MATCH_NOMATCH);              if (fi >= max) RRETURN(MATCH_NOMATCH);
3795              if (eptr >= md->end_subject)              if (eptr >= md->end_subject)
3796                {                {
3797                SCHECK_PARTIAL();                SCHECK_PARTIAL();
3798                MRRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
3799                }                }
3800              GETCHARINC(d, eptr);              GETCHARINC(d, eptr);
3801              if (fc == d) MRRETURN(MATCH_NOMATCH);              if (fc == d) RRETURN(MATCH_NOMATCH);
3802              }              }
3803            }            }
3804          else          else
3805  #endif  #endif
3806          /* Not UTF-8 mode */          /* Not UTF mode */
3807            {            {
3808            for (fi = min;; fi++)            for (fi = min;; fi++)
3809              {              {
3810              RMATCH(eptr, ecode, offset_top, md, eptrb, RM33);              RMATCH(eptr, ecode, offset_top, md, eptrb, RM33);
3811              if (rrc != MATCH_NOMATCH) RRETURN(rrc);              if (rrc != MATCH_NOMATCH) RRETURN(rrc);
3812              if (fi >= max) MRRETURN(MATCH_NOMATCH);              if (fi >= max) RRETURN(MATCH_NOMATCH);
3813              if (eptr >= md->end_subject)              if (eptr >= md->end_subject)
3814                {                {
3815                SCHECK_PARTIAL();                SCHECK_PARTIAL();
3816                MRRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
3817                }                }
3818              if (fc == *eptr++) MRRETURN(MATCH_NOMATCH);              if (fc == *eptr++) RRETURN(MATCH_NOMATCH);
3819              }              }
3820            }            }
3821          /* Control never gets here */          /* Control never gets here */
# Line 3583  for (;;) Line 3827  for (;;)
3827          {          {
3828          pp = eptr;          pp = eptr;
3829    
3830  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
3831          /* UTF-8 mode */          if (utf)
         if (utf8)  
3832            {            {
3833            register unsigned int d;            register unsigned int d;
3834            for (i = min; i < max; i++)            for (i = min; i < max; i++)
# Line 3611  for (;;) Line 3854  for (;;)
3854            }            }
3855          else          else
3856  #endif  #endif
3857          /* Not UTF-8 mode */          /* Not UTF mode */
3858            {            {
3859            for (i = min; i < max; i++)            for (i = min; i < max; i++)
3860              {              {
# Line 3632  for (;;) Line 3875  for (;;)
3875              }              }
3876            }            }
3877    
3878          MRRETURN(MATCH_NOMATCH);          RRETURN(MATCH_NOMATCH);
3879          }          }
3880        }        }
3881      /* Control never gets here */      /* Control never gets here */
# Line 3644  for (;;) Line 3887  for (;;)
3887      case OP_TYPEEXACT:      case OP_TYPEEXACT:
3888      min = max = GET2(ecode, 1);      min = max = GET2(ecode, 1);
3889      minimize = TRUE;      minimize = TRUE;
3890      ecode += 3;      ecode += 1 + IMM2_SIZE;
3891      goto REPEATTYPE;      goto REPEATTYPE;
3892    
3893      case OP_TYPEUPTO:      case OP_TYPEUPTO:
# Line 3652  for (;;) Line 3895  for (;;)
3895      min = 0;      min = 0;
3896      max = GET2(ecode, 1);      max = GET2(ecode, 1);
3897      minimize = *ecode == OP_TYPEMINUPTO;      minimize = *ecode == OP_TYPEMINUPTO;
3898      ecode += 3;      ecode += 1 + IMM2_SIZE;
3899      goto REPEATTYPE;      goto REPEATTYPE;
3900    
3901      case OP_TYPEPOSSTAR:      case OP_TYPEPOSSTAR:
# Line 3680  for (;;) Line 3923  for (;;)
3923      possessive = TRUE;      possessive = TRUE;
3924      min = 0;      min = 0;
3925      max = GET2(ecode, 1);      max = GET2(ecode, 1);
3926      ecode += 3;      ecode += 1 + IMM2_SIZE;
3927      goto REPEATTYPE;      goto REPEATTYPE;
3928    
3929      case OP_TYPESTAR:      case OP_TYPESTAR:
# Line 3726  for (;;) Line 3969  for (;;)
3969          switch(prop_type)          switch(prop_type)
3970            {            {
3971            case PT_ANY:            case PT_ANY:
3972            if (prop_fail_result) MRRETURN(MATCH_NOMATCH);            if (prop_fail_result) RRETURN(MATCH_NOMATCH);
3973            for (i = 1; i <= min; i++)            for (i = 1; i <= min; i++)
3974              {              {
3975              if (eptr >= md->end_subject)              if (eptr >= md->end_subject)
3976                {                {
3977                SCHECK_PARTIAL();                SCHECK_PARTIAL();
3978                MRRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
3979                }                }
3980              GETCHARINCTEST(c, eptr);              GETCHARINCTEST(c, eptr);
3981              }              }
# Line 3741  for (;;) Line 3984  for (;;)
3984            case PT_LAMP:            case PT_LAMP:
3985            for (i = 1; i <= min; i++)            for (i = 1; i <= min; i++)
3986              {              {
3987                int chartype;
3988              if (eptr >= md->end_subject)              if (eptr >= md->end_subject)
3989                {                {
3990                SCHECK_PARTIAL();                SCHECK_PARTIAL();
3991                MRRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
3992                }                }
3993              GETCHARINCTEST(c, eptr);              GETCHARINCTEST(c, eptr);
3994              prop_chartype = UCD_CHARTYPE(c);              chartype = UCD_CHARTYPE(c);
3995              if ((prop_chartype == ucp_Lu ||              if ((chartype == ucp_Lu ||
3996                   prop_chartype == ucp_Ll ||                   chartype == ucp_Ll ||
3997                   prop_chartype == ucp_Lt) == prop_fail_result)                   chartype == ucp_Lt) == prop_fail_result)
3998                MRRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
3999              }              }
4000            break;            break;
4001    
# Line 3761  for (;;) Line 4005  for (;;)
4005              if (eptr >= md->end_subject)              if (eptr >= md->end_subject)
4006                {                {
4007                SCHECK_PARTIAL();                SCHECK_PARTIAL();
4008                MRRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
4009                }                }
4010              GETCHARINCTEST(c, eptr);              GETCHARINCTEST(c, eptr);
4011              prop_category = UCD_CATEGORY(c);              if ((UCD_CATEGORY(c) == prop_value) == prop_fail_result)
4012              if ((prop_category == prop_value) == prop_fail_result)                RRETURN(MATCH_NOMATCH);
               MRRETURN(MATCH_NOMATCH);  
4013              }              }
4014            break;            break;
4015    
# Line 3776  for (;;) Line 4019  for (;;)
4019              if (eptr >= md->end_subject)              if (eptr >= md->end_subject)
4020                {                {
4021                SCHECK_PARTIAL();                SCHECK_PARTIAL();
4022                MRRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
4023                }                }
4024              GETCHARINCTEST(c, eptr);              GETCHARINCTEST(c, eptr);
4025              prop_chartype = UCD_CHARTYPE(c);              if ((UCD_CHARTYPE(c) == prop_value) == prop_fail_result)
4026              if ((prop_chartype == prop_value) == prop_fail_result)                RRETURN(MATCH_NOMATCH);
               MRRETURN(MATCH_NOMATCH);  
4027              }              }
4028            break;            break;
4029    
# Line 3791  for (;;) Line 4033  for (;;)
4033              if (eptr >= md->end_subject)              if (eptr >= md->end_subject)
4034                {                {
4035                SCHECK_PARTIAL();                SCHECK_PARTIAL();
4036                MRRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
4037                }                }
4038              GETCHARINCTEST(c, eptr);              GETCHARINCTEST(c, eptr);
4039              prop_script = UCD_SCRIPT(c);              if ((UCD_SCRIPT(c) == prop_value) == prop_fail_result)
4040              if ((prop_script == prop_value) == prop_fail_result)                RRETURN(MATCH_NOMATCH);
               MRRETURN(MATCH_NOMATCH);  
4041              }              }
4042            break;            break;
4043    
4044            case PT_ALNUM:            case PT_ALNUM:
4045            for (i = 1; i <= min; i++)            for (i = 1; i <= min; i++)
4046              {              {
4047                int category;
4048              if (eptr >= md->end_subject)              if (eptr >= md->end_subject)
4049                {                {
4050                SCHECK_PARTIAL();                SCHECK_PARTIAL();
4051                MRRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
4052                }                }
4053              GETCHARINCTEST(c, eptr);              GETCHARINCTEST(c, eptr);
4054              prop_category = UCD_CATEGORY(c);              category = UCD_CATEGORY(c);
4055              if ((prop_category == ucp_L || prop_category == ucp_N)              if ((category == ucp_L || category == ucp_N) == prop_fail_result)
4056                     == prop_fail_result)                RRETURN(MATCH_NOMATCH);
               MRRETURN(MATCH_NOMATCH);  
4057              }              }
4058            break;            break;
4059    
# Line 3822  for (;;) Line 4063  for (;;)
4063              if (eptr >= md->end_subject)              if (eptr >= md->end_subject)
4064                {                {
4065                SCHECK_PARTIAL();                SCHECK_PARTIAL();
4066                MRRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
4067                }                }
4068              GETCHARINCTEST(c, eptr);              GETCHARINCTEST(c, eptr);
4069              prop_category = UCD_CATEGORY(c);              if ((UCD_CATEGORY(c) == ucp_Z || c == CHAR_HT || c == CHAR_NL ||
             if ((prop_category == ucp_Z || c == CHAR_HT || c == CHAR_NL ||  
4070                   c == CHAR_FF || c == CHAR_CR)                   c == CHAR_FF || c == CHAR_CR)
4071                     == prop_fail_result)                     == prop_fail_result)
4072                MRRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
4073              }              }
4074            break;            break;
4075    
# Line 3839  for (;;) Line 4079  for (;;)
4079              if (eptr >= md->end_subject)              if (eptr >= md->end_subject)
4080                {                {
4081                SCHECK_PARTIAL();                SCHECK_PARTIAL();
4082                MRRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
4083                }                }
4084              GETCHARINCTEST(c, eptr);              GETCHARINCTEST(c, eptr);
4085              prop_category = UCD_CATEGORY(c);              if ((UCD_CATEGORY(c) == ucp_Z || c == CHAR_HT || c == CHAR_NL ||
             if ((prop_category == ucp_Z || c == CHAR_HT || c == CHAR_NL ||  
4086                   c == CHAR_VT || c == CHAR_FF || c == CHAR_CR)                   c == CHAR_VT || c == CHAR_FF || c == CHAR_CR)
4087                     == prop_fail_result)                     == prop_fail_result)
4088                MRRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
4089              }              }
4090            break;            break;
4091    
4092            case PT_WORD:            case PT_WORD:
4093            for (i = 1; i <= min; i++)            for (i = 1; i <= min; i++)
4094              {              {
4095                int category;
4096              if (eptr >= md->end_subject)              if (eptr >= md->end_subject)
4097                {                {
4098                SCHECK_PARTIAL();                SCHECK_PARTIAL();
4099                MRRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
4100                }                }
4101              GETCHARINCTEST(c, eptr);              GETCHARINCTEST(c, eptr);
4102              prop_category = UCD_CATEGORY(c);              category = UCD_CATEGORY(c);
4103              if ((prop_category == ucp_L || prop_category == ucp_N ||              if ((category == ucp_L || category == ucp_N || c == CHAR_UNDERSCORE)
                  c == CHAR_UNDERSCORE)  
4104                     == prop_fail_result)                     == prop_fail_result)
4105                MRRETURN(MATCH_NOMATCH);                RRETURN(MATCH_NOMATCH);
4106              }              }
4107            break;            break;
4108    
# Line 3884  for (;;) Line 4123  for (;;)
4123            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
4124              {              {
4125              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4126              MRRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4127              }              }
4128            GETCHARINCTEST(c, eptr);            GETCHARINCTEST(c, eptr);
4129            prop_category = UCD_CATEGORY(c);            if (UCD_CATEGORY(c) == ucp_M) RRETURN(MATCH_NOMATCH);
           if (prop_category == ucp_M) MRRETURN(MATCH_NOMATCH);  
4130            while (eptr < md->end_subject)            while (eptr < md->end_subject)
4131              {              {
4132              int len = 1;              int len = 1;
4133              if (!utf8) c = *eptr;              if (!utf) c = *eptr; else { GETCHARLEN(c, eptr, len); }
4134                else { GETCHARLEN(c, eptr, len); }              if (UCD_CATEGORY(c) != ucp_M) break;
             prop_category = UCD_CATEGORY(c);  
             if (prop_category != ucp_M) break;  
4135              eptr += len;              eptr += len;
4136              }              }
4137            }            }
# Line 3906  for (;;) Line 4142  for (;;)
4142    
4143  /* Handle all other cases when the coding is UTF-8 */  /* Handle all other cases when the coding is UTF-8 */
4144    
4145  #ifdef SUPPORT_UTF8  #ifdef SUPPORT_UTF
4146        if (utf8) switch(ctype)        if (utf) switch(ctype)
4147          {          {
4148          case OP_ANY:          case OP_ANY:
4149          for (i = 1; i <= min; i++)          for (i = 1; i <= min; i++)
# Line 3915  for (;;) Line 4151  for (;;)
4151            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
4152              {              {
4153              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4154              MRRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4155              }              }
4156            if (IS_NEWLINE(eptr)) MRRETURN(MATCH_NOMATCH);            if (IS_NEWLINE(eptr)) RRETURN(MATCH_NOMATCH);
4157            eptr++;            eptr++;
4158            while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;            ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);
4159            }            }
4160          break;          break;
4161    
# Line 3929  for (;;) Line 4165  for (;;)
4165            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
4166              {              {
4167              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4168              MRRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4169              }              }
4170            eptr++;            eptr++;
4171            while (eptr < md->end_subject && (*eptr & 0xc0) == 0x80) eptr++;            ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);
4172            }            }
4173          break;          break;
4174    
4175          case OP_ANYBYTE:          case OP_ANYBYTE:
4176          if (eptr > md->end_subject - min) MRRETURN(MATCH_NOMATCH);          if (eptr > md->end_subject - min) RRETURN(MATCH_NOMATCH);
4177          eptr += min;          eptr += min;
4178          break;          break;
4179    
# Line 3947  for (;;) Line 4183  for (;;)
4183            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
4184              {              {
4185              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4186              MRRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4187              }              }
4188            GETCHARINC(c, eptr);            GETCHARINC(c, eptr);
4189            switch(c)            switch(c)
4190              {              {
4191              default: MRRETURN(MATCH_NOMATCH);              default: RRETURN(MATCH_NOMATCH);
4192    
4193              case 0x000d:              case 0x000d:
4194              if (eptr < md->end_subject && *eptr == 0x0a) eptr++;              if (eptr < md->end_subject && *eptr == 0x0a) eptr++;
4195              break;              break;
# Line 3966  for (;;) Line 4202  for (;;)
4202              case 0x0085:              case 0x0085:
4203              case 0x2028:              case 0x2028:
4204              case 0x2029:              case 0x2029:
4205              if (md->bsr_anycrlf) MRRETURN(MATCH_NOMATCH);              if (md->bsr_anycrlf) RRETURN(MATCH_NOMATCH);
4206              break;              break;
4207              }              }
4208            }            }
# Line 3978  for (;;) Line 4214  for (;;)
4214            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
4215              {              {
4216              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4217              MRRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4218              }              }
4219            GETCHARINC(c, eptr);            GETCHARINC(c, eptr);
4220            switch(c)            switch(c)
# Line 4003  for (;;) Line 4239  for (;;)
4239              case 0x202f:    /* NARROW NO-BREAK SPACE */              case 0x202f:    /* NARROW NO-BREAK SPACE */
4240              case 0x205f:    /* MEDIUM MATHEMATICAL SPACE */              case 0x205f:    /* MEDIUM MATHEMATICAL SPACE */
4241              case 0x3000:    /* IDEOGRAPHIC SPACE */              case 0x3000:    /* IDEOGRAPHIC SPACE */
4242              MRRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4243              }              }
4244            }            }
4245          break;          break;
# Line 4014  for (;;) Line 4250  for (;;)
4250            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
4251              {              {
4252              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4253              MRRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4254              }              }
4255            GETCHARINC(c, eptr);            GETCHARINC(c, eptr);
4256            switch(c)            switch(c)
4257              {              {
4258              default: MRRETURN(MATCH_NOMATCH);              default: RRETURN(MATCH_NOMATCH);
4259              case 0x09:      /* HT */              case 0x09:      /* HT */
4260              case 0x20:      /* SPACE */              case 0x20:      /* SPACE */
4261              case 0xa0:      /* NBSP */              case 0xa0:      /* NBSP */
# Line 4050  for (;;) Line 4286  for (;;)
4286            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
4287              {              {
4288              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4289              MRRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4290              }              }
4291            GETCHARINC(c, eptr);            GETCHARINC(c, eptr);
4292            switch(c)            switch(c)
# Line 4063  for (;;) Line 4299  for (;;)
4299              case 0x85:      /* NEL */              case 0x85:      /* NEL */
4300              case 0x2028:    /* LINE SEPARATOR */              case 0x2028:    /* LINE SEPARATOR */
4301              case 0x2029:    /* PARAGRAPH SEPARATOR */              case 0x2029:    /* PARAGRAPH SEPARATOR */
4302              MRRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4303              }              }
4304            }            }
4305          break;          break;
# Line 4074  for (;;) Line 4310  for (;;)
4310            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
4311              {              {
4312              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4313              MRRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4314              }              }
4315            GETCHARINC(c, eptr);            GETCHARINC(c, eptr);
4316            switch(c)            switch(c)
4317              {              {
4318              default: MRRETURN(MATCH_NOMATCH);              default: RRETURN(MATCH_NOMATCH);
4319              case 0x0a:      /* LF */              case 0x0a:      /* LF */
4320              case 0x0b:      /* VT */              case 0x0b:      /* VT */
4321              case 0x0c:      /* FF */              case 0x0c:      /* FF */
# Line 4098  for (;;) Line 4334  for (;;)
4334            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
4335              {              {
4336              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4337              MRRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4338              }              }
4339            GETCHARINC(c, eptr);            GETCHARINC(c, eptr);
4340            if (c < 128 && (md->ctypes[c] & ctype_digit) != 0)            if (c < 128 && (md->ctypes[c] & ctype_digit) != 0)
4341              MRRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4342            }            }
4343          break;          break;
4344    
# Line 4112  for (;;) Line 4348  for (;;)
4348            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
4349              {              {
4350              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4351              MRRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4352              }              }
4353            if (*eptr >= 128 || (md->ctypes[*eptr++] & ctype_digit) == 0)            if (*eptr >= 128 || (md->ctypes[*eptr] & ctype_digit) == 0)
4354              MRRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4355              eptr++;
4356            /* No need to skip more bytes - we know it's a 1-byte character */            /* No need to skip more bytes - we know it's a 1-byte character */
4357            }            }
4358          break;          break;
# Line 4126  for (;;) Line 4363  for (;;)
4363            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
4364              {              {
4365              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4366              MRRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4367              }              }
4368            if (*eptr < 128 && (md->ctypes[*eptr] & ctype_space) != 0)            if (*eptr < 128 && (md->ctypes[*eptr] & ctype_space) != 0)
4369              MRRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4370            while (++eptr < md->end_subject && (*eptr & 0xc0) == 0x80);            eptr++;
4371              ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);
4372            }            }
4373          break;          break;
4374    
# Line 4140  for (;;) Line 4378  for (;;)
4378            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
4379              {              {
4380              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4381              MRRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4382              }              }
4383            if (*eptr >= 128 || (md->ctypes[*eptr++] & ctype_space) == 0)            if (*eptr >= 128 || (md->ctypes[*eptr] & ctype_space) == 0)
4384              MRRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4385              eptr++;
4386            /* No need to skip more bytes - we know it's a 1-byte character */            /* No need to skip more bytes - we know it's a 1-byte character */
4387            }            }
4388          break;          break;
# Line 4154  for (;;) Line 4393  for (;;)
4393            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
4394              {              {
4395              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4396              MRRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4397              }              }
4398            if (*eptr < 128 && (md->ctypes[*eptr] & ctype_word) != 0)            if (*eptr < 128 && (md->ctypes[*eptr] & ctype_word) != 0)
4399              MRRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4400            while (++eptr < md->end_subject && (*eptr & 0xc0) == 0x80);            eptr++;
4401              ACROSSCHAR(eptr < md->end_subject, *eptr, eptr++);
4402            }            }
4403          break;          break;
4404    
# Line 4168  for (;;) Line 4408  for (;;)
4408            if (eptr >= md->end_subject)            if (eptr >= md->end_subject)
4409              {              {
4410              SCHECK_PARTIAL();              SCHECK_PARTIAL();
4411              MRRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4412              }              }
4413            if (*eptr >= 128 || (md->ctypes[*eptr++] & ctype_word) == 0)            if (*eptr >= 128 || (md->ctypes[*eptr] & ctype_word) == 0)
4414              MRRETURN(MATCH_NOMATCH);              RRETURN(MATCH_NOMATCH);
4415              eptr++;
4416            /* No need to skip more bytes - we know it's a 1-byte character */            /* No need to skip more bytes - we know it's a 1-byte character */
4417            }            }
4418          break;          break;
# Line 4181  for (;;) Line 4422  for (;;)
4422          }  /* End switch(ctype) */          }  /* End switch(ctype) */
4423    
4424        else        else
4425  #endif     /* SUPPORT_UTF8 */  #endif     /* SUPPORT_UTF */
4426    
4427        /* Code for the non-UTF-8 case for minimum matching of operators other        /* Code for the non-UTF-8 case for minimum matching of operators other
4428        than OP_PROP and OP_NOTPROP. */        than OP_PROP and OP_NOTPROP. */
# Line 4194  for (;;) Line 4435  for (;;)
4435